Merge branch 'master' into sverk/hipe-line-table-bug/master/OTP-13282

290 files changed, 23059 insertions, 27265 deletions

diff --git a/erts/aclocal.m4 b/erts/aclocal.m4
index 01541aff72..f6d8f20e4e 100644
--- a/erts/aclocal.m4
+++ b/erts/aclocal.m4
@@ -74,21 +74,6 @@ AC_ARG_VAR(erl_xcomp_clock_gettime_cpu_time, [clock_gettime() can be used for re
 AC_ARG_VAR(erl_xcomp_after_morecore_hook, [__after_morecore_hook can track malloc()s core memory usage: yes|no (only used when cross compiling)])
 AC_ARG_VAR(erl_xcomp_dlsym_brk_wrappers, [dlsym(RTLD_NEXT, _) brk wrappers can track malloc()s core memory usage: yes|no (only used when cross compiling)])
 
-dnl Cross compilation variables for OSE
-AC_ARG_VAR(erl_xcomp_ose_ldflags_pass1, [Linker flags for the OSE module (pass 1) (only used when cross compiling for OSE)])
-AC_ARG_VAR(erl_xcomp_ose_ldflags_pass2, [Linker flags for the OSE module (pass 2) (only used when cross compiling for OSE)])
-AC_ARG_VAR(erl_xcomp_ose_OSEROOT, [OSE installation root directory (only used when cross compiling for OSE)])
-AC_ARG_VAR(erl_xcomp_ose_STRIP, [Strip utility shipped with the OSE distribution(only used when cross compiling for OSE)])
-AC_ARG_VAR(erl_xcomp_ose_LM_POST_LINK, [OSE postlink tool (only used when cross compiling for OSE)])
-AC_ARG_VAR(erl_xcomp_ose_LM_SET_CONF, [Sets the configuration for an OSE load module (only used when cross compiling for OSE)])
-AC_ARG_VAR(erl_xcomp_ose_LM_ELF_SIZE, [Prints the section size information for an OSE load module (only used when cross compiling for OSE)])
-AC_ARG_VAR(erl_xcomp_ose_LM_LCF, [OSE load module linker configuration file (only used when cross compiling for OSE)])
-AC_ARG_VAR(erl_xcomp_ose_BEAM_LM_CONF, [BEAM OSE load module default configuration file (only used when cross compiling for OSE)])
-AC_ARG_VAR(erl_xcomp_ose_EPMD_LM_CONF, [EPMD OSE load module default configuration file (only used when cross compiling for OSE)])
-AC_ARG_VAR(erl_xcomp_ose_RUN_ERL_LM_CONF, [run_erl_lm OSE load module default configuration file (only used when cross compiling for OSE)])
-AC_ARG_VAR(erl_xcomp_ose_CONFD, [OSE confd source file])
-AC_ARG_VAR(erl_xcomp_ose_CRT0_LM, [OSE crt0 lm source file])
-
 ])
 
 AC_DEFUN(ERL_XCOMP_SYSROOT_INIT,
@@ -142,18 +127,18 @@ MIXED_MSYS=no
 
 AC_MSG_CHECKING(for mixed cygwin or msys and native VC++ environment)
 if test "X$host" = "Xwin32" -a "x$GCC" != "xyes"; then
-	if test -x /usr/bin/cygpath; then
-		CFLAGS="-O2"
-		MIXED_CYGWIN=yes
-		AC_MSG_RESULT([Cygwin and VC])
-		MIXED_CYGWIN_VC=yes
-		CPPFLAGS="$CPPFLAGS -DERTS_MIXED_CYGWIN_VC"
-	elif test -x /usr/bin/msysinfo; then
+	if test -x /usr/bin/msys-?.0.dll; then
 	        CFLAGS="-O2"
 		MIXED_MSYS=yes
 		AC_MSG_RESULT([MSYS and VC])
 		MIXED_MSYS_VC=yes
 		CPPFLAGS="$CPPFLAGS -DERTS_MIXED_MSYS_VC"
+	elif test -x /usr/bin/cygpath; then
+		CFLAGS="-O2"
+		MIXED_CYGWIN=yes
+		AC_MSG_RESULT([Cygwin and VC])
+		MIXED_CYGWIN_VC=yes
+		CPPFLAGS="$CPPFLAGS -DERTS_MIXED_CYGWIN_VC"
 	else		    
 		AC_MSG_RESULT([undeterminable])
 		AC_MSG_ERROR(Seems to be mixed windows but not with cygwin, cannot handle this!)
@@ -503,8 +488,6 @@ AC_CACHE_VAL(ac_cv_sys_ipv6_support,
 #ifdef __WIN32__
 #include <winsock2.h>
 #include <ws2tcpip.h>
-#elif __OSE__
-#error "no ipv6"
 #else
 #include <netinet/in.h>
 #endif],
@@ -517,8 +500,6 @@ else
 #ifdef __WIN32__
 #include <winsock2.h>
 #include <ws2tcpip.h>
-#elif __OSE__
-#error "no ipv6"
 #else
 #include <netinet/in.h>
 #endif],
@@ -726,9 +707,15 @@ esac
 
 AC_DEFUN(ERL_MONOTONIC_CLOCK,
 [
-  default_resolution_clock_gettime_monotonic="CLOCK_HIGHRES CLOCK_BOOTTIME CLOCK_MONOTONIC"
-  low_resolution_clock_gettime_monotonic="CLOCK_MONOTONIC_COARSE CLOCK_MONOTONIC_FAST"
-  high_resolution_clock_gettime_monotonic="CLOCK_MONOTONIC_PRECISE"
+  if test "$3" = "yes"; then
+     default_resolution_clock_gettime_monotonic="CLOCK_HIGHRES CLOCK_BOOTTIME CLOCK_MONOTONIC"
+     low_resolution_clock_gettime_monotonic="CLOCK_MONOTONIC_COARSE CLOCK_MONOTONIC_FAST"
+     high_resolution_clock_gettime_monotonic="CLOCK_MONOTONIC_PRECISE"
+  else
+     default_resolution_clock_gettime_monotonic="CLOCK_HIGHRES CLOCK_UPTIME CLOCK_MONOTONIC"
+     low_resolution_clock_gettime_monotonic="CLOCK_MONOTONIC_COARSE CLOCK_UPTIME_FAST"
+     high_resolution_clock_gettime_monotonic="CLOCK_UPTIME_PRECISE"
+  fi
 
   case "$1" in
     high_resolution)
@@ -744,8 +731,8 @@ AC_DEFUN(ERL_MONOTONIC_CLOCK,
 	prefer_resolution_clock_gettime_monotonic="$2"
 	;;
     *)
-	check_msg=""
-	prefer_resolution_clock_gettime_monotonic=
+	check_msg="custom "
+	prefer_resolution_clock_gettime_monotonic="$2"
 	;;
   esac
 
@@ -934,16 +921,16 @@ AC_DEFUN(ERL_WALL_CLOCK,
 
   erl_wall_clock_low_resolution=no
   erl_wall_clock_id=
-  case $erl_cv_clock_gettime_wall_$1-$erl_cv_mach_clock_get_time_wall-$ac_cv_func_gettimeofday-$host_os in
-    *-*-*-win32)
+  case $1-$erl_cv_clock_gettime_wall_$1-$erl_cv_mach_clock_get_time_wall-$ac_cv_func_gettimeofday-$host_os in
+    *-*-*-*-win32)
       erl_wall_clock_func=WindowsAPI
       erl_wall_clock_low_resolution=yes
       ;;
-    no-yes-*-*)
+    high_resolution-no-yes-*-*)
       erl_wall_clock_func=mach_clock_get_time
       erl_wall_clock_id=CALENDAR_CLOCK
       ;;
-    CLOCK_*-*-*-*)
+    *-CLOCK_*-*-*-*)
       erl_wall_clock_func=clock_gettime
       erl_wall_clock_id=$erl_cv_clock_gettime_wall_$1
       for low_res_id in $low_resolution_clock_gettime_wall; do
@@ -953,7 +940,7 @@ AC_DEFUN(ERL_WALL_CLOCK,
 	  fi
       done
       ;;
-    no-no-yes-*)
+    *-no-*-yes-*)
       erl_wall_clock_func=gettimeofday
       ;;
     *)
@@ -985,12 +972,6 @@ if test "X$host_os" = "Xwin32"; then
     THR_LIBS=
     THR_LIB_NAME=win32_threads
     THR_LIB_TYPE=win32_threads
-elif test "X$host_os" = "Xose"; then
-    AC_MSG_RESULT(yes)
-    THR_DEFS="-DOSE_THREADS"
-    THR_LIBS=
-    THR_LIB_NAME=ose_threads
-    THR_LIB_TYPE=ose_threads
 else
     AC_MSG_RESULT(no)
     THR_DEFS=
@@ -1466,7 +1447,7 @@ AC_ARG_WITH(with_sparc_memory_order,
 LM_CHECK_THR_LIB
 ERL_INTERNAL_LIBS
 
-ERL_MONOTONIC_CLOCK(high_resolution)
+ERL_MONOTONIC_CLOCK(try_find_pthread_compatible, CLOCK_HIGHRES CLOCK_MONOTONIC, no)
 
 case $erl_monotonic_clock_func in
   clock_gettime)
@@ -1577,22 +1558,9 @@ case "$THR_LIB_NAME" in
 	fi
 	;;
 
-    pthread|ose_threads)
-        case "$THR_LIB_NAME" in
-	     pthread)
-		ETHR_THR_LIB_BASE_DIR=pthread
-		AC_DEFINE(ETHR_PTHREADS, 1, [Define if you have pthreads])
-		;;
-	     ose_threads)
-		AC_DEFINE(ETHR_OSE_THREADS, 1,
-		   [Define if you have OSE style threads])
-		ETHR_THR_LIB_BASE_DIR=ose
-		AC_CHECK_HEADER(ose_spi/ose_spi.h,
-		  AC_DEFINE(HAVE_OSE_SPI_H, 1,
-		    [Define if you have the "ose_spi/ose_spi.h" header file.]))
-		;;
-	esac
-	if test "x$THR_LIB_NAME" = "xpthread"; then
+    pthread)
+	ETHR_THR_LIB_BASE_DIR=pthread
+	AC_DEFINE(ETHR_PTHREADS, 1, [Define if you have pthreads])
 	case $host_os in
 	    openbsd*)
 		# The default stack size is insufficient for our needs
@@ -1651,7 +1619,6 @@ case "$THR_LIB_NAME" in
 	    *) ;;
 	esac
 
-	fi
 	dnl We sometimes need ETHR_DEFS in order to find certain headers
 	dnl (at least for pthread.h on osf1).
 	saved_cppflags="$CPPFLAGS"
@@ -1696,7 +1663,6 @@ case "$THR_LIB_NAME" in
 	dnl
 	dnl Check for functions
 	dnl
-	if test "x$THR_LIB_NAME" = "xpthread"; then
 	AC_CHECK_FUNC(pthread_spin_lock, \
 			[ethr_have_native_spinlock=yes \
 			 AC_DEFINE(ETHR_HAVE_PTHREAD_SPIN_LOCK, 1, \
@@ -1900,7 +1866,7 @@ case "$THR_LIB_NAME" in
                      #define _DARWIN_C_SOURCE
                      #include <pthread.h>],
                     [char buff[256]; pthread_getname_np(pthread_self(), buff, 256);],
-                    pthread_getname=normal)
+                    pthread_getname=linux)
 	AC_TRY_LINK([#define __USE_GNU
                      #define _DARWIN_C_SOURCE
                      #include <pthread.h>],
@@ -1916,8 +1882,6 @@ case "$THR_LIB_NAME" in
 	esac
 	CFLAGS=$old_CFLAGS
 
-        fi ## test "x$THR_LIB_NAME" = "xpthread"
-
 	if test "X$disable_native_ethr_impls" = "Xyes"; then
 	    ethr_have_native_atomics=no
 	else
@@ -2202,7 +2166,7 @@ AC_DEFUN(ERL_TIME_CORRECTION,
 
 AC_ARG_WITH(clock-resolution,
 AS_HELP_STRING([--with-clock-resolution=high|low|default],
-               [specify wanted clock resolution)]))
+               [specify wanted clock resolution]))
 
 AC_ARG_WITH(clock-gettime-realtime-id,
 AS_HELP_STRING([--with-clock-gettime-realtime-id=CLOCKID],
@@ -2212,6 +2176,24 @@ AC_ARG_WITH(clock-gettime-monotonic-id,
 AS_HELP_STRING([--with-clock-gettime-monotonic-id=CLOCKID],
                [specify clock id to use with clock_gettime() for monotonic time)]))
 
+AC_ARG_ENABLE(prefer-elapsed-monotonic-time-during-suspend,
+AS_HELP_STRING([--enable-prefer-elapsed-monotonic-time-during-suspend],
+               [Prefer an OS monotonic time source with elapsed time during suspend])
+AS_HELP_STRING([--disable-prefer-elapsed-monotonic-time-during-suspend],
+               [Do not prefer an OS monotonic time source with elapsed time during suspend]),
+[ case "$enableval" in
+    yes) prefer_elapsed_monotonic_time_during_suspend=yes ;;
+    *)  prefer_elapsed_monotonic_time_during_suspend=no ;;
+  esac ], prefer_elapsed_monotonic_time_during_suspend=no)
+
+AC_ARG_ENABLE(gettimeofday-as-os-system-time,
+	      AS_HELP_STRING([--enable-gettimeofday-as-os-system-time],
+                             [Force usage of gettimeofday() for OS system time]),
+[ case "$enableval" in
+    yes) force_gettimeofday_os_system_time=yes ;;
+    *)  force_gettimeofday_os_system_time=no ;;
+  esac ], force_gettimeofday_os_system_time=no)
+
 case "$with_clock_resolution" in
    ""|no|yes)
      with_clock_resolution=default;;
@@ -2222,6 +2204,17 @@ case "$with_clock_resolution" in
      ;;
 esac
 
+if test "$force_gettimeofday_os_system_time" = "yes"; then
+
+  AC_CHECK_FUNCS([gettimeofday])
+  if test "$ac_cv_func_gettimeofday" = "yes"; then
+    AC_DEFINE(OS_SYSTEM_TIME_GETTIMEOFDAY,  [1], [Define if you want to implement erts_os_system_time() using gettimeofday()])
+  else
+    AC_MSG_ERROR([No gettimeofday() available])
+  fi
+
+else # $force_gettimeofday_os_system_time != yes
+
 case "$with_clock_gettime_realtime_id" in
    ""|no)
      with_clock_gettime_realtime_id=no
@@ -2239,23 +2232,6 @@ case "$with_clock_gettime_realtime_id" in
      ;;
 esac
 
-case "$with_clock_gettime_monotonic_id" in
-   ""|no)
-     with_clock_gettime_monotonic_id=no
-     ;;
-   CLOCK_*CPUTIME*)
-     AC_MSG_ERROR([Invalid clock_gettime() monotonic clock id: Refusing to use the cputime clock id $with_clock_gettime_monotonic_id as monotonic clock id])
-     ;;
-   CLOCK_REALTIME*|CLOCK_TAI*)
-     AC_MSG_ERROR([Invalid clock_gettime() monotonic clock id: Refusing to use the realtime clock id $with_clock_gettime_monotonic_id as monotonic clock id])
-     ;;
-   CLOCK_*)
-     ;;
-   *)
-     AC_MSG_ERROR([Invalid clock_gettime() clock id: $with_clock_gettime_monotonic_id])
-     ;;
-esac
-
 case "$with_clock_resolution-$with_clock_gettime_realtime_id" in
   high-no)
 	ERL_WALL_CLOCK(high_resolution);;
@@ -2278,6 +2254,9 @@ case "$erl_wall_clock_func-$erl_wall_clock_id-$with_clock_gettime_realtime_id" i
 esac
 
 case $erl_wall_clock_func in
+  none)
+    AC_MSG_ERROR([No wall clock source found])
+    ;;
   mach_clock_get_time)
     AC_DEFINE(OS_SYSTEM_TIME_USING_MACH_CLOCK_GET_TIME, [1], [Define if you want to implement erts_os_system_time() using mach clock_get_time()])
     ;;
@@ -2296,15 +2275,34 @@ if test "x$erl_wall_clock_id" != "x"; then
     AC_DEFINE_UNQUOTED(WALL_CLOCK_ID, [$erl_wall_clock_id], [Define to wall clock id to use])
 fi
 
+fi # $force_gettimeofday_os_system_time != yes
+
+case "$with_clock_gettime_monotonic_id" in
+   ""|no)
+     with_clock_gettime_monotonic_id=no
+     ;;
+   CLOCK_*CPUTIME*)
+     AC_MSG_ERROR([Invalid clock_gettime() monotonic clock id: Refusing to use the cputime clock id $with_clock_gettime_monotonic_id as monotonic clock id])
+     ;;
+   CLOCK_REALTIME*|CLOCK_TAI*)
+     AC_MSG_ERROR([Invalid clock_gettime() monotonic clock id: Refusing to use the realtime clock id $with_clock_gettime_monotonic_id as monotonic clock id])
+     ;;
+   CLOCK_*)
+     ;;
+   *)
+     AC_MSG_ERROR([Invalid clock_gettime() clock id: $with_clock_gettime_monotonic_id])
+     ;;
+esac
+
 case "$with_clock_resolution-$with_clock_gettime_monotonic_id" in
   high-no)
-	ERL_MONOTONIC_CLOCK(high_resolution);;
+	ERL_MONOTONIC_CLOCK(high_resolution, undefined, $prefer_elapsed_monotonic_time_during_suspend);;
   low-no)
-	ERL_MONOTONIC_CLOCK(low_resolution);;
+	ERL_MONOTONIC_CLOCK(low_resolution, undefined, $prefer_elapsed_monotonic_time_during_suspend);;
   default-no)
-	ERL_MONOTONIC_CLOCK(default_resolution);;
+	ERL_MONOTONIC_CLOCK(default_resolution, undefined, $prefer_elapsed_monotonic_time_during_suspend);;
   *)
-	ERL_MONOTONIC_CLOCK(custom_resolution, $with_clock_gettime_monotonic_id);;
+	ERL_MONOTONIC_CLOCK(custom_resolution, $with_clock_gettime_monotonic_id, $prefer_elapsed_monotonic_time_during_suspend);;
 esac
 
 case "$erl_monotonic_clock_func-$erl_monotonic_clock_id-$with_clock_gettime_monotonic_id" in
@@ -2352,7 +2350,7 @@ if test $erl_cv_clock_gettime_monotonic_raw = yes; then
   AC_DEFINE(HAVE_CLOCK_GETTIME_MONOTONIC_RAW, [1], [Define if you have clock_gettime(CLOCK_MONOTONIC_RAW, _)])
 fi
 
-ERL_MONOTONIC_CLOCK(high_resolution)
+ERL_MONOTONIC_CLOCK(high_resolution, undefined, no)
 
 case $$erl_monotonic_clock_low_resolution-$erl_monotonic_clock_func in
   no-mach_clock_get_time)
diff --git a/erts/configure.in b/erts/configure.in
index 22ca7ec17d..2f19c0f760 100644
--- a/erts/configure.in
+++ b/erts/configure.in
@@ -106,7 +106,6 @@ AC_CONFIG_HEADER($host/config.h:config.h.in include/internal/$host/ethread_heade
 dnl ----------------------------------------------------------------------
 dnl Optional features.
 dnl ----------------------------------------------------------------------
-enable_child_waiter_thread=no
 ENABLE_ALLOC_TYPE_VARS=
 AC_SUBST(ENABLE_ALLOC_TYPE_VARS)
 
@@ -143,14 +142,6 @@ AS_HELP_STRING([--enable-dirty-schedulers], [enable dirty scheduler support]),
     *)  enable_dirty_schedulers=yes ;;
   esac ], enable_dirty_schedulers=no)
 
-AC_ARG_ENABLE(halfword-emulator,
-AS_HELP_STRING([--enable-halfword-emulator],
-               [enable halfword emulator (only for 64bit builds). Note: Halfword emulator is marked as deprecated and scheduled for removal in future major release.]),
-[ case "$enableval" in
-    no) enable_halfword_emualtor=no ;;
-    *)  enable_halfword_emulator=yes ;;
-  esac ], enable_halfword_emulator=unknown)
-
 AC_ARG_ENABLE(smp-support,
 AS_HELP_STRING([--enable-smp-support], [enable smp support])
 AS_HELP_STRING([--disable-smp-support], [disable smp support]),
@@ -342,6 +333,15 @@ AS_HELP_STRING([--enable-systemd], [enable systemd support in epmd]),
 [],
 [enable_systemd=no])
 
+AC_ARG_ENABLE(saved-compile-time,
+AS_HELP_STRING([--disable-saved-compile-time], [disable saved compile time]),
+[ case "$enableval" in
+    no) save_compile_time=0 ;;
+    *)  save_compile_time=1 ;;
+  esac ], save_compile_time=1)
+
+AC_DEFINE_UNQUOTED(ERTS_SAVED_COMPILE_TIME, $save_compile_time, [Save compile time?])
+
 dnl Magic test for clearcase.
 OTP_RELEASE=
 if test "${ERLANG_COMMERCIAL_BUILD}" != ""; then
@@ -787,36 +787,18 @@ esac
 	
 AC_SUBST(LIBCARBON)
 
-dnl Check if we should/can build a halfword emulator
-
-AC_MSG_CHECKING(if we are building a halfword emulator (32bit heap on 64bit machine))
-if test "$enable_halfword_emulator" = "yes"; then
-	if test "$ARCH" = "amd64"; then
-		AC_DEFINE(HALFWORD_HEAP_EMULATOR, [1],
-			  [Define if building a halfword-heap 64bit emulator])
-		ENABLE_ALLOC_TYPE_VARS="$ENABLE_ALLOC_TYPE_VARS halfword"
-		AC_MSG_RESULT([yes])
-
-		test -f "$ERL_TOP/erts/CONF_INFO" ||
-		    echo "" > "$ERL_TOP/erts/CONF_INFO"
-		cat >> $ERL_TOP/erts/CONF_INFO <<EOF
-
-                 The HALFWORD emulator has been enabled.
-                 This is a DEPRECATED feature scheduled for removal
-                 in a future major release.
+dnl Check if we should/can build a sharing-preserving emulator
 
-EOF
-	else
-		AC_MSG_ERROR(no; halfword emulator not supported on this architecture)
-	fi
+AC_MSG_CHECKING(if we are building a sharing-preserving emulator)
+if test "$enable_sharing_preserving" = "yes"; then
+        AC_DEFINE(SHCOPY, [1],
+		  [Define if building a sharing-preserving emulator])
+	AC_MSG_RESULT([yes])
 else
 	AC_MSG_RESULT([no])
 fi
 
 
-
-
-
 dnl some tests below will call this if we haven't already - and autoconf
 dnl can't handle those tests being done conditionally at runtime
 AC_PROG_CPP
@@ -935,10 +917,7 @@ dnl what the user say. This might not be the right way to do it, but
 dnl for now that is the way we do it.
 USER_LD=$LD
 USER_LDFLAGS="$LDFLAGS"
-case $host in
-   *ose) ;;
-   *) LD='$(CC)' ;;
-esac
+LD='$(CC)'
 AC_SUBST(LD)
 
 LDFLAG_RUNTIME_LIBRARY_PATH="$CFLAG_RUNTIME_LIBRARY_PATH"
@@ -953,8 +932,6 @@ dnl This is the os flavour, should be unix, ose, vxworks or win32
 case $host in
    win32)
       ERLANG_OSTYPE=win32 ;;
-   *ose)
-      ERLANG_OSTYPE=ose ;;
    *)
       ERLANG_OSTYPE=unix ;;
 esac
@@ -1263,7 +1240,7 @@ case "$enable_threads"-"$found_threads" in
 	AC_MSG_RESULT(yes; enabled by user) ;;
     unknown-yes)
 	case $host_os in
-	    solaris*|linux*|darwin*|win32|ose)
+	    solaris*|linux*|darwin*|win32)
 		emu_threads=yes
 		AC_MSG_RESULT(yes; default on this platform)
 		;;
@@ -1325,11 +1302,7 @@ else
 	    AC_MSG_RESULT(no)
 	fi
 
-	disable_child_waiter_thread=no
 	case $host_os in
-	    solaris*)
-		enable_child_waiter_thread=yes
-		;;
 	    linux*)
 		AC_MSG_CHECKING([whether dlopen() needs to be called before first call to dlerror()])
 		if test "x$ETHR_THR_LIB_BASE_TYPE" != "xposix_nptl"; then
@@ -1339,16 +1312,6 @@ else
 		else
 			AC_MSG_RESULT(no)
 		fi
-		if test "x$ETHR_THR_LIB_BASE_TYPE" != "xposix_nptl"; then
-		    # Child waiter thread cannot be enabled
-		    disable_child_waiter_thread=yes
-		    enable_child_waiter_thread=no
-		fi
-		;;
-	    win32|ose)
-		# Child waiter thread cannot be enabled
-		disable_child_waiter_thread=yes
-		enable_child_waiter_thread=no
 		;;
 	    *)
 		;;
@@ -1368,24 +1331,6 @@ else
 	    esac
 	done
 	EMU_THR_DEFS=$new_emu_thr_defs
-
-	AC_MSG_CHECKING(whether the child waiter thread should be enabled)
-	if test $enable_child_waiter_thread = yes; then
-	    AC_DEFINE(ENABLE_CHILD_WAITER_THREAD,[1],
-		[Define if you want to enable child waiter thread])
-	    AC_MSG_RESULT(yes)
-	else
-	    case $ERTS_BUILD_SMP_EMU-$disable_child_waiter_thread in
-		yes-no)
-		    AC_MSG_RESULT([yes on SMP build, but not on non-SMP build]);;
-		*-yes)
-	    	    AC_DEFINE(DISABLE_CHILD_WAITER_THREAD,[1],
-			[Define if you want to disable child waiter thread])
-		    AC_MSG_RESULT(no);;
-		*)
-		    AC_MSG_RESULT(no);;
-	    esac
-	fi
 fi
 
 AC_SUBST(EMU_THR_LIB_NAME)
@@ -1517,19 +1462,27 @@ dnl
 #	   if -lsocket doesn't work by itself.
 #--------------------------------------------------------------------
 
+tk_oldLibs=$LIBS
 erl_checkBoth=0
+SOCKET_LIBS=""
 AC_CHECK_FUNC(connect, erl_checkSocket=0, erl_checkSocket=1)
 if test "$erl_checkSocket" = 1; then
-    AC_CHECK_LIB(socket, main, LIBS="$LIBS -lsocket", erl_checkBoth=1)
+    AC_CHECK_LIB(socket, main, SOCKET_LIBS="-lsocket", erl_checkBoth=1)
 fi
+
 if test "$erl_checkBoth" = 1; then
-    tk_oldLibs=$LIBS
     LIBS="$LIBS -lsocket -lnsl"
-    AC_CHECK_FUNC(accept, erl_checkNsl=0, [LIBS=$tk_oldLibs])
+    AC_CHECK_FUNC(accept, SOCKET_LIBS="-lsocket -lnsl")
 fi
-AC_CHECK_FUNC(gethostbyname, , AC_CHECK_LIB(nsl, main, [LIBS="$LIBS -lnsl"]))
+
+LIBS="$tk_oldLibs $SOCKET_LIBS"
+AC_CHECK_FUNC(gethostbyname, , AC_CHECK_LIB(nsl, main, [SOCKET_LIBS="$SOCKET_LIBS -lnsl"]))
 AC_CHECK_FUNC(gethostbyname_r,have_gethostbyname_r=yes)
 
+LIBS="$tk_oldLibs $SOCKET_LIBS"
+
+AC_SUBST(SOCKET_LIBS)
+
 dnl
 dnl These gethostbyname thingies use old style AC_DEFINE for BC with ancient 
 dnl autoconf...
@@ -2109,7 +2062,7 @@ AC_CHECK_FUNCS([getipnodebyname getipnodebyaddr gethostbyname2])
 AC_CHECK_FUNCS([ieee_handler fpsetmask finite isnan isinf res_gethostbyname dlopen \
 		pread pwrite memmove strerror strerror_r strncasecmp \
 		gethrtime localtime_r gmtime_r inet_pton \
-		memcpy mallopt sbrk _sbrk __sbrk brk _brk __brk \
+		mmap mremap memcpy mallopt sbrk _sbrk __sbrk brk _brk __brk \
 		flockfile fstat strlcpy strlcat setsid posix2time time2posix \
 		setlocale nl_langinfo poll mlockall ppoll])
 
@@ -2127,8 +2080,7 @@ fi
 
 case X$erl_xcomp_posix_memalign in
      Xno) ;;
-     Xyes) AC_DEFINE(HAVE_POSIX_MEMALIGN,[1],
-                [Define to 1 if you have the `posix_memalign' function.]) ;;
+     Xyes) have_posix_memalign=yes ;;
      *)
 	AC_CHECK_FUNC(
 		[posix_memalign],
@@ -2143,15 +2095,19 @@ int main(void) {
       return error;
     return 0;
 }
-],AC_DEFINE(HAVE_POSIX_MEMALIGN,[1],
-    [Define to 1 if you have the `posix_memalign' function.])
+],have_posix_memalign=yes
 )
 		 else
-			AC_DEFINE(HAVE_POSIX_MEMALIGN,[1],
-		[Define to 1 if you have the `posix_memalign' function.])
+			have_posix_memalign=yes
 		 fi]);;
 esac
 
+if test $have_posix_memalign = yes; then
+   AC_DEFINE(HAVE_POSIX_MEMALIGN,[1],
+	     [Define to 1 if you have the `posix_memalign' function.])
+fi
+
+
 dnl writev on OS X snow leopard is broken for files > 4GB
 case $host_os in
         darwin10.8.0)
@@ -2161,17 +2117,6 @@ case $host_os in
 	    AC_CHECK_FUNCS([writev]) ;;
 esac
 
-case $host_os in
-        *ose)
-	    AC_MSG_CHECKING([for mmap])
-	    AC_MSG_RESULT(not using for OSE)
-	    AC_MSG_CHECKING([for mremap])
-	    AC_MSG_RESULT(not using for OSE) ;;
-        *)
-	    AC_CHECK_FUNCS([mmap mremap]) ;;
-esac
-
-
 AC_CHECK_DECLS([posix2time, time2posix],,,[#include <time.h>])
 
 disable_vfork=false
@@ -2820,17 +2765,26 @@ LM_SYS_IPV6
 LM_SYS_MULTICAST
 ERL_TIME_CORRECTION
 AC_CHECK_PROG(M4, m4, m4)
+
+
+dnl HiPE cannot run on 64-bit without MAP_FIXED and MAP_NORESERVE
+if test X${enable_hipe} != Xno && test X$ac_cv_sizeof_void_p != X4; then
+   AC_CHECK_DECLS([MAP_FIXED, MAP_NORESERVE], [], [], [#include <sys/mman.h>])
+   if test X$ac_cv_have_decl_MAP_FIXED != Xyes || test X$ac_cv_have_decl_MAP_NORESERVE != Xyes; then
+      if test X${enable_hipe} = Xyes; then
+	 AC_MSG_ERROR([HiPE on 64-bit needs MAP_FIXED and MAP_NORESERVE flags for mmap()])
+      else
+	 enable_hipe=no
+	 AC_MSG_WARN([Disable HiPE due to lack of MAP_FIXED and MAP_NORESERVE flags for mmap()])
+      fi
+   fi
+fi
+
 dnl check to auto-enable hipe here...
 if test "$cross_compiling" != "yes" && test X${enable_hipe} != Xno; then
   if test -z "$M4"; then
 	enable_hipe=no
    	AC_MSG_NOTICE([HiPE disabled as no valid m4 is found in PATH])
-  elif test "$enable_halfword_emulator" = "yes"; then
-       if test X${enable_hipe} = Xyes; then
-              AC_MSG_ERROR([HiPE can not be combined with halfword emulator (yet)])
-       else		    
-   	      AC_MSG_NOTICE([HiPE auto-disabled on halfword emulator])
-       fi	
   else
   	case "$ARCH-$OPSYS" in
 	    x86-linux|amd64-linux|x86-darwin*|amd64-darwin*|ppc-linux|ppc64-linux|ppc-darwin|arm-linux|amd64-freebsd|x86-freebsd|x86-sol2|amd64-sol2|ultrasparc-linux)
@@ -3759,7 +3713,7 @@ dnl crypto
 #
 #--------------------------------------------------------------------
 
-DED_SYS_INCLUDE="-I${ERL_TOP}/erts/emulator/beam -I${ERL_TOP}/erts/include -I${ERL_TOP}/erts/include/$host -I${ERL_TOP}/erts/include/internal -I${ERL_TOP}/erts/include/internal/$host -I${ERL_TOP}/erts/emulator/sys/$ERLANG_OSTYPE"
+DED_SYS_INCLUDE="-I${ERL_TOP}/erts/emulator/beam -I${ERL_TOP}/erts/include -I${ERL_TOP}/erts/include/$host -I${ERL_TOP}/erts/include/internal -I${ERL_TOP}/erts/include/internal/$host -I${ERL_TOP}/erts/emulator/sys/$ERLANG_OSTYPE -I${ERL_TOP}/erts/emulator/sys/common"
 
 if test "X$ETHR_DEFS" = "X"; then
     DED_THR_DEFS="-D_THREAD_SAFE -D_REENTRANT"
@@ -4931,7 +4885,6 @@ AC_OUTPUT(
   Makefile:Makefile.in
   ../make/$host/otp.mk:../make/otp.mk.in
   ../make/$host/otp_ded.mk:../make/otp_ded.mk.in
-  ../make/$host/ose_lm.mk:../make/ose_lm.mk.in
 dnl
 dnl The ones below should be moved to their respective lib
 dnl
diff --git a/erts/doc/src/absform.xml b/erts/doc/src/absform.xml
index 547d5e583d..ccdecf44ec 100644
--- a/erts/doc/src/absform.xml
+++ b/erts/doc/src/absform.xml
@@ -4,14 +4,14 @@
 <chapter>
   <header>
     <copyright>
-      <year>2001</year><year>2015</year>
+      <year>2001</year><year>2016</year>
       <holder>Ericsson AB. All Rights Reserved.</holder>
     </copyright>
     <legalnotice>
       Licensed under the Apache License, Version 2.0 (the "License");
       you may not use this file except in compliance with the License.
       You may obtain a copy of the License at
- 
+
           http://www.apache.org/licenses/LICENSE-2.0
 
       Unless required by applicable law or agreed to in writing, software
@@ -19,7 +19,7 @@
       WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
       See the License for the specific language governing permissions and
       limitations under the License.
-    
+
     </legalnotice>
 
     <title>The Abstract Format</title>
@@ -35,24 +35,24 @@
   <p></p>
   <p>This document describes the standard representation of parse trees for Erlang
     programs as Erlang terms. This representation is known as the <em>abstract format</em>.
-    Functions dealing with such parse trees are <c><![CDATA[compile:forms/[1,2]]]></c>
+    Functions dealing with such parse trees are <c>compile:forms/[1,2]</c>
     and functions in the modules
-    <c><![CDATA[epp]]></c>,
-    <c><![CDATA[erl_eval]]></c>,
-    <c><![CDATA[erl_lint]]></c>,
-    <c><![CDATA[erl_pp]]></c>,
-    <c><![CDATA[erl_parse]]></c>,
+    <c>epp</c>,
+    <c>erl_eval</c>,
+    <c>erl_lint</c>,
+    <c>erl_pp</c>,
+    <c>erl_parse</c>,
     and
-    <c><![CDATA[io]]></c>.
+    <c>io</c>.
     They are also used as input and output for parse transforms (see the module
-    <c><![CDATA[compile]]></c>).</p>
-  <p>We use the function <c><![CDATA[Rep]]></c> to denote the mapping from an Erlang source
-    construct <c><![CDATA[C]]></c> to its abstract format representation <c><![CDATA[R]]></c>, and write
-    <c><![CDATA[R = Rep(C)]]></c>.
+    <c>compile</c>).</p>
+  <p>We use the function <c>Rep</c> to denote the mapping from an Erlang source
+    construct <c>C</c> to its abstract format representation <c>R</c>, and write
+    <c>R = Rep(C)</c>.
     </p>
-  <p>The word <c><![CDATA[LINE]]></c> below represents an integer, and denotes the
+  <p>The word <c>LINE</c> below represents an integer, and denotes the
     number of the line in the source file where the construction occurred.
-    Several instances of <c><![CDATA[LINE]]></c> in the same construction may denote
+    Several instances of <c>LINE</c> in the same construction may denote
     different lines.</p>
   <p>Since operators are not terms in their own right, when operators are
     mentioned below, the representation of an operator should be taken to
@@ -61,229 +61,114 @@
     </p>
 
   <section>
-    <title>Module declarations and forms</title>
+    <title>Module Declarations and Forms</title>
     <p>A module declaration consists of a sequence of forms that are either
       function declarations or attributes.</p>
     <list type="bulleted">
       <item>If D is a module declaration consisting of the forms
-      <c><![CDATA[F_1]]></c>, ..., <c><![CDATA[F_k]]></c>, then
-       Rep(D) = <c><![CDATA[[Rep(F_1), ..., Rep(F_k)]]]></c>.</item>
-      <item>If F is an attribute <c><![CDATA[-module(Mod)]]></c>, then
-       Rep(F) = <c><![CDATA[{attribute,LINE,module,Mod}]]></c>.</item>
-      <item>If F is an attribute <c><![CDATA[-export([Fun_1/A_1, ..., Fun_k/A_k])]]></c>, then
-       Rep(F) = <c><![CDATA[{attribute,LINE,export,[{Fun_1,A_1}, ..., {Fun_k,A_k}]}]]></c>.</item>
-      <item>If F is an attribute <c><![CDATA[-import(Mod,[Fun_1/A_1, ..., Fun_k/A_k])]]></c>, then
-       Rep(F) = <c><![CDATA[{attribute,LINE,import,{Mod,[{Fun_1,A_1}, ..., {Fun_k,A_k}]}}]]></c>.</item>
-      <item>If F is an attribute <c><![CDATA[-compile(Options)]]></c>, then
-       Rep(F) = <c><![CDATA[{attribute,LINE,compile,Options}]]></c>.</item>
-      <item>If F is an attribute <c><![CDATA[-file(File,Line)]]></c>, then
-       Rep(F) = <c><![CDATA[{attribute,LINE,file,{File,Line}}]]></c>.</item>
-      <item>If F is a record declaration <c><![CDATA[-record(Name,{V_1, ..., V_k})]]></c>, then
-       Rep(F) =
-      <c><![CDATA[{attribute,LINE,record,{Name,[Rep(V_1), ..., Rep(V_k)]}}]]></c>. For Rep(V), see below.</item>
-      <item>If F is a type attribute (i.e. <c><![CDATA[opaque]]></c> or
-       <c><![CDATA[type]]></c>)
-       <c><![CDATA[-Attr Name(A_1, ..., A_k) :: T]]></c> where each
-       <c><![CDATA[A_i]]></c> is a variable, then Rep(F) =
-       <c><![CDATA[{attribute,LINE,Attr,{Name,Rep(T),[Rep(A_1), ..., Rep(A_k)]}}]]></c>.
-       For Rep(T), see below.</item>
-      <item>If F is a type spec (i.e. <c><![CDATA[callback]]></c> or
-       <c><![CDATA[spec]]></c>)
-       <c><![CDATA[-Attr F Tc_1; ...; Tc_k]]></c>,
-       where each <c><![CDATA[Tc_i]]></c> is a fun type clause with an
-       argument sequence of the same length <c><![CDATA[Arity]]></c>, then
-       Rep(F) =
-       <c><![CDATA[{Attr,LINE,{{F,Arity},[Rep(Tc_1), ..., Rep(Tc_k)]}}]]></c>.
-       For Rep(Tc_i), see below.</item>
-      <item>If F is a type spec (i.e. <c><![CDATA[callback]]></c> or
-       <c><![CDATA[spec]]></c>)
-       <c><![CDATA[-Attr Mod:F Tc_1; ...; Tc_k]]></c>,
-       where each <c><![CDATA[Tc_i]]></c> is a fun type clause with an
-       argument sequence of the same length <c><![CDATA[Arity]]></c>, then
-       Rep(F) =
-       <c><![CDATA[{Attr,LINE,{{Mod,F,Arity},[Rep(Tc_1), ..., Rep(Tc_k)]}}]]></c>.
-       For Rep(Tc_i), see below.</item>
-      <item>If F is a wild attribute <c><![CDATA[-A(T)]]></c>, then
-       Rep(F) = <c><![CDATA[{attribute,LINE,A,T}]]></c>.
+      <c>F_1</c>, ..., <c>F_k</c>, then
+       Rep(D) = <c>[Rep(F_1), ..., Rep(F_k)]</c>.</item>
+      <item>If F is an attribute <c>-behavior(Behavior)</c>, then
+       Rep(F) = <c>{attribute,LINE,behavior,Behavior}</c>.</item>
+      <item>If F is an attribute <c>-behaviour(Behaviour)</c>, then
+       Rep(F) = <c>{attribute,LINE,behaviour,Behaviour}</c>.</item>
+      <item>If F is an attribute <c>-compile(Options)</c>, then
+       Rep(F) = <c>{attribute,LINE,compile,Options}</c>.</item>
+      <item>If F is an attribute <c>-export([Fun_1/A_1, ..., Fun_k/A_k])</c>, then
+       Rep(F) = <c>{attribute,LINE,export,[{Fun_1,A_1}, ..., {Fun_k,A_k}]}</c>.</item>
+      <item>If F is an attribute <c>-export_type([Type_1/A_1, ..., Type_k/A_k])</c>, then
+       Rep(F) = <c>{attribute,LINE,export_type,[{Type_1,A_1}, ..., {Type_k,A_k}]}</c>.</item>
+      <item>If F is an attribute <c>-import(Mod,[Fun_1/A_1, ..., Fun_k/A_k])</c>, then
+       Rep(F) = <c>{attribute,LINE,import,{Mod,[{Fun_1,A_1}, ..., {Fun_k,A_k}]}}</c>.</item>
+      <item>If F is an attribute <c>-module(Mod)</c>, then
+       Rep(F) = <c>{attribute,LINE,module,Mod}</c>.</item>
+      <item>If F is an attribute <c>-optional_callbacks([Fun_1/A_1, ..., Fun_k/A_k])</c>, then
+       Rep(F) = <c>{attribute,LINE,optional_callbacks,[{Fun_1,A_1}, ..., {Fun_k,A_k}]}</c>.</item>
+      <item>If F is an attribute <c>-file(File,Line)</c>, then
+       Rep(F) = <c>{attribute,LINE,file,{File,Line}}</c>.</item>
+      <item>If F is a function declaration
+       <c>Name Fc_1 ; ... ; Name Fc_k</c>,
+       where each <c>Fc_i</c> is a function clause with a
+       pattern sequence of the same length <c>Arity</c>, then
+       Rep(F) = <c>{function,LINE,Name,Arity,[Rep(Fc_1), ...,Rep(Fc_k)]}</c>.
+      </item>
+      <item>If F is a function specification
+       <c>-Spec Name Ft_1; ...; Ft_k</c>,
+       where <c>Spec</c> is either the atom <c>spec</c> or the atom
+       <c>callback</c>, and each <c>Ft_i</c> is a possibly constrained
+       function type with an argument sequence of the same length
+       <c>Arity</c>, then Rep(F) =
+       <c>{attribute,Line,Spec,{{Name,Arity},[Rep(Ft_1), ..., Rep(Ft_k)]}}</c>.
+      </item>
+      <item>If F is a function specification
+       <c>-spec Mod:Name Ft_1; ...; Ft_k</c>,
+       where each <c>Ft_i</c> is a possibly constrained
+       function type with an argument sequence of the same length
+       <c>Arity</c>, then Rep(F) =
+       <c>{attribute,Line,spec,{{Mod,Name,Arity},[Rep(Ft_1), ..., Rep(Ft_k)]}}</c>.
+      </item>
+      <item>If F is a record declaration
+       <c>-record(Name,{V_1, ..., V_k})</c>,
+       where each <c>V_i</c> is a record field, then Rep(F) =
+       <c>{attribute,LINE,record,{Name,[Rep(V_1), ..., Rep(V_k)]}}</c>.
+       For Rep(V), see below.</item>
+      <item>If F is a type declaration
+       <c>-Type Name(V_1, ..., V_k) :: T</c>, where
+       <c>Type</c> is either the atom <c>type</c> or the atom <c>opaque</c>,
+       each <c>V_i</c> is a variable, and <c>T</c> is a type, then Rep(F) =
+       <c>{attribute,LINE,Type,{Name,Rep(T),[Rep(V_1), ..., Rep(V_k)]}}</c>.
+      </item>
+      <item>If F is a wild attribute <c>-A(T)</c>, then
+       Rep(F) = <c>{attribute,LINE,A,T}</c>.
       <br></br></item>
-      <item>If F is a function declaration <c><![CDATA[Name Fc_1 ; ... ; Name Fc_k]]></c>, 
-       where each <c><![CDATA[Fc_i]]></c> is a function clause with a
-       pattern sequence of the same length <c><![CDATA[Arity]]></c>, then
-       Rep(F) = <c><![CDATA[{function,LINE,Name,Arity,[Rep(Fc_1), ...,Rep(Fc_k)]}]]></c>.</item>
     </list>
 
     <section>
-      <title>Type clauses</title>
-      <list type="bulleted">
-       <item>If T is a fun type clause
-        <c><![CDATA[(A_1, ..., A_n) -> Ret]]></c>, where each
-        <c><![CDATA[A_i]]></c> and <c><![CDATA[Ret]]></c> are types, then
-        Rep(T) =
-        <c><![CDATA[{type,LINE,'fun',[{type,LINE,product,[Rep(A_1), ..., Rep(A_n)]},Rep(Ret)]}]]></c>.
-        </item>
-       <item>If T is a bounded fun type clause <c><![CDATA[Tc when Tg]]></c>,
-        where <c><![CDATA[Tc]]></c> is an unbounded fun type clause and
-        <c><![CDATA[Tg]]></c> is a type guard sequence, then Rep(T) =
-        <c><![CDATA[{type,LINE,bounded_fun,[Rep(Tc),Rep(Tg)]}]]></c>.</item>
-      </list>
-    </section>
-
-    <section>
-      <title>Type guards</title>
-      <list type="bulleted">
-       <item>If G is a constraint <c><![CDATA[F(A_1, ..., A_k)]]></c>, where
-        <c><![CDATA[F]]></c> is an atom and each <c><![CDATA[A_i]]></c> is a
-        type, then Rep(G) =
-        <c><![CDATA[{type,LINE,constraint,[Rep(F),[Rep(A_1), ..., Rep(A_k)]]}]]></c>.
-        </item>
-       <item>If G is a type definition <c><![CDATA[Name :: Type]]></c>,
-        where <c><![CDATA[Name]]></c> is a variable and
-        <c><![CDATA[Type]]></c> is a type, then Rep(G) =
-        <c><![CDATA[{type,LINE,constraint,[{atom,LINE,is_subtype},[Rep(Name),Rep(Type)]]}]]></c>.</item>
-      </list>
-    </section>
-
-    <section>
-      <title>Types</title>
-      <list type="bulleted">
-       <item>If T is a type definition <c><![CDATA[Name :: Type]]></c>,
-        where <c><![CDATA[Name]]></c> is a variable and
-        <c><![CDATA[Type]]></c> is a type, then Rep(T) =
-        <c><![CDATA[{ann_type,LINE,[Rep(Name),Rep(Type)]}]]></c>.</item>
-       <item>If T is a type union <c><![CDATA[A_1 | ... | A_k]]></c>,
-        where each <c><![CDATA[A_i]]></c> is a type, then Rep(T) =
-        <c><![CDATA[{type,LINE,union,[Rep(A_1), ..., Rep(A_k)]}]]></c>.</item>
-       <item>If T is a type range <c><![CDATA[L .. R]]></c>,
-        where <c><![CDATA[L]]></c> and <c><![CDATA[R]]></c> are types, then
-        Rep(T) = <c><![CDATA[{type,LINE,range,[Rep(L), Rep(R)]}]]></c>.</item>
-       <item>If T is a binary operation <c><![CDATA[L Op R]]></c>,
-        where <c><![CDATA[Op]]></c> is an arithmetic or bitwise binary operator
-        and <c><![CDATA[L]]></c> and <c><![CDATA[R]]></c> are types, then
-        Rep(T) = <c><![CDATA[{op,LINE,Op,Rep(L),Rep(R)}]]></c>.</item>
-       <item>If T is <c><![CDATA[Op A]]></c>, where <c><![CDATA[Op]]></c> is an
-        arithmetic or bitwise unary operator and <c><![CDATA[A]]></c> is a
-        type, then Rep(T) = <c><![CDATA[{op,LINE,Op,Rep(A)}]]></c>.</item>
-       <item>If T is a fun type <c><![CDATA[fun()]]></c>, then Rep(T) =
-        <c><![CDATA[{type,LINE,'fun',[]}]]></c>.</item>
-       <item>If T is a variable <c><![CDATA[V]]></c>, then Rep(T) =
-        <c><![CDATA[{var,LINE,A}]]></c>, where <c><![CDATA[A]]></c> is an atom
-        with a printname consisting of the same characters as
-        <c><![CDATA[V]]></c>.</item>
-       <item>If T is an atomic literal L and L is not a string literal, then
-        Rep(T) = Rep(L).</item>
-       <item>If T is a tuple or map type <c><![CDATA[F()]]></c> (i.e.
-        <c><![CDATA[tuple]]></c> or <c><![CDATA[map]]></c>), then Rep(T) =
-        <c><![CDATA[{type,LINE,F,any}]]></c>.</item>
-       <item>If T is a type <c><![CDATA[F(A_1, ..., A_k)]]></c>, where each
-        <c><![CDATA[A_i]]></c> is a type, then Rep(T) =
-        <c><![CDATA[{user_type,LINE,F,[Rep(A_1), ..., Rep(A_k)]}]]></c>.</item>
-       <item>If T is a remote type <c><![CDATA[M:F(A_1, ..., A_k)]]></c>, where
-        each <c><![CDATA[A_i]]></c> is a type and <c><![CDATA[M]]></c> and
-        <c><![CDATA[F]]></c>, then Rep(T) =
-        <c><![CDATA[{remote_type,LINE,[Rep(M),Rep(F),[Rep(A_1), ..., Rep(A_k)]]}]]></c>.
-        </item>
-       <item>If T is the nil type <c><![CDATA[[]]]></c>, then Rep(T) =
-        <c><![CDATA[{type,LINE,nil,[]}]]></c>.</item>
-       <item>If T is a list type <c><![CDATA[[A]]]></c>, where
-        <c><![CDATA[A]]></c> is a type, then Rep(T) =
-        <c><![CDATA[{type,LINE,list,[Rep(A)]}]]></c>.</item>
-       <item>If T is a non-empty list type <c><![CDATA[[A, ...]]]></c>, where
-        <c><![CDATA[A]]></c> is a type, then Rep(T) =
-        <c><![CDATA[{type,LINE,nonempty_list,[Rep(A)]}]]></c>.</item>
-       <item>If T is a map type <c><![CDATA[#{P_1, ..., P_k}]]></c>, where each
-        <c><![CDATA[P_i]]></c> is a map pair type, then Rep(T) =
-        <c><![CDATA[{type,LINE,map,[Rep(P_1), ..., Rep(P_k)]}]]></c>.</item>
-       <item>If T is a map pair type <c><![CDATA[K => V]]></c>, where
-        <c><![CDATA[K]]></c> and <c><![CDATA[V]]></c> are types,
-        then Rep(T) =
-        <c><![CDATA[{type,LINE,map_field_assoc,[Rep(K),Rep(V)]}]]></c>.</item>
-       <item>If T is a tuple type <c><![CDATA[{A_1, ..., A_k}]]></c>, where
-        each <c><![CDATA[A_i]]></c> is a type, then Rep(T) =
-        <c><![CDATA[{type,LINE,tuple,[Rep(A_1), ..., Rep(A_k)]}]]></c>.</item>
-       <item>If T is a record type <c><![CDATA[#Name{}]]></c>, where
-        <c><![CDATA[Name]]></c> is an atom, then Rep(T) =
-        <c><![CDATA[{type,LINE,record,[Rep(Name)]}]]></c>.</item>
-       <item>If T is a record type <c><![CDATA[#Name{F_1, ..., F_k}]]></c>,
-        where <c><![CDATA[Name]]></c> is an atom, then Rep(T) =
-        <c><![CDATA[{type,LINE,record,[Rep(Name),[Rep(F_1), ..., Rep(F_k)]]}]]></c>.
-        </item>
-       <item>If T is a record field type <c><![CDATA[Name :: Type]]></c>,
-        where <c><![CDATA[Name]]></c> is an atom, then Rep(T) =
-        <c><![CDATA[{type,LINE,field_type,[Rep(Name),Rep(Type)]}]]></c>.</item>
-       <item>If T is a record field type <c><![CDATA[<<>>]]></c>, then Rep(T) =
-        <c><![CDATA[{type,LINE,binary,[{integer,LINE,0},{integer,LINE,0}]}]]></c>.
-        </item>
-       <item>If T is a binary type <c><![CDATA[<< _ : B >>]]></c>, where
-        <c><![CDATA[B]]></c> is a type, then Rep(T) =
-        <c><![CDATA[{type,LINE,binary,[Rep(B),{integer,LINE,0}]}]]></c>.</item>
-       <item>If T is a binary type <c><![CDATA[<< _ : _ * U >>]]></c>,
-        where <c><![CDATA[U]]></c> is a type, then Rep(T) =
-        <c><![CDATA[{type,LINE,binary,[{integer,LINE,0},Rep(U)]}]]></c>.</item>
-       <item>If T is a binary type <c><![CDATA[<< _ : B , _ : _ * U >>]]></c>,
-        where <c><![CDATA[B]]></c> and <c><![CDATA[U]]></c> is a type, then
-        Rep(T) =
-        <c><![CDATA[{type,LINE,binary,[Rep(B),Rep(U)]}]]></c>.</item>
-
-       <item>If T is a fun type <c><![CDATA[fun((...) -> Ret)]]></c>, then
-        Rep(T) = <c><![CDATA[{type,LINE,'fun',[{type,LINE,product,[]},Rep(Ret)]}]]></c>.
-        </item>
-       <item>If T is a fun type <c><![CDATA[fun(Tc)]]></c>, where
-        <c><![CDATA[Tc]]></c> is an unbounded fun type clause,
-        then Rep(T) = <c><![CDATA[Rep(Tc)]]></c>.</item>
-      </list>
-    </section>
-
-    <section>
-      <title>Record fields</title>
+      <title>Record Fields</title>
       <p>Each field in a record declaration may have an optional
-        explicit default initializer expression</p>
+        explicit default initializer expression, as well as an
+        optional type.</p>
       <list type="bulleted">
-        <item>If V is <c><![CDATA[A]]></c>, then
-         Rep(V) = <c><![CDATA[{record_field,LINE,Rep(A)}]]></c>.</item>
-        <item>If V is <c><![CDATA[A = E]]></c>, then
-         Rep(V) = <c><![CDATA[{record_field,LINE,Rep(A),Rep(E)}]]></c>.</item>
-        <item>If V is <c><![CDATA[A :: T]]></c>, where <c><![CDATA[A]]></c> is
-         an atom and <c><![CDATA[T]]></c> is a type and it does not contain
-         <c><![CDATA[undefined]]></c> syntactically, then Rep(V) =
-         <c><![CDATA[{typed_record_field,{record_field,LINE,Rep(A)},Rep(undefined | T)}]]></c>.
-         Note that if <![CDATA[T]]> is an annotated type, it will be wrapped in
-         parentheses.</item>
-        <item>If V is <c><![CDATA[A :: T]]></c>, where <c><![CDATA[A]]></c> is
-         an atom and <c><![CDATA[T]]></c> is a type, then Rep(V) =
-         <c><![CDATA[{typed_record_field,{record_field,LINE,Rep(A)},Rep(T)}]]></c>.
-         </item>
-        <item>If V is <c><![CDATA[A = E :: T]]></c>, where <c><![CDATA[A]]></c>
-         is an atom, <c><![CDATA[E]]></c> is an expression and
-         <c><![CDATA[T]]></c> is a type, then Rep(V) =
-         <c><![CDATA[{typed_record_field,{record_field,LINE,Rep(A),Rep(E)},Rep(T)}]]></c>.
-         </item>
+        <item>If V is <c>A</c>, then
+         Rep(V) = <c>{record_field,LINE,Rep(A)}</c>.</item>
+        <item>If V is <c>A = E</c>,
+	 where <c>E</c> is an expression, then
+         Rep(V) = <c>{record_field,LINE,Rep(A),Rep(E)}</c>.</item>
+        <item>If V is <c>A :: T</c>, where <c>T</c> is a type, then Rep(V) =
+          <c>{typed_record_field,{record_field,LINE,Rep(A)},Rep(T)}</c>.
+          </item>
+        <item>If V is <c>A = E :: T</c>, where
+          <c>E</c> is an expression and <c>T</c> is a type, then Rep(V) =
+          <c>{typed_record_field,{record_field,LINE,Rep(A),Rep(E)},Rep(T)}</c>.
+        </item>
       </list>
     </section>
 
     <section>
-      <title>Representation of parse errors and end of file</title>
+      <title>Representation of Parse Errors and End-of-file</title>
       <p>In addition to the representations of forms, the list that represents
-        a module declaration (as returned by functions in <c><![CDATA[erl_parse]]></c> and
-        <c><![CDATA[epp]]></c>) may contain tuples <c><![CDATA[{error,E}]]></c> and <c><![CDATA[{warning,W}]]></c>, denoting
-        syntactically incorrect forms and warnings, and <c><![CDATA[{eof,LINE}]]></c>, denoting an end
-        of stream encountered before a complete form had been parsed.</p>
+        a module declaration (as returned by functions in <c>erl_parse</c> and
+        <c>epp</c>) may contain tuples <c>{error,E}</c> and
+	<c>{warning,W}</c>, denoting syntactically incorrect forms and
+	warnings, and <c>{eof,LINE}</c>, denoting an end-of-stream
+	encountered before a complete form had been parsed.</p>
     </section>
   </section>
 
   <section>
-    <title>Atomic literals</title>
+    <title>Atomic Literals</title>
     <p>There are five kinds of atomic literals, which are represented in the
       same way in patterns, expressions and guards:</p>
     <list type="bulleted">
-      <item>If L is an integer or character literal, then
-       Rep(L) = <c><![CDATA[{integer,LINE,L}]]></c>.</item>
+      <item>If L is an atom literal, then
+       Rep(L) = <c>{atom,LINE,L}</c>.</item>
       <item>If L is a float literal, then
-       Rep(L) = <c><![CDATA[{float,LINE,L}]]></c>.</item>
+       Rep(L) = <c>{float,LINE,L}</c>.</item>
+      <item>If L is an integer or character literal, then
+       Rep(L) = <c>{integer,LINE,L}</c>.</item>
       <item>If L is a string literal consisting of the characters
-      <c><![CDATA[C_1]]></c>, ..., <c><![CDATA[C_k]]></c>, then
-       Rep(L) = <c><![CDATA[{string,LINE,[C_1, ..., C_k]}]]></c>.</item>
-      <item>If L is an atom literal, then
-       Rep(L) = <c><![CDATA[{atom,LINE,L}]]></c>.</item>
+      <c>C_1</c>, ..., <c>C_k</c>, then
+       Rep(L) = <c>{string,LINE,[C_1, ..., C_k]}</c>.</item>
     </list>
     <p>Note that negative integer and float literals do not occur as such; they are
       parsed as an application of the unary negation operator.</p>
@@ -291,47 +176,59 @@
 
   <section>
     <title>Patterns</title>
-    <p>If <c><![CDATA[Ps]]></c> is a sequence of patterns <c><![CDATA[P_1, ..., P_k]]></c>, then
-      Rep(Ps) = <c><![CDATA[[Rep(P_1), ..., Rep(P_k)]]]></c>. Such sequences occur as the
+    <p>If Ps is a sequence of patterns <c>P_1, ..., P_k</c>, then
+      Rep(Ps) = <c>[Rep(P_1), ..., Rep(P_k)]</c>. Such sequences occur as the
       list of arguments to a function or fun.</p>
     <p>Individual patterns are represented as follows:</p>
     <list type="bulleted">
-      <item>If P is an atomic literal L, then Rep(P) = Rep(L).</item>
-      <item>If P is a compound pattern <c><![CDATA[P_1 = P_2]]></c>, then
-       Rep(P) = <c><![CDATA[{match,LINE,Rep(P_1),Rep(P_2)}]]></c>.</item>
-      <item>If P is a variable pattern <c><![CDATA[V]]></c>, then
-       Rep(P) = <c><![CDATA[{var,LINE,A}]]></c>,
-       where A is an atom with a printname consisting of the same characters as
-      <c><![CDATA[V]]></c>.</item>
-      <item>If P is a universal pattern <c><![CDATA[_]]></c>, then
-       Rep(P) = <c><![CDATA[{var,LINE,'_'}]]></c>.</item>
-      <item>If P is a tuple pattern <c><![CDATA[{P_1, ..., P_k}]]></c>, then
-       Rep(P) = <c><![CDATA[{tuple,LINE,[Rep(P_1), ..., Rep(P_k)]}]]></c>.</item>
-      <item>If P is a nil pattern <c><![CDATA[[]]]></c>, then
-       Rep(P) = <c><![CDATA[{nil,LINE}]]></c>.</item>
-      <item>If P is a cons pattern <c><![CDATA[[P_h | P_t]]]></c>, then
-       Rep(P) = <c><![CDATA[{cons,LINE,Rep(P_h),Rep(P_t)}]]></c>.</item>
-      <item>If E is a binary pattern <c><![CDATA[<<P_1:Size_1/TSL_1, ..., P_k:Size_k/TSL_k>>]]></c>, then
-       Rep(E) = <c><![CDATA[{bin,LINE,[{bin_element,LINE,Rep(P_1),Rep(Size_1),Rep(TSL_1)}, ..., {bin_element,LINE,Rep(P_k),Rep(Size_k),Rep(TSL_k)}]}]]></c>.
+      <item>If P is an atomic literal <c>L</c>, then Rep(P) = Rep(L).</item>
+      <item>If P is a binary pattern
+       <c>&lt;&lt;P_1:Size_1/TSL_1, ..., P_k:Size_k/TSL_k>></c>, where each
+       <c>Size_i</c> is an expression that can be evaluated to an integer
+       and each <c>TSL_i</c> is a type specificer list, then
+       Rep(P) = <c>{bin,LINE,[{bin_element,LINE,Rep(P_1),Rep(Size_1),Rep(TSL_1)}, ..., {bin_element,LINE,Rep(P_k),Rep(Size_k),Rep(TSL_k)}]}</c>.
        For Rep(TSL), see below.
-       An omitted <c><![CDATA[Size]]></c> is represented by <c><![CDATA[default]]></c>. An omitted <c><![CDATA[TSL]]></c>
-       (type specifier list) is represented by <c><![CDATA[default]]></c>.</item>
-      <item>If P is <c><![CDATA[P_1 Op P_2]]></c>, where <c><![CDATA[Op]]></c> is a binary operator (this
-       is either an occurrence of <c><![CDATA[++]]></c> applied to a literal string or character
+       An omitted <c>Size_i</c> is represented by <c>default</c>.
+       An omitted <c>TSL_i</c> is represented by <c>default</c>.</item>
+      <item>If P is a compound pattern <c>P_1 = P_2</c>, then
+       Rep(P) = <c>{match,LINE,Rep(P_1),Rep(P_2)}</c>.</item>
+      <item>If P is a cons pattern <c>[P_h | P_t]</c>, then
+       Rep(P) = <c>{cons,LINE,Rep(P_h),Rep(P_t)}</c>.</item>
+      <item>If P is a map pattern <c>#{A_1, ..., A_k}</c>, where each
+       <c>A_i</c> is an association <c>P_i_1 := P_i_2</c>, then Rep(P) =
+       <c>{map,LINE,[Rep(A_1), ..., Rep(A_k)]}</c>. For Rep(A), see
+       below.</item>
+      <item>If P is a nil pattern <c>[]</c>, then
+       Rep(P) = <c>{nil,LINE}</c>.</item>
+      <item>If P is an operator pattern <c>P_1 Op P_2</c>,
+       where <c>Op</c> is a binary operator (this is either an occurrence
+       of <c>++</c> applied to a literal string or character
        list, or an occurrence of an expression that can be evaluated to a number
        at compile time),
-       then Rep(P) = <c><![CDATA[{op,LINE,Op,Rep(P_1),Rep(P_2)}]]></c>.</item>
-      <item>If P is <c><![CDATA[Op P_0]]></c>, where <c><![CDATA[Op]]></c> is a unary operator (this is an
-       occurrence of an expression that can be evaluated to a number at compile
-       time), then Rep(P) = <c><![CDATA[{op,LINE,Op,Rep(P_0)}]]></c>.</item>
-      <item>If P is a record pattern <c><![CDATA[#Name{Field_1=P_1, ..., Field_k=P_k}]]></c>,
-       then Rep(P) =
-      <c><![CDATA[{record,LINE,Name, [{record_field,LINE,Rep(Field_1),Rep(P_1)}, ..., {record_field,LINE,Rep(Field_k),Rep(P_k)}]}]]></c>.</item>
-      <item>If P is <c><![CDATA[#Name.Field]]></c>, then
-       Rep(P) = <c><![CDATA[{record_index,LINE,Name,Rep(Field)}]]></c>.</item>
-      <item>If P is <c><![CDATA[( P_0 )]]></c>, then
-       Rep(P) = <c><![CDATA[Rep(P_0)]]></c>,
-       i.e., patterns cannot be distinguished from their bodies.</item>
+       then Rep(P) = <c>{op,LINE,Op,Rep(P_1),Rep(P_2)}</c>.</item>
+      <item>If P is an operator pattern <c>Op P_0</c>,
+       where <c>Op</c> is a unary operator (this is an occurrence of
+       an expression that can be evaluated to a number at compile
+       time), then Rep(P) = <c>{op,LINE,Op,Rep(P_0)}</c>.</item>
+      <item>If P is a parenthesized pattern <c>( P_0 )</c>, then
+       Rep(P) = <c>Rep(P_0)</c>,
+       that is, parenthesized patterns cannot be distinguished from their
+       bodies.</item>
+      <item>If P is a record field index pattern <c>#Name.Field</c>,
+       where <c>Field</c> is an atom, then
+       Rep(P) = <c>{record_index,LINE,Name,Rep(Field)}</c>.</item>
+      <item>If P is a record pattern
+       <c>#Name{Field_1=P_1, ..., Field_k=P_k}</c>,
+       where each <c>Field_i</c> is an atom or <c>_</c>, then Rep(P) =
+       <c>{record,LINE,Name,[{record_field,LINE,Rep(Field_1),Rep(P_1)}, ..., {record_field,LINE,Rep(Field_k),Rep(P_k)}]}</c>.</item>
+      <item>If P is a tuple pattern <c>{P_1, ..., P_k}</c>, then
+       Rep(P) = <c>{tuple,LINE,[Rep(P_1), ..., Rep(P_k)]}</c>.</item>
+      <item>If P is a universal pattern <c>_</c>, then
+       Rep(P) = <c>{var,LINE,'_'}</c>.</item>
+      <item>If P is a variable pattern <c>V</c>, then
+       Rep(P) = <c>{var,LINE,A}</c>,
+       where A is an atom with a printname consisting of the same characters as
+      <c>V</c>.</item>
     </list>
     <p>Note that every pattern has the same source form as some expression, and is
       represented the same way as the corresponding expression.</p>
@@ -339,180 +236,185 @@
 
   <section>
     <title>Expressions</title>
-    <p>A body B is a sequence of expressions <c><![CDATA[E_1, ..., E_k]]></c>, and
-      Rep(B) = <c><![CDATA[[Rep(E_1), ..., Rep(E_k)]]]></c>.</p>
+    <p>A body B is a nonempty sequence of expressions <c>E_1, ..., E_k</c>,
+      and Rep(B) = <c>[Rep(E_1), ..., Rep(E_k)]</c>.</p>
     <p>An expression E is one of the following alternatives:</p>
     <list type="bulleted">
-      <item>If P is an atomic literal <c><![CDATA[L]]></c>, then
-       Rep(P) = Rep(L).</item>
-      <item>If E is <c><![CDATA[P = E_0]]></c>, then
-       Rep(E) = <c><![CDATA[{match,LINE,Rep(P),Rep(E_0)}]]></c>.</item>
-      <item>If E is a variable <c><![CDATA[V]]></c>, then
-       Rep(E) = <c><![CDATA[{var,LINE,A}]]></c>,
-       where <c><![CDATA[A]]></c> is an atom with a printname consisting of the same
-       characters as <c><![CDATA[V]]></c>.</item>
-      <item>If E is a tuple skeleton <c><![CDATA[{E_1, ..., E_k}]]></c>, then
-       Rep(E) = <c><![CDATA[{tuple,LINE,[Rep(E_1), ..., Rep(E_k)]}]]></c>.</item>
-      <item>If E is <c><![CDATA[[]]]></c>, then
-       Rep(E) = <c><![CDATA[{nil,LINE}]]></c>.</item>
-      <item>If E is a cons skeleton <c><![CDATA[[E_h | E_t]]]></c>, then
-       Rep(E) = <c><![CDATA[{cons,LINE,Rep(E_h),Rep(E_t)}]]></c>.</item>
-      <item>If E is a binary constructor <c><![CDATA[<<V_1:Size_1/TSL_1, ..., V_k:Size_k/TSL_k>>]]></c>, then
-       Rep(E) = <c><![CDATA[{bin,LINE,[{bin_element,LINE,Rep(V_1),Rep(Size_1),Rep(TSL_1)}, ..., {bin_element,LINE,Rep(V_k),Rep(Size_k),Rep(TSL_k)}]}]]></c>.
+      <item>If E is an atomic literal <c>L</c>, then Rep(E) = Rep(L).</item>
+      <item>If E is a binary comprehension
+       <c>&lt;&lt;E_0 || Q_1, ..., Q_k>></c>,
+       where each <c>Q_i</c> is a qualifier, then
+       Rep(E) = <c>{bc,LINE,Rep(E_0),[Rep(Q_1), ..., Rep(Q_k)]}</c>.
+       For Rep(Q), see below.</item>
+      <item>If E is a binary constructor <c>&lt;&lt;E_1:Size_1/TSL_1, ..., E_k:Size_k/TSL_k>></c>,
+       where each <c>Size_i</c> is an expression and each
+       <c>TSL_i</c> is a type specificer list, then Rep(E) =
+       <c>{bin,LINE,[{bin_element,LINE,Rep(E_1),Rep(Size_1),Rep(TSL_1)}, ..., {bin_element,LINE,Rep(E_k),Rep(Size_k),Rep(TSL_k)}]}</c>.
        For Rep(TSL), see below.
-       An omitted <c><![CDATA[Size]]></c> is represented by <c><![CDATA[default]]></c>. An omitted <c><![CDATA[TSL]]></c>
-       (type specifier list) is represented by <c><![CDATA[default]]></c>.</item>
-      <item>If E is <c><![CDATA[E_1 Op E_2]]></c>, where <c><![CDATA[Op]]></c> is a binary operator,
-       then Rep(E) = <c><![CDATA[{op,LINE,Op,Rep(E_1),Rep(E_2)}]]></c>.</item>
-      <item>If E is <c><![CDATA[Op E_0]]></c>, where <c><![CDATA[Op]]></c> is a unary operator, then
-       Rep(E) = <c><![CDATA[{op,LINE,Op,Rep(E_0)}]]></c>.</item>
-      <item>If E is <c><![CDATA[#Name{Field_1=E_1, ..., Field_k=E_k}]]></c>, then
-       Rep(E) =
-      <c><![CDATA[{record,LINE,Name,  [{record_field,LINE,Rep(Field_1),Rep(E_1)}, ..., {record_field,LINE,Rep(Field_k),Rep(E_k)}]}]]></c>.</item>
-      <item>If E is <c><![CDATA[E_0#Name{Field_1=E_1, ..., Field_k=E_k}]]></c>, then
-       Rep(E) =
-      <c><![CDATA[{record,LINE,Rep(E_0),Name,  [{record_field,LINE,Rep(Field_1),Rep(E_1)}, ..., {record_field,LINE,Rep(Field_k),Rep(E_k)}]}]]></c>.</item>
-      <item>If E is <c><![CDATA[#Name.Field]]></c>, then
-       Rep(E) = <c><![CDATA[{record_index,LINE,Name,Rep(Field)}]]></c>.</item>
-      <item>If E is <c><![CDATA[E_0#Name.Field]]></c>, then
-       Rep(E) = <c><![CDATA[{record_field,LINE,Rep(E_0),Name,Rep(Field)}]]></c>.</item>
-      <item>If E is <c><![CDATA[#{W_1, ..., W_k}]]></c> where each
-       <c><![CDATA[W_i]]></c> is a map assoc or exact field, then Rep(E) =
-       <c><![CDATA[{map,LINE,[Rep(W_1), ..., Rep(W_k)]}]]></c>. For Rep(W), see
-       below.</item>
-      <item>If E is <c><![CDATA[E_0#{W_1, ..., W_k}]]></c> where
-       <c><![CDATA[W_i]]></c> is a map assoc or exact field, then Rep(E) =
-       <c><![CDATA[{map,LINE,Rep(E_0),[Rep(W_1), ..., Rep(W_k)]}]]></c>. For
-       Rep(W), see below.</item>
-      <item>If E is <c><![CDATA[catch E_0]]></c>, then
-       Rep(E) = <c><![CDATA[{'catch',LINE,Rep(E_0)}]]></c>.</item>
-      <item>If E is <c><![CDATA[E_0(E_1, ..., E_k)]]></c>, then
-       Rep(E) = <c><![CDATA[{call,LINE,Rep(E_0),[Rep(E_1), ..., Rep(E_k)]}]]></c>.</item>
-      <item>If E is <c><![CDATA[E_m:E_0(E_1, ..., E_k)]]></c>, then
-       Rep(E) =
-      <c><![CDATA[{call,LINE,{remote,LINE,Rep(E_m),Rep(E_0)},[Rep(E_1), ...,  Rep(E_k)]}]]></c>.</item>
-      <item>If E is a list comprehension <c><![CDATA[[E_0 || W_1, ..., W_k]]]></c>,
-       where each <c><![CDATA[W_i]]></c> is a generator or a filter, then
-       Rep(E) = <c><![CDATA[{lc,LINE,Rep(E_0),[Rep(W_1), ..., Rep(W_k)]}]]></c>. For Rep(W), see
+       An omitted <c>Size_i</c> is represented by <c>default</c>.
+       An omitted <c>TSL_i</c> is represented by <c>default</c>.</item>
+      <item>If E is a block expression <c>begin B end</c>,
+       where <c>B</c> is a body, then
+       Rep(E) = <c>{block,LINE,Rep(B)}</c>.</item>
+      <item>If E is a case expression <c>case E_0 of Cc_1 ; ... ; Cc_k end</c>,
+       where <c>E_0</c> is an expression and each <c>Cc_i</c> is a
+       case clause then Rep(E) =
+      <c>{'case',LINE,Rep(E_0),[Rep(Cc_1), ..., Rep(Cc_k)]}</c>.</item>
+      <item>If E is a catch expression <c>catch E_0</c>, then
+       Rep(E) = <c>{'catch',LINE,Rep(E_0)}</c>.</item>
+      <item>If E is a cons skeleton <c>[E_h | E_t]</c>, then
+       Rep(E) = <c>{cons,LINE,Rep(E_h),Rep(E_t)}</c>.</item>
+      <item>If E is a fun expression <c>fun Name/Arity</c>, then
+       Rep(E) = <c>{'fun',LINE,{function,Name,Arity}}</c>.</item>
+      <item>If E is a fun expression
+       <c>fun Module:Name/Arity</c>, then Rep(E) =
+       <c>{'fun',LINE,{function,Rep(Module),Rep(Name),Rep(Arity)}}</c>.
+       (Before the R15 release: Rep(E) =
+       <c>{'fun',LINE,{function,Module,Name,Arity}}</c>.)</item>
+      <item>If E is a fun expression <c>fun Fc_1 ; ... ; Fc_k end</c>,
+       where each <c>Fc_i</c> is a function clause then Rep(E) =
+      <c>{'fun',LINE,{clauses,[Rep(Fc_1), ..., Rep(Fc_k)]}}</c>.</item>
+      <item>If E is a fun expression
+       <c>fun Name Fc_1 ; ... ; Name Fc_k end</c>,
+       where <c>Name</c> is a variable and each
+       <c>Fc_i</c> is a function clause then Rep(E) =
+       <c>{named_fun,LINE,Name,[Rep(Fc_1), ..., Rep(Fc_k)]}</c>.
+       </item>
+      <item>If E is a function call <c>E_0(E_1, ..., E_k)</c>, then
+       Rep(E) = <c>{call,LINE,Rep(E_0),[Rep(E_1), ..., Rep(E_k)]}</c>.</item>
+      <item>If E is a function call <c>E_m:E_0(E_1, ..., E_k)</c>,
+       then Rep(E) =
+       <c>{call,LINE,{remote,LINE,Rep(E_m),Rep(E_0)},[Rep(E_1), ...,  Rep(E_k)]}</c>.
+      </item>
+      <item>If E is an if expression <c>if Ic_1 ; ... ; Ic_k  end</c>,
+       where each <c>Ic_i</c> is an if clause then Rep(E) =
+      <c>{'if',LINE,[Rep(Ic_1), ..., Rep(Ic_k)]}</c>.</item>
+      <item>If E is a list comprehension <c>[E_0 || Q_1, ..., Q_k]</c>,
+       where each <c>Q_i</c> is a qualifier, then Rep(E) =
+       <c>{lc,LINE,Rep(E_0),[Rep(Q_1), ..., Rep(Q_k)]}</c>. For Rep(Q), see
        below.</item>
-      <item>If E is a binary comprehension <c><![CDATA[<<E_0 || W_1, ..., W_k>>]]></c>,
-       where each <c><![CDATA[W_i]]></c> is a generator or a filter, then
-       Rep(E) = <c><![CDATA[{bc,LINE,Rep(E_0),[Rep(W_1), ..., Rep(W_k)]}]]></c>. For Rep(W), see
+      <item>If E is a map creation <c>#{A_1, ..., A_k}</c>,
+       where each <c>A_i</c> is an association <c>E_i_1 => E_i_2</c>
+       or <c>E_i_1 := E_i_2</c>, then Rep(E) =
+       <c>{map,LINE,[Rep(A_1), ..., Rep(A_k)]}</c>. For Rep(A), see
        below.</item>
-      <item>If E is <c><![CDATA[begin B end]]></c>, where <c><![CDATA[B]]></c> is a body, then
-       Rep(E) = <c><![CDATA[{block,LINE,Rep(B)}]]></c>.</item>
-      <item>If E is <c><![CDATA[if Ic_1 ; ... ; Ic_k  end]]></c>,
-       where each <c><![CDATA[Ic_i]]></c> is an if clause then
+      <item>If E is a map update <c>E_0#{A_1, ..., A_k}</c>,
+       where each <c>A_i</c> is an association <c>E_i_1 => E_i_2</c>
+       or <c>E_i_1 := E_i_2</c>, then Rep(E) =
+       <c>{map,LINE,Rep(E_0),[Rep(A_1), ..., Rep(A_k)]}</c>.
+       For Rep(A), see below.</item>
+      <item>If E is a match operator expression <c>P = E_0</c>,
+       where <c>P</c> is a pattern, then
+       Rep(E) = <c>{match,LINE,Rep(P),Rep(E_0)}</c>.</item>
+      <item>If E is nil, <c>[]</c>, then
+       Rep(E) = <c>{nil,LINE}</c>.</item>
+      <item>If E is an operator expression <c>E_1 Op E_2</c>,
+       where <c>Op</c> is a binary operator other than the match
+       operator <c>=</c>, then
+       Rep(E) = <c>{op,LINE,Op,Rep(E_1),Rep(E_2)}</c>.</item>
+      <item>If E is an operator expression <c>Op E_0</c>,
+       where <c>Op</c> is a unary operator, then
+       Rep(E) = <c>{op,LINE,Op,Rep(E_0)}</c>.</item>
+      <item>If E is a parenthesized expression <c>( E_0 )</c>, then
+       Rep(E) = <c>Rep(E_0)</c>, that is, parenthesized
+       expressions cannot be distinguished from their bodies.</item>
+      <item>If E is a receive expression <c>receive Cc_1 ; ... ; Cc_k end</c>,
+       where each <c>Cc_i</c> is a case clause then Rep(E) =
+      <c>{'receive',LINE,[Rep(Cc_1), ..., Rep(Cc_k)]}</c>.</item>
+      <item>If E is a receive expression
+       <c>receive Cc_1 ; ... ; Cc_k after E_0 -> B_t end</c>,
+       where each <c>Cc_i</c> is a case clause,
+      <c>E_0</c> is an expression and <c>B_t</c> is a body, then Rep(E) =
+      <c>{'receive',LINE,[Rep(Cc_1), ..., Rep(Cc_k)],Rep(E_0),Rep(B_t)}</c>.</item>
+      <item>If E is a record creation
+       <c>#Name{Field_1=E_1, ..., Field_k=E_k}</c>,
+       where each <c>Field_i</c> is an atom or <c>_</c>, then Rep(E) =
+      <c>{record,LINE,Name,[{record_field,LINE,Rep(Field_1),Rep(E_1)}, ..., {record_field,LINE,Rep(Field_k),Rep(E_k)}]}</c>.</item>
+      <item>If E is a record field access <c>E_0#Name.Field</c>,
+       where <c>Field</c> is an atom, then
+       Rep(E) = <c>{record_field,LINE,Rep(E_0),Name,Rep(Field)}</c>.</item>
+      <item>If E is a record field index <c>#Name.Field</c>,
+       where <c>Field</c> is an atom, then
+       Rep(E) = <c>{record_index,LINE,Name,Rep(Field)}</c>.</item>
+      <item>If E is a record update
+       <c>E_0#Name{Field_1=E_1, ..., Field_k=E_k}</c>,
+       where each <c>Field_i</c> is an atom, then Rep(E) =
+      <c>{record,LINE,Rep(E_0),Name,[{record_field,LINE,Rep(Field_1),Rep(E_1)}, ..., {record_field,LINE,Rep(Field_k),Rep(E_k)}]}</c>.</item>
+      <item>If E is a tuple skeleton <c>{E_1, ..., E_k}</c>, then
+       Rep(E) = <c>{tuple,LINE,[Rep(E_1), ..., Rep(E_k)]}</c>.</item>
+      <item>If E is a try expression <c>try B catch Tc_1 ; ... ; Tc_k end</c>,
+       where <c>B</c> is a body and each <c>Tc_i</c> is a catch clause then
        Rep(E) =
-      <c><![CDATA[{'if',LINE,[Rep(Ic_1), ..., Rep(Ic_k)]}]]></c>.</item>
-      <item>If E is <c><![CDATA[case E_0 of Cc_1 ; ... ; Cc_k end]]></c>,
-       where <c><![CDATA[E_0]]></c> is an expression and each <c><![CDATA[Cc_i]]></c> is a
-       case clause then
+      <c>{'try',LINE,Rep(B),[],[Rep(Tc_1), ..., Rep(Tc_k)],[]}</c>.</item>
+      <item>If E is a try expression
+       <c>try B of Cc_1 ; ... ; Cc_k catch Tc_1 ; ... ; Tc_n end</c>,
+       where <c>B</c> is a body,
+       each <c>Cc_i</c> is a case clause and
+       each <c>Tc_j</c> is a catch clause then Rep(E) =
+      <c>{'try',LINE,Rep(B),[Rep(Cc_1), ..., Rep(Cc_k)],[Rep(Tc_1), ..., Rep(Tc_n)],[]}</c>.</item>
+      <item>If E is a try expression <c>try B after A end</c>,
+       where <c>B</c> and <c>A</c> are bodies then Rep(E) =
+      <c>{'try',LINE,Rep(B),[],[],Rep(A)}</c>.</item>
+      <item>If E is a try expression
+       <c>try B of Cc_1 ; ... ; Cc_k after A end</c>,
+       where <c>B</c> and <c>A</c> are a bodies and
+       each <c>Cc_i</c> is a case clause then Rep(E) =
+      <c>{'try',LINE,Rep(B),[Rep(Cc_1), ..., Rep(Cc_k)],[],Rep(A)}</c>.</item>
+      <item>If E is a try expression
+       <c>try B catch Tc_1 ; ... ; Tc_k after A end</c>,
+       where <c>B</c> and <c>A</c> are bodies and
+       each <c>Tc_i</c> is a catch clause then Rep(E) =
+      <c>{'try',LINE,Rep(B),[],[Rep(Tc_1), ..., Rep(Tc_k)],Rep(A)}</c>.</item>
+      <item>If E is a try expression
+       <c>try B of Cc_1 ; ... ; Cc_k  catch Tc_1 ; ... ; Tc_n after A end</c>,
+       where <c>B</c> and <c>A</c> are a bodies,
+       each <c>Cc_i</c> is a case clause, and
+       each <c>Tc_j</c> is a catch clause then
        Rep(E) =
-      <c><![CDATA[{'case',LINE,Rep(E_0),[Rep(Cc_1), ..., Rep(Cc_k)]}]]></c>.</item>
-      <item>If E is <c><![CDATA[try B catch Tc_1 ; ... ; Tc_k end]]></c>,
-       where <c><![CDATA[B]]></c> is a body and each <c><![CDATA[Tc_i]]></c> is a catch clause then 
-       Rep(E) =
-      <c><![CDATA[{'try',LINE,Rep(B),[],[Rep(Tc_1), ..., Rep(Tc_k)],[]}]]></c>.</item>
-      <item>If E is <c><![CDATA[try B of Cc_1 ; ... ; Cc_k catch Tc_1 ; ... ; Tc_n end]]></c>,
-       where <c><![CDATA[B]]></c> is a body, 
-       each <c><![CDATA[Cc_i]]></c> is a case clause and 
-       each <c><![CDATA[Tc_j]]></c> is a catch clause then 
-       Rep(E) =
-      <c><![CDATA[{'try',LINE,Rep(B),[Rep(Cc_1), ..., Rep(Cc_k)],[Rep(Tc_1), ..., Rep(Tc_n)],[]}]]></c>.</item>
-      <item>If E is <c><![CDATA[try B after A end]]></c>,
-       where <c><![CDATA[B]]></c> and <c><![CDATA[A]]></c> are bodies then
-       Rep(E) =
-      <c><![CDATA[{'try',LINE,Rep(B),[],[],Rep(A)}]]></c>.</item>
-      <item>If E is <c><![CDATA[try B of Cc_1 ; ... ; Cc_k after A end]]></c>,
-       where <c><![CDATA[B]]></c> and <c><![CDATA[A]]></c> are a bodies and
-       each <c><![CDATA[Cc_i]]></c> is a case clause then 
-       Rep(E) =
-      <c><![CDATA[{'try',LINE,Rep(B),[Rep(Cc_1), ..., Rep(Cc_k)],[],Rep(A)}]]></c>.</item>
-      <item>If E is <c><![CDATA[try B catch Tc_1 ; ... ; Tc_k after A end]]></c>,
-       where <c><![CDATA[B]]></c> and <c><![CDATA[A]]></c> are bodies and
-       each <c><![CDATA[Tc_i]]></c> is a catch clause then 
-       Rep(E) =
-      <c><![CDATA[{'try',LINE,Rep(B),[],[Rep(Tc_1), ..., Rep(Tc_k)],Rep(A)}]]></c>.</item>
-      <item>If E is <c><![CDATA[try B of Cc_1 ; ... ; Cc_k  catch Tc_1 ; ... ; Tc_n after A end]]></c>,
-       where <c><![CDATA[B]]></c> and <c><![CDATA[A]]></c> are a bodies, 
-       each <c><![CDATA[Cc_i]]></c> is a case clause and
-       each <c><![CDATA[Tc_j]]></c> is a catch clause then 
-       Rep(E) =
-      <c><![CDATA[{'try',LINE,Rep(B),[Rep(Cc_1), ..., Rep(Cc_k)],[Rep(Tc_1),  ..., Rep(Tc_n)],Rep(A)}]]></c>.</item>
-      <item>If E is <c><![CDATA[receive Cc_1 ; ... ; Cc_k end]]></c>,
-       where each <c><![CDATA[Cc_i]]></c> is a case clause then
-       Rep(E) =
-      <c><![CDATA[{'receive',LINE,[Rep(Cc_1), ..., Rep(Cc_k)]}]]></c>.</item>
-      <item>If E is <c><![CDATA[receive Cc_1 ; ... ; Cc_k after E_0 -> B_t end]]></c>,
-       where each <c><![CDATA[Cc_i]]></c> is a case clause, 
-      <c><![CDATA[E_0]]></c> is an expression and <c><![CDATA[B_t]]></c> is a body, then
-       Rep(E) =
-      <c><![CDATA[{'receive',LINE,[Rep(Cc_1), ..., Rep(Cc_k)],Rep(E_0),Rep(B_t)}]]></c>.</item>
-      <item>If E is <c><![CDATA[fun Name / Arity]]></c>, then
-       Rep(E) = <c><![CDATA[{'fun',LINE,{function,Name,Arity}}]]></c>.</item>
-      <item>If E is <c><![CDATA[fun Module:Name/Arity]]></c>, then
-       Rep(E) = <c><![CDATA[{'fun',LINE,{function,Rep(Module),Rep(Name),Rep(Arity)}}]]></c>.
-      (Before the R15 release: Rep(E) = <c><![CDATA[{'fun',LINE,{function,Module,Name,Arity}}]]></c>.)</item>
-      <item>If E is <c><![CDATA[fun Fc_1 ; ... ; Fc_k end]]></c> 
-       where each <c><![CDATA[Fc_i]]></c> is a function clause then Rep(E) =
-      <c><![CDATA[{'fun',LINE,{clauses,[Rep(Fc_1), ..., Rep(Fc_k)]}}]]></c>.</item>
-      <item>If E is <c><![CDATA[fun Name Fc_1 ; ... ; Name Fc_k end]]></c>
-       where <c><![CDATA[Name]]></c> is a variable and each
-       <c><![CDATA[Fc_i]]></c> is a function clause then Rep(E) =
-       <c><![CDATA[{named_fun,LINE,Name,[Rep(Fc_1), ..., Rep(Fc_k)]}]]></c>.
-       </item>
-      <item>If E is <c><![CDATA[query [E_0 || W_1, ..., W_k] end]]></c>,
-       where each <c><![CDATA[W_i]]></c> is a generator or a filter, then
-       Rep(E) = <c><![CDATA[{'query',LINE,{lc,LINE,Rep(E_0),[Rep(W_1), ..., Rep(W_k)]}}]]></c>.
-       For Rep(W), see below.</item>
-      <item>If E is <c><![CDATA[E_0.Field]]></c>, a Mnesia record access
-       inside a query, then
-       Rep(E) = <c><![CDATA[{record_field,LINE,Rep(E_0),Rep(Field)}]]></c>.</item>
-      <item>If E is <c><![CDATA[( E_0 )]]></c>, then
-       Rep(E) = <c><![CDATA[Rep(E_0)]]></c>,
-       i.e., parenthesized expressions cannot be distinguished from their bodies.</item>
+      <c>{'try',LINE,Rep(B),[Rep(Cc_1), ..., Rep(Cc_k)],[Rep(Tc_1),  ..., Rep(Tc_n)],Rep(A)}</c>.</item>
+      <item>If E is a variable <c>V</c>, then Rep(E) = <c>{var,LINE,A}</c>,
+       where <c>A</c> is an atom with a printname consisting of the same
+       characters as <c>V</c>.</item>
     </list>
 
     <section>
-      <title>Generators and filters</title>
-      <p>When W is a generator or a filter (in the body of a list or binary comprehension), then:</p>
+      <title>Qualifiers</title>
+      <p>A qualifier Q is one of the following alternatives:</p>
       <list type="bulleted">
-        <item>If W is a generator <c><![CDATA[P <- E]]></c>, where <c><![CDATA[P]]></c> is a pattern and <c><![CDATA[E]]></c>
-         is an expression, then
-         Rep(W) = <c><![CDATA[{generate,LINE,Rep(P),Rep(E)}]]></c>.</item>
-        <item>If W is a generator <c><![CDATA[P <= E]]></c>, where <c><![CDATA[P]]></c> is a pattern and <c><![CDATA[E]]></c>
-         is an expression, then
-         Rep(W) = <c><![CDATA[{b_generate,LINE,Rep(P),Rep(E)}]]></c>.</item>
-        <item>If W is a filter <c><![CDATA[E]]></c>, which is an expression, then
-         Rep(W) = <c><![CDATA[Rep(E)]]></c>.</item>
+        <item>If Q is a filter <c>E</c>, where <c>E</c> is an expression, then
+         Rep(Q) = <c>Rep(E)</c>.</item>
+        <item>If Q is a generator <c>P &lt;- E</c>, where <c>P</c> is
+	 a pattern and <c>E</c> is an expression, then
+         Rep(Q) = <c>{generate,LINE,Rep(P),Rep(E)}</c>.</item>
+        <item>If Q is a generator <c>P &lt;= E</c>, where <c>P</c> is
+	 a pattern and <c>E</c> is an expression, then
+         Rep(Q) = <c>{b_generate,LINE,Rep(P),Rep(E)}</c>.</item>
       </list>
     </section>
 
     <section>
-      <title>Binary element type specifiers</title>
+      <title>Binary Element Type Specifiers</title>
       <p>A type specifier list TSL for a binary element is a sequence of type
-        specifiers <c><![CDATA[TS_1 - ... - TS_k]]></c>.
-        Rep(TSL) = <c><![CDATA[[Rep(TS_1), ..., Rep(TS_k)]]]></c>.</p>
-      <p>When TS is a type specifier for a binary element, then:</p>
+        specifiers <c>TS_1 - ... - TS_k</c>, and
+        Rep(TSL) = <c>[Rep(TS_1), ..., Rep(TS_k)]</c>.</p>
       <list type="bulleted">
-        <item>If TS is an atom <c><![CDATA[A]]></c>, Rep(TS) = <c><![CDATA[A]]></c>.</item>
-        <item>If TS is a couple <c><![CDATA[A:Value]]></c> where <c><![CDATA[A]]></c> is an atom and <c><![CDATA[Value]]></c>
-         is an integer, Rep(TS) = <c><![CDATA[{A, Value}]]></c>.</item>
+        <item>If TS is a type specifier <c>A</c>, where <c>A</c> is an atom,
+	 then Rep(TS) = <c>A</c>.</item>
+        <item>If TS is a type specifier <c>A:Value</c>,
+	 where <c>A</c> is an atom and <c>Value</c> is an integer,
+	 then Rep(TS) = <c>{A,Value}</c>.</item>
       </list>
     </section>
 
     <section>
-      <title>Map assoc and exact fields</title>
-      <p>When W is an assoc or exact field (in the body of a map), then:</p>
+      <title>Associations</title>
+      <p>An association A is one of the following alternatives:</p>
       <list type="bulleted">
-        <item>If W is an assoc field <c><![CDATA[K => V]]></c>, where
-         <c><![CDATA[K]]></c> and <c><![CDATA[V]]></c> are both expressions,
-         then Rep(W) = <c><![CDATA[{map_field_assoc,LINE,Rep(K),Rep(V)}]]></c>.
+	<item>If A is an association <c>K => V</c>,
+	 then Rep(A) = <c>{map_field_assoc,LINE,Rep(K),Rep(V)}</c>.
          </item>
-        <item>If W is an exact field <c><![CDATA[K := V]]></c>, where
-         <c><![CDATA[K]]></c> and <c><![CDATA[V]]></c> are both expressions,
-         then Rep(W) = <c><![CDATA[{map_field_exact,LINE,Rep(K),Rep(V)}]]></c>.
+	<item>If A is an association <c>K := V</c>,
+         then Rep(A) = <c>{map_field_exact,LINE,Rep(K),Rep(V)}</c>.
          </item>
       </list>
     </section>
@@ -520,112 +422,245 @@
 
   <section>
     <title>Clauses</title>
-    <p>There are function clauses, if clauses, case clauses 
+    <p>There are function clauses, if clauses, case clauses
       and catch clauses.</p>
-    <p>A clause <c><![CDATA[C]]></c> is one of the following alternatives:</p>
+    <p>A clause <c>C</c> is one of the following alternatives:</p>
     <list type="bulleted">
-      <item>If C is a function clause <c><![CDATA[( Ps ) -> B]]></c> 
-       where <c><![CDATA[Ps]]></c> is a pattern sequence and <c><![CDATA[B]]></c> is a body, then 
-       Rep(C) = <c><![CDATA[{clause,LINE,Rep(Ps),[],Rep(B)}]]></c>.</item>
-      <item>If C is a function clause <c><![CDATA[( Ps ) when Gs -> B]]></c> 
-       where <c><![CDATA[Ps]]></c> is a pattern sequence, 
-      <c><![CDATA[Gs]]></c> is a guard sequence and <c><![CDATA[B]]></c> is a body, then 
-       Rep(C) = <c><![CDATA[{clause,LINE,Rep(Ps),Rep(Gs),Rep(B)}]]></c>.</item>
-      <item>If C is an if clause <c><![CDATA[Gs -> B]]></c> 
-       where <c><![CDATA[Gs]]></c> is a guard sequence and <c><![CDATA[B]]></c> is a body, then 
-       Rep(C) = <c><![CDATA[{clause,LINE,[],Rep(Gs),Rep(B)}]]></c>.</item>
-      <item>If C is a case clause <c><![CDATA[P -> B]]></c> 
-       where <c><![CDATA[P]]></c> is a pattern and <c><![CDATA[B]]></c> is a body, then 
-       Rep(C) = <c><![CDATA[{clause,LINE,[Rep(P)],[],Rep(B)}]]></c>.</item>
-      <item>If C is a case clause <c><![CDATA[P when Gs -> B]]></c> 
-       where <c><![CDATA[P]]></c> is a pattern, 
-      <c><![CDATA[Gs]]></c> is a guard sequence and <c><![CDATA[B]]></c> is a body, then 
-       Rep(C) = <c><![CDATA[{clause,LINE,[Rep(P)],Rep(Gs),Rep(B)}]]></c>.</item>
-      <item>If C is a catch clause <c><![CDATA[P -> B]]></c> 
-       where <c><![CDATA[P]]></c> is a pattern and <c><![CDATA[B]]></c> is a body, then 
-       Rep(C) = <c><![CDATA[{clause,LINE,[Rep({throw,P,_})],[],Rep(B)}]]></c>.</item>
-      <item>If C is a catch clause <c><![CDATA[X : P -> B]]></c> 
-       where <c><![CDATA[X]]></c> is an atomic literal or a variable pattern, 
-      <c><![CDATA[P]]></c> is a pattern and <c><![CDATA[B]]></c> is a body, then 
-       Rep(C) = <c><![CDATA[{clause,LINE,[Rep({X,P,_})],[],Rep(B)}]]></c>.</item>
-      <item>If C is a catch clause <c><![CDATA[P when Gs -> B]]></c> 
-       where <c><![CDATA[P]]></c> is a pattern, <c><![CDATA[Gs]]></c> is a guard sequence 
-       and <c><![CDATA[B]]></c> is a body, then 
-       Rep(C) = <c><![CDATA[{clause,LINE,[Rep({throw,P,_})],Rep(Gs),Rep(B)}]]></c>.</item>
-      <item>If C is a catch clause <c><![CDATA[X : P when Gs -> B]]></c> 
-       where <c><![CDATA[X]]></c> is an atomic literal or a variable pattern, 
-      <c><![CDATA[P]]></c> is a pattern, <c><![CDATA[Gs]]></c> is a guard sequence 
-       and <c><![CDATA[B]]></c> is a body, then 
-       Rep(C) = <c><![CDATA[{clause,LINE,[Rep({X,P,_})],Rep(Gs),Rep(B)}]]></c>.</item>
+      <item>If C is a case clause <c>P -> B</c>,
+       where <c>P</c> is a pattern and <c>B</c> is a body, then
+       Rep(C) = <c>{clause,LINE,[Rep(P)],[],Rep(B)}</c>.</item>
+      <item>If C is a case clause <c>P when Gs -> B</c>,
+       where <c>P</c> is a pattern,
+      <c>Gs</c> is a guard sequence and <c>B</c> is a body, then
+       Rep(C) = <c>{clause,LINE,[Rep(P)],Rep(Gs),Rep(B)}</c>.</item>
+      <item>If C is a catch clause <c>P -> B</c>,
+       where <c>P</c> is a pattern and <c>B</c> is a body, then
+       Rep(C) = <c>{clause,LINE,[Rep({throw,P,_})],[],Rep(B)}</c>.</item>
+      <item>If C is a catch clause <c>X : P -> B</c>,
+       where <c>X</c> is an atomic literal or a variable pattern,
+      <c>P</c> is a pattern, and <c>B</c> is a body, then
+       Rep(C) = <c>{clause,LINE,[Rep({X,P,_})],[],Rep(B)}</c>.</item>
+      <item>If C is a catch clause <c>P when Gs -> B</c>,
+       where <c>P</c> is a pattern, <c>Gs</c> is a guard sequence,
+       and <c>B</c> is a body, then
+       Rep(C) = <c>{clause,LINE,[Rep({throw,P,_})],Rep(Gs),Rep(B)}</c>.</item>
+      <item>If C is a catch clause <c>X : P when Gs -> B</c>,
+       where <c>X</c> is an atomic literal or a variable pattern,
+      <c>P</c> is a pattern, <c>Gs</c> is a guard sequence,
+       and <c>B</c> is a body, then
+       Rep(C) = <c>{clause,LINE,[Rep({X,P,_})],Rep(Gs),Rep(B)}</c>.</item>
+      <item>If C is a function clause <c>( Ps ) -> B</c>,
+       where <c>Ps</c> is a pattern sequence and <c>B</c> is a body, then
+       Rep(C) = <c>{clause,LINE,Rep(Ps),[],Rep(B)}</c>.</item>
+      <item>If C is a function clause <c>( Ps ) when Gs -> B</c>,
+       where <c>Ps</c> is a pattern sequence,
+      <c>Gs</c> is a guard sequence and <c>B</c> is a body, then
+       Rep(C) = <c>{clause,LINE,Rep(Ps),Rep(Gs),Rep(B)}</c>.</item>
+      <item>If C is an if clause <c>Gs -> B</c>,
+       where <c>Gs</c> is a guard sequence and <c>B</c> is a body, then
+       Rep(C) = <c>{clause,LINE,[],Rep(Gs),Rep(B)}</c>.</item>
     </list>
   </section>
 
   <section>
     <title>Guards</title>
-    <p>A guard sequence Gs is a sequence of guards <c><![CDATA[G_1; ...; G_k]]></c>, and
-      Rep(Gs) = <c><![CDATA[[Rep(G_1), ..., Rep(G_k)]]]></c>. If the guard sequence is
-      empty, Rep(Gs) = <c><![CDATA[[]]]></c>.</p>
-    <p>A guard G is a nonempty sequence of guard tests <c><![CDATA[Gt_1, ..., Gt_k]]></c>, and
-      Rep(G) = <c><![CDATA[[Rep(Gt_1), ..., Rep(Gt_k)]]]></c>.</p>
-    <p>A guard test <c><![CDATA[Gt]]></c> is one of the following alternatives:</p>
+    <p>A guard sequence Gs is a sequence of guards <c>G_1; ...; G_k</c>, and
+      Rep(Gs) = <c>[Rep(G_1), ..., Rep(G_k)]</c>. If the guard sequence is
+      empty, Rep(Gs) = <c>[]</c>.</p>
+    <p>A guard G is a nonempty sequence of guard tests
+      <c>Gt_1, ..., Gt_k</c>, and Rep(G) =
+      <c>[Rep(Gt_1), ..., Rep(Gt_k)]</c>.</p>
+    <p>A guard test <c>Gt</c> is one of the following alternatives:</p>
     <list type="bulleted">
-      <item>If Gt is an atomic literal L, then Rep(Gt) = Rep(L).</item>
-      <item>If Gt is a variable pattern <c><![CDATA[V]]></c>, then
-       Rep(Gt) = <c><![CDATA[{var,LINE,A}]]></c>,
-       where A is an atom with a printname consisting of the same characters as
-      <c><![CDATA[V]]></c>.</item>
-      <item>If Gt is a tuple skeleton <c><![CDATA[{Gt_1, ..., Gt_k}]]></c>, then
-       Rep(Gt) = <c><![CDATA[{tuple,LINE,[Rep(Gt_1), ..., Rep(Gt_k)]}]]></c>.</item>
-      <item>If Gt is <c><![CDATA[[]]]></c>, then
-       Rep(Gt) = <c><![CDATA[{nil,LINE}]]></c>.</item>
-      <item>If Gt is a cons skeleton <c><![CDATA[[Gt_h | Gt_t]]]></c>, then
-       Rep(Gt) = <c><![CDATA[{cons,LINE,Rep(Gt_h),Rep(Gt_t)}]]></c>.</item>
-      <item>If Gt is a binary constructor <c><![CDATA[<<Gt_1:Size_1/TSL_1, ..., Gt_k:Size_k/TSL_k>>]]></c>, then
-       Rep(Gt) = <c><![CDATA[{bin,LINE,[{bin_element,LINE,Rep(Gt_1),Rep(Size_1),Rep(TSL_1)}, ..., {bin_element,LINE,Rep(Gt_k),Rep(Size_k),Rep(TSL_k)}]}]]></c>.
+      <item>If Gt is an atomic literal <c>L</c>, then Rep(Gt) = Rep(L).</item>
+      <item>If Gt is a binary constructor
+       <c>&lt;&lt;Gt_1:Size_1/TSL_1, ..., Gt_k:Size_k/TSL_k>></c>,
+       where each <c>Size_i</c> is a guard test and each
+       <c>TSL_i</c> is a type specificer list, then
+       Rep(Gt) = <c>{bin,LINE,[{bin_element,LINE,Rep(Gt_1),Rep(Size_1),Rep(TSL_1)}, ..., {bin_element,LINE,Rep(Gt_k),Rep(Size_k),Rep(TSL_k)}]}</c>.
        For Rep(TSL), see above.
-       An omitted <c><![CDATA[Size]]></c> is represented by <c><![CDATA[default]]></c>. An omitted <c><![CDATA[TSL]]></c>
-       (type specifier list) is represented by <c><![CDATA[default]]></c>.</item>
-      <item>If Gt is <c><![CDATA[Gt_1 Op Gt_2]]></c>, where <c><![CDATA[Op]]></c>
-       is a binary operator, then Rep(Gt) = <c><![CDATA[{op,LINE,Op,Rep(Gt_1),Rep(Gt_2)}]]></c>.</item>
-      <item>If Gt is <c><![CDATA[Op Gt_0]]></c>, where <c><![CDATA[Op]]></c> is a unary operator, then
-       Rep(Gt) = <c><![CDATA[{op,LINE,Op,Rep(Gt_0)}]]></c>.</item>
-      <item>If Gt is <c><![CDATA[#Name{Field_1=Gt_1, ..., Field_k=Gt_k}]]></c>, then
-       Rep(E) =
-      <c><![CDATA[{record,LINE,Name,  [{record_field,LINE,Rep(Field_1),Rep(Gt_1)}, ..., {record_field,LINE,Rep(Field_k),Rep(Gt_k)}]}]]></c>.</item>
-      <item>If Gt is <c><![CDATA[#Name.Field]]></c>, then
-       Rep(Gt) = <c><![CDATA[{record_index,LINE,Name,Rep(Field)}]]></c>.</item>
-      <item>If Gt is <c><![CDATA[Gt_0#Name.Field]]></c>, then
-       Rep(Gt) = <c><![CDATA[{record_field,LINE,Rep(Gt_0),Name,Rep(Field)}]]></c>.</item>
-      <item>If Gt is <c><![CDATA[A(Gt_1, ..., Gt_k)]]></c>, where <c><![CDATA[A]]></c> is an atom, then
-       Rep(Gt) = <c><![CDATA[{call,LINE,Rep(A),[Rep(Gt_1), ..., Rep(Gt_k)]}]]></c>.</item>
-      <item>If Gt is <c><![CDATA[A_m:A(Gt_1, ..., Gt_k)]]></c>, where <c><![CDATA[A_m]]></c> is 
-       the atom <c><![CDATA[erlang]]></c> and <c><![CDATA[A]]></c> is an atom or an operator, then
-       Rep(Gt) = <c><![CDATA[{call,LINE,{remote,LINE,Rep(A_m),Rep(A)},[Rep(Gt_1), ..., Rep(Gt_k)]}]]></c>.</item>
-      <item>If Gt is <c><![CDATA[{A_m,A}(Gt_1, ..., Gt_k)]]></c>, where <c><![CDATA[A_m]]></c> is 
-       the atom <c><![CDATA[erlang]]></c> and <c><![CDATA[A]]></c> is an atom or an operator, then
-       Rep(Gt) = <c><![CDATA[{call,LINE,Rep({A_m,A}),[Rep(Gt_1), ..., Rep(Gt_k)]}]]></c>.</item>
-      <item>If Gt is <c><![CDATA[( Gt_0 )]]></c>, then
-       Rep(Gt) = <c><![CDATA[Rep(Gt_0)]]></c>,
-       i.e., parenthesized guard tests cannot be distinguished from their bodies.</item>
+       An omitted <c>Size_i</c> is represented by <c>default</c>.
+       An omitted <c>TSL_i</c> is represented by <c>default</c>.</item>
+      <item>If Gt is a cons skeleton <c>[Gt_h | Gt_t]</c>, then
+       Rep(Gt) = <c>{cons,LINE,Rep(Gt_h),Rep(Gt_t)}</c>.</item>
+      <item>If Gt is a function call <c>A(Gt_1, ..., Gt_k)</c>,
+       where <c>A</c> is an atom, then Rep(Gt) =
+       <c>{call,LINE,Rep(A),[Rep(Gt_1), ..., Rep(Gt_k)]}</c>.</item>
+      <item>If Gt is a function call <c>A_m:A(Gt_1, ..., Gt_k)</c>,
+       where <c>A_m</c> is the atom <c>erlang</c> and <c>A</c> is
+       an atom or an operator, then Rep(Gt) =
+       <c>{call,LINE,{remote,LINE,Rep(A_m),Rep(A)},[Rep(Gt_1), ..., Rep(Gt_k)]}</c>.</item>
+      <item>If Gt is a map creation <c>#{A_1, ..., A_k}</c>,
+       where each <c>A_i</c> is an association <c>Gt_i_1 => Gt_i_2</c>
+       or <c>Gt_i_1 := Gt_i_2</c>, then Rep(Gt) =
+       <c>{map,LINE,[Rep(A_1), ..., Rep(A_k)]}</c>. For Rep(A), see
+       above.</item>
+      <item>If Gt is a map update <c>Gt_0#{A_1, ..., A_k}</c>, where each
+       <c>A_i</c> is an association <c>Gt_i_1 => Gt_i_2</c>
+       or <c>Gt_i_1 := Gt_i_2</c>, then Rep(Gt) =
+       <c>{map,LINE,Rep(Gt_0),[Rep(A_1), ..., Rep(A_k)]}</c>.
+       For Rep(A), see above.</item>
+      <item>If Gt is nil, <c>[]</c>,
+       then Rep(Gt) = <c>{nil,LINE}</c>.</item>
+      <item>If Gt is an operator guard test <c>Gt_1 Op Gt_2</c>,
+       where <c>Op</c> is a binary operator other than the match
+       operator <c>=</c>, then
+       Rep(Gt) = <c>{op,LINE,Op,Rep(Gt_1),Rep(Gt_2)}</c>.</item>
+      <item>If Gt is an operator guard test <c>Op Gt_0</c>,
+       where <c>Op</c> is a unary operator, then
+       Rep(Gt) = <c>{op,LINE,Op,Rep(Gt_0)}</c>.</item>
+      <item>If Gt is a parenthesized guard test <c>( Gt_0 )</c>, then
+       Rep(Gt) = <c>Rep(Gt_0)</c>, that is, parenthesized
+       guard tests cannot be distinguished from their bodies.</item>
+      <item>If Gt is a record creation
+       <c>#Name{Field_1=Gt_1, ..., Field_k=Gt_k}</c>,
+       where each <c>Field_i</c> is an atom or <c>_</c>, then Rep(Gt) =
+      <c>{record,LINE,Name,[{record_field,LINE,Rep(Field_1),Rep(Gt_1)}, ..., {record_field,LINE,Rep(Field_k),Rep(Gt_k)}]}</c>.</item>
+      <item>If Gt is  a record field access <c>Gt_0#Name.Field</c>,
+       where <c>Field</c> is an atom, then
+       Rep(Gt) = <c>{record_field,LINE,Rep(Gt_0),Name,Rep(Field)}</c>.</item>
+      <item>If Gt is a record field index <c>#Name.Field</c>,
+       where <c>Field</c> is an atom, then
+       Rep(Gt) = <c>{record_index,LINE,Name,Rep(Field)}</c>.</item>
+      <item>If Gt is a tuple skeleton <c>{Gt_1, ..., Gt_k}</c>, then
+       Rep(Gt) = <c>{tuple,LINE,[Rep(Gt_1), ..., Rep(Gt_k)]}</c>.</item>
+      <item>If Gt is a variable pattern <c>V</c>, then
+       Rep(Gt) = <c>{var,LINE,A}</c>, where A is an atom with
+       a printname consisting of the same characters as <c>V</c>.</item>
     </list>
     <p>Note that every guard test has the same source form as some expression,
       and is represented the same way as the corresponding expression.</p>
   </section>
 
   <section>
-    <title>The abstract format after preprocessing</title>
-    <p>The compilation option <c><![CDATA[debug_info]]></c> can be given to the
-      compiler to have the abstract code stored in 
-      the <c><![CDATA[abstract_code]]></c> chunk in the BEAM file
+    <title>Types</title>
+    <list type="bulleted">
+      <item>If T is an annotated type <c>A :: T_0</c>,
+       where <c>A</c> is a variable, then Rep(T) =
+       <c>{ann_type,LINE,[Rep(A),Rep(T_0)]}</c>.</item>
+      <item>If T is an atom or integer literal L, then Rep(T) = Rep(L).
+      </item>
+      <item>If T is a bitstring type <c>&lt;&lt;_:M,_:_*N>></c>,
+       where <c>M</c> and <c>N</c> are singleton integer types, then Rep(T) =
+       <c>{type,LINE,binary,[Rep(M),Rep(N)]}</c>.</item>
+      <item>If T is the empty list type <c>[]</c>, then Rep(T) =
+       <c>{type,Line,nil,[]}</c>.</item>
+      <item>If T is a fun type <c>fun()</c>, then Rep(T) =
+       <c>{type,LINE,'fun',[]}</c>.</item>
+      <item>If T is a fun type <c>fun((...) -> T_0)</c>, then
+       Rep(T) = <c>{type,LINE,'fun',[{type,LINE,any},Rep(T_0)]}</c>.
+       </item>
+      <item>If T is a fun type <c>fun(Ft)</c>, where
+       <c>Ft</c> is a function type,
+       then Rep(T) = <c>Rep(Ft)</c>. For Rep(Ft), see below.</item>
+      <item>If T is an integer range type <c>L .. H</c>,
+       where <c>L</c> and <c>H</c> are singleton integer types, then
+       Rep(T) = <c>{type,LINE,range,[Rep(L),Rep(H)]}</c>.</item>
+      <item>If T is a map type <c>map()</c>, then Rep(T) =
+       <c>{type,LINE,map,any}</c>.</item>
+      <item>If T is a map type <c>#{A_1, ..., A_k}</c>, where each
+       <c>A_i</c> is an association type, then Rep(T) =
+       <c>{type,LINE,map,[Rep(A_1), ..., Rep(A_k)]}</c>.
+       For Rep(A), see below.</item>
+      <item>If T is an operator type <c>T_1 Op T_2</c>,
+       where <c>Op</c> is a binary operator (this is an occurrence of
+       an expression that can be evaluated to an integer at compile
+       time), then
+       Rep(T) = <c>{op,LINE,Op,Rep(T_1),Rep(T_2)}</c>.</item>
+      <item>If T is an operator type <c>Op T_0</c>, where <c>Op</c> is a
+       unary operator (this is an occurrence of
+       an expression that can be evaluated to an integer at compile time),
+       then Rep(T) = <c>{op,LINE,Op,Rep(T_0)}</c>.</item>
+      <item>If T is <c>( T_0 )</c>, then Rep(T) = <c>Rep(T_0)</c>,
+       that is, parenthesized types cannot be distinguished from their
+       bodies.</item>
+      <item>If T is a predefined (or built-in) type <c>N(T_1, ..., T_k)</c>,
+       then Rep(T) =
+       <c>{type,LINE,N,[Rep(T_1), ..., Rep(T_k)]}</c>.</item>
+      <item>If T is a record type <c>#Name{F_1, ..., F_k}</c>,
+       where each <c>F_i</c> is a record field type, then Rep(T) =
+       <c>{type,LINE,record,[Rep(Name),Rep(F_1), ..., Rep(F_k)]}</c>.
+       For Rep(F), see below.</item>
+      <item>If T is a remote type <c>M:N(T_1, ..., T_k)</c>, then Rep(T) =
+       <c>{remote_type,LINE,[Rep(M),Rep(N),[Rep(T_1), ..., Rep(T_k)]]}</c>.
+       </item>
+      <item>If T is a tuple type <c>tuple()</c>, then Rep(T) =
+       <c>{type,LINE,tuple,any}</c>.</item>
+      <item>If T is a tuple type <c>{T_1, ..., T_k}</c>, then Rep(T) =
+       <c>{type,LINE,tuple,[Rep(T_1), ..., Rep(T_k)]}</c>.</item>
+      <item>If T is a type union <c>T_1 | ... | T_k</c>, then Rep(T) =
+       <c>{type,LINE,union,[Rep(T_1), ..., Rep(T_k)]}</c>.</item>
+      <item>If T is a type variable <c>V</c>, then Rep(T) =
+       <c>{var,LINE,A}</c>, where <c>A</c> is an atom with a printname
+       consisting of the same characters as <c>V</c>. A type variable
+       is any variable except underscore (<c>_</c>).</item>
+      <item>If T is a user-defined type <c>N(T_1, ..., T_k)</c>,
+       then Rep(T) =
+       <c>{user_type,LINE,N,[Rep(T_1), ..., Rep(T_k)]}</c>.</item>
+    </list>
+
+    <section>
+      <title>Function Types</title>
+      <p>A function type Ft is one of the following alternatives:</p>
+      <list type="bulleted">
+       <item>If Ft is a constrained function type <c>Ft_1 when Fc</c>,
+	where <c>Ft_1</c> is a function type and
+	<c>Fc</c> is a function constraint, then Rep(T) =
+	<c>{type,LINE,bounded_fun,[Rep(Ft_1),Rep(Fc)]}</c>.
+        For Rep(Fc), see below.</item>
+       <item>If Ft is a function type <c>(T_1, ..., T_n) -> T_0</c>,
+        where each <c>T_i</c> is a type, then
+	Rep(Ft) = <c>{type,LINE,'fun',[{type,LINE,product,[Rep(T_1),
+	..., Rep(T_n)]},Rep(T_0)]}</c>.</item>
+      </list>
+    </section>
+
+    <section>
+      <title>Function Constraints</title>
+      <p>A function constraint Fc is a nonempty sequence of constraints
+	<c>C_1, ..., C_k</c>, and
+	Rep(Fc) = <c>[Rep(C_1), ..., Rep(C_k)]</c>.</p>
+      <list type="bulleted">
+       <item>If C is a constraint <c>is_subtype(V, T)</c> or <c>V :: T</c>,
+        where <c>V</c> is a type variable and <c>T</c> is a type, then
+        Rep(C) = <c>{type,LINE,constraint,[{atom,LINE,is_subtype},[Rep(V),Rep(T)]]}</c>.
+       </item>
+      </list>
+    </section>
+
+    <section>
+      <title>Association Types</title>
+      <list type="bulleted">
+       <item>If A is an association type <c>K => V</c>, where
+        <c>K</c> and <c>V</c> are types, then Rep(A) =
+        <c>{type,LINE,map_field_assoc,[Rep(K),Rep(V)]}</c>.</item>
+      </list>
+    </section>
+
+    <section>
+      <title>Record Field Types</title>
+      <list type="bulleted">
+       <item>If F is a record field type <c>Name :: Type</c>,
+	where <c>Type</c> is a type, then Rep(F) =
+	<c>{type,LINE,field_type,[Rep(Name),Rep(Type)]}</c>.</item>
+      </list>
+    </section>
+  </section>
+
+  <section>
+    <title>The Abstract Format After Preprocessing</title>
+    <p>The compilation option <c>debug_info</c> can be given to the
+      compiler to have the abstract code stored in
+      the <c>abstract_code</c> chunk in the BEAM file
       (for debugging purposes).</p>
-    <p>In OTP R9C and later, the <c><![CDATA[abstract_code]]></c> chunk will
+    <p>In OTP R9C and later, the <c>abstract_code</c> chunk will
       contain</p>
-    <p><c><![CDATA[{raw_abstract_v1,AbstractCode}]]></c></p>
-    <p>where <c><![CDATA[AbstractCode]]></c> is the abstract code as described
+    <p><c>{raw_abstract_v1,AbstractCode}</c></p>
+    <p>where <c>AbstractCode</c> is the abstract code as described
       in this document.</p>
     <p>In releases of OTP prior to R9C, the abstract code after some more
       processing was stored in the BEAM file. The first element of the
-      tuple would be either <c><![CDATA[abstract_v1]]></c> (R7B) or <c><![CDATA[abstract_v2]]></c>
+      tuple would be either <c>abstract_v1</c> (R7B) or <c>abstract_v2</c>
       (R8B).</p>
   </section>
 </chapter>
diff --git a/erts/doc/src/driver_entry.xml b/erts/doc/src/driver_entry.xml
index 30772c68fe..ae7f264d0c 100644
--- a/erts/doc/src/driver_entry.xml
+++ b/erts/doc/src/driver_entry.xml
@@ -4,7 +4,7 @@
 <cref>
   <header>
     <copyright>
-      <year>2001</year><year>2013</year>
+      <year>2001</year><year>2015</year>
       <holder>Ericsson AB. All Rights Reserved.</holder>
     </copyright>
     <legalnotice>
@@ -126,7 +126,7 @@
   <section>
     <title>DATA TYPES</title>
     <taglist>
-    <tag><b>ErlDrvEntry</b></tag>
+    <tag><em>ErlDrvEntry</em></tag>
     <item>
     <p/>
     <code type="none">
@@ -235,6 +235,7 @@ typedef struct erl_drv_entry {
       </item>
 
       <tag><marker id="ready_input"/>void (*ready_input)(ErlDrvData drv_data, ErlDrvEvent event)</tag>
+      <item/>
       <tag><marker id="ready_output"/>void (*ready_output)(ErlDrvData drv_data, ErlDrvEvent event)</tag>
       <item>
         <p>This is called when a driver event (given in the
@@ -246,14 +247,10 @@ typedef struct erl_drv_entry {
           something that the <c>WaitForMultipleObjects</c> API
           function understands). (Some trickery in the emulator allows
           more than the built-in limit of 64 <c>Events</c> to be used.)</p>
-	  <p>On Enea OSE the <c>event</c> is one or more signals that can
-	  be retrieved using <seealso marker="ose:ose_erl_driver#erl_drv_ose_get_signal">erl_drv_ose_get_signal</seealso>.</p>
         <p>To use this with threads and asynchronous routines, create a
-          pipe on unix, an Event on Windows or a unique signal number on
-	  Enea OSE. When the routine
+          pipe on unix and an Event on Windows. When the routine
           completes, write to the pipe (use <c>SetEvent</c> on
-          Windows or send a message to the emulator process on Enea OSE),
-	  this will make the emulator call
+          Windows), this will make the emulator call
           <c>ready_input</c> or <c>ready_output</c>.</p>
           <p>Spurious events may happen. That is, calls to <c>ready_input</c>
           or <c>ready_output</c> even though no real events are signaled. In
@@ -440,7 +437,14 @@ typedef struct erl_drv_entry {
 	  <seealso marker="erl_driver#erl_drv_busy_msgq_limits">erl_drv_busy_msgq_limits()</seealso>
 	  function.
           </item>
-        </taglist> 
+          <tag><c>ERL_DRV_FLAG_USE_INIT_ACK</c></tag>
+          <item>When this flag is given the linked-in driver has to manually
+          acknowledge that the port has been successfully started using 
+	  <seealso marker="erl_driver#erl_drv_init_ack">erl_drv_init_ack()</seealso>.
+          This allows the implementor to make the erlang:open_port exit with
+          badarg after some initial asynchronous initialization has been done.
+          </item>
+        </taglist>
       </item>
       <tag>void *handle2</tag>
       <item>
diff --git a/erts/doc/src/erl.xml b/erts/doc/src/erl.xml
index b0322b7d43..b6fa4c254c 100644
--- a/erts/doc/src/erl.xml
+++ b/erts/doc/src/erl.xml
@@ -4,7 +4,7 @@
 <comref>
   <header>
     <copyright>
-      <year>1996</year><year>2013</year>
+      <year>1996</year><year>2015</year>
       <holder>Ericsson AB. All Rights Reserved.</holder>
     </copyright>
     <legalnotice>
@@ -138,7 +138,7 @@
           see <seealso marker="kernel:app">app(4)</seealso> and
           <seealso marker="kernel:application">application(3)</seealso>.</p>
       </item>
-      <tag><marker id="args_file"><c><![CDATA[-args_file FileName]]></c></marker></tag>
+      <tag><marker id="args_file"/><c><![CDATA[-args_file FileName]]></c></tag>
       <item>
         <p>Command line arguments are read from the file <c><![CDATA[FileName]]></c>.
           The arguments read from the file replace the
@@ -203,7 +203,7 @@
           <seealso marker="kernel:app">app(4)</seealso> and
           <seealso marker="kernel:application">application(3)</seealso>.</p>
       </item>
-      <tag><marker id="connect_all"><c><![CDATA[-connect_all false]]></c></marker></tag>
+      <tag><marker id="connect_all"/><c><![CDATA[-connect_all false]]></c></tag>
       <item>
         <p>If this flag is present, <c><![CDATA[global]]></c> will not maintain a
           fully connected network of distributed Erlang nodes, and then
@@ -288,7 +288,7 @@
         <p>Makes <c><![CDATA[init]]></c> write some debug information while
           interpreting the boot script.</p>
       </item>
-      <tag><marker id="instr"><c><![CDATA[-instr]]></c>(emulator flag)</marker></tag>
+      <tag><marker id="instr"/><c><![CDATA[-instr]]></c>(emulator flag)</tag>
       <item>
         <p>Selects an instrumented Erlang runtime system (virtual
           machine) to run, instead of the ordinary one. When running an
@@ -371,7 +371,7 @@
           path, similar to <c><![CDATA[code:add_pathsa/1]]></c>. See
           <seealso marker="kernel:code">code(3)</seealso>.
 	  As an alternative to <c>-pa</c>, if several directories are
-	  to be prepended to the code and the directories have a
+	  to be prepended to the code path and the directories have a
 	  common parent directory, that parent directory could be
 	  specified in the <c>ERL_LIBS</c> environment variable.
 	  See <seealso marker="kernel:code">code(3)</seealso>.</p>
@@ -436,7 +436,7 @@
           flag and those running with the <c><![CDATA[-name]]></c> flag, as node
           names must be unique in distributed Erlang systems.</p>
       </item>
-      <tag><marker id="smp"><c><![CDATA[-smp [enable|auto|disable]]]></c></marker></tag>
+      <tag><marker id="smp"/><c><![CDATA[-smp [enable|auto|disable]]]></c></tag>
       <item>
         <p><c>-smp enable</c> and <c>-smp</c> starts the Erlang runtime
 	  system with SMP support enabled. This may fail if no runtime
@@ -462,7 +462,7 @@
     <p><c><![CDATA[erl]]></c> invokes the code for the Erlang emulator (virtual
       machine), which supports the following flags:</p>
     <taglist>
-      <tag><marker id="async_thread_stack_size"><c><![CDATA[+a size]]></c></marker></tag>
+      <tag><marker id="async_thread_stack_size"/><c><![CDATA[+a size]]></c></tag>
       <item>
         <p>Suggested stack size, in kilowords, for threads in the
           async-thread pool. Valid range is 16-8192 kilowords. The
@@ -477,7 +477,7 @@
           suggestion, and it might even be ignored on some
           platforms.</p>
       </item>
-      <tag><marker id="async_thread_pool_size"><c><![CDATA[+A size]]></c></marker></tag>
+      <tag><marker id="async_thread_pool_size"/><c><![CDATA[+A size]]></c></tag>
       <item>
         <p>Sets the number of threads in async thread pool, valid range
           is 0-1024. If thread support is available, the default is 10.</p>
@@ -496,7 +496,7 @@
           <c><![CDATA[werl]]></c>, not <c><![CDATA[erl]]></c> (<c><![CDATA[oldshell]]></c>). Note also that
           <c><![CDATA[Ctrl-Break]]></c> is used instead of <c><![CDATA[Ctrl-C]]></c> on Windows.</p>
       </item>
-      <tag><marker id="+c"><c><![CDATA[+c true | false]]></c></marker></tag>
+      <tag><marker id="+c"/><c><![CDATA[+c true | false]]></c></tag>
       <item>
         <p>Enable or disable
 	<seealso marker="time_correction#Time_Correction">time correction</seealso>:</p>
@@ -512,7 +512,7 @@
 	This is interpreted as <c>+c false</c>.
 	</p>
       </item>
-      <tag><marker id="+C_"><c><![CDATA[+C no_time_warp | single_time_warp | multi_time_warp]]></c></marker></tag>
+      <tag><marker id="+C_"/><c><![CDATA[+C no_time_warp | single_time_warp | multi_time_warp]]></c></tag>
       <item>
         <p>Set
 	<seealso marker="time_correction#Time_Warp_Modes">time warp mode</seealso>:
@@ -540,7 +540,7 @@
 	produce a crash dump. On Unix systems, sending an emulator process
 	a SIGUSR1 signal will also force a crash dump.</p>
       </item>
-      <tag><marker id="+e"><c><![CDATA[+e Number]]></c></marker></tag>
+      <tag><marker id="+e"/><c><![CDATA[+e Number]]></c></tag>
       <item>
         <p>Set max number of ETS tables.</p>
       </item>
@@ -625,7 +625,7 @@
 	  information about the file names and line numbers.
           </p>
       </item>
-      <tag><marker id="erts_alloc"><c><![CDATA[+MFlag Value]]></c></marker></tag>
+      <tag><marker id="erts_alloc"/><c><![CDATA[+MFlag Value]]></c></tag>
       <item>
         <p>Memory allocator specific flags, see
           <seealso marker="erts_alloc">erts_alloc(3)</seealso> for
@@ -664,10 +664,10 @@
 	     debugging.</item>
 	   </taglist>
       </item>
-      <tag><marker id="+pc"/><marker id="printable_character_range"><c><![CDATA[+pc Range]]></c></marker></tag>
+      <tag><marker id="+pc"/><marker id="printable_character_range"/><c><![CDATA[+pc Range]]></c></tag>
       <item>
 	<p>Sets the range of characters that the system will consider printable in heuristic detection of strings. This typically affects the shell, debugger and io:format functions (when ~tp is used in the format string).</p> 
-	<p>Currently two values for the <c>Range</c> are supported:
+	<p>Currently two values for the <c>Range</c> are supported:</p>
 	<taglist>
 	  <tag><c>latin1</c></tag> <item>The default. Only characters
 	  in the ISO-latin-1 range can be considered printable, which means
@@ -682,11 +682,10 @@
 	  example your font does not cover all Unicode
 	  characters.</item>
 	</taglist>
-	</p>
 	<p>Se also <seealso marker="stdlib:io#printable_range/0">
 	io:printable_range/0</seealso>.</p>
       </item>
-      <tag><marker id="+P"/><marker id="max_processes"><c><![CDATA[+P Number|legacy]]></c></marker></tag>
+      <tag><marker id="+P"/><marker id="max_processes"/><c><![CDATA[+P Number|legacy]]></c></tag>
       <item>
 	<p>Sets the maximum number of simultaneously existing processes for this
           system if a <c>Number</c> is passed as value. Valid range for
@@ -706,7 +705,7 @@
 	circumstances be extremely expensive. The legacy algoritm is deprecated,
 	and the <c>legacy</c> option is scheduled for removal in OTP-R18.</p>
       </item>
-      <tag><marker id="+Q"/><marker id="max_ports"><c><![CDATA[+Q Number|legacy]]></c></marker></tag>
+      <tag><marker id="+Q"/><marker id="max_ports"/><c><![CDATA[+Q Number|legacy]]></c></tag>
       <item>
 	<p>Sets the maximum number of simultaneously existing ports for this
           system if a Number is passed as value. Valid range for <c>Number</c>
@@ -737,7 +736,7 @@
 	circumstances be extremely expensive. The legacy algoritm is deprecated,
 	and the <c>legacy</c> option is scheduled for removal in OTP-R18.</p>
       </item>
-      <tag><marker id="compat_rel"><c><![CDATA[+R ReleaseNumber]]></c></marker></tag>
+      <tag><marker id="compat_rel"/><c><![CDATA[+R ReleaseNumber]]></c></tag>
       <item>
         <p>Sets the compatibility mode.</p>
         <p>The distribution mechanism is not backwards compatible by
@@ -757,7 +756,7 @@
       <item>
         <p>Force ets memory block to be moved on realloc.</p>
       </item>
-      <tag><marker id="+rg"><c><![CDATA[+rg ReaderGroupsLimit]]></c></marker></tag>
+      <tag><marker id="+rg"/><c><![CDATA[+rg ReaderGroupsLimit]]></c></tag>
       <item>
         <p>Limits the amount of reader groups used by read/write locks
 	   optimized for read operations in the Erlang runtime system. By
@@ -775,7 +774,7 @@
 	   schedulers to logical processors</seealso>, since the reader groups
 	   are distributed better between schedulers.</p>
       </item>
-      <tag><marker id="+S"><c><![CDATA[+S Schedulers:SchedulerOnline]]></c></marker></tag>
+      <tag><marker id="+S"/><c><![CDATA[+S Schedulers:SchedulerOnline]]></c></tag>
       <item>
         <p>Sets the number of scheduler threads to create and scheduler
 	  threads to set online when SMP support has been enabled. The maximum for
@@ -800,7 +799,7 @@
           SMP support enabled (see the <seealso marker="#smp">-smp</seealso>
           flag).</p>
       </item>
-      <tag><marker id="+SP"><c><![CDATA[+SP SchedulersPercentage:SchedulersOnlinePercentage]]></c></marker></tag>
+      <tag><marker id="+SP"/><c><![CDATA[+SP SchedulersPercentage:SchedulersOnlinePercentage]]></c></tag>
       <item>
         <p>Similar to <seealso marker="#+S">+S</seealso> but uses percentages to set the
 	  number of scheduler threads to create, based on logical processors configured,
@@ -821,7 +820,7 @@
           SMP support enabled (see the <seealso marker="#smp">-smp</seealso>
           flag).</p>
       </item>
-      <tag><marker id="+SDcpu"><c><![CDATA[+SDcpu DirtyCPUSchedulers:DirtyCPUSchedulersOnline]]></c></marker></tag>
+      <tag><marker id="+SDcpu"/><c><![CDATA[+SDcpu DirtyCPUSchedulers:DirtyCPUSchedulersOnline]]></c></tag>
       <item>
           <p>Sets the number of dirty CPU scheduler threads to create and dirty
             CPU scheduler threads to set online when threading support has been
@@ -845,7 +844,7 @@
             enabled (it's disabled by default).
           </p>
       </item>
-      <tag><marker id="+SDPcpu"><c><![CDATA[+SDPcpu DirtyCPUSchedulersPercentage:DirtyCPUSchedulersOnlinePercentage]]></c></marker></tag>
+      <tag><marker id="+SDPcpu"/><c><![CDATA[+SDPcpu DirtyCPUSchedulersPercentage:DirtyCPUSchedulersOnlinePercentage]]></c></tag>
       <item>
           <p>Similar to <seealso marker="#+SDcpu">+SDcpu</seealso> but uses percentages to set the
             number of dirty CPU scheduler threads to create and number of dirty CPU scheduler threads
@@ -868,7 +867,7 @@
             enabled (it's disabled by default).
           </p>
       </item>
-      <tag><marker id="+SDio"><c><![CDATA[+SDio IOSchedulers]]></c></marker></tag>
+      <tag><marker id="+SDio"/><c><![CDATA[+SDio IOSchedulers]]></c></tag>
       <item>
           <p>Sets the number of dirty I/O scheduler threads to create when threading
             support has been enabled. The valid range is 0-1024. By default, the number
@@ -886,7 +885,7 @@
       <item>
         <p>Scheduling specific flags.</p>
         <taglist>
-          <tag><marker id="+sbt"><c>+sbt BindType</c></marker></tag>
+          <tag><marker id="+sbt"/><c>+sbt BindType</c></tag>
           <item>
             <p>Set scheduler bind type.</p>
 	    <p>Schedulers can also be bound using the
@@ -1010,7 +1009,7 @@
 	       <seealso marker="erlang#system_info_scheduler_bindings">erlang:system_info(scheduler_bindings)</seealso>.
 	    </p>
           </item>
-	  <tag><marker id="+sbwt"><c>+sbwt none|very_short|short|medium|long|very_long</c></marker></tag>
+	  <tag><marker id="+sbwt"/><c>+sbwt none|very_short|short|medium|long|very_long</c></tag>
 	  <item>
              <p>Set scheduler busy wait threshold. Default is <c>medium</c>.
 	        The threshold determines how long schedulers should busy
@@ -1020,7 +1019,7 @@
 	        without prior notice.
 	     </p>
 	  </item>
-          <tag><marker id="+scl"><c>+scl true|false</c></marker></tag>
+          <tag><marker id="+scl"/><c>+scl true|false</c></tag>
           <item>
 	    <p>Enable or disable scheduler compaction of load. By default
 	    scheduler compaction of load is enabled. When enabled, load
@@ -1037,7 +1036,7 @@
 	    between schedulers.
 	    </p>
           </item>
-          <tag><marker id="+sct"><c>+sct CpuTopology</c></marker></tag>
+          <tag><marker id="+sct"/><c>+sct CpuTopology</c></tag>
           <item>
             <list type="bulleted">
               <item><c><![CDATA[<Id> = integer(); when 0 =< <Id> =< 65535]]></c></item>
@@ -1159,7 +1158,7 @@
 	    <p>For more information, see
 	    <seealso marker="erlang#system_info_cpu_topology">erlang:system_info(cpu_topology)</seealso>.</p>
           </item>
-          <tag><marker id="+secio"><c>+secio true|false</c></marker></tag>
+          <tag><marker id="+secio"/><c>+secio true|false</c></tag>
           <item>
 	    <p>Enable or disable eager check I/O scheduling. The default
 	    is currently <c>true</c>. The default was changed from <c>false</c>
@@ -1176,7 +1175,7 @@
 	    <p><seealso marker="erlang#system_info_eager_check_io"><c>erlang:system_info(eager_check_io)</c></seealso>
 	    returns the value of this parameter used when starting the VM.</p>
           </item>
-          <tag><marker id="+sfwi"><c>+sfwi Interval</c></marker></tag>
+          <tag><marker id="+sfwi"/><c>+sfwi Interval</c></tag>
           <item>
             <p>Set scheduler forced wakeup interval. All run queues will
 	    be scanned each <c>Interval</c> milliseconds. While there are
@@ -1185,12 +1184,12 @@
 	    disables this feature, which also is the default.
 	    </p>
 	    <p>This feature has been introduced as a temporary workaround
-	    for lengthy executing native code, and native code that do not
+	    for long-executing native code, and native code that does not
 	    bump reductions properly in OTP. When these bugs have be fixed
 	    the <c>+sfwi</c> flag will be removed.
 	    </p>
 	  </item>
-          <tag><marker id="+stbt"><c>+stbt BindType</c></marker></tag>
+          <tag><marker id="+stbt"/><c>+stbt BindType</c></tag>
           <item>
             <p>Try to set scheduler bind type. The same as the
 	    <seealso marker="#+sbt">+sbt</seealso> flag with the exception of
@@ -1198,7 +1197,7 @@
 	    documentation of the <seealso marker="#+sbt">+sbt</seealso> flag.
 	    </p>
 	  </item>
-          <tag><marker id="+sub"><c>+sub true|false</c></marker></tag>
+          <tag><marker id="+sub"/><c>+sub true|false</c></tag>
           <item>
 	    <p>Enable or disable
 	    <seealso marker="erts:erlang#statistics_scheduler_wall_time">scheduler
@@ -1211,7 +1210,7 @@
 	    balance scheduler utilization between schedulers. That is,
 	    strive for equal scheduler utilization on all schedulers.
 	    <br/>&nbsp;&nbsp;&nbsp;<c>+sub true</c> is only supported on
-	    systems where the runtime system detects and use a monotonically
+      systems where the runtime system detects and uses a monotonically
 	    increasing high resolution clock. On other systems, the runtime
 	    system will fail to start.
 	    <br/>&nbsp;&nbsp;&nbsp;<c>+sub true</c> implies
@@ -1221,7 +1220,7 @@
 	    utilization.
 	    </p>
           </item>
-	  <tag><marker id="+swct"><c>+swct very_eager|eager|medium|lazy|very_lazy</c></marker></tag>
+	  <tag><marker id="+swct"/><c>+swct very_eager|eager|medium|lazy|very_lazy</c></tag>
 	  <item>
              <p>
 	       Set scheduler wake cleanup threshold. Default is <c>medium</c>.
@@ -1235,7 +1234,7 @@
              <p><em>NOTE:</em> This flag may be removed or changed at any time without prior notice.
 	     </p>
 	  </item>
-	  <tag><marker id="+sws"><c>+sws default|legacy</c></marker></tag>
+	  <tag><marker id="+sws"/><c>+sws default|legacy</c></tag>
 	  <item>
              <p>
 	       Set scheduler wakeup strategy. Default strategy changed in erts-5.10/OTP-R16A. This strategy was previously known as <c>proposal</c> in OTP-R15. The <c>legacy</c> strategy was used as default from R13 up to and including R15.
@@ -1243,7 +1242,7 @@
              <p><em>NOTE:</em> This flag may be removed or changed at any time without prior notice.
 	     </p>
 	  </item>
-	  <tag><marker id="+swt"><c>+swt very_low|low|medium|high|very_high</c></marker></tag>
+	  <tag><marker id="+swt"/><c>+swt very_low|low|medium|high|very_high</c></tag>
 	  <item>
              <p>Set scheduler wakeup threshold. Default is <c>medium</c>.
 	        The threshold determines when to wake up sleeping schedulers
@@ -1257,7 +1256,7 @@
 	        without prior notice.
 	     </p>
 	  </item>
-          <tag><marker id="+spp"><c>+spp Bool</c></marker></tag>
+          <tag><marker id="+spp"/><c>+spp Bool</c></tag>
           <item>
 	    <p>Set default scheduler hint for port parallelism. If set to
 	    <c>true</c>, the VM will schedule port tasks when doing so will
@@ -1273,7 +1272,7 @@
 	    option to <seealso
 	    marker="erlang#open_port/2">open_port/2</seealso></p>.
 	  </item>
-	  <tag><marker id="sched_thread_stack_size"><c><![CDATA[+sss size]]></c></marker></tag>
+	  <tag><marker id="sched_thread_stack_size"/><c><![CDATA[+sss size]]></c></tag>
 	  <item>
 	    <p>Suggested stack size, in kilowords, for scheduler threads.
             Valid range is 4-8192 kilowords. The default stack size
@@ -1281,11 +1280,11 @@
 	  </item>
         </taglist>
       </item>
-      <tag><marker id="+t"><c><![CDATA[+t size]]></c></marker></tag>
+      <tag><marker id="+t"/><c><![CDATA[+t size]]></c></tag>
       <item>
         <p>Set the maximum number of atoms the VM can handle. Default is 1048576.</p>
       </item>
-      <tag><marker id="+T"><c><![CDATA[+T Level]]></c></marker></tag>
+      <tag><marker id="+T"/><c><![CDATA[+T Level]]></c></tag>
       <item>
         <p>Enables modified timing and sets the modified timing level.
           Currently valid range is 0-9. The timing of the runtime system
@@ -1335,11 +1334,26 @@
           <seealso marker="kernel:error_logger#warning_map/0">error_logger(3)</seealso>
           for further information.</p>
       </item>
+      <tag><c><![CDATA[+xFlag Value]]></c></tag>
+      <item>
+        <p>Default process flag settings.</p>
+        <taglist>
+          <tag><marker id="+xmqd"><c>+xmqd off_heap|on_heap|mixed</c></marker></tag>
+          <item><p>
+	    Sets the default value for the process flag
+	    <c>message_queue_data</c>. If <c>+xmqd</c> is not
+	    passed, <c>mixed</c> will be the default. For more information,
+	    see the documentation of
+            <seealso marker="erlang#process_flag_message_queue_data"><c>process_flag(message_queue_data,
+	    MQD)</c></seealso>.
+	  </p></item>
+	</taglist>
+      </item>
       <tag><c><![CDATA[+zFlag Value]]></c></tag>
       <item>
         <p>Miscellaneous flags.</p>
         <taglist>
-          <tag><marker id="+zdbbl"><c>+zdbbl size</c></marker></tag>
+          <tag><marker id="+zdbbl"/><c>+zdbbl size</c></tag>
           <item>
             <p>Set the distribution buffer busy limit
 	    (<seealso marker="erlang#system_info_dist_buf_busy_limit">dist_buf_busy_limit</seealso>)
@@ -1352,7 +1366,7 @@
             give lower latency and higher throughput at the expense
             of higher memory usage.</p>
           </item>
-          <tag><marker id="+zdntgc"><c>+zdntgc time</c></marker></tag>
+          <tag><marker id="+zdntgc"/><c>+zdntgc time</c></tag>
           <item>
             <p>Set the delayed node table garbage collection time
 	    (<seealso marker="erlang#system_info_delayed_node_table_gc">delayed_node_table_gc</seealso>)
@@ -1426,7 +1440,7 @@
 		  </item>
 	  </taglist>
       </item>
-      <tag><marker id="ERL_AFLAGS"><c><![CDATA[ERL_AFLAGS]]></c></marker></tag>
+      <tag><marker id="ERL_AFLAGS"/><c><![CDATA[ERL_AFLAGS]]></c></tag>
       <item>
         <p>The content of this environment variable will be added to the
           beginning of the command line for <c><![CDATA[erl]]></c>.</p>
@@ -1436,7 +1450,7 @@
           the <c><![CDATA[-extra]]></c> section, i.e. the end of the command line
           following after an <c><![CDATA[-extra]]></c> flag.</p>
       </item>
-      <tag><marker id="ERL_ZFLAGS"><c><![CDATA[ERL_ZFLAGS]]></c></marker> and <marker id="ERL_FLAGS"><c><![CDATA[ERL_FLAGS]]></c></marker></tag>
+      <tag><marker id="ERL_ZFLAGS"/><c><![CDATA[ERL_ZFLAGS]]></c> and <marker id="ERL_FLAGS"/><c><![CDATA[ERL_FLAGS]]></c></tag>
       <item>
         <p>The content of these environment variables will be added to the
           end of the command line for <c><![CDATA[erl]]></c>.</p>
diff --git a/erts/doc/src/erl_dist_protocol.xml b/erts/doc/src/erl_dist_protocol.xml
index e1a58856f3..b435d5c9b4 100644
--- a/erts/doc/src/erl_dist_protocol.xml
+++ b/erts/doc/src/erl_dist_protocol.xml
@@ -5,7 +5,7 @@
   <header>
     <copyright>
       <year>2007</year>
-      <year>2013</year>
+      <year>2015</year>
       <holder>Ericsson AB, All Rights Reserved</holder>
     </copyright>
     <legalnotice>
@@ -549,10 +549,10 @@ If Result > 0, the packet only consists of [119, Result].
 -->
 
     </section>
-    <marker id="distribution_handshake"/>
       <section>
 	<title>Distribution Handshake</title>
 	  <p>
+            <marker id="distribution_handshake"/>
 	    This section describes the distribution handshake protocol
 	    introduced in the OTP-R6 release of Erlang/OTP. This
 	    description was previously located in
diff --git a/erts/doc/src/erl_driver.xml b/erts/doc/src/erl_driver.xml
index 1f7fe0f961..cade732c56 100644
--- a/erts/doc/src/erl_driver.xml
+++ b/erts/doc/src/erl_driver.xml
@@ -4,7 +4,7 @@
 <cref>
   <header>
     <copyright>
-      <year>2001</year><year>2014</year>
+      <year>2001</year><year>2015</year>
       <holder>Ericsson AB. All Rights Reserved.</holder>
     </copyright>
     <legalnotice>
@@ -223,7 +223,7 @@
        asynchronous function calls, using a thread pool provided by
        Erlang. There is also a select call, that can be used for
        asynchronous drivers.</item>
-      <tag><marker id="multi_threading">Multi-threading</marker></tag>
+      <tag><marker id="multi_threading"/>Multi-threading</tag>
       <item>
       <p>A POSIX thread like API for multi-threading is provided. The
          Erlang driver thread API only provide a subset of the functionality
@@ -297,7 +297,7 @@
       <item><p>A driver can add and later remove drivers.</p></item>
       <tag>Monitoring processes</tag>
       <item><p>A driver can monitor a process that does not own a port.</p></item>
-      <tag><marker id="version_management">Version management</marker></tag>
+      <tag><marker id="version_management"/>Version management</tag>
       <item>
         <p>Version management is enabled for drivers that have set the
           <seealso marker="driver_entry#extended_marker">extended_marker</seealso>
@@ -384,12 +384,12 @@
       <item>
 	<p>
 	  Rewrite driver callback
-	  <c><seealso marker="driver_entry#control">control</seealso></c>
+	  <seealso marker="driver_entry#control"><c>control</c></seealso>
 	  to use return type <c>ErlDrvSSizeT</c> instead of <c>int</c>.
 	</p>
 	<p>
 	  Rewrite driver callback
-	  <c><seealso marker="driver_entry#call">call</seealso></c>
+	  <seealso marker="driver_entry#call"><c>call</c></seealso>
 	  to use return type <c>ErlDrvSSizeT</c> instead of <c>int</c>.
 	</p>
 	<note>
@@ -407,19 +407,19 @@
       <item>
 	<p>
 	  Driver callback
-	  <c><seealso marker="driver_entry#output">output</seealso></c>
+	  <seealso marker="driver_entry#output"><c>output</c></seealso>
 	  now gets <c>ErlDrvSizeT</c> as 3rd argument instead
 	  of previously <c>int</c>.
 	</p>
 	<p>
 	  Driver callback
-	  <c><seealso marker="driver_entry#control">control</seealso></c>
+	  <seealso marker="driver_entry#control"><c>control</c></seealso>
 	  now gets <c>ErlDrvSizeT</c> as 4th and 6th arguments instead
 	  of previously <c>int</c>.
 	</p>
 	<p>
 	  Driver callback
-	  <c><seealso marker="driver_entry#call">call</seealso></c>
+	  <seealso marker="driver_entry#call"><c>call</c></seealso>
 	  now gets <c>ErlDrvSizeT</c> as 4th and 6th arguments instead
 	  of previously <c>int</c>.
 	</p>
@@ -1044,9 +1044,7 @@ typedef struct ErlIOVec {
           <c>select</c>/<c>poll</c> can use).
           On windows, the Win32 API function <c>WaitForMultipleObjects</c>
           is used. This places other restrictions on the event object.
-          Refer to the Win32 SDK documentation.
-	  On Enea OSE, the receive function is used. See the <seealso
-	  marker="ose:ose_erl_driver"></seealso> for more details.</p>
+          Refer to the Win32 SDK documentation.</p>
         <p>The <c>on</c> parameter should be <c>1</c> for setting events
           and <c>0</c> for clearing them.</p>
         <p>The <c>mode</c> argument is a bitwise-or combination of
@@ -1058,7 +1056,7 @@ typedef struct ErlIOVec {
           <seealso marker="driver_entry#ready_output">ready_output</seealso>.
           </p>
     <note>
-      <p>Some OS (Windows and Enea OSE) do not differentiate between read and write events.
+      <p>Some OS (Windows) do not differentiate between read and write events.
          The call-back for a fired event then only depends on the value of <c>mode</c>.</p>
     </note>
         <p><c>ERL_DRV_USE</c> specifies if we are using the event object or if we want to close it.
@@ -2133,6 +2131,53 @@ ERL_DRV_MAP          int sz
     </func>
 
     <func>
+      <name><ret>void</ret><nametext>erl_drv_init_ack(ErlDrvPort port, ErlDrvData res)</nametext></name>
+      <fsummary>Acknowledge the start of the port</fsummary>
+      <desc>
+        <marker id="erl_drv_init_ack"></marker>
+        <p>Arguments:</p>
+        <taglist>
+          <tag><c>port</c></tag>
+          <item>The port handle of the port (driver instance) creating
+           doing the acknowledgment.
+	  </item>
+          <tag><c>res</c></tag>
+          <item>The result of the port initialization. This can be the same values
+          as the return value of <seealso marker="driver_entry#start">start</seealso>,
+          i.e any of the error codes or the ErlDrvData that is to be used for this
+          port.
+          </item>
+        </taglist>
+        <p>
+          When this function is called the initiating erlang:open_port call is
+          returned as if the <seealso marker="driver_entry#start">start</seealso>
+          function had just been called. It can only be used when the
+          <seealso marker="driver_entry#driver_flags">ERL_DRV_FLAG_USE_INIT_ACK</seealso>
+          flag has been set on the linked-in driver.
+        </p>
+      </desc>
+    </func>
+
+    <func>
+      <name><ret>void</ret><nametext>erl_drv_set_os_pid(ErlDrvPort port, ErlDrvSInt pid)</nametext></name>
+      <fsummary>Set the os_pid for the port</fsummary>
+      <desc>
+        <marker id="erl_drv_set_os_pid"></marker>
+        <p>Arguments:</p>
+        <taglist>
+          <tag><c>port</c></tag>
+          <item>The port handle of the port (driver instance) to set the pid on.
+	  </item>
+          <tag><c>pid</c></tag>
+          <item>The pid to set.</item>
+        </taglist>
+        <p>
+          Set the os_pid seen when doing erlang:port_info/2 on this port.
+        </p>
+      </desc>
+    </func>
+
+    <func>
       <name><ret>int</ret><nametext>erl_drv_thread_create(char *name,
 				      ErlDrvTid *tid,
 				      void * (*func)(void *),
@@ -2811,7 +2856,7 @@ ERL_DRV_MAP          int sz
     </func>
 
     <func>
-      <name><ret>int</ret><nametext>erl_drv_putenv(char *key, char *value)</nametext></name>
+      <name><ret>int</ret><nametext>erl_drv_putenv(const char *key, char *value)</nametext></name>
       <fsummary>Set the value of an environment variable</fsummary>
       <desc>
         <marker id="erl_drv_putenv"></marker>
@@ -2840,7 +2885,7 @@ ERL_DRV_MAP          int sz
       </desc>
     </func>
     <func>
-      <name><ret>int</ret><nametext>erl_drv_getenv(char *key, char *value, size_t *value_size)</nametext></name>
+      <name><ret>int</ret><nametext>erl_drv_getenv(const char *key, char *value, size_t *value_size)</nametext></name>
       <fsummary>Get the value of an environment variable</fsummary>
       <desc>
         <marker id="erl_drv_getenv"></marker>
diff --git a/erts/doc/src/erl_ext_dist.xml b/erts/doc/src/erl_ext_dist.xml
index caf1e812c4..2ac974f497 100644
--- a/erts/doc/src/erl_ext_dist.xml
+++ b/erts/doc/src/erl_ext_dist.xml
@@ -5,7 +5,7 @@
   <header>
     <copyright>
       <year>2007</year>
-      <year>2014</year>
+      <year>2015</year>
       <holder>Ericsson AB, All Rights Reserved</holder>
     </copyright>
     <legalnotice>
@@ -150,10 +150,10 @@
     </note>
   </section>
 
-  <marker id="distribution_header"/>
   <section>  
     <title>Distribution header</title>
     <p>
+      <marker id="distribution_header"/>
       As of erts version 5.7.2 the old atom cache protocol was
       dropped and a new one was introduced. This atom cache protocol
       introduced the distribution header. Nodes with erts versions
diff --git a/erts/doc/src/erl_nif.xml b/erts/doc/src/erl_nif.xml
index 23c3d5fcee..2d8706169f 100644
--- a/erts/doc/src/erl_nif.xml
+++ b/erts/doc/src/erl_nif.xml
@@ -4,7 +4,7 @@
 <cref>
   <header>
     <copyright>
-      <year>2001</year><year>2013</year>
+      <year>2001</year><year>2015</year>
       <holder>Ericsson AB. All Rights Reserved.</holder>
     </copyright>
     <legalnotice>
@@ -791,6 +791,10 @@ typedef enum {
       and return true, or return false if <c>term</c> is not an unsigned integer or is
       outside the bounds of type <c>unsigned long</c>.</p></desc>
     </func>
+    <func><name><ret>int</ret><nametext>enif_getenv(const char* key, char* value, size_t *value_size)</nametext></name>
+      <fsummary>Get the value of an environment variable</fsummary>
+      <desc><p>Same as <seealso marker="erl_driver#erl_drv_getenv">erl_drv_getenv</seealso>.</p></desc>
+    </func>
     <func><name><ret>int</ret><nametext>enif_has_pending_exception(ErlNifEnv* env, ERL_NIF_TERM* reason)</nametext></name>
       <fsummary>Check if an exception has been raised</fsummary>
       <desc><p>Return true if a pending exception is associated
@@ -833,9 +837,10 @@ typedef enum {
       <fsummary>Determine if a term is an empty list</fsummary>
       <desc><p>Return true if <c>term</c> is an empty list.</p></desc>
     </func>
-    <marker id="enif_is_exception"/><func><name><ret>int</ret><nametext>enif_is_exception(ErlNifEnv* env, ERL_NIF_TERM term)</nametext></name>
+    <func><name><ret>int</ret><nametext>enif_is_exception(ErlNifEnv* env, ERL_NIF_TERM term)</nametext></name>
       <fsummary>Determine if a term is an exception</fsummary>
-      <desc><p>Return true if <c>term</c> is an exception.</p></desc>
+      <desc><marker id="enif_is_exception"/>
+        <p>Return true if <c>term</c> is an exception.</p></desc>
     </func>
     <func><name><ret>int</ret><nametext>enif_is_map(ErlNifEnv* env, ERL_NIF_TERM term)</nametext></name>
       <fsummary>Determine if a term is a map</fsummary>
diff --git a/erts/doc/src/erl_prim_loader.xml b/erts/doc/src/erl_prim_loader.xml
index d05f0d9aea..8f66e07ae1 100644
--- a/erts/doc/src/erl_prim_loader.xml
+++ b/erts/doc/src/erl_prim_loader.xml
@@ -36,17 +36,11 @@
       the system. The start script is also fetched with this low level
       loader.</p>
     <p><c>erl_prim_loader</c> knows about the environment and how to
-      fetch modules. The loader could, for example, fetch files using
-      the file system (with absolute file names as input), or a
-      database (where the binary format of a module is stored).</p>
+      fetch modules.</p>
     <p>The <c>-loader Loader</c> command line flag can be used to
       choose the method used by the <c>erl_prim_loader</c>. Two
       <c>Loader</c> methods are supported by the Erlang runtime system:
-      <c>efile</c> and <c>inet</c>. If another loader is required, then
-      it has to be implemented by the user. The <c>Loader</c> provided
-      by the user must fulfill the protocol defined below, and it is
-      started with the <c>erl_prim_loader</c> by evaluating
-      <c>open_port({spawn,Loader},[binary])</c>.</p>
+      <c>efile</c> and <c>inet</c>.</p>
 
     <warning><p>The support for loading of code from archive files is
      experimental. The sole purpose of releasing it before it is ready
@@ -56,39 +50,9 @@
      <c>-loader_debug</c> are also experimental</p></warning>
 
   </description>
-  <datatypes>
-    <datatype>
-      <name name="host"/>
-    </datatype>
-  </datatypes>
 
   <funcs>
     <func>
-      <name name="start" arity="3"/>
-      <fsummary>Start the Erlang low level loader</fsummary>
-      <desc>
-        <p>Starts the Erlang low level loader. This function is called
-          by the <c>init</c> process (and module). The <c>init</c>
-          process reads the command line flags <c>-id <anno>Id</anno></c>,
-          <c>-loader <anno>Loader</anno></c>, and <c>-hosts <anno>Hosts</anno></c>. These are
-          the arguments supplied to the <c>start/3</c> function.</p>
-        <p>If <c>-loader</c> is not given, the default loader is
-          <c>efile</c> which tells the system to read from the file
-          system.</p>
-        <p>If <c>-loader</c> is <c>inet</c>, the <c>-id <anno>Id</anno></c>,
-          <c>-hosts <anno>Hosts</anno></c>, and <c>-setcookie Cookie</c> flags must
-          also be supplied. <c><anno>Hosts</anno></c> identifies hosts which this
-          node can contact in order to load modules. One Erlang
-          runtime system with a <c>erl_boot_server</c> process must be
-          started on each of hosts given in <c><anno>Hosts</anno></c> in order to
-          answer the requests. See <seealso
-          marker="kernel:erl_boot_server">erl_boot_server(3)</seealso>.</p>
-        <p>If <c>-loader</c> is something else, the given port program
-          is started. The port program is supposed to follow
-          the protocol specified below.</p>
-      </desc>
-    </func>
-    <func>
       <name name="get_file" arity="1"/>
       <fsummary>Get a file</fsummary>
       <desc>
@@ -96,8 +60,6 @@
           <c><anno>Filename</anno></c> is either an absolute file name or just the name
           of the file, for example <c>"lists.beam"</c>. If an internal
           path is set to the loader, this path is used to find the file.
-          If a user supplied loader is used, the path can be stripped
-          off if it is obsolete, and the loader does not use a path.
           <c><anno>FullName</anno></c> is the complete name of the fetched file.
           <c><anno>Bin</anno></c> is the contents of the file as a binary.</p>
 
@@ -175,22 +137,6 @@
   </funcs>
 
   <section>
-    <title>Protocol</title>
-    <p>The following protocol must be followed if a user provided
-      loader port program is used. The <c>Loader</c> port program is
-      started with the command
-      <c>open_port({spawn,Loader},[binary])</c>. The protocol is as
-      follows:</p>
-    <pre>
-Function          Send               Receive
--------------------------------------------------------------
-get_file          [102 | FileName]   [121 | BinaryFile] (on success)
-                                     [122]              (failure)
-
-stop              eof                terminate</pre>
-  </section>
-
-  <section>
     <title>Command Line Flags</title>
     <p>The <c>erl_prim_loader</c> module interprets the following
       command line flags:</p>
@@ -199,10 +145,8 @@ stop              eof                terminate</pre>
       <item>
         <p>Specifies the name of the loader used by
           <c>erl_prim_loader</c>. <c>Loader</c> can be <c>efile</c>
-          (use the local file system), or <c>inet</c> (load using
-          the <c>boot_server</c> on another Erlang node). If
-          <c>Loader</c> is user defined, the defined <c>Loader</c> port
-          program is started.</p>
+          (use the local file system) or <c>inet</c> (load using
+          the <c>boot_server</c> on another Erlang node).</p>
         <p>If the <c>-loader</c> flag is omitted, it defaults to
           <c>efile</c>.</p>
       </item>
@@ -216,17 +160,12 @@ stop              eof                terminate</pre>
         <p>Specifies which other Erlang nodes the <c>inet</c> loader
           can use. This flag is mandatory if the <c>-loader inet</c>
           flag is present. On each host, there must be on Erlang node
-          with the <c>erl_boot_server</c> which handles the load
-          requests. <c>Hosts</c> is a list of IP addresses (hostnames
+          with the <seealso
+          marker="kernel:erl_boot_server">erl_boot_server(3)</seealso>
+	  which handles the load requests.
+	  <c>Hosts</c> is a list of IP addresses (hostnames
           are not acceptable).</p>
       </item>
-      <tag><c>-id Id</c></tag>
-      <item>
-        <p>Specifies the identity of the Erlang runtime system. If
-          the system runs as a distributed node, <c>Id</c> must be
-          identical to the name supplied with the <c>-sname</c> or
-          <c>-name</c> distribution flags.</p>
-      </item>
       <tag><c>-setcookie Cookie</c></tag>
       <item>
         <p>Specifies the cookie of the Erlang runtime system. This flag
diff --git a/erts/doc/src/erlang.xml b/erts/doc/src/erlang.xml
index 37f0aa289e..c3e2ed61e9 100644
--- a/erts/doc/src/erlang.xml
+++ b/erts/doc/src/erlang.xml
@@ -4,7 +4,7 @@
 <erlref>
   <header>
     <copyright>
-      <year>1996</year><year>2013</year>
+      <year>1996</year><year>2015</year>
       <holder>Ericsson AB. All Rights Reserved.</holder>
     </copyright>
     <legalnotice>
@@ -30,42 +30,49 @@
     <file>erlang.xml</file>
   </header>
   <module>erlang</module>
-  <modulesummary>The Erlang BIFs</modulesummary>
+  <modulesummary>The Erlang BIFs.</modulesummary>
   <description>
-    <p>By convention, most built-in functions (BIFs) are seen as being
-      in the module <c>erlang</c>. A number of the BIFs are viewed more
+    <p>By convention, most Built-In Functions (BIFs) are seen as being
+      in this module. Some of the BIFs are viewed more
       or less as part of the Erlang programming language and are
-      <em>auto-imported</em>. Thus, it is not necessary to specify
-      the module name and both the calls <c>atom_to_list(Erlang)</c> and
-      <c>erlang:atom_to_list(Erlang)</c> are identical.</p>
-    <p>In the text, auto-imported BIFs are listed without module prefix.
+      <em>auto-imported</em>. Thus, it is not necessary to specify the
+      module name. For example, the calls <c>atom_to_list(Erlang)</c>
+      and <c>erlang:atom_to_list(Erlang)</c> are identical.</p>
+    <p>Auto-imported BIFs are listed without module prefix.
       BIFs listed with module prefix are not auto-imported.</p>
-    <p>BIFs may fail for a variety of reasons. All BIFs fail with
+    <p>BIFs can fail for various reasons. All BIFs fail with
       reason <c>badarg</c> if they are called with arguments of an
-      incorrect type. The other reasons that may make BIFs fail are
-      described in connection with the description of each individual
-      BIF.</p>
-    <p>Some BIFs may be used in guard tests, these are marked with
+      incorrect type. The other reasons are described in the
+      description of each individual BIF.</p>
+    <p>Some BIFs can be used in guard tests and are marked with
       "Allowed in guard tests".</p>
   </description>
 
   <datatypes>
     <datatype>
-      <name><marker id="type-ext_binary">ext_binary()</marker></name>
+      <name>ext_binary()</name>
       <desc>
+      <marker id="type-ext_binary"></marker>
         <p>A binary data object, structured according to
            the Erlang external term format.</p>
       </desc>
     </datatype>
+
+    <datatype>
+      <name name="message_queue_data"></name>
+      <desc><p>See <seealso marker="#process_flag_message_queue_data"><c>erlang:process_flag(message_queue_data, MQD)</c></seealso>.</p>
+      </desc>
+    </datatype>
+
     <datatype>
       <name name="timestamp"></name>
       <desc><p>See <seealso marker="#timestamp/0">erlang:timestamp/0</seealso>.</p>
       </desc>
     </datatype>
-    <marker id="type_time_unit"/>
     <datatype>
       <name name="time_unit"></name>
-      <desc><p>Currently supported time unit representations:</p>
+      <desc><p><marker id="type_time_unit"/>
+        Supported time unit representations:</p>
         <taglist>
 	  <tag><c>PartsPerSecond :: integer() >= 1</c></tag>
 	  <item><p>Time unit expressed in parts per second. That is,
@@ -92,11 +99,11 @@
 	  used by the Erlang runtime system.</p>
 
 	  <p>The <c>native</c> time unit is determined at
-	  runtime system start, and will remain the same until
+	  runtime system start, and remains the same until
 	  the runtime system terminates. If a runtime system
 	  is stopped and then started again (even on the same
 	  machine), the <c>native</c> time unit of the new
-	  runtime system instance may differ from the
+	  runtime system instance can differ from the
 	  <c>native</c> time unit of the old runtime system
 	  instance.</p>
 
@@ -105,8 +112,7 @@
 	  seconds, native)</c>. The result equals the number
 	  of whole <c>native</c> time units per second. In case
 	  the number of <c>native</c> time units per second does
-	  not add up to a whole number, the result will be
-	  rounded downwards.</p>
+	  not add up to a whole number, the result is rounded downwards.</p>
 
 	  <note>
 	    <p>The value of the <c>native</c> time unit gives
@@ -120,7 +126,7 @@
 	    but it gives absolutely no information at all about the
 	    <seealso marker="time_correction#Time_Accuracy">accuracy</seealso>
 	    of time values. The resolution of the <c>native</c> time
-	    unit and the resolution of time values may differ
+	    unit and the resolution of time values can differ
 	    significantly.</p>
 	  </note>
 	  </item>
@@ -139,12 +145,15 @@
     <func>
       <name name="abs" arity="1" clause_i="1"/>
       <name name="abs" arity="1" clause_i="2"/>
-      <type variable="Float" name_i="1"/>
-      <type variable="Int" name_i="2"/>
-      <fsummary>Arithmetical absolute value</fsummary>
-      <desc>
-        <p>Returns an integer or float which is the arithmetical
-          absolute value of <c><anno>Float</anno></c> or <c><anno>Int</anno></c>.</p>
+      <fsummary>Arithmetical absolute value.</fsummary>
+      <type>
+        <v>Float = float()</v>
+        <v>Int = integer()</v>
+      </type>
+      <desc>
+        <p>Returns an integer or float that is the arithmetical
+          absolute value of <c><anno>Float</anno></c> or
+          <c><anno>Int</anno></c>, for example:</p>
         <pre>
 > <input>abs(-3.33).</input>
 3.33
@@ -153,206 +162,214 @@
         <p>Allowed in guard tests.</p>
       </desc>
     </func>
+
     <func>
       <name name="adler32" arity="1"/>
-      <fsummary>Compute adler32 checksum</fsummary>
+      <fsummary>Computes adler32 checksum.</fsummary>
       <desc>
-        <p>Computes and returns the adler32 checksum for <c><anno>Data</anno></c>.</p>
+        <p>Computes and returns the adler32 checksum for
+        <c><anno>Data</anno></c>.</p>
       </desc>
     </func>
+
     <func>
       <name name="adler32" arity="2"/>
-      <fsummary>Compute adler32 checksum</fsummary>
+      <fsummary>Computes adler32 checksum.</fsummary>
       <desc>
-        <p>Continue computing the adler32 checksum by combining 
-	the previous checksum, <c><anno>OldAdler</anno></c>, with the checksum of
-	<c><anno>Data</anno></c>.</p>
-	<p>The following code:</p>
-	<code>
-	X = erlang:adler32(Data1),
-	Y = erlang:adler32(X,Data2).
-	</code>
-	<p>- would assign the same value to <c>Y</c> as this would:</p>
-	<code>
-	Y = erlang:adler32([Data1,Data2]).
-	</code>
+        <p>Continues computing the adler32 checksum by combining 
+        the previous checksum, <c><anno>OldAdler</anno></c>, with
+	the checksum of <c><anno>Data</anno></c>.</p>
+        <p>The following code:</p>
+        <code>
+        X = erlang:adler32(Data1),
+        Y = erlang:adler32(X,Data2).</code>
+	<p>assigns the same value to <c>Y</c> as this:</p>
+        <code>
+        Y = erlang:adler32([Data1,Data2]).</code>
       </desc>
     </func>
+
     <func>
       <name name="adler32_combine" arity="3"/>
-      <fsummary>Combine two adler32 checksums</fsummary>
-      <desc>
-        <p>Combines two previously computed adler32 checksums. 
-	This computation requires the size of the data object for 
-	the second checksum to be known.</p>
-	<p>The following code:</p>
-	<code>
-	Y = erlang:adler32(Data1),
-	Z = erlang:adler32(Y,Data2).
-	</code>
-	<p>- would assign the same value to <c>Z</c> as this would:</p>
-	<code>
-	X = erlang:adler32(Data1),
-	Y = erlang:adler32(Data2),
-	Z = erlang:adler32_combine(X,Y,iolist_size(Data2)).
-	</code>
+      <fsummary>Combines two adler32 checksums.</fsummary>
+      <desc>
+        <p>Combines two previously computed adler32 checksums.
+        This computation requires the size of the data object for
+        the second checksum to be known.</p>
+        <p>The following code:</p>
+        <code>
+        Y = erlang:adler32(Data1),
+        Z = erlang:adler32(Y,Data2).</code>
+        <p>assigns the same value to <c>Z</c> as this:</p>
+        <code>
+        X = erlang:adler32(Data1),
+        Y = erlang:adler32(Data2),
+        Z = erlang:adler32_combine(X,Y,iolist_size(Data2)).</code>
       </desc>
     </func>
+
     <func>
       <name name="append_element" arity="2"/>
-      <fsummary>Append an extra element to a tuple</fsummary>
-      <desc>
-        <p>Returns a new tuple which has one element more than
-          <c><anno>Tuple1</anno></c>, and contains the elements in <c><anno>Tuple1</anno></c>
-          followed by <c><anno>Term</anno></c> as the last element. Semantically
-          equivalent to
-          <c>list_to_tuple(tuple_to_list(<anno>Tuple1</anno>) ++ [<anno>Term</anno>])</c>, but much
-          faster.</p>
+      <fsummary>Appends an extra element to a tuple.</fsummary>
+      <desc>
+        <p>Returns a new tuple that has one element more than
+          <c><anno>Tuple1</anno></c>, and contains the elements in
+	  <c><anno>Tuple1</anno></c>
+          followed by <c><anno>Term</anno></c> as the last element.
+	  Semantically equivalent to
+          <c>list_to_tuple(tuple_to_list(<anno>Tuple1</anno>) ++
+	  [<anno>Term</anno>])</c>, but much faster.</p>
+        <p>Example:</p>
         <pre>
 > <input>erlang:append_element({one, two}, three).</input>
 {one,two,three}</pre>
       </desc>
     </func>
+
     <func>
       <name name="apply" arity="2"/>
-      <fsummary>Apply a function to an argument list</fsummary>
+      <fsummary>Applies a function to an argument list.</fsummary>
       <desc>
-        <p>Call a fun, passing the elements in <c><anno>Args</anno></c> as
-          arguments.</p>
-        <p>Note: If the number of elements in the arguments are known at
-          compile-time, the call is better written as
+        <p>Calls a fun, passing the elements in <c><anno>Args</anno></c>
+	  as arguments.</p>
+        <p>If the number of elements in the arguments are known at
+          compile time, the call is better written as
           <c><anno>Fun</anno>(Arg1, Arg2, ... ArgN)</c>.</p>
         <warning>
           <p>Earlier, <c><anno>Fun</anno></c> could also be given as
             <c>{Module, Function}</c>, equivalent to
-            <c>apply(Module, Function, Args)</c>. This usage is
-            deprecated and will stop working in a future release of
-            Erlang/OTP.</p>
+            <c>apply(Module, Function, Args)</c>. This use is
+            deprecated and will stop working in a future release.</p>
         </warning>
       </desc>
     </func>
+
     <func>
       <name name="apply" arity="3"/>
-      <fsummary>Apply a function to an argument list</fsummary>
+      <fsummary>Applies a function to an argument list.</fsummary>
       <desc>
         <p>Returns the result of applying <c>Function</c> in
-          <c><anno>Module</anno></c> to <c><anno>Args</anno></c>. The applied function must
+          <c><anno>Module</anno></c> to <c><anno>Args</anno></c>.
+	  The applied function must
           be exported from <c>Module</c>. The arity of the function is
           the length of <c>Args</c>.</p>
+        <p>Example:</p>
         <pre>
 > <input>apply(lists, reverse, [[a, b, c]]).</input>
-[c,b,a]</pre>
-        <p><c>apply</c> can be used to evaluate BIFs by using
-          the module name <c>erlang</c>.</p>
-        <pre>
+[c,b,a]
 > <input>apply(erlang, atom_to_list, ['Erlang']).</input>
 "Erlang"</pre>
-        <p>Note: If the number of arguments are known at compile-time,
+        <p>If the number of arguments are known at compile time,
           the call is better written as
           <c><anno>Module</anno>:<anno>Function</anno>(Arg1, Arg2, ..., ArgN)</c>.</p>
         <p>Failure: <c>error_handler:undefined_function/3</c> is called
           if the applied function is not exported. The error handler
           can be redefined (see
           <seealso marker="#process_flag/2">process_flag/2</seealso>).
-          If the <c>error_handler</c> is undefined, or if the user has
+          If <c>error_handler</c> is undefined, or if the user has
           redefined the default <c>error_handler</c> so the replacement
           module is undefined, an error with the reason <c>undef</c>
           is generated.</p>
       </desc>
     </func>
+
     <func>
       <name name="atom_to_binary" arity="2"/>
-      <fsummary>Return the binary representation of an atom</fsummary>
-      <desc>
-        <p>Returns a binary which corresponds to the text
-          representation of <c><anno>Atom</anno></c>. If <c><anno>Encoding</anno></c>
-	  is <c>latin1</c>, there will be one byte for each character
-	  in the text representation. If <c><anno>Encoding</anno></c> is 
-	  <c>utf8</c> or
-	  <c>unicode</c>, the characters will be encoded using UTF-8
-	  (meaning that characters from 16#80 up to 0xFF will be
-	  encoded in two bytes).</p>
-
-	<note><p>Currently, <c>atom_to_binary(<anno>Atom</anno>, latin1)</c> can
-	never fail because the text representation of an atom can only contain
-	characters from 0 to 16#FF. In a future release, the text representation
-	of atoms might be allowed to contain any Unicode character
-	and <c>atom_to_binary(<anno>Atom</anno>, latin1)</c> will fail if the
-	text representation for the <c><anno>Atom</anno></c> contains a Unicode
-	character greater than 16#FF.</p></note>
-
+      <fsummary>Returns the binary representation of an atom.</fsummary>
+      <desc>
+        <p>Returns a binary corresponding to the text
+          representation of <c><anno>Atom</anno></c>.
+          If <c><anno>Encoding</anno></c>
+          is <c>latin1</c>, there is one byte for each character
+          in the text representation. If <c><anno>Encoding</anno></c> is
+          <c>utf8</c> or
+          <c>unicode</c>, the characters are encoded using UTF-8
+          (that is, characters from 128 through 255 are
+          encoded in two bytes).</p>
+	<note><p><c>atom_to_binary(<anno>Atom</anno>, latin1)</c> never
+        fails because the text representation of an atom can only
+        contain characters from 0 through 255. In a future release,
+        the text representation
+        of atoms can be allowed to contain any Unicode character and
+        <c>atom_to_binary(<anno>Atom</anno>, latin1)</c> will then fail if the
+        text representation for <c><anno>Atom</anno></c> contains a Unicode
+        character greater than 255.</p></note>
+        <p>Example:</p>
         <pre>
 > <input>atom_to_binary('Erlang', latin1).</input>
 &lt;&lt;"Erlang"&gt;&gt;</pre>
       </desc>
     </func>
+
     <func>
       <name name="atom_to_list" arity="1"/>
-      <fsummary>Text representation of an atom</fsummary>
+      <fsummary>Text representation of an atom.</fsummary>
       <desc>
-        <p>Returns a string which corresponds to the text
-          representation of <c><anno>Atom</anno></c>.</p>
+        <p>Returns a string corresponding to the text
+          representation of <c><anno>Atom</anno></c>, for example:</p>
         <pre>
 > <input>atom_to_list('Erlang').</input>
 "Erlang"</pre>
       </desc>
     </func>
+
     <func>
       <name name="binary_part" arity="2"/>
-      <fsummary>Extracts a part of a binary</fsummary>
+      <fsummary>Extracts a part of a binary.</fsummary>
       <desc>
-      <p>Extracts the part of the binary described by <c><anno>PosLen</anno></c>.</p>
-
-      <p>Negative length can be used to extract bytes at the end of a binary:</p>
-
+      <p>Extracts the part of the binary described by
+        <c><anno>PosLen</anno></c>.</p>
+      <p>Negative length can be used to extract bytes at the end
+      of a binary, for example:</p>
 <code>
 1> Bin = &lt;&lt;1,2,3,4,5,6,7,8,9,10&gt;&gt;.
 2> binary_part(Bin,{byte_size(Bin), -5}).
-&lt;&lt;6,7,8,9,10&gt;&gt;
-</code>
-
-      <p>If <c><anno>PosLen</anno></c> in any way references outside the binary, a <c>badarg</c> exception is raised.</p>
-
-      <p><c><anno>Start</anno></c> is zero-based, i.e.:</p>
+&lt;&lt;6,7,8,9,10&gt;&gt;</code>
+      <p>Failure: <c>badarg</c> if <c><anno>PosLen</anno></c> in any way
+        references outside the binary.</p>
+      <p><c><anno>Start</anno></c> is zero-based, that is:</p>
 <code>
 1> Bin = &lt;&lt;1,2,3&gt;&gt;
 2> binary_part(Bin,{0,2}).
-&lt;&lt;1,2&gt;&gt;
-</code>
-
-      <p>See the STDLIB module <c>binary</c> for details about the <c><anno>PosLen</anno></c> semantics.</p>
-
+&lt;&lt;1,2&gt;&gt;</code>
+      <p>For details about the <c><anno>PosLen</anno></c> semantics, see the
+        <seealso marker="stdlib:binary">binary</seealso>
+        manual page in <c>STDLIB</c>.</p>
       <p>Allowed in guard tests.</p>
       </desc>
     </func>
+
     <func>
       <name name="binary_part" arity="3"/>
-      <fsummary>Extracts a part of a binary</fsummary>
+      <fsummary>Extracts a part of a binary.</fsummary>
       <desc>
-      <p>The same as <c>binary_part(<anno>Subject</anno>, {<anno>Start</anno>, <anno>Length</anno>})</c>.</p>
-
+      <p>The same as <c>binary_part(<anno>Subject</anno>,
+      {<anno>Start</anno>, <anno>Length</anno>})</c>.</p>
       <p>Allowed in guard tests.</p>
       </desc>
     </func>
+
     <func>
       <name name="binary_to_atom" arity="2"/>
-      <fsummary>Convert from text representation to an atom</fsummary>
+      <fsummary>Converts from text representation to an atom.</fsummary>
       <desc>
         <p>Returns the atom whose text representation is
-	<c><anno>Binary</anno></c>.  If <c><anno>Encoding</anno></c> is <c>latin1</c>, no
-	translation of bytes in the binary is done. If <c><anno>Encoding</anno></c>
-	is <c>utf8</c> or <c>unicode</c>, the binary must contain
-	valid UTF-8 sequences; furthermore, only Unicode characters up
-	to 0xFF are allowed.</p>
-
-	<note><p><c>binary_to_atom(<anno>Binary</anno>, utf8)</c> will fail if
-	the binary contains Unicode characters greater than 16#FF.
-	In a future release, such Unicode characters might be allowed
-	and <c>binary_to_atom(<anno>Binary</anno>, utf8)</c>
-	will not fail in that case. For more information on Unicode support in atoms
-	see <seealso marker="erl_ext_dist#utf8_atoms">note on UTF-8 encoded atoms</seealso>
-	in the chapter about the external term format in the ERTS User's Guide.</p></note>
-
+        <c><anno>Binary</anno></c>.
+        If <c><anno>Encoding</anno></c> is <c>latin1</c>, no
+        translation of bytes in the binary is done.
+        If <c><anno>Encoding</anno></c>
+        is <c>utf8</c> or <c>unicode</c>, the binary must contain
+        valid UTF-8 sequences. Only Unicode characters up
+        to 255 are allowed.</p>
+        <note><p><c>binary_to_atom(<anno>Binary</anno>, utf8)</c> fails if
+        the binary contains Unicode characters greater than 255.
+        In a future release, such Unicode characters can be allowed
+        and <c>binary_to_atom(<anno>Binary</anno>, utf8)</c> does then not fail.
+        For more information on Unicode support in atoms, see the
+        <seealso marker="erl_ext_dist#utf8_atoms">note on UTF-8
+        encoded atoms</seealso>
+        in Section "External Term Format" in the User's Guide.</p></note>
+        <p>Examples:</p>
         <pre>
 > <input>binary_to_atom(&lt;&lt;"Erlang"&gt;&gt;, latin1).</input>
 'Erlang'
@@ -362,20 +379,24 @@
         called as binary_to_atom(&lt;&lt;208,128&gt;&gt;,utf8)</pre>
       </desc>
     </func>
+
     <func>
       <name name="binary_to_existing_atom" arity="2"/>
-      <fsummary>Convert from text representation to an atom</fsummary>
+      <fsummary>Converts from text representation to an atom.</fsummary>
       <desc>
-        <p>Works like <seealso marker="#binary_to_atom/2">binary_to_atom/2</seealso>,
-          but the atom must already exist.</p>
-        <p>Failure: <c>badarg</c> if the atom does not already exist.</p>
+        <p>As
+          <seealso marker="#binary_to_atom/2">binary_to_atom/2</seealso>,
+          but the atom must exist.</p>
+        <p>Failure: <c>badarg</c> if the atom does not exist.</p>
       </desc>
     </func>
+
     <func>
       <name name="binary_to_float" arity="1"/>
-      <fsummary>Convert from text representation to a float</fsummary>
+      <fsummary>Converts from text representation to a float.</fsummary>
       <desc>
-        <p>Returns the float whose text representation is <c><anno>Binary</anno></c>.</p>
+        <p>Returns the float whose text representation is
+          <c><anno>Binary</anno></c>, for example:</p>
         <pre>
 > <input>binary_to_float(&lt;&lt;"2.2017764e+0">>).</input>
 2.2017764</pre>
@@ -383,12 +404,13 @@
           representation of a float.</p>
       </desc>
     </func>
+
     <func>
       <name name="binary_to_integer" arity="1"/>
-      <fsummary>Convert from text representation to an integer</fsummary>
+      <fsummary>Converts from text representation to an integer.</fsummary>
       <desc>
         <p>Returns an integer whose text representation is
-          <c><anno>Binary</anno></c>.</p>
+          <c><anno>Binary</anno></c>, for example:</p>
         <pre>
 > <input>binary_to_integer(&lt;&lt;"123"&gt;&gt;).</input>
 123</pre>
@@ -396,12 +418,13 @@
           representation of an integer.</p>
       </desc>
     </func>
+
     <func>
       <name name="binary_to_integer" arity="2"/>
-      <fsummary>Convert from text representation to an integer</fsummary>
+      <fsummary>Converts from text representation to an integer.</fsummary>
       <desc>
         <p>Returns an integer whose text representation in base
-          <c><anno>Base</anno></c> is <c><anno>Binary</anno></c>.</p>
+          <c><anno>Base</anno></c> is <c><anno>Binary</anno></c>, for example:</p>
         <pre>
 > <input>binary_to_integer(&lt;&lt;"3FF"&gt;&gt;, 16).</input>
 1023</pre>
@@ -409,93 +432,101 @@
           representation of an integer.</p>
       </desc>
     </func>
+
     <func>
       <name name="binary_to_list" arity="1"/>
-      <fsummary>Convert a binary to a list</fsummary>
+      <fsummary>Converts a binary to a list.</fsummary>
       <desc>
-        <p>Returns a list of integers which correspond to the bytes of
+        <p>Returns a list of integers corresponding to the bytes of
           <c><anno>Binary</anno></c>.</p>
       </desc>
     </func>
+
     <func>
       <name name="binary_to_list" arity="3"/>
-      <fsummary>Convert part of a binary to a list</fsummary>
-      <type_desc variable="Start">1..byte_size(<anno>Binary</anno>)</type_desc>
+      <fsummary>Converts part of a binary to a list.</fsummary>
+      <type_desc variable="Start">1..byte_size(<c><anno>Binary</anno></c>)</type_desc>
       <desc>
         <p>As <c>binary_to_list/1</c>, but returns a list of integers
           corresponding to the bytes from position <c><anno>Start</anno></c> to
-          position <c><anno>Stop</anno></c> in <c><anno>Binary</anno></c>. Positions in the
+          position <c><anno>Stop</anno></c> in <c><anno>Binary</anno></c>.
+	  The positions in the
           binary are numbered starting from 1.</p>
-
-	  <note><p>This function's indexing style of using one-based indices for
-	  binaries is deprecated. New code should use the functions in
-	  the STDLIB module <c>binary</c> instead. They consequently
-	  use the same (zero-based) style of indexing.</p></note>
+          <note><p>The one-based indexing for binaries used by
+	  this function is deprecated. New code is to use
+          <seealso marker="stdlib:binary#bin_to_list/3">binary:bin_to_list/3</seealso>
+	  in <c>STDLIB</c> instead. All functions in module
+	  <c>binary</c> consistently use zero-based indexing.</p></note>
       </desc>
     </func>
+
     <func>
       <name name="bitstring_to_list" arity="1"/>
-      <fsummary>Convert a bitstring to a list</fsummary>
+      <fsummary>Converts a bitstring to a list.</fsummary>
       <desc>
-        <p>Returns a list of integers which correspond to the bytes of
-          <c><anno>Bitstring</anno></c>. If the number of bits in the binary is not
-	  divisible by 8, the last element of the list will be a bitstring
-	  containing the remaining bits (1 up to 7 bits).</p>
+        <p>Returns a list of integers corresponding to the bytes of
+          <c><anno>Bitstring</anno></c>. If the number of bits in the binary
+          is not divisible by 8, the last element of the list is a bitstring
+          containing the remaining 1-7 bits.</p>
       </desc>
     </func>
+
     <func>
       <name name="binary_to_term" arity="1"/>
-      <fsummary>Decode an Erlang external term format binary</fsummary>
+      <fsummary>Decodes an Erlang external term format binary.</fsummary>
       <desc>
-        <p>Returns an Erlang term which is the result of decoding
-          the binary object <c><anno>Binary</anno></c>, which must be encoded
+        <p>Returns an Erlang term that is the result of decoding
+          binary object <c><anno>Binary</anno></c>, which must be encoded
           according to the Erlang external term format.</p>
-        <warning>
-	  <p>When decoding binaries from untrusted sources, consider using
-	     <c>binary_to_term/2</c> to prevent denial of service attacks.</p>
-        </warning>
-	<p>See also
-	   <seealso marker="#term_to_binary/1">term_to_binary/1</seealso>
-           and
-	   <seealso marker="#binary_to_term/2">binary_to_term/2</seealso>.</p>
+        <warning><p>When decoding binaries from untrusted sources,
+        consider using <c>binary_to_term/2</c> to prevent Denial
+        of Service attacks.</p></warning>
+        <p>See also
+          <seealso marker="#term_to_binary/1">term_to_binary/1</seealso>
+          and
+          <seealso marker="#binary_to_term/2">binary_to_term/2</seealso>.</p>
       </desc>
     </func>
+
     <func>
       <name name="binary_to_term" arity="2"/>
-      <fsummary>Decode an Erlang external term format binary</fsummary>
+      <fsummary>Decodes an Erlang external term format binary.</fsummary>
       <desc>
         <p>As <c>binary_to_term/1</c>, but takes options that affect decoding
            of the binary.</p>
         <taglist>
           <tag><c>safe</c></tag>
           <item>
-	    <p>Use this option when receiving binaries from an untrusted
+            <p>Use this option when receiving binaries from an untrusted
                source.</p>
-	    <p>When enabled, it prevents decoding data that may be used to
-	       attack the Erlang system.  In the event of receiving unsafe
-               data, decoding fails with a badarg error.</p>
-	    <p>Currently, this prevents creation of new atoms directly,
-	       creation of new atoms indirectly (as they are embedded in
-	       certain structures like pids, refs, funs, etc.), and creation of
-	       new external function references.  None of those resources are
-	       currently garbage collected, so unchecked creation of them can
-               exhaust available memory.</p>
+            <p>When enabled, it prevents decoding data that can be used to
+               attack the Erlang system. In the event of receiving unsafe
+               data, decoding fails with a <c>badarg</c> error.</p>
+            <p>This prevents creation of new atoms directly,
+               creation of new atoms indirectly (as they are embedded in
+               certain structures, such as process identifiers,
+               refs, and funs), and
+               creation of new external function references.
+               None of those resources are garbage collected, so unchecked
+               creation of them can exhaust available memory.</p>
           </item>
         </taglist>
-	<p>Failure: <c>badarg</c> if <c>safe</c> is specified and unsafe data
-           is decoded.</p>
+        <p>Failure: <c>badarg</c> if <c>safe</c> is specified and unsafe
+           data is decoded.</p>
         <p>See also
            <seealso marker="#term_to_binary/1">term_to_binary/1</seealso>,
            <seealso marker="#binary_to_term/1">binary_to_term/1</seealso>,
-           and <seealso marker="#list_to_existing_atom/1">
-             list_to_existing_atom/1</seealso>.</p>
+           and
+           <seealso marker="#list_to_existing_atom/1">list_to_existing_atom/1</seealso>.</p>
       </desc>
     </func>
+
     <func>
       <name name="bit_size" arity="1"/>
-      <fsummary>Return the size of a bitstring</fsummary>
+      <fsummary>Returns the size of a bitstring.</fsummary>
       <desc>
-        <p>Returns an integer which is the size in bits of <c><anno>Bitstring</anno></c>.</p>
+        <p>Returns an integer that is the size in bits of
+          <c><anno>Bitstring</anno></c>, for example:</p>
         <pre>
 > <input>bit_size(&lt;&lt;433:16,3:3&gt;&gt;).</input>
 19
@@ -504,30 +535,34 @@
         <p>Allowed in guard tests.</p>
       </desc>
     </func>
+
     <func>
       <name name="bump_reductions" arity="1"/>
-      <fsummary>Increment the reduction counter</fsummary>
+      <fsummary>Increments the reduction counter.</fsummary>
       <desc>
         <p>This implementation-dependent function increments
           the reduction counter for the calling process. In the Beam
           emulator, the reduction counter is normally incremented by
-          one for each function and BIF call, and a context switch is
-          forced when the counter reaches the maximum number of reductions
-	  for a process (2000 reductions in R12B).</p>
+          one for each function and BIF call. A context switch is
+          forced when the counter reaches the maximum number of
+          reductions for a process (2000 reductions in OTP R12B).</p>
         <warning>
-          <p>This BIF might be removed in a future version of the Beam
+          <p>This BIF can be removed in a future version of the Beam
             machine without prior warning. It is unlikely to be
             implemented in other Erlang implementations.</p>
         </warning>
       </desc>
     </func>
+
     <func>
       <name name="byte_size" arity="1"/>
-      <fsummary>Return the size of a bitstring (or binary)</fsummary>
+      <fsummary>Returns the size of a bitstring (or binary).</fsummary>
       <desc>
-        <p>Returns an integer which is the number of bytes needed to contain
-	<c><anno>Bitstring</anno></c>. (That is, if the number of bits in <c><anno>Bitstring</anno></c> is not
-	divisible by 8, the resulting number of bytes will be rounded <em>up</em>.)</p>
+        <p>Returns an integer that is the number of bytes needed to
+          contain <c><anno>Bitstring</anno></c>. That is, if the number of bits
+          in <c><anno>Bitstring</anno></c> is not divisible by 8, the resulting
+          number of bytes is rounded <em>up</em>.</p>
+        <p>Examples:</p>
         <pre>
 > <input>byte_size(&lt;&lt;433:16,3:3&gt;&gt;).</input>
 3
@@ -536,18 +571,19 @@
         <p>Allowed in guard tests.</p>
       </desc>
     </func>
+
     <func>
       <name name="cancel_timer" arity="2"/>
-      <fsummary>Cancel a timer</fsummary>
+      <fsummary>Cancels a timer.</fsummary>
       <desc>
         <p>
-	  Cancels a timer that has been created by either
+	  Cancels a timer that has been created by
 	  <seealso marker="#start_timer/4"><c>erlang:start_timer()</c></seealso>,
 	  or <seealso marker="#send_after/4"><c>erlang:send_after()</c></seealso>.
 	  <c><anno>TimerRef</anno></c> identifies the timer, and
 	  was returned by the BIF that created the timer.
 	</p>
-	<p>Currently available <c><anno>Option</anno></c>s:</p>
+	<p>Available <c><anno>Option</anno></c>s:</p>
         <taglist>
           <tag><c>{async, Async}</c></tag>
           <item>
@@ -556,7 +592,7 @@
 	      defaults to <c>false</c> which will cause the
 	      cancellation to be performed synchronously. When
 	      <c>Async</c> is set to <c>true</c>, the cancel
-	      operation will be performed asynchronously. That is,
+	      operation is performed asynchronously. That is,
 	      <c>erlang:cancel_timer()</c> will send an asynchronous
 	      request for cancellation to the timer service that
 	      manages the timer, and then return <c>ok</c>.
@@ -567,17 +603,17 @@
 	    <p>
 	      Request information about the <c><anno>Result</anno></c>
 	      of the cancellation. <c>Info</c> defaults to <c>true</c>
-	      which means that the <c><anno>Result</anno></c> will
-	      be given. When <c>Info</c> is set to <c>false</c>, no
+	      which means the <c><anno>Result</anno></c> is
+	      given. When <c>Info</c> is set to <c>false</c>, no
 	      information about the result of the cancellation
-	      will be given. When the operation is performed</p>
+	      is given. When the operation is performed</p>
               <taglist>
 		<tag>synchronously</tag>
 		<item>
 		  <p>
-		    If <c>Info</c> is <c>true</c>, the <c>Result</c> will
+		    If <c>Info</c> is <c>true</c>, the <c>Result</c> is
 		    returned by <c>erlang:cancel_timer()</c>; otherwise,
-		    <c>ok</c> will be returned.
+		    <c>ok</c> is returned.
 		  </p>
 		</item>
 		<tag>asynchronously</tag>
@@ -585,10 +621,10 @@
 		  <p>
 		    If <c>Info</c> is <c>true</c>, a message on the form
 		    <c>{cancel_timer, <anno>TimerRef</anno>,
-		    <anno>Result</anno>}</c> will be sent to the
+		    <anno>Result</anno>}</c> is sent to the
 		    caller of <c>erlang:cancel_timer()</c> when the
 		    cancellation operation has been performed; otherwise,
-		    no message will be sent.
+		    no message is sent.
 		  </p>
 		</item>
             </taglist>
@@ -597,30 +633,30 @@
 	<p>
 	  More <c><anno>Option</anno></c>s may be added in the future.
 	</p>
+	<p>If <c><anno>Result</anno></c> is an integer, it represents
+	  the time in milli-seconds left until the canceled timer would
+	  have expired.</p>
 	<p>
-	  When the <c><anno>Result</anno></c> equals <c>false</c>, a
+	  If <c><anno>Result</anno></c> is <c>false</c>, a
 	  timer corresponding to <c><anno>TimerRef</anno></c> could not
 	  be found. This can be either because the timer had expired,
 	  already had been canceled, or because <c><anno>TimerRef</anno></c>
-	  never has corresponded to a timer. If the timer has expired,
-	  the timeout message has been sent, but it does not tell you
-	  whether or not it has arrived at its destination yet. When the
-	  <c><anno>Result</anno></c> is an integer, it represents the
-	  time in milli-seconds left until the timer will expire.
+	  never corresponded to a timer. Even if the timer had expired,
+	  it does not tell you whether or not the timeout message has
+	  arrived at its destination yet.
 	</p>
 	<note>
 	  <p>
 	    The timer service that manages the timer may be co-located
 	    with another scheduler than the scheduler that the calling
 	    process is executing on. If this is the case, communication
-	    with the timer service will take much longer time than if it
+	    with the timer service takes much longer time than if it
 	    is located locally. If the calling process is in critical
 	    path, and can do other things while waiting for the result
 	    of this operation, or is not interested in the result of
-	    the operation, you want to use the <c>{async, true}</c>
-	    option. If using the <c>{async, false}</c> option, the calling
-	    process will be blocked until the operation has been
-	    performed.
+	    the operation, you want to use option <c>{async, true}</c>.
+	    If using option <c>{async, false}</c>, the calling
+	    process blocks until the operation has been performed.
 	  </p>
 	</note>
         <p>See also 
@@ -632,7 +668,7 @@
     </func>
     <func>
       <name name="cancel_timer" arity="1"/>
-      <fsummary>Cancel a timer</fsummary>
+      <fsummary>Cancels a timer.</fsummary>
       <desc>
         <p>Cancels a timer. The same as calling
 	<seealso marker="#cancel_timer/2"><c>erlang:cancel_timer(TimerRef,
@@ -641,103 +677,102 @@
     </func>
     <func>
       <name name="check_old_code" arity="1"/>
-      <fsummary>Check if a module has old code</fsummary>
+      <fsummary>Checks if a module has old code.</fsummary>
       <desc>
-        <p>Returns <c>true</c> if the <c><anno>Module</anno></c> has old code,
-	and <c>false</c> otherwise.</p>
+        <p>Returns <c>true</c> if <c><anno>Module</anno></c> has old code,
+	  otherwise <c>false</c>.</p>
         <p>See also <seealso marker="kernel:code">code(3)</seealso>.</p>
       </desc>
     </func>
+
     <func>
       <name name="check_process_code" arity="2"/>
-      <fsummary>Check if a process is executing old code for a module</fsummary>
+      <fsummary>Checks if a process executes old code for a module.</fsummary>
       <desc>
         <p>The same as
-	<seealso marker="#check_process_code/3"><c>erlang:check_process_code(<anno>Pid</anno>,
-	<anno>Module</anno>, [])</c></seealso>.</p>
+        <seealso marker="#check_process_code/3"><c>erlang:check_process_code(<anno>Pid</anno>, <anno>Module</anno>, [])</c></seealso>.</p>
       </desc>
     </func>
+
     <func>
       <name name="check_process_code" arity="3"/>
-      <fsummary>Check if a process is executing old code for a module</fsummary>
+      <fsummary>Checks if a process executes old code for a module.</fsummary>
       <desc>
-	<p>Check if the node local process identified by <c><anno>Pid</anno></c>
-	is executing old code for <c><anno>Module</anno></c>.</p>
-	<p>Currently available <c><anno>Option</anno>s</c>:</p>
+        <p>Checks if the node local process identified by <c><anno>Pid</anno></c>
+        executes old code for <c><anno>Module</anno></c>.</p>
+        <p>The available <c><anno>Option</anno></c>s are as follows:</p>
         <taglist>
           <tag><c>{allow_gc, boolean()}</c></tag>
           <item>
-	    Determines if garbage collection is allowed when performing
-	    the operation. If <c>{allow_gc, false}</c> is passed, and
-	    a garbage collection is needed in order to determine the
-	    result of the operation, the operation will be aborted
-	    (see information on <c><anno>CheckResult</anno></c> below).
-	    The default is to allow garbage collection, i.e.,
-	    <c>{allow_gc, true}</c>.
+            <p>Determines if garbage collection is allowed when performing
+            the operation. If <c>{allow_gc, false}</c> is passed, and
+            a garbage collection is needed to determine the
+            result of the operation, the operation is aborted (see
+            information on <c><anno>CheckResult</anno></c> in the following).
+            The default is to allow garbage collection, that is,
+            <c>{allow_gc, true}</c>.</p>
 	  </item>
           <tag><c>{async, RequestId}</c></tag>
           <item>
-	    The <c>check_process_code/3</c> function will return
-	    the value <c>async</c> immediately after the request
-	    has been sent. When the request has been processed, the
-	    process that called this function will be passed a
-	    message on the form:<br/>
-	    <c>{check_process_code, <anno>RequestId</anno>, <anno>CheckResult</anno>}</c>.
+            <p>The function <c>check_process_code/3</c> returns
+            the value <c>async</c> immediately after the request
+            has been sent. When the request has been processed, the
+            process that called this function is passed a
+            message on the form
+            <c>{check_process_code, <anno>RequestId</anno>, <anno>CheckResult</anno>}</c>.</p>
 	  </item>
         </taglist>
-	<p>If <c><anno>Pid</anno></c> equals <c>self()</c>, and
-	no <c>async</c> option has been passed, the operation will
-	be performed at once. In all other cases a request for
-	the operation will be sent to the process identified by
-	<c><anno>Pid</anno></c>, and will be handled when
-	appropriate. If no <c>async</c> option has been passed,
-	the caller will block until <c><anno>CheckResult</anno></c>
-	is available and can be returned.</p>
-	<p><c><anno>CheckResult</anno></c> informs about the result of
-	the request:</p>
+        <p>If <c><anno>Pid</anno></c> equals <c>self()</c>, and
+        no <c>async</c> option has been passed, the operation
+        is performed at once. Otherwise a request for
+        the operation is sent to the process identified by
+        <c><anno>Pid</anno></c>, and is handled when
+        appropriate. If no <c>async</c> option has been passed,
+        the caller blocks until <c><anno>CheckResult</anno></c>
+        is available and can be returned.</p>
+        <p><c><anno>CheckResult</anno></c> informs about the result of
+        the request as follows:</p>
         <taglist>
           <tag><c>true</c></tag>
           <item>
-	    The process identified by <c><anno>Pid</anno></c> is
-	    executing old code for <c><anno>Module</anno></c>.
-	    That is, the current call of the process executes old
-	    code for this module, or the process has references
-	    to old code for this module, or the process contains
-	    funs that references old code for this module.
+            <p>The process identified by <c><anno>Pid</anno></c>
+            executes old code for <c><anno>Module</anno></c>.
+            That is, the current call of the process executes old
+            code for this module, or the process has references
+            to old code for this module, or the process contains
+            funs that references old code for this module.</p>
 	  </item>
           <tag><c>false</c></tag>
           <item>
-	    The process identified by <c><anno>Pid</anno></c> is
-	    not executing old code for <c><anno>Module</anno></c>.
+            <p>The process identified by <c><anno>Pid</anno></c> does
+            not execute old code for <c><anno>Module</anno></c>.</p>
 	  </item>
           <tag><c>aborted</c></tag>
           <item>
-	    The operation was aborted since the process needed to
-	    be garbage collected in order to determine the result
-	    of the operation, and the operation was requested
-	    by passing the <c>{allow_gc, false}</c> option.</item>
+            <p>The operation was aborted, as the process needed to
+            be garbage collected to determine the operation result,
+            and the operation was requested
+            by passing option <c>{allow_gc, false}</c>.</p></item>
         </taglist>
         <p>See also <seealso marker="kernel:code">code(3)</seealso>.</p>
         <p>Failures:</p>
         <taglist>
           <tag><c>badarg</c></tag>
-          <item>
-	    If <c><anno>Pid</anno></c> is not a node local process identifier.
+          <item>If <c><anno>Pid</anno></c> is not a node local process identifier.
 	  </item>
           <tag><c>badarg</c></tag>
-          <item>
-	    If <c><anno>Module</anno></c> is not an atom.
+          <item>If <c><anno>Module</anno></c> is not an atom.
 	  </item>
           <tag><c>badarg</c></tag>
-          <item>
-	    If <c><anno>OptionList</anno></c> is not a valid list of options.
+          <item>If <c><anno>OptionList</anno></c> is an invalid list of options.
 	  </item>
         </taglist>
       </desc>
     </func>
+
     <func>
       <name name="convert_time_unit" arity="3"/>
-      <fsummary>Convert time unit of a time value</fsummary>
+      <fsummary>Converts time unit of a time value.</fsummary>
       <desc>
 	<p>Converts the <c><anno>Time</anno></c> value of time unit
 	<c><anno>FromUnit</anno></c> to the corresponding
@@ -753,100 +788,102 @@
     </func>
     <func>
       <name name="crc32" arity="1"/>
-      <fsummary>Compute crc32 (IEEE 802.3) checksum</fsummary>
+      <fsummary>Computes crc32 (IEEE 802.3) checksum.</fsummary>
       <desc>
-        <p>Computes and returns the crc32 (IEEE 802.3 style) checksum for <c><anno>Data</anno></c>.</p>
+        <p>Computes and returns the crc32 (IEEE 802.3 style) checksum
+          for <c><anno>Data</anno></c>.</p>
       </desc>
     </func>
+
     <func>
       <name name="crc32" arity="2"/>
-      <fsummary>Compute crc32 (IEEE 802.3) checksum</fsummary>
+      <fsummary>Computes crc32 (IEEE 802.3) checksum.</fsummary>
       <desc>
-        <p>Continue computing the crc32 checksum by combining 
-	the previous checksum, <c><anno>OldCrc</anno></c>, with the checksum of
-	<c><anno>Data</anno></c>.</p>
-	<p>The following code:</p>
-	<code>
-	X = erlang:crc32(Data1),
-	Y = erlang:crc32(X,Data2).
-	</code>
-	<p>- would assign the same value to <c>Y</c> as this would:</p>
-	<code>
-	Y = erlang:crc32([Data1,Data2]).
-	</code>
+        <p>Continues computing the crc32 checksum by combining
+        the previous checksum, <c><anno>OldCrc</anno></c>, with the checksum of
+        <c><anno>Data</anno></c>.</p>
+        <p>The following code:</p>
+        <code>
+        X = erlang:crc32(Data1),
+        Y = erlang:crc32(X,Data2).</code>
+        <p>assigns the same value to <c>Y</c> as this:</p>
+        <code>
+        Y = erlang:crc32([Data1,Data2]).</code>
       </desc>
     </func>
+
     <func>
       <name name="crc32_combine" arity="3"/>
-      <fsummary>Combine two crc32 (IEEE 802.3) checksums</fsummary>
-      <desc>
-        <p>Combines two previously computed crc32 checksums. 
-	This computation requires the size of the data object for 
-	the second checksum to be known.</p>
-	<p>The following code:</p>
-	<code>
-	Y = erlang:crc32(Data1),
-	Z = erlang:crc32(Y,Data2).
-	</code>
-	<p>- would assign the same value to <c>Z</c> as this would:</p>
+      <fsummary>Combines two crc32 (IEEE 802.3) checksums.</fsummary>
+      <desc>
+        <p>Combines two previously computed crc32 checksums.
+        This computation requires the size of the data object for
+        the second checksum to be known.</p>
+        <p>The following code:</p>
+        <code>
+        Y = erlang:crc32(Data1),
+        Z = erlang:crc32(Y,Data2).</code>
+        <p>assigns the same value to <c>Z</c> as this:</p>
 	<code>
-	X = erlang:crc32(Data1),
-	Y = erlang:crc32(Data2),
-	Z = erlang:crc32_combine(X,Y,iolist_size(Data2)).
-	</code>
+        X = erlang:crc32(Data1),
+        Y = erlang:crc32(Data2),
+        Z = erlang:crc32_combine(X,Y,iolist_size(Data2)).</code>
       </desc>
     </func>
+
     <func>
       <name name="date" arity="0"/>
-      <fsummary>Current date</fsummary>
+      <fsummary>Current date.</fsummary>
       <desc>
         <p>Returns the current date as <c>{Year, Month, Day}</c>.</p>
-        <p>The time zone and daylight saving time correction depend on
+        <p>The time zone and Daylight Saving Time correction depend on
           the underlying OS.</p>
+        <p>Example:</p>
         <pre>
 > <input>date().</input>
 {1995,2,19}</pre>
       </desc>
     </func>
+
     <func>
       <name name="decode_packet" arity="3"/>
-      <fsummary>Extracts a protocol packet from a binary</fsummary>
+      <fsummary>Extracts a protocol packet from a binary.</fsummary>
       <desc>
-
         <p>Decodes the binary <c><anno>Bin</anno></c> according to the packet
-        protocol specified by <c><anno>Type</anno></c>. Very similar to the packet
-        handling done by sockets with the option {packet,<anno>Type</anno>}.</p>
-        <p>If an entire packet is contained in <c><anno>Bin</anno></c> it is
+        protocol specified by <c><anno>Type</anno></c>. Similar to the packet
+        handling done by sockets with option {packet,<anno>Type</anno>}.</p>
+        <p>If an entire packet is contained in <c><anno>Bin</anno></c>, it is
         returned together with the remainder of the binary as
         <c>{ok,<anno>Packet</anno>,<anno>Rest</anno>}</c>.</p>
         <p>If <c><anno>Bin</anno></c> does not contain the entire packet,
-        <c>{more,<anno>Length</anno>}</c> is returned. <c><anno>Length</anno></c> is either the
-        expected <em>total size</em> of the packet or <c>undefined</c>
-        if the expected packet size is not known. <c>decode_packet</c>
+        <c>{more,<anno>Length</anno>}</c> is returned.
+        <c><anno>Length</anno></c> is either the
+        expected <em>total size</em> of the packet, or <c>undefined</c>
+        if the expected packet size is unknown. <c>decode_packet</c>
         can then be called again with more data added.</p>
-        <p>If the packet does not conform to the protocol format
+        <p>If the packet does not conform to the protocol format,
         <c>{error,<anno>Reason</anno>}</c> is returned.</p>
-        <p>The following values of <c><anno>Type</anno></c> are valid:</p>
+        <p>The following <c>Type</c>s are valid:</p>
         <taglist>
           <tag><c>raw | 0</c></tag>
           <item>
-            <p>No packet handling is done. Entire binary is
+            <p>No packet handling is done. The entire binary is
             returned unless it is empty.</p>
           </item>
           <tag><c>1 | 2 | 4</c></tag>
           <item>
             <p>Packets consist of a header specifying the number of
               bytes in the packet, followed by that number of bytes.
-              The length of header can be one, two, or four bytes;
+              The length of the header can be one, two, or four bytes;
               the order of the bytes is big-endian. The header
-              will be stripped off when the packet is returned.</p>
+              is stripped off when the packet is returned.</p>
           </item>
           <tag><c>line</c></tag>
           <item>
-            <p>A packet is a line terminated with newline. The
-              newline character is included in the returned packet
-              unless the line was truncated according to the option
-              <c>line_length</c>.</p> 
+            <p>A packet is a line terminated by a delimiter byte,
+            default is the latin1 newline character. The delimiter
+            byte is included in the returned packet unless the line
+            was truncated according to option <c>line_length</c>.</p>
           </item>
           <tag><c>asn1 | cdr | sunrm | fcgi | tpkt</c></tag>
           <item>
@@ -864,41 +901,50 @@
           <item>
             <p>The Hypertext Transfer Protocol. The packets
                are returned with the format according to
-               <c><anno>HttpPacket</anno></c> described above. A packet is either a
-               request, a response, a header or an end of header
-               mark. Invalid lines are returned as <c><anno>HttpError</anno></c>.</p>
-            <p>Recognized request methods and header fields are returned as atoms.
-               Others are returned as strings. Strings of unrecognized header fields
-	       are formatted with only capital letters first and after hyphen characters
-	       (like <c>"Sec-Websocket-Key"</c>).</p>
-            <p>The protocol type <c>http</c> should only be used for
-               the first line when a <c><anno>HttpRequest</anno></c> or a
-               <c><anno>HttpResponse</anno></c> is expected. The following calls
-               should use <c>httph</c> to get <c><anno>HttpHeader</anno></c>'s until
-               <c>http_eoh</c> is returned that marks the end of the
+               <c><anno>HttpPacket</anno></c> described earlier.
+               A packet is either a
+               request, a response, a header, or an end of header
+               mark. Invalid lines are returned as
+               <c><anno>HttpError</anno></c>.</p>
+            <p>Recognized request methods and header fields are returned
+               as atoms. Others are returned as strings. Strings of
+               unrecognized header fields are formatted with only
+               capital letters first and after hyphen characters, for
+               example, <c>"Sec-Websocket-Key"</c>.</p>
+            <p>The protocol type <c>http</c> is only to be used for
+               the first line when an <c><anno>HttpRequest</anno></c> or an
+               <c><anno>HttpResponse</anno></c> is expected.
+               The following calls are to use <c>httph</c> to get
+               <c><anno>HttpHeader</anno></c>s until
+               <c>http_eoh</c> is returned, which marks the end of the
                headers and the beginning of any following message body.</p> 
-            <p>The variants <c>http_bin</c> and <c>httph_bin</c> will return
+            <p>The variants <c>http_bin</c> and <c>httph_bin</c> return
                strings (<c>HttpString</c>) as binaries instead of lists.</p>
           </item>
         </taglist>
         <p>The following options are available:</p>
             <taglist>
               <tag><c>{packet_size, integer() >= 0}</c></tag>
-              <item><p>Sets the max allowed size of the packet body. If
-                the packet header indicates that the length of the
-                packet is longer than the max allowed length, the packet
-                is considered invalid. Default is 0 which means no
-                size limit.</p>
+              <item><p>Sets the maximum allowed size of the packet body.
+                If the packet header indicates that the length of the
+                packet is longer than the maximum allowed length, the
+                packet is considered invalid. Default is 0, which means
+                no size limit.</p>
               </item>
               <tag><c>{line_length, integer() >= 0}</c></tag>
-              <item><p>For packet type <c>line</c>, truncate lines longer
-	      than the indicated length.</p>
-	      <p>Option <c>line_length</c> also applies to <c>http*</c> 
-	      packet types as an alias for option <c>packet_size</c> in the
-	      case when <c>packet_size</c> itself is not set. This usage is
-	      only intended for backward compatibility.</p>
+              <item><p>For packet type <c>line</c>, lines longer than
+              the indicated length are truncated.</p>
+              <p>Option <c>line_length</c> also applies to <c>http*</c>
+                packet types as an alias for option <c>packet_size</c>
+                if <c>packet_size</c> itself is not set. This use is
+                only intended for backward compatibility.</p>
+              </item>
+              <tag><c>{line_delimiter, 0 =&lt; byte() =&lt; 255}</c></tag>
+              <item><p>For packet type <c>line</c>, sets the delimiting byte.
+                Default is the latin1 character <c>$\n</c>.</p>
               </item>
             </taglist>
+        <p>Examples:</p>
         <pre>
 > <input>erlang:decode_packet(1,&lt;&lt;3,"abcd"&gt;&gt;,[]).</input>
 {ok,&lt;&lt;"abc"&gt;&gt;,&lt;&lt;"d"&gt;&gt;}
@@ -909,13 +955,11 @@
 
     <func>
       <name name="delete_element" arity="2"/>
-      <fsummary>Delete element at index in a tuple</fsummary>
+      <fsummary>Deletes element at index in a tuple.</fsummary>
       <type_desc variable="Index">1..tuple_size(<anno>Tuple1</anno>)</type_desc>
       <desc>
-		  <p>
-			  Returns a new tuple with element at <c><anno>Index</anno></c> removed from
-			  tuple <c><anno>Tuple1</anno></c>.
-		  </p>
+        <p>Returns a new tuple with element at <c><anno>Index</anno></c>
+          removed from tuple <c><anno>Tuple1</anno></c>, for example:</p>
         <pre>
 > <input>erlang:delete_element(2, {one, two, three}).</input>
 {one,three}</pre>
@@ -924,45 +968,49 @@
 
     <func>
       <name name="delete_module" arity="1"/>
-      <fsummary>Make the current code for a module old</fsummary>
+      <fsummary>Makes the current code for a module old.</fsummary>
       <desc>
-        <p>Makes the current code for <c><anno>Module</anno></c> become old code, and
-          deletes all references for this module from the export table.
+        <p>Makes the current code for <c><anno>Module</anno></c> become old code,
+          and deletes all references for this module from the export table.
           Returns <c>undefined</c> if the module does not exist,
           otherwise <c>true</c>.</p>
         <warning>
           <p>This BIF is intended for the code server (see
-            <seealso marker="kernel:code">code(3)</seealso>) and should not be
-            used elsewhere.</p>
+            <seealso marker="kernel:code">code(3)</seealso>) and is not
+            to be used elsewhere.</p>
         </warning>
-        <p>Failure: <c>badarg</c> if there is already an old version of
+        <p>Failure: <c>badarg</c> if there already is an old version of
           <c>Module</c>.</p>
       </desc>
     </func>
+
     <func>
       <name name="demonitor" arity="1"/>
-      <fsummary>Stop monitoring</fsummary>
+      <fsummary>Stops monitoring.</fsummary>
       <desc>
-        <p>If <c><anno>MonitorRef</anno></c> is a reference which the calling process
-          obtained by calling
+        <p>If <c><anno>MonitorRef</anno></c> is a reference that the
+          calling process obtained by calling
           <seealso marker="#monitor/2">monitor/2</seealso>,
           this monitoring is turned off. If the monitoring is already
           turned off, nothing happens.</p>
-        <p>Once <c>demonitor(<anno>MonitorRef</anno>)</c> has returned it is
-          guaranteed that no <c>{'DOWN', <anno>MonitorRef</anno>, _, _, _}</c> message
-          due to the monitor will be placed in the caller's message queue
-          in the future. A <c>{'DOWN', <anno>MonitorRef</anno>, _, _, _}</c> message
-          might have been placed in the caller's message queue prior to
-          the call, though. Therefore, in most cases, it is advisable
+        <p>Once <c>demonitor(<anno>MonitorRef</anno>)</c> has returned, it is
+          guaranteed that no <c>{'DOWN',
+          <anno>MonitorRef</anno>, _, _, _}</c> message,
+          because of the monitor, will be placed in the caller message queue
+          in the future. A <c>{'DOWN',
+          <anno>MonitorRef</anno>, _, _, _}</c> message
+          can have been placed in the caller message queue before
+          the call, though. It is therefore usually advisable
           to remove such a <c>'DOWN'</c> message from the message queue
-          after monitoring has been stopped. 
-          <seealso marker="#demonitor/2">demonitor(<anno>MonitorRef</anno>, [flush])</seealso> can be used instead of
+          after monitoring has been stopped.
+          <seealso marker="#demonitor/2"><c>demonitor(<anno>MonitorRef</anno>, [flush])</c></seealso>
+          can be used instead of
           <c>demonitor(<anno>MonitorRef</anno>)</c> if this cleanup is wanted.</p>
         <note>
-          <p>Prior to OTP release R11B (erts version 5.5) <c>demonitor/1</c>
-            behaved completely asynchronous, i.e., the monitor was active
+          <p>Prior to OTP release R11B (ERTS version 5.5) <c>demonitor/1</c>
+            behaved completely asynchronously, i.e., the monitor was active
             until the "demonitor signal" reached the monitored entity. This
-            had one undesirable effect, though. You could never know when
+            had one undesirable effect. You could never know when
             you were guaranteed <em>not</em> to receive a <c>DOWN</c> message
             due to the monitor.</p>
           <p>Current behavior can be viewed as two combined operations:
@@ -971,31 +1019,31 @@
         </note>
         <p>Failure: It is an error if <c><anno>MonitorRef</anno></c> refers to a
           monitoring started by another process. Not all such cases are
-          cheap to check; if checking is cheap, the call fails with
-          <c>badarg</c> (for example if <c><anno>MonitorRef</anno></c> is a remote
-          reference).</p>
+          cheap to check. If checking is cheap, the call fails with
+          <c>badarg</c> for example, if <c><anno>MonitorRef</anno></c> is a
+          remote reference.</p>
       </desc>
     </func>
+
     <func>
       <name name="demonitor" arity="2"/>
-      <fsummary>Stop monitoring</fsummary>
+      <fsummary>Stops monitoring.</fsummary>
       <desc>
         <p>The returned value is <c>true</c> unless <c>info</c> is part
-	   of <c><anno>OptionList</anno></c>.
-	</p>
+          of <c><anno>OptionList</anno></c>.</p>
         <p><c>demonitor(<anno>MonitorRef</anno>, [])</c> is equivalent to
-          <seealso marker="#demonitor/1">demonitor(<anno>MonitorRef</anno>)</seealso>.</p>
-        <p>Currently the following <c><anno>Option</anno></c>s are valid:</p>
+          <seealso marker="#demonitor/1"><c>demonitor(<anno>MonitorRef</anno>)</c></seealso>.</p>
+        <p>The available <c><anno>Option</anno></c>s are as follows:</p>
         <taglist>
           <tag><c>flush</c></tag>
           <item>
-            <p>Remove (one) <c>{_, <anno>MonitorRef</anno>, _, _, _}</c> message,
-              if there is one, from the caller's message queue after
+            <p>Removes (one) <c>{_,
+              <anno>MonitorRef</anno>, _, _, _}</c> message,
+              if there is one, from the caller message queue after
               monitoring has been stopped.</p>
             <p>Calling <c>demonitor(<anno>MonitorRef</anno>, [flush])</c>
               is equivalent to the following, but more efficient:</p>
             <code type="none">
-
     demonitor(MonitorRef),
     receive
         {_, MonitorRef, _, _, _} ->
@@ -1006,78 +1054,90 @@
           </item>
           <tag><c>info</c></tag>
           <item>
-	    <p>The returned value is one of the following:</p>
-	    <taglist>
-	      <tag><c>true</c></tag>
-	      <item><p>The monitor was found and removed. In this case
-		       no <c>'DOWN'</c> message due to this monitor have
-		       been nor will be placed in the message queue
-		       of the caller.
-		    </p>
-	      </item>
-	      <tag><c>false</c></tag>
-	      <item><p>The monitor was not found and could not be removed.
-	      	       This probably because someone already has placed a
-		       <c>'DOWN'</c> message corresponding to this monitor
-		       in the caller's message queue.
-		    </p>
-	      </item>
-	    </taglist>
-	    <p>If the <c>info</c> option is combined with the <c>flush</c>
-	       option, <c>false</c> will be returned if a flush was needed;
-	       otherwise, <c>true</c>.
-	    </p>
+            <p>The returned value is one of the following:</p>
+            <taglist>
+              <tag><c>true</c></tag>
+              <item>The monitor was found and removed. In this case,
+                no <c>'DOWN'</c> message corresponding to this
+                monitor has been delivered and will not be delivered.
+              </item>
+              <tag><c>false</c></tag>
+              <item>The monitor was not found and could not be removed.
+                This probably because someone already has placed a
+                <c>'DOWN'</c> message corresponding to this monitor
+                in the caller message queue.
+              </item>
+            </taglist>
+            <p>If option <c>info</c> is combined with option <c>flush</c>,
+              <c>false</c> is returned if a flush was needed,
+              otherwise <c>true</c>.</p>
           </item>
         </taglist>
         <note>
-          <p>More options may be added in the future.</p>
+          <p>More options can be added in a future release.</p>
         </note>
-        <p>Failure: <c>badarg</c> if <c><anno>OptionList</anno></c> is not a list, or
-          if <c><anno>Option</anno></c> is not a valid option, or the same failure as for
-          <seealso marker="#demonitor/1">demonitor/1</seealso></p>
+        <p>Failures:</p>
+        <taglist>
+          <tag><c>badarg</c></tag>
+          <item>If <c><anno>OptionList</anno></c> is not a list.
+	  </item>
+          <tag><c>badarg</c></tag>
+          <item>If <c><anno>Option</anno></c> is an invalid option.
+	  </item>
+          <tag><c>badarg</c></tag>
+          <item>The same failure as for
+            <seealso marker="#demonitor/1">demonitor/1</seealso>.
+	  </item>
+        </taglist>
       </desc>
     </func>
+
     <func>
       <name name="disconnect_node" arity="1"/>
-      <fsummary>Force the disconnection of a node</fsummary>
+      <fsummary>Forces the disconnection of a node.</fsummary>
       <desc>
-        <p>Forces the disconnection of a node. This will appear to
-          the node <c><anno>Node</anno></c> as if the local node has crashed. This
-          BIF is mainly used in the Erlang network authentication
-          protocols. Returns <c>true</c> if disconnection succeeds,
+        <p>Forces the disconnection of a node. This appears to
+          the node <c><anno>Node</anno></c> as if the local node has crashed.
+          This BIF is mainly used in the Erlang network authentication
+          protocols.</p>
+        <p>Returns <c>true</c> if disconnection succeeds,
           otherwise <c>false</c>. If the local node is not alive,
-          the function returns <c>ignored</c>.</p>
+          <c>ignored</c> is returned.</p>
       </desc>
     </func>
+
     <func>
       <name name="display" arity="1"/>
-      <fsummary>Print a term on standard output</fsummary>
+      <fsummary>Prints a term on standard output.</fsummary>
       <desc>
-        <p>Prints a text representation of <c><anno>Term</anno></c> on the standard
-          output. On OSE the term is printed to the ramlog.</p>
+        <p>Prints a text representation of <c><anno>Term</anno></c> on the
+          standard output.</p>
         <warning>
           <p>This BIF is intended for debugging only.</p>
         </warning>
       </desc>
     </func>
+
     <func>
       <name name="element" arity="2"/>
+      <fsummary>Returns the Nth element of a tuple.</fsummary>
       <type_desc variable="N">1..tuple_size(<anno>Tuple</anno>)</type_desc>
-      <fsummary>Get Nth element of a tuple</fsummary>
       <desc>
         <p>Returns the <c><anno>N</anno></c>th element (numbering from 1) of
-          <c><anno>Tuple</anno></c>.</p>
+          <c><anno>Tuple</anno></c>, for example:</p>
         <pre>
 > <input>element(2, {a, b, c}).</input>
 b</pre>
         <p>Allowed in guard tests.</p>
       </desc>
     </func>
+
     <func>
       <name name="erase" arity="0"/>
-      <fsummary>Return and delete the process dictionary</fsummary>
+      <fsummary>Returns and deletes the process dictionary.</fsummary>
       <desc>
-        <p>Returns the process dictionary and deletes it.</p>
+        <p>Returns the process dictionary and deletes it, for
+          example:</p>
         <pre>
 > <input>put(key1, {1, 2, 3}),</input>
 <input>put(key2, [a, b, c]),</input>
@@ -1085,13 +1145,16 @@ b</pre>
 [{key1,{1,2,3}},{key2,[a,b,c]}]</pre>
       </desc>
     </func>
+
     <func>
       <name name="erase" arity="1"/>
-      <fsummary>Return and delete a value from the process dictionary</fsummary>
+      <fsummary>Returns and deletes a value from the process dictionary.</fsummary>
       <desc>
-        <p>Returns the value <c><anno>Val</anno></c> associated with <c><anno>Key</anno></c> and
-          deletes it from the process dictionary. Returns
-          <c>undefined</c> if no value is associated with <c><anno>Key</anno></c>.</p>
+        <p>Returns the value <c><anno>Val</anno></c> associated with
+          <c><anno>Key</anno></c> and deletes it from the process dictionary.
+          Returns <c>undefined</c> if no value is associated with
+          <c><anno>Key</anno></c>.</p>
+        <p>Example:</p>
         <pre>
 > <input>put(key1, {merry, lambs, are, playing}),</input>
 <input>X = erase(key1),</input>
@@ -1099,16 +1162,19 @@ b</pre>
 {{merry,lambs,are,playing},undefined}</pre>
       </desc>
     </func>
+
     <func>
       <name name="error" arity="1"/>
-      <fsummary>Stop execution with a given reason</fsummary>
+      <fsummary>Stops execution with a given reason.</fsummary>
       <desc>
         <p>Stops the execution of the calling process with the reason
-          <c><anno>Reason</anno></c>, where <c><anno>Reason</anno></c> is any term. The actual
-          exit reason will be <c>{<anno>Reason</anno>, Where}</c>, where <c>Where</c>
+          <c><anno>Reason</anno></c>, where <c><anno>Reason</anno></c>
+          is any term. The exit reason is
+          <c>{<anno>Reason</anno>, Where}</c>, where <c>Where</c>
           is a list of the functions most recently called (the current
           function first). Since evaluating this function causes
           the process to terminate, it has no return value.</p>
+        <p>Example:</p>
         <pre>
 > <input>catch error(foobar).</input>
 {'EXIT',{foobar,[{erl_eval,do_apply,5},
@@ -1118,29 +1184,34 @@ b</pre>
                  {shell,eval_loop,3}]}}</pre>
       </desc>
     </func>
+
     <func>
       <name name="error" arity="2"/>
-      <fsummary>Stop execution with a given reason</fsummary>
+      <fsummary>Stops execution with a given reason.</fsummary>
       <desc>
         <p>Stops the execution of the calling process with the reason
-          <c><anno>Reason</anno></c>, where <c><anno>Reason</anno></c> is any term. The actual
-          exit reason will be <c>{<anno>Reason</anno>, Where}</c>, where <c>Where</c>
+          <c><anno>Reason</anno></c>, where <c><anno>Reason</anno></c>
+          is any term. The exit reason is
+          <c>{<anno>Reason</anno>, Where}</c>, where <c>Where</c>
           is a list of the functions most recently called (the current
-          function first). <c><anno>Args</anno></c> is expected to be the list of
-          arguments for the current function; in Beam it will be used
-          to provide the actual arguments for the current function in
-          the <c>Where</c> term. Since evaluating this function causes
+          function first). <c><anno>Args</anno></c> is expected to be the
+          list of arguments for the current function; in Beam it is used
+          to provide the arguments for the current function in
+          the term <c>Where</c>. Since evaluating this function causes
           the process to terminate, it has no return value.</p>
       </desc>
     </func>
+
     <func>
       <name name="exit" arity="1"/>
-      <fsummary>Stop execution with a given reason</fsummary>
+      <fsummary>Stops execution with a given reason.</fsummary>
       <desc>
-        <p>Stops the execution of the calling process with the exit
-          reason <c><anno>Reason</anno></c>, where <c><anno>Reason</anno></c> is any term. Since
+        <p>Stops the execution of the calling process with exit reason
+          <c><anno>Reason</anno></c>, where <c><anno>Reason</anno></c>
+          is any term. Since
           evaluating this function causes the process to terminate, it
           has no return value.</p>
+        <p>Example:</p>
         <pre>
 > <input>exit(foobar).</input>
 ** exception exit: foobar
@@ -1148,110 +1219,117 @@ b</pre>
 {'EXIT',foobar}</pre>
       </desc>
     </func>
+
     <func>
       <name name="exit" arity="2"/>
-      <fsummary>Send an exit signal to a process or a port</fsummary>
+      <fsummary>Sends an exit signal to a process or a port.</fsummary>
       <desc>
         <p>Sends an exit signal with exit reason <c><anno>Reason</anno></c> to
           the process or port identified by <c><anno>Pid</anno></c>.</p>
-        <p>The following behavior apply if <c><anno>Reason</anno></c> is any term
-          except <c>normal</c> or <c>kill</c>:</p>
-        <p>If <c><anno>Pid</anno></c> is not trapping exits, <c><anno>Pid</anno></c> itself will
-          exit with exit reason <c><anno>Reason</anno></c>. If <c><anno>Pid</anno></c> is trapping
-          exits, the exit signal is transformed into a message
-          <c>{'EXIT', From, <anno>Reason</anno>}</c> and delivered to the message
-          queue of <c><anno>Pid</anno></c>. <c>From</c> is the pid of the process
-          which sent the exit signal. See also
-          <seealso marker="#process_flag/2">process_flag/2</seealso>.</p>
-        <p>If <c><anno>Reason</anno></c> is the atom <c>normal</c>, <c><anno>Pid</anno></c> will
-          not exit. If it is trapping exits, the exit signal is
-          transformed into a message <c>{'EXIT', From, normal}</c>
-          and delivered to its message queue.</p>
-        <p>If <c><anno>Reason</anno></c> is the atom <c>kill</c>, that is if
-          <c>exit(<anno>Pid</anno>, kill)</c> is called, an untrappable exit signal
-          is sent to <c><anno>Pid</anno></c> which will unconditionally exit with
-          exit reason <c>killed</c>.</p>
+        <p>The following behavior applies if <c><anno>Reason</anno></c>
+          is any term, except <c>normal</c> or <c>kill</c>:</p>
+          <list type="bulleted">
+            <item>If <c><anno>Pid</anno></c> is not trapping exits,
+              <c><anno>Pid</anno></c>
+              itself exits with exit reason <c><anno>Reason</anno></c>.
+            </item>
+            <item>If <c><anno>Pid</anno></c> is trapping exits, the exit
+              signal is transformed into a message
+              <c>{'EXIT', From, <anno>Reason</anno>}</c>
+              and delivered to the message queue of <c><anno>Pid</anno></c>.
+            </item>
+            <item><c>From</c> is the process identifier of the process
+              that sent the exit signal. See also
+              <seealso marker="#process_flag/2">process_flag/2</seealso>.
+            </item>
+          </list>
+          <p>If <c><anno>Reason</anno></c> is the atom <c>normal</c>,
+            <c><anno>Pid</anno></c>
+            does not exit. If it is trapping exits, the exit signal is
+            transformed into a message <c>{'EXIT', From, normal}</c>
+            and delivered to its message queue.</p>
+          <p>If <c><anno>Reason</anno></c> is the atom <c>kill</c>,
+            that is, if <c>exit(<anno>Pid</anno>, kill)</c> is called,
+            an untrappable exit signal is sent to <c><anno>Pid</anno></c>,
+            which unconditionally exits with exit reason <c>killed</c>.
+          </p>
       </desc>
     </func>
+
     <func>
       <name name="external_size" arity="1"/>
-      <fsummary>Calculate the maximum size for a term encoded in the Erlang
-        external term format</fsummary>
+      <fsummary>Calculates the maximum size for a term encoded in the Erlang external term format.</fsummary>
       <desc>
         <p>Calculates, without doing the encoding, the maximum byte size for
           a term encoded in the Erlang external term format. The following
           condition applies always:</p>
-        <p>
           <pre>
 > <input>Size1 = byte_size(term_to_binary(<anno>Term</anno>)),</input>
 > <input>Size2 = erlang:external_size(<anno>Term</anno>),</input>
 > <input>true = Size1 =&lt; Size2.</input>
-true
-          </pre>
-        </p>
-        <p>This is equivalent to a call to: <code>erlang:external_size(<anno>Term</anno>, [])
-          </code></p>
+true</pre>
+        <p>This is equivalent to a call to:</p>
+<code>erlang:external_size(<anno>Term</anno>, [])</code>
       </desc>
     </func>
+
     <func>
       <name name="external_size" arity="2"/>
-      <fsummary>Calculate the maximum size for a term encoded in the Erlang
-        external term format</fsummary>
+      <fsummary>Calculates the maximum size for a term encoded in the Erlang external term format.</fsummary>
       <desc>
         <p>Calculates, without doing the encoding, the maximum byte size for
           a term encoded in the Erlang external term format. The following
           condition applies always:</p>
-        <p>
           <pre>
 > <input>Size1 = byte_size(term_to_binary(<anno>Term</anno>, <anno>Options</anno>)),</input>
 > <input>Size2 = erlang:external_size(<anno>Term</anno>, <anno>Options</anno>),</input>
 > <input>true = Size1 =&lt; Size2.</input>
-true
-          </pre>
-        </p>
-        <p>The option <c>{minor_version, <anno>Version</anno>}</c> specifies how floats
-          are encoded. See
-          <seealso marker="#term_to_binary/2">term_to_binary/2</seealso> for
-          a more detailed description.
-        </p>
+true</pre>
+        <p>Option <c>{minor_version, <anno>Version</anno>}</c> specifies how
+          floats are encoded. For a detailed description, see
+          <seealso marker="#term_to_binary/2">term_to_binary/2</seealso>.</p>
       </desc>
     </func>
+
     <func>
       <name name="float" arity="1"/>
-      <fsummary>Convert a number to a float</fsummary>
+      <fsummary>Converts a number to a float.</fsummary>
       <desc>
-        <p>Returns a float by converting <c><anno>Number</anno></c> to a float.</p>
+        <p>Returns a float by converting <c><anno>Number</anno></c> to a float,
+          for example:</p>
         <pre>
 > <input>float(55).</input>
 55.0</pre>
         <p>Allowed in guard tests.</p>
         <note>
-          <p>Note that if used on the top-level in a guard, it will
-            test whether the argument is a floating point number; for
-            clarity, use
+          <p>If used on the top level in a guard, it tests whether the
+            argument is a floating point number; for clarity, use
             <seealso marker="#is_float/1">is_float/1</seealso> instead.</p>
           <p>When <c>float/1</c> is used in an expression in a guard,
             such as '<c>float(A) == 4.0</c>', it converts a number as
-            described above.</p>
+            described earlier.</p>
         </note>
       </desc>
     </func>
+
     <func>
       <name name="float_to_binary" arity="1"/>
-      <fsummary>Text representation of a float</fsummary>
+      <fsummary>Text representation of a float.</fsummary>
       <desc>
-        <p>The same as <c>float_to_binary(<anno>Float</anno>,[{scientific,20}])</c>.</p>
+        <p>The same as
+          <c>float_to_binary(<anno>Float</anno>,[{scientific,20}])</c>.</p>
       </desc>
     </func>
+
     <func>
       <name name="float_to_binary" arity="2"/>
-      <fsummary>Text representation of a float formatted using given options</fsummary>
+      <fsummary>Text representation of a float formatted using given options.</fsummary>
       <desc>
-        <p>Returns a binary which corresponds to the text
+        <p>Returns a binary corresponding to the text
           representation of <c><anno>Float</anno></c> using fixed decimal 
-	  point formatting. The <c><anno>Options</anno></c> behave in the same
-	  way as <seealso marker="#float_to_list/2">float_to_list/2</seealso>.
-	</p>
+          point formatting. <c><anno>Options</anno></c> behaves in the same
+          way as <seealso marker="#float_to_list/2">float_to_list/2</seealso>.</p>
+        <p>Examples:</p>
         <pre>
 > <input>float_to_binary(7.12, [{decimals, 4}]).</input>
 &lt;&lt;"7.1200">>
@@ -1259,31 +1337,42 @@ true
 &lt;&lt;"7.12">></pre>
       </desc>
     </func>
+
     <func>
       <name name="float_to_list" arity="1"/>
-      <fsummary>Text representation of a float</fsummary>
+      <fsummary>Text representation of a float.</fsummary>
       <desc>
-	<p>The same as <c>float_to_list(<anno>Float</anno>,[{scientific,20}])</c>.</p>
+        <p>The same as
+          <c>float_to_list(<anno>Float</anno>,[{scientific,20}])</c>.</p>
       </desc>
     </func>
+
     <func>
       <name name="float_to_list" arity="2"/>
-      <fsummary>Text representation of a float formatted using given options</fsummary>
-      <desc>
-        <p>Returns a string which corresponds to the text
-          representation of <c>Float</c> using fixed decimal point formatting.
-          When <c>decimals</c> option is specified
-          the returned value will contain at most <c>Decimals</c> number of
-          digits past the decimal point.  If the number doesn't fit in the
-          internal static buffer of 256 bytes, the function throws <c>badarg</c>.
-          When <c>compact</c> option is provided
-          the trailing zeros at the end of the list are truncated (this option is
-          only meaningful together with the <c>decimals</c> option). When
-          <c>scientific</c> option is provided, the float will be formatted using
-          scientific notation with <c>Decimals</c> digits of precision. If
-          <c>Options</c> is <c>[]</c> the function behaves like 
-	  <c><seealso marker="#float_to_list/1">float_to_list/1</seealso></c>.
-          </p>
+      <fsummary>Text representation of a float formatted using given options.</fsummary>
+      <desc>
+        <p>Returns a string corresponding to the text representation
+          of <c>Float</c> using fixed decimal point formatting. The
+          options are as follows:</p>
+        <list type="bulleted">
+          <item>If option <c>decimals</c> is specified, the returned value
+            contains at most <c>Decimals</c> number of digits past the
+            decimal point.  If the number does not fit in the internal
+            static buffer of 256 bytes, the function throws <c>badarg</c>.
+          </item>
+          <item>If option <c>compact</c> is provided, the trailing zeros
+            at the end of the list are truncated. This option is only
+            meaningful together with option <c>decimals</c>.
+          </item>
+          <item>If option <c>scientific</c> is provided, the float is
+            formatted using scientific notation with <c>Decimals</c>
+            digits of precision.
+          </item>
+          <item>If <c>Options</c> is <c>[]</c>, the function behaves as
+            <seealso marker="#float_to_list/1">float_to_list/1</seealso>.
+          </item>
+        </list>
+        <p>Examples:</p>
         <pre>
 > <input>float_to_list(7.12, [{decimals, 4}]).</input>
 "7.1200"
@@ -1291,36 +1380,40 @@ true
 "7.12"</pre>
       </desc>
     </func>
+
     <func>
       <name name="fun_info" arity="1"/>
-      <fsummary>Information about a fun</fsummary>
+      <fsummary>Information about a fun.</fsummary>
       <desc>
-        <p>Returns a list containing information about the fun
-          <c><anno>Fun</anno></c>. Each element of the list is a tuple. The order of
-          the tuples is not defined, and more tuples may be added in a
+        <p>Returns a list with information about the fun
+          <c><anno>Fun</anno></c>. Each list element is a tuple. The order
+          of the tuples is undefined, and more tuples can be added in a
           future release.</p>
         <warning>
           <p>This BIF is mainly intended for debugging, but it can
-            occasionally be useful in library functions that might need
-            to verify, for instance, the arity of a fun.</p>
+            sometimes be useful in library functions that need
+            to verify, for example, the arity of a fun.</p>
         </warning>
-        <p>There are two types of funs with slightly different
-          semantics:</p>
-        <p>A fun created by <c>fun M:F/A</c> is called an
-          <em>external</em> fun. Calling it will always call the
-          function <c>F</c> with arity <c>A</c> in the latest code for
-          module <c>M</c>. Note that module <c>M</c> does not even need
-          to be loaded when the fun <c>fun M:F/A</c> is created.</p>
-        <p>All other funs are called <em>local</em>. When a local fun
-          is called, the same version of the code that created the fun
-          will be called (even if newer version of the module has been
-          loaded).</p>
-        <p>The following elements will always be present in the list
+        <p>Two types of funs have slightly different semantics:</p>
+        <list type="bulleted">
+          <item>A fun created by <c>fun M:F/A</c> is called an
+            <em>external</em> fun. Calling it will always call the
+            function <c>F</c> with arity <c>A</c> in the latest code for
+            module <c>M</c>. Notice that module <c>M</c> does not even
+            need to be loaded when the fun <c>fun M:F/A</c> is created.
+          </item>
+          <item>All other funs are called <em>local</em>. When a local fun
+            is called, the same version of the code that created the fun
+            is called (even if a newer version of the module has been
+            loaded).
+          </item>
+        </list>
+        <p>The following elements are always present in the list
           for both local and external funs:</p>
         <taglist>
           <tag><c>{type, Type}</c></tag>
           <item>
-            <p><c>Type</c> is either <c>local</c> or <c>external</c>.</p>
+            <p><c>Type</c> is <c>local</c> or <c>external</c>.</p>
           </item>
           <tag><c>{module, Module}</c></tag>
           <item>
@@ -1335,148 +1428,154 @@ true
             <p><c>Name</c> (an atom) is a function name.</p>
             <p>If <c>Fun</c> is a local fun, <c>Name</c> is the name
               of the local function that implements the fun.
-              (This name was generated by the compiler, and is generally
+              (This name was generated by the compiler, and is
               only of informational use. As it is a local function, it
-              is not possible to call it directly.)
+              cannot be called directly.)
               If no code is currently loaded for the fun, <c>[]</c>
-              will be returned instead of an atom.</p>
+              is returned instead of an atom.</p>
             <p>If <c>Fun</c> is an external fun, <c>Name</c> is the name
               of the exported function that the fun refers to.</p>
           </item>
           <tag><c>{arity, Arity}</c></tag>
           <item>
             <p><c>Arity</c> is the number of arguments that the fun
-              should be called with.</p>
+              is to be called with.</p>
           </item>
           <tag><c>{env, Env}</c></tag>
           <item>
             <p><c>Env</c> (a list) is the environment or free variables
-              for the fun. (For external funs, the returned list is
-              always empty.)</p>
+              for the fun. For external funs, the returned list is
+              always empty.</p>
           </item>
         </taglist>
-        <p>The following elements will only be present in the list if
+        <p>The following elements are only present in the list if
           <c>Fun</c> is local:</p>
         <taglist>
           <tag><c>{pid, Pid}</c></tag>
           <item>
-            <p><c>Pid</c> is the pid of the process that originally
-              created the fun.</p>
+            <p><c>Pid</c> is the process identifier of the process
+              that originally created the fun.</p>
           </item>
           <tag><c>{index, Index}</c></tag>
           <item>
-            <p><c>Index</c> (an integer) is an index into the module's
+            <p><c>Index</c> (an integer) is an index into the module
               fun table.</p>
           </item>
           <tag><c>{new_index, Index}</c></tag>
           <item>
-            <p><c>Index</c> (an integer) is an index into the module's
+            <p><c>Index</c> (an integer) is an index into the module
               fun table.</p>
           </item>
           <tag><c>{new_uniq, Uniq}</c></tag>
           <item>
-            <p><c>Uniq</c> (a binary) is a unique value for this fun.
-	    It is calculated from the compiled code for the entire module.</p>
+            <p><c>Uniq</c> (a binary) is a unique value for this fun. It
+            is calculated from the compiled code for the entire module.</p>
           </item>
           <tag><c>{uniq, Uniq}</c></tag>
           <item>
             <p><c>Uniq</c> (an integer) is a unique value for this fun.
-	    Starting in the R15 release, this integer is calculated from
-	    the compiled code for the entire module. Before R15, this
-	    integer was based on only the body of the fun.
-	    </p>
+            As from OTP R15, this integer is calculated from the
+            compiled code for the entire module. Before OTP R15, this
+            integer was based on only the body of the fun.</p>
           </item>
         </taglist>
       </desc>
     </func>
+
     <func>
       <name name="fun_info" arity="2"/>
+      <fsummary>Information about a fun.</fsummary>
       <type name="fun_info_item"/>
-      <fsummary>Information about a fun</fsummary>
       <desc>
         <p>Returns information about <c><anno>Fun</anno></c> as specified by
-          <c><anno>Item</anno></c>, in the form <c>{<anno>Item</anno>,<anno>Info</anno>}</c>.</p>
+          <c><anno>Item</anno></c>, in the form
+          <c>{<anno>Item</anno>,<anno>Info</anno>}</c>.</p>
         <p>For any fun, <c><anno>Item</anno></c> can be any of the atoms
-          <c>module</c>, <c>name</c>, <c>arity</c>, <c>env</c>, or <c>type</c>.</p>
-        <p>For a local fun, <c><anno>Item</anno></c> can also be any of the atoms
-          <c>index</c>, <c>new_index</c>, <c>new_uniq</c>,
+          <c>module</c>, <c>name</c>, <c>arity</c>, <c>env</c>, or
+          <c>type</c>.</p>
+        <p>For a local fun, <c><anno>Item</anno></c> can also be any of the
+          atoms <c>index</c>, <c>new_index</c>, <c>new_uniq</c>,
           <c>uniq</c>, and <c>pid</c>. For an external fun, the value
           of any of these items is always the atom <c>undefined</c>.</p>
         <p>See
           <seealso marker="#fun_info/1">erlang:fun_info/1</seealso>.</p>
       </desc>
     </func>
+
     <func>
       <name name="fun_to_list" arity="1"/>
-      <fsummary>Text representation of a fun</fsummary>
+      <fsummary>Text representation of a fun.</fsummary>
       <desc>
-        <p>Returns a string which corresponds to the text
+        <p>Returns a string corresponding to the text
           representation of <c><anno>Fun</anno></c>.</p>
       </desc>
     </func>
+
     <func>
       <name name="function_exported" arity="3"/>
-      <fsummary>Check if a function is exported and loaded</fsummary>
+      <fsummary>Checks if a function is exported and loaded.</fsummary>
       <desc>
         <p>Returns <c>true</c> if the module <c><anno>Module</anno></c> is loaded
           and contains an exported function <c><anno>Function</anno>/<anno>Arity</anno></c>,
 	  or if there is a BIF (a built-in function implemented in C)
-	  with the given name; otherwise returns <c>false</c>.</p>
+	  with the given name, otherwise returns <c>false</c>.</p>
 	  <note><p>This function used to return false for built-in
 	  functions before the 18.0 release.</p></note>
       </desc>
     </func>
+
     <func>
       <name name="garbage_collect" arity="0"/>
-      <fsummary>Force an immediate garbage collection of the calling process</fsummary>
+      <fsummary>Forces an immediate garbage collection of the calling process.</fsummary>
       <desc>
-        <p>Forces an immediate garbage collection of the currently
-          executing process. The function should not be used, unless
-          it has been noticed -- or there are good reasons to suspect --
+        <p>Forces an immediate garbage collection of the
+          executing process. The function is not to be used unless
+          it has been noticed (or there are good reasons to suspect)
           that the spontaneous garbage collection will occur too late
-          or not at all. Improper use may seriously degrade system
-          performance.</p>
+          or not at all.</p>
+        <warning>
+          <p>Improper use can seriously degrade system performance.</p>
+        </warning>
       </desc>
     </func>
+
     <func>
       <name name="garbage_collect" arity="1"/>
-      <fsummary>Garbage collect a process</fsummary>
+      <fsummary>Garbage collects a process.</fsummary>
       <desc>
         <p>The same as
 	<seealso marker="#garbage_collect/2"><c>garbage_collect(<anno>Pid</anno>, [])</c></seealso>.</p>
       </desc>
     </func>
+
     <func>
       <name name="garbage_collect" arity="2"/>
-      <fsummary>Garbage collect a process</fsummary>
+      <fsummary>Garbage collects a process.</fsummary>
       <desc>
-	<p>Garbage collect the node local process identified by
-	<c><anno>Pid</anno></c>.</p>
-	<p>Currently available <c><anno>Option</anno></c>s:</p>
+        <p>Garbage collects the node local process identified by
+        <c><anno>Pid</anno></c>.</p>
+        <p>The available <c><anno>Option</anno></c>s are as follows:</p>
         <taglist>
           <tag><c>{async, RequestId}</c></tag>
-          <item>
-	    The <c>garbage_collect/2</c> function will return
+          <item>The function <c>garbage_collect/2</c> returns
 	    the value <c>async</c> immediately after the request
 	    has been sent. When the request has been processed, the
-	    process that called this function will be passed a
-	    message on the form:<br/>
-	    <c>{garbage_collect, <anno>RequestId</anno>, <anno>GCResult</anno>}</c>.
-	  </item>
+	    process that called this function is passed a message on
+            the form <c>{garbage_collect,
+            <anno>RequestId</anno>, <anno>GCResult</anno>}</c>.
+          </item>
         </taglist>
 	<p>If <c><anno>Pid</anno></c> equals <c>self()</c>, and
 	no <c>async</c> option has been passed, the garbage
-	collection will be performed at once, i.e. the same as
-	calling
+	collection is performed at once, that is, the same as calling
 	<seealso marker="#garbage_collect/0">garbage_collect/0</seealso>.
-	In all other cases a request for garbage collection will
-	be sent to the process identified by <c><anno>Pid</anno></c>,
+	Otherwise a request for garbage collection
+	is sent to the process identified by <c><anno>Pid</anno></c>,
 	and will be handled when appropriate. If no <c>async</c>
-	option has been passed, the caller will block until
-	<c><anno>GCResult</anno></c> is available and can be
-	returned.</p>
+	option has been passed, the caller blocks until
+	<c><anno>GCResult</anno></c> is available and can be returned.</p>
 	<p><c><anno>GCResult</anno></c> informs about the result of
-	the garbage collection request:</p>
+	the garbage collection request as follows:</p>
         <taglist>
           <tag><c>true</c></tag>
           <item>
@@ -1485,14 +1584,13 @@ true
 	  </item>
           <tag><c>false</c></tag>
           <item>
-	    No garbage collection was performed. This since the
+	    No garbage collection was performed, as
 	    the process identified by <c><anno>Pid</anno></c>
 	    terminated before the request could be satisfied.
 	  </item>
         </taglist>
-        <p>Note that the same caveats as for
-	<seealso marker="#garbage_collect/0">garbage_collect/0</seealso>
-	apply.</p>
+        <p>Notice that the same caveats apply as for
+	<seealso marker="#garbage_collect/0">garbage_collect/0</seealso>.</p>
         <p>Failures:</p>
         <taglist>
           <tag><c>badarg</c></tag>
@@ -1501,17 +1599,18 @@ true
 	  </item>
           <tag><c>badarg</c></tag>
           <item>
-	    If <c><anno>OptionList</anno></c> is not a valid list of options.
+	    If <c><anno>OptionList</anno></c> is an invalid list of options.
 	  </item>
         </taglist>
       </desc>
     </func>
+
     <func>
       <name name="get" arity="0"/>
-      <fsummary>Return the process dictionary</fsummary>
+      <fsummary>Returns the process dictionary.</fsummary>
       <desc>
         <p>Returns the process dictionary as a list of
-          <c>{<anno>Key</anno>, <anno>Val</anno>}</c> tuples.</p>
+          <c>{<anno>Key</anno>, <anno>Val</anno>}</c> tuples, for example:</p>
         <pre>
 > <input>put(key1, merry),</input>
 <input>put(key2, lambs),</input>
@@ -1520,13 +1619,15 @@ true
 [{key1,merry},{key2,lambs},{key3,{are,playing}}]</pre>
       </desc>
     </func>
+
     <func>
       <name name="get" arity="1"/>
-      <fsummary>Return a value from the process dictionary</fsummary>
+      <fsummary>Returns a value from the process dictionary.</fsummary>
       <desc>
         <p>Returns the value <c><anno>Val</anno></c> associated with <c><anno>Key</anno></c> in
           the process dictionary, or <c>undefined</c> if <c><anno>Key</anno></c>
           does not exist.</p>
+        <p>Example:</p>
         <pre>
 > <input>put(key1, merry),</input>
 <input>put(key2, lambs),</input>
@@ -1535,14 +1636,16 @@ true
 {are,playing}</pre>
       </desc>
     </func>
+
     <func>
       <name name="get_cookie" arity="0"/>
-      <fsummary>Get the magic cookie of the local node</fsummary>
+      <fsummary>Gets the magic cookie of the local node.</fsummary>
       <desc>
-        <p>Returns the magic cookie of the local node, if the node is
-          alive; otherwise the atom <c>nocookie</c>.</p>
+        <p>Returns the magic cookie of the local node if the node is
+          alive, otherwise the atom <c>nocookie</c>.</p>
       </desc>
     </func>
+
     <func>
       <name name="get_keys" arity="0"/>
       <fsummary>Return a list of all keys from the process dictionary</fsummary>
@@ -1558,10 +1661,10 @@ true
     </func>
     <func>
       <name name="get_keys" arity="1"/>
-      <fsummary>Return a list of keys from the process dictionary</fsummary>
+      <fsummary>Returns a list of keys from the process dictionary.</fsummary>
       <desc>
-        <p>Returns a list of keys which are associated with the value
-          <c><anno>Val</anno></c> in the process dictionary.</p>
+        <p>Returns a list of keys that are associated with the value
+          <c><anno>Val</anno></c> in the process dictionary, for example:</p>
         <pre>
 > <input>put(mary, {1, 2}),</input>
 <input>put(had, {1, 2}),</input>
@@ -1573,40 +1676,40 @@ true
 [mary,had,a,little,lamb]</pre>
       </desc>
     </func>
+
     <func>
       <name name="get_stacktrace" arity="0"/>
-      <fsummary>Get the call stack back-trace of the last exception</fsummary>
+      <fsummary>Gets the call stack back-trace of the last exception.</fsummary>
       <type name="stack_item"/>
       <desc>
-        <p>Get the call stack back-trace (<em>stacktrace</em>) of the last
-          exception in the calling process as a list of 
+        <p>Gets the call stack back-trace (<em>stacktrace</em>) of the
+          last exception in the calling process as a list of
           <c>{<anno>Module</anno>,<anno>Function</anno>,<anno>Arity</anno>,<anno>Location</anno>}</c> tuples.
-          The <c><anno>Arity</anno></c> field in the first tuple may be the argument
-          list of that function call instead of an arity integer,
+          Field <c><anno>Arity</anno></c> in the first tuple can be the
+          argument list of that function call instead of an arity integer,
           depending on the exception.</p>
         <p>If there has not been any exceptions in a process, the
           stacktrace is <c>[]</c>. After a code change for the process,
-          the stacktrace may also be reset to [].</p>
+          the stacktrace can also be reset to <c>[]</c>.</p>
         <p>The stacktrace is the same data as the <c>catch</c> operator
           returns, for example:</p>
         <p><c>{'EXIT',{badarg,Stacktrace}} = catch abs(x)</c></p>
-	<p><c><anno>Location</anno></c> is a (possibly empty) list of two-tuples that
-	  may indicate the location in the source code of the function.
-	  The first element is an atom that describes the type of
-	  information in the second element. Currently the following
-	  items may occur:</p>
+	<p><c><anno>Location</anno></c> is a (possibly empty) list
+          of two-tuples that
+	  can indicate the location in the source code of the function.
+	  The first element is an atom describing the type of
+	  information in the second element. The following
+	  items can occur:</p>
 	<taglist>
 	  <tag><c>file</c></tag>
-	  <item>
-	    <p>The second element of the tuple is a string (list of
-	      characters) representing the filename of the source file
-	      of the function.</p>
+	  <item>The second element of the tuple is a string (list of
+	      characters) representing the file name of the source file
+	      of the function.
 	  </item>
 	  <tag><c>line</c></tag>
-	  <item>
-	    <p>The second element of the tuple is the line number
+	  <item>The second element of the tuple is the line number
 	      (an integer greater than zero) in the source file
-	      where the exception occurred or the function was called.</p>
+	      where the exception occurred or the function was called.
 	  </item>
 	</taglist>
         <p>See also
@@ -1614,49 +1717,56 @@ true
           <seealso marker="#error/2">erlang:error/2</seealso>.</p>
       </desc>
     </func>
+
     <func>
       <name name="group_leader" arity="0"/>
-      <fsummary>Get the group leader for the calling process</fsummary>
+      <fsummary>Gets the group leader for the calling process.</fsummary>
       <desc>
-        <p>Returns the pid of the group leader for the process which
-          evaluates the function.</p>
+        <p>Returns the process identifier of the group leader for the
+          process evaluating the function.</p>
         <p>Every process is a member of some process group and all
-          groups have a <em>group leader</em>. All IO from the group
+          groups have a <em>group leader</em>. All I/O from the group
           is channeled to the group leader. When a new process is
           spawned, it gets the same group leader as the spawning
           process. Initially, at system start-up, <c>init</c> is both
           its own group leader and the group leader of all processes.</p>
       </desc>
     </func>
+
     <func>
       <name name="group_leader" arity="2"/>
-      <fsummary>Set the group leader for a process</fsummary>
+      <fsummary>Sets the group leader for a process.</fsummary>
       <desc>
-        <p>Sets the group leader of <c><anno>Pid</anno></c> to <c><anno>GroupLeader</anno></c>.
-          Typically, this is used when a processes started from a
-          certain shell should have another group leader than
+        <p>Sets the group leader of <c><anno>Pid</anno></c>
+          to <c><anno>GroupLeader</anno></c>.
+          Typically, this is used when a process started from a
+          certain shell is to have another group leader than
           <c>init</c>.</p>
         <p>See also
           <seealso marker="#group_leader/0">group_leader/0</seealso>.</p>
       </desc>
     </func>
+
     <func>
       <name name="halt" arity="0"/>
-      <fsummary>Halt the Erlang runtime system and indicate normal exit to the calling environment</fsummary>
+      <fsummary>Halts the Erlang runtime system and indicates normal exit to the calling environment.</fsummary>
       <desc>
 	<p>The same as
 	  <seealso marker="#halt/2"><c>halt(0, [])</c></seealso>.</p>
+        <p>Example:</p>
         <pre>
 > <input>halt().</input>
 os_prompt% </pre>
       </desc>
     </func>
+
     <func>
       <name name="halt" arity="1"/>
-      <fsummary>Halt the Erlang runtime system</fsummary>
+      <fsummary>Halts the Erlang runtime system.</fsummary>
       <desc>
 	<p>The same as
 	  <seealso marker="#halt/2"><c>halt(<anno>Status</anno>, [])</c></seealso>.</p>
+        <p>Example:</p>
         <pre>
 > <input>halt(17).</input>
 os_prompt% <input>echo $?</input>
@@ -1664,178 +1774,188 @@ os_prompt% <input>echo $?</input>
 os_prompt% </pre>
       </desc>
     </func>
+
     <func>
       <name name="halt" arity="2"/>
-      <fsummary>Halt the Erlang runtime system</fsummary>
+      <fsummary>Halts the Erlang runtime system.</fsummary>
       <desc>
         <p><c><anno>Status</anno></c> must be a non-negative integer, a string,
 	  or the atom <c>abort</c>.
 	  Halts the Erlang runtime system. Has no return value.
-	  Depending on <c><anno>Status</anno></c>:
-	</p>
+	  Depending on <c><anno>Status</anno></c>, the following occurs:</p>
 	<taglist>
 	  <tag>integer()</tag>
-	  <item>The runtime system exits with the integer value <c><anno>Status</anno></c>
-	  as status code to the calling environment (operating system).
+	  <item>The runtime system exits with integer value
+          <c><anno>Status</anno></c>
+	  as status code to the calling environment (OS).
 	  </item>
 	  <tag>string()</tag>
-	  <item>An erlang crash dump is produced with <c><anno>Status</anno></c> as slogan,
-	  and then the runtime system exits with status code <c>1</c>.
+	  <item>An Erlang crash dump is produced with <c><anno>Status</anno></c>
+	  as slogan. Then the runtime system exits with status code <c>1</c>.
 	  </item>
 	  <tag><c>abort</c></tag>
 	  <item>
 	    The runtime system aborts producing a core dump, if that is
-	    enabled in the operating system.
+	    enabled in the OS.
 	  </item>
 	</taglist>
-	<p>Note that on many platforms, only the status codes 0-255 are
-	  supported by the operating system.
-	</p>
-	<p>For integer <c><anno>Status</anno></c> the Erlang runtime system closes all ports
-	  and allows async threads to finish their operations before exiting.
-	  To exit without such flushing use
-	  <c><anno>Option</anno></c> as <c>{flush,false}</c>.
-	</p>
-	<p>For statuses <c>string()</c> and <c>abort</c> the <c>flush</c>
-	  option is ignored and flushing is <em>not</em> done.
-	</p>
+        <note><p>On many platforms, the OS supports only status
+        codes 0-255.</p></note>
+	<p>For integer <c><anno>Status</anno></c>, the Erlang runtime system
+          closes all ports and allows async threads to finish their
+          operations before exiting. To exit without such flushing, use
+	  <c><anno>Option</anno></c> as <c>{flush,false}</c>.</p>
+	<p>For statuses <c>string()</c> and <c>abort</c>, option
+          <c>flush</c> is ignored and flushing is <em>not</em> done.</p>
       </desc>
     </func>
+
     <func>
       <name name="hash" arity="2"/>
-      <fsummary>Hash function (deprecated)</fsummary>
+      <fsummary>Hash function (deprecated).</fsummary>
       <desc>
         <p>Returns a hash value for <c><anno>Term</anno></c> within the range
-          <c>1..<anno>Range</anno></c>. The allowed range is 1..2^27-1.</p>
+          <c>1..<anno>Range</anno></c>. The maximum range is 1..2^27-1.</p>
         <warning>
-          <p>This BIF is deprecated as the hash value may differ on
-            different architectures. Also the hash values for integer
-            terms larger than 2^27 as well as large binaries are very
+          <p>This BIF is deprecated, as the hash value can differ on
+            different architectures. The hash values for integer
+            terms higher than 2^27 and large binaries are
             poor. The BIF is retained for backward compatibility
-            reasons (it may have been used to hash records into a file),
-            but all new code should use one of the BIFs
+            reasons (it can have been used to hash records into a file),
+            but all new code is to use one of the BIFs
             <c>erlang:phash/2</c> or <c>erlang:phash2/1,2</c> instead.</p>
         </warning>
       </desc>
     </func>
+
     <func>
       <name name="hd" arity="1"/>
-      <fsummary>Head of a list</fsummary>
+      <fsummary>Head of a list.</fsummary>
       <desc>
-        <p>Returns the head of <c><anno>List</anno></c>, that is, the first element.</p>
+        <p>Returns the head of <c><anno>List</anno></c>, that is,
+          the first element, for example:</p>
         <pre>
 > <input>hd([1,2,3,4,5]).</input>
 1</pre>
         <p>Allowed in guard tests.</p>
-        <p>Failure: <c>badarg</c> if <c><anno>List</anno></c> is the empty list [].</p>
+        <p>Failure: <c>badarg</c> if <c><anno>List</anno></c> is the empty
+          list <c>[]</c>.</p>
       </desc>
     </func>
+
     <func>
       <name name="hibernate" arity="3"/>
-      <fsummary>Hibernate a process until a message is sent to it</fsummary>
+      <fsummary>Hibernates a process until a message is sent to it.</fsummary>
       <desc>
         <p>Puts the calling process into a wait state where its memory
-          allocation has been reduced as much as possible, which is
+          allocation has been reduced as much as possible. This is
           useful if the process does not expect to receive any messages
-          in the near future.</p>
-        <p>The process will be awaken when a message is sent to it, and
-          control will resume in <c><anno>Module</anno>:<anno>Function</anno></c> with
-          the arguments given by <c><anno>Args</anno></c> with the call stack
-          emptied, meaning that the process will terminate when that
-          function returns. Thus <c>erlang:hibernate/3</c> will never
-          return to its caller.</p>
+          soon.</p>
+        <p>The process is awaken when a message is sent to it, and control
+          resumes in <c><anno>Module</anno>:<anno>Function</anno></c> with
+          the arguments given by <c><anno>Args</anno></c> with the call
+          stack emptied, meaning that the process terminates when that
+          function returns. Thus <c>erlang:hibernate/3</c> never
+          returns to its caller.</p>
         <p>If the process has any message in its message queue,
-          the process will be awaken immediately in the same way as
-          described above.</p>
+          the process is awakened immediately in the same way as
+          described earlier.</p>
         <p>In more technical terms, what <c>erlang:hibernate/3</c> does
-          is the following. It discards the call stack for the process.
-          Then it garbage collects the process. After the garbage
-          collection, all live data is in one continuous heap. The heap
+          is the following. It discards the call stack for the process,
+          and then garbage collects the process. After this,
+          all live data is in one continuous heap. The heap
           is then shrunken to the exact same size as the live data
-          which it holds (even if that size is less than the minimum
+          that it holds (even if that size is less than the minimum
           heap size for the process).</p>
         <p>If the size of the live data in the process is less than
           the minimum heap size, the first garbage collection occurring
-          after the process has been awaken will ensure that the heap
+          after the process is awakened ensures that the heap
           size is changed to a size not smaller than the minimum heap
           size.</p>
-        <p>Note that emptying the call stack means that any surrounding
-          <c>catch</c> is removed and has to be re-inserted after
+        <p>Notice that emptying the call stack means that any surrounding
+          <c>catch</c> is removed and must be reinserted after
           hibernation. One effect of this is that processes started
           using <c>proc_lib</c> (also indirectly, such as
-          <c>gen_server</c> processes), should use
+          <c>gen_server</c> processes), are to use
           <seealso marker="stdlib:proc_lib#hibernate/3">proc_lib:hibernate/3</seealso>
-          instead to ensure that the exception handler continues to work
+          instead, to ensure that the exception handler continues to work
           when the process wakes up.</p>
       </desc>
     </func>
 
     <func>
       <name name="insert_element" arity="3"/>
-      <fsummary>Insert an element at index in a tuple</fsummary>
+      <fsummary>Inserts an element at index in a tuple.</fsummary>
       <type_desc variable="Index">1..tuple_size(<anno>Tuple1</anno>) + 1</type_desc>
       <desc>
-		  <p>
-			  Returns a new tuple with element <c><anno>Term</anno></c> insert at position
-			  <c><anno>Index</anno></c> in tuple <c><anno>Tuple1</anno></c>.
-			  All elements from position <c><anno>Index</anno></c> and upwards are subsequently
-			  pushed one step higher in the new tuple <c><anno>Tuple2</anno></c>.
-		  </p>
+        <p>Returns a new tuple with element <c><anno>Term</anno></c>
+        inserted at position
+        <c><anno>Index</anno></c> in tuple <c><anno>Tuple1</anno></c>.
+        All elements from position <c><anno>Index</anno></c> and upwards are
+        pushed one step higher in the new tuple <c><anno>Tuple2</anno></c>.</p>
+        <p>Example:</p>
         <pre>
 > <input>erlang:insert_element(2, {one, two, three}, new).</input>
 {one,new,two,three}</pre>
       </desc>
     </func>
+
     <func>
       <name name="integer_to_binary" arity="1"/>
-      <fsummary>Text representation of an integer</fsummary>
+      <fsummary>Text representation of an integer.</fsummary>
       <desc>
-        <p>Returns a binary which corresponds to the text
-          representation of <c><anno>Integer</anno></c>.</p>
+        <p>Returns a binary corresponding to the text
+          representation of <c><anno>Integer</anno></c>, for example:</p>
         <pre>
 > <input>integer_to_binary(77).</input>
 &lt;&lt;"77">></pre>
       </desc>
     </func>
+
     <func>
       <name name="integer_to_binary" arity="2"/>
-      <fsummary>Text representation of an integer</fsummary>
+      <fsummary>Text representation of an integer.</fsummary>
       <desc>
-        <p>Returns a binary which corresponds to the text
-          representation of <c><anno>Integer</anno></c> in base <c><anno>Base</anno></c>.</p>
+        <p>Returns a binary corresponding to the text
+        representation of <c><anno>Integer</anno></c> in base
+        <c><anno>Base</anno></c>, for example:</p>
         <pre>
 > <input>integer_to_binary(1023, 16).</input>
 &lt;&lt;"3FF">></pre>
       </desc>
     </func>
+
     <func>
       <name name="integer_to_list" arity="1"/>
-      <fsummary>Text representation of an integer</fsummary>
+      <fsummary>Text representation of an integer.</fsummary>
       <desc>
-        <p>Returns a string which corresponds to the text
-          representation of <c><anno>Integer</anno></c>.</p>
+        <p>Returns a string corresponding to the text
+          representation of <c><anno>Integer</anno></c>, for example:</p>
         <pre>
 > <input>integer_to_list(77).</input>
 "77"</pre>
       </desc>
     </func>
+
     <func>
       <name name="integer_to_list" arity="2"/>
-      <fsummary>Text representation of an integer</fsummary>
+      <fsummary>Text representation of an integer.</fsummary>
       <desc>
-        <p>Returns a string which corresponds to the text
-          representation of <c><anno>Integer</anno></c> in base <c><anno>Base</anno></c>.</p>
+        <p>Returns a string corresponding to the text
+          representation of <c><anno>Integer</anno></c> in base
+          <c><anno>Base</anno></c>, for example:</p>
         <pre>
 > <input>integer_to_list(1023, 16).</input>
 "3FF"</pre>
       </desc>
     </func>
+
     <func>
       <name name="iolist_to_binary" arity="1"/>
-      <fsummary>Convert an iolist to a binary</fsummary>
+      <fsummary>Converts an iolist to a binary.</fsummary>
       <desc>
-        <p>Returns a binary which is made from the integers and
-          binaries in <c><anno>IoListOrBinary</anno></c>.</p>
+        <p>Returns a binary that is made from the integers and
+          binaries in <c><anno>IoListOrBinary</anno></c>, for example:</p>
         <pre>
 > <input>Bin1 = &lt;&lt;1,2,3&gt;&gt;.</input>
 &lt;&lt;1,2,3&gt;&gt;
@@ -1847,278 +1967,311 @@ os_prompt% </pre>
 &lt;&lt;1,2,3,1,2,3,4,5,4,6&gt;&gt;</pre>
       </desc>
     </func>
+
     <func>
       <name name="iolist_size" arity="1"/>
-      <fsummary>Size of an iolist</fsummary>
+      <fsummary>Size of an iolist.</fsummary>
       <desc>
-        <p>Returns an integer which is the size in bytes
-          of the binary that would be the result of 
-          <c>iolist_to_binary(<anno>Item</anno>)</c>.</p>
+        <p>Returns an integer that is the size in bytes
+          of the binary that would be the result of
+          <c>iolist_to_binary(<anno>Item</anno>)</c>, for example:</p>
         <pre>
 > <input>iolist_size([1,2|&lt;&lt;3,4>>]).</input>
 4</pre>
       </desc>
     </func>
+
     <func>
       <name name="is_alive" arity="0"/>
-      <fsummary>Check whether the local node is alive</fsummary>
+      <fsummary>Checks whether the local node is alive.</fsummary>
       <desc>
-        <p>Returns <c>true</c> if the local node is alive; that is, if
-          the node can be part of a distributed system. Otherwise, it
-          returns <c>false</c>.</p>
+        <p>Returns <c>true</c> if the local node is alive (that is, if
+          the node can be part of a distributed system), otherwise
+          <c>false</c>.</p>
       </desc>
     </func>
+
     <func>
       <name name="is_atom" arity="1"/>
-      <fsummary>Check whether a term is an atom</fsummary>
+      <fsummary>Checks whether a term is an atom.</fsummary>
       <desc>
-        <p>Returns <c>true</c> if <c><anno>Term</anno></c> is an atom;
-          otherwise returns <c>false</c>.</p>
+        <p>Returns <c>true</c> if <c><anno>Term</anno></c> is an atom,
+          otherwise <c>false</c>.</p>
         <p>Allowed in guard tests.</p>
       </desc>
     </func>
+
     <func>
       <name name="is_binary" arity="1"/>
-      <fsummary>Check whether a term is a binary</fsummary>
+      <fsummary>Checks whether a term is a binary.</fsummary>
       <desc>
-        <p>Returns <c>true</c> if <c><anno>Term</anno></c> is a binary;
-          otherwise returns <c>false</c>.</p>
-
+        <p>Returns <c>true</c> if <c><anno>Term</anno></c> is a binary,
+          otherwise <c>false</c>.</p>
 	<p>A binary always contains a complete number of bytes.</p>
-
         <p>Allowed in guard tests.</p>
       </desc>
     </func>
+
     <func>
       <name name="is_bitstring" arity="1"/>
-      <fsummary>Check whether a term is a bitstring</fsummary>
+      <fsummary>Checks whether a term is a bitstring.</fsummary>
       <desc>
-        <p>Returns <c>true</c> if <c><anno>Term</anno></c> is a bitstring (including a binary);
-          otherwise returns <c>false</c>.</p>
-
+        <p>Returns <c>true</c> if <c><anno>Term</anno></c> is a
+          bitstring (including a binary), otherwise <c>false</c>.</p>
         <p>Allowed in guard tests.</p>
       </desc>
     </func>
+
     <func>
       <name name="is_boolean" arity="1"/>
-      <fsummary>Check whether a term is a boolean</fsummary>
+      <fsummary>Checks whether a term is a boolean.</fsummary>
       <desc>
-        <p>Returns <c>true</c> if <c><anno>Term</anno></c> is
-          either the atom <c>true</c> or the atom <c>false</c>
-          (i.e. a boolean); otherwise returns <c>false</c>.</p>
+        <p>Returns <c>true</c> if <c><anno>Term</anno></c> is the
+          atom <c>true</c> or the atom <c>false</c> (that is, a boolean).
+          Otherwise returns <c>false</c>.</p>
         <p>Allowed in guard tests.</p>
       </desc>
     </func>
+
     <func>
       <name name="is_builtin" arity="3"/>
-      <fsummary>Check if a function is a BIF implemented in C</fsummary>
+      <fsummary>Checks if a function is a BIF implemented in C.</fsummary>
       <desc>
-        <p>Returns <c>true</c> if <c><anno>Module</anno>:<anno>Function</anno>/<anno>Arity</anno></c> is
-          a BIF implemented in C; otherwise returns <c>false</c>.
-          This BIF is useful for builders of cross reference tools.</p>
+        <p>This BIF is useful for builders of cross-reference tools.</p>
+        <p>Returns <c>true</c> if
+          <c><anno>Module</anno>:<anno>Function</anno>/<anno>Arity</anno></c>
+          is a BIF implemented in C, otherwise  <c>false</c>.</p>
       </desc>
     </func>
+
     <func>
       <name name="is_float" arity="1"/>
-      <fsummary>Check whether a term is a float</fsummary>
+      <fsummary>Checks whether a term is a float.</fsummary>
       <desc>
         <p>Returns <c>true</c> if <c><anno>Term</anno></c> is a floating point
-          number; otherwise returns <c>false</c>.</p>
+          number, otherwise <c>false</c>.</p>
         <p>Allowed in guard tests.</p>
       </desc>
     </func>
+
     <func>
       <name name="is_function" arity="1"/>
-      <fsummary>Check whether a term is a fun</fsummary>
+      <fsummary>Checks whether a term is a fun.</fsummary>
       <desc>
-        <p>Returns <c>true</c> if <c><anno>Term</anno></c> is a fun; otherwise
-          returns <c>false</c>.</p>
+        <p>Returns <c>true</c> if <c><anno>Term</anno></c> is a fun, otherwise
+          <c>false</c>.</p>
         <p>Allowed in guard tests.</p>
       </desc>
     </func>
+
     <func>
       <name name="is_function" arity="2"/>
-      <fsummary>Check whether a term is a fun with a given arity</fsummary>
+      <fsummary>Checks whether a term is a fun with a given arity.</fsummary>
       <desc>
         <p>Returns <c>true</c> if <c><anno>Term</anno></c> is a fun that can be
-          applied with <c><anno>Arity</anno></c> number of arguments; otherwise
-          returns <c>false</c>.</p>
+          applied with <c><anno>Arity</anno></c> number of arguments, otherwise
+          <c>false</c>.</p>
         <p>Allowed in guard tests.</p>
       </desc>
     </func>
+
     <func>
       <name name="is_integer" arity="1"/>
-      <fsummary>Check whether a term is an integer</fsummary>
+      <fsummary>Checks whether a term is an integer.</fsummary>
       <desc>
-        <p>Returns <c>true</c> if <c><anno>Term</anno></c> is an integer;
-          otherwise returns <c>false</c>.</p>
+        <p>Returns <c>true</c> if <c><anno>Term</anno></c> is an integer,
+          otherwise <c>false</c>.</p>
         <p>Allowed in guard tests.</p>
       </desc>
     </func>
+
     <func>
       <name name="is_list" arity="1"/>
-      <fsummary>Check whether a term is a list</fsummary>
+      <fsummary>Checks whether a term is a list.</fsummary>
       <desc>
         <p>Returns <c>true</c> if <c><anno>Term</anno></c> is a list with
-          zero or more elements; otherwise returns <c>false</c>.</p>
+          zero or more elements, otherwise <c>false</c>.</p>
         <p>Allowed in guard tests.</p>
       </desc>
     </func>
+
     <func>
       <name name="is_map" arity="1"/>
-      <fsummary>Check whether a term is a map</fsummary>
+      <fsummary>Checks whether a term is a map.</fsummary>
       <desc>
-        <p>Returns <c>true</c> if <c><anno>Term</anno></c> is a map;
-          otherwise returns <c>false</c>.</p>
+        <p>Returns <c>true</c> if <c><anno>Term</anno></c> is a map,
+          otherwise <c>false</c>.</p>
         <p>Allowed in guard tests.</p>
       </desc>
     </func>
+
     <func>
       <name name="is_number" arity="1"/>
-      <fsummary>Check whether a term is a number</fsummary>
+      <fsummary>Checks whether a term is a number.</fsummary>
       <desc>
-        <p>Returns <c>true</c> if <c><anno>Term</anno></c> is either an integer or a
-          floating point number; otherwise returns <c>false</c>.</p>
+        <p>Returns <c>true</c> if <c><anno>Term</anno></c> is an integer or a
+          floating point number. Otherwise returns <c>false</c>.</p>
         <p>Allowed in guard tests.</p>
       </desc>
     </func>
+
     <func>
       <name name="is_pid" arity="1"/>
-      <fsummary>Check whether a term is a pid</fsummary>
+      <fsummary>Checks whether a term is a process identifier.</fsummary>
       <desc>
-        <p>Returns <c>true</c> if <c><anno>Term</anno></c> is a pid (process
-          identifier); otherwise returns <c>false</c>.</p>
+        <p>Returns <c>true</c> if <c><anno>Term</anno></c> is a process
+          identifier, otherwise <c>false</c>.</p>
         <p>Allowed in guard tests.</p>
       </desc>
     </func>
+
     <func>
       <name name="is_port" arity="1"/>
-      <fsummary>Check whether a term is a port</fsummary>
+      <fsummary>Checks whether a term is a port.</fsummary>
       <desc>
-        <p>Returns <c>true</c> if <c><anno>Term</anno></c> is a port identifier;
-          otherwise returns <c>false</c>.</p>
+        <p>Returns <c>true</c> if <c><anno>Term</anno></c> is a port identifier,
+          otherwise <c>false</c>.</p>
         <p>Allowed in guard tests.</p>
       </desc>
     </func>
+
     <func>
       <name name="is_process_alive" arity="1"/>
-      <fsummary>Check whether a process is alive</fsummary>
+      <fsummary>Checks whether a process is alive.</fsummary>
       <desc>
-        <p>
-	  <c><anno>Pid</anno></c> must refer to a process at the local node.
-          Returns <c>true</c> if the process exists and is alive, that
-          is, is not exiting and has not exited. Otherwise, returns
+        <p><c><anno>Pid</anno></c> must refer to a process at the local node.</p>
+        <p>Returns <c>true</c> if the process exists and is alive, that
+          is, is not exiting and has not exited. Otherwise returns
 	  <c>false</c>.
 	</p>
       </desc>
     </func>
+
     <func>
       <name name="is_record" arity="2"/>
-      <fsummary>Check whether a term appears to be a record</fsummary>
+      <fsummary>Checks whether a term appears to be a record.</fsummary>
       <desc>
-        <p>Returns <c>true</c> if <c><anno>Term</anno></c> is a tuple and its first
-          element is <c><anno>RecordTag</anno></c>. Otherwise, returns <c>false</c>.</p>
+        <p>Returns <c>true</c> if <c><anno>Term</anno></c> is a tuple and its
+          first element is <c><anno>RecordTag</anno></c>.
+          Otherwise returns <c>false</c>.</p>
         <note>
           <p>Normally the compiler treats calls to <c>is_record/2</c>
-            specially. It emits code to verify that <c><anno>Term</anno></c> is a
-            tuple, that its first element is <c><anno>RecordTag</anno></c>, and that
-            the size is correct. However, if the <c><anno>RecordTag</anno></c> is
-            not a literal atom, the <c>is_record/2</c> BIF will be
-            called instead and the size of the tuple will not be
-            verified.</p>
+            specially. It emits code to verify that <c><anno>Term</anno></c>
+            is a tuple, that its first element is
+            <c><anno>RecordTag</anno></c>, and that the
+            size is correct. However, if <c><anno>RecordTag</anno></c> is
+            not a literal atom, the BIF <c>is_record/2</c> is called
+            instead and the size of the tuple is not verified.</p>
         </note>
-        <p>Allowed in guard tests, if <c><anno>RecordTag</anno></c> is a literal
-          atom.</p>
+        <p>Allowed in guard tests, if <c><anno>RecordTag</anno></c> is
+          a literal atom.</p>
       </desc>
     </func>
+
     <func>
       <name name="is_record" arity="3"/>
-      <fsummary>Check whether a term appears to be a record</fsummary>
-      <desc>
-        <p><c><anno>RecordTag</anno></c> must be an atom. Returns <c>true</c> if
-          <c><anno>Term</anno></c> is a tuple, its first element is <c><anno>RecordTag</anno></c>,
-          and its size is <c><anno>Size</anno></c>. Otherwise, returns <c>false</c>.</p>
-        <p>Allowed in guard tests, provided that <c><anno>RecordTag</anno></c> is
+      <fsummary>Checks whether a term appears to be a record.</fsummary>
+      <desc>
+        <p><c><anno>RecordTag</anno></c> must be an atom.</p>
+        <p>Returns <c>true</c> if
+          <c><anno>Term</anno></c> is a tuple,
+          its first element is <c><anno>RecordTag</anno></c>,
+          and its size is <c><anno>Size</anno></c>.
+          Otherwise returns <c>false</c>.</p>
+        <p>Allowed in guard tests if <c><anno>RecordTag</anno></c> is
           a literal atom and <c>Size</c> is a literal integer.</p>
         <note>
-          <p>This BIF is documented for completeness. In most cases
-            <c>is_record/2</c> should be used.</p>
+          <p>This BIF is documented for completeness. Usually
+            <c>is_record/2</c> is to be used.</p>
         </note>
       </desc>
     </func>
+
     <func>
       <name name="is_reference" arity="1"/>
-      <fsummary>Check whether a term is a reference</fsummary>
+      <fsummary>Checks whether a term is a reference.</fsummary>
       <desc>
-        <p>Returns <c>true</c> if <c><anno>Term</anno></c> is a reference;
-          otherwise returns <c>false</c>.</p>
+        <p>Returns <c>true</c> if <c><anno>Term</anno></c> is a reference,
+          otherwise <c>false</c>.</p>
         <p>Allowed in guard tests.</p>
       </desc>
     </func>
+
     <func>
       <name name="is_tuple" arity="1"/>
-      <fsummary>Check whether a term is a tuple</fsummary>
+      <fsummary>Checks whether a term is a tuple.</fsummary>
       <desc>
-        <p>Returns <c>true</c> if <c><anno>Term</anno></c> is a tuple;
-          otherwise returns <c>false</c>.</p>
+        <p>Returns <c>true</c> if <c><anno>Term</anno></c> is a tuple,
+          otherwise <c>false</c>.</p>
         <p>Allowed in guard tests.</p>
       </desc>
     </func>
+
     <func>
       <name name="length" arity="1"/>
-      <fsummary>Length of a list</fsummary>
+      <fsummary>Length of a list.</fsummary>
       <desc>
-        <p>Returns the length of <c><anno>List</anno></c>.</p>
+        <p>Returns the length of <c><anno>List</anno></c>, for example:</p>
         <pre>
 > <input>length([1,2,3,4,5,6,7,8,9]).</input>
 9</pre>
         <p>Allowed in guard tests.</p>
       </desc>
     </func>
+
     <func>
       <name name="link" arity="1"/>
-      <fsummary>Create a link to another process (or port)</fsummary>
+      <fsummary>Creates a link to another process (or port).</fsummary>
       <desc>
         <p>Creates a link between the calling process and another
-          process (or port) <c><anno>PidOrPort</anno></c>, if there is not such a link
+          process (or port) <c><anno>PidOrPort</anno></c>, if there is
+          not such a link
           already. If a process attempts to create a link to itself,
           nothing is done. Returns <c>true</c>.</p>
-        <p>If <c><anno>PidOrPort</anno></c> does not exist, the behavior of the BIF depends
-          on if the calling process is trapping exits or not (see
+        <p>If <c><anno>PidOrPort</anno></c> does not exist, the behavior
+          of the BIF
+          depends on if the calling process is trapping exits or not (see
           <seealso marker="#process_flag/2">process_flag/2</seealso>):</p>
         <list type="bulleted">
           <item>If the calling process is not trapping exits, and
-           checking <c><anno>PidOrPort</anno></c> is cheap -- that is, if <c><anno>PidOrPort</anno></c> is
-           local -- <c>link/1</c> fails with reason <c>noproc</c>.</item>
+           checking <c><anno>PidOrPort</anno></c> is cheap
+          (that is, if <c><anno>PidOrPort</anno></c>
+           is local), <c>link/1</c> fails with reason <c>noproc</c>.</item>
           <item>Otherwise, if the calling process is trapping exits,
-           and/or <c><anno>PidOrPort</anno></c> is remote, <c>link/1</c> returns
-          <c>true</c>, but an exit signal with reason <c>noproc</c>
+           and/or <c><anno>PidOrPort</anno></c> is remote, <c>link/1</c>
+           returns <c>true</c>, but an exit signal with reason <c>noproc</c>
            is sent to the calling process.</item>
         </list>
       </desc>
     </func>
+
     <func>
       <name name="list_to_atom" arity="1"/>
-      <fsummary>Convert from text representation to an atom</fsummary>
-      <desc>
-        <p>Returns the atom whose text representation is <c><anno>String</anno></c>.</p>
-	<p><c><anno>String</anno></c> may only contain ISO-latin-1
-	characters (i.e. numbers below 256) as the current
-	implementation does not allow unicode characters >= 256 in
-	atoms. For more information on Unicode support in atoms
-	see <seealso marker="erl_ext_dist#utf8_atoms">note on UTF-8 encoded atoms</seealso>
-	in the chapter about the external term format in the ERTS User's Guide.</p>
+      <fsummary>Converts from text representation to an atom.</fsummary>
+      <desc>
+        <p>Returns the atom whose text representation is
+          <c><anno>String</anno></c>.</p>
+	<p><c><anno>String</anno></c> can only contain ISO-latin-1
+          characters (that is,
+          numbers less than 256) as the implementation does not
+          allow unicode characters equal to or above 256 in atoms.
+	  For more information on Unicode support in atoms, see
+	  <seealso marker="erl_ext_dist#utf8_atoms">note on UTF-8
+          encoded atoms</seealso>
+	  in Section "External Term Format" in the User's Guide.</p>
+        <p>Example:</p>
         <pre>
 > <input>list_to_atom("Erlang").</input>
 'Erlang'</pre>
       </desc>
     </func>
+
     <func>
       <name name="list_to_binary" arity="1"/>
-      <fsummary>Convert a list to a binary</fsummary>
+      <fsummary>Converts a list to a binary.</fsummary>
       <desc>
-        <p>Returns a binary which is made from the integers and
-          binaries in <c><anno>IoList</anno></c>.</p>
+        <p>Returns a binary that is made from the integers and
+          binaries in <c><anno>IoList</anno></c>, for example:</p>
         <pre>
 > <input>Bin1 = &lt;&lt;1,2,3&gt;&gt;.</input>
 &lt;&lt;1,2,3&gt;&gt;
@@ -2130,40 +2283,46 @@ os_prompt% </pre>
 &lt;&lt;1,2,3,1,2,3,4,5,4,6&gt;&gt;</pre>
       </desc>
     </func>
+
     <func>
       <name name="list_to_bitstring" arity="1"/>
+      <fsummary>Converts a list to a bitstring.</fsummary>
       <type name="bitstring_list"/>
-      <fsummary>Convert a list to a bitstring</fsummary>
       <desc>
-        <p>Returns a bitstring which is made from the integers and
-          bitstrings in <c><anno>BitstringList</anno></c>. (The last tail in <c><anno>BitstringList</anno></c>
-	  is allowed to be a bitstring.)</p>
+        <p>Returns a bitstring that is made from the integers and
+          bitstrings in <c><anno>BitstringList</anno></c>. (The last tail in
+          <c><anno>BitstringList</anno></c> is allowed to be a bitstring.)</p>
+        <p>Example:</p>
         <pre>
 > <input>Bin1 = &lt;&lt;1,2,3&gt;&gt;.</input>
 &lt;&lt;1,2,3&gt;&gt;
 > <input>Bin2 = &lt;&lt;4,5&gt;&gt;.</input>
 &lt;&lt;4,5&gt;&gt;
-> <input>Bin3 = &lt;&lt;6,7:4,&gt;&gt;.</input>
-&lt;&lt;6&gt;&gt;
+> <input>Bin3 = &lt;&lt;6,7:4&gt;&gt;.</input>
+&lt;&lt;6,7:4&gt;&gt;
 > <input>list_to_bitstring([Bin1,1,[2,3,Bin2],4|Bin3]).</input>
-&lt;&lt;1,2,3,1,2,3,4,5,4,6,7:46&gt;&gt;</pre>
+&lt;&lt;1,2,3,1,2,3,4,5,4,6,7:4&gt;&gt;</pre>
       </desc>
     </func>
+
     <func>
       <name name="list_to_existing_atom" arity="1"/>
-      <fsummary>Convert from text representation to an atom</fsummary>
+      <fsummary>Converts from text representation to an atom.</fsummary>
       <desc>
-        <p>Returns the atom whose text representation is <c><anno>String</anno></c>,
+        <p>Returns the atom whose text representation is
+          <c><anno>String</anno></c>,
           but only if there already exists such atom.</p>
         <p>Failure: <c>badarg</c> if there does not already exist an atom
           whose text representation is <c><anno>String</anno></c>.</p>
       </desc>
     </func>
+
     <func>
       <name name="list_to_float" arity="1"/>
-      <fsummary>Convert from text representation to a float</fsummary>
+      <fsummary>Converts from text representation to a float.</fsummary>
       <desc>
-        <p>Returns the float whose text representation is <c><anno>String</anno></c>.</p>
+        <p>Returns the float whose text representation is
+          <c><anno>String</anno></c>, for example:</p>
         <pre>
 > <input>list_to_float("2.2017764e+0").</input>
 2.2017764</pre>
@@ -2171,12 +2330,13 @@ os_prompt% </pre>
           representation of a float.</p>
       </desc>
     </func>
+
     <func>
       <name name="list_to_integer" arity="1"/>
-      <fsummary>Convert from text representation to an integer</fsummary>
+      <fsummary>Converts from text representation to an integer.</fsummary>
       <desc>
         <p>Returns an integer whose text representation is
-          <c><anno>String</anno></c>.</p>
+          <c><anno>String</anno></c>, for example:</p>
         <pre>
 > <input>list_to_integer("123").</input>
 123</pre>
@@ -2184,12 +2344,14 @@ os_prompt% </pre>
           representation of an integer.</p>
       </desc>
     </func>
+
     <func>
       <name name="list_to_integer" arity="2"/>
-      <fsummary>Convert from text representation to an integer</fsummary>
+      <fsummary>Converts from text representation to an integer.</fsummary>
       <desc>
         <p>Returns an integer whose text representation in base
-          <c><anno>Base</anno></c> is <c><anno>String</anno></c>.</p>
+          <c><anno>Base</anno></c> is <c><anno>String</anno></c>,
+          for example:</p>
         <pre>
 > <input>list_to_integer("3FF", 16).</input>
 1023</pre>
@@ -2197,47 +2359,52 @@ os_prompt% </pre>
           representation of an integer.</p>
       </desc>
     </func>
+
     <func>
       <name name="list_to_pid" arity="1"/>
-      <fsummary>Convert from text representation to a pid</fsummary>
+      <fsummary>Converts from text representation to a pid.</fsummary>
       <desc>
-        <p>Returns a pid whose text representation is <c><anno>String</anno></c>.</p>
-        <warning>
-          <p>This BIF is intended for debugging and for use in
-            the Erlang operating system. It should not be used in
-            application programs.</p>
-        </warning>
+        <p>Returns a process identifier whose text representation is a 
+          <c><anno>String</anno></c>, for example:</p>
         <pre>
 > <input>list_to_pid("&lt;0.4.1>").</input>
 &lt;0.4.1></pre>
         <p>Failure: <c>badarg</c> if <c><anno>String</anno></c> contains a bad
-          representation of a pid.</p>
+          representation of a process identifier.</p>
+        <warning>
+          <p>This BIF is intended for debugging and is not to be used
+	  in application programs.</p>
+        </warning>
       </desc>
     </func>
+
     <func>
       <name name="list_to_tuple" arity="1"/>
-      <fsummary>Convert a list to a tuple</fsummary>
+      <fsummary>Converts a list to a tuple.</fsummary>
       <desc>
-        <p>Returns a tuple which corresponds to <c><anno>List</anno></c>. <c><anno>List</anno></c>
-          can contain any Erlang terms.</p>
+        <p>Returns a tuple corresponding to <c><anno>List</anno></c>,
+          for example</p>
         <pre>
 > <input>list_to_tuple([share, ['Ericsson_B', 163]]).</input>
 {share, ['Ericsson_B', 163]}</pre>
+        <p><c><anno>List</anno></c> can contain any Erlang terms.</p>
       </desc>
     </func>
+
     <func>
       <name name="load_module" arity="2"/>
-      <fsummary>Load object code for a module</fsummary>
+      <fsummary>Loads object code for a module.</fsummary>
       <desc>
-        <p>If <c><anno>Binary</anno></c> contains the object code for the module
-          <c><anno>Module</anno></c>, this BIF loads that object code. Also, if
-          the code for the module <c><anno>Module</anno></c> already exists, all
+        <p>If <c><anno>Binary</anno></c> contains the object code for module
+          <c><anno>Module</anno></c>, this BIF loads that object code. If
+          the code for module <c><anno>Module</anno></c> already exists, all
           export references are replaced so they point to the newly
           loaded code. The previously loaded code is kept in the system
-          as old code, as there may still be processes which are
-          executing that code. It returns either
-          <c>{module, <anno>Module</anno>}</c>, or <c>{error, <anno>Reason</anno>}</c> if loading
-          fails. <c><anno>Reason</anno></c> is one of the following:</p>
+          as old code, as there can still be processes executing
+          that code.</p>
+        <p>Returns either <c>{module, <anno>Module</anno>}</c>, or
+          <c>{error, <anno>Reason</anno>}</c> if loading fails.
+          <c><anno>Reason</anno></c> is any of the following:</p>
         <taglist>
           <tag><c>badfile</c></tag>
           <item>
@@ -2247,118 +2414,122 @@ os_prompt% </pre>
           </item>
           <tag><c>not_purged</c></tag>
           <item>
-            <p><c><anno>Binary</anno></c> contains a module which cannot be loaded
-              because old code for this module already exists.</p>
+            <p><c><anno>Binary</anno></c> contains a module that cannot be
+              loaded because old code for this module already exists.</p>
           </item>
         </taglist>
         <warning>
           <p>This BIF is intended for the code server (see
-            <seealso marker="kernel:code">code(3)</seealso>) and should not be
-            used elsewhere.</p>
+            <seealso marker="kernel:code">code(3)</seealso>)
+            and is not to be used elsewhere.</p>
         </warning>
       </desc>
     </func>
+
     <func>
       <name name="load_nif" arity="2"/>
-      <fsummary>Load NIF library</fsummary>
+      <fsummary>Loads NIF library.</fsummary>
       <desc>
         <note>
-          <p>In releases older than OTP R14B, NIFs were an
-          experimental feature. Versions of OTP older than R14B might
+          <p>Before OTP R14B, NIFs were an
+          experimental feature. Versions before OTP R14B can
           have different and possibly incompatible NIF semantics and
-          interfaces. For example, in R13B03 the return value on
-          failure was
-          <c>{error,Reason,Text}</c>.</p>
+          interfaces. For example, in OTP R13B03 the return value on
+          failure was <c>{error,Reason,Text}</c>.</p>
         </note>
         <p>Loads and links a dynamic library containing native
-        implemented functions (NIFs) for a module. <c><anno>Path</anno></c> is a
-        file path to the sharable object/dynamic library file minus
-        the OS-dependent file extension (.so for Unix and .dll for
-        Windows). See <seealso marker="erl_nif">erl_nif</seealso>
-        on how to implement a NIF library.</p> 
-        <p><c><anno>LoadInfo</anno></c> can be any term. It will be passed on to
+        implemented functions (NIFs) for a module. <c><anno>Path</anno></c>
+        is a file path to the shareable object/dynamic library file minus
+        the OS-dependent file extension (<c>.so</c> for Unix and
+        <c>.dll</c> for Windows. For information on how to
+        implement a NIF library, see
+        <seealso marker="erl_nif">erl_nif</seealso>.</p> 
+        <p><c><anno>LoadInfo</anno></c> can be any term. It is passed on to
         the library as part of the initialization. A good practice is
         to include a module version number to support future code
         upgrade scenarios.</p>
         <p>The call to <c>load_nif/2</c> must be made
         <em>directly</em> from the Erlang code of the module that the
-        NIF library belongs to.</p>
-        <p>It returns either <c>ok</c>, or <c>{error,{<anno>Reason</anno>,Text}}</c>
-        if loading fails. <c><anno>Reason</anno></c> is one of the atoms below,
-        while <c><anno>Text</anno></c> is a human readable string that may give
-        some more information about the failure.</p>
+        NIF library belongs to. It returns either <c>ok</c>, or
+        <c>{error,{<anno>Reason</anno>,Text}}</c> if loading fails.
+        <c><anno>Reason</anno></c> is one of the following atoms
+        while <c><anno>Text</anno></c> is a human readable string that
+        can give more information about the failure:</p>
         <taglist>
           <tag><c>load_failed</c></tag>
-          <item>
-            <p>The OS failed to load the NIF library.</p>
+          <item>The OS failed to load the NIF library.
           </item>
           <tag><c>bad_lib</c></tag>
-          <item>
-            <p>The library did not fulfil the requirements as a NIF
-            library of the calling module.</p> 
+          <item>The library did not fulfill the requirements as a NIF
+            library of the calling module.
           </item>
           <tag><c>load | reload | upgrade</c></tag>
-          <item>
-            <p>The corresponding library callback was not successful.</p>
+          <item>The corresponding library callback was unsuccessful.
           </item>
           <tag><c>old_code</c></tag>
-          <item>
-            <p>The call to <c>load_nif/2</c> was made from the old
-            code of a module that has been upgraded. This is not
-            allowed.</p>
+          <item>The call to <c>load_nif/2</c> was made from the old
+            code of a module that has been upgraded; this is not
+            allowed.
           </item>
         </taglist>
       </desc>
     </func>
+
     <func>
       <name name="loaded" arity="0"/>
-      <fsummary>List of all loaded modules</fsummary>
+      <fsummary>Lists all loaded modules.</fsummary>
       <desc>
-        <p>Returns a list of all loaded Erlang modules (current and/or
+        <p>Returns a list of all loaded Erlang modules (current and
           old code), including preloaded modules.</p>
         <p>See also <seealso marker="kernel:code">code(3)</seealso>.</p>
       </desc>
     </func>
+
     <func>
       <name name="localtime" arity="0"/>
-      <fsummary>Current local date and time</fsummary>
+      <fsummary>Current local date and time.</fsummary>
       <desc>
-        <p>Returns the current local date and time
-          <c>{{Year, Month, Day}, {Hour, Minute, Second}}</c>.</p>
-        <p>The time zone and daylight saving time correction depend
-          on the underlying OS.</p>
+        <p>Returns the current local date and time,
+          <c>{{Year, Month, Day}, {Hour, Minute, Second}}</c>,
+          for example:</p>
         <pre>
 > <input>erlang:localtime().</input>
 {{1996,11,6},{14,45,17}}</pre>
+        <p>The time zone and Daylight Saving Time correction depend
+          on the underlying OS.</p>
       </desc>
     </func>
+
     <func>
       <name name="localtime_to_universaltime" arity="1"/>
-      <fsummary>Convert from local to Universal Time Coordinated (UTC) date and time</fsummary>
+      <fsummary>Converts from local to Universal Time Coordinated (UTC) date and time.</fsummary>
       <desc>
         <p>Converts local date and time to Universal Time Coordinated
-          (UTC), if this is supported by the underlying OS. Otherwise,
-          no conversion is done and <c><anno>Localtime</anno></c> is returned.</p>
+          (UTC), if supported by the underlying OS. Otherwise
+          no conversion is done and <c><anno>Localtime</anno></c>
+          is returned.</p>
+        <p>Example:</p>
         <pre>
 > <input>erlang:localtime_to_universaltime({{1996,11,6},{14,45,17}}).</input>
 {{1996,11,6},{13,45,17}}</pre>
-        <p>Failure: <c>badarg</c> if <c><anno>Localtime</anno></c> does not denote
-          a valid date and time.</p>
+        <p>Failure: <c>badarg</c> if <c><anno>Localtime</anno></c> denotes an
+          invalid date and time.</p>
       </desc>
     </func>
+
     <func>
       <name name="localtime_to_universaltime" arity="2"/>
-      <fsummary>Convert from local to Universal Time Coordinated (UTC) date and time</fsummary>
+      <fsummary>Converts from local to Universal Time Coordinated (UTC) date and time.</fsummary>
       <desc>
         <p>Converts local date and time to Universal Time Coordinated
-          (UTC) just like <c>erlang:localtime_to_universaltime/1</c>,
-          but the caller decides if daylight saving time is active or
-          not.</p>
-        <p>If <c><anno>IsDst</anno> == true</c> the <c><anno>Localtime</anno></c> is during
-          daylight saving time, if <c><anno>IsDst</anno> == false</c> it is not,
-          and if <c><anno>IsDst</anno> == undefined</c> the underlying OS may
+          (UTC) as <c>erlang:localtime_to_universaltime/1</c>,
+          but the caller decides if Daylight Saving Time is active.</p>
+        <p>If <c><anno>IsDst</anno> == true</c>, <c><anno>Localtime</anno></c> is
+          during Daylight Saving Time, if <c><anno>IsDst</anno> == false</c> it is
+          not. If <c><anno>IsDst</anno> == undefined</c>, the underlying OS can
           guess, which is the same as calling
           <c>erlang:localtime_to_universaltime(<anno>Localtime</anno>)</c>.</p>
+        <p>Examples:</p>
         <pre>
 > <input>erlang:localtime_to_universaltime({{1996,11,6},{14,45,17}}, true).</input>
 {{1996,11,6},{12,45,17}}
@@ -2366,15 +2537,16 @@ os_prompt% </pre>
 {{1996,11,6},{13,45,17}}
 > <input>erlang:localtime_to_universaltime({{1996,11,6},{14,45,17}}, undefined).</input>
 {{1996,11,6},{13,45,17}}</pre>
-        <p>Failure: <c>badarg</c> if <c><anno>Localtime</anno></c> does not denote
-          a valid date and time.</p>
+        <p>Failure: <c>badarg</c> if <c><anno>Localtime</anno></c> denotes an
+          invalid date and time.</p>
       </desc>
     </func>
+
     <func>
       <name name="make_ref" arity="0"/>
-      <fsummary>Return a unique reference</fsummary>
+      <fsummary>Returns a unique reference.</fsummary>
       <desc>
-	<p>Return a <seealso marker="doc/efficiency_guide:advanced#unique_references">unique
+	<p>Returns a <seealso marker="doc/efficiency_guide:advanced#unique_references">unique
 	reference</seealso>. The reference is unique among
 	connected nodes.</p>
 	<warning><p>Known issue: When a node is restarted multiple
@@ -2383,200 +2555,209 @@ os_prompt% </pre>
 	created on an older node with the same node name.</p></warning>
       </desc>
     </func>
+
     <func>
       <name name="make_tuple" arity="2"/>
-      <fsummary>Create a new tuple of a given arity</fsummary>
+      <fsummary>Creates a new tuple of a given arity.</fsummary>
       <desc>
-        <p>Returns a new tuple of the given <c><anno>Arity</anno></c>, where all
-          elements are <c><anno>InitialValue</anno></c>.</p>
+        <p>Creates a new tuple of the given <c><anno>Arity</anno></c>, where all
+          elements are <c><anno>InitialValue</anno></c>, for example:</p>
         <pre>
 > <input>erlang:make_tuple(4, []).</input>
 {[],[],[],[]}</pre>
       </desc>
     </func>
+
     <func>
       <name name="make_tuple" arity="3"/>
-      <fsummary>Create a new tuple with given arity and contents</fsummary>
-      <desc>
-        <p><c>erlang:make_tuple</c> first creates a tuple of size <c><anno>Arity</anno></c>
-	  where each element has the value <c><anno>DefaultValue</anno></c>. It then fills
-	  in values from <c><anno>InitList</anno></c>. Each list element in <c><anno>InitList</anno></c>
-	  must be a two-tuple where the first element is a position in the
-	  newly created tuple and the second element is any term. If a position
-	  occurs more than once in the list, the term corresponding to
-	  last occurrence will be used.</p>
+      <fsummary>Creates a new tuple with given arity and contents.</fsummary>
+      <desc>
+        <p>Creates a tuple of size <c><anno>Arity</anno></c>, where each element
+          has value <c><anno>DefaultValue</anno></c>, and then fills in
+	  values from <c><anno>InitList</anno></c>.
+          Each list element in <c><anno>InitList</anno></c>
+	  must be a two-tuple, where the first element is a position in the
+	  newly created tuple and the second element is any term. If a
+	  position occurs more than once in the list, the term corresponding
+	  to the last occurrence is used.</p>
+        <p>Example:</p>
         <pre>
 > <input>erlang:make_tuple(5, [], [{2,ignored},{5,zz},{2,aa}]).</input>
 {{[],aa,[],[],zz}</pre>
       </desc>
     </func>
+
     <func>
       <name name="map_size" arity="1"/>
-      <fsummary>Return the size of a map</fsummary>
+      <fsummary>Returns the size of a map.</fsummary>
       <desc>
-        <p>Returns an integer which is the number of key-value pairs in <c><anno>Map</anno></c>.</p>
+        <p>Returns an integer, which is the number of key-value pairs
+          in <c><anno>Map</anno></c>, for example:</p>
         <pre>
 > <input>map_size(#{a=>1, b=>2, c=>3}).</input>
 3</pre>
         <p>Allowed in guard tests.</p>
       </desc>
     </func>
+
     <func>
       <name name="max" arity="2"/>
-      <fsummary>Return the largest of two term</fsummary>
+      <fsummary>Returns the largest of two terms.</fsummary>
       <desc>
-        <p>Return the largest of <c><anno>Term1</anno></c> and <c><anno>Term2</anno></c>;
-	if the terms compare equal, <c><anno>Term1</anno></c> will be returned.</p>
+        <p>Returns the largest of <c><anno>Term1</anno></c> and
+          <c><anno>Term2</anno></c>.
+	  If the terms are equal, <c><anno>Term1</anno></c> is returned.</p>
       </desc>
     </func>
+
     <func>
       <name name="md5" arity="1"/>
-      <fsummary>Compute an MD5 message digest</fsummary>
+      <fsummary>Computes an MD5 message digest.</fsummary>
       <desc>
-        <p>Computes an <c>MD5</c> message digest from <c><anno>Data</anno></c>, where
-          the length of the digest is 128 bits (16 bytes). <c><anno>Data</anno></c>
+        <p>Computes an MD5 message digest from <c><anno>Data</anno></c>, where
+          the length of the digest is 128 bits (16 bytes).
+          <c><anno>Data</anno></c>
           is a binary or a list of small integers and binaries.</p>
-        <p>See The MD5 Message Digest Algorithm (RFC 1321) for more
-          information about MD5.</p>
-	<warning><p>The MD5 Message Digest Algorithm is <em>not</em> considered
-	safe for code-signing or software integrity purposes.</p></warning>
+        <p>For more information about MD5, see RFC 1321 - The
+          MD5 Message-Digest Algorithm.</p>
+	<warning><p>The MD5 Message-Digest Algorithm is <em>not</em> considered
+	safe for code-signing or software-integrity purposes.</p></warning>
       </desc>
     </func>
+
     <func>
       <name name="md5_final" arity="1"/>
-      <fsummary>Finish the update of an MD5 context and return the computed MD5 message digest</fsummary>
+      <fsummary>Finishes the update of an MD5 context and returns the computed MD5 message digest.</fsummary>
       <desc>
         <p>Finishes the update of an MD5 <c><anno>Context</anno></c> and returns
           the computed <c>MD5</c> message digest.</p>
       </desc>
     </func>
+
     <func>
       <name name="md5_init" arity="0"/>
-      <fsummary>Create an MD5 context</fsummary>
+      <fsummary>Creates an MD5 context.</fsummary>
       <desc>
         <p>Creates an MD5 context, to be used in subsequent calls to
           <c>md5_update/2</c>.</p>
       </desc>
     </func>
+
     <func>
       <name name="md5_update" arity="2"/>
-      <fsummary>Update an MD5 context with data, and return a new context</fsummary>
+      <fsummary>Updates an MD5 context with data and returns a new context.</fsummary>
       <desc>
-        <p>Updates an MD5 <c><anno>Context</anno></c> with <c><anno>Data</anno></c>, and returns
-          a <c><anno>NewContext</anno></c>.</p>
+        <p>Updates an MD5 <c><anno>Context</anno></c> with
+          <c><anno>Data</anno></c> and returns a
+          <c><anno>NewContext</anno></c>.</p>
       </desc>
     </func>
+
     <func>
       <name name="memory" arity="0"/>
+      <fsummary>Information about dynamically allocated memory.</fsummary>
       <type name="memory_type"/>
-      <fsummary>Information about dynamically allocated memory</fsummary>
-      <desc>
-        <p>Returns a list containing information about memory
-          dynamically allocated by the Erlang emulator. Each element of
-          the list is a tuple <c>{Type, Size}</c>. The first element
-          <c><anno>Type</anno></c>is an atom describing memory type. The second
-          element <c><anno>Size</anno></c>is memory size in bytes. A description of
-          each memory type follows:</p>
+      <desc>
+        <p>Returns a list with information about memory
+          dynamically allocated by the Erlang emulator. Each list
+          element is a tuple <c>{Type, Size}</c>. The first element
+          <c><anno>Type</anno></c> is an atom describing memory type. The second
+          element <c><anno>Size</anno></c> is the memory size in bytes.</p>
+        <p>The memory types are as follows:</p>
         <taglist>
           <tag><c>total</c></tag>
           <item>
-            <p>The total amount of memory currently allocated, which is
-              the same as the sum of memory size for <c>processes</c>
+            <p>The total amount of memory currently allocated. This is
+              the same as the sum of the memory size for <c>processes</c>
               and <c>system</c>.</p>
           </item>
           <tag><c>processes</c></tag>
           <item>
-            <p>The total amount of memory currently allocated by
+            <p>The total amount of memory currently allocated for
               the Erlang processes.</p>
           </item>
           <tag><c>processes_used</c></tag>
           <item>
             <p>The total amount of memory currently used by the Erlang
-              processes.</p>
-            <p>This memory is part of the memory presented as
+              processes. This is part of the memory presented as
               <c>processes</c> memory.</p>
           </item>
           <tag><c>system</c></tag>
           <item>
-            <p>The total amount of memory currently allocated by
+            <p>The total amount of memory currently allocated for
               the emulator that is not directly related to any Erlang
-              process.</p>
-            <p>Memory presented as <c>processes</c> is not included in
-              this memory.</p>
+              process. Memory presented as <c>processes</c> is not
+              included in this memory.</p>
           </item>
           <tag><c>atom</c></tag>
           <item>
-            <p>The total amount of memory currently allocated for atoms.</p>
-            <p>This memory is part of the memory presented as
+            <p>The total amount of memory currently allocated for atoms.
+              This memory is part of the memory presented as
               <c>system</c> memory.</p>
           </item>
           <tag><c>atom_used</c></tag>
           <item>
-            <p>The total amount of memory currently used for atoms.</p>
-            <p>This memory is part of the memory presented as
+            <p>The total amount of memory currently used for atoms.
+              This memory is part of the memory presented as
               <c>atom</c> memory.</p>
           </item>
           <tag><c>binary</c></tag>
           <item>
             <p>The total amount of memory currently allocated for
-              binaries.</p>
-            <p>This memory is part of the memory presented as
-              <c>system</c> memory.</p>
+              binaries. This memory is part of the memory presented
+              as <c>system</c> memory.</p>
           </item>
           <tag><c>code</c></tag>
           <item>
             <p>The total amount of memory currently allocated for
-              Erlang code.</p>
-            <p>This memory is part of the memory presented as
-              <c>system</c> memory.</p>
+              Erlang code. This memory is part of the memory presented
+              as <c>system</c> memory.</p>
           </item>
           <tag><c>ets</c></tag>
           <item>
             <p>The total amount of memory currently allocated for ets
-              tables.</p>
-            <p>This memory is part of the memory presented as
+              tables. This memory is part of the memory presented as
               <c>system</c> memory.</p>
           </item>
           <tag><c>low</c></tag>
           <item>
-            <p>Only on 64-bit halfword emulator.</p>
-            <p>The total amount of memory allocated in low memory areas
-	      that are restricted to less than 4 Gb even though
-	      the system may have more physical memory.</p>
-	    <p>May be removed in future releases of halfword emulator.</p>
+            <p>Only on 64-bit halfword emulator.
+              The total amount of memory allocated in low memory areas
+	      that are restricted to less than 4 GB, although
+	      the system can have more memory.</p>
+	    <p>Can be removed in a future release of the halfword
+              emulator.</p>
           </item>
           <tag><c>maximum</c></tag>
           <item>
             <p>The maximum total amount of memory allocated since
-              the emulator was started.</p>
-            <p>This tuple is only present when the emulator is run with
-              instrumentation.</p>
+              the emulator was started. This tuple is only present
+              when the emulator is run with instrumentation.</p>
             <p>For information on how to run the emulator with
-              instrumentation see
+              instrumentation, see
               <seealso marker="tools:instrument">instrument(3)</seealso>
               and/or <seealso marker="erts:erl">erl(1)</seealso>.</p>
           </item>
         </taglist>
         <note>
           <p>The <c>system</c> value is not complete. Some allocated
-            memory that should be part of the <c>system</c> value are
-            not.</p>
+            memory that is to be part of this value is not.</p>
           <p>When the emulator is run with instrumentation,
             the <c>system</c> value is more accurate, but memory
-            directly allocated by <c>malloc</c> (and friends) are still
+            directly allocated for <c>malloc</c> (and friends) is still
             not part of the <c>system</c> value. Direct calls to
-            <c>malloc</c> are only done from OS specific runtime
-            libraries and perhaps from user implemented Erlang drivers
+            <c>malloc</c> are only done from OS-specific runtime
+            libraries and perhaps from user-implemented Erlang drivers
             that do not use the memory allocation functions in
             the driver interface.</p>
-          <p>Since the <c>total</c> value is the sum of <c>processes</c>
-            and <c>system</c> the error in <c>system</c> will propagate
+          <p>As the <c>total</c> value is the sum of <c>processes</c>
+            and <c>system</c>, the error in <c>system</c> propagates
             to the <c>total</c> value.</p>
 	  <p>The different amounts of memory that are summed are
-	    <em>not</em> gathered atomically which also introduce
+	    <em>not</em> gathered atomically, which introduces
 	    an error in the result.</p>
         </note>
-        <p>The different values has the following relation to each
+        <p>The different values have the following relation to each
           other. Values beginning with an uppercase letter is not part
           of the result.</p>
         <code type="none">
@@ -2584,69 +2765,62 @@ os_prompt% </pre>
         processes = processes_used + ProcessesNotUsed
         system = atom + binary + code + ets + OtherSystem
         atom = atom_used + AtomNotUsed
-
         RealTotal = processes + RealSystem
         RealSystem = system + MissedSystem</code>
-        <p>More tuples in the returned list may be added in the future.</p>
+        <p>More tuples in the returned list can be added in a
+          future release.</p>
         <note>
           <p>The <c>total</c> value is supposed to be the total amount
             of memory dynamically allocated by the emulator. Shared
             libraries, the code of the emulator itself, and
-            the emulator stack(s) are not supposed to be included. That
+            the emulator stacks are not supposed to be included. That
             is, the <c>total</c> value is <em>not</em> supposed to be
-            equal to the total size of all pages mapped to the emulator.
-            Furthermore, due to fragmentation and pre-reservation of
-            memory areas, the size of the memory segments which contain
-            the dynamically allocated memory blocks can be substantially
+            equal to the total size of all pages mapped to the emulator.</p>
+          <p>Furthermore, because of fragmentation and prereservation of
+            memory areas, the size of the memory segments containing
+            the dynamically allocated memory blocks can be much
             larger than the total size of the dynamically allocated
             memory blocks.</p>
         </note>
 	<note>
-	  <p>
-	    Since erts version 5.6.4 <c>erlang:memory/0</c> requires that
+	  <p>As from <c>ERTS</c> 5.6.4, <c>erlang:memory/0</c> requires that
 	    all <seealso marker="erts:erts_alloc">erts_alloc(3)</seealso>
-	    allocators are enabled (default behaviour).
-	  </p>
+	    allocators are enabled (default behavior).</p>
 	</note>
-        <p>Failure:</p>
-        <taglist>
-          <tag><c>notsup</c></tag>
-          <item>
-	    If an <seealso marker="erts:erts_alloc">erts_alloc(3)</seealso>
-	    allocator has been disabled.
-	  </item>
-	</taglist>
+        <p>Failure: <c>notsup</c> if an
+          <seealso marker="erts:erts_alloc">erts_alloc(3)</seealso>
+	  allocator has been disabled.</p>
       </desc>
     </func>
+
     <func>
       <name name="memory" arity="1" clause_i="1"/>
       <name name="memory" arity="1" clause_i="2"/>
+      <fsummary>Information about dynamically allocated memory.</fsummary>
       <type name="memory_type"/>
-      <fsummary>Information about dynamically allocated memory</fsummary>
       <desc>
         <p>Returns the memory size in bytes allocated for memory of
           type <c><anno>Type</anno></c>. The argument can also be given as a list
           of <c>memory_type()</c> atoms, in which case a corresponding list of
           <c>{memory_type(), Size :: integer >= 0}</c> tuples is returned.</p>
 	<note>
-	  <p>
-	    Since erts version 5.6.4 <c>erlang:memory/1</c> requires that
+	  <p>As from <c>ERTS</c> version 5.6.4,
+            <c>erlang:memory/1</c> requires that
 	    all <seealso marker="erts:erts_alloc">erts_alloc(3)</seealso>
-	    allocators are enabled (default behaviour).
-	  </p>
+	    allocators are enabled (default behavior).</p>
 	</note>
         <p>Failures:</p>
         <taglist>
           <tag><c>badarg</c></tag>
           <item>
-	    If <c><anno>Type</anno></c> is not one of the memory types listed in the
-	    documentation of
+	    If <c><anno>Type</anno></c> is not one of the memory types
+            listed in the description of
 	    <seealso marker="#memory/0">erlang:memory/0</seealso>.
 	  </item>
           <tag><c>badarg</c></tag>
           <item>
-	    If <c>maximum</c> is passed as <c><anno>Type</anno></c> and the emulator
-	    is not run in instrumented mode.
+	    If <c>maximum</c> is passed as <c><anno>Type</anno></c> and
+            the emulator is not run in instrumented mode.
 	  </item>
           <tag><c>notsup</c></tag>
           <item>
@@ -2658,35 +2832,39 @@ os_prompt% </pre>
           <seealso marker="#memory/0">erlang:memory/0</seealso>.</p>
       </desc>
     </func>
+
     <func>
       <name name="min" arity="2"/>
-      <fsummary>Return the smallest of two term</fsummary>
+      <fsummary>Returns the smallest of two terms.</fsummary>
       <desc>
-        <p>Return the smallest of <c><anno>Term1</anno></c> and <c><anno>Term2</anno></c>;
-	if the terms compare equal, <c><anno>Term1</anno></c> will be returned.</p>
+        <p>Returns the smallest of <c><anno>Term1</anno></c> and
+          <c><anno>Term2</anno></c>.
+	  If the terms are equal, <c><anno>Term1</anno></c> is returned.</p>
       </desc>
     </func>
+
     <func>
       <name name="module_loaded" arity="1"/>
-      <fsummary>Check if a module is loaded</fsummary>
+      <fsummary>Checks if a module is loaded.</fsummary>
       <desc>
-        <p>Returns <c>true</c> if the module <c><anno>Module</anno></c> is loaded,
-          otherwise returns <c>false</c>. It does not attempt to load
+        <p>Returns <c>true</c> if the module <c><anno>Module</anno></c>
+          is loaded, otherwise <c>false</c>. It does not attempt to load
           the module.</p>
         <warning>
           <p>This BIF is intended for the code server (see
-            <seealso marker="kernel:code">code(3)</seealso>) and should not be
+            <seealso marker="kernel:code">code(3)</seealso>) and is not to be
             used elsewhere.</p>
         </warning>
       </desc>
     </func>
+
     <func>
       <name name="monitor" arity="2" clause_i="1"/>
       <name name="monitor" arity="2" clause_i="2"/>
+      <fsummary>Starts monitoring.</fsummary>
       <type name="registered_name"/>
       <type name="registered_process_identifier"/>
       <type name="monitor_process_identifier"/>
-      <fsummary>Start monitoring</fsummary>
       <desc>
 	<p>Send a monitor request of type <c><anno>Type</anno></c> to the
 	entity identified by <c><anno>Item</anno></c>. The caller of
@@ -2694,15 +2872,15 @@ os_prompt% </pre>
 	following format if the monitored state is changed:</p>
         <code type="none">{Tag, <anno>MonitorRef</anno>, <anno>Type</anno>, Object, Info}</code>
 	<note><p>The monitor request is an asynchronous signal. That is, it
-	takes time before the signal reach its destination.</p></note>
-	<p>Currently valid <c><anno>Type</anno></c>s:</p>
+	takes time before the signal reaches its destination.</p></note>
+	<p>Valid <c><anno>Type</anno></c>s:</p>
 	<taglist>
 	  <tag><marker id="monitor_process"/><c>process</c></tag>
 	  <item>
 	    <p>Monitor the existence of the process identified by
-	    <c><anno>Item</anno></c>. Currently valid
+	    <c><anno>Item</anno></c>. Valid
 	    <c><anno>Item</anno></c>s in combination with the
-	    <c>process <anno>Type</anno></c>:</p>
+	    <c>process <anno>Type</anno></c> can be any of the following:</p>
             <taglist>
               <tag><c>pid()</c></tag>
               <item>
@@ -2721,10 +2899,10 @@ os_prompt% </pre>
 		will become monitored.</p>
               </item>
             </taglist>
-            <note><p>When a process is monitored by registered name, the
-	    process that has the registered name at the time when the
+            <note><p>When a registered name is used, the
+	    process that has the registered name when the
             monitor request reach its destination will be monitored.
-            The monitor will not be effected, if the registered name is
+            The monitor is not effected if the registered name is
             unregistered, or unregistered and later registered on another
 	    process.</p></note>
 	    <p>The monitor is triggered either when the monitored process
@@ -2732,22 +2910,22 @@ os_prompt% </pre>
 	    lost. In the case the connection to it is lost, we do not know
 	    if it still exist or not. After this type of monitor has been
 	    triggered, the monitor is automatically removed.</p>
-            <p>When the monitor is triggered a <c>'DOWN'</c> message will
-	    be sent to the monitoring process. A <c>'DOWN'</c> message has
+            <p>When the monitor is triggered a <c>'DOWN'</c> message is
+	    sent to the monitoring process. A <c>'DOWN'</c> message has
 	    the following pattern:</p>
             <code type="none">{'DOWN', MonitorRef, Type, Object, Info}</code>
-            <p>where <c>MonitorRef</c> and <c>Type</c> are the same as
-            described above, and:</p>
+            <p>Here <c>MonitorRef</c> and <c>Type</c> are the same as
+            described earlier, and:</p>
             <taglist>
               <tag><c>Object</c></tag>
               <item>
 		<p>equals:</p>
 		<taglist>
 		  <tag><c><anno>Item</anno></c></tag>
-		  <item>If <c><anno>Item</anno></c> was specified by a
-		  pid.</item>
+		  <item>If <c><anno>Item</anno></c> is specified by a
+		  process identifier.</item>
 		  <tag><c>{RegisteredName, Node}</c></tag>
-		  <item>If <c><anno>Item</anno></c> was specified as
+		  <item>If <c><anno>Item</anno></c> is specified as
 		  <c>RegisteredName</c>, or <c>{RegisteredName, Node}</c>
 		  where <c>Node</c> corresponds to the node that the
 		  monitored process resides on.</item>
@@ -2760,26 +2938,26 @@ os_prompt% </pre>
 		connection to the node where the monitored process
 		resides).</p></item>
 	    </taglist>
-	    <p>The monitoring is turned off either when the <c>'DOWN'</c>
-	    message is sent, or when
+	    <p>The monitoring is turned off when the <c>'DOWN'</c>
+	    message is sent or when
 	    <seealso marker="#demonitor/1">demonitor/1</seealso>
 	    is called.</p>
 	    <p>If an attempt is made to monitor a process on an older node
-	    (where remote process monitoring is not implemented or one
+	    (where remote process monitoring is not implemented or
 	    where remote process monitoring by registered name is not
 	    implemented), the call fails with <c>badarg</c>.</p>
 	    <note>
-	      <p>The format of the <c>'DOWN'</c> message changed in the 5.2
-	      version of the emulator (OTP release R9B) for monitor
-	      <em>by registered name</em>. The <c>Object</c> element of
+	      <p>The format of the <c>'DOWN'</c> message changed in ERTS
+	      version 5.2 (OTP R9B) for monitoring
+	      <em>by registered name</em>. Element <c>Object</c> of
 	      the <c>'DOWN'</c> message could in earlier versions
-	      sometimes be the pid of the monitored process and sometimes
-	      be the registered name. Now the <c>Object</c> element is
+	      sometimes be the process identifier of the monitored process and sometimes
+	      be the registered name. Now element <c>Object</c> is
 	      always a tuple consisting of the registered name and
-	      the node name. Processes on new nodes (emulator version 5.2
-	      or greater) will always get <c>'DOWN'</c> messages on
+	      the node name. Processes on new nodes (ERTS version 5.2
+	      or higher) always get <c>'DOWN'</c> messages on
 	      the new format even if they are monitoring processes on old
-	      nodes. Processes on old nodes will always get <c>'DOWN'</c>
+	      nodes. Processes on old nodes always get <c>'DOWN'</c>
 	      messages on the old format.</p>
 	    </note>
 	  </item>
@@ -2807,8 +2985,8 @@ os_prompt% </pre>
 	    <seealso marker="time_correction#Single_Time_Warp_Mode">single
 	    time warp mode</seealso> is used. When a change from preliminary
 	    to final time offset is made, the monitor will be triggered once
-	    regardless of whether the time offset value was changed due to
-	    the finalization or not.</p>
+	    regardless of whether the time offset value was actually changed
+	    or not.</p>
 
 	    <p>If the runtime system is in
 	    <seealso marker="time_correction#Multi_Time_Warp_Mode">multi
@@ -2836,7 +3014,7 @@ os_prompt% </pre>
 	    <p>When the <c>'CHANGE'</c> message has been received you are
 	    guaranteed not to retrieve the old time offset when calling
 	    <seealso marker="#time_offset/0"><c>erlang:time_offset()</c></seealso>.
-	    Note that you may observe the change of the time offset
+	    Note that you can observe the change of the time offset
 	    when calling <c>erlang:time_offset()</c> before you
 	    get the <c>'CHANGE'</c> message.</p>
 
@@ -2844,67 +3022,71 @@ os_prompt% </pre>
 	</taglist>
 	<p>Making several calls to <c>monitor/2</c> for the same
 	<c><anno>Item</anno></c> and/or <c><anno>Type</anno></c> is not
-	an error; it results in many, completely independent,
-	monitorings.</p>
+	an error; it results in as many independent monitoring instances.</p>
 	<p>The monitor functionality is expected to be extended. That is,
 	other <c><anno>Type</anno></c>s and <c><anno>Item</anno></c>s
-	are expected to be supported in	the future.</p>
+	are expected to be supported in	a future release.</p>
         <note>
-          <p>If/when <c>monitor/2</c> is extended, other
-            possible values for <c>Tag</c>, <c>Object</c>, and
+          <p>If or when <c>monitor/2</c> is extended, other
+            possible values for <c>Tag</c>, <c>Object</c> and
 	    <c>Info</c> in the monitor message will be introduced.</p>
         </note>
       </desc>
     </func>
+
     <func>
       <name name="monitor_node" arity="2"/>
-      <fsummary>Monitor the status of a node</fsummary>
+      <fsummary>Monitors the status of a node.</fsummary>
       <desc>
-        <p>Monitors the status of the node <c><anno>Node</anno></c>. If <c><anno>Flag</anno></c>
-          is <c>true</c>, monitoring is turned on; if <c><anno>Flag</anno></c> is
-          <c>false</c>, monitoring is turned off.</p>
+        <p>Monitors the status of the node <c><anno>Node</anno></c>.
+          If <c><anno>Flag</anno></c>
+          is <c>true</c>, monitoring is turned on. If <c><anno>Flag</anno></c>
+          is <c>false</c>, monitoring is turned off.</p>
         <p>Making several calls to <c>monitor_node(Node, true)</c> for
-          the same <c><anno>Node</anno></c> is not an error; it results in as many,
-          completely independent, monitorings.</p>
+          the same <c><anno>Node</anno></c> is not an error; it results
+          in as many independent monitoring instances.</p>
         <p>If <c><anno>Node</anno></c> fails or does not exist, the message
           <c>{nodedown, Node}</c> is delivered to the process. If a
           process has made two calls to <c>monitor_node(Node, true)</c>
-          and <c><anno>Node</anno></c> terminates, two <c>nodedown</c> messages are
-          delivered to the process. If there is no connection to
-          <c><anno>Node</anno></c>, there will be an attempt to create one. If this
-          fails, a <c>nodedown</c> message is delivered.</p>
+          and <c><anno>Node</anno></c> terminates, two <c>nodedown</c> messages
+          are delivered to the process. If there is no connection to
+          <c><anno>Node</anno></c>, an attempt is made to create one.
+          If this fails, a <c>nodedown</c> message is delivered.</p>
         <p>Nodes connected through hidden connections can be monitored
-          as any other node.</p>
+          as any other nodes.</p>
         <p>Failure: <c>badarg</c> if the local node is not alive.</p>
       </desc>
     </func>
+
     <func>
       <name name="monitor_node" arity="3"/>
-      <fsummary>Monitor the status of a node</fsummary>
+      <fsummary>Monitors the status of a node.</fsummary>
       <desc>
-        <p>Behaves as <c>monitor_node/2</c> except that it allows an
+        <p>Behaves as
+          <seealso marker="#monitor_node/2">monitor_node/2</seealso>
+          except that it allows an
           extra option to be given, namely <c>allow_passive_connect</c>.
-          The option allows the BIF to wait the normal net connection
-          timeout for the <em>monitored node</em> to connect itself,
+          This option allows the BIF to wait the normal network connection
+          time-out for the <em>monitored node</em> to connect itself,
           even if it cannot be actively connected from this node
-          (i.e. it is blocked). The state where this might be useful can
-          only be achieved by using the kernel option
-          <c>dist_auto_connect once</c>. If that kernel option is not
-          used, the <c>allow_passive_connect</c> option has no
-          effect.</p>
+          (that is, it is blocked). The state where this can be useful
+          can only be achieved by using the <c>Kernel</c> option
+          <c>dist_auto_connect once</c>. If that option is not
+          used, option <c>allow_passive_connect</c> has no effect.</p>
         <note>
-          <p>The <c>allow_passive_connect</c> option is used
+          <p>Option <c>allow_passive_connect</c> is used
             internally and is seldom needed in applications where the
-            network topology and the kernel options in effect is known in
-            advance.</p>
+            network topology and the <c>Kernel</c> options in effect
+            are known in advance.</p>
         </note>
         <p>Failure: <c>badarg</c> if the local node is not alive or the
           option list is malformed.</p>
       </desc>
     </func>
+
     <func>
       <name name="monotonic_time" arity="0"/>
-      <fsummary>Current Erlang monotonic time</fsummary>
+      <fsummary>Current Erlang monotonic time.</fsummary>
       <desc>
 	<p>Returns the current
 	<seealso marker="time_correction#Erlang_Monotonic_Time">Erlang
@@ -2917,7 +3099,7 @@ os_prompt% </pre>
 	<seealso marker="time_correction#Monotonically_Increasing">monotonically increasing</seealso> time, but <em>not</em> a
 	<seealso marker="time_correction#Strictly_Monotonically_Increasing">strictly monotonically increasing</seealso>
 	time. That is, consecutive calls to
-	<c>erlang:monotonic_time/0</c> may produce the same result.</p>
+	<c>erlang:monotonic_time/0</c> can produce the same result.</p>
 
 	<p>Different runtime system instances will use different
 	unspecified points in time as base for their Erlang monotonic clocks.
@@ -2925,9 +3107,9 @@ os_prompt% </pre>
 	different runtime system instances. Different runtime system instances
 	may also place this unspecified point in time different relative
 	runtime system start. It may be placed in the future (time at start
-	will be a negative value), the past (time at start will be a
-	positive value), or the runtime system start (time at start will
-	be zero). The monotonic time as of runtime system start can be
+	is a negative value), the past (time at start is a
+	positive value), or the runtime system start (time at start is
+	zero). The monotonic time at runtime system start can be
 	retrieved by calling
 	<seealso marker="#system_info_start_time"><c>erlang:system_info(start_time)</c></seealso>.</p></note>
       </desc>
@@ -2949,61 +3131,68 @@ os_prompt% </pre>
     </func>
     <func>
       <name name="nif_error" arity="1"/>
-      <fsummary>Stop execution with a given reason</fsummary>
+      <fsummary>Stops execution with a given reason.</fsummary>
       <desc>
         <p>Works exactly like
-	 <seealso marker="#error/1">erlang:error/1</seealso>,
-	 but Dialyzer thinks that this BIF will return an arbitrary term.
-	 When used in a stub function for a NIF to generate an
-	 exception when the NIF library is not loaded, Dialyzer
-	 will not generate false warnings.</p>
+	 <seealso marker="#error/1">erlang:error/1</seealso>, but
+	 <c>Dialyzer</c> thinks that this BIF will return an arbitrary
+	 term. When used in a stub function for a NIF to generate an
+	 exception when the NIF library is not loaded, <c>Dialyzer</c>
+	 does not generate false warnings.</p>
       </desc>
     </func>
+
     <func>
       <name name="nif_error" arity="2"/>
-      <fsummary>Stop execution with a given reason</fsummary>
+      <fsummary>Stops execution with a given reason.</fsummary>
       <desc>
         <p>Works exactly like
-	 <seealso marker="#error/2">erlang:error/2</seealso>,
-	 but Dialyzer thinks that this BIF will return an arbitrary term.
-	 When used in a stub function for a NIF to generate an
-	 exception when the NIF library is not loaded, Dialyzer
-	 will not generate false warnings.</p>
+	 <seealso marker="#error/2">erlang:error/2</seealso>, but
+	 <c>Dialyzer</c> thinks that this BIF will return an arbitrary
+	 term. When used in a stub function for a NIF to generate an
+	 exception when the NIF library is not loaded, <c>Dialyzer</c>
+	 does not generate false warnings.</p>
       </desc>
     </func>
+
     <func>
       <name name="node" arity="0"/>
-      <fsummary>Name of the local node</fsummary>
+      <fsummary>Name of the local node.</fsummary>
       <desc>
         <p>Returns the name of the local node. If the node is not alive,
           <c>nonode@nohost</c> is returned instead.</p>
         <p>Allowed in guard tests.</p>
       </desc>
     </func>
+
     <func>
       <name name="node" arity="1"/>
-      <fsummary>At which node is a pid, port or reference located</fsummary>
+      <fsummary>At which node a pid, port, or reference originates.</fsummary>
       <desc>
-        <p>Returns the node where <c><anno>Arg</anno></c> is located. <c><anno>Arg</anno></c> can
-          be a pid, a reference, or a port. If the local node is not
+        <p>Returns the node where <c><anno>Arg</anno></c> originates.
+          <c><anno>Arg</anno></c> can
+          be a process identifier, a reference, or a port.
+          If the local node is not
           alive, <c>nonode@nohost</c> is returned.</p>
         <p>Allowed in guard tests.</p>
       </desc>
     </func>
+
     <func>
       <name name="nodes" arity="0"/>
-      <fsummary>All visible nodes in the system</fsummary>
+      <fsummary>All visible nodes in the system.</fsummary>
       <desc>
-        <p>Returns a list of all visible nodes in the system, excluding
+        <p>Returns a list of all visible nodes in the system, except
           the local node. Same as <c>nodes(visible)</c>.</p>
       </desc>
     </func>
+
     <func>
       <name name="nodes" arity="1"/>
-      <fsummary>All nodes of a certain type in the system</fsummary>
+      <fsummary>All nodes of a certain type in the system.</fsummary>
       <desc>
-        <p>Returns a list of nodes according to argument given.
-          The result returned when the argument is a list, is the list
+        <p>Returns a list of nodes according to the argument given.
+          The returned result when the argument is a list, is the list
           of nodes satisfying the disjunction(s) of the list elements.</p>
         <p><c><anno>NodeType</anno></c> can be any of the following:</p>
         <taglist>
@@ -3025,22 +3214,26 @@ os_prompt% </pre>
           </item>
           <tag><c>known</c></tag>
           <item>
-            <p>Nodes which are known to this node, i.e., connected,
-              previously connected, etc.</p>
+            <p>Nodes that are known to this node. That is, connected
+	    nodes and nodes referred to by process identifiers, port
+	    identifiers and references located on this node.
+	    The set of known nodes is garbage collected. Notice that
+	    this garbage collection can be delayed. For more
+	    information, see
+	    <seealso marker="erlang#system_info_delayed_node_table_gc">delayed_node_table_gc</seealso>.
+	    </p>
           </item>
         </taglist>
         <p>Some equalities: <c>[node()] = nodes(this)</c>,
           <c>nodes(connected) = nodes([visible, hidden])</c>, and
           <c>nodes() = nodes(visible)</c>.</p>
-        <p>If the local node is not alive,
-          <c>nodes(this) == nodes(known) == [nonode@nohost]</c>, for
-          any other <c><anno>Arg</anno></c> the empty list [] is returned.</p>
       </desc>
     </func>
+
     <func>
       <name name="now" arity="0"/>
+      <fsummary>Elapsed time since 00:00 GMT.</fsummary>
       <type name="timestamp"/>
-      <fsummary>Elapsed time since 00:00 GMT</fsummary>
       <desc>
 	<warning><p><em>This function is deprecated! Do not use it!</em>
 	See the users guide chapter
@@ -3050,107 +3243,101 @@ os_prompt% </pre>
 	section for information on what to use instead of <c>erlang:now/0</c>.
 	</p></warning>
         <p>Returns the tuple <c>{MegaSecs, Secs, MicroSecs}</c> which is
-          the elapsed time since 00:00 GMT, January 1, 1970 (zero hour)
+          the elapsed time since 00:00 GMT, January 1, 1970 (zero hour),
           on the assumption that the underlying OS supports this.
-          Otherwise, some other point in time is chosen. It is also
-          guaranteed that subsequent calls to this BIF returns
+          Otherwise some other point in time is chosen. It is also
+          guaranteed that subsequent calls to this BIF return
           continuously increasing values. Hence, the return value from
-          <c>now()</c> can be used to generate unique time-stamps,
-          and if it is called in a tight loop on a fast machine
+          <c>now()</c> can be used to generate unique time-stamps.
+          If it is called in a tight loop on a fast machine,
           the time of the node can become skewed.</p>
-        <p>It can only be used to check the local time of day if
-          the time-zone info of the underlying operating system is
+        <p>Can only be used to check the local time of day if
+          the time-zone information of the underlying OS is
           properly configured.</p>
       </desc>
     </func>
+
     <func>
       <name name="open_port" arity="2"/>
-      <fsummary>Open a port</fsummary>
+      <fsummary>Opens a port.</fsummary>
       <desc>
         <p>Returns a port identifier as the result of opening a
           new Erlang port. A port can be seen as an external Erlang
-          process.
-	</p>
+          process.</p>
 	<p>The name of the executable as well as the arguments
-	given in <c>cd</c>, <c>env</c>, <c>args</c> and <c>arg0</c> is subject to
-	Unicode file name translation if the system is running
+	given in <c>cd</c>, <c>env</c>, <c>args</c>, and <c>arg0</c> are
+        subject to Unicode file name translation if the system is running
 	in Unicode file name mode. To avoid
-	translation or force i.e. UTF-8, supply the executable
+	translation or to force, for example UTF-8, supply the executable
 	and/or arguments as a binary in the correct
-	encoding. See the <seealso
-	marker="kernel:file">file</seealso> module, the
-	<seealso marker="kernel:file#native_name_encoding/0">
-	  file:native_name_encoding/0</seealso> function and the
-	  <seealso marker="stdlib:unicode_usage">stdlib users guide
-	</seealso> for details.</p>
-
-	<note><p>The characters in the name (if given as a list)
-	can only be &gt; 255 if the Erlang VM is started in
-	Unicode file name translation mode, otherwise the name
+	encoding. For details, see the module
+        <seealso marker="kernel:file">file</seealso>, the function
+	<seealso marker="kernel:file#native_name_encoding/0">file:native_name_encoding/0</seealso>, and the
+	<seealso marker="stdlib:unicode_usage">STDLIB </seealso>
+        User's Guide.</p>
+	<note><p>The characters in the name (if given as a list) can
+	only be higher than 255 if the Erlang Virtual Machine is started
+	in Unicode file name translation mode. Otherwise the name
 	of the executable is limited to the ISO-latin-1
 	character set.</p></note>
-
-	<p><c><anno>PortName</anno></c> is one of the following:</p>
+	<p><c><anno>PortName</anno></c> can be any of the following:</p>
         <taglist>
           <tag><c>{spawn, <anno>Command</anno>}</c></tag>
           <item>
-            <p>Starts an external program. <c><anno>Command</anno></c> is the name
-              of the external program which will be run. <c><anno>Command</anno></c>
+            <p>Starts an external program. <c><anno>Command</anno></c>
+              is the name of the external program to be run.
+              <c><anno>Command</anno></c>
               runs outside the Erlang work space unless an Erlang
-              driver with the name <c><anno>Command</anno></c> is found. If found,
-              that driver will be started. A driver runs in the Erlang
-              workspace, which means that it is linked with the Erlang
+              driver with the name <c><anno>Command</anno></c> is found.
+              If found, that driver is started. A driver runs in the Erlang
+              work space, which means that it is linked with the Erlang
               runtime system.</p>
             <p>When starting external programs on Solaris, the system
               call <c>vfork</c> is used in preference to <c>fork</c>
               for performance reasons, although it has a history of
-              being less robust. If there are problems with using
-              <c>vfork</c>, setting the environment variable
-              <c>ERL_NO_VFORK</c> to any value will cause <c>fork</c>
+              being less robust. If there are problems using
+              <c>vfork</c>, setting environment variable
+              <c>ERL_NO_VFORK</c> to any value causes <c>fork</c>
               to be used instead.</p>
-
-	      <p>For external programs, the <c>PATH</c> is searched
+	      <p>For external programs, <c>PATH</c> is searched
 	      (or an equivalent method is used to find programs,
-	      depending on operating system). This is done by invoking
-	      the shell on certain platforms. The first space
-	      separated token of the command will be considered as the
+	      depending on OS). This is done by invoking
+	      the shell on certain platforms. The first space-separated
+	      token of the command is considered as the
 	      name of the executable (or driver). This (among other
 	      things) makes this option unsuitable for running
-	      programs having spaces in file or directory names. Use
-	      {spawn_executable, <anno>Command</anno>} instead if spaces in executable
-	      file names is desired.</p>
+	      programs having spaces in file names or directory names.
+              If spaces in executable file names are desired, use
+	      <c>{spawn_executable, <anno>Command</anno>}</c> instead.</p>
           </item>
           <tag><c>{spawn_driver, <anno>Command</anno>}</c></tag>
           <item>
 	      <p>Works like <c>{spawn, <anno>Command</anno>}</c>, but demands the
-	      first (space separated) token of the command to be the name of a
+	      first (space-separated) token of the command to be the name of a
 	      loaded driver. If no driver with that name is loaded, a
 	      <c>badarg</c> error is raised.</p>
           </item>
           <tag><c>{spawn_executable, <anno>FileName</anno>}</c></tag>
           <item>
-
 	      <p>Works like <c>{spawn, <anno>FileName</anno>}</c>, but only runs
-	      external executables. The <c><anno>FileName</anno></c> in its whole
-	      is used as the name of the executable, including any
-	      spaces. If arguments are to be passed, the
-	      <c>args</c> and <c>arg0</c> <c><anno>PortSettings</anno></c> can be used.</p>
-
-	      <p>The shell is not usually invoked to start the
-	      program, it's executed directly. Neither is the
-	      <c>PATH</c> (or equivalent) searched. To find a program
-	      in the PATH to execute, use <seealso
-	      marker="kernel:os#find_executable/1">os:find_executable/1</seealso>.</p>
+	      external executables. <c><anno>FileName</anno></c> in its whole
+	      is used as the name of the executable, including any spaces.
+	      If arguments are to be passed, the <c><anno>PortSettings</anno></c>
+	      <c>args</c> and <c>arg0</c> can be used.</p>
+	      <p>The shell is usually not invoked to start the
+	      program, it is executed directly. <c>PATH</c> (or
+              equivalent) is not searched. To find a program
+	      in <c>PATH</c> to execute, use
+              <seealso marker="kernel:os#find_executable/1">os:find_executable/1</seealso>.</p>
 	      <p>Only if a shell script or <c>.bat</c> file is
-	      executed, the appropriate command interpreter will
-	      implicitly be invoked, but there will still be no
-	      command argument expansion or implicit PATH search.</p>
-  
-	      <p>If the <c><anno>FileName</anno></c> cannot be run, an error
-	      exception, with the posix error code as the reason, is
-	      raised. The error reason may differ between operating
-	      systems. Typically the error <c>enoent</c> is raised
-	      when one tries to run a program that is not found and
+	      executed, the appropriate command interpreter is
+	      invoked implicitly, but there is still no
+	      command argument expansion or implicit <c>PATH</c> search.</p>
+	      <p>If <c><anno>FileName</anno></c> cannot be run, an error
+	      exception is raised, with the POSIX error code as the reason.
+	      The error reason can differ between OSs.
+	      Typically the error <c>enoent</c> is raised when an
+	      attempt is made to run a program that is not found and
 	      <c>eacces</c> is raised when the given file is not
 	      executable.</p>
           </item>
@@ -3160,19 +3347,18 @@ os_prompt% </pre>
               file descriptors used by Erlang. The file descriptor
               <c><anno>In</anno></c> can be used for standard input, and the file
               descriptor <c><anno>Out</anno></c> for standard output. It is only
-              used for various servers in the Erlang operating system
-              (<c>shell</c> and <c>user</c>). Hence, its use is very
-              limited.</p>
+              used for various servers in the Erlang OS (<c>shell</c>
+              and <c>user</c>). Hence, its use is limited.</p>
           </item>
         </taglist>
         <p><c><anno>PortSettings</anno></c> is a list of settings for the port.
-          Valid settings are:</p>
+          The valid settings are as follows:</p>
         <taglist>
           <tag><c>{packet, <anno>N</anno>}</c></tag>
           <item>
             <p>Messages are preceded by their length, sent in <c><anno>N</anno></c>
-              bytes, with the most significant byte first. Valid values
-              for <c>N</c> are 1, 2, or 4.</p>
+              bytes, with the most significant byte first. The valid values
+              for <c>N</c> are 1, 2, and 4.</p>
           </item>
           <tag><c>stream</c></tag>
           <item>
@@ -3183,116 +3369,108 @@ os_prompt% </pre>
           <tag><c>{line, <anno>L</anno>}</c></tag>
           <item>
             <p>Messages are delivered on a per line basis. Each line
-              (delimited by the OS-dependent newline sequence) is
-              delivered in one single message. The message data format
-              is <c>{Flag, Line}</c>, where <c>Flag</c> is either
-              <c>eol</c> or <c>noeol</c> and <c>Line</c> is the actual
-              data delivered (without the newline sequence).</p>
+              (delimited by the OS-dependent new line sequence) is
+              delivered in a single message. The message data format
+              is <c>{Flag, Line}</c>, where <c>Flag</c> is
+              <c>eol</c> or <c>noeol</c>, and <c>Line</c> is the
+              data delivered (without the new line sequence).</p>
             <p><c><anno>L</anno></c> specifies the maximum line length in bytes.
-              Lines longer than this will be delivered in more than one
-              message, with the <c>Flag</c> set to <c>noeol</c> for all
+              Lines longer than this are delivered in more than one
+              message, with <c>Flag</c> set to <c>noeol</c> for all
               but the last message. If end of file is encountered
-              anywhere else than immediately following a newline
-              sequence, the last line will also be delivered with
-              the <c>Flag</c> set to <c>noeol</c>. In all other cases,
+              anywhere else than immediately following a new line
+              sequence, the last line is also delivered with
+              <c>Flag</c> set to <c>noeol</c>. Otherwise
               lines are delivered with <c>Flag</c> set to <c>eol</c>.</p>
-            <p>The <c>{packet, <anno>N</anno>}</c> and <c>{line, <anno>L</anno>}</c> settings are
-              mutually exclusive.</p>
+            <p>The <c>{packet, <anno>N</anno>}</c> and <c>{line,
+              <anno>L</anno>}</c> settings are mutually exclusive.</p>
           </item>
           <tag><c>{cd, <anno>Dir</anno>}</c></tag>
           <item>
-            <p>This is only valid for <c>{spawn, <anno>Command</anno>}</c> and
-	    <c>{spawn_executable, <anno>FileName</anno>}</c>.
+            <p>Only valid for <c>{spawn, <anno>Command</anno>}</c> and
+	      <c>{spawn_executable, <anno>FileName</anno>}</c>.
               The external program starts using <c><anno>Dir</anno></c> as its
-              working directory. <c><anno>Dir</anno></c> must be a string.
-		  </p>
+              working directory. <c><anno>Dir</anno></c> must be a string.</p>
           </item>
           <tag><c>{env, <anno>Env</anno>}</c></tag>
           <item>
-            <p>This is only valid for <c>{spawn, <anno>Command</anno>}</c> and
+            <p>Only valid for <c>{spawn, <anno>Command</anno>}</c> and
 	    <c>{spawn_executable, <anno>FileName</anno>}</c>.
               The environment of the started process is extended using
               the environment specifications in <c><anno>Env</anno></c>.</p>
-            <p><c><anno>Env</anno></c> should be a list of tuples <c>{<anno>Name</anno>, <anno>Val</anno>}</c>,
-              where <c><anno>Name</anno></c> is the name of an environment variable,
-              and <c><anno>Val</anno></c> is the value it is to have in the spawned
-              port process. Both <c><anno>Name</anno></c> and <c><anno>Val</anno></c> must be
-              strings. The one exception is <c><anno>Val</anno></c> being the atom
+            <p><c><anno>Env</anno></c> is to be a list of tuples
+              <c>{<anno>Name</anno>, <anno>Val</anno>}</c>,
+              where <c><anno>Name</anno></c> is the name of an
+              environment variable, and <c><anno>Val</anno></c> is the
+              value it is to have in the spawned
+              port process. Both <c><anno>Name</anno></c> and
+              <c><anno>Val</anno></c> must be strings. The one
+              exception is <c><anno>Val</anno></c> being the atom
               <c>false</c> (in analogy with <c>os:getenv/1</c>), which
-              removes the environment variable.
-	    </p>
+              removes the environment variable.</p>
           </item>
 	  <tag><c>{args, [ string() | binary() ]}</c></tag>
           <item>
-
-	    <p>This option is only valid for <c>{spawn_executable, <anno>FileName</anno>}</c>
+	    <p>Only valid for <c>{spawn_executable, <anno>FileName</anno>}</c>
 	    and specifies arguments to the executable. Each argument
 	    is given as a separate string and (on Unix) eventually
 	    ends up as one element each in the argument vector. On
-	    other platforms, similar behavior is mimicked.</p>
-
-	    <p>The arguments are not expanded by the shell prior to
-	      being supplied to the executable, most notably this
-	      means that file wildcard expansion will not happen. Use
-	      <seealso
-	      marker="stdlib:filelib#wildcard/1">filelib:wildcard/1</seealso>
-	      to expand wildcards for the arguments. Note that even if
+	    other platforms, a similar behavior is mimicked.</p>
+	    <p>The arguments are not expanded by the shell before
+	      being supplied to the executable. Most notably this
+	      means that file wild card expansion does not happen.
+              To expand wild cards for the arguments, use
+	      <seealso marker="stdlib:filelib#wildcard/1">filelib:wildcard/1</seealso>.
+	      Notice that even if
 	      the program is a Unix shell script, meaning that the
-	      shell will ultimately be invoked, wildcard expansion
-	      will not happen and the script will be provided with the
-	      untouched arguments. On Windows&reg;, wildcard expansion
-	      is always up to the program itself, why this isn't an
-	      issue.</p>
-
-	    <p>Note also that the actual executable name (a.k.a. <c>argv[0]</c>)
-	    should not be given in this list. The proper executable name will
-	    automatically be used as argv[0] where applicable.</p>
-
-	    <p>If one, for any reason, wants to explicitly set the
-	    program name in the argument vector, the <c>arg0</c>
-	    option can be used.</p>
-
+	      shell ultimately is invoked, wild card expansion
+	      does not happen, and the script is provided with the
+	      untouched arguments. On Windows, wild card expansion
+	      is always up to the program itself, therefore this is
+	      not an issue issue.</p>
+	    <p>The executable name (also known as <c>argv[0]</c>)
+	      is not to be given in this list. The proper executable name
+	      is automatically used as argv[0], where applicable.</p>
+	    <p>If you explicitly want to set the
+	      program name in the argument vector, option <c>arg0</c>
+	      can be used.</p>
           </item>
 	  <tag><c>{arg0, string() | binary()}</c></tag>
           <item>
-
-	    <p>This option is only valid for <c>{spawn_executable, <anno>FileName</anno>}</c>
+	    <p>Only valid for <c>{spawn_executable, <anno>FileName</anno>}</c>
 	    and explicitly specifies the program name argument when
-	    running an executable. This might in some circumstances,
-	    on some operating systems, be desirable. How the program
-	    responds to this is highly system dependent and no specific 
+	    running an executable. This can in some circumstances,
+	    on some OSs, be desirable. How the program
+	    responds to this is highly system-dependent and no specific
 	    effect is guaranteed.</p>
-
 	  </item> 
-
           <tag><c>exit_status</c></tag>
           <item>
-            <p>This is only valid for <c>{spawn, <anno>Command</anno>}</c> where
-              <c><anno>Command</anno></c> refers to an external program, and for
-	      <c>{spawn_executable, <anno>FileName</anno>}</c>.</p>
+            <p>Only valid for <c>{spawn, <anno>Command</anno>}</c>, where
+              <c><anno>Command</anno></c> refers to an external program, and
+	      for <c>{spawn_executable, <anno>FileName</anno>}</c>.</p>
             <p>When the external process connected to the port exits, a
               message of the form <c>{Port,{exit_status,Status}}</c> is
               sent to the connected process, where <c>Status</c> is the
               exit status of the external process. If the program
-              aborts, on Unix the same convention is used as the shells
-              do (i.e., 128+signal).</p>
-            <p>If the <c>eof</c> option has been given as well,
-              the <c>eof</c> message and the <c>exit_status</c> message
-              appear in an unspecified order.</p>
-            <p>If the port program closes its stdout without exiting,
-              the <c>exit_status</c> option will not work.</p>
+              aborts on Unix, the same convention is used as the shells
+              do (that is, 128+signal).</p>
+            <p>If option <c>eof</c> is also given, the messages <c>eof</c>
+              and <c>exit_status</c> appear in an unspecified order.</p>
+            <p>If the port program closes its <c>stdout</c> without exiting,
+              option <c>exit_status</c> does not work.</p>
           </item>
           <tag><c>use_stdio</c></tag>
           <item>
-            <p>This is only valid for <c>{spawn, <anno>Command</anno>}</c> and
+            <p>Only valid for <c>{spawn, <anno>Command</anno>}</c> and
 	    <c>{spawn_executable, <anno>FileName</anno>}</c>. It
               allows the standard input and output (file descriptors 0
-              and 1) of the spawned (UNIX) process for communication
+              and 1) of the spawned (Unix) process for communication
               with Erlang.</p>
           </item>
           <tag><c>nouse_stdio</c></tag>
           <item>
-            <p>The opposite of <c>use_stdio</c>. Uses file descriptors
+            <p>The opposite of <c>use_stdio</c>. It uses file descriptors
               3 and 4 for communication with Erlang.</p>
           </item>
           <tag><c>stderr_to_stdout</c></tag>
@@ -3304,14 +3482,15 @@ os_prompt% </pre>
           </item>
           <tag><c>overlapped_io</c></tag>
           <item>
-            <p>Affects ports to external programs on Windows&reg; only.
-	    The standard input and standard output handles of the port program
-	    will, if this option is supplied, be opened with the flag
-	    FILE_FLAG_OVERLAPPED, so that the port program can (and has to) do
+            <p>Affects ports to external programs on Windows only. The
+	    standard input and standard output handles of the port program
+	    are, if this option is supplied, opened with flag
+	    <c>FILE_FLAG_OVERLAPPED</c>, so that the port program can
+            (and must) do
 	    overlapped I/O on its standard handles. This is not normally
 	    the case for simple port programs, but an option of value for the
-	    experienced Windows programmer. <em>On all other platforms, this 
-	    option is silently discarded</em>.</p> 
+	    experienced Windows programmer. <em>On all other platforms, this
+	    option is silently discarded.</em></p>
           </item>
           <tag><c>in</c></tag>
           <item>
@@ -3323,345 +3502,354 @@ os_prompt% </pre>
           </item>
           <tag><c>binary</c></tag>
           <item>
-            <p>All IO from the port are binary data objects as opposed
+            <p>All I/O from the port is binary data objects as opposed
               to lists of bytes.</p>
           </item>
           <tag><c>eof</c></tag>
           <item>
-            <p>The port will not be closed at the end of the file and
-              produce an exit signal. Instead, it will remain open and
-              a <c>{Port, eof}</c> message will be sent to the process
+            <p>The port is not closed at the end of the file and does not
+              produce an exit signal. Instead, it remains open and
+              a <c>{Port, eof}</c> message is sent to the process
               holding the port.</p>
           </item>
           <tag><c>hide</c></tag>
           <item>
-            <p>When running on Windows, suppress creation of a new
+            <p>When running on Windows, suppresses creation of a new
 	      console window when spawning the port program.
 	      (This option has no effect on other platforms.)</p>
           </item>
-          <tag><marker id="open_port_parallelism"><c>{parallelism, Boolean}</c></marker></tag>
+          <tag><c>{parallelism, Boolean}</c></tag>
           <item>
-            <p>Set scheduler hint for port parallelism. If set to <c>true</c>,
-	    the VM will schedule port tasks when doing so will improve
-	    parallelism in the system. If set to <c>false</c>, the VM will
-	    try to perform port tasks immediately, improving latency at the
-            expense of parallelism. The default can be set on system startup
-            by passing the
-	    <seealso marker="erl#+spp">+spp</seealso> command line argument
-	    to <seealso marker="erl">erl(1)</seealso>.
-	    </p>
+          <marker id="open_port_parallelism"></marker>
+            <p>Sets scheduler hint for port parallelism. If set to
+            <c>true</c>, the Virtual Machine schedules port tasks;
+            when doing so, it improves parallelism in the system. If set
+            to <c>false</c>, the Virtual Machine tries to
+	    perform port tasks immediately, improving latency at the
+            expense of parallelism. The default can be set at system startup
+            by passing command-line argument
+	    <seealso marker="erl#+spp">+spp</seealso> to <c>erl(1)</c>.</p>
           </item>
         </taglist>
-        <p>The default is <c>stream</c> for all types of port and
+        <p>Default is <c>stream</c> for all port types and
           <c>use_stdio</c> for spawned ports.</p>
         <p>Failure: If the port cannot be opened, the exit reason is
-	  <c>badarg</c>, <c>system_limit</c>, or the Posix error code which
-	  most closely describes the error, or <c>einval</c> if no Posix code
-	  is appropriate:</p>
+	  <c>badarg</c>, <c>system_limit</c>, or the POSIX error code that
+	  most closely describes the error, or <c>einval</c> if no POSIX
+          code is appropriate:</p>
         <taglist>
           <tag><c>badarg</c></tag>
-          <item>
-            <p>Bad input arguments to <c>open_port</c>.</p>
+          <item>Bad input arguments to <c>open_port</c>.
           </item>
           <tag><c>system_limit</c></tag>
-          <item>
-            <p>All available ports in the Erlang emulator are in use.</p>
+          <item>All available ports in the Erlang emulator are in use.
           </item>
           <tag><c>enomem</c></tag>
-          <item>
-            <p>There was not enough memory to create the port.</p>
+          <item>Not enough memory to create the port.
           </item>
           <tag><c>eagain</c></tag>
-          <item>
-            <p>There are no more available operating system processes.</p>
+          <item>No more available OS processes.
           </item>
           <tag><c>enametoolong</c></tag>
-          <item>
-            <p>The external command given was too long.</p>
+          <item>Too long external command.
           </item>
           <tag><c>emfile</c></tag>
-          <item>
-            <p>There are no more available file descriptors (for the operating system process
-	    that the Erlang emulator runs in).</p>
+          <item>No more available file descriptors (for the
+            OS process that the Erlang emulator runs in).
           </item>
           <tag><c>enfile</c></tag>
-          <item>
-            <p>The file table is full (for the entire operating system).</p>
+          <item>Full file table (for the entire OS).
           </item>
           <tag><c>eacces</c></tag>
-          <item>
-            <p>The <c>Command</c> given in <c>{spawn_executable, Command}</c> does not point out an executable file.</p>
+          <item><c>Command</c> given in <c>{spawn_executable, Command}</c>
+            does not point out an executable file.
           </item>
           <tag><c>enoent</c></tag>
-          <item>
-            <p>The <c><anno>FileName</anno></c> given in <c>{spawn_executable, <anno>FileName</anno>}</c> does not point out an existing file.</p>
+          <item><c><anno>FileName</anno></c> given in
+            <c>{spawn_executable, <anno>FileName</anno>}</c>
+            does not point out an existing file.
           </item>
         </taglist>
         <p>During use of a port opened using <c>{spawn, Name}</c>,
-          <c>{spawn_driver, Name}</c> or <c>{spawn_executable, Name}</c>,
+          <c>{spawn_driver, Name}</c>, or <c>{spawn_executable, Name}</c>,
           errors arising when sending messages to it are reported to
           the owning process using signals of the form
-          <c>{'EXIT', Port, PosixCode}</c>. See <c>file(3)</c> for
-          possible values of <c>PosixCode</c>.</p>
+          <c>{'EXIT', Port, PosixCode}</c>. For the possible values of
+          <c>PosixCode</c>, see the
+          <seealso marker="kernel:file">file(3)</seealso>
+          manual page in <c>Kernel</c>.</p>
 	  <p>The maximum number of ports that can be open at the same
-          time can be configured by passing the 
-	  <seealso marker="erl#max_ports"><c>+Q</c></seealso>
-	  command line flag to
-	  <seealso marker="erl"><c>erl(1)</c></seealso>.</p>
+          time can be configured by passing command-line flag
+	  <seealso marker="erl#max_ports"><c>+Q</c></seealso> to
+	  <c>erl(1)</c>.</p>
       </desc>
     </func>
+
     <func>
       <name name="phash" arity="2"/>
+      <fsummary>Portable hash function.</fsummary>
       <type_desc variable="Range">Range = 1..2^32, Hash = 1..Range</type_desc>
-      <fsummary>Portable hash function</fsummary>
       <desc>
-        <p>Portable hash function that will give the same hash for
+        <p>Portable hash function that gives the same hash for
           the same Erlang term regardless of machine architecture and
-          ERTS version (the BIF was introduced in ERTS 4.9.1.1). Range
-          can be between 1 and 2^32, the function returns a hash value
-          for <c><anno>Term</anno></c> within the range <c>1..<anno>Range</anno></c>.</p>
-        <p>This BIF could be used instead of the old deprecated
-          <c>erlang:hash/2</c> BIF, as it calculates better hashes for
-          all data-types, but consider using <c>phash2/1,2</c> instead.</p>
+          <c>ERTS</c> version (the BIF was introduced in <c>ERTS</c> 4.9.1.1).
+          The function returns a hash value for
+          <c><anno>Term</anno></c> within the range
+          <c>1..<anno>Range</anno></c>. The maximum value for
+	  <c><anno>Range</anno></c> is 2^32.</p>
+        <p>This BIF can be used instead of the old deprecated BIF
+          <c>erlang:hash/2</c>, as it calculates better hashes for
+          all data types, but consider using <c>phash2/1,2</c> instead.</p>
       </desc>
     </func>
+
     <func>
       <name name="phash2" arity="1"/>
       <name name="phash2" arity="2"/>
+      <fsummary>Portable hash function.</fsummary>
       <type_desc variable="Range">1..2^32</type_desc>
       <type_desc variable="Hash">0..Range-1</type_desc>
-      <fsummary>Portable hash function</fsummary>
       <desc>
-        <p>Portable hash function that will give the same hash for
+        <p>Portable hash function that gives the same hash for
           the same Erlang term regardless of machine architecture and
-          ERTS version (the BIF was introduced in ERTS 5.2). Range can
-          be between 1 and 2^32, the function returns a hash value for
-          <c><anno>Term</anno></c> within the range <c>0..<anno>Range</anno>-1</c>. When called
-          without the <c><anno>Range</anno></c> argument, a value in the range
-          <c>0..2^27-1</c> is returned.</p>
-        <p>This BIF should always be used for hashing terms. It
+          <c>ERTS</c> version (the BIF was introduced in <c>ERTS</c> 5.2).
+          The function returns a hash value for
+	  <c><anno>Term</anno></c> within the range
+	  <c>0..<anno>Range</anno>-1</c>. The maximum value for
+	  <c><anno>Range</anno></c> is 2^32. When without argument
+          <c><anno>Range</anno></c>, a value in the range
+          0..2^27-1 is returned.</p>
+        <p>This BIF is always to be used for hashing terms. It
           distributes small integers better than <c>phash/2</c>, and
           it is faster for bignums and binaries.</p>
-        <p>Note that the range <c>0..<anno>Range</anno>-1</c> is different from
-          the range of <c>phash/2</c> (<c>1..<anno>Range</anno></c>).</p>
+        <p>Notice that the range <c>0..<anno>Range</anno>-1</c> is
+          different from the range of <c>phash/2</c>, which is
+          <c>1..<anno>Range</anno></c>.</p>
       </desc>
     </func>
+
     <func>
       <name name="pid_to_list" arity="1"/>
-      <fsummary>Text representation of a pid</fsummary>
+      <fsummary>Text representation of a pid.</fsummary>
       <desc>
-        <p>Returns a string which corresponds to the text
+        <p>Returns a string corresponding to the text
           representation of <c><anno>Pid</anno></c>.</p>
         <warning>
-          <p>This BIF is intended for debugging and for use in
-            the Erlang operating system. It should not be used in
-            application programs.</p>
+          <p>This BIF is intended for debugging and is not to be used
+	  in application programs.</p>
         </warning>
       </desc>
     </func>
+
     <func>
       <name name="port_close" arity="1"/>
-      <fsummary>Close an open port</fsummary>
+      <fsummary>Closes an open port.</fsummary>
       <desc>
         <p>Closes an open port. Roughly the same as
-          <c><anno>Port</anno> ! {self(), close}</c> except for the error behaviour
-          (see below), being synchronous, and that the port does
-	  <em>not</em> reply with <c>{Port, closed}</c>. Any process may
+          <c><anno>Port</anno> ! {self(), close}</c> except for the error behavior
+          (see the following), being synchronous, and that the port does
+	  <em>not</em> reply with <c>{Port, closed}</c>. Any process can
 	  close a port with <c>port_close/1</c>, not only the port owner
 	  (the connected process). If the calling process is linked to
-	  port identified by <c><anno>Port</anno></c>, an exit signal due
-	  to that link will be received by the process prior to the return
-	  from <c>port_close/1</c>.</p>
-        <p>For comparison: <c><anno>Port</anno> ! {self(), close}</c> fails with
-          <c>badarg</c> if <c><anno>Port</anno></c> cannot be sent to (i.e.,
-          <c><anno>Port</anno></c> refers neither to a port nor to a process). If
-          <c><anno>Port</anno></c> is a closed port nothing happens. If <c><anno>Port</anno></c>
+          the port identified by <c><anno>Port</anno></c>, the exit
+	  signal from the port is guaranteed to be delivered before
+	  <c>port_close/1</c> returns.</p> 
+        <p>For comparison: <c><anno>Port</anno> ! {self(), close}</c>
+	  only fails with <c>badarg</c> if <c><anno>Port</anno></c> does
+	  not refer to a port or a process. If <c><anno>Port</anno></c>
+	  is a closed port, nothing happens. If <c><anno>Port</anno></c>
           is an open port and the calling process is the port owner,
-          the port replies with <c>{Port, closed}</c> when all buffers
-          have been flushed and the port really closes, but if
-          the calling process is not the port owner the <em>port owner</em> fails with <c>badsig</c>.</p>
-
-        <p>Note that any process can close a port using
-          <c><anno>Port</anno> ! {PortOwner, close}</c> just as if it itself was
+	  the port replies with <c>{Port, closed}</c> when all buffers
+	  have been flushed and the port really closes. If the calling
+	  process is not the port owner, the <em>port owner</em> fails
+	  with <c>badsig</c>.</p> 
+        <p>Notice that any process can close a port using
+          <c><anno>Port</anno> ! {PortOwner, close}</c> as if it itself was
           the port owner, but the reply always goes to the port owner.</p>
-	<p>As of OTP-R16 <c><anno>Port</anno> ! {PortOwner, close}</c> is truly
-	   asynchronous. Note that this operation has always been
+	<p>As from OTP R16, <c><anno>Port</anno> ! {PortOwner, close}</c> is truly
+	   asynchronous. Notice that this operation has always been
 	   documented as an asynchronous operation, while the underlying
 	   implementation has been synchronous. <c>port_close/1</c> is
-	   however still fully synchronous. This due to its error
+	   however still fully synchronous. This because of its error
 	   behavior.</p>
-        <p>Failure:</p>
-        <taglist>
-          <tag><c>badarg</c></tag>
-          <item>
-	     If <c><anno>Port</anno></c> is not an identifier of an open
-	     port, or the registered name of an open port. If the calling
-	     process was linked to the previously open port identified by
-	     <c><anno>Port</anno></c>, an exit signal due to this link
-	     was received by the process prior to this exception.
-	  </item>
-	</taglist>
+        <p>Failure: <c>badarg</c> if <c><anno>Port</anno></c> is not an identifier
+           of an open port, or the registered name of an open port.
+	   If the calling process was previously linked to the closed
+	   port, identified by <c><anno>Port</anno></c>, the exit
+	   signal from the port is guaranteed to be delivered before
+	   this <c>badarg</c> exception occurs.</p>
       </desc>
     </func>
+
     <func>
       <name name="port_command" arity="2"/>
-      <fsummary>Send data to a port</fsummary>
+      <fsummary>Sends data to a port.</fsummary>
       <desc>
         <p>Sends data to a port. Same as
-          <c><anno>Port</anno> ! {PortOwner, {command, Data}}</c> except for the error
-          behaviour and being synchronous (see below). Any process may
-	  send data to a port with <c>port_command/2</c>, not only the
+          <c><anno>Port</anno> ! {PortOwner, {command, Data}}</c> except
+          for the error
+          behavior and being synchronous (see the following). Any process
+	  can send data to a port with <c>port_command/2</c>, not only the
 	  port owner (the connected process).</p>
         <p>For comparison: <c><anno>Port</anno> ! {PortOwner, {command, Data}}</c>
-          fails with <c>badarg</c> if <c><anno>Port</anno></c> cannot be sent to
-          (i.e., <c><anno>Port</anno></c> refers neither to a port nor to a process).
-          If <c><anno>Port</anno></c> is a closed port the data message disappears
+          only fails with <c>badarg</c> if <c><anno>Port</anno></c>
+	  does not refer to a port or a process. If
+	  <c><anno>Port</anno></c> is a closed port, the data message
+	  disappears 
           without a sound. If <c><anno>Port</anno></c> is open and the calling
           process is not the port owner, the <em>port owner</em> fails
           with <c>badsig</c>. The port owner fails with <c>badsig</c>
-          also if <c><anno>Data</anno></c> is not a valid IO list.</p>
-        <p>Note that any process can send to a port using
-          <c><anno>Port</anno> ! {PortOwner, {command, <anno>Data</anno>}}</c> just as if it
-          itself was the port owner.</p>
-	<p>If the port is busy, the calling process will be suspended
-	   until the port is not busy anymore.</p>
-	<p>As of OTP-R16 <c><anno>Port</anno> ! {PortOwner, {command, Data}}</c> is
-	   truly asynchronous. Note that this operation has always been
+          also if <c><anno>Data</anno></c> is an invalid I/O list.</p>
+        <p>Notice that any process can send to a port using
+          <c><anno>Port</anno> ! {PortOwner, {command, <anno>Data</anno>}}</c>
+          as if it itself was the port owner.</p>
+	<p>If the port is busy, the calling process is suspended
+	   until the port is not busy any more.</p>
+	<p>As from OTP-R16, <c><anno>Port</anno> ! {PortOwner, {command, Data}}</c>
+	   is truly asynchronous. Notice that this operation has always been
 	   documented as an asynchronous operation, while the underlying
 	   implementation has been synchronous. <c>port_command/2</c> is
-	   however still fully synchronous. This due to its error
+	   however still fully synchronous. This because of its error
 	   behavior.</p>
         <p>Failures:</p>
         <taglist>
           <tag><c>badarg</c></tag>
           <item>
 	     If <c><anno>Port</anno></c> is not an identifier of an open
-	     port, or the registered name of an open port. If the calling
-	     process was linked to the previously open port identified by
-	     <c><anno>Port</anno></c>, an exit signal due to this link
-	     was received by the process prior to this exception.
+	     port, or the registered name of an open port. If the
+	     calling process was previously linked to the closed port,
+	     identified by <c><anno>Port</anno></c>, the exit signal
+	     from the port is guaranteed to be delivered before this
+	     <c>badarg</c> exception occurs.
 	  </item>
           <tag><c>badarg</c></tag>
           <item>
-	     If <c><anno>Data</anno></c> is not a valid io list.
+	     If <c><anno>Data</anno></c> is an invalid I/O list.
 	  </item>
 	</taglist>
       </desc>
     </func>
+
     <func>
       <name name="port_command" arity="3"/>
-      <fsummary>Send data to a port</fsummary>
+      <fsummary>Sends data to a port.</fsummary>
       <desc>
         <p>Sends data to a port. <c>port_command(Port, Data, [])</c>
 	  equals <c>port_command(Port, Data)</c>.</p>
-	<p>If the port command is aborted <c>false</c> is returned;
-	   otherwise, <c>true</c> is returned.</p>
-	<p>If the port is busy, the calling process will be suspended
-	   until the port is not busy anymore.</p>
-        <p>Currently the following <c><anno>Option</anno></c>s are valid:</p>
+	<p>If the port command is aborted, <c>false</c> is returned,
+	   otherwise <c>true</c>.</p>
+	<p>If the port is busy, the calling process is suspended
+	   until the port is not busy any more.</p>
+        <p>The following <c><anno>Option</anno></c>s are valid:</p>
         <taglist>
           <tag><c>force</c></tag>
-          <item>The calling process will not be suspended if the port is
-	        busy; instead, the port command is forced through. The
-		call will fail with a <c>notsup</c> exception if the
+          <item>The calling process is not suspended if the port is
+	        busy, instead the port command is forced through. The
+		call fails with a <c>notsup</c> exception if the
 		driver of the port does not support this. For more
-		information see the
-		<seealso marker="driver_entry#driver_flags"><![CDATA[ERL_DRV_FLAG_SOFT_BUSY]]></seealso>
-		driver flag.
+		information, see driver flag
+		<seealso marker="driver_entry#driver_flags"><![CDATA[ERL_DRV_FLAG_SOFT_BUSY]]></seealso>.
           </item>
           <tag><c>nosuspend</c></tag>
-          <item>The calling process will not be suspended if the port is
-	        busy; instead, the port command is aborted and
+          <item>The calling process is not suspended if the port is
+	        busy, instead the port command is aborted and
 		<c>false</c> is returned. 
           </item>
         </taglist>
         <note>
-          <p>More options may be added in the future.</p>
+          <p>More options can be added in a future release.</p>
         </note>
         <p>Failures:</p>
         <taglist>
           <tag><c>badarg</c></tag>
           <item>
 	     If <c><anno>Port</anno></c> is not an identifier of an open
-	     port, or the registered name of an open port. If the calling
-	     process was linked to the previously open port identified by
-	     <c><anno>Port</anno></c>, an exit signal due to this link
-	     was received by the process prior to this exception.
+	     port, or the registered name of an open port. If the
+	     calling process was previously linked to the closed port,
+	     identified by <c><anno>Port</anno></c>, the exit signal
+	     from the port is guaranteed to be delivered before this
+	     <c>badarg</c> exception occurs.
 	  </item>
           <tag><c>badarg</c></tag>
           <item>
-	     If <c><anno>Data</anno></c> is not a valid io list.
+	     If <c><anno>Data</anno></c> is an invalid I/O list.
 	  </item>
           <tag><c>badarg</c></tag>
           <item>
-	    If <c><anno>OptionList</anno></c> is not a valid option list.
+	    If <c><anno>OptionList</anno></c> is an invalid option list.
 	  </item>
           <tag><c>notsup</c></tag>
           <item>
-	    If the <c>force</c> option has been passed, but the
+	    If option <c>force</c> has been passed, but the
 	    driver of the port does not allow forcing through
 	    a busy port.
 	  </item>
 	</taglist>
       </desc>
     </func>
+
     <func>
       <name name="port_connect" arity="2"/>
-      <fsummary>Set the owner of a port</fsummary>
+      <fsummary>Sets the owner of a port.</fsummary>
       <desc>
         <p>Sets the port owner (the connected port) to <c><anno>Pid</anno></c>.
-          Roughly the same as <c><anno>Port</anno> ! {Owner, {connect, <anno>Pid</anno>}}</c>
+          Roughly the same as
+          <c><anno>Port</anno> ! {Owner, {connect, <anno>Pid</anno>}}</c>
           except for the following:</p>
         <list type="bulleted">
           <item>
-            <p>The error behavior differs, see below.</p>
+            <p>The error behavior differs, see the following.</p>
           </item>
           <item>
             <p>The port does <em>not</em> reply with
               <c>{Port,connected}</c>.</p>
           </item>
           <item>
-            <p><c>port_connect/1</c> is synchronous, see below.</p>
+            <p><c>port_connect/1</c> is synchronous, see the following.</p>
           </item>
           <item>
             <p>The new port owner gets linked to the port.</p>
           </item>
         </list>
-        <p>The old port owner stays linked to the port and have to call
-          <c>unlink(Port)</c> if this is not desired. Any process may
+        <p>The old port owner stays linked to the port and must call
+          <c>unlink(Port)</c> if this is not desired. Any process can
           set the port owner to be any process with
           <c>port_connect/2</c>.</p>
-        <p>For comparison: <c><anno>Port</anno> ! {self(), {connect, <anno>Pid</anno>}}</c> fails
-          with <c>badarg</c> if <c><anno>Port</anno></c> cannot be sent to (i.e.,
-          <c><anno>Port</anno></c> refers neither to a port nor to a process). If
-          <c><anno>Port</anno></c> is a closed port nothing happens. If <c><anno>Port</anno></c>
+        <p>For comparison:
+          <c><anno>Port</anno> ! {self(), {connect, <anno>Pid</anno>}}</c>
+	  only fails with <c>badarg</c> if <c><anno>Port</anno></c>
+          does not refer to a port or a process. If
+          <c><anno>Port</anno></c> is a closed port, nothing happens.
+          If <c><anno>Port</anno></c>
           is an open port and the calling process is the port owner,
           the port replies with <c>{Port, connected}</c> to the old
-          port owner. Note that the old port owner is still linked to
-          the port, and that the new is not. If <c><anno>Port</anno></c> is an open
+          port owner. Notice that the old port owner is still linked to
+          the port, while the new is not. If <c><anno>Port</anno></c> is an open
           port and the calling process is not the port owner,
           the <em>port owner</em> fails with <c>badsig</c>. The port
           owner fails with <c>badsig</c> also if <c><anno>Pid</anno></c> is not an
-          existing local pid.</p>
-        <p>Note that any process can set the port owner using
-          <c><anno>Port</anno> ! {PortOwner, {connect, <anno>Pid</anno>}}</c> just as if it
-          itself was the port owner, but the reply always goes to
+          existing local process identifier.</p>
+        <p>Notice that any process can set the port owner using
+          <c><anno>Port</anno> ! {PortOwner, {connect, <anno>Pid</anno>}}</c>
+          as if it itself was the port owner, but the reply always goes to
           the port owner.</p>
-	<p>As of OTP-R16 <c><anno>Port</anno> ! {PortOwner, {connect, <anno>Pid</anno>}}</c> is
-	   truly asynchronous. Note that this operation has always been
+	<p>As from OTP-R16,
+           <c><anno>Port</anno> ! {PortOwner, {connect, <anno>Pid</anno>}}</c> is
+	   truly asynchronous. Notice that this operation has always been
 	   documented as an asynchronous operation, while the underlying
 	   implementation has been synchronous. <c>port_connect/2</c> is
-	   however still fully synchronous. This due to its error
+	   however still fully synchronous. This because of its error
 	   behavior.</p>
         <p>Failures:</p>
         <taglist>
           <tag><c>badarg</c></tag>
           <item>
-	     If <c><anno>Port</anno></c> is not an identifier of an open
-	     port, or the registered name of an open port. If the calling
-	     process was linked to the previously open port identified by
-	     <c><anno>Port</anno></c>, an exit signal due to this link
-	     was received by the process prior to this exception.
+	     If <c><anno>Port</anno></c> is not an identifier of an open port, or
+	     the registered name of an open port. If the calling
+	     process was previously linked to the closed port,
+	     identified by <c><anno>Port</anno></c>, the exit signal
+	     from the port is guaranteed to be delivered before this
+	     <c>badarg</c> exception occurs.
 	  </item>
           <tag><c>badarg</c></tag>
 	  <item>If process identified by <c>Pid</c> is not an existing
@@ -3669,53 +3857,75 @@ os_prompt% </pre>
 	</taglist>
       </desc>
     </func>
+
     <func>
       <name name="port_control" arity="3"/>
-      <fsummary>Perform a synchronous control operation on a port</fsummary>
+      <fsummary>Performs a synchronous control operation on a port.</fsummary>
       <desc>
         <p>Performs a synchronous control operation on a port.
-          The meaning of <c><anno>Operation</anno></c> and <c><anno>Data</anno></c> depends on
-          the port, i.e., on the port driver. Not all port drivers
+          The meaning of <c><anno>Operation</anno></c> and
+          <c><anno>Data</anno></c> depends on
+          the port, that is, on the port driver. Not all port drivers
           support this control feature.</p>
-        <p>Returns: a list of integers in the range 0 through 255, or a
+        <p>Returns a list of integers in the range 0..255, or a
           binary, depending on the port driver. The meaning of
           the returned data also depends on the port driver.</p>
-        <p>Failure: <c>badarg</c> if <c><anno>Port</anno></c> is not an open port or
-          the registered name of an open port, if <c><anno>Operation</anno></c>
-          cannot fit in a 32-bit integer, if the port driver does not
-          support synchronous control operations, or if the port driver
-          so decides for any reason (probably something wrong with
-          <c><anno>Operation</anno></c> or <c><anno>Data</anno></c>).</p>
+        <p>Failures:</p>
+        <taglist>
+          <tag><c>badarg</c></tag>
+          <item>
+	     If <c><anno>Port</anno></c> is not an open port or the registered
+             name of an open port.
+	  </item>
+          <tag><c>badarg</c></tag>
+          <item>
+	     If <c><anno>Operation</anno></c> cannot fit in a 32-bit integer.
+	  </item>
+          <tag><c>badarg</c></tag>
+          <item>
+	    If the port driver does not support synchronous control
+            operations.
+	  </item>
+          <tag><c>badarg</c></tag>
+          <item>
+	    If the port driver so decides for any reason (probably
+            something wrong with <c><anno>Operation</anno></c> or
+            <c><anno>Data</anno></c>).
+	  </item>
+	</taglist>
       </desc>
     </func>
+
     <func>
       <name name="port_call" arity="3"/>
-      <fsummary>Synchronous call to a port with term data</fsummary>
+      <fsummary>Performs a synchronous call to a port with term data.</fsummary>
       <desc>
         <p>Performs a synchronous call to a port. The meaning of
-          <c><anno>Operation</anno></c> and <c><anno>Data</anno></c> depends on the port, i.e.,
+          <c><anno>Operation</anno></c> and <c><anno>Data</anno></c>
+          depends on the port, that is,
           on the port driver. Not all port drivers support this feature.</p>
-        <p><c><anno>Port</anno></c> is a port identifier, referring to a driver.</p>
+        <p><c><anno>Port</anno></c> is a port identifier,
+          referring to a driver.</p>
         <p><c><anno>Operation</anno></c> is an integer, which is passed on to
           the driver.</p>
-        <p><c><anno>Data</anno></c> is any Erlang term. This data is converted to
-          binary term format and sent to the port.</p>
-        <p>Returns: a term from the driver. The meaning of the returned
+        <p><c><anno>Data</anno></c> is any Erlang term. This data is converted
+          to binary term format and sent to the port.</p>
+        <p>Returns a term from the driver. The meaning of the returned
           data also depends on the port driver.</p>
         <p>Failures:</p>
         <taglist>
           <tag><c>badarg</c></tag>
           <item>
-	     If <c><anno>Port</anno></c> is not an identifier of an open
-	     port, or the registered name of an open port. If the calling
-	     process was linked to the previously open port identified by
-	     <c><anno>Port</anno></c>, an exit signal due to this link
-	     was received by the process prior to this exception.
+	     If <c><anno>Port</anno></c> is not an identifier of an open port,
+	     or the registered name of an open port. If the calling
+	     process was previously linked to the closed port,
+	     identified by <c><anno>Port</anno></c>, the exit signal
+	     from the port is guaranteed to be delivered before this
+	     <c>badarg</c> exception occurs.
 	  </item>
           <tag><c>badarg</c></tag>
 	  <item>
-	     If <c><anno>Operation</anno></c> does not fit in a
-	     32-bit integer.
+	     If <c><anno>Operation</anno></c> does not fit in a 32-bit integer.
 	  </item>
           <tag><c>badarg</c></tag>
 	  <item>
@@ -3725,171 +3935,183 @@ os_prompt% </pre>
           <tag><c>badarg</c></tag>
 	  <item>
 	     If the port driver so decides for any reason (probably
-	     something wrong with <c><anno>Operation</anno></c>, or
-	     <c><anno>Data</anno></c>).
+	     something wrong with <c><anno>Operation</anno></c>
+             or <c><anno>Data</anno></c>).
 	  </item>
 	</taglist>
       </desc>
     </func>
+
     <func>
       <name name="port_info" arity="1"/>
-      <fsummary>Information about a port</fsummary>
+      <fsummary>Information about a port.</fsummary>
       <desc>
         <p>Returns a list containing tuples with information about
-          the <c><anno>Port</anno></c>, or <c>undefined</c> if the port is not open.
-          The order of the tuples is not defined, nor are all the
-          tuples mandatory.
-	  If <c>undefined</c> is returned and the calling process
-	  was linked to a previously open port identified by
-	  <c><anno>Port</anno></c>, an exit signal due to this link
-	  was received by the process prior to the return from
-	  <c>port_info/1</c>.</p>
-	<p>Currently the result will containt information about the
-	following <c>Item</c>s: <c>registered_name</c> (if the port has
-	a registered name), <c>id</c>, <c>connected</c>, <c>links</c>,
-	<c>name</c>, <c>input</c>, and <c>output</c>. For more information
-	about the different <c>Item</c>s, see
+          <c><anno>Port</anno></c>, or <c>undefined</c> if the port is not open.
+          The order of the tuples is undefined, and all the
+          tuples are not mandatory.
+	  If the port is closed and the calling process
+	  was previously linked to the port, the exit signal from the
+	  port is guaranteed to be delivered before <c>port_info/1</c>
+	  returns <c>undefined</c>.</p>
+	<p>The result contains information about the following
+          <c>Item</c>s:</p>
+        <list type="bulleted">
+          <item><c>registered_name</c> (if the port has a registered
+            name)</item>
+          <item><c>id</c></item>
+          <item><c>connected</c></item>
+          <item><c>links</c></item>
+          <item><c>name</c></item>
+          <item><c>input</c></item>
+          <item><c>output</c></item>
+        </list>
+        <p>For more information about the different <c>Item</c>s, see
 	<seealso marker="#port_info/2">port_info/2</seealso>.</p>
         <p>Failure: <c>badarg</c> if <c>Port</c> is not a local port
 	identifier, or an atom.</p>
       </desc>
     </func>
+
     <func>
       <name name="port_info" arity="2" clause_i="1"/>
-      <fsummary>Information about the connected process of a port</fsummary>
+      <fsummary>Information about the connected process of a port.</fsummary>
       <desc>
 	<p><c><anno>Pid</anno></c> is the process identifier of the process
 	connected to the port.</p>
 	<p>If the port identified by <c><anno>Port</anno></c> is not open,
-	<c>undefined</c> is returned. If <c>undefined</c> is returned and
-	the calling process was linked to a previously open port identified
-	by <c><anno>Port</anno></c>, an exit signal due to this link
-	was received by the process prior to the return from
-	<c>port_info/2</c>.</p>
+	<c>undefined</c> is returned. If the port is closed and the
+	calling process was previously linked to the port, the exit
+	signal from the port is guaranteed to be delivered before
+	<c>port_info/2</c> returns <c>undefined</c>.</p>
         <p>Failure: <c>badarg</c> if <c><anno>Port</anno></c> is not a local
 	port identifier, or an atom.</p>
       </desc>
     </func>
+
     <func>
       <name name="port_info" arity="2" clause_i="2"/>
-      <fsummary>Information about the internal index of a port</fsummary>
+      <fsummary>Information about the internal index of a port.</fsummary>
       <desc>
 	<p><c><anno>Index</anno></c> is the internal index of the port. This
-	index may be used to separate ports.</p>
+	index can be used to separate ports.</p>
 	<p>If the port identified by <c><anno>Port</anno></c> is not open,
-	<c>undefined</c> is returned. If <c>undefined</c> is returned and
-	the calling process was linked to a previously open port identified
-	by <c><anno>Port</anno></c>, an exit signal due to this link
-	was received by the process prior to the return from
-	<c>port_info/2</c>.</p>
+	<c>undefined</c> is returned. If the port is closed and the
+	calling process was previously linked to the port, the exit
+	signal from the port is guaranteed to be delivered before
+	<c>port_info/2</c> returns <c>undefined</c>.</p>
         <p>Failure: <c>badarg</c> if <c><anno>Port</anno></c> is not a local
 	port identifier, or an atom.</p>
       </desc>
     </func>
+
     <func>
       <name name="port_info" arity="2" clause_i="3"/>
-      <fsummary>Information about the input of a port</fsummary>
+      <fsummary>Information about the input of a port.</fsummary>
       <desc>
 	<p><c><anno>Bytes</anno></c> is the total number of bytes
 	read from the port.</p>
 	<p>If the port identified by <c><anno>Port</anno></c> is not open,
-	<c>undefined</c> is returned. If <c>undefined</c> is returned and
-	the calling process was linked to a previously open port identified
-	by <c><anno>Port</anno></c>, an exit signal due to this link
-	was received by the process prior to the return from
-	<c>port_info/2</c>.</p>
+	<c>undefined</c> is returned. If the port is closed and the
+	calling process was previously linked to the port, the exit
+	signal from the port is guaranteed to be delivered before
+	<c>port_info/2</c> returns <c>undefined</c>.</p>
         <p>Failure: <c>badarg</c> if <c><anno>Port</anno></c> is not a local
 	port identifier, or an atom.</p>
       </desc>
     </func>
+
     <func>
       <name name="port_info" arity="2" clause_i="4"/>
-      <fsummary>Information about the links of a port</fsummary>
+      <fsummary>Information about the links of a port.</fsummary>
       <desc>
 	<p><c><anno>Pids</anno></c> is a list of the process identifiers
 	of the processes that the port is linked to.</p>
 	<p>If the port identified by <c><anno>Port</anno></c> is not open,
-	<c>undefined</c> is returned. If <c>undefined</c> is returned and
-	the calling process was linked to a previously open port identified
-	by <c><anno>Port</anno></c>, an exit signal due to this link
-	was received by the process prior to the return from
-	<c>port_info/2</c>.</p>
+	<c>undefined</c> is returned. If the port is closed and the
+	calling process was previously linked to the port, the exit
+	signal from the port is guaranteed to be delivered before
+	<c>port_info/2</c> returns <c>undefined</c>.</p>
         <p>Failure: <c>badarg</c> if <c><anno>Port</anno></c> is not a local
 	port identifier, or an atom.</p>
       </desc>
     </func>
+
     <func>
       <name name="port_info" arity="2" clause_i="5"/>
-      <fsummary>Information about the locking of a port</fsummary>
+      <fsummary>Information about the locking of a port.</fsummary>
       <desc>
-	<p><c><anno>Locking</anno></c> is currently either <c>false</c>
-	(emulator without SMP support), <c>port_level</c> (port specific
-	locking), or <c>driver_level</c> (driver specific locking). Note
-	that these results are highly implementation specific and might
-	change in the future.</p>
+	<p><c><anno>Locking</anno></c> is one of the following:</p>
+        <list type="bulleted">
+          <item><c>false</c> (emulator without SMP support)</item>
+          <item><c>port_level</c> (port-specific locking)</item>
+          <item><c>driver_level</c> (driver-specific locking)</item>
+        </list>
+        <p>Notice that these results are highly implementation-specific
+          and can change in a future release.</p>
 	<p>If the port identified by <c><anno>Port</anno></c> is not open,
-	<c>undefined</c> is returned. If <c>undefined</c> is returned and
-	the calling process was linked to a previously open port identified
-	by <c><anno>Port</anno></c>, an exit signal due to this link
-	was received by the process prior to the return from
-	<c>port_info/2</c>.</p>
+	<c>undefined</c> is returned. If the port is closed and the
+	calling process was previously linked to the port, the exit
+	signal from the port is guaranteed to be delivered before
+	<c>port_info/2</c> returns <c>undefined</c>.</p>
         <p>Failure: <c>badarg</c> if <c><anno>Port</anno></c> is not a local
 	port identifier, or an atom.</p>
       </desc>
     </func>
+
     <func>
       <name name="port_info" arity="2" clause_i="6"/>
-      <fsummary>Information about the memory size of a port</fsummary>
+      <fsummary>Information about the memory size of a port.</fsummary>
       <desc>
-	<p><c><anno>Bytes</anno></c> is the total amount of memory,
-	in bytes, allocated for this port by the runtime system. Note
-	that the port itself might have allocated memory which is not
+	<p><c><anno>Bytes</anno></c> is the total number of
+	bytes allocated for this port by the runtime system. The
+	port itself can have allocated memory that is not
 	included in <c><anno>Bytes</anno></c>.</p>
 	<p>If the port identified by <c><anno>Port</anno></c> is not open,
-	<c>undefined</c> is returned. If <c>undefined</c> is returned and
-	the calling process was linked to a previously open port identified
-	by <c><anno>Port</anno></c>, an exit signal due to this link
-	was received by the process prior to the return from
-	<c>port_info/2</c>.</p>
+	<c>undefined</c> is returned. If the port is closed and the
+	calling process was previously linked to the port, the exit
+	signal from the port is guaranteed to be delivered before
+	<c>port_info/2</c> returns <c>undefined</c>.</p>
         <p>Failure: <c>badarg</c> if <c><anno>Port</anno></c> is not a local
 	port identifier, or an atom.</p>
       </desc>
     </func>
+
     <func>
       <name name="port_info" arity="2" clause_i="7"/>
-      <fsummary>Information about the monitors of a port</fsummary>
+      <fsummary>Information about the monitors of a port.</fsummary>
       <desc>
 	<p><c><anno>Monitors</anno></c> represent processes that this port
-	is monitoring.</p>
+	monitors.</p>
 	<p>If the port identified by <c><anno>Port</anno></c> is not open,
-	<c>undefined</c> is returned. If <c>undefined</c> is returned and
-	the calling process was linked to a previously open port identified
-	by <c><anno>Port</anno></c>, an exit signal due to this link
-	was received by the process prior to the return from
-	<c>port_info/2</c>.</p>
+	<c>undefined</c> is returned. If the port is closed and the
+	calling process was previously linked to the port, the exit
+	signal from the port is guaranteed to be delivered before
+	<c>port_info/2</c> returns <c>undefined</c>.</p>
         <p>Failure: <c>badarg</c> if <c><anno>Port</anno></c> is not a local
 	port identifier, or an atom.</p>
       </desc>
     </func>
+
     <func>
       <name name="port_info" arity="2" clause_i="8"/>
-      <fsummary>Information about the name of a port</fsummary>
+      <fsummary>Information about the name of a port.</fsummary>
       <desc>
 	<p><c><anno>Name</anno></c> is the command name set by
 	<seealso marker="#open_port/2">open_port/2</seealso>.</p>
 	<p>If the port identified by <c><anno>Port</anno></c> is not open,
-	<c>undefined</c> is returned. If <c>undefined</c> is returned and
-	the calling process was linked to a previously open port identified
-	by <c><anno>Port</anno></c>, an exit signal due to this link
-	was received by the process prior to the return from
-	<c>port_info/2</c>.</p>
+	<c>undefined</c> is returned. If the port is closed and the
+	calling process was previously linked to the port, the exit
+	signal from the port is guaranteed to be delivered before
+	<c>port_info/2</c> returns <c>undefined</c>.</p>
         <p>Failure: <c>badarg</c> if <c><anno>Port</anno></c> is not a local
 	port identifier, or an atom.</p>
       </desc>
     </func>
+
     <func>
       <name name="port_info" arity="2" clause_i="9"/>
-      <fsummary>Information about the OS pid of a port</fsummary>
+      <fsummary>Information about the OS pid of a port.</fsummary>
       <desc>
 	<p><c><anno>OsPid</anno></c> is the process identifier (or equivalent)
 	of an OS process created with
@@ -3897,432 +4119,516 @@ os_prompt% </pre>
 	Command}, Options)</seealso>. If the port is not the result of spawning
 	an OS process, the value is <c>undefined</c>.</p>
 	<p>If the port identified by <c><anno>Port</anno></c> is not open,
-	<c>undefined</c> is returned. If <c>undefined</c> is returned and
-	the calling process was linked to a previously open port identified
-	by <c><anno>Port</anno></c>, an exit signal due to this link
-	was received by the process prior to the return from
-	<c>port_info/2</c>.</p>
+	<c>undefined</c> is returned. If the port is closed and the
+	calling process was previously linked to the port, the exit
+	signal from the port is guaranteed to be delivered before
+	<c>port_info/2</c> returns <c>undefined</c>.</p>
         <p>Failure: <c>badarg</c> if <c><anno>Port</anno></c> is not a local
 	port identifier, or an atom.</p>
       </desc>
     </func>
+
     <func>
       <name name="port_info" arity="2" clause_i="10"/>
-      <fsummary>Information about the output of a port</fsummary>
+      <fsummary>Information about the output of a port.</fsummary>
       <desc>
 	<p><c><anno>Bytes</anno></c> is the total number of bytes written
-	to the port from Erlang processes using either
+	to the port from Erlang processes using
 	<seealso marker="#port_command/2">port_command/2</seealso>,
 	<seealso marker="#port_command/3">port_command/3</seealso>,
-	or <c><anno>Port</anno> ! {Owner, {command, Data}</c>.
-	</p>
+	or <c><anno>Port</anno> ! {Owner, {command, Data}</c>.</p>
 	<p>If the port identified by <c><anno>Port</anno></c> is not open,
-	<c>undefined</c> is returned. If <c>undefined</c> is returned and
-	the calling process was linked to a previously open port identified
-	by <c><anno>Port</anno></c>, an exit signal due to this link
-	was received by the process prior to the return from
-	<c>port_info/2</c>.</p>
+	<c>undefined</c> is returned. If the port is closed and the
+	calling process was previously linked to the port, the exit
+	signal from the port is guaranteed to be delivered before
+	<c>port_info/2</c> returns <c>undefined</c>.</p>
         <p>Failure: <c>badarg</c> if <c><anno>Port</anno></c> is not a local
 	port identifier, or an atom.</p>
       </desc>
     </func>
+
     <func>
       <name name="port_info" arity="2" clause_i="11"/>
-      <fsummary>Information about the parallelism hint of a port</fsummary>
+      <fsummary>Information about the parallelism hint of a port.</fsummary>
       <desc>
 	<p><c><anno>Boolean</anno></c> corresponds to the port parallelism
-	hint being used by this port. For more information see
-	the <seealso marker="#open_port_parallelism">parallelism</seealso>
-	option of <seealso marker="#open_port/2">open_port/2</seealso>.</p>
+	hint being used by this port. For more information, see option
+	<seealso marker="#open_port_parallelism">parallelism</seealso>
+	of <seealso marker="#open_port/2">open_port/2</seealso>.</p>
       </desc>
     </func>
+
     <func>
       <name name="port_info" arity="2" clause_i="12"/>
-      <fsummary>Information about the queue size of a port</fsummary>
+      <fsummary>Information about the queue size of a port.</fsummary>
       <desc>
-	<p><c><anno>Bytes</anno></c> is the total amount of data,
-	in bytes, queued by the port using the ERTS driver queue
+	<p><c><anno>Bytes</anno></c> is the total number
+	of bytes queued by the port using the <c>ERTS</c> driver queue
 	implementation.</p>
 	<p>If the port identified by <c><anno>Port</anno></c> is not open,
-	<c>undefined</c> is returned. If <c>undefined</c> is returned and
-	the calling process was linked to a previously open port identified
-	by <c><anno>Port</anno></c>, an exit signal due to this link
-	was received by the process prior to the return from
-	<c>port_info/2</c>.</p>
+	<c>undefined</c> is returned. If the port is closed and the
+	calling process was previously linked to the port, the exit
+	signal from the port is guaranteed to be delivered before
+	<c>port_info/2</c> returns <c>undefined</c>.</p>
         <p>Failure: <c>badarg</c> if <c><anno>Port</anno></c> is not a local
 	port identifier, or an atom.</p>
       </desc>
     </func>
+
     <func>
       <name name="port_info" arity="2" clause_i="13"/>
-      <fsummary>Information about the registered name of a port</fsummary>
+      <fsummary>Information about the registered name of a port.</fsummary>
       <desc>
 	<p><c><anno>RegisteredName</anno></c> is the registered name of
 	the port. If the port has no registered name, <c>[]</c> is returned.</p>
 	<p>If the port identified by <c><anno>Port</anno></c> is not open,
-	<c>undefined</c> is returned. If <c>undefined</c> is returned and
-	the calling process was linked to a previously open port identified
-	by <c><anno>Port</anno></c>, an exit signal due to this link
-	was received by the process prior to the return from
-	<c>port_info/2</c>.</p>
+	<c>undefined</c> is returned. If the port is closed and the
+	calling process was previously linked to the port, the exit
+	signal from the port is guaranteed to be delivered before
+	<c>port_info/2</c> returns <c>undefined</c>.</p>
         <p>Failure: <c>badarg</c> if <c><anno>Port</anno></c> is not a local
 	port identifier, or an atom.</p>
       </desc>
     </func>
+
     <func>
       <name name="port_to_list" arity="1"/>
-      <fsummary>Text representation of a port identifier</fsummary>
+      <fsummary>Text representation of a port identifier.</fsummary>
       <desc>
-        <p>Returns a string which corresponds to the text
+        <p>Returns a string corresponding to the text
           representation of the port identifier <c><anno>Port</anno></c>.</p>
         <warning>
-          <p>This BIF is intended for debugging and for use in
-            the Erlang operating system. It should not be used in
-            application programs.</p>
+          <p>This BIF is intended for debugging. It is not to be used
+	  in application programs.</p>
         </warning>
       </desc>
     </func>
+
     <func>
       <name name="ports" arity="0"/>
-      <fsummary>All open ports</fsummary>
+      <fsummary>Lists all existing ports.</fsummary>
       <desc>
 	<p>Returns a list of port identifiers corresponding to all the
-	ports currently existing on the local node.</p>
-
-	<p>Note that a port that is exiting, exists but is not open.</p>
+	ports existing on the local node.</p>
+	<p>Notice that an exiting port exists, but is not open.</p>
       </desc>
     </func>
+
     <func>
       <name name="pre_loaded" arity="0"/>
-      <fsummary>List of all pre-loaded modules</fsummary>
+      <fsummary>Lists all pre-loaded modules.</fsummary>
       <desc>
-        <p>Returns a list of Erlang modules which are pre-loaded in
+        <p>Returns a list of Erlang modules that are preloaded in
           the system. As all loading of code is done through the file
           system, the file system must have been loaded previously.
-          Hence, at least the module <c>init</c> must be pre-loaded.</p>
+          Hence, at least the module <c>init</c> must be preloaded.</p>
       </desc>
     </func>
+
     <func>
       <name name="process_display" arity="2"/>
-      <fsummary>Write information about a local process on standard error</fsummary>
+      <fsummary>Writes information about a local process on standard error.</fsummary>
       <desc>
         <p>Writes information about the local process <c><anno>Pid</anno></c> on
-          standard error. The currently allowed value for the atom
+          standard error. The only allowed value for the atom
           <c><anno>Type</anno></c> is <c>backtrace</c>, which shows the contents of
           the call stack, including information about the call chain, with
           the current function printed first. The format of the output
           is not further defined.</p>
       </desc>
     </func>
+
     <func>
       <name name="process_flag" arity="2" clause_i="1"/>
-      <fsummary>Set process flag trap_exit for the calling process</fsummary>
+      <fsummary>Sets process flag <c>trap_exit</c> for the calling process.</fsummary>
       <desc>
         <p>When <c>trap_exit</c> is set to <c>true</c>, exit signals
-          arriving to a process are converted to <c>{'EXIT', From, Reason}</c> messages, which can be received as ordinary
+          arriving to a process are converted to <c>{'EXIT', From, Reason}</c>
+          messages, which can be received as ordinary
           messages. If <c>trap_exit</c> is set to <c>false</c>, the
           process exits if it receives an exit signal other than
           <c>normal</c> and the exit signal is propagated to its
-          linked processes. Application processes should normally
-          not trap exits.</p>
+          linked processes. Application processes are normally
+          not to trap exits.</p>
         <p>Returns the old value of the flag.</p>
         <p>See also <seealso marker="#exit/2">exit/2</seealso>.</p>
       </desc>
     </func>
+
     <func>
       <name name="process_flag" arity="2" clause_i="2"/>
-      <fsummary>Set process flag error_handler for the calling process</fsummary>
+      <fsummary>Sets process flag <c>error_handler</c> for the calling process.</fsummary>
       <desc>
-        <p>This is used by a process to redefine the error handler
+        <p>Used by a process to redefine the error handler
           for undefined function calls and undefined registered
-          processes.  Inexperienced users should not use this flag
-          since code auto-loading is dependent on the correct
+          processes. Inexperienced users are not to use this flag,
+          as code auto-loading depends on the correct
           operation of the error handling module.</p>
         <p>Returns the old value of the flag.</p>
       </desc>
     </func>
+
     <func>
       <name name="process_flag" arity="2" clause_i="3"/>
-      <fsummary>Set process flag min_heap_size for the calling process</fsummary>
+      <fsummary>Sets process flag <c>min_heap_size</c> for the calling process.</fsummary>
       <desc>
-        <p>This changes the minimum heap size for the calling
-          process.</p>
+        <p>Changes the minimum heap size for the calling process.</p>
         <p>Returns the old value of the flag.</p>
       </desc>
     </func>
+
     <func>
       <name name="process_flag" arity="2" clause_i="4"/>
-      <fsummary>Set process flag min_bin_vheap_size for the calling process</fsummary>
+      <fsummary>Sets process flag <c>min_bin_vheap_size</c> for the calling process.</fsummary>
       <desc>
-        <p>This changes the minimum binary virtual heap size for the calling
+        <p>Changes the minimum binary virtual heap size for the calling
           process.</p>
-        <p>Returns the old value of the flag.</p>      </desc>
+        <p>Returns the old value of the flag.</p>
+      </desc>
     </func>
+    <marker id="process_flag_message_queue_data"/>
     <func>
       <name name="process_flag" arity="2" clause_i="5"/>
+      <fsummary>Set process flag <c>message_queue_data</c> for the calling process</fsummary>
+      <type name="message_queue_data"/>
+      <desc>
+	<p>This flag determines how messages in the message queue
+	are stored. When the flag is:</p>
+        <taglist>
+          <tag><c>off_heap</c></tag>
+	  <item><p>
+	    <em>All</em> messages in the message queue will be stored
+	    outside of the process heap. This implies that <em>no</em>
+	    messages in the message queue will be part of a garbage
+	    collection of the process.
+	  </p></item>
+          <tag><c>on_heap</c></tag>
+	  <item><p>
+	    All messages in the message queue will eventually be
+	    placed on heap. They may however temporarily be stored
+	    off heap. This is how messages always have been stored
+	    up until ERTS version 8.0.
+	  </p></item>
+          <tag><c>mixed</c></tag>
+	  <item><p>
+	    Messages may be placed either on the heap or outside
+	    of the heap.
+	  </p></item>
+        </taglist>
+	<p>
+	  The default <c>message_queue_data</c> process flag is determined
+	  by the <seealso marker="erl#+xmqd"><c>+xmqd</c></seealso>
+	  <c>erl</c> command line argument.
+	</p>
+	<p>
+	  If the process potentially may get a hugh amount of messages,
+	  you are recommended to set the flag to <c>off_heap</c>. This
+	  since a garbage collection with lots of messages placed on
+	  the heap may become extremly expensive and the process may
+	  consume large amounts of memory. Performance of the
+	  actual message passing is however generally better when not
+	  using the <c>off_heap</c> flag.
+	</p>
+	<p>
+	  When changing this flag messages will be moved. This work
+	  has been initiated but not completed when this function
+	  call returns.
+	</p>
+        <p>Returns the old value of the flag.</p>
+      </desc>
+    </func>
+    <func>
+      <name name="process_flag" arity="2" clause_i="6"/>
+      <fsummary>Sets process flag <c>priority</c> for the calling process.</fsummary>
       <type name="priority_level"/>
-      <fsummary>Set process flag priority for the calling process</fsummary>
       <desc>
         <p><marker id="process_flag_priority"></marker>
-          This sets the process priority. <c><anno>Level</anno></c> is an atom.
-          There are currently four priority levels: <c>low</c>,
-          <c>normal</c>, <c>high</c>, and <c>max</c>. The default
-          priority level is <c>normal</c>. <em>NOTE</em>: The
-          <c>max</c> priority level is reserved for internal use in
-          the Erlang runtime system, and should <em>not</em> be used
-          by others.
-        </p>
-        <p>Internally in each priority level processes are scheduled
-          in a round robin fashion.
-        </p>
+          Sets the process priority. <c><anno>Level</anno></c> is an atom.
+          There are four priority levels: <c>low</c>,
+          <c>normal</c>, <c>high</c>, and <c>max</c>. Default
+          is <c>normal</c>.</p>
+        <note>
+          <p>Priority level <c>max</c> is reserved for internal use in
+          the Erlang runtime system, and is <em>not</em> to be used
+          by others.</p>
+        </note>
+        <p>Internally in each priority level, processes are scheduled
+          in a round robin fashion.</p>
         <p>Execution of processes on priority <c>normal</c> and
-          priority <c>low</c> will be interleaved. Processes on
-          priority <c>low</c> will be selected for execution less
-          frequently than processes on priority <c>normal</c>.
-        </p>
-        <p>When there are runnable processes on priority <c>high</c>
-          no processes on priority <c>low</c>, or <c>normal</c> will
-          be selected for execution. Note, however, that this does
-          <em>not</em> mean that no processes on priority <c>low</c>,
-          or <c>normal</c> will be able to run when there are
-          processes on priority <c>high</c> running. On the runtime
-          system with SMP support there might be more processes running
-          in parallel than processes on priority <c>high</c>, i.e.,
-          a <c>low</c>, and a <c>high</c> priority process might
-          execute at the same time.
-        </p>
-        <p>When there are runnable processes on priority <c>max</c>
+          <c>low</c> are interleaved. Processes on priority
+          <c>low</c> are selected for execution less
+          frequently than processes on priority <c>normal</c>.</p>
+        <p>When there are runnable processes on priority <c>high</c>,
+          no processes on priority <c>low</c> or <c>normal</c> are
+          selected for execution. Notice however, that this does
+          <em>not</em> mean that no processes on priority <c>low</c>
+          or <c>normal</c> can run when there are processes
+          running on priority <c>high</c>. On the runtime
+          system with SMP support, more processes can be running
+          in parallel than processes on priority <c>high</c>, that is,
+          a <c>low</c> and a <c>high</c> priority process can
+          execute at the same time.</p>
+        <p>When there are runnable processes on priority <c>max</c>,
           no processes on priority <c>low</c>, <c>normal</c>, or
-          <c>high</c> will be selected for execution. As with the
-          <c>high</c> priority, processes on lower priorities might
-          execute in parallel with processes on priority <c>max</c>.
-        </p>
+          <c>high</c> are selected for execution. As with priority
+          <c>high</c>, processes on lower priorities can
+          execute in parallel with processes on priority <c>max</c>.</p>
         <p>Scheduling is preemptive. Regardless of priority, a process
-          is preempted when it has consumed more than a certain amount
+          is preempted when it has consumed more than a certain number
           of reductions since the last time it was selected for
-          execution.
-        </p>
-        <p><em>NOTE</em>: You should not depend on the scheduling
+          execution.</p>
+        <note>
+          <p>Do not depend on the scheduling
           to remain exactly as it is today. Scheduling, at least on
-          the runtime system with SMP support, is very likely to be
-          modified in the future in order to better utilize available
-          processor cores.
-        </p>
-        <p>There is currently <em>no</em> automatic mechanism for
-          avoiding priority inversion, such as priority inheritance,
-          or priority ceilings. When using priorities you have
-          to take this into account and handle such scenarios by
-          yourself.
-        </p>
+          the runtime system with SMP support, is likely to be
+          changed in a future release to use available
+          processor cores better.</p>
+        </note>
+        <p>There is <em>no</em> automatic mechanism for
+          avoiding priority inversion, such as priority inheritance
+          or priority ceilings. When using priorities,
+          take this into account and handle such scenarios by
+          yourself.</p>
         <p>Making calls from a <c>high</c> priority process into code
-          that you don't have control over may cause the <c>high</c>
-          priority process to wait for a processes with lower
-          priority, i.e., effectively decreasing the priority of the
+          that you have no control over can cause the <c>high</c>
+          priority process to wait for a process with lower
+          priority. That is, effectively decreasing the priority of the
           <c>high</c> priority process during the call. Even if this
-          isn't the case with one version of the code that you don't
-          have under your control, it might be the case in a future
-          version of it. This might, for example, happen if a
-          <c>high</c> priority process triggers code loading, since
-          the code server runs on priority <c>normal</c>.
-        </p>
+          is not the case with one version of the code that you have no
+          control over, it can be the case in a future
+          version of it. This can, for example, occur if a
+          <c>high</c> priority process triggers code loading, as
+          the code server runs on priority <c>normal</c>.</p>
         <p>Other priorities than <c>normal</c> are normally not needed.
-          When other priorities are used, they need to be used
-          with care, especially the <c>high</c> priority <em>must</em>
-          be used with care. A process on <c>high</c> priority should
-          only perform work for short periods of time. Busy looping for
-          long periods of time in a <c>high</c> priority process will
-          most likely cause problems, since there are important servers
-          in OTP running on priority <c>normal</c>.
-        </p>
+          When other priorities are used, use them with care,
+          <em>especially</em> priority <c>high</c>. A
+          process on priority <c>high</c> is only
+          to perform work for short periods. Busy looping for
+          long periods in a <c>high</c> priority process does
+          most likely cause problems, as important OTP servers
+          run on priority <c>normal</c>.</p>
         <p>Returns the old value of the flag.</p>
       </desc>
     </func>
+
     <func>
-      <name name="process_flag" arity="2" clause_i="6"/>
-      <fsummary>Set process flag save_calls for the calling process</fsummary>
+      <name name="process_flag" arity="2" clause_i="7"/>
+      <fsummary>Sets process flag <c>save_calls</c> for the calling process.</fsummary>
       <desc>
         <p><c><anno>N</anno></c> must be an integer in the interval 0..10000.
-          If <c><anno>N</anno></c> &gt; 0, call saving is made active for the
-          process, which means that information about the <c><anno>N</anno></c>
-          most recent global function calls, BIF calls, sends and
+          If <c><anno>N</anno></c> is greater than 0, call saving is made
+          active for the
+          process. This means that information about the <c><anno>N</anno></c>
+          most recent global function calls, BIF calls, sends, and
           receives made by the process are saved in a list, which
           can be retrieved with
           <c>process_info(Pid, last_calls)</c>. A global function
           call is one in which the module of the function is
           explicitly mentioned. Only a fixed amount of information
-          is saved: a tuple <c>{Module, Function, Arity}</c> for
-          function calls, and the mere atoms <c>send</c>,
-          <c>'receive'</c> and <c>timeout</c> for sends and receives
-          (<c>'receive'</c> when a message is received and
-          <c>timeout</c> when a receive times out). If <c>N</c> = 0,
+          is saved, as follows:</p>
+        <list type="bulleted">
+          <item>A tuple <c>{Module, Function, Arity}</c> for
+            function calls</item>
+          <item> The atoms <c>send</c>, <c>'receive'</c>, and
+            <c>timeout</c> for sends and receives (<c>'receive'</c>
+            when a message is received and <c>timeout</c> when a
+            receive times out)</item>
+        </list>
+        <p>If <c>N</c> = 0,
           call saving is disabled for the process, which is the
           default. Whenever the size of the call saving list is set,
           its contents are reset.</p>
         <p>Returns the old value of the flag.</p>
       </desc>
     </func>
+
     <func>
-      <name name="process_flag" arity="2" clause_i="7"/>
-      <fsummary>Set process flag sensitive for the calling process</fsummary>
+      <name name="process_flag" arity="2" clause_i="8"/>
+      <fsummary>Sets process flag <c>sensitive</c> for the calling process.</fsummary>
       <desc>
-        <p>Set or clear the <c>sensitive</c> flag for the current process.
+        <p>Sets or clears flag <c>sensitive</c> for the current process.
           When a process has been marked as sensitive by calling
-          <c>process_flag(sensitive, true)</c>, features in the run-time
-          system that can be used for examining the data and/or inner working
+          <c>process_flag(sensitive, true)</c>, features in the runtime
+          system that can be used for examining the data or inner working
           of the process are silently disabled.</p>
         <p>Features that are disabled include (but are not limited to)
           the following:</p>
-        <p>Tracing: Trace flags can still be set for the process, but no
-          trace messages of any kind will be generated.
-          (If the <c>sensitive</c> flag is turned off, trace messages will
-          again be generated if there are any trace flags set.)</p>
-        <p>Sequential tracing: The sequential trace token will be propagated
-          as usual, but no sequential trace messages will be generated.</p>
-        <p><c>process_info/1,2</c> cannot be used to read out the message
-          queue or the process dictionary (both will be returned as empty lists).</p>
+        <list type="bulleted">
+          <item>Tracing: Trace flags can still be set for the process,
+            but no trace messages of any kind are generated. (If flag
+            <c>sensitive</c> is turned off, trace messages are again
+            generated if any trace flags are set.)</item>
+          <item>Sequential tracing: The sequential trace token is
+            propagated as usual, but no sequential trace messages are
+            generated.</item>
+        </list>
+        <p><c>process_info/1,2</c> cannot be used to read out the
+          message queue or the process dictionary (both are returned
+          as empty lists).</p>
         <p>Stack back-traces cannot be displayed for the process.</p>
         <p>In crash dumps, the stack, messages, and the process dictionary
-          will be omitted.</p>
+          are omitted.</p>
         <p>If <c>{save_calls,N}</c> has been set for the process, no
-          function calls will be saved to the call saving list.
-          (The call saving list will not be cleared; furthermore, send, receive,
-          and timeout events will still be added to the list.)</p>
+          function calls are saved to the call saving list.
+          (The call saving list is not cleared. Furthermore, send, receive,
+          and timeout events are still added to the list.)</p>
         <p>Returns the old value of the flag.</p>
       </desc>
     </func>
+
     <func>
       <name name="process_flag" arity="3"/>
-      <fsummary>Set process flags for a process</fsummary>
+      <fsummary>Sets process flags for a process.</fsummary>
       <desc>
-        <p>Sets certain flags for the process <c><anno>Pid</anno></c>, in the same
-          manner as
+        <p>Sets certain flags for the process <c><anno>Pid</anno></c>,
+          in the same manner as
           <seealso marker="#process_flag/2">process_flag/2</seealso>.
-          Returns the old value of the flag. The allowed values for
+          Returns the old value of the flag. The valid values for
           <c><anno>Flag</anno></c> are only a subset of those allowed in
-          <c>process_flag/2</c>, namely: <c>save_calls</c>.</p>
-        <p>Failure: <c>badarg</c> if <c><anno>Pid</anno></c> is not a local process.</p>
+          <c>process_flag/2</c>, namely <c>save_calls</c>.</p>
+        <p>Failure: <c>badarg</c> if <c><anno>Pid</anno></c>
+          is not a local process.</p>
       </desc>
     </func>
+
     <func>
       <name name="process_info" arity="1"/>
+      <fsummary>Information about a process.</fsummary>
       <type name="process_info_result_item"/>
       <type name="priority_level"/>
       <type name="stack_item"/>
-      <fsummary>Information about a process</fsummary>
+      <type name="message_queue_data" />
       <desc>
         <p>Returns a list containing <c><anno>InfoTuple</anno></c>s with
 	  miscellaneous information about the process identified by
-	  <c>Pid</c>, or <c>undefined</c> if the process is not alive.
-	</p>
-	<p>
-	  The order of the <c><anno>InfoTuple</anno></c>s is not defined, nor
-	  are all the <c><anno>InfoTuple</anno></c>s mandatory. The <c><anno>InfoTuple</anno></c>s
-	  part of the result may be changed without prior notice.
-	  Currently <c><anno>InfoTuple</anno></c>s with the following items
-	  are part of the result:
-	  <c>current_function</c>, <c>initial_call</c>, <c>status</c>,
-	  <c>message_queue_len</c>, <c>messages</c>, <c>links</c>,
-	  <c>dictionary</c>, <c>trap_exit</c>, <c>error_handler</c>,
-	  <c>priority</c>, <c>group_leader</c>, <c>total_heap_size</c>,
-	  <c>heap_size</c>, <c>stack_size</c>, <c>reductions</c>, and
-	  <c>garbage_collection</c>.
-	  If the process identified by <c><anno>Pid</anno></c> has a registered name
-	  also an <c><anno>InfoTuple</anno></c> with the item <c>registered_name</c>
-	  will appear.
-	</p>
-	<p>See <seealso marker="#process_info/2">process_info/2</seealso>
-	   for information about specific <c><anno>InfoTuple</anno></c>s.</p>
+	  <c>Pid</c>, or <c>undefined</c> if the process is not alive.</p>
+	<p>The order of the <c><anno>InfoTuple</anno></c>s is undefined and
+	  all <c><anno>InfoTuple</anno></c>s are not mandatory.
+          The <c><anno>InfoTuple</anno></c>s
+	  part of the result can be changed without prior notice.</p>
+	<p>The <c><anno>InfoTuple</anno></c>s with the following items
+	  are part of the result:</p>
+        <list type="bulleted">
+          <item><c>current_function</c></item>
+          <item><c>initial_call</c></item>
+          <item><c>status</c></item>
+	  <item><c>message_queue_len</c></item>
+          <item><c>messages</c></item>
+          <item><c>links</c></item>
+	  <item><c>dictionary</c></item>
+          <item><c>trap_exit</c></item>
+          <item><c>error_handler</c></item>
+	  <item><c>priority</c></item>
+          <item><c>group_leader</c></item>
+          <item><c>total_heap_size</c></item>
+	  <item><c>heap_size</c></item>
+          <item><c>stack_size</c></item>
+          <item><c>reductions</c></item>
+	  <item><c>garbage_collection</c></item>
+        </list>
+        <p>If the process identified by <c><anno>Pid</anno></c> has a
+          registered name,
+	  also an <c><anno>InfoTuple</anno></c> with item <c>registered_name</c>
+	  appears.</p>
+	<p>For information about specific <c><anno>InfoTuple</anno></c>s, see
+          <seealso marker="#process_info/2">process_info/2</seealso>.</p>
         <warning>
-          <p>This BIF is intended for <em>debugging only</em>, use
-	  <seealso marker="#process_info/2">process_info/2</seealso>
-	  for all other purposes.
-	  </p>
+          <p>This BIF is intended for <em>debugging only</em>. For
+            all other purposes, use
+	    <seealso marker="#process_info/2">process_info/2</seealso>.</p>
         </warning>
-        <p>Failure: <c>badarg</c> if <c>Pid</c> is not a local process.</p>
+        <p>Failure: <c>badarg</c> if <c><anno>Pid</anno></c> is not a
+          local process.</p>
       </desc>
     </func>
+
     <func>
       <name name="process_info" arity="2" clause_i="1"/>
       <name name="process_info" arity="2" clause_i="2"/>
+      <fsummary>Information about a process.</fsummary>
       <type name="process_info_item"/>
       <type name="process_info_result_item"/>
       <type name="stack_item"/>
       <type name="priority_level"/>
-      <fsummary>Information about a process</fsummary>
-      <desc>
-        <p>Returns information about the process identified by <c><anno>Pid</anno></c>
-	   as specified by the <c><anno>Item</anno></c> or the <c><anno>ItemList</anno></c>, or <c>undefined</c> if the
-	   process is not alive.
-	</p>
-	<p>If the process is alive and a single <c><anno>Item</anno></c> is given,
-           the returned value is the corresponding
-	   <c><anno>InfoTuple</anno></c> unless <c>Item =:= registered_name</c>
-	   and the process has no registered name. In this case
-	   <c>[]</c> is returned. This strange behavior is due to
-	   historical reasons, and is kept for backward compatibility.
-	</p>
-	<p>If an  <c>ItemList</c> is given, the result is an
-	   <c><anno>InfoTupleList</anno></c>. The <c><anno>InfoTuple</anno></c>s in the
-	   <c><anno>InfoTupleList</anno></c> will appear with the corresponding
-	   <c><anno>Item</anno></c>s in the same order as the <c><anno>Item</anno></c>s appeared
-	   in the <c><anno>ItemList</anno></c>. Valid <c><anno>Item</anno></c>s may appear multiple
-	   times in the <c><anno>ItemList</anno></c>.
-	</p>
-	 <note><p>If <c>registered_name</c> is part of an <c><anno>ItemList</anno></c>
+      <type name="message_queue_data" />
+      <desc>
+        <p>Returns information about the process identified by
+           <c><anno>Pid</anno></c>, as specified by
+           <c><anno>Item</anno></c> or <c><anno>ItemList</anno></c>.
+           Returns <c>undefined</c> if the process is not alive.</p>
+	<p>If the process is alive and a single <c><anno>Item</anno></c>
+           is given, the returned value is the corresponding
+	   <c><anno>InfoTuple</anno></c>, unless <c>Item =:= registered_name</c>
+	   and the process has no registered name. In this case,
+	   <c>[]</c> is returned. This strange behavior is because of
+	   historical reasons, and is kept for backward compatibility.</p>
+	<p>If <c><anno>ItemList</anno></c> is given, the result is
+	   <c><anno>InfoTupleList</anno></c>.
+           The <c><anno>InfoTuple</anno></c>s in
+	   <c><anno>InfoTupleList</anno></c> appear with the corresponding
+	   <c><anno>Item</anno></c>s in the same order as the
+           <c><anno>Item</anno></c>s appeared
+	   in <c><anno>ItemList</anno></c>. Valid <c><anno>Item</anno></c>s can
+           appear multiple times in <c><anno>ItemList</anno></c>.</p>
+	<note><p>If <c>registered_name</c> is part of <c><anno>ItemList</anno></c>
 	          and the process has no name registered a
-		  <c>{registered_name, []}</c> <c><anno>InfoTuple</anno></c> <em>will</em>
-		  appear in the resulting <c><anno>InfoTupleList</anno></c>. This
-		  behavior is different than when a single
-		  <c>Item =:= registered_name</c> is given, and than when
-		  <c>process_info/1</c> is used.
-	</p></note>
-	<p>Currently the following <c><anno>InfoTuple</anno></c>s with corresponding
+		  <c>{registered_name, []}</c>, <c><anno>InfoTuple</anno></c>
+                  <em>will</em> appear in the resulting
+                  <c><anno>InfoTupleList</anno></c>. This
+		  behavior is different when a single
+		  <c>Item =:= registered_name</c> is given, and when
+		  <c>process_info/1</c> is used.</p>
+	</note>
+	<p>The following <c><anno>InfoTuple</anno></c>s with corresponding
 	   <c><anno>Item</anno></c>s are valid:</p>
         <taglist>
           <tag><c>{backtrace, <anno>Bin</anno>}</c></tag>
           <item>
-            <p>The binary <c><anno>Bin</anno></c> contains the same information as
-              the output from
+            <p>Binary <c><anno>Bin</anno></c> contains the same information
+              as the output from
               <c>erlang:process_display(<anno>Pid</anno>, backtrace)</c>. Use
               <c>binary_to_list/1</c> to obtain the string of characters
               from the binary.</p>
           </item>
           <tag><c>{binary, <anno>BinInfo</anno>}</c></tag>
           <item>
-            <p><c><anno>BinInfo</anno></c> is a list containing miscellaneous information
-	      about binaries currently being referred to by this process.
-	      This <c><anno>InfoTuple</anno></c> may be changed or removed without prior
-	      notice.</p>
+            <p><c><anno>BinInfo</anno></c> is a list containing miscellaneous
+              information about binaries currently being referred to by this
+              process. This <c><anno>InfoTuple</anno></c> can be changed or
+              removed without prior notice.</p>
           </item>
           <tag><c>{catchlevel, <anno>CatchLevel</anno>}</c></tag>
           <item>
             <p><c><anno>CatchLevel</anno></c> is the number of currently active
-	    catches in this process. This <c><anno>InfoTuple</anno></c> may be
+	    catches in this process. This <c><anno>InfoTuple</anno></c> can be
 	    changed or removed without prior notice.</p>
           </item>
-          <tag><c>{current_function, {<anno>Module</anno>, <anno>Function</anno>, <anno>Arity</anno>}}</c></tag>
+          <tag><c>{current_function, {<anno>Module</anno>,
+            <anno>Function</anno>, Arity}}</c></tag>
           <item>
-            <p><c><anno>Module</anno></c>, <c><anno>Function</anno></c>, <c><anno>Arity</anno></c> is
+            <p><c><anno>Module</anno></c>, <c><anno>Function</anno></c>,
+              <c><anno>Arity</anno></c> is
               the current function call of the process.</p>
           </item>
-          <tag><c>{current_location, {<anno>Module</anno>, <anno>Function</anno>, <anno>Arity</anno>, <anno>Location</anno>}}</c></tag>
+          <tag><c>{current_location, {<anno>Module</anno>,
+            <anno>Function</anno>, <anno>Arity</anno>,
+            <anno>Location</anno>}}</c></tag>
           <item>
-            <p><c><anno>Module</anno></c>, <c><anno>Function</anno></c>, <c><anno>Arity</anno></c> is
+            <p><c><anno>Module</anno></c>, <c><anno>Function</anno></c>,
+              <c><anno>Arity</anno></c> is
               the current function call of the process.
-	      <c><anno>Location</anno></c> is a list of two-tuples that describes the
-	      location in the source code.
-	    </p>
+	      <c><anno>Location</anno></c> is a list of two-tuples describing the
+	      location in the source code.</p>
           </item>
           <tag><c>{current_stacktrace, <anno>Stack</anno>}</c></tag>
           <item>
-            <p>Return the current call stack back-trace (<em>stacktrace</em>)
+            <p>Returns the current call stack back-trace (<em>stacktrace</em>)
               of the process. The stack has the same format as returned by
-	      <seealso marker="#get_stacktrace/0">erlang:get_stacktrace/0</seealso>.
-	    </p>
+	      <seealso marker="#get_stacktrace/0">erlang:get_stacktrace/0</seealso>.</p>
           </item>
           <tag><c>{dictionary, <anno>Dictionary</anno>}</c></tag>
           <item>
-            <p><c><anno>Dictionary</anno></c> is the dictionary of the process.</p>
+            <p><c><anno>Dictionary</anno></c> is the process dictionary.</p>
           </item>
           <tag><c>{error_handler, <anno>Module</anno>}</c></tag>
           <item>
@@ -4331,34 +4637,36 @@ os_prompt% </pre>
           </item>
           <tag><c>{garbage_collection, <anno>GCInfo</anno>}</c></tag>
           <item>
-            <p><c><anno>GCInfo</anno></c> is a list which contains miscellaneous
+            <p><c><anno>GCInfo</anno></c> is a list containing miscellaneous
 	       information about garbage collection for this process.
-	       The content of <c><anno>GCInfo</anno></c> may be changed without
+	       The content of <c><anno>GCInfo</anno></c> can be changed without
 	       prior notice.</p>
           </item>
           <tag><c>{group_leader, <anno>GroupLeader</anno>}</c></tag>
           <item>
-            <p><c><anno>GroupLeader</anno></c> is group leader for the IO of
+            <p><c><anno>GroupLeader</anno></c> is group leader for the I/O of
               the process.</p>
           </item>
           <tag><c>{heap_size, <anno>Size</anno>}</c></tag>
           <item>
-            <p><c><anno>Size</anno></c> is the size in words of youngest heap generation
-	    of the process. This generation currently include the stack
-	    of the process. This information is highly implementation
-	    dependent, and may change if the implementation change.
-	    </p>
+            <p><c><anno>Size</anno></c> is the size in words of the youngest heap
+              generation of the process. This generation includes
+	      the process stack. This information is highly
+              implementation-dependent, and can change if the
+              implementation changes.</p>
           </item>
-          <tag><c>{initial_call, {Module, Function, Arity}}</c></tag>
+          <tag><c>{initial_call, {<anno>Module</anno>, <anno>Function</anno>,
+            <anno>Arity</anno>}}</c></tag>
           <item>
-            <p><c>Module</c>, <c>Function</c>, <c>Arity</c> is
+            <p><c><anno>Module</anno></c>, <c><anno>Function</anno></c>,
+              <c><anno>Arity</anno></c> is
               the initial function call with which the process was
               spawned.</p>
           </item>
           <tag><c>{links, <anno>PidsAndPorts</anno>}</c></tag>
           <item>
-            <p><c><anno>PidsAndPorts</anno></c> is a list of pids and 
-	    port identifiers, with processes or ports to which the process 
+            <p><c><anno>PidsAndPorts</anno></c> is a list of process identifiers 
+	    and port identifiers, with processes or ports to which the process
 	    has a link.</p>
           </item>
           <tag><c>{last_calls, false|Calls}</c></tag>
@@ -4372,14 +4680,14 @@ os_prompt% </pre>
           <tag><c>{memory, <anno>Size</anno>}</c></tag>
           <item>
             <p><c><anno>Size</anno></c> is the size in bytes of the process. This
-              includes call stack, heap and internal structures.</p>
+              includes call stack, heap, and internal structures.</p>
           </item>
           <tag><c>{message_queue_len, <anno>MessageQueueLen</anno>}</c></tag>
           <item>
             <p><c><anno>MessageQueueLen</anno></c> is the number of messages
               currently in the message queue of the process. This is
               the length of the list <c><anno>MessageQueue</anno></c> returned as
-              the info item <c>messages</c> (see below).</p>
+              the information item <c>messages</c> (see the following).</p>
           </item>
           <tag><c>{messages, <anno>MessageQueue</anno>}</c></tag>
           <item>
@@ -4388,31 +4696,44 @@ os_prompt% </pre>
           </item>
           <tag><c>{min_heap_size, <anno>MinHeapSize</anno>}</c></tag>
           <item>
-            <p><c><anno>MinHeapSize</anno></c> is the minimum heap size for the process.</p>
+            <p><c><anno>MinHeapSize</anno></c> is the minimum heap size
+              for the process.</p>
           </item>
           <tag><c>{min_bin_vheap_size, <anno>MinBinVHeapSize</anno>}</c></tag>
           <item>
-            <p><c><anno>MinBinVHeapSize</anno></c> is the minimum binary virtual heap size for the process.</p>
+            <p><c><anno>MinBinVHeapSize</anno></c> is the minimum binary virtual
+              heap size for the process.</p>
           </item>
           <tag><c>{monitored_by, <anno>Pids</anno>}</c></tag>
           <item>
-            <p>A list of pids that are monitoring the process (with
+            <p>A list of process identifiers monitoring the process (with
               <c>monitor/2</c>).</p>
           </item>
           <tag><c>{monitors, <anno>Monitors</anno>}</c></tag>
           <item>
             <p>A list of monitors (started by <c>monitor/2</c>)
               that are active for the process. For a local process
-              monitor or a remote process monitor by pid, the list item
-              is <c>{process, <anno>Pid</anno>}</c>, and for a remote process
+              monitor or a remote process monitor by a process
+              identifier, the list item is <c>{process, <anno>Pid</anno>}</c>.
+              For a remote process
               monitor by name, the list item is
               <c>{process, {<anno>RegName</anno>, <anno>Node</anno>}}</c>.</p>
           </item>
-          <tag><c>{priority, Level}</c></tag>
+          <tag><c>{message_queue_data, <anno>MQD</anno>}</c></tag>
+          <item>
+            <p>Returns the current state of the <c>message_queue_data</c>
+	    process flag. <c><anno>MQD</anno></c> is either <c>off_heap</c>,
+	    <c>on_heap</c>, or <c>mixed</c>. For more information, see the
+	    documentation of
+            <seealso marker="#process_flag_message_queue_data"><c>process_flag(message_queue_data,
+	    MQD)</c></seealso>.</p>
+          </item>
+          <tag><c>{priority, <anno>Level</anno>}</c></tag>
           <item>
             <p><c><anno>Level</anno></c> is the current priority level for
-	      the process. For more information on priorities see
-              <seealso marker="#process_flag_priority">process_flag(priority, Level)</seealso>.</p>
+	      the process. For more information on priorities, see
+              <seealso marker="#process_flag_priority">process_flag(priority,
+              Level)</seealso>.</p>
           </item>
           <tag><c>{reductions, <anno>Number</anno>}</c></tag>
           <item>
@@ -4427,166 +4748,207 @@ os_prompt% </pre>
           </item>
           <tag><c>{sequential_trace_token, [] | <anno>SequentialTraceToken</anno>}</c></tag>
           <item>
-            <p><c><anno>SequentialTraceToken</anno></c> the sequential trace token for
-	    the process. This <c><anno>InfoTuple</anno></c> may be changed or removed
-	    without prior notice.</p>
+            <p><c><anno>SequentialTraceToken</anno></c> is the sequential trace
+            token for the process. This <c><anno>InfoTuple</anno></c> can be
+            changed or removed without prior notice.</p>
           </item>
           <tag><c>{stack_size, <anno>Size</anno>}</c></tag>
           <item>
-            <p><c><anno>Size</anno></c> is the stack size of the process in words.</p>
+            <p><c><anno>Size</anno></c> is the stack size, in words,
+              of the process.</p>
           </item>
           <tag><c>{status, <anno>Status</anno>}</c></tag>
           <item>
-            <p><c><anno>Status</anno></c> is the status of the process. <c><anno>Status</anno></c>
-              is <c>exiting</c>, <c>garbage_collecting</c>,
-              <c>waiting</c> (for a message), <c>running</c>,
-              <c>runnable</c> (ready to run, but another process is
-              running), or <c>suspended</c> (suspended on a "busy" port
-              or by the <c>erlang:suspend_process/[1,2]</c> BIF).</p>
+            <p><c><anno>Status</anno></c> is the status of the process and is one
+              of the following:</p>
+	    <list type="bulleted">
+              <item><c>exiting</c></item>
+              <item><c>garbage_collecting</c></item>
+              <item><c>waiting</c> (for a message)</item>
+              <item><c>running</c></item>
+              <item><c>runnable</c> (ready to run, but another process is
+                running)</item>
+              <item><c>suspended</c> (suspended on a "busy" port
+                or by the BIF <c>erlang:suspend_process/[1,2]</c>)</item>
+            </list>
           </item>
           <tag><c>{suspending, <anno>SuspendeeList</anno>}</c></tag>
           <item>
-            <p><c><anno>SuspendeeList</anno></c> is a list of <c>{<anno>Suspendee</anno>,
-	    <anno>ActiveSuspendCount</anno>, <anno>OutstandingSuspendCount</anno>}</c> tuples.
-	    <c><anno>Suspendee</anno></c> is the pid of a process that have been or is to
-	    be suspended by the process identified by <c><anno>Pid</anno></c> via the
-	    <seealso marker="#suspend_process/2">erlang:suspend_process/2</seealso>
-	    BIF, or the
-	    <seealso marker="#suspend_process/1">erlang:suspend_process/1</seealso>
-	    BIF. <c><anno>ActiveSuspendCount</anno></c> is the number of times the
-	    <c><anno>Suspendee</anno></c> has been suspended by <c><anno>Pid</anno></c>.
+            <p><c><anno>SuspendeeList</anno></c> is a list of
+            <c>{<anno>Suspendee</anno>, <anno>ActiveSuspendCount</anno>,
+            <anno>OutstandingSuspendCount</anno>}</c> tuples.
+	    <c><anno>Suspendee</anno></c> is the process identifier of a
+            process that has been, or is to be,
+            suspended by the process identified by <c><anno>Pid</anno></c>
+            through one of the following BIFs:</p>
+            <list type="bulleted">
+              <item>
+	      <seealso marker="#suspend_process/2">erlang:suspend_process/2</seealso>
+              </item>
+	      <item>
+	      <seealso marker="#suspend_process/1">erlang:suspend_process/1</seealso>
+              </item>
+            </list>
+	    <p><c><anno>ActiveSuspendCount</anno></c> is the number of
+            times <c><anno>Suspendee</anno></c> has been suspended by
+            <c><anno>Pid</anno></c>.
 	    <c><anno>OutstandingSuspendCount</anno></c> is the number of not yet
-	    completed suspend requests sent by <c><anno>Pid</anno></c>. That is,
-	    if <c><anno>ActiveSuspendCount</anno> =/= 0</c>, <c><anno>Suspendee</anno></c> is
-	    currently in the suspended state, and if
-	    <c><anno>OutstandingSuspendCount</anno> =/= 0</c> the <c>asynchronous</c>
-	    option of <c>erlang:suspend_process/2</c> has been used and
-	    the suspendee has not yet been suspended by <c><anno>Pid</anno></c>.
-	    Note that the <c><anno>ActiveSuspendCount</anno></c> and
-	    <c><anno>OutstandingSuspendCount</anno></c> are not the total suspend count
-	    on <c><anno>Suspendee</anno></c>, only the parts contributed by <c>Pid</c>.
-	    </p>
+	    completed suspend requests sent by <c><anno>Pid</anno></c>, that is:</p>
+            <list type="bulleted">
+              <item>If <c><anno>ActiveSuspendCount</anno> =/= 0</c>,
+                <c><anno>Suspendee</anno></c> is
+	        currently in the suspended state.
+              </item>
+              <item>If <c><anno>OutstandingSuspendCount</anno> =/= 0</c>, option
+                <c>asynchronous</c> of <c>erlang:suspend_process/2</c>
+                has been used and the suspendee has not yet been
+                suspended by <c><anno>Pid</anno></c>.
+              </item>
+            </list>
+	    <p>Notice that <c><anno>ActiveSuspendCount</anno></c> and
+	    <c><anno>OutstandingSuspendCount</anno></c> are not the
+            total suspend count on <c><anno>Suspendee</anno></c>,
+            only the parts contributed by <c><anno>Pid</anno></c>.</p>
           </item>
           <tag><c>{total_heap_size, <anno>Size</anno>}</c></tag>
           <item>
-            <p><c><anno>Size</anno></c> is the total size in words of all heap
-	    fragments of the process. This currently include the stack
-	    of the process.
-	    </p>
+            <p><c><anno>Size</anno></c> is the total size, in words, of all heap
+	    fragments of the process. This includes the process stack.</p>
           </item>
           <tag><c>{trace, <anno>InternalTraceFlags</anno>}</c></tag>
           <item>
-            <p><c><anno>InternalTraceFlags</anno></c> is an integer representing
-	    internal trace flag for this process. This <c><anno>InfoTuple</anno></c>
-	    may be changed or removed without prior notice.</p>
+            <p><c><anno>InternalTraceFlags</anno></c> is an integer
+            representing the internal trace flag for this process.
+            This <c><anno>InfoTuple</anno></c>
+	    can be changed or removed without prior notice.</p>
           </item>
           <tag><c>{trap_exit, <anno>Boolean</anno>}</c></tag>
           <item>
-            <p><c><anno>Boolean</anno></c> is <c>true</c> if the process is trapping
-              exits, otherwise it is <c>false</c>.</p>
+            <p><c><anno>Boolean</anno></c> is <c>true</c> if the process
+            is trapping exits, otherwise <c>false</c>.</p>
           </item>
         </taglist>
-        <p>Note however, that not all implementations support every one
-          of the above <c><anno>Item</anno></c>s.</p>
-        <p>Failure: <c>badarg</c> if <c><anno>Pid</anno></c> is not a local process,
-	or if <c><anno>Item</anno></c> is not a valid <c><anno>Item</anno></c>.</p>
+        <p>Notice that not all implementations support all
+          these <c><anno>Item</anno></c>s.</p>
+        <p>Failures:</p>
+        <taglist>
+          <tag><c>badarg</c></tag>
+            <item>If <c><anno>Pid</anno></c> is not a local process.</item>
+          <tag><c>badarg</c></tag>
+            <item>If <c><anno>Item</anno></c> is an invalid item.</item>
+        </taglist>
       </desc>
     </func>
+
     <func>
       <name name="processes" arity="0"/>
-      <fsummary>All processes</fsummary>
+      <fsummary>All processes.</fsummary>
       <desc>
         <p>Returns a list of process identifiers corresponding to
-	   all the processes currently existing on the local node.
-	</p>
-	<p>Note that a process that is exiting, exists but is not alive, i.e.,
-	   <c>is_process_alive/1</c> will return <c>false</c> for a process
-	   that is exiting, but its process identifier will be part
-	   of the result returned from <c>processes/0</c>.
-	</p>
+	   all the processes currently existing on the local node.</p>
+	<p>Notice that an exiting process exists, but is not alive.
+           That is, <c>is_process_alive/1</c> returns <c>false</c>
+           for an exiting process, but its process identifier is part
+	   of the result returned from <c>processes/0</c>.</p>
+        <p>Example:</p>
         <pre>
 > <input>processes().</input>
 [&lt;0.0.0&gt;,&lt;0.2.0&gt;,&lt;0.4.0&gt;,&lt;0.5.0&gt;,&lt;0.7.0&gt;,&lt;0.8.0&gt;]</pre>
       </desc>
     </func>
+
     <func>
       <name name="purge_module" arity="1"/>
-      <fsummary>Remove old code for a module</fsummary>
+      <fsummary>Removes old code for a module.</fsummary>
       <desc>
-        <p>Removes old code for <c><anno>Module</anno></c>. Before this BIF is used,
-          <c>erlang:check_process_code/2</c> should be called to check
-          that no processes are executing old code in the module.</p>
+        <p>Removes old code for <c><anno>Module</anno></c>.
+          Before this BIF is used,
+          <c>erlang:check_process_code/2</c> is to be called to check
+          that no processes execute old code in the module.</p>
         <warning>
           <p>This BIF is intended for the code server (see
-            <seealso marker="kernel:code">code(3)</seealso>) and should not be
-            used elsewhere.</p>
+            <seealso marker="kernel:code">code(3)</seealso>)
+            and is not to be used elsewhere.</p>
         </warning>
+        <note>
+	  <p>As from <c>ERTS</c> 8.0 (OTP 19), any lingering processes
+	   that still execute the old code will be killed by this function.
+	   In earlier versions, such incorrect use could cause much
+	   more fatal failures, like emulator crash.</p>
+        </note>
         <p>Failure: <c>badarg</c> if there is no old code for
           <c><anno>Module</anno></c>.</p>
       </desc>
     </func>
+
     <func>
       <name name="put" arity="2"/>
-      <fsummary>Add a new value to the process dictionary</fsummary>
-      <desc>
-        <p>Adds a new <c><anno>Key</anno></c> to the process dictionary, associated
-          with the value <c><anno>Val</anno></c>, and returns <c>undefined</c>. If
-          <c><anno>Key</anno></c> already exists, the old value is deleted and
-          replaced by <c><anno>Val</anno></c> and the function returns the old value.</p>
-        <note>
-          <p>The values stored when <c>put</c> is evaluated within
-            the scope of a <c>catch</c> will not be retracted if a
-            <c>throw</c> is evaluated, or if an error occurs.</p>
-        </note>
+      <fsummary>Adds a new value to the process dictionary.</fsummary>
+      <desc>
+        <p>Adds a new <c><anno>Key</anno></c> to the process dictionary,
+          associated with the value <c><anno>Val</anno></c>, and returns
+          <c>undefined</c>. If <c><anno>Key</anno></c> exists, the old
+          value is deleted and replaced by <c><anno>Val</anno></c>, and
+          the function returns the old value.</p>
+        <p>Example:</p>
         <pre>
 > <input>X = put(name, walrus), Y = put(name, carpenter),</input>
 <input>Z = get(name),</input>
 <input>{X, Y, Z}.</input>
 {undefined,walrus,carpenter}</pre>
+        <note>
+          <p>The values stored when <c>put</c> is evaluated within
+            the scope of a <c>catch</c> are not retracted if a
+            <c>throw</c> is evaluated, or if an error occurs.</p>
+        </note>
       </desc>
     </func>
+
     <func>
       <name name="raise" arity="3"/>
+      <fsummary>Stops execution with an exception of given class, reason, and call stack backtrace.</fsummary>
       <type name="raise_stacktrace"/>
-      <fsummary>Stop execution with an exception of given class, reason and call stack backtrace</fsummary>
       <desc>
         <p>Stops the execution of the calling process with an
-          exception of given class, reason and call stack backtrace
+          exception of given class, reason, and call stack backtrace
           (<em>stacktrace</em>).</p>
         <warning>
-          <p>This BIF is intended for debugging and for use in
-            the Erlang operating system. In general, it should
-            be avoided in applications, unless you know
-            very well what you are doing.</p>
+          <p>This BIF is intended for debugging. Avoid to use it in applications,
+            unless you really know what you are doing.</p>
         </warning>
-        <p><c><anno>Class</anno></c> is one of <c>error</c>, <c>exit</c> or
-          <c>throw</c>, so if it were not for the stacktrace
-          <c>erlang:raise(<anno>Class</anno>, <anno>Reason</anno>, <anno>Stacktrace</anno>)</c> is
-          equivalent to <c>erlang:<anno>Class</anno>(<anno>Reason</anno>)</c>.
-          <c><anno>Reason</anno></c> is any term and <c><anno>Stacktrace</anno></c> is a list as
-          returned from <c>get_stacktrace()</c>, that is a list of
-          4-tuples <c>{Module, Function, Arity | Args,
-          Location}</c> where <c>Module</c> and <c>Function</c>
-          are atoms and the third element is an integer arity or an
-          argument list. The stacktrace may also contain <c>{Fun,
-          Args, Location}</c> tuples where
-          <c>Fun</c> is a local fun and <c>Args</c> is an argument list.</p>
-	<p>The <c>Location</c> element at the end is optional.
+        <p><c><anno>Class</anno></c> is <c>error</c>, <c>exit</c>, or
+          <c>throw</c>. So, if it were not for the stacktrace,
+          <c>erlang:raise(<anno>Class</anno>, <anno>Reason</anno>,
+          <anno>Stacktrace</anno>)</c> is
+          equivalent to <c>erlang:<anno>Class</anno>(<anno>Reason</anno>)</c>.</p>
+        <p><c><anno>Reason</anno></c> is any term.
+          <c><anno>Stacktrace</anno></c> is a list as
+          returned from <c>get_stacktrace()</c>, that is, a list of
+          four-tuples <c>{Module, Function, Arity | Args,
+          Location}</c>, where <c>Module</c> and <c>Function</c>
+          are atoms, and the third element is an integer arity or an
+          argument list. The stacktrace can also contain <c>{Fun,
+          Args, Location}</c> tuples, where <c>Fun</c> is a local
+          fun and <c>Args</c> is an argument list.</p>
+	<p>Element <c>Location</c> at the end is optional.
 	  Omitting it is equivalent to specifying an empty list.</p>
         <p>The stacktrace is used as the exception stacktrace for the
-          calling process; it will be truncated to the current
+          calling process; it is truncated to the current
           maximum stacktrace depth.</p>
-        <p>Because evaluating this function causes the process to
-          terminate, it has no return value - unless the arguments are
-          invalid, in which case the function <em>returns the error reason</em>, that is <c>badarg</c>. If you want to be
-          really sure not to return you can call
-          <c>error(erlang:raise(<anno>Class</anno>, <anno>Reason</anno>, <anno>Stacktrace</anno>))</c>
+        <p>Since evaluating this function causes the process to
+          terminate, it has no return value unless the arguments are
+          invalid, in which case the function <em>returns the error
+          reason</em> <c>badarg</c>. If you want to be
+          sure not to return, you can call
+          <c>error(erlang:raise(<anno>Class</anno>, <anno>Reason</anno>,
+          <anno>Stacktrace</anno>))</c>
           and hope to distinguish exceptions later.</p>
       </desc>
     </func>
+
     <func>
       <name name="read_timer" arity="2"/>
-      <fsummary>Read the state of a timer</fsummary>
+      <fsummary>Reads the state of a timer.</fsummary>
       <desc>
         <p>
 	  Read the state of a timer that has been created by either
@@ -4595,7 +4957,7 @@ os_prompt% </pre>
 	  <c><anno>TimerRef</anno></c> identifies the timer, and
 	  was returned by the BIF that created the timer.
 	</p>
-	<p>Currently available <c><anno>Option</anno>s</c>:</p>
+	<p>Available <c><anno>Option</anno>s</c>:</p>
         <taglist>
           <tag><c>{async, Async}</c></tag>
           <item>
@@ -4603,12 +4965,12 @@ os_prompt% </pre>
 	      Asynchronous request for state information. <c>Async</c>
 	      defaults to <c>false</c> which will cause the operation
 	      to be performed synchronously. In this case, the <c>Result</c>
-	      will be returned by <c>erlang:read_timer()</c>. When
-	      <c>Async</c> is set to <c>true</c>, <c>erlang:read_timer()</c>
-	      will send an asynchronous request for the state information
-	      to the timer service that manages the timer, and then return
+	      is returned by <c>erlang:read_timer()</c>. When
+	      <c>Async</c> is <c>true</c>, <c>erlang:read_timer()</c>
+	      sends an asynchronous request for the state information
+	      to the timer service that manages the timer, and then returns
 	      <c>ok</c>. A message on the format <c>{read_timer,
-	      <anno>TimerRef</anno>, <anno>Result</anno>}</c> will be
+	      <anno>TimerRef</anno>, <anno>Result</anno>}</c> is
 	      sent to the caller of <c>erlang:read_timer()</c> when the
 	      operation has been processed.
 	    </p>
@@ -4618,26 +4980,27 @@ os_prompt% </pre>
 	  More <c><anno>Option</anno></c>s may be added in the future.
 	</p>
 	<p>
-	  When the <c><anno>Result</anno></c> equals <c>false</c>, a
+	  If <c><anno>Result</anno></c> is an integer, it represents the
+	  time in milli-seconds left until the timer expires.</p>
+	<p>
+	  If <c><anno>Result</anno></c> is <c>false</c>, a
 	  timer corresponding to <c><anno>TimerRef</anno></c> could not
-	  be found. This can be either because the timer had expired,
-	  had been canceled, or because <c><anno>TimerRef</anno></c>
-	  never has corresponded to a timer. If the timer has expired,
-	  the timeout message has been sent, but it does not tell you
-	  whether or not it has arrived at its destination yet. When the
-	  <c><anno>Result</anno></c> is an integer, it represents the
-	  time in milli-seconds left until the timer will expire.
+	  be found. This can be because the timer had expired,
+	  it had been canceled, or because <c><anno>TimerRef</anno></c>
+	  never has corresponded to a timer. Even if the timer has expired,
+	  it does not tell you whether or not the timeout message has
+	  arrived at its destination yet.
 	</p>
 	<note>
 	  <p>
 	    The timer service that manages the timer may be co-located
 	    with another scheduler than the scheduler that the calling
 	    process is executing on. If this is the case, communication
-	    with the timer service will take much longer time than if it
+	    with the timer service takes much longer time than if it
 	    is located locally. If the calling process is in critical
 	    path, and can do other things while waiting for the result
-	    of this operation you want to use the <c>{async, true}</c>
-	    option. If using the <c>{async, false}</c> option, the calling
+	    of this operation, you want to use option <c>{async, true}</c>.
+	    If using option <c>{async, false}</c>, the calling
 	    process will be blocked until the operation has been
 	    performed.
 	  </p>
@@ -4651,70 +5014,86 @@ os_prompt% </pre>
     </func>
     <func>
       <name name="read_timer" arity="1"/>
-      <fsummary>Read the state of a timer</fsummary>
+      <fsummary>Reads the state of a timer.</fsummary>
       <desc>
         <p>Read the state of a timer. The same as calling
 	<seealso marker="#read_timer/2"><c>erlang:read_timer(TimerRef,
 	[])</c></seealso>.</p>
       </desc>
     </func>
+
     <func>
       <name name="ref_to_list" arity="1"/>
-      <fsummary>Text representation of a reference</fsummary>
+      <fsummary>Text representation of a reference.</fsummary>
       <desc>
-        <p>Returns a string which corresponds to the text
+        <p>Returns a string corresponding to the text
           representation of <c><anno>Ref</anno></c>.</p>
         <warning>
-          <p>This BIF is intended for debugging and for use in
-            the Erlang operating system. It should not be used in
-            application programs.</p>
+          <p>This BIF is intended for debugging and is not to be used
+	  in application programs.</p>
         </warning>
       </desc>
     </func>
+
     <func>
       <name name="register" arity="2"/>
-      <fsummary>Register a name for a pid (or port)</fsummary>
+      <fsummary>Registers a name for a pid (or port).</fsummary>
       <desc>
-        <p>Associates the name <c><anno>RegName</anno></c> with a pid or a port
-          identifier. <c><anno>RegName</anno></c>, which must be an atom, can be used
-          instead of the pid / port identifier in the send operator
+        <p>Associates the name <c><anno>RegName</anno></c> with a process
+          identifier (pid) or a port identifier.
+          <c><anno>RegName</anno></c>, which must be an atom, can be used
+          instead of the pid or port identifier in send operator
           (<c><anno>RegName</anno> ! Message</c>).</p>
+        <p>Example:</p>
         <pre>
 > <input>register(db, Pid).</input>
 true</pre>
-        <p>Failure: <c>badarg</c> if <c><anno>PidOrPort</anno></c> is not an existing,
-          local process or port, if <c><anno>RegName</anno></c> is already in use,
-          if the process or port is already registered (already has a
-          name), or if <c><anno>RegName</anno></c> is the atom <c>undefined</c>.</p>
+        <p>Failures:</p>
+        <taglist>
+          <tag><c>badarg</c></tag>
+            <item>If <c><anno>PidOrPort</anno></c> is not an existing local
+              process or port.</item>
+          <tag><c>badarg</c></tag>
+            <item>If <c><anno>RegName</anno></c> is already in use.</item>
+          <tag><c>badarg</c></tag>
+            <item>If the process or port is already registered
+              (already has a name).</item>
+          <tag><c>badarg</c></tag>
+            <item>If <c><anno>RegName</anno></c> is the atom
+              <c>undefined</c>.</item>
+        </taglist>
       </desc>
     </func>
+
     <func>
       <name name="registered" arity="0"/>
-      <fsummary>All registered names</fsummary>
+      <fsummary>All registered names.</fsummary>
       <desc>
-        <p>Returns a list of names which have been registered using
-          <seealso marker="#register/2">register/2</seealso>.</p>
+        <p>Returns a list of names that have been registered using
+          <seealso marker="#register/2">register/2</seealso>, for
+          example:</p>
         <pre>
 > <input>registered().</input>
 [code_server, file_server, init, user, my_db]</pre>
       </desc>
     </func>
+
     <func>
       <name name="resume_process" arity="1"/>
-      <fsummary>Resume a suspended process</fsummary>
+      <fsummary>Resumes a suspended process.</fsummary>
       <desc>
         <p>Decreases the suspend count on the process identified by
-	<c><anno>Suspendee</anno></c>. <c><anno>Suspendee</anno></c> should previously have been
-	suspended via
-	<seealso marker="#suspend_process/2">erlang:suspend_process/2</seealso>,
+	<c><anno>Suspendee</anno></c>. <c><anno>Suspendee</anno></c>
+        is previously to have been suspended through
+	<seealso marker="#suspend_process/2">erlang:suspend_process/2</seealso>
 	or
 	<seealso marker="#suspend_process/1">erlang:suspend_process/1</seealso>
-	by the process calling <c>erlang:resume_process(<anno>Suspendee</anno>)</c>. When
-	the suspend count on <c><anno>Suspendee</anno></c> reach zero, <c><anno>Suspendee</anno></c>
-	will be resumed, i.e., the state of the <c>Suspendee</c> is changed
-	from suspended into the state <c><anno>Suspendee</anno></c> was in before it was
-	suspended.
-	</p>
+	by the process calling
+        <c>erlang:resume_process(<anno>Suspendee</anno>)</c>. When the
+        suspend count on <c><anno>Suspendee</anno></c> reaches zero,
+        <c><anno>Suspendee</anno></c> is resumed, that is, its state
+        is changed from suspended into the state it had before it was
+        suspended.</p>
         <warning>
           <p>This BIF is intended for debugging only.</p>
         </warning>
@@ -4722,7 +5101,7 @@ true</pre>
         <taglist>
           <tag><c>badarg</c></tag>
           <item>
-	  If <c><anno>Suspendee</anno></c> isn't a process identifier.
+	  If <c><anno>Suspendee</anno></c> is not a process identifier.
 	  </item>
           <tag><c>badarg</c></tag>
           <item>
@@ -4732,58 +5111,65 @@ true</pre>
 	  </item>
           <tag><c>badarg</c></tag>
           <item>
-	  If the process identified by <c><anno>Suspendee</anno></c> is not alive.
+	  If the process identified by <c><anno>Suspendee</anno></c>
+           is not alive.
 	  </item>
         </taglist>
       </desc>
     </func>
+
     <func>
       <name name="round" arity="1"/>
-      <fsummary>Return an integer by rounding a number</fsummary>
+      <fsummary>Returns an integer by rounding a number.</fsummary>
       <desc>
-        <p>Returns an integer by rounding <c><anno>Number</anno></c>.</p>
+        <p>Returns an integer by rounding <c><anno>Number</anno></c>,
+        for example:</p>
         <pre>
-> <input>round(5.5).</input>
+<input>round(5.5).</input>
 6</pre>
         <p>Allowed in guard tests.</p>
       </desc>
     </func>
+
     <func>
       <name name="self" arity="0"/>
-      <fsummary>Pid of the calling process</fsummary>
+      <fsummary>Returns pid of the calling process.</fsummary>
       <desc>
-        <p>Returns the pid (process identifier) of the calling process.</p>
+        <p>Returns the process identifier of the calling process, for
+          example:</p>
         <pre>
 > <input>self().</input>
 &lt;0.26.0></pre>
         <p>Allowed in guard tests.</p>
       </desc>
     </func>
+
     <func>
       <name name="send" arity="2"/>
-      <fsummary>Send a message</fsummary>
+      <fsummary>Sends a message.</fsummary>
       <type name="dst"/>
       <desc>
-        <p>Sends a message and returns <c><anno>Msg</anno></c>. This is the same as
-          <c><anno>Dest</anno> ! <anno>Msg</anno></c>.</p>
-        <p><c><anno>Dest</anno></c> may be a remote or local pid, a (local) port, a
-          locally registered name, or a tuple <c>{<anno>RegName</anno>, <anno>Node</anno>}</c>
+        <p>Sends a message and returns <c><anno>Msg</anno></c>. This
+          is the same as <c><anno>Dest</anno> ! <anno>Msg</anno></c>.</p>
+        <p><c><anno>Dest</anno></c> can be a remote or local process identifier,
+          a (local) port, a locally registered name, or a tuple
+          <c>{<anno>RegName</anno>, <anno>Node</anno>}</c>
           for a registered name at another node.</p>
       </desc>
     </func>
+
     <func>
       <name name="send" arity="3"/>
+      <fsummary>Sends a message conditionally.</fsummary>
       <type name="dst"/>
-      <fsummary>Send a message conditionally</fsummary>
-      <desc>
-        <p>Sends a message and returns <c>ok</c>, or does not send
-          the message but returns something else (see below). Otherwise
-          the same as
-          <seealso marker="#send/2">erlang:send/2</seealso>. See
-          also
-          <seealso marker="#send_nosuspend/2">erlang:send_nosuspend/2,3</seealso>.
-          for more detailed explanation and warnings.</p>
-        <p>The possible options are:</p>
+      <desc>
+        <p>Either sends a message and returns <c>ok</c>, or does not send
+          the message but returns something else (see the following).
+          Otherwise the same as
+          <seealso marker="#send/2">erlang:send/2</seealso>.
+          For more detailed explanation and warnings, see
+          <seealso marker="#send_nosuspend/2">erlang:send_nosuspend/2,3</seealso>.</p>
+        <p>The options are as follows:</p>
         <taglist>
           <tag><c>nosuspend</c></tag>
           <item>
@@ -4793,314 +5179,372 @@ true</pre>
           <tag><c>noconnect</c></tag>
           <item>
             <p>If the destination node would have to be auto-connected
-              before doing the send, <c>noconnect</c> is returned
+              to do the send, <c>noconnect</c> is returned
               instead.</p>
           </item>
         </taglist>
         <warning>
-          <p>As with <c>erlang:send_nosuspend/2,3</c>: Use with extreme
-            care!</p>
+          <p>As with <c>erlang:send_nosuspend/2,3</c>: use with extreme
+            care.</p>
         </warning>
       </desc>
     </func>
+
     <func>
       <name name="send_after" arity="4"/>
       <fsummary>Start a timer</fsummary>
       <desc>
         <p>
 	  Starts a timer. When the timer expires, the message
-          <c><anno>Msg</anno></c> will be sent to the process
-	  identified by <c><anno>Dest</anno></c>. Appart from
-	  the format of the message sent to
-	  <c><anno>Dest</anno></c> when the timer expires
+          <c><anno>Msg</anno></c> is sent to the process
+	  identified by <c><anno>Dest</anno></c>. Apart from
+	  the format of the timeout message,
 	  <c>erlang:send_after/4</c> works exactly as
 	  <seealso marker="#start_timer/4"><c>erlang:start_timer/4</c></seealso>.</p>
       </desc>
     </func>
     <func>
       <name name="send_after" arity="3"/>
-      <fsummary>Start a timer</fsummary>
+      <fsummary>Starts a timer.</fsummary>
       <desc>
         <p>Starts a timer. The same as calling
 	<seealso marker="#send_after/4"><c>erlang:send_after(<anno>Time</anno>,
 	<anno>Dest</anno>, <anno>Msg</anno>, [])</c></seealso>.</p>
       </desc>
     </func>
+
     <func>
       <name name="send_nosuspend" arity="2"/>
-      <fsummary>Try to send a message without ever blocking</fsummary>
+      <fsummary>Tries to send a message without ever blocking.</fsummary>
       <type name="dst"/>
       <desc>
         <p>The same as
-          <seealso marker="#send/3">erlang:send(<anno>Dest</anno>, <anno>Msg</anno>, [nosuspend])</seealso>, but returns <c>true</c> if
+          <seealso marker="#send/3">erlang:send(<anno>Dest</anno>,
+          <anno>Msg</anno>, [nosuspend])</seealso>,
+          but returns <c>true</c> if
           the message was sent and <c>false</c> if the message was not
           sent because the sender would have had to be suspended.</p>
-        <p>This function is intended for send operations towards an
+        <p>This function is intended for send operations to an
           unreliable remote node without ever blocking the sending
           (Erlang) process. If the connection to the remote node
           (usually not a real Erlang node, but a node written in C or
-          Java) is overloaded, this function <em>will not send the message</em> but return <c>false</c> instead.</p>
-        <p>The same happens, if <c><anno>Dest</anno></c> refers to a local port that
-          is busy. For all other destinations (allowed for the ordinary
-          send operator <c>'!'</c>) this function sends the message and
+          Java) is overloaded, this function <em>does not send the message</em>
+          and returns <c>false</c>.</p>
+        <p>The same occurs if <c><anno>Dest</anno></c> refers to a local port
+          that is busy. For all other destinations (allowed for the ordinary
+          send operator <c>'!'</c>), this function sends the message and
           returns <c>true</c>.</p>
-        <p>This function is only to be used in very rare circumstances
+        <p>This function is only to be used in rare circumstances
           where a process communicates with Erlang nodes that can
-          disappear without any trace causing the TCP buffers and
-          the drivers queue to be over-full before the node will actually
-          be shut down (due to tick timeouts) by <c>net_kernel</c>. The
-          normal reaction to take when this happens is some kind of
+          disappear without any trace, causing the TCP buffers and
+          the drivers queue to be over-full before the node is
+          shut down (because of tick time-outs) by <c>net_kernel</c>.
+          The normal reaction to take when this occurs is some kind of
           premature shutdown of the other node.</p>
-        <p>Note that ignoring the return value from this function would
-          result in <em>unreliable</em> message passing, which is
+        <p>Notice that ignoring the return value from this function would
+          result in an <em>unreliable</em> message passing, which is
           contradictory to the Erlang programming model. The message is
           <em>not</em> sent if this function returns <c>false</c>.</p>
-        <p>Note also that in many systems, transient states of
+        <p>In many systems, transient states of
           overloaded queues are normal. The fact that this function
-          returns <c>false</c> does not in any way mean that the other
+          returns <c>false</c> does not mean that the other
           node is guaranteed to be non-responsive, it could be a
-          temporary overload. Also a return value of <c>true</c> does
-          only mean that the message could be sent on the (TCP) channel
-          without blocking, the message is not guaranteed to have
-          arrived at the remote node. Also in the case of a disconnected
+          temporary overload. Also, a return value of <c>true</c> does
+          only mean that the message can be sent on the (TCP) channel
+          without blocking, the message is not guaranteed to
+          arrive at the remote node. For a disconnected
           non-responsive node, the return value is <c>true</c> (mimics
-          the behaviour of the <c>!</c> operator). The expected
-          behaviour as well as the actions to take when the function
-          returns <c>false</c> are application and hardware specific.</p>
+          the behavior of operator <c>!</c>). The expected
+          behavior and the actions to take when the function
+          returns <c>false</c> are application- and hardware-specific.</p>
         <warning>
-          <p>Use with extreme care!</p>
+          <p>Use with extreme care.</p>
         </warning>
       </desc>
     </func>
+
     <func>
       <name name="send_nosuspend" arity="3"/>
-      <fsummary>Try to send a message without ever blocking</fsummary>
+      <fsummary>Tries to send a message without ever blocking.</fsummary>
       <type name="dst"/>
       <desc>
         <p>The same as
-          <seealso marker="#send/3">erlang:send(<anno>Dest</anno>, <anno>Msg</anno>, [nosuspend | <anno>Options</anno>])</seealso>,
-          but with boolean return value.</p>
+          <seealso marker="#send/3">erlang:send(<anno>Dest</anno>,
+          <anno>Msg</anno>, [nosuspend | <anno>Options</anno>])</seealso>,
+          but with a Boolean return value.</p>
         <p>This function behaves like
-          <seealso marker="#send_nosuspend/2">erlang:send_nosuspend/2)</seealso>,
-          but takes a third parameter, a list of options. The only
-          currently implemented option is <c>noconnect</c>. The option
-          <c>noconnect</c> makes the function return <c>false</c> if
+          <seealso marker="#send_nosuspend/2">erlang:send_nosuspend/2</seealso>,
+          but takes a third parameter, a list of options.
+          The only option is <c>noconnect</c>, which
+          makes the function return <c>false</c> if
           the remote node is not currently reachable by the local
-          node. The normal behaviour is to try to connect to the node,
-          which may stall the process for a shorter period. The use of
-          the <c>noconnect</c> option makes it possible to be
-          absolutely sure not to get even the slightest delay when
+          node. The normal behavior is to try to connect to the node,
+          which can stall the process during a short period. The use of
+          option <c>noconnect</c> makes it possible to be
+          sure not to get the slightest delay when
           sending to a remote process. This is especially useful when
-          communicating with nodes who expect to always be
-          the connecting part (i.e. nodes written in C or Java).</p>
+          communicating with nodes that expect to always be
+          the connecting part (that is, nodes written in C or Java).</p>
         <p>Whenever the function returns <c>false</c> (either when a
           suspend would occur or when <c>noconnect</c> was specified and
           the node was not already connected), the message is guaranteed
           <em>not</em> to have been sent.</p>
         <warning>
-          <p>Use with extreme care!</p>
+          <p>Use with extreme care.</p>
         </warning>
       </desc>
     </func>
+
     <func>
       <name name="set_cookie" arity="2"/>
-      <fsummary>Set the magic cookie of a node</fsummary>
+      <fsummary>Sets the magic cookie of a node.</fsummary>
       <desc>
         <p>Sets the magic cookie of <c><anno>Node</anno></c> to the atom
-          <c><anno>Cookie</anno></c>. If <c><anno>Node</anno></c> is the local node, the function
+          <c><anno>Cookie</anno></c>. If <c><anno>Node</anno></c> is the
+          local node, the function
           also sets the cookie of all other unknown nodes to
-          <c><anno>Cookie</anno></c> (see
-          <seealso marker="doc/reference_manual:distributed">Distributed Erlang</seealso> in the Erlang Reference Manual).</p>
+          <c><anno>Cookie</anno></c> (see Section
+          <seealso marker="doc/reference_manual:distributed">Distributed Erlang</seealso>
+          in the Erlang Reference Manual in System Documentation).</p>
         <p>Failure: <c>function_clause</c> if the local node is not
           alive.</p>
       </desc>
     </func>
+
     <func>
       <name name="setelement" arity="3"/>
-      <type_desc variable="Index">1..tuple_size(<anno>Tuple1</anno>)</type_desc>
-      <fsummary>Set Nth element of a tuple</fsummary>
+      <fsummary>Sets the Nth element of a tuple.</fsummary>
+      <type_desc variable="Index">1..tuple_size(<anno>Tuple1</anno></type_desc>
       <desc>
-        <p>Returns a tuple which is a copy of the argument <c><anno>Tuple1</anno></c>
-          with the element given by the integer argument <c><anno>Index</anno></c>
+        <p>Returns a tuple that is a copy of argument
+          <c><anno>Tuple1</anno></c>
+          with the element given by integer argument
+          <c><anno>Index</anno></c>
           (the first element is the element with index 1) replaced by
-          the argument <c><anno>Value</anno></c>.</p>
+          argument <c><anno>Value</anno></c>, for example:</p>
         <pre>
 > <input>setelement(2, {10, green, bottles}, red).</input>
 {10,red,bottles}</pre>
       </desc>
     </func>
+
     <func>
       <name name="size" arity="1"/>
-      <fsummary>Size of a tuple or binary</fsummary>
+      <fsummary>Size of a tuple or binary.</fsummary>
       <desc>
-        <p>Returns an integer which is the size of the argument
-          <c><anno>Item</anno></c>,  which must be either a tuple or a binary.</p>
+        <p>Returns the number of elements in a tuple or the number of
+	bytes in a binary or bitstring, for example:</p>
         <pre>
 > <input>size({morni, mulle, bwange}).</input>
-3</pre>
+3
+> <input>size(&lt;&lt;11, 22, 33&gt;&gt;).</input>
+3
+</pre>
+        <p>For bitstrings the number of whole bytes is returned. That is, if the number of bits
+          in the bitstring is not divisible by 8, the resulting
+          number of bytes is rounded <em>down</em>.</p>
         <p>Allowed in guard tests.</p>
+        <p>See also
+          <seealso marker="#tuple_size/1"><c>tuple_size/1</c></seealso>,
+          <seealso marker="#byte_size/1"><c>byte_size/1</c></seealso>
+	  and
+          <seealso marker="#bit_size/1"><c>bit_size/1</c></seealso>.</p>
       </desc>
     </func>
+
     <func>
       <name name="spawn" arity="1"/>
-      <fsummary>Create a new process with a fun as entry point</fsummary>
+      <fsummary>Creates a new process with a fun as entry point.</fsummary>
       <desc>
-        <p>Returns the pid of a new process started by the application
-          of <c><anno>Fun</anno></c> to the empty list <c>[]</c>. Otherwise works
-          like <seealso marker="#spawn/3">spawn/3</seealso>.</p>
+        <p>Returns the process identifier of a new process started by the
+          application of <c><anno>Fun</anno></c> to the empty list
+          <c>[]</c>. Otherwise
+          works like <seealso marker="#spawn/3">spawn/3</seealso>.</p>
       </desc>
     </func>
+
     <func>
       <name name="spawn" arity="2"/>
-      <fsummary>Create a new process with a fun as entry point on a given node</fsummary>
+      <fsummary>Creates a new process with a fun as entry point on a given node.</fsummary>
       <desc>
-        <p>Returns the pid of a new process started by the application
-          of <c><anno>Fun</anno></c> to the empty list <c>[]</c> on <c><anno>Node</anno></c>. If
-          <c><anno>Node</anno></c> does not exist, a useless pid is returned.
-          Otherwise works like
+        <p>Returns the process identifier of a new process started
+          by the application of <c><anno>Fun</anno></c> to the
+          empty list <c>[]</c> on <c><anno>Node</anno></c>. If
+          <c><anno>Node</anno></c> does not exist, a useless pid is
+          returned. Otherwise works like
           <seealso marker="#spawn/3">spawn/3</seealso>.</p>
       </desc>
     </func>
+
     <func>
       <name name="spawn" arity="3"/>
-      <fsummary>Create a new process with a function as entry point</fsummary>
-      <desc>
-        <p>Returns the pid of a new process started by the application
-          of <c><anno>Module</anno>:<anno>Function</anno></c> to <c><anno>Args</anno></c>. The new process
-          created will be placed in the system scheduler queue and be
-          run some time later.</p>
-        <p><c>error_handler:undefined_function(<anno>Module</anno>, <anno>Function</anno>, <anno>Args</anno>)</c> is evaluated by the new process if
-          <c><anno>Module</anno>:<anno>Function</anno>/Arity</c> does not exist (where
-          <c>Arity</c> is the length of <c><anno>Args</anno></c>). The error handler
+      <fsummary>Creates a new process with a function as entry point.</fsummary>
+      <desc>
+        <p>Returns the process identifier of a new process started by
+          the application of <c><anno>Module</anno>:<anno>Function</anno></c>
+          to <c><anno>Args</anno></c>.</p>
+        <p><c>error_handler:undefined_function(<anno>Module</anno>,
+          <anno>Function</anno>, <anno>Args</anno>)</c>
+          is evaluated by the new process if
+          <c><anno>Module</anno>:<anno>Function</anno>/Arity</c>
+          does not exist (where <c>Arity</c> is the length of
+          <c><anno>Args</anno></c>). The error handler
           can be redefined (see
           <seealso marker="#process_flag/2">process_flag/2</seealso>).
           If <c>error_handler</c> is undefined, or the user has
-          redefined the default <c>error_handler</c> its replacement is
-          undefined, a failure with the reason <c>undef</c> will occur.</p>
+          redefined the default <c>error_handler</c> and its replacement is
+          undefined, a failure with reason <c>undef</c> occurs.</p>
+        <p>Example:</p>
         <pre>
 > <input>spawn(speed, regulator, [high_speed, thin_cut]).</input>
 &lt;0.13.1></pre>
       </desc>
     </func>
+
     <func>
       <name name="spawn" arity="4"/>
-      <fsummary>Create a new process with a function as entry point on a given node</fsummary>
+      <fsummary>Creates a new process with a function as entry point on a given node.</fsummary>
       <desc>
-        <p>Returns the pid of a new process started by the application
-          of <c><anno>Module</anno>:<anno>Function</anno></c> to <c><anno>Args</anno></c> on <c>Node</c>. If
-          <c><anno>Node</anno></c> does not exists, a useless pid is returned.
+        <p>Returns the process identifier (pid) of a new process started
+          by the application
+          of <c><anno>Module</anno>:<anno>Function</anno></c>
+          to <c><anno>Args</anno></c> on <c><anno>Node</anno></c>. If
+          <c><anno>Node</anno></c> does not exist, a useless pid is returned.
           Otherwise works like
           <seealso marker="#spawn/3">spawn/3</seealso>.</p>
       </desc>
     </func>
+
     <func>
       <name name="spawn_link" arity="1"/>
-      <fsummary>Create and link to a new process with a fun as entry point</fsummary>
+      <fsummary>Creates and links to a new process with a fun as entry point.</fsummary>
       <desc>
-        <p>Returns the pid of a new process started by the application
-          of <c><anno>Fun</anno></c> to the empty list []. A link is created between
+        <p>Returns the process identifier of a new process started by
+          the application of <c><anno>Fun</anno></c> to the empty list
+          <c>[]</c>. A link is created between
           the calling process and the new process, atomically.
           Otherwise works like
           <seealso marker="#spawn/3">spawn/3</seealso>.</p>
       </desc>
     </func>
+
     <func>
       <name name="spawn_link" arity="2"/>
-      <fsummary>Create and link to a new process with a fun as entry point on a specified node</fsummary>
+      <fsummary>Creates and links to a new process with a fun as entry point on a specified node.</fsummary>
       <desc>
-        <p>Returns the pid of a new process started by the application
-          of <c><anno>Fun</anno></c> to the empty list [] on <c><anno>Node</anno></c>. A link is
+        <p>Returns the process identifier (pid) of a new process started
+          by the application of <c><anno>Fun</anno></c> to the empty
+          list <c>[]</c> on <c><anno>Node</anno></c>. A link is
           created between the calling process and the new process,
-          atomically. If <c><anno>Node</anno></c> does not exist, a useless pid is
-          returned (and due to the link, an exit signal with exit
-          reason <c>noconnection</c> will be received). Otherwise works
-          like <seealso marker="#spawn/3">spawn/3</seealso>.</p>
+          atomically. If <c><anno>Node</anno></c> does not exist,
+          a useless pid is returned and an exit signal with
+          reason <c>noconnection</c> is sent to the calling
+	  process. Otherwise works like <seealso marker="#spawn/3">spawn/3</seealso>.</p>
       </desc>
     </func>
+
     <func>
       <name name="spawn_link" arity="3"/>
-      <fsummary>Create and link to a new process with a function as entry point</fsummary>
+      <fsummary>Creates and links to a new process with a function as entry point.</fsummary>
       <desc>
-        <p>Returns the pid of a new process started by the application
-          of <c><anno>Module</anno>:<anno>Function</anno></c> to <c><anno>Args</anno></c>. A link is created
+        <p>Returns the process identifier of a new process started by
+          the application of <c><anno>Module</anno>:<anno>Function</anno></c>
+          to <c><anno>Args</anno></c>. A link is created
           between the calling process and the new process, atomically.
           Otherwise works like
           <seealso marker="#spawn/3">spawn/3</seealso>.</p>
       </desc>
     </func>
+
     <func>
       <name name="spawn_link" arity="4"/>
-      <fsummary>Create and link to a new process with a function as entry point on a given node</fsummary>
+      <fsummary>Creates and links to a new process with a function as entry point on a given node.</fsummary>
       <desc>
-        <p>Returns the pid of a new process started by the application
-          of <c><anno>Module</anno>:<anno>Function</anno></c> to <c><anno>Args</anno></c> on <c>Node</c>. A
+        <p>Returns the process identifier (pid) of a new process
+          started by the application
+          of <c><anno>Module</anno>:<anno>Function</anno></c>
+          to <c><anno>Args</anno></c> on <c><anno>Node</anno></c>. A
           link is created between the calling process and the new
-          process, atomically. If <c><anno>Node</anno></c> does not exist, a useless
-          pid is returned (and due to the link, an exit signal with exit
-          reason <c>noconnection</c> will be received). Otherwise works
-          like <seealso marker="#spawn/3">spawn/3</seealso>.</p>
+          process, atomically. If <c><anno>Node</anno></c> does
+          not exist, a useless pid is returned and an exit signal with
+          reason <c>noconnection</c> is sent to the calling
+	  process. Otherwise works like <seealso marker="#spawn/3">spawn/3</seealso>.</p>
       </desc>
     </func>
+
     <func>
       <name name="spawn_monitor" arity="1"/>
-      <fsummary>Create and monitor a new process with a fun as entry point</fsummary>
+      <fsummary>Creates and monitors a new process with a fun as entry point.</fsummary>
       <desc>
-        <p>Returns the pid of a new process started by the application
-          of <c><anno>Fun</anno></c> to the empty list [] and reference for a monitor
-          created to the new process.
+        <p>Returns the process identifier of a new process, started by
+          the application of <c><anno>Fun</anno></c> to the empty list
+          <c>[]</c>,
+          and a reference for a monitor created to the new process.
           Otherwise works like
           <seealso marker="#spawn/3">spawn/3</seealso>.</p>
       </desc>
     </func>
+
     <func>
       <name name="spawn_monitor" arity="3"/>
-      <fsummary>Create and monitor a new process with a function as entry point</fsummary>
+      <fsummary>Creates and monitors a new process with a function as entry point.</fsummary>
       <desc>
         <p>A new process is started by the application
-          of <c><anno>Module</anno>:<anno>Function</anno></c> to <c><anno>Args</anno></c>, and the process is
-          monitored at the same time. Returns the pid and a reference
-          for the monitor.
-          Otherwise works like
+          of <c><anno>Module</anno>:<anno>Function</anno></c>
+          to <c><anno>Args</anno></c>. The process is
+          monitored at the same time. Returns the process identifier
+          and a reference for the monitor. Otherwise works like
           <seealso marker="#spawn/3">spawn/3</seealso>.</p>
       </desc>
     </func>
+
     <func>
       <name name="spawn_opt" arity="2"/>
-      <type name="priority_level" />
-      <fsummary>Create a new process with a fun as entry point</fsummary>
+      <fsummary>Creates a new process with a fun as entry point.</fsummary>
+      <type name="priority_level"/>
+      <type name="message_queue_data" />
+      <type name="spawn_opt_option" />
       <desc>
-        <p>Returns the pid of a new process started by the application
-          of <c><anno>Fun</anno></c> to the empty list <c>[]</c>. Otherwise
-          works like
+        <p>Returns the process identifier (pid) of a new process
+          started by the application of <c><anno>Fun</anno></c>
+          to the empty list <c>[]</c>. Otherwise works like
           <seealso marker="#spawn_opt/4">spawn_opt/4</seealso>.</p>
-        <p>If the option <c>monitor</c> is given, the newly created
-          process will be monitored and both the pid and reference for
-          the monitor will be returned.</p>
+        <p>If option <c>monitor</c> is given, the newly created
+          process is monitored, and both the pid and reference for
+          the monitor is returned.</p>
       </desc>
     </func>
+
     <func>
       <name name="spawn_opt" arity="3"/>
-      <type name="priority_level" />
-      <fsummary>Create a new process with a fun as entry point on a given node</fsummary>
-      <desc>
-        <p>Returns the pid of a new process started by the application
-          of <c><anno>Fun</anno></c> to the empty list <c>[]</c> on <c><anno>Node</anno></c>. If
-          <c><anno>Node</anno></c> does not exist, a useless pid is returned.
-          Otherwise works like
+      <fsummary>Creates a new process with a fun as entry point on a given node.</fsummary>
+      <type name="priority_level"/>
+      <type name="message_queue_data" />
+      <type name="spawn_opt_option" />
+      <desc>
+        <p>Returns the process identifier (pid) of a new process started
+          by the application of <c><anno>Fun</anno></c> to the
+          empty list <c>[]</c> on <c><anno>Node</anno></c>. If
+          <c><anno>Node</anno></c> does not exist, a useless pid is
+          returned. Otherwise works like
           <seealso marker="#spawn_opt/4">spawn_opt/4</seealso>.</p>
       </desc>
     </func>
+
     <func>
       <name name="spawn_opt" arity="4"/>
-      <type name="priority_level" />
-      <fsummary>Create a new process with a function as entry point</fsummary>
+      <fsummary>Creates a new process with a function as entry point.</fsummary>
+      <type name="priority_level"/>
+      <type name="message_queue_data" />
+      <type name="spawn_opt_option" />
       <desc>
-        <p>Works exactly like
+        <p>Works as
           <seealso marker="#spawn/3">spawn/3</seealso>, except that an
           extra option list is given when creating the process.</p>
-        <p>If the option <c>monitor</c> is given, the newly created
-          process will be monitored and both the pid and reference for
-          the monitor will be returned.</p>
+        <p>If option <c>monitor</c> is given, the newly created
+          process is monitored, and both the pid and reference for
+          the monitor is returned.</p>
+        <p>The options are as follows:</p>
         <taglist>
           <tag><c>link</c></tag>
           <item>
@@ -5109,112 +5553,137 @@ true</pre>
           </item>
           <tag><c>monitor</c></tag>
           <item>
-            <p>Monitor the new process (just like
+            <p>Monitors the new process (like
               <seealso marker="#monitor/2">monitor/2</seealso> does).</p>
           </item>
-          <tag><c>{priority, <anno>Level</anno>}</c></tag>
+          <tag><c>{priority, <anno>Level</anno></c></tag>
           <item>
             <p>Sets the priority of the new process. Equivalent to
               executing
-              <seealso marker="#process_flag_priority">process_flag(priority, <anno>Level</anno>)</seealso> in the start function of the new process,
-              except that the priority will be set before the process is
-              selected for execution for the first time. For more information
-              on priorities see
-              <seealso marker="#process_flag_priority">process_flag(priority, Level)</seealso>.</p>
+              <seealso marker="#process_flag_priority">process_flag(priority,
+              <anno>Level</anno>)</seealso>
+              in the start function of the new process,
+              except that the priority is set before the process is
+              selected for execution for the first time. For more
+              information on priorities, see
+              <seealso marker="#process_flag_priority">process_flag(priority,
+              <anno>Level</anno>)</seealso>.</p>
           </item>
           <tag><c>{fullsweep_after, <anno>Number</anno>}</c></tag>
           <item>
-            <p>This option is only useful for performance tuning.
-              In general, you should not use this option unless you
-              know that there is problem with execution times and/or
-              memory consumption, and you should measure to make sure
-              that the option improved matters.
-              </p>
+            <p>Useful only for performance tuning. Do not use this
+              option unless you
+              know that there is problem with execution times or
+              memory consumption, and ensure
+              that the option improves matters.</p>
             <p>The Erlang runtime system uses a generational garbage
               collection scheme, using an "old heap" for data that has
               survived at least one garbage collection. When there is
               no more room on the old heap, a fullsweep garbage
-              collection will be done.</p>
-            <p>The <c>fullsweep_after</c> option makes it possible to
+              collection is done.</p>
+            <p>Option <c>fullsweep_after</c> makes it possible to
               specify the maximum number of generational collections
-              before forcing a fullsweep even if there is still room on
-              the old heap. Setting the number to zero effectively
-              disables the general collection algorithm, meaning that
+              before forcing a fullsweep, even if there is room on
+              the old heap. Setting the number to zero
+              disables the general collection algorithm, that is,
               all live data is copied at every garbage collection.</p>
-            <p>Here are a few cases when it could be useful to change
-              <c>fullsweep_after</c>. Firstly, if binaries that are no
-              longer used should be thrown away as soon as possible.
-              (Set <c><anno>Number</anno></c> to zero.) Secondly, a process that
-              mostly have short-lived data will be fullsweeped seldom
-              or never, meaning that the old heap will contain mostly
-              garbage. To ensure a fullsweep once in a while, set
-              <c><anno>Number</anno></c> to a suitable value such as 10 or 20.
-              Thirdly, in embedded systems with limited amount of RAM
-              and no virtual memory, one might want to preserve memory
-              by setting <c><anno>Number</anno></c> to zero. (The value may be set
-              globally, see
-              <seealso marker="#system_flag/2">erlang:system_flag/2</seealso>.)</p>
+            <p>A few cases when it can be useful to change
+              <c>fullsweep_after</c>:</p>
+            <list type="bulleted">
+              <item>If binaries that are no longer used are to be
+                thrown away as soon as possible. (Set
+                <c><anno>Number</anno></c> to zero.)
+              </item>
+              <item>A process that mostly have short-lived data is
+                fullsweeped seldom or never, that is, the old heap
+                contains mostly garbage. To ensure a fullsweep
+                occasionally, set <c><anno>Number</anno></c> to a
+                suitable value, such as 10 or 20.
+              </item>
+              <item>In embedded systems with a limited amount of RAM
+              and no virtual memory, you might want to preserve memory
+              by setting <c><anno>Number</anno></c> to zero.
+              (The value can be set globally, see
+              <seealso marker="#system_flag/2">erlang:system_flag/2</seealso>.)
+              </item>
+            </list>
           </item>
           <tag><c>{min_heap_size, <anno>Size</anno>}</c></tag>
           <item>
-            <p>This option is only useful for performance tuning.
-              In general, you should not use this option unless you
-              know that there is problem with execution times and/or
-              memory consumption, and you should measure to make sure
-              that the option improved matters.
-              </p>
-            <p>Gives a minimum heap size in words. Setting this value
-              higher than the system default might speed up some
+            <p>Useful only for performance tuning. Do not use this
+              option unless you know that there is problem with
+              execution times or memory consumption, and
+              ensure that the option improves matters.</p>
+            <p>Gives a minimum heap size, in words. Setting this value
+              higher than the system default can speed up some
               processes because less garbage collection is done.
-              Setting too high value, however, might waste memory and
-              slow down the system due to worse data locality.
-              Therefore, it is recommended to use this option only for
+              However, setting a too high value can waste memory and
+              slow down the system because of worse data locality.
+              Therefore, use this option only for
               fine-tuning an application and to measure the execution
               time with various <c><anno>Size</anno></c> values.</p>
           </item>
           <tag><c>{min_bin_vheap_size, <anno>VSize</anno>}</c></tag>
           <item>
-            <p>This option is only useful for performance tuning.
-              In general, you should not use this option unless you
-              know that there is problem with execution times and/or
-              memory consumption, and you should measure to make sure
-              that the option improved matters.
-              </p>
-	      <p>Gives a minimum binary virtual heap size in words. Setting this value
-              higher than the system default might speed up some
+            <p>Useful only for performance tuning. Do not use this
+              option unless you know that there is problem with
+              execution times or memory consumption, and
+              ensure that the option improves matters.</p>
+	      <p>Gives a minimum binary virtual heap size, in words.
+              Setting this value
+              higher than the system default can speed up some
               processes because less garbage collection is done.
-              Setting too high value, however, might waste memory.
-              Therefore, it is recommended to use this option only for
+              However, setting a too high value can waste memory.
+              Therefore, use this option only for
               fine-tuning an application and to measure the execution
               time with various <c><anno>VSize</anno></c> values.</p>
           </item>
-
+          <tag><c>{message_queue_data, <anno>MQD</anno>}</c></tag>
+          <item>
+            <p>Sets the state of the <c>message_queue_data</c> process
+	    flag. <c><anno>MQD</anno></c> should be either <c>off_heap</c>,
+	    <c>on_heap</c>, or <c>mixed</c>. The default
+	    <c>message_queue_data</c> process flag is determined by the
+	    <seealso marker="erl#+xmqd"><c>+xmqd</c></seealso> <c>erl</c>
+	    command line argument. For more information, see the
+	    documentation of
+            <seealso marker="#process_flag_message_queue_data"><c>process_flag(message_queue_data,
+	    <anno>MQD</anno>)</c></seealso>.</p>
+          </item>
         </taglist>
       </desc>
     </func>
+
     <func>
       <name name="spawn_opt" arity="5"/>
-      <type name="priority_level" />
-      <fsummary>Create a new process with a function as entry point on a given node</fsummary>
+      <fsummary>Creates a new process with a function as entry point on a given node.</fsummary>
+      <type name="priority_level"/>
+      <type name="message_queue_data" />
+      <type name="spawn_opt_option" />
       <desc>
-        <p>Returns the pid of a new process started by the application
-          of <c><anno>Module</anno>:<anno>Function</anno></c> to <c><anno>Args</anno></c> on <c>Node</c>. If
+        <p>Returns the process identifier (pid) of a new process started
+          by the application
+          of <c><anno>Module</anno>:<anno>Function</anno></c> to
+          <c><anno>Args</anno></c> on <c><anno>Node</anno></c>. If
           <c><anno>Node</anno></c> does not exist, a useless pid is returned.
           Otherwise works like
           <seealso marker="#spawn_opt/4">spawn_opt/4</seealso>.</p>
-	  <note><p>The <c>monitor</c> option is currently not supported by
+	  <note><p>Option <c>monitor</c> is not supported by
 	  <c>spawn_opt/5</c>.</p></note>
       </desc>
     </func>
+
     <func>
       <name name="split_binary" arity="2"/>
+      <fsummary>Splits a binary into two.</fsummary>
       <type_desc variable="Pos">0..byte_size(Bin)</type_desc>
-      <fsummary>Split a binary into two</fsummary>
       <desc>
-        <p>Returns a tuple containing the binaries which are the result
-          of splitting <c><anno>Bin</anno></c> into two parts at position <c><anno>Pos</anno></c>.
+        <p>Returns a tuple containing the binaries that are the result
+          of splitting <c><anno>Bin</anno></c> into two parts at
+          position <c><anno>Pos</anno></c>.
           This is not a destructive operation. After the operation,
-          there will be three binaries altogether.</p>
+          there are three binaries altogether.</p>
+        <p>Example:</p>
         <pre>
 > <input>B = list_to_binary("0123456789").</input>
 &lt;&lt;"0123456789">>
@@ -5228,31 +5697,35 @@ true</pre>
 7</pre>
       </desc>
     </func>
+
     <func>
       <name name="start_timer" arity="4"/>
-      <fsummary>Start a timer</fsummary>
+      <fsummary>Starts a timer.</fsummary>
       <desc>
         <p>
 	  Starts a timer. When the timer expires, the message
           <c>{timeout, <anno>TimerRef</anno>, <anno>Msg</anno>}</c>
-	  will be sent to the process identified by
+	  is sent to the process identified by
 	  <c><anno>Dest</anno></c>.
 	</p>
-	<p>Currently available <c><anno>Option</anno></c>s:</p>
+	<p>Available <c><anno>Option</anno></c>s:</p>
         <taglist>
-          <tag><c>{abs, Abs}</c></tag>
+          <tag><c>{abs, false}</c></tag>
           <item>
 	    <p>
-	      Absolute <c><anno>Time</anno></c> value. <c>Abs</c>
-	      defaults to <c>false</c> which means that the
-	      <c><anno>Time</anno></c> value will be interpreted
-	      as a time in milli-seconds relative current
+	      This is the default. It means the
+	      <c><anno>Time</anno></c> value is interpreted
+	      as a time in milli-seconds <em>relative</em> current
 	      <seealso marker="time_correction#Erlang_Monotonic_Time">Erlang
-	      monotonic time</seealso>. When <c>Abs</c> is set to
-	      <c>true</c>, the <c><anno>Time</anno></c> value will
-	      be interpreted as an absolute Erlang monotonic time of
-	      milli-seconds
-	      <seealso marker="#type_time_unit">time unit</seealso>.
+	      monotonic time</seealso>.
+	    </p>
+	  </item>
+          <tag><c>{abs, true}</c></tag>
+          <item>
+	    <p>
+	      Absolute <c><anno>Time</anno></c> value. The
+	      <c><anno>Time</anno></c> value is interpreted as an
+	      absolute Erlang monotonic time in milli-seconds.
 	    </p>
 	  </item>
         </taglist>
@@ -5260,7 +5733,7 @@ true</pre>
 	  More <c><anno>Option</anno></c>s may be added in the future.
 	</p>
 	<p>
-	  The absolute point in time that the timer is set to expire on
+	  The absolute point in time, the timer is set to expire on,
 	  has to be in the interval
 	  <c>[</c><seealso marker="#system_info_start_time"><c>erlang:system_info(start_time)</c></seealso><c>,
 	  </c><seealso marker="#system_info_end_time"><c>erlang:system_info(end_time)</c></seealso><c>]</c>.
@@ -5268,21 +5741,21 @@ true</pre>
 	  is not allowed to be negative.
 	</p>
 	<p>
-	  If <c><anno>Dest</anno></c> is a <c>pid()</c>, it has to
+	  If <c><anno>Dest</anno></c> is a <c>pid()</c>, it must
 	  be a <c>pid()</c> of a process created on the current
 	  runtime system instance. This process may or may not
 	  have terminated. If <c><anno>Dest</anno></c> is an
-	  <c>atom()</c>, it will be interpreted as the name of a
+	  <c>atom()</c>, it is interpreted as the name of a
 	  locally registered process. The process referred to by the
 	  name is looked up at the time of timer expiration. No error
 	  is given if the name does not refer to a process.
 	</p>
         <p>
-	  If <c><anno>Dest</anno></c> is a <c>pid()</c>, the timer will
-	  be automatically canceled if the process referred to by the
-	  <c>pid()</c> is not alive, or when the process exits. This
-	  feature was introduced in erts version 5.4.11. Note that
-	  timers will not be automatically canceled when
+	  If <c><anno>Dest</anno></c> is a <c>pid()</c>, the timer is
+	  automatically canceled if the process referred to by the
+	  <c>pid()</c> is not alive, or if the process exits. This
+	  feature was introduced in ERTS version 5.4.11. Notice that
+	  timers are not automatically canceled when
 	  <c><anno>Dest</anno></c> is an <c>atom()</c>.
 	</p>
         <p>See also
@@ -5290,141 +5763,203 @@ true</pre>
           <seealso marker="#cancel_timer/2"><c>erlang:cancel_timer/2</c></seealso>,
           and
           <seealso marker="#read_timer/2"><c>erlang:read_timer/2</c></seealso>.</p>
-        <p>Failure: <c>badarg</c> if the arguments does not satisfy
-          the requirements specified above.</p>
+        <p>Failure: <c>badarg</c> if the arguments do not satisfy
+          the requirements specified here.</p>
       </desc>
     </func>
+
     <func>
       <name name="start_timer" arity="3"/>
-      <fsummary>Start a timer</fsummary>
+      <fsummary>Starts a timer.</fsummary>
       <desc>
         <p>Starts a timer. The same as calling
 	<seealso marker="#start_timer/4"><c>erlang:start_timer(<anno>Time</anno>,
 	<anno>Dest</anno>, <anno>Msg</anno>, [])</c></seealso>.</p>
       </desc>
     </func>
+
     <func>
       <name name="statistics" arity="1" clause_i="1"/>
-      <fsummary>Information about context switches</fsummary>
-      <desc>
-        <p><c><anno>ContextSwitches</anno></c> is the total number of context
-          switches since the system started.</p>
+      <fsummary>Information about active processes and ports.</fsummary>
+      <desc><marker id="statistics_active_tasks"></marker>
+	<p>
+	  Returns a list where each element represents the amount
+	  of active processes and ports on each run queue and its
+	  associated scheduler. That is, the number of processes and
+	  ports that are ready to run, or are currently running. The
+	  element location in the list corresponds to the scheduler
+	  and its run queue. The first element corresponds to scheduler
+	  number 1 and so on. The information is <em>not</em> gathered
+	  atomically. That is, the result is not necessarily a
+	  consistent snapshot of the state, but instead quite
+	  efficiently gathered. See also,
+	  <seealso marker="#statistics_total_active_tasks"><c>statistics(total_active_tasks)</c></seealso>,
+	  <seealso marker="#statistics_run_queue_lengths"><c>statistics(run_queue_lengths)</c></seealso>, and
+	  <seealso marker="#statistics_total_run_queue_lengths"><c>statistics(total_run_queue_lengths)</c></seealso>.
+	</p>
       </desc>
     </func>
+
     <func>
       <name name="statistics" arity="1" clause_i="2"/>
-      <fsummary>Information about exact reductions</fsummary>
+      <fsummary>Information about context switches.</fsummary>
       <desc>
-        <marker id="statistics_exact_reductions"></marker>
-			<note><p><c>statistics(exact_reductions)</c> is
-					a more expensive operation than
-					<seealso marker="#statistics_reductions">statistics(reductions)</seealso>
-					especially on an Erlang machine with SMP support.</p>
-			</note>
+        <p>Returns the total number of context switches since the
+          system started.</p>
       </desc>
     </func>
+
     <func>
       <name name="statistics" arity="1" clause_i="3"/>
-      <fsummary>Information about garbage collection</fsummary>
+      <fsummary>Information about exact reductions.</fsummary>
       <desc>
-        <p>This information may not be valid for all implementations.</p>
-	  <pre>
-> <input>statistics(garbage_collection).</input>
-{85,23961,0}
-</pre>
+        <marker id="statistics_exact_reductions"></marker>
+        <p>Returns the number of exact reductions.</p>
+      <note><p><c>statistics(exact_reductions)</c> is
+        a more expensive operation than
+        <seealso marker="#statistics_reductions">statistics(reductions)</seealso>,
+        especially on an Erlang machine with SMP support.</p>
+      </note>
       </desc>
     </func>
+
     <func>
       <name name="statistics" arity="1" clause_i="4"/>
-      <fsummary>Information about io</fsummary>
+      <fsummary>Information about garbage collection.</fsummary>
       <desc>
-        <p><c><anno>Input</anno></c> is the total number of bytes received
-          through ports, and <c><anno>Output</anno></c> is the total number of
-          bytes output to ports.</p>
+        <p>Returns information about garbage collection, for example:</p>
+	  <pre>
+> <input>statistics(garbage_collection).</input>
+{85,23961,0}</pre>
+        <p>This information can be invalid for some implementations.</p>
       </desc>
     </func>
+
     <func>
       <name name="statistics" arity="1" clause_i="5"/>
-      <fsummary>Information about reductions</fsummary>
+      <fsummary>Information about I/O.</fsummary>
       <desc>
-        <marker id="statistics_reductions"></marker>
-			<note>
-				<p>Since erts-5.5 (OTP release R11B)
-					this value does not include reductions performed in current
-					time slices of currently scheduled processes. If an
-					exact value is wanted, use
-					<seealso marker="#statistics_exact_reductions">statistics(exact_reductions)</seealso>.</p>
-			</note>
-        <pre>
-> <input>statistics(reductions).</input>
-{2046,11}
-</pre>
+        <p>Returns <c><anno>Input</anno></c>,
+           which is the total number of bytes
+           received through ports, and <c><anno>Output</anno></c>,
+           which is the total number of bytes output to ports.</p>
       </desc>
     </func>
+
     <func>
       <name name="statistics" arity="1" clause_i="6"/>
-      <fsummary>Information about the run-queue</fsummary>
+      <fsummary>Information about reductions.</fsummary>
       <desc>
-        <p>Returns the total length of the run queues, that is, the number
-          of processes that are ready to run on all available run queues.</p>
+        <marker id="statistics_reductions"></marker>
+        <p>Returns information about reductions, for example:</p>
+        <pre>
+> <input>statistics(reductions).</input>
+{2046,11}</pre>
+        <note><p>As from <c>ERTS</c> 5.5 (OTP R11B),
+          this value does not include reductions performed in current
+          time slices of currently scheduled processes. If an
+          exact value is wanted, use
+          <seealso marker="#statistics_exact_reductions">statistics(exact_reductions)</seealso>.</p>
+        </note>
       </desc>
     </func>
+
     <func>
       <name name="statistics" arity="1" clause_i="7"/>
-      <fsummary>Information about run-time</fsummary>
-      <desc>
-        <p>Note that the run-time is the sum of the run-time for all
-        threads in the Erlang run-time system and may therefore be greater
-        than the wall-clock time. The time is returned in milliseconds.</p>
-        <pre>
-> <input>statistics(runtime).</input>
-{1690,1620}
-</pre>
+      <fsummary>Information about the run-queues.</fsummary>
+      <desc><marker id="statistics_run_queue"></marker>
+        <p>
+	  Returns the total length of the run-queues. That is, the number
+          of processes and ports that are ready to run on all available
+	  run-queues. The information is gathered atomically. That
+	  is, the result is a consistent snapshot of the state, but
+	  this operation is much more expensive compared to
+	  <seealso marker="#statistics_total_run_queue_lengths"><c>statistics(total_run_queue_lengths)</c></seealso>.
+	  This especially when a large amount of schedulers is used.
+	</p>
       </desc>
     </func>
+
     <func>
       <name name="statistics" arity="1" clause_i="8"/>
-      <fsummary>Information about each schedulers work time</fsummary>
-      <desc>
-		  <marker id="statistics_scheduler_wall_time"></marker>
-		  <p>
-			  Returns a list of tuples with <c>{<anno>SchedulerId</anno>,
-				  <anno>ActiveTime</anno>, <anno>TotalTime</anno>}</c>, where
-			  <c>SchedulerId</c> is an integer id of the scheduler, <c>ActiveTime</c> is
-			  the duration the scheduler has been busy, <c>TotalTime</c> is the total time duration since
-			  <seealso marker="#system_flag_scheduler_wall_time">scheduler_wall_time</seealso>
-			  activation. The time unit is not defined and may be subject to change
-			  between releases, operating systems and system restarts.
-			  <c>scheduler_wall_time</c> should only be used to calculate relative
-			  values for scheduler-utilization. <c>ActiveTime</c> can never exceed <c>TotalTime</c>.
-		  </p>
-
-		  <p>The definition of a busy scheduler is when it is not idle or not
-			  scheduling (selecting) a process or port, meaning; executing process
-			  code, executing linked-in-driver or NIF code, executing
-			  built-in-functions or any other runtime handling, garbage collecting
-			  or handling any other memory management. Note, a scheduler may also be
-			  busy even if the operating system has scheduled out the scheduler
-			  thread.
-		  </p>
+      <fsummary>Information about the run-queue lengths.</fsummary>
+      <desc><marker id="statistics_run_queue_lengths"></marker>
+	<p>
+	  Returns a list where each element represents the amount
+	  of processes and ports ready to run for each run queue. The
+	  element location in the list corresponds to the run queue
+	  of a scheduler. The first element corresponds to the run
+	  queue of scheduler number 1 and so on. The information is
+	  <em>not</em> gathered atomically. That is, the result is
+	  not necessarily a consistent snapshot of the state, but
+	  instead quite efficiently gathered. See also,
+	  <seealso marker="#statistics_total_run_queue_lengths"><c>statistics(total_run_queue_lengths)</c></seealso>,
+	  <seealso marker="#statistics_active_tasks"><c>statistics(active_tasks)</c></seealso>, and
+	  <seealso marker="#statistics_total_active_tasks"><c>statistics(total_active_tasks)</c></seealso>.
+	</p>
+      </desc>
+    </func>
 
-		  <p>
-			  Returns <c>undefined</c> if the system flag
-			  <seealso marker="#system_flag_scheduler_wall_time">scheduler_wall_time</seealso>
-			  is turned off.
-		  </p>
+    <func>
+      <name name="statistics" arity="1" clause_i="9"/>
+      <fsummary>Information about runtime.</fsummary>
+      <desc>
+        <p>Returns information about runtime, in milliseconds.</p>
+        <p>This is the sum of the runtime for all threads
+        in the Erlang runtime system and can therefore be greater
+        than the wall clock time.</p>
+        <p>Example:</p>
+        <pre>
+> <input>statistics(runtime).</input>
+{1690,1620}</pre>
+      </desc>
+    </func>
 
-		  <p>The list of scheduler information is unsorted and may appear in different order
-			  between calls.
-		  </p>
-		  <p>Using <c>scheduler_wall_time</c> to calculate scheduler utilization.</p>
+    <func>
+      <name name="statistics" arity="1" clause_i="10"/>
+      <fsummary>Information about each schedulers work time.</fsummary>
+      <desc>
+      <marker id="statistics_scheduler_wall_time"></marker>
+        <p>Returns a list of tuples with
+          <c>{<anno>SchedulerId</anno>, <anno>ActiveTime</anno>,
+          <anno>TotalTime</anno>}</c>, where
+          <c><anno>SchedulerId</anno></c> is an integer ID of the scheduler,
+          <c><anno>ActiveTime</anno></c> is
+          the duration the scheduler has been busy, and
+          <c><anno>TotalTime</anno></c> is the total time duration since
+          <seealso marker="#system_flag_scheduler_wall_time">scheduler_wall_time</seealso>
+          activation. The time unit is undefined and can be subject
+          to change between releases, OSs, and system restarts.
+          <c>scheduler_wall_time</c> is only to be used to
+          calculate relative values for scheduler-utilization.
+          <c><anno>ActiveTime</anno></c> can never exceed
+          <c><anno>TotalTime</anno></c>.</p>
+        <p>The definition of a busy scheduler is when it is not idle
+          and is not scheduling (selecting) a process or port,
+          that is:</p>
+        <list type="bulleted">
+          <item>Executing process code</item>
+          <item>Executing linked-in-driver or NIF code</item>
+          <item>Executing built-in-functions, or any other runtime
+            handling</item>
+          <item>Garbage collecting</item>
+          <item>Handling any other memory management</item>
+        </list>
+          <p>Notice that a scheduler can also be busy even if the
+            OS has scheduled out the scheduler thread.</p>
+        <p>Returns <c>undefined</c> if system flag
+          <seealso marker="#system_flag_scheduler_wall_time">scheduler_wall_time</seealso>
+          is turned off.</p>
+        <p>The list of scheduler information is unsorted and can
+          appear in different order between calls.</p>
+        <p>Using <c>scheduler_wall_time</c> to calculate scheduler-utilization:</p>
 <pre>
 > <input>erlang:system_flag(scheduler_wall_time, true).</input>
 false
 > <input>Ts0 = lists:sort(erlang:statistics(scheduler_wall_time)), ok.</input>
-ok
-</pre>
-		<p>Some time later we will take another snapshot and calculate scheduler-utilization per scheduler.</p>
+ok</pre>
+        <p>Some time later the user takes another snapshot and calculates
+          scheduler-utilization per scheduler, for example:</p>
 <pre>
 > <input>Ts1 = lists:sort(erlang:statistics(scheduler_wall_time)), ok.</input>
 ok
@@ -5437,86 +5972,127 @@ ok
  {5,0.9717956667018103},
  {6,0.9739235846420741},
  {7,0.973237033077876},
- {8,0.9741297293248656}]
-</pre>
-		<p>Using the same snapshots to calculate a total scheduler-utilization.</p>
+ {8,0.9741297293248656}]</pre>
+		<p>Using the same snapshots to calculate a total scheduler-utilization:</p>
 <pre>
 > <input>{A, T} = lists:foldl(fun({{_, A0, T0}, {_, A1, T1}}, {Ai,Ti}) ->
 	{Ai + (A1 - A0), Ti + (T1 - T0)} end, {0, 0}, lists:zip(Ts0,Ts1)), A/T.</input>
-0.9769136803764825
-</pre>
+0.9769136803764825</pre>
         <note>
-          <p><c>scheduler_wall_time</c> is by default disabled. Use <c>erlang:system_flag(scheduler_wall_time, true)</c> to enable it. </p>
+          <p><c>scheduler_wall_time</c> is by default disabled. To
+            enable it, use
+            <c>erlang:system_flag(scheduler_wall_time, true)</c>.</p>
         </note>
       </desc>
     </func>
+
     <func>
-      <name name="statistics" arity="1" clause_i="9"/>
-      <fsummary>Information about wall-clock</fsummary>
+      <name name="statistics" arity="1" clause_i="11"/>
+      <fsummary>Information about active processes and ports.</fsummary>
+      <desc><marker id="statistics_total_active_tasks"></marker>
+	<p>
+	  Returns the total amount of active processes and ports in
+	  the system. That is, the number of processes and ports that
+	  are ready to run, or are currently running. The information
+	  is <em>not</em> gathered atomically. That is, the result
+	  is not necessarily a consistent snapshot of the state, but
+	  instead quite efficiently gathered. See also,
+	  <seealso marker="#statistics_active_tasks"><c>statistics(active_tasks)</c></seealso>,
+	  <seealso marker="#statistics_run_queue_lengths"><c>statistics(run_queue_lengths)</c></seealso>, and
+	  <seealso marker="#statistics_total_run_queue_lengths"><c>statistics(total_run_queue_lengths)</c></seealso>.
+	</p>
+      </desc>
+    </func>
+
+    <func>
+      <name name="statistics" arity="1" clause_i="12"/>
+      <fsummary>Information about the run-queue lengths.</fsummary>
+      <desc><marker id="statistics_total_run_queue_lengths"></marker>
+        <p>
+	  Returns the total length of the run-queues. That is, the number
+          of processes and ports that are ready to run on all available
+	  run-queues. The information is <em>not</em> gathered atomically.
+	  That is, the result is not necessarily a consistent snapshot of
+	  the state, but much more efficiently gathered compared to
+	  <seealso marker="#statistics_run_queue"><c>statistics(run_queue)</c></seealso>.
+	  See also,
+	  <seealso marker="#statistics_run_queue_lengths"><c>statistics(run_queue_lengths)</c></seealso>,
+	  <seealso marker="#statistics_total_active_tasks"><c>statistics(total_active_tasks)</c></seealso>, and
+	  <seealso marker="#statistics_active_tasks"><c>statistics(active_tasks)</c></seealso>.
+	</p>
+      </desc>
+    </func>
+
+    <func>
+      <name name="statistics" arity="1" clause_i="13"/>
+      <fsummary>Information about wall clock.</fsummary>
       <desc>
-        <p><c>wall_clock</c> can be used in the same manner as
+        <p>Returns information about wall clock. <c>wall_clock</c> can
+          be used in the same manner as
           <c>runtime</c>, except that real time is measured as
           opposed to runtime or CPU time.</p>
       </desc>
     </func>
+
     <func>
       <name name="suspend_process" arity="2"/>
-      <fsummary>Suspend a process</fsummary>
+      <fsummary>Suspends a process.</fsummary>
       <desc>
         <p>Increases the suspend count on the process identified by
-	<c><anno>Suspendee</anno></c> and puts it in the suspended state if it isn't
-	already in the suspended state. A suspended process will not be
-	scheduled for execution until the process has been resumed.
-	</p>
-
+	<c><anno>Suspendee</anno></c> and puts it in the suspended
+        state if it is not
+	already in that state. A suspended process will not be
+	scheduled for execution until the process has been resumed.</p>
 	<p>A process can be suspended by multiple processes and can
 	be suspended multiple times by a single process. A suspended
-	process will not leave the suspended state until its suspend 
-	count reach zero. The suspend count of <c><anno>Suspendee</anno></c>
-	is decreased when
+	process does not leave the suspended state until its suspend 
+	count reaches zero. The suspend count of
+        <c><anno>Suspendee</anno></c> is decreased when
 	<seealso marker="#resume_process/1">erlang:resume_process(<anno>Suspendee</anno>)</seealso>
 	is called by the same process that called
-	<c>erlang:suspend_process(<anno>Suspendee</anno>)</c>. All increased suspend
-	counts on other processes acquired by a process will automatically be
+	<c>erlang:suspend_process(<anno>Suspendee</anno>)</c>.
+        All increased suspend
+	counts on other processes acquired by a process are automatically
 	decreased when the process terminates.</p>
-
-	<p>Currently the following options (<c><anno>Opt</anno></c>s) are available:</p>
+	<p>The options (<c><anno>Opt</anno></c>s) are as follows:</p>
         <taglist>
           <tag><c>asynchronous</c></tag>
           <item>
 	  A suspend request is sent to the process identified by
-	  <c><anno>Suspendee</anno></c>. <c><anno>Suspendee</anno></c> will eventually suspend
-	  unless it is resumed before it was able to suspend. The caller
-	  of <c>erlang:suspend_process/2</c> will return immediately,
-	  regardless of whether the <c><anno>Suspendee</anno></c> has suspended yet
-	  or not. Note that the point in time when the <c><anno>Suspendee</anno></c>
-	  will actually suspend cannot be deduced from other events
-	  in the system. The only guarantee given is that the
-	  <c><anno>Suspendee</anno></c> will <em>eventually</em> suspend (unless it
-	  is resumed). If the <c>asynchronous</c> option has <em>not</em>
-	  been passed, the caller of <c>erlang:suspend_process/2</c> will
-	  be blocked until the <c><anno>Suspendee</anno></c> has actually suspended.
+	  <c><anno>Suspendee</anno></c>. <c><anno>Suspendee</anno></c>
+          eventually suspends
+	  unless it is resumed before it could suspend. The caller
+	  of <c>erlang:suspend_process/2</c> returns immediately,
+	  regardless of whether <c><anno>Suspendee</anno></c> has
+          suspended yet or not. The point in time when
+          <c><anno>Suspendee</anno></c> suspends cannot be deduced
+          from other events in the system. It is only guaranteed that
+          <c><anno>Suspendee</anno></c> <em>eventually</em> suspends
+          (unless it
+	  is resumed). If option <c>asynchronous</c> has <em>not</em>
+	  been passed, the caller of <c>erlang:suspend_process/2</c> is
+	  blocked until <c><anno>Suspendee</anno></c> has suspended.
 	  </item>
           <tag><c>unless_suspending</c></tag>
           <item>
-	  The process identified by <c><anno>Suspendee</anno></c> will be suspended
-	  unless the calling process already is suspending the
-	  <c><anno>Suspendee</anno></c>. If <c>unless_suspending</c> is combined
-	  with the <c>asynchronous</c> option, a suspend request will be
-	  sent unless the calling process already is suspending the
-	  <c><anno>Suspendee</anno></c> or if a suspend request already has been sent
-	  and is in transit. If the calling process already is suspending
-	  the <c><anno>Suspendee</anno></c>, or if combined with the <c>asynchronous</c>
-	  option and a send request already is in transit,
-	  <c>false</c> is returned and the suspend count on <c><anno>Suspendee</anno></c>
-	  will remain unchanged.
+	  The process identified by <c><anno>Suspendee</anno></c> is
+          suspended unless the calling process already is suspending
+	  <c><anno>Suspendee</anno></c>.
+          If <c>unless_suspending</c> is combined
+	  with option <c>asynchronous</c>, a suspend request is
+	  sent unless the calling process already is suspending
+	  <c><anno>Suspendee</anno></c> or if a suspend request
+          already has been sent and is in transit. If the calling
+          process already is suspending <c><anno>Suspendee</anno></c>,
+          or if combined with option <c>asynchronous</c>
+	  and a send request already is in transit,
+	  <c>false</c> is returned and the suspend count on
+          <c><anno>Suspendee</anno></c> remains unchanged.
 	  </item>
         </taglist>
-
 	<p>If the suspend count on the process identified by
-	<c><anno>Suspendee</anno></c> was increased, <c>true</c> is returned; otherwise,
-	<c>false</c> is returned.</p>
-
+	<c><anno>Suspendee</anno></c> is increased, <c>true</c>
+        is returned, otherwise <c>false</c>.</p>
         <warning>
           <p>This BIF is intended for debugging only.</p>
         </warning>
@@ -5524,310 +6100,322 @@ ok
         <taglist>
           <tag><c>badarg</c></tag>
           <item>
-	  If <c><anno>Suspendee</anno></c> isn't a process identifier.
+	  If <c><anno>Suspendee</anno></c> is not a process identifier.
 	  </item>
           <tag><c>badarg</c></tag>
           <item>
-	  If the process identified by <c><anno>Suspendee</anno></c> is same the process as
-	  the process calling <c>erlang:suspend_process/2</c>.
+	  If the process identified by <c><anno>Suspendee</anno></c>
+          is the same process
+	  as the process calling <c>erlang:suspend_process/2</c>.
 	  </item>
           <tag><c>badarg</c></tag>
           <item>
-	  If the process identified by <c><anno>Suspendee</anno></c> is not alive.
+	  If the process identified by <c><anno>Suspendee</anno></c>
+          is not alive.
 	  </item>
           <tag><c>badarg</c></tag>
           <item>
-	  If the process identified by <c><anno>Suspendee</anno></c> resides on another node.
+	  If the process identified by <c><anno>Suspendee</anno></c>
+          resides on another node.
 	  </item>
           <tag><c>badarg</c></tag>
           <item>
-	  If <c><anno>OptList</anno></c> isn't a proper list of valid <c><anno>Opt</anno></c>s.
+	  If <c><anno>OptList</anno></c> is not a proper list of valid
+          <c><anno>Opt</anno></c>s.
 	  </item>
           <tag><c>system_limit</c></tag>
           <item>
-	  If the process identified by <c><anno>Suspendee</anno></c> has been suspended more
-	  times by the calling process than can be represented by the
-	  currently used internal data structures. The current system limit
-	  is larger than 2 000 000 000 suspends, and it will never be less
-	  than that.
+	  If the process identified by <c><anno>Suspendee</anno></c>
+          has been suspended
+	  more times by the calling process than can be represented by the
+	  currently used internal data structures. The system limit is
+	  higher than 2,000,000,000 suspends and will never be lower.
 	  </item>
         </taglist>
       </desc>
     </func>
+
     <func>
       <name name="suspend_process" arity="1"/>
-      <fsummary>Suspend a process</fsummary>
+      <fsummary>Suspends a process.</fsummary>
       <desc>
-        <p>Suspends the process identified by <c><anno>Suspendee</anno></c>. The
-	same as calling
-	<seealso marker="#suspend_process/2">erlang:suspend_process(<anno>Suspendee</anno>, [])</seealso>. For more information see the documentation of <seealso marker="#suspend_process/2">erlang:suspend_process/2</seealso>.
-	</p>
+        <p>Suspends the process identified by
+        <c><anno>Suspendee</anno></c>. The same as calling
+	<seealso marker="#suspend_process/2">erlang:suspend_process(<anno>Suspendee</anno>,
+        [])</seealso>.</p>
         <warning>
           <p>This BIF is intended for debugging only.</p>
         </warning>
       </desc>
     </func>
+
     <func>
       <name name="system_flag" arity="2" clause_i="1"/>
-      <fsummary>Set system flag backtrace_depth</fsummary>
+      <fsummary>Sets system flag <c>backtrace_depth</c>.</fsummary>
       <desc>
         <p>Sets the maximum depth of call stack back-traces in the
           exit reason element of <c>'EXIT'</c> tuples.</p>
         <p>Returns the old value of the flag.</p>
       </desc>
     </func>
+
     <func>
       <name name="system_flag" arity="2" clause_i="2"/>
+      <fsummary>Sets system flag <c>cpu_topology</c>.</fsummary>
       <type name="cpu_topology"/>
       <type name="level_entry"/>
       <type name="level_tag"/>
       <type name="sub_level"/>
       <type name="info_list"/>
-      <fsummary>Set system flag cpu_topology</fsummary>
       <desc>
         <warning>
           <p><marker id="system_flag_cpu_topology"></marker>
-          This argument is <em>deprecated</em> and
-          scheduled for removal in erts-5.10/OTP-R16. Instead of using
-          this argument you are advised to use the <c>erl</c> command
-          line argument <seealso marker="erts:erl#+sct">+sct</seealso>.
-          When this argument has been removed a final CPU topology to use
-          will be determined at emulator boot time.</p>
+          This argument is <em>deprecated</em> and scheduled for
+          removal in <c>ERTS</c> 5.10/OTP R16. Instead of using this
+          argument, use command-line argument
+          <seealso marker="erts:erl#+sct">+sct</seealso> in
+          <c>erl(1)</c>.</p>
+          <p>When this argument is removed, a final CPU topology
+          to use is determined at emulator boot time.</p>
         </warning>
-        <p>Sets the user defined <c><anno>CpuTopology</anno></c>. The user defined
-           CPU topology will override any automatically detected
-           CPU topology. By passing <c>undefined</c> as <c><anno>CpuTopology</anno></c>
-           the system will revert back to the CPU topology automatically
+        <p>Sets the user-defined <c><anno>CpuTopology</anno></c>.
+           The user-defined
+           CPU topology overrides any automatically detected
+           CPU topology. By passing <c>undefined</c> as
+           <c><anno>CpuTopology</anno></c>,
+           the system reverts to the CPU topology automatically
            detected. The returned value equals the value returned
            from <c>erlang:system_info(cpu_topology)</c> before the
-           change was made.
-        </p>
+           change was made.</p>
         <p>Returns the old value of the flag.</p>
         <p>The CPU topology is used when binding schedulers to logical
            processors. If schedulers are already bound when the CPU
-           topology is changed, the schedulers will be sent a request
-           to rebind according to the new CPU topology.
-        </p>
-        <p>The user defined CPU topology can also be set by passing
-           the <seealso marker="erts:erl#+sct">+sct</seealso> command
-           line argument to <c>erl</c>.
-        </p>
-        <p>For information on the <c><anno>CpuTopology</anno></c> type
-           and more, see the documentation of
-           <seealso marker="#system_info_cpu_topology">erlang:system_info(cpu_topology)</seealso>,
-           and the <c>erl</c> <seealso marker="erts:erl#+sct">+sct</seealso>
-           and <seealso marker="erts:erl#+sbt">+sbt</seealso>
-           command line flags.
-        </p>
+           topology is changed, the schedulers are sent a request
+           to rebind according to the new CPU topology.</p>
+        <p>The user-defined CPU topology can also be set by passing
+           command-line argument
+           <seealso marker="erts:erl#+sct">+sct</seealso> to
+           <c>erl(1)</c>.</p>
+        <p>For information on type <c><anno>CpuTopology</anno></c>
+           and more, see
+           <seealso marker="#system_info_cpu_topology">erlang:system_info(cpu_topology)</seealso>
+           as well as the command-line flags
+           <seealso marker="erts:erl#+sct">+sct</seealso> and
+           <seealso marker="erts:erl#+sbt">+sbt</seealso> in
+           <c>erl(1)</c>.</p>
       </desc>
     </func>
+
     <func>
       <name name="system_flag" arity="2" clause_i="3"/>
-      <fsummary>Set system flag dirty CPU schedulers online</fsummary>
+      <fsummary>Sets <c>system_flag_dirty_cpu_schedulers_online</c>.</fsummary>
       <desc>
         <p><marker id="system_flag_dirty_cpu_schedulers_online"></marker>
-        Sets the amount of dirty CPU schedulers online. Valid range is
-        <![CDATA[1 <= DirtyCPUSchedulersOnline <= N]]> where <c>N</c> is the
-        lesser of the return values of <c>erlang:system_info(dirty_cpu_schedulers)</c> and
-        <c>erlang:system_info(schedulers_online)</c>.
-        </p>
+        Sets the number of dirty CPU schedulers online. Range is
+        <![CDATA[1 <= DirtyCPUSchedulersOnline <= N]]>, where <c>N</c>
+        is the smallest of the return values of
+        <c>erlang:system_info(dirty_cpu_schedulers)</c> and
+        <c>erlang:system_info(schedulers_online)</c>.</p>
         <p>Returns the old value of the flag.</p>
-        <p>Note that the number of dirty CPU schedulers online may change if the number of
-        schedulers online changes. For example, if there are 12 schedulers and all are
-        online, and 6 dirty CPU schedulers, all online as well, and <c>system_flag/2</c>
-        is used to set the number of schedulers online to 6, then the number of dirty
-        CPU schedulers online is automatically decreased by half as well, down to 3.
-        Similarly, the number of dirty CPU schedulers online increases proportionally
-        to increases in the number of schedulers online.</p>
-        <p><em>Note that the dirty schedulers functionality is experimental</em>, and
-        that you have to enable support for dirty schedulers when building OTP in order
-        to try out the functionality.</p>
-        <p>For more information see
+        <p>The number of dirty CPU schedulers online can change if the
+          number of schedulers online changes. For example, if 12
+          schedulers and 6 dirty CPU schedulers are online, and
+          <c>system_flag/2</c> is used to set the number of
+          schedulers online to 6, then the number of dirty CPU
+          schedulers online is automatically decreased by half as well,
+          down to 3. Similarly, the number of dirty CPU schedulers
+          online increases proportionally to increases in the number of
+          schedulers online.</p>
+        <note><p>The dirty schedulers functionality is experimental.
+          Enable support for dirty schedulers when building OTP to
+          try out the functionality.</p>
+        </note>
+        <p>For more information, see
         <seealso marker="#system_info_dirty_cpu_schedulers">erlang:system_info(dirty_cpu_schedulers)</seealso>
         and
-        <seealso marker="#system_info_dirty_cpu_schedulers_online">erlang:system_info(dirty_cpu_schedulers_online)</seealso>.
-        </p>
+        <seealso marker="#system_info_dirty_cpu_schedulers_online">erlang:system_info(dirty_cpu_schedulers_online)</seealso>.</p>
       </desc>
     </func>
+
     <func>
       <name name="system_flag" arity="2" clause_i="4"/>
-      <fsummary>Set system flag fullsweep_after</fsummary>
+      <fsummary>Sets system flag <c>fullsweep_after</c>.</fsummary>
       <desc>
-        <p><c><anno>Number</anno></c> is a non-negative integer which indicates
+        <p>Sets system flag <c>fullsweep_after</c>.
+          <c><anno>Number</anno></c> is a non-negative integer indicating
           how many times generational garbage collections can be
           done without forcing a fullsweep collection. The value
-          applies to new processes; processes already running are
+          applies to new processes, while processes already running are
           not affected.</p>
         <p>Returns the old value of the flag.</p>
         <p>In low-memory systems (especially without virtual
-          memory), setting the value to 0 can help to conserve
+          memory), setting the value to <c>0</c> can help to conserve
           memory.</p>
-        <p>An alternative way to set this value is through the
-          (operating system) environment variable
-          <c>ERL_FULLSWEEP_AFTER</c>.</p>
+        <p>This value can also be set through (OS)
+          environment variable <c>ERL_FULLSWEEP_AFTER</c>.</p>
       </desc>
     </func>
+
     <func>
       <name name="system_flag" arity="2" clause_i="5"/>
-      <fsummary>Set system flag min_heap_size</fsummary>
-      <desc>
-        <p>Sets the default minimum heap size for processes. The
-          size is given in words. The new <c>min_heap_size</c> only
-          effects processes spawned after the change of
-          <c>min_heap_size</c> has been made.
-          The <c>min_heap_size</c> can be set for individual
-          processes by use of
+      <fsummary>Sets system flag <c>min_heap_size</c>.</fsummary>
+      <desc>
+        <p>Sets the default minimum heap size for processes. The size
+          is given in words. The new <c>min_heap_size</c> effects
+          only processes spawned after the change of
+          <c>min_heap_size</c> has been made. <c>min_heap_size</c>
+          can be set for individual processes by using
           <seealso marker="#spawn_opt/4">spawn_opt/N</seealso> or
-          <seealso marker="#process_flag/2">process_flag/2</seealso>. </p>
+          <seealso marker="#process_flag/2">process_flag/2</seealso>.</p>
         <p>Returns the old value of the flag.</p>
       </desc>
     </func>
+
     <func>
       <name name="system_flag" arity="2" clause_i="6"/>
-      <fsummary>Set system flag min_bin_vheap_size</fsummary>
+      <fsummary>Sets system flag <c>min_bin_vheap_size</c>.</fsummary>
       <desc>
-        <p>Sets the default minimum binary virtual heap size for processes. The
-          size is given in words. The new <c>min_bin_vhheap_size</c> only
-          effects processes spawned after the change of
+        <p>Sets the default minimum binary virtual heap size for
+          processes. The size is given in words.
+          The new <c>min_bin_vhheap_size</c> effects only
+          processes spawned after the change of
           <c>min_bin_vhheap_size</c> has been made.
-          The <c>min_bin_vheap_size</c> can be set for individual
-          processes by use of
+          <c>min_bin_vheap_size</c> can be set for individual
+          processes by using
           <seealso marker="#spawn_opt/4">spawn_opt/N</seealso> or
-          <seealso marker="#process_flag/2">process_flag/2</seealso>. </p>
+          <seealso marker="#process_flag/2">process_flag/2</seealso>.</p>
         <p>Returns the old value of the flag.</p>
       </desc>
     </func>
+
     <func>
       <name name="system_flag" arity="2" clause_i="7"/>
-      <fsummary>Set system flag multi_scheduling</fsummary>
+      <fsummary>Sets system flag <c>multi_scheduling</c>.</fsummary>
       <desc>
         <p><marker id="system_flag_multi_scheduling"></marker>
           If multi-scheduling is enabled, more than one scheduler
           thread is used by the emulator. Multi-scheduling can be
-          blocked. When multi-scheduling has been blocked, only
-          one scheduler thread will schedule Erlang processes.</p>
-        <p>If <c><anno>BlockState</anno> =:= block</c>, multi-scheduling will
-          be blocked. If <c><anno>BlockState</anno> =:= unblock</c> and no-one
-          else is blocking multi-scheduling and this process has
-          only blocked one time, multi-scheduling will be unblocked.
-          One process can block multi-scheduling multiple times.
-          If a process has blocked multiple times, it has to
+          blocked. When multi-scheduling is blocked, only
+          one scheduler thread schedules Erlang processes.</p>
+        <p>If <c><anno>BlockState</anno> =:= block</c>, multi-scheduling is
+          blocked. If <c><anno>BlockState</anno> =:= unblock</c> and no one
+          else blocks multi-scheduling, and this process has
+          blocked only once, multi-scheduling is unblocked.</p>
+        <p>One process can block multi-scheduling multiple times.
+          If a process has blocked multiple times, it must
           unblock exactly as many times as it has blocked before it
           has released its multi-scheduling block. If a process that
-          has blocked multi-scheduling exits, it will release its
+          has blocked multi-scheduling exits, it releases its
           blocking of multi-scheduling.</p>
         <p>The return values are <c>disabled</c>, <c>blocked</c>,
           or <c>enabled</c>. The returned value describes the
           state just after the call to
           <c>erlang:system_flag(multi_scheduling, <anno>BlockState</anno>)</c>
-          has been made. The return values are described in the
-          documentation of <seealso marker="#system_info_multi_scheduling">erlang:system_info(multi_scheduling)</seealso>.</p>
-        <p><em>NOTE</em>: Blocking of multi-scheduling should normally
-          not be needed. If you feel that you need to
-          block multi-scheduling, think through the
-          problem at least a couple of times again.
-          Blocking multi-scheduling should only be used
-          as a last resort since it will most likely be
-          a <em>very inefficient</em> way to solve the
-          problem.</p>
-        <p>See also <seealso marker="#system_info_multi_scheduling">erlang:system_info(multi_scheduling)</seealso>,
+          has been made. For information about the return values, see
+          <seealso marker="#system_info_multi_scheduling">erlang:system_info(multi_scheduling)</seealso>.</p>
+        <note><p>Blocking of multi-scheduling is normally not needed.
+          If you feel that you need to block multi-scheduling,
+          consider it a few more times again. Blocking multi-scheduling
+          is only to be used as a last resort, as it is most likely
+          a <em>very inefficient</em> way to solve the problem.</p>
+        </note>
+        <p>See also
+          <seealso marker="#system_info_multi_scheduling">erlang:system_info(multi_scheduling)</seealso>,
           <seealso marker="#system_info_multi_scheduling_blockers">erlang:system_info(multi_scheduling_blockers)</seealso>, and
           <seealso marker="#system_info_schedulers">erlang:system_info(schedulers)</seealso>.</p>
       </desc>
     </func>
+
     <func>
       <name name="system_flag" arity="2" clause_i="8"/>
+      <fsummary>Sets system flag <c>scheduler_bind_type</c>.</fsummary>
       <type name="scheduler_bind_type"/>
-      <fsummary>Set system flag scheduler_bind_type</fsummary>
       <desc>
         <warning>
           <p><marker id="system_flag_scheduler_bind_type"></marker>
-          This argument is <em>deprecated</em> and
-          scheduled for removal in erts-5.10/OTP-R16. Instead of using
-          this argument you are advised to use the <c>erl</c> command
-          line argument <seealso marker="erts:erl#+sbt">+sbt</seealso>.
-          When this argument has been removed a final scheduler bind type
-          to use will be determined at emulator boot time.</p>
+          This argument is <em>deprecated</em> and scheduled for
+          removal in <c>ERTS</c> 5.10/OTP R16. Instead of using this
+          argument, use command-line argument
+          <seealso marker="erts:erl#+sbt">+sbt</seealso> in <c>erl(1)</c>.
+          When this argument is removed, a final scheduler bind
+          type to use is determined at emulator boot time.</p>
         </warning>
         <p>Controls if and how schedulers are bound to logical
            processors.</p>
-        <p>When <c>erlang:system_flag(scheduler_bind_type, <anno>How</anno>)</c> is
-           called, an asynchronous signal is sent to all schedulers
-           online which causes them to try to bind or unbind as requested.
-           <em>NOTE:</em> If a scheduler fails to bind, this
-           will often be silently ignored. This since it isn't always
-           possible to verify valid logical processor identifiers. If
-           an error is reported, it will be reported to the
-           <c>error_logger</c>. If you want to verify that the
-           schedulers actually have bound as requested, call
-           <seealso marker="#system_info_scheduler_bindings">erlang:system_info(scheduler_bindings)</seealso>.
-        </p>
-        <p>Schedulers can currently only be bound on newer Linux,
+        <p>When <c>erlang:system_flag(scheduler_bind_type, <anno>How</anno>)</c>
+           is called, an asynchronous signal is sent to all schedulers
+           online, causing them to try to bind or unbind as requested.</p>
+        <note><p>If a scheduler fails to bind, this is often silently
+           ignored, as it is not always possible to verify valid
+           logical processor identifiers. If an error is reported,
+           it is reported to <c>error_logger</c>. To verify that the
+           schedulers have bound as requested, call
+           <seealso marker="#system_info_scheduler_bindings">erlang:system_info(scheduler_bindings)</seealso>.</p>
+        </note>
+        <p>Schedulers can be bound on newer Linux,
            Solaris, FreeBSD, and Windows systems, but more systems will be
-           supported in the future.
-        </p>
+           supported in future releases.</p>
         <p>In order for the runtime system to be able to bind schedulers,
-           the CPU topology needs to be known. If the runtime system fails
-           to automatically detect the CPU topology, it can be defined.
+           the CPU topology must be known. If the runtime system fails
+           to detect the CPU topology automatically, it can be defined.
            For more information on how to define the CPU topology, see
-           the <c>erl</c> <seealso marker="erts:erl#+sct">+sct</seealso> command
-           line flag.
-        </p>
-        <p>The runtime system will by default <em>not</em> bind schedulers
-           to logical processors.
-        </p>
-        <p><em>NOTE:</em> If the Erlang runtime system is the only
-           operating system process that binds threads to logical processors,
-           this improves the performance of the runtime system. However,
-           if other operating system processes (as for example another Erlang
-           runtime system) also bind threads to logical processors, there
-           might be a performance penalty instead. In some cases this
-           performance penalty might be severe. If this is the case, you
-           are advised to not bind the schedulers.</p>
-        <p>Schedulers can be bound in different ways. The <c><anno>How</anno></c>
-           argument determines how schedulers are bound. <c><anno>How</anno></c> can
-           currently be one of:</p>
+           command-line flag <seealso marker="erts:erl#+sct">+sct</seealso>
+           in <c>erl(1)</c>.</p>
+        <p>The runtime system does by default <em>not</em> bind schedulers
+           to logical processors.</p>
+        <note><p>If the Erlang runtime system is the only OS
+           process binding threads to logical processors, this
+           improves the performance of the runtime system. However,
+           if other OS processes (for example, another Erlang
+           runtime system) also bind threads to logical processors,
+           there can be a performance penalty instead. Sometimes this
+           performance penalty can be severe. If so, it is recommended
+           to not bind the schedulers.</p>
+        </note>
+        <p>Schedulers can be bound in different ways. Argument
+           <c><anno>How</anno></c> determines how schedulers are
+           bound and can be any of the following:</p>
         <taglist>
           <tag><c>unbound</c></tag>
-          <item><p>Same as the <c>erl</c> command line argument
-          <seealso marker="erts:erl#+sbt">+sbt u</seealso>.
+          <item><p>Same as command-line argument
+          <seealso marker="erts:erl#+sbt">+sbt u</seealso> in <c>erl(1)</c>.
           </p></item>
           <tag><c>no_spread</c></tag>
-          <item><p>Same as the <c>erl</c> command line argument
-          <seealso marker="erts:erl#+sbt">+sbt ns</seealso>.
+          <item><p>Same as command-line argument
+          <seealso marker="erts:erl#+sbt">+sbt ns</seealso> in <c>erl(1)</c>.
           </p></item>
           <tag><c>thread_spread</c></tag>
-          <item><p>Same as the <c>erl</c> command line argument
-          <seealso marker="erts:erl#+sbt">+sbt ts</seealso>.
+          <item><p>Same as command-line argument
+          <seealso marker="erts:erl#+sbt">+sbt ts</seealso> in <c>erl(1)</c>.
           </p></item>
           <tag><c>processor_spread</c></tag>
-          <item><p>Same as the <c>erl</c> command line argument
-          <seealso marker="erts:erl#+sbt">+sbt ps</seealso>.
+          <item><p>Same as command-line argument
+          <seealso marker="erts:erl#+sbt">+sbt ps</seealso> in <c>erl(1)</c>.
           </p></item>
           <tag><c>spread</c></tag>
-          <item><p>Same as the <c>erl</c> command line argument
-          <seealso marker="erts:erl#+sbt">+sbt s</seealso>.
+          <item><p>Same as command-line argument
+          <seealso marker="erts:erl#+sbt">+sbt s</seealso> in <c>erl(1)</c>.
           </p></item>
           <tag><c>no_node_thread_spread</c></tag>
-          <item><p>Same as the <c>erl</c> command line argument
-          <seealso marker="erts:erl#+sbt">+sbt nnts</seealso>.
+          <item><p>Same as command-line argument
+          <seealso marker="erts:erl#+sbt">+sbt nnts</seealso> in <c>erl(1)</c>.
           </p></item>
           <tag><c>no_node_processor_spread</c></tag>
-          <item><p>Same as the <c>erl</c> command line argument
-          <seealso marker="erts:erl#+sbt">+sbt nnps</seealso>.
+          <item><p>Same as command-line argument
+          <seealso marker="erts:erl#+sbt">+sbt nnps</seealso> in <c>erl(1)</c>.
           </p></item>
           <tag><c>thread_no_node_processor_spread</c></tag>
-          <item><p>Same as the <c>erl</c> command line argument
-          <seealso marker="erts:erl#+sbt">+sbt tnnps</seealso>.
+          <item><p>Same as command-line argument
+          <seealso marker="erts:erl#+sbt">+sbt tnnps</seealso> in <c>erl(1)</c>.
           </p></item>
           <tag><c>default_bind</c></tag>
-          <item><p>Same as the <c>erl</c> command line argument
-          <seealso marker="erts:erl#+sbt">+sbt db</seealso>.
+          <item><p>Same as command-line argument
+          <seealso marker="erts:erl#+sbt">+sbt db</seealso> in <c>erl(1)</c>.
           </p></item>
         </taglist>
-        <p>The value returned equals <c><anno>How</anno></c> before the
-           <c>scheduler_bind_type</c> flag was changed.</p>
-        <p>Failure:</p>
+        <p>The returned value equals <c><anno>How</anno></c> before flag
+           <c>scheduler_bind_type</c> was changed.</p>
+        <p>Failures:</p>
         <taglist>
           <tag><c>notsup</c></tag>
           <item>
@@ -5835,168 +6423,171 @@ ok
           </item>
           <tag><c>badarg</c></tag>
           <item>
-            <p>If <c>How</c> isn't one of the documented alternatives.</p>
+            <p>If <c><anno>How</anno></c> is not one of the documented
+            alternatives.</p>
           </item>
           <tag><c>badarg</c></tag>
           <item>
-            <p>If no CPU topology information is available.</p>
+            <p>If CPU topology information is unavailable.</p>
           </item>
         </taglist>
         <p>The scheduler bind type can also be set by passing
-           the <seealso marker="erts:erl#+sbt">+sbt</seealso> command
-           line argument to <c>erl</c>.
-        </p>
+           command-line argument
+           <seealso marker="erts:erl#+sbt">+sbt</seealso> to <c>erl(1)</c>.</p>
         <p>For more information, see
            <seealso marker="#system_info_scheduler_bind_type">erlang:system_info(scheduler_bind_type)</seealso>,
            <seealso marker="#system_info_scheduler_bindings">erlang:system_info(scheduler_bindings)</seealso>,
-           the <c>erl</c> <seealso marker="erts:erl#+sbt">+sbt</seealso>
-           and <seealso marker="erts:erl#+sct">+sct</seealso> command line
-           flags.
-        </p>
+           as well as command-line flags
+           <seealso marker="erts:erl#+sbt">+sbt</seealso>
+           and <seealso marker="erts:erl#+sct">+sct</seealso>
+           in <c>erl(1)</c>.</p>
       </desc>
     </func>
+
     <func>
       <name name="system_flag" arity="2" clause_i="9"/>
-      <fsummary>Set system flag scheduler_wall_time</fsummary>
+      <fsummary>Sets system flag <c>scheduler_wall_time</c>.</fsummary>
       <desc><p><marker id="system_flag_scheduler_wall_time"></marker>
-      Turns on/off scheduler wall time measurements. </p>
-      <p>For more information see,
-      <seealso marker="#statistics_scheduler_wall_time">erlang:statistics(scheduler_wall_time)</seealso>.
-      </p>
+      Turns on or off scheduler wall time measurements.</p>
+      <p>For more information, see
+      <seealso marker="#statistics_scheduler_wall_time">erlang:statistics(scheduler_wall_time)</seealso>.</p>
       </desc>
     </func>
+
     <func>
       <name name="system_flag" arity="2" clause_i="10"/>
-      <fsummary>Set system flag schedulers_online</fsummary>
+      <fsummary>Sets system flag <c>schedulers_online</c>.</fsummary>
       <desc>
         <p><marker id="system_flag_schedulers_online"></marker>
-        Sets the amount of schedulers online. Valid range is
-        <![CDATA[1 <= SchedulersOnline <= erlang:system_info(schedulers)]]>.
-        </p>
+        Sets the number of schedulers online. Range is
+        <![CDATA[1 <= SchedulersOnline <= erlang:system_info(schedulers)]]>.</p>
         <p>Returns the old value of the flag.</p>
-        <p>Note that if the emulator was built with support for <seealso
-        marker="#system_flag_dirty_cpu_schedulers_online">dirty schedulers</seealso>,
-        changing the number of schedulers online can also change the number of dirty
-        CPU schedulers online. For example, if there are 12 schedulers and all are
-        online, and 6 dirty CPU schedulers, all online as well, and <c>system_flag/2</c>
-        is used to set the number of schedulers online to 6, then the number of dirty
-        CPU schedulers online is automatically decreased by half as well, down to 3.
-        Similarly, the number of dirty CPU schedulers online increases proportionally
-        to increases in the number of schedulers online.</p>
-        <p>For more information see,
-        <seealso marker="#system_info_schedulers">erlang:system_info(schedulers)</seealso>,
+        <p>If the emulator was built with support for
+        <seealso marker="#system_flag_dirty_cpu_schedulers_online">dirty schedulers</seealso>,
+        changing the number of schedulers online can also change the
+        number of dirty CPU schedulers online. For example, if 12
+        schedulers and 6 dirty CPU schedulers are online, and
+        <c>system_flag/2</c> is used to set the number of schedulers
+        online to 6, then the number of dirty CPU schedulers online
+        is automatically decreased by half as well, down to 3.
+        Similarly, the number of dirty CPU schedulers online increases
+        proportionally to increases in the number of schedulers online.</p>
+        <p>For more information, see
+        <seealso marker="#system_info_schedulers">erlang:system_info(schedulers)</seealso>
         and
-        <seealso marker="#system_info_schedulers_online">erlang:system_info(schedulers_online)</seealso>.
-        </p>
+        <seealso marker="#system_info_schedulers_online">erlang:system_info(schedulers_online)</seealso>.</p>
       </desc>
     </func>
+
     <func>
       <name name="system_flag" arity="2" clause_i="11"/>
-      <fsummary>Set system flag trace_control_word</fsummary>
+      <fsummary>Sets system flag <c>trace_control_word</c>.</fsummary>
       <desc>
-        <p>Sets the value of the node's trace control word to
-          <c><anno>TCW</anno></c>. <c><anno>TCW</anno></c> should be an unsigned integer. For
-          more information see documentation of the
+        <p>Sets the value of the node trace control word to
+          <c><anno>TCW</anno></c>, which is to be an unsigned integer.
+          For more information, see the function
           <seealso marker="erts:match_spec#set_tcw">set_tcw</seealso>
-          function in the match specification documentation in the
-          ERTS User's Guide.</p>
+          in Section "Match Specifications in Erlang" in the
+          User's Guide.</p>
         <p>Returns the old value of the flag.</p>
       </desc>
     </func>
-    <marker id="system_flag_time_offset"/>
+
     <func>
       <name name="system_flag" arity="2" clause_i="12"/>
       <fsummary>Finalize the Time Offset</fsummary>
       <desc>
-        <p>Finalizes the <seealso marker="#time_offset/0">time offset</seealso>
-	when the <seealso marker="time_correction#Single_Time_Warp_Mode">single
-	time warp mode</seealso> is being used. If another time warp mode than
-	the "single time warp mode" is used, the time offset state will be left
-	unchanged.</p>
-        <p>Returns the old state identifier. That is, if:</p>
+        <p><marker id="system_flag_time_offset"></marker>
+	Finalizes the <seealso marker="#time_offset/0">time offset</seealso>
+	when <seealso marker="time_correction#Single_Time_Warp_Mode">single
+	time warp mode</seealso> is used. If another time warp mode
+	is used, the time offset state is left unchanged.</p>
+        <p>Returns the old state identifier. That is:</p>
 	<list>
-	  <item><p><c>preliminary</c> is returned, finalization was
+	  <item><p>If <c>preliminary</c> is returned, finalization was
 	  performed and the time offset is now final.</p></item>
 
-	  <item><p><c>final</c> is returned, the time offset was
-	  already in the final state. This either due to another
+	  <item><p>If <c>final</c> is returned, the time offset was
+	  already in the final state. This either because another
 	  <c>erlang:system_flag(time_offset, finalize)</c> call, or
-	  due to the
-	  <seealso marker="time_correction#No_Time_Warp_Mode">no
-	  time warp mode</seealso> being used.</p></item>
+	  because <seealso marker="time_correction#No_Time_Warp_Mode">no
+	  time warp mode</seealso> is used.</p></item>
 
-	  <item><p><c>volatile</c> is returned, the time offset
-	  cannot be finalized due to the 
+	  <item><p>If <c>volatile</c> is returned, the time offset
+	  cannot be finalized because
 	  <seealso marker="time_correction#Multi_Time_Warp_Mode">multi
-	  time warp mode</seealso> being used.</p></item>
+	  time warp mode</seealso> is used.</p></item>
 	</list>
       </desc>
     </func>
+
     <func>
       <name name="system_info" arity="1" clause_i="1"/>
       <name name="system_info" arity="1" clause_i="2"/>
       <name name="system_info" arity="1" clause_i="3"/>
       <name name="system_info" arity="1" clause_i="4"/>
       <name name="system_info" arity="1" clause_i="5"/>
+      <fsummary>Information about the system allocators.</fsummary>
       <type variable="Allocator" name_i="2"/>
       <type variable="Version" name_i="2"/>
       <type variable="Features" name_i="2"/>
       <type variable="Settings" name_i="2"/>
       <type variable="Alloc" name_i="3"/>
-      <fsummary>Information about the allocators of the system</fsummary>
       <desc>
-        <p>
-          Returns various information about the
-          <marker id="system_info_allocator_tags">allocators</marker> of the
+      <marker id="system_info_allocator_tags"></marker>
+        <p>Returns various information about the allocators of the
           current system (emulator) as specified by
           <c><anno>Item</anno></c>:</p>
+        <marker id="system_info_allocated_areas"></marker>
         <taglist>
-          <tag><marker id="system_info_allocated_areas"><c>allocated_areas</c></marker></tag>
+          <tag><c>allocated_areas</c></tag>
           <item>
             <p>Returns a list of tuples with information about
               miscellaneous allocated memory areas.</p>
-            <p>Each tuple contains an atom describing type of memory as
-              first element and amount of allocated memory in bytes as
-              second element. In those cases when there is information
-              present about allocated and used memory, a third element
-              is present. This third element contains the amount of
+            <p>Each tuple contains an atom describing the type of
+              memory as first element and the amount of allocated
+              memory in bytes as second element. When information
+              about allocated and used memory is present, also a
+              third element is present, containing the amount of
               used memory in bytes.</p>
             <p><c>erlang:system_info(allocated_areas)</c> is intended
-              for debugging, and the content is highly implementation
-              dependent. The content of the results will therefore
-              change when needed without prior notice.</p>
-            <p><em>Note:</em> The sum of these values is <em>not</em>
+              for debugging, and the content is highly
+              implementation-dependent. The content of the results
+             therefore changes when needed without prior notice.</p>
+            <p>Notice that the sum of these values is <em>not</em>
               the total amount of memory allocated by the emulator.
               Some values are part of other values, and some memory
-              areas are not part of the result. If you are interested
-              in the total amount of memory allocated by the emulator
-              see <seealso marker="#memory/0">erlang:memory/0,1</seealso>.</p>
+              areas are not part of the result. For information about
+              the total amount of memory allocated by the emulator, see
+              <seealso marker="#memory/0">erlang:memory/0,1</seealso>.</p>
           </item>
-          <tag><marker id="system_info_allocator"><c>allocator</c></marker></tag>
+          <tag><c>allocator</c></tag>
           <item>
-            <p>Returns <c>{<anno>Allocator</anno>, <anno>Version</anno>, <anno>Features</anno>, <anno>Settings</anno>}.</c></p>
-            <p>Explanation:</p>
+          <marker id="system_info_allocator"></marker>
+            <p>Returns <c>{<anno>Allocator</anno>, <anno>Version</anno>,
+              <anno>Features</anno>, <anno>Settings</anno></c>, where:</p>
             <list type="bulleted">
               <item>
-                <p><c><anno>Allocator</anno></c> corresponds to the <c>malloc()</c>
-                  implementation used. If <c><anno>Allocator</anno></c> equals
+                <p><c><anno>Allocator</anno></c> corresponds to the
+                  <c>malloc()</c> implementation used. If
+                  <c><anno>Allocator</anno></c> equals
                   <c>undefined</c>, the <c>malloc()</c> implementation
-                  used could not be identified. Currently
-                  <c>glibc</c> can be identified.</p>
+                  used cannot be identified. <c>glibc</c> can be
+                  identified.</p>
               </item>
               <item>
-                <p><c><anno>Version</anno></c> is a list of integers (but not a
-                  string) representing the version of
+                <p><c><anno>Version</anno></c> is a list of integers
+                  (but not a string) representing the version of
                   the <c>malloc()</c> implementation used.</p>
               </item>
               <item>
-                <p><c><anno>Features</anno></c> is a list of atoms representing
-                  allocation features used.</p>
+                <p><c><anno>Features</anno></c> is a list of atoms
+                  representing the allocation features used.</p>
               </item>
               <item>
-                <p><c><anno>Settings</anno></c> is a list of subsystems, their
-                  configurable parameters, and used values. Settings
-                  may differ between different combinations of
+                <p><c><anno>Settings</anno></c> is a list of subsystems,
+                  their configurable parameters, and used values. Settings
+                  can differ between different combinations of
                   platforms, allocators, and allocation features.
                   Memory sizes are given in bytes.</p>
               </item>
@@ -6004,165 +6595,169 @@ ok
             <p>See also "System Flags Effecting erts_alloc" in
               <seealso marker="erts:erts_alloc#flags">erts_alloc(3)</seealso>.</p>
           </item>
-          <tag><marker id="system_info_alloc_util_allocators"><c>alloc_util_allocators</c></marker></tag>
+          <tag><c>alloc_util_allocators</c></tag>
           <item>
-	     <p>Returns a list of the names of all allocators
-	        using the ERTS internal <c>alloc_util</c> framework
-		as atoms. For more information see the
-	        <seealso marker="erts:erts_alloc#alloc_util">"the
-		alloc_util framework" section in the
-		erts_alloc(3)</seealso> documentation.
-	     </p>
+          <marker id="system_info_alloc_util_allocators"></marker>
+	     <p>Returns a list of the names of all allocators using
+	        the <c>ERTS</c> internal <c>alloc_util</c> framework
+		as atoms. For more information, see Section
+	        <seealso marker="erts:erts_alloc#alloc_util">"The
+		alloc_util framework" in erts_alloc(3)</seealso>.</p>
           </item>
-          <tag><marker id="system_info_allocator_tuple"><c>{allocator, <anno>Alloc</anno>}</c></marker></tag>
+          <tag><c>{allocator, <anno>Alloc</anno>}</c></tag>
           <item>
+          <marker id="system_info_allocator_tuple"></marker>
             <p>Returns information about the specified allocator.
-	       As of erts version 5.6.1 the return value is a list
-	       of <c>{instance, InstanceNo, InstanceInfo}</c> tuples
+	       As from <c>ERTS</c> 5.6.1, the return value is a list
+	       of <c>{instance, InstanceNo, InstanceInfo}</c> tuples,
 	       where <c>InstanceInfo</c> contains information about
-	       a specific instance of the allocator. As of erts version
-	       5.10.4 the returned list when calling
+	       a specific instance of the allocator. As from
+               <c>ERTS</c> 5.10.4, the returned list when calling
 	       <c>erlang:system_info({allocator, mseg_alloc})</c> also
-	       include an <c>{erts_mmap, _}</c> tuple as one element
-	       in the list.
-	       If <c><anno>Alloc</anno></c> is not a recognized allocator,
-              <c>undefined</c> is returned. If <c><anno>Alloc</anno></c> is disabled,
+	       includes an <c>{erts_mmap, _}</c> tuple as one element
+	       in the list. If <c><anno>Alloc</anno></c> is not a
+               recognized allocator, <c>undefined</c> is returned.
+               If <c><anno>Alloc</anno></c> is disabled,
               <c>false</c> is returned.</p>
-            <p><em>Note:</em> The information returned is highly
-              implementation dependent and may be changed, or removed
+            <p>Notice that the information returned is highly
+              implementation-dependent and can be changed or removed
               at any time without prior notice. It was initially
               intended as a tool when developing new allocators, but
-              since it might be of interest for others it has been
+              as it can be of interest for others it has been
               briefly documented.</p>
             <p>The recognized allocators are listed in
               <seealso marker="erts:erts_alloc">erts_alloc(3)</seealso>.
               After reading the <c>erts_alloc(3)</c> documentation,
 	      the returned information
-              should more or less speak for itself. But it can be worth
+              more or less speaks for itself, but it can be worth
               explaining some things. Call counts are presented by two
-              values. The first value is giga calls, and the second
-              value is calls. <c>mbcs</c>, and <c>sbcs</c> are
-              abbreviations for, respectively, multi-block carriers, and
-              single-block carriers. Sizes are presented in bytes. When
-              it is not a size that is presented, it is the amount of
-              something. Sizes and amounts are often presented by three
-              values, the first is current value, the second is maximum
-              value since the last call to
-              <c>erlang:system_info({allocator, Alloc})</c>, and
-              the third is maximum value since the emulator was started.
-              If only one value is present, it is the current value.
+              values, the first value is giga calls, and the second
+              value is calls. <c>mbcs</c> and <c>sbcs</c> denote
+              multi-block carriers, and single-block carriers,
+              respectively. Sizes are presented in bytes. When a
+              size is not presented, it is the amount of something.
+              Sizes and amounts are often presented by three values:</p>
+            <list type="bulleted">
+              <item>The first is the current value.</item>
+              <item>The second is the maximum value since the last call
+                to <c>erlang:system_info({allocator, Alloc})</c>.</item>
+              <item>The third is the maximum value since the emulator
+                was started.</item>
+            </list>
+              <p>If only one value is present, it is the current value.
               <c>fix_alloc</c> memory block types are presented by two
-              values. The first value is memory pool size and
-              the second value used memory size.</p>
+              values. The first value is the memory pool size and
+              the second value is the used memory size.</p>
           </item>
-          <tag><marker id="system_info_allocator_sizes"><c>{allocator_sizes, <anno>Alloc</anno>}</c></marker></tag>
+          <tag><c>{allocator_sizes, <anno>Alloc</anno>}</c></tag>
           <item>
+          <marker id="system_info_allocator_sizes"></marker>
 	    <p>Returns various size information for the specified
 	    allocator. The information returned is a subset of the
 	    information returned by
-	    <seealso marker="#system_info_allocator_tuple">erlang:system_info({allocator, <anno>Alloc</anno>})</seealso>.
+	    <seealso marker="#system_info_allocator_tuple"><c>erlang:system_info({allocator, <anno>Alloc</anno>})</c></seealso>.
 	    </p>
           </item>
         </taglist>
       </desc>
     </func>
+
     <func>
       <name name="system_info" arity="1" clause_i="10"/>
       <name name="system_info" arity="1" clause_i="11"/>
+      <fsummary>Information about the CPU topology of the system.</fsummary>
       <type name="cpu_topology"/>
       <type name="level_entry"/>
       <type_desc name="cpu_topology">
-        <marker id="system_info_cpu_topology"></marker>
         All <c><anno>LevelEntry</anno></c>s of a list
         must contain the same <c><anno>LevelTag</anno></c>, except
         on the top level where both <c>node</c> and
-        <c>processor</c> <c><anno>LevelTag</anno></c>s may co-exist.
+        <c>processor</c> <c><anno>LevelTag</anno></c>s can coexist.
       </type_desc>
       <type_desc name="level_entry">
-       <c>{<anno>LevelTag</anno>, <anno>SubLevel</anno>} == {<anno>LevelTag</anno>, [], <anno>SubLevel</anno>}</c>
+        <c>{<anno>LevelTag</anno>,
+        <anno>SubLevel</anno>} == {<anno>LevelTag</anno>, [],
+        <anno>SubLevel</anno>}</c>
       </type_desc>
       <type name="level_tag"/>
       <type_desc name="level_tag">
-        More <c><anno>LevelTag</anno></c>s may be introduced in the future.
+        More <c><anno>LevelTag</anno></c>s can be introduced in a
+        future release.
       </type_desc>
       <type name="sub_level"/>
       <type name="info_list"/>
       <type_desc name="info_list">
-        The <c>info_list()</c> may be extended in the future.
+        The <c>info_list()</c> can be extended in a future release.
       </type_desc>
-      <fsummary>Information about the CPU topology of the system</fsummary>
       <desc>
-        <p>Returns various information about the
-          <marker id="system_info_cpu_topology_tags">CPU topology</marker>
-          of the current system
-          (emulator) as specified by <c><anno>Item</anno></c>:</p>
+      <marker id="system_info_cpu_topology_tags"></marker>
+      <marker id="system_info_cpu_topology"></marker>
+        <p>Returns various information about the CPU topology of
+          the current system (emulator) as specified by
+          <c><anno>Item</anno></c>:</p>
         <taglist>
           <tag><c>cpu_topology</c></tag>
           <item>
-	    <p>Returns the <c><anno>CpuTopology</anno></c> which currently is used by the
-	       emulator. The CPU topology is used when binding schedulers
+	    <p>Returns the <c><anno>CpuTopology</anno></c> currently used by
+	       the emulator. The CPU topology is used when binding schedulers
 	       to logical processors. The CPU topology used is the
-	       <seealso marker="erlang#system_info_cpu_topology_defined">user
-	       defined CPU topology</seealso> if such exists; otherwise, the
-	       <seealso marker="erlang#system_info_cpu_topology_detected">automatically
-	       detected CPU topology</seealso> if such exists. If no CPU topology
+	       <seealso marker="erlang#system_info_cpu_topology_defined">user-defined CPU topology</seealso>,
+               if such exists, otherwise the
+	       <seealso marker="erlang#system_info_cpu_topology_detected">automatically detected CPU topology</seealso>,
+               if such exists. If no CPU topology
 	       exists, <c>undefined</c> is returned.</p>
-	    <p><c>node</c> refers to NUMA (non-uniform memory access)
-	       nodes, and <c>thread</c> refers to hardware threads
-	       (e.g. Intels hyper-threads).</p>
-            <p>A level in the <c><anno>CpuTopology</anno></c> term can be omitted if
-	       only one entry exists and the <c><anno>InfoList</anno></c> is empty.
-	    </p>
+	    <p><c>node</c> refers to Non-Uniform Memory Access (NUMA)
+	       nodes. <c>thread</c> refers to hardware threads
+	       (for example, Intel hyper-threads).</p>
+            <p>A level in term <c><anno>CpuTopology</anno></c> can be
+               omitted if only one entry exists and
+               <c><anno>InfoList</anno></c> is empty.</p>
 	    <p><c>thread</c> can only be a sub level to <c>core</c>.
-	       <c>core</c> can be a sub level to either <c>processor</c>
-	       or <c>node</c>.  <c>processor</c> can either be on the
+	       <c>core</c> can be a sub level to <c>processor</c>
+	       or <c>node</c>. <c>processor</c> can be on the
 	       top level or a sub level to <c>node</c>. <c>node</c>
-	       can either be on the top level or a sub level to
+	       can be on the top level or a sub level to
 	       <c>processor</c>. That is, NUMA nodes can be processor
 	       internal or processor external. A CPU topology can
 	       consist of a mix of processor internal and external
-	       NUMA nodes, as long as each logical CPU belongs to one
-	       and only one NUMA node. Cache hierarchy is not part of
-	       the <c><anno>CpuTopology</anno></c> type yet, but will be in the
-	       future. Other things may also make it into the CPU
-	       topology in the future. In other words, expect the
-	       <c><anno>CpuTopology</anno></c> type to change.
-	    </p>
-          </item>
-          <tag><marker id="system_info_cpu_topology_defined"><c>{cpu_topology, defined}</c></marker></tag>
-          <item>
-            <p>Returns the user defined <c><anno>CpuTopology</anno></c>. For more
-	       information see the documentation of
-	       the <c>erl</c> <seealso marker="erts:erl#+sct">+sct</seealso> command
-	       line flag, and the documentation of the
-	       <seealso marker="#system_info_cpu_topology">cpu_topology</seealso>
-	       argument.
-	    </p>
-          </item>
-          <tag><marker id="system_info_cpu_topology_detected"><c>{cpu_topology, detected}</c></marker></tag>
-          <item>
-            <p>Returns the automatically detected <c><anno>CpuTopology</anno></c>. The
-	       emulator currently only detects the CPU topology on some newer
-	       Linux, Solaris, FreeBSD, and Windows systems. On Windows system with
-	       more than 32 logical processors the CPU topology is not detected.
-	    </p>
-            <p>For more information see the documentation of the
-	       <seealso marker="#system_info_cpu_topology">cpu_topology</seealso>
-	       argument.
-	    </p>
+	       NUMA nodes, as long as each logical CPU belongs to
+	       <em>one</em> NUMA node. Cache hierarchy is not part of
+	       the <c><anno>CpuTopology</anno></c> type, but will be in a
+               future release. Other things can also make it into the CPU
+	       topology in a future release. In other words, expect the
+	       <c><anno>CpuTopology</anno></c> type to change.</p>
+          </item>
+          <tag><c>{cpu_topology, defined}</c></tag>
+          <item>
+          <marker id="system_info_cpu_topology_defined"></marker>
+            <p>Returns the user-defined <c><anno>CpuTopology</anno></c>.
+               For more information, see command-line flag
+	       <seealso marker="erts:erl#+sct">+sct</seealso> in
+	       <c>erl(1)</c> and argument
+	       <seealso marker="#system_info_cpu_topology">cpu_topology</seealso>.</p>
+          </item>
+          <tag><c>{cpu_topology, detected}</c></tag>
+          <item>
+          <marker id="system_info_cpu_topology_detected"></marker>
+            <p>Returns the automatically detected
+               <c><anno>CpuTopology</anno>y</c>. The
+	       emulator detects the CPU topology on some newer
+	       Linux, Solaris, FreeBSD, and Windows systems.
+               On Windows system with more than 32 logical processors,
+               the CPU topology is not detected.</p>
+            <p>For more information, see argument
+	       <seealso marker="#system_info_cpu_topology">cpu_topology</seealso>.</p>
           </item>
           <tag><c>{cpu_topology, used}</c></tag>
           <item>
-            <p>Returns the <c><anno>CpuTopology</anno></c> which is used by the
-	       emulator. For more information see the 
-               documentation of the
-	       <seealso marker="#system_info_cpu_topology">cpu_topology</seealso>
-	       argument.
-	    </p>
+            <p>Returns <c><anno>CpuTopology</anno></c> used by the emulator.
+              For more information, see argument
+	       <seealso marker="#system_info_cpu_topology">cpu_topology</seealso>.</p>
           </item>
         </taglist>
       </desc>
     </func>
+
     <func>
       <name name="system_info" arity="1" clause_i="6"/>
       <name name="system_info" arity="1" clause_i="7"/>
@@ -6224,7 +6819,8 @@ ok
       <name name="system_info" arity="1" clause_i="65"/>
       <name name="system_info" arity="1" clause_i="66"/>
       <name name="system_info" arity="1" clause_i="67"/>
-      <fsummary>Information about the system</fsummary>
+      <name name="system_info" arity="1" clause_i="68"/>
+      <fsummary>Information about the system.</fsummary>
       <desc>
         <p>Returns various information about the current system
           (emulator) as specified by <c><anno>Item</anno></c>:</p>
@@ -6241,8 +6837,7 @@ ok
                Other possible return values are <c>debug</c>, <c>purify</c>,
 	       <c>quantify</c>, <c>purecov</c>, <c>gcov</c>, <c>valgrind</c>,
 	       <c>gprof</c>, and <c>lcnt</c>. Possible return values
-	       may be added and/or removed at any time without prior notice.
-	     </p>
+	       can be added or removed at any time without prior notice.</p>
           </item>
           <tag><c>c_compiler_used</c></tag>
           <item>
@@ -6250,26 +6845,25 @@ ok
 	       compiling the runtime system. The first element is an
 	       atom describing the name of the compiler, or <c>undefined</c>
 	       if unknown. The second element is a term describing the
-	       version of the compiler, or <c>undefined</c> if unknown.
-	     </p>
+	       version of the compiler, or <c>undefined</c> if unknown.</p>
           </item>
           <tag><c>check_io</c></tag>
           <item>
             <p>Returns a list containing miscellaneous information
-              regarding the emulators internal I/O checking. Note,
-              the content of the returned list may vary between
-              platforms and over time. The only thing guaranteed is
+              about the emulators internal I/O checking. Notice that
+              the content of the returned list can vary between
+              platforms and over time. It is only guaranteed
               that a list is returned.</p>
           </item>
           <tag><c>compat_rel</c></tag>
           <item>
             <p>Returns the compatibility mode of the local node as
               an integer. The integer returned represents the
-              Erlang/OTP release which the current emulator has been
+              Erlang/OTP release that the current emulator has been
               set to be backward compatible with. The compatibility
-              mode can be configured at startup by using the command
-              line flag <c>+R</c>, see
-              <seealso marker="erts:erl#compat_rel">erl(1)</seealso>.</p>
+              mode can be configured at startup by using command-line flag
+              <seealso marker="erts:erl#compat_rel">+R</seealso> in
+              <c>erl(1)</c>.</p>
           </item>
           <tag><c>cpu_topology</c></tag>
           <item>
@@ -6282,144 +6876,150 @@ ok
               creation of a node is stored in process identifiers, port
               identifiers, and references. This makes it (to some
               extent) possible to distinguish between identifiers from
-              different incarnations of a node. Currently valid
-              creations are integers in the range 1..3, but this may
-              (probably will) change in the future. If the node is not
-              alive, 0 is returned.</p>
+              different incarnations of a node. The valid
+              creations are integers in the range 1..3, but this will
+              probably change in a future release. If the node is not
+              alive, <c>0</c> is returned.</p>
           </item>
           <tag><c>debug_compiled</c></tag>
           <item>
             <p>Returns <c>true</c> if the emulator has been debug
-	       compiled; otherwise, <c>false</c>.
-	    </p>
+	       compiled, otherwise <c>false</c>.</p>
           </item>
-          <tag><marker id="system_info_delayed_node_table_gc"><c>delayed_node_table_gc</c></marker></tag>
+          <tag><c>delayed_node_table_gc</c></tag>
           <item>
-            <p>Returns the amount of time in seconds that garbage collection
-	    of an entry in a node table will be delayed. This limit can be set
-	    on startup by passing the
-	    <seealso marker="erts:erl#+zdntgc">+zdntgc</seealso> command line
-	    flag to <c>erl</c>. For more information see the documentation of the
+            <marker id="system_info_delayed_node_table_gc"></marker>
+            <p>Returns the amount of time in seconds garbage collection
+	    of an entry in a node table is delayed. This limit can be set
+	    on startup by passing the command line flag
+	    <seealso marker="erts:erl#+zdntgc">+zdntgc</seealso>
+	    to <c>erl</c>. For more information see the documentation of the
 	    command line flag.</p>
           </item>
-          <tag><marker id="system_info_dirty_cpu_schedulers"><c>dirty_cpu_schedulers</c></marker></tag>
+          <tag><c>dirty_cpu_schedulers</c></tag>
           <item>
+          <marker id="system_info_dirty_cpu_schedulers"></marker>
             <p>Returns the number of dirty CPU scheduler threads used by
               the emulator. Dirty CPU schedulers execute CPU-bound
-              native functions such as NIFs, linked-in driver code, and BIFs
-              that cannot be managed cleanly by the emulator's normal schedulers.
-            </p>
-            <p>The number of dirty CPU scheduler threads is determined at emulator
-              boot time and cannot be changed after that. The number of dirty CPU
-              scheduler threads online can however be changed at any time. The number of
-              dirty CPU schedulers can be set on startup by passing
-              the <seealso marker="erts:erl#+SDcpu">+SDcpu</seealso> or
-              <seealso marker="erts:erl#+SDPcpu">+SDPcpu</seealso> command line flags,
-              see <seealso marker="erts:erl#+SDcpu">erl(1)</seealso>.
-            </p>
-	    <p><em>Note that the dirty schedulers functionality is experimental</em>, and
-	    that you have to enable support for dirty schedulers when building OTP in
-	    order to try out the functionality.</p>
-            <p>See also <seealso marker="#system_flag_dirty_cpu_schedulers_online">erlang:system_flag(dirty_cpu_schedulers_online, DirtyCPUSchedulersOnline)</seealso>,
+              native functions, such as NIFs, linked-in driver code,
+              and BIFs that cannot be managed cleanly by the normal
+              emulator schedulers.</p>
+            <p>The number of dirty CPU scheduler threads is determined
+              at emulator boot time and cannot be changed after that.
+              However, the number of dirty CPU scheduler threads online
+              can be changed at any time. The number of dirty CPU
+              schedulers can be set at startup by passing
+              command-line flag
+              <seealso marker="erts:erl#+SDcpu">+SDcpu</seealso> or
+              <seealso marker="erts:erl#+SDPcpu">+SDPcpu</seealso> in
+              <c>erl(1)</c>.</p>
+	    <p>Notice that the dirty schedulers functionality is
+              experimental. Enable support for dirty schedulers when
+              building OTP to try out the functionality.</p>
+            <p>See also
+              <seealso marker="#system_flag_dirty_cpu_schedulers_online">erlang:system_flag(dirty_cpu_schedulers_online, DirtyCPUSchedulersOnline)</seealso>,
               <seealso marker="#system_info_dirty_cpu_schedulers_online">erlang:system_info(dirty_cpu_schedulers_online)</seealso>,
               <seealso marker="#system_info_dirty_io_schedulers">erlang:system_info(dirty_io_schedulers)</seealso>,
               <seealso marker="#system_info_schedulers">erlang:system_info(schedulers)</seealso>,
 	      <seealso marker="#system_info_schedulers_online">erlang:system_info(schedulers_online)</seealso>, and
               <seealso marker="#system_flag_schedulers_online">erlang:system_flag(schedulers_online, SchedulersOnline)</seealso>.</p>
           </item>
-          <tag><marker id="system_info_dirty_cpu_schedulers_online"><c>dirty_cpu_schedulers_online</c></marker></tag>
-          <item>
-            <p>Returns the number of dirty CPU schedulers online. The return value
-              satisfies the following relationship:
-              <c><![CDATA[1 <= DirtyCPUSchedulersOnline <= N]]></c>, where <c>N</c> is
-              the lesser of the return values of <c>erlang:system_info(dirty_cpu_schedulers)</c> and
-              <c>erlang:system_info(schedulers_online)</c>.
-            </p>
-            <p>The number of dirty CPU schedulers online can be set on startup by passing
-              the <seealso marker="erts:erl#+SDcpu">+SDcpu</seealso> command line flag, see
-              <seealso marker="erts:erl#+SDcpu">erl(1)</seealso>.
-            </p>
-	    <p><em>Note that the dirty schedulers functionality is experimental</em>, and
-	    that you have to enable support for dirty schedulers when building OTP in
-	    order to try out the functionality.</p>
+          <tag><c>dirty_cpu_schedulers_online</c></tag>
+          <item>
+          <marker id="system_info_dirty_cpu_schedulers_online"></marker>
+            <p>Returns the number of dirty CPU schedulers online.
+              The return value satisfies
+              <c><![CDATA[1 <= DirtyCPUSchedulersOnline <= N]]></c>,
+              where <c>N</c> is the smallest of the return values of
+              <c>erlang:system_info(dirty_cpu_schedulers)</c> and
+              <c>erlang:system_info(schedulers_online)</c>.</p>
+            <p>The number of dirty CPU schedulers online can be set at
+              startup by passing command-line flag
+              <seealso marker="erts:erl#+SDcpu">+SDcpu</seealso> in
+              <c>erl(1)</c>.</p>
+	    <p>Notice that the dirty schedulers functionality is
+              experimental. Enable support for dirty schedulers when
+              building OTP to try out the functionality.</p>
             <p>For more information, see
               <seealso marker="#system_info_dirty_cpu_schedulers">erlang:system_info(dirty_cpu_schedulers)</seealso>,
               <seealso marker="#system_info_dirty_io_schedulers">erlang:system_info(dirty_io_schedulers)</seealso>,
               <seealso marker="#system_info_schedulers_online">erlang:system_info(schedulers_online)</seealso>, and
-              <seealso marker="#system_flag_dirty_cpu_schedulers_online">erlang:system_flag(dirty_cpu_schedulers_online, DirtyCPUSchedulersOnline)</seealso>.
-            </p>
-          </item>
-          <tag><marker id="system_info_dirty_io_schedulers"><c>dirty_io_schedulers</c></marker></tag>
-          <item>
-            <p>Returns the number of dirty I/O schedulers as an integer. Dirty I/O schedulers
-              execute I/O-bound native functions such as NIFs and linked-in driver code that
-              cannot be managed cleanly by the emulator's normal schedulers.
-            </p>
-            <p>This value can be set on startup by passing
-              the <seealso marker="erts:erl#+SDio">+SDio</seealso> command line flag, see
-              <seealso marker="erts:erl#+SDio">erl(1)</seealso>.
-            </p>
-	    <p><em>Note that the dirty schedulers functionality is experimental</em>, and
-	    that you have to enable support for dirty schedulers when building OTP in
-	    order to try out the functionality.</p>
+              <seealso marker="#system_flag_dirty_cpu_schedulers_online">erlang:system_flag(dirty_cpu_schedulers_online, DirtyCPUSchedulersOnline)</seealso>.</p>
+          </item>
+          <tag><c>dirty_io_schedulers</c></tag>
+          <item>
+          <marker id="system_info_dirty_io_schedulers"></marker>
+            <p>Returns the number of dirty I/O schedulers as an integer.
+              Dirty I/O schedulers execute I/O-bound native functions,
+              such as NIFs and linked-in driver code, which cannot be
+              managed cleanly by the normal emulator schedulers.</p>
+            <p>This value can be set at startup by passing command-line
+              argument <seealso marker="erts:erl#+SDio">+SDio</seealso>
+              in <c>erl(1)</c>.</p>
+	    <p>Notice that the dirty schedulers functionality is
+              experimental. Enable support for dirty schedulers when
+              building OTP to try out the functionality.</p>
             <p>For more information, see
               <seealso marker="#system_info_dirty_cpu_schedulers">erlang:system_info(dirty_cpu_schedulers)</seealso>,
               <seealso marker="#system_info_dirty_cpu_schedulers_online">erlang:system_info(dirty_cpu_schedulers_online)</seealso>, and
-              <seealso marker="#system_flag_dirty_cpu_schedulers_online">erlang:system_flag(dirty_cpu_schedulers_online, DirtyCPUSchedulersOnline)</seealso>.
-            </p>
+              <seealso marker="#system_flag_dirty_cpu_schedulers_online">erlang:system_flag(dirty_cpu_schedulers_online, DirtyCPUSchedulersOnline)</seealso>.</p>
           </item>
           <tag><c>dist</c></tag>
           <item>
             <p>Returns a binary containing a string of distribution
               information formatted as in Erlang crash dumps. For more
-              information see the <seealso marker="erts:crash_dump">"How to interpret the Erlang crash dumps"</seealso>
-              chapter in the ERTS User's Guide.</p>
+              information, see Section
+              <seealso marker="erts:crash_dump">"How to interpret the Erlang crash dumps"</seealso>
+              in the User's Guide.</p>
           </item>
-          <tag><marker id="system_info_dist_buf_busy_limit"><c>dist_buf_busy_limit</c></marker></tag>
+          <tag><c>dist_buf_busy_limit</c></tag>
           <item>
+          <marker id="system_info_dist_buf_busy_limit"></marker>
             <p>Returns the value of the distribution buffer busy limit
-	    in bytes. This limit can be set on startup by passing the
-	    <seealso marker="erts:erl#+zdbbl">+zdbbl</seealso> command line
-	    flag to <c>erl</c>.</p>
+	    in bytes. This limit can be set at startup by passing
+            command-line flag
+	    <seealso marker="erts:erl#+zdbbl">+zdbbl</seealso>
+            to <c>erl</c>.</p>
           </item>
           <tag><c>dist_ctrl</c></tag>
           <item>
             <p>Returns a list of tuples
-              <c>{Node, ControllingEntity}</c>, one entry for each
-              connected remote node. The <c><anno>Node</anno></c> is the name of the
-              node and the <c><anno>ControllingEntity</anno></c> is the port or pid
-              responsible for the communication to that node. More
-              specifically, the <c><anno>ControllingEntity</anno></c> for nodes
-              connected via TCP/IP (the normal case) is the socket
-              actually used in communication with the specific node.</p>
+              <c>{<anno>Node</anno>, <anno>ControllingEntity</anno>}</c>,
+              one entry for each connected remote node.
+              <c><anno>Node</anno></c> is the node name
+              and <c><anno>ControllingEntity</anno></c> is the port or process
+              identifier responsible for the communication to that node.
+              More specifically, <c><anno>ControllingEntity</anno></c> for
+              nodes connected through TCP/IP (the normal case) is the socket
+              used in communication with the specific node.</p>
           </item>
           <tag><c>driver_version</c></tag>
           <item>
-            <p>Returns a string containing the erlang driver version
-              used by the runtime system. It will be on the form
+            <p>Returns a string containing the Erlang driver version
+              used by the runtime system. It has the form
               <seealso marker="erts:erl_driver#version_management">"&lt;major ver&gt;.&lt;minor ver&gt;"</seealso>.</p>
           </item>
           <tag><c>dynamic_trace</c></tag>
           <item>
             <p>Returns an atom describing the dynamic trace framework
-            compiled into the virtual machine. It can currently be either
-            <c>dtrace</c>, <c>systemtap</c> or <c>none</c>. For a
-            commercial or standard build, this is always <c>none</c>,
-            the other return values indicate a custom configuration
-            (e.g. <c>./configure --with-dynamic-trace=dtrace</c>). See
-            the <seealso marker="runtime_tools:dyntrace">dyntrace
-            </seealso> manual page and the
+            compiled into the virtual machine. It can be
+            <c>dtrace</c>, <c>systemtap</c>, or <c>none</c>. For a
+            commercial or standard build, it is always <c>none</c>.
+            The other return values indicate a custom configuration
+            (for example, <c>./configure --with-dynamic-trace=dtrace</c>).
+            For more information about dynamic tracing, see the
+            <seealso marker="runtime_tools:dyntrace">dyntrace</seealso>
+            manual page and the
             <c>README.dtrace</c>/<c>README.systemtap</c> files in the
-            Erlang source code top directory for more information
-            about dynamic tracing.</p>
+            Erlang source code top directory.</p>
           </item>
           <tag><c>dynamic_trace_probes</c></tag>
           <item>
-            <p>Returns a <c>boolean()</c> indicating if dynamic trace probes
-            (either dtrace or systemtap) are built into the
-            emulator. This can only be <c>true</c> if the virtual
-            machine was built for dynamic tracing
-            (i.e. <c>system_info(dynamic_trace)</c> returns
+            <p>Returns a <c>boolean()</c> indicating if dynamic trace
+            probes (<c>dtrace</c> or <c>systemtap</c>) are built into
+            the emulator. This can only be <c>true</c> if the Virtual
+            Machine was built for dynamic tracing (that is,
+            <c>system_info(dynamic_trace)</c> returns
             <c>dtrace</c> or <c>systemtap</c>).</p>
           </item>
 	  <tag><marker id="system_info_end_time"/><c>end_time</c></tag>
@@ -6433,42 +7033,43 @@ ok
           <tag><c>elib_malloc</c></tag>
           <item>
             <p>This option will be removed in a future release.
-	      The return value will always be <c>false</c> since
-	      the elib_malloc allocator has been removed.</p>
+	      The return value will always be <c>false</c>, as the
+	      <c>elib_malloc</c> allocator has been removed.</p>
           </item>
-          <tag><marker id="system_info_eager_check_io"><c>eager_check_io</c></marker></tag>
+          <tag><marker id="system_info_eager_check_io"/><c>eager_check_io</c></tag>
           <item>
             <p>
-	      Returns the value of the <c>erl</c>
-	      <seealso marker="erl#+secio">+secio</seealso> command line
-	      flag which is either <c>true</c> or <c>false</c>. See the
+	      Returns the value of the <c>erl</c> command line flag
+	      <seealso marker="erl#+secio">+secio</seealso>
+	      which is either <c>true</c> or <c>false</c>. See the
 	      documentation of the command line flag for information about
 	      the different values.
 	    </p>
           </item>
           <tag><c>ets_limit</c></tag>
           <item>
-            <p>Returns the maximum number of ETS tables allowed. This limit
-              can be increased on startup by passing the <seealso
-              marker="erts:erl#+e">+e</seealso> command line flag to
-              <c>erl</c> or by setting the environment variable
-              <c>ERL_MAX_ETS_TABLES</c> before starting the Erlang runtime
-              system.</p>
+            <p>Returns the maximum number of ETS tables allowed. This
+              limit can be increased at startup by passing
+              command-line flag
+              <seealso marker="erts:erl#+e">+e</seealso> to
+              <c>erl(1)</c> or by setting environment variable
+              <c>ERL_MAX_ETS_TABLES</c> before starting the Erlang
+              runtime system.</p>
           </item>
           <tag><c>fullsweep_after</c></tag>
           <item>
-            <p>Returns <c>{fullsweep_after, integer() >= 0}</c> which is the
-              <c>fullsweep_after</c> garbage collection setting used
-              by default. For more information see
-              <c>garbage_collection</c> described below.</p>
+            <p>Returns <c>{fullsweep_after, integer() >= 0}</c>, which is
+              the <c>fullsweep_after</c> garbage collection setting used
+              by default. For more information, see
+              <c>garbage_collection</c> described in the following.</p>
           </item>
           <tag><c>garbage_collection</c></tag>
           <item>
             <p>Returns a list describing the default garbage collection
               settings. A process spawned on the local node by a
-              <c>spawn</c> or <c>spawn_link</c> will use these
+              <c>spawn</c> or <c>spawn_link</c> uses these
               garbage collection settings. The default settings can be
-              changed by use of
+              changed by using
               <seealso marker="#system_flag/2">system_flag/2</seealso>.
               <seealso marker="#spawn_opt/4">spawn_opt/4</seealso>
               can spawn a process that does not use the default
@@ -6482,8 +7083,8 @@ ok
           </item>
           <tag><c>heap_type</c></tag>
           <item>
-            <p>Returns the heap type used by the current emulator.
-              Currently only the following heap type exists:</p>
+            <p>Returns the heap type used by the current emulator. One
+              heap type exists:</p>
             <taglist>
               <tag><c>private</c></tag>
               <item>
@@ -6498,51 +7099,51 @@ ok
           <item>
             <p>Returns a binary containing a string of miscellaneous
               system information formatted as in Erlang crash dumps.
-              For more information see the
-              <seealso marker="erts:crash_dump">"How to interpret the Erlang crash dumps"</seealso> chapter in the ERTS
-              User's Guide.</p>
+              For more information, see Section
+              <seealso marker="erts:crash_dump">"How to interpret the Erlang crash dumps"</seealso>
+              in the User's Guide.</p>
           </item>
           <tag><c>kernel_poll</c></tag>
           <item>
             <p>Returns <c>true</c> if the emulator uses some kind of
-              kernel-poll implementation; otherwise, <c>false</c>.</p>
+              kernel-poll implementation, otherwise <c>false</c>.</p>
           </item>
           <tag><c>loaded</c></tag>
           <item>
             <p>Returns a binary containing a string of loaded module
               information formatted as in Erlang crash dumps. For more
-              information see the <seealso marker="erts:crash_dump">"How to interpret the Erlang crash dumps"</seealso> chapter
-              in the ERTS User's Guide.</p>
+              information, see Section
+              <seealso marker="erts:crash_dump">"How to interpret the Erlang crash dumps"</seealso>
+              in the User's Guide.</p>
           </item>
-	  <tag><marker id="logical_processors"><c>logical_processors</c></marker></tag>
+	  <tag><c>logical_processors</c></tag>
           <item>
+          <marker id="logical_processors"></marker>
             <p>Returns the detected number of logical processors configured
-	       on the system. The return value is either an integer, or
-	       the atom <c>unknown</c> if the emulator wasn't able to
-	       detect logical processors configured.
-	    </p>
+	       in the system. The return value is either an integer, or
+	       the atom <c>unknown</c> if the emulator cannot
+	       detect the configured logical processors.</p>
           </item>
-	  <tag><marker id="logical_processors_available"><c>logical_processors_available</c></marker></tag>
+	  <tag><c>logical_processors_available</c></tag>
           <item>
-            <p>Returns the detected number of logical processors available to
-	       the Erlang runtime system. The return value is either an
-	       integer, or the atom <c>unknown</c> if the emulator wasn't
-	       able to detect logical processors available. The number
-	       of logical processors available is less than or equal to
-	       the number of <seealso marker="#logical_processors_online">logical
-	       processors online</seealso>.
-	    </p>
+          <marker id="logical_processors_available"></marker>
+            <p>Returns the detected number of logical processors available
+	       to the Erlang runtime system. The return value is either an
+	       integer, or the atom <c>unknown</c> if the emulator
+	       cannot detect the available logical processors. The number
+	       of available logical processors is less than or equal to
+	       the number of
+               <seealso marker="#logical_processors_online">logical processors online</seealso>.</p>
           </item>
-	  <tag><marker id="logical_processors_online"><c>logical_processors_online</c></marker></tag>
+	  <tag><c>logical_processors_online</c></tag>
           <item>
+          <marker id="logical_processors_online"></marker>
             <p>Returns the detected number of logical processors online on
 	       the system. The return value is either an integer,
-	       or the atom <c>unknown</c> if the emulator wasn't able to
+	       or the atom <c>unknown</c> if the emulator cannot
 	       detect logical processors online. The number of logical
 	       processors online is less than or equal to the number of
-	       <seealso marker="#logical_processors">logical processors
-	       configured</seealso>.
-	    </p>
+	       <seealso marker="#logical_processors">logical processors configured</seealso>.</p>
           </item>
           <tag><c>machine</c></tag>
           <item>
@@ -6550,27 +7151,30 @@ ok
           </item>
 	  <tag><c>min_heap_size</c></tag>
           <item>
-	      <p>Returns <c>{min_heap_size, <anno>MinHeapSize</anno>}</c> where <c><anno>MinHeapSize</anno></c> is the current system wide
-	      minimum heap size for spawned processes.</p>
+	      <p>Returns <c>{min_heap_size, <anno>MinHeapSize</anno>}</c>,
+              where <c><anno>MinHeapSize</anno></c> is the current
+              system-wide minimum heap size for spawned processes.</p>
           </item>
 	  <tag><c>min_bin_vheap_size</c></tag>
           <item>
-	      <p>Returns <c>{min_bin_vheap_size, <anno>MinBinVHeapSize</anno>}</c> where <c><anno>MinBinVHeapSize</anno></c> is the current system wide
+	      <p>Returns <c>{min_bin_vheap_size,
+              <anno>MinBinVHeapSize</anno>}</c>, where
+              <c><anno>MinBinVHeapSize</anno></c> is the current system-wide
 	      minimum binary virtual heap size for spawned processes.</p>
           </item>
           <tag><c>modified_timing_level</c></tag>
           <item>
-            <p>Returns the modified timing level (an integer) if
-              modified timing has been enabled; otherwise,
-              <c>undefined</c>. See the <c>+T</c> command line flag
-              in the documentation of the
-              <seealso marker="erts:erl#+T">erl(1)</seealso>
-              command for more information on modified timing.</p>
+            <p>Returns the modified timing-level (an integer) if
+              modified timing is enabled, otherwise, <c>undefined</c>.
+              For more information about modified timing, see
+              command-line flag
+              <seealso marker="erts:erl#+T">+T</seealso>
+              in <c>erl(1)</c></p>
           </item>
-          <tag><marker id="system_info_multi_scheduling"><c>multi_scheduling</c></marker></tag>
+          <tag><c>multi_scheduling</c></tag>
           <item>
-            <p>Returns <c>disabled</c>, <c>blocked</c>, or <c>enabled</c>.
-              A description of the return values:</p>
+          <marker id="system_info_multi_scheduling"></marker>
+            <p>Returns <c>disabled</c>, <c>blocked</c>, or <c>enabled</c>:</p>
             <taglist>
               <tag><c>disabled</c></tag>
               <item>
@@ -6581,93 +7185,109 @@ ok
               <tag><c>blocked</c></tag>
               <item>
                 <p>The emulator has more than one scheduler thread,
-                  but all scheduler threads but one have been blocked,
-                  i.e., only one scheduler thread will schedule
-                  Erlang processes and execute Erlang code.</p>
+                  but all scheduler threads except one are blocked,
+                  that is, only one scheduler thread schedules
+                  Erlang processes and executes Erlang code.</p>
               </item>
               <tag><c>enabled</c></tag>
               <item>
                 <p>The emulator has more than one scheduler thread,
-                  and no scheduler threads have been blocked, i.e.,
-                  all available scheduler threads will schedule
+                  and no scheduler threads are blocked, that is,
+                  all available scheduler threads schedule
                   Erlang processes and execute Erlang code.</p>
               </item>
             </taglist>
-            <p>See also <seealso marker="#system_flag_multi_scheduling">erlang:system_flag(multi_scheduling, BlockState)</seealso>,
-              <seealso marker="#system_info_multi_scheduling_blockers">erlang:system_info(multi_scheduling_blockers)</seealso>, and
+            <p>See also
+              <seealso marker="#system_flag_multi_scheduling">erlang:system_flag(multi_scheduling, BlockState)</seealso>,
+              <seealso marker="#system_info_multi_scheduling_blockers">erlang:system_info(multi_scheduling_blockers)</seealso>,
+              and
               <seealso marker="#system_info_schedulers">erlang:system_info(schedulers)</seealso>.</p>
           </item>
-          <tag><marker id="system_info_multi_scheduling_blockers"><c>multi_scheduling_blockers</c></marker></tag>
+          <tag><c>multi_scheduling_blockers</c></tag>
           <item>
-            <p>Returns a list of <c><anno>PID</anno></c>s when multi-scheduling
-              is blocked; otherwise, the empty list. The <c><anno>PID</anno></c>s
-              in the list is <c><anno>PID</anno></c>s of the processes currently
-              blocking multi-scheduling. A <c><anno>PID</anno></c> will only be
-              present once in the list, even if the corresponding
+          <marker id="system_info_multi_scheduling_blockers"></marker>
+            <p>Returns a list of <c><anno>Pid</anno></c>s when
+              multi-scheduling is blocked, otherwise the empty list is
+              returned. The <c><anno>Pid</anno></c>s in the list
+	      represent all the processes currently
+              blocking multi-scheduling. A <c><anno>Pid</anno></c> occurs
+              only once in the list, even if the corresponding
               process has blocked multiple times.</p>
-            <p>See also <seealso marker="#system_flag_multi_scheduling">erlang:system_flag(multi_scheduling, BlockState)</seealso>,
-              <seealso marker="#system_info_multi_scheduling">erlang:system_info(multi_scheduling)</seealso>, and
+            <p>See also
+              <seealso marker="#system_flag_multi_scheduling">erlang:system_flag(multi_scheduling, BlockState)</seealso>,
+              <seealso marker="#system_info_multi_scheduling">erlang:system_info(multi_scheduling)</seealso>,
+              and
               <seealso marker="#system_info_schedulers">erlang:system_info(schedulers)</seealso>.</p>
           </item>
           <tag><c>nif_version</c></tag>
           <item>
-            <p>Returns a string containing the erlang NIF version
-              used by the runtime system. It will be on the form "&lt;major ver&gt;.&lt;minor ver&gt;".</p>
+            <p>Returns a string containing the version of the Erlang NIF interface
+              used by the runtime system. It is on the form
+	      "&lt;major ver&gt;.&lt;minor ver&gt;".</p>
+          </item>
+          <tag><marker id="system_info_message_queue_data"><c>message_queue_data</c></marker></tag>
+          <item>
+            <p>Returns the default value of the <c>message_queue_data</c>
+	    process flag which is either <c>off_heap</c>, <c>on_heap</c>, or <c>mixed</c>.
+	    This default is set by the <c>erl</c> command line argument
+	    <seealso marker="erl#+xmqd"><c>+xmqd</c></seealso>. For more information on the
+	    <c>message_queue_data</c> process flag, see documentation of
+            <seealso marker="#process_flag_message_queue_data"><c>process_flag(message_queue_data,
+	    MQD)</c></seealso>.</p>
           </item>
           <tag><marker id="system_info_otp_release"><c>otp_release</c></marker></tag>
           <item>
+          <marker id="system_info_otp_release"></marker>
             <p>Returns a string containing the OTP release number of the
-	    OTP release that the currently executing ERTS application is
+	    OTP release that the currently executing <c>ERTS</c> application is
             part of.</p>
-            <p>As of OTP release 17, the OTP release number corresponds to
-            the major OTP version number. There is no
-	    <c>erlang:system_info()</c> argument giving the exact OTP
-	    version. This since the exact OTP version in the general case
-	    is hard to determine.  For more information see 
-	    <seealso marker="doc/system_principles:versions">the
-	    documentation of versions in the system principles
-	    guide</seealso>.</p>
+            <p>As from OTP 17, the OTP release number corresponds to
+            the major OTP version number. No
+	    <c>erlang:system_info()</c> argument gives the exact OTP
+	    version. This is because the exact OTP version in the general case
+	    is difficult to determine. For more information, see the description
+	    of versions in <seealso marker="doc/system_principles:versions">
+	    System principles</seealso> in System Documentation.</p>
           </item>
-          <tag><marker id="system_info_os_monotonic_time_source"><c>os_monotonic_time_source</c></marker></tag>
+          <tag><marker id="system_info_os_monotonic_time_source"/><c>os_monotonic_time_source</c></tag>
           <item>
             <p>Returns a list containing information about the source of
 	    <seealso marker="erts:time_correction#OS_Monotonic_Time">OS
 	    monotonic time</seealso> that is used by the runtime system.</p>
-	    <p>In case <c>[]</c> is returned, no OS monotonic time is
+	    <p>If <c>[]</c> is returned, no OS monotonic time is
 	    available. The list contains two-tuples with <c>Key</c>s
 	    as first element, and <c>Value</c>s as second element. The
-	    order if these tuples is undefined. Currently the following
-	    tuples may be part of the list, but more tuples may be
+	    order of these tuples is undefined. The following
+	    tuples can be part of the list, but more tuples can be
 	    introduced in the future:</p>
 	    <taglist>
 	      <tag><c>{function, Function}</c></tag>
-	      <item><p><c>Function</c> is the name of the funcion
+	      <item><p><c>Function</c> is the name of the function
 	      used. This tuple always exist if OS monotonic time is
 	      available to the runtime system.</p></item>
 
 	      <tag><c>{clock_id, ClockId}</c></tag>
 	      <item><p>This tuple only exist if <c>Function</c>
 	      can be used with different clocks. <c>ClockId</c>
-	      corresponds to the clock identifer used when calling
+	      corresponds to the clock identifier used when calling
 	      <c>Function</c>.</p></item>
 
 	      <tag><c>{resolution, OsMonotonicTimeResolution}</c></tag>
 	      <item><p>Highest possible
 	      <seealso marker="time_correction#Time_Resolution">resolution</seealso>
 	      of current OS monotonic time source as parts per
-	      second. If no resolution information can be retreived
-	      from the OS, <c>OsMonotonicTimeResolution</c> will be
+	      second. If no resolution information can be retrieved
+	      from the OS, <c>OsMonotonicTimeResolution</c> is
 	      set to the resolution of the time unit of
 	      <c>Function</c>s return value. That is, the actual
-	      resolution may be lower than
+	      resolution can be lower than
 	      <c>OsMonotonicTimeResolution</c>. Also note that
 	      the resolution does not say anything about the
 	      <seealso marker="time_correction#Time_Accuracy">accuracy</seealso>,
-	      and that the
+	      and whether the
 	      <seealso marker="time_correction#Time_Precision">precision</seealso>
-	      might not align with the resolution. You do,
-	      however, know that the precision won't be
-	      better than
+	      do align with the resolution. You do,
+	      however, know that the precision is not better than
 	      <c>OsMonotonicTimeResolution</c>.</p></item>
 
 	      <tag><c>{extended, Extended}</c></tag>
@@ -6692,15 +7312,15 @@ ok
 	      <seealso marker="#type_time_unit">time unit</seealso>.</p></item>
 	    </taglist>
           </item>
-          <tag><marker id="system_info_os_system_time_source"><c>os_system_time_source</c></marker></tag>
+          <tag><marker id="system_info_os_system_time_source"/><c>os_system_time_source</c></tag>
           <item>
             <p>Returns a list containing information about the source of
 	    <seealso marker="erts:time_correction#OS_System_Time">OS
 	    system time</seealso> that is used by the runtime system.</p>
 	    <p>The list contains two-tuples with <c>Key</c>s
 	    as first element, and <c>Value</c>s as second element. The
-	    order if these tuples is undefined. Currently the following
-	    tuples may be part of the list, but more tuples may be
+	    order if these tuples is undefined. The following
+	    tuples can be part of the list, but more tuples can be
 	    introduced in the future:</p>
 	    <taglist>
 	      <tag><c>{function, Function}</c></tag>
@@ -6710,26 +7330,25 @@ ok
 	      <tag><c>{clock_id, ClockId}</c></tag>
 	      <item><p>This tuple only exist if <c>Function</c>
 	      can be used with different clocks. <c>ClockId</c>
-	      corresponds to the clock identifer used when calling
+	      corresponds to the clock identifier used when calling
 	      <c>Function</c>.</p></item>
 
 	      <tag><c>{resolution, OsSystemTimeResolution}</c></tag>
 	      <item><p>Highest possible
 	      <seealso marker="time_correction#Time_Resolution">resolution</seealso>
 	      of current OS system time source as parts per
-	      second. If no resolution information can be retreived
-	      from the OS, <c>OsSystemTimeResolution</c> will be
+	      second. If no resolution information can be retrieved
+	      from the OS, <c>OsSystemTimeResolution</c> is
 	      set to the resolution of the time unit of
 	      <c>Function</c>s return value. That is, the actual
 	      resolution may be lower than
 	      <c>OsSystemTimeResolution</c>. Also note that
 	      the resolution does not say anything about the
 	      <seealso marker="time_correction#Time_Accuracy">accuracy</seealso>,
-	      and that the
+	      and whether the
 	      <seealso marker="time_correction#Time_Precision">precision</seealso>
-	      might not align with the resolution. You do,
-	      however, know that the precision won't be
-	      better than
+	      do align with the resolution. You do,
+	      however, know that the precision is not better than
 	      <c>OsSystemTimeResolution</c>.</p></item>
 
 	      <tag><c>{parallel, Parallel}</c></tag>
@@ -6745,130 +7364,136 @@ ok
 	      <seealso marker="#type_time_unit">time unit</seealso>.</p></item>
 	    </taglist>
           </item>
-	  <tag><marker id="system_info_port_parallelism"><c>port_parallelism</c></marker></tag>
-	  <item><p>Returns the default port parallelism scheduling hint used.
-	  For more information see the
-	  <seealso marker="erl#+spp">+spp</seealso> command line argument
-	  of <seealso marker="erl">erl(1)</seealso>.</p></item>
+	  <tag><c>port_parallelism</c></tag>
+	  <item>
+	  <marker id="system_info_port_parallelism"></marker>
+	  <p>Returns the default port parallelism scheduling hint used.
+	  For more information, see command-line argument
+	  <seealso marker="erl#+spp">+spp</seealso> in <c>erl(1)</c>.</p></item>
           <tag><marker id="system_info_port_count"/><c>port_count</c></tag>
           <item>
-            <p>Returns the number of ports currently existing at
-              the local node as an integer. The same value as
-              <c>length(erlang:ports())</c> returns, but more efficient.</p>
+            <p>Returns the number of ports currently existing at the
+              local node. The value is given as an integer. This is
+              the same value as returned by
+              <c>length(erlang:ports())</c>, but more efficient.</p>
           </item>
-          <tag><marker id="system_info_port_limit"><c>port_limit</c></marker></tag>
+          <tag><c>port_limit</c></tag>
           <item>
+          <marker id="system_info_port_limit"></marker>
             <p>Returns the maximum number of simultaneously existing
-              ports at the local node as an integer. This limit
-              can be configured at startup by using the
-              <seealso marker="erl#+Q">+Q</seealso>
-	      command line flag of
-	      <seealso marker="erl">erl(1)</seealso>.</p>
+              ports at the local node as an integer. This limit can be
+              configured at startup by using command-line flag
+              <seealso marker="erl#+Q">+Q</seealso> in <c>erl(1)</c>.</p>
           </item>
           <tag><marker id="system_info_process_count"/><c>process_count</c></tag>
           <item>
-            <p>Returns the number of processes currently existing at
-              the local node as an integer. The same value as
-              <c>length(processes())</c> returns, but more efficient.</p>
+            <p>Returns the number of processes currently existing at the
+              local node. The value is given as an integer. This is
+              the same value as returned by
+              <c>length(processes())</c>, but more efficient.</p>
           </item>
-          <tag><marker id="system_info_process_limit"><c>process_limit</c></marker></tag>
+          <tag><c>process_limit</c></tag>
           <item>
+          <marker id="system_info_process_limit"></marker>
             <p>Returns the maximum number of simultaneously existing
-              processes at the local node as an integer. This limit
-              can be configured at startup by using the
-              <seealso marker="erl#+P">+P</seealso>
-	      command line flag of
-	      <seealso marker="erl">erl(1)</seealso>.</p>
+              processes at the local node. The value is given as an
+              integer. This limit can be configured at startup by using
+              command-line flag <seealso marker="erl#+P">+P</seealso>
+	      in <c>erl(1)</c>.</p>
           </item>
           <tag><c>procs</c></tag>
           <item>
             <p>Returns a binary containing a string of process and port
               information formatted as in Erlang crash dumps. For more
-              information see the <seealso marker="erts:crash_dump">"How to interpret the Erlang crash dumps"</seealso> chapter
-              in the ERTS User's Guide.</p>
+              information, see Section
+              <seealso marker="erts:crash_dump">"How to interpret the Erlang crash dumps"</seealso>
+              in the User's Guide.</p>
           </item>
-          <tag><marker id="system_info_scheduler_bind_type"><c>scheduler_bind_type</c></marker></tag>
+          <tag><c>scheduler_bind_type</c></tag>
           <item>
-	    <p>Returns information on how user has requested
+          <marker id="system_info_scheduler_bind_type"></marker>
+	    <p>Returns information about how the user has requested
 	       schedulers to be bound or not bound.</p>
-	    <p><em>NOTE:</em> Even though user has requested
-	       schedulers to be bound, they might have silently failed
-	       to bind. In order to inspect actual scheduler bindings call
-	       <seealso marker="#system_info_scheduler_bindings">erlang:system_info(scheduler_bindings)</seealso>.
-	    </p>
-	    <p>For more information, see
-	       the <c>erl</c> <seealso marker="erts:erl#+sbt">+sbt</seealso>
-	       command line argument, and
-	       <seealso marker="#system_info_scheduler_bindings">erlang:system_info(scheduler_bindings)</seealso>.
-	    </p>
-          </item>
-          <tag><marker id="system_info_scheduler_bindings"><c>scheduler_bindings</c></marker></tag>
-          <item>
-	    <p>Returns information on currently used scheduler
+	    <p>Notice that even though a user has requested
+	       schedulers to be bound, they can silently have failed
+	       to bind. To inspect the scheduler bindings, call
+	       <seealso marker="#system_info_scheduler_bindings">erlang:system_info(scheduler_bindings)</seealso>.</p>
+	    <p>For more information, see command-line argument
+	       <seealso marker="erts:erl#+sbt">+sbt</seealso>
+	       in <c>erl(1)</c> and
+	       <seealso marker="#system_info_scheduler_bindings">erlang:system_info(scheduler_bindings)</seealso>.</p>
+          </item>
+          <tag><c>scheduler_bindings</c></tag>
+          <item>
+          <marker id="system_info_scheduler_bindings"></marker>
+	    <p>Returns information about the currently used scheduler
 	       bindings.</p>
 	    <p>A tuple of a size equal to
-	       <seealso marker="#system_info_schedulers">erlang:system_info(schedulers)</seealso> is returned. The elements of the tuple are integers
+	       <seealso marker="#system_info_schedulers">erlang:system_info(schedulers)</seealso>
+               is returned. The tuple elements are integers
 	       or the atom <c>unbound</c>. Logical processor identifiers
 	       are represented as integers. The <c>N</c>th
 	       element of the tuple equals the current binding for
 	       the scheduler with the scheduler identifier equal to
-	       <c>N</c>. E.g., if the schedulers have been bound,
+	       <c>N</c>. For example, if the schedulers are bound,
 	       <c>element(erlang:system_info(scheduler_id),
-	       erlang:system_info(scheduler_bindings))</c> will return
+	       erlang:system_info(scheduler_bindings))</c> returns
 	       the identifier of the logical processor that the calling
-	       process is executing on.
-	    </p>
-	    <p>Note that only schedulers online can be bound to logical
+	       process is executing on.</p>
+	    <p>Notice that only schedulers online can be bound to logical
 	       processors.</p>
-	    <p>For more information, see
-	       the <c>erl</c> <seealso marker="erts:erl#+sbt">+sbt</seealso>
-	       command line argument,
+	    <p>For more information, see command-line argument
+	       <seealso marker="erts:erl#+sbt">+sbt</seealso>
+	       in <c>erl(1)</c> and
 	       <seealso marker="#system_info_schedulers_online">erlang:system_info(schedulers_online)</seealso>.
 	    </p>
           </item>
-          <tag><marker id="system_info_scheduler_id"><c>scheduler_id</c></marker></tag>
+          <tag><c>scheduler_id</c></tag>
           <item>
-            <p>Returns the scheduler id (<c>SchedulerId</c>) of the
+          <marker id="system_info_scheduler_id"></marker>
+            <p>Returns the scheduler ID (<c>SchedulerId</c>) of the
               scheduler thread that the calling process is executing
-              on. <c><anno>SchedulerId</anno></c> is a positive integer; where
-              <c><![CDATA[1 <= SchedulerId <= erlang:system_info(schedulers)]]></c>. See also
+              on. <c><anno>SchedulerId</anno></c> is a positive integer,
+              where
+              <c><![CDATA[1 <= SchedulerId <= erlang:system_info(schedulers)]]></c>.
+              See also
               <seealso marker="#system_info_schedulers">erlang:system_info(schedulers)</seealso>.</p>
           </item>
-          <tag><marker id="system_info_schedulers"><c>schedulers</c></marker></tag>
+          <tag><c>schedulers</c></tag>
           <item>
+          <marker id="system_info_schedulers"></marker>
             <p>Returns the number of scheduler threads used by
               the emulator. Scheduler threads online schedules Erlang
               processes and Erlang ports, and execute Erlang code
-              and Erlang linked in driver code.</p>
+              and Erlang linked-in driver code.</p>
             <p>The number of scheduler threads is determined at
-              emulator boot time and cannot be changed after
-              that. The amount of schedulers online can
-	      however be changed at any time.</p>
-            <p>See also <seealso marker="#system_flag_schedulers_online">erlang:system_flag(schedulers_online, SchedulersOnline)</seealso>,
+              emulator boot time and cannot be changed later.
+              However, the number of schedulers online can
+	      be changed at any time.</p>
+            <p>See also
+              <seealso marker="#system_flag_schedulers_online">erlang:system_flag(schedulers_online, SchedulersOnline)</seealso>,
 	      <seealso marker="#system_info_schedulers_online">erlang:system_info(schedulers_online)</seealso>,
 	      <seealso marker="#system_info_scheduler_id">erlang:system_info(scheduler_id)</seealso>,
               <seealso marker="#system_flag_multi_scheduling">erlang:system_flag(multi_scheduling, BlockState)</seealso>,
-              <seealso marker="#system_info_multi_scheduling">erlang:system_info(multi_scheduling)</seealso>, and
-              and <seealso marker="#system_info_multi_scheduling_blockers">erlang:system_info(multi_scheduling_blockers)</seealso>.</p>
+              <seealso marker="#system_info_multi_scheduling">erlang:system_info(multi_scheduling)</seealso>,
+              and
+              <seealso marker="#system_info_multi_scheduling_blockers">erlang:system_info(multi_scheduling_blockers)</seealso>.</p>
           </item>
-          <tag><marker id="system_info_schedulers_online"><c>schedulers_online</c></marker></tag>
+          <tag><c>schedulers_online</c></tag>
           <item>
-            <p>Returns the amount of schedulers online. The scheduler
-	    identifiers of schedulers online satisfy the following
-	    relationship:
-	    <c><![CDATA[1 <= SchedulerId <= erlang:system_info(schedulers_online)]]></c>.
-	    </p>
+          <marker id="system_info_schedulers_online"></marker>
+            <p>Returns the number of schedulers online. The scheduler
+	    identifiers of schedulers online satisfy the relationship
+	    <c><![CDATA[1 <= SchedulerId <= erlang:system_info(schedulers_online)]]></c>.</p>
 	    <p>For more information, see
-	    <seealso marker="#system_info_schedulers">erlang:system_info(schedulers)</seealso>,
+	    <seealso marker="#system_info_schedulers">erlang:system_info(schedulers)</seealso>
 	    and
-	    <seealso marker="#system_flag_schedulers_online">erlang:system_flag(schedulers_online, SchedulersOnline)</seealso>.
-	    </p>      <name name="system_info" arity="1" clause_i="49"/>
-
+	    <seealso marker="#system_flag_schedulers_online">erlang:system_flag(schedulers_online, SchedulersOnline)</seealso>.</p>
           </item>
           <tag><c>smp_support</c></tag>
           <item>
             <p>Returns <c>true</c> if the emulator has been compiled
-              with smp support; otherwise, <c>false</c>.</p>
+              with SMP support, otherwise <c>false</c> is returned.</p>
           </item>
 	  <tag><marker id="system_info_start_time"/><c>start_time</c></tag>
 	  <item><p>The <seealso marker="#monotonic_time/0">Erlang monotonic
@@ -6880,7 +7505,7 @@ ok
           <tag><c>system_version</c></tag>
           <item>
             <p>Returns a string containing version number and
-	    some important properties such as the number of schedulers.</p>
+	    some important properties, such as the number of schedulers.</p>
           </item>
           <tag><c>system_architecture</c></tag>
           <item>
@@ -6890,23 +7515,28 @@ ok
           <tag><c>threads</c></tag>
           <item>
             <p>Returns <c>true</c> if the emulator has been compiled
-              with thread support; otherwise, <c>false</c> is
-              returned.</p>
+              with thread support, otherwise <c>false</c> is returned.</p>
           </item>
-          <tag><marker id="system_info_thread_pool_size"><c>thread_pool_size</c></marker></tag>
+          <tag><c>thread_pool_size</c></tag>
           <item>
+          <marker id="system_info_thread_pool_size"></marker>
             <p>Returns the number of async threads in the async thread
               pool used for asynchronous driver calls
-              (<seealso marker="erts:erl_driver#driver_async">driver_async()</seealso>)
-              as an integer.</p>
+              (<seealso marker="erts:erl_driver#driver_async">driver_async()</seealso>).
+              The value is given as an integer.</p>
           </item>
-	  <tag><marker id="system_info_time_correction"/><c>time_correction</c></tag>
-	  <item><p>Returns a boolean value indicating whether
+
+	  <tag><c>time_correction</c></tag>
+	  <item>
+	    <marker id="system_info_time_correction"></marker>
+	    <p>Returns a boolean value indicating whether
 	  <seealso marker="time_correction#Time_Correction">time correction</seealso>
 	  is enabled or not.
 	  </p></item>
-	  <tag><marker id="system_info_time_offset"/><c>time_offset</c></tag>
-	  <item><p>Returns the state of the time offset:</p>
+	  <tag><c>time_offset</c></tag>
+	  <item>
+	    <marker id="system_info_time_offset"></marker>
+	    <p>Returns the state of the time offset:</p>
 	  <taglist>
 	    <tag><c>preliminary</c></tag>
 	    <item><p>The time offset is preliminary, and will be changed
@@ -6916,19 +7546,18 @@ ok
 	    time warp mode</seealso>.</p></item>
 
 	    <tag><c>final</c></tag>
-	    <item><p>The time offset is final. This
-	    either due to the use of the
+	    <item><p>The time offset is final. This either because
 	    <seealso marker="time_correction#No_Time_Warp_Mode">no
-	    time warp mode</seealso>, or due to the time offset having
-	    been finalized when using the
+	    time warp mode</seealso> is used, or because the time
+	    offset have been finalized when
 	    <seealso marker="time_correction#Single_Time_Warp_Mode">single
-	    time warp mode</seealso>.</p></item>
+	    time warp mode</seealso> is used.</p></item>
 
 	    <tag><c>volatile</c></tag>
-	    <item><p>The time offset is volatile. That is, it may
-	    change at any time. This due to the
+	    <item><p>The time offset is volatile. That is, it can
+	    change at any time. This is because
 	    <seealso marker="time_correction#Multi_Time_Warp_Mode">multi
-	    time warp mode</seealso> being used.</p></item>
+	    time warp mode</seealso> is used.</p></item>
 	  </taglist>
 	  </item>
 	  <tag><marker id="system_info_time_warp_mode"/><c>time_warp_mode</c></tag>
@@ -6938,19 +7567,20 @@ ok
 	  <taglist>
 	    <tag><c>no_time_warp</c></tag>
 	    <item><p>The <seealso marker="time_correction#No_Time_Warp_Mode">no
-	    time warp mode</seealso> is being used.</p></item>
+	    time warp mode</seealso> is used.</p></item>
 
 	    <tag><c>single_time_warp</c></tag>
 	    <item><p>The <seealso marker="time_correction#Single_Time_Warp_Mode">single
-	    time warp mode</seealso> is being used.</p></item>
+	    time warp mode</seealso> is used.</p></item>
 
 	    <tag><c>multi_time_warp</c></tag>
 	    <item><p>The <seealso marker="time_correction#Multi_Time_Warp_Mode">multi
-	    time warp mode</seealso> is being used.</p></item>
+	    time warp mode</seealso> is used.</p></item>
 	  </taglist>
 	  </item>
-          <tag><marker id="system_info_tolerant_timeofday"><c>tolerant_timeofday</c></marker></tag>
+          <tag><c>tolerant_timeofday</c></tag>
           <item>
+          <marker id="system_info_tolerant_timeofday"></marker>
             <p>Returns whether a pre erts-7.0 backwards compatible compensation
 	    for sudden changes of system time is <c>enabled</c> or <c>disabled</c>.
 	    Such compensation is <c>enabled</c> when the
@@ -6961,90 +7591,92 @@ ok
           </item>
           <tag><c>trace_control_word</c></tag>
           <item>
-            <p>Returns the value of the node's trace control word.
-              For more information see documentation of the function
-              <c>get_tcw</c> in "Match Specifications in Erlang",
-              <seealso marker="erts:match_spec#get_tcw">ERTS User's Guide</seealso>.</p>
+            <p>Returns the value of the node trace control word. For
+              more information, see function <c>get_tcw</c> in Section
+              <seealso marker="erts:match_spec#get_tcw">Match Specifications in Erlang</seealso> in the User's Guide.</p>
           </item>
-          <tag><marker id="update_cpu_info"><c>update_cpu_info</c></marker></tag>
+          <tag><c>update_cpu_info</c></tag>
           <item>
-            <p>The runtime system rereads the CPU information available and
-	       updates its internally stored information about the
-	       <seealso marker="#system_info_cpu_topology_detected">detected CPU
-	       topology</seealso> and the amount of logical processors
+          <marker id="update_cpu_info"></marker>
+            <p>The runtime system rereads the CPU information available
+	       and updates its internally stored information about the
+	       <seealso marker="#system_info_cpu_topology_detected">detected
+               CPU topology</seealso> and the number of logical processors
 	       <seealso marker="#logical_processors">configured</seealso>,
 	       <seealso marker="#logical_processors_online">online</seealso>, and
-	       <seealso marker="#logical_processors_available">available</seealso>.
-	       If the CPU information has changed since the last time it was read,
-	       the atom <c>changed</c> is returned; otherwise, the atom
-	       <c>unchanged</c> is returned. If the CPU information has changed
+	       <seealso marker="#logical_processors_available">available</seealso>.</p>
+               <p>If the CPU information has changed since the last time
+               it was read, the atom <c>changed</c> is returned, otherwise
+               the atom <c>unchanged</c>. If the CPU information has changed,
 	       you probably want to
-	       <seealso marker="#system_flag_schedulers_online">adjust the amount
-	       of schedulers online</seealso>. You typically want to have as
-	       many schedulers online as
-	       <seealso marker="#logical_processors_available">logical processors
-	       available</seealso>.
-	    </p>
+	       <seealso marker="#system_flag_schedulers_online">adjust the
+               number of schedulers online</seealso>. You typically want
+               to have as many schedulers online as
+	       <seealso marker="#logical_processors_available">logical
+               processors available</seealso>.</p>
           </item>
-          <tag><marker id="system_info_version"><c>version</c></marker></tag>
+          <tag><c>version</c></tag>
           <item>
+          <marker id="system_info_version"></marker>
             <p>Returns a string containing the version number of the
               emulator.</p>
           </item>
           <tag><c>wordsize</c></tag>
           <item>
-            <p>Same as <c>{wordsize, internal}.</c></p>
+            <p>Same as <c>{wordsize, internal}</c>.</p>
           </item>
           <tag><c>{wordsize, internal}</c></tag>
           <item>
             <p>Returns the size of Erlang term words in bytes as an
-              integer, i.e. on a 32-bit architecture 4 is returned,
-              and on a pure 64-bit architecture 8 is returned. On a
+              integer, that is, 4 is returned on a 32-bit architecture,
+              and 8 is returned on a pure 64-bit architecture. On a
               halfword 64-bit emulator, 4 is returned, as the Erlang
-              terms are stored using a virtual wordsize of half the
-              system's wordsize.</p>
+              terms are stored using a virtual word size of half the
+              system word size.</p>
           </item>
           <tag><c>{wordsize, external}</c></tag>
           <item>
-	    <p>Returns the true wordsize of the emulator, i.e. the size
-	       of a pointer, in bytes as an integer. On a pure 32-bit
-	       architecture 4 is returned, on both a halfword and pure
+	    <p>Returns the true word size of the emulator, that is,
+               the size of a pointer. The value is given in bytes
+               as an integer. On a pure 32-bit architecture, 4 is
+               returned. On both a half word and on a pure
 	       64-bit architecture, 8 is returned.</p>
           </item>
         </taglist>
         <note>
-          <p>The <c>scheduler</c> argument has changed name to
-            <c>scheduler_id</c>. This in order to avoid mixup with
-            the <c>schedulers</c> argument. The <c>scheduler</c>
-            argument was introduced in ERTS version 5.5 and renamed
-            in ERTS version 5.5.1.</p>
+          <p>Argument <c>scheduler</c> has changed name to
+            <c>scheduler_id</c> to avoid mix up with argument
+            <c>schedulers</c>. Argument <c>scheduler</c> was
+            introduced in <c>ERTS</c> 5.5 and renamed in
+            <c>ERTS</c> 5.5.1.</p>
         </note>
       </desc>
     </func>
 
     <func>
       <name name="system_monitor" arity="0"/>
+      <fsummary>Current system performance monitoring settings.</fsummary>
       <type name="system_monitor_option"/>
-      <fsummary>Current system performance monitoring settings</fsummary>
       <desc>
         <p>Returns the current system monitoring settings set by
           <seealso marker="#system_monitor/2">erlang:system_monitor/2</seealso>
-          as <c>{<anno>MonitorPid</anno>, <anno>Options</anno>}</c>, or <c>undefined</c> if there
-          are no settings. The order of the options may be different
+          as <c>{<anno>MonitorPid</anno>, <anno>Options</anno>}</c>,
+          or <c>undefined</c> if there
+          are no settings. The order of the options can be different
           from the one that was set.</p>
       </desc>
     </func>
 
     <func>
       <name name="system_monitor" arity="1"/>
+      <fsummary>Sets or clears system performance monitoring options.</fsummary>
       <type name="system_monitor_option"/>
-      <fsummary>Set or clear system performance monitoring options</fsummary>
       <desc>
-        <p>When called with the argument <c>undefined</c>, all
+        <p>When called with argument <c>undefined</c>, all
           system performance monitoring settings are cleared.</p>
-        <p>Calling the function with <c>{<anno>MonitorPid</anno>, <anno>Options</anno>}</c> as
-          argument, is the same as calling
-          <seealso marker="#system_monitor/2">erlang:system_monitor(<anno>MonitorPid</anno>, <anno>Options</anno>)</seealso>.</p>
+        <p>Calling the function with <c>{<anno>MonitorPid</anno>,
+          <anno>Options</anno>}</c> as argument is the same as calling
+          <seealso marker="#system_monitor/2"><c>erlang:system_monitor(<anno>MonitorPid</anno>, <anno>Options</anno>)</c></seealso>.</p>
         <p>Returns the previous system monitor settings just like
           <seealso marker="#system_monitor/0">erlang:system_monitor/0</seealso>.</p>
       </desc>
@@ -7052,102 +7684,101 @@ ok
 
     <func>
       <name name="system_monitor" arity="2"/>
+      <fsummary>Sets system performance monitoring options.</fsummary>
       <type name="system_monitor_option"/>
-      <fsummary>Set system performance monitoring options</fsummary>
       <desc>
-        <p>Sets system performance monitoring options. <c><anno>MonitorPid</anno></c>
-          is a local pid that will receive system monitor messages, and
-          the second argument is a list of monitoring options:</p>
+        <p>Sets the system performance monitoring options.
+          <c><anno>MonitorPid</anno></c> is a local process identifier (pid)
+          receiving system monitor messages. The
+          second argument is a list of monitoring options:</p>
         <taglist>
           <tag><c>{long_gc, Time}</c></tag>
           <item>
             <p>If a garbage collection in the system takes at least
-              <c>Time</c> wallclock milliseconds, a message
+              <c>Time</c> wall clock milliseconds, a message
               <c>{monitor, GcPid, long_gc, Info}</c> is sent to
-              <c><anno>MonitorPid</anno></c>. <c>GcPid</c> is the pid that was
-              garbage collected and <c>Info</c> is a list of two-element
-              tuples describing the result of the garbage collection.
-              One of the tuples is <c>{timeout, GcTime}</c> where
-              <c>GcTime</c> is the actual time for the garbage
+              <c><anno>MonitorPid</anno></c>. <c>GcPid</c> is the pid that
+              was garbage collected. <c>Info</c> is a list of two-element
+              tuples describing the result of the garbage collection.</p>
+            <p>One of the tuples is <c>{timeout, GcTime}</c>, where
+              <c>GcTime</c> is the time for the garbage
               collection in milliseconds. The other tuples are
-              tagged with <c>heap_size</c>, <c>heap_block_size</c>,
-	      <c>stack_size</c>, <c>mbuf_size</c>, <c>old_heap_size</c>,
-	      and <c>old_heap_block_size</c>. These tuples are
-	      explained in the documentation of the
-	      <seealso marker="#gc_start">gc_start</seealso>
-	      trace message (see
-	      <seealso marker="#trace/3">erlang:trace/3</seealso>).
-	      New tuples may be added, and the order of the tuples in
-	      the <c>Info</c> list may be changed at any time without prior
-	      notice.
-	      </p>
+              tagged with <c>heap_size</c>, <c>heap_block_size</c>
+              <c>stack_size</c>, <c>mbuf_size</c>, <c>old_heap_size</c>,
+              and <c>old_heap_block_size</c>. These tuples are
+              explained in the description of trace message
+              <seealso marker="#gc_start">gc_start</seealso> (see
+              <seealso marker="#trace/3">erlang:trace/3</seealso>).
+              New tuples can be added, and the order of the tuples in
+              the <c>Info</c> list can be changed at any time without
+              prior notice.</p>
           </item>
           <tag><c>{long_schedule, Time}</c></tag>
           <item>
-              <p>If a process or port in the system runs uninterrupted
+            <p>If a process or port in the system runs uninterrupted
               for at least <c>Time</c> wall clock milliseconds, a
               message <c>{monitor, PidOrPort, long_schedule, Info}</c>
               is sent to <c>MonitorPid</c>. <c>PidOrPort</c> is the
-              process or port that was running and <c>Info</c> is a
-              list of two-element tuples describing the event. In case
-              of a <c>pid()</c>, the tuples <c>{timeout, Millis}</c>,
-              <c>{in, Location}</c> and <c>{out, Location}</c> will be
+              process or port that was running. <c>Info</c> is a
+              list of two-element tuples describing the event.</p>
+            <p>If a <c>pid()</c>, the tuples <c>{timeout, Millis}</c>,
+              <c>{in, Location}</c>, and <c>{out, Location}</c> are
               present, where <c>Location</c> is either an MFA
               (<c>{Module, Function, Arity}</c>) describing the
               function where the process was scheduled in/out, or the
-              atom <c>undefined</c>. In case of a <c>port()</c>, the
+              atom <c>undefined</c>.</p>
+            <p>If a <c>port()</c>, the
               tuples <c>{timeout, Millis}</c> and <c>{port_op,Op}</c>
-              will be present. <c>Op</c> will be one of <c>proc_sig</c>,
+              are present. <c>Op</c> is one of <c>proc_sig</c>,
               <c>timeout</c>, <c>input</c>, <c>output</c>,
-              <c>event</c> or <c>dist_cmd</c>, depending on which
-              driver callback was executing. <c>proc_sig</c> is an
-              internal operation and should never appear, while the
+              <c>event</c>, or <c>dist_cmd</c>, depending on which
+              driver callback was executing.</p>
+            <p><c>proc_sig</c> is an
+              internal operation and is never to appear, while the
               others represent the corresponding driver callbacks
               <c>timeout</c>, <c>ready_input</c>, <c>ready_output</c>,
-              <c>event</c> and finally <c>outputv</c> (when the port
-              is used by distribution).  The <c>Millis</c> value in
-              the <c>timeout</c> tuple will tell you the actual
-              uninterrupted execution time of the process or port,
-              which will always be <c>&gt;=</c> the <c>Time</c> value
-              supplied when starting the trace. New tuples may be
-              added to the <c>Info</c> list in the future, and the
-              order of the tuples in the list may be changed at any
-              time without prior notice.
-	      </p>
-	      <p>This can be used to detect problems with NIF's or
-	      drivers that take too long to execute. Generally, 1 ms
-	      is considered a good maximum time for a driver callback
-	      or a NIF. However, a time sharing system should usually
-	      consider everything below 100 ms as "possible" and
-	      fairly "normal". Schedule times above that might however
-	      indicate swapping or a NIF/driver that is
-	      misbehaving. Misbehaving NIF's and drivers could cause
-	      bad resource utilization and bad overall performance of
-	      the system.</p>
+              <c>event</c>, and <c>outputv</c> (when the port
+              is used by distribution). Value <c>Millis</c> in
+              the <c>timeout</c> tuple informs about the
+              uninterrupted execution time of the process or port, which
+              always is equal to or higher than the <c>Time</c> value
+              supplied when starting the trace. New tuples can be
+              added to the <c>Info</c> list in a future release. The
+              order of the tuples in the list can be changed at any
+              time without prior notice.</p>
+            <p>This can be used to detect problems with NIFs or
+              drivers that take too long to execute. 1 ms is
+              considered a good maximum time for a driver callback
+              or a NIF. However, a time-sharing system is usually to
+              consider everything below 100 ms as "possible" and
+              fairly "normal". However, longer schedule times can
+              indicate swapping or a misbehaving NIF/driver.
+              Misbehaving NIFs and drivers can cause bad resource
+              utilization and bad overall system performance.</p>
           </item>
           <tag><c>{large_heap, Size}</c></tag>
           <item>
             <p>If a garbage collection in the system results in
               the allocated size of a heap being at least <c>Size</c>
               words, a message <c>{monitor, GcPid, large_heap, Info}</c>
-              is sent to <c><anno>MonitorPid</anno></c>. <c>GcPid</c> and <c>Info</c>
-              are the same as for <c>long_gc</c> above, except that
-              the tuple tagged with <c>timeout</c> is not present.
-	      <em>Note</em>: As of erts version 5.6 the monitor message
-	      is sent if the sum of the sizes of all memory blocks allocated
-	      for all heap generations is equal to or larger than <c>Size</c>.
-	      Previously the monitor message was sent if the memory block
-	      allocated for the youngest generation was equal to or larger
-	      than <c>Size</c>.
-	    </p>
+              is sent to <c><anno>MonitorPid</anno></c>.
+              <c>GcPid</c> and <c>Info</c>
+              are the same as for <c>long_gc</c> earlier, except that
+              the tuple tagged with <c>timeout</c> is not present.</p>
+            <p>As of <c>ERTS</c> 5.6, the monitor message is sent
+              if the sum of the sizes of all memory blocks allocated
+              for all heap generations is equal to or higher than <c>Size</c>.
+              Previously the monitor message was sent if the memory block
+              allocated for the youngest generation was equal to or higher
+              than <c>Size</c>.</p>
           </item>
           <tag><c>busy_port</c></tag>
           <item>
             <p>If a process in the system gets suspended because it
               sends to a busy port, a message
               <c>{monitor, SusPid, busy_port, Port}</c> is sent to
-              <c><anno>MonitorPid</anno></c>. <c>SusPid</c> is the pid that got
-              suspended when sending to <c>Port</c>.</p>
+              <c><anno>MonitorPid</anno></c>. <c>SusPid</c> is the pid
+              that got suspended when sending to <c>Port</c>.</p>
           </item>
           <tag><c>busy_dist_port</c></tag>
           <item>
@@ -7155,8 +7786,8 @@ ok
               sends to a process on a remote node whose inter-node
               communication was handled by a busy port, a message
               <c>{monitor, SusPid, busy_dist_port, Port}</c> is sent to
-              <c><anno>MonitorPid</anno></c>. <c>SusPid</c> is the pid that got
-              suspended when sending through the inter-node
+              <c><anno>MonitorPid</anno></c>. <c>SusPid</c> is the pid
+              that got suspended when sending through the inter-node
               communication port <c>Port</c>.</p>
           </item>
         </taglist>
@@ -7165,74 +7796,76 @@ ok
         <note>
           <p>If a monitoring process gets so large that it itself
             starts to cause system monitor messages when garbage
-            collecting, the messages will enlarge the process's
+            collecting, the messages enlarge the process
             message queue and probably make the problem worse.</p>
           <p>Keep the monitoring process neat and do not set the system
             monitor limits too tight.</p>
         </note>
-        <p>Failure: <c>badarg</c> if <c><anno>MonitorPid</anno></c> does not exist or is not a local process.</p>
+        <p>Failures:</p>
+        <taglist>
+          <tag><c>badarg</c></tag>
+          <item>If <c><anno>MonitorPid</anno></c> does not exist.</item>
+          <tag><c>badarg</c></tag>
+          <item>If <c><anno>MonitorPid</anno></c> is not a local process.</item>
+        </taglist>
       </desc>
     </func>
 
     <func>
       <name name="system_profile" arity="0"/>
+      <fsummary>Current system profiling settings.</fsummary>
       <type name="system_profile_option"/>
-      <fsummary>Current system profiling settings</fsummary>
       <desc>
         <p>Returns the current system profiling settings set by
           <seealso marker="#system_profile/2">erlang:system_profile/2</seealso>
-          as <c>{<anno>ProfilerPid</anno>, <anno>Options</anno>}</c>, or <c>undefined</c> if there
-          are no settings. The order of the options may be different
+          as <c>{<anno>ProfilerPid</anno>, <anno>Options</anno>}</c>,
+          or <c>undefined</c> if there
+          are no settings. The order of the options can be different
           from the one that was set.</p>
       </desc>
     </func>
 
     <func>
       <name name="system_profile" arity="2"/>
+      <fsummary>Current system profiling settings.</fsummary>
       <type name="system_profile_option"/>
-      <fsummary>Current system profiling settings</fsummary>
       <desc>
         <p>Sets system profiler options. <c><anno>ProfilerPid</anno></c>
-          is a local pid or port that will receive profiling messages. The
-	  receiver is excluded from all profiling.
+          is a local process identifier (pid) or port receiving profiling
+          messages. The receiver is excluded from all profiling.
           The second argument is a list of profiling options:</p>
         <taglist>
           <tag><c>exclusive</c></tag>
           <item>
-          <p>
-             If a synchronous call to a port from a process is done, the
+          <p>If a synchronous call to a port from a process is done, the
 	     calling process is considered not runnable during the call
              runtime to the port. The calling process is notified as
-             <c>inactive</c> and subsequently <c>active</c> when the port
-	     callback returns.
-          </p>
+             <c>inactive</c>, and later <c>active</c> when the port
+	     callback returns.</p>
           </item>
           <tag><c>runnable_procs</c></tag>
           <item>
-          <p>If a process is put into or removed from the run queue a message, 
-             <c>{profile, Pid, State, Mfa, Ts}</c>, is sent to
-	     <c><anno>ProfilerPid</anno></c>. Running processes that is reinserted into the
-	     run queue after having been preemptively scheduled out will not trigger this
-	     message.
-          </p>
+          <p>If a process is put into or removed from the run queue, a
+             message, <c>{profile, Pid, State, Mfa, Ts}</c>, is sent to
+	     <c><anno>ProfilerPid</anno></c>. Running processes that
+             are reinserted into the run queue after having been
+	     preempted do not trigger this message.</p>
           </item>
           <tag><c>runnable_ports</c></tag>
           <item>
-          <p>If a port is put into or removed from the run queue a message, 
-             <c>{profile, Port, State, 0, Ts}</c>, is sent to
-	     <c><anno>ProfilerPid</anno></c>.
-          </p>
+          <p>If a port is put into or removed from the run queue, a 
+             message, <c>{profile, Port, State, 0, Ts}</c>, is sent to
+	     <c><anno>ProfilerPid</anno></c>.</p>
           </item>
           <tag><c>scheduler</c></tag>
           <item>
-          <p>If a scheduler is put to sleep or awoken a message,
-             <c>{profile, scheduler, Id, State, NoScheds, Ts}</c>, is sent
-             to <c><anno>ProfilerPid</anno></c>.
-          </p>
+          <p>If a scheduler is put to sleep or awoken, a message,
+             <c>{profile, scheduler, Id, State, NoScheds, Ts}</c>, is
+             sent to <c><anno>ProfilerPid</anno></c>.</p>
           </item>
         </taglist>
-        <note><p><c>erlang:system_profile</c> is considered experimental and
-        its behaviour may change in the future.</p>
+        <note><p><c>erlang:system_profile</c> is considered experimental
+          and its behavior can change in a future release.</p>
         </note>
       </desc>
     </func>
@@ -7276,11 +7909,12 @@ ok
     </func>
     <func>
       <name name="term_to_binary" arity="1"/>
-      <fsummary>Encode a term to an Erlang external term format binary</fsummary>
+      <fsummary>Encodes a term to an Erlang external term format binary.</fsummary>
       <desc>
-        <p>Returns a binary data object which is the result of encoding
-          <c><anno>Term</anno></c> according to the Erlang external term format.</p>
-        <p>This can be used for a variety of purposes, for example
+        <p>Returns a binary data object that is the result of encoding
+          <c><anno>Term</anno></c> according to the Erlang external
+          term format.</p>
+        <p>This can be used for various purposes, for example,
           writing a term to a file in an efficient way, or sending an
           Erlang term to some type of communications channel not
           supported by distributed Erlang.</p>
@@ -7288,67 +7922,81 @@ ok
           <seealso marker="#binary_to_term/1">binary_to_term/1</seealso>.</p>
       </desc>
     </func>
+
     <func>
       <name name="term_to_binary" arity="2"/>
-      <fsummary>Encode a term to en Erlang external term format binary</fsummary>
-      <desc>
-        <p>Returns a binary data object which is the result of encoding
-          <c><anno>Term</anno></c> according to the Erlang external term format.</p>
-        <p>If the option <c>compressed</c> is provided, the external
-          term format will be compressed. The compressed format is
-          automatically recognized by <c>binary_to_term/1</c> in R7B and later.</p>
-        <p>It is also possible to specify a compression level by giving
-          the option <c>{compressed, <anno>Level</anno>}</c>, where <c><anno>Level</anno></c> is an
-          integer from 0 through 9. <c>0</c> means that no compression
-          will be done (it is the same as not giving any <c>compressed</c> option);
-          <c>1</c> will take the least time but may not compress as well as
-          the higher levels; <c>9</c> will take the most time and may produce
-          a smaller result. Note the "mays" in the preceding sentence; depending
-          on the input term, level 9 compression may or may not produce a smaller
-          result than level 1 compression.</p>
-        <p>Currently, <c>compressed</c> gives the same result as
-          <c>{compressed, 6}</c>.</p>
-        <p>The option <c>{minor_version, <anno>Version</anno>}</c> can be use to control
-          some details of the encoding. This option was
-          introduced in R11B-4. Currently, the allowed values for <c><anno>Version</anno></c>
-          are <c>0</c> and <c>1</c>.</p>
-        <p><c>{minor_version, 1}</c> is since 17.0 the default, it forces any floats in
-	  the term to be encoded
-          in a more space-efficient and exact way (namely in the 64-bit IEEE format,
-          rather than converted to a textual representation). <c>binary_to_term/1</c>
-          in R11B-4 and later is able decode this representation.</p>
-        <p><c>{minor_version, 0}</c> meaning that floats
-          will be encoded using a textual representation; this option is useful if
-          you want to ensure that releases prior to R11B-4 can decode resulting
+      <fsummary>Encodes a term to en Erlang external term format binary.</fsummary>
+      <desc>
+        <p>Returns a binary data object that is the result of encoding
+          <c><anno>Term</anno></c> according to the Erlang external
+          term format.</p>
+        <p>If option <c>compressed</c> is provided, the external term
+          format is compressed. The compressed format is automatically
+          recognized by <c>binary_to_term/1</c> as from Erlang R7B.</p>
+        <p>A compression level can be specified by giving option
+          <c>{compressed, <anno>Level</anno>}</c>.
+          <c><anno>Level</anno></c> is an integer
+          with range 0..9, where:</p>
+        <list type="bulleted">
+          <item><c>0</c> - No compression is done (it is the same as
+            giving no <c>compressed</c> option).</item>
+          <item><c>1</c> - Takes least time but may not compress
+            as well as the higher levels.</item>
+          <item><c>6</c> - Default level when option <c>compressed</c>
+	    is provided.</item>
+          <item><c>9</c> - Takes most time and tries to produce a smaller
+            result. Notice "tries" in the preceding sentence; depending
+            on the input term, level 9 compression either does or does
+            not produce a smaller result than level 1 compression.</item>
+        </list>
+        <p>Option <c>{minor_version, <anno>Version</anno>}</c>
+          can be used to control
+          some encoding details. This option was introduced in OTP R11B-4.
+          The valid values for <c><anno>Version</anno></c> are
+          <c>0</c> and <c>1</c>.</p>
+        <p>As from OTP 17.0, <c>{minor_version, 1}</c> is the default. It
+          forces any floats in the term to be encoded in a more
+          space-efficient and exact way (namely in the 64-bit IEEE format,
+          rather than converted to a textual representation).</p>
+        <p>As from OTP R11B-4, <c>binary_to_term/1</c> can decode this
+          representation.</p>
+        <p><c>{minor_version, 0}</c> means that floats are encoded
+          using a textual representation. This option is useful to
+          ensure that releases before OTP R11B-4 can decode resulting
           binary.</p>
         <p>See also
           <seealso marker="#binary_to_term/1">binary_to_term/1</seealso>.</p>
       </desc>
     </func>
+
     <func>
       <name name="throw" arity="1"/>
-      <fsummary>Throw an exception</fsummary>
+      <fsummary>Throws an exception.</fsummary>
       <desc>
         <p>A non-local return from a function. If evaluated within a
-          <c>catch</c>, <c>catch</c> will return the value <c><anno>Any</anno></c>.</p>
+          <c>catch</c>, <c>catch</c> returns value <c><anno>Any</anno></c>.</p>
+        <p>Example:</p>
         <pre>
 > <input>catch throw({hello, there}).</input>
 {hello,there}</pre>
         <p>Failure: <c>nocatch</c> if not evaluated within a catch.</p>
       </desc>
     </func>
+
     <func>
       <name name="time" arity="0"/>
-      <fsummary>Current time</fsummary>
+      <fsummary>Current time.</fsummary>
       <desc>
         <p>Returns the current time as <c>{Hour, Minute, Second}</c>.</p>
-        <p>The time zone and daylight saving time correction depend on
+        <p>The time zone and Daylight Saving Time correction depend on
           the underlying OS.</p>
+        <p>Example:</p>
         <pre>
 > <input>time().</input>
 {9,42,44}</pre>
       </desc>
     </func>
+
     <func>
       <name name="time_offset" arity="0"/>
       <fsummary>Current time offset</fsummary>
@@ -7399,14 +8047,14 @@ ok
     </func>
     <func>
       <name name="timestamp" arity="0"/>
-      <type name="timestamp"/>
       <fsummary>Current Erlang System time</fsummary>
+      <type name="timestamp"/>
       <desc>
 	<p>Returns current
 	<seealso marker="time_correction#Erlang_System_Time">Erlang system time</seealso>
 	on the format <c>{MegaSecs, Secs, MicroSecs}</c>. This format is 
-	the same that <seealso marker="kernel:os#timestamp/0"><c>os:timestamp/0</c></seealso>
-	and the now deprecated <seealso marker="#now/0"><c>erlang:now/0</c></seealso>
+	the same as <seealso marker="kernel:os#timestamp/0"><c>os:timestamp/0</c></seealso>
+	and the deprecated <seealso marker="#now/0"><c>erlang:now/0</c></seealso>
 	uses. The reason for the existence of <c>erlang:timestamp()</c> is
 	purely to simplify usage for existing code that assumes this timestamp
 	format. Current Erlang system time can more efficiently be retrieved in
@@ -7420,9 +8068,9 @@ timestamp() ->
     Secs = ErlangSystemTime div 1000000 - MegaSecs*1000000,
     MicroSecs = ErlangSystemTime rem 1000000,
     {MegaSecs, Secs, MicroSecs}.</code>
-        <p>It however use a native implementation which does
-	not build garbage on the heap and with slightly better
-	performance.</p>
+    <p>It, however, uses a native implementation which does
+    not build garbage on the heap and with slightly better
+    performance.</p>
 
 	<note><p>This time is <em>not</em> a monotonically increasing time
 	in the general case. For more information, see the documentation of
@@ -7433,186 +8081,190 @@ timestamp() ->
     </func>
     <func>
       <name name="tl" arity="1"/>
-      <fsummary>Tail of a list</fsummary>
+      <fsummary>Tail of a list.</fsummary>
       <desc>
-        <p>Returns the tail of <c><anno>List</anno></c>, that is, the list minus
-          the first element.</p>
+        <p>Returns the tail of <c><anno>List</anno></c>, that is,
+          the list minus the first element, for example:</p>
         <pre>
 > <input>tl([geesties, guilies, beasties]).</input>
 [guilies, beasties]</pre>
         <p>Allowed in guard tests.</p>
-        <p>Failure: <c>badarg</c> if <c><anno>List</anno></c> is the empty list [].</p>
+        <p>Failure: <c>badarg</c> if <c><anno>List</anno></c>
+          is the empty list <c>[]</c>.</p>
       </desc>
     </func>
+
     <func>
       <name name="trace" arity="3"/>
+      <fsummary>Sets trace flags for a process or processes.</fsummary>
       <type name="trace_flag"/>
-      <fsummary>Set trace flags for a process or processes</fsummary>
       <desc>
         <p>Turns on (if <c><anno>How</anno> == true</c>) or off (if
-          <c><anno>How</anno> == false</c>) the trace flags in <c><anno>FlagList</anno></c> for
-          the process or processes represented by <c><anno>PidSpec</anno></c>.</p>
-        <p><c><anno>PidSpec</anno></c> is either a pid for a local process, or one of
-          the following atoms:</p>
+          <c><anno>How</anno> == false</c>) the trace flags in
+          <c><anno>FlagList</anno></c> for
+          the process or processes represented by
+          <c><anno>PidSpec</anno></c>.</p>
+        <p><c><anno>PidSpec</anno></c> is either a process identifier
+          (pid) for a local process, or one of the following atoms:</p>
         <taglist>
           <tag><c>existing</c></tag>
           <item>
-            <p>All processes currently existing.</p>
+            <p>All currently existing processes.</p>
           </item>
           <tag><c>new</c></tag>
           <item>
-            <p>All processes that will be created in the future.</p>
+            <p>All processes that are created in the future.</p>
           </item>
           <tag><c>all</c></tag>
           <item>
             <p>All currently existing processes and all processes that
-              will be created in the future.</p>
+              are created in the future.</p>
           </item>
         </taglist>
-        <p><c><anno>FlagList</anno></c> can contain any number of the following
-          flags (the "message tags" refers to the list of messages
-          following below):</p>
+        <p><c><anno>FlagList</anno></c> can contain any number of the
+          following flags (the "message tags" refers to the list of
+	  <seealso marker="#trace_3_trace_messages">trace messages</seealso>):</p>
         <taglist>
           <tag><c>all</c></tag>
           <item>
-            <p>Set all trace flags except <c>{tracer, Tracer}</c> and
-              <c>cpu_timestamp</c> that are in their nature different
+            <p>Sets all trace flags except <c>{tracer, Tracer}</c> and
+              <c>cpu_timestamp</c>, which are in their nature different
               than the others.</p>
           </item>
           <tag><c>send</c></tag>
           <item>
-            <p>Trace sending of messages.</p>
-            <p>Message tags: <c>send</c>,
+            <p>Traces sending of messages.</p>
+            <p>Message tags: <c>send</c> and
               <c>send_to_non_existing_process</c>.</p>
           </item>
           <tag><c>'receive'</c></tag>
           <item>
-            <p>Trace receiving of messages.</p>
+            <p>Traces receiving of messages.</p>
             <p>Message tags: <c>'receive'</c>.</p>
           </item>
           <tag><c>procs</c></tag>
           <item>
-            <p>Trace process related events.</p>
+            <p>Traces process-related events.</p>
             <p>Message tags: <c>spawn</c>, <c>exit</c>,
               <c>register</c>, <c>unregister</c>, <c>link</c>,
-              <c>unlink</c>, <c>getting_linked</c>,
+              <c>unlink</c>, <c>getting_linked</c>, and
               <c>getting_unlinked</c>.</p>
           </item>
           <tag><c>call</c></tag>
           <item>
-            <p>Trace certain function calls. Specify which function
+            <p>Traces certain function calls. Specify which function
               calls to trace by calling
               <seealso marker="#trace_pattern/3">erlang:trace_pattern/3</seealso>.</p>
-            <p>Message tags: <c>call</c>, <c>return_from</c>.</p>
+            <p>Message tags: <c>call</c> and <c>return_from</c>.</p>
           </item>
           <tag><c>silent</c></tag>
           <item>
-            <p>Used in conjunction with the <c>call</c> trace flag.
-              The <c>call</c>, <c>return_from</c> and <c>return_to</c>
-              trace messages are inhibited if this flag is set,
-              but if there are match specs they are executed as normal.</p>
+            <p>Used with the <c>call</c> trace flag.
+              The <c>call</c>, <c>return_from</c>, and <c>return_to</c>
+              trace messages are inhibited if this flag is set, but they
+              are executed as normal if there are match specifications.</p>
             <p>Silent mode is inhibited by executing
               <c>erlang:trace(_, false, [silent|_])</c>,
-              or by a match spec executing the <c>{silent, false}</c>
-              function.</p>
+              or by a match specification executing the function
+              <c>{silent, false}</c>.</p>
             <p>The <c>silent</c> trace flag facilitates setting up
               a trace on many or even all processes in the system.
-              Then the interesting trace can be activated and
-              deactivated using the <c>{silent,Bool}</c>
-              match spec function, giving a high degree
-              of control of which functions with which 
-              arguments that triggers the trace.</p>
-            <p>Message tags: <c>call</c>, <c>return_from</c>, 
+              The trace can then be activated and deactivated using the match
+              specification function <c>{silent,Bool}</c>, giving
+              a high degree of control of which functions with which 
+              arguments that trigger the trace.</p>
+            <p>Message tags: <c>call</c>, <c>return_from</c>, and 
               <c>return_to</c>. Or rather, the absence of.</p>
           </item>
           <tag><c>return_to</c></tag>
           <item>
-            <p>Used in conjunction with the <c>call</c> trace flag.
-              Trace the actual return from a traced function back to
+            <p>Used with the <c>call</c> trace flag.
+              Traces the return from a traced function back to
               its caller. Only works for functions traced with
-              the <c>local</c> option to
+              option <c>local</c> to
               <seealso marker="#trace_pattern/3">erlang:trace_pattern/3</seealso>.</p>
             <p>The semantics is that a trace message is sent when a
-              call traced function actually returns, that is, when a
-              chain of tail recursive calls is ended. There will be
-              only one trace message sent per chain of tail recursive
-              calls, why the properties of tail recursiveness for
+              call traced function returns, that is, when a
+              chain of tail recursive calls ends. Only one trace
+              message is sent per chain of tail recursive calls,
+              so the properties of tail recursiveness for
               function calls are kept while tracing with this flag.
               Using <c>call</c> and <c>return_to</c> trace together
               makes it possible to know exactly in which function a
               process executes at any time.</p>
             <p>To get trace messages containing return values from
-              functions, use the <c>{return_trace}</c> match_spec
-              action instead.</p>
+              functions, use the <c>{return_trace}</c> match
+              specification action instead.</p>
             <p>Message tags: <c>return_to</c>.</p>
           </item>
           <tag><c>running</c></tag>
           <item>
-            <p>Trace scheduling of processes.</p>
-            <p>Message tags: <c>in</c>, and <c>out</c>.</p>
+            <p>Traces scheduling of processes.</p>
+            <p>Message tags: <c>in</c> and <c>out</c>.</p>
           </item>
           <tag><c>exiting</c></tag>
           <item>
-            <p>Trace scheduling of an exiting processes.</p>
+            <p>Traces scheduling of exiting processes.</p>
             <p>Message tags: <c>in_exiting</c>, <c>out_exiting</c>, and
 	    <c>out_exited</c>.</p>
           </item>
           <tag><c>garbage_collection</c></tag>
           <item>
-            <p>Trace garbage collections of processes.</p>
-            <p>Message tags: <c>gc_start</c>, <c>gc_end</c>.</p>
+            <p>Traces garbage collections of processes.</p>
+            <p>Message tags: <c>gc_start</c> and <c>gc_end</c>.</p>
           </item>
           <tag><c>timestamp</c></tag>
           <item>
-            <p>Include a time stamp in all trace messages. The time
-              stamp (Ts) is of the same form as returned by
+            <p>Includes a time-stamp in all trace messages. The
+              time-stamp (Ts) has the same form as returned by
               <c>erlang:now()</c>.</p>
           </item>
           <tag><c>cpu_timestamp</c></tag>
           <item>
             <p>A global trace flag for the Erlang node that makes all
-              trace timestamps be in CPU time, not wallclock. It is
-              only allowed with <c>PidSpec==all</c>. If the host
-              machine operating system does not support high resolution
+              trace time-stamps to be in CPU time, not wall clock time.
+              Only allowed with <c>PidSpec==all</c>. If the host
+              machine OS does not support high-resolution
               CPU time measurements, <c>trace/3</c> exits with
-              <c>badarg</c>. Note that most operating systems do
+              <c>badarg</c>. Notice that most OS do
               not synchronize this value across cores, so be prepared
               that time might seem to go backwards when using this option.</p>
           </item>
           <tag><c>arity</c></tag>
           <item>
-            <p>Used in conjunction with the <c>call</c> trace flag.
-              <c>{M, F, Arity}</c> will be specified instead of
+            <p>Used with the <c>call</c> trace flag.
+              <c>{M, F, Arity}</c> is specified instead of
               <c>{M, F, Args}</c> in call trace messages.</p>
           </item>
           <tag><c>set_on_spawn</c></tag>
           <item>
             <p>Makes any process created by a traced process inherit
-              its trace flags, including the <c>set_on_spawn</c> flag.</p>
+              its trace flags, including flag <c>set_on_spawn</c>.</p>
           </item>
           <tag><c>set_on_first_spawn</c></tag>
           <item>
             <p>Makes the first process created by a traced process
-              inherit its trace flags, excluding
-              the <c>set_on_first_spawn</c> flag.</p>
+              inherit its trace flags, excluding flag
+              <c>set_on_first_spawn</c>.</p>
           </item>
           <tag><c>set_on_link</c></tag>
           <item>
             <p>Makes any process linked by a traced process inherit its
-              trace flags, including the <c>set_on_link</c> flag.</p>
+              trace flags, including flag <c>set_on_link</c>.</p>
           </item>
           <tag><c>set_on_first_link</c></tag>
           <item>
             <p>Makes the first process linked to by a traced process
-              inherit its trace flags, excluding
-              the <c>set_on_first_link</c> flag.</p>
+              inherit its trace flags, excluding flag
+              <c>set_on_first_link</c>.</p>
           </item>
           <tag><c>{tracer, Tracer}</c></tag>
           <item>
-            <p>Specify where to send the trace messages. <c>Tracer</c>
-              must be the pid of a local process or the port identifier
+            <p>Specifies where to send the trace messages. <c>Tracer</c>
+              must be the process identifier of a local process
+              or the port identifier
               of a local port. If this flag is not given, trace
-              messages will be sent to the process that called
+              messages are sent to the process that called
               <c>erlang:trace/3</c>.</p>
           </item>
         </taglist>
@@ -7620,27 +8272,27 @@ timestamp() ->
           <c>set_on_link</c> is the same as having
           <c>set_on_first_link</c> alone. Likewise for
           <c>set_on_spawn</c> and <c>set_on_first_spawn</c>.</p>
-        <p>If the <c>timestamp</c> flag is not given, the tracing
-          process will receive the trace messages described below.
-          <c>Pid</c> is the pid of the traced process in which
-          the traced event has occurred. The third element of the tuple
-          is the message tag.</p>
-        <p>If the <c>timestamp</c> flag is given, the first element of
-          the tuple will be <c>trace_ts</c> instead and the timestamp
-          is added last in the tuple.</p>
+        <p>The tracing process receives the <em>trace messages</em> described
+	in the following list. <c>Pid</c> is the process identifier of the
+	traced process in which the traced event has occurred. The
+	third tuple element is the message tag.</p>
+        <p>If flag <c>timestamp</c> is given, the first tuple
+          element is <c>trace_ts</c> instead, and the time-stamp
+          is added last in the message tuple.</p>
+        <marker id="trace_3_trace_messages"></marker>
         <taglist>
           <tag><c>{trace, Pid, 'receive', Msg}</c></tag>
           <item>
-            <p>When <c>Pid</c> receives the message <c>Msg</c>.</p>
+            <p>When <c>Pid</c> receives message <c>Msg</c>.</p>
           </item>
           <tag><c>{trace, Pid, send, Msg, To}</c></tag>
           <item>
-            <p>When <c>Pid</c> sends the message <c>Msg</c> to
-              the process <c>To</c>.</p>
+            <p>When <c>Pid</c> sends message <c>Msg</c> to
+              process <c>To</c>.</p>
           </item>
           <tag><c>{trace, Pid, send_to_non_existing_process, Msg, To}</c></tag>
           <item>
-            <p>When <c>Pid</c> sends the message <c>Msg</c> to
+            <p>When <c>Pid</c> sends message <c>Msg</c> to
               the non-existing process <c>To</c>.</p>
           </item>
           <tag><c>{trace, Pid, call, {M, F, Args}}</c></tag>
@@ -7648,7 +8300,7 @@ timestamp() ->
             <p>When <c>Pid</c> calls a traced function. The return
               values of calls are never supplied, only the call and its
               arguments.</p>
-            <p>Note that the trace flag <c>arity</c> can be used to
+            <p>Trace flag <c>arity</c> can be used to
               change the contents of this message, so that <c>Arity</c>
               is specified instead of <c>Args</c>.</p>
           </item>
@@ -7656,35 +8308,34 @@ timestamp() ->
           <item>
             <p>When <c>Pid</c> returns <em>to</em> the specified
               function. This trace message is sent if both
-              the <c>call</c> and the <c>return_to</c> flags are set,
+              the flags <c>call</c> and <c>return_to</c> are set,
               and the function is set to be traced on <em>local</em>
               function calls. The message is only sent when returning
-              from a chain of tail recursive function calls where at
+              from a chain of tail recursive function calls, where at
               least one call generated a <c>call</c> trace message
-              (that is, the functions match specification matched and
+              (that is, the functions match specification matched, and
               <c>{message, false}</c> was not an action).</p>
           </item>
           <tag><c>{trace, Pid, return_from, {M, F, Arity}, ReturnValue}</c></tag>
           <item>
             <p>When <c>Pid</c> returns <em>from</em> the specified
-              function. This trace message is sent if the <c>call</c>
-              flag is set, and the function has a match specification
+              function. This trace message is sent if flag <c>call</c>
+              is set, and the function has a match specification
               with a <c>return_trace</c> or <c>exception_trace</c> action.</p>
           </item>
           <tag><c>{trace, Pid, exception_from, {M, F, Arity}, {Class, Value}}</c></tag>
           <item>
             <p>When <c>Pid</c> exits <em>from</em> the specified
-              function due to an exception. This trace message is sent
-              if the <c>call</c> flag is set, and the function has 
+              function because of an exception. This trace message is
+              sent if flag <c>call</c> is set, and the function has 
               a match specification with an <c>exception_trace</c> action.</p>
           </item>
           <tag><c>{trace, Pid, spawn, Pid2, {M, F, Args}}</c></tag>
           <item>
             <p>When <c>Pid</c> spawns a new process <c>Pid2</c> with
               the specified function call as entry point.</p>
-            <p>Note that <c>Args</c> is supposed to be the argument
-              list, but may be any term in the case of an erroneous
-              spawn.</p>
+            <p><c>Args</c> is supposed to be the argument list,
+              but can be any term if the spawn is erroneous.</p>
           </item>
           <tag><c>{trace, Pid, exit, Reason}</c></tag>
           <item>
@@ -7714,148 +8365,158 @@ timestamp() ->
           <tag><c>{trace, Pid, unregister, RegName}</c></tag>
           <item>
             <p>When <c>Pid</c> gets the name <c>RegName</c> unregistered.
-              Note that this is done automatically when a registered
+              This is done automatically when a registered
               process exits.</p>
           </item>
           <tag><c>{trace, Pid, in, {M, F, Arity} | 0}</c></tag>
           <item>
-            <p>When <c>Pid</c> is scheduled to run. The process will
-              run in function <c>{M, F, Arity}</c>. On some rare
-              occasions the current function cannot be determined, then
-              the last element <c>Arity</c> is 0.</p>
+            <p>When <c>Pid</c> is scheduled to run. The process
+              runs in function <c>{M, F, Arity}</c>. On some rare
+              occasions, the current function cannot be determined,
+              then the last element is <c>0</c>.</p>
           </item>
           <tag><c>{trace, Pid, out, {M, F, Arity} | 0}</c></tag>
           <item>
             <p>When <c>Pid</c> is scheduled out. The process was
-              running in function {M, F, Arity}. On some rare occasions
+              running in function {M, F, Arity}. On some rare occasions,
               the current function cannot be determined, then the last
-              element <c>Arity</c> is 0.</p>
+              element is <c>0</c>.</p>
           </item>
-          <tag><marker id="gc_start"><c>{trace, Pid, gc_start, Info}</c></marker></tag>
+          <tag><c>{trace, Pid, gc_start, Info}</c></tag>
           <item>
+          <marker id="gc_start"></marker>
             <p>Sent when garbage collection is about to be started.
               <c>Info</c> is a list of two-element tuples, where
               the first element is a key, and the second is the value.
-              You should not depend on the tuples have any defined
-              order. Currently, the following keys are defined:</p>
+              Do not depend on any order of the tuples.
+              The following keys are defined:</p>
             <taglist>
               <tag><c>heap_size</c></tag>
               <item>The size of the used part of the heap.</item>
 	      <tag><c>heap_block_size</c></tag>
 	      <item>The size of the memory block used for storing
-	            the heap and the stack.</item>
+	        the heap and the stack.</item>
               <tag><c>old_heap_size</c></tag>
               <item>The size of the used part of the old heap.</item>
 	      <tag><c>old_heap_block_size</c></tag>
 	      <item>The size of the memory block used for storing
-	            the old heap.</item>
+                the old heap.</item>
               <tag><c>stack_size</c></tag>
-              <item>The actual size of the stack.</item>
+              <item>The size of the stack.</item>
               <tag><c>recent_size</c></tag>
               <item>The size of the data that survived the previous garbage
-               collection.</item>
+                collection.</item>
               <tag><c>mbuf_size</c></tag>
               <item>The combined size of message buffers associated with
-               the process.</item>
-
+                the process.</item>
               <tag><c>bin_vheap_size</c></tag>
-              <item>The total size  of unique off-heap binaries referenced from the process heap.</item>
+              <item>The total size of unique off-heap binaries referenced
+                from the process heap.</item>
               <tag><c>bin_vheap_block_size</c></tag>
-	      <item>The total size of binaries, in words, allowed in the virtual
-	       heap in the process before doing a garbage collection. </item>
+	      <item>The total size of binaries allowed in the virtual
+	        heap in the process before doing a garbage collection.</item>
               <tag><c>bin_old_vheap_size</c></tag>
-              <item>The total size of unique off-heap binaries referenced from the process old heap.</item>
-              <tag><c>bin_vheap_block_size</c></tag>
-	      <item>The total size of binaries, in words, allowed in the virtual
-	       old heap in the process before doing a garbage collection. </item>
-
-
+              <item>The total size of unique off-heap binaries referenced
+                from the process old heap.</item>
+              <tag><c>bin_old_vheap_block_size</c></tag>
+	      <item>The total size of binaries allowed in the virtual
+	        old heap in the process before doing a garbage collection.</item>
             </taglist>
             <p>All sizes are in words.</p>
           </item>
           <tag><c>{trace, Pid, gc_end, Info}</c></tag>
           <item>
             <p>Sent when garbage collection is finished. <c>Info</c>
-              contains the same kind of list as in the <c>gc_start</c>
-              message, but the sizes reflect the new sizes after
+              contains the same kind of list as in message <c>gc_start</c>,
+              but the sizes reflect the new sizes after
               garbage collection.</p>
           </item>
         </taglist>
-        <p>If the tracing process dies, the flags will be silently
+        <p>If the tracing process dies, the flags are silently
           removed.</p>
-        <p>Only one process can trace a particular process. For this
-          reason, attempts to trace an already traced process will fail.</p>
+        <p>Only one process can trace a particular process. Therefore,
+          attempts to trace an already traced process fail.</p>
         <p>Returns: A number indicating the number of processes that
-          matched <c><anno>PidSpec</anno></c>. If <c><anno>PidSpec</anno></c> is a pid,
-          the return value will be <c>1</c>. If <c><anno>PidSpec</anno></c> is
-          <c>all</c> or <c>existing</c> the return value will be
+          matched <c><anno>PidSpec</anno></c>.
+          If <c><anno>PidSpec</anno></c> is a process
+          identifier, the return value is <c>1</c>.
+          If <c><anno>PidSpec</anno></c>
+          is <c>all</c> or <c>existing</c>, the return value is
           the number of processes running, excluding tracer processes.
-          If <c><anno>PidSpec</anno></c> is <c>new</c>, the return value will be
+          If <c><anno>PidSpec</anno></c> is <c>new</c>, the return value is
           <c>0</c>.</p>
-        <p>Failure: If specified arguments are not supported. For
-          example <c>cpu_timestamp</c> is not supported on all
-          platforms.</p>
+        <p>Failure: <c>badarg</c> if the specified arguments are
+          not supported. For example, <c>cpu_timestamp</c> is not
+          supported on all platforms.</p>
       </desc>
     </func>
+
     <func>
       <name name="trace_delivered" arity="1"/>
-      <fsummary>Notification when trace has been delivered</fsummary>
+      <fsummary>Notification when trace has been delivered.</fsummary>
       <desc>
         <p>The delivery of trace messages is dislocated on the time-line
-          compared to other events in the system. If you know that the
-          <c><anno>Tracee</anno></c> has passed some specific point in its execution,
+          compared to other events in the system. If you know that
+          <c><anno>Tracee</anno></c> has passed some specific point
+          in its execution,
           and you want to know when at least all trace messages
-          corresponding to events up to this point have reached the tracer
-          you can use <c>erlang:trace_delivered(<anno>Tracee</anno>)</c>. A
+          corresponding to events up to this point have reached the
+          tracer, use <c>erlang:trace_delivered(<anno>Tracee</anno>)</c>. A
           <c>{trace_delivered, <anno>Tracee</anno>, <anno>Ref</anno>}</c> message is sent to
           the caller of <c>erlang:trace_delivered(<anno>Tracee</anno>)</c> when it
-          is guaranteed that all trace messages have been delivered to
-          the tracer up to the point that the <c><anno>Tracee</anno></c> had reached
+          is guaranteed that all trace messages are delivered to
+          the tracer up to the point that <c><anno>Tracee</anno></c> reached
           at the time of the call to
           <c>erlang:trace_delivered(<anno>Tracee</anno>)</c>.</p>
-        <p>Note that the <c>trace_delivered</c> message does <em>not</em>
-          imply that trace messages have been delivered; instead, it implies
-          that all trace messages that <em>should</em> be delivered have
-          been delivered. It is not an error if <c><anno>Tracee</anno></c> isn't, and
-          hasn't been traced by someone, but if this is the case,
-          <em>no</em> trace messages will have been delivered when the
+        <p>Notice that message <c>trace_delivered</c> does <em>not</em>
+          imply that trace messages have been delivered.
+          Instead it implies that all trace messages that
+          <em>are to be delivered</em> have been delivered.
+          It is not an error if <c><anno>Tracee</anno></c> is not, and
+          has not been traced by someone, but if this is the case,
+          <em>no</em> trace messages have been delivered when the
           <c>trace_delivered</c> message arrives.</p>
-        <p>Note that <c><anno>Tracee</anno></c> has to refer to a process currently,
+        <p>Notice that that <c><anno>Tracee</anno></c> must refer
+          to a process currently,
           or previously existing on the same node as the caller of
           <c>erlang:trace_delivered(<anno>Tracee</anno>)</c> resides on.
-          The special <c><anno>Tracee</anno></c> atom <c>all</c> denotes all processes
+          The special <c><anno>Tracee</anno></c> atom <c>all</c>
+          denotes all processes
           that currently are traced in the node.</p>
-        <p>An example: Process <c>A</c> is <c><anno>Tracee</anno></c>, port <c>B</c> is
-          tracer, and process <c>C</c> is the port owner of <c>B</c>.
-          <c>C</c> wants to close <c>B</c> when <c>A</c> exits. <c>C</c>
-          can ensure that the trace isn't truncated by calling
-          <c>erlang:trace_delivered(A)</c> when <c>A</c> exits and wait
-          for the <c>{trace_delivered, A, <anno>Ref</anno>}</c> message before closing
-          <c>B</c>.</p>
-        <p>Failure: <c>badarg</c> if <c><anno>Tracee</anno></c> does not refer to a
+        <p>Example: Process <c>A</c> is <c><anno>Tracee</anno></c>,
+	  port <c>B</c> is tracer, and process <c>C</c> is the port
+	  owner of <c>B</c>. <c>C</c> wants to close <c>B</c> when
+	  <c>A</c> exits. To ensure that the trace is not truncated,
+	  <c>C</c> can call <c>erlang:trace_delivered(A)</c>, when
+	  <c>A</c> exits, and wait for message <c>{trace_delivered, A,
+	  <anno>Ref</anno>}</c> before closing <c>B</c>.</p>
+        <p>Failure: <c>badarg</c> if <c><anno>Tracee</anno></c>
+          does not refer to a
           process (dead or alive) on the same node as the caller of
           <c>erlang:trace_delivered(<anno>Tracee</anno>)</c> resides on.</p>
       </desc>
     </func>
+
     <func>
       <name name="trace_info" arity="2"/>
+     <fsummary>Trace information about a process or function.</fsummary>
       <type name="trace_info_return"/>
       <type name="trace_info_item_result"/>
       <type name="trace_info_flag"/>
       <type name="trace_match_spec"/>
-      <fsummary>Trace information about a process or function</fsummary>
       <desc>
         <p>Returns trace information about a process or function.</p>
-        <p>To get information about a process, <c><anno>PidOrFunc</anno></c> should
-          be a pid or the atom <c>new</c>. The atom <c>new</c> means
-          that the default trace state for processes to be created will
-          be returned. <c><anno>Item</anno></c> must have one of the following
-          values:</p>
+        <p>To get information about a process,
+         <c><anno>PidOrFunc</anno></c> is to
+          be a process identifier (pid) or the atom <c>new</c>.
+          The atom <c>new</c> means that the default trace state for
+          processes to be created is returned.</p>
+        <p>The following <c>Item</c>s are valid:</p>
         <taglist>
           <tag><c>flags</c></tag>
           <item>
-            <p>Return a list of atoms indicating what kind of traces is
-              enabled for the process. The list will be empty if no
+            <p>Returns a list of atoms indicating what kind of traces is
+              enabled for the process. The list is empty if no
               traces are enabled, and one or more of the followings
               atoms if traces are enabled: <c>send</c>,
               <c>'receive'</c>, <c>set_on_spawn</c>, <c>call</c>,
@@ -7866,337 +8527,354 @@ timestamp() ->
           </item>
           <tag><c>tracer</c></tag>
           <item>
-            <p>Return the identifier for process or port tracing this
+            <p>Returns the identifier for process or port tracing this
               process. If this process is not being traced, the return
-              value will be <c>[]</c>.</p>
+              value is <c>[]</c>.</p>
           </item>
         </taglist>
-        <p>To get information about a function, <c>PidOrFunc</c> should
-          be a three-element tuple: <c>{Module, Function, Arity}</c> or
-          the atom <c>on_load</c>. No wildcards are allowed. Returns
-          <c>undefined</c> if the function does not exist or
-          <c>false</c> if the function is not traced at all. <c>Item</c>
-          must have one of the following values:</p>
+        <p>To get information about a function, <c>PidOrFunc</c> is to
+          be the three-element tuple <c>{Module, Function, Arity}</c> or
+          the atom <c>on_load</c>. No wild cards are allowed. Returns
+          <c>undefined</c> if the function does not exist, or
+          <c>false</c> if the function is not traced.</p>
+         <p>The following <c>Item</c>s are valid::</p>
         <taglist>
           <tag><c>traced</c></tag>
           <item>
-            <p>Return <c>global</c> if this function is traced on
+            <p>Returns <c>global</c> if this function is traced on
               global function calls, <c>local</c> if this function is
-              traced on local function calls (i.e local and global
-              function calls), and <c>false</c> if neither local nor
-              global function calls are traced.</p>
+              traced on local function calls (that is, local and global
+              function calls), and <c>false</c> if local or
+              global function calls are not traced.</p>
           </item>
           <tag><c>match_spec</c></tag>
           <item>
-            <p>Return the match specification for this function, if it
+            <p>Returns the match specification for this function, if it
               has one. If the function is locally or globally traced but
               has no match specification defined, the returned value
               is <c>[]</c>.</p>
           </item>
           <tag><c>meta</c></tag>
           <item>
-            <p>Return the meta trace tracer process or port for this
-              function, if it has one. If the function is not meta
-              traced the returned value is <c>false</c>, and if
-              the function is meta traced but has once detected that
-              the tracer proc is invalid, the returned value is [].</p>
+            <p>Returns the meta-trace tracer process or port for this
+              function, if it has one. If the function is not
+              meta-traced, the returned value is <c>false</c>. If
+              the function is meta-traced but has once detected that
+              the tracer process is invalid, the returned value is [].</p>
           </item>
           <tag><c>meta_match_spec</c></tag>
           <item>
-            <p>Return the meta trace match specification for this
-              function, if it has one. If the function is meta traced
+            <p>Returns the meta-trace match specification for this
+              function, if it has one. If the function is meta-traced
               but has no match specification defined, the returned
               value is <c>[]</c>.</p>
           </item>
           <tag><c>call_count</c></tag>
           <item>
-            <p>Return the call count value for this function or
+            <p>Returns the call count value for this function or
               <c>true</c> for the pseudo function <c>on_load</c> if call
-              count tracing is active. Return <c>false</c> otherwise.
+              count tracing is active. Otherwise <c>false</c> is returned.
               See also
               <seealso marker="#trace_pattern/3">erlang:trace_pattern/3</seealso>.</p>
           </item>
           <tag><c>call_time</c></tag>
           <item>
-	      <p>Return the call time values for this function or
+	      <p>Returns the call time values for this function or
               <c>true</c> for the pseudo function <c>on_load</c> if call
-	      time tracing is active. Returns <c>false</c> otherwise.
+	      time tracing is active. Otherwise <c>false</c> is returned.
 	      The call time values returned, <c>[{Pid, Count, S, Us}]</c>,
-	      is a list of each process that has executed the function and its specific counters.
-              See also
+	      is a list of each process that executed the function
+              and its specific counters. See also
               <seealso marker="#trace_pattern/3">erlang:trace_pattern/3</seealso>.</p>
           </item>
 
           <tag><c>all</c></tag>
           <item>
-            <p>Return a list containing the <c>{<anno>Item</anno>, Value}</c> tuples
-              for all other items, or return <c>false</c> if no tracing
+            <p>Returns a list containing the
+              <c>{<anno>Item</anno>, Value}</c> tuples
+              for all other items, or returns <c>false</c> if no tracing
               is active for this function.</p>
           </item>
         </taglist>
-        <p>The actual return value will be <c>{<anno>Item</anno>, Value}</c>, where
-          <c>Value</c> is the requested information as described above.
+        <p>The return value is <c>{<anno>Item</anno>, Value}</c>, where
+          <c>Value</c> is the requested information as described earlier.
           If a pid for a dead process was given, or the name of a
-          non-existing function, <c>Value</c> will be <c>undefined</c>.</p>
-        <p>If <c><anno>PidOrFunc</anno></c> is the <c>on_load</c>, the information
+          non-existing function, <c>Value</c> is <c>undefined</c>.</p>
+        <p>If <c><anno>PidOrFunc</anno></c> is <c>on_load</c>, the information
           returned refers to the default value for code that will be
           loaded.</p>
       </desc>
     </func>
+
     <func>
       <name name="trace_pattern" arity="2" clause_i="1"/>
+      <fsummary>Sets trace patterns for global call tracing.</fsummary>
       <type name="trace_pattern_mfa"/>
       <type name="trace_match_spec"/>
-      <fsummary>Set trace patterns for global call tracing</fsummary>
       <desc>
         <p>The same as
           <seealso marker="#trace_pattern/3">erlang:trace_pattern(MFA, MatchSpec, [])</seealso>,
           retained for backward compatibility.</p>
       </desc>
     </func>
+
     <func>
       <name name="trace_pattern" arity="3"/>
+      <fsummary>Sets trace patterns for tracing of function calls.</fsummary>
       <type name="trace_pattern_mfa"/>
       <type name="trace_match_spec"/>
       <type name="trace_pattern_flag"/>
-      <fsummary>Set trace patterns for tracing of function calls</fsummary>
       <desc>
-        <p>This BIF is used to enable or disable call tracing for
-          exported functions. It must be combined with
+        <p>Enables or disables call tracing for
+          one or more functions. Must be combined with
           <seealso marker="#trace/3">erlang:trace/3</seealso>
           to set the <c>call</c> trace flag for one or more processes.</p>
-        <p>Conceptually, call tracing works like this: Inside
-          the Erlang virtual machine there is a set of processes to be
-          traced and a set of functions to be traced. Tracing will be
-          enabled on the intersection of the set. That is, if a process
-          included in the traced process set calls a function included
-          in the traced function set, the trace action will be taken.
-          Otherwise, nothing will happen.</p>
-        <p>Use
-          <seealso marker="#trace/3">erlang:trace/3</seealso> to
-          add or remove one or more processes to the set of traced
-          processes. Use <c>erlang:trace_pattern/2</c> to add or remove
-          exported functions to the set of traced functions.</p>
-        <p>The <c>erlang:trace_pattern/3</c> BIF can also add match
-          specifications to an exported function. A match specification
-          comprises a pattern that the arguments to the function must
-          match, a guard expression which must evaluate to <c>true</c>
+        <p>Conceptually, call tracing works as follows. Inside
+          the Erlang Virtual Machine, a set of processes and
+          a set of functions are to be traced. If a traced process
+          calls a traced function, the trace action is taken.
+          Otherwise, nothing happens.</p>
+        <p>To add or remove one or more processes to the set of traced
+          processes, use
+          <seealso marker="#trace/3">erlang:trace/3</seealso>.</p>
+        <p>To add or remove functions to the set of traced
+          functions, use <c>erlang:trace_pattern/3</c>.</p>
+        <p>The BIF <c>erlang:trace_pattern/3</c> can also add match
+          specifications to a function. A match specification
+          comprises a pattern that the function arguments must
+          match, a guard expression that must evaluate to <c>true</c>,
           and an action to be performed. The default action is to send a
           trace message. If the pattern does not match or the guard
-          fails, the action will not be executed.</p>
-        <p>The <c><anno>MFA</anno></c> argument should be a tuple like
-          <c>{Module, Function, Arity}</c> or the atom <c>on_load</c>
-          (described below). It can be the module, function, and arity
-          for an exported function (or a BIF in any module).
-          The <c>'_'</c> atom can be used to mean any of that kind.
-          Wildcards can be used in any of the following ways:</p>
+          fails, the action is not executed.</p>
+        <p>Argument <c><anno>MFA</anno></c> is to be a tuple, such as
+          <c>{Module, Function, Arity}</c>, or the atom <c>on_load</c>
+          (described in the following). It can be the module, function,
+          and arity for a function (or a BIF in any module).
+          The atom <c>'_'</c> can be used as a wild card in any of the
+	  following ways:</p>
         <taglist>
           <tag><c>{Module,Function,'_'}</c></tag>
           <item>
-            <p>All exported functions of any arity named <c>Function</c>
+            <p>All functions of any arity named <c>Function</c>
               in module <c>Module</c>.</p>
           </item>
           <tag><c>{Module,'_','_'}</c></tag>
           <item>
-            <p>All exported functions in module <c>Module</c>.</p>
+            <p>All functions in module <c>Module</c>.</p>
           </item>
           <tag><c>{'_','_','_'}</c></tag>
           <item>
-            <p>All exported functions in all loaded modules.</p>
+            <p>All functions in all loaded modules.</p>
           </item>
         </taglist>
         <p>Other combinations, such as <c>{Module,'_',Arity}</c>, are
-          not allowed. Local functions will match wildcards only if
-          the <c>local</c> option is in the <c><anno>FlagList</anno></c>.</p>
-        <p>If the <c><anno>MFA</anno></c> argument is the atom <c>on_load</c>,
-          the match specification and flag list will be used on all
+          not allowed. Local functions match wild cards only if
+          option <c>local</c> is in <c><anno>FlagList</anno></c>.</p>
+        <p>If argument <c><anno>MFA</anno></c> is the atom <c>on_load</c>,
+          the match specification and flag list are used on all
           modules that are newly loaded.</p>
-        <p>The <c><anno>MatchSpec</anno></c> argument can take any of the following
-          forms:</p>
+        <p>Argument <c><anno>MatchSpec</anno></c> can take the
+          following forms:</p>
         <taglist>
           <tag><c>false</c></tag>
           <item>
-            <p>Disable tracing for the matching function(s). Any match
-              specification will be removed.</p>
+            <p>Disables tracing for the matching functions.
+              Any match specification is removed.</p>
           </item>
           <tag><c>true</c></tag>
           <item>
-            <p>Enable tracing for the matching function(s).</p>
+            <p>Enables tracing for the matching functions.</p>
           </item>
           <tag><c><anno>MatchSpecList</anno></c></tag>
           <item>
             <p>A list of match specifications. An empty list is
-              equivalent to <c>true</c>. See the ERTS User's Guide
-              for a description of match specifications.</p>
+              equivalent to <c>true</c>. For a description of match
+              specifications, see the User's Guide.</p>
           </item>
           <tag><c>restart</c></tag>
           <item>
-	      <p>For the <c><anno>FlagList</anno></c> option <c>call_count</c> and <c>call_time</c>:
-              restart the existing counters. The behaviour is undefined
+            <p>For the <c><anno>FlagList</anno></c> options <c>call_count</c>
+              and <c>call_time</c>: restarts
+              the existing counters. The behavior is undefined
               for other <c><anno>FlagList</anno></c> options.</p>
           </item>
           <tag><c>pause</c></tag>
           <item>
-	      <p>For the <c><anno>FlagList</anno></c> option <c>call_count</c> and <c>call_time</c>: pause
-              the existing counters. The behaviour is undefined for
-              other <c>FlagList</c> options.</p>
+	      <p>For the <c><anno>FlagList</anno></c> options
+              <c>call_count</c> and <c>call_time</c>: pauses
+              the existing counters. The behavior is undefined for
+              other <c><anno>FlagList</anno></c> options.</p>
           </item>
         </taglist>
-        <p>The <c><anno>FlagList</anno></c> parameter is a list of options.
-          The following options are allowed:</p>
+        <p>Parameter <c><anno>FlagList</anno></c> is a list of options.
+          The following are the valid options:</p>
         <taglist>
           <tag><c>global</c></tag>
           <item>
-            <p>Turn on or off call tracing for global function calls
+            <p>Turns on or off call tracing for global function calls
               (that is, calls specifying the module explicitly). Only
-              exported functions will match and only global calls will
-              generate trace messages. This is the default.</p>
+              exported functions match and only global calls
+              generate trace messages. <em>This is the default</em>.</p>
           </item>
           <tag><c>local</c></tag>
           <item>
-            <p>Turn on or off call tracing for all types of function
-              calls. Trace messages will be sent whenever any of
+            <p>Turns on or off call tracing for all types of function
+              calls. Trace messages are sent whenever any of
               the specified functions are called, regardless of how they
-              are called. If the <c>return_to</c> flag is set for
-              the process, a <c>return_to</c> message will also be sent
+              are called. If flag <c>return_to</c> is set for
+              the process, a <c>return_to</c> message is also sent
               when this function returns to its caller.</p>
           </item>
           <tag><c>meta | {meta, <anno>Pid</anno>}</c></tag>
           <item>
-            <p>Turn on or off meta tracing for all types of function
-              calls. Trace messages will be sent to the tracer process
+            <p>Turns on or off meta-tracing for all types of function
+              calls. Trace messages are sent to the tracer process
               or port <c><anno>Pid</anno></c> whenever any of the specified
               functions are called, regardless of how they are called.
-              If no <c><anno>Pid</anno></c> is specified, <c>self()</c> is used as a
-              default tracer process.</p>
-            <p>Meta tracing traces all processes and does not care
+              If no <c><anno>Pid</anno></c> is specified,
+              <c>self()</c> is used as a default tracer process.</p>
+            <p>Meta-tracing traces all processes and does not care
               about the process trace flags set by <c>trace/3</c>,
               the trace flags are instead fixed to
               <c>[call, timestamp]</c>.</p>
-            <p>The match spec function <c>{return_trace}</c> works with
-              meta trace and send its trace message to the same tracer
-              process.</p>
+            <p>The match specification function <c>{return_trace}</c>
+              works with meta-trace and sends its trace message to the
+              same tracer process.</p>
           </item>
           <tag><c>call_count</c></tag>
           <item>
             <p>Starts (<c><anno>MatchSpec</anno> == true</c>) or stops
-              (<c><anno>MatchSpec</anno> == false</c>) call count tracing for all
-              types of function calls. For every function a counter is
+              (<c><anno>MatchSpec</anno> == false</c>)
+              call count tracing for all
+              types of function calls. For every function, a counter is
               incremented when the function is called, in any process.
               No process trace flags need to be activated.</p>
             <p>If call count tracing is started while already running,
-              the count is restarted from zero. Running counters can be
-              paused with <c><anno>MatchSpec</anno> == pause</c>. Paused and running
-              counters can be restarted from zero with
+              the count is restarted from zero. To pause running
+              counters, use <c><anno>MatchSpec</anno> == pause</c>.
+              Paused and running counters can be restarted from zero with
               <c><anno>MatchSpec</anno> == restart</c>.</p>
-            <p>The counter value can be read with
+            <p>To read the counter value, use
               <seealso marker="#trace_info/2">erlang:trace_info/2</seealso>.</p>
           </item>
           <tag><c>call_time</c></tag>
           <item>
             <p>Starts (<c><anno>MatchSpec</anno> == true</c>) or stops
-              (<c><anno>MatchSpec</anno> == false</c>) call time tracing for all
-              types of function calls. For every function a counter is
-	      incremented when the function is called. Time spent in the function
-	      is accumulated in two other counters, seconds and micro-seconds.
+              (<c><anno>MatchSpec</anno> == false</c>) call time
+              tracing for all
+              types of function calls. For every function, a counter is
+	      incremented when the function is called.
+              Time spent in the function is accumulated in
+              two other counters, seconds and microseconds.
 	      The counters are stored for each call traced process.</p>
             <p>If call time tracing is started while already running,
-              the count and time is restarted from zero. Running counters can be
-              paused with <c><anno>MatchSpec</anno> == pause</c>. Paused and running
-              counters can be restarted from zero with
+              the count and time is restarted from zero. To pause
+              running counters, use <c><anno>MatchSpec</anno> == pause</c>.
+              Paused and running counters can be restarted from zero with
               <c><anno>MatchSpec</anno> == restart</c>.</p>
-            <p>The counter value can be read with
+            <p>To read the counter value, use
               <seealso marker="#trace_info/2">erlang:trace_info/2</seealso>.</p>
           </item>
-
         </taglist>
-        <p>The <c>global</c> and <c>local</c> options are mutually
-          exclusive and <c>global</c> is the default (if no options are
-          specified). The <c>call_count</c> and <c>meta</c> options
-          perform a kind of local tracing, and can also not be combined
-          with <c>global</c>. A function can be either globally or
+        <p>The options <c>global</c> and <c>local</c> are mutually
+          exclusive, and <c>global</c> is the default (if no options are
+          specified). The options <c>call_count</c> and <c>meta</c>
+          perform a kind of local tracing, and cannot be combined
+          with <c>global</c>. A function can be globally or
           locally traced. If global tracing is specified for a
-          specified set of functions; local, meta, call time and call count
-          tracing for the matching set of local functions will be
-          disabled, and vice versa.</p>
+          set of functions, then local, meta, call time, and call count
+          tracing for the matching set of local functions is
+          disabled, and conversely.</p>
         <p>When disabling trace, the option must match the type of trace
-          that is set on the function, so that local tracing must be
-          disabled with the <c>local</c> option and global tracing with
-          the <c>global</c> option (or no option at all), and so forth.</p>
-        <p>There is no way to directly change part of a match
-          specification list. If a function has a match specification,
-          you can replace it with a completely new one. If you need to
-          change an existing match specification, use the
+          set on the function. That is, local tracing must be
+          disabled with option <c>local</c> and global tracing with
+          option <c>global</c> (or no option), and so forth.</p>
+        <p>Part of a match specification list cannot be changed directly.
+          If a function has a match specification, it can be replaced
+          with a new one. To change an existing match specification,
+          use the BIF
           <seealso marker="#trace_info/2">erlang:trace_info/2</seealso>
-          BIF to retrieve the existing match specification.</p>
-        <p>Returns the number of exported functions that matched
-          the <c><anno>MFA</anno></c> argument. This will be zero if none matched at
-          all.</p>
+          to retrieve the existing match specification.</p>
+        <p>Returns the number of functions matching
+          argument <c><anno>MFA</anno></c>. This is zero if none matched.</p>
       </desc>
     </func>
+
     <func>
       <name name="trunc" arity="1"/>
-      <fsummary>Return an integer by the truncating a number</fsummary>
+      <fsummary>Returns an integer by truncating a number</fsummary>
       <desc>
-        <p>Returns an integer by the truncating <c><anno>Number</anno></c>.</p>
+        <p>Returns an integer by truncating <c><anno>Number</anno></c>,
+          for example:</p>
         <pre>
 > <input>trunc(5.5).</input>
 5</pre>
         <p>Allowed in guard tests.</p>
       </desc>
     </func>
+
     <func>
       <name name="tuple_size" arity="1"/>
-      <fsummary>Return the size of a tuple</fsummary>
+      <fsummary>Returns the size of a tuple.</fsummary>
       <desc>
-        <p>Returns an integer which is the number of elements in <c><anno>Tuple</anno></c>.</p>
+        <p>Returns an integer that is the number of elements in
+          <c><anno>Tuple</anno></c>, for example:</p>
         <pre>
 > <input>tuple_size({morni, mulle, bwange}).</input>
 3</pre>
         <p>Allowed in guard tests.</p>
       </desc>
     </func>
+
     <func>
       <name name="tuple_to_list" arity="1"/>
-      <fsummary>Convert a tuple to a list</fsummary>
+      <fsummary>Converts a tuple to a list.</fsummary>
       <desc>
-        <p>Returns a list which corresponds to <c><anno>Tuple</anno></c>.
-          <c><anno>Tuple</anno></c> may contain any Erlang terms.</p>
+        <p>Returns a list corresponding to <c><anno>Tuple</anno></c>.
+          <c><anno>Tuple</anno></c> can contain any Erlang terms.</p>
+        <p>Example:</p>
         <pre>
 > <input>tuple_to_list({share, {'Ericsson_B', 163}}).</input>
 [share,{'Ericsson_B',163}]</pre>
       </desc>
     </func>
+
     <func>
       <name name="universaltime" arity="0"/>
-      <fsummary>Current date and time according to Universal Time Coordinated (UTC)</fsummary>
+      <fsummary>Current date and time according to Universal Time Coordinated (UTC).</fsummary>
       <desc>
         <p>Returns the current date and time according to Universal
-          Time Coordinated (UTC), also called GMT, in the form
+          Time Coordinated (UTC) in the form
           <c>{{Year, Month, Day}, {Hour, Minute, Second}}</c> if
-          supported by the underlying operating system. If not,
-          <c>erlang:universaltime()</c> is equivalent to
+          supported by the underlying OS.
+          Otherwise <c>erlang:universaltime()</c> is equivalent to
           <c>erlang:localtime()</c>.</p>
+        <p>Example:</p>
         <pre>
 > <input>erlang:universaltime().</input>
 {{1996,11,6},{14,18,43}}</pre>
       </desc>
     </func>
+
     <func>
       <name name="universaltime_to_localtime" arity="1"/>
-      <fsummary>Convert from Universal Time Coordinated (UTC) to local date and time</fsummary>
+      <fsummary>Converts from Universal Time Coordinated (UTC) to local date and time.</fsummary>
       <desc>
         <p>Converts Universal Time Coordinated (UTC) date and time to
-          local date and time, if this is supported by the underlying
-          OS. Otherwise, no conversion is done, and
+          local date and time in the form
+          <c>{{Year, Month, Day}, {Hour, Minute, Second}}</c> if
+          supported by the underlying OS.
+          Otherwise no conversion is done, and
           <c><anno>Universaltime</anno></c> is returned.</p>
+        <p>Example:</p>
         <pre>
 > <input>erlang:universaltime_to_localtime({{1996,11,6},{14,18,43}}).</input>
 {{1996,11,7},{15,18,43}}</pre>
-        <p>Failure: <c>badarg</c> if <c>Universaltime</c> does not denote
-          a valid date and time.</p>
+        <p>Failure: <c>badarg</c> if <c>Universaltime</c> denotes
+          an invalid date and time.</p>
       </desc>
     </func>
+
     <func>
       <name name="unique_integer" arity="0"/>
       <fsummary>Get a unique integer value</fsummary>
@@ -8221,23 +8899,23 @@ timestamp() ->
 	Each integer value can of course be constructed
 	by other means.</p>
 
-	<p>By default, i.e. when <c>[]</c> is passed as
+	<p>By default, when <c>[]</c> is passed as
 	<c><anno>ModifierList</anno></c>, both negative and
-	positive integers will be returned. This is order
-	to be able to utilize the range of integers that do
-	not need to be heap allocated as much as possible.
+	positive integers can be returned. This in order
+	to utilize the range of integers that do
+	not need heap memory allocation as much as possible.
 	By default the returned integers are also only
-	guaranteed to be unique, i.e., any integer returned
-	may be either smaller, or larger than previously
+	guaranteed to be unique, that is, any returned integer
+	can be smaller or larger than previously
 	returned integers.</p>
 
-	<p>Currently valid <c><anno>Modifier</anno></c>s:</p>
+	<p>Valid <c><anno>Modifier</anno></c>s:</p>
 	<taglist>
 
 	  <tag>positive</tag>
 	  <item><p>Return only positive integers.</p>
 	  <p>Note that by passing the <c>positive</c> modifier
-	  you will get heap allocated integers (big-nums)
+	  you will get heap allocated integers (bignums)
 	  quicker.</p>
 	  </item>
 
@@ -8249,7 +8927,7 @@ timestamp() ->
 	  returned will always be larger than previously
 	  returned integers on the current runtime system
 	  instance.</p>
-	  <p>These values can be used when ordering events
+	  <p>These values can be used to determine order between events
 	  on the runtime system instance. That is, if both
 	  <c>X = erlang:unique_integer([monotonic])</c> and
 	  <c>Y = erlang:unique_integer([monotonic])</c> are
@@ -8259,21 +8937,21 @@ timestamp() ->
 	  before <c>Y</c>.</p>
 	  <warning><p>Strictly monotonically increasing values
 	  are inherently quite expensive to generate and scales
-	  poorly. This since the values needs to be
-	  synchronized. That is, do not pass the <c>monotonic</c>
+	  poorly. This is because the values need to be
+	  synchronized between cpu cores. That is, do not pass the <c>monotonic</c>
 	  modifier unless you really need strictly monotonically
 	  increasing values.</p></warning>
 	  </item>
 
 	</taglist>
 
-	<p>All currently valid <c><anno>Modifier</anno></c>s
+	<p>All valid <c><anno>Modifier</anno></c>s
 	can be combined. Repeated (valid)
 	<c><anno>Modifier</anno></c>s in the <c>ModifierList</c>
 	are ignored.</p>
 
 	<note><p>Note that the set of integers returned by
-	<c>unique_integer/1</c> using diffrent sets of
+	<c>unique_integer/1</c> using different sets of
 	<c><anno>Modifier</anno></c>s <em>will overlap</em>.
 	For example, by calling	<c>unique_integer([monotonic])</c>,
 	and <c>unique_integer([positive, monotonic])</c>
@@ -8293,25 +8971,30 @@ timestamp() ->
     </func>
     <func>
       <name name="unlink" arity="1"/>
-      <fsummary>Remove a link, if there is one, to another process or port</fsummary>
+      <fsummary>Removes a link to another process or port.</fsummary>
       <desc>
         <p>Removes the link, if there is one, between the calling
-          process and the process or port referred to by <c><anno>Id</anno></c>.</p>
+          process and the process or port referred to by
+          <c><anno>Id</anno></c>.</p>
         <p>Returns <c>true</c> and does not fail, even if there is no
-          link to <c><anno>Id</anno></c>, or if <c><anno>Id</anno></c> does not exist.</p>
-        <p>Once <c>unlink(<anno>Id</anno>)</c> has returned it is guaranteed that
+          link to <c><anno>Id</anno></c>, or if <c><anno>Id</anno></c>
+          does not exist.</p>
+        <p>Once <c>unlink(<anno>Id</anno>)</c> has returned,
+          it is guaranteed that
           the link between the caller and the entity referred to by
-          <c><anno>Id</anno></c> has no effect on the caller in the future (unless
-          the link is setup again). If caller is trapping exits, an
-          <c>{'EXIT', <anno>Id</anno>, _}</c> message due to the link might have
-          been placed in the caller's message queue prior to the call,
-          though. Note, the <c>{'EXIT', <anno>Id</anno>, _}</c> message can be the
-          result of the link, but can also be the result of <c><anno>Id</anno></c>
-          calling <c>exit/2</c>. Therefore, it <em>may</em> be
-          appropriate to cleanup the message queue when trapping exits
-          after the call to <c>unlink(<anno>Id</anno>)</c>, as follow:</p>
+          <c><anno>Id</anno></c> has no effect on the caller
+          in the future (unless
+          the link is setup again). If the caller is trapping exits, an
+          <c>{'EXIT', <anno>Id</anno>, _}</c> message from the link
+	  can have been placed in the caller's message queue before
+	  the call.</p>
+        <p>Notice that the <c>{'EXIT', <anno>Id</anno>, _}</c>
+          message can be the
+          result of the link, but can also be the result of <c>Id</c>
+          calling <c>exit/2</c>. Therefore, it <em>can</em> be
+          appropriate to clean up the message queue when trapping exits
+          after the call to <c>unlink(<anno>Id</anno>)</c>, as follows:</p>
         <code type="none">
-
     unlink(Id),
     receive
         {'EXIT', Id, _} ->
@@ -8320,23 +9003,25 @@ timestamp() ->
             true
     end</code>
         <note>
-          <p>Prior to OTP release R11B (erts version 5.5) <c>unlink/1</c>
-            behaved completely asynchronous, i.e., the link was active
+          <p>Prior to OTP release R11B (ERTS version 5.5) <c>unlink/1</c>
+            behaved completely asynchronously, i.e., the link was active
             until the "unlink signal" reached the linked entity. This
-            had one undesirable effect, though. You could never know when
+            had an undesirable effect, as you could never know when
             you were guaranteed <em>not</em> to be effected by the link.</p>
-          <p>Current behavior can be viewed as two combined operations:
+          <p>The current behavior can be viewed as two combined operations:
             asynchronously send an "unlink signal" to the linked entity
             and ignore any future results of the link.</p>
         </note>
       </desc>
     </func>
+
     <func>
       <name name="unregister" arity="1"/>
-      <fsummary>Remove the registered name for a process (or port)</fsummary>
+      <fsummary>Removes the registered name for a process (or port).</fsummary>
       <desc>
-        <p>Removes the registered name <c><anno>RegName</anno></c>, associated with a
-          pid or a port identifier.</p>
+        <p>Removes the registered name <c><anno>RegName</anno></c>
+          associated with a
+          process identifier or a port identifier, for example:</p>
         <pre>
 > <input>unregister(db).</input>
 true</pre>
@@ -8345,31 +9030,34 @@ true</pre>
           name.</p>
       </desc>
     </func>
+
     <func>
       <name name="whereis" arity="1"/>
-      <fsummary>Get the pid (or port) with a given registered name</fsummary>
+      <fsummary>Gets the pid (or port) with a given registered name.</fsummary>
       <desc>
-        <p>Returns the pid or port identifier with the registered name
-          <c>RegName</c>. Returns <c>undefined</c> if the name is not
-          registered.</p>
+        <p>Returns the process identifier or port identifier with
+          the registered name <c>RegName</c>. Returns <c>undefined</c>
+          if the name is not registered.</p>
+        <p>Example:</p>
         <pre>
 > <input>whereis(db).</input>
 &lt;0.43.0></pre>
       </desc>
     </func>
+
     <func>
       <name name="yield" arity="0"/>
-      <fsummary>Let other processes get a chance to execute</fsummary>
+      <fsummary>Lets other processes get a chance to execute.</fsummary>
       <desc>
-        <p>Voluntarily let other processes (if any) get a chance to
+        <p>Voluntarily lets other processes (if any) get a chance to
           execute. Using <c>erlang:yield()</c> is similar to
           <c>receive after 1 -> ok end</c>, except that <c>yield()</c>
           is faster.</p>
 	<warning><p>There is seldom or never any need to use this BIF,
-	especially in the SMP-emulator as other	processes will have a
-	chance to run in another scheduler thread anyway.
-	Using this BIF without a thorough grasp of how the scheduler
-	works may cause performance degradation.</p></warning>
+        especially in the SMP emulator, as other processes have a
+        chance to run in another scheduler thread anyway.
+        Using this BIF without a thorough grasp of how the scheduler
+        works can cause performance degradation.</p></warning>
       </desc>
     </func>
   </funcs>
diff --git a/erts/doc/src/erts_alloc.xml b/erts/doc/src/erts_alloc.xml
index 376cae4a95..15b78ffa10 100644
--- a/erts/doc/src/erts_alloc.xml
+++ b/erts/doc/src/erts_alloc.xml
@@ -4,7 +4,7 @@
 <cref>
   <header>
     <copyright>
-      <year>2002</year><year>2014</year>
+      <year>2002</year><year>2015</year>
       <holder>Ericsson AB. All Rights Reserved.</holder>
     </copyright>
     <legalnotice>
@@ -260,19 +260,19 @@
     <p>The following flags are available for configuration of
       <c>mseg_alloc</c>:</p>
     <taglist>
-      <tag><marker id="MMamcbf"><c><![CDATA[+MMamcbf <size>]]></c></marker></tag>
+      <tag><marker id="MMamcbf"/><c><![CDATA[+MMamcbf <size>]]></c></tag>
       <item>
        Absolute max cache bad fit (in kilobytes). A segment in the
        memory segment cache is not reused if its size exceeds the
        requested size with more than the value of this
        parameter. Default value is 4096. </item>
-      <tag><marker id="MMrmcbf"><c><![CDATA[+MMrmcbf <ratio>]]></c></marker></tag>
+      <tag><marker id="MMrmcbf"/><c><![CDATA[+MMrmcbf <ratio>]]></c></tag>
       <item>
        Relative max cache bad fit (in percent). A segment in the
        memory segment cache is not reused if its size exceeds the
        requested size with more than relative max cache bad fit
        percent of the requested size. Default value is 20.</item>
-      <tag><marker id="MMsco"><c><![CDATA[+MMsco true|false]]></c></marker></tag>
+      <tag><marker id="MMsco"/><c><![CDATA[+MMsco true|false]]></c></tag>
       <item>
 	Set <seealso marker="#MMscs">super carrier</seealso> only flag. This
 	flag defaults to <c>true</c>. When a super carrier is used and this
@@ -292,7 +292,7 @@
 	disabled on halfword heap systems. This flag will be
 	ignored on halfword heap systems.
       </item>
-      <tag><marker id="MMscrfsd"><c><![CDATA[+MMscrfsd <amount>]]></c></marker></tag>
+      <tag><marker id="MMscrfsd"/><c><![CDATA[+MMscrfsd <amount>]]></c></tag>
       <item>
 	Set <seealso marker="#MMscs">super carrier</seealso> reserved
 	free segment descriptors. This parameter defaults to <c>65536</c>.
@@ -305,7 +305,7 @@
 	<c>erts_mmap</c> tuple part of the result from calling
 	<seealso marker="erts:erlang#system_info_allocator_tuple">erlang:system_info({allocator, mseg_alloc})</seealso>.
       </item>
-      <tag><marker id="MMscrpm"><c><![CDATA[+MMscrpm true|false]]></c></marker></tag>
+      <tag><marker id="MMscrpm"/><c><![CDATA[+MMscrpm true|false]]></c></tag>
       <item>
 	Set <seealso marker="#MMscs">super carrier</seealso> reserve physical
 	memory flag. This flag defaults	to <c>true</c>. When this flag is
@@ -328,7 +328,7 @@
 	disabled on halfword heap systems. This flag will be
 	ignored on halfword heap systems.
       </item>
-      <tag><marker id="MMscs"><c><![CDATA[+MMscs <size in MB>]]></c></marker></tag>
+      <tag><marker id="MMscs"/><c><![CDATA[+MMscs <size in MB>]]></c></tag>
       <item>
 	Set super carrier size (in MB). The super carrier size defaults to
 	zero; i.e, the super carrier is by default disabled. The super
@@ -343,7 +343,7 @@
 	disabled on halfword heap systems. This flag will be
 	ignored on halfword heap systems.
       </item>
-      <tag><marker id="MMmcs"><c><![CDATA[+MMmcs <amount>]]></c></marker></tag>
+      <tag><marker id="MMmcs"/><c><![CDATA[+MMmcs <amount>]]></c></tag>
       <item>
        Max cached segments. The maximum number of memory segments
        stored in the memory segment cache. Valid range is
@@ -352,15 +352,15 @@
     <p>The following flags are available for configuration of
       <c>sys_alloc</c>:</p>
     <taglist>
-      <tag><marker id="MYe"><c>+MYe true</c></marker></tag>
+      <tag><marker id="MYe"/><c>+MYe true</c></tag>
       <item>
        Enable <c>sys_alloc</c>. Note: <c>sys_alloc</c> cannot be disabled.</item>
-      <tag><marker id="MYm"><c>+MYm libc</c></marker></tag>
+      <tag><marker id="MYm"/><c>+MYm libc</c></tag>
       <item>
       <c>malloc</c> library to use. Currently only
       <c>libc</c> is available. <c>libc</c> enables the standard
       <c>libc</c> malloc implementation. By default <c>libc</c> is used.</item>
-      <tag><marker id="MYtt"><c><![CDATA[+MYtt <size>]]></c></marker></tag>
+      <tag><marker id="MYtt"/><c><![CDATA[+MYtt <size>]]></c></tag>
       <item>
        Trim threshold size (in kilobytes). This is the maximum amount
        of free memory at the top of the heap (allocated by
@@ -372,7 +372,7 @@
        trim threshold is 128. <em>Note:</em> This flag will
        only have any effect when the emulator has been linked with
        the GNU C library, and uses its <c>malloc</c> implementation.</item>
-      <tag><marker id="MYtp"><c><![CDATA[+MYtp <size>]]></c></marker></tag>
+      <tag><marker id="MYtp"/><c><![CDATA[+MYtp <size>]]></c></tag>
       <item>
        Top pad size (in kilobytes). This is the amount of extra
        memory that will be allocated by <c>malloc</c> when
@@ -390,7 +390,7 @@
        subsystem identifier, only the specific allocator identified will be
        effected:</p>
     <taglist>
-      <tag><marker id="M_acul"><c><![CDATA[+M<S>acul <utilization>|de]]></c></marker></tag>
+      <tag><marker id="M_acul"/><c><![CDATA[+M<S>acul <utilization>|de]]></c></tag>
       <item>
 	Abandon carrier utilization limit. A valid
 	<c><![CDATA[<utilization>]]></c> is an integer in the range
@@ -422,7 +422,7 @@
 	allocators based on the <c>alloc_util</c> framework with the
 	exception of <c>temp_alloc</c> (which would be pointless).
       </item>
-      <tag><marker id="M_as"><c><![CDATA[+M<S>as bf|aobf|aoff|aoffcbf|aoffcaobf|gf|af]]></c></marker></tag>
+      <tag><marker id="M_as"/><c><![CDATA[+M<S>as bf|aobf|aoff|aoffcbf|aoffcaobf|gf|af]]></c></tag>
       <item>
        Allocation strategy. Valid strategies are <c>bf</c> (best fit),
       <c>aobf</c> (address order best fit), <c>aoff</c> (address order first fit),
@@ -430,7 +430,7 @@
       <c>aoffcaobf</c> (address order first fit carrier address order best fit),
       <c>gf</c> (good fit), and <c>af</c> (a fit). See 
       <seealso marker="#strategy">the description of allocation strategies</seealso> in "the <c>alloc_util</c> framework" section.</item>
-      <tag><marker id="M_asbcst"><c><![CDATA[+M<S>asbcst <size>]]></c></marker></tag>
+      <tag><marker id="M_asbcst"/><c><![CDATA[+M<S>asbcst <size>]]></c></tag>
       <item>
        Absolute singleblock carrier shrink threshold (in
        kilobytes). When a block located in an
@@ -438,23 +438,23 @@
        will be left unchanged if the amount of unused memory is less
        than this threshold; otherwise, the carrier will be shrunk.
        See also <seealso marker="#M_rsbcst">rsbcst</seealso>.</item>
-      <tag><marker id="M_e"><c><![CDATA[+M<S>e true|false]]></c></marker></tag>
+      <tag><marker id="M_e"/><c><![CDATA[+M<S>e true|false]]></c></tag>
       <item>
        Enable allocator <c><![CDATA[<S>]]></c>.</item>
-      <tag><marker id="M_lmbcs"><c><![CDATA[+M<S>lmbcs <size>]]></c></marker></tag>
+      <tag><marker id="M_lmbcs"/><c><![CDATA[+M<S>lmbcs <size>]]></c></tag>
       <item>
        Largest (<c>mseg_alloc</c>) multiblock carrier size (in
        kilobytes).  See <seealso marker="#mseg_mbc_sizes">the description
        on how sizes for mseg_alloc multiblock carriers are decided</seealso>
        in "the <c>alloc_util</c> framework" section. On 32-bit Unix style OS
        this limit can not be set higher than 128 megabyte.</item>
-      <tag><marker id="M_mbcgs"><c><![CDATA[+M<S>mbcgs <ratio>]]></c></marker></tag>
+      <tag><marker id="M_mbcgs"/><c><![CDATA[+M<S>mbcgs <ratio>]]></c></tag>
       <item>
        (<c>mseg_alloc</c>) multiblock carrier growth stages. See
       <seealso marker="#mseg_mbc_sizes">the description on how sizes for
        mseg_alloc multiblock carriers are decided</seealso>
        in "the <c>alloc_util</c> framework" section.</item>
-      <tag><marker id="M_mbsd"><c><![CDATA[+M<S>mbsd <depth>]]></c></marker></tag>
+      <tag><marker id="M_mbsd"/><c><![CDATA[+M<S>mbsd <depth>]]></c></tag>
       <item>
        Max block search depth. This flag has effect only if the
        good fit strategy has been selected for allocator
@@ -464,40 +464,40 @@
        search depth sets a limit on the maximum number of blocks to
        inspect in a free list during a search for suitable block
        satisfying the request.</item>
-      <tag><marker id="M_mmbcs"><c><![CDATA[+M<S>mmbcs <size>]]></c></marker></tag>
+      <tag><marker id="M_mmbcs"/><c><![CDATA[+M<S>mmbcs <size>]]></c></tag>
       <item>
        Main multiblock carrier size. Sets the size of the main
        multiblock carrier for allocator <c><![CDATA[<S>]]></c>. The main
        multiblock carrier is allocated via <c><![CDATA[sys_alloc]]></c> and is
        never deallocated.</item>
-      <tag><marker id="M_mmmbc"><c><![CDATA[+M<S>mmmbc <amount>]]></c></marker></tag>
+      <tag><marker id="M_mmmbc"/><c><![CDATA[+M<S>mmmbc <amount>]]></c></tag>
       <item>
        Max <c>mseg_alloc</c> multiblock carriers. Maximum number of
        multiblock carriers allocated via <c>mseg_alloc</c> by
        allocator <c><![CDATA[<S>]]></c>. When this limit has been reached,
        new multiblock carriers will be allocated via
       <c>sys_alloc</c>.</item>
-      <tag><marker id="M_mmsbc"><c><![CDATA[+M<S>mmsbc <amount>]]></c></marker></tag>
+      <tag><marker id="M_mmsbc"/><c><![CDATA[+M<S>mmsbc <amount>]]></c></tag>
       <item>
        Max <c>mseg_alloc</c> singleblock carriers. Maximum number of
        singleblock carriers allocated via <c>mseg_alloc</c> by
        allocator <c><![CDATA[<S>]]></c>. When this limit has been reached,
        new singleblock carriers will be allocated via
       <c>sys_alloc</c>.</item>
-      <tag><marker id="M_ramv"><c><![CDATA[+M<S>ramv <bool>]]></c></marker></tag>
+      <tag><marker id="M_ramv"/><c><![CDATA[+M<S>ramv <bool>]]></c></tag>
       <item>
        Realloc always moves. When enabled, reallocate operations will
        more or less be translated into an allocate, copy, free sequence.
        This often reduce memory fragmentation, but costs performance.
       </item>
-      <tag><marker id="M_rmbcmt"><c><![CDATA[+M<S>rmbcmt <ratio>]]></c></marker></tag>
+      <tag><marker id="M_rmbcmt"/><c><![CDATA[+M<S>rmbcmt <ratio>]]></c></tag>
       <item>
        Relative multiblock carrier move threshold (in percent). When
        a block located in a multiblock carrier is shrunk,
        the block will be moved if the ratio of the size of the returned
        memory compared to the previous size is more than this threshold;
        otherwise, the block will be shrunk at current location.</item>
-      <tag><marker id="M_rsbcmt"><c><![CDATA[+M<S>rsbcmt <ratio>]]></c></marker></tag>
+      <tag><marker id="M_rsbcmt"/><c><![CDATA[+M<S>rsbcmt <ratio>]]></c></tag>
       <item>
        Relative singleblock carrier move threshold (in percent). When
        a block located in a singleblock carrier is shrunk to
@@ -506,7 +506,7 @@
        the block will be left unchanged in the singleblock carrier if
        the ratio of unused memory is less than this threshold;
        otherwise, it will be moved into a multiblock carrier. </item>
-      <tag><marker id="M_rsbcst"><c><![CDATA[+M<S>rsbcst <ratio>]]></c></marker></tag>
+      <tag><marker id="M_rsbcst"/><c><![CDATA[+M<S>rsbcst <ratio>]]></c></tag>
       <item>
        Relative singleblock carrier shrink threshold (in
        percent). When a block located in an <c>mseg_alloc</c>
@@ -514,20 +514,20 @@
        unchanged if the ratio of unused memory is less than this
        threshold; otherwise, the carrier will be shrunk.
        See also <seealso marker="#M_asbcst">asbcst</seealso>.</item>
-      <tag><marker id="M_sbct"><c><![CDATA[+M<S>sbct <size>]]></c></marker></tag>
+      <tag><marker id="M_sbct"/><c><![CDATA[+M<S>sbct <size>]]></c></tag>
       <item>
        Singleblock carrier threshold. Blocks larger than this
        threshold will be placed in singleblock carriers. Blocks
        smaller than this threshold will be placed in multiblock
        carriers. On 32-bit Unix style OS this threshold can not be set higher
        than 8 megabytes.</item>
-      <tag><marker id="M_smbcs"><c><![CDATA[+M<S>smbcs <size>]]></c></marker></tag>
+      <tag><marker id="M_smbcs"/><c><![CDATA[+M<S>smbcs <size>]]></c></tag>
       <item>
        Smallest (<c>mseg_alloc</c>) multiblock carrier size (in
        kilobytes). See <seealso marker="#mseg_mbc_sizes">the description
        on how sizes for mseg_alloc multiblock carriers are decided</seealso>
        in "the <c>alloc_util</c> framework" section.</item>
-      <tag><marker id="M_t"><c><![CDATA[+M<S>t true|false]]></c></marker></tag>
+      <tag><marker id="M_t"/><c><![CDATA[+M<S>t true|false]]></c></tag>
       <item>
        <p>Multiple, thread specific instances of the allocator.
        This option will only have any effect on the runtime system
@@ -544,20 +544,20 @@
       <c>alloc_util</c>, i.e. all allocators based on <c>alloc_util</c>
       will be effected:</p>
     <taglist>
-      <tag><marker id="Muycs"><c><![CDATA[+Muycs <size>]]></c></marker></tag>
+      <tag><marker id="Muycs"/><c><![CDATA[+Muycs <size>]]></c></tag>
       <item>
       <c>sys_alloc</c> carrier size. Carriers allocated via
       <c>sys_alloc</c> will be allocated in sizes which are
        multiples of the <c>sys_alloc</c> carrier size. This is not
        true for main multiblock carriers and carriers allocated
        during a memory shortage, though.</item>
-      <tag><marker id="Mummc"><c><![CDATA[+Mummc <amount>]]></c></marker></tag>
+      <tag><marker id="Mummc"/><c><![CDATA[+Mummc <amount>]]></c></tag>
       <item>
        Max <c>mseg_alloc</c> carriers. Maximum number of carriers
        placed in separate memory segments. When this limit has been
        reached, new carriers will be placed in memory retrieved from
       <c>sys_alloc</c>.</item>
-      <tag><marker id="Musac"><c><![CDATA[+Musac <bool>]]></c></marker></tag>
+      <tag><marker id="Musac"/><c><![CDATA[+Musac <bool>]]></c></tag>
       <item>
 	Allow <c>sys_alloc</c> carriers. By default <c>true</c>. If
 	set to <c>false</c>, <c>sys_alloc</c> carriers will never be
@@ -565,19 +565,19 @@
     </taglist>
     <p>Instrumentation flags:</p>
     <taglist>
-      <tag><marker id="Mim"><c>+Mim true|false</c></marker></tag>
+      <tag><marker id="Mim"/><c>+Mim true|false</c></tag>
       <item>
        A map over current allocations is kept by the emulator. The
        allocation map can be retrieved via the <c>instrument</c>
        module. <c>+Mim true</c> implies <c>+Mis true</c>.
       <c>+Mim true</c> is the same as
       <seealso marker="erl#instr">-instr</seealso>.</item>
-      <tag><marker id="Mis"><c>+Mis true|false</c></marker></tag>
+      <tag><marker id="Mis"/><c>+Mis true|false</c></tag>
       <item>
        Status over allocated memory is kept by the emulator. The
        allocation status can be retrieved via the <c>instrument</c>
        module.</item>
-      <tag><marker id="Mit"><c>+Mit X</c></marker></tag>
+      <tag><marker id="Mit"/><c>+Mit X</c></tag>
       <item>
        Reserved for future use. Do <em>not</em> use this flag.</item>
     </taglist>
@@ -587,7 +587,7 @@
     </note>
     <p>Other flags:</p>
     <taglist>
-      <tag><marker id="Mea"><c>+Mea min|max|r9c|r10b|r11b|config</c></marker></tag>
+      <tag><marker id="Mea"/><c>+Mea min|max|r9c|r10b|r11b|config</c></tag>
       <item>
       <taglist>
         <tag><c>min</c></tag>
@@ -617,7 +617,7 @@
         </item>
       </taglist>
       </item>
-      <tag><marker id="Mlpm"><c>+Mlpm all|no</c></marker></tag>
+      <tag><marker id="Mlpm"/><c>+Mlpm all|no</c></tag>
       <item>Lock physical memory. The default value is <c>no</c>, i.e.,
       no physical memory will be locked. If set to <c>all</c>, all
       memory mappings made by the runtime system, will be locked into
diff --git a/erts/doc/src/escript.xml b/erts/doc/src/escript.xml
index 46110333f9..f12f76890c 100644
--- a/erts/doc/src/escript.xml
+++ b/erts/doc/src/escript.xml
@@ -4,7 +4,7 @@
 <comref>
   <header>
     <copyright>
-      <year>2007</year><year>2014</year>
+      <year>2007</year><year>2015</year>
       <holder>Ericsson AB. All Rights Reserved.</holder>
     </copyright>
     <legalnotice>
@@ -96,8 +96,8 @@ $ <input>escript factorial 5</input>        </pre>
         <note><p>
           The encoding specified by the above mentioned comment
           applies to the script itself. The encoding of the
-          I/O-server, however, has to be set explicitly like this:
-<code>io:setopts([{encoding, unicode}])</code></p>
+          I/O-server, however, has to be set explicitly like this:</p>
+<code>io:setopts([{encoding, unicode}])</code>
           <p>The default encoding of the I/O-server for <c>standard_io</c>
           is <c>latin1</c>
           since the script runs in a non-interactive terminal
diff --git a/erts/doc/src/init.xml b/erts/doc/src/init.xml
index fe26df61f7..2a33096d04 100644
--- a/erts/doc/src/init.xml
+++ b/erts/doc/src/init.xml
@@ -247,10 +247,7 @@
           <c>Expr</c> during system initialization. If any of these
           steps fail (syntax error, parse error or exception during
           evaluation), Erlang stops with an error message. Here is an
-          example that seeds the random number generator:</p>
-        <pre>
-% <input>erl -eval '{X,Y,Z} = now(), random:seed(X,Y,Z).'</input></pre>
-        <p>This example uses Erlang as a hexadecimal calculator:</p>
+          example that uses Erlang as a hexadecimal calculator:</p>
         <pre>
 % <input>erl -noshell -eval 'R = 16#1F+16#A0, io:format("~.16B~n", [R])' \\</input>
 <input>-s erlang halt</input>
diff --git a/erts/doc/src/notes.xml b/erts/doc/src/notes.xml
index 2d96ed6105..a726cc7b97 100644
--- a/erts/doc/src/notes.xml
+++ b/erts/doc/src/notes.xml
@@ -4,7 +4,7 @@
 <chapter>
   <header>
     <copyright>
-      <year>2004</year><year>2013</year>
+      <year>2004</year><year>2015</year>
       <holder>Ericsson AB. All Rights Reserved.</holder>
     </copyright>
     <legalnotice>
@@ -31,6 +31,378 @@
   </header>
   <p>This document describes the changes made to the ERTS application.</p>
 
+
+<section><title>Erts 7.2.1</title>
+
+    <section><title>Fixed Bugs and Malfunctions</title>
+      <list>
+        <item>
+          <p>
+	    Revert "Fix erroneous splitting of emulator path"</p>
+          <p>
+	    Own Id: OTP-13202</p>
+        </item>
+        <item>
+          <p>
+	    Fix HiPE enabled emulator for FreeBSD.</p>
+          <p>
+	    Own Id: OTP-13204 Aux Id: pr926 </p>
+        </item>
+      </list>
+    </section>
+
+</section>
+
+<section><title>Erts 7.2</title>
+
+    <section><title>Fixed Bugs and Malfunctions</title>
+      <list>
+        <item>
+          <p>
+	    Small documentation fixes</p>
+          <p>
+	    Own Id: OTP-13017</p>
+        </item>
+        <item>
+          <p>
+	    Fix memory corruption bug caused by disabling
+	    distribution and then re-enable distribution with a node
+	    name that has previously been used by a remote node.</p>
+          <p>
+	    Own Id: OTP-13076 Aux Id: seq12959 </p>
+        </item>
+        <item>
+          <p>
+	    Renamed variables with name bool as Visual Studio 2015
+	    now treats this is a keyword.</p>
+          <p>
+	    Own Id: OTP-13079</p>
+        </item>
+        <item>
+	    <p><c>erl_prim_loader</c> has not supported custom
+	    loaders for several releases. In the documentation for
+	    <c>erl_prim_loader</c>, all references to custom loaders
+	    have now been removed.</p>
+          <p>
+	    Own Id: OTP-13102</p>
+        </item>
+        <item>
+          <p>
+	    Fixed compilation of erts together with libc versions
+	    that do not define __uint32_t.</p>
+          <p>
+	    Own Id: OTP-13105</p>
+        </item>
+        <item>
+          <p>
+	    erl -make now returns non-zero exit codes on failure</p>
+          <p>
+	    Own Id: OTP-13107</p>
+        </item>
+        <item>
+          <p>
+	    Fix crash on init:restart in embedded mode caused by
+	    on_load handler process not being relaunched leading to
+	    load failure for modules such as crypto and asn1rt_nif
+	    that need it to be present for correct NIF loading.</p>
+          <p>
+	    Own Id: OTP-13115</p>
+        </item>
+        <item>
+          <p>
+	    Fix maps decode in erlang:binary_to_term/1</p>
+	    <p>Decoding a term with a large (HAMT) map in an small
+	    (FLAT) map could cause a critical error if the external
+	    format was not produced by beam.</p>
+          <p>
+	    Own Id: OTP-13125</p>
+        </item>
+        <item>
+          <p>
+	    Fix very rare bug in GC when big maps with a lot of hash
+	    collisions from a remote node are waiting in inner
+	    message queue.</p>
+          <p>
+	    Own Id: OTP-13146</p>
+        </item>
+        <item>
+          <p>
+	    Fixed a bug that could cause a crash dump to become
+	    almost empty.</p>
+          <p>
+	    Own Id: OTP-13150</p>
+        </item>
+      </list>
+    </section>
+
+
+    <section><title>Improvements and New Features</title>
+      <list>
+        <item>
+	    <p> Updated the xmllint target to just check the xml
+	    files with real documentation content.<br/> Corrected
+	    some errors and added some missing target in the DTD's.
+	    </p>
+          <p>
+	    Own Id: OTP-13026</p>
+        </item>
+        <item>
+          <p>
+	    Add function enif_getenv to read OS environment variables
+	    in a portable way from NIFs.</p>
+          <p>
+	    Own Id: OTP-13147</p>
+        </item>
+      </list>
+    </section>
+
+</section>
+
+<section><title>Erts 7.1</title>
+    <section><title>Fixed Bugs and Malfunctions</title>
+      <list>
+        <item>
+          <p>
+	    Fix bug in ETS that could cause stray objects marked for
+	    deletion to occasionally be missed by the cleanup done by
+	    <c>safe_fixtable(_,false)</c>.</p>
+          <p>
+	    Own Id: OTP-12870</p>
+        </item>
+        <item>
+          <p>
+	    Fixed VM crash that could occur if a trace port was
+	    linked to a process, and the trace port terminated
+	    abnormally while handling a trace message. This bug has
+	    always existed in the runtime system with SMP support.</p>
+          <p>
+	    Own Id: OTP-12901</p>
+        </item>
+        <item>
+          <p>
+	    Instead of aborting, the vm now creates a crash dump when
+	    a system process is terminated.</p>
+          <p>
+	    Own Id: OTP-12934</p>
+        </item>
+        <item>
+          <p>
+	    Fixed a rare emulator dead lock that occurred when
+	    erlang:process_flag(priority,...) was called by a process
+	    that was also scheduled for an internal system activity.</p>
+          <p>
+	    Own Id: OTP-12943</p>
+        </item>
+        <item>
+          <p>
+	    The runtime system on various posix platforms (except for
+	    Linux and Solaris) could crash when large amounts of
+	    file-descriptors were in use.</p>
+          <p>
+	    Own Id: OTP-12954</p>
+        </item>
+        <item>
+          <p>
+	    A beam file compiled by hipe for an incompatible runtime
+	    system was sometimes not rejected by the loader, which
+	    could lead to vm crash. This fix will also allow the same
+	    hipe compiler to be used by both normal and debug-built
+	    vm.</p>
+          <p>
+	    Own Id: OTP-12962</p>
+        </item>
+        <item>
+          <p>
+	    Fix bug in <c>maps:merge/2</c> when called by hipe
+	    compiled code that could cause vm crash. Bug exists since
+	    erts-7.0 (OTP 18.0).</p>
+          <p>
+	    Own Id: OTP-12965</p>
+        </item>
+        <item>
+          <p>
+	    When tracing with <c>process_dump</c> option, the VM
+	    could abort if there was an ongoing binary match
+	    somewhere in the call stack of the traced process.</p>
+          <p>
+	    Own Id: OTP-12968</p>
+        </item>
+        <item>
+          <p>
+	    Fixed possible output deadlock in tty driver when hitting
+	    "CTRL-C" in a non-smp emulator shell on unix.</p>
+          <p>
+	    Own Id: OTP-12987 Aux Id: Seq12947 </p>
+        </item>
+        <item>
+          <p>
+	    Fix <c>binary_to_integer</c> to throw badarg for "+" and
+	    "-" similar to <c>list_to_integer</c>.</p>
+          <p>
+	    Own Id: OTP-12988</p>
+        </item>
+        <item>
+          <p>
+	    Suppress warning of unused argument when using macro
+	    enif_make_pid.</p>
+          <p>
+	    Own Id: OTP-12989</p>
+        </item>
+      </list>
+    </section>
+
+
+    <section><title>Improvements and New Features</title>
+      <list>
+        <item>
+          <p>
+	    Changed default clock source used for OS system time on
+	    MacOS X to <c>gettimeofday()</c> in order to improve
+	    performance. The system can be configured during build to
+	    use the previously used higher resolution clock source by
+	    passing the switch <seealso
+	    marker="doc/installation_guide:INSTALL#Advanced-configuration-and-build-of-ErlangOTP_Configuring"><c>--with-clock-resolution=high</c></seealso>
+	    when configuring the build.</p>
+          <p>
+	    Own Id: OTP-12945 Aux Id: OTP-12892 </p>
+        </item>
+        <item>
+          <p>
+	    Added the <c>configure</c> option <seealso
+	    marker="doc/installation_guide:INSTALL#Advanced-configuration-and-build-of-ErlangOTP_Configuring"><c>--disable-saved-compile-time</c></seealso>
+	    which disables saving of compile date and time in the
+	    emulator binary.</p>
+          <p>
+	    Own Id: OTP-12971</p>
+        </item>
+      </list>
+    </section>
+
+</section>
+
+<section><title>Erts 7.0.3</title>
+
+    <section><title>Fixed Bugs and Malfunctions</title>
+      <list>
+        <item>
+          <p>
+	    Fixed a binary memory leak when printing to shell using
+	    the tty driver (i.e. not -oldshell).</p>
+          <p>
+	    Own Id: OTP-12941</p>
+        </item>
+        <item>
+          <p>
+	    Fix a bug where the standard error port sometimes crashes
+	    with eagain as the reason.</p>
+          <p>
+	    Own Id: OTP-12942</p>
+        </item>
+        <item>
+           <p>
+	    When tracing with <c>process_dump</c> option, the VM
+	    could abort if there was an ongoing binary match
+	    somewhere in the call stack of the traced process./</p>
+          <p>
+	    Own Id: OTP-12968</p>
+        </item>
+      </list>
+    </section>
+
+</section>
+
+<section><title>Erts 7.0.2</title>
+
+    <section><title>Fixed Bugs and Malfunctions</title>
+      <list>
+        <item>
+          <p>
+	    A process could end up in an inconsistent half exited
+	    state in the runtime system without SMP support. This
+	    could occur if the processes was traced by a port that it
+	    also was linked to, and the port terminated abnormally
+	    while handling a trace message for the process.</p>
+          <p>
+	    This bug has always existed in the runtime system without
+	    SMP support, but never in the runtime system with SMP
+	    support.</p>
+          <p>
+	    Own Id: OTP-12889 Aux Id: seq12885 </p>
+        </item>
+        <item>
+          <p>
+	    Removed unnecessary copying of data when retrieving
+	    corrected Erlang monotonic time.</p>
+          <p>
+	    Own Id: OTP-12894</p>
+        </item>
+        <item>
+          <p>
+	    Changed default OS monotonic clock source chosen at build
+	    time. This in order to improve performance. The behavior
+	    will now on most systems be that (both OS and Erlang)
+	    monotonic time stops when the system is suspended.</p>
+          <p>
+	    If you prefer that monotonic time elapse during suspend
+	    of the machine, you can pass the command line argument
+	    <c>--enable-prefer-elapsed-monotonic-time-during-suspend</c>
+	    to <c>configure</c> when building Erlang/OTP. The
+	    configuration stage will try to find such a clock source,
+	    but might not be able to find it. Note that there might
+	    be a performance penalty associated with such a clock
+	    source.</p>
+          <p>
+	    *** POTENTIAL INCOMPATIBILITY ***</p>
+          <p>
+	    Own Id: OTP-12895</p>
+        </item>
+        <item>
+          <p>
+	    <c>erlang:system_info(end_time)</c> returned a faulty
+	    value on 32-bit architectures.</p>
+          <p>
+	    Own Id: OTP-12896</p>
+        </item>
+      </list>
+    </section>
+
+
+    <section><title>Improvements and New Features</title>
+      <list>
+        <item>
+          <p>
+	    The <c>configure</c> command line argument
+	    <c>--enable-gettimeofday-as-os-system-time</c> has been
+	    added which force usage of <c>gettimeofday()</c> for OS
+	    system time. This will improve performance of
+	    <c>os:system_time()</c> and <c>os:timestamp()</c> on
+	    MacOS X, at the expense of worse accuracy, resolution and
+	    precision of Erlang monotonic time, Erlang system time,
+	    and OS system time.</p>
+          <p>
+	    Own Id: OTP-12892</p>
+        </item>
+      </list>
+    </section>
+
+</section>
+
+<section><title>Erts 7.0.1</title>
+
+    <section><title>Fixed Bugs and Malfunctions</title>
+      <list>
+        <item>
+          <p>
+	    Fix a rare hanging of the VM seen to happen just after
+	    emulator start. Bug exists since R14.</p>
+          <p>
+	    Own Id: OTP-12859 Aux Id: seq12882 </p>
+        </item>
+      </list>
+    </section>
+
+</section>
+
 <section><title>Erts 7.0</title>
 
     <section><title>Fixed Bugs and Malfunctions</title>
@@ -462,19 +834,20 @@
         </item>
         <item>
           <p>
-	    Use persistent hashmaps for large Maps <p>Maps will use a
+	    Use persistent hashmaps for large Maps</p>
+          <p>Maps will use a
 	    persistent hashmap implementation when the number of
 	    pairs in a Map becomes sufficiently large. The change
 	    will occur when a Map reaches 33 pairs in size but this
-	    limit might change in the future.</p></p>
-	    <p>The most significant impact for the user by this
+	    limit might change in the future.</p>
+          <p>The most significant impact for the user by this
 	    change is speed, and to a lesser degree memory
 	    consumption and introspection of Maps. Memory consumption
 	    size is probalistic but lesser than <c>gb_trees</c> or
 	    <c>dict</c> for instance. Any other impacts will be
 	    transparent for the user except for the following
 	    changes.</p>
-	    <p>Semantics of Maps have changed in two incompatible
+          <p>Semantics of Maps have changed in two incompatible
 	    ways compared to the experimental implementation in OTP
 	    17:</p> <list> <item>Hashing of maps is done different by
 	    <c>erlang:phash2/1,2</c>, <c>erlang:phash/1</c> and
@@ -734,6 +1107,42 @@
 
 </section>
 
+<section><title>Erts 6.4.1.5</title>
+
+    <section><title>Fixed Bugs and Malfunctions</title>
+      <list>
+        <item>
+          <p>
+	    Fixed a bug that could cause a crash dump to become
+	    almost empty.</p>
+          <p>
+	    Own Id: OTP-13150</p>
+        </item>
+      </list>
+    </section>
+
+</section>
+
+<section><title>Erts 6.4.1.4</title>
+
+    <section><title>Fixed Bugs and Malfunctions</title>
+      <list>
+        <item>
+          <p>
+	    The 'raw' socket option could not be used multiple times
+	    in one call to any e.g gen_tcp function because only one
+	    of the occurrences were used. This bug has been fixed,
+	    and also a small bug concerning propagating error codes
+	    from within inet:setopts/2.</p>
+          <p>
+	    Own Id: OTP-11482 Aux Id: seq12872 </p>
+        </item>
+      </list>
+    </section>
+
+</section>
+
+
 <section><title>Erts 6.4.1</title>
 
     <section><title>Fixed Bugs and Malfunctions</title>
@@ -1122,7 +1531,7 @@
           <p>
 	    Improved support for atomic memory operations provided by
 	    the <url
-	    href="https://github.com/ivmai/libatomic_ops/"><c>libatomic_ops</c></url>
+	    href="https://github.com/ivmai/libatomic_ops/">libatomic_ops</url>
 	    library. Most importantly support for use of native
 	    double word atomics when implemented by
 	    <c>libatomic_ops</c> (for example, implemented for ARM).</p>
@@ -2089,22 +2498,28 @@
           <p>
 	    EEP43: New data type - Maps</p>
           <p>
-	    With Maps you may for instance: <taglist> <item><c>M0 =
-	    #{ a =&gt; 1, b =&gt; 2}, % create
-	    associations</c></item> <item><c>M1 = M0#{ a := 10 }, %
-	    update values</c></item> <item><c>M2 = M1#{ "hi" =&gt;
-	    "hello"}, % add new associations</c></item> <item><c>#{
-	    "hi" := V1, a := V2, b := V3} = M2. % match keys with
-	    values</c></item> </taglist></p>
+	    With Maps you may for instance:</p>
+          <taglist>
+            <tag/> <item><c>M0 = #{ a =&gt; 1, b =&gt; 2}, % create
+	      associations</c></item>
+            <tag/><item><c>M1 = M0#{ a := 10 }, % update values</c></item>
+            <tag/><item><c>M2 = M1#{ "hi" =&gt;
+	      "hello"}, % add new associations</c></item>
+            <tag/><item><c>#{ "hi" := V1, a := V2, b := V3} = M2.
+               % match keys with values</c></item>
+          </taglist>
           <p>
 	    For information on how to use Maps please see Map Expressions in the
 		<seealso marker="doc/reference_manual:expressions#map_expressions">
 			Reference Manual</seealso>.</p>
           <p>
 	    The current implementation is without the following
-	    features: <taglist> <item>No variable keys</item>
-	    <item>No single value access</item> <item>No map
-	    comprehensions</item> </taglist></p>
+	    features:</p>
+          <taglist>
+            <tag/><item>No variable keys</item>
+	    <tag/><item>No single value access</item>
+            <tag/><item>No map comprehensions</item>
+          </taglist>
           <p>
 	    Note that Maps is <em>experimental</em> during OTP 17.0.</p>
           <p>
@@ -4264,8 +4679,7 @@
           <p>
 	    Fix erl_prim_loader errors in handling of primary
 	    archive. The following errors have been corrected:</p>
-          <p>
-	    <list> <item> If primary archive was named "xxx", then a
+          <list> <item> If primary archive was named "xxx", then a
 	    file in the same directory named "xxxyyy" would be
 	    interpreted as a file named "yyy" inside the archive.
 	    </item> <item> erl_prim_loader did not correctly create
@@ -4280,7 +4694,8 @@
 	    erl_prim_loader:list_dir/1 would sometimes return an
 	    empty string inside the file list. This was a virtual
 	    element representing the top directory of the archive.
-	    This has been removed. </item> </list></p>
+	    This has been removed. </item>
+          </list>
           <p>
 	    Thanks to Tuncer Ayaz and Shunichi Shinohara for
 	    reporting and co-authoring corrections.</p>
@@ -6723,12 +7138,12 @@
 	    Own Id: OTP-8726 Aux Id: seq11617 </p>
         </item>
         <item>
-	    <p>Fix libm linking with --as-needed flag
+	    <p>Fix libm linking with --as-needed flag</p>
           <p>
 	    When building with "--as-needed" linker flags on Linux
 	    the build will fail. This has now been fixed.</p>
           <p>
-	    (Thanks to Christian Faulhammer)</p></p>
+	    (Thanks to Christian Faulhammer)</p>
           <p>
 	    Own Id: OTP-8728</p>
         </item>
diff --git a/erts/doc/src/run_erl.xml b/erts/doc/src/run_erl.xml
index 0a5b2c6136..faec3c68c1 100644
--- a/erts/doc/src/run_erl.xml
+++ b/erts/doc/src/run_erl.xml
@@ -59,7 +59,7 @@
            first argument to run_erl on the command line.</item>
           <tag>pipe_dir</tag>
           <item>This is where to put the named pipe, usually
-          <c><![CDATA[/tmp/]]></c> on Unix or <c><![CDATA[/pipe/]]></c> on OSE. It shall be suffixed by a <c><![CDATA[/]]></c> (slash),
+          <c><![CDATA[/tmp/]]></c>. It shall be suffixed by a <c><![CDATA[/]]></c> (slash),
            i.e. not <c><![CDATA[/tmp/epipies]]></c>, but <c><![CDATA[/tmp/epipes/]]></c>. </item>
           <tag>log_dir</tag>
           <item>This is where the log files are written. There will be one
diff --git a/erts/doc/src/time_correction.xml b/erts/doc/src/time_correction.xml
index aed38fbb92..236fe679cb 100644
--- a/erts/doc/src/time_correction.xml
+++ b/erts/doc/src/time_correction.xml
@@ -4,7 +4,7 @@
 <chapter>
   <header>
     <copyright>
-      <year>1999</year><year>2014</year>
+      <year>1999</year><year>2015</year>
       <holder>Ericsson AB. All Rights Reserved.</holder>
     </copyright>
     <legalnotice>
@@ -35,31 +35,35 @@
 
   <section>
     <title>New Extended Time Functionality</title>
-    <note><p>As of OTP 18 (ERTS version 7.0) the time functionality of
-    Erlang has been extended. This both includes a
+    <note><p>As of OTP 18 (<c>ERTS</c> version 7.0) the time functionality of
+    Erlang has been extended. This includes a
     <seealso marker="#The_New_Time_API">new API</seealso>
-    for time, as well as
+    for time and
     <seealso marker="#Time_Warp_Modes">time warp
-    modes</seealso> which alters the behavior of the system when
+    modes</seealso> that alter the system behavior when
     system time changes.</p>
+
     <p>The <seealso marker="#No_Time_Warp_Mode">default
     time warp mode</seealso> has the same behavior as before, and the
-    old API will still work, so you are not required to change
+    old API still works. Thus, you are not required to change
     anything unless you want to. However, <em>you are strongly
     encouraged to use the new API</em> instead of the old API based
     on <seealso marker="erlang#now/0"><c>erlang:now/0</c></seealso>.
-    <c>erlang:now/0</c> has been deprecated since it is and forever
-    will be a scalability bottleneck. By using the new API you will
+    <c>erlang:now/0</c> is deprecated, as it is and
+    will be a scalability bottleneck.</p>
+
+    <p>By using the new API, you
     automatically get scalability and performance improvements. This
-    will also enable you to use the
-    <seealso marker="#Multi_Time_Warp_Mode">multi time warp mode</seealso>
-    which improves accuracy, and precision of time measurements.</p></note>
+    also enables you to use the
+    <seealso marker="#Multi_Time_Warp_Mode">multi-time warp mode</seealso>
+    that improves accuracy and precision of time measurements.</p>
+    </note>
   </section>
 
   <section>
-    <title>Some Terminology</title>
-    <p>In order to make it easier to understand this document we first
-    define some terminology. This is a mixture of our own terminology
+    <title>Terminology</title>
+    <p>To make it easier to understand this section, some terms
+    are defined. This is a mix of our own terminology
     (Erlang/OS system time, Erlang/OS monotonic time, time warp)
     and globally accepted terminology.</p>
 
@@ -67,7 +71,7 @@
     <section>
       <title>Monotonically Increasing</title>
       <p>In a monotonically increasing sequence of values, all values
-      that have a predecessor are either larger than, or equal to its
+      that have a predecessor are either larger than or equal to its
       predecessor.</p>
     </section>
 
@@ -82,19 +86,19 @@
     <marker id="UT1"/>
     <section>
       <title>UT1</title>
-      <p>Universal Time. Based on the rotation of the earth. Conceptually
-      mean solar time at 0° longitude.</p>
+      <p>Universal Time. UT1 is based on the rotation of the earth
+      and conceptually means solar time at 0° longitude.</p>
     </section>
 
     <marker id="UTC"/>
     <section>
       <title>UTC</title>
-      <p>Coordinated Universal Time. UTC almost align with
-      <seealso marker="#UT1">UT1</seealso>, however, UTC uses the
-      SI definition of a second which is not exactly of the same length
+      <p>Coordinated Universal Time. UTC almost aligns with
+      <seealso marker="#UT1">UT1</seealso>. However, UTC uses the
+      SI definition of a second, which has not exactly the same length
       as the second used by UT1. This means that UTC slowly drifts from
-      UT1. In order to keep UTC relatively in sync with UT1, leap seconds
-      are inserted, and potentially also deleted. That is, an UTC day may
+      UT1. To keep UTC relatively in sync with UT1, leap seconds
+      are inserted, and potentially also deleted. That is, an UTC day can
       be 86400, 86401, or 86399 seconds long.</p>
     </section>
 
@@ -104,14 +108,15 @@
       <p>Time since
       <url href="http://pubs.opengroup.org/onlinepubs/9699919799/xrat/V4_xbd_chap03.html#tag_21_03_00_17">Epoch</url>.
       Epoch is defined to be 00:00:00 <seealso marker="#UTC">UTC</seealso>,
-      January 1, 1970.
+      1970-01-01.
       <url href="http://pubs.opengroup.org/onlinepubs/009604499/basedefs/xbd_chap04.html#tag_04_14">A day in POSIX time</url>
       is defined to be exactly 86400 seconds long. Strangely enough
-      Epoch is defined to be a time in UTC, and UTC have another
+      Epoch is defined to be a time in UTC, and UTC has another
       definition of how long a day is. Quoting the Open Group
-      <url href="http://pubs.opengroup.org/onlinepubs/9699919799/xrat/V4_xbd_chap04.html#tag_21_04_15">"POSIX time is therefore not necessarily UTC, despite its appearance"</url>. The effect of this is that when an UTC leap second is
+      <url href="http://pubs.opengroup.org/onlinepubs/9699919799/xrat/V4_xbd_chap04.html#tag_21_04_15">"POSIX time is therefore not necessarily UTC, despite its appearance"</url>.
+      The effect of this is that when an UTC leap second is
       inserted, POSIX time either stops for a second, or repeats the
-      last second. If an UTC leap second would be deleted (has never
+      last second. If an UTC leap second would be deleted (which has not
       happened yet), POSIX time would make a one second leap forward.</p>
     </section>
 
@@ -125,11 +130,11 @@
     <marker id="Time_Precision"/>
     <section>
       <title>Time Precision</title>
-      <p>The shortest time interval that can be be distinguished
+      <p>The shortest time interval that can be distinguished
       repeatedly and reliably when reading time values. Precision
       is limited by the
       <seealso marker="#Time_Resolution">resolution</seealso>, but
-      resolution and precision might differ significantly.</p>
+      resolution and precision can differ significantly.</p>
     </section>
 
     <marker id="Time_Accuracy"/>
@@ -143,21 +148,23 @@
       <title>Time Warp</title>
       <p>A time warp is a leap forwards or backwards in time. That
       is, the difference of time values taken before and after the
-      time warp will not correspond to the actual elapsed time.</p>
+      time warp does not correspond to the actual elapsed time.</p>
     </section>
 
     <marker id="OS_System_Time"/>
     <section>
       <title>OS System Time</title>
       <p>The operating systems view of
-      <seealso marker="#POSIX_Time">POSIX time</seealso>. It can be
-      retrieved by calling
+      <seealso marker="#POSIX_Time">POSIX time</seealso>. To
+      retrieve it, call
       <seealso marker="kernel:os#system_time/0"><c>os:system_time()</c></seealso>.
       This may or may not be an accurate view of POSIX time. This time
       may typically be adjusted both backwards and forwards without
       limitation. That is, <seealso marker="#Time_Warp">time warps</seealso>
-      may be observed. You can get information about the Erlang runtime
-      system's source of OS system time by calling
+      may be observed.</p>
+
+      <p>To get information about the Erlang runtime
+      system's source of OS system time, call
       <seealso marker="erlang#system_info_os_system_time_source"><c>erlang:system_info(os_system_time_source)</c></seealso>.</p>
     </section>
 
@@ -165,15 +172,17 @@
     <section>
       <title>OS Monotonic Time</title>
       <p>A monotonically increasing time provided by the operating
-      system. This time does not leap and have a relatively steady
+      system. This time does not leap and has a relatively steady
       frequency although not completely correct. However, it is not
-      uncommon that the OS monotonic time stops if the system is
-      suspended. This time typically increase since some
+      uncommon that OS monotonic time stops if the system is
+      suspended. This time typically increases since some
       unspecified point in time that is not connected to
-      <seealso marker="#OS_System_Time">OS system time</seealso>. Note
-      that this type of time is not necessarily provided by all
-      operating systems. You can get information about the Erlang
-      runtime system's source of OS monotonic time by calling
+      <seealso marker="#OS_System_Time">OS system time</seealso>.
+      This type of time is not necessarily provided by all
+      operating systems.</p>
+
+      <p>To get information about the Erlang
+      runtime system's source of OS monotonic time, call
       <seealso marker="erlang#system_info_os_monotonic_time_source"><c>erlang:system_info(os_monotonic_time_source)</c></seealso>.</p>
     </section>
 
@@ -181,14 +190,17 @@
     <section>
       <title>Erlang System Time</title>
       <p>The Erlang runtime systems view of
-      <seealso marker="#POSIX_Time">POSIX time</seealso>. It can be
-      retrieved by calling
-      <seealso marker="erlang#system_time/0"><c>erlang:system_time()</c></seealso>.
-      This time may or may not be an accurate view of POSIX time, and may
+      <seealso marker="#POSIX_Time">POSIX time</seealso>. To
+      retrieve it, call
+      <seealso marker="erlang#system_time/0"><c>erlang:system_time()</c></seealso>.</p>
+
+      <p>This time may or may not be an accurate view of POSIX time,
+      and may
       or may not align with <seealso marker="#OS_System_Time">OS system
       time</seealso>. The runtime system works towards aligning the two
-      system times. Depending on <seealso marker="#Time_Warp_Modes">time
-      warp mode</seealso> used, this may be achieved by letting the Erlang
+      system times. Depending on the
+      <seealso marker="#Time_Warp_Modes">time warp mode</seealso> used,
+      this can be achieved by letting Erlang
       system time perform a <seealso marker="#Time_Warp">time
       warp</seealso>.</p>
     </section>
@@ -197,35 +209,43 @@
     <section>
       <title>Erlang Monotonic Time</title>
       <p>A monotonically increasing time provided by the
-      Erlang runtime system. The Erlang monotonic time increase since
-      some unspecified point in time. It can be retrieved by calling
+      Erlang runtime system. Erlang monotonic time increases since
+      some unspecified point in time. To retrieve it, call
       <seealso marker="erlang#monotonic_time/0"><c>erlang:monotonic_time()</c></seealso>.
-      The
-      <seealso marker="#Time_Accuracy">accuracy</seealso>, and
+      </p>
+
+      <p>The <seealso marker="#Time_Accuracy">accuracy</seealso> and
       <seealso marker="#Time_Precision">precision</seealso> of Erlang
-      monotonic time heavily depends on the accuracy and precision of
-      <seealso marker="#OS_Monotonic_Time">OS monotonic time</seealso>,
-      the accuracy and precision of
-      <seealso marker="#OS_System_Time">OS system time</seealso> as well
-      as on the
-      <seealso marker="#Time_Warp_Modes">time warp mode</seealso>
-      used. On a system that is lacking OS monotonic time, the Erlang
-      monotonic time can only guarantee monotonicity and can more or less
-      not give any other guarantees. The frequency adjustments made to
-      the Erlang monotonic time depends on the time warp mode
-      used.</p>
-
-      <p>Internally in the runtime system the Erlang monotonic
+      monotonic time heavily depends on the following:</p>
+
+      <list type="bulleted">
+        <item>Accuracy and precision of
+          <seealso marker="#OS_Monotonic_Time">OS monotonic time</seealso>
+        </item>
+        <item>Accuracy and precision of
+          <seealso marker="#OS_System_Time">OS system time</seealso>
+        </item>
+          <item><seealso marker="#Time_Warp_Modes">time warp mode</seealso> used
+        </item>
+      </list>
+
+      <p>On a system without OS monotonic time, Erlang monotonic
+      time guarantees monotonicity, but cannot give
+      other guarantees. The frequency adjustments made to
+      Erlang monotonic time depend on the time warp mode used.</p>
+
+      <p>Internally in the runtime system, Erlang monotonic
       time is the "time engine" that is used for more or less
-      everything that has anything to do with time. All timers
+      everything that has anything to do with time. All timers,
       regardless of it is a <c>receive ... after</c> timer, BIF timer,
-      or a timer in the <c>timer</c> module are triggered
+      or a timer in the <c>timer</c> module, are triggered
       relative Erlang monotonic time. Even
       <seealso marker="#Erlang_System_Time">Erlang system
       time</seealso> is based on Erlang monotonic time.
       By adding current Erlang monotonic time with current time
-      offset you get current Erlang system time. Current time
-      offset can be retrieved by calling
+      offset, you get current Erlang system time.</p>
+
+      <p>To retrieve current time offset, call
       <seealso marker="erlang#time_offset/0"><c>erlang:time_offset/0</c></seealso>.
       </p>
     </section>
@@ -234,176 +254,169 @@
 
   <section>
     <title>Introduction</title>
-
   <p>Time is vital to an Erlang program and, more importantly, <em>correct</em>
   time is vital to an Erlang program. As Erlang is a language with
-  soft real time properties and we have the possibility to express
-  time in our programs, the Virtual Machine and the language has to be
-  very careful about what is considered a correct point in time and in
+  soft real-time properties and we can express
+  time in our programs, the Virtual Machine and the language must be
+  careful about what is considered a correct time and in
   how time functions behave.</p>
 
-  <p>In the beginning, Erlang was constructed assuming that the wall
+  <p>When Erlang was designed, it was assumed that the wall
   clock time in the system showed a monotonic time moving forward at
-  exactly the same pace as the definition of time. That more or less
-  meant that an atomic clock (or better) was expected to be attached
+  exactly the same pace as the definition of time. This more or less meant
+  that an atomic clock (or better time source) was expected to be attached
   to your hardware and that the hardware was then expected to be
-  locked away from any human (or unearthly) tinkering for all
-  eternity. While this might be a compelling thought, it's simply
-  never the case.</p>
-
-  <p>A "normal" modern computer can not keep time. Not on itself and
-  not unless you actually have a chip level atomic clock wired to
-  it. Time, as perceived by your computer, will normally need to be
-  corrected. Hence the NTP protocol that together with the ntpd
-  process will do it's best to keep your computers time in sync with
-  the "real" time in the universe. Between NTP corrections, usually a
+  locked away from any human tinkering forever. While this can be a
+  compelling thought, it is simply never the case.</p>
+
+  <p>A "normal" modern computer cannot keep time, not on itself and
+  not unless you have a chip-level atomic clock wired to it. Time,
+  as perceived by your computer, must normally be corrected. Hence
+  the Network Time Protocol (NTP) protocol, together with the <c>ntpd</c>
+  process, does its best to keep your computer time in sync with
+  the correct time. Between NTP corrections, usually a
   less potent time-keeper than an atomic clock is used.</p>
 
-  <p>But NTP is not fail safe. The NTP server can be unavailable, the
-  ntp.conf can be wrongly configured or your computer may from time to
-  time be disconnected from the internet. Furthermore you can have a
-  user (or even system administrator) on your system that thinks the
-  right way to handle daylight saving time is to adjust the clock one
-  hour two times a year (a tip, that is not the right way to do
-  it...). To further complicate things, this user fetched your
-  software from the internet and has never ever thought about what's
-  the correct time as perceived by a computer. The user simply does
-  not care about keeping the wall clock in sync with the rest of the
-  universe. The user expects your program to have omnipotent knowledge
+  <p>However, NTP is not fail-safe. The NTP server can be unavailable,
+  <c>ntp.conf</c> can be wrongly configured, or your computer may
+  sometimes be disconnected from Internet. Furthermore, you can have a
+  user (or even system administrator) who thinks the correct
+  way to handle Daylight Saving Time is to adjust the clock one
+  hour two times a year (which is the incorrect way to do it).
+  To complicate things further, this user fetched your
+  software from Internet and has not considered what
+  the correct time is as perceived by a computer. The user does
+  not care about keeping the wall clock in sync with the correct
+  time. The user expects your program to have unlimited knowledge
   about the time.</p>
 
   <p>Most programmers also expect time to be reliable, at least until
-  they realize that the wall clock time on their workstation is of by
-  a minute. Then they simply set it to the correct time, maybe or
-  maybe not in a smooth way. Most probably not in a smooth way.</p>
+  they realize that the wall clock time on their workstation is off by
+  a minute. Then they set it to the correct time, but most probably
+  not in a smooth way.</p>
 
-  <p>The amount of problems that arise when you expect the wall clock
-  time on the system to always be correct may be immense. Therefore Erlang
+  <p>The number of problems that arise when you always expect the wall clock
+  time on the system to be correct can be immense. Erlang therefore
   introduced the "corrected estimate of time", or the "time
-  correction" many years ago. The time correction relies on the fact
+  correction", many years ago. The time correction relies on the fact
   that most operating systems have some kind of monotonic clock,
-  either a real time extension or some built in "tick counter" that is
-  independent of the wall clock settings. This counter may have
-  microsecond resolution or much less, but generally it has a drift
-  that is not to be ignored.</p>
-
+  either a real-time extension or some built-in "tick counter" that is
+  independent of the wall clock settings. This counter can have
+  microsecond resolution or much less, but it has a drift that cannot
+  be ignored.</p>
   </section>
 
-  <marker id="Time_Correction"/>
   <section>
+    <marker id="Time_Correction"/>
     <title>Time Correction</title>
     <p>If time correction is enabled, the Erlang runtime system
-    will make use of both
+    makes use of both
     <seealso marker="#OS_System_Time">OS system time</seealso>
     and <seealso marker="#OS_Monotonic_Time">OS monotonic time</seealso>,
-    in order to make adjustments of the frequency of the Erlang
-    monotonic clock. Time correction will ensure that
+    to adjust the frequency of the Erlang
+    monotonic clock. Time correction ensures that
     <seealso marker="#Erlang_Monotonic_Time">Erlang monotonic time</seealso>
-    will not warp, and that the frequency is relatively accurate.
-    The type of adjustments made to the frequency depends on the
-    time warp mode used. This will be discussed in more details in
-    the <seealso marker="#Time_Warp_Modes">time warp modes</seealso>
-    section below.</p>
-
-    <p>By default time correction will be enabled if support for
-    it on the specific platform exist. Support for it includes
-    both an OS monotonic time provided by the OS, and an
-    implementation in the Erlang runtime system utilizing the
-    OS monotonic time. You can check if your system has support
-    for OS monotonic time by calling
-    <seealso marker="erlang#system_info_os_monotonic_time_source"><c>erlang:system_info(os_monotonic_time_source)</c></seealso>,
-    and you can check if time correction is enabled on your
-    system by calling
+    does not warp and that the frequency is relatively accurate.
+    The type of frequency adjustments depends on the time warp mode used.
+    Section <seealso marker="#Time_Warp_Modes">Time Warp Modes</seealso>
+    provides more details.</p>
+
+    <p>By default time correction is enabled if support for
+    it exists on the specific platform. Support for it includes
+    both OS monotonic time, provided by the OS, and an
+    implementation in the Erlang runtime system using
+    OS monotonic time. To check if your system has support
+    for OS monotonic time, call
+    <seealso marker="erlang#system_info_os_monotonic_time_source"><c>erlang:system_info(os_monotonic_time_source)</c></seealso>.
+    To check if time correction is enabled on your system, call
     <seealso marker="erlang#system_info_time_correction"><c>erlang:system_info(time_correction)</c></seealso>.</p>
 
-    <p>Time correction is enabled or disabled by passing the
+    <p>To enable or disable time correction, pass command-line argument
     <seealso marker="erl#+c"><c>+c [true|false]</c></seealso>
-    command line argument to <c>erl</c>.</p>
+    to <c>erl</c>.</p>
 
     <p>If time correction is disabled, Erlang monotonic time
-    may warp forwards, it may stop and even freeze for extended
-    periods of time, and there are no guarantees that the frequency
+    may warp forwards or stop, or even freeze for extended
+    periods of time. There are then no guarantees that the frequency
     of the Erlang monotonic clock is accurate or stable.</p>
 
     <p><em>You typically never want to disable time correction</em>.
-    Previously there was a performance penalty associated with time
-    correction, but nowadays it is most often the other way around.
-    By disabling time correction you are likely to get bad scalability,
+    Previously a performance penalty was associated with time
+    correction, but nowadays it is usually the other way around.
+    If time correction is disabled, you probably get bad scalability,
     bad performance, and bad time measurements.</p>
   </section>
 
-
-  <marker id="Time_Warp_Safe_Code"/>
   <section>
+    <marker id="Time_Warp_Safe_Code"/>
     <title>Time Warp Safe Code</title>
-    <p>Time warp safe code is code that is able to handle
+    <p>Time warp safe code can handle
     a <seealso marker="#Time_Warp">time warp</seealso> of
-    <seealso marker="#Erlang_System_Time">Erlang system time</seealso>.
-    </p>
+    <seealso marker="#Erlang_System_Time">Erlang system time</seealso>.</p>
 
     <p><seealso marker="erlang#now/0"><c>erlang:now/0</c></seealso>
-    behaves very bad when Erlang system time warps. When Erlang
-    system time do a time warp backwards, the values returned
-    from <c>erlang:now/0</c> will freeze (if you disregard the
-    micro second increments made due to the actual call) until
-    OS system time reach the point of the last value returned by
-    <c>erlang:now/0</c>. This freeze might continue for very
-    long periods of time. It might take years, decades,
-    and even longer than this until the freeze stops.</p>
+    behaves bad when Erlang system time warps. When Erlang
+    system time does a time warp backwards, the values returned
+    from <c>erlang:now/0</c> freeze (if you disregard the
+    microsecond increments made because of the actual call) until
+    OS system time reaches the point of the last value returned by
+    <c>erlang:now/0</c>. This freeze can continue for a long time. It
+    can take years, decades, and even longer until the freeze stops.</p>
 
     <p>All uses of <c>erlang:now/0</c> are not necessarily
     time warp unsafe. If you do not use it to get time, it
-    will be time warp safe. However <em>all uses of
+    is time warp safe. However, <em>all uses of
     <c>erlang:now/0</c> are suboptimal</em> from a performance
     and scalability perspective. So you really want to replace
-    the usage of it with other functionality. For examples
-    of how to replace the usage of <c>erlang:now/0</c>,
-    see the <seealso marker="#Dos_and_Donts">Dos and Donts</seealso>
-    section.</p>
+    the use of it with other functionality. For examples
+    of how to replace the use of <c>erlang:now/0</c>, see Section
+    <seealso marker="#Dos_and_Donts">How to Work with the New
+    API</seealso>.</p>
   </section>
 
-  <marker id="Time_Warp_Modes"/>
   <section>
     <title>Time Warp Modes</title>
-
+    <marker id="Time_Warp_Modes"/>
     <p>Current <seealso marker="#Erlang_System_Time">Erlang system
     time</seealso> is determined by adding current
     <seealso marker="erlang#monotonic_time/0">Erlang monotonic time</seealso>
     with current
     <seealso marker="erlang#time_offset/0">time offset</seealso>. The
     time offset is managed differently depending on which time
-    warp mode you use. The time warp mode is set by passing the
+    warp mode you use.</p>
+
+    <p>To set the time warp mode, pass command-line argument
     <seealso marker="erl#+C_"><c>+C
     [no_time_warp|single_time_warp|multi_time_warp]</c></seealso>
-    command line argument to <c>erl</c>.</p>
+    to <c>erl</c>.</p>
 
     <marker id="No_Time_Warp_Mode"/>
     <section>
       <title>No Time Warp Mode</title>
       <p>The time offset is determined at runtime system start
-      and will after this not change. This is the default behavior.
-      Not because it is the best mode (which it isn't). It is
+      and does not change later. This is the default behavior, but
+      not because it is the best mode (which it is not). It is
       default <em>only</em> because this is how the runtime system
-      always has behaved up until ERTS version 7.0, and you have to
-      ensure that your Erlang code that may execute during a time
+      behaved until <c>ERTS</c> 7.0.
+      Ensure that your Erlang code that may execute during a time
       warp is <seealso marker="#Time_Warp_Safe_Code">time warp
-      safe</seealso> before you can enable other modes.</p>
+      safe</seealso> before enabling other modes.</p>
 
-      <p>Since the time offset is not allowed to change, time
-      correction needs to adjust the frequency of the Erlang
-      monotonic clock in order to smoothly align Erlang system
-      time with OS system time. A big downside of this approach
+      <p>As the time offset is not allowed to change, time
+      correction must adjust the frequency of the Erlang
+      monotonic clock to align Erlang system time with OS
+      system time smoothly. A significant downside of this approach
       is that we on purpose will use a faulty frequency on the
       Erlang monotonic clock if adjustments are needed. This
-      error may be as big as 1%. This error will show up in all
+      error can be as large as 1%. This error will show up in all
       time measurements in the runtime system.</p>
 
-      <p>If time correction is not enabled, the Erlang monotonic
-      time will freeze when the OS system time leap backwards.
-      The freeze of the monotonic time will continue until
-      OS system time catch up. The freeze may continue for
-      a very long time. When OS system time leaps forwards,
-      Erlang monotonic time will also leap forward.</p>
+      <p>If time correction is not enabled, Erlang monotonic
+      time freezes when OS system time leaps backwards.
+      The freeze of monotonic time continues until
+      OS system time catches up. The freeze can continue for
+      a long time. When OS system time leaps forwards,
+      Erlang monotonic time also leaps forward.</p>
     </section>
 
     <marker id="Single_Time_Warp_Mode"/>
@@ -411,26 +424,27 @@
       <title>Single Time Warp Mode</title>
       <p>This mode is more or less a backwards compatibility mode
       as of its introduction.</p>
+
       <p>On an embedded system it is not uncommon that the system
-      has no power supply at all, not even a battery, when it is
-      shut off. The system clock on such a system will typically
-      be way off when the system boots. If the
+      has no power supply, not even a battery, when it is
+      shut off. The system clock on such a system is typically
+      way off when the system boots. If
       <seealso marker="#No_Time_Warp_Mode">no time warp mode</seealso>
       is used, and the Erlang runtime system is started before
-      the OS system time has been corrected, the Erlang system
-      time may be wrong for a very long time, even centuries or
-      more.</p>
-      <p>If you for some reason need to use Erlang code that
-      is not
+      OS system time has been corrected, Erlang system time
+      can be wrong for a long time, centuries or even longer.</p>
+
+      <p>If you need to use Erlang code that is not
       <seealso marker="#Time_Warp_Safe_Code">time warp safe</seealso>,
-      and you need to start the Erlang runtime system before the OS
+      and you need to start the Erlang runtime system before OS
       system time has been corrected, you may want to use the single
-      time warp mode. Note that there are limitations to when you can
+      time warp mode.</p>
+
+      <note><p>There are limitations to when you can
       execute time warp unsafe code using this mode. If it is possible
-      to only utilize time warp safe code, it is <em>much</em> better
-      to use the <seealso marker="#Multi_Time_Warp_Mode">multi time
-      warp mode</seealso> instead.
-      </p>
+      to use time warp safe code only, it is <em>much</em> better
+      to use the <seealso marker="#Multi_Time_Warp_Mode">multi-time
+      warp mode</seealso> instead.</p></note>
 
       <p>Using the single time warp mode, the time offset is
       handled in two phases:</p>
@@ -438,158 +452,150 @@
       <taglist>
 	<tag>Preliminary Phase</tag>
 	<item>
-	  <p>The preliminary phase starts when the runtime
+	  <p>This phase starts when the runtime
 	  system starts. A preliminary time offset based on
-	  current OS system time is determined. This offset will
-	  from now on be fixed during the whole preliminary phase.</p>
+	  current OS system time is determined. This offset is from
+	  now on to be fixed during the whole preliminary phase.</p>
 
 	  <p>If time correction is enabled, adjustments to the
-	  Erlang monotonic clock will be made to keep its
-	  frequency as correct as possible, but <em>no</em>
-	  adjustments will be made trying to align Erlang system
-	  time and OS system time. That is, during the preliminary
-	  Erlang system time and OS system time might diverge
-	  from each other, and no attempt to prevent this will
-	  be made.</p>
+	  Erlang monotonic clock are made to keep its
+	  frequency as correct as possible. However, <em>no</em>
+	  adjustments are made trying to align Erlang system
+	  time and OS system time. That is, during the preliminary phase
+	  Erlang system time and OS system time can diverge
+	  from each other, and no attempt is made to prevent this.</p>
 
 	  <p>If time correction is disabled, changes in OS system
-	  time will effect the monotonic clock the same way as
+	  time affects the monotonic clock the same way as
 	  when the <seealso marker="#No_Time_Warp_Mode">no time warp
 	  mode</seealso> is used.</p>
 	</item>
 
 	<tag>Final Phase</tag>
 	<item>
-
-	  <p>The final phase begin when the user finalize the time
+	  <p>This phase begins when the user finalizes the time
 	  offset by calling
 	  <seealso marker="erlang#system_flag_time_offset"><c>erlang:system_flag(time_offset, finalize)</c></seealso>.
-	  The finalization can only be performed once.
-	  </p>
+	  The finalization can only be performed once.</p>
 
 	  <p>During finalization, the time offset is adjusted and
-	  fixated so that current Erlang system time align with
-	  current OS system time. Since the time offset may
-	  change during the finalization, the Erlang system time
-	  may do a time warp at this point. The time offset will
-	  from now on be fixed until the runtime system terminates.
+	  fixated so that current Erlang system time aligns with
+	  current OS system time. As the time offset can
+	  change during the finalization, Erlang system time
+	  can do a time warp at this point. The time offset is
+	  from now on fixed until the runtime system terminates.
 	  If time correction has been enabled, the time
-	  correction will from now on also make adjustments
-	  in order to align Erlang system time with OS system
-	  time. When the system is in the final phase it behaves
+	  correction from now on also makes adjustments
+	  to align Erlang system time with OS system
+	  time. When the system is in the final phase, it behaves
 	  exactly as in the <seealso marker="#No_Time_Warp_Mode">no
 	  time warp mode</seealso>.</p>
-
 	</item>
       </taglist>
 
-      <p>In order for this to work properly there are two
-      requirements that the user needs to ensure are
-      satisfied:</p>
+      <p>In order for this to work properly, the user must ensure
+      that the following two requirements are satisfied:</p>
 
       <taglist>
 	<tag>Forward Time Warp</tag>
 	<item><p>The time warp made when finalizing the time offset
 	can only be done forwards without encountering problems.
-	This implies that the user has to ensure that the OS
+	This implies that the user must ensure that OS
 	system time is set to a time earlier or equal to actual
-	POSIX time before starting the Erlang runtime system. If
-	you are not completely sure the OS system time is correct,
+	POSIX time before starting the Erlang runtime system.</p>
+
+	<p>If you are not sure that OS system time is correct,
 	set it to a time that is guaranteed to be earlier than
 	actual POSIX time before starting the Erlang runtime
-	system just to be safe.</p></item>
+	system, just to be safe.</p>
+        </item>
 
 	<tag>Finalize Correct OS System Time</tag>
-	<item><p>The OS system time needs to be correct when the
-	the user finalizes the time offset.</p></item>
+	<item><p>OS system time must be correct when
+	the user finalizes the time offset.</p>
+        </item>
       </taglist>
 
       <p>If these requirements are not fulfilled, the system
-      may behave very bad.
-      </p>
-
-      <p>Assuming that the requirements above are fulfilled,
-      time correction is enabled, and that the OS system time
-      is adjusted using some time adjustment protocol like NTP
-      or similar, only small adjustments of the Erlang monotonic
-      time should be needed in order to keep system times
-      aligned after finilization. As long as the system is not
-      suspended, the largest adjustments needed should be for
+      may behave very bad.</p>
+
+      <p>Assuming that these requirements are fulfilled,
+      time correction is enabled, and that OS system time
+      is adjusted using a time adjustment protocol such as NTP,
+      only small adjustments of Erlang monotonic
+      time are needed to keep system times
+      aligned after finalization. As long as the system is not
+      suspended, the largest adjustments needed are for
       inserted (or deleted) leap seconds.</p>
 
-      <warning><p>In order to be able to use this mode you have
-      to ensure that all Erlang code that will execute in
-      both phases are
+      <warning><p>To use this mode, ensure that
+      all Erlang code that will execute in both phases are
       <seealso marker="#Time_Warp_Safe_Code">time warp
       safe</seealso>.</p>
-      <p>Code that only execute in the final phase does not have
+      <p>Code executing only in the final phase does not have
       to be able to cope with the time warp.</p></warning>
-
     </section>
 
     <marker id="Multi_Time_Warp_Mode"/>
     <section>
-      <title>Multi Time Warp Mode</title>
-
-      <p><em>Multi time warp mode in combination with time
-      correction is the preferred configuration</em>. This since,
-      on almost all platforms, the Erlang runtime system will have
-      better performance, will scale better, will behave better,
-      and since the accuracy, and precision of time measurements
-      will be better. Only Erlang runtime systems executing on
-      ancient platforms will benefit from another configuration.</p>
+      <title>Multi-Time Warp Mode</title>
+      <p><em>Multi-time warp mode in combination with time
+      correction is the preferred configuration</em>. This as
+      the Erlang runtime system have better performance, scale
+      better, and behave better on almost all platforms. In
+      addition, the accuracy and precision of time measurements
+      are better. Only Erlang runtime systems executing on
+      ancient platforms benefit from another configuration.</p>
 
       <p>The time offset may change at any time without limitations.
       That is, Erlang system time may perform time warps both
-      forwards and backwards at <em>any</em> time. Since we align
-      the Erlang system time with the OS system time by changing
+      forwards and backwards at <em>any</em> time. As we align
+      Erlang system time with OS system time by changing
       the time offset, we can enable a time correction that tries
       to adjust the frequency of the Erlang monotonic clock to be as
-      correct as possible. This will make time measurements using
-      the Erlang monotonic time more accurate and precise.</p>
+      correct as possible. This makes time measurements using
+      Erlang monotonic time more accurate and precise.</p>
 
       <p>If time correction is disabled, Erlang monotonic time
-      will leap forward if OS system time leaps forward. If the
-      OS system time leaps backwards, Erlang monotonic time will
-      stop briefly but it does not freeze for extended periods
-      of time. This since the time offset is changed in order to
+      leaps forward if OS system time leaps forward. If
+      OS system time leaps backwards, Erlang monotonic time
+      stops briefly, but it does not freeze for extended periods
+      of time. This as the time offset is changed to
       align Erlang system time with OS system time.</p>
 
-      <warning><p>In order to be able to use this mode you have
-      to ensure that all Erlang code that will execute on the
-      runtime system is
+      <warning><p>To use this mode, ensure that all
+      Erlang code that will execute on the runtime system is
       <seealso marker="#Time_Warp_Safe_Code">time warp
       safe</seealso>.</p></warning>
     </section>
   </section>
 
-  <marker id="The_New_Time_API"/>
   <section>
-    <title>The New Time API</title>
-
+    <title>New Time API</title>
+    <marker id="The_New_Time_API"/>
     <p>The old time API is based on
     <seealso marker="erlang#now/0"><c>erlang:now/0</c></seealso>.
-    The major issue with <c>erlang:now/0</c> is that it was
-    intended to be used for so many unrelated things. This
-    tied these unrelated operations together and unnecessarily
-    caused performance, scalability as well as accuracy, and
-    precision issues for operations that do not need to have
-    such issues. The new API spreads different functionality
-    over multiple functions in order to improve on this.</p>
-
-    <p>In order to be backwards compatible <c>erlang:now/0</c> will
-    remain as is, but <em>you are strongly discouraged from using
-    it</em>. A lot of uses of <c>erlang:now/0</c> will also
-    prevent you from using the new
-    <seealso marker="#Multi_Time_Warp_Mode">multi time warp
-    mode</seealso> which is an important part of this
+    <c>erlang:now/0</c> was intended to be used for many unrelated
+    things. This tied these unrelated operations together and
+    caused issues with performance, scalability, accuracy, and
+    precision for operations that did not need to have
+    such issues. To improve this, the new API spreads different
+    functionality over multiple functions.</p>
+
+    <p>To be backwards compatible, <c>erlang:now/0</c>
+    remains as is, but <em>you are strongly discouraged from using
+    it</em>. Many use cases of <c>erlang:now/0</c>
+    prevents you from using the new
+    <seealso marker="#Multi_Time_Warp_Mode">multi-time warp
+    mode</seealso>, which is an important part of this
     new time functionality improvement.</p>
 
     <p>Some of the new BIFs on some systems, perhaps surprisingly,
-    return negative integer values on a newly started run time
-    system. This is not a bug, but a memory usage optimization.</p>
+    return negative integer values on a newly started runtime
+    system. This is not a bug, but a memory use optimization.</p>
+
+    <p>The new API consists of the following new BIFs:</p>
 
-    <p>The new API consists of a number of new BIFs:</p>
     <list>
       <item><p><seealso marker="erlang#convert_time_unit/3"><c>erlang:convert_time_unit/3</c></seealso></p></item>
       <item><p><seealso marker="erlang#monotonic_time/0"><c>erlang:monotonic_time/0</c></seealso></p></item>
@@ -604,7 +610,9 @@
       <item><p><seealso marker="kernel:os#system_time/0"><c>os:system_time/0</c></seealso></p></item>
       <item><p><seealso marker="kernel:os#system_time/1"><c>os:system_time/1</c></seealso></p></item>
     </list>
-    <p>and a number of extensions of existing BIFs:</p>
+
+    <p>The new API also consists of extensions of the following existing BIFs:</p>
+
     <list>
       <item><p><seealso marker="erlang#monitor/2"><c>erlang:monitor(time_offset, clock_service)</c></seealso></p></item>
       <item><p><seealso marker="erlang#system_flag_time_offset"><c>erlang:system_flag(time_offset, finalize)</c></seealso></p></item>
@@ -619,102 +627,99 @@
 
     <marker id="The_New_Erlang_Monotonic_Time"/>
     <section>
-      <title>The New Erlang Monotonic Time</title>
-      <p>The Erlang monotonic time as such is new as of ERTS
-      version 7.0. It has been introduced in order to be able
-      to detach time measurements such as elapsed time from
-      calender time. It is very common that one is interested
-      in measuring elapsed time or specifying a time relative
-      to another point in time without having any need to know
-      what the involved times are in UTC or any other
-      globally defined time scale. By introducing a time scale
-      that has a local definition of where it starts, it is
-      possible to manage time that do not concern calender
-      time on that time scale. Erlang monotonic time use
-      such a time scale with a locally defined start.</p>
-
-      <p>The introduction of Erlang monotonic time gives us
-      the possibility to adjust the two Erlang times (Erlang
+      <title>New Erlang Monotonic Time</title>
+      <p>Erlang monotonic time as such is new as of <c>ERTS</c> 7.0.
+      It is introduced to detach time measurements, such as elapsed
+      time from calendar time. In many use cases there is a need to
+      measure elapsed time or specify a time relative to another point
+      in time without the need to know the involved times in UTC or
+      any other globally defined time scale. By introducing a time
+      scale with a local definition of where it starts, time that do
+      not concern calendar time can be managed on that time
+      scale. Erlang monotonic time uses such a time scale with a
+      locally defined start.</p>
+
+      <p>The introduction of Erlang monotonic time allows
+      us to adjust the two Erlang times (Erlang
       monotonic time and Erlang system time) separately. By
-      doing this, accuracy of elapsed time does not have to
+      doing this, the accuracy of elapsed time does not have to
       suffer just because the system time happened to be
       wrong at some point in time. Separate adjustments
       of the two times are only performed in the time warp
       modes, and only fully separated in the
-      <seealso marker="#Multi_Time_Warp_Mode">multi
-      time warp mode</seealso>. All other modes than the
-      multi time warp mode are there for backwards
-      compatibility reasons, and when using these the
-      accuracy of Erlang monotonic time suffer since
+      <seealso marker="#Multi_Time_Warp_Mode">multi-time
+      warp mode</seealso>. All other modes than the
+      multi-time warp mode are for backwards
+      compatibility reasons. When using these modes, the
+      accuracy of Erlang monotonic time suffer, as
       the adjustments of Erlang monotonic time in these
-      modes are more or less tied to the Erlang system
-      time.</p>
+      modes are more or less tied to Erlang system time.</p>
 
       <p>The adjustment of system time could have been made
       smother than using a time warp approach, but we think
-      that would be a bad choice. Since we are able to
-      express and measure time that aren't connected to
-      calender time by the use of Erlang monotonic time, it
+      that would be a bad choice. As we can
+      express and measure time that is not connected to
+      calendar time by the use of Erlang monotonic time, it
       is better to expose the change in Erlang system time
-      immediately. This since it makes it possible for the
-      Erlang applications executing on the system to react
-      on the change in system time as soon as possible. This
-      is also more or less exactly how most OSes handle this
+      immediately. This as the Erlang applications
+      executing on the system can react on the change in
+      system time as soon as possible. This is also more or
+      less exactly how most operating systems handle this
       (OS monotonic time and OS system time). By adjusting
-      system time smoothly we would just hide the fact that
+      system time smoothly, we would just hide the fact that
       system time changed and make it harder for the Erlang
       applications to react to the change in a sensible way.</p>
 
-      <p>In order to be able to react to a change in Erlang
-      system time you have to be able to detect that it
+      <p>To be able to react to a change in Erlang
+      system time, you must be able to detect that it
       happened. The change in Erlang system time occurs when
       current time offset is changed. We have therefore
-      introduced the possibility to monitor the time offset
-      using
-      <seealso marker="erlang#monitor/2"><c>erlang:monitor(time_offset, clock_service)</c></seealso>. A process monitoring the time
-      offset will be sent a message on the following format
+      introduced the possibility to monitor the time offset using
+      <seealso marker="erlang#monitor/2"><c>erlang:monitor(time_offset, clock_service)</c></seealso>.
+      A process monitoring the time
+      offset is sent a message on the following format
       when the time offset is changed:</p>
+
       <code type="none">{'CHANGE', MonitorReference, time_offset, clock_service, NewTimeOffset}</code>
     </section>
 
     <marker id="Unique_Values"/>
     <section>
       <title>Unique Values</title>
-      <p>Besides reporting time <c>erlang:now/0</c> also
-      produce unique and strictly monotonically increasing
-      values. In order to detach this functionality from
-      time measurements we have introduced
+      <p>Besides reporting time, <c>erlang:now/0</c> also
+      produces unique and strictly monotonically increasing
+      values. To detach this functionality from
+      time measurements, we have introduced
       <seealso marker="erlang#unique_integer/1"><c>erlang:unique_integer()</c></seealso>.
       </p>
     </section>
 
     <marker id="Dos_and_Donts"/>
     <section>
-      <title>Dos and Don'ts</title>
+      <title>How to Work with the New API</title>
       <p>Previously <c>erlang:now/0</c> was the only option for doing
-      quite a lot of things. We will look at a few different things
-      <c>erlang:now/0</c> could be used for, and how you want to do
-      this using the new API:</p>
+      many things. This section deals with some things that
+      <c>erlang:now/0</c> can be used for, and how you are to
+      these using the new API.</p>
 
       <marker id="Dos_and_Donts_Retrieve_Erlang_System_Time"/>
       <section>
 	<title>Retrieve Erlang System Time</title>
 	<dont>
 	  <p>
-	    use <c>erlang:now/0</c> in order to retrieve current Erlang
-	    system time.
+	    Use <c>erlang:now/0</c> to retrieve current Erlang system time.
 	  </p>
 	</dont>
 	<do>
 	  <p>
-	    use
+	    Use
 	    <seealso marker="erlang#system_time/1"><c>erlang:system_time/1</c></seealso>
-	    in order to retrieve current Erlang system time on the
+	    to retrieve current Erlang system time on the
 	    <seealso marker="erlang#type_time_unit">time unit</seealso>
 	    of your choice.</p>
 	    <p>If you want the same format as returned by <c>erlang:now/0</c>, use
 	    <seealso marker="erlang#timestamp/0"><c>erlang:timestamp/0</c></seealso>.
-	    </p>
+	   </p>
 	</do>
       </section>
 
@@ -723,26 +728,27 @@
 	<title>Measure Elapsed Time</title>
 	<dont>
 	  <p>
-	    take timestamps with <c>erlang:now/0</c> and calculate
+	    Take timestamps with <c>erlang:now/0</c> and calculate
 	    the difference in time with
 	    <seealso marker="stdlib:timer#now_diff/2"><c>timer:now_diff/2</c></seealso>.
 	  </p>
 	</dont>
 	<do>
 	  <p>
-	    take timestamps with
+	    Take timestamps with
 	    <seealso marker="erlang#monotonic_time/0"><c>erlang:monotonic_time/0</c></seealso>
 	    and calculate the time difference using ordinary subtraction.
 	    The result will be in <c>native</c>
 	    <seealso marker="erlang#type_time_unit">time unit</seealso>.
 	    If you want to convert the
-	    result to another time unit you can do this using
+	    result to another time unit, you can use
 	    <seealso marker="erlang#convert_time_unit/3"><c>erlang:convert_time_unit/3</c></seealso>.
 	    </p>
-	    <p>Another easier way of doing this is to use
+
+	    <p>An easier way to do this is to use
 	    <seealso marker="erlang#monotonic_time/1"><c>erlang:monotonic_time/1</c></seealso>
-	    with desired time unit. However, you may lose accuracy,
-	    and precision this way.
+	    with the desired time unit. However, you can then lose accuracy
+	    and precision.
 	    </p>
 	</do>
       </section>
@@ -752,16 +758,16 @@
 	<title>Determine Order of Events</title>
 	<dont>
 	  <p>
-	    determine the order of events by saving a timestamp
-	    with <c>erlang:now/0</c> when the event happens.
+	    Determine the order of events by saving a timestamp
+	    with <c>erlang:now/0</c> when the event occurs.
 	  </p>
 	</dont>
 	<do>
 	  <p>
-	    determine the order of events by saving the integer
+	    Determine the order of events by saving the integer
 	    returned by
 	    <seealso marker="erlang#unique_integer/1"><c>erlang:unique_integer([monotonic])</c></seealso>
-	    when the event happens. These integers will be strictly
+	    when the event occurs. These integers will be strictly
 	    monotonically ordered on current runtime system instance
 	    corresponding to creation time.
 	  </p>
@@ -770,40 +776,43 @@
 
       <marker id="Dos_and_Donts_Determine_Order_of_Events_With_Time_of_the_Event"/>
       <section>
-	<title>Determine Order of Events With Time of the Event</title>
+	<title>Determine Order of Events with Time of the Event</title>
 	<dont>
 	  <p>
-	    determine the order of events by saving a timestamp
-	    with <c>erlang:now/0</c> when the event happens.
+	    Determine the order of events by saving a timestamp
+	    with <c>erlang:now/0</c> when the event occurs.
 	  </p>
 	</dont>
 	<do>
 	  <p>
-	    determine the order of events by saving a tuple
-	    containing
+	    Determine the order of events by saving a tuple containing
 	    <seealso marker="erlang#monotonic_time/0">monotonic time</seealso>
 	    and a <seealso marker="erlang#unique_integer/1">strictly
-	    monotonically increasing integer</seealso> like this:</p>
+	    monotonically increasing integer</seealso> as follows:</p>
+
 	    <code type="none">
 Time = erlang:monotonic_time(),
 UMI = erlang:unique_integer([monotonic]),
 EventTag = {Time, UMI}</code>
+
 	    <p>These tuples will be strictly monotonically ordered
-	    on the current runtime system instance according to
-	    creation time. Note that it is important that the
+	    on current runtime system instance according to
+	    creation time. It is important that the
 	    monotonic time is in the first element (the most
 	    significant element when comparing 2-tuples). Using
 	    the monotonic time in the tuples, you can calculate time
 	    between events.</p>
-	    <p>If you are interested in the Erlang system time at the
-	    time when the event occurred you can also save the time
+
+	    <p>If you are interested in Erlang system time at the
+	    time when the event occurred, you can also save the time
 	    offset before or after saving the events using
 	    <seealso marker="erlang#time_offset/0"><c>erlang:time_offset/0</c></seealso>.
 	    Erlang monotonic time added with the time
 	    offset corresponds to Erlang system time.</p>
+
 	    <p>If you are executing in a mode where time offset
-	    may change and you want to be able to get the actual
-	    Erlang system time when the event occurred you can
+	    can change, and you want to get the actual
+	    Erlang system time when the event occurred, you can
 	    save the time offset as a third element in the tuple
 	    (the least significant element when comparing 3-tuples).</p>
 	</do>
@@ -814,16 +823,15 @@ EventTag = {Time, UMI}</code>
 	<title>Create a Unique Name</title>
 	<dont>
 	  <p>
-	    use the values returned from <c>erlang:now/0</c>
-	    in order to create a name unique on the current
-	    runtime system instance.
+	    Use the values returned from <c>erlang:now/0</c>
+	    to create a name unique on the current runtime system instance.
 	  </p>
 	</dont>
 	<do>
 	  <p>
-	    use the value returned from
+	    Use the value returned from
 	    <seealso marker="erlang#unique_integer/0"><c>erlang:unique_integer/0</c></seealso>
-	    in order to create a name unique on the current runtime system
+	    to create a name unique on the current runtime system
 	    instance. If you only want positive integers, you can use
 	    <seealso marker="erlang#unique_integer/1"><c>erlang:unique_integer([positive])</c></seealso>.
 	  </p>
@@ -832,49 +840,63 @@ EventTag = {Time, UMI}</code>
 
       <marker id="Dos_and_Donts_Seed_Random_Number_Generation_With_a_Unique_Value"/>
       <section>
-	<title>Seed Random Number Generation With a Unique Value</title>
+	<title>Seed Random Number Generation with a Unique Value</title>
 	<dont>
 	  <p>
-	    seed random number generation using <c>erlang:now()</c>.
+	    Seed random number generation using <c>erlang:now()</c>.
 	  </p>
 	</dont>
 	<do>
 	  <p>
-	    seed random number generation using a combination of
+	    Seed random number generation using a combination of
 	    <seealso marker="erlang#monotonic_time/0"><c>erlang:monotonic_time()</c></seealso>,
 	    <seealso marker="erlang#time_offset/0"><c>erlang:time_offset()</c></seealso>,
-	    <seealso marker="erlang#unique_integer/0"><c>erlang:unique_integer()</c></seealso>, and other functionality.
+	    <seealso marker="erlang#unique_integer/0"><c>erlang:unique_integer()</c></seealso>,
+            and other functionality.
 	  </p>
 	  </do>
       </section>
 
-      <p>To sum this section up: <em>Don't use <c>erlang:now/0</c>!</em></p>
+      <p>To sum up this section: <em>Do not use <c>erlang:now/0</c>.</em></p>
     </section>
   </section>
 
-  <marker id="Supporting_Both_New_and_Old_OTP_Releases"/>
   <section>
-    <title>Supporting Both New and Old OTP Releases</title>
-    <p>Your code may be required to be able to run on a variety 
+    <marker id="Supporting_Both_New_and_Old_OTP_Releases"/>
+    <title>Support of Both New and Old OTP Releases</title>
+    <p>It can be required that your code must run on a variety
     of OTP installations of different OTP releases. If so, you
-    can not just use the new API out of the box, since it will
+    cannot use the new API out of the box, as it will
     not be available on old pre OTP 18 releases. The solution
-    is <em>not</em> to avoid using the new API, since your
-    code then won't be able to benefit from the scalability
-    and accuracy improvements made. Instead you want to use the
+    is <em>not</em> to avoid using the new API, as your
+    code then would not benefit from the scalability
+    and accuracy improvements made. Instead, use the
     new API when available, and fall back on <c>erlang:now/0</c>
-    when it is not available. Fortunately almost all of the new
-    API can easily be implemented using existing primitives
-    (except for
-    <seealso marker="erlang#system_info_start_time"><c>erlang:system_info(start_time)</c></seealso>,
-    <seealso marker="erlang#system_info_end_time"><c>erlang:system_info(end_time)</c></seealso>,
-    <seealso marker="erlang#system_info_os_monotonic_time_source"><c>erlang:system_info(os_monotonic_time_source)</c></seealso>, and
-    <seealso marker="erlang#system_info_os_system_time_source"><c>erlang:system_info(os_system_time_source)</c></seealso>).
-    By wrapping the API with functions that fall back on
-    <c>erlang:now/0</c> when the new API is not available,
-    and using these wrappers instead of using the API directly
-    the problem is solved. These wrappers can for example
+    when the new API is unavailable.</p>
+
+    <p>Fortunately most of the new API can easily be
+    implemented using existing primitives, except for:</p>
+
+    <list type="bulleted">
+      <item>
+        <seealso marker="erlang#system_info_start_time"><c>erlang:system_info(start_time)</c></seealso>
+      </item>
+      <item>
+        <seealso marker="erlang#system_info_end_time"><c>erlang:system_info(end_time)</c></seealso>
+      </item>
+      <item>
+        <seealso marker="erlang#system_info_os_monotonic_time_source"><c>erlang:system_info(os_monotonic_time_source)</c></seealso>
+      </item>
+      <item>
+        <seealso marker="erlang#system_info_os_system_time_source"><c>erlang:system_info(os_system_time_source)</c></seealso>)
+      </item>
+    </list>
+
+    <p>By wrapping the API with functions that fall back on
+    <c>erlang:now/0</c> when the new API is unavailable,
+    and using these wrappers instead of using the API directly,
+    the problem is solved. These wrappers can, for example,
     be implemented as in
-    <url href="time_compat.erl"><c>$ERL_TOP/erts/example/time_compat.erl</c></url>.</p>
+    <url href="time_compat.erl">$ERL_TOP/erts/example/time_compat.erl</url>.</p>
   </section>
 </chapter>
diff --git a/erts/emulator/Makefile.in b/erts/emulator/Makefile.in
index a919f0e3ac..f4b806fae9 100644
--- a/erts/emulator/Makefile.in
+++ b/erts/emulator/Makefile.in
@@ -23,10 +23,6 @@ include ../vsn.mk
 include $(ERL_TOP)/make/$(TARGET)/otp.mk
 -include $(TARGET)/gen_git_version.mk
 
-ifeq ($(findstring ose,$(TARGET)),ose)
-include $(ERL_TOP)/make/$(TARGET)/ose_lm.mk
-endif
-
 ENABLE_ALLOC_TYPE_VARS = @ENABLE_ALLOC_TYPE_VARS@
 HIPE_ENABLED=@HIPE_ENABLED@
 DTRACE_ENABLED=@DTRACE_ENABLED@
@@ -245,9 +241,7 @@ HCC     = @HCC@
 LD      = @LD@
 DEXPORT = @DEXPORT@
 RANLIB  = @RANLIB@
-ifneq ($(findstring ose,$(TARGET)),ose)
 STRIP   = strip
-endif
 PERL    = @PERL@
 RM	= @RM@
 MKDIR	= @MKDIR@
@@ -323,7 +317,7 @@ else
 CS_CFLAGS = $(CS_CFLAGS_)
 endif
 CS_LDFLAGS = $(LDFLAGS)
-CS_LIBS = -L../lib/internal/$(TARGET) -lerts_internal$(TYPEMARKER) @ERTS_INTERNAL_X_LIBS@
+CS_LIBS = -L../lib/internal/$(TARGET) -lerts_internal$(TYPEMARKER) @ERTS_INTERNAL_X_LIBS@ @SOCKET_LIBS@
 
 LIBS	+= @TERMCAP_LIB@ -L../lib/internal/$(TARGET) @ERTS_INTERNAL_X_LIBS@
 
@@ -404,7 +398,7 @@ EMULATOR_EXECUTABLE = beam$(TF_MARKER).dll
 else
 EMULATOR_EXECUTABLE = beam$(TF_MARKER)
 endif
-CS_EXECUTABLE = child_setup$(TYPEMARKER)
+CS_EXECUTABLE = erl_child_setup$(TYPEMARKER)
 
 # ----------------------------------------------------------------------
 
@@ -597,6 +591,7 @@ ifeq ($(TARGET),win32)
 PRELOAD_OBJ = $(OBJDIR)/beams.$(RES_EXT)
 PRELOAD_SRC = $(TARGET)/beams.rc
 $(PRELOAD_SRC): $(ERL_TOP)/erts/preloaded/ebin/otp_ring0.beam \
+			$(ERL_TOP)/erts/preloaded/ebin/erts_code_purger.beam \
 			$(ERL_TOP)/erts/preloaded/ebin/init.beam \
 			$(ERL_TOP)/erts/preloaded/ebin/prim_eval.beam \
 			$(ERL_TOP)/erts/preloaded/ebin/prim_inet.beam \
@@ -606,11 +601,13 @@ $(PRELOAD_SRC): $(ERL_TOP)/erts/preloaded/ebin/otp_ring0.beam \
 			$(ERL_TOP)/erts/preloaded/ebin/erl_prim_loader.beam \
 			$(ERL_TOP)/erts/preloaded/ebin/erlang.beam \
 			$(ERL_TOP)/erts/preloaded/ebin/erts_internal.beam
+
 	$(gen_verbose)LANG=C $(PERL) utils/make_preload $(MAKE_PRELOAD_EXTRA) -rc $^ > $@
 else
 PRELOAD_OBJ = $(OBJDIR)/preload.o
 PRELOAD_SRC = $(TARGET)/preload.c
 $(PRELOAD_SRC): $(ERL_TOP)/erts/preloaded/ebin/otp_ring0.beam \
+			$(ERL_TOP)/erts/preloaded/ebin/erts_code_purger.beam \
 			$(ERL_TOP)/erts/preloaded/ebin/init.beam \
 			$(ERL_TOP)/erts/preloaded/ebin/prim_eval.beam \
 			$(ERL_TOP)/erts/preloaded/ebin/prim_inet.beam \
@@ -684,14 +681,6 @@ $(OBJDIR)/%.o: $(TTF_DIR)/%.c
 $(OBJDIR)/%.o: sys/$(ERLANG_OSTYPE)/%.c
 	$(V_CC) $(CFLAGS) $(INCLUDES) -c $< -o $@
 
-ifeq ($(findstring ose,$(TARGET)),ose)
-$(OBJDIR)/ose_confd.o: $(OSE_CONFD)
-	$(V_CC) $(CFLAGS) $(INCLUDES) -c $< -o $@
-
-$(OBJDIR)/crt0_lm.o: $(CRT0_LM)
-	$(V_CC) $(CFLAGS) $(INCLUDES) -c $< -o $@
-endif
-
 $(OBJDIR)/%.o: sys/common/%.c
 	$(V_CC) $(subst -O2, $(GEN_OPT_FLGS), $(CFLAGS)) $(INCLUDES) -c $< -o $@
 
@@ -704,11 +693,11 @@ $(OBJDIR)/%.o: drivers/$(ERLANG_OSTYPE)/%.c
 # ----------------------------------------------------------------------
 # Specials
 #
-CS_SRC = sys/$(ERLANG_OSTYPE)/erl_child_setup.c
+CS_OBJ = $(OBJDIR)/erl_child_setup.o $(OBJDIR)/sys_uds.o $(OBJDIR)/hash.o
 
-$(BINDIR)/$(CS_EXECUTABLE): $(TTF_DIR)/GENERATED $(PRELOAD_SRC) $(CS_SRC) $(ERTS_LIB)
-	$(ld_verbose)$(CS_PURIFY) $(CC) $(CS_LDFLAGS) -o $(BINDIR)/$(CS_EXECUTABLE) \
-               $(CS_CFLAGS) $(COMMON_INCLUDES) $(CS_SRC) $(CS_LIBS)
+$(BINDIR)/$(CS_EXECUTABLE): $(TTF_DIR)/GENERATED $(PRELOAD_SRC) $(CS_OBJ) $(ERTS_LIB)
+	$(ld_verbose)$(CS_PURIFY) $(LD) $(CS_LDFLAGS) -o $(BINDIR)/$(CS_EXECUTABLE) \
+               $(CS_CFLAGS) $(COMMON_INCLUDES) $(CS_OBJ) $(CS_LIBS)
 
 $(OBJDIR)/%.kp.o: sys/common/%.c
 	$(V_CC) -DERTS_KERNEL_POLL_VERSION $(subst -O2, $(GEN_OPT_FLGS), $(CFLAGS)) $(INCLUDES) -c $< -o $@
@@ -810,31 +799,10 @@ OS_OBJS = \
 	$(OBJDIR)/dosmap.o
 
 else
-ifeq ($(findstring ose,$(TARGET)),ose)
-OS_OBJS = \
-	$(OBJDIR)/sys.o \
-	$(OBJDIR)/driver_tab.o \
-	$(OBJDIR)/ose_efile.o \
-	$(OBJDIR)/gzio.o \
-	$(OBJDIR)/elib_memmove.o
-
-OS_OBJS += $(OBJDIR)/ose_confd.o \
-	$(OBJDIR)/crt0_lm.o
-
-OS_OBJS += $(OBJDIR)/sys_float.o \
-					 $(OBJDIR)/sys_time.o
-
-DRV_OBJS = \
-	$(OBJDIR)/efile_drv.o \
-	$(OBJDIR)/ose_signal_drv.o \
-	$(OBJDIR)/inet_drv.o \
-	$(OBJDIR)/zlib_drv.o \
-	$(OBJDIR)/ram_file_drv.o \
-	$(OBJDIR)/ttsl_drv.o
-
-else
 OS_OBJS = \
 	$(OBJDIR)/sys.o \
+	$(OBJDIR)/sys_drivers.o \
+	$(OBJDIR)/sys_uds.o \
 	$(OBJDIR)/driver_tab.o \
 	$(OBJDIR)/unix_efile.o \
 	$(OBJDIR)/gzio.o \
@@ -849,7 +817,6 @@ DRV_OBJS = \
 	$(OBJDIR)/ram_file_drv.o \
 	$(OBJDIR)/ttsl_drv.o
 endif
-endif
 
 ifneq ($(STATIC_NIFS),no)
 STATIC_NIF_LIBS = $(STATIC_NIFS)
@@ -1022,19 +989,12 @@ $(BINDIR)/$(EMULATOR_EXECUTABLE): $(INIT_OBJS) $(OBJS) $(DEPLIBS)
 	$(STATIC_DRIVER_LIBS) $(LIBS)
 
 else
-ifeq ($(findstring ose,$(TARGET)),ose)
-$(BINDIR)/$(EMULATOR_EXECUTABLE): $(INIT_OBJS) $(OBJS) $(DEPLIBS) $(LCF)
-	$(call build-ose-load-module, $@, $(INIT_OBJS) $(OBJS), $(STATIC_NIF_LIBS) \
-	$(STATIC_DRIVER_LIBS) $(LIBS), $(BEAM_LMCONF))
-
-else
 $(BINDIR)/$(EMULATOR_EXECUTABLE): $(INIT_OBJS) $(OBJS) $(DEPLIBS)
 	$(ld_verbose)$(PURIFY) $(LD) -o $(BINDIR)/$(EMULATOR_EXECUTABLE) \
 	$(HIPEBEAMLDFLAGS) $(LDFLAGS) $(DEXPORT) $(INIT_OBJS) $(OBJS) \
 	$(STATIC_NIF_LIBS) $(STATIC_DRIVER_LIBS) $(LIBS)
 
 endif
-endif
 
 # ----------------------------------------------------------------------
 # Dependencies
diff --git a/erts/emulator/beam/atom.c b/erts/emulator/beam/atom.c
index fe91134ef4..fd2adac676 100644
--- a/erts/emulator/beam/atom.c
+++ b/erts/emulator/beam/atom.c
@@ -435,6 +435,9 @@ init_atom_table(void)
     f.cmp  = (HCMP_FUN) atom_cmp;
     f.alloc = (HALLOC_FUN) atom_alloc;
     f.free = (HFREE_FUN) atom_free;
+    f.meta_alloc = (HMALLOC_FUN) erts_alloc;
+    f.meta_free = (HMFREE_FUN) erts_free;
+    f.meta_print = (HMPRINT_FUN) erts_print;
 
     atom_text_pos = NULL;
     atom_text_end = NULL;
diff --git a/erts/emulator/beam/atom.h b/erts/emulator/beam/atom.h
index ead56c83d8..2c002ca92f 100644
--- a/erts/emulator/beam/atom.h
+++ b/erts/emulator/beam/atom.h
@@ -129,6 +129,7 @@ typedef enum {
   (erts_is_atom_utf8_bytes((byte *) LSTR, sizeof(LSTR) - 1, (TERM)))
 #define ERTS_DECL_AM(S) Eterm AM_ ## S = am_atom_put(#S, sizeof(#S) - 1)
 #define ERTS_INIT_AM(S) AM_ ## S = am_atom_put(#S, sizeof(#S) - 1)
+#define ERTS_MAKE_AM(Str) am_atom_put(Str, sizeof(Str) - 1)
 
 int atom_table_size(void);	/* number of elements */
 int atom_table_sz(void);	/* table size in bytes, excluding stored objects */
diff --git a/erts/emulator/beam/atom.names b/erts/emulator/beam/atom.names
index f9a2f3e33e..ae7f405cce 100644
--- a/erts/emulator/beam/atom.names
+++ b/erts/emulator/beam/atom.names
@@ -71,6 +71,7 @@ atom absoluteURI
 atom ac
 atom accessor
 atom active
+atom active_tasks
 atom all
 atom all_but_first
 atom all_names
@@ -117,10 +118,9 @@ atom bif_timer_server
 atom binary
 atom binary_bin_to_list_trap
 atom binary_copy_trap
+atom binary_find_trap
 atom binary_longest_prefix_trap
 atom binary_longest_suffix_trap
-atom binary_match_trap
-atom binary_matches_trap
 atom binary_to_list_continue
 atom binary_to_term_trap
 atom block
@@ -210,6 +210,7 @@ atom dsend_continue_trap
 atom dunlink
 atom duplicate_bag
 atom dupnames
+atom einval
 atom elib_malloc
 atom emulator
 atom enable_trace
@@ -321,6 +322,7 @@ atom ldflags
 atom Le='=<'
 atom lf
 atom line
+atom line_delimiter
 atom line_length
 atom linked_in_driver
 atom links
@@ -350,6 +352,7 @@ atom memory_internal
 atom memory_types
 atom message
 atom message_binary
+atom message_queue_data
 atom message_queue_len
 atom messages
 atom merge_trap
@@ -361,6 +364,7 @@ atom min_heap_size
 atom min_bin_vheap_size
 atom minor_version
 atom Minus='-'
+atom mixed
 atom module
 atom module_info
 atom monitored_by
@@ -423,10 +427,12 @@ atom notify
 atom notsup
 atom nouse_stdio
 atom objects
+atom off_heap
 atom offset
 atom ok
 atom old_heap_block_size
 atom old_heap_size
+atom on_heap
 atom on_load
 atom open
 atom open_error
@@ -437,13 +443,6 @@ atom orelse
 atom os_pid
 atom os_type
 atom os_version
-atom ose_bg_proc
-atom ose_int_proc
-atom ose_phantom
-atom ose_pri_proc
-atom ose_process_prio
-atom ose_process_type
-atom ose_ti_proc
 atom out
 atom out_exited
 atom out_exiting
@@ -511,6 +510,7 @@ atom return_from
 atom return_to
 atom return_trace
 atom run_queue
+atom run_queue_lengths
 atom runnable
 atom runnable_ports
 atom runnable_procs
@@ -578,12 +578,16 @@ atom timeout_value
 atom Times='*'
 atom timestamp
 atom total
+atom total_active_tasks
 atom total_heap_size
+atom total_run_queue_lengths
 atom tpkt
 atom trace trace_ts traced 
 atom trace_control_word
 atom tracer
 atom trap_exit
+atom trim
+atom trim_all
 atom try_clause
 atom true
 atom tuple
diff --git a/erts/emulator/beam/beam_bif_load.c b/erts/emulator/beam/beam_bif_load.c
index 0e192b1ebd..a000935388 100644
--- a/erts/emulator/beam/beam_bif_load.c
+++ b/erts/emulator/beam/beam_bif_load.c
@@ -38,9 +38,9 @@
 #include "erl_thr_progress.h"
 
 static void set_default_trace_pattern(Eterm module);
-static Eterm check_process_code(Process* rp, Module* modp, int allow_gc, int *redsp);
+static Eterm check_process_code(Process* rp, Module* modp, Uint flags, int *redsp);
 static void delete_code(Module* modp);
-static void decrement_refc(BeamInstr* code);
+static void decrement_refc(BeamCodeHeader*);
 static int any_heap_ref_ptrs(Eterm* start, Eterm* end, char* mod_start, Uint mod_size);
 static int any_heap_refs(Eterm* start, Eterm* end, char* mod_start, Uint mod_size);
 
@@ -58,8 +58,8 @@ BIF_RETTYPE code_is_module_native_1(BIF_ALIST_1)
 	return am_undefined;
     }
     erts_rlock_old_code(code_ix);
-    res = (erts_is_module_native(modp->curr.code) ||
-           erts_is_module_native(modp->old.code)) ?
+    res = (erts_is_module_native(modp->curr.code_hdr) ||
+           erts_is_module_native(modp->old.code_hdr)) ?
 		am_true : am_false;
     erts_runlock_old_code(code_ix);
     return res;
@@ -81,7 +81,7 @@ BIF_RETTYPE code_make_stub_module_3(BIF_ALIST_3)
     modp = erts_get_module(BIF_ARG_1, erts_active_code_ix());
 
     if (modp && modp->curr.num_breakpoints > 0) {
-	ASSERT(modp->curr.code != NULL);
+	ASSERT(modp->curr.code_hdr != NULL);
 	erts_clear_module_break(modp);
 	ASSERT(modp->curr.num_breakpoints == 0);
     }
@@ -154,7 +154,7 @@ static struct /* Protected by code_write_permission */
 {
     Process* stager;
     ErtsThrPrgrLaterOp lop;
-}commiter_state;
+} committer_state;
 #endif
 
 static Eterm
@@ -281,7 +281,7 @@ finish_loading_1(BIF_ALIST_1)
     exceptions = 0;
     for (i = 0; i < n; i++) {
 	p[i].exception = 0;
-	if (p[i].modp->curr.code && p[i].modp->old.code) {
+	if (p[i].modp->curr.code_hdr && p[i].modp->old.code_hdr) {
 	    p[i].exception = 1;
 	    exceptions++;
 	}
@@ -367,9 +367,9 @@ staging_epilogue(Process* c_p, int commit, Eterm res, int is_blocking,
 	 * schedulers to read active code_ix in a safe way while executing
 	 * without any memory barriers at all. 
 	 */
-	ASSERT(commiter_state.stager == NULL);
-	commiter_state.stager = c_p;
-	erts_schedule_thr_prgr_later_op(smp_code_ix_commiter, NULL, &commiter_state.lop);
+	ASSERT(committer_state.stager == NULL);
+	committer_state.stager = c_p;
+	erts_schedule_thr_prgr_later_op(smp_code_ix_commiter, NULL, &committer_state.lop);
 	erts_proc_inc_refc(c_p);
 	erts_suspend(c_p, ERTS_PROC_LOCK_MAIN, NULL);
 	/*
@@ -385,11 +385,11 @@ staging_epilogue(Process* c_p, int commit, Eterm res, int is_blocking,
 #ifdef ERTS_SMP
 static void smp_code_ix_commiter(void* null)
 {
-    Process* p = commiter_state.stager;
+    Process* p = committer_state.stager;
 
     erts_commit_staging_code_ix();
 #ifdef DEBUG
-    commiter_state.stager = NULL;
+    committer_state.stager = NULL;
 #endif
     erts_release_code_write_permission();
     erts_smp_proc_lock(p, ERTS_PROC_LOCK_STATUS);
@@ -417,7 +417,7 @@ check_old_code_1(BIF_ALIST_1)
     modp = erts_get_module(BIF_ARG_1, code_ix);
     if (modp != NULL) {
 	erts_rlock_old_code(code_ix);
-	if (modp->old.code != NULL) {
+	if (modp->old.code_hdr) {
 	    res = am_true;
 	}
 	erts_runlock_old_code(code_ix);
@@ -426,7 +426,7 @@ check_old_code_1(BIF_ALIST_1)
 }
 
 Eterm
-erts_check_process_code(Process *c_p, Eterm module, int allow_gc, int *redsp)
+erts_check_process_code(Process *c_p, Eterm module, Uint flags, int *redsp)
 {
     Module* modp;
     Eterm res;
@@ -441,7 +441,8 @@ erts_check_process_code(Process *c_p, Eterm module, int allow_gc, int *redsp)
     if (!modp)
 	return am_false;
     erts_rlock_old_code(code_ix);
-    res = modp->old.code ? check_process_code(c_p, modp, allow_gc, redsp) : am_false;
+    res = (!modp->old.code_hdr ? am_false :
+           check_process_code(c_p, modp, flags, redsp));
     erts_runlock_old_code(code_ix);
 
     return res;
@@ -450,49 +451,21 @@ erts_check_process_code(Process *c_p, Eterm module, int allow_gc, int *redsp)
 BIF_RETTYPE erts_internal_check_process_code_2(BIF_ALIST_2)
 {
     int reds = 0;
+    Uint flags;
     Eterm res;
-    Eterm olist = BIF_ARG_2;
-    int allow_gc = 1;
 
     if (is_not_atom(BIF_ARG_1))
 	goto badarg;
 
-    while (is_list(olist)) {
-	Eterm *lp = list_val(olist);
-	Eterm opt = CAR(lp);
-	if (is_tuple(opt)) {
-	    Eterm* tp = tuple_val(opt);
-	    switch (arityval(tp[0])) {
-	    case 2:
-		switch (tp[1]) {
-		case am_allow_gc:
-		    switch (tp[2]) {
-		    case am_false:
-			allow_gc = 0;
-			break;
-		    case am_true:
-			allow_gc = 1;
-			break;
-		    default:
-			goto badarg;
-		    }
-		    break;
-		default:
-		    goto badarg;
-		}
-		break;
-	    default:
-		goto badarg;
-	    }
-	}
-	else
-	    goto badarg;
-	olist = CDR(lp);
+    if (is_not_small(BIF_ARG_2))
+        goto badarg;
+
+    flags = unsigned_val(BIF_ARG_2);
+    if (flags & ~ERTS_CPC_ALL) {
+        goto badarg;
     }
-    if (is_not_nil(olist))
-	goto badarg;
 
-    res = erts_check_process_code(BIF_P, BIF_ARG_1, allow_gc, &reds);
+    res = erts_check_process_code(BIF_P, BIF_ARG_1, flags, &reds);
 
     ASSERT(is_value(res));
 
@@ -525,7 +498,7 @@ BIF_RETTYPE delete_module_1(BIF_ALIST_1)
 	if (!modp) {
 	    res = am_undefined;
 	}
-	else if (modp->old.code != 0) {
+	else if (modp->old.code_hdr) {
 	    erts_dsprintf_buf_t *dsbufp = erts_create_logger_dsbuf();
 	    erts_dsprintf(dsbufp, "Module %T must be purged before loading\n",
 			  BIF_ARG_1);
@@ -563,8 +536,8 @@ BIF_RETTYPE module_loaded_1(BIF_ALIST_1)
     }
     code_ix = erts_active_code_ix();
     if ((modp = erts_get_module(BIF_ARG_1, code_ix)) != NULL) {
-	if (modp->curr.code != NULL
-	    && modp->curr.code[MI_ON_LOAD_FUNCTION_PTR] == 0) {
+	if (modp->curr.code_hdr
+	    && modp->curr.code_hdr->on_load_function_ptr == NULL) {
 	    res = am_true;
 	}
     }
@@ -611,8 +584,8 @@ BIF_RETTYPE call_on_load_function_1(BIF_ALIST_1)
 {
     Module* modp = erts_get_module(BIF_ARG_1, erts_active_code_ix());
 
-    if (modp && modp->curr.code) {
-	BIF_TRAP_CODE_PTR_0(BIF_P, modp->curr.code[MI_ON_LOAD_FUNCTION_PTR]);
+    if (modp && modp->curr.code_hdr) {
+	BIF_TRAP_CODE_PTR_0(BIF_P, modp->curr.code_hdr->on_load_function_ptr);
     }
     else {
 	BIF_ERROR(BIF_P, BADARG);
@@ -623,7 +596,6 @@ BIF_RETTYPE finish_after_on_load_2(BIF_ALIST_2)
 {
     ErtsCodeIndex code_ix;
     Module* modp;
-    Eterm on_load;
 
     if (!erts_try_seize_code_write_permission(BIF_P)) {
 	ERTS_BIF_YIELD2(bif_export[BIF_finish_after_on_load_2],
@@ -638,14 +610,14 @@ BIF_RETTYPE finish_after_on_load_2(BIF_ALIST_2)
     code_ix = erts_active_code_ix();
     modp = erts_get_module(BIF_ARG_1, code_ix);
 
-    if (!modp || modp->curr.code == 0) {
+    if (!modp || !modp->curr.code_hdr) {
     error:
 	erts_smp_thr_progress_unblock();
         erts_smp_proc_lock(BIF_P, ERTS_PROC_LOCK_MAIN);
 	erts_release_code_write_permission();
 	BIF_ERROR(BIF_P, BADARG);
     }
-    if ((on_load = modp->curr.code[MI_ON_LOAD_FUNCTION_PTR]) == 0) {
+    if (modp->curr.code_hdr->on_load_function_ptr == NULL) {
 	goto error;
     }
     if (BIF_ARG_2 != am_false && BIF_ARG_2 != am_true) {
@@ -667,7 +639,7 @@ BIF_RETTYPE finish_after_on_load_2(BIF_ALIST_2)
 		ep->code[4] = 0;
 	    }
 	}
-	modp->curr.code[MI_ON_LOAD_FUNCTION_PTR] = 0;
+	modp->curr.code_hdr->on_load_function_ptr = NULL;
 	set_default_trace_pattern(BIF_ARG_1);
     } else if (BIF_ARG_2 == am_false) {
 	BeamInstr* code;
@@ -679,13 +651,16 @@ BIF_RETTYPE finish_after_on_load_2(BIF_ALIST_2)
 	 * the current code; the old code is not touched.
 	 */
 	erts_total_code_size -= modp->curr.code_length;
-	code = modp->curr.code;
-	end = (BeamInstr *)((char *)code + modp->curr.code_length);
+	code = (BeamInstr*) modp->curr.code_hdr;
+	end = (BeamInstr *) ((char *)code + modp->curr.code_length);
 	erts_cleanup_funs_on_purge(code, end);
 	beam_catches_delmod(modp->curr.catches, code, modp->curr.code_length,
 			    erts_active_code_ix());
-	erts_free(ERTS_ALC_T_CODE, (void *) code);
-	modp->curr.code = NULL;
+        if (modp->curr.code_hdr->literals_start) {
+            erts_free(ERTS_ALC_T_LITERAL, modp->curr.code_hdr->literals_start);
+        }
+	erts_free(ERTS_ALC_T_CODE, modp->curr.code_hdr);
+	modp->curr.code_hdr = NULL;
 	modp->curr.code_length = 0;
 	modp->curr.catches = BEAM_CATCHES_NIL;
 	erts_remove_from_ranges(code);
@@ -721,31 +696,50 @@ set_default_trace_pattern(Eterm module)
     }
 }
 
+static ERTS_INLINE int
+check_mod_funs(Process *p, ErlOffHeap *off_heap, char *area, size_t area_size)
+{
+    struct erl_off_heap_header* oh;
+    for (oh = off_heap->first; oh; oh = oh->next) {
+	if (thing_subtag(oh->thing_word) == FUN_SUBTAG) {
+	    ErlFunThing* funp = (ErlFunThing*) oh;
+	    if (ErtsInArea(funp->fe->address, area, area_size))
+		return !0;
+	}
+    }
+    return 0;
+}
+
+
 static Eterm
-check_process_code(Process* rp, Module* modp, int allow_gc, int *redsp)
+check_process_code(Process* rp, Module* modp, Uint flags, int *redsp)
 {
     BeamInstr* start;
+    char* literals;
+    Uint lit_bsize;
     char* mod_start;
     Uint mod_size;
-    BeamInstr* end;
     Eterm* sp;
-    struct erl_off_heap_header* oh;
     int done_gc = 0;
+    int need_gc = 0;
+    ErtsMessage *msgp;
+    ErlHeapFragment *hfrag;
 
-#define INSIDE(a) (start <= (a) && (a) < end)
+#define ERTS_ORDINARY_GC__ (1 << 0)
+#define ERTS_LITERAL_GC__  (1 << 1)
 
     /*
      * Pick up limits for the module.
      */
-    start = modp->old.code;
-    end = (BeamInstr *)((char *)start + modp->old.code_length);
+    start = (BeamInstr*) modp->old.code_hdr;
     mod_start = (char *) start;
     mod_size = modp->old.code_length;
 
     /*
      * Check if current instruction or continuation pointer points into module.
      */
-    if (INSIDE(rp->i) || INSIDE(rp->cp)) {
+    if (ErtsInArea(rp->i, mod_start, mod_size)
+	|| ErtsInArea(rp->cp, mod_start, mod_size)) {
 	return am_true;
     }
 
@@ -753,7 +747,7 @@ check_process_code(Process* rp, Module* modp, int allow_gc, int *redsp)
      * Check all continuation pointers stored on the stack.
      */
     for (sp = rp->stop; sp < STACK_START(rp); sp++) {
-	if (is_CP(*sp) && INSIDE(cp_val(*sp))) {
+	if (is_CP(*sp) && ErtsInArea(cp_val(*sp), mod_start, mod_size)) {
 	    return am_true;
 	}
     }
@@ -767,15 +761,15 @@ check_process_code(Process* rp, Module* modp, int allow_gc, int *redsp)
 	struct StackTrace *s;
 	ASSERT(is_list(rp->ftrace));
 	s = (struct StackTrace *) big_val(CDR(list_val(rp->ftrace)));
-	if ((s->pc && INSIDE(s->pc)) ||
-	    (s->current && INSIDE(s->current))) {
+	if ((s->pc && ErtsInArea(s->pc, mod_start, mod_size)) ||
+	    (s->current && ErtsInArea(s->current, mod_start, mod_size))) {
 	    rp->freason = EXC_NULL;
 	    rp->fvalue = NIL;
 	    rp->ftrace = NIL;
 	} else {
 	    int i;
 	    for (i = 0;  i < s->depth;  i++) {
-		if (INSIDE(s->trace[i])) {
+		if (ErtsInArea(s->trace[i], mod_start, mod_size)) {
 		    rp->freason = EXC_NULL;
 		    rp->fvalue = NIL;
 		    rp->ftrace = NIL;
@@ -796,111 +790,147 @@ check_process_code(Process* rp, Module* modp, int allow_gc, int *redsp)
     }
 
     /*
-     * See if there are funs that refer to the old version of the module.
+     * Message queue can contains funs, but (at least currently) no
+     * constants. If we got references to this module from the message
+     * queue, a GC cannot remove these...
      */
 
- rescan:
-    for (oh = MSO(rp).first; oh; oh = oh->next) {
-	if (thing_subtag(oh->thing_word) == FUN_SUBTAG) {
-	    ErlFunThing* funp = (ErlFunThing*) oh;
+    erts_smp_proc_lock(rp, ERTS_PROC_LOCK_MSGQ);
+    ERTS_SMP_MSGQ_MV_INQ2PRIVQ(rp);
+    erts_smp_proc_unlock(rp, ERTS_PROC_LOCK_MSGQ);
 
-	    if (INSIDE((BeamInstr *) funp->fe->address)) {
-		if (done_gc) {
-		    return am_true;
-		} else {
-		    if (!allow_gc)
-			return am_aborted;
-		    /*
-		    * Try to get rid of this fun by garbage collecting.
-		    * Clear both fvalue and ftrace to make sure they
-		    * don't hold any funs.
-		    */
-		    rp->freason = EXC_NULL;
-		    rp->fvalue = NIL;
-		    rp->ftrace = NIL;
-		    done_gc = 1;
-		    FLAGS(rp) |= F_NEED_FULLSWEEP;
-		    *redsp += erts_garbage_collect(rp, 0, rp->arg_reg, rp->arity);
-		    goto rescan;
-		}
-	    }
+    literals = (char*) modp->old.code_hdr->literals_start;
+    lit_bsize = (char*) modp->old.code_hdr->literals_end - literals;
+
+    for (msgp = rp->msg.first; msgp; msgp = msgp->next) {
+	if (msgp->data.attached == ERTS_MSG_COMBINED_HFRAG)
+	    hfrag = &msgp->hfrag;
+	else if (is_value(ERL_MESSAGE_TERM(msgp)) && msgp->data.heap_frag)
+	    hfrag = msgp->data.heap_frag;
+	else
+	    continue;
+	for (; hfrag; hfrag = hfrag->next) {
+	    if (check_mod_funs(rp, &hfrag->off_heap, mod_start, mod_size))
+		return am_true;
+	    /* Should not contain any constants... */
+	    ASSERT(!any_heap_refs(&hfrag->mem[0],
+				  &hfrag->mem[hfrag->used_size],
+                                  literals,
+				  lit_bsize));
 	}
     }
 
-    /*
-     * See if there are constants inside the module referenced by the process.
-     */
-    done_gc = 0;
-    for (;;) {
-	ErlMessage* mp;
-
-	if (any_heap_ref_ptrs(&rp->fvalue, &rp->fvalue+1, mod_start, mod_size)) {
+    while (1) {
+
+	/* Check heap, stack etc... */
+	if (check_mod_funs(rp, &rp->off_heap, mod_start, mod_size))
+	    goto try_gc;
+        if (!(flags & ERTS_CPC_COPY_LITERALS)) {
+            /* Process ok. May contain old literals but we will be called
+             * again before module is purged.
+             */
+            return am_false;
+        }
+	if (any_heap_ref_ptrs(&rp->fvalue, &rp->fvalue+1, literals, lit_bsize)) {
 	    rp->freason = EXC_NULL;
 	    rp->fvalue = NIL;
 	    rp->ftrace = NIL;
 	}
-	if (any_heap_ref_ptrs(rp->stop, rp->hend, mod_start, mod_size)) {
-	    goto need_gc;
-	}
-	if (any_heap_refs(rp->heap, rp->htop, mod_start, mod_size)) {
-	    goto need_gc;
-	}
-
-	if (any_heap_refs(rp->old_heap, rp->old_htop, mod_start, mod_size)) {
-	    goto need_gc;
+	if (any_heap_ref_ptrs(rp->stop, rp->hend, literals, lit_bsize))
+	    goto try_literal_gc;
+	if (any_heap_refs(rp->heap, rp->htop, literals, lit_bsize))
+	    goto try_literal_gc;
+	if (any_heap_refs(rp->old_heap, rp->old_htop, literals, lit_bsize))
+	    goto try_literal_gc;
+
+	/* Check dictionary */
+	if (rp->dictionary) {
+	    Eterm* start = ERTS_PD_START(rp->dictionary);
+	    Eterm* end = start + ERTS_PD_SIZE(rp->dictionary);
+
+	    if (any_heap_ref_ptrs(start, end, literals, lit_bsize))
+		goto try_literal_gc;
 	}
 
-	if (rp->dictionary != NULL) {
-	    Eterm* start = rp->dictionary->data;
-	    Eterm* end = start + rp->dictionary->used;
+	/* Check heap fragments */
+	for (hfrag = rp->mbuf; hfrag; hfrag = hfrag->next) {
+	    Eterm *hp, *hp_end;
+	    /* Off heap lists should already have been moved into process */
+	    ASSERT(!check_mod_funs(rp, &hfrag->off_heap, mod_start, mod_size));
 
-	    if (any_heap_ref_ptrs(start, end, mod_start, mod_size)) {
-		goto need_gc;
-	    }
+	    hp = &hfrag->mem[0];
+	    hp_end = &hfrag->mem[hfrag->used_size];
+	    if (any_heap_refs(hp, hp_end, literals, lit_bsize))
+		goto try_literal_gc;
 	}
 
-	for (mp = rp->msg.first; mp != NULL; mp = mp->next) {
-	    if (any_heap_ref_ptrs(mp->m, mp->m+2, mod_start, mod_size)) {
-		goto need_gc;
+#ifdef DEBUG
+	/*
+	 * Message buffer fragments should not have any references
+	 * to constants, and off heap lists should already have
+	 * been moved into process off heap structure.
+	 */
+	for (msgp = rp->msg_frag; msgp; msgp = msgp->next) {
+	    if (msgp->data.attached == ERTS_MSG_COMBINED_HFRAG)
+		hfrag = &msgp->hfrag;
+	    else
+		hfrag = msgp->data.heap_frag;
+	    for (; hfrag; hfrag = hfrag->next) {
+		Eterm *hp, *hp_end;
+		ASSERT(!check_mod_funs(rp, &hfrag->off_heap, mod_start, mod_size));
+
+		hp = &hfrag->mem[0];
+		hp_end = &hfrag->mem[hfrag->used_size];
+		ASSERT(!any_heap_refs(hp, hp_end, literals, lit_bsize));
 	    }
 	}
-	break;
 
-    need_gc:
-	if (done_gc) {
+#endif
+
+	return am_false;
+
+    try_literal_gc:
+	need_gc |= ERTS_LITERAL_GC__;
+
+    try_gc:
+	need_gc |= ERTS_ORDINARY_GC__;
+
+	if ((done_gc & need_gc) == need_gc)
 	    return am_true;
-	} else {
-	    Eterm* literals;
-	    Uint lit_size;
-	    struct erl_off_heap_header* oh;
 
-	    if (!allow_gc)
-		return am_aborted;
+	if (!(flags & ERTS_CPC_ALLOW_GC))
+	    return am_aborted;
 
-	    /*
-	     * Try to get rid of constants by by garbage collecting.
-	     * Clear both fvalue and ftrace.
-	     */
-	    rp->freason = EXC_NULL;
-	    rp->fvalue = NIL;
-	    rp->ftrace = NIL;
-	    done_gc = 1;
+	need_gc &= ~done_gc;
+
+	/*
+	 * Try to get rid of constants by by garbage collecting.
+	 * Clear both fvalue and ftrace.
+	 */
+
+	rp->freason = EXC_NULL;
+	rp->fvalue = NIL;
+	rp->ftrace = NIL;
+
+	if (need_gc & ERTS_ORDINARY_GC__) {
 	    FLAGS(rp) |= F_NEED_FULLSWEEP;
-	    *redsp += erts_garbage_collect(rp, 0, rp->arg_reg, rp->arity);
-	    literals = (Eterm *) modp->old.code[MI_LITERALS_START];
-	    lit_size = (Eterm *) modp->old.code[MI_LITERALS_END] - literals;
-	    oh = (struct erl_off_heap_header *)
-		modp->old.code[MI_LITERALS_OFF_HEAP];
-	    *redsp += lit_size / 10; /* Need, better value... */
-	    erts_garbage_collect_literals(rp, literals, lit_size, oh);
+	    *redsp += erts_garbage_collect_nobump(rp, 0, rp->arg_reg, rp->arity);
+	    done_gc |= ERTS_ORDINARY_GC__;
 	}
+	if (need_gc & ERTS_LITERAL_GC__) {
+	    struct erl_off_heap_header* oh;
+	    oh = modp->old.code_hdr->literals_off_heap;
+	    *redsp += lit_bsize / 64; /* Need, better value... */
+	    erts_garbage_collect_literals(rp, (Eterm*)literals, lit_bsize, oh);
+	    done_gc |= ERTS_LITERAL_GC__;
+	}
+	need_gc = 0;
     }
-    return am_false;
-#undef INSIDE
-}
 
-#define in_area(ptr,start,nbytes) \
-    ((UWord)((char*)(ptr) - (char*)(start)) < (nbytes))
+#undef ERTS_ORDINARY_GC__
+#undef ERTS_LITERAL_GC__
+
+}
 
 static int
 any_heap_ref_ptrs(Eterm* start, Eterm* end, char* mod_start, Uint mod_size)
@@ -913,7 +943,7 @@ any_heap_ref_ptrs(Eterm* start, Eterm* end, char* mod_start, Uint mod_size)
 	switch (primary_tag(val)) {
 	case TAG_PRIMARY_BOXED:
 	case TAG_PRIMARY_LIST:
-	    if (in_area(EXPAND_POINTER(val), mod_start, mod_size)) {
+	    if (ErtsInArea(val, mod_start, mod_size)) {
 		return 1;
 	    }
 	    break;
@@ -933,7 +963,7 @@ any_heap_refs(Eterm* start, Eterm* end, char* mod_start, Uint mod_size)
 	switch (primary_tag(val)) {
 	case TAG_PRIMARY_BOXED:
 	case TAG_PRIMARY_LIST:
-	    if (in_area(EXPAND_POINTER(val), mod_start, mod_size)) {
+	    if (ErtsInArea(val, mod_start, mod_size)) {
 		return 1;
 	    }
 	    break;
@@ -943,7 +973,7 @@ any_heap_refs(Eterm* start, Eterm* end, char* mod_start, Uint mod_size)
                 if (header_is_bin_matchstate(val)) {
                     ErlBinMatchState *ms = (ErlBinMatchState*) p;
                     ErlBinMatchBuffer *mb = &(ms->mb);
-                    if (in_area(EXPAND_POINTER(mb->orig), mod_start, mod_size)) {
+                    if (ErtsInArea(mb->orig, mod_start, mod_size)) {
                         return 1;
                     }
                 }
@@ -958,7 +988,104 @@ any_heap_refs(Eterm* start, Eterm* end, char* mod_start, Uint mod_size)
 
 #undef in_area
 
-BIF_RETTYPE purge_module_1(BIF_ALIST_1)
+#ifdef ERTS_SMP
+static void copy_literals_commit(void*);
+#endif
+
+copy_literals_t erts_clrange = {NULL, 0, THE_NON_VALUE};
+
+/* copy literals
+ *
+ * copy_literals.ptr = LitPtr
+ * copy_literals.sz  = LitSz
+ * ------ THR PROG COMMIT -----
+ *
+ *  - check process code
+ *  - check process code
+ *  ...
+ * copy_literals.ptr = NULL
+ * copy_literals.sz  = 0
+ * ------ THR PROG COMMIT -----
+ *  ...
+ */
+
+
+BIF_RETTYPE erts_internal_copy_literals_2(BIF_ALIST_2)
+{
+    ErtsCodeIndex code_ix;
+    Eterm res = am_true;
+
+    if (is_not_atom(BIF_ARG_1) || (am_true != BIF_ARG_2 && am_false != BIF_ARG_2)) {
+        BIF_ERROR(BIF_P, BADARG);
+    }
+
+    if (!erts_try_seize_code_write_permission(BIF_P)) {
+	ERTS_BIF_YIELD2(bif_export[BIF_erts_internal_copy_literals_2],
+                        BIF_P, BIF_ARG_1, BIF_ARG_2);
+    }
+
+    code_ix = erts_active_code_ix();
+
+    if (BIF_ARG_2 == am_true) {
+        Module* modp = erts_get_module(BIF_ARG_1, code_ix);
+        if (!modp || !modp->old.code_hdr) {
+            res = am_false;
+            goto done;
+        }
+        if (erts_clrange.ptr != NULL
+            && !(BIF_P->static_flags & ERTS_STC_FLG_SYSTEM_PROC)) {
+            res = am_aborted;
+            goto done;
+        }
+        erts_clrange.ptr = modp->old.code_hdr->literals_start;
+        erts_clrange.sz  = modp->old.code_hdr->literals_end - erts_clrange.ptr;
+        erts_clrange.pid = BIF_P->common.id;
+    } else if (BIF_ARG_2 == am_false) {
+        if (erts_clrange.pid != BIF_P->common.id) {
+            res = am_false;
+            goto done;
+        }
+        erts_clrange.ptr = NULL;
+        erts_clrange.sz  = 0;
+        erts_clrange.pid = THE_NON_VALUE;
+    }
+
+#ifdef ERTS_SMP
+    ASSERT(committer_state.stager == NULL);
+    committer_state.stager = BIF_P;
+    erts_schedule_thr_prgr_later_op(copy_literals_commit, NULL, &committer_state.lop);
+    erts_proc_inc_refc(BIF_P);
+    erts_suspend(BIF_P, ERTS_PROC_LOCK_MAIN, NULL);
+    ERTS_BIF_YIELD_RETURN(BIF_P, am_true);
+#endif
+done:
+    erts_release_code_write_permission();
+    BIF_RET(res);
+}
+
+#ifdef ERTS_SMP
+static void copy_literals_commit(void* null) {
+    Process* p = committer_state.stager;
+#ifdef DEBUG
+    committer_state.stager = NULL;
+#endif
+    erts_release_code_write_permission();
+    erts_smp_proc_lock(p, ERTS_PROC_LOCK_STATUS);
+    if (!ERTS_PROC_IS_EXITING(p)) {
+	erts_resume(p, ERTS_PROC_LOCK_STATUS);
+    }
+    erts_smp_proc_unlock(p, ERTS_PROC_LOCK_STATUS);
+    erts_proc_dec_refc(p);
+}
+#endif /* ERTS_SMP */
+
+
+/* Do the actualy module purging and return:
+ * true for success
+ * false if no such old module
+ * BADARG if not an atom
+ */
+BIF_RETTYPE erts_internal_purge_module_1(BIF_ALIST_1)
 {
     ErtsCodeIndex code_ix;
     BeamInstr* code;
@@ -972,7 +1099,8 @@ BIF_RETTYPE purge_module_1(BIF_ALIST_1)
     }
 
     if (!erts_try_seize_code_write_permission(BIF_P)) {
-	ERTS_BIF_YIELD1(bif_export[BIF_purge_module_1], BIF_P, BIF_ARG_1);
+	ERTS_BIF_YIELD1(bif_export[BIF_erts_internal_purge_module_1],
+                        BIF_P, BIF_ARG_1);
     }
 
     code_ix = erts_active_code_ix();
@@ -982,7 +1110,7 @@ BIF_RETTYPE purge_module_1(BIF_ALIST_1)
      */
 
     if ((modp = erts_get_module(BIF_ARG_1, code_ix)) == NULL) {
-	ERTS_BIF_PREP_ERROR(ret, BIF_P, BADARG);
+        ERTS_BIF_PREP_RET(ret, am_false);
     }
     else {
 	erts_rwlock_old_code(code_ix);
@@ -990,8 +1118,8 @@ BIF_RETTYPE purge_module_1(BIF_ALIST_1)
 	/*
 	 * Any code to purge?
 	 */
-	if (modp->old.code == 0) {
-	    ERTS_BIF_PREP_ERROR(ret, BIF_P, BADARG);
+	if (!modp->old.code_hdr) {
+            ERTS_BIF_PREP_RET(ret, am_false);
 	}
 	else {
 	    /*
@@ -1013,14 +1141,17 @@ BIF_RETTYPE purge_module_1(BIF_ALIST_1)
 	     */
 	    ASSERT(erts_total_code_size >= modp->old.code_length);
 	    erts_total_code_size -= modp->old.code_length;
-	    code = modp->old.code;
+	    code = (BeamInstr*) modp->old.code_hdr;
 	    end = (BeamInstr *)((char *)code + modp->old.code_length);
 	    erts_cleanup_funs_on_purge(code, end);
 	    beam_catches_delmod(modp->old.catches, code, modp->old.code_length,
 				code_ix);
-	    decrement_refc(code);
+	    decrement_refc(modp->old.code_hdr);
+            if (modp->old.code_hdr->literals_start) {
+                erts_free(ERTS_ALC_T_LITERAL, modp->old.code_hdr->literals_start);
+            }
 	    erts_free(ERTS_ALC_T_CODE, (void *) code);
-	    modp->old.code = NULL;
+	    modp->old.code_hdr = NULL;
 	    modp->old.code_length = 0;
 	    modp->old.catches = BEAM_CATCHES_NIL;
 	    erts_remove_from_ranges(code);
@@ -1037,10 +1168,9 @@ BIF_RETTYPE purge_module_1(BIF_ALIST_1)
 }
 
 static void
-decrement_refc(BeamInstr* code)
+decrement_refc(BeamCodeHeader* code_hdr)
 {
-    struct erl_off_heap_header* oh =
-	(struct erl_off_heap_header *) code[MI_LITERALS_OFF_HEAP];
+    struct erl_off_heap_header* oh = code_hdr->literals_off_heap;
 
     while (oh) {
 	Binary* bptr;
@@ -1089,7 +1219,7 @@ delete_code(Module* modp)
     ASSERT(modp->curr.num_breakpoints == 0);
     ASSERT(modp->curr.num_traced_exports == 0);
     modp->old = modp->curr;
-    modp->curr.code = NULL;
+    modp->curr.code_hdr = NULL;
     modp->curr.code_length = 0;
     modp->curr.catches = BEAM_CATCHES_NIL;
     modp->curr.nif = NULL;
@@ -1106,9 +1236,9 @@ beam_make_current_old(Process *c_p, ErtsProcLocks c_p_locks, Eterm module)
      * if not, delete old code; error if old code already exists.
      */
 
-    if (modp->curr.code != NULL && modp->old.code != NULL)  {
+    if (modp->curr.code_hdr && modp->old.code_hdr)  {
 	return am_not_purged;
-    } else if (modp->old.code == NULL) { /* Make the current version old. */
+    } else if (!modp->old.code_hdr) { /* Make the current version old. */
 	delete_code(modp);
     }
     return NIL;
diff --git a/erts/emulator/beam/beam_bp.c b/erts/emulator/beam/beam_bp.c
index 016d0aaa32..2a23da4f25 100644
--- a/erts/emulator/beam/beam_bp.c
+++ b/erts/emulator/beam/beam_bp.c
@@ -75,6 +75,16 @@ extern BeamInstr beam_return_time_trace[1]; /* OpCode(i_return_time_trace) */
 erts_smp_atomic32_t erts_active_bp_index;
 erts_smp_atomic32_t erts_staging_bp_index;
 
+/*
+ * Inlined helpers
+ */
+
+static ERTS_INLINE ErtsMonotonicTime
+get_mtime(Process *c_p)
+{
+    return erts_get_monotonic_time(ERTS_PROC_GET_SCHDATA(c_p));
+}
+
 /* *************************************************************************
 ** Local prototypes
 */
@@ -97,9 +107,6 @@ static int clear_function_break(BeamInstr *pc, Uint break_flags);
 
 static BpDataTime* get_time_break(BeamInstr *pc);
 static GenericBpData* check_break(BeamInstr *pc, Uint break_flags);
-static void bp_time_diff(bp_data_time_item_t *item,
-			 process_breakpoint_time_t *pbt,
-			 Uint ms, Uint s, Uint us);
 
 static void bp_meta_unref(BpMetaPid* bmp);
 static void bp_count_unref(BpCount* bcp);
@@ -110,13 +117,8 @@ static void uninstall_breakpoint(BeamInstr* pc);
 /* bp_hash */
 #define BP_TIME_ADD(pi0, pi1)                       \
     do {                                            \
-	Uint r;                                     \
 	(pi0)->count   += (pi1)->count;             \
-	(pi0)->s_time  += (pi1)->s_time;            \
-	(pi0)->us_time += (pi1)->us_time;           \
-	r = (pi0)->us_time / 1000000;               \
-	(pi0)->s_time  += r;                        \
-	(pi0)->us_time  = (pi0)->us_time % 1000000; \
+	(pi0)->time    += (pi1)->time;              \
     } while(0)
 
 static void bp_hash_init(bp_time_hash_t *hash, Uint n);
@@ -152,8 +154,8 @@ erts_bp_match_functions(BpFunctions* f, Eterm mfa[3], int specified)
     num_modules = 0;
     for (current = 0; current < max_modules; current++) {
 	modp = module_code(current, code_ix);
-	if (modp->curr.code) {
-	    max_funcs += modp->curr.code[MI_NUM_FUNCTIONS];
+	if (modp->curr.code_hdr) {
+	    max_funcs += modp->curr.code_hdr->num_functions;
 	    module[num_modules++] = modp;
 	}
     }
@@ -161,9 +163,9 @@ erts_bp_match_functions(BpFunctions* f, Eterm mfa[3], int specified)
     f->matching = (BpFunction *) Alloc(max_funcs*sizeof(BpFunction));
     i = 0;
     for (current = 0; current < num_modules; current++) {
-	BeamInstr** code_base = (BeamInstr **) module[current]->curr.code;
+	BeamCodeHeader* code_hdr = module[current]->curr.code_hdr;
 	BeamInstr* code;
-	Uint num_functions = (Uint)(UWord) code_base[MI_NUM_FUNCTIONS];
+	Uint num_functions = (Uint)(UWord) code_hdr->num_functions;
 	Uint fi;
 
 	if (specified > 0) {
@@ -177,7 +179,7 @@ erts_bp_match_functions(BpFunctions* f, Eterm mfa[3], int specified)
 	    BeamInstr* pc;
 	    int wi;
 
-	    code = code_base[MI_FUNCTIONS+fi];
+	    code = code_hdr->functions[fi];
 	    ASSERT(code[0] == (BeamInstr) BeamOp(op_i_func_info_IaaI));
 	    pc = code+5;
 	    if (erts_is_native_break(pc)) {
@@ -547,21 +549,21 @@ erts_clear_all_breaks(BpFunctions* f)
 
 int 
 erts_clear_module_break(Module *modp) {
-    BeamInstr** code_base;
+    BeamCodeHeader* code_hdr;
     Uint n;
     Uint i;
 
     ERTS_SMP_LC_ASSERT(erts_smp_thr_progress_is_blocking());
     ASSERT(modp);
-    code_base = (BeamInstr **) modp->curr.code;
-    if (code_base == NULL) {
+    code_hdr = modp->curr.code_hdr;
+    if (!code_hdr) {
 	return 0;
     }
-    n = (Uint)(UWord) code_base[MI_NUM_FUNCTIONS];
+    n = (Uint)(UWord) code_hdr->num_functions;
     for (i = 0; i < n; ++i) {
 	BeamInstr* pc;
 
-	pc = code_base[MI_FUNCTIONS+i] + 5;
+	pc = code_hdr->functions[i] + 5;
 	if (erts_is_native_break(pc)) {
 	    continue;
 	}
@@ -573,7 +575,7 @@ erts_clear_module_break(Module *modp) {
     for (i = 0; i < n; ++i) {
 	BeamInstr* pc;
 
-	pc = code_base[MI_FUNCTIONS+i] + 5;
+	pc = code_hdr->functions[i] + 5;
 	if (erts_is_native_break(pc)) {
 	    continue;
 	}
@@ -948,7 +950,7 @@ do_call_trace(Process* c_p, BeamInstr* I, Eterm* reg,
 void
 erts_trace_time_call(Process* c_p, BeamInstr* I, BpDataTime* bdt)
 {
-    Uint ms,s,us;
+    ErtsMonotonicTime time;
     process_breakpoint_time_t *pbt = NULL;
     bp_data_time_item_t sitem, *item = NULL;
     bp_time_hash_t *h = NULL;
@@ -961,7 +963,7 @@ erts_trace_time_call(Process* c_p, BeamInstr* I, BpDataTime* bdt)
      * from the process psd  */
 
     pbt = ERTS_PROC_GET_CALL_TIME(c_p);
-    get_sys_now(&ms, &s, &us);
+    time = get_mtime(c_p);
 
     /* get pbt
      * timestamp = t0
@@ -976,7 +978,7 @@ erts_trace_time_call(Process* c_p, BeamInstr* I, BpDataTime* bdt)
     } else {
 	ASSERT(pbt->pc);
 	/* add time to previous code */
-	bp_time_diff(&sitem, pbt, ms, s, us);
+	sitem.time = time - pbt->time;
 	sitem.pid = c_p->common.id;
 	sitem.count = 0;
 
@@ -1002,8 +1004,7 @@ erts_trace_time_call(Process* c_p, BeamInstr* I, BpDataTime* bdt)
     /* Add count to this code */
     sitem.pid     = c_p->common.id;
     sitem.count   = 1;
-    sitem.s_time  = 0;
-    sitem.us_time = 0;
+    sitem.time    = 0;
 
     /* this breakpoint */
     ASSERT(bdt);
@@ -1020,15 +1021,13 @@ erts_trace_time_call(Process* c_p, BeamInstr* I, BpDataTime* bdt)
     }
 
     pbt->pc = I;
-    pbt->ms = ms;
-    pbt->s  = s;
-    pbt->us = us;
+    pbt->time = time;
 }
 
 void
 erts_trace_time_return(Process *p, BeamInstr *pc)
 {
-    Uint ms,s,us;
+    ErtsMonotonicTime time;
     process_breakpoint_time_t *pbt = NULL;
     bp_data_time_item_t sitem, *item = NULL;
     bp_time_hash_t *h = NULL;
@@ -1041,7 +1040,7 @@ erts_trace_time_return(Process *p, BeamInstr *pc)
      * from the process psd  */
 
     pbt = ERTS_PROC_GET_CALL_TIME(p);
-    get_sys_now(&ms,&s,&us);
+    time = get_mtime(p);
 
     /* get pbt
      * lookup bdt from code
@@ -1057,7 +1056,7 @@ erts_trace_time_return(Process *p, BeamInstr *pc)
 	 */
 	ASSERT(pbt->pc);
 
-	bp_time_diff(&sitem, pbt, ms, s, us);
+	sitem.time = time - pbt->time;
 	sitem.pid   = p->common.id;
 	sitem.count = 0;
 
@@ -1080,9 +1079,7 @@ erts_trace_time_return(Process *p, BeamInstr *pc)
 	}
 
 	pbt->pc = pc;
-	pbt->ms = ms;
-	pbt->s  = s;
-	pbt->us = us;
+	pbt->time = time;
     }
 }
 
@@ -1183,10 +1180,14 @@ int erts_is_time_break(Process *p, BeamInstr *pc, Eterm *retval) {
 		for(ix = 0; ix < hash.n; ix++) {
 		    item = &(hash.item[ix]);
 		    if (item->pid != NIL) {
+			ErtsMonotonicTime sec, usec;
+			usec = ERTS_MONOTONIC_TO_USEC(item->time);
+			sec = usec / 1000000;
+			usec = usec - sec*1000000;
 			t = TUPLE4(hp, item->pid,
 				make_small(item->count),
-				make_small(item->s_time),
-				make_small(item->us_time));
+				   make_small((Uint) sec),
+				   make_small((Uint) usec));
 			hp += 5;
 			*retval = CONS(hp, t, *retval); hp += 2;
 		    }
@@ -1204,17 +1205,17 @@ int erts_is_time_break(Process *p, BeamInstr *pc, Eterm *retval) {
 BeamInstr *
 erts_find_local_func(Eterm mfa[3]) {
     Module *modp;
-    BeamInstr** code_base;
+    BeamCodeHeader* code_hdr;
     BeamInstr* code_ptr;
     Uint i,n;
 
     if ((modp = erts_get_module(mfa[0], erts_active_code_ix())) == NULL)
 	return NULL;
-    if ((code_base = (BeamInstr **) modp->curr.code) == NULL)
+    if ((code_hdr = modp->curr.code_hdr) == NULL)
 	return NULL;
-    n = (BeamInstr) code_base[MI_NUM_FUNCTIONS];
+    n = (BeamInstr) code_hdr->num_functions;
     for (i = 0; i < n; ++i) {
-	code_ptr = code_base[MI_FUNCTIONS+i];
+	code_ptr = code_hdr->functions[i];
 	ASSERT(((BeamInstr) BeamOp(op_i_func_info_IaaI)) == code_ptr[0]);
 	ASSERT(mfa[0] == ((Eterm) code_ptr[2]) ||
 	       is_nil((Eterm) code_ptr[2]));
@@ -1266,8 +1267,7 @@ static void bp_hash_rehash(bp_time_hash_t *hash, Uint n) {
 	    }
 	    item[hval].pid     = hash->item[ix].pid;
 	    item[hval].count   = hash->item[ix].count;
-	    item[hval].s_time  = hash->item[ix].s_time;
-	    item[hval].us_time = hash->item[ix].us_time;
+	    item[hval].time    = hash->item[ix].time;
 	}
     }
 
@@ -1315,8 +1315,7 @@ static ERTS_INLINE bp_data_time_item_t * bp_hash_put(bp_time_hash_t *hash, bp_da
     item = &(hash->item[hval]);
 
     item->pid     = sitem->pid;
-    item->s_time  = sitem->s_time;
-    item->us_time = sitem->us_time;
+    item->time    = sitem->time;
     item->count   = sitem->count;
     hash->used++;
 
@@ -1330,41 +1329,7 @@ static void bp_hash_delete(bp_time_hash_t *hash) {
     hash->item = NULL;
 }
 
-static void bp_time_diff(bp_data_time_item_t *item, /* out */
-	process_breakpoint_time_t *pbt,             /* in  */
-	Uint ms, Uint s, Uint us) {
-    int ds,dus;
-#ifdef DEBUG
-    int dms;
-
-
-    dms = ms - pbt->ms;
-#endif
-    ds  = s  - pbt->s;
-    dus = us - pbt->us;
-
-    /* get_sys_now may return zero difftime,
-     * this is ok.
-     */
-
-#ifdef DEBUG
-    ASSERT(dms >= 0 || ds >= 0 || dus >= 0);
-#endif
-
-    if (dus < 0) {
-	dus += 1000000;
-	ds  -= 1;
-    }
-    if (ds < 0) {
-	ds += 1000000;
-    }
-
-    item->s_time  = ds;
-    item->us_time = dus;
-}
-
 void erts_schedule_time_break(Process *p, Uint schedule) {
-    Uint ms, s, us;
     process_breakpoint_time_t *pbt = NULL;
     bp_data_time_item_t sitem, *item = NULL;
     bp_time_hash_t *h = NULL;
@@ -1387,8 +1352,7 @@ void erts_schedule_time_break(Process *p, Uint schedule) {
 
 	    pbdt = get_time_break(pbt->pc);
 	    if (pbdt) {
-		get_sys_now(&ms,&s,&us);
-		bp_time_diff(&sitem, pbt, ms, s, us);
+		sitem.time = get_mtime(p) - pbt->time;
 		sitem.pid   = p->common.id;
 		sitem.count = 0;
 
@@ -1410,10 +1374,7 @@ void erts_schedule_time_break(Process *p, Uint schedule) {
 	     * timestamp it and remove the previous
 	     * timestamp in the psd.
 	     */
-	    get_sys_now(&ms,&s,&us);
-	    pbt->ms = ms;
-	    pbt->s  = s;
-	    pbt->us = us;
+	    pbt->time = get_mtime(p);
 	    break;
 	default :
 	    ASSERT(0);
diff --git a/erts/emulator/beam/beam_bp.h b/erts/emulator/beam/beam_bp.h
index 97d0539ac7..2b89d6fc71 100644
--- a/erts/emulator/beam/beam_bp.h
+++ b/erts/emulator/beam/beam_bp.h
@@ -29,8 +29,7 @@
 typedef struct {
     Eterm pid;
     Sint  count;
-    Uint  s_time;
-    Uint  us_time;
+    ErtsMonotonicTime time;
 } bp_data_time_item_t;
 
 typedef struct {
@@ -46,9 +45,7 @@ typedef struct bp_data_time {     /* Call time */
 } BpDataTime;
 
 typedef struct {
-    Uint ms;
-    Uint s;
-    Uint us;
+    ErtsMonotonicTime time;
     BeamInstr *pc;
 } process_breakpoint_time_t; /* used within psd */
 
diff --git a/erts/emulator/beam/beam_debug.c b/erts/emulator/beam/beam_debug.c
index c756de8c8e..e37bd4d78c 100644
--- a/erts/emulator/beam/beam_debug.c
+++ b/erts/emulator/beam/beam_debug.c
@@ -73,13 +73,47 @@ erts_debug_flat_size_1(BIF_ALIST_1)
     }
 }
 
+BIF_RETTYPE
+erts_debug_size_shared_1(BIF_ALIST_1)
+{
+    Process* p = BIF_P;
+    Eterm term = BIF_ARG_1;
+    Uint size = size_shared(term);
+
+    if (IS_USMALL(0, size)) {
+	BIF_RET(make_small(size));
+    } else {
+	Eterm* hp = HAlloc(p, BIG_UINT_HEAP_SIZE);
+	BIF_RET(uint_to_big(size, hp));
+    }
+}
+
+BIF_RETTYPE
+erts_debug_copy_shared_1(BIF_ALIST_1)
+{
+    Process* p = BIF_P;
+    Eterm term = BIF_ARG_1;
+    Uint size;
+    Eterm* hp;
+    Eterm copy;
+    erts_shcopy_t info;
+    INITIALIZE_SHCOPY(info);
+
+    size = copy_shared_calculate(term, &info);
+    if (size > 0) {
+      hp = HAlloc(p, size);
+    }
+    copy = copy_shared_perform(term, size, &info, &hp, &p->off_heap);
+    DESTROY_SHCOPY(info);
+    BIF_RET(copy);
+}
 
 BIF_RETTYPE
 erts_debug_breakpoint_2(BIF_ALIST_2)
 {
     Process* p = BIF_P;
     Eterm MFA = BIF_ARG_1;
-    Eterm bool = BIF_ARG_2;
+    Eterm boolean = BIF_ARG_2;
     Eterm* tp;
     Eterm mfa[3];
     int i;
@@ -87,7 +121,7 @@ erts_debug_breakpoint_2(BIF_ALIST_2)
     Eterm res;
     BpFunctions f;
 
-    if (bool != am_true && bool != am_false)
+    if (boolean != am_true && boolean != am_false)
 	goto error;
 
     if (is_not_tuple(MFA)) {
@@ -124,7 +158,7 @@ erts_debug_breakpoint_2(BIF_ALIST_2)
     erts_smp_thr_progress_block();
 
     erts_bp_match_functions(&f, mfa, specified);
-    if (bool == am_true) {
+    if (boolean == am_true) {
 	erts_set_debug_break(&f);
 	erts_install_breakpoints(&f);
 	erts_commit_staged_bp();
@@ -208,7 +242,7 @@ erts_debug_disassemble_1(BIF_ALIST_1)
     Eterm bin;
     Eterm mfa;
     BeamInstr* funcinfo = NULL;	/* Initialized to eliminate warning. */
-    BeamInstr* code_base;
+    BeamCodeHeader* code_hdr;
     BeamInstr* code_ptr = NULL;	/* Initialized to eliminate warning. */
     BeamInstr instr;
     BeamInstr uaddr;
@@ -258,12 +292,12 @@ erts_debug_disassemble_1(BIF_ALIST_1)
 	     */
 	    code_ptr = ((BeamInstr *) ep->addressv[code_ix]) - 5;
 	    funcinfo = code_ptr+2;
-	} else if (modp == NULL || (code_base = modp->curr.code) == NULL) {
+	} else if (modp == NULL || (code_hdr = modp->curr.code_hdr) == NULL) {
 	    BIF_RET(am_undef);
 	} else {
-	    n = code_base[MI_NUM_FUNCTIONS];
+	    n = code_hdr->num_functions;
 	    for (i = 0; i < n; i++) {
-		code_ptr = (BeamInstr *) code_base[MI_FUNCTIONS+i];
+		code_ptr = code_hdr->functions[i];
 		if (code_ptr[3] == name && code_ptr[4] == arity) {
 		    funcinfo = code_ptr+2;
 		    break;
@@ -298,8 +332,8 @@ erts_debug_disassemble_1(BIF_ALIST_1)
     (void) erts_bld_uword(NULL, &hsz, (BeamInstr) code_ptr);
     hp = HAlloc(p, hsz);
     addr = erts_bld_uword(&hp, NULL, (BeamInstr) code_ptr);
-    ASSERT(is_atom(funcinfo[0]));
-    ASSERT(is_atom(funcinfo[1]));
+    ASSERT(is_atom(funcinfo[0]) || funcinfo[0] == NIL);
+    ASSERT(is_atom(funcinfo[1]) || funcinfo[1] == NIL);
     mfa = TUPLE3(hp, (Eterm) funcinfo[0], (Eterm) funcinfo[1], make_small((Eterm) funcinfo[2]));
     hp += 4;
     return TUPLE3(hp, addr, bin, mfa);
@@ -432,39 +466,33 @@ print_op(int to, void *to_arg, int op, int size, BeamInstr* addr)
     while (*sign) {
 	switch (*sign) {
 	case 'r':		/* x(0) */
-	    erts_print(to, to_arg, "x(0)");
+	    erts_print(to, to_arg, "r(0)");
 	    break;
 	case 'x':		/* x(N) */
-	    if (reg_index(ap[0]) == 0) {
-		erts_print(to, to_arg, "x[0]");
-	    } else {
-		erts_print(to, to_arg, "x(%d)", reg_index(ap[0]));
+	    {
+		Uint n = ap[0] / sizeof(Eterm);
+		erts_print(to, to_arg, "x(%d)", n);
+		ap++;
 	    }
-	    ap++;
 	    break;
 	case 'y':		/* y(N) */
-	    erts_print(to, to_arg, "y(%d)", reg_index(ap[0]) - CP_SIZE);
-	    ap++;
+	    {
+		Uint n = ap[0] / sizeof(Eterm) - CP_SIZE;
+		erts_print(to, to_arg, "y(%d)", n);
+		ap++;
+	    }
 	    break;
 	case 'n':		/* Nil */
 	    erts_print(to, to_arg, "[]");
 	    break;
 	case 's':		/* Any source (tagged constant or register) */
-	    tag = beam_reg_tag(*ap);
-	    if (tag == X_REG_DEF) {
-		if (reg_index(*ap) == 0) {
-		    erts_print(to, to_arg, "x[0]");
-		} else {
-		    erts_print(to, to_arg, "x(%d)", reg_index(*ap));
-		}
+	    tag = loader_tag(*ap);
+	    if (tag == LOADER_X_REG) {
+		erts_print(to, to_arg, "x(%d)", loader_x_reg_index(*ap));
 		ap++;
 		break;
-	    } else if (tag == Y_REG_DEF) {
-		erts_print(to, to_arg, "y(%d)", reg_index(*ap) - CP_SIZE);
-		ap++;
-		break;
-	    } else if (tag == R_REG_DEF) {
-		erts_print(to, to_arg, "x(0)");
+	    } else if (tag == LOADER_Y_REG) {
+		erts_print(to, to_arg, "y(%d)", loader_y_reg_index(*ap) - CP_SIZE);
 		ap++;
 		break;
 	    }
@@ -481,20 +509,12 @@ print_op(int to, void *to_arg, int op, int size, BeamInstr* addr)
 	    ap++;
 	    break;
 	case 'd':		/* Destination (x(0), x(N), y(N)) */
-	    switch (beam_reg_tag(*ap)) {
-	    case X_REG_DEF:
-		if (reg_index(*ap) == 0) {
-		    erts_print(to, to_arg, "x[0]");
-		} else {
-		    erts_print(to, to_arg, "x(%d)", reg_index(*ap));
-		}
-		break;
-	    case Y_REG_DEF:
-		erts_print(to, to_arg, "y(%d)", reg_index(*ap) - CP_SIZE);
-		break;
-	    case R_REG_DEF:
-		erts_print(to, to_arg, "x(0)");
-		break;
+	    if (*ap & 1) {
+		erts_print(to, to_arg, "y(%d)",
+			   *ap / sizeof(Eterm) - CP_SIZE);
+	    } else {
+		erts_print(to, to_arg, "x(%d)",
+			   *ap / sizeof(Eterm));
 	    }
 	    ap++;
 	    break;
@@ -561,7 +581,7 @@ print_op(int to, void *to_arg, int op, int size, BeamInstr* addr)
 	    ap++;
 	    break;
 	case 'l':		/* fr(N) */
-	    erts_print(to, to_arg, "fr(%d)", reg_index(ap[0]));
+	    erts_print(to, to_arg, "fr(%d)", loader_reg_index(ap[0]));
 	    ap++;
 	    break;
 	default:
@@ -580,7 +600,6 @@ print_op(int to, void *to_arg, int op, int size, BeamInstr* addr)
     unpacked = ap;
     ap = addr + size;
     switch (op) {
-    case op_i_select_val_lins_rfI:
     case op_i_select_val_lins_xfI:
     case op_i_select_val_lins_yfI:
 	{
@@ -600,7 +619,6 @@ print_op(int to, void *to_arg, int op, int size, BeamInstr* addr)
 	    }
 	}
 	break;
-    case op_i_select_val_bins_rfI:
     case op_i_select_val_bins_xfI:
     case op_i_select_val_bins_yfI:
 	{
@@ -614,7 +632,6 @@ print_op(int to, void *to_arg, int op, int size, BeamInstr* addr)
 	    }
 	}
 	break;
-    case op_i_select_tuple_arity_rfI:
     case op_i_select_tuple_arity_xfI:
     case op_i_select_tuple_arity_yfI:
         {
@@ -639,7 +656,6 @@ print_op(int to, void *to_arg, int op, int size, BeamInstr* addr)
             }
         }
         break;
-    case op_i_jump_on_val_rfII:
     case op_i_jump_on_val_xfII:
     case op_i_jump_on_val_yfII:
 	{
@@ -651,7 +667,6 @@ print_op(int to, void *to_arg, int op, int size, BeamInstr* addr)
 	    }
 	}
 	break;
-    case op_i_jump_on_val_zero_rfI:
     case op_i_jump_on_val_zero_xfI:
     case op_i_jump_on_val_zero_yfI:
 	{
@@ -663,29 +678,47 @@ print_op(int to, void *to_arg, int op, int size, BeamInstr* addr)
 	    }
 	}
 	break;
-    case op_i_put_tuple_rI:
     case op_i_put_tuple_xI:
     case op_i_put_tuple_yI:
     case op_new_map_dII:
     case op_update_map_assoc_jsdII:
     case op_update_map_exact_jsdII:
-    case op_i_get_map_elements_fsI:
 	{
 	    int n = unpacked[-1];
 
 	    while (n > 0) {
-		if (!is_header(ap[0])) {
+		switch (loader_tag(ap[0])) {
+		case LOADER_X_REG:
+		    erts_print(to, to_arg, " x(%d)", loader_x_reg_index(ap[0]));
+		    break;
+		case LOADER_Y_REG:
+		    erts_print(to, to_arg, " x(%d)", loader_y_reg_index(ap[0]));
+		    break;
+		default:
 		    erts_print(to, to_arg, " %T", (Eterm) ap[0]);
+		    break;
+		}
+		ap++, size++, n--;
+	    }
+	}
+	break;
+    case op_i_get_map_elements_fsI:
+	{
+	    int n = unpacked[-1];
+
+	    while (n > 0) {
+		if (n % 3 == 1) {
+		    erts_print(to, to_arg, " %X", ap[0]);
 		} else {
-		    switch ((ap[0] >> 2) & 0x03) {
-		    case R_REG_DEF:
-			erts_print(to, to_arg, " x(0)");
+		    switch (loader_tag(ap[0])) {
+		    case LOADER_X_REG:
+			erts_print(to, to_arg, " x(%d)", loader_x_reg_index(ap[0]));
 			break;
-		    case X_REG_DEF:
-			erts_print(to, to_arg, " x(%d)", ap[0] >> 4);
+		    case LOADER_Y_REG:
+			erts_print(to, to_arg, " y(%d)", loader_y_reg_index(ap[0]));
 			break;
-		    case Y_REG_DEF:
-			erts_print(to, to_arg, " y(%d)", ap[0] >> 4);
+		    default:
+			erts_print(to, to_arg, " %T", (Eterm) ap[0]);
 			break;
 		    }
 		}
diff --git a/erts/emulator/beam/beam_emu.c b/erts/emulator/beam/beam_emu.c
index 8b409e139b..4d7b00b032 100644
--- a/erts/emulator/beam/beam_emu.c
+++ b/erts/emulator/beam/beam_emu.c
@@ -99,10 +99,7 @@ do {									\
 do {                                     \
   int i_;                                \
   int Arity_ = PC[-1];                   \
-  if (Arity_ > 0) {                      \
-	CHECK_TERM(r(0));                \
-  }                                      \
-  for (i_ = 1; i_ < Arity_; i_++) {      \
+  for (i_ = 0; i_ < Arity_; i_++) {      \
 	CHECK_TERM(x(i_));               \
   }                                      \
 } while (0)
@@ -151,25 +148,21 @@ do {                                     \
    ASSERT(VALID_INSTR(* (Eterm *)(ip))); \
    I = (ip)
 
-#define FetchArgs(S1, S2) tmp_arg1 = (S1); tmp_arg2 = (S2)
+/*
+ * Register target (X or Y register).
+ */
+#define REG_TARGET(Target) (*(((Target) & 1) ? &yb(Target-1) : &xb(Target)))
 
 /*
  * Store a result into a register given a destination descriptor.
  */
 
-#define StoreResult(Result, DestDesc)               \
-  do {                                              \
-    Eterm stb_reg;                                  \
-    stb_reg = (DestDesc);                           \
-    CHECK_TERM(Result);                             \
-    switch (beam_reg_tag(stb_reg)) {                \
-    case R_REG_DEF:                                 \
-      r(0) = (Result); break;                       \
-    case X_REG_DEF:                                 \
-      xb(x_reg_offset(stb_reg)) = (Result); break;  \
-    default:                                        \
-      yb(y_reg_offset(stb_reg)) = (Result); break;  \
-    }                                               \
+#define StoreResult(Result, DestDesc)		\
+  do {						\
+    Eterm stb_reg;				\
+    stb_reg = (DestDesc);			\
+    CHECK_TERM(Result);				\
+    REG_TARGET(stb_reg) = (Result);		\
   } while (0)
 
 #define StoreSimpleDest(Src, Dest) Dest = (Src)
@@ -180,22 +173,16 @@ do {                                     \
  * be just before the next instruction.
  */
  
-#define StoreBifResult(Dst, Result)                          \
-  do {                                                       \
-    BeamInstr* stb_next;                                         \
-    Eterm stb_reg;                                           \
-    stb_reg = Arg(Dst);                                      \
-    I += (Dst) + 2;                                          \
-    stb_next = (BeamInstr *) *I;                                 \
-    CHECK_TERM(Result);                                      \
-    switch (beam_reg_tag(stb_reg)) {                         \
-    case R_REG_DEF:                                          \
-      r(0) = (Result); Goto(stb_next);                       \
-    case X_REG_DEF:                                          \
-      xb(x_reg_offset(stb_reg)) = (Result); Goto(stb_next);  \
-    default:                                                 \
-      yb(y_reg_offset(stb_reg)) = (Result); Goto(stb_next);  \
-    }                                                        \
+#define StoreBifResult(Dst, Result)		\
+  do {						\
+    BeamInstr* stb_next;			\
+    Eterm stb_reg;				\
+    stb_reg = Arg(Dst);				\
+    I += (Dst) + 2;				\
+    stb_next = (BeamInstr *) *I;		\
+    CHECK_TERM(Result);				\
+    REG_TARGET(stb_reg) = (Result);		\
+    Goto(stb_next);				\
   } while (0)
 
 #define ClauseFail() goto jump_f
@@ -250,6 +237,14 @@ void** beam_ops;
     HEAP_TOP(c_p) = HTOP;  \
     c_p->stop = E
 
+#define HEAVY_SWAPIN       \
+    SWAPIN;		   \
+    FCALLS = c_p->fcalls
+
+#define HEAVY_SWAPOUT      \
+    SWAPOUT;		   \
+    c_p->fcalls = FCALLS
+
 /*
  * Use LIGHT_SWAPOUT when the called function
  * will call HeapOnlyAlloc() (and never HAlloc()).
@@ -293,7 +288,7 @@ void** beam_ops;
 #define Ib(N) (N)
 #define x(N) reg[N]
 #define y(N) E[N]
-#define r(N) x##N
+#define r(N) x(N)
 
 /*
  * Makes sure that there are StackNeed + HeapNeed + 1 words available
@@ -309,12 +304,10 @@ void** beam_ops;
      needed = (StackNeed) + 1; \
      if (E - HTOP < (needed + (HeapNeed))) { \
            SWAPOUT; \
-           reg[0] = r(0); \
            PROCESS_MAIN_CHK_LOCKS(c_p); \
-           FCALLS -= erts_garbage_collect(c_p, needed + (HeapNeed), reg, (M)); \
+           FCALLS -= erts_garbage_collect_nobump(c_p, needed + (HeapNeed), reg, (M)); \
            ERTS_VERIFY_UNUSED_TEMP_ALLOC(c_p); \
            PROCESS_MAIN_CHK_LOCKS(c_p); \
-           r(0) = reg[0]; \
            SWAPIN; \
      } \
      E -= needed; \
@@ -363,12 +356,10 @@ void** beam_ops;
     unsigned need = (Nh);                                       		\
     if ((E - HTOP < need) || (MSO(c_p).overhead + (VNh) >= BIN_VHEAP_SZ(c_p))) {\
        SWAPOUT;                                                 		\
-       reg[0] = r(0);                                           		\
        PROCESS_MAIN_CHK_LOCKS(c_p);                             		\
-       FCALLS -= erts_garbage_collect(c_p, need, reg, (Live));  		\
+       FCALLS -= erts_garbage_collect_nobump(c_p, need, reg, (Live));  		\
        ERTS_VERIFY_UNUSED_TEMP_ALLOC(c_p);					\
        PROCESS_MAIN_CHK_LOCKS(c_p);                             		\
-       r(0) = reg[0];                                           		\
        SWAPIN;                                                  		\
     }                                                           		\
     HEAP_SPACE_VERIFIED(need);                                                  \
@@ -386,12 +377,10 @@ void** beam_ops;
     unsigned need = (Nh);                                       \
     if (E - HTOP < need) {                                      \
        SWAPOUT;                                                 \
-       reg[0] = r(0);                                           \
        PROCESS_MAIN_CHK_LOCKS(c_p);                             \
-       FCALLS -= erts_garbage_collect(c_p, need, reg, (Live));  \
+       FCALLS -= erts_garbage_collect_nobump(c_p, need, reg, (Live));\
        ERTS_VERIFY_UNUSED_TEMP_ALLOC(c_p);			\
        PROCESS_MAIN_CHK_LOCKS(c_p);                             \
-       r(0) = reg[0];                                           \
        SWAPIN;                                                  \
     }                                                           \
     HEAP_SPACE_VERIFIED(need);                             \
@@ -408,15 +397,11 @@ void** beam_ops;
     unsigned need = (Nh);						\
     if (E - HTOP < need) {						\
        SWAPOUT;								\
-       reg[0] = r(0);							\
        reg[Live] = Extra;						\
        PROCESS_MAIN_CHK_LOCKS(c_p);					\
-       FCALLS -= erts_garbage_collect(c_p, need, reg, (Live)+1);	\
+       FCALLS -= erts_garbage_collect_nobump(c_p, need, reg, (Live)+1);	\
        ERTS_VERIFY_UNUSED_TEMP_ALLOC(c_p);				\
        PROCESS_MAIN_CHK_LOCKS(c_p);					\
-       if (Live > 0) {							\
-	   r(0) = reg[0];						\
-       }								\
        Extra = reg[Live];						\
        SWAPIN;								\
     }									\
@@ -438,10 +423,9 @@ void** beam_ops;
 
 #define MakeFun(FunP, NumFree)					\
   do {								\
-     SWAPOUT;							\
-     reg[0] = r(0);						\
+     HEAVY_SWAPOUT;						\
      r(0) = new_fun(c_p, reg, (ErlFunEntry *) FunP, NumFree);	\
-     SWAPIN;							\
+     HEAVY_SWAPIN;						\
   } while (0)
 
 #define PutTuple(Dst, Arity)			\
@@ -531,15 +515,19 @@ void** beam_ops;
     ASSERT(VALID_INSTR(Dst));  \
     Goto(Dst)
 
-#define GetR(pos, tr) \
-   do { \
-     tr = Arg(pos); \
-     switch (beam_reg_tag(tr)) { \
-     case R_REG_DEF: tr = r(0); break; \
-     case X_REG_DEF: tr = xb(x_reg_offset(tr)); break; \
-     case Y_REG_DEF: ASSERT(y_reg_offset(tr) >= 1); tr = yb(y_reg_offset(tr)); break; \
-     } \
-     CHECK_TERM(tr); \
+#define GetR(pos, tr)				\
+   do {						\
+     tr = Arg(pos);				\
+     switch (loader_tag(tr)) {			\
+     case LOADER_X_REG:				\
+        tr = x(loader_x_reg_index(tr));		\
+        break;					\
+     case LOADER_Y_REG:				\
+        ASSERT(loader_y_reg_index(tr) >= 1);	\
+        tr = y(loader_y_reg_index(tr));		\
+        break;					\
+     }						\
+     CHECK_TERM(tr);				\
    } while (0)
 
 #define GetArg1(N, Dst) GetR((N), Dst)
@@ -557,24 +545,93 @@ void** beam_ops;
    HTOP += 2;                      \
   } while (0)
 
+#define Swap(R1, R2)				\
+  do {						\
+    Eterm V = R1;				\
+    R1 = R2;					\
+    R2 = V;					\
+  } while (0)
+
+#define SwapTemp(R1, R2, Tmp)			\
+  do {						\
+    Eterm V = R1;				\
+    R1 = R2;					\
+    R2 = Tmp = V;				\
+  } while (0)
+
 #define Move(Src, Dst, Store)      \
    do {                            \
        Eterm term = (Src);         \
        Store(term, Dst);           \
    } while (0)
 
-#define Move2(S1, D1, S2, D2) D1 = (S1); D2 = (S2)
+#define Move2Par(S1, D1, S2, D2)		\
+  do {						\
+      Eterm V1, V2;				\
+      V1 = (S1); V2 = (S2); D1 = V1; D2 = V2;	\
+  } while (0)
+
+#define MoveShift(Src, SD, D)			\
+  do {						\
+    Eterm V;					\
+    V = Src; D = SD; SD = V;			\
+  } while (0)
+
+#define MoveDup(Src, D1, D2)			\
+  do {						\
+    D1 = D2 = (Src);				\
+  } while (0)
+
 #define Move3(S1, D1, S2, D2, S3, D3) D1 = (S1); D2 = (S2); D3 = (S3)
 
-#define MoveGenDest(src, dstp) \
-   if ((dstp) == NULL) { r(0) = (src); } else { *(dstp) = src; }
+#define MoveWindow3(S1, S2, S3, D)		\
+  do {						\
+      Eterm xt0, xt1, xt2;			\
+      Eterm *y = &D;				\
+      xt0  = S1;				\
+      xt1  = S2;				\
+      xt2  = S3;				\
+      y[0] = xt0;				\
+      y[1] = xt1;				\
+      y[2] = xt2;				\
+ } while (0)
 
-#define MoveReturn(Src, Dest)       \
-    (Dest) = (Src);                 \
-    I = c_p->cp;                    \
-    ASSERT(VALID_INSTR(*c_p->cp));  \
-    c_p->cp = 0;                    \
-    CHECK_TERM(r(0));               \
+#define MoveWindow4(S1, S2, S3, S4, D)		\
+  do {						\
+      Eterm xt0, xt1, xt2, xt3;			\
+      Eterm *y = &D;				\
+      xt0  = S1;				\
+      xt1  = S2;				\
+      xt2  = S3;				\
+      xt3  = S4;				\
+      y[0] = xt0;				\
+      y[1] = xt1;				\
+      y[2] = xt2;				\
+      y[3] = xt3;				\
+ } while (0)
+
+#define MoveWindow5(S1, S2, S3, S4, S5, D)	\
+  do {						\
+      Eterm xt0, xt1, xt2, xt3, xt4;		\
+      Eterm *y = &D;				\
+      xt0  = S1;				\
+      xt1  = S2;				\
+      xt2  = S3;				\
+      xt3  = S4;				\
+      xt4  = S5;				\
+      y[0] = xt0;				\
+      y[1] = xt1;				\
+      y[2] = xt2;				\
+      y[3] = xt3;				\
+      y[4] = xt4;				\
+ } while (0)
+
+#define MoveReturn(Src)				\
+    x(0) = (Src);				\
+    I = c_p->cp;				\
+    ASSERT(VALID_INSTR(*c_p->cp));		\
+    c_p->cp = 0;				\
+    CHECK_TERM(r(0));				\
     Goto(*I)
 
 #define DeallocateReturn(Deallocate)       \
@@ -586,26 +643,26 @@ void** beam_ops;
     Goto(*I);                              \
   } while (0)
 
-#define MoveDeallocateReturn(Src, Dest, Deallocate)  \
-    (Dest) = (Src);                                  \
+#define MoveDeallocateReturn(Src, Deallocate)	\
+    x(0) = (Src);				\
     DeallocateReturn(Deallocate)
 
-#define MoveCall(Src, Dest, CallDest, Size)	\
-    (Dest) = (Src);				\
+#define MoveCall(Src, CallDest, Size)		\
+    x(0) = (Src);				\
     SET_CP(c_p, I+Size+1);			\
-    SET_I((BeamInstr *) CallDest);			\
+    SET_I((BeamInstr *) CallDest);		\
     Dispatch();
 
-#define MoveCallLast(Src, Dest, CallDest, Deallocate)	\
-    (Dest) = (Src);					\
-    RESTORE_CP(E);					\
-    E = ADD_BYTE_OFFSET(E, (Deallocate));		\
-    SET_I((BeamInstr *) CallDest);				\
+#define MoveCallLast(Src, CallDest, Deallocate)	\
+    x(0) = (Src);				\
+    RESTORE_CP(E);				\
+    E = ADD_BYTE_OFFSET(E, (Deallocate));	\
+    SET_I((BeamInstr *) CallDest);		\
     Dispatch();
 
-#define MoveCallOnly(Src, Dest, CallDest)	\
-    (Dest) = (Src);				\
-    SET_I((BeamInstr *) CallDest);			\
+#define MoveCallOnly(Src, CallDest)		\
+    x(0) = (Src);				\
+    SET_I((BeamInstr *) CallDest);		\
     Dispatch();
 
 #define MoveJump(Src)				\
@@ -613,58 +670,56 @@ void** beam_ops;
      SET_I((BeamInstr *) Arg(0));		\
      Goto(*I);
 
-#define GetList(Src, H, T) do {			\
-   Eterm* tmp_ptr = list_val(Src);		\
-   H = CAR(tmp_ptr);				\
-   T = CDR(tmp_ptr); } while (0)
-
-#define GetTupleElement(Src, Element, Dest)				\
-  do {									\
-    tmp_arg1 = (Eterm) COMPRESS_POINTER(((unsigned char *) tuple_val(Src)) + 	\
-				(Element));				\
-    (Dest) = (*(Eterm *) EXPAND_POINTER(tmp_arg1));			\
-  } while (0)
-
-#define ExtractNextElement(Dest)					  \
-    tmp_arg1 += sizeof(Eterm);						  \
-    (Dest) = (* (Eterm *) (((unsigned char *) EXPAND_POINTER(tmp_arg1))))
-
-#define ExtractNextElement2(Dest)				\
-  do {								\
-    Eterm* ene_dstp = &(Dest);					\
-    ene_dstp[0] = ((Eterm *) EXPAND_POINTER(tmp_arg1))[1];	\
-    ene_dstp[1] = ((Eterm *) EXPAND_POINTER(tmp_arg1))[2];	\
-    tmp_arg1 += sizeof(Eterm) + sizeof(Eterm);			\
+#define GetList(Src, H, T)			\
+  do {						\
+    Eterm* tmp_ptr = list_val(Src);		\
+    Eterm hd, tl;				\
+    hd = CAR(tmp_ptr);				\
+    tl = CDR(tmp_ptr);				\
+    H = hd; T = tl;				\
   } while (0)
 
-#define ExtractNextElement3(Dest)		\
-  do {						\
-    Eterm* ene_dstp = &(Dest);			\
-    ene_dstp[0] = ((Eterm *) EXPAND_POINTER(tmp_arg1))[1];	\
-    ene_dstp[1] = ((Eterm *) EXPAND_POINTER(tmp_arg1))[2];	\
-    ene_dstp[2] = ((Eterm *) EXPAND_POINTER(tmp_arg1))[3];	\
-    tmp_arg1 += 3*sizeof(Eterm);		\
+#define GetTupleElement(Src, Element, Dest)		\
+  do {							\
+    Eterm* src;						\
+    src = ADD_BYTE_OFFSET(tuple_val(Src), (Element));	\
+    (Dest) = *src;					\
   } while (0)
 
-#define ExtractNextElement4(Dest)		\
-  do {						\
-    Eterm* ene_dstp = &(Dest);			\
-    ene_dstp[0] = ((Eterm *) EXPAND_POINTER(tmp_arg1))[1];	\
-    ene_dstp[1] = ((Eterm *) EXPAND_POINTER(tmp_arg1))[2];	\
-    ene_dstp[2] = ((Eterm *) EXPAND_POINTER(tmp_arg1))[3];	\
-    ene_dstp[3] = ((Eterm *) EXPAND_POINTER(tmp_arg1))[4];	\
-    tmp_arg1 += 4*sizeof(Eterm);		\
+#define GetTupleElement2(Src, Element, Dest)		\
+  do {							\
+    Eterm* src;						\
+    Eterm* dst;						\
+    Eterm E1, E2;					\
+    src = ADD_BYTE_OFFSET(tuple_val(Src), (Element));	\
+    dst = &(Dest);					\
+    E1 = src[0];					\
+    E2 = src[1];					\
+    dst[0] = E1;					\
+    dst[1] = E2;					\
   } while (0)
 
-#define ExtractElement(Element, Dest)		\
-  do {						\
-     tmp_arg1 += (Element);			\
-     (Dest) = (* (Eterm *) EXPAND_POINTER(tmp_arg1));		\
+#define GetTupleElement3(Src, Element, Dest)		\
+  do {							\
+    Eterm* src;						\
+    Eterm* dst;						\
+    Eterm E1, E2, E3;					\
+    src = ADD_BYTE_OFFSET(tuple_val(Src), (Element));	\
+    dst = &(Dest);					\
+    E1 = src[0];					\
+    E2 = src[1];					\
+    E3 = src[2];					\
+    dst[0] = E1;					\
+    dst[1] = E2;					\
+    dst[2] = E3;					\
   } while (0)
 
 #define EqualImmed(X, Y, Action) if (X != Y) { Action; }
 #define NotEqualImmed(X, Y, Action) if (X == Y) { Action; }
 #define EqualExact(X, Y, Action) if (!EQ(X,Y)) { Action; }
+#define NotEqualExact(X, Y, Action) if (EQ(X,Y)) { Action; }
+#define Equal(X, Y, Action) if (!CMP_EQ(X,Y)) { Action; }
+#define NotEqual(X, Y, Action) if (!CMP_NE(X,Y)) { Action; }
 #define IsLessThan(X, Y, Action) if (CMP_GE(X, Y)) { Action; }
 #define IsGreaterEqual(X, Y, Action) if (CMP_LT(X, Y)) { Action; }
 
@@ -690,18 +745,26 @@ void** beam_ops;
     if (is_not_list(Src)) { Fail; }                     \
     A(Need, Alive)
 
-#define IsNonemptyListTestHeap(Src, Need, Alive, Fail)  \
-    if (is_not_list(Src)) { Fail; }                     \
+#define IsNonemptyListTestHeap(Need, Alive, Fail)  \
+    if (is_not_list(x(0))) { Fail; }			\
     TestHeap(Need, Alive)
 
+#define IsNonemptyListGetList(Src, H, T, Fail)	\
+    if (is_not_list(Src)) {			\
+         Fail;					\
+    } else {					\
+       Eterm* tmp_ptr = list_val(Src);		\
+       Eterm hd, tl;				\
+       hd = CAR(tmp_ptr);			\
+       tl = CDR(tmp_ptr);			\
+       H = hd; T = tl;				\
+    }
+
 #define IsTuple(X, Action) if (is_not_tuple(X)) Action
 
-#define IsArity(Pointer, Arity, Fail)					  \
-    if (*(Eterm *)							  \
-	EXPAND_POINTER(tmp_arg1 = (Eterm) 				  \
-		       COMPRESS_POINTER(tuple_val(Pointer))) != (Arity))  \
-    { 									  \
-        Fail; 								  \
+#define IsArity(Pointer, Arity, Fail)		\
+    if (*tuple_val(Pointer) != (Arity)) {	\
+        Fail;					\
     }
 
 #define IsMap(Src, Fail) if (!is_map(Src)) { Fail; }
@@ -738,15 +801,21 @@ void** beam_ops;
      }							\
   } while (0)
 
-#define IsTupleOfArity(Src, Arity, Fail)				      \
-  do {									      \
-    if (is_not_tuple(Src) || 						      \
-	*(Eterm *)							      \
-	EXPAND_POINTER(tmp_arg1 = 					      \
-		       (Eterm) COMPRESS_POINTER(tuple_val(Src))) != Arity) { \
-        Fail;								      \
-    }									      \
+#ifdef DEBUG
+#define IsTupleOfArity(Src, Arityval, Fail)			\
+  do {								\
+    if (!(is_tuple(Src) && *tuple_val(Src) == Arityval)) {	\
+        Fail;							\
+    }								\
+  } while (0)
+#else
+#define IsTupleOfArity(Src, Arityval, Fail)			\
+  do {								\
+    if (!(is_boxed(Src) && *tuple_val(Src) == Arityval)) {	\
+        Fail;							\
+    }								\
   } while (0)
+#endif
 
 #define IsBoolean(X, Fail) if ((X) != am_true && (X) != am_false) { Fail; }
 
@@ -756,7 +825,7 @@ void** beam_ops;
 #define IsBitstring(Src, Fail) \
   if (is_not_binary(Src)) { Fail; }
 
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
 #define BsSafeMul(A, B, Fail, Target)		\
    do { Uint64 _res = (A) * (B);		\
       if (_res / B != A) { Fail; }		\
@@ -773,6 +842,7 @@ void** beam_ops;
 #define BsGetFieldSize(Bits, Unit, Fail, Target)	\
    do {							\
       Sint _signed_size; Uint _uint_size;		\
+      Uint temp_bits;					\
       if (is_small(Bits)) {				\
         _signed_size = signed_val(Bits);		\
          if (_signed_size < 0) { Fail; }		\
@@ -787,6 +857,7 @@ void** beam_ops;
 #define BsGetUncheckedFieldSize(Bits, Unit, Fail, Target)	\
    do {								\
       Sint _signed_size; Uint _uint_size;			\
+      Uint temp_bits;						\
       if (is_small(Bits)) {					\
         _signed_size = signed_val(Bits);			\
          if (_signed_size < 0) { Fail; }			\
@@ -998,23 +1069,17 @@ init_emulator(void)
  */
 
 #if defined(__GNUC__) && defined(sparc) && !defined(DEBUG)
-#  define REG_x0 asm("%l0")
 #  define REG_xregs asm("%l1")
 #  define REG_htop asm("%l2")
 #  define REG_stop asm("%l3")
 #  define REG_I asm("%l4")
 #  define REG_fcalls asm("%l5")
-#  define REG_tmp_arg1 asm("%l6")
-#  define REG_tmp_arg2 asm("%l7")
 #else
-#  define REG_x0
 #  define REG_xregs
 #  define REG_htop
 #  define REG_stop
 #  define REG_I
 #  define REG_fcalls
-#  define REG_tmp_arg1
-#  define REG_tmp_arg2
 #endif
 
 #ifdef USE_VM_PROBES
@@ -1132,11 +1197,6 @@ void process_main(void)
     ERTS_DECLARE_DUMMY(Eterm pid);
 #endif
 
-    /*
-     * X register zero; also called r(0)
-     */
-    register Eterm x0 REG_x0 = NIL;
-
     /* Pointer to X registers: x(1)..x(N); reg[0] is used when doing GC,
      * in all other cases x0 is used.
      */
@@ -1164,17 +1224,6 @@ void process_main(void)
     register Sint FCALLS REG_fcalls = 0;
 
     /*
-     * Temporaries used for picking up arguments for instructions.
-     */
-    register Eterm tmp_arg1 REG_tmp_arg1 = NIL;
-    register Eterm tmp_arg2 REG_tmp_arg2 = NIL;
-#if HEAP_ON_C_STACK
-    Eterm tmp_big[2];           /* Temporary buffer for small bignums if HEAP_ON_C_STACK. */
-#else
-    Eterm *tmp_big;		/* Temporary buffer for small bignums if !HEAP_ON_C_STACK. */
-#endif
-
-    /*
      * X registers and floating point registers are located in
      * scheduler specific data.
      */
@@ -1185,8 +1234,6 @@ void process_main(void)
      */
     int neg_o_reds = 0;
 
-    Eterm (*arith_func)(Process* p, Eterm* reg, Uint live);
-
 #ifdef ERTS_OPCODE_COUNTER_SUPPORT
     static void* counting_opcodes[] = { DEFINE_COUNTING_OPCODES };
 #else
@@ -1197,8 +1244,6 @@ void process_main(void)
 #endif
 #endif
 
-    Uint temp_bits; /* Temporary used by BsSkipBits2 & BsGetInteger2 */
-
     Eterm pt_arity;		/* Used by do_put_tuple */
 
     Uint64 start_time = 0;          /* Monitor long schedule */
@@ -1247,9 +1292,6 @@ void process_main(void)
 
     PROCESS_MAIN_CHK_LOCKS(c_p);
     ERTS_SMP_UNREQ_PROC_MAIN_LOCK(c_p);
-#if HALFWORD_HEAP
-    ASSERT(erts_get_scheduler_data()->num_tmp_heap_used == 0);
-#endif
     ERTS_VERIFY_UNUSED_TEMP_ALLOC(c_p);
     c_p = schedule(c_p, reds_used);
     ERTS_VERIFY_UNUSED_TEMP_ALLOC(c_p);
@@ -1267,9 +1309,6 @@ void process_main(void)
 
     reg = ERTS_PROC_GET_SCHDATA(c_p)->x_reg_array;
     freg = ERTS_PROC_GET_SCHDATA(c_p)->f_reg_array;
-#if !HEAP_ON_C_STACK
-    tmp_big = ERTS_PROC_GET_SCHDATA(c_p)->beam_emu_tmp_heap;
-#endif
     ERL_BITS_RELOAD_STATEP(c_p);
     {
 	int reds;
@@ -1278,7 +1317,7 @@ void process_main(void)
 	int i;
 
 	argp = c_p->arg_reg;
-	for (i = c_p->arity - 1; i > 0; i--) {
+	for (i = c_p->arity - 1; i >= 0; i--) {
 	    reg[i] = argp[i];
 	    CHECK_TERM(reg[i]);
 	}
@@ -1302,12 +1341,6 @@ void process_main(void)
 	}
 
 	next = (BeamInstr *) *I;
-	r(0) = c_p->arg_reg[0];
-#ifdef HARDDEBUG
-	if (c_p->arity > 0) {
-	    CHECK_TERM(r(0));
-	}
-#endif
 	SWAPIN;
 	ASSERT(VALID_INSTR(next));
 
@@ -1346,26 +1379,24 @@ void process_main(void)
 #endif
 #include "beam_hot.h"
 
-#define STORE_ARITH_RESULT(res) StoreBifResult(2, (res));
-#define ARITH_FUNC(name) erts_gc_##name
-
 	{
 	    Eterm increment_reg_val;
 	    Eterm increment_val;
 	    Uint live;
 	    Eterm result;
 
-	OpCase(i_increment_yIId):
-	    increment_reg_val = yb(Arg(0));
+	OpCase(i_increment_rIId):
+	    increment_reg_val = x(0);
+	    I--;
 	    goto do_increment;
 
 	OpCase(i_increment_xIId):
 	    increment_reg_val = xb(Arg(0));
 	    goto do_increment;
 
-	OpCase(i_increment_rIId):
-	    increment_reg_val = r(0);
-	    I--;
+	OpCase(i_increment_yIId):
+	    increment_reg_val = yb(Arg(0));
+	    goto do_increment;
 
 	do_increment:
 	    increment_val = Arg(1);
@@ -1374,229 +1405,133 @@ void process_main(void)
 		ASSERT(MY_IS_SSMALL(i) == IS_SSMALL(i));
 		if (MY_IS_SSMALL(i)) {
 		    result = make_small(i);
-		store_result:
 		    StoreBifResult(3, result);
 		}
 	    }
 
 	    live = Arg(2);
-	    SWAPOUT;
-	    reg[0] = r(0);
+	    HEAVY_SWAPOUT;
 	    reg[live] = increment_reg_val;
 	    reg[live+1] = make_small(increment_val);
 	    result = erts_gc_mixed_plus(c_p, reg, live);
-	    r(0) = reg[0];
-	    SWAPIN;
+	    HEAVY_SWAPIN;
 	    ERTS_HOLE_CHECK(c_p);
 	    if (is_value(result)) {
-		goto store_result;
+		StoreBifResult(3, result);
 	    }
 	    ASSERT(c_p->freason != BADMATCH || is_value(c_p->fvalue));
 	    goto find_func_info;
 	}
 
-#define DO_BIG_ARITH(Func,Arg1,Arg2)     \
-    do {                                 \
-        Uint live = Arg(1);              \
-        SWAPOUT;                         \
-        reg[0] = r(0);                   \
-        reg[live] = (Arg1);              \
-        reg[live+1] = (Arg2);            \
-        result = (Func)(c_p, reg, live); \
-        r(0) = reg[0];                   \
-        SWAPIN;                          \
-        ERTS_HOLE_CHECK(c_p);            \
-        if (is_value(result)) {          \
-            StoreBifResult(4,result);    \
-        }                                \
-        goto lb_Cl_error;                \
-    } while(0)
+#define DO_OUTLINED_ARITH_2(name, Op1, Op2)	\
+ do {						\
+     Eterm result;				\
+     Uint live = Arg(1);			\
+						\
+     HEAVY_SWAPOUT;				\
+     reg[live] = Op1;				\
+     reg[live+1] = Op2;				\
+     result = erts_gc_##name(c_p, reg, live);	\
+     HEAVY_SWAPIN;				\
+     ERTS_HOLE_CHECK(c_p);			\
+     if (is_value(result)) {			\
+	 StoreBifResult(4, result);		\
+     }						\
+     goto lb_Cl_error;				\
+ } while (0)
 
- OpCase(i_plus_jIxxd):
  {
+     Eterm PlusOp1, PlusOp2;
      Eterm result;
 
-     if (is_both_small(xb(Arg(2)), xb(Arg(3)))) {
-	 Sint i = signed_val(xb(Arg(2))) + signed_val(xb(Arg(3)));
+ OpCase(i_plus_jIxxd):
+     PlusOp1 = xb(Arg(2));
+     PlusOp2 = xb(Arg(3));
+     goto do_plus;
+
+ OpCase(i_plus_jIxyd):
+     PlusOp1 = xb(Arg(2));
+     PlusOp2 = yb(Arg(3));
+     goto do_plus;
+
+ OpCase(i_plus_jIssd):
+     GetArg2(2, PlusOp1, PlusOp2);
+     goto do_plus;
+
+ do_plus:
+     if (is_both_small(PlusOp1, PlusOp2)) {
+	 Sint i = signed_val(PlusOp1) + signed_val(PlusOp2);
 	 ASSERT(MY_IS_SSMALL(i) == IS_SSMALL(i));
 	 if (MY_IS_SSMALL(i)) {
 	     result = make_small(i);
              StoreBifResult(4, result);
 	 }
      }
-     DO_BIG_ARITH(ARITH_FUNC(mixed_plus), xb(Arg(2)), xb(Arg(3)));
+     DO_OUTLINED_ARITH_2(mixed_plus, PlusOp1, PlusOp2);
  }
 
- OpCase(i_plus_jId):
  {
+     Eterm MinusOp1, MinusOp2;
      Eterm result;
 
-     if (is_both_small(tmp_arg1, tmp_arg2)) {
-	 Sint i = signed_val(tmp_arg1) + signed_val(tmp_arg2);
-	 ASSERT(MY_IS_SSMALL(i) == IS_SSMALL(i));
-	 if (MY_IS_SSMALL(i)) {
-	     result = make_small(i);
-	     STORE_ARITH_RESULT(result);
-	 }
-     }
-     arith_func = ARITH_FUNC(mixed_plus);
-     goto do_big_arith2;
- }
-
  OpCase(i_minus_jIxxd):
- {
-     Eterm result;
+     MinusOp1 = xb(Arg(2));
+     MinusOp2 = xb(Arg(3));
+     goto do_minus;
 
-     if (is_both_small(xb(Arg(2)), xb(Arg(3)))) {
-	 Sint i = signed_val(xb(Arg(2))) - signed_val(xb(Arg(3)));
-	 ASSERT(MY_IS_SSMALL(i) == IS_SSMALL(i));
-	 if (MY_IS_SSMALL(i)) {
-	     result = make_small(i);
-             StoreBifResult(4, result);
-	 }
-     }
-     DO_BIG_ARITH(ARITH_FUNC(mixed_minus), xb(Arg(2)), xb(Arg(3)));
- }
+ OpCase(i_minus_jIssd):
+     GetArg2(2, MinusOp1, MinusOp2);
+     goto do_minus;
 
- OpCase(i_minus_jId):
- {
-     Eterm result;
-
-     if (is_both_small(tmp_arg1, tmp_arg2)) {
-	 Sint i = signed_val(tmp_arg1) - signed_val(tmp_arg2);
+ do_minus:
+     if (is_both_small(MinusOp1, MinusOp2)) {
+	 Sint i = signed_val(MinusOp1) - signed_val(MinusOp2);
 	 ASSERT(MY_IS_SSMALL(i) == IS_SSMALL(i));
 	 if (MY_IS_SSMALL(i)) {
 	     result = make_small(i);
-	     STORE_ARITH_RESULT(result);
+             StoreBifResult(4, result);
 	 }
      }
-     arith_func = ARITH_FUNC(mixed_minus);
-     goto do_big_arith2;
+     DO_OUTLINED_ARITH_2(mixed_minus, MinusOp1, MinusOp2);
  }
 
- OpCase(i_is_lt_f):
-    if (CMP_GE(tmp_arg1, tmp_arg2)) {
-	ClauseFail();
-    }
-    Next(1);
-
- OpCase(i_is_ge_f):
-    if (CMP_LT(tmp_arg1, tmp_arg2)) {
-	ClauseFail();
-    }
-    Next(1);
-
- OpCase(i_is_eq_f):
-    if (CMP_NE(tmp_arg1, tmp_arg2)) {
-	ClauseFail();
-    }
-    Next(1);
-
- OpCase(i_is_ne_f):
-    if (CMP_EQ(tmp_arg1, tmp_arg2)) {
-	ClauseFail();
-    }
-    Next(1);
-
- OpCase(i_is_eq_exact_f):
-    if (!EQ(tmp_arg1, tmp_arg2)) {
-	ClauseFail();
-    }
-    Next(1);
-
     {
 	Eterm is_eq_exact_lit_val;
 
-    OpCase(i_is_eq_exact_literal_xfc):
-	is_eq_exact_lit_val = xb(Arg(0));
-	I++;
+    OpCase(i_is_eq_exact_literal_fxc):
+	is_eq_exact_lit_val = xb(Arg(1));
 	goto do_is_eq_exact_literal;
 
-    OpCase(i_is_eq_exact_literal_yfc):
-	is_eq_exact_lit_val = yb(Arg(0));
-	I++;
+    OpCase(i_is_eq_exact_literal_fyc):
+	is_eq_exact_lit_val = yb(Arg(1));
 	goto do_is_eq_exact_literal;
 
-    OpCase(i_is_eq_exact_literal_rfc):
-	is_eq_exact_lit_val = r(0);
-
     do_is_eq_exact_literal:
-	if (!eq(Arg(1), is_eq_exact_lit_val)) {
+	if (!eq(Arg(2), is_eq_exact_lit_val)) {
 	    ClauseFail();
 	}
-	Next(2);
+	Next(3);
     }
 
     {
 	Eterm is_ne_exact_lit_val;
 
-    OpCase(i_is_ne_exact_literal_xfc):
-	is_ne_exact_lit_val = xb(Arg(0));
-	I++;
+    OpCase(i_is_ne_exact_literal_fxc):
+	is_ne_exact_lit_val = xb(Arg(1));
 	goto do_is_ne_exact_literal;
 
-    OpCase(i_is_ne_exact_literal_yfc):
-	is_ne_exact_lit_val = yb(Arg(0));
-	I++;
+    OpCase(i_is_ne_exact_literal_fyc):
+	is_ne_exact_lit_val = yb(Arg(1));
 	goto do_is_ne_exact_literal;
 
-    OpCase(i_is_ne_exact_literal_rfc):
-	is_ne_exact_lit_val = r(0);
-
     do_is_ne_exact_literal:
-	if (eq(Arg(1), is_ne_exact_lit_val)) {
+	if (eq(Arg(2), is_ne_exact_lit_val)) {
 	    ClauseFail();
 	}
-	Next(2);
+	Next(3);
     }
 
- OpCase(move_window3_xxxy): {
-     BeamInstr *next;
-     Eterm xt0, xt1, xt2;
-     Eterm *y = (Eterm *)(((unsigned char *)E) + (Arg(3)));
-     PreFetch(4, next);
-     xt0  = xb(Arg(0));
-     xt1  = xb(Arg(1));
-     xt2  = xb(Arg(2));
-     y[0] = xt0;
-     y[1] = xt1;
-     y[2] = xt2;
-     NextPF(4, next);
- }
- OpCase(move_window4_xxxxy): {
-     BeamInstr *next;
-     Eterm xt0, xt1, xt2, xt3;
-     Eterm *y = (Eterm *)(((unsigned char *)E) + (Arg(4)));
-     PreFetch(5, next);
-     xt0  = xb(Arg(0));
-     xt1  = xb(Arg(1));
-     xt2  = xb(Arg(2));
-     xt3  = xb(Arg(3));
-     y[0] = xt0;
-     y[1] = xt1;
-     y[2] = xt2;
-     y[3] = xt3;
-     NextPF(5, next);
- }
- OpCase(move_window5_xxxxxy): {
-     BeamInstr *next;
-     Eterm xt0, xt1, xt2, xt3, xt4;
-     Eterm *y = (Eterm *)(((unsigned char *)E) + (Arg(5)));
-     PreFetch(6, next);
-     xt0  = xb(Arg(0));
-     xt1  = xb(Arg(1));
-     xt2  = xb(Arg(2));
-     xt3  = xb(Arg(3));
-     xt4  = xb(Arg(4));
-     y[0] = xt0;
-     y[1] = xt1;
-     y[2] = xt2;
-     y[3] = xt3;
-     y[4] = xt4;
-     NextPF(6, next);
- }
-
- OpCase(i_move_call_only_fcr): {
+ OpCase(i_move_call_only_fc): {
      r(0) = Arg(1);
  }
  /* FALL THROUGH */
@@ -1606,7 +1541,7 @@ void process_main(void)
      Dispatch();
  }
 
- OpCase(i_move_call_last_fPcr): {
+ OpCase(i_move_call_last_fPc): {
      r(0) = Arg(2);
  }
  /* FALL THROUGH */
@@ -1618,7 +1553,7 @@ void process_main(void)
      Dispatch();
  }
 
- OpCase(i_move_call_crf): {
+ OpCase(i_move_call_cf): {
      r(0) = Arg(0);
      I++;
  }
@@ -1630,7 +1565,7 @@ void process_main(void)
      Dispatch();
  }
 
- OpCase(i_move_call_ext_last_ePcr): {
+ OpCase(i_move_call_ext_last_ePc): {
      r(0) = Arg(2);
  }
  /* FALL THROUGH */
@@ -1646,7 +1581,7 @@ void process_main(void)
     DTRACE_GLOBAL_CALL_FROM_EXPORT(c_p, Arg(0));
     Dispatchx();
 
- OpCase(i_move_call_ext_cre): {
+ OpCase(i_move_call_ext_ce): {
      r(0) = Arg(0);
      I++;
  }
@@ -1656,7 +1591,7 @@ void process_main(void)
     DTRACE_GLOBAL_CALL_FROM_EXPORT(c_p, Arg(0));
     Dispatchx();
 
- OpCase(i_move_call_ext_only_ecr): {
+ OpCase(i_move_call_ext_only_ec): {
      r(0) = Arg(1);
  }
  /* FALL THROUGH */
@@ -1722,17 +1657,11 @@ void process_main(void)
 
      PRE_BIF_SWAPOUT(c_p);
      c_p->fcalls = FCALLS - 1;
-     reg[0] = r(0);
      result = erl_send(c_p, r(0), x(1));
      PreFetch(0, next);
      ERTS_VERIFY_UNUSED_TEMP_ALLOC(c_p);
      ERTS_SMP_REQ_PROC_MAIN_LOCK(c_p);
      PROCESS_MAIN_CHK_LOCKS(c_p);
-     if (c_p->mbuf || MSO(c_p).overhead >= BIN_VHEAP_SZ(c_p)) {
-	 result = erts_gc_after_bif_call(c_p, result, reg, 2);
-	 r(0) = reg[0];
-	 E = c_p->stop;
-     }
      HTOP = HEAP_TOP(c_p);
      FCALLS = c_p->fcalls;
      if (is_value(result)) {
@@ -1743,7 +1672,6 @@ void process_main(void)
 	 SET_CP(c_p, I+1);
 	 SET_I(c_p->i);
 	 SWAPIN;
-	 r(0) = reg[0];
 	 Dispatch();
      }
      goto find_func_info;
@@ -1753,29 +1681,23 @@ void process_main(void)
 	Eterm element_index;
 	Eterm element_tuple;
 
-    OpCase(i_element_xjsd):
-	element_tuple = xb(Arg(0));
-	I++;
+    OpCase(i_element_jxsd):
+	element_tuple = xb(Arg(1));
 	goto do_element;
 
-    OpCase(i_element_yjsd):
-	element_tuple = yb(Arg(0));
-	I++;
+    OpCase(i_element_jysd):
+	element_tuple = yb(Arg(1));
 	goto do_element;
 
-    OpCase(i_element_rjsd):
-	element_tuple = r(0);
-	/* Fall through */
-
     do_element:
-	GetArg1(1, element_index);
+	GetArg1(2, element_index);
 	if (is_small(element_index) && is_tuple(element_tuple)) {
 	    Eterm* tp = tuple_val(element_tuple);
 
 	    if ((signed_val(element_index) >= 1) &&
 		(signed_val(element_index) <= arityval(*tp))) {
 		Eterm result = tp[signed_val(element_index)];
-		StoreBifResult(2, result);
+		StoreBifResult(3, result);
 	    }
 	}
     }
@@ -1789,29 +1711,24 @@ void process_main(void)
     {
 	Eterm fast_element_tuple;
 
-    OpCase(i_fast_element_rjId):
-	fast_element_tuple = r(0);
+    OpCase(i_fast_element_jxId):
+     fast_element_tuple = xb(Arg(1));
+     goto do_fast_element;
+
+    OpCase(i_fast_element_jyId):
+     fast_element_tuple = yb(Arg(1));
+     goto do_fast_element;
 
     do_fast_element:
 	if (is_tuple(fast_element_tuple)) {
 	    Eterm* tp = tuple_val(fast_element_tuple);
-	    Eterm pos = Arg(1);	/* Untagged integer >= 1 */
+	    Eterm pos = Arg(2);	/* Untagged integer >= 1 */
 	    if (pos <= arityval(*tp)) {
 		Eterm result = tp[pos];
-		StoreBifResult(2, result);
+		StoreBifResult(3, result);
 	    }
 	}
      goto badarg;
-
-    OpCase(i_fast_element_xjId):
-     fast_element_tuple = xb(Arg(0));
-     I++;
-     goto do_fast_element;
-
-    OpCase(i_fast_element_yjId):
-     fast_element_tuple = yb(Arg(0));
-     I++;
-     goto do_fast_element;
  }
 
  OpCase(catch_yf):
@@ -1832,11 +1749,10 @@ void process_main(void)
 		 SWAPIN;
 	     }
 	     /* only x(2) is included in the rootset here */
-	     if (E - HTOP < 3 || c_p->mbuf) {	/* Force GC in case add_stacktrace()
-						 * created heap fragments */
+	     if (E - HTOP < 3) {
 		 SWAPOUT;
 		 PROCESS_MAIN_CHK_LOCKS(c_p);
-		 FCALLS -= erts_garbage_collect(c_p, 3, reg+2, 1);
+		 FCALLS -= erts_garbage_collect_nobump(c_p, 3, reg+2, 1);
 		 ERTS_VERIFY_UNUSED_TEMP_ALLOC(c_p);
 		 PROCESS_MAIN_CHK_LOCKS(c_p);
 		 SWAPIN;
@@ -1917,13 +1833,20 @@ void process_main(void)
      * Pick up the next message and place it in x(0).
      * If no message, jump to a wait or wait_timeout instruction.
      */
- OpCase(i_loop_rec_fr):
+ OpCase(i_loop_rec_f):
  {
      BeamInstr *next;
-     ErlMessage* msgp;
+     ErtsMessage* msgp;
 
- loop_rec__:
+     /*
+      * We need to disable GC while matching messages
+      * in the queue. This since messages with data outside
+      * the heap will be corrupted by a GC.
+      */
+     ASSERT(!(c_p->flags & F_DELAY_GC));
+     c_p->flags |= F_DELAY_GC;
 
+ loop_rec__:
      PROCESS_MAIN_CHK_LOCKS(c_p);
 
      msgp = PEEK_MESSAGE(c_p);
@@ -1935,6 +1858,7 @@ void process_main(void)
 	 if (ERTS_PROC_PENDING_EXIT(c_p)) {
 	     erts_smp_proc_unlock(c_p, ERTS_PROC_LOCKS_MSG_RECEIVE);
 	     SWAPOUT;
+	     c_p->flags &= ~F_DELAY_GC;
 	     goto do_schedule; /* Will be rescheduled for exit */
 	 }
 	 ERTS_SMP_MSGQ_MV_INQ2PRIVQ(c_p);
@@ -1944,32 +1868,27 @@ void process_main(void)
 	 else
 #endif
 	 {
+	     c_p->flags &= ~F_DELAY_GC;
 	     SET_I((BeamInstr *) Arg(0));
 	     Goto(*I);		/* Jump to a wait or wait_timeout instruction */
 	 }
      }
-     ErtsMoveMsgAttachmentIntoProc(msgp, c_p, E, HTOP, FCALLS,
-				   {
-				       SWAPOUT;
-				       reg[0] = r(0);
-				       PROCESS_MAIN_CHK_LOCKS(c_p);
-				   },
-				   {
-				       ERTS_VERIFY_UNUSED_TEMP_ALLOC(c_p);
-				       PROCESS_MAIN_CHK_LOCKS(c_p);
-				       r(0) = reg[0];
-				       SWAPIN;
-				   });
      if (is_non_value(ERL_MESSAGE_TERM(msgp))) {
-	 /*
-	  * A corrupt distribution message that we weren't able to decode;
-	  * remove it...
-	  */
-	 ASSERT(!msgp->data.attached);
-         /* TODO: Add DTrace probe for this bad message situation? */
-	 UNLINK_MESSAGE(c_p, msgp);
-	 free_message(msgp);
-	 goto loop_rec__;
+	 SWAPOUT; /* erts_decode_dist_message() may write to heap... */
+	 if (!erts_decode_dist_message(c_p, ERTS_PROC_LOCK_MAIN, msgp, 0)) {
+	     /*
+	      * A corrupt distribution message that we weren't able to decode;
+	      * remove it...
+	      */
+	     /* No swapin should be needed */
+	     ASSERT(HTOP == c_p->htop && E == c_p->stop);
+	     /* TODO: Add DTrace probe for this bad message situation? */
+	     UNLINK_MESSAGE(c_p, msgp);
+	     msgp->next = NULL;
+	     erts_cleanup_messages(msgp);
+	     goto loop_rec__;
+	 }
+	 SWAPIN;
      }
      PreFetch(1, next);
      r(0) = ERL_MESSAGE_TERM(msgp);
@@ -1981,8 +1900,7 @@ void process_main(void)
   */
  OpCase(remove_message): {
      BeamInstr *next;
-     ErlMessage* msgp;
-
+     ErtsMessage* msgp;
      PROCESS_MAIN_CHK_LOCKS(c_p);
 
      PreFetch(0, next);
@@ -1996,20 +1914,7 @@ void process_main(void)
 	 if (DT_UTAG(c_p) != NIL) {
 	     if (DT_UTAG_FLAGS(c_p) & DT_UTAG_PERMANENT) {
 		 SEQ_TRACE_TOKEN(c_p) = am_have_dt_utag;
-#ifdef DTRACE_TAG_HARDDEBUG
-		 if (DT_UTAG_FLAGS(c_p) & DT_UTAG_SPREADING) 
-		     erts_fprintf(stderr,
-				  "Dtrace -> (%T) stop spreading "
-				  "tag %T with message %T\r\n",
-				  c_p->common.id,DT_UTAG(c_p),ERL_MESSAGE_TERM(msgp));
-#endif
 	     } else {
-#ifdef DTRACE_TAG_HARDDEBUG
-		 erts_fprintf(stderr,
-			      "Dtrace -> (%T) kill tag %T with "
-			      "message %T\r\n",
-			      c_p->common.id,DT_UTAG(c_p),ERL_MESSAGE_TERM(msgp));
-#endif
 		 DT_UTAG(c_p) = NIL;
 		 SEQ_TRACE_TOKEN(c_p) = NIL;
 	     }
@@ -2029,12 +1934,6 @@ void process_main(void)
 		 DT_UTAG(c_p) = ERL_MESSAGE_DT_UTAG(msgp);
 	     }
 	     DT_UTAG_FLAGS(c_p) |= DT_UTAG_SPREADING;
-#ifdef DTRACE_TAG_HARDDEBUG
-	     erts_fprintf(stderr,
-			  "Dtrace -> (%T) receive tag (%T) "
-			  "with message %T\r\n",
-			  c_p->common.id, DT_UTAG(c_p), ERL_MESSAGE_TERM(msgp));
-#endif
 	 } else {
 #endif
 	     ASSERT(is_tuple(SEQ_TRACE_TOKEN(c_p)));
@@ -2064,7 +1963,7 @@ void process_main(void)
 
          dtrace_proc_str(c_p, receiver_name);
          token2 = SEQ_TRACE_TOKEN(c_p);
-         if (token2 != NIL && token2 != am_have_dt_utag) {
+         if (have_seqtrace(token2)) {
              tok_label = signed_val(SEQ_TRACE_T_LABEL(token2));
              tok_lastcnt = signed_val(SEQ_TRACE_T_LASTCNT(token2));
              tok_serial = signed_val(SEQ_TRACE_T_SERIAL(token2));
@@ -2077,11 +1976,21 @@ void process_main(void)
      UNLINK_MESSAGE(c_p, msgp);
      JOIN_MESSAGE(c_p);
      CANCEL_TIMER(c_p);
-     free_message(msgp);
+
+     erts_save_message_in_proc(c_p, msgp);
+     c_p->flags &= ~F_DELAY_GC;
+
+     if (ERTS_IS_GC_DESIRED_INTERNAL(c_p, HTOP, E)) {
+	 /*
+	  * We want to GC soon but we leave a few
+	  * reductions giving the message some time
+	  * to turn into garbage.
+	  */
+	 ERTS_VBUMP_LEAVE_REDS_INTERNAL(c_p, 5, FCALLS);
+     }
 
      ERTS_VERIFY_UNUSED_TEMP_ALLOC(c_p);
      PROCESS_MAIN_CHK_LOCKS(c_p);
-
      NextPF(0, next);
  }
 
@@ -2090,9 +1999,22 @@ void process_main(void)
      * message didn't match), then jump to the loop_rec instruction.
      */
  OpCase(loop_rec_end_f): {
+
+     ASSERT(c_p->flags & F_DELAY_GC);
+
      SET_I((BeamInstr *) Arg(0));
      SAVE_MESSAGE(c_p);
-     goto loop_rec__;
+     if (FCALLS > 0 || FCALLS > neg_o_reds) {
+	 FCALLS--;
+	 goto loop_rec__;
+     }
+
+     c_p->flags &= ~F_DELAY_GC;
+     c_p->i = I;
+     SWAPOUT;
+     c_p->arity = 0;
+     c_p->current = NULL;
+     goto do_schedule;
  }
     /*
      * Prepare to wait for a message or a timeout, whichever occurs first.
@@ -2131,8 +2053,6 @@ void process_main(void)
 		  * c_p->def_arg_reg[0].  Note that it is safe to use this
 		  * location because there are no living x registers in
 		  * a receive statement.
-		  * Note that for the halfword emulator, the two first elements
-		  * of the array are used.
 		  */
 		 BeamInstr** pi = (BeamInstr**) c_p->def_arg_reg;
 		 *pi = I+3;
@@ -2163,6 +2083,22 @@ void process_main(void)
 	 OpCase(wait_f):
 
 	 wait2: {
+#ifndef ERTS_SMP
+	     if (ERTS_PROC_IS_EXITING(c_p)) {
+		 /*
+		  * I non smp case:
+		  *
+		  * Currently executing process might be sent an exit
+		  * signal if it is traced by a port that it also is
+		  * linked to, and the port terminates during the
+		  * trace. In this case we do *not* want to clear
+		  * the active flag, which will make the process hang
+		  * in limbo forever.
+		  */
+		 SWAPOUT;
+		 goto do_schedule;
+	     }
+#endif
 	     c_p->i = (BeamInstr *) Arg(0); /* L1 */
 	     SWAPOUT;
 	     c_p->arity = 0;
@@ -2233,10 +2169,6 @@ void process_main(void)
      select_val2 = xb(Arg(0));
      goto do_select_tuple_arity2;
 
- OpCase(i_select_tuple_arity2_rfAAff):
-     select_val2 = r(0);
-     I--;
-
  do_select_tuple_arity2:
      if (is_not_tuple(select_val2)) {
 	 goto select_val2_fail;
@@ -2252,10 +2184,6 @@ void process_main(void)
      select_val2 = xb(Arg(0));
      goto do_select_val2;
 
- OpCase(i_select_val2_rfccff):
-     select_val2 = r(0);
-     I--;
-
  do_select_val2:
      if (select_val2 == Arg(2)) {
 	 I += 3;
@@ -2279,10 +2207,6 @@ void process_main(void)
      select_val = yb(Arg(0));
      goto do_select_tuple_arity;
 
- OpCase(i_select_tuple_arity_rfI):
-     select_val = r(0);
-     I--;
-
  do_select_tuple_arity:
      if (is_tuple(select_val)) {
 	 select_val = *tuple_val(select_val);
@@ -2299,10 +2223,6 @@ void process_main(void)
      select_val = yb(Arg(0));
      goto do_linear_search;
 
- OpCase(i_select_val_lins_rfI):
-     select_val = r(0);
-     I--;
-
  do_linear_search: {
      BeamInstr *vs = &Arg(3);
      int ix = 0;
@@ -2329,10 +2249,6 @@ void process_main(void)
      select_val = yb(Arg(0));
      goto do_binary_search;
      
- OpCase(i_select_val_bins_rfI):
-     select_val = r(0);
-     I--;
-
  do_binary_search:
  {
      struct Pairs {
@@ -2393,10 +2309,6 @@ void process_main(void)
      jump_on_val_zero_index = xb(Arg(0));
      goto do_jump_on_val_zero_index;
 
- OpCase(i_jump_on_val_zero_rfI):
-     jump_on_val_zero_index = r(0);
-     I--;
-
  do_jump_on_val_zero_index:
      if (is_small(jump_on_val_zero_index)) {
 	 jump_on_val_zero_index = signed_val(jump_on_val_zero_index);
@@ -2421,10 +2333,6 @@ void process_main(void)
      jump_on_val_index = xb(Arg(0));
      goto do_jump_on_val_index;
 
- OpCase(i_jump_on_val_rfII):
-     jump_on_val_index = r(0);
-     I--;
-
  do_jump_on_val_index:
      if (is_small(jump_on_val_index)) {
 	 jump_on_val_index = (Uint) (signed_val(jump_on_val_index) - Arg(3));
@@ -2444,15 +2352,12 @@ void process_main(void)
 
      do {
 	 Eterm term = *I++;
-	 switch (term & _TAG_IMMED1_MASK) {
-	 case (R_REG_DEF << _TAG_PRIMARY_SIZE) | TAG_PRIMARY_HEADER:
-	     *hp++ = r(0);
-	     break;
-	 case (X_REG_DEF << _TAG_PRIMARY_SIZE) | TAG_PRIMARY_HEADER:
-	     *hp++ = x(term >> _TAG_IMMED1_SIZE);
+	 switch (loader_tag(term)) {
+	 case LOADER_X_REG:
+	     *hp++ = x(loader_x_reg_index(term));
 	     break;
-	 case (Y_REG_DEF << _TAG_PRIMARY_SIZE) | TAG_PRIMARY_HEADER:
-	     *hp++ = y(term >> _TAG_IMMED1_SIZE);
+	 case LOADER_Y_REG:
+	     *hp++ = y(loader_y_reg_index(term));
 	     break;
 	 default:
 	     *hp++ = term;
@@ -2466,31 +2371,26 @@ void process_main(void)
  OpCase(new_map_dII): {
      Eterm res;
 
-     x(0) = r(0);
-     SWAPOUT;
+     HEAVY_SWAPOUT;
      res = new_map(c_p, reg, I-1);
-     SWAPIN;
-     r(0) = x(0);
+     HEAVY_SWAPIN;
      StoreResult(res, Arg(0));
      Next(3+Arg(2));
  }
 
-#define PUT_TERM_REG(term, desc)				\
-do {								\
-    switch ((desc) & _TAG_IMMED1_MASK) {			\
-    case (R_REG_DEF << _TAG_PRIMARY_SIZE) | TAG_PRIMARY_HEADER:	\
-	r(0) = (term);						\
-	break;							\
-    case (X_REG_DEF << _TAG_PRIMARY_SIZE) | TAG_PRIMARY_HEADER:	\
-	x((desc) >> _TAG_IMMED1_SIZE) = (term);			\
-	break;							\
-    case (Y_REG_DEF << _TAG_PRIMARY_SIZE) | TAG_PRIMARY_HEADER:	\
-	y((desc) >> _TAG_IMMED1_SIZE) = (term);			\
-	break;							\
-    default:							\
-	ASSERT(0);						\
-	break;							\
-    }								\
+#define PUT_TERM_REG(term, desc)		\
+do {						\
+    switch (loader_tag(desc)) {			\
+    case LOADER_X_REG:				\
+	x(loader_x_reg_index(desc)) = (term);	\
+	break;					\
+    case LOADER_Y_REG:				\
+	y(loader_y_reg_index(desc)) = (term);	\
+	break;					\
+    default:					\
+	ASSERT(0);				\
+	break;					\
+    }						\
 } while(0)
 
  OpCase(i_get_map_elements_fsI): {
@@ -2561,12 +2461,10 @@ do {								\
      Eterm map;
 
      GetArg1(1, map);
-     x(0) = r(0);
-     SWAPOUT;
+     HEAVY_SWAPOUT;
      res = update_map_assoc(c_p, reg, map, I);
-     SWAPIN;
+     HEAVY_SWAPIN;
      if (is_value(res)) {
-	 r(0) = x(0);
 	 StoreResult(res, Arg(2));
 	 Next(5+Arg(4));
      } else {
@@ -2585,12 +2483,10 @@ do {								\
      Eterm map;
 
      GetArg1(1, map);
-     x(0) = r(0);
-     SWAPOUT;
+     HEAVY_SWAPOUT;
      res = update_map_exact(c_p, reg, map, I);
-     SWAPIN;
+     HEAVY_SWAPIN;
      if (is_value(res)) {
-	 r(0) = x(0);
 	 StoreResult(res, Arg(2));
 	 Next(5+Arg(4));
      } else {
@@ -2672,13 +2568,10 @@ do {								\
     {
 	typedef Eterm (*GcBifFunction)(Process*, Eterm*, Uint);
 	GcBifFunction bf;
-	Eterm arg;
 	Eterm result;
 	Uint live = (Uint) Arg(3);
 
-	GetArg1(2, arg);
-	reg[0] = r(0);
-	reg[live] = arg;
+	GetArg1(2, x(live));
 	bf = (GcBifFunction) Arg(1);
 	c_p->fcalls = FCALLS;
 	SWAPOUT;
@@ -2689,7 +2582,6 @@ do {								\
 	ERTS_SMP_REQ_PROC_MAIN_LOCK(c_p);
 	PROCESS_MAIN_CHK_LOCKS(c_p);
 	SWAPIN;
-	r(0) = reg[0];
 	ERTS_HOLE_CHECK(c_p);
 	FCALLS = c_p->fcalls;
 	if (is_value(result)) {
@@ -2699,12 +2591,12 @@ do {								\
 	    SET_I((BeamInstr *) Arg(0));
 	    Goto(*I);
 	}
-	reg[0] = arg;
+	x(0) = x(live);
 	I = handle_error(c_p, I, reg, translate_gc_bif((void *) bf));
 	goto post_error_handling;
     }
 
- OpCase(i_gc_bif2_jIId): /* Note, one less parameter than the i_gc_bif1
+ OpCase(i_gc_bif2_jIIssd): /* Note, one less parameter than the i_gc_bif1
 			    and i_gc_bif3 */
     {
 	typedef Eterm (*GcBifFunction)(Process*, Eterm*, Uint);
@@ -2712,9 +2604,13 @@ do {								\
 	Eterm result;
 	Uint live = (Uint) Arg(2);
 
-	reg[0] = r(0);
-	reg[live++] = tmp_arg1;
-	reg[live] = tmp_arg2;
+	GetArg2(3, x(live), x(live+1));
+	/*
+	 * XXX This calling convention does not make sense. 'live'
+	 * should point out the first argument, not the second
+	 * (i.e. 'live' should not be incremented below).
+	 */
+	live++;
 	bf = (GcBifFunction) Arg(1);
 	c_p->fcalls = FCALLS;
 	SWAPOUT;
@@ -2725,35 +2621,37 @@ do {								\
 	ERTS_SMP_REQ_PROC_MAIN_LOCK(c_p);
 	PROCESS_MAIN_CHK_LOCKS(c_p);
 	SWAPIN;
-	r(0) = reg[0];
 	ERTS_HOLE_CHECK(c_p);
 	FCALLS = c_p->fcalls;
 	if (is_value(result)) {
-	    StoreBifResult(3, result);
+	    StoreBifResult(5, result);
 	}
 	if (Arg(0) != 0) {
 	    SET_I((BeamInstr *) Arg(0));
 	    Goto(*I);
 	}
-	reg[0] = tmp_arg1;
-	reg[1] = tmp_arg2;
+	live--;
+	x(0) = x(live);
+	x(1) = x(live+1);
 	I = handle_error(c_p, I, reg, translate_gc_bif((void *) bf));
 	goto post_error_handling;
     }
 
- OpCase(i_gc_bif3_jIsId):
+ OpCase(i_gc_bif3_jIIssd):
     {
 	typedef Eterm (*GcBifFunction)(Process*, Eterm*, Uint);
 	GcBifFunction bf;
-	Eterm arg;
 	Eterm result;
-	Uint live = (Uint) Arg(3);
+	Uint live = (Uint) Arg(2);
 
-	GetArg1(2, arg);
-	reg[0] = r(0);
-	reg[live++] = arg;
-	reg[live++] = tmp_arg1;
-	reg[live] = tmp_arg2;
+	x(live) = x(SCRATCH_X_REG);
+	GetArg2(3, x(live+1), x(live+2));
+	/*
+	 * XXX This calling convention does not make sense. 'live'
+	 * should point out the first argument, not the third
+	 * (i.e. 'live' should not be incremented below).
+	 */
+	live += 2;
 	bf = (GcBifFunction) Arg(1);
 	c_p->fcalls = FCALLS;
 	SWAPOUT;
@@ -2764,19 +2662,19 @@ do {								\
 	ERTS_SMP_REQ_PROC_MAIN_LOCK(c_p);
 	PROCESS_MAIN_CHK_LOCKS(c_p);
 	SWAPIN;
-	r(0) = reg[0];
 	ERTS_HOLE_CHECK(c_p);
 	FCALLS = c_p->fcalls;
 	if (is_value(result)) {
-	    StoreBifResult(4, result);
+	    StoreBifResult(5, result);
 	}
 	if (Arg(0) != 0) {
 	    SET_I((BeamInstr *) Arg(0));
 	    Goto(*I);
 	}
-	reg[0] = arg;
-	reg[1] = tmp_arg1;
-	reg[2] = tmp_arg2;
+	live -= 2;
+	x(0) = x(live);
+	x(1) = x(live+1);
+	x(2) = x(live+2);
 	I = handle_error(c_p, I, reg, translate_gc_bif((void *) bf));
 	goto post_error_handling;
     }
@@ -2784,12 +2682,13 @@ do {								\
  /*
   * Guards bifs and, or, xor in guards.
   */
- OpCase(i_bif2_fbd):
+ OpCase(i_bif2_fbssd):
     {
-	Eterm tmp_reg[2] = {tmp_arg1, tmp_arg2};
+	Eterm tmp_reg[2];
 	Eterm (*bf)(Process*, Eterm*);
 	Eterm result;
 
+	GetArg2(2, tmp_reg[0], tmp_reg[1]);
 	bf = (BifFunction) Arg(1);
 	c_p->fcalls = FCALLS;
 	PROCESS_MAIN_CHK_LOCKS(c_p);
@@ -2801,7 +2700,7 @@ do {								\
 	ERTS_HOLE_CHECK(c_p);
 	FCALLS = c_p->fcalls;
 	if (is_value(result)) {
-	    StoreBifResult(2, result);
+	    StoreBifResult(4, result);
 	}
 	SET_I((BeamInstr *) Arg(0));
 	Goto(*I);
@@ -2810,12 +2709,13 @@ do {								\
  /*
   * Guards bifs and, or, xor, relational operators in body.
   */
- OpCase(i_bif2_body_bd):
+ OpCase(i_bif2_body_bssd):
     {
-	Eterm tmp_reg[2] = {tmp_arg1, tmp_arg2};
+	Eterm tmp_reg[2];
 	Eterm (*bf)(Process*, Eterm*);
 	Eterm result;
 
+	GetArg2(1, tmp_reg[0], tmp_reg[1]);
 	bf = (BifFunction) Arg(0);
 	PROCESS_MAIN_CHK_LOCKS(c_p);
 	ASSERT(!ERTS_PROC_IS_EXITING(c_p));
@@ -2826,10 +2726,10 @@ do {								\
 	ERTS_HOLE_CHECK(c_p);
 	if (is_value(result)) {
 	    ASSERT(!is_CP(result));
-	    StoreBifResult(1, result);
+	    StoreBifResult(3, result);
 	}
-	reg[0] = tmp_arg1;
-	reg[1] = tmp_arg2;
+	reg[0] = tmp_reg[0];
+	reg[1] = tmp_reg[1];
 	SWAPOUT;
 	I = handle_error(c_p, I, reg, bf);
 	goto post_error_handling;
@@ -2844,6 +2744,7 @@ do {								\
 	Eterm (*bf)(Process*, Eterm*, BeamInstr*) = GET_BIF_ADDRESS(Arg(0));
 	Eterm result;
 	BeamInstr *next;
+	ErlHeapFragment *live_hf_end;
 
 	PRE_BIF_SWAPOUT(c_p);
 	c_p->fcalls = FCALLS - 1;
@@ -2853,18 +2754,18 @@ do {								\
 	PreFetch(1, next);
 	ASSERT(!ERTS_PROC_IS_EXITING(c_p));
 	ERTS_VERIFY_UNUSED_TEMP_ALLOC(c_p);
-	reg[0] = r(0);
+	live_hf_end = c_p->mbuf;
 	result = (*bf)(c_p, reg, I);
 	ASSERT(!ERTS_PROC_IS_EXITING(c_p) || is_non_value(result));
 	ERTS_VERIFY_UNUSED_TEMP_ALLOC(c_p);
 	ERTS_HOLE_CHECK(c_p);
 	ERTS_SMP_REQ_PROC_MAIN_LOCK(c_p);
-	PROCESS_MAIN_CHK_LOCKS(c_p);
-	if (c_p->mbuf || MSO(c_p).overhead >= BIN_VHEAP_SZ(c_p)) {
+	if (ERTS_IS_GC_DESIRED(c_p)) {
 	    Uint arity = ((Export *)Arg(0))->code[2];
-	    result = erts_gc_after_bif_call(c_p, result, reg, arity);
+	    result = erts_gc_after_bif_call_lhf(c_p, live_hf_end, result, reg, arity);
 	    E = c_p->stop;
 	}
+	PROCESS_MAIN_CHK_LOCKS(c_p);
 	HTOP = HEAP_TOP(c_p);
 	FCALLS = c_p->fcalls;
 	if (is_value(result)) {
@@ -2875,7 +2776,6 @@ do {								\
 	    SET_CP(c_p, I+2);
 	    SET_I(c_p->i);
 	    SWAPIN;
-	    r(0) = reg[0];
 	    Dispatch();
 	}
 
@@ -2891,111 +2791,81 @@ do {								\
   * Arithmetic operations.
   */
 
- OpCase(i_times_jId):
+ OpCase(i_times_jIssd):
  {
-     arith_func = ARITH_FUNC(mixed_times);
-     goto do_big_arith2;
+     Eterm Op1, Op2;
+     GetArg2(2, Op1, Op2);
+     DO_OUTLINED_ARITH_2(mixed_times, Op1, Op2);
  }
 
- OpCase(i_m_div_jId):
+ OpCase(i_m_div_jIssd):
  {
-     arith_func = ARITH_FUNC(mixed_div);
-     goto do_big_arith2;
+     Eterm Op1, Op2;
+     GetArg2(2, Op1, Op2);
+     DO_OUTLINED_ARITH_2(mixed_div, Op1, Op2);
  }
 
- OpCase(i_int_div_jId):
+ OpCase(i_int_div_jIssd):
  {
-     Eterm result;
+     Eterm Op1, Op2;
 
-     if (tmp_arg2 == SMALL_ZERO) {
+     GetArg2(2, Op1, Op2);
+     if (Op2 == SMALL_ZERO) {
 	 goto badarith;
-     } else if (is_both_small(tmp_arg1, tmp_arg2)) {
-	 Sint ires = signed_val(tmp_arg1) / signed_val(tmp_arg2);
+     } else if (is_both_small(Op1, Op2)) {
+	 Sint ires = signed_val(Op1) / signed_val(Op2);
 	 if (MY_IS_SSMALL(ires)) {
-	     result = make_small(ires);
-	     STORE_ARITH_RESULT(result);
+	     Eterm result = make_small(ires);
+	     StoreBifResult(4, result);
 	 }
      }
-     arith_func = ARITH_FUNC(int_div);
-     goto do_big_arith2;
+     DO_OUTLINED_ARITH_2(int_div, Op1, Op2);
  }
 
- OpCase(i_rem_jIxxd):
  {
-     Eterm result;
+     Eterm RemOp1, RemOp2;
 
-     if (xb(Arg(3)) == SMALL_ZERO) {
-	 goto badarith;
-     } else if (is_both_small(xb(Arg(2)), xb(Arg(3)))) {
-	 result = make_small(signed_val(xb(Arg(2))) % signed_val(xb(Arg(3))));
+ OpCase(i_rem_jIxxd):
+     RemOp1 = xb(Arg(2));
+     RemOp2 = xb(Arg(3));
+     goto do_rem;
+
+ OpCase(i_rem_jIssd):
+     GetArg2(2, RemOp1, RemOp2);
+     goto do_rem;
+
+ do_rem:
+     if (RemOp2 == SMALL_ZERO) {
+         goto badarith;
+     } else if (is_both_small(RemOp1, RemOp2)) {
+         Eterm result = make_small(signed_val(RemOp1) % signed_val(RemOp2));
          StoreBifResult(4, result);
+     } else {
+	 DO_OUTLINED_ARITH_2(int_rem, RemOp1, RemOp2);
      }
-     DO_BIG_ARITH(ARITH_FUNC(int_rem),xb(Arg(2)),xb(Arg(3)));
  }
 
- OpCase(i_rem_jId):
  {
-     Eterm result;
-
-     if (tmp_arg2 == SMALL_ZERO) {
-	 goto badarith;
-     } else if (is_both_small(tmp_arg1, tmp_arg2)) {
-	 result = make_small(signed_val(tmp_arg1) % signed_val(tmp_arg2));
-	 STORE_ARITH_RESULT(result);
-     } else {
-	 arith_func = ARITH_FUNC(int_rem);
-	 goto do_big_arith2;
-     }
- }
+     Eterm BandOp1, BandOp2;
 
  OpCase(i_band_jIxcd):
- {
-     Eterm result;
+     BandOp1 = xb(Arg(2));
+     BandOp2 = Arg(3);
+     goto do_band;
+
+ OpCase(i_band_jIssd):
+     GetArg2(2, BandOp1, BandOp2);
+     goto do_band;
 
-     if (is_both_small(xb(Arg(2)), Arg(3))) {
+ do_band:
+     if (is_both_small(BandOp1, BandOp2)) {
          /*
           * No need to untag -- TAG & TAG == TAG.
           */
-         result = xb(Arg(2)) & Arg(3);
+         Eterm result = BandOp1 & BandOp2;
          StoreBifResult(4, result);
      }
-     DO_BIG_ARITH(ARITH_FUNC(band),xb(Arg(2)),Arg(3));
- }
-
- OpCase(i_band_jId):
- {
-     Eterm result;
-
-     if (is_both_small(tmp_arg1, tmp_arg2)) {
-	 /*
-	  * No need to untag -- TAG & TAG == TAG.
-	  */
-	 result = tmp_arg1 & tmp_arg2;
-	 STORE_ARITH_RESULT(result);
-     }
-     arith_func = ARITH_FUNC(band);
-     goto do_big_arith2;
- }
-
-#undef DO_BIG_ARITH
-
- do_big_arith2:
- {
-     Eterm result;
-     Uint live = Arg(1);
-
-     SWAPOUT;
-     reg[0] = r(0);
-     reg[live] = tmp_arg1;
-     reg[live+1] = tmp_arg2;
-     result = arith_func(c_p, reg, live);
-     r(0) = reg[0];
-     SWAPIN;
-     ERTS_HOLE_CHECK(c_p);
-     if (is_value(result)) {
-	 STORE_ARITH_RESULT(result);
-     }
-     goto lb_Cl_error;
+     DO_OUTLINED_ARITH_2(band, BandOp1, BandOp2);
  }
 
  /*
@@ -3016,97 +2886,105 @@ do {								\
      goto find_func_info;
  }
 
- OpCase(i_bor_jId):
+ OpCase(i_bor_jIssd):
  {
-     Eterm result;
+     Eterm Op1, Op2;
 
-     if (is_both_small(tmp_arg1, tmp_arg2)) {
+     GetArg2(2, Op1, Op2);
+     if (is_both_small(Op1, Op2)) {
 	 /*
 	  * No need to untag -- TAG | TAG == TAG.
 	  */
-	 result = tmp_arg1 | tmp_arg2;
-	 STORE_ARITH_RESULT(result);
+	 Eterm result = Op1 | Op2;
+	 StoreBifResult(4, result);
      }
-     arith_func = ARITH_FUNC(bor);
-     goto do_big_arith2;
+     DO_OUTLINED_ARITH_2(bor, Op1, Op2);
  }
 
- OpCase(i_bxor_jId):
+ OpCase(i_bxor_jIssd):
  {
-     Eterm result;
+     Eterm Op1, Op2;
 
-     if (is_both_small(tmp_arg1, tmp_arg2)) {
+     GetArg2(2, Op1, Op2);
+     if (is_both_small(Op1, Op2)) {
 	 /*
 	  * We could extract the tag from one argument, but a tag extraction
 	  * could mean a shift.  Therefore, play it safe here.
 	  */
-	 result = make_small(signed_val(tmp_arg1) ^ signed_val(tmp_arg2));
-	 STORE_ARITH_RESULT(result);
+	 Eterm result = make_small(signed_val(Op1) ^ signed_val(Op2));
+	 StoreBifResult(4, result);
      }
-     arith_func = ARITH_FUNC(bxor);
-     goto do_big_arith2;
+     DO_OUTLINED_ARITH_2(bxor, Op1, Op2);
  }
 
  {
+     Eterm Op1, Op2;
      Sint i;
      Sint ires;
      Eterm* bigp;
+     Eterm tmp_big[2];
 
-     OpCase(i_bsr_jId):
-	 if (is_small(tmp_arg2)) {
-	     i = -signed_val(tmp_arg2);
-	     if (is_small(tmp_arg1)) {
+     OpCase(i_bsr_jIssd):
+         GetArg2(2, Op1, Op2);
+	 if (is_small(Op2)) {
+	     i = -signed_val(Op2);
+	     if (is_small(Op1)) {
 		 goto small_shift;
-	     } else if (is_big(tmp_arg1)) {
+	     } else if (is_big(Op1)) {
 		 if (i == 0) {
-		     StoreBifResult(2, tmp_arg1);
+		     StoreBifResult(4, Op1);
 		 }
 		 goto big_shift;
 	     }
-	 } else if (is_big(tmp_arg2)) {
+	 } else if (is_big(Op2)) {
 	     /*
 	      * N bsr NegativeBigNum == N bsl MAX_SMALL
 	      * N bsr PositiveBigNum == N bsl MIN_SMALL
 	      */
-	     tmp_arg2 = make_small(bignum_header_is_neg(*big_val(tmp_arg2)) ?
+	     Op2 = make_small(bignum_header_is_neg(*big_val(Op2)) ?
 				   MAX_SMALL : MIN_SMALL);
 	     goto do_bsl;
 	}
      goto badarith;
      
-     OpCase(i_bsl_jId):
+     OpCase(i_bsl_jIssd):
+         GetArg2(2, Op1, Op2);
+
  do_bsl:
-	 if (is_small(tmp_arg2)) {
-	     i = signed_val(tmp_arg2);
+	 if (is_small(Op2)) {
+	     i = signed_val(Op2);
 
-	     if (is_small(tmp_arg1)) {
+	     if (is_small(Op1)) {
 	     small_shift:
-		 ires = signed_val(tmp_arg1);
+		 ires = signed_val(Op1);
 	     
 		 if (i == 0 || ires == 0) {
-		     StoreBifResult(2, tmp_arg1);
+		     StoreBifResult(4, Op1);
 		 } else if (i < 0)  { /* Right shift */
 		     i = -i;
 		     if (i >= SMALL_BITS-1) {
-			 tmp_arg1 = (ires < 0) ? SMALL_MINUS_ONE : SMALL_ZERO;
+			 Op1 = (ires < 0) ? SMALL_MINUS_ONE : SMALL_ZERO;
 		     } else {
-			 tmp_arg1 = make_small(ires >> i);
+			 Op1 = make_small(ires >> i);
 		     }
-		     StoreBifResult(2, tmp_arg1);
+		     StoreBifResult(4, Op1);
 		 } else if (i < SMALL_BITS-1) { /* Left shift */
 		     if ((ires > 0 && ((~(Uint)0 << ((SMALL_BITS-1)-i)) & ires) == 0) ||
 			 ((~(Uint)0 << ((SMALL_BITS-1)-i)) & ~ires) == 0) {
-			 tmp_arg1 = make_small(ires << i);
-			 StoreBifResult(2, tmp_arg1);
+			 Op1 = make_small(ires << i);
+			 StoreBifResult(4, Op1);
 		     }
 		 }
-		 tmp_arg1 = small_to_big(ires, tmp_big);
+		 Op1 = small_to_big(ires, tmp_big);
+#ifdef TAG_LITERAL_PTR
+		 Op1 |= TAG_LITERAL_PTR;
+#endif
 
 	     big_shift:
 		 if (i > 0) {	/* Left shift. */
-		     ires = big_size(tmp_arg1) + (i / D_EXP);
+		     ires = big_size(Op1) + (i / D_EXP);
 		 } else {	/* Right shift. */
-		     ires = big_size(tmp_arg1);
+		     ires = big_size(Op1);
 		     if (ires <= (-i / D_EXP))
 			 ires = 3; /* ??? */
 		     else
@@ -3124,14 +3002,14 @@ do {								\
 			 c_p->freason = SYSTEM_LIMIT;
 			 goto lb_Cl_error;
 		     }
-		     TestHeapPreserve(ires+1, Arg(1), tmp_arg1);
+		     TestHeapPreserve(ires+1, Arg(1), Op1);
 		     bigp = HTOP;
-		     tmp_arg1 = big_lshift(tmp_arg1, i, bigp);
-		     if (is_big(tmp_arg1)) {
+		     Op1 = big_lshift(Op1, i, bigp);
+		     if (is_big(Op1)) {
 			 HTOP += bignum_header_arity(*HTOP) + 1;
 		     }
 		     HEAP_SPACE_VERIFIED(0);
-		     if (is_nil(tmp_arg1)) {
+		     if (is_nil(Op1)) {
 			 /*
 			  * This result must have been only slight larger
 			  * than allowed since it wasn't caught by the
@@ -3141,25 +3019,25 @@ do {								\
 			 goto lb_Cl_error;
 		     }
 		     ERTS_HOLE_CHECK(c_p);
-		     StoreBifResult(2, tmp_arg1);
+		     StoreBifResult(4, Op1);
 		 }
-	     } else if (is_big(tmp_arg1)) {
+	     } else if (is_big(Op1)) {
 		 if (i == 0) {
-		     StoreBifResult(2, tmp_arg1);
+		     StoreBifResult(4, Op1);
 		 }
 		 goto big_shift;
 	     }
-	 } else if (is_big(tmp_arg2)) {
-	     if (bignum_header_is_neg(*big_val(tmp_arg2))) {
+	 } else if (is_big(Op2)) {
+	     if (bignum_header_is_neg(*big_val(Op2))) {
 		 /*
 		  * N bsl NegativeBigNum is either 0 or -1, depending on
 		  * the sign of N. Since we don't believe this case
 		  * is common, do the calculation with the minimum
 		  * amount of code.
 		  */
-		 tmp_arg2 = make_small(MIN_SMALL);
+		 Op2 = make_small(MIN_SMALL);
 		 goto do_bsl;
-	     } else if (is_small(tmp_arg1) || is_big(tmp_arg1)) {
+	     } else if (is_small(Op1) || is_big(Op1)) {
 		 /*
 		  * N bsl PositiveBigNum is too large to represent.
 		  */
@@ -3183,12 +3061,10 @@ do {								\
 	 bnot_val = make_small(~signed_val(bnot_val));
      } else {
 	 Uint live = Arg(2);
-	 SWAPOUT;
-	 reg[0] = r(0);
+	 HEAVY_SWAPOUT;
 	 reg[live] = bnot_val;
 	 bnot_val = erts_gc_bnot(c_p, reg, live);
-	 r(0) = reg[0];
-	 SWAPIN;
+	 HEAVY_SWAPIN;
 	 ERTS_HOLE_CHECK(c_p);
 	 if (is_nil(bnot_val)) {
 	     goto lb_Cl_error;
@@ -3203,11 +3079,10 @@ do {								\
 
  OpCase(i_apply): {
      BeamInstr *next;
-     SWAPOUT;
+     HEAVY_SWAPOUT;
      next = apply(c_p, r(0), x(1), x(2), reg);
-     SWAPIN;
+     HEAVY_SWAPIN;
      if (next != NULL) {
-	 r(0) = reg[0];
 	 SET_CP(c_p, I+1);
 	 SET_I(next);
 	 Dispatch();
@@ -3218,12 +3093,11 @@ do {								\
 
  OpCase(i_apply_last_P): {
      BeamInstr *next;
-     SWAPOUT;
+     HEAVY_SWAPOUT;
      next = apply(c_p, r(0), x(1), x(2), reg);
-     SWAPIN;
+     HEAVY_SWAPIN;
      if (next != NULL) {
-	 r(0) = reg[0];
-	 SET_CP(c_p, (BeamInstr *) EXPAND_POINTER(E[0]));
+	 SET_CP(c_p, (BeamInstr *) E[0]);
 	 E = ADD_BYTE_OFFSET(E, Arg(0));
 	 SET_I(next);
 	 Dispatch();
@@ -3234,11 +3108,10 @@ do {								\
 
  OpCase(i_apply_only): {
      BeamInstr *next;
-     SWAPOUT;
+     HEAVY_SWAPOUT;
      next = apply(c_p, r(0), x(1), x(2), reg);
-     SWAPIN;
+     HEAVY_SWAPIN;
      if (next != NULL) {
-	 r(0) = reg[0];
 	 SET_I(next);
 	 Dispatch();
      }
@@ -3249,12 +3122,10 @@ do {								\
  OpCase(apply_I): {
      BeamInstr *next;
 
-     reg[0] = r(0);
-     SWAPOUT;
+     HEAVY_SWAPOUT;
      next = fixed_apply(c_p, reg, Arg(0));
-     SWAPIN;
+     HEAVY_SWAPIN;
      if (next != NULL) {
-	 r(0) = reg[0];
 	 SET_CP(c_p, I+2);
 	 SET_I(next);
 	 Dispatch();
@@ -3266,13 +3137,11 @@ do {								\
  OpCase(apply_last_IP): {
      BeamInstr *next;
 
-     reg[0] = r(0);
-     SWAPOUT;
+     HEAVY_SWAPOUT;
      next = fixed_apply(c_p, reg, Arg(0));
-     SWAPIN;
+     HEAVY_SWAPIN;
      if (next != NULL) {
-	 r(0) = reg[0];
-	 SET_CP(c_p, (BeamInstr *) EXPAND_POINTER(E[0]));
+	 SET_CP(c_p, (BeamInstr *) E[0]);
 	 E = ADD_BYTE_OFFSET(E, Arg(1));
 	 SET_I(next);
 	 Dispatch();
@@ -3284,11 +3153,10 @@ do {								\
  OpCase(i_apply_fun): {
      BeamInstr *next;
 
-     SWAPOUT;
+     HEAVY_SWAPOUT;
      next = apply_fun(c_p, r(0), x(1), reg);
-     SWAPIN;
+     HEAVY_SWAPIN;
      if (next != NULL) {
-	 r(0) = reg[0];
 	 SET_CP(c_p, I+1);
 	 SET_I(next);
 	 Dispatchfun();
@@ -3299,12 +3167,11 @@ do {								\
  OpCase(i_apply_fun_last_P): {
      BeamInstr *next;
 
-     SWAPOUT;
+     HEAVY_SWAPOUT;
      next = apply_fun(c_p, r(0), x(1), reg);
-     SWAPIN;
+     HEAVY_SWAPIN;
      if (next != NULL) {
-	 r(0) = reg[0];
-	 SET_CP(c_p, (BeamInstr *) EXPAND_POINTER(E[0]));
+	 SET_CP(c_p, (BeamInstr *) E[0]);
 	 E = ADD_BYTE_OFFSET(E, Arg(0));
 	 SET_I(next);
 	 Dispatchfun();
@@ -3315,11 +3182,10 @@ do {								\
  OpCase(i_apply_fun_only): {
      BeamInstr *next;
 
-     SWAPOUT;
+     HEAVY_SWAPOUT;
      next = apply_fun(c_p, r(0), x(1), reg);
-     SWAPIN;
+     HEAVY_SWAPIN;
      if (next != NULL) {
-	 r(0) = reg[0];
 	 SET_I(next);
 	 Dispatchfun();
      }
@@ -3329,13 +3195,10 @@ do {								\
  OpCase(i_call_fun_I): {
      BeamInstr *next;
 
-     SWAPOUT;
-     reg[0] = r(0);
-
+     HEAVY_SWAPOUT;
      next = call_fun(c_p, Arg(0), reg, THE_NON_VALUE);
-     SWAPIN;
+     HEAVY_SWAPIN;
      if (next != NULL) {
-	 r(0) = reg[0];
 	 SET_CP(c_p, I+2);
 	 SET_I(next);
 	 Dispatchfun();
@@ -3346,13 +3209,11 @@ do {								\
  OpCase(i_call_fun_last_IP): {
      BeamInstr *next;
 
-     SWAPOUT;
-     reg[0] = r(0);
+     HEAVY_SWAPOUT;
      next = call_fun(c_p, Arg(0), reg, THE_NON_VALUE);
-     SWAPIN;
+     HEAVY_SWAPIN;
      if (next != NULL) {
-	r(0) = reg[0];
-	SET_CP(c_p, (BeamInstr *) EXPAND_POINTER(E[0]));
+	SET_CP(c_p, (BeamInstr *) E[0]);
 	E = ADD_BYTE_OFFSET(E, Arg(1));
 	SET_I(next);
 	Dispatchfun();
@@ -3436,10 +3297,9 @@ do {								\
       */
 
      argp = c_p->arg_reg;
-     for (i = c_p->arity - 1; i > 0; i--) {
+     for (i = c_p->arity - 1; i >= 0; i--) {
 	 argp[i] = reg[i];
      }
-     c_p->arg_reg[0] = r(0);
      SWAPOUT;
      c_p->i = I;
      goto do_schedule1;
@@ -3452,19 +3312,14 @@ do {								\
      Eterm* p;
      
      PreFetch(3, next);
-     GetArg2(0, element, tuple);
+     GetArg1(0, element);
+     tuple = REG_TARGET(Arg(1));
      ASSERT(is_tuple(tuple));
      p = (Eterm *) ((unsigned char *) tuple_val(tuple) + Arg(2));
      *p = element;
      NextPF(3, next);
  }
 
- OpCase(i_is_ne_exact_f):
-    if (EQ(tmp_arg1, tmp_arg2)) {
-	ClauseFail();
-    }
-    Next(1);
-
  OpCase(normal_exit): {
      SWAPOUT;
      c_p->freason = EXC_NORMAL;
@@ -3531,25 +3386,14 @@ do {								\
     {
 	Eterm badmatch_val;
 
-    OpCase(badmatch_y):
-	badmatch_val = yb(Arg(0));
-	goto do_badmatch;
-
     OpCase(badmatch_x):
 	badmatch_val = xb(Arg(0));
-	goto do_badmatch;
-
-    OpCase(badmatch_r):
-	badmatch_val = r(0);
-
-    do_badmatch:
 	c_p->fvalue = badmatch_val;
 	c_p->freason = BADMATCH;
     }
  /* Fall through here */
 
  find_func_info: {
-     reg[0] = r(0);
      SWAPOUT;
      I = handle_error(c_p, I, reg, NULL);
      goto post_error_handling;
@@ -3566,11 +3410,9 @@ do {								\
      * code[3]: &&call_error_handler
      * code[4]: Not used
      */
-    SWAPOUT;
-    reg[0] = r(0);
+    HEAVY_SWAPOUT;
     I = call_error_handler(c_p, I-3, reg, am_undefined_function);
-    r(0) = reg[0];
-    SWAPIN;
+    HEAVY_SWAPIN;
     if (I) {
 	Goto(*I);
     }
@@ -3578,18 +3420,13 @@ do {								\
  /* Fall through */
  OpCase(error_action_code): {
     handle_error:
-     reg[0] = r(0);
      SWAPOUT;
      I = handle_error(c_p, NULL, reg, NULL);
  post_error_handling:
      if (I == 0) {
 	 goto do_schedule;
      } else {
-	 r(0) = reg[0];
 	 ASSERT(!is_value(r(0)));
-	 if (c_p->mbuf) {
-	     erts_garbage_collect(c_p, 0, reg+1, 3);
-	 }
 	 SWAPIN;
 	 Goto(*I);
      }
@@ -3613,6 +3450,7 @@ do {								\
 	     * I[3]: Function pointer to dirty NIF
 	     */
 	    BifFunction vbf;
+	    ErlHeapFragment *live_hf_end;
 
 	    DTRACE_NIF_ENTRY(c_p, (Eterm)I[-3], (Eterm)I[-2], (Uint)I[-1]);
 	    c_p->current = I-3; /* current and vbf set to please handle_error */ 
@@ -3628,7 +3466,7 @@ do {								\
 		NifF* fp = vbf = (NifF*) I[1];
 		struct enif_environment_t env;
 		erts_pre_nif(&env, c_p, (struct erl_module_nif*)I[2]);
-		reg[0] = r(0);
+		live_hf_end = c_p->mbuf;
 		nif_bif_result = (*fp)(&env, bif_nif_arity, reg);
 		if (env.exception_thrown)
 		    nif_bif_result = THE_NON_VALUE;
@@ -3667,10 +3505,11 @@ do {								\
 	    bif_nif_arity = I[-1];
             ASSERT(bif_nif_arity <= 4);
 	    ERTS_SMP_UNREQ_PROC_MAIN_LOCK(c_p);
-	    reg[0] = r(0);
+	    ERTS_VERIFY_UNUSED_TEMP_ALLOC(c_p);
 	    {
 		Eterm (*bf)(Process*, Eterm*, BeamInstr*) = vbf;
 		ASSERT(!ERTS_PROC_IS_EXITING(c_p));
+		live_hf_end = c_p->mbuf;
 		nif_bif_result = (*bf)(c_p, reg, I);
 		ASSERT(!ERTS_PROC_IS_EXITING(c_p) ||
 		       is_non_value(nif_bif_result));
@@ -3683,9 +3522,10 @@ do {								\
 	apply_bif_or_nif_epilogue:
 	    ERTS_SMP_REQ_PROC_MAIN_LOCK(c_p);
 	    ERTS_HOLE_CHECK(c_p);
-	    if (c_p->mbuf) {
-		nif_bif_result = erts_gc_after_bif_call(c_p, nif_bif_result,
-						  reg, bif_nif_arity);
+	    if (ERTS_IS_GC_DESIRED(c_p)) {
+		nif_bif_result = erts_gc_after_bif_call_lhf(c_p, live_hf_end,
+							    nif_bif_result,
+							    reg, bif_nif_arity);
 	    }
 	    SWAPIN;  /* There might have been a garbage collection. */
 	    FCALLS = c_p->fcalls;
@@ -3697,7 +3537,6 @@ do {								\
 		Goto(*I);
 	    } else if (c_p->freason == TRAP) {
 		SET_I(c_p->i);
-		r(0) = reg[0];
 		if (c_p->flags & F_HIBERNATE_SCHED) {
 		    c_p->flags &= ~F_HIBERNATE_SCHED;
 		    goto do_schedule;
@@ -3719,21 +3558,21 @@ do {								\
 	StoreBifResult(1, result);
     }
 
+ OpCase(i_get_hash_cId):
+    {
+	Eterm arg;
+	Eterm result;
+
+	GetArg1(0, arg);
+	result = erts_pd_hash_get_with_hx(c_p, Arg(1), arg);
+	StoreBifResult(2, result);
+    }
+
     {
 	Eterm case_end_val;
 
     OpCase(case_end_x):
 	case_end_val = xb(Arg(0));
-	goto do_case_end;
-
-    OpCase(case_end_y):
-	case_end_val = yb(Arg(0));
-	goto do_case_end;
-
-    OpCase(case_end_r):
-	case_end_val = r(0);
-
-    do_case_end:
 	c_p->fvalue = case_end_val;
 	c_p->freason = EXC_CASE_CLAUSE;
 	goto find_func_info;
@@ -3781,19 +3620,13 @@ do {								\
 	 goto do_bs_init_bits_known;
      }
 
-     OpCase(i_bs_init_bits_fail_heap_IjId): {
-	 /* tmp_arg1 was fetched by an i_fetch instruction */
-	 num_bits_term = tmp_arg1;
-	 alloc = Arg(0);
-	 I++;
+     OpCase(i_bs_init_bits_fail_heap_sIjId): {
+	 GetArg1(0, num_bits_term);
+	 alloc = Arg(1);
+	 I += 2;
 	 goto do_bs_init_bits;
      }
 
-     OpCase(i_bs_init_bits_fail_rjId): {
-	 num_bits_term = r(0);
-	 alloc = 0;
-	 goto do_bs_init_bits;
-     }
      OpCase(i_bs_init_bits_fail_yjId): {
 	 num_bits_term = yb(Arg(0));
 	 I++;
@@ -3913,52 +3746,48 @@ do {								\
  }
 
  {
-     OpCase(i_bs_init_fail_heap_IjId): {
-	 /* tmp_arg1 was fetched by an i_fetch instruction */
-	 tmp_arg2 = Arg(0);
-	 I++;
-	 goto do_bs_init;
-     }
+     Eterm BsOp1, BsOp2;
 
-     OpCase(i_bs_init_fail_rjId): {
-	 tmp_arg1 = r(0);
-	 tmp_arg2 = 0;
+     OpCase(i_bs_init_fail_heap_sIjId): {
+	 GetArg1(0, BsOp1);
+	 BsOp2 = Arg(1);
+	 I += 2;
 	 goto do_bs_init;
      }
 
      OpCase(i_bs_init_fail_yjId): {
-	 tmp_arg1 = yb(Arg(0));
-	 tmp_arg2 = 0;
+	 BsOp1 = yb(Arg(0));
+	 BsOp2 = 0;
 	 I++;
 	 goto do_bs_init;
      }
 
      OpCase(i_bs_init_fail_xjId): {
-	 tmp_arg1 = xb(Arg(0));
-	 tmp_arg2 = 0;
+	 BsOp1 = xb(Arg(0));
+	 BsOp2 = 0;
 	 I++;
      }
 	 /* FALL THROUGH */
      do_bs_init:
-         if (is_small(tmp_arg1)) {
-	     Sint size = signed_val(tmp_arg1);
+         if (is_small(BsOp1)) {
+	     Sint size = signed_val(BsOp1);
 	     if (size < 0) {
 		 goto badarg;
 	     }
-	     tmp_arg1 = (Eterm) size;
+	     BsOp1 = (Eterm) size;
 	 } else {
 	     Uint bytes;
 
-	     if (!term_to_Uint(tmp_arg1, &bytes)) {
+	     if (!term_to_Uint(BsOp1, &bytes)) {
 		 c_p->freason = bytes;
 		 goto lb_Cl_error;
 	     }
 	     if ((bytes >> (8*sizeof(Uint)-3)) != 0) {
 		 goto system_limit;
 	     }
-	     tmp_arg1 = (Eterm) bytes;
+	     BsOp1 = (Eterm) bytes;
 	 }
-	 if (tmp_arg1 <= ERL_ONHEAP_BIN_LIMIT) {
+	 if (BsOp1 <= ERL_ONHEAP_BIN_LIMIT) {
 	     goto do_heap_bin_alloc;
 	 } else {
 	     goto do_proc_bin_alloc;
@@ -3966,15 +3795,15 @@ do {								\
 
 
      OpCase(i_bs_init_heap_IIId): {
-	 tmp_arg1 = Arg(0);
-	 tmp_arg2 = Arg(1);
+	 BsOp1 = Arg(0);
+	 BsOp2 = Arg(1);
 	 I++;
 	 goto do_proc_bin_alloc;
      }
 
      OpCase(i_bs_init_IId): {
-	 tmp_arg1 = Arg(0);
-	 tmp_arg2 = 0;
+	 BsOp1 = Arg(0);
+	 BsOp2 = 0;
      }
      /* FALL THROUGH */
      do_proc_bin_alloc: {
@@ -3983,13 +3812,13 @@ do {								\
 
 	 erts_bin_offset = 0;
 	 erts_writable_bin = 0;
-	 TestBinVHeap(tmp_arg1 / sizeof(Eterm), 
-	 	      tmp_arg2 + PROC_BIN_SIZE + ERL_SUB_BIN_SIZE, Arg(1));
+	 TestBinVHeap(BsOp1 / sizeof(Eterm),
+		      BsOp2 + PROC_BIN_SIZE + ERL_SUB_BIN_SIZE, Arg(1));
 
 	 /*
 	  * Allocate the binary struct itself.
 	  */
-	 bptr = erts_bin_nrml_alloc(tmp_arg1);
+	 bptr = erts_bin_nrml_alloc(BsOp1);
 	 erts_refc_init(&bptr->refc, 1);
 	 erts_current_bin = (byte *) bptr->orig_bytes;
 
@@ -3999,28 +3828,28 @@ do {								\
 	 pb = (ProcBin *) HTOP;
 	 HTOP += PROC_BIN_SIZE;
 	 pb->thing_word = HEADER_PROC_BIN;
-	 pb->size = tmp_arg1;
+	 pb->size = BsOp1;
 	 pb->next = MSO(c_p).first;
 	 MSO(c_p).first = (struct erl_off_heap_header*) pb;
 	 pb->val = bptr;
 	 pb->bytes = (byte*) bptr->orig_bytes;
 	 pb->flags = 0;
 	 
-	 OH_OVERHEAD(&(MSO(c_p)), tmp_arg1 / sizeof(Eterm));
+	 OH_OVERHEAD(&(MSO(c_p)), BsOp1 / sizeof(Eterm));
 
 	 StoreBifResult(2, make_binary(pb));
      }
 
      OpCase(i_bs_init_heap_bin_heap_IIId): {
-	 tmp_arg1 = Arg(0);
-	 tmp_arg2 = Arg(1);
+	 BsOp1 = Arg(0);
+	 BsOp2 = Arg(1);
 	 I++;
 	 goto do_heap_bin_alloc;
      }
 
      OpCase(i_bs_init_heap_bin_IId): {
-	 tmp_arg1 = Arg(0);
-	 tmp_arg2 = 0;
+	 BsOp1 = Arg(0);
+	 BsOp2 = 0;
      }
      /* Fall through */
      do_heap_bin_alloc:
@@ -4028,33 +3857,36 @@ do {								\
 	     ErlHeapBin* hb;
 	     Uint bin_need;
 
-	     bin_need = heap_bin_size(tmp_arg1);
+	     bin_need = heap_bin_size(BsOp1);
 	     erts_bin_offset = 0;
 	     erts_writable_bin = 0;
-	     TestHeap(bin_need+tmp_arg2+ERL_SUB_BIN_SIZE, Arg(1));
+	     TestHeap(bin_need+BsOp2+ERL_SUB_BIN_SIZE, Arg(1));
 	     hb = (ErlHeapBin *) HTOP;
 	     HTOP += bin_need;
-	     hb->thing_word = header_heap_bin(tmp_arg1);
-	     hb->size = tmp_arg1;
+	     hb->thing_word = header_heap_bin(BsOp1);
+	     hb->size = BsOp1;
 	     erts_current_bin = (byte *) hb->data;
-	     tmp_arg1 = make_binary(hb);
-	     StoreBifResult(2, tmp_arg1);
+	     BsOp1 = make_binary(hb);
+	     StoreBifResult(2, BsOp1);
 	 }
  }
 
- OpCase(i_bs_add_jId): {
-     Uint Unit = Arg(1);
-     if (is_both_small(tmp_arg1, tmp_arg2)) {
-	 Sint Arg1 = signed_val(tmp_arg1);
-	 Sint Arg2 = signed_val(tmp_arg2);
+ OpCase(bs_add_jssId): {
+     Eterm Op1, Op2;
+     Uint Unit = Arg(3);
+
+     GetArg2(1, Op1, Op2);
+     if (is_both_small(Op1, Op2)) {
+	 Sint Arg1 = signed_val(Op1);
+	 Sint Arg2 = signed_val(Op2);
 
 	 if (Arg1 >= 0 && Arg2 >= 0) {
-	     BsSafeMul(Arg2, Unit, goto system_limit, tmp_arg1);
-	     tmp_arg1 += Arg1;
+	     BsSafeMul(Arg2, Unit, goto system_limit, Op1);
+	     Op1 += Arg1;
 
 	 store_bs_add_result:
-	     if (MY_IS_SSMALL((Sint) tmp_arg1)) {
-		 tmp_arg1 = make_small(tmp_arg1);
+	     if (MY_IS_SSMALL((Sint) Op1)) {
+		 Op1 = make_small(Op1);
 	     } else {
 		 /*
 		  * May generate a heap fragment, but in this
@@ -4066,10 +3898,10 @@ do {								\
 		  * references (such as the heap).
 		  */
 		 SWAPOUT;
-		 tmp_arg1 = erts_make_integer(tmp_arg1, c_p);
+		 Op1 = erts_make_integer(Op1, c_p);
 		 HTOP = HEAP_TOP(c_p);
 	     }
-	     StoreBifResult(2, tmp_arg1);
+	     StoreBifResult(4, Op1);
 	 }
 	 goto badarg;
      } else {
@@ -4092,16 +3924,16 @@ do {								\
 	  * an Uint, the reason is SYSTEM_LIMIT.
 	  */
 
-	 if (!term_to_Uint(tmp_arg1, &a)) {
+	 if (!term_to_Uint(Op1, &a)) {
 	     if (a == BADARG) {
 		 goto badarg;
 	     }
-	     if (!term_to_Uint(tmp_arg2, &b)) {
+	     if (!term_to_Uint(Op2, &b)) {
 		 c_p->freason = b;
 		 goto lb_Cl_error;
 	     }
 	     goto system_limit;
-	 } else if (!term_to_Uint(tmp_arg2, &b)) {
+	 } else if (!term_to_Uint(Op2, &b)) {
 	     c_p->freason = b;
 	     goto lb_Cl_error;
 	 }
@@ -4111,8 +3943,8 @@ do {								\
 	  */
 	 
 	 BsSafeMul(b, Unit, goto system_limit, c);
-	 tmp_arg1 = a + c;
-	 if (tmp_arg1 < a) {
+	 Op1 = a + c;
+	 if (Op1 < a) {
 	     /*
 	      * If the result is less than one of the
 	      * arguments, there must have been an overflow.
@@ -4134,52 +3966,47 @@ do {								\
     }
 
  /*
-  * tmp_arg1 = Number of bytes to build
-  * tmp_arg2 = Source binary
-  * Operands: Fail ExtraHeap Live Unit Dst
+  * x(SCRATCH_X_REG);
+  * Operands: Fail ExtraHeap Live Unit Size Dst
   */
 
- OpCase(i_bs_append_jIIId): {
+ OpCase(i_bs_append_jIIIsd): {
      Uint live = Arg(2);
      Uint res;
+     Eterm Size;
 
-     SWAPOUT;
-     reg[0] = r(0);
-     reg[live] = tmp_arg2;
-     res = erts_bs_append(c_p, reg, live, tmp_arg1, Arg(1), Arg(3));
-     r(0) = reg[0];
-     SWAPIN;
+     GetArg1(4, Size);
+     HEAVY_SWAPOUT;
+     reg[live] = x(SCRATCH_X_REG);
+     res = erts_bs_append(c_p, reg, live, Size, Arg(1), Arg(3));
+     HEAVY_SWAPIN;
      if (is_non_value(res)) {
 	 /* c_p->freason is already set (may be either BADARG or SYSTEM_LIMIT). */
 	 goto lb_Cl_error;
      }
-     StoreBifResult(4, res);
+     StoreBifResult(5, res);
  }
 
  /*
-  * tmp_arg1 = Number of bytes to build
-  * tmp_arg2 = Source binary
-  * Operands: Fail Unit Dst
+  * Operands: Fail Size Src Unit Dst
   */
- OpCase(i_bs_private_append_jId): {
+ OpCase(i_bs_private_append_jIssd): {
      Eterm res;
+     Eterm Size, Src;
 
-     res = erts_bs_private_append(c_p, tmp_arg2, tmp_arg1, Arg(1));
+     GetArg2(2, Size, Src);
+     res = erts_bs_private_append(c_p, Src, Size, Arg(1));
      if (is_non_value(res)) {
 	 /* c_p->freason is already set (may be either BADARG or SYSTEM_LIMIT). */
 	 goto lb_Cl_error;
      }
-     StoreBifResult(2, res);
+     StoreBifResult(4, res);
  }
 
- /*
-  * tmp_arg1 = Initial size of writable binary
-  * Operands: Live Dst
-  */
  OpCase(bs_init_writable): {
-     SWAPOUT;
+     HEAVY_SWAPOUT;
      r(0) = erts_bs_init_writable(c_p, r(0));
-     SWAPIN;
+     HEAVY_SWAPIN;
      Next(0);
  }
 
@@ -4270,26 +4097,29 @@ do {								\
 
  /*
   * Only used for validating a value matched out. 
-  *
-  * tmp_arg1 = Integer to validate
-  * tmp_arg2 = Match context
   */
- OpCase(i_bs_validate_unicode_retract_j): {
-     /*
-      * There is no need to untag the integer, but it IS necessary
-      * to make sure it is small (a bignum pointer could fall in
-      * the valid range).
-      */
-     if (is_not_small(tmp_arg1) || tmp_arg1 > make_small(0x10FFFFUL) ||
-	 (make_small(0xD800UL) <= tmp_arg1 &&
-	  tmp_arg1 <= make_small(0xDFFFUL))) {
-	 ErlBinMatchBuffer *mb = ms_matchbuffer(tmp_arg2);
+ OpCase(i_bs_validate_unicode_retract_jss): {
+	Eterm i; 		/* Integer to validate */
 
-	 mb->offset -= 32;
-	 goto badarg;
-     }
-     Next(1);
- }
+	/*
+	 * There is no need to untag the integer, but it IS necessary
+	 * to make sure it is small (a bignum pointer could fall in
+	 * the valid range).
+	 */
+
+	GetArg1(1, i);
+	if (is_not_small(i) || i > make_small(0x10FFFFUL) ||
+	    (make_small(0xD800UL) <= i && i <= make_small(0xDFFFUL))) {
+	    Eterm ms;		/* Match context */
+	    ErlBinMatchBuffer* mb;
+
+	    GetArg1(2, ms);
+	    mb = ms_matchbuffer(ms);
+	    mb->offset -= 32;
+	    goto badarg;
+	}
+	Next(3);
+    }
 
  /*
   * Matching of binaries.
@@ -4301,9 +4131,6 @@ do {								\
      Uint slots;
      Eterm context;
 
-     OpCase(i_bs_start_match2_rfIId): {
-	 context = r(0);
-
      do_start_match:
 	 slots = Arg(2);
 	 if (!is_boxed(context)) {
@@ -4350,7 +4177,7 @@ do {								\
 	     ClauseFail();
 	 }
 	 NextPF(4, next);
-     }
+
      OpCase(i_bs_start_match2_xfIId): {
 	 context = xb(Arg(0));
 	 I++;
@@ -4363,18 +4190,6 @@ do {								\
      }
  }
 
- OpCase(bs_test_zero_tail2_fr): {
-     BeamInstr *next;
-     ErlBinMatchBuffer *_mb;
-     
-     PreFetch(1, next);
-     _mb = (ErlBinMatchBuffer*) ms_matchbuffer(r(0));
-     if (_mb->size != _mb->offset) {
-	 ClauseFail();
-     }
-     NextPF(1, next);
- }
-
  OpCase(bs_test_zero_tail2_fx): {
      BeamInstr *next;
      ErlBinMatchBuffer *_mb;
@@ -4387,16 +4202,6 @@ do {								\
      NextPF(2, next);
  }
 
- OpCase(bs_test_tail_imm2_frI): {
-     BeamInstr *next;
-     ErlBinMatchBuffer *_mb;
-     PreFetch(2, next);
-     _mb = ms_matchbuffer(r(0));
-     if (_mb->size - _mb->offset != Arg(1)) {
-	 ClauseFail();
-     }
-     NextPF(2, next);
- }
  OpCase(bs_test_tail_imm2_fxI): {
      BeamInstr *next;
      ErlBinMatchBuffer *_mb;
@@ -4408,16 +4213,6 @@ do {								\
      NextPF(3, next);
  }
 
- OpCase(bs_test_unit_frI): {
-     BeamInstr *next;
-     ErlBinMatchBuffer *_mb;
-     PreFetch(2, next);
-     _mb = ms_matchbuffer(r(0));
-     if ((_mb->size - _mb->offset) % Arg(1)) {
-	 ClauseFail();
-     }
-     NextPF(2, next);
- }
  OpCase(bs_test_unit_fxI): {
      BeamInstr *next;
      ErlBinMatchBuffer *_mb;
@@ -4429,16 +4224,6 @@ do {								\
      NextPF(3, next);
  }
 
- OpCase(bs_test_unit8_fr): {
-     BeamInstr *next;
-     ErlBinMatchBuffer *_mb;
-     PreFetch(1, next);
-     _mb = ms_matchbuffer(r(0));
-     if ((_mb->size - _mb->offset) & 7) {
-	 ClauseFail();
-     }
-     NextPF(1, next);
- }
  OpCase(bs_test_unit8_fx): {
      BeamInstr *next;
      ErlBinMatchBuffer *_mb;
@@ -4453,19 +4238,11 @@ do {								\
  {
      Eterm bs_get_integer8_context;
 
- OpCase(i_bs_get_integer_8_rfd): {
-	 bs_get_integer8_context = r(0);
-	 goto do_bs_get_integer_8;
-     }
-
  OpCase(i_bs_get_integer_8_xfd): {
-	 bs_get_integer8_context = xb(Arg(0));
-	 I++;
-     }
-
- do_bs_get_integer_8: {
 	 ErlBinMatchBuffer *_mb;
 	 Eterm _result;
+	 bs_get_integer8_context = xb(Arg(0));
+	 I++;
 	 _mb = ms_matchbuffer(bs_get_integer8_context);
 	 if (_mb->size - _mb->offset < 8) {
 	     ClauseFail();
@@ -4483,15 +4260,10 @@ do {								\
  {
      Eterm bs_get_integer_16_context;
 
- OpCase(i_bs_get_integer_16_rfd):
-     bs_get_integer_16_context = r(0);
-     goto do_bs_get_integer_16;
-
  OpCase(i_bs_get_integer_16_xfd):
      bs_get_integer_16_context = xb(Arg(0));
      I++;
 
- do_bs_get_integer_16:
      {
 	 ErlBinMatchBuffer *_mb;
 	 Eterm _result;
@@ -4512,17 +4284,10 @@ do {								\
  {
      Eterm bs_get_integer_32_context;
 
- OpCase(i_bs_get_integer_32_rfId):
-     bs_get_integer_32_context = r(0);
-     goto do_bs_get_integer_32;
-
-     
  OpCase(i_bs_get_integer_32_xfId):
      bs_get_integer_32_context = xb(Arg(0));
      I++;
 
-
- do_bs_get_integer_32:
      {
 	 ErlBinMatchBuffer *_mb;
 	 Uint32 _integer;
@@ -4535,11 +4300,11 @@ do {								\
 	     _integer = get_int32(_mb->base + _mb->offset/8);
 	 }
 	 _mb->offset += 32;
-#if !defined(ARCH_64) || HALFWORD_HEAP
+#if !defined(ARCH_64)
 	 if (IS_USMALL(0, _integer)) {
 #endif
 	     _result = make_small(_integer);
-#if !defined(ARCH_64) || HALFWORD_HEAP
+#if !defined(ARCH_64)
 	 } else {
 	     TestHeap(BIG_UINT_HEAP_SIZE, Arg(1));
 	     _result = uint_to_big((Uint) _integer, HTOP);
@@ -4551,103 +4316,82 @@ do {								\
      }
  }
 
- /* Operands: Size Live Fail Flags Dst */
- OpCase(i_bs_get_integer_imm_rIIfId): {
-     tmp_arg1 = r(0);
-     /* Operands: Size Live Fail Flags Dst */
-     goto do_bs_get_integer_imm_test_heap;
- }
+ {
+     Eterm Ms, Sz;
 
- /* Operands: x(Reg) Size Live Fail Flags Dst */
+     /* Operands: x(Reg) Size Live Fail Flags Dst */
  OpCase(i_bs_get_integer_imm_xIIfId): {
-     tmp_arg1 = xb(Arg(0));
-     I++;
-     /* Operands: Size Live Fail Flags Dst */
-     goto do_bs_get_integer_imm_test_heap;
- }
-
- /*
-  * tmp_arg1 = match context
-  * Operands: Size Live Fail Flags Dst
-  */
- do_bs_get_integer_imm_test_heap: {
-     Uint wordsneeded;
-     tmp_arg2 = Arg(0);
-     wordsneeded = 1+WSIZE(NBYTES(tmp_arg2));
-     TestHeapPreserve(wordsneeded, Arg(1), tmp_arg1);
-     I += 2;
-     /* Operands: Fail Flags Dst */
-     goto do_bs_get_integer_imm;
- }
-
- /* Operands: Size Fail Flags Dst */
- OpCase(i_bs_get_integer_small_imm_rIfId): {
-     tmp_arg1 = r(0);
-     tmp_arg2 = Arg(0);
-     I++;
-     /* Operands: Fail Flags Dst */
-     goto do_bs_get_integer_imm;
- }
+	 Uint wordsneeded;
+	 Ms = xb(Arg(0));
+	 Sz = Arg(1);
+	 wordsneeded = 1+WSIZE(NBYTES(Sz));
+	 TestHeapPreserve(wordsneeded, Arg(2), Ms);
+	 I += 3;
+	 /* Operands: Fail Flags Dst */
+	 goto do_bs_get_integer_imm;
+     }
 
- /* Operands: x(Reg) Size Fail Flags Dst */
+     /* Operands: x(Reg) Size Fail Flags Dst */
  OpCase(i_bs_get_integer_small_imm_xIfId): {
-     tmp_arg1 = xb(Arg(0));
-     tmp_arg2 = Arg(1);
-     I += 2;
-     /* Operands: Fail Flags Dst */
-     goto do_bs_get_integer_imm;
- }
+	 Ms = xb(Arg(0));
+	 Sz = Arg(1);
+	 I += 2;
+	 /* Operands: Fail Flags Dst */
+	 goto do_bs_get_integer_imm;
+     }
  
- /*
-  * tmp_arg1 = match context
-  * tmp_arg2 = size of field
-  * Operands: Fail Flags Dst
-  */
+     /*
+      * Ms = match context
+      * Sz = size of field
+      * Operands: Fail Flags Dst
+      */
  do_bs_get_integer_imm: {
-     ErlBinMatchBuffer* mb;
-     Eterm result;
+	 ErlBinMatchBuffer* mb;
+	 Eterm result;
 
-     mb = ms_matchbuffer(tmp_arg1);
-     LIGHT_SWAPOUT;
-     result = erts_bs_get_integer_2(c_p, tmp_arg2, Arg(1), mb);
-     LIGHT_SWAPIN;
-     HEAP_SPACE_VERIFIED(0);
-     if (is_non_value(result)) {
-	 ClauseFail();
+	 mb = ms_matchbuffer(Ms);
+	 LIGHT_SWAPOUT;
+	 result = erts_bs_get_integer_2(c_p, Sz, Arg(1), mb);
+	 LIGHT_SWAPIN;
+	 HEAP_SPACE_VERIFIED(0);
+	 if (is_non_value(result)) {
+	     ClauseFail();
+	 }
+	 StoreBifResult(2, result);
      }
-     StoreBifResult(2, result);
  }
 
  /*
-  * tmp_arg1 = Match context
-  * tmp_arg2 = Size field
-  * Operands: Fail Live FlagsAndUnit Dst
+  * Operands: Fail Live FlagsAndUnit Ms Sz Dst
   */
- OpCase(i_bs_get_integer_fIId): {
+ OpCase(i_bs_get_integer_fIIssd): {
      Uint flags;
      Uint size;
+     Eterm Ms;
+     Eterm Sz;
      ErlBinMatchBuffer* mb;
      Eterm result;
 
      flags = Arg(2);
-     BsGetFieldSize(tmp_arg2, (flags >> 3), ClauseFail(), size);
+     GetArg2(3, Ms, Sz);
+     BsGetFieldSize(Sz, (flags >> 3), ClauseFail(), size);
      if (size >= SMALL_BITS) {
 	 Uint wordsneeded;
-	 /* check bits size before potential gc.
+	 /* Check bits size before potential gc.
 	  * We do not want a gc and then realize we don't need
-	  * the allocated space (i.e. if the op fails)
+	  * the allocated space (i.e. if the op fails).
 	  *
-	  * remember to reacquire the matchbuffer after gc.
+	  * Remember to re-acquire the matchbuffer after gc.
 	  */
 
-	 mb = ms_matchbuffer(tmp_arg1);
+	 mb = ms_matchbuffer(Ms);
 	 if (mb->size - mb->offset < size) {
 	     ClauseFail();
 	 }
 	 wordsneeded = 1+WSIZE(NBYTES((Uint) size));
-	 TestHeapPreserve(wordsneeded, Arg(1), tmp_arg1);
+	 TestHeapPreserve(wordsneeded, Arg(1), Ms);
      }
-     mb = ms_matchbuffer(tmp_arg1);
+     mb = ms_matchbuffer(Ms);
      LIGHT_SWAPOUT;
      result = erts_bs_get_integer_2(c_p, size, flags, mb);
      LIGHT_SWAPIN;
@@ -4655,18 +4399,13 @@ do {								\
      if (is_non_value(result)) {
 	 ClauseFail();
      }
-     StoreBifResult(3, result);
+     StoreBifResult(5, result);
  }
 
  {
      Eterm get_utf8_context;
 
      /* Operands: MatchContext Fail Dst */
- OpCase(i_bs_get_utf8_rfd): {
-	 get_utf8_context = r(0);
-	 goto do_bs_get_utf8;
-     }
-
  OpCase(i_bs_get_utf8_xfd): {
 	 get_utf8_context = xb(Arg(0));
 	 I++;
@@ -4677,7 +4416,7 @@ do {								\
       * Operands: Fail Dst
       */
 
- do_bs_get_utf8: {
+     {
 	 Eterm result = erts_bs_get_utf8(ms_matchbuffer(get_utf8_context));
 	 if (is_non_value(result)) {
 	     ClauseFail();
@@ -4690,12 +4429,7 @@ do {								\
      Eterm get_utf16_context;
 
      /* Operands: MatchContext Fail Flags Dst */
- OpCase(i_bs_get_utf16_rfId): {
-	 get_utf16_context = r(0);
-	 goto do_bs_get_utf16;
-     }
-
- OpCase(i_bs_get_utf16_xfId): {
+     OpCase(i_bs_get_utf16_xfId): {
 	 get_utf16_context = xb(Arg(0));
 	 I++;
      }
@@ -4704,7 +4438,7 @@ do {								\
       * get_utf16_context = match_context
       * Operands: Fail Flags Dst
       */
- do_bs_get_utf16: {
+     {
 	 Eterm result = erts_bs_get_utf16(ms_matchbuffer(get_utf16_context),
 					  Arg(1));
 	 if (is_non_value(result)) {
@@ -4723,26 +4457,10 @@ do {								\
      Uint orig;
      Uint hole_size;
 
-     OpCase(bs_context_to_binary_r): {
-	 context_to_binary_context = x0;
-	 I -= 2;
-	 goto do_context_to_binary;
-     }
-
-     /* Unfortunately, inlining can generate this instruction. */
-     OpCase(bs_context_to_binary_y): {
-	 context_to_binary_context = yb(Arg(0));
-	 goto do_context_to_binary0;
-     }
-
-     OpCase(bs_context_to_binary_x): {
+     OpCase(bs_context_to_binary_x):
 	 context_to_binary_context = xb(Arg(0));
-     
-     do_context_to_binary0:
 	 I--;
-     }
 
- do_context_to_binary:
      if (is_boxed(context_to_binary_context) &&
 	 header_is_bin_matchstate(*boxed_val(context_to_binary_context))) {
 	 ErlBinMatchState* ms;
@@ -4754,17 +4472,11 @@ do {								\
      }
      Next(2);
 
-     OpCase(i_bs_get_binary_all_reuse_rfI): {
-	 context_to_binary_context = x0;
-	 goto do_bs_get_binary_all_reuse;
-     }
-
      OpCase(i_bs_get_binary_all_reuse_xfI): {
 	 context_to_binary_context = xb(Arg(0));
 	 I++;
      }
 
- do_bs_get_binary_all_reuse:
      mb = ms_matchbuffer(context_to_binary_context);
      size = mb->size - mb->offset;
      if (size % Arg(1) != 0) {
@@ -4792,16 +4504,11 @@ do {								\
  {
      Eterm match_string_context;
 
-     OpCase(i_bs_match_string_rfII): {
-	 match_string_context = r(0);
-	 goto do_bs_match_string;
-     }
      OpCase(i_bs_match_string_xfII): {
 	 match_string_context = xb(Arg(0));
 	 I++;
      }
 
- do_bs_match_string:
      {
 	 BeamInstr *next;
 	 byte* bytes;
@@ -4829,14 +4536,6 @@ do {								\
      }
  }
 
- OpCase(i_bs_save2_rI): {
-     BeamInstr *next;
-     ErlBinMatchState *_ms;
-     PreFetch(1, next);
-     _ms = (ErlBinMatchState*) boxed_val((Eterm) r(0));
-     _ms->save_offset[Arg(0)] = _ms->mb.offset;
-     NextPF(1, next);
- }
  OpCase(i_bs_save2_xI): {
      BeamInstr *next;
      ErlBinMatchState *_ms;
@@ -4846,14 +4545,6 @@ do {								\
      NextPF(2, next);
  }
 
- OpCase(i_bs_restore2_rI): {
-     BeamInstr *next;
-     ErlBinMatchState *_ms;
-     PreFetch(1, next);
-     _ms = (ErlBinMatchState*) boxed_val((Eterm) r(0));
-     _ms->mb.offset = _ms->save_offset[Arg(0)];
-     NextPF(1, next);
- }
  OpCase(i_bs_restore2_xI): {
      BeamInstr *next;
      ErlBinMatchState *_ms;
@@ -4901,9 +4592,7 @@ do {								\
      BeamInstr real_I;
      ASSERT(I[-5] == (BeamInstr) BeamOp(op_i_func_info_IaaI));
      SWAPOUT;
-     reg[0] = r(0);
      real_I = erts_generic_breakpoint(c_p, I, reg);
-     r(0) = reg[0];
      SWAPIN;
      ASSERT(VALID_INSTR(real_I));
      Goto(real_I);
@@ -4963,7 +4652,7 @@ do {								\
      BeamInstr *next;
 
      PreFetch(2, next);
-     GetR(0, targ1);
+     targ1 = REG_TARGET(Arg(0));
      /* Arg(0) == HEADER_FLONUM */
      GET_DOUBLE(targ1, *(FloatDef*)ADD_BYTE_OFFSET(freg, fr));
      NextPF(2, next);
@@ -4983,7 +4672,7 @@ do {								\
      Eterm fr = Arg(1);
      BeamInstr *next;
 
-     GetR(0, targ1);
+     targ1 = REG_TARGET(Arg(0));
      PreFetch(2, next);
      if (is_small(targ1)) {
 	 fb(fr) = (double) signed_val(targ1);
@@ -5079,7 +4768,12 @@ do {								\
 
 #ifdef HIPE
  {
-     unsigned cmd;
+#define HIPE_MODE_SWITCH(Cmd)			\
+     SWAPOUT;					\
+     c_p->fcalls = FCALLS;			\
+     c_p->def_arg_reg[4] = -neg_o_reds;		\
+     c_p = hipe_mode_switch(c_p, Cmd, reg);     \
+     goto L_post_hipe_mode_switch
 
      OpCase(hipe_trap_call): {
 	 /*
@@ -5093,52 +4787,45 @@ do {								\
 	  */
 	 ASSERT(I[-5] == (Uint) OpCode(i_func_info_IaaI));
 	 c_p->hipe.u.ncallee = (void(*)(void)) I[-4];
-	 cmd = HIPE_MODE_SWITCH_CMD_CALL | (I[-1] << 8);
 	 ++hipe_trap_count;
-	 goto L_hipe_mode_switch;
+	 HIPE_MODE_SWITCH(HIPE_MODE_SWITCH_CMD_CALL | (I[-1] << 8));
      }
      OpCase(hipe_trap_call_closure): {
        ASSERT(I[-5] == (Uint) OpCode(i_func_info_IaaI));
        c_p->hipe.u.ncallee = (void(*)(void)) I[-4];
-       cmd = HIPE_MODE_SWITCH_CMD_CALL_CLOSURE | (I[-1] << 8);
        ++hipe_trap_count;
-       goto L_hipe_mode_switch;
+       HIPE_MODE_SWITCH(HIPE_MODE_SWITCH_CMD_CALL_CLOSURE | (I[-1] << 8));
      }
      OpCase(hipe_trap_return): {
-	 cmd = HIPE_MODE_SWITCH_CMD_RETURN;
-	 goto L_hipe_mode_switch;
+	 HIPE_MODE_SWITCH(HIPE_MODE_SWITCH_CMD_RETURN);
      }
      OpCase(hipe_trap_throw): {
-	 cmd = HIPE_MODE_SWITCH_CMD_THROW;
-	 goto L_hipe_mode_switch;
+	 HIPE_MODE_SWITCH(HIPE_MODE_SWITCH_CMD_THROW);
      }
      OpCase(hipe_trap_resume): {
-	 cmd = HIPE_MODE_SWITCH_CMD_RESUME;
-	 goto L_hipe_mode_switch;
+	 HIPE_MODE_SWITCH(HIPE_MODE_SWITCH_CMD_RESUME);
      }
- L_hipe_mode_switch:
-     /* XXX: this abuse of def_arg_reg[] is horrid! */
-     SWAPOUT;
-     c_p->fcalls = FCALLS;
-     c_p->def_arg_reg[4] = -neg_o_reds;
-     reg[0] = r(0);
-     c_p = hipe_mode_switch(c_p, cmd, reg);
+#undef HIPE_MODE_SWITCH
+
+ L_post_hipe_mode_switch:
      reg = ERTS_PROC_GET_SCHDATA(c_p)->x_reg_array;
      freg = ERTS_PROC_GET_SCHDATA(c_p)->f_reg_array;
      ERL_BITS_RELOAD_STATEP(c_p);
+     /* XXX: this abuse of def_arg_reg[] is horrid! */
      neg_o_reds = -c_p->def_arg_reg[4];
      FCALLS = c_p->fcalls;
      SWAPIN;
-     switch( c_p->def_arg_reg[3] ) { /* Halfword wont work with hipe yet! */
+     switch( c_p->def_arg_reg[3] ) {
        case HIPE_MODE_SWITCH_RES_RETURN:
 	 ASSERT(is_value(reg[0]));
-	 MoveReturn(reg[0], r(0));
+	 SET_I(c_p->cp);
+	 c_p->cp = 0;
+	 Goto(*I);
        case HIPE_MODE_SWITCH_RES_CALL_EXPORTED:
 	 c_p->i = c_p->hipe.u.callee_exp->addressv[erts_active_code_ix()];
 	 /*fall through*/
        case HIPE_MODE_SWITCH_RES_CALL_BEAM:
 	 SET_I(c_p->i);
-	 r(0) = reg[0];
 	 Dispatch();
        case HIPE_MODE_SWITCH_RES_CALL_CLOSURE:
 	 /* This can be used to call any function value, but currently it's
@@ -5147,9 +4834,8 @@ do {								\
 	     BeamInstr *next;
 
 	     next = call_fun(c_p, c_p->arity - 1, reg, THE_NON_VALUE);
-	     SWAPIN;
+	     HEAVY_SWAPIN;
 	     if (next != NULL) {
-		 r(0) = reg[0];
 		 SET_I(next);
 		 Dispatchfun();
 	     }
@@ -5197,22 +4883,22 @@ do {								\
  }
 
  OpCase(i_hibernate): {
-     SWAPOUT;
+     HEAVY_SWAPOUT;
      if (erts_hibernate(c_p, r(0), x(1), x(2), reg)) {
+	 FCALLS = c_p->fcalls;
 	 c_p->flags &= ~F_HIBERNATE_SCHED;
 	 goto do_schedule;
      } else {
+	 HEAVY_SWAPIN;
 	 I = handle_error(c_p, I, reg, hibernate_3);
 	 goto post_error_handling;
      }
  }
 
  OpCase(i_debug_breakpoint): {
-     SWAPOUT;
-     reg[0] = r(0);
+     HEAVY_SWAPOUT;
      I = call_error_handler(c_p, I-3, reg, am_breakpoint);
-     r(0) = reg[0];
-     SWAPIN;
+     HEAVY_SWAPIN;
      if (I) {
 	 Goto(*I);
      }
@@ -6219,6 +5905,23 @@ erts_hibernate(Process* c_p, Eterm module, Eterm function, Eterm args, Eterm* re
     int arity;
     Eterm tmp;
 
+#ifndef ERTS_SMP
+    if (ERTS_PROC_IS_EXITING(c_p)) {
+	/*
+	 * I non smp case:
+	 *
+	 * Currently executing process might be sent an exit
+	 * signal if it is traced by a port that it also is
+	 * linked to, and the port terminates during the
+	 * trace. In this case we do *not* want to clear
+	 * the active flag, which will make the process hang
+	 * in limbo forever. Get out of here and terminate
+	 * the process...
+	 */
+	return -1;
+    }
+#endif
+
     if (is_not_atom(module) || is_not_atom(function)) {
 	/*
 	 * No need to test args here -- done below.
@@ -6295,7 +5998,16 @@ erts_hibernate(Process* c_p, Eterm module, Eterm function, Eterm args, Eterm* re
 	ERTS_VERIFY_UNUSED_TEMP_ALLOC(c_p);
 	PROCESS_MAIN_CHK_LOCKS(c_p);
 	erts_smp_proc_lock(c_p, ERTS_PROC_LOCK_MSGQ|ERTS_PROC_LOCK_STATUS);
-#ifdef ERTS_SMP
+#ifndef ERTS_SMP
+	if (ERTS_PROC_IS_EXITING(c_p)) {
+	    /*
+	     * See comment in the begining of the function...
+	     *
+	     * This second test is needed since gc might be traced.
+	     */
+	    return -1;
+	}
+#else /* ERTS_SMP */
 	ERTS_SMP_MSGQ_MV_INQ2PRIVQ(c_p);
 	if (!c_p->msg.len)
 #endif
@@ -6397,7 +6109,7 @@ call_fun(Process* p,		/* Current process. */
 		 */
 		module = fe->module;
 		if ((modp = erts_get_module(module, code_ix)) != NULL
-		    && modp->curr.code != NULL) {
+		    && modp->curr.code_hdr != NULL) {
 		    /*
 		     * There is a module loaded, but obviously the fun is not
 		     * defined in it. We must not call the error_handler
@@ -6602,23 +6314,20 @@ static Eterm get_map_element_hash(Eterm map, Eterm key, Uint32 hx)
     return vs ? *vs : THE_NON_VALUE;
 }
 
-#define GET_TERM(term, dest)					\
-do {								\
-    Eterm src = (Eterm)(term);					\
-    switch (src & _TAG_IMMED1_MASK) {				\
-    case (R_REG_DEF << _TAG_PRIMARY_SIZE) | TAG_PRIMARY_HEADER:	\
-	dest = x(0);						\
-	break;							\
-    case (X_REG_DEF << _TAG_PRIMARY_SIZE) | TAG_PRIMARY_HEADER:	\
-	dest = x(src >> _TAG_IMMED1_SIZE);			\
-	break;							\
-    case (Y_REG_DEF << _TAG_PRIMARY_SIZE) | TAG_PRIMARY_HEADER:	\
-	dest = y(src >> _TAG_IMMED1_SIZE);			\
-	break;							\
-    default:							\
-	dest = src;						\
-	break;							\
-    }								\
+#define GET_TERM(term, dest)			\
+do {						\
+    Eterm src = (Eterm)(term);			\
+    switch (loader_tag(src)) {			\
+    case LOADER_X_REG:				\
+        dest = x(loader_x_reg_index(src));	\
+	break;					\
+    case LOADER_Y_REG:				\
+        dest = y(loader_y_reg_index(src));	\
+	break;					\
+    default:					\
+	dest = src;				\
+	break;					\
+    }						\
 } while(0)
 
 
@@ -6657,13 +6366,6 @@ new_map(Process* p, Eterm* reg, BeamInstr* I)
         erts_factory_proc_init(&factory, p);
         res = erts_hashmap_from_array(&factory, thp, n/2, 0);
         erts_factory_close(&factory);
-        if (p->mbuf) {
-            Uint live = Arg(2);
-            reg[live] = res;
-            erts_garbage_collect(p, 0, reg, live+1);
-            res       = reg[live];
-            E = p->stop;
-        }
         return res;
     }
 
@@ -6729,13 +6431,6 @@ update_map_assoc(Process* p, Eterm* reg, Eterm map, BeamInstr* I)
 	    hx = hashmap_make_hash(new_key);
 
 	    res = erts_hashmap_insert(p, hx, new_key, val, res,  0);
-	    if (p->mbuf) {
-		Uint live = Arg(3);
-		reg[live] = res;
-		erts_garbage_collect(p, 0, reg, live+1);
-		res       = reg[live];
-		E = p->stop;
-	    }
 
 	    new_p += 2;
 	}
@@ -6895,12 +6590,6 @@ update_map_assoc(Process* p, Eterm* reg, Eterm map, BeamInstr* I)
     /* The expensive case, need to build a hashmap */
     if (n > MAP_SMALL_MAP_LIMIT) {
 	res = erts_hashmap_from_ks_and_vs(p,flatmap_get_keys(mp),flatmap_get_values(mp),n);
-	if (p->mbuf) {
-	    Uint live = Arg(3);
-	    reg[live] = res;
-	    erts_garbage_collect(p, 0, reg, live+1);
-	    res       = reg[live];
-	}
     }
     return res;
 }
@@ -6956,14 +6645,6 @@ update_map_exact(Process* p, Eterm* reg, Eterm map, BeamInstr* I)
 		return res;
 	    }
 
-	    if (p->mbuf) {
-		Uint live = Arg(3);
-		reg[live] = res;
-		erts_garbage_collect(p, 0, reg, live+1);
-		res       = reg[live];
-		E = p->stop;
-	    }
-
 	    new_p += 2;
 	}
 	return res;
@@ -7072,6 +6753,15 @@ erts_is_builtin(Eterm Mod, Eterm Name, int arity)
     Export e;
     Export* ep;
 
+    if (Mod == am_erlang && Name == am_apply && arity == 3) {
+	/*
+	 * Special case. apply/3 is built-in (implemented in C),
+	 * but implemented in a different way than all other
+	 * BIFs.
+	 */
+	return 1;
+    }
+
     e.code[0] = Mod;
     e.code[1] = Name;
     e.code[2] = arity;
diff --git a/erts/emulator/beam/beam_load.c b/erts/emulator/beam/beam_load.c
index a6dce2d1d2..4d67587b10 100644
--- a/erts/emulator/beam/beam_load.c
+++ b/erts/emulator/beam/beam_load.c
@@ -80,7 +80,7 @@ ErlDrvBinary* erts_gzinflate_buffer(char*, int);
 
 typedef struct {
     Uint value;			/* Value of label (NULL if not known yet). */
-    Uint patches;		/* Index (into code buffer) to first location
+    Sint patches;		/* Index (into code buffer) to first location
 				 * which must be patched with the value of this label.
 				 */
 #ifdef ERTS_SMP
@@ -284,9 +284,10 @@ typedef struct LoaderState {
     int specific_op;		/* Specific opcode (-1 if not found). */
     int num_functions;		/* Number of functions in module. */
     int num_labels;		/* Number of labels. */
-    int code_buffer_size;	/* Size of code buffer in words.  */
-    BeamInstr* code;		/* Loaded code. */
-    int ci;			/* Current index into loaded code. */
+    BeamCodeHeader* hdr;	/* Loaded code header */
+    BeamInstr* codev;	        /* Loaded code buffer */
+    int        codev_size;      /* Size of code buffer in words. */
+    int ci;			/* Current index into loaded code buffer. */
     Label* labels;
     StringPatch* string_patches; /* Linked list of position into string table to patch. */
     BeamInstr catches;		/* Linked list of catch_yf instructions. */
@@ -480,7 +481,7 @@ static void free_literal_fragment(ErlHeapFragment*);
 static void loader_state_dtor(Binary* magic);
 static Eterm insert_new_code(Process *c_p, ErtsProcLocks c_p_locks,
 			     Eterm group_leader, Eterm module,
-			     BeamInstr* code, Uint size);
+			     BeamCodeHeader* code, Uint size);
 static int init_iff_file(LoaderState* stp, byte* code, Uint size);
 static int scan_iff_file(LoaderState* stp, Uint* chunk_types,
 			 Uint num_types, Uint num_mandatory);
@@ -526,15 +527,15 @@ static void new_string_patch(LoaderState* stp, int pos);
 static Uint new_literal(LoaderState* stp, Eterm** hpp, Uint heap_size);
 static int genopargcompare(GenOpArg* a, GenOpArg* b);
 static Eterm get_module_info(Process* p, ErtsCodeIndex code_ix,
-                             BeamInstr* code, Eterm module, Eterm what);
+                             BeamCodeHeader*, Eterm module, Eterm what);
 static Eterm exported_from_module(Process* p, ErtsCodeIndex code_ix,
                                   Eterm mod);
-static Eterm functions_in_module(Process* p, BeamInstr* code);
-static Eterm attributes_for_module(Process* p, BeamInstr* code);
-static Eterm compilation_info_for_module(Process* p, BeamInstr* code);
-static Eterm md5_of_module(Process* p, BeamInstr* code);
-static Eterm has_native(BeamInstr* code);
-static Eterm native_addresses(Process* p, BeamInstr* code);
+static Eterm functions_in_module(Process* p, BeamCodeHeader*);
+static Eterm attributes_for_module(Process* p, BeamCodeHeader*);
+static Eterm compilation_info_for_module(Process* p, BeamCodeHeader*);
+static Eterm md5_of_module(Process* p, BeamCodeHeader*);
+static Eterm has_native(BeamCodeHeader*);
+static Eterm native_addresses(Process* p, BeamCodeHeader*);
 int patch_funentries(Eterm Patchlist);
 int patch(Eterm Addresses, Uint fe);
 static int safe_mul(UWord a, UWord b, UWord* resp);
@@ -601,6 +602,7 @@ extern void check_allocated_block(Uint type, void *blk);
 #define CHKBLK(TYPE,BLK) /* nothing */
 #endif
 
+
 Eterm
 erts_prepare_loading(Binary* magic, Process *c_p, Eterm group_leader,
 		     Eterm* modp, byte* code, Uint unloaded_size)
@@ -641,20 +643,27 @@ erts_prepare_loading(Binary* magic, Process *c_p, Eterm group_leader,
     /*
      * Initialize code area.
      */
-    stp->code_buffer_size = 2048 + stp->num_functions;
-    stp->code = (BeamInstr *) erts_alloc(ERTS_ALC_T_CODE,
-				    sizeof(BeamInstr) * stp->code_buffer_size);
+    stp->codev_size = 2048 + stp->num_functions;
+    stp->hdr = (BeamCodeHeader*) erts_alloc(ERTS_ALC_T_CODE,
+                                            (offsetof(BeamCodeHeader,functions)
+                                             + sizeof(BeamInstr) * stp->codev_size));
 
-    stp->code[MI_NUM_FUNCTIONS] = stp->num_functions;
-    stp->ci = MI_FUNCTIONS + stp->num_functions + 1;
+    stp->hdr->num_functions = stp->num_functions;
 
-    stp->code[MI_ATTR_PTR] = 0;
-    stp->code[MI_ATTR_SIZE] = 0;
-    stp->code[MI_ATTR_SIZE_ON_HEAP] = 0;
-    stp->code[MI_COMPILE_PTR] = 0;
-    stp->code[MI_COMPILE_SIZE] = 0;
-    stp->code[MI_COMPILE_SIZE_ON_HEAP] = 0;
-    stp->code[MI_MD5_PTR] = 0;
+    /* Let the codev array start at functions[0] in order to index
+     * both function pointers and the loaded code itself that follows.
+     */
+    stp->codev = (BeamInstr*) &stp->hdr->functions;
+    stp->ci = stp->num_functions + 1;
+
+    stp->hdr->attr_ptr = NULL;
+    stp->hdr->attr_size = 0;
+    stp->hdr->attr_size_on_heap = 0;
+    stp->hdr->compile_ptr = NULL;
+    stp->hdr->compile_size = 0;
+    stp->hdr->compile_size_on_heap = 0;
+    stp->hdr->literals_start = NULL;
+    stp->hdr->md5_ptr = NULL;
 
     /*
      * Read the atom table.
@@ -776,7 +785,7 @@ erts_finish_loading(Binary* magic, Process* c_p,
 
     CHKBLK(ERTS_ALC_T_CODE,stp->code);
     retval = insert_new_code(c_p, c_p_locks, stp->group_leader, stp->module,
-			     stp->code, stp->loaded_size);
+			     stp->hdr, stp->loaded_size);
     if (retval != NIL) {
 	goto load_error;
     }
@@ -799,7 +808,8 @@ erts_finish_loading(Binary* magic, Process* c_p,
     debug_dump_code(stp->code,stp->ci);
 #endif
 #endif
-    stp->code = NULL;		/* Prevent code from being freed. */
+    stp->hdr = NULL;		/* Prevent code from being freed. */
+    stp->codev = NULL;
     *modp = stp->module;
 
     /*
@@ -831,7 +841,8 @@ erts_alloc_loader_state(void)
     stp->specific_op = -1;
     stp->genop = NULL;
     stp->atom = NULL;
-    stp->code = NULL;
+    stp->hdr = NULL;
+    stp->codev = NULL;
     stp->labels = NULL;
     stp->import = NULL;
     stp->export = NULL;
@@ -870,7 +881,7 @@ erts_module_for_prepared_code(Binary* magic)
 	return NIL;
     }
     stp = ERTS_MAGIC_BIN_DATA(magic);
-    if (stp->code != 0) {
+    if (stp->hdr != 0) {
 	return stp->module;
     } else {
 	return NIL;
@@ -891,7 +902,7 @@ static ErlHeapFragment* new_literal_fragment(Uint size)
     ErlHeapFragment* bp;
     bp = (ErlHeapFragment*) ERTS_HEAP_ALLOC(ERTS_ALC_T_PREPARED_CODE,
 					    ERTS_HEAP_FRAG_SIZE(size));
-    ERTS_INIT_HEAP_FRAG(bp, size);
+    ERTS_INIT_HEAP_FRAG(bp, size, size);
     return bp;
 }
 
@@ -920,9 +931,13 @@ loader_state_dtor(Binary* magic)
 	driver_free_binary(stp->bin);
 	stp->bin = 0;
     }
-    if (stp->code != 0) {
-	erts_free(ERTS_ALC_T_CODE, stp->code);
-	stp->code = 0;
+    if (stp->hdr != 0) {
+        if (stp->hdr->literals_start) {
+            erts_free(ERTS_ALC_T_LITERAL, stp->hdr->literals_start);
+        }
+	erts_free(ERTS_ALC_T_CODE, stp->hdr);
+	stp->hdr = 0;
+        stp->codev = 0;
     }
     if (stp->labels != 0) {
 	erts_free(ERTS_ALC_T_PREPARED_CODE, (void *) stp->labels);
@@ -995,7 +1010,7 @@ loader_state_dtor(Binary* magic)
 
 static Eterm
 insert_new_code(Process *c_p, ErtsProcLocks c_p_locks,
-		Eterm group_leader, Eterm module, BeamInstr* code,
+		Eterm group_leader, Eterm module, BeamCodeHeader* code_hdr,
 		Uint size)
 {
     Module* modp;
@@ -1016,7 +1031,7 @@ insert_new_code(Process *c_p, ErtsProcLocks c_p_locks,
 
     erts_total_code_size += size;
     modp = erts_put_module(module);
-    modp->curr.code = code;
+    modp->curr.code_hdr = code_hdr;
     modp->curr.code_length = size;
     modp->curr.catches = BEAM_CATCHES_NIL; /* Will be filled in later. */
 
@@ -1024,7 +1039,7 @@ insert_new_code(Process *c_p, ErtsProcLocks c_p_locks,
      * Update ranges (used for finding a function from a PC value).
      */
 
-    erts_update_ranges(code, size);
+    erts_update_ranges((BeamInstr*)modp->curr.code_hdr, size);
     return NIL;
 }
 
@@ -1373,7 +1388,7 @@ read_export_table(LoaderState* stp)
 	if (value == 0) {
 	    LoadError2(stp, "export table entry %d: label %d not resolved", i, n);
 	}
-	stp->export[i].address = address = stp->code + value;
+	stp->export[i].address = address = stp->codev + value;
 
 	/*
 	 * Find out if there is a BIF with the same name.
@@ -1392,7 +1407,7 @@ read_export_table(LoaderState* stp)
 	 * any other functions that walk through all local functions.
 	 */
 
-	if (stp->labels[n].patches) {
+	if (stp->labels[n].patches >= 0) {
 	    LoadError3(stp, "there are local calls to the stub for "
 		       "the BIF %T:%T/%d",
 		       stp->module, func, arity);
@@ -1513,8 +1528,8 @@ read_literal_table(LoaderState* stp)
 	}
 
         if (heap_size > 0) {
-            erts_factory_message_init(&factory, NULL, NULL,
-                                      new_literal_fragment(heap_size));
+            erts_factory_heap_frag_init(&factory,
+					new_literal_fragment(heap_size));
 	    factory.alloc_type = ERTS_ALC_T_PREPARED_CODE;
             val = erts_decode_ext(&factory, &p);
 
@@ -1735,7 +1750,7 @@ read_code_header(LoaderState* stp)
 				       stp->num_labels * sizeof(Label));
     for (i = 0; i < stp->num_labels; i++) {
 	stp->labels[i].value = 0;
-	stp->labels[i].patches = 0;
+	stp->labels[i].patches = -1;
 #ifdef ERTS_SMP
 	stp->labels[i].looprec_targeted = 0;
 #endif
@@ -1754,13 +1769,14 @@ read_code_header(LoaderState* stp)
     } else {}
 
 #define CodeNeed(w) do {						\
-    ASSERT(ci <= code_buffer_size);					\
-    if (code_buffer_size < ci+(w)) {					\
-        code_buffer_size = 2*ci+(w);					\
-	stp->code = code =						\
-	    (BeamInstr *) erts_realloc(ERTS_ALC_T_CODE,			\
-				     (void *) code,			\
-				     code_buffer_size * sizeof(BeamInstr));	\
+    ASSERT(ci <= codev_size);					        \
+    if (codev_size < ci+(w)) {					        \
+        codev_size = 2*ci+(w);	                                        \
+	stp->hdr = (BeamCodeHeader*) erts_realloc(ERTS_ALC_T_CODE,	\
+            (void *) stp->hdr,	                                        \
+            (offsetof(BeamCodeHeader,functions)                         \
+             + codev_size * sizeof(BeamInstr)));                        \
+        code = stp->codev = (BeamInstr*) &stp->hdr->functions;          \
     } 									\
 } while (0)
     
@@ -1776,7 +1792,7 @@ load_code(LoaderState* stp)
     int arg;			/* Number of current argument. */
     int num_specific;		/* Number of specific ops for current. */
     BeamInstr* code;
-    int code_buffer_size;
+    int codev_size;
     int specific;
     Uint last_label = 0;	/* Number of last label. */
     Uint function_number = 0;
@@ -1793,15 +1809,15 @@ load_code(LoaderState* stp)
 	FUNC_INFO_SZ = 5
     };
 
-    code = stp->code;
-    code_buffer_size = stp->code_buffer_size;
+    code = stp->codev;
+    codev_size = stp->codev_size;
     ci = stp->ci;
 
     for (;;) {
 	int new_op;
 	GenOp* tmp_op;
 
-	ASSERT(ci <= code_buffer_size);
+	ASSERT(ci <= codev_size);
 
     get_next_instr:
 	GetByte(stp, new_op);
@@ -1847,9 +1863,7 @@ load_code(LoaderState* stp)
 	    case TAG_o:
 		break;
 	    case TAG_x:
-		if (last_op->a[arg].val == 0) {
-		    last_op->a[arg].type = TAG_r;
-		} else if (last_op->a[arg].val >= MAX_REG) {
+		if (last_op->a[arg].val >= MAX_REG) {
 		    LoadError1(stp, "invalid x register number: %u",
 			       last_op->a[arg].val);
 		}
@@ -1894,15 +1908,14 @@ load_code(LoaderState* stp)
 			 */
 			{
 			    Eterm* hp;
-/* XXX:PaN - Halfword should use ARCH_64 variant instead */
-#if !defined(ARCH_64) || HALFWORD_HEAP
+#if !defined(ARCH_64)
 			    Uint high, low;
 # endif
 			    last_op->a[arg].val = new_literal(stp, &hp,
 							      FLOAT_SIZE_OBJECT);
 			    hp[0] = HEADER_FLONUM;
 			    last_op->a[arg].type = TAG_q;
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
 			    GetInt(stp, 8, hp[1]);
 # else
 			    GetInt(stp, 4, high);
@@ -2056,7 +2069,42 @@ load_code(LoaderState* stp)
 		if (((opc[specific].mask[0] & mask[0]) == mask[0]) &&
 		    ((opc[specific].mask[1] & mask[1]) == mask[1]) &&
 		    ((opc[specific].mask[2] & mask[2]) == mask[2])) {
-		    break;
+
+		    if (!opc[specific].involves_r) {
+			break;	/* No complications - match */
+		    }
+
+		    /*
+		     * The specific operation uses the 'r' operand,
+		     * which is shorthand for x(0). Now things
+		     * get complicated. First we must check whether
+		     * all operands that should be of type 'r' use
+		     * x(0) (as opposed to some other X register).
+		     */
+		    for (arg = 0; arg < arity; arg++) {
+			if (opc[specific].involves_r & (1 << arg) &&
+			    tmp_op->a[arg].type == TAG_x) {
+			    if (tmp_op->a[arg].val != 0) {
+				break; /* Other X register than 0 */
+			    }
+			}
+		    }
+
+		    if (arg == arity) {
+			/*
+			 * All 'r' operands use x(0) in the generic
+			 * operation. That means a match. Now we
+			 * will need to rewrite the generic instruction
+			 * to actually use 'r' instead of 'x(0)'.
+			 */
+			for (arg = 0; arg < arity; arg++) {
+			    if (opc[specific].involves_r & (1 << arg) &&
+				tmp_op->a[arg].type == TAG_x) {
+				tmp_op->a[arg].type = TAG_r;
+			    }
+			}
+			break;	/* Match */
+		    }
 		}
 		specific++;
 	    }
@@ -2167,14 +2215,11 @@ load_code(LoaderState* stp)
 		break;
 	    case 's':	/* Any source (tagged constant or register) */
 		switch (tag) {
-		case TAG_r:
-		    code[ci++] = make_rreg();
-		    break;
 		case TAG_x:
-		    code[ci++] = make_xreg(tmp_op->a[arg].val);
+		    code[ci++] = make_loader_x_reg(tmp_op->a[arg].val);
 		    break;
 		case TAG_y:
-		    code[ci++] = make_yreg(tmp_op->a[arg].val);
+		    code[ci++] = make_loader_y_reg(tmp_op->a[arg].val);
 		    break;
 		case TAG_i:
 		    code[ci++] = (BeamInstr) make_small((Uint)tmp_op->a[arg].val);
@@ -2185,6 +2230,10 @@ load_code(LoaderState* stp)
 		case TAG_n:
 		    code[ci++] = NIL;
 		    break;
+		case TAG_q:
+		    new_literal_patch(stp, ci);
+		    code[ci++] = tmp_op->a[arg].val;
+		    break;
 		default:
 		    LoadError1(stp, "bad tag %d for general source",
 			       tmp_op->a[arg].type);
@@ -2193,14 +2242,11 @@ load_code(LoaderState* stp)
 		break;
 	    case 'd':	/* Destination (x(0), x(N), y(N) */
 		switch (tag) {
-		case TAG_r:
-		    code[ci++] = make_rreg();
-		    break;
 		case TAG_x:
-		    code[ci++] = make_xreg(tmp_op->a[arg].val);
+		    code[ci++] = tmp_op->a[arg].val * sizeof(Eterm);
 		    break;
 		case TAG_y:
-		    code[ci++] = make_yreg(tmp_op->a[arg].val);
+		    code[ci++] = tmp_op->a[arg].val * sizeof(Eterm) + 1;
 		    break;
 		default:
 		    LoadError1(stp, "bad tag %d for destination",
@@ -2357,20 +2403,13 @@ load_code(LoaderState* stp)
 		stp->labels[tmp_op->a[arg].val].patches = ci;
 		ci++;
 		break;
-	    case TAG_r:
-		CodeNeed(1);
-		code[ci++] = (R_REG_DEF << _TAG_PRIMARY_SIZE) |
-		    TAG_PRIMARY_HEADER;
-		break;
 	    case TAG_x:
 		CodeNeed(1);
-		code[ci++] = (tmp_op->a[arg].val << _TAG_IMMED1_SIZE) |
-		    (X_REG_DEF << _TAG_PRIMARY_SIZE) | TAG_PRIMARY_HEADER;
+		code[ci++] = make_loader_x_reg(tmp_op->a[arg].val);
 		break;
 	    case TAG_y:
 		CodeNeed(1);
-		code[ci++] = (tmp_op->a[arg].val << _TAG_IMMED1_SIZE) |
-		    (Y_REG_DEF << _TAG_PRIMARY_SIZE) | TAG_PRIMARY_HEADER;
+		code[ci++] = make_loader_y_reg(tmp_op->a[arg].val);
 		break;
 	    case TAG_n:
 		CodeNeed(1);
@@ -2393,8 +2432,7 @@ load_code(LoaderState* stp)
 	switch (stp->specific_op) {
 	case op_i_func_info_IaaI:
 	    {
-		Uint offset;
-
+		Sint offset;
 		if (function_number >= stp->num_functions) {
 		    LoadError1(stp, "too many functions in module (header said %d)",
 			       stp->num_functions); 
@@ -2436,15 +2474,15 @@ load_code(LoaderState* stp)
 		stp->arity = code[ci-1];
 
 		ASSERT(stp->labels[last_label].value == ci - FUNC_INFO_SZ);
-		offset = MI_FUNCTIONS + function_number;
-		code[offset] = stp->labels[last_label].patches;
+		stp->hdr->functions[function_number] = (BeamInstr*) stp->labels[last_label].patches;
+		offset = function_number;
 		stp->labels[last_label].patches = offset;
 		function_number++;
 		if (stp->arity > MAX_ARG) {
 		    LoadError1(stp, "too many arguments: %d", stp->arity);
 		}
 #ifdef DEBUG
-		ASSERT(stp->labels[0].patches == 0); /* Should not be referenced. */
+		ASSERT(stp->labels[0].patches < 0); /* Should not be referenced. */
 		for (i = 1; i < stp->num_labels; i++) {
 		    ASSERT(stp->labels[i].patches < ci);
 		}
@@ -2458,7 +2496,6 @@ load_code(LoaderState* stp)
 	    stp->on_load = ci;
 	    break;
 	case op_bs_put_string_II:
-	case op_i_bs_match_string_rfII:
 	case op_i_bs_match_string_xfII:
 	    new_string_patch(stp, ci-1);
 	    break;
@@ -2515,7 +2552,7 @@ load_code(LoaderState* stp)
 	     * End of code found.
 	     */
 	case op_int_code_end:
-	    stp->code_buffer_size = code_buffer_size;
+	    stp->codev_size = codev_size;
 	    stp->ci = ci;
 	    stp->function = THE_NON_VALUE;
 	    stp->genop = NULL;
@@ -2680,6 +2717,12 @@ same_label(LoaderState* stp, GenOpArg Target, GenOpArg Label)
 	Target.val == Label.val;
 }
 
+static int
+is_killed(LoaderState* stp, GenOpArg Reg, GenOpArg Live)
+{
+    return Reg.type == TAG_x && Live.type == TAG_u &&
+	Live.val <= Reg.val;
+}
 
 /*
  * Generate an instruction for element/2.
@@ -2696,17 +2739,17 @@ gen_element(LoaderState* stp, GenOpArg Fail, GenOpArg Index,
     op->next = NULL;
 
     if (Index.type == TAG_i && Index.val > 0 &&
-	(Tuple.type == TAG_r || Tuple.type == TAG_x || Tuple.type == TAG_y)) {
+	(Tuple.type == TAG_x || Tuple.type == TAG_y)) {
 	op->op = genop_i_fast_element_4;
-	op->a[0] = Tuple;
-	op->a[1] = Fail;
+	op->a[0] = Fail;
+	op->a[1] = Tuple;
 	op->a[2].type = TAG_u;
 	op->a[2].val = Index.val;
 	op->a[3] = Dst;
     } else {
 	op->op = genop_i_element_4;
-	op->a[0] = Tuple;
-	op->a[1] = Fail;
+	op->a[0] = Fail;
+	op->a[1] = Tuple;
 	op->a[2] = Index;
 	op->a[3] = Dst;
     }
@@ -2832,23 +2875,16 @@ gen_get_integer2(LoaderState* stp, GenOpArg Fail, GenOpArg Ms, GenOpArg Live,
 	    goto generic;
 	}
     } else {
-	GenOp* op2;
-	NEW_GENOP(stp, op2);
-	
-	op->op = genop_i_fetch_2;
-	op->arity = 2;
-	op->a[0] = Ms;
-	op->a[1] = Size;
-	op->next = op2;
-
-	op2->op = genop_i_bs_get_integer_4;
-	op2->arity = 4;
-	op2->a[0] = Fail;
-	op2->a[1] = Live;
-	op2->a[2].type = TAG_u;
-	op2->a[2].val = (Unit.val << 3) | Flags.val;
-	op2->a[3] = Dst;
-	op2->next = NULL;
+	op->op = genop_i_bs_get_integer_6;
+	op->arity = 6;
+	op->a[0] = Fail;
+	op->a[1] = Live;
+	op->a[2].type = TAG_u;
+	op->a[2].val = (Unit.val << 3) | Flags.val;
+	op->a[3] = Ms;
+	op->a[4] = Size;
+	op->a[5] = Dst;
+	op->next = NULL;
 	return op;
     }
     op->next = NULL;
@@ -3272,14 +3308,14 @@ gen_literal_timeout(LoaderState* stp, GenOpArg Fail, GenOpArg Time)
     op->a[1].type = TAG_u;
     
     if (Time.type == TAG_i && (timeout = Time.val) >= 0 &&
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
 	(timeout >> 32) == 0
 #else
 	1
 #endif
 	) {
 	op->a[1].val = timeout;
-#if !defined(ARCH_64) || HALFWORD_HEAP
+#if !defined(ARCH_64)
     } else if (Time.type == TAG_q) {
 	Eterm big;
 
@@ -3296,7 +3332,7 @@ gen_literal_timeout(LoaderState* stp, GenOpArg Fail, GenOpArg Time)
 	}
 #endif
     } else {
-#if !defined(ARCH_64) || HALFWORD_HEAP
+#if !defined(ARCH_64)
     error:
 #endif
 	op->op = genop_i_wait_error_0;
@@ -3319,14 +3355,14 @@ gen_literal_timeout_locked(LoaderState* stp, GenOpArg Fail, GenOpArg Time)
     op->a[1].type = TAG_u;
     
     if (Time.type == TAG_i && (timeout = Time.val) >= 0 &&
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
 	(timeout >> 32) == 0
 #else
 	1
 #endif
 	) {
 	op->a[1].val = timeout;
-#if !defined(ARCH_64) || HALFWORD_HEAP
+#if !defined(ARCH_64)
     } else if (Time.type == TAG_q) {
 	Eterm big;
 
@@ -3343,7 +3379,7 @@ gen_literal_timeout_locked(LoaderState* stp, GenOpArg Fail, GenOpArg Time)
 	}
 #endif
     } else {
-#if !defined(ARCH_64) || HALFWORD_HEAP
+#if !defined(ARCH_64)
     error:
 #endif
 	op->op = genop_i_wait_error_locked_0;
@@ -3899,9 +3935,7 @@ gen_make_fun2(LoaderState* stp, GenOpArg idx)
 /*
  * Rewrite gc_bifs with one parameter (the common case). Utilized
  * in ops.tab to rewrite instructions calling bif's in guards
- * to use a garbage collecting implementation. The instructions
- * are sometimes once again rewritten to handle literals (putting the
- * parameter in the mostly unused r[0] before the instruction is executed).
+ * to use a garbage collecting implementation.
  */
 static GenOp*
 gen_guard_bif1(LoaderState* stp, GenOpArg Fail, GenOpArg Live, GenOpArg Bif,
@@ -3956,10 +3990,6 @@ gen_guard_bif1(LoaderState* stp, GenOpArg Fail, GenOpArg Live, GenOpArg Bif,
 
 /*
  * This is used by the ops.tab rule that rewrites gc_bifs with two parameters.
- * The instruction returned is then again rewritten to an i_load instruction
- * followed by i_gc_bif2_jIId, to handle literals properly.
- * As opposed to the i_gc_bif1_jIsId, the instruction  i_gc_bif2_jIId is
- * always rewritten, regardless of if there actually are any literals.
  */
 static GenOp*
 gen_guard_bif2(LoaderState* stp, GenOpArg Fail, GenOpArg Live, GenOpArg Bif,
@@ -3986,23 +4016,19 @@ gen_guard_bif2(LoaderState* stp, GenOpArg Fail, GenOpArg Live, GenOpArg Bif,
 	op->a[2].val = stp->import[Bif.val].arity;
 	return op;
     }
-    op->op = genop_ii_gc_bif2_6;
+    op->op = genop_i_gc_bif2_6;
     op->arity = 6;
     op->a[0] = Fail;
     op->a[1].type = TAG_u;
-    op->a[2] = S1;
-    op->a[3] = S2;
-    op->a[4] = Live;
+    op->a[2] = Live;
+    op->a[3] = S1;
+    op->a[4] = S2;
     op->a[5] = Dst;
     return op;
 }
 
 /*
  * This is used by the ops.tab rule that rewrites gc_bifs with three parameters.
- * The instruction returned is then again rewritten to a move instruction that
- * uses r[0] for temp storage, followed by an i_load instruction,
- * followed by i_gc_bif3_jIsId, to handle literals properly. Rewriting
- * always occur, as with the gc_bif2 counterpart.
  */
 static GenOp*
 gen_guard_bif3(LoaderState* stp, GenOpArg Fail, GenOpArg Live, GenOpArg Bif,
@@ -4033,10 +4059,10 @@ gen_guard_bif3(LoaderState* stp, GenOpArg Fail, GenOpArg Live, GenOpArg Bif,
     op->arity = 7;
     op->a[0] = Fail;
     op->a[1].type = TAG_u;
-    op->a[2] = S1;
-    op->a[3] = S2;
-    op->a[4] = S3;
-    op->a[5] = Live;
+    op->a[2] = Live;
+    op->a[3] = S1;
+    op->a[4] = S2;
+    op->a[5] = S3;
     op->a[6] = Dst;
     op->next = NULL;
     return op;
@@ -4258,6 +4284,53 @@ gen_get_map_element(LoaderState* stp, GenOpArg Fail, GenOpArg Src,
     return op;
 }
 
+static int
+hash_internal_genop_arg(LoaderState* stp, GenOpArg Key, Uint32* hx)
+{
+    switch (Key.type) {
+    case TAG_a:
+	*hx = atom_tab(atom_val(Key.val))->slot.bucket.hvalue;
+	return 1;
+    case TAG_i:
+	*hx = Key.val;
+	return 1;
+    case TAG_n:
+	*hx = make_internal_hash(NIL);
+	return 1;
+    case TAG_q:
+	*hx = make_internal_hash(stp->literals[Key.val].term);
+	return 1;
+    default:
+	return 0;
+    }
+}
+
+
+static GenOp*
+gen_get(LoaderState* stp, GenOpArg Src, GenOpArg Dst)
+{
+    GenOp* op;
+    Uint32 hx = 0;
+
+    NEW_GENOP(stp, op);
+    op->next = NULL;
+    if (hash_internal_genop_arg(stp, Src, &hx)) {
+	op->arity = 3;
+	op->op = genop_i_get_hash_3;
+        op->a[0] = Src;
+	op->a[1].type = TAG_u;
+	op->a[1].val = (BeamInstr) hx;
+	op->a[2] = Dst;
+    } else {
+	op->arity = 2;
+	op->op = genop_i_get_2;
+        op->a[0] = Src;
+	op->a[1] = Dst;
+    }
+    return op;
+}
+
+
 static GenOp*
 gen_get_map_elements(LoaderState* stp, GenOpArg Fail, GenOpArg Src,
 		     GenOpArg Size, GenOpArg* Rest)
@@ -4323,7 +4396,7 @@ gen_has_map_fields(LoaderState* stp, GenOpArg Fail, GenOpArg Src,
     for (i = 0; i < n; i++) {
 	op->a[3+2*i] = Rest[i];
 	op->a[3+2*i+1].type = TAG_x;
-	op->a[3+2*i+1].val = 0;	/* x(0); normally not used */
+	op->a[3+2*i+1].val = SCRATCH_X_REG; /* Ignore result */
     }
     return op;
 }
@@ -4336,8 +4409,8 @@ gen_has_map_fields(LoaderState* stp, GenOpArg Fail, GenOpArg Src,
 static int
 freeze_code(LoaderState* stp)
 {
-    BeamInstr* code = stp->code;
-    Uint *literal_end = NULL;
+    BeamCodeHeader* code_hdr = stp->hdr;
+    BeamInstr* codev = (BeamInstr*) &stp->hdr->functions;
     int i;
     byte* str_table;
     unsigned strtab_size = stp->chunks[STR_CHUNK].size;
@@ -4362,77 +4435,76 @@ freeze_code(LoaderState* stp)
     if (stp->line_instr == 0) {
 	line_size = 0;
     } else {
-	line_size = (MI_LINE_FUNC_TAB + (stp->num_functions + 1) +
-		     (stp->current_li+1) + stp->num_fnames) *
-	    sizeof(Eterm) + (stp->current_li+1) * stp->loc_size;
+	line_size = (offsetof(BeamCodeLineTab,func_tab)
+                     + (stp->num_functions + 1) * sizeof(BeamInstr**) /* func_tab */
+                     + (stp->current_li + 1) * sizeof(BeamInstr*)     /* line items */
+                     + stp->num_fnames * sizeof(Eterm)                /* fname table */
+                     + (stp->current_li + 1) * stp->loc_size);        /* loc_tab */
     }
-    size = (stp->ci * sizeof(BeamInstr)) +
-	(stp->total_literal_size * sizeof(Eterm)) +
+    size = offsetof(BeamCodeHeader,functions) + (stp->ci * sizeof(BeamInstr)) +
 	strtab_size + attr_size + compile_size + MD5_SIZE + line_size;
 
     /*
      * Move the code to its final location.
      */
 
-    code = (BeamInstr *) erts_realloc(ERTS_ALC_T_CODE, (void *) code, size);
-    CHKBLK(ERTS_ALC_T_CODE,code);
+    code_hdr = (BeamCodeHeader*) erts_realloc(ERTS_ALC_T_CODE, (void *) code_hdr, size);
+    codev = (BeamInstr*) &code_hdr->functions;
+    CHKBLK(ERTS_ALC_T_CODE,code_hdr);
     /*
      * Place a pointer to the op_int_code_end instruction in the
      * function table in the beginning of the file.
      */
 
-    code[MI_FUNCTIONS+stp->num_functions] = (BeamInstr) (code + stp->ci - 1);
-    CHKBLK(ERTS_ALC_T_CODE,code);
+    code_hdr->functions[stp->num_functions] = (codev + stp->ci - 1);
+    CHKBLK(ERTS_ALC_T_CODE,code_hdr);
 
     /*
      * Store the pointer to the on_load function.
      */
 
     if (stp->on_load) {
-	code[MI_ON_LOAD_FUNCTION_PTR] = (BeamInstr) (code + stp->on_load);
+	code_hdr->on_load_function_ptr = codev + stp->on_load;
     } else {
-	code[MI_ON_LOAD_FUNCTION_PTR] = 0;
+	code_hdr->on_load_function_ptr = NULL;
     }
-    CHKBLK(ERTS_ALC_T_CODE,code);
+    CHKBLK(ERTS_ALC_T_CODE,code_hdr);
 
-    literal_end = (Uint *) (code+stp->ci);
     /*
-     * Place the literal heap directly after the code and fix up all
-     * instructions that refer to it.
+     * Place the literals in their own allocated heap (for fast range check)
+     * and fix up all instructions that refer to it.
      */
     {
-	Uint* ptr;
-	Uint* low;
-	Uint* high;
+	Eterm* ptr;
 	LiteralPatch* lp;
         ErlOffHeap code_off_heap;
 
         ERTS_INIT_OFF_HEAP(&code_off_heap);
 
-	low = (Uint *) (code+stp->ci);
-	high = low + stp->total_literal_size;
-	code[MI_LITERALS_START] = (BeamInstr) low;
-	code[MI_LITERALS_END] = (BeamInstr) high;
-	ptr = low;
+        ptr = (Eterm*)erts_alloc(ERTS_ALC_T_LITERAL,
+                                 stp->total_literal_size*sizeof(Eterm));
+	code_hdr->literals_start = ptr;
+	code_hdr->literals_end = ptr + stp->total_literal_size;
 	for (i = 0; i < stp->num_literals; i++) {
-            if (stp->literals[i].heap_frags) {
-                move_multi_frags(&ptr, &code_off_heap, stp->literals[i].heap_frags,
-                                 &stp->literals[i].term, 1);
+            if (is_not_immed(stp->literals[i].term)) {
+                erts_move_multi_frags(&ptr, &code_off_heap,
+				      stp->literals[i].heap_frags,
+				      &stp->literals[i].term, 1, 1);
+                ASSERT(erts_is_literal(stp->literals[i].term,
+				       ptr_val(stp->literals[i].term)));
             }
-            else ASSERT(is_immed(stp->literals[i].term));
 	}
-        code[MI_LITERALS_OFF_HEAP] = (BeamInstr) code_off_heap.first;
+	code_hdr->literals_off_heap = code_off_heap.first;
 	lp = stp->literal_patches;
 	while (lp != 0) {
 	    BeamInstr* op_ptr;
 	    Literal* lit;
 
-	    op_ptr = code + lp->pos;
+	    op_ptr = codev + lp->pos;
 	    lit = &stp->literals[op_ptr[0]];
 	    op_ptr[0] = lit->term;
 	    lp = lp->next;
 	}
-	literal_end += stp->total_literal_size;
     }
     CHKBLK(ERTS_ALC_T_CODE,code);
 
@@ -4440,52 +4512,49 @@ freeze_code(LoaderState* stp)
      * If there is line information, place it here.
      */
     if (stp->line_instr == 0) {
-	code[MI_LINE_TABLE] = (BeamInstr) 0;
-	str_table = (byte *) literal_end;
+	code_hdr->line_table = NULL;
+	str_table = (byte *) (codev + stp->ci);
     } else {
-	Eterm* line_tab = (Eterm *) literal_end;
-	Eterm* p;
-	int ftab_size = stp->num_functions;
-	int num_instrs = stp->current_li;
-	Eterm* first_line_item;
+	BeamCodeLineTab* const line_tab = (BeamCodeLineTab *) (codev+stp->ci);
+	const int ftab_size = stp->num_functions;
+        const int num_instrs = stp->current_li;
+        const BeamInstr** const line_items =
+            (const BeamInstr**) &line_tab->func_tab[ftab_size + 1];
 
-	code[MI_LINE_TABLE] = (BeamInstr) line_tab;
-	p = line_tab + MI_LINE_FUNC_TAB;
+	code_hdr->line_table = line_tab;
 
-	first_line_item = (p + ftab_size + 1);
 	for (i = 0; i < ftab_size; i++) {
-	     *p++ = (Eterm) (BeamInstr) (first_line_item + stp->func_line[i]);
+            line_tab->func_tab[i] = line_items + stp->func_line[i];
 	}
-	*p++ = (Eterm) (BeamInstr) (first_line_item + num_instrs);
-	ASSERT(p == first_line_item);
+	line_tab->func_tab[i] = line_items + num_instrs;
+
 	for (i = 0; i < num_instrs; i++) {
-		*p++ = (Eterm) (BeamInstr) (code + stp->line_instr[i].pos);
+            line_items[i] = codev + stp->line_instr[i].pos;
 	}
-	*p++ = (Eterm) (BeamInstr) (code + stp->ci - 1);
+	line_items[i] = codev + stp->ci - 1;
 
-	line_tab[MI_LINE_FNAME_PTR] = (Eterm) (BeamInstr) p;
-	memcpy(p, stp->fname, stp->num_fnames*sizeof(Eterm));
-	p += stp->num_fnames;
+	line_tab->fname_ptr = (Eterm*) &line_items[i + 1];
+	memcpy(line_tab->fname_ptr, stp->fname, stp->num_fnames*sizeof(Eterm));
 
-	line_tab[MI_LINE_LOC_TAB] = (Eterm) (BeamInstr) p;
-	line_tab[MI_LINE_LOC_SIZE] = stp->loc_size;
+	line_tab->loc_size = stp->loc_size;
 	if (stp->loc_size == 2) {
-	    Uint16* locp = (Uint16 *) p;
-	    for (i = 0; i < num_instrs; i++) {
+            Uint16* locp = (Uint16 *) &line_tab->fname_ptr[stp->num_fnames];
+            line_tab->loc_tab.p2 = locp;
+            for (i = 0; i < num_instrs; i++) {
 		*locp++ = (Uint16) stp->line_instr[i].loc;
-	    }
-	    *locp++ = LINE_INVALID_LOCATION;
-	    str_table = (byte *) locp;
+            }
+            *locp++ = LINE_INVALID_LOCATION;
+            str_table = (byte *) locp;
 	} else {
-	    Uint32* locp = (Uint32 *) p;
-	    ASSERT(stp->loc_size == 4);
+	    Uint32* locp = (Uint32 *) &line_tab->fname_ptr[stp->num_fnames];
+            ASSERT(stp->loc_size == 4);
+            line_tab->loc_tab.p4 = locp;
 	    for (i = 0; i < num_instrs; i++) {
 		*locp++ = stp->line_instr[i].loc;
 	    }
 	    *locp++ = LINE_INVALID_LOCATION;
-	    str_table = (byte *) locp;
+            str_table = (byte *) locp;
 	}
-
 	CHKBLK(ERTS_ALC_T_CODE,code);
     }
 
@@ -4497,13 +4566,13 @@ freeze_code(LoaderState* stp)
     if (attr_size) {
 	byte* attr = str_table + strtab_size;
 	sys_memcpy(attr, stp->chunks[ATTR_CHUNK].start, stp->chunks[ATTR_CHUNK].size);
-	code[MI_ATTR_PTR] = (BeamInstr) attr;
-	code[MI_ATTR_SIZE] = (BeamInstr) stp->chunks[ATTR_CHUNK].size;
+	code_hdr->attr_ptr = attr;
+	code_hdr->attr_size = (BeamInstr) stp->chunks[ATTR_CHUNK].size;
 	decoded_size = erts_decode_ext_size(attr, attr_size);
 	if (decoded_size < 0) {
  	    LoadError0(stp, "bad external term representation of module attributes");
  	}
-	code[MI_ATTR_SIZE_ON_HEAP] = decoded_size;
+	code_hdr->attr_size_on_heap = decoded_size;
     }
     CHKBLK(ERTS_ALC_T_CODE,code);
     if (compile_size) {
@@ -4513,9 +4582,9 @@ freeze_code(LoaderState* stp)
 	       stp->chunks[COMPILE_CHUNK].size);
 
 	CHKBLK(ERTS_ALC_T_CODE,code);
-	code[MI_COMPILE_PTR] = (BeamInstr) compile_info;
+	code_hdr->compile_ptr = compile_info;
 	CHKBLK(ERTS_ALC_T_CODE,code);
-	code[MI_COMPILE_SIZE] = (BeamInstr) stp->chunks[COMPILE_CHUNK].size;
+	code_hdr->compile_size = (BeamInstr) stp->chunks[COMPILE_CHUNK].size;
 	CHKBLK(ERTS_ALC_T_CODE,code);
 	decoded_size = erts_decode_ext_size(compile_info, compile_size);
 	CHKBLK(ERTS_ALC_T_CODE,code);
@@ -4523,7 +4592,7 @@ freeze_code(LoaderState* stp)
  	    LoadError0(stp, "bad external term representation of compilation information");
  	}
 	CHKBLK(ERTS_ALC_T_CODE,code);
-	code[MI_COMPILE_SIZE_ON_HEAP] = decoded_size;
+	code_hdr->compile_size_on_heap = decoded_size;
     }
     CHKBLK(ERTS_ALC_T_CODE,code);
     {
@@ -4531,7 +4600,7 @@ freeze_code(LoaderState* stp)
 	CHKBLK(ERTS_ALC_T_CODE,code);
 	sys_memcpy(md5_sum, stp->mod_md5, MD5_SIZE);
 	CHKBLK(ERTS_ALC_T_CODE,code);
-	code[MI_MD5_PTR] = (BeamInstr) md5_sum;
+	code_hdr->md5_ptr = md5_sum;
 	CHKBLK(ERTS_ALC_T_CODE,code);
     }
     CHKBLK(ERTS_ALC_T_CODE,code);
@@ -4540,7 +4609,7 @@ freeze_code(LoaderState* stp)
      * Make sure that we have not overflowed the allocated code space.
      */
     ASSERT(str_table + strtab_size + attr_size + compile_size + MD5_SIZE ==
-	   ((byte *) code) + size);
+	   ((byte *) code_hdr) + size);
 
     /*
      * Patch all instructions that refer to the string table.
@@ -4552,46 +4621,47 @@ freeze_code(LoaderState* stp)
 	    BeamInstr* op_ptr;
 	    byte* strp;
 
-	    op_ptr = code + sp->pos;
+	    op_ptr = codev + sp->pos;
 	    strp = str_table + op_ptr[0];
 	    op_ptr[0] = (BeamInstr) strp;
 	    sp = sp->next;
 	}
     }
-    CHKBLK(ERTS_ALC_T_CODE,code);
+    CHKBLK(ERTS_ALC_T_CODE,code_hdr);
 
     /*
      * Resolve all labels.
      */
 
     for (i = 0; i < stp->num_labels; i++) {
-	Uint this_patch;
-	Uint next_patch;
+	Sint this_patch;
+	Sint next_patch;
 	Uint value = stp->labels[i].value;
 	
-	if (value == 0 && stp->labels[i].patches != 0) {
+	if (value == 0 && stp->labels[i].patches >= 0) {
 	    LoadError1(stp, "label %d not resolved", i);
 	}
 	ASSERT(value < stp->ci);
 	this_patch = stp->labels[i].patches;
-	while (this_patch != 0) {
+	while (this_patch >= 0) {
 	    ASSERT(this_patch < stp->ci);
-	    next_patch = code[this_patch];
+	    next_patch = codev[this_patch];
 	    ASSERT(next_patch < stp->ci);
-	    code[this_patch] = (BeamInstr) (code + value);
+	    codev[this_patch] = (BeamInstr) (codev + value);
 	    this_patch = next_patch;
 	}
     }
-    CHKBLK(ERTS_ALC_T_CODE,code);
+    CHKBLK(ERTS_ALC_T_CODE,code_hdr);
 
     /*
      * Save the updated code pointer and code size.
      */
 
-    stp->code = code;
+    stp->hdr = code_hdr;
+    stp->codev = codev;
     stp->loaded_size = size;
 
-    CHKBLK(ERTS_ALC_T_CODE,code);
+    CHKBLK(ERTS_ALC_T_CODE,code_hdr);
     return 1;
 
  load_error:
@@ -4599,7 +4669,8 @@ freeze_code(LoaderState* stp)
      * Make sure that the caller frees the newly reallocated block, and
      * not the old one (in case it has moved).
      */
-    stp->code = code;
+    stp->hdr = code_hdr;
+    stp->codev = codev;
     return 0;
 }
 
@@ -4610,7 +4681,7 @@ final_touch(LoaderState* stp)
     int on_load = stp->on_load;
     unsigned catches;
     Uint index;
-    BeamInstr* code = stp->code;
+    BeamInstr* codev = stp->codev;
     Module* modp;
 
     /*
@@ -4620,10 +4691,10 @@ final_touch(LoaderState* stp)
     index = stp->catches;
     catches = BEAM_CATCHES_NIL;
     while (index != 0) {
-	BeamInstr next = code[index];
-	code[index] = BeamOpCode(op_catch_yf);
-	catches = beam_catches_cons((BeamInstr *)code[index+2], catches);
-	code[index+2] = make_catch(catches);
+	BeamInstr next = codev[index];
+	codev[index] = BeamOpCode(op_catch_yf);
+	catches = beam_catches_cons((BeamInstr *)codev[index+2], catches);
+	codev[index+2] = make_catch(catches);
 	index = next;
     }
     modp = erts_put_module(stp->module);
@@ -4674,8 +4745,8 @@ final_touch(LoaderState* stp)
 	current = stp->import[i].patches;
 	while (current != 0) {
 	    ASSERT(current < stp->ci);
-	    next = stp->code[current];
-	    stp->code[current] = import;
+	    next = stp->codev[current];
+	    stp->codev[current] = import;
 	    current = next;
 	}
     }
@@ -4688,7 +4759,7 @@ final_touch(LoaderState* stp)
 	for (i = 0; i < stp->num_lambdas; i++) {
 	    unsigned entry_label = stp->lambdas[i].label;
 	    ErlFunEntry* fe = stp->lambdas[i].fe;
-	    BeamInstr* code_ptr = (BeamInstr *) (stp->code + stp->labels[entry_label].value);
+	    BeamInstr* code_ptr = stp->codev + stp->labels[entry_label].value;
 
 	    if (fe->address[0] != 0) {
 		/*
@@ -4801,7 +4872,8 @@ transform_engine(LoaderState* st)
 	    if (var[i].type != instr->a[ap].type)
 		goto restart;
 	    switch (var[i].type) {
-	    case TAG_r: case TAG_n: break;
+	    case TAG_n:
+		break;
 	    default:
 		if (var[i].val != instr->a[ap].val)
 		    goto restart;
@@ -5311,7 +5383,7 @@ new_label(LoaderState* stp)
 					 (void *) stp->labels,
 					 stp->num_labels * sizeof(Label));
     stp->labels[num].value = 0;
-    stp->labels[num].patches = 0;
+    stp->labels[num].patches = -1;
     return num;
 }
 
@@ -5371,7 +5443,7 @@ erts_module_info_0(Process* p, Eterm module)
 {
     Module* modp;
     ErtsCodeIndex code_ix = erts_active_code_ix();
-    BeamInstr* code;
+    BeamCodeHeader* code_hdr;
     Eterm *hp;
     Eterm list = NIL;
     Eterm tup;
@@ -5385,13 +5457,13 @@ erts_module_info_0(Process* p, Eterm module)
 	return THE_NON_VALUE;
     }
 
-    code = modp->curr.code;
-    if (code == NULL) {
+    code_hdr = modp->curr.code_hdr;
+    if (code_hdr == NULL) {
         return THE_NON_VALUE;
     }
 
 #define BUILD_INFO(What) \
-    tup = get_module_info(p, code_ix, code, module, What); \
+    tup = get_module_info(p, code_ix, code_hdr, module, What); \
     hp = HAlloc(p, 5); \
     tup = TUPLE2(hp, What, tup); \
     hp += 3; \
@@ -5414,7 +5486,7 @@ erts_module_info_1(Process* p, Eterm module, Eterm what)
 {
     Module* modp;
     ErtsCodeIndex code_ix = erts_active_code_ix();
-    BeamInstr* code;
+    BeamCodeHeader* code_hdr;
 
     if (is_not_atom(module)) {
         return THE_NON_VALUE;
@@ -5425,34 +5497,34 @@ erts_module_info_1(Process* p, Eterm module, Eterm what)
         return THE_NON_VALUE;
     }
 
-    code = modp->curr.code;
-    if (code == NULL) {
+    code_hdr = modp->curr.code_hdr;
+    if (code_hdr == NULL) {
         return THE_NON_VALUE;
     }
 
-    return get_module_info(p, code_ix, code, module, what);
+    return get_module_info(p, code_ix, code_hdr, module, what);
 }
 
 static Eterm
-get_module_info(Process* p, ErtsCodeIndex code_ix, BeamInstr* code,
+get_module_info(Process* p, ErtsCodeIndex code_ix, BeamCodeHeader* code_hdr,
                 Eterm module, Eterm what)
 {
     if (what == am_module) {
 	return module;
     } else if (what == am_md5) {
-	return md5_of_module(p, code);
+	return md5_of_module(p, code_hdr);
     } else if (what == am_exports) {
 	return exported_from_module(p, code_ix, module);
     } else if (what == am_functions) {
-	return functions_in_module(p, code);
+	return functions_in_module(p, code_hdr);
     } else if (what == am_attributes) {
-	return attributes_for_module(p, code);
+	return attributes_for_module(p, code_hdr);
     } else if (what == am_compile) {
-	return compilation_info_for_module(p, code);
+	return compilation_info_for_module(p, code_hdr);
     } else if (what == am_native_addresses) {
-	return native_addresses(p, code);
+	return native_addresses(p, code_hdr);
     } else if (what == am_native) {
-	return has_native(code);
+	return has_native(code_hdr);
     }
     return THE_NON_VALUE;
 }
@@ -5464,7 +5536,7 @@ get_module_info(Process* p, ErtsCodeIndex code_ix, BeamInstr* code,
 
 Eterm
 functions_in_module(Process* p, /* Process whose heap to use. */
-		    BeamInstr* code)
+		    BeamCodeHeader* code_hdr)
 {
     int i;
     Uint num_functions;
@@ -5473,12 +5545,12 @@ functions_in_module(Process* p, /* Process whose heap to use. */
     Eterm* hp_end;
     Eterm result = NIL;
 
-    num_functions = code[MI_NUM_FUNCTIONS];
+    num_functions = code_hdr->num_functions;
     need = 5*num_functions;
     hp = HAlloc(p, need);
     hp_end = hp + need;
     for (i = num_functions-1; i >= 0 ; i--) {
-	BeamInstr* func_info = (BeamInstr *) code[MI_FUNCTIONS+i];
+	BeamInstr* func_info = code_hdr->functions[i];
 	Eterm name = (Eterm) func_info[3];
 	int arity = (int) func_info[4];
 	Eterm tuple;
@@ -5504,11 +5576,11 @@ functions_in_module(Process* p, /* Process whose heap to use. */
  */
 
 static Eterm
-has_native(BeamInstr *code)
+has_native(BeamCodeHeader *code_hdr)
 {
     Eterm result = am_false;
 #ifdef HIPE
-    if (erts_is_module_native(code)) {
+    if (erts_is_module_native(code_hdr)) {
         result = am_true;
     }
 #endif
@@ -5516,15 +5588,15 @@ has_native(BeamInstr *code)
 }
 
 int
-erts_is_module_native(BeamInstr* code)
+erts_is_module_native(BeamCodeHeader* code_hdr)
 {
     Uint i, num_functions;
 
     /* Check NativeAdress of first real function in module */
-    if (code != NULL) {
-        num_functions = code[MI_NUM_FUNCTIONS];
+    if (code_hdr != NULL) {
+        num_functions = code_hdr->num_functions;
         for (i=0; i<num_functions; i++) {
-            BeamInstr* func_info = (BeamInstr *) code[MI_FUNCTIONS+i];
+            BeamInstr* func_info = (BeamInstr *) code_hdr->functions[i];
             Eterm name = (Eterm) func_info[3];
             if (is_atom(name)) {
                 return func_info[1] != 0;
@@ -5541,7 +5613,7 @@ erts_is_module_native(BeamInstr* code)
  */
 
 static Eterm
-native_addresses(Process* p, BeamInstr* code)
+native_addresses(Process* p, BeamCodeHeader* code_hdr)
 {
     int i;
     Eterm* hp;
@@ -5550,12 +5622,12 @@ native_addresses(Process* p, BeamInstr* code)
     Eterm* hp_end;
     Eterm result = NIL;
 
-    num_functions = code[MI_NUM_FUNCTIONS];
+    num_functions = code_hdr->num_functions;
     need = (6+BIG_UINT_HEAP_SIZE)*num_functions;
     hp = HAlloc(p, need);
     hp_end = hp + need;
     for (i = num_functions-1; i >= 0 ; i--) {
-	BeamInstr* func_info = (BeamInstr *) code[MI_FUNCTIONS+i];
+	BeamInstr* func_info = code_hdr->functions[i];
 	Eterm name = (Eterm) func_info[3];
 	int arity = (int) func_info[4];
 	Eterm tuple;
@@ -5621,15 +5693,15 @@ exported_from_module(Process* p, /* Process whose heap to use. */
 
 Eterm
 attributes_for_module(Process* p, /* Process whose heap to use. */
-                      BeamInstr* code)
+                      BeamCodeHeader* code_hdr)
 {
     byte* ext;
     Eterm result = NIL;
 
-    ext = (byte *) code[MI_ATTR_PTR];
+    ext = code_hdr->attr_ptr;
     if (ext != NULL) {
 	ErtsHeapFactory factory;
-	erts_factory_proc_prealloc_init(&factory, p, code[MI_ATTR_SIZE_ON_HEAP]);
+	erts_factory_proc_prealloc_init(&factory, p, code_hdr->attr_size_on_heap);
 	result = erts_decode_ext(&factory, &ext);
 	if (is_value(result)) {
 	    erts_factory_close(&factory);
@@ -5644,15 +5716,15 @@ attributes_for_module(Process* p, /* Process whose heap to use. */
 
 Eterm
 compilation_info_for_module(Process* p, /* Process whose heap to use. */
-                            BeamInstr* code)
+                            BeamCodeHeader* code_hdr)
 {
     byte* ext;
     Eterm result = NIL;
 
-    ext = (byte *) code[MI_COMPILE_PTR];
+    ext = code_hdr->compile_ptr;
     if (ext != NULL) {
 	ErtsHeapFactory factory;
-	erts_factory_proc_prealloc_init(&factory, p, code[MI_COMPILE_SIZE_ON_HEAP]);
+	erts_factory_proc_prealloc_init(&factory, p, code_hdr->compile_size_on_heap);
 	result = erts_decode_ext(&factory, &ext);
 	if (is_value(result)) {
 	    erts_factory_close(&factory);
@@ -5667,9 +5739,9 @@ compilation_info_for_module(Process* p, /* Process whose heap to use. */
 
 Eterm
 md5_of_module(Process* p, /* Process whose heap to use. */
-              BeamInstr* code)
+              BeamCodeHeader* code_hdr)
 {
-    return new_binary(p, (byte *) code[MI_MD5_PTR], MD5_SIZE);
+    return new_binary(p, code_hdr->md5_ptr, MD5_SIZE);
 }
 
 /*
@@ -5868,7 +5940,7 @@ make_stub(BeamInstr* fp, Eterm mod, Eterm func, Uint arity, Uint native, BeamIns
     fp[4] = arity;
 #ifdef HIPE
     if (native) {
-	fp[5] = BeamOpCode(op_move_return_nr);
+	fp[5] = BeamOpCode(op_move_return_n);
 	hipe_mfa_save_orig_beam_op(mod, func, arity, fp+5);
     }
 #endif
@@ -5880,7 +5952,7 @@ static byte*
 stub_copy_info(LoaderState* stp,
 	       int chunk,	/* Chunk: ATTR_CHUNK or COMPILE_CHUNK */
 	       byte* info,	/* Where to store info. */
-	       BeamInstr* ptr_word,	/* Where to store pointer into info. */
+	       byte** ptr_word,	/* Where to store pointer into info. */
 	       BeamInstr* size_word,	/* Where to store size into info. */
 	       BeamInstr* size_on_heap_word) /* Where to store size on heap. */
 {
@@ -5888,7 +5960,7 @@ stub_copy_info(LoaderState* stp,
     Uint size = stp->chunks[chunk].size;
     if (size != 0) {
 	memcpy(info, stp->chunks[chunk].start, size);
-	*ptr_word = (BeamInstr) info;
+	*ptr_word = info;
 	decoded_size = erts_decode_ext_size(info, size);
 	if (decoded_size < 0) {
  	    return 0;
@@ -6143,11 +6215,10 @@ erts_make_stub_module(Process* p, Eterm Mod, Eterm Beam, Eterm Info)
     BeamInstr Patchlist;
     Eterm MD5Bin;
     Eterm* tp;
-    BeamInstr* code = NULL;
-    BeamInstr* ptrs;
+    BeamCodeHeader* code_hdr;
+    BeamInstr* code_base;
     BeamInstr* fp;
     byte* info;
-    Uint ci;
     int n;
     int code_size;
     int rval;
@@ -6225,40 +6296,40 @@ erts_make_stub_module(Process* p, Eterm Mod, Eterm Beam, Eterm Info)
      * Allocate memory for the stub module.
      */
 
-    code_size = ((WORDS_PER_FUNCTION+1)*n + MI_FUNCTIONS + 2) * sizeof(BeamInstr);
-    code_size += stp->chunks[ATTR_CHUNK].size;
-    code_size += stp->chunks[COMPILE_CHUNK].size;
-    code_size += MD5_SIZE;
-    code = erts_alloc_fnf(ERTS_ALC_T_CODE, code_size);
-    if (!code) {
+    code_size = (offsetof(BeamCodeHeader,functions)
+                 + ((n+1) * sizeof(BeamInstr*))
+                 + (WORDS_PER_FUNCTION*n + 1) * sizeof(BeamInstr)
+                 + stp->chunks[ATTR_CHUNK].size
+                 + stp->chunks[COMPILE_CHUNK].size
+                 + MD5_SIZE);
+    code_hdr = erts_alloc_fnf(ERTS_ALC_T_CODE, code_size);
+    if (!code_hdr) {
 	goto error;
     }
 
     /*
-     * Initialize code area.
+     * Initialize code header.
      */
 
-    code[MI_NUM_FUNCTIONS] = n;
-    code[MI_ATTR_PTR] = 0;
-    code[MI_ATTR_SIZE] = 0;
-    code[MI_ATTR_SIZE_ON_HEAP] = 0;
-    code[MI_COMPILE_PTR] = 0;
-    code[MI_COMPILE_SIZE] = 0;
-    code[MI_COMPILE_SIZE_ON_HEAP] = 0;
-    code[MI_LITERALS_START] = 0;
-    code[MI_LITERALS_END] = 0;
-    code[MI_LITERALS_OFF_HEAP] = 0;
-    code[MI_ON_LOAD_FUNCTION_PTR] = 0;
-    code[MI_LINE_TABLE] = 0;
-    code[MI_MD5_PTR] = 0;
-    ci = MI_FUNCTIONS + n + 1;
+    code_hdr->num_functions = n;
+    code_hdr->attr_ptr = NULL;
+    code_hdr->attr_size = 0;
+    code_hdr->attr_size_on_heap = 0;
+    code_hdr->compile_ptr = NULL;
+    code_hdr->compile_size = 0;
+    code_hdr->compile_size_on_heap = 0;
+    code_hdr->literals_start = NULL;
+    code_hdr->literals_end = NULL;
+    code_hdr->literals_off_heap = 0;
+    code_hdr->on_load_function_ptr = NULL;
+    code_hdr->line_table = NULL;
+    code_hdr->md5_ptr = NULL;
 
     /*
      * Make stubs for all functions.
      */
 
-    ptrs = code + MI_FUNCTIONS;
-    fp = code + ci;
+    fp = code_base = (BeamInstr*) &code_hdr->functions[n+1];
     for (i = 0; i < n; i++) {
 	Eterm* listp;
 	Eterm tuple;
@@ -6301,11 +6372,11 @@ erts_make_stub_module(Process* p, Eterm Mod, Eterm Beam, Eterm Info)
 	 * Set the pointer and make the stub. Put a return instruction
 	 * as the body until we know what kind of trap we should put there.
 	 */
-	ptrs[i] = (BeamInstr) fp;
+	code_hdr->functions[i] = fp;
 #ifdef HIPE
 	op = (Eterm) BeamOpCode(op_hipe_trap_call); /* Might be changed later. */
 #else
-	op = (Eterm) BeamOpCode(op_move_return_nr);
+	op = (Eterm) BeamOpCode(op_move_return_n);
 #endif
 	fp = make_stub(fp, Mod, func, arity, (Uint)native_address, op);
     }
@@ -6314,7 +6385,7 @@ erts_make_stub_module(Process* p, Eterm Mod, Eterm Beam, Eterm Info)
      * Insert the last pointer and the int_code_end instruction.
      */
 
-    ptrs[i] = (BeamInstr) fp;
+    code_hdr->functions[i] = fp;
     *fp++ = (BeamInstr) BeamOp(op_int_code_end);
 
     /*
@@ -6323,16 +6394,16 @@ erts_make_stub_module(Process* p, Eterm Mod, Eterm Beam, Eterm Info)
 
     info = (byte *) fp;
     info = stub_copy_info(stp, ATTR_CHUNK, info,
-			  code+MI_ATTR_PTR,
-			  code+MI_ATTR_SIZE,
-			  code+MI_ATTR_SIZE_ON_HEAP);
+			  &code_hdr->attr_ptr,
+			  &code_hdr->attr_size,
+			  &code_hdr->attr_size_on_heap);
     if (info == NULL) {
 	goto error;
     }
     info = stub_copy_info(stp, COMPILE_CHUNK, info,
-			  code+MI_COMPILE_PTR,
-			  code+MI_COMPILE_SIZE,
-			  code+MI_COMPILE_SIZE_ON_HEAP);
+			  &code_hdr->compile_ptr,
+			  &code_hdr->compile_size,
+			  &code_hdr->compile_size_on_heap);
     if (info == NULL) {
 	goto error;
     }
@@ -6341,7 +6412,7 @@ erts_make_stub_module(Process* p, Eterm Mod, Eterm Beam, Eterm Info)
       byte *md5 = NULL;
       if ((md5 = erts_get_aligned_binary_bytes(MD5Bin, &tmp)) != NULL) {
         sys_memcpy(info, md5, MD5_SIZE);
-        code[MI_MD5_PTR] = (BeamInstr) info;
+        code_hdr->md5_ptr = info;
       }
       erts_free_aligned_binary_bytes(tmp);
     }
@@ -6350,7 +6421,7 @@ erts_make_stub_module(Process* p, Eterm Mod, Eterm Beam, Eterm Info)
      * Insert the module in the module table.
      */
 
-    rval = insert_new_code(p, 0, p->group_leader, Mod, code, code_size);
+    rval = insert_new_code(p, 0, p->group_leader, Mod, code_hdr, code_size);
     if (rval != NIL) {
 	goto error;
     }
@@ -6359,7 +6430,7 @@ erts_make_stub_module(Process* p, Eterm Mod, Eterm Beam, Eterm Info)
      * Export all stub functions and insert the correct type of HiPE trap.
      */
 
-    fp = code + ci;
+    fp = code_base;
     for (i = 0; i < n; i++) {
 	stub_final_touch(stp, fp);
 	fp += WORDS_PER_FUNCTION;
diff --git a/erts/emulator/beam/beam_load.h b/erts/emulator/beam/beam_load.h
index d5af634fad..22ab71c868 100644
--- a/erts/emulator/beam/beam_load.h
+++ b/erts/emulator/beam/beam_load.h
@@ -24,7 +24,6 @@
 #include "beam_opcodes.h"
 #include "erl_process.h"
 
-int erts_is_module_native(BeamInstr* code);
 Eterm beam_make_current_old(Process *c_p, ErtsProcLocks c_p_locks,
 			    Eterm module);
 
@@ -52,6 +51,7 @@ extern BeamInstr* em_call_error_handler;
 extern BeamInstr* em_apply_bif;
 extern BeamInstr* em_call_nif;
 
+
 /*
  * The following variables keep a sorted list of address ranges for
  * each module.  It allows us to quickly find a function given an
@@ -60,72 +60,81 @@ extern BeamInstr* em_call_nif;
 
 /* Total code size in bytes */
 extern Uint erts_total_code_size;
-/*
- * Index into start of code chunks which contains additional information
- * about the loaded module.
- *
- * First number of functions.
- */
 
-#define MI_NUM_FUNCTIONS     0
+typedef struct BeamCodeLineTab_ BeamCodeLineTab;
 
 /*
- * The attributes retrieved by Mod:module_info(attributes).
- */
-
-#define MI_ATTR_PTR          1
-#define MI_ATTR_SIZE	     2
-#define MI_ATTR_SIZE_ON_HEAP 3
-
-/*
- * The compilation information retrieved by Mod:module_info(compile).
- */
-
-#define MI_COMPILE_PTR          4
-#define MI_COMPILE_SIZE         5
-#define MI_COMPILE_SIZE_ON_HEAP 6
-
-/*
- * Literal area (constant pool).
- */
-#define MI_LITERALS_START	7
-#define MI_LITERALS_END		8
-#define MI_LITERALS_OFF_HEAP	9
-
-/*
- * Pointer to the on_load function (or NULL if none).
- */
-#define MI_ON_LOAD_FUNCTION_PTR 10
-
-/*
- * Pointer to the line table (or NULL if none).
- */
-#define MI_LINE_TABLE 11
-
-/*
- * Pointer to the module MD5 sum (16 bytes)
- */
-#define MI_MD5_PTR 12
-
-/*
- * Start of function pointer table.  This table contains pointers to
- * all functions in the module plus an additional pointer just beyond
- * the end of the last function.
- *
- * The actual loaded code (for the first function) start just beyond
- * this table.
+ * Header of code chunks which contains additional information
+ * about the loaded module.
  */
-
-#define MI_FUNCTIONS         13
+typedef struct beam_code_header {
+    /*
+     * Number of functions.
+     */
+    UWord num_functions;
+
+    /*
+     * The attributes retrieved by Mod:module_info(attributes).
+     */
+    byte* attr_ptr;
+    UWord attr_size;
+    UWord attr_size_on_heap;
+
+    /*
+     * The compilation information retrieved by Mod:module_info(compile).
+     */
+    byte* compile_ptr;
+    UWord compile_size;
+    UWord compile_size_on_heap;
+
+    /*
+     * Literal area (constant pool).
+     */
+    Eterm* literals_start;
+    Eterm* literals_end;
+    struct erl_off_heap_header* literals_off_heap;
+
+    /*
+     * Pointer to the on_load function (or NULL if none).
+     */
+    BeamInstr* on_load_function_ptr;
+
+    /*
+     * Pointer to the line table (or NULL if none).
+     */
+    BeamCodeLineTab* line_table;
+
+    /*
+     * Pointer to the module MD5 sum (16 bytes)
+     */
+    byte* md5_ptr;
+
+    /*
+     * Start of function pointer table.  This table contains pointers to
+     * all functions in the module plus an additional pointer just beyond
+     * the end of the last function.
+     *
+     * The actual loaded code (for the first function) start just beyond
+     * this table.
+     */
+    BeamInstr* functions[1];
+
+}BeamCodeHeader;
+
+int erts_is_module_native(BeamCodeHeader* code);
 
 /*
  * Layout of the line table.
  */
-
-#define MI_LINE_FNAME_PTR 0
-#define MI_LINE_LOC_TAB 1
-#define MI_LINE_LOC_SIZE 2
-#define MI_LINE_FUNC_TAB 3
+struct BeamCodeLineTab_ {
+    Eterm* fname_ptr;
+    int loc_size;
+    union {
+        Uint16* p2;
+        Uint32* p4;
+    }loc_tab;
+    const BeamInstr** func_tab[1];
+};
 
 #define LINE_INVALID_LOCATION (0)
 
diff --git a/erts/emulator/beam/beam_ranges.c b/erts/emulator/beam/beam_ranges.c
index 19079ba150..5a2b66727a 100644
--- a/erts/emulator/beam/beam_ranges.c
+++ b/erts/emulator/beam/beam_ranges.c
@@ -37,8 +37,8 @@ typedef struct {
 #define RANGE_END(R) ((BeamInstr*)erts_smp_atomic_read_nob(&(R)->end))
 
 static Range* find_range(BeamInstr* pc);
-static void lookup_loc(FunctionInfo* fi, BeamInstr* pc,
-		       BeamInstr* modp, int idx);
+static void lookup_loc(FunctionInfo* fi, const BeamInstr* pc,
+		       BeamCodeHeader*, int idx);
 
 /*
  * The following variables keep a sorted list of address ranges for
@@ -241,6 +241,7 @@ erts_lookup_function_info(FunctionInfo* fi, BeamInstr* pc, int full_info)
     BeamInstr** high;
     BeamInstr** mid;
     Range* rp;
+    BeamCodeHeader* hdr;
 
     fi->current = NULL;
     fi->needed = 5;
@@ -249,9 +250,10 @@ erts_lookup_function_info(FunctionInfo* fi, BeamInstr* pc, int full_info)
     if (rp == 0) {
 	return;
     }
+    hdr = (BeamCodeHeader*) rp->start;
 
-    low = (BeamInstr **) (rp->start + MI_FUNCTIONS);
-    high = low + rp->start[MI_NUM_FUNCTIONS];
+    low = hdr->functions;
+    high = low + hdr->num_functions;
     while (low < high) {
 	mid = low + (high-low) / 2;
 	if (pc < mid[0]) {
@@ -259,10 +261,9 @@ erts_lookup_function_info(FunctionInfo* fi, BeamInstr* pc, int full_info)
 	} else if (pc < mid[1]) {
 	    fi->current = mid[0]+2;
 	    if (full_info) {
-		BeamInstr** fp = (BeamInstr **) (rp->start +
-						 MI_FUNCTIONS);
+		BeamInstr** fp = hdr->functions;
 		int idx = mid - fp;
-		lookup_loc(fi, pc, rp->start, idx);
+		lookup_loc(fi, pc, hdr, idx);
 	    }
 	    return;
 	} else {
@@ -295,39 +296,34 @@ find_range(BeamInstr* pc)
 }
 
 static void
-lookup_loc(FunctionInfo* fi, BeamInstr* orig_pc, BeamInstr* modp, int idx)
+lookup_loc(FunctionInfo* fi, const BeamInstr* pc,
+           BeamCodeHeader* code_hdr, int idx)
 {
-    Eterm* line = (Eterm *) modp[MI_LINE_TABLE];
-    Eterm* low;
-    Eterm* high;
-    Eterm* mid;
-    Eterm pc;
+    BeamCodeLineTab* lt = code_hdr->line_table;
+    const BeamInstr** low;
+    const BeamInstr** high;
+    const BeamInstr** mid;
 
-    if (line == 0) {
+    if (lt == NULL) {
 	return;
     }
 
-    pc = (Eterm) (BeamInstr) orig_pc;
-    fi->fname_ptr = (Eterm *) (BeamInstr) line[MI_LINE_FNAME_PTR];
-    low = (Eterm *) (BeamInstr) line[MI_LINE_FUNC_TAB+idx];
-    high = (Eterm *) (BeamInstr) line[MI_LINE_FUNC_TAB+idx+1];
+    fi->fname_ptr = lt->fname_ptr;
+    low = lt->func_tab[idx];
+    high = lt->func_tab[idx+1];
     while (high > low) {
 	mid = low + (high-low) / 2;
 	if (pc < mid[0]) {
 	    high = mid;
 	} else if (pc < mid[1]) {
 	    int file;
-	    int index = mid - (Eterm *) (BeamInstr) line[MI_LINE_FUNC_TAB];
+	    int index = mid - lt->func_tab[0];
 
-	    if (line[MI_LINE_LOC_SIZE] == 2) {
-		Uint16* loc_table =
-		    (Uint16 *) (BeamInstr) line[MI_LINE_LOC_TAB];
-		fi->loc = loc_table[index];
+	    if (lt->loc_size == 2) {
+		fi->loc = lt->loc_tab.p2[index];
 	    } else {
-		Uint32* loc_table =
-		    (Uint32 *) (BeamInstr) line[MI_LINE_LOC_TAB];
-		ASSERT(line[MI_LINE_LOC_SIZE] == 4);
-		fi->loc = loc_table[index];
+		ASSERT(lt->loc_size == 4);
+		fi->loc = lt->loc_tab.p4[index];
 	    }
 	    if (fi->loc == LINE_INVALID_LOCATION) {
 		return;
diff --git a/erts/emulator/beam/bif.c b/erts/emulator/beam/bif.c
index 4e3a1cef69..bb9165cd79 100644
--- a/erts/emulator/beam/bif.c
+++ b/erts/emulator/beam/bif.c
@@ -72,7 +72,7 @@ BIF_RETTYPE spawn_3(BIF_ALIST_3)
     ErlSpawnOpts so;
     Eterm pid;
 
-    so.flags = 0;
+    so.flags = erts_default_spo_flags;
     pid = erl_create_process(BIF_P, BIF_ARG_1, BIF_ARG_2, BIF_ARG_3, &so);
     if (is_non_value(pid)) {
 	BIF_ERROR(BIF_P, so.error_code);
@@ -589,7 +589,7 @@ erts_queue_monitor_message(Process *p,
     Eterm reason_copy, ref_copy, item_copy;
     Uint reason_size, ref_size, item_size, heap_size;
     ErlOffHeap *ohp;
-    ErlHeapFragment *bp;
+    ErtsMessage *msgp;
 
     reason_size = IS_CONST(reason) ? 0 : size_object(reason);
     item_size   = IS_CONST(item) ? 0 : size_object(item);
@@ -597,11 +597,8 @@ erts_queue_monitor_message(Process *p,
 
     heap_size = 6+reason_size+ref_size+item_size;
 
-    hp = erts_alloc_message_heap(heap_size,
-				 &bp,
-				 &ohp,
-				 p,
-				 p_locksp);
+    msgp = erts_alloc_message_heap(p, p_locksp, heap_size,
+				   &hp, &ohp);
 
     reason_copy = (IS_CONST(reason)
 		   ? reason
@@ -612,11 +609,11 @@ erts_queue_monitor_message(Process *p,
     ref_copy    = copy_struct(ref, ref_size, &hp, ohp);
 
     tup = TUPLE5(hp, am_DOWN, ref_copy, type, item_copy, reason_copy);
-    erts_queue_message(p, p_locksp, bp, tup, NIL);
+    erts_queue_message(p, p_locksp, msgp, tup, NIL);
 }
 
 static BIF_RETTYPE
-local_pid_monitor(Process *p, Eterm target, Eterm mon_ref, int bool)
+local_pid_monitor(Process *p, Eterm target, Eterm mon_ref, int boolean)
 {
     BIF_RETTYPE ret;
     Process *rp;
@@ -634,7 +631,7 @@ local_pid_monitor(Process *p, Eterm target, Eterm mon_ref, int bool)
     if (!rp) {
 	erts_smp_proc_unlock(p, ERTS_PROC_LOCK_LINK);
 	p_locks &= ~ERTS_PROC_LOCK_LINK;
-	if (bool)
+	if (boolean)
 	    ret = am_false;
 	else
 	    erts_queue_monitor_message(p, &p_locks,
@@ -643,7 +640,7 @@ local_pid_monitor(Process *p, Eterm target, Eterm mon_ref, int bool)
     else {
 	ASSERT(rp != p);
 
-	if (bool)
+	if (boolean)
 	    ret = am_true;
 
 	erts_add_monitor(&ERTS_P_MONITORS(p), MON_ORIGIN, mon_ref, target, NIL);
@@ -841,7 +838,7 @@ BIF_RETTYPE spawn_link_3(BIF_ALIST_3)
     ErlSpawnOpts so;
     Eterm pid;
 
-    so.flags = SPO_LINK;
+    so.flags = erts_default_spo_flags|SPO_LINK;
     pid = erl_create_process(BIF_P, BIF_ARG_1, BIF_ARG_2, BIF_ARG_3, &so);
     if (is_non_value(pid)) {
 	BIF_ERROR(BIF_P, so.error_code);
@@ -878,7 +875,7 @@ BIF_RETTYPE spawn_opt_1(BIF_ALIST_1)
     /*
      * Store default values for options.
      */
-    so.flags          = SPO_USE_ARGS;
+    so.flags          = erts_default_spo_flags|SPO_USE_ARGS;
     so.min_heap_size  = H_MIN_SIZE;
     so.min_vheap_size = BIN_VH_MIN_SIZE;
     so.priority       = PRIORITY_NORMAL;
@@ -913,6 +910,22 @@ BIF_RETTYPE spawn_opt_1(BIF_ALIST_1)
 		    so.priority = PRIORITY_LOW;
 		else
 		    goto error;
+	    } else if (arg == am_message_queue_data) {
+		switch (val) {
+		case am_mixed:
+		    so.flags &= ~(SPO_OFF_HEAP_MSGQ|SPO_ON_HEAP_MSGQ);
+		    break;
+		case am_on_heap:
+		    so.flags &= ~SPO_OFF_HEAP_MSGQ;
+		    so.flags |= SPO_ON_HEAP_MSGQ;
+		    break;
+		case am_off_heap:
+		    so.flags &= ~SPO_ON_HEAP_MSGQ;
+		    so.flags |= SPO_OFF_HEAP_MSGQ;
+		    break;
+		default:
+		    goto error;
+		}
 	    } else if (arg == am_min_heap_size && is_small(val)) {
 		Sint min_heap_size = signed_val(val);
 		if (min_heap_size < 0) {
@@ -1691,6 +1704,12 @@ BIF_RETTYPE process_flag_2(BIF_ALIST_2)
        }
        BIF_RET(old_value);
    }
+   else if (BIF_ARG_1 == am_message_queue_data) {
+       old_value = erts_change_message_queue_management(BIF_P, BIF_ARG_2);
+       if (is_non_value(old_value))
+	   goto error;
+       BIF_RET(old_value);
+   }
    else if (BIF_ARG_1 == am_sensitive) {
        Uint is_sensitive;
        if (BIF_ARG_2 == am_true) {
@@ -1931,7 +1950,7 @@ do_send(Process *p, Eterm to, Eterm msg, Eterm *refp, ErtsSendContext* ctx)
     } else if (is_atom(to)) {
 	Eterm id = erts_whereis_name_to_id(p, to);
 
-	rp = erts_proc_lookup(id);
+	rp = erts_proc_lookup_raw(id);
 	if (rp) {
 	    if (IS_TRACED(p))
 		trace_send(p, to, msg);
@@ -2023,11 +2042,7 @@ do_send(Process *p, Eterm to, Eterm msg, Eterm *refp, ErtsSendContext* ctx)
 	if (ERTS_PROC_GET_SAVED_CALLS_BUF(p))
 	    save_calls(p, &exp_send);
 	
-	if (SEQ_TRACE_TOKEN(p) != NIL
-#ifdef USE_VM_PROBES
-	    && SEQ_TRACE_TOKEN(p) != am_have_dt_utag
-#endif
-	    ) {
+        if (have_seqtrace(SEQ_TRACE_TOKEN(p))) {
 	    seq_trace_update_send(p);
 	    seq_trace_output(SEQ_TRACE_TOKEN(p), msg, 
 			     SEQ_TRACE_SEND, portid, p);
@@ -3824,11 +3839,9 @@ BIF_RETTYPE now_0(BIF_ALIST_0)
 
 BIF_RETTYPE garbage_collect_0(BIF_ALIST_0)
 {
-    int reds;
-
     FLAGS(BIF_P) |= F_NEED_FULLSWEEP;
-    reds = erts_garbage_collect(BIF_P, 0, NULL, 0);
-    BIF_RET2(am_true, reds);
+    erts_garbage_collect(BIF_P, 0, NULL, 0);
+    BIF_RET(am_true);
 }
 
 /**********************************************************************/
@@ -4117,16 +4130,9 @@ BIF_RETTYPE make_fun_3(BIF_ALIST_3)
     if (arity < 0) {
 	goto error;
     }
-#if HALFWORD_HEAP
-    hp = HAlloc(BIF_P, 3);
-    hp[0] = HEADER_EXPORT;
-    /* Yes, May be misaligned, but X86_64 will fix it... */
-    *((Export **) (hp+1)) = erts_export_get_or_make_stub(BIF_ARG_1, BIF_ARG_2, (Uint) arity);
-#else
     hp = HAlloc(BIF_P, 2);
     hp[0] = HEADER_EXPORT;
     hp[1] = (Eterm) erts_export_get_or_make_stub(BIF_ARG_1, BIF_ARG_2, (Uint) arity);
-#endif
     BIF_RET(make_export(hp));
 }
 
@@ -4231,6 +4237,7 @@ BIF_RETTYPE list_to_pid_1(BIF_ALIST_1)
 	  goto bad;
       
       enp = erts_find_or_insert_node(dep->sysname, dep->creation);
+      ASSERT(enp != erts_this_node);
 
       etp = (ExternalThing *) HAlloc(BIF_P, EXTERNAL_THING_HEAD_SIZE + 1);
       etp->header = make_external_pid_header(1);
@@ -4486,7 +4493,7 @@ BIF_RETTYPE system_flag_2(BIF_ALIST_2)
 	}
     } else if (BIF_ARG_1 == make_small(1)) {
 	int i, max;
-	ErlMessage* mp;
+	ErtsMessage* mp;
 	erts_smp_proc_unlock(BIF_P, ERTS_PROC_LOCK_MAIN);
 	erts_smp_thr_progress_block();
 
@@ -4631,7 +4638,7 @@ BIF_RETTYPE hash_2(BIF_ALIST_2)
     if ((range = signed_val(BIF_ARG_2)) <= 0) {  /* [1..MAX_SMALL] */
 	BIF_ERROR(BIF_P, BADARG);
     }
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
     if (range > ((1L << 27) - 1))
 	BIF_ERROR(BIF_P, BADARG);
 #endif
@@ -4703,7 +4710,7 @@ BIF_RETTYPE phash2_2(BIF_ALIST_2)
     /*
      * Return either a small or a big. Use the heap for bigs if there is room.
      */
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
     BIF_RET(make_small(final_hash));
 #else
     if (IS_USMALL(0, final_hash)) {
diff --git a/erts/emulator/beam/bif.h b/erts/emulator/beam/bif.h
index c6ed60376a..a62eddf36b 100644
--- a/erts/emulator/beam/bif.h
+++ b/erts/emulator/beam/bif.h
@@ -54,22 +54,24 @@ extern Export *erts_convert_time_unit_trap;
 	(p)->fcalls = -CONTEXT_REDS;		\
 } while(0)
 
-
-#define ERTS_VBUMP_ALL_REDS(p)						\
+#define ERTS_VBUMP_ALL_REDS_INTERNAL(p, fcalls)				\
 do {									\
     if (!ERTS_PROC_GET_SAVED_CALLS_BUF((p))) {				\
-	if ((p)->fcalls > 0)						\
-	    ERTS_PROC_GET_SCHDATA((p))->virtual_reds += (p)->fcalls;	\
-	(p)->fcalls = 0;						\
+	if ((fcalls) > 0)						\
+	    ERTS_PROC_GET_SCHDATA((p))->virtual_reds += (fcalls);	\
+	(fcalls) = 0;							\
     }									\
     else {								\
-	if ((p)->fcalls > -CONTEXT_REDS)				\
+	if ((fcalls) > -CONTEXT_REDS)					\
 	    ERTS_PROC_GET_SCHDATA((p))->virtual_reds			\
-		+= ((p)->fcalls - (-CONTEXT_REDS));			\
-	(p)->fcalls = -CONTEXT_REDS;					\
+		+= ((fcalls) - (-CONTEXT_REDS));			\
+	(fcalls) = -CONTEXT_REDS;					\
     }									\
 } while(0)
 
+#define ERTS_VBUMP_ALL_REDS(p) \
+    ERTS_VBUMP_ALL_REDS_INTERNAL((p), (p)->fcalls)
+
 #define BUMP_REDS(p, gc) do {			   \
      ASSERT(p);		 			   \
      ERTS_SMP_LC_ASSERT(ERTS_PROC_LOCK_MAIN & erts_proc_lc_my_proc_locks(p));\
@@ -110,10 +112,34 @@ do {									\
     }									\
 } while(0)
 
-#define ERTS_BIF_REDS_LEFT(p)						\
+#define ERTS_VBUMP_LEAVE_REDS_INTERNAL(P, Reds, FCalls)			\
+    do {								\
+	if (ERTS_PROC_GET_SAVED_CALLS_BUF((P))) {			\
+	    int nreds__ = ((int)(Reds)) - CONTEXT_REDS;			\
+	    if ((FCalls) > nreds__) {					\
+		ERTS_PROC_GET_SCHDATA((P))->virtual_reds		\
+		    += (FCalls) - nreds__;				\
+		(FCalls) = nreds__;					\
+	    }								\
+	}								\
+	else {								\
+	    if ((FCalls) > (Reds)) {					\
+		ERTS_PROC_GET_SCHDATA((P))->virtual_reds		\
+		    += (FCalls) - (Reds);				\
+		(FCalls) = (Reds);					\
+	    }								\
+	}								\
+    } while (0)
+
+#define ERTS_VBUMP_LEAVE_REDS(P, Reds)					\
+    ERTS_VBUMP_LEAVE_REDS_INTERNAL(P, Reds, (P)->fcalls)
+
+#define ERTS_REDS_LEFT(p, FCalls)					\
   (ERTS_PROC_GET_SAVED_CALLS_BUF((p))					\
-   ? ((p)->fcalls > -CONTEXT_REDS ? ((p)->fcalls - (-CONTEXT_REDS)) : 0)\
-   : ((p)->fcalls > 0 ? (p)->fcalls : 0))
+   ? ((FCalls) > -CONTEXT_REDS ? ((FCalls) - (-CONTEXT_REDS)) : 0)	\
+   : ((FCalls) > 0 ? (FCalls) : 0))
+
+#define ERTS_BIF_REDS_LEFT(p) ERTS_REDS_LEFT(p, p->fcalls)
 
 #define BIF_RET2(x, gc) do {			\
     BUMP_REDS(BIF_P, (gc));			\
diff --git a/erts/emulator/beam/bif.tab b/erts/emulator/beam/bif.tab
index 4f0656d174..1b8ae8cef5 100644
--- a/erts/emulator/beam/bif.tab
+++ b/erts/emulator/beam/bif.tab
@@ -115,7 +115,7 @@ bif erlang:time_offset/0
 bif erlang:time_offset/1
 bif erlang:timestamp/0
 
-bif erlang:open_port/2
+bif erts_internal:open_port/2
 
 bif erlang:pid_to_list/1
 bif erlang:ports/0
@@ -125,7 +125,6 @@ bif erlang:process_flag/3
 bif erlang:process_info/1
 bif erlang:process_info/2
 bif erlang:processes/0
-bif erlang:purge_module/1
 bif erlang:put/2
 bif erlang:register/2
 bif erlang:registered/0
@@ -167,7 +166,7 @@ bif erts_internal:request_system_task/3
 bif erts_internal:check_process_code/2
 
 bif erts_internal:map_to_tuple_keys/1
-bif erts_internal:map_type/1
+bif erts_internal:term_type/1
 bif erts_internal:map_hashmap_children/1
 
 bif erts_internal:time_unit/0
@@ -639,6 +638,17 @@ bif ets:update_counter/4
 bif erts_debug:map_info/1
 
 #
+# New in 19.0
+#
+
+bif erts_internal:copy_literals/2
+bif erts_internal:purge_module/1
+bif binary:split/2
+bif binary:split/3
+bif erts_debug:size_shared/1
+bif erts_debug:copy_shared/1
+
+#
 # Obsolete
 #
 
diff --git a/erts/emulator/beam/big.c b/erts/emulator/beam/big.c
index 044bf6a34e..02d37e24df 100644
--- a/erts/emulator/beam/big.c
+++ b/erts/emulator/beam/big.c
@@ -1487,20 +1487,8 @@ Eterm uint_to_big(Uint x, Eterm *y)
 
 Eterm uword_to_big(UWord x, Eterm *y)
 {
-#if HALFWORD_HEAP
-    Uint upper = x >> 32;
-    Uint lower = x & 0xFFFFFFFFUL;
-    if (upper == 0) {
-	*y = make_pos_bignum_header(1);
-    } else {
-	*y = make_pos_bignum_header(2);
-	BIG_DIGIT(y, 1) = upper;
-    }
-    BIG_DIGIT(y, 0) = lower;
-#else
     *y = make_pos_bignum_header(1);
     BIG_DIGIT(y, 0) = x;
-#endif
     return make_big(y);
 }
 
@@ -1525,7 +1513,7 @@ Eterm small_to_big(Sint x, Eterm *y)
 Eterm erts_uint64_to_big(Uint64 x, Eterm **hpp)
 {
     Eterm *hp = *hpp;
-#if defined(ARCH_32) || HALFWORD_HEAP
+#if defined(ARCH_32)
     if (x >= (((Uint64) 1) << 32)) {
 	*hp = make_pos_bignum_header(2);
 	BIG_DIGIT(hp, 0) = (Uint) (x & ((Uint) 0xffffffff));
@@ -1555,7 +1543,7 @@ Eterm erts_sint64_to_big(Sint64 x, Eterm **hpp)
 	neg = 1;
 	ux = -(Uint64)x;
     }
-#if defined(ARCH_32) || HALFWORD_HEAP
+#if defined(ARCH_32)
     if (ux >= (((Uint64) 1) << 32)) {
 	if (neg)
 	    *hp = make_neg_bignum_header(2);
@@ -1588,7 +1576,7 @@ erts_uint64_array_to_big(Uint **hpp, int neg, int len, Uint64 *array)
 
     pot_digits = digits = 0;
     for (i = 0; i < len; i++) {
-#if defined(ARCH_32) || HALFWORD_HEAP
+#if defined(ARCH_32)
 	Uint low_val = array[i] & ((Uint) 0xffffffff);
 	Uint high_val = (array[i] >> 32) & ((Uint) 0xffffffff);
 	BIG_DIGIT(headerp, pot_digits) = low_val;
@@ -1651,8 +1639,6 @@ big_to_double(Wterm x, double* resp)
 /*
  * Logic has been copied from erl_bif_guard.c and slightly
  * modified to use a static instead of dynamic heap
- *
- * HALFWORD: Return relative term with 'heap' as base.
  */
 Eterm
 double_to_big(double x, Eterm *heap, Uint hsz)
@@ -1683,7 +1669,7 @@ double_to_big(double x, Eterm *heap, Uint hsz)
     sz = BIG_NEED_SIZE(ds); /* number of words including arity */
 
     hp = heap;
-    res = make_big_rel(hp, heap);
+    res = make_big(hp);
     xp = (ErtsDigit*) (hp + 1);
 
     ASSERT(ds < hsz);
@@ -2618,6 +2604,9 @@ Eterm erts_chars_to_integer(Process *BIF_P, char *bytes,
 	size--;
     }
 
+    if (size == 0)
+	goto bytebuf_to_integer_1_error;
+
     if (size < SMALL_DIGITS && base <= 10) {
 	/* *
 	 * Take shortcut if we know that all chars are '0' < b < '9' and
diff --git a/erts/emulator/beam/big.h b/erts/emulator/beam/big.h
index 4aa9724ae3..94f9bce10e 100644
--- a/erts/emulator/beam/big.h
+++ b/erts/emulator/beam/big.h
@@ -35,7 +35,7 @@
 
 typedef Uint     ErtsDigit;
 
-#if ((SIZEOF_VOID_P == 4) || HALFWORD_HEAP) && defined(SIZEOF_LONG_LONG) && (SIZEOF_LONG_LONG == 8)
+#if (SIZEOF_VOID_P == 4) && defined(SIZEOF_LONG_LONG) && (SIZEOF_LONG_LONG == 8)
 /* Assume 32-bit machine with long long support */
 typedef Uint64   ErtsDoubleDigit;
 typedef Uint16   ErtsHalfDigit;
@@ -90,13 +90,9 @@ typedef Uint  dsize_t;	 /* Vector size type */
 
 #define BIG_UINT_HEAP_SIZE (1 + 1)	/* always, since sizeof(Uint) <= sizeof(Eterm) */
 
-#if HALFWORD_HEAP
-#define BIG_UWORD_HEAP_SIZE(UW) (((UW) >> (sizeof(Uint) * 8)) ? 3 : 2)
-#else
 #define BIG_UWORD_HEAP_SIZE(UW) BIG_UINT_HEAP_SIZE
-#endif
 
-#if defined(ARCH_32) || HALFWORD_HEAP
+#if defined(ARCH_32)
 
 #define ERTS_UINT64_BIG_HEAP_SIZE__(X) \
   ((X) >= (((Uint64) 1) << 32) ? (1 + 2) : (1 + 1))
@@ -178,4 +174,3 @@ Eterm erts_sint64_to_big(Sint64, Eterm **);
 Eterm erts_chars_to_integer(Process *, char*, Uint, const int);
 
 #endif
-
diff --git a/erts/emulator/beam/binary.c b/erts/emulator/beam/binary.c
index e670fbf31c..d3e481c7f9 100644
--- a/erts/emulator/beam/binary.c
+++ b/erts/emulator/beam/binary.c
@@ -34,6 +34,9 @@
 #include "erl_binary.h"
 #include "erl_bits.h"
 
+#define L2B_B2L_MIN_EXEC_REDS (CONTEXT_REDS/4)
+#define L2B_B2L_RESCHED_REDS (CONTEXT_REDS/40)
+
 static Export binary_to_list_continue_export;
 static Export list_to_binary_continue_export;
 
@@ -415,10 +418,10 @@ binary_to_list_chunk(Process *c_p,
 }
 
 static ERTS_INLINE BIF_RETTYPE
-binary_to_list(Process *c_p, Eterm *hp, Eterm tail, byte *bytes, Uint size, Uint bitoffs)
+binary_to_list(Process *c_p, Eterm *hp, Eterm tail, byte *bytes,
+	       Uint size, Uint bitoffs, int reds_left, int one_chunk)
 {
-    int reds_left = ERTS_BIF_REDS_LEFT(c_p);
-    if (size < reds_left*ERTS_B2L_BYTES_PER_REDUCTION) {
+    if (one_chunk) {
 	Eterm res;
 	BIF_RETTYPE ret;
 	int bump_reds = (size - 1)/ERTS_B2L_BYTES_PER_REDUCTION + 1;
@@ -472,11 +475,29 @@ BIF_RETTYPE binary_to_list_1(BIF_ALIST_1)
     Uint size;
     Uint bitsize;
     Uint bitoffs;
+    int reds_left;
+    int one_chunk;
 
     if (is_not_binary(BIF_ARG_1)) {
 	goto error;
     }
+
     size = binary_size(BIF_ARG_1);
+    reds_left = ERTS_BIF_REDS_LEFT(BIF_P);
+    one_chunk = size < reds_left*ERTS_B2L_BYTES_PER_REDUCTION;
+    if (!one_chunk) {
+	if (size < L2B_B2L_MIN_EXEC_REDS*ERTS_B2L_BYTES_PER_REDUCTION) {
+	    if (reds_left <= L2B_B2L_RESCHED_REDS) {
+		/* Yield and do it with full context reds... */
+		ERTS_BIF_YIELD1(bif_export[BIF_binary_to_list_1],
+				BIF_P, BIF_ARG_1);
+	    }
+	    /* Allow a bit more reductions... */
+	    one_chunk = 1;
+	    reds_left = L2B_B2L_MIN_EXEC_REDS;
+	}
+    }
+
     ERTS_GET_REAL_BIN(BIF_ARG_1, real_bin, offset, bitoffs, bitsize);
     if (bitsize != 0) {
 	goto error;
@@ -486,7 +507,8 @@ BIF_RETTYPE binary_to_list_1(BIF_ALIST_1)
     } else {
 	Eterm* hp = HAlloc(BIF_P, 2 * size);
 	byte* bytes = binary_bytes(real_bin)+offset;
-	return binary_to_list(BIF_P, hp, NIL, bytes, size, bitoffs);
+	return binary_to_list(BIF_P, hp, NIL, bytes, size,
+			      bitoffs, reds_left, one_chunk);
     }
 
     error:
@@ -505,6 +527,8 @@ BIF_RETTYPE binary_to_list_3(BIF_ALIST_3)
     Uint start;
     Uint stop;
     Eterm* hp;
+    int reds_left;
+    int one_chunk;
 
     if (is_not_binary(BIF_ARG_1)) {
 	goto error;
@@ -513,6 +537,21 @@ BIF_RETTYPE binary_to_list_3(BIF_ALIST_3)
 	goto error;
     }
     size = binary_size(BIF_ARG_1);
+    reds_left = ERTS_BIF_REDS_LEFT(BIF_P);
+    one_chunk = size < reds_left*ERTS_B2L_BYTES_PER_REDUCTION;
+    if (!one_chunk) {
+	if (size < L2B_B2L_MIN_EXEC_REDS*ERTS_B2L_BYTES_PER_REDUCTION) {
+	    if (reds_left <= L2B_B2L_RESCHED_REDS) {
+		/* Yield and do it with full context reds... */
+		ERTS_BIF_YIELD3(bif_export[BIF_binary_to_list_3],
+				BIF_P, BIF_ARG_1, BIF_ARG_2, BIF_ARG_3);
+	    }
+	    /* Allow a bit more reductions... */
+	    one_chunk = 1;
+	    reds_left = L2B_B2L_MIN_EXEC_REDS;
+	}
+    }
+
     ERTS_GET_BINARY_BYTES(BIF_ARG_1, bytes, bitoffs, bitsize);
     if (start < 1 || start > size || stop < 1 ||
 	stop > size || stop < start ) {
@@ -520,7 +559,8 @@ BIF_RETTYPE binary_to_list_3(BIF_ALIST_3)
     }
     i = stop-start+1;
     hp = HAlloc(BIF_P, 2*i);
-    return binary_to_list(BIF_P, hp, NIL, bytes+start-1, i, bitoffs);
+    return binary_to_list(BIF_P, hp, NIL, bytes+start-1, i,
+			  bitoffs, reds_left, one_chunk);
     error:
 	BIF_ERROR(BIF_P, BADARG);
 }
@@ -537,11 +577,27 @@ BIF_RETTYPE bitstring_to_list_1(BIF_ALIST_1)
     byte* bytes;
     Eterm previous = NIL;
     Eterm* hp;
+    int reds_left;
+    int one_chunk;
 
     if (is_not_binary(BIF_ARG_1)) {
 	BIF_ERROR(BIF_P, BADARG);
     }
     size = binary_size(BIF_ARG_1);
+    reds_left = ERTS_BIF_REDS_LEFT(BIF_P);
+    one_chunk = size < reds_left*ERTS_B2L_BYTES_PER_REDUCTION;
+    if (!one_chunk) {
+	if (size < L2B_B2L_MIN_EXEC_REDS*ERTS_B2L_BYTES_PER_REDUCTION) {
+	    if (reds_left <= L2B_B2L_RESCHED_REDS) {
+		/* Yield and do it with full context reds... */
+		ERTS_BIF_YIELD1(bif_export[BIF_bitstring_to_list_1],
+				BIF_P, BIF_ARG_1);
+	    }
+	    /* Allow a bit more reductions... */
+	    one_chunk = 1;
+	    reds_left = L2B_B2L_MIN_EXEC_REDS;
+	}
+    }
     ERTS_GET_REAL_BIN(BIF_ARG_1, real_bin, offset, bitoffs, bitsize);
     bytes = binary_bytes(real_bin)+offset;
     if (bitsize == 0) {
@@ -566,7 +622,8 @@ BIF_RETTYPE bitstring_to_list_1(BIF_ALIST_1)
 	hp += 2;
     }
 
-    return binary_to_list(BIF_P, hp, previous, bytes, size, bitoffs);
+    return binary_to_list(BIF_P, hp, previous, bytes, size,
+			  bitoffs, reds_left, one_chunk);
 }
 
 
@@ -795,8 +852,19 @@ static BIF_RETTYPE list_to_binary_continue(BIF_ALIST_1)
 
 BIF_RETTYPE erts_list_to_binary_bif(Process *c_p, Eterm arg, Export *bif)
 {
+    int orig_reds_left = ERTS_BIF_REDS_LEFT(c_p);
     BIF_RETTYPE ret;
 
+    if (orig_reds_left < L2B_B2L_MIN_EXEC_REDS) {
+	if (orig_reds_left <= L2B_B2L_RESCHED_REDS) {
+	    /* Yield and do it with full context reds... */
+	    ERTS_BIF_PREP_YIELD1(ret, bif, c_p, arg);
+	    return ret;
+	}
+	/* Allow a bit more reductions... */
+	orig_reds_left = L2B_B2L_MIN_EXEC_REDS;
+    }
+
     if (is_nil(arg))
 	ERTS_BIF_PREP_RET(ret, new_binary(c_p, (byte *) "", 0));
     else if (is_not_list(arg))
@@ -818,7 +886,6 @@ BIF_RETTYPE erts_list_to_binary_bif(Process *c_p, Eterm arg, Export *bif)
 						       bif,
 						       erts_iolist_size_yielding,
 						       erts_iolist_to_buf_yielding);
-	    int orig_reds_left = ERTS_BIF_REDS_LEFT(c_p);
 
 	    /*
 	     * First try to do it all at once without having to use
diff --git a/erts/emulator/beam/break.c b/erts/emulator/beam/break.c
index 64c8bc5e58..2c8ecf04be 100644
--- a/erts/emulator/beam/break.c
+++ b/erts/emulator/beam/break.c
@@ -252,7 +252,7 @@ print_process_info(int to, void *to_arg, Process *p)
 
     /* display the message queue only if there is anything in it */
     if (!ERTS_IS_CRASH_DUMPING && p->msg.first != NULL && !garbing) {
-	ErlMessage* mp;
+	ErtsMessage* mp;
 	erts_print(to, to_arg, "Message queue: [");
 	for (mp = p->msg.first; mp; mp = mp->next)
 	    erts_print(to, to_arg, mp->next ? "%T," : "%T", ERL_MESSAGE_TERM(mp));
@@ -323,7 +323,7 @@ print_process_info(int to, void *to_arg, Process *p)
     erts_print(to, to_arg, "Heap unused: %bpu\n", (p->hend - p->htop));
     erts_print(to, to_arg, "OldHeap unused: %bpu\n",
 	       (OLD_HEAP(p) == NULL) ? 0 : (OLD_HEND(p) - OLD_HTOP(p)) );
-    erts_print(to, to_arg, "Memory: %beu\n", erts_process_memory(p));
+    erts_print(to, to_arg, "Memory: %beu\n", erts_process_memory(p, !0));
 
     if (garbing) {
 	print_garb_info(to, to_arg, p);
@@ -381,7 +381,7 @@ loaded(int to, void *to_arg)
     int i;
     int old = 0;
     int cur = 0;
-    BeamInstr* code;
+    BeamCodeHeader* code;
     Module* modp;
     ErtsCodeIndex code_ix;
 
@@ -439,30 +439,30 @@ loaded(int to, void *to_arg)
 		erts_print(to, to_arg, "\n");
 		erts_print(to, to_arg, "Current size: %d\n",
 			   modp->curr.code_length);
-		code = modp->curr.code;
-		if (code != NULL && code[MI_ATTR_PTR]) {
+		code = modp->curr.code_hdr;
+		if (code != NULL && code->attr_ptr) {
 		    erts_print(to, to_arg, "Current attributes: ");
-		    dump_attributes(to, to_arg, (byte *) code[MI_ATTR_PTR],
-				    code[MI_ATTR_SIZE]);
+		    dump_attributes(to, to_arg, code->attr_ptr,
+				    code->attr_size);
 		}
-		if (code != NULL && code[MI_COMPILE_PTR]) {
+		if (code != NULL && code->compile_ptr) {
 		    erts_print(to, to_arg, "Current compilation info: ");
-		    dump_attributes(to, to_arg, (byte *) code[MI_COMPILE_PTR],
-				    code[MI_COMPILE_SIZE]);
+		    dump_attributes(to, to_arg, code->compile_ptr,
+				    code->compile_size);
 		}
 
 		if (modp->old.code_length != 0) {
 		    erts_print(to, to_arg, "Old size: %d\n", modp->old.code_length);
-		    code = modp->old.code;
-		    if (code[MI_ATTR_PTR]) {
+		    code = modp->old.code_hdr;
+		    if (code->attr_ptr) {
 			erts_print(to, to_arg, "Old attributes: ");
-			dump_attributes(to, to_arg, (byte *) code[MI_ATTR_PTR],
-					code[MI_ATTR_SIZE]);
+			dump_attributes(to, to_arg, code->attr_ptr,
+					code->attr_size);
 		    }
-		    if (code[MI_COMPILE_PTR]) {
+		    if (code->compile_ptr) {
 			erts_print(to, to_arg, "Old compilation info: ");
-			dump_attributes(to, to_arg, (byte *) code[MI_COMPILE_PTR],
-					code[MI_COMPILE_SIZE]);
+			dump_attributes(to, to_arg, code->compile_ptr,
+					code->compile_size);
 		    }
 		}
 	    }
@@ -536,7 +536,9 @@ do_break(void)
 	    erts_printf("Erlang (%s) emulator version "
 		       ERLANG_VERSION "\n",
 		       EMULATOR);
+#if ERTS_SAVED_COMPILE_TIME
 	    erts_printf("Compiled on " ERLANG_COMPILE_DATE "\n");
+#endif
 	    return;
 	case 'd':
 	    distribution_info(ERTS_PRINT_STDOUT, NULL);
@@ -774,7 +776,9 @@ erl_crash_dump_v(char *file, int line, char* fmt, va_list args)
     }
     erts_fdprintf(fd, "System version: ");
     erts_print_system_version(fd, NULL, NULL);
+#if ERTS_SAVED_COMPILE_TIME
     erts_fdprintf(fd, "%s\n", "Compiled: " ERLANG_COMPILE_DATE);
+#endif
 
     erts_fdprintf(fd, "Taints: ");
     erts_print_nif_taints(fd, NULL);
diff --git a/erts/emulator/beam/copy.c b/erts/emulator/beam/copy.c
index 8849dadd00..67a96f6442 100644
--- a/erts/emulator/beam/copy.c
+++ b/erts/emulator/beam/copy.c
@@ -35,56 +35,59 @@
 #include "erl_bits.h"
 #include "dtrace-wrapper.h"
 
-static void move_one_frag(Eterm** hpp, ErlHeapFragment*, ErlOffHeap*);
+static void move_one_frag(Eterm** hpp, ErlHeapFragment*, ErlOffHeap*, int);
 
 /*
  *  Copy object "obj" to process p.
  */
-Eterm
-copy_object(Eterm obj, Process* to)
-{
-    Uint size = size_object(obj);
-    Eterm* hp = HAlloc(to, size);
-    Eterm res;
+Eterm copy_object_x(Eterm obj, Process* to, Uint extra) {
+    if (!is_immed(obj)) {
+        Uint size = size_object(obj);
+        Eterm* hp = HAllocX(to, size, extra);
+        Eterm res;
 
 #ifdef USE_VM_PROBES
-    if (DTRACE_ENABLED(copy_object)) {
-        DTRACE_CHARBUF(proc_name, 64);
+        if (DTRACE_ENABLED(copy_object)) {
+            DTRACE_CHARBUF(proc_name, 64);
 
-        erts_snprintf(proc_name, sizeof(DTRACE_CHARBUF_NAME(proc_name)),
-                      "%T", to->common.id);
-        DTRACE2(copy_object, proc_name, size);
-    }
+            erts_snprintf(proc_name, sizeof(DTRACE_CHARBUF_NAME(proc_name)),
+                    "%T", to->common.id);
+            DTRACE2(copy_object, proc_name, size);
+        }
 #endif
-    res = copy_struct(obj, size, &hp, &to->off_heap);
+        res = copy_struct(obj, size, &hp, &to->off_heap);
 #ifdef DEBUG
-    if (eq(obj, res) == 0) {
-	erl_exit(ERTS_ABORT_EXIT, "copy not equal to source\n");
-    }
+        if (eq(obj, res) == 0) {
+            erl_exit(ERTS_ABORT_EXIT, "copy not equal to source\n");
+        }
 #endif
-    return res;
+        return res;
+    }
+    return obj;
 }
 
 /*
  * Return the "flat" size of the object.
  */
 
-#if HALFWORD_HEAP
-Uint size_object_rel(Eterm obj, Eterm* base)
-#else
 Uint size_object(Eterm obj)
-#endif
 {
     Uint sum = 0;
     Eterm* ptr;
     int arity;
+#ifdef DEBUG
+    Eterm mypid = erts_get_current_pid();
+#endif
 
     DECLARE_ESTACK(s);
+
+    VERBOSE(DEBUG_SHCOPY, ("[pid=%T] size_object %p\n", mypid, obj));
+
     for (;;) {
 	switch (primary_tag(obj)) {
 	case TAG_PRIMARY_LIST:
 	    sum += 2;
-	    ptr = list_val_rel(obj,base);
+	    ptr = list_val(obj);
 	    obj = *ptr++;
 	    if (!IS_CONST(obj)) {
 		ESTACK_PUSH(s, obj);
@@ -93,11 +96,11 @@ Uint size_object(Eterm obj)
 	    break;
 	case TAG_PRIMARY_BOXED:
 	    {
-		Eterm hdr = *boxed_val_rel(obj,base);
+		Eterm hdr = *boxed_val(obj);
 		ASSERT(is_header(hdr));
 		switch (hdr & _TAG_HEADER_MASK) {
 		case ARITYVAL_SUBTAG:
-		    ptr = tuple_val_rel(obj,base);
+		    ptr = tuple_val(obj);
 		    arity = header_arity(hdr);
 		    sum += arity + 1;
 		    if (arity == 0) { /* Empty tuple -- unusual. */
@@ -113,7 +116,7 @@ Uint size_object(Eterm obj)
 		    break;
 		case FUN_SUBTAG:
 		    {
-			Eterm* bptr = fun_val_rel(obj,base);
+			Eterm* bptr = fun_val(obj);
 			ErlFunThing* funp = (ErlFunThing *) bptr;
 			unsigned eterms = 1 /* creator */ + funp->num_free;
 			unsigned sz = thing_arityval(hdr);
@@ -134,7 +137,7 @@ Uint size_object(Eterm obj)
                             {
                                 Uint n;
                                 flatmap_t *mp;
-                                mp  = (flatmap_t*)flatmap_val_rel(obj,base);
+                                mp  = (flatmap_t*)flatmap_val(obj);
                                 ptr = (Eterm *)mp;
                                 n   = flatmap_get_size(mp) + 1;
                                 sum += n + 2;
@@ -153,7 +156,7 @@ Uint size_object(Eterm obj)
 			    {
 				Eterm *head;
 				Uint sz;
-				head  = hashmap_val_rel(obj, base);
+				head  = hashmap_val(obj);
 				sz    = hashmap_bitcount(MAP_HEADER_VAL(hdr));
 				sum  += 1 + sz + header_arity(hdr);
 				head += 1 + header_arity(hdr);
@@ -182,7 +185,7 @@ Uint size_object(Eterm obj)
 			Uint bitoffs;
 			Uint extra_bytes;
 			Eterm hdr;
-			ERTS_GET_REAL_BIN_REL(obj, real_bin, offset, bitoffs, bitsize, base);
+			ERTS_GET_REAL_BIN(obj, real_bin, offset, bitoffs, bitsize);
 			if ((bitsize + bitoffs) > 8) {
 			    sum += ERL_SUB_BIN_SIZE;
 			    extra_bytes = 2;
@@ -192,11 +195,11 @@ Uint size_object(Eterm obj)
 			} else {
 			    extra_bytes = 0;
 			}
-			hdr = *binary_val_rel(real_bin,base);
+			hdr = *binary_val(real_bin);
 			if (thing_subtag(hdr) == REFC_BINARY_SUBTAG) {
 			    sum += PROC_BIN_SIZE;
 			} else {
-			    sum += heap_bin_size(binary_size_rel(obj,base)+extra_bytes);
+			    sum += heap_bin_size(binary_size(obj)+extra_bytes);
 			}
 			goto pop_next;
 		    }
@@ -214,6 +217,7 @@ Uint size_object(Eterm obj)
 	pop_next:
 	    if (ESTACK_ISEMPTY(s)) {
 		DESTROY_ESTACK(s);
+		VERBOSE(DEBUG_SHCOPY, ("[pid=%T] size was: %u\n", mypid, sum));
 		return sum;
 	    }
 	    obj = ESTACK_POP(s);
@@ -225,14 +229,377 @@ Uint size_object(Eterm obj)
 }
 
 /*
+ *  Machinery for sharing preserving information
+ *  Using a WSTACK but not very transparently; consider refactoring
+ */
+
+#define DECLARE_BITSTORE(s)						\
+    DECLARE_WSTACK(s);							\
+    int WSTK_CONCAT(s,_bitoffs) = 0;					\
+    int WSTK_CONCAT(s,_offset) = 0;					\
+    UWord WSTK_CONCAT(s,_buffer) = 0
+
+#define DESTROY_BITSTORE(s) DESTROY_WSTACK(s)
+#define BITSTORE_PUT(s,i)						\
+do {									\
+    WSTK_CONCAT(s,_buffer) |= i << WSTK_CONCAT(s,_bitoffs);		\
+    WSTK_CONCAT(s,_bitoffs) += 2;					\
+    if (WSTK_CONCAT(s,_bitoffs) >= 8*sizeof(UWord)) {			\
+	WSTACK_PUSH(s, WSTK_CONCAT(s,_buffer));				\
+	WSTK_CONCAT(s,_bitoffs) = 0;					\
+	WSTK_CONCAT(s,_buffer) = 0;					\
+    }									\
+} while(0)
+#define BITSTORE_CLOSE(s)						\
+do {									\
+    if (WSTK_CONCAT(s,_bitoffs) > 0) {					\
+	WSTACK_PUSH(s, WSTK_CONCAT(s,_buffer));				\
+	WSTK_CONCAT(s,_bitoffs) = 0;					\
+    }									\
+} while(0)
+
+#define BITSTORE_FETCH(s,dst)                                           \
+do {                                                                    \
+    UWord result;                                                       \
+    if (WSTK_CONCAT(s,_bitoffs) <= 0) {                                 \
+        WSTK_CONCAT(s,_buffer) = s.wstart[WSTK_CONCAT(s,_offset)];      \
+        WSTK_CONCAT(s,_offset)++;                                       \
+        WSTK_CONCAT(s,_bitoffs) = 8*sizeof(UWord);                      \
+    }                                                                   \
+    WSTK_CONCAT(s,_bitoffs) -= 2;                                       \
+    result = WSTK_CONCAT(s,_buffer) & 3;                                \
+    WSTK_CONCAT(s,_buffer) >>= 2;                                       \
+    (dst) = result;                                                     \
+} while(0)
+
+#define BOXED_VISITED_MASK	 ((Eterm) 3)
+#define BOXED_VISITED		 ((Eterm) 1)
+#define BOXED_SHARED_UNPROCESSED ((Eterm) 2)
+#define BOXED_SHARED_PROCESSED	 ((Eterm) 3)
+
+#define COUNT_OFF_HEAP (0)
+
+#define IN_LITERAL_PURGE_AREA(info, ptr)                 \
+    ((info)->range_ptr && (                              \
+        (info)->range_ptr <= (ptr) &&                    \
+        (ptr) < ((info)->range_ptr + (info)->range_sz)))
+/*
+ *  Return the real size of an object and find sharing information
+ *  This currently returns the same as erts_debug:size/1.
+ *  It is argued whether the size of subterms in constant pools
+ *  should be counted or not.
+ */
+
+Uint size_shared(Eterm obj)
+{
+    Eterm saved_obj = obj;
+    Uint sum = 0;
+    Eterm* ptr;
+
+    DECLARE_EQUEUE(s);
+    DECLARE_BITSTORE(b);
+
+    for (;;) {
+	switch (primary_tag(obj)) {
+	case TAG_PRIMARY_LIST: {
+	    Eterm head, tail;
+	    ptr = list_val(obj);
+	    /* we're not counting anything that's outside our heap */
+	    if (!COUNT_OFF_HEAP && erts_is_literal(obj,ptr)) {
+		goto pop_next;
+	    }
+	    head = CAR(ptr);
+	    tail = CDR(ptr);
+	    /* if it's visited, don't count it */
+	    if (primary_tag(tail) == TAG_PRIMARY_HEADER ||
+		primary_tag(head) == TAG_PRIMARY_HEADER) {
+		goto pop_next;
+	    }
+	    /* else make it visited now */
+	    switch (primary_tag(tail)) {
+	    case TAG_PRIMARY_LIST:
+		ptr[1] = (tail - TAG_PRIMARY_LIST) | TAG_PRIMARY_HEADER;
+		break;
+	    case TAG_PRIMARY_IMMED1:
+		CAR(ptr) = (head - primary_tag(head)) | TAG_PRIMARY_HEADER;
+		CDR(ptr) = (tail - TAG_PRIMARY_IMMED1) | primary_tag(head);
+		break;
+	    case TAG_PRIMARY_BOXED:
+		BITSTORE_PUT(b, primary_tag(head));
+		CAR(ptr) = (head - primary_tag(head)) | TAG_PRIMARY_HEADER;
+		CDR(ptr) = (tail - TAG_PRIMARY_BOXED) | TAG_PRIMARY_HEADER;
+		break;
+	    }
+	    /* and count it */
+	    sum += 2;
+	    if (!IS_CONST(head)) {
+		EQUEUE_PUT(s, head);
+	    }
+	    obj = tail;
+	    break;
+	}
+	case TAG_PRIMARY_BOXED: {
+	    Eterm hdr;
+	    ptr = boxed_val(obj);
+	    /* we're not counting anything that's outside our heap */
+	    if (!COUNT_OFF_HEAP && erts_is_literal(obj,ptr)) {
+		goto pop_next;
+	    }
+	    hdr = *ptr;
+	    /* if it's visited, don't count it */
+	    if (primary_tag(hdr) != TAG_PRIMARY_HEADER) {
+		goto pop_next;
+	    }
+	    /* else make it visited now */
+	    *ptr = (hdr - primary_tag(hdr)) + BOXED_VISITED;
+	    /* and count it */
+	    ASSERT(is_header(hdr));
+	    switch (hdr & _TAG_HEADER_MASK) {
+	    case ARITYVAL_SUBTAG: {
+		int arity = header_arity(hdr);
+		sum += arity + 1;
+		if (arity == 0) { /* Empty tuple -- unusual. */
+		    goto pop_next;
+		}
+		while (arity-- > 0) {
+		    obj = *++ptr;
+		    if (!IS_CONST(obj)) {
+			EQUEUE_PUT(s, obj);
+		    }
+		}
+		goto pop_next;
+	    }
+	    case FUN_SUBTAG: {
+		ErlFunThing* funp = (ErlFunThing *) ptr;
+		unsigned eterms = 1 /* creator */ + funp->num_free;
+		unsigned sz = thing_arityval(hdr);
+		sum += 1 /* header */ + sz + eterms;
+		ptr += 1 /* header */ + sz;
+		while (eterms-- > 0) {
+		    obj = *ptr++;
+		    if (!IS_CONST(obj)) {
+			EQUEUE_PUT(s, obj);
+		    }
+		}
+		goto pop_next;
+	    }
+	    case SUB_BINARY_SUBTAG: {
+		ErlSubBin* sb = (ErlSubBin *) ptr;
+		Uint extra_bytes;
+		Eterm hdr;
+		ASSERT((sb->thing_word & ~BOXED_VISITED_MASK) == HEADER_SUB_BIN);
+		if (sb->bitsize + sb->bitoffs > 8) {
+		    sum += ERL_SUB_BIN_SIZE;
+		    extra_bytes = 2;
+		} else if (sb->bitsize + sb->bitoffs > 0) {
+		    sum += ERL_SUB_BIN_SIZE;
+		    extra_bytes = 1;
+		} else {
+		    extra_bytes = 0;
+		}
+		ptr = binary_val(sb->orig);
+		hdr = (*ptr) & ~BOXED_VISITED_MASK;
+		if (thing_subtag(hdr) == REFC_BINARY_SUBTAG) {
+		    sum += PROC_BIN_SIZE;
+		} else {
+		    ASSERT(thing_subtag(hdr) == HEAP_BINARY_SUBTAG);
+		    sum += heap_bin_size(binary_size(obj) + extra_bytes);
+		}
+		goto pop_next;
+	    }
+            case MAP_SUBTAG:
+                switch (MAP_HEADER_TYPE(hdr)) {
+                    case MAP_HEADER_TAG_FLATMAP_HEAD : {
+                        flatmap_t *mp  = (flatmap_t*)flatmap_val(obj);
+                        Uint n = flatmap_get_size(mp) + 1;
+                        ptr  = (Eterm *)mp;
+                        sum += n + 2;
+                        ptr += 2; /* hdr + size words */
+                        while (n--) {
+                            obj = *ptr++;
+                            if (!IS_CONST(obj)) {
+                                EQUEUE_PUT(s, obj);
+                            }
+                        }
+                        goto pop_next;
+                    }
+                    case MAP_HEADER_TAG_HAMT_HEAD_BITMAP :
+                    case MAP_HEADER_TAG_HAMT_HEAD_ARRAY :
+                    case MAP_HEADER_TAG_HAMT_NODE_BITMAP : {
+                        Uint n = hashmap_bitcount(MAP_HEADER_VAL(hdr));
+                        sum += 1 + n + header_arity(hdr);
+                        ptr += 1 + header_arity(hdr);
+                        while (n--) {
+                            obj = *ptr++;
+                            if (!IS_CONST(obj)) {
+                                EQUEUE_PUT(s, obj);
+                            }
+                        }
+                        goto pop_next;
+                    }
+                    default:
+                        erl_exit(ERTS_ABORT_EXIT, "size_shared: bad hashmap type %d\n", MAP_HEADER_TYPE(hdr));
+                }
+	    case BIN_MATCHSTATE_SUBTAG:
+		erl_exit(ERTS_ABORT_EXIT,
+			 "size_shared: matchstate term not allowed");
+	    default:
+		sum += thing_arityval(hdr) + 1;
+		goto pop_next;
+	    }
+	    break;
+	}
+	case TAG_PRIMARY_IMMED1:
+	pop_next:
+	    if (EQUEUE_ISEMPTY(s)) {
+		goto cleanup;
+	    }
+	    obj = EQUEUE_GET(s);
+	    break;
+	default:
+	    erl_exit(ERTS_ABORT_EXIT, "size_shared: bad tag for %#x\n", obj);
+	}
+    }
+
+cleanup:
+    obj = saved_obj;
+    BITSTORE_CLOSE(b);
+    for (;;) {
+	switch (primary_tag(obj)) {
+	case TAG_PRIMARY_LIST: {
+	    Eterm head, tail;
+	    ptr = list_val(obj);
+	    if (!COUNT_OFF_HEAP && erts_is_literal(obj,ptr)) {
+		goto cleanup_next;
+	    }
+	    head = CAR(ptr);
+	    tail = CDR(ptr);
+	    /* if not already clean, clean it up */
+	    if (primary_tag(tail) == TAG_PRIMARY_HEADER) {
+		if (primary_tag(head) == TAG_PRIMARY_HEADER) {
+		    Eterm saved;
+                    BITSTORE_FETCH(b, saved);
+		    CAR(ptr) = head = (head - TAG_PRIMARY_HEADER) | saved;
+		    CDR(ptr) = tail = (tail - TAG_PRIMARY_HEADER) | TAG_PRIMARY_BOXED;
+		} else {
+		    CDR(ptr) = tail = (tail - TAG_PRIMARY_HEADER) | TAG_PRIMARY_LIST;
+		}
+	    } else if (primary_tag(head) == TAG_PRIMARY_HEADER) {
+		CAR(ptr) = head = (head - TAG_PRIMARY_HEADER) | primary_tag(tail);
+		CDR(ptr) = tail = (tail - primary_tag(tail)) | TAG_PRIMARY_IMMED1;
+	    } else {
+		goto cleanup_next;
+	    }
+	    /* and its children too */
+	    if (!IS_CONST(head)) {
+		EQUEUE_PUT_UNCHECKED(s, head);
+	    }
+	    obj = tail;
+	    break;
+	}
+	case TAG_PRIMARY_BOXED: {
+	    Eterm hdr;
+	    ptr = boxed_val(obj);
+	    if (!COUNT_OFF_HEAP && erts_is_literal(obj,ptr)) {
+		goto cleanup_next;
+	    }
+	    hdr = *ptr;
+	    /* if not already clean, clean it up */
+	    if (primary_tag(hdr) == TAG_PRIMARY_HEADER) {
+		goto cleanup_next;
+	    }
+	    else {
+		ASSERT(primary_tag(hdr) == BOXED_VISITED);
+		*ptr = hdr = (hdr - BOXED_VISITED) + TAG_PRIMARY_HEADER;
+	    }
+	    /* and its children too */
+	    switch (hdr & _TAG_HEADER_MASK) {
+	    case ARITYVAL_SUBTAG: {
+		int arity = header_arity(hdr);
+		if (arity == 0) { /* Empty tuple -- unusual. */
+		    goto cleanup_next;
+		}
+		while (arity-- > 0) {
+		    obj = *++ptr;
+		    if (!IS_CONST(obj)) {
+			EQUEUE_PUT_UNCHECKED(s, obj);
+		    }
+		}
+		goto cleanup_next;
+	    }
+	    case FUN_SUBTAG: {
+		ErlFunThing* funp = (ErlFunThing *) ptr;
+		unsigned eterms = 1 /* creator */ + funp->num_free;
+		unsigned sz = thing_arityval(hdr);
+		ptr += 1 /* header */ + sz;
+		while (eterms-- > 0) {
+		    obj = *ptr++;
+		    if (!IS_CONST(obj)) {
+			EQUEUE_PUT_UNCHECKED(s, obj);
+		    }
+		}
+		goto cleanup_next;
+	    }
+            case MAP_SUBTAG:
+                switch (MAP_HEADER_TYPE(hdr)) {
+                    case MAP_HEADER_TAG_FLATMAP_HEAD : {
+                        flatmap_t *mp = (flatmap_t *) ptr;
+                        Uint n = flatmap_get_size(mp) + 1;
+                        ptr += 2; /* hdr + size words */
+                        while (n--) {
+                            obj = *ptr++;
+                            if (!IS_CONST(obj)) {
+                                EQUEUE_PUT_UNCHECKED(s, obj);
+                            }
+                        }
+                        goto cleanup_next;
+                    }
+                    case MAP_HEADER_TAG_HAMT_HEAD_BITMAP :
+                    case MAP_HEADER_TAG_HAMT_HEAD_ARRAY :
+                    case MAP_HEADER_TAG_HAMT_NODE_BITMAP : {
+                        Uint n = hashmap_bitcount(MAP_HEADER_VAL(hdr));
+                        sum += 1 + n + header_arity(hdr);
+                        ptr += 1 + header_arity(hdr);
+                        while (n--) {
+                            obj = *ptr++;
+                            if (!IS_CONST(obj)) {
+                                EQUEUE_PUT_UNCHECKED(s, obj);
+                            }
+                        }
+                        goto cleanup_next;
+                    }
+                    default:
+                        erl_exit(ERTS_ABORT_EXIT, "size_shared: bad hashmap type %d\n", MAP_HEADER_TYPE(hdr));
+                }
+	    default:
+		goto cleanup_next;
+	    }
+	    break;
+	}
+	case TAG_PRIMARY_IMMED1:
+	cleanup_next:
+	    if (EQUEUE_ISEMPTY(s)) {
+		goto all_clean;
+	    }
+	    obj = EQUEUE_GET(s);
+	    break;
+	default:
+	    erl_exit(ERTS_ABORT_EXIT, "size_shared: bad tag for %#x\n", obj);
+	}
+    }
+
+ all_clean:
+    /* Return the result */
+    DESTROY_EQUEUE(s);
+    DESTROY_BITSTORE(b);
+    return sum;
+}
+
+
+/*
  *  Copy a structure to a heap.
  */
-#if HALFWORD_HEAP
-Eterm copy_struct_rel(Eterm obj, Uint sz, Eterm** hpp, ErlOffHeap* off_heap,
-                      Eterm* src_base, Eterm* dst_base)
-#else
-Eterm copy_struct(Eterm obj, Uint sz, Eterm** hpp, ErlOffHeap* off_heap)
-#endif
+Eterm copy_struct_x(Eterm obj, Uint sz, Eterm** hpp, ErlOffHeap* off_heap, Uint *bsz)
 {
     char* hstart;
     Uint hsize;
@@ -247,19 +614,23 @@ Eterm copy_struct(Eterm obj, Uint sz, Eterm** hpp, ErlOffHeap* off_heap)
     Eterm* argp;
     Eterm* const_tuple;
     Eterm hdr;
+    Eterm *hend;
     int i;
 #ifdef DEBUG
     Eterm org_obj = obj;
     Uint org_sz = sz;
+    Eterm mypid = erts_get_current_pid();
 #endif
 
     if (IS_CONST(obj))
 	return obj;
 
+    VERBOSE(DEBUG_SHCOPY, ("[pid=%T] copy_struct %p\n", mypid, obj));
+
     DTRACE1(copy_struct, (int32_t)sz);
 
     hp = htop = *hpp;
-    hbot   = htop + sz;
+    hbot = hend = htop + sz;
     hstart = (char *)htop;
     hsize = (char*) hbot - hstart;
     const_tuple = 0;
@@ -268,7 +639,7 @@ Eterm copy_struct(Eterm obj, Uint sz, Eterm** hpp, ErlOffHeap* off_heap)
     switch (primary_tag(obj)) {
     case TAG_PRIMARY_LIST:
 	argp = &res;
-	objp = list_val_rel(obj,src_base);
+	objp = list_val(obj);
 	goto L_copy_list;
     case TAG_PRIMARY_BOXED: argp = &res; goto L_copy_boxed;
     default:
@@ -286,14 +657,11 @@ Eterm copy_struct(Eterm obj, Uint sz, Eterm** hpp, ErlOffHeap* off_heap)
 	    hp++;
 	    break;
 	case TAG_PRIMARY_LIST:
-	    objp = list_val_rel(obj,src_base);
-	#if !HALFWORD_HEAP || defined(DEBUG)
-	    if (in_area(objp,hstart,hsize)) {
-		ASSERT(!HALFWORD_HEAP);
+	    objp = list_val(obj);
+	    if (ErtsInArea(objp,hstart,hsize)) {
 		hp++;
 		break;
 	    }
-	#endif
 	    argp = hp++;
 	    /* Fall through */
 
@@ -309,23 +677,15 @@ Eterm copy_struct(Eterm obj, Uint sz, Eterm** hpp, ErlOffHeap* off_heap)
 		}
 		else {
 		    CAR(htop) = elem;
-		#if HALFWORD_HEAP
-		    CDR(htop) = CDR(objp);
-		    *tailp = make_list_rel(htop,dst_base);
-		    htop += 2;
-		    goto L_copy;
-		#else
 		    tailp = &CDR(htop);
 		    htop += 2;
-		#endif
 		}
-		ASSERT(!HALFWORD_HEAP || tp < hp || tp >= hbot);
-		*tp = make_list_rel(tailp - 1, dst_base);
+		*tp = make_list(tailp - 1);
 		obj = CDR(objp);
 		if (!is_list(obj)) {
 		    break;
 		}
-		objp = list_val_rel(obj,src_base);
+		objp = list_val(obj);
 	    }
 	    switch (primary_tag(obj)) {
 	    case TAG_PRIMARY_IMMED1: *tailp = obj; goto L_copy;
@@ -337,24 +697,21 @@ Eterm copy_struct(Eterm obj, Uint sz, Eterm** hpp, ErlOffHeap* off_heap)
 	    }
 	    
 	case TAG_PRIMARY_BOXED:
-	#if !HALFWORD_HEAP || defined(DEBUG)
-	    if (in_area(boxed_val_rel(obj,src_base),hstart,hsize)) {
-		ASSERT(!HALFWORD_HEAP);
+	    if (ErtsInArea(boxed_val(obj),hstart,hsize)) {
 		hp++;
 		break;
 	    }
-	#endif
 	    argp = hp++;
 
 	L_copy_boxed:
-	    objp = boxed_val_rel(obj, src_base);
+	    objp = boxed_val(obj);
 	    hdr = *objp;
 	    switch (hdr & _TAG_HEADER_MASK) {
 	    case ARITYVAL_SUBTAG:
 		{
 		    int const_flag = 1; /* assume constant tuple */
 		    i = arityval(hdr);
-		    *argp = make_tuple_rel(htop, dst_base);
+		    *argp = make_tuple(htop);
 		    tp = htop;	/* tp is pointer to new arity value */
 		    *htop++ = *objp++; /* copy arity value */
 		    while (i--) {
@@ -383,7 +740,7 @@ Eterm copy_struct(Eterm obj, Uint sz, Eterm** hpp, ErlOffHeap* off_heap)
 		    while (i--)  {
 			*tp++ = *objp++;
 		    }
-		    *argp = make_binary_rel(hbot, dst_base);
+		    *argp = make_binary(hbot);
 		    pb = (ProcBin*) hbot;
 		    erts_refc_inc(&pb->val->refc, 2);
 		    pb->next = off_heap->first;
@@ -410,7 +767,7 @@ Eterm copy_struct(Eterm obj, Uint sz, Eterm** hpp, ErlOffHeap* off_heap)
 			extra_bytes = 0;
 		    } 
 		    real_size = size+extra_bytes;
-		    objp = binary_val_rel(real_bin,src_base);
+		    objp = binary_val(real_bin);
 		    if (thing_subtag(*objp) == HEAP_BINARY_SUBTAG) {
 			ErlHeapBin* from = (ErlHeapBin *) objp;
 			ErlHeapBin* to;
@@ -440,7 +797,7 @@ Eterm copy_struct(Eterm obj, Uint sz, Eterm** hpp, ErlOffHeap* off_heap)
 			off_heap->first = (struct erl_off_heap_header*) to;
 			OH_OVERHEAD(off_heap, to->size / sizeof(Eterm));
 		    }
-		    *argp = make_binary_rel(hbot, dst_base);
+		    *argp = make_binary(hbot);
 		    if (extra_bytes != 0) {
 			ErlSubBin* res;
 			hbot -= ERL_SUB_BIN_SIZE;
@@ -452,7 +809,7 @@ Eterm copy_struct(Eterm obj, Uint sz, Eterm** hpp, ErlOffHeap* off_heap)
 			res->offs = 0;
 			res->is_writable = 0;
 			res->orig = *argp;
-			*argp = make_binary_rel(hbot, dst_base);
+			*argp = make_binary(hbot);
 		    }
 		    break;
 		}
@@ -470,7 +827,7 @@ Eterm copy_struct(Eterm obj, Uint sz, Eterm** hpp, ErlOffHeap* off_heap)
 		    funp->next = off_heap->first;
 		    off_heap->first = (struct erl_off_heap_header*) funp;
 		    erts_refc_inc(&funp->fe->refc, 2);
-		    *argp = make_fun_rel(tp, dst_base);
+		    *argp = make_fun(tp);
 		}
 		break;
 	    case EXTERNAL_PID_SUBTAG:
@@ -490,7 +847,7 @@ Eterm copy_struct(Eterm obj, Uint sz, Eterm** hpp, ErlOffHeap* off_heap)
 		  off_heap->first = (struct erl_off_heap_header*)etp;
 		  erts_refc_inc(&etp->node->refc, 2);
 
-		  *argp = make_external_rel(tp, dst_base);
+		  *argp = make_external(tp);
 		}
 		break;
 	    case MAP_SUBTAG:
@@ -498,7 +855,7 @@ Eterm copy_struct(Eterm obj, Uint sz, Eterm** hpp, ErlOffHeap* off_heap)
 		switch (MAP_HEADER_TYPE(hdr)) {
 		    case MAP_HEADER_TAG_FLATMAP_HEAD :
                         i = flatmap_get_size(objp) + 3;
-                        *argp = make_flatmap_rel(htop, dst_base);
+                        *argp = make_flatmap(htop);
                         while (i--) {
                             *htop++ = *objp++;
                         }
@@ -509,7 +866,7 @@ Eterm copy_struct(Eterm obj, Uint sz, Eterm** hpp, ErlOffHeap* off_heap)
 		    case MAP_HEADER_TAG_HAMT_NODE_BITMAP :
 			i = 1 + hashmap_bitcount(MAP_HEADER_VAL(hdr));
 			while (i--)  { *htop++ = *objp++; }
-			*argp = make_hashmap_rel(tp, dst_base);
+			*argp = make_hashmap(tp);
 			break;
 		    default:
 			erl_exit(ERTS_ABORT_EXIT, "copy_struct: bad hashmap type %d\n", MAP_HEADER_TYPE(hdr));
@@ -522,7 +879,7 @@ Eterm copy_struct(Eterm obj, Uint sz, Eterm** hpp, ErlOffHeap* off_heap)
 		i = thing_arityval(hdr)+1;
 		hbot -= i;
 		tp = hbot;
-		*argp = make_boxed_rel(hbot, dst_base);
+		*argp = make_boxed(hbot);
 		while (i--) {
 		    *tp++ = *objp++;
 		}
@@ -538,22 +895,870 @@ Eterm copy_struct(Eterm obj, Uint sz, Eterm** hpp, ErlOffHeap* off_heap)
 	}
     }
 
+    if (bsz) {
+        *hpp = htop;
+        *bsz = hend - hbot;
+    } else {
 #ifdef DEBUG
-    if (htop != hbot)
-	erl_exit(ERTS_ABORT_EXIT,
-		 "Internal error in copy_struct() when copying %T:"
-		 " htop=%p != hbot=%p (sz=%beu)\n",
-		 org_obj, htop, hbot, org_sz); 
+        if (htop != hbot)
+            erl_exit(ERTS_ABORT_EXIT,
+                    "Internal error in copy_struct() when copying %T:"
+                    " htop=%p != hbot=%p (sz=%beu)\n",
+                    org_obj, htop, hbot, org_sz);
 #else
-    if (htop > hbot) {
-	erl_exit(ERTS_ABORT_EXIT,
-		 "Internal error in copy_struct(): htop, hbot overrun\n");
-    }
+        if (htop > hbot) {
+            erl_exit(ERTS_ABORT_EXIT,
+                    "Internal error in copy_struct(): htop, hbot overrun\n");
+        }
 #endif
-    *hpp = (Eterm *) (hstart+hsize);
+        *hpp = (Eterm *) (hstart+hsize);
+    }
+    VERBOSE(DEBUG_SHCOPY, ("[pid=%T] result is at %p\n", mypid, res));
     return res;
 }
 
+
+/*
+ *  Machinery for the table used by the sharing preserving copier
+ *  Using an ESTACK but not very transparently; consider refactoring
+ */
+
+#define DECLARE_SHTABLE(s)					\
+    DECLARE_ESTACK(s);						\
+    Uint ESTK_CONCAT(s,_offset) = 0
+#define DESTROY_SHTABLE(s) DESTROY_ESTACK(s)
+#define SHTABLE_INCR 4
+#define SHTABLE_NEXT(s)	ESTK_CONCAT(s,_offset)
+#define SHTABLE_PUSH(s,x,y,b)					\
+do {								\
+    if (s.sp > s.end - SHTABLE_INCR) {				\
+        erl_grow_estack(&(s), SHTABLE_INCR);	       		\
+    }								\
+    *s.sp++ = (x);						\
+    *s.sp++ = (y);						\
+    *s.sp++ = (Eterm) NULL;					\
+    *s.sp++ = (Eterm) (b);		                	\
+    ESTK_CONCAT(s,_offset) += SHTABLE_INCR;			\
+} while(0)
+#define SHTABLE_X(s,e) (s.start[e])
+#define SHTABLE_Y(s,e) (s.start[(e)+1])
+#define SHTABLE_FWD(s,e) ((Eterm *) (s.start[(e)+2]))
+#define SHTABLE_FWD_UPD(s,e,p) (s.start[(e)+2] = (Eterm) (p))
+#define SHTABLE_REV(s,e) ((Eterm *) (s.start[(e)+3]))
+
+#define LIST_SHARED_UNPROCESSED ((Eterm) 0)
+#define LIST_SHARED_PROCESSED	((Eterm) 1)
+
+#define HEAP_ELEM_TO_BE_FILLED	_unchecked_make_list(NULL)
+
+
+/*
+ *  Specialized macros for using/reusing the persistent state
+ */
+
+#define DECLARE_EQUEUE_INIT_INFO(q, info)		\
+    UWord* EQUE_DEF_QUEUE(q) = info->queue_default;	\
+    ErtsEQueue q = {					\
+        EQUE_DEF_QUEUE(q), /* start */			\
+        EQUE_DEF_QUEUE(q), /* front */			\
+        EQUE_DEF_QUEUE(q), /* back */			\
+        1,                 /* possibly_empty */		\
+        EQUE_DEF_QUEUE(q) + DEF_EQUEUE_SIZE, /* end */	\
+        ERTS_ALC_T_ESTACK  /* alloc_type */		\
+    }
+
+#define DECLARE_EQUEUE_FROM_INFO(q, info)		\
+    /* no EQUE_DEF_QUEUE(q), read-only */		\
+    ErtsEQueue q = {					\
+        info->queue_start,      /* start */		\
+        info->queue_start,      /* front */		\
+        info->queue_start,      /* back */		\
+        1,                      /* possibly_empty */	\
+        info->queue_end,        /* end */		\
+        info->queue_alloc_type  /* alloc_type */	\
+    }
+
+#define DECLARE_BITSTORE_INIT_INFO(s, info)		\
+    UWord* WSTK_DEF_STACK(s) = info->bitstore_default;	\
+    ErtsWStack s = {					\
+        WSTK_DEF_STACK(s),  /* wstart */		\
+        WSTK_DEF_STACK(s),  /* wsp */			\
+        WSTK_DEF_STACK(s) + DEF_WSTACK_SIZE, /* wend */	\
+        WSTK_DEF_STACK(s),  /* wdflt */ 		\
+        ERTS_ALC_T_ESTACK /* alloc_type */		\
+    };							\
+    int WSTK_CONCAT(s,_bitoffs) = 0;			\
+    /* no WSTK_CONCAT(s,_offset), write-only */		\
+    UWord WSTK_CONCAT(s,_buffer) = 0
+
+#define DECLARE_BITSTORE_FROM_INFO(s, info)		\
+    /* no WSTK_DEF_STACK(s), read-only */		\
+    ErtsWStack s = {					\
+        info->bitstore_start,  /* wstart */		\
+        NULL,                  /* wsp,  read-only */	\
+        NULL,                  /* wend, read-only */	\
+        NULL,                  /* wdef, read-only */	\
+        info->bitstore_alloc_type /* alloc_type */	\
+    };							\
+    int WSTK_CONCAT(s,_bitoffs) = 0;			\
+    int WSTK_CONCAT(s,_offset) = 0;			\
+    UWord WSTK_CONCAT(s,_buffer) = 0
+
+#define DECLARE_SHTABLE_INIT_INFO(s, info)		\
+    Eterm* ESTK_DEF_STACK(s) = info->shtable_default;	\
+    ErtsEStack s = {					\
+        ESTK_DEF_STACK(s),  /* start */			\
+        ESTK_DEF_STACK(s),  /* sp */			\
+        ESTK_DEF_STACK(s) + DEF_ESTACK_SIZE, /* end */	\
+        ESTK_DEF_STACK(s),  /* default */		\
+        ERTS_ALC_T_ESTACK /* alloc_type */		\
+    };							\
+    Uint ESTK_CONCAT(s,_offset) = 0
+
+#define DECLARE_SHTABLE_FROM_INFO(s, info)		\
+    /* no ESTK_DEF_STACK(s), read-only */		\
+    ErtsEStack s = {					\
+        info->shtable_start,     /* start */		\
+        NULL,                    /* sp,  read-only */	\
+        NULL,                    /* end, read-only */	\
+        NULL,                    /* def, read-only */	\
+        info->shtable_alloc_type /* alloc_type */	\
+    };							\
+    /* no ESTK_CONCAT(s,_offset), read-only */
+
+/*
+ *  Copy object "obj" preserving sharing.
+ *  First half: count size and calculate sharing.
+ */
+Uint copy_shared_calculate(Eterm obj, erts_shcopy_t *info)
+{
+    Uint sum;
+    Uint e;
+    unsigned sz;
+    Eterm* ptr;
+#ifdef DEBUG
+    Eterm mypid = erts_get_current_pid();
+#endif
+
+    DECLARE_EQUEUE_INIT_INFO(s, info);
+    DECLARE_BITSTORE_INIT_INFO(b, info);
+    DECLARE_SHTABLE_INIT_INFO(t, info);
+
+    /* step #0:
+       -------------------------------------------------------
+       get rid of the easy cases first:
+       - copying constants
+       - if not a proper process, do flat copy
+    */
+
+    if (IS_CONST(obj))
+	return 0;
+
+    VERBOSE(DEBUG_SHCOPY, ("[pid=%T] copy_shared_calculate %p\n", mypid, obj));
+    VERBOSE(DEBUG_SHCOPY, ("[pid=%T] message is %T\n", mypid, obj));
+
+    /* step #1:
+       -------------------------------------------------------
+       traverse the term and calculate the size;
+       when traversing, transform as you do in size_shared
+       but when you find shared objects:
+
+       a. add entry in the table, indexed by i
+       b. mark them:
+	  b1. boxed terms, set header to (i | 11)
+	      store (old header, NONV, NULL, backptr) in the entry
+	  b2. cons cells, set CDR to NONV, set CAR to i
+	      store (old CAR, old CDR, NULL, backptr) in the entry
+    */
+
+    sum = 0;
+
+    for (;;) {
+	switch (primary_tag(obj)) {
+	case TAG_PRIMARY_LIST: {
+	    Eterm head, tail;
+	    ptr = list_val(obj);
+	    /* off heap list pointers are copied verbatim */
+	    if (erts_is_literal(obj,ptr)) {
+		VERBOSE(DEBUG_SHCOPY, ("[pid=%T] bypassed copying %p is %T\n", mypid, ptr, obj));
+                if (IN_LITERAL_PURGE_AREA(info,ptr))
+                    info->literal_size += size_object(obj);
+		goto pop_next;
+	    }
+	    head = CAR(ptr);
+	    tail = CDR(ptr);
+	    /* if it's visited, don't count it;
+	       if not already shared, make it shared and store it in the table */
+	    if (primary_tag(tail) == TAG_PRIMARY_HEADER ||
+		primary_tag(head) == TAG_PRIMARY_HEADER) {
+		if (tail != THE_NON_VALUE) {
+		    e = SHTABLE_NEXT(t);
+		    VERBOSE(DEBUG_SHCOPY, ("[pid=%T] tabling L %p\n", mypid, ptr));
+		    SHTABLE_PUSH(t, head, tail, ptr);
+		    CAR(ptr) = (e << _TAG_PRIMARY_SIZE) | LIST_SHARED_UNPROCESSED;
+		    CDR(ptr) = THE_NON_VALUE;
+		}
+		goto pop_next;
+	    }
+	    /* else make it visited now */
+	    switch (primary_tag(tail)) {
+	    case TAG_PRIMARY_LIST:
+		VERBOSE(DEBUG_SHCOPY, ("[pid=%T] mangling L/L %p\n", mypid, ptr));
+		CDR(ptr) = (tail - TAG_PRIMARY_LIST) | TAG_PRIMARY_HEADER;
+		break;
+	    case TAG_PRIMARY_IMMED1:
+		VERBOSE(DEBUG_SHCOPY, ("[pid=%T] mangling L/I %p\n", mypid, ptr));
+		CAR(ptr) = (head - primary_tag(head)) | TAG_PRIMARY_HEADER;
+		CDR(ptr) = (tail - TAG_PRIMARY_IMMED1) | primary_tag(head);
+		break;
+	    case TAG_PRIMARY_BOXED:
+		BITSTORE_PUT(b, primary_tag(head));
+		VERBOSE(DEBUG_SHCOPY, ("[pid=%T] mangling L/B %p\n", mypid, ptr));
+		CAR(ptr) = (head - primary_tag(head)) | TAG_PRIMARY_HEADER;
+		CDR(ptr) = (tail - TAG_PRIMARY_BOXED) | TAG_PRIMARY_HEADER;
+		break;
+	    }
+	    /* and count it */
+	    sum += 2;
+	    if (!IS_CONST(head)) {
+		EQUEUE_PUT(s, head);
+	    }
+	    obj = tail;
+	    break;
+	}
+	case TAG_PRIMARY_BOXED: {
+	    Eterm hdr;
+	    ptr = boxed_val(obj);
+	    /* off heap pointers to boxes are copied verbatim */
+	    if (erts_is_literal(obj,ptr)) {
+		VERBOSE(DEBUG_SHCOPY, ("[pid=%T] bypassed copying %p is %T\n", mypid, ptr, obj));
+                if (IN_LITERAL_PURGE_AREA(info,ptr))
+                    info->literal_size += size_object(obj);
+		goto pop_next;
+	    }
+	    hdr = *ptr;
+	    /* if it's visited, don't count it;
+	       if not already shared, make it shared and store it in the table */
+	    if (primary_tag(hdr) != TAG_PRIMARY_HEADER) {
+		if (primary_tag(hdr) == BOXED_VISITED) {
+		    e = SHTABLE_NEXT(t);
+		    VERBOSE(DEBUG_SHCOPY, ("[pid=%T] tabling B %p\n", mypid, ptr));
+		    SHTABLE_PUSH(t, hdr, THE_NON_VALUE, ptr);
+		    *ptr = (e << _TAG_PRIMARY_SIZE) | BOXED_SHARED_UNPROCESSED;
+		}
+		goto pop_next;
+	    }
+	    /* else make it visited now */
+	    VERBOSE(DEBUG_SHCOPY, ("[pid=%T] mangling B %p\n", mypid, ptr));
+	    *ptr = (hdr - primary_tag(hdr)) + BOXED_VISITED;
+	    /* and count it */
+	    ASSERT(is_header(hdr));
+	    switch (hdr & _TAG_HEADER_MASK) {
+	    case ARITYVAL_SUBTAG: {
+		int arity = header_arity(hdr);
+		sum += arity + 1;
+		if (arity == 0) { /* Empty tuple -- unusual. */
+		    goto pop_next;
+		}
+		while (arity-- > 0) {
+		    obj = *++ptr;
+		    if (!IS_CONST(obj)) {
+			EQUEUE_PUT(s, obj);
+		    }
+		}
+		goto pop_next;
+	    }
+	    case FUN_SUBTAG: {
+		ErlFunThing* funp = (ErlFunThing *) ptr;
+		unsigned eterms = 1 /* creator */ + funp->num_free;
+		sz = thing_arityval(hdr);
+		sum += 1 /* header */ + sz + eterms;
+		ptr += 1 /* header */ + sz;
+		while (eterms-- > 0) {
+		    obj = *ptr++;
+		    if (!IS_CONST(obj)) {
+			EQUEUE_PUT(s, obj);
+		    }
+		}
+		goto pop_next;
+	    }
+	    case SUB_BINARY_SUBTAG: {
+		ErlSubBin* sb = (ErlSubBin *) ptr;
+		Eterm real_bin = sb->orig;
+		Uint bit_offset = sb->bitoffs;
+		Uint bit_size = sb->bitsize;
+		size_t size = sb->size;
+		Uint extra_bytes;
+		Eterm hdr;
+		if (bit_size + bit_offset > 8) {
+		    sum += ERL_SUB_BIN_SIZE;
+		    extra_bytes = 2;
+		} else if (bit_size + bit_offset > 0) {
+		    sum += ERL_SUB_BIN_SIZE;
+		    extra_bytes = 1;
+		} else {
+		    extra_bytes = 0;
+		}
+		ASSERT(is_boxed(real_bin) &&
+		       (((*boxed_val(real_bin)) &
+			 (_TAG_HEADER_MASK - _BINARY_XXX_MASK - BOXED_VISITED_MASK))
+			== _TAG_HEADER_REFC_BIN));
+		hdr = *_unchecked_binary_val(real_bin) & ~BOXED_VISITED_MASK;
+		if (thing_subtag(hdr) == HEAP_BINARY_SUBTAG) {
+		    sum += heap_bin_size(size+extra_bytes);
+		} else {
+		    ASSERT(thing_subtag(hdr) == REFC_BINARY_SUBTAG);
+		    sum += PROC_BIN_SIZE;
+		}
+		goto pop_next;
+	    }
+            case MAP_SUBTAG:
+                switch (MAP_HEADER_TYPE(hdr)) {
+                    case MAP_HEADER_TAG_FLATMAP_HEAD : {
+                        flatmap_t *mp = (flatmap_t *) ptr;
+                        Uint n = flatmap_get_size(mp) + 1;
+                        sum += n + 2;
+                        ptr += 2; /* hdr + size words */
+                        while (n--) {
+                            obj = *ptr++;
+                            if (!IS_CONST(obj)) {
+                                EQUEUE_PUT(s, obj);
+                            }
+                        }
+                        goto pop_next;
+                    }
+                    case MAP_HEADER_TAG_HAMT_HEAD_BITMAP :
+                    case MAP_HEADER_TAG_HAMT_HEAD_ARRAY :
+                    case MAP_HEADER_TAG_HAMT_NODE_BITMAP : {
+                        Uint n = hashmap_bitcount(MAP_HEADER_VAL(hdr));
+                        sum += 1 + n + header_arity(hdr);
+                        ptr += 1 + header_arity(hdr);
+
+                        if (n == 0) {
+                            goto pop_next;
+                        }
+                        while(n--) {
+                            obj = *ptr++;
+                            if (!IS_CONST(obj)) {
+                                EQUEUE_PUT(s, obj);
+                            }
+                        }
+                        goto pop_next;
+                    }
+                    default:
+                        erl_exit(ERTS_ABORT_EXIT, "copy_shared_calculate: bad hashmap type %d\n", MAP_HEADER_TYPE(hdr));
+                }
+            case BIN_MATCHSTATE_SUBTAG:
+		erl_exit(ERTS_ABORT_EXIT,
+			 "size_shared: matchstate term not allowed");
+	    default:
+		sum += thing_arityval(hdr) + 1;
+		goto pop_next;
+	    }
+	    break;
+	}
+	case TAG_PRIMARY_IMMED1:
+	pop_next:
+	    if (EQUEUE_ISEMPTY(s)) {
+                /* add sentinel to the table */
+                SHTABLE_PUSH(t, THE_NON_VALUE, THE_NON_VALUE, NULL);
+                /* store persistent info */
+                BITSTORE_CLOSE(b);
+                info->queue_start = s.start;
+                info->queue_end = s.end;
+                info->queue_alloc_type = s.alloc_type;
+                info->bitstore_start = b.wstart;
+                info->bitstore_alloc_type = b.alloc_type;
+                info->shtable_start = t.start;
+                info->shtable_alloc_type = t.alloc_type;
+                /* single point of return: the size of the object */
+                VERBOSE(DEBUG_SHCOPY, ("[pid=%T] size was: %u\n", mypid, sum));
+                return sum + info->literal_size;
+	    }
+	    obj = EQUEUE_GET(s);
+	    break;
+	default:
+	    erl_exit(ERTS_ABORT_EXIT, "[pid=%T] size_shared: bad tag for %#x\n", obj);
+	}
+    }
+}
+
+/*
+ *  Copy object "obj" preserving sharing.
+ *  Second half: copy and restore the object.
+ */
+Uint copy_shared_perform(Eterm obj, Uint size, erts_shcopy_t *info,
+                         Eterm** hpp, ErlOffHeap* off_heap) {
+    Uint e;
+    unsigned sz;
+    Eterm* ptr;
+    Eterm* hp;
+    Eterm* hscan;
+    Eterm result;
+    Eterm* resp;
+    Eterm *hbot, *hend;
+    unsigned remaining;
+#ifdef DEBUG
+    Eterm mypid = erts_get_current_pid();
+    Eterm saved_obj = obj;
+#endif
+
+    DECLARE_EQUEUE_FROM_INFO(s, info);
+    DECLARE_BITSTORE_FROM_INFO(b, info);
+    DECLARE_SHTABLE_FROM_INFO(t, info);
+
+    /* step #0:
+       -------------------------------------------------------
+       get rid of the easy cases first:
+       - copying constants
+       - if not a proper process, do flat copy
+    */
+
+    if (IS_CONST(obj))
+	return obj;
+
+    VERBOSE(DEBUG_SHCOPY, ("[pid=%T] copy_shared_perform %p\n", mypid, obj));
+
+    /* step #2: was performed before this function was called
+       -------------------------------------------------------
+       allocate new space
+    */
+
+    hscan = hp = *hpp;
+    hbot  = hend = hp + size;
+
+    /* step #3:
+       -------------------------------------------------------
+       traverse the term a second time and when traversing:
+       a. if the object is marked as shared
+	  a1. if the entry contains a forwarding ptr, use that
+	  a2. otherwise, copy it to the new space and store the
+	      forwarding ptr to the entry
+      b. otherwise, reverse-transform as you do in size_shared
+	 and copy to the new space
+    */
+
+    resp = &result;
+    remaining = 0;
+    for (;;) {
+	switch (primary_tag(obj)) {
+	case TAG_PRIMARY_LIST: {
+	    Eterm head, tail;
+	    ptr = list_val(obj);
+	    /* off heap list pointers are copied verbatim */
+	    if (erts_is_literal(obj,ptr)) {
+                if (!IN_LITERAL_PURGE_AREA(info,ptr)) {
+                    *resp = obj;
+                } else {
+                    Uint bsz = 0;
+                    *resp = copy_struct_x(obj, hbot - hp, &hp, off_heap, &bsz);
+                    hbot -= bsz;
+                }
+		goto cleanup_next;
+	    }
+	    head = CAR(ptr);
+	    tail = CDR(ptr);
+	    /* if it is shared */
+	    if (tail == THE_NON_VALUE) {
+		e = head >> _TAG_PRIMARY_SIZE;
+		/* if it has been processed, just use the forwarding pointer */
+		if (primary_tag(head) == LIST_SHARED_PROCESSED) {
+		    *resp = make_list(SHTABLE_FWD(t, e));
+		    goto cleanup_next;
+		}
+		/* else, let's process it now,
+		   copy it and keep the forwarding pointer */
+		else {
+		    CAR(ptr) = (head - primary_tag(head)) + LIST_SHARED_PROCESSED;
+		    head = SHTABLE_X(t, e);
+		    tail = SHTABLE_Y(t, e);
+		    ptr = &(SHTABLE_X(t, e));
+		    VERBOSE(DEBUG_SHCOPY, ("[pid=%T] tabled L %p is %p\n", mypid, ptr, SHTABLE_REV(t, e)));
+		    SHTABLE_FWD_UPD(t, e, hp);
+		}
+	    }
+	    /* if not already clean, clean it up and copy it */
+	    if (primary_tag(tail) == TAG_PRIMARY_HEADER) {
+		if (primary_tag(head) == TAG_PRIMARY_HEADER) {
+		    Eterm saved;
+                    BITSTORE_FETCH(b, saved);
+		    VERBOSE(DEBUG_SHCOPY, ("[pid=%T] unmangling L/B %p\n", mypid, ptr));
+		    CAR(ptr) = head = (head - TAG_PRIMARY_HEADER) + saved;
+		    CDR(ptr) = tail = (tail - TAG_PRIMARY_HEADER) + TAG_PRIMARY_BOXED;
+		} else {
+		    VERBOSE(DEBUG_SHCOPY, ("[pid=%T] unmangling L/L %p\n", mypid, ptr));
+		    CDR(ptr) = tail = (tail - TAG_PRIMARY_HEADER) + TAG_PRIMARY_LIST;
+		}
+	    } else if (primary_tag(head) == TAG_PRIMARY_HEADER) {
+		VERBOSE(DEBUG_SHCOPY, ("[pid=%T] unmangling L/I %p\n", mypid, ptr));
+		CAR(ptr) = head = (head - TAG_PRIMARY_HEADER) | primary_tag(tail);
+		CDR(ptr) = tail = (tail - primary_tag(tail)) | TAG_PRIMARY_IMMED1;
+	    } else {
+		ASSERT(0 && "cannot come here");
+		goto cleanup_next;
+	    }
+	    /* and its children too */
+	    if (IS_CONST(head)) {
+		CAR(hp) = head;
+	    } else {
+		EQUEUE_PUT_UNCHECKED(s, head);
+		CAR(hp) = HEAP_ELEM_TO_BE_FILLED;
+	    }
+	    *resp = make_list(hp);
+	    resp = &(CDR(hp));
+	    hp += 2;
+	    obj = tail;
+	    break;
+	}
+	case TAG_PRIMARY_BOXED: {
+	    Eterm hdr;
+	    ptr = boxed_val(obj);
+	    /* off heap pointers to boxes are copied verbatim */
+	    if (erts_is_literal(obj,ptr)) {
+                if (!IN_LITERAL_PURGE_AREA(info,ptr)) {
+                    *resp = obj;
+                } else {
+                    Uint bsz = 0;
+                    *resp = copy_struct_x(obj, hbot - hp, &hp, off_heap, &bsz);
+                    hbot -= bsz;
+                }
+		goto cleanup_next;
+	    }
+	    hdr = *ptr;
+	    /* clean it up, unless it's already clean or shared and processed */
+	    switch (primary_tag(hdr)) {
+	    case TAG_PRIMARY_HEADER:
+		ASSERT(0 && "cannot come here");
+	    /* if it is shared and has been processed,
+	       just use the forwarding pointer */
+	    case BOXED_SHARED_PROCESSED:
+		e = hdr >> _TAG_PRIMARY_SIZE;
+		*resp = make_boxed(SHTABLE_FWD(t, e));
+		goto cleanup_next;
+	    /* if it is shared but has not been processed yet, let's process
+	       it now: copy it and keep the forwarding pointer */
+	    case BOXED_SHARED_UNPROCESSED:
+		e = hdr >> _TAG_PRIMARY_SIZE;
+		*ptr = (hdr - primary_tag(hdr)) + BOXED_SHARED_PROCESSED;
+		hdr = SHTABLE_X(t, e);
+		ASSERT(primary_tag(hdr) == BOXED_VISITED);
+		VERBOSE(DEBUG_SHCOPY, ("[pid=%T] tabled B %p is %p\n", mypid, ptr, SHTABLE_REV(t, e)));
+		VERBOSE(DEBUG_SHCOPY, ("[pid=%T] unmangling B %p\n", mypid, ptr));
+		SHTABLE_X(t, e) = hdr = (hdr - BOXED_VISITED) + TAG_PRIMARY_HEADER;
+		SHTABLE_FWD_UPD(t, e, hp);
+		break;
+	    case BOXED_VISITED:
+		VERBOSE(DEBUG_SHCOPY, ("[pid=%T] unmangling B %p\n", mypid, ptr));
+		*ptr = hdr = (hdr - BOXED_VISITED) + TAG_PRIMARY_HEADER;
+		break;
+	    }
+	    /* and its children too */
+	    switch (hdr & _TAG_HEADER_MASK) {
+	    case ARITYVAL_SUBTAG: {
+		int arity = header_arity(hdr);
+		*resp = make_boxed(hp);
+		*hp++ = hdr;
+		while (arity-- > 0) {
+		    obj = *++ptr;
+		    if (IS_CONST(obj)) {
+			*hp++ = obj;
+		    } else {
+			EQUEUE_PUT_UNCHECKED(s, obj);
+			*hp++ = HEAP_ELEM_TO_BE_FILLED;
+		    }
+		}
+		goto cleanup_next;
+	    }
+	    case FUN_SUBTAG: {
+		ErlFunThing* funp = (ErlFunThing *) ptr;
+		unsigned eterms = 1 /* creator */ + funp->num_free;
+		sz = thing_arityval(hdr);
+		funp = (ErlFunThing *) hp;
+		*resp = make_fun(hp);
+		*hp++ = hdr;
+		ptr++;
+		while (sz-- > 0) {
+		    *hp++ = *ptr++;
+		}
+		while (eterms-- > 0) {
+		    obj = *ptr++;
+		    if (IS_CONST(obj)) {
+			*hp++ = obj;
+		    } else {
+			EQUEUE_PUT_UNCHECKED(s, obj);
+			*hp++ = HEAP_ELEM_TO_BE_FILLED;
+		    }
+		}
+		funp->next = off_heap->first;
+		off_heap->first = (struct erl_off_heap_header*) funp;
+		erts_refc_inc(&funp->fe->refc, 2);
+		goto cleanup_next;
+	    }
+	    case MAP_SUBTAG:
+                *resp  = make_flatmap(hp);
+                *hp++  = hdr;
+                switch (MAP_HEADER_TYPE(hdr)) {
+                    case MAP_HEADER_TAG_FLATMAP_HEAD : {
+                        flatmap_t *mp = (flatmap_t *) ptr;
+                        Uint n = flatmap_get_size(mp) + 1;
+                        *hp++  = *++ptr; /* keys */
+                        while (n--) {
+                            obj = *++ptr;
+                            if (IS_CONST(obj)) {
+                                *hp++ = obj;
+                            } else {
+                                EQUEUE_PUT_UNCHECKED(s, obj);
+                                *hp++ = HEAP_ELEM_TO_BE_FILLED;
+                            }
+                        }
+                        goto cleanup_next;
+                    }
+                    case MAP_HEADER_TAG_HAMT_HEAD_BITMAP :
+                    case MAP_HEADER_TAG_HAMT_HEAD_ARRAY :
+			*hp++ = *++ptr; /* total map size */
+                    case MAP_HEADER_TAG_HAMT_NODE_BITMAP : {
+                         Uint n = hashmap_bitcount(MAP_HEADER_VAL(hdr));
+                         while (n--)  {
+                             obj = *++ptr;
+                             if (IS_CONST(obj)) {
+                                 *hp++ = obj;
+                             } else {
+                                 EQUEUE_PUT_UNCHECKED(s, obj);
+                                 *hp++ = HEAP_ELEM_TO_BE_FILLED;
+                             }
+                         }
+                        goto cleanup_next;
+                    }
+                    default:
+                        erl_exit(ERTS_ABORT_EXIT, "copy_shared_perform: bad hashmap type %d\n", MAP_HEADER_TYPE(hdr));
+                }
+	    case REFC_BINARY_SUBTAG: {
+		ProcBin* pb = (ProcBin *) ptr;
+		sz = thing_arityval(hdr);
+		if (pb->flags) {
+		    erts_emasculate_writable_binary(pb);
+		}
+		pb = (ProcBin *) hp;
+		*resp = make_binary(hp);
+		*hp++ = hdr;
+		ptr++;
+		while (sz-- > 0) {
+		    *hp++ = *ptr++;
+		}
+		erts_refc_inc(&pb->val->refc, 2);
+		pb->next = off_heap->first;
+		pb->flags = 0;
+		off_heap->first = (struct erl_off_heap_header*) pb;
+		OH_OVERHEAD(off_heap, pb->size / sizeof(Eterm));
+		goto cleanup_next;
+	    }
+	    case SUB_BINARY_SUBTAG: {
+		ErlSubBin* sb = (ErlSubBin *) ptr;
+		Eterm real_bin = sb->orig;
+		Uint bit_offset = sb->bitoffs;
+		Uint bit_size = sb->bitsize;
+		Uint offset = sb->offs;
+		size_t size = sb->size;
+		Uint extra_bytes;
+		Uint real_size;
+		if ((bit_size + bit_offset) > 8) {
+		    extra_bytes = 2;
+		} else if ((bit_size + bit_offset) > 0) {
+		    extra_bytes = 1;
+		} else {
+		    extra_bytes = 0;
+		}
+		real_size = size+extra_bytes;
+		ASSERT(is_boxed(real_bin) &&
+		       (((*boxed_val(real_bin)) &
+			 (_TAG_HEADER_MASK - _BINARY_XXX_MASK - BOXED_VISITED_MASK))
+			== _TAG_HEADER_REFC_BIN));
+		ptr = _unchecked_binary_val(real_bin);
+		*resp = make_binary(hp);
+		if (extra_bytes != 0) {
+		    ErlSubBin* res = (ErlSubBin *) hp;
+		    hp += ERL_SUB_BIN_SIZE;
+		    res->thing_word = HEADER_SUB_BIN;
+		    res->size = size;
+		    res->bitsize = bit_size;
+		    res->bitoffs = bit_offset;
+		    res->offs = 0;
+		    res->is_writable = 0;
+		    res->orig = make_binary(hp);
+		}
+		if (thing_subtag(*ptr & ~BOXED_VISITED_MASK) == HEAP_BINARY_SUBTAG) {
+		    ErlHeapBin* from = (ErlHeapBin *) ptr;
+		    ErlHeapBin* to = (ErlHeapBin *) hp;
+		    hp += heap_bin_size(real_size);
+		    to->thing_word = header_heap_bin(real_size);
+		    to->size = real_size;
+		    sys_memcpy(to->data, ((byte *)from->data)+offset, real_size);
+		} else {
+		    ProcBin* from = (ProcBin *) ptr;
+		    ProcBin* to = (ProcBin *) hp;
+		    ASSERT(thing_subtag(*ptr & ~BOXED_VISITED_MASK) == REFC_BINARY_SUBTAG);
+		    if (from->flags) {
+			erts_emasculate_writable_binary(from);
+		    }
+		    hp += PROC_BIN_SIZE;
+		    to->thing_word = HEADER_PROC_BIN;
+		    to->size = real_size;
+		    to->val = from->val;
+		    erts_refc_inc(&to->val->refc, 2);
+		    to->bytes = from->bytes + offset;
+		    to->next = off_heap->first;
+		    to->flags = 0;
+		    off_heap->first = (struct erl_off_heap_header*) to;
+		    OH_OVERHEAD(off_heap, to->size / sizeof(Eterm));
+		}
+		goto cleanup_next;
+	    }
+	    case EXTERNAL_PID_SUBTAG:
+	    case EXTERNAL_PORT_SUBTAG:
+	    case EXTERNAL_REF_SUBTAG: {
+		ExternalThing *etp = (ExternalThing *) hp;
+		sz = thing_arityval(hdr);
+		*resp = make_external(hp);
+		*hp++ = hdr;
+		ptr++;
+		while (sz-- > 0) {
+		    *hp++ = *ptr++;
+		}
+		etp->next = off_heap->first;
+		off_heap->first = (struct erl_off_heap_header*) etp;
+		erts_refc_inc(&etp->node->refc, 2);
+		goto cleanup_next;
+	    }
+	    default:
+		sz = thing_arityval(hdr);
+		*resp = make_boxed(hp);
+		*hp++ = hdr;
+		ptr++;
+		while (sz-- > 0) {
+		    *hp++ = *ptr++;
+		}
+		goto cleanup_next;
+	    }
+	    break;
+	}
+	case TAG_PRIMARY_IMMED1:
+	    *resp = obj;
+	cleanup_next:
+	    if (EQUEUE_ISEMPTY(s)) {
+		goto all_clean;
+	    }
+	    obj = EQUEUE_GET(s);
+	    for (;;) {
+		ASSERT(hscan < hp);
+		if (remaining == 0) {
+		    if (*hscan == HEAP_ELEM_TO_BE_FILLED) {
+			resp = hscan;
+			hscan += 2;
+			break; /* scanning loop */
+		    } else if (primary_tag(*hscan) == TAG_PRIMARY_HEADER) {
+			switch (*hscan & _TAG_HEADER_MASK) {
+			case ARITYVAL_SUBTAG:
+			    remaining = header_arity(*hscan);
+			    hscan++;
+			    break;
+			case FUN_SUBTAG: {
+			    ErlFunThing* funp = (ErlFunThing *) hscan;
+			    hscan += 1 + thing_arityval(*hscan);
+			    remaining = 1 + funp->num_free;
+			    break;
+			}
+			case MAP_SUBTAG:
+                            switch (MAP_HEADER_TYPE(*hscan)) {
+                                case MAP_HEADER_TAG_FLATMAP_HEAD : {
+                                    flatmap_t *mp = (flatmap_t *) hscan;
+                                    remaining = flatmap_get_size(mp) + 1;
+                                    hscan += 2;
+                                    break;
+                                }
+                                case MAP_HEADER_TAG_HAMT_HEAD_BITMAP :
+                                case MAP_HEADER_TAG_HAMT_HEAD_ARRAY :
+                                case MAP_HEADER_TAG_HAMT_NODE_BITMAP :
+                                    remaining = hashmap_bitcount(MAP_HEADER_VAL(*hscan));
+                                    hscan += MAP_HEADER_ARITY(*hscan) + 1;
+                                    break;
+                                default:
+                                    erl_exit(ERTS_ABORT_EXIT,
+                                            "copy_shared_perform: bad hashmap type %d\n",
+                                            MAP_HEADER_TYPE(*hscan));
+                            }
+                            break;
+			case SUB_BINARY_SUBTAG:
+			    ASSERT(((ErlSubBin *) hscan)->bitoffs +
+				   ((ErlSubBin *) hscan)->bitsize > 0);
+			    hscan += ERL_SUB_BIN_SIZE;
+			    break;
+			default:
+			    hscan += 1 + thing_arityval(*hscan);
+			    break;
+			}
+		    } else {
+			hscan++;
+		    }
+		} else if (*hscan == HEAP_ELEM_TO_BE_FILLED) {
+		    resp = hscan++;
+		    remaining--;
+		    break; /* scanning loop */
+		} else {
+		    hscan++;
+		    remaining--;
+		}
+	    }
+	    ASSERT(resp < hp);
+	    break;
+	default:
+	    erl_exit(ERTS_ABORT_EXIT, "size_shared: bad tag for %#x\n", obj);
+	}
+    }
+
+    /* step #4:
+       -------------------------------------------------------
+       traverse the table and reverse-transform all stored entries
+    */
+
+all_clean:
+    for (e = 0; ; e += SHTABLE_INCR) {
+	ptr = SHTABLE_REV(t, e);
+	if (ptr == NULL)
+	    break;
+	VERBOSE(DEBUG_SHCOPY, ("[copy] restoring shared: %x\n", ptr));
+	/* entry was a list */
+	if (SHTABLE_Y(t, e) != THE_NON_VALUE) {
+	    VERBOSE(DEBUG_SHCOPY, ("[pid=%T] untabling L %p\n", mypid, ptr));
+	    CAR(ptr) = SHTABLE_X(t, e);
+	    CDR(ptr) = SHTABLE_Y(t, e);
+	}
+	/* entry was boxed */
+	else {
+	    VERBOSE(DEBUG_SHCOPY, ("[pid=%T] untabling B %p\n", mypid, ptr));
+	    *ptr = SHTABLE_X(t, e);
+	    ASSERT(primary_tag(*ptr) == TAG_PRIMARY_HEADER);
+	}
+    }
+
+#ifdef DEBUG
+    if (eq(saved_obj, result) == 0) {
+	erts_fprintf(stderr, "original = %T\n", saved_obj);
+	erts_fprintf(stderr, "copy = %T\n", result);
+	erl_exit(ERTS_ABORT_EXIT, "copy (shared) not equal to source\n");
+    }
+#endif
+
+    VERBOSE(DEBUG_SHCOPY, ("[pid=%T] original was %T\n", mypid, saved_obj));
+    VERBOSE(DEBUG_SHCOPY, ("[pid=%T] copy is %T\n", mypid, result));
+    VERBOSE(DEBUG_SHCOPY, ("[pid=%T] result is at %p\n", mypid, result));
+
+    ASSERT(hbot == hp);
+    ASSERT(size == ((hp - *hpp) + (hend - hbot)));
+    *hpp = hend;
+    return result;
+}
+
+
 /*
  * Copy a term that is guaranteed to be contained in a single
  * heap block. The heap block is copied word by word, and any
@@ -563,21 +1768,12 @@ Eterm copy_struct(Eterm obj, Uint sz, Eterm** hpp, ErlOffHeap* off_heap)
  *
  * NOTE: Assumes that term is a tuple (ptr is an untagged tuple ptr).
  */
-#if HALFWORD_HEAP
-Eterm copy_shallow_rel(Eterm* ptr, Uint sz, Eterm** hpp, ErlOffHeap* off_heap,
-		       Eterm* src_base)
-#else
 Eterm copy_shallow(Eterm* ptr, Uint sz, Eterm** hpp, ErlOffHeap* off_heap)
-#endif
 {
     Eterm* tp = ptr;
     Eterm* hp = *hpp;
     const Eterm res = make_tuple(hp);
-#if HALFWORD_HEAP
-    const Sint offs = COMPRESS_POINTER(hp - (tp - src_base));
-#else
     const Sint offs = (hp - tp) * sizeof(Eterm);
-#endif
 
     while (sz--) {
 	Eterm val = *tp++;
@@ -652,17 +1848,24 @@ Eterm copy_shallow(Eterm* ptr, Uint sz, Eterm** hpp, ErlOffHeap* off_heap)
  * move markers.
  * Typically used to copy a multi-fragmented message (from NIF).
  */
-void move_multi_frags(Eterm** hpp, ErlOffHeap* off_heap, ErlHeapFragment* first,
-		      Eterm* refs, unsigned nrefs)
+void erts_move_multi_frags(Eterm** hpp, ErlOffHeap* off_heap, ErlHeapFragment* first,
+			   Eterm* refs, unsigned nrefs, int literals)
 {
     ErlHeapFragment* bp;
     Eterm* hp_start = *hpp;
     Eterm* hp_end;
     Eterm* hp;
     unsigned i;
+    Eterm literal_tag;
+
+#ifdef TAG_LITERAL_PTR
+    literal_tag = (Eterm) literals ? TAG_LITERAL_PTR : 0;
+#else
+    literal_tag = (Eterm) 0;
+#endif
 
     for (bp=first; bp!=NULL; bp=bp->next) {
-	move_one_frag(hpp, bp, off_heap);
+	move_one_frag(hpp, bp, off_heap, literals);
     }
     hp_end = *hpp;
     for (hp=hp_start; hp<hp_end; ++hp) {
@@ -675,6 +1878,9 @@ void move_multi_frags(Eterm** hpp, ErlOffHeap* off_heap, ErlHeapFragment* first,
 	    val = *ptr;
 	    if (IS_MOVED_BOXED(val)) {
 		ASSERT(is_boxed(val));
+#ifdef TAG_LITERAL_PTR
+		val |= literal_tag;
+#endif
 		*hp = val;
 	    }
 	    break;
@@ -682,7 +1888,11 @@ void move_multi_frags(Eterm** hpp, ErlOffHeap* off_heap, ErlHeapFragment* first,
 	    ptr = list_val(gval);
 	    val = *ptr;
 	    if (IS_MOVED_CONS(val)) {
-		*hp = ptr[1];
+		val = ptr[1];
+#ifdef TAG_LITERAL_PTR
+		val |= literal_tag;
+#endif
+		*hp = val;
 	    }
 	    break;
 	case TAG_PRIMARY_HEADER:
@@ -693,12 +1903,12 @@ void move_multi_frags(Eterm** hpp, ErlOffHeap* off_heap, ErlHeapFragment* first,
 	}
     }
     for (i=0; i<nrefs; ++i) {
-	refs[i] = follow_moved(refs[i]);
+	refs[i] = follow_moved(refs[i], literal_tag);
     }
 }
 
 static void
-move_one_frag(Eterm** hpp, ErlHeapFragment* frag, ErlOffHeap* off_heap)
+move_one_frag(Eterm** hpp, ErlHeapFragment* frag, ErlOffHeap* off_heap, int literals)
 {
     Eterm* ptr = frag->mem;
     Eterm* end = ptr + frag->used_size;
@@ -735,4 +1945,3 @@ move_one_frag(Eterm** hpp, ErlHeapFragment* frag, ErlOffHeap* off_heap)
     OH_OVERHEAD(off_heap, frag->off_heap.overhead);
     frag->off_heap.first = NULL;
 }
-
diff --git a/erts/emulator/beam/dist.c b/erts/emulator/beam/dist.c
index 23897a49ae..e31ef29562 100644
--- a/erts/emulator/beam/dist.c
+++ b/erts/emulator/beam/dist.c
@@ -45,6 +45,8 @@
 #include "erl_thr_progress.h"
 #include "dtrace-wrapper.h"
 
+#define DIST_CTL_DEFAULT_SIZE 64
+
 /* Turn this on to get printouts of all distribution messages
  * which go on the line
  */
@@ -66,9 +68,13 @@ static void bw(byte *buf, ErlDrvSizeT sz)
 static void
 dist_msg_dbg(ErtsDistExternal *edep, char *what, byte *buf, int sz)
 {
+    ErtsHeapFactory factory;
+    DeclareTmpHeapNoproc(ctl_default,DIST_CTL_DEFAULT_SIZE);
+    Eterm* ctl = ctl_default;
     byte *extp = edep->extp;
     Eterm msg;
-    Sint size = erts_decode_dist_ext_size(edep);
+    Sint ctl_len;
+    Sint size = ctl_len = erts_decode_dist_ext_size(edep);
     if (size < 0) {
 	erts_fprintf(stderr,
 		     "DIST MSG DEBUG: erts_decode_dist_ext_size(%s) failed:\n",
@@ -76,10 +82,9 @@ dist_msg_dbg(ErtsDistExternal *edep, char *what, byte *buf, int sz)
 	bw(buf, sz);
     }
     else {
-	Eterm *hp;
 	ErlHeapFragment *mbuf = new_message_buffer(size);
-	hp = mbuf->mem;
-	msg = erts_decode_dist_ext(&hp, &mbuf->off_heap, edep);
+	erts_factory_static_init(&factory, ctl, ctl_len, &mbuf->off_heap);
+	msg = erts_decode_dist_ext(&factory, edep);
 	if (is_value(msg))
 	    erts_fprintf(stderr, "    %s: %T\n", what, msg);
 	else {
@@ -373,10 +378,11 @@ static void doit_node_link_net_exits(ErtsLink *lnk, void *vnecp)
     ASSERT(lnk->type == LINK_NODE);
     if (is_internal_pid(lnk->pid)) {
 	ErtsProcLocks rp_locks = ERTS_PROC_LOCK_LINK;
-	rp = erts_pid2proc(NULL, 0, lnk->pid, rp_locks);
-	if (!rp) {
+	ErlOffHeap *ohp;
+	rp = erts_proc_lookup(lnk->pid);
+	if (!rp)
 	    goto done;
-	}
+	erts_smp_proc_lock(rp, rp_locks);
 	rlnk = erts_remove_link(&ERTS_P_LINKS(rp), name);
 	if (rlnk != NULL) {
 	    ASSERT(is_atom(rlnk->pid) && (rlnk->type == LINK_NODE));
@@ -384,12 +390,14 @@ static void doit_node_link_net_exits(ErtsLink *lnk, void *vnecp)
 	}
 	n = ERTS_LINK_REFC(lnk);
 	for (i = 0; i < n; ++i) {
-	    ErlHeapFragment* bp;
-	    ErlOffHeap *ohp;
 	    Eterm tup;
-	    Eterm *hp = erts_alloc_message_heap(3,&bp,&ohp,rp,&rp_locks);
+	    Eterm *hp;
+	    ErtsMessage *msgp;
+
+	    msgp = erts_alloc_message_heap(rp, &rp_locks,
+					   3, &hp, &ohp);
 	    tup = TUPLE2(hp, am_nodedown, name);
-	    erts_queue_message(rp, &rp_locks, bp, tup, NIL);
+	    erts_queue_message(rp, &rp_locks, msgp, tup, NIL);
 	}
 	erts_smp_proc_unlock(rp, rp_locks);
     }
@@ -732,19 +740,11 @@ Eterm erts_dsend_export_trap_context(Process* p, ErtsSendContext* ctx)
     Binary* ctx_bin = erts_create_magic_binary(sizeof(struct exported_ctx),
 					       erts_dsend_context_dtor);
     struct exported_ctx* dst = ERTS_MAGIC_BIN_DATA(ctx_bin);
-    Uint ctl_size = !HALFWORD_HEAP ? 0 : (arityval(ctx->ctl_heap[0]) + 1);
-    Eterm* hp = HAlloc(p, ctl_size + PROC_BIN_SIZE);
+    Eterm* hp = HAlloc(p, PROC_BIN_SIZE);
 
     sys_memcpy(&dst->ctx, ctx, sizeof(ErtsSendContext));
     ASSERT(ctx->dss.ctl == make_tuple(ctx->ctl_heap));
-#if !HALFWORD_HEAP
     dst->ctx.dss.ctl = make_tuple(dst->ctx.ctl_heap);
-#else
-    /* Must put control tuple in low mem */
-    sys_memcpy(hp, ctx->ctl_heap,  ctl_size*sizeof(Eterm));
-    dst->ctx.dss.ctl = make_tuple(hp);
-    hp += ctl_size;
-#endif
     if (ctx->dss.acmp) {
 	sys_memcpy(&dst->acm, ctx->dss.acmp, sizeof(ErtsAtomCacheMap));
 	dst->ctx.dss.acmp = &dst->acm;
@@ -881,11 +881,7 @@ erts_dsig_send_msg(Eterm remote, Eterm message, ErtsSendContext* ctx)
     DTRACE_CHARBUF(receiver_name, 64);
 #endif
 
-    if (SEQ_TRACE_TOKEN(sender) != NIL
-#ifdef USE_VM_PROBES
-	&& SEQ_TRACE_TOKEN(sender) != am_have_dt_utag 
-#endif
-	) {
+    if (have_seqtrace(SEQ_TRACE_TOKEN(sender))) {
 	seq_trace_update_send(sender);
 	token = SEQ_TRACE_TOKEN(sender);
 	seq_trace_output(token, message, SEQ_TRACE_SEND, remote, sender);
@@ -900,7 +896,7 @@ erts_dsig_send_msg(Eterm remote, Eterm message, ErtsSendContext* ctx)
         erts_snprintf(receiver_name, sizeof(DTRACE_CHARBUF_NAME(receiver_name)),
                       "%T", remote);
         msize = size_object(message);
-        if (token != NIL && token != am_have_dt_utag) {
+        if (have_seqtrace(token)) {
             tok_label = signed_val(SEQ_TRACE_T_LABEL(token));
             tok_lastcnt = signed_val(SEQ_TRACE_T_LASTCNT(token));
             tok_serial = signed_val(SEQ_TRACE_T_SERIAL(token));
@@ -942,11 +938,7 @@ erts_dsig_send_reg_msg(Eterm remote_name, Eterm message,
     DTRACE_CHARBUF(receiver_name, 128);
 #endif
 
-    if (SEQ_TRACE_TOKEN(sender) != NIL
-#ifdef USE_VM_PROBES
-	&& SEQ_TRACE_TOKEN(sender) != am_have_dt_utag 
-#endif
-	) {
+    if (have_seqtrace(SEQ_TRACE_TOKEN(sender))) {
 	seq_trace_update_send(sender);
 	token = SEQ_TRACE_TOKEN(sender);
 	seq_trace_output(token, message, SEQ_TRACE_SEND, remote_name, sender);
@@ -961,7 +953,7 @@ erts_dsig_send_reg_msg(Eterm remote_name, Eterm message,
         erts_snprintf(receiver_name, sizeof(DTRACE_CHARBUF_NAME(receiver_name)),
                       "{%T,%s}", remote_name, node_name);
         msize = size_object(message);
-        if (token != NIL && token != am_have_dt_utag) {
+        if (have_seqtrace(token)) {
             tok_label = signed_val(SEQ_TRACE_T_LABEL(token));
             tok_lastcnt = signed_val(SEQ_TRACE_T_LASTCNT(token));
             tok_serial = signed_val(SEQ_TRACE_T_SERIAL(token));
@@ -1006,11 +998,7 @@ erts_dsig_send_exit_tt(ErtsDSigData *dsdp, Eterm local, Eterm remote,
 #endif
 
     UseTmpHeapNoproc(6);
-    if (token != NIL 
-#ifdef USE_VM_PROBES
-	&& token != am_have_dt_utag
-#endif
-	) {	
+    if (have_seqtrace(token)) {
 	seq_trace_update_send(dsdp->proc);
 	seq_trace_output_exit(token, reason, SEQ_TRACE_SEND, remote, local);
 	ctl = TUPLE5(&ctl_heap[0],
@@ -1029,7 +1017,7 @@ erts_dsig_send_exit_tt(ErtsDSigData *dsdp, Eterm local, Eterm remote,
                       "{%T,%s}", remote, node_name);
         erts_snprintf(reason_str, sizeof(DTRACE_CHARBUF_NAME(reason_str)),
                       "%T", reason);
-        if (token != NIL && token != am_have_dt_utag) {
+        if (have_seqtrace(token)) {
             tok_label = signed_val(SEQ_TRACE_T_LABEL(token));
             tok_lastcnt = signed_val(SEQ_TRACE_T_LASTCNT(token));
             tok_serial = signed_val(SEQ_TRACE_T_SERIAL(token));
@@ -1136,7 +1124,6 @@ int erts_net_message(Port *prt,
 		     byte *buf,
 		     ErlDrvSizeT len)
 {
-#define DIST_CTL_DEFAULT_SIZE 64
     ErtsDistExternal ede;
     byte *t;
     Sint ctl_len;
@@ -1149,7 +1136,6 @@ int erts_net_message(Port *prt,
     Process* rp;
     DeclareTmpHeapNoproc(ctl_default,DIST_CTL_DEFAULT_SIZE);
     Eterm* ctl = ctl_default;
-    ErlOffHeap off_heap;
     ErtsHeapFactory factory;
     Eterm* hp;
     Sint type;
@@ -1164,9 +1150,6 @@ int erts_net_message(Port *prt,
 #endif
 
     UseTmpHeapNoproc(DIST_CTL_DEFAULT_SIZE);
-    /* Thanks to Luke Gorrie */
-    off_heap.first = NULL;
-    off_heap.overhead = 0;
 
     ERTS_SMP_CHK_NO_PROC_LOCKS;
 
@@ -1227,15 +1210,15 @@ int erts_net_message(Port *prt,
     }
     hp = ctl;
 
-    erts_factory_static_init(&factory, ctl, ctl_len, &off_heap);
+    erts_factory_tmp_init(&factory, ctl, ctl_len, ERTS_ALC_T_DCTRL_BUF);
     arg = erts_decode_dist_ext(&factory, &ede);
     if (is_non_value(arg)) {
 #ifdef ERTS_DIST_MSG_DBG
-	erts_fprintf(stderr, "DIST MSG DEBUG: erts_dist_ext_size(CTL) failed:\n");
+	erts_fprintf(stderr, "DIST MSG DEBUG: erts_decode_dist_ext(CTL) failed:\n");
 	bw(buf, orig_len);
 #endif
 	PURIFY_MSG("data error");
-	goto data_error;
+	goto decode_error;
     }
     ctl_len = t - buf;
 
@@ -1466,7 +1449,7 @@ int erts_net_message(Port *prt,
 		ErlOffHeap *ohp;
 		ASSERT(xsize);
 		heap_frag = erts_dist_ext_trailer(ede_copy);
-		ERTS_INIT_HEAP_FRAG(heap_frag, token_size);
+		ERTS_INIT_HEAP_FRAG(heap_frag, token_size, token_size);
 		hp = heap_frag->mem;
 		ohp = &heap_frag->off_heap;
 		token = tuple[5];
@@ -1515,7 +1498,7 @@ int erts_net_message(Port *prt,
 		ErlOffHeap *ohp;
 		ASSERT(xsize);
 		heap_frag = erts_dist_ext_trailer(ede_copy);
-		ERTS_INIT_HEAP_FRAG(heap_frag, token_size);
+		ERTS_INIT_HEAP_FRAG(heap_frag, token_size, token_size);
 		hp = heap_frag->mem;
 		ohp = &heap_frag->off_heap;
 		token = tuple[4];
@@ -1715,7 +1698,7 @@ int erts_net_message(Port *prt,
 	goto invalid_message;
     }
 
-    erts_cleanup_offheap(&off_heap);
+    erts_factory_close(&factory);
     if (ctl != ctl_default) {
 	erts_free(ERTS_ALC_T_DCTRL_BUF, (void *) ctl);
     }
@@ -1728,12 +1711,13 @@ int erts_net_message(Port *prt,
 	erts_dsprintf(dsbufp, "Invalid distribution message: %.200T", arg);
 	erts_send_error_to_logger_nogl(dsbufp);
     }
- data_error:
+decode_error:
     PURIFY_MSG("data error");
-    erts_cleanup_offheap(&off_heap);
+    erts_factory_close(&factory);
     if (ctl != ctl_default) {
 	erts_free(ERTS_ALC_T_DCTRL_BUF, (void *) ctl);
     }
+data_error:
     UnUseTmpHeapNoproc(DIST_CTL_DEFAULT_SIZE);
     erts_deliver_port_exit(prt, dep->cid, am_killed, 0);
     ERTS_SMP_CHK_NO_PROC_LOCKS;
@@ -1790,8 +1774,8 @@ erts_dsig_send(ErtsDSigData *dsdp, struct erts_dsig_send_context* ctx)
 
     #ifdef ERTS_DIST_MSG_DBG
 	    erts_fprintf(stderr, ">>%s CTL: %T\n", ctx->pass_through_size ? "P" : " ", ctx->ctl);
-	    if (is_value(msg))
-		erts_fprintf(stderr, "    MSG: %T\n", msg);
+        if (is_value(ctx->msg))
+            erts_fprintf(stderr, "    MSG: %T\n", ctx->msg);
     #endif
 
 	    ctx->data_size = ctx->pass_through_size;
@@ -2061,9 +2045,9 @@ dist_port_commandv(Port *prt, ErtsDistOutputBuf *obuf)
 }
 
 
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
 #define ERTS_PORT_REDS_MASK__ 0x003fffffffffffffL
-#elif defined(ARCH_32) || HALFWORD_HEAP
+#elif defined(ARCH_32)
 #define ERTS_PORT_REDS_MASK__ 0x003fffff
 #else
 #  error "Ohh come on ... !?!"
@@ -2572,7 +2556,9 @@ int distribution_info(int to, void *arg)	/* Called by break handler */
     }
 
     for (dep = erts_not_connected_dist_entries; dep; dep = dep->next) {
-	info_dist_entry(to, arg, dep, 0, 0);
+        if (dep != erts_this_dist_entry) {
+            info_dist_entry(to, arg, dep, 0, 0);
+        }
     }
 
     return(0);
@@ -2650,13 +2636,8 @@ BIF_RETTYPE setnode_2(BIF_ALIST_2)
     if (!net_kernel)
 	goto error;
 
-    /* By setting dist_entry==erts_this_dist_entry and DISTRIBUTION on
-       net_kernel do_net_exist will be called when net_kernel
-       is terminated !! */
-    (void) ERTS_PROC_SET_DIST_ENTRY(net_kernel,
-				    ERTS_PROC_LOCK_MAIN,
-				    erts_this_dist_entry);
-    erts_refc_inc(&erts_this_dist_entry->refc, 2);
+    /* By setting F_DISTRIBUTION on net_kernel,
+     * do_net_exist will be called when net_kernel is terminated !! */
     net_kernel->flags |= F_DISTRIBUTION;
 
     if (net_kernel != BIF_P)
@@ -3017,11 +2998,11 @@ BIF_RETTYPE nodes_1(BIF_ALIST_1)
 
     erts_smp_rwmtx_rlock(&erts_dist_table_rwmtx);
 
-    ASSERT(erts_no_of_not_connected_dist_entries >= 0);
+    ASSERT(erts_no_of_not_connected_dist_entries > 0);
     ASSERT(erts_no_of_hidden_dist_entries >= 0);
     ASSERT(erts_no_of_visible_dist_entries >= 0);
     if(not_connected)
-      length += erts_no_of_not_connected_dist_entries;
+      length += (erts_no_of_not_connected_dist_entries - 1);
     if(hidden)
       length += erts_no_of_hidden_dist_entries;
     if(visible)
@@ -3043,8 +3024,10 @@ BIF_RETTYPE nodes_1(BIF_ALIST_1)
 #endif
     if(not_connected)
       for(dep = erts_not_connected_dist_entries; dep; dep = dep->next) {
-	result = CONS(hp, dep->sysname, result);
-	hp += 2;
+          if (dep != erts_this_dist_entry) {
+            result = CONS(hp, dep->sysname, result);
+            hp += 2;
+          }
       }
     if(hidden)
       for(dep = erts_hidden_dist_entries; dep; dep = dep->next) {
@@ -3275,11 +3258,16 @@ send_nodes_mon_msg(Process *rp,
 		   Uint sz)
 {
     Eterm msg;
-    ErlHeapFragment* bp;
+    Eterm *hp;
+    ErtsMessage *mp;
     ErlOffHeap *ohp;
-    Eterm *hp = erts_alloc_message_heap(sz, &bp, &ohp, rp, rp_locksp);
 #ifdef DEBUG
-    Eterm *hend = hp + sz;
+    Eterm *hend;
+#endif
+
+    mp = erts_alloc_message_heap(rp, rp_locksp, sz, &hp, &ohp);
+#ifdef DEBUG
+    hend = hp + sz;
 #endif
 
     if (!nmp->opts) {
@@ -3325,7 +3313,7 @@ send_nodes_mon_msg(Process *rp,
     }
 
     ASSERT(hend == hp);
-    erts_queue_message(rp, rp_locksp, bp, msg, NIL);
+    erts_queue_message(rp, rp_locksp, mp, msg, NIL);
 }
 
 static void
diff --git a/erts/emulator/beam/erl_alloc.c b/erts/emulator/beam/erl_alloc.c
index 55c164bf11..5544712e8d 100644
--- a/erts/emulator/beam/erl_alloc.c
+++ b/erts/emulator/beam/erl_alloc.c
@@ -123,10 +123,6 @@ typedef union {
 
 static ErtsAllocatorState_t std_alloc_state;
 static ErtsAllocatorState_t ll_alloc_state;
-#if HALFWORD_HEAP
-static ErtsAllocatorState_t std_low_alloc_state;
-static ErtsAllocatorState_t ll_low_alloc_state;
-#endif
 static ErtsAllocatorState_t sl_alloc_state;
 static ErtsAllocatorState_t temp_alloc_state;
 static ErtsAllocatorState_t eheap_alloc_state;
@@ -134,6 +130,8 @@ static ErtsAllocatorState_t binary_alloc_state;
 static ErtsAllocatorState_t ets_alloc_state;
 static ErtsAllocatorState_t driver_alloc_state;
 static ErtsAllocatorState_t fix_alloc_state;
+static ErtsAllocatorState_t literal_alloc_state;
+static ErtsAllocatorState_t test_alloc_state;
 
 typedef struct {
     erts_smp_atomic32_t refc;
@@ -150,24 +148,10 @@ typedef struct {
 #define ERTS_ALC_INFO_A_MSEG_ALLOC (ERTS_ALC_A_MAX + 2)
 #define ERTS_ALC_INFO_A_MAX ERTS_ALC_INFO_A_MSEG_ALLOC
 
-#if !HALFWORD_HEAP
 ERTS_SCHED_PREF_QUICK_ALLOC_IMPL(aireq,
-				 ErtsAllocInfoReq,
-				 5,
-				 ERTS_ALC_T_AINFO_REQ)
-#else
-static ERTS_INLINE ErtsAllocInfoReq *
-aireq_alloc(void)
-{
-    return erts_alloc(ERTS_ALC_T_AINFO_REQ, sizeof(ErtsAllocInfoReq));
-}
-
-static ERTS_INLINE void
-aireq_free(ErtsAllocInfoReq *ptr)
-{
-    erts_free(ERTS_ALC_T_AINFO_REQ, ptr);
-}
-#endif
+                                 ErtsAllocInfoReq,
+                                 5,
+                                 ERTS_ALC_T_AINFO_REQ)
 
 ErtsAlcType_t erts_fix_core_allocator_ix;
 
@@ -229,10 +213,8 @@ typedef struct {
     struct au_init ets_alloc;
     struct au_init driver_alloc;
     struct au_init fix_alloc;
-#if HALFWORD_HEAP
-    struct au_init std_low_alloc;
-    struct au_init ll_low_alloc;
-#endif
+    struct au_init literal_alloc;
+    struct au_init test_alloc;
 } erts_alc_hndl_args_init_t;
 
 #define ERTS_AU_INIT__ {0, 0, 1, GOODFIT, DEFAULT_ALLCTR_INIT, {1,1,1,1}}
@@ -259,10 +241,6 @@ set_default_sl_alloc_opts(struct au_init *ip)
 #endif
     ip->init.util.ts 		= ERTS_ALC_MTA_SHORT_LIVED;
     ip->init.util.rsbcst	= 80;
-#if HALFWORD_HEAP
-    ip->init.util.force         = 1;
-    ip->init.util.low_mem       = 1;
-#endif
     ip->init.util.acul		= ERTS_ALC_DEFAULT_ACUL;
 }
 
@@ -298,9 +276,9 @@ set_default_ll_alloc_opts(struct au_init *ip)
     ip->init.util.name_prefix	= "ll_";
     ip->init.util.alloc_no	= ERTS_ALC_A_LONG_LIVED;
 #ifndef SMALL_MEMORY
-    ip->init.util.mmbcs 	= 2*1024*1024 - 40; /* Main carrier size */
+    ip->init.util.mmbcs 	= 2*1024*1024; /* Main carrier size */
 #else
-    ip->init.util.mmbcs 	= 1*1024*1024 - 40; /* Main carrier size */
+    ip->init.util.mmbcs 	= 1*1024*1024; /* Main carrier size */
 #endif
     ip->init.util.ts 		= ERTS_ALC_MTA_LONG_LIVED;
     ip->init.util.asbcst	= 0;
@@ -311,6 +289,51 @@ set_default_ll_alloc_opts(struct au_init *ip)
 }
 
 static void
+set_default_literal_alloc_opts(struct au_init *ip)
+{
+    SET_DEFAULT_ALLOC_OPTS(ip);
+    ip->enable			= 1;
+    ip->thr_spec		= 0;
+    ip->atype			= BESTFIT;
+    ip->init.bf.ao		= 1;
+    ip->init.util.ramv		= 0;
+    ip->init.util.mmsbc		= 0;
+    ip->init.util.sbct		= ~((UWord) 0);
+    ip->init.util.name_prefix	= "literal_";
+    ip->init.util.alloc_no	= ERTS_ALC_A_LITERAL;
+#ifndef SMALL_MEMORY
+    ip->init.util.mmbcs 	= 2*1024*1024; /* Main carrier size */
+#else
+    ip->init.util.mmbcs 	= 1*1024*1024; /* Main carrier size */
+#endif
+    ip->init.util.ts 		= ERTS_ALC_MTA_LITERAL;
+    ip->init.util.asbcst	= 0;
+    ip->init.util.rsbcst	= 0;
+    ip->init.util.rsbcmt	= 0;
+    ip->init.util.rmbcmt	= 0;
+    ip->init.util.acul		= 0;
+
+#if defined(ARCH_32)
+# if HAVE_ERTS_MSEG
+    ip->init.util.mseg_alloc   = &erts_alcu_literal_32_mseg_alloc;
+    ip->init.util.mseg_realloc = &erts_alcu_literal_32_mseg_realloc;
+    ip->init.util.mseg_dealloc = &erts_alcu_literal_32_mseg_dealloc;
+# endif
+    ip->init.util.sys_alloc    = &erts_alcu_literal_32_sys_alloc;
+    ip->init.util.sys_realloc  = &erts_alcu_literal_32_sys_realloc;
+    ip->init.util.sys_dealloc  = &erts_alcu_literal_32_sys_dealloc;
+#elif defined(ARCH_64)
+# ifdef ERTS_HAVE_OS_PHYSICAL_MEMORY_RESERVATION
+    ip->init.util.mseg_alloc    = &erts_alcu_literal_64_mseg_alloc;
+    ip->init.util.mseg_realloc  = &erts_alcu_literal_64_mseg_realloc;
+    ip->init.util.mseg_dealloc  = &erts_alcu_literal_64_mseg_dealloc;
+# endif
+#else
+# error Unknown architecture
+#endif
+}
+
+static void
 set_default_temp_alloc_opts(struct au_init *ip)
 {
     SET_DEFAULT_ALLOC_OPTS(ip);
@@ -328,10 +351,6 @@ set_default_temp_alloc_opts(struct au_init *ip)
     ip->init.util.ts 		= ERTS_ALC_MTA_TEMPORARY;
     ip->init.util.rsbcst	= 90;
     ip->init.util.rmbcmt	= 100;
-#if HALFWORD_HEAP
-    ip->init.util.force         = 1;
-    ip->init.util.low_mem       = 1;
-#endif
 }
 
 static void
@@ -350,10 +369,6 @@ set_default_eheap_alloc_opts(struct au_init *ip)
 #endif
     ip->init.util.ts 		= ERTS_ALC_MTA_EHEAP;
     ip->init.util.rsbcst	= 50;
-#if HALFWORD_HEAP
-    ip->init.util.force         = 1;
-    ip->init.util.low_mem       = 1;
-#endif
     ip->init.util.acul		= ERTS_ALC_DEFAULT_ACUL_EHEAP_ALLOC;
 }
 
@@ -432,6 +447,33 @@ set_default_fix_alloc_opts(struct au_init *ip,
     ip->init.util.acul		= ERTS_ALC_DEFAULT_ACUL;
 }
 
+static void
+set_default_test_alloc_opts(struct au_init *ip)
+{
+    SET_DEFAULT_ALLOC_OPTS(ip);
+    ip->enable			= 0; /* Disabled by default */
+    ip->thr_spec		= -1 * erts_no_schedulers;
+    ip->atype			= AOFIRSTFIT;
+    ip->init.aoff.flavor        = AOFF_BF;
+    ip->init.util.name_prefix	= "test_";
+    ip->init.util.alloc_no	= ERTS_ALC_A_TEST;
+    ip->init.util.mmbcs 	= 0; /* Main carrier size */
+    ip->init.util.ts 		= ERTS_ALC_MTA_TEST;
+    ip->init.util.acul		= ERTS_ALC_DEFAULT_ACUL;
+
+    /* Use a constant minimal MBC size */
+#if ERTS_SA_MB_CARRIERS
+    ip->init.util.smbcs = ERTS_SACRR_UNIT_SZ;
+    ip->init.util.lmbcs = ERTS_SACRR_UNIT_SZ;
+    ip->init.util.sbct  = ERTS_SACRR_UNIT_SZ;
+#else
+    ip->init.util.smbcs = 1 << 12;
+    ip->init.util.lmbcs = 1 << 12;
+    ip->init.util.sbct  = 1 << 12;
+#endif
+}
+
+
 #ifdef ERTS_SMP
 
 static void
@@ -560,18 +602,16 @@ erts_alloc_init(int *argc, char **argv, ErtsAllocInitOpts *eaiop)
 
     fix_type_sizes[ERTS_ALC_FIX_TYPE_IX(ERTS_ALC_T_PROC)]
 	= sizeof(Process);
-#if !HALFWORD_HEAP
     fix_type_sizes[ERTS_ALC_FIX_TYPE_IX(ERTS_ALC_T_MONITOR_SH)]
 	= ERTS_MONITOR_SH_SIZE * sizeof(Uint);
     fix_type_sizes[ERTS_ALC_FIX_TYPE_IX(ERTS_ALC_T_NLINK_SH)]
 	= ERTS_LINK_SH_SIZE * sizeof(Uint);
-#endif
     fix_type_sizes[ERTS_ALC_FIX_TYPE_IX(ERTS_ALC_T_DRV_EV_D_STATE)]
 	= sizeof(ErtsDrvEventDataState);
     fix_type_sizes[ERTS_ALC_FIX_TYPE_IX(ERTS_ALC_T_DRV_SEL_D_STATE)]
 	= sizeof(ErtsDrvSelectDataState);
     fix_type_sizes[ERTS_ALC_FIX_TYPE_IX(ERTS_ALC_T_MSG_REF)]
-	= sizeof(ErlMessage);
+	= sizeof(ErtsMessageRef);
 #ifdef ERTS_SMP
     fix_type_sizes[ERTS_ALC_FIX_TYPE_IX(ERTS_ALC_T_THR_Q_EL_SL)]
 	= sizeof(ErtsThrQElement_t);
@@ -613,6 +653,8 @@ erts_alloc_init(int *argc, char **argv, ErtsAllocInitOpts *eaiop)
     set_default_driver_alloc_opts(&init.driver_alloc);
     set_default_fix_alloc_opts(&init.fix_alloc,
 			       fix_type_sizes);
+    set_default_literal_alloc_opts(&init.literal_alloc);
+    set_default_test_alloc_opts(&init.test_alloc);
 
     if (argc && argv)
 	handle_args(argc, argv, &init);
@@ -640,6 +682,7 @@ erts_alloc_init(int *argc, char **argv, ErtsAllocInitOpts *eaiop)
     init.ets_alloc.thr_spec = 0;
     init.driver_alloc.thr_spec = 0;
     init.fix_alloc.thr_spec = 0;
+    init.literal_alloc.thr_spec = 0;
 #endif
 
     /* Make adjustments for carrier migration support */
@@ -652,6 +695,7 @@ erts_alloc_init(int *argc, char **argv, ErtsAllocInitOpts *eaiop)
     adjust_carrier_migration_support(&init.ets_alloc);
     adjust_carrier_migration_support(&init.driver_alloc);
     adjust_carrier_migration_support(&init.fix_alloc);
+    adjust_carrier_migration_support(&init.literal_alloc);
 
     if (init.erts_alloc_config) {
 	/* Adjust flags that erts_alloc_config won't like */
@@ -666,6 +710,7 @@ erts_alloc_init(int *argc, char **argv, ErtsAllocInitOpts *eaiop)
 	init.ets_alloc.thr_spec = 0;
 	init.driver_alloc.thr_spec = 0;
 	init.fix_alloc.thr_spec = 0;
+        init.literal_alloc.thr_spec = 0;
 
 	/* No carrier migration */
 	init.temp_alloc.init.util.acul = 0;
@@ -677,6 +722,7 @@ erts_alloc_init(int *argc, char **argv, ErtsAllocInitOpts *eaiop)
 	init.ets_alloc.init.util.acul = 0;
 	init.driver_alloc.init.util.acul = 0;
 	init.fix_alloc.init.util.acul = 0;
+        init.literal_alloc.init.util.acul = 0;
     }
 
 #ifdef ERTS_SMP
@@ -693,6 +739,7 @@ erts_alloc_init(int *argc, char **argv, ErtsAllocInitOpts *eaiop)
     adjust_tpref(&init.ets_alloc, erts_no_schedulers);
     adjust_tpref(&init.driver_alloc, erts_no_schedulers);
     adjust_tpref(&init.fix_alloc, erts_no_schedulers);
+    adjust_tpref(&init.literal_alloc, erts_no_schedulers);
 
 #else
     /* No thread specific if not smp */
@@ -711,6 +758,7 @@ erts_alloc_init(int *argc, char **argv, ErtsAllocInitOpts *eaiop)
     refuse_af_strategy(&init.ets_alloc);
     refuse_af_strategy(&init.driver_alloc);
     refuse_af_strategy(&init.fix_alloc);
+    refuse_af_strategy(&init.literal_alloc);
 
 #ifdef ERTS_SMP 
     if (!init.temp_alloc.thr_spec)
@@ -721,6 +769,9 @@ erts_alloc_init(int *argc, char **argv, ErtsAllocInitOpts *eaiop)
 #if HAVE_ERTS_MSEG
     init.mseg.nos = erts_no_schedulers;
     erts_mseg_init(&init.mseg);
+# if defined(ARCH_64) && defined(ERTS_HAVE_OS_PHYSICAL_MEMORY_RESERVATION)
+    erts_mmap_init(&erts_literal_mmapper, &init.mseg.literal_mmap);
+# endif
 #endif
 
     erts_alcu_init(&init.alloc_util);
@@ -745,24 +796,6 @@ erts_alloc_init(int *argc, char **argv, ErtsAllocInitOpts *eaiop)
     erts_allctrs[ERTS_ALC_A_SYSTEM].free		= erts_sys_free;
     erts_allctrs_info[ERTS_ALC_A_SYSTEM].enabled	= 1;
 
-#if HALFWORD_HEAP
-    /* Init low memory variants by cloning */
-    init.std_low_alloc = init.std_alloc;
-    init.std_low_alloc.init.util.name_prefix	= "std_low_";
-    init.std_low_alloc.init.util.alloc_no = ERTS_ALC_A_STANDARD_LOW;
-    init.std_low_alloc.init.util.force = 1;
-    init.std_low_alloc.init.util.low_mem = 1;
-
-    init.ll_low_alloc = init.ll_alloc;
-    init.ll_low_alloc.init.util.name_prefix	= "ll_low_";
-    init.ll_low_alloc.init.util.alloc_no = ERTS_ALC_A_LONG_LIVED_LOW;
-    init.ll_low_alloc.init.util.force = 1;
-    init.ll_low_alloc.init.util.low_mem = 1;
-
-    set_au_allocator(ERTS_ALC_A_STANDARD_LOW, &init.std_low_alloc, ncpu);
-    set_au_allocator(ERTS_ALC_A_LONG_LIVED_LOW, &init.ll_low_alloc, ncpu);
-#endif /* HALFWORD */
-
     set_au_allocator(ERTS_ALC_A_TEMPORARY, &init.temp_alloc, ncpu);
     set_au_allocator(ERTS_ALC_A_SHORT_LIVED, &init.sl_alloc, ncpu);
     set_au_allocator(ERTS_ALC_A_STANDARD, &init.std_alloc, ncpu);
@@ -772,6 +805,8 @@ erts_alloc_init(int *argc, char **argv, ErtsAllocInitOpts *eaiop)
     set_au_allocator(ERTS_ALC_A_ETS, &init.ets_alloc, ncpu);
     set_au_allocator(ERTS_ALC_A_DRIVER, &init.driver_alloc, ncpu);
     set_au_allocator(ERTS_ALC_A_FIXED_SIZE, &init.fix_alloc, ncpu);
+    set_au_allocator(ERTS_ALC_A_LITERAL, &init.literal_alloc, ncpu);
+    set_au_allocator(ERTS_ALC_A_TEST, &init.test_alloc, ncpu);
 
     for (i = ERTS_ALC_A_MIN; i <= ERTS_ALC_A_MAX; i++) {
 	if (!erts_allctrs[i].alloc)
@@ -805,14 +840,6 @@ erts_alloc_init(int *argc, char **argv, ErtsAllocInitOpts *eaiop)
     start_au_allocator(ERTS_ALC_A_LONG_LIVED,
 		       &init.ll_alloc,
 		       &ll_alloc_state);
-#if HALFWORD_HEAP
-    start_au_allocator(ERTS_ALC_A_LONG_LIVED_LOW,
-		       &init.ll_low_alloc,
-		       &ll_low_alloc_state);
-    start_au_allocator(ERTS_ALC_A_STANDARD_LOW,
-		       &init.std_low_alloc,
-		       &std_low_alloc_state);
-#endif
     start_au_allocator(ERTS_ALC_A_EHEAP,
 		       &init.eheap_alloc,
 		       &eheap_alloc_state);
@@ -832,13 +859,18 @@ erts_alloc_init(int *argc, char **argv, ErtsAllocInitOpts *eaiop)
     start_au_allocator(ERTS_ALC_A_FIXED_SIZE,
 		       &init.fix_alloc,
 		       &fix_alloc_state);
+    start_au_allocator(ERTS_ALC_A_LITERAL,
+                       &init.literal_alloc,
+                       &literal_alloc_state);
+
+    start_au_allocator(ERTS_ALC_A_TEST,
+		       &init.test_alloc,
+		       &test_alloc_state);
 
     erts_mtrace_install_wrapper_functions();
     extra_block_size += erts_instr_init(init.instr.stat, init.instr.map);
 
-#if !HALFWORD_HEAP
     init_aireq_alloc();
-#endif
 
 #ifdef DEBUG
     extra_block_size += install_debug_functions();
@@ -1016,7 +1048,7 @@ start_au_allocator(ErtsAlcType_t alctr_n,
     }
 
     for (i = 0; i < size; i++) {
-	void *as;
+	Allctr_t *as;
 	atype = init->atype;
 
 	if (!init->thr_spec)
@@ -1053,22 +1085,22 @@ start_au_allocator(ErtsAlcType_t alctr_n,
 
 	switch (atype) {
 	case GOODFIT:
-	    as = (void *) erts_gfalc_start((GFAllctr_t *) as0,
+	    as = erts_gfalc_start((GFAllctr_t *) as0,
 					   &init->init.gf,
 					   &init->init.util);
 	    break;
 	case BESTFIT:
-	    as = (void *) erts_bfalc_start((BFAllctr_t *) as0,
+	    as = erts_bfalc_start((BFAllctr_t *) as0,
 					   &init->init.bf,
 					   &init->init.util);
 	    break;
 	case AFIT:
-	    as = (void *) erts_afalc_start((AFAllctr_t *) as0,
+	    as = erts_afalc_start((AFAllctr_t *) as0,
 					   &init->init.af,
 					   &init->init.util);
 	    break;
     	case AOFIRSTFIT:
-	    as = (void *) erts_aoffalc_start((AOFFAllctr_t *) as0,
+	    as = erts_aoffalc_start((AOFFAllctr_t *) as0,
 					     &init->init.aoff,
 					     &init->init.util);
 	    break;
@@ -1238,9 +1270,6 @@ get_acul_value(struct au_init *auip, char *param_end, char** argv, int* ip)
     if (sys_strcmp(value, "de") == 0) {
 	switch (auip->init.util.alloc_no) {
 	case ERTS_ALC_A_LONG_LIVED:
-#if HALFWORD_HEAP
-	case ERTS_ALC_A_LONG_LIVED_LOW:
-#endif
 	    return ERTS_ALC_DEFAULT_ENABLED_ACUL_LL_ALLOC;
 	case ERTS_ALC_A_EHEAP:
 	    return ERTS_ALC_DEFAULT_ENABLED_ACUL_EHEAP_ALLOC;
@@ -1418,6 +1447,7 @@ handle_args(int *argc, char **argv, erts_alc_hndl_args_init_t *init)
 	&init->fix_alloc,
 	&init->sl_alloc,
 	&init->temp_alloc
+	/* test_alloc not affected by +Mea??? or +Mu???  */
     };
     int aui_sz = (int) sizeof(aui)/sizeof(aui[0]);
     char *arg;
@@ -1473,25 +1503,25 @@ handle_args(int *argc, char **argv, erts_alc_hndl_args_init_t *init)
 		    }
 		    else if (has_prefix("scs", argv[i]+3)) {
 #if HAVE_ERTS_MSEG
-			init->mseg.mmap.scs =
+			init->mseg.dflt_mmap.scs =
 #endif
 			    get_mb_value(argv[i]+6, argv, &i);
 		    }
 		    else if (has_prefix("sco", argv[i]+3)) {
 #if HAVE_ERTS_MSEG
-			init->mseg.mmap.sco =
+			init->mseg.dflt_mmap.sco =
 #endif
 			    get_bool_value(argv[i]+6, argv, &i);
 		    }
 		    else if (has_prefix("scrpm", argv[i]+3)) {
 #if HAVE_ERTS_MSEG
-			init->mseg.mmap.scrpm =
+			init->mseg.dflt_mmap.scrpm =
 #endif
 			    get_bool_value(argv[i]+8, argv, &i);
 		    }
 		    else if (has_prefix("scrfsd", argv[i]+3)) {
 #if HAVE_ERTS_MSEG
-			init->mseg.mmap.scrfsd =
+			init->mseg.dflt_mmap.scrfsd =
 #endif
 			    get_amount_value(argv[i]+9, argv, &i);
 		    }
@@ -1508,6 +1538,9 @@ handle_args(int *argc, char **argv, erts_alc_hndl_args_init_t *init)
 		case 'T':
 		    handle_au_arg(&init->temp_alloc, &argv[i][3], argv, &i, 0);
 		    break;
+		case 'Z':
+		    handle_au_arg(&init->test_alloc, &argv[i][3], argv, &i, 0);
+		    break;
 		case 'Y': { /* sys_alloc */
 		    if (has_prefix("tt", param+2)) {
 			/* set trim threshold */
@@ -1957,48 +1990,6 @@ alcu_size(ErtsAlcType_t ai, ErtsAlcUFixInfo_t *fi, int fisz)
     return res;
 }
 
-#if HALFWORD_HEAP
-static ERTS_INLINE int
-alcu_is_low(ErtsAlcType_t ai)
-{
-    int is_low = 0;
-    ASSERT(erts_allctrs_info[ai].enabled);
-    ASSERT(erts_allctrs_info[ai].alloc_util);
-
-    if (!erts_allctrs_info[ai].thr_spec) {
-	Allctr_t *allctr = erts_allctrs_info[ai].extra;
-	is_low = allctr->mseg_opt.low_mem;
-    }
-    else {
-	ErtsAllocatorThrSpec_t *tspec = &erts_allctr_thr_spec[ai];
-	int i;
-# ifdef DEBUG
-	int found_one = 0;
-# endif
-
-	ASSERT(tspec->enabled);
-
-	for (i = tspec->size - 1; i >= 0; i--) {
-	    Allctr_t *allctr = tspec->allctr[i];
-	    if (allctr) {
-# ifdef DEBUG
-		if (!found_one) {
-		    is_low = allctr->mseg_opt.low_mem;
-		    found_one = 1;
-		}
-		else ASSERT(is_low == allctr->mseg_opt.low_mem);
-# else
-		is_low = allctr->mseg_opt.low_mem;
-		break;
-# endif
-	    }
-	}
-	ASSERT(found_one);
-    }
-    return is_low;
-}
-#endif /* HALFWORD */
-
 static ERTS_INLINE void
 add_fix_values(UWord *ap, UWord *up, ErtsAlcUFixInfo_t *fi, ErtsAlcType_t type)
 {
@@ -2028,9 +2019,6 @@ erts_memory(int *print_to_p, void *print_to_arg, void *proc, Eterm earg)
 	int code;
 	int ets;
 	int maximum;
-#if HALFWORD_HEAP
-	int low;
-#endif
     } want = {0};
     struct {
 	UWord total;
@@ -2043,9 +2031,6 @@ erts_memory(int *print_to_p, void *print_to_arg, void *proc, Eterm earg)
 	UWord code;
 	UWord ets;
 	UWord maximum;
-#if HALFWORD_HEAP
-	UWord low;
-#endif
     } size = {0};
     Eterm atoms[sizeof(size)/sizeof(UWord)];
     UWord *uintps[sizeof(size)/sizeof(UWord)];
@@ -2104,11 +2089,6 @@ erts_memory(int *print_to_p, void *print_to_arg, void *proc, Eterm earg)
 	    atoms[length] = am_maximum;
 	    uintps[length++] = &size.maximum;
 	}
-#if HALFWORD_HEAP
-	want.low = 1;
-	atoms[length] = am_low;
-	uintps[length++] = &size.low;
-#endif
     }
     else {
 	DeclareTmpHeapNoproc(tmp_heap,2);
@@ -2202,15 +2182,6 @@ erts_memory(int *print_to_p, void *print_to_arg, void *proc, Eterm earg)
 		    return am_badarg;
 		}
 		break;
-#if HALFWORD_HEAP
-	    case am_low:
-		if (!want.low) {
-		    want.low = 1;
-		    atoms[length] = am_low;
-		    uintps[length++] = &size.low;
-		}
-		break;
-#endif
 	    default:
 		UnUseTmpHeapNoproc(2);
 		return am_badarg;
@@ -2222,11 +2193,12 @@ erts_memory(int *print_to_p, void *print_to_arg, void *proc, Eterm earg)
 	    return am_badarg;
     }
 
-    /* All alloc_util allocators *have* to be enabled */
+    /* All alloc_util allocators *have* to be enabled, except test_alloc */
     
     for (ai = ERTS_ALC_A_MIN; ai <= ERTS_ALC_A_MAX; ai++) {
 	switch (ai) {
 	case ERTS_ALC_A_SYSTEM:
+        case ERTS_ALC_A_TEST:
 	    break;
 	default:
 	    if (!erts_allctrs_info[ai].enabled
@@ -2267,6 +2239,8 @@ erts_memory(int *print_to_p, void *print_to_arg, void *proc, Eterm earg)
 		      * contain any allocated memory.
 		      */
 		    continue;
+                case ERTS_ALC_A_TEST:
+                    continue;
 		case ERTS_ALC_A_EHEAP:
 		    save = &size.processes;
 		    break;
@@ -2288,11 +2262,6 @@ erts_memory(int *print_to_p, void *print_to_arg, void *proc, Eterm earg)
 		if (save)
 		    *save = asz;
 		size.total += asz;
-#if HALFWORD_HEAP
-		if (alcu_is_low(ai)) {
-		    size.low += asz;
-		}
-#endif
 	    }
 	}
     }
@@ -2319,7 +2288,6 @@ erts_memory(int *print_to_p, void *print_to_arg, void *proc, Eterm earg)
 		       &size.processes_used,
 		       fi,
 		       ERTS_ALC_T_PROC);
-#if !HALFWORD_HEAP
 	add_fix_values(&size.processes,
 		       &size.processes_used,
 		       fi,
@@ -2329,7 +2297,6 @@ erts_memory(int *print_to_p, void *print_to_arg, void *proc, Eterm earg)
 		       &size.processes_used,
 		       fi,
 		       ERTS_ALC_T_NLINK_SH);
-#endif
 	add_fix_values(&size.processes,
 		       &size.processes_used,
 		       fi,
@@ -2675,14 +2642,17 @@ erts_alloc_util_allocators(void *proc)
     /*
      * Currently all allocators except sys_alloc are
      * alloc_util allocators.
+     * Also hide test_alloc which is disabled by default
+     * and only intended for our own testing.
      */
-    sz = ((ERTS_ALC_A_MAX + 1 - ERTS_ALC_A_MIN) - 1)*2;
+    sz = ((ERTS_ALC_A_MAX + 1 - ERTS_ALC_A_MIN) - 2)*2;
     ASSERT(sz > 0);
     hp = HAlloc((Process *) proc, sz);
     res = NIL;
     for (i = ERTS_ALC_A_MAX; i >= ERTS_ALC_A_MIN; i--) {
 	switch (i) {
 	case ERTS_ALC_A_SYSTEM:
+        case ERTS_ALC_A_TEST:
 	    break;
 	default: {
 	    char *alc_str = (char *) ERTS_ALC_A2AD(i);
@@ -2765,7 +2735,7 @@ erts_allocator_info(int to, void *arg)
 	    erts_mseg_info(i, &to, arg, 0, NULL, NULL);
 	}
 	erts_print(to, arg, "=allocator:mseg_alloc.erts_mmap\n");
-	erts_mmap_info(&to, arg, NULL, NULL, &emis);
+	erts_mmap_info(&erts_dflt_mmapper, &to, arg, NULL, NULL, &emis);
     }
 #endif
 
@@ -2991,12 +2961,12 @@ reply_alloc_info(void *vair)
     int global_instances = air->req_sched == sched_id;
     ErtsProcLocks rp_locks;
     Process *rp = air->proc;
-    Eterm ref_copy = NIL, ai_list, msg;
-    Eterm *hp = NULL, *hp_end = NULL, *hp_start = NULL;
+    Eterm ref_copy = NIL, ai_list, msg = NIL;
+    Eterm *hp = NULL, *hp_start = NULL, *hp_end = NULL;
     Eterm **hpp;
     Uint sz, *szp;
     ErlOffHeap *ohp = NULL;
-    ErlHeapFragment *bp = NULL;
+    ErtsMessage *mp = NULL;
     struct erts_mmap_info_struct emis;
     int i;
     Eterm (*info_func)(Allctr_t *,
@@ -3117,7 +3087,8 @@ reply_alloc_info(void *vair)
 
 		    ai_list = erts_bld_cons(hpp, szp,
 					    ainfo, ai_list);
-		    ainfo = (air->only_sz ? NIL : erts_mmap_info(NULL, NULL, hpp, szp, &emis));
+		    ainfo = (air->only_sz ? NIL :
+                             erts_mmap_info(&erts_dflt_mmapper, NULL, NULL, hpp, szp, &emis));
 		    ainfo = erts_bld_tuple3(hpp, szp,
                                             alloc_atom,
                                             erts_bld_atom(hpp,szp,"erts_mmap"),
@@ -3197,20 +3168,17 @@ reply_alloc_info(void *vair)
 	if (hpp)
 	    break;
 
-	hp = erts_alloc_message_heap(sz, &bp, &ohp, rp, &rp_locks);
+	mp = erts_alloc_message_heap(rp, &rp_locks, sz, &hp, &ohp);
 	hp_start = hp;
 	hp_end = hp + sz;
 	szp = NULL;
 	hpp = &hp;
     }
-    if (bp)
-	bp = erts_resize_message_buffer(bp, hp - hp_start, &msg, 1);
-    else {
-	ASSERT(hp);
-	HRelease(rp, hp_end, hp);	    
-    }
 
-    erts_queue_message(rp, &rp_locks, bp, msg, NIL);
+    if (hp != hp_end)
+	erts_shrink_message_heap(&mp, rp, hp_start, hp, hp_end, &msg, 1);
+
+    erts_queue_message(rp, &rp_locks, mp, msg, NIL);
 
     if (air->req_sched == sched_id)
 	rp_locks &= ~ERTS_PROC_LOCK_MAIN;
@@ -3566,6 +3534,41 @@ UWord erts_alc_test(UWord op, UWord a1, UWord a2, UWord a3)
 #else
 	case 0xf13: return (UWord) 0;
 #endif
+	case 0xf14: return (UWord) erts_alloc(ERTS_ALC_T_TEST, (Uint)a1);
+
+	case 0xf15: erts_free(ERTS_ALC_T_TEST, (void*)a1); return 0;
+
+	case 0xf16: {
+            Uint extra_hdr_sz = UNIT_CEILING((Uint)a1);
+	    ErtsAllocatorThrSpec_t* ts = &erts_allctr_thr_spec[ERTS_ALC_A_TEST];
+	    Uint offset = ts->allctr[0]->mbc_header_size;
+	    void* orig_creating_mbc = ts->allctr[0]->creating_mbc;
+	    void* orig_destroying_mbc = ts->allctr[0]->destroying_mbc;
+	    void* new_creating_mbc = *(void**)a2; /* inout arg */
+	    void* new_destroying_mbc = *(void**)a3; /* inout arg */
+	    int i;
+
+	    for (i=0; i < ts->size; i++) {
+		Allctr_t* ap = ts->allctr[i];
+		if (ap->mbc_header_size != offset
+		    || ap->creating_mbc != orig_creating_mbc
+		    || ap->destroying_mbc != orig_destroying_mbc
+		    || ap->mbc_list.first != NULL)
+		    return -1;
+	    }
+	    for (i=0; i < ts->size; i++) {
+		ts->allctr[i]->mbc_header_size += extra_hdr_sz;
+		ts->allctr[i]->creating_mbc = new_creating_mbc;
+		ts->allctr[i]->destroying_mbc = new_destroying_mbc;
+	    }
+	    *(void**)a2 = orig_creating_mbc;
+	    *(void**)a3 = orig_destroying_mbc;
+	    return offset;
+	}
+	case 0xf17: {
+	    ErtsAllocatorThrSpec_t* ts = &erts_allctr_thr_spec[ERTS_ALC_A_TEST];
+	    return ts->allctr[0]->largest_mbc_size;
+	}
 	default:
 	    break;
 	}
diff --git a/erts/emulator/beam/erl_alloc.h b/erts/emulator/beam/erl_alloc.h
index f540bae20d..14e80960f5 100644
--- a/erts/emulator/beam/erl_alloc.h
+++ b/erts/emulator/beam/erl_alloc.h
@@ -30,6 +30,7 @@
 #ifdef USE_THREADS
 #include "erl_threads.h"
 #endif
+#include "erl_mmap.h"
 
 #ifdef DEBUG
 #  undef ERTS_ALC_WANT_INLINE
@@ -43,9 +44,11 @@
 #if ERTS_CAN_INLINE && ERTS_ALC_WANT_INLINE
 #  define ERTS_ALC_DO_INLINE 1
 #  define ERTS_ALC_INLINE static ERTS_INLINE
+#  define ERTS_ALC_FORCE_INLINE static ERTS_FORCE_INLINE
 #else
 #  define ERTS_ALC_DO_INLINE 0
 #  define ERTS_ALC_INLINE
+#  define ERTS_ALC_FORCE_INLINE
 #endif
 
 #define ERTS_ALC_NO_FIXED_SIZES \
@@ -177,6 +180,12 @@ void  sys_free(void *)               __deprecated; /* erts_free()        */
 void *sys_alloc(Uint )               __deprecated; /* erts_alloc_fnf()   */
 void *sys_realloc(void *, Uint)      __deprecated; /* erts_realloc_fnf() */
 
+#undef ERTS_HAVE_IS_IN_LITERAL_RANGE
+#if defined(ARCH_32) || defined(ERTS_HAVE_OS_PHYSICAL_MEMORY_RESERVATION)
+#  define ERTS_HAVE_IS_IN_LITERAL_RANGE
+#endif
+
+
 /*
  * erts_alloc[_fnf](), erts_realloc[_fnf](), erts_free() works as
  * malloc(), realloc(), and free() with the following exceptions:
@@ -204,6 +213,9 @@ void erts_free(ErtsAlcType_t type, void *ptr);
 void *erts_alloc_fnf(ErtsAlcType_t type, Uint size);
 void *erts_realloc_fnf(ErtsAlcType_t type, void *ptr, Uint size);
 int erts_is_allctr_wrapper_prelocked(void);
+#ifdef ERTS_HAVE_IS_IN_LITERAL_RANGE
+int erts_is_in_literal_range(void* ptr);
+#endif
 
 #endif /* #if !ERTS_ALC_DO_INLINE */
 
@@ -281,6 +293,28 @@ int erts_is_allctr_wrapper_prelocked(void)
 	&& !!erts_tsd_get(erts_allctr_prelock_tsd_key);  /* by me  */
 }
 
+#ifdef ERTS_HAVE_IS_IN_LITERAL_RANGE
+
+ERTS_ALC_FORCE_INLINE
+int erts_is_in_literal_range(void* ptr)
+{
+#if defined(ARCH_32)
+    Uint ix = (UWord)ptr >> ERTS_MMAP_SUPERALIGNED_BITS;
+
+    return erts_literal_vspace_map[ix / ERTS_VSPACE_WORD_BITS]
+                  & ((UWord)1 << (ix % ERTS_VSPACE_WORD_BITS));
+
+#elif defined(ARCH_64)
+    extern char* erts_literals_start;
+    extern UWord erts_literals_size;
+    return ErtsInArea(ptr, erts_literals_start, erts_literals_size);
+#else
+# error No ARCH_xx
+#endif
+}
+
+#endif /* ERTS_HAVE_IS_IN_LITERAL_RANGE */
+
 #endif /* #if ERTS_ALC_DO_INLINE || defined(ERTS_ALC_INTERNAL__) */
 
 #define ERTS_ALC_GET_THR_IX() ((int) erts_get_scheduler_id())
diff --git a/erts/emulator/beam/erl_alloc.types b/erts/emulator/beam/erl_alloc.types
index b1d511ab78..e0bc71c88a 100644
--- a/erts/emulator/beam/erl_alloc.types
+++ b/erts/emulator/beam/erl_alloc.types
@@ -86,11 +86,7 @@ allocator	LONG_LIVED	true		ll_alloc
 allocator	EHEAP		true		eheap_alloc
 allocator	ETS		true		ets_alloc
 allocator	FIXED_SIZE	true		fix_alloc
-
-+if halfword
-allocator	LONG_LIVED_LOW	true		ll_low_alloc
-allocator	STANDARD_LOW	true		std_low_alloc
-+endif
+allocator	LITERAL 	true		literal_alloc
 
 +else # Non smp build
 
@@ -101,18 +97,14 @@ allocator	LONG_LIVED	false		ll_alloc
 allocator	EHEAP		false		eheap_alloc
 allocator	ETS		false		ets_alloc
 allocator	FIXED_SIZE	false		fix_alloc
-
-+if halfword
-allocator	LONG_LIVED_LOW	false		ll_low_alloc
-allocator	STANDARD_LOW	false		std_low_alloc
-+endif
+allocator	LITERAL 	false		literal_alloc
 
 +endif
 
 allocator	BINARY		true		binary_alloc
 allocator	DRIVER		true		driver_alloc
 
-
+allocator	TEST            true            test_alloc
 
 # --- Class declarations -----------------------------------------------------
 #
@@ -160,6 +152,8 @@ type	OLD_HEAP	EHEAP		PROCESSES	old_heap
 type	HEAP_FRAG	EHEAP		PROCESSES	heap_frag
 type	TMP_HEAP	TEMPORARY	PROCESSES	tmp_heap
 type	MSG_REF		FIXED_SIZE	PROCESSES	msg_ref
+type	MSG		EHEAP		PROCESSES	message
+type	MSGQ_CHNG	SHORT_LIVED	PROCESSES	messages_queue_change
 type	MSG_ROOTS	TEMPORARY	PROCESSES	msg_roots
 type	ROOTSET		TEMPORARY	PROCESSES	root_set
 type	LOADER_TMP	TEMPORARY	CODE		loader_tmp
@@ -349,33 +343,11 @@ type	SSB		SHORT_LIVED	PROCESSES	ssb
 
 +endif
 
-
-+if halfword
-
-type	DDLL_PROCESS	STANDARD_LOW	SYSTEM		ddll_processes
-type	MONITOR_LH	STANDARD_LOW	PROCESSES	monitor_lh
-type	NLINK_LH	STANDARD_LOW	PROCESSES	nlink_lh
-type	CODE		LONG_LIVED_LOW	CODE		code
-type	DB_HEIR_DATA	STANDARD_LOW	ETS		db_heir_data
-type	DB_MS_PSDO_PROC	LONG_LIVED_LOW	ETS		db_match_pseudo_proc
-type	SCHDLR_DATA	LONG_LIVED_LOW	SYSTEM		scheduler_data
-type  	LL_TEMP_TERM 	LONG_LIVED_LOW	SYSTEM		ll_temp_term
-
-type	NIF_TRAP_EXPORT	STANDARD_LOW	CODE		nif_trap_export_entry
-type	EXPORT		LONG_LIVED_LOW	CODE		export_entry
-type	MONITOR_SH	STANDARD_LOW	PROCESSES	monitor_sh
-type	NLINK_SH	STANDARD_LOW	PROCESSES	nlink_sh
-type	AINFO_REQ	STANDARD_LOW	SYSTEM		alloc_info_request
-type	SCHED_WTIME_REQ	STANDARD_LOW	SYSTEM		sched_wall_time_request
-type	GC_INFO_REQ	STANDARD_LOW	SYSTEM		gc_info_request
-type	PORT_DATA_HEAP	STANDARD_LOW	SYSTEM		port_data_heap
-
-+else # "fullword"
-
 type	DDLL_PROCESS	STANDARD	SYSTEM		ddll_processes
 type	MONITOR_LH	STANDARD	PROCESSES	monitor_lh
 type	NLINK_LH	STANDARD	PROCESSES	nlink_lh
 type	CODE		LONG_LIVED	CODE		code
+type	LITERAL		LITERAL 	CODE		literal
 type	DB_HEIR_DATA	STANDARD	ETS		db_heir_data
 type	DB_MS_PSDO_PROC	LONG_LIVED	ETS		db_match_pseudo_proc
 type	SCHDLR_DATA	LONG_LIVED	SYSTEM		scheduler_data
@@ -390,9 +362,6 @@ type	SCHED_WTIME_REQ	SHORT_LIVED	SYSTEM		sched_wall_time_request
 type	GC_INFO_REQ	SHORT_LIVED	SYSTEM		gc_info_request
 type	PORT_DATA_HEAP	STANDARD	SYSTEM		port_data_heap
 
-+endif
-
-
 #
 # Types used by system specific code
 #
@@ -430,21 +399,6 @@ type	SYS_WRITE_BUF	BINARY		SYSTEM		sys_write_buf
 
 +endif
 
-+if ose
-
-type	SYS_READ_BUF	TEMPORARY	SYSTEM		sys_read_buf
-type	FD_TAB		LONG_LIVED	SYSTEM		fd_tab
-type	FD_ENTRY_BUF	STANDARD	SYSTEM		fd_entry_buf
-type	FD_SIG_LIST	SHORT_LIVED	SYSTEM		fd_sig_list
-type    DRV_EV		STANDARD	SYSTEM		driver_event
-type	CS_PROG_PATH	LONG_LIVED	SYSTEM		cs_prog_path
-type	ENVIRONMENT	TEMPORARY	SYSTEM		environment
-type	PUTENV_STR	SYSTEM		SYSTEM		putenv_string
-type	PRT_REP_EXIT	STANDARD	SYSTEM		port_report_exit
-
-+endif
-
-
 +if win32
 
 type	DRV_DATA_BUF	SYSTEM		SYSTEM		drv_data_buf
@@ -456,4 +410,7 @@ type	CON_VPRINTF_BUF	TEMPORARY	SYSTEM		con_vprintf_buf
 
 +endif
 
+# This type should only be used for test
+type	TEST		TEST		SYSTEM		testing
+
 # ----------------------------------------------------------------------------
diff --git a/erts/emulator/beam/erl_alloc_util.c b/erts/emulator/beam/erl_alloc_util.c
index 236ee35d18..eedfd1e13d 100644
--- a/erts/emulator/beam/erl_alloc_util.c
+++ b/erts/emulator/beam/erl_alloc_util.c
@@ -751,12 +751,77 @@ internal_free(void *ptr)
 
 #endif
 
+#ifdef ARCH_32
+
+/*
+ * Bit vector for the entire 32-bit virtual address space
+ * with one bit for each super aligned memory segment.
+ */
+
+#define VSPACE_MAP_BITS  (1 << (32 - ERTS_MMAP_SUPERALIGNED_BITS))
+#define VSPACE_MAP_SZ    (VSPACE_MAP_BITS / ERTS_VSPACE_WORD_BITS)
+
+static ERTS_INLINE void set_bit(UWord* map, Uint ix)
+{
+    ASSERT(ix / ERTS_VSPACE_WORD_BITS < VSPACE_MAP_SZ);
+    map[ix / ERTS_VSPACE_WORD_BITS]
+        |= ((UWord)1 << (ix % ERTS_VSPACE_WORD_BITS));
+}
+
+static ERTS_INLINE void clr_bit(UWord* map, Uint ix)
+{
+    ASSERT(ix / ERTS_VSPACE_WORD_BITS < VSPACE_MAP_SZ);
+    map[ix / ERTS_VSPACE_WORD_BITS]
+        &= ~((UWord)1 << (ix % ERTS_VSPACE_WORD_BITS));
+}
+
+static ERTS_INLINE int is_bit_set(UWord* map, Uint ix)
+{
+    ASSERT(ix / ERTS_VSPACE_WORD_BITS < VSPACE_MAP_SZ);
+    return map[ix / ERTS_VSPACE_WORD_BITS]
+        & ((UWord)1 << (ix % ERTS_VSPACE_WORD_BITS));
+}
+
+UWord erts_literal_vspace_map[VSPACE_MAP_SZ];
+
+static void set_literal_range(void* start, Uint size)
+{
+    Uint ix = (UWord)start >> ERTS_MMAP_SUPERALIGNED_BITS;
+    Uint n = size >> ERTS_MMAP_SUPERALIGNED_BITS;
+
+    ASSERT(!((UWord)start & ERTS_INV_SUPERALIGNED_MASK));
+    ASSERT(!((UWord)size & ERTS_INV_SUPERALIGNED_MASK));
+    ASSERT(n);
+    while (n--) {
+        ASSERT(!is_bit_set(erts_literal_vspace_map, ix));
+        set_bit(erts_literal_vspace_map, ix);
+        ix++;
+    }
+}
+
+static void clear_literal_range(void* start, Uint size)
+{
+    Uint ix = (UWord)start >> ERTS_MMAP_SUPERALIGNED_BITS;
+    Uint n = size >> ERTS_MMAP_SUPERALIGNED_BITS;
+
+    ASSERT(!((UWord)start & ERTS_INV_SUPERALIGNED_MASK));
+    ASSERT(!((UWord)size & ERTS_INV_SUPERALIGNED_MASK));
+    ASSERT(n);
+    while (n--) {
+        ASSERT(is_bit_set(erts_literal_vspace_map, ix));
+        clr_bit(erts_literal_vspace_map, ix);
+        ix++;
+    }
+}
+
+#endif /* ARCH_32 */
+
 /* mseg ... */
 
 #if HAVE_ERTS_MSEG
 
-static ERTS_INLINE void *
-alcu_mseg_alloc(Allctr_t *allctr, Uint *size_p, Uint flags)
+void*
+erts_alcu_mseg_alloc(Allctr_t *allctr, Uint *size_p, Uint flags)
 {
     void *res;
     UWord size = (UWord) *size_p;
@@ -766,8 +831,9 @@ alcu_mseg_alloc(Allctr_t *allctr, Uint *size_p, Uint flags)
     return res;
 }
 
-static ERTS_INLINE void *
-alcu_mseg_realloc(Allctr_t *allctr, void *seg, Uint old_size, Uint *new_size_p)
+void*
+erts_alcu_mseg_realloc(Allctr_t *allctr, void *seg,
+                       Uint old_size, Uint *new_size_p)
 {
     void *res;
     UWord new_size = (UWord) *new_size_p;
@@ -778,17 +844,103 @@ alcu_mseg_realloc(Allctr_t *allctr, void *seg, Uint old_size, Uint *new_size_p)
     return res;
 }
 
-static ERTS_INLINE void
-alcu_mseg_dealloc(Allctr_t *allctr, void *seg, Uint size, Uint flags)
+void
+erts_alcu_mseg_dealloc(Allctr_t *allctr, void *seg, Uint size, Uint flags)
 {
     erts_mseg_dealloc_opt(allctr->alloc_no, seg, (UWord) size, flags, &allctr->mseg_opt);
     INC_CC(allctr->calls.mseg_dealloc);
 }
 
-#endif
 
-static ERTS_INLINE void *
-alcu_sys_alloc(Allctr_t *allctr, Uint size, int superalign)
+#if defined(ARCH_32)
+
+void*
+erts_alcu_literal_32_mseg_alloc(Allctr_t *allctr, Uint *size_p, Uint flags)
+{
+    void* res;
+    ERTS_LC_ASSERT(allctr->alloc_no == ERTS_ALC_A_LITERAL
+                   && !allctr->t && allctr->thread_safe);
+
+    res = erts_alcu_mseg_alloc(allctr, size_p, flags);
+    if (res)
+        set_literal_range(res, *size_p);
+    return res;
+}
+
+void*
+erts_alcu_literal_32_mseg_realloc(Allctr_t *allctr, void *seg,
+                               Uint old_size, Uint *new_size_p)
+{
+    void* res;
+    ERTS_LC_ASSERT(allctr->alloc_no == ERTS_ALC_A_LITERAL
+                   && !allctr->t && allctr->thread_safe);
+
+    if (seg && old_size)
+        clear_literal_range(seg, old_size);
+    res = erts_alcu_mseg_realloc(allctr, seg, old_size, new_size_p);
+    if (res)
+        set_literal_range(res, *new_size_p);
+    return res;
+}
+
+void
+erts_alcu_literal_32_mseg_dealloc(Allctr_t *allctr, void *seg, Uint size,
+                               Uint flags)
+{
+    ERTS_LC_ASSERT(allctr->alloc_no == ERTS_ALC_A_LITERAL
+                   && !allctr->t && allctr->thread_safe);
+
+    erts_alcu_mseg_dealloc(allctr, seg, size, flags);
+
+    clear_literal_range(seg, size);
+}
+
+#elif defined(ARCH_64) && defined(ERTS_HAVE_OS_PHYSICAL_MEMORY_RESERVATION)
+
+void*
+erts_alcu_literal_64_mseg_alloc(Allctr_t *allctr, Uint *size_p, Uint flags)
+{
+    void* res;
+    UWord size = (UWord) *size_p;
+    Uint32 mmap_flags = ERTS_MMAPFLG_SUPERCARRIER_ONLY;
+    if (flags & ERTS_MSEG_FLG_2POW)
+        mmap_flags |= ERTS_MMAPFLG_SUPERALIGNED;
+
+    res = erts_mmap(&erts_literal_mmapper, mmap_flags, &size);
+    *size_p = (Uint)size;
+    INC_CC(allctr->calls.mseg_alloc);
+    return res;
+}
+
+void*
+erts_alcu_literal_64_mseg_realloc(Allctr_t *allctr, void *seg,
+                               Uint old_size, Uint *new_size_p)
+{
+    void *res;
+    UWord new_size = (UWord) *new_size_p;
+    res = erts_mremap(&erts_literal_mmapper, ERTS_MSEG_FLG_NONE, seg, old_size, &new_size);
+    *new_size_p = (Uint) new_size;
+    INC_CC(allctr->calls.mseg_realloc);
+    return res;
+}
+
+void
+erts_alcu_literal_64_mseg_dealloc(Allctr_t *allctr, void *seg, Uint size,
+                               Uint flags)
+{
+    Uint32 mmap_flags = ERTS_MMAPFLG_SUPERCARRIER_ONLY;
+    if (flags & ERTS_MSEG_FLG_2POW)
+        mmap_flags |= ERTS_MMAPFLG_SUPERALIGNED;
+
+    erts_munmap(&erts_literal_mmapper, mmap_flags, seg, (UWord)size);
+    INC_CC(allctr->calls.mseg_dealloc);
+}
+#endif /* ARCH_64 && ERTS_HAVE_OS_PHYSICAL_MEMORY_RESERVATION */
+
+#endif /* HAVE_ERTS_MSEG */
+
+void*
+erts_alcu_sys_alloc(Allctr_t *allctr, Uint size, int superalign)
 {
     void *res;
 #if ERTS_SA_MB_CARRIERS && ERTS_HAVE_ERTS_SYS_ALIGNED_ALLOC
@@ -803,8 +955,8 @@ alcu_sys_alloc(Allctr_t *allctr, Uint size, int superalign)
     return res;
 }
 
-static ERTS_INLINE void *
-alcu_sys_realloc(Allctr_t *allctr, void *ptr, Uint size, Uint old_size, int superalign)
+void*
+erts_alcu_sys_realloc(Allctr_t *allctr, void *ptr, Uint size, Uint old_size, int superalign)
 {
     void *res;
 
@@ -824,8 +976,8 @@ alcu_sys_realloc(Allctr_t *allctr, void *ptr, Uint size, Uint old_size, int supe
     return res;
 }
 
-static ERTS_INLINE void
-alcu_sys_free(Allctr_t *allctr, void *ptr, int superalign)
+void
+erts_alcu_sys_dealloc(Allctr_t *allctr, void *ptr, Uint size, int superalign)
 {
 #if ERTS_SA_MB_CARRIERS && ERTS_HAVE_ERTS_SYS_ALIGNED_ALLOC
     if (superalign)
@@ -838,6 +990,49 @@ alcu_sys_free(Allctr_t *allctr, void *ptr, int superalign)
 	erts_mtrace_crr_free(allctr->alloc_no, ERTS_ALC_A_SYSTEM, ptr);
 }
 
+#ifdef ARCH_32
+
+void*
+erts_alcu_literal_32_sys_alloc(Allctr_t *allctr, Uint size, int superalign)
+{
+    void* res;
+    ERTS_LC_ASSERT(allctr->alloc_no == ERTS_ALC_A_LITERAL
+                   && !allctr->t && allctr->thread_safe);
+
+    res = erts_alcu_sys_alloc(allctr, size, 1);
+    if (res)
+        set_literal_range(res, size);
+    return res;
+}
+
+void*
+erts_alcu_literal_32_sys_realloc(Allctr_t *allctr, void *ptr, Uint size, Uint old_size, int superalign)
+{
+    void* res;
+    ERTS_LC_ASSERT(allctr->alloc_no == ERTS_ALC_A_LITERAL
+                   && !allctr->t && allctr->thread_safe);
+
+    if (ptr && old_size)
+        clear_literal_range(ptr, old_size);
+    res = erts_alcu_sys_realloc(allctr, ptr, size, old_size, 1);
+    if (res)
+        set_literal_range(res, size);
+    return res;
+}
+
+void
+erts_alcu_literal_32_sys_dealloc(Allctr_t *allctr, void *ptr, Uint size, int superalign)
+{
+    ERTS_LC_ASSERT(allctr->alloc_no == ERTS_ALC_A_LITERAL
+                   && !allctr->t && allctr->thread_safe);
+
+    erts_alcu_sys_dealloc(allctr, ptr, size, 1);
+
+    clear_literal_range(ptr, size);
+}
+
+#endif /* ARCH_32 */
+
 static Uint
 get_next_mbc_size(Allctr_t *allctr)
 {
@@ -1437,6 +1632,16 @@ erts_alcu_fix_alloc_shrink(Allctr_t *allctr, erts_aint32_t flgs)
 
 static void dealloc_carrier(Allctr_t *allctr, Carrier_t *crr, int superaligned);
 
+static ERTS_INLINE void
+dealloc_mbc(Allctr_t *allctr, Carrier_t *crr)
+{
+    ASSERT(IS_MB_CARRIER(crr));
+    if (allctr->destroying_mbc)
+        allctr->destroying_mbc(allctr, crr);
+
+    dealloc_carrier(allctr, crr, 1);
+}
+
 #ifdef ERTS_SMP
 
 static ERTS_INLINE Allctr_t*
@@ -2074,7 +2279,7 @@ mbc_alloc_block(Allctr_t *allctr, Uint size, Uint *blk_szp)
 
     if (!blk) {
 	blk = create_carrier(allctr, get_blk_sz, CFLG_MBC);
-#if !HALFWORD_HEAP && !ERTS_SUPER_ALIGNED_MSEG_ONLY
+#if !ERTS_SUPER_ALIGNED_MSEG_ONLY
 	if (!blk) {
 	    /* Emergency! We couldn't create the carrier as we wanted.
 	       Try to place it in a sys_alloced sbc. */
@@ -3149,7 +3354,7 @@ cpool_fetch(Allctr_t *allctr, UWord size)
 	cpool_entrance = sentinel;
 	cpdp = cpool_aint2cpd(cpool_read(&cpool_entrance->prev));
 	if (cpdp == sentinel)
-	    return NULL;
+	    goto check_dc_list;
     }
 
     has_passed_sentinel = 0;
@@ -3160,18 +3365,18 @@ cpool_fetch(Allctr_t *allctr, UWord size)
 	    if (cpool_entrance == sentinel) {
 		cpdp = cpool_aint2cpd(cpool_read(&cpdp->prev));
 		if (cpdp == sentinel)
-		    return NULL;
+		    break;
 	    }
 	    i = 0; /* Last one to inspect */
 	}
 	else if (cpdp == sentinel) {
 	    if (has_passed_sentinel) {
 		/* We been here before. cpool_entrance must have been removed */
-		return NULL;
+		break;
 	    }
 	    cpdp = cpool_aint2cpd(cpool_read(&cpdp->prev));
 	    if (cpdp == sentinel)
-		return NULL;
+                break;
 	    has_passed_sentinel = 1;
 	}
 	crr = (Carrier_t *)(((char *)cpdp) - offsetof(Carrier_t, cpool));
@@ -3195,10 +3400,12 @@ cpool_fetch(Allctr_t *allctr, UWord size)
 	    return NULL;
     }
 
+check_dc_list:
     /* Last; check our own pending dealloc carrier list... */
     crr = allctr->cpool.dc_list.last;
     while (crr) {
 	if (erts_atomic_read_nob(&crr->cpool.max_size) >= size) {
+	    Block_t* blk;
 	    unlink_carrier(&allctr->cpool.dc_list, crr);
 #ifdef ERTS_ALC_CPOOL_DEBUG
 	    ERTS_ALC_CPOOL_ASSERT(erts_smp_atomic_xchg_nob(&crr->allctr,
@@ -3207,6 +3414,9 @@ cpool_fetch(Allctr_t *allctr, UWord size)
 #else
 	    erts_smp_atomic_set_nob(&crr->allctr, ((erts_aint_t) allctr));
 #endif
+	    blk = MBC_TO_FIRST_BLK(allctr, crr);
+	    ASSERT(FBLK_TO_MBC(blk) == crr);
+	    allctr->link_free_block(allctr, blk);
 	    return crr;
 	}
 	crr = crr->prev;
@@ -3237,7 +3447,7 @@ check_pending_dealloc_carrier(Allctr_t *allctr,
 
 	    dcrr = crr;
 	    crr = crr->next;
-	    dealloc_carrier(allctr, dcrr, 1);
+	    dealloc_mbc(allctr, dcrr);
 	    i++;
 	} while (crr && i < ERTS_ALC_MAX_DEALLOC_CARRIER);
 
@@ -3268,18 +3478,20 @@ static void
 schedule_dealloc_carrier(Allctr_t *allctr, Carrier_t *crr)
 {
     Allctr_t *orig_allctr;
+    Block_t *blk;
     int check_pending_dealloc;
     erts_aint_t max_size;
 
+    ASSERT(IS_MB_CARRIER(crr));
+
     if (!ERTS_ALC_IS_CPOOL_ENABLED(allctr)) {
-	dealloc_carrier(allctr, crr, 1);
+	dealloc_mbc(allctr, crr);
 	return;
     }
 
     orig_allctr = crr->cpool.orig_allctr;
 
     if (allctr != orig_allctr) {
-	Block_t *blk = MBC_TO_FIRST_BLK(allctr, crr);
 	int cinit = orig_allctr->dd.ix - allctr->dd.ix;
 
 	/*
@@ -3296,6 +3508,7 @@ schedule_dealloc_carrier(Allctr_t *allctr, Carrier_t *crr)
 	 * since the block is an mbc block that is free and last
 	 * in the carrier.
 	 */
+	blk = MBC_TO_FIRST_BLK(allctr, crr);
 	ERTS_ALC_CPOOL_ASSERT(IS_FREE_LAST_MBC_BLK(blk));
 
 	ERTS_ALC_CPOOL_ASSERT(IS_MBC_FIRST_ABLK(allctr, blk));
@@ -3315,11 +3528,13 @@ schedule_dealloc_carrier(Allctr_t *allctr, Carrier_t *crr)
 
     if (crr->cpool.thr_prgr == ERTS_THR_PRGR_INVALID
 	|| erts_thr_progress_has_reached(crr->cpool.thr_prgr)) {
-	dealloc_carrier(allctr, crr, 1);
+	dealloc_mbc(allctr, crr);
 	return;
     }
 
-    max_size = (erts_aint_t) allctr->largest_fblk_in_mbc(allctr, crr);
+    blk = MBC_TO_FIRST_BLK(allctr, crr);
+    ASSERT(IS_FREE_LAST_MBC_BLK(blk));
+    max_size = (erts_aint_t) MBC_FBLK_SZ(blk);
     erts_atomic_set_nob(&crr->cpool.max_size, max_size);
 
     crr->next = NULL;
@@ -3511,8 +3726,7 @@ create_carrier(Allctr_t *allctr, Uint umem_sz, UWord flags)
     int is_mseg = 0;
 #endif
 
-    if (HALFWORD_HEAP
-	|| (ERTS_SUPER_ALIGNED_MSEG_ONLY && (flags & CFLG_MBC))
+    if ((ERTS_SUPER_ALIGNED_MSEG_ONLY && (flags & CFLG_MBC))
 	|| !allow_sys_alloc_carriers) {
 	flags |= CFLG_FORCE_MSEG;
 	flags &= ~CFLG_FORCE_SYS_ALLOC;
@@ -3520,6 +3734,8 @@ create_carrier(Allctr_t *allctr, Uint umem_sz, UWord flags)
 	return NULL;
 #endif
     }
+    flags |= allctr->crr_set_flgs;
+    flags &= ~allctr->crr_clr_flgs;
 
     ASSERT((flags & CFLG_SBC && !(flags & CFLG_MBC))
 	   || (flags & CFLG_MBC && !(flags & CFLG_SBC)));
@@ -3535,7 +3751,21 @@ create_carrier(Allctr_t *allctr, Uint umem_sz, UWord flags)
 	return NULL;
     }
 
-    blk_sz = UMEMSZ2BLKSZ(allctr, umem_sz);
+    if (flags & CFLG_MAIN_CARRIER) {
+        ASSERT(flags & CFLG_MBC);
+        ASSERT(flags & CFLG_NO_CPOOL);
+        ASSERT(umem_sz == allctr->main_carrier_size);
+        ERTS_UNDEF(blk_sz, 0);
+
+        if (allctr->main_carrier_size < allctr->min_mbc_size)
+            allctr->main_carrier_size = allctr->min_mbc_size;
+        crr_sz = bcrr_sz = allctr->main_carrier_size;
+    }
+    else {
+        ERTS_UNDEF(bcrr_sz, 0);
+	ERTS_UNDEF(crr_sz, 0);
+        blk_sz = UMEMSZ2BLKSZ(allctr, umem_sz);
+    }
 
 #ifdef ERTS_SMP
     allctr->cpool.disable_abandon = ERTS_ALC_CPOOL_MAX_DISABLE_ABANDON;
@@ -3581,13 +3811,15 @@ create_carrier(Allctr_t *allctr, Uint umem_sz, UWord flags)
 	mseg_flags = ERTS_MSEG_FLG_NONE;
     }
     else {
-	crr_sz = (*allctr->get_next_mbc_size)(allctr);
-	if (crr_sz < MBC_HEADER_SIZE(allctr) + blk_sz)
-	    crr_sz = MBC_HEADER_SIZE(allctr) + blk_sz;
-	mseg_flags = ERTS_MSEG_FLG_2POW;
+        if (!(flags & CFLG_MAIN_CARRIER)) {
+            crr_sz = (*allctr->get_next_mbc_size)(allctr);
+            if (crr_sz < MBC_HEADER_SIZE(allctr) + blk_sz)
+                crr_sz = MBC_HEADER_SIZE(allctr) + blk_sz;
+        }
+        mseg_flags = ERTS_MSEG_FLG_2POW;
     }
 
-    crr = (Carrier_t *) alcu_mseg_alloc(allctr, &crr_sz, mseg_flags);
+    crr = (Carrier_t *) allctr->mseg_alloc(allctr, &crr_sz, mseg_flags);
     if (!crr) {
 	have_tried_mseg = 1;
 	if (!(have_tried_sys_alloc || flags & CFLG_FORCE_MSEG))
@@ -3619,23 +3851,22 @@ create_carrier(Allctr_t *allctr, Uint umem_sz, UWord flags)
     if (flags & CFLG_SBC) {
 	bcrr_sz = blk_sz + SBC_HEADER_SIZE;
     }
-    else {
+    else if (!(flags & CFLG_MAIN_CARRIER)) {
 	bcrr_sz = MBC_HEADER_SIZE(allctr) + blk_sz;
-	if (!(flags & CFLG_MAIN_CARRIER)
-	    && bcrr_sz < allctr->smallest_mbc_size)
-	    bcrr_sz = allctr->smallest_mbc_size;
+	if (bcrr_sz < allctr->smallest_mbc_size)
+            bcrr_sz = allctr->smallest_mbc_size;
     }
 
     crr_sz = (flags & CFLG_FORCE_SIZE
 	      ? UNIT_CEILING(bcrr_sz)
 	      : SYS_ALLOC_CARRIER_CEILING(bcrr_sz));
 
-    crr = (Carrier_t *) alcu_sys_alloc(allctr, crr_sz, flags & CFLG_MBC);
+    crr = (Carrier_t *) allctr->sys_alloc(allctr, crr_sz, flags & CFLG_MBC);
 	
     if (!crr) {
 	if (crr_sz > UNIT_CEILING(bcrr_sz)) {
 	    crr_sz = UNIT_CEILING(bcrr_sz);
-	    crr = (Carrier_t *) alcu_sys_alloc(allctr, crr_sz, flags & CFLG_MBC);
+	    crr = (Carrier_t *) allctr->sys_alloc(allctr, crr_sz, flags & CFLG_MBC);
 	}
 	if (!crr) {
 #if HAVE_ERTS_MSEG
@@ -3734,7 +3965,7 @@ resize_carrier(Allctr_t *allctr, Block_t *old_blk, Uint umem_sz, UWord flags)
 
 	    new_crr_sz = new_blk_sz + SBC_HEADER_SIZE;
 	    new_crr_sz = ERTS_SACRR_UNIT_CEILING(new_crr_sz);
-	    new_crr = (Carrier_t *) alcu_mseg_realloc(allctr,
+	    new_crr = (Carrier_t *) allctr->mseg_realloc(allctr,
 						      old_crr,
 						      old_crr_sz,
 						      &new_crr_sz);
@@ -3749,11 +3980,6 @@ resize_carrier(Allctr_t *allctr, Block_t *old_blk, Uint umem_sz, UWord flags)
 		DEBUG_SAVE_ALIGNMENT(new_crr);
 		return new_blk;
 	    }
-#if HALFWORD_HEAP
-	    /* Old carrier unchanged; restore stat */
-	    STAT_MSEG_SBC_ALLOC(allctr, old_crr_sz, old_blk_sz);
-	    return NULL;
-#endif	    
 	    create_flags |= CFLG_FORCE_SYS_ALLOC; /* since mseg_realloc()
 						     failed */
 	}
@@ -3764,7 +3990,7 @@ resize_carrier(Allctr_t *allctr, Block_t *old_blk, Uint umem_sz, UWord flags)
 		       (void *) BLK2UMEM(old_blk),
 		       MIN(new_blk_sz, old_blk_sz) - ABLK_HDR_SZ);
 	    unlink_carrier(&allctr->sbc_list, old_crr);
-	    alcu_mseg_dealloc(allctr, old_crr, old_crr_sz, ERTS_MSEG_FLG_NONE);
+            allctr->mseg_dealloc(allctr, old_crr, old_crr_sz, ERTS_MSEG_FLG_NONE);
 	}
 	else {
 	    /* Old carrier unchanged; restore stat */
@@ -3781,7 +4007,7 @@ resize_carrier(Allctr_t *allctr, Block_t *old_blk, Uint umem_sz, UWord flags)
 			  ? UNIT_CEILING(new_bcrr_sz)
 			  : SYS_ALLOC_CARRIER_CEILING(new_bcrr_sz));
 
-	    new_crr = (Carrier_t *) alcu_sys_realloc(allctr,
+	    new_crr = (Carrier_t *) allctr->sys_realloc(allctr,
 						     (void *) old_crr,
 						     new_crr_sz,
 						     old_crr_sz,
@@ -3802,7 +4028,7 @@ resize_carrier(Allctr_t *allctr, Block_t *old_blk, Uint umem_sz, UWord flags)
 	    else if (new_crr_sz > UNIT_CEILING(new_bcrr_sz)) {
 		new_crr_sz = new_blk_sz + SBC_HEADER_SIZE;
 		new_crr_sz = UNIT_CEILING(new_crr_sz);
-		new_crr = (Carrier_t *) alcu_sys_realloc(allctr,
+		new_crr = (Carrier_t *) allctr->sys_realloc(allctr,
 							 (void *) old_crr,
 							 new_crr_sz,
 							 old_crr_sz,
@@ -3825,7 +4051,7 @@ resize_carrier(Allctr_t *allctr, Block_t *old_blk, Uint umem_sz, UWord flags)
 		       (void *) BLK2UMEM(old_blk),
 		       MIN(new_blk_sz, old_blk_sz) - ABLK_HDR_SZ);
 	    unlink_carrier(&allctr->sbc_list, old_crr);
-	    alcu_sys_free(allctr, old_crr, 0);
+	    allctr->sys_dealloc(allctr, old_crr, CARRIER_SZ(old_crr), 0);
 	}
 	else {
 	    /* Old carrier unchanged; restore... */
@@ -3841,13 +4067,13 @@ dealloc_carrier(Allctr_t *allctr, Carrier_t *crr, int superaligned)
 {
 #if HAVE_ERTS_MSEG
     if (IS_MSEG_CARRIER(crr))
-	alcu_mseg_dealloc(allctr, crr, CARRIER_SZ(crr),
+	allctr->mseg_dealloc(allctr, crr, CARRIER_SZ(crr),
 			  (superaligned
 			   ? ERTS_MSEG_FLG_2POW
 			   : ERTS_MSEG_FLG_NONE));
     else
 #endif
-	alcu_sys_free(allctr, crr, superaligned);
+	allctr->sys_dealloc(allctr, crr, CARRIER_SZ(crr), superaligned);
 }
 
 static void
@@ -3894,9 +4120,6 @@ destroy_carrier(Allctr_t *allctr, Block_t *blk, Carrier_t **busy_pcrr_pp)
 	}
 #endif
 
-	if (allctr->destroying_mbc)
-	    (*allctr->destroying_mbc)(allctr, crr);
-
 #ifdef ERTS_SMP
 	if (busy_pcrr_pp && *busy_pcrr_pp) {
 	    ERTS_ALC_CPOOL_ASSERT(*busy_pcrr_pp == crr);
@@ -3920,12 +4143,15 @@ destroy_carrier(Allctr_t *allctr, Block_t *blk, Carrier_t **busy_pcrr_pp)
 	    else
 #endif
 		STAT_SYS_ALLOC_MBC_FREE(allctr, crr_sz);
+
+            if (allctr->remove_mbc)
+                allctr->remove_mbc(allctr, crr);
 	}
 
 #ifdef ERTS_SMP
 	schedule_dealloc_carrier(allctr, crr);
 #else
-	dealloc_carrier(allctr, crr, 1);
+	dealloc_mbc(allctr, crr);
 #endif
     }
 }
@@ -3942,9 +4168,6 @@ static struct {
     Eterm e;
     Eterm t;
     Eterm ramv;
-#if HALFWORD_HEAP
-    Eterm low;
-#endif
     Eterm sbct;
 #if HAVE_ERTS_MSEG
     Eterm asbcst;
@@ -4035,9 +4258,6 @@ init_atoms(Allctr_t *allctr)
 	AM_INIT(e);
 	AM_INIT(t);
 	AM_INIT(ramv);
-#if HALFWORD_HEAP
-	AM_INIT(low);
-#endif
 	AM_INIT(sbct);
 #if HAVE_ERTS_MSEG
 	AM_INIT(asbcst);
@@ -4643,9 +4863,6 @@ info_options(Allctr_t *allctr,
 		   "option e: true\n"
 		   "option t: %s\n"
 		   "option ramv: %s\n"
-#if HALFWORD_HEAP
-		   "option low: %s\n"
-#endif
 		   "option sbct: %beu\n"
 #if HAVE_ERTS_MSEG
 		   "option asbcst: %bpu\n"
@@ -4664,9 +4881,6 @@ info_options(Allctr_t *allctr,
 		   "option acul: %d\n",
 		   topt,
 		   allctr->ramv ? "true" : "false",
-#if HALFWORD_HEAP
-		   allctr->mseg_opt.low_mem ? "true" : "false",
-#endif
 		   allctr->sbc_threshold,
 #if HAVE_ERTS_MSEG
 		   allctr->mseg_opt.abs_shrink_th,
@@ -4729,9 +4943,6 @@ info_options(Allctr_t *allctr,
 	add_2tup(hpp, szp, &res,
 		 am.sbct,
 		 bld_uint(hpp, szp, allctr->sbc_threshold));
-#if HALFWORD_HEAP
-	add_2tup(hpp, szp, &res, am.low, allctr->mseg_opt.low_mem ? am_true : am_false);
-#endif
 	add_2tup(hpp, szp, &res, am.ramv, allctr->ramv ? am_true : am_false);
 	add_2tup(hpp, szp, &res, am.t, (allctr->t ? am_true : am_false));
 	add_2tup(hpp, szp, &res, am.e, am_true);
@@ -5416,11 +5627,7 @@ do_erts_alcu_realloc(ErtsAlcType_t type,
 	Block_t *new_blk;
 	if(IS_SBC_BLK(blk)) {
 	do_carrier_resize:
-#if HALFWORD_HEAP
-	    new_blk = resize_carrier(allctr, blk, size, CFLG_SBC | CFLG_FORCE_MSEG);
-#else
 	    new_blk = resize_carrier(allctr, blk, size, CFLG_SBC);
-#endif
 	    res = new_blk ? BLK2UMEM(new_blk) : NULL;
 	}
 	else if (alcu_flgs & ERTS_ALCU_FLG_FAIL_REALLOC_MOVE)
@@ -5714,11 +5921,8 @@ erts_alcu_start(Allctr_t *allctr, AllctrInit_t *init)
 #ifdef ERTS_SMP
     if (init->tspec || init->tpref)
 	allctr->mseg_opt.sched_spec = 1;
-#endif
-# if HALFWORD_HEAP
-    allctr->mseg_opt.low_mem = init->low_mem;
-# endif
-#endif
+#endif /* ERTS_SMP */
+#endif /* HAVE_ERTS_MSEG */
 
     allctr->name_prefix			= init->name_prefix;
     if (!allctr->name_prefix)
@@ -5867,17 +6071,52 @@ erts_alcu_start(Allctr_t *allctr, AllctrInit_t *init)
 					    + ABLK_HDR_SZ)
 			       - ABLK_HDR_SZ);
 
+    if (init->sys_alloc) {
+        ASSERT(init->sys_realloc && init->sys_dealloc);
+        allctr->sys_alloc   = init->sys_alloc;
+        allctr->sys_realloc = init->sys_realloc;
+        allctr->sys_dealloc = init->sys_dealloc;
+    }
+    else {
+        ASSERT(!init->sys_realloc && !init->sys_dealloc);
+        allctr->sys_alloc   = &erts_alcu_sys_alloc;
+        allctr->sys_realloc = &erts_alcu_sys_realloc;
+        allctr->sys_dealloc = &erts_alcu_sys_dealloc;
+    }
+#if HAVE_ERTS_MSEG
+    if (init->mseg_alloc) {
+        ASSERT(init->mseg_realloc && init->mseg_dealloc);
+        allctr->mseg_alloc   = init->mseg_alloc;
+        allctr->mseg_realloc = init->mseg_realloc;
+        allctr->mseg_dealloc = init->mseg_dealloc;
+    }
+    else {
+        ASSERT(!init->mseg_realloc && !init->mseg_dealloc);
+        allctr->mseg_alloc   = &erts_alcu_mseg_alloc;
+        allctr->mseg_realloc = &erts_alcu_mseg_realloc;
+        allctr->mseg_dealloc = &erts_alcu_mseg_dealloc;
+    }
+    /* If a custom carrier alloc function is specified, make sure it's used */
+    if (init->mseg_alloc && !init->sys_alloc) {
+        allctr->crr_set_flgs = CFLG_FORCE_MSEG;
+        allctr->crr_clr_flgs = CFLG_FORCE_SYS_ALLOC;
+    }
+    else if (!init->mseg_alloc && init->sys_alloc) {
+        allctr->crr_set_flgs = CFLG_FORCE_SYS_ALLOC;
+        allctr->crr_clr_flgs = CFLG_FORCE_MSEG;
+    }
+#endif
+
     if (allctr->main_carrier_size) {
 	Block_t *blk;
 
 	blk = create_carrier(allctr,
 			     allctr->main_carrier_size,
-			     CFLG_MBC
+                             (ERTS_SUPER_ALIGNED_MSEG_ONLY
+                              ? CFLG_FORCE_MSEG : CFLG_FORCE_SYS_ALLOC)
+                             | CFLG_MBC
 			     | CFLG_FORCE_SIZE
 			     | CFLG_NO_CPOOL
-#if !HALFWORD_HEAP && !ERTS_SUPER_ALIGNED_MSEG_ONLY
-			     | CFLG_FORCE_SYS_ALLOC
-#endif
 			     | CFLG_MAIN_CARRIER);
 	if (!blk) {
 #ifdef USE_THREADS
@@ -6054,6 +6293,16 @@ erts_alcu_test(UWord op, UWord a1, UWord a2)
     case 0x023: return (UWord) 0;
     case 0x024: return (UWord) 0;
 #endif
+    case 0x025: /* UMEM2BLK_TEST*/
+#ifdef DEBUG
+# ifdef HARD_DEBUG
+	return (UWord)UMEM2BLK(a1-3*sizeof(UWord));
+# else
+	return (UWord)UMEM2BLK(a1-2*sizeof(UWord));
+# endif
+#else
+	return (UWord)UMEM2BLK(a1);
+#endif
 
     default:	ASSERT(0); return ~((UWord) 0);
     }
diff --git a/erts/emulator/beam/erl_alloc_util.h b/erts/emulator/beam/erl_alloc_util.h
index df1f0aa65a..b7d717ed23 100644
--- a/erts/emulator/beam/erl_alloc_util.h
+++ b/erts/emulator/beam/erl_alloc_util.h
@@ -24,6 +24,12 @@
 #define ERTS_ALCU_VSN_STR "3.0"
 
 #include "erl_alloc_types.h"
+#ifdef USE_THREADS
+#define ERL_THREADS_EMU_INTERNAL__
+#include "erl_threads.h"
+#endif
+
+#include "erl_mseg.h"
 
 #define ERTS_AU_PREF_ALLOC_BITS 11
 #define ERTS_AU_MAX_PREF_ALLOC_INSTANCES (1 << ERTS_AU_PREF_ALLOC_BITS)
@@ -45,7 +51,6 @@ typedef struct {
     int tspec;
     int tpref;
     int ramv;
-    int low_mem;   /* HALFWORD only */
     UWord sbct;
     UWord asbcst;
     UWord rsbcst;
@@ -61,6 +66,15 @@ typedef struct {
 
     void *fix;
     size_t *fix_type_size;
+
+#if HAVE_ERTS_MSEG
+    void* (*mseg_alloc)(Allctr_t*, Uint *size_p, Uint flags);
+    void* (*mseg_realloc)(Allctr_t*, void *seg, Uint old_size, Uint *new_size_p);
+    void  (*mseg_dealloc)(Allctr_t*, void *seg, Uint size, Uint flags);
+#endif
+    void* (*sys_alloc)(Allctr_t *allctr, Uint size, int superalign);
+    void* (*sys_realloc)(Allctr_t *allctr, void *ptr, Uint size, Uint old_size, int superalign);
+    void  (*sys_dealloc)(Allctr_t *allctr, void *ptr, Uint size, int superalign);
 } AllctrInit_t;
 
 typedef struct {
@@ -90,7 +104,6 @@ typedef struct {
     0,			/* (bool)   tspec:  thread specific              */\
     0,			/* (bool)   tpref:  thread preferred             */\
     0,			/* (bool)   ramv:   realloc always moves         */\
-    0,			/* (bool)   low_mem: HALFWORD only               */\
     512*1024,		/* (bytes)  sbct:   sbc threshold                */\
     2*1024*2024,	/* (amount) asbcst: abs sbc shrink threshold     */\
     20,			/* (%)      rsbcst: rel sbc shrink threshold     */\
@@ -125,7 +138,6 @@ typedef struct {
     0,			/* (bool)   tspec:  thread specific              */\
     0,			/* (bool)   tpref:  thread preferred             */\
     0,			/* (bool)   ramv:   realloc always moves         */\
-    0,			/* (bool)   low_mem: HALFWORD only               */\
     64*1024,		/* (bytes)  sbct:   sbc threshold                */\
     2*1024*2024,	/* (amount) asbcst: abs sbc shrink threshold     */\
     20,			/* (%)      rsbcst: rel sbc shrink threshold     */\
@@ -176,20 +188,44 @@ void    erts_alcu_check_delayed_dealloc(Allctr_t *, int, int *, ErtsThrPrgrVal *
 #endif
 erts_aint32_t erts_alcu_fix_alloc_shrink(Allctr_t *, erts_aint32_t);
 
+#ifdef ARCH_32
+extern UWord erts_literal_vspace_map[];
+# define ERTS_VSPACE_WORD_BITS (sizeof(UWord)*8)
+#endif
+
+void* erts_alcu_mseg_alloc(Allctr_t*, Uint *size_p, Uint flags);
+void* erts_alcu_mseg_realloc(Allctr_t*, void *seg, Uint old_size, Uint *new_size_p);
+void  erts_alcu_mseg_dealloc(Allctr_t*, void *seg, Uint size, Uint flags);
+
+#if HAVE_ERTS_MSEG
+# if defined(ARCH_32)
+void* erts_alcu_literal_32_mseg_alloc(Allctr_t*, Uint *size_p, Uint flags);
+void* erts_alcu_literal_32_mseg_realloc(Allctr_t*, void *seg, Uint old_size, Uint *new_size_p);
+void  erts_alcu_literal_32_mseg_dealloc(Allctr_t*, void *seg, Uint size, Uint flags);
+
+# elif defined(ARCH_64) && defined(ERTS_HAVE_OS_PHYSICAL_MEMORY_RESERVATION)
+void* erts_alcu_literal_64_mseg_alloc(Allctr_t*, Uint *size_p, Uint flags);
+void* erts_alcu_literal_64_mseg_realloc(Allctr_t*, void *seg, Uint old_size, Uint *new_size_p);
+void  erts_alcu_literal_64_mseg_dealloc(Allctr_t*, void *seg, Uint size, Uint flags);
+# endif
+#endif /* HAVE_ERTS_MSEG */
+
+void* erts_alcu_sys_alloc(Allctr_t*, Uint size, int superalign);
+void* erts_alcu_sys_realloc(Allctr_t*, void *ptr, Uint size, Uint old_size, int superalign);
+void  erts_alcu_sys_dealloc(Allctr_t*, void *ptr, Uint size, int superalign);
+#ifdef ARCH_32
+void* erts_alcu_literal_32_sys_alloc(Allctr_t*, Uint size, int superalign);
+void* erts_alcu_literal_32_sys_realloc(Allctr_t*, void *ptr, Uint size, Uint old_size, int superalign);
+void  erts_alcu_literal_32_sys_dealloc(Allctr_t*, void *ptr, Uint size, int superalign);
 #endif
 
+#endif /* !ERL_ALLOC_UTIL__ */
+
 #if defined(GET_ERL_ALLOC_UTIL_IMPL) && !defined(ERL_ALLOC_UTIL_IMPL__)
 #define ERL_ALLOC_UTIL_IMPL__
 
 #define ERTS_ALCU_FLG_FAIL_REALLOC_MOVE		(((Uint32) 1) << 0)
 
-#ifdef USE_THREADS
-#define ERL_THREADS_EMU_INTERNAL__
-#include "erl_threads.h"
-#endif
-
-#include "erl_mseg.h"
-
 #undef ERTS_ALLOC_UTIL_HARD_DEBUG
 #ifdef DEBUG
 #  if 0
@@ -223,7 +259,7 @@ erts_aint32_t erts_alcu_fix_alloc_shrink(Allctr_t *, erts_aint32_t);
 
 #if ERTS_HAVE_MSEG_SUPER_ALIGNED \
     || (!HAVE_ERTS_MSEG && ERTS_HAVE_ERTS_SYS_ALIGNED_ALLOC)
-#  ifndef MSEG_ALIGN_BITS
+#  ifdef MSEG_ALIGN_BITS
 #    define ERTS_SUPER_ALIGN_BITS MSEG_ALIGN_BITS
 #  else
 #    define ERTS_SUPER_ALIGN_BITS 18
@@ -277,7 +313,7 @@ typedef struct ErtsDoubleLink_t_ {
 typedef struct {
     erts_atomic_t next;
     erts_atomic_t prev;
-    Allctr_t *orig_allctr;
+    Allctr_t *orig_allctr;      /* read-only while carrier is alive */
     ErtsThrPrgrVal thr_prgr;
     erts_atomic_t max_size;
     UWord abandon_limit;
@@ -501,6 +537,8 @@ struct Allctr_t_ {
     Uint		min_mbc_size;
     Uint		min_mbc_first_free_size;
     Uint		min_block_size;
+    UWord               crr_set_flgs;
+    UWord               crr_clr_flgs;
 
     /* Carriers */
     CarrierList_t	mbc_list;
@@ -546,6 +584,15 @@ struct Allctr_t_ {
     void		(*remove_mbc)	        (Allctr_t *, Carrier_t *);
     UWord		(*largest_fblk_in_mbc)  (Allctr_t *, Carrier_t *);
 
+#if HAVE_ERTS_MSEG
+    void*               (*mseg_alloc)(Allctr_t*, Uint *size_p, Uint flags);
+    void*               (*mseg_realloc)(Allctr_t*, void *seg, Uint old_size, Uint *new_size_p);
+    void                (*mseg_dealloc)(Allctr_t*, void *seg, Uint size, Uint flags);
+#endif
+    void*               (*sys_alloc)(Allctr_t *allctr, Uint size, int superalign);
+    void*               (*sys_realloc)(Allctr_t *allctr, void *ptr, Uint size, Uint old_size, int superalign);
+    void                (*sys_dealloc)(Allctr_t *allctr, void *ptr, Uint size, int superalign);
+
     void		(*init_atoms)		(void);
 
 #ifdef ERTS_ALLOC_UTIL_HARD_DEBUG
diff --git a/erts/emulator/beam/erl_ao_firstfit_alloc.c b/erts/emulator/beam/erl_ao_firstfit_alloc.c
index 7c2a5c3323..19420af8ab 100644
--- a/erts/emulator/beam/erl_ao_firstfit_alloc.c
+++ b/erts/emulator/beam/erl_ao_firstfit_alloc.c
@@ -209,7 +209,9 @@ static Block_t*	aoff_get_free_block(Allctr_t *, Uint, Block_t *, Uint);
 static void aoff_link_free_block(Allctr_t *, Block_t*);
 static void aoff_unlink_free_block(Allctr_t *allctr, Block_t *del);
 static void aoff_creating_mbc(Allctr_t*, Carrier_t*);
+#ifdef DEBUG
 static void aoff_destroying_mbc(Allctr_t*, Carrier_t*);
+#endif
 static void aoff_add_mbc(Allctr_t*, Carrier_t*);
 static void aoff_remove_mbc(Allctr_t*, Carrier_t*);
 static UWord aoff_largest_fblk_in_mbc(Allctr_t*, Carrier_t*);
@@ -271,7 +273,11 @@ erts_aoffalc_start(AOFFAllctr_t *alc,
 
     allctr->get_next_mbc_size		= NULL;
     allctr->creating_mbc		= aoff_creating_mbc;
+#ifdef DEBUG
     allctr->destroying_mbc		= aoff_destroying_mbc;
+#else
+    allctr->destroying_mbc		= NULL;
+#endif
     allctr->add_mbc                     = aoff_add_mbc;
     allctr->remove_mbc                  = aoff_remove_mbc;
     allctr->largest_fblk_in_mbc         = aoff_largest_fblk_in_mbc;
@@ -885,17 +891,18 @@ static void aoff_creating_mbc(Allctr_t *allctr, Carrier_t *carrier)
     HARD_CHECK_TREE(NULL, 0, *root, 0);
 }
 
+#define IS_CRR_IN_TREE(CRR,ROOT) \
+    ((CRR)->rbt_node.parent || (ROOT) == &(CRR)->rbt_node)
+
+#ifdef DEBUG
 static void aoff_destroying_mbc(Allctr_t *allctr, Carrier_t *carrier)
 {
     AOFFAllctr_t *alc = (AOFFAllctr_t *) allctr;
     AOFF_Carrier_t *crr = (AOFF_Carrier_t*) carrier;
-    AOFF_RBTree_t *root = alc->mbc_root;
 
-    if (crr->rbt_node.parent || &crr->rbt_node == root) {
-	aoff_remove_mbc(allctr, carrier);
-    }
-    /*else already removed */
+    ASSERT(!IS_CRR_IN_TREE(crr, alc->mbc_root));
 }
+#endif
 
 static void aoff_add_mbc(Allctr_t *allctr, Carrier_t *carrier)
 {
@@ -903,6 +910,7 @@ static void aoff_add_mbc(Allctr_t *allctr, Carrier_t *carrier)
     AOFF_Carrier_t *crr = (AOFF_Carrier_t*) carrier;
     AOFF_RBTree_t **root = &alc->mbc_root;
 
+    ASSERT(!IS_CRR_IN_TREE(crr, *root));
     HARD_CHECK_TREE(NULL, 0, *root, 0);   
 
     /* Link carrier in address order tree
@@ -919,6 +927,10 @@ static void aoff_remove_mbc(Allctr_t *allctr, Carrier_t *carrier)
     AOFF_RBTree_t **root = &alc->mbc_root;
 
     ASSERT(allctr == ERTS_ALC_CARRIER_TO_ALLCTR(carrier));
+
+    if (!IS_CRR_IN_TREE(crr,*root))
+        return;
+
     HARD_CHECK_TREE(NULL, 0, *root, 0);
 
     rbt_delete(root, &crr->rbt_node);
diff --git a/erts/emulator/beam/erl_arith.c b/erts/emulator/beam/erl_arith.c
index b8c5ef9b09..3671025d22 100644
--- a/erts/emulator/beam/erl_arith.c
+++ b/erts/emulator/beam/erl_arith.c
@@ -42,15 +42,8 @@
 #  define MAX(x, y) (((x) > (y)) ? (x) : (y))
 #endif
 
-#if !HEAP_ON_C_STACK
-#  define DECLARE_TMP(VariableName,N,P)  \
-     Eterm *VariableName = ((ERTS_PROC_GET_SCHDATA(P)->erl_arith_tmp_heap) + (2 * N))
-#else
-#  define DECLARE_TMP(VariableName,N,P) \
-     Eterm VariableName[2]
-#endif
-#  define ARG_IS_NOT_TMP(Arg,Tmp) ((Arg) != make_big((Tmp)))
-
+#define DECLARE_TMP(VariableName,N,P)  Eterm VariableName[2]
+#define ARG_IS_NOT_TMP(Arg,Tmp) ((Arg) != make_big((Tmp)))
 
 static Eterm shift(Process* p, Eterm arg1, Eterm arg2, int right);
 
diff --git a/erts/emulator/beam/erl_async.c b/erts/emulator/beam/erl_async.c
index f071898046..be0bc0cfec 100644
--- a/erts/emulator/beam/erl_async.c
+++ b/erts/emulator/beam/erl_async.c
@@ -167,7 +167,6 @@ async_ready_q(Uint sched_id)
 
 #endif
 
-
 void
 erts_init_async(void)
 {
diff --git a/erts/emulator/beam/erl_bif_binary.c b/erts/emulator/beam/erl_bif_binary.c
index 134aa2d396..aec72bd61a 100644
--- a/erts/emulator/beam/erl_bif_binary.c
+++ b/erts/emulator/beam/erl_bif_binary.c
@@ -55,10 +55,8 @@
 
 /* Init and local variables */
 
-static Export binary_match_trap_export;
-static BIF_RETTYPE binary_match_trap(BIF_ALIST_3);
-static Export binary_matches_trap_export;
-static BIF_RETTYPE binary_matches_trap(BIF_ALIST_3);
+static Export binary_find_trap_export;
+static BIF_RETTYPE binary_find_trap(BIF_ALIST_3);
 static Export binary_longest_prefix_trap_export;
 static BIF_RETTYPE binary_longest_prefix_trap(BIF_ALIST_3);
 static Export binary_longest_suffix_trap_export;
@@ -70,19 +68,15 @@ static BIF_RETTYPE binary_copy_trap(BIF_ALIST_2);
 static Uint max_loop_limit;
 
 static BIF_RETTYPE
-binary_match(Process *p, Eterm arg1, Eterm arg2, Eterm arg3);
+binary_match(Process *p, Eterm arg1, Eterm arg2, Eterm arg3, Uint flags);
 static BIF_RETTYPE
-binary_matches(Process *p, Eterm arg1, Eterm arg2, Eterm arg3);
+binary_split(Process *p, Eterm arg1, Eterm arg2, Eterm arg3);
 
 void erts_init_bif_binary(void)
 {
-    erts_init_trap_export(&binary_match_trap_export,
-			  am_erlang, am_binary_match_trap, 3,
-			  &binary_match_trap);
-
-    erts_init_trap_export(&binary_matches_trap_export,
-			  am_erlang, am_binary_matches_trap, 3,
-			  &binary_matches_trap);
+    erts_init_trap_export(&binary_find_trap_export,
+			  am_erlang, am_binary_find_trap, 3,
+			  &binary_find_trap);
 
     erts_init_trap_export(&binary_longest_prefix_trap_export,
 			  am_erlang, am_binary_longest_prefix_trap, 3,
@@ -314,8 +308,8 @@ static BMData *create_bmdata(MyAllocator *my, byte *x, Uint len,
 /*
  * Aho Corasick - Build a Trie and fill in the failure functions
  * when all strings are added.
- * The algorithm is nicely described by Dieter B�hler of University of
- * T�bingen:
+ * The algorithm is nicely described by Dieter Bühler of University of
+ * Tübingen:
  * http://www-sr.informatik.uni-tuebingen.de/~buehler/AC/AC.html
  */
 
@@ -537,21 +531,23 @@ static void ac_init_find_all(ACFindAllState *state, ACTrie *act, Sint startpos,
     state->out = NULL;
 }
 
-static void ac_restore_find_all(ACFindAllState *state, char *buff)
+static void ac_restore_find_all(ACFindAllState *state,
+                                const ACFindAllState *src)
 {
-    memcpy(state,buff,sizeof(ACFindAllState));
+    memcpy(state, src, sizeof(ACFindAllState));
     if (state->allocated > 0) {
 	state->out = erts_alloc(ERTS_ALC_T_TMP, sizeof(FindallData) * (state->allocated));
-	memcpy(state->out,buff+sizeof(ACFindAllState),sizeof(FindallData)*state->m);
+	memcpy(state->out, src+1, sizeof(FindallData)*state->m);
     } else {
 	state->out = NULL;
     }
 }
 
-static void ac_serialize_find_all(ACFindAllState *state, char *buff)
+static void ac_serialize_find_all(const ACFindAllState *state,
+                                  ACFindAllState *dst)
 {
-    memcpy(buff,state,sizeof(ACFindAllState));
-    memcpy(buff+sizeof(ACFindAllState),state->out,sizeof(FindallData)*state->m);
+    memcpy(dst, state, sizeof(ACFindAllState));
+    memcpy(dst+1, state->out, sizeof(FindallData)*state->m);
 }
 
 static void ac_clean_find_all(ACFindAllState *state)
@@ -565,9 +561,6 @@ static void ac_clean_find_all(ACFindAllState *state)
 #endif
 }
 
-#define SIZEOF_AC_SERIALIZED_FIND_ALL_STATE(S) \
-          (sizeof(ACFindAllState)+(sizeof(FindallData)*(S).m))
-
 /*
  * Differs to the find_first function in that it stores all matches and the values
  * arte returned only in the state.
@@ -814,24 +807,24 @@ static void bm_init_find_all(BMFindAllState *state, Sint startpos, Uint len)
     state->out = NULL;
 }
 
-static void bm_restore_find_all(BMFindAllState *state, char *buff)
+static void bm_restore_find_all(BMFindAllState *state,
+                                const BMFindAllState *src)
 {
-    memcpy(state,buff,sizeof(BMFindAllState));
+    memcpy(state, src, sizeof(BMFindAllState));
     if (state->allocated > 0) {
 	state->out = erts_alloc(ERTS_ALC_T_TMP, sizeof(FindallData) *
 				(state->allocated));
-	memcpy(state->out,buff+sizeof(BMFindAllState),
-	       sizeof(FindallData)*state->m);
+	memcpy(state->out, src+1, sizeof(FindallData)*state->m);
     } else {
 	state->out = NULL;
     }
 }
 
-static void bm_serialize_find_all(BMFindAllState *state, char *buff)
+static void bm_serialize_find_all(const BMFindAllState *state,
+                                  BMFindAllState *dst)
 {
-    memcpy(buff,state,sizeof(BMFindAllState));
-    memcpy(buff+sizeof(BMFindAllState),state->out,
-	   sizeof(FindallData)*state->m);
+    memcpy(dst, state, sizeof(BMFindAllState));
+    memcpy(dst+1, state->out, sizeof(FindallData)*state->m);
 }
 
 static void bm_clean_find_all(BMFindAllState *state)
@@ -845,9 +838,6 @@ static void bm_clean_find_all(BMFindAllState *state)
 #endif
 }
 
-#define SIZEOF_BM_SERIALIZED_FIND_ALL_STATE(S) \
-          (sizeof(BMFindAllState)+(sizeof(FindallData)*(S).m))
-
 /*
  * Differs to the find_first function in that it stores all matches and the
  * values are returned only in the state.
@@ -1030,283 +1020,142 @@ BIF_RETTYPE binary_compile_pattern_1(BIF_ALIST_1)
 #define DO_BIN_MATCH_BADARG -1
 #define DO_BIN_MATCH_RESTART -2
 
-static int do_binary_match(Process *p, Eterm subject, Uint hsstart, Uint hsend,
-			   Eterm type, Binary *bin, Eterm state_term,
-			   Eterm *res_term)
-{
-    byte *bytes;
-    Uint bitoffs, bitsize;
-    byte *temp_alloc = NULL;
-
-    ERTS_GET_BINARY_BYTES(subject, bytes, bitoffs, bitsize);
-    if (bitsize != 0) {
-	goto badarg;
-    }
-    if (bitoffs != 0) {
-	bytes = erts_get_aligned_binary_bytes(subject, &temp_alloc);
-    }
-    if (state_term != NIL) {
-	Eterm *ptr = big_val(state_term);
-	type = ptr[1];
-    }
-
-    if (type == am_bm) {
-	BMData *bm;
-	Sint pos;
-	Eterm ret;
-	Eterm *hp;
-	BMFindFirstState state;
-	Uint reds = get_reds(p, BM_LOOP_FACTOR);
-	Uint save_reds = reds;
-
-	bm = (BMData *) ERTS_MAGIC_BIN_DATA(bin);
-#ifdef HARDDEBUG
-	dump_bm_data(bm);
-#endif
-	if (state_term == NIL) {
-	    bm_init_find_first_match(&state, hsstart, hsend);
-	} else {
-	    Eterm *ptr = big_val(state_term);
-	    memcpy(&state,ptr+2,sizeof(state));
-	}
-#ifdef HARDDEBUG
-	erts_printf("(bm) state->pos = %ld, state->len = %lu\n",state.pos,
-		    state.len);
-#endif
-	pos = bm_find_first_match(&state, bm, bytes, &reds);
-	if (pos == BM_NOT_FOUND) {
-	    ret = am_nomatch;
-	} else if (pos == BM_RESTART) {
-	    int x = (sizeof(BMFindFirstState) / sizeof(Eterm)) +
-		!!(sizeof(BMFindFirstState) % sizeof(Eterm));
-#ifdef HARDDEBUG
-	    erts_printf("Trap bm!\n");
-#endif
-	    hp = HAlloc(p,x+2);
-	    hp[0] = make_pos_bignum_header(x+1);
-	    hp[1] = type;
-	    memcpy(hp+2,&state,sizeof(state));
-	    *res_term = make_big(hp);
-	    erts_free_aligned_binary_bytes(temp_alloc);
-	    return DO_BIN_MATCH_RESTART;
-	} else {
-	    Eterm erlen = erts_make_integer((Uint) bm->len, p);
-	    ret = erts_make_integer(pos,p);
-	    hp = HAlloc(p,3);
-	    ret = TUPLE2(hp, ret, erlen);
-	}
-	erts_free_aligned_binary_bytes(temp_alloc);
-	BUMP_REDS(p, (save_reds - reds) / BM_LOOP_FACTOR);
-	*res_term = ret;
-	return DO_BIN_MATCH_OK;
-    } else if (type == am_ac) {
-	ACTrie *act;
-	Uint pos, rlen;
-	int acr;
-	ACFindFirstState state;
-	Eterm ret;
-	Eterm *hp;
-	Uint reds = get_reds(p, AC_LOOP_FACTOR);
-	Uint save_reds = reds;
+#define BINARY_FIND_ALL		0x01
+#define BINARY_SPLIT_TRIM	0x02
+#define BINARY_SPLIT_TRIM_ALL	0x04
 
-	act = (ACTrie *) ERTS_MAGIC_BIN_DATA(bin);
-#ifdef HARDDEBUG
-	dump_ac_trie(act);
-#endif
-	if (state_term == NIL) {
-	    ac_init_find_first_match(&state, act, hsstart, hsend);
-	} else {
-	    Eterm *ptr = big_val(state_term);
-	    memcpy(&state,ptr+2,sizeof(state));
-	}
-	acr = ac_find_first_match(&state, bytes, &pos, &rlen, &reds);
-	if (acr == AC_NOT_FOUND) {
-	    ret = am_nomatch;
-	} else if (acr == AC_RESTART) {
-	    int x = (sizeof(state) / sizeof(Eterm)) +
-		!!(sizeof(ACFindFirstState) % sizeof(Eterm));
-#ifdef HARDDEBUG
-	    erts_printf("Trap ac!\n");
-#endif
-	    hp = HAlloc(p,x+2);
-	    hp[0] = make_pos_bignum_header(x+1);
-	    hp[1] = type;
-	    memcpy(hp+2,&state,sizeof(state));
-	    *res_term = make_big(hp);
-	    erts_free_aligned_binary_bytes(temp_alloc);
-	    return DO_BIN_MATCH_RESTART;
-	} else {
-	    Eterm epos = erts_make_integer(pos,p);
-	    Eterm erlen = erts_make_integer(rlen,p);
-	    hp = HAlloc(p,3);
-	    ret = TUPLE2(hp, epos, erlen);
-	}
-	erts_free_aligned_binary_bytes(temp_alloc);
-	BUMP_REDS(p, (save_reds - reds) / AC_LOOP_FACTOR);
-	*res_term = ret;
-	return DO_BIN_MATCH_OK;
-    }
- badarg:
-    return DO_BIN_MATCH_BADARG;
-}
+typedef struct BinaryFindState {
+    Eterm type;
+    Uint  flags;
+    Uint  hsstart;
+    Uint  hsend;
+    Eterm (*not_found_result) (Process *, Eterm, struct BinaryFindState *);
+    Eterm (*single_result)    (Process *, Eterm, struct BinaryFindState *, Sint, Sint);
+    Eterm (*global_result)    (Process *, Eterm, struct BinaryFindState *, FindallData *, Uint);
+} BinaryFindState;
+
+typedef struct BinaryFindState_bignum {
+    Eterm           bignum_hdr;
+    BinaryFindState bfs;
+    union {
+	BMFindFirstState bmffs;
+	BMFindAllState   bmfas;
+	ACFindFirstState acffs;
+	ACFindAllState   acfas;
+    } data;
+} BinaryFindState_bignum;
+
+#define SIZEOF_BINARY_FIND_STATE(S) \
+	  (sizeof(BinaryFindState)+sizeof(S))
+
+#define SIZEOF_BINARY_FIND_ALL_STATE(S) \
+	  (sizeof(BinaryFindState)+sizeof(S)+(sizeof(FindallData)*(S).m))
+
+static Eterm do_match_not_found_result(Process *p, Eterm subject, BinaryFindState *bfs);
+static Eterm do_match_single_result(Process *p, Eterm subject, BinaryFindState *bfs,
+				    Sint pos, Sint len);
+static Eterm do_match_global_result(Process *p, Eterm subject, BinaryFindState *bfs,
+				    FindallData *fad, Uint fad_sz);
+static Eterm do_split_not_found_result(Process *p, Eterm subject, BinaryFindState *bfs);
+static Eterm do_split_single_result(Process *p, Eterm subject, BinaryFindState *bfs,
+				    Sint pos, Sint len);
+static Eterm do_split_global_result(Process *p, Eterm subject, BinaryFindState *bfs,
+				    FindallData *fad, Uint fad_sz);
 
-static int do_binary_matches(Process *p, Eterm subject, Uint hsstart,
-			     Uint hsend, Eterm type, Binary *bin,
-			     Eterm state_term, Eterm *res_term)
+static int parse_match_opts_list(Eterm l, Eterm bin, Uint *posp, Uint *endp)
 {
-    byte *bytes;
-    Uint bitoffs, bitsize;
-    byte *temp_alloc = NULL;
-
-    ERTS_GET_BINARY_BYTES(subject, bytes, bitoffs, bitsize);
-    if (bitsize != 0) {
-	goto badarg;
-    }
-    if (bitoffs != 0) {
-	bytes = erts_get_aligned_binary_bytes(subject, &temp_alloc);
-    }
-    if (state_term != NIL) {
-	Eterm *ptr = big_val(state_term);
-	type = ptr[1];
-    }
-
-    if (type == am_bm) {
-	BMData *bm;
-	Sint pos;
-	Eterm ret,tpl;
-	Eterm *hp;
-	BMFindAllState state;
-	Uint reds = get_reds(p, BM_LOOP_FACTOR);
-	Uint save_reds = reds;
-
-	bm = (BMData *) ERTS_MAGIC_BIN_DATA(bin);
-#ifdef HARDDEBUG
-	dump_bm_data(bm);
-#endif
-	if (state_term == NIL) {
-	    bm_init_find_all(&state, hsstart, hsend);
-	} else {
-	    Eterm *ptr = big_val(state_term);
-	    bm_restore_find_all(&state,(char *) (ptr+2));
-	}
-
-	pos = bm_find_all_non_overlapping(&state, bm, bytes, &reds);
-	if (pos == BM_NOT_FOUND) {
-	    ret = NIL;
-	} else if (pos == BM_RESTART) {
-	    int x =
-		(SIZEOF_BM_SERIALIZED_FIND_ALL_STATE(state) / sizeof(Eterm)) +
-		!!(SIZEOF_BM_SERIALIZED_FIND_ALL_STATE(state) % sizeof(Eterm));
-#ifdef HARDDEBUG
-	    erts_printf("Trap bm!\n");
-#endif
-	    hp = HAlloc(p,x+2);
-	    hp[0] = make_pos_bignum_header(x+1);
-	    hp[1] = type;
-	    bm_serialize_find_all(&state, (char *) (hp+2));
-	    *res_term = make_big(hp);
-	    erts_free_aligned_binary_bytes(temp_alloc);
-	    bm_clean_find_all(&state);
-	    return DO_BIN_MATCH_RESTART;
-	} else {
-	    FindallData *fad = state.out;
-	    int i;
-	    for (i = 0; i < state.m; ++i) {
-		fad[i].epos = erts_make_integer(fad[i].pos,p);
-		fad[i].elen = erts_make_integer(fad[i].len,p);
+    Eterm *tp;
+    Uint pos;
+    Sint len;
+    if (l == THE_NON_VALUE || l == NIL) {
+	/* Invalid term or NIL, we're called from binary_match(es)_2 or
+	   have no options*/
+	*posp = 0;
+	*endp = binary_size(bin);
+	return 0;
+    } else if (is_list(l)) {
+	while(is_list(l)) {
+	    Eterm t = CAR(list_val(l));
+	    Uint orig_size;
+	    if (!is_tuple(t)) {
+		goto badarg;
 	    }
-	    hp = HAlloc(p,state.m * (3 + 2));
-	    ret = NIL;
-	    for (i = state.m - 1; i >= 0; --i) {
-		tpl = TUPLE2(hp, fad[i].epos, fad[i].elen);
-		hp +=3;
-		ret = CONS(hp,tpl,ret);
-		hp += 2;
+	    tp = tuple_val(t);
+	    if (arityval(*tp) != 2) {
+		goto badarg;
 	    }
-	}
-	erts_free_aligned_binary_bytes(temp_alloc);
-	bm_clean_find_all(&state);
-	BUMP_REDS(p, (save_reds - reds) / BM_LOOP_FACTOR);
-	*res_term = ret;
-	return DO_BIN_MATCH_OK;
-    } else if (type == am_ac) {
-	ACTrie *act;
-	int acr;
-	ACFindAllState state;
-	Eterm ret,tpl;
-	Eterm *hp;
-	Uint reds = get_reds(p, AC_LOOP_FACTOR);
-	Uint save_reds = reds;
-
-	act = (ACTrie *) ERTS_MAGIC_BIN_DATA(bin);
-#ifdef HARDDEBUG
-	dump_ac_trie(act);
-#endif
-	if (state_term == NIL) {
-	    ac_init_find_all(&state, act, hsstart, hsend);
-	} else {
-	    Eterm *ptr = big_val(state_term);
-	    ac_restore_find_all(&state,(char *) (ptr+2));
-	}
-	acr = ac_find_all_non_overlapping(&state, bytes, &reds);
-	if (acr == AC_NOT_FOUND) {
-	    ret = NIL;
-	} else if (acr == AC_RESTART) {
-	    int x =
-		(SIZEOF_AC_SERIALIZED_FIND_ALL_STATE(state) / sizeof(Eterm)) +
-		!!(SIZEOF_AC_SERIALIZED_FIND_ALL_STATE(state) % sizeof(Eterm));
-#ifdef HARDDEBUG
-	    erts_printf("Trap ac!\n");
-#endif
-	    hp = HAlloc(p,x+2);
-	    hp[0] = make_pos_bignum_header(x+1);
-	    hp[1] = type;
-	    ac_serialize_find_all(&state, (char *) (hp+2));
-	    *res_term = make_big(hp);
-	    erts_free_aligned_binary_bytes(temp_alloc);
-	    ac_clean_find_all(&state);
-	    return DO_BIN_MATCH_RESTART;
-	} else {
-	    FindallData *fad = state.out;
-	    int i;
-	    for (i = 0; i < state.m; ++i) {
-		fad[i].epos = erts_make_integer(fad[i].pos,p);
-		fad[i].elen = erts_make_integer(fad[i].len,p);
+	    if (tp[1] != am_scope || is_not_tuple(tp[2])) {
+		goto badarg;
 	    }
-	    hp = HAlloc(p,state.m * (3 + 2));
-	    ret = NIL;
-	    for (i = state.m - 1; i >= 0; --i) {
-		tpl = TUPLE2(hp, fad[i].epos, fad[i].elen);
-		hp +=3;
-		ret = CONS(hp,tpl,ret);
-		hp += 2;
+	    tp = tuple_val(tp[2]);
+	    if (arityval(*tp) != 2) {
+		goto badarg;
+	    }
+	    if (!term_to_Uint(tp[1], &pos)) {
+		goto badarg;
 	    }
+	    if (!term_to_Sint(tp[2], &len)) {
+		goto badarg;
+	    }
+	    if (len < 0) {
+		Uint lentmp = -(Uint)len;
+		/* overflow */
+		if ((Sint)lentmp < 0) {
+		    goto badarg;
+		}
+		len = lentmp;
+		pos -= len;
+	    }
+	    /* overflow */
+	    if ((pos + len) < pos || (len > 0 && (pos + len) == pos)) {
+		goto badarg;
+	    }
+	    *endp = len + pos;
+	    *posp = pos;
+	    if ((orig_size = binary_size(bin)) < pos ||
+		orig_size < (*endp)) {
+		goto badarg;
+	    }
+	    l = CDR(list_val(l));
 	}
-	erts_free_aligned_binary_bytes(temp_alloc);
-	ac_clean_find_all(&state);
-	BUMP_REDS(p, (save_reds - reds) / AC_LOOP_FACTOR);
-	*res_term = ret;
-	return DO_BIN_MATCH_OK;
+	return 0;
+    } else {
+    badarg:
+	return 1;
     }
- badarg:
-    return DO_BIN_MATCH_BADARG;
 }
 
-static int parse_match_opts_list(Eterm l, Eterm bin, Uint *posp, Uint *endp)
+static int parse_split_opts_list(Eterm l, Eterm bin, Uint *posp, Uint *endp, Uint *optp)
 {
     Eterm *tp;
     Uint pos;
     Sint len;
-    if (l == ((Eterm) 0) || l == NIL) {
-	/* Invalid term or NIL, we're called from binary_match(es)_2 or
-	   have no options*/
-	*posp = 0;
-	*endp = binary_size(bin);
+    *optp = 0;
+    *posp = 0;
+    *endp = binary_size(bin);
+    if (l == THE_NON_VALUE || l == NIL) {
 	return 0;
     } else if (is_list(l)) {
 	while(is_list(l)) {
 	    Eterm t = CAR(list_val(l));
 	    Uint orig_size;
+	    if (is_atom(t)) {
+		if (t == am_global) {
+		    *optp |= BINARY_FIND_ALL;
+		    l = CDR(list_val(l));
+		    continue;
+		}
+		if (t == am_trim) {
+		    *optp |= BINARY_SPLIT_TRIM;
+		    l = CDR(list_val(l));
+		    continue;
+		}
+		if (t == am_trim_all) {
+		    *optp |= BINARY_SPLIT_TRIM_ALL;
+		    l = CDR(list_val(l));
+		    continue;
+		}
+	    }
 	    if (!is_tuple(t)) {
 		goto badarg;
 	    }
@@ -1355,48 +1204,207 @@ static int parse_match_opts_list(Eterm l, Eterm bin, Uint *posp, Uint *endp)
     }
 }
 
-static BIF_RETTYPE binary_match_trap(BIF_ALIST_3)
+static int do_binary_find(Process *p, Eterm subject, BinaryFindState *bfs, Binary *bin,
+			  Eterm state_term, Eterm *res_term)
 {
-    int runres;
-    Eterm result;
-    Binary *bin = ((ProcBin *) binary_val(BIF_ARG_3))->val;
-    runres = do_binary_match(BIF_P,BIF_ARG_1,0,0,NIL,bin,BIF_ARG_2,&result);
-    if (runres == DO_BIN_MATCH_OK) {
-	BIF_RET(result);
-    } else {
-	BUMP_ALL_REDS(BIF_P);
-	BIF_TRAP3(&binary_match_trap_export, BIF_P, BIF_ARG_1, result,
-		  BIF_ARG_3);
+    byte *bytes;
+    Uint bitoffs, bitsize;
+    byte *temp_alloc = NULL;
+    BinaryFindState_bignum *state_ptr = NULL;
+
+    ERTS_GET_BINARY_BYTES(subject, bytes, bitoffs, bitsize);
+    if (bitsize != 0) {
+	goto badarg;
+    }
+    if (bitoffs != 0) {
+	bytes = erts_get_aligned_binary_bytes(subject, &temp_alloc);
+    }
+    if (state_term != NIL) {
+	state_ptr = (BinaryFindState_bignum *)(big_val(state_term));
+	bfs = &(state_ptr->bfs);
     }
-}
 
-static BIF_RETTYPE binary_matches_trap(BIF_ALIST_3)
-{
-    int runres;
-    Eterm result;
-    Binary *bin = ((ProcBin *) binary_val(BIF_ARG_3))->val;
-    runres = do_binary_matches(BIF_P,BIF_ARG_1,0,0,NIL,bin,BIF_ARG_2,&result);
-    if (runres == DO_BIN_MATCH_OK) {
-	BIF_RET(result);
+    if (bfs->flags & BINARY_FIND_ALL) {
+	if (bfs->type == am_bm) {
+	    BMData *bm;
+	    Sint pos;
+	    BMFindAllState state;
+	    Uint reds = get_reds(p, BM_LOOP_FACTOR);
+	    Uint save_reds = reds;
+
+	    bm = (BMData *) ERTS_MAGIC_BIN_DATA(bin);
+#ifdef HARDDEBUG
+	    dump_bm_data(bm);
+#endif
+	    if (state_term == NIL) {
+		bm_init_find_all(&state, bfs->hsstart, bfs->hsend);
+	    } else {
+		bm_restore_find_all(&state, &(state_ptr->data.bmfas));
+	    }
+
+	    pos = bm_find_all_non_overlapping(&state, bm, bytes, &reds);
+	    if (pos == BM_NOT_FOUND) {
+		*res_term = bfs->not_found_result(p, subject, bfs);
+	    } else if (pos == BM_RESTART) {
+		int x =
+		    (SIZEOF_BINARY_FIND_ALL_STATE(state) / sizeof(Eterm)) +
+		    !!(SIZEOF_BINARY_FIND_ALL_STATE(state) % sizeof(Eterm));
+#ifdef HARDDEBUG
+		erts_printf("Trap bm!\n");
+#endif
+		state_ptr = (BinaryFindState_bignum*) HAlloc(p, x+1);
+		state_ptr->bignum_hdr = make_pos_bignum_header(x);
+		memcpy(&state_ptr->bfs, bfs, sizeof(BinaryFindState));
+		bm_serialize_find_all(&state, &state_ptr->data.bmfas);
+		*res_term = make_big(&state_ptr->bignum_hdr);
+		erts_free_aligned_binary_bytes(temp_alloc);
+		bm_clean_find_all(&state);
+		return DO_BIN_MATCH_RESTART;
+	    } else {
+		*res_term = bfs->global_result(p, subject, bfs, state.out, state.m);
+	    }
+	    erts_free_aligned_binary_bytes(temp_alloc);
+	    bm_clean_find_all(&state);
+	    BUMP_REDS(p, (save_reds - reds) / BM_LOOP_FACTOR);
+	    return DO_BIN_MATCH_OK;
+	} else if (bfs->type == am_ac) {
+	    ACTrie *act;
+	    int acr;
+	    ACFindAllState state;
+	    Uint reds = get_reds(p, AC_LOOP_FACTOR);
+	    Uint save_reds = reds;
+
+	    act = (ACTrie *) ERTS_MAGIC_BIN_DATA(bin);
+#ifdef HARDDEBUG
+	    dump_ac_trie(act);
+#endif
+	    if (state_term == NIL) {
+		ac_init_find_all(&state, act, bfs->hsstart, bfs->hsend);
+	    } else {
+		ac_restore_find_all(&state, &(state_ptr->data.acfas));
+	    }
+	    acr = ac_find_all_non_overlapping(&state, bytes, &reds);
+	    if (acr == AC_NOT_FOUND) {
+		*res_term = bfs->not_found_result(p, subject, bfs);
+	    } else if (acr == AC_RESTART) {
+		int x =
+		    (SIZEOF_BINARY_FIND_ALL_STATE(state) / sizeof(Eterm)) +
+		    !!(SIZEOF_BINARY_FIND_ALL_STATE(state) % sizeof(Eterm));
+#ifdef HARDDEBUG
+		erts_printf("Trap ac!\n");
+#endif
+                state_ptr = (BinaryFindState_bignum*) HAlloc(p, x+1);
+		state_ptr->bignum_hdr = make_pos_bignum_header(x);
+		memcpy(&state_ptr->bfs, bfs, sizeof(BinaryFindState));
+		ac_serialize_find_all(&state, &state_ptr->data.acfas);
+		*res_term = make_big(&state_ptr->bignum_hdr);
+		erts_free_aligned_binary_bytes(temp_alloc);
+		ac_clean_find_all(&state);
+		return DO_BIN_MATCH_RESTART;
+	    } else {
+		*res_term = bfs->global_result(p, subject, bfs, state.out, state.m);
+	    }
+	    erts_free_aligned_binary_bytes(temp_alloc);
+	    ac_clean_find_all(&state);
+	    BUMP_REDS(p, (save_reds - reds) / AC_LOOP_FACTOR);
+	    return DO_BIN_MATCH_OK;
+	}
     } else {
-	BUMP_ALL_REDS(BIF_P);
-	BIF_TRAP3(&binary_matches_trap_export, BIF_P, BIF_ARG_1, result,
-		  BIF_ARG_3);
-    }
-}
+	if (bfs->type == am_bm) {
+	    BMData *bm;
+	    Sint pos;
+	    BMFindFirstState state;
+	    Uint reds = get_reds(p, BM_LOOP_FACTOR);
+	    Uint save_reds = reds;
+
+	    bm = (BMData *) ERTS_MAGIC_BIN_DATA(bin);
+#ifdef HARDDEBUG
+	    dump_bm_data(bm);
+#endif
+	    if (state_term == NIL) {
+		bm_init_find_first_match(&state, bfs->hsstart, bfs->hsend);
+	    } else {
+		memcpy(&state, &state_ptr->data.bmffs, sizeof(BMFindFirstState));
+	    }
 
-BIF_RETTYPE binary_match_3(BIF_ALIST_3)
-{
-    return binary_match(BIF_P, BIF_ARG_1, BIF_ARG_2, BIF_ARG_3);
+#ifdef HARDDEBUG
+	    erts_printf("(bm) state->pos = %ld, state->len = %lu\n",state.pos,
+			state.len);
+#endif
+	    pos = bm_find_first_match(&state, bm, bytes, &reds);
+	    if (pos == BM_NOT_FOUND) {
+		*res_term = bfs->not_found_result(p, subject, bfs);
+	    } else if (pos == BM_RESTART) {
+		int x =
+		    (SIZEOF_BINARY_FIND_STATE(state) / sizeof(Eterm)) +
+		    !!(SIZEOF_BINARY_FIND_STATE(state) % sizeof(Eterm));
+#ifdef HARDDEBUG
+		erts_printf("Trap bm!\n");
+#endif
+                state_ptr = (BinaryFindState_bignum*) HAlloc(p, x+1);
+                state_ptr->bignum_hdr = make_pos_bignum_header(x);
+		memcpy(&state_ptr->bfs, bfs, sizeof(BinaryFindState));
+		memcpy(&state_ptr->data.acffs, &state, sizeof(BMFindFirstState));
+		*res_term = make_big(&state_ptr->bignum_hdr);
+		erts_free_aligned_binary_bytes(temp_alloc);
+		return DO_BIN_MATCH_RESTART;
+	    } else {
+		*res_term = bfs->single_result(p, subject, bfs, pos, bm->len);
+	    }
+	    erts_free_aligned_binary_bytes(temp_alloc);
+	    BUMP_REDS(p, (save_reds - reds) / BM_LOOP_FACTOR);
+	    return DO_BIN_MATCH_OK;
+	} else if (bfs->type == am_ac) {
+	    ACTrie *act;
+	    Uint pos, rlen;
+	    int acr;
+	    ACFindFirstState state;
+	    Uint reds = get_reds(p, AC_LOOP_FACTOR);
+	    Uint save_reds = reds;
+
+	    act = (ACTrie *) ERTS_MAGIC_BIN_DATA(bin);
+#ifdef HARDDEBUG
+	    dump_ac_trie(act);
+#endif
+	    if (state_term == NIL) {
+		ac_init_find_first_match(&state, act, bfs->hsstart, bfs->hsend);
+	    } else {
+		memcpy(&state, &state_ptr->data.acffs, sizeof(ACFindFirstState));
+	    }
+	    acr = ac_find_first_match(&state, bytes, &pos, &rlen, &reds);
+	    if (acr == AC_NOT_FOUND) {
+		*res_term = bfs->not_found_result(p, subject, bfs);
+	    } else if (acr == AC_RESTART) {
+		int x =
+		    (SIZEOF_BINARY_FIND_STATE(state) / sizeof(Eterm)) +
+		    !!(SIZEOF_BINARY_FIND_STATE(state) % sizeof(Eterm));
+#ifdef HARDDEBUG
+		erts_printf("Trap ac!\n");
+#endif
+		state_ptr = (BinaryFindState_bignum*) HAlloc(p, x+1);
+		state_ptr->bignum_hdr = make_pos_bignum_header(x);
+		memcpy(&state_ptr->bfs, bfs, sizeof(BinaryFindState));
+		memcpy(&state_ptr->data.acffs, &state, sizeof(ACFindFirstState));
+		*res_term = make_big(&state_ptr->bignum_hdr);
+		erts_free_aligned_binary_bytes(temp_alloc);
+		return DO_BIN_MATCH_RESTART;
+	    } else {
+		*res_term = bfs->single_result(p, subject, bfs, pos, rlen);
+	    }
+	    erts_free_aligned_binary_bytes(temp_alloc);
+	    BUMP_REDS(p, (save_reds - reds) / AC_LOOP_FACTOR);
+	    return DO_BIN_MATCH_OK;
+	}
+    }
+ badarg:
+    return DO_BIN_MATCH_BADARG;
 }
 
 static BIF_RETTYPE
-binary_match(Process *p, Eterm arg1, Eterm arg2, Eterm arg3)
+binary_match(Process *p, Eterm arg1, Eterm arg2, Eterm arg3, Uint flags)
 {
-    Uint hsstart;
-    Uint hsend;
+    BinaryFindState bfs;
     Eterm *tp;
-    Eterm type;
     Binary *bin;
     Eterm bin_term = NIL;
     int runres;
@@ -1405,11 +1413,12 @@ binary_match(Process *p, Eterm arg1, Eterm arg2, Eterm arg3)
     if (is_not_binary(arg1)) {
 	goto badarg;
     }
-    if (parse_match_opts_list(arg3,arg1,&hsstart,&hsend)) {
+    bfs.flags = flags;
+    if (parse_match_opts_list(arg3, arg1, &(bfs.hsstart), &(bfs.hsend))) {
 	goto badarg;
     }
-    if (hsend == 0) {
-	BIF_RET(am_nomatch);
+    if (bfs.hsend == 0) {
+	BIF_RET(do_match_not_found_result(p, arg1, &bfs));
     }
     if (is_tuple(arg2)) {
 	tp = tuple_val(arg2);
@@ -1420,21 +1429,24 @@ binary_match(Process *p, Eterm arg1, Eterm arg2, Eterm arg3)
 	    !ERTS_TERM_IS_MAGIC_BINARY(tp[2])) {
 	    goto badarg;
 	}
-	type = tp[1];
+	bfs.type = tp[1];
 	bin = ((ProcBin *) binary_val(tp[2]))->val;
-	if (type == am_bm &&
+	if (bfs.type == am_bm &&
 	    ERTS_MAGIC_BIN_DESTRUCTOR(bin) != cleanup_my_data_bm) {
 	    goto badarg;
 	}
-	if (type == am_ac &&
+	if (bfs.type == am_ac &&
 	    ERTS_MAGIC_BIN_DESTRUCTOR(bin) != cleanup_my_data_ac) {
 	    goto badarg;
 	}
 	bin_term = tp[2];
-    } else if (do_binary_match_compile(arg2,&type,&bin)) {
+    } else if (do_binary_match_compile(arg2, &(bfs.type), &bin)) {
 	goto badarg;
     }
-    runres = do_binary_match(p,arg1,hsstart,hsend,type,bin,NIL,&result);
+    bfs.not_found_result = &do_match_not_found_result;
+    bfs.single_result = &do_match_single_result;
+    bfs.global_result = &do_match_global_result;
+    runres = do_binary_find(p, arg1, &bfs, bin, NIL, &result);
     if (runres == DO_BIN_MATCH_RESTART && bin_term == NIL) {
 	Eterm *hp = HAlloc(p, PROC_BIN_SIZE);
 	bin_term = erts_mk_magic_binary_term(&hp, &MSO(p), bin);
@@ -1446,7 +1458,7 @@ binary_match(Process *p, Eterm arg1, Eterm arg2, Eterm arg3)
 	BIF_RET(result);
     case DO_BIN_MATCH_RESTART:
 	BUMP_ALL_REDS(p);
-	BIF_TRAP3(&binary_match_trap_export, p, arg1, result, bin_term);
+	BIF_TRAP3(&binary_find_trap_export, p, arg1, result, bin_term);
     default:
 	goto badarg;
     }
@@ -1454,17 +1466,31 @@ binary_match(Process *p, Eterm arg1, Eterm arg2, Eterm arg3)
     BIF_ERROR(p,BADARG);
 }
 
+BIF_RETTYPE binary_match_2(BIF_ALIST_2)
+{
+    return binary_match(BIF_P, BIF_ARG_1, BIF_ARG_2, THE_NON_VALUE, 0);
+}
+
+BIF_RETTYPE binary_match_3(BIF_ALIST_3)
+{
+    return binary_match(BIF_P, BIF_ARG_1, BIF_ARG_2, BIF_ARG_3, 0);
+}
+
+BIF_RETTYPE binary_matches_2(BIF_ALIST_2)
+{
+    return binary_match(BIF_P, BIF_ARG_1, BIF_ARG_2, THE_NON_VALUE, BINARY_FIND_ALL);
+}
+
 BIF_RETTYPE binary_matches_3(BIF_ALIST_3)
 {
-    return binary_matches(BIF_P, BIF_ARG_1, BIF_ARG_2, BIF_ARG_3);
+    return binary_match(BIF_P, BIF_ARG_1, BIF_ARG_2, BIF_ARG_3, BINARY_FIND_ALL);
 }
 
 static BIF_RETTYPE
-binary_matches(Process *p, Eterm arg1, Eterm arg2, Eterm arg3)
+binary_split(Process *p, Eterm arg1, Eterm arg2, Eterm arg3)
 {
-    Uint hsstart, hsend;
+    BinaryFindState bfs;
     Eterm *tp;
-    Eterm type;
     Binary *bin;
     Eterm bin_term = NIL;
     int runres;
@@ -1473,11 +1499,12 @@ binary_matches(Process *p, Eterm arg1, Eterm arg2, Eterm arg3)
     if (is_not_binary(arg1)) {
 	goto badarg;
     }
-    if (parse_match_opts_list(arg3,arg1,&hsstart,&hsend)) {
+    if (parse_split_opts_list(arg3, arg1, &(bfs.hsstart), &(bfs.hsend), &(bfs.flags))) {
 	goto badarg;
     }
-    if (hsend == 0) {
-	BIF_RET(NIL);
+    if (bfs.hsend == 0) {
+	result = do_split_not_found_result(p, arg1, &bfs);
+	BIF_RET(result);
     }
     if (is_tuple(arg2)) {
 	tp = tuple_val(arg2);
@@ -1488,54 +1515,267 @@ binary_matches(Process *p, Eterm arg1, Eterm arg2, Eterm arg3)
 	    !ERTS_TERM_IS_MAGIC_BINARY(tp[2])) {
 	    goto badarg;
 	}
-	type = tp[1];
+	bfs.type = tp[1];
 	bin = ((ProcBin *) binary_val(tp[2]))->val;
-	if (type == am_bm &&
+	if (bfs.type == am_bm &&
 	    ERTS_MAGIC_BIN_DESTRUCTOR(bin) != cleanup_my_data_bm) {
 	    goto badarg;
 	}
-	if (type == am_ac &&
+	if (bfs.type == am_ac &&
 	    ERTS_MAGIC_BIN_DESTRUCTOR(bin) != cleanup_my_data_ac) {
 	    goto badarg;
 	}
 	bin_term = tp[2];
-    } else if (do_binary_match_compile(arg2,&type,&bin)) {
+    } else if (do_binary_match_compile(arg2, &(bfs.type), &bin)) {
 	goto badarg;
     }
-    runres = do_binary_matches(p,arg1,hsstart,hsend,type,bin,
-			       NIL,&result);
+    bfs.not_found_result = &do_split_not_found_result;
+    bfs.single_result = &do_split_single_result;
+    bfs.global_result = &do_split_global_result;
+    runres = do_binary_find(p, arg1, &bfs, bin, NIL, &result);
     if (runres == DO_BIN_MATCH_RESTART && bin_term == NIL) {
 	Eterm *hp = HAlloc(p, PROC_BIN_SIZE);
 	bin_term = erts_mk_magic_binary_term(&hp, &MSO(p), bin);
     } else if (bin_term == NIL) {
 	erts_bin_free(bin);
     }
-    switch (runres) {
+    switch(runres) {
     case DO_BIN_MATCH_OK:
 	BIF_RET(result);
     case DO_BIN_MATCH_RESTART:
-	BUMP_ALL_REDS(p);
-	BIF_TRAP3(&binary_matches_trap_export, p, arg1, result,
-		  bin_term);
+	BIF_TRAP3(&binary_find_trap_export, p, arg1, result, bin_term);
     default:
 	goto badarg;
     }
  badarg:
-    BIF_ERROR(p,BADARG);
+    BIF_ERROR(p, BADARG);
 }
 
+BIF_RETTYPE binary_split_2(BIF_ALIST_2)
+{
+    return binary_split(BIF_P, BIF_ARG_1, BIF_ARG_2, THE_NON_VALUE);
+}
 
-BIF_RETTYPE binary_match_2(BIF_ALIST_2)
+BIF_RETTYPE binary_split_3(BIF_ALIST_3)
 {
-    return binary_match(BIF_P,BIF_ARG_1,BIF_ARG_2,((Eterm) 0));
+    return binary_split(BIF_P, BIF_ARG_1, BIF_ARG_2, BIF_ARG_3);
+}
+
+static Eterm do_match_not_found_result(Process *p, Eterm subject, BinaryFindState *bfs)
+{
+    if (bfs->flags & BINARY_FIND_ALL) {
+	return NIL;
+    } else {
+	return am_nomatch;
+    }
 }
 
+static Eterm do_match_single_result(Process *p, Eterm subject, BinaryFindState *bfs,
+				    Sint pos, Sint len)
+{
+    Eterm erlen;
+    Eterm *hp;
+    Eterm ret;
 
-BIF_RETTYPE binary_matches_2(BIF_ALIST_2)
+    erlen = erts_make_integer((Uint)(len), p);
+    ret = erts_make_integer(pos, p);
+    hp = HAlloc(p, 3);
+    ret = TUPLE2(hp, ret, erlen);
+
+    return ret;
+}
+
+static Eterm do_match_global_result(Process *p, Eterm subject, BinaryFindState *bfs,
+				    FindallData *fad, Uint fad_sz)
+{
+    Sint i;
+    Eterm tpl;
+    Eterm *hp;
+    Eterm ret;
+
+    for (i = 0; i < fad_sz; ++i) {
+	fad[i].epos = erts_make_integer(fad[i].pos, p);
+	fad[i].elen = erts_make_integer(fad[i].len, p);
+    }
+    hp = HAlloc(p, fad_sz * (3 + 2));
+    ret = NIL;
+    for (i = fad_sz - 1; i >= 0; --i) {
+	tpl = TUPLE2(hp, fad[i].epos, fad[i].elen);
+	hp += 3;
+	ret = CONS(hp, tpl, ret);
+	hp += 2;
+    }
+
+    return ret;
+}
+
+static Eterm do_split_not_found_result(Process *p, Eterm subject, BinaryFindState *bfs)
 {
-    return binary_matches(BIF_P,BIF_ARG_1,BIF_ARG_2,((Eterm) 0));
+    Eterm *hp;
+    Eterm ret;
+
+    if (bfs->flags & (BINARY_SPLIT_TRIM | BINARY_SPLIT_TRIM_ALL)
+        && binary_size(subject) == 0) {
+        return NIL;
+    }
+    hp = HAlloc(p, 2);
+    ret = CONS(hp, subject, NIL);
+
+    return ret;
 }
 
+static Eterm do_split_single_result(Process *p, Eterm subject, BinaryFindState *bfs,
+				    Sint pos, Sint len)
+{
+    size_t orig_size;
+    Eterm orig;
+    Uint offset;
+    Uint bit_offset;
+    Uint bit_size;
+    ErlSubBin *sb1;
+    ErlSubBin *sb2;
+    Eterm *hp;
+    Eterm ret;
+
+    orig_size = binary_size(subject);
+
+    if ((bfs->flags & (BINARY_SPLIT_TRIM | BINARY_SPLIT_TRIM_ALL)) &&
+	(orig_size - pos - len) == 0) {
+	if (pos == 0) {
+	    ret = NIL;
+	} else {
+	    hp = HAlloc(p, (ERL_SUB_BIN_SIZE + 2));
+	    ERTS_GET_REAL_BIN(subject, orig, offset, bit_offset, bit_size);
+	    sb1 = (ErlSubBin *) hp;
+	    sb1->thing_word = HEADER_SUB_BIN;
+	    sb1->size = pos;
+	    sb1->offs = offset;
+	    sb1->orig = orig;
+	    sb1->bitoffs = bit_offset;
+	    sb1->bitsize = bit_size;
+	    sb1->is_writable = 0;
+	    hp += ERL_SUB_BIN_SIZE;
+
+	    ret = CONS(hp, make_binary(sb1), NIL);
+	    hp += 2;
+	}
+    } else {
+	if ((bfs->flags & BINARY_SPLIT_TRIM_ALL) && (pos == 0)) {
+	    hp = HAlloc(p, 1 * (ERL_SUB_BIN_SIZE + 2));
+	    ERTS_GET_REAL_BIN(subject, orig, offset, bit_offset, bit_size);
+	    sb1 = NULL;
+	} else {
+	    hp = HAlloc(p, 2 * (ERL_SUB_BIN_SIZE + 2));
+	    ERTS_GET_REAL_BIN(subject, orig, offset, bit_offset, bit_size);
+	    sb1 = (ErlSubBin *) hp;
+	    sb1->thing_word = HEADER_SUB_BIN;
+	    sb1->size = pos;
+	    sb1->offs = offset;
+	    sb1->orig = orig;
+	    sb1->bitoffs = bit_offset;
+	    sb1->bitsize = 0;
+	    sb1->is_writable = 0;
+	    hp += ERL_SUB_BIN_SIZE;
+	}
+
+	sb2 = (ErlSubBin *) hp;
+	sb2->thing_word = HEADER_SUB_BIN;
+	sb2->size = orig_size - pos - len;
+	sb2->offs = offset + pos + len;
+	sb2->orig = orig;
+	sb2->bitoffs = bit_offset;
+	sb2->bitsize = bit_size;
+	sb2->is_writable = 0;
+	hp += ERL_SUB_BIN_SIZE;
+
+	ret = CONS(hp, make_binary(sb2), NIL);
+	hp += 2;
+	if (sb1 != NULL) {
+	    ret = CONS(hp, make_binary(sb1), ret);
+	    hp += 2;
+	}
+    }
+    return ret;
+}
+
+static Eterm do_split_global_result(Process *p, Eterm subject, BinaryFindState *bfs,
+				    FindallData *fad, Uint fad_sz)
+{
+    size_t orig_size;
+    Eterm orig;
+    Uint offset;
+    Uint bit_offset;
+    Uint bit_size;
+    ErlSubBin *sb;
+    Sint i;
+    Sint tail;
+    Uint list_size;
+    Uint end_pos;
+    Uint do_trim = bfs->flags & (BINARY_SPLIT_TRIM | BINARY_SPLIT_TRIM_ALL);
+    Eterm *hp;
+    Eterm *hendp;
+    Eterm ret;
+
+    tail = fad_sz - 1;
+    list_size = fad_sz + 1;
+    orig_size = binary_size(subject);
+    end_pos = (Uint)(orig_size);
+
+    hp = HAlloc(p, list_size * (ERL_SUB_BIN_SIZE + 2));
+    hendp = hp + list_size * (ERL_SUB_BIN_SIZE + 2);
+    ERTS_GET_REAL_BIN(subject, orig, offset, bit_offset, bit_size);
+    ASSERT(bit_size == 0);
+
+    ret = NIL;
+
+    for (i = tail; i >= 0; --i) {
+	sb = (ErlSubBin *)(hp);
+	sb->size = end_pos - (fad[i].pos + fad[i].len);
+	if (!(sb->size == 0 && do_trim)) {
+	    sb->thing_word = HEADER_SUB_BIN;
+	    sb->offs = offset + fad[i].pos + fad[i].len;
+	    sb->orig = orig;
+	    sb->bitoffs = bit_offset;
+	    sb->bitsize = 0;
+	    sb->is_writable = 0;
+	    hp += ERL_SUB_BIN_SIZE;
+	    ret = CONS(hp, make_binary(sb), ret);
+	    hp += 2;
+	    do_trim &= ~BINARY_SPLIT_TRIM;
+	}
+	end_pos = fad[i].pos;
+    }
+
+    sb = (ErlSubBin *)(hp);
+    sb->size = fad[0].pos;
+    if (!(sb->size == 0 && do_trim)) {
+	sb->thing_word = HEADER_SUB_BIN;
+	sb->offs = offset;
+	sb->orig = orig;
+	sb->bitoffs = bit_offset;
+	sb->bitsize = 0;
+	sb->is_writable = 0;
+	hp += ERL_SUB_BIN_SIZE;
+	ret = CONS(hp, make_binary(sb), ret);
+	hp += 2;
+    }
+    HRelease(p, hendp, hp);
+    return ret;
+}
+
+static BIF_RETTYPE binary_find_trap(BIF_ALIST_3)
+{
+    int runres;
+    Eterm result;
+    Binary *bin = ((ProcBin *) binary_val(BIF_ARG_3))->val;
+    runres = do_binary_find(BIF_P, BIF_ARG_1, NULL, bin, BIF_ARG_2, &result);
+    if (runres == DO_BIN_MATCH_OK) {
+	BIF_RET(result);
+    } else {
+	BUMP_ALL_REDS(BIF_P);
+	BIF_TRAP3(&binary_find_trap_export, BIF_P, BIF_ARG_1, result, BIF_ARG_3);
+    }
+}
 
 BIF_RETTYPE erts_binary_part(Process *p, Eterm binary, Eterm epos, Eterm elen)
 {
@@ -2550,7 +2790,6 @@ BIF_RETTYPE binary_referenced_byte_size_1(BIF_ALIST_1)
     }
     pb = (ProcBin *) binary_val(bin);
     if (pb->thing_word == HEADER_PROC_BIN) {
-	/* XXX:PaN - Halfword - orig_size is a long, we should handle that */
 	res = erts_make_integer((Uint) pb->val->orig_size, BIF_P);
     } else { /* heap binary */
 	res = erts_make_integer((Uint) ((ErlHeapBin *) pb)->size, BIF_P);
@@ -2568,7 +2807,7 @@ BIF_RETTYPE binary_referenced_byte_size_1(BIF_ALIST_1)
 #endif
 
 static int get_need(Uint u) {
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
     if (u > 0xFFFFFFFFUL) {
 	if (u > 0xFFFFFFFFFFFFUL) {
 	    if (u > 0xFFFFFFFFFFFFFFUL) {
diff --git a/erts/emulator/beam/erl_bif_ddll.c b/erts/emulator/beam/erl_bif_ddll.c
index 28bec6325c..2b1d875bfe 100644
--- a/erts/emulator/beam/erl_bif_ddll.c
+++ b/erts/emulator/beam/erl_bif_ddll.c
@@ -1707,18 +1707,19 @@ static void notify_proc(Process *proc, Eterm ref, Eterm driver_name, Eterm type,
     Eterm mess;
     Eterm r;
     Eterm *hp;
-    ErlHeapFragment *bp;
-    ErlOffHeap *ohp;
+    ErtsMessage *mp;
     ErtsProcLocks rp_locks = 0;
+    ErlOffHeap *ohp;
     ERTS_SMP_CHK_NO_PROC_LOCKS;
 
     assert_drv_list_rwlocked();
     if (errcode != 0) {
 	int need = load_error_need(errcode);
 	Eterm e;
-	hp = erts_alloc_message_heap(6 /* tuple */ + 3 /* Error tuple */ + 
-				     REF_THING_SIZE + need, &bp, &ohp, 
-				     proc, &rp_locks);
+	mp = erts_alloc_message_heap(proc, &rp_locks,
+				     (6 /* tuple */ + 3 /* Error tuple */ + 
+				      REF_THING_SIZE + need),
+				     &hp, &ohp);
 	r = copy_ref(ref,hp);
 	hp += REF_THING_SIZE;
 	e = build_load_error_hp(hp, errcode);
@@ -1727,12 +1728,14 @@ static void notify_proc(Process *proc, Eterm ref, Eterm driver_name, Eterm type,
 	hp += 3;
 	mess = TUPLE5(hp,type,r,am_driver,driver_name,mess);
     } else {	
-	hp = erts_alloc_message_heap(6 /* tuple */ + REF_THING_SIZE, &bp, &ohp, proc, &rp_locks);
+	mp = erts_alloc_message_heap(proc, &rp_locks,
+				     6 /* tuple */ + REF_THING_SIZE,
+				     &hp, &ohp);
 	r = copy_ref(ref,hp);
 	hp += REF_THING_SIZE;
 	mess = TUPLE5(hp,type,r,am_driver,driver_name,tag);
     }
-    erts_queue_message(proc, &rp_locks, bp, mess, am_undefined);
+    erts_queue_message(proc, &rp_locks, mp, mess, am_undefined);
     erts_smp_proc_unlock(proc, rp_locks);
     ERTS_SMP_CHK_NO_PROC_LOCKS;
 }
diff --git a/erts/emulator/beam/erl_bif_info.c b/erts/emulator/beam/erl_bif_info.c
index b44382cde8..e69908bbef 100644
--- a/erts/emulator/beam/erl_bif_info.c
+++ b/erts/emulator/beam/erl_bif_info.c
@@ -73,9 +73,6 @@ static char otp_version[] = ERLANG_OTP_VERSION;
 static char erts_system_version[] = ("Erlang/OTP " ERLANG_OTP_RELEASE
 				     "%s"
 				     " [erts-" ERLANG_VERSION "]"
-#if !HEAP_ON_C_STACK && !HALFWORD_HEAP
-				     " [no-c-stack-objects]"
-#endif
 #ifndef OTP_RELEASE
 #ifdef ERLANG_GIT_VERSION
 				     " [source-" ERLANG_GIT_VERSION "]"
@@ -84,12 +81,8 @@ static char erts_system_version[] = ("Erlang/OTP " ERLANG_OTP_RELEASE
 #endif
 #endif	
 #ifdef ARCH_64
-#if HALFWORD_HEAP
-				     " [64-bit halfword]"
-#else
 				     " [64-bit]"
 #endif
-#endif
 #ifdef ERTS_SMP
 				     " [smp:%beu:%beu]"
 #endif
@@ -137,6 +130,9 @@ static char erts_system_version[] = ("Erlang/OTP " ERLANG_OTP_RELEASE
 #ifdef USE_SYSTEMTAP
 				     " [systemtap]"
 #endif
+#ifdef SHCOPY
+				     " [sharing-preserving]"
+#endif
 				     "\n");
 
 #define ASIZE(a) (sizeof(a)/sizeof(a[0]))
@@ -596,6 +592,7 @@ static Eterm pi_args[] = {
     am_min_bin_vheap_size,
     am_current_location,
     am_current_stacktrace,
+    am_message_queue_data
 };
 
 #define ERTS_PI_ARGS ((int) (sizeof(pi_args)/sizeof(Eterm)))
@@ -643,6 +640,7 @@ pi_arg2ix(Eterm arg)
     case am_min_bin_vheap_size:			return 28;
     case am_current_location:			return 29;
     case am_current_stacktrace:			return 30;
+    case am_message_queue_data:			return 31;
     default:					return -1;
     }
 }
@@ -671,18 +669,12 @@ static Eterm pi_1_keys[] = {
 #define ERTS_PI_1_NO_OF_KEYS (sizeof(pi_1_keys)/sizeof(Eterm))
 
 static Eterm pi_1_keys_list;
-#if HEAP_ON_C_STACK
 static Eterm pi_1_keys_list_heap[2*ERTS_PI_1_NO_OF_KEYS];
-#endif
 
 static void
 process_info_init(void)
 {
-#if HEAP_ON_C_STACK
     Eterm *hp = &pi_1_keys_list_heap[0];
-#else
-    Eterm *hp = erts_alloc(ERTS_ALC_T_LL_TEMP_TERM,sizeof(Eterm)*2*ERTS_PI_1_NO_OF_KEYS);
-#endif
     int i;
 
     pi_1_keys_list = NIL;
@@ -731,9 +723,10 @@ pi_pid2proc(Process *c_p, Eterm pid, ErtsProcLocks info_locks)
 
 
 
-BIF_RETTYPE
+static BIF_RETTYPE
 process_info_aux(Process *BIF_P,
 		 Process *rp,
+		 ErtsProcLocks rp_locks,
 		 Eterm rpid,
 		 Eterm item,
 		 int always_wrap);
@@ -824,10 +817,31 @@ process_info_list(Process *c_p, Eterm pid, Eterm list, int always_wrap,
 	*fail_type = ERTS_PI_FAIL_TYPE_AWAIT_EXIT;
 	goto done;
     }
-    else if (!(locks & ERTS_PROC_LOCK_STATUS)) {
-	erts_smp_proc_unlock(rp, ERTS_PROC_LOCK_STATUS);
+    else {
+	ErtsProcLocks unlock_locks = 0;
+
+	if (c_p == rp)
+	    locks |= ERTS_PROC_LOCK_MAIN;
+
+	if (!(locks & ERTS_PROC_LOCK_STATUS))
+	    unlock_locks |= ERTS_PROC_LOCK_STATUS;
+
+	if (locks & ERTS_PROC_LOCK_MSGQ) {
+	    /*
+	     * Move in queue into private queue and
+	     * release msgq lock, enabling others to
+	     * send messages to the process while it
+	     * is being inspected...
+	     */
+	    ASSERT(locks & ERTS_PROC_LOCK_MAIN);
+	    ERTS_SMP_MSGQ_MV_INQ2PRIVQ(rp);
+	    locks &= ~ERTS_PROC_LOCK_MSGQ;
+	    unlock_locks |= ERTS_PROC_LOCK_MSGQ;
+	}
+
+	if (unlock_locks)
+	    erts_smp_proc_unlock(rp, unlock_locks);
     }
-	
 
     /*
      * We always handle 'messages' first if it should be part
@@ -839,7 +853,7 @@ process_info_list(Process *c_p, Eterm pid, Eterm list, int always_wrap,
     if (want_messages) {
 	ix = pi_arg2ix(am_messages);
 	ASSERT(part_res[ix] == THE_NON_VALUE);
-	part_res[ix] = process_info_aux(c_p, rp, pid, am_messages, always_wrap);
+	part_res[ix] = process_info_aux(c_p, rp, locks, pid, am_messages, always_wrap);
 	ASSERT(part_res[ix] != THE_NON_VALUE);
     }
 
@@ -847,7 +861,7 @@ process_info_list(Process *c_p, Eterm pid, Eterm list, int always_wrap,
 	ix = res_elem_ix[res_elem_ix_ix];
 	if (part_res[ix] == THE_NON_VALUE) {
 	    arg = pi_ix2arg(ix);
-	    part_res[ix] = process_info_aux(c_p, rp, pid, arg, always_wrap);
+	    part_res[ix] = process_info_aux(c_p, rp, locks, pid, arg, always_wrap);
 	    ASSERT(part_res[ix] != THE_NON_VALUE);
 	}
     }
@@ -978,9 +992,31 @@ BIF_RETTYPE process_info_2(BIF_ALIST_2)
 	ERTS_BIF_AWAIT_X_DATA_TRAP(BIF_P, BIF_ARG_1, am_undefined);
     }
     else {
+	ErtsProcLocks unlock_locks = 0;
+
+	if (BIF_P == rp)
+	    info_locks |= ERTS_PROC_LOCK_MAIN;
+
 	if (!(info_locks & ERTS_PROC_LOCK_STATUS))
-	    erts_smp_proc_unlock(rp, ERTS_PROC_LOCK_STATUS);
-	res = process_info_aux(BIF_P, rp, pid, BIF_ARG_2, 0);
+	    unlock_locks |= ERTS_PROC_LOCK_STATUS;
+
+	if (info_locks & ERTS_PROC_LOCK_MSGQ) {
+	    /*
+	     * Move in queue into private queue and
+	     * release msgq lock, enabling others to
+	     * send messages to the process while it
+	     * is being inspected...
+	     */
+	    ASSERT(info_locks & ERTS_PROC_LOCK_MAIN);
+	    ERTS_SMP_MSGQ_MV_INQ2PRIVQ(rp);
+	    info_locks &= ~ERTS_PROC_LOCK_MSGQ;
+	    unlock_locks |= ERTS_PROC_LOCK_MSGQ;
+	}
+
+	if (unlock_locks)
+	    erts_smp_proc_unlock(rp, unlock_locks);
+
+	res = process_info_aux(BIF_P, rp, info_locks, pid, BIF_ARG_2, 0);
     }
     ASSERT(is_value(res));
 
@@ -998,6 +1034,7 @@ BIF_RETTYPE process_info_2(BIF_ALIST_2)
 Eterm
 process_info_aux(Process *BIF_P,
 		 Process *rp,
+		 ErtsProcLocks rp_locks,
 		 Eterm rpid,
 		 Eterm item,
 		 int always_wrap)
@@ -1069,171 +1106,55 @@ process_info_aux(Process *BIF_P,
 	break;
 
     case am_messages: {
-	ErlMessage* mp;
-	int n;
 
-	ERTS_SMP_MSGQ_MV_INQ2PRIVQ(rp);
-	n = rp->msg.len;
-
-	if (n == 0 || ERTS_TRACE_FLAGS(rp) & F_SENSITIVE) {
+	if (rp->msg.len == 0 || ERTS_TRACE_FLAGS(rp) & F_SENSITIVE) {
 	    hp = HAlloc(BIF_P, 3);
 	} else {
-	    int remove_bad_messages = 0;
-	    struct {
-		Uint copy_struct_size;
-		ErlMessage* msgp;
-	    } *mq = erts_alloc(ERTS_ALC_T_TMP, n*sizeof(*mq));
-	    Sint i = 0;
-	    Uint heap_need = 3;
+	    ErtsMessageInfo *mip;
+	    Sint i;
+	    Uint heap_need;
+#ifdef DEBUG
 	    Eterm *hp_end;
+#endif
 
-	    for (mp = rp->msg.first; mp; mp = mp->next) {
-		heap_need += 2;
-		mq[i].msgp = mp;
-		if (rp != BIF_P) {
-		    Eterm msg = ERL_MESSAGE_TERM(mq[i].msgp);
-		    if (is_value(msg)) {
-			mq[i].copy_struct_size = (is_immed(msg)? 0 :
-						  size_object(msg));
-		    }
-		    else if (mq[i].msgp->data.attached) {
-			mq[i].copy_struct_size
-			    = erts_msg_attached_data_size(mq[i].msgp);
-		    }
-		    else {
-			/* Bad distribution message; ignore */
-			remove_bad_messages = 1;
-			mq[i].copy_struct_size = 0;
-		    }
-		    heap_need += mq[i].copy_struct_size;
-		}
-		else {
-		    mq[i].copy_struct_size = mp->data.attached ?
-                        erts_msg_attached_data_size(mp) : 0;
-		}
-		i++;
-	    }
+	    mip = erts_alloc(ERTS_ALC_T_TMP,
+			     rp->msg.len*sizeof(ErtsMessageInfo));
 
-            if (rp != BIF_P) {
-                hp = HAlloc(BIF_P, heap_need);
-                hp_end = hp + heap_need;
-                ASSERT(i == n);
-                for (i--; i >= 0; i--) {
-                    Eterm msg = ERL_MESSAGE_TERM(mq[i].msgp);
-		    if (is_value(msg)) {
-			if (mq[i].copy_struct_size)
-			    msg = copy_struct(msg,
-					      mq[i].copy_struct_size,
-					      &hp,
-					      &MSO(BIF_P));
-		    }
-		    else if (mq[i].msgp->data.attached) {
-			ErlHeapFragment *hfp;
-			/*
-			 * Decode it into a message buffer and attach it
-			 * to the message instead of the attached external
-			 * term.
-			 *
-			 * Note that we may not pass a process pointer
-			 * to erts_msg_distext2heap(), since it would then
-			 * try to alter locks on that process.
-			 */
-			msg = erts_msg_distext2heap(
-			    NULL, NULL, &hfp, &ERL_MESSAGE_TOKEN(mq[i].msgp),
-			    mq[i].msgp->data.dist_ext);
-
-			ERL_MESSAGE_TERM(mq[i].msgp) = msg;
-			mq[i].msgp->data.heap_frag = hfp;
-
-			if (is_non_value(msg)) {
-			    ASSERT(!mq[i].msgp->data.heap_frag);
-			    /* Bad distribution message; ignore */
-			    remove_bad_messages = 1;
-			    continue;
-			}
-			else {
-			    /* Make our copy of the message */
-			    ASSERT(size_object(msg) == erts_used_frag_sz(hfp));
-			    msg = copy_struct(msg,
-					      erts_used_frag_sz(hfp),
-					      &hp,
-					      &MSO(BIF_P));
-			}
-		    }
-		    else {
-			/* Bad distribution message; ignore */
-			remove_bad_messages = 1;
-			continue;
-		    }
-                    res = CONS(hp, msg, res);
-                    hp += 2;
-		}
-                HRelease(BIF_P, hp_end, hp+3);
-            }
-	    else {
-                for (i--; i >= 0; i--) {
-                    ErtsHeapFactory factory;
-                    Eterm msg = ERL_MESSAGE_TERM(mq[i].msgp);
-
-                    erts_factory_proc_prealloc_init(&factory, BIF_P,
-                                                    mq[i].copy_struct_size+2);
-		    if (mq[i].msgp->data.attached) {
-			/* Decode it on the heap */
-			erts_move_msg_attached_data_to_heap(&factory,
-							    mq[i].msgp);
-			msg = ERL_MESSAGE_TERM(mq[i].msgp);
-			ASSERT(!mq[i].msgp->data.attached);
-                    }
-                    if (is_value(msg)) {
-                        hp = erts_produce_heap(&factory, 2, 0);
-                        res = CONS(hp, msg, res);
-                    }
-                    else {
-                        /* Bad distribution message; ignore */
-                        remove_bad_messages = 1;
-                        continue;
-                    }
-                    erts_factory_close(&factory);
-		}
-                hp = HAlloc(BIF_P, 3);
-	    }
-	    erts_free(ERTS_ALC_T_TMP, mq);
-	    if (remove_bad_messages) {
-		ErlMessage **mpp;
-		/*
-		 * We need to remove bad distribution messages from
-		 * the queue, so that the value returned for
-		 * 'message_queue_len' is consistent with the value
-		 * returned for 'messages'.
-		 */
-		mpp = &rp->msg.first;
-		mp = rp->msg.first;
-		while (mp) {
-		    if (is_value(ERL_MESSAGE_TERM(mp))) {
-			mpp = &mp->next;
-			mp = mp->next;
-		    }
-		    else {
-			ErlMessage* bad_mp = mp;
-			ASSERT(!mp->data.attached);
-			if (rp->msg.save == &mp->next)
-			    rp->msg.save = mpp;
-			if (rp->msg.last == &mp->next)
-			    rp->msg.last = mpp;
-			*mpp = mp->next;
-			mp = mp->next;
-			rp->msg.len--;
-			free_message(bad_mp);
-		    }
-		}
+	    /*
+	     * Note that message queue may shrink when calling
+	     * erts_prep_msgq_for_inspection() since it removes
+	     * corrupt distribution messages.
+	     */
+	    heap_need = erts_prep_msgq_for_inspection(BIF_P, rp, rp_locks, mip);
+	    heap_need += 3; /* top 2-tuple */
+	    heap_need += rp->msg.len*2; /* Cons cells */
+
+	    hp = HAlloc(BIF_P, heap_need); /* heap_need is exact */
+#ifdef DEBUG
+	    hp_end = hp + heap_need;
+#endif
+
+	    /* Build list of messages... */
+	    for (i = rp->msg.len - 1, res = NIL; i >= 0; i--) {
+		Eterm msg = ERL_MESSAGE_TERM(mip[i].msgp);
+		Uint sz = mip[i].size;
+
+		if (sz != 0)
+		    msg = copy_struct(msg, sz, &hp, &BIF_P->off_heap);
+
+		res = CONS(hp, msg, res);
+		hp += 2;
 	    }
+
+	    ASSERT(hp_end == hp + 3);
+
+	    erts_free(ERTS_ALC_T_TMP, mip);
 	}
 	break;
     }
 
     case am_message_queue_len:
 	hp = HAlloc(BIF_P, 3);
-	ERTS_SMP_MSGQ_MV_INQ2PRIVQ(rp);
 	res = make_small(rp->msg.len);
 	break;
 
@@ -1421,7 +1342,7 @@ process_info_aux(Process *BIF_P,
     }
 
     case am_total_heap_size: {
-	ErlMessage *mp;
+	ErtsMessage *mp;
 	Uint total_heap_size;
 	Uint hsz = 3;
 
@@ -1431,8 +1352,6 @@ process_info_aux(Process *BIF_P,
 
 	total_heap_size += rp->mbuf_sz;
 
-	ERTS_SMP_MSGQ_MV_INQ2PRIVQ(rp);
-
 	for (mp = rp->msg.first; mp; mp = mp->next)
 	    if (mp->data.attached)
 		total_heap_size += erts_msg_attached_data_size(mp);
@@ -1454,7 +1373,7 @@ process_info_aux(Process *BIF_P,
 
     case am_memory: { /* Memory consumed in bytes */
 	Uint hsz = 3;
-	Uint size = erts_process_memory(rp);
+	Uint size = erts_process_memory(rp, 0);
 	(void) erts_bld_uint(NULL, &hsz, size);
 	hp = HAlloc(BIF_P, hsz);
 	res = erts_bld_uint(&hp, NULL, size);
@@ -1580,6 +1499,25 @@ process_info_aux(Process *BIF_P,
 	break;
     }
 
+    case am_message_queue_data:
+	switch (rp->flags & (F_OFF_HEAP_MSGQ|F_ON_HEAP_MSGQ)) {
+	case F_OFF_HEAP_MSGQ:
+	    res = am_off_heap;
+	    break;
+	case F_ON_HEAP_MSGQ:
+	    res = am_on_heap;
+	    break;
+	case 0:
+	    res = am_mixed;
+	    break;
+	default:
+	    res = am_error;
+	    ERTS_INTERNAL_ERROR("Inconsistent message queue management state");
+	    break;
+	}
+	hp = HAlloc(BIF_P, 3);
+	break;
+
     default:
 	return THE_NON_VALUE; /* will produce badarg */
 
@@ -2741,6 +2679,19 @@ BIF_RETTYPE system_info_1(BIF_ALIST_1)
 	BIF_RET(am_true);
     }
 #endif
+    else if (BIF_ARG_1 == am_message_queue_data) {
+	switch (erts_default_spo_flags & (SPO_ON_HEAP_MSGQ|SPO_OFF_HEAP_MSGQ)) {
+	case SPO_OFF_HEAP_MSGQ:
+	    BIF_RET(am_off_heap);
+	case SPO_ON_HEAP_MSGQ:
+	    BIF_RET(am_on_heap);
+	case 0:
+	    BIF_RET(am_mixed);
+	default:
+	    ERTS_INTERNAL_ERROR("Inconsistent message queue management state");
+	    BIF_RET(am_error);
+	}
+    }
     else if (ERTS_IS_ATOM_STR("compile_info",BIF_ARG_1)) {
 	Uint  sz;
 	Eterm res = NIL, tup, text;
@@ -2779,6 +2730,20 @@ BIF_RETTYPE system_info_1(BIF_ALIST_1)
     else if (ERTS_IS_ATOM_STR("eager_check_io",BIF_ARG_1)) {
 	BIF_RET(erts_eager_check_io ? am_true : am_false);
     }
+    else if (ERTS_IS_ATOM_STR("literal_test",BIF_ARG_1)) {
+#ifdef ERTS_HAVE_IS_IN_LITERAL_RANGE
+#ifdef ARCH_64
+	DECL_AM(range);
+	BIF_RET(AM_range);
+#else /* ARCH_32 */
+	DECL_AM(range_bitmask);
+	BIF_RET(AM_range_bitmask);
+#endif /* ARCH_32 */
+#else  /* ! ERTS_HAVE_IS_IN_LITERAL_RANGE */
+	DECL_AM(tag);
+	BIF_RET(AM_tag);
+#endif
+    }
 
     BIF_ERROR(BIF_P, BADARG);
 }
@@ -3234,6 +3199,39 @@ BIF_RETTYPE statistics_1(BIF_ALIST_1)
 	if (is_non_value(res))
 	    BIF_RET(am_undefined);
 	BIF_TRAP1(gather_sched_wall_time_res_trap, BIF_P, res);
+    } else if (BIF_ARG_1 == am_total_active_tasks
+	       || BIF_ARG_1 == am_total_run_queue_lengths) {
+	Uint no = erts_run_queues_len(NULL, 0, BIF_ARG_1 == am_total_active_tasks);
+	if (IS_USMALL(0, no))
+	    res = make_small(no);
+	else {
+	    Eterm *hp = HAlloc(BIF_P, BIG_UINT_HEAP_SIZE);
+	    res = uint_to_big(no, hp);
+	}
+	BIF_RET(res);
+    } else if (BIF_ARG_1 == am_active_tasks
+	       || BIF_ARG_1 == am_run_queue_lengths) {
+	Eterm res, *hp, **hpp;
+	Uint sz, *szp;
+	int no_qs = erts_no_run_queues;
+	Uint *qszs = erts_alloc(ERTS_ALC_T_TMP,sizeof(Uint)*no_qs*2);
+	(void) erts_run_queues_len(qszs, 0, BIF_ARG_1 == am_active_tasks);
+	sz = 0;
+	szp = &sz;
+	hpp = NULL;
+	while (1) {
+	    int i;
+	    for (i = 0; i < no_qs; i++)
+		qszs[no_qs+i] = erts_bld_uint(hpp, szp, qszs[i]);
+	    res = erts_bld_list(hpp, szp, no_qs, &qszs[no_qs]);
+	    if (hpp) {
+		erts_free(ERTS_ALC_T_TMP, qszs);
+		BIF_RET(res);
+	    }
+	    hp = HAlloc(BIF_P, sz);
+	    szp = NULL;
+	    hpp = &hp;
+	}
     } else if (BIF_ARG_1 == am_context_switches) {
 	Eterm cs = erts_make_integer(erts_get_total_context_switches(), BIF_P);
 	hp = HAlloc(BIF_P, 3);
@@ -3282,7 +3280,7 @@ BIF_RETTYPE statistics_1(BIF_ALIST_1)
 	res = TUPLE2(hp, b1, b2);
 	BIF_RET(res);
     } else if (BIF_ARG_1 ==  am_run_queue) {
-	res = erts_run_queues_len(NULL);
+	res = erts_run_queues_len(NULL, 1, 0);
 	BIF_RET(make_small(res));
     } else if (BIF_ARG_1 == am_wall_clock) {
 	UWord w1, w2;
@@ -3302,7 +3300,7 @@ BIF_RETTYPE statistics_1(BIF_ALIST_1)
 	Uint sz, *szp;
 	int no_qs = erts_no_run_queues;
 	Uint *qszs = erts_alloc(ERTS_ALC_T_TMP,sizeof(Uint)*no_qs*2);
-	(void) erts_run_queues_len(qszs);
+	(void) erts_run_queues_len(qszs, 0, 0);
 	sz = 0;
 	szp = &sz;
 	hpp = NULL;
@@ -3454,7 +3452,7 @@ BIF_RETTYPE erts_debug_get_internal_state_1(BIF_ALIST_1)
 	    BIF_RET(res);
 	}
         else if (ERTS_IS_ATOM_STR("mmap", BIF_ARG_1)) {
-            BIF_RET(erts_mmap_debug_info(BIF_P));
+            BIF_RET(erts_mmap_debug_info(&erts_dflt_mmapper, BIF_P));
         }
 	else if (ERTS_IS_ATOM_STR("unique_monotonic_integer_state", BIF_ARG_1)) {
 	    BIF_RET(erts_debug_get_unique_monotonic_integer_state(BIF_P));
@@ -3864,9 +3862,7 @@ BIF_RETTYPE erts_debug_set_internal_state_2(BIF_ALIST_2)
 		BIF_RET(am_false);
 	    }
 	    else {
-		FLAGS(rp) |= F_FORCE_GC;
-		if (BIF_P != rp)
-		    erts_smp_proc_unlock(rp, ERTS_PROC_LOCK_MAIN);
+		ERTS_FORCE_GC(BIF_P);
 		BIF_RET(am_true);
 	    }
 	}
@@ -4085,6 +4081,17 @@ BIF_RETTYPE erts_debug_set_internal_state_2(BIF_ALIST_2)
 		BIF_RET(am_ok);
 	    }
 	}
+        else if (ERTS_IS_ATOM_STR("fill_heap", BIF_ARG_1)) {
+            UWord left = HeapWordsLeft(BIF_P);
+            if (left > 1) {
+                Eterm* hp = HAlloc(BIF_P, left);
+                *hp = make_pos_bignum_header(left - 1);
+            }
+            if (BIF_ARG_2 == am_true) {
+                FLAGS(BIF_P) |= F_NEED_FULLSWEEP;
+            }
+            BIF_RET(am_ok);
+        }
     }
 
     BIF_ERROR(BIF_P, BADARG);
@@ -4334,14 +4341,17 @@ static void os_info_init(void)
     os_flavor(buf, 1024);
     flav = erts_atom_put((byte *) buf, strlen(buf), ERTS_ATOM_ENC_LATIN1, 1);
     erts_free(ERTS_ALC_T_TMP, (void *) buf);
-    hp = erts_alloc(ERTS_ALC_T_LL_TEMP_TERM, (3+4)*sizeof(Eterm));
+    hp = erts_alloc(ERTS_ALC_T_LITERAL, (3+4)*sizeof(Eterm));
     os_type_tuple = TUPLE2(hp, type, flav);
+    erts_set_literal_tag(&os_type_tuple, hp, 3);
+
     hp += 3;
     os_version(&major, &minor, &build);
     os_version_tuple = TUPLE3(hp,
 			      make_small(major),
 			      make_small(minor),
 			      make_small(build));
+    erts_set_literal_tag(&os_version_tuple, hp, 4);
 }
 
 void
diff --git a/erts/emulator/beam/erl_bif_op.c b/erts/emulator/beam/erl_bif_op.c
index c9192fc420..d53a9e11ca 100644
--- a/erts/emulator/beam/erl_bif_op.c
+++ b/erts/emulator/beam/erl_bif_op.c
@@ -258,7 +258,7 @@ Eterm erl_is_function(Process* p, Eterm arg1, Eterm arg2)
 	    BIF_RET(am_true);
 	}
     } else if (is_export(arg1)) {
-	Export* exp = (Export *) EXPAND_POINTER((export_val(arg1))[1]);
+	Export* exp = (Export *) (export_val(arg1)[1]);
 
 	if (exp->code[2] == (Uint) arity) {
 	    BIF_RET(am_true);
diff --git a/erts/emulator/beam/erl_bif_port.c b/erts/emulator/beam/erl_bif_port.c
index 3ff54c7a60..839abd0424 100644
--- a/erts/emulator/beam/erl_bif_port.c
+++ b/erts/emulator/beam/erl_bif_port.c
@@ -41,6 +41,7 @@
 #include "external.h"
 #include "packet_parser.h"
 #include "erl_bits.h"
+#include "erl_bif_unique.h"
 #include "dtrace-wrapper.h"
 
 static Port *open_port(Process* p, Eterm name, Eterm settings, int *err_typep, int *err_nump);
@@ -50,10 +51,10 @@ static void free_args(char **);
 
 char *erts_default_arg0 = "default";
 
-BIF_RETTYPE open_port_2(BIF_ALIST_2)
+BIF_RETTYPE erts_internal_open_port_2(BIF_ALIST_2)
 {
     Port *port;
-    Eterm port_id;
+    Eterm res;
     char *str;
     int err_type, err_num;
 
@@ -61,27 +62,58 @@ BIF_RETTYPE open_port_2(BIF_ALIST_2)
     if (!port) {
 	if (err_type == -3) {
 	    ASSERT(err_num == BADARG || err_num == SYSTEM_LIMIT);
-	    BIF_ERROR(BIF_P, err_num);
+	    if (err_num == BADARG)
+                res = am_badarg;
+            else if (err_num == SYSTEM_LIMIT)
+                res = am_system_limit;
+            else
+                /* this is only here to silence gcc, it should not happen */
+                BIF_ERROR(BIF_P, EXC_INTERNAL_ERROR);
 	} else if (err_type == -2) {
 	    str = erl_errno_id(err_num);
+            res = erts_atom_put((byte *) str, strlen(str), ERTS_ATOM_ENC_LATIN1, 1);
 	} else {
-	    str = "einval";
+	    res = am_einval;
 	}
-	BIF_P->fvalue = erts_atom_put((byte *) str, strlen(str), ERTS_ATOM_ENC_LATIN1, 1);
-	BIF_ERROR(BIF_P, EXC_ERROR);
-    }
+        BIF_RET(res);
+    }
+
+    if (port->drv_ptr->flags & ERL_DRV_FLAG_USE_INIT_ACK) {
+        /* Copied from erl_port_task.c */
+        port->async_open_port = erts_alloc(ERTS_ALC_T_PRTSD,
+                                           sizeof(*port->async_open_port));
+        erts_make_ref_in_array(port->async_open_port->ref);
+        port->async_open_port->to = BIF_P->common.id;
+
+        erts_smp_proc_lock(BIF_P, ERTS_PROC_LOCKS_MSG_RECEIVE | ERTS_PROC_LOCK_LINK);
+        if (ERTS_PROC_PENDING_EXIT(BIF_P)) {
+	    /* need to exit caller instead */
+	    erts_smp_proc_unlock(BIF_P, ERTS_PROC_LOCKS_MSG_RECEIVE | ERTS_PROC_LOCK_LINK);
+	    KILL_CATCHES(BIF_P);
+	    BIF_P->freason = EXC_EXIT;
+            erts_port_release(port);
+            BIF_RET(am_badarg);
+	}
+
+	ERTS_SMP_MSGQ_MV_INQ2PRIVQ(BIF_P);
+	BIF_P->msg.save = BIF_P->msg.last;
 
-    erts_smp_proc_lock(BIF_P, ERTS_PROC_LOCK_LINK);
+	erts_smp_proc_unlock(BIF_P, ERTS_PROC_LOCKS_MSG_RECEIVE);
+
+        res = erts_proc_store_ref(BIF_P, port->async_open_port->ref);
+    } else {
+        res = port->common.id;
+        erts_smp_proc_lock(BIF_P, ERTS_PROC_LOCK_LINK);
+    }
 
-    port_id = port->common.id;
     erts_add_link(&ERTS_P_LINKS(port), LINK_PID, BIF_P->common.id);
-    erts_add_link(&ERTS_P_LINKS(BIF_P), LINK_PID, port_id);
+    erts_add_link(&ERTS_P_LINKS(BIF_P), LINK_PID, port->common.id);
 
     erts_smp_proc_unlock(BIF_P, ERTS_PROC_LOCK_LINK);
 
     erts_port_release(port);
 
-    BIF_RET(port_id);
+    BIF_RET(res);
 }
 
 static ERTS_INLINE Port *
@@ -1329,7 +1361,8 @@ BIF_RETTYPE decode_packet_3(BIF_ALIST_3)
     ErlSubBin* rest;
     Eterm res;
     Eterm options;
-    int code;
+    int   code;
+    char  delimiter = '\n';
 
     if (!is_binary(BIF_ARG_2) || 
         (!is_list(BIF_ARG_3) && !is_nil(BIF_ARG_3))) {
@@ -1370,6 +1403,11 @@ BIF_RETTYPE decode_packet_3(BIF_ALIST_3)
                 case am_line_length:
                     trunc_len = val;
                     goto next_option;
+                case am_line_delimiter:
+                    if (type == TCP_PB_LINE_LF && val >= 0 && val <= 255) {
+                        delimiter = (char)val;
+                        goto next_option;
+                    }
                 }
             }
         }
@@ -1390,7 +1428,7 @@ BIF_RETTYPE decode_packet_3(BIF_ALIST_3)
         pca.aligned_ptr = bin_ptr;
     }
     packet_sz = packet_get_length(type, (char*)pca.aligned_ptr, pca.bin_sz,
-                                  max_plen, trunc_len, &http_state);
+                                  max_plen, trunc_len, delimiter, &http_state);
     if (!(packet_sz > 0 && packet_sz <= pca.bin_sz)) {
         if (packet_sz < 0) {
 	    goto error;
diff --git a/erts/emulator/beam/erl_bif_re.c b/erts/emulator/beam/erl_bif_re.c
index 86951f32b0..7f7cd376ac 100644
--- a/erts/emulator/beam/erl_bif_re.c
+++ b/erts/emulator/beam/erl_bif_re.c
@@ -100,7 +100,7 @@ Sint erts_re_set_loop_limit(Sint limit)
 
 static int term_to_int(Eterm term, int *sp)
 {
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
 
     if (is_small(term)) {
 	Uint x = signed_val(term);
@@ -151,7 +151,7 @@ static int term_to_int(Eterm term, int *sp)
 
 static Eterm make_signed_integer(int x, Process *p)
 {
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
     return make_small(x);
 #else
     Eterm* hp;
diff --git a/erts/emulator/beam/erl_bif_trace.c b/erts/emulator/beam/erl_bif_trace.c
index 03f51132b1..4d67e39e7e 100644
--- a/erts/emulator/beam/erl_bif_trace.c
+++ b/erts/emulator/beam/erl_bif_trace.c
@@ -1879,11 +1879,7 @@ new_seq_trace_token(Process* p)
 {
     Eterm* hp;
 
-    if (SEQ_TRACE_TOKEN(p) == NIL
-#ifdef USE_VM_PROBES
-	|| SEQ_TRACE_TOKEN(p) == am_have_dt_utag
-#endif
-	) {
+    if (have_no_seqtrace(SEQ_TRACE_TOKEN(p))) {
 	hp = HAlloc(p, 6);
 	SEQ_TRACE_TOKEN(p) = TUPLE5(hp, make_small(0),		/* Flags  */ 
 				    make_small(0),		/* Label  */
@@ -1903,12 +1899,8 @@ BIF_RETTYPE erl_seq_trace_info(Process *p, Eterm item)
 	BIF_ERROR(p, BADARG);
     }
 
-    if (SEQ_TRACE_TOKEN(p) == NIL
-#ifdef USE_VM_PROBES
-	|| SEQ_TRACE_TOKEN(p) == am_have_dt_utag
-#endif
-	) {
-	if ((item == am_send) || (item == am_receive) || 
+    if (have_no_seqtrace(SEQ_TRACE_TOKEN(p))) {
+	if ((item == am_send)  || (item == am_receive) || 
 	    (item == am_print) || (item == am_timestamp)) {
 	    hp = HAlloc(p,3);
 	    res = TUPLE2(hp, item, am_false);
@@ -1964,11 +1956,7 @@ BIF_RETTYPE seq_trace_info_1(BIF_ALIST_1)
  */
 BIF_RETTYPE seq_trace_print_1(BIF_ALIST_1)    
 {
-    if (SEQ_TRACE_TOKEN(BIF_P) == NIL 
-#ifdef USE_VM_PROBES
-	|| SEQ_TRACE_TOKEN(BIF_P) == am_have_dt_utag
-#endif
-	) {
+    if (have_no_seqtrace(SEQ_TRACE_TOKEN(BIF_P))) {
 	BIF_RET(am_false);
     }
     seq_trace_update_send(BIF_P);
@@ -1987,11 +1975,7 @@ BIF_RETTYPE seq_trace_print_1(BIF_ALIST_1)
  */
 BIF_RETTYPE seq_trace_print_2(BIF_ALIST_2)    
 {
-    if (SEQ_TRACE_TOKEN(BIF_P) == NIL
-#ifdef USE_VM_PROBES
-	|| SEQ_TRACE_TOKEN(BIF_P) == am_have_dt_utag
-#endif
-	) {
+    if (have_no_seqtrace(SEQ_TRACE_TOKEN(BIF_P))) {
 	BIF_RET(am_false);
     }
     if (!(is_atom(BIF_ARG_1) || is_small(BIF_ARG_1))) {
diff --git a/erts/emulator/beam/erl_bif_unique.c b/erts/emulator/beam/erl_bif_unique.c
index 5eca09c5a6..c4a39b8897 100644
--- a/erts/emulator/beam/erl_bif_unique.c
+++ b/erts/emulator/beam/erl_bif_unique.c
@@ -338,7 +338,7 @@ static struct {
     } w;
 } raw_unique_monotonic_integer erts_align_attribute(ERTS_CACHE_LINE_SIZE);
 
-#if defined(ARCH_32) || HALFWORD_HEAP
+#if defined(ARCH_32)
 #  define ERTS_UNIQUE_MONOTONIC_OFFSET ERTS_SINT64_MIN
 #else
 #  define ERTS_UNIQUE_MONOTONIC_OFFSET MIN_SMALL
@@ -368,7 +368,7 @@ get_unique_monotonic_integer_heap_size(Uint64 raw, int positive)
 	Sint64 value = ((Sint64) raw) + ERTS_UNIQUE_MONOTONIC_OFFSET;
 	if (IS_SSMALL(value))
 	    return 0;
-#if defined(ARCH_32) || HALFWORD_HEAP
+#if defined(ARCH_32)
 	return ERTS_SINT64_HEAP_SIZE(value);
 #else
 	return ERTS_UINT64_HEAP_SIZE((Uint64) value);
@@ -393,7 +393,7 @@ make_unique_monotonic_integer_value(Eterm *hp, Uint hsz, Uint64 raw, int positiv
 	if (hsz == 0)
 	    res = make_small(value);
 	else {
-#if defined(ARCH_32) || HALFWORD_HEAP
+#if defined(ARCH_32)
 	    res = erts_sint64_to_big(value, &hp);
 #else 
 	    res = erts_uint64_to_big((Uint64) value, &hp);
diff --git a/erts/emulator/beam/erl_binary.h b/erts/emulator/beam/erl_binary.h
index ea01bf08f0..e181b5555d 100644
--- a/erts/emulator/beam/erl_binary.h
+++ b/erts/emulator/beam/erl_binary.h
@@ -72,7 +72,6 @@ typedef struct erl_heap_bin {
  */
 
 #define binary_size(Bin) (binary_val(Bin)[1])
-#define binary_size_rel(Bin,BasePtr) (binary_val_rel(Bin,BasePtr)[1])
 
 #define binary_bitsize(Bin)			\
   ((*binary_val(Bin) == HEADER_SUB_BIN) ?	\
@@ -95,12 +94,9 @@ typedef struct erl_heap_bin {
  * Bitsize: output variable (Uint)
  */
 
-#define ERTS_GET_BINARY_BYTES(Bin,Bytep,Bitoffs,Bitsize) \
-     ERTS_GET_BINARY_BYTES_REL(Bin,Bytep,Bitoffs,Bitsize,NULL)
-
-#define ERTS_GET_BINARY_BYTES_REL(Bin,Bytep,Bitoffs,Bitsize,BasePtr)    \
+#define ERTS_GET_BINARY_BYTES(Bin,Bytep,Bitoffs,Bitsize)                \
 do {									\
-    Eterm* _real_bin = binary_val_rel(Bin,BasePtr);			\
+    Eterm* _real_bin = binary_val(Bin);		                	\
     Uint _offs = 0;							\
     Bitoffs = Bitsize = 0;						\
     if (*_real_bin == HEADER_SUB_BIN) {					\
@@ -108,7 +104,7 @@ do {									\
 	_offs = _sb->offs;						\
         Bitoffs = _sb->bitoffs;						\
         Bitsize = _sb->bitsize;						\
-	_real_bin = binary_val_rel(_sb->orig,BasePtr);			\
+	_real_bin = binary_val(_sb->orig);	        		\
     }									\
     if (*_real_bin == HEADER_PROC_BIN) {				\
 	Bytep = ((ProcBin *) _real_bin)->bytes + _offs;			\
@@ -131,11 +127,8 @@ do {									\
  */
 
 #define ERTS_GET_REAL_BIN(Bin, RealBin, ByteOffset, BitOffset, BitSize) \
-     ERTS_GET_REAL_BIN_REL(Bin, RealBin, ByteOffset, BitOffset, BitSize, NULL)
-
-#define ERTS_GET_REAL_BIN_REL(Bin, RealBin, ByteOffset, BitOffset, BitSize, BasePtr) \
   do {									\
-    ErlSubBin* _sb = (ErlSubBin *) binary_val_rel(Bin,BasePtr);	        \
+    ErlSubBin* _sb = (ErlSubBin *) binary_val(Bin);	                \
     if (_sb->thing_word == HEADER_SUB_BIN) {				\
       RealBin = _sb->orig;						\
       ByteOffset = _sb->offs;						\
diff --git a/erts/emulator/beam/erl_bits.c b/erts/emulator/beam/erl_bits.c
index 01734c55d7..11d83686a3 100644
--- a/erts/emulator/beam/erl_bits.c
+++ b/erts/emulator/beam/erl_bits.c
@@ -282,7 +282,7 @@ erts_bs_get_integer_2(Process *p, Uint num_bits, unsigned flags, ErlBinMatchBuff
 	 * Simply shift whole bytes into the result.
 	 */
 	switch (BYTE_OFFSET(n)) {
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
 	case 7: w = (w << 8) | *bp++;
 	case 6: w = (w << 8) | *bp++;
 	case 5: w = (w << 8) | *bp++;
@@ -387,7 +387,7 @@ erts_bs_get_integer_2(Process *p, Uint num_bits, unsigned flags, ErlBinMatchBuff
     case 3: 
 	v32 = LSB[0] + (LSB[1]<<8) + (LSB[2]<<16); 
 	goto big_small;
-#if !defined(ARCH_64) || HALFWORD_HEAP
+#if !defined(ARCH_64)
     case 4:
 	v32 = (LSB[0] + (LSB[1]<<8) + (LSB[2]<<16) + (LSB[3]<<24));
 	if (!IS_USMALL(sgn, v32)) {
diff --git a/erts/emulator/beam/erl_db.c b/erts/emulator/beam/erl_db.c
index 878ee32b47..3030c1c91a 100644
--- a/erts/emulator/beam/erl_db.c
+++ b/erts/emulator/beam/erl_db.c
@@ -1774,15 +1774,9 @@ BIF_RETTYPE ets_delete_1(BIF_ALIST_1)
 	 * (it looks like an continuation pointer), but that is will crash the
 	 * emulator if this BIF is call traced.
 	 */
-#if HALFWORD_HEAP
-	Eterm *hp = HAlloc(BIF_P, 3);
-	hp[0] = make_pos_bignum_header(2);
-	*((UWord *) (UWord) (hp+1)) = (UWord) tb;
-#else
 	Eterm *hp = HAlloc(BIF_P, 2);
 	hp[0] = make_pos_bignum_header(1);
 	hp[1] = (Eterm) tb;
-#endif
 	BIF_TRAP1(&ets_delete_continue_exp, BIF_P, make_big(hp));
     }
     else {
@@ -1838,7 +1832,7 @@ BIF_RETTYPE ets_give_away_3(BIF_ALIST_3)
 			     tb->common.id,
 			     from_pid,
 			     BIF_ARG_3), 
-		      0);
+                      0);
     erts_smp_proc_unlock(to_proc, to_locks);
     UnUseTmpHeap(5,BIF_P);
     BIF_RET(am_true);
@@ -2840,7 +2834,7 @@ BIF_RETTYPE ets_match_spec_run_r_3(BIF_ALIST_3)
 	    BIF_TRAP3(bif_export[BIF_ets_match_spec_run_r_3],
 		      BIF_P,lst,BIF_ARG_2,ret);
 	}
-	res = db_prog_match(BIF_P, mp, CAR(list_val(lst)), NULL, NULL, 0,
+	res = db_prog_match(BIF_P, mp, CAR(list_val(lst)), NULL, 0,
 			    ERTS_PAM_COPY_RESULT, &dummy);
 	if (is_value(res)) {
 	    hp = HAlloc(BIF_P, 2);
@@ -3217,7 +3211,7 @@ retry:
 			     tb->common.id,
 			     p->common.id,
 			     heir_data), 
-		      0);
+                      0);
     erts_smp_proc_unlock(to_proc, to_locks);
     return !0;
 }
@@ -3652,11 +3646,8 @@ static BIF_RETTYPE ets_delete_trap(BIF_ALIST_1)
     Eterm* ptr = big_val(cont);
     DbTable *tb = *((DbTable **) (UWord) (ptr + 1));
 
-#if HALFWORD_HEAP
-    ASSERT(*ptr == make_pos_bignum_header(2));
-#else
     ASSERT(*ptr == make_pos_bignum_header(1));
-#endif
+
     db_lock(tb, LCK_WRITE);
     trap = free_table_cont(p, tb, 0, 1);
     db_unlock(tb, LCK_WRITE);
diff --git a/erts/emulator/beam/erl_db_hash.c b/erts/emulator/beam/erl_db_hash.c
index 81b0c4465c..cff65f244d 100644
--- a/erts/emulator/beam/erl_db_hash.c
+++ b/erts/emulator/beam/erl_db_hash.c
@@ -148,8 +148,11 @@ static ERTS_INLINE Uint hash_to_ix(DbTableHash* tb, HashValue hval)
 }
 
 /* Remember a slot containing a pseudo-deleted item (INVALID_HASH)
-*/
-static ERTS_INLINE void add_fixed_deletion(DbTableHash* tb, int ix)
+ * Return false if we got raced by unfixing thread
+ * and the object should be deleted for real.
+ */
+static ERTS_INLINE int add_fixed_deletion(DbTableHash* tb, int ix,
+                                          erts_aint_t fixated_by_me)
 {
     erts_aint_t was_next;
     erts_aint_t exp_next;
@@ -160,12 +163,18 @@ static ERTS_INLINE void add_fixed_deletion(DbTableHash* tb, int ix)
     fixd->slot = ix;
     was_next = erts_smp_atomic_read_acqb(&tb->fixdel);
     do { /* Lockless atomic insertion in linked list: */
-	exp_next = was_next;
+        if (NFIXED(tb) <= fixated_by_me) {
+            erts_db_free(ERTS_ALC_T_DB_FIX_DEL, (DbTable*)tb,
+                         fixd, sizeof(FixedDeletion));
+            return 0; /* raced by unfixer */
+        }
+        exp_next = was_next;
 	fixd->next = (FixedDeletion*) exp_next;
-	was_next = erts_smp_atomic_cmpxchg_relb(&tb->fixdel,
-						(erts_aint_t) fixd,
-						exp_next);
+	was_next = erts_smp_atomic_cmpxchg_mb(&tb->fixdel,
+                                              (erts_aint_t) fixd,
+                                              exp_next);
     }while (was_next != exp_next);
+    return 1;
 }
 
 
@@ -460,9 +469,6 @@ static ERTS_INLINE void try_shrink(DbTableHash* tb)
     }
 }	
 
-#define EQ_REL(x,y,y_base) \
-    (is_same(x,NULL,y,y_base) || (is_not_both_immed((x),(y)) && eq_rel((x),NULL,(y),y_base)))
-
 /* Is this a live object (not pseodo-deleted) with the specified key? 
 */
 static ERTS_INLINE int has_live_key(DbTableHash* tb, HashDbTerm* b,
@@ -472,7 +478,7 @@ static ERTS_INLINE int has_live_key(DbTableHash* tb, HashDbTerm* b,
     else {
 	Eterm itemKey = GETKEY(tb, b->dbterm.tpl);
 	ASSERT(!is_header(itemKey));
-	return EQ_REL(key, itemKey, b->dbterm.tpl);
+	return EQ(key, itemKey);
     }
 }
 
@@ -485,7 +491,7 @@ static ERTS_INLINE int has_key(DbTableHash* tb, HashDbTerm* b,
     else {
 	Eterm itemKey = GETKEY(tb, b->dbterm.tpl);
 	ASSERT(!is_header(itemKey));
-	return EQ_REL(key, itemKey, b->dbterm.tpl);
+	return EQ(key, itemKey);
     }
 }
 
@@ -607,8 +613,8 @@ void db_unfix_table_hash(DbTableHash *tb)
 		       || (erts_smp_lc_rwmtx_is_rlocked(&tb->common.rwlock)
 			   && !tb->common.is_thread_safe));
 restart:
-    fixdel = (FixedDeletion*) erts_smp_atomic_xchg_acqb(&tb->fixdel,
-							(erts_aint_t) NULL);
+    fixdel = (FixedDeletion*) erts_smp_atomic_xchg_mb(&tb->fixdel,
+                                                      (erts_aint_t) NULL);
     while (fixdel != NULL) {
 	FixedDeletion *fx = fixdel;
 	int ix = fx->slot;
@@ -1142,9 +1148,9 @@ int db_erase_hash(DbTable *tbl, Eterm key, Eterm *ret)
     while(b != 0) {
 	if (has_live_key(tb,b,key,hval)) {
 	    --nitems_diff;
-	    if (nitems_diff == -1 && IS_FIXED(tb)) {
+	    if (nitems_diff == -1 && IS_FIXED(tb)
+                && add_fixed_deletion(tb, ix, 0)) {
 		/* Pseudo remove (no need to keep several of same key) */
-		add_fixed_deletion(tb, ix);
 		b->hvalue = INVALID_HASH;
 	    } else {
 		*bp = b->next;
@@ -1196,9 +1202,8 @@ static int db_erase_object_hash(DbTable *tbl, Eterm object, Eterm *ret)
 	    ++nkeys;
 	    if (db_eq(&tb->common,object, &b->dbterm)) {
 		--nitems_diff;
-		if (nkeys==1 && IS_FIXED(tb)) { /* Pseudo remove */
-		    add_fixed_deletion(tb,ix);
-		    b->hvalue = INVALID_HASH;
+		if (nkeys==1 && IS_FIXED(tb) && add_fixed_deletion(tb,ix,0)) {
+		    b->hvalue = INVALID_HASH;        /* Pseudo remove */
 		    bp = &b->next;
 		    b = b->next;
 		} else {
@@ -1820,14 +1825,17 @@ static int db_select_delete_hash(Process *p,
 	    int did_erase = 0;
 	    if (db_match_dbterm(&tb->common, p, mpi.mp, 0,
 				&(*current)->dbterm, NULL, 0) == am_true) {
+                HashDbTerm *del;
 		if (NFIXED(tb) > fixated_by_me) { /* fixated by others? */
 		    if (slot_ix != last_pseudo_delete) {
-			add_fixed_deletion(tb, slot_ix);
-			last_pseudo_delete = slot_ix;
+                        if (!add_fixed_deletion(tb, slot_ix, fixated_by_me))
+                            goto do_erase;
+                        last_pseudo_delete = slot_ix;
 		    }
 		    (*current)->hvalue = INVALID_HASH;
 		} else {
-		    HashDbTerm *del = *current;
+                do_erase:
+		    del = *current;
 		    *current = (*current)->next;
 		    free_term(tb, del);
 		    did_erase = 1;
@@ -1931,14 +1939,17 @@ static int db_select_delete_continue_hash(Process *p,
 	    int did_erase = 0;
 	    if (db_match_dbterm(&tb->common, p, mp, 0,
 				&(*current)->dbterm, NULL, 0) == am_true) {
+                HashDbTerm *del;
 		if (NFIXED(tb) > fixated_by_me) { /* fixated by others? */
 		    if (slot_ix != last_pseudo_delete) {
-			add_fixed_deletion(tb, slot_ix);
+                        if (!add_fixed_deletion(tb, slot_ix, fixated_by_me))
+                            goto do_erase;
 			last_pseudo_delete = slot_ix;
 		    }
 		    (*current)->hvalue = INVALID_HASH;
 		} else {
-		    HashDbTerm *del = *current;
+                do_erase:
+		    del = *current;
 		    *current = (*current)->next;
 		    free_term(tb, del);
 		    did_erase = 1;
@@ -2089,9 +2100,9 @@ static int db_take_hash(Process *p, DbTable *tbl, Eterm key, Eterm *ret)
             *ret = get_term_list(p, tb, key, hval, b, &bend);
             while (b != bend) {
                 --nitems_diff;
-                if (nitems_diff == -1 && IS_FIXED(tb)) {
+                if (nitems_diff == -1 && IS_FIXED(tb)
+                    && add_fixed_deletion(tb, ix, 0)) {
                     /* Pseudo remove (no need to keep several of same key) */
-                    add_fixed_deletion(tb, ix);
                     bp = &b->next;
                     b->hvalue = INVALID_HASH;
                     b = b->next;
@@ -2131,7 +2142,7 @@ int db_mark_all_deleted_hash(DbTable *tbl)
 
     for (i = 0; i < NACTIVE(tb); i++) {
 	if ((list = BUCKET(tb,i)) != NULL) {
-	    add_fixed_deletion(tb, i);
+	    add_fixed_deletion(tb, i, 0);
 	    do {
 		list->hvalue = INVALID_HASH;
 		list = list->next;
@@ -2190,11 +2201,11 @@ static void db_print_hash(int to, void *to_arg, int show, DbTable *tbl)
 		    erts_print(to, to_arg, "*");
 		if (tb->common.compress) {
 		    Eterm key = GETKEY(tb, list->dbterm.tpl);
-		    erts_print(to, to_arg, "key=%R", key, list->dbterm.tpl);
+		    erts_print(to, to_arg, "key=%T", key);
 		}
 		else {
-		    Eterm obj = make_tuple_rel(list->dbterm.tpl,list->dbterm.tpl);
-		    erts_print(to, to_arg, "%R", obj, list->dbterm.tpl);
+		    Eterm obj = make_tuple(list->dbterm.tpl);
+		    erts_print(to, to_arg, "%T", obj);
 		}
 		if (list->next != 0)
 		    erts_print(to, to_arg, ",");
@@ -2885,9 +2896,6 @@ Ldone:
     handle->dbterm = &b->dbterm;
     handle->flags = flags;
     handle->new_size = b->dbterm.size;
-#if HALFWORD_HEAP
-    handle->abs_vec = NULL;
-#endif
     handle->lck = lck;
     return 1;
 }
@@ -2908,8 +2916,8 @@ db_finalize_dbterm_hash(int cret, DbUpdateHandle* handle)
     ASSERT((&b->dbterm == handle->dbterm) == !(tb->common.compress && handle->flags & DB_MUST_RESIZE));
 
     if (handle->flags & DB_NEW_OBJECT && cret != DB_ERROR_NONE) {
-        if (IS_FIXED(tb)) {
-            add_fixed_deletion(tb, hash_to_ix(tb, b->hvalue));
+        if (IS_FIXED(tb) && add_fixed_deletion(tb, hash_to_ix(tb, b->hvalue),
+                                               0)) {
             b->hvalue = INVALID_HASH;
         } else {
             *bp = b->next;
diff --git a/erts/emulator/beam/erl_db_tree.c b/erts/emulator/beam/erl_db_tree.c
index 465aa566ad..311b69d114 100644
--- a/erts/emulator/beam/erl_db_tree.c
+++ b/erts/emulator/beam/erl_db_tree.c
@@ -280,7 +280,7 @@ struct select_delete_context {
 /*
 ** Forward declarations 
 */
-static TreeDbTerm *linkout_tree(DbTableTree *tb, Eterm key, Eterm* key_base);
+static TreeDbTerm *linkout_tree(DbTableTree *tb, Eterm key);
 static TreeDbTerm *linkout_object_tree(DbTableTree *tb, 
 				       Eterm object);
 static int do_free_tree_cont(DbTableTree *tb, int num_left);
@@ -291,15 +291,15 @@ static int delsub(TreeDbTerm **this);
 static TreeDbTerm *slot_search(Process *p, DbTableTree *tb, Sint slot);
 static TreeDbTerm *find_node(DbTableTree *tb, Eterm key);
 static TreeDbTerm **find_node2(DbTableTree *tb, Eterm key);
-static TreeDbTerm *find_next(DbTableTree *tb, DbTreeStack*, Eterm key, Eterm* kbase);
-static TreeDbTerm *find_prev(DbTableTree *tb, DbTreeStack*, Eterm key, Eterm* kbase);
+static TreeDbTerm *find_next(DbTableTree *tb, DbTreeStack*, Eterm key);
+static TreeDbTerm *find_prev(DbTableTree *tb, DbTreeStack*, Eterm key);
 static TreeDbTerm *find_next_from_pb_key(DbTableTree *tb, DbTreeStack*,
 					 Eterm key);
 static TreeDbTerm *find_prev_from_pb_key(DbTableTree *tb, DbTreeStack*,
 					 Eterm key);
 static void traverse_backwards(DbTableTree *tb,
 			       DbTreeStack*,
-			       Eterm lastkey, Eterm* lk_base,
+			       Eterm lastkey,
 			       int (*doit)(DbTableTree *tb,
 					   TreeDbTerm *,
 					   void *,
@@ -307,7 +307,7 @@ static void traverse_backwards(DbTableTree *tb,
 			       void *context); 
 static void traverse_forward(DbTableTree *tb,
 			     DbTreeStack*,
-			     Eterm lastkey, Eterm* lk_base,
+			     Eterm lastkey,
 			     int (*doit)(DbTableTree *tb,
 					 TreeDbTerm *,
 					 void *,
@@ -315,8 +315,8 @@ static void traverse_forward(DbTableTree *tb,
 			     void *context); 
 static int key_given(DbTableTree *tb, Eterm pattern, TreeDbTerm **ret,
 		     Eterm *partly_bound_key);
-static Sint cmp_partly_bound(Eterm partly_bound_key, Eterm bound_key, Eterm* bk_base);
-static Sint do_cmp_partly_bound(Eterm a, Eterm b, Eterm* b_base, int *done);
+static Sint cmp_partly_bound(Eterm partly_bound_key, Eterm bound_key);
+static Sint do_cmp_partly_bound(Eterm a, Eterm b, int *done);
 
 static int analyze_pattern(DbTableTree *tb, Eterm pattern, 
 			   struct mp_info *mpi);
@@ -517,7 +517,7 @@ static int db_next_tree(Process *p, DbTable *tbl, Eterm key, Eterm *ret)
     if (is_atom(key) && key == am_EOT)
 	return DB_ERROR_BADKEY;
     stack = get_any_stack(tb);
-    this = find_next(tb, stack, key, NULL);
+    this = find_next(tb, stack, key);
     release_stack(tb,stack);
     if (this == NULL) {
 	*ret = am_EOT;
@@ -563,7 +563,7 @@ static int db_prev_tree(Process *p, DbTable *tbl, Eterm key, Eterm *ret)
     if (is_atom(key) && key == am_EOT)
 	return DB_ERROR_BADKEY;
     stack = get_any_stack(tb);
-    this = find_prev(tb, stack, key, NULL);
+    this = find_prev(tb, stack, key);
     release_stack(tb,stack);
     if (this == NULL) {
 	*ret = am_EOT;
@@ -573,19 +573,13 @@ static int db_prev_tree(Process *p, DbTable *tbl, Eterm key, Eterm *ret)
     return DB_ERROR_NONE;
 }
 
-static ERTS_INLINE Sint cmp_key(DbTableTree* tb, Eterm key, Eterm* key_base,
-			       TreeDbTerm* obj)
-{
-    return cmp_rel(key, key_base,
-		   GETKEY(tb,obj->dbterm.tpl), obj->dbterm.tpl);
+static ERTS_INLINE Sint cmp_key(DbTableTree* tb, Eterm key, TreeDbTerm* obj) {
+    return CMP(key, GETKEY(tb,obj->dbterm.tpl));
 }
 
-static ERTS_INLINE int cmp_key_eq(DbTableTree* tb, Eterm key, Eterm* key_base,
-				  TreeDbTerm* obj)
-{
+static ERTS_INLINE int cmp_key_eq(DbTableTree* tb, Eterm key, TreeDbTerm* obj) {
     Eterm obj_key = GETKEY(tb,obj->dbterm.tpl);
-    return is_same(key, key_base, obj_key, obj->dbterm.tpl)
-	|| cmp_rel(key, key_base, obj_key, obj->dbterm.tpl) == 0;
+    return is_same(key, obj_key) || CMP(key, obj_key) == 0;
 }
 
 static int db_put_tree(DbTable *tbl, Eterm obj, int key_clash_fail)
@@ -619,7 +613,7 @@ static int db_put_tree(DbTable *tbl, Eterm obj, int key_clash_fail)
 	    (*this)->balance = 0;
 	    (*this)->left = (*this)->right = NULL;
 	    break;
-	} else if ((c = cmp_key(tb, key, NULL, *this)) < 0) {
+	} else if ((c = cmp_key(tb, key, *this)) < 0) {
 	    /* go lefts */
 	    dstack[dpos++] = DIR_LEFT;
 	    tstack[tpos++] = this;
@@ -774,7 +768,7 @@ static int db_erase_tree(DbTable *tbl, Eterm key, Eterm *ret)
 
     *ret = am_true;
 
-    if ((res = linkout_tree(tb, key, NULL)) != NULL) {
+    if ((res = linkout_tree(tb, key)) != NULL) {
 	free_term(tb, res);
     }
     return DB_ERROR_NONE;
@@ -970,15 +964,15 @@ static int db_select_continue_tree(Process *p,
     stack = get_any_stack(tb);
     if (chunk_size) {
 	if (reverse) {
-	    traverse_backwards(tb, stack, lastkey, NULL, &doit_select_chunk, &sc);
+	    traverse_backwards(tb, stack, lastkey, &doit_select_chunk, &sc);
 	} else {
-	    traverse_forward(tb, stack, lastkey, NULL, &doit_select_chunk, &sc);
+	    traverse_forward(tb, stack, lastkey, &doit_select_chunk, &sc);
 	}
     } else {
 	if (reverse) {
-	    traverse_forward(tb, stack, lastkey, NULL, &doit_select, &sc);
+	    traverse_forward(tb, stack, lastkey, &doit_select, &sc);
 	} else {
-	    traverse_backwards(tb, stack, lastkey, NULL, &doit_select, &sc);
+	    traverse_backwards(tb, stack, lastkey, &doit_select, &sc);
 	}
     }
     release_stack(tb,stack);
@@ -1003,9 +997,9 @@ static int db_select_continue_tree(Process *p,
 	    }
 
 	    key = GETKEY(tb, sc.lastobj);
-	    sz = size_object_rel(key,sc.lastobj);
+	    sz = size_object(key);
 	    hp = HAlloc(p, 9 + sz);
-	    key = copy_struct_rel(key, sz, &hp, &MSO(p), sc.lastobj, NULL);
+	    key = copy_struct(key, sz, &hp, &MSO(p));
 	    continuation = TUPLE8
 		(hp,
 		 tptr[1],
@@ -1026,8 +1020,8 @@ static int db_select_continue_tree(Process *p,
     key = GETKEY(tb, sc.lastobj);
     if (chunk_size) {
 	if (end_condition != NIL && 
-	    ((!reverse && cmp_partly_bound(end_condition,key,sc.lastobj) < 0) ||
-	     (reverse && cmp_partly_bound(end_condition,key,sc.lastobj) > 0))) {
+	    ((!reverse && cmp_partly_bound(end_condition,key) < 0) ||
+	     (reverse && cmp_partly_bound(end_condition,key) > 0))) {
 	    /* done anyway */
 	    if (!sc.got) {
 		RET_TO_BIF(am_EOT, DB_ERROR_NONE);
@@ -1039,16 +1033,16 @@ static int db_select_continue_tree(Process *p,
 	}
     } else {
 	if (end_condition != NIL && 
-	    ((!reverse && cmp_partly_bound(end_condition,key,sc.lastobj) > 0) ||
-	     (reverse && cmp_partly_bound(end_condition,key,sc.lastobj) < 0))) {
+	    ((!reverse && cmp_partly_bound(end_condition,key) > 0) ||
+	     (reverse && cmp_partly_bound(end_condition,key) < 0))) {
 	    /* done anyway */
 	    RET_TO_BIF(sc.accum,DB_ERROR_NONE);
 	}
     }
     /* Not done yet, let's trap. */
-    sz = size_object_rel(key,sc.lastobj);
+    sz = size_object(key);
     hp = HAlloc(p, 9 + sz);
-    key = copy_struct_rel(key, sz, &hp, &MSO(p), sc.lastobj, NULL);
+    key = copy_struct(key, sz, &hp, &MSO(p));
     continuation = TUPLE8
 	(hp,
 	 tptr[1],
@@ -1075,7 +1069,6 @@ static int db_select_tree(Process *p, DbTable *tbl,
     struct select_context sc;
     struct mp_info mpi;
     Eterm lastkey = THE_NON_VALUE;
-    Eterm* lk_base = NULL;
     Eterm key;
     Eterm continuation;
     unsigned sz;
@@ -1127,20 +1120,18 @@ static int db_select_tree(Process *p, DbTable *tbl,
 	if (mpi.some_limitation) {
 	    if ((this = find_prev_from_pb_key(tb, stack, mpi.least)) != NULL) {
 		lastkey = GETKEY(tb, this->dbterm.tpl);
-		lk_base = this->dbterm.tpl;
 	    }
 	    sc.end_condition = mpi.most;
 	}
-	traverse_forward(tb, stack, lastkey, lk_base, &doit_select, &sc);
+	traverse_forward(tb, stack, lastkey, &doit_select, &sc);
     } else {
 	if (mpi.some_limitation) {
 	    if ((this = find_next_from_pb_key(tb, stack, mpi.most)) != NULL) {
 		lastkey = GETKEY(tb, this->dbterm.tpl);
-		lk_base = this->dbterm.tpl;
 	    }
 	    sc.end_condition = mpi.least;
 	}
-	traverse_backwards(tb, stack, lastkey, lk_base, &doit_select, &sc);
+	traverse_backwards(tb, stack, lastkey, &doit_select, &sc);
     }
     release_stack(tb,stack);
 #ifdef HARDDEBUG
@@ -1153,9 +1144,9 @@ static int db_select_tree(Process *p, DbTable *tbl,
     }
 
     key = GETKEY(tb, sc.lastobj);
-    sz = size_object_rel(key, sc.lastobj);
+    sz = size_object(key);
     hp = HAlloc(p, 9 + sz + PROC_BIN_SIZE);
-    key = copy_struct_rel(key, sz, &hp, &MSO(p), sc.lastobj, NULL);
+    key = copy_struct(key, sz, &hp, &MSO(p));
     if (mpi.all_objects)
 	(mpi.mp)->flags |= BIN_FLAG_ALL_OBJECTS;
     mpb=db_make_mp_binary(p,mpi.mp,&hp);
@@ -1236,7 +1227,7 @@ static int db_select_count_continue_tree(Process *p,
     }
 
     stack = get_any_stack(tb);
-    traverse_backwards(tb, stack, lastkey, NULL, &doit_select_count, &sc);
+    traverse_backwards(tb, stack, lastkey, &doit_select_count, &sc);
     release_stack(tb,stack);
 
     BUMP_REDS(p, 1000 - sc.max);
@@ -1246,12 +1237,12 @@ static int db_select_count_continue_tree(Process *p,
     }	
     key = GETKEY(tb, sc.lastobj);
     if (end_condition != NIL && 
-	(cmp_partly_bound(end_condition,key,sc.lastobj) > 0)) {
+	(cmp_partly_bound(end_condition,key) > 0)) {
 	/* done anyway */
 	RET_TO_BIF(make_small(sc.got),DB_ERROR_NONE);
     }
     /* Not done yet, let's trap. */
-    sz = size_object_rel(key, sc.lastobj);
+    sz = size_object(key);
     if (IS_USMALL(0, sc.got)) {
 	hp = HAlloc(p, sz + 6);
 	egot = make_small(sc.got);
@@ -1261,7 +1252,7 @@ static int db_select_count_continue_tree(Process *p,
 	egot = uint_to_big(sc.got, hp);
 	hp += BIG_UINT_HEAP_SIZE;
     }
-    key = copy_struct_rel(key, sz, &hp, &MSO(p), sc.lastobj, NULL);
+    key = copy_struct(key, sz, &hp, &MSO(p));
     continuation = TUPLE5
 	(hp,
 	 tptr[1],
@@ -1284,7 +1275,6 @@ static int db_select_count_tree(Process *p, DbTable *tbl,
     struct select_count_context sc;
     struct mp_info mpi;
     Eterm lastkey = THE_NON_VALUE;
-    Eterm* lk_base = NULL;
     Eterm key;
     Eterm continuation;
     unsigned sz;
@@ -1334,12 +1324,11 @@ static int db_select_count_tree(Process *p, DbTable *tbl,
     if (mpi.some_limitation) {
 	if ((this = find_next_from_pb_key(tb, stack, mpi.most)) != NULL) {
 	    lastkey = GETKEY(tb, this->dbterm.tpl);
-	    lk_base = this->dbterm.tpl;
 	}
 	sc.end_condition = mpi.least;
     }
     
-    traverse_backwards(tb, stack, lastkey, lk_base, &doit_select_count, &sc);
+    traverse_backwards(tb, stack, lastkey, &doit_select_count, &sc);
     release_stack(tb,stack);
     BUMP_REDS(p, 1000 - sc.max);
     if (sc.max > 0) {
@@ -1347,7 +1336,7 @@ static int db_select_count_tree(Process *p, DbTable *tbl,
     }
 
     key = GETKEY(tb, sc.lastobj);
-    sz = size_object_rel(key, sc.lastobj);
+    sz = size_object(key);
     if (IS_USMALL(0, sc.got)) {
 	hp = HAlloc(p, sz + PROC_BIN_SIZE + 6);
 	egot = make_small(sc.got);
@@ -1357,7 +1346,7 @@ static int db_select_count_tree(Process *p, DbTable *tbl,
 	egot = uint_to_big(sc.got, hp);
 	hp += BIG_UINT_HEAP_SIZE;
     }
-    key = copy_struct_rel(key, sz, &hp, &MSO(p), sc.lastobj, NULL);
+    key = copy_struct(key, sz, &hp, &MSO(p));
     if (mpi.all_objects)
 	(mpi.mp)->flags |= BIN_FLAG_ALL_OBJECTS;
     mpb = db_make_mp_binary(p,mpi.mp,&hp);
@@ -1388,7 +1377,6 @@ static int db_select_chunk_tree(Process *p, DbTable *tbl,
     struct select_context sc;
     struct mp_info mpi;
     Eterm lastkey = THE_NON_VALUE;
-    Eterm* lk_base = NULL;
     Eterm key;
     Eterm continuation;
     unsigned sz;
@@ -1445,20 +1433,18 @@ static int db_select_chunk_tree(Process *p, DbTable *tbl,
 	if (mpi.some_limitation) {
 	    if ((this = find_next_from_pb_key(tb, stack, mpi.most)) != NULL) {
 		lastkey = GETKEY(tb, this->dbterm.tpl);
-		lk_base = this->dbterm.tpl;
 	    }
 	    sc.end_condition = mpi.least;
 	}
-	traverse_backwards(tb, stack, lastkey, lk_base, &doit_select_chunk, &sc);
+	traverse_backwards(tb, stack, lastkey, &doit_select_chunk, &sc);
     } else {
 	if (mpi.some_limitation) {
 	    if ((this = find_prev_from_pb_key(tb, stack, mpi.least)) != NULL) {
 		lastkey = GETKEY(tb, this->dbterm.tpl);
-		lk_base = this->dbterm.tpl;
 	    }
 	    sc.end_condition = mpi.most;
 	}
-	traverse_forward(tb, stack, lastkey, lk_base, &doit_select_chunk, &sc);
+	traverse_forward(tb, stack, lastkey, &doit_select_chunk, &sc);
     }
     release_stack(tb,stack);
 
@@ -1483,9 +1469,9 @@ static int db_select_chunk_tree(Process *p, DbTable *tbl,
 	}
 
 	key = GETKEY(tb, sc.lastobj);
-	sz = size_object_rel(key, sc.lastobj);
+	sz = size_object(key);
 	hp = HAlloc(p, 9 + sz + PROC_BIN_SIZE);
-	key = copy_struct_rel(key, sz, &hp, &MSO(p), sc.lastobj, NULL);
+	key = copy_struct(key, sz, &hp, &MSO(p));
 	if (mpi.all_objects)
 	    (mpi.mp)->flags |= BIN_FLAG_ALL_OBJECTS;
 	mpb = db_make_mp_binary(p,mpi.mp,&hp);
@@ -1508,9 +1494,9 @@ static int db_select_chunk_tree(Process *p, DbTable *tbl,
     }
 
     key = GETKEY(tb, sc.lastobj);
-    sz = size_object_rel(key, sc.lastobj);
+    sz = size_object(key);
     hp = HAlloc(p, 9 + sz + PROC_BIN_SIZE);
-    key = copy_struct_rel(key, sz, &hp, &MSO(p), sc.lastobj, NULL);
+    key = copy_struct(key, sz, &hp, &MSO(p));
 
     if (mpi.all_objects)
 	(mpi.mp)->flags |= BIN_FLAG_ALL_OBJECTS;
@@ -1586,7 +1572,7 @@ static int db_select_delete_continue_tree(Process *p,
     sc.keypos = tb->common.keypos;
 
     ASSERT(!erts_smp_atomic_read_nob(&tb->is_stack_busy));
-    traverse_backwards(tb, &tb->static_stack, lastkey, NULL, &doit_select_delete, &sc);
+    traverse_backwards(tb, &tb->static_stack, lastkey, &doit_select_delete, &sc);
 
     BUMP_REDS(p, 1000 - sc.max);
 
@@ -1595,11 +1581,11 @@ static int db_select_delete_continue_tree(Process *p,
     }	
     key = GETKEY(tb, (sc.lastterm)->dbterm.tpl);
     if (end_condition != NIL && 
-	cmp_partly_bound(end_condition,key,sc.lastterm->dbterm.tpl) > 0) { /* done anyway */
+	cmp_partly_bound(end_condition,key) > 0) { /* done anyway */
 	RET_TO_BIF(erts_make_integer(sc.accum,p),DB_ERROR_NONE);
     }
     /* Not done yet, let's trap. */
-    sz = size_object_rel(key, sc.lastterm->dbterm.tpl);
+    sz = size_object(key);
     if (IS_USMALL(0, sc.accum)) {
 	hp = HAlloc(p, sz + 6);
 	eaccsum = make_small(sc.accum);
@@ -1609,7 +1595,7 @@ static int db_select_delete_continue_tree(Process *p,
 	eaccsum = uint_to_big(sc.accum, hp);
 	hp += BIG_UINT_HEAP_SIZE;
     }
-    key = copy_struct_rel(key, sz, &hp, &MSO(p), sc.lastterm->dbterm.tpl, NULL);
+    key = copy_struct(key, sz, &hp, &MSO(p));
     continuation = TUPLE5
 	(hp,
 	 tptr[1],
@@ -1630,7 +1616,6 @@ static int db_select_delete_tree(Process *p, DbTable *tbl,
     struct select_delete_context sc;
     struct mp_info mpi;
     Eterm lastkey = THE_NON_VALUE;
-    Eterm* lk_base = NULL;
     Eterm key;
     Eterm continuation;
     unsigned sz;
@@ -1683,12 +1668,11 @@ static int db_select_delete_tree(Process *p, DbTable *tbl,
     if (mpi.some_limitation) {
 	if ((this = find_next_from_pb_key(tb, &tb->static_stack, mpi.most)) != NULL) {
 	    lastkey = GETKEY(tb, this->dbterm.tpl);
-	    lk_base = this->dbterm.tpl;
 	}
 	sc.end_condition = mpi.least;
     }
 
-    traverse_backwards(tb, &tb->static_stack, lastkey, lk_base, &doit_select_delete, &sc);
+    traverse_backwards(tb, &tb->static_stack, lastkey, &doit_select_delete, &sc);
     BUMP_REDS(p, 1000 - sc.max);
 
     if (sc.max > 0) {
@@ -1696,7 +1680,7 @@ static int db_select_delete_tree(Process *p, DbTable *tbl,
     }
 
     key = GETKEY(tb, (sc.lastterm)->dbterm.tpl);
-    sz = size_object_rel(key, sc.lastterm->dbterm.tpl);
+    sz = size_object(key);
     if (IS_USMALL(0, sc.accum)) {
 	hp = HAlloc(p, sz + PROC_BIN_SIZE + 6);
 	eaccsum = make_small(sc.accum);
@@ -1706,7 +1690,7 @@ static int db_select_delete_tree(Process *p, DbTable *tbl,
 	eaccsum = uint_to_big(sc.accum, hp);
 	hp += BIG_UINT_HEAP_SIZE;
     }
-    key = copy_struct_rel(key, sz, &hp, &MSO(p), sc.lastterm->dbterm.tpl, NULL);
+    key = copy_struct(key, sz, &hp, &MSO(p));
     mpb = db_make_mp_binary(p,mpi.mp,&hp);
     
     continuation = TUPLE5
@@ -1734,7 +1718,7 @@ static int db_take_tree(Process *p, DbTable *tbl, Eterm key, Eterm *ret)
     TreeDbTerm *this;
 
     *ret = NIL;
-    this = linkout_tree(tb, key, NULL);
+    this = linkout_tree(tb, key);
     if (this) {
         Eterm copy, *hp, *hend;
 
@@ -1845,9 +1829,7 @@ do_db_tree_foreach_offheap(TreeDbTerm *tdbt,
     do_db_tree_foreach_offheap(tdbt->right, func, arg);
 }
 
-static TreeDbTerm *linkout_tree(DbTableTree *tb, 
-				Eterm key, Eterm* key_base)
-{
+static TreeDbTerm *linkout_tree(DbTableTree *tb, Eterm key) {
     TreeDbTerm **tstack[STACK_NEED];
     int tpos = 0;
     int dstack[STACK_NEED+1];
@@ -1869,7 +1851,7 @@ static TreeDbTerm *linkout_tree(DbTableTree *tb,
     for (;;) {
 	if (!*this) { /* Failure */
 	    return NULL;
-	} else if ((c = cmp_key(tb, key, key_base, *this)) < 0) {
+	} else if ((c = cmp_key(tb, key, *this)) < 0) {
 	    dstack[dpos++] = DIR_LEFT;
 	    tstack[tpos++] = this;
 	    this = &((*this)->left);
@@ -1933,7 +1915,7 @@ static TreeDbTerm *linkout_object_tree(DbTableTree *tb,
     for (;;) {
 	if (!*this) { /* Failure */
 	    return NULL;
-	} else if ((c = cmp_key(tb,key,NULL,*this)) < 0) {
+	} else if ((c = cmp_key(tb,key,*this)) < 0) {
 	    dstack[dpos++] = DIR_LEFT;
 	    tstack[tpos++] = this;
 	    this = &((*this)->left);
@@ -2319,15 +2301,13 @@ done:
  * Find next and previous in sort order
  */
 
-static TreeDbTerm *find_next(DbTableTree *tb, DbTreeStack* stack,
-			     Eterm key, Eterm* key_base)
-{
+static TreeDbTerm *find_next(DbTableTree *tb, DbTreeStack* stack, Eterm key) {
     TreeDbTerm *this;
     TreeDbTerm *tmp;
     Sint c;
 
     if(( this = TOP_NODE(stack)) != NULL) {
-	if (!cmp_key_eq(tb,key,key_base,this)) {
+	if (!cmp_key_eq(tb,key,this)) {
 	    /* Start from the beginning */
 	    stack->pos = stack->slot = 0;
 	}
@@ -2337,7 +2317,7 @@ static TreeDbTerm *find_next(DbTableTree *tb, DbTreeStack* stack,
 	    return NULL;
 	for (;;) {
 	    PUSH_NODE(stack, this);
-	    if (( c = cmp_key(tb,key,key_base,this) ) > 0) {
+	    if (( c = cmp_key(tb,key,this) ) > 0) {
 		if (this->right == NULL) /* We are at the previos 
 					    and the element does
 					    not exist */
@@ -2377,15 +2357,13 @@ static TreeDbTerm *find_next(DbTableTree *tb, DbTreeStack* stack,
     return this;
 }
 
-static TreeDbTerm *find_prev(DbTableTree *tb, DbTreeStack* stack,
-			     Eterm key, Eterm* key_base)
-{
+static TreeDbTerm *find_prev(DbTableTree *tb, DbTreeStack* stack, Eterm key) {
     TreeDbTerm *this;
     TreeDbTerm *tmp;
     Sint c;
 
     if(( this = TOP_NODE(stack)) != NULL) {
-	if (!cmp_key_eq(tb,key,key_base,this)) {
+	if (!cmp_key_eq(tb,key,this)) {
 	    /* Start from the beginning */
 	    stack->pos = stack->slot = 0;
 	}
@@ -2395,7 +2373,7 @@ static TreeDbTerm *find_prev(DbTableTree *tb, DbTreeStack* stack,
 	    return NULL;
 	for (;;) {
 	    PUSH_NODE(stack, this);
-	    if (( c = cmp_key(tb,key,key_base,this) ) < 0) {
+	    if (( c = cmp_key(tb,key,this) ) < 0) {
 		if (this->left == NULL) /* We are at the next 
 					   and the element does
 					   not exist */
@@ -2448,8 +2426,7 @@ static TreeDbTerm *find_next_from_pb_key(DbTableTree *tb, DbTreeStack* stack,
 	return NULL;
     for (;;) {
 	PUSH_NODE(stack, this);
-	if (( c = cmp_partly_bound(key,GETKEY(tb, this->dbterm.tpl),
-				   this->dbterm.tpl) ) >= 0) {
+	if (( c = cmp_partly_bound(key,GETKEY(tb, this->dbterm.tpl))) >= 0) {
 	    if (this->right == NULL) {
 		do {
 		    tmp = POP_NODE(stack);
@@ -2482,8 +2459,7 @@ static TreeDbTerm *find_prev_from_pb_key(DbTableTree *tb, DbTreeStack* stack,
 	return NULL;
     for (;;) {
 	PUSH_NODE(stack, this);
-	if (( c = cmp_partly_bound(key,GETKEY(tb, this->dbterm.tpl),
-				   this->dbterm.tpl) ) <= 0) {
+	if (( c = cmp_partly_bound(key,GETKEY(tb, this->dbterm.tpl))) <= 0) {
 	    if (this->left == NULL) {
 		do {
 		    tmp = POP_NODE(stack);
@@ -2514,10 +2490,10 @@ static TreeDbTerm *find_node(DbTableTree *tb, Eterm key)
     DbTreeStack* stack = get_static_stack(tb);
 
     if(!stack || EMPTY_NODE(stack)
-       || !cmp_key_eq(tb, key, NULL, (this=TOP_NODE(stack)))) {
+       || !cmp_key_eq(tb, key, (this=TOP_NODE(stack)))) {
 
 	this = tb->root;
-	while (this != NULL && (res = cmp_key(tb,key,NULL,this)) != 0) {
+	while (this != NULL && (res = cmp_key(tb,key,this)) != 0) {
 	    if (res < 0)
 		this = this->left;
 	    else
@@ -2539,7 +2515,7 @@ static TreeDbTerm **find_node2(DbTableTree *tb, Eterm key)
     Sint res;
 
     this = &tb->root;
-    while ((*this) != NULL && (res = cmp_key(tb, key, NULL, *this)) != 0) {
+    while ((*this) != NULL && (res = cmp_key(tb, key, *this)) != 0) {
 	if (res < 0)
 	    this = &((*this)->left);
 	else
@@ -2589,9 +2565,6 @@ db_lookup_dbterm_tree(Process *p, DbTable *tbl, Eterm key, Eterm obj,
     handle->flags = flags;
     handle->bp = (void**) pp;
     handle->new_size = (*pp)->dbterm.size;
-#if HALFWORD_HEAP
-    handle->abs_vec = NULL;
-#endif
     return 1;
 }
 
@@ -2622,7 +2595,7 @@ db_finalize_dbterm_tree(int cret, DbUpdateHandle *handle)
  */
 static void traverse_backwards(DbTableTree *tb,
 			       DbTreeStack* stack,
-			       Eterm lastkey, Eterm* lk_base,
+			       Eterm lastkey,
 			       int (*doit)(DbTableTree *,
 					   TreeDbTerm *,
 					   void *,
@@ -2641,16 +2614,15 @@ static void traverse_backwards(DbTableTree *tb,
 	    this = this->right;
 	}
 	this = TOP_NODE(stack);
-	next = find_prev(tb, stack, GETKEY(tb, this->dbterm.tpl),
-			 this->dbterm.tpl);
+	next = find_prev(tb, stack, GETKEY(tb, this->dbterm.tpl));
 	if (!((*doit)(tb, this, context, 0)))
 	    return;
     } else {
-	next = find_prev(tb, stack, lastkey, lk_base);
+	next = find_prev(tb, stack, lastkey);
     }
 
     while ((this = next) != NULL) {
-	next = find_prev(tb, stack, GETKEY(tb, this->dbterm.tpl), this->dbterm.tpl);
+	next = find_prev(tb, stack, GETKEY(tb, this->dbterm.tpl));
 	if (!((*doit)(tb, this, context, 0)))
 	    return;
     }
@@ -2661,7 +2633,7 @@ static void traverse_backwards(DbTableTree *tb,
  */
 static void traverse_forward(DbTableTree *tb,
 			     DbTreeStack* stack,
-			     Eterm lastkey, Eterm* lk_base,
+			     Eterm lastkey,
 			     int (*doit)(DbTableTree *,
 					 TreeDbTerm *,
 					 void *,
@@ -2680,15 +2652,15 @@ static void traverse_forward(DbTableTree *tb,
 	    this = this->left;
 	}
 	this = TOP_NODE(stack);
-	next = find_next(tb, stack, GETKEY(tb, this->dbterm.tpl), this->dbterm.tpl);
+	next = find_next(tb, stack, GETKEY(tb, this->dbterm.tpl));
 	if (!((*doit)(tb, this, context, 1)))
 	    return;
     } else {
-	next = find_next(tb, stack, lastkey, lk_base);
+	next = find_next(tb, stack, lastkey);
     }
 
     while ((this = next) != NULL) {
-	next = find_next(tb, stack, GETKEY(tb, this->dbterm.tpl), this->dbterm.tpl);
+	next = find_next(tb, stack, GETKEY(tb, this->dbterm.tpl));
 	if (!((*doit)(tb, this, context, 1)))
 	    return;
     }
@@ -2725,7 +2697,7 @@ static int key_given(DbTableTree *tb, Eterm pattern, TreeDbTerm **ret,
 
 
 
-static Sint do_cmp_partly_bound(Eterm a, Eterm b, Eterm* b_base, int *done)
+static Sint do_cmp_partly_bound(Eterm a, Eterm b, int *done)
 {
     Eterm* aa;
     Eterm* bb;
@@ -2739,44 +2711,44 @@ static Sint do_cmp_partly_bound(Eterm a, Eterm b, Eterm* b_base, int *done)
 	*done = 1;
 	return 0;
     }
-    if (is_same(a,NULL,b,b_base))
+    if (is_same(a,b))
 	return 0;
     
     switch (a & _TAG_PRIMARY_MASK) {
     case TAG_PRIMARY_LIST:
 	if (!is_list(b)) {
-	    return cmp_rel(a,NULL,b,b_base);
+	    return CMP(a,b);
 	}
 	aa = list_val(a);
-	bb = list_val_rel(b,b_base);
+	bb = list_val(b);
 	while (1) {
-	    if ((j = do_cmp_partly_bound(*aa++, *bb++, b_base, done)) != 0 || *done)
+	    if ((j = do_cmp_partly_bound(*aa++, *bb++, done)) != 0 || *done)
 		return j;
-	    if (is_same(*aa, NULL, *bb, b_base))
+	    if (is_same(*aa, *bb))
 		return 0;
 	    if (is_not_list(*aa) || is_not_list(*bb))
-		return do_cmp_partly_bound(*aa, *bb, b_base, done);
+		return do_cmp_partly_bound(*aa, *bb, done);
 	    aa = list_val(*aa);
-	    bb = list_val_rel(*bb,b_base);
+	    bb = list_val(*bb);
 	}
     case TAG_PRIMARY_BOXED:
 	if ((b & _TAG_PRIMARY_MASK) != TAG_PRIMARY_BOXED) {
-	    return cmp_rel(a,NULL,b,b_base);
+	    return CMP(a,b);
 	}
 	a_hdr = ((*boxed_val(a)) & _TAG_HEADER_MASK) >> _TAG_PRIMARY_SIZE;
-	b_hdr = ((*boxed_val_rel(b,b_base)) & _TAG_HEADER_MASK) >> _TAG_PRIMARY_SIZE;
+	b_hdr = ((*boxed_val(b)) & _TAG_HEADER_MASK) >> _TAG_PRIMARY_SIZE;
 	if (a_hdr != b_hdr) {
-	    return cmp_rel(a, NULL, b, b_base);
+	    return CMP(a,b);
 	}
 	if (a_hdr == (_TAG_HEADER_ARITYVAL >> _TAG_PRIMARY_SIZE)) {
 	    aa = tuple_val(a);
-	    bb = tuple_val_rel(b, b_base);
+	    bb = tuple_val(b);
 	    /* compare the arities */
 	    i = arityval(*aa);	/* get the arity*/
 	    if (i < arityval(*bb)) return(-1);
 	    if (i > arityval(*bb)) return(1);
 	    while (i--) {
-		if ((j = do_cmp_partly_bound(*++aa, *++bb, b_base, done)) != 0
+		if ((j = do_cmp_partly_bound(*++aa, *++bb, done)) != 0
 		    || *done) 
 		    return j;
 	    }
@@ -2784,14 +2756,13 @@ static Sint do_cmp_partly_bound(Eterm a, Eterm b, Eterm* b_base, int *done)
 	}
 	/* Drop through */
       default:
-	  return cmp_rel(a, NULL, b, b_base);
+	  return CMP(a,b);
     }
 }
 
-static Sint cmp_partly_bound(Eterm partly_bound_key, Eterm bound_key, Eterm* bk_base)
-{
+static Sint cmp_partly_bound(Eterm partly_bound_key, Eterm bound_key) {
     int done = 0;
-    Sint ret = do_cmp_partly_bound(partly_bound_key, bound_key, bk_base, &done);
+    Sint ret = do_cmp_partly_bound(partly_bound_key, bound_key, &done);
 #ifdef HARDDEBUG
     erts_fprintf(stderr,"\ncmp_partly_bound: %T", partly_bound_key);
     if (ret < 0)
@@ -2800,7 +2771,7 @@ static Sint cmp_partly_bound(Eterm partly_bound_key, Eterm bound_key, Eterm* bk_
 	erts_fprintf(stderr," > ");
     else
 	erts_fprintf(stderr," == ");
-    erts_fprintf(stderr,"%R\n", bound_key, bk_base);
+    erts_fprintf(stderr,"%T\n", bound_key);
 #endif
     return ret;
 }
@@ -3017,12 +2988,10 @@ static int doit_select(DbTableTree *tb, TreeDbTerm *this, void *ptr,
     if (sc->end_condition != NIL && 
 	((forward && 
 	  cmp_partly_bound(sc->end_condition, 
-			   GETKEY_WITH_POS(sc->keypos, this->dbterm.tpl),
-			   this->dbterm.tpl) < 0) ||
+			   GETKEY_WITH_POS(sc->keypos, this->dbterm.tpl)) < 0) ||
 	 (!forward && 
 	  cmp_partly_bound(sc->end_condition, 
-			   GETKEY_WITH_POS(sc->keypos, this->dbterm.tpl),
-			   this->dbterm.tpl) > 0))) {
+			   GETKEY_WITH_POS(sc->keypos, this->dbterm.tpl)) > 0))) {
 	return 0;
     }
     ret = db_match_dbterm(&tb->common,sc->p,sc->mp,sc->all_objects,
@@ -3054,8 +3023,7 @@ static int doit_select_count(DbTableTree *tb, TreeDbTerm *this, void *ptr,
     /* Always backwards traversing */
     if (sc->end_condition != NIL && 
 	(cmp_partly_bound(sc->end_condition, 
-			  GETKEY_WITH_POS(sc->keypos, this->dbterm.tpl),
-			  this->dbterm.tpl) > 0)) {
+			  GETKEY_WITH_POS(sc->keypos, this->dbterm.tpl)) > 0)) {
 	return 0;
     }
     ret = db_match_dbterm(&tb->common, sc->p, sc->mp, 0,
@@ -3081,12 +3049,10 @@ static int doit_select_chunk(DbTableTree *tb, TreeDbTerm *this, void *ptr,
     if (sc->end_condition != NIL && 
 	((forward && 
 	  cmp_partly_bound(sc->end_condition, 
-			   GETKEY_WITH_POS(sc->keypos, this->dbterm.tpl),
-			   this->dbterm.tpl) < 0) ||
+			   GETKEY_WITH_POS(sc->keypos, this->dbterm.tpl)) < 0) ||
 	 (!forward && 
 	  cmp_partly_bound(sc->end_condition, 
-			   GETKEY_WITH_POS(sc->keypos, this->dbterm.tpl),
-			   this->dbterm.tpl) > 0))) {
+			   GETKEY_WITH_POS(sc->keypos, this->dbterm.tpl)) > 0))) {
 	return 0;
     }
 
@@ -3124,14 +3090,13 @@ static int doit_select_delete(DbTableTree *tb, TreeDbTerm *this, void *ptr,
     
     if (sc->end_condition != NIL && 
 	cmp_partly_bound(sc->end_condition, 
-			 GETKEY_WITH_POS(sc->keypos, this->dbterm.tpl),
-			 this->dbterm.tpl) > 0)
+			 GETKEY_WITH_POS(sc->keypos, this->dbterm.tpl)) > 0)
 	return 0;
     ret = db_match_dbterm(&tb->common, sc->p, sc->mp, 0,
 			  &this->dbterm, NULL, 0);
     if (ret == am_true) {
 	key = GETKEY(sc->tb, this->dbterm.tpl);
-	linkout_tree(sc->tb, key, this->dbterm.tpl);
+	linkout_tree(sc->tb, key);
 	sc->erase_lastterm = 1;
 	++sc->accum;
     }
@@ -3157,11 +3122,10 @@ static void do_dump_tree2(DbTableTree* tb, int to, void *to_arg, int show,
 	}
 	else {
 	    prefix = "";
-	    term = make_tuple_rel(t->dbterm.tpl,t->dbterm.tpl);
+	    term = make_tuple(t->dbterm.tpl);
 	}
-	erts_print(to, to_arg, "%*s%s%R (addr = %p, bal = %d)\n",
-		   offset, "", prefix, term, t->dbterm.tpl,
-		   t, t->balance);
+	erts_print(to, to_arg, "%*s%s%T (addr = %p, bal = %d)\n",
+		   offset, "", prefix, term, t, t->balance);
     }
     do_dump_tree2(tb, to, to_arg, show, t->left, offset + 4); 
 }
diff --git a/erts/emulator/beam/erl_db_util.c b/erts/emulator/beam/erl_db_util.c
index dab357a079..399be6058c 100644
--- a/erts/emulator/beam/erl_db_util.c
+++ b/erts/emulator/beam/erl_db_util.c
@@ -239,11 +239,7 @@ typedef enum {
     matchCall2,
     matchCall3,
     matchPushV,
-#if HALFWORD_HEAP
-    matchPushVGuard,    /* First guard-only variable reference */
-#endif
-    matchPushVResult, /* First variable reference in result, or (if HALFWORD)
-                         in guard if also referenced in result */
+    matchPushVResult, /* First variable reference in result */
     matchPushExpr, /* Push the whole expression we're matching ('$_') */
     matchPushArrayAsList, /* Only when parameter is an Array and 
 			     not an erlang term  (DCOMP_TRACE) */
@@ -310,9 +306,6 @@ DMC_DECLARE_STACK_TYPE(unsigned);
 typedef struct DMCVariable {
     int is_bound;
     int is_in_body;
-#if HALFWORD_HEAP
-    int first_guard_label;  /* to maybe change from PushVGuard to PushVResult */
-#endif
 } DMCVariable;
 
 typedef struct DMCHeap {
@@ -374,7 +367,6 @@ typedef struct MatchVariable {
     Eterm term;
 #ifdef DEBUG
     Process* proc;
-    Eterm* base;
 #endif
 } MatchVariable;
 
@@ -415,7 +407,7 @@ cleanup_match_pseudo_process(ErtsMatchPseudoProcess *mpsp, int keep_heap)
 	else {
 	    int i;
 	    for (i = 0; i < ERTS_DEFAULT_MS_HEAP_SIZE; i++) {
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
 		mpsp->default_heap[i] = (Eterm) 0xdeadbeefdeadbeef;
 #else
 		mpsp->default_heap[i] = (Eterm) 0xdeadbeef;
@@ -1240,7 +1232,7 @@ Eterm erts_match_set_run(Process *p, Binary *mpsp,
 {
     Eterm ret;
 
-    ret = db_prog_match(p, mpsp, NIL, NULL, args, num_args,
+    ret = db_prog_match(p, mpsp, NIL, args, num_args,
 			in_flags, return_flags);
 #if defined(HARDDEBUG)
     if (is_non_value(ret)) {
@@ -1265,7 +1257,7 @@ static Eterm erts_match_set_run_ets(Process *p, Binary *mpsp,
 {
     Eterm ret;
 
-    ret = db_prog_match(p, mpsp, args, NULL, NULL, num_args,
+    ret = db_prog_match(p, mpsp, args, NULL, num_args,
 			ERTS_PAM_COPY_RESULT,
 			return_flags);
 #if defined(HARDDEBUG)
@@ -1755,71 +1747,6 @@ static Eterm dpm_array_to_list(Process *psp, Eterm *arr, int arity)
     return ret;
 }
 
-
-#if HALFWORD_HEAP
-struct heap_checkpoint_t
-{
-    Process *p;
-    Eterm* htop;
-    ErlHeapFragment* mbuf;
-    unsigned used_size;
-    ErlOffHeap off_heap;
-};
-
-static void heap_checkpoint_init(Process* p, struct heap_checkpoint_t* hcp)
-{
-    hcp->p = p;
-    hcp->htop = HEAP_TOP(p);
-    hcp->mbuf = MBUF(p);
-    hcp->used_size = hcp->mbuf ? hcp->mbuf->used_size : 0;
-    hcp->off_heap = MSO(p);
-}
-
-static void heap_checkpoint_revert(struct heap_checkpoint_t* hcp)
-{
-    struct erl_off_heap_header* oh = MSO(hcp->p).first;
-
-    if (oh != hcp->off_heap.first) {
-	ASSERT(oh != NULL);
-	if (hcp->off_heap.first) {
-	    while (oh->next != hcp->off_heap.first) {
-		oh = oh->next;
-	    }
-	    oh->next = NULL;
-	}
-	erts_cleanup_offheap(&MSO(hcp->p));
-	MSO(hcp->p) = hcp->off_heap;
-    }
-    if (MBUF(hcp->p) != hcp->mbuf) {
-	ErlHeapFragment* hf = MBUF(hcp->p);
-	ASSERT(hf != NULL);
-	if (hcp->mbuf) {
-	    while (hf->next != hcp->mbuf) {
-		hf = hf->next;
-	    }
-	    hf->next = NULL;
-	}
-	free_message_buffer(MBUF(hcp->p));
-	MBUF(hcp->p) = hcp->mbuf;
-    }
-    if (hcp->mbuf != NULL && hcp->mbuf->used_size != hcp->used_size) {
-	hcp->mbuf->used_size = hcp->used_size;
-    }
-    HEAP_TOP(hcp->p) = hcp->htop;
-}
-#endif /* HALFWORD_HEAP */
-
-static ERTS_INLINE Eterm copy_object_rel(Process* p, Eterm term, Eterm* base)
-{
-    if (!is_immed(term)) {
-	Uint sz = size_object_rel(term, base);
-	Eterm* top = HAllocX(p, sz, HEAP_XTRA);
-	return copy_struct_rel(term, sz, &top, &MSO(p), base, NULL);
-    }
-    return term;
-}
-
-
 /*
 ** Execution of the match program, this is Pam.
 ** May return THE_NON_VALUE, which is a bailout.
@@ -1827,7 +1754,7 @@ static ERTS_INLINE Eterm copy_object_rel(Process* p, Eterm term, Eterm* base)
 ** i.e. 'DCOMP_TRACE' was specified 
 */
 Eterm db_prog_match(Process *c_p, Binary *bprog,
-		    Eterm term, Eterm* base,
+		    Eterm term,
 		    Eterm *termp,
 		    int arity,
 		    enum erts_pam_run_flags in_flags,
@@ -1855,17 +1782,12 @@ Eterm db_prog_match(Process *c_p, Binary *bprog,
     Eterm bif_args[3];
     int fail_label;
     int atomic_trace;
-#if HALFWORD_HEAP
-    struct heap_checkpoint_t c_p_checkpoint = {};
-#endif
 #ifdef DMC_DEBUG
     Uint *heap_fence;
     Uint *stack_fence;
     Uint save_op;
 #endif /* DMC_DEBUG */
 
-    ASSERT(base==NULL || HALFWORD_HEAP);
-
     mpsp = get_match_pseudo_process(c_p, prog->heap_size);
     psp = &mpsp->process;
 
@@ -1916,11 +1838,7 @@ Eterm db_prog_match(Process *c_p, Binary *bprog,
 			     do_catch != 0 */
 
     *return_flags = 0U;
-
     variables = mpsp->u.variables;
-#if HALFWORD_HEAP
-    c_p_checkpoint.p = NULL;
-#endif
 
 restart:
     ep = &term;
@@ -1931,14 +1849,12 @@ restart:
     fail_label = -1;
     build_proc = psp;
     esdp->current_process = psp;
-    ASSERT_HALFWORD(!c_p_checkpoint.p);
 
 #ifdef DEBUG
     ASSERT(variables == mpsp->u.variables);
     for (i=0; i<prog->num_bindings; i++) {
 	variables[i].term = THE_NON_VALUE;
 	variables[i].proc = NULL;
-	variables[i].base = base;
     }
 #endif
 
@@ -1974,9 +1890,9 @@ restart:
 	    variables[n].term = dpm_array_to_list(psp, termp, arity);
 	    break;
 	case matchTuple: /* *ep is a tuple of arity n */
-	    if (!is_tuple_rel(*ep,base))
+	    if (!is_tuple(*ep))
 		FAIL();
-	    ep = tuple_val_rel(*ep,base);
+	    ep = tuple_val(*ep);
 	    n = *pc++;
 	    if (arityval(*ep) != n)
 		FAIL();
@@ -1984,9 +1900,9 @@ restart:
 	    break;
 	case matchPushT: /* *ep is a tuple of arity n, 
 			    push ptr to first element */
-	    if (!is_tuple_rel(*ep,base))
+	    if (!is_tuple(*ep))
 		FAIL();
-	    tp = tuple_val_rel(*ep,base);
+	    tp = tuple_val(*ep);
 	    n = *pc++;
 	    if (arityval(*tp) != n)
 		FAIL();
@@ -1996,51 +1912,51 @@ restart:
 	case matchList:
 	    if (!is_list(*ep))
 		FAIL();
-	    ep = list_val_rel(*ep,base);
+	    ep = list_val(*ep);
 	    break;
 	case matchPushL:
 	    if (!is_list(*ep))
 		FAIL();
-	    *sp++ = list_val_rel(*ep,base);
+	    *sp++ = list_val(*ep);
 	    ++ep;
 	    break;
         case matchMap:
-            if (!is_map_rel(*ep, base)) {
+            if (!is_map(*ep)) {
                 FAIL();
             }
             n = *pc++;
-            if (is_flatmap_rel(*ep,base)) {
-		if (flatmap_get_size(flatmap_val_rel(*ep, base)) < n) {
+            if (is_flatmap(*ep)) {
+		if (flatmap_get_size(flatmap_val(*ep)) < n) {
 		    FAIL();
 		}
             } else {
-		ASSERT(is_hashmap_rel(*ep,base));
-		if (hashmap_size_rel(*ep, base) < n) {
+		ASSERT(is_hashmap(*ep));
+		if (hashmap_size(*ep) < n) {
 		    FAIL();
 		}
 	    }
-            ep = flatmap_val_rel(*ep, base);
+            ep = flatmap_val(*ep);
             break;
         case matchPushM:
-            if (!is_map_rel(*ep, base)) {
+            if (!is_map(*ep)) {
                 FAIL();
             }
             n = *pc++;
-            if (is_flatmap_rel(*ep,base)) {
-		if (flatmap_get_size(flatmap_val_rel(*ep, base)) < n) {
+            if (is_flatmap(*ep)) {
+		if (flatmap_get_size(flatmap_val(*ep)) < n) {
 		    FAIL();
 		}
 	    } else {
-		ASSERT(is_hashmap_rel(*ep,base));
-		if (hashmap_size_rel(*ep, base) < n) {
+		ASSERT(is_hashmap(*ep));
+		if (hashmap_size(*ep) < n) {
 		    FAIL();
 		}
 	    }
-            *sp++ = flatmap_val_rel(*ep++, base);
+            *sp++ = flatmap_val(*ep++);
             break;
         case matchKey:
             t = (Eterm) *pc++;
-            tp = erts_maps_get_rel(t, make_boxed_rel(ep, base), base);
+            tp = erts_maps_get(t, make_boxed(ep));
             if (!tp) {
                 FAIL();
             }
@@ -2061,29 +1977,29 @@ restart:
 	    break;
 	case matchCmp:
 	    n = *pc++;
-	    if (!eq_rel(variables[n].term, base, *ep, base))
+	    if (!EQ(variables[n].term, *ep))
 		FAIL();
 	    ++ep;
 	    break;
 	case matchEqBin:
 	    t = (Eterm) *pc++;
-	    if (!eq_rel(t,NULL,*ep,base))
+	    if (!EQ(t,*ep))
 		FAIL();
 	    ++ep;
 	    break;
 	case matchEqFloat:
-	    if (!is_float_rel(*ep,base))
+	    if (!is_float(*ep))
 		FAIL();
-	    if (memcmp(float_val_rel(*ep,base) + 1, pc, sizeof(double)))
+	    if (memcmp(float_val(*ep) + 1, pc, sizeof(double)))
 		FAIL();
 	    pc += TermWords(2);
 	    ++ep;
 	    break;
 	case matchEqRef: {
 	    Eterm* epc = (Eterm*)pc;
-	    if (!is_ref_rel(*ep,base))
+	    if (!is_ref(*ep))
 		FAIL();
-	    if (!eq_rel(make_internal_ref_rel(epc, epc), epc, *ep, base)) {
+	    if (!EQ(make_internal_ref(epc), *ep)) {
 		FAIL();
 	    }
 	    i = thing_arityval(*epc);
@@ -2092,9 +2008,9 @@ restart:
 	    break;
 	}
 	case matchEqBig:
-	    if (!is_big_rel(*ep,base))
+	    if (!is_big(*ep))
 		FAIL();
-	    tp = big_val_rel(*ep,base);
+	    tp = big_val(*ep);
 	    {
 		Eterm *epc = (Eterm *) pc;
 		if (*tp != *epc)
@@ -2236,62 +2152,36 @@ restart:
 	    esp -= 2;
 	    esp[-1] = t;
 	    break;
-
-	#if HALFWORD_HEAP
-	case matchPushVGuard:
-	    if (!base) goto case_matchPushV;
-	    /* Build NULL-based copy on pseudo heap for easy disposal */
-	    n = *pc++;
-	    ASSERT(is_value(variables[n].term));
-	    ASSERT(!variables[n].proc);
-	    variables[n].term = copy_object_rel(psp, variables[n].term, base);
-	    *esp++ = variables[n].term;
-	    #ifdef DEBUG
-	    variables[n].proc = psp;
-	    variables[n].base = NULL;
-	    #endif
-	    break;
-	#endif
 	case matchPushVResult:
 	    if (!(in_flags & ERTS_PAM_COPY_RESULT)) goto case_matchPushV;
-
-	    /* Build (NULL-based) copy on callers heap */
-	#if HALFWORD_HEAP
-	    if (!do_catch && !c_p_checkpoint.p) {
-		heap_checkpoint_init(c_p, &c_p_checkpoint);
-	    }
-        #endif
+	    /* Build copy on callers heap */
 	    n = *pc++;
 	    ASSERT(is_value(variables[n].term));
 	    ASSERT(!variables[n].proc);
-	    variables[n].term = copy_object_rel(c_p, variables[n].term, base);
+	    variables[n].term = copy_object_x(variables[n].term, c_p, HEAP_XTRA);
 	    *esp++ = variables[n].term;
 	    #ifdef DEBUG
 	    variables[n].proc = c_p;
-	    variables[n].base = NULL;
 	    #endif
 	    break;
 	case matchPushV:
 	case_matchPushV:
 	    n = *pc++;
 	    ASSERT(is_value(variables[n].term));
-	    ASSERT(!variables[n].base);
 	    *esp++ = variables[n].term;
 	    break;
 	case matchPushExpr:
 	    if (in_flags & ERTS_PAM_COPY_RESULT) {
 		Uint sz;
 		Eterm* top;
-		sz = size_object_rel(term, base);
+		sz = size_object(term);
 		top = HAllocX(build_proc, sz, HEAP_XTRA);
 		if (in_flags & ERTS_PAM_CONTIGUOUS_TUPLE) {
-		    ASSERT(is_tuple_rel(term,base));
-		    *esp++ = copy_shallow_rel(tuple_val_rel(term,base), sz,
-					      &top, &MSO(build_proc), base);
+		    ASSERT(is_tuple(term));
+		    *esp++ = copy_shallow(tuple_val(term), sz, &top, &MSO(build_proc));
 		}
 		else {
-		    *esp++ = copy_struct_rel(term, sz, &top, &MSO(build_proc),
-					     base, NULL);
+		    *esp++ = copy_struct(term, sz, &top, &MSO(build_proc));
 		}
 	    }
 	    else {
@@ -2299,7 +2189,6 @@ restart:
 	    }
 	    break;
 	case matchPushArrayAsList:
-	    ASSERT_HALFWORD(base == NULL);
 	    n = arity; /* Only happens when 'term' is an array */
 	    tp = termp;
 	    ehp = HAllocX(build_proc, n*2, HEAP_XTRA);
@@ -2315,7 +2204,6 @@ restart:
 	    break;
 	case matchPushArrayAsListU:
 	    /* This instruction is NOT efficient. */
-	    ASSERT_HALFWORD(base == NULL);
 	    *esp++  = dpm_array_to_list(build_proc, termp, arity);
 	    break;
 	case matchTrue:
@@ -2420,11 +2308,7 @@ restart:
 	    *esp++ = am_true;
 	    break;
 	case matchIsSeqTrace:
-	    if (SEQ_TRACE_TOKEN(c_p) != NIL
-#ifdef USE_VM_PROBES
-		&& SEQ_TRACE_TOKEN(c_p) != am_have_dt_utag
-#endif
-		)
+            if (have_seqtrace(SEQ_TRACE_TOKEN(c_p)))
 		*esp++ = am_true;
 	    else
 		*esp++ = am_false;
@@ -2448,11 +2332,7 @@ restart:
 	    --esp;
 	    break;
 	case matchGetSeqToken:
-	    if (SEQ_TRACE_TOKEN(c_p) == NIL
-#ifdef USE_VM_PROBES
-		|| SEQ_TRACE_TOKEN(c_p) == am_have_dt_utag
-#endif
-		) 
+            if (have_no_seqtrace(SEQ_TRACE_TOKEN(c_p)))
 		*esp++ = NIL;
 	    else {
 		Eterm sender = SEQ_TRACE_TOKEN_SENDER(c_p);
@@ -2619,13 +2499,6 @@ restart:
 	}
     }
 fail:
-#if HALFWORD_HEAP
-    if (c_p_checkpoint.p) {
-	/* Dispose garbage built by guards on caller heap */
-	heap_checkpoint_revert(&c_p_checkpoint);
-	c_p_checkpoint.p = NULL;
-    }
-#endif
     *return_flags = 0U;
     if (fail_label >= 0) { /* We failed during a "TryMeElse",
 			      lets restart, with the next match
@@ -2788,13 +2661,6 @@ Wterm db_do_read_element(DbUpdateHandle* handle, Sint position)
 {
     Eterm elem = handle->dbterm->tpl[position];
     if (!is_header(elem)) {
-#if HALFWORD_HEAP
-	if (!is_immed(elem)
-	    && !handle->tb->common.compress
-	    && !(handle->abs_vec && handle->abs_vec[position])) {
-	    return rterm2wterm(elem, handle->dbterm->tpl);
-	}
-#endif
 	return elem;
     }
 
@@ -2822,9 +2688,6 @@ void db_do_update_element(DbUpdateHandle* handle,
     Eterm* oldp;
     Uint newval_sz;
     Uint oldval_sz;
-#if HALFWORD_HEAP
-    Eterm* old_base;
-#endif
 
     if (is_both_immed(newval,oldval)) {
 	handle->dbterm->tpl[position] = newval;
@@ -2841,15 +2704,8 @@ void db_do_update_element(DbUpdateHandle* handle,
 						       handle->dbterm);
 	    handle->flags |= DB_MUST_RESIZE;
 	    oldval = handle->dbterm->tpl[position];
-        #if HALFWORD_HEAP
-	    old_base = NULL;
-	#endif
 	}
 	else {
-	#if HALFWORD_HEAP
-	    ASSERT(!handle->abs_vec);
-	    old_base = handle->dbterm->tpl;
-	#endif
 	    if (is_boxed(newval)) {
 		newp = boxed_val(newval);
 		switch (*newp & _TAG_HEADER_MASK) {
@@ -2859,7 +2715,7 @@ void db_do_update_element(DbUpdateHandle* handle,
 		case _TAG_HEADER_HEAP_BIN:
 		    newval_sz = header_arity(*newp) + 1;
 		    if (is_boxed(oldval)) {
-			oldp = boxed_val_rel(oldval,old_base);
+			oldp = boxed_val(oldval);
 			switch (*oldp & _TAG_HEADER_MASK) {
 			case _TAG_HEADER_POS_BIG:
 			case _TAG_HEADER_NEG_BIG:
@@ -2879,20 +2735,13 @@ void db_do_update_element(DbUpdateHandle* handle,
 	    }
 	}
     }
-#if HALFWORD_HEAP
-    else {
-	old_base = (handle->tb->common.compress
-		    || (handle->abs_vec && handle->abs_vec[position])) ?
-	    NULL : handle->dbterm->tpl;
-    }
-#endif
     /* Not possible for simple memcpy or dbterm is already non-contiguous, */
     /* need to realloc... */
 
     newval_sz = is_immed(newval) ? 0 : size_object(newval);
 new_size_set:
 
-    oldval_sz = is_immed(oldval) ? 0 : size_object_rel(oldval,old_base);
+    oldval_sz = is_immed(oldval) ? 0 : size_object(oldval);
 both_size_set:
 
     handle->new_size = handle->new_size - oldval_sz + newval_sz;
@@ -2900,19 +2749,6 @@ both_size_set:
     /* write new value in old dbterm, finalize will make a flat copy */
     handle->dbterm->tpl[position] = newval;
     handle->flags |= DB_MUST_RESIZE;
-
-#if HALFWORD_HEAP
-    if (old_base && newval_sz > 0) {
-	ASSERT(!handle->tb->common.compress);
-	if (!handle->abs_vec) {
-	    int i = header_arity(handle->dbterm->tpl[0]);
-	    handle->abs_vec = erts_alloc(ERTS_ALC_T_TMP, (i+1)*sizeof(char));
-	    sys_memset(handle->abs_vec, 0, i+1);
-	    /* abs_vec[0] not used */
-	}
-	handle->abs_vec[position] = 1;
-    }
-#endif
 }
 
 static ERTS_INLINE byte* db_realloc_term(DbTableCommon* tb, void* old,
@@ -3011,7 +2847,7 @@ static void* copy_to_comp(DbTableCommon* tb, Eterm obj, DbTerm* dest,
     tpl[arity + 1] = alloc_size;
 
     tmp_offheap.first = NULL;
-    tpl[tb->keypos] = copy_struct_rel(key, size_object(key), &top.ep, &tmp_offheap, NULL, tpl);
+    tpl[tb->keypos] = copy_struct(key, size_object(key), &top.ep, &tmp_offheap);
     dest->first_oh = tmp_offheap.first;
     for (i=1; i<=arity; i++) {
 	if (i != tb->keypos) {
@@ -3030,7 +2866,7 @@ static void* copy_to_comp(DbTableCommon* tb, Eterm obj, DbTerm* dest,
 	Eterm* dbg_top = erts_alloc(ERTS_ALC_T_DB_TERM, dest->size * sizeof(Eterm));
 	dest->debug_clone = dbg_top;
 	tmp_offheap.first = dest->first_oh;
-	copy_struct_rel(obj, dest->size, &dbg_top, &tmp_offheap, NULL, dbg_top);
+	copy_struct(obj, dest->size, &dbg_top, &tmp_offheap);
 	dest->first_oh = tmp_offheap.first;
 	ASSERT(dbg_top == dest->debug_clone + dest->size);
     }
@@ -3077,7 +2913,7 @@ void* db_store_term(DbTableCommon *tb, DbTerm* old, Uint offset, Eterm obj)
     newp->size = size;
     top = newp->tpl;
     tmp_offheap.first  = NULL;
-    copy_struct_rel(obj, size, &top, &tmp_offheap, NULL, top);
+    copy_struct(obj, size, &top, &tmp_offheap);
     newp->first_oh = tmp_offheap.first;
 #ifdef DEBUG_CLONE
     newp->debug_clone = NULL;
@@ -3160,29 +2996,8 @@ void db_finalize_resize(DbUpdateHandle* handle, Uint offset)
 
 	tmp_offheap.first = NULL;
 
-    #if HALFWORD_HEAP
-	if (handle->abs_vec) {
-	    int i, arity = header_arity(handle->dbterm->tpl[0]);
-
-	    top[0] = tpl[0];
-	    top += arity + 1;
-	    for (i=1; i<=arity; i++) {
-		Eterm* src_base = handle->abs_vec[i] ? NULL : tpl;
-
-		newDbTerm->tpl[i] = copy_struct_rel(tpl[i],
-						    size_object_rel(tpl[i],src_base),
-						    &top, &tmp_offheap, src_base,
-						    newDbTerm->tpl);
-	    }
-	    newDbTerm->first_oh = tmp_offheap.first;
-	    ASSERT((byte*)top <= (newp + alloc_sz));
-	    erts_free(ERTS_ALC_T_TMP, handle->abs_vec);
-	}
-	else
-    #endif /* HALFWORD_HEAP */
 	{
-	    copy_struct_rel(make_tuple_rel(tpl,tpl), handle->new_size, &top,
-			    &tmp_offheap, tpl, top);
+	    copy_struct(make_tuple(tpl), handle->new_size, &top, &tmp_offheap);
 	    newDbTerm->first_oh = tmp_offheap.first;
 	    ASSERT((byte*)top == (newp + alloc_sz));
 	}
@@ -3199,9 +3014,9 @@ Eterm db_copy_from_comp(DbTableCommon* tb, DbTerm* bp, Eterm** hpp,
     hp[0] = bp->tpl[0];
     *hpp += arity + 1;
 
-    hp[tb->keypos] = copy_struct_rel(bp->tpl[tb->keypos],
-				     size_object_rel(bp->tpl[tb->keypos], bp->tpl),
-				     hpp, off_heap, bp->tpl, NULL);
+    hp[tb->keypos] = copy_struct(bp->tpl[tb->keypos],
+                                 size_object(bp->tpl[tb->keypos]),
+                                 hpp, off_heap);
 
     erts_factory_static_init(&factory, *hpp, bp->size - (arity+1), off_heap);
 
@@ -3221,7 +3036,7 @@ Eterm db_copy_from_comp(DbTableCommon* tb, DbTerm* bp, Eterm** hpp,
 
     ASSERT((*hpp - hp) <= bp->size);
 #ifdef DEBUG_CLONE
-    ASSERT(eq_rel(make_tuple(hp),NULL,make_tuple(bp->debug_clone),bp->debug_clone));
+    ASSERT(EQ(make_tuple(hp),make_tuple(bp->debug_clone)));
 #endif
     return make_tuple(hp);
 }
@@ -3245,14 +3060,14 @@ Eterm db_copy_element_from_ets(DbTableCommon* tb, Process* p,
 	*hpp = erts_produce_heap(&factory, extra, 0);
         erts_factory_close(&factory);
 #ifdef DEBUG_CLONE
-	ASSERT(eq_rel(copy, NULL, obj->debug_clone[pos], obj->debug_clone));
+	ASSERT(EQ(copy, obj->debug_clone[pos]));
 #endif
 	return copy;
     }
     else {
-	Uint sz = size_object_rel(obj->tpl[pos], obj->tpl);
+	Uint sz = size_object(obj->tpl[pos]);
 	*hpp = HAlloc(p, sz + extra);
-	return copy_struct_rel(obj->tpl[pos], sz, hpp, &MSO(p), obj->tpl, NULL);
+	return copy_struct(obj->tpl[pos], sz, hpp, &MSO(p));
     }
 }
 
@@ -3599,26 +3414,10 @@ static DMCRet dmc_one_term(DMCContext *context,
 	{
 	    Eterm* ref_val = internal_ref_val(c);
 	    DMC_PUSH(*text, matchEqRef);
-#if HALFWORD_HEAP
-	    {
-		union {
-		    UWord u;
-		    Uint t[2];
-		} fiddle;
-		ASSERT(thing_arityval(ref_val[0]) == 3);
-		fiddle.t[0] = ref_val[0];
-		fiddle.t[1] = ref_val[1];
-		DMC_PUSH(*text, fiddle.u);
-		fiddle.t[0] = ref_val[2];
-		fiddle.t[1] = ref_val[3];
-		DMC_PUSH(*text, fiddle.u);
-	    }
-#else
 	    n = thing_arityval(ref_val[0]);
 	    for (i = 0; i <= n; ++i) {
 		DMC_PUSH(*text, ref_val[i]);
 	    }
-#endif
 	    break;
 	}
 	case (_TAG_HEADER_POS_BIG >> _TAG_PRIMARY_SIZE):
@@ -3627,53 +3426,19 @@ static DMCRet dmc_one_term(DMCContext *context,
 	    Eterm* bval = big_val(c);
 	    n = thing_arityval(bval[0]);
 	    DMC_PUSH(*text, matchEqBig);
-#if HALFWORD_HEAP
-	    {
-		union {
-		    UWord u;
-		    Uint t[2];
-		} fiddle;
-		ASSERT(n >= 1);
-		fiddle.t[0] = bval[0];
-		fiddle.t[1] = bval[1];
-		DMC_PUSH(*text, fiddle.u);
-		for (i = 2; i <= n; ++i) {
-		    fiddle.t[0] = bval[i];
-		    if (++i <= n) {
-			fiddle.t[1] = bval[i];
-		    } else {
-			fiddle.t[1] = (Uint) 0;
-		    }
-		    DMC_PUSH(*text, fiddle.u);
-		}
-	    }
-#else
 	    for (i = 0; i <= n; ++i) {
 		DMC_PUSH(*text, (Uint) bval[i]);
 	    }
-#endif
 	    break;
 	}
 	case (_TAG_HEADER_FLOAT >> _TAG_PRIMARY_SIZE):
 	    DMC_PUSH(*text,matchEqFloat);
-#if HALFWORD_HEAP
-	    {
-		union {
-		    UWord u;
-		    Uint t[2];
-		} fiddle;
-		fiddle.t[0] = float_val(c)[1];
-		fiddle.t[1] = float_val(c)[2];
-		DMC_PUSH(*text, fiddle.u);
-	    }
-#else
 	    DMC_PUSH(*text, (Uint) float_val(c)[1]);
 #ifdef ARCH_64
 	    DMC_PUSH(*text, (Uint) 0);
 #else
 	    DMC_PUSH(*text, (Uint) float_val(c)[2]);
 #endif
-#endif
 	    break;
 	default: /* BINARY, FUN, VECTOR, or EXTERNAL */
 	    DMC_PUSH(*text, matchEqBin);
@@ -3997,24 +3762,8 @@ static void dmc_add_pushv_variant(DMCContext *context, DMCHeap *heap,
     MatchOps instr = matchPushV;
 
     ASSERT(n < heap->vars_used && v->is_bound);
-    if (context->is_guard) {
-	#if HALFWORD_HEAP
-	if (!v->first_guard_label) {
-	    v->first_guard_label = DMC_STACK_NUM(*text);
-	    ASSERT(v->first_guard_label);
-	    instr = matchPushVGuard; /* may be changed to PushVResult below */
-	}
-	#endif
-    }
-    else { /* body */
-	#if HALFWORD_HEAP
-	if (v->first_guard_label) {
-	    /* Avoid double-copy, copy to result heap at first encounter in guard */
-	    DMC_POKE(*text, v->first_guard_label, matchPushVResult);
-	    v->is_in_body = 1;
-	}
-	#endif
-	if (!v->is_in_body) {
+    if (!context->is_guard) {
+        if(!v->is_in_body) {
 	    instr = matchPushVResult;
 	    v->is_in_body = 1;
 	}
@@ -5443,16 +5192,13 @@ Eterm db_match_dbterm(DbTableCommon* tb, Process* c_p, Binary* bprog,
 			     int all, DbTerm* obj, Eterm** hpp, Uint extra)
 {
     Uint32 dummy;
-    Eterm* base;
     Eterm res;
 
     if (tb->compress) {
 	obj = db_alloc_tmp_uncompressed(tb, obj);
-	base = NULL;
     }
-    else base = HALFWORD_HEAP ? obj->tpl : NULL;
 
-    res = db_prog_match(c_p, bprog, make_tuple_rel(obj->tpl,base), base, NULL, 0,
+    res = db_prog_match(c_p, bprog, make_tuple(obj->tpl), NULL, 0,
 			ERTS_PAM_COPY_RESULT|ERTS_PAM_CONTIGUOUS_TUPLE, &dummy);
 
     if (is_value(res) && hpp!=NULL) {
@@ -5573,7 +5319,7 @@ void db_match_dis(Binary *bp)
 			first = 0;
 		    else
 			erts_printf(", ");
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
 		    erts_printf("0x%016bex", rt->data.ui[ri]);
 #else
 		    erts_printf("0x%08bex", rt->data.ui[ri]);
@@ -5597,7 +5343,7 @@ void db_match_dis(Binary *bp)
 			first = 0;
 		    else
 			erts_printf(", ");
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
 		    erts_printf("0x%016bex", *et);
 #else
 		    erts_printf("0x%08bex", *et);
@@ -5720,13 +5466,6 @@ void db_match_dis(Binary *bp)
 	    ++t;
 	    erts_printf("PushV\t%beu\n", n);
 	    break;
-    #if HALFWORD_HEAP
-	case matchPushVGuard:
-	    n = (Uint) *++t;
-	    ++t;
-	    erts_printf("PushVGuard\t%beu\n", n);
-	    break;
-    #endif
 	case matchPushVResult:
 	    n = (Uint) *++t;
 	    ++t;
diff --git a/erts/emulator/beam/erl_db_util.h b/erts/emulator/beam/erl_db_util.h
index 1ccdc0305b..c0966655cd 100644
--- a/erts/emulator/beam/erl_db_util.h
+++ b/erts/emulator/beam/erl_db_util.h
@@ -90,9 +90,6 @@ typedef struct {
     Uint new_size;
     int flags;
     void* lck;
-#if HALFWORD_HEAP
-    unsigned char* abs_vec;  /* [i] true if dbterm->tpl[i] is absolute Eterm */
-#endif
 } DbUpdateHandle;
 
 
@@ -287,10 +284,10 @@ ERTS_GLB_INLINE Eterm db_copy_key(Process* p, DbTable* tb, DbTerm* obj)
     Eterm key = GETKEY(tb, obj->tpl);
     if IS_CONST(key) return key;
     else {
-	Uint size = size_object_rel(key, obj->tpl);
+	Uint size = size_object(key);
 	Eterm* hp = HAlloc(p, size);
-	Eterm res = copy_struct_rel(key, size, &hp, &MSO(p), obj->tpl, NULL);
-	ASSERT(eq_rel(res,NULL,key,obj->tpl));
+	Eterm res = copy_struct(key, size, &hp, &MSO(p));
+	ASSERT(EQ(res,key));
 	return res;
     }
 }
@@ -302,14 +299,14 @@ ERTS_GLB_INLINE Eterm db_copy_object_from_ets(DbTableCommon* tb, DbTerm* bp,
 	return db_copy_from_comp(tb, bp, hpp, off_heap);
     }
     else {
-	return copy_shallow_rel(bp->tpl, bp->size, hpp, off_heap, bp->tpl);
+	return copy_shallow(bp->tpl, bp->size, hpp, off_heap);
     }
 }
 
 ERTS_GLB_INLINE int db_eq(DbTableCommon* tb, Eterm a, DbTerm* b)
 {
     if (!tb->compress) {
-	return eq_rel(a, NULL, make_tuple_rel(b->tpl,b->tpl), b->tpl);
+	return EQ(a, make_tuple(b->tpl));
     }
     else {
 	return db_eq_comp(tb, a, b);
@@ -435,7 +432,7 @@ Binary *db_match_compile(Eterm *matchexpr, Eterm *guards,
 Eterm db_match_dbterm(DbTableCommon* tb, Process* c_p, Binary* bprog,
 		      int all, DbTerm* obj, Eterm** hpp, Uint extra);
 
-Eterm db_prog_match(Process *p, Binary *prog, Eterm term, Eterm* base,
+Eterm db_prog_match(Process *p, Binary *prog, Eterm term,
 		    Eterm *termp, int arity,
 		    enum erts_pam_run_flags in_flags,
 		    Uint32 *return_flags /* Zeroed on enter */);
diff --git a/erts/emulator/beam/erl_debug.c b/erts/emulator/beam/erl_debug.c
index 77a1e3d7cb..7b13cf2c83 100644
--- a/erts/emulator/beam/erl_debug.c
+++ b/erts/emulator/beam/erl_debug.c
@@ -189,6 +189,9 @@ pdisplay1(int to, void *to_arg, Process* p, Eterm obj)
     case BINARY_DEF:
 	erts_print(to, to_arg, "#Bin");
 	break;
+    case MATCHSTATE_DEF:
+        erts_print(to, to_arg, "#Matchstate");
+        break;
     default:
 	erts_print(to, to_arg, "unknown object %x", obj);
     }
@@ -309,6 +312,8 @@ void erts_check_for_holes(Process* p)
     p->last_htop = HEAP_TOP(p);
 
     for (hf = MBUF(p); hf != 0; hf = hf->next) {
+	if (hf == p->heap_hfrag)
+	    continue;
 	if (hf == p->last_mbuf) {
 	    break;
 	}
@@ -399,7 +404,7 @@ void verify_process(Process *p)
         erl_exit(1,"Wild pointer found in " name " of %T!\n",p->common.id); }
 
 
-    ErlMessage* mp = p->msg.first;
+    ErtsMessage* mp = p->msg.first;
 
     VERBOSE(DEBUG_MEMORY,("Verify process: %T...\n",p->common.id));
 
@@ -413,7 +418,7 @@ void verify_process(Process *p)
     erts_check_heap(p);
 
     if (p->dictionary)
-        VERIFY_AREA("dictionary",p->dictionary->data, p->dictionary->used);
+        VERIFY_AREA("dictionary", ERTS_PD_START(p->dictionary), ERTS_PD_SIZE(p->dictionary));
     VERIFY_ETERM("seq trace token",p->seq_trace_token);
     VERIFY_ETERM("group leader",p->group_leader);
     VERIFY_ETERM("fvalue",p->fvalue);
@@ -528,7 +533,7 @@ static void print_process_memory(Process *p)
                 PTR_SIZE, "PCB", dashes, dashes, dashes, dashes);
 
     if (p->msg.first != NULL) {
-        ErlMessage* mp;
+        ErtsMessage* mp;
         erts_printf("  Message Queue:\n");
         mp = p->msg.first;
         while (mp != NULL) {
@@ -539,8 +544,8 @@ static void print_process_memory(Process *p)
     }
 
     if (p->dictionary != NULL) {
-        int n = p->dictionary->used;
-        Eterm *ptr = p->dictionary->data;
+        int n = ERTS_PD_SIZE(p->dictionary);
+        Eterm *ptr = ERTS_PD_START(p->dictionary);
         erts_printf("  Dictionary: ");
         while (n--) erts_printf("0x%0*lx ",PTR_SIZE,(unsigned long)ptr++);
         erts_printf("\n");
@@ -628,29 +633,4 @@ void print_memory_info(Process *p)
     }
     erts_printf("+-----------------%s-%s-%s-%s-+\n",dashes,dashes,dashes,dashes);
 }
-#if !HEAP_ON_C_STACK && defined(DEBUG)
-Eterm *erts_debug_allocate_tmp_heap(int size, Process *p)
-{
-    ErtsSchedulerData *sd = ((p == NULL) ? erts_get_scheduler_data() : ERTS_PROC_GET_SCHDATA(p));
-    int offset = sd->num_tmp_heap_used;
-
-    ASSERT(offset+size <= TMP_HEAP_SIZE);
-    return (sd->tmp_heap)+offset;
-}
-void erts_debug_use_tmp_heap(int size, Process *p)
-{
-    ErtsSchedulerData *sd = ((p == NULL) ? erts_get_scheduler_data() : ERTS_PROC_GET_SCHDATA(p));
-
-    sd->num_tmp_heap_used += size;
-    ASSERT(sd->num_tmp_heap_used <= TMP_HEAP_SIZE);
-}
-void erts_debug_unuse_tmp_heap(int size, Process *p)
-{
-    ErtsSchedulerData *sd = ((p == NULL) ? erts_get_scheduler_data() : ERTS_PROC_GET_SCHDATA(p));
-
-    sd->num_tmp_heap_used -= size;
-    ASSERT(sd->num_tmp_heap_used >= 0);
-}
-#endif
 #endif
-
diff --git a/erts/emulator/beam/erl_debug.h b/erts/emulator/beam/erl_debug.h
index 4905e64f07..029320691d 100644
--- a/erts/emulator/beam/erl_debug.h
+++ b/erts/emulator/beam/erl_debug.h
@@ -48,6 +48,7 @@
 #define DEBUG_THREADS      0x0010    /* Thread-related stuff                 */
 #define DEBUG_PROCESSES    0x0020    /* Process creation and removal         */
 #define DEBUG_MEMORY       0x0040    /* Display results of memory checks     */
+#define DEBUG_SHCOPY       0x0080    /* Sharing-preserving copying of terms  */
 
 extern Uint32 verbose;
 
@@ -92,10 +93,4 @@ extern void print_tagged_memory(Eterm *start, Eterm *end);
 extern void print_untagged_memory(Eterm *start, Eterm *end);
 extern void print_memory(Process *p);
 extern void print_memory_info(Process *p);
-#if defined(DEBUG) && !HEAP_ON_C_STACK
-extern Eterm *erts_debug_allocate_tmp_heap(int, Process *);
-extern void erts_debug_use_tmp_heap(int, Process *);
-extern void erts_debug_unuse_tmp_heap(int, Process *);
-#endif
-
 #endif /* _ERL_DEBUG_H_ */
diff --git a/erts/emulator/beam/erl_driver.h b/erts/emulator/beam/erl_driver.h
index 6b406d069c..0636171ad1 100644
--- a/erts/emulator/beam/erl_driver.h
+++ b/erts/emulator/beam/erl_driver.h
@@ -57,10 +57,6 @@
 #  define SIZEOF_LONG_LONG_SAVED__ SIZEOF_LONG_LONG
 #  undef SIZEOF_LONG_LONG
 #endif
-#ifdef HALFWORD_HEAP_EMULATOR
-#  define HALFWORD_HEAP_EMULATOR_SAVED__ HALFWORD_HEAP_EMULATOR
-#  undef HALFWORD_HEAP_EMULATOR
-#endif
 #include "erl_int_sizes_config.h"
 #if defined(SIZEOF_CHAR_SAVED__) && SIZEOF_CHAR_SAVED__ != SIZEOF_CHAR
 #  error SIZEOF_CHAR mismatch
@@ -78,11 +74,6 @@
 #  error SIZEOF_LONG_LONG mismatch
 #endif
 
-/* This is OK to override by the NIF/driver implementor */
-#if defined(HALFWORD_HEAP_EMULATOR_SAVED__) && !defined(HALFWORD_HEAP_EMULATOR)
-#define HALFWORD_HEAP_EMULATOR HALFWORD_HEAP_EMULATOR_SAVED__
-#endif
-
 #include "erl_drv_nif.h"
 
 #include <stdlib.h>
@@ -134,7 +125,7 @@ typedef struct {
 
 #define ERL_DRV_EXTENDED_MARKER		(0xfeeeeeed)
 #define ERL_DRV_EXTENDED_MAJOR_VERSION	3
-#define ERL_DRV_EXTENDED_MINOR_VERSION	2
+#define ERL_DRV_EXTENDED_MINOR_VERSION	3
 
 /*
  * The emulator will refuse to load a driver with a major version
@@ -172,6 +163,7 @@ typedef struct {
 #define ERL_DRV_FLAG_USE_PORT_LOCKING	(1 << 0)
 #define ERL_DRV_FLAG_SOFT_BUSY		(1 << 1)
 #define ERL_DRV_FLAG_NO_BUSY_MSGQ	(1 << 2)
+#define ERL_DRV_FLAG_USE_INIT_ACK	(1 << 3)
 
 /*
  * Integer types
@@ -386,22 +378,23 @@ typedef struct erl_drv_entry {
 
 #ifdef STATIC_ERLANG_DRIVER
 #  define ERLANG_DRIVER_NAME(NAME) NAME ## _driver_init
+#  define ERL_DRIVER_EXPORT
 #else
 #  define ERLANG_DRIVER_NAME(NAME) driver_init
+#  if defined(__GNUC__) && __GNUC__ >= 4
+#    define ERL_DRIVER_EXPORT __attribute__ ((visibility("default")))
+#  elif defined (__SUNPRO_C) && (__SUNPRO_C >= 0x550)
+#    define ERL_DRIVER_EXPORT __global
+#  else
+#    define ERL_DRIVER_EXPORT
+#  endif
 #endif
 
-/* For windows dynamic drivers */
 #ifndef ERL_DRIVER_TYPES_ONLY
 
-#if defined(__WIN32__)
-#  define DRIVER_INIT(DRIVER_NAME) \
-  __declspec(dllexport) ErlDrvEntry* ERLANG_DRIVER_NAME(DRIVER_NAME)(void);	\
-    __declspec(dllexport) ErlDrvEntry* ERLANG_DRIVER_NAME(DRIVER_NAME)(void)
-#else 
-#  define DRIVER_INIT(DRIVER_NAME) \
-    ErlDrvEntry* ERLANG_DRIVER_NAME(DRIVER_NAME)(void); \
-    ErlDrvEntry* ERLANG_DRIVER_NAME(DRIVER_NAME)(void)
-#endif
+#define DRIVER_INIT(DRIVER_NAME) \
+    ERL_DRIVER_EXPORT ErlDrvEntry* ERLANG_DRIVER_NAME(DRIVER_NAME)(void); \
+    ERL_DRIVER_EXPORT ErlDrvEntry* ERLANG_DRIVER_NAME(DRIVER_NAME)(void)
 
 #define ERL_DRV_BUSY_MSGQ_DISABLED	(~((ErlDrvSizeT) 0))
 #define ERL_DRV_BUSY_MSGQ_READ_ONLY	((ErlDrvSizeT) 0)
@@ -696,18 +689,14 @@ EXTERN int driver_dl_close(void *);
 EXTERN char *driver_dl_error(void);
 
 /* environment */
-EXTERN int erl_drv_putenv(char *key, char *value);
-EXTERN int erl_drv_getenv(char *key, char *value, size_t *value_size);
-
-#ifdef __OSE__
-typedef ErlDrvUInt ErlDrvOseEventId;
-EXTERN union SIGNAL *erl_drv_ose_get_signal(ErlDrvEvent ev);
-EXTERN ErlDrvEvent erl_drv_ose_event_alloc(SIGSELECT sig, ErlDrvOseEventId handle,
-					   ErlDrvOseEventId (*resolve_signal)(union SIGNAL *sig), void *extra);
-EXTERN void erl_drv_ose_event_free(ErlDrvEvent ev);
-EXTERN void erl_drv_ose_event_fetch(ErlDrvEvent ev, SIGSELECT *sig,
-                  ErlDrvOseEventId *handle, void **extra);
-#endif
+EXTERN int erl_drv_putenv(const char *key, char *value);
+EXTERN int erl_drv_getenv(const char *key, char *value, size_t *value_size);
+
+/* spawn start init ack */
+EXTERN void erl_drv_init_ack(ErlDrvPort ix, ErlDrvData res);
+
+/* set the pid seen in port_info */
+EXTERN void erl_drv_set_os_pid(ErlDrvPort ix, ErlDrvSInt pid);
 
 #endif /* !ERL_DRIVER_TYPES_ONLY */
 
diff --git a/erts/emulator/beam/erl_fun.c b/erts/emulator/beam/erl_fun.c
index 4268e2d40a..cff476694c 100644
--- a/erts/emulator/beam/erl_fun.c
+++ b/erts/emulator/beam/erl_fun.c
@@ -66,6 +66,9 @@ erts_init_fun_table(void)
     f.cmp  = (HCMP_FUN) fun_cmp;
     f.alloc = (HALLOC_FUN) fun_alloc;
     f.free = (HFREE_FUN) fun_free;
+    f.meta_alloc = (HMALLOC_FUN) erts_alloc;
+    f.meta_free = (HMFREE_FUN) erts_free;
+    f.meta_print = (HMPRINT_FUN) erts_print;
 
     hash_init(ERTS_ALC_T_FUN_TABLE, &erts_fun_table, "fun_table", 16, f);
 }
diff --git a/erts/emulator/beam/erl_gc.c b/erts/emulator/beam/erl_gc.c
index d2604f1595..0f86b9a25c 100644
--- a/erts/emulator/beam/erl_gc.c
+++ b/erts/emulator/beam/erl_gc.c
@@ -47,6 +47,17 @@
 #define ERTS_INACT_WR_PB_LEAVE_LIMIT 10
 #define ERTS_INACT_WR_PB_LEAVE_PERCENTAGE 10
 
+#if defined(DEBUG) || 0
+#define ERTS_GC_DEBUG
+#else
+#undef ERTS_GC_DEBUG
+#endif
+#ifdef ERTS_GC_DEBUG
+#  define ERTS_GC_ASSERT ASSERT
+#else
+#  define ERTS_GC_ASSERT(B) ((void) 1)
+#endif
+
 /*
  * Returns number of elements in an array.
  */
@@ -67,10 +78,10 @@
 #define ErtsGcQuickSanityCheck(P)					\
 do {									\
     ASSERT((P)->heap < (P)->hend);					\
-    ASSERT((P)->heap_sz == (P)->hend - (P)->heap);			\
+    ASSERT((p)->abandoned_heap || (P)->heap_sz == (P)->hend - (P)->heap); \
     ASSERT((P)->heap <= (P)->htop && (P)->htop <= (P)->hend);		\
     ASSERT((P)->heap <= (P)->stop && (P)->stop <= (P)->hend);		\
-    ASSERT((P)->heap <= (P)->high_water && (P)->high_water <= (P)->hend);\
+    ASSERT((p)->abandoned_heap || ((P)->heap <= (P)->high_water && (P)->high_water <= (P)->hend)); \
     OverRunCheck((P));							\
 } while (0)
 #else
@@ -98,18 +109,33 @@ typedef struct {
 
 static Uint setup_rootset(Process*, Eterm*, int, Rootset*);
 static void cleanup_rootset(Rootset *rootset);
-static Uint combined_message_size(Process* p);
 static void remove_message_buffers(Process* p);
-static int major_collection(Process* p, int need, Eterm* objv, int nobj, Uint *recl);
-static int minor_collection(Process* p, int need, Eterm* objv, int nobj, Uint *recl);
-static void do_minor(Process *p, Uint new_sz, Eterm* objv, int nobj);
-static Eterm* sweep_rootset(Rootset *rootset, Eterm* htop, char* src, Uint src_size);
-static Eterm* sweep_one_area(Eterm* n_hp, Eterm* n_htop, char* src, Uint src_size);
-static Eterm* sweep_one_heap(Eterm* heap_ptr, Eterm* heap_end, Eterm* htop,
-			     char* src, Uint src_size);
-static Eterm* collect_heap_frags(Process* p, Eterm* heap,
-				 Eterm* htop, Eterm* objv, int nobj);
-static void adjust_after_fullsweep(Process *p, int need, Eterm *objv, int nobj);
+static Eterm *full_sweep_heaps(Process *p,
+			       int hibernate,
+			       Eterm *n_heap, Eterm* n_htop,
+			       char *oh, Uint oh_size,
+			       Eterm *objv, int nobj);
+static int garbage_collect(Process* p, ErlHeapFragment *live_hf_end,
+			   int need, Eterm* objv, int nobj);
+static int major_collection(Process* p, ErlHeapFragment *live_hf_end,
+			    int need, Eterm* objv, int nobj, Uint *recl);
+static int minor_collection(Process* p, ErlHeapFragment *live_hf_end,
+			    int need, Eterm* objv, int nobj, Uint *recl);
+static void do_minor(Process *p, ErlHeapFragment *live_hf_end,
+		     char *mature, Uint mature_size,
+		     Uint new_sz, Eterm* objv, int nobj);
+static Eterm *sweep_new_heap(Eterm *n_hp, Eterm *n_htop,
+			     char* old_heap, Uint old_heap_size);
+static Eterm *sweep_heaps(Eterm *n_hp, Eterm *n_htop,
+			  char* old_heap, Uint old_heap_size);
+static Eterm* sweep_literal_area(Eterm* n_hp, Eterm* n_htop,
+				 char* old_heap, Uint old_heap_size,
+				 char* src, Uint src_size);
+static Eterm* sweep_literals_to_old_heap(Eterm* heap_ptr, Eterm* heap_end, Eterm* htop,
+					 char* src, Uint src_size);
+static Eterm* collect_live_heap_frags(Process* p, ErlHeapFragment *live_hf_end,
+				      Eterm* heap, Eterm* htop, Eterm* objv, int nobj);
+static int adjust_after_fullsweep(Process *p, int need, Eterm *objv, int nobj);
 static void shrink_new_heap(Process *p, Uint new_sz, Eterm *objv, int nobj);
 static void grow_new_heap(Process *p, Uint new_sz, Eterm* objv, int nobj);
 static void sweep_off_heap(Process *p, int fullsweep);
@@ -119,16 +145,16 @@ static void offset_rootset(Process *p, Sint offs, char* area, Uint area_size,
 			   Eterm* objv, int nobj);
 static void offset_off_heap(Process* p, Sint offs, char* area, Uint area_size);
 static void offset_mqueue(Process *p, Sint offs, char* area, Uint area_size);
+static void move_msgq_to_heap(Process *p);
 
 static void init_gc_info(ErtsGCInfo *gcip);
 
 #ifdef HARDDEBUG
 static void disallow_heap_frag_ref_in_heap(Process* p);
 static void disallow_heap_frag_ref_in_old_heap(Process* p);
-static void disallow_heap_frag_ref(Process* p, Eterm* n_htop, Eterm* objv, int nobj);
 #endif
 
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
 # define MAX_HEAP_SIZES 154
 #else
 # define MAX_HEAP_SIZES 59
@@ -147,26 +173,24 @@ typedef struct {
     erts_smp_atomic32_t refc;
 } ErtsGCInfoReq;
 
-#if !HALFWORD_HEAP
-ERTS_SCHED_PREF_QUICK_ALLOC_IMPL(gcireq,
-				 ErtsGCInfoReq,
-				 5,
-				 ERTS_ALC_T_GC_INFO_REQ)
-#else
-static ERTS_INLINE ErtsGCInfoReq *
-gcireq_alloc(void)
+static ERTS_INLINE int
+gc_cost(Uint gc_moved_live_words, Uint resize_moved_words)
 {
-    return erts_alloc(ERTS_ALC_T_GC_INFO_REQ,
-		      sizeof(ErtsGCInfoReq));
-}
+    Sint reds;
 
-static ERTS_INLINE void
-gcireq_free(ErtsGCInfoReq *ptr)
-{
-    erts_free(ERTS_ALC_T_GC_INFO_REQ, ptr);
+    reds = gc_moved_live_words/10;
+    reds += resize_moved_words/100;
+    if (reds < 1)
+	return 1;
+    if (reds > INT_MAX)
+	return INT_MAX;
+    return (int) reds;
 }
-#endif
 
+ERTS_SCHED_PREF_QUICK_ALLOC_IMPL(gcireq,
+                                 ErtsGCInfoReq,
+                                 5,
+                                 ERTS_ALC_T_GC_INFO_REQ)
 /*
  * Initialize GC global data.
  */
@@ -208,7 +232,7 @@ erts_init_gc(void)
     }
 
 
-    /* for 32 bit we want max_heap_size to be MAX(32bit) / 4 [words] (and halfword)
+    /* for 32 bit we want max_heap_size to be MAX(32bit) / 4 [words]
      * for 64 bit we want max_heap_size to be MAX(52bit) / 8 [words]
      */
 
@@ -232,10 +256,7 @@ erts_init_gc(void)
       init_gc_info(&esdp->gc_info);
     }
 
-#if !HALFWORD_HEAP
     init_gcireq_alloc();
-#endif
-
 }
 
 /*
@@ -351,10 +372,19 @@ erts_offset_off_heap(ErlOffHeap *ohp, Sint offs, Eterm* low, Eterm* high)
 #undef ptr_within
 
 Eterm
-erts_gc_after_bif_call(Process* p, Eterm result, Eterm* regs, Uint arity)
+erts_gc_after_bif_call_lhf(Process* p, ErlHeapFragment *live_hf_end,
+			   Eterm result, Eterm* regs, Uint arity)
 {
     int cost;
 
+    if (p->flags & F_HIBERNATE_SCHED) {
+	/*
+	 * We just hibernated. We do *not* want to mess
+	 * up the hibernation by an ordinary GC...
+	 */
+	return result;
+    }
+
     if (is_non_value(result)) {
 	if (p->freason == TRAP) {
 	  #if HIPE
@@ -362,21 +392,28 @@ erts_gc_after_bif_call(Process* p, Eterm result, Eterm* regs, Uint arity)
 		regs = ERTS_PROC_GET_SCHDATA(p)->x_reg_array;
 	    }
 	  #endif
-	    cost = erts_garbage_collect(p, 0, regs, p->arity);
+	    cost = garbage_collect(p, live_hf_end, 0, regs, p->arity);
 	} else {
-	    cost = erts_garbage_collect(p, 0, regs, arity);
+	    cost = garbage_collect(p, live_hf_end, 0, regs, arity);
 	}
     } else {
 	Eterm val[1];
 
 	val[0] = result;
-	cost = erts_garbage_collect(p, 0, val, 1);
+	cost = garbage_collect(p, live_hf_end, 0, val, 1);
 	result = val[0];
     }
     BUMP_REDS(p, cost);
     return result;
 }
 
+Eterm
+erts_gc_after_bif_call(Process* p, Eterm result, Eterm* regs, Uint arity)
+{
+    return erts_gc_after_bif_call_lhf(p, ERTS_INVALID_HFRAG_PTR,
+				      result, regs, arity);
+}
+
 static ERTS_INLINE void reset_active_writer(Process *p)
 {
     struct erl_off_heap_header* ptr;
@@ -390,6 +427,139 @@ static ERTS_INLINE void reset_active_writer(Process *p)
     }
 }
 
+#define ERTS_DELAY_GC_EXTRA_FREE 40
+#define ERTS_ABANDON_HEAP_COST 10
+
+static int
+delay_garbage_collection(Process *p, ErlHeapFragment *live_hf_end, int need)
+{
+    ErlHeapFragment *hfrag;
+    Eterm *orig_heap, *orig_hend, *orig_htop, *orig_stop;
+    Eterm *stop, *hend;
+    Uint hsz, ssz;
+    int reds_left;
+
+    ERTS_HOLE_CHECK(p);
+
+    if ((p->flags & F_DISABLE_GC)
+	&& p->live_hf_end == ERTS_INVALID_HFRAG_PTR) {
+	/*
+	 * A BIF yielded with disabled GC. Remember
+	 * heap fragments created by the BIF until we
+	 * do next GC.
+	 */
+	p->live_hf_end = live_hf_end;
+    }
+
+    if (need == 0)
+	return 1;
+
+    /*
+     * Satisfy need in a heap fragment...
+     */
+    ASSERT(need > 0);
+
+    orig_heap = p->heap;
+    orig_hend = p->hend;
+    orig_htop = p->htop;
+    orig_stop = p->stop;
+
+    ssz = orig_hend - orig_stop;
+    hsz = ssz + need + ERTS_DELAY_GC_EXTRA_FREE;
+
+    hfrag = new_message_buffer(hsz);
+    hfrag->next = p->mbuf;
+    p->mbuf = hfrag;
+    p->mbuf_sz += hsz;
+    p->heap = p->htop = &hfrag->mem[0];
+    p->hend = hend = &hfrag->mem[hsz];
+    p->stop = stop = hend - ssz;
+    sys_memcpy((void *) stop, (void *) orig_stop, ssz * sizeof(Eterm));
+
+    if (p->abandoned_heap) {
+	/* Active heap already in a fragment; adjust it... */
+	ErlHeapFragment *hfrag = ((ErlHeapFragment *)
+				  (((char *) orig_heap)
+				   - offsetof(ErlHeapFragment, mem)));
+	Uint unused = orig_hend - orig_htop;
+	ASSERT(hfrag->used_size == hfrag->alloc_size);
+	ASSERT(hfrag->used_size >= unused);
+	hfrag->used_size -= unused;
+	p->mbuf_sz -= unused;
+    }
+    else {
+	/* Do not leave a hole in the abandoned heap... */
+	if (orig_htop < orig_hend) {
+	    *orig_htop = make_pos_bignum_header(orig_hend-orig_htop-1);
+	    if (orig_htop + 1 < orig_hend) {
+		orig_hend[-1] = (Uint) (orig_htop - orig_heap);
+		p->flags |= F_ABANDONED_HEAP_USE;
+	    }
+	}
+	p->abandoned_heap = orig_heap;
+    }
+
+#ifdef CHECK_FOR_HOLES
+    p->last_htop = p->htop;
+    p->heap_hfrag = hfrag;
+#endif
+
+    /* Make sure that we do a proper GC as soon as possible... */
+    p->flags |= F_FORCE_GC;
+    reds_left = ERTS_BIF_REDS_LEFT(p);
+    if (reds_left > ERTS_ABANDON_HEAP_COST) {
+	int vreds = reds_left - ERTS_ABANDON_HEAP_COST;
+	ERTS_VBUMP_REDS(p, vreds);
+    }
+    return ERTS_ABANDON_HEAP_COST;
+}
+
+static ERTS_FORCE_INLINE Uint
+young_gen_usage(Process *p)
+{
+    Uint hsz;
+    Eterm *aheap;
+
+    hsz = p->mbuf_sz;
+
+    if (p->flags & F_ON_HEAP_MSGQ) {
+	ErtsMessage *mp;
+	for (mp = p->msg.first; mp; mp = mp->next)
+	    if (mp->data.attached)
+		hsz += erts_msg_attached_data_size(mp);
+    }
+
+    aheap = p->abandoned_heap;
+    if (!aheap)
+	hsz += p->htop - p->heap;
+    else {
+	/* used in orig heap */
+	if (p->flags & F_ABANDONED_HEAP_USE)
+	    hsz += aheap[p->heap_sz-1];
+	else
+	    hsz += p->heap_sz;
+	/* Remove unused part in latest fragment */
+	hsz -= p->hend - p->htop;
+    }
+    return hsz;
+}
+
+#define ERTS_GET_ORIG_HEAP(Proc, Heap, HTop)			\
+    do {							\
+	Eterm *aheap__ = (Proc)->abandoned_heap;		\
+	if (!aheap__) {						\
+	    (Heap) = (Proc)->heap;				\
+	    (HTop) = (Proc)->htop;				\
+	}							\
+	else {							\
+	    (Heap) = aheap__;					\
+	    if ((Proc)->flags & F_ABANDONED_HEAP_USE)		\
+		(HTop) = aheap__ + aheap__[(Proc)->heap_sz-1];	\
+	    else						\
+		(HTop) = aheap__ + (Proc)->heap_sz;		\
+	}							\
+    } while (0)
+
 /*
  * Garbage collect a process.
  *
@@ -398,21 +568,25 @@ static ERTS_INLINE void reset_active_writer(Process *p)
  * objv: Array of terms to add to rootset; that is to preserve.
  * nobj: Number of objects in objv.
  */
-int
-erts_garbage_collect(Process* p, int need, Eterm* objv, int nobj)
+static int
+garbage_collect(Process* p, ErlHeapFragment *live_hf_end,
+		int need, Eterm* objv, int nobj)
 {
     Uint reclaimed_now = 0;
-    int done = 0;
+    int reds;
     ErtsMonotonicTime start_time = 0; /* Shut up faulty warning... */
     ErtsSchedulerData *esdp;
 #ifdef USE_VM_PROBES
     DTRACE_CHARBUF(pidbuf, DTRACE_TERM_BUF_SIZE);
 #endif
 
-    if (p->flags & F_DISABLE_GC) {
-	ASSERT(need == 0);
-	return 1;
-    }
+    if (p->flags & (F_DISABLE_GC|F_DELAY_GC))
+	return delay_garbage_collection(p, live_hf_end, need);
+
+    if (p->abandoned_heap)
+	live_hf_end = ERTS_INVALID_HFRAG_PTR;
+    else if (p->live_hf_end != ERTS_INVALID_HFRAG_PTR)
+	live_hf_end = p->live_hf_end;
 
     esdp = erts_get_scheduler_data();
 
@@ -420,16 +594,14 @@ erts_garbage_collect(Process* p, int need, Eterm* objv, int nobj)
         trace_gc(p, am_gc_start);
     }
 
-    (void) erts_smp_atomic32_read_bor_nob(&p->state, ERTS_PSFLG_GC);
+    erts_smp_atomic32_read_bor_nob(&p->state, ERTS_PSFLG_GC);
     if (erts_system_monitor_long_gc != 0)
 	start_time = erts_get_monotonic_time(esdp);
 
     ERTS_CHK_OFFHEAP(p);
 
     ErtsGcQuickSanityCheck(p);
-    if (GEN_GCS(p) >= MAX_GEN_GCS(p)) {
-        FLAGS(p) |= F_NEED_FULLSWEEP;
-    }
+
 #ifdef USE_VM_PROBES
     *pidbuf = '\0';
     if (DTRACE_ENABLED(gc_major_start)
@@ -442,17 +614,21 @@ erts_garbage_collect(Process* p, int need, Eterm* objv, int nobj)
     /*
      * Test which type of GC to do.
      */
-    while (!done) {
-	if ((FLAGS(p) & F_NEED_FULLSWEEP) != 0) {
-	    DTRACE2(gc_major_start, pidbuf, need);
-	    done = major_collection(p, need, objv, nobj, &reclaimed_now);
-	    DTRACE2(gc_major_end, pidbuf, reclaimed_now);
-	} else {
-	    DTRACE2(gc_minor_start, pidbuf, need);
-	    done = minor_collection(p, need, objv, nobj, &reclaimed_now);
-	    DTRACE2(gc_minor_end, pidbuf, reclaimed_now);
-	}
+
+    if (GEN_GCS(p) < MAX_GEN_GCS(p) && !(FLAGS(p) & F_NEED_FULLSWEEP)) {
+	DTRACE2(gc_minor_start, pidbuf, need);
+	reds = minor_collection(p, live_hf_end, need, objv, nobj, &reclaimed_now);
+	DTRACE2(gc_minor_end, pidbuf, reclaimed_now);
+	if (reds < 0)
+	    goto do_major_collection;
+    }
+    else {
+    do_major_collection:
+	DTRACE2(gc_major_start, pidbuf, need);
+	reds = major_collection(p, live_hf_end, need, objv, nobj, &reclaimed_now);
+	DTRACE2(gc_major_end, pidbuf, reclaimed_now);
     }
+
     reset_active_writer(p);
 
     /*
@@ -491,6 +667,7 @@ erts_garbage_collect(Process* p, int need, Eterm* objv, int nobj)
     esdp->gc_info.reclaimed += reclaimed_now;
     
     FLAGS(p) &= ~F_FORCE_GC;
+    p->live_hf_end = ERTS_INVALID_HFRAG_PTR;
 
 #ifdef CHECK_FOR_HOLES
     /*
@@ -512,15 +689,20 @@ erts_garbage_collect(Process* p, int need, Eterm* objv, int nobj)
     p->last_old_htop = p->old_htop;
 #endif
 
-    /* FIXME: This function should really return an Sint, i.e., a possibly
-       64 bit wide signed integer, but that requires updating all the code
-       that calls it. For now, we just return INT_MAX if the result is too
-       large for an int. */
-    {
-      Sint result = (HEAP_TOP(p) - HEAP_START(p)) / 10;
-      if (result >= INT_MAX) return INT_MAX;
-      else return (int) result;
-    }
+    return reds;
+}
+
+int
+erts_garbage_collect_nobump(Process* p, int need, Eterm* objv, int nobj)
+{
+    return garbage_collect(p, ERTS_INVALID_HFRAG_PTR, need, objv, nobj);
+}
+
+void
+erts_garbage_collect(Process* p, int need, Eterm* objv, int nobj)
+{
+    int reds = garbage_collect(p, ERTS_INVALID_HFRAG_PTR, need, objv, nobj);
+    BUMP_REDS(p, reds);
 }
 
 /*
@@ -533,13 +715,11 @@ erts_garbage_collect_hibernate(Process* p)
     Uint heap_size;
     Eterm* heap;
     Eterm* htop;
-    Rootset rootset;
-    char* src;
-    Uint src_size;
     Uint actual_size;
     char* area;
     Uint area_size;
     Sint offs;
+    int reds;
 
     if (p->flags & F_DISABLE_GC)
 	ERTS_INTERNAL_ERROR("GC disabled");
@@ -549,55 +729,34 @@ erts_garbage_collect_hibernate(Process* p)
      */
     erts_smp_atomic32_read_bor_nob(&p->state, ERTS_PSFLG_GC);
     ErtsGcQuickSanityCheck(p);
-    ASSERT(p->mbuf_sz == 0);
-    ASSERT(p->mbuf == 0);
+    ASSERT(p->mbuf == NULL);
     ASSERT(p->stop == p->hend);	/* Stack must be empty. */
+    ASSERT(!p->abandoned_heap);
 
     /*
      * Do it.
      */
 
 
-    heap_size = p->heap_sz + (p->old_htop - p->old_heap);
+    heap_size = p->heap_sz + (p->old_htop - p->old_heap) + p->mbuf_sz;
     heap = (Eterm*) ERTS_HEAP_ALLOC(ERTS_ALC_T_TMP_HEAP,
 				    sizeof(Eterm)*heap_size);
     htop = heap;
 
-    (void) setup_rootset(p, p->arg_reg, p->arity, &rootset);
-#if HIPE
-    hipe_empty_nstack(p);
-#endif
-
-    src = (char *) p->heap;
-    src_size = (char *) p->htop - src;
-    htop = sweep_rootset(&rootset, htop, src, src_size);
-    htop = sweep_one_area(heap, htop, src, src_size);
+    htop = full_sweep_heaps(p,
+			    1,
+			    heap,
+			    htop,
+			    (char *) p->old_heap,
+			    (char *) p->old_htop - (char *) p->old_heap,
+			    p->arg_reg,
+			    p->arity);
 
-    if (p->old_heap) {
-	src = (char *) p->old_heap;
-	src_size = (char *) p->old_htop - src;
-	htop = sweep_rootset(&rootset, htop, src, src_size);
-	htop = sweep_one_area(heap, htop, src, src_size);
-    }
-
-    cleanup_rootset(&rootset);
-
-    if (MSO(p).first) {
-        sweep_off_heap(p, 1);
-    }
-
-    /*
-     *  Update all pointers.
-     */
     ERTS_HEAP_FREE(ERTS_ALC_T_HEAP,
-		   (void*)HEAP_START(p),
-		   HEAP_SIZE(p) * sizeof(Eterm));
-    if (p->old_heap) {
-	ERTS_HEAP_FREE(ERTS_ALC_T_OLD_HEAP,
-		       (void*)p->old_heap,
-		       (p->old_hend - p->old_heap) * sizeof(Eterm));
-	p->old_heap = p->old_htop = p->old_hend = 0;
-    }
+		   (p->abandoned_heap
+		    ? p->abandoned_heap
+		    : p->heap),
+		   p->heap_sz * sizeof(Eterm));
 
     p->heap = heap;
     p->high_water = htop;
@@ -612,6 +771,7 @@ erts_garbage_collect_hibernate(Process* p)
     }
 
     FLAGS(p) &= ~F_FORCE_GC;
+    p->live_hf_end = ERTS_INVALID_HFRAG_PTR;
 
     /*
      * Move the heap to its final destination.
@@ -631,6 +791,8 @@ erts_garbage_collect_hibernate(Process* p)
     sys_memcpy((void *) heap, (void *) p->heap, actual_size*sizeof(Eterm));
     ERTS_HEAP_FREE(ERTS_ALC_T_TMP_HEAP, p->heap, p->heap_sz*sizeof(Eterm));
 
+    remove_message_buffers(p);
+
     p->stop = p->hend = heap + heap_size;
 
     offs = heap - p->heap;
@@ -659,15 +821,18 @@ erts_garbage_collect_hibernate(Process* p)
     ErtsGcQuickSanityCheck(p);
 
     erts_smp_atomic32_read_band_nob(&p->state, ~ERTS_PSFLG_GC);
+
+    reds = gc_cost(actual_size, actual_size);
+    BUMP_REDS(p, reds);
 }
 
 
 void
 erts_garbage_collect_literals(Process* p, Eterm* literals,
-			      Uint lit_size,
+			      Uint byte_lit_size,
 			      struct erl_off_heap_header* oh)
 {
-    Uint byte_lit_size = sizeof(Eterm)*lit_size;
+    Uint lit_size = byte_lit_size / sizeof(Eterm);
     Uint old_heap_size;
     Eterm* temp_lit;
     Sint offs;
@@ -743,7 +908,7 @@ erts_garbage_collect_literals(Process* p, Eterm* literals,
                 if (IS_MOVED_BOXED(val)) {
 		    ASSERT(is_boxed(val));
                     *g_ptr++ = val;
-		} else if (in_area(ptr, area, area_size)) {
+		} else if (ErtsInArea(ptr, area, area_size)) {
                     MOVE_BOXED(ptr,val,old_htop,g_ptr++);
 		} else {
 		    g_ptr++;
@@ -754,7 +919,7 @@ erts_garbage_collect_literals(Process* p, Eterm* literals,
                 val = *ptr;
                 if (IS_MOVED_CONS(val)) { /* Moved */
                     *g_ptr++ = ptr[1];
-		} else if (in_area(ptr, area, area_size)) {
+		} else if (ErtsInArea(ptr, area, area_size)) {
                     MOVE_CONS(ptr,val,old_htop,g_ptr++);
                 } else {
 		    g_ptr++;
@@ -774,8 +939,10 @@ erts_garbage_collect_literals(Process* p, Eterm* literals,
      * Now we'll have to go through all heaps updating all other references.
      */
 
-    old_htop = sweep_one_heap(p->heap, p->htop, old_htop, area, area_size);
-    old_htop = sweep_one_area(p->old_heap, old_htop, area, area_size);
+    old_htop = sweep_literals_to_old_heap(p->heap, p->htop, old_htop, area, area_size);
+    old_htop = sweep_literal_area(p->old_heap, old_htop,
+				  (char *) p->old_heap, sizeof(Eterm)*old_heap_size,
+				  area, area_size);
     ASSERT(p->old_htop <= old_htop && old_htop <= p->old_hend);
     p->old_htop = old_htop;
 
@@ -834,15 +1001,18 @@ erts_garbage_collect_literals(Process* p, Eterm* literals,
 }
 
 static int
-minor_collection(Process* p, int need, Eterm* objv, int nobj, Uint *recl)
+minor_collection(Process* p, ErlHeapFragment *live_hf_end,
+		 int need, Eterm* objv, int nobj, Uint *recl)
 {
-    Uint mature = HIGH_WATER(p) - HEAP_START(p);
+    Eterm *mature = p->abandoned_heap ? p->abandoned_heap : p->heap;
+    Uint mature_size = p->high_water - mature;
+    Uint size_before = young_gen_usage(p);
 
     /*
      * Allocate an old heap if we don't have one and if we'll need one.
      */
 
-    if (OLD_HEAP(p) == NULL && mature != 0) {
+    if (OLD_HEAP(p) == NULL && mature_size != 0) {
         Eterm* n_old;
 
         /* Note: We choose a larger heap size than strictly needed,
@@ -850,7 +1020,7 @@ minor_collection(Process* p, int need, Eterm* objv, int nobj, Uint *recl)
          * This improved Estone by more than 1200 estones on my computer
          * (Ultra Sparc 10).
          */
-        Uint new_sz = erts_next_heap_size(HEAP_TOP(p) - HEAP_START(p), 1);
+        Uint new_sz = erts_next_heap_size(size_before, 1);
 
         /* Create new, empty old_heap */
         n_old = (Eterm *) ERTS_HEAP_ALLOC(ERTS_ALC_T_OLD_HEAP,
@@ -866,41 +1036,32 @@ minor_collection(Process* p, int need, Eterm* objv, int nobj, Uint *recl)
      */
 
     if (OLD_HEAP(p) &&
-	    ((mature <= OLD_HEND(p) - OLD_HTOP(p)) &&
-	    ((BIN_VHEAP_MATURE(p) < ( BIN_OLD_VHEAP_SZ(p) - BIN_OLD_VHEAP(p)))) &&
-	    ((BIN_OLD_VHEAP_SZ(p) > BIN_OLD_VHEAP(p))) ) ) {
-	ErlMessage *msgp;
-	Uint size_after;
-	Uint need_after;
-	const Uint stack_size = STACK_SZ_ON_HEAP(p);
-	const Uint size_before = MBUF_SIZE(p) + (HEAP_TOP(p) - HEAP_START(p));
-	Uint new_sz = HEAP_SIZE(p) + MBUF_SIZE(p) + combined_message_size(p);
+	((mature_size <= OLD_HEND(p) - OLD_HTOP(p)) &&
+	 ((BIN_OLD_VHEAP_SZ(p) > BIN_OLD_VHEAP(p))) ) ) {
+	Eterm *prev_old_htop;
+	Uint stack_size, size_after, adjust_size, need_after, new_sz, new_mature;
+
+	stack_size = p->hend - p->stop;
+	new_sz = stack_size + size_before;
         new_sz = next_heap_size(p, new_sz, 0);
 
-        do_minor(p, new_sz, objv, nobj);
+	prev_old_htop = p->old_htop;
+        do_minor(p, live_hf_end, (char *) mature, mature_size*sizeof(Eterm),
+		 new_sz, objv, nobj);
+
+	if (p->flags & F_ON_HEAP_MSGQ)
+	    move_msgq_to_heap(p);
 
-        size_after = HEAP_TOP(p) - HEAP_START(p);
+	new_mature = p->old_htop - prev_old_htop;
+
+	size_after = new_mature;
+        size_after += HEAP_TOP(p) - HEAP_START(p) + p->mbuf_sz;
         *recl += (size_before - size_after);
 
-        /*
-	 * Copy newly received message onto the end of the new heap.
-	 */
-        ErtsGcQuickSanityCheck(p);
-        for (msgp = p->msg.first; msgp; msgp = msgp->next) {
-            if (msgp->data.attached) {
-                ErtsHeapFactory factory;
-                erts_factory_proc_prealloc_init(&factory, p,
-                                                erts_msg_attached_data_size(msgp));
-                erts_move_msg_attached_data_to_heap(&factory, msgp);
-                erts_factory_close(&factory);
-                ErtsGcQuickSanityCheck(p);
-            }
-        }
 	ErtsGcQuickSanityCheck(p);
 
         GEN_GCS(p)++;
         need_after = ((HEAP_TOP(p) - HEAP_START(p))
-                      + erts_used_frag_sz(MBUF(p))
                       + need
                       + stack_size);
 	
@@ -913,6 +1074,8 @@ minor_collection(Process* p, int need, Eterm* objv, int nobj, Uint *recl)
          * the heap size is substantial, we don't want to shrink.
          */
 
+	adjust_size = 0;
+
         if ((HEAP_SIZE(p) > 3000) && (4 * need_after < HEAP_SIZE(p)) &&
             ((HEAP_SIZE(p) > 8000) ||
              (HEAP_SIZE(p) > (OLD_HEND(p) - OLD_HEAP(p))))) {
@@ -934,28 +1097,33 @@ minor_collection(Process* p, int need, Eterm* objv, int nobj, Uint *recl)
 					       : next_heap_size(p, wanted, 0);
             if (wanted < HEAP_SIZE(p)) {
                 shrink_new_heap(p, wanted, objv, nobj);
+		adjust_size = p->htop - p->heap;
             }
 
-            ASSERT(HEAP_SIZE(p) == next_heap_size(p, HEAP_SIZE(p), 0));
-            ASSERT(MBUF(p) == NULL);
-	    return 1;		/* We are done. */
+	    goto done;
         }
 
         if (HEAP_SIZE(p) >= need_after) {
 	    /*
 	     * The heap size turned out to be just right. We are done.
 	     */
-            ASSERT(HEAP_SIZE(p) == next_heap_size(p, HEAP_SIZE(p), 0));
-            ASSERT(MBUF(p) == NULL);
-            return 1;
+	    goto done;
 	}
+
+        grow_new_heap(p, next_heap_size(p, need_after, 0), objv, nobj);
+	adjust_size = p->htop - p->heap;
+
+    done:
+	ASSERT(HEAP_SIZE(p) == next_heap_size(p, HEAP_SIZE(p), 0));
+	ASSERT(MBUF(p) == NULL);
+
+	return gc_cost(size_after, adjust_size);
     }
 
     /*
-     * Still not enough room after minor collection. Must force a major collection.
+     * Not enough room for a minor collection. Must force a major collection.
      */
-    FLAGS(p) |= F_NEED_FULLSWEEP;
-    return 0;
+    return -1;
 }
 
 /*
@@ -1009,7 +1177,9 @@ static ERTS_INLINE void offset_nstack(Process* p, Sint offs,
 #endif /* HIPE */
 
 static void
-do_minor(Process *p, Uint new_sz, Eterm* objv, int nobj)
+do_minor(Process *p, ErlHeapFragment *live_hf_end,
+	 char *mature, Uint mature_size,
+	 Uint new_sz, Eterm* objv, int nobj)
 {
     Rootset rootset;            /* Rootset for GC (stack, dictionary, etc). */
     Roots* roots;
@@ -1018,17 +1188,24 @@ do_minor(Process *p, Uint new_sz, Eterm* objv, int nobj)
     Eterm* ptr;
     Eterm val;
     Eterm gval;
-    char* heap = (char *) HEAP_START(p);
-    Uint heap_size = (char *) HEAP_TOP(p) - heap;
-    Uint mature_size = (char *) HIGH_WATER(p) - heap;
     Eterm* old_htop = OLD_HTOP(p);
     Eterm* n_heap;
+    char* oh = (char *) OLD_HEAP(p);
+    Uint oh_size = (char *) OLD_HTOP(p) - oh;
+
+    VERBOSE(DEBUG_SHCOPY, ("[pid=%T] MINOR GC: %p %p %p %p\n", p->common.id,
+                           HEAP_START(p), HEAP_END(p), OLD_HEAP(p), OLD_HEND(p)));
 
     n_htop = n_heap = (Eterm*) ERTS_HEAP_ALLOC(ERTS_ALC_T_HEAP,
 					       sizeof(Eterm)*new_sz);
 
-    if (MBUF(p) != NULL) {
-	n_htop = collect_heap_frags(p, n_heap, n_htop, objv, nobj);
+    if (live_hf_end != ERTS_INVALID_HFRAG_PTR) {
+	/*
+	 * Move heap frags that we know are completely live
+	 * directly into the new young heap generation.
+	 */
+	n_htop = collect_live_heap_frags(p, live_hf_end, n_heap, n_htop,
+					 objv, nobj);
     }
 
     n = setup_rootset(p, objv, nobj, &rootset);
@@ -1051,9 +1228,9 @@ do_minor(Process *p, Uint new_sz, Eterm* objv, int nobj)
                 if (IS_MOVED_BOXED(val)) {
 		    ASSERT(is_boxed(val));
                     *g_ptr++ = val;
-                } else if (in_area(ptr, heap, mature_size)) {
+                } else if (ErtsInArea(ptr, mature, mature_size)) {
                     MOVE_BOXED(ptr,val,old_htop,g_ptr++);
-                } else if (in_area(ptr, heap, heap_size)) {
+                } else if (ErtsInYoungGen(gval, ptr, oh, oh_size)) {
                     MOVE_BOXED(ptr,val,n_htop,g_ptr++);
                 } else {
 		    g_ptr++;
@@ -1066,9 +1243,9 @@ do_minor(Process *p, Uint new_sz, Eterm* objv, int nobj)
                 val = *ptr;
                 if (IS_MOVED_CONS(val)) { /* Moved */
                     *g_ptr++ = ptr[1];
-                } else if (in_area(ptr, heap, mature_size)) {
+                } else if (ErtsInArea(ptr, mature, mature_size)) {
                     MOVE_CONS(ptr,val,old_htop,g_ptr++);
-                } else if (in_area(ptr, heap, heap_size)) {
+                } else if (ErtsInYoungGen(gval, ptr, oh, oh_size)) {
                     MOVE_CONS(ptr,val,n_htop,g_ptr++);
                 } else {
 		    g_ptr++;
@@ -1093,7 +1270,7 @@ do_minor(Process *p, Uint new_sz, Eterm* objv, int nobj)
      */
 
     if (mature_size == 0) {
-	n_htop = sweep_one_area(n_heap, n_htop, heap, heap_size);
+	n_htop = sweep_new_heap(n_heap, n_htop, oh, oh_size);
     } else {
 	Eterm* n_hp = n_heap;
 	Eterm* ptr;
@@ -1110,9 +1287,9 @@ do_minor(Process *p, Uint new_sz, Eterm* objv, int nobj)
 		if (IS_MOVED_BOXED(val)) {
 		    ASSERT(is_boxed(val));
 		    *n_hp++ = val;
-		} else if (in_area(ptr, heap, mature_size)) {
+		} else if (ErtsInArea(ptr, mature, mature_size)) {
 		    MOVE_BOXED(ptr,val,old_htop,n_hp++);
-		} else if (in_area(ptr, heap, heap_size)) {
+		} else if (ErtsInYoungGen(gval, ptr, oh, oh_size)) {
 		    MOVE_BOXED(ptr,val,n_htop,n_hp++);
 		} else {
 		    n_hp++;
@@ -1124,9 +1301,9 @@ do_minor(Process *p, Uint new_sz, Eterm* objv, int nobj)
 		val = *ptr;
 		if (IS_MOVED_CONS(val)) {
 		    *n_hp++ = ptr[1];
-		} else if (in_area(ptr, heap, mature_size)) {
+		} else if (ErtsInArea(ptr, mature, mature_size)) {
 		    MOVE_CONS(ptr,val,old_htop,n_hp++);
-		} else if (in_area(ptr, heap, heap_size)) {
+		} else if (ErtsInYoungGen(gval, ptr, oh, oh_size)) {
 		    MOVE_CONS(ptr,val,n_htop,n_hp++);
 		} else {
 		    n_hp++;
@@ -1146,10 +1323,10 @@ do_minor(Process *p, Uint new_sz, Eterm* objv, int nobj)
 			if (IS_MOVED_BOXED(val)) {
 			    *origptr = val;
 			    mb->base = binary_bytes(val);
-			} else if (in_area(ptr, heap, mature_size)) {
+			} else if (ErtsInArea(ptr, mature, mature_size)) {
 			    MOVE_BOXED(ptr,val,old_htop,origptr);
 			    mb->base = binary_bytes(mb->orig);
-			} else if (in_area(ptr, heap, heap_size)) {
+			} else if (ErtsInYoungGen(*origptr, ptr, oh, oh_size)) {
 			    MOVE_BOXED(ptr,val,n_htop,origptr);
 			    mb->base = binary_bytes(mb->orig);
 			}
@@ -1170,9 +1347,8 @@ do_minor(Process *p, Uint new_sz, Eterm* objv, int nobj)
      * may point to the old (soon to be deleted) new_heap.
      */
 
-    if (OLD_HTOP(p) < old_htop) {
-	old_htop = sweep_one_area(OLD_HTOP(p), old_htop, heap, heap_size);
-    }
+    if (OLD_HTOP(p) < old_htop)
+	old_htop = sweep_new_heap(OLD_HTOP(p), old_htop, oh, oh_size);
     OLD_HTOP(p) = old_htop;
     HIGH_WATER(p) = n_htop;
 
@@ -1204,8 +1380,12 @@ do_minor(Process *p, Uint new_sz, Eterm* objv, int nobj)
 #endif
 
     ERTS_HEAP_FREE(ERTS_ALC_T_HEAP,
-		   (void*)HEAP_START(p),
+		   (p->abandoned_heap
+		    ? p->abandoned_heap
+		    : HEAP_START(p)),
 		   HEAP_SIZE(p) * sizeof(Eterm));
+    p->abandoned_heap = NULL;
+    p->flags &= ~F_ABANDONED_HEAP_USE;
     HEAP_START(p) = n_heap;
     HEAP_TOP(p) = n_htop;
     HEAP_SIZE(p) = new_sz;
@@ -1222,30 +1402,30 @@ do_minor(Process *p, Uint new_sz, Eterm* objv, int nobj)
  */
 
 static int
-major_collection(Process* p, int need, Eterm* objv, int nobj, Uint *recl)
+major_collection(Process* p, ErlHeapFragment *live_hf_end,
+		 int need, Eterm* objv, int nobj, Uint *recl)
 {
-    Rootset rootset;
-    Roots* roots;
-    const Uint size_before = ((HEAP_TOP(p) - HEAP_START(p))
-                              + (OLD_HTOP(p) - OLD_HEAP(p))
-                              + MBUF_SIZE(p));
+    Uint size_before, size_after, stack_size;
     Eterm* n_heap;
     Eterm* n_htop;
-    char* src = (char *) HEAP_START(p);
-    Uint src_size = (char *) HEAP_TOP(p) - src;
     char* oh = (char *) OLD_HEAP(p);
     Uint oh_size = (char *) OLD_HTOP(p) - oh;
-    Uint n;
-    Uint new_sz;
+    Uint new_sz, stk_sz;
+    int adjusted;
+
+    VERBOSE(DEBUG_SHCOPY, ("[pid=%T] MAJOR GC: %p %p %p %p\n", p->common.id,
+                           HEAP_START(p), HEAP_END(p), OLD_HEAP(p), OLD_HEND(p)));
 
     /*
      * Do a fullsweep GC. First figure out the size of the heap
      * to receive all live data.
      */
 
-    new_sz = (HEAP_SIZE(p) + MBUF_SIZE(p)
-              + combined_message_size(p)
-              + (OLD_HTOP(p) - OLD_HEAP(p)));
+    size_before = young_gen_usage(p);
+    size_before += p->old_htop - p->old_heap;
+    stack_size = p->hend - p->stop;
+
+    new_sz = stack_size + size_before;
     new_sz = next_heap_size(p, new_sz, 0);
 
     /*
@@ -1259,13 +1439,76 @@ major_collection(Process* p, int need, Eterm* objv, int nobj, Uint *recl)
     n_htop = n_heap = (Eterm *) ERTS_HEAP_ALLOC(ERTS_ALC_T_HEAP,
 						sizeof(Eterm)*new_sz);
 
-    /*
-     * Get rid of heap fragments.
-     */
+    if (live_hf_end != ERTS_INVALID_HFRAG_PTR) {
+	/*
+	 * Move heap frags that we know are completely live
+	 * directly into the heap.
+	 */
+	n_htop = collect_live_heap_frags(p, live_hf_end, n_heap, n_htop,
+					 objv, nobj);
+    }
 
-    if (MBUF(p) != NULL) {
-	n_htop = collect_heap_frags(p, n_heap, n_htop, objv, nobj);
+    n_htop = full_sweep_heaps(p, 0, n_heap, n_htop, oh, oh_size, objv, nobj);
+
+    /* Move the stack to the end of the heap */
+    stk_sz = HEAP_END(p) - p->stop;
+    sys_memcpy(n_heap + new_sz - stk_sz, p->stop, stk_sz * sizeof(Eterm));
+    p->stop = n_heap + new_sz - stk_sz;
+
+#ifdef USE_VM_PROBES
+    if (HEAP_SIZE(p) != new_sz && DTRACE_ENABLED(process_heap_grow)) {
+        DTRACE_CHARBUF(pidbuf, DTRACE_TERM_BUF_SIZE);
+
+        dtrace_proc_str(p, pidbuf);
+        DTRACE3(process_heap_grow, pidbuf, HEAP_SIZE(p), new_sz);
     }
+#endif
+
+    ERTS_HEAP_FREE(ERTS_ALC_T_HEAP,
+		   (p->abandoned_heap
+		    ? p->abandoned_heap
+		    : HEAP_START(p)),
+		   p->heap_sz * sizeof(Eterm));
+    p->abandoned_heap = NULL;
+    p->flags &= ~F_ABANDONED_HEAP_USE;
+    HEAP_START(p) = n_heap;
+    HEAP_TOP(p) = n_htop;
+    HEAP_SIZE(p) = new_sz;
+    HEAP_END(p) = n_heap + new_sz;
+    GEN_GCS(p) = 0;
+
+    HIGH_WATER(p) = HEAP_TOP(p);
+
+    remove_message_buffers(p);
+
+    if (p->flags & F_ON_HEAP_MSGQ)
+	move_msgq_to_heap(p);
+
+    ErtsGcQuickSanityCheck(p);
+
+    size_after = HEAP_TOP(p) - HEAP_START(p) + p->mbuf_sz;
+    *recl += size_before - size_after;
+
+    adjusted = adjust_after_fullsweep(p, need, objv, nobj);
+
+#ifdef HARDDEBUG
+    disallow_heap_frag_ref_in_heap(p);
+#endif
+    ErtsGcQuickSanityCheck(p);
+
+    return gc_cost(size_after, adjusted ? size_after : 0);
+}
+
+static Eterm *
+full_sweep_heaps(Process *p,
+		 int hibernate,
+		 Eterm *n_heap, Eterm* n_htop,
+		 char *oh, Uint oh_size,
+		 Eterm *objv, int nobj)
+{
+    Rootset rootset;
+    Roots *roots;
+    Uint n;
 
     /*
      * Copy all top-level terms directly referenced by the rootset to
@@ -1273,7 +1516,14 @@ major_collection(Process* p, int need, Eterm* objv, int nobj, Uint *recl)
      */
 
     n = setup_rootset(p, objv, nobj, &rootset);
-    n_htop = fullsweep_nstack(p, n_htop);
+
+#ifdef HIPE
+    if (hibernate)
+	hipe_empty_nstack(p);
+    else
+	n_htop = fullsweep_nstack(p, n_htop);
+#endif
+
     roots = rootset.roots;
     while (n--) {
 	Eterm* g_ptr = roots->v;
@@ -1293,7 +1543,7 @@ major_collection(Process* p, int need, Eterm* objv, int nobj, Uint *recl)
 		if (IS_MOVED_BOXED(val)) {
 		    ASSERT(is_boxed(val));
 		    *g_ptr++ = val;
-		} else if (in_area(ptr, src, src_size) || in_area(ptr, oh, oh_size)) {
+		} else if (!erts_is_literal(gval, ptr)) {
 		    MOVE_BOXED(ptr,val,n_htop,g_ptr++);
 		} else {
 		    g_ptr++;
@@ -1306,7 +1556,7 @@ major_collection(Process* p, int need, Eterm* objv, int nobj, Uint *recl)
 		val = *ptr;
 		if (IS_MOVED_CONS(val)) {
 		    *g_ptr++ = ptr[1];
-		} else if (in_area(ptr, src, src_size) || in_area(ptr, oh, oh_size)) {
+		} else if (!erts_is_literal(gval, ptr)) {
 		    MOVE_CONS(ptr,val,n_htop,g_ptr++);
 		} else {
 		    g_ptr++;
@@ -1331,74 +1581,7 @@ major_collection(Process* p, int need, Eterm* objv, int nobj, Uint *recl)
      * until all is copied.
      */
 
-    if (oh_size == 0) {
-	n_htop = sweep_one_area(n_heap, n_htop, src, src_size);
-    } else {
-	Eterm* n_hp = n_heap;
-
-	while (n_hp != n_htop) {
-	    Eterm* ptr;
-	    Eterm val;
-	    Eterm gval = *n_hp;
-
-	    switch (primary_tag(gval)) {
-	    case TAG_PRIMARY_BOXED: {
-		ptr = boxed_val(gval);
-		val = *ptr;
-		if (IS_MOVED_BOXED(val)) {
-		    ASSERT(is_boxed(val));
-		    *n_hp++ = val;
-		} else if (in_area(ptr, src, src_size) || in_area(ptr, oh, oh_size)) {
-		    MOVE_BOXED(ptr,val,n_htop,n_hp++);
-		} else {
-		    n_hp++;
-		}
-		break;
-	    }
-	    case TAG_PRIMARY_LIST: {
-		ptr = list_val(gval);
-		val = *ptr;
-		if (IS_MOVED_CONS(val)) {
-		    *n_hp++ = ptr[1];
-		} else if (in_area(ptr, src, src_size) || in_area(ptr, oh, oh_size)) {
-		    MOVE_CONS(ptr,val,n_htop,n_hp++);
-		} else {
-		    n_hp++;
-		}
-		break;
-	    }
-	    case TAG_PRIMARY_HEADER: {
-		if (!header_is_thing(gval))
-		    n_hp++;
-		else {
-		    if (header_is_bin_matchstate(gval)) {
-			ErlBinMatchState *ms = (ErlBinMatchState*) n_hp;
-			ErlBinMatchBuffer *mb = &(ms->mb);
-			Eterm* origptr;	
-			origptr = &(mb->orig);
-			ptr = boxed_val(*origptr);
-			val = *ptr;
-			if (IS_MOVED_BOXED(val)) {
-			    *origptr = val;
-			    mb->base = binary_bytes(*origptr);
-			} else if (in_area(ptr, src, src_size) ||
-				   in_area(ptr, oh, oh_size)) {
-			    MOVE_BOXED(ptr,val,n_htop,origptr); 
-			    mb->base = binary_bytes(*origptr);
-			    ptr = boxed_val(*origptr);
-			    val = *ptr;
-			}
-		    }
-		    n_hp += (thing_arityval(gval)+1);
-		}
-		break;
-	    }
-	    default:
-		n_hp++;
-		break;
-	    }
-	}
-    }
+    n_htop = sweep_heaps(n_heap, n_htop, oh, oh_size);
 
     if (MSO(p).first) {
 	sweep_off_heap(p, 1);
@@ -1411,67 +1594,13 @@ major_collection(Process* p, int need, Eterm* objv, int nobj, Uint *recl)
 	OLD_HEAP(p) = OLD_HTOP(p) = OLD_HEND(p) = NULL;
     }
 
-    /* Move the stack to the end of the heap */
-    n = HEAP_END(p) - p->stop;
-    sys_memcpy(n_heap + new_sz - n, p->stop, n * sizeof(Eterm));
-    p->stop = n_heap + new_sz - n;
-
-#ifdef USE_VM_PROBES
-    if (HEAP_SIZE(p) != new_sz && DTRACE_ENABLED(process_heap_grow)) {
-        DTRACE_CHARBUF(pidbuf, DTRACE_TERM_BUF_SIZE);
-
-        dtrace_proc_str(p, pidbuf);
-        DTRACE3(process_heap_grow, pidbuf, HEAP_SIZE(p), new_sz);
-    }
-#endif
-
-    ERTS_HEAP_FREE(ERTS_ALC_T_HEAP,
-		   (void *) HEAP_START(p),
-		   (HEAP_END(p) - HEAP_START(p)) * sizeof(Eterm));
-    HEAP_START(p) = n_heap;
-    HEAP_TOP(p) = n_htop;
-    HEAP_SIZE(p) = new_sz;
-    HEAP_END(p) = n_heap + new_sz;
-    GEN_GCS(p) = 0;
-
-    HIGH_WATER(p) = HEAP_TOP(p);
-
-    ErtsGcQuickSanityCheck(p);
-
-    *recl += size_before - (HEAP_TOP(p) - HEAP_START(p));
-
-    {
-	ErlMessage *msgp;
-
-	/*
-	 * Copy newly received message onto the end of the new heap.
-	 */
-        for (msgp = p->msg.first; msgp; msgp = msgp->next) {
-	    if (msgp->data.attached) {
-                ErtsHeapFactory factory;
-                erts_factory_proc_prealloc_init(&factory, p,
-                                                erts_msg_attached_data_size(msgp));
-		erts_move_msg_attached_data_to_heap(&factory, msgp);
-                erts_factory_close(&factory);
-		ErtsGcQuickSanityCheck(p);
-	    }
-	}
-    }
-
-    adjust_after_fullsweep(p, need, objv, nobj);
-
-#ifdef HARDDEBUG
-    disallow_heap_frag_ref_in_heap(p);
-#endif
-    remove_message_buffers(p);
-
-    ErtsGcQuickSanityCheck(p);
-    return 1;			/* We are done. */
+    return n_htop;
 }
 
-static void
+static int
 adjust_after_fullsweep(Process *p, int need, Eterm *objv, int nobj)
 {
+    int adjusted = 0;
     Uint wanted, sz, need_after;
     Uint stack_size = STACK_SZ_ON_HEAP(p);
     
@@ -1484,6 +1613,7 @@ adjust_after_fullsweep(Process *p, int need, Eterm *objv, int nobj)
         /* Too small - grow to match requested need */
         sz = next_heap_size(p, need_after, 0);
         grow_new_heap(p, sz, objv, nobj);
+	adjusted = 1;
     } else if (3 * HEAP_SIZE(p) < 4 * need_after){
         /* Need more than 75% of current, postpone to next GC.*/
         FLAGS(p) |= F_HEAP_GROW;
@@ -1500,25 +1630,10 @@ adjust_after_fullsweep(Process *p, int need, Eterm *objv, int nobj)
 
         if (sz < HEAP_SIZE(p)) {
             shrink_new_heap(p, sz, objv, nobj);
+	    adjusted = 1;
         }
     }
-}
-
-/*
- * Return the size of all message buffers that are NOT linked in the
- * mbuf list.
- */
-static Uint
-combined_message_size(Process* p)
-{
-    Uint sz;
-    ErlMessage *msgp;
-
-    for (sz = 0, msgp = p->msg.first; msgp; msgp = msgp->next) {
-	if (msgp->data.attached)
-	    sz += erts_msg_attached_data_size(msgp);
-    }
-    return sz;
+    return adjusted;
 }
 
 /*
@@ -1531,6 +1646,10 @@ remove_message_buffers(Process* p)
 	free_message_buffer(MBUF(p));
 	MBUF(p) = NULL;
     }    
+    if (p->msg_frag) {
+	erts_cleanup_messages(p->msg_frag);
+	p->msg_frag = NULL;
+    }
     MBUF_SIZE(p) = 0;    
 }
 #ifdef HARDDEBUG
@@ -1542,64 +1661,6 @@ remove_message_buffers(Process* p)
  * For performance reasons, we use _unchecked_list_val(), _unchecked_boxed_val(),
  * and so on to avoid a function call.
  */
- 
-static void
-disallow_heap_frag_ref(Process* p, Eterm* n_htop, Eterm* objv, int nobj)
-{
-    ErlHeapFragment* mbuf;
-    ErlHeapFragment* qb;
-    Eterm gval;
-    Eterm* ptr;
-    Eterm val;
-
-    ASSERT(p->htop != NULL);
-    mbuf = MBUF(p);
-
-    while (nobj--) {
-	gval = *objv;
-	
-	switch (primary_tag(gval)) {
-
-	case TAG_PRIMARY_BOXED: {
-	    ptr = _unchecked_boxed_val(gval);
-	    val = *ptr;
-	    if (IS_MOVED_BOXED(val)) {
-		ASSERT(is_boxed(val));
-		objv++;
-	    } else {
- 		for (qb = mbuf; qb != NULL; qb = qb->next) {
-		    if (in_area(ptr, qb->mem, qb->alloc_size*sizeof(Eterm))) {
-			abort();
-		    }
-		}
-		objv++;
-	    }
-	    break;
-	}
-
-	case TAG_PRIMARY_LIST: {
-	    ptr = _unchecked_list_val(gval);
-	    val = *ptr;
-	    if (IS_MOVED_CONS(val)) {
-		objv++;
-	    } else {
-		for (qb = mbuf; qb != NULL; qb = qb->next) {
-		    if (in_area(ptr, qb->mem, qb->alloc_size*sizeof(Eterm))) {
-			abort();
-		    }
-		}
-		objv++;
-	    }
-	    break;
-	}
-
-	default: {
-	    objv++;
-	    break;
-	}
-	}
-    }
-}
 
 static void
 disallow_heap_frag_ref_in_heap(Process* p)
@@ -1627,9 +1688,9 @@ disallow_heap_frag_ref_in_heap(Process* p)
 	switch (primary_tag(val)) {
 	case TAG_PRIMARY_BOXED:
 	    ptr = _unchecked_boxed_val(val);
-	    if (!in_area(ptr, heap, heap_size)) {
+	    if (!ErtsInArea(ptr, heap, heap_size)) {
 		for (qb = MBUF(p); qb != NULL; qb = qb->next) {
-		    if (in_area(ptr, qb->mem, qb->alloc_size*sizeof(Eterm))) {
+		    if (ErtsInArea(ptr, qb->mem, qb->alloc_size*sizeof(Eterm))) {
 			abort();
 		    }
 		}
@@ -1637,9 +1698,9 @@ disallow_heap_frag_ref_in_heap(Process* p)
 	    break;
 	case TAG_PRIMARY_LIST:
 	    ptr = _unchecked_list_val(val);
-	    if (!in_area(ptr, heap, heap_size)) {
+	    if (!ErtsInArea(ptr, heap, heap_size)) {
 		for (qb = MBUF(p); qb != NULL; qb = qb->next) {
-		    if (in_area(ptr, qb->mem, qb->alloc_size*sizeof(Eterm))) {
+		    if (ErtsInArea(ptr, qb->mem, qb->alloc_size*sizeof(Eterm))) {
 			abort();
 		    }
 		}
@@ -1681,26 +1742,26 @@ disallow_heap_frag_ref_in_old_heap(Process* p)
 	val = *hp++;
 	switch (primary_tag(val)) {
 	case TAG_PRIMARY_BOXED:
-	    ptr = (Eterm *) EXPAND_POINTER(val);
-	    if (!in_area(ptr, old_heap, old_heap_size)) {
-		if (in_area(ptr, new_heap, new_heap_size)) {
+	    ptr = (Eterm *) val;
+	    if (!ErtsInArea(ptr, old_heap, old_heap_size)) {
+		if (ErtsInArea(ptr, new_heap, new_heap_size)) {
 		    abort();
 		}
 		for (qb = MBUF(p); qb != NULL; qb = qb->next) {
-		    if (in_area(ptr, qb->mem, qb->alloc_size*sizeof(Eterm))) {
+		    if (ErtsInArea(ptr, qb->mem, qb->alloc_size*sizeof(Eterm))) {
 			abort();
 		    }
 		}
 	    }
 	    break;
 	case TAG_PRIMARY_LIST:
-	    ptr = (Eterm *) EXPAND_POINTER(val);
-	    if (!in_area(ptr, old_heap, old_heap_size)) {
-		if (in_area(ptr, new_heap, new_heap_size)) {
+	    ptr = (Eterm *) val;
+	    if (!ErtsInArea(ptr, old_heap, old_heap_size)) {
+		if (ErtsInArea(ptr, new_heap, new_heap_size)) {
 		    abort();
 		}
 		for (qb = MBUF(p); qb != NULL; qb = qb->next) {
-		    if (in_area(ptr, qb->mem, qb->alloc_size*sizeof(Eterm))) {
+		    if (ErtsInArea(ptr, qb->mem, qb->alloc_size*sizeof(Eterm))) {
 			abort();
 		    }
 		}
@@ -1709,7 +1770,7 @@ disallow_heap_frag_ref_in_old_heap(Process* p)
 	case TAG_PRIMARY_HEADER:
 	    if (header_is_thing(val)) {
 		hp += _unchecked_thing_arityval(val);
-		if (!in_area(hp, old_heap, old_heap_size+1)) {
+		if (!ErtsInArea(hp, old_heap, old_heap_size+1)) {
 		    abort();
 		}
 	    }
@@ -1719,66 +1780,30 @@ disallow_heap_frag_ref_in_old_heap(Process* p)
 }
 #endif
 
-static Eterm*
-sweep_rootset(Rootset* rootset, Eterm* htop, char* src, Uint src_size)
+typedef enum {
+    ErtsSweepNewHeap,
+    ErtsSweepHeaps,
+    ErtsSweepLiteralArea
+} ErtsSweepType;
+
+static ERTS_FORCE_INLINE Eterm *
+sweep(Eterm *n_hp, Eterm *n_htop,
+      ErtsSweepType type,
+      char *oh, Uint ohsz,
+      char *src, Uint src_size)
 {
-    Roots* roots = rootset->roots;
-    Uint n = rootset->num_roots;
     Eterm* ptr;
-    Eterm gval;
     Eterm val;
+    Eterm gval;
 
-    while (n--) {
-        Eterm* g_ptr = roots->v;
-        Uint g_sz = roots->sz;
-
-	roots++;
-        while (g_sz--) {
-            gval = *g_ptr;
-
-            switch (primary_tag(gval)) {
-	    case TAG_PRIMARY_BOXED: {
-		ptr = boxed_val(gval);
-                val = *ptr;
-                if (IS_MOVED_BOXED(val)) {
-		    ASSERT(is_boxed(val));
-                    *g_ptr++ = val;
-                } else if (in_area(ptr, src, src_size)) {
-                    MOVE_BOXED(ptr,val,htop,g_ptr++);
-                } else {
-		    g_ptr++;
-		}
-                break;
-	    }
-	    case TAG_PRIMARY_LIST: {
-                ptr = list_val(gval);
-                val = *ptr;
-                if (IS_MOVED_CONS(val)) {
-                    *g_ptr++ = ptr[1];
-                } else if (in_area(ptr, src, src_size)) {
-                    MOVE_CONS(ptr,val,htop,g_ptr++);
-                } else {
-		    g_ptr++;
-		}
-		break;
-	    }
-
-	    default:
-                g_ptr++;
-		break;
-            }
-        }
-    }
-    return htop;
-}
-
+#undef ERTS_IS_IN_SWEEP_AREA
 
-static Eterm*
-sweep_one_area(Eterm* n_hp, Eterm* n_htop, char* src, Uint src_size)
-{
-    Eterm* ptr;
-    Eterm val;
-    Eterm gval;
+#define ERTS_IS_IN_SWEEP_AREA(TPtr, Ptr)				\
+    (type == ErtsSweepHeaps						\
+     ? !erts_is_literal((TPtr), (Ptr))					\
+     : (type == ErtsSweepNewHeap					\
+	? ErtsInYoungGen((TPtr), (Ptr), oh, ohsz)			\
+	: ErtsInArea((Ptr), src, src_size)))
 
     while (n_hp != n_htop) {
 	ASSERT(n_hp < n_htop);
@@ -1790,7 +1815,7 @@ sweep_one_area(Eterm* n_hp, Eterm* n_htop, char* src, Uint src_size)
 	    if (IS_MOVED_BOXED(val)) {
 		ASSERT(is_boxed(val));
 		*n_hp++ = val;
-	    } else if (in_area(ptr, src, src_size)) {
+	    } else if (ERTS_IS_IN_SWEEP_AREA(gval, ptr)) {
 		MOVE_BOXED(ptr,val,n_htop,n_hp++);
 	    } else {
 		n_hp++;
@@ -1802,7 +1827,7 @@ sweep_one_area(Eterm* n_hp, Eterm* n_htop, char* src, Uint src_size)
 	    val = *ptr;
 	    if (IS_MOVED_CONS(val)) {
 		*n_hp++ = ptr[1];
-	    } else if (in_area(ptr, src, src_size)) {
+	    } else if (ERTS_IS_IN_SWEEP_AREA(gval, ptr)) {
 		MOVE_CONS(ptr,val,n_htop,n_hp++);
 	    } else {
 		n_hp++;
@@ -1823,7 +1848,7 @@ sweep_one_area(Eterm* n_hp, Eterm* n_htop, char* src, Uint src_size)
 		    if (IS_MOVED_BOXED(val)) {
 			*origptr = val;
 			mb->base = binary_bytes(*origptr);
-		    } else if (in_area(ptr, src, src_size)) {
+		    } else if (ERTS_IS_IN_SWEEP_AREA(*origptr, ptr)) {
 			MOVE_BOXED(ptr,val,n_htop,origptr); 
 			mb->base = binary_bytes(*origptr);
 		    }
@@ -1838,10 +1863,41 @@ sweep_one_area(Eterm* n_hp, Eterm* n_htop, char* src, Uint src_size)
 	}
     }
     return n_htop;
+#undef ERTS_IS_IN_SWEEP_AREA
+}
+
+static Eterm *
+sweep_new_heap(Eterm *n_hp, Eterm *n_htop, char* old_heap, Uint old_heap_size)
+{
+    return sweep(n_hp, n_htop,
+		 ErtsSweepNewHeap,
+		 old_heap, old_heap_size,
+		 NULL, 0);
+}
+
+static Eterm *
+sweep_heaps(Eterm *n_hp, Eterm *n_htop, char* old_heap, Uint old_heap_size)
+{
+    return sweep(n_hp, n_htop,
+		 ErtsSweepHeaps,
+		 old_heap, old_heap_size,
+		 NULL, 0);
+}
+
+static Eterm *
+sweep_literal_area(Eterm *n_hp, Eterm *n_htop,
+		   char* old_heap, Uint old_heap_size,
+		   char* src, Uint src_size)
+{
+    return sweep(n_hp, n_htop,
+		 ErtsSweepLiteralArea,
+		 old_heap, old_heap_size,
+		 src, src_size);
 }
 
 static Eterm*
-sweep_one_heap(Eterm* heap_ptr, Eterm* heap_end, Eterm* htop, char* src, Uint src_size)
+sweep_literals_to_old_heap(Eterm* heap_ptr, Eterm* heap_end, Eterm* htop,
+			   char* src, Uint src_size)
 {
     while (heap_ptr < heap_end) {
 	Eterm* ptr;
@@ -1855,7 +1911,7 @@ sweep_one_heap(Eterm* heap_ptr, Eterm* heap_end, Eterm* htop, char* src, Uint sr
 	    if (IS_MOVED_BOXED(val)) {
 		ASSERT(is_boxed(val));
 		*heap_ptr++ = val;
-	    } else if (in_area(ptr, src, src_size)) {
+	    } else if (ErtsInArea(ptr, src, src_size)) {
 		MOVE_BOXED(ptr,val,htop,heap_ptr++);
 	    } else {
 		heap_ptr++;
@@ -1867,7 +1923,7 @@ sweep_one_heap(Eterm* heap_ptr, Eterm* heap_end, Eterm* htop, char* src, Uint sr
 	    val = *ptr;
 	    if (IS_MOVED_CONS(val)) {
 		*heap_ptr++ = ptr[1];
-	    } else if (in_area(ptr, src, src_size)) {
+	    } else if (ErtsInArea(ptr, src, src_size)) {
 		MOVE_CONS(ptr,val,htop,heap_ptr++);
 	    } else {
 		heap_ptr++;
@@ -1888,7 +1944,7 @@ sweep_one_heap(Eterm* heap_ptr, Eterm* heap_end, Eterm* htop, char* src, Uint sr
 		    if (IS_MOVED_BOXED(val)) {
 			*origptr = val;
 			mb->base = binary_bytes(*origptr);
-		    } else if (in_area(ptr, src, src_size)) {
+		    } else if (ErtsInArea(ptr, src, src_size)) {
 			MOVE_BOXED(ptr,val,htop,origptr);
 			mb->base = binary_bytes(*origptr);
 		    }
@@ -1939,32 +1995,21 @@ move_one_area(Eterm* n_htop, char* src, Uint src_size)
  */
 
 static Eterm*
-collect_heap_frags(Process* p, Eterm* n_hstart, Eterm* n_htop,
-		   Eterm* objv, int nobj)
+collect_live_heap_frags(Process* p, ErlHeapFragment *live_hf_end,
+			Eterm* n_hstart, Eterm* n_htop,
+			Eterm* objv, int nobj)
 {
     ErlHeapFragment* qb;
     char* frag_begin;
     Uint frag_size;
 
     /*
-     * We don't allow references to a heap fragments from the stack, heap,
-     * or process dictionary.
-     */
-#ifdef HARDDEBUG
-    disallow_heap_frag_ref(p, n_htop, p->stop, STACK_START(p) - p->stop);
-    if (p->dictionary != NULL) {
-	disallow_heap_frag_ref(p, n_htop, p->dictionary->data, p->dictionary->used);
-    }
-    disallow_heap_frag_ref_in_heap(p);
-#endif
-
-    /*
      * Move the heap fragments to the new heap. Note that no GC is done on
      * the heap fragments. Any garbage will thus be moved as well and survive
      * until next GC.  
      */ 
     qb = MBUF(p);
-    while (qb != NULL) {
+    while (qb != live_hf_end) {
         ASSERT(!qb->off_heap.first);  /* process fragments use the MSO(p) list */
 	frag_size = qb->used_size * sizeof(Eterm);
 	if (frag_size != 0) {
@@ -1976,12 +2021,177 @@ collect_heap_frags(Process* p, Eterm* n_hstart, Eterm* n_htop,
     return n_htop;
 }
 
+static ERTS_INLINE void
+copy_one_frag(Eterm** hpp, ErlOffHeap* off_heap,
+	      ErlHeapFragment *bp, Eterm *refs, int nrefs)
+{
+    Uint sz;
+    int i;
+    Sint offs;
+    struct erl_off_heap_header* oh;
+    Eterm *fhp, *hp;
+
+    OH_OVERHEAD(off_heap, bp->off_heap.overhead);
+    sz = bp->used_size;
+
+    fhp = bp->mem;
+    hp = *hpp;
+    offs = hp - fhp;
+
+    oh = NULL;
+    while (sz--) {
+	Uint cpy_sz;
+	Eterm val = *fhp++;
+
+	switch (primary_tag(val)) {
+	case TAG_PRIMARY_IMMED1:
+	    *hp++ = val;
+	    break;
+	case TAG_PRIMARY_LIST:
+	case TAG_PRIMARY_BOXED:
+	    *hp++ = offset_ptr(val, offs);
+	    break;
+	case TAG_PRIMARY_HEADER:
+	    *hp++ = val;
+	    switch (val & _HEADER_SUBTAG_MASK) {
+	    case ARITYVAL_SUBTAG:
+		break;
+	    case REFC_BINARY_SUBTAG:
+	    case FUN_SUBTAG:
+	    case EXTERNAL_PID_SUBTAG:
+	    case EXTERNAL_PORT_SUBTAG:
+	    case EXTERNAL_REF_SUBTAG:
+		oh = (struct erl_off_heap_header*) (hp-1);
+		cpy_sz = thing_arityval(val);
+		goto cpy_words;
+	    default:
+		cpy_sz = header_arity(val);
+
+	    cpy_words:
+		ASSERT(sz >= cpy_sz);
+		sz -= cpy_sz;
+		while (cpy_sz >= 8) {
+		    cpy_sz -= 8;
+		    *hp++ = *fhp++;
+		    *hp++ = *fhp++;
+		    *hp++ = *fhp++;
+		    *hp++ = *fhp++;
+		    *hp++ = *fhp++;
+		    *hp++ = *fhp++;
+		    *hp++ = *fhp++;
+		    *hp++ = *fhp++;
+		}
+		switch (cpy_sz) {
+		case 7: *hp++ = *fhp++;
+		case 6: *hp++ = *fhp++;
+		case 5: *hp++ = *fhp++;
+		case 4: *hp++ = *fhp++;
+		case 3: *hp++ = *fhp++;
+		case 2: *hp++ = *fhp++;
+		case 1: *hp++ = *fhp++;
+		default: break;
+		}
+		if (oh) {
+		    /* Add to offheap list */
+		    oh->next = off_heap->first;
+		    off_heap->first = oh;
+		    ASSERT(*hpp <= (Eterm*)oh);
+		    ASSERT(hp > (Eterm*)oh);
+		    oh = NULL;
+		}
+		break;
+	    }
+	    break;
+	}
+    }
+
+    ASSERT(bp->used_size == hp - *hpp);
+    *hpp = hp;
+
+    for (i = 0; i < nrefs; i++) {
+	if (is_not_immed(refs[i]))
+	    refs[i] = offset_ptr(refs[i], offs);
+    }
+    bp->off_heap.first = NULL;
+}
+
+static void
+move_msgq_to_heap(Process *p)
+{
+    ErtsMessage **mpp = &p->msg.first;
+
+    while (*mpp) {
+	ErtsMessage *mp = *mpp;
+
+	if (mp->data.attached) {
+	    ErlHeapFragment *bp;
+	    ErtsHeapFactory factory;
+
+	    erts_factory_proc_prealloc_init(&factory, p,
+					    erts_msg_attached_data_size(mp));
+
+	    if (is_non_value(ERL_MESSAGE_TERM(mp))) {
+		if (mp->data.dist_ext) {
+		    ASSERT(mp->data.dist_ext->heap_size >= 0);
+		    if (is_not_nil(ERL_MESSAGE_TOKEN(mp))) {
+			bp = erts_dist_ext_trailer(mp->data.dist_ext);
+			ERL_MESSAGE_TOKEN(mp) = copy_struct(ERL_MESSAGE_TOKEN(mp),
+							    bp->used_size,
+							    &factory.hp,
+							    factory.off_heap);
+			erts_cleanup_offheap(&bp->off_heap);
+		    }
+		    ERL_MESSAGE_TERM(mp) = erts_decode_dist_ext(&factory,
+								mp->data.dist_ext);
+		    erts_free_dist_ext_copy(mp->data.dist_ext);
+		    mp->data.dist_ext = NULL;
+		}
+	    }
+	    else {
+
+		if (mp->data.attached == ERTS_MSG_COMBINED_HFRAG)
+		    bp = &mp->hfrag;
+		else
+		    bp = mp->data.heap_frag;
+
+		if (bp->next)
+		    erts_move_multi_frags(&factory.hp, factory.off_heap, bp,
+					  mp->m, ERL_MESSAGE_REF_ARRAY_SZ, 0);
+		else
+		    copy_one_frag(&factory.hp, factory.off_heap, bp,
+				  mp->m, ERL_MESSAGE_REF_ARRAY_SZ);
+
+		if (mp->data.attached != ERTS_MSG_COMBINED_HFRAG) {
+		    mp->data.heap_frag = NULL;
+		    free_message_buffer(bp);
+		}
+		else {
+		    ErtsMessage *tmp = erts_alloc_message(0, NULL);
+		    sys_memcpy((void *) tmp->m, (void *) mp->m,
+			       sizeof(Eterm)*ERL_MESSAGE_REF_ARRAY_SZ); 
+		    tmp->next = mp->next;
+		    if (p->msg.save == &mp->next)
+			p->msg.save = &tmp->next;
+		    if (p->msg.last == &mp->next)
+			p->msg.last = &tmp->next;
+		    *mpp = tmp;
+		    mp->next = NULL;
+		    erts_cleanup_messages(mp);
+		    mp = tmp;
+		}
+	    }
+
+	    erts_factory_close(&factory);
+	}
+
+	mpp = &(*mpp)->next;
+    }
+}
+
 static Uint
 setup_rootset(Process *p, Eterm *objv, int nobj, Rootset *rootset)
 {
-    Uint avail;
     Roots* roots;
-    ErlMessage* mp;
     Uint n;
 
     n = 0;
@@ -1993,8 +2203,8 @@ setup_rootset(Process *p, Eterm *objv, int nobj, Rootset *rootset)
     ++n;
 
     if (p->dictionary != NULL) {
-        roots[n].v = p->dictionary->data;
-        roots[n].sz = p->dictionary->used;
+        roots[n].v = ERTS_PD_START(p->dictionary);
+        roots[n].sz = ERTS_PD_SIZE(p->dictionary);
         ++n;
     }
     if (nobj > 0) {
@@ -2004,7 +2214,7 @@ setup_rootset(Process *p, Eterm *objv, int nobj, Rootset *rootset)
     }
 
     ASSERT((is_nil(p->seq_trace_token) ||
-	    is_tuple(follow_moved(p->seq_trace_token)) ||
+	    is_tuple(follow_moved(p->seq_trace_token, (Eterm) 0)) ||
 	    is_atom(p->seq_trace_token)));
     if (is_not_immed(p->seq_trace_token)) {
 	roots[n].v = &p->seq_trace_token;
@@ -2022,7 +2232,7 @@ setup_rootset(Process *p, Eterm *objv, int nobj, Rootset *rootset)
 	   is_internal_pid(ERTS_TRACER_PROC(p)) ||
 	   is_internal_port(ERTS_TRACER_PROC(p)));
 
-    ASSERT(is_pid(follow_moved(p->group_leader)));
+    ASSERT(is_pid(follow_moved(p->group_leader, (Eterm) 0)));
     if (is_not_immed(p->group_leader)) {
 	roots[n].v  = &p->group_leader;
 	roots[n].sz = 1;
@@ -2059,31 +2269,47 @@ setup_rootset(Process *p, Eterm *objv, int nobj, Rootset *rootset)
 
     ASSERT(n <= rootset->size);
 
-    mp = p->msg.first;
-    avail = rootset->size - n;
-    while (mp != NULL) {
-	if (avail == 0) {
-	    Uint new_size = 2*rootset->size;
-	    if (roots == rootset->def) {
-		roots = erts_alloc(ERTS_ALC_T_ROOTSET,
-				   new_size*sizeof(Roots));
-		sys_memcpy(roots, rootset->def, sizeof(rootset->def));
-	    } else {
-		roots = erts_realloc(ERTS_ALC_T_ROOTSET,
-				     (void *) roots,
-				     new_size*sizeof(Roots));
-	    }
+    switch (p->flags & (F_OFF_HEAP_MSGQ|F_OFF_HEAP_MSGQ_CHNG)) {
+    case F_OFF_HEAP_MSGQ|F_OFF_HEAP_MSGQ_CHNG:
+	(void) erts_move_messages_off_heap(p);
+    case F_OFF_HEAP_MSGQ:
+	break;
+    case F_OFF_HEAP_MSGQ_CHNG:
+    case 0: {
+	/*
+	 * We do not have off heap message queue enabled, i.e. we
+	 * need to add message queue to rootset...
+	 */
+	ErtsMessage *mp;
+
+	/* Ensure large enough rootset... */
+	if (n + p->msg.len > rootset->size) {
+	    Uint new_size = n + p->msg.len;
+	    ERTS_GC_ASSERT(roots == rootset->def);
+	    roots = erts_alloc(ERTS_ALC_T_ROOTSET,
+			       new_size*sizeof(Roots));
+	    sys_memcpy(roots, rootset->def, n*sizeof(Roots));
 	    rootset->size = new_size;
-	    avail = new_size - n;
 	}
-	if (mp->data.attached == NULL) {
-	    roots[n].v = mp->m;
-	    roots[n].sz = 2;
-	    n++;
-	    avail--;
+
+	for (mp = p->msg.first; mp; mp = mp->next) {
+
+	    if (!mp->data.attached) {
+		/*
+		 * Message may refer data on heap;
+		 * add it to rootset...
+		 */
+		roots[n].v = mp->m;
+		roots[n].sz = ERL_MESSAGE_REF_ARRAY_SZ;
+		n++;
+	    }
 	}
-        mp = mp->next;
+	break;
     }
+    }
+
+    ASSERT(rootset->size >= n);
+
     rootset->roots = roots;
     rootset->num_roots = n;
     return n;
@@ -2323,11 +2549,11 @@ sweep_off_heap(Process *p, int fullsweep)
      */
     while (ptr) {
 	if (IS_MOVED_BOXED(ptr->thing_word)) {
-	    ASSERT(!in_area(ptr, oheap, oheap_sz));
+	    ASSERT(!ErtsInArea(ptr, oheap, oheap_sz));
 	    *prev = ptr = (struct erl_off_heap_header*) boxed_val(ptr->thing_word);
 	    ASSERT(!IS_MOVED_BOXED(ptr->thing_word));
 	    if (ptr->thing_word == HEADER_PROC_BIN) {
-		int to_new_heap = !in_area(ptr, oheap, oheap_sz);
+		int to_new_heap = !ErtsInArea(ptr, oheap, oheap_sz);
 		ASSERT(to_new_heap == !seen_mature || (!to_new_heap && (seen_mature=1)));
 		if (to_new_heap) {
 		    bin_vheap += ptr->size / sizeof(Eterm);
@@ -2341,7 +2567,7 @@ sweep_off_heap(Process *p, int fullsweep)
 		ptr = ptr->next;
 	    }
 	}
-	else if (!in_area(ptr, oheap, oheap_sz)) {
+	else if (!ErtsInArea(ptr, oheap, oheap_sz)) {
 	    /* garbage */
 	    switch (thing_subtag(ptr->thing_word)) {
 	    case REFC_BINARY_SUBTAG:
@@ -2373,7 +2599,7 @@ sweep_off_heap(Process *p, int fullsweep)
      * generational collection - keep objects in list.
      */
     while (ptr) {
-	ASSERT(in_area(ptr, oheap, oheap_sz));
+	ASSERT(ErtsInArea(ptr, oheap, oheap_sz));
 	ASSERT(!IS_MOVED_BOXED(ptr->thing_word));       
 	if (ptr->thing_word == HEADER_PROC_BIN) {
 	    BIN_OLD_VHEAP(p) += ptr->size / sizeof(Eterm); /* for binary gc (words)*/
@@ -2392,7 +2618,6 @@ sweep_off_heap(Process *p, int fullsweep)
     }
     BIN_VHEAP_SZ(p)     = next_vheap_size(p, bin_vheap, BIN_VHEAP_SZ(p));
     MSO(p).overhead     = bin_vheap;
-    BIN_VHEAP_MATURE(p) = bin_vheap;
 
     /*
      * If we got any shrink candidates, check them out.
@@ -2463,7 +2688,7 @@ offset_heap(Eterm* hp, Uint sz, Sint offs, char* area, Uint area_size)
 	switch (primary_tag(val)) {
 	  case TAG_PRIMARY_LIST:
 	  case TAG_PRIMARY_BOXED:
-	      if (in_area(ptr_val(val), area, area_size)) {
+	      if (ErtsInArea(ptr_val(val), area, area_size)) {
 		  *hp = offset_ptr(val, offs);
 	      }
 	      hp++;
@@ -2485,7 +2710,7 @@ offset_heap(Eterm* hp, Uint sz, Sint offs, char* area, Uint area_size)
 		  {
 		      struct erl_off_heap_header* oh = (struct erl_off_heap_header*) hp;
 
-		      if (in_area(oh->next, area, area_size)) {
+		      if (ErtsInArea(oh->next, area, area_size)) {
 			  Eterm** uptr = (Eterm **) (void *) &oh->next;
 			  *uptr += offs; /* Patch the mso chain */
 		      }
@@ -2495,7 +2720,7 @@ offset_heap(Eterm* hp, Uint sz, Sint offs, char* area, Uint area_size)
 		{	
 		  ErlBinMatchState *ms = (ErlBinMatchState*) hp;
 		  ErlBinMatchBuffer *mb = &(ms->mb);
-		  if (in_area(ptr_val(mb->orig), area, area_size)) {
+		  if (ErtsInArea(ptr_val(mb->orig), area, area_size)) {
 		      mb->orig = offset_ptr(mb->orig, offs);
 		      mb->base = binary_bytes(mb->orig);
 		  }
@@ -2525,7 +2750,7 @@ offset_heap_ptr(Eterm* hp, Uint sz, Sint offs, char* area, Uint area_size)
 	switch (primary_tag(val)) {
 	case TAG_PRIMARY_LIST:
 	case TAG_PRIMARY_BOXED:
-	    if (in_area(ptr_val(val), area, area_size)) {
+	    if (ErtsInArea(ptr_val(val), area, area_size)) {
 		*hp = offset_ptr(val, offs);
 	    }
 	    hp++;
@@ -2540,7 +2765,7 @@ offset_heap_ptr(Eterm* hp, Uint sz, Sint offs, char* area, Uint area_size)
 static void
 offset_off_heap(Process* p, Sint offs, char* area, Uint area_size)
 {
-    if (MSO(p).first && in_area((Eterm *)MSO(p).first, area, area_size)) {
+    if (MSO(p).first && ErtsInArea((Eterm *)MSO(p).first, area, area_size)) {
         Eterm** uptr = (Eterm**) (void *) &MSO(p).first;
         *uptr += offs;
     }
@@ -2552,35 +2777,39 @@ offset_off_heap(Process* p, Sint offs, char* area, Uint area_size)
 static void
 offset_mqueue(Process *p, Sint offs, char* area, Uint area_size)
 {
-    ErlMessage* mp = p->msg.first;
-
-    while (mp != NULL) {
-        Eterm mesg = ERL_MESSAGE_TERM(mp);
-	if (is_value(mesg)) {
-	    switch (primary_tag(mesg)) {
-	    case TAG_PRIMARY_LIST:
-	    case TAG_PRIMARY_BOXED:
-		if (in_area(ptr_val(mesg), area, area_size)) {
-		    ERL_MESSAGE_TERM(mp) = offset_ptr(mesg, offs);
+    ErtsMessage* mp = p->msg.first;
+
+    if ((p->flags & (F_OFF_HEAP_MSGQ|F_OFF_HEAP_MSGQ_CHNG)) != F_OFF_HEAP_MSGQ) {
+
+	while (mp != NULL) {
+	    Eterm mesg = ERL_MESSAGE_TERM(mp);
+	    if (is_value(mesg)) {
+		switch (primary_tag(mesg)) {
+		case TAG_PRIMARY_LIST:
+		case TAG_PRIMARY_BOXED:
+		    if (ErtsInArea(ptr_val(mesg), area, area_size)) {
+			ERL_MESSAGE_TERM(mp) = offset_ptr(mesg, offs);
+		    }
+		    break;
 		}
-		break;
 	    }
-	}
-	mesg = ERL_MESSAGE_TOKEN(mp);
-	if (is_boxed(mesg) && in_area(ptr_val(mesg), area, area_size)) {
-	    ERL_MESSAGE_TOKEN(mp) = offset_ptr(mesg, offs);
-        }
+	    mesg = ERL_MESSAGE_TOKEN(mp);
+	    if (is_boxed(mesg) && ErtsInArea(ptr_val(mesg), area, area_size)) {
+		ERL_MESSAGE_TOKEN(mp) = offset_ptr(mesg, offs);
+	    }
 #ifdef USE_VM_PROBES
-	mesg = ERL_MESSAGE_DT_UTAG(mp);
-	if (is_boxed(mesg) && in_area(ptr_val(mesg), area, area_size)) {
-	    ERL_MESSAGE_DT_UTAG(mp) = offset_ptr(mesg, offs);
-        }
+	    mesg = ERL_MESSAGE_DT_UTAG(mp);
+	    if (is_boxed(mesg) && ErtsInArea(ptr_val(mesg), area, area_size)) {
+		ERL_MESSAGE_DT_UTAG(mp) = offset_ptr(mesg, offs);
+	    }
 #endif	
 	
-        ASSERT((is_nil(ERL_MESSAGE_TOKEN(mp)) ||
-		is_tuple(ERL_MESSAGE_TOKEN(mp)) ||
-		is_atom(ERL_MESSAGE_TOKEN(mp))));
-        mp = mp->next;
+	    ASSERT((is_nil(ERL_MESSAGE_TOKEN(mp)) ||
+		    is_tuple(ERL_MESSAGE_TOKEN(mp)) ||
+		    is_atom(ERL_MESSAGE_TOKEN(mp))));
+	    mp = mp->next;
+	}
+
     }
 }
 
@@ -2589,8 +2818,8 @@ offset_one_rootset(Process *p, Sint offs, char* area, Uint area_size,
 	       Eterm* objv, int nobj)
 {
     if (p->dictionary)  {
-	offset_heap(p->dictionary->data, 
-		    p->dictionary->used, 
+	offset_heap(ERTS_PD_START(p->dictionary),
+		    ERTS_PD_SIZE(p->dictionary),
 		    offs, area, area_size);
     }
 
@@ -2639,7 +2868,7 @@ reply_gc_info(void *vgcirp)
     Eterm **hpp;
     Uint sz, *szp;
     ErlOffHeap *ohp = NULL;
-    ErlHeapFragment *bp = NULL;
+    ErtsMessage *mp = NULL;
 
     ASSERT(esdp);
 
@@ -2665,12 +2894,13 @@ reply_gc_info(void *vgcirp)
 	if (hpp)
 	  break;
 	
-	hp = erts_alloc_message_heap(sz, &bp, &ohp, rp, &rp_locks);
+	mp = erts_alloc_message_heap(rp, &rp_locks, sz, &hp, &ohp);
+
 	szp = NULL;
 	hpp = &hp;
     }
 
-    erts_queue_message(rp, &rp_locks, bp, msg, NIL);
+    erts_queue_message(rp, &rp_locks, mp, msg, NIL);
 
     if (gcirp->req_sched == esdp->no)
 	rp_locks &= ~ERTS_PROC_LOCK_MAIN;
@@ -2722,36 +2952,49 @@ erts_gc_info_request(Process *c_p)
 static int
 within2(Eterm *ptr, Process *p, Eterm *real_htop)
 {
-    ErlHeapFragment* bp = MBUF(p);
-    ErlMessage* mp = p->msg.first;
-    Eterm *htop = real_htop ? real_htop : HEAP_TOP(p);
+    ErlHeapFragment* bp;
+    ErtsMessage* mp;
+    Eterm *htop, *heap;
+
+    if (p->abandoned_heap)
+	ERTS_GET_ORIG_HEAP(p, heap, htop);
+    else {
+	heap = p->heap;
+	htop = real_htop ? real_htop : HEAP_TOP(p);
+    }
 
     if (OLD_HEAP(p) && (OLD_HEAP(p) <= ptr && ptr < OLD_HEND(p))) {
         return 1;
     }
-    if (HEAP_START(p) <= ptr && ptr < htop) {
+    if (heap <= ptr && ptr < htop) {
         return 1;
     }
-    while (bp != NULL) {
-        if (bp->mem <= ptr && ptr < bp->mem + bp->used_size) {
-            return 1;
-        }
-        bp = bp->next;
-    }
+
+    mp = p->msg_frag;
+    bp = p->mbuf;
+
+    if (bp)
+	goto search_heap_frags;
+
     while (mp) {
-	if (mp->data.attached) {
-	    ErlHeapFragment *hfp;
-	    if (is_value(ERL_MESSAGE_TERM(mp)))
-		hfp = mp->data.heap_frag;
-	    else if (is_not_nil(ERL_MESSAGE_TOKEN(mp)))
-		hfp = erts_dist_ext_trailer(mp->data.dist_ext);
-	    else
-		hfp = NULL;
-	    if (hfp && hfp->mem <= ptr && ptr < hfp->mem + hfp->used_size)
+
+	if (mp->data.attached == ERTS_MSG_COMBINED_HFRAG)
+	    bp = &mp->hfrag;
+	else
+	    bp = mp->data.heap_frag;
+
+	mp = mp->next;
+
+    search_heap_frags:
+
+	while (bp) {
+	    if (bp->mem <= ptr && ptr < bp->mem + bp->used_size) {
 		return 1;
+	    }
+	    bp = bp->next;
 	}
-        mp = mp->next;
     }
+
     return 0;
 }
 
@@ -2773,11 +3016,11 @@ do {							\
 		 __FILE__, __LINE__, #EXP);		\
 } while (0)
 
+
 #ifdef ERTS_OFFHEAP_DEBUG_CHK_CIRCULAR_LIST
 #  define ERTS_OFFHEAP_VISITED_BIT ((Eterm) 1 << 31)
 #endif
 
-
 void
 erts_check_off_heap2(Process *p, Eterm *htop)
 {
@@ -2806,7 +3049,7 @@ erts_check_off_heap2(Process *p, Eterm *htop)
 	}
 	ERTS_CHK_OFFHEAP_ASSERT(refc >= 1);
 #ifdef ERTS_OFFHEAP_DEBUG_CHK_CIRCULAR_LIST
-	ERTS_CHK_OFFHEAP_ASSERT(!(u.hdr->thing_word & ERTS_EXTERNAL_VISITED_BIT));
+	ERTS_CHK_OFFHEAP_ASSERT(!(u.hdr->thing_word & ERTS_OFFHEAP_VISITED_BIT));
 	u.hdr->thing_word |= ERTS_OFFHEAP_VISITED_BIT;
 #endif
 	if (old) {
@@ -2819,7 +3062,7 @@ erts_check_off_heap2(Process *p, Eterm *htop)
 	}
     }
 
-#ifdef ERTS_OFFHEAP_DEBUG_CHK_CIRCULAR_EXTERNAL_LIST
+#ifdef ERTS_OFFHEAP_DEBUG_CHK_CIRCULAR_LIST
     for (u.hdr = MSO(p).first; u.hdr; u.hdr = u.hdr->next)
 	u.hdr->thing_word &= ~ERTS_OFFHEAP_VISITED_BIT;
 #endif
diff --git a/erts/emulator/beam/erl_gc.h b/erts/emulator/beam/erl_gc.h
index ecd1bf4d22..d603866cbf 100644
--- a/erts/emulator/beam/erl_gc.h
+++ b/erts/emulator/beam/erl_gc.h
@@ -69,17 +69,18 @@ do {                                                                    \
     while (nelts--) *HTOP++ = *PTR++;                                   \
 } while(0)
 
-#define in_area(ptr,start,nbytes) \
- ((UWord)((char*)(ptr) - (char*)(start)) < (nbytes))
-
 #if defined(DEBUG) || defined(ERTS_OFFHEAP_DEBUG)
 int within(Eterm *ptr, Process *p);
 #endif
 
-ERTS_GLB_INLINE Eterm follow_moved(Eterm term);
+#define ErtsInYoungGen(TPtr, Ptr, OldHeap, OldHeapSz)			\
+    (!erts_is_literal((TPtr), (Ptr))					\
+     & !ErtsInArea((Ptr), (OldHeap), (OldHeapSz)))
+
+ERTS_GLB_INLINE Eterm follow_moved(Eterm term, Eterm xptr_tag);
 
 #if ERTS_GLB_INLINE_INCL_FUNC_DEF
-ERTS_GLB_INLINE Eterm follow_moved(Eterm term)
+ERTS_GLB_INLINE Eterm follow_moved(Eterm term, Eterm xptr_tag)
 {
     Eterm* ptr;
     switch (primary_tag(term)) {
@@ -87,17 +88,18 @@ ERTS_GLB_INLINE Eterm follow_moved(Eterm term)
 	break;
     case TAG_PRIMARY_BOXED:
 	ptr = boxed_val(term);
-	if (IS_MOVED_BOXED(*ptr)) term = *ptr;
+	if (IS_MOVED_BOXED(*ptr)) term = (*ptr) | xptr_tag;
 	break;
     case TAG_PRIMARY_LIST:
 	ptr = list_val(term);
-	if (IS_MOVED_CONS(ptr[0])) term = ptr[1];
+	if (IS_MOVED_CONS(ptr[0])) term = (ptr[1]) | xptr_tag;
 	break;
     default:
 	ASSERT(!"strange tag in follow_moved");
     }
     return term;
 }
+
 #endif
 
 #endif /* ERL_GC_C__ || HIPE_GC_C__ */
@@ -106,6 +108,23 @@ ERTS_GLB_INLINE Eterm follow_moved(Eterm term)
  * Global exported
  */
 
+#define ERTS_IS_GC_DESIRED_INTERNAL(Proc, HTop, STop)			\
+    ((((STop) - (HTop) < (Proc)->mbuf_sz))				\
+     | ((Proc)->off_heap.overhead > (Proc)->bin_vheap_sz)		\
+     | !!((Proc)->flags & F_FORCE_GC))
+
+#define ERTS_IS_GC_DESIRED(Proc)					\
+    ERTS_IS_GC_DESIRED_INTERNAL((Proc), (Proc)->htop, (Proc)->stop)
+
+#define ERTS_FORCE_GC_INTERNAL(Proc, FCalls)				\
+    do {								\
+	(Proc)->flags |= F_FORCE_GC;					\
+	ERTS_VBUMP_ALL_REDS_INTERNAL((Proc), (FCalls));			\
+    } while (0)
+
+#define ERTS_FORCE_GC(Proc)						\
+    ERTS_FORCE_GC_INTERNAL((Proc), (Proc)->fcalls)
+
 extern Uint erts_test_long_gc_sleep;
 
 typedef struct {
@@ -115,8 +134,11 @@ typedef struct {
 
 void erts_gc_info(ErtsGCInfo *gcip);
 void erts_init_gc(void);
-int erts_garbage_collect(struct process*, int, Eterm*, int);
+int erts_garbage_collect_nobump(struct process*, int, Eterm*, int);
+void erts_garbage_collect(struct process*, int, Eterm*, int);
 void erts_garbage_collect_hibernate(struct process* p);
+Eterm erts_gc_after_bif_call_lhf(struct process* p, ErlHeapFragment *live_hf_end,
+				 Eterm result, Eterm* regs, Uint arity);
 Eterm erts_gc_after_bif_call(struct process* p, Eterm result, Eterm* regs, Uint arity);
 void erts_garbage_collect_literals(struct process* p, Eterm* literals,
 				   Uint lit_size,
diff --git a/erts/emulator/beam/erl_hl_timer.c b/erts/emulator/beam/erl_hl_timer.c
index 51a0d68247..734057a12c 100644
--- a/erts/emulator/beam/erl_hl_timer.c
+++ b/erts/emulator/beam/erl_hl_timer.c
@@ -1245,7 +1245,9 @@ hlt_bif_timer_timeout(ErtsHLTimer *tmr, Uint32 roflgs)
 	 * the middle of tree destruction).
 	 */
 	if (!ERTS_PROC_IS_EXITING(proc)) {
-	    erts_queue_message(proc, &proc_locks, tmr->btm.bp,
+	    ErtsMessage *mp = erts_alloc_message(0, NULL);
+	    mp->data.heap_frag = tmr->btm.bp;
+	    erts_queue_message(proc, &proc_locks, mp,
 			       tmr->btm.message, NIL);
 	    erts_smp_proc_unlock(proc, ERTS_PROC_LOCKS_MSG_SEND);
 	    queued_message = 1;
@@ -1926,36 +1928,31 @@ access_sched_local_btm(Process *c_p, Eterm pid,
 
     if (proc) {
 	Uint hsz;
-	ErlOffHeap *ohp;
-	ErlHeapFragment* bp;
+	ErtsMessage *mp;
 	Eterm *hp, msg, ref, result;
+	ErlOffHeap *ohp;
+	Uint32 *refn;
 #ifdef ERTS_HLT_DEBUG
 	Eterm *hp_end;
 #endif
 
-	hsz = 3; /* 2-tuple */
-	if (!async)
-	    hsz += REF_THING_SIZE;
+	hsz = REF_THING_SIZE;
+	if (async) {
+	    refn = trefn; /* timer ref */
+	    hsz += 4; /* 3-tuple */
+	}
 	else {
-	    if (is_non_value(tref) || proc != c_p)
-		hsz += REF_THING_SIZE;
-	    hsz += 1; /* upgrade to 3-tuple */
+	    refn = rrefn; /* request ref */
+	    hsz += 3; /* 2-tuple */
 	}
+
+	ERTS_HLT_ASSERT(refn);
+
 	if (time_left > (Sint64) MAX_SMALL)
 	    hsz += ERTS_SINT64_HEAP_SIZE(time_left);
 
-	if (proc == c_p) {
-	    bp = NULL;
-	    ohp = NULL;
-	    hp = HAlloc(c_p, hsz);
-	}
-	else {
-	    hp = erts_alloc_message_heap(hsz,
-					 &bp,
-					 &ohp,
-					 proc,
-					 &proc_locks);
-	}
+	mp = erts_alloc_message_heap(proc, &proc_locks,
+				     hsz, &hp, &ohp);
 
 #ifdef ERTS_HLT_DEBUG
 	hp_end = hp + hsz;
@@ -1968,35 +1965,22 @@ access_sched_local_btm(Process *c_p, Eterm pid,
 	else
 	    result = erts_sint64_to_big(time_left, &hp);
 
-	if (!async) {
-	    write_ref_thing(hp,
-			    rrefn[0],
-			    rrefn[1],
-			    rrefn[2]);
-	    ref = make_internal_ref(hp);
-	    hp += REF_THING_SIZE;
-	    msg = TUPLE2(hp, ref, result);
+	write_ref_thing(hp,
+			refn[0],
+			refn[1],
+			refn[2]);
+	ref = make_internal_ref(hp);
+	hp += REF_THING_SIZE;
 
-	    ERTS_HLT_ASSERT(hp + 3 == hp_end);
-	}
-	else {
-	    Eterm tag = cancel ? am_cancel_timer : am_read_timer;
-	    if (is_value(tref) && proc == c_p)
-		ref = tref;
-	    else {
-		write_ref_thing(hp,
-				trefn[0],
-				trefn[1],
-				trefn[2]);
-		ref = make_internal_ref(hp);
-		hp += REF_THING_SIZE;
-	    }
-	    msg = TUPLE3(hp, tag, ref, result);
+	msg = (async
+	       ? TUPLE3(hp, (cancel
+			     ? am_cancel_timer
+			     : am_read_timer), ref, result)
+	       : TUPLE2(hp, ref, result));
 
-	    ERTS_HLT_ASSERT(hp + 4 == hp_end);
+	ERTS_HLT_ASSERT(hp + (async ? 4 : 3) == hp_end);
 
-	}
-	erts_queue_message(proc, &proc_locks, bp, msg, NIL);
+	erts_queue_message(proc, &proc_locks, mp, msg, NIL);
 
 	if (c_p)
 	    proc_locks &= ~ERTS_PROC_LOCK_MAIN;
@@ -2034,7 +2018,7 @@ bif_timer_access_request(void *vreq)
 static int
 try_access_sched_remote_btm(ErtsSchedulerData *esdp,
 			    Process *c_p, Uint32 sid,
-			    Eterm tref, Uint32 *trefn,
+			    Uint32 *trefn,
 			    int async, int cancel,
 			    int info, Eterm *resp)
 {
@@ -2093,21 +2077,31 @@ try_access_sched_remote_btm(ErtsSchedulerData *esdp,
 	}
     }
     else {
-	Eterm tag, res, msg;
+	ErtsMessage *mp;
+	Eterm tag, res, msg, tref;
 	Uint hsz;
 	Eterm *hp;
 	ErtsProcLocks proc_locks = ERTS_PROC_LOCK_MAIN;
+	ErlOffHeap *ohp;
 
-	hsz = 4;
+	hsz = 4 + REF_THING_SIZE;
 	if (time_left > (Sint64) MAX_SMALL)
 	    hsz += ERTS_SINT64_HEAP_SIZE(time_left);
 
-	hp = HAlloc(c_p, hsz);
+	mp = erts_alloc_message_heap(c_p, &proc_locks,
+				     hsz, &hp, &ohp);
 	if (cancel)
 	    tag = am_cancel_timer;
 	else
 	    tag = am_read_timer;
 
+	write_ref_thing(hp,
+			trefn[0],
+			trefn[1],
+			trefn[2]);
+	tref = make_internal_ref(hp);
+	hp += REF_THING_SIZE;
+
 	if (time_left < 0)
 	    res = am_false;
 	else if (time_left <= (Sint64) MAX_SMALL)
@@ -2117,7 +2111,7 @@ try_access_sched_remote_btm(ErtsSchedulerData *esdp,
 
 	msg = TUPLE3(hp, tag, tref, res);
 
-	erts_queue_message(c_p, &proc_locks, NULL, msg, NIL);
+	erts_queue_message(c_p, &proc_locks, mp, msg, NIL);
 
 	proc_locks &= ~ERTS_PROC_LOCK_MAIN;
 	if (proc_locks)
@@ -2158,8 +2152,8 @@ access_bif_timer(Process *c_p, Eterm tref, int cancel, int async, int info)
 				     info);
 	ERTS_BIF_PREP_RET(ret, res);
     }
-    else if (try_access_sched_remote_btm(esdp, c_p, sid,
-					 tref, trefn,
+    else if (try_access_sched_remote_btm(esdp, c_p,
+					 sid, trefn,
 					 async, cancel,
 					 info, &res)) {
 	ERTS_BIF_PREP_RET(ret, res);
diff --git a/erts/emulator/beam/erl_init.c b/erts/emulator/beam/erl_init.c
index d9c3b0dcf4..42aca726bf 100644
--- a/erts/emulator/beam/erl_init.c
+++ b/erts/emulator/beam/erl_init.c
@@ -91,11 +91,6 @@ const int etp_arch_bits = 32;
 #else
 # error "Not 64-bit, nor 32-bit arch"
 #endif
-#if HALFWORD_HEAP
-const int etp_halfword = 1;
-#else
-const int etp_halfword = 0;
-#endif
 #ifdef HIPE
 const int etp_hipe = 1;
 #else
@@ -197,7 +192,7 @@ Uint32 verbose;             /* See erl_debug.h for information about verbose */
 
 int erts_atom_table_size = ATOM_LIMIT;	/* Maximum number of atoms */
 
-int erts_pd_initial_size = 10;
+int erts_pd_initial_size = 8;  /* must be power of 2 */
 
 int erts_modified_timing_level;
 
@@ -393,6 +388,7 @@ erl_init(int ncpu,
     erts_mseg_late_init(); /* Must be after timer (erts_init_time()) and thread
 			      initializations */
 #endif
+    erl_sys_late_init();
 #ifdef HIPE
     hipe_mode_switch_init(); /* Must be after init_load/beam_catches/init */
 #endif
@@ -436,13 +432,36 @@ erl_first_process_otp(char* modname, void* code, unsigned size, int argc, char**
     hp += 2;
     args = CONS(hp, env, args);
 
-    so.flags = SPO_SYSTEM_PROC;
+    so.flags = erts_default_spo_flags|SPO_SYSTEM_PROC;
     res = erl_create_process(&parent, start_mod, am_start, args, &so);
     erts_smp_proc_unlock(&parent, ERTS_PROC_LOCK_MAIN);
     erts_cleanup_empty_process(&parent);
     return res;
 }
 
+static Eterm
+erl_system_process_otp(Eterm parent_pid, char* modname)
+{
+    Eterm start_mod;
+    Process* parent;
+    ErlSpawnOpts so;
+    Eterm res;
+
+    start_mod = erts_atom_put((byte *) modname, sys_strlen(modname), ERTS_ATOM_ENC_LATIN1, 1);
+    if (erts_find_function(start_mod, am_start, 0,
+			   erts_active_code_ix()) == NULL) {
+	erl_exit(5, "No function %s:start/0\n", modname);
+    }
+
+    parent = erts_pid2proc(NULL, 0, parent_pid, ERTS_PROC_LOCK_MAIN);
+
+    so.flags = erts_default_spo_flags|SPO_SYSTEM_PROC;
+    res = erl_create_process(parent, start_mod, am_start, NIL, &so);
+    erts_smp_proc_unlock(parent, ERTS_PROC_LOCK_MAIN);
+    return res;
+}
+
+
 Eterm
 erts_preloaded(Process* p)
 {
@@ -635,6 +654,8 @@ void erts_usage(void)
 
     erts_fprintf(stderr, "-W<i|w|e>      set error logger warnings mapping,\n");
     erts_fprintf(stderr, "               see error_logger documentation for details\n");
+    erts_fprintf(stderr, "-xmqd  val     set default message queue data flag for processes,\n");
+    erts_fprintf(stderr, "               valid values are: off_heap | on_heap | mixed\n");
     erts_fprintf(stderr, "-zdbbl size    set the distribution buffer busy limit in kilobytes\n");
     erts_fprintf(stderr, "               valid range is [1-%d]\n", INT_MAX/1024);
     erts_fprintf(stderr, "-zdntgc time   set delayed node table gc in seconds\n");
@@ -1236,6 +1257,7 @@ erl_start(int argc, char **argv)
     ErtsTimeWarpMode time_warp_mode;
     int node_tab_delete_delay = ERTS_NODE_TAB_DELAY_GC_DEFAULT;
     ErtsDbSpinCount db_spin_count = ERTS_DB_SPNCNT_NORMAL;
+    Eterm otp_ring0_pid;
 
     set_default_time_adj(&time_correction,
 			 &time_warp_mode);
@@ -1405,6 +1427,7 @@ erl_start(int argc, char **argv)
 		    case 't': verbose |= DEBUG_THREADS; break;
 		    case 'p': verbose |= DEBUG_PROCESSES; break;
 		    case 'm': verbose |= DEBUG_MESSAGES; break;
+		    case 'c': verbose |= DEBUG_SHCOPY; break;
 		    default : erts_fprintf(stderr,"Unknown verbose option: %c\n",*ch);
 		    }
 		}
@@ -1417,6 +1440,7 @@ erl_start(int argc, char **argv)
 	    if (verbose & DEBUG_THREADS) erts_printf("THREADS ");
 	    if (verbose & DEBUG_PROCESSES) erts_printf("PROCESSES ");
 	    if (verbose & DEBUG_MESSAGES) erts_printf("MESSAGES ");
+	    if (verbose & DEBUG_SHCOPY) erts_printf("SHCOPY ");
             erts_printf("\n");
 #else
 	    erts_fprintf(stderr, "warning: -v (only in debug compiled code)\n");
@@ -1479,7 +1503,7 @@ erl_start(int argc, char **argv)
 		VERBOSE(DEBUG_SYSTEM, ("using minimum heap size %d\n", H_MIN_SIZE));
 	    } else if (has_prefix("pds", sub_param)) {
 		arg = get_arg(sub_param+3, argv[i+1], &i);
-		if ((erts_pd_initial_size = atoi(arg)) <= 0) {
+		if (!erts_pd_set_initial_size(atoi(arg))) {
 		    erts_fprintf(stderr, "bad initial process dictionary size %s\n", arg);
 		    erts_usage();
 		}
@@ -2020,6 +2044,32 @@ erl_start(int argc, char **argv)
 	    }
 	    break;
 
+	case 'x': {
+	    char *sub_param = argv[i]+2;
+	    if (has_prefix("mqd", sub_param)) {
+		arg = get_arg(sub_param+3, argv[i+1], &i);
+		if (sys_strcmp(arg, "mixed") == 0)
+		    erts_default_spo_flags &= ~(SPO_ON_HEAP_MSGQ|SPO_OFF_HEAP_MSGQ);
+		else if (sys_strcmp(arg, "on_heap") == 0) {
+		    erts_default_spo_flags &= ~SPO_OFF_HEAP_MSGQ;
+		    erts_default_spo_flags |= SPO_ON_HEAP_MSGQ;
+		}
+		else if (sys_strcmp(arg, "off_heap") == 0) {
+		    erts_default_spo_flags &= ~SPO_ON_HEAP_MSGQ;
+		    erts_default_spo_flags |= SPO_OFF_HEAP_MSGQ;
+		}
+		else {
+		    erts_fprintf(stderr,
+				 "Invalid message_queue_data flag: %s\n", arg);
+		    erts_usage();
+		}
+	    } else {
+		erts_fprintf(stderr, "bad -x option %s\n", argv[i]);
+		erts_usage();
+	    }
+	    break;
+        }
+
 	case 'z': {
 	    char *sub_param = argv[i]+2;
 
@@ -2073,7 +2123,8 @@ erl_start(int argc, char **argv)
 				 "Invalid ets busy wait threshold: %s\n", arg);
 		    erts_usage();
 		}
-	    } else {
+	    }
+	    else {
 		erts_fprintf(stderr, "bad -z option %s\n", argv[i]);
 		erts_usage();
 	    }
@@ -2156,8 +2207,10 @@ erl_start(int argc, char **argv)
 
     erts_initialized = 1;
 
-    (void) erl_first_process_otp("otp_ring0", NULL, 0,
-				 boot_argc, boot_argv);
+    otp_ring0_pid = erl_first_process_otp("otp_ring0", NULL, 0,
+					  boot_argc, boot_argv);
+
+    (void) erl_system_process_otp(otp_ring0_pid, "erts_code_purger");
 
 #ifdef ERTS_SMP
     erts_start_schedulers();
diff --git a/erts/emulator/beam/erl_lock_check.c b/erts/emulator/beam/erl_lock_check.c
index 3b3b247020..f7b4bd8041 100644
--- a/erts/emulator/beam/erl_lock_check.c
+++ b/erts/emulator/beam/erl_lock_check.c
@@ -113,9 +113,6 @@ static erts_lc_lock_order_t erts_lock_order[] = {
     {	"environ",				NULL			},
 #endif
     {	"efile_drv",				"address"		},
-#if defined(ENABLE_CHILD_WAITER_THREAD) || defined(ERTS_SMP)
-    {	"child_status",				NULL			},
-#endif
     {	"drv_ev_state_grow",			NULL,   		},
     {	"drv_ev_state",				"address"		},
     {	"safe_hash",				"address"		},
@@ -156,9 +153,6 @@ static erts_lc_lock_order_t erts_lock_order[] = {
     {	"instr",				NULL			},
     {	"alcu_allocator",			"index"			},
     {	"mseg",					NULL			},
-#if HALFWORD_HEAP
-    {	"pmmap",				NULL			},
-#endif
 #ifdef ERTS_SMP
     {	"port_task_pre_alloc_lock",		"address"		},
     {	"proclist_pre_alloc_lock",		"address"		},
diff --git a/erts/emulator/beam/erl_map.c b/erts/emulator/beam/erl_map.c
index a91e36e3c5..d0ffb11e79 100644
--- a/erts/emulator/beam/erl_map.c
+++ b/erts/emulator/beam/erl_map.c
@@ -32,7 +32,9 @@
 #include "global.h"
 #include "erl_process.h"
 #include "error.h"
+#define ERL_WANT_HIPE_BIF_WRAPPER__
 #include "bif.h"
+#undef ERL_WANT_HIPE_BIF_WRAPPER__
 #include "erl_binary.h"
 
 #include "erl_map.h"
@@ -170,26 +172,22 @@ BIF_RETTYPE maps_to_list_1(BIF_ALIST_1) {
  */
 
 const Eterm *
-#if HALFWORD_HEAP
-erts_maps_get_rel(Eterm key, Eterm map, Eterm *map_base)
-#else
 erts_maps_get(Eterm key, Eterm map)
-#endif
 {
     Uint32 hx;
-    if (is_flatmap_rel(map, map_base)) {
+    if (is_flatmap(map)) {
 	Eterm *ks, *vs;
 	flatmap_t *mp;
 	Uint n, i;
 
-	mp  = (flatmap_t *)flatmap_val_rel(map, map_base);
+	mp  = (flatmap_t *)flatmap_val(map);
 	n   = flatmap_get_size(mp);
 
 	if (n == 0) {
 	    return NULL;
 	}
 
-	ks  = (Eterm *)tuple_val_rel(mp->keys, map_base) + 1;
+	ks  = (Eterm *)tuple_val(mp->keys) + 1;
 	vs  = flatmap_get_values(mp);
 
 	if (is_immed(key)) {
@@ -201,16 +199,16 @@ erts_maps_get(Eterm key, Eterm map)
 	}
 
 	for (i = 0; i < n; i++) {
-	    if (eq_rel(ks[i], map_base, key, NULL)) {
+	    if (EQ(ks[i], key)) {
 		return &vs[i];
 	    }
 	}
 	return NULL;
     }
-    ASSERT(is_hashmap_rel(map, map_base));
+    ASSERT(is_hashmap(map));
     hx = hashmap_make_hash(key);
 
-    return erts_hashmap_get_rel(hx, key, map, map_base);
+    return erts_hashmap_get(hx, key, map);
 }
 
 BIF_RETTYPE maps_find_2(BIF_ALIST_2) {
@@ -952,8 +950,11 @@ BIF_RETTYPE maps_keys_1(BIF_ALIST_1) {
     BIF_P->fvalue = BIF_ARG_1;
     BIF_ERROR(BIF_P, BADMAP);
 }
+
 /* maps:merge/2 */
 
+HIPE_WRAPPER_BIF_DISABLE_GC(maps_merge, 2)
+
 BIF_RETTYPE maps_merge_2(BIF_ALIST_2) {
     if (is_flatmap(BIF_ARG_1)) {
 	if (is_flatmap(BIF_ARG_2)) {
@@ -1993,11 +1994,7 @@ Eterm* hashmap_iterator_prev(ErtsWStack* s) {
 }
 
 const Eterm *
-#if HALFWORD_HEAP
-erts_hashmap_get_rel(Uint32 hx, Eterm key, Eterm node, Eterm *map_base)
-#else
 erts_hashmap_get(Uint32 hx, Eterm key, Eterm node)
-#endif
 {
     Eterm *ptr, hdr, *res;
     Uint ix, lvl = 0;
@@ -2006,7 +2003,7 @@ erts_hashmap_get(Uint32 hx, Eterm key, Eterm node)
     UseTmpHeapNoproc(2);
 
     ASSERT(is_boxed(node));
-    ptr = boxed_val_rel(node, map_base);
+    ptr = boxed_val(node);
     hdr = *ptr;
     ASSERT(is_header(hdr));
     ASSERT(is_hashmap_header_head(hdr));
@@ -2027,15 +2024,15 @@ erts_hashmap_get(Uint32 hx, Eterm key, Eterm node)
         node  = ptr[ix+1];
 
         if (is_list(node)) { /* LEAF NODE [K|V] */
-            ptr = list_val_rel(node,map_base);
-            res = eq_rel(CAR(ptr), map_base, key, NULL) ? &(CDR(ptr)) : NULL;
+            ptr = list_val(node);
+            res = EQ(CAR(ptr), key) ? &(CDR(ptr)) : NULL;
             break;
         }
 
         hx = hashmap_shift_hash(th,hx,lvl,key);
 
         ASSERT(is_boxed(node));
-        ptr = boxed_val_rel(node, map_base);
+        ptr = boxed_val(node);
         hdr = *ptr;
         ASSERT(is_header(hdr));
         ASSERT(!is_hashmap_header_head(hdr));
@@ -2701,32 +2698,88 @@ BIF_RETTYPE erts_internal_map_to_tuple_keys_1(BIF_ALIST_1) {
 }
 
 /*
- * erts_internal:map_type/1
+ * erts_internal:term_type/1
  *
  * Used in erts_debug:size/1
  */
 
-BIF_RETTYPE erts_internal_map_type_1(BIF_ALIST_1) {
-    DECL_AM(hashmap);
-    DECL_AM(hashmap_node);
-    DECL_AM(flatmap);
-    if (is_map(BIF_ARG_1)) {
-        Eterm hdr = *(boxed_val(BIF_ARG_1));
-        ASSERT(is_header(hdr));
-        switch (hdr & _HEADER_MAP_SUBTAG_MASK) {
-            case HAMT_SUBTAG_HEAD_FLATMAP:
-                BIF_RET(AM_flatmap);
-            case HAMT_SUBTAG_HEAD_ARRAY:
-            case HAMT_SUBTAG_HEAD_BITMAP:
-                BIF_RET(AM_hashmap);
-            case HAMT_SUBTAG_NODE_BITMAP:
-                BIF_RET(AM_hashmap_node);
-            default:
-                erl_exit(1, "bad header");
+BIF_RETTYPE erts_internal_term_type_1(BIF_ALIST_1) {
+    Eterm obj = BIF_ARG_1;
+    switch (primary_tag(obj)) {
+        case TAG_PRIMARY_LIST:
+            BIF_RET(ERTS_MAKE_AM("list"));
+        case TAG_PRIMARY_BOXED: {
+            Eterm hdr = *boxed_val(obj);
+            ASSERT(is_header(hdr));
+            switch (hdr & _TAG_HEADER_MASK) {
+                case ARITYVAL_SUBTAG:
+                    BIF_RET(ERTS_MAKE_AM("tuple"));
+                case EXPORT_SUBTAG:
+                    BIF_RET(ERTS_MAKE_AM("export"));
+                case FUN_SUBTAG:
+                    BIF_RET(ERTS_MAKE_AM("fun"));
+                case MAP_SUBTAG:
+                    switch (MAP_HEADER_TYPE(hdr)) {
+                        case MAP_HEADER_TAG_FLATMAP_HEAD :
+                            BIF_RET(ERTS_MAKE_AM("flatmap"));
+                        case MAP_HEADER_TAG_HAMT_HEAD_BITMAP :
+                        case MAP_HEADER_TAG_HAMT_HEAD_ARRAY :
+                            BIF_RET(ERTS_MAKE_AM("hashmap"));
+                        case MAP_HEADER_TAG_HAMT_NODE_BITMAP :
+                            BIF_RET(ERTS_MAKE_AM("hashmap_node"));
+                        default:
+                            erl_exit(ERTS_ABORT_EXIT, "term_type: bad map header type %d\n", MAP_HEADER_TYPE(hdr));
+                    }
+                case REFC_BINARY_SUBTAG:
+                    BIF_RET(ERTS_MAKE_AM("refc_binary"));
+                case HEAP_BINARY_SUBTAG:
+                    BIF_RET(ERTS_MAKE_AM("heap_binary"));
+                case SUB_BINARY_SUBTAG:
+                    BIF_RET(ERTS_MAKE_AM("sub_binary"));
+                case BIN_MATCHSTATE_SUBTAG:
+                    BIF_RET(ERTS_MAKE_AM("matchstate"));
+                case POS_BIG_SUBTAG:
+                case NEG_BIG_SUBTAG:
+                    BIF_RET(ERTS_MAKE_AM("bignum"));
+                case REF_SUBTAG:
+                    BIF_RET(ERTS_MAKE_AM("reference"));
+                case EXTERNAL_REF_SUBTAG:
+                    BIF_RET(ERTS_MAKE_AM("external_reference"));
+                case EXTERNAL_PID_SUBTAG:
+                    BIF_RET(ERTS_MAKE_AM("external_pid"));
+                case EXTERNAL_PORT_SUBTAG:
+                    BIF_RET(ERTS_MAKE_AM("external_port"));
+                case FLOAT_SUBTAG:
+                    BIF_RET(ERTS_MAKE_AM("hfloat"));
+                default:
+                    erl_exit(ERTS_ABORT_EXIT, "term_type: Invalid tag (0x%X)\n", hdr);
+            }
         }
+        case TAG_PRIMARY_IMMED1:
+            switch (obj & _TAG_IMMED1_MASK) {
+                case _TAG_IMMED1_SMALL:
+                    BIF_RET(ERTS_MAKE_AM("fixnum"));
+                case _TAG_IMMED1_PID:
+                    BIF_RET(ERTS_MAKE_AM("pid"));
+                case _TAG_IMMED1_PORT:
+                    BIF_RET(ERTS_MAKE_AM("port"));
+                case _TAG_IMMED1_IMMED2:
+                    switch (obj & _TAG_IMMED2_MASK) {
+                        case _TAG_IMMED2_ATOM:
+                            BIF_RET(ERTS_MAKE_AM("atom"));
+                        case _TAG_IMMED2_CATCH:
+                            BIF_RET(ERTS_MAKE_AM("catch"));
+                        case _TAG_IMMED2_NIL:
+                            BIF_RET(ERTS_MAKE_AM("nil"));
+                        default:
+                            erl_exit(ERTS_ABORT_EXIT, "term_type: Invalid tag (0x%X)\n", obj);
+                    }
+                default:
+                    erl_exit(ERTS_ABORT_EXIT, "term_type: Invalid tag (0x%X)\n", obj);
+            }
+        default:
+            erl_exit(ERTS_ABORT_EXIT, "term_type: Invalid tag (0x%X)\n", obj);
     }
-    BIF_P->fvalue = BIF_ARG_1;
-    BIF_ERROR(BIF_P, BADMAP);
 }
 
 /*
diff --git a/erts/emulator/beam/erl_map.h b/erts/emulator/beam/erl_map.h
index c391de3f11..052fa99f03 100644
--- a/erts/emulator/beam/erl_map.h
+++ b/erts/emulator/beam/erl_map.h
@@ -57,7 +57,6 @@ typedef struct flatmap_s {
 
 
 #define hashmap_size(x)               (((hashmap_head_t*) hashmap_val(x))->size)
-#define hashmap_size_rel(RTERM, BASE) hashmap_size(rterm2wterm(RTERM, BASE))
 #define hashmap_make_hash(Key)        make_internal_hash(Key)
 
 #define hashmap_restore_hash(Heap,Lvl,Key) \
@@ -104,23 +103,9 @@ Eterm  erts_hashmap_from_array(ErtsHeapFactory*, Eterm *leafs, Uint n, int rejec
 Eterm  erts_hashmap_from_ks_and_vs_extra(Process *p, Eterm *ks, Eterm *vs, Uint n,
 					 Eterm k, Eterm v);
 
-const Eterm *
-#if HALFWORD_HEAP
-erts_maps_get_rel(Eterm key, Eterm map, Eterm *map_base);
-#  define erts_maps_get(A, B) erts_maps_get_rel(A, B, NULL)
-#else
-erts_maps_get(Eterm key, Eterm map);
-#  define erts_maps_get_rel(A, B, B_BASE) erts_maps_get(A, B)
-#endif
+const Eterm *erts_maps_get(Eterm key, Eterm map);
 
-const Eterm *
-#if HALFWORD_HEAP
-erts_hashmap_get_rel(Uint32 hx, Eterm key, Eterm node, Eterm *map_base);
-#  define erts_hashmap_get(Hx, K, M) erts_hashmap_get_rel(Hx, K, M, NULL)
-#else
-erts_hashmap_get(Uint32 hx, Eterm key, Eterm map);
-#  define erts_hashmap_get_rel(Hx, K, M, M_BASE) erts_hashmap_get(Hx, K, M)
-#endif
+const Eterm *erts_hashmap_get(Uint32 hx, Eterm key, Eterm map);
 
 /* hamt nodes v2.0
  *
@@ -195,14 +180,17 @@ typedef struct hashmap_head_s {
    [one cons cell + one list term in parent node] per key
    [one header + one boxed term in parent node] per inner node
    [one header + one size word] for root node
+   Observed average number of nodes per key is about 0.35.
 */
-#define HASHMAP_HEAP_SIZE(KEYS,NODES) ((KEYS)*3 + (NODES)*2)
+#define HASHMAP_WORDS_PER_KEY 3
+#define HASHMAP_WORDS_PER_NODE 2
 #ifdef DEBUG
-#  define HASHMAP_ESTIMATED_NODE_COUNT(KEYS) (KEYS)
+#  define HASHMAP_ESTIMATED_TOT_NODE_SIZE(KEYS) \
+    (HASHMAP_WORDS_PER_NODE * (KEYS) * 3/10)   /* slightly under estimated */
 #else
-#  define HASHMAP_ESTIMATED_NODE_COUNT(KEYS) (2*(KEYS)/5)
+#  define HASHMAP_ESTIMATED_TOT_NODE_SIZE(KEYS) \
+    (HASHMAP_WORDS_PER_NODE * (KEYS) * 4/10)   /* slightly over estimated */
 #endif
 #define HASHMAP_ESTIMATED_HEAP_SIZE(KEYS) \
-        HASHMAP_HEAP_SIZE(KEYS,HASHMAP_ESTIMATED_NODE_COUNT(KEYS))
-
+        ((KEYS)*HASHMAP_WORDS_PER_KEY + HASHMAP_ESTIMATED_TOT_NODE_SIZE(KEYS))
 #endif
diff --git a/erts/emulator/beam/erl_message.c b/erts/emulator/beam/erl_message.c
index ef52823287..b3e74e3e6a 100644
--- a/erts/emulator/beam/erl_message.c
+++ b/erts/emulator/beam/erl_message.c
@@ -33,8 +33,8 @@
 #include "erl_binary.h"
 #include "dtrace-wrapper.h"
 
-ERTS_SCHED_PREF_QUICK_ALLOC_IMPL(message,
-				 ErlMessage,
+ERTS_SCHED_PREF_QUICK_ALLOC_IMPL(message_ref,
+				 ErtsMessageRef,
 				 ERL_MESSAGE_BUF_SZ,
 				 ERTS_ALC_T_MSG_REF)
 
@@ -44,27 +44,20 @@ ERTS_SCHED_PREF_QUICK_ALLOC_IMPL(message,
 #undef HARD_DEBUG
 #endif
 
-
-
-
-#ifdef DEBUG
-static ERTS_INLINE int in_heapfrag(const Eterm* ptr, const ErlHeapFragment *bp)
+void
+init_message(void)
 {
-    return ((unsigned)(ptr - bp->mem) < bp->used_size);
+    init_message_ref_alloc();
 }
-#endif
-
 
-void
-init_message(void)
+void *erts_alloc_message_ref(void)
 {
-    init_message_alloc();
+    return (void *) message_ref_alloc();
 }
 
-void
-free_message(ErlMessage* mp)
+void erts_free_message_ref(void *mp)
 {
-    message_free(mp);
+    message_ref_free((ErtsMessageRef *) mp);
 }
 
 /* Allocate message buffer (size in words) */
@@ -74,7 +67,8 @@ new_message_buffer(Uint size)
     ErlHeapFragment* bp;
     bp = (ErlHeapFragment*) ERTS_HEAP_ALLOC(ERTS_ALC_T_HEAP_FRAG,
 					    ERTS_HEAP_FRAG_SIZE(size));
-    ERTS_INIT_HEAP_FRAG(bp, size);
+    ERTS_INIT_HEAP_FRAG(bp, size, size);
+    VERBOSE(DEBUG_SHCOPY, ("[pid=%T] new message buffer %p\n", erts_get_current_pid(), bp->mem));
     return bp;
 }
 
@@ -203,83 +197,57 @@ free_message_buffer(ErlHeapFragment* bp)
     }while (bp != NULL);
 }
 
-static ERTS_INLINE void
-link_mbuf_to_proc(Process *proc, ErlHeapFragment *bp)
+void
+erts_cleanup_messages(ErtsMessage *msgp)
 {
-    if (bp) {
-	/* Link the message buffer */
-	bp->next = MBUF(proc);
-	MBUF(proc) = bp;
-	MBUF_SIZE(proc) += bp->used_size;
-	FLAGS(proc) |= F_FORCE_GC;
-
-	/* Move any off_heap's into the process */
-	if (bp->off_heap.first != NULL) {
-	    struct erl_off_heap_header** next_p = &bp->off_heap.first;
-	    while (*next_p != NULL) {
-		next_p = &((*next_p)->next);
+    ErtsMessage *mp = msgp;
+    while (mp) {
+	ErtsMessage *fmp;
+	ErlHeapFragment *bp;
+	if (is_non_value(ERL_MESSAGE_TERM(mp))) {
+	    if (is_not_immed(ERL_MESSAGE_TOKEN(mp))) {
+		bp = (ErlHeapFragment *) mp->data.dist_ext->ext_endp;
+		erts_cleanup_offheap(&bp->off_heap);
+	    }
+	    if (mp->data.dist_ext)
+		erts_free_dist_ext_copy(mp->data.dist_ext);
+	}
+	else {
+	    if (mp->data.attached != ERTS_MSG_COMBINED_HFRAG)
+		bp = mp->data.heap_frag;
+	    else {
+		bp = mp->hfrag.next;
+		erts_cleanup_offheap(&mp->hfrag.off_heap);
 	    }
-	    *next_p = MSO(proc).first;
-	    MSO(proc).first = bp->off_heap.first;
-	    bp->off_heap.first = NULL;
-	    OH_OVERHEAD(&(MSO(proc)), bp->off_heap.overhead);
+	    if (bp)
+		free_message_buffer(bp);
 	}
+	fmp = mp;
+	mp = mp->next;
+	erts_free_message(fmp);
     }
 }
 
-Eterm
-erts_msg_distext2heap(Process *pp,
-		      ErtsProcLocks *plcksp,
-		      ErlHeapFragment **bpp,
-		      Eterm *tokenp,
-		      ErtsDistExternal *dist_extp)
+ErtsMessage *
+erts_realloc_shrink_message(ErtsMessage *mp, Uint sz, Eterm *brefs, Uint brefs_size)
 {
-    Eterm msg;
-    Uint tok_sz = 0;
-    Eterm *hp = NULL;
-    ErtsHeapFactory factory;
-    Sint sz;
-
-    *bpp = NULL;
-    sz = erts_decode_dist_ext_size(dist_extp);
-    if (sz < 0)
-	goto decode_error;
-    if (is_not_nil(*tokenp)) {
-	ErlHeapFragment *heap_frag = erts_dist_ext_trailer(dist_extp);
-	tok_sz = heap_frag->used_size;
-	sz += tok_sz;
-    }
-    if (pp) {
-	ErlOffHeap *ohp;
-	hp = erts_alloc_message_heap(sz, bpp, &ohp, pp, plcksp);
-    }
-    else {
-	*bpp = new_message_buffer(sz);
-	hp = (*bpp)->mem;
-    }
-    erts_factory_message_init(&factory, pp, hp, *bpp);
-    msg = erts_decode_dist_ext(&factory, dist_extp);
-    if (is_non_value(msg))
-	goto decode_error;
-    if (is_not_nil(*tokenp)) {
-	ErlHeapFragment *heap_frag = erts_dist_ext_trailer(dist_extp);
-	hp = erts_produce_heap(&factory, tok_sz, 0);
-	*tokenp = copy_struct(*tokenp, tok_sz, &hp, factory.off_heap);
-	erts_cleanup_offheap(&heap_frag->off_heap);
+    ErtsMessage *nmp = erts_realloc(ERTS_ALC_T_MSG, mp,
+				    sizeof(ErtsMessage) + (sz - 1)*sizeof(Eterm));
+    if (nmp != mp) {
+	Eterm *sp = &mp->hfrag.mem[0];
+	Eterm *ep = sp + sz;
+	Sint offs = &nmp->hfrag.mem[0] - sp;
+	erts_offset_off_heap(&nmp->hfrag.off_heap, offs, sp, ep);
+	erts_offset_heap(&nmp->hfrag.mem[0], sz, offs, sp, ep);
+	if (brefs && brefs_size)
+	    erts_offset_heap_ptr(brefs, brefs_size, offs, sp, ep);
     }
-    erts_free_dist_ext_copy(dist_extp);
-    erts_factory_close(&factory);
-    return msg;
 
- decode_error:
-    if (is_not_nil(*tokenp)) {
-	ErlHeapFragment *heap_frag = erts_dist_ext_trailer(dist_extp);
-	erts_cleanup_offheap(&heap_frag->off_heap);
-    }
-    erts_free_dist_ext_copy(dist_extp);
-    *bpp = NULL;
-    return THE_NON_VALUE;
- }
+    nmp->hfrag.used_size = sz;
+    nmp->hfrag.alloc_size = sz;
+
+    return nmp;
+}
 
 void
 erts_queue_dist_message(Process *rcvr,
@@ -287,7 +255,7 @@ erts_queue_dist_message(Process *rcvr,
 			ErtsDistExternal *dist_ext,
 			Eterm token)
 {
-    ErlMessage* mp;
+    ErtsMessage* mp;
 #ifdef USE_VM_PROBES
     Sint tok_label = 0;
     Sint tok_lastcnt = 0;
@@ -299,7 +267,17 @@ erts_queue_dist_message(Process *rcvr,
 
     ERTS_SMP_LC_ASSERT(*rcvr_locks == erts_proc_lc_my_proc_locks(rcvr));
 
-    mp = message_alloc();
+    mp = erts_alloc_message(0, NULL);
+    mp->data.dist_ext = dist_ext;
+
+    ERL_MESSAGE_TERM(mp) = THE_NON_VALUE;
+#ifdef USE_VM_PROBES
+    ERL_MESSAGE_DT_UTAG(mp) = NIL;
+    if (token == am_have_dt_utag)
+	ERL_MESSAGE_TOKEN(mp) = NIL;
+    else
+#endif
+	ERL_MESSAGE_TOKEN(mp) = token;
 
 #ifdef ERTS_SMP
     if (!(*rcvr_locks & ERTS_PROC_LOCK_MSGQ)) {
@@ -318,64 +296,46 @@ erts_queue_dist_message(Process *rcvr,
 	if (!(*rcvr_locks & ERTS_PROC_LOCK_MSGQ))
 	    erts_smp_proc_unlock(rcvr, ERTS_PROC_LOCK_MSGQ);
 	/* Drop message if receiver is exiting or has a pending exit ... */
-	if (is_not_nil(token)) {
-	    ErlHeapFragment *heap_frag;
-	    heap_frag = erts_dist_ext_trailer(mp->data.dist_ext);
-	    erts_cleanup_offheap(&heap_frag->off_heap);
-	}
-	erts_free_dist_ext_copy(dist_ext);
-	message_free(mp);
+	erts_cleanup_messages(mp);
     }
     else
 #endif
     if (IS_TRACED_FL(rcvr, F_TRACE_RECEIVE)) {
 	/* Ahh... need to decode it in order to trace it... */
-	ErlHeapFragment *mbuf;
-	Eterm msg;
 	if (!(*rcvr_locks & ERTS_PROC_LOCK_MSGQ))
 	    erts_smp_proc_unlock(rcvr, ERTS_PROC_LOCK_MSGQ);
-	message_free(mp);
-	msg = erts_msg_distext2heap(rcvr, rcvr_locks, &mbuf, &token, dist_ext);
-	if (is_value(msg))
+	if (!erts_decode_dist_message(rcvr, *rcvr_locks, mp, 0))
+	    erts_free_message(mp);
+	else {
+	    Eterm msg = ERL_MESSAGE_TERM(mp);
+	    token = ERL_MESSAGE_TOKEN(mp);
 #ifdef USE_VM_PROBES
-            if (DTRACE_ENABLED(message_queued)) {
-                DTRACE_CHARBUF(receiver_name, DTRACE_TERM_BUF_SIZE);
-
-                dtrace_proc_str(rcvr, receiver_name);
-                if (token != NIL && token != am_have_dt_utag) {
-                    tok_label = signed_val(SEQ_TRACE_T_LABEL(token));
-                    tok_lastcnt = signed_val(SEQ_TRACE_T_LASTCNT(token));
-                    tok_serial = signed_val(SEQ_TRACE_T_SERIAL(token));
-                }
-                DTRACE6(message_queued,
-                        receiver_name, size_object(msg), rcvr->msg.len,
-                        tok_label, tok_lastcnt, tok_serial);
-            }
+	    if (DTRACE_ENABLED(message_queued)) {
+		DTRACE_CHARBUF(receiver_name, DTRACE_TERM_BUF_SIZE);
+
+		dtrace_proc_str(rcvr, receiver_name);
+                if (have_seqtrace(token)) {
+		    tok_label = signed_val(SEQ_TRACE_T_LABEL(token));
+		    tok_lastcnt = signed_val(SEQ_TRACE_T_LASTCNT(token));
+		    tok_serial = signed_val(SEQ_TRACE_T_SERIAL(token));
+		}
+		DTRACE6(message_queued,
+			receiver_name, size_object(msg), rcvr->msg.len,
+			tok_label, tok_lastcnt, tok_serial);
+	    }
 #endif
-	    erts_queue_message(rcvr, rcvr_locks, mbuf, msg, token);
+	    erts_queue_message(rcvr, rcvr_locks, mp, msg, token);
+	}
     }
     else {
 	/* Enqueue message on external format */
 
-	ERL_MESSAGE_TERM(mp) = THE_NON_VALUE;
-#ifdef USE_VM_PROBES
-	ERL_MESSAGE_DT_UTAG(mp) = NIL;
-	if (token == am_have_dt_utag) {
-	    ERL_MESSAGE_TOKEN(mp) = NIL;
-	} else {
-#endif
-	    ERL_MESSAGE_TOKEN(mp) = token;
-#ifdef USE_VM_PROBES
-	}
-#endif
-	mp->next = NULL;
-
 #ifdef USE_VM_PROBES
         if (DTRACE_ENABLED(message_queued)) {
             DTRACE_CHARBUF(receiver_name, DTRACE_TERM_BUF_SIZE);
 
             dtrace_proc_str(rcvr, receiver_name);
-            if (token != NIL && token != am_have_dt_utag) {
+            if (have_seqtrace(token)) {
                 tok_label = signed_val(SEQ_TRACE_T_LABEL(token));
                 tok_lastcnt = signed_val(SEQ_TRACE_T_LASTCNT(token));
                 tok_serial = signed_val(SEQ_TRACE_T_SERIAL(token));
@@ -388,7 +348,7 @@ erts_queue_dist_message(Process *rcvr,
                     tok_label, tok_lastcnt, tok_serial);
         }
 #endif
-	mp->data.dist_ext = dist_ext;
+
 	LINK_MESSAGE(rcvr, mp);
 
 	if (!(*rcvr_locks & ERTS_PROC_LOCK_MSGQ))
@@ -408,9 +368,9 @@ erts_queue_dist_message(Process *rcvr,
 static Sint
 queue_message(Process *c_p,
 	      Process* receiver,
-	      ErtsProcLocks *receiver_locks,
 	      erts_aint32_t *receiver_state,
-	      ErlHeapFragment* bp,
+	      ErtsProcLocks *receiver_locks,
+	      ErtsMessage* mp,
 	      Eterm message,
 	      Eterm seq_trace_token
 #ifdef USE_VM_PROBES
@@ -419,31 +379,24 @@ queue_message(Process *c_p,
     )
 {
     Sint res;
-    ErlMessage* mp;
     int locked_msgq = 0;
-    erts_aint_t state;
-
-#ifndef ERTS_SMP
-    ASSERT(bp != NULL || receiver->mbuf == NULL);
-#endif
+    erts_aint32_t state;
 
     ERTS_SMP_LC_ASSERT(*receiver_locks == erts_proc_lc_my_proc_locks(receiver));
 
-    mp = message_alloc();
-
-    if (receiver_state)
-	state = *receiver_state;
-    else
-	state = erts_smp_atomic32_read_acqb(&receiver->state);
-
 #ifdef ERTS_SMP
 
-    if (state & (ERTS_PSFLG_EXITING|ERTS_PSFLG_PENDING_EXIT))
-	goto exiting;
-
     if (!(*receiver_locks & ERTS_PROC_LOCK_MSGQ)) {
 	if (erts_smp_proc_trylock(receiver, ERTS_PROC_LOCK_MSGQ) == EBUSY) {
 	    ErtsProcLocks need_locks = ERTS_PROC_LOCK_MSGQ;
+
+	    if (receiver_state)
+		state = *receiver_state;
+	    else
+		state = erts_smp_atomic32_read_nob(&receiver->state);
+	    if (state & (ERTS_PSFLG_EXITING|ERTS_PSFLG_PENDING_EXIT))
+		goto exiting;
+
 	    if (*receiver_locks & ERTS_PROC_LOCK_STATUS) {
 		erts_smp_proc_unlock(receiver, ERTS_PROC_LOCK_STATUS);
 		need_locks |= ERTS_PROC_LOCK_STATUS;
@@ -451,13 +404,12 @@ queue_message(Process *c_p,
 	    erts_smp_proc_lock(receiver, need_locks);
 	}
 	locked_msgq = 1;
-	state = erts_smp_atomic32_read_nob(&receiver->state);
-	if (receiver_state)
-	    *receiver_state = state;
     }
 
 #endif
 
+    state = erts_smp_atomic32_read_nob(&receiver->state);
+
     if (state & (ERTS_PSFLG_PENDING_EXIT|ERTS_PSFLG_EXITING)) {
 #ifdef ERTS_SMP
     exiting:
@@ -465,9 +417,7 @@ queue_message(Process *c_p,
 	/* Drop message if receiver is exiting or has a pending exit... */
 	if (locked_msgq)
 	    erts_smp_proc_unlock(receiver, ERTS_PROC_LOCK_MSGQ);
-	if (bp)
-	    free_message_buffer(bp);
-	message_free(mp);
+	erts_cleanup_messages(mp);
 	return 0;
     }
 
@@ -476,13 +426,9 @@ queue_message(Process *c_p,
 #ifdef USE_VM_PROBES
     ERL_MESSAGE_DT_UTAG(mp) = dt_utag;
 #endif
-    mp->next = NULL;
-    mp->data.heap_frag = bp;
 
-#ifndef ERTS_SMP
     res = receiver->msg.len;
-#else
-    res = receiver->msg_inq.len;
+#ifdef ERTS_SMP
     if (*receiver_locks & ERTS_PROC_LOCK_MAIN) {
 	/*
 	 * We move 'in queue' to 'private queue' and place
@@ -492,7 +438,7 @@ queue_message(Process *c_p,
 	 * we don't need to include the 'in queue' in
 	 * the root set when garbage collecting.
 	 */
-	res += receiver->msg.len;
+	res += receiver->msg_inq.len;
 	ERTS_SMP_MSGQ_MV_INQ2PRIVQ(receiver);
 	LINK_MESSAGE_PRIVQ(receiver, mp);
     }
@@ -544,19 +490,19 @@ queue_message(Process *c_p,
 void
 #ifdef USE_VM_PROBES
 erts_queue_message_probe(Process* receiver, ErtsProcLocks *receiver_locks,
-                         ErlHeapFragment* bp,
+                         ErtsMessage* mp,
                          Eterm message, Eterm seq_trace_token, Eterm dt_utag)
 #else
 erts_queue_message(Process* receiver, ErtsProcLocks *receiver_locks,
-                   ErlHeapFragment* bp,
+                   ErtsMessage* mp,
                    Eterm message, Eterm seq_trace_token)
 #endif
 {
     queue_message(NULL,
 		  receiver,
-		  receiver_locks,
 		  NULL,
-		  bp,
+		  receiver_locks,
+		  mp,
 		  message,
 		  seq_trace_token
 #ifdef USE_VM_PROBES
@@ -566,245 +512,37 @@ erts_queue_message(Process* receiver, ErtsProcLocks *receiver_locks,
 }
 
 void
-erts_link_mbuf_to_proc(struct process *proc, ErlHeapFragment *bp)
-{
-    Eterm* htop = HEAP_TOP(proc);
-
-    link_mbuf_to_proc(proc, bp);
-    if (htop < HEAP_LIMIT(proc)) {
-	*htop = make_pos_bignum_header(HEAP_LIMIT(proc)-htop-1);
-	HEAP_TOP(proc) = HEAP_LIMIT(proc);
-    }
-}
-
-/*
- * Moves content of message buffer attached to a message into a heap.
- * The message buffer is deallocated.
- */
-void
-erts_move_msg_mbuf_to_heap(Eterm** hpp, ErlOffHeap* off_heap, ErlMessage *msg)
+erts_link_mbuf_to_proc(Process *proc, ErlHeapFragment *first_bp)
 {
-    struct erl_off_heap_header* oh;
-    Eterm term, token, *fhp, *hp;
-    Sint offs;
-    Uint sz;
-    ErlHeapFragment *bp;
-#ifdef USE_VM_PROBES
-    Eterm utag;
-#endif
-
-#ifdef HARD_DEBUG
-    struct erl_off_heap_header* dbg_oh_start = off_heap->first;
-    Eterm dbg_term, dbg_token;
-    ErlHeapFragment *dbg_bp;
-    Uint *dbg_hp, *dbg_thp_start;
-    Uint dbg_term_sz, dbg_token_sz;
-#ifdef USE_VM_PROBES
-    Eterm dbg_utag;
-    Uint dbg_utag_sz;
-#endif
-#endif
-
-    bp = msg->data.heap_frag;
-    term = ERL_MESSAGE_TERM(msg);
-    token = ERL_MESSAGE_TOKEN(msg);
-#ifdef USE_VM_PROBES
-    utag = ERL_MESSAGE_DT_UTAG(msg);
-#endif
-    if (!bp) {
-#ifdef USE_VM_PROBES
-	ASSERT(is_immed(term) && is_immed(token) && is_immed(utag));
-#else
-	ASSERT(is_immed(term) && is_immed(token));
-#endif
-	return;
-    }
-
-#ifdef HARD_DEBUG
-    dbg_term_sz = size_object(term);
-    dbg_token_sz = size_object(token);
-    dbg_bp = new_message_buffer(dbg_term_sz + dbg_token_sz);
-#ifdef USE_VM_PROBES
-    dbg_utag_sz = size_object(utag);
-    dbg_bp = new_message_buffer(dbg_term_sz + dbg_token_sz + dbg_utag_sz );
-#endif
-    /*ASSERT(dbg_term_sz + dbg_token_sz == erts_msg_used_frag_sz(msg));
-      Copied size may be smaller due to removed SubBins's or garbage.
-      Copied size may be larger due to duplicated shared terms.
-    */
-    dbg_hp = dbg_bp->mem;
-    dbg_term = copy_struct(term, dbg_term_sz, &dbg_hp, &dbg_bp->off_heap);
-    dbg_token = copy_struct(token, dbg_token_sz, &dbg_hp, &dbg_bp->off_heap);
-#ifdef USE_VM_PROBES
-    dbg_utag = copy_struct(utag, dbg_utag_sz, &dbg_hp, &dbg_bp->off_heap);
-#endif
-   dbg_thp_start = *hpp;
-#endif
-
-    if (bp->next != NULL) {
-	move_multi_frags(hpp, off_heap, bp, msg->m, 
-#ifdef USE_VM_PROBES
-			 3
-#else
-			 2
-#endif
-			 );
-	goto copy_done;
-    }
-
-    OH_OVERHEAD(off_heap, bp->off_heap.overhead);
-    sz = bp->used_size;
-
-    ASSERT(is_immed(term) || in_heapfrag(ptr_val(term),bp));
-    ASSERT(is_immed(token) || in_heapfrag(ptr_val(token),bp));
-
-    fhp = bp->mem;
-    hp = *hpp;
-    offs = hp - fhp;
-
-    oh = NULL;
-    while (sz--) {
-	Uint cpy_sz;
-	Eterm val = *fhp++;
-
-	switch (primary_tag(val)) {
-	case TAG_PRIMARY_IMMED1:
-	    *hp++ = val;
-	    break;
-	case TAG_PRIMARY_LIST:
-	case TAG_PRIMARY_BOXED:
-	    ASSERT(in_heapfrag(ptr_val(val), bp));
-	    *hp++ = offset_ptr(val, offs);
-	    break;
-	case TAG_PRIMARY_HEADER:
-	    *hp++ = val;
-	    switch (val & _HEADER_SUBTAG_MASK) {
-	    case ARITYVAL_SUBTAG:
-		break;
-	    case REFC_BINARY_SUBTAG:
-	    case FUN_SUBTAG:
-	    case EXTERNAL_PID_SUBTAG:
-	    case EXTERNAL_PORT_SUBTAG:
-	    case EXTERNAL_REF_SUBTAG:
-		oh = (struct erl_off_heap_header*) (hp-1);
-		cpy_sz = thing_arityval(val);
-		goto cpy_words;
-	    default:
-		cpy_sz = header_arity(val);
-
-	    cpy_words:
-		ASSERT(sz >= cpy_sz);
-		sz -= cpy_sz;
-		while (cpy_sz >= 8) {
-		    cpy_sz -= 8;
-		    *hp++ = *fhp++;
-		    *hp++ = *fhp++;
-		    *hp++ = *fhp++;
-		    *hp++ = *fhp++;
-		    *hp++ = *fhp++;
-		    *hp++ = *fhp++;
-		    *hp++ = *fhp++;
-		    *hp++ = *fhp++;
-		}
-		switch (cpy_sz) {
-		case 7: *hp++ = *fhp++;
-		case 6: *hp++ = *fhp++;
-		case 5: *hp++ = *fhp++;
-		case 4: *hp++ = *fhp++;
-		case 3: *hp++ = *fhp++;
-		case 2: *hp++ = *fhp++;
-		case 1: *hp++ = *fhp++;
-		default: break;
+    if (first_bp) {
+	ErlHeapFragment *bp = first_bp;
+
+	while (1) {
+	    /* Move any off_heap's into the process */
+	    if (bp->off_heap.first != NULL) {
+		struct erl_off_heap_header** next_p = &bp->off_heap.first;
+		while (*next_p != NULL) {
+		    next_p = &((*next_p)->next);
 		}
-		if (oh) {
-		    /* Add to offheap list */
-		    oh->next = off_heap->first;
-		    off_heap->first = oh;
-		    ASSERT(*hpp <= (Eterm*)oh);
-		    ASSERT(hp > (Eterm*)oh);
-		    oh = NULL;
-		}
-		break;
+		*next_p = MSO(proc).first;
+		MSO(proc).first = bp->off_heap.first;
+		bp->off_heap.first = NULL;
+		OH_OVERHEAD(&(MSO(proc)), bp->off_heap.overhead);
 	    }
-	    break;
+	    MBUF_SIZE(proc) += bp->used_size;
+	    if (!bp->next)
+		break;
+	    bp = bp->next;
 	}
-    }
-
-    ASSERT(bp->used_size == hp - *hpp);
-    *hpp = hp;
-
-    if (is_not_immed(token)) {
-	ASSERT(in_heapfrag(ptr_val(token), bp));
-	ERL_MESSAGE_TOKEN(msg) = offset_ptr(token, offs);
-#ifdef HARD_DEBUG
-	ASSERT(dbg_thp_start <= ptr_val(ERL_MESSAGE_TOKEN(msg)));
-	ASSERT(hp > ptr_val(ERL_MESSAGE_TOKEN(msg)));
-#endif
-    }
-
-    if (is_not_immed(term)) {
-	ASSERT(in_heapfrag(ptr_val(term),bp));
-	ERL_MESSAGE_TERM(msg) = offset_ptr(term, offs);
-#ifdef HARD_DEBUG
-	ASSERT(dbg_thp_start <= ptr_val(ERL_MESSAGE_TERM(msg)));
-	ASSERT(hp > ptr_val(ERL_MESSAGE_TERM(msg)));
-#endif
-    }
-#ifdef USE_VM_PROBES
-    if (is_not_immed(utag)) {
-	ASSERT(in_heapfrag(ptr_val(utag), bp));
-	ERL_MESSAGE_DT_UTAG(msg) = offset_ptr(utag, offs);
-#ifdef HARD_DEBUG
-	ASSERT(dbg_thp_start <= ptr_val(ERL_MESSAGE_DT_UTAG(msg)));
-	ASSERT(hp > ptr_val(ERL_MESSAGE_DT_UTAG(msg)));
-#endif
-    }
-#endif
-
-copy_done:
 
-#ifdef HARD_DEBUG
-    {
-	int i, j;
-	ErlHeapFragment* frag;
-	{
-	    struct erl_off_heap_header* dbg_oh = off_heap->first;
-	    i = j = 0;
-	    while (dbg_oh != dbg_oh_start) {
-		dbg_oh = dbg_oh->next;
-		i++;
-	    }
-	    for (frag=bp; frag; frag=frag->next) {
-		dbg_oh = frag->off_heap.first;
-		while (dbg_oh) {
-		    dbg_oh = dbg_oh->next;
-		    j++;
-		}
-	    }
-	    ASSERT(i == j);
-	}
+	/* Link the message buffer */
+	bp->next = MBUF(proc);
+	MBUF(proc) = first_bp;
     }
-#endif
-	    
-
-    bp->off_heap.first = NULL;
-    free_message_buffer(bp);
-    msg->data.heap_frag = NULL;
-
-#ifdef HARD_DEBUG
-    ASSERT(eq(ERL_MESSAGE_TERM(msg), dbg_term));
-    ASSERT(eq(ERL_MESSAGE_TOKEN(msg), dbg_token));
-#ifdef USE_VM_PROBES
-    ASSERT(eq(ERL_MESSAGE_DT_UTAG(msg), dbg_utag));
-#endif
-    free_message_buffer(dbg_bp);
-#endif
-
 }
 
-
 Uint
-erts_msg_attached_data_size_aux(ErlMessage *msg)
+erts_msg_attached_data_size_aux(ErtsMessage *msg)
 {
     Sint sz;
     ASSERT(is_non_value(ERL_MESSAGE_TERM(msg)));
@@ -833,29 +571,87 @@ erts_msg_attached_data_size_aux(ErlMessage *msg)
     return sz;
 }
 
-void
-erts_move_msg_attached_data_to_heap(ErtsHeapFactory* factory,
-				    ErlMessage *msg)
+ErtsMessage *
+erts_try_alloc_message_on_heap(Process *pp,
+			       erts_aint32_t *psp,
+			       ErtsProcLocks *plp,
+			       Uint sz,
+			       Eterm **hpp,
+			       ErlOffHeap **ohpp,
+			       int *on_heap_p)
 {
-    if (is_value(ERL_MESSAGE_TERM(msg)))
-	erts_move_msg_mbuf_to_heap(&factory->hp, factory->off_heap, msg);
-    else if (msg->data.dist_ext) {
-	ASSERT(msg->data.dist_ext->heap_size >= 0);
-	if (is_not_nil(ERL_MESSAGE_TOKEN(msg))) {
-	    ErlHeapFragment *heap_frag;
-	    heap_frag = erts_dist_ext_trailer(msg->data.dist_ext);
-	    ERL_MESSAGE_TOKEN(msg) = copy_struct(ERL_MESSAGE_TOKEN(msg),
-						 heap_frag->used_size,
-						 &factory->hp,
-						 factory->off_heap);
-	    erts_cleanup_offheap(&heap_frag->off_heap);
+#ifdef ERTS_SMP
+    int locked_main = 0;
+#endif
+    ErtsMessage *mp;
+
+    ASSERT(!(*psp & ERTS_PSFLG_OFF_HEAP_MSGQ));
+
+    if (
+#if defined(ERTS_SMP)
+	*plp & ERTS_PROC_LOCK_MAIN
+#else
+	1
+#endif
+	) {
+#ifdef ERTS_SMP
+    try_on_heap:
+#endif
+	if ((*psp & (ERTS_PSFLG_EXITING|ERTS_PSFLG_PENDING_EXIT))
+	    || (pp->flags & F_DISABLE_GC)
+	    || HEAP_LIMIT(pp) - HEAP_TOP(pp) <= sz) {
+	    /*
+	     * The heap is either potentially in an inconsistent
+	     * state, or not large enough.
+	     */
+#ifdef ERTS_SMP
+	    if (locked_main) {
+		*plp &= ~ERTS_PROC_LOCK_MAIN;
+		erts_smp_proc_unlock(pp, ERTS_PROC_LOCK_MAIN);
+	    }
+#endif
+	    goto in_message_fragment;
 	}
-	ERL_MESSAGE_TERM(msg) = erts_decode_dist_ext(factory,
-						     msg->data.dist_ext);
-	erts_free_dist_ext_copy(msg->data.dist_ext);
-	msg->data.dist_ext = NULL;
+
+	*hpp = HEAP_TOP(pp);
+	HEAP_TOP(pp) = *hpp + sz;
+	*ohpp = &MSO(pp);
+	mp = erts_alloc_message(0, NULL);
+	mp->data.attached = NULL;
+	*on_heap_p = !0;
+    }
+#ifdef ERTS_SMP
+    else if (erts_smp_proc_trylock(pp, ERTS_PROC_LOCK_MAIN) == 0) {
+	locked_main = 1;
+	*psp = erts_smp_atomic32_read_nob(&pp->state);
+	*plp |= ERTS_PROC_LOCK_MAIN;
+	goto try_on_heap;
     }
-    /* else: bad external detected when calculating size */
+#endif
+    else {
+    in_message_fragment:
+	if (!((*psp) & ERTS_PSFLG_ON_HEAP_MSGQ)) {
+	    mp = erts_alloc_message(sz, hpp);
+	    *ohpp = sz == 0 ? NULL : &mp->hfrag.off_heap;
+	}
+	else {
+	    mp = erts_alloc_message(0, NULL);
+	    if (!sz) {
+		*hpp = NULL;
+		*ohpp = NULL;
+	    }
+	    else {
+		ErlHeapFragment *bp;
+		bp = new_message_buffer(sz);
+		*hpp = &bp->mem[0];
+		mp->data.heap_frag = bp;
+		*ohpp = &bp->off_heap;
+	    }
+	}
+	*on_heap_p = 0;
+    }
+
+    return mp;
 }
 
 /*
@@ -870,7 +666,8 @@ erts_send_message(Process* sender,
 		  unsigned flags)
 {
     Uint msize;
-    ErlHeapFragment* bp = NULL;
+    ErtsMessage* mp;
+    ErlOffHeap *ohp;
     Eterm token = NIL;
     Sint res = 0;
 #ifdef USE_VM_PROBES
@@ -879,67 +676,96 @@ erts_send_message(Process* sender,
     Sint tok_label = 0;
     Sint tok_lastcnt = 0;
     Sint tok_serial = 0;
+    Eterm utag = NIL;
+#endif
+    erts_aint32_t receiver_state;
+#ifdef SHCOPY_SEND
+    erts_shcopy_t info;
 #endif
     BM_STOP_TIMER(system);
     BM_MESSAGE(message,sender,receiver);
     BM_START_TIMER(send);
 
- #ifdef USE_VM_PROBES
+#ifdef USE_VM_PROBES
     *sender_name = *receiver_name = '\0';
-   if (DTRACE_ENABLED(message_send)) {
+    if (DTRACE_ENABLED(message_send)) {
         erts_snprintf(sender_name, sizeof(DTRACE_CHARBUF_NAME(sender_name)),
 		      "%T", sender->common.id);
         erts_snprintf(receiver_name, sizeof(DTRACE_CHARBUF_NAME(receiver_name)),
 		      "%T", receiver->common.id);
     }
 #endif
+
+    receiver_state = erts_smp_atomic32_read_nob(&receiver->state);
+
     if (SEQ_TRACE_TOKEN(sender) != NIL && !(flags & ERTS_SND_FLG_NO_SEQ_TRACE)) {
         Eterm* hp;
 	Eterm stoken = SEQ_TRACE_TOKEN(sender);
 	Uint seq_trace_size = 0;
 #ifdef USE_VM_PROBES
 	Uint dt_utag_size = 0;
-	Eterm utag = NIL;
-#endif
-
-	BM_SWAP_TIMER(send,size);
-	msize = size_object(message);
-	BM_SWAP_TIMER(size,send);
-
-#ifdef USE_VM_PROBES
-	if (stoken != am_have_dt_utag) {
 #endif
-
+        BM_SWAP_TIMER(send,size);
+
+        /* SHCOPY corrupts the heap between
+         * copy_shared_calculate, and
+         * copy_shared_perform. (it inserts move_markers like the gc).
+         * Make sure we don't use the heap between those instances.
+         */
+        if (have_seqtrace(stoken)) {
 	    seq_trace_update_send(sender);
 	    seq_trace_output(stoken, message, SEQ_TRACE_SEND, 
 			     receiver->common.id, sender);
 	    seq_trace_size = 6; /* TUPLE5 */
-#ifdef USE_VM_PROBES
-	}
-	if (DT_UTAG_FLAGS(sender) & DT_UTAG_SPREADING) {
-	    dt_utag_size = size_object(DT_UTAG(sender));
-	} else if (stoken == am_have_dt_utag ) {
-	    stoken = NIL;
 	}
+#ifdef USE_VM_PROBES
+        if (DT_UTAG_FLAGS(sender) & DT_UTAG_SPREADING) {
+            dt_utag_size = size_object(DT_UTAG(sender));
+        } else if (stoken == am_have_dt_utag ) {
+            stoken = NIL;
+        }
 #endif
 
-	bp = new_message_buffer(msize + seq_trace_size 
+#ifdef SHCOPY_SEND
+        INITIALIZE_SHCOPY(info);
+        msize = copy_shared_calculate(message, &info);
+#else
+        msize = size_object(message);
+#endif
+        BM_SWAP_TIMER(size,send);
+
+        mp = erts_alloc_message_heap_state(receiver,
+                                           &receiver_state,
+                                           receiver_locks,
+                                           (msize
 #ifdef USE_VM_PROBES
-				+ dt_utag_size
+                                            + dt_utag_size
 #endif
-				);
-	hp = bp->mem;
+                                            + seq_trace_size),
+                                           &hp,
+                                           &ohp);
 
         BM_SWAP_TIMER(send,copy);
-	token = copy_struct(stoken,
-			    seq_trace_size,
-			    &hp,
-			    &bp->off_heap);
 
-	message = copy_struct(message, msize, &hp, &bp->off_heap);
+#ifdef SHCOPY_SEND
+	if (is_not_immed(message))
+            message = copy_shared_perform(message, msize, &info, &hp, ohp);
+        DESTROY_SHCOPY(info);
+#else
+	if (is_not_immed(message))
+            message = copy_struct(message, msize, &hp, ohp);
+#endif
+	if (is_immed(stoken))
+	    token = stoken;
+	else
+	    token = copy_struct(stoken, seq_trace_size, &hp, ohp);
+
 #ifdef USE_VM_PROBES
 	if (DT_UTAG_FLAGS(sender) & DT_UTAG_SPREADING) {
-	    utag = copy_struct(DT_UTAG(sender), dt_utag_size, &hp, &bp->off_heap);
+	    if (is_immed(DT_UTAG(sender)))
+		utag = DT_UTAG(sender);
+	    else
+		utag = copy_struct(DT_UTAG(sender), dt_utag_size, &hp, ohp);
 #ifdef DTRACE_TAG_HARDDEBUG
 	    erts_fprintf(stderr,
 			 "Dtrace -> (%T) Spreading tag (%T) with "
@@ -952,7 +778,7 @@ erts_send_message(Process* sender,
 
 #ifdef USE_VM_PROBES
         if (DTRACE_ENABLED(message_send)) {
-	    if (stoken != NIL && stoken != am_have_dt_utag) {
+            if (have_seqtrace(stoken)) {
 		tok_label = signed_val(SEQ_TRACE_T_LABEL(stoken));
 		tok_lastcnt = signed_val(SEQ_TRACE_T_LASTCNT(stoken));
 		tok_serial = signed_val(SEQ_TRACE_T_SERIAL(stoken));
@@ -961,103 +787,63 @@ erts_send_message(Process* sender,
 		    msize, tok_label, tok_lastcnt, tok_serial);
         }
 #endif
-        res = queue_message(NULL,
-			    receiver,
-			    receiver_locks,
-			    NULL,
-			    bp,
-			    message,
-			    token
-#ifdef USE_VM_PROBES
-			    , utag
-#endif
-	    );
-        BM_SWAP_TIMER(send,system);
-    } else if (sender == receiver) {
-	/* Drop message if receiver has a pending exit ... */
-#ifdef ERTS_SMP
-	ErtsProcLocks need_locks = (~(*receiver_locks)
-				    & (ERTS_PROC_LOCK_MSGQ
-				       | ERTS_PROC_LOCK_STATUS));
-	if (need_locks) {
-	    *receiver_locks |= need_locks;
-	    if (erts_smp_proc_trylock(receiver, need_locks) == EBUSY) {
-		if (need_locks == ERTS_PROC_LOCK_MSGQ) {
-		    erts_smp_proc_unlock(receiver, ERTS_PROC_LOCK_STATUS);
-		    need_locks = ERTS_PROC_LOCK_MSGQ|ERTS_PROC_LOCK_STATUS;
-		}
-		erts_smp_proc_lock(receiver, need_locks);
-	    }
+    } else {
+        Eterm *hp;
+
+	if (receiver == sender && !(receiver_state & ERTS_PSFLG_OFF_HEAP_MSGQ)) {
+	    mp = erts_alloc_message(0, NULL);
+	    msize = 0;
 	}
-	if (!ERTS_PROC_PENDING_EXIT(receiver))
+	else {
+	    BM_SWAP_TIMER(send,size);
+#ifdef SHCOPY_SEND
+            INITIALIZE_SHCOPY(info);
+            msize = copy_shared_calculate(message, &info);
+#else
+            msize = size_object(message);
 #endif
-	{
-	    ErlMessage* mp = message_alloc();
-
-            DTRACE6(message_send, sender_name, receiver_name,
-                    size_object(message), tok_label, tok_lastcnt, tok_serial);
-	    mp->data.attached = NULL;
-	    ERL_MESSAGE_TERM(mp) = message;
-	    ERL_MESSAGE_TOKEN(mp) = NIL;
-#ifdef USE_VM_PROBES
-	    ERL_MESSAGE_DT_UTAG(mp) = NIL;
+	    BM_SWAP_TIMER(size,send);
+
+	    mp = erts_alloc_message_heap_state(receiver,
+					       &receiver_state,
+					       receiver_locks,
+					       msize,
+					       &hp,
+					       &ohp);
+	    BM_SWAP_TIMER(send,copy);
+#ifdef SHCOPY_SEND
+            if (is_not_immed(message))
+                message = copy_shared_perform(message, msize, &info, &hp, ohp);
+            DESTROY_SHCOPY(info);
+#else
+            if (is_not_immed(message))
+                message = copy_struct(message, msize, &hp, ohp);
 #endif
-	    mp->next = NULL;
-	    /*
-	     * We move 'in queue' to 'private queue' and place
-	     * message at the end of 'private queue' in order
-	     * to ensure that the 'in queue' doesn't contain
-	     * references into the heap. By ensuring this,
-	     * we don't need to include the 'in queue' in
-	     * the root set when garbage collecting.
-	     */
-	    
-	    ERTS_SMP_MSGQ_MV_INQ2PRIVQ(receiver);
-	    LINK_MESSAGE_PRIVQ(receiver, mp);
-
-	    res = receiver->msg.len;
-
-	    if (IS_TRACED_FL(receiver, F_TRACE_RECEIVE)) {
-		trace_receive(receiver, message);
-	    }
+	    BM_MESSAGE_COPIED(msz);
+	    BM_SWAP_TIMER(copy,send);
 	}
-        BM_SWAP_TIMER(send,system);
-    } else {
-	ErlOffHeap *ohp;
-        Eterm *hp;
-	erts_aint32_t state;
-
-	BM_SWAP_TIMER(send,size);
-	msize = size_object(message);
-	BM_SWAP_TIMER(size,send);
-	hp = erts_alloc_message_heap_state(msize,
-					   &bp,
-					   &ohp,
-					   receiver,
-					   receiver_locks,
-					   &state);
-	BM_SWAP_TIMER(send,copy);
-	message = copy_struct(message, msize, &hp, ohp);
-	BM_MESSAGE_COPIED(msz);
-	BM_SWAP_TIMER(copy,send);
         DTRACE6(message_send, sender_name, receiver_name,
                 msize, tok_label, tok_lastcnt, tok_serial);
-	res = queue_message(sender,
-			    receiver,
-			    receiver_locks,
-			    &state,
-			    bp,
-			    message,
-			    token
+    }
+
+    res = queue_message(sender,
+			receiver,
+			&receiver_state,
+			receiver_locks,
+			mp,
+			message,
+			token
 #ifdef USE_VM_PROBES
-			    , NIL
+			, utag
 #endif
-	    );
-        BM_SWAP_TIMER(send,system);
-    }
-   return res;
+	);
+
+    BM_SWAP_TIMER(send,system);
+    
+    return res;
 }
 
+
 /*
  * This function delivers an EXIT message to a process
  * which is trapping EXITs.
@@ -1075,46 +861,572 @@ erts_deliver_exit_message(Eterm from, Process *to, ErtsProcLocks *to_locksp,
     Uint sz_from;
     Eterm* hp;
     Eterm temptoken;
-    ErlHeapFragment* bp = NULL;
-
-    if (token != NIL
-#ifdef USE_VM_PROBES
-	&& token != am_have_dt_utag
+    ErtsMessage* mp;
+    ErlOffHeap *ohp;
+#ifdef SHCOPY_SEND
+    erts_shcopy_t info;
 #endif
-	) {
 
+    if (have_seqtrace(token)) {
 	ASSERT(is_tuple(token));
-	sz_reason = size_object(reason);
 	sz_token = size_object(token);
 	sz_from = size_object(from);
-	bp = new_message_buffer(sz_reason + sz_from + sz_token + 4);
-	hp = bp->mem;
-	mess = copy_struct(reason, sz_reason, &hp, &bp->off_heap);
-	from_copy = copy_struct(from, sz_from, &hp, &bp->off_heap);
+#ifdef SHCOPY_SEND
+        INITIALIZE_SHCOPY(info);
+        sz_reason = copy_shared_calculate(reason, &info);
+#else
+	sz_reason = size_object(reason);
+#endif
+	mp = erts_alloc_message_heap(to, to_locksp,
+				     sz_reason + sz_from + sz_token + 4,
+				     &hp, &ohp);
+#ifdef SHCOPY_SEND
+        mess = copy_shared_perform(reason, sz_reason, &info, &hp, ohp);
+        DESTROY_SHCOPY(info);
+#else
+	mess = copy_struct(reason, sz_reason, &hp, ohp);
+#endif
+	from_copy = copy_struct(from, sz_from, &hp, ohp);
 	save = TUPLE3(hp, am_EXIT, from_copy, mess);
 	hp += 4;
 	/* the trace token must in this case be updated by the caller */
 	seq_trace_output(token, save, SEQ_TRACE_SEND, to->common.id, NULL);
-	temptoken = copy_struct(token, sz_token, &hp, &bp->off_heap);
-	erts_queue_message(to, to_locksp, bp, save, temptoken);
+	temptoken = copy_struct(token, sz_token, &hp, ohp);
+	erts_queue_message(to, to_locksp, mp, save, temptoken);
     } else {
-	ErlOffHeap *ohp;
-	sz_reason = size_object(reason);
 	sz_from = IS_CONST(from) ? 0 : size_object(from);
+#ifdef SHCOPY_SEND
+        INITIALIZE_SHCOPY(info);
+        sz_reason = copy_shared_calculate(reason, &info);
+#else
+	sz_reason = size_object(reason);
+#endif
+	mp = erts_alloc_message_heap(to, to_locksp,
+				     sz_reason+sz_from+4, &hp, &ohp);
 
-	hp = erts_alloc_message_heap(sz_reason+sz_from+4,
-				     &bp,
-				     &ohp,
-				     to,
-				     to_locksp);
-
+#ifdef SHCOPY_SEND
+        mess = copy_shared_perform(reason, sz_reason, &info, &hp, ohp);
+        DESTROY_SHCOPY(info);
+#else
 	mess = copy_struct(reason, sz_reason, &hp, ohp);
+#endif
 	from_copy = (IS_CONST(from)
 		     ? from
 		     : copy_struct(from, sz_from, &hp, ohp));
 	save = TUPLE3(hp, am_EXIT, from_copy, mess);
-	erts_queue_message(to, to_locksp, bp, save, NIL);
+	erts_queue_message(to, to_locksp, mp, save, NIL);
+    }
+}
+
+void erts_save_message_in_proc(Process *p, ErtsMessage *msgp)
+{
+    ErlHeapFragment *hfp;
+
+    if (msgp->data.attached == ERTS_MSG_COMBINED_HFRAG)
+	hfp = &msgp->hfrag;
+    else if (msgp->data.attached) {
+	hfp = msgp->data.heap_frag;
+    }
+    else {
+	erts_free_message(msgp);
+	return; /* Nothing to save */
+    }
+
+    while (1) {
+	struct erl_off_heap_header *ohhp = hfp->off_heap.first;
+	if (ohhp) {
+	    for ( ; ohhp->next; ohhp = ohhp->next)
+		;
+	    ohhp->next = p->off_heap.first;
+	    p->off_heap.first = hfp->off_heap.first;
+	    hfp->off_heap.first = NULL;
+	}
+	p->off_heap.overhead += hfp->off_heap.overhead;
+	hfp->off_heap.overhead = 0;
+	p->mbuf_sz += hfp->used_size;
+
+	if (!hfp->next)
+	    break;
+	hfp = hfp->next;
+    }
+
+    msgp->next = p->msg_frag;
+    p->msg_frag = msgp;
+}
+
+Sint
+erts_move_messages_off_heap(Process *c_p)
+{
+    int reds = 1;
+    /*
+     * Move all messages off heap. This *only* occurs when the
+     * process had off heap message disabled and just enabled
+     * it...
+     */
+    ErtsMessage *mp;
+
+    reds += c_p->msg.len / 10;
+
+    ASSERT(erts_smp_atomic32_read_nob(&c_p->state)
+	   & ERTS_PSFLG_OFF_HEAP_MSGQ);
+    ASSERT(c_p->flags & F_OFF_HEAP_MSGQ_CHNG);
+
+    for (mp = c_p->msg.first; mp; mp = mp->next) {
+	Uint msg_sz, token_sz;
+#ifdef USE_VM_PROBES
+	Uint utag_sz;
+#endif
+	Eterm *hp;
+	ErlHeapFragment *hfrag;
+
+	if (mp->data.attached)
+	    continue;
+
+	if (is_immed(ERL_MESSAGE_TERM(mp))
+#ifdef USE_VM_PROBES
+	    && is_immed(ERL_MESSAGE_DT_UTAG(mp))
+#endif
+	    && is_not_immed(ERL_MESSAGE_TOKEN(mp)))
+	    continue;
+
+	/*
+	 * The message refers into the heap. Copy the message
+	 * from the heap into a heap fragment and attach
+	 * it to the message...
+	 */
+	msg_sz = size_object(ERL_MESSAGE_TERM(mp));
+#ifdef USE_VM_PROBES
+	utag_sz = size_object(ERL_MESSAGE_DT_UTAG(mp));
+#endif
+	token_sz = size_object(ERL_MESSAGE_TOKEN(mp));
+
+	hfrag = new_message_buffer(msg_sz
+#ifdef USE_VM_PROBES
+				   + utag_sz
+#endif
+				   + token_sz);
+	hp = hfrag->mem;
+	if (is_not_immed(ERL_MESSAGE_TERM(mp)))
+	    ERL_MESSAGE_TERM(mp) = copy_struct(ERL_MESSAGE_TERM(mp),
+					       msg_sz, &hp,
+					       &hfrag->off_heap);
+	if (is_not_immed(ERL_MESSAGE_TOKEN(mp)))
+	    ERL_MESSAGE_TOKEN(mp) = copy_struct(ERL_MESSAGE_TOKEN(mp),
+						token_sz, &hp,
+						&hfrag->off_heap);
+#ifdef USE_VM_PROBES
+	if (is_not_immed(ERL_MESSAGE_DT_UTAG(mp)))
+	    ERL_MESSAGE_DT_UTAG(mp) = copy_struct(ERL_MESSAGE_DT_UTAG(mp),
+						  utag_sz, &hp,
+						  &hfrag->off_heap);
+#endif
+	mp->data.heap_frag = hfrag;
+	reds += 1;
+    }
+
+    return reds;
+}
+
+Sint
+erts_complete_off_heap_message_queue_change(Process *c_p)
+{
+    int reds = 1;
+
+    ERTS_SMP_LC_ASSERT(ERTS_PROC_LOCK_MAIN == erts_proc_lc_my_proc_locks(c_p));
+    ASSERT(c_p->flags & F_OFF_HEAP_MSGQ_CHNG);
+    ASSERT(erts_smp_atomic32_read_nob(&c_p->state) & ERTS_PSFLG_OFF_HEAP_MSGQ);
+
+    /*
+     * This job was first initiated when the process changed to off heap
+     * message queue management. Since then ERTS_PSFLG_OFF_HEAP_MSGQ
+     * has been set. However, the management state might have been changed
+     * again (multiple times) since then. Check users last requested state
+     * (the flags F_OFF_HEAP_MSGQ, and F_ON_HEAP_MSGQ), and make the state
+     * consistent with that.
+     */
+
+    if (!(c_p->flags & F_OFF_HEAP_MSGQ))
+	erts_smp_atomic32_read_band_nob(&c_p->state,
+					~ERTS_PSFLG_OFF_HEAP_MSGQ);
+    else {
+	reds += 2;
+	erts_smp_proc_lock(c_p, ERTS_PROC_LOCK_MSGQ);
+	ERTS_SMP_MSGQ_MV_INQ2PRIVQ(c_p);
+	erts_smp_proc_unlock(c_p, ERTS_PROC_LOCK_MSGQ);
+	reds += erts_move_messages_off_heap(c_p);
+    }
+    c_p->flags &= ~F_OFF_HEAP_MSGQ_CHNG;
+    return reds;
+}
+
+typedef struct {
+    Eterm pid;
+    ErtsThrPrgrLaterOp lop;
+} ErtsChangeOffHeapMessageQueue;
+
+static void
+change_off_heap_msgq(void *vcohmq)
+{
+    ErtsChangeOffHeapMessageQueue *cohmq;
+    /*
+     * Now we've waited thread progress which ensures that all
+     * messages to the process are enqueued off heap. Schedule
+     * completion of this change as a system task on the process
+     * itself. This in order to avoid lock contention on its
+     * main lock. We will be called in
+     * erts_complete_off_heap_message_queue_change() (above) when
+     * the system task has been selected for execution.
+     */
+    cohmq = (ErtsChangeOffHeapMessageQueue *) vcohmq;
+    erts_schedule_complete_off_heap_message_queue_change(cohmq->pid);
+    erts_free(ERTS_ALC_T_MSGQ_CHNG, vcohmq);
+}
+
+Eterm
+erts_change_message_queue_management(Process *c_p, Eterm new_state)
+{
+    Eterm res;
+
+#ifdef DEBUG
+    if (c_p->flags & F_OFF_HEAP_MSGQ) {
+	ASSERT(erts_smp_atomic32_read_nob(&c_p->state)
+	       & ERTS_PSFLG_OFF_HEAP_MSGQ);
+    }
+    else {
+	if (c_p->flags & F_OFF_HEAP_MSGQ_CHNG) {
+	    ASSERT(erts_smp_atomic32_read_nob(&c_p->state)
+		   & ERTS_PSFLG_OFF_HEAP_MSGQ);
+	}
+	else {
+	    ASSERT(!(erts_smp_atomic32_read_nob(&c_p->state)
+		     & ERTS_PSFLG_OFF_HEAP_MSGQ));
+	}
+    }
+#endif
+
+    switch (c_p->flags & (F_OFF_HEAP_MSGQ|F_ON_HEAP_MSGQ)) {
+
+    case F_OFF_HEAP_MSGQ:
+	res = am_off_heap;
+
+	switch (new_state) {
+	case am_off_heap:
+	    break;
+	case am_on_heap:
+	    c_p->flags |= F_ON_HEAP_MSGQ;
+	    erts_smp_atomic32_read_bor_nob(&c_p->state,
+					   ERTS_PSFLG_ON_HEAP_MSGQ);
+	    /* fall through */
+	case am_mixed:
+	    c_p->flags &= ~F_OFF_HEAP_MSGQ;
+	    /*
+	     * We are not allowed to clear ERTS_PSFLG_OFF_HEAP_MSGQ
+	     * if a off heap change is ongoing. It will be adjusted
+	     * when the change completes...
+	     */
+	    if (!(c_p->flags & F_OFF_HEAP_MSGQ_CHNG)) {
+		/* Safe to clear ERTS_PSFLG_OFF_HEAP_MSGQ... */
+		erts_smp_atomic32_read_band_nob(&c_p->state,
+						~ERTS_PSFLG_OFF_HEAP_MSGQ);
+	    }
+	    break;
+	default:
+	    res = THE_NON_VALUE; /* badarg */
+	    break;
+	}
+	break;
+
+    case F_ON_HEAP_MSGQ:
+	res = am_on_heap;
+
+	switch (new_state) {
+	case am_on_heap:
+	    break;
+	case am_mixed:
+	    c_p->flags &= ~F_ON_HEAP_MSGQ;
+	    erts_smp_atomic32_read_band_nob(&c_p->state,
+					    ~ERTS_PSFLG_ON_HEAP_MSGQ);
+	    break;
+	case am_off_heap:
+	    c_p->flags &= ~F_ON_HEAP_MSGQ;
+	    erts_smp_atomic32_read_band_nob(&c_p->state,
+					    ~ERTS_PSFLG_ON_HEAP_MSGQ);
+	    goto change_to_off_heap;
+	default:
+	    res = THE_NON_VALUE; /* badarg */
+	    break;
+	}
+	break;
+
+    case 0:
+	res = am_mixed;
+
+	switch (new_state) {
+	case am_mixed:
+	    break;
+	case am_on_heap:
+	    c_p->flags |= F_ON_HEAP_MSGQ;
+	    erts_smp_atomic32_read_bor_nob(&c_p->state,
+					   ERTS_PSFLG_ON_HEAP_MSGQ);
+	    break;
+	case am_off_heap:
+	    goto change_to_off_heap;
+	default:
+	    res = THE_NON_VALUE; /* badarg */
+	    break;
+	}
+	break;
+
+    default:
+	res = am_error;
+	ERTS_INTERNAL_ERROR("Inconsistent message queue management state");
+	break;
+    }
+
+    return res;
+
+change_to_off_heap:
+
+    c_p->flags |= F_OFF_HEAP_MSGQ;
+
+    /*
+     * We do not have to schedule a change if
+     * we have an ongoing off heap change...
+     */
+    if (!(c_p->flags & F_OFF_HEAP_MSGQ_CHNG)) {
+	ErtsChangeOffHeapMessageQueue *cohmq;
+	/*
+	 * Need to set ERTS_PSFLG_OFF_HEAP_MSGQ and wait
+	 * thread progress before completing the change in
+	 * order to ensure that all senders observe that
+	 * messages should be passed off heap. When the
+	 * change has completed, GC does not need to inspect
+	 * the message queue at all.
+	 */
+	erts_smp_atomic32_read_bor_nob(&c_p->state,
+				       ERTS_PSFLG_OFF_HEAP_MSGQ);
+	c_p->flags |= F_OFF_HEAP_MSGQ_CHNG;
+	cohmq = erts_alloc(ERTS_ALC_T_MSGQ_CHNG,
+			   sizeof(ErtsChangeOffHeapMessageQueue));
+	cohmq->pid = c_p->common.id;
+	erts_schedule_thr_prgr_later_op(change_off_heap_msgq,
+					(void *) cohmq,
+					&cohmq->lop);
+    }
+
+    return res;
+}
+
+int
+erts_decode_dist_message(Process *proc, ErtsProcLocks proc_locks,
+			 ErtsMessage *msgp, int force_off_heap)
+{
+    ErtsHeapFactory factory;
+    Eterm msg;
+    ErlHeapFragment *bp;
+    Sint need;
+    int decode_in_heap_frag;
+
+    decode_in_heap_frag = (force_off_heap
+			   || !(proc_locks & ERTS_PROC_LOCK_MAIN)
+			   || (proc->flags & F_OFF_HEAP_MSGQ));
+
+    if (msgp->data.dist_ext->heap_size >= 0)
+	need = msgp->data.dist_ext->heap_size;
+    else {
+	need = erts_decode_dist_ext_size(msgp->data.dist_ext);
+	if (need < 0) {
+	    /* bad msg; remove it... */
+	    if (is_not_immed(ERL_MESSAGE_TOKEN(msgp))) {
+		bp = erts_dist_ext_trailer(msgp->data.dist_ext);
+		erts_cleanup_offheap(&bp->off_heap);
+	    }
+	    erts_free_dist_ext_copy(msgp->data.dist_ext);
+	    msgp->data.dist_ext = NULL;
+	    return 0;
+	}
+
+	msgp->data.dist_ext->heap_size = need;
+    }
+
+    if (is_not_immed(ERL_MESSAGE_TOKEN(msgp))) {
+	bp = erts_dist_ext_trailer(msgp->data.dist_ext);
+	need += bp->used_size;
+    }
+
+    if (decode_in_heap_frag)
+	erts_factory_heap_frag_init(&factory, new_message_buffer(need));
+    else
+	erts_factory_proc_prealloc_init(&factory, proc, need);
+
+    ASSERT(msgp->data.dist_ext->heap_size >= 0);
+    if (is_not_immed(ERL_MESSAGE_TOKEN(msgp))) {
+	ErlHeapFragment *heap_frag;
+	heap_frag = erts_dist_ext_trailer(msgp->data.dist_ext);
+	ERL_MESSAGE_TOKEN(msgp) = copy_struct(ERL_MESSAGE_TOKEN(msgp),
+					      heap_frag->used_size,
+					      &factory.hp,
+					      factory.off_heap);
+	erts_cleanup_offheap(&heap_frag->off_heap);
     }
+
+    msg = erts_decode_dist_ext(&factory, msgp->data.dist_ext);
+    ERL_MESSAGE_TERM(msgp) = msg;
+    erts_free_dist_ext_copy(msgp->data.dist_ext);
+    msgp->data.attached = NULL;
+
+    if (is_non_value(msg)) {
+	erts_factory_undo(&factory);
+	return 0;
+    }
+
+    erts_factory_trim_and_close(&factory, msgp->m,
+				ERL_MESSAGE_REF_ARRAY_SZ);
+
+    ASSERT(!msgp->data.heap_frag);
+
+    if (decode_in_heap_frag)
+	msgp->data.heap_frag = factory.heap_frags;
+
+    return 1;
+}
+
+/*
+ * ERTS_INSPECT_MSGQ_KEEP_OH_MSGS == 0 will move off heap messages
+ * into the heap of the inspected process if off_heap_message_queue
+ * is false when process_info(_, messages) is called. That is, the
+ * following GC will have more data in the rootset compared to the
+ * scenario when process_info(_, messages) had not been called.
+ *
+ * ERTS_INSPECT_MSGQ_KEEP_OH_MSGS != 0 will keep off heap messages
+ * off heap when process_info(_, messages) is called regardless of
+ * the off_heap_message_queue setting of the process. That is, it
+ * will change the following execution of the process as little as
+ * possible.
+ */
+#define ERTS_INSPECT_MSGQ_KEEP_OH_MSGS 1
+
+Uint
+erts_prep_msgq_for_inspection(Process *c_p, Process *rp,
+			      ErtsProcLocks rp_locks, ErtsMessageInfo *mip)
+{
+    Uint tot_heap_size;
+    ErtsMessage* mp;
+    Sint i;
+    int self_on_heap;
+    
+    /*
+     * Prepare the message queue for inspection
+     * by process_info().
+     *
+     *
+     * - Decode all messages on external format
+     * - Remove all corrupt dist messages from queue
+     * - Save pointer to, and heap size need of each
+     *   message in the mip array.
+     * - Return total heap size need for all messages
+     *   that needs to be copied.
+     *
+     * If ERTS_INSPECT_MSGQ_KEEP_OH_MSGS == 0:
+     * - In case off heap messages is disabled and
+     *   we are inspecting our own queue, move all
+     *   off heap data into the heap.
+     */
+
+    self_on_heap = c_p == rp && !(c_p->flags & F_OFF_HEAP_MSGQ);
+
+    tot_heap_size = 0;
+    i = 0;
+    mp = rp->msg.first;
+    while (mp) {
+	Eterm msg = ERL_MESSAGE_TERM(mp);
+
+	mip[i].size = 0;
+
+	if (is_non_value(msg)) {
+	    /* Dist message on external format; decode it... */
+	    if (mp->data.attached)
+		erts_decode_dist_message(rp, rp_locks, mp,
+					 ERTS_INSPECT_MSGQ_KEEP_OH_MSGS);
+
+	    msg = ERL_MESSAGE_TERM(mp);
+
+	    if (is_non_value(msg)) {
+		ErtsMessage **mpp;
+		ErtsMessage *bad_mp = mp;
+		/*
+		 * Bad distribution message; remove
+		 * it from the queue...
+		 */
+		ASSERT(!mp->data.attached);
+
+		mpp = i == 0 ? &rp->msg.first : &mip[i-1].msgp->next;
+
+		if (rp->msg.save == &bad_mp->next)
+		    rp->msg.save = mpp;
+		if (rp->msg.last == &bad_mp->next)
+		    rp->msg.last = mpp;
+		mp = mp->next;
+		*mpp = mp;
+		rp->msg.len--;
+		bad_mp->next = NULL;
+		erts_cleanup_messages(bad_mp);
+		continue;
+	    }
+	}
+
+	ASSERT(is_value(msg));
+
+#if ERTS_INSPECT_MSGQ_KEEP_OH_MSGS
+	if (is_not_immed(msg) && (!self_on_heap || mp->data.attached)) {
+	    Uint sz = size_object(msg);
+	    mip[i].size = sz;
+	    tot_heap_size += sz;
+	}
+#else
+	if (self_on_heap) {
+	    if (mp->data.attached) {
+		ErtsMessage *tmp = NULL;
+		if (mp->data.attached != ERTS_MSG_COMBINED_HFRAG) {
+		    erts_link_mbuf_to_proc(rp, mp->data.heap_frag);
+		    mp->data.attached = NULL;
+		}
+		else {
+		    /*
+		     * Need to replace the message reference since
+		     * we will get references to the message data
+		     * from the heap...
+		     */
+		    ErtsMessage **mpp;
+		    tmp = erts_alloc_message(0, NULL);
+		    sys_memcpy((void *) tmp->m, (void *) mp->m,
+			       sizeof(Eterm)*ERL_MESSAGE_REF_ARRAY_SZ); 
+		    mpp = i == 0 ? &rp->msg.first : &mip[i-1].msgp->next;
+		    tmp->next = mp->next;
+		    if (rp->msg.save == &mp->next)
+			rp->msg.save = &tmp->next;
+		    if (rp->msg.last == &mp->next)
+			rp->msg.last = &tmp->next;
+		    *mpp = tmp;
+		    erts_save_message_in_proc(rp, mp);
+		    mp = tmp;
+		}
+	    }
+	}
+	else if (is_not_immed(msg)) {
+	    Uint sz = size_object(msg);
+	    mip[i].size = sz;
+	    tot_heap_size += sz;
+	}
+
+#endif
+
+	mip[i].msgp = mp;
+	i++;
+	mp = mp->next;
+    }
+
+    return tot_heap_size;
 }
 
 void erts_factory_proc_init(ErtsHeapFactory* factory,
@@ -1127,53 +1439,147 @@ void erts_factory_proc_prealloc_init(ErtsHeapFactory* factory,
 				     Process* p,
 				     Sint size)
 {
+    ErlHeapFragment *bp = p->mbuf;
     factory->mode     = FACTORY_HALLOC;
     factory->p        = p;
     factory->hp_start = HAlloc(p, size);
     factory->hp       = factory->hp_start;
     factory->hp_end   = factory->hp_start + size;
     factory->off_heap = &p->off_heap;
+    factory->message  = NULL;
     factory->off_heap_saved.first    = p->off_heap.first;
     factory->off_heap_saved.overhead = p->off_heap.overhead;
-    factory->heap_frags_saved = p->mbuf;
+    factory->heap_frags_saved = bp;
+    factory->heap_frags_saved_used = bp ? bp->used_size : 0;
     factory->heap_frags = NULL; /* not used */
     factory->alloc_type = 0; /* not used */
 }
 
-void erts_factory_message_init(ErtsHeapFactory* factory,
-			       Process* rp,
-			       Eterm* hp,
-			       ErlHeapFragment* bp)
+void erts_factory_heap_frag_init(ErtsHeapFactory* factory,
+				 ErlHeapFragment* bp)
+{
+    factory->mode     = FACTORY_HEAP_FRAGS;
+    factory->p        = NULL;
+    factory->hp_start = bp->mem;
+    factory->hp       = bp->mem;
+    factory->hp_end   = bp->mem + bp->alloc_size;
+    factory->off_heap = &bp->off_heap;
+    factory->message  = NULL;
+    factory->heap_frags = bp;
+    factory->heap_frags_saved = NULL;
+    factory->heap_frags_saved_used = 0;
+    factory->alloc_type = ERTS_ALC_T_HEAP_FRAG;
+    ASSERT(!bp->next);
+    factory->off_heap_saved.first    = factory->off_heap->first;
+    factory->off_heap_saved.overhead = factory->off_heap->overhead;
+
+    ASSERT(factory->hp >= factory->hp_start && factory->hp <= factory->hp_end);
+}
+
+
+ErtsMessage *
+erts_factory_message_create(ErtsHeapFactory* factory,
+			    Process *proc,
+			    ErtsProcLocks *proc_locksp,
+			    Uint sz)
+{
+    Eterm *hp;
+    ErlOffHeap *ohp;
+    ErtsMessage *msgp;
+    int on_heap;
+    erts_aint32_t state;
+
+    state = erts_smp_atomic32_read_nob(&proc->state);
+
+    if (state & ERTS_PSFLG_OFF_HEAP_MSGQ) {
+	msgp = erts_alloc_message(sz, &hp);
+	ohp = sz == 0 ? NULL : &msgp->hfrag.off_heap;
+	on_heap = 0;
+    }
+    else {
+	msgp = erts_try_alloc_message_on_heap(proc, &state,
+					      proc_locksp,
+					      sz, &hp, &ohp,
+					      &on_heap);
+    }
+
+    if (on_heap) {
+	ASSERT(*proc_locksp & ERTS_PROC_LOCK_MAIN);
+	ASSERT(ohp == &proc->off_heap);
+	factory->mode = FACTORY_HALLOC;
+	factory->p = proc;
+	factory->heap_frags_saved = proc->mbuf;
+	factory->heap_frags_saved_used = proc->mbuf ? proc->mbuf->used_size : 0;
+    }
+    else {
+	factory->mode = FACTORY_MESSAGE;
+	factory->p = NULL;
+	factory->heap_frags_saved = NULL;
+	factory->heap_frags_saved_used = 0;
+
+	if (msgp->data.attached == ERTS_MSG_COMBINED_HFRAG) {
+	    ASSERT(!msgp->hfrag.next);
+	    factory->heap_frags = NULL;
+	}
+	else {
+	    ASSERT(!msgp->data.heap_frag
+		   || !msgp->data.heap_frag->next);
+	    factory->heap_frags = msgp->data.heap_frag;
+	}
+    }
+    factory->hp_start = hp;
+    factory->hp       = hp;
+    factory->hp_end   = hp + sz;
+    factory->message  = msgp;
+    factory->off_heap = ohp;
+    factory->alloc_type = ERTS_ALC_T_HEAP_FRAG;
+    if (ohp) {
+	factory->off_heap_saved.first    = ohp->first;
+	factory->off_heap_saved.overhead = ohp->overhead;
+    }
+    else {
+	factory->off_heap_saved.first    = NULL;
+	factory->off_heap_saved.overhead = 0;
+    }
+
+    ASSERT(factory->hp >= factory->hp_start && factory->hp <= factory->hp_end);
+
+    return msgp;
+}
+
+void erts_factory_selfcontained_message_init(ErtsHeapFactory* factory,
+					     ErtsMessage *msgp,
+					     Eterm *hp)
 {
-    if (bp) {
-	factory->mode     = FACTORY_HEAP_FRAGS;
-	factory->p        = NULL;
-	factory->hp_start = bp->mem;
-	factory->hp       = hp ? hp : bp->mem;
-	factory->hp_end   = bp->mem + bp->alloc_size;
-	factory->off_heap = &bp->off_heap;
-        factory->heap_frags = bp;
-        factory->heap_frags_saved = bp;
-        factory->alloc_type = ERTS_ALC_T_HEAP_FRAG;
-        ASSERT(!bp->next);
+    ErlHeapFragment* bp;
+    if (msgp->data.attached == ERTS_MSG_COMBINED_HFRAG) {
+	bp = &msgp->hfrag;
+	factory->heap_frags = NULL;
     }
     else {
-	factory->mode     = FACTORY_HALLOC;
-	factory->p        = rp;
-	factory->hp_start = hp;
-	factory->hp       = hp;
-	factory->hp_end   = HEAP_TOP(rp);
-	factory->off_heap = &rp->off_heap;
-        factory->heap_frags_saved = rp->mbuf;
-        factory->heap_frags = NULL; /* not used */
-        factory->alloc_type = 0; /* not used */
+	bp = msgp->data.heap_frag;
+	factory->heap_frags = bp;
     }
+    factory->mode     = FACTORY_MESSAGE;
+    factory->p        = NULL;
+    factory->hp_start = bp->mem;
+    factory->hp       = hp;
+    factory->hp_end   = bp->mem + bp->alloc_size;
+    factory->message  = msgp;
+    factory->off_heap = &bp->off_heap;
+    factory->heap_frags_saved = NULL;
+    factory->heap_frags_saved_used = 0;
+    factory->alloc_type = ERTS_ALC_T_HEAP_FRAG;
+    ASSERT(!bp->next);
     factory->off_heap_saved.first    = factory->off_heap->first;
     factory->off_heap_saved.overhead = factory->off_heap->overhead;
 
     ASSERT(factory->hp >= factory->hp_start && factory->hp <= factory->hp_end);
 }
 
+/* One static sized heap that must suffice.
+   No extra heap fragments will be allocated.
+*/
 void erts_factory_static_init(ErtsHeapFactory* factory,
 			     Eterm* hp,
 			     Uint size,
@@ -1188,6 +1594,23 @@ void erts_factory_static_init(ErtsHeapFactory* factory,
     factory->off_heap_saved.overhead = factory->off_heap->overhead;
 }
 
+/* A temporary heap with default buffer allocated/freed by client.
+ * factory_close is same as factory_undo
+ */
+void erts_factory_tmp_init(ErtsHeapFactory* factory, Eterm* hp, Uint size,
+			   Uint32 atype)
+{
+    factory->mode     = FACTORY_TMP;
+    factory->hp_start = hp;
+    factory->hp       = hp;
+    factory->hp_end   = hp + size;
+    factory->heap_frags = NULL;
+    factory->off_heap_saved.first    = NULL;
+    factory->off_heap_saved.overhead = 0;
+    factory->off_heap = &factory->off_heap_saved;
+    factory->alloc_type = atype;
+}
+
 /* When we know the term is an immediate and need no heap.
 */
 void erts_factory_dummy_init(ErtsHeapFactory* factory)
@@ -1230,8 +1653,17 @@ static void reserve_heap(ErtsHeapFactory* factory, Uint need, Uint xtra)
 	factory->hp_end = factory->hp + need;
 	return;
 
-    case FACTORY_HEAP_FRAGS:
-	bp = factory->heap_frags;
+    case FACTORY_MESSAGE:
+	if (!factory->heap_frags) {
+	    ASSERT(factory->message->data.attached == ERTS_MSG_COMBINED_HFRAG);
+	    bp = &factory->message->hfrag;
+	}
+	else {
+	    /* Fall through */
+	case FACTORY_HEAP_FRAGS:
+	case FACTORY_TMP:
+	    bp = factory->heap_frags;
+	}
 
         if (bp) {
 	    ASSERT(factory->hp > bp->mem);
@@ -1269,8 +1701,23 @@ void erts_factory_close(ErtsHeapFactory* factory)
 	HRelease(factory->p, factory->hp_end, factory->hp);
 	break;
 
-    case FACTORY_HEAP_FRAGS:
-	bp = factory->heap_frags;
+    case FACTORY_MESSAGE:
+	if (!factory->heap_frags) {
+	    if (factory->message->data.attached == ERTS_MSG_COMBINED_HFRAG)
+		bp = &factory->message->hfrag;
+	    else
+		bp = NULL;
+	}
+	else {
+	    if (factory->message->data.attached	== ERTS_MSG_COMBINED_HFRAG)
+		factory->message->hfrag.next = factory->heap_frags;
+	    else
+		factory->message->data.heap_frag = factory->heap_frags;
+
+	    /* Fall through */
+	case FACTORY_HEAP_FRAGS:
+	    bp = factory->heap_frags;
+	}
 
 	if (bp) {
 	    ASSERT(factory->hp >= bp->mem);
@@ -1280,6 +1727,9 @@ void erts_factory_close(ErtsHeapFactory* factory)
 	    bp->used_size = factory->hp - bp->mem;
         }
 	break;
+    case FACTORY_TMP:
+	erts_factory_undo(factory);
+	break;
     case FACTORY_STATIC: break;
     case FACTORY_CLOSED: break;
     default:
@@ -1291,17 +1741,47 @@ void erts_factory_close(ErtsHeapFactory* factory)
 void erts_factory_trim_and_close(ErtsHeapFactory* factory,
 				 Eterm *brefs, Uint brefs_size)
 {
-    if (factory->mode == FACTORY_HEAP_FRAGS) {
-        ErlHeapFragment* bp = factory->heap_frags;
+    ErlHeapFragment *bp;
+
+    switch (factory->mode) {
+    case FACTORY_MESSAGE: {
+	ErtsMessage *mp = factory->message;
+	if (mp->data.attached == ERTS_MSG_COMBINED_HFRAG) {
+	    if (!mp->hfrag.next) {
+		Uint sz = factory->hp - factory->hp_start;
+		mp = erts_shrink_message(mp, sz, brefs, brefs_size);
+		factory->message = mp;
+		factory->mode = FACTORY_CLOSED;
+		return;
+	    }
+	    /*else we don't trim multi fragmented messages for now (off_heap...) */
+	    break;
+	}
+	/* Fall through... */
+    }
+    case FACTORY_HEAP_FRAGS:
+	bp = factory->heap_frags;
+	if (!bp)
+	    break;
         if (bp->next == NULL) {
             Uint used_sz = factory->hp - bp->mem;
             ASSERT(used_sz <= bp->alloc_size);
-            factory->heap_frags = erts_resize_message_buffer(bp, used_sz,
-                                                             brefs, brefs_size);
+	    if (used_sz > 0)
+		bp = erts_resize_message_buffer(bp, used_sz,
+						brefs, brefs_size);
+	    else {
+		free_message_buffer(bp);
+		bp = NULL;
+	    }
+	    factory->heap_frags = bp;
+	    if (factory->mode == FACTORY_MESSAGE)
+		factory->message->data.heap_frag = bp;
             factory->mode = FACTORY_CLOSED;
             return;
         }
-        /*else we don't trim multi fragmented messages for now */
+        /*else we don't trim multi fragmented messages for now (off_heap...) */
+    default:
+	break;
     }
     erts_factory_close(factory);
 }
@@ -1349,38 +1829,47 @@ void erts_factory_undo(ErtsHeapFactory* factory)
 
             /* Rollback heap top
 	     */
-            if (factory->heap_frags_saved == NULL) { /* No heap frags when we started */
-                ASSERT(factory->hp_start >= HEAP_START(factory->p));
-                ASSERT(factory->hp_start <= HEAP_LIMIT(factory->p));
 
-                HEAP_TOP(factory->p) = factory->hp_start;
-            }
-            else {
+	    if (HEAP_START(factory->p) <= factory->hp_start
+		&& factory->hp_start <= HEAP_LIMIT(factory->p)) {
+		HEAP_TOP(factory->p) = factory->hp_start;
+	    }
+
+	    /* Fix last heap frag */
+            if (factory->heap_frags_saved) {
                 ASSERT(factory->heap_frags_saved == factory->p->mbuf);
-                if (factory->hp_start == factory->heap_frags_saved->mem) {
+                if (factory->hp_start != factory->heap_frags_saved->mem)
+                    factory->heap_frags_saved->used_size = factory->heap_frags_saved_used;
+		else {
                     factory->p->mbuf = factory->p->mbuf->next;
                     ERTS_HEAP_FREE(ERTS_ALC_T_HEAP_FRAG, factory->heap_frags_saved,
                                    ERTS_HEAP_FRAG_SIZE(factory->heap_frags_saved->alloc_size));
                 }
-                else if (factory->hp_start != factory->hp_end) {
-                    unsigned remains = factory->hp_start - factory->heap_frags_saved->mem;
-                    ASSERT(remains > 0 && remains < factory->heap_frags_saved->used_size);
-                    factory->heap_frags_saved->used_size = remains;
-                }
             }
         }
         break;
 
+    case FACTORY_MESSAGE:
+	if (factory->message->data.attached == ERTS_MSG_COMBINED_HFRAG)
+	    factory->message->hfrag.next = factory->heap_frags;
+	else
+	    factory->message->data.heap_frag = factory->heap_frags;
+	erts_cleanup_messages(factory->message);
+	break;
+    case FACTORY_TMP:
     case FACTORY_HEAP_FRAGS:
+	erts_cleanup_offheap(factory->off_heap);
+	factory->off_heap->first = NULL;
+
         bp = factory->heap_frags;
-        do {
+        while (bp != NULL) {
             ErlHeapFragment* next_bp = bp->next;
 
-            erts_cleanup_offheap(&bp->off_heap);
+            ASSERT(bp->off_heap.first == NULL);
             ERTS_HEAP_FREE(factory->alloc_type, (void *) bp,
-                           ERTS_HEAP_FRAG_SIZE(bp->size));
+                           ERTS_HEAP_FRAG_SIZE(bp->alloc_size));
             bp = next_bp;
-        }while (bp != NULL);
+        }
 	break;
 
     case FACTORY_CLOSED: break;
@@ -1394,4 +1883,3 @@ void erts_factory_undo(ErtsHeapFactory* factory)
     factory->heap_frags = NULL;
 #endif
 }
-
diff --git a/erts/emulator/beam/erl_message.h b/erts/emulator/beam/erl_message.h
index fbdf3fb0e2..60035d15ae 100644
--- a/erts/emulator/beam/erl_message.h
+++ b/erts/emulator/beam/erl_message.h
@@ -24,6 +24,8 @@
 struct proc_bin;
 struct external_thing_;
 
+typedef struct erl_mesg ErtsMessage;
+
 /*
  * This struct represents data that must be updated by structure copy,
  * but is stored outside of any heap.
@@ -32,9 +34,6 @@ struct external_thing_;
 struct erl_off_heap_header {
     Eterm thing_word;
     Uint size;
-#if HALFWORD_HEAP
-    void* dummy_ptr_padding__;
-#endif
     struct erl_off_heap_header* next;
 };
 
@@ -57,15 +56,19 @@ typedef struct {
     enum {
         FACTORY_CLOSED = 0,
         FACTORY_HALLOC,
+        FACTORY_MESSAGE,
         FACTORY_HEAP_FRAGS,
-        FACTORY_STATIC
+        FACTORY_STATIC,
+        FACTORY_TMP
     } mode;
     Process* p;
     Eterm* hp_start;
     Eterm* hp;
     Eterm* hp_end;
+    ErtsMessage *message;
     struct erl_heap_fragment* heap_frags;
     struct erl_heap_fragment* heap_frags_saved;
+    Uint heap_frags_saved_used;
     ErlOffHeap* off_heap;
     ErlOffHeap off_heap_saved;
     Uint32 alloc_type;
@@ -73,8 +76,12 @@ typedef struct {
 
 void erts_factory_proc_init(ErtsHeapFactory*, Process*);
 void erts_factory_proc_prealloc_init(ErtsHeapFactory*, Process*, Sint size);
-void erts_factory_message_init(ErtsHeapFactory*, Process*, Eterm* hp, struct erl_heap_fragment*);
+void erts_factory_heap_frag_init(ErtsHeapFactory*, struct erl_heap_fragment*);
+ErtsMessage *erts_factory_message_create(ErtsHeapFactory *, Process *,
+					  ErtsProcLocks *, Uint sz);
+void erts_factory_selfcontained_message_init(ErtsHeapFactory*, ErtsMessage *, Eterm *);
 void erts_factory_static_init(ErtsHeapFactory*, Eterm* hp, Uint size, ErlOffHeap*);
+void erts_factory_tmp_init(ErtsHeapFactory*, Eterm* hp, Uint size, Uint32 atype);
 void erts_factory_dummy_init(ErtsHeapFactory*);
 
 Eterm* erts_produce_heap(ErtsHeapFactory*, Uint need, Uint xtra);
@@ -94,6 +101,8 @@ void erts_factory_undo(ErtsHeapFactory*);
 #include "external.h"
 #include "erl_process.h"
 
+#define ERTS_INVALID_HFRAG_PTR ((ErlHeapFragment *) ~((UWord) 7))
+
 /*
  * This struct represents a heap fragment, which is used when there
  * isn't sufficient room in the process heap and we can't do a GC.
@@ -108,33 +117,53 @@ struct erl_heap_fragment {
     Eterm mem[1];		/* Data */
 };
 
-typedef struct erl_mesg {
-    struct erl_mesg* next;	/* Next message */
-    union {
-	ErtsDistExternal *dist_ext;
-	ErlHeapFragment *heap_frag;
-	void *attached;
-    } data;
+/* m[0] = message, m[1] = seq trace token */
+#define ERL_MESSAGE_REF_ARRAY_SZ 2
+#define ERL_MESSAGE_TERM(mp) ((mp)->m[0])
+#define ERL_MESSAGE_TOKEN(mp) ((mp)->m[1])
+
 #ifdef USE_VM_PROBES
-    Eterm m[3];                 /* m[0] = message, m[1] = seq trace token, m[3] = dynamic trace user tag */
+/* m[2] = dynamic trace user tag */
+#undef ERL_MESSAGE_REF_ARRAY_SZ
+#define ERL_MESSAGE_REF_ARRAY_SZ 3
+#define ERL_MESSAGE_DT_UTAG(mp) ((mp)->m[2])
 #else
-    Eterm m[2];			/* m[0] = message, m[1] = seq trace token */
 #endif
-} ErlMessage;
 
-#define ERL_MESSAGE_TERM(mp) ((mp)->m[0])
-#define ERL_MESSAGE_TOKEN(mp) ((mp)->m[1])
 #ifdef USE_VM_PROBES
-#define ERL_MESSAGE_DT_UTAG(mp) ((mp)->m[2])
+#define have_no_seqtrace(T) ((T) == NIL || (T) == am_have_dt_utag)
+#else
+#define have_no_seqtrace(T) ((T) == NIL)
 #endif
+#define have_seqtrace(T)    (!have_no_seqtrace(T))
+
+#define ERL_MESSAGE_REF_FIELDS__			\
+    ErtsMessage *next;	/* Next message */		\
+    union {						\
+	ErtsDistExternal *dist_ext;			\
+	ErlHeapFragment *heap_frag;			\
+	void *attached;					\
+    } data;						\
+    Eterm m[ERL_MESSAGE_REF_ARRAY_SZ]
+
+
+typedef struct erl_msg_ref__ {
+    ERL_MESSAGE_REF_FIELDS__;
+} ErtsMessageRef;
+
+struct erl_mesg {
+    ERL_MESSAGE_REF_FIELDS__;
+
+    ErlHeapFragment hfrag;
+};
 
 /* Size of default message buffer (erl_message.c) */
 #define ERL_MESSAGE_BUF_SZ 500
 
 typedef struct {
-    ErlMessage* first;
-    ErlMessage** last;  /* point to the last next pointer */
-    ErlMessage** save;
+    ErtsMessage* first;
+    ErtsMessage** last;  /* point to the last next pointer */
+    ErtsMessage** save;
     Sint len;            /* queue length */
 
     /*
@@ -142,14 +171,14 @@ typedef struct {
      * recv_set/1 instructions.
      */
     BeamInstr* mark;		/* address to rec_loop/2 instruction */
-    ErlMessage** saved_last;	/* saved last pointer */
+    ErtsMessage** saved_last;	/* saved last pointer */
 } ErlMessageQueue;
 
 #ifdef ERTS_SMP
 
 typedef struct {
-    ErlMessage* first;
-    ErlMessage** last;  /* point to the last next pointer */
+    ErtsMessage* first;
+    ErtsMessage** last;  /* point to the last next pointer */
     Sint len;            /* queue length */
 } ErlMessageInQueue;
 
@@ -200,7 +229,7 @@ do {									\
 
 /* Unlink current message */
 #define UNLINK_MESSAGE(p,msgp) do { \
-     ErlMessage* __mp = (msgp)->next; \
+     ErtsMessage* __mp = (msgp)->next; \
      *(p)->msg.save = __mp; \
      (p)->msg.len--; \
      if (__mp == NULL) \
@@ -216,76 +245,33 @@ do {									\
 #define SAVE_MESSAGE(p) \
      (p)->msg.save = &(*(p)->msg.save)->next
 
-/*
- * ErtsMoveMsgAttachmentIntoProc() moves data attached to a message
- * onto the heap of a process. The attached data is the content of
- * the the message either on the internal format or on the external
- * format, and also possibly a seq trace token on the internal format.
- * If the message content is on the external format, the decode might
- * fail. If the decoding fails, ERL_MESSAGE_TERM(M) will contain
- * THE_NON_VALUE. That is, ERL_MESSAGE_TERM(M) *has* to be checked
- * afterwards and taken care of appropriately.
- *
- * ErtsMoveMsgAttachmentIntoProc() will shallow copy to heap if
- * possible; otherwise, move to heap via garbage collection.
- *
- * ErtsMoveMsgAttachmentIntoProc() is used when receiveing messages
- * in process_main() and in hipe_check_get_msg().
- */
-
-#define ErtsMoveMsgAttachmentIntoProc(M, P, ST, HT, FC, SWPO, SWPI)	\
-do {									\
-    if ((M)->data.attached) {						\
-	Uint need__ = erts_msg_attached_data_size((M));			\
-        { SWPO ; }							\
- 	if ((ST) - (HT) >= need__) {					\
-            ErtsHeapFactory factory__;                                  \
-            erts_factory_proc_prealloc_init(&factory__, (P), need__);   \
-	    erts_move_msg_attached_data_to_heap(&factory__, (M));       \
-            erts_factory_close(&factory__);                             \
-            if ((P)->mbuf != NULL) {                                    \
-                /* Heap was exhausted by messages. This is a rare case */   \
-                /* that can currently (OTP 18) only happen if hamts are */  \
-                /* far exceeding the estimated heap size. Do GC. */         \
-                (FC) -= erts_garbage_collect((P), 0, NULL, 0);		\
-            }								\
-	}								\
-	else {								\
-	    (FC) -= erts_garbage_collect((P), 0, NULL, 0);		\
-	}								\
-        { SWPI ; }							\
-	ASSERT(!(M)->data.attached);					\
-    }									\
-} while (0)
-
 #define ERTS_SND_FLG_NO_SEQ_TRACE		(((unsigned) 1) << 0)
 
 #define ERTS_HEAP_FRAG_SIZE(DATA_WORDS) \
    (sizeof(ErlHeapFragment) - sizeof(Eterm) + (DATA_WORDS)*sizeof(Eterm))
 
-#define ERTS_INIT_HEAP_FRAG(HEAP_FRAG_P, DATA_WORDS)	\
-do {							\
-    (HEAP_FRAG_P)->next = NULL;				\
-    (HEAP_FRAG_P)->alloc_size = (DATA_WORDS);		\
-    (HEAP_FRAG_P)->used_size = (DATA_WORDS);            \
-    (HEAP_FRAG_P)->off_heap.first = NULL; 	        \
-    (HEAP_FRAG_P)->off_heap.overhead = 0;		\
-} while (0)
+#define ERTS_INIT_HEAP_FRAG(HEAP_FRAG_P, USED_WORDS, DATA_WORDS)	\
+    do {								\
+	(HEAP_FRAG_P)->next = NULL;					\
+	(HEAP_FRAG_P)->alloc_size = (DATA_WORDS);			\
+	(HEAP_FRAG_P)->used_size = (USED_WORDS);			\
+	(HEAP_FRAG_P)->off_heap.first = NULL;				\
+	(HEAP_FRAG_P)->off_heap.overhead = 0;				\
+    } while (0)
 
 void init_message(void);
-void free_message(ErlMessage *);
 ErlHeapFragment* new_message_buffer(Uint);
 ErlHeapFragment* erts_resize_message_buffer(ErlHeapFragment *, Uint,
 					    Eterm *, Uint);
 void free_message_buffer(ErlHeapFragment *);
 void erts_queue_dist_message(Process*, ErtsProcLocks*, ErtsDistExternal *, Eterm);
 #ifdef USE_VM_PROBES
-void erts_queue_message_probe(Process*, ErtsProcLocks*, ErlHeapFragment*,
+void erts_queue_message_probe(Process*, ErtsProcLocks*, ErtsMessage*,
                               Eterm message, Eterm seq_trace_token, Eterm dt_utag);
 #define erts_queue_message(RP,RL,BP,Msg,SEQ) \
     erts_queue_message_probe((RP),(RL),(BP),(Msg),(SEQ),NIL)
 #else
-void erts_queue_message(Process*, ErtsProcLocks*, ErlHeapFragment*,
+void erts_queue_message(Process*, ErtsProcLocks*, ErtsMessage*,
                         Eterm message, Eterm seq_trace_token);
 #define erts_queue_message_probe(RP,RL,BP,Msg,SEQ,TAG) \
     erts_queue_message((RP),(RL),(BP),(Msg),(SEQ))
@@ -294,20 +280,141 @@ void erts_deliver_exit_message(Eterm, Process*, ErtsProcLocks *, Eterm, Eterm);
 Sint erts_send_message(Process*, Process*, ErtsProcLocks*, Eterm, unsigned);
 void erts_link_mbuf_to_proc(Process *proc, ErlHeapFragment *bp);
 
-void erts_move_msg_mbuf_to_heap(Eterm**, ErlOffHeap*, ErlMessage *);
-
-Uint erts_msg_attached_data_size_aux(ErlMessage *msg);
-void erts_move_msg_attached_data_to_heap(ErtsHeapFactory*, ErlMessage *);
-Eterm erts_msg_distext2heap(Process *, ErtsProcLocks *, ErlHeapFragment **,
-			    Eterm *, ErtsDistExternal *);
+Uint erts_msg_attached_data_size_aux(ErtsMessage *msg);
 
 void erts_cleanup_offheap(ErlOffHeap *offheap);
+void erts_save_message_in_proc(Process *p, ErtsMessage *msg);
+Sint erts_move_messages_off_heap(Process *c_p);
+Sint erts_complete_off_heap_message_queue_change(Process *c_p);
+Eterm erts_change_message_queue_management(Process *c_p, Eterm new_state);
+
+int erts_decode_dist_message(Process *, ErtsProcLocks, ErtsMessage *, int);
+
+void erts_cleanup_messages(ErtsMessage *mp);
+
+typedef struct {
+    Uint size;
+    ErtsMessage *msgp;
+} ErtsMessageInfo;
+
+Uint erts_prep_msgq_for_inspection(Process *c_p,
+				   Process *rp,
+				   ErtsProcLocks rp_locks,
+				   ErtsMessageInfo *mip);
+
+void *erts_alloc_message_ref(void);
+void erts_free_message_ref(void *);
 
+#define ERTS_SMALL_FIX_MSG_SZ 10
+#define ERTS_MEDIUM_FIX_MSG_SZ 20
+#define ERTS_LARGE_FIX_MSG_SZ 30
 
+void *erts_alloc_small_message(void);
+void erts_free_small_message(void *mp);
+
+typedef struct {
+    ErtsMessage m;
+    Eterm data[ERTS_SMALL_FIX_MSG_SZ-1];
+} ErtsSmallFixSzMessage;
+
+typedef struct {
+    ErtsMessage m;
+    Eterm data[ERTS_MEDIUM_FIX_MSG_SZ-1];
+} ErtsMediumFixSzMessage;
+
+typedef struct {
+    ErtsMessage m;
+    Eterm data[ERTS_LARGE_FIX_MSG_SZ-1];
+} ErtsLargeFixSzMessage;
+
+ErtsMessage *erts_try_alloc_message_on_heap(Process *pp,
+					    erts_aint32_t *psp,
+					    ErtsProcLocks *plp,
+					    Uint sz,
+					    Eterm **hpp,
+					    ErlOffHeap **ohpp,
+					    int *on_heap_p);
+ErtsMessage *erts_realloc_shrink_message(ErtsMessage *mp, Uint sz,
+					 Eterm *brefs, Uint brefs_size);
+
+ERTS_GLB_FORCE_INLINE ErtsMessage *erts_alloc_message(Uint sz, Eterm **hpp);
+ERTS_GLB_FORCE_INLINE ErtsMessage *erts_shrink_message(ErtsMessage *mp, Uint sz,
+						       Eterm *brefs, Uint brefs_size);
+ERTS_GLB_FORCE_INLINE void erts_free_message(ErtsMessage *mp);
 ERTS_GLB_INLINE Uint erts_used_frag_sz(const ErlHeapFragment*);
-ERTS_GLB_INLINE Uint erts_msg_attached_data_size(ErlMessage *msg);
+ERTS_GLB_INLINE Uint erts_msg_attached_data_size(ErtsMessage *msg);
+
+#define ERTS_MSG_COMBINED_HFRAG ((void *) 0x1)
 
 #if ERTS_GLB_INLINE_INCL_FUNC_DEF
+
+ERTS_GLB_FORCE_INLINE ErtsMessage *erts_alloc_message(Uint sz, Eterm **hpp)
+{
+    ErtsMessage *mp;
+
+    if (sz == 0) {
+	mp = erts_alloc_message_ref();
+	mp->next = NULL;
+	ERL_MESSAGE_TERM(mp) = NIL;
+	mp->data.attached = NULL;
+	if (hpp)
+	    *hpp = NULL;
+	return mp;
+    }
+
+    mp = erts_alloc(ERTS_ALC_T_MSG,
+		    sizeof(ErtsMessage) + (sz - 1)*sizeof(Eterm));
+
+    mp->next = NULL;
+    ERL_MESSAGE_TERM(mp) = NIL;
+    mp->data.attached = ERTS_MSG_COMBINED_HFRAG;
+    ERTS_INIT_HEAP_FRAG(&mp->hfrag, sz, sz);
+
+    if (hpp)
+	*hpp = &mp->hfrag.mem[0];
+
+    return mp;
+}
+
+ERTS_GLB_FORCE_INLINE ErtsMessage *
+erts_shrink_message(ErtsMessage *mp, Uint sz, Eterm *brefs, Uint brefs_size)
+{
+    if (sz == 0) {
+	ErtsMessage *nmp;
+	if (!mp->data.attached)
+	    return mp;
+	ASSERT(mp->data.attached == ERTS_MSG_COMBINED_HFRAG);
+	nmp = erts_alloc_message_ref();
+#ifdef DEBUG
+	if (brefs && brefs_size) {
+	    int i;
+	    for (i = 0; i < brefs_size; i++)
+		ASSERT(is_non_value(brefs[i]) || is_immed(brefs[i]));
+	}
+#endif
+	erts_free(ERTS_ALC_T_MSG, mp);
+	return nmp;
+    }
+
+    ASSERT(mp->data.attached == ERTS_MSG_COMBINED_HFRAG);
+    ASSERT(mp->hfrag.used_size >= sz);
+
+    if (sz >= (mp->hfrag.alloc_size - mp->hfrag.alloc_size / 16)) {
+	mp->hfrag.used_size = sz;
+	return mp;
+    }
+
+    return erts_realloc_shrink_message(mp, sz, brefs, brefs_size);
+}
+
+ERTS_GLB_FORCE_INLINE void erts_free_message(ErtsMessage *mp)
+{
+    if (mp->data.attached != ERTS_MSG_COMBINED_HFRAG)
+	erts_free_message_ref(mp);
+    else
+	erts_free(ERTS_ALC_T_MSG, mp);
+}
+
 ERTS_GLB_INLINE Uint erts_used_frag_sz(const ErlHeapFragment* bp)
 {
     Uint sz = 0;
@@ -317,11 +424,17 @@ ERTS_GLB_INLINE Uint erts_used_frag_sz(const ErlHeapFragment* bp)
     return sz;
 }
 
-ERTS_GLB_INLINE Uint erts_msg_attached_data_size(ErlMessage *msg)
+ERTS_GLB_INLINE Uint erts_msg_attached_data_size(ErtsMessage *msg)
 {
     ASSERT(msg->data.attached);
-    if (is_value(ERL_MESSAGE_TERM(msg)))
-	return erts_used_frag_sz(msg->data.heap_frag);
+    if (is_value(ERL_MESSAGE_TERM(msg))) {
+	ErlHeapFragment *bp;
+	if (msg->data.attached == ERTS_MSG_COMBINED_HFRAG)
+	    bp = &msg->hfrag;
+	else
+	    bp = msg->data.heap_frag;
+	return erts_used_frag_sz(bp);
+    }
     else if (msg->data.dist_ext->heap_size < 0)
 	return erts_msg_attached_data_size_aux(msg);
     else {
diff --git a/erts/emulator/beam/erl_nif.c b/erts/emulator/beam/erl_nif.c
index add4a66f90..d659788f30 100644
--- a/erts/emulator/beam/erl_nif.c
+++ b/erts/emulator/beam/erl_nif.c
@@ -314,6 +314,7 @@ int enif_send(ErlNifEnv* env, const ErlNifPid* to_pid,
     ErtsProcLocks rp_locks = 0;
     Process* rp;
     Process* c_p;
+    ErtsMessage *mp;
     ErlHeapFragment* frags;
     Eterm receiver = to_pid->pid;
     int flush_me = 0;
@@ -347,7 +348,7 @@ int enif_send(ErlNifEnv* env, const ErlNifPid* to_pid,
     ASSERT(frags == MBUF(&menv->phony_proc));
     if (frags != NULL) {
 	/* Move all offheap's from phony proc to the first fragment.
-	   Quick and dirty, but erts_move_msg_mbuf_to_heap doesn't care. */
+	   Quick and dirty... */
 	ASSERT(!is_offheap(&frags->off_heap));
 	frags->off_heap = MSO(&menv->phony_proc);
 	clear_offheap(&MSO(&menv->phony_proc));
@@ -359,7 +360,9 @@ int enif_send(ErlNifEnv* env, const ErlNifPid* to_pid,
     if (flush_me) {	
 	flush_env(env); /* Needed for ERTS_HOLE_CHECK */ 
     }
-    erts_queue_message(rp, &rp_locks, frags, msg, am_undefined);
+    mp = erts_alloc_message(0, NULL);
+    mp->data.heap_frag = frags;
+    erts_queue_message(rp, &rp_locks, mp, msg, am_undefined);
     if (c_p == rp)
 	rp_locks &= ~ERTS_PROC_LOCK_MAIN;
     if (rp_locks)
@@ -376,9 +379,19 @@ ERL_NIF_TERM enif_make_copy(ErlNifEnv* dst_env, ERL_NIF_TERM src_term)
 {
     Uint sz;
     Eterm* hp;
+#ifdef SHCOPY
+    erts_shcopy_t info;
+    INITIALIZE_SHCOPY(info);
+    sz = copy_shared_calculate(src_term, &info);
+    hp = alloc_heap(dst_env, sz);
+    src_term = copy_shared_perform(src_term, sz, &info, &hp, &MSO(dst_env->proc));
+    DESTROY_SHCOPY(info);
+    return src_term;
+#else
     sz = size_object(src_term);
     hp = alloc_heap(dst_env, sz);
     return copy_struct(src_term, sz, &hp, &MSO(dst_env->proc));
+#endif
 }
 
 
@@ -1173,6 +1186,7 @@ ErlNifTid enif_thread_self(void) { return erl_drv_thread_self(); }
 int enif_equal_tids(ErlNifTid tid1, ErlNifTid tid2) { return erl_drv_equal_tids(tid1,tid2); }
 void enif_thread_exit(void *resp) { erl_drv_thread_exit(resp); }
 int enif_thread_join(ErlNifTid tid, void **respp) { return erl_drv_thread_join(tid,respp); }
+int enif_getenv(const char *key, char *value, size_t *value_size) { return erl_drv_getenv(key, value, value_size); }
 
 int enif_fprintf(void* filep, const char* format, ...) 
 { 
@@ -2230,17 +2244,17 @@ int enif_map_iterator_get_pair(ErlNifEnv *env,
  ***************************************************************************/
 
 
-static BeamInstr** get_func_pp(BeamInstr* mod_code, Eterm f_atom, unsigned arity)
+static BeamInstr** get_func_pp(BeamCodeHeader* mod_code, Eterm f_atom, unsigned arity)
 {
-    int n = (int) mod_code[MI_NUM_FUNCTIONS];
+    int n = (int) mod_code->num_functions;
     int j;
     for (j = 0; j < n; ++j) {
-	BeamInstr* code_ptr = (BeamInstr*) mod_code[MI_FUNCTIONS+j];
+	BeamInstr* code_ptr = (BeamInstr*) mod_code->functions[j];
 	ASSERT(code_ptr[0] == (BeamInstr) BeamOp(op_i_func_info_IaaI));
 	if (f_atom == ((Eterm) code_ptr[3])
 	    && arity == ((unsigned) code_ptr[4])) {
 
-	    return (BeamInstr**) &mod_code[MI_FUNCTIONS+j];
+	    return (BeamInstr**) &mod_code->functions[j];
 	}
     }
     return NULL;
@@ -2423,8 +2437,8 @@ BIF_RETTYPE load_nif_2(BIF_ALIST_2)
     if (init_func != NULL)
       handle = init_func;
 
-    if (!in_area(caller, mod->curr.code, mod->curr.code_length)) {
-	ASSERT(in_area(caller, mod->old.code, mod->old.code_length));
+    if (!in_area(caller, mod->curr.code_hdr, mod->curr.code_length)) {
+	ASSERT(in_area(caller, mod->old.code_hdr, mod->old.code_length));
 
 	ret = load_nif_error(BIF_P, "old_code", "Calling load_nif from old "
 			     "module '%T' not allowed", mod_atom);
@@ -2478,7 +2492,7 @@ BIF_RETTYPE load_nif_2(BIF_ALIST_2)
 	for (i=0; i < entry->num_of_funcs && ret==am_ok; i++) {
 	    BeamInstr** code_pp;
 	    if (!erts_atom_get(f->name, sys_strlen(f->name), &f_atom, ERTS_ATOM_ENC_LATIN1)
-		|| (code_pp = get_func_pp(mod->curr.code, f_atom, f->arity))==NULL) {
+		|| (code_pp = get_func_pp(mod->curr.code_hdr, f_atom, f->arity))==NULL) {
 		ret = load_nif_error(BIF_P,bad_lib,"Function not found %T:%s/%u",
 				     mod_atom, f->name, f->arity);
 	    }
@@ -2621,7 +2635,7 @@ BIF_RETTYPE load_nif_2(BIF_ALIST_2)
 	{
 	    BeamInstr* code_ptr;
 	    erts_atom_get(f->name, sys_strlen(f->name), &f_atom, ERTS_ATOM_ENC_LATIN1);
-	    code_ptr = *get_func_pp(mod->curr.code, f_atom, f->arity);
+	    code_ptr = *get_func_pp(mod->curr.code_hdr, f_atom, f->arity);
 
 	    if (code_ptr[1] == 0) {
 		code_ptr[5+0] = (BeamInstr) BeamOp(op_call_nif);
diff --git a/erts/emulator/beam/erl_nif.h b/erts/emulator/beam/erl_nif.h
index 7d880126f8..40c2ad6f08 100644
--- a/erts/emulator/beam/erl_nif.h
+++ b/erts/emulator/beam/erl_nif.h
@@ -48,9 +48,10 @@
 **           add ErlNifEntry options
 **           add ErlNifFunc flags
 ** 2.8: 18.0 add enif_has_pending_exception
+** 2.9: 18.2 enif_getenv
 */
 #define ERL_NIF_MAJOR_VERSION 2
-#define ERL_NIF_MINOR_VERSION 8
+#define ERL_NIF_MINOR_VERSION 9
 
 /*
  * The emulator will refuse to load a nif-lib with a major version
@@ -86,10 +87,6 @@
 #  define SIZEOF_LONG_LONG_SAVED__ SIZEOF_LONG_LONG
 #  undef SIZEOF_LONG_LONG
 #endif
-#ifdef HALFWORD_HEAP_EMULATOR
-#  define HALFWORD_HEAP_EMULATOR_SAVED__ HALFWORD_HEAP_EMULATOR
-#  undef HALFWORD_HEAP_EMULATOR
-#endif
 #include "erl_int_sizes_config.h"
 
 #ifdef __cplusplus
@@ -109,16 +106,11 @@ typedef long long ErlNifSInt64;
 #error No 64-bit integer type
 #endif
 
-#ifdef HALFWORD_HEAP_EMULATOR
-#  define ERL_NIF_VM_VARIANT "beam.halfword" 
-typedef unsigned int ERL_NIF_TERM;
-#else
-#  define ERL_NIF_VM_VARIANT "beam.vanilla" 
-#  if SIZEOF_LONG == SIZEOF_VOID_P
+#define ERL_NIF_VM_VARIANT "beam.vanilla" 
+#if SIZEOF_LONG == SIZEOF_VOID_P
 typedef unsigned long ERL_NIF_TERM;
-#  elif SIZEOF_LONG_LONG == SIZEOF_VOID_P
+#elif SIZEOF_LONG_LONG == SIZEOF_VOID_P
 typedef unsigned long long ERL_NIF_TERM;
-#  endif
 #endif
 
 typedef ERL_NIF_TERM ERL_NIF_UINT;
@@ -231,6 +223,7 @@ typedef enum {
 #  define ERL_NIF_API_FUNC_DECL(RET_TYPE, NAME, ARGS) RET_TYPE (*NAME) ARGS
 typedef struct {
 #  include "erl_nif_api_funcs.h"
+   void* erts_alc_test;
 } TWinDynNifCallbacks;
 extern TWinDynNifCallbacks WinDynNifCallbacks;
 #  undef ERL_NIF_API_FUNC_DECL
@@ -250,22 +243,28 @@ extern TWinDynNifCallbacks WinDynNifCallbacks;
 
 #if (defined(__WIN32__) || defined(_WIN32) || defined(_WIN32_))
 #  define ERL_NIF_INIT_GLOB TWinDynNifCallbacks WinDynNifCallbacks;
-#  ifdef STATIC_ERLANG_NIF
-#    define ERL_NIF_INIT_DECL(MODNAME) __declspec(dllexport) ErlNifEntry* MODNAME ## _nif_init(TWinDynNifCallbacks* callbacks)
-#  else
-#    define ERL_NIF_INIT_DECL(MODNAME) __declspec(dllexport) ErlNifEntry* nif_init(TWinDynNifCallbacks* callbacks)
-#  endif
+#  define ERL_NIF_INIT_ARGS TWinDynNifCallbacks* callbacks
 #  define ERL_NIF_INIT_BODY memcpy(&WinDynNifCallbacks,callbacks,sizeof(TWinDynNifCallbacks))
+#  define ERL_NIF_INIT_EXPORT __declspec(dllexport)
 #else 
 #  define ERL_NIF_INIT_GLOB
+#  define ERL_NIF_INIT_ARGS void
 #  define ERL_NIF_INIT_BODY
-#  ifdef STATIC_ERLANG_NIF
-#    define ERL_NIF_INIT_DECL(MODNAME)  ErlNifEntry* MODNAME ## _nif_init(void)
+#  if defined(__GNUC__) && __GNUC__ >= 4
+#    define ERL_NIF_INIT_EXPORT __attribute__ ((visibility("default")))
+#  elif defined (__SUNPRO_C) && (__SUNPRO_C >= 0x550)
+#    define ERL_NIF_INIT_EXPORT __global
 #  else
-#    define ERL_NIF_INIT_DECL(MODNAME)  ErlNifEntry* nif_init(void)
+#    define ERL_NIF_INIT_EXPORT
 #  endif
 #endif
 
+#ifdef STATIC_ERLANG_NIF
+#  define ERL_NIF_INIT_DECL(MODNAME) ErlNifEntry* MODNAME ## _nif_init(ERL_NIF_INIT_ARGS)
+#else
+#  define ERL_NIF_INIT_DECL(MODNAME) ERL_NIF_INIT_EXPORT ErlNifEntry* nif_init(ERL_NIF_INIT_ARGS)
+#endif
+
 #ifdef ERL_NIF_DIRTY_SCHEDULER_SUPPORT
 #  define ERL_NIF_ENTRY_OPTIONS ERL_NIF_DIRTY_NIF_OPTION
 #else
diff --git a/erts/emulator/beam/erl_nif_api_funcs.h b/erts/emulator/beam/erl_nif_api_funcs.h
index f93152c921..08b9afc6af 100644
--- a/erts/emulator/beam/erl_nif_api_funcs.h
+++ b/erts/emulator/beam/erl_nif_api_funcs.h
@@ -159,6 +159,7 @@ ERL_NIF_API_FUNC_DECL(int, enif_map_iterator_get_pair, (ErlNifEnv *env, ErlNifMa
 ERL_NIF_API_FUNC_DECL(ERL_NIF_TERM,enif_schedule_nif,(ErlNifEnv*,const char*,int,ERL_NIF_TERM (*)(ErlNifEnv*,int,const ERL_NIF_TERM[]),int,const ERL_NIF_TERM[]));
 ERL_NIF_API_FUNC_DECL(int, enif_has_pending_exception, (ErlNifEnv *env, ERL_NIF_TERM* reason));
 ERL_NIF_API_FUNC_DECL(ERL_NIF_TERM, enif_raise_exception, (ErlNifEnv *env, ERL_NIF_TERM reason));
+ERL_NIF_API_FUNC_DECL(int,enif_getenv,(const char* key, char* value, size_t* value_size));
 
 /*
 ** ADD NEW ENTRIES HERE (before this comment) !!!
@@ -310,6 +311,7 @@ ERL_NIF_API_FUNC_DECL(int,enif_is_on_dirty_scheduler,(ErlNifEnv*));
 #  define enif_schedule_nif ERL_NIF_API_FUNC_MACRO(enif_schedule_nif)
 #  define enif_has_pending_exception ERL_NIF_API_FUNC_MACRO(enif_has_pending_exception)
 #  define enif_raise_exception ERL_NIF_API_FUNC_MACRO(enif_raise_exception)
+#  define enif_getenv ERL_NIF_API_FUNC_MACRO(enif_getenv)
 
 /*
 ** ADD NEW ENTRIES HERE (before this comment)
@@ -543,7 +545,7 @@ static ERL_NIF_INLINE ERL_NIF_TERM enif_make_list9(ErlNifEnv* env,
 
 #ifndef enif_make_pid
 
-#  define enif_make_pid(ENV, PID) ((const ERL_NIF_TERM)((PID)->pid))
+#  define enif_make_pid(ENV, PID) ((void)(ENV),(const ERL_NIF_TERM)((PID)->pid))
 
 #if SIZEOF_LONG == 8
 #  define enif_get_int64 enif_get_long
diff --git a/erts/emulator/beam/erl_node_container_utils.h b/erts/emulator/beam/erl_node_container_utils.h
index 211b1a0090..b6a3531e28 100644
--- a/erts/emulator/beam/erl_node_container_utils.h
+++ b/erts/emulator/beam/erl_node_container_utils.h
@@ -255,7 +255,7 @@ extern ErtsPTab erts_port;
  * Refs                                                                    *
 \*                                                                         */
 
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
 
 #define internal_ref_no_of_numbers(x)					\
   (internal_ref_data((x))[0])
@@ -328,8 +328,6 @@ extern ErtsPTab erts_port;
 					 : external_ref_channel_no((x)))
 #define is_ref(x)			(is_internal_ref((x))		\
 					 || is_external_ref((x)))
-#define is_ref_rel(x,Base)		(is_internal_ref_rel((x),Base)	\
-					 || is_external_ref_rel((x),Base))
 #define is_not_ref(x)			(!is_ref(x))
 
 #endif
diff --git a/erts/emulator/beam/erl_node_tables.c b/erts/emulator/beam/erl_node_tables.c
index 2fb790b953..13f14adbab 100644
--- a/erts/emulator/beam/erl_node_tables.c
+++ b/erts/emulator/beam/erl_node_tables.c
@@ -37,18 +37,18 @@ erts_smp_rwmtx_t erts_node_table_rwmtx;
 
 DistEntry *erts_hidden_dist_entries;
 DistEntry *erts_visible_dist_entries;
-DistEntry *erts_not_connected_dist_entries;
+DistEntry *erts_not_connected_dist_entries; /* including erts_this_dist_entry */
 Sint erts_no_of_hidden_dist_entries;
 Sint erts_no_of_visible_dist_entries;
-Sint erts_no_of_not_connected_dist_entries;
+Sint erts_no_of_not_connected_dist_entries; /* including erts_this_dist_entry */
 
 DistEntry *erts_this_dist_entry;
 ErlNode *erts_this_node;
 char erts_this_node_sysname_BUFFER[256],
     *erts_this_node_sysname = "uninitialized yet";
 
-static Uint node_entries;
-static Uint dist_entries;
+static Uint node_entries = 0;
+static Uint dist_entries = 0;
 
 static int references_atoms_need_init = 1;
 
@@ -91,9 +91,6 @@ dist_table_alloc(void *dep_tmpl)
     erts_smp_rwmtx_opt_t rwmtx_opt = ERTS_SMP_RWMTX_OPT_DEFAULT_INITER;
     rwmtx_opt.type = ERTS_SMP_RWMTX_TYPE_FREQUENT_READ;
 
-    if(((DistEntry *) dep_tmpl) == erts_this_dist_entry)
-	return dep_tmpl;
-
     sysname = ((DistEntry *) dep_tmpl)->sysname;
     chnl_nr = make_small((Uint) atom_val(sysname));
     dep = (DistEntry *) erts_alloc(ERTS_ALC_T_DIST_ENTRY, sizeof(DistEntry));
@@ -132,7 +129,9 @@ dist_table_alloc(void *dep_tmpl)
 
     /* Link in */
 
-    /* All new dist entries are "not connected" */
+    /* All new dist entries are "not connected".
+     * erts_this_dist_entry is also always included among "not connected"
+     */
     dep->next = erts_not_connected_dist_entries;
     if(erts_not_connected_dist_entries) {
 	ASSERT(erts_not_connected_dist_entries->prev == NULL);
@@ -149,9 +148,6 @@ dist_table_free(void *vdep)
 {
     DistEntry *dep = (DistEntry *) vdep;
 
-    if(dep == erts_this_dist_entry)
-	return;
-
     ASSERT(is_nil(dep->cid));
     ASSERT(dep->nlinks == NULL);
     ASSERT(dep->node_links == NULL);
@@ -186,7 +182,7 @@ dist_table_free(void *vdep)
 #endif
     erts_free(ERTS_ALC_T_DIST_ENTRY, (void *) dep);
 
-    ASSERT(dist_entries > 1);
+    ASSERT(dist_entries > 0);
     dist_entries--;
 }
 
@@ -306,7 +302,7 @@ static void try_delete_dist_entry(void *vdep)
      * thread incremented refc twice. Once for the new reference
      * and once for this thread.
      *
-     * If refc reach -1, noone has used the entry since we
+     * If refc reach -1, no one has used the entry since we
      * set up the timer. Delete the entry.
      *
      * If refc reach 0, the entry is currently not in use
@@ -369,8 +365,7 @@ erts_dist_table_size(void)
 
     ASSERT(dist_entries == (erts_no_of_visible_dist_entries
 			    + erts_no_of_hidden_dist_entries
-			    + erts_no_of_not_connected_dist_entries
-			    + 1 /* erts_this_dist_entry */));
+			    + erts_no_of_not_connected_dist_entries));
 #endif
 
     res = (hash_table_sz(&erts_dist_table)
@@ -543,9 +538,6 @@ node_table_alloc(void *venp_tmpl)
 {
     ErlNode *enp;
 
-    if(((ErlNode *) venp_tmpl) == erts_this_node)
-	return venp_tmpl;
-
     enp = (ErlNode *) erts_alloc(ERTS_ALC_T_NODE_ENTRY, sizeof(ErlNode));
 
     node_entries++;
@@ -563,8 +555,7 @@ node_table_free(void *venp)
 {
     ErlNode *enp = (ErlNode *) venp;
 
-    if(enp == erts_this_node)
-	return;
+    ERTS_SMP_LC_ASSERT(enp != erts_this_node || erts_thr_progress_is_blocking());
 
     erts_deref_dist_entry(enp->dist_entry);
 #ifdef DEBUG
@@ -572,7 +563,7 @@ node_table_free(void *venp)
 #endif
     erts_free(ERTS_ALC_T_NODE_ENTRY, venp);
 
-    ASSERT(node_entries > 1);
+    ASSERT(node_entries > 0);
     node_entries--;
 }
 
@@ -650,7 +641,7 @@ static void try_delete_node(void *venp)
      * thread incremented refc twice. Once for the new reference
      * and once for this thread.
      *
-     * If refc reach -1, noone has used the entry since we
+     * If refc reach -1, no one has used the entry since we
      * set up the timer. Delete the entry.
      *
      * If refc reach 0, the entry is currently not in use
@@ -747,25 +738,24 @@ void erts_print_node_info(int to,
 void
 erts_set_this_node(Eterm sysname, Uint creation)
 {
-    erts_smp_rwmtx_rwlock(&erts_node_table_rwmtx);
-    erts_smp_rwmtx_rwlock(&erts_dist_table_rwmtx);
+    ERTS_SMP_LC_ASSERT(erts_thr_progress_is_blocking());
+    ASSERT(erts_refc_read(&erts_this_dist_entry->refc, 2));
 
-    (void) hash_erase(&erts_dist_table, (void *) erts_this_dist_entry);
-    erts_this_dist_entry->sysname = sysname;
-    erts_this_dist_entry->creation = creation;
-    (void) hash_put(&erts_dist_table, (void *) erts_this_dist_entry);
+    if (erts_refc_dectest(&erts_this_node->refc, 0) == 0)
+        try_delete_node(erts_this_node);
 
-    (void) hash_erase(&erts_node_table, (void *) erts_this_node);
-    erts_this_node->sysname = sysname;
-    erts_this_node->creation = creation;
-    erts_this_node_sysname = erts_this_node_sysname_BUFFER;
-    erts_snprintf(erts_this_node_sysname, sizeof(erts_this_node_sysname_BUFFER),
-                  "%T", sysname);
-    (void) hash_put(&erts_node_table, (void *) erts_this_node);
+    if (erts_refc_dectest(&erts_this_dist_entry->refc, 0) == 0)
+        try_delete_dist_entry(erts_this_dist_entry);
 
-    erts_smp_rwmtx_rwunlock(&erts_dist_table_rwmtx);
-    erts_smp_rwmtx_rwunlock(&erts_node_table_rwmtx);
+    erts_this_node = NULL; /* to make sure refc is bumped for this node */
+    erts_this_node = erts_find_or_insert_node(sysname, creation);
+    erts_this_dist_entry = erts_this_node->dist_entry;
+
+    erts_refc_inc(&erts_this_dist_entry->refc, 2);
 
+    erts_this_node_sysname = erts_this_node_sysname_BUFFER;
+    erts_snprintf(erts_this_node_sysname, sizeof(erts_this_node_sysname_BUFFER),
+		  "%T", sysname);
 }
 
 Uint
@@ -782,6 +772,7 @@ void erts_init_node_tables(int dd_sec)
 {
     erts_smp_rwmtx_opt_t rwmtx_opt = ERTS_SMP_RWMTX_OPT_DEFAULT_INITER;
     HashFunctions f;
+    ErlNode node_tmpl;
 
     if (dd_sec == ERTS_NODE_TAB_DELAY_GC_INFINITY)
 	node_tab_delete_delay = (ErtsMonotonicTime) -1;
@@ -793,16 +784,24 @@ void erts_init_node_tables(int dd_sec)
     rwmtx_opt.type = ERTS_SMP_RWMTX_TYPE_FREQUENT_READ;
     rwmtx_opt.lived = ERTS_SMP_RWMTX_LONG_LIVED;
 
-    f.hash  = (H_FUN)			dist_table_hash;
-    f.cmp   = (HCMP_FUN)		dist_table_cmp;
-    f.alloc = (HALLOC_FUN)		dist_table_alloc;
-    f.free  = (HFREE_FUN)		dist_table_free;
-
-    erts_this_dist_entry = erts_alloc(ERTS_ALC_T_DIST_ENTRY, sizeof(DistEntry));
-    dist_entries = 1;
+    erts_smp_rwmtx_init_opt(&erts_node_table_rwmtx, &rwmtx_opt, "node_table");
+    erts_smp_rwmtx_init_opt(&erts_dist_table_rwmtx, &rwmtx_opt, "dist_table");
 
+    f.hash       = (H_FUN)		dist_table_hash;
+    f.cmp        = (HCMP_FUN)		dist_table_cmp;
+    f.alloc      = (HALLOC_FUN)		dist_table_alloc;
+    f.free       = (HFREE_FUN)		dist_table_free;
+    f.meta_alloc = (HMALLOC_FUN) 	erts_alloc;
+    f.meta_free  = (HMFREE_FUN) 	erts_free;
+    f.meta_print = (HMPRINT_FUN) 	erts_print;
     hash_init(ERTS_ALC_T_DIST_TABLE, &erts_dist_table, "dist_table", 11, f);
 
+    f.hash  = (H_FUN)      			node_table_hash;
+    f.cmp   = (HCMP_FUN)   			node_table_cmp;
+    f.alloc = (HALLOC_FUN) 			node_table_alloc;
+    f.free  = (HFREE_FUN)  			node_table_free;
+    hash_init(ERTS_ALC_T_NODE_TABLE, &erts_node_table, "node_table", 11, f);
+
     erts_hidden_dist_entries				= NULL;
     erts_visible_dist_entries				= NULL;
     erts_not_connected_dist_entries			= NULL;
@@ -810,69 +809,23 @@ void erts_init_node_tables(int dd_sec)
     erts_no_of_visible_dist_entries			= 0;
     erts_no_of_not_connected_dist_entries		= 0;
 
-    erts_this_dist_entry->next				= NULL;
-    erts_this_dist_entry->prev				= NULL;
-    erts_refc_init(&erts_this_dist_entry->refc, 1); /* erts_this_node */
-
-    erts_smp_rwmtx_init_opt_x(&erts_this_dist_entry->rwmtx,
-			      &rwmtx_opt,
-			      "dist_entry",
-			      make_small(ERST_INTERNAL_CHANNEL_NO));
-    erts_this_dist_entry->sysname			= am_Noname;
-    erts_this_dist_entry->cid				= NIL;
-    erts_this_dist_entry->connection_id			= 0;
-    erts_this_dist_entry->status			= 0;
-    erts_this_dist_entry->flags				= 0;
-    erts_this_dist_entry->version			= 0;
-
-    erts_smp_mtx_init_x(&erts_this_dist_entry->lnk_mtx,
-			"dist_entry_links",
-			make_small(ERST_INTERNAL_CHANNEL_NO));
-    erts_this_dist_entry->node_links			= NULL;
-    erts_this_dist_entry->nlinks			= NULL;
-    erts_this_dist_entry->monitors			= NULL;
-
-    erts_smp_mtx_init_x(&erts_this_dist_entry->qlock,
-			"dist_entry_out_queue",
-			make_small(ERST_INTERNAL_CHANNEL_NO));
-    erts_this_dist_entry->qflgs				= 0;
-    erts_this_dist_entry->qsize				= 0;
-    erts_this_dist_entry->out_queue.first		= NULL;
-    erts_this_dist_entry->out_queue.last		= NULL;
-    erts_this_dist_entry->suspended			= NULL;
-
-    erts_this_dist_entry->finalized_out_queue.first	= NULL;
-    erts_this_dist_entry->finalized_out_queue.last	= NULL;
-    erts_smp_atomic_init_nob(&erts_this_dist_entry->dist_cmd_scheduled, 0);
-    erts_port_task_handle_init(&erts_this_dist_entry->dist_cmd);
-    erts_this_dist_entry->send				= NULL;
-    erts_this_dist_entry->cache				= NULL;
-
-    (void) hash_put(&erts_dist_table, (void *) erts_this_dist_entry);
-
-    f.hash  = (H_FUN)      			node_table_hash;
-    f.cmp   = (HCMP_FUN)   			node_table_cmp;
-    f.alloc = (HALLOC_FUN) 			node_table_alloc;
-    f.free  = (HFREE_FUN)  			node_table_free;
+    node_tmpl.sysname = am_Noname;
+    node_tmpl.creation = 0;
+    erts_this_node = hash_put(&erts_node_table, &node_tmpl);
+     /* +1 for erts_this_node */
+    erts_refc_init(&erts_this_node->refc, 1);
 
-    hash_init(ERTS_ALC_T_NODE_TABLE, &erts_node_table, "node_table", 11, f);
+    ASSERT(erts_this_node->dist_entry != NULL);
+    erts_this_dist_entry = erts_this_node->dist_entry;
+    /* +1 for erts_this_dist_entry */
+    /* +1 for erts_this_node->dist_entry */
+    erts_refc_init(&erts_this_dist_entry->refc, 2);
 
-    erts_this_node = erts_alloc(ERTS_ALC_T_NODE_ENTRY, sizeof(ErlNode));
-    node_entries = 1;
 
-    erts_refc_init(&erts_this_node->refc, 1); /* The system itself */
-    erts_this_node->sysname			= am_Noname;
-    erts_this_node->creation			= 0;
-    erts_this_node->dist_entry			= erts_this_dist_entry;
     erts_this_node_sysname = erts_this_node_sysname_BUFFER;
     erts_snprintf(erts_this_node_sysname, sizeof(erts_this_node_sysname_BUFFER),
                   "%T", erts_this_node->sysname);
 
-    (void) hash_put(&erts_node_table, (void *) erts_this_node);
-
-    erts_smp_rwmtx_init_opt(&erts_node_table_rwmtx, &rwmtx_opt, "node_table");
-    erts_smp_rwmtx_init_opt(&erts_dist_table_rwmtx, &rwmtx_opt, "dist_table");
-
     references_atoms_need_init = 1;
 }
 
@@ -1207,23 +1160,11 @@ insert_offheap(ErlOffHeap *oh, int type, Eterm id)
 			break;
 		    }
 		if (insert_bin) {
-#if HALFWORD_HEAP
-		    UWord val = (UWord) u.pb->val;
-		    DeclareTmpHeapNoproc(id_heap,BIG_UINT_HEAP_SIZE*2); /* extra place allocated */
-#else
 		    DeclareTmpHeapNoproc(id_heap,BIG_UINT_HEAP_SIZE);
-#endif
 		    Uint *hp = &id_heap[0];
 		    InsertedBin *nib;
-#if HALFWORD_HEAP
-		    int actual_need = BIG_UWORD_HEAP_SIZE(val);
-		    ASSERT(actual_need <= (BIG_UINT_HEAP_SIZE*2));
-		    UseTmpHeapNoproc(actual_need);
-		    a.id = erts_bld_uword(&hp, NULL, (UWord) val);
-#else
 		    UseTmpHeapNoproc(BIG_UINT_HEAP_SIZE);
 		    a.id = erts_bld_uint(&hp, NULL, (Uint) u.pb->val);
-#endif
 		    erts_match_prog_foreach_offheap(u.pb->val,
 						    insert_offheap2,
 						    (void *) &a);
@@ -1231,11 +1172,7 @@ insert_offheap(ErlOffHeap *oh, int type, Eterm id)
 		    nib->bin_val = u.pb->val;
 		    nib->next = inserted_bins;
 		    inserted_bins = nib;
-#if HALFWORD_HEAP
-		    UnUseTmpHeapNoproc(actual_need);
-#else
 		    UnUseTmpHeapNoproc(BIG_UINT_HEAP_SIZE);
-#endif
 		}
 	    }		
 	    break;
@@ -1410,6 +1347,10 @@ setup_reference_table(void)
 		SYSTEM_REF,
 		TUPLE2(&heap[0], AM_system, am_undefined));
 
+    insert_dist_entry(erts_this_dist_entry,
+                      SYSTEM_REF,
+                      TUPLE2(&heap[0], AM_system, am_undefined),
+                      erts_this_node->creation);
     UnUseTmpHeapNoproc(3);
 
     max = erts_ptab_max(&erts_proc);
@@ -1417,67 +1358,55 @@ setup_reference_table(void)
     for (i = 0; i < max; i++) {
 	Process *proc = erts_pix2proc(i);
 	if (proc) {
-	    ErlMessage *msg;
+	    int mli;
+	    ErtsMessage *msg_list[] = {
+		proc->msg.first,
+#ifdef ERTS_SMP
+		proc->msg_inq.first,
+#endif
+		proc->msg_frag};
 
 	    /* Insert Heap */
 	    insert_offheap(&(proc->off_heap),
 			   HEAP_REF,
 			   proc->common.id);
-	    /* Insert message buffers */
+	    /* Insert heap fragments buffers */
 	    for(hfp = proc->mbuf; hfp; hfp = hfp->next)
 		insert_offheap(&(hfp->off_heap),
 			       HEAP_REF,
 			       proc->common.id);
-	    /* Insert msg msg buffers */
-	    for (msg = proc->msg.first; msg; msg = msg->next) {
-		ErlHeapFragment *heap_frag = NULL;
-		if (msg->data.attached) {
-		    if (is_value(ERL_MESSAGE_TERM(msg)))
-			heap_frag = msg->data.heap_frag;
-		    else {
-			if (msg->data.dist_ext->dep)
-			    insert_dist_entry(msg->data.dist_ext->dep,
-					      HEAP_REF, proc->common.id, 0);
-			if (is_not_nil(ERL_MESSAGE_TOKEN(msg)))
-			    heap_frag = erts_dist_ext_trailer(msg->data.dist_ext);
+
+	    /* Insert msg buffers */
+	    for (mli = 0; mli < sizeof(msg_list)/sizeof(msg_list[0]); mli++) {
+		ErtsMessage *msg;
+		for (msg = msg_list[mli]; msg; msg = msg->next) {
+		    ErlHeapFragment *heap_frag = NULL;
+		    if (msg->data.attached) {
+			if (msg->data.attached == ERTS_MSG_COMBINED_HFRAG)
+			    heap_frag = &msg->hfrag;
+			else if (is_value(ERL_MESSAGE_TERM(msg)))
+			    heap_frag = msg->data.heap_frag;
+			else {
+			    if (msg->data.dist_ext->dep)
+				insert_dist_entry(msg->data.dist_ext->dep,
+						  HEAP_REF, proc->common.id, 0);
+			    if (is_not_nil(ERL_MESSAGE_TOKEN(msg)))
+				heap_frag = erts_dist_ext_trailer(msg->data.dist_ext);
+			}
 		    }
-		}
-		if (heap_frag)
-		    insert_offheap(&(heap_frag->off_heap),
-				   HEAP_REF,
-				   proc->common.id);
-	    }
-#ifdef ERTS_SMP
-	    for (msg = proc->msg_inq.first; msg; msg = msg->next) {
-		ErlHeapFragment *heap_frag = NULL;
-		if (msg->data.attached) {
-		    if (is_value(ERL_MESSAGE_TERM(msg)))
-			heap_frag = msg->data.heap_frag;
-		    else {
-			if (msg->data.dist_ext->dep)
-			    insert_dist_entry(msg->data.dist_ext->dep,
-					      HEAP_REF, proc->common.id, 0);
-			if (is_not_nil(ERL_MESSAGE_TOKEN(msg)))
-			    heap_frag = erts_dist_ext_trailer(msg->data.dist_ext);
+		    while (heap_frag) {
+			insert_offheap(&(heap_frag->off_heap),
+				       HEAP_REF,
+				       proc->common.id);
+			heap_frag = heap_frag->next;
 		    }
 		}
-		if (heap_frag)
-		    insert_offheap(&(heap_frag->off_heap),
-				   HEAP_REF,
-				   proc->common.id);
 	    }
-#endif
 	    /* Insert links */
 	    if (ERTS_P_LINKS(proc))
 		insert_links(ERTS_P_LINKS(proc), proc->common.id);
 	    if (ERTS_P_MONITORS(proc))
 		insert_monitors(ERTS_P_MONITORS(proc), proc->common.id);
-	    /* Insert controller */
-	    {
-		DistEntry *dep = ERTS_PROC_GET_DIST_ENTRY(proc);
-		if (dep)
-		    insert_dist_entry(dep, CTRL_REF, proc->common.id, 0);
-	    }
 	}
     }
     
diff --git a/erts/emulator/beam/erl_node_tables.h b/erts/emulator/beam/erl_node_tables.h
index 694ac84232..fb2f2a5407 100644
--- a/erts/emulator/beam/erl_node_tables.h
+++ b/erts/emulator/beam/erl_node_tables.h
@@ -41,6 +41,7 @@
 
 #include "sys.h"
 #include "hash.h"
+#include "erl_alloc.h"
 #include "erl_process.h"
 #include "erl_monitors.h"
 #include "erl_smp.h"
@@ -66,7 +67,7 @@
 #define ERTS_DE_QFLGS_ALL			(ERTS_DE_QFLG_BUSY \
 						 | ERTS_DE_QFLG_EXIT)
 
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
 #define ERTS_DIST_OUTPUT_BUF_DBG_PATTERN ((Uint) 0xf713f713f713f713UL)
 #else
 #define ERTS_DIST_OUTPUT_BUF_DBG_PATTERN ((Uint) 0xf713f713)
diff --git a/erts/emulator/beam/erl_port.h b/erts/emulator/beam/erl_port.h
index acd68ef0ad..fa97707a87 100644
--- a/erts/emulator/beam/erl_port.h
+++ b/erts/emulator/beam/erl_port.h
@@ -187,6 +187,11 @@ struct _erl_drv_port {
 
     ErtsPrtSD *psd;		 /* Port specific data */
     int reds; /* Only used while executing driver callbacks */
+
+    struct {
+        Eterm to;
+        Uint32 ref[ERTS_MAX_REF_NUMBERS];
+    } *async_open_port;         /* Reference used with async open port */
 };
 
 
@@ -687,7 +692,7 @@ erts_drvport2port_state(ErlDrvPort drvport, erts_aint32_t *statep)
     Port *prt = ERTS_ErlDrvPort2Port(drvport);
     erts_aint32_t state;
     ASSERT(prt);
-    ERTS_LC_ASSERT(erts_lc_is_emu_thr());
+//    ERTS_LC_ASSERT(erts_lc_is_emu_thr());
     if (prt == ERTS_INVALID_ERL_DRV_PORT)
 	return ERTS_INVALID_ERL_DRV_PORT;
     ERTS_SMP_LC_ASSERT(erts_lc_is_port_locked(prt)
@@ -944,4 +949,9 @@ ErtsPortOpResult erts_port_control(Process *, Port *, unsigned int, Eterm, Eterm
 ErtsPortOpResult erts_port_call(Process *, Port *, unsigned int, Eterm, Eterm *);
 ErtsPortOpResult erts_port_info(Process *, Port *, Eterm, Eterm *);
 
+/*
+ * Signals from ports to ports. Used by sys drivers.
+ */
+int erl_drv_port_control(Eterm, char, char*, ErlDrvSizeT);
+
 #endif
diff --git a/erts/emulator/beam/erl_port_task.c b/erts/emulator/beam/erl_port_task.c
index 5c38db1cbc..2c09834d19 100644
--- a/erts/emulator/beam/erl_port_task.c
+++ b/erts/emulator/beam/erl_port_task.c
@@ -1734,7 +1734,7 @@ erts_port_task_execute(ErtsRunQueue *runq, Port **curr_port_pp)
 	    reds = ERTS_PORT_REDS_INPUT;
 	    ASSERT((state & ERTS_PORT_SFLGS_DEAD) == 0);
             DTRACE_DRIVER(driver_ready_input, pp);
-	    /* NOTE some windows/ose drivers use ->ready_input
+	    /* NOTE some windows drivers use ->ready_input
 	       for input and output */
 	    (*pp->drv_ptr->ready_input)((ErlDrvData) pp->drv_data,
 					ptp->u.alive.td.io.event);
diff --git a/erts/emulator/beam/erl_printf_term.c b/erts/emulator/beam/erl_printf_term.c
index 267c0b3ff4..45ba4371dc 100644
--- a/erts/emulator/beam/erl_printf_term.c
+++ b/erts/emulator/beam/erl_printf_term.c
@@ -117,13 +117,12 @@ do {									\
 /* return 0 if list is not a non-empty flat list of printable characters */
 
 static int
-is_printable_string(Eterm list, Eterm* base)
-{
+is_printable_string(Eterm list) {
     int len = 0;
     int c;
 
     while(is_list(list)) {
-	Eterm* consp = list_val_rel(list, base);
+	Eterm* consp = list_val(list);
 	Eterm hd = CAR(consp);
 
 	if (!is_byte(hd))
@@ -260,9 +259,7 @@ static char *format_binary(Uint16 x, char *b) {
 #endif
 
 static int
-print_term(fmtfn_t fn, void* arg, Eterm obj, long *dcount,
-	   Eterm* obj_base) /* ignored if !HALFWORD_HEAP */
-{
+print_term(fmtfn_t fn, void* arg, Eterm obj, long *dcount) {
     DECLARE_WSTACK(s);
     int res;
     int i;
@@ -308,7 +305,7 @@ print_term(fmtfn_t fn, void* arg, Eterm obj, long *dcount,
 		obj = (Eterm) popped.word;
 	    L_print_one_cons:
 		{
-		    Eterm* cons = list_val_rel(obj, obj_base);
+		    Eterm* cons = list_val(obj);
 		    Eterm tl;
 		    
 		    obj = CAR(cons);
@@ -344,11 +341,7 @@ print_term(fmtfn_t fn, void* arg, Eterm obj, long *dcount,
 	    PRINT_CHAR(res, fn, arg, '>');
 	    goto L_done;
 	}
-#if HALFWORD_HEAP
-	wobj = is_immed(obj) ? (Wterm)obj : rterm2wterm(obj, obj_base);
-#else
 	wobj = (Wterm)obj;
-#endif
 	switch (tag_val_def(wobj)) {
 	case NIL_DEF:
 	    PRINT_STRING(res, fn, arg, "[]");
@@ -424,10 +417,10 @@ print_term(fmtfn_t fn, void* arg, Eterm obj, long *dcount,
 	    PRINT_CHAR(res, fn, arg, '>');
 	    break;
 	case LIST_DEF:
-	    if (is_printable_string(obj, obj_base)) {
+	    if (is_printable_string(obj)) {
 		int c;
 		PRINT_CHAR(res, fn, arg, '"');
-		nobj = list_val_rel(obj, obj_base);
+		nobj = list_val(obj);
 		while (1) {
 		    if ((*dcount)-- <= 0)
 			goto L_done;
@@ -441,7 +434,7 @@ print_term(fmtfn_t fn, void* arg, Eterm obj, long *dcount,
 		    }
 		    if (is_not_list(*nobj))
 			break;
-		    nobj = list_val_rel(*nobj, obj_base);
+		    nobj = list_val(*nobj);
 		}
 		PRINT_CHAR(res, fn, arg, '"');
 	    } else {
@@ -467,16 +460,13 @@ print_term(fmtfn_t fn, void* arg, Eterm obj, long *dcount,
 	}
 	    break;
 	case BINARY_DEF:
-	    if (header_is_bin_matchstate(*boxed_val(wobj))) {
-		PRINT_STRING(res, fn, arg, "#MatchState");
-	    }
-	    else {
+	    {
 		byte* bytep;
-		Uint bytesize = binary_size_rel(obj,obj_base);
+		Uint bytesize = binary_size(obj);
 		Uint bitoffs;
 		Uint bitsize;
 		byte octet;
-		ERTS_GET_BINARY_BYTES_REL(obj, bytep, bitoffs, bitsize, obj_base);
+		ERTS_GET_BINARY_BYTES(obj, bytep, bitoffs, bitsize);
 
 		if (bitsize || !bytesize
 		    || !is_printable_ascii(bytep, bytesize, bitoffs)) {
@@ -633,6 +623,9 @@ print_term(fmtfn_t fn, void* arg, Eterm obj, long *dcount,
                 }
             }
             break;
+	case MATCHSTATE_DEF:
+	    PRINT_STRING(res, fn, arg, "#MatchState");
+	    break;
         default:
 	    PRINT_STRING(res, fn, arg, "<unknown:");
 	    PRINT_POINTER(res, fn, arg, wobj);
@@ -648,13 +641,11 @@ print_term(fmtfn_t fn, void* arg, Eterm obj, long *dcount,
 
 
 int
-erts_printf_term(fmtfn_t fn, void* arg, ErlPfEterm term, long precision,
-		 ErlPfEterm* term_base)
-{
+erts_printf_term(fmtfn_t fn, void* arg, ErlPfEterm term, long precision) {
     int res;
     ERTS_CT_ASSERT(sizeof(ErlPfEterm) == sizeof(Eterm));
 
-    res = print_term(fn, arg, (Eterm)term, &precision, (Eterm*)term_base);
+    res = print_term(fn, arg, (Eterm)term, &precision);
     if (res < 0)
 	return res;
     if (precision <= 0)
diff --git a/erts/emulator/beam/erl_printf_term.h b/erts/emulator/beam/erl_printf_term.h
index 87618a36d7..4618457f27 100644
--- a/erts/emulator/beam/erl_printf_term.h
+++ b/erts/emulator/beam/erl_printf_term.h
@@ -22,6 +22,5 @@
 #define ERL_PRINTF_TERM_H__
 
 #include "erl_printf_format.h"
-int erts_printf_term(fmtfn_t fn, void* arg, ErlPfEterm term, long precision,
-		     ErlPfEterm* term_base);
+int erts_printf_term(fmtfn_t fn, void* arg, ErlPfEterm term, long precision);
 #endif
diff --git a/erts/emulator/beam/erl_process.c b/erts/emulator/beam/erl_process.c
index 7b3d12ce09..c386d63a9e 100644
--- a/erts/emulator/beam/erl_process.c
+++ b/erts/emulator/beam/erl_process.c
@@ -58,11 +58,7 @@
 
 #define ERTS_PROC_MIN_CONTEXT_SWITCH_REDS_COST (CONTEXT_REDS/10)
 
-#ifndef ERTS_SCHED_MIN_SPIN
 #define ERTS_SCHED_SPIN_UNTIL_YIELD 100
-#else
-#define ERTS_SCHED_SPIN_UNTIL_YIELD 1
-#endif
 
 #define ERTS_SCHED_SYS_SLEEP_SPINCOUNT_VERY_LONG 40
 #define ERTS_SCHED_AUX_WORK_SLEEP_SPINCOUNT_FACT_VERY_LONG 1000
@@ -119,7 +115,7 @@
 #define RUNQ_SET_RQ(X, RQ) erts_smp_atomic_set_nob((X), (erts_aint_t) (RQ))
 
 #ifdef DEBUG
-#  if defined(ARCH_64) && !HALFWORD_HEAP
+#  if defined(ARCH_64)
 #    define ERTS_DBG_SET_INVALID_RUNQP(RQP, N) \
     (RUNQ_SET_RQ((RQP), (0xdeadbeefdead0003LL | ((N) << 4)))
 #  define ERTS_DBG_VERIFY_VALID_RUNQP(RQP) \
@@ -152,12 +148,8 @@ extern BeamInstr beam_apply[];
 extern BeamInstr beam_exit[];
 extern BeamInstr beam_continue_exit[];
 
-#ifdef __OSE__
-/* Eager check I/O not supported on OSE yet. */
-int erts_eager_check_io = 0;
-#else
+int erts_default_spo_flags = 0;
 int erts_eager_check_io = 1;
-#endif
 int erts_sched_compact_load;
 int erts_sched_balance_util = 0;
 Uint erts_no_schedulers;
@@ -343,7 +335,6 @@ static ErtsAlignedSchedulerSleepInfo *aligned_dirty_io_sched_sleep_info;
 
 static Uint last_reductions;
 static Uint last_exact_reductions;
-Uint erts_default_process_flags;
 Eterm erts_system_monitor;
 Eterm erts_system_monitor_long_gc;
 Uint erts_system_monitor_long_schedule;
@@ -360,7 +351,8 @@ struct erts_system_profile_flags_t erts_system_profile_flags;
 
 typedef enum {
     ERTS_PSTT_GC,	/* Garbage Collect */
-    ERTS_PSTT_CPC	/* Check Process Code */
+    ERTS_PSTT_CPC,	/* Check Process Code */
+    ERTS_PSTT_COHMQ     /* Change off heap message queue */
 } ErtsProcSysTaskType;
 
 #define ERTS_MAX_PROC_SYS_TASK_ARGS 2
@@ -506,9 +498,6 @@ dbg_chk_aux_work_val(erts_aint32_t value)
 #if HAVE_ERTS_MSEG
     valid |= ERTS_SSI_AUX_WORK_MSEG_CACHE_CHECK;
 #endif
-#ifdef ERTS_SMP_SCHEDULERS_NEED_TO_CHECK_CHILDREN
-    valid |= ERTS_SSI_AUX_WORK_CHECK_CHILDREN;
-#endif
 #ifdef ERTS_SSI_AUX_WORK_REAP_PORTS
     valid |= ERTS_SSI_AUX_WORK_REAP_PORTS;
 #endif
@@ -595,8 +584,6 @@ erts_pre_init_process(void)
 	= "MISC_THR_PRGR";
     erts_aux_work_flag_descr[ERTS_SSI_AUX_WORK_MISC_IX]
 	= "MISC";
-    erts_aux_work_flag_descr[ERTS_SSI_AUX_WORK_CHECK_CHILDREN_IX]
-	= "CHECK_CHILDREN";
     erts_aux_work_flag_descr[ERTS_SSI_AUX_WORK_SET_TMO_IX]
 	= "SET_TMO";
     erts_aux_work_flag_descr[ERTS_SSI_AUX_WORK_MSEG_CACHE_CHECK_IX]
@@ -625,11 +612,6 @@ erts_pre_init_process(void)
      erts_psd_required_locks[ERTS_PSD_SCHED_ID].set_locks
 	 = ERTS_PSD_SCHED_ID_SET_LOCKS;
 
-     erts_psd_required_locks[ERTS_PSD_DIST_ENTRY].get_locks
-	 = ERTS_PSD_DIST_ENTRY_GET_LOCKS;
-     erts_psd_required_locks[ERTS_PSD_DIST_ENTRY].set_locks
-	 = ERTS_PSD_DIST_ENTRY_SET_LOCKS;
-
      erts_psd_required_locks[ERTS_PSD_CALL_TIME_BP].get_locks
 	 = ERTS_PSD_CALL_TIME_BP_GET_LOCKS;
      erts_psd_required_locks[ERTS_PSD_CALL_TIME_BP].set_locks
@@ -689,7 +671,6 @@ erts_init_process(int ncpu, int proc_tab_size, int legacy_proc_tab)
 
     last_reductions = 0;
     last_exact_reductions = 0;
-    erts_default_process_flags = 0;
 }
 
 void
@@ -971,25 +952,10 @@ typedef struct {
     erts_smp_atomic32_t refc;
 } ErtsSchedWallTimeReq;
 
-#if !HALFWORD_HEAP
 ERTS_SCHED_PREF_QUICK_ALLOC_IMPL(swtreq,
-				 ErtsSchedWallTimeReq,
-				 5,
-				 ERTS_ALC_T_SCHED_WTIME_REQ)
-#else
-static ERTS_INLINE ErtsSchedWallTimeReq *
-swtreq_alloc(void)
-{
-    return erts_alloc(ERTS_ALC_T_SCHED_WTIME_REQ,
-		      sizeof(ErtsSchedWallTimeReq));
-}
-
-static ERTS_INLINE void
-swtreq_free(ErtsSchedWallTimeReq *ptr)
-{
-    erts_free(ERTS_ALC_T_SCHED_WTIME_REQ, ptr);
-}
-#endif
+                                 ErtsSchedWallTimeReq,
+                                 5,
+                                 ERTS_ALC_T_SCHED_WTIME_REQ)
 
 static void
 reply_sched_wall_time(void *vswtrp)
@@ -1006,7 +972,7 @@ reply_sched_wall_time(void *vswtrp)
     Eterm **hpp;
     Uint sz, *szp;
     ErlOffHeap *ohp = NULL;
-    ErlHeapFragment *bp = NULL;
+    ErtsMessage *mp = NULL;
 
     ASSERT(esdp);
 #ifdef ERTS_DIRTY_SCHEDULERS
@@ -1062,12 +1028,12 @@ reply_sched_wall_time(void *vswtrp)
 	if (hpp)
 	    break;
 
-	hp = erts_alloc_message_heap(sz, &bp, &ohp, rp, &rp_locks);
+	mp = erts_alloc_message_heap(rp, &rp_locks, sz, &hp, &ohp);
 	szp = NULL;
 	hpp = &hp;
     }
 
-    erts_queue_message(rp, &rp_locks, bp, msg, NIL);
+    erts_queue_message(rp, &rp_locks, mp, msg, NIL);
 
     if (swtrp->req_sched == esdp->no)
 	rp_locks &= ~ERTS_PROC_LOCK_MAIN;
@@ -2124,34 +2090,6 @@ erts_debug_wait_completed(Process *c_p, int flags)
 }
 
 
-#ifdef ERTS_SMP_SCHEDULERS_NEED_TO_CHECK_CHILDREN
-void
-erts_smp_notify_check_children_needed(void)
-{
-    int i;
-    for (i = 0; i < erts_no_schedulers; i++)
-	set_aux_work_flags_wakeup_nob(ERTS_SCHED_SLEEP_INFO_IX(i),
-				      ERTS_SSI_AUX_WORK_CHECK_CHILDREN);
-#ifdef ERTS_DIRTY_SCHEDULERS
-    for (i = 0; i < erts_no_dirty_cpu_schedulers; i++)
-	set_aux_work_flags_wakeup_nob(ERTS_DIRTY_CPU_SCHED_SLEEP_INFO_IX(i),
-				      ERTS_SSI_AUX_WORK_CHECK_CHILDREN);
-    for (i = 0; i < erts_no_dirty_io_schedulers; i++)
-	set_aux_work_flags_wakeup_nob(ERTS_DIRTY_IO_SCHED_SLEEP_INFO_IX(i),
-				      ERTS_SSI_AUX_WORK_CHECK_CHILDREN);
-#endif
-}
-
-static ERTS_INLINE erts_aint32_t
-handle_check_children(ErtsAuxWorkData *awdp, erts_aint32_t aux_work, int waiting)
-{
-    unset_aux_work_flags(awdp->ssi, ERTS_SSI_AUX_WORK_CHECK_CHILDREN);
-    erts_check_children();
-    return aux_work & ~ERTS_SSI_AUX_WORK_CHECK_CHILDREN;
-}
-
-#endif
-
 static void
 notify_reap_ports_relb(void)
 {
@@ -2244,6 +2182,7 @@ handle_aux_work(ErtsAuxWorkData *awdp, erts_aint32_t orig_aux_work, int waiting)
     erts_aint32_t aux_work = orig_aux_work;
     erts_aint32_t ignore = 0;
 
+    ASSERT(!awdp->esdp || !ERTS_SCHEDULER_IS_DIRTY(awdp->esdp));
 #ifdef ERTS_SMP
     haw_thr_prgr_current_reset(awdp);
 #endif
@@ -2305,10 +2244,6 @@ handle_aux_work(ErtsAuxWorkData *awdp, erts_aint32_t orig_aux_work, int waiting)
     HANDLE_AUX_WORK(ERTS_SSI_AUX_WORK_MISC,
 		    handle_misc_aux_work);
 
-#ifdef ERTS_SMP_SCHEDULERS_NEED_TO_CHECK_CHILDREN
-    HANDLE_AUX_WORK(ERTS_SSI_AUX_WORK_CHECK_CHILDREN,
-		    handle_check_children);
-#endif
 
     HANDLE_AUX_WORK(ERTS_SSI_AUX_WORK_SET_TMO,
 		    handle_setup_aux_work_timer);
@@ -2536,19 +2471,10 @@ try_set_sys_scheduling(void)
 #endif
 
 static ERTS_INLINE int
-prepare_for_sys_schedule(ErtsSchedulerData *esdp, int non_blocking)
+prepare_for_sys_schedule(int non_blocking)
 {
     if (non_blocking && erts_eager_check_io) {
 #ifdef ERTS_SMP
-#ifdef ERTS_SCHED_ONLY_POLL_SCHED_1
-	if (esdp->no != 1) {
-	    /* If we are not scheduler 1 and ERTS_SCHED_ONLY_POLL_SCHED_1 is used
-	       then we make sure to wake scheduler 1 */
-	    ErtsRunQueue *rq = ERTS_RUNQ_IX(0);
-	    wake_scheduler(rq);
-	    return 0;
-	}
-#endif
 	return try_set_sys_scheduling();
 #else
 	return 1;
@@ -2558,16 +2484,6 @@ prepare_for_sys_schedule(ErtsSchedulerData *esdp, int non_blocking)
 #ifdef ERTS_SMP
 	while (!erts_port_task_have_outstanding_io_tasks()
 	       && try_set_sys_scheduling()) {
-#ifdef ERTS_SCHED_ONLY_POLL_SCHED_1
-	    if (esdp->no != 1) {
-		/* If we are not scheduler 1 and ERTS_SCHED_ONLY_POLL_SCHED_1 is used
-		   then we make sure to wake scheduler 1 */
-		ErtsRunQueue *rq = ERTS_RUNQ_IX(0);
-		clear_sys_scheduling();
-		wake_scheduler(rq);
-		return 0;
-	    }
-#endif
 	    if (!erts_port_task_have_outstanding_io_tasks())
 		return 1;
 	    clear_sys_scheduling();
@@ -2891,8 +2807,6 @@ aux_thread(void *unused)
 		erts_thr_progress_active(NULL, thr_prgr_active = 0);
 	    erts_thr_progress_prepare_wait(NULL);
 
-	    ERTS_SCHED_FAIR_YIELD();
-
 	    flgs = sched_spin_wait(ssi, 0);
 
 	    if (flgs & ERTS_SSI_FLG_SLEEPING) {
@@ -2960,7 +2874,7 @@ scheduler_wait(int *fcalls, ErtsSchedulerData *esdp, ErtsRunQueue *rq)
      * be waiting in erl_sys_schedule()
      */
 
-    if (ERTS_SCHEDULER_IS_DIRTY(esdp) || !prepare_for_sys_schedule(esdp, 0)) {
+    if (ERTS_SCHEDULER_IS_DIRTY(esdp) || !prepare_for_sys_schedule(0)) {
 
 	sched_waiting(esdp->no, rq);
 
@@ -2977,14 +2891,13 @@ scheduler_wait(int *fcalls, ErtsSchedulerData *esdp, ErtsRunQueue *rq)
 	    ErtsMonotonicTime current_time;
 
 	    aux_work = erts_atomic32_read_acqb(&ssi->aux_work);
-	    if (aux_work) {
-		if (!ERTS_SCHEDULER_IS_DIRTY(esdp) && !thr_prgr_active) {
+	    if (aux_work && !ERTS_SCHEDULER_IS_DIRTY(esdp)) {
+		if (!thr_prgr_active) {
 		    erts_thr_progress_active(esdp, thr_prgr_active = 1);
 		    sched_wall_time_change(esdp, 1);
 		}
 		aux_work = handle_aux_work(&esdp->aux_work_data, aux_work, 1);
-		if (aux_work && !ERTS_SCHEDULER_IS_DIRTY(esdp)
-		    && erts_thr_progress_update(esdp))
+		if (aux_work && erts_thr_progress_update(esdp))
 		    erts_thr_progress_leader_update(esdp);
 	    }
 
@@ -3025,8 +2938,6 @@ scheduler_wait(int *fcalls, ErtsSchedulerData *esdp, ErtsRunQueue *rq)
 			erts_thr_progress_prepare_wait(esdp);
 		    }
 
-		    ERTS_SCHED_FAIR_YIELD();
-
 		    flgs = sched_spin_wait(ssi, spincount);
 		    if (flgs & ERTS_SSI_FLG_SLEEPING) {
 			ASSERT(flgs & ERTS_SSI_FLG_WAITING);
@@ -3097,13 +3008,8 @@ scheduler_wait(int *fcalls, ErtsSchedulerData *esdp, ErtsRunQueue *rq)
 #ifdef ERTS_DIRTY_SCHEDULERS
 	ASSERT(!ERTS_SCHEDULER_IS_DIRTY(esdp));
 #endif
-
-#ifdef ERTS_SCHED_ONLY_POLL_SCHED_1
-	ASSERT(esdp->no == 1);
-#endif
 	sched_waiting_sys(esdp->no, rq);
 
-
 	erts_smp_runq_unlock(rq);
 
 	ASSERT(working);
@@ -3136,25 +3042,22 @@ scheduler_wait(int *fcalls, ErtsSchedulerData *esdp, ErtsRunQueue *rq)
 #endif
 
 	    aux_work = erts_atomic32_read_acqb(&ssi->aux_work);
-	    if (aux_work) {
-		if (!ERTS_SCHEDULER_IS_DIRTY(esdp)) {
-		    if (!working)
-			sched_wall_time_change(esdp, working = 1);
+	    if (aux_work && !ERTS_SCHEDULER_IS_DIRTY(esdp)) {
+		if (!working)
+		    sched_wall_time_change(esdp, working = 1);
 #ifdef ERTS_SMP
-		    if (!thr_prgr_active)
-			erts_thr_progress_active(esdp, thr_prgr_active = 1);
+		if (!thr_prgr_active)
+		    erts_thr_progress_active(esdp, thr_prgr_active = 1);
 #endif
-		}
 		aux_work = handle_aux_work(&esdp->aux_work_data, aux_work, 1);
 #ifdef ERTS_SMP
-		if (!ERTS_SCHEDULER_IS_DIRTY(esdp) && aux_work &&
-		    erts_thr_progress_update(esdp))
+		if (aux_work && erts_thr_progress_update(esdp))
 		    erts_thr_progress_leader_update(esdp);
 #endif
 	    }
 
 #ifndef ERTS_SMP
-	    if (rq->len != 0 || rq->misc.start)
+	    if (erts_smp_atomic32_read_dirty(&rq->len) != 0 || rq->misc.start)
 		goto sys_woken;
 #else
 	    flgs = erts_smp_atomic32_read_acqb(&ssi->flags);
@@ -3173,7 +3076,7 @@ scheduler_wait(int *fcalls, ErtsSchedulerData *esdp, ErtsRunQueue *rq)
 		 * Got to check that we still got I/O tasks; otherwise
 		 * we have to continue checking for I/O...
 		 */
-		if (!prepare_for_sys_schedule(esdp, 0)) {
+		if (!prepare_for_sys_schedule(0)) {
 		    spincount *= ERTS_SCHED_TSE_SLEEP_SPINCOUNT_FACT;
 		    goto tse_wait;
 		}
@@ -3195,7 +3098,7 @@ scheduler_wait(int *fcalls, ErtsSchedulerData *esdp, ErtsRunQueue *rq)
 	     * Got to check that we still got I/O tasks; otherwise
 	     * we have to wait in erl_sys_schedule() after all...
 	     */
-	    if (!prepare_for_sys_schedule(esdp, 0)) {
+	    if (!prepare_for_sys_schedule(0)) {
 		/*
 		 * Not allowed to wait in erl_sys_schedule;
 		 * do tse wait instead...
@@ -3253,7 +3156,7 @@ scheduler_wait(int *fcalls, ErtsSchedulerData *esdp, ErtsRunQueue *rq)
 	}
 
 #ifndef ERTS_SMP
-	if (rq->len == 0 && !rq->misc.start)
+	if (erts_smp_atomic32_read_dirty(&rq->len) == 0 && !rq->misc.start)
 	    goto sys_aux_work;
     sys_woken:
 #else
@@ -4970,7 +4873,7 @@ erts_fprintf(stderr, "--------------------------------\n");
 	rq->out_of_work_count = 0;
 	(void) ERTS_RUNQ_FLGS_READ_BSET(rq, ERTS_RUNQ_FLGS_MIGRATION_INFO, flags);
 
-	rq->max_len = rq->len;
+	rq->max_len = erts_smp_atomic32_read_dirty(&rq->len);
 	for (pix = 0; pix < ERTS_NO_PRIO_LEVELS; pix++) {
 	    ErtsRunQueueInfo *rqi;
 	    rqi = (pix == ERTS_PORT_PRIO_LEVEL
@@ -5128,7 +5031,7 @@ wakeup_other_check(ErtsRunQueue *rq, Uint32 flags)
 {
     int wo_reds = rq->wakeup_other_reds;
     if (wo_reds) {
-	int left_len = rq->len - 1;
+	int left_len = erts_smp_atomic32_read_dirty(&rq->len) - 1;
 	if (left_len < 1) {
 	    int wo_reduce = wo_reds << wakeup_other.dec_shift;
 	    wo_reduce &= wakeup_other.dec_mask;
@@ -5201,7 +5104,7 @@ wakeup_other_check_legacy(ErtsRunQueue *rq, Uint32 flags)
 {
     int wo_reds = rq->wakeup_other_reds;
     if (wo_reds) {
-	erts_aint32_t len = rq->len;
+	erts_aint32_t len = erts_smp_atomic32_read_dirty(&rq->len);
 	if (len < 2) {
 	    rq->wakeup_other -= ERTS_WAKEUP_OTHER_DEC_LEGACY*wo_reds;
 	    if (rq->wakeup_other < 0)
@@ -5297,7 +5200,7 @@ runq_supervisor(void *unused)
 	    ErtsRunQueue *rq = ERTS_RUNQ_IX(ix);
 	    if (ERTS_RUNQ_FLGS_GET(rq) & ERTS_RUNQ_FLG_NONEMPTY) {
 		erts_smp_runq_lock(rq);
-		if (rq->len != 0)
+		if (erts_smp_atomic32_read_dirty(&rq->len) != 0)
 		    wake_scheduler_on_empty_runq(rq); /* forced wakeup... */
 		erts_smp_runq_unlock(rq);
 	    }
@@ -5316,17 +5219,11 @@ erts_early_init_scheduling(int no_schedulers)
     wakeup_other.threshold = ERTS_SCHED_WAKEUP_OTHER_THRESHOLD_MEDIUM;
     wakeup_other.type = ERTS_SCHED_WAKEUP_OTHER_TYPE_DEFAULT;
 #endif
-#ifndef ERTS_SCHED_MIN_SPIN
     sched_busy_wait.sys_schedule = ERTS_SCHED_SYS_SLEEP_SPINCOUNT_MEDIUM;
     sched_busy_wait.tse = (ERTS_SCHED_SYS_SLEEP_SPINCOUNT_MEDIUM
 			   * ERTS_SCHED_TSE_SLEEP_SPINCOUNT_FACT);
     sched_busy_wait.aux_work = (ERTS_SCHED_SYS_SLEEP_SPINCOUNT_MEDIUM
 				* ERTS_SCHED_AUX_WORK_SLEEP_SPINCOUNT_FACT_MEDIUM);
-#else
-    sched_busy_wait.sys_schedule = ERTS_SCHED_SYS_SLEEP_SPINCOUNT_NONE;
-    sched_busy_wait.tse = ERTS_SCHED_SYS_SLEEP_SPINCOUNT_NONE;
-    sched_busy_wait.aux_work = ERTS_SCHED_SYS_SLEEP_SPINCOUNT_NONE;
-#endif
 }
 
 int
@@ -5510,9 +5407,6 @@ init_scheduler_data(ErtsSchedulerData* esdp, int num,
     esdp->f_reg_array =
 	erts_alloc_permanent_cache_aligned(ERTS_ALC_T_BEAM_REGISTER,
 					   MAX_REG * sizeof(FloatDef));
-#if !HEAP_ON_C_STACK
-    esdp->num_tmp_heap_used = 0;
-#endif
 #ifdef ERTS_DIRTY_SCHEDULERS
     if (ERTS_RUNQ_IX_IS_DIRTY(runq->ix)) {
 	esdp->no = 0;
@@ -5647,7 +5541,7 @@ erts_init_scheduling(int no_schedulers, int no_schedulers_online
 	}
 	rq->out_of_work_count = 0;
 	rq->max_len = 0;
-	rq->len = 0;
+	erts_smp_atomic32_set_nob(&rq->len, 0);
 	rq->wakeup_other = 0;
 	rq->wakeup_other_reds = 0;
 	rq->halt_in_progress = 0;
@@ -5797,9 +5691,7 @@ erts_init_scheduling(int no_schedulers, int no_schedulers_online
 #endif
 
     init_misc_aux_work();
-#if !HALFWORD_HEAP
     init_swtreq_alloc();
-#endif
 
     erts_atomic32_init_nob(&debug_wait_completed_count, 0); /* debug only */
     debug_wait_completed_flags = 0;
@@ -6357,22 +6249,99 @@ erts_schedule_process(Process *p, erts_aint32_t state, ErtsProcLocks locks)
     schedule_process(p, state, locks);
 }
 
-static void
-schedule_process_sys_task(Process *p, erts_aint32_t state, Process *proxy)
+static int
+schedule_process_sys_task(Process *p, erts_aint32_t prio, ErtsProcSysTask *st)
 {
-    /*
-     * Expects status lock to be locked when called, and
-     * returns with status lock unlocked...
-     */
-    erts_aint32_t a = state, n, enq_prio = -1;
+    int res;
+    int locked;
+    ErtsProcSysTaskQs *stqs, *free_stqs;
+    erts_aint32_t state, a, n, enq_prio;
     int enqueue; /* < 0 -> use proxy */
-    unsigned int prof_runnable_procs = erts_system_profile_flags.runnable_procs;
+    unsigned int prof_runnable_procs;
+
+    res = 1; /* prepare for success */
+    st->next = st->prev = st; /* Prep for empty prio queue */
+    state = erts_smp_atomic32_read_nob(&p->state);
+    prof_runnable_procs = erts_system_profile_flags.runnable_procs;
+    locked = 0;
+    free_stqs = NULL;
+    if (state & ERTS_PSFLG_ACTIVE_SYS)
+	stqs = NULL;
+    else {
+    alloc_qs:
+	stqs = proc_sys_task_queues_alloc();
+	stqs->qmask = 1 << prio;
+	stqs->ncount = 0;
+	stqs->q[PRIORITY_MAX] = NULL;
+	stqs->q[PRIORITY_HIGH] = NULL;
+	stqs->q[PRIORITY_NORMAL] = NULL;
+	stqs->q[PRIORITY_LOW] = NULL;
+	stqs->q[prio] = st;
+    }
+
+    if (!locked) {
+	locked = 1;
+	erts_smp_proc_lock(p, ERTS_PROC_LOCK_STATUS);
+
+	state = erts_smp_atomic32_read_nob(&p->state);
+	if (state & ERTS_PSFLG_EXITING) {
+	    free_stqs = stqs;
+	    res = 0;
+	    goto cleanup;
+	}
+    }
+
+    if (!p->sys_task_qs) {
+	if (stqs)
+	    p->sys_task_qs = stqs;
+	else
+	    goto alloc_qs;
+    }
+    else {
+	free_stqs = stqs;
+	stqs = p->sys_task_qs;
+	if (!stqs->q[prio]) {
+	    stqs->q[prio] = st;
+	    stqs->qmask |= 1 << prio;
+	}
+	else {
+	    st->next = stqs->q[prio];
+	    st->prev = stqs->q[prio]->prev;
+	    st->next->prev = st;
+	    st->prev->next = st;
+	    ASSERT(stqs->qmask & (1 << prio));
+	}
+    }
+
+    if (ERTS_PSFLGS_GET_ACT_PRIO(state) > prio) {
+	erts_aint32_t n, a, e;
+	/* Need to elevate actual prio */
+
+	a = state;
+	do {
+	    if (ERTS_PSFLGS_GET_ACT_PRIO(a) <= prio) {
+		n = a;
+		break;
+	    }
+	    n = e = a;
+	    n &= ~ERTS_PSFLGS_ACT_PRIO_MASK;
+	    n |= (prio << ERTS_PSFLGS_ACT_PRIO_OFFSET);
+	    a = erts_smp_atomic32_cmpxchg_nob(&p->state, n, e);    
+	} while (a != e);
+	state = n;
+    }
+
+
+    a = state;
+    enq_prio = -1;
 
     /* Status lock prevents out of order "runnable proc" trace msgs */
     ERTS_SMP_LC_ASSERT(ERTS_PROC_LOCK_STATUS & erts_proc_lc_my_proc_locks(p));
 
-    if (!prof_runnable_procs)
+    if (!prof_runnable_procs) {
 	erts_smp_proc_unlock(p, ERTS_PROC_LOCK_STATUS);
+	locked = 0;
+    }
 
     ASSERT(!(state & ERTS_PSFLG_PROXY));
 
@@ -6380,8 +6349,10 @@ schedule_process_sys_task(Process *p, erts_aint32_t state, Process *proxy)
 	erts_aint32_t e;
 	n = e = a;
 
-	if (a & ERTS_PSFLG_FREE)
+	if (a & ERTS_PSFLG_FREE) {
+	    res = 0;
 	    goto cleanup; /* We don't want to schedule free processes... */
+	}
 
 	enqueue = ERTS_ENQUEUE_NOT;
 	n |= ERTS_PSFLG_ACTIVE_SYS;
@@ -6405,29 +6376,24 @@ schedule_process_sys_task(Process *p, erts_aint32_t state, Process *proxy)
 	}
 
 	erts_smp_proc_unlock(p, ERTS_PROC_LOCK_STATUS);
-	prof_runnable_procs = 0;
+	locked = 0;
     }
 
-    if (enqueue != ERTS_ENQUEUE_NOT) {
-	Process *sched_p;
-	if (enqueue > 0)
-	    sched_p = p;
-	else {
-	    sched_p = make_proxy_proc(proxy, p, enq_prio);
-	    proxy = NULL;
-	}
-	add2runq(sched_p, n, enq_prio);
-    }
+    if (enqueue != ERTS_ENQUEUE_NOT)
+	add2runq(enqueue > 0 ? p : make_proxy_proc(NULL, p, enq_prio),
+		 n, enq_prio);
 
 cleanup:
 
-    if (prof_runnable_procs)
+    if (locked)
 	erts_smp_proc_unlock(p, ERTS_PROC_LOCK_STATUS);
 
-    if (proxy)
-	free_proxy_proc(proxy);
+    if (free_stqs)
+	proc_sys_task_queues_free(free_stqs);
 
     ERTS_SMP_LC_ASSERT(!(ERTS_PROC_LOCK_STATUS & erts_proc_lc_my_proc_locks(p)));
+
+    return res;
 }
 
 static ERTS_INLINE int
@@ -6803,18 +6769,19 @@ suspend_scheduler(ErtsSchedulerData *esdp)
 			 & ERTS_RUNQ_FLGS_QMASK);
 		aux_work = erts_atomic32_read_acqb(&ssi->aux_work);
 		if (aux_work|qmask) {
-		    if (!ERTS_SCHEDULER_IS_DIRTY(esdp) && !thr_prgr_active) {
-			erts_thr_progress_active(esdp, thr_prgr_active = 1);
-			sched_wall_time_change(esdp, 1);
+		    if (!ERTS_SCHEDULER_IS_DIRTY(esdp)) {
+			if (!thr_prgr_active) {
+			    erts_thr_progress_active(esdp, thr_prgr_active = 1);
+			    sched_wall_time_change(esdp, 1);
+			}
+			if (aux_work)
+			    aux_work = handle_aux_work(&esdp->aux_work_data,
+						       aux_work,
+						       1);
+
+			if (aux_work && erts_thr_progress_update(esdp))
+			    erts_thr_progress_leader_update(esdp);
 		    }
-		    if (aux_work)
-			aux_work = handle_aux_work(&esdp->aux_work_data,
-						   aux_work,
-						   1);
-
-		    if (!ERTS_SCHEDULER_IS_DIRTY(esdp) &&
-			(aux_work && erts_thr_progress_update(esdp)))
-			erts_thr_progress_leader_update(esdp);
 		    if (qmask) {
 #ifdef ERTS_DIRTY_SCHEDULERS
 			if (ERTS_SCHEDULER_IS_DIRTY(esdp)) {
@@ -7031,17 +6998,18 @@ suspend_scheduler(ErtsSchedulerData *esdp)
 			 & ERTS_RUNQ_FLGS_QMASK);
 		aux_work = erts_atomic32_read_acqb(&ssi->aux_work);
 		if (aux_work|qmask) {
-		    if (!ERTS_SCHEDULER_IS_DIRTY(esdp) && !thr_prgr_active) {
-			erts_thr_progress_active(esdp, thr_prgr_active = 1);
-			sched_wall_time_change(esdp, 1);
+		    if (!ERTS_SCHEDULER_IS_DIRTY(esdp)) {
+			if (!thr_prgr_active) {
+			    erts_thr_progress_active(esdp, thr_prgr_active = 1);
+			    sched_wall_time_change(esdp, 1);
+			}
+			if (aux_work)
+			    aux_work = handle_aux_work(&esdp->aux_work_data,
+						       aux_work,
+						       1);
+			if (aux_work && erts_thr_progress_update(esdp))
+			    erts_thr_progress_leader_update(esdp);
 		    }
-		    if (aux_work)
-			aux_work = handle_aux_work(&esdp->aux_work_data,
-						   aux_work,
-						   1);
-		    if (!ERTS_SCHEDULER_IS_DIRTY(esdp) && aux_work &&
-			erts_thr_progress_update(esdp))
-			erts_thr_progress_leader_update(esdp);
 		    if (qmask) {
 			erts_smp_runq_lock(esdp->run_queue);
 			evacuate_run_queue(esdp->run_queue, &sbp);
@@ -7938,11 +7906,19 @@ sched_thread_func(void *vesdp)
     ErtsThrPrgrCallbacks callbacks;
     ErtsSchedulerData *esdp = vesdp;
     Uint no = esdp->no;
+#ifdef ERTS_SMP
+    erts_tse_t *tse;
+#endif
 
     erts_sched_init_time_sup(esdp);
 
+    (void) ERTS_RUNQ_FLGS_SET_NOB(esdp->run_queue,
+				  ERTS_RUNQ_FLG_EXEC);
+
 #ifdef ERTS_SMP
-    ERTS_SCHED_SLEEP_INFO_IX(no - 1)->event = erts_tse_fetch();
+    tse = erts_tse_fetch();
+    erts_tse_prepare_timed(tse);
+    ERTS_SCHED_SLEEP_INFO_IX(no - 1)->event = tse;
     callbacks.arg = (void *) esdp->ssi;
     callbacks.wakeup = thr_prgr_wakeup;
     callbacks.prepare_wait = thr_prgr_prep_wait;
@@ -8204,18 +8180,12 @@ erts_start_schedulers(void)
 
 	erts_snprintf(opts.name, 16, "%lu_scheduler", actual + 1);
 
-#ifdef __OSE__
-        /* This should be done in the bind strategy */
-	opts.coreNo = (actual+1) % ose_num_cpus();
-#endif
-
 	res = ethr_thr_create(&esdp->tid, sched_thread_func, (void*)esdp, &opts);
 
 	if (res != 0) {
            break;
 	}
     }
-
     erts_no_schedulers = actual;
 
 #ifdef ERTS_DIRTY_SCHEDULERS
@@ -8244,10 +8214,6 @@ erts_start_schedulers(void)
 
     erts_snprintf(opts.name, 16, "aux");
 
-#ifdef __OSE__
-    opts.coreNo = 0;
-#endif /* __OSE__ */
-
     res = ethr_thr_create(&aux_tid, aux_thread, NULL, &opts);
     if (res != 0)
 	erl_exit(1, "Failed to create aux thread\n");
@@ -8267,7 +8233,6 @@ erts_start_schedulers(void)
 		      actual, actual == 1 ? " was" : "s were");
 	erts_send_error_to_logger_nogl(dsbufp);
     }
-
 }
 
 #endif /* ERTS_SMP */
@@ -8286,7 +8251,7 @@ add_pend_suspend(Process *suspendee,
 					 sizeof(ErtsPendingSuspend));
     psp->next = NULL;
 #ifdef DEBUG
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
     psp->end = (ErtsPendingSuspend *) 0xdeaddeaddeaddead;
 #else
     psp->end = (ErtsPendingSuspend *) 0xdeaddead;
@@ -8947,24 +8912,39 @@ resume_process_1(BIF_ALIST_1)
 }
 
 Uint
-erts_run_queues_len(Uint *qlen)
+erts_run_queues_len(Uint *qlen, int atomic_queues_read, int incl_active_sched)
 {
     int i = 0;
     Uint len = 0;
-    ERTS_ATOMIC_FOREACH_RUNQ(rq,
-    {
-	Sint pqlen = 0;
-	int pix;
-	for (pix = 0; pix < ERTS_NO_PROC_PRIO_LEVELS; pix++)
-	    pqlen += RUNQ_READ_LEN(&rq->procs.prio_info[pix].len);
+    if (atomic_queues_read)
+	ERTS_ATOMIC_FOREACH_RUNQ(rq,
+	 {
+	     Sint rq_len = (Sint) erts_smp_atomic32_read_dirty(&rq->len);
+	     ASSERT(rq_len >= 0);
+	     if (incl_active_sched
+		 && (ERTS_RUNQ_FLGS_GET_NOB(rq) & ERTS_RUNQ_FLG_EXEC)) {
+		 rq_len++;
+	     }
+	     if (qlen)
+		 qlen[i++] = rq_len;
+	     len += (Uint) rq_len;
+	 }
+	    );
+    else {
+	for (i = 0; i < erts_no_run_queues; i++) {
+	    ErtsRunQueue *rq = ERTS_RUNQ_IX(i);
+	    Sint rq_len = (Sint) erts_smp_atomic32_read_nob(&rq->len);
+	    ASSERT(rq_len >= 0);
+	     if (incl_active_sched
+		 && (ERTS_RUNQ_FLGS_GET_NOB(rq) & ERTS_RUNQ_FLG_EXEC)) {
+		 rq_len++;
+	     }
+	    if (qlen)
+		qlen[i] = rq_len;
+	    len += (Uint) rq_len;
+	}
 
-	if (pqlen < 0)
-	    pqlen = 0;
- 	if (qlen)
-	    qlen[i++] = pqlen;
-	len += pqlen;
     }
-	);
     return len;
 }
 
@@ -9177,6 +9157,10 @@ erts_set_process_priority(Process *p, Eterm value)
 
 	    a = erts_smp_atomic32_cmpxchg_mb(&p->state, n, e);
 	} while (a != e);
+
+        if (slocked)
+            erts_smp_proc_unlock(p, ERTS_PROC_LOCK_STATUS);
+
     }
 
     switch (oprio) {
@@ -9253,6 +9237,8 @@ Process *schedule(Process *p, int calls)
     } else {
     sched_out_proc:
 
+	ASSERT(!(p->flags & F_DELAY_GC));
+
 #ifdef ERTS_SMP
 	ERTS_SMP_CHK_HAVE_ONLY_MAIN_PROC_LOCK(p);
 	esdp = p->scheduler_data;
@@ -9387,8 +9373,10 @@ Process *schedule(Process *p, int calls)
 
 	if (flags & (ERTS_RUNQ_FLG_CHK_CPU_BIND|ERTS_RUNQ_FLG_SUSPENDED)) {
 	    if (flags & ERTS_RUNQ_FLG_SUSPENDED) {
+		(void) ERTS_RUNQ_FLGS_UNSET_NOB(rq, ERTS_RUNQ_FLG_EXEC);
 		suspend_scheduler(esdp);
-		flags = ERTS_RUNQ_FLGS_GET_NOB(rq);
+		flags = ERTS_RUNQ_FLGS_SET_NOB(rq, ERTS_RUNQ_FLG_EXEC);
+		flags |= ERTS_RUNQ_FLG_EXEC;
 	    }
 	    if (flags & ERTS_RUNQ_FLG_CHK_CPU_BIND) {
 		flags = ERTS_RUNQ_FLGS_UNSET(rq, ERTS_RUNQ_FLG_CHK_CPU_BIND);
@@ -9403,24 +9391,20 @@ Process *schedule(Process *p, int calls)
 	    suspend_scheduler(esdp);
 #endif
 
-	{
+	if (!ERTS_SCHEDULER_IS_DIRTY(esdp)) {
 	    erts_aint32_t aux_work;
-	    int leader_update = ERTS_SCHEDULER_IS_DIRTY(esdp) ? 0
-		: erts_thr_progress_update(esdp);
+	    int leader_update = erts_thr_progress_update(esdp);
 	    aux_work = erts_atomic32_read_acqb(&esdp->ssi->aux_work);
-	    if (aux_work | leader_update | ERTS_SCHED_FAIR) {
+	    if (aux_work | leader_update) {
 		erts_smp_runq_unlock(rq);
 		if (leader_update)
 		    erts_thr_progress_leader_update(esdp);
-		else if (ERTS_SCHED_FAIR)
-		  ERTS_SCHED_FAIR_YIELD();
 		if (aux_work)
 		    handle_aux_work(&esdp->aux_work_data, aux_work, 0);
 		erts_smp_runq_lock(rq);
 	    }
 
-	    ERTS_SMP_LC_ASSERT(ERTS_SCHEDULER_IS_DIRTY(esdp)
-			       || !erts_thr_progress_is_blocking());
+	    ERTS_SMP_LC_ASSERT(!erts_thr_progress_is_blocking());
 	}
 	ERTS_SMP_LC_ASSERT(erts_smp_lc_runq_is_locked(rq));
 
@@ -9479,7 +9463,10 @@ Process *schedule(Process *p, int calls)
 	    }
 #endif
 
+	    (void) ERTS_RUNQ_FLGS_UNSET(rq, ERTS_RUNQ_FLG_EXEC);
 	    scheduler_wait(&fcalls, esdp, rq);
+	    flags = ERTS_RUNQ_FLGS_SET_NOB(rq, ERTS_RUNQ_FLG_EXEC);
+	    flags |= ERTS_RUNQ_FLG_EXEC;
 
 #ifdef ERTS_SMP
 	    non_empty_runq(rq);
@@ -9489,7 +9476,7 @@ Process *schedule(Process *p, int calls)
 	}
 	else if (!ERTS_SCHEDULER_IS_DIRTY(esdp) &&
 		 (fcalls > input_reductions &&
-		  prepare_for_sys_schedule(esdp, !0))) {
+		  prepare_for_sys_schedule(!0))) {
 	    ErtsMonotonicTime current_time;
 	    /*
 	     * Schedule system-level activities.
@@ -9768,13 +9755,13 @@ Process *schedule(Process *p, int calls)
 	    }
 	}
 
-	if (!(state & ERTS_PSFLG_EXITING)
-	    && ((FLAGS(p) & F_FORCE_GC)
-		|| (MSO(p).overhead > BIN_VHEAP_SZ(p)))) {
-	    reds -= erts_garbage_collect(p, 0, p->arg_reg, p->arity);
-	    if (reds <= 0) {
-		p->fcalls = reds;
-		goto sched_out_proc;
+	if (ERTS_IS_GC_DESIRED(p)) {
+	    if (!(state & ERTS_PSFLG_EXITING) && !(p->flags & F_DISABLE_GC)) {
+		reds -= erts_garbage_collect_nobump(p, 0, p->arg_reg, p->arity);
+		if (reds <= 0) {
+		    p->fcalls = reds;
+		    goto sched_out_proc;
+		}
 	    }
 	}
 
@@ -9814,7 +9801,7 @@ notify_sys_task_executed(Process *c_p, ErtsProcSysTask *st, Eterm st_result)
     if (rp) {
 	ErtsProcLocks rp_locks;
 	ErlOffHeap *ohp;
-	ErlHeapFragment* bp;
+	ErtsMessage *mp;
 	Eterm *hp, msg, req_id, result;
 	Uint st_result_sz, hsz;
 #ifdef DEBUG
@@ -9826,11 +9813,7 @@ notify_sys_task_executed(Process *c_p, ErtsProcSysTask *st, Eterm st_result)
 	st_result_sz = is_immed(st_result) ? 0 : size_object(st_result);
 	hsz = st->req_id_sz + st_result_sz + 4 /* 3-tuple */;
 
-	hp = erts_alloc_message_heap(hsz,
-				     &bp,
-				     &ohp,
-				     rp,
-				     &rp_locks);
+	mp = erts_alloc_message_heap(rp, &rp_locks, hsz, &hp, &ohp);
 
 #ifdef DEBUG
 	hp_start = hp;
@@ -9855,7 +9838,7 @@ notify_sys_task_executed(Process *c_p, ErtsProcSysTask *st, Eterm st_result)
 	ASSERT(hp_start + hsz == hp);
 #endif
 
-	erts_queue_message(rp, &rp_locks, bp, msg, NIL);
+	erts_queue_message(rp, &rp_locks, mp, msg, NIL);
 
 	if (c_p == rp)
 	    rp_locks &= ~ERTS_PROC_LOCK_MAIN;
@@ -10057,10 +10040,10 @@ execute_sys_tasks(Process *c_p, erts_aint32_t *statep, int in_reds)
 	    else {
 		if (!garbage_collected) {
 		    FLAGS(c_p) |= F_NEED_FULLSWEEP;
-		    reds += erts_garbage_collect(c_p,
-						 0,
-						 c_p->arg_reg,
-						 c_p->arity);
+		    reds += erts_garbage_collect_nobump(c_p,
+							0,
+							c_p->arg_reg,
+							c_p->arity);
 		    garbage_collected = 1;
 		}
 		st_res = am_true;
@@ -10069,7 +10052,7 @@ execute_sys_tasks(Process *c_p, erts_aint32_t *statep, int in_reds)
 	case ERTS_PSTT_CPC:
 	    st_res = erts_check_process_code(c_p,
 					     st->arg[0],
-					     st->arg[1] == am_true,
+                                             unsigned_val(st->arg[1]),
 					     &reds);
 	    if (is_non_value(st_res)) {
 		/* Needed gc, but gc was disabled */
@@ -10077,6 +10060,10 @@ execute_sys_tasks(Process *c_p, erts_aint32_t *statep, int in_reds)
 		st = NULL;
 	    }
 	    break;
+	case ERTS_PSTT_COHMQ:
+	    reds += erts_complete_off_heap_message_queue_change(c_p);
+	    st_res = am_true;
+	    break;
 	default:
 	    ERTS_INTERNAL_ERROR("Invalid process sys task type");
 	    st_res = am_false;
@@ -10119,6 +10106,9 @@ cleanup_sys_tasks(Process *c_p, erts_aint32_t in_state, int in_reds)
 	case ERTS_PSTT_CPC:
 	    st_res = am_false;
 	    break;
+	case ERTS_PSTT_COHMQ:
+	    st_res = am_false;
+	    break;
 	default:
 	    ERTS_INTERNAL_ERROR("Invalid process sys task type");
 	    st_res = am_false;
@@ -10137,10 +10127,8 @@ BIF_RETTYPE
 erts_internal_request_system_task_3(BIF_ALIST_3)
 {
     Process *rp = erts_proc_lookup(BIF_ARG_1);
-    ErtsProcSysTaskQs *stqs, *free_stqs = NULL;
     ErtsProcSysTask *st = NULL;
-    erts_aint32_t prio, rp_state;
-    int rp_locked;
+    erts_aint32_t prio;
     Eterm noproc_res, req_type;
 
     if (!rp && !is_internal_pid(BIF_ARG_1)) {
@@ -10197,7 +10185,6 @@ erts_internal_request_system_task_3(BIF_ALIST_3)
 	}
 	st = erts_alloc(ERTS_ALC_T_PROC_SYS_TSK,
 			ERTS_PROC_SYS_TASK_SIZE(tot_sz));
-	st->next = st->prev = st; /* Prep for empty prio queue */
 	ERTS_INIT_OFF_HEAP(&st->off_heap);
 	hp = &st->heap[0];
 
@@ -10229,7 +10216,7 @@ erts_internal_request_system_task_3(BIF_ALIST_3)
     case am_check_process_code:
 	if (is_not_atom(st->arg[0]))
 	    goto badarg;
-	if (st->arg[1] != am_true && st->arg[1] != am_false)
+	if (is_not_small(st->arg[1]) || (unsigned_val(st->arg[1]) & ~ERTS_CPC_ALL))
 	    goto badarg;
 	noproc_res = am_false;
 	st->type = ERTS_PSTT_CPC;
@@ -10241,95 +10228,11 @@ erts_internal_request_system_task_3(BIF_ALIST_3)
 	goto badarg;
     }
 
-    rp_state = erts_smp_atomic32_read_nob(&rp->state);
-
-    rp_locked = 0;
-
-    free_stqs = NULL;
-    if (rp_state & ERTS_PSFLG_ACTIVE_SYS)
-	stqs = NULL;
-    else {
-    alloc_qs:
-	stqs = proc_sys_task_queues_alloc();
-	stqs->qmask = 1 << prio;
-	stqs->ncount = 0;
-	stqs->q[PRIORITY_MAX] = NULL;
-	stqs->q[PRIORITY_HIGH] = NULL;
-	stqs->q[PRIORITY_NORMAL] = NULL;
-	stqs->q[PRIORITY_LOW] = NULL;
-	stqs->q[prio] = st;
-    }
-
-    if (!rp_locked) {
-	rp_locked = 1;
-	erts_smp_proc_lock(rp, ERTS_PROC_LOCK_STATUS);
-
-	rp_state = erts_smp_atomic32_read_nob(&rp->state);
-	if (rp_state & ERTS_PSFLG_EXITING) {
-	    erts_smp_proc_unlock(rp, ERTS_PROC_LOCK_STATUS);
-	    rp = NULL;
-	    free_stqs = stqs;
-	    goto noproc;
-	}
-    }
-
-    if (!rp->sys_task_qs) {
-	if (stqs)
-	    rp->sys_task_qs = stqs;
-	else
-	    goto alloc_qs;
-    }
-    else {
-	if (stqs)
-	    free_stqs = stqs;
-	stqs = rp->sys_task_qs;
-	if (!stqs->q[prio]) {
-	    stqs->q[prio] = st;
-	    stqs->qmask |= 1 << prio;
-	}
-	else {
-	    st->next = stqs->q[prio];
-	    st->prev = stqs->q[prio]->prev;
-	    st->next->prev = st;
-	    st->prev->next = st;
-	    ASSERT(stqs->qmask & (1 << prio));
-	}
-    }
-
-    if (ERTS_PSFLGS_GET_ACT_PRIO(rp_state) > prio) {
-	erts_aint32_t n, a, e;
-	/* Need to elevate actual prio */
-
-	a = rp_state;
-	do {
-	    if (ERTS_PSFLGS_GET_ACT_PRIO(a) <= prio) {
-		n = a;
-		break;
-	    }
-	    n = e = a;
-	    n &= ~ERTS_PSFLGS_ACT_PRIO_MASK;
-	    n |= (prio << ERTS_PSFLGS_ACT_PRIO_OFFSET);
-	    a = erts_smp_atomic32_cmpxchg_nob(&rp->state, n, e);    
-	} while (a != e);
-	rp_state = n;
+    if (!schedule_process_sys_task(rp, prio, st)) {
+    noproc:
+	notify_sys_task_executed(BIF_P, st, noproc_res);
     }
 
-    /*
-     * schedule_process_sys_task() unlocks status
-     * lock on process.
-     */
-    schedule_process_sys_task(rp, rp_state, NULL);
-
-    if (free_stqs)
-	proc_sys_task_queues_free(free_stqs);
-
-    BIF_RET(am_ok);
-
-noproc:
-
-    notify_sys_task_executed(BIF_P, st, noproc_res);
-    if (free_stqs)
-	proc_sys_task_queues_free(free_stqs);
     BIF_RET(am_ok);
 
 badarg:
@@ -10338,11 +10241,35 @@ badarg:
 	erts_cleanup_offheap(&st->off_heap);
 	erts_free(ERTS_ALC_T_PROC_SYS_TSK, st);
     }
-    if (free_stqs)
-	proc_sys_task_queues_free(free_stqs);
     BIF_ERROR(BIF_P, BADARG);
 }
 
+void
+erts_schedule_complete_off_heap_message_queue_change(Eterm pid)
+{
+    Process *rp = erts_proc_lookup(pid);
+    if (rp) {
+	ErtsProcSysTask *st;
+	erts_aint32_t state;
+	int i;
+
+	st = erts_alloc(ERTS_ALC_T_PROC_SYS_TSK,
+			ERTS_PROC_SYS_TASK_SIZE(0));
+	st->type = ERTS_PSTT_COHMQ;
+	st->requester = NIL;
+	st->reply_tag = NIL;
+	st->req_id = NIL;
+	st->req_id_sz = 0;
+	for (i = 0; i < ERTS_MAX_PROC_SYS_TASK_ARGS; i++)
+	    st->arg[i] = NIL;
+	ERTS_INIT_OFF_HEAP(&st->off_heap);
+	state = erts_smp_atomic32_read_nob(&rp->state);
+
+	if (!schedule_process_sys_task(rp, ERTS_PSFLGS_GET_USR_PRIO(state), st))
+	    erts_free(ERTS_ALC_T_PROC_SYS_TSK, st);
+    }
+}
+
 static void
 save_gc_task(Process *c_p, ErtsProcSysTask *st, int prio)
 {
@@ -10788,6 +10715,7 @@ erl_create_process(Process* parent, /* Parent of process (default group leader).
 		   Eterm args,	/* Arguments for function (must be well-formed list). */
 		   ErlSpawnOpts* so) /* Options for spawn. */
 {
+    Uint flags = 0;
     ErtsRunQueue *rq = NULL;
     Process *p;
     Sint arity;			/* Number of arguments. */
@@ -10797,6 +10725,10 @@ erl_create_process(Process* parent, /* Parent of process (default group leader).
     Eterm res = THE_NON_VALUE;
     erts_aint32_t state = 0;
     erts_aint32_t prio = (erts_aint32_t) PRIORITY_NORMAL;
+#ifdef SHCOPY_SPAWN
+    erts_shcopy_t info;
+    INITIALIZE_SHCOPY(info);
+#endif
 
 #ifdef ERTS_SMP
     erts_smp_proc_lock(parent, ERTS_PROC_LOCKS_ALL_MINOR);
@@ -10825,6 +10757,15 @@ erl_create_process(Process* parent, /* Parent of process (default group leader).
     state |= (((prio & ERTS_PSFLGS_PRIO_MASK) << ERTS_PSFLGS_ACT_PRIO_OFFSET)
 	      | ((prio & ERTS_PSFLGS_PRIO_MASK) << ERTS_PSFLGS_USR_PRIO_OFFSET));
 
+    if (so->flags & SPO_OFF_HEAP_MSGQ) {
+	state |= ERTS_PSFLG_OFF_HEAP_MSGQ;
+	flags |= F_OFF_HEAP_MSGQ;
+    }
+    else if (so->flags & SPO_ON_HEAP_MSGQ) {
+	state |= ERTS_PSFLG_ON_HEAP_MSGQ;
+	flags |= F_ON_HEAP_MSGQ;
+    }
+
     if (!rq)
 	rq = erts_get_runq_proc(parent);
 
@@ -10843,11 +10784,15 @@ erl_create_process(Process* parent, /* Parent of process (default group leader).
     BM_COUNT(processes_spawned);
 
     BM_SWAP_TIMER(system,size);
+#ifdef SHCOPY_SPAWN
+    arg_size = copy_shared_calculate(args, &info);
+#else
     arg_size = size_object(args);
+#endif
     BM_SWAP_TIMER(size,system);
     heap_need = arg_size;
 
-    p->flags = erts_default_process_flags;
+    p->flags = flags;
 
     p->static_flags = 0;
     if (so->flags & SPO_SYSTEM_PROC)
@@ -10896,12 +10841,13 @@ erl_create_process(Process* parent, /* Parent of process (default group leader).
     p->stop = p->hend = p->heap + sz;
     p->htop = p->heap;
     p->heap_sz = sz;
+    p->abandoned_heap = NULL;
+    p->live_hf_end = ERTS_INVALID_HFRAG_PTR;
     p->catches = 0;
 
     p->bin_vheap_sz     = p->min_vheap_size;
     p->bin_old_vheap_sz = p->min_vheap_size;
     p->bin_old_vheap    = 0;
-    p->bin_vheap_mature = 0;
 
     p->sys_task_qs = NULL;
 
@@ -10920,7 +10866,12 @@ erl_create_process(Process* parent, /* Parent of process (default group leader).
     BM_MESSAGE(args,p,parent);
     BM_START_TIMER(system);
     BM_SWAP_TIMER(system,copy);
+#ifdef SHCOPY_SPAWN
+    p->arg_reg[2] = copy_shared_perform(args, arg_size, &info, &p->htop, &p->off_heap);
+    DESTROY_SHCOPY(info);
+#else
     p->arg_reg[2] = copy_struct(args, arg_size, &p->htop, &p->off_heap);
+#endif
     BM_MESSAGE_COPIED(arg_size);
     BM_SWAP_TIMER(copy,system);
     p->arity = 3;
@@ -10966,6 +10917,7 @@ erl_create_process(Process* parent, /* Parent of process (default group leader).
     p->accessor_bif_timers = NULL;
 #endif
     p->mbuf = NULL;
+    p->msg_frag = NULL;
     p->mbuf_sz = 0;
     p->psd = NULL;
     p->dictionary = NULL;
@@ -11101,6 +11053,8 @@ void erts_init_empty_process(Process *p)
     p->stop = NULL;
     p->hend = NULL;
     p->heap = NULL;
+    p->abandoned_heap = NULL;
+    p->live_hf_end = ERTS_INVALID_HFRAG_PTR;
     p->gen_gcs = 0;
     p->max_gen_gcs = 0;
     p->min_heap_size = 0;
@@ -11121,7 +11075,6 @@ void erts_init_empty_process(Process *p)
     p->bin_old_vheap_sz = BIN_VH_MIN_SIZE;
     p->bin_old_vheap = 0;
     p->sys_task_qs = NULL;
-    p->bin_vheap_mature = 0;
     ERTS_PTMR_INIT(p);
     p->next = NULL;
     p->off_heap.first = NULL;
@@ -11133,6 +11086,7 @@ void erts_init_empty_process(Process *p)
     p->old_htop = NULL;
     p->old_heap = NULL;
     p->mbuf = NULL;
+    p->msg_frag = NULL;
     p->mbuf_sz = 0;
     p->psd = NULL;
     ERTS_P_MONITORS(p) = NULL;
@@ -11221,6 +11175,8 @@ erts_debug_verify_clean_empty_process(Process* p)
     ASSERT(p->htop == NULL);
     ASSERT(p->stop == NULL);
     ASSERT(p->hend == NULL);
+    ASSERT(p->abandoned_heap == NULL);
+    ASSERT(p->live_hf_end == ERTS_INVALID_HFRAG_PTR);
     ASSERT(p->heap == NULL);
     ASSERT(p->common.id == ERTS_INVALID_PID);
     ASSERT(ERTS_TRACER_PROC(p) == NIL);
@@ -11298,9 +11254,10 @@ erts_cleanup_empty_process(Process* p)
 static void
 delete_process(Process* p)
 {
-    ErlMessage* mp;
-
+    Eterm *heap;
     VERBOSE(DEBUG_PROCESSES, ("Removing process: %T\n",p->common.id));
+    VERBOSE(DEBUG_SHCOPY, ("[pid=%T] delete process: %p %p %p %p\n", p->common.id,
+                           HEAP_START(p), HEAP_END(p), OLD_HEAP(p), OLD_HEND(p)));
 
     /* Cleanup psd */
 
@@ -11324,16 +11281,17 @@ delete_process(Process* p)
      * Release heaps. Clobber contents in DEBUG build.
      */
 
-
-#ifdef DEBUG
-    sys_memset(p->heap, DEBUG_BAD_BYTE, p->heap_sz*sizeof(Eterm));
-#endif
-
 #ifdef HIPE
     hipe_delete_process(&p->hipe);
 #endif
 
-    ERTS_HEAP_FREE(ERTS_ALC_T_HEAP, (void*) p->heap, p->heap_sz*sizeof(Eterm));
+    heap = p->abandoned_heap ? p->abandoned_heap : p->heap;
+
+#ifdef DEBUG
+    sys_memset(heap, DEBUG_BAD_BYTE, p->heap_sz*sizeof(Eterm));
+#endif
+
+    ERTS_HEAP_FREE(ERTS_ALC_T_HEAP, (void*) heap, p->heap_sz*sizeof(Eterm));
     if (p->old_heap != NULL) {
 
 #ifdef DEBUG
@@ -11352,27 +11310,14 @@ delete_process(Process* p)
 	free_message_buffer(p->mbuf);
     }
 
+    if (p->msg_frag)
+	erts_cleanup_messages(p->msg_frag);
+
     erts_erase_dicts(p);
 
     /* free all pending messages */
-    mp = p->msg.first;
-    while(mp != NULL) {
-	ErlMessage* next_mp = mp->next;
-	if (mp->data.attached) {
-	    if (is_value(mp->m[0]))
-		free_message_buffer(mp->data.heap_frag);
-	    else {
-		if (is_not_nil(mp->m[1])) {
-		    ErlHeapFragment *heap_frag;
-		    heap_frag = (ErlHeapFragment *) mp->data.dist_ext->ext_endp;
-		    erts_cleanup_offheap(&heap_frag->off_heap);
-		}
-		erts_free_dist_ext_copy(mp->data.dist_ext);
-	    }
-	}
-	free_message(mp);
-	mp = next_mp;
-    }
+    erts_cleanup_messages(p->msg.first);
+    p->msg.first = NULL;
 
     ASSERT(!p->nodes_monitors);
     ASSERT(!p->suspend_monitors);
@@ -11409,6 +11354,22 @@ set_proc_exiting(Process *p,
     KILL_CATCHES(p);
     p->i = (BeamInstr *) beam_exit;
 
+#ifndef ERTS_SMP
+    if (state & (ERTS_PSFLG_RUNNING|ERTS_PSFLG_RUNNING_SYS)) {
+	/*
+	 * I non smp case:
+	 *
+	 * Currently executing process might be sent an exit
+	 * signal if it is traced by a port that it also is
+	 * linked to, and the port terminates during the
+	 * trace. In this case we want schedule out the
+	 * process as quickly as possible in order to detect
+	 * the event as fast as possible.
+	 */
+	ERTS_VBUMP_ALL_REDS(p);
+    }
+#endif
+
     if (enqueue)
 	add2runq(enqueue > 0 ? p : make_proxy_proc(NULL, p, enq_prio),
 		 state,
@@ -11506,10 +11467,14 @@ save_pending_exiter(Process *p)
 {
     ErtsProcList *plp;
     ErtsRunQueue *rq;
+    ErtsSchedulerData *esdp = erts_get_scheduler_data();
 
     ERTS_SMP_LC_ASSERT(ERTS_PROC_LOCK_STATUS & erts_proc_lc_my_proc_locks(p));
 
-    rq = erts_get_runq_current(NULL);
+    if (!esdp)
+        rq = RUNQ_READ_RQ(&p->run_queue);
+    else
+        rq = esdp->run_queue;
 
     plp = proclist_create(p);
 
@@ -11526,6 +11491,7 @@ save_pending_exiter(Process *p)
     else
 #endif
 	wake_scheduler(rq);
+
 }
 
 #endif
@@ -11539,35 +11505,46 @@ static ERTS_INLINE void
 send_exit_message(Process *to, ErtsProcLocks *to_locksp,
 		  Eterm exit_term, Uint term_size, Eterm token)
 {
-    if (token == NIL 
-#ifdef USE_VM_PROBES
-	|| token == am_have_dt_utag
-#endif
-	) {
-	Eterm* hp;
-	Eterm mess;
-	ErlHeapFragment* bp;
-	ErlOffHeap *ohp;
-
-	hp = erts_alloc_message_heap(term_size, &bp, &ohp, to, to_locksp);
+    ErtsMessage *mp;
+    ErlOffHeap *ohp;
+    Eterm* hp;
+    Eterm mess;
+#ifdef SHCOPY_SEND
+    erts_shcopy_t info;
+#endif
+
+    if (!have_seqtrace(token)) {
+#ifdef SHCOPY_SEND
+        INITIALIZE_SHCOPY(info);
+        term_size = copy_shared_calculate(exit_term, &info);
+	mp = erts_alloc_message_heap(to, to_locksp, term_size, &hp, &ohp);
+        mess = copy_shared_perform(exit_term, term_size, &info, &hp, ohp);
+        DESTROY_SHCOPY(info);
+#else
+	mp = erts_alloc_message_heap(to, to_locksp, term_size, &hp, &ohp);
 	mess = copy_struct(exit_term, term_size, &hp, ohp);
-	erts_queue_message(to, to_locksp, bp, mess, NIL);
+#endif
+	erts_queue_message(to, to_locksp, mp, mess, NIL);
     } else {
-	ErlHeapFragment* bp;
-	Eterm* hp;
-	Eterm mess;
 	Eterm temp_token;
 	Uint sz_token;
 
 	ASSERT(is_tuple(token));
 	sz_token = size_object(token);
-	bp = new_message_buffer(term_size+sz_token);
-	hp = bp->mem;
-	mess = copy_struct(exit_term, term_size, &hp, &bp->off_heap);
+#ifdef SHCOPY_SEND
+        INITIALIZE_SHCOPY(info);
+        term_size = copy_shared_calculate(exit_term, &info);
+	mp = erts_alloc_message_heap(to, to_locksp, term_size+sz_token, &hp, &ohp);
+        mess = copy_shared_perform(exit_term, term_size, &info, &hp, ohp);
+        DESTROY_SHCOPY(info);
+#else
+	mp = erts_alloc_message_heap(to, to_locksp, term_size+sz_token, &hp, &ohp);
+	mess = copy_struct(exit_term, term_size, &hp, ohp);
+#endif
 	/* the trace token must in this case be updated by the caller */
 	seq_trace_output(token, mess, SEQ_TRACE_SEND, to->common.id, NULL);
-	temp_token = copy_struct(token, sz_token, &hp, &bp->off_heap);
-	erts_queue_message(to, to_locksp, bp, mess, temp_token);
+	temp_token = copy_struct(token, sz_token, &hp, ohp);
+	erts_queue_message(to, to_locksp, mp, mess, temp_token);
     }
 }
 
@@ -11676,11 +11653,7 @@ send_exit_signal(Process *c_p,		/* current process if and only
 
     if ((state & ERTS_PSFLG_TRAP_EXIT)
 	&& (reason != am_kill || (flags & ERTS_XSIG_FLG_IGN_KILL))) {
-	if (is_not_nil(token) 
-#ifdef USE_VM_PROBES
-	    && token != am_have_dt_utag
-#endif
-	    && token_update)
+        if (have_seqtrace(token) && token_update)
 	    seq_trace_update_send(token_update);
 	if (is_value(exit_tuple))
 	    send_exit_message(rp, rp_locks, exit_tuple, exit_tuple_sz, token);
@@ -11714,23 +11687,21 @@ send_exit_signal(Process *c_p,		/* current process if and only
 		if (need_locks
 		    && erts_smp_proc_trylock(rp, need_locks) == EBUSY) {
 		    /* ... but we havn't got all locks on it ... */
-		    save_pending_exiter(rp);
+		    save_pending_exiter(rp); 
 		    /*
 		     * The pending exit will be discovered when next
 		     * process is scheduled in
 		     */
-		    goto set_pending_exit;
-		}
-		else {
-		    /* ...and we have all locks on it... */
-		    *rp_locks = ERTS_PROC_LOCKS_ALL;
-		    set_proc_exiting(rp,
-				     state,
-				     (is_immed(rsn)
-				      ? rsn
-				      : copy_object(rsn, rp)),
-				     NULL);
+                    goto set_pending_exit;
 		}
+                /* ...and we have all locks on it... */
+                *rp_locks = ERTS_PROC_LOCKS_ALL;
+                set_proc_exiting(rp,
+                                 state,
+                                 (is_immed(rsn)
+                                  ? rsn
+                                  : copy_object(rsn, rp)),
+                                 NULL);
 	    }
 	    else { /* Process running... */
 
@@ -12110,7 +12081,8 @@ erts_do_exit_process(Process* p, Eterm reason)
 #endif
 
     if (p->static_flags & ERTS_STC_FLG_SYSTEM_PROC)
-	erl_exit(1, "System process %T terminated: %T\n", p->common.id, reason);
+	erl_exit(ERTS_DUMP_EXIT, "System process %T terminated: %T\n",
+                 p->common.id, reason);
 
 #ifdef ERTS_SMP
     ERTS_SMP_CHK_HAVE_ONLY_MAIN_PROC_LOCK(p);
@@ -12344,9 +12316,7 @@ erts_continue_exit_process(Process *p)
 	    erts_proc_dec_refc(p);
     }
     
-    dep = ((p->flags & F_DISTRIBUTION)
-	   ? ERTS_PROC_SET_DIST_ENTRY(p, ERTS_PROC_LOCKS_ALL, NULL)
-	   : NULL);
+    dep = (p->flags & F_DISTRIBUTION) ? erts_this_dist_entry : NULL;
     scb = ERTS_PROC_SET_SAVED_CALLS_BUF(p, ERTS_PROC_LOCKS_ALL, NULL);
     pbt = ERTS_PROC_SET_CALL_TIME(p, ERTS_PROC_LOCKS_ALL, NULL);
     nif_export = ERTS_PROC_SET_NIF_TRAP_EXPORT(p, ERTS_PROC_LOCKS_ALL, NULL);
@@ -12358,8 +12328,6 @@ erts_continue_exit_process(Process *p)
 
     if (dep) {
 	erts_do_net_exits(dep, reason);
-	if(dep)
-	    erts_deref_dist_entry(dep);
     }
 
     /*
@@ -12518,7 +12486,7 @@ stack_element_dump(int to, void *to_arg, Eterm* sp, int yreg)
     }
 
     if (is_CP(x)) {
-        erts_print(to, to_arg, "Return addr %p (", (Eterm *) EXPAND_POINTER(x));
+        erts_print(to, to_arg, "Return addr %p (", (Eterm *) x);
         print_function_from_pc(to, to_arg, cp_val(x));
         erts_print(to, to_arg, ")\n");
         yreg = 0;
diff --git a/erts/emulator/beam/erl_process.h b/erts/emulator/beam/erl_process.h
index 20ffe7ea7c..b50fa5a10f 100644
--- a/erts/emulator/beam/erl_process.h
+++ b/erts/emulator/beam/erl_process.h
@@ -170,8 +170,10 @@ extern int erts_sched_thread_suggested_stack_size;
   (((Uint32) 1) << (ERTS_RUNQ_FLG_BASE2 + 5))
 #define ERTS_RUNQ_FLG_PROTECTED \
   (((Uint32) 1) << (ERTS_RUNQ_FLG_BASE2 + 6))
+#define ERTS_RUNQ_FLG_EXEC \
+  (((Uint32) 1) << (ERTS_RUNQ_FLG_BASE2 + 7))
 
-#define ERTS_RUNQ_FLG_MAX (ERTS_RUNQ_FLG_BASE2 + 7)
+#define ERTS_RUNQ_FLG_MAX (ERTS_RUNQ_FLG_BASE2 + 8)
 
 #define ERTS_RUNQ_FLGS_MIGRATION_QMASKS	\
   (ERTS_RUNQ_FLGS_EMIGRATE_QMASK	\
@@ -215,6 +217,9 @@ extern int erts_sched_thread_suggested_stack_size;
 #define ERTS_RUNQ_FLGS_SET(RQ, FLGS)					\
     ((Uint32) erts_smp_atomic32_read_bor_relb(&(RQ)->flags,		\
 					      (erts_aint32_t) (FLGS)))
+#define ERTS_RUNQ_FLGS_SET_NOB(RQ, FLGS)				\
+    ((Uint32) erts_smp_atomic32_read_bor_nob(&(RQ)->flags,		\
+					     (erts_aint32_t) (FLGS)))
 #define ERTS_RUNQ_FLGS_BSET(RQ, MSK, FLGS)				\
     ((Uint32) erts_smp_atomic32_read_bset_relb(&(RQ)->flags,		\
 					       (erts_aint32_t) (MSK),	\
@@ -222,6 +227,9 @@ extern int erts_sched_thread_suggested_stack_size;
 #define ERTS_RUNQ_FLGS_UNSET(RQ, FLGS)					\
     ((Uint32) erts_smp_atomic32_read_band_relb(&(RQ)->flags,		\
 					       (erts_aint32_t) ~(FLGS)))
+#define ERTS_RUNQ_FLGS_UNSET_NOB(RQ, FLGS)					\
+    ((Uint32) erts_smp_atomic32_read_band_nob(&(RQ)->flags,		\
+					      (erts_aint32_t) ~(FLGS)))
 #define ERTS_RUNQ_FLGS_GET(RQ)						\
     ((Uint32) erts_smp_atomic32_read_acqb(&(RQ)->flags))
 #define ERTS_RUNQ_FLGS_GET_NOB(RQ)					\
@@ -293,7 +301,6 @@ typedef enum {
     ERTS_SSI_AUX_WORK_ASYNC_READY_CLEAN_IX,
     ERTS_SSI_AUX_WORK_MISC_THR_PRGR_IX,
     ERTS_SSI_AUX_WORK_MISC_IX,
-    ERTS_SSI_AUX_WORK_CHECK_CHILDREN_IX,
     ERTS_SSI_AUX_WORK_SET_TMO_IX,
     ERTS_SSI_AUX_WORK_MSEG_CACHE_CHECK_IX,
     ERTS_SSI_AUX_WORK_REAP_PORTS_IX,
@@ -326,8 +333,6 @@ typedef enum {
     (((erts_aint32_t) 1) << ERTS_SSI_AUX_WORK_MISC_THR_PRGR_IX)
 #define ERTS_SSI_AUX_WORK_MISC \
     (((erts_aint32_t) 1) << ERTS_SSI_AUX_WORK_MISC_IX)
-#define ERTS_SSI_AUX_WORK_CHECK_CHILDREN \
-    (((erts_aint32_t) 1) << ERTS_SSI_AUX_WORK_CHECK_CHILDREN_IX)
 #define ERTS_SSI_AUX_WORK_SET_TMO \
     (((erts_aint32_t) 1) << ERTS_SSI_AUX_WORK_SET_TMO_IX)
 #define ERTS_SSI_AUX_WORK_MSEG_CACHE_CHECK \
@@ -467,7 +472,7 @@ struct ErtsRunQueue_ {
     int full_reds_history[ERTS_FULL_REDS_HISTORY_SIZE];
     int out_of_work_count;
     erts_aint32_t max_len;
-    erts_aint32_t len;
+    erts_smp_atomic32_t len;
     int wakeup_other;
     int wakeup_other_reds;
     int halt_in_progress;
@@ -607,7 +612,7 @@ typedef enum {
 typedef union {
     struct {
 	ErtsDirtySchedulerType type: 1;
-	unsigned num: 31;
+	Uint num: sizeof(Uint)*8 - 1;
     } s;
     Uint no;
 } ErtsDirtySchedId;
@@ -632,12 +637,6 @@ struct ErtsSchedulerData_ {
     Process *free_process;
     ErtsThrPrgrData thr_progress_data;
 #endif
-#if !HEAP_ON_C_STACK
-    Eterm tmp_heap[TMP_HEAP_SIZE];
-    int num_tmp_heap_used;
-    Eterm beam_emu_tmp_heap[BEAM_EMU_TMP_HEAP_SIZE];
-    Eterm erl_arith_tmp_heap[ERL_ARITH_TMP_HEAP_SIZE];
-#endif
     ErtsSchedulerSleepInfo *ssi;
     Process *current_process;
     Uint no;			/* Scheduler number for normal schedulers */
@@ -691,13 +690,6 @@ extern ErtsAlignedSchedulerData *erts_aligned_dirty_io_scheduler_data;
 extern ErtsSchedulerData *erts_scheduler_data;
 #endif
 
-#ifdef ERTS_SCHED_FAIR
-#define ERTS_SCHED_FAIR_YIELD() ETHR_YIELD()
-#else
-#define ERTS_SCHED_FAIR 0
-#define ERTS_SCHED_FAIR_YIELD()
-#endif
-
 #if defined(ERTS_SMP) && defined(ERTS_ENABLE_LOCK_CHECK)
 int erts_smp_lc_runq_is_locked(ErtsRunQueue *);
 #endif
@@ -728,7 +720,19 @@ erts_smp_inc_runq_len(ErtsRunQueue *rq, ErtsRunQueueInfo *rqi, int prio)
 
     ERTS_SMP_LC_ASSERT(erts_smp_lc_runq_is_locked(rq));
 
-    len = erts_smp_atomic32_read_nob(&rqi->len);
+    len = erts_smp_atomic32_read_dirty(&rq->len);
+
+#ifdef ERTS_SMP
+    if (len == 0)
+	erts_non_empty_runq(rq);
+#endif
+    len++;
+    if (rq->max_len < len)
+	rq->max_len = len;
+    ASSERT(len > 0);
+    erts_smp_atomic32_set_nob(&rq->len, len);
+
+    len = erts_smp_atomic32_read_dirty(&rqi->len);
     ASSERT(len >= 0);
     if (len == 0) {
 	ASSERT((erts_smp_atomic32_read_nob(&rq->flags)
@@ -741,15 +745,6 @@ erts_smp_inc_runq_len(ErtsRunQueue *rq, ErtsRunQueueInfo *rqi, int prio)
 	rqi->max_len = len;
 
     erts_smp_atomic32_set_relb(&rqi->len, len);
-
-#ifdef ERTS_SMP
-    if (rq->len == 0)
-	erts_non_empty_runq(rq);
-#endif
-    rq->len++;
-    if (rq->max_len < rq->len)
-	rq->max_len = len;
-    ASSERT(rq->len > 0);
 }
 
 ERTS_GLB_INLINE void
@@ -759,7 +754,12 @@ erts_smp_dec_runq_len(ErtsRunQueue *rq, ErtsRunQueueInfo *rqi, int prio)
 
     ERTS_SMP_LC_ASSERT(erts_smp_lc_runq_is_locked(rq));
 
-    len = erts_smp_atomic32_read_nob(&rqi->len);
+    len = erts_smp_atomic32_read_dirty(&rq->len);
+    len--;
+    ASSERT(len >= 0);
+    erts_smp_atomic32_set_nob(&rq->len, len);
+
+    len = erts_smp_atomic32_read_dirty(&rqi->len);
     len--;
     ASSERT(len >= 0);
     if (len == 0) {
@@ -770,8 +770,6 @@ erts_smp_dec_runq_len(ErtsRunQueue *rq, ErtsRunQueueInfo *rqi, int prio)
     }
     erts_smp_atomic32_set_relb(&rqi->len, len);
 
-    rq->len--;
-    ASSERT(rq->len >= 0);
 }
 
 ERTS_GLB_INLINE void
@@ -781,7 +779,7 @@ erts_smp_reset_max_len(ErtsRunQueue *rq, ErtsRunQueueInfo *rqi)
 
     ERTS_SMP_LC_ASSERT(erts_smp_lc_runq_is_locked(rq));
 
-    len = erts_smp_atomic32_read_nob(&rqi->len);
+    len = erts_smp_atomic32_read_dirty(&rqi->len);
     ASSERT(rqi->max_len >= len);
     rqi->max_len = len;
 }
@@ -801,12 +799,11 @@ erts_smp_reset_max_len(ErtsRunQueue *rq, ErtsRunQueueInfo *rqi)
 #define ERTS_PSD_ERROR_HANDLER			0
 #define ERTS_PSD_SAVED_CALLS_BUF		1
 #define ERTS_PSD_SCHED_ID			2
-#define ERTS_PSD_DIST_ENTRY			3
-#define ERTS_PSD_CALL_TIME_BP			4
-#define ERTS_PSD_DELAYED_GC_TASK_QS		5
-#define ERTS_PSD_NIF_TRAP_EXPORT		6
+#define ERTS_PSD_CALL_TIME_BP			3
+#define ERTS_PSD_DELAYED_GC_TASK_QS		4
+#define ERTS_PSD_NIF_TRAP_EXPORT		5
 
-#define ERTS_PSD_SIZE				7
+#define ERTS_PSD_SIZE				6
 
 typedef struct {
     void *data[ERTS_PSD_SIZE];
@@ -824,9 +821,6 @@ typedef struct {
 #define ERTS_PSD_SCHED_ID_GET_LOCKS ERTS_PROC_LOCK_STATUS
 #define ERTS_PSD_SCHED_ID_SET_LOCKS ERTS_PROC_LOCK_STATUS
 
-#define ERTS_PSD_DIST_ENTRY_GET_LOCKS ERTS_PROC_LOCK_MAIN
-#define ERTS_PSD_DIST_ENTRY_SET_LOCKS ERTS_PROC_LOCK_MAIN
-
 #define ERTS_PSD_CALL_TIME_BP_GET_LOCKS ERTS_PROC_LOCK_MAIN
 #define ERTS_PSD_CALL_TIME_BP_SET_LOCKS ERTS_PROC_LOCK_MAIN
 
@@ -911,7 +905,6 @@ struct ErtsPendingSuspend_ {
 
 #  define MIN_VHEAP_SIZE(p)   (p)->min_vheap_size
 #  define BIN_VHEAP_SZ(p)     (p)->bin_vheap_sz
-#  define BIN_VHEAP_MATURE(p) (p)->bin_vheap_mature
 #  define BIN_OLD_VHEAP_SZ(p) (p)->bin_old_vheap_sz
 #  define BIN_OLD_VHEAP(p)    (p)->bin_old_vheap
 
@@ -929,6 +922,7 @@ struct process {
     Eterm* stop;		/* Stack top */
     Eterm* heap;		/* Heap start */
     Eterm* hend;		/* Heap end */
+    Eterm* abandoned_heap;
     Uint heap_sz;		/* Size of heap in words */
     Uint min_heap_size;         /* Minimum size of heap (in words). */
     Uint min_vheap_size;        /* Minimum size of virtual heap (in words). */
@@ -1025,12 +1019,13 @@ struct process {
     Uint16 gen_gcs;		/* Number of (minor) generational GCs. */
     Uint16 max_gen_gcs;		/* Max minor gen GCs before fullsweep. */
     ErlOffHeap off_heap;	/* Off-heap data updated by copy_struct(). */
-    ErlHeapFragment* mbuf;	/* Pointer to message buffer list */
-    Uint mbuf_sz;		/* Size of all message buffers */
+    ErlHeapFragment* mbuf;	/* Pointer to heap fragment list */
+    ErlHeapFragment* live_hf_end;
+    ErtsMessage *msg_frag;	/* Pointer to message fragment list */
+    Uint mbuf_sz;		/* Total size of heap fragments and message fragments */
     ErtsPSD *psd;		/* Rarely used process specific data */
 
     Uint64 bin_vheap_sz;	/* Virtual heap block size for binaries */
-    Uint64 bin_vheap_mature;	/* Virtual heap block size for binaries */
     Uint64 bin_old_vheap_sz;	/* Virtual old heap block size for binaries */
     Uint64 bin_old_vheap;	/* Virtual old heap size for binaries */
 
@@ -1054,6 +1049,7 @@ struct process {
 #ifdef CHECK_FOR_HOLES
     Eterm* last_htop;		/* No need to scan the heap below this point. */
     ErlHeapFragment* last_mbuf;	/* No need to scan beyond this mbuf. */
+    ErlHeapFragment* heap_hfrag; /* Heap abandoned, htop now lives in this frag */
 #endif
 
 #ifdef DEBUG
@@ -1077,6 +1073,7 @@ extern const Process erts_invalid_process;
 do {						\
   (p)->last_htop = 0;				\
   (p)->last_mbuf = 0;				\
+  (p)->heap_hfrag = NULL;			\
 } while (0)
 
 # define ERTS_HOLE_CHECK(p) erts_check_for_holes((p))
@@ -1154,14 +1151,16 @@ void erts_check_for_holes(Process* p);
 #define ERTS_PSFLG_RUNNING_SYS		ERTS_PSFLG_BIT(15)
 #define ERTS_PSFLG_PROXY		ERTS_PSFLG_BIT(16)
 #define ERTS_PSFLG_DELAYED_SYS		ERTS_PSFLG_BIT(17)
+#define ERTS_PSFLG_OFF_HEAP_MSGQ	ERTS_PSFLG_BIT(18)
+#define ERTS_PSFLG_ON_HEAP_MSGQ		ERTS_PSFLG_BIT(19)
 #ifdef ERTS_DIRTY_SCHEDULERS
-#define ERTS_PSFLG_DIRTY_CPU_PROC	ERTS_PSFLG_BIT(18)
-#define ERTS_PSFLG_DIRTY_IO_PROC	ERTS_PSFLG_BIT(19)
-#define ERTS_PSFLG_DIRTY_CPU_PROC_IN_Q	ERTS_PSFLG_BIT(20)
-#define ERTS_PSFLG_DIRTY_IO_PROC_IN_Q	ERTS_PSFLG_BIT(21)
-#define ERTS_PSFLG_MAX  (ERTS_PSFLGS_ZERO_BIT_OFFSET + 22)
+#define ERTS_PSFLG_DIRTY_CPU_PROC	ERTS_PSFLG_BIT(20)
+#define ERTS_PSFLG_DIRTY_IO_PROC	ERTS_PSFLG_BIT(21)
+#define ERTS_PSFLG_DIRTY_CPU_PROC_IN_Q	ERTS_PSFLG_BIT(22)
+#define ERTS_PSFLG_DIRTY_IO_PROC_IN_Q	ERTS_PSFLG_BIT(23)
+#define ERTS_PSFLG_MAX  (ERTS_PSFLGS_ZERO_BIT_OFFSET + 24)
 #else
-#define ERTS_PSFLG_MAX  (ERTS_PSFLGS_ZERO_BIT_OFFSET + 18)
+#define ERTS_PSFLG_MAX  (ERTS_PSFLGS_ZERO_BIT_OFFSET + 20)
 #endif
 
 #define ERTS_PSFLGS_IN_PRQ_MASK 	(ERTS_PSFLG_IN_PRQ_MAX		\
@@ -1209,12 +1208,16 @@ void erts_check_for_holes(Process* p);
 #define SPO_USE_ARGS 2
 #define SPO_MONITOR 4
 #define SPO_SYSTEM_PROC 8
+#define SPO_OFF_HEAP_MSGQ 16
+#define SPO_ON_HEAP_MSGQ 32
+
+extern int erts_default_spo_flags;
 
 /*
  * The following struct contains options for a process to be spawned.
  */
 typedef struct {
-    Uint flags;
+    int flags;
     int error_code;		/* Error code returned from create_process(). */
     Eterm mref;			/* Monitor ref returned (if SPO_MONITOR was given). */
 
@@ -1254,7 +1257,6 @@ Eterm* erts_heap_alloc(Process* p, Uint need, Uint xtra);
 Eterm* erts_set_hole_marker(Eterm* ptr, Uint sz);
 #endif
 
-extern Uint erts_default_process_flags;
 extern erts_smp_rwmtx_t erts_cpu_bind_rwmtx;
 /* If any of the erts_system_monitor_* variables are set (enabled),
 ** erts_system_monitor must be != NIL, to allow testing on just
@@ -1295,7 +1297,23 @@ extern struct erts_system_profile_flags_t erts_system_profile_flags;
 #define F_HAVE_BLCKD_MSCHED  (1 <<  8) /* Process has blocked multi-scheduling */
 #define F_P2PNR_RESCHED      (1 <<  9) /* Process has been rescheduled via erts_pid2proc_not_running() */
 #define F_FORCE_GC           (1 << 10) /* Force gc at process in-scheduling */
-#define F_DISABLE_GC         (1 << 11) /* Disable GC */
+#define F_DISABLE_GC         (1 << 11) /* Disable GC (see below) */
+#define F_OFF_HEAP_MSGQ      (1 << 12) /* Off heap msg queue */
+#define F_ON_HEAP_MSGQ       (1 << 13) /* Off heap msg queue */
+#define F_OFF_HEAP_MSGQ_CHNG (1 << 14) /* Off heap msg queue changing */
+#define F_ABANDONED_HEAP_USE (1 << 15) /* Have usage of abandoned heap */
+#define F_DELAY_GC           (1 << 16) /* Similar to disable GC (see below) */
+
+/*
+ * F_DISABLE_GC and F_DELAY_GC are similar. Both will prevent
+ * GC of the process, but it is important to use the right
+ * one:
+ * - F_DISABLE_GC should *only* be used by BIFs. This when
+ *   the BIF needs to yield while preventig a GC.
+ * - F_DELAY_GC should only be used when GC is temporarily
+ *   disabled while the process is scheduled. A process must
+ *   not be scheduled out while F_DELAY_GC is set.
+ */
 
 /* process trace_flags */
 #define F_SENSITIVE          (1 << 0)
@@ -1629,6 +1647,7 @@ void erts_schedule_thr_prgr_later_cleanup_op(void (*)(void *),
 					     void *,
 					     ErtsThrPrgrLaterOp *,
 					     UWord);
+void erts_schedule_complete_off_heap_message_queue_change(Eterm pid);
 
 #if defined(ERTS_SMP) && defined(ERTS_ENABLE_LOCK_CHECK)
 int erts_dbg_check_halloc_lock(Process *p);
@@ -1658,11 +1677,8 @@ Eterm erts_multi_scheduling_blockers(Process *);
 void erts_start_schedulers(void);
 void erts_alloc_notify_delayed_dealloc(int);
 void erts_alloc_ensure_handle_delayed_dealloc_call(int);
-#ifdef ERTS_SMP
 void erts_notify_canceled_timer(ErtsSchedulerData *, int);
 #endif
-void erts_smp_notify_check_children_needed(void);
-#endif
 #if ERTS_USE_ASYNC_READY_Q
 void erts_notify_check_async_ready_queue(void *);
 #endif
@@ -1682,7 +1698,7 @@ void erts_sched_notify_check_cpu_bind(void);
 Uint erts_active_schedulers(void);
 void erts_init_process(int, int, int);
 Eterm erts_process_status(Process *, ErtsProcLocks, Process *, Eterm);
-Uint erts_run_queues_len(Uint *);
+Uint erts_run_queues_len(Uint *, int, int);
 void erts_add_to_runq(Process *);
 Eterm erts_bound_schedulers_term(Process *c_p);
 Eterm erts_get_cpu_topology_term(Process *c_p, Eterm which);
@@ -1756,7 +1772,7 @@ Uint erts_debug_nbalance(void);
 
 int erts_debug_wait_completed(Process *c_p, int flags);
 
-Uint erts_process_memory(Process *c_p);
+Uint erts_process_memory(Process *c_p, int incl_msg_inq);
 
 #ifdef ERTS_SMP
 #  define ERTS_GET_SCHEDULER_DATA_FROM_PROC(PROC) ((PROC)->scheduler_data)
@@ -1887,11 +1903,6 @@ erts_psd_set(Process *p, ErtsProcLocks plocks, int ix, void *data)
 #define ERTS_PROC_SCHED_ID(P, L, ID) \
   ((UWord) erts_psd_set((P), (L), ERTS_PSD_SCHED_ID, (void *) (ID)))
 
-#define ERTS_PROC_GET_DIST_ENTRY(P) \
-  ((DistEntry *) erts_psd_get((P), ERTS_PSD_DIST_ENTRY))
-#define ERTS_PROC_SET_DIST_ENTRY(P, L, D) \
-  ((DistEntry *) erts_psd_set((P), (L), ERTS_PSD_DIST_ENTRY, (void *) (D)))
-
 #define ERTS_PROC_GET_SAVED_CALLS_BUF(P) \
   ((struct saved_calls *) erts_psd_get((P), ERTS_PSD_SAVED_CALLS_BUF))
 #define ERTS_PROC_SET_SAVED_CALLS_BUF(P, L, SCB) \
@@ -2071,6 +2082,22 @@ ERTS_GLB_INLINE void erts_smp_xrunq_unlock(ErtsRunQueue *rq, ErtsRunQueue *xrq);
 ERTS_GLB_INLINE void erts_smp_runqs_lock(ErtsRunQueue *rq1, ErtsRunQueue *rq2);
 ERTS_GLB_INLINE void erts_smp_runqs_unlock(ErtsRunQueue *rq1, ErtsRunQueue *rq2);
 
+ERTS_GLB_INLINE ErtsMessage *erts_alloc_message_heap_state(Process *pp,
+							   erts_aint32_t *psp,
+							   ErtsProcLocks *plp,
+							   Uint sz,
+							   Eterm **hpp,
+							   ErlOffHeap **ohpp);
+ERTS_GLB_INLINE ErtsMessage *erts_alloc_message_heap(Process *pp,
+						     ErtsProcLocks *plp,
+						     Uint sz,
+						     Eterm **hpp,
+						     ErlOffHeap **ohpp);
+
+ERTS_GLB_INLINE void erts_shrink_message_heap(ErtsMessage **msgpp, Process *pp,
+					      Eterm *start_hp, Eterm *used_hp, Eterm *end_hp,
+					      Eterm *brefs, Uint brefs_size);
+
 #if ERTS_GLB_INLINE_INCL_FUNC_DEF
 
 ERTS_GLB_INLINE
@@ -2219,6 +2246,63 @@ erts_smp_runqs_unlock(ErtsRunQueue *rq1, ErtsRunQueue *rq2)
 #endif
 }
 
+ERTS_GLB_INLINE ErtsMessage *
+erts_alloc_message_heap_state(Process *pp,
+			      erts_aint32_t *psp,
+			      ErtsProcLocks *plp,
+			      Uint sz,
+			      Eterm **hpp,
+			      ErlOffHeap **ohpp)
+{
+    int on_heap;
+
+    if ((*psp) & ERTS_PSFLG_OFF_HEAP_MSGQ) {
+	ErtsMessage *mp = erts_alloc_message(sz, hpp);
+	*ohpp = sz == 0 ? NULL : &mp->hfrag.off_heap;
+	return mp;
+    }
+
+    return erts_try_alloc_message_on_heap(pp, psp, plp, sz, hpp, ohpp, &on_heap);
+}
+
+ERTS_GLB_INLINE ErtsMessage *
+erts_alloc_message_heap(Process *pp,
+			ErtsProcLocks *plp,
+			Uint sz,
+			Eterm **hpp,
+			ErlOffHeap **ohpp)
+{
+    erts_aint32_t state = erts_smp_atomic32_read_nob(&pp->state);
+    return erts_alloc_message_heap_state(pp, &state, plp, sz, hpp, ohpp);
+}
+
+ERTS_GLB_INLINE void
+erts_shrink_message_heap(ErtsMessage **msgpp, Process *pp,
+			 Eterm *start_hp, Eterm *used_hp, Eterm *end_hp,
+			 Eterm *brefs, Uint brefs_size)
+{
+    ASSERT(start_hp <= used_hp && used_hp <= end_hp);
+    if ((*msgpp)->data.attached == ERTS_MSG_COMBINED_HFRAG)
+	*msgpp = erts_shrink_message(*msgpp, used_hp - start_hp,
+				     brefs, brefs_size);
+    else if (!(*msgpp)->data.attached) {
+	ERTS_SMP_LC_ASSERT(ERTS_PROC_LOCK_MAIN
+			   & erts_proc_lc_my_proc_locks(pp));
+	HRelease(pp, end_hp, used_hp);
+    }
+    else {
+	ErlHeapFragment *hfrag = (*msgpp)->data.heap_frag;
+	if (start_hp != used_hp)
+	    hfrag = erts_resize_message_buffer(hfrag, used_hp - start_hp,
+					       brefs, brefs_size);
+	else {
+	    free_message_buffer(hfrag);
+	    hfrag = NULL;
+	}
+	(*msgpp)->data.heap_frag = hfrag;
+    }
+}
+
 #endif /* #if ERTS_GLB_INLINE_INCL_FUNC_DEF */
 
 ERTS_GLB_INLINE ErtsAtomCacheMap *erts_get_atom_cache_map(Process *c_p);
diff --git a/erts/emulator/beam/erl_process_dict.c b/erts/emulator/beam/erl_process_dict.c
index 8606371bdf..84cd81aecf 100644
--- a/erts/emulator/beam/erl_process_dict.c
+++ b/erts/emulator/beam/erl_process_dict.c
@@ -51,16 +51,19 @@
 #define INITIAL_SIZE        (erts_pd_initial_size)
 
 /* Hash utility macros  */
-#define HASH_RANGE(PDict) ((PDict)->homeSize + (PDict)->splitPosition)
+#define HASH_RANGE(PDict) ((PDict)->usedSlots)
 
-#define MAKE_HASH(Term) 				\
-((is_small(Term)) ? unsigned_val(Term) :		\
- ((is_atom(Term)) ? 					\
-  (atom_tab(atom_val(term))->slot.bucket.hvalue) :	\
-  make_hash2(Term)))
+#define MAKE_HASH(Term)                                \
+    ((is_small(Term)) ? unsigned_val(Term) :           \
+     ((is_atom(Term)) ?                                \
+      (atom_tab(atom_val(Term))->slot.bucket.hvalue) : \
+      make_internal_hash(Term)))
 
 #define PD_SZ2BYTES(Sz) (sizeof(ProcDict) + ((Sz) - 1)*sizeof(Eterm))
 
+#define pd_hash_value(Pdict, Key) \
+    pd_hash_value_to_ix(Pdict, MAKE_HASH((Key)))
+
 /* Memory allocation macros */
 #define PD_ALLOC(Sz)				\
      erts_alloc(ERTS_ALC_T_PROC_DICT, (Sz))
@@ -74,14 +77,19 @@
 #define TCDR(Term) CDR(list_val(Term))
 
 /* Array access macro */ 
-#define ARRAY_GET(PDict, Index) (((PDict)->size > (Index)) ? \
-				 (PDict)->data[Index] : NIL)
+#define ARRAY_GET(PDict, Index) (ASSERT((Index) < (PDict)->arraySize), \
+				 (PDict)->data[Index])
+#define ARRAY_PUT(PDict, Index, Val) (ASSERT((Index) < (PDict)->arraySize), \
+                                      (PDict)->data[Index] = (Val))
+
+#define IS_POW2(X) ((X) && !((X) & ((X)-1)))
 
 /*
  * Forward decalarations
  */
 static void pd_hash_erase(Process *p, Eterm id, Eterm *ret);
 static void pd_hash_erase_all(Process *p);
+static Eterm pd_hash_get_with_hval(Process *p, Eterm bucket, Eterm id);
 static Eterm pd_hash_get_keys(Process *p, Eterm value);
 static Eterm pd_hash_get_all_keys(Process *p, ProcDict *pd);
 static Eterm pd_hash_get_all(Process *p, ProcDict *pd);
@@ -91,9 +99,9 @@ static void shrink(Process *p, Eterm* ret);
 static void grow(Process *p);
 
 static void array_shrink(ProcDict **ppd, unsigned int need);
-static Eterm array_put(ProcDict **ppdict, unsigned int ndx, Eterm term);
+static void ensure_array_size(ProcDict**, unsigned int size);
 
-static unsigned int pd_hash_value(ProcDict *pdict, Eterm term);
+static unsigned int pd_hash_value_to_ix(ProcDict *pdict, Uint32 hx);
 static unsigned int next_array_size(unsigned int need);
 
 /*
@@ -134,6 +142,16 @@ static void pd_check(ProcDict *pd);
 ** External interface
 */
 
+int
+erts_pd_set_initial_size(int size)
+{
+    if (size <= 0)
+	return 0;
+
+    erts_pd_initial_size = 1 << erts_fit_in_bits_uint(size-1);
+    return 1;
+}
+
 /*
  * Called from break handler
  */
@@ -146,9 +164,9 @@ erts_dictionary_dump(int to, void *to_arg, ProcDict *pd)
     /*PD_CHECK(pd);*/
     if (pd == NULL)
 	return;
-    erts_print(to, to_arg, "(size = %d, used = %d, homeSize = %d, "
+    erts_print(to, to_arg, "(size = %d, usedSlots = %d, "
 	       "splitPosition = %d, numElements = %d)\n",
-	       pd->size, pd->used, pd->homeSize, 
+	       pd->arraySize, pd->usedSlots,
 	       pd->splitPosition, (unsigned int) pd->numElements);
     for (i = 0; i < HASH_RANGE(pd); ++i) {
 	erts_print(to, to_arg, "%d: %T\n", i, ARRAY_GET(pd, i));
@@ -203,7 +221,7 @@ erts_dicts_mem_size(Process *p)
 {
     Uint size = 0;
     if (p->dictionary)
-	size += PD_SZ2BYTES(p->dictionary->size);
+	size += PD_SZ2BYTES(p->dictionary->arraySize);
     return size;
 }
 
@@ -345,7 +363,7 @@ static void pd_hash_erase(Process *p, Eterm id, Eterm *ret)
     if (is_boxed(old)) {	/* Tuple */
 	ASSERT(is_tuple(old));
 	if (EQ(tuple_val(old)[1], id)) {
-	    array_put(&(p->dictionary), hval, NIL);
+	    ARRAY_PUT(p->dictionary, hval, NIL);
 	    --(p->dictionary->numElements);
 	    *ret = tuple_val(old)[2];
 	}
@@ -365,7 +383,7 @@ static void pd_hash_erase(Process *p, Eterm id, Eterm *ret)
 	old = ARRAY_GET(p->dictionary, hval);
 	ASSERT(is_list(old));
 	if (is_nil(TCDR(old))) {
-	    array_put(&p->dictionary, hval, TCAR(old));
+	    ARRAY_PUT(p->dictionary, hval, TCAR(old));
 	}
     } else if (is_not_nil(old)) {
 #ifdef DEBUG
@@ -390,40 +408,56 @@ static void pd_hash_erase_all(Process *p)
     }
 }
 
+Eterm erts_pd_hash_get_with_hx(Process *p, Uint32 hx, Eterm id)
+{
+    unsigned int hval;
+    ProcDict *pd = p->dictionary;
+
+    ASSERT(hx == MAKE_HASH(id));
+    if (pd == NULL)
+	return am_undefined;
+    hval = pd_hash_value_to_ix(pd, hx);
+    return pd_hash_get_with_hval(p, ARRAY_GET(pd, hval), id);
+}
+
 Eterm erts_pd_hash_get(Process *p, Eterm id) 
 {
     unsigned int hval;
-    Eterm tmp;
     ProcDict *pd = p->dictionary;
 
     if (pd == NULL)
 	return am_undefined;
     hval = pd_hash_value(pd, id);
-    tmp = ARRAY_GET(pd, hval);
-    if (is_boxed(tmp)) {	/* Tuple */
-	ASSERT(is_tuple(tmp));
-	if (EQ(tuple_val(tmp)[1], id)) {
-	    return tuple_val(tmp)[2];
+    return pd_hash_get_with_hval(p, ARRAY_GET(pd, hval), id);
+}
+
+Eterm pd_hash_get_with_hval(Process *p, Eterm bucket, Eterm id)
+{
+    if (is_boxed(bucket)) {     /* Tuple */
+	ASSERT(is_tuple(bucket));
+	if (EQ(tuple_val(bucket)[1], id)) {
+	    return tuple_val(bucket)[2];
 	}
-    } else if (is_list(tmp)) {
-	for (; tmp != NIL && !EQ(tuple_val(TCAR(tmp))[1], id); tmp = TCDR(tmp)) {
+    } else if (is_list(bucket)) {
+	for (; bucket != NIL && !EQ(tuple_val(TCAR(bucket))[1], id); bucket = TCDR(bucket)) {
 	    ;
 	}
-	if (tmp != NIL) {
-	    return tuple_val(TCAR(tmp))[2];
+	if (bucket != NIL) {
+	    return tuple_val(TCAR(bucket))[2];
 	}
-    } else if (is_not_nil(tmp)) {
+    } else if (is_not_nil(bucket)) {
 #ifdef DEBUG
 	erts_fprintf(stderr,
 		     "Process dictionary for process %T is broken, trying to "
 		     "display term found in line %d:\n"
-		     "%T\n", p->common.id, __LINE__, tmp);
+		     "%T\n", p->common.id, __LINE__, bucket);
 #endif
 	erl_exit(1, "Damaged process dictionary found during get/1.");
     }
     return am_undefined;
 }
 
+
 #define PD_GET_TKEY(Dst,Src)                            \
 do {                                                    \
     ASSERT(is_tuple((Src)));                            \
@@ -549,8 +583,11 @@ static Eterm pd_hash_put(Process *p, Eterm id, Eterm value)
 
     if (p->dictionary == NULL) {
 	/* Create it */
-	array_put(&(p->dictionary), INITIAL_SIZE - 1, NIL);
-	p->dictionary->homeSize = INITIAL_SIZE;
+        ensure_array_size(&p->dictionary, INITIAL_SIZE);
+	p->dictionary->usedSlots = INITIAL_SIZE;
+        p->dictionary->sizeMask = INITIAL_SIZE*2 - 1;
+        p->dictionary->splitPosition = 0;
+        p->dictionary->numElements = 0;
     }	
     hval = pd_hash_value(p->dictionary, id);
     old = ARRAY_GET(p->dictionary, hval);
@@ -583,7 +620,7 @@ static Eterm pd_hash_put(Process *p, Eterm id, Eterm value)
 	root[0] = id;
 	root[1] = value;
 	root[2] = old;
-	BUMP_REDS(p, erts_garbage_collect(p, needed, root, 3));
+	erts_garbage_collect(p, needed, root, 3);
 	id = root[0];
 	value = root[1];
 	old = root[2];
@@ -602,19 +639,19 @@ static Eterm pd_hash_put(Process *p, Eterm id, Eterm value)
      * Update the dictionary.
      */
     if (is_nil(old)) {
-	array_put(&(p->dictionary), hval, tpl);
+	ARRAY_PUT(p->dictionary, hval, tpl);
 	++(p->dictionary->numElements);
     } else if (is_boxed(old)) {
 	ASSERT(is_tuple(old));
 	if (EQ(tuple_val(old)[1],id)) {
-	    array_put(&(p->dictionary), hval, tpl);
+	    ARRAY_PUT(p->dictionary, hval, tpl);
 	    return tuple_val(old)[2];
 	} else {
 	    hp = HeapOnlyAlloc(p, 4);
 	    tmp = CONS(hp, old, NIL);
 	    hp += 2;
 	    ++(p->dictionary->numElements);
-	    array_put(&(p->dictionary), hval, CONS(hp, tpl, tmp));
+	    ARRAY_PUT(p->dictionary, hval, CONS(hp, tpl, tmp));
 	    hp += 2;
 	    ASSERT(hp <= hp_limit);
 	}
@@ -624,7 +661,7 @@ static Eterm pd_hash_put(Process *p, Eterm id, Eterm value)
 	     * New key. Simply prepend the tuple to the beginning of the list.
 	     */
 	    hp = HeapOnlyAlloc(p, 2);
-	    array_put(&(p->dictionary), hval, CONS(hp, tpl, old));
+	    ARRAY_PUT(p->dictionary, hval, CONS(hp, tpl, old));
 	    hp += 2;
 	    ASSERT(hp <= hp_limit);
 	    ++(p->dictionary->numElements);
@@ -659,7 +696,7 @@ static Eterm pd_hash_put(Process *p, Eterm id, Eterm value)
 	    nlist = CONS(hp, tpl, nlist);
 	    hp += 2;
 	    ASSERT(hp <= hp_limit);
-	    array_put(&(p->dictionary), hval, nlist);
+	    ARRAY_PUT(p->dictionary, hval, nlist);
 	    return tuple_val(TCAR(tmp))[2];
 	}
     } else {
@@ -684,6 +721,7 @@ static Eterm pd_hash_put(Process *p, Eterm id, Eterm value)
 
 static void shrink(Process *p, Eterm* ret) 
 {
+    ProcDict *pd = p->dictionary;
     unsigned int range = HASH_RANGE(p->dictionary);
     unsigned int steps = (range*3) / 10;
     Eterm hi, lo, tmp;
@@ -698,25 +736,26 @@ static void shrink(Process *p, Eterm* ret)
     }
 
     for (i = 0; i < steps; ++i) {
-	ProcDict *pd = p->dictionary;
 	if (pd->splitPosition == 0) {
-	    pd->homeSize /= 2;
-	    pd->splitPosition = pd->homeSize;
+            ASSERT(IS_POW2(pd->usedSlots));
+            pd->sizeMask = pd->usedSlots - 1;
+	    pd->splitPosition = pd->usedSlots / 2;
 	}
 	--(pd->splitPosition);
-	hi = ARRAY_GET(pd, (pd->splitPosition + pd->homeSize));
+        /* Must wait to decrement 'usedSlots' for GC rootset below */
+	hi = ARRAY_GET(pd, pd->usedSlots - 1);
 	lo = ARRAY_GET(pd, pd->splitPosition);
 	if (hi != NIL) {
 	    if (lo == NIL) {
-		array_put(&(p->dictionary), pd->splitPosition, hi);
+		ARRAY_PUT(pd, pd->splitPosition, hi);
 	    } else {
 		int needed = 4;
 		if (is_list(hi) && is_list(lo)) {
 		    needed = 2*erts_list_length(hi);
 		}
 		if (HeapWordsLeft(p) < needed) {
-		    BUMP_REDS(p, erts_garbage_collect(p, needed, ret, 1));
-		    hi = pd->data[(pd->splitPosition + pd->homeSize)];
+		    erts_garbage_collect(p, needed, ret, 1);
+		    hi = pd->data[pd->usedSlots - 1];
 		    lo = pd->data[pd->splitPosition];
 		}
 #ifdef DEBUG
@@ -727,13 +766,13 @@ static void shrink(Process *p, Eterm* ret)
 			hp = HeapOnlyAlloc(p, 4);
 			tmp = CONS(hp, hi, NIL);
 			hp += 2;
-			array_put(&(p->dictionary), pd->splitPosition, 
+			ARRAY_PUT(pd, pd->splitPosition,
 				  CONS(hp,lo,tmp));
 			hp += 2;
 			ASSERT(hp <= hp_limit);
 		    } else { /* hi is a list */
 			hp = HeapOnlyAlloc(p, 2);
-			array_put(&(p->dictionary), pd->splitPosition, 
+			ARRAY_PUT(pd, pd->splitPosition,
 				  CONS(hp, lo, hi));
 			hp += 2;
 			ASSERT(hp <= hp_limit);
@@ -741,7 +780,7 @@ static void shrink(Process *p, Eterm* ret)
 		} else { /* lo is a list */
 		    if (is_tuple(hi)) {
 			hp = HeapOnlyAlloc(p, 2);
-			array_put(&(p->dictionary), pd->splitPosition, 
+			ARRAY_PUT(pd, pd->splitPosition,
 				  CONS(hp, hi, lo));
 			hp += 2;
 			ASSERT(hp <= hp_limit);
@@ -753,14 +792,15 @@ static void shrink(Process *p, Eterm* ret)
 			    hp += 2;
 			}
 			ASSERT(hp <= hp_limit);
-			array_put(&(p->dictionary), pd->splitPosition, lo);
+			ARRAY_PUT(pd, pd->splitPosition, lo);
 		    }
 		}
 	    }
 	}
-	array_put(&(p->dictionary), (pd->splitPosition + pd->homeSize), NIL);
+        --pd->usedSlots;
+	ARRAY_PUT(pd, pd->usedSlots, NIL);
     }
-    if (HASH_RANGE(p->dictionary) <= (p->dictionary->size / 4)) {
+    if (HASH_RANGE(p->dictionary) <= (p->dictionary->arraySize / 4)) {
 	array_shrink(&(p->dictionary), (HASH_RANGE(p->dictionary) * 3) / 2);
     }
 }
@@ -768,14 +808,14 @@ static void shrink(Process *p, Eterm* ret)
 static void grow(Process *p)
 {
     unsigned int i,j;
-    unsigned int steps = p->dictionary->homeSize / 5;
+    unsigned int steps = (p->dictionary->usedSlots / 4) & 0xf;
     Eterm l1,l2;
     Eterm l;
     Eterm *hp;
     unsigned int pos;
     unsigned int homeSize;
     int needed = 0;
-    ProcDict *pd;
+    ProcDict *pd = p->dictionary;
 #ifdef DEBUG
     Eterm *hp_limit;
 #endif
@@ -784,18 +824,20 @@ static void grow(Process *p)
     if (steps == 0)
 	steps = 1;
     /* Dont grow over MAX_HASH */
-    if ((MAX_HASH - steps) <= HASH_RANGE(p->dictionary)) {
+    if ((MAX_HASH - steps) <= HASH_RANGE(pd)) {
 	return;
     }
 
+    ensure_array_size(&p->dictionary, HASH_RANGE(pd) + steps);
+    pd = p->dictionary;
+
     /*
      * Calculate total number of heap words needed, and garbage collect
      * if necessary.
      */
 
-    pd = p->dictionary;
     pos = pd->splitPosition;
-    homeSize = pd->homeSize;
+    homeSize = pd->usedSlots - pd->splitPosition;
     for (i = 0; i < steps; ++i) {
 	if (pos == homeSize) {
 	    homeSize *= 2;
@@ -811,7 +853,7 @@ static void grow(Process *p)
 	}
     }
     if (HeapWordsLeft(p) < needed) {
-	BUMP_REDS(p, erts_garbage_collect(p, needed, 0, 0));
+	erts_garbage_collect(p, needed, 0, 0);
     }
 #ifdef DEBUG
     hp_limit = p->htop + needed;
@@ -820,21 +862,22 @@ static void grow(Process *p)
     /*
      * Now grow.
      */
-
+    homeSize = pd->usedSlots - pd->splitPosition;
     for (i = 0; i < steps; ++i) {
-	ProcDict *pd = p->dictionary;
-	if (pd->splitPosition == pd->homeSize) {
-	    pd->homeSize *= 2;
-	    pd->splitPosition = 0;
+	if (pd->splitPosition == homeSize) {
+	    homeSize *= 2;
+            pd->sizeMask = homeSize*2 - 1;
+            pd->splitPosition = 0;
 	}
 	pos = pd->splitPosition;
 	++pd->splitPosition; /* For the hashes */
+        ++pd->usedSlots;
+        ASSERT(pos + homeSize == pd->usedSlots - 1);
 	l = ARRAY_GET(pd, pos);
 	if (is_tuple(l)) {
 	    if (pd_hash_value(pd, tuple_val(l)[1]) != pos) {
-		array_put(&(p->dictionary), pos + 
-			  p->dictionary->homeSize, l);
-		array_put(&(p->dictionary), pos, NIL);
+		ARRAY_PUT(pd, pos + homeSize, l);
+		ARRAY_PUT(pd, pos, NIL);
 	    }
 	} else {
 	    l2 = NIL;
@@ -856,10 +899,8 @@ static void grow(Process *p)
 	    if (l2 != NIL && TCDR(l2) == NIL)
 		l2 = TCAR(l2);
 	    ASSERT(hp <= hp_limit);
-	    /* After array_put pd is no longer valid */
-	    array_put(&(p->dictionary), pos, l1);
-	    array_put(&(p->dictionary), pos + 
-		      p->dictionary->homeSize, l2);
+	    ARRAY_PUT(pd, pos, l1);
+	    ARRAY_PUT(pd, pos + homeSize, l2);
 	}
     }
 
@@ -876,73 +917,65 @@ static void array_shrink(ProcDict **ppd, unsigned int need)
 {
     unsigned int siz = next_array_size(need);
 
-    HDEBUGF(("array_shrink: size = %d, used = %d, need = %d",
-	     (*ppd)->size, (*ppd)->used, need));
+    HDEBUGF(("array_shrink: size = %d, need = %d",
+	     (*ppd)->arraySize, need));
 
-    if (siz > (*ppd)->size)
+    if (siz >= (*ppd)->arraySize)
 	return; /* Only shrink */
 
     *ppd = PD_REALLOC(((void *) *ppd),
-		      PD_SZ2BYTES((*ppd)->size),
+		      PD_SZ2BYTES((*ppd)->arraySize),
 		      PD_SZ2BYTES(siz));
 
-    (*ppd)->size = siz;
-    if ((*ppd)->size < (*ppd)->used)
-	(*ppd)->used = (*ppd)->size;
+    (*ppd)->arraySize = siz;
 }
     
 			
-static Eterm array_put(ProcDict **ppdict, unsigned int ndx, Eterm term)
+static void ensure_array_size(ProcDict **ppdict, unsigned int size)
 {
+    ProcDict *pd = *ppdict;
     unsigned int i;
-    Eterm ret;
-    if (*ppdict == NULL) {
-	Uint siz = next_array_size(ndx+1);
-	ProcDict *p;
 
-        p = PD_ALLOC(PD_SZ2BYTES(siz));
+    if (pd == NULL) {
+	Uint siz = next_array_size(size);
+
+        pd = PD_ALLOC(PD_SZ2BYTES(siz));
 	for (i = 0; i < siz; ++i) 
-	    p->data[i] = NIL;
-	p->size = siz;
-	p->homeSize = p->splitPosition = p->numElements = p->used = 0;
-	*ppdict = p;
-    } else if (ndx >= (*ppdict)->size) {
-	Uint osize = (*ppdict)->size;
-	Uint nsize = next_array_size(ndx+1);
-	*ppdict = PD_REALLOC(((void *) *ppdict),
+	    pd->data[i] = NIL;
+	pd->arraySize = siz;
+        *ppdict = pd;
+    } else if (size > pd->arraySize) {
+	Uint osize = pd->arraySize;
+	Uint nsize = next_array_size(size);
+	pd = PD_REALLOC(((void *) pd),
 			     PD_SZ2BYTES(osize),
 			     PD_SZ2BYTES(nsize));
 	for (i = osize; i < nsize; ++i)
-	    (*ppdict)->data[i] = NIL;
-	(*ppdict)->size = nsize;
+	    pd->data[i] = NIL;
+	pd->arraySize = nsize;
+        *ppdict = pd;
     }
-    ret = (*ppdict)->data[ndx];
-    (*ppdict)->data[ndx] = term;
-    if ((ndx + 1) > (*ppdict)->used)
-	(*ppdict)->used = ndx + 1;
-#ifdef HARDDEBUG
-    HDEBUGF(("array_put: (*ppdict)->size = %d, (*ppdict)->used = %d, ndx = %d",
-	     (*ppdict)->size, (*ppdict)->used, ndx));
-    erts_fprintf(stderr, "%T", term);
-#endif /* HARDDEBUG */
-    return ret;
 }
 
 /*
 ** Basic utilities
 */
 
-static unsigned int pd_hash_value(ProcDict *pdict, Eterm term) 
+static unsigned int pd_hash_value_to_ix(ProcDict *pdict, Uint32 hx) 
 { 
-    Uint hash, high;
+    Uint high;
 
-    hash = MAKE_HASH(term);
-    high = hash % (pdict->homeSize*2);
+    ASSERT(IS_POW2(pdict->sizeMask+1));
+    ASSERT(HASH_RANGE(pdict) >= (pdict->sizeMask >> 1));
+    ASSERT(HASH_RANGE(pdict) <= (pdict->sizeMask + 1));
+
+    high = hx & pdict->sizeMask;
     if (high >= HASH_RANGE(pdict))
-	return hash % pdict->homeSize;
+	return hx & (pdict->sizeMask >> 1);
     return high;
 }
 
+
 static unsigned int next_array_size(unsigned int need)
 {
     static unsigned int tab[] =
@@ -1002,24 +1035,28 @@ static unsigned int next_array_size(unsigned int need)
 static void pd_check(ProcDict *pd) 
 {
     unsigned int i;
+    unsigned int used;
     Uint num;
     if (pd == NULL)
 	return;
-    ASSERT(pd->size >= pd->used);
+    used = HASH_RANGE(pd);
+    ASSERT(pd->arraySize >= used);
     ASSERT(HASH_RANGE(pd) <= MAX_HASH);
-    for (i = 0, num = 0; i < pd->used; ++i) {
+    for (i = 0, num = 0; i < used; ++i) {
 	Eterm t = pd->data[i];
 	if (is_nil(t)) {
 	    continue;
 	} else if (is_tuple(t)) {
 	    ++num;
 	    ASSERT(arityval(*tuple_val(t)) == 2);
+            ASSERT(pd_hash_value(pd, tuple_val(t)[1]) == i);
 	    continue;
 	} else if (is_list(t)) {
 	    while (t != NIL) {
 		++num;
 		ASSERT(is_tuple(TCAR(t)));
 		ASSERT(arityval(*(tuple_val(TCAR(t)))) == 2);
+                ASSERT(pd_hash_value(pd, tuple_val(TCAR(t))[1]) == i);
 		t = TCDR(t);
 	    }
 	    continue;
@@ -1030,7 +1067,7 @@ static void pd_check(ProcDict *pd)
 	}
     }
     ASSERT(num == pd->numElements);
-    ASSERT(pd->splitPosition <= pd->homeSize);
+    ASSERT(pd->usedSlots >= pd->splitPosition*2);
 }
 
 #endif /* DEBUG */
diff --git a/erts/emulator/beam/erl_process_dict.h b/erts/emulator/beam/erl_process_dict.h
index cc53800eb5..dac214c8a1 100644
--- a/erts/emulator/beam/erl_process_dict.h
+++ b/erts/emulator/beam/erl_process_dict.h
@@ -23,14 +23,18 @@
 #include "sys.h"
 
 typedef struct proc_dict {
-    unsigned int size;
-    unsigned int used;
-    unsigned int homeSize;
+    unsigned int sizeMask;
+    unsigned int usedSlots;
+    unsigned int arraySize;
     unsigned int splitPosition;
     Uint numElements;
     Eterm data[1]; /* The beginning of an array of erlang terms */
 } ProcDict;
 
+#define ERTS_PD_START(PD) ((PD)->data)
+#define ERTS_PD_SIZE(PD)  ((PD)->usedSlots)
+
+int erts_pd_set_initial_size(int size);
 Uint erts_dicts_mem_size(struct process *p);
 void erts_erase_dicts(struct process *p);
 void erts_dictionary_dump(int to, void *to_arg, ProcDict *pd);
@@ -39,5 +43,6 @@ void erts_deep_dictionary_dump(int to, void *to_arg,
 Eterm erts_dictionary_copy(struct process *p, ProcDict *pd);
 
 Eterm erts_pd_hash_get(struct process *p, Eterm id);
+Eterm erts_pd_hash_get_with_hx(Process *p, Uint32 hx, Eterm id);
 
 #endif
diff --git a/erts/emulator/beam/erl_process_dump.c b/erts/emulator/beam/erl_process_dump.c
index 25f0b1ed38..71396561a3 100644
--- a/erts/emulator/beam/erl_process_dump.c
+++ b/erts/emulator/beam/erl_process_dump.c
@@ -78,13 +78,14 @@ erts_deep_process_dump(int to, void *to_arg)
     dump_binaries(to, to_arg, all_binaries);
 }
 
-Uint erts_process_memory(Process *p) {
-  ErlMessage *mp;
+Uint erts_process_memory(Process *p, int incl_msg_inq) {
+  ErtsMessage *mp;
   Uint size = 0;
   struct saved_calls *scb;
   size += sizeof(Process);
 
-  ERTS_SMP_MSGQ_MV_INQ2PRIVQ(p);
+  if (incl_msg_inq)
+      ERTS_SMP_MSGQ_MV_INQ2PRIVQ(p);
 
   erts_doforall_links(ERTS_P_LINKS(p), &erts_one_link_size, &size);
   erts_doforall_monitors(ERTS_P_MONITORS(p), &erts_one_mon_size, &size);
@@ -92,7 +93,7 @@ Uint erts_process_memory(Process *p) {
   if (p->old_hend && p->old_heap)
     size += (p->old_hend - p->old_heap) * sizeof(Eterm);
 
-  size += p->msg.len * sizeof(ErlMessage);
+  size += p->msg.len * sizeof(ErtsMessage);
 
   for (mp = p->msg.first; mp; mp = mp->next)
     if (mp->data.attached)
@@ -119,7 +120,7 @@ static void
 dump_process_info(int to, void *to_arg, Process *p)
 {
     Eterm* sp;
-    ErlMessage* mp;
+    ErtsMessage* mp;
     int yreg = -1;
 
     ERTS_SMP_MSGQ_MV_INQ2PRIVQ(p);
@@ -365,7 +366,7 @@ heap_dump(int to, void *to_arg, Eterm x)
 
     while (x != OUR_NIL) {
 	if (is_CP(x)) {
-	    next = (Eterm *) EXPAND_POINTER(x);
+	    next = (Eterm *) x;
 	} else if (is_list(x)) {
 	    ptr = list_val(x);
 	    if (ptr[0] != OUR_NIL) {
@@ -378,7 +379,7 @@ heap_dump(int to, void *to_arg, Eterm x)
 		    ptr[1] = make_small(0);
 		}
 		x = ptr[0];
-		ptr[0] = (Eterm) COMPRESS_POINTER(next);
+		ptr[0] = (Eterm) next;
 		next = ptr + 1;
 		continue;
 	    }
@@ -408,7 +409,7 @@ heap_dump(int to, void *to_arg, Eterm x)
 			ptr[0] = OUR_NIL;
 		    } else {
 			x = ptr[arity];
-			ptr[0] = (Eterm) COMPRESS_POINTER(next);
+			ptr[0] = (Eterm) next;
 			next = ptr + arity - 1;
 			continue;
 		    }
@@ -657,6 +658,8 @@ erts_dump_extended_process_state(int to, void *to_arg, erts_aint32_t psflg) {
                 erts_print(to, to_arg, "PROXY"); break;
             case ERTS_PSFLG_DELAYED_SYS:
                 erts_print(to, to_arg, "DELAYED_SYS"); break;
+            case ERTS_PSFLG_OFF_HEAP_MSGQ:
+                erts_print(to, to_arg, "OFF_HEAP_MSGQ"); break;
 #ifdef ERTS_DIRTY_SCHEDULERS
             case ERTS_PSFLG_DIRTY_CPU_PROC:
                 erts_print(to, to_arg, "DIRTY_CPU_PROC"); break;
diff --git a/erts/emulator/beam/erl_process_lock.h b/erts/emulator/beam/erl_process_lock.h
index 788348e613..a64c993e8f 100644
--- a/erts/emulator/beam/erl_process_lock.h
+++ b/erts/emulator/beam/erl_process_lock.h
@@ -854,9 +854,6 @@ ERTS_GLB_INLINE void erts_proc_dec_refc(Process *p)
 #endif
     if (!referred) {
 	ASSERT(ERTS_PROC_IS_EXITING(p));
-	ASSERT(ERTS_AINT_NULL
-	       == erts_ptab_pix2intptr_ddrb(&erts_proc,
-					    internal_pid_index(p->common.id)));
 	erts_free_proc(p);
     }
 }
@@ -872,9 +869,6 @@ ERTS_GLB_INLINE void erts_proc_add_refc(Process *p, Sint add_refc)
 #endif
     if (!referred) {
 	ASSERT(ERTS_PROC_IS_EXITING(p));
-	ASSERT(ERTS_AINT_NULL
-	       == erts_ptab_pix2intptr_ddrb(&erts_proc,
-					    internal_pid_index(p->common.id)));
 	erts_free_proc(p);
     }
 }
diff --git a/erts/emulator/beam/erl_term.c b/erts/emulator/beam/erl_term.c
index 89459fb278..e31d3a951d 100644
--- a/erts/emulator/beam/erl_term.c
+++ b/erts/emulator/beam/erl_term.c
@@ -28,6 +28,37 @@
 #include <stdlib.h>
 #include <stdio.h>
 
+void
+erts_set_literal_tag(Eterm *term, Eterm *hp_start, Eterm hsz)
+{
+#ifdef TAG_LITERAL_PTR
+    Eterm *hp_end, *hp;
+    
+    hp_end = hp_start + hsz;
+    hp = hp_start;
+
+    while (hp < hp_end) {
+	switch (primary_tag(*hp)) {
+	case TAG_PRIMARY_BOXED:
+	case TAG_PRIMARY_LIST:
+	    *hp |= TAG_LITERAL_PTR;
+	    break;
+	case TAG_PRIMARY_HEADER:
+	    if (header_is_thing(*hp)) {
+		hp += thing_arityval(*hp);
+	    }
+	    break;
+	default:
+	    break;
+	}
+	
+	hp++;
+    }
+    if (is_boxed(*term) || is_list(*term))
+	*term |= TAG_LITERAL_PTR;
+#endif
+}
+
 __decl_noreturn static void __noreturn
 et_abort(const char *expr, const char *file, unsigned line)
 {
@@ -91,6 +122,7 @@ unsigned tag_val_def(Wterm x)
 	    case (_TAG_HEADER_REFC_BIN >> _TAG_PRIMARY_SIZE):	return BINARY_DEF;
 	    case (_TAG_HEADER_HEAP_BIN >> _TAG_PRIMARY_SIZE):	return BINARY_DEF;
 	    case (_TAG_HEADER_SUB_BIN >> _TAG_PRIMARY_SIZE):	return BINARY_DEF;
+	    case (_TAG_HEADER_BIN_MATCHSTATE >> _TAG_PRIMARY_SIZE): return MATCHSTATE_DEF;
 	  }
  
 	  break;
@@ -173,9 +205,7 @@ ET_DEFINE_CHECKED(Uint,external_thing_data_words,ExternalThing*,is_thing_ptr);
 ET_DEFINE_CHECKED(Eterm,make_cp,UWord *,_is_taggable_pointer);
 ET_DEFINE_CHECKED(UWord *,cp_val,Eterm,is_CP);
 ET_DEFINE_CHECKED(Uint,catch_val,Eterm,is_catch);
-ET_DEFINE_CHECKED(Uint,x_reg_offset,Uint,_is_xreg);
-ET_DEFINE_CHECKED(Uint,y_reg_offset,Uint,_is_yreg);
-ET_DEFINE_CHECKED(Uint,x_reg_index,Uint,_is_xreg);
-ET_DEFINE_CHECKED(Uint,y_reg_index,Uint,_is_yreg);
+ET_DEFINE_CHECKED(Uint,loader_x_reg_index,Uint,_is_loader_x_reg);
+ET_DEFINE_CHECKED(Uint,loader_y_reg_index,Uint,_is_loader_y_reg);
 
 #endif	/* ET_DEBUG */
diff --git a/erts/emulator/beam/erl_term.h b/erts/emulator/beam/erl_term.h
index 90e35151b0..2b28762db5 100644
--- a/erts/emulator/beam/erl_term.h
+++ b/erts/emulator/beam/erl_term.h
@@ -21,34 +21,10 @@
 #ifndef __ERL_TERM_H
 #define __ERL_TERM_H
 
-#include "sys.h" /* defines HALFWORD_HEAP */
+#include "erl_mmap.h"
 
 typedef UWord Wterm;  /* Full word terms */
 
-#if HALFWORD_HEAP
-#  define HEAP_ON_C_STACK 0
-#  if HALFWORD_ASSERT
-#    ifdef ET_DEBUG
-#      undef ET_DEBUG
-#    endif
-#    define ET_DEBUG 1
-#  endif
-#  if 1
-#    define CHECK_POINTER_MASK 0xFFFFFFFF00000000UL
-#    define COMPRESS_POINTER(APointer) ((Eterm) (UWord) (APointer))
-#    define EXPAND_POINTER(AnEterm) ((UWord) (AnEterm))
-#  else
-#    define CHECK_POINTER_MASK 0x0UL
-#    define COMPRESS_POINTER(AnUint) (AnUint)
-#    define EXPAND_POINTER(APointer) (APointer)
-#  endif
-#else
-#  define HEAP_ON_C_STACK 1
-#  define CHECK_POINTER_MASK 0x0UL
-#  define COMPRESS_POINTER(AnUint) (AnUint)
-#  define EXPAND_POINTER(APointer) (APointer)
-#endif
-
 struct erl_node_; /* Declared in erl_node_tables.h */
 
 /*
@@ -74,6 +50,24 @@ struct erl_node_; /* Declared in erl_node_tables.h */
 #define _ET_APPLY(F,X)	_unchecked_##F(X)
 #endif
 
+#if defined(ARCH_64)
+#  define TAG_PTR_MASK__	0x7
+#  if !defined(ERTS_HAVE_OS_PHYSICAL_MEMORY_RESERVATION)
+#    ifdef HIPE
+#      error Hipe on 64-bit needs a real mmap as it does not support the literal tag
+#    endif
+#    define TAG_LITERAL_PTR	0x4
+#  else
+#    undef TAG_LITERAL_PTR
+#  endif
+#elif defined(ARCH_32)
+#  define TAG_PTR_MASK__	0x3
+#  undef TAG_LITERAL_PTR
+#else
+#  error Not supported arch
+#endif
+
+
 #define _TAG_PRIMARY_SIZE	2
 #define _TAG_PRIMARY_MASK	0x3
 #define TAG_PRIMARY_HEADER	0x0
@@ -190,15 +184,13 @@ struct erl_node_; /* Declared in erl_node_tables.h */
 
 
 /* boxed object access methods */
-#if HALFWORD_HEAP
-#define _is_taggable_pointer(x)	 (((UWord)(x) & (CHECK_POINTER_MASK | 0x3)) == 0)
-#define _boxed_precond(x)        (is_boxed(x))
-#else
-#define _is_taggable_pointer(x)	 (((Uint)(x) & 0x3) == 0)
+
+#define _is_taggable_pointer(x)	 (((Uint)(x) & TAG_PTR_MASK__) == 0)
+
 #define  _boxed_precond(x)       (is_boxed(x))
-#endif
-#define _is_aligned(x)		(((Uint)(x) & 0x3) == 0)
-#define _unchecked_make_boxed(x) ((Uint) COMPRESS_POINTER(x) + TAG_PRIMARY_BOXED)
+
+#define _is_aligned(x)		(((Uint)(x) & TAG_PTR_MASK__) == 0)
+#define _unchecked_make_boxed(x) ((Uint)(x) + TAG_PRIMARY_BOXED)
 _ET_DECLARE_CHECKED(Eterm,make_boxed,const Eterm*)
 #define make_boxed(x)		_ET_APPLY(make_boxed,(x))
 #if 1
@@ -209,12 +201,16 @@ _ET_DECLARE_CHECKED(int,is_boxed,Eterm)
 #else
 #define is_boxed(x)		(((x) & _TAG_PRIMARY_MASK) == TAG_PRIMARY_BOXED)
 #endif
-#define _unchecked_boxed_val(x) ((Eterm*) EXPAND_POINTER(((x) - TAG_PRIMARY_BOXED)))
+#ifdef TAG_LITERAL_PTR
+#define _unchecked_boxed_val(x) _unchecked_ptr_val(x)
+#else
+#define _unchecked_boxed_val(x) ((Eterm*) ((x) - TAG_PRIMARY_BOXED))
+#endif
 _ET_DECLARE_CHECKED(Eterm*,boxed_val,Wterm)
 #define boxed_val(x)		_ET_APPLY(boxed_val,(x))
 
 /* cons cell ("list") access methods */
-#define _unchecked_make_list(x)	((Uint) COMPRESS_POINTER(x) + TAG_PRIMARY_LIST)
+#define _unchecked_make_list(x)	((Uint)(x) + TAG_PRIMARY_LIST)
 _ET_DECLARE_CHECKED(Eterm,make_list,const Eterm*)
 #define make_list(x)		_ET_APPLY(make_list,(x))
 #if 1
@@ -226,12 +222,12 @@ _ET_DECLARE_CHECKED(int,is_not_list,Eterm)
 #define is_list(x)		(((x) & _TAG_PRIMARY_MASK) == TAG_PRIMARY_LIST)
 #define is_not_list(x)		(!is_list((x)))
 #endif
-#if HALFWORD_HEAP
 #define _list_precond(x)        (is_list(x))
+#ifdef TAG_LITERAL_PTR
+#define _unchecked_list_val(x) _unchecked_ptr_val(x)
 #else
-#define _list_precond(x)       (is_list(x))
+#define _unchecked_list_val(x) ((Eterm*) ((x) - TAG_PRIMARY_LIST))
 #endif
-#define _unchecked_list_val(x) ((Eterm*) EXPAND_POINTER((x) - TAG_PRIMARY_LIST))
 _ET_DECLARE_CHECKED(Eterm*,list_val,Wterm)
 #define list_val(x)		_ET_APPLY(list_val,(x))
 
@@ -242,15 +238,22 @@ _ET_DECLARE_CHECKED(Eterm*,list_val,Wterm)
 #define CDR(x)  ((x)[1])
 
 /* generic tagged pointer (boxed or list) access methods */
-#define _unchecked_ptr_val(x)	((Eterm*) EXPAND_POINTER((x) & ~((Uint) 0x3)))
+#define _unchecked_ptr_val(x)	((Eterm*) ((x) & ~((Uint) TAG_PTR_MASK__)))
 #define ptr_val(x)		_unchecked_ptr_val((x))	/*XXX*/
 #define _unchecked_offset_ptr(x,offs)	((x)+((offs)*sizeof(Eterm)))
 #define offset_ptr(x,offs)	_unchecked_offset_ptr(x,offs)	/*XXX*/
 #define _unchecked_byte_offset_ptr(x,byte_offs)	((x)+(offs))
 #define byte_offset_ptr(x,offs) _unchecked_byte_offset_ptr(x,offs)  /*XXX*/
 
+#ifdef TAG_LITERAL_PTR
+#define _unchecked_is_not_literal_ptr(x) (!((x) & TAG_LITERAL_PTR))
+#define is_not_literal_ptr(x) _unchecked_is_not_literal_ptr((x)) /*XXX*/
+#define is_literal_ptr(x) (!is_not_literal_ptr((x))) /*XXX*/
+#endif
+
+
 /* fixnum ("small") access methods */
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
 #define SMALL_BITS	(64-4)
 #define SMALL_DIGITS	(17)
 #else
@@ -396,11 +399,7 @@ _ET_DECLARE_CHECKED(Eterm*,fun_val,Wterm)
 _ET_DECLARE_CHECKED(Eterm*,export_val,Wterm)
 #define export_val(x)	_ET_APPLY(export_val,(x))
 #define is_export_header(x)	((x) == HEADER_EXPORT)
-#if HALFWORD_HEAP
-#define HEADER_EXPORT   _make_header(2,_TAG_HEADER_EXPORT)
-#else
 #define HEADER_EXPORT   _make_header(1,_TAG_HEADER_EXPORT)
-#endif
 
 /* bignum access methods */
 #define make_pos_bignum_header(sz)	_make_header((sz),_TAG_HEADER_POS_BIG)
@@ -424,7 +423,7 @@ _ET_DECLARE_CHECKED(Eterm*,big_val,Wterm)
 #define big_val(x)		_ET_APPLY(big_val,(x))
 
 /* flonum ("float") access methods */
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
 #define HEADER_FLONUM   _make_header(1,_TAG_HEADER_FLOAT)
 #else
 #define HEADER_FLONUM	_make_header(2,_TAG_HEADER_FLOAT)
@@ -445,12 +444,12 @@ typedef union float_def
     byte   fb[sizeof(ieee754_8)];
     Uint16 fs[sizeof(ieee754_8) / sizeof(Uint16)];
     Uint32 fw[sizeof(ieee754_8) / sizeof(Uint32)];
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
     Uint   fdw;
 #endif
 } FloatDef;
 
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
 
 #define FLOAT_VAL_GET_DOUBLE(fval, f) (f).fdw = *((fval)+1)
 
@@ -727,7 +726,7 @@ _ET_DECLARE_CHECKED(struct erl_node_*,internal_port_node,Eterm)
 #define ERTS_MAX_REF_NUMBERS	3
 #define ERTS_REF_NUMBERS	ERTS_MAX_REF_NUMBERS
 
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
 #  define ERTS_REF_WORDS	(ERTS_REF_NUMBERS/2 + 1)
 #  define ERTS_REF_32BIT_WORDS  (ERTS_REF_NUMBERS+1)
 #else
@@ -749,7 +748,7 @@ typedef struct {
 #define make_ref_thing_header(DW) \
   _make_header((DW)+REF_THING_HEAD_SIZE-1,_TAG_HEADER_REF)
 
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
 
 /*
  * Ref layout on a 64-bit little endian machine:
@@ -1009,27 +1008,20 @@ _ET_DECLARE_CHECKED(struct erl_node_*,external_ref_node,Eterm)
 #define MAP_HEADER_VAL(Hdr)   (((Hdr) >> (_HEADER_ARITY_OFFS + MAP_HEADER_TAG_SZ + MAP_HEADER_ARITY_SZ)) & (0xffff))
 
 #define make_hashmap(x)		      make_boxed((Eterm*)(x))
-#define make_hashmap_rel 	      make_boxed_rel
 #define is_hashmap(x)		      (is_boxed((x)) && is_hashmap_header(*boxed_val((x))))
 #define is_not_hashmap(x)             (!is_hashmap(x))
-#define is_hashmap_rel(RTERM,BASE)    is_hashmap(rterm2wterm(RTERM,BASE))
 #define is_hashmap_header(x)	      (((x) & (_HEADER_MAP_HASHMAP_HEAD_MASK)) == HAMT_SUBTAG_HEAD_ARRAY)
 #define hashmap_val(x)		      _unchecked_boxed_val((x))
-#define hashmap_val_rel(RTERM, BASE)  hashmap_val(rterm2wterm(RTERM, BASE))
 
 #define make_flatmap(x)               make_boxed((Eterm*)(x))
-#define make_flatmap_rel(x, BASE)     make_boxed_rel((Eterm*)(x),(BASE))
 #define is_flatmap(x)                 (is_boxed((x)) && is_flatmap_header(*boxed_val((x))))
-#define is_flatmap_rel(RTERM,BASE)    is_flatmap(rterm2wterm(RTERM,BASE))
 #define is_not_flatmap(x)             (!is_flatmap((x)))
 #define is_flatmap_header(x)          (((x) & (_HEADER_MAP_SUBTAG_MASK)) == HAMT_SUBTAG_HEAD_FLATMAP)
 #define flatmap_val(x)                (_unchecked_boxed_val((x)))
-#define flatmap_val_rel(RTERM, BASE)  flatmap_val(rterm2wterm(RTERM, BASE))
 
 #define is_map_header(x)       (((x) & (_TAG_HEADER_MASK)) == _TAG_HEADER_MAP)
 #define is_map(x)              (is_boxed((x)) && is_map_header(*boxed_val(x)))
 #define is_not_map(x)          (!is_map(x))
-#define is_map_rel(RTERM,BASE) is_map(rterm2wterm(RTERM,BASE))
 
 /* number tests */
 
@@ -1048,14 +1040,14 @@ _ET_DECLARE_CHECKED(struct erl_node_*,external_ref_node,Eterm)
 #error "fix yer arch, like"
 #endif
 
-#define _unchecked_make_cp(x)	((Eterm) COMPRESS_POINTER(x))
+#define _unchecked_make_cp(x)	((Eterm)(x))
 _ET_DECLARE_CHECKED(Eterm,make_cp,BeamInstr*)
 #define make_cp(x)	_ET_APPLY(make_cp,(x))
 
 #define is_not_CP(x)	((x) & _CPMASK)
 #define is_CP(x)	(!is_not_CP(x))
 
-#define _unchecked_cp_val(x)	((BeamInstr*) EXPAND_POINTER(x))
+#define _unchecked_cp_val(x)	((BeamInstr*) (x))
 _ET_DECLARE_CHECKED(BeamInstr*,cp_val,Eterm)
 #define cp_val(x)	_ET_APPLY(cp_val,(x))
 
@@ -1071,44 +1063,40 @@ _ET_DECLARE_CHECKED(Uint,catch_val,Eterm)
 /*
  * Overloaded tags.
  *
- * SMALL = 15
- * ATOM/NIL=7
+ * In the loader, we want to tag a term in a way so that it can
+ * be any literal (atom/integer/float/tuple/list/binary) or a
+ * register.
  *
- * Note that the two least significant bits in SMALL/ATOM/NIL always are 3;
- * thus, we can distinguish register from literals by looking at only these
- * two bits.
+ * We can achive that by overloading the PID and PORT tags to
+ * mean X and Y registers. That works because there are no
+ * pid or port literals.
  */
 
-#define X_REG_DEF	0
-#define Y_REG_DEF	1
-#define R_REG_DEF	2
+#define _LOADER_TAG_XREG _TAG_IMMED1_PID
+#define _LOADER_TAG_YREG _TAG_IMMED1_PORT
+#define _LOADER_TAG_SIZE _TAG_IMMED1_SIZE
+#define _LOADER_MASK _TAG_IMMED1_MASK
 
-#define beam_reg_tag(x)	((x) & 3)
+#define LOADER_X_REG _LOADER_TAG_XREG
+#define LOADER_Y_REG _LOADER_TAG_YREG
 
-#define make_rreg()	R_REG_DEF
-#define make_xreg(ix)	(((ix) * sizeof(Eterm)) | X_REG_DEF)
-#define make_yreg(ix)	(((ix) * sizeof(Eterm)) | Y_REG_DEF)
+#define make_loader_x_reg(R) (((R) << _LOADER_TAG_SIZE) | _LOADER_TAG_XREG)
+#define make_loader_y_reg(R) (((R) << _LOADER_TAG_SIZE) | _LOADER_TAG_YREG)
 
-#define _is_xreg(x)	(beam_reg_tag(x) == X_REG_DEF)
-#define _is_yreg(x)	(beam_reg_tag(x) == Y_REG_DEF)
+#define loader_reg_index(R) ((R) >> _LOADER_TAG_SIZE)
 
-#define _unchecked_x_reg_offset(R)	((R) - X_REG_DEF)
-_ET_DECLARE_CHECKED(Uint,x_reg_offset,Uint)
-#define x_reg_offset(R)	_ET_APPLY(x_reg_offset,(R))
+#define loader_tag(T) ((T) & _LOADER_MASK)
 
-#define _unchecked_y_reg_offset(R)	((R) - Y_REG_DEF)
-_ET_DECLARE_CHECKED(Uint,y_reg_offset,Uint)
-#define y_reg_offset(R)	_ET_APPLY(y_reg_offset,(R))
+#define _is_loader_x_reg(x) (loader_tag(x) == _LOADER_TAG_XREG)
+#define _is_loader_y_reg(x) (loader_tag(x) == _LOADER_TAG_YREG)
 
-#define reg_index(R) ((R) / sizeof(Eterm))
+#define _unchecked_loader_x_reg_index(R) ((R) >> _LOADER_TAG_SIZE)
+_ET_DECLARE_CHECKED(Uint,loader_x_reg_index,Uint)
+#define loader_x_reg_index(R) _ET_APPLY(loader_x_reg_index,(R))
 
-#define _unchecked_x_reg_index(R)	((R) >> 2)
-_ET_DECLARE_CHECKED(Uint,x_reg_index,Uint)
-#define x_reg_index(R)	_ET_APPLY(x_reg_index,(R))
-
-#define _unchecked_y_reg_index(R)	((R) >> 2)
-_ET_DECLARE_CHECKED(Uint,y_reg_index,Uint)
-#define y_reg_index(R)	_ET_APPLY(y_reg_index,(R))
+#define _unchecked_loader_y_reg_index(R)  ((R) >> _LOADER_TAG_SIZE)
+_ET_DECLARE_CHECKED(Uint,loader_y_reg_index,Uint)
+#define loader_y_reg_index(R) _ET_APPLY(loader_y_reg_index,(R))
 
 /*
  * Backwards compatibility definitions:
@@ -1136,8 +1124,9 @@ _ET_DECLARE_CHECKED(Uint,y_reg_index,Uint)
 #define FLOAT_DEF		0xe
 #define BIG_DEF			0xf
 #define SMALL_DEF		0x10
+#define MATCHSTATE_DEF          0x11   /* not a "real" term */
 
-#define FIRST_VACANT_TAG_DEF    0x11
+#define FIRST_VACANT_TAG_DEF    0x12
 
 #if ET_DEBUG
 extern unsigned tag_val_def_debug(Wterm, const char*, unsigned);
@@ -1159,82 +1148,9 @@ extern unsigned tag_val_def(Wterm);
 #define FLOAT_BIG 	_NUMBER_CODE(FLOAT_DEF,BIG_DEF)
 #define FLOAT_FLOAT	_NUMBER_CODE(FLOAT_DEF,FLOAT_DEF)
 
-#if HALFWORD_HEAP
-#define ptr2rel(PTR,BASE) ((Eterm*)((char*)(PTR) - (char*)(BASE)))
-#define rterm2wterm(REL,BASE) ((Wterm)(REL) + (Wterm)(BASE))
-
-#else /* HALFWORD_HEAP */
-
-#define ptr2rel(PTR,BASE) (PTR)
-#define rterm2wterm(REL,BASE) (REL)
-
-#endif /* !HALFWORD_HEAP */
-
-#define make_list_rel(PTR, BASE) make_list(ptr2rel(PTR,BASE))
-#define make_boxed_rel(PTR, BASE) make_boxed(ptr2rel(PTR,BASE))
-#define make_fun_rel make_boxed_rel
-#define make_binary_rel make_boxed_rel
-#define make_tuple_rel make_boxed_rel
-#define make_external_rel make_boxed_rel
-#define make_internal_ref_rel make_boxed_rel
-#define make_big_rel make_boxed_rel
-
-#define binary_val_rel(RTERM, BASE) binary_val(rterm2wterm(RTERM, BASE))
-#define list_val_rel(RTERM, BASE) list_val(rterm2wterm(RTERM, BASE))
-#define boxed_val_rel(RTERM, BASE) boxed_val(rterm2wterm(RTERM, BASE))
-#define tuple_val_rel(RTERM, BASE) tuple_val(rterm2wterm(RTERM, BASE))
-#define export_val_rel(RTERM, BASE) export_val(rterm2wterm(RTERM, BASE))
-#define fun_val_rel(RTERM, BASE) fun_val(rterm2wterm(RTERM, BASE))
-#define big_val_rel(RTERM,BASE)	big_val(rterm2wterm(RTERM,BASE))
-#define float_val_rel(RTERM,BASE) float_val(rterm2wterm(RTERM,BASE))
-#define internal_ref_val_rel(RTERM,BASE) internal_ref_val(rterm2wterm(RTERM,BASE))
-
-#define external_thing_ptr_rel(RTERM, BASE) external_thing_ptr(rterm2wterm(RTERM, BASE))
-#define external_data_words_rel(RTERM,BASE) external_data_words(rterm2wterm(RTERM,BASE))
-
-#define external_port_node_rel(RTERM,BASE) external_port_node(rterm2wterm(RTERM,BASE))
-#define external_port_data_rel(RTERM,BASE) external_port_data(rterm2wterm(RTERM,BASE))
-
-#define is_external_pid_rel(RTERM,BASE) is_external_pid(rterm2wterm(RTERM,BASE))
-#define external_pid_node_rel(RTERM,BASE) external_pid_node(rterm2wterm(RTERM,BASE))
-#define external_pid_data_rel(RTERM,BASE) external_pid_data(rterm2wterm(RTERM,BASE))
+#define is_same(A,B) ((A)==(B))
 
-#define is_binary_rel(RTERM,BASE) is_binary(rterm2wterm(RTERM,BASE))
-#define is_float_rel(RTERM,BASE) is_float(rterm2wterm(RTERM,BASE))
-#define is_fun_rel(RTERM,BASE) is_fun(rterm2wterm(RTERM,BASE))
-#define is_big_rel(RTERM,BASE) is_big(rterm2wterm(RTERM,BASE))
-#define is_export_rel(RTERM,BASE) is_export(rterm2wterm(RTERM,BASE))
-#define is_tuple_rel(RTERM,BASE) is_tuple(rterm2wterm(RTERM,BASE))
-
-#define GET_DOUBLE_REL(RTERM, f, BASE) GET_DOUBLE(rterm2wterm(RTERM,BASE), f)
-
-#define ref_thing_ptr_rel(RTERM,BASE) ref_thing_ptr(rterm2wterm(RTERM,BASE))
-#define is_internal_ref_rel(RTERM,BASE) is_internal_ref(rterm2wterm(RTERM,BASE))
-#define is_external_rel(RTERM,BASE) is_external(rterm2wterm(RTERM,BASE))
-#define is_external_port_rel(RTERM,BASE) is_external_port(rterm2wterm(RTERM,BASE))
-#define is_external_ref_rel(RTERM,BASE) is_external_ref(rterm2wterm(RTERM,BASE))
-
-#define external_node_rel(RTERM,BASE) external_node(rterm2wterm(RTERM,BASE))
-
-
-#if HALFWORD_HEAP
-ERTS_GLB_INLINE int is_same(Eterm a, Eterm* a_base, Eterm b, Eterm* b_base);
-
-#if ERTS_GLB_INLINE_INCL_FUNC_DEF
-ERTS_GLB_INLINE int is_same(Eterm a, Eterm* a_base, Eterm b, Eterm* b_base)
-{
-    /* If bases differ, assume a and b are on different "heaps",
-       ie can only be same if immed */
-    ASSERT(a_base == b_base || is_immed(a) || is_immed(b)
-	   || rterm2wterm(a,a_base) != rterm2wterm(b,b_base));
-
-    return a == b && (a_base == b_base || is_immed(a));
-}
-#endif
-
-#else /* !HALFWORD_HEAP */
-#define is_same(A,A_BASE,B,B_BASE) ((A)==(B))
-#endif
+void erts_set_literal_tag(Eterm *term, Eterm *hp_start, Eterm hsz);
 
 #endif	/* __ERL_TERM_H */
 
diff --git a/erts/emulator/beam/erl_threads.h b/erts/emulator/beam/erl_threads.h
index 5347979372..34f91e2ec8 100644
--- a/erts/emulator/beam/erl_threads.h
+++ b/erts/emulator/beam/erl_threads.h
@@ -649,6 +649,7 @@ ERTS_GLB_INLINE void erts_tsd_set(erts_tsd_key_t key, void *value);
 ERTS_GLB_INLINE void * erts_tsd_get(erts_tsd_key_t key);
 ERTS_GLB_INLINE erts_tse_t *erts_tse_fetch(void);
 ERTS_GLB_INLINE void erts_tse_return(erts_tse_t *ep);
+ERTS_GLB_INLINE void erts_tse_prepare_timed(erts_tse_t *ep);
 ERTS_GLB_INLINE void erts_tse_set(erts_tse_t *ep);
 ERTS_GLB_INLINE void erts_tse_reset(erts_tse_t *ep);
 ERTS_GLB_INLINE int erts_tse_wait(erts_tse_t *ep);
@@ -3461,6 +3462,15 @@ ERTS_GLB_INLINE void erts_tse_return(erts_tse_t *ep)
 #endif
 }
 
+ERTS_GLB_INLINE void erts_tse_prepare_timed(erts_tse_t *ep)
+{
+#ifdef USE_THREADS
+    int res = ethr_event_prepare_timed(&((ethr_ts_event *) ep)->event);
+    if (res != 0)
+	erts_thr_fatal_error(res, "prepare timed");
+#endif
+}
+
 ERTS_GLB_INLINE void erts_tse_set(erts_tse_t *ep)
 {
 #ifdef USE_THREADS
diff --git a/erts/emulator/beam/erl_time.h b/erts/emulator/beam/erl_time.h
index 36a3d52264..43e543e035 100644
--- a/erts/emulator/beam/erl_time.h
+++ b/erts/emulator/beam/erl_time.h
@@ -345,8 +345,10 @@ erts_time_unit_conversion(Uint64 value,
 #endif /* !ERTS_COMPILE_TIME_MONOTONIC_TIME_UNIT */
 
 #define ERTS_MONOTONIC_TIME_END_EXTERNAL				\
-    (ERTS_MONOTONIC_TIME_START_EXTERNAL					\
-     + (ERTS_MONOTONIC_END - ERTS_MONOTONIC_BEGIN))
+    (ERTS_MONOTONIC_TIME_START_EXTERNAL < 0				\
+     ? (ERTS_MONOTONIC_TIME_START_EXTERNAL				\
+	+ (ERTS_MONOTONIC_END - ERTS_MONOTONIC_BEGIN))			\
+     : (ERTS_MONOTONIC_END - ERTS_MONOTONIC_TIME_START_EXTERNAL))
 
 #define ERTS_MSEC_TO_CLKTCKS__(MON) \
     ((MON) * (ERTS_CLKTCK_RESOLUTION/1000))
diff --git a/erts/emulator/beam/erl_time_sup.c b/erts/emulator/beam/erl_time_sup.c
index 7f8f560681..7ec64506e8 100644
--- a/erts/emulator/beam/erl_time_sup.c
+++ b/erts/emulator/beam/erl_time_sup.c
@@ -124,7 +124,11 @@ typedef struct {
 
 typedef struct {
     ErtsMonotonicCorrectionInstance prev;
-    ErtsMonotonicCorrectionInstance curr;    
+    ErtsMonotonicCorrectionInstance curr;
+} ErtsMonotonicCorrectionInstances;
+
+typedef struct {
+    ErtsMonotonicCorrectionInstances insts;
     ErtsMonotonicDriftData drift;
     ErtsMonotonicTime last_check;
     int short_check_interval;
@@ -272,27 +276,24 @@ static ERTS_INLINE ErtsMonotonicTime
 read_corrected_time(int os_drift_corrected)
 {
     ErtsMonotonicTime os_mtime;
-    ErtsMonotonicCorrectionData cdata;
-    ErtsMonotonicCorrectionInstance *cip;
+    ErtsMonotonicCorrectionInstance ci;
 
     erts_smp_rwmtx_rlock(&time_sup.inf.c.parmon.rwmtx);
 
     os_mtime = erts_os_monotonic_time();
 
-    cdata = time_sup.inf.c.parmon.cdata;
-
-    erts_smp_rwmtx_runlock(&time_sup.inf.c.parmon.rwmtx);
-
-    if (os_mtime >= cdata.curr.os_mtime)
-	cip = &cdata.curr;
+    if (os_mtime >= time_sup.inf.c.parmon.cdata.insts.curr.os_mtime)
+	ci = time_sup.inf.c.parmon.cdata.insts.curr;
     else {
-	if (os_mtime < cdata.prev.os_mtime)
+	if (os_mtime < time_sup.inf.c.parmon.cdata.insts.prev.os_mtime)
 	    erl_exit(ERTS_ABORT_EXIT,
 		     "OS monotonic time stepped backwards\n");
-	cip = &cdata.prev;
+	ci = time_sup.inf.c.parmon.cdata.insts.prev;
     }
 
-    return calc_corrected_erl_mtime(os_mtime, cip, NULL,
+    erts_smp_rwmtx_runlock(&time_sup.inf.c.parmon.rwmtx);
+
+    return calc_corrected_erl_mtime(os_mtime, &ci, NULL,
 				    os_drift_corrected);
 }
 
@@ -360,9 +361,8 @@ check_time_correction(void *vesdp)
 {
     int init_drift_adj = !vesdp;
     ErtsSchedulerData *esdp = (ErtsSchedulerData *) vesdp;
-    ErtsMonotonicCorrectionData cdata;
     ErtsMonotonicCorrection new_correction;
-    ErtsMonotonicCorrectionInstance *cip;
+    ErtsMonotonicCorrectionInstance ci;
     ErtsMonotonicTime mdiff, sdiff, os_mtime, erl_mtime, os_stime,
 	erl_stime, time_offset, timeout_pos;
     Uint timeout;
@@ -373,16 +373,15 @@ check_time_correction(void *vesdp)
 
     erts_os_times(&os_mtime, &os_stime);
 
-    cdata = time_sup.inf.c.parmon.cdata;
+    ci = time_sup.inf.c.parmon.cdata.insts.curr;
 
     erts_smp_rwmtx_runlock(&time_sup.inf.c.parmon.rwmtx);
 
-    if (os_mtime < cdata.curr.os_mtime)
+    if (os_mtime < ci.os_mtime)
 	erl_exit(ERTS_ABORT_EXIT,
 		 "OS monotonic time stepped backwards\n");
-    cip = &cdata.curr;
 
-    erl_mtime = calc_corrected_erl_mtime(os_mtime, cip, &mdiff,
+    erl_mtime = calc_corrected_erl_mtime(os_mtime, &ci, &mdiff,
 					 os_drift_corrected);
     time_offset = get_time_offset();
     erl_stime = erl_mtime + time_offset;
@@ -397,7 +396,7 @@ check_time_correction(void *vesdp)
 	    time_sup.inf.c.shadow_offset = 0;
     }
 
-    new_correction = cip->correction;
+    new_correction = ci.correction;
 
     if (time_sup.r.o.warp_mode == ERTS_MULTI_TIME_WARP_MODE
 	&& (sdiff < -2*time_sup.r.o.adj.small_diff
@@ -408,7 +407,7 @@ check_time_correction(void *vesdp)
 	set_time_offset(time_offset);
 	schedule_send_time_offset_changed_notifications(time_offset);
 	begin_short_intervals = 1;
-	if (cdata.curr.correction.error != 0) {
+	if (ci.correction.error != 0) {
 	    set_new_correction = 1;
 	    new_correction.error = 0;
 	}
@@ -425,12 +424,12 @@ check_time_correction(void *vesdp)
 	time_sup.inf.c.shadow_offset -= sdiff;
 	sdiff = 0;
 	begin_short_intervals = 1;
-	if (cdata.curr.correction.error != 0) {
+	if (ci.correction.error != 0) {
 	    set_new_correction = 1;
 	    new_correction.error = 0;
 	}
     }
-    else if (cdata.curr.correction.error == 0) {
+    else if (ci.correction.error == 0) {
 	if (sdiff < -time_sup.r.o.adj.small_diff) {
 	    set_new_correction = 1;
 	    if (sdiff < -time_sup.r.o.adj.large_diff)
@@ -446,9 +445,9 @@ check_time_correction(void *vesdp)
 		new_correction.error = -ERTS_TCORR_ERR_SMALL_ADJ;
 	}
     }
-    else if (cdata.curr.correction.error > 0) {
+    else if (ci.correction.error > 0) {
 	if (sdiff < 0) {
-	    if (cdata.curr.correction.error != ERTS_TCORR_ERR_LARGE_ADJ
+	    if (ci.correction.error != ERTS_TCORR_ERR_LARGE_ADJ
 		&& sdiff < -time_sup.r.o.adj.large_diff) {
 		new_correction.error = ERTS_TCORR_ERR_LARGE_ADJ;
 		set_new_correction = 1;
@@ -466,9 +465,9 @@ check_time_correction(void *vesdp)
 	    new_correction.error = 0;
 	}
     }
-    else /* if (cdata.curr.correction.error < 0) */ { 
+    else /* if (ci.correction.error < 0) */ { 
 	if (0 < sdiff) {
-	    if (cdata.curr.correction.error != -ERTS_TCORR_ERR_LARGE_ADJ
+	    if (ci.correction.error != -ERTS_TCORR_ERR_LARGE_ADJ
 		&& time_sup.r.o.adj.large_diff < sdiff) {
 		new_correction.error = -ERTS_TCORR_ERR_LARGE_ADJ;
 		set_new_correction = 1;
@@ -631,8 +630,8 @@ check_time_correction(void *vesdp)
 #ifdef ERTS_TIME_CORRECTION_PRINT
     print_correction(set_new_correction,
 		     sdiff,
-		     cip->correction.error,
-		     cip->correction.drift,
+		     ci.correction.error,
+		     ci.correction.drift,
 		     new_correction.error,
 		     new_correction.drift,
 		     timeout);
@@ -644,7 +643,7 @@ check_time_correction(void *vesdp)
 	os_mtime = erts_os_monotonic_time();
 
 	/* Save previous correction instance */
-	time_sup.inf.c.parmon.cdata.prev = *cip;
+	time_sup.inf.c.parmon.cdata.insts.prev = ci;
 
 	/*
 	 * Current correction instance begin when
@@ -657,15 +656,15 @@ check_time_correction(void *vesdp)
 	 * next OS monotonic time using previous
 	 * correction.
 	 */
-	erl_mtime = calc_corrected_erl_mtime(os_mtime, cip, NULL,
+	erl_mtime = calc_corrected_erl_mtime(os_mtime, &ci, NULL,
 					     os_drift_corrected);
 
 	/*
 	 * Save new current correction instance.
 	 */
-	time_sup.inf.c.parmon.cdata.curr.erl_mtime = erl_mtime;
-	time_sup.inf.c.parmon.cdata.curr.os_mtime = os_mtime;
-	time_sup.inf.c.parmon.cdata.curr.correction = new_correction;
+	time_sup.inf.c.parmon.cdata.insts.curr.erl_mtime = erl_mtime;
+	time_sup.inf.c.parmon.cdata.insts.curr.os_mtime = os_mtime;
+	time_sup.inf.c.parmon.cdata.insts.curr.correction = new_correction;
 
 	erts_smp_rwmtx_rwunlock(&time_sup.inf.c.parmon.rwmtx);
     }
@@ -784,24 +783,22 @@ static ErtsMonotonicTime
 finalize_corrected_time_offset(ErtsSystemTime *stimep)
 {
     ErtsMonotonicTime os_mtime;
-    ErtsMonotonicCorrectionData cdata;
-    ErtsMonotonicCorrectionInstance *cip;
+    ErtsMonotonicCorrectionInstance ci;
     int os_drift_corrected = time_sup.r.o.os_corrected_monotonic_time;
 
     erts_smp_rwmtx_rlock(&time_sup.inf.c.parmon.rwmtx);
 
     erts_os_times(&os_mtime, stimep);
 
-    cdata = time_sup.inf.c.parmon.cdata;
+    ci = time_sup.inf.c.parmon.cdata.insts.curr;
 
     erts_smp_rwmtx_runlock(&time_sup.inf.c.parmon.rwmtx);
 
-    if (os_mtime < cdata.curr.os_mtime)
+    if (os_mtime < ci.os_mtime)
 	erl_exit(ERTS_ABORT_EXIT,
 		 "OS monotonic time stepped backwards\n");
-    cip = &cdata.curr;
 
-    return calc_corrected_erl_mtime(os_mtime, cip, NULL,
+    return calc_corrected_erl_mtime(os_mtime, &ci, NULL,
 				    os_drift_corrected);
 }
 
@@ -1128,13 +1125,13 @@ erts_init_time_sup(int time_correction, ErtsTimeWarpMode time_warp_mode)
     
 	cdatap->drift.intervals[0].time.sys = time_sup.inf.c.sinit;
 	cdatap->drift.intervals[0].time.mon = time_sup.inf.c.minit;
-	cdatap->curr.correction.drift = 0;
-	cdatap->curr.correction.error = 0;
-	cdatap->curr.erl_mtime = ERTS_MONOTONIC_BEGIN;
-	cdatap->curr.os_mtime = time_sup.inf.c.minit;
+	cdatap->insts.curr.correction.drift = 0;
+	cdatap->insts.curr.correction.error = 0;
+	cdatap->insts.curr.erl_mtime = ERTS_MONOTONIC_BEGIN;
+	cdatap->insts.curr.os_mtime = time_sup.inf.c.minit;
 	cdatap->last_check = time_sup.inf.c.minit;
 	cdatap->short_check_interval = ERTS_INIT_SHORT_INTERVAL_COUNTER;
-	cdatap->prev = cdatap->curr;
+	cdatap->insts.prev = cdatap->insts.curr;
 
 	if (!time_sup.r.o.os_corrected_monotonic_time)
 	    time_sup.r.o.get_time = get_corrected_time;
@@ -1922,15 +1919,16 @@ send_time_offset_changed_notifications(void *new_offsetp)
 		ErtsProcLocks rp_locks = ERTS_PROC_LOCK_LINK;
 		erts_smp_proc_lock(rp, ERTS_PROC_LOCK_LINK);
 		if (erts_lookup_monitor(ERTS_P_MONITORS(rp), ref)) {
-		    ErlHeapFragment *bp;
+		    ErtsMessage *mp;
 		    ErlOffHeap *ohp;
 		    Eterm message;
 
-		    hp = erts_alloc_message_heap(hsz, &bp, &ohp, rp, &rp_locks);
+		    mp = erts_alloc_message_heap(rp, &rp_locks,
+						 hsz, &hp, &ohp);
 		    *patch_refp = ref;
 		    ASSERT(hsz == size_object(message_template));
 		    message = copy_struct(message_template, hsz, &hp, ohp);
-		    erts_queue_message(rp, &rp_locks, bp, message, NIL);
+		    erts_queue_message(rp, &rp_locks, mp, message, NIL);
 		}
 		erts_smp_proc_unlock(rp, rp_locks);
 	    }
diff --git a/erts/emulator/beam/erl_trace.c b/erts/emulator/beam/erl_trace.c
index e1b03a057f..b774ba97af 100644
--- a/erts/emulator/beam/erl_trace.c
+++ b/erts/emulator/beam/erl_trace.c
@@ -114,15 +114,10 @@ void erts_init_trace(void) {
 
 static Eterm system_seq_tracer;
 
-#ifdef ERTS_SMP
 #define ERTS_ALLOC_SYSMSG_HEAP(SZ, BPP, OHPP, UNUSED) \
   (*(BPP) = new_message_buffer((SZ)), \
    *(OHPP) = &(*(BPP))->off_heap, \
    (*(BPP))->mem)
-#else
-#define ERTS_ALLOC_SYSMSG_HEAP(SZ, BPP, OHPP, RPP) \
-  erts_alloc_message_heap((SZ), (BPP), (OHPP), (RPP), 0)
-#endif
 
 #ifdef ERTS_SMP
 #define ERTS_ENQ_TRACE_MSG(FPID, TPID, MSG, BP) \
@@ -131,8 +126,12 @@ do { \
     enqueue_sys_msg_unlocked(SYS_MSG_TYPE_TRACE, (FPID), (TPID), (MSG), (BP)); \
 } while(0)
 #else
-#define ERTS_ENQ_TRACE_MSG(FPID, TPROC, MSG, BP) \
-    erts_queue_message((TPROC), NULL, (BP), (MSG), NIL)
+#define ERTS_ENQ_TRACE_MSG(FPID, TPROC, MSG, BP)		\
+    do {							\
+	ErtsMessage *mp__ = erts_alloc_message(0, NULL);		\
+	mp__->data.heap_frag = (BP);				\
+	erts_queue_message((TPROC), NULL, mp__, (MSG), NIL);	\
+    } while (0)
 #endif
 
 /*
@@ -591,11 +590,9 @@ send_to_port(Process *c_p, Eterm message,
 static void 
 profile_send(Eterm from, Eterm message) {
     Uint sz = 0;
-    ErlHeapFragment *bp = NULL;
     Uint *hp = NULL;
     Eterm msg = NIL;
     Process *profile_p = NULL;
-    ErlOffHeap *off_heap = NULL;
 
     Eterm profiler = erts_get_system_profile();
 
@@ -621,6 +618,7 @@ profile_send(Eterm from, Eterm message) {
 	}
     	
     } else {
+	ErtsMessage *mp;
 	ASSERT(is_internal_pid(profiler));
         
 	profile_p = erts_proc_lookup(profiler);
@@ -629,10 +627,13 @@ profile_send(Eterm from, Eterm message) {
 	    return;
 
 	sz = size_object(message);
-	hp = erts_alloc_message_heap(sz, &bp, &off_heap, profile_p, 0);
-	msg = copy_struct(message, sz, &hp, &bp->off_heap);
-	
-        erts_queue_message(profile_p, NULL, bp, msg, NIL);
+	mp = erts_alloc_message(sz, &hp);
+	if (sz == 0)
+	    msg = message;
+	else
+	    msg = copy_struct(message, sz, &hp, &mp->hfrag.off_heap);
+
+        erts_queue_message(profile_p, NULL, mp, msg, NIL);
     }
 }
 
@@ -1045,11 +1046,7 @@ seq_trace_update_send(Process *p)
 {
     Eterm seq_tracer = erts_get_system_seq_tracer();
     ASSERT((is_tuple(SEQ_TRACE_TOKEN(p)) || is_nil(SEQ_TRACE_TOKEN(p))));
-    if ( (p->common.id == seq_tracer) || (SEQ_TRACE_TOKEN(p) == NIL)
-#ifdef USE_VM_PROBES
-	 || (SEQ_TRACE_TOKEN(p) == am_have_dt_utag)
-#endif
-	 ) {
+    if ((p->common.id == seq_tracer) || have_no_seqtrace(SEQ_TRACE_TOKEN(p))) {
 	return 0;
     }
     SEQ_TRACE_TOKEN_SENDER(p) = p->common.id;
@@ -1233,7 +1230,11 @@ seq_trace_output_generic(Eterm token, Eterm msg, Uint type,
 	erts_smp_mtx_unlock(&smq_mtx);
 #else
         /* trace_token must be NIL here */
-	erts_queue_message(tracer, NULL, bp, mess, NIL);
+	{
+	    ErtsMessage *mp = erts_alloc_message(0, NULL);
+	    mp->data.heap_frag = bp;
+	    erts_queue_message(tracer, NULL, mp, mess, NIL);
+	}
 #endif
     }
 }
@@ -1677,12 +1678,7 @@ erts_call_trace(Process* p, BeamInstr mfa[3], Binary *match_spec,
     args = transformed_args;
 
     if (is_internal_port(*tracer_pid)) {
-#if HEAP_ON_C_STACK
 	Eterm local_heap[64+MAX_ARG];
-#else
-	Eterm *local_heap = erts_alloc(ERTS_ALC_T_TEMP_TERM,
-				       sizeof(Eterm)*(64+MAX_ARG));
-#endif
 	hp = local_heap;
 
 	if (!erts_is_valid_tracer_port(*tracer_pid)) {
@@ -1697,9 +1693,6 @@ erts_call_trace(Process* p, BeamInstr mfa[3], Binary *match_spec,
 	    if (is_not_nil(tracee))
 		erts_smp_proc_unlock(p, ERTS_PROC_LOCKS_ALL_MINOR);
 #endif
-#if !HEAP_ON_C_STACK
-	    erts_free(ERTS_ALC_T_TEMP_TERM,local_heap);
-#endif
 	    UnUseTmpHeap(ERL_SUB_BIN_SIZE,p);
 	    return 0;
 	}
@@ -1727,9 +1720,6 @@ erts_call_trace(Process* p, BeamInstr mfa[3], Binary *match_spec,
 					    ERTS_PAM_TMP_RESULT, &return_flags);
 	    if (is_non_value(pam_result)) {
 		erts_match_set_release_result(p);
-#if !HEAP_ON_C_STACK
-		erts_free(ERTS_ALC_T_TEMP_TERM,local_heap);
-#endif
 		UnUseTmpHeap(ERL_SUB_BIN_SIZE,p);
 		return 0;
 	    }
@@ -1738,9 +1728,6 @@ erts_call_trace(Process* p, BeamInstr mfa[3], Binary *match_spec,
 	    /* Meta trace */
 	    if (pam_result == am_false) {
 		erts_match_set_release_result(p);
-#if !HEAP_ON_C_STACK
-		erts_free(ERTS_ALC_T_TEMP_TERM,local_heap);
-#endif
 		UnUseTmpHeap(ERL_SUB_BIN_SIZE,p);
 		return return_flags;
 	    }
@@ -1748,17 +1735,11 @@ erts_call_trace(Process* p, BeamInstr mfa[3], Binary *match_spec,
 	    /* Non-meta trace */
 	    if (*tracee_flags & F_TRACE_SILENT) { 
 		erts_match_set_release_result(p);
-#if !HEAP_ON_C_STACK
-		erts_free(ERTS_ALC_T_TEMP_TERM,local_heap);
-#endif
 		UnUseTmpHeap(ERL_SUB_BIN_SIZE,p);
 		return 0;
 	    }
 	    if (pam_result == am_false) {
 		erts_match_set_release_result(p);
-#if !HEAP_ON_C_STACK
-		erts_free(ERTS_ALC_T_TEMP_TERM,local_heap);
-#endif
 		UnUseTmpHeap(ERL_SUB_BIN_SIZE,p);
 		return return_flags;
 	    }
@@ -1802,9 +1783,6 @@ erts_call_trace(Process* p, BeamInstr mfa[3], Binary *match_spec,
 	send_to_port(p, mess, tracer_pid, tracee_flags);
 	erts_smp_mtx_unlock(&smq_mtx);
 	erts_match_set_release_result(p);
-#if !HEAP_ON_C_STACK
-	erts_free(ERTS_ALC_T_TEMP_TERM,local_heap);
-#endif
 	UnUseTmpHeap(ERL_SUB_BIN_SIZE,p);
 	return *tracer_pid == NIL ? 0 : return_flags;
 
@@ -1823,7 +1801,6 @@ erts_call_trace(Process* p, BeamInstr mfa[3], Binary *match_spec,
 #ifdef DEBUG
 	Eterm* limit;
 #endif
-
 	ASSERT(is_internal_pid(*tracer_pid));
 	
 	tracer = erts_pid2proc(p, ERTS_PROC_LOCK_MAIN,
@@ -2332,7 +2309,11 @@ monitor_long_schedule_proc(Process *p, BeamInstr *in_fp, BeamInstr *out_fp, Uint
 #ifdef ERTS_SMP
     enqueue_sys_msg(SYS_MSG_TYPE_SYSMON, p->common.id, NIL, msg, bp);
 #else
-    erts_queue_message(monitor_p, NULL, bp, msg, NIL);
+    {
+	ErtsMessage *mp = erts_alloc_message(0, NULL);
+	mp->data.heap_frag = bp;
+	erts_queue_message(monitor_p, NULL, mp, msg, NIL);
+    }
 #endif
 }
 void 
@@ -2393,7 +2374,11 @@ monitor_long_schedule_port(Port *pp, ErtsPortTaskType type, Uint time)
 #ifdef ERTS_SMP
     enqueue_sys_msg(SYS_MSG_TYPE_SYSMON, pp->common.id, NIL, msg, bp);
 #else
-    erts_queue_message(monitor_p, NULL, bp, msg, NIL);
+    {
+	ErtsMessage *mp = erts_alloc_message(0, NULL);
+	mp->data.heap_frag = bp;
+	erts_queue_message(monitor_p, NULL, mp, msg, NIL);
+    }
 #endif
 }
 
@@ -2464,7 +2449,11 @@ monitor_long_gc(Process *p, Uint time) {
 #ifdef ERTS_SMP
     enqueue_sys_msg(SYS_MSG_TYPE_SYSMON, p->common.id, NIL, msg, bp);
 #else
-    erts_queue_message(monitor_p, NULL, bp, msg, NIL);
+    {
+	ErtsMessage *mp = erts_alloc_message(0, NULL);
+	mp->data.heap_frag = bp;
+	erts_queue_message(monitor_p, NULL, mp, msg, NIL);
+    }
 #endif
 }
 
@@ -2535,7 +2524,11 @@ monitor_large_heap(Process *p) {
 #ifdef ERTS_SMP
     enqueue_sys_msg(SYS_MSG_TYPE_SYSMON, p->common.id, NIL, msg, bp);
 #else
-    erts_queue_message(monitor_p, NULL, bp, msg, NIL);
+    {
+	ErtsMessage *mp = erts_alloc_message(0, NULL);
+	mp->data.heap_frag = bp;
+	erts_queue_message(monitor_p, NULL, mp, msg, NIL);
+    }
 #endif
 }
 
@@ -2563,7 +2556,11 @@ monitor_generic(Process *p, Eterm type, Eterm spec) {
 #ifdef ERTS_SMP
     enqueue_sys_msg(SYS_MSG_TYPE_SYSMON, p->common.id, NIL, msg, bp);
 #else
-    erts_queue_message(monitor_p, NULL, bp, msg, NIL);
+    {
+	ErtsMessage *mp = erts_alloc_message(0, NULL);
+	mp->data.heap_frag = bp;
+	erts_queue_message(monitor_p, NULL, mp, msg, NIL);
+    }
 #endif
 
 }
@@ -3297,8 +3294,6 @@ sys_msg_dispatcher_func(void *unused)
 	    if (erts_thr_progress_update(NULL))
 		erts_thr_progress_leader_update(NULL);
 
-	    ERTS_SCHED_FAIR_YIELD();
-
 #ifdef DEBUG_PRINTOUTS
 	    print_msg_type(smqp);
 #endif
@@ -3357,8 +3352,11 @@ sys_msg_dispatcher_func(void *unused)
 		    goto failure;
 		}
 		else {
+		    ErtsMessage *mp;
 		queue_proc_msg:
-		    erts_queue_message(proc,&proc_locks,smqp->bp,smqp->msg,NIL);
+		    mp = erts_alloc_message(0, NULL);
+		    mp->data.heap_frag = smqp->bp;
+		    erts_queue_message(proc,&proc_locks,mp,smqp->msg,NIL);
 #ifdef DEBUG_PRINTOUTS
 		    erts_fprintf(stderr, "delivered\n");
 #endif
@@ -3453,9 +3451,6 @@ static void
 init_sys_msg_dispatcher(void)
 {
     erts_smp_thr_opts_t thr_opts = ERTS_SMP_THR_OPTS_DEFAULT_INITER;
-#ifdef __OSE__
-    thr_opts.coreNo   = 0;
-#endif
     thr_opts.detached = 1;
     thr_opts.name = "sys_msg_dispatcher";
     init_smq_element_alloc();
@@ -3463,7 +3458,6 @@ init_sys_msg_dispatcher(void)
     sys_message_queue_end = NULL;
     erts_smp_cnd_init(&smq_cnd);
     erts_smp_mtx_init(&smq_mtx, "sys_msg_q");
-
     erts_smp_thr_create(&sys_msg_dispatcher_tid,
 			sys_msg_dispatcher_func,
 			NULL,
diff --git a/erts/emulator/beam/erl_utils.h b/erts/emulator/beam/erl_utils.h
index 6dab3bf297..4058d63eaf 100644
--- a/erts/emulator/beam/erl_utils.h
+++ b/erts/emulator/beam/erl_utils.h
@@ -157,32 +157,15 @@ void erts_init_utils_mem(void);
 erts_dsprintf_buf_t *erts_create_tmp_dsbuf(Uint);
 void erts_destroy_tmp_dsbuf(erts_dsprintf_buf_t *);
 
-#if HALFWORD_HEAP
-int eq_rel(Eterm a, Eterm* a_base, Eterm b, Eterm* b_base);
-#  define eq(A,B) eq_rel(A,NULL,B,NULL)
-#else
 int eq(Eterm, Eterm);
-#  define eq_rel(A,A_BASE,B,B_BASE) eq(A,B)
-#endif
 
 #define EQ(x,y) (((x) == (y)) || (is_not_both_immed((x),(y)) && eq((x),(y))))
 
-#if HALFWORD_HEAP
-Sint erts_cmp_rel_opt(Eterm, Eterm*, Eterm, Eterm*, int, int);
-#define cmp_rel(A,A_BASE,B,B_BASE)       erts_cmp_rel_opt(A,A_BASE,B,B_BASE,0,0)
-#define cmp_rel_term(A,A_BASE,B,B_BASE)  erts_cmp_rel_opt(A,A_BASE,B,B_BASE,1,0)
-#define CMP(A,B)                         erts_cmp_rel_opt(A,NULL,B,NULL,0,0)
-#define CMP_TERM(A,B)                    erts_cmp_rel_opt(A,NULL,B,NULL,1,0)
-#define CMP_EQ_ONLY(A,B)                 erts_cmp_rel_opt(A,NULL,B,NULL,0,1)
-#else
 Sint erts_cmp(Eterm, Eterm, int, int);
 Sint cmp(Eterm a, Eterm b);
-#define cmp_rel(A,A_BASE,B,B_BASE)       erts_cmp(A,B,0,0)
-#define cmp_rel_term(A,A_BASE,B,B_BASE)  erts_cmp(A,B,1,0)
 #define CMP(A,B)                         erts_cmp(A,B,0,0)
 #define CMP_TERM(A,B)                    erts_cmp(A,B,1,0)
 #define CMP_EQ_ONLY(A,B)                 erts_cmp(A,B,0,1)
-#endif
 
 #define cmp_lt(a,b)          (CMP((a),(b)) <  0)
 #define cmp_le(a,b)          (CMP((a),(b)) <= 0)
diff --git a/erts/emulator/beam/erl_vm.h b/erts/emulator/beam/erl_vm.h
index 8948ca2ea9..98f27a1725 100644
--- a/erts/emulator/beam/erl_vm.h
+++ b/erts/emulator/beam/erl_vm.h
@@ -40,14 +40,6 @@
 #define MAX_ARG 255	        /* Max number of arguments allowed */
 #define MAX_REG 1024            /* Max number of x(N) registers used */
 
-/* Scheduler stores data for temporary heaps if
-   !HEAP_ON_C_STACK. Macros (*TmpHeap*) in global.h selects if we put temporary
-   heap data on the C stack or if we use the buffers in the scheduler data. */
-#define TMP_HEAP_SIZE 128            /* Number of Eterm in the schedulers
-				        small heap for transient heap data */
-#define ERL_ARITH_TMP_HEAP_SIZE 4    /* as does erl_arith... */
-#define BEAM_EMU_TMP_HEAP_SIZE  2    /* and beam_emu... */
-
 /*
  * The new arithmetic operations need some extra X registers in the register array.
  * so does the gc_bif's (i_gc_bif3 need 3 extra).
@@ -117,12 +109,8 @@
 #define HeapWordsLeft(p) (HEAP_LIMIT(p) - HEAP_TOP(p))
 
 #if defined(DEBUG) || defined(CHECK_FOR_HOLES)
-#if HALFWORD_HEAP
-# define ERTS_HOLE_MARKER (0xdeadbeef)
-#else
 # define ERTS_HOLE_MARKER (((0xdeadbeef << 24) << 8) | 0xdeadbeef)
-#endif
-#endif
+#endif /* egil: 32-bit ? */
 
 /*
  * Allocate heap memory on the ordinary heap, NEVER in a heap
@@ -153,6 +141,7 @@
 typedef struct op_entry {
    char* name;			/* Name of instruction. */
    Uint32 mask[3];		/* Signature mask. */
+   unsigned involves_r;		/* Needs special attention when matching. */
    int sz;			/* Number of loaded words. */
    char* pack;			/* Instructions for packing engine. */
    char* sign;			/* Signature string. */
diff --git a/erts/emulator/beam/export.c b/erts/emulator/beam/export.c
index 2420df36b5..581efe6eec 100644
--- a/erts/emulator/beam/export.c
+++ b/erts/emulator/beam/export.c
@@ -184,6 +184,9 @@ init_export_table(void)
     f.cmp  = (HCMP_FUN) export_cmp;
     f.alloc = (HALLOC_FUN) export_alloc;
     f.free = (HFREE_FUN) export_free;
+    f.meta_alloc = (HMALLOC_FUN) erts_alloc;
+    f.meta_free = (HMFREE_FUN) erts_free;
+    f.meta_print = (HMPRINT_FUN) erts_print;
 
     for (i=0; i<ERTS_NUM_CODE_IX; i++) {
 	erts_index_init(ERTS_ALC_T_EXPORT_TABLE, &export_tables[i], "export_list",
diff --git a/erts/emulator/beam/external.c b/erts/emulator/beam/external.c
index c6d7e3fcc5..58010dcb72 100644
--- a/erts/emulator/beam/external.c
+++ b/erts/emulator/beam/external.c
@@ -1179,7 +1179,7 @@ typedef struct {
     ErtsHeapFactory factory;
     int remaining_n;
     char* remaining_bytes;
-    Eterm* maps_list;
+    ErtsWStack flat_maps;
     ErtsPStack hamt_array;
 } B2TDecodeContext;
 
@@ -1519,7 +1519,7 @@ static BIF_RETTYPE binary_to_term_int(Process* p, Uint32 flags, Eterm bin, Binar
             ctx->u.dc.res = (Eterm) (UWord) NULL;
             ctx->u.dc.next = &ctx->u.dc.res;
 	    erts_factory_proc_prealloc_init(&ctx->u.dc.factory, p, ctx->heap_size);
-	    ctx->u.dc.maps_list = NULL;
+	    ctx->u.dc.flat_maps.wstart = NULL;
 	    ctx->u.dc.hamt_array.pstart = NULL;
             ctx->state = B2TDecode;
             /*fall through*/
@@ -2339,10 +2339,6 @@ enc_term_int(TTBEncodeContext* ctx, ErtsAtomCacheMap *acmp, Eterm obj, byte* ep,
     Eterm val;
     FloatDef f;
     Sint r = 0;
-#if HALFWORD_HEAP
-    UWord wobj;
-#endif
-
 
     if (ctx) {
 	WSTACK_CHANGE_ALLOCATOR(s, ERTS_ALC_T_SAVED_ESTACK);
@@ -2362,11 +2358,8 @@ enc_term_int(TTBEncodeContext* ctx, ErtsAtomCacheMap *acmp, Eterm obj, byte* ep,
 
  outer_loop:
     while (!WSTACK_ISEMPTY(s)) {
-#if HALFWORD_HEAP
-	obj = (Eterm) (wobj = WSTACK_POP(s));
-#else
 	obj = WSTACK_POP(s);
-#endif
+
 	switch (val = WSTACK_POP(s)) {
 	case ENC_TERM:
 	    break;
@@ -2384,11 +2377,7 @@ enc_term_int(TTBEncodeContext* ctx, ErtsAtomCacheMap *acmp, Eterm obj, byte* ep,
 	    break;
 	case ENC_PATCH_FUN_SIZE:
 	    {
-#if HALFWORD_HEAP
-		byte* size_p = (byte *) wobj;
-#else
 		byte* size_p = (byte *) obj;
-#endif
 		put_int32(ep - size_p, size_p);
 	    }
 	    goto outer_loop;
@@ -2435,21 +2424,13 @@ enc_term_int(TTBEncodeContext* ctx, ErtsAtomCacheMap *acmp, Eterm obj, byte* ep,
 	case ENC_LAST_ARRAY_ELEMENT:
 	    /* obj is the tuple */
 	    {
-#if HALFWORD_HEAP
-		Eterm* ptr = (Eterm *) wobj;
-#else
 		Eterm* ptr = (Eterm *) obj;
-#endif
 		obj = *ptr;
 	    }
 	    break;
 	default:		/* ENC_LAST_ARRAY_ELEMENT+1 and upwards */
 	    {
-#if HALFWORD_HEAP
-		Eterm* ptr = (Eterm *) wobj;
-#else
 		Eterm* ptr = (Eterm *) obj;
-#endif
 		obj = *ptr++;
 		WSTACK_PUSH2(s, val-1, (UWord)ptr);
 	    }
@@ -2938,7 +2919,7 @@ dec_term(ErtsDistExternal *edep,
     int n;
     ErtsAtomEncoding char_enc;
     register Eterm* hp;        /* Please don't take the address of hp */
-    Eterm *maps_list;   /* for preprocessing of small maps */
+    DECLARE_WSTACK(flat_maps); /* for preprocessing of small maps */
     Eterm* next;
     SWord reds;
 #ifdef DEBUG
@@ -2950,7 +2931,6 @@ dec_term(ErtsDistExternal *edep,
         next     = ctx->u.dc.next;
         ep       = ctx->u.dc.ep;
 	factory  = &ctx->u.dc.factory;
-	maps_list = ctx->u.dc.maps_list;
 
         if (ctx->state != B2TDecode) {
             int n_limit = reds;
@@ -2977,7 +2957,7 @@ dec_term(ErtsDistExternal *edep,
             case B2TDecodeList:
                 objp = next - 2;
                 while (n > 0) {
-                    objp[0] = (Eterm) COMPRESS_POINTER(next);
+                    objp[0] = (Eterm) next;
                     objp[1] = make_list(next);
                     next = objp;
                     objp -= 2;
@@ -2988,7 +2968,7 @@ dec_term(ErtsDistExternal *edep,
             case B2TDecodeTuple:
                 objp = next - 1;
                 while (n-- > 0) {
-                    objp[0] = (Eterm) COMPRESS_POINTER(next);
+                    objp[0] = (Eterm) next;
                     next = objp;
                     objp--;
                 }
@@ -3026,22 +3006,25 @@ dec_term(ErtsDistExternal *edep,
             }
         }
 	PSTACK_CHANGE_ALLOCATOR(hamt_array, ERTS_ALC_T_SAVED_ESTACK);
+        WSTACK_CHANGE_ALLOCATOR(flat_maps, ERTS_ALC_T_SAVED_ESTACK);
 	if (ctx->u.dc.hamt_array.pstart) {
 	    PSTACK_RESTORE(hamt_array, &ctx->u.dc.hamt_array);
 	}
+	if (ctx->u.dc.flat_maps.wstart) {
+	    WSTACK_RESTORE(flat_maps, &ctx->u.dc.flat_maps);
+	}
     }
     else {
         reds = ERTS_SWORD_MAX;
         next = objp;
         *next = (Eterm) (UWord) NULL;
-	maps_list = NULL;
     }
     hp = factory->hp;
 
     while (next != NULL) {
 
 	objp = next;
-	next = (Eterm *) EXPAND_POINTER(*objp);
+	next = (Eterm *) *objp;
 
 	switch (*ep++) {
 	case INTEGER_EXT:
@@ -3049,7 +3032,7 @@ dec_term(ErtsDistExternal *edep,
 		Sint sn = get_int32(ep);
 
 		ep += 4;
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
 		*objp = make_small(sn);
 #else
 		if (MY_IS_SSMALL(sn)) {
@@ -3172,7 +3155,7 @@ dec_term_atom_common:
 		reds -= n;
 	    }
 	    while (n-- > 0) {
-		objp[0] = (Eterm) COMPRESS_POINTER(next);
+		objp[0] = (Eterm) next;
 		next = objp;
 		objp--;
 	    }
@@ -3190,8 +3173,8 @@ dec_term_atom_common:
 	    *objp = make_list(hp);
             hp += 2 * n;
 	    objp = hp - 2;
-	    objp[0] = (Eterm) COMPRESS_POINTER((objp+1));
-	    objp[1] = (Eterm) COMPRESS_POINTER(next);
+	    objp[0] = (Eterm) (objp+1);
+	    objp[1] = (Eterm) next;
 	    next = objp;
 	    objp -= 2;
             n--;
@@ -3204,7 +3187,7 @@ dec_term_atom_common:
 		reds -= n;
 	    }
             while (n > 0) {
-		objp[0] = (Eterm) COMPRESS_POINTER(next);
+		objp[0] = (Eterm) next;
 		objp[1] = make_list(next);
 		next = objp;
 		objp -= 2;
@@ -3371,7 +3354,7 @@ dec_term_atom_common:
 		    RefThing *rtp = (RefThing *) hp;
 		    ref_num = (Uint32 *) (hp + REF_THING_HEAD_SIZE);
 
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
 		    hp += REF_THING_HEAD_SIZE + ref_words/2 + 1;
 		    rtp->header = make_ref_thing_header(ref_words/2 + 1);
 #else
@@ -3382,13 +3365,13 @@ dec_term_atom_common:
 		}
 		else {
 		    ExternalThing *etp = (ExternalThing *) hp;
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
 		    hp += EXTERNAL_THING_HEAD_SIZE + ref_words/2 + 1;
 #else
 		    hp += EXTERNAL_THING_HEAD_SIZE + ref_words;
 #endif
 
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
 		    etp->header = make_external_ref_header(ref_words/2 + 1);
 #else
 		    etp->header = make_external_ref_header(ref_words);
@@ -3401,7 +3384,7 @@ dec_term_atom_common:
 		    ref_num = &(etp->data.ui32[0]);
 		}
 
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
 		*(ref_num++) = ref_words /* 32-bit arity */;
 #endif
 		ref_num[0] = r0;
@@ -3409,7 +3392,7 @@ dec_term_atom_common:
 		    ref_num[i] = get_int32(ep);
 		    ep += 4;
 		}
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
 		if ((1 + ref_words) % 2)
 		    ref_num[ref_words] = 0;
 #endif
@@ -3561,12 +3544,7 @@ dec_term_atom_common:
                 }
 		*objp = make_export(hp);
 		*hp++ = HEADER_EXPORT;
-#if HALFWORD_HEAP
-		*((UWord *) (UWord) hp) =  (UWord) erts_export_get_or_make_stub(mod, name, arity);
-		hp += 2;
-#else
 		*hp++ = (Eterm) erts_export_get_or_make_stub(mod, name, arity);
-#endif
 		break;
 	    }
 	    break;
@@ -3595,21 +3573,15 @@ dec_term_atom_common:
                      * vptr, last word for values
                      */
 
-                    /*
-                     * Use thing_word to link through decoded maps.
-                     * The list of maps is for later validation.
-                     */
-
-                    mp->thing_word = (Eterm) COMPRESS_POINTER(maps_list);
-                    maps_list      = (Eterm *) mp;
-
+                    WSTACK_PUSH(flat_maps, (UWord)mp);
+                    mp->thing_word = MAP_HEADER_FLATMAP;
                     mp->size       = size;
                     mp->keys       = keys;
                     *objp          = make_flatmap(mp);
 
                     for (n = size; n; n--) {
-                        *vptr = (Eterm) COMPRESS_POINTER(next);
-                        *kptr = (Eterm) COMPRESS_POINTER(vptr);
+                        *vptr = (Eterm) next;
+                        *kptr = (Eterm) vptr;
                         next  = kptr;
                         vptr--;
                         kptr--;
@@ -3623,8 +3595,8 @@ dec_term_atom_common:
                     hamt->leaf_array = hp;
 
                     for (n = size; n; n--) {
-                        CDR(hp) = (Eterm) COMPRESS_POINTER(next);
-                        CAR(hp) = (Eterm) COMPRESS_POINTER(&CDR(hp));
+                        CDR(hp) = (Eterm) next;
+                        CAR(hp) = (Eterm) &CDR(hp);
                         next = &CAR(hp);
                         hp += 2;
                     }
@@ -3701,11 +3673,11 @@ dec_term_atom_common:
 
 		/* Environment */
 		for (i = num_free-1; i >= 0; i--) {
-		    funp->env[i] = (Eterm) COMPRESS_POINTER(next);
+		    funp->env[i] = (Eterm) next;
 		    next = funp->env + i;
 		}
 		/* Creator */
-		funp->creator = (Eterm) COMPRESS_POINTER(next);
+		funp->creator = (Eterm) next;
 		next = &(funp->creator);
 		break;
 	    }
@@ -3774,7 +3746,7 @@ dec_term_atom_common:
 
 		/* Environment */
 		for (i = num_free-1; i >= 0; i--) {
-		    funp->env[i] = (Eterm) COMPRESS_POINTER(next);
+		    funp->env[i] = (Eterm) next;
 		    next = funp->env + i;
 		}
 		break;
@@ -3851,7 +3823,9 @@ dec_term_atom_common:
                     ctx->u.dc.ep = ep;
                     ctx->u.dc.next = next;
                     ctx->u.dc.factory.hp = hp;
-		    ctx->u.dc.maps_list = maps_list;
+		    if (!WSTACK_ISEMPTY(flat_maps)) {
+			WSTACK_SAVE(flat_maps, &ctx->u.dc.flat_maps);
+		    }
 		    if (!PSTACK_IS_EMPTY(hamt_array)) {
 			PSTACK_SAVE(hamt_array, &ctx->u.dc.hamt_array);
 		    }
@@ -3865,18 +3839,6 @@ dec_term_atom_common:
         }
     }
 
-    /* Iterate through all the maps and check for validity and sort keys
-     * - done here for when we know it is complete.
-     */
-
-    while (maps_list) {
-	next  = (Eterm *)(EXPAND_POINTER(*maps_list));
-	*maps_list = MAP_HEADER_FLATMAP;
-	if (!erts_validate_and_sort_flatmap((flatmap_t*)maps_list))
-	    goto error;
-	maps_list  = next;
-    }
-
     ASSERT(hp <= factory->hp_end
            || (factory->mode == FACTORY_CLOSED && is_immed(*dbg_resultp)));
     factory->hp = hp;
@@ -3885,22 +3847,33 @@ dec_term_atom_common:
      */
 
     if (!PSTACK_IS_EMPTY(hamt_array)) {
-	do {
-	    struct dec_term_hamt* hamt = PSTACK_TOP(hamt_array);
-
-	    *hamt->objp = erts_hashmap_from_array(factory,
-						  hamt->leaf_array,
-						  hamt->size,
-						  1);
-	    if (is_non_value(*hamt->objp))
-		goto error_hamt;
-
-	    (void) PSTACK_POP(hamt_array);
-	} while (!PSTACK_IS_EMPTY(hamt_array));
-	PSTACK_DESTROY(hamt_array);
+        do {
+            struct dec_term_hamt* hamt = PSTACK_TOP(hamt_array);
+
+            *hamt->objp = erts_hashmap_from_array(factory,
+                                                  hamt->leaf_array,
+                                                  hamt->size,
+                                                  1);
+            if (is_non_value(*hamt->objp))
+                goto error_hamt;
+
+            (void) PSTACK_POP(hamt_array);
+        } while (!PSTACK_IS_EMPTY(hamt_array));
+        PSTACK_DESTROY(hamt_array);
+    }
+
+    /* Iterate through all the (flat)maps and check for validity and sort keys
+     * - done here for when we know it is complete.
+     */
+
+    while(!WSTACK_ISEMPTY(flat_maps)) {
+        next = (Eterm *)WSTACK_POP(flat_maps);
+        if (!erts_validate_and_sort_flatmap((flatmap_t*)next))
+            goto error;
     }
+    WSTACK_DESTROY(flat_maps);
 
-    ASSERT((Eterm*)EXPAND_POINTER(*dbg_resultp) != NULL);
+    ASSERT((Eterm*)*dbg_resultp != NULL);
 
     if (ctx) {
         ctx->state = B2TDone;
@@ -3924,6 +3897,7 @@ error_hamt:
 	ctx->state = B2TDecodeFail;
 	ctx->reds = reds;
     }
+    WSTACK_DESTROY(flat_maps);
         
     return NULL;
 }
@@ -4192,11 +4166,7 @@ encode_size_struct_int(TTBSizeContext* ctx, ErtsAtomCacheMap *acmp, Eterm obj,
 	case EXPORT_DEF:
 	    {
 		Export* ep = *((Export **) (export_val(obj) + 1));
-#if HALFWORD_HEAP
-		result += 2;
-#else
 		result += 1;
-#endif
 		result += encode_size_struct2(acmp, ep->code[0], dflags);
 		result += encode_size_struct2(acmp, ep->code[1], dflags);
 		result += encode_size_struct2(acmp, make_small(ep->code[2]), dflags);
@@ -4311,7 +4281,7 @@ init_done:
 	switch (tag) {
 	case INTEGER_EXT:
 	    SKIP(4);
-#if !defined(ARCH_64) || HALFWORD_HEAP
+#if !defined(ARCH_64)
 	    heap_size += BIG_UINT_HEAP_SIZE;
 #endif
 	    break;
@@ -4400,7 +4370,7 @@ init_done:
 		ep += 2;
 		atom_extra_skip = 1 + 4*id_words;
 		/* In case it is an external ref */
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
 		heap_size += EXTERNAL_THING_HEAD_SIZE + id_words/2 + 1;
 #else
 		heap_size += EXTERNAL_THING_HEAD_SIZE + id_words;
@@ -4486,11 +4456,7 @@ init_done:
 	    break;
 	case EXPORT_EXT:
 	    terms += 3;
-#if HALFWORD_HEAP
-	    heap_size += 3;
-#else
 	    heap_size += 2;
-#endif
 	    break;
 	case NEW_FUN_EXT:
 	    {
diff --git a/erts/emulator/beam/global.h b/erts/emulator/beam/global.h
index ec9296d034..3f5925765d 100644
--- a/erts/emulator/beam/global.h
+++ b/erts/emulator/beam/global.h
@@ -310,9 +310,6 @@ typedef union {
 typedef struct proc_bin {
     Eterm thing_word;		/* Subtag REFC_BINARY_SUBTAG. */
     Uint size;			/* Binary size in bytes. */
-#if HALFWORD_HEAP
-    void* dummy_ptr_padding__;
-#endif
     struct erl_off_heap_header *next;
     Binary *val;		/* Pointer to Binary structure. */
     byte *bytes;		/* Pointer to the actual data bytes. */
@@ -383,7 +380,7 @@ extern int bif_reductions;      /* reductions + fcalls (when doing call_bif) */
 extern int stackdump_on_exit;
 
 /*
- * Here is an implementation of a lightweiht stack.
+ * Here is an implementation of a lightweight stack.
  *
  * Use it like this:
  *
@@ -848,6 +845,94 @@ do {\
 } while(0)
 
 
+/*
+ *  An implementation of lightweight unbounded queues,
+ *  using a circular dynamic array.
+ *  It does not include support for change_allocator.
+ *
+ *  Use it like this:
+ *
+ *  DECLARE_EQUEUE(Queue)	(At the start of a block)
+ *  ...
+ *  EQUEUE_PUT(Queue, Term)
+ *  ...
+ *  if (EQUEUE_ISEMPTY(Queue)) {
+ *     Queue is empty
+ *  } else {
+ *     Term = EQUEUE_GET(Stack);
+ *     Process popped Term here
+ *  }
+ *  ...
+ *  DESTROY_EQUEUE(Queue)
+ */
+
+typedef struct {
+    Eterm* start;
+    Eterm* front;
+    Eterm* back;
+    int possibly_empty;
+    Eterm* end;
+    ErtsAlcType_t alloc_type;
+} ErtsEQueue;
+
+#define DEF_EQUEUE_SIZE (16)
+
+void erl_grow_equeue(ErtsEQueue*, Eterm* def_queue);
+#define EQUE_CONCAT(a,b) a##b
+#define EQUE_DEF_QUEUE(q) EQUE_CONCAT(q,_default_equeue)
+
+#define DECLARE_EQUEUE(q)				\
+    UWord EQUE_DEF_QUEUE(q)[DEF_EQUEUE_SIZE];     	\
+    ErtsEQueue q = {					\
+        EQUE_DEF_QUEUE(q), /* start */			\
+        EQUE_DEF_QUEUE(q), /* front */			\
+        EQUE_DEF_QUEUE(q), /* back */			\
+        1,                 /* possibly_empty */		\
+        EQUE_DEF_QUEUE(q) + DEF_EQUEUE_SIZE, /* end */	\
+        ERTS_ALC_T_ESTACK  /* alloc_type */		\
+    }
+
+#define DESTROY_EQUEUE(q)				\
+do {							\
+    if (q.start != EQUE_DEF_QUEUE(q)) {			\
+      erts_free(q.alloc_type, q.start);			\
+    }							\
+} while(0)
+
+#define EQUEUE_PUT_UNCHECKED(q, x)			\
+do {							\
+    q.possibly_empty = 0;				\
+    *(q.back) = (x);                    		\
+    if (++(q.back) == q.end) {				\
+	q.back = q.start;				\
+    }							\
+} while(0)
+
+#define EQUEUE_PUT(q, x)				\
+do {							\
+    if (q.back == q.front && !q.possibly_empty) {	\
+        erl_grow_equeue(&q, EQUE_DEF_QUEUE(q));		\
+    }							\
+    EQUEUE_PUT_UNCHECKED(q, x);				\
+} while(0)
+
+#define EQUEUE_ISEMPTY(q) (q.back == q.front && q.possibly_empty)
+
+ERTS_GLB_INLINE Eterm erts_equeue_get(ErtsEQueue *q);
+
+#if ERTS_GLB_INLINE_INCL_FUNC_DEF
+ERTS_GLB_INLINE Eterm erts_equeue_get(ErtsEQueue *q) {
+    Eterm x;
+    q->possibly_empty = 1;
+    x = *(q->front);
+    if (++(q->front) == q->end) {
+        q->front = q->start;
+    }
+    return x;
+}
+#endif
+#define EQUEUE_GET(q) erts_equeue_get(&(q));
+
 /* binary.c */
 
 void erts_emasculate_writable_binary(ProcBin* pb);
@@ -900,8 +985,18 @@ Eterm erl_send(Process *p, Eterm to, Eterm msg);
 Eterm erl_is_function(Process* p, Eterm arg1, Eterm arg2);
 
 /* beam_bif_load.c */
-Eterm erts_check_process_code(Process *c_p, Eterm module, int allow_gc, int *redsp);
+#define ERTS_CPC_ALLOW_GC      (1 << 0)
+#define ERTS_CPC_COPY_LITERALS (1 << 1)
+#define ERTS_CPC_ALL           (ERTS_CPC_ALLOW_GC | ERTS_CPC_COPY_LITERALS)
+Eterm erts_check_process_code(Process *c_p, Eterm module, Uint flags, int *redsp);
+
+typedef struct {
+    Eterm *ptr;
+    Uint   sz;
+    Eterm pid;
+} copy_literals_t;
 
+extern copy_literals_t erts_clrange;
 
 /* beam_load.c */
 typedef struct {
@@ -956,35 +1051,71 @@ __decl_noreturn void __noreturn erl_exit(int n, char*, ...);
 __decl_noreturn void __noreturn erl_exit_flush_async(int n, char*, ...);
 void erl_error(char*, va_list);
 
-/* copy.c */
-Eterm copy_object(Eterm, Process*);
+/* This controls whether sharing-preserving copy is used by Erlang */
 
-#if HALFWORD_HEAP
-Uint size_object_rel(Eterm, Eterm*);
-#  define size_object(A) size_object_rel(A,NULL)
+#ifdef SHCOPY
+#define SHCOPY_SEND
+#define SHCOPY_SPAWN
+#endif
 
-Eterm copy_struct_rel(Eterm, Uint, Eterm**, ErlOffHeap*, Eterm* src_base, Eterm* dst_base);
-#  define copy_struct(OBJ,SZ,HPP,OH) copy_struct_rel(OBJ,SZ,HPP,OH, NULL,NULL)
+/* The persistent state while the sharing-preserving copier works */
 
-Eterm copy_shallow_rel(Eterm*, Uint, Eterm**, ErlOffHeap*, Eterm* src_base);
-#  define copy_shallow(A,B,C,D) copy_shallow_rel(A,B,C,D,NULL)
+typedef struct {
+    Eterm  queue_default[DEF_EQUEUE_SIZE];
+    Eterm* queue_start;
+    Eterm* queue_end;
+    ErtsAlcType_t queue_alloc_type;
+    UWord  bitstore_default[DEF_WSTACK_SIZE];
+    UWord* bitstore_start;
+    ErtsAlcType_t bitstore_alloc_type;
+    Eterm  shtable_default[DEF_ESTACK_SIZE];
+    Eterm* shtable_start;
+    ErtsAlcType_t shtable_alloc_type;
+    Uint literal_size;
+    Eterm *range_ptr;
+    Uint  range_sz;
+} erts_shcopy_t;
+
+#define INITIALIZE_SHCOPY(info)                         \
+do {                                                    \
+    info.queue_start = info.queue_default;              \
+    info.bitstore_start = info.bitstore_default;        \
+    info.shtable_start = info.shtable_default;          \
+    info.literal_size = 0;                              \
+    info.range_ptr = erts_clrange.ptr;                  \
+    info.range_sz  = erts_clrange.sz;                   \
+} while(0)
 
-#else /* !HALFWORD_HEAP */
+#define DESTROY_SHCOPY(info)                                            \
+do {                                                                    \
+    if (info.queue_start != info.queue_default) {                       \
+        erts_free(info.queue_alloc_type, info.queue_start);             \
+    }                                                                   \
+    if (info.bitstore_start != info.bitstore_default) {                 \
+        erts_free(info.bitstore_alloc_type, info.bitstore_start);       \
+    }                                                                   \
+    if (info.shtable_start != info.shtable_default) {                   \
+        erts_free(info.shtable_alloc_type, info.shtable_start);         \
+    }                                                                   \
+} while(0)
+
+/* copy.c */
+Eterm copy_object_x(Eterm, Process*, Uint);
+#define copy_object(Term, Proc) copy_object_x(Term,Proc,0)
 
 Uint size_object(Eterm);
-#  define size_object_rel(A,B) size_object(A)
+Uint copy_shared_calculate(Eterm, erts_shcopy_t*);
+Eterm copy_shared_perform(Eterm, Uint, erts_shcopy_t*, Eterm**, ErlOffHeap*);
 
-Eterm copy_struct(Eterm, Uint, Eterm**, ErlOffHeap*);
-#  define copy_struct_rel(OBJ,SZ,HPP,OH, SB,DB) copy_struct(OBJ,SZ,HPP,OH)
+Uint size_shared(Eterm);
 
+Eterm copy_struct_x(Eterm, Uint, Eterm**, ErlOffHeap*, Uint* bsz);
+#define copy_struct(Obj,Sz,HPP,OH) \
+    copy_struct_x(Obj,Sz,HPP,OH,NULL)
 Eterm copy_shallow(Eterm*, Uint, Eterm**, ErlOffHeap*);
-#  define copy_shallow_rel(A,B,C,D, BASE) copy_shallow(A,B,C,D)
-
-#endif
 
-
-void move_multi_frags(Eterm** hpp, ErlOffHeap*, ErlHeapFragment* first,
-		      Eterm* refs, unsigned nrefs);
+void erts_move_multi_frags(Eterm** hpp, ErlOffHeap*, ErlHeapFragment* first,
+			   Eterm* refs, unsigned nrefs, int literals);
 
 /* Utilities */
 extern void erts_delete_nodes_monitors(Process *, ErtsProcLocks);
@@ -1180,7 +1311,7 @@ void bin_write(int, void*, byte*, size_t);
 int intlist_to_buf(Eterm, char*, int); /* most callers pass plain char*'s */
 
 struct Sint_buf {
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
     char s[22];
 #else
     char s[12];
@@ -1296,6 +1427,27 @@ int erts_print_system_version(int to, void *arg, Process *c_p);
 
 int erts_hibernate(Process* c_p, Eterm module, Eterm function, Eterm args, Eterm* reg);
 
+ERTS_GLB_FORCE_INLINE int erts_is_literal(Eterm tptr, Eterm *ptr);
+
+#if ERTS_GLB_INLINE_INCL_FUNC_DEF
+
+ERTS_GLB_FORCE_INLINE int erts_is_literal(Eterm tptr, Eterm *ptr)
+{
+    ASSERT(is_boxed(tptr) || is_list(tptr));
+    ASSERT(ptr == ptr_val(tptr));
+
+#if defined(ERTS_HAVE_IS_IN_LITERAL_RANGE)
+    return erts_is_in_literal_range(ptr);
+#elif defined(TAG_LITERAL_PTR)
+    return is_literal_ptr(tptr);
+#else
+#  error Not able to detect literals...
+#endif
+
+}
+
+#endif
+
 /*
 ** Call_trace uses this API for the parameter matching functions
 */
@@ -1344,191 +1496,21 @@ extern void erts_match_prog_foreach_offheap(Binary *b,
 
 extern erts_driver_t vanilla_driver;
 extern erts_driver_t spawn_driver;
+extern erts_driver_t forker_driver;
 extern erts_driver_t fd_driver;
 
 int erts_beam_jump_table(void);
 
-/* Should maybe be placed in erl_message.h, but then we get an include mess. */
-ERTS_GLB_INLINE Eterm *
-erts_alloc_message_heap_state(Uint size,
-			      ErlHeapFragment **bpp,
-			      ErlOffHeap **ohpp,
-			      Process *receiver,
-			      ErtsProcLocks *receiver_locks,
-			      erts_aint32_t *statep);
-
-ERTS_GLB_INLINE Eterm *
-erts_alloc_message_heap(Uint size,
-			ErlHeapFragment **bpp,
-			ErlOffHeap **ohpp,
-			Process *receiver,
-			ErtsProcLocks *receiver_locks);
-
-#if ERTS_GLB_INLINE_INCL_FUNC_DEF
-
-/*
- * NOTE: erts_alloc_message_heap() releases msg q and status
- *       lock on receiver without ensuring that other locks are
- *       held. User is responsible to ensure that the receiver
- *       pointer cannot become invalid until after message has
- *       been passed. This is normal done either by increasing
- *       reference count on process (preferred) or by holding
- *       main or link lock over the whole message passing
- *       operation.
- */
-
-ERTS_GLB_INLINE Eterm *
-erts_alloc_message_heap_state(Uint size,
-			      ErlHeapFragment **bpp,
-			      ErlOffHeap **ohpp,
-			      Process *receiver,
-			      ErtsProcLocks *receiver_locks,
-			      erts_aint32_t *statep)
-{
-    Eterm *hp;
-    erts_aint32_t state;
-#ifdef ERTS_SMP
-    int locked_main = 0;
-    state = erts_smp_atomic32_read_acqb(&receiver->state);
-    if (statep)
-	*statep = state;
-    if (state & (ERTS_PSFLG_EXITING
-		 | ERTS_PSFLG_PENDING_EXIT))
-	goto allocate_in_mbuf;
-#endif
-
-    if (size > (Uint) INT_MAX)
-	erl_exit(ERTS_ABORT_EXIT, "HUGE size (%beu)\n", size);
-
-    if (
-#if defined(ERTS_SMP)
-	*receiver_locks & ERTS_PROC_LOCK_MAIN
-#else
-	1
-#endif
-	) {
-#ifdef ERTS_SMP
-    try_allocate_on_heap:
-#endif
-	state = erts_smp_atomic32_read_nob(&receiver->state);
-	if (statep)
-	    *statep = state;
-	if ((state & (ERTS_PSFLG_EXITING
-		      | ERTS_PSFLG_PENDING_EXIT))
-	    || (receiver->flags & F_DISABLE_GC)
-	    || HEAP_LIMIT(receiver) - HEAP_TOP(receiver) <= size) {
-	    /*
-	     * The heap is either potentially in an inconsistent
-	     * state, or not large enough.
-	     */
-#ifdef ERTS_SMP
-	    if (locked_main) {
-		*receiver_locks &= ~ERTS_PROC_LOCK_MAIN;
-		erts_smp_proc_unlock(receiver, ERTS_PROC_LOCK_MAIN);
-	    }
-#endif
-	    goto allocate_in_mbuf;
-	}
-	hp = HEAP_TOP(receiver);
-	HEAP_TOP(receiver) = hp + size;
-	*bpp = NULL;
-	*ohpp = &MSO(receiver);
-    }
-#ifdef ERTS_SMP
-    else if (erts_smp_proc_trylock(receiver, ERTS_PROC_LOCK_MAIN) == 0) {
-	locked_main = 1;
-	*receiver_locks |= ERTS_PROC_LOCK_MAIN;
-	goto try_allocate_on_heap;
-    }
-#endif
-    else {
-	ErlHeapFragment *bp;
-    allocate_in_mbuf:
-	bp = new_message_buffer(size);
-	hp = bp->mem;
-	*bpp = bp;
-	*ohpp = &bp->off_heap;
-    }
-
-    return hp;
-}
-
-ERTS_GLB_INLINE Eterm *
-erts_alloc_message_heap(Uint size,
-			ErlHeapFragment **bpp,
-			ErlOffHeap **ohpp,
-			Process *receiver,
-			ErtsProcLocks *receiver_locks)
-{
-    return erts_alloc_message_heap_state(size, bpp, ohpp, receiver,
-					 receiver_locks, NULL);
-}
-
-#endif /* #if ERTS_GLB_INLINE_INCL_FUNC_DEF */
-
-#if !HEAP_ON_C_STACK
-#  if defined(DEBUG)
-#    define DeclareTmpHeap(VariableName,Size,Process) \
-       Eterm *VariableName = erts_debug_allocate_tmp_heap(Size,Process)
-#    define DeclareTypedTmpHeap(Type,VariableName,Process)		\
-      Type *VariableName = (Type *) erts_debug_allocate_tmp_heap(sizeof(Type)/sizeof(Eterm),Process)
-#    define DeclareTmpHeapNoproc(VariableName,Size) \
-       Eterm *VariableName = erts_debug_allocate_tmp_heap(Size,NULL)
-#    define UseTmpHeap(Size,Proc) \
-       do { \
-         erts_debug_use_tmp_heap((Size),(Proc)); \
-       } while (0)
-#    define UnUseTmpHeap(Size,Proc) \
-       do { \
-         erts_debug_unuse_tmp_heap((Size),(Proc)); \
-       } while (0)
-#    define UseTmpHeapNoproc(Size) \
-       do { \
-         erts_debug_use_tmp_heap(Size,NULL); \
-       } while (0)
-#    define UnUseTmpHeapNoproc(Size) \
-       do { \
-         erts_debug_unuse_tmp_heap(Size,NULL); \
-       } while (0)
-#  else
-#    define DeclareTmpHeap(VariableName,Size,Process) \
-       Eterm *VariableName = (ERTS_PROC_GET_SCHDATA(Process)->tmp_heap)+(ERTS_PROC_GET_SCHDATA(Process)->num_tmp_heap_used)
-#    define DeclareTypedTmpHeap(Type,VariableName,Process)		\
-      Type *VariableName = (Type *) (ERTS_PROC_GET_SCHDATA(Process)->tmp_heap)+(ERTS_PROC_GET_SCHDATA(Process)->num_tmp_heap_used)
-#    define DeclareTmpHeapNoproc(VariableName,Size) \
-       Eterm *VariableName = (erts_get_scheduler_data()->tmp_heap)+(erts_get_scheduler_data()->num_tmp_heap_used)
-#    define UseTmpHeap(Size,Proc) \
-       do { \
-         ERTS_PROC_GET_SCHDATA(Proc)->num_tmp_heap_used += (Size); \
-       } while (0)
-#    define UnUseTmpHeap(Size,Proc) \
-       do { \
-         ERTS_PROC_GET_SCHDATA(Proc)->num_tmp_heap_used -= (Size); \
-       } while (0)
-#    define UseTmpHeapNoproc(Size) \
-       do { \
-         erts_get_scheduler_data()->num_tmp_heap_used += (Size); \
-       } while (0)
-#    define UnUseTmpHeapNoproc(Size) \
-       do { \
-         erts_get_scheduler_data()->num_tmp_heap_used -= (Size); \
-       } while (0)
-
-
-#  endif
-
-#else
-#  define DeclareTmpHeap(VariableName,Size,Process) \
+#define DeclareTmpHeap(VariableName,Size,Process) \
      Eterm VariableName[Size]
-#  define DeclareTypedTmpHeap(Type,VariableName,Process)	\
+#define DeclareTypedTmpHeap(Type,VariableName,Process)	\
      Type VariableName[1]
-#  define DeclareTmpHeapNoproc(VariableName,Size) \
+#define DeclareTmpHeapNoproc(VariableName,Size) \
      Eterm VariableName[Size]
-#  define UseTmpHeap(Size,Proc) /* Nothing */
-#  define UnUseTmpHeap(Size,Proc) /* Nothing */
-#  define UseTmpHeapNoproc(Size) /* Nothing */
-#  define UnUseTmpHeapNoproc(Size) /* Nothing */
-#endif /* HEAP_ON_C_STACK */
+#define UseTmpHeap(Size,Proc) /* Nothing */
+#define UnUseTmpHeap(Size,Proc) /* Nothing */
+#define UseTmpHeapNoproc(Size) /* Nothing */
+#define UnUseTmpHeapNoproc(Size) /* Nothing */
 
 ERTS_GLB_INLINE void dtrace_pid_str(Eterm pid, char *process_buf);
 ERTS_GLB_INLINE void dtrace_proc_str(Process *process, char *process_buf);
@@ -1576,6 +1558,7 @@ dtrace_fun_decode(Process *process,
     erts_snprintf(mfa_buf, DTRACE_TERM_BUF_SIZE, "%T:%T/%d",
                   module, function, arity);
 }
+
 #endif /* #if ERTS_GLB_INLINE_INCL_FUNC_DEF */
 
 #endif /* !__GLOBAL_H__ */
diff --git a/erts/emulator/beam/hash.c b/erts/emulator/beam/hash.c
index e0fde337f2..5a0b93f693 100644
--- a/erts/emulator/beam/hash.c
+++ b/erts/emulator/beam/hash.c
@@ -27,8 +27,6 @@
 #endif
 
 #include "sys.h"
-#include "erl_vm.h"
-#include "global.h"
 #include "hash.h"
 
 /*
@@ -66,6 +64,7 @@ void hash_get_info(HashInfo *hi, Hash *h)
     int i;
     int max_depth = 0;
     int objects = 0;
+    int used = 0;
 
     for (i = 0; i < size; i++) {
 	int depth = 0;
@@ -76,14 +75,18 @@ void hash_get_info(HashInfo *hi, Hash *h)
 	    depth++;
 	    b = b->next;
 	}
-	if (depth > max_depth)
-	    max_depth = depth;
+        if (depth) {
+            used++;
+            if (depth > max_depth)
+                max_depth = depth;
+        }
     }
+    ASSERT(objects == h->nobjs);
 
     hi->name  = h->name;
     hi->size  = h->size;
-    hi->used  = h->used;
-    hi->objs  = objects;
+    hi->used  = used;
+    hi->objs  = h->nobjs;
     hi->depth = max_depth;
 }
 
@@ -98,11 +101,11 @@ void hash_info(int to, void *arg, Hash* h)
 
     hash_get_info(&hi, h);
 
-    erts_print(to, arg, "=hash_table:%s\n", hi.name);
-    erts_print(to, arg, "size: %d\n",       hi.size);
-    erts_print(to, arg, "used: %d\n",       hi.used);
-    erts_print(to, arg, "objs: %d\n",       hi.objs);
-    erts_print(to, arg, "depth: %d\n",      hi.depth);
+    h->fun.meta_print(to, arg, "=hash_table:%s\n", hi.name);
+    h->fun.meta_print(to, arg, "size: %d\n",       hi.size);
+    h->fun.meta_print(to, arg, "used: %d\n",       hi.used);
+    h->fun.meta_print(to, arg, "objs: %d\n",       hi.objs);
+    h->fun.meta_print(to, arg, "depth: %d\n",      hi.depth);
 }
 
 
@@ -119,47 +122,56 @@ hash_table_sz(Hash *h)
 }
 
 
+static ERTS_INLINE void set_thresholds(Hash* h)
+{
+    h->grow_threshold = (8*h->size)/5;   /* grow at 160% load */
+    if (h->size_ix > h->min_size_ix)
+        h->shrink_threshold = h->size / 5;  /* shrink at 20% load */
+    else
+        h->shrink_threshold = -1;  /* never shrink below inital size */
+}
+
 /*
 ** init a pre allocated or static hash structure
 ** and allocate buckets.
 */
-Hash* hash_init(ErtsAlcType_t type, Hash* h, char* name, int size, HashFunctions fun)
+Hash* hash_init(int type, Hash* h, char* name, int size, HashFunctions fun)
 {
     int sz;
     int ix = 0;
 
-    h->type = type;
+    h->meta_alloc_type = type;
 
     while (h_size_table[ix] != -1 && h_size_table[ix] < size)
 	ix++;
     if (h_size_table[ix] == -1)
-	erl_exit(1, "panic: too large hash table size (%d)\n", size);
+	return NULL;
 
     size = h_size_table[ix];
     sz = size*sizeof(HashBucket*);
 
-    h->bucket = (HashBucket**) erts_alloc(h->type, sz);
+    h->bucket = (HashBucket**) fun.meta_alloc(h->meta_alloc_type, sz);
 
     sys_memzero(h->bucket, sz);
     h->is_allocated = 0;
     h->name = name;
     h->fun = fun;
     h->size = size;
-    h->size20percent = h->size/5;
-    h->size80percent = (4*h->size)/5;
-    h->ix = ix;
-    h->used = 0;
+    h->size_ix = ix;
+    h->min_size_ix = ix;
+    h->nobjs = 0;
+    set_thresholds(h);
     return h;
 }
 
 /*
 ** Create a new hash table
 */
-Hash* hash_new(ErtsAlcType_t type, char* name, int size, HashFunctions fun)
+Hash* hash_new(int type, char* name, int size, HashFunctions fun)
 {
     Hash* h;
 
-    h = erts_alloc(type, sizeof(Hash));
+    h = fun.meta_alloc(type, sizeof(Hash));
 
     h = hash_init(type, h, name, size, fun);
     h->is_allocated =  1;
@@ -183,9 +195,9 @@ void hash_delete(Hash* h)
 	    b = b_next;
 	}
     }
-    erts_free(h->type, h->bucket);
+    h->fun.meta_free(h->meta_alloc_type, h->bucket);
     if (h->is_allocated)
-	erts_free(h->type, (void*) h);
+	h->fun.meta_free(h->meta_alloc_type, (void*) h);
 }
 
 /*
@@ -199,39 +211,34 @@ static void rehash(Hash* h, int grow)
     int i;
 
     if (grow) {
-	if ((h_size_table[h->ix+1]) == -1)
+	if ((h_size_table[h->size_ix+1]) == -1)
 	    return;
-	h->ix++;
+	h->size_ix++;
     }
     else {
-	if (h->ix == 0)
+	if (h->size_ix == 0)
 	    return;
-	h->ix--;
+	h->size_ix--;
     }
-    h->size = h_size_table[h->ix];
-    h->size20percent = h->size/5;
-    h->size80percent = (4*h->size)/5;
+    h->size = h_size_table[h->size_ix];
     sz = h->size*sizeof(HashBucket*);
 
-    new_bucket = (HashBucket **) erts_alloc(h->type, sz);
+    new_bucket = (HashBucket **) h->fun.meta_alloc(h->meta_alloc_type, sz);
     sys_memzero(new_bucket, sz);
 
-    h->used = 0;
-
     for (i = 0; i < old_size; i++) {
 	HashBucket* b = h->bucket[i];
 	while (b != (HashBucket*) 0) {
 	    HashBucket* b_next = b->next;
 	    int ix = b->hvalue % h->size;
-	    if (new_bucket[ix] == NULL)
-		h->used++;
 	    b->next = new_bucket[ix];
 	    new_bucket[ix] = b;
 	    b = b_next;
 	}
     }
-    erts_free(h->type, (void *) h->bucket);
+    h->fun.meta_free(h->meta_alloc_type, (void *) h->bucket);
     h->bucket = new_bucket;
+    set_thresholds(h);
 }
 
 /*
@@ -268,68 +275,15 @@ void* hash_put(Hash* h, void* tmpl)
     }
     b = (HashBucket*) h->fun.alloc(tmpl);
 
-    if (h->bucket[ix] == NULL)
-	h->used++;
-
     b->hvalue = hval;
     b->next = h->bucket[ix];
     h->bucket[ix] = b;
 
-    if (h->used > h->size80percent)  /* rehash at 80% */
+    if (++h->nobjs > h->grow_threshold)
 	rehash(h, 1);
     return (void*) b;
 }
 
-static void
-hash_insert_entry(Hash* h, HashBucket* entry)
-{
-    HashValue hval = entry->hvalue;
-    int ix = hval % h->size;
-    HashBucket* b = h->bucket[ix];
-
-    while (b != (HashBucket*) 0) {
-	if ((b->hvalue == hval) && (h->fun.cmp((void*)entry, (void*)b) == 0)) {
-	    abort();		/* Should not happen */
-	}
-	b = b->next;
-    }
-
-    if (h->bucket[ix] == NULL)
-	h->used++;
-
-    entry->next = h->bucket[ix];
-    h->bucket[ix] = entry;
-
-    if (h->used > h->size80percent)  /* rehash at 80% */
-	rehash(h, 1);
-}
-
-
-/*
- * Move all entries in src into dst; empty src.
- * Entries in src must not exist in dst.
- */
-void
-erts_hash_merge(Hash* src, Hash* dst)
-{
-    int limit = src->size;
-    HashBucket** bucket = src->bucket;
-    int i;
-
-    src->used = 0;
-    for (i = 0; i < limit; i++) {
-	HashBucket* b = bucket[i];
-	HashBucket* next;
-
-	bucket[i] = NULL;
-	while (b) {
-	    next = b->next;
-	    hash_insert_entry(dst, b);
-	    b = next;
-	}
-    }
-}
-
 /*
 ** Erase hash entry return template if erased
 ** return 0 if not erased
@@ -348,9 +302,7 @@ void* hash_erase(Hash* h, void* tmpl)
 	    else
 		h->bucket[ix] = b->next;
 	    h->fun.free((void*)b);
-	    if (h->bucket[ix] == NULL)
-		h->used--;
-	    if (h->used < h->size20percent)  /* rehash at 20% */
+	    if (--h->nobjs < h->shrink_threshold)
 		rehash(h, 0);
 	    return tmpl;
 	}
@@ -381,9 +333,7 @@ hash_remove(Hash *h, void *tmpl)
 		prev->next = b->next;
 	    else
 		h->bucket[ix] = b->next;
-	    if (h->bucket[ix] == NULL)
-		h->used--;
-	    if (h->used < h->size20percent)  /* rehash at 20% */
+	    if (--h->nobjs < h->shrink_threshold)
 		rehash(h, 0);
 	    return (void *) b;
 	}
diff --git a/erts/emulator/beam/hash.h b/erts/emulator/beam/hash.h
index 87fdb360e3..e94aaa0a84 100644
--- a/erts/emulator/beam/hash.h
+++ b/erts/emulator/beam/hash.h
@@ -29,14 +29,17 @@
 #include "sys.h"
 #endif
 
-#include "erl_alloc.h"
-
 typedef unsigned long HashValue;
+typedef struct hash Hash;
 
 typedef int (*HCMP_FUN)(void*, void*);
 typedef HashValue (*H_FUN)(void*);
 typedef void* (*HALLOC_FUN)(void*);
 typedef void (*HFREE_FUN)(void*);
+/* Meta functions */
+typedef void* (*HMALLOC_FUN)(int,size_t);
+typedef void (*HMFREE_FUN)(int,void*);
+typedef int (*HMPRINT_FUN)(int,void*,char*, ...);
 
 /*
 ** This bucket must be placed in top of 
@@ -55,6 +58,9 @@ typedef struct hash_functions
     HCMP_FUN cmp;
     HALLOC_FUN alloc;
     HFREE_FUN free;
+    HMALLOC_FUN meta_alloc;
+    HMFREE_FUN meta_free;
+    HMPRINT_FUN meta_print;
 } HashFunctions;
 
 typedef struct {
@@ -65,22 +71,23 @@ typedef struct {
   int   depth;
 } HashInfo;
 
-typedef struct hash
+struct hash
 {
     HashFunctions fun;   /* Function block */
     int is_allocated;    /* 0 iff hash structure is on stack or is static */
-    ErtsAlcType_t type;
+    int meta_alloc_type; /* argument to pass to meta_alloc and meta_free */
     char* name;          /* Table name (static string, for debugging) */
     int size;		 /* Number of slots */
-    int size20percent;   /* 20 percent of number of slots */
-    int size80percent;   /* 80 percent of number of slots */
-    int ix;              /* Size index in size table */
-    int used;		 /* Number of slots used */
+    int shrink_threshold;
+    int grow_threshold;
+    int size_ix;         /* Size index in size table */
+    int min_size_ix;     /* Never shrink table smaller than this */
+    int nobjs;		 /* Number of objects in table */
     HashBucket** bucket; /* Vector of bucket pointers (objects) */
-} Hash;
+};
 
-Hash* hash_new(ErtsAlcType_t, char*, int, HashFunctions);
-Hash* hash_init(ErtsAlcType_t, Hash*, char*, int, HashFunctions);
+Hash* hash_new(int, char*, int, HashFunctions);
+Hash* hash_init(int, Hash*, char*, int, HashFunctions);
 
 void  hash_delete(Hash*);
 void  hash_get_info(HashInfo*, Hash*);
@@ -93,6 +100,4 @@ void* hash_erase(Hash*, void*);
 void* hash_remove(Hash*, void*);
 void  hash_foreach(Hash*, void (*func)(void *, void *), void *);
 
-void erts_hash_merge(Hash* src, Hash* dst);
-
 #endif
diff --git a/erts/emulator/beam/index.h b/erts/emulator/beam/index.h
index 14fab41026..99b2bdfab0 100644
--- a/erts/emulator/beam/index.h
+++ b/erts/emulator/beam/index.h
@@ -30,6 +30,10 @@
 #include "hash.h"
 #endif
 
+#ifndef ERL_ALLOC_H__
+#include "erl_alloc.h"
+#endif
+
 typedef struct index_slot 
 {
     HashBucket bucket;
diff --git a/erts/emulator/beam/io.c b/erts/emulator/beam/io.c
index 900616c981..93c591b124 100644
--- a/erts/emulator/beam/io.c
+++ b/erts/emulator/beam/io.c
@@ -52,10 +52,11 @@
 #include "erl_hl_timer.h"
 
 extern ErlDrvEntry fd_driver_entry;
-#ifndef __OSE__
 extern ErlDrvEntry vanilla_driver_entry;
-#endif
 extern ErlDrvEntry spawn_driver_entry;
+#ifndef __WIN32__
+extern ErlDrvEntry forker_driver_entry;
+#endif
 extern ErlDrvEntry *driver_tab[]; /* table of static drivers, only used during initialization */
 
 erts_driver_t *driver_list; /* List of all drivers, static and dynamic. */
@@ -73,6 +74,9 @@ const Port erts_invalid_port = {{ERTS_INVALID_PORT}};
 
 erts_driver_t vanilla_driver;
 erts_driver_t spawn_driver;
+#ifndef __WIN32__
+erts_driver_t forker_driver;
+#endif
 erts_driver_t fd_driver;
 
 int erts_port_synchronous_ops = 0;
@@ -86,6 +90,7 @@ static void deliver_result(Eterm sender, Eterm pid, Eterm res);
 static int init_driver(erts_driver_t *, ErlDrvEntry *, DE_Handle *);
 static void terminate_port(Port *p);
 static void pdl_init(void);
+static int driver_failure_term(ErlDrvPort ix, Eterm term, int eof);
 #ifdef ERTS_SMP
 static void driver_monitor_lock_pdl(Port *p);
 static void driver_monitor_unlock_pdl(Port *p);
@@ -307,12 +312,9 @@ static Port *create_port(char *name,
     size_t port_size, busy_port_queue_size, size;
     erts_aint32_t state = ERTS_PORT_SFLG_CONNECTED;
     erts_aint32_t x_pts_flgs = 0;
-#ifdef DEBUG
-    /* Make sure the debug flags survives until port is freed */
-    state |= ERTS_PORT_SFLG_PORT_DEBUG;
-#endif
 
 #ifdef ERTS_SMP
+    ErtsRunQueue *runq;
     if (!driver_lock) {
 	/* Align size for mutex following port struct */
 	port_size = size = ERTS_ALC_DATA_ALIGN_SIZE(sizeof(Port));
@@ -322,6 +324,12 @@ static Port *create_port(char *name,
 #endif
 	port_size = size = ERTS_ALC_DATA_ALIGN_SIZE(sizeof(Port));
 
+#ifdef DEBUG
+    /* Make sure the debug flags survives until port is freed */
+    state |= ERTS_PORT_SFLG_PORT_DEBUG;
+#endif
+
+
     busy_port_queue_size
 	= ((driver->flags & ERL_DRV_FLAG_NO_BUSY_MSGQ)
 	   ? 0
@@ -357,8 +365,12 @@ static Port *create_port(char *name,
 	p += sizeof(erts_mtx_t);
 	state |= ERTS_PORT_SFLG_PORT_SPECIFIC_LOCK;
     }
-    erts_smp_atomic_set_nob(&prt->run_queue,
-			    (erts_aint_t) erts_get_runq_current(NULL));
+    if (erts_get_scheduler_data())
+        runq = erts_get_runq_current(NULL);
+    else
+        runq = ERTS_RUNQ_IX(0);
+    erts_smp_atomic_set_nob(&prt->run_queue, (erts_aint_t) runq);
+
     prt->xports = NULL;
 #else
     erts_atomic32_init_nob(&prt->refc, 1);
@@ -385,6 +397,7 @@ static Port *create_port(char *name,
     ERTS_PTMR_INIT(prt);
     erts_port_task_handle_init(&prt->timeout_task);
     prt->psd = NULL;
+    prt->async_open_port = NULL;
     prt->drv_data = (SWord) 0;
     prt->os_pid = -1;
 
@@ -466,6 +479,11 @@ erts_port_free(Port *prt)
 
     erts_port_task_fini_sched(&prt->sched);
 
+    if (prt->async_open_port) {
+        erts_free(ERTS_ALC_T_PRTSD, prt->async_open_port);
+        prt->async_open_port = NULL;
+    }
+
 #ifdef ERTS_SMP
     ASSERT(prt->lock);
     if (state & ERTS_PORT_SFLG_PORT_SPECIFIC_LOCK)
@@ -1105,7 +1123,7 @@ io_list_vec_len(Eterm obj, int* vsize, Uint* csize,
     Uint p_v_size = 0;
     Uint p_c_size = 0;
     Uint p_in_clist = 0;
-    Uint total; /* Uint due to halfword emulator */
+    Uint total;
 
     goto L_jump_start;  /* avoid a push */
 
@@ -1412,7 +1430,7 @@ queue_port_sched_op_reply(Process *rp,
 
     erts_factory_trim_and_close(factory, &msg, 1);
 
-    erts_queue_message(rp, rp_locksp, factory->heap_frags, msg, NIL);
+    erts_queue_message(rp, rp_locksp, factory->message, msg, NIL);
 }
 
 static void
@@ -1420,12 +1438,9 @@ port_sched_op_reply(Eterm to, Uint32 *ref_num, Eterm msg)
 {
     Process *rp = erts_proc_lookup_raw(to);
     if (rp) {
-	ErlOffHeap *ohp;
-	ErlHeapFragment* bp;
         ErtsHeapFactory factory;
 	Eterm msg_copy;
 	Uint hsz, msg_sz;
-	Eterm *hp;
 	ErtsProcLocks rp_locks = 0;
 
 	hsz = ERTS_QUEUE_PORT_SCHED_OP_REPLY_SIZE;
@@ -1436,18 +1451,13 @@ port_sched_op_reply(Eterm to, Uint32 *ref_num, Eterm msg)
 	    hsz += msg_sz;
 	}
 
-	hp = erts_alloc_message_heap(hsz,
-						&bp,
-						&ohp,
-						rp,
-						&rp_locks);
-        erts_factory_message_init(&factory, rp, hp, bp);
-	if (is_immed(msg))
-	    msg_copy = msg;
-	else {
-	    msg_copy = copy_struct(msg, msg_sz, &hp, ohp);
-            factory.hp = hp;
-        }
+	(void) erts_factory_message_create(&factory, rp,
+					   &rp_locks, hsz);
+	msg_copy = (is_immed(msg)
+		    ? msg
+		    : copy_struct(msg, msg_sz,
+				  &factory.hp,
+				  factory.off_heap));
 
 	queue_port_sched_op_reply(rp,
 				  &rp_locks,
@@ -1535,6 +1545,26 @@ erts_schedule_proc2port_signal(Process *c_p,
     return ERTS_PORT_OP_SCHEDULED;
 }
 
+static int
+erts_schedule_port2port_signal(Eterm port_num, ErtsProc2PortSigData *sigdp,
+                               int task_flags,
+                               ErtsProc2PortSigCallback callback)
+{
+    Port *prt = erts_port_lookup_raw(port_num);
+
+    if (!prt)
+        return -1;
+
+    sigdp->caller = ERTS_INVALID_PID;
+
+    return erts_port_task_schedule(prt->common.id,
+                                   NULL,
+                                   ERTS_PORT_TASK_PROC_SIG,
+                                   sigdp,
+                                   callback,
+                                   task_flags);
+}
+
 static ERTS_INLINE void
 send_badsig(Port *prt) {
     ErtsProcLocks rp_locks = ERTS_PROC_LOCKS_XSIG_SEND;
@@ -2370,6 +2400,11 @@ erts_port_exit(Process *c_p,
 		      | ERTS_PORT_SIG_FLG_BROKEN_LINK
 		      | ERTS_PORT_SIG_FLG_FORCE_SCHED)) == 0);
 
+#ifndef __WIN32__
+    if (prt->drv_ptr == &forker_driver)
+        return ERTS_PORT_OP_DROPPED;
+#endif
+
     if (!(flags & ERTS_PORT_SIG_FLG_FORCE_SCHED)) {
 	ErtsTryImmDrvCallState try_call_state
 	    = ERTS_INIT_TRY_IMM_DRV_CALL_STATE(c_p,
@@ -2734,6 +2769,72 @@ erts_port_link(Process *c_p, Port *prt, Eterm to, Eterm *refp)
 					  port_sig_link);
 }
 
+static void
+init_ack_send_reply(Port *port, Eterm resp)
+{
+
+    if (!is_internal_port(resp)) {
+        Process *rp = erts_proc_lookup_raw(port->async_open_port->to);
+        erts_smp_proc_lock(rp, ERTS_PROC_LOCK_LINK);
+        erts_remove_link(&ERTS_P_LINKS(port), port->async_open_port->to);
+        erts_remove_link(&ERTS_P_LINKS(rp), port->common.id);
+        erts_smp_proc_unlock(rp, ERTS_PROC_LOCK_LINK);
+    }
+    port_sched_op_reply(port->async_open_port->to,
+                        port->async_open_port->ref,
+                        resp);
+
+    erts_free(ERTS_ALC_T_PRTSD, port->async_open_port);
+    port->async_open_port = NULL;
+}
+
+void
+erl_drv_init_ack(ErlDrvPort ix, ErlDrvData res) {
+    Port *port = erts_drvport2port(ix);
+    SWord err_type = (SWord)res;
+    Eterm resp;
+
+    if (port == ERTS_INVALID_ERL_DRV_PORT && port->async_open_port)
+        return;
+
+    if (port->async_open_port) {
+        switch(err_type) {
+        case -3:
+            resp = am_badarg;
+            break;
+        case -2: {
+            char *str = erl_errno_id(errno);
+            resp = erts_atom_put((byte *) str, strlen(str),
+                                 ERTS_ATOM_ENC_LATIN1, 1);
+            break;
+        }
+        case -1:
+            resp = am_einval;
+            break;
+        default:
+            resp = port->common.id;
+            break;
+        }
+
+        init_ack_send_reply(port, resp);
+
+        if (err_type == -1 || err_type == -2 || err_type == -3)
+            driver_failure_term(ix, am_normal, 0);
+        port->drv_data = err_type;
+    }
+}
+
+void
+erl_drv_set_os_pid(ErlDrvPort ix, ErlDrvSInt pid) {
+    Port *port = erts_drvport2port(ix);
+
+    if (port == ERTS_INVALID_ERL_DRV_PORT)
+        return;
+
+    port->os_pid = (SWord)pid;
+
+}
+
 void erts_init_io(int port_tab_size,
 		  int port_tab_size_ignore_files,
 		  int legacy_port_tab)
@@ -2794,10 +2895,11 @@ void erts_init_io(int port_tab_size,
     erts_smp_rwmtx_rwlock(&erts_driver_list_lock);
 
     init_driver(&fd_driver, &fd_driver_entry, NULL);
-#ifndef __OSE__
     init_driver(&vanilla_driver, &vanilla_driver_entry, NULL);
-#endif
     init_driver(&spawn_driver, &spawn_driver_entry, NULL);
+#ifndef __WIN32__
+    init_driver(&forker_driver, &forker_driver_entry, NULL);
+#endif
     erts_init_static_drivers();
     for (dp = driver_tab; *dp != NULL; dp++)
 	erts_add_driver_entry(*dp, NULL, 1);
@@ -2859,6 +2961,9 @@ void erts_lcnt_enable_io_lock_count(int enable) {
 
     lcnt_enable_drv_lock_count(&vanilla_driver, enable);
     lcnt_enable_drv_lock_count(&spawn_driver, enable);
+#ifndef __WIN32__
+    lcnt_enable_drv_lock_count(&forker_driver, enable);
+#endif
     lcnt_enable_drv_lock_count(&fd_driver, enable);
     /* enable lock counting in all drivers */
     for (dp = driver_list; dp; dp = dp->next) {
@@ -3054,16 +3159,17 @@ deliver_result(Eterm sender, Eterm pid, Eterm res)
 
     if (rp) {
 	Eterm tuple;
-	ErlHeapFragment *bp;
+	ErtsMessage *mp;
 	ErlOffHeap *ohp;
 	Eterm* hp;
 	Uint sz_res;
 
 	sz_res = size_object(res);
-	hp = erts_alloc_message_heap(sz_res + 3, &bp, &ohp, rp, &rp_locks);
+	mp = erts_alloc_message_heap(rp, &rp_locks,
+				     sz_res + 3, &hp, &ohp);
 	res = copy_struct(res, sz_res, &hp, ohp);
 	tuple = TUPLE2(hp, sender, res);
-	erts_queue_message(rp, &rp_locks, bp, tuple, NIL);
+	erts_queue_message(rp, &rp_locks, mp, tuple, NIL);
 
 	if (rp_locks)
 	    erts_smp_proc_unlock(rp, rp_locks);
@@ -3091,7 +3197,7 @@ static void deliver_read_message(Port* prt, erts_aint32_t state, Eterm to,
     Eterm tuple;
     Process* rp;
     Eterm* hp;
-    ErlHeapFragment *bp;
+    ErtsMessage *mp;
     ErlOffHeap *ohp;
     ErtsProcLocks rp_locks = 0;
     int scheduler = erts_get_scheduler_id() != 0;
@@ -3117,7 +3223,7 @@ static void deliver_read_message(Port* prt, erts_aint32_t state, Eterm to,
     if (!rp)
 	return;
 
-    hp = erts_alloc_message_heap(need, &bp, &ohp, rp, &rp_locks);
+    mp = erts_alloc_message_heap(rp, &rp_locks, need, &hp, &ohp);
 
     listp = NIL;
     if ((state & ERTS_PORT_SFLG_BINARY_IO) == 0) {
@@ -3159,7 +3265,7 @@ static void deliver_read_message(Port* prt, erts_aint32_t state, Eterm to,
     tuple = TUPLE2(hp, prt->common.id, tuple);
     hp += 3;
 
-    erts_queue_message(rp, &rp_locks, bp, tuple, am_undefined);
+    erts_queue_message(rp, &rp_locks, mp, tuple, am_undefined);
     if (rp_locks)
 	erts_smp_proc_unlock(rp, rp_locks);
     if (!scheduler)
@@ -3233,7 +3339,7 @@ deliver_vec_message(Port* prt,			/* Port */
     Eterm tuple;
     Process* rp;
     Eterm* hp;
-    ErlHeapFragment *bp;
+    ErtsMessage *mp;
     ErlOffHeap *ohp;
     ErtsProcLocks rp_locks = 0;
     int scheduler = erts_get_scheduler_id() != 0;
@@ -3265,7 +3371,7 @@ deliver_vec_message(Port* prt,			/* Port */
 	need += (hlen+csize)*2;
     }
 
-    hp = erts_alloc_message_heap(need, &bp, &ohp, rp, &rp_locks);
+    mp = erts_alloc_message_heap(rp, &rp_locks, need, &hp, &ohp);
 
     listp = NIL;
     iov += vsize;
@@ -3326,7 +3432,7 @@ deliver_vec_message(Port* prt,			/* Port */
     tuple = TUPLE2(hp, prt->common.id, tuple);
     hp += 3;
 
-    erts_queue_message(rp, &rp_locks, bp, tuple, am_undefined);
+    erts_queue_message(rp, &rp_locks, mp, tuple, am_undefined);
     erts_smp_proc_unlock(rp, rp_locks);
     if (!scheduler)
 	erts_proc_dec_refc(rp);
@@ -3817,7 +3923,6 @@ write_port_control_result(int control_flags,
 			  ErlDrvSizeT resp_size,
 			  char *pre_alloc_buf,
 			  Eterm **hpp,
-			  ErlHeapFragment *bp,
 			  ErlOffHeap *ohp)
 {
     Eterm res;
@@ -3891,16 +3996,13 @@ port_sig_control(Port *prt,
 
 	if (res == ERTS_PORT_OP_DONE) {
 	    Eterm msg;
-	    Eterm *hp;
-	    ErlHeapFragment *bp;
-	    ErlOffHeap *ohp;
             ErtsHeapFactory factory;
 	    Process *rp;
 	    ErtsProcLocks rp_locks = 0;
 	    Uint hsz, rsz;
 	    int control_flags;
 
-	    rp = erts_proc_lookup_raw(sigdp->caller);
+	    rp = sigdp->caller == ERTS_INVALID_PID ? NULL : erts_proc_lookup_raw(sigdp->caller);
 	    if (!rp)
 		goto done;
 
@@ -3913,22 +4015,15 @@ port_sig_control(Port *prt,
 	    hsz = rsz + ERTS_QUEUE_PORT_SCHED_OP_REPLY_SIZE;
 
 
-	    hp = erts_alloc_message_heap(hsz,
-						    &bp,
-						    &ohp,
-						    rp,
-						    &rp_locks);
-            erts_factory_message_init(&factory, rp, hp, bp);
+	    (void) erts_factory_message_create(&factory, rp,
+					       &rp_locks, hsz);
 
 	    msg = write_port_control_result(control_flags,
 					    resp_bufp,
 					    resp_size,
 					    &resp_buf[0],
-					    &hp,
-					    bp,
-					    ohp);
-            factory.hp = hp;
-
+					    &factory.hp,
+					    factory.off_heap);
 	    queue_port_sched_op_reply(rp,
 				      &rp_locks,
                                       &factory,
@@ -3943,7 +4038,8 @@ port_sig_control(Port *prt,
 
     /* failure */
 
-    port_sched_op_reply(sigdp->caller, sigdp->ref, am_badarg);
+    if (sigdp->caller != ERTS_INVALID_PID)
+        port_sched_op_reply(sigdp->caller, sigdp->ref, am_badarg);
 
 done:
 
@@ -3953,6 +4049,23 @@ done:
     return ERTS_PORT_REDS_CONTROL;
 }
 
+/*
+ * This is an asynchronous control call. I.e. it will not return anything
+ * to the caller.
+ */
+int
+erl_drv_port_control(Eterm port_num, char cmd, char* buff, ErlDrvSizeT size)
+{
+    ErtsProc2PortSigData *sigdp = erts_port_task_alloc_p2p_sig_data();
+
+    sigdp->flags = ERTS_P2P_SIG_TYPE_CONTROL | ERTS_P2P_SIG_DATA_FLG_REPLY;
+    sigdp->u.control.binp = NULL;
+    sigdp->u.control.command = cmd;
+    sigdp->u.control.bufp = buff;
+    sigdp->u.control.size = size;
+
+    return erts_schedule_port2port_signal(port_num, sigdp, 0, port_sig_control);
+}
 
 ErtsPortOpResult
 erts_port_control(Process* c_p,
@@ -4069,7 +4182,6 @@ erts_port_control(Process* c_p,
 						 resp_size,
 						 &resp_buf[0],
 						 &hp,
-						 NULL,
 						 &c_p->off_heap);
 	    BUMP_REDS(c_p, ERTS_PORT_REDS_CONTROL);
 	    return ERTS_PORT_OP_DONE;
@@ -4091,10 +4203,10 @@ erts_port_control(Process* c_p,
     binp = NULL;
 
     if (is_binary(data) && binary_bitoffset(data) == 0) {
-	Eterm *ebinp = binary_val_rel(data, NULL);
+	Eterm *ebinp = binary_val(data);
 	ASSERT(!tmp_alloced);
 	if (*ebinp == HEADER_SUB_BIN)
-	    ebinp = binary_val_rel(((ErlSubBin *) ebinp)->orig, NULL);
+	    ebinp = binary_val(((ErlSubBin *) ebinp)->orig);
 	if (*ebinp != HEADER_PROC_BIN)
 	    copy = 1;
 	else {
@@ -4228,21 +4340,14 @@ port_sig_call(Port *prt,
 
 	    hsz = erts_decode_ext_size((byte *) resp_bufp, resp_size);
 	    if (hsz >= 0) {
-                ErlHeapFragment* bp;
-                ErlOffHeap* ohp;
                 ErtsHeapFactory factory;
 		byte *endp;
 
 		hsz += 3; /* ok tuple */
 		hsz += ERTS_QUEUE_PORT_SCHED_OP_REPLY_SIZE;
 
-                hp = erts_alloc_message_heap(hsz,
-                                             &bp,
-                                             &ohp,
-                                             rp,
-                                             &rp_locks);
+                (void) erts_factory_message_create(&factory, rp, &rp_locks, hsz);
 		endp = (byte *) resp_bufp;
-                erts_factory_message_init(&factory, rp, hp, bp);
 		msg = erts_decode_ext(&factory, &endp);
 		if (is_value(msg)) {
                     hp = erts_produce_heap(&factory,
@@ -4503,7 +4608,9 @@ port_sig_info(Port *prt,
 				    sigdp->u.info.item);
 	if (is_value(value)) {
             ErtsHeapFactory factory;
-            erts_factory_message_init(&factory, NULL, hp, bp);
+	    ErtsMessage *mp = erts_alloc_message(0, NULL);
+	    mp->data.heap_frag = bp;
+            erts_factory_selfcontained_message_init(&factory, mp, hp);
 	    queue_port_sched_op_reply(rp,
 				      &rp_locks,
                                       &factory,
@@ -4591,8 +4698,8 @@ reply_io_bytes(void *vreq)
 
     rp = erts_proc_lookup(req->pid);
     if (rp) {
-	ErlOffHeap *ohp = NULL;
-	ErlHeapFragment *bp = NULL;
+	ErlOffHeap *ohp;
+	ErtsMessage *mp;
 	ErtsProcLocks rp_locks;
 	Eterm ref, msg, ein, eout, *hp;
 	Uint64 in, out;
@@ -4614,7 +4721,7 @@ reply_io_bytes(void *vreq)
 	erts_bld_uint64(NULL, &hsz, in);
 	erts_bld_uint64(NULL, &hsz, out);
 
-	hp = erts_alloc_message_heap(hsz, &bp, &ohp, rp, &rp_locks);
+	mp = erts_alloc_message_heap(rp, &rp_locks, hsz, &hp, &ohp);
 
 	ref = make_internal_ref(hp);
 	write_ref_thing(hp, req->refn[0], req->refn[1], req->refn[2]);
@@ -4624,7 +4731,7 @@ reply_io_bytes(void *vreq)
 	eout = erts_bld_uint64(&hp, NULL, out);
 
 	msg = TUPLE4(hp, ref, make_small(sched_id), ein, eout);
-	erts_queue_message(rp, &rp_locks, bp, msg, NIL);
+	erts_queue_message(rp, &rp_locks, mp, msg, NIL);
 
 	if (req->sched_id == sched_id)
 	    rp_locks &= ~ERTS_PROC_LOCK_MAIN;
@@ -4730,6 +4837,10 @@ print_port_info(Port *p, int to, void *arg)
 	erts_print(to, arg, "Port is a file: %s\n",p->name);
     } else if (p->drv_ptr == &spawn_driver) {
 	erts_print(to, arg, "Port controls external process: %s\n",p->name);
+#ifndef __WIN32__
+    } else if (p->drv_ptr == &forker_driver) {
+	erts_print(to, arg, "Port controls forker process: %s\n",p->name);
+#endif
     } else {
 	erts_print(to, arg, "Port controls linked-in driver: %s\n",p->name);
     }
@@ -5069,11 +5180,11 @@ ErlDrvTermData driver_mk_term_nil(void)
 void driver_report_exit(ErlDrvPort ix, int status)
 {
    Eterm* hp;
+   ErlOffHeap *ohp;
    Eterm tuple;
    Process *rp;
    Eterm pid;
-   ErlHeapFragment *bp = NULL;
-   ErlOffHeap *ohp;
+   ErtsMessage *mp;
    ErtsProcLocks rp_locks = 0;
    int scheduler = erts_get_scheduler_id() != 0;
    Port* prt = erts_drvport2port(ix);
@@ -5093,13 +5204,13 @@ void driver_report_exit(ErlDrvPort ix, int status)
    if (!rp)
        return;
 
-   hp = erts_alloc_message_heap(3+3, &bp, &ohp, rp, &rp_locks);
+   mp = erts_alloc_message_heap(rp, &rp_locks, 3+3, &hp, &ohp);
 
    tuple = TUPLE2(hp, am_exit_status, make_small(status));
    hp += 3;
    tuple = TUPLE2(hp, prt->common.id, tuple);
 
-   erts_queue_message(rp, &rp_locks, bp, tuple, am_undefined);
+   erts_queue_message(rp, &rp_locks, mp, tuple, am_undefined);
 
    erts_smp_proc_unlock(rp, rp_locks);
    if (!scheduler)
@@ -5209,7 +5320,6 @@ driver_deliver_term(Eterm to, ErlDrvTermData* data, int len)
     ErtsProcLocks rp_locks = 0;
     struct b2t_states__ b2t;
     int scheduler;
-    int is_heap_need_limited = 1;
     ErtsSchedulerData *esdp = erts_get_scheduler_data();
 
     ERTS_UNDEF(mess,NIL);
@@ -5248,25 +5358,17 @@ driver_deliver_term(Eterm to, ErlDrvTermData* data, int len)
 	    break;
 	case ERL_DRV_INT:  /* signed int argument */
 	    ERTS_DDT_CHK_ENOUGH_ARGS(1);
-#if HALFWORD_HEAP
-	    erts_bld_sint64(NULL, &need, (Sint64)ptr[0]);
-#else
 	    /* check for bignum */
 	    if (!IS_SSMALL((Sint)ptr[0]))
 		need += BIG_UINT_HEAP_SIZE;  /* use small_to_big */
-#endif
 	    ptr++;
 	    depth++;
 	    break;
 	case ERL_DRV_UINT:  /* unsigned int argument */
 	    ERTS_DDT_CHK_ENOUGH_ARGS(1);
-#if HALFWORD_HEAP
-	    erts_bld_uint64(NULL, &need, (Uint64)ptr[0]);
-#else
 	    /* check for bignum */
 	    if (!IS_USMALL(0, (Uint)ptr[0]))
 		need += BIG_UINT_HEAP_SIZE;  /* use small_to_big */
-#endif
 	    ptr++;
 	    depth++;
 	    break;
@@ -5386,9 +5488,6 @@ driver_deliver_term(Eterm to, ErlDrvTermData* data, int len)
 	    need += hsz;
 	    ptr += 2;
 	    depth++;
-            if (size > MAP_SMALL_MAP_LIMIT*3) { /* may contain big map */
-                is_heap_need_limited = 0;
-            }
 	    break;
 	}
 	case ERL_DRV_MAP: { /* int */
@@ -5396,7 +5495,6 @@ driver_deliver_term(Eterm to, ErlDrvTermData* data, int len)
 	    if ((int) ptr[0] < 0) ERTS_DDT_FAIL;
             if (ptr[0] > MAP_SMALL_MAP_LIMIT) {
                 need += HASHMAP_ESTIMATED_HEAP_SIZE(ptr[0]);
-                is_heap_need_limited = 0;
             } else {
                 need += MAP_HEADER_FLATMAP_SZ + 1 + 2*ptr[0];
             }
@@ -5435,17 +5533,7 @@ driver_deliver_term(Eterm to, ErlDrvTermData* data, int len)
 	goto done;
     }
 
-    /* Try copy directly to destination heap if we know there are no big maps */
-    if (is_heap_need_limited) {
-        ErlOffHeap *ohp;
-        ErlHeapFragment* bp;
-        Eterm* hp = erts_alloc_message_heap(need, &bp, &ohp, rp, &rp_locks);
-        erts_factory_message_init(&factory, rp, hp, bp);
-    }
-    else {
-        erts_factory_message_init(&factory, NULL, NULL,
-                                  new_message_buffer(need));
-    }
+    (void) erts_factory_message_create(&factory, rp, &rp_locks, need);
 
     /*
      * Interpret the instructions and build the term.
@@ -5465,10 +5553,6 @@ driver_deliver_term(Eterm to, ErlDrvTermData* data, int len)
 	    break;
 
 	case ERL_DRV_INT:  /* signed int argument */
-#if HALFWORD_HEAP
-	    erts_reserve_heap(&factory, BIG_NEED_SIZE(2));
-	    mess = erts_bld_sint64(&factory.hp, NULL, (Sint64)ptr[0]);
-#else
 	    erts_reserve_heap(&factory, BIG_UINT_HEAP_SIZE);
 	    if (IS_SSMALL((Sint)ptr[0]))
 		mess = make_small((Sint)ptr[0]);
@@ -5476,15 +5560,10 @@ driver_deliver_term(Eterm to, ErlDrvTermData* data, int len)
 		mess = small_to_big((Sint)ptr[0], factory.hp);
 		factory.hp += BIG_UINT_HEAP_SIZE;
 	    }
-#endif
 	    ptr++;
 	    break;
 
 	case ERL_DRV_UINT:  /* unsigned int argument */
-#if HALFWORD_HEAP
-	    erts_reserve_heap(&factory, BIG_NEED_FOR_BITS(64));
-	    mess = erts_bld_uint64(&factory.hp, NULL, (Uint64)ptr[0]);
-#else
 	    erts_reserve_heap(&factory, BIG_UINT_HEAP_SIZE);
 	    if (IS_USMALL(0, (Uint)ptr[0]))
 		mess = make_small((Uint)ptr[0]);
@@ -5492,7 +5571,6 @@ driver_deliver_term(Eterm to, ErlDrvTermData* data, int len)
 		mess = uint_to_big((Uint)ptr[0], factory.hp);
 		factory.hp += BIG_UINT_HEAP_SIZE;
 	    }
-#endif
 	    ptr++;
 	    break;
 
@@ -5724,9 +5802,9 @@ driver_deliver_term(Eterm to, ErlDrvTermData* data, int len)
 
     if (res > 0) {
 	mess = ESTACK_POP(stack);  /* get resulting value */
-	erts_factory_close(&factory);
+	erts_factory_trim_and_close(&factory, &mess, 1);
 	/* send message */
-	erts_queue_message(rp, &rp_locks, factory.heap_frags, mess, am_undefined);
+	erts_queue_message(rp, &rp_locks, factory.message, mess, am_undefined);
     }
     else {
 	if (b2t.ix > b2t.used)
@@ -6809,7 +6887,7 @@ int driver_monitor_process(ErlDrvPort drvport,
 {
     Port *prt;
     int ret;
-#if !HEAP_ON_C_STACK || (defined(ERTS_SMP) && defined(ERTS_ENABLE_LOCK_CHECK))
+#if defined(ERTS_SMP) && defined(ERTS_ENABLE_LOCK_CHECK)
     ErtsSchedulerData *sched = erts_get_scheduler_data();
 #endif
 
@@ -6820,16 +6898,6 @@ int driver_monitor_process(ErlDrvPort drvport,
     /* Now (in SMP) we should have either the port lock (if we have a scheduler) or the port data lock
        (if we're a driver thread) */
     ERTS_SMP_LC_ASSERT((sched != NULL || prt->port_data_lock));
-
-#if !HEAP_ON_C_STACK
-    if (!sched) {
-	/* Need a separate allocation for the ref :( */
-	Eterm *buf = erts_alloc(ERTS_ALC_T_TEMP_TERM,
-				sizeof(Eterm)*REF_THING_SIZE);
-	ret = do_driver_monitor_process(prt,buf,process,monitor);
-	erts_free(ERTS_ALC_T_TEMP_TERM,buf);
-    } else
-#endif
     {
 	DeclareTmpHeapNoproc(buf,REF_THING_SIZE);
 	UseTmpHeapNoproc(REF_THING_SIZE);
@@ -6882,7 +6950,7 @@ int driver_demonitor_process(ErlDrvPort drvport,
 {
     Port *prt;
     int ret;
-#if !HEAP_ON_C_STACK || (defined(ERTS_SMP) && defined(ERTS_ENABLE_LOCK_CHECK))
+#if defined(ERTS_SMP) && defined(ERTS_ENABLE_LOCK_CHECK)
     ErtsSchedulerData *sched = erts_get_scheduler_data();
 #endif
 
@@ -6893,15 +6961,6 @@ int driver_demonitor_process(ErlDrvPort drvport,
     /* Now we should have either the port lock (if we have a scheduler) or the port data lock
        (if we're a driver thread) */
     ERTS_SMP_LC_ASSERT((sched != NULL || prt->port_data_lock));
-#if !HEAP_ON_C_STACK
-    if (!sched) {
-	/* Need a separate allocation for the ref :( */
-	Eterm *buf = erts_alloc(ERTS_ALC_T_TEMP_TERM,
-				sizeof(Eterm)*REF_THING_SIZE);
-	ret = do_driver_demonitor_process(prt,buf,monitor);
-	erts_free(ERTS_ALC_T_TEMP_TERM,buf);
-    } else
-#endif
     {
 	DeclareTmpHeapNoproc(buf,REF_THING_SIZE);
 	UseTmpHeapNoproc(REF_THING_SIZE);
@@ -6937,7 +6996,7 @@ ErlDrvTermData driver_get_monitored_process(ErlDrvPort drvport,
 {
     Port *prt;
     ErlDrvTermData ret;
-#if !HEAP_ON_C_STACK || (defined(ERTS_SMP) && defined(ERTS_ENABLE_LOCK_CHECK))
+#if defined(ERTS_SMP) && defined(ERTS_ENABLE_LOCK_CHECK)
     ErtsSchedulerData *sched = erts_get_scheduler_data();
 #endif
 
@@ -6948,16 +7007,6 @@ ErlDrvTermData driver_get_monitored_process(ErlDrvPort drvport,
     /* Now we should have either the port lock (if we have a scheduler) or the port data lock
        (if we're a driver thread) */
     ERTS_SMP_LC_ASSERT((sched != NULL || prt->port_data_lock));
-
-#if !HEAP_ON_C_STACK
-    if (!sched) {
-	/* Need a separate allocation for the ref :( */
-	Eterm *buf = erts_alloc(ERTS_ALC_T_TEMP_TERM,
-				sizeof(Eterm)*REF_THING_SIZE);
-	ret = do_driver_get_monitored_process(prt,buf,monitor);
-	erts_free(ERTS_ALC_T_TEMP_TERM,buf);
-    } else
-#endif
     {
 	DeclareTmpHeapNoproc(buf,REF_THING_SIZE);
 	UseTmpHeapNoproc(REF_THING_SIZE);
@@ -7023,6 +7072,9 @@ driver_failure_term(ErlDrvPort ix, Eterm term, int eof)
     if (prt == ERTS_INVALID_ERL_DRV_PORT)
 	return -1;
     ERTS_SMP_LC_ASSERT(erts_lc_is_port_locked(prt));
+
+    if (prt->async_open_port)
+        init_ack_send_reply(prt, prt->common.id);
     if (eof)
 	flush_linebuf_messages(prt, state);
     if (state & ERTS_PORT_SFLG_CLOSING) {
@@ -7596,15 +7648,15 @@ int null_func(void)
 }
 
 int
-erl_drv_putenv(char *key, char *value)
+erl_drv_putenv(const char *key, char *value)
 {
-    return erts_sys_putenv_raw(key, value);
+    return erts_sys_putenv_raw((char*)key, value);
 }
 
 int
-erl_drv_getenv(char *key, char *value, size_t *value_size)
+erl_drv_getenv(const char *key, char *value, size_t *value_size)
 {
-    return erts_sys_getenv_raw(key, value, value_size);
+    return erts_sys_getenv_raw((char*)key, value, value_size);
 }
 
 /* get heart_port
diff --git a/erts/emulator/beam/module.c b/erts/emulator/beam/module.c
index 86dd3b5aac..f5c7b177d3 100644
--- a/erts/emulator/beam/module.c
+++ b/erts/emulator/beam/module.c
@@ -74,8 +74,8 @@ static Module* module_alloc(Module* tmpl)
     erts_smp_atomic_add_nob(&tot_module_bytes, sizeof(Module));
 
     obj->module = tmpl->module;
-    obj->curr.code = 0;
-    obj->old.code = 0;
+    obj->curr.code_hdr = 0;
+    obj->old.code_hdr = 0;
     obj->curr.code_length = 0;
     obj->old.code_length = 0;
     obj->slot.index = -1;
@@ -103,6 +103,9 @@ void init_module_table(void)
     f.cmp  = (HCMP_FUN) module_cmp;
     f.alloc = (HALLOC_FUN) module_alloc;
     f.free = (HFREE_FUN) module_free;
+    f.meta_alloc = (HMALLOC_FUN) erts_alloc;
+    f.meta_free = (HMFREE_FUN) erts_free;
+    f.meta_print = (HMPRINT_FUN) erts_print;
 
     for (i = 0; i < ERTS_NUM_CODE_IX; i++) {
 	erts_index_init(ERTS_ALC_T_MODULE_TABLE, &module_tables[i], "module_code",
diff --git a/erts/emulator/beam/module.h b/erts/emulator/beam/module.h
index c8a6351b04..e66d628ca9 100644
--- a/erts/emulator/beam/module.h
+++ b/erts/emulator/beam/module.h
@@ -26,7 +26,7 @@
 #endif
 
 struct erl_module_instance {
-    BeamInstr* code;
+    BeamCodeHeader* code_hdr;
     int code_length;		/* Length of loaded code in bytes. */
     unsigned catches;
     struct erl_module_nif* nif;
diff --git a/erts/emulator/beam/ops.tab b/erts/emulator/beam/ops.tab
index 1d32e72247..081c4108a0 100644
--- a/erts/emulator/beam/ops.tab
+++ b/erts/emulator/beam/ops.tab
@@ -39,8 +39,8 @@ too_old_compiler | never() =>
 # necessary.) Since the instructions don't work correctly in R12B, simply
 # refuse to load the module.
 
-func_info M=a a==am_module_info A=u==0 | label L | move n r => too_old_compiler
-func_info M=a a==am_module_info A=u==1 | label L | move n r => too_old_compiler
+func_info M=a a==am_module_info A=u==0 | label L | move n x==0 => too_old_compiler
+func_info M=a a==am_module_info A=u==1 | label L | move n x==0 => too_old_compiler
 
 # The undocumented and unsupported guard BIF is_constant/1 was removed
 # in R13. The is_constant/2 operation is marked as obsolete in genop.tab,
@@ -76,17 +76,6 @@ return
 # with the following call instruction, we need to make sure that
 # there is no line/1 instruction between the move and the call.
 #
-
-move S r | line Loc | call_ext Ar Func => \
-     line Loc | move S r | call_ext Ar Func
-move S r | line Loc | call_ext_last Ar Func=u$is_bif D => \
-     line Loc | move S r | call_ext_last Ar Func D
-move S r | line Loc | call_ext_only Ar Func=u$is_bif => \
-     line Loc | move S r | call_ext_only Ar Func
-move S r | line Loc | call Ar Func => \
-     line Loc | move S r | call Ar Func
-
-#
 # A tail-recursive call to an external function (non-BIF) will
 # never be saved on the stack, so there is no reason to keep
 # the line instruction. (The compiler did not remove the line
@@ -94,10 +83,14 @@ move S r | line Loc | call Ar Func => \
 # BIFs and ordinary Erlang functions.)
 #
 
-line Loc | call_ext_last Ar Func=u$is_not_bif D => \
-     call_ext_last Ar Func D
-line Loc | call_ext_only Ar Func=u$is_not_bif => \
-     call_ext_only Ar Func
+move S X0=x==0 | line Loc | call_ext Ar Func => \
+     line Loc | move S X0 | call_ext Ar Func
+move S X0=x==0 | line Loc | call_ext_last Ar Func=u$is_not_bif D => \
+     move S X0 | call_ext_last Ar Func D
+move S X0=x==0 | line Loc | call_ext_only Ar Func=u$is_not_bif => \
+     move S X0 | call_ext_only Ar Func
+move S X0=x==0 | line Loc | call Ar Func => \
+     line Loc | move S X0 | call Ar Func
 
 line Loc | func_info M F A => func_info M F A | line Loc
 
@@ -167,31 +160,24 @@ is_tuple Fail=f S | select_tuple_arity S=d Fail=f Size=u Rest=* => \
 select_tuple_arity S=d Fail=f Size=u Rest=* => \
   gen_select_tuple_arity(S, Fail, Size, Rest)
 
-i_select_val_bins r f I
 i_select_val_bins x f I
 i_select_val_bins y f I
 
-i_select_val_lins r f I
 i_select_val_lins x f I
 i_select_val_lins y f I
 
-i_select_val2 r f c c f f
 i_select_val2 x f c c f f
 i_select_val2 y f c c f f
 
-i_select_tuple_arity r f I
 i_select_tuple_arity x f I
 i_select_tuple_arity y f I
 
-i_select_tuple_arity2 r f A A f f
 i_select_tuple_arity2 x f A A f f
 i_select_tuple_arity2 y f A A f f
 
-i_jump_on_val_zero r f I
 i_jump_on_val_zero x f I
 i_jump_on_val_zero y f I
 
-i_jump_on_val r f I I
 i_jump_on_val x f I I
 i_jump_on_val y f I I
 
@@ -203,30 +189,21 @@ is_ne_exact L1 S1 S2 | jump Fail | label L2 | same_label(L1, L2) => \
 %macro: get_list GetList -pack
 get_list x x x
 get_list x x y
-get_list x x r
 get_list x y x
 get_list x y y
-get_list x y r
-get_list x r x
-get_list x r y
 
 get_list y x x
 get_list y x y
-get_list y x r
 get_list y y x
 get_list y y y
-get_list y y r
-get_list y r x
-get_list y r y
 
+# The following get_list instructions using x(0) are frequently used.
 get_list r x x
+get_list r r y
+get_list x r x
 get_list r x y
-get_list r x r
-get_list r y x
-get_list r y y
 get_list r y r
-get_list r r x
-get_list r r y
+get_list r x r
 
 # Old-style catch.
 catch y f
@@ -237,33 +214,31 @@ try Y F => catch Y F
 try_case Y => try_end Y
 try_end y
 
-try_case_end Literal=q => move Literal x | try_case_end x
 try_case_end s
 
 # Destructive set tuple element
 
-set_tuple_element Lit=q Tuple Pos => move Lit x | set_tuple_element x Tuple Pos
 set_tuple_element s d P
 
 # Get tuple element
 
 %macro: i_get_tuple_element GetTupleElement -pack
 i_get_tuple_element x P x
-i_get_tuple_element r P x
 i_get_tuple_element y P x
-i_get_tuple_element x P r
-i_get_tuple_element y P r
 
 %cold
-i_get_tuple_element r P r
 i_get_tuple_element x P y
-i_get_tuple_element r P y
 i_get_tuple_element y P y
 %hot
 
+%macro: i_get_tuple_element2 GetTupleElement2 -pack
+i_get_tuple_element2 x P x
+
+%macro: i_get_tuple_element3 GetTupleElement3 -pack
+i_get_tuple_element3 x P x
+
 %macro: is_number IsNumber -fail_action
 %cold
-is_number f r
 is_number f x
 is_number f y
 %hot
@@ -273,16 +248,12 @@ is_number Fail Literal=q => move Literal x | is_number Fail x
 
 jump f
 
-case_end Literal=cq => move Literal x | case_end x
-badmatch Literal=cq => move Literal x | badmatch x
+case_end NotInX=cy => move NotInX x | case_end x
+badmatch NotInX=cy => move NotInX x | badmatch x
 
-case_end r
 case_end x
-case_end y
 
-badmatch r
 badmatch x
-badmatch y
 
 if_end
 raise s s
@@ -291,7 +262,7 @@ raise s s
 badarg j
 system_limit j
 
-move C=cxy r | jump Lbl => move_jump Lbl C
+move C=cxy x==0 | jump Lbl => move_jump Lbl C
 
 %macro: move_jump MoveJump -nonext
 move_jump f n
@@ -309,8 +280,6 @@ move_window/6
 
 # x -> y
 
-move S1=r S2=y | move X1=x Y1=y => move2 S1 S2 X1 Y1
-
 move X1=x Y1=y | move X2=x Y2=y | move X3=x Y3=y | succ(Y1,Y2) | succ(Y2,Y3) => \
     move_window X1 X2 X3 Y1 Y3
 
@@ -323,24 +292,76 @@ move_window X1=x X2=x X3=x X4=x Y1=y Y4=y | move X5=x Y5=y | succ(Y4,Y5) => \
 move_window X1=x X2=x X3=x Y1=y Y3=y => move_window3 X1 X2 X3 Y1
 move_window X1=x X2=x X3=x X4=x Y1=y Y4=y => move_window4 X1 X2 X3 X4 Y1
 
+%macro: move_window3 MoveWindow3 -pack
+%macro: move_window4 MoveWindow4 -pack
+%macro: move_window5 MoveWindow5 -pack
+
 move_window3 x x x y
 move_window4 x x x x y
 move_window5 x x x x x y
 
-move X1=x Y1=y | move X2=x Y2=y => move2 X1 Y1 X2 Y2
-move Y1=y X1=x | move Y2=y X2=x => move2 Y1 X1 Y2 X2
-move X1=x X2=x | move X3=x X4=x => move2 X1 X2 X3 X4
+# Swap registers.
+move R1=x Tmp=x | move R2=xy R1 | move Tmp R2 => swap_temp R1 R2 Tmp
+
+swap_temp R1 R2 Tmp | line Loc | apply Live | is_killed(Tmp, Live) => \
+  swap R1 R2 | line Loc | apply Live
+
+swap_temp R1 R2 Tmp | line Loc | call Live Addr | is_killed(Tmp, Live) => \
+  swap R1 R2 | line Loc | call Live Addr
+swap_temp R1 R2 Tmp | call_only Live Addr | \
+  is_killed(Tmp, Live) => swap R1 R2 | call_only Live Addr
+swap_temp R1 R2 Tmp | call_last Live Addr D | \
+  is_killed(Tmp, Live) => swap R1 R2 | call_last Live Addr D
 
-move S1=x S2=r | move S3=x S4=x => move2 S1 S2 S3 S4
-move S1=x S2=r | move X1=x Y1=y => move2 S1 S2 X1 Y1
-move S1=y S2=r | move X1=x Y1=y => move2 S1 S2 X1 Y1
+swap_temp R1 R2 Tmp | line Loc | call_ext Live Addr | is_killed(Tmp, Live) => \
+  swap R1 R2 | line Loc | call_ext Live Addr
+swap_temp R1 R2 Tmp | line Loc | call_ext_only Live Addr | \
+  is_killed(Tmp, Live) => swap R1 R2 | line Loc | call_ext_only Live Addr
+swap_temp R1 R2 Tmp | line Loc | call_ext_last Live Addr D | \
+  is_killed(Tmp, Live) => swap R1 R2 | line Loc | call_ext_last Live Addr D
 
-move Y1=y X1=x | move S1=r D1=x => move2 Y1 X1 S1 D1
-move S1=r D1=x | move Y1=y X1=x => move2 S1 D1 Y1 X1
+%macro: swap_temp SwapTemp -pack
+swap_temp x x x
+swap_temp x y x
 
-move2 X1=x Y1=y X2=x Y2=y | move X3=x Y3=y => move3 X1 Y1 X2 Y2 X3 Y3
-move2 Y1=y X1=x Y2=y X2=x | move Y3=y X3=x => move3 Y1 X1 Y2 X2 Y3 X3
-move2 X1=x X2=x X3=x X4=x | move X5=x X6=x => move3 X1 X2 X3 X4 X5 X6
+%macro: swap Swap -pack
+swap x x
+swap x y
+
+move Src=x D1=x | move Src=x D2=x => move_dup Src D1 D2
+move Src=x SD=x | move SD=x D=x   => move_dup Src SD D
+move Src=x D1=x | move Src=x D2=y => move_dup Src D1 D2
+move Src=y SD=x | move SD=x D=y   => move_dup Src SD D
+move Src=x SD=x | move SD=x D=y   => move_dup Src SD D
+move Src=y SD=x | move SD=x D=x   => move_dup Src SD D
+
+move SD=x D=x | move Src=xy SD=x => move_shift Src SD D
+move SD=y D=x | move Src=x  SD=y => move_shift Src SD D
+move SD=x D=y | move Src=x  SD=x => move_shift Src SD D
+
+# The transformations above guarantee that the source for
+# the second move is not the same as the destination for
+# the first move. That means that we can do the moves in
+# parallel (fetch both values, then store them) which could
+# be faster.
+
+move X1=x Y1=y | move X2=x Y2=y => move2_par X1 Y1 X2 Y2
+move Y1=y X1=x | move Y2=y X2=x => move2_par Y1 X1 Y2 X2
+
+move X1=x X2=x | move X3=x X4=x => move2_par X1 X2 X3 X4
+
+move X1=x X2=x | move X3=x Y1=y => move2_par X1 X2 X3 Y1
+
+move S1=x S2=x | move X1=x Y1=y => move2_par S1 S2 X1 Y1
+
+move S1=y S2=x | move X1=x Y1=y => move2_par S1 S2 X1 Y1
+
+move Y1=y X1=x | move S1=x D1=x => move2_par Y1 X1 S1 D1
+move S1=x D1=x | move Y1=y X1=x => move2_par S1 D1 Y1 X1
+
+move2_par X1=x Y1=y X2=x Y2=y | move X3=x Y3=y => move3 X1 Y1 X2 Y2 X3 Y3
+move2_par Y1=y X1=x Y2=y X2=x | move Y3=y X3=x => move3 Y1 X1 Y2 X2 Y3 X3
+move2_par X1=x X2=x X3=x X4=x | move X5=x X6=x => move3 X1 X2 X3 X4 X5 X6
 
 move C=aiq X=x==1 => move_x1 C
 move C=aiq X=x==2 => move_x2 C
@@ -348,21 +369,32 @@ move C=aiq X=x==2 => move_x2 C
 move_x1 c
 move_x2 c
 
-%macro: move2 Move2 -pack
-move2 x y x y
-move2 y x y x
-move2 x x x x
+%macro: move_shift MoveShift -pack
+move_shift x x x
+move_shift y x x
+move_shift x y x
+move_shift x x y
+
+%macro: move_dup MoveDup -pack
+move_dup x x x
+move_dup x x y
+move_dup y x x
+move_dup y x y
 
-move2 x r x x
+%macro: move2_par Move2Par -pack
 
-move2 x r x y
-move2 r y x y
-move2 y r x y
+move2_par x y x y
+move2_par y x y x
+move2_par x x x x
 
-move2 r x y x
-move2 y x r x
+move2_par x x x y
 
-%macro: move3 Move3
+move2_par y x x y
+
+move2_par x x y x
+move2_par y x x x
+
+%macro: move3 Move3 -pack
 move3 x y x y x y
 move3 y x y x y x
 move3 x x x x x x
@@ -375,20 +407,22 @@ move S=c D=y => move S x | move x D
 %macro:move Move -pack -gen_dest
 move x x
 move x y
-move x r
 move y x
-move y r
-move r x
-move r y
-move c r
 move c x
 move n x
-move n r
 move y y
 
+# The following move instructions using x(0) are frequently used.
+
+move x r
+move r x
+move y r
+move c r
+move r y
+
 # Receive operations.
 
-loop_rec Fail Src | smp_mark_target_label(Fail) => i_loop_rec Fail Src
+loop_rec Fail x==0 | smp_mark_target_label(Fail) => i_loop_rec Fail
 
 label L | wait_timeout Fail Src | smp_already_locked(L) => label L | i_wait_timeout_locked Fail Src
 wait_timeout Fail Src => i_wait_timeout Fail Src
@@ -403,7 +437,7 @@ label L | timeout | smp_already_locked(L) => label L | timeout_locked
 remove_message
 timeout
 timeout_locked
-i_loop_rec f r
+i_loop_rec f
 loop_rec_end f
 wait f
 wait_locked f
@@ -421,93 +455,57 @@ send
 # Optimized comparisons with one immediate/literal operand.
 #
 
-is_eq_exact Lbl R=rxy C=ian => i_is_eq_exact_immed Lbl R C
-is_eq_exact Lbl R=rxy C=q => i_is_eq_exact_literal R Lbl C
+is_eq_exact Lbl R=xy C=ian => i_is_eq_exact_immed Lbl R C
+is_eq_exact Lbl R=xy C=q => i_is_eq_exact_literal Lbl R C
 
-is_ne_exact Lbl R=rxy C=ian => i_is_ne_exact_immed Lbl R C
-is_ne_exact Lbl R=rxy C=q => i_is_ne_exact_literal R Lbl C
+is_ne_exact Lbl R=xy C=ian => i_is_ne_exact_immed Lbl R C
+is_ne_exact Lbl R=xy C=q => i_is_ne_exact_literal Lbl R C
 
 %macro: i_is_eq_exact_immed EqualImmed -fail_action
 i_is_eq_exact_immed f r c
 i_is_eq_exact_immed f x c
 i_is_eq_exact_immed f y c
 
-i_is_eq_exact_literal r f c
-i_is_eq_exact_literal x f c
-i_is_eq_exact_literal y f c
+i_is_eq_exact_literal f x c
+i_is_eq_exact_literal f y c
 
 %macro: i_is_ne_exact_immed NotEqualImmed -fail_action
-i_is_ne_exact_immed f r c
 i_is_ne_exact_immed f x c
 i_is_ne_exact_immed f y c
 
-i_is_ne_exact_literal r f c
-i_is_ne_exact_literal x f c
-i_is_ne_exact_literal y f c
-
-#
-# Common Compare Specializations
-# We don't do all of them since we want
-# to keep the instruction set small-ish
-#
-
-is_eq_exact Lbl S1=xy S2=r => is_eq_exact Lbl S2 S1
-is_eq_exact Lbl S1=rx S2=xy => i_is_eq_exact_spec Lbl S1 S2
-%macro: i_is_eq_exact_spec EqualExact -fail_action
-
-i_is_eq_exact_spec f x x
-i_is_eq_exact_spec f x y
-i_is_eq_exact_spec f r x
-i_is_eq_exact_spec f r y
-%cold
-i_is_eq_exact_spec f r r
-%hot
+i_is_ne_exact_literal f x c
+i_is_ne_exact_literal f y c
 
-is_lt Lbl S1=rxc S2=rxc  => i_is_lt_spec Lbl S1 S2
+is_eq_exact Lbl Y=y X=x => is_eq_exact Lbl X Y
+%macro: is_eq_exact EqualExact -fail_action -pack
+is_eq_exact f x x
+is_eq_exact f x y
+is_eq_exact f s s
 
-%macro: i_is_lt_spec IsLessThan -fail_action
-
-i_is_lt_spec f x x
-i_is_lt_spec f x r
-i_is_lt_spec f x c
-i_is_lt_spec f r x
-i_is_lt_spec f r c
-i_is_lt_spec f c x
-i_is_lt_spec f c r
+%macro: is_lt IsLessThan -fail_action
+is_lt f x x
+is_lt f x c
+is_lt f c x
 %cold
-i_is_lt_spec f r r
-i_is_lt_spec f c c
+is_lt f s s
 %hot
 
-is_ge Lbl S1=xc S2=xc => i_is_ge_spec Lbl S1 S2
-
-%macro: i_is_ge_spec IsGreaterEqual -fail_action
-
-i_is_ge_spec f x x
-i_is_ge_spec f x c
-i_is_ge_spec f c x
+%macro: is_ge IsGreaterEqual -fail_action
+is_ge f x x
+is_ge f x c
+is_ge f c x
 %cold
-i_is_ge_spec f c c
+is_ge f s s
 %hot
 
-#
-# All other comparisons.
-#
+%macro: is_ne_exact NotEqualExact -fail_action
+is_ne_exact f s s
 
-is_eq_exact Lbl S1 S2 => i_fetch S1 S2 | i_is_eq_exact Lbl
-is_ne_exact Lbl S1 S2 => i_fetch S1 S2 | i_is_ne_exact Lbl
+%macro: is_eq Equal -fail_action
+is_eq f s s
 
-is_lt Lbl S1 S2 => i_fetch S1 S2 | i_is_lt Lbl
-is_ge Lbl S1 S2 => i_fetch S1 S2 | i_is_ge Lbl
-is_eq Lbl S1 S2 => i_fetch S1 S2 | i_is_eq Lbl
-is_ne Lbl S1 S2 => i_fetch S1 S2 | i_is_ne Lbl
-
-i_is_eq_exact f
-i_is_ne_exact f
-i_is_lt f
-i_is_ge f
-i_is_eq f
-i_is_ne f
+%macro: is_ne NotEqual -fail_action
+is_ne f s s
 
 #
 # Putting things.
@@ -525,7 +523,6 @@ i_put_tuple Dst Arity Puts=* | put S => \
 i_put_tuple/2
 
 %macro:i_put_tuple PutTuple -pack -goto:do_put_tuple
-i_put_tuple r I
 i_put_tuple x I
 i_put_tuple y I
 
@@ -542,74 +539,36 @@ put_list x n x
 put_list y n x
 put_list x x x
 put_list y x x
-put_list x x r
-put_list y r r
 
 put_list y y x
 put_list x y x
-put_list r x x
-put_list r y x
-put_list r x r
-put_list y y r
-put_list y r x
-put_list r n x
 
-put_list x r x
-put_list x y r
-put_list y x r
 put_list y x x
 
-put_list x r r
-
 # put_list SrcReg Constant Dst
-put_list r c r
-put_list r c x
-put_list r c y
 
-put_list x c r
 put_list x c x
 put_list x c y
 
-put_list y c r
 put_list y c x
-put_list y c y
 
 # put_list Constant SrcReg Dst
-put_list c r r
-put_list c r x
-put_list c r y
 
-put_list c x r
 put_list c x x
-put_list c x y
-
-put_list c y r
 put_list c y x
-put_list c y y
 
-%cold
-put_list s s d
-%hot
+# The following put_list instructions using x(0) are frequently used.
 
-%macro: i_fetch FetchArgs -pack
-i_fetch c r
-i_fetch c x
-i_fetch c y
-i_fetch r c
-i_fetch r x
-i_fetch r y
-i_fetch x c
-i_fetch x r
-i_fetch x x
-i_fetch x y
-i_fetch y c
-i_fetch y r
-i_fetch y x
-i_fetch y y
+put_list y r r
+put_list x r r
+put_list r n r
+put_list r n x
+put_list r x x
+put_list r x r
+put_list x x r
 
 %cold
-i_fetch c c
-i_fetch s s
+put_list s s d
 %hot
 
 #
@@ -631,27 +590,27 @@ return_trace
 # Note: There is no 'move_return y r', since there never are any y registers
 # when we do move_return (if we have y registers, we must do move_deallocate_return).
 
-move S r | return => move_return S r
+move S x==0 | return => move_return S
 
 %macro: move_return MoveReturn -nonext
-move_return x r
-move_return c r
-move_return n r
+move_return x
+move_return c
+move_return n
 
-move S r | deallocate D | return => move_deallocate_return S r D
+move S x==0 | deallocate D | return => move_deallocate_return S D
 
 %macro: move_deallocate_return MoveDeallocateReturn -pack -nonext
-move_deallocate_return x r Q
-move_deallocate_return y r Q
-move_deallocate_return c r Q
-move_deallocate_return n r Q
+move_deallocate_return x Q
+move_deallocate_return y Q
+move_deallocate_return c Q
+move_deallocate_return n Q
 
 deallocate D | return => deallocate_return D
 
 %macro: deallocate_return DeallocateReturn -nonext
 deallocate_return Q
 
-test_heap Need u==1 | put_list Y=y r r => test_heap_1_put_list Need Y
+test_heap Need u==1 | put_list Y=y x==0 x==0 => test_heap_1_put_list Need Y
 
 %macro: test_heap_1_put_list TestHeapPutList -pack
 test_heap_1_put_list I y
@@ -660,18 +619,18 @@ test_heap_1_put_list I y
 
 is_tuple Fail Literal=q => move Literal x | is_tuple Fail x
 is_tuple Fail=f c => jump Fail
-is_tuple Fail=f S=rxy | test_arity Fail=f S=rxy Arity => is_tuple_of_arity Fail S Arity
+is_tuple Fail=f S=xy | test_arity Fail=f S=xy Arity => is_tuple_of_arity Fail S Arity
 
 %macro:is_tuple_of_arity IsTupleOfArity -fail_action
 
+is_tuple_of_arity f r A
 is_tuple_of_arity f x A
 is_tuple_of_arity f y A
-is_tuple_of_arity f r A
 
 %macro: is_tuple IsTuple -fail_action
+is_tuple f r
 is_tuple f x
 is_tuple f y
-is_tuple f r
 
 test_arity Fail Literal=q Arity => move Literal x | test_arity Fail x Arity
 test_arity Fail=f c Arity => jump Fail
@@ -679,95 +638,63 @@ test_arity Fail=f c Arity => jump Fail
 %macro: test_arity IsArity -fail_action
 test_arity f x A
 test_arity f y A
-test_arity f r A
-
-is_tuple_of_arity Fail=f Reg Arity | get_tuple_element Reg P=u==0 Dst=xy => \
-  is_tuple_of_arity Fail Reg Arity | extract_next_element Dst | original_reg Reg P
-
-test_arity Fail Reg Arity | get_tuple_element Reg P=u==0 Dst=xy => \
-  test_arity Fail Reg Arity | extract_next_element Dst | original_reg Reg P
-
-original_reg Reg P1 | get_tuple_element Reg P2 Dst=xy | succ(P1, P2) => \
-  extract_next_element Dst | original_reg Reg P2
-
-get_tuple_element Reg P Dst => i_get_tuple_element Reg P Dst | original_reg Reg P
-
-original_reg Reg Pos =>
-
-original_reg/2
-
-extract_next_element D1=xy | original_reg Reg P1 | get_tuple_element Reg P2 D2=xy | \
-succ(P1, P2) | succ(D1, D2) => \
-  extract_next_element2 D1 | original_reg Reg P2
-
-extract_next_element2 D1=xy | original_reg Reg P1 | get_tuple_element Reg P2 D2=xy | \
-succ(P1, P2) | succ2(D1, D2) => \
-  extract_next_element3 D1 | original_reg Reg P2
-
-#extract_next_element3 D1=xy | original_reg Reg P1 | get_tuple_element Reg P2 D2=xy | \
-#succ(P1, P2) | succ3(D1, D2) => \
-#  extract_next_element4 D1 | original_reg Reg P2
-
-%macro: extract_next_element ExtractNextElement -pack
-extract_next_element x
-extract_next_element y
 
-%macro: extract_next_element2 ExtractNextElement2 -pack
-extract_next_element2 x
-extract_next_element2 y
+get_tuple_element Reg=x P1 D1=x | get_tuple_element Reg=x P2 D2=x | \
+   get_tuple_element Reg=x P3 D3=x | \
+   succ(P1, P2) | succ(P2, P3) | \
+   succ(D1, D2) | succ(D2, D3) => i_get_tuple_element3 Reg P1 D1
 
-%macro: extract_next_element3 ExtractNextElement3 -pack
-extract_next_element3 x
-extract_next_element3 y
+get_tuple_element Reg=x P1 D1=x | get_tuple_element Reg=x P2 D2=x | \
+   succ(P1, P2) | succ(D1, D2) => i_get_tuple_element2 Reg P1 D1
 
-#%macro: extract_next_element4 ExtractNextElement4 -pack
-#extract_next_element4 x
-#extract_next_element4 y
+get_tuple_element Reg P Dst => i_get_tuple_element Reg P Dst
 
 is_integer Fail=f i =>
 is_integer Fail=f an => jump Fail
 is_integer Fail Literal=q => move Literal x | is_integer Fail x
 
-is_integer Fail=f S=rx | allocate Need Regs => is_integer_allocate Fail S Need Regs
+is_integer Fail=f S=x | allocate Need Regs => is_integer_allocate Fail S Need Regs
 
 %macro: is_integer_allocate IsIntegerAllocate -fail_action
 is_integer_allocate f x I I
-is_integer_allocate f r I I
 
 %macro: is_integer IsInteger -fail_action
 is_integer f x
 is_integer f y
-is_integer f r
 
 is_list Fail=f n =>
 is_list Fail Literal=q => move Literal x | is_list Fail x
 is_list Fail=f c => jump Fail
 %macro: is_list IsList -fail_action
-is_list f r
 is_list f x
 %cold
 is_list f y
 %hot
 
-is_nonempty_list Fail=f S=rx | allocate Need Rs => is_nonempty_list_allocate Fail S Need Rs
+is_nonempty_list Fail=f S=x | allocate Need Rs => is_nonempty_list_allocate Fail S Need Rs
 
 %macro:is_nonempty_list_allocate IsNonemptyListAllocate -fail_action -pack
-is_nonempty_list_allocate f x I t
 is_nonempty_list_allocate f r I t
+is_nonempty_list_allocate f x I t
 
-is_nonempty_list F=f r | test_heap I1 I2 => is_non_empty_list_test_heap F r I1 I2
+is_nonempty_list F=f x==0 | test_heap I1 I2 => is_non_empty_list_test_heap F I1 I2
 
 %macro: is_non_empty_list_test_heap IsNonemptyListTestHeap -fail_action -pack
-is_non_empty_list_test_heap f r I t
+is_non_empty_list_test_heap f I t
+
+is_nonempty_list Fail=f S=x | get_list S D1=x D2=x => \
+  is_nonempty_list_get_list Fail S D1 D2
+
+%macro: is_nonempty_list_get_list IsNonemptyListGetList -fail_action -pack
+is_nonempty_list_get_list f r x x
+is_nonempty_list_get_list f x x x
 
 %macro: is_nonempty_list IsNonemptyList -fail_action
 is_nonempty_list f x
 is_nonempty_list f y
-is_nonempty_list f r
 
 %macro: is_atom IsAtom -fail_action
 is_atom f x
-is_atom f r
 %cold
 is_atom f y
 %hot
@@ -775,7 +702,6 @@ is_atom Fail=f a =>
 is_atom Fail=f niq => jump Fail
 
 %macro: is_float IsFloat -fail_action
-is_float f r
 is_float f x
 %cold
 is_float f y
@@ -789,12 +715,10 @@ is_nil Fail=f qia => jump Fail
 %macro: is_nil IsNil -fail_action
 is_nil f x
 is_nil f y
-is_nil f r
 
 is_binary Fail Literal=q => move Literal x | is_binary Fail x
 is_binary Fail=f c => jump Fail
 %macro: is_binary IsBinary -fail_action
-is_binary f r
 is_binary f x
 %cold
 is_binary f y
@@ -806,7 +730,6 @@ is_bitstr Fail Term => is_bitstring Fail Term
 is_bitstring Fail Literal=q => move Literal x | is_bitstring Fail x
 is_bitstring Fail=f c => jump Fail
 %macro: is_bitstring IsBitstring -fail_action
-is_bitstring f r
 is_bitstring f x
 %cold
 is_bitstring f y
@@ -814,7 +737,6 @@ is_bitstring f y
 
 is_reference Fail=f cq => jump Fail
 %macro: is_reference IsRef -fail_action
-is_reference f r
 is_reference f x
 %cold
 is_reference f y
@@ -822,7 +744,6 @@ is_reference f y
 
 is_pid Fail=f cq => jump Fail
 %macro: is_pid IsPid -fail_action
-is_pid f r
 is_pid f x
 %cold
 is_pid f y
@@ -830,7 +751,6 @@ is_pid f y
 
 is_port Fail=f cq => jump Fail
 %macro: is_port IsPort -fail_action
-is_port f r
 is_port f x
 %cold
 is_port f y
@@ -842,14 +762,12 @@ is_boolean Fail=f ac => jump Fail
 
 %cold
 %macro: is_boolean IsBoolean -fail_action
-is_boolean f r
 is_boolean f x
 is_boolean f y
 %hot
 
 is_function2 Fail=f acq Arity => jump Fail
 is_function2 Fail=f Fun a => jump Fail
-is_function2 Fail Fun Literal=q => move Literal x | is_function2 Fail Fun x
 
 is_function2 f s s
 %macro: is_function2 IsFunction2 -fail_action
@@ -989,76 +907,76 @@ call_ext_only u==3 u$func:erlang:hibernate/3 => i_hibernate
 %unless USE_VM_PROBES
 
 call_ext Arity u$func:erlang:dt_get_tag/0 => \
-    move a=am_undefined r
+    move a=am_undefined x=0
 call_ext_last Arity u$func:erlang:dt_get_tag/0 D => \
-    move a=am_undefined r | deallocate D | return
+    move a=am_undefined x=0 | deallocate D | return
 call_ext_only Arity u$func:erlang:dt_get_tag/0 => \
-    move a=am_undefined r | return
-
-move Any r | call_ext Arity u$func:erlang:dt_put_tag/1 => \
-    move a=am_undefined r
-move Any r | call_ext_last Arity u$func:erlang:dt_put_tag/1 D => \
-    move a=am_undefined r | deallocate D | return
-move Any r | call_ext_only Arity u$func:erlang:dt_put_tag/1 => \
-    move a=am_undefined r | return
+    move a=am_undefined x=0 | return
+
+move Any x==0 | call_ext Arity u$func:erlang:dt_put_tag/1 => \
+    move a=am_undefined x=0
+move Any x==0 | call_ext_last Arity u$func:erlang:dt_put_tag/1 D => \
+    move a=am_undefined x=0 | deallocate D | return
+move Any x==0 | call_ext_only Arity u$func:erlang:dt_put_tag/1 => \
+    move a=am_undefined x=0 | return
 call_ext Arity u$func:erlang:dt_put_tag/1 => \
-    move a=am_undefined r
+    move a=am_undefined x=0
 call_ext_last Arity u$func:erlang:dt_put_tag/1 D => \
-    move a=am_undefined r | deallocate D | return
+    move a=am_undefined x=0 | deallocate D | return
 call_ext_only Arity u$func:erlang:dt_put_tag/1 => \
-    move a=am_undefined r | return
+    move a=am_undefined x=0 | return
 
 call_ext Arity u$func:erlang:dt_get_tag_data/0 => \
-    move a=am_undefined r
+    move a=am_undefined x=0
 call_ext_last Arity u$func:erlang:dt_get_tag_data/0 D => \
-    move a=am_undefined r | deallocate D | return
+    move a=am_undefined x=0 | deallocate D | return
 call_ext_only Arity u$func:erlang:dt_get_tag_data/0 => \
-    move a=am_undefined r | return
-
-move Any r | call_ext Arity u$func:erlang:dt_spread_tag/1 => \
-    move a=am_true r
-move Any r | call_ext_last Arity u$func:erlang:dt_spread_tag/1 D => \
-    move a=am_true r | deallocate D | return
-move Any r | call_ext_only Arity u$func:erlang:dt_spread_tag/1 => \
-    move a=am_true r | return
+    move a=am_undefined x=0 | return
+
+move Any x==0 | call_ext Arity u$func:erlang:dt_spread_tag/1 => \
+    move a=am_true x=0
+move Any x==0 | call_ext_last Arity u$func:erlang:dt_spread_tag/1 D => \
+    move a=am_true x=0 | deallocate D | return
+move Any x==0 | call_ext_only Arity u$func:erlang:dt_spread_tag/1 => \
+    move a=am_true x=0 | return
 call_ext Arity u$func:erlang:dt_spread_tag/1 => \
-    move a=am_true r
+    move a=am_true x=0
 call_ext_last Arity u$func:erlang:dt_spread_tag/1 D => \
-    move a=am_true r | deallocate D | return
+    move a=am_true x=0 | deallocate D | return
 call_ext_only Arity u$func:erlang:dt_spread_tag/1 => \
-    move a=am_true r | return
-
-move Any r | call_ext Arity u$func:erlang:dt_restore_tag/1 => \
-    move a=am_true r
-move Any r | call_ext_last Arity u$func:erlang:dt_restore_tag/1 D => \
-    move a=am_true r | deallocate D | return
-move Any r | call_ext_only Arity u$func:erlang:dt_restore_tag/1 => \
-    move a=am_true r | return
+    move a=am_true x=0 | return
+
+move Any x==0 | call_ext Arity u$func:erlang:dt_restore_tag/1 => \
+    move a=am_true x=0
+move Any x==0 | call_ext_last Arity u$func:erlang:dt_restore_tag/1 D => \
+    move a=am_true x=0 | deallocate D | return
+move Any x==0 | call_ext_only Arity u$func:erlang:dt_restore_tag/1 => \
+    move a=am_true x=0 | return
 call_ext Arity u$func:erlang:dt_restore_tag/1 => \
-    move a=am_true r
+    move a=am_true x=0
 call_ext_last Arity u$func:erlang:dt_restore_tag/1 D => \
-    move a=am_true r | deallocate D | return
+    move a=am_true x=0 | deallocate D | return
 call_ext_only Arity u$func:erlang:dt_restore_tag/1 => \
-    move a=am_true r | return
-
-move Any r | call_ext Arity u$func:erlang:dt_prepend_vm_tag_data/1 => \
-    move Any r
-move Any r | call_ext_last Arity u$func:erlang:dt_prepend_vm_tag_data/1 D => \
-    move Any r | deallocate D | return
-move Any r | call_ext_only Arity u$func:erlang:dt_prepend_vm_tag_data/1 => \
-    move Any r | return
+    move a=am_true x=0 | return
+
+move Any x==0 | call_ext Arity u$func:erlang:dt_prepend_vm_tag_data/1 => \
+    move Any x=0
+move Any x==0 | call_ext_last Arity u$func:erlang:dt_prepend_vm_tag_data/1 D => \
+    move Any x=0 | deallocate D | return
+move Any x==0 | call_ext_only Arity u$func:erlang:dt_prepend_vm_tag_data/1 => \
+    move Any x=0 | return
 call_ext Arity u$func:erlang:dt_prepend_vm_tag_data/1 =>
 call_ext_last Arity u$func:erlang:dt_prepend_vm_tag_data/1 D => \
     deallocate D | return
 call_ext_only Arity u$func:erlang:dt_prepend_vm_tag_data/1 => \
     return
 
-move Any r | call_ext Arity u$func:erlang:dt_append_vm_tag_data/1 => \
-    move Any r
-move Any r | call_ext_last Arity u$func:erlang:dt_append_vm_tag_data/1 D => \
-    move Any r | deallocate D | return
-move Any r | call_ext_only Arity u$func:erlang:dt_append_vm_tag_data/1 => \
-    move Any r | return
+move Any x==0 | call_ext Arity u$func:erlang:dt_append_vm_tag_data/1 => \
+    move Any x=0
+move Any x==0 | call_ext_last Arity u$func:erlang:dt_append_vm_tag_data/1 D => \
+    move Any x=0 | deallocate D | return
+move Any x==0 | call_ext_only Arity u$func:erlang:dt_append_vm_tag_data/1 => \
+    move Any x=0 | return
 call_ext Arity u$func:erlang:dt_append_vm_tag_data/1 =>
 call_ext_last Arity u$func:erlang:dt_append_vm_tag_data/1 D => \
     deallocate D | return
@@ -1066,7 +984,7 @@ call_ext_only Arity u$func:erlang:dt_append_vm_tag_data/1 => \
     return
 
 # Can happen after one of the transformations above.
-move Discarded r | move Something r => move Something r
+move Discarded x==0 | move Something x==0 => move Something x=0
 
 %endif
 
@@ -1091,9 +1009,9 @@ call_ext_only Ar=u Bif=u$is_bif => \
 # with call instructions.
 #
 
-move S=c r | call_ext Ar=u Func=u$is_not_bif => i_move_call_ext S r Func
-move S=c r | call_ext_last Ar=u Func=u$is_not_bif D => i_move_call_ext_last Func D S r
-move S=c r | call_ext_only Ar=u Func=u$is_not_bif => i_move_call_ext_only Func S r 
+move S=c x==0 | call_ext Ar=u Func=u$is_not_bif => i_move_call_ext S Func
+move S=c x==0 | call_ext_last Ar=u Func=u$is_not_bif D => i_move_call_ext_last Func D S
+move S=c x==0 | call_ext_only Ar=u Func=u$is_not_bif => i_move_call_ext_only Func S
 
 call_ext Ar Func        => i_call_ext Func
 call_ext_last Ar Func D => i_call_ext_last Func D
@@ -1118,80 +1036,74 @@ call_bif e
 bif0 u$bif:erlang:self/0 Dst=d => self Dst
 bif0 u$bif:erlang:node/0 Dst=d => node Dst
 
-bif1 Fail Bif=u$bif:erlang:get/1 Src=s Dst=d => i_get Src Dst
+bif1 Fail Bif=u$bif:erlang:get/1 Src=s Dst=d => gen_get(Src, Dst)
 
-bif2 Jump=j u$bif:erlang:element/2 S1=s S2=rxy Dst=d => gen_element(Jump, S1, S2, Dst)
+bif2 Jump=j u$bif:erlang:element/2 S1=s S2=xy Dst=d => gen_element(Jump, S1, S2, Dst)
 
-bif1 Fail Bif Literal=q Dst => move Literal x | bif1 Fail Bif x Dst
 bif1 p Bif S1 Dst => bif1_body Bif S1 Dst
 
-bif1_body Bif Literal=q Dst => move Literal x | bif1_body Bif x Dst
-
-bif2 p Bif S1 S2 Dst => i_fetch S1 S2 | i_bif2_body Bif Dst
-bif2 Fail Bif S1 S2 Dst => i_fetch S1 S2 | i_bif2 Fail Bif Dst
+bif2 p Bif S1 S2 Dst => i_bif2_body Bif S1 S2 Dst
+bif2 Fail Bif S1 S2 Dst => i_bif2 Fail Bif S1 S2 Dst
 
+i_get_hash c I d
 i_get s d
 
 %macro: self Self
-self r
 self x
 self y
 
 %macro: node Node
-node r
 node x
 %cold
 node y
 %hot
 
-i_fast_element r j I d
-i_fast_element x j I d
-i_fast_element y j I d
+i_fast_element j x I d
+i_fast_element j y I d
 
-i_element r j s d
-i_element x j s d
-i_element y j s d
+i_element j x s d
+i_element j y s d
 
 bif1 f b s d
 bif1_body b s d
-i_bif2 f b d
-i_bif2_body b d
+i_bif2 f b s s d
+i_bif2_body b s s d
 
 #
 # Internal calls.
 #
 
-move S=c r | call Ar P=f => i_move_call S r P
-move S=s r | call Ar P=f => move_call S r P
+move S=c x==0 | call Ar P=f => i_move_call S P
+move S=s x==0 | call Ar P=f => move_call S P
 
-i_move_call c r f
+i_move_call c f
 
 %macro:move_call MoveCall -arg_f -size -nonext
-move_call/3
+move_call/2
 
-move_call x r f
-move_call y r f
+move_call x f
+move_call y f
 
-move S=c r | call_last Ar P=f D => i_move_call_last P D S r
-move S r | call_last Ar P=f D => move_call_last S r P D
+move S=c x==0 | call_last Ar P=f D => i_move_call_last P D S
+move S x==0 | call_last Ar P=f D => move_call_last S P D
 
-i_move_call_last f P c r
+i_move_call_last f P c
 
 %macro:move_call_last MoveCallLast -arg_f -nonext -pack
 
-move_call_last/4
-move_call_last x r f Q
-move_call_last y r f Q
+move_call_last/3
+move_call_last x f Q
+move_call_last y f Q
 
-move S=c r | call_only Ar P=f => i_move_call_only P S r
-move S=x r | call_only Ar P=f => move_call_only S r P
+move S=c x==0 | call_only Ar P=f => i_move_call_only P S
+move S=x x==0 | call_only Ar P=f => move_call_only S P
 
-i_move_call_only f c r
+i_move_call_only f c
 
 %macro:move_call_only MoveCallOnly -arg_f -nonext
-move_call_only/3
+move_call_only/2
 
-move_call_only x r f
+move_call_only x f
 
 call Ar Func        => i_call Func
 call_last Ar Func D => i_call_last Func D
@@ -1205,9 +1117,9 @@ i_call_ext e
 i_call_ext_last e P
 i_call_ext_only e
 
-i_move_call_ext c r e
-i_move_call_ext_last e P c r
-i_move_call_ext_only e c r
+i_move_call_ext c e
+i_move_call_ext_last e P c
+i_move_call_ext_only e c
 
 # Fun calls.
 
@@ -1227,7 +1139,6 @@ i_make_fun I t
 %macro: is_function IsFunction -fail_action
 is_function f x
 is_function f y
-is_function f r
 is_function Fail=f c => jump Fail
 
 func_info M F A => i_func_info u M F A
@@ -1239,131 +1150,105 @@ func_info M F A => i_func_info u M F A
 %cold
 bs_start_match2 Fail=f ica X Y D => jump Fail
 bs_start_match2 Fail Bin X Y D => i_bs_start_match2 Bin Fail X Y D
-i_bs_start_match2 r f I I d
 i_bs_start_match2 x f I I d
 i_bs_start_match2 y f I I d
 
 bs_save2 Reg Index => gen_bs_save(Reg, Index)
-i_bs_save2 r I
 i_bs_save2 x I
 
 bs_restore2 Reg Index => gen_bs_restore(Reg, Index)
-i_bs_restore2 r I
 i_bs_restore2 x I
 
 # Matching integers
 bs_match_string Fail Ms Bits Val => i_bs_match_string Ms Fail Bits Val
 
-i_bs_match_string r f I I
 i_bs_match_string x f I I
 
 # Fetching integers from binaries.
-bs_get_integer2 Fail=f Ms=rx Live=u Sz=sq Unit=u Flags=u Dst=d => \
+bs_get_integer2 Fail=f Ms=x Live=u Sz=sq Unit=u Flags=u Dst=d => \
 			gen_get_integer2(Fail, Ms, Live, Sz, Unit, Flags, Dst)
 
-i_bs_get_integer_small_imm r I f I d
 i_bs_get_integer_small_imm x I f I d
-i_bs_get_integer_imm r I I f I d
 i_bs_get_integer_imm x I I f I d
-i_bs_get_integer f I I d
-i_bs_get_integer_8 r f d
+i_bs_get_integer f I I s s d
 i_bs_get_integer_8 x f d
-i_bs_get_integer_16 r f d
 i_bs_get_integer_16 x f d
-i_bs_get_integer_32 r f I d
 i_bs_get_integer_32 x f I d
 
 # Fetching binaries from binaries.
-bs_get_binary2 Fail=f Ms=rx Live=u Sz=sq Unit=u Flags=u Dst=d => \
+bs_get_binary2 Fail=f Ms=x Live=u Sz=sq Unit=u Flags=u Dst=d => \
 			gen_get_binary2(Fail, Ms, Live, Sz, Unit, Flags, Dst)
 
 %macro: i_bs_get_binary_imm2 BsGetBinaryImm_2 -fail_action -gen_dest
 %macro: i_bs_get_binary2 BsGetBinary_2 -fail_action -gen_dest
 %macro: i_bs_get_binary_all2 BsGetBinaryAll_2 -fail_action -gen_dest
 
-i_bs_get_binary_imm2 f r I I I d
 i_bs_get_binary_imm2 f x I I I d
-i_bs_get_binary2 f r I s I d
 i_bs_get_binary2 f x I s I d
-i_bs_get_binary_all2 f r I I d
 i_bs_get_binary_all2 f x I I d
-i_bs_get_binary_all_reuse r f I
 i_bs_get_binary_all_reuse x f I
 
 # Fetching float from binaries.
-bs_get_float2 Fail=f Ms=rx Live=u Sz=s Unit=u Flags=u Dst=d => \
+bs_get_float2 Fail=f Ms=x Live=u Sz=s Unit=u Flags=u Dst=d => \
 		gen_get_float2(Fail, Ms, Live, Sz, Unit, Flags, Dst)
 
-bs_get_float2 Fail=f Ms=rx Live=u Sz=q Unit=u Flags=u Dst=d => jump Fail
+bs_get_float2 Fail=f Ms=x Live=u Sz=q Unit=u Flags=u Dst=d => jump Fail
 
 %macro: i_bs_get_float2 BsGetFloat2 -fail_action -gen_dest
-i_bs_get_float2 f r I s I d
 i_bs_get_float2 f x I s I d
 
 # Miscellanous
 
-bs_skip_bits2 Fail=f Ms=rx Sz=s Unit=u Flags=u => \
-			gen_skip_bits2(Fail, Ms, Sz, Unit, Flags)
-bs_skip_bits2 Fail=f Ms=rx Sz=q Unit=u Flags=u => \
+bs_skip_bits2 Fail=f Ms=x Sz=sq Unit=u Flags=u => \
 			gen_skip_bits2(Fail, Ms, Sz, Unit, Flags)
 
 %macro: i_bs_skip_bits_imm2 BsSkipBitsImm2 -fail_action
-i_bs_skip_bits_imm2 f r I
 i_bs_skip_bits_imm2 f x I
 
 %macro: i_bs_skip_bits2 BsSkipBits2 -fail_action
-i_bs_skip_bits2 f r x I
-i_bs_skip_bits2 f r y I
 i_bs_skip_bits2 f x x I
-i_bs_skip_bits2 f x r I
 i_bs_skip_bits2 f x y I
 
 %macro: i_bs_skip_bits_all2 BsSkipBitsAll2 -fail_action
-i_bs_skip_bits_all2 f r I
 i_bs_skip_bits_all2 f x I 
 
-bs_test_tail2 Fail=f Ms=rx Bits=u==0 => bs_test_zero_tail2 Fail Ms
-bs_test_tail2 Fail=f Ms=rx Bits=u => bs_test_tail_imm2 Fail Ms Bits
-bs_test_zero_tail2 f r
+bs_test_tail2 Fail=f Ms=x Bits=u==0 => bs_test_zero_tail2 Fail Ms
+bs_test_tail2 Fail=f Ms=x Bits=u => bs_test_tail_imm2 Fail Ms Bits
 bs_test_zero_tail2 f x
-bs_test_tail_imm2 f r I
 bs_test_tail_imm2 f x I
 
 bs_test_unit F Ms Unit=u==8 => bs_test_unit8 F Ms
-bs_test_unit f r I
 bs_test_unit f x I
-bs_test_unit8 f r
 bs_test_unit8 f x
 
-bs_context_to_binary r
+# An y register operand for bs_context_to_binary is rare,
+# but can happen because of inlining.
+
+bs_context_to_binary Y=y => move Y x | bs_context_to_binary x
+
 bs_context_to_binary x
-bs_context_to_binary y
 
 #
 # Utf8/utf16/utf32 support. (R12B-5)
 #
-bs_get_utf8 Fail=f Ms=rx u u Dst=d => i_bs_get_utf8 Ms Fail Dst
-i_bs_get_utf8 r f d
+bs_get_utf8 Fail=f Ms=x u u Dst=d => i_bs_get_utf8 Ms Fail Dst
 i_bs_get_utf8 x f d
 
-bs_skip_utf8 Fail=f Ms=rx u u => i_bs_get_utf8 Ms Fail x
+bs_skip_utf8 Fail=f Ms=x u u => i_bs_get_utf8 Ms Fail x
 
-bs_get_utf16 Fail=f Ms=rx u Flags=u Dst=d => i_bs_get_utf16 Ms Fail Flags Dst
-bs_skip_utf16 Fail=f Ms=rx u Flags=u => i_bs_get_utf16 Ms Fail Flags x
+bs_get_utf16 Fail=f Ms=x u Flags=u Dst=d => i_bs_get_utf16 Ms Fail Flags Dst
+bs_skip_utf16 Fail=f Ms=x u Flags=u => i_bs_get_utf16 Ms Fail Flags x
 
-i_bs_get_utf16 r f I d
 i_bs_get_utf16 x f I d
 
-bs_get_utf32 Fail=f Ms=rx Live=u Flags=u Dst=d => \
+bs_get_utf32 Fail=f Ms=x Live=u Flags=u Dst=d => \
 	bs_get_integer2 Fail Ms Live i=32 u=1 Flags Dst | \
-	i_fetch Dst Ms | \
-	i_bs_validate_unicode_retract Fail
-bs_skip_utf32 Fail=f Ms=rx Live=u Flags=u => \
+	i_bs_validate_unicode_retract Fail Dst Ms
+bs_skip_utf32 Fail=f Ms=x Live=u Flags=u => \
 	bs_get_integer2 Fail Ms Live i=32 u=1 Flags x | \
-	i_fetch x Ms | \
-	i_bs_validate_unicode_retract Fail
+	i_bs_validate_unicode_retract Fail x Ms
 
-i_bs_validate_unicode_retract j
+i_bs_validate_unicode_retract j s s
 %hot
 
 #
@@ -1385,13 +1270,12 @@ bs_init2 Fail Sz=u Words Regs Flags Dst => \
 bs_init2 Fail Sz Words=u==0 Regs Flags Dst => \
   i_bs_init_fail Sz Fail Regs Dst
 bs_init2 Fail Sz Words Regs Flags Dst => \
-  i_fetch Sz r | i_bs_init_fail_heap Words Fail Regs Dst
+  i_bs_init_fail_heap Sz Words Fail Regs Dst
 
-i_bs_init_fail r j I d
 i_bs_init_fail x j I d
 i_bs_init_fail y j I d
 
-i_bs_init_fail_heap I j I d
+i_bs_init_fail_heap s I j I d
 
 i_bs_init I I d
 i_bs_init_heap_bin I I d
@@ -1408,39 +1292,35 @@ bs_init_bits Fail Sz=u Words Regs Flags Dst =>  i_bs_init_bits_heap Sz Words Reg
 bs_init_bits Fail Sz Words=u==0 Regs Flags Dst => \
   i_bs_init_bits_fail Sz Fail Regs Dst
 bs_init_bits Fail Sz Words Regs Flags Dst => \
-  i_fetch Sz r | i_bs_init_bits_fail_heap Words Fail Regs Dst
+  i_bs_init_bits_fail_heap Sz Words Fail Regs Dst
 
-i_bs_init_bits_fail r j I d
 i_bs_init_bits_fail x j I d
 i_bs_init_bits_fail y j I d
 
-i_bs_init_bits_fail_heap I j I d
+i_bs_init_bits_fail_heap s I j I d
 
 i_bs_init_bits I I d
 i_bs_init_bits_heap I I I d
 
 bs_add Fail S1=i==0 S2 Unit=u==1 D => move S2 D
-bs_add Fail S1 S2 Unit D => i_fetch S1 S2 | i_bs_add Fail Unit D
 
-i_bs_add j I d
+bs_add j s s I d
 
 bs_append Fail Size Extra Live Unit Bin Flags Dst => \
-  i_fetch Size Bin | i_bs_append Fail Extra Live Unit Dst
+  move Bin x | i_bs_append Fail Extra Live Unit Size Dst
 
 bs_private_append Fail Size Unit Bin Flags Dst => \
-  i_fetch Size Bin | i_bs_private_append Fail Unit Dst
+  i_bs_private_append Fail Unit Size Bin Dst
 
 bs_init_writable
 
-i_bs_append j I I I d
-i_bs_private_append j I d
+i_bs_append j I I I s d
+i_bs_private_append j I s s d
 
 #
 # Storing integers into binaries.
 #
 
-bs_put_integer Fail=j Sz=s Unit=u Flags=u Literal=q => \
-	move Literal x | bs_put_integer Fail Sz Unit Flags x
 bs_put_integer Fail=j Sz=sq Unit=u Flags=u Src=s => \
 			gen_put_integer(Fail, Sz, Unit, Flags, Src)
 
@@ -1454,32 +1334,22 @@ i_new_bs_put_integer_imm j I I s
 # Utf8/utf16/utf32 support. (R12B-5)
 #
 
-bs_utf8_size Fail Literal=q Dst=d => \
-	move Literal x | bs_utf8_size Fail x Dst
 bs_utf8_size j Src=s Dst=d => i_bs_utf8_size Src Dst
 
 i_bs_utf8_size s d
 
-bs_utf16_size Fail Literal=q Dst=d => \
-	move Literal x | bs_utf16_size Fail x Dst
 bs_utf16_size j Src=s Dst=d => i_bs_utf16_size Src Dst
 
 i_bs_utf16_size s d
 
-bs_put_utf8 Fail=j Flags=u Literal=q => \
-	move Literal x | bs_put_utf8 Fail Flags x
 bs_put_utf8 Fail u Src=s => i_bs_put_utf8 Fail Src
 
 i_bs_put_utf8 j s
 
-bs_put_utf16 Fail=j Flags=u Literal=q => \
-	move Literal x | bs_put_utf16 Fail Flags x
 bs_put_utf16 Fail Flags=u Src=s => i_bs_put_utf16 Fail Flags Src
 
 i_bs_put_utf16 j I s
 
-bs_put_utf32 Fail=j Flags=u Literal=q => \
-	move Literal x | bs_put_utf32 Fail Flags x
 bs_put_utf32 Fail=j Flags=u Src=s => \
    i_bs_validate_unicode Fail Src | bs_put_integer Fail i=32 u=1 Flags Src
 
@@ -1490,9 +1360,6 @@ i_bs_validate_unicode j s
 #
 bs_put_float Fail Sz=q Unit Flags Val => badarg Fail
 
-bs_put_float Fail=j Sz Unit=u Flags=u Literal=q => \
-	move Literal x | bs_put_float Fail Sz Unit Flags x
-
 bs_put_float Fail=j Sz=s Unit=u Flags=u Src=s => \
 			gen_put_float(Fail, Sz, Unit, Flags, Src)
 
@@ -1506,8 +1373,6 @@ i_new_bs_put_float_imm j I I s
 # Storing binaries into binaries.
 #
 
-bs_put_binary Fail Sz Unit Flags Literal=q => \
- 	move Literal x | bs_put_binary Fail Sz Unit Flags x
 bs_put_binary Fail=j Sz=s Unit=u Flags=u Src=s => \
 			gen_put_binary(Fail, Sz, Unit, Flags, Src)
 
@@ -1600,7 +1465,6 @@ is_map Fail Lit=q | literal_is_map(Lit) =>
 is_map Fail cq => jump Fail
 
 %macro: is_map IsMap -fail_action
-is_map f r
 is_map f x
 is_map f y
 
@@ -1611,101 +1475,120 @@ has_map_fields Fail Src Size Rest=* => \
 
 ## Transform get_map_elements(s) #{ K1 := V1, K2 := V2 }
 
-get_map_elements Fail Src=rxy Size=u==2 Rest=* => \
+get_map_elements Fail Src=xy Size=u==2 Rest=* => \
     gen_get_map_element(Fail, Src, Size, Rest)
 get_map_elements Fail Src Size Rest=* | map_key_sort(Size, Rest) => \
    gen_get_map_elements(Fail, Src, Size, Rest)
 
 i_get_map_elements f s I
 
-i_get_map_element Fail Src=rxy Key=ry Dst => \
+i_get_map_element Fail Src=xy Key=y Dst => \
     move Key x | i_get_map_element Fail Src x Dst
 
 %macro: i_get_map_element_hash GetMapElementHash -fail_action
-i_get_map_element_hash f r c I r
-i_get_map_element_hash f x c I r
-i_get_map_element_hash f y c I r
-i_get_map_element_hash f r c I x
 i_get_map_element_hash f x c I x
 i_get_map_element_hash f y c I x
-i_get_map_element_hash f r c I y
 i_get_map_element_hash f x c I y
 i_get_map_element_hash f y c I y
 
 %macro: i_get_map_element GetMapElement -fail_action
-i_get_map_element f r x r
-i_get_map_element f x x r
-i_get_map_element f y x r
-i_get_map_element f r x x
 i_get_map_element f x x x
 i_get_map_element f y x x
-i_get_map_element f r x y
 i_get_map_element f x x y
 i_get_map_element f y x y
 
 #
+# Convert the plus operations to a generic plus instruction.
+#
+gen_plus/5
+gen_minus/5
+
+gc_bif1 Fail Live u$bif:erlang:splus/1 Src Dst => \
+   gen_plus Fail Live Src i Dst
+gc_bif2 Fail Live u$bif:erlang:splus/2 S1 S2 Dst => \
+   gen_plus Fail Live S1 S2 Dst
+
+gc_bif1 Fail Live u$bif:erlang:sminus/1 Src Dst => \
+   gen_minus Fail Live i Src Dst
+gc_bif2 Fail Live u$bif:erlang:sminus/2 S1 S2 Dst => \
+   gen_minus Fail Live S1 S2 Dst
+
+#
 # Optimize addition and subtraction of small literals using
 # the i_increment/4 instruction (in bodies, not in guards).
 #
 
-gc_bif2 p Live u$bif:erlang:splus/2 Int=i Reg=d Dst => \
+gen_plus p Live Int=i Reg=d Dst => \
 	gen_increment(Reg, Int, Live, Dst)
-gc_bif2 p Live u$bif:erlang:splus/2 Reg=d Int=i Dst => \
+gen_plus p Live Reg=d Int=i Dst => \
 	gen_increment(Reg, Int, Live, Dst)
 
-gc_bif2 p Live u$bif:erlang:sminus/2 Reg=d Int=i Dst | \
-	negation_is_small(Int) => \
+gen_minus p Live Reg=d Int=i Dst | negation_is_small(Int) => \
 	gen_increment_from_minus(Reg, Int, Live, Dst)
 
 #
 # GCing arithmetic instructions.
 #
 
-gc_bif2 Fail I u$bif:erlang:splus/2 S1=x S2=x Dst=d => i_plus Fail I S1 S2 Dst
-gc_bif2 Fail I u$bif:erlang:splus/2 S1 S2 Dst=d => i_fetch S1 S2 | i_plus Fail I Dst
-gc_bif2 Fail I u$bif:erlang:sminus/2 S1=x S2=x Dst=d => i_minus Fail I S1 S2 Dst
-gc_bif2 Fail I u$bif:erlang:sminus/2 S1 S2 Dst=d => i_fetch S1 S2 | i_minus Fail I Dst
-gc_bif2 Fail I u$bif:erlang:stimes/2 S1 S2 Dst=d => i_fetch S1 S2 | i_times Fail I Dst
-gc_bif2 Fail I u$bif:erlang:div/2 S1 S2 Dst=d => i_fetch S1 S2 | i_m_div Fail I Dst
+gen_plus Fail Live Y=y X=x Dst => i_plus Fail Live X Y Dst
+gen_plus Fail Live S1 S2 Dst => i_plus Fail Live S1 S2 Dst
 
-gc_bif2 Fail I u$bif:erlang:intdiv/2 S1 S2 Dst=d => i_fetch S1 S2 | i_int_div Fail I Dst
-gc_bif2 Fail I u$bif:erlang:rem/2 S1=x S2=x Dst=d => i_rem Fail I S1 S2 Dst
-gc_bif2 Fail I u$bif:erlang:rem/2 S1 S2 Dst=d => i_fetch S1 S2 | i_rem Fail I Dst
+gen_minus Fail Live S1 S2 Dst => i_minus Fail Live S1 S2 Dst
 
-gc_bif2 Fail I u$bif:erlang:bsl/2 S1 S2 Dst=d => i_fetch S1 S2 | i_bsl Fail I Dst
-gc_bif2 Fail I u$bif:erlang:bsr/2 S1 S2 Dst=d => i_fetch S1 S2 | i_bsr Fail I Dst
+gc_bif2 Fail Live u$bif:erlang:stimes/2 S1 S2 Dst => \
+  i_times Fail Live S1 S2 Dst
 
-gc_bif2 Fail I u$bif:erlang:band/2 S1=x S2=c Dst=d => i_band Fail I S1 S2 Dst
-gc_bif2 Fail I u$bif:erlang:band/2 S1 S2 Dst=d => i_fetch S1 S2 | i_band Fail I Dst
-gc_bif2 Fail I u$bif:erlang:bor/2 S1 S2 Dst=d => i_fetch S1 S2 | i_bor Fail I Dst
-gc_bif2 Fail I u$bif:erlang:bxor/2 S1 S2 Dst=d => i_fetch S1 S2 | i_bxor Fail I Dst
+gc_bif2 Fail Live u$bif:erlang:div/2 S1 S2 Dst => \
+  i_m_div Fail Live S1 S2 Dst
+gc_bif2 Fail Live u$bif:erlang:intdiv/2 S1 S2 Dst => \
+  i_int_div Fail Live S1 S2 Dst
 
-gc_bif1 Fail I u$bif:erlang:bnot/1 Src Dst=d => i_int_bnot Fail Src I Dst
+gc_bif2 Fail Live u$bif:erlang:rem/2 S1 S2 Dst => \
+  i_rem Fail Live S1 S2 Dst
+
+gc_bif2 Fail Live u$bif:erlang:bsl/2 S1 S2 Dst => \
+  i_bsl Fail Live S1 S2 Dst
+gc_bif2 Fail Live u$bif:erlang:bsr/2 S1 S2 Dst => \
+  i_bsr Fail Live S1 S2 Dst
+
+gc_bif2 Fail Live u$bif:erlang:band/2 S1 S2 Dst => \
+  i_band Fail Live S1 S2 Dst
+
+gc_bif2 Fail Live u$bif:erlang:bor/2 S1 S2 Dst => \
+  i_bor Fail Live S1 S2 Dst
 
-gc_bif1 Fail I u$bif:erlang:sminus/1 Src Dst=d => i_fetch i Src | i_minus Fail I Dst
-gc_bif1 Fail I u$bif:erlang:splus/1 Src Dst=d => i_fetch i Src | i_plus Fail I Dst
+gc_bif2 Fail Live u$bif:erlang:bxor/2 S1 S2 Dst => \
+  i_bxor Fail Live S1 S2 Dst
+
+gc_bif1 Fail I u$bif:erlang:bnot/1 Src Dst=d => i_int_bnot Fail Src I Dst
 
 i_increment r I I d
 i_increment x I I d
 i_increment y I I d
 
 i_plus j I x x d
-i_plus j I d
+i_plus j I x y d
+i_plus j I s s d
+
 i_minus j I x x d
-i_minus j I d
-i_times j I d
-i_m_div j I d
-i_int_div j I d
+i_minus j I s s d
+
+i_times j I s s d
+
+i_m_div j I s s d
+i_int_div j I s s d
+
 i_rem j I x x d
-i_rem j I d
+i_rem j I s s d
 
-i_bsl j I d
-i_bsr j I d
+i_bsl j I s s d
+i_bsr j I s s d
 
 i_band j I x c d
-i_band j I d
-i_bor j I d
-i_bxor j I d
+i_band j I s s d
+
+i_bor j I s s d
+i_bxor j I s s d
 
 i_int_bnot j s I d
 
@@ -1731,21 +1614,18 @@ gc_bif2 Fail I Bif S1 S2 Dst => \
 gc_bif3 Fail I Bif S1 S2 S3 Dst => \
 	gen_guard_bif3(Fail, I, Bif, S1, S2, S3, Dst)
 
-i_gc_bif1 Fail Bif V=q Live D => move V x | i_gc_bif1 Fail Bif x Live D
-
 i_gc_bif1 j I s I d
 
-ii_gc_bif2/6
-
-ii_gc_bif2 Fail Bif S1 S2 Live D => i_fetch S1 S2 | i_gc_bif2 Fail Bif Live D
-
-i_gc_bif2 j I I d
+i_gc_bif2 j I I s s d
 
 ii_gc_bif3/7
 
-ii_gc_bif3 Fail Bif S1 S2 S3 Live D => move S1 x | i_fetch S2 S3 | i_gc_bif3 Fail Bif x Live D
+# A specific instruction can only have 6 operands, so we must
+# pass one of the arguments in an x register.
+ii_gc_bif3 Fail Bif Live S1 S2 S3 Dst => \
+  move S1 x | i_gc_bif3 Fail Bif Live S2 S3 Dst
 
-i_gc_bif3 j I s I d
+i_gc_bif3 j I I s s d
 
 #
 # The following instruction is specially handled in beam_load.c
diff --git a/erts/emulator/beam/packet_parser.c b/erts/emulator/beam/packet_parser.c
index 2dd421a9e9..a737a86f14 100644
--- a/erts/emulator/beam/packet_parser.c
+++ b/erts/emulator/beam/packet_parser.c
@@ -256,6 +256,7 @@ int packet_get_length(enum PacketParseType htype,
                       const char* ptr, unsigned n, /* Bytes read so far */
                       unsigned max_plen,     /* Max packet length, 0=no limit */
                       unsigned trunc_len,    /* Truncate (lines) if longer, 0=no limit */
+                      char     delimiter,    /* Line delimiting character */
                       int*     statep)       /* Protocol specific state */
 {
     unsigned hlen, plen;
@@ -299,9 +300,9 @@ int packet_get_length(enum PacketParseType htype,
         goto remain;
 
     case TCP_PB_LINE_LF: {
-        /* TCP_PB_LINE_LF:  [Data ... \n]  */
+        /* TCP_PB_LINE_LF:  [Data ... Delimiter]  */
         const char* ptr2;
-        if ((ptr2 = memchr(ptr, '\n', n)) == NULL) {
+        if ((ptr2 = memchr(ptr, delimiter, n)) == NULL) {
             if (n > max_plen && max_plen != 0) { /* packet full */
                 DEBUGF((" => packet full (no NL)=%d\r\n", n));
                 goto error;
diff --git a/erts/emulator/beam/packet_parser.h b/erts/emulator/beam/packet_parser.h
index ff158ff8b8..717d905fad 100644
--- a/erts/emulator/beam/packet_parser.h
+++ b/erts/emulator/beam/packet_parser.h
@@ -105,7 +105,8 @@ int packet_get_length(enum PacketParseType htype,
 		      const char* ptr, unsigned n,  /* Bytes read so far */
 		      unsigned max_plen,      /* Packet max length, 0=no limit */
 		      unsigned trunc_len,     /* Truncate (lines) if longer, 0=no limit */
-		      int* statep);           /* Internal protocol state */
+		      char     delimiter,     /* Line delimiting character */
+		      int*     statep);       /* Internal protocol state */
 
 ERTS_GLB_INLINE
 void packet_get_body(enum PacketParseType htype,
diff --git a/erts/emulator/beam/register.c b/erts/emulator/beam/register.c
index 7ade8bca0f..fdb6cbc813 100644
--- a/erts/emulator/beam/register.c
+++ b/erts/emulator/beam/register.c
@@ -151,6 +151,9 @@ void init_register_table(void)
     f.cmp  = (HCMP_FUN) reg_cmp;
     f.alloc = (HALLOC_FUN) reg_alloc;
     f.free = (HFREE_FUN) reg_free;
+    f.meta_alloc = (HMALLOC_FUN) erts_alloc;
+    f.meta_free = (HMFREE_FUN) erts_free;
+    f.meta_print = (HMPRINT_FUN) erts_print;
 
     hash_init(ERTS_ALC_T_REG_TABLE, &process_reg, "process_reg",
 	      PREG_HASH_SIZE, f);
diff --git a/erts/emulator/beam/sys.h b/erts/emulator/beam/sys.h
index bb871b05ba..53f8313daa 100644
--- a/erts/emulator/beam/sys.h
+++ b/erts/emulator/beam/sys.h
@@ -21,6 +21,19 @@
 #ifndef __SYS_H__
 #define __SYS_H__
 
+#if !defined(__GNUC__)
+#  define ERTS_AT_LEAST_GCC_VSN__(MAJ, MIN, PL) 0
+#elif !defined(__GNUC_MINOR__)
+#  define ERTS_AT_LEAST_GCC_VSN__(MAJ, MIN, PL) \
+  ((__GNUC__ << 24) >= (((MAJ) << 24) | ((MIN) << 12) | (PL)))
+#elif !defined(__GNUC_PATCHLEVEL__)
+#  define ERTS_AT_LEAST_GCC_VSN__(MAJ, MIN, PL) \
+  (((__GNUC__ << 24) | (__GNUC_MINOR__ << 12)) >= (((MAJ) << 24) | ((MIN) << 12) | (PL)))
+#else
+#  define ERTS_AT_LEAST_GCC_VSN__(MAJ, MIN, PL) \
+  (((__GNUC__ << 24) | (__GNUC_MINOR__ << 12) | __GNUC_PATCHLEVEL__) >= (((MAJ) << 24) | ((MIN) << 12) | (PL)))
+#endif
+
 #ifdef ERTS_INLINE
 #  ifndef ERTS_CAN_INLINE
 #    define ERTS_CAN_INLINE 1
@@ -38,6 +51,17 @@
 #  endif
 #endif
 
+#ifndef ERTS_FORCE_INLINE
+#  if ERTS_AT_LEAST_GCC_VSN__(3,1,1)
+#    define ERTS_FORCE_INLINE __inline__ __attribute__((__always_inline__))
+#  elif defined(__WIN32__)
+#    define ERTS_FORCE_INLINE __forceinline
+#  endif
+#  ifndef ERTS_FORCE_INLINE
+#    define ERTS_FORCE_INLINE ERTS_INLINE
+#  endif
+#endif
+
 #if defined(DEBUG) || defined(ERTS_ENABLE_LOCK_CHECK)
 #  undef ERTS_CAN_INLINE
 #  define ERTS_CAN_INLINE 0
@@ -46,8 +70,10 @@
 #endif
 
 #if ERTS_CAN_INLINE
+#define ERTS_GLB_FORCE_INLINE static ERTS_FORCE_INLINE
 #define ERTS_GLB_INLINE static ERTS_INLINE
 #else
+#define ERTS_GLB_FORCE_INLINE
 #define ERTS_GLB_INLINE
 #endif
 
@@ -61,22 +87,14 @@
 #  define NO_FPE_SIGNALS
 #endif
 
-#ifdef DISABLE_CHILD_WAITER_THREAD
-#undef ENABLE_CHILD_WAITER_THREAD
-#endif
-
-#if defined(ERTS_SMP) && !defined(DISABLE_CHILD_WAITER_THREAD)
-#undef ENABLE_CHILD_WAITER_THREAD
-#define ENABLE_CHILD_WAITER_THREAD 1
-#endif
-
 #define ERTS_I64_LITERAL(X) X##LL
 
+#define ErtsInArea(ptr,start,nbytes) \
+    ((UWord)((char*)(ptr) - (char*)(start)) < (nbytes))
+
 #if defined (__WIN32__)
 #  include "erl_win_sys.h"
-#elif defined (__OSE__)
-#  include "erl_ose_sys.h"
-#else 
+#else
 #  include "erl_unix_sys.h"
 #ifndef UNIX
 #  define UNIX 1
@@ -111,19 +129,6 @@ typedef int ErtsSysFdType;
 typedef ERTS_SYS_FD_TYPE ErtsSysFdType;
 #endif
 
-#if !defined(__GNUC__)
-#  define ERTS_AT_LEAST_GCC_VSN__(MAJ, MIN, PL) 0
-#elif !defined(__GNUC_MINOR__)
-#  define ERTS_AT_LEAST_GCC_VSN__(MAJ, MIN, PL) \
-  ((__GNUC__ << 24) >= (((MAJ) << 24) | ((MIN) << 12) | (PL)))
-#elif !defined(__GNUC_PATCHLEVEL__)
-#  define ERTS_AT_LEAST_GCC_VSN__(MAJ, MIN, PL) \
-  (((__GNUC__ << 24) | (__GNUC_MINOR__ << 12)) >= (((MAJ) << 24) | ((MIN) << 12) | (PL)))
-#else
-#  define ERTS_AT_LEAST_GCC_VSN__(MAJ, MIN, PL) \
-  (((__GNUC__ << 24) | (__GNUC_MINOR__ << 12) | __GNUC_PATCHLEVEL__) >= (((MAJ) << 24) | ((MIN) << 12) | (PL)))
-#endif
-
 #if ERTS_AT_LEAST_GCC_VSN__(2, 96, 0)
 #  define ERTS_LIKELY(BOOL)   __builtin_expect((BOOL), !0)
 #  define ERTS_UNLIKELY(BOOL) __builtin_expect((BOOL), 0)
@@ -285,62 +290,11 @@ __decl_noreturn void __noreturn erl_assert_error(const char* expr, const char *f
 #else
 #error Neither 32 nor 64 bit architecture
 #endif
-#if defined(ARCH_64) && defined(HALFWORD_HEAP_EMULATOR)
-#    define HALFWORD_HEAP 1
-#    define HALFWORD_ASSERT 0
-#    define ASSERT_HALFWORD(COND) ASSERT(COND)
-#    undef ERTS_SIZEOF_TERM
-#    define ERTS_SIZEOF_TERM 4
-#else
-#    define HALFWORD_HEAP 0
-#    define HALFWORD_ASSERT 0
-#    define ASSERT_HALFWORD(COND)
-#endif
 
 #if SIZEOF_VOID_P != SIZEOF_SIZE_T
 #error sizeof(void*) != sizeof(size_t)
 #endif
 
-#if HALFWORD_HEAP
-
-#if SIZEOF_INT == 4
-typedef unsigned int Eterm;
-typedef unsigned int Uint;
-typedef int          Sint;
-#define ERTS_UINT_MAX UINT_MAX
-#define ERTS_SIZEOF_ETERM SIZEOF_INT
-#define ErtsStrToSint strtol
-#else
-#error Found no appropriate type to use for 'Eterm', 'Uint' and 'Sint'
-#endif
-
-#if SIZEOF_VOID_P == SIZEOF_LONG
-typedef unsigned long UWord;
-typedef long          SWord;
-#define SWORD_CONSTANT(Const) Const##L
-#define UWORD_CONSTANT(Const) Const##UL
-#define ERTS_UWORD_MAX ULONG_MAX
-#define ERTS_SWORD_MAX LONG_MAX
-#elif SIZEOF_VOID_P == SIZEOF_INT
-typedef unsigned int UWord;
-typedef int          SWord;
-#define SWORD_CONSTANT(Const) Const
-#define UWORD_CONSTANT(Const) Const##U
-#define ERTS_UWORD_MAX UINT_MAX
-#define ERTS_SWORD_MAX INT_MAX
-#elif SIZEOF_VOID_P == SIZEOF_LONG_LONG
-typedef unsigned long long UWord;
-typedef long long          SWord;
-#define SWORD_CONSTANT(Const) Const##LL
-#define UWORD_CONSTANT(Const) Const##ULL
-#define ERTS_UWORD_MAX ULLONG_MAX
-#define ERTS_SWORD_MAX LLONG_MAX
-#else
-#error Found no appropriate type to use for 'Eterm', 'Uint' and 'Sint'
-#endif
-
-#else /* !HALFWORD_HEAP */
-
 #if SIZEOF_VOID_P == SIZEOF_LONG
 typedef unsigned long Eterm;
 typedef unsigned long Uint;
@@ -383,8 +337,6 @@ typedef Uint UWord;
 typedef Sint SWord;
 #define ERTS_UINT_MAX ERTS_UWORD_MAX
 
-#endif /* HALFWORD_HEAP */
-
 typedef UWord BeamInstr;
 
 #ifndef HAVE_INT64
@@ -634,7 +586,6 @@ Uint erts_sys_misc_mem_sz(void);
 /* Io constants to erts_print and erts_putc */
 #define ERTS_PRINT_STDERR	(2)
 #define ERTS_PRINT_STDOUT	(1)
-#define ERTS_PRINT_INVALID	(0) /* Don't want to use 0 since CBUF was 0 */
 #define ERTS_PRINT_FILE		(-1)
 #define ERTS_PRINT_SBUF		(-2)
 #define ERTS_PRINT_SNBUF	(-3)
@@ -786,6 +737,7 @@ void erts_sys_main_thread(void);
 extern int erts_sys_prepare_crash_dump(int secs);
 extern void erts_sys_pre_init(void);
 extern void erl_sys_init(void);
+extern void erl_sys_late_init(void);
 extern void erl_sys_args(int *argc, char **argv);
 extern void erl_sys_schedule(int);
 void sys_tty_reset(int);
@@ -1099,7 +1051,6 @@ extern int erts_use_kernel_poll;
 
 #define put_int8(i, s) do {((unsigned char*)(s))[0] = (i) & 0xff;} while (0)
 
-
 /*
  * Use DEBUGF as you would use printf, but use double parentheses:
  *
diff --git a/erts/emulator/beam/utils.c b/erts/emulator/beam/utils.c
index 342e91e983..ef851d840d 100644
--- a/erts/emulator/beam/utils.c
+++ b/erts/emulator/beam/utils.c
@@ -110,7 +110,6 @@ Eterm*
 erts_heap_alloc(Process* p, Uint need, Uint xtra)
 {
     ErlHeapFragment* bp;
-    Eterm* htop;
     Uint n;
 #if defined(DEBUG) || defined(CHECK_FOR_HOLES)
     Uint i;
@@ -156,16 +155,6 @@ erts_heap_alloc(Process* p, Uint need, Uint xtra)
     n--;
 #endif
 
-    /*
-     * When we have created a heap fragment, we are no longer allowed
-     * to store anything more on the heap. 
-     */
-    htop = HEAP_TOP(p);
-    if (htop < HEAP_LIMIT(p)) {
-	*htop = make_pos_bignum_header(HEAP_LIMIT(p)-htop-1);
-	HEAP_TOP(p) = HEAP_LIMIT(p);
-    }
-
     bp->next = MBUF(p);
     MBUF(p) = bp;
     bp->alloc_size = n;
@@ -269,6 +258,31 @@ erl_grow_pstack(ErtsPStack* s, void* default_pstack, unsigned need_bytes)
     s->psp = s->pstart + sp_offs;
 }
 
+/*
+ * Helper function for the EQUEUE macros defined in global.h.
+ */
+
+void
+erl_grow_equeue(ErtsEQueue* q, Eterm* default_equeue)
+{
+    Uint old_size = (q->end - q->start);
+    Uint new_size = old_size * 2;
+    Uint first_part = (q->end - q->front);
+    Uint second_part = (q->back - q->start);
+    Eterm* new_ptr = erts_alloc(q->alloc_type, new_size*sizeof(Eterm));
+    ASSERT(q->back == q->front);   // of course the queue is full now!
+    if (first_part > 0)
+      sys_memcpy(new_ptr, q->front, first_part*sizeof(Eterm));
+    if (second_part > 0)
+      sys_memcpy(new_ptr+first_part, q->start, second_part*sizeof(Eterm));
+    if (q->start != default_equeue)
+      erts_free(q->alloc_type, q->start);
+    q->start = new_ptr;
+    q->end = q->start + new_size;
+    q->front = q->start;
+    q->back = q->start + old_size;
+}
+
 /* CTYPE macros */
 
 #define LATIN1
@@ -399,9 +413,6 @@ erts_print(int to, void *arg, char *format, ...)
 	case ERTS_PRINT_DSBUF:
 	    res = erts_vdsprintf((erts_dsprintf_buf_t *) arg, format, arg_list);
 	    break;
-	case ERTS_PRINT_INVALID:
-	    res = -EINVAL;
-	    break;
 	default:
 	    res = erts_vfdprintf((int) to, format, arg_list);
 	    break;
@@ -844,7 +855,6 @@ Uint32 make_hash(Eterm term_arg)
     Eterm hash = 0;
     unsigned op;
 
-    /* Must not collide with the real tag_val_def's: */
 #define MAKE_HASH_TUPLE_OP      (FIRST_VACANT_TAG_DEF)
 #define MAKE_HASH_TERM_ARRAY_OP (FIRST_VACANT_TAG_DEF+1)
 #define MAKE_HASH_CDR_PRE_OP    (FIRST_VACANT_TAG_DEF+2)
@@ -898,7 +908,7 @@ tail_recur:
 	    Uint y2 = y1 < 0 ? -(Uint)y1 : y1;
 
 	    UINT32_HASH_STEP(y2, FUNNY_NUMBER2);
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
 	    if (y2 >> 32)
 		UINT32_HASH_STEP(y2 >> 32, FUNNY_NUMBER2);
 #endif
@@ -1019,7 +1029,7 @@ tail_recur:
 	    }
 	    d = BIG_DIGIT(ptr, k);
 	    k = sizeof(ErtsDigit);
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
 	    if (!(d >> 32))
 		k /= 2;
 #endif
@@ -1989,7 +1999,7 @@ tail_recur:
 	    (atom_tab(atom_val(term))->slot.bucket.hvalue);
 	break;
     case SMALL_DEF:
-#if defined(ARCH_64) && !HALFWORD_HEAP
+#if defined(ARCH_64)
     {
 	Sint y1 = signed_val(term);
 	Uint y2 = y1 < 0 ? -(Uint)y1 : y1;
@@ -2285,7 +2295,11 @@ static void do_send_logger_message(Eterm *hp, ErlOffHeap *ohp, ErlHeapFragment *
 	erts_queue_error_logger_message(from, message, bp);
     }
 #else
-    erts_queue_message(p, NULL /* only used for smp build */, bp, message, NIL);
+    {
+	ErtsMessage *mp = erts_alloc_message(0, NULL);
+	mp->data.heap_frag = bp;
+	erts_queue_message(p, NULL /* only used for smp build */, mp, message, NIL);
+    }
 #endif
 }
 
@@ -2598,11 +2612,7 @@ erts_destroy_tmp_dsbuf(erts_dsprintf_buf_t *dsbufp)
  * Test for equality of two terms.
  * Returns 0 if not equal, or a non-zero value otherwise.
  */
-#if HALFWORD_HEAP
-int eq_rel(Eterm a, Eterm* a_base, Eterm b, Eterm* b_base)
-#else
 int eq(Eterm a, Eterm b)
-#endif
 {
     DECLARE_WSTACK(stack);
     Sint sz;
@@ -2610,18 +2620,18 @@ int eq(Eterm a, Eterm b)
     Eterm* bb;
 
 tailrecur:
-    if (is_same(a, a_base, b, b_base)) goto pop_next;
+    if (is_same(a, b)) goto pop_next;
 tailrecur_ne:
 
     switch (primary_tag(a)) {
     case TAG_PRIMARY_LIST:
 	if (is_list(b)) {
-	    Eterm* aval = list_val_rel(a, a_base);
-	    Eterm* bval = list_val_rel(b, b_base);
+	    Eterm* aval = list_val(a);
+	    Eterm* bval = list_val(b);
 	    while (1) {
 		Eterm atmp = CAR(aval);
 		Eterm btmp = CAR(bval);
-		if (!is_same(atmp,a_base,btmp,b_base)) {
+		if (!is_same(atmp,btmp)) {
 		    WSTACK_PUSH2(stack,(UWord) CDR(bval),(UWord) CDR(aval));
 		    a = atmp;
 		    b = btmp;
@@ -2629,7 +2639,7 @@ tailrecur_ne:
 		}
 		atmp = CDR(aval);
 		btmp = CDR(bval);
-		if (is_same(atmp,a_base,btmp,b_base)) {
+		if (is_same(atmp,btmp)) {
 		    goto pop_next;
 		}
 		if (is_not_list(atmp) || is_not_list(btmp)) {
@@ -2637,22 +2647,22 @@ tailrecur_ne:
 		    b = btmp;
 		    goto tailrecur_ne;
 		}
-		aval = list_val_rel(atmp, a_base);
-		bval = list_val_rel(btmp, b_base);
+		aval = list_val(atmp);
+		bval = list_val(btmp);
 	    }
 	}
 	break; /* not equal */
 
     case TAG_PRIMARY_BOXED:
 	{	
-	    Eterm hdr = *boxed_val_rel(a,a_base);
+	    Eterm hdr = *boxed_val(a);
 	    switch (hdr & _TAG_HEADER_MASK) {
 	    case ARITYVAL_SUBTAG:
 		{
-		    aa = tuple_val_rel(a, a_base);
-		    if (!is_boxed(b) || *boxed_val_rel(b,b_base) != *aa)
+		    aa = tuple_val(a);
+		    if (!is_boxed(b) || *boxed_val(b) != *aa)
 			goto not_equal;
-		    bb = tuple_val_rel(b,b_base);
+		    bb = tuple_val(b);
 		    if ((sz = arityval(*aa)) == 0) goto pop_next;
 		    ++aa;
 		    ++bb;
@@ -2671,16 +2681,16 @@ tailrecur_ne:
 		    Uint a_bitoffs;
 		    Uint b_bitoffs;
 		    
-		    if (!is_binary_rel(b,b_base)) {
+		    if (!is_binary(b)) {
 			goto not_equal;
 		    }
-		    a_size = binary_size_rel(a,a_base);
-		    b_size = binary_size_rel(b,b_base);
+		    a_size = binary_size(a);
+		    b_size = binary_size(b);
 		    if (a_size != b_size) {
 			goto not_equal;
 		    }
-		    ERTS_GET_BINARY_BYTES_REL(a, a_ptr, a_bitoffs, a_bitsize, a_base);
-		    ERTS_GET_BINARY_BYTES_REL(b, b_ptr, b_bitoffs, b_bitsize, b_base);
+		    ERTS_GET_BINARY_BYTES(a, a_ptr, a_bitoffs, a_bitsize);
+		    ERTS_GET_BINARY_BYTES(b, b_ptr, b_bitoffs, b_bitsize);
 		    if ((a_bitsize | b_bitsize | a_bitoffs | b_bitoffs) == 0) {
 			if (sys_memcmp(a_ptr, b_ptr, a_size) == 0) goto pop_next;
 		    } else if (a_bitsize == b_bitsize) {
@@ -2691,9 +2701,9 @@ tailrecur_ne:
 		}
 	    case EXPORT_SUBTAG:
 		{
-		    if (is_export_rel(b,b_base)) {
-			Export* a_exp = *((Export **) (export_val_rel(a,a_base) + 1));
-			Export* b_exp = *((Export **) (export_val_rel(b,b_base) + 1));
+		    if (is_export(b)) {
+			Export* a_exp = *((Export **) (export_val(a) + 1));
+			Export* b_exp = *((Export **) (export_val(b) + 1));
 			if (a_exp == b_exp) goto pop_next;
 		    }
 		    break; /* not equal */
@@ -2703,10 +2713,10 @@ tailrecur_ne:
 		    ErlFunThing* f1;
 		    ErlFunThing* f2;
   
-		    if (!is_fun_rel(b,b_base))
+		    if (!is_fun(b))
 			goto not_equal;
-		    f1 = (ErlFunThing *) fun_val_rel(a,a_base);
-		    f2 = (ErlFunThing *) fun_val_rel(b,b_base);
+		    f1 = (ErlFunThing *) fun_val(a);
+		    f2 = (ErlFunThing *) fun_val(b);
 		    if (f1->fe->module != f2->fe->module ||
 			f1->fe->old_index != f2->fe->old_index ||
 			f1->fe->old_uniq != f2->fe->old_uniq ||
@@ -2724,15 +2734,15 @@ tailrecur_ne:
 		ExternalThing *ap;
 		ExternalThing *bp;
 
-		if(!is_external_rel(b,b_base))
+		if(!is_external(b))
 		    goto not_equal;
 
-		ap = external_thing_ptr_rel(a,a_base);
-		bp = external_thing_ptr_rel(b,b_base);
+		ap = external_thing_ptr(a);
+		bp = external_thing_ptr(b);
 
 		if(ap->header == bp->header && ap->node == bp->node) {
-		    ASSERT(1 == external_data_words_rel(a,a_base));
-		    ASSERT(1 == external_data_words_rel(b,b_base));
+		    ASSERT(1 == external_data_words(a));
+		    ASSERT(1 == external_data_words(b));
 		    
 		    if (ap->data.ui[0] == bp->data.ui[0]) goto pop_next;
 		}
@@ -2753,11 +2763,11 @@ tailrecur_ne:
 		ExternalThing* athing;
 		ExternalThing* bthing;
 
-		if(!is_external_ref_rel(b,b_base))
+		if(!is_external_ref(b))
 		    goto not_equal;
 
-		athing = external_thing_ptr_rel(a,a_base);
-		bthing = external_thing_ptr_rel(b,b_base);
+		athing = external_thing_ptr(a);
+		bthing = external_thing_ptr(b);
 
 		if(athing->node != bthing->node)
 		    goto not_equal;
@@ -2769,12 +2779,12 @@ tailrecur_ne:
 
 		goto ref_common;
 	    case REF_SUBTAG:
-		    if (!is_internal_ref_rel(b,b_base))
+		    if (!is_internal_ref(b))
 			goto not_equal;
 
 		    {
-			RefThing* athing = ref_thing_ptr_rel(a,a_base);
-			RefThing* bthing = ref_thing_ptr_rel(b,b_base);
+			RefThing* athing = ref_thing_ptr(a);
+			RefThing* bthing = ref_thing_ptr(b);
 			alen = internal_thing_ref_no_of_numbers(athing);
 			blen = internal_thing_ref_no_of_numbers(bthing);
 			anum = internal_thing_ref_numbers(athing);
@@ -2824,10 +2834,10 @@ tailrecur_ne:
 		{
 		    int i;
   
-		    if (!is_big_rel(b,b_base))
+		    if (!is_big(b))
 			goto not_equal;
-		    aa = big_val_rel(a,a_base);
-		    bb = big_val_rel(b,b_base);
+		    aa = big_val(a);
+		    bb = big_val(b);
 		    if (*aa != *bb)
 			goto not_equal;
 		    i = BIG_ARITY(aa);
@@ -2842,19 +2852,19 @@ tailrecur_ne:
 		    FloatDef af;
 		    FloatDef bf;
   
-		    if (is_float_rel(b,b_base)) {
-			GET_DOUBLE_REL(a, af, a_base);
-			GET_DOUBLE_REL(b, bf, b_base);
+		    if (is_float(b)) {
+			GET_DOUBLE(a, af);
+			GET_DOUBLE(b, bf);
 			if (af.fd == bf.fd) goto pop_next;
 		    }
 		    break; /* not equal */
 		}
 	    case MAP_SUBTAG:
-                if (is_flatmap_rel(a, a_base)) {
-		    aa = flatmap_val_rel(a, a_base);
-		    if (!is_boxed(b) || *boxed_val_rel(b,b_base) != *aa)
+                if (is_flatmap(a)) {
+		    aa = flatmap_val(a);
+		    if (!is_boxed(b) || *boxed_val(b) != *aa)
 			goto not_equal;
-		    bb = flatmap_val_rel(b,b_base);
+		    bb = flatmap_val(b);
 		    sz = flatmap_get_size((flatmap_t*)aa);
 
 		    if (sz != flatmap_get_size((flatmap_t*)bb)) goto not_equal;
@@ -2866,11 +2876,11 @@ tailrecur_ne:
 		    goto term_array;
 
                 } else {
-		    if (!is_boxed(b) || *boxed_val_rel(b,b_base) != hdr)
+		    if (!is_boxed(b) || *boxed_val(b) != hdr)
 			goto not_equal;
 
-		    aa = hashmap_val_rel(a, a_base) + 1;
-		    bb = hashmap_val_rel(b, b_base) + 1;
+		    aa = hashmap_val(a) + 1;
+		    bb = hashmap_val(b) + 1;
 		    switch (hdr & _HEADER_MAP_SUBTAG_MASK) {
 		    case HAMT_SUBTAG_HEAD_ARRAY:
 			aa++; bb++;
@@ -2903,7 +2913,7 @@ term_array: /* arrays in 'aa' and 'bb', length in 'sz' */
 	Eterm* bp = bb;
 	Sint i = sz;
 	for (;;) {
-	    if (!is_same(*ap,a_base,*bp,b_base)) break;
+	    if (!is_same(*ap,*bp)) break;
 	    if (--i == 0) goto pop_next;
 	    ++ap;
 	    ++bp;
@@ -2992,7 +3002,6 @@ static int cmp_atoms(Eterm a, Eterm b)
 		    bb->name+3, bb->len-3);
 }
 
-#if !HALFWORD_HEAP
 /* cmp(Eterm a, Eterm b)
  *  For compatibility with HiPE - arith-based compare.
  */
@@ -3000,39 +3009,22 @@ Sint cmp(Eterm a, Eterm b)
 {
     return erts_cmp(a, b, 0, 0);
 }
-#endif
 
-#if HALFWORD_HEAP
-static Sint erts_cmp_compound_rel_opt(Eterm a, Eterm* a_base,
-                                      Eterm b, Eterm* b_base,
-                                      int exact, int eq_only);
-#else
 static Sint erts_cmp_compound(Eterm a, Eterm b, int exact, int eq_only);
-#endif
 
-#if HALFWORD_HEAP
-Sint erts_cmp_rel_opt(Eterm a, Eterm* a_base,
-                      Eterm b, Eterm* b_base,
-                      int exact, int eq_only)
-#else
 Sint erts_cmp(Eterm a, Eterm b, int exact, int eq_only)
-#endif
 {
     if (is_atom(a) && is_atom(b)) {
         return cmp_atoms(a, b);
     } else if (is_both_small(a, b)) {
         return (signed_val(a) - signed_val(b));
-    } else if (is_float_rel(a, a_base) && is_float_rel(b, b_base)) {
+    } else if (is_float(a) && is_float(b)) {
         FloatDef af, bf;
-        GET_DOUBLE_REL(a, af, a_base);
-        GET_DOUBLE_REL(b, bf, b_base);
+        GET_DOUBLE(a, af);
+        GET_DOUBLE(b, bf);
         return float_comp(af.fd, bf.fd);
     }
-#if HALFWORD_HEAP
-    return erts_cmp_compound_rel_opt(a,a_base,b,b_base,exact,eq_only);
-#else
     return erts_cmp_compound(a,b,exact,eq_only);
-#endif
 }
 
 
@@ -3040,13 +3032,7 @@ Sint erts_cmp(Eterm a, Eterm b, int exact, int eq_only)
  * exact = 1 -> term-based compare
  * exact = 0 -> arith-based compare
  */
-#if HALFWORD_HEAP
-static Sint erts_cmp_compound_rel_opt(Eterm a, Eterm* a_base,
-                                      Eterm b, Eterm* b_base,
-                                      int exact, int eq_only)
-#else
 static Sint erts_cmp_compound(Eterm a, Eterm b, int exact, int eq_only)
-#endif
 {
 #define PSTACK_TYPE struct erts_cmp_hashmap_state
     struct erts_cmp_hashmap_state {
@@ -3113,7 +3099,7 @@ static Sint erts_cmp_compound(Eterm a, Eterm b, int exact, int eq_only)
 bodyrecur:
     j = 0;
 tailrecur:
-    if (is_same(a,a_base,b,b_base)) {	/* Equal values or pointers. */
+    if (is_same(a,b)) {	/* Equal values or pointers. */
 	goto pop_next;
     }
 tailrecur_ne:
@@ -3139,9 +3125,9 @@ tailrecur_ne:
 	    if (is_internal_port(b)) {
 		bnode = erts_this_node;
 		bdata = internal_port_data(b);
-	    } else if (is_external_port_rel(b,b_base)) {
-		bnode = external_port_node_rel(b,b_base);
-		bdata = external_port_data_rel(b,b_base);
+	    } else if (is_external_port(b)) {
+		bnode = external_port_node(b);
+		bdata = external_port_data(b);
 	    } else {
 		a_tag = PORT_DEF;
 		goto mixed_types;
@@ -3157,9 +3143,9 @@ tailrecur_ne:
 	    if (is_internal_pid(b)) {
 		bnode = erts_this_node;
 		bdata = internal_pid_data(b);
-	    } else if (is_external_pid_rel(b,b_base)) {
-		bnode = external_pid_node_rel(b,b_base);
-		bdata = external_pid_data_rel(b,b_base);
+	    } else if (is_external_pid(b)) {
+		bnode = external_pid_node(b);
+		bdata = external_pid_data(b);
 	    } else {
 		a_tag = PID_DEF;
 		goto mixed_types;
@@ -3192,12 +3178,12 @@ tailrecur_ne:
 	    a_tag = LIST_DEF;
 	    goto mixed_types;
 	}
-	aa = list_val_rel(a,a_base);
-	bb = list_val_rel(b,b_base);
+	aa = list_val(a);
+	bb = list_val(b);
 	while (1) {
 	    Eterm atmp = CAR(aa);
 	    Eterm btmp = CAR(bb);
-	    if (!is_same(atmp,a_base,btmp,b_base)) {
+	    if (!is_same(atmp,btmp)) {
 		WSTACK_PUSH2(stack,(UWord) CDR(bb),(UWord) CDR(aa));
 		a = atmp;
 		b = btmp;
@@ -3205,7 +3191,7 @@ tailrecur_ne:
 	    }
 	    atmp = CDR(aa);
 	    btmp = CDR(bb);
-	    if (is_same(atmp,a_base,btmp,b_base)) {
+	    if (is_same(atmp,btmp)) {
 		goto pop_next;
 	    }
 	    if (is_not_list(atmp) || is_not_list(btmp)) {
@@ -3213,20 +3199,20 @@ tailrecur_ne:
 		b = btmp;
 		goto tailrecur_ne;
 	    }
-	    aa = list_val_rel(atmp,a_base);
-	    bb = list_val_rel(btmp,b_base);
+	    aa = list_val(atmp);
+	    bb = list_val(btmp);
 	}
     case TAG_PRIMARY_BOXED:
 	{
-	    Eterm ahdr = *boxed_val_rel(a,a_base);
+	    Eterm ahdr = *boxed_val(a);
 	    switch ((ahdr & _TAG_HEADER_MASK) >> _TAG_PRIMARY_SIZE) {
 	    case (_TAG_HEADER_ARITYVAL >> _TAG_PRIMARY_SIZE):
-		if (!is_tuple_rel(b,b_base)) {
+		if (!is_tuple(b)) {
 		    a_tag = TUPLE_DEF;
 		    goto mixed_types;
 		}
-		aa = tuple_val_rel(a,a_base);
-		bb = tuple_val_rel(b,b_base);
+		aa = tuple_val(a);
+		bb = tuple_val(b);
 		/* compare the arities */
 		i = arityval(ahdr);	/* get the arity*/
 		if (i != arityval(*bb)) {
@@ -3242,18 +3228,18 @@ tailrecur_ne:
 		{
                     struct erts_cmp_hashmap_state* sp;
                     if (is_flatmap_header(ahdr)) {
-                        if (!is_flatmap_rel(b,b_base)) {
-                            if (is_hashmap_rel(b,b_base)) {
-                                aa = (Eterm *)flatmap_val_rel(a,a_base);
-                                i = flatmap_get_size((flatmap_t*)aa) - hashmap_size_rel(b,b_base);
+                        if (!is_flatmap(b)) {
+                            if (is_hashmap(b)) {
+                                aa = (Eterm *)flatmap_val(a);
+                                i = flatmap_get_size((flatmap_t*)aa) - hashmap_size(b);
                                 ASSERT(i != 0);
                                 RETURN_NEQ(i);
                             }
                             a_tag = MAP_DEF;
                             goto mixed_types;
                         }
-                        aa = (Eterm *)flatmap_val_rel(a,a_base);
-                        bb = (Eterm *)flatmap_val_rel(b,b_base);
+                        aa = (Eterm *)flatmap_val(a);
+                        bb = (Eterm *)flatmap_val(b);
 
                         i = flatmap_get_size((flatmap_t*)aa);
                         if (i != flatmap_get_size((flatmap_t*)bb)) {
@@ -3280,21 +3266,21 @@ tailrecur_ne:
                             goto bodyrecur;
                         }
                     }
-		    if (!is_hashmap_rel(b,b_base)) {
-                        if (is_flatmap_rel(b,b_base)) {
-                            bb = (Eterm *)flatmap_val_rel(b,b_base);
-                            i = hashmap_size_rel(a,a_base) - flatmap_get_size((flatmap_t*)bb);
+		    if (!is_hashmap(b)) {
+                        if (is_flatmap(b)) {
+                            bb = (Eterm *)flatmap_val(b);
+                            i = hashmap_size(a) - flatmap_get_size((flatmap_t*)bb);
                             ASSERT(i != 0);
                             RETURN_NEQ(i);
                         }
 			a_tag = MAP_DEF;
 			goto mixed_types;
 		    }
-		    i = hashmap_size_rel(a,a_base) - hashmap_size_rel(b,b_base);
+		    i = hashmap_size(a) - hashmap_size(b);
 		    if (i) {
 			RETURN_NEQ(i);
 		    }
-                    if (hashmap_size_rel(a,a_base) == 0) {
+                    if (hashmap_size(a) == 0) {
                         goto pop_next;
                     }
 
@@ -3329,31 +3315,31 @@ tailrecur_ne:
                     goto bodyrecur;
 		}
 	    case (_TAG_HEADER_FLOAT >> _TAG_PRIMARY_SIZE):
-		if (!is_float_rel(b,b_base)) {
+		if (!is_float(b)) {
 		    a_tag = FLOAT_DEF;
 		    goto mixed_types;
 		} else {
 		    FloatDef af;
 		    FloatDef bf; 
 
-		    GET_DOUBLE_REL(a, af, a_base);
-		    GET_DOUBLE_REL(b, bf, b_base);
+		    GET_DOUBLE(a, af);
+		    GET_DOUBLE(b, bf);
 		    ON_CMP_GOTO(float_comp(af.fd, bf.fd));
 		}
 	    case (_TAG_HEADER_POS_BIG >> _TAG_PRIMARY_SIZE):
 	    case (_TAG_HEADER_NEG_BIG >> _TAG_PRIMARY_SIZE):
-		if (!is_big_rel(b,b_base)) {
+		if (!is_big(b)) {
 		    a_tag = BIG_DEF;
 		    goto mixed_types;
 		}
-		ON_CMP_GOTO(big_comp(rterm2wterm(a,a_base), rterm2wterm(b,b_base)));
+		ON_CMP_GOTO(big_comp(a, b));
 	    case (_TAG_HEADER_EXPORT >> _TAG_PRIMARY_SIZE):
-		if (!is_export_rel(b,b_base)) {
+		if (!is_export(b)) {
 		    a_tag = EXPORT_DEF;
 		    goto mixed_types;
 		} else {
-		    Export* a_exp = *((Export **) (export_val_rel(a,a_base) + 1));
-		    Export* b_exp = *((Export **) (export_val_rel(b,b_base) + 1));
+		    Export* a_exp = *((Export **) (export_val(a) + 1));
+		    Export* b_exp = *((Export **) (export_val(b) + 1));
 
 		    if ((j = cmp_atoms(a_exp->code[0], b_exp->code[0])) != 0) {
 			RETURN_NEQ(j);
@@ -3365,12 +3351,12 @@ tailrecur_ne:
 		}
 		break;
 	    case (_TAG_HEADER_FUN >> _TAG_PRIMARY_SIZE):
-		if (!is_fun_rel(b,b_base)) {
+		if (!is_fun(b)) {
 		    a_tag = FUN_DEF;
 		    goto mixed_types;
 		} else {
-		    ErlFunThing* f1 = (ErlFunThing *) fun_val_rel(a,a_base);
-		    ErlFunThing* f2 = (ErlFunThing *) fun_val_rel(b,b_base);
+		    ErlFunThing* f1 = (ErlFunThing *) fun_val(a);
+		    ErlFunThing* f2 = (ErlFunThing *) fun_val(b);
 		    Sint diff;
 
 		    diff = cmpbytes(atom_tab(atom_val(f1->fe->module))->name,
@@ -3402,29 +3388,29 @@ tailrecur_ne:
 		if (is_internal_pid(b)) {
 		    bnode = erts_this_node;
 		    bdata = internal_pid_data(b);
-		} else if (is_external_pid_rel(b,b_base)) {
-		    bnode = external_pid_node_rel(b,b_base);
-		    bdata = external_pid_data_rel(b,b_base);
+		} else if (is_external_pid(b)) {
+		    bnode = external_pid_node(b);
+		    bdata = external_pid_data(b);
 		} else {
 		    a_tag = EXTERNAL_PID_DEF;
 		    goto mixed_types;
 		}
-		anode = external_pid_node_rel(a,a_base);
-		adata = external_pid_data_rel(a,a_base);
+		anode = external_pid_node(a);
+		adata = external_pid_data(a);
 		goto pid_common;
 	    case (_TAG_HEADER_EXTERNAL_PORT >> _TAG_PRIMARY_SIZE):
 		if (is_internal_port(b)) {
 		    bnode = erts_this_node;
 		    bdata = internal_port_data(b);
-		} else if (is_external_port_rel(b,b_base)) {
-		    bnode = external_port_node_rel(b,b_base);
-		    bdata = external_port_data_rel(b,b_base);
+		} else if (is_external_port(b)) {
+		    bnode = external_port_node(b);
+		    bdata = external_port_data(b);
 		} else {
 		    a_tag = EXTERNAL_PORT_DEF;
 		    goto mixed_types;
 		}
-		anode = external_port_node_rel(a,a_base);
-		adata = external_port_data_rel(a,a_base);
+		anode = external_port_node(a);
+		adata = external_port_data(a);
 		goto port_common;
 	    case (_TAG_HEADER_REF >> _TAG_PRIMARY_SIZE):
 		/*
@@ -3432,14 +3418,13 @@ tailrecur_ne:
 		 * (32-bit words), *not* ref data words.
 		 */
 
-		
-		if (is_internal_ref_rel(b,b_base)) {
-		    RefThing* bthing = ref_thing_ptr_rel(b,b_base);
+		if (is_internal_ref(b)) {
+		    RefThing* bthing = ref_thing_ptr(b);
 		    bnode = erts_this_node;
 		    bnum = internal_thing_ref_numbers(bthing);
 		    blen = internal_thing_ref_no_of_numbers(bthing);
-		} else if(is_external_ref_rel(b,b_base)) {
-		    ExternalThing* bthing = external_thing_ptr_rel(b,b_base);
+		} else if(is_external_ref(b)) {
+		    ExternalThing* bthing = external_thing_ptr(b);
 		    bnode = bthing->node;
 		    bnum = external_thing_ref_numbers(bthing);
 		    blen = external_thing_ref_no_of_numbers(bthing);
@@ -3448,7 +3433,7 @@ tailrecur_ne:
 		    goto mixed_types;
 		}
 		{
-		    RefThing* athing = ref_thing_ptr_rel(a,a_base);
+		    RefThing* athing = ref_thing_ptr(a);
 		    anode = erts_this_node;
 		    anum = internal_thing_ref_numbers(athing);
 		    alen = internal_thing_ref_no_of_numbers(athing);
@@ -3481,13 +3466,13 @@ tailrecur_ne:
 			RETURN_NEQ((Sint32) (anum[i] - bnum[i]));
 		goto pop_next;
 	    case (_TAG_HEADER_EXTERNAL_REF >> _TAG_PRIMARY_SIZE):
-		if (is_internal_ref_rel(b,b_base)) {
-		    RefThing* bthing = ref_thing_ptr_rel(b,b_base);
+		if (is_internal_ref(b)) {
+		    RefThing* bthing = ref_thing_ptr(b);
 		    bnode = erts_this_node;
 		    bnum = internal_thing_ref_numbers(bthing);
 		    blen = internal_thing_ref_no_of_numbers(bthing);
-		} else if (is_external_ref_rel(b,b_base)) {
-		    ExternalThing* bthing = external_thing_ptr_rel(b,b_base);
+		} else if (is_external_ref(b)) {
+		    ExternalThing* bthing = external_thing_ptr(b);
 		    bnode = bthing->node;
 		    bnum = external_thing_ref_numbers(bthing);
 		    blen = external_thing_ref_no_of_numbers(bthing);
@@ -3496,7 +3481,7 @@ tailrecur_ne:
 		    goto mixed_types;
 		}
 		{
-		    ExternalThing* athing = external_thing_ptr_rel(a,a_base);
+		    ExternalThing* athing = external_thing_ptr(a);
 		    anode = athing->node;
 		    anum = external_thing_ref_numbers(athing);
 		    alen = external_thing_ref_no_of_numbers(athing);
@@ -3504,13 +3489,13 @@ tailrecur_ne:
 		goto ref_common;
 	    default:
 		/* Must be a binary */
-		ASSERT(is_binary_rel(a,a_base));
-		if (!is_binary_rel(b,b_base)) {
+		ASSERT(is_binary(a));
+		if (!is_binary(b)) {
 		    a_tag = BINARY_DEF;
 		    goto mixed_types;
 		} else {
-		    Uint a_size = binary_size_rel(a,a_base);
-		    Uint b_size = binary_size_rel(b,b_base);
+		    Uint a_size = binary_size(a);
+		    Uint b_size = binary_size(b);
 		    Uint a_bitsize;
 		    Uint b_bitsize;
 		    Uint a_bitoffs;
@@ -3519,8 +3504,8 @@ tailrecur_ne:
 		    int cmp;
 		    byte* a_ptr;
 		    byte* b_ptr;
-		    ERTS_GET_BINARY_BYTES_REL(a, a_ptr, a_bitoffs, a_bitsize, a_base);
-		    ERTS_GET_BINARY_BYTES_REL(b, b_ptr, b_bitoffs, b_bitsize, b_base);
+		    ERTS_GET_BINARY_BYTES(a, a_ptr, a_bitoffs, a_bitsize);
+		    ERTS_GET_BINARY_BYTES(b, b_ptr, b_bitoffs, b_bitsize);
 		    if ((a_bitsize | b_bitsize | a_bitoffs | b_bitoffs) == 0) {
 			min_size = (a_size < b_size) ? a_size : b_size;
 			if ((cmp = sys_memcmp(a_ptr, b_ptr, min_size)) != 0) {
@@ -3551,13 +3536,8 @@ tailrecur_ne:
     {
 	FloatDef f1, f2;
 	Eterm big;
-#if HALFWORD_HEAP
-	Wterm aw = is_immed(a) ? a : rterm2wterm(a,a_base);
-	Wterm bw = is_immed(b) ? b : rterm2wterm(b,b_base);
-#else
 	Eterm aw = a;
 	Eterm bw = b;
-#endif
 #define MAX_LOSSLESS_FLOAT ((double)((1LL << 53) - 2))
 #define MIN_LOSSLESS_FLOAT ((double)(((1LL << 53) - 2)*-1))
 #define BIG_ARITY_FLOAT_MAX (1024 / D_EXP) /* arity of max float as a bignum */
@@ -3629,7 +3609,7 @@ tailrecur_ne:
 		}
 	    } else {
 		big = double_to_big(f2.fd, big_buf, sizeof(big_buf)/sizeof(Eterm));
-		j = big_comp(aw, rterm2wterm(big,big_buf));
+		j = big_comp(aw, big);
 	    }
 	    if (_NUMBER_CODE(a_tag, b_tag) == FLOAT_BIG) {
 		j = -j;
@@ -3677,7 +3657,7 @@ term_array: /* arrays in 'aa' and 'bb', length in 'i' */
     while (--i) {
 	a = *aa++;
 	b = *bb++;
-	if (!is_same(a,a_base, b,b_base)) {
+	if (!is_same(a, b)) {
 	    if (is_atom(a) && is_atom(b)) {
 		if ((j = cmp_atoms(a, b)) != 0) {
 		    goto not_equal;
@@ -4369,7 +4349,7 @@ iolist_size(const int yield_support, ErtsIOListState *state, Eterm obj, ErlDrvSi
 {
     int res, init_yield_count, yield_count;
     Eterm* objp;
-    Uint size = (Uint) *sizep; /* Intentionally Uint due to halfword heap */
+    Uint size = (Uint) *sizep;
     DECLARE_ESTACK(s);
 
     if (!yield_support)
diff --git a/erts/emulator/drivers/common/efile_drv.c b/erts/emulator/drivers/common/efile_drv.c
index 8aff6c1865..a82abc3e9f 100644
--- a/erts/emulator/drivers/common/efile_drv.c
+++ b/erts/emulator/drivers/common/efile_drv.c
@@ -101,15 +101,9 @@
 #  include "config.h"
 #endif
 
-#ifndef __OSE__
 #include <ctype.h>
 #include <sys/types.h>
 #include <stdlib.h>
-#else
-#include "ctype.h"
-#include "sys/types.h"
-#include "stdlib.h"
-#endif
 
 /* Need (NON)BLOCKING macros for sendfile */
 #ifndef WANT_NONBLOCKING
@@ -895,7 +889,7 @@ static void reply_Uint_posix_error(file_descriptor *desc, Uint num,
     TRACE_C('N');
 
     response[0] = FILE_RESP_NUMERR;
-#if SIZEOF_VOID_P == 4 || HALFWORD_HEAP
+#if SIZEOF_VOID_P == 4
     put_int32(0, response+1);
 #else
     put_int32(num>>32, response+1);
@@ -964,7 +958,7 @@ static int reply_Uint(file_descriptor *desc, Uint result) {
     TRACE_C('R');
 
     tmp[0] = FILE_RESP_NUMBER;
-#if SIZEOF_VOID_P == 4 || HALFWORD_HEAP
+#if SIZEOF_VOID_P == 4
     put_int32(0, tmp+1);
 #else
     put_int32(result>>32, tmp+1);
@@ -2581,7 +2575,6 @@ file_async_ready(ErlDrvData e, ErlDrvThreadData data)
       case FILE_CLOSE_ON_PORT_EXIT:
 	  /* See file_stop. However this is never invoked after the port is killed. */
 	  free_data(data);
-	  EF_FREE(desc);
 	  desc = NULL;
 	  /* This is it for this port, so just send dtrace and return, avoid doing anything to the freed data */
 	  DTRACE6(efile_drv_return, sched_i1, sched_i2, sched_utag,
diff --git a/erts/emulator/drivers/common/inet_drv.c b/erts/emulator/drivers/common/inet_drv.c
index 549de6503c..e3f5060117 100644
--- a/erts/emulator/drivers/common/inet_drv.c
+++ b/erts/emulator/drivers/common/inet_drv.c
@@ -1,7 +1,7 @@
 /*
  * %CopyrightBegin%
  *
- * Copyright Ericsson AB 1997-2013. All Rights Reserved.
+ * Copyright Ericsson AB 1997-2015. All Rights Reserved.
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
@@ -94,10 +94,6 @@ typedef unsigned long long llu_t;
 #define INT16_MAX (32767)
 #endif
 
-#ifdef __OSE__
-#include "inet.h"
-#endif
-
 #ifdef __WIN32__
 #define  STRNCASECMP strncasecmp
 
@@ -298,139 +294,7 @@ static unsigned long one_value = 1;
 #define TCP_SHUT_RD    SD_RECEIVE
 #define TCP_SHUT_RDWR  SD_BOTH
 
-#elif defined (__OSE__)
-
-/*
- * Some notes about how inet (currently only tcp) works on OSE.
- * The driver uses OSE signals to communicate with the one_inet
- * process. Because of the difference in how signals and file descriptors
- * work the whole select/deselect mechanic is very different.
- * In ose when a sock_select is done a function is called. That function
- * notes the changes that the driver want to do, but does not act on it.
- * later when the function returns the new desired state is compared
- * to the previous state and the apprioriate actions are taken. The action
- * is usually to either request more data from the stack or stop requesting
- * data.
- *
- * One thing to note is that the driver never does select/deselect. It always
- * listens for the signals. Flow of data is regulated by sending or not sending
- * signals to the ose inet process.
- *
- * The interesting functions to look at are:
- *  * inet_driver_select : called when sock_select is called
- *  * tcp_inet_ose_dispatch_signal : checks state changes and sends new signals
- *  * tcp_inet_drv_output_ose : ready output callback, reads signals and calls
- *                              dispatch_signal
- *  * tcp_inet_drv_input_ose : ready input callback.
- */
-
-#include "efs.h"
-#include "sys/socket.h"
-#include "sys/uio.h"
-#include "sfk/sys/sfk_uio.h"
-#include "netinet/in.h"
-#include "netinet/tcp.h"
-#include "netdb.h"
-#include "ose_spi/socket.sig"
-
-
-static ssize_t writev_fallback(int fd, const struct iovec *iov, int iovcnt, int max_sz);
-
-#define INVALID_SOCKET -1
-#define INVALID_EVENT  -1
-#define SOCKET_ERROR   -1
-
-#define SOCKET int
-#define HANDLE int
-#define FD_READ         ERL_DRV_READ
-#define FD_WRITE        ERL_DRV_WRITE
-#define FD_CLOSE        0
-#define FD_CONNECT      (1<<4)
-#define FD_ACCEPT       (1<<5)
-#define SOCK_FD_ERROR   (1<<6)
-
-#define sock_connect(s, addr, len)  connect((s), (addr), (len))
-#define sock_listen(s, b)           listen((s), (b))
-#define sock_bind(s, addr, len)     bind((s), (addr), (len))
-#define sock_getopt(s,t,n,v,l)      getsockopt((s),(t),(n),(v),(l))
-#define sock_setopt(s,t,n,v,l)      setsockopt((s),(t),(n),(v),(l))
-#define sock_name(s, addr, len)     getsockname((s), (addr), (len))
-#define sock_peer(s, addr, len)     getpeername((s), (addr), (len))
-#define sock_ntohs(x)               ntohs((x))
-#define sock_ntohl(x)               ntohl((x))
-#define sock_htons(x)               htons((x))
-#define sock_htonl(x)               htonl((x))
-
-#define sock_accept(s, addr, len)   accept((s), (addr), (len))
-#define sock_send(s,buf,len,flag)   inet_send((s),(buf),(len),(flag))
-#define sock_sendto(s,buf,blen,flag,addr,alen) \
-	    sendto((s),(buf),(blen),(flag),(addr),(alen))
-#define sock_sendv(s, vec, size, np, flag) \
-	    (*(np) = writev_fallback((s), (struct iovec*)(vec), (size), (*(np))))
-#define sock_sendmsg(s,msghdr,flag) sendmsg((s),(msghdr),(flag))
-
-#define sock_open(af, type, proto)  socket((af), (type), (proto))
-#define sock_close(s)               close((s))
-#define sock_dup(s)                 dup((s))
-#define sock_shutdown(s, how)       shutdown((s), (how))
-
-#define sock_hostname(buf, len)     gethostname((buf), (len))
-#define sock_getservbyname(name,proto) getservbyname((name), (proto))
-#define sock_getservbyport(port,proto) getservbyport((port), (proto))
-
-#define sock_recv(s,buf,len,flag)   recv((s),(buf),(len),(flag))
-#define sock_recvfrom(s,buf,blen,flag,addr,alen) \
-                recvfrom((s),(buf),(blen),(flag),(addr),(alen))
-#define sock_recvmsg(s,msghdr,flag) recvmsg((s),(msghdr),(flag))
-
-#define sock_errno()                errno
-#define sock_create_event(d)        ((d)->s) /* return file descriptor */
-#define sock_close_event(e)                  /* do nothing */
-
-#ifndef WANT_NONBLOCKING
-#define WANT_NONBLOCKING
-#endif
-#include "sys.h"
-
-typedef unsigned long  u_long;
-#define  IN_CLASSA(a)      ((((in_addr_t)(a)) & 0x80000000) == 0)
-#define  IN_CLASSA_NET     0xff000000
-#define  IN_CLASSA_NSHIFT  24
-#define  IN_CLASSA_HOST    (0xffffffff & ~IN_CLASSA_NET)
-#define  IN_CLASSA_MAX     128
-
-#define  IN_CLASSB(a)      ((((in_addr_t)(a)) & 0xc0000000) == 0x80000000)
-#define  IN_CLASSB_NET     0xffff0000
-#define  IN_CLASSB_NSHIFT  16
-#define  IN_CLASSB_HOST    (0xffffffff & ~IN_CLASSB_NET)
-#define  IN_CLASSB_MAX     65536
-
-#define  IN_CLASSC(a)      ((((in_addr_t)(a)) & 0xe0000000) == 0xc0000000)
-#define  IN_CLASSC_NET     0xffffff00
-#define  IN_CLASSC_NSHIFT  8
-#define  IN_CLASSC_HOST    (0xffffffff & ~IN_CLASSC_NET)
-
-#define  IN_CLASSD(a)      ((((in_addr_t)(a)) & 0xf0000000) == 0xe0000000)
-#define  IN_MULTICAST(a)      IN_CLASSD(a)
-
-#define  IN_EXPERIMENTAL(a)   ((((in_addr_t)(a)) & 0xe0000000) == 0xe0000000)
-#define  IN_BADCLASS(a)    ((((in_addr_t)(a)) & 0xf0000000) == 0xf0000000)
-
-#define sock_select(d, flags, onoff) do { \
-        ASSERT(!(d)->is_ignored); \
-        (d)->event_mask = (onoff) ? \
-                 ((d)->event_mask | (flags)) : \
-                 ((d)->event_mask & ~(flags)); \
-        DEBUGF(("(%s / %d) sock_select(%ld): flags=%02X, onoff=%d, event_mask=%02lX, s=%d\r\n", \
-		__FILE__, __LINE__, (long) (d)->port, (flags), (onoff), (unsigned long) (d)->event_mask, (d)->s)); \
-        inet_driver_select((d), (flags), (onoff)); \
-   } while(0)
-
-#define TCP_SHUT_WR    SHUT_WR
-#define TCP_SHUT_RD    SHUT_RD
-#define TCP_SHUT_RDWR  SHUT_RDWR
-
-#else /* !__OSE__ && !__WIN32__ */
+#else /* !__WIN32__ */
 
 #include <sys/time.h>
 #ifdef NETDB_H_NEEDS_IN_H
@@ -704,7 +568,7 @@ static int my_strncasecmp(const char *s1, const char *s2, size_t n)
 #define TCP_SHUT_RD    SHUT_RD
 #define TCP_SHUT_RDWR  SHUT_RDWR
 
-#endif /* !__WIN32__ && !__OSE__ */
+#endif /* !__WIN32__ */
 
 #ifdef HAVE_SOCKLEN_T
 #  define SOCKLEN_T socklen_t
@@ -885,6 +749,7 @@ static int my_strncasecmp(const char *s1, const char *s2, size_t n)
 #define INET_LOPT_MSGQ_LOWTRMRK     37  /* set local msgq low watermark */
 #define INET_LOPT_NETNS             38  /* Network namespace pathname */
 #define INET_LOPT_TCP_SHOW_ECONNRESET 39  /* tell user about incoming RST */
+#define INET_LOPT_LINE_DELIM        40  /* Line delimiting char */
 /* SCTP options: a separate range, from 100: */
 #define SCTP_OPT_RTOINFO		100
 #define SCTP_OPT_ASSOCINFO		101
@@ -1154,6 +1019,7 @@ typedef struct {
 #else
     Uint32        send_oct[2];  /* number of octets sent, 64 bits */
 #endif
+    char          delimiter;    /* Line delimiting character (def: '\n')  */
     unsigned long send_cnt;     /* number of packets sent */
     unsigned long send_max;     /* maximum packet send */
     double send_avg;            /* average packet size sent */
@@ -1166,13 +1032,6 @@ typedef struct {
     char *netns;                /* Socket network namespace name
 				   as full file path */
 #endif
-#ifdef __OSE__
-    int select_state;           /* state to keep track of whether we
-				   should trigger another read/write
-				   request at end of ready_input/output */
-    ErlDrvEvent events[6];
-#endif
-
 } inet_descriptor;
 
 
@@ -1188,10 +1047,8 @@ static void tcp_inet_stop(ErlDrvData);
 static void tcp_inet_command(ErlDrvData, char*, ErlDrvSizeT);
 static void tcp_inet_commandv(ErlDrvData, ErlIOVec*);
 static void tcp_inet_flush(ErlDrvData drv_data);
-#ifndef __OSE__
 static void tcp_inet_drv_input(ErlDrvData, ErlDrvEvent);
 static void tcp_inet_drv_output(ErlDrvData data, ErlDrvEvent event);
-#endif
 static ErlDrvData tcp_inet_start(ErlDrvPort, char* command);
 static ErlDrvSSizeT tcp_inet_ctl(ErlDrvData, unsigned int,
 				 char*, ErlDrvSizeT, char**, ErlDrvSizeT);
@@ -1204,71 +1061,6 @@ static void tcp_inet_event(ErlDrvData, ErlDrvEvent);
 static void find_dynamic_functions(void);
 #endif
 
-#ifdef __OSE__
-/* The structure of the signal used for requesting asynchronous 
- * notification from the stack. Under normal circumstances the network stack 
- * shouldn't overwrite the value set in the fd field by the sender 
- * of the request */
-struct OseAsyncSig {
-    struct FmEvent event;
-    int fd;
-};
-
-union SIGNAL {
-    SIGSELECT signo;
-    struct OseAsyncSig async;
-};
-
-static ErlDrvSSizeT tcp_inet_ctl_ose(ErlDrvData e, unsigned int cmd,
-				     char* buf, ErlDrvSizeT len,
-				     char** rbuf, ErlDrvSizeT rsize);
-static void tcp_inet_commandv_ose(ErlDrvData e, ErlIOVec* ev);
-static void tcp_inet_drv_output_ose(ErlDrvData data, ErlDrvEvent event);
-static void tcp_inet_drv_input_ose(ErlDrvData data, ErlDrvEvent event);
-static ErlDrvOseEventId inet_resolve_signal(union SIGNAL *sig);
-
-#ifdef INET_DRV_DEBUG
-
-static char *read_req = "SO_EVENT_READ_REQUEST";
-static char *read_rep = "SO_EVENT_READ_REPLY";
-static char *write_req = "SO_EVENT_WRITE_REQUEST";
-static char *write_rep = "SO_EVENT_WRITE_REPLY";
-static char *eof_req = "SO_EVENT_EOF_REQUEST";
-static char *eof_rep = "SO_EVENT_EOF_REPLY";
-static char *accept_req = "SO_EVENT_ACCEPT_REQUEST";
-static char *accept_rep = "SO_EVENT_ACCEPT_REPLY";
-static char *connect_req = "SO_EVENT_CONNECT_REQUEST";
-static char *connect_rep = "SO_EVENT_CONNECT_REPLY";
-static char *error_req = "SO_EVENT_ERROR_REQUEST";
-static char *error_rep = "SO_EVENT_ERROR_REPLY";
-static char signo_tmp[32];
-
-static char *signo_to_string(SIGSELECT signo) {
-  switch (signo) {
-  case SO_EVENT_READ_REQUEST: { return read_req; }
-  case SO_EVENT_READ_REPLY: { return read_rep; }
-  case SO_EVENT_WRITE_REQUEST: { return write_req; }
-  case SO_EVENT_WRITE_REPLY: { return write_rep; }
-  case SO_EVENT_EOF_REQUEST: { return eof_req; }
-  case SO_EVENT_EOF_REPLY: { return eof_rep; }
-  case SO_EVENT_ACCEPT_REQUEST: { return accept_req; }
-  case SO_EVENT_ACCEPT_REPLY: { return accept_rep; }
-  case SO_EVENT_CONNECT_REQUEST: { return connect_req; }
-  case SO_EVENT_CONNECT_REPLY: { return connect_rep; }
-  case SO_EVENT_ERROR_REQUEST: { return error_req; }
-  case SO_EVENT_ERROR_REPLY: { return error_rep; }
-  }
-
-  snprintf(signo_tmp,32,"0x%x",signo);
-
-  return signo_tmp;
-}
-
-#endif
-
-#endif /* __OSE__ */
-
-
 static struct erl_drv_entry tcp_inet_driver_entry =
 {
     tcp_inet_init,  /* inet_init will add this driver !! */
@@ -1278,9 +1070,6 @@ static struct erl_drv_entry tcp_inet_driver_entry =
 #ifdef __WIN32__
     tcp_inet_event,
     NULL,
-#elif defined(__OSE__)
-    tcp_inet_drv_input_ose, /*ready_input*/
-    tcp_inet_drv_output_ose, /*ready_output*/
 #else
     tcp_inet_drv_input,
     tcp_inet_drv_output,
@@ -1288,17 +1077,9 @@ static struct erl_drv_entry tcp_inet_driver_entry =
     "tcp_inet",
     NULL,
     NULL,
-#ifdef __OSE__
-    tcp_inet_ctl_ose,
-#else
     tcp_inet_ctl,
-#endif
     tcp_inet_timeout,
-#ifdef __OSE__
-    tcp_inet_commandv_ose,
-#else
     tcp_inet_commandv,
-#endif
     NULL,
     tcp_inet_flush,
     NULL,
@@ -1450,14 +1231,6 @@ static int packet_inet_output(udp_descriptor* udesc, HANDLE event);
 /* convert descriptor pointer to inet_descriptor pointer */
 #define INETP(d) (&(d)->inet)
 
-#ifdef __OSE__
-static void inet_driver_select(inet_descriptor* desc,
-			       int flags, int onoff);
-static void tcp_inet_ose_dispatch_signals(tcp_descriptor *desc,
-					  int prev_select_state,
-					  union SIGNAL *sig);
-#endif
-
 static int async_ref = 0;          /* async reference id generator */
 #define NEW_ASYNC_ID() ((async_ref++) & 0xffff)
 
@@ -2749,7 +2522,7 @@ int ssl_tls_inetdrv(void* arg, unsigned type, unsigned major, unsigned minor,
 {
     tcp_descriptor* desc = (tcp_descriptor*) arg;
     int i = 0;
-    ErlDrvTermData spec[28];
+    ErlDrvTermData spec[30];
     ErlDrvTermData caller = ERL_DRV_NIL;
     ErlDrvBinary* bin;
     int ret;
@@ -2790,11 +2563,11 @@ int ssl_tls_inetdrv(void* arg, unsigned type, unsigned major, unsigned minor,
     if (desc->inet.active == INET_PASSIVE) {
         i = LOAD_TUPLE(spec, i, 2);
         i = LOAD_TUPLE(spec, i, 4);
-        ASSERT(i <= 28);
+        ASSERT(i <= sizeof(spec)/sizeof(*spec));
         ret = erl_drv_send_term(desc->inet.dport, caller, spec, i);
     }
     else {
-        ASSERT(i <= 28);
+        ASSERT(i <= sizeof(spec)/sizeof(*spec));
         ret = erl_drv_output_term(desc->inet.dport, spec, i);
     }
 done:
@@ -4353,16 +4126,6 @@ static void desc_close(inet_descriptor* desc)
 	desc->forced_events = 0;
 	desc->send_would_block = 0;
 #endif
-#ifdef __OSE__
-	if (desc->events[0]) {
-	  driver_select(desc->port,desc->events[0],FD_READ|FD_WRITE|ERL_DRV_USE,0);
-	  driver_select(desc->port,desc->events[1],FD_READ|FD_WRITE|ERL_DRV_USE,0);
-	  driver_select(desc->port,desc->events[2],FD_READ|FD_WRITE|ERL_DRV_USE,0);
-	  driver_select(desc->port,desc->events[3],FD_READ|FD_WRITE|ERL_DRV_USE,0);
-	  driver_select(desc->port,desc->events[4],FD_READ|FD_WRITE|ERL_DRV_USE,0);
-	  driver_select(desc->port,desc->events[5],FD_READ|FD_WRITE|ERL_DRV_USE,0);
-	}
-#else
 	/*
 	 * We should close the fd here, but the other driver might still
 	 * be selecting on it.
@@ -4372,7 +4135,6 @@ static void desc_close(inet_descriptor* desc)
 			  ERL_DRV_USE, 0);
 	else
 	  inet_stop_select((ErlDrvEvent)(long)desc->event,NULL);
-#endif
 	desc->event = INVALID_EVENT; /* closed by stop_select callback */
 	desc->s = INVALID_SOCKET;
 	desc->event_mask = 0;
@@ -4414,65 +4176,6 @@ static int erl_inet_close(inet_descriptor* desc)
     return 0;
 }
 
-#ifdef __OSE__
-static void inet_select_init(inet_descriptor* desc)
-{
-    desc->events[0] =
-        erl_drv_ose_event_alloc(SO_EVENT_READ_REPLY,
-                desc->s,
-                inet_resolve_signal,
-                NULL);
-    driver_select(desc->port, desc->events[0],
-            ERL_DRV_READ|ERL_DRV_USE, 1);
-
-    desc->events[1] =
-        erl_drv_ose_event_alloc(SO_EVENT_EOF_REPLY,
-                desc->s,
-                inet_resolve_signal,
-                NULL);
-    driver_select(desc->port, desc->events[1],
-            ERL_DRV_READ|ERL_DRV_USE, 1);
-
-    desc->events[2] =
-        erl_drv_ose_event_alloc(SO_EVENT_ACCEPT_REPLY,
-                desc->s,
-                inet_resolve_signal,
-                NULL);
-    driver_select(desc->port, desc->events[2],
-            ERL_DRV_READ|ERL_DRV_USE, 1);
-
-    /* trigger tcp_inet_input */
-    desc->events[3] =
-        erl_drv_ose_event_alloc(SO_EVENT_WRITE_REPLY,
-                desc->s,
-                inet_resolve_signal,
-                NULL);
-    driver_select(desc->port, desc->events[3],
-            ERL_DRV_WRITE|ERL_DRV_USE, 1);
-
-    desc->events[4] =
-        erl_drv_ose_event_alloc(SO_EVENT_CONNECT_REPLY,
-                desc->s,
-                inet_resolve_signal,
-                NULL);
-    driver_select(desc->port, desc->events[4],
-            ERL_DRV_WRITE|ERL_DRV_USE, 1);
-
-    desc->events[5] =
-        erl_drv_ose_event_alloc(SO_EVENT_ERROR_REPLY,
-                desc->s,
-                inet_resolve_signal,
-                NULL);
-    driver_select(desc->port, desc->events[5],
-            ERL_DRV_WRITE|ERL_DRV_USE, 1);
-
-    /* Issue a select on error event before any other select to be sure we are
-       prepared to receive error notifications from the stack, even in the
-       situations when select isn't issued */
-    sock_select(desc, SOCK_FD_ERROR, 1);
-}
-#endif
-
 static ErlDrvSSizeT inet_ctl_open(inet_descriptor* desc, int domain, int type,
 				  char** rbuf, ErlDrvSizeT rsize)
 {
@@ -4555,9 +4258,6 @@ static ErlDrvSSizeT inet_ctl_open(inet_descriptor* desc, int domain, int type,
 #ifdef __WIN32__
     driver_select(desc->port, desc->event, ERL_DRV_READ, 1);
 #endif
-#ifdef __OSE__
-    inet_select_init(desc);
-#endif
 
     desc->state = INET_STATE_OPEN;
     desc->stype = type;
@@ -4581,13 +4281,7 @@ static ErlDrvSSizeT inet_ctl_fdopen(inet_descriptor* desc, int domain, int type,
         if (name.sa.sa_family != domain)
             return ctl_error(EINVAL, rbuf, rsize);
     }
-#ifdef __OSE__        
-    /* for fdopen duplicating the sd will allow to uniquely identify
-       the signal from OSE with erlang port */
-    desc->s = sock_dup(s);    
-#else
     desc->s = s;
-#endif
 
     if ((desc->event = sock_create_event(desc)) == INVALID_EVENT)
 	return ctl_error(sock_errno(), rbuf, rsize);
@@ -4605,12 +4299,6 @@ static ErlDrvSSizeT inet_ctl_fdopen(inet_descriptor* desc, int domain, int type,
 	sz = sizeof(name);
 	if (!IS_SOCKET_ERROR(sock_peer(s, (struct sockaddr*) &name, &sz))) {
 	    desc->state = INET_STATE_CONNECTED;
-#ifdef __OSE__
-            /* since we are dealing with different descriptors (i.e. inet and
-               socket) the select part should be initialized with the right
-               values */
-            inet_select_init(desc);
-#endif
         }
     }
 
@@ -6074,9 +5762,9 @@ static int inet_set_opts(inet_descriptor* desc, char* ptr, int len)
     int arg_sz;
     enum PacketParseType old_htype = desc->htype;
     int old_active = desc->active;
-    int propagate = 0; /* Set to 1 if failure to set this option
-			  should be propagated to erlang (not all 
-			  errors can be propagated for BC reasons) */
+    int propagate; /* Set to 1 if failure to set this option
+		      should be propagated to erlang (not all
+		      errors can be propagated for BC reasons) */
     int res;
 #ifdef HAVE_SCTP
     /* SCTP sockets are treated completely separately: */
@@ -6093,6 +5781,7 @@ static int inet_set_opts(inet_descriptor* desc, char* ptr, int len)
 	arg_ptr = (char*) &ival;
 	arg_sz = sizeof(ival);
 	proto = SOL_SOCKET;
+	propagate = 0;
 
 	switch(opt) {
 	case INET_LOPT_HEADER:
@@ -6276,6 +5965,12 @@ static int inet_set_opts(inet_descriptor* desc, char* ptr, int len)
 	    }
 	    continue;
 
+	case INET_LOPT_LINE_DELIM:
+	    DEBUGF(("inet_set_opts(%ld): s=%d, LINE_DELIM=%d\r\n",
+		    (long)desc->port, desc->s, ival));
+	    desc->delimiter = (char)ival;
+	    continue;
+
 	case INET_OPT_REUSEADDR: 
 #ifdef __WIN32__
 	    continue;  /* Bjorn says */
@@ -8371,6 +8066,7 @@ static ErlDrvData inet_start(ErlDrvPort port, int size, int protocol)
     desc->deliver = INET_DELIVER_TERM; /* standard term format */
     desc->active  = INET_PASSIVE;      /* start passive */
     desc->active_count = 0;
+    desc->delimiter    = '\n';         /* line delimiting char */
     desc->oph = NULL;
     desc->opt = NULL;
 
@@ -8404,15 +8100,6 @@ static ErlDrvData inet_start(ErlDrvPort port, int size, int protocol)
 #ifdef HAVE_SETNS
     desc->netns = NULL;
 #endif
-#ifdef __OSE__
-    desc->select_state = 0;
-    desc->events[0] = NULL;
-    desc->events[1] = NULL;
-    desc->events[2] = NULL;
-    desc->events[3] = NULL;
-    desc->events[4] = NULL;
-    desc->events[5] = NULL;
-#endif
 
     return (ErlDrvData)desc;
 }
@@ -9139,10 +8826,6 @@ static tcp_descriptor* tcp_inet_copy(tcp_descriptor* desc,SOCKET s,
     copy_desc->inet.port = port;
     copy_desc->inet.dport = driver_mk_port(port);
 
-#ifdef __OSE__
-    inet_select_init(&copy_desc->inet);
-#endif
-
     *err = 0;
     return copy_desc;
 }
@@ -9204,23 +8887,6 @@ static void tcp_inet_stop(ErlDrvData e)
     inet_stop(INETP(desc));
 }
 
-#ifdef __OSE__
-
-static ErlDrvSSizeT tcp_inet_ctl_ose(ErlDrvData e, unsigned int cmd,
-				     char* buf, ErlDrvSizeT len,
-				     char** rbuf, ErlDrvSizeT rsize) {
-
-  tcp_descriptor* desc = (tcp_descriptor*)e;
-  int prev_select_state = INETP(desc)->select_state;
-
-  ErlDrvSSizeT res = tcp_inet_ctl(e,cmd,buf,len,rbuf,rsize);
-
-  tcp_inet_ose_dispatch_signals((tcp_descriptor*)e,prev_select_state,NULL);
-
-  return res;
-}
-#endif
-
 /* TCP requests from Erlang */
 static ErlDrvSSizeT tcp_inet_ctl(ErlDrvData e, unsigned int cmd,
 				 char* buf, ErlDrvSizeT len,
@@ -9662,17 +9328,6 @@ static void tcp_inet_command(ErlDrvData e, char *buf, ErlDrvSizeT len)
     DEBUGF(("tcp_inet_command(%ld) }\r\n", (long)desc->inet.port)); 
 }
 
-#ifdef __OSE__
-
-static void tcp_inet_commandv_ose(ErlDrvData e, ErlIOVec* ev) {
-  int prev_select_state = INETP((tcp_descriptor*)e)->select_state;
-  tcp_inet_commandv(e, ev);
-  tcp_inet_ose_dispatch_signals((tcp_descriptor*)e,prev_select_state,NULL);
-}
-
-#endif
-
-
 static void tcp_inet_commandv(ErlDrvData e, ErlIOVec* ev)
 {
     tcp_descriptor* desc = (tcp_descriptor*)e;
@@ -9745,22 +9400,6 @@ static void inet_stop_select(ErlDrvEvent event, void* _)
 {
 #ifdef __WIN32__
     WSACloseEvent((HANDLE)event);
-#elif defined(__OSE__)
-    ErlDrvOseEventId id;
-    union SIGNAL *sig;
-    erl_drv_ose_event_fetch(event, NULL, &id,NULL);
-    DEBUGF(("inet_stop_select(?#?) {s=%d\n",id));
-    sock_close((int)id);
-    /* On socket close all the signals waiting to be processed as part of the
-       select should be deallocated */
-    while((sig = erl_drv_ose_get_signal(event))) {
-      DEBUGF(("inet_stop_select(?#?): Freeing signal %s\n",
-	      signo_to_string(sig->signo)));
-      free_buf(&sig);
-    }
-    erl_drv_ose_event_free(event);
-    DEBUGF(("inet_stop_select(?#?) }\n"));
-      
 #else
     sock_close((SOCKET)(long)event);
 #endif
@@ -9882,7 +9521,7 @@ static int tcp_remain(tcp_descriptor* desc, int* len)
 
     tlen = packet_get_length(desc->inet.htype, ptr, n, 
                              desc->inet.psize, desc->i_bufsz,
-                             &desc->http_state);
+                             desc->inet.delimiter, &desc->http_state);
 
     DEBUGF(("tcp_remain(%ld): s=%d, n=%d, nfill=%d nsz=%d, tlen %d\r\n",
 	    (long)desc->inet.port, desc->inet.s, n, nfill, nsz, tlen));
@@ -10309,146 +9948,7 @@ static void tcp_inet_event(ErlDrvData e, ErlDrvEvent event)
     return;
 }
 
-#elif defined(__OSE__) /* !__WIN32__ */
-/* The specific resolve signal function. It will return the socket descriptor
-   for which the select was issued */
-static ErlDrvOseEventId inet_resolve_signal(union SIGNAL *sig) {
-  DEBUGF(("%s(?#?): s=%d got signal %s, status = %d, extra = %d, sender = 0x%x\n",
-	  __FUNCTION__,sig->async.fd,signo_to_string(sig->signo),
-	  sig->async.event.status,
-	  sig->async.event.extra,sender(&sig)));
-  if (sig->signo == SO_EVENT_READ_REPLY ||
-      sig->signo == SO_EVENT_ACCEPT_REPLY ||
-      sig->signo == SO_EVENT_EOF_REPLY ||
-      sig->signo == SO_EVENT_WRITE_REPLY ||
-      sig->signo == SO_EVENT_ERROR_REPLY ||
-      sig->signo == SO_EVENT_CONNECT_REPLY ) {
-    return sig->async.fd;
-  }
-
-  return -1;
-}
-
-static void inet_driver_select(inet_descriptor* desc,
-			       int flags, int onoff) {
-    ASSERT(!desc->is_ignored);
-
-    if(onoff) {
-      desc->select_state |= flags;
-    } else {
-      desc->select_state &= ~flags;
-    }
-}
-
-static ssize_t writev_fallback(int fd, const struct iovec *iov, int iovcnt, int max_sz)
-{
-    size_t data_len = 0;
-    size_t sent = 0;
-    ssize_t n;
-    int i;
-
-    for(i = 0; i < iovcnt; i++)
-    {
-        data_len = iov[i].iov_len;
-tryagain:        
-        n = sock_send(fd, iov[i].iov_base, data_len, 0);
-        if (IS_SOCKET_ERROR(n)) {
-            /* If buffer length is greater than the amount stack is able to
-             * send out then try to send at least max_sz (this comes with
-             * SO_EVENT_WRITE_REPLY signal*/
-            if ((errno == EMSGSIZE) && (max_sz > 0) && (data_len > max_sz)) {
-                data_len = max_sz;
-                goto tryagain;
-            }
-            break;
-        }
-        sent += n;
-    }
-    return sent;
-}
-
-#define OSE_EVENT_REQ(TCP_DESC,EVENT) do {				\
-    union SIGNAL *sig = alloc(sizeof(struct OseAsyncSig), EVENT);	\
-    sig->async.fd = INETP(TCP_DESC)->s;					\
-    ose_request_event(INETP(TCP_DESC)->s, &sig, 1);			\
-    DEBUGF(("%s(%ld): s=%d sent %s\r\n",__FUNCTION__,			\
-      INETP(TCP_DESC)->port,INETP(TCP_DESC)->s,signo_to_string(EVENT))); \
-  } while(0)
-
-static void tcp_inet_ose_dispatch_signals(tcp_descriptor *desc,
-					  int prev_select_state,
-					  union SIGNAL *sig) {
-  if (sig) {
-    DEBUGF(("tcp_inet_ose_dispatch_signals(%ld) {s=%d resend\r\n",
-	    (long)INETP(desc)->port,INETP(desc)->s));
-    /* We are reacting to a signal, which means that if
-       the select_state for that signal is still activated
-       we should send a new signal */
-    switch (sig->signo) {
-    case SO_EVENT_READ_REPLY: {
-      if (INETP(desc)->select_state & FD_READ)
-	OSE_EVENT_REQ(desc,SO_EVENT_READ_REQUEST);
-      break;
-    }
-    case SO_EVENT_WRITE_REPLY: {
-      if (INETP(desc)->select_state & FD_WRITE)
-	OSE_EVENT_REQ(desc,SO_EVENT_WRITE_REQUEST);
-      break;
-    }
-    case SO_EVENT_CONNECT_REPLY: {
-      if (INETP(desc)->select_state & FD_CONNECT)
-	OSE_EVENT_REQ(desc,SO_EVENT_CONNECT_REQUEST);
-      break;
-    }
-    case SO_EVENT_ACCEPT_REPLY: {
-      if (INETP(desc)->select_state & FD_ACCEPT)
-	OSE_EVENT_REQ(desc,SO_EVENT_ACCEPT_REQUEST);
-      break;
-    }
-    case SO_EVENT_ERROR_REPLY: {
-      if (INETP(desc)->select_state & SOCK_FD_ERROR)
-	OSE_EVENT_REQ(desc,SO_EVENT_ERROR_REQUEST);
-      break;
-    }
-
-    }
-    DEBUGF(("tcp_inet_ose_dispatch_signals(%ld) }\r\n",
-	    (long)INETP(desc)->port));
-  }
-
-  if (INETP(desc)->select_state != prev_select_state) {
-    /* If the select state has changed we have to issue signals for
-       the state parts that have changed. */
-    int xor_select_state = INETP(desc)->select_state ^ prev_select_state;
-    DEBUGF(("tcp_inet_ose_dispatch_signals(%ld) {s=%d select change\r\n",
-	    (long)INETP(desc)->port,INETP(desc)->s));
-    if ((xor_select_state & FD_READ) &&
-	(INETP(desc)->select_state & FD_READ)) {
-      OSE_EVENT_REQ(desc,SO_EVENT_READ_REQUEST);
-    }
-    if ((xor_select_state & FD_WRITE) &&
-	(INETP(desc)->select_state & FD_WRITE)) {
-      OSE_EVENT_REQ(desc,SO_EVENT_WRITE_REQUEST);
-    }
-    if ((xor_select_state & FD_CONNECT) &&
-	(INETP(desc)->select_state & FD_CONNECT)) {
-      OSE_EVENT_REQ(desc,SO_EVENT_CONNECT_REQUEST);
-    }
-    if ((xor_select_state & FD_ACCEPT) &&
-	(INETP(desc)->select_state & FD_ACCEPT)) {
-      OSE_EVENT_REQ(desc,SO_EVENT_ACCEPT_REQUEST);
-    }
-    if ((xor_select_state & SOCK_FD_ERROR) &&
-	(INETP(desc)->select_state & SOCK_FD_ERROR)) {
-      OSE_EVENT_REQ(desc,SO_EVENT_ERROR_REQUEST);
-    }
-
-    DEBUGF(("tcp_inet_ose_dispatch_signals(%ld) }\r\n",
-	    (long)INETP(desc)->port));
-  }
-}
-
-#endif /* __OSE__ */
+#endif /* __WIN32__ */
 
 
 /* socket has input:
@@ -10935,49 +10435,6 @@ static void tcp_shutdown_async(tcp_descriptor* desc)
 	desc->tcp_add_flags |= TCP_ADDF_SHUTDOWN_WR_DONE;
 }
 
-#ifdef __OSE__
-
-static void tcp_inet_drv_output_ose(ErlDrvData data, ErlDrvEvent event)
-{
-  union SIGNAL *event_sig = erl_drv_ose_get_signal(event);
-
-  while (event_sig) {
-      int prev_select_state = INETP((tcp_descriptor*)data)->select_state;
-      int res = tcp_inet_output((tcp_descriptor*)data, (HANDLE)event_sig);
-      if (res != -1) {
-	tcp_inet_ose_dispatch_signals((tcp_descriptor*)data,
-				      prev_select_state,event_sig);
-	free_buf(&event_sig);
-	event_sig = erl_drv_ose_get_signal(event);
-      } else {
-	/* NOTE: here the event object could have been deallocated!!!!
-	   inet_stop_select is called when doing driver_select(ERL_DRV_USE,0)
-	 */
-	free_buf(&event_sig);
-	return;
-      }
-  }
-}
-
-static void tcp_inet_drv_input_ose(ErlDrvData data, ErlDrvEvent event)
-{
-  union SIGNAL *event_sig = erl_drv_ose_get_signal(event);
-
-  while (event_sig) {
-    int prev_select_state = INETP((tcp_descriptor*)data)->select_state;
-    int res = tcp_inet_input((tcp_descriptor*)data, (HANDLE)event);
-    if (res != -1) {
-      tcp_inet_ose_dispatch_signals((tcp_descriptor*)data, prev_select_state,
-				    event_sig);
-      free_buf(&event_sig);
-      event_sig = erl_drv_ose_get_signal(event);
-    } else {
-      free_buf(&event_sig);
-      return;
-    }
-  }
-}
-#else
 static void tcp_inet_drv_output(ErlDrvData data, ErlDrvEvent event)
 {
     (void)tcp_inet_output((tcp_descriptor*)data, (HANDLE)event);
@@ -10987,7 +10444,6 @@ static void tcp_inet_drv_input(ErlDrvData data, ErlDrvEvent event)
 {
     (void)tcp_inet_input((tcp_descriptor*)data, (HANDLE)event);
 }
-#endif
 
 /* socket ready for ouput:
 ** 1. INET_STATE_CONNECTING => non block connect ?
@@ -11053,13 +10509,6 @@ static int tcp_inet_output(tcp_descriptor* desc, HANDLE event)
 	    ssize_t n;
 	    SysIOVec* iov;
 
-#ifdef __OSE__
-            /* For large size buffers case the amount of data that the stack is
-               able to send out (received in the .extra field) should be passed
-               down to writev_fallback */
-            n = event ? ((union SIGNAL*)event)->async.event.extra : 0;
-#endif
-
 	    if ((iov = driver_peekq(ix, &vsize)) == NULL) {
 		sock_select(INETP(desc), FD_WRITE, 0);
 		send_empty_out_q_msgs(INETP(desc));
@@ -11087,12 +10536,6 @@ static int tcp_inet_output(tcp_descriptor* desc, HANDLE event)
 				    sizes > (max 32 bit signed int) */
 	      size_t howmuch = 0x7FFFFFFF; /* max signed 32 bit */
 	      int x;
-#ifdef __OSE__
-              /* For EWOULDBLOCK sock_sendv returns 0 so we have to be sure it
-                 wasn't the case */
-              if(sock_errno() == ERRNO_BLOCK) 
-                    goto done;
-#endif
 	      for(x = 0; x < vsize && iov[x].iov_len == 0; ++x)
 		;
 	      if (x < vsize) {
diff --git a/erts/emulator/drivers/ose/ose_efile.c b/erts/emulator/drivers/ose/ose_efile.c
deleted file mode 100644
index c8337a95d5..0000000000
--- a/erts/emulator/drivers/ose/ose_efile.c
+++ /dev/null
@@ -1,1125 +0,0 @@
-/*
- * %CopyrightBegin%
- *
- * Copyright Ericsson AB 1997-2012. All Rights Reserved.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- *
- * %CopyrightEnd%
- */
-/*
- * Purpose: Provides file and directory operations for OSE.
- */
-#ifdef HAVE_CONFIG_H
-#  include "config.h"
-#endif
-#if defined(HAVE_POSIX_FALLOCATE) && !defined(__sun) && !defined(__sun__)
-#define _XOPEN_SOURCE 600
-#endif
-#if !defined(_GNU_SOURCE) && defined(HAVE_LINUX_FALLOC_H)
-#define _GNU_SOURCE
-#endif
-#include "sys.h"
-#include "erl_driver.h"
-#include "erl_efile.h"
-#if defined(DARWIN) || defined(HAVE_LINUX_FALLOC_H) || defined(HAVE_POSIX_FALLOCATE)
-#include "fcntl.h"
-#endif
-#include "ose.h"
-#include "unistd.h"
-#include "sys/stat.h"
-#include "dirent.h"
-#include "sys/time.h"
-#include "time.h"
-#include "assert.h"
-
-/* Find a definition of MAXIOV, that is used in the code later. */
-#if defined IOV_MAX
-#define MAXIOV IOV_MAX
-#elif defined UIO_MAXIOV
-#define MAXIOV UIO_MAXIOV
-#else
-#define MAXIOV 16
-#endif
-
-/*
- * Macros for testing file types.
- */
-
-#define ISDIR(st) (((st).st_mode & S_IFMT) == S_IFDIR)
-#define ISREG(st) (((st).st_mode & S_IFMT) == S_IFREG)
-#define ISDEV(st) \
-    (((st).st_mode&S_IFMT) == S_IFCHR || ((st).st_mode&S_IFMT) == S_IFBLK)
-#define ISLNK(st) (((st).st_mode & S_IFLNK) == S_IFLNK)
-#ifdef NO_UMASK
-#define FILE_MODE 0644
-#define DIR_MODE  0755
-#else
-#define FILE_MODE 0666
-#define DIR_MODE  0777
-#endif
-
-#define IS_DOT_OR_DOTDOT(s) \
-    (s[0] == '.' && (s[1] == '\0' || (s[1] == '.' && s[2] == '\0')))
-
-/*
- * Macros for handling local file descriptors
- * and mutexes.
- *
- * Handling of files like this is necessary because OSE
- * does not allow seeking after the end of a file. So
- * what we do it emulate this by keeping track of the size
- * of the file and where the file's positions is. If a
- * write happens after eof then we pad it.
- *
- * Given time this should be rewritten to get rid of the
- * mutex and use the port lock to protect the data. This
- * could be done be done by adapting the efile api for some
- * calls to allow some meta-data to be associated with the
- * open file.
- */
-
-#define L_FD_IS_VALID(fd_data) ((fd_data)->beyond_eof > 0)
-#define L_FD_INVALIDATE(fd_data) (fd_data)->beyond_eof = 0
-#define L_FD_CUR(fd_data) (fd_data)->pos
-#define L_FD_OFFS_BEYOND_EOF(fd_data, offs) \
-    (((fd_data)->size > offs) ? 0 : 1)
-
-#define L_FD_FAIL -1
-#define L_FD_SUCCESS 1
-#define L_FD_PAD_SIZE 255
-
-struct fd_meta {
-    ErlDrvMutex *meta_mtx;
-    struct fd_data *fd_data_list;
-};
-
-struct fd_data {
-    int fd;
-    struct fd_data *next;
-    struct fd_data *prev;
-    int pos;
-    int beyond_eof;
-    size_t size;
-#ifdef DEBUG
-    PROCESS owner;
-#endif
-};
-
-static int l_invalidate_local_fd(int fd);
-static int l_pad_file(struct fd_data *fd_data, off_t offset);
-static int check_error(int result, Efile_error* errInfo);
-static struct fd_data* l_new_fd(void);
-static int l_remove_local_fd(int fd);
-static struct fd_data* l_find_local_fd(int fd);
-static int l_update_local_fd(int fd, int pos, int size);
-
-static struct fd_meta* fdm = NULL;
-
-
-/***************************************************************************/
-
-static int
-l_remove_local_fd(int fd)
-{
-    struct fd_data *fd_data;
-    fd_data = l_find_local_fd(fd);
-
-    if (fd_data == NULL) {
-        return L_FD_FAIL;
-    }
-#ifdef DEBUG
-    assert(fd_data->owner == current_process());
-#endif
-    erl_drv_mutex_lock(fdm->meta_mtx);
-    /* head ? */
-    if (fd_data == fdm->fd_data_list) {
-        if (fd_data->next != NULL) {
-            /* remove link to head */
-            fd_data->next->prev = NULL;
-            /* set new head */
-            fdm->fd_data_list = fd_data->next;
-        }
-        else {
-            /* head is lonely */
-            fdm->fd_data_list = NULL;
-        }
-    }
-    else { /* not head */
-        if (fd_data->prev == NULL) {
-            erl_drv_mutex_unlock(fdm->meta_mtx);
-            return L_FD_FAIL;
-        }
-        else {
-            if (fd_data->next != NULL) {
-                fd_data->next->prev = fd_data->prev;
-                fd_data->prev->next = fd_data->next;
-            }
-            else {
-                fd_data->prev->next = NULL;
-            }
-        }
-    }
-
-    /* scramble values */
-    fd_data->beyond_eof = -1;
-    fd_data->next = NULL;
-    fd_data->prev = NULL;
-    fd_data->fd = -1;
-
-    /* unlock and clean */
-    driver_free(fd_data);
-    erl_drv_mutex_unlock(fdm->meta_mtx);
-
-    return L_FD_SUCCESS;
-}
-
-/***************************************************************************/
-
-static int
-l_invalidate_local_fd(int fd) {
-    struct fd_data *fd_data;
-
-    if ((fd_data = l_find_local_fd(fd)) == NULL) {
-        return L_FD_FAIL;
-    }
-
-    fd_data->beyond_eof = 0;
-    return L_FD_SUCCESS;
-}
-
-/****************************************************************************/
-
-static struct fd_data*
-l_find_local_fd(int fd) {
-    struct fd_data *fd_data;
-
-    fd_data = NULL;
-    erl_drv_mutex_lock(fdm->meta_mtx);
-    for (fd_data = fdm->fd_data_list; fd_data != NULL; ) {
-        if (fd_data->fd == fd) {
-#ifdef DEBUG
-            assert(fd_data->owner == current_process());
-#endif
-            break;
-        }
-        fd_data = fd_data->next;
-    }
-    erl_drv_mutex_unlock(fdm->meta_mtx);
-    return fd_data;
-}
-
-/***************************************************************************/
-
-static struct fd_data*
-l_new_fd(void) {
-    struct fd_data *fd_data;
-
-    fd_data = driver_alloc(sizeof(struct fd_data));
-    if (fd_data == NULL) {
-        return NULL;
-    }
-    erl_drv_mutex_lock(fdm->meta_mtx);
-    if (fdm->fd_data_list == NULL) {
-        fdm->fd_data_list = fd_data;
-        fdm->fd_data_list->prev = NULL;
-        fdm->fd_data_list->next = NULL;
-    }
-    else {
-        fd_data->next = fdm->fd_data_list;
-        fdm->fd_data_list = fd_data;
-        fdm->fd_data_list->prev = NULL;
-    }
-#ifdef DEBUG
-    fd_data->owner = current_process();
-#endif
-    erl_drv_mutex_unlock(fdm->meta_mtx);
-    return fd_data;
-}
-
-/***************************************************************************/
-
-static int
-l_update_local_fd(int fd, int pos, int size) {
-    struct fd_data *fd_data = NULL;
-
-    fd_data = l_find_local_fd(fd);
-    /* new fd to handle? */
-    if (fd_data == NULL) {
-        fd_data = l_new_fd();
-        if (fd_data == NULL) {
-            /* out of memory */
-            return L_FD_FAIL;
-        }
-    }
-    fd_data->size = size;
-    fd_data->pos = pos;
-    fd_data->fd = fd;
-    fd_data->beyond_eof = 1;
-
-    return L_FD_SUCCESS;
-}
-
-/***************************************************************************/
-
-static int
-l_pad_file(struct fd_data *fd_data, off_t offset) {
-    int size_dif;
-    int written = 0;
-    int ret_val = L_FD_SUCCESS;
-    char padding[L_FD_PAD_SIZE];
-
-    size_dif = (offset - fd_data->size);
-    memset(&padding, '\0', L_FD_PAD_SIZE);
-
-    while (size_dif > 0) {
-        written = write(fd_data->fd, padding,
-                (size_dif < L_FD_PAD_SIZE) ?
-                size_dif : L_FD_PAD_SIZE);
-        if (written < 0 && errno != EINTR && errno != EAGAIN) {
-            ret_val = -1;
-            break;
-        }
-        size_dif -= written;
-    }
-    L_FD_INVALIDATE(fd_data);
-    return ret_val;
-}
-
-/***************************************************************************/
-
-static int
-check_error(int result, Efile_error *errInfo) {
-    if (result < 0) {
-        errInfo->posix_errno = errInfo->os_errno = errno;
-        return 0;
-    }
-    return 1;
-}
-
-/***************************************************************************/
-
-int
-efile_init() {
-    fdm = driver_alloc(sizeof(struct fd_meta));
-    if (fdm == NULL) {
-        return L_FD_FAIL;
-    }
-    fdm->meta_mtx = erl_drv_mutex_create("ose_efile local fd mutex\n");
-    erl_drv_mutex_lock(fdm->meta_mtx);
-    fdm->fd_data_list = NULL;
-    erl_drv_mutex_unlock(fdm->meta_mtx);
-    return L_FD_SUCCESS;
-}
-
-/***************************************************************************/
-
-int
-efile_mkdir(Efile_error* errInfo,       /* Where to return error codes. */
-        char* name)                 /* Name of directory to create. */
-{
-#ifdef NO_MKDIR_MODE
-    return check_error(mkdir(name), errInfo);
-#else
-    int res = mkdir(name, DIR_MODE);
-    if (res < 0 && errno == EINVAL) {
-      errno = ENOENT;
-    }
-    return check_error(res, errInfo);
-#endif
-}
-
-/***************************************************************************/
-
-int
-efile_rmdir(Efile_error* errInfo,       /* Where to return error codes. */
-        char* name)                 /* Name of directory to delete. */
-{
-    if (rmdir(name) == 0) {
-        return 1;
-    }
-    if (errno == ENOTEMPTY) {
-        errno = EEXIST;
-    }
-    if (errno == EEXIST || errno == EINVAL) {
-        int saved_errno = errno;
-        struct stat file_stat;
-        struct stat cwd_stat;
-
-        if(stat(name, &file_stat) != 0) {
-            errno = ENOENT;
-            return check_error(-1, errInfo);
-        }
-        /*
-         *  The error code might be wrong if this is the current directory.
-         */
-        if (stat(name, &file_stat) == 0 && stat(".", &cwd_stat) == 0 &&
-                file_stat.st_ino == cwd_stat.st_ino &&
-                file_stat.st_dev == cwd_stat.st_dev) {
-            saved_errno = EACCES;
-        }
-        errno = saved_errno;
-    }
-    return check_error(-1, errInfo);
-}
-
-/***************************************************************************/
-
-int
-efile_delete_file(Efile_error* errInfo, /* Where to return error codes. */
-        char* name)           /* Name of file to delete. */
-{
-    struct stat statbuf;
-
-    if (stat(name, &statbuf) >= 0) {
-        /* Do not let unlink() remove directories */
-        if (ISDIR(statbuf)) {
-            errno = EPERM;
-            return check_error(-1, errInfo);
-        }
-
-        if (unlink(name) == 0) {
-            return 1;
-        }
-
-        if (errno == EISDIR) {
-            errno = EPERM;
-            return check_error(-1, errInfo);
-        }
-    }
-    else {
-        if (errno == EINVAL) {
-            errno = ENOENT;
-            return check_error(-1, errInfo);
-        }
-    }
-    return check_error(-1, errInfo);
-}
-
-/*
- *---------------------------------------------------------------------------
- *
- *      Changes the name of an existing file or directory, from src to dst.
- *      If src and dst refer to the same file or directory, does nothing
- *      and returns success.  Otherwise if dst already exists, it will be
- *      deleted and replaced by src subject to the following conditions:
- *          If src is a directory, dst may be an empty directory.
- *          If src is a file, dst may be a file.
- *      In any other situation where dst already exists, the rename will
- *      fail.
- *
- * Results:
- *      If the directory was successfully created, returns 1.
- *      Otherwise the return value is 0 and errno is set to
- *      indicate the error.  Some possible values for errno are:
- *
- *      EACCES:     src or dst parent directory can't be read and/or written.
- *      EEXIST:     dst is a non-empty directory.
- *      EINVAL:     src is a root directory or dst is a subdirectory of src.
- *      EISDIR:     dst is a directory, but src is not.
- *      ENOENT:     src doesn't exist, or src or dst is "".
- *      ENOTDIR:    src is a directory, but dst is not.
- *      EXDEV:      src and dst are on different filesystems.
- *
- * Side effects:
- *      The implementation of rename may allow cross-filesystem renames,
- *      but the caller should be prepared to emulate it with copy and
- *      delete if errno is EXDEV.
- *
- *---------------------------------------------------------------------------
- */
-
-int
-efile_rename(Efile_error* errInfo,      /* Where to return error codes. */
-        char* src,                 /* Original name. */
-        char* dst)                 /* New name. */
-{
-
-    /* temporary fix AFM does not recognize ./<file name>
-     * in destination remove pending on adaption of AFM fix
-     */
-
-    char *dot_str;
-    if (dst != NULL) {
-        dot_str = strchr(dst, '.');
-        if (dot_str && dot_str == dst && dot_str[1] == '/') {
-            dst = dst+2;
-        }
-    }
-
-    if (rename(src, dst) == 0) {
-        return 1;
-    }
-    if (errno == ENOTEMPTY) {
-        errno = EEXIST;
-    }
-    if (errno == EINVAL) {
-        struct stat file_stat;
-
-        if (stat(dst, &file_stat)== 0) {
-            if (ISDIR(file_stat)) {
-                errno = EISDIR;
-            }
-            else if (ISREG(file_stat)) {
-                errno = ENOTDIR;
-            }
-            else {
-                errno = EINVAL;
-            }
-        }
-        else {
-            errno = EINVAL;
-        }
-    }
-
-    if (strcmp(src, "/") == 0) {
-        errno = EINVAL;
-    }
-    return check_error(-1, errInfo);
-}
-
-/***************************************************************************/
-
-int
-efile_chdir(Efile_error* errInfo,   /* Where to return error codes. */
-        char* name)             /* Name of directory to make current. */
-{
-    return check_error(chdir(name), errInfo);
-}
-
-/***************************************************************************/
-
-int
-efile_getdcwd(Efile_error* errInfo,     /* Where to return error codes. */
-        int drive,                /* 0 - current, 1 - A, 2 - B etc. */
-        char* buffer,             /* Where to return the current
-                                     directory. */
-        size_t size)              /* Size of buffer. */
-{
-    if (drive == 0) {
-        if (getcwd(buffer, size) == NULL)
-            return check_error(-1, errInfo);
-
-        return 1;
-    }
-
-    /*
-     * Drives other than 0 is not supported on Unix.
-     */
-
-    errno = ENOTSUP;
-    return check_error(-1, errInfo);
-}
-
-/***************************************************************************/
-
-int
-efile_readdir(Efile_error* errInfo,     /* Where to return error codes. */
-        char* name,               /* Name of directory to open. */
-        EFILE_DIR_HANDLE* p_dir_handle,   /* Pointer to directory
-                                             handle of
-                                             open directory.*/
-        char* buffer,             /* Pointer to buffer for
-                                     one filename. */
-        size_t *size)             /* in-out Size of buffer, length
-                                     of name. */
-{
-    DIR *dp;                    /* Pointer to directory structure. */
-    struct dirent* dirp;        /* Pointer to directory entry. */
-
-    /*
-     * If this is the first call, we must open the directory.
-     */
-
-    if (*p_dir_handle == NULL) {
-        dp = opendir(name);
-        if (dp == NULL)
-            return check_error(-1, errInfo);
-        *p_dir_handle = (EFILE_DIR_HANDLE) dp;
-    }
-
-    /*
-     * Retrieve the name of the next file using the directory handle.
-     */
-
-    dp = *((DIR **)((void *)p_dir_handle));
-    for (;;) {
-        dirp = readdir(dp);
-        if (dirp == NULL) {
-            closedir(dp);
-            return 0;
-        }
-        if (IS_DOT_OR_DOTDOT(dirp->d_name))
-            continue;
-        buffer[0] = '\0';
-        strncat(buffer, dirp->d_name, (*size)-1);
-        *size = strlen(dirp->d_name);
-        return 1;
-    }
-}
-
-/***************************************************************************/
-
-int
-efile_openfile(Efile_error* errInfo,    /* Where to return error codes. */
-        char* name,              /* Name of directory to open. */
-        int flags,               /* Flags to user for opening. */
-        int* pfd,                /* Where to store the file
-                                    descriptor. */
-        Sint64 *pSize)           /* Where to store the size of the
-                                    file. */
-{
-    struct stat statbuf;
-    int fd;
-    int mode;                   /* Open mode. */
-
-    if (stat(name, &statbuf) >= 0 && !ISREG(statbuf)) {
-        errno = EISDIR;
-        return check_error(-1, errInfo);
-    }
-
-    switch (flags & (EFILE_MODE_READ|EFILE_MODE_WRITE)) {
-        case EFILE_MODE_READ:
-            mode = O_RDONLY;
-            break;
-        case EFILE_MODE_WRITE:
-            if (flags & EFILE_NO_TRUNCATE)
-                mode = O_WRONLY | O_CREAT;
-            else
-                mode = O_WRONLY | O_CREAT | O_TRUNC;
-            break;
-        case EFILE_MODE_READ_WRITE:
-            mode = O_RDWR | O_CREAT;
-            break;
-        default:
-            errno = EINVAL;
-            return check_error(-1, errInfo);
-    }
-
-
-    if (flags & EFILE_MODE_APPEND) {
-        mode &= ~O_TRUNC;
-        mode |= O_APPEND;
-    }
-
-    if (flags & EFILE_MODE_EXCL) {
-        mode |= O_EXCL;
-    }
-
-    fd = open(name, mode, FILE_MODE);
-
-    if (!check_error(fd, errInfo))
-        return 0;
-
-    *pfd = fd;
-    if (pSize) {
-        *pSize = statbuf.st_size;
-    }
-    return 1;
-}
-
-/***************************************************************************/
-
-int
-efile_may_openfile(Efile_error* errInfo, char *name) {
-    struct stat statbuf;        /* Information about the file */
-    int result;
-
-    result = stat(name, &statbuf);
-    if (!check_error(result, errInfo))
-        return 0;
-    if (!ISREG(statbuf)) {
-        errno = EISDIR;
-        return check_error(-1, errInfo);
-    }
-    return 1;
-}
-
-/***************************************************************************/
-
-void
-efile_closefile(int fd)
-{
-    if (l_find_local_fd(fd) != NULL) {
-        l_remove_local_fd(fd);
-    }
-    close(fd);
-}
-
-/***************************************************************************/
-
-int
-efile_fdatasync(Efile_error *errInfo, /* Where to return error codes. */
-        int fd)               /* File descriptor for file to sync data. */
-{
-    return efile_fsync(errInfo, fd);
-}
-
-/***************************************************************************/
-
-int
-efile_fsync(Efile_error *errInfo, /* Where to return error codes. */
-        int fd)               /* File descriptor for file to sync. */
-{
-    return check_error(fsync(fd), errInfo);
-}
-
-/***************************************************************************/
-
-int
-efile_fileinfo(Efile_error* errInfo, Efile_info* pInfo,
-        char* name, int info_for_link)
-{
-    struct stat statbuf;        /* Information about the file */
-    int result;
-
-    result = stat(name, &statbuf);
-    if (!check_error(result, errInfo)) {
-        return 0;
-    }
-
-#if SIZEOF_OFF_T == 4
-    pInfo->size_high = 0;
-#else
-    pInfo->size_high = (Uint32)(statbuf.st_size >> 32);
-#endif
-    pInfo->size_low = (Uint32)statbuf.st_size;
-
-#ifdef NO_ACCESS
-    /* Just look at read/write access for owner. */
-
-    pInfo->access = ((statbuf.st_mode >> 6) & 07) >> 1;
-
-#else
-    pInfo->access = FA_NONE;
-    if (access(name, R_OK) == 0)
-        pInfo->access |= FA_READ;
-    if (access(name, W_OK) == 0)
-        pInfo->access |= FA_WRITE;
-
-#endif
-
-    if (ISDEV(statbuf))
-        pInfo->type = FT_DEVICE;
-    else if (ISDIR(statbuf))
-        pInfo->type = FT_DIRECTORY;
-    else if (ISREG(statbuf))
-        pInfo->type = FT_REGULAR;
-    else if (ISLNK(statbuf))
-        pInfo->type = FT_SYMLINK;
-    else
-        pInfo->type = FT_OTHER;
-
-    pInfo->accessTime   = statbuf.st_atime;
-    pInfo->modifyTime   = statbuf.st_mtime;
-    pInfo->cTime        = statbuf.st_ctime;
-
-    pInfo->mode         = statbuf.st_mode;
-    pInfo->links        = statbuf.st_nlink;
-    pInfo->major_device = statbuf.st_dev;
-    pInfo->inode        = statbuf.st_ino;
-    pInfo->uid          = statbuf.st_uid;
-    pInfo->gid          = statbuf.st_gid;
-
-    return 1;
-}
-
-/***************************************************************************/
-
-int
-efile_write_info(Efile_error *errInfo, Efile_info *pInfo, char *name)
-{
-    /*
-     * On some systems chown will always fail for a non-root user unless
-     * POSIX_CHOWN_RESTRICTED is not set.  Others will succeed as long as
-     * you don't try to chown a file to someone besides youself.
-     */
-    if (pInfo->mode != -1) {
-        mode_t newMode = pInfo->mode & (S_ISUID | S_ISGID |
-                S_IRWXU | S_IRWXG | S_IRWXO);
-        if (chmod(name, newMode)) {
-            newMode &= ~(S_ISUID | S_ISGID);
-            if (chmod(name, newMode)) {
-                return check_error(-1, errInfo);
-            }
-        }
-    }
-
-    return 1;
-}
-
-/***************************************************************************/
-
-int
-efile_write(Efile_error* errInfo,       /* Where to return error codes. */
-        int flags,                  /* Flags given when file was
-                                       opened. */
-        int fd,                     /* File descriptor to write to. */
-        char* buf,                  /* Buffer to write. */
-        size_t count)               /* Number of bytes to write. */
-{
-    ssize_t written;                    /* Bytes written in last operation. */
-    struct fd_data *fd_data;
-
-    if ((fd_data = l_find_local_fd(fd)) != NULL) {
-        if (L_FD_IS_VALID(fd_data)) {
-            /* we are beyond eof and need to pad*/
-            if (l_pad_file(fd_data, L_FD_CUR(fd_data)) < 0) {
-                return check_error(-1, errInfo);
-            }
-        }
-    }
-
-    while (count > 0)  {
-        if ((written = write(fd, buf, count)) < 0) {
-            if (errno != EINTR) {
-                return check_error(-1, errInfo);
-            }
-            else {
-                written = 0;
-            }
-        }
-        ASSERT(written <= count);
-        buf += written;
-        count -= written;
-    }
-    return 1;
-}
-
-/***************************************************************************/
-
-int
-efile_writev(Efile_error* errInfo,   /* Where to return error codes */
-        int flags,              /* Flags given when file was
-                                 * opened */
-        int fd,                 /* File descriptor to write to */
-        SysIOVec* iov,          /* Vector of buffer structs.
-                                 * The structs may be changed i.e.
-                                 * due to incomplete writes */
-        int iovcnt)             /* Number of structs in vector */
-{
-    struct fd_data *fd_data;
-    int cnt = 0;                     /* Buffers so far written */
-
-    ASSERT(iovcnt >= 0);
-    if ((fd_data = l_find_local_fd(fd)) != NULL) {
-        if (L_FD_IS_VALID(fd_data)) {
-            /* we are beyond eof and need to pad*/
-            if (l_pad_file(fd_data, L_FD_CUR(fd_data)) < 0) {
-                return check_error(-1, errInfo);
-            }
-        }
-    }
-    while (cnt < iovcnt) {
-        if ((! iov[cnt].iov_base) || (iov[cnt].iov_len <= 0)) {
-            /* Empty buffer - skip */
-            cnt++;
-        }
-        else { /* Non-empty buffer */
-            ssize_t w;                   /* Bytes written in this call */
-            do {
-                w = write(fd, iov[cnt].iov_base, iov[cnt].iov_len);
-            } while (w < 0 && errno == EINTR);
-
-            ASSERT(w <= iov[cnt].iov_len || w == -1);
-
-            if (w < 0) {
-                return check_error(-1, errInfo);
-            }
-            /* Move forward to next buffer to write */
-            for (; cnt < iovcnt && w > 0; cnt++) {
-                if (iov[cnt].iov_base && iov[cnt].iov_len > 0) {
-                    if (w < iov[cnt].iov_len) {
-                        /* Adjust the buffer for next write */
-                        iov[cnt].iov_len -= w;
-                        iov[cnt].iov_base += w;
-                        w = 0;
-                        break;
-                    }
-                    else {
-                        w -= iov[cnt].iov_len;
-                    }
-                }
-            }
-            ASSERT(w == 0);
-        } /* else Non-empty buffer */
-    } /* while (cnt< iovcnt) */
-    return 1;
-}
-
-/***************************************************************************/
-
-int
-efile_read(Efile_error* errInfo,     /* Where to return error codes. */
-        int flags,                  /* Flags given when file was opened. */
-        int fd,                     /* File descriptor to read from. */
-        char* buf,                  /* Buffer to read into. */
-        size_t count,       /* Number of bytes to read. */
-        size_t *pBytesRead)         /* Where to return number of
-                                       bytes read. */
-{
-    ssize_t n;
-    struct fd_data *fd_data;
-
-    if ((fd_data = l_find_local_fd(fd)) != NULL) {
-        if (L_FD_IS_VALID(fd_data)) {
-            *pBytesRead = 0;
-            return 1;
-        }
-    }
-    for (;;)  {
-        if ((n = read(fd, buf, count)) >= 0) {
-            break;
-        }
-        else if (errno != EINTR) {
-            return check_error(-1, errInfo);
-        }
-    }
-    if (fd_data != NULL && L_FD_IS_VALID(fd_data)) {
-        L_FD_INVALIDATE(fd_data);
-    }
-    *pBytesRead = (size_t) n;
-    return 1;
-}
-
-/* pread() and pwrite()                                                   */
-/* Some unix systems, notably Solaris has these syscalls                  */
-/* It is especially nice for i.e. the dets module to have support         */
-/* for this, even if the underlying OS dosn't support it, it is           */
-/* reasonably easy to work around by first calling seek, and then         */
-/* calling read().                                                        */
-/* This later strategy however changes the file pointer, which pread()    */
-/* does not do. We choose to ignore this and say that the location        */
-/* of the file pointer is undefined after a call to any of the p functions*/
-
-
-int
-efile_pread(Efile_error* errInfo,     /* Where to return error codes. */
-        int fd,                /* File descriptor to read from. */
-        Sint64 offset,            /* Offset in bytes from BOF. */
-        char* buf,                     /* Buffer to read into. */
-        size_t count,          /* Number of bytes to read. */
-        size_t *pBytesRead)            /* Where to return
-                                          number of bytes read. */
-{
-    int res = efile_seek(errInfo, fd, offset, EFILE_SEEK_SET, NULL);
-    if (res) {
-        return efile_read(errInfo, 0, fd, buf, count, pBytesRead);
-    } else {
-        return res;
-    }
-}
-
-
-/***************************************************************************/
-
-int
-efile_pwrite(Efile_error* errInfo,  /* Where to return error codes. */
-        int fd,                /* File descriptor to write to. */
-        char* buf,             /* Buffer to write. */
-        size_t count,          /* Number of bytes to write. */
-        Sint64 offset)         /* where to write it */
-{
-    int res = efile_seek(errInfo, fd, offset, EFILE_SEEK_SET, NULL);
-
-    if (res) {
-        return efile_write(errInfo, 0, fd, buf, count);
-    } else {
-        return res;
-    }
-}
-
-/***************************************************************************/
-
-int
-efile_seek(Efile_error* errInfo,      /* Where to return error codes. */
-        int fd,                    /* File descriptor to do the seek on. */
-        Sint64 offset,             /* Offset in bytes from the given
-                                      origin. */
-        int origin,                /* Origin of seek (SEEK_SET, SEEK_CUR,
-                                      SEEK_END). */
-        Sint64 *new_location)      /* Resulting new location in file. */
-{
-    off_t off, result;
-    off = (off_t) offset;
-
-    switch (origin) {
-        case EFILE_SEEK_SET:
-            origin = SEEK_SET;
-            break;
-        case EFILE_SEEK_CUR:
-            origin = SEEK_CUR;
-            break;
-        case EFILE_SEEK_END:
-            origin = SEEK_END;
-            break;
-        default:
-            errno = EINVAL;
-            return check_error(-1, errInfo);
-    }
-
-    if (off != offset) {
-        errno = EINVAL;
-        return check_error(-1, errInfo);
-    }
-
-    errno = 0;
-    result = lseek(fd, off, origin);
-
-    if (result >= 0) {
-        l_invalidate_local_fd(fd);
-    }
-
-    if (result < 0)
-    {
-        if (errno == ENOSYS) {
-            int size, cur_pos;
-
-            if (off < 0) {
-                errno = EINVAL;
-                return check_error(-1, errInfo);
-            }
-
-            cur_pos = lseek(fd, 0, SEEK_CUR);
-            size = lseek(fd, 0, SEEK_END);
-
-            if (origin == SEEK_SET) {
-                result = offset;
-            }
-            else if (origin == SEEK_CUR) {
-                result = offset + cur_pos;
-            }
-            else if (origin == SEEK_END) {
-                result = size + offset;
-            }
-
-            /* sanity check our result */
-            if (size > result) {
-                return check_error(-1, errInfo);
-            }
-
-            /* store the data localy */
-            l_update_local_fd(fd, result, size);
-
-            /* reset the original file position */
-            if (origin != SEEK_END) {
-                lseek(fd, cur_pos, SEEK_SET);
-            }
-        }
-        else if (errno == 0) {
-            errno = EINVAL;
-        }
-    }
-
-    if (new_location) {
-        *new_location = result;
-    }
-
-    return 1;
-}
-
-/***************************************************************************/
-
-int
-efile_truncate_file(Efile_error* errInfo, int *fd, int flags)
-{
-    off_t offset;
-    struct fd_data *fd_data;
-
-    if ((fd_data = l_find_local_fd(*fd)) != NULL && L_FD_IS_VALID(fd_data)) {
-        offset = L_FD_CUR(fd_data);
-    }
-    else {
-        offset = lseek(*fd, 0, SEEK_CUR);
-    }
-
-    return check_error(((offset >= 0) &&
-                (ftruncate(*fd, offset) == 0)) ? 1 : -1, errInfo);
-}
-
-/***************************************************************************/
-
-int
-efile_readlink(Efile_error* errInfo, char* name, char* buffer, size_t size)
-{
-    errno = ENOTSUP;
-    return check_error(-1, errInfo);
-}
-
-/***************************************************************************/
-
-int
-efile_altname(Efile_error* errInfo, char* name, char* buffer, size_t size)
-{
-    errno = ENOTSUP;
-    return check_error(-1, errInfo);
-}
-
-/***************************************************************************/
-
-int
-efile_link(Efile_error* errInfo, char* old, char* new)
-{
-    errno = ENOTSUP;
-    return check_error(-1, errInfo);
-}
-
-/***************************************************************************/
-
-int
-efile_symlink(Efile_error* errInfo, char* old, char* new)
-{
-    errno = ENOTSUP;
-    return check_error(-1, errInfo);
-}
-
-/***************************************************************************/
-
-int
-efile_fadvise(Efile_error* errInfo, int fd, Sint64 offset,
-        Sint64 length, int advise)
-{
-    return check_error(posix_fadvise(fd, offset, length, advise), errInfo);
-}
-
-/***************************************************************************/
-
-static int
-call_posix_fallocate(int fd, Sint64 offset, Sint64 length)
-{
-    int ret;
-
-    /*
-     * On Linux and Solaris for example, posix_fallocate() returns
-     * a positive error number on error and it does not set errno.
-     * On FreeBSD however (9.0 at least), it returns -1 on error
-     * and it sets errno.
-     */
-    do {
-        ret = posix_fallocate(fd, (off_t) offset, (off_t) length);
-        if (ret > 0) {
-            errno = ret;
-            ret = -1;
-        }
-    } while (ret != 0 && errno == EINTR);
-
-    return ret;
-}
-
-/***************************************************************************/
-
-int
-efile_fallocate(Efile_error* errInfo, int fd, Sint64 offset, Sint64 length)
-{
-    return check_error(call_posix_fallocate(fd, offset, length), errInfo);
-}
diff --git a/erts/emulator/drivers/ose/ose_signal_drv.c b/erts/emulator/drivers/ose/ose_signal_drv.c
deleted file mode 100644
index 2be9462a47..0000000000
--- a/erts/emulator/drivers/ose/ose_signal_drv.c
+++ /dev/null
@@ -1,897 +0,0 @@
-/*
- * %CopyrightBegin%
- *
- * Copyright Ericsson AB 2013-2013. All Rights Reserved.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- *
- * %CopyrightEnd%
- */
-#ifdef HAVE_CONFIG_H
-#  include "config.h"
-#endif
-
-#include "errno.h"
-#include "stdio.h"
-#include "string.h"
-#include "stddef.h"
-
-#include "sys.h"
-#include "erl_driver.h"
-#include "ose.h"
-
-
-#ifdef HAVE_OSE_SPI_H
-#include "ose_spi/ose_spi.h"
-#endif
-
-#define DEBUG_ATTACH   0
-#define DEBUG_HUNT     0
-#define DEBUG_SEND     0
-#define DEBUG_LISTEN   0
-
-#if 0
-#define DEBUGP(FMT,...) printf(FMT, __VA_ARGS__)
-#else
-#define DEBUGP(FMT,...)
-#endif
-
-#if DEBUG_ATTACH
-#define DEBUGP_ATTACH(...) DEBUGP( __VA_ARGS__)
-#else
-#define DEBUGP_ATTACH(...)
-#endif
-
-#if DEBUG_HUNT
-#define DEBUGP_HUNT(...) DEBUGP( __VA_ARGS__)
-#else
-#define DEBUGP_HUNT(...)
-#endif
-
-#if DEBUG_LISTEN
-#define DEBUGP_LISTEN(...) DEBUGP( __VA_ARGS__)
-#else
-#define DEBUGP_LISTEN(...)
-#endif
-
-#if DEBUG_SEND
-#define DEBUGP_SEND(...) DEBUGP( __VA_ARGS__)
-#else
-#define DEBUGP_SEND(...)
-#endif
-
-
-#define DRIVER_NAME "ose_signal_drv"
-#define GET_SPID        1
-#define GET_NAME        2
-#define HUNT          100
-#define DEHUNT        101
-#define ATTACH        102
-#define DETACH        103
-#define SEND          104
-#define SEND_W_S      105
-#define LISTEN        106
-#define OPEN          200
-
-#define REF_SEGMENT_SIZE 8
-
-struct async {
-  SIGSELECT signo;
-  ErlDrvTermData port;
-  ErlDrvTermData proc;
-  PROCESS spid;
-  PROCESS target;
-  Uint32 ref;
-};
-
-/**
- * OSE signals
- **/
-union SIGNAL {
-  SIGSELECT signo;
-  struct async async;
-};
-
-/**
- * The driver's context
- **/
-typedef struct _driver_context {
-  ErlDrvPort port;
-  PROCESS spid;
-  ErlDrvEvent perm_events[2];
-  ErlDrvEvent *events;
-  Uint32 event_cnt;
-  Uint32 ref;
-  Uint32 *outstanding_refs;
-  Uint32 outstanding_refs_max;
-  Uint32 outstanding_refs_cnt;
-} driver_context_t;
-
-/**
- * Global variables
- **/
-static ErlDrvTermData a_ok;
-static ErlDrvTermData a_error;
-static ErlDrvTermData a_enomem;
-static ErlDrvTermData a_enoent;
-static ErlDrvTermData a_badarg;
-static ErlDrvTermData a_mailbox_up;
-static ErlDrvTermData a_mailbox_down;
-static ErlDrvTermData a_ose_drv_reply;
-static ErlDrvTermData a_message;
-static PROCESS proxy_proc;
-
-
-/**
- * Serialize/unserialize unsigned 32-bit values
- **/
-static char *put_u32(unsigned int value, char *ptr) {
-  *ptr++ = (value & 0xff000000) >> 24;
-  *ptr++ = (value & 0x00ff0000) >> 16;
-  *ptr++ = (value & 0x0000ff00) >> 8;
-  *ptr++ = (value & 0xff);
-
-  return ptr;
-}
-
-static unsigned int get_u32(char *ptr) {
-  unsigned int result = 0;
-  result += (ptr[0] & 0xff) << 24;
-  result += (ptr[1] & 0xff) << 16;
-  result += (ptr[2] & 0xff) << 8;
-  result += (ptr[3] & 0xff);
-
-  return result;
-}
-
-
-/* Stolen from efile_drv.c */
-
-/* char EV_CHAR_P(ErlIOVec *ev, int p, int q) */
-#define EV_CHAR_P(ev, p, q)                   \
-    (((char *)(ev)->iov[(q)].iov_base) + (p))
-
-/* int EV_GET_CHAR(ErlIOVec *ev, char *p, int *pp, int *qp) */
-#define EV_GET_CHAR(ev, p, pp, qp) ev_get_char(ev, p ,pp, qp)
-static int
-ev_get_char(ErlIOVec *ev, char *p, int *pp, int *qp) {
-  if (*(pp)+1 <= (ev)->iov[*(qp)].iov_len) {
-    *(p) = *EV_CHAR_P(ev, *(pp), *(qp));
-    if (*(pp)+1 < (ev)->iov[*(qp)].iov_len)
-      *(pp) = *(pp)+1;
-    else {
-      (*(qp))++;
-      *pp = 0;
-    }
-    return !0;
-  }
-  return 0;
-}
-
-/* Uint32 EV_UINT32(ErlIOVec *ev, int p, int q)*/
-#define EV_UINT32(ev, p, q) \
-    ((Uint32) *(((unsigned char *)(ev)->iov[(q)].iov_base) + (p)))
-
-/* int EV_GET_UINT32(ErlIOVec *ev, Uint32 *p, int *pp, int *qp) */
-#define EV_GET_UINT32(ev, p, pp, qp) ev_get_uint32(ev,(Uint32*)(p),pp,qp)
-static int
-ev_get_uint32(ErlIOVec *ev, Uint32 *p, int *pp, int *qp) {
-  if (*(pp)+4 <= (ev)->iov[*(qp)].iov_len) {
-    *(p) = (EV_UINT32(ev, *(pp),   *(qp)) << 24)
-      | (EV_UINT32(ev, *(pp)+1, *(qp)) << 16)
-      | (EV_UINT32(ev, *(pp)+2, *(qp)) << 8)
-      | (EV_UINT32(ev, *(pp)+3, *(qp)));
-    if (*(pp)+4 < (ev)->iov[*(qp)].iov_len)
-      *(pp) = *(pp)+4;
-    else {
-      (*(qp))++;
-      *pp = 0;
-    }
-    return !0;
-  }
-  return 0;
-}
-
-/**
- * Convinience macros
- **/
-#define send_response(port,output) erl_drv_send_term(driver_mk_port(port),\
-    driver_caller(port), output, sizeof(output) / sizeof(output[0]));
-
-void iov_memcpy(void *dest,ErlIOVec *ev,int ind,int off);
-void iov_memcpy(void *dest,ErlIOVec *ev,int ind,int off) {
-  int i;
-  memcpy(dest,ev->iov[ind].iov_base+off,ev->iov[ind].iov_len-off);
-  for (i = ind+1; i < ev->vsize; i++)
-    memcpy(dest,ev->iov[i].iov_base,ev->iov[i].iov_len);
-}
-
-/**
- * Reference handling
- **/
-
-static int add_reference(driver_context_t *ctxt, Uint32 ref) {
-
-  /*
-   * Premature optimizations may be evil, but they sure are fun.
-   */
-
-  if (ctxt->outstanding_refs == NULL) {
-    /* First ref to be ignored */
-    ctxt->outstanding_refs = driver_alloc(REF_SEGMENT_SIZE*sizeof(Uint32));
-    if (!ctxt->outstanding_refs)
-      return 1;
-
-    memset(ctxt->outstanding_refs,0,REF_SEGMENT_SIZE*sizeof(Uint32));
-    ctxt->outstanding_refs_max += REF_SEGMENT_SIZE;
-    ctxt->outstanding_refs[ctxt->outstanding_refs_cnt++] = ref;
-  } else if (ctxt->outstanding_refs_cnt == ctxt->outstanding_refs_max) {
-    /* Expand ref array */
-    Uint32 *new_array;
-    ctxt->outstanding_refs_max += REF_SEGMENT_SIZE;
-    new_array = driver_realloc(ctxt->outstanding_refs,
-			       ctxt->outstanding_refs_max*sizeof(Uint32));
-
-    if (!new_array) {
-      ctxt->outstanding_refs_max -= REF_SEGMENT_SIZE;
-      return 1;
-    }
-
-    ctxt->outstanding_refs = new_array;
-
-    memset(ctxt->outstanding_refs+ctxt->outstanding_refs_cnt,0,
-	   REF_SEGMENT_SIZE*sizeof(Uint32));
-    ctxt->outstanding_refs[ctxt->outstanding_refs_cnt++] = ref;
-
-  } else {
-    /* Find an empty slot:
-     *   First we try current index,
-     *   then we scan for a slot.
-     */
-    if (!ctxt->outstanding_refs[ctxt->outstanding_refs_cnt]) {
-      ctxt->outstanding_refs[ctxt->outstanding_refs_cnt++] = ref;
-    } else {
-      int i;
-      ASSERT(ctxt->outstanding_refs_cnt < ctxt->outstanding_refs_max);
-      for (i = 0; i < ctxt->outstanding_refs_max; i++)
-	if (!ctxt->outstanding_refs[i])
-	  break;
-      ASSERT(ctxt->outstanding_refs[i] == 0);
-      ctxt->outstanding_refs[i] = ref;
-      ctxt->outstanding_refs_cnt++;
-    }
-  }
-  return 0;
-}
-
-/* Return 0 if removed, 1 if does not exist, */
-static int remove_reference(driver_context_t *ctxt, Uint32 ref) {
-  int i,j;
-
-  if (ctxt->outstanding_refs_max == 0 && ctxt->outstanding_refs_cnt == 0) {
-    ASSERT(ctxt->outstanding_refs == NULL);
-    return 1;
-  }
-
-  for (i = 0; i < ctxt->outstanding_refs_max; i++) {
-    if (ctxt->outstanding_refs[i] == ref) {
-      ctxt->outstanding_refs[i] = 0;
-      ctxt->outstanding_refs_cnt--;
-      i = -1;
-      break;
-    }
-  }
-
-  if (i != -1)
-    return 1;
-
-  if (ctxt->outstanding_refs_cnt == 0) {
-    driver_free(ctxt->outstanding_refs);
-    ctxt->outstanding_refs = NULL;
-    ctxt->outstanding_refs_max = 0;
-  } else if (ctxt->outstanding_refs_cnt == (ctxt->outstanding_refs_max - REF_SEGMENT_SIZE)) {
-    Uint32 *new_array;
-    for (i = 0, j = 0; i < ctxt->outstanding_refs_cnt; i++) {
-      if (ctxt->outstanding_refs[i] == 0) {
-	for (j = i+1; j < ctxt->outstanding_refs_max; j++)
-	  if (ctxt->outstanding_refs[j]) {
-	    ctxt->outstanding_refs[i] = ctxt->outstanding_refs[j];
-	    ctxt->outstanding_refs[j] = 0;
-	    break;
-	  }
-      }
-    }
-    ctxt->outstanding_refs_max -= REF_SEGMENT_SIZE;
-    new_array = driver_realloc(ctxt->outstanding_refs,
-			       ctxt->outstanding_refs_max*sizeof(Uint32));
-    if (!new_array) {
-      ctxt->outstanding_refs_max += REF_SEGMENT_SIZE;
-      return 2;
-    }
-
-    ctxt->outstanding_refs = new_array;
-
-  }
-
-  return 0;
-}
-
-/**
- * The OSE proxy process. This only handles ERTS_SIGNAL_OSE_DRV_ATTACH.
- * The process is needed because signals triggered by attach ignore
- * redir tables.
- *
- * We have one global proxy process to save memory. An attempt to make each
- * port phantom into a proxy was made, but that used way to much memory.
- */
-static OS_PROCESS(driver_proxy_process) {
-  SIGSELECT sigs[] = {1,ERTS_SIGNAL_OSE_DRV_ATTACH};
-  PROCESS master = 0;
-
-  while (1) {
-    union SIGNAL *sig = receive(sigs);
-
-    if (sig->signo == ERTS_SIGNAL_OSE_DRV_ATTACH) {
-
-      /* The first message is used to determine who to send messages to. */
-      if (master == 0)
-	master = sender(&sig);
-
-      if (sig->async.target == 0) {
-	PROCESS from = sender(&sig);
-	restore(sig);
-	DEBUGP_ATTACH("0x%x: got attach 0x%x, sending to 0x%x\n",
-		      current_process(),from,master);
-	sig->async.target = from;
-	send(&sig,master);
-      } else {
-	PROCESS target = sig->async.target;
-	restore(sig);
-	sig->async.target = 0;
-	DEBUGP_ATTACH("0x%x: doing attach on 0x%x\n",current_process(),target);
-	attach(&sig,target);
-      }
-    }
-  }
-}
-
-
-/**
- * Init routine for the driver
- **/
-static int drv_init(void) {
-
-  a_ok = driver_mk_atom("ok");
-  a_error = driver_mk_atom("error");
-  a_enomem = driver_mk_atom("enomem");
-  a_enoent = driver_mk_atom("enoent");
-  a_badarg = driver_mk_atom("badarg");
-  a_mailbox_up = driver_mk_atom("mailbox_up");
-  a_mailbox_down = driver_mk_atom("mailbox_down");
-  a_ose_drv_reply = driver_mk_atom("ose_drv_reply");
-  a_message = driver_mk_atom("message");
-
-  proxy_proc = create_process(get_ptype(current_process()),
-			      "ose_signal_driver_proxy",
-			      driver_proxy_process, 10000,
-			      get_pri(current_process()),
-			      0, 0, NULL, 0, 0);
-
-#ifdef DEBUG
-  efs_clone(proxy_proc);
-#endif
-  start(proxy_proc);
-
-  return 0;
-}
-
-/* Signal resolution callback */
-static ErlDrvOseEventId resolve_signal(union SIGNAL* osig) {
-  union SIGNAL *sig = osig;
-  if (sig->signo == ERTS_SIGNAL_OSE_DRV_HUNT ||
-      sig->signo == ERTS_SIGNAL_OSE_DRV_ATTACH) {
-    return sig->async.spid;
-  }
-  DEBUGP("%p: Got signal %d sent to %p from 0x%p\n",
-	 current_process(),sig->signo,addressee(&sig),sender(&sig));
-  return addressee(&sig);
-}
-
-
-/**
- * Start routine for the driver
- **/
-static ErlDrvData drv_start(ErlDrvPort port, char *command)
-{
-  driver_context_t *ctxt = driver_alloc(sizeof(driver_context_t));
-
-  ctxt->perm_events[0] = NULL;
-  ctxt->perm_events[1] = NULL;
-
-  ctxt->spid = 0;
-  ctxt->port = port;
-  ctxt->event_cnt = 0;
-  ctxt->events = NULL;
-  ctxt->ref = 0;
-  ctxt->outstanding_refs = NULL;
-  ctxt->outstanding_refs_max = 0;
-  ctxt->outstanding_refs_cnt = 0;
-
-
-  /* Set the communication protocol to Erlang to be binary */
-  set_port_control_flags(port, PORT_CONTROL_FLAG_BINARY);
-
-  /* Everything ok */
-  return (ErlDrvData)ctxt;
-}
-
-/**
- * Stop routine for the driver
- **/
-static void drv_stop(ErlDrvData driver_data)
-{
-  driver_context_t *ctxt = (driver_context_t *)driver_data;
-  int i;
-
-  /* HUNT + ATTACH */
-  if (ctxt->perm_events[0])
-    driver_select(ctxt->port, ctxt->perm_events[0],
-		  ERL_DRV_USE|ERL_DRV_READ, 0);
-  if (ctxt->perm_events[1])
-    driver_select(ctxt->port, ctxt->perm_events[1],
-		  ERL_DRV_USE|ERL_DRV_READ, 0);
-
-  for (i = 0; i < ctxt->event_cnt; i++) {
-    driver_select(ctxt->port, ctxt->events[i], ERL_DRV_USE|ERL_DRV_READ, 0);
-  }
-
-  if (ctxt->spid != 0)
-    kill_proc(ctxt->spid);
-  DEBUGP("0x%x: stopped\n",ctxt->spid);
-  if (ctxt->events)
-    driver_free(ctxt->events);
-  if (ctxt->outstanding_refs)
-    driver_free(ctxt->outstanding_refs);
-
-  driver_free(ctxt);
-}
-
-/**
- * Output from Erlang
- **/
-static void outputv(ErlDrvData driver_data, ErlIOVec *ev)
-{
-  driver_context_t *ctxt = (driver_context_t *)driver_data;
-  int p = 0, q = 1;
-  char cmd;
-
-  if (! EV_GET_CHAR(ev,&cmd,&p,&q)) {
-    ErlDrvTermData output[] = {
-      ERL_DRV_ATOM, a_ose_drv_reply,
-      ERL_DRV_PORT, driver_mk_port(ctxt->port),
-      ERL_DRV_ATOM, a_badarg,
-      ERL_DRV_TUPLE, 3};
-    send_response(ctxt->port, output);
-    return;
-  }
-
-  /* Command is in the buffer's first byte */
-  switch(cmd) {
-
-  case OPEN: {
-    char *name = driver_alloc(ev->size - 1+1);
-    struct OS_redir_entry redir[2];
-
-    redir[0].sig = 1;
-    redir[0].pid = current_process();
-
-    iov_memcpy(name,ev,q,p);
-    name[ev->size-1] = '\0';
-
-    ctxt->spid = create_process(OS_PHANTOM, name, NULL, 0,
-				0, 0, 0, redir, 0, 0);
-
-    DEBUGP("0x%x: open\n",ctxt->spid);
-
-    ctxt->perm_events[1] =
-      erl_drv_ose_event_alloc(ERTS_SIGNAL_OSE_DRV_ATTACH,(int)ctxt->spid,
-			      resolve_signal, NULL);
-    driver_select(ctxt->port,ctxt->perm_events[1],ERL_DRV_READ|ERL_DRV_USE,1);
-
-    ctxt->perm_events[0] =
-      erl_drv_ose_event_alloc(ERTS_SIGNAL_OSE_DRV_HUNT,(int)ctxt->spid,
-			      resolve_signal, NULL);
-    driver_select(ctxt->port,ctxt->perm_events[0],ERL_DRV_READ|ERL_DRV_USE,1);
-
-    start(ctxt->spid);
-
-    {
-      ErlDrvTermData output[] = {
-	ERL_DRV_ATOM, a_ose_drv_reply,
-	ERL_DRV_PORT, driver_mk_port(ctxt->port),
-	ERL_DRV_ATOM, a_ok,
-	ERL_DRV_TUPLE, 3};
-
-      send_response(ctxt->port, output);
-    }
-
-    break;
-
-  }
-
-  case ATTACH:
-  case HUNT:
-    {
-      union SIGNAL *sig = alloc(sizeof(union SIGNAL),
-				cmd == HUNT ? ERTS_SIGNAL_OSE_DRV_HUNT:ERTS_SIGNAL_OSE_DRV_ATTACH);
-
-      sig->async.port = driver_mk_port(ctxt->port);
-      sig->async.proc = driver_caller(ctxt->port);
-      sig->async.spid = ctxt->spid;
-      sig->async.ref = ++ctxt->ref;
-
-      if (add_reference(ctxt,ctxt->ref)) {
-	ErlDrvTermData output[] = {
-	  ERL_DRV_ATOM, a_ose_drv_reply,
-	  ERL_DRV_PORT, driver_mk_port(ctxt->port),
-	  ERL_DRV_ATOM, a_enomem,
-	  ERL_DRV_TUPLE, 3};
-	send_response(ctxt->port, output);
-	free_buf(&sig);
-      } else {
-	ErlDrvTermData output[] = {
-	  ERL_DRV_ATOM, a_ose_drv_reply,
-	  ERL_DRV_PORT, driver_mk_port(ctxt->port),
-	  ERL_DRV_PORT, driver_mk_port(ctxt->port),
-	  ERL_DRV_INT, (ErlDrvUInt)ctxt->ref,
-	  ERL_DRV_TUPLE, 2,
-	  ERL_DRV_TUPLE, 3};
-	send_response(ctxt->port, output);
-
-	if (cmd == HUNT) {
-	  char *huntname = driver_alloc(sizeof(char)*((ev->size-1)+1));
-
-	  iov_memcpy(huntname,ev,q,p);
-	  huntname[ev->size-1] = '\0';
-
-	  DEBUGP_HUNT("0x%x: hunt %s -> %u (%u,%u)\n",
-		      ctxt->spid,huntname,ctxt->ref,
-		      ctxt->outstanding_refs_cnt,
-		      ctxt->outstanding_refs_max);
-
-	  hunt(huntname, 0, NULL, &sig);
-
-	  driver_free(huntname);
-	} else {
-	  EV_GET_UINT32(ev,&sig->async.target,&p,&q);
-	  DEBUGP_ATTACH("0x%x: attach %u -> %u (%u,%u)\n",
-			ctxt->spid,sig->async.target,
-			ctxt->ref,
-			ctxt->outstanding_refs_cnt,
-			ctxt->outstanding_refs_max);
-
-	  send(&sig,proxy_proc);
-	}
-
-      }
-
-      break;
-    }
-
-  case DETACH:
-  case DEHUNT:
-    {
-
-        Uint32 ref;
-
-	EV_GET_UINT32(ev,&ref,&p,&q);
-	if (cmd == DETACH) {
-	  DEBUGP_ATTACH("0x%x: detach %u (%u,%u)\n",ctxt->spid,ref,
-			ctxt->outstanding_refs_cnt,
-			ctxt->outstanding_refs_max);
-	} else {
-	  DEBUGP_HUNT("0x%x: dehunt %u (%u,%u)\n",ctxt->spid,ref,
-		      ctxt->outstanding_refs_cnt,
-		      ctxt->outstanding_refs_max);
-	}
-
-	if (remove_reference(ctxt,ref)) {
-	  ErlDrvTermData output[] = {
-	    ERL_DRV_ATOM, a_ose_drv_reply,
-	    ERL_DRV_PORT, driver_mk_port(ctxt->port),
-	    ERL_DRV_ATOM, a_error,
-	    ERL_DRV_ATOM, a_enoent,
-	    ERL_DRV_TUPLE, 2,
-	    ERL_DRV_TUPLE, 3};
-
-	  send_response(ctxt->port, output);
-	} else {
-	  ErlDrvTermData output[] = {
-	    ERL_DRV_ATOM, a_ose_drv_reply,
-	    ERL_DRV_PORT, driver_mk_port(ctxt->port),
-	    ERL_DRV_ATOM, a_ok,
-	    ERL_DRV_TUPLE, 3};
-
-	  send_response(ctxt->port, output);
-	}
-
-      break;
-    }
-
-  case SEND:
-  case SEND_W_S:
-    {
-      PROCESS spid;
-      PROCESS sender;
-      SIGSELECT signo;
-      OSBUFSIZE size = ev->size-9;
-      union SIGNAL *sig;
-
-      EV_GET_UINT32(ev,&spid,&p,&q);
-
-      if (cmd == SEND_W_S) {
-	EV_GET_UINT32(ev,&sender,&p,&q);
-	size -= 4;
-      } else {
-	sender = ctxt->spid;
-      }
-
-      EV_GET_UINT32(ev,&signo,&p,&q);
-
-      sig = alloc(size + sizeof(SIGSELECT),signo);
-
-      if (cmd == SEND_W_S) {
-	DEBUGP_SEND("0x%x: send_w_s(%u,%u,%u)\n",ctxt->spid,spid,signo,sender);
-      } else {
-	DEBUGP_SEND("0x%x: send(%u,%u)\n",ctxt->spid,spid,signo);
-      }
-
-      iov_memcpy(((char *)&sig->signo) + sizeof(SIGSELECT),ev,q,p);
-
-      send_w_s(&sig, sender, spid);
-
-      break;
-    }
-
-    case LISTEN:
-    {
-      int i,j,event_cnt = (ev->size - 1)/4;
-      ErlDrvEvent *events = NULL;
-      SIGSELECT signo,tmp_signo;
-
-      if (event_cnt == 0) {
-	for (i = 0; i < ctxt->event_cnt; i++)
-	  driver_select(ctxt->port,ctxt->events[i],ERL_DRV_READ|ERL_DRV_USE,0);
-	if (ctxt->events)
-	  driver_free(ctxt->events);
-      } else {
-	events = driver_alloc(sizeof(ErlDrvEvent)*event_cnt);
-	EV_GET_UINT32(ev,&signo,&p,&q);
-	for (i = 0, j = 0; i < event_cnt || j < ctxt->event_cnt; ) {
-
-	  if (ctxt->events)
-	    erl_drv_ose_event_fetch(ctxt->events[j],&tmp_signo,NULL,NULL);
-
-	  if (signo == tmp_signo) {
-	    events[i++] = ctxt->events[j++];
-	    EV_GET_UINT32(ev,&signo,&p,&q);
-	  } else if (signo < tmp_signo || !ctxt->events) {
-	    /* New signal to select on */
-	    events[i] = erl_drv_ose_event_alloc(signo,(int)ctxt->spid,
-						resolve_signal, NULL);
-	    driver_select(ctxt->port,events[i++],ERL_DRV_READ|ERL_DRV_USE,1);
-	    EV_GET_UINT32(ev,&signo,&p,&q);
-	  } else {
-	    /* Remove old signal to select on */
-	    driver_select(ctxt->port,ctxt->events[j++],ERL_DRV_READ|ERL_DRV_USE,0);
-	  }
-	}
-	if (ctxt->events)
-	  driver_free(ctxt->events);
-      }
-      ctxt->events = events;
-      ctxt->event_cnt = event_cnt;
-
-      {
-	ErlDrvTermData output[] = {
-	  ERL_DRV_ATOM, a_ose_drv_reply,
-	  ERL_DRV_PORT, driver_mk_port(ctxt->port),
-	  ERL_DRV_ATOM, a_ok,
-	  ERL_DRV_TUPLE, 3};
-	send_response(ctxt->port, output);
-      }
-      break;
-    }
-
-    default:
-    {
-      DEBUGP("Warning: 'ose_signal_drv' unknown command '%d'\n", cmd);
-      break;
-    }
-  }
-}
-
-/**
- * Handler for when OSE signal arrives
- **/
-static void ready_input(ErlDrvData driver_data, ErlDrvEvent event)
-{
-  driver_context_t *ctxt = (driver_context_t *)driver_data;
-  union SIGNAL *sig = erl_drv_ose_get_signal(event);
-
-  while (sig != NULL) {
-
-    switch(sig->signo)
-      {
-	/* Remote process is available */
-      case ERTS_SIGNAL_OSE_DRV_HUNT:
-	{
-	  const PROCESS spid = sender(&sig);
-
-	  if (remove_reference(ctxt,sig->async.ref)) {
-	    DEBUGP_HUNT("0x%x: Got hunt from 0x%x -> %u (CANCELLED) (%u,%u)\n",
-			ctxt->spid,spid,sig->async.ref,
-			ctxt->outstanding_refs_cnt,
-			ctxt->outstanding_refs_max);
-	    /* Already removed by dehunt */
-	  } else {
-	    ErlDrvTermData reply[] = {
-	      ERL_DRV_ATOM, a_mailbox_up,
-	      ERL_DRV_PORT, sig->async.port,
-	      ERL_DRV_PORT, sig->async.port,
-	      ERL_DRV_UINT, (ErlDrvUInt)sig->async.ref,
-	      ERL_DRV_TUPLE, 2,
-	      ERL_DRV_UINT, (ErlDrvUInt)spid,
-	      ERL_DRV_TUPLE, 4};
-	    DEBUGP_HUNT("0x%x: Got hunt from 0x%x -> %u (%u,%u)\n",
-			ctxt->spid,spid,sig->async.ref,
-			ctxt->outstanding_refs_cnt,
-			ctxt->outstanding_refs_max);
-	    erl_drv_send_term(sig->async.port, sig->async.proc, reply,
-			      sizeof(reply) / sizeof(reply[0]));
-	  }
-	  break;
-	}
-
-	/* Remote process is down */
-      case ERTS_SIGNAL_OSE_DRV_ATTACH:
-	{
-	  PROCESS spid = sig->async.target;
-
-	  if (remove_reference(ctxt,sig->async.ref)) {
-	    DEBUGP_ATTACH("0x%x: Got attach from 0x%x -> %u (CANCELLED) (%u,%u)\n",
-			  ctxt->spid,spid,sig->async.ref,
-			  ctxt->outstanding_refs_cnt,
-			  ctxt->outstanding_refs_max);
-	    /* Already removed by detach */
-	  } else {
-	    ErlDrvTermData reply[] = {
-	      ERL_DRV_ATOM, a_mailbox_down,
-	      ERL_DRV_PORT, sig->async.port,
-	      ERL_DRV_PORT, sig->async.port,
-	      ERL_DRV_UINT, sig->async.ref,
-	      ERL_DRV_TUPLE, 2,
-	      ERL_DRV_UINT, (ErlDrvUInt)spid,
-	      ERL_DRV_TUPLE, 4};
-	     DEBUGP_ATTACH("0x%x: Got attach from 0x%x -> %u (%u,%u)\n",
-			   ctxt->spid,spid,sig->async.ref,
-			   ctxt->outstanding_refs_cnt,
-			   ctxt->outstanding_refs_max);
-	    erl_drv_send_term(sig->async.port, sig->async.proc, reply,
-			      sizeof(reply) / sizeof(reply[0]));
-	  }
-	  break;
-	}
-
-	/* Received user defined signal */
-      default:
-	{
-	  const PROCESS spid = sender(&sig);
-	  const OSBUFSIZE size = sigsize(&sig) - sizeof(SIGSELECT);
-	  const char *sig_data = ((char *)&sig->signo) + sizeof(SIGSELECT);
-
-	  ErlDrvTermData reply[] = {
-	    ERL_DRV_ATOM, a_message,
-	    ERL_DRV_PORT, driver_mk_port(ctxt->port),
-	    ERL_DRV_UINT, (ErlDrvUInt)spid,
-	    ERL_DRV_UINT, (ErlDrvUInt)ctxt->spid,
-	    ERL_DRV_UINT, (ErlDrvUInt)sig->signo,
-	    ERL_DRV_BUF2BINARY, (ErlDrvTermData)sig_data, (ErlDrvUInt)size,
-	    ERL_DRV_TUPLE, 4,
-	    ERL_DRV_TUPLE, 3};
-
-	  DEBUGP_SEND("0x%x: Got 0x%u\r\n", spid, sig->signo);
-
-	  erl_drv_output_term(driver_mk_port(ctxt->port), reply,
-			      sizeof(reply) / sizeof(reply[0]));
-	  break;
-	}
-      }
-
-    free_buf(&sig);
-    sig = erl_drv_ose_get_signal(event);
-  }
-}
-
-/**
- * Handler for 'port_control'
- **/
-static ErlDrvSSizeT control(ErlDrvData driver_data, unsigned int cmd,
-                            char *buf, ErlDrvSizeT len,
-                            char **rbuf, ErlDrvSizeT rlen)
-{
-  driver_context_t *ctxt = (driver_context_t *)driver_data;
-
-  switch(cmd)
-  {
-    case GET_SPID:
-    {
-      const PROCESS spid = ctxt->spid;
-      put_u32(spid, *rbuf);
-      return sizeof(PROCESS);
-    }
-
-#ifdef HAVE_OSE_SPI_H
-    case GET_NAME:
-    {
-      const PROCESS spid = get_u32(buf);
-      char *name = (char*)get_pid_info(spid,OSE_PI_NAME);
-      int n;
-      if (!name) {
-	*rbuf = NULL;
-	return 0;
-      }
-
-      if (rlen < (n = strlen(name))) {
-	ErlDrvBinary *bin = driver_alloc_binary(n);
-	strncpy(bin->orig_bytes,name,n);
-	*rbuf = (char*)bin;
-      } else
-	strncpy(*rbuf,name,n);
-      free_buf((union SIGNAL**)&name);
-
-      return n;
-    }
-#endif
-    default:
-    {
-      /* Unknown command */
-      return (ErlDrvSSizeT)ERL_DRV_ERROR_GENERAL;
-      break;
-    }
-  }
-}
-
-static void stop_select(ErlDrvEvent event, void *reserved)
-{
-  erl_drv_ose_event_free(event);
-}
-
-/**
- * Setup the driver entry for the Erlang runtime
- **/
-ErlDrvEntry ose_signal_driver_entry = {
-  .init                         = drv_init,
-  .start                        = drv_start,
-  .stop                         = drv_stop,
-  .outputv                      = outputv,
-  .ready_input                  = ready_input,
-  .driver_name                  = DRIVER_NAME,
-  .control                      = control,
-  .extended_marker              = ERL_DRV_EXTENDED_MARKER,
-  .major_version                = ERL_DRV_EXTENDED_MAJOR_VERSION,
-  .minor_version                = ERL_DRV_EXTENDED_MINOR_VERSION,
-  .driver_flags                 = ERL_DRV_FLAG_USE_PORT_LOCKING,
-  .stop_select                  = stop_select
-};
-
diff --git a/erts/emulator/drivers/ose/ttsl_drv.c b/erts/emulator/drivers/ose/ttsl_drv.c
deleted file mode 100644
index f759b47984..0000000000
--- a/erts/emulator/drivers/ose/ttsl_drv.c
+++ /dev/null
@@ -1,69 +0,0 @@
-/*
- * %CopyrightBegin%
- *
- * Copyright Ericsson AB 1996-2013. All Rights Reserved.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- *
- * %CopyrightEnd%
- */
-/*
- * Stub tty driver because group/user depend on this.
- */
-
-#ifdef HAVE_CONFIG_H
-#  include "config.h"
-#endif
-
-#include "erl_driver.h"
-
-static int ttysl_init(void);
-static ErlDrvData ttysl_start(ErlDrvPort, char*);
-
-/* Define the driver table entry. */
-struct erl_drv_entry ttsl_driver_entry = {
-    ttysl_init,
-    ttysl_start,
-    NULL,
-    NULL,
-    NULL,
-    NULL,
-    "tty_sl",
-    NULL,
-    NULL,
-    NULL,
-    NULL, /* timeout */
-    NULL, /* outputv */
-    NULL, /* ready_async */
-    NULL, /* flush */
-    NULL, /* call */
-    NULL, /* event */
-    ERL_DRV_EXTENDED_MARKER,
-    ERL_DRV_EXTENDED_MAJOR_VERSION,
-    ERL_DRV_EXTENDED_MINOR_VERSION,
-    0, /* ERL_DRV_FLAGs */
-    NULL,
-    NULL, /* process_exit */
-    NULL
-};
-
-
-static int ttysl_init(void)
-{
-    return 0;
-}
-
-static ErlDrvData ttysl_start(ErlDrvPort port, char* buf)
-{
-    return ERL_DRV_ERROR_GENERAL;
-}
diff --git a/erts/emulator/drivers/unix/ttsl_drv.c b/erts/emulator/drivers/unix/ttsl_drv.c
index 0f773b69fb..25cad37e25 100644
--- a/erts/emulator/drivers/unix/ttsl_drv.c
+++ b/erts/emulator/drivers/unix/ttsl_drv.c
@@ -720,6 +720,7 @@ static void ttysl_from_erlang(ErlDrvData ttysl_data, char* buf, ErlDrvSizeT coun
     }
 
     driver_enq_bin(ttysl_port,putcbuf,0,putcpos);
+    driver_free_binary(putcbuf);
 
     if (sz == 0) {
         for (;;) {
@@ -732,13 +733,13 @@ static void ttysl_from_erlang(ErlDrvData ttysl_data, char* buf, ErlDrvSizeT coun
             else
                 written = 0;
             if (written < 0) {
-                if (errno == EAGAIN) {
+                if (errno == ERRNO_BLOCK || errno == EINTR) {
                     driver_select(ttysl_port,(ErlDrvEvent)(long)ttysl_fd,
                                   ERL_DRV_USE|ERL_DRV_WRITE,1);
                     break;
                 } else {
-                    /* we ignore all other errors */
-                    break;
+		    driver_failure_posix(ttysl_port, errno);
+		    return;
                 }
             } else {
                 if (driver_deq(ttysl_port, written) == 0)
@@ -778,11 +779,12 @@ static void ttysl_to_tty(ErlDrvData ttysl_data, ErlDrvEvent fd) {
         else
             written = 0;
         if (written < 0) {
-            if (errno == EAGAIN) {
-                break;
-            } else {
-                /* we ignore all other errors */
+            if (errno == EINTR) {
+	        continue;
+	    } else if (errno != ERRNO_BLOCK){
+	        driver_failure_posix(ttysl_port, errno);
             }
+            break;
         } else {
             sz = driver_deq(ttysl_port, written);
             if (sz < TTY_BUFFSIZE && ttysl_send_ok) {
@@ -1207,6 +1209,7 @@ static int outc(int c)
     putcbuf->orig_bytes[putcpos++] = c;
     if (putcpos == putclen) {
         driver_enq_bin(ttysl_port,putcbuf,0,putclen);
+        driver_free_binary(putcbuf);
         putcpos = 0;
         putclen = TTY_BUFFSIZE;
         putcbuf = driver_alloc_binary(BUFSIZ);
diff --git a/erts/emulator/drivers/unix/unix_efile.c b/erts/emulator/drivers/unix/unix_efile.c
index 06ba986044..46eccc6568 100644
--- a/erts/emulator/drivers/unix/unix_efile.c
+++ b/erts/emulator/drivers/unix/unix_efile.c
@@ -45,10 +45,10 @@
 #endif
 
 #if defined(__APPLE__) && defined(__MACH__) && !defined(__DARWIN__)
-#define DARWIN 1
+#define __DARWIN__ 1
 #endif
 
-#if defined(DARWIN) || defined(HAVE_LINUX_FALLOC_H) || defined(HAVE_POSIX_FALLOCATE)
+#if defined(__DARWIN__) || defined(HAVE_LINUX_FALLOC_H) || defined(HAVE_POSIX_FALLOCATE)
 #include <fcntl.h>
 #endif
 
@@ -476,11 +476,11 @@ efile_fsync(Efile_error *errInfo, /* Where to return error codes. */
 #ifdef NO_FSYNC
   undefined fsync /* XXX: Really? */
 #else
-#if defined(DARWIN) && defined(F_FULLFSYNC)
+#if defined(__DARWIN__) && defined(F_FULLFSYNC)
     return check_error(fcntl(fd, F_FULLFSYNC), errInfo);
 #else
     return check_error(fsync(fd), errInfo);
-#endif /* DARWIN */
+#endif /* __DARWIN__ */
 #endif /* NO_FSYNC */
 }
 
@@ -962,7 +962,7 @@ efile_sendfile(Efile_error* errInfo, int in_fd, int out_fd,
 	  retval = len;
       }
     } while (len == SENDFILE_CHUNK_SIZE);
-#elif defined(DARWIN)
+#elif defined(__DARWIN__)
     int retval;
     off_t len;
     do {
diff --git a/erts/emulator/hipe/hipe_bif0.c b/erts/emulator/hipe/hipe_bif0.c
index efbe951ce5..6f495b8825 100644
--- a/erts/emulator/hipe/hipe_bif0.c
+++ b/erts/emulator/hipe/hipe_bif0.c
@@ -488,7 +488,7 @@ static void *const_term_alloc(void *tmpl)
     alloc_size = size + (offsetof(struct const_term, mem)/sizeof(Eterm));
     hipe_constants_size += alloc_size;
 
-    p = (struct const_term*)erts_alloc(ERTS_ALC_T_HIPE, alloc_size * sizeof(Eterm));
+    p = (struct const_term*)erts_alloc(ERTS_ALC_T_LITERAL, alloc_size * sizeof(Eterm));
 
     /* I have absolutely no idea if having a private 'off_heap'
        works or not. _Some_ off_heap object is required for
@@ -497,6 +497,8 @@ static void *const_term_alloc(void *tmpl)
     hp = &p->mem[0];
     p->val = copy_struct(obj, size, &hp, &const_term_table_off_heap);
 
+    erts_set_literal_tag(&p->val, &p->mem[0], size);
+
     return &p->bucket;
 }
 
@@ -507,6 +509,9 @@ static void init_const_term_table(void)
     f.cmp = (HCMP_FUN) const_term_cmp;
     f.alloc = (HALLOC_FUN) const_term_alloc;
     f.free = (HFREE_FUN) NULL;
+    f.meta_alloc = (HMALLOC_FUN) erts_alloc;
+    f.meta_free = (HMFREE_FUN) erts_free;
+    f.meta_print = (HMPRINT_FUN) erts_print;
     hash_init(ERTS_ALC_T_HIPE, &const_term_table, "const_term_table", 97, f);
 }
 
@@ -574,15 +579,15 @@ static void print_mfa(Eterm mod, Eterm fun, unsigned int ari)
 static Uint *hipe_find_emu_address(Eterm mod, Eterm name, unsigned int arity)
 {
     Module *modp;
-    Uint *code_base;
+    BeamCodeHeader* code_hdr;
     int i, n;
 
     modp = erts_get_module(mod, erts_active_code_ix());
-    if (modp == NULL || (code_base = modp->curr.code) == NULL)
+    if (modp == NULL || (code_hdr = modp->curr.code_hdr) == NULL)
 	return NULL;
-    n = code_base[MI_NUM_FUNCTIONS];
+    n = code_hdr->num_functions;
     for (i = 0; i < n; ++i) {
-	Uint *code_ptr = (Uint*)code_base[MI_FUNCTIONS+i];
+	Uint *code_ptr = (Uint*)code_hdr->functions[i];
 	ASSERT(code_ptr[0] == BeamOpCode(op_i_func_info_IaaI));
 	if (code_ptr[3] == name && code_ptr[4] == arity)
 	    return code_ptr+5;
@@ -715,6 +720,9 @@ static void init_nbif_table(void)
     f.cmp = (HCMP_FUN) nbif_cmp;
     f.alloc = (HALLOC_FUN) nbif_alloc;
     f.free = NULL;
+    f.meta_alloc = (HMALLOC_FUN) erts_alloc;
+    f.meta_free = (HMFREE_FUN) erts_free;
+    f.meta_print = (HMPRINT_FUN) erts_print;
 
     hash_init(ERTS_ALC_T_NBIF_TABLE, &nbif_table, "nbif_table", 500, f);
 
@@ -808,6 +816,9 @@ static void init_primop_table(void)
     f.cmp = (HCMP_FUN) primop_cmp;
     f.alloc = (HALLOC_FUN) primop_alloc;
     f.free = NULL;
+    f.meta_alloc = (HMALLOC_FUN) erts_alloc;
+    f.meta_free = (HMFREE_FUN) erts_free;
+    f.meta_print = (HMPRINT_FUN) erts_print;
 
     hash_init(ERTS_ALC_T_HIPE, &primop_table, "primop_table", 50, f);
 
@@ -1741,7 +1752,7 @@ BIF_RETTYPE hipe_bifs_check_crc_1(BIF_ALIST_1)
 
     if (!term_to_Uint(BIF_ARG_1, &crc))
 	BIF_ERROR(BIF_P, BADARG);
-    if (crc == HIPE_SYSTEM_CRC)
+    if (crc == HIPE_ERTS_CHECKSUM)
 	BIF_RET(am_true);
     BIF_RET(am_false);
 }
@@ -1826,6 +1837,9 @@ static void init_modinfo_table(void)
     f.cmp = (HCMP_FUN) modinfo_cmp;
     f.alloc = (HALLOC_FUN) modinfo_alloc;
     f.free = (HFREE_FUN) NULL;
+    f.meta_alloc = (HMALLOC_FUN) erts_alloc;
+    f.meta_free = (HMFREE_FUN) erts_free;
+    f.meta_print = (HMPRINT_FUN) erts_print;
     hash_init(ERTS_ALC_T_HIPE, &modinfo_table, "modinfo_table", 11, f);
 }
 
diff --git a/erts/emulator/hipe/hipe_bif_list.m4 b/erts/emulator/hipe/hipe_bif_list.m4
index b3bd5e4357..6aa0c9a32e 100644
--- a/erts/emulator/hipe/hipe_bif_list.m4
+++ b/erts/emulator/hipe/hipe_bif_list.m4
@@ -269,21 +269,23 @@ noproc_primop_interface_1(nbif_atomic_inc, hipe_atomic_inc)
 /* BIFs that disable GC while trapping are called via a wrapper
  * to reserve stack space for the "trap frame".
  */
-define(CFUN,`ifelse($1,term_to_binary_1,hipe_wrapper_term_to_binary_1,
-ifelse($1,term_to_binary_2,hipe_wrapper_term_to_binary_2,
-ifelse($1,binary_to_term_1,hipe_wrapper_binary_to_term_1,
-ifelse($1,binary_to_term_2,hipe_wrapper_binary_to_term_2,
-ifelse($1,binary_to_list_1,hipe_wrapper_binary_to_list_1,
-ifelse($1,binary_to_list_3,hipe_wrapper_binary_to_list_3,
-ifelse($1,bitstring_to_list_1,hipe_wrapper_bitstring_to_list_1,
-ifelse($1,list_to_binary_1,hipe_wrapper_list_to_binary_1,
-ifelse($1,iolist_to_binary_1,hipe_wrapper_iolist_to_binary_1,
-ifelse($1,binary_list_to_bin_1,hipe_wrapper_binary_list_to_bin_1,
-ifelse($1,list_to_bitstring_1,hipe_wrapper_list_to_bitstring_1,
-ifelse($1,send_2,hipe_wrapper_send_2,
-ifelse($1,send_3,hipe_wrapper_send_3,
-ifelse($1,ebif_bang_2,hipe_wrapper_ebif_bang_2,
-$1))))))))))))))')
+define(CFUN,`ifelse(
+$1, term_to_binary_1, hipe_wrapper_$1,
+$1, term_to_binary_2, hipe_wrapper_$1,
+$1, binary_to_term_1, hipe_wrapper_$1,
+$1, binary_to_term_2, hipe_wrapper_$1,
+$1, binary_to_list_1, hipe_wrapper_$1,
+$1, binary_to_list_3, hipe_wrapper_$1,
+$1, bitstring_to_list_1, hipe_wrapper_$1,
+$1, list_to_binary_1, hipe_wrapper_$1,
+$1, iolist_to_binary_1, hipe_wrapper_$1,
+$1, binary_list_to_bin_1, hipe_wrapper_$1,
+$1, list_to_bitstring_1, hipe_wrapper_$1,
+$1, send_2, hipe_wrapper_$1,
+$1, send_3, hipe_wrapper_$1,
+$1, ebif_bang_2, hipe_wrapper_$1,
+$1, maps_merge_2, hipe_wrapper_$1,
+$1)')
 
 define(BIF_LIST,`standard_bif_interface_$3(nbif_$4, CFUN($4))')
 include(TARGET/`erl_bif_list.h')
diff --git a/erts/emulator/hipe/hipe_gc.c b/erts/emulator/hipe/hipe_gc.c
index 2c747771ac..2e19bf88bf 100644
--- a/erts/emulator/hipe/hipe_gc.c
+++ b/erts/emulator/hipe/hipe_gc.c
@@ -46,10 +46,6 @@ Eterm *fullsweep_nstack(Process *p, Eterm *n_htop)
     /* arch-specific nstack walk state */
     struct nstack_walk_state walk_state;
 
-    /* fullsweep-specific state */
-    char *src, *oh;
-    Uint src_size, oh_size;
-
     if (!p->hipe.nstack) {
 	ASSERT(!p->hipe.nsp && !p->hipe.nstend);
 	return n_htop;
@@ -66,11 +62,6 @@ Eterm *fullsweep_nstack(Process *p, Eterm *n_htop)
 
     sdesc = nstack_walk_init_sdesc(p, &walk_state);
 
-    src = (char*)HEAP_START(p);
-    src_size = (char*)HEAP_TOP(p) - src;
-    oh = (char*)OLD_HEAP(p);
-    oh_size = (char*)OLD_HTOP(p) - oh;
-
     for (;;) {
 	if (nstack_walk_nsp_reached_end(nsp, nsp_end)) {
 	    if (nsp == nsp_end) {
@@ -97,8 +88,7 @@ Eterm *fullsweep_nstack(Process *p, Eterm *n_htop)
 		    if (IS_MOVED_BOXED(val)) {
 			ASSERT(is_boxed(val));
 			*nsp_i = val;
-		    } else if (in_area(ptr, src, src_size) ||
-			       in_area(ptr, oh, oh_size)) {
+		    } else if (!erts_is_literal(gval, ptr)) {
 			MOVE_BOXED(ptr, val, n_htop, nsp_i);
 		    }
 		} else if (is_list(gval)) {
@@ -106,8 +96,7 @@ Eterm *fullsweep_nstack(Process *p, Eterm *n_htop)
 		    Eterm val = *ptr;
 		    if (IS_MOVED_CONS(val)) {
 			*nsp_i = ptr[1];
-		    } else if (in_area(ptr, src, src_size) ||
-			       in_area(ptr, oh, oh_size)) {
+		    } else if (!erts_is_literal(gval, ptr)) {
 			ASSERT(within(ptr, p));
 			MOVE_CONS(ptr, val, n_htop, nsp_i);
 		    }
@@ -139,11 +128,13 @@ void gensweep_nstack(Process *p, Eterm **ptr_old_htop, Eterm **ptr_n_htop)
     unsigned int mask;
     /* arch-specific nstack walk state */
     struct nstack_walk_state walk_state;
+    char *oh;
+    Uint oh_size;
 
     /* gensweep-specific state */
     Eterm *old_htop, *n_htop;
-    char *heap;
-    Uint heap_size, mature_size;
+    char *mature;
+    Uint mature_size;
 
     if (!p->hipe.nstack) {
 	ASSERT(!p->hipe.nsp && !p->hipe.nstend);
@@ -168,9 +159,10 @@ void gensweep_nstack(Process *p, Eterm **ptr_old_htop, Eterm **ptr_n_htop)
 
     old_htop = *ptr_old_htop;
     n_htop = *ptr_n_htop;
-    heap = (char*)HEAP_START(p);
-    heap_size = (char*)HEAP_TOP(p) - heap;
-    mature_size = (char*)HIGH_WATER(p) - heap;
+    mature = (char *) (p->abandoned_heap ? p->abandoned_heap : p->heap);
+    mature_size = (char*)HIGH_WATER(p) - mature;
+    oh = (char*)OLD_HEAP(p);
+    oh_size = (char*)OLD_HTOP(p) - oh;
 
     for (;;) {
 	if (nstack_walk_nsp_reached_end(nsp, nsp_end)) {
@@ -209,9 +201,9 @@ void gensweep_nstack(Process *p, Eterm **ptr_old_htop, Eterm **ptr_n_htop)
 		    if (IS_MOVED_BOXED(val)) {
 			ASSERT(is_boxed(val));
 			*nsp_i = val;
-		    } else if (in_area(ptr, heap, mature_size)) {
+		    } else if (ErtsInArea(ptr, mature, mature_size)) {
 			MOVE_BOXED(ptr, val, old_htop, nsp_i);
-		    } else if (in_area(ptr, heap, heap_size)) {
+		    } else if (ErtsInYoungGen(gval, ptr, oh, oh_size)) {
 			ASSERT(within(ptr, p));
 			MOVE_BOXED(ptr, val, n_htop, nsp_i);
 		    }
@@ -220,9 +212,9 @@ void gensweep_nstack(Process *p, Eterm **ptr_old_htop, Eterm **ptr_n_htop)
 		    Eterm val = *ptr;
 		    if (IS_MOVED_CONS(val)) {
 			*nsp_i = ptr[1];
-		    } else if (in_area(ptr, heap, mature_size)) {
+		    } else if (ErtsInArea(ptr, mature, mature_size)) {
 			MOVE_CONS(ptr, val, old_htop, nsp_i);
-		    } else if (in_area(ptr, heap, heap_size)) {
+		    } else if (ErtsInYoungGen(gval, ptr, oh, oh_size)) {
 			ASSERT(within(ptr, p));
 			MOVE_CONS(ptr, val, n_htop, nsp_i);
 		    }
diff --git a/erts/emulator/hipe/hipe_mkliterals.c b/erts/emulator/hipe/hipe_mkliterals.c
index aa12df2932..dfa5313739 100644
--- a/erts/emulator/hipe/hipe_mkliterals.c
+++ b/erts/emulator/hipe/hipe_mkliterals.c
@@ -269,9 +269,6 @@ static const struct literal {
     /* freason codes */
     { "FREASON_TRAP", TRAP },
 
-    /* special Erlang constants */
-    { "THE_NON_VALUE", (int)THE_NON_VALUE },
-
     /* funs */
 #ifdef HIPE
     { "EFE_NATIVE_ADDRESS", offsetof(struct erl_fun_entry, native_address) },
@@ -526,6 +523,8 @@ static const struct rts_param rts_params[] = {
     { 49, "P_MSG_FIRST", 1, offsetof(struct process, msg.first) },
     { 50, "P_MSG_SAVE", 1, offsetof(struct process, msg.save) },
     { 51, "P_CALLEE_EXP", 1, offsetof(struct process, hipe.u.callee_exp) },
+
+    { 52, "THE_NON_VALUE", 1, (int)THE_NON_VALUE },
 };
 
 #define NR_PARAMS	ARRAY_SIZE(rts_params)
@@ -543,6 +542,8 @@ static void compute_crc(void)
 	crc_value = crc_update_int(crc_value, &literals[i].value);
     crc_value &= 0x07FFFFFF;
     literals_crc = crc_value;
+
+    crc_value = crc_init();
     for (i = 0; i < NR_PARAMS; ++i)
 	if (rts_params[i].is_defined)
 	    crc_value = crc_update_int(crc_value, &rts_params[i].value);
@@ -628,6 +629,7 @@ static int do_c(FILE *fp, const char* this_exe)
     print_params(fp, c_define_param);
     fprintf(fp, "#define HIPE_LITERALS_CRC %uU\n", literals_crc);
     fprintf(fp, "#define HIPE_SYSTEM_CRC %uU\n", system_crc);
+    fprintf(fp, "#define HIPE_ERTS_CHECKSUM (HIPE_LITERALS_CRC ^ HIPE_SYSTEM_CRC)\n");
     fprintf(fp, "\n");
     fprintf(fp, "#define RTS_PARAMS_CASES");
     print_params(fp, c_case_param);
@@ -645,12 +647,14 @@ static int do_e(FILE *fp, const char* this_exe)
     fprintf(fp, "\n");
     print_params(fp, e_define_param);
     fprintf(fp, "\n");
+    fprintf(fp, "-define(HIPE_LITERALS_CRC, %u).\n", literals_crc);
     if (is_xcomp) {
 	fprintf(fp, "-define(HIPE_SYSTEM_CRC, %u).\n", system_crc);
     }
     else {
 	fprintf(fp, "-define(HIPE_SYSTEM_CRC, hipe_bifs:system_crc()).\n");
     }
+    fprintf(fp, "-define(HIPE_ERTS_CHECKSUM, (?HIPE_LITERALS_CRC bxor ?HIPE_SYSTEM_CRC)).\n");
     return 0;
 }
 
diff --git a/erts/emulator/hipe/hipe_mode_switch.c b/erts/emulator/hipe/hipe_mode_switch.c
index 968452a641..976180cc30 100644
--- a/erts/emulator/hipe/hipe_mode_switch.c
+++ b/erts/emulator/hipe/hipe_mode_switch.c
@@ -196,7 +196,7 @@ hipe_push_beam_trap_frame(Process *p, Eterm reg[], unsigned arity)
 	ASSERT(!(p->flags & F_DISABLE_GC));
 	if ((p->stop - 2) < p->htop) {
 	    DPRINTF("calling gc to increase BEAM stack size");
-	    p->fcalls -= erts_garbage_collect(p, 2, reg, arity);
+	    erts_garbage_collect(p, 2, reg, arity);
 	    ASSERT(!((p->stop - 2) < p->htop));
 	}
 	p->stop -= 2;
diff --git a/erts/emulator/hipe/hipe_mode_switch.h b/erts/emulator/hipe/hipe_mode_switch.h
index bc863a4f36..620cc6356b 100644
--- a/erts/emulator/hipe/hipe_mode_switch.h
+++ b/erts/emulator/hipe/hipe_mode_switch.h
@@ -95,7 +95,7 @@ ERTS_GLB_INLINE void hipe_reserve_beam_trap_frame(Process *p, Eterm reg[], unsig
 
     /* ensure that at least 2 words are available on the BEAM stack */
     if ((p->stop - 2) < p->htop) {
-	p->fcalls -= erts_garbage_collect(p, 2, reg, arity);
+	erts_garbage_collect(p, 2, reg, arity);
 	ASSERT(!((p->stop - 2) < p->htop));
     }
     p->stop -= 2;
diff --git a/erts/emulator/hipe/hipe_native_bif.c b/erts/emulator/hipe/hipe_native_bif.c
index 98bda43f0e..1bfee94e9e 100644
--- a/erts/emulator/hipe/hipe_native_bif.c
+++ b/erts/emulator/hipe/hipe_native_bif.c
@@ -80,7 +80,7 @@ Eterm hipe_show_nstack_1(BIF_ALIST_1)
 void hipe_gc(Process *p, Eterm need)
 {
     hipe_set_narity(p, 1);
-    p->fcalls -= erts_garbage_collect(p, unsigned_val(need), NULL, 0);
+    erts_garbage_collect(p, unsigned_val(need), NULL, 0);
     hipe_set_narity(p, 0);
 }
 
@@ -93,9 +93,6 @@ BIF_RETTYPE hipe_set_timeout(BIF_ALIST_1)
 {
     Process* p = BIF_P;
     Eterm timeout_value = BIF_ARG_1;
-#if !defined(ARCH_64)
-    Uint time_val;
-#endif
     /* XXX: This should be converted to follow BEAM conventions,
      * but that requires some compiler changes.
      *
@@ -160,13 +157,22 @@ BIF_RETTYPE hipe_set_timeout(BIF_ALIST_1)
  */
 void hipe_select_msg(Process *p)
 {
-    ErlMessage *msgp;
+    ErtsMessage *msgp;
 
     msgp = PEEK_MESSAGE(p);
     UNLINK_MESSAGE(p, msgp);	/* decrements global 'erts_proc_tot_mem' variable */
     JOIN_MESSAGE(p);
     CANCEL_TIMER(p);		/* calls erts_cancel_proc_timer() */
-    free_message(msgp);
+    erts_save_message_in_proc(p, msgp);
+    p->flags &= ~F_DELAY_GC;
+    if (ERTS_IS_GC_DESIRED(p)) {
+	/*
+	 * We want to GC soon but we leave a few
+	 * reductions giving the message some time
+	 * to turn into garbage.
+	 */
+	ERTS_VBUMP_LEAVE_REDS(p, 5);
+    }
 }
 
 void hipe_fclearerror_error(Process *p)
@@ -511,8 +517,9 @@ int hipe_bs_validate_unicode_retract(ErlBinMatchBuffer* mb, Eterm arg)
  */
 Eterm hipe_check_get_msg(Process *c_p)
 {
-    Eterm ret;
-    ErlMessage *msgp;
+    ErtsMessage *msgp;
+
+    c_p->flags |= F_DELAY_GC;
 
  next_message:
 
@@ -534,25 +541,29 @@ Eterm hipe_check_get_msg(Process *c_p)
 	    /* XXX: BEAM doesn't need this */
 	    c_p->hipe_smp.have_receive_locks = 1;
 #endif
+	    c_p->flags &= ~F_DELAY_GC;
 	    return THE_NON_VALUE;
 #ifdef ERTS_SMP
 	}
 #endif
     }
-    ErtsMoveMsgAttachmentIntoProc(msgp, c_p, c_p->stop, HEAP_TOP(c_p),
-				  c_p->fcalls, (void) 0, (void) 0);
-    ret = ERL_MESSAGE_TERM(msgp);
-    if (is_non_value(ret)) {
+
+    if (is_non_value(ERL_MESSAGE_TERM(msgp))
+	&& !erts_decode_dist_message(c_p, ERTS_PROC_LOCK_MAIN, msgp, 0)) {
 	/*
 	 * A corrupt distribution message that we weren't able to decode;
 	 * remove it...
 	 */
 	ASSERT(!msgp->data.attached);
 	UNLINK_MESSAGE(c_p, msgp);
-	free_message(msgp);
+	msgp->next = NULL;
+	erts_cleanup_messages(msgp);
 	goto next_message;
     }
-    return ret;
+
+    ASSERT(is_value(ERL_MESSAGE_TERM(msgp)));
+
+    return ERL_MESSAGE_TERM(msgp);
 }
 
 /*
diff --git a/erts/emulator/hipe/hipe_x86_signal.c b/erts/emulator/hipe/hipe_x86_signal.c
index bb8a3f041f..b7dae88417 100644
--- a/erts/emulator/hipe/hipe_x86_signal.c
+++ b/erts/emulator/hipe/hipe_x86_signal.c
@@ -198,7 +198,7 @@ static void do_init(void)
 #define INIT()	do { if (!init_done()) do_init(); } while (0)
 #endif /* __DARWIN__ */
 
-#if !defined(__GLIBC__) && !defined(__DARWIN__) && !defined(__NetBSD__)
+#if defined(__sun__)
 /*
  * Assume Solaris/x86 2.8.
  * There is a number of sigaction() procedures in libc:
@@ -232,7 +232,56 @@ static void do_init(void)
 }
 #define _NSIG NSIG
 #define INIT()	do { if (!init_done()) do_init(); } while (0)
-#endif	/* not glibc or darwin */
+#endif /* __sun__ */
+
+#if defined(__FreeBSD__)
+/*
+ * This is a copy of Darwin code for FreeBSD.
+ * CAVEAT: detailed semantics are not verified yet.
+ */
+#include <dlfcn.h>
+static int (*__next_sigaction)(int, const struct sigaction*, struct sigaction*);
+#define init_done()	(__next_sigaction != 0)
+extern int _sigaction(int, const struct sigaction*, struct sigaction*);
+#define __SIGACTION _sigaction
+static void do_init(void)
+{
+    __next_sigaction = dlsym(RTLD_NEXT, "sigaction");
+    if (__next_sigaction != 0)
+	return;
+    perror("dlsym_freebsd");
+    abort();
+}
+#define _NSIG NSIG
+#define INIT()	do { if (!init_done()) do_init(); } while (0)
+#endif /* __FreeBSD__ */
+
+#if !(defined(__GLIBC__) || defined(__DARWIN__) || defined(__NetBSD__) || defined(__FreeBSD__) || defined(__sun__))
+/*
+ * Unknown libc -- assume musl.  Note: musl deliberately does not provide a musl-specific
+ * feature test macro, so we cannot check for it.
+ *
+ * sigaction is a weak alias for __sigaction, which is a wrapper for __libc_sigaction.
+ * There are libc-internal calls to __libc_sigaction which install handlers, so we must
+ * override __libc_sigaction rather than __sigaction.
+ */
+#include <dlfcn.h>
+static int (*__next_sigaction)(int, const struct sigaction*, struct sigaction*);
+#define init_done()	(__next_sigaction != 0)
+#define __SIGACTION __libc_sigaction
+static void do_init(void)
+{
+    __next_sigaction = dlsym(RTLD_NEXT, "__libc_sigaction");
+    if (__next_sigaction != 0)
+	return;
+    perror("dlsym");
+    abort();
+}
+#ifndef _NSIG
+#define _NSIG NSIG
+#endif
+#define INIT()	do { if (!init_done()) do_init(); } while (0)
+#endif	/* !(__GLIBC__ || __DARWIN__ || __NetBSD__ || __FreeBSD__ || __sun__) */
 
 #if !defined(__NetBSD__)
 /*
@@ -272,7 +321,7 @@ int __SIGACTION(int signum, const struct sigaction *act, struct sigaction *oldac
 /*
  * This catches the application's own sigaction() calls.
  */
-#if !defined(__DARWIN__) && !defined(__NetBSD__)
+#if !defined(__DARWIN__) && !defined(__NetBSD__) && !defined(__FreeBSD__)
 int sigaction(int signum, const struct sigaction *act, struct sigaction *oldact)
 {
     return my_sigaction(signum, act, oldact);
diff --git a/erts/emulator/sys/common/erl_check_io.c b/erts/emulator/sys/common/erl_check_io.c
index 105b129065..f87196d724 100644
--- a/erts/emulator/sys/common/erl_check_io.c
+++ b/erts/emulator/sys/common/erl_check_io.c
@@ -1337,11 +1337,7 @@ print_select_op(erts_dsprintf_buf_t *dsbufp,
 {
     Port *pp = erts_drvport2port(ix);
     erts_dsprintf(dsbufp,
-#ifdef __OSE__
-		  "driver_select(%p, %d,%s%s%s%s | %d, %d) "
-#else
 		  "driver_select(%p, %d,%s%s%s%s, %d) "
-#endif
 		  "by ",
 		  ix,
 		  (int) GET_FD(fd),
@@ -1861,25 +1857,6 @@ stale_drv_select(Eterm id, ErtsDrvEventState *state, int mode)
 
 #ifndef ERTS_SYS_CONTINOUS_FD_NUMBERS
 
-#ifdef __OSE__
-static SafeHashValue drv_ev_state_hash(void *des)
-{
-    ErtsSysFdType fd = ((ErtsDrvEventState *) des)->fd;
-    /* We use hash on signo ^ id in order for steal to happen when the
-       same signo + fd is selected on by two different ports */
-    SafeHashValue val = (SafeHashValue)(fd->signo ^ fd->id);
-    return val ^ (val >> 8);
-}
-
-static int drv_ev_state_cmp(void *des1, void *des2)
-{
-    ErtsSysFdType fd1 = ((ErtsDrvEventState *) des1)->fd;
-    ErtsSysFdType fd2 = ((ErtsDrvEventState *) des2)->fd;
-    if (fd1->signo == fd2->signo && fd1->id == fd2->id)
-      return 0;
-    return 1;
-}
-#else /* !__OSE__ && !ERTS_SYS_CONTINOUS_FD_NUMBERS i.e. probably windows */
 static SafeHashValue drv_ev_state_hash(void *des)
 {
     SafeHashValue val = (SafeHashValue) ((ErtsDrvEventState *) des)->fd;
@@ -1891,7 +1868,6 @@ static int drv_ev_state_cmp(void *des1, void *des2)
     return ( ((ErtsDrvEventState *) des1)->fd == ((ErtsDrvEventState *) des2)->fd
 	    ? 0 : 1);
 }
-#endif
 
 static void *drv_ev_state_alloc(void *des_tmpl)
 {
diff --git a/erts/emulator/sys/common/erl_mmap.c b/erts/emulator/sys/common/erl_mmap.c
index 754047829f..03ca080c14 100644
--- a/erts/emulator/sys/common/erl_mmap.c
+++ b/erts/emulator/sys/common/erl_mmap.c
@@ -51,23 +51,22 @@
 #endif
 
 /*
- * `mmap_state.sa.bot` and `mmap_state.sua.top` are read only after
+ * `mm->sa.bot` and `mm->sua.top` are read only after
  * initialization, but the other pointers are not; i.e., only
  * ERTS_MMAP_IN_SUPERCARRIER() is allowed without the mutex held.
  */
 #define ERTS_MMAP_IN_SUPERCARRIER(PTR)					\
-    (((UWord) (PTR)) - ((UWord) mmap_state.sa.bot)			\
-     < ((UWord) mmap_state.sua.top) - ((UWord) mmap_state.sa.bot))
+    (((UWord) (PTR)) - ((UWord) mm->sa.bot)			\
+     < ((UWord) mm->sua.top) - ((UWord) mm->sa.bot))
 #define ERTS_MMAP_IN_SUPERALIGNED_AREA(PTR)				\
-    (ERTS_SMP_LC_ASSERT(erts_lc_mtx_is_locked(&mmap_state.mtx)),	\
-     (((UWord) (PTR)) - ((UWord) mmap_state.sa.bot)			\
-      < ((UWord) mmap_state.sa.top) - ((UWord) mmap_state.sa.bot)))
+    (ERTS_SMP_LC_ASSERT(erts_lc_mtx_is_locked(&mm->mtx)),	\
+     (((UWord) (PTR)) - ((UWord) mm->sa.bot)			\
+      < ((UWord) mm->sa.top) - ((UWord) mm->sa.bot)))
 #define ERTS_MMAP_IN_SUPERUNALIGNED_AREA(PTR)				\
-    (ERTS_SMP_LC_ASSERT(erts_lc_mtx_is_locked(&mmap_state.mtx)),	\
-     (((UWord) (PTR)) - ((UWord) mmap_state.sua.bot)			\
-      < ((UWord) mmap_state.sua.top) - ((UWord) mmap_state.sua.bot)))
+    (ERTS_SMP_LC_ASSERT(erts_lc_mtx_is_locked(&mm->mtx)),	\
+     (((UWord) (PTR)) - ((UWord) mm->sua.bot)			\
+      < ((UWord) mm->sua.top) - ((UWord) mm->sua.bot)))
 
-int erts_have_erts_mmap;
 UWord erts_page_inv_mask;
 
 #if defined(DEBUG) || defined(ERTS_MMAP_DEBUG)
@@ -197,10 +196,10 @@ static ErtsMMapOp mmap_ops[ERTS_MMAP_OP_RINGBUF_SZ];
 
 #define ERTS_MMAP_OP_LCK(RES, IN_SZ, OUT_SZ)			\
     do {							\
-	erts_smp_mtx_lock(&mmap_state.mtx);			\
+	erts_smp_mtx_lock(&mm->mtx);			\
 	ERTS_MMAP_OP_START((IN_SZ));				\
 	ERTS_MMAP_OP_END((RES), (OUT_SZ));			\
-	erts_smp_mtx_unlock(&mmap_state.mtx);			\
+	erts_smp_mtx_unlock(&mm->mtx);			\
     } while (0)
 
 #define ERTS_MUNMAP_OP(PTR, SZ)					\
@@ -219,9 +218,9 @@ static ErtsMMapOp mmap_ops[ERTS_MMAP_OP_RINGBUF_SZ];
 
 #define ERTS_MUNMAP_OP_LCK(PTR, SZ)				\
     do {							\
-	erts_smp_mtx_lock(&mmap_state.mtx);			\
+	erts_smp_mtx_lock(&mm->mtx);			\
 	ERTS_MUNMAP_OP((PTR), (SZ));				\
-	erts_smp_mtx_unlock(&mmap_state.mtx);			\
+	erts_smp_mtx_unlock(&mm->mtx);			\
     } while (0)
 
 #define ERTS_MREMAP_OP_START(OLD_PTR, OLD_SZ, IN_SZ)		\
@@ -247,10 +246,10 @@ static ErtsMMapOp mmap_ops[ERTS_MMAP_OP_RINGBUF_SZ];
 
 #define ERTS_MREMAP_OP_LCK(RES, OLD_PTR, OLD_SZ, IN_SZ, OUT_SZ)	\
     do {							\
-	erts_smp_mtx_lock(&mmap_state.mtx);			\
+	erts_smp_mtx_lock(&mm->mtx);			\
 	ERTS_MREMAP_OP_START((OLD_PTR), (OLD_SZ), (IN_SZ));	\
 	ERTS_MREMAP_OP_END((RES), (OUT_SZ));			\
-	erts_smp_mtx_unlock(&mmap_state.mtx);			\
+	erts_smp_mtx_unlock(&mm->mtx);			\
     } while (0)
 
 #define ERTS_MMAP_OP_ABORT()					\
@@ -294,7 +293,7 @@ typedef struct {
     Uint nseg;
 }ErtsFreeSegMap;
 
-static struct {
+struct ErtsMemMapper_ {
     int (*reserve_physical)(char *, UWord);
     void (*unreserve_physical)(char *, UWord);
     int supercarrier;
@@ -346,54 +345,62 @@ static struct {
 	    UWord used;
 	} os;
     } size;
-} mmap_state;
+};
+
+ErtsMemMapper erts_dflt_mmapper;
+
+#if defined(ARCH_64) && defined(ERTS_HAVE_OS_PHYSICAL_MEMORY_RESERVATION)
+ErtsMemMapper erts_literal_mmapper;
+char* erts_literals_start;
+UWord erts_literals_size;
+#endif
 
 #define ERTS_MMAP_SIZE_SC_SA_INC(SZ) 						\
     do {									\
-	mmap_state.size.supercarrier.used.total += (SZ);			\
-	mmap_state.size.supercarrier.used.sa += (SZ);				\
-	ERTS_MMAP_ASSERT(mmap_state.size.supercarrier.used.total		\
-			 <= mmap_state.size.supercarrier.total);		\
-	ERTS_MMAP_ASSERT(mmap_state.size.supercarrier.used.sa			\
-			 <= mmap_state.size.supercarrier.used.total);		\
+	mm->size.supercarrier.used.total += (SZ);			\
+	mm->size.supercarrier.used.sa += (SZ);				\
+	ERTS_MMAP_ASSERT(mm->size.supercarrier.used.total		\
+			 <= mm->size.supercarrier.total);		\
+	ERTS_MMAP_ASSERT(mm->size.supercarrier.used.sa			\
+			 <= mm->size.supercarrier.used.total);		\
     } while (0)
 #define ERTS_MMAP_SIZE_SC_SA_DEC(SZ) 						\
     do {									\
-	ERTS_MMAP_ASSERT(mmap_state.size.supercarrier.used.total >= (SZ));	\
-	mmap_state.size.supercarrier.used.total -= (SZ);			\
-	ERTS_MMAP_ASSERT(mmap_state.size.supercarrier.used.sa >= (SZ));		\
-	mmap_state.size.supercarrier.used.sa -= (SZ);				\
+	ERTS_MMAP_ASSERT(mm->size.supercarrier.used.total >= (SZ));	\
+	mm->size.supercarrier.used.total -= (SZ);			\
+	ERTS_MMAP_ASSERT(mm->size.supercarrier.used.sa >= (SZ));		\
+	mm->size.supercarrier.used.sa -= (SZ);				\
     } while (0)
 #define ERTS_MMAP_SIZE_SC_SUA_INC(SZ) 						\
     do {									\
-	mmap_state.size.supercarrier.used.total += (SZ);			\
-	mmap_state.size.supercarrier.used.sua += (SZ);				\
-	ERTS_MMAP_ASSERT(mmap_state.size.supercarrier.used.total		\
-			 <= mmap_state.size.supercarrier.total);		\
-	ERTS_MMAP_ASSERT(mmap_state.size.supercarrier.used.sua			\
-			 <= mmap_state.size.supercarrier.used.total);		\
+	mm->size.supercarrier.used.total += (SZ);			\
+	mm->size.supercarrier.used.sua += (SZ);				\
+	ERTS_MMAP_ASSERT(mm->size.supercarrier.used.total		\
+			 <= mm->size.supercarrier.total);		\
+	ERTS_MMAP_ASSERT(mm->size.supercarrier.used.sua			\
+			 <= mm->size.supercarrier.used.total);		\
     } while (0)
 #define ERTS_MMAP_SIZE_SC_SUA_DEC(SZ) 						\
     do {									\
-	ERTS_MMAP_ASSERT(mmap_state.size.supercarrier.used.total >= (SZ));	\
-	mmap_state.size.supercarrier.used.total -= (SZ);			\
-	ERTS_MMAP_ASSERT(mmap_state.size.supercarrier.used.sua >= (SZ));	\
-	mmap_state.size.supercarrier.used.sua -= (SZ);				\
+	ERTS_MMAP_ASSERT(mm->size.supercarrier.used.total >= (SZ));	\
+	mm->size.supercarrier.used.total -= (SZ);			\
+	ERTS_MMAP_ASSERT(mm->size.supercarrier.used.sua >= (SZ));	\
+	mm->size.supercarrier.used.sua -= (SZ);				\
     } while (0)
 #define ERTS_MMAP_SIZE_OS_INC(SZ)						\
     do {									\
-	ERTS_MMAP_ASSERT(mmap_state.size.os.used + (SZ) >= (SZ));		\
-	mmap_state.size.os.used += (SZ);					\
+	ERTS_MMAP_ASSERT(mm->size.os.used + (SZ) >= (SZ));		\
+	mm->size.os.used += (SZ);					\
     } while (0)
 #define ERTS_MMAP_SIZE_OS_DEC(SZ) 						\
     do {									\
-	ERTS_MMAP_ASSERT(mmap_state.size.os.used >= (SZ));			\
-	mmap_state.size.os.used -= (SZ);					\
+	ERTS_MMAP_ASSERT(mm->size.os.used >= (SZ));			\
+	mm->size.os.used -= (SZ);					\
     } while (0)
 
 
 static void
-add_free_desc_area(char *start, char *end)
+add_free_desc_area(ErtsMemMapper* mm, char *start, char *end)
 {
     ERTS_MMAP_ASSERT(end == (void *) 0 || end > start);
     if (sizeof(ErtsFreeSegDesc) <= ((UWord) end) - ((UWord) start)) {
@@ -403,7 +410,7 @@ add_free_desc_area(char *start, char *end)
 
 	no = 1;
 	prev_desc = (ErtsFreeSegDesc *) start;
-	prev_desc->start = mmap_state.desc.free_list;
+	prev_desc->start = mm->desc.free_list;
 	desc = (ErtsFreeSegDesc *) (start + sizeof(ErtsFreeSegDesc));
 	desc_end = start + 2*sizeof(ErtsFreeSegDesc);
 
@@ -414,59 +421,59 @@ add_free_desc_area(char *start, char *end)
 	    desc_end += sizeof(ErtsFreeSegDesc);
 	    no++;
 	}
-	mmap_state.desc.free_list = (char *) prev_desc;
-	mmap_state.no.free_seg_descs += no;
+	mm->desc.free_list = (char *) prev_desc;
+	mm->no.free_seg_descs += no;
     }
 }
 
 static ErtsFreeSegDesc *
-add_unused_free_desc_area(void)
+add_unused_free_desc_area(ErtsMemMapper* mm)
 {
     char *ptr;
-    if (!mmap_state.desc.unused_start)
+    if (!mm->desc.unused_start)
 	return NULL;
 
-    ERTS_MMAP_ASSERT(mmap_state.desc.unused_end);
+    ERTS_MMAP_ASSERT(mm->desc.unused_end);
     ERTS_MMAP_ASSERT(ERTS_PAGEALIGNED_SIZE
-		     <= mmap_state.desc.unused_end - mmap_state.desc.unused_start);
+		     <= mm->desc.unused_end - mm->desc.unused_start);
 
-    ptr = mmap_state.desc.unused_start + ERTS_PAGEALIGNED_SIZE;
-    add_free_desc_area(mmap_state.desc.unused_start, ptr);
+    ptr = mm->desc.unused_start + ERTS_PAGEALIGNED_SIZE;
+    add_free_desc_area(mm, mm->desc.unused_start, ptr);
 
-    if ((mmap_state.desc.unused_end - ptr) >= ERTS_PAGEALIGNED_SIZE)
-	mmap_state.desc.unused_start = ptr;
+    if ((mm->desc.unused_end - ptr) >= ERTS_PAGEALIGNED_SIZE)
+	mm->desc.unused_start = ptr;
     else
-	mmap_state.desc.unused_end = mmap_state.desc.unused_start = NULL;
+	mm->desc.unused_end = mm->desc.unused_start = NULL;
 
-    ERTS_MMAP_ASSERT(mmap_state.desc.free_list);
-    return (ErtsFreeSegDesc *) mmap_state.desc.free_list;
+    ERTS_MMAP_ASSERT(mm->desc.free_list);
+    return (ErtsFreeSegDesc *) mm->desc.free_list;
 }
 
 static ERTS_INLINE ErtsFreeSegDesc *
-alloc_desc(void)
+alloc_desc(ErtsMemMapper* mm)
 {
     ErtsFreeSegDesc *res;
-    res = (ErtsFreeSegDesc *) mmap_state.desc.free_list;
+    res = (ErtsFreeSegDesc *) mm->desc.free_list;
     if (!res) {
-	res = add_unused_free_desc_area();
+	res = add_unused_free_desc_area(mm);
 	if (!res)
 	    return NULL;
     }
-    mmap_state.desc.free_list = res->start;
-    ASSERT(mmap_state.no.free_segs.curr < mmap_state.no.free_seg_descs);
-    mmap_state.no.free_segs.curr++;
-    if (mmap_state.no.free_segs.max < mmap_state.no.free_segs.curr)
-	mmap_state.no.free_segs.max = mmap_state.no.free_segs.curr;
+    mm->desc.free_list = res->start;
+    ASSERT(mm->no.free_segs.curr < mm->no.free_seg_descs);
+    mm->no.free_segs.curr++;
+    if (mm->no.free_segs.max < mm->no.free_segs.curr)
+	mm->no.free_segs.max = mm->no.free_segs.curr;
     return res;
 }
 
 static ERTS_INLINE void
-free_desc(ErtsFreeSegDesc *desc)
+free_desc(ErtsMemMapper* mm, ErtsFreeSegDesc *desc)
 {
-    desc->start = mmap_state.desc.free_list;
-    mmap_state.desc.free_list = (char *) desc;
-    ERTS_MMAP_ASSERT(mmap_state.no.free_segs.curr > 0);
-    mmap_state.no.free_segs.curr--;
+    desc->start = mm->desc.free_list;
+    mm->desc.free_list = (char *) desc;
+    ERTS_MMAP_ASSERT(mm->no.free_segs.curr > 0);
+    mm->no.free_segs.curr--;
 }
 
 static ERTS_INLINE ErtsFreeSegDesc* anode_to_desc(RBTNode* anode)
@@ -1233,7 +1240,7 @@ Eterm build_free_seg_list(Process* p, ErtsFreeSegMap* map)
 #    define ERTS_MMAP_FD		(-1)
 #  else
 #    define ERTS_MMAP_FLAGS		(MAP_PRIVATE)
-#    define ERTS_MMAP_FD		mmap_state.mmap_fd
+#    define ERTS_MMAP_FD		mm->mmap_fd
 #  endif
 #endif
 
@@ -1378,11 +1385,12 @@ static void unreserve_noop(char *ptr, UWord size)
 }
 
 static UWord
-alloc_desc_insert_free_seg(ErtsFreeSegMap *map, char* start, char* end)
+alloc_desc_insert_free_seg(ErtsMemMapper* mm,
+                           ErtsFreeSegMap *map, char* start, char* end)
 {
     char *ptr;
     ErtsFreeSegMap *da_map;
-    ErtsFreeSegDesc *desc = alloc_desc();
+    ErtsFreeSegDesc *desc = alloc_desc(mm);
     if (desc) {
 	insert_free_seg(map, desc, start, end);
 	return 0;
@@ -1395,13 +1403,13 @@ alloc_desc_insert_free_seg(ErtsFreeSegMap *map, char* start, char* end)
      */
 
 #if ERTS_HAVE_OS_MMAP
-    if (!mmap_state.no_os_mmap) {
-        ptr = os_mmap(mmap_state.desc.new_area_hint, ERTS_PAGEALIGNED_SIZE, 0);
+    if (!mm->no_os_mmap) {
+        ptr = os_mmap(mm->desc.new_area_hint, ERTS_PAGEALIGNED_SIZE, 0);
         if (ptr) {
-            mmap_state.desc.new_area_hint = ptr+ERTS_PAGEALIGNED_SIZE;
+            mm->desc.new_area_hint = ptr+ERTS_PAGEALIGNED_SIZE;
             ERTS_MMAP_SIZE_OS_INC(ERTS_PAGEALIGNED_SIZE);
-            add_free_desc_area(ptr, ptr+ERTS_PAGEALIGNED_SIZE);
-            desc = alloc_desc();
+            add_free_desc_area(mm, ptr, ptr+ERTS_PAGEALIGNED_SIZE);
+            desc = alloc_desc(mm);
             ERTS_MMAP_ASSERT(desc);
             insert_free_seg(map, desc, start, end);
             return 0;
@@ -1412,20 +1420,20 @@ alloc_desc_insert_free_seg(ErtsFreeSegMap *map, char* start, char* end)
     /*
      * ...then try to find a good place in the supercarrier...
      */
-    da_map = &mmap_state.sua.map;
+    da_map = &mm->sua.map;
     desc = lookup_free_seg(da_map, ERTS_PAGEALIGNED_SIZE);
     if (desc) {
-	if (mmap_state.reserve_physical(desc->start, ERTS_PAGEALIGNED_SIZE))
+	if (mm->reserve_physical(desc->start, ERTS_PAGEALIGNED_SIZE))
 	    ERTS_MMAP_SIZE_SC_SUA_INC(ERTS_PAGEALIGNED_SIZE);
 	else
 	    desc = NULL;
 	
     }
     else {
-	da_map = &mmap_state.sa.map;
+	da_map = &mm->sa.map;
 	desc = lookup_free_seg(da_map, ERTS_PAGEALIGNED_SIZE);
 	if (desc) {
-	    if (mmap_state.reserve_physical(desc->start, ERTS_PAGEALIGNED_SIZE))
+	    if (mm->reserve_physical(desc->start, ERTS_PAGEALIGNED_SIZE))
 		ERTS_MMAP_SIZE_SC_SA_INC(ERTS_PAGEALIGNED_SIZE);
 	    else
 		desc = NULL;
@@ -1433,15 +1441,15 @@ alloc_desc_insert_free_seg(ErtsFreeSegMap *map, char* start, char* end)
     }
     if (desc) {
 	char *da_end = desc->start + ERTS_PAGEALIGNED_SIZE;
-	add_free_desc_area(desc->start, da_end);
+	add_free_desc_area(mm, desc->start, da_end);
 	if (da_end != desc->end)
 	    resize_free_seg(da_map, desc, da_end, desc->end);
 	else {
 	    delete_free_seg(da_map, desc);
-	    free_desc(desc);
+	    free_desc(mm, desc);
 	}
 
-	desc = alloc_desc();
+	desc = alloc_desc(mm);
 	ERTS_MMAP_ASSERT(desc);
 	insert_free_seg(map, desc, start, end);
 	return 0;
@@ -1454,10 +1462,10 @@ alloc_desc_insert_free_seg(ErtsFreeSegMap *map, char* start, char* end)
     ptr = start + ERTS_PAGEALIGNED_SIZE;
     ERTS_MMAP_ASSERT(ptr <= end);
 
-    add_free_desc_area(start, ptr);
+    add_free_desc_area(mm, start, ptr);
 
     if (ptr != end) {
-	desc = alloc_desc();
+	desc = alloc_desc(mm);
 	ERTS_MMAP_ASSERT(desc);
 	insert_free_seg(map, desc, ptr, end);
     }
@@ -1466,46 +1474,46 @@ alloc_desc_insert_free_seg(ErtsFreeSegMap *map, char* start, char* end)
 }
 
 void *
-erts_mmap(Uint32 flags, UWord *sizep)
+erts_mmap(ErtsMemMapper* mm, Uint32 flags, UWord *sizep)
 {
     char *seg;
     UWord asize = ERTS_PAGEALIGNED_CEILING(*sizep);
 
     /* Map in premapped supercarrier */
-    if (mmap_state.supercarrier && !(ERTS_MMAPFLG_OS_ONLY & flags)) {
+    if (mm->supercarrier && !(ERTS_MMAPFLG_OS_ONLY & flags)) {
 	char *end;
 	ErtsFreeSegDesc *desc;
 	Uint32 superaligned = (ERTS_MMAPFLG_SUPERALIGNED & flags);
 
-	erts_smp_mtx_lock(&mmap_state.mtx);
+	erts_smp_mtx_lock(&mm->mtx);
 
 	ERTS_MMAP_OP_START(*sizep);
 
 	if (!superaligned) {
-	    desc = lookup_free_seg(&mmap_state.sua.map, asize);
+	    desc = lookup_free_seg(&mm->sua.map, asize);
 	    if (desc) {
 		seg = desc->start;
 		end = seg+asize;
-		if (!mmap_state.reserve_physical(seg, asize))
+		if (!mm->reserve_physical(seg, asize))
 		    goto supercarrier_reserve_failure;
 		if (desc->end == end) {
-		    delete_free_seg(&mmap_state.sua.map, desc);
-		    free_desc(desc);
+		    delete_free_seg(&mm->sua.map, desc);
+		    free_desc(mm, desc);
 		}
 		else {
 		    ERTS_MMAP_ASSERT(end < desc->end);
-		    resize_free_seg(&mmap_state.sua.map, desc, end, desc->end);
+		    resize_free_seg(&mm->sua.map, desc, end, desc->end);
 		}
 		ERTS_MMAP_SIZE_SC_SUA_INC(asize);
 		goto supercarrier_success;
 	    }
 
-	    if (asize <= mmap_state.sua.bot - mmap_state.sa.top) {
-		if (!mmap_state.reserve_physical(mmap_state.sua.bot - asize,
+	    if (asize <= mm->sua.bot - mm->sa.top) {
+		if (!mm->reserve_physical(mm->sua.bot - asize,
 						 asize))
 		    goto supercarrier_reserve_failure;
-		mmap_state.sua.bot -= asize;
-		seg = mmap_state.sua.bot;
+		mm->sua.bot -= asize;
+		seg = mm->sua.bot;
 		ERTS_MMAP_SIZE_SC_SUA_INC(asize);
 		goto supercarrier_success;
 	    }
@@ -1513,84 +1521,84 @@ erts_mmap(Uint32 flags, UWord *sizep)
 
 	asize = ERTS_SUPERALIGNED_CEILING(asize);
 
-	desc = lookup_free_seg(&mmap_state.sa.map, asize);
+	desc = lookup_free_seg(&mm->sa.map, asize);
 	if (desc) {
 	    char *start = seg = desc->start;
 	    seg = (char *) ERTS_SUPERALIGNED_CEILING(seg);
 	    end = seg+asize;
-	    if (!mmap_state.reserve_physical(start, (UWord) (end - start)))
+	    if (!mm->reserve_physical(start, (UWord) (end - start)))
 		goto supercarrier_reserve_failure;
 	    ERTS_MMAP_SIZE_SC_SA_INC(asize);
 	    if (desc->end == end) {
 		if (start != seg)
-		    resize_free_seg(&mmap_state.sa.map, desc, start, seg);
+		    resize_free_seg(&mm->sa.map, desc, start, seg);
 		else {
-		    delete_free_seg(&mmap_state.sa.map, desc);
-		    free_desc(desc);
+		    delete_free_seg(&mm->sa.map, desc);
+		    free_desc(mm, desc);
 		}
 	    }
 	    else {
 		ERTS_MMAP_ASSERT(end < desc->end);
-		resize_free_seg(&mmap_state.sa.map, desc, end, desc->end);
+		resize_free_seg(&mm->sa.map, desc, end, desc->end);
 		if (start != seg) {
 		    UWord ad_sz;
-		    ad_sz = alloc_desc_insert_free_seg(&mmap_state.sua.map,
+		    ad_sz = alloc_desc_insert_free_seg(mm, &mm->sua.map,
 						       start, seg);
 		    start += ad_sz;
 		    if (start != seg)
-			mmap_state.unreserve_physical(start, (UWord) (seg - start));
+			mm->unreserve_physical(start, (UWord) (seg - start));
 		}
 	    }
 	    goto supercarrier_success;
 	}
 
 	if (superaligned) {
-	    char *start = mmap_state.sa.top;
+	    char *start = mm->sa.top;
 	    seg = (char *) ERTS_SUPERALIGNED_CEILING(start);
 
-	    if (asize + (seg - start) <= mmap_state.sua.bot - start) {
+	    if (asize + (seg - start) <= mm->sua.bot - start) {
 		end = seg + asize;
-		if (!mmap_state.reserve_physical(start, (UWord) (end - start)))
+		if (!mm->reserve_physical(start, (UWord) (end - start)))
 		    goto supercarrier_reserve_failure;
-		mmap_state.sa.top = end;
+		mm->sa.top = end;
 		ERTS_MMAP_SIZE_SC_SA_INC(asize);
 		if (start != seg) {
 		    UWord ad_sz;
-		    ad_sz = alloc_desc_insert_free_seg(&mmap_state.sua.map,
+		    ad_sz = alloc_desc_insert_free_seg(mm, &mm->sua.map,
 						       start, seg);
 		    start += ad_sz;
 		    if (start != seg)
-			mmap_state.unreserve_physical(start, (UWord) (seg - start));
+			mm->unreserve_physical(start, (UWord) (seg - start));
 		}
 		goto supercarrier_success;
 	    }
 
-	    desc = lookup_free_seg(&mmap_state.sua.map, asize + ERTS_SUPERALIGNED_SIZE);
+	    desc = lookup_free_seg(&mm->sua.map, asize + ERTS_SUPERALIGNED_SIZE);
 	    if (desc) {
 		char *org_start = desc->start;
 		char *org_end = desc->end;
 
 		seg = (char *) ERTS_SUPERALIGNED_CEILING(org_start);
 		end = seg + asize;
-		if (!mmap_state.reserve_physical(seg, (UWord) (org_end - seg)))
+		if (!mm->reserve_physical(seg, (UWord) (org_end - seg)))
 		    goto supercarrier_reserve_failure;
 		ERTS_MMAP_SIZE_SC_SUA_INC(asize);
 		if (org_start != seg) {
 		    ERTS_MMAP_ASSERT(org_start < seg);
-		    resize_free_seg(&mmap_state.sua.map, desc, org_start, seg);
+		    resize_free_seg(&mm->sua.map, desc, org_start, seg);
 		    desc = NULL;
 		}
 		if (end != org_end) {
 		    UWord ad_sz = 0;
 		    ERTS_MMAP_ASSERT(end < org_end);
 		    if (desc)
-			resize_free_seg(&mmap_state.sua.map, desc, end, org_end);
+			resize_free_seg(&mm->sua.map, desc, end, org_end);
 		    else
-			ad_sz = alloc_desc_insert_free_seg(&mmap_state.sua.map,
+			ad_sz = alloc_desc_insert_free_seg(mm, &mm->sua.map,
 							   end, org_end);
 		    end += ad_sz;
 		    if (end != org_end)
-			mmap_state.unreserve_physical(end,
+			mm->unreserve_physical(end,
 						      (UWord) (org_end - end));
 		}
 		goto supercarrier_success;
@@ -1598,12 +1606,12 @@ erts_mmap(Uint32 flags, UWord *sizep)
 	}
 
 	ERTS_MMAP_OP_ABORT();
-	erts_smp_mtx_unlock(&mmap_state.mtx);
+	erts_smp_mtx_unlock(&mm->mtx);
     }
 
 #if ERTS_HAVE_OS_MMAP
     /* Map using OS primitives */
-    if (!(ERTS_MMAPFLG_SUPERCARRIER_ONLY & flags) && !mmap_state.no_os_mmap) {
+    if (!(ERTS_MMAPFLG_SUPERCARRIER_ONLY & flags) && !mm->no_os_mmap) {
 	if (!(ERTS_MMAPFLG_SUPERALIGNED & flags)) {
 	    seg = os_mmap(NULL, asize, 0);
 	    if (!seg)
@@ -1661,25 +1669,25 @@ supercarrier_success:
 #endif
 
     ERTS_MMAP_OP_END(seg, asize);
-    erts_smp_mtx_unlock(&mmap_state.mtx);
+    erts_smp_mtx_unlock(&mm->mtx);
 
     *sizep = asize;
     return (void *) seg;
 
 supercarrier_reserve_failure:
-    erts_smp_mtx_unlock(&mmap_state.mtx);
+    erts_smp_mtx_unlock(&mm->mtx);
     *sizep = 0;
     return NULL;
 }
 
 void
-erts_munmap(Uint32 flags, void *ptr, UWord size)
+erts_munmap(ErtsMemMapper* mm, Uint32 flags, void *ptr, UWord size)
 {
     ERTS_MMAP_ASSERT(ERTS_IS_PAGEALIGNED(ptr));
     ERTS_MMAP_ASSERT(ERTS_IS_PAGEALIGNED(size));
 
     if (!ERTS_MMAP_IN_SUPERCARRIER(ptr)) {
-	ERTS_MMAP_ASSERT(!mmap_state.no_os_mmap);
+	ERTS_MMAP_ASSERT(!mm->no_os_mmap);
 #if ERTS_HAVE_OS_MMAP
 	ERTS_MUNMAP_OP_LCK(ptr, size);
 	ERTS_MMAP_SIZE_OS_DEC(size);
@@ -1692,45 +1700,45 @@ erts_munmap(Uint32 flags, void *ptr, UWord size)
 	ErtsFreeSegDesc *prev, *next, *desc;
 	UWord ad_sz = 0;
 
-	ERTS_MMAP_ASSERT(mmap_state.supercarrier);
+	ERTS_MMAP_ASSERT(mm->supercarrier);
 
 	start = (char *) ptr;
 	end = start + size;
 
-	erts_smp_mtx_lock(&mmap_state.mtx);
+	erts_smp_mtx_lock(&mm->mtx);
 
 	ERTS_MUNMAP_OP(ptr, size);
 
 	if (ERTS_MMAP_IN_SUPERALIGNED_AREA(ptr)) {
 
-	    map = &mmap_state.sa.map;
+	    map = &mm->sa.map;
 	    adjacent_free_seg(map, start, end, &prev, &next);
 
 	    ERTS_MMAP_SIZE_SC_SA_DEC(size);
-	    if (end == mmap_state.sa.top) {
+	    if (end == mm->sa.top) {
 		ERTS_MMAP_ASSERT(!next);
 		if (prev) {
 		    start = prev->start;
 		    delete_free_seg(map, prev);
-		    free_desc(prev);
+		    free_desc(mm, prev);
 		}
-		mmap_state.sa.top = start;
+		mm->sa.top = start;
 		goto supercarrier_success;
 	    }
 	}
 	else {
-	    map = &mmap_state.sua.map;
+	    map = &mm->sua.map;
 	    adjacent_free_seg(map, start, end, &prev, &next);
 
 	    ERTS_MMAP_SIZE_SC_SUA_DEC(size);
-	    if (start == mmap_state.sua.bot) {
+	    if (start == mm->sua.bot) {
 		ERTS_MMAP_ASSERT(!prev);
 		if (next) {
 		    end = next->end;
 		    delete_free_seg(map, next);
-		    free_desc(next);
+		    free_desc(mm, next);
 		}
-		mmap_state.sua.bot = end;
+		mm->sua.bot = end;
 		goto supercarrier_success;
 	    }
 	}
@@ -1742,7 +1750,7 @@ erts_munmap(Uint32 flags, void *ptr, UWord size)
 	    end = next->end;
 	    if (prev) {
 		delete_free_seg(map, next);
-		free_desc(next);
+		free_desc(mm, next);
 		goto save_prev;
 	    }
 	    desc = next;
@@ -1756,7 +1764,7 @@ erts_munmap(Uint32 flags, void *ptr, UWord size)
 	if (desc)
 	    resize_free_seg(map, desc, start, end);
 	else
-	    ad_sz = alloc_desc_insert_free_seg(map, start, end);
+	    ad_sz = alloc_desc_insert_free_seg(mm, map, start, end);
 	
     supercarrier_success: {
 	    UWord unres_sz;
@@ -1764,30 +1772,32 @@ erts_munmap(Uint32 flags, void *ptr, UWord size)
 	    ERTS_MMAP_ASSERT(size >= ad_sz);
 	    unres_sz = size - ad_sz;
 	    if (unres_sz)
-		mmap_state.unreserve_physical(((char *) ptr) + ad_sz, unres_sz);
+		mm->unreserve_physical(((char *) ptr) + ad_sz, unres_sz);
 
-            erts_smp_mtx_unlock(&mmap_state.mtx);
+            erts_smp_mtx_unlock(&mm->mtx);
 	}
     }
 }
 
 static void *
-remap_move(Uint32 flags, void *ptr, UWord old_size, UWord *sizep)
+remap_move(ErtsMemMapper* mm,
+           Uint32 flags, void *ptr, UWord old_size, UWord *sizep)
 {
     UWord size = *sizep;
-    void *new_ptr = erts_mmap(flags, &size);
+    void *new_ptr = erts_mmap(mm, flags, &size);
     if (!new_ptr)
 	return NULL;
     *sizep = size;
     if (old_size < size)
 	size = old_size;
     sys_memcpy(new_ptr, ptr, (size_t) size);
-    erts_munmap(flags, ptr, old_size);
+    erts_munmap(mm, flags, ptr, old_size);
     return new_ptr;
 }
 
 void *
-erts_mremap(Uint32 flags, void *ptr, UWord old_size, UWord *sizep)
+erts_mremap(ErtsMemMapper* mm,
+            Uint32 flags, void *ptr, UWord old_size, UWord *sizep)
 {
     void *new_ptr;
     Uint32 superaligned;
@@ -1799,11 +1809,11 @@ erts_mremap(Uint32 flags, void *ptr, UWord old_size, UWord *sizep)
 
     if (!ERTS_MMAP_IN_SUPERCARRIER(ptr)) {
 
-	ERTS_MMAP_ASSERT(!mmap_state.no_os_mmap);
+	ERTS_MMAP_ASSERT(!mm->no_os_mmap);
 
-	if (!(ERTS_MMAPFLG_OS_ONLY & flags) && mmap_state.supercarrier) {
-	    new_ptr = remap_move(ERTS_MMAPFLG_SUPERCARRIER_ONLY|flags, ptr,
-				 old_size, sizep);
+	if (!(ERTS_MMAPFLG_OS_ONLY & flags) && mm->supercarrier) {
+	    new_ptr = remap_move(mm, ERTS_MMAPFLG_SUPERCARRIER_ONLY|flags,
+                                 ptr, old_size, sizep);
 	    if (new_ptr)
 		return new_ptr;
 	}
@@ -1850,7 +1860,7 @@ erts_mremap(Uint32 flags, void *ptr, UWord old_size, UWord *sizep)
 #endif
 #if ERTS_HAVE_OS_MREMAP
 	if (superaligned)
-	    return remap_move(flags, new_ptr, old_size, sizep);
+	    return remap_move(mm, flags, new_ptr, old_size, sizep);
 	else {
 	    new_ptr = os_mremap(ptr, old_size, asize, 0);
 	    if (!new_ptr)
@@ -1872,10 +1882,10 @@ erts_mremap(Uint32 flags, void *ptr, UWord old_size, UWord *sizep)
 	ErtsFreeSegDesc *prev, *next;
 	UWord ad_sz = 0;
 
-	ERTS_MMAP_ASSERT(mmap_state.supercarrier);
+	ERTS_MMAP_ASSERT(mm->supercarrier);
 
 	if (ERTS_MMAPFLG_OS_ONLY & flags)
-	    return remap_move(flags, ptr, old_size, sizep);
+	    return remap_move(mm, flags, ptr, old_size, sizep);
 
 	superaligned = (ERTS_MMAPFLG_SUPERALIGNED & flags);
 
@@ -1883,19 +1893,19 @@ erts_mremap(Uint32 flags, void *ptr, UWord old_size, UWord *sizep)
 		 ? ERTS_SUPERALIGNED_CEILING(*sizep)
 		 : ERTS_PAGEALIGNED_CEILING(*sizep));
 
-	erts_smp_mtx_lock(&mmap_state.mtx);
+	erts_smp_mtx_lock(&mm->mtx);
 
 	if (ERTS_MMAP_IN_SUPERALIGNED_AREA(ptr)
-	    ? (!superaligned && lookup_free_seg(&mmap_state.sua.map, asize))
-	    : (superaligned && lookup_free_seg(&mmap_state.sa.map, asize))) {
-	    erts_smp_mtx_unlock(&mmap_state.mtx);
+	    ? (!superaligned && lookup_free_seg(&mm->sua.map, asize))
+	    : (superaligned && lookup_free_seg(&mm->sa.map, asize))) {
+	    erts_smp_mtx_unlock(&mm->mtx);
 	    /*
 	     * Segment currently in wrong area (due to a previous memory
 	     * shortage), move it to the right area.
 	     * (remap_move() will succeed)
 	     */
-	    return remap_move(ERTS_MMAPFLG_SUPERCARRIER_ONLY|flags, ptr,
-			      old_size, sizep);
+	    return remap_move(mm, ERTS_MMAPFLG_SUPERCARRIER_ONLY|flags,
+                              ptr, old_size, sizep);
 	}
 
 	ERTS_MREMAP_OP_START(ptr, old_size, *sizep);
@@ -1917,18 +1927,18 @@ erts_mremap(Uint32 flags, void *ptr, UWord old_size, UWord *sizep)
 	    UWord unres_sz;
 	    new_ptr = ptr;
 	    if (!ERTS_MMAP_IN_SUPERALIGNED_AREA(ptr)) {
-		map = &mmap_state.sua.map;
+		map = &mm->sua.map;
 		ERTS_MMAP_SIZE_SC_SUA_DEC(old_size - asize);
 	    }
 	    else {
-		if (end == mmap_state.sa.top) {
-		    mmap_state.sa.top = new_end;
-		    mmap_state.unreserve_physical(((char *) ptr) + asize,
+		if (end == mm->sa.top) {
+		    mm->sa.top = new_end;
+		    mm->unreserve_physical(((char *) ptr) + asize,
 						  old_size - asize);
 		    goto supercarrier_resize_success;
 		}
 		ERTS_MMAP_SIZE_SC_SA_DEC(old_size - asize);
-		map = &mmap_state.sa.map;
+		map = &mm->sa.map;
 	    }
 	    
 	    adjacent_free_seg(map, start, end, &prev, &next);
@@ -1936,11 +1946,11 @@ erts_mremap(Uint32 flags, void *ptr, UWord old_size, UWord *sizep)
 	    if (next)
 		resize_free_seg(map, next, new_end, next->end);
 	    else
-		ad_sz = alloc_desc_insert_free_seg(map, new_end, end);
+		ad_sz = alloc_desc_insert_free_seg(mm, map, new_end, end);
 	    ERTS_MMAP_ASSERT(old_size - asize >= ad_sz);
 	    unres_sz = old_size - asize - ad_sz;
 	    if (unres_sz)
-		mmap_state.unreserve_physical(((char *) ptr) + asize + ad_sz,
+		mm->unreserve_physical(((char *) ptr) + asize + ad_sz,
 					      unres_sz);
 	    goto supercarrier_resize_success;
 	}
@@ -1950,17 +1960,17 @@ erts_mremap(Uint32 flags, void *ptr, UWord old_size, UWord *sizep)
 	    ERTS_MMAP_ASSERT(ERTS_IS_PAGEALIGNED(old_size));
 	    ERTS_MMAP_ASSERT(ERTS_IS_PAGEALIGNED(asize));
 
-	    adjacent_free_seg(&mmap_state.sua.map, start, end, &prev, &next);
+	    adjacent_free_seg(&mm->sua.map, start, end, &prev, &next);
 
 	    if (next && new_end <= next->end) {
-		if (!mmap_state.reserve_physical(((char *) ptr) + old_size,
+		if (!mm->reserve_physical(((char *) ptr) + old_size,
 						 asize - old_size))
 		    goto supercarrier_reserve_failure;
 		if (new_end < next->end)
-		    resize_free_seg(&mmap_state.sua.map, next, new_end, next->end);
+		    resize_free_seg(&mm->sua.map, next, new_end, next->end);
 		else {
-		    delete_free_seg(&mmap_state.sua.map, next);
-		    free_desc(next);
+		    delete_free_seg(&mm->sua.map, next);
+		    free_desc(mm, next);
 		}
 		new_ptr = ptr;
 		ERTS_MMAP_SIZE_SC_SUA_INC(asize - old_size);
@@ -1969,28 +1979,28 @@ erts_mremap(Uint32 flags, void *ptr, UWord old_size, UWord *sizep)
 	}
 	else { /* Superaligned area */
 
-	    if (end == mmap_state.sa.top) {
-		if (new_end <= mmap_state.sua.bot) {
-		    if (!mmap_state.reserve_physical(((char *) ptr) + old_size,
+	    if (end == mm->sa.top) {
+		if (new_end <= mm->sua.bot) {
+		    if (!mm->reserve_physical(((char *) ptr) + old_size,
 						     asize - old_size))
 			goto supercarrier_reserve_failure;
-		    mmap_state.sa.top = new_end;
+		    mm->sa.top = new_end;
 		    new_ptr = ptr;
 		    ERTS_MMAP_SIZE_SC_SA_INC(asize - old_size);
 		    goto supercarrier_resize_success;
 		}
 	    }
 	    else {
-		adjacent_free_seg(&mmap_state.sa.map, start, end, &prev, &next);
+		adjacent_free_seg(&mm->sa.map, start, end, &prev, &next);
 		if (next && new_end <= next->end) {
-		    if (!mmap_state.reserve_physical(((char *) ptr) + old_size,
+		    if (!mm->reserve_physical(((char *) ptr) + old_size,
 						     asize - old_size))
 			goto supercarrier_reserve_failure;
 		    if (new_end < next->end)
-			resize_free_seg(&mmap_state.sa.map, next, new_end, next->end);
+			resize_free_seg(&mm->sa.map, next, new_end, next->end);
 		    else {
-			delete_free_seg(&mmap_state.sa.map, next);
-			free_desc(next);
+			delete_free_seg(&mm->sa.map, next);
+			free_desc(mm, next);
 		    }
 		    new_ptr = ptr;
 		    ERTS_MMAP_SIZE_SC_SA_INC(asize - old_size);
@@ -2000,12 +2010,12 @@ erts_mremap(Uint32 flags, void *ptr, UWord old_size, UWord *sizep)
 	}
 
 	ERTS_MMAP_OP_ABORT();
-	erts_smp_mtx_unlock(&mmap_state.mtx);
+	erts_smp_mtx_unlock(&mm->mtx);
 
 	/* Failed to resize... */
     }
 
-    return remap_move(flags, ptr, old_size, sizep);
+    return remap_move(mm, flags, ptr, old_size, sizep);
 
 supercarrier_resize_success:
 
@@ -2022,25 +2032,24 @@ supercarrier_resize_success:
 #endif
 
     ERTS_MREMAP_OP_END(new_ptr, asize);
-    erts_smp_mtx_unlock(&mmap_state.mtx);
+    erts_smp_mtx_unlock(&mm->mtx);
 
     *sizep = asize;
     return new_ptr;
 
 supercarrier_reserve_failure:
     ERTS_MREMAP_OP_END(NULL, old_size);
-    erts_smp_mtx_unlock(&mmap_state.mtx);
+    erts_smp_mtx_unlock(&mm->mtx);
     *sizep = old_size;
     return NULL;
     
 }
 
-int erts_mmap_in_supercarrier(void *ptr)
+int erts_mmap_in_supercarrier(ErtsMemMapper* mm, void *ptr)
 {
     return ERTS_MMAP_IN_SUPERCARRIER(ptr);
 }
 
-
 static struct {
     Eterm total;
     Eterm total_sa;
@@ -2103,8 +2112,9 @@ static void hard_dbg_mseg_init(void);
 #endif
 
 void
-erts_mmap_init(ErtsMMapInit *init)
+erts_mmap_init(ErtsMemMapper* mm, ErtsMMapInit *init)
 {
+    static int is_first_call = 1;
     int virtual_map = 0;
     char *start = NULL, *end = NULL;
     UWord pagesize;
@@ -2131,20 +2141,20 @@ erts_mmap_init(ErtsMMapInit *init)
 
     ERTS_MMAP_OP_RINGBUF_INIT();
 
-    erts_have_erts_mmap = 0;
-
-    mmap_state.supercarrier = 0;
-    mmap_state.reserve_physical = reserve_noop;
-    mmap_state.unreserve_physical = unreserve_noop;
+    mm->supercarrier = 0;
+    mm->reserve_physical = reserve_noop;
+    mm->unreserve_physical = unreserve_noop;
 
 #if HAVE_MMAP && !defined(MAP_ANON)
-    mmap_state.mmap_fd = open("/dev/zero", O_RDWR);
-    if (mmap_state.mmap_fd < 0)
+    mm->mmap_fd = open("/dev/zero", O_RDWR);
+    if (mm->mmap_fd < 0)
 	erl_exit(-1, "erts_mmap: Failed to open /dev/zero\n");
 #endif
 
-    erts_smp_mtx_init(&mmap_state.mtx, "erts_mmap");
-    erts_mtx_init(&am.init_mutex, "mmap_init_atoms");
+    erts_smp_mtx_init(&mm->mtx, "erts_mmap");
+    if (is_first_call) {
+        erts_mtx_init(&am.init_mutex, "mmap_init_atoms");
+    }
 
 #ifdef ERTS_HAVE_OS_PHYSICAL_MEMORY_RESERVATION
     if (init->virtual_range.start) {
@@ -2160,8 +2170,8 @@ erts_mmap_init(ErtsMMapInit *init)
 	sz = start - ptr;
 	if (sz)
 	    os_munmap(end, sz);
-	mmap_state.reserve_physical = os_reserve_physical;
-	mmap_state.unreserve_physical = os_unreserve_physical;
+	mm->reserve_physical = os_reserve_physical;
+	mm->unreserve_physical = os_unreserve_physical;
 	virtual_map = 1;
     }
     else
@@ -2179,8 +2189,8 @@ erts_mmap_init(ErtsMMapInit *init)
 #ifdef ERTS_HAVE_OS_PHYSICAL_MEMORY_RESERVATION
 	if (!init->scrpm) {
 	    start = os_mmap_virtual(NULL, sz);
-	    mmap_state.reserve_physical = os_reserve_physical;
-	    mmap_state.unreserve_physical = os_unreserve_physical;
+	    mm->reserve_physical = os_reserve_physical;
+	    mm->unreserve_physical = os_unreserve_physical;
 	    virtual_map = 1;
 	}
 	else
@@ -2206,34 +2216,32 @@ erts_mmap_init(ErtsMMapInit *init)
 	}
 #endif
     }
-    if (!mmap_state.no_os_mmap)
-	erts_have_erts_mmap |= ERTS_HAVE_ERTS_OS_MMAP;
 #endif
 
-    mmap_state.no.free_seg_descs = 0;
-    mmap_state.no.free_segs.curr = 0;
-    mmap_state.no.free_segs.max = 0;
+    mm->no.free_seg_descs = 0;
+    mm->no.free_segs.curr = 0;
+    mm->no.free_segs.max = 0;
 
-    mmap_state.size.supercarrier.total = 0;
-    mmap_state.size.supercarrier.used.total = 0;
-    mmap_state.size.supercarrier.used.sa = 0;
-    mmap_state.size.supercarrier.used.sua = 0;
-    mmap_state.size.os.used = 0;
+    mm->size.supercarrier.total = 0;
+    mm->size.supercarrier.used.total = 0;
+    mm->size.supercarrier.used.sa = 0;
+    mm->size.supercarrier.used.sua = 0;
+    mm->size.os.used = 0;
 
-    mmap_state.desc.new_area_hint = NULL;
+    mm->desc.new_area_hint = NULL;
 
     if (!start) {
-	mmap_state.sa.bot = NULL;
-	mmap_state.sua.top = NULL;
-	mmap_state.sa.bot = NULL;
-	mmap_state.sua.top = NULL;
-	mmap_state.no_os_mmap = 0;
-        mmap_state.supercarrier = 0;
+	mm->sa.bot = NULL;
+	mm->sua.top = NULL;
+	mm->sa.bot = NULL;
+	mm->sua.top = NULL;
+	mm->no_os_mmap = 0;
+        mm->supercarrier = 0;
     }
     else {
 	size_t desc_size;
 
-	mmap_state.no_os_mmap = init->sco;
+	mm->no_os_mmap = init->sco;
 
 	desc_size = init->scrfsd;
 	if (desc_size < 100)
@@ -2244,66 +2252,73 @@ erts_mmap_init(ErtsMMapInit *init)
 	     + ERTS_PAGEALIGNED_SIZE) > end - start)
 	    erl_exit(-1, "erts_mmap: No space for segments in super carrier\n");
 
-	mmap_state.sa.bot = start;
-	mmap_state.sa.bot += desc_size;
-	mmap_state.sa.bot = (char *) ERTS_SUPERALIGNED_CEILING(mmap_state.sa.bot);
-	mmap_state.sa.top = mmap_state.sa.bot;
-	mmap_state.sua.top = end;
-	mmap_state.sua.bot = mmap_state.sua.top;
+	mm->sa.bot = start;
+	mm->sa.bot += desc_size;
+	mm->sa.bot = (char *) ERTS_SUPERALIGNED_CEILING(mm->sa.bot);
+	mm->sa.top = mm->sa.bot;
+	mm->sua.top = end;
+	mm->sua.bot = mm->sua.top;
 
-	mmap_state.size.supercarrier.used.total += (UWord) (mmap_state.sa.bot - start);
+	mm->size.supercarrier.used.total += (UWord) (mm->sa.bot - start);
 
-	mmap_state.desc.free_list = NULL;
-        mmap_state.desc.reserved = 0;
+	mm->desc.free_list = NULL;
+        mm->desc.reserved = 0;
 
 	if (end == (void *) 0) {
 	    /*
 	     * Very unlikely, but we need a guarantee
-	     * that `mmap_state.sua.top` always will
+	     * that `mm->sua.top` always will
 	     * compare as larger than all segment pointers
 	     * into the super carrier...
 	     */
-	    mmap_state.sua.top -= ERTS_PAGEALIGNED_SIZE;
-	    mmap_state.size.supercarrier.used.total += ERTS_PAGEALIGNED_SIZE;
+	    mm->sua.top -= ERTS_PAGEALIGNED_SIZE;
+	    mm->size.supercarrier.used.total += ERTS_PAGEALIGNED_SIZE;
 #ifdef ERTS_HAVE_OS_PHYSICAL_MEMORY_RESERVATION
-	    if (!virtual_map || os_reserve_physical(mmap_state.sua.top, ERTS_PAGEALIGNED_SIZE))
+	    if (!virtual_map || os_reserve_physical(mm->sua.top, ERTS_PAGEALIGNED_SIZE))
 #endif
-		add_free_desc_area(mmap_state.sua.top, end);
-            mmap_state.desc.reserved += (end - mmap_state.sua.top) / sizeof(ErtsFreeSegDesc);
+		add_free_desc_area(mm, mm->sua.top, end);
+            mm->desc.reserved += (end - mm->sua.top) / sizeof(ErtsFreeSegDesc);
 	}
 
-	mmap_state.size.supercarrier.total = (UWord) (mmap_state.sua.top - start);
+	mm->size.supercarrier.total = (UWord) (mm->sua.top - start);
 
 	/*
 	 * Area before (and after) super carrier
 	 * will be used for free segment descritors.
 	 */
 #ifdef ERTS_HAVE_OS_PHYSICAL_MEMORY_RESERVATION
-	if (virtual_map && !os_reserve_physical(start, mmap_state.sa.bot - start))
+	if (virtual_map && !os_reserve_physical(start, mm->sa.bot - start))
 	    erl_exit(-1, "erts_mmap: Failed to reserve physical memory for descriptors\n");
 #endif
-	mmap_state.desc.unused_start = start;
-	mmap_state.desc.unused_end = mmap_state.sa.bot;
-        mmap_state.desc.reserved += ((mmap_state.desc.unused_end - start)
+	mm->desc.unused_start = start;
+	mm->desc.unused_end = mm->sa.bot;
+        mm->desc.reserved += ((mm->desc.unused_end - start)
                                      / sizeof(ErtsFreeSegDesc));
 
-	init_free_seg_map(&mmap_state.sa.map, SA_SZ_ADDR_ORDER);
-	init_free_seg_map(&mmap_state.sua.map, SZ_REVERSE_ADDR_ORDER);
+	init_free_seg_map(&mm->sa.map, SA_SZ_ADDR_ORDER);
+	init_free_seg_map(&mm->sua.map, SZ_REVERSE_ADDR_ORDER);
 
-	mmap_state.supercarrier = 1;
-	erts_have_erts_mmap |= ERTS_HAVE_ERTS_SUPERCARRIER_MMAP;
+	mm->supercarrier = 1;
 
-	mmap_state.desc.new_area_hint = end;
+	mm->desc.new_area_hint = end;
 
     }
 
 #if !ERTS_HAVE_OS_MMAP
-    mmap_state.no_os_mmap = 1;
+    mm->no_os_mmap = 1;
 #endif
 
 #ifdef HARD_DEBUG_MSEG
     hard_dbg_mseg_init();
 #endif
+
+#if defined(ARCH_64) && defined(ERTS_HAVE_OS_PHYSICAL_MEMORY_RESERVATION)
+   if (mm == &erts_literal_mmapper) {
+       erts_literals_start = erts_literal_mmapper.sa.bot;
+       erts_literals_size  = erts_literal_mmapper.sua.top - erts_literals_start;
+    }
+#endif
+    is_first_call = 0;
 }
 
 
@@ -2313,7 +2328,8 @@ add_2tup(Uint **hpp, Uint *szp, Eterm *lp, Eterm el1, Eterm el2)
     *lp = erts_bld_cons(hpp, szp, erts_bld_tuple(hpp, szp, 2, el1, el2), *lp);
 }
 
-Eterm erts_mmap_info(int *print_to_p,
+Eterm erts_mmap_info(ErtsMemMapper* mm,
+                     int *print_to_p,
                      void *print_to_arg,
                      Eterm** hpp, Uint* szp,
                      struct erts_mmap_info_struct* emis)
@@ -2328,29 +2344,29 @@ Eterm erts_mmap_info(int *print_to_p,
     Eterm res = THE_NON_VALUE;
 
     if (!hpp) {
-        erts_smp_mtx_lock(&mmap_state.mtx);
-        emis->sizes[0] = mmap_state.size.supercarrier.total;
-        emis->sizes[1] = mmap_state.sa.top  - mmap_state.sa.bot;
-        emis->sizes[2] = mmap_state.sua.top - mmap_state.sua.bot;
-        emis->sizes[3] = mmap_state.size.supercarrier.used.total;
-        emis->sizes[4] = mmap_state.size.supercarrier.used.sa;
-        emis->sizes[5] = mmap_state.size.supercarrier.used.sua;
+        erts_smp_mtx_lock(&mm->mtx);
+        emis->sizes[0] = mm->size.supercarrier.total;
+        emis->sizes[1] = mm->sa.top  - mm->sa.bot;
+        emis->sizes[2] = mm->sua.top - mm->sua.bot;
+        emis->sizes[3] = mm->size.supercarrier.used.total;
+        emis->sizes[4] = mm->size.supercarrier.used.sa;
+        emis->sizes[5] = mm->size.supercarrier.used.sua;
 
-        emis->segs[0] = mmap_state.no.free_segs.curr;
-        emis->segs[1] = mmap_state.no.free_segs.max;
-        emis->segs[2] = mmap_state.no.free_seg_descs;
-        emis->segs[3] = mmap_state.desc.reserved;
-        emis->segs[4] = mmap_state.sa.map.nseg;
-        emis->segs[5] = mmap_state.sua.map.nseg;
+        emis->segs[0] = mm->no.free_segs.curr;
+        emis->segs[1] = mm->no.free_segs.max;
+        emis->segs[2] = mm->no.free_seg_descs;
+        emis->segs[3] = mm->desc.reserved;
+        emis->segs[4] = mm->sa.map.nseg;
+        emis->segs[5] = mm->sua.map.nseg;
 
-        emis->os_used = mmap_state.size.os.used;
-        erts_smp_mtx_unlock(&mmap_state.mtx);
+        emis->os_used = mm->size.os.used;
+        erts_smp_mtx_unlock(&mm->mtx);
     }
 
     if (print_to_p) {
         int to = *print_to_p;
 	void *arg = print_to_arg;
-        if (mmap_state.supercarrier) {
+        if (mm->supercarrier) {
             const char* prefix = "supercarrier ";
             erts_print(to, arg, "%stotal size: %bpu\n", prefix, emis->sizes[0]);
             erts_print(to, arg, "%stotal sa size: %bpu\n", prefix, emis->sizes[1]);
@@ -2365,7 +2381,7 @@ Eterm erts_mmap_info(int *print_to_p,
             erts_print(to, arg, "%ssa free segs: %bpu\n", prefix, emis->segs[4]);
             erts_print(to, arg, "%ssua free segs: %bpu\n", prefix, emis->segs[5]);
         }
-        if (!mmap_state.no_os_mmap) {
+        if (!mm->no_os_mmap) {
             erts_print(to, arg, "os mmap size used: %bpu\n", emis->os_used);
         }
     }
@@ -2377,7 +2393,7 @@ Eterm erts_mmap_info(int *print_to_p,
         }
 
         lix = 0;
-        if (mmap_state.supercarrier) {
+        if (mm->supercarrier) {
             group[0] = erts_bld_atom_uword_2tup_list(hpp, szp,
                                                      sizeof(size_tags)/sizeof(Eterm),
                                                      size_tags, emis->sizes);
@@ -2389,7 +2405,7 @@ Eterm erts_mmap_info(int *print_to_p,
             lix++;
         }
 
-        if (!mmap_state.no_os_mmap) {
+        if (!mm->no_os_mmap) {
             group[0] = erts_bld_atom_uword_2tup_list(hpp, szp,
                                                       1, &am.used, &emis->os_used);
             list[lix] = erts_bld_2tup_list(hpp, szp, 1, group_tags, group);
@@ -2401,25 +2417,26 @@ Eterm erts_mmap_info(int *print_to_p,
     return res;
 }
 
-Eterm erts_mmap_info_options(char *prefix,
+Eterm erts_mmap_info_options(ErtsMemMapper* mm,
+                             char *prefix,
                              int *print_to_p,
                              void *print_to_arg,
                              Uint **hpp,
                              Uint *szp)
 {
-    const UWord scs = mmap_state.sua.top - mmap_state.sa.bot;
-    const Eterm sco = mmap_state.no_os_mmap ? am_true : am_false;
-    const Eterm scrpm = (mmap_state.reserve_physical == reserve_noop) ? am_true : am_false;
+    const UWord scs = mm->sua.top - mm->sa.bot;
+    const Eterm sco = mm->no_os_mmap ? am_true : am_false;
+    const Eterm scrpm = (mm->reserve_physical == reserve_noop) ? am_true : am_false;
     Eterm res = THE_NON_VALUE;
 
     if (print_to_p) {
         int to = *print_to_p;
 	void *arg = print_to_arg;
         erts_print(to, arg, "%sscs: %bpu\n", prefix, scs);
-        if (mmap_state.supercarrier) {
+        if (mm->supercarrier) {
             erts_print(to, arg, "%ssco: %T\n", prefix, sco);
             erts_print(to, arg, "%sscrpm: %T\n", prefix, scrpm);
-            erts_print(to, arg, "%sscrfsd: %beu\n", prefix, mmap_state.desc.reserved);
+            erts_print(to, arg, "%sscrfsd: %beu\n", prefix, mm->desc.reserved);
         }
     }
 
@@ -2429,9 +2446,9 @@ Eterm erts_mmap_info_options(char *prefix,
         }
 
         res = NIL;
-        if (mmap_state.supercarrier) {
+        if (mm->supercarrier) {
             add_2tup(hpp, szp, &res, am.scrfsd,
-                     erts_bld_uint(hpp,szp, mmap_state.desc.reserved));
+                     erts_bld_uint(hpp,szp, mm->desc.reserved));
             add_2tup(hpp, szp, &res, am.scrpm, scrpm);
             add_2tup(hpp, szp, &res, am.sco, sco);
         }
@@ -2441,9 +2458,9 @@ Eterm erts_mmap_info_options(char *prefix,
 }
 
 
-Eterm erts_mmap_debug_info(Process* p)
+Eterm erts_mmap_debug_info(ErtsMemMapper* mm, Process* p)
 {
-    if (mmap_state.supercarrier) {
+    if (mm->supercarrier) {
         ERTS_DECL_AM(sabot);
         ERTS_DECL_AM(satop);
         ERTS_DECL_AM(suabot);
@@ -2453,18 +2470,17 @@ Eterm erts_mmap_debug_info(Process* p)
         UWord values[4];
         Eterm *hp, *hp_end;
         Uint may_need;
-        const Uint PTR_BIG_SZ = HALFWORD_HEAP ? 3 : 2;
-
-        erts_smp_mtx_lock(&mmap_state.mtx);
-        values[0] = (UWord)mmap_state.sa.bot;
-        values[1] = (UWord)mmap_state.sa.top;
-        values[2] = (UWord)mmap_state.sua.bot;
-        values[3] = (UWord)mmap_state.sua.top;
-        sa_list = build_free_seg_list(p, &mmap_state.sa.map);
-        sua_list = build_free_seg_list(p, &mmap_state.sua.map);
-        erts_smp_mtx_unlock(&mmap_state.mtx);
-
-        may_need = 4*(2+3+PTR_BIG_SZ) + 2*(2+3);
+
+        erts_smp_mtx_lock(&mm->mtx);
+        values[0] = (UWord)mm->sa.bot;
+        values[1] = (UWord)mm->sa.top;
+        values[2] = (UWord)mm->sua.bot;
+        values[3] = (UWord)mm->sua.top;
+        sa_list = build_free_seg_list(p, &mm->sa.map);
+        sua_list = build_free_seg_list(p, &mm->sua.map);
+        erts_smp_mtx_unlock(&mm->mtx);
+
+        may_need = 4*(2+3+2) + 2*(2+3);
         hp = HAlloc(p, may_need);
         hp_end = hp + may_need;
 
diff --git a/erts/emulator/sys/common/erl_mmap.h b/erts/emulator/sys/common/erl_mmap.h
index 66619c5161..61d912fd28 100644
--- a/erts/emulator/sys/common/erl_mmap.h
+++ b/erts/emulator/sys/common/erl_mmap.h
@@ -30,9 +30,6 @@
 #define ERTS_MMAPFLG_SUPERCARRIER_ONLY		(((Uint32) 1) << 1)
 #define ERTS_MMAPFLG_SUPERALIGNED		(((Uint32) 1) << 2)
 
-#define ERTS_HAVE_ERTS_OS_MMAP			(1 << 0)
-#define ERTS_HAVE_ERTS_SUPERCARRIER_MMAP	(1 << 1)
-extern int erts_have_erts_mmap;
 extern UWord erts_page_inv_mask;
 
 typedef struct {
@@ -53,23 +50,29 @@ typedef struct {
 #define ERTS_MMAP_INIT_DEFAULT_INITER \
     {{NULL, NULL}, {NULL, NULL}, 0, 1, (1 << 16), 1}
 
-void *erts_mmap(Uint32 flags, UWord *sizep);
-void erts_munmap(Uint32 flags, void *ptr, UWord size);
-void *erts_mremap(Uint32 flags, void *ptr, UWord old_size, UWord *sizep);
-int erts_mmap_in_supercarrier(void *ptr);
-void erts_mmap_init(ErtsMMapInit*);
+#define ERTS_MMAP_INIT_LITERAL_INITER \
+    {{NULL, NULL}, {NULL, NULL}, 1024*1024*1024, 1, (1 << 16), 0}
+
+typedef struct ErtsMemMapper_ ErtsMemMapper;
+
+void *erts_mmap(ErtsMemMapper*, Uint32 flags, UWord *sizep);
+void erts_munmap(ErtsMemMapper*, Uint32 flags, void *ptr, UWord size);
+void *erts_mremap(ErtsMemMapper*, Uint32 flags, void *ptr, UWord old_size, UWord *sizep);
+int erts_mmap_in_supercarrier(ErtsMemMapper*, void *ptr);
+void erts_mmap_init(ErtsMemMapper*, ErtsMMapInit*);
 struct erts_mmap_info_struct
 {
     UWord sizes[6];
     UWord segs[6];
     UWord os_used;
 };
-Eterm erts_mmap_info(int *print_to_p, void *print_to_arg,
+Eterm erts_mmap_info(ErtsMemMapper*, int *print_to_p, void *print_to_arg,
                      Eterm** hpp, Uint* szp, struct erts_mmap_info_struct*);
-Eterm erts_mmap_info_options(char *prefix, int *print_to_p, void *print_to_arg,
+Eterm erts_mmap_info_options(ErtsMemMapper*,
+                             char *prefix, int *print_to_p, void *print_to_arg,
                              Uint **hpp, Uint *szp);
 struct process;
-Eterm erts_mmap_debug_info(struct process*);
+Eterm erts_mmap_debug_info(ErtsMemMapper*, struct process*);
 
 #define ERTS_SUPERALIGNED_SIZE \
     (1 << ERTS_MMAP_SUPERALIGNED_BITS)
@@ -121,6 +124,11 @@ Eterm erts_mmap_debug_info(struct process*);
 #  define ERTS_HAVE_OS_MMAP 1
 #endif
 
+extern ErtsMemMapper erts_dflt_mmapper;
+#if defined(ARCH_64) && defined(ERTS_HAVE_OS_PHYSICAL_MEMORY_RESERVATION)
+extern ErtsMemMapper erts_literal_mmapper;
+#endif
+
 /*#define HARD_DEBUG_MSEG*/
 #ifdef HARD_DEBUG_MSEG
 #  define HARD_DBG_INSERT_MSEG hard_dbg_insert_mseg
diff --git a/erts/emulator/sys/common/erl_mseg.c b/erts/emulator/sys/common/erl_mseg.c
index 0d51aad863..20695899eb 100644
--- a/erts/emulator/sys/common/erl_mseg.c
+++ b/erts/emulator/sys/common/erl_mseg.c
@@ -99,17 +99,12 @@ static const int debruijn[32] = {
 
 static int atoms_initialized;
 
-typedef struct mem_kind_t MemKind;
-
 const ErtsMsegOpt_t erts_mseg_default_opt = {
     1,			/* Use cache		     */
     1,			/* Preserv data		     */
     0,			/* Absolute shrink threshold */
     0,			/* Relative shrink threshold */
     0			/* Scheduler specific        */
-#if HALFWORD_HEAP
-    ,0                  /* need low memory */
-#endif
 };
 
 
@@ -142,7 +137,14 @@ struct cache_t_ {
 
 typedef struct ErtsMsegAllctr_t_ ErtsMsegAllctr_t;
 
-struct mem_kind_t {
+struct ErtsMsegAllctr_t_ {
+    int ix;
+
+    int is_init_done;
+    int is_thread_safe;
+    erts_mtx_t mtx;
+
+    int is_cache_check_scheduled;
 
     cache_t cache[MAX_CACHE_SIZE];
     cache_t cache_unpowered_node;
@@ -168,29 +170,6 @@ struct mem_kind_t {
 	} max_ever;
     } segments;
 
-    ErtsMsegAllctr_t *ma;
-    const char* name;
-    MemKind* next;
-};/*MemKind*/
-
-struct ErtsMsegAllctr_t_ {
-    int ix;
-
-    int is_init_done;
-    int is_thread_safe;
-    erts_mtx_t mtx;
-
-    int is_cache_check_scheduled;
-
-    MemKind* mk_list;
-
-#if HALFWORD_HEAP
-    MemKind low_mem;
-    MemKind hi_mem;
-#else
-    MemKind the_mem;
-#endif
-
     Uint max_cache_size;
     Uint abs_max_cache_bad_fit;
     Uint rel_max_cache_bad_fit;
@@ -302,22 +281,17 @@ schedule_cache_check(ErtsMsegAllctr_t *ma) {
 /* #define ERTS_PRINT_ERTS_MMAP */
 
 static ERTS_INLINE void *
-mseg_create(ErtsMsegAllctr_t *ma, Uint flags, MemKind* mk, UWord *sizep)
+mseg_create(ErtsMsegAllctr_t *ma, Uint flags, UWord *sizep)
 {
 #ifdef ERTS_PRINT_ERTS_MMAP
     UWord req_size = *sizep;
 #endif
     void *seg;
     Uint32 mmap_flags = 0;
-#if HALFWORD_HEAP
-    mmap_flags |= ((mk == &ma->low_mem)
-		   ? ERTS_MMAPFLG_SUPERCARRIER_ONLY
-		   : ERTS_MMAPFLG_OS_ONLY);
-#endif
     if (MSEG_FLG_IS_2POW(flags))
 	mmap_flags |= ERTS_MMAPFLG_SUPERALIGNED;
 
-    seg = erts_mmap(mmap_flags, sizep);
+    seg = erts_mmap(&erts_dflt_mmapper, mmap_flags, sizep);
 
 #ifdef ERTS_PRINT_ERTS_MMAP
     erts_fprintf(stderr, "%p = erts_mmap(%s, {%bpu, %bpu});\n", seg,
@@ -331,18 +305,13 @@ mseg_create(ErtsMsegAllctr_t *ma, Uint flags, MemKind* mk, UWord *sizep)
 }
 
 static ERTS_INLINE void
-mseg_destroy(ErtsMsegAllctr_t *ma, Uint flags, MemKind* mk, void *seg_p, UWord size) {
+mseg_destroy(ErtsMsegAllctr_t *ma, Uint flags, void *seg_p, UWord size) {
     
     Uint32 mmap_flags = 0;
-#if HALFWORD_HEAP
-    mmap_flags |= ((mk == &ma->low_mem)
-		   ? ERTS_MMAPFLG_SUPERCARRIER_ONLY
-		   : ERTS_MMAPFLG_OS_ONLY);
-#endif
     if (MSEG_FLG_IS_2POW(flags))
 	 mmap_flags |= ERTS_MMAPFLG_SUPERALIGNED;
 
-    erts_munmap(mmap_flags, seg_p, size);
+    erts_munmap(&erts_dflt_mmapper, mmap_flags, seg_p, size);
 #ifdef ERTS_PRINT_ERTS_MMAP
     erts_fprintf(stderr, "erts_munmap(%s, %p, %bpu);\n",
 		 (mmap_flags & ERTS_MMAPFLG_SUPERALIGNED) ? "sa" : "sua",
@@ -353,22 +322,17 @@ mseg_destroy(ErtsMsegAllctr_t *ma, Uint flags, MemKind* mk, void *seg_p, UWord s
 }
 
 static ERTS_INLINE void *
-mseg_recreate(ErtsMsegAllctr_t *ma, Uint flags, MemKind* mk, void *old_seg, UWord old_size, UWord *sizep)
+mseg_recreate(ErtsMsegAllctr_t *ma, Uint flags, void *old_seg, UWord old_size, UWord *sizep)
 {
 #ifdef ERTS_PRINT_ERTS_MMAP
     UWord req_size = *sizep;	
 #endif
     void *new_seg;
     Uint32 mmap_flags = 0;
-#if HALFWORD_HEAP
-    mmap_flags |= ((mk == &ma->low_mem)
-		   ? ERTS_MMAPFLG_SUPERCARRIER_ONLY
-		   : ERTS_MMAPFLG_OS_ONLY);
-#endif
     if (MSEG_FLG_IS_2POW(flags))
 	mmap_flags |= ERTS_MMAPFLG_SUPERALIGNED;
 
-    new_seg = erts_mremap(mmap_flags, old_seg, old_size, sizep);
+    new_seg = erts_mremap(&erts_dflt_mmapper, mmap_flags, old_seg, old_size, sizep);
 
 #ifdef ERTS_PRINT_ERTS_MMAP
     erts_fprintf(stderr, "%p = erts_mremap(%s, %p, %bpu, {%bpu, %bpu});\n",
@@ -392,11 +356,8 @@ do {									\
 		       || erts_smp_thr_progress_is_blocking()		\
 		       || ERTS_IS_CRASH_DUMPING);			\
 } while (0)
-#define ERTS_DBG_MK_CHK_THR_ACCESS(MK) \
-  ERTS_DBG_MA_CHK_THR_ACCESS((MK)->ma)
 #else
 #define ERTS_DBG_MA_CHK_THR_ACCESS(MA)
-#define ERTS_DBG_MK_CHK_THR_ACCESS(MK)
 #endif
 
 /* Cache interface */
@@ -409,10 +370,10 @@ static ERTS_INLINE void mseg_cache_clear_node(cache_t *c) {
     c->prev = c;
 }
 
-static ERTS_INLINE int cache_bless_segment(MemKind *mk, void *seg, UWord size, Uint flags) {
+static ERTS_INLINE int cache_bless_segment(ErtsMsegAllctr_t *ma, void *seg, UWord size, Uint flags) {
 
     cache_t *c;
-    ERTS_DBG_MK_CHK_THR_ACCESS(mk);
+    ERTS_DBG_MA_CHK_THR_ACCESS(ma);
 
     ASSERT(!MSEG_FLG_IS_2POW(flags) || (MSEG_FLG_IS_2POW(flags) && MAP_IS_ALIGNED(seg) && IS_2POW(size)));
 
@@ -421,11 +382,11 @@ static ERTS_INLINE int cache_bless_segment(MemKind *mk, void *seg, UWord size, U
      * Large blocks has no such cache and it is up to mseg to cache them to speed things up.
      */
 
-    if (!erts_circleq_is_empty(&(mk->cache_free))) {
+    if (!erts_circleq_is_empty(&(ma->cache_free))) {
 
 	/* We have free slots, use one to cache the segment */
 
-	c = erts_circleq_head(&(mk->cache_free));
+	c = erts_circleq_head(&(ma->cache_free));
 	erts_circleq_remove(c);
 
 	c->seg  = seg;
@@ -437,29 +398,28 @@ static ERTS_INLINE int cache_bless_segment(MemKind *mk, void *seg, UWord size, U
 	    ASSERT(ix < CACHE_AREAS);
 	    ASSERT((1 << (ix + MSEG_ALIGN_BITS)) == size);
 
-	    erts_circleq_push_head(&(mk->cache_powered_node[ix]), c);
+	    erts_circleq_push_head(&(ma->cache_powered_node[ix]), c);
 
 	} else
-	    erts_circleq_push_head(&(mk->cache_unpowered_node), c);
+	    erts_circleq_push_head(&(ma->cache_unpowered_node), c);
 
-	mk->cache_size++;
-	ASSERT(mk->cache_size <= mk->ma->max_cache_size);
+	ma->cache_size++;
 
 	return 1;
-    } else if (!MSEG_FLG_IS_2POW(flags) && !erts_circleq_is_empty(&(mk->cache_unpowered_node))) {
+    } else if (!MSEG_FLG_IS_2POW(flags) && !erts_circleq_is_empty(&(ma->cache_unpowered_node))) {
 	/* No free slots.
 	 * Evict oldest slot from unpowered cache so we can cache an unpowered (sbc) segment  */
 
-	c = erts_circleq_tail(&(mk->cache_unpowered_node));
+	c = erts_circleq_tail(&(ma->cache_unpowered_node));
 	erts_circleq_remove(c);
 
-	mseg_destroy(mk->ma, ERTS_MSEG_FLG_NONE, mk, c->seg, c->size);
+	mseg_destroy(ma, ERTS_MSEG_FLG_NONE, c->seg, c->size);
 	mseg_cache_clear_node(c);
 
 	c->seg  = seg;
 	c->size = size;
 
-	erts_circleq_push_head(&(mk->cache_unpowered_node), c);
+	erts_circleq_push_head(&(ma->cache_unpowered_node), c);
 
 	return 1;
     } else if (!MSEG_FLG_IS_2POW(flags)) {
@@ -473,20 +433,20 @@ static ERTS_INLINE int cache_bless_segment(MemKind *mk, void *seg, UWord size, U
 	int i;
 
 	for( i = 0; i < CACHE_AREAS; i++) {
-	    if (erts_circleq_is_empty(&(mk->cache_powered_node[i])))
+	    if (erts_circleq_is_empty(&(ma->cache_powered_node[i])))
 		continue;
 
-	    c = erts_circleq_tail(&(mk->cache_powered_node[i]));
+	    c = erts_circleq_tail(&(ma->cache_powered_node[i]));
 	    erts_circleq_remove(c);
 
-	    mseg_destroy(mk->ma, ERTS_MSEG_FLG_2POW, mk, c->seg, c->size);
+	    mseg_destroy(ma, ERTS_MSEG_FLG_2POW, c->seg, c->size);
 
 	    mseg_cache_clear_node(c);
 
 	    c->seg  = seg;
 	    c->size = size;
 
-	    erts_circleq_push_head(&(mk->cache_unpowered_node), c);
+	    erts_circleq_push_head(&(ma->cache_unpowered_node), c);
 
 	    return 1;
 	}
@@ -495,11 +455,11 @@ static ERTS_INLINE int cache_bless_segment(MemKind *mk, void *seg, UWord size, U
     return 0;
 }
 
-static ERTS_INLINE void *cache_get_segment(MemKind *mk, UWord *size_p, Uint flags) {
+static ERTS_INLINE void *cache_get_segment(ErtsMsegAllctr_t *ma, UWord *size_p, Uint flags) {
 
     UWord size = *size_p;
 
-    ERTS_DBG_MK_CHK_THR_ACCESS(mk);
+    ERTS_DBG_MA_CHK_THR_ACCESS(ma);
 
     if (MSEG_FLG_IS_2POW(flags)) {
 
@@ -512,10 +472,10 @@ static ERTS_INLINE void *cache_get_segment(MemKind *mk, UWord *size_p, Uint flag
 
 	for( i = ix; i < CACHE_AREAS; i++) {
 
-	    if (erts_circleq_is_empty(&(mk->cache_powered_node[i])))
+	    if (erts_circleq_is_empty(&(ma->cache_powered_node[i])))
 		continue;
 
-	    c = erts_circleq_head(&(mk->cache_powered_node[i]));
+	    c = erts_circleq_head(&(ma->cache_powered_node[i]));
 	    erts_circleq_remove(c);
 
 	    ASSERT(IS_2POW(c->size));
@@ -524,31 +484,31 @@ static ERTS_INLINE void *cache_get_segment(MemKind *mk, UWord *size_p, Uint flag
 	    csize = c->size;
 	    seg   = (char*) c->seg;
 
-	    mk->cache_size--;
-	    mk->cache_hits++;
+	    ma->cache_size--;
+	    ma->cache_hits++;
 
 	    /* link to free cache list */
 	    mseg_cache_clear_node(c);
-	    erts_circleq_push_head(&(mk->cache_free), c);
+	    erts_circleq_push_head(&(ma->cache_free), c);
 
-	    ASSERT(!(mk->cache_size < 0));
+	    ASSERT(!(ma->cache_size < 0));
 
 	    if (csize != size)
-		mseg_destroy(mk->ma, ERTS_MSEG_FLG_2POW, mk, seg + size, csize - size);
+		mseg_destroy(ma, ERTS_MSEG_FLG_2POW, seg + size, csize - size);
 
 	    return seg;
 	}
     } 
-    else if (!erts_circleq_is_empty(&(mk->cache_unpowered_node))) {
+    else if (!erts_circleq_is_empty(&(ma->cache_unpowered_node))) {
 	void *seg;
 	cache_t *c;
 	cache_t *best = NULL;
 	UWord bdiff = 0;
 	UWord csize;
-	UWord bad_max_abs = mk->ma->abs_max_cache_bad_fit;
-	UWord bad_max_rel = mk->ma->rel_max_cache_bad_fit;
+	UWord bad_max_abs = ma->abs_max_cache_bad_fit;
+	UWord bad_max_rel = ma->rel_max_cache_bad_fit;
 
-	erts_circleq_foreach(c, &(mk->cache_unpowered_node)) {
+	erts_circleq_foreach(c, &(ma->cache_unpowered_node)) {
 	    csize = c->size;
 	    if (csize >= size) {
 		if (((csize - size)*100 < bad_max_rel*size) && (csize - size) < bad_max_abs ) {
@@ -557,11 +517,11 @@ static ERTS_INLINE void *cache_get_segment(MemKind *mk, UWord *size_p, Uint flag
 
 		    erts_circleq_remove(c);
 
-		    mk->cache_size--;
-		    mk->cache_hits++;
+		    ma->cache_size--;
+		    ma->cache_hits++;
 
 		    mseg_cache_clear_node(c);
-		    erts_circleq_push_head(&(mk->cache_free), c);
+		    erts_circleq_push_head(&(ma->cache_free), c);
 
 		    *size_p = csize;
 
@@ -584,7 +544,7 @@ static ERTS_INLINE void *cache_get_segment(MemKind *mk, UWord *size_p, Uint flag
 	    ASSERT(best->seg);
 	    ASSERT(best->size > 0);
 
-	    mk->cache_hits++;
+	    ma->cache_hits++;
 
 	    /* Use current cache placement for remaining segment space */
 
@@ -608,7 +568,7 @@ static ERTS_INLINE void *cache_get_segment(MemKind *mk, UWord *size_p, Uint flag
  * using callbacks from aux-work in the scheduler.
  */
 
-static ERTS_INLINE Uint mseg_drop_one_memkind_cache_size(MemKind *mk, Uint flags, cache_t *head) {
+static ERTS_INLINE Uint mseg_drop_one_cache_size(ErtsMsegAllctr_t *ma, Uint flags, cache_t *head) {
     cache_t *c = NULL;
 
     c = erts_circleq_tail(head);
@@ -617,19 +577,19 @@ static ERTS_INLINE Uint mseg_drop_one_memkind_cache_size(MemKind *mk, Uint flags
     if (erts_mtrace_enabled)
 	erts_mtrace_crr_free(SEGTYPE, SEGTYPE, c->seg);
 
-    mseg_destroy(mk->ma, flags, mk, c->seg, c->size);
+    mseg_destroy(ma, flags, c->seg, c->size);
     mseg_cache_clear_node(c);
-    erts_circleq_push_head(&(mk->cache_free), c);
+    erts_circleq_push_head(&(ma->cache_free), c);
 
-    mk->segments.current.watermark--;
-    mk->cache_size--;
+    ma->segments.current.watermark--;
+    ma->cache_size--;
 
-    ASSERT( mk->cache_size >= 0 );
+    ASSERT(ma->cache_size >= 0);
 
-    return mk->cache_size;
+    return ma->cache_size;
 }
 
-static ERTS_INLINE Uint mseg_drop_memkind_cache_size(MemKind *mk, Uint flags, cache_t *head) {
+static ERTS_INLINE Uint mseg_drop_cache_size(ErtsMsegAllctr_t *ma, Uint flags, cache_t *head) {
     cache_t *c = NULL;
 
     while (!erts_circleq_is_empty(head)) {
@@ -640,58 +600,52 @@ static ERTS_INLINE Uint mseg_drop_memkind_cache_size(MemKind *mk, Uint flags, ca
 	if (erts_mtrace_enabled)
 	    erts_mtrace_crr_free(SEGTYPE, SEGTYPE, c->seg);
 
-	mseg_destroy(mk->ma, flags, mk, c->seg, c->size);
+	mseg_destroy(ma, flags, c->seg, c->size);
 
 	mseg_cache_clear_node(c);
-	erts_circleq_push_head(&(mk->cache_free), c);
-
-	mk->segments.current.watermark--;
-	mk->cache_size--;
+	erts_circleq_push_head(&(ma->cache_free), c);
 
+	ma->segments.current.watermark--;
+	ma->cache_size--;
     }
 
-    ASSERT( mk->cache_size >= 0 );
+    ASSERT(ma->cache_size >= 0);
 
-    return mk->cache_size;
+    return ma->cache_size;
 }
 
-/* mseg_check_memkind_cache
- * - Check if we can empty some cached segments in this
- *   MemKind.
+/* mseg_check_cache
+ * - Check if we can empty some cached segments in this allocator
  */
 
 
-static Uint mseg_check_memkind_cache(MemKind *mk) {
+static Uint mseg_check_cache(ErtsMsegAllctr_t *ma) {
     int i;
 
-    ERTS_DBG_MK_CHK_THR_ACCESS(mk);
+    ERTS_DBG_MA_CHK_THR_ACCESS(ma);
 
     for (i = 0; i < CACHE_AREAS; i++) {
-	if (!erts_circleq_is_empty(&(mk->cache_powered_node[i])))
-	    return mseg_drop_one_memkind_cache_size(mk, ERTS_MSEG_FLG_2POW, &(mk->cache_powered_node[i]));
+	if (!erts_circleq_is_empty(&(ma->cache_powered_node[i])))
+	    return mseg_drop_one_cache_size(ma, ERTS_MSEG_FLG_2POW, &(ma->cache_powered_node[i]));
     }
 
-    if (!erts_circleq_is_empty(&(mk->cache_unpowered_node)))
-	return mseg_drop_one_memkind_cache_size(mk, ERTS_MSEG_FLG_NONE, &(mk->cache_unpowered_node));
+    if (!erts_circleq_is_empty(&(ma->cache_unpowered_node)))
+	return mseg_drop_one_cache_size(ma, ERTS_MSEG_FLG_NONE, &(ma->cache_unpowered_node));
 
     return 0;
 }
 
 /* mseg_cache_check
  * - Check if we have some cache we can purge
- *   in any of the memkinds.
  */
 
 static void mseg_cache_check(ErtsMsegAllctr_t *ma) {
-    MemKind* mk;
     Uint empty_cache = 1;
 
     ERTS_MSEG_LOCK(ma);
 
-    for (mk = ma->mk_list; mk; mk = mk->next) {
-	if (mseg_check_memkind_cache(mk))
-	    empty_cache = 0;
-    }
+    if (mseg_check_cache(ma))
+        empty_cache = 0;
 
     /* If all MemKinds caches are empty,
      * remove aux-work callback
@@ -709,7 +663,7 @@ static void mseg_cache_check(ErtsMsegAllctr_t *ma) {
 /* erts_mseg_cache_check
  * - This is a callback that is scheduled as aux-work from
  *   schedulers and is called at some interval if we have a cache
- *   on this mseg-allocator and memkind.
+ *   on this mseg-allocator.
  * - Purpose: Empty cache slowly so we don't collect mapped areas
  *   and bloat memory.
  */
@@ -719,42 +673,32 @@ void erts_mseg_cache_check(void) {
 }
 
 
-/* *_mseg_clear_*_cache
+/* mseg_clear_cache
  * Remove cached segments from the allocator completely
  */
 
-static void mseg_clear_memkind_cache(MemKind *mk) {
+
+static void mseg_clear_cache(ErtsMsegAllctr_t *ma) {
     int i;
+    ERTS_MSEG_LOCK(ma);
+    ERTS_DBG_MA_CHK_THR_ACCESS(ma);
 
     /* drop pow2 caches */
     for (i = 0; i < CACHE_AREAS; i++) {
-	if (erts_circleq_is_empty(&(mk->cache_powered_node[i])))
+	if (erts_circleq_is_empty(&(ma->cache_powered_node[i])))
 	    continue;
 
-	mseg_drop_memkind_cache_size(mk, ERTS_MSEG_FLG_2POW, &(mk->cache_powered_node[i]));
-	ASSERT(erts_circleq_is_empty(&(mk->cache_powered_node[i])));
+	mseg_drop_cache_size(ma, ERTS_MSEG_FLG_2POW, &(ma->cache_powered_node[i]));
+	ASSERT(erts_circleq_is_empty(&(ma->cache_powered_node[i])));
     }
     /* drop varied caches */
-    if (!erts_circleq_is_empty(&(mk->cache_unpowered_node)))
-	mseg_drop_memkind_cache_size(mk, ERTS_MSEG_FLG_NONE, &(mk->cache_unpowered_node));
-
-    ASSERT(erts_circleq_is_empty(&(mk->cache_unpowered_node)));
-    ASSERT(mk->cache_size == 0);
-}
-
-static void mseg_clear_cache(ErtsMsegAllctr_t *ma) {
-    MemKind* mk;
-
-    ERTS_MSEG_LOCK(ma);
-    ERTS_DBG_MA_CHK_THR_ACCESS(ma);
+    if (!erts_circleq_is_empty(&(ma->cache_unpowered_node)))
+	mseg_drop_cache_size(ma, ERTS_MSEG_FLG_NONE, &(ma->cache_unpowered_node));
 
-
-    for (mk = ma->mk_list; mk; mk = mk->next) {
-	mseg_clear_memkind_cache(mk);
-    }
+    ASSERT(erts_circleq_is_empty(&(ma->cache_unpowered_node)));
+    ASSERT(ma->cache_size == 0);
 
     INC_CC(ma, clear_cache);
-
     ERTS_MSEG_UNLOCK(ma);
 }
 
@@ -763,25 +707,12 @@ void erts_mseg_clear_cache(void) {
     mseg_clear_cache(ERTS_MSEG_ALLCTR_IX(0));
 }
 
-
-
-static ERTS_INLINE MemKind* memkind(ErtsMsegAllctr_t *ma,
-				    const ErtsMsegOpt_t *opt)
-{
-#if HALFWORD_HEAP
-    return opt->low_mem ? &ma->low_mem : &ma->hi_mem;
-#else
-    return &ma->the_mem;
-#endif
-}
-
 static void *
 mseg_alloc(ErtsMsegAllctr_t *ma, ErtsAlcType_t atype, UWord *size_p,
 	   Uint flags, const ErtsMsegOpt_t *opt)
 {
     UWord size;
     void *seg;
-    MemKind* mk = memkind(ma, opt);
 
     INC_CC(ma, alloc);
 
@@ -795,10 +726,10 @@ mseg_alloc(ErtsMsegAllctr_t *ma, ErtsAlcType_t atype, UWord *size_p,
 	}
     }
 
-    if (opt->cache && mk->cache_size > 0 && (seg = cache_get_segment(mk, &size, flags)) != NULL)
+    if (opt->cache && ma->cache_size > 0 && (seg = cache_get_segment(ma, &size, flags)) != NULL)
 	goto done;
 
-    seg = mseg_create(ma, flags, mk, &size);
+    seg = mseg_create(ma, flags, &size);
 
     if (!seg)
 	*size_p = 0;
@@ -808,7 +739,7 @@ done:
 	if (erts_mtrace_enabled)
 	    erts_mtrace_crr_alloc(seg, atype, ERTS_MTRACE_SEGMENT_ID, size);
 
-	ERTS_MSEG_ALLOC_STAT(mk,size);
+	ERTS_MSEG_ALLOC_STAT(ma,size);
     }
 
     return seg;
@@ -819,11 +750,9 @@ static void
 mseg_dealloc(ErtsMsegAllctr_t *ma, ErtsAlcType_t atype, void *seg, UWord size,
 	     Uint flags, const ErtsMsegOpt_t *opt)
 {
-    MemKind* mk = memkind(ma, opt);
+    ERTS_MSEG_DEALLOC_STAT(ma,size);
 
-    ERTS_MSEG_DEALLOC_STAT(mk,size);
-
-    if (opt->cache && cache_bless_segment(mk, seg, size, flags)) {
+    if (opt->cache && cache_bless_segment(ma, seg, size, flags)) {
 	schedule_cache_check(ma);
 	goto done;
     }
@@ -831,7 +760,7 @@ mseg_dealloc(ErtsMsegAllctr_t *ma, ErtsAlcType_t atype, void *seg, UWord size,
     if (erts_mtrace_enabled)
 	erts_mtrace_crr_free(atype, SEGTYPE, seg);
 
-    mseg_destroy(ma, flags, mk, seg, size);
+    mseg_destroy(ma, flags, seg, size);
 
 done:
 
@@ -842,7 +771,6 @@ static void *
 mseg_realloc(ErtsMsegAllctr_t *ma, ErtsAlcType_t atype, void *seg,
 	     UWord old_size, UWord *new_size_p, Uint flags, const ErtsMsegOpt_t *opt)
 {
-    MemKind* mk;
     void *new_seg;
     UWord new_size;
 
@@ -861,7 +789,6 @@ mseg_realloc(ErtsMsegAllctr_t *ma, ErtsAlcType_t atype, void *seg,
 	return NULL;
     }
 
-    mk = memkind(ma, opt);
     new_seg = seg;
 
     if (!MSEG_FLG_IS_2POW(flags))
@@ -876,7 +803,7 @@ mseg_realloc(ErtsMsegAllctr_t *ma, ErtsAlcType_t atype, void *seg,
 
     if (new_size > old_size) {
 	if (opt->preserv) {
-	    new_seg = mseg_recreate(ma, flags, mk, (void *) seg, old_size, &new_size);
+	    new_seg = mseg_recreate(ma, flags, (void *) seg, old_size, &new_size);
 	    if (!new_seg)
 		new_size = old_size;
 	}
@@ -896,7 +823,7 @@ mseg_realloc(ErtsMsegAllctr_t *ma, ErtsAlcType_t atype, void *seg,
 	    new_size = old_size;
 	}
 	else {
-	    new_seg = mseg_recreate(ma, flags, mk, (void *) seg, old_size, &new_size);
+	    new_seg = mseg_recreate(ma, flags, (void *) seg, old_size, &new_size);
 	    if (!new_seg)
 		new_size = old_size;
 	}
@@ -910,7 +837,7 @@ mseg_realloc(ErtsMsegAllctr_t *ma, ErtsAlcType_t atype, void *seg,
     ASSERT(!MSEG_FLG_IS_2POW(flags) || IS_2POW(new_size));
     *new_size_p = new_size;
 
-    ERTS_MSEG_REALLOC_STAT(mk, old_size, new_size);
+    ERTS_MSEG_REALLOC_STAT(ma, old_size, new_size);
 
     return new_seg;
 }
@@ -1070,7 +997,8 @@ info_options(ErtsMsegAllctr_t *ma,
 {
     Eterm res;
 
-    res = erts_mmap_info_options(prefix, print_to_p, print_to_arg, hpp, szp);
+    res = erts_mmap_info_options(&erts_dflt_mmapper,
+                                 prefix, print_to_p, print_to_arg, hpp, szp);
 
     if (print_to_p) {
 	int to = *print_to_p;
@@ -1180,63 +1108,63 @@ info_calls(ErtsMsegAllctr_t *ma, int *print_to_p, void *print_to_arg, Uint **hpp
 }
 
 static Eterm
-info_status(ErtsMsegAllctr_t *ma, MemKind* mk, int *print_to_p, void *print_to_arg,
+info_status(ErtsMsegAllctr_t *ma, int *print_to_p, void *print_to_arg,
 	    int begin_new_max_period, Uint **hpp, Uint *szp)
 {
     Eterm res = THE_NON_VALUE;
     
-    if (mk->segments.max_ever.no < mk->segments.max.no)
-	mk->segments.max_ever.no = mk->segments.max.no;
-    if (mk->segments.max_ever.sz < mk->segments.max.sz)
-	mk->segments.max_ever.sz = mk->segments.max.sz;
+    if (ma->segments.max_ever.no < ma->segments.max.no)
+	ma->segments.max_ever.no = ma->segments.max.no;
+    if (ma->segments.max_ever.sz < ma->segments.max.sz)
+	ma->segments.max_ever.sz = ma->segments.max.sz;
 
     if (print_to_p) {
 	int to = *print_to_p;
 	void *arg = print_to_arg;
 
-	erts_print(to, arg, "cached_segments: %beu\n", mk->cache_size);
-	erts_print(to, arg, "cache_hits: %beu\n", mk->cache_hits);
+	erts_print(to, arg, "cached_segments: %beu\n", ma->cache_size);
+	erts_print(to, arg, "cache_hits: %beu\n", ma->cache_hits);
 	erts_print(to, arg, "segments: %beu %beu %beu\n",
-		   mk->segments.current.no, mk->segments.max.no, mk->segments.max_ever.no);
+		   ma->segments.current.no, ma->segments.max.no, ma->segments.max_ever.no);
 	erts_print(to, arg, "segments_size: %beu %beu %beu\n",
-		   mk->segments.current.sz, mk->segments.max.sz, mk->segments.max_ever.sz);
+		   ma->segments.current.sz, ma->segments.max.sz, ma->segments.max_ever.sz);
 	erts_print(to, arg, "segments_watermark: %beu\n",
-		   mk->segments.current.watermark);
+		   ma->segments.current.watermark);
     }
 
     if (hpp || szp) {
 	res = NIL;
 	add_2tup(hpp, szp, &res,
 		 am.segments_watermark,
-		 bld_unstable_uint(hpp, szp, mk->segments.current.watermark));
+		 bld_unstable_uint(hpp, szp, ma->segments.current.watermark));
 	add_4tup(hpp, szp, &res,
 		 am.segments_size,
-		 bld_unstable_uint(hpp, szp, mk->segments.current.sz),
-		 bld_unstable_uint(hpp, szp, mk->segments.max.sz),
-		 bld_unstable_uint(hpp, szp, mk->segments.max_ever.sz));
+		 bld_unstable_uint(hpp, szp, ma->segments.current.sz),
+		 bld_unstable_uint(hpp, szp, ma->segments.max.sz),
+		 bld_unstable_uint(hpp, szp, ma->segments.max_ever.sz));
 	add_4tup(hpp, szp, &res,
 		 am.segments,
-		 bld_unstable_uint(hpp, szp, mk->segments.current.no),
-		 bld_unstable_uint(hpp, szp, mk->segments.max.no),
-		 bld_unstable_uint(hpp, szp, mk->segments.max_ever.no));
+		 bld_unstable_uint(hpp, szp, ma->segments.current.no),
+		 bld_unstable_uint(hpp, szp, ma->segments.max.no),
+		 bld_unstable_uint(hpp, szp, ma->segments.max_ever.no));
 	add_2tup(hpp, szp, &res,
 		 am.cache_hits,
-		 bld_unstable_uint(hpp, szp, mk->cache_hits));
+		 bld_unstable_uint(hpp, szp, ma->cache_hits));
 	add_2tup(hpp, szp, &res,
 		 am.cached_segments,
-		 bld_unstable_uint(hpp, szp, mk->cache_size));
+		 bld_unstable_uint(hpp, szp, ma->cache_size));
 
     }
 
     if (begin_new_max_period) {
-	mk->segments.max.no = mk->segments.current.no;
-	mk->segments.max.sz = mk->segments.current.sz;
+	ma->segments.max.no = ma->segments.current.no;
+	ma->segments.max.sz = ma->segments.current.sz;
     }
 
     return res;
 }
 
-static Eterm info_memkind(ErtsMsegAllctr_t *ma, MemKind* mk, int *print_to_p, void *print_to_arg,
+static Eterm info_memkind(ErtsMsegAllctr_t *ma, int *print_to_p, void *print_to_arg,
 			  int begin_max_per, Uint **hpp, Uint *szp)
 {
     Eterm res = THE_NON_VALUE;
@@ -1244,15 +1172,15 @@ static Eterm info_memkind(ErtsMsegAllctr_t *ma, MemKind* mk, int *print_to_p, vo
     Eterm values[3];
 
     if (print_to_p) {
-	erts_print(*print_to_p, print_to_arg, "memory kind: %s\n", mk->name);
+	erts_print(*print_to_p, print_to_arg, "memory kind: %s\n", "all memory");
     }
     if (hpp || szp) {
 	atoms[0] = am.name;
 	atoms[1] = am.status;
 	atoms[2] = am.calls;
-	values[0] = erts_bld_string(hpp, szp, mk->name);
+	values[0] = erts_bld_string(hpp, szp, "all memory");
     }
-    values[1] = info_status(ma, mk, print_to_p, print_to_arg, begin_max_per, hpp, szp);
+    values[1] = info_status(ma, print_to_p, print_to_arg, begin_max_per, hpp, szp);
     values[2] = info_calls(ma, print_to_p, print_to_arg, hpp, szp);
 
     if (hpp || szp)
@@ -1261,7 +1189,6 @@ static Eterm info_memkind(ErtsMsegAllctr_t *ma, MemKind* mk, int *print_to_p, vo
     return res;
 }
 
-
 static Eterm
 info_version(ErtsMsegAllctr_t *ma, int *print_to_p, void *print_to_arg, Uint **hpp, Uint *szp)
 {
@@ -1326,12 +1253,7 @@ erts_mseg_info(int ix,
     ERTS_MSEG_LOCK(ma);
     ERTS_DBG_MA_CHK_THR_ACCESS(ma);
 
-#if HALFWORD_HEAP
-    values[n++] = info_memkind(ma, &ma->low_mem, print_to_p, print_to_arg, begin_max_per, hpp, szp);
-    values[n++] = info_memkind(ma, &ma->hi_mem, print_to_p, print_to_arg, begin_max_per, hpp, szp);
-#else
-    values[n++] = info_memkind(ma, &ma->the_mem, print_to_p, print_to_arg, begin_max_per, hpp, szp);
-#endif
+    values[n++] = info_memkind(ma, print_to_p, print_to_arg, begin_max_per, hpp, szp);
     if (hpp || szp)
 	res = bld_2tup_list(hpp, szp, n, atoms, values);
 
@@ -1408,13 +1330,10 @@ Uint
 erts_mseg_no(const ErtsMsegOpt_t *opt)
 {
     ErtsMsegAllctr_t *ma = ERTS_MSEG_ALLCTR_OPT(opt);
-    MemKind* mk;
-    Uint n = 0;
+    Uint n;
     ERTS_MSEG_LOCK(ma);
     ERTS_DBG_MA_CHK_THR_ACCESS(ma);
-    for (mk=ma->mk_list; mk; mk=mk->next) {
-	n += mk->segments.current.no;
-    }
+    n = ma->segments.current.no;
     ERTS_MSEG_UNLOCK(ma);
     return n;
 }
@@ -1426,16 +1345,16 @@ erts_mseg_unit_size(void)
 }
 
 
-static void mem_kind_init(ErtsMsegAllctr_t *ma, MemKind* mk, const char* name)
+static void mem_cache_init(ErtsMsegAllctr_t *ma)
 {
     int i;
 
     /* Clear all cache headers */
-    mseg_cache_clear_node(&(mk->cache_free));
-    mseg_cache_clear_node(&(mk->cache_unpowered_node));
+    mseg_cache_clear_node(&(ma->cache_free));
+    mseg_cache_clear_node(&(ma->cache_unpowered_node));
 
     for (i = 0; i < CACHE_AREAS; i++) {
-	mseg_cache_clear_node(&(mk->cache_powered_node[i]));
+	mseg_cache_clear_node(&(ma->cache_powered_node[i]));
     }
 
     /* Populate cache free list */
@@ -1443,25 +1362,20 @@ static void mem_kind_init(ErtsMsegAllctr_t *ma, MemKind* mk, const char* name)
     ASSERT(ma->max_cache_size <= MAX_CACHE_SIZE);
 
     for (i = 0; i < ma->max_cache_size; i++) {
-	mseg_cache_clear_node(&(mk->cache[i]));
-	erts_circleq_push_head(&(mk->cache_free), &(mk->cache[i]));
+	mseg_cache_clear_node(&(ma->cache[i]));
+	erts_circleq_push_head(&(ma->cache_free), &(ma->cache[i]));
     }
 
-    mk->cache_size = 0;
-    mk->cache_hits = 0;
-
-    mk->segments.current.watermark = 0;
-    mk->segments.current.no = 0;
-    mk->segments.current.sz = 0;
-    mk->segments.max.no = 0;
-    mk->segments.max.sz = 0;
-    mk->segments.max_ever.no = 0;
-    mk->segments.max_ever.sz = 0;
-
-    mk->ma = ma;
-    mk->name = name;
-    mk->next = ma->mk_list;
-    ma->mk_list = mk;
+    ma->cache_size = 0;
+    ma->cache_hits = 0;
+
+    ma->segments.current.watermark = 0;
+    ma->segments.current.no = 0;
+    ma->segments.current.sz = 0;
+    ma->segments.max.no = 0;
+    ma->segments.max.sz = 0;
+    ma->segments.max_ever.no = 0;
+    ma->segments.max_ever.sz = 0;
 }
 
 void
@@ -1488,16 +1402,7 @@ erts_mseg_init(ErtsMsegInit_t *init)
 
     erts_mtx_init(&init_atoms_mutex, "mseg_init_atoms");
 
-#if HALFWORD_HEAP
-    if (sizeof(void *) != 8)
-	erl_exit(-1,"Halfword emulator cannot be run in 32bit mode");
-
-    init->mmap.virtual_range.start = (char *) sbrk(0);
-    init->mmap.virtual_range.end = (char *) 0x100000000UL;
-    init->mmap.sco = 0;
-#endif
-
-    erts_mmap_init(&init->mmap);
+    erts_mmap_init(&erts_dflt_mmapper, &init->dflt_mmap);
 
     if (!IS_2POW(GET_PAGE_SIZE))
 	erl_exit(ERTS_ABORT_EXIT, "erts_mseg: Unexpected page_size %beu\n", GET_PAGE_SIZE);
@@ -1529,14 +1434,7 @@ erts_mseg_init(ErtsMsegInit_t *init)
 	if (ma->max_cache_size > MAX_CACHE_SIZE)
 	    ma->max_cache_size = MAX_CACHE_SIZE;
 
-	ma->mk_list = NULL;
-
-#if HALFWORD_HEAP
-	mem_kind_init(ma, &ma->low_mem, "low memory");
-	mem_kind_init(ma, &ma->hi_mem, "high memory");
-#else
-	mem_kind_init(ma, &ma->the_mem, "all memory");
-#endif
+	mem_cache_init(ma);
 
 	sys_memzero((void *) &ma->calls, sizeof(ErtsMsegCalls));
     }
@@ -1545,13 +1443,8 @@ erts_mseg_init(ErtsMsegInit_t *init)
 
 static ERTS_INLINE Uint tot_cache_size(ErtsMsegAllctr_t *ma)
 {
-    MemKind* mk;
-    Uint sz = 0;
     ERTS_DBG_MA_CHK_THR_ACCESS(ma);
-    for (mk=ma->mk_list; mk; mk=mk->next) {
-	sz += mk->cache_size;
-    }
-    return sz;
+    return ma->cache_size;
 }
 
 /*
diff --git a/erts/emulator/sys/common/erl_mseg.h b/erts/emulator/sys/common/erl_mseg.h
index ba04e919fc..2acd8f8505 100644
--- a/erts/emulator/sys/common/erl_mseg.h
+++ b/erts/emulator/sys/common/erl_mseg.h
@@ -42,16 +42,6 @@
 
 #if ERTS_HAVE_MSEG_SUPER_ALIGNED
 #  define MSEG_ALIGN_BITS ERTS_MMAP_SUPERALIGNED_BITS
-#else
-/* If we don't use super aligned multiblock carriers
- * we will mmap with page size alignment (and thus use corresponding
- * align bits).
- *
- * Current implementation needs this to be a constant and
- * only uses this for user dev testing so setting page size
- * to 4096 (12 bits) is fine.
- */
-#  define MSEG_ALIGN_BITS       (12)
 #endif
 
 #if HAVE_ERTS_MSEG
@@ -69,7 +59,8 @@ typedef struct {
     Uint rmcbf;
     Uint mcs;
     Uint nos;
-    ErtsMMapInit mmap;
+    ErtsMMapInit dflt_mmap;
+    ErtsMMapInit literal_mmap;
 } ErtsMsegInit_t;
 
 #define ERTS_MSEG_INIT_DEFAULT_INITIALIZER				\
@@ -78,7 +69,8 @@ typedef struct {
     20,			/* rmcbf: Relative max cache bad fit	*/	\
     10,			/* mcs:   Max cache size		*/	\
     1000,		/* cci:   Cache check interval		*/	\
-    ERTS_MMAP_INIT_DEFAULT_INITER					\
+    ERTS_MMAP_INIT_DEFAULT_INITER,					\
+    ERTS_MMAP_INIT_LITERAL_INITER                                       \
 }
 
 typedef struct {
@@ -87,9 +79,6 @@ typedef struct {
     UWord abs_shrink_th;
     UWord rel_shrink_th;
     int sched_spec;
-#if HALFWORD_HEAP
-    int low_mem;
-#endif
 } ErtsMsegOpt_t;
 
 extern const ErtsMsegOpt_t erts_mseg_default_opt;
diff --git a/erts/emulator/sys/common/erl_poll.c b/erts/emulator/sys/common/erl_poll.c
index 884c403da4..21fa214364 100644
--- a/erts/emulator/sys/common/erl_poll.c
+++ b/erts/emulator/sys/common/erl_poll.c
@@ -431,7 +431,7 @@ static ERTS_INLINE int
 is_interrupted_reset(ErtsPollSet ps)
 {
 #if defined(USE_THREADS) || ERTS_POLL_ASYNC_INTERRUPT_SUPPORT
-    return (erts_atomic32_xchg_nob(&ps->wakeup_state, ERTS_POLL_NOT_WOKEN)
+    return (erts_atomic32_xchg_acqb(&ps->wakeup_state, ERTS_POLL_NOT_WOKEN)
 	    == ERTS_POLL_WOKEN_INTR);
 #else
     return 0;
@@ -442,7 +442,7 @@ static ERTS_INLINE void
 woke_up(ErtsPollSet ps)
 {
 #if defined(USE_THREADS) || ERTS_POLL_ASYNC_INTERRUPT_SUPPORT
-    erts_aint32_t wakeup_state = erts_atomic32_read_nob(&ps->wakeup_state);
+    erts_aint32_t wakeup_state = erts_atomic32_read_acqb(&ps->wakeup_state);
     if (wakeup_state == ERTS_POLL_NOT_WOKEN)
 	(void) erts_atomic32_cmpxchg_nob(&ps->wakeup_state,
 					 ERTS_POLL_WOKEN,
@@ -469,14 +469,9 @@ wake_poller(ErtsPollSet ps, int interrupted, int async_signal_safe)
 	    wakeup_state = erts_atomic32_cmpxchg_relb(&ps->wakeup_state,
 						      ERTS_POLL_WOKEN,
 						      ERTS_POLL_NOT_WOKEN);
-	else {
-	    /*
-	     * We might unnecessarily write to the pipe, however,
-	     * that isn't problematic.
-	     */
-	    wakeup_state = erts_atomic32_read_nob(&ps->wakeup_state);
-	    erts_atomic32_set_relb(&ps->wakeup_state, ERTS_POLL_WOKEN_INTR);
-	}
+	else
+            wakeup_state = erts_atomic32_xchg_relb(&ps->wakeup_state,
+                                                   ERTS_POLL_WOKEN_INTR);
 	wake = wakeup_state == ERTS_POLL_NOT_WOKEN;
     }
     /*
diff --git a/erts/emulator/sys/common/erl_poll.h b/erts/emulator/sys/common/erl_poll.h
index 19ce582154..bc2c681876 100644
--- a/erts/emulator/sys/common/erl_poll.h
+++ b/erts/emulator/sys/common/erl_poll.h
@@ -93,7 +93,7 @@
 #  if defined(ERTS_USE_POLL)
 #    undef ERTS_POLL_USE_POLL
 #    define ERTS_POLL_USE_POLL 1
-#  elif !defined(__WIN32__) && !defined(__OSE__)
+#  elif !defined(__WIN32__)
 #    undef ERTS_POLL_USE_SELECT
 #    define ERTS_POLL_USE_SELECT 1
 #  endif
@@ -104,31 +104,13 @@
 typedef Uint32 ErtsPollEvents;
 #undef ERTS_POLL_EV_E2N
 
-#if defined(__WIN32__) || defined(__OSE__)	/* --- win32 or ose -------- */
+#if defined(__WIN32__)	/* --- win32  --------------------------------------- */
 
 #define ERTS_POLL_EV_IN   1
 #define ERTS_POLL_EV_OUT  2
 #define ERTS_POLL_EV_ERR  4
 #define ERTS_POLL_EV_NVAL 8
 
-#ifdef __OSE__
-
-typedef struct ErtsPollOseMsgList_ {
-  struct ErtsPollOseMsgList_ *next;
-  union SIGNAL *data;
-} ErtsPollOseMsgList;
-
-struct erts_sys_fd_type {
-    SIGSELECT signo;
-    ErlDrvOseEventId id;
-    ErtsPollOseMsgList *msgs;
-    ErlDrvOseEventId (*resolve_signal)(union SIGNAL *sig);
-    ethr_mutex mtx;
-    void *extra;
-};
-
-#endif
-
 #elif ERTS_POLL_USE_EPOLL	/* --- epoll ------------------------------- */
 
 #include <sys/epoll.h>
@@ -140,7 +122,7 @@ struct erts_sys_fd_type {
 #endif
 
 #define ERTS_POLL_EV_E2N(EV) \
-  ((__uint32_t) (EV))
+  ((uint32_t) (EV))
 #define ERTS_POLL_EV_N2E(EV) \
   ((ErtsPollEvents) (EV))
 
diff --git a/erts/emulator/sys/ose/beam.lmconf b/erts/emulator/sys/ose/beam.lmconf
deleted file mode 100644
index 4ad46b01d9..0000000000
--- a/erts/emulator/sys/ose/beam.lmconf
+++ /dev/null
@@ -1,26 +0,0 @@
-OSE_LM_STACK_SIZES=256,512,1024,2048,4096,8192,16384,65536
-OSE_LM_SIGNAL_SIZES=31,63,127,255,1023,4095,16383,65535
-OSE_LM_POOL_SIZE=0x200000
-OSE_LM_MAIN_NAME=main
-OSE_LM_MAIN_STACK_SIZE=0xF000
-OSE_LM_MAIN_PRIORITY=20
-## Has to be of a type that allows MAM
-OSE_LM_PROGRAM_TYPE=APP_RAM
-OSE_LM_DATA_INIT=YES
-OSE_LM_BSS_INIT=YES
-OSE_LM_EXEC_MODEL=SHARED
-HEAP_MAX_SIZE=1000000000
-HEAP_SMALL_BUF_INIT_SIZE=20971520
-HEAP_LARGE_BUF_THRESHOLD=16000000
-HEAP_LOCK_TYPE=2
-
-ERTS_DEFAULT_PRIO=24
-ERTS_SCHEDULER_PRIO=24
-ERTS_ASYNC_PRIO=22
-ERTS_AUX_PRIO=24
-ERTS_SYS_MSG_DISPATCHER_PRIO=21
-
-# Setting the environment variable EFS_RESOLVE_TMO on the block to 0.
-# This will eliminiate delays when trying to open files on not mounted
-# volumes.
-EFS_RESOLVE_TMO=0
diff --git a/erts/emulator/sys/ose/driver_int.h b/erts/emulator/sys/ose/driver_int.h
deleted file mode 100644
index 4a5b7171d1..0000000000
--- a/erts/emulator/sys/ose/driver_int.h
+++ /dev/null
@@ -1,42 +0,0 @@
-/*
- * %CopyrightBegin%
- *
- * Copyright Ericsson AB 1997-2009. All Rights Reserved.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- *
- * %CopyrightEnd%
- */
-/*
- * System dependant driver declarations
- */
-
-#ifndef __DRIVER_INT_H__
-#define __DRIVER_INT_H__
-
-#ifdef HAVE_SYS_UIO_H
-#include <sys/types.h>
-#include <sys/uio.h>
-
-typedef struct iovec SysIOVec;
-
-#else
-
-typedef struct {
-    char* iov_base;
-    int   iov_len;
-} SysIOVec;
-
-#endif
-
-#endif
diff --git a/erts/emulator/sys/ose/erl_main.c b/erts/emulator/sys/ose/erl_main.c
deleted file mode 100644
index 877e85f43a..0000000000
--- a/erts/emulator/sys/ose/erl_main.c
+++ /dev/null
@@ -1,54 +0,0 @@
-/*
- * %CopyrightBegin%
- *
- * Copyright Ericsson AB 2000-2009. All Rights Reserved.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- *
- * %CopyrightEnd%
- */
-
-#ifdef HAVE_CONFIG_H
-#  include "config.h"
-#endif
-#include <stdlib.h>
-
-#include "sys.h"
-#include "erl_vm.h"
-#include "global.h"
-#include "ose.h"
-
-int
-main(int argc, char **argv) {
-
-  (void)stdin;(void)stdout;(void)stderr;
-
-  /* When starting using pm_create -c ARGV="-- -root ..", argv[0] is the first
-     part of ARGV and not the name of the executable. So we shuffle some
-     pointers here to make erl_start happy. */
-  if (argv[0][0] == '-') {
-    int i;
-    char **tmp_argv = malloc(sizeof(char*)*(argc+1));
-    for (i = 0; i < argc; i++)
-      tmp_argv[i+1] = argv[i];
-    tmp_argv[0] = "beam";
-    erl_start(argc+1,tmp_argv);
-    free(tmp_argv);
-  } else {
-   erl_start(argc,argv);
-  }
-
-   stop(current_process());
-
-   return 0;
-}
diff --git a/erts/emulator/sys/ose/erl_ose_sys.h b/erts/emulator/sys/ose/erl_ose_sys.h
deleted file mode 100644
index d0cd3180bf..0000000000
--- a/erts/emulator/sys/ose/erl_ose_sys.h
+++ /dev/null
@@ -1,356 +0,0 @@
-/*
- * %CopyrightBegin%
- *
- * Copyright Ericsson AB 1997-2011. All Rights Reserved.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- *
- * %CopyrightEnd%
- *
- * This file handles differences between different Unix systems.
- * This should be the only place with conditional compilation
- * depending on the type of OS.
- */
-
-#ifndef _ERL_OSE_SYS_H
-#define _ERL_OSE_SYS_H
-
-#include "ose.h"
-#undef NIL
-#include "ramlog.h"
-#include "erts.sig"
-
-#include "fcntl.h"
-#include "math.h"
-#include "stdio.h"
-#include "stdlib.h"
-#include "string.h"
-#include "sys/param.h"
-#include "sys/time.h"
-#include "time.h"
-#include "dirent.h"
-#include "ethread.h"
-
-/* FIXME: configuration options */
-#define ERTS_SCHED_MIN_SPIN 1
-#define ERTS_SCHED_ONLY_POLL_SCHED_1 1
-#define ERTS_SCHED_FAIR 1
-#define NO_SYSCONF 1
-#define OPEN_MAX FOPEN_MAX
-
-#define MAP_ANON MAP_ANONYMOUS
-
-#ifndef HAVE_MMAP
-#   define HAVE_MMAP 0
-#endif
-
-#if HAVE_MMAP
-#   include "sys/mman.h"
-#endif
-
-/*
- * Min number of async threads
- */
-#define ERTS_MIN_NO_OF_ASYNC_THREADS 1
-
-/*
- * Our own type of "FD's"
- */
-#define ERTS_SYS_FD_TYPE struct erts_sys_fd_type*
-#define NO_FSTAT_ON_SYS_FD_TYPE 1 /* They are signals, not files */
-
-#include "sys/stat.h"
-
-/* FIXME mremap is not defined in OSE - POSIX issue */
-extern void *mremap (void *__addr, size_t __old_len, size_t __new_len,
-                        int __flags, ...);
-
-/* FIXME: mremap constants */
-#define MREMAP_MAYMOVE  1
-#define MREMAP_FIXED    2
-
-typedef void *GETENV_STATE;
-
-/*
-** For the erl_timer_sup module.
-*/
-#define HAVE_GETHRTIME
-
-typedef long long SysHrTime;
-extern SysHrTime sys_gethrtime(void);
-
-void sys_init_hrtime(void);
-
-typedef time_t erts_time_t;
-
-typedef struct timeval SysTimeval;
-
-#define sys_gettimeofday(Arg) ((void) gettimeofday((Arg), NULL))
-
-typedef struct {
-    clock_t tms_utime;
-    clock_t tms_stime;
-    clock_t tms_cutime;
-    clock_t tms_cstime;
-} SysTimes;
-
-extern int erts_ticks_per_sec;
-
-#define SYS_CLK_TCK (erts_ticks_per_sec)
-
-extern clock_t sys_times(SysTimes *buffer);
-
-/* No use in having other resolutions than 1 Ms. */
-#define SYS_CLOCK_RESOLUTION 1
-
-#define erts_isfinite finite
-
-#ifdef NO_FPE_SIGNALS
-
-#define erts_get_current_fp_exception() NULL
-#ifdef ERTS_SMP
-#define erts_thread_init_fp_exception() do{}while(0)
-#endif
-#  define __ERTS_FP_CHECK_INIT(fpexnp) do {} while (0)
-#  define __ERTS_FP_ERROR(fpexnp, f, Action) if (!finite(f)) { Action; } else {}
-#  define __ERTS_FP_ERROR_THOROUGH(fpexnp, f, Action) __ERTS_FP_ERROR(fpexnp, f, Action)
-#  define __ERTS_SAVE_FP_EXCEPTION(fpexnp)
-#  define __ERTS_RESTORE_FP_EXCEPTION(fpexnp)
-
-#define erts_sys_block_fpe() 0
-#define erts_sys_unblock_fpe(x) do{}while(0)
-
-#else /* !NO_FPE_SIGNALS */
-
-extern volatile unsigned long *erts_get_current_fp_exception(void);
-#ifdef ERTS_SMP
-extern void erts_thread_init_fp_exception(void);
-#endif
-#  if (defined(__i386__) || defined(__x86_64__)) && defined(__GNUC__)
-#    define erts_fwait(fpexnp,f) \
-	__asm__ __volatile__("fwait" : "=m"(*(fpexnp)) : "m"(f))
-#  elif (defined(__powerpc__) || defined(__ppc__)) && defined(__GNUC__)
-#    define erts_fwait(fpexnp,f) \
-	__asm__ __volatile__("" : "=m"(*(fpexnp)) : "fm"(f))
-#  elif defined(__sparc__) && defined(__linux__) && defined(__GNUC__)
-#    define erts_fwait(fpexnp,f) \
-	__asm__ __volatile__("" : "=m"(*(fpexnp)) : "em"(f))
-#  else
-#    define erts_fwait(fpexnp,f) \
-	__asm__ __volatile__("" : "=m"(*(fpexnp)) : "g"(f))
-#  endif
-#  if (defined(__i386__) || defined(__x86_64__)) && defined(__GNUC__)
-     extern void erts_restore_fpu(void);
-#  else
-#    define erts_restore_fpu() /*empty*/
-#  endif
-#  if (!defined(__GNUC__) || \
-       (__GNUC__ < 2) || \
-       (__GNUC__ == 2 && __GNUC_MINOR < 96)) && \
-      !defined(__builtin_expect)
-#    define __builtin_expect(x, expected_value) (x)
-#  endif
-static __inline__ int erts_check_fpe(volatile unsigned long *fp_exception, double f)
-{
-    erts_fwait(fp_exception, f);
-    if (__builtin_expect(*fp_exception == 0, 1))
-       return 0;
-    *fp_exception = 0;
-    erts_restore_fpu();
-    return 1;
-}
-#  undef erts_fwait
-#  undef erts_restore_fpu
-extern void erts_fp_check_init_error(volatile unsigned long *fp_exception);
-static __inline__ void __ERTS_FP_CHECK_INIT(volatile unsigned long *fp_exception)
-{
-    if (__builtin_expect(*fp_exception == 0, 1))
-	return;
-    erts_fp_check_init_error(fp_exception);
-}
-#  define __ERTS_FP_ERROR(fpexnp, f, Action) do { if (erts_check_fpe((fpexnp),(f))) { Action; } } while (0)
-#  define __ERTS_SAVE_FP_EXCEPTION(fpexnp) unsigned long old_erl_fp_exception = *(fpexnp)
-#  define __ERTS_RESTORE_FP_EXCEPTION(fpexnp) \
-              do { *(fpexnp) = old_erl_fp_exception; } while (0)
-   /* This is for library calls where we don't trust the external
-      code to always throw floating-point exceptions on errors. */
-static __inline__ int erts_check_fpe_thorough(volatile unsigned long *fp_exception, double f)
-{
-    return erts_check_fpe(fp_exception, f) || !finite(f);
-}
-#  define __ERTS_FP_ERROR_THOROUGH(fpexnp, f, Action) \
-  do { if (erts_check_fpe_thorough((fpexnp),(f))) { Action; } } while (0)
-
-int erts_sys_block_fpe(void);
-void erts_sys_unblock_fpe(int);
-
-#endif /* !NO_FPE_SIGNALS */
-
-#define ERTS_FP_CHECK_INIT(p)		__ERTS_FP_CHECK_INIT(&(p)->fp_exception)
-#define ERTS_FP_ERROR(p, f, A)		__ERTS_FP_ERROR(&(p)->fp_exception, f, A)
-#define ERTS_FP_ERROR_THOROUGH(p, f, A)	__ERTS_FP_ERROR_THOROUGH(&(p)->fp_exception, f, A)
-
-/* FIXME: force HAVE_GETPAGESIZE and stub getpagesize */
-#ifndef HAVE_GETPAGESIZE
-#define HAVE_GETPAGESIZE 1
-#endif
-
-extern int getpagesize(void);
-
-#ifndef HZ
-#define HZ 60
-#endif
-
-/* OSE5 doesn't provide limits.h so a number of macros should be
- * added manually */
-
-#ifndef CHAR_BIT
-#define CHAR_BIT 8
-#endif
-
-/* Minimum and maximum values a `signed int' can hold.  */
-#ifndef INT_MAX
-#define INT_MAX	2147483647
-#endif
-
-#ifndef INT_MIN
-#define INT_MIN	(-INT_MAX - 1)
-#endif
-
-#ifndef UINT_MAX
-#  define UINT_MAX   4294967295U
-#endif
-
-/*
-static void erts_ose_sys_send(union SIGNAL **signal,PROCESS dst,
-			      char* file,int line) {
-  SIGSELECT **ziggy = (SIGSELECT**)signal;
-  printf("%s:%d 0x%x Send signal 0x%x(0x%x) to 0x%x\r\n",
-	 file,line,current_process(),ziggy[0][0],*ziggy,dst);
-  send(signal,dst);
-}
-#define send(signal,dst) erts_ose_sys_send(signal,dst,__FILE__,__LINE__)
-
-static void erts_ose_sys_send_w_sender(union SIGNAL **signal,
-				       PROCESS sender,PROCESS dst,
-				       char* file,int line) {
-  SIGSELECT **ziggy = (SIGSELECT**)signal;
-  printf("%s:%d 0x%x Send signal 0x%x(0x%x) to 0x%x as 0x%x\r\n",
-	 file,line,current_process(),ziggy[0][0],*ziggy,dst,sender);
-  send_w_sender(signal,sender,dst);
-}
-#define send_w_sender(signal,sender,dst) \
-  erts_ose_sys_send_w_sender(signal,sender,dst,__FILE__,__LINE__)
-
-
-static union SIGNAL *erts_ose_sys_receive(SIGSELECT *sigsel,
-					  char *file,
-					  int line) {
-  SIGSELECT *sig;
-  int i;
-
-  printf("%s:%d 0x%x receive({%d,",file,line,current_process(),sigsel[0]);
-  for (i = 1; i < sigsel[0]; i++)
-    printf("0x%x, ",sigsel[i]);
-  if (sigsel[0] != 0)
-    printf("0x%x",sigsel[i]);
-  printf("})\n");
-  sig = (SIGSELECT*)receive(sigsel);
-  printf("%s:%d 0x%x got 0x%x from 0x%x\n",file,line,current_process(),
-	 *sig,sender((union SIGNAL**)(&sig)));
-  return (union SIGNAL*)sig;
-}
-#define receive(SIGSEL) erts_ose_sys_receive(SIGSEL,__FILE__,__LINE__)
-
-static union SIGNAL *erts_ose_sys_receive_w_tmo(OSTIME tmo,SIGSELECT *sigsel,
-						char *file,int line) {
-  SIGSELECT *sig;
-  int i;
-  if (tmo == 0) {
-    sig = (SIGSELECT*)receive_w_tmo(tmo,sigsel);
-    if (sig != NULL) {
-      printf("%s:%d 0x%x receive_w_tmo(0,{%d,",file,line,current_process(),
-	     sigsel[0]);
-      for (i = 1; i < sigsel[0]; i++)
-	printf("0x%x, ",sigsel[i]);
-      if (sigsel[0] != 0)
-	printf("0x%x",sigsel[i]);
-      printf("})\n");
-      printf("%s:%d 0x%x got 0x%x from 0x%x\n",file,line,current_process(),
-	     *sig,sender((union SIGNAL**)(&sig)));
-    }
-  } else {
-    printf("%s:%d 0x%x receive_w_tmo(%u,{%d,",file,line,current_process(),tmo,
-	   sigsel[0]);
-      for (i = 1; i < sigsel[0]; i++)
-	printf("0x%x, ",sigsel[i]);
-      if (sigsel[0] != 0)
-	printf("0x%x",sigsel[i]);
-      printf("})\n");
-      sig = (SIGSELECT*)receive_w_tmo(tmo,sigsel);
-      printf("%s:%d 0x%x got ",file,line,current_process());
-      if (sig == NULL)
-	printf("TIMEOUT\n");
-      else
-	printf("0x%x from 0x%x\n",*sig,sender((union SIGNAL**)(&sig)));
-  }
-
-  return (union SIGNAL*)sig;
-}
-
-#define receive_w_tmo(tmo,sigsel) erts_ose_sys_receive_w_tmo(tmo,sigsel, \
-							     __FILE__,__LINE__)
-
-static union SIGNAL *erts_ose_sys_receive_fsem(OSTIME tmo,SIGSELECT *sigsel,
-					       OSFSEMVAL fsem,
-					       char *file,int line) {
-  SIGSELECT *sig;
-  int i;
-  if (tmo == 0) {
-    sig = (SIGSELECT*)receive_fsem(tmo,sigsel,fsem);
-    if (sig != NULL && sig != OS_RCV_FSEM) {
-      printf("%s:%d 0x%x receive_fsem(0,{%d,",file,line,current_process(),
-	     sigsel[0]);
-      for (i = 1; i < sigsel[0]; i++)
-	printf("0x%x, ",sigsel[i]);
-      if (sigsel[0] != 0)
-	printf("0x%x",sigsel[i]);
-      printf("},%d)\n",fsem);
-      printf("%s:%d 0x%x got 0x%x from 0x%x\n",file,line,current_process(),
-	     *sig,sender((union SIGNAL**)(&sig)));
-    }
-  } else {
-    printf("%s:%d 0x%x receive_fsem(%u,{%d,",file,line,current_process(),tmo,
-	   sigsel[0]);
-      for (i = 1; i < sigsel[0]; i++)
-	printf("0x%x, ",sigsel[i]);
-      if (sigsel[0] != 0)
-	printf("0x%x",sigsel[i]);
-      printf("},%d)\n",fsem);
-      sig = (SIGSELECT*)receive_fsem(tmo,sigsel,fsem);
-      printf("%s:%d 0x%x got ",file,line,current_process());
-      if (sig == NULL)
-	printf("TIMEOUT\n");
-      else if (sig == OS_RCV_FSEM)
-	printf("FSEM\n");
-      else
-	printf("0x%x from 0x%x\n",*sig,sender((union SIGNAL**)(&sig)));
-  }
-
-  return (union SIGNAL*)sig;
-}
-
-#define receive_fsem(tmo,sigsel,fsem) \
-  erts_ose_sys_receive_fsem(tmo,sigsel,fsem,__FILE__,__LINE__)
-*/
-#endif  /* _ERL_OSE_SYS_H */
diff --git a/erts/emulator/sys/ose/erl_ose_sys_ddll.c b/erts/emulator/sys/ose/erl_ose_sys_ddll.c
deleted file mode 100644
index 5051f7fcc1..0000000000
--- a/erts/emulator/sys/ose/erl_ose_sys_ddll.c
+++ /dev/null
@@ -1,127 +0,0 @@
-/*
- * %CopyrightBegin%
- *
- * Copyright Ericsson AB 2006-2013. All Rights Reserved.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- *
- * %CopyrightEnd%
- */
-
-/*
- * Interface functions to the dynamic linker using dl* functions.
- * (No support in OSE, we use static linkage instead)
- */
-
-#ifdef HAVE_CONFIG_H
-#  include "config.h"
-#endif
-
-#include "sys.h"
-#include "erl_vm.h"
-#include "global.h"
-
-
-void erl_sys_ddll_init(void) {
-}
-
-/*
- * Open a shared object
- */
-int erts_sys_ddll_open(const char *full_name, void **handle, ErtsSysDdllError* err)
-{
-   return ERL_DE_ERROR_NO_DDLL_FUNCTIONALITY;
-}
-
-int erts_sys_ddll_open_noext(char *dlname, void **handle, ErtsSysDdllError* err)
-{
-   return ERL_DE_ERROR_NO_DDLL_FUNCTIONALITY;
-}
-
-/*
- * Find a symbol in the shared object
- */
-int erts_sys_ddll_sym2(void *handle, const char *func_name, void **function,
-		       ErtsSysDdllError* err)
-{
-   return ERL_DE_ERROR_NO_DDLL_FUNCTIONALITY;
-}
-
-/* XXX:PaN These two will be changed with new driver interface! */
-
-/*
- * Load the driver init function, might appear under different names depending on object arch...
- */
-
-int erts_sys_ddll_load_driver_init(void *handle, void **function)
-{
-    void *fn;
-    int res;
-    if ((res = erts_sys_ddll_sym2(handle, "driver_init", &fn, NULL)) != ERL_DE_NO_ERROR) {
-	res = erts_sys_ddll_sym2(handle, "_driver_init", &fn, NULL);
-    }
-    if (res == ERL_DE_NO_ERROR) {
-	*function = fn;
-    }
-    return res;
-}
-
-int erts_sys_ddll_load_nif_init(void *handle, void **function, ErtsSysDdllError* err)
-{
-    void *fn;
-    int res;
-    if ((res = erts_sys_ddll_sym2(handle, "nif_init", &fn, err)) != ERL_DE_NO_ERROR) {
-	res = erts_sys_ddll_sym2(handle, "_nif_init", &fn, err);
-    }
-    if (res == ERL_DE_NO_ERROR) {
-	*function = fn;
-    }
-    return res;
-}
-
-/*
- * Call the driver_init function, whatever it's really called, simple on unix...
-*/
-void *erts_sys_ddll_call_init(void *function) {
-    void *(*initfn)(void) = function;
-    return (*initfn)();
-}
-void *erts_sys_ddll_call_nif_init(void *function) {
-    return erts_sys_ddll_call_init(function);
-}
-
-
-
-/*
- * Close a chared object
- */
-int erts_sys_ddll_close2(void *handle, ErtsSysDdllError* err)
-{
-   return ERL_DE_ERROR_NO_DDLL_FUNCTIONALITY;
-}
-
-
-/*
- * Return string that describes the (current) error
- */
-char *erts_sys_ddll_error(int code)
-{
-    return "Unspecified error";
-}
-
-void erts_sys_ddll_free_error(ErtsSysDdllError* err)
-{
-    if (err->str != NULL) {
-	erts_free(ERTS_ALC_T_DDLL_TMP_BUF, err->str);
-    }
-}
diff --git a/erts/emulator/sys/ose/erl_poll.c b/erts/emulator/sys/ose/erl_poll.c
deleted file mode 100644
index 5cee582a00..0000000000
--- a/erts/emulator/sys/ose/erl_poll.c
+++ /dev/null
@@ -1,818 +0,0 @@
-/*
- * %CopyrightBegin%
- *
- * Copyright Ericsson AB 2006-2012. All Rights Reserved.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- *
- * %CopyrightEnd%
- */
-
-/*
- * Description:	Poll interface suitable for ERTS on OSE with or without
- *              SMP support.
- *
- *		The interface is currently implemented using:
- *                - receive + receive_fsem
- *
- * Author: 	Lukas Larsson
- */
-
-#ifdef HAVE_CONFIG_H
-#  include "config.h"
-#endif
-
-#include "erl_thr_progress.h"
-#include "erl_driver.h"
-#include "erl_alloc.h"
-#include "erl_poll.h"
-
-#define NOFILE    4096
-
-/*
- * Some debug macros
- */
-
-/* #define HARDDEBUG
-#define HARDTRACE*/
-#ifdef HARDDEBUG
-#ifdef HARDTRACE
-#define HARDTRACEF(X, ...) { fprintf(stderr, X, __VA_ARGS__); fprintf(stderr,"\r\n"); }
-#else
-#define HARDTRACEF(...)
-#endif
-
-#else
-#define HARDTRACEF(X,...)
-#define HARDDEBUGF(...)
-#endif
-
-#if 0
-#define ERTS_POLL_DEBUG_PRINT
-#endif
-
-#if defined(DEBUG) && 0
-#define HARD_DEBUG
-#endif
-
-#  define SEL_ALLOC	erts_alloc
-#  define SEL_REALLOC	realloc_wrap
-#  define SEL_FREE	erts_free
-
-#ifdef ERTS_SMP
-
-#define ERTS_POLLSET_LOCK(PS) \
-  erts_smp_mtx_lock(&(PS)->mtx)
-#define ERTS_POLLSET_UNLOCK(PS) \
-  erts_smp_mtx_unlock(&(PS)->mtx)
-
-#else
-
-#define ERTS_POLLSET_LOCK(PS)
-#define ERTS_POLLSET_UNLOCK(PS)
-
-#endif
-
-/*
- * --- Data types ------------------------------------------------------------
- */
-
-union SIGNAL {
-    SIGSELECT sig_no;
-};
-
-typedef struct erts_sigsel_item_ ErtsSigSelItem;
-
-struct erts_sigsel_item_ {
-    ErtsSigSelItem *next;
-    ErtsSysFdType fd;
-    ErtsPollEvents events;
-};
-
-typedef struct erts_sigsel_info_ ErtsSigSelInfo;
-
-struct erts_sigsel_info_ {
-    ErtsSigSelInfo *next;
-    SIGSELECT signo;
-    ErlDrvOseEventId (*decode)(union SIGNAL* sig);
-    ErtsSigSelItem *fds;
-};
-
-struct ErtsPollSet_ {
-    SIGSELECT *sigs;
-    ErtsSigSelInfo *info;
-    Uint sig_count;
-    Uint item_count;
-    PROCESS interrupt;
-    erts_atomic32_t wakeup_state;
-    erts_atomic64_t timeout_time;
-#ifdef ERTS_SMP
-    erts_smp_mtx_t mtx;
-#endif
-};
-
-static int max_fds = -1;
-
-static ERTS_INLINE void
-init_timeout_time(ErtsPollSet ps)
-{
-    erts_atomic64_init_nob(&ps->timeout_time,
-			   (erts_aint64_t) ERTS_MONOTONIC_TIME_MAX);
-}
-
-static ERTS_INLINE void
-set_timeout_time(ErtsPollSet ps, ErtsMonotonicTime time)
-{
-    erts_atomic64_set_relb(&ps->timeout_time,
-			   (erts_aint64_t) time);
-}
-
-static ERTS_INLINE ErtsMonotonicTime
-get_timeout_time(ErtsPollSet ps)
-{
-    return (ErtsMonotonicTime) erts_atomic64_read_acqb(&ps->timeout_time);
-}
-
-#define ERTS_POLL_NOT_WOKEN		((erts_aint32_t) (1 << 0))
-#define ERTS_POLL_WOKEN_INTR		((erts_aint32_t) (1 << 1))
-#define ERTS_POLL_WOKEN_TIMEDOUT	((erts_aint32_t) (1 << 2))
-#define ERTS_POLL_WOKEN_IO_READY	((erts_aint32_t) (1 << 3))
-#define ERTS_POLL_SLEEPING	        ((erts_aint32_t) (1 << 4))
-
-/* signal list prototypes */
-static ErtsSigSelInfo *get_sigsel_info(ErtsPollSet ps, SIGSELECT signo);
-static ErtsSigSelItem *get_sigsel_item(ErtsPollSet ps, ErtsSysFdType fd);
-static ErtsSigSelInfo *add_sigsel_info(ErtsPollSet ps, ErtsSysFdType fd,
-				       ErlDrvOseEventId (*decode)(union SIGNAL* sig));
-static ErtsSigSelItem *add_sigsel_item(ErtsPollSet ps, ErtsSysFdType fd,
-				       ErlDrvOseEventId (*decode)(union SIGNAL* sig));
-static int del_sigsel_info(ErtsPollSet ps, ErtsSigSelInfo *info);
-static int del_sigsel_item(ErtsPollSet ps, ErtsSigSelItem *item);
-static int update_sigsel(ErtsPollSet ps);
-
-static ErtsSigSelInfo *
-get_sigsel_info(ErtsPollSet ps, SIGSELECT signo) {
-    ErtsSigSelInfo *curr = ps->info;
-    while (curr != NULL) {
-	if (curr->signo == signo)
-	    return curr;
-	curr = curr->next;
-    }
-    return NULL;
-}
-
-static ErtsSigSelItem *
-get_sigsel_item(ErtsPollSet ps, ErtsSysFdType fd) {
-    ErtsSigSelInfo *info = get_sigsel_info(ps,fd->signo);
-    ErtsSigSelItem *curr;
-
-    if (info == NULL)
-	return NULL;
-
-    curr = info->fds;
-
-    while (curr != NULL) {
-	if (curr->fd->id == fd->id) {
-	    ASSERT(curr->fd->signo == fd->signo);
-	    return curr;
-	}
-	curr = curr->next;
-    }
-    return NULL;
-}
-
-static ErtsSigSelInfo *
-add_sigsel_info(ErtsPollSet ps, ErtsSysFdType fd,
-		ErlDrvOseEventId (*decode)(union SIGNAL* sig)) {
-    ErtsSigSelInfo *info = SEL_ALLOC(ERTS_ALC_T_POLLSET,
-		       sizeof(ErtsSigSelInfo));
-    info->next = ps->info;
-    info->fds = NULL;
-    info->signo = fd->signo;
-    info->decode = decode;
-    ps->info = info;
-    ps->sig_count++;
-    return info;
-}
-
-static ErtsSigSelItem *
-add_sigsel_item(ErtsPollSet ps, ErtsSysFdType fd,
-		ErlDrvOseEventId (*decode)(union SIGNAL* sig)) {
-    ErtsSigSelInfo *info = get_sigsel_info(ps,fd->signo);
-    ErtsSigSelItem *item = SEL_ALLOC(ERTS_ALC_T_POLLSET,
-			   sizeof(ErtsSigSelItem));
-    if (info == NULL)
-	info = add_sigsel_info(ps, fd, decode);
-    if (info->decode != decode) {
-	erts_dsprintf_buf_t *dsbufp = erts_create_logger_dsbuf();
-	erts_dsprintf(dsbufp, "erts_poll_control() inconsistency: multiple resolve_signal functions for same signal (%d)\n",
-	    fd->signo);
-	erts_send_error_to_logger_nogl(dsbufp);
-    }
-    ASSERT(info->decode == decode);
-    item->next = info->fds;
-    item->fd = fd;
-    item->events = 0;
-    info->fds = item;
-    ps->item_count++;
-    return item;
-}
-
-static int del_sigsel_info(ErtsPollSet ps, ErtsSigSelInfo *info) {
-    ErtsSigSelInfo *curr, *prev;
-
-    if (ps->info == info) {
-	ps->info = ps->info->next;
-    } else {
-	curr = ps->info->next;
-	prev = ps->info;
-
-	while (curr != info) {
-	    if (curr == NULL)
-		return 1;
-	    prev = curr;
-	    curr = curr->next;
-	}
-	prev->next = curr->next;
-    }
-
-    ps->sig_count--;
-    SEL_FREE(ERTS_ALC_T_POLLSET, info);
-    return 0;
-}
-
-static int del_sigsel_item(ErtsPollSet ps, ErtsSigSelItem *item) {
-    ErtsSigSelInfo *info = get_sigsel_info(ps,item->fd->signo);
-    ErtsSigSelItem *curr, *prev;
-
-    ps->item_count--;
-    ASSERT(ps->item_count >= 0);
-
-    if (info->fds == item) {
-	info->fds = info->fds->next;
-	SEL_FREE(ERTS_ALC_T_POLLSET,item);
-	if (info->fds == NULL)
-	    return del_sigsel_info(ps,info);
-	return 0;
-    }
-
-    curr = info->fds->next;
-    prev = info->fds;
-
-    while (curr != item) {
-	if (curr == NULL) {
-	    /* We did not find an item to delete so we have to
-	     * increment item count again.
-	     */
-	    ps->item_count++;
-	    return 1;
-	}
-	prev = curr;
-	curr = curr->next;
-    }
-    prev->next = curr->next;
-    SEL_FREE(ERTS_ALC_T_POLLSET,item);
-    return 0;
-}
-
-#ifdef ERTS_SMP
-
-static void update_redir_tables(ErtsPollSet ps) {
-  struct OS_redir_entry *redir_table;
-  PROCESS sched_1 = ERTS_SCHEDULER_IX(0)->tid.id;
-  int i;
-  redir_table = SEL_ALLOC(ERTS_ALC_T_POLLSET,
-			  sizeof(struct OS_redir_entry)*(ps->sig_count+1));
-
-  redir_table[0].sig = ps->sig_count+1;
-  redir_table[0].pid = 0;
-
-  for (i = 1; i < ps->sig_count+1; i++) {
-    redir_table[i].sig = ps->sigs[i];
-    redir_table[i].pid = sched_1;
-  }
-
-  for (i = 1; i < erts_no_schedulers; i++) {
-    ErtsSchedulerData *esdp = ERTS_SCHEDULER_IX(i);
-    set_redirection(esdp->tid.id,redir_table);
-  }
-
-  SEL_FREE(ERTS_ALC_T_POLLSET,redir_table);
-}
-
-#endif
-
-static int update_sigsel(ErtsPollSet ps) {
-    ErtsSigSelInfo *info = ps->info;
-
-    int i;
-
-    if (ps->sigs != NULL)
-	SEL_FREE(ERTS_ALC_T_POLLSET,ps->sigs);
-
-    if (ps->sig_count == 0) {
-	/* If there are no signals we place a non-valid signal to make sure that
-	 * we do not trigger on a any unrelated signals which are sent to the
-	 * process.
-	 */
-	ps->sigs = SEL_ALLOC(ERTS_ALC_T_POLLSET,sizeof(SIGSELECT)*(2));
-	ps->sigs[0] = 1;
-	ps->sigs[1] = ERTS_SIGNAL_INVALID;
-	return 0;
-    }
-
-    ps->sigs = SEL_ALLOC(ERTS_ALC_T_POLLSET,sizeof(SIGSELECT)*(ps->sig_count+1));
-    ps->sigs[0] = ps->sig_count;
-
-    for (i = 1; info != NULL; i++, info = info->next)
-	ps->sigs[i] = info->signo;
-
-#ifdef ERTS_SMP
-    update_redir_tables(ps);
-#endif
-
-    return 0;
-}
-
-static ERTS_INLINE void
-wake_poller(ErtsPollSet ps)
-{
-  erts_aint32_t wakeup_state;
-
-  ERTS_THR_MEMORY_BARRIER;
-  wakeup_state = erts_atomic32_read_nob(&ps->wakeup_state);
-  while (wakeup_state != ERTS_POLL_WOKEN_IO_READY
-	 && wakeup_state != ERTS_POLL_WOKEN_INTR) {
-    erts_aint32_t act = erts_atomic32_cmpxchg_nob(&ps->wakeup_state,
-						  ERTS_POLL_WOKEN_INTR,
-						  wakeup_state);
-    if (act == wakeup_state) {
-      wakeup_state = act;
-      break;
-    }
-    wakeup_state = act;
-  }
-  if (wakeup_state == ERTS_POLL_SLEEPING) {
-    /*
-     * Since we don't know the internals of signal_fsem() we issue
-     * a memory barrier as a safety precaution ensuring that
-     * the store we just made to wakeup_state wont be reordered
-     * with loads in signal_fsem().
-     */
-    ERTS_THR_MEMORY_BARRIER;
-    signal_fsem(ps->interrupt);
-  }
-}
-
-static ERTS_INLINE void
-reset_interrupt(ErtsPollSet ps)
-{
-    /* We need to keep io-ready if set */
-    erts_aint32_t wakeup_state = erts_atomic32_read_nob(&ps->wakeup_state);
-    while (wakeup_state != ERTS_POLL_NOT_WOKEN &&
-	   wakeup_state != ERTS_POLL_SLEEPING) {
-	erts_aint32_t act = erts_atomic32_cmpxchg_nob(&ps->wakeup_state,
-						      ERTS_POLL_NOT_WOKEN,
-						      wakeup_state);
-	if (wakeup_state == act)
-	    break;
-	wakeup_state = act;
-    }
-    ERTS_THR_MEMORY_BARRIER;
-}
-
-static ERTS_INLINE void
-set_interrupt(ErtsPollSet ps)
-{
-    wake_poller(ps);
-}
-
-void erts_poll_interrupt(ErtsPollSet ps,int set) {
-    HARDTRACEF("erts_poll_interrupt called!\n");
-
-    if (!set)
-	reset_interrupt(ps);
-    else
-	set_interrupt(ps);
-
-}
-
-void erts_poll_interrupt_timed(ErtsPollSet ps,
-			       int set,
-			       ErtsTimeoutTime timeout_time) {
-    HARDTRACEF("erts_poll_interrupt_timed called!\n");
-
-    if (!set)
-	reset_interrupt(ps);
-    else if (get_timeout_time(ps) > timeout_time)
-        set_interrupt(ps);
-}
-
-ErtsPollEvents erts_poll_control(ErtsPollSet ps, ErtsSysFdType fd,
-	ErtsPollEvents pe, int on, int* do_wake) {
-    ErtsSigSelItem *curr;
-    ErtsPollEvents new_events;
-    int old_sig_count;
-
-    HARDTRACEF(
-	    "%ux: In erts_poll_control, fd = %d, pe = %d, on = %d, *do_wake = %d, curr = 0x%xu",
-	    ps, fd, pe, on, do_wake, curr);
-
-    ERTS_POLLSET_LOCK(ps);
-
-    if (on && (pe & ERTS_POLL_EV_IN) && (pe & ERTS_POLL_EV_OUT)) {
-      /* Check to make sure both in and out are not used at the same time */
-      new_events = ERTS_POLL_EV_NVAL;
-      goto done;
-    }
-
-    curr = get_sigsel_item(ps, fd);
-    old_sig_count = ps->sig_count;
-
-    if (curr == NULL && on) {
-	curr = add_sigsel_item(ps, fd, fd->resolve_signal);
-    } else if (curr == NULL && !on) {
-        new_events = ERTS_POLL_EV_NVAL;
-	goto done;
-    }
-
-    new_events = curr->events;
-
-    if (pe == 0) {
-	*do_wake = 0;
-	goto done;
-    }
-
-    if (on) {
-	new_events |= pe;
-	curr->events = new_events;
-    } else {
-	new_events &= ~pe;
-	curr->events = new_events;
-	if (new_events == 0 && del_sigsel_item(ps, curr)) {
-	    new_events = ERTS_POLL_EV_NVAL;
-	    goto done;
-	}
-    }
-
-    if (ps->sig_count != old_sig_count) {
-      if (update_sigsel(ps))
-	new_events = ERTS_POLL_EV_NVAL;
-    }
-done:
-    ERTS_POLLSET_UNLOCK(ps);
-    HARDTRACEF("%ux: Out erts_poll_control", ps);
-    return new_events;
-}
-
-int erts_poll_wait(ErtsPollSet ps,
-		   ErtsPollResFd pr[],
-		   int *len,
-		   ErtsMonotonicTime timeout_time)
-{
-    int res = ETIMEDOUT, no_fds, currid = 0;
-    OSTIME timeout;
-    union SIGNAL *sig;
-    ErtsMonotonicTime current_time, diff_time, timeout;
-    // HARDTRACEF("%ux: In erts_poll_wait",ps);
-    if (ps->interrupt == (PROCESS)0)
-      ps->interrupt = current_process();
-
-    ASSERT(current_process() == ps->interrupt);
-    ASSERT(get_fsem(current_process()) == 0);
-    ASSERT(erts_atomic32_read_nob(&ps->wakeup_state) &
-	   (ERTS_POLL_NOT_WOKEN | ERTS_POLL_WOKEN_INTR));
-    /* Max no of spots avable in pr */
-    no_fds = *len;
-
-    *len = 0;
-
-    /* erts_printf("Entering erts_poll_wait(), timeout_time=%bps\n",
-		   timeout_time); */
-
-    if (timeout_time == ERTS_POLL_NO_TIMEOUT) {
-    no_timeout:
-	timeout = (OSTIME) 0;
-	save_timeout_time = ERTS_MONOTONIC_TIME_MIN;
-    }
-    else {
-	ErtsMonotonicTime current_time, diff_time;
-	current_time = erts_get_monotonic_time(NULL);
-	diff_time = timeout_time - current_time;
-	if (diff_time <= 0)
-	    goto no_timeout;
-	diff_time = (ERTS_MONOTONIC_TO_MSEC(diff_time - 1) + 1);
-	if (diff_time > INT_MAX)
-	    diff_time = INT_MAX;
-	timeout = (OSTIME) diff_time;
-	save_timeout_time = current_time;
-	save_timeout_time += ERTS_MSEC_TO_MONOTONIC(diff_time);
-    }
-
-    set_timeout_time(ps, save_timeout_time);
-
-    while (currid < no_fds) {
-      if (timeout > 0) {
-	erts_aint32_t act = erts_atomic32_cmpxchg_nob(&ps->wakeup_state,
-						      ERTS_POLL_SLEEPING,
-						      ERTS_POLL_NOT_WOKEN);
-	if (act == ERTS_POLL_NOT_WOKEN) {
-#ifdef ERTS_SMP
-	  erts_thr_progress_prepare_wait(NULL);
-#endif
-	  sig = receive_fsem(timeout, ps->sigs, 1);
-#ifdef ERTS_SMP
-	  erts_thr_progress_finalize_wait(NULL);
-#endif
-	} else {
-	  ASSERT(act == ERTS_POLL_WOKEN_INTR);
-	  sig = OS_RCV_FSEM;
-	}
-      } else
-	  sig = receive_w_tmo(0, ps->sigs);
-
-	if (sig == NULL) {
-	  if (timeout > 0) {
-	    erts_aint32_t act = erts_atomic32_cmpxchg_nob(&ps->wakeup_state,
-							  ERTS_POLL_WOKEN_TIMEDOUT,
-							  ERTS_POLL_SLEEPING);
-	    if (act == ERTS_POLL_WOKEN_INTR)
-	      /* Restore fsem as it was signaled but we got a timeout */
-	      wait_fsem(1);
-	    } else
-	    erts_atomic32_cmpxchg_nob(&ps->wakeup_state,
-				      ERTS_POLL_WOKEN_TIMEDOUT,
-				      ERTS_POLL_NOT_WOKEN);
-	    break;
-	} else if (sig == OS_RCV_FSEM) {
-	  ASSERT(erts_atomic32_read_nob(&ps->wakeup_state) == ERTS_POLL_WOKEN_INTR);
-	  break;
-	}
-       {
-          ErtsSigSelInfo *info = get_sigsel_info(ps, sig->sig_no);
-          struct erts_sys_fd_type fd = { sig->sig_no, info->decode(sig) };
-          ErtsSigSelItem *item = get_sigsel_item(ps, &fd);
-
-	  ASSERT(sig);
-          if (currid == 0 && timeout > 0) {
-	    erts_aint32_t act = erts_atomic32_cmpxchg_nob(&ps->wakeup_state,
-							  ERTS_POLL_WOKEN_IO_READY,
-							  ERTS_POLL_SLEEPING);
-	    if (act == ERTS_POLL_WOKEN_INTR) {
-	      /* Restore fsem as it was signaled but we got a msg */
-	      wait_fsem(1);
-	      act = erts_atomic32_cmpxchg_nob(&ps->wakeup_state,
-					      ERTS_POLL_WOKEN_IO_READY,
-					      ERTS_POLL_WOKEN_INTR);
-	    }
-	  } else if (currid == 0) {
-	    erts_atomic32_set_nob(&ps->wakeup_state,
-				  ERTS_POLL_WOKEN_IO_READY);
-	  }
-
-	  if (item == NULL) {
-	    erts_dsprintf_buf_t *dsbufp = erts_create_logger_dsbuf();
-	    erts_dsprintf(
-                dsbufp,
-                "erts_poll_wait() failed: found unkown signal id %d (signo %u) "
-		"(curr_proc 0x%x)\n",
-                fd.id, fd.signo, current_process());
-             erts_send_error_to_logger_nogl(dsbufp);
-	     timeout = 0;
-             /* Under normal circumstances the signal is deallocated by the
-              * driver that issued the select operation. But in this case
-              * there's no driver waiting for such signal so we have to
-              * deallocate it here */
-             if (sig)
-                 free_buf(&sig);
-	  } else {
-	    int i;
-	    struct erts_sys_fd_type *fd = NULL;
-	    ErtsPollOseMsgList *tl,*new;
-
-	    /* Check if this fd has already been triggered by a previous signal */
-	    for (i = 0; i < currid;i++) {
-	      if (pr[i].fd == item->fd) {
-		fd = pr[i].fd;
-		pr[i].events |= item->events;
-		break;
-	      }
-	    }
-
-	    /* First time this fd is triggered */
-	    if (fd == NULL) {
-	      pr[currid].fd = item->fd;
-	      pr[currid].events = item->events;
-	      fd = item->fd;
-	      timeout = 0;
-	      currid++;
-	    }
-
-	    /* Insert new signal in approriate list */
-	    new = erts_alloc(ERTS_ALC_T_FD_SIG_LIST,sizeof(ErtsPollOseMsgList));
-	    new->next = NULL;
-	    new->data = sig;
-
-	    ethr_mutex_lock(&fd->mtx);
-	    tl = fd->msgs;
-
-	    if (tl == NULL) {
-	      fd->msgs = new;
-	    } else {
-	      while (tl->next != NULL)
-		tl = tl->next;
-	      tl->next = new;
-	    }
-	    ethr_mutex_unlock(&fd->mtx);
-          }
-
-       }
-    }
-
-    {
-       erts_aint32_t wakeup_state = erts_atomic32_read_nob(&ps->wakeup_state);
-
-       switch (wakeup_state) {
-          case ERTS_POLL_WOKEN_IO_READY:
-             res = 0;
-             break;
-          case ERTS_POLL_WOKEN_INTR:
-             res = EINTR;
-             break;
-          case ERTS_POLL_WOKEN_TIMEDOUT:
-             res = ETIMEDOUT;
-             break;
-          case ERTS_POLL_NOT_WOKEN:
-             /* This happens when we get an invalid signal only */
-             res = EINVAL;
-             break;
-          default:
-             res = 0;
-             erl_exit(ERTS_ABORT_EXIT,
-                      "%s:%d: Internal error: Invalid wakeup_state=%d\n",
-                      __FILE__, __LINE__, (int) wakeup_state);
-       }
-    }
-
-    erts_atomic32_set_nob(&ps->wakeup_state, ERTS_POLL_NOT_WOKEN);
-    set_timeout_time(ps, ERTS_MONOTONIC_TIME_MAX);
-
-    *len = currid;
-
-    // HARDTRACEF("%ux: Out erts_poll_wait",ps);
-    return res;
-}
-
-int erts_poll_max_fds(void)
-{
-
-    HARDTRACEF("In/Out erts_poll_max_fds -> %d",max_fds);
-    return max_fds;
-}
-
-void erts_poll_info(ErtsPollSet ps,
-		    ErtsPollInfo *pip)
-{
-    Uint size = 0;
-    Uint num_events = 0;
-
-    size += sizeof(struct ErtsPollSet_);
-    size += sizeof(ErtsSigSelInfo)*ps->sig_count;
-    size += sizeof(ErtsSigSelItem)*ps->item_count;
-    size += sizeof(SIGSELECT)*(ps->sig_count+1);
-
-    pip->primary = "receive_fsem";
-
-    pip->fallback = NULL;
-
-    pip->kernel_poll = NULL;
-
-    pip->memory_size = size;
-
-    pip->poll_set_size = num_events;
-
-    pip->fallback_poll_set_size = 0;
-
-    pip->lazy_updates = 0;
-
-    pip->pending_updates = 0;
-
-    pip->batch_updates = 0;
-
-    pip->concurrent_updates = 0;
-
-
-    pip->max_fds = erts_poll_max_fds();
-    HARDTRACEF("%ux: Out erts_poll_info",ps);
-
-}
-
-ErtsPollSet erts_poll_create_pollset(void)
-{
-    ErtsPollSet ps = SEL_ALLOC(ERTS_ALC_T_POLLSET,
-			       sizeof(struct ErtsPollSet_));
-
-    ps->sigs       = NULL;
-    ps->sig_count  = 0;
-    ps->item_count = 0;
-    ps->info       = NULL;
-    ps->interrupt  = (PROCESS)0;
-    erts_atomic32_init_nob(&ps->wakeup_state, ERTS_POLL_NOT_WOKEN);
-    init_timeout_time(ps);
-#ifdef ERTS_SMP
-    erts_smp_mtx_init(&ps->mtx, "pollset");
-#endif
-    update_sigsel(ps);
-    HARDTRACEF("%ux: Out erts_poll_create_pollset",ps);
-    return ps;
-}
-
-void erts_poll_destroy_pollset(ErtsPollSet ps)
-{
-    ErtsSigSelInfo *info;
-    for (info = ps->info; ps->info != NULL; info = ps->info, ps->info = ps->info->next) {
-	ErtsSigSelItem *item;
-	for (item = info->fds; info->fds != NULL; item = info->fds, info->fds = info->fds->next)
-	    SEL_FREE(ERTS_ALC_T_POLLSET, item);
-	SEL_FREE(ERTS_ALC_T_POLLSET, info);
-    }
-
-    SEL_FREE(ERTS_ALC_T_POLLSET,ps->sigs);
-
-#ifdef ERTS_SMP
-    erts_smp_mtx_destroy(&ps->mtx);
-#endif
-
-    SEL_FREE(ERTS_ALC_T_POLLSET,ps);
-}
-
-void  erts_poll_init(void)
-{
-    HARDTRACEF("In %s", __FUNCTION__);
-    max_fds = 256;
-
-    HARDTRACEF("Out %s", __FUNCTION__);
-}
-
-
-/* OSE driver functions */
-
-union SIGNAL *erl_drv_ose_get_signal(ErlDrvEvent drv_ev) {
-    struct erts_sys_fd_type *ev = (struct erts_sys_fd_type *)drv_ev;
-    ethr_mutex_lock(&ev->mtx);
-    if (ev->msgs == NULL) {
-      ethr_mutex_unlock(&ev->mtx);
-      return NULL;
-    } else {
-      ErtsPollOseMsgList *msg = ev->msgs;
-      union SIGNAL *sig = (union SIGNAL*)msg->data;
-      ASSERT(msg->data);
-      ev->msgs = msg->next;
-      ethr_mutex_unlock(&ev->mtx);
-      erts_free(ERTS_ALC_T_FD_SIG_LIST,msg);
-      restore(sig);
-      return sig;
-    }
-}
-
-ErlDrvEvent
-erl_drv_ose_event_alloc(SIGSELECT signo, ErlDrvOseEventId id,
-			ErlDrvOseEventId (*resolve_signal)(union SIGNAL *sig), void *extra) {
-  struct erts_sys_fd_type *ev = erts_alloc(ERTS_ALC_T_DRV_EV,
-					   sizeof(struct erts_sys_fd_type));
-  ev->signo = signo;
-  ev->extra = extra;
-  ev->id = id;
-  ev->msgs = NULL;
-  ev->resolve_signal = resolve_signal;
-  ethr_mutex_init(&ev->mtx);
-  return (ErlDrvEvent)ev;
-}
-
-void erl_drv_ose_event_free(ErlDrvEvent drv_ev) {
-  struct erts_sys_fd_type *ev = (struct erts_sys_fd_type *)drv_ev;
-  ASSERT(ev->msgs == NULL);
-  ethr_mutex_destroy(&ev->mtx);
-  erts_free(ERTS_ALC_T_DRV_EV,ev);
-}
-
-void erl_drv_ose_event_fetch(ErlDrvEvent drv_ev, SIGSELECT *signo,
-                             ErlDrvOseEventId *id, void **extra) {
-  struct erts_sys_fd_type *ev = (struct erts_sys_fd_type *)drv_ev;
-  if (signo)
-    *signo = ev->signo;
-  if (extra)
-    *extra = ev->extra;
-  if (id)
-    *id = ev->id;
-}
diff --git a/erts/emulator/sys/ose/erts.sig b/erts/emulator/sys/ose/erts.sig
deleted file mode 100644
index 78b883ee6c..0000000000
--- a/erts/emulator/sys/ose/erts.sig
+++ /dev/null
@@ -1,17 +0,0 @@
-#ifndef ERTS_OSE_SIGNALS
-#define ERTS_OSE_SIGNALS
-
-#ifndef ERTS_OSE_SIGNAL_BASE
-#define ERTS_OSE_SIGNAL_BASE 0x01900280
-#endif
-
-#define ERTS_SIGNAL_INVALID        ERTS_OSE_SIGNAL_BASE
-#define ERTS_SIGNAL_FD_DRV_CONFIG  ERTS_OSE_SIGNAL_BASE+1
-#define ERTS_SIGNAL_FD_DRV_ASYNC   ERTS_OSE_SIGNAL_BASE+2
-#define ERTS_SIGNAL_OSE_DRV_ATTACH ERTS_OSE_SIGNAL_BASE+3
-#define ERTS_SIGNAL_OSE_DRV_HUNT   ERTS_OSE_SIGNAL_BASE+4
-
-#define ERTS_SIGNAL_RUN_ERL_SETUP  ERTS_OSE_SIGNAL_BASE+100
-#define ERTS_SIGNAL_RUN_ERL_DAEMON ERTS_OSE_SIGNAL_BASE+101
-
-#endif
diff --git a/erts/emulator/sys/ose/gcc_4.4.3_lm_ppc.lcf b/erts/emulator/sys/ose/gcc_4.4.3_lm_ppc.lcf
deleted file mode 100644
index a19d23facf..0000000000
--- a/erts/emulator/sys/ose/gcc_4.4.3_lm_ppc.lcf
+++ /dev/null
@@ -1,182 +0,0 @@
-/*******************************************************************************
- * Copyright (C) 2013-2014 by Enea Software AB,
- * All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions are met:
- *
- * 1. Redistributions of source code must retain the above copyright notice,
- *    this list of conditions and the following disclaimer.
- *
- * 2. Redistributions in binary form must reproduce the above copyright notice,
- *    this list of conditions and the following disclaimer in the documentation
- *    and/or other materials provided with the distribution.
- *
- * 3. Neither the name of the copyright holder nor the names of its contributors
- *    may be used to endorse or promote products derived from this software
- *    without specific prior written permission.
- *
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
- * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
- * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
- * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
- * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
- * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
- * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
- * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
- * POSSIBILITY OF SUCH DAMAGE.
- ******************************************************************************/
-
-OUTPUT_FORMAT("elf32-powerpc", "elf32-powerpc", "elf32-powerpc")
-OUTPUT_ARCH("powerpc")
-ENTRY("crt0_lm")
-MEMORY
-{
- rom : ORIGIN = 0x01000000, LENGTH = 0x01000000
- ram : ORIGIN = 0x02000000, LENGTH = 0x01000000
-}
-PHDRS
-{
-  ph_conf PT_LOAD ;
- ph_rom PT_LOAD ;
- ph_ram PT_LOAD ;
-}
-SECTIONS
-{
- .text :
- {
-  *(.text_first)
-  *(.text)
-  *(.text.*)
-  *(.stub)
-  *(oscode)
-  *(.init*)
-  *(.fini*)
-  *(.gnu.warning)
-  *(.gnu.linkonce.t.*)
-  *(.glue_7t)
-  *(.glue_7)
- } > rom :ph_rom = 0
- .ose_sfk_biosentry :
- {
-  *(.ose_sfk_biosentry)
-        } > rom :ph_rom
- .ctors :
- {
-  __CTOR_LIST__ = .;
-  *(.ctors)
-  *(SORT(.ctors.*))
-  __CTOR_END__ = .;
- } > rom :ph_rom
- .dtors :
- {
-  __DTOR_LIST__ = .;
-  *(.dtors)
-  *(SORT(.dtors.*))
-  __DTOR_END__ = .;
- } > rom :ph_rom
- OSESYMS :
- {
-  *(.osesyms)
- } > rom :ph_rom
- .rodata :
- {
-  *(.rodata)
-  *(.rodata.*)
-  *(.gnu.linkonce.r.*)
- } > rom :ph_rom
- .eh_frame_hdr :
- {
-  *(.eh_frame_hdr)
- } > rom :ph_rom
- .eh_frame :
- {
-  __EH_FRAME_BEGIN__ = .;
-  *(.eh_frame)
-  LONG(0)
-  __EH_FRAME_END__ = .;
- } > rom :ph_rom
- .gcc_except_table :
- {
-  *(.gcc_except_table .gcc_except_table.*)
- } > rom :ph_rom
- .sdata2 :
- {
-  PROVIDE (_SDA2_BASE_ = .);
-  *(.sdata2)
-  *(.sdata2.*)
-  *(.gnu.linkonce.s2.*)
- } > rom :ph_rom
- .sbss2 :
- {
-  *(.sbss2)
-  *(.sbss2.*)
-  *(.gnu.linkonce.sb2.*)
- } > rom :ph_rom
- LMCONF :
- {
-  obj/?*?/ose_confd.o(.rodata)
-  *(LMCONF)
- } > rom :ph_conf
- .data :
- {
-  LONG(0xDEADBABE)
-  *(.data)
-  *(.data.*)
-  *(.gnu.linkonce.d.*)
-  SORT(CONSTRUCTORS)
-  . = ALIGN(0x10);
- } > ram :ph_ram = 0
-        .sdata2 :
-   {
-         _SDA2_BASE_ = .;
-             *(.sdata2 .sdata2.* .gnu.linkonce.s2.*)
-   }> ram :ph_ram
- .sdata :
- {
-         PROVIDE (_SDA_BASE_ = .);
-  *(.sdata)
-  *(.sdata.*)
-  *(.gnu.linkonce.s.*)
- } > ram :ph_ram
- .sbss :
- {
-  *(.sbss)
-  *(.sbss.*)
-  *(.scommon)
-  *(.gnu.linkonce.sb.*)
- } > ram :ph_ram
- .bss (NOLOAD) :
- {
-  *(.bss)
-  *(.bss.*)
-  *(COMMON)
-  *(.gnu.linkonce.b.*)
-  *(.osvars)
- } > ram :ph_ram
- .ignore (NOLOAD) :
- {
-  *(.rel.dyn)
- } > ram :ph_ram
- .debug 0 : { *(.debug) }
- .line 0 : { *(.line) }
- .debug_srcinfo 0 : { *(.debug_srcinfo) }
- .debug_sfnames 0 : { *(.debug_sfnames) }
- .debug_aranges 0 : { *(.debug_aranges) }
- .debug_pubnames 0 : { *(.debug_pubnames) }
- .debug_info 0 : { *(.debug_info) *(.gnu.linkonce.wi.*) }
- .debug_abbrev 0 : { *(.debug_abbrev) }
- .debug_line 0 : { *(.debug_line) }
- .debug_frame 0 : { *(.debug_frame) }
- .debug_str 0 : { *(.debug_str) }
- .debug_loc 0 : { *(.debug_loc) }
- .debug_macinfo 0 : { *(.debug_macinfo) }
- .debug_weaknames 0 : { *(.debug_weaknames) }
- .debug_funcnames 0 : { *(.debug_funcnames) }
- .debug_typenames 0 : { *(.debug_typenames) }
- .debug_varnames 0 : { *(.debug_varnames) }
-}
-__OSESYMS_START = ADDR(OSESYMS);
-__OSESYMS_END = ADDR(OSESYMS) + SIZEOF(OSESYMS);
diff --git a/erts/emulator/sys/ose/gcc_4.6.3_lm_ppc.lcf b/erts/emulator/sys/ose/gcc_4.6.3_lm_ppc.lcf
deleted file mode 100644
index 3440c2961b..0000000000
--- a/erts/emulator/sys/ose/gcc_4.6.3_lm_ppc.lcf
+++ /dev/null
@@ -1,242 +0,0 @@
-/*******************************************************************************
- * Copyright (C) 2013-2014 by Enea Software AB,
- * All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions are met:
- *
- * 1. Redistributions of source code must retain the above copyright notice,
- *    this list of conditions and the following disclaimer.
- *
- * 2. Redistributions in binary form must reproduce the above copyright notice,
- *    this list of conditions and the following disclaimer in the documentation
- *    and/or other materials provided with the distribution.
- *
- * 3. Neither the name of the copyright holder nor the names of its contributors
- *    may be used to endorse or promote products derived from this software
- *    without specific prior written permission.
- *
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
- * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
- * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
- * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
- * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
- * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
- * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
- * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
- * POSSIBILITY OF SUCH DAMAGE.
- ******************************************************************************/
-
-OUTPUT_FORMAT("elf32-powerpc", "elf32-powerpc", "elf32-powerpc")
-OUTPUT_ARCH("powerpc")
-
-ENTRY("crt0_lm")
-
-/* Note:
- * You may have to increase the length of the "rom" memory region and the
- * origin and length of the "ram" memory region below depending on the size
- * of the code and data in your load module.
- */
-
-MEMORY
-{
-	conf : ORIGIN = 0x00100000, LENGTH = 0x00030000
-	rom  : ORIGIN = 0x01000000, LENGTH = 0x01000000
-	ram  : ORIGIN = 0x03000000, LENGTH = 0x01000000
-}
-
-PHDRS
-{
-	ph_conf PT_LOAD ;
-	ph_rom PT_LOAD ;
-	ph_ram PT_LOAD ;
-}
-
-SECTIONS
-{
-/*---------------------------------------------------------------------------
- *		Load module configuration area
- *-------------------------------------------------------------------------*/
-
-	/* Load module configuration section. */
-	LMCONF :
-	{
-		obj/?*?/ose_confd.o(.rodata)
-		*(LMCONF)
-	} > conf :ph_conf
-
-/*---------------------------------------------------------------------------
- *		Read-only area
- *-------------------------------------------------------------------------*/
-
-	/* Code section. */
-	.text :
-	{
-		*(.text)
-		*(.text.*)
-		*(.stub)
-		*(oscode)
-		*(.init*)
-		*(.fini*)
-		*(.gnu.warning)
-		*(.gnu.linkonce.t.*)
-	} > rom :ph_rom = 0
-
-	/* OSE symbols section. */
-	OSESYMS :
-	{
-		*(.osesyms)
-	} > rom :ph_rom
-
-	/* Read-only data section. */
-	.rodata :
-	{
-		*(.rodata)
-		*(.rodata.*)
-		*(.gnu.linkonce.r.*)
-	} > rom :ph_rom
-
-	/* C++ exception handling section. */
-	.eh_frame :
-	{
-		__EH_FRAME_BEGIN__ = .;
-		*(.eh_frame)
-		LONG(0)
-		__EH_FRAME_END__ = .;
-	} > rom :ph_rom
-
-	/* C++ exception handling section. */
-	.gcc_except_table :
-	{
-		*(.gcc_except_table .gcc_except_table.*)
-	} > rom :ph_rom
-
-	/* PowerPC EABI initialized read-only data section. */
-	.sdata2 :
-	{
-		PROVIDE (_SDA2_BASE_ = .);
-		*(.sdata2)
-		*(.sdata2.*)
-		*(.gnu.linkonce.s2.*)
-	} > rom :ph_rom
-
-	/* PowerPC EABI uninitialized read-only data section. */
-	.sbss2 :
-	{
-		*(.sbss2)
-		*(.sbss2.*)
-		*(.gnu.linkonce.sb2.*)
-	} > rom :ph_rom
-
-/*---------------------------------------------------------------------------
- *		Read-write area
- *-------------------------------------------------------------------------*/
-
-	/*-------------------------------------------------------------------
-	 * Initialized data (copied by PM)
-	 *-----------------------------------------------------------------*/
-
-	/* Data section. */
-	.data :
-	{
-		*(.data)
-		*(.data.*)
-		*(.gnu.linkonce.d.*)
-		SORT(CONSTRUCTORS)
-	} > ram :ph_ram
-
-	/* C++ constructor section. */
-	.ctors :
-	{
-		__CTOR_LIST__ = .;
-		*(.ctors)
-		*(SORT(.ctors.*))
-		__CTOR_END__ = .;
-	} > ram :ph_ram
-
-	/* C++ destructor section. */
-	.dtors :
-	{
-		__DTOR_LIST__ = .;
-		*(.dtors)
-		*(SORT(.dtors.*))
-		__DTOR_END__ = .;
-	} > ram :ph_ram
-
-
-	/* Small data section. */
-	.sdata ALIGN(0x10) :
-	{
-		PROVIDE (_SDA_BASE_ = .);
-		*(.sdata)
-		*(.sdata.*)
-		*(.gnu.linkonce.s.*)
-	} > ram :ph_ram
-
-	/*-------------------------------------------------------------------
-	 * Uninitialized data (cleared by PM)
-	 *-----------------------------------------------------------------*/
-
-	/* Small bss section. */
-	.sbss :
-	{
-		*(.sbss)
-		*(.sbss.*)
-		*(.scommon)
-		*(.gnu.linkonce.sb.*)
-	} > ram :ph_ram
-
-	/* Bss section. */
-	.bss :
-	{
-		*(.bss)
-		*(.bss.*)
-		*(COMMON)
-		*(.gnu.linkonce.b.*)
-	} > ram :ph_ram
-
-/*---------------------------------------------------------------------------
- *		Debug information
- *-------------------------------------------------------------------------*/
-
-	/*
-	 * Stabs debug sections.
-	 */
-
-	.stab            0 : { *(.stab) }
-	.stabstr         0 : { *(.stabstr) }
-	.stab.excl       0 : { *(.stab.excl) }
-	.stab.exclstr    0 : { *(.stab.exclstr) }
-	.stab.index      0 : { *(.stab.index) }
-	.stab.indexstr   0 : { *(.stab.indexstr) }
-	.comment         0 : { *(.comment) }
-
-	/*
-	 * DWARF debug sections.
-	 */
-
-	/* DWARF 1 */
-	.debug           0 : { *(.debug) }
-	.line            0 : { *(.line) }
-	/* GNU DWARF 1 extensions */
-	.debug_srcinfo   0 : { *(.debug_srcinfo) }
-	.debug_sfnames   0 : { *(.debug_sfnames) }
-	/* DWARF 1.1 and DWARF 2 */
-	.debug_aranges   0 : { *(.debug_aranges) }
-	.debug_pubnames  0 : { *(.debug_pubnames) }
-	/* DWARF 2 */
-	.debug_info      0 : { *(.debug_info) *(.gnu.linkonce.wi.*) }
-	.debug_abbrev    0 : { *(.debug_abbrev) }
-	.debug_line      0 : { *(.debug_line) }
-	.debug_frame     0 : { *(.debug_frame) }
-	.debug_str       0 : { *(.debug_str) }
-	.debug_loc       0 : { *(.debug_loc) }
-	.debug_macinfo   0 : { *(.debug_macinfo) }
-	/* SGI/MIPS DWARF 2 extensions */
-	.debug_weaknames 0 : { *(.debug_weaknames) }
-	.debug_funcnames 0 : { *(.debug_funcnames) }
-	.debug_typenames 0 : { *(.debug_typenames) }
-	.debug_varnames  0 : { *(.debug_varnames) }
-}
diff --git a/erts/emulator/sys/ose/sys.c b/erts/emulator/sys/ose/sys.c
deleted file mode 100644
index bcd0ffa0b6..0000000000
--- a/erts/emulator/sys/ose/sys.c
+++ /dev/null
@@ -1,1847 +0,0 @@
-/*
- * %CopyrightBegin%
- *
- * Copyright Ericsson AB 1996-2013. All Rights Reserved.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- *
- * %CopyrightEnd%
- */
-
-#ifdef HAVE_CONFIG_H
-#  include "config.h"
-#endif
-#include "sys/time.h"
-#include "time.h"
-#include "sys/uio.h"
-#include "termios.h"
-#include "ctype.h"
-#include "termios.h"
-
-#ifdef HAVE_FCNTL_H
-#include "fcntl.h"
-#endif
-
-#ifdef HAVE_SYS_IOCTL_H
-#include "sys/ioctl.h"
-#endif
-
-#define ERTS_WANT_BREAK_HANDLING
-#define WANT_NONBLOCKING
-#include "sys.h"
-#include "erl_thr_progress.h"
-
-#ifdef USE_THREADS
-#include "erl_threads.h"
-#endif
-
-#include "erl_mseg.h"
-
-#include "unistd.h"
-#include "efs.h"
-#include "erl_printf.h"
-#include "aio.h"
-#include "pm.h"
-#include "fcntl.h"
-
-/* Set the define to 1 to get some logging */
-#if 0
-#include "ramlog.h"
-#define LOG(output) ramlog_printf output
-#else
-#define LOG(output)
-#endif
-
-extern char **environ;
-static erts_smp_rwmtx_t environ_rwmtx;
-static PROCESS sig_proxy_pid = 0;
-
-#define MAX_VSIZE 16		/* Max number of entries allowed in an I/O
-				 * vector sock_sendv().
-				 */
-/*
- * Don't need global.h, but bif_table.h (included by bif.h),
- * won't compile otherwise
- */
-#include "global.h"
-#include "bif.h"
-
-#include "erl_sys_driver.h"
-#include "erl_check_io.h"
-#include "erl_cpu_topology.h"
-
-/* The priority for reader/writer processes */
-#define FD_PROC_PRI get_pri(current_process())
-
-typedef struct ErtsSysReportExit_ ErtsSysReportExit;
-struct ErtsSysReportExit_ {
-    ErtsSysReportExit  *next;
-    Eterm              port;
-    int                pid;
-    int                ifd;
-    int                ofd;
-    ErlDrvEvent        attach_event;
-    ErlDrvEvent        input_event;
-    ErlDrvEvent        output_event;
-};
-
-/* This data is shared by these drivers - initialized by spawn_init() */
-static struct driver_data {
-   ErlDrvPort          port_num;
-   int                 ofd;
-   int                 ifd;
-   int                 packet_bytes;
-   ErtsSysReportExit   *report_exit;
-   int                 pid;
-   int                 alive;
-   int                 status;
-   ErlDrvEvent         input_event;
-   ErlDrvEvent         output_event;
-   struct aiocb        aiocb;
-   FmHandle            handle;
-   char                *install_handle;
-} *driver_data;			/* indexed by fd */
-
-struct async {
-  SIGSELECT            signo;
-  ErlDrvTermData       port;
-  ErlDrvTermData       proc;
-  PROCESS              spid;
-  PROCESS              target;
-  Uint32               ref;
-};
-
-static ErtsSysReportExit *report_exit_list;
-static ERTS_INLINE void report_exit_status(ErtsSysReportExit *rep, int status);
-
-extern int  driver_interrupt(int, int);
-extern void do_break(void);
-
-extern void erl_sys_args(int*, char**);
-
-/* The following two defs should probably be moved somewhere else */
-
-extern void erts_sys_init_float(void);
-
-extern void erl_crash_dump(char* file, int line, char* fmt, ...);
-
-#define DIR_SEPARATOR_CHAR    '/'
-
-#if defined(DEBUG)
-#define ERL_BUILD_TYPE_MARKER ".debug"
-#else /* opt */
-#define ERL_BUILD_TYPE_MARKER
-#endif
-
-#define CHILD_SETUP_PROG_NAME	"child_setup" ERL_BUILD_TYPE_MARKER
-
-#ifdef DEBUG
-static int debug_log = 0;
-#endif
-
-#ifdef ERTS_SMP
-static erts_smp_atomic32_t have_prepared_crash_dump;
-#define ERTS_PREPARED_CRASH_DUMP \
-  ((int) erts_smp_atomic32_xchg_nob(&have_prepared_crash_dump, 1))
-#else
-static volatile int have_prepared_crash_dump;
-#define ERTS_PREPARED_CRASH_DUMP \
-  (have_prepared_crash_dump++)
-#endif
-
-static erts_smp_atomic_t sys_misc_mem_sz;
-
-#if defined(ERTS_SMP)
-erts_mtx_t chld_stat_mtx;
-#endif
-
-#if defined(ERTS_SMP) /* ------------------------------------------------- */
-#define CHLD_STAT_LOCK		erts_mtx_lock(&chld_stat_mtx)
-#define CHLD_STAT_UNLOCK	erts_mtx_unlock(&chld_stat_mtx)
-
-#else /* ------------------------------------------------------------------- */
-#define CHLD_STAT_LOCK
-#define CHLD_STAT_UNLOCK
-static volatile int children_died;
-#endif
-
-#define SET_AIO(REQ,FD,SIZE,BUFF)                                       \
-   memset(&(REQ),0,sizeof(REQ));                                        \
-   (REQ).aio_fildes = FD;                                               \
-   (REQ).aio_offset = FM_POSITION_CURRENT;                              \
-   (REQ).aio_nbytes = SIZE;                                             \
-   (REQ).aio_buf = BUFF;                                                \
-   (REQ).aio_sigevent.sigev_notify = SIGEV_NONE
-
-/* the first sizeof(struct aiocb *) bytes of the write buffer
- * will contain the pointer to the aiocb struct, this needs
- * to be freed between asynchronous writes.
- * A write of 0 bytes is ignored. */
-#define WRITE_AIO(FD,SIZE,BUFF) do {                                    \
-   if (SIZE > 0) {                                                      \
-      struct aiocb *write_req = driver_alloc(sizeof(struct aiocb));     \
-      char *write_buff = driver_alloc((sizeof(char)*SIZE)+1+            \
-            (sizeof(struct aiocb *)));                                  \
-      *(struct aiocb **)write_buff = (struct aiocb *)write_req;         \
-      write_buff += sizeof(struct aiocb *);                             \
-      memcpy(write_buff,BUFF,SIZE+1);                                   \
-      SET_AIO(*write_req,FD,SIZE,write_buff);                           \
-      if (aio_write(write_req))						\
-	ramlog_printf("%s:%d: write failed with %d\n",			\
-		      __FILE__,__LINE__,errno);				\
-   }                                                                    \
-} while(0)
-
-/* free the write_buffer and write_req
- * created in the WRITE_AIO() request macro */
-#define FREE_AIO(ptr) do {                                              \
-   struct aiocb *aiocb_ptr;                                             \
-   char *buffer_ptr;                                                    \
-   aiocb_ptr = *(struct aiocb **)((ptr)-sizeof(struct aiocb *));        \
-   buffer_ptr = (((char*)ptr)-sizeof(struct aiocb *));			\
-   driver_free(aiocb_ptr);                                              \
-   driver_free(buffer_ptr);                                             \
-} while(0)
-
-#define DISPATCH_AIO(sig) do {                                          \
-   if (aio_dispatch(sig))						\
-     ramlog_printf("%s:%d: dispatch failed with %d\n",			\
-		   __FILE__,__LINE__,errno);				\
-   } while(0)
-
-#define AIO_PIPE_SIZE 1024
-
-/* debug print macros */
-#define DEBUG_RES 0
-
-#ifdef DEBUG_RES
-#define DEBUG_CHECK_RES(actual, expected) \
-   do { \
-      if (actual != expected ) { \
-         ramlog_printf("Result check failed" \
-                       " got: 0x%08x expected:0x%08x\nat: %s:%d\n", \
-               actual, expected, __FILE__, __LINE__); \
-         abort();  /* This might perhaps be too harsh? */ \
-      } \
-   } while(0)
-#else
-#define DEBUG_CHECK_RES
-#endif
-
-static struct fd_data {
-    char  pbuf[4];   /* hold partial packet bytes */
-    int   psz;       /* size of pbuf */
-    char  *buf;
-    char  *cpos;
-    int   sz;
-    int   remain;  /* for input on fd */
-} *fd_data;			/* indexed by fd */
-
-/********************* General functions ****************************/
-
-/* This is used by both the drivers and general I/O, must be set early */
-static int max_files = -1;
-
-/*
- * a few variables used by the break handler
- */
-#ifdef ERTS_SMP
-erts_smp_atomic32_t erts_break_requested;
-#define ERTS_SET_BREAK_REQUESTED \
-  erts_smp_atomic32_set_nob(&erts_break_requested, (erts_aint32_t) 1)
-#define ERTS_UNSET_BREAK_REQUESTED \
-  erts_smp_atomic32_set_nob(&erts_break_requested, (erts_aint32_t) 0)
-#else
-volatile int erts_break_requested = 0;
-#define ERTS_SET_BREAK_REQUESTED (erts_break_requested = 1)
-#define ERTS_UNSET_BREAK_REQUESTED (erts_break_requested = 0)
-#endif
-/* set early so the break handler has access to initial mode */
-static struct termios initial_tty_mode;
-static int replace_intr = 0;
-/* assume yes initially, ttsl_init will clear it */
-int using_oldshell = 1;
-static PROCESS get_signal_proxy_pid(void);
-
-static void
-init_check_io(void)
-{
-    erts_init_check_io();
-    max_files = erts_check_io_max_files();
-}
-
-#ifdef ERTS_POLL_NEED_ASYNC_INTERRUPT_SUPPORT
-#define ERTS_CHK_IO_AS_INTR()	erts_check_io_async_sig_interrupt()
-#else
-#define ERTS_CHK_IO_AS_INTR()	erts_check_io_interrupt(1)
-#endif
-#define ERTS_CHK_IO_INTR	erts_check_io_interrupt
-#define ERTS_CHK_IO_INTR_TMD	erts_check_io_interrupt_timed
-#define ERTS_CHK_IO		erts_check_io
-#define ERTS_CHK_IO_SZ		erts_check_io_size
-
-
-void
-erts_sys_schedule_interrupt(int set)
-{
-    ERTS_CHK_IO_INTR(set);
-}
-
-#ifdef ERTS_SMP
-void
-erts_sys_schedule_interrupt_timed(int set, ErtsMonotonicTime timeout_time)
-{
-    ERTS_CHK_IO_INTR_TMD(set, timeout_time);
-}
-#endif
-
-Uint
-erts_sys_misc_mem_sz(void)
-{
-    Uint res = ERTS_CHK_IO_SZ();
-    res += erts_smp_atomic_read_mb(&sys_misc_mem_sz);
-    return res;
-}
-
-/*
- * reset the terminal to the original settings on exit
- */
-void sys_tty_reset(int exit_code)
-{
-  if (using_oldshell && !replace_intr) {
-    SET_BLOCKING(0);
-  }
-  else if (isatty(0)) {
-    tcsetattr(0,TCSANOW,&initial_tty_mode);
-  }
-}
-
-#ifdef USE_THREADS
-
-typedef struct {
-    int sched_bind_data;
-} erts_thr_create_data_t;
-
-/*
- * thr_create_prepare() is called in parent thread before thread creation.
- * Returned value is passed as argument to thr_create_cleanup().
- */
-static void *
-thr_create_prepare(void)
-{
-    erts_thr_create_data_t *tcdp;
-
-    tcdp = erts_alloc(ERTS_ALC_T_TMP, sizeof(erts_thr_create_data_t));
-
-    tcdp->sched_bind_data = erts_sched_bind_atthrcreate_prepare();
-
-    return (void *) tcdp;
-}
-
-
-/* thr_create_cleanup() is called in parent thread after thread creation. */
-static void
-thr_create_cleanup(void *vtcdp)
-{
-    erts_thr_create_data_t *tcdp = (erts_thr_create_data_t *) vtcdp;
-
-    erts_sched_bind_atthrcreate_parent(tcdp->sched_bind_data);
-
-    erts_free(ERTS_ALC_T_TMP, tcdp);
-}
-
-static void
-thr_create_prepare_child(void *vtcdp)
-{
-    erts_thr_create_data_t *tcdp = (erts_thr_create_data_t *) vtcdp;
-
-#ifdef ERTS_ENABLE_LOCK_COUNT
-    erts_lcnt_thread_setup();
-#endif
-
-    erts_sched_bind_atthrcreate_child(tcdp->sched_bind_data);
-}
-
-#endif /* #ifdef USE_THREADS */
-
-/* The two functions below are stolen from win_con.c
-   They have to use malloc/free/realloc directly becasue
-   we want to do able to do erts_printf very early on.
- */
-#define VPRINTF_BUF_INC_SIZE 128
-static erts_dsprintf_buf_t *
-grow_vprintf_buf(erts_dsprintf_buf_t *dsbufp, size_t need)
-{
-    char *buf;
-    size_t size;
-
-    ASSERT(dsbufp);
-
-    if (!dsbufp->str) {
-	size = (((need + VPRINTF_BUF_INC_SIZE - 1)
-		 / VPRINTF_BUF_INC_SIZE)
-		* VPRINTF_BUF_INC_SIZE);
-	buf = (char *) malloc(size * sizeof(char));
-    }
-    else {
-	size_t free_size = dsbufp->size - dsbufp->str_len;
-
-	if (need <= free_size)
-	    return dsbufp;
-
-	size = need - free_size + VPRINTF_BUF_INC_SIZE;
-	size = (((size + VPRINTF_BUF_INC_SIZE - 1)
-		 / VPRINTF_BUF_INC_SIZE)
-		* VPRINTF_BUF_INC_SIZE);
-	size += dsbufp->size;
-        buf = (char *) realloc((void *) dsbufp->str,
-			       size * sizeof(char));
-    }
-    if (!buf)
-	return NULL;
-    if (buf != dsbufp->str)
-	dsbufp->str = buf;
-    dsbufp->size = size;
-    return dsbufp;
-}
-
-static int erts_sys_ramlog_printf(char *format, va_list arg_list)
-{
-    int res,i;
-    erts_dsprintf_buf_t dsbuf = ERTS_DSPRINTF_BUF_INITER(grow_vprintf_buf);
-    res = erts_vdsprintf(&dsbuf, format, arg_list);
-    if (res >= 0) {
-      for (i = 0; i < dsbuf.str_len; i+= 50)
-	/* We print 50 characters at a time because otherwise
-	   the ramlog looks broken */
-        ramlog_printf("%.*s",dsbuf.str_len-50 < 0?dsbuf.str_len:50,dsbuf.str+i);
-    }
-    if (dsbuf.str)
-      free((void *) dsbuf.str);
-    return res;
-}
-
-void
-erts_sys_pre_init(void)
-{
-    erts_printf_add_cr_to_stdout = 1;
-    erts_printf_add_cr_to_stderr = 1;
-#ifdef USE_THREADS
-    {
-    erts_thr_init_data_t eid = ERTS_THR_INIT_DATA_DEF_INITER;
-
-    eid.thread_create_child_func = thr_create_prepare_child;
-    /* Before creation in parent */
-    eid.thread_create_prepare_func = thr_create_prepare;
-    /* After creation in parent */
-    eid.thread_create_parent_func = thr_create_cleanup,
-
-    erts_thr_init(&eid);
-
-    report_exit_list = NULL;
-
-#ifdef ERTS_ENABLE_LOCK_COUNT
-    erts_lcnt_init();
-#endif
-
-#if defined(ERTS_SMP)
-    erts_mtx_init(&chld_stat_mtx, "child_status");
-#endif
-    }
-#ifdef ERTS_SMP
-    erts_smp_atomic32_init_nob(&erts_break_requested, 0);
-    erts_smp_atomic32_init_nob(&have_prepared_crash_dump, 0);
-#else
-    erts_break_requested = 0;
-    have_prepared_crash_dump = 0;
-#endif
-#if !defined(ERTS_SMP)
-    children_died = 0;
-#endif
-#endif /* USE_THREADS */
-
-    erts_printf_stdout_func = erts_sys_ramlog_printf;
-
-    erts_smp_atomic_init_nob(&sys_misc_mem_sz, 0);
-}
-
-void
-erl_sys_init(void)
-{
-
-#ifdef USE_SETLINEBUF
-    setlinebuf(stdout);
-#else
-    setvbuf(stdout, (char *)NULL, _IOLBF, BUFSIZ);
-#endif
-
-    erts_sys_init_float();
-
-    /* we save this so the break handler can set and reset it properly */
-    /* also so that we can reset on exit (break handler or not) */
-    if (isatty(0)) {
-	tcgetattr(0,&initial_tty_mode);
-    }
-    tzset(); /* Required at least for NetBSD with localtime_r() */
-}
-
-static ERTS_INLINE int
-prepare_crash_dump(int secs)
-{
-#define NUFBUF (3)
-    int i, max;
-    char env[21]; /* enough to hold any 64-bit integer */
-    size_t envsz;
-    /*DeclareTmpHeapNoproc(heap,NUFBUF);*/
-    /*Eterm *hp = heap;*/
-    /*Eterm list = NIL;*/
-    int has_heart = 0;
-
-    UseTmpHeapNoproc(NUFBUF);
-
-    if (ERTS_PREPARED_CRASH_DUMP)
-	return 0; /* We have already been called */
-
-
-    /* Positive secs means an alarm must be set
-     * 0 or negative means no alarm
-     *
-     * Set alarm before we try to write to a port
-     * we don't want to hang on a port write with
-     * no alarm.
-     *
-     */
-
-#if 0 /*ose TBD!!!*/
-    if (secs >= 0) {
-	alarm((unsigned int)secs);
-    }
-#endif
-
-    /* Make sure we unregister at epmd (unknown fd) and get at least
-       one free filedescriptor (for erl_crash.dump) */
-
-    max = max_files;
-    if (max < 1024)
-	max = 1024;
-    for (i = 3; i < max; i++) {
-	close(i);
-    }
-
-    envsz = sizeof(env);
-    i = erts_sys_getenv__("ERL_CRASH_DUMP_NICE", env, &envsz);
-    if (i >= 0) {
-	int nice_val;
-	nice_val = i != 0 ? 0 : atoi(env);
-	if (nice_val > 39) {
-	    nice_val = 39;
-	}
-	set_pri(nice_val);
-    }
-
-    UnUseTmpHeapNoproc(NUFBUF);
-#undef NUFBUF
-    return has_heart;
-}
-
-int erts_sys_prepare_crash_dump(int secs)
-{
-    return prepare_crash_dump(secs);
-}
-
-static ERTS_INLINE void
-break_requested(void)
-{
-  /*
-   * just set a flag - checked for and handled by
-   * scheduler threads erts_check_io() (not signal handler).
-   */
-#ifdef DEBUG
-  fprintf(stderr,"break!\n");
-#endif
-  if (ERTS_BREAK_REQUESTED)
-      erl_exit(ERTS_INTR_EXIT, "");
-
-  ERTS_SET_BREAK_REQUESTED;
-  ERTS_CHK_IO_AS_INTR(); /* Make sure we don't sleep in poll */
-}
-
-/* Disable break */
-void erts_set_ignore_break(void) {
-
-}
-
-/* Don't use ctrl-c for break handler but let it be
-   used by the shell instead (see user_drv.erl) */
-void erts_replace_intr(void) {
-  struct termios mode;
-
-  if (isatty(0)) {
-    tcgetattr(0, &mode);
-
-    /* here's an example of how to replace ctrl-c with ctrl-u */
-    /* mode.c_cc[VKILL] = 0;
-       mode.c_cc[VINTR] = CKILL; */
-
-    mode.c_cc[VINTR] = 0;	/* disable ctrl-c */
-    tcsetattr(0, TCSANOW, &mode);
-    replace_intr = 1;
-  }
-}
-
-void init_break_handler(void)
-{
-
-}
-
-int sys_max_files(void)
-{
-   return(max_files);
-}
-
-
-/************************** OS info *******************************/
-
-/* Used by erlang:info/1. */
-/* (This code was formerly in drv.XXX/XXX_os_drv.c) */
-
-char os_type[] = "ose";
-
-void
-os_flavor(char* namebuf,	/* Where to return the name. */
-	  unsigned size)	/* Size of name buffer. */
-{
-#if 0
-    struct utsname uts;		/* Information about the system. */
-    char* s;
-
-    (void) uname(&uts);
-    for (s = uts.sysname; *s; s++) {
-	if (isupper((int) *s)) {
-	    *s = tolower((int) *s);
-	}
-    }
-    strcpy(namebuf, uts.sysname);
-#else
-    strncpy(namebuf, "release", size);
-#endif
-}
-
-void
-os_version(pMajor, pMinor, pBuild)
-int* pMajor;			/* Pointer to major version. */
-int* pMinor;			/* Pointer to minor version. */
-int* pBuild;			/* Pointer to build number. */
-{
-    *pMajor = 5;
-    *pMinor = 7;
-    *pBuild = 0;
-}
-
-void init_getenv_state(GETENV_STATE *state)
-{
-   erts_smp_rwmtx_rlock(&environ_rwmtx);
-   *state = NULL;
-}
-
-char **environ; /*ose - needs replacement*/
-
-char *getenv_string(GETENV_STATE *state0)
-{
-   char **state = (char **) *state0;
-   char *cp;
-
-   ERTS_SMP_LC_ASSERT(erts_smp_lc_rwmtx_is_rlocked(&environ_rwmtx));
-
-   if (state == NULL)
-      state = environ;
-
-   cp = *state++;
-   *state0 = (GETENV_STATE) state;
-
-   return cp;
-}
-
-void fini_getenv_state(GETENV_STATE *state)
-{
-   *state = NULL;
-   erts_smp_rwmtx_runlock(&environ_rwmtx);
-}
-
-
-/************************** Port I/O *******************************/
-
-/* I. Common stuff */
-
-union SIGNAL {
-    SIGSELECT sig_no;
-    struct FmReadPtr fm_read_reply;
-    struct FmWritePtr fm_write_reply;
-    struct async async;
-};
-
-/* II. The spawn/fd drivers */
-
-/*
- * Decreasing the size of it below 16384 is not allowed.
- */
-#define ERTS_SYS_READ_BUF_SZ (64*1024)
-
-/* Driver interfaces */
-static ErlDrvData spawn_start(ErlDrvPort, char*, SysDriverOpts*);
-static ErlDrvData fd_start(ErlDrvPort, char*, SysDriverOpts*);
-static ErlDrvSSizeT fd_control(ErlDrvData, unsigned int, char *, ErlDrvSizeT,
-			       char **, ErlDrvSizeT);
-static int spawn_init(void);
-static void fd_stop(ErlDrvData);
-static void erl_stop(ErlDrvData);
-static void ready_input(ErlDrvData, ErlDrvEvent);
-static void ready_output(ErlDrvData, ErlDrvEvent);
-static void output(ErlDrvData, char*, ErlDrvSizeT);
-static void stop_select(ErlDrvEvent, void*);
-
-static PROCESS
-get_signal_proxy_pid(void) {
-   union SIGNAL *sig;
-   SIGSELECT any_sig[] = {1,ERTS_SIGNAL_OSE_DRV_ATTACH};
-
-   if (!sig_proxy_pid) {
-      sig = alloc(sizeof(union SIGNAL), ERTS_SIGNAL_OSE_DRV_ATTACH);
-      hunt("ose_signal_driver_proxy", 0, NULL, &sig);
-      sig = receive(any_sig);
-      sig_proxy_pid = sender(&sig);
-      free_buf(&sig);
-   }
-   ASSERT(sig_proxy_pid);
-   return sig_proxy_pid;
-}
-
-static ErlDrvOseEventId
-resolve_signal(union SIGNAL* sig) {
-   switch(sig->sig_no) {
-
-      case FM_READ_PTR_REPLY:
-            return (ErlDrvOseEventId)sig->fm_read_reply.handle;
-
-       case FM_WRITE_PTR_REPLY:
-            return (ErlDrvOseEventId)sig->fm_write_reply.handle;
-
-       case ERTS_SIGNAL_OSE_DRV_ATTACH:
-            return (ErlDrvOseEventId)sig->async.target;
-
-       default:
-            break;
-    }
-    return (ErlDrvOseEventId)-1;
-}
-
-struct erl_drv_entry spawn_driver_entry = {
-    spawn_init,
-    spawn_start,
-    NULL, /* erl_stop, */
-    output,
-    ready_input,
-    ready_output,
-    "spawn",
-    NULL,
-    NULL,
-    NULL,
-    NULL,
-    NULL,
-    NULL,
-    NULL,
-    NULL,
-    NULL,
-    ERL_DRV_EXTENDED_MARKER,
-    ERL_DRV_EXTENDED_MAJOR_VERSION,
-    ERL_DRV_EXTENDED_MINOR_VERSION,
-    ERL_DRV_FLAG_USE_PORT_LOCKING,
-    NULL, NULL,
-    stop_select
-};
-struct erl_drv_entry fd_driver_entry = {
-    NULL,
-    fd_start,
-    fd_stop,
-    output,
-    ready_input,
-    ready_output,
-    "fd",
-    NULL,
-    NULL,
-    fd_control,
-    NULL,
-    NULL,
-    NULL, /* ready_async */
-    NULL, /* flush */
-    NULL, /* call */
-    NULL, /* event */
-    ERL_DRV_EXTENDED_MARKER,
-    ERL_DRV_EXTENDED_MAJOR_VERSION,
-    ERL_DRV_EXTENDED_MINOR_VERSION,
-    0, /* ERL_DRV_FLAGs */
-    NULL, /* handle2 */
-    NULL, /* process_exit */
-    stop_select
-};
-
-static void
-set_spawn_fd(int local_fd, int remote_fd, PROCESS remote_pid) {
-   PROCESS vm_pid;
-   FmHandle handle;
-   char env_val[55];
-   char env_name[10];
-   EfsStatus efs_res;
-
-   /* get pid of pipevm and handle of chosen fd */
-   efs_res = efs_examine_fd(local_fd, FLIB_FD_VMPID, &vm_pid, 0);
-   DEBUG_CHECK_RES(efs_res, EFS_SUCCESS);
-
-   /* setup the file descriptor to buffer per line */
-   efs_res = efs_config_fd(local_fd, FLIB_FD_BUFMODE, FM_BUFF_LINE,
-                    FLIB_FD_BUFSIZE, 80, 0);
-   DEBUG_CHECK_RES(efs_res, EFS_SUCCESS);
-
-   /* duplicate handle  and set spawn pid owner */
-   efs_res = efs_dup_to(local_fd, remote_pid, &handle);
-   DEBUG_CHECK_RES(efs_res, EFS_SUCCESS);
-
-   sprintf(env_name, "FD%d", remote_fd);
-
-   /* Syntax of the environment variable:
-    * "FD#" "<pid of pipevm>,<handle>,<buffer mode>,<buff size>,<omode>" */
-   sprintf(env_val, "0x%lx,0x%lx,%lu,%lu,0x%x",
-                    vm_pid, handle,
-                    FM_BUFF_LINE, 80,
-                    O_APPEND);
-
-   set_env(remote_pid, env_name, env_val);
-}
-
-static ErlDrvData
-set_driver_data(ErlDrvPort port_num,
-			   int ifd,
-			   int ofd,
-			   int packet_bytes,
-			   int read_write,
-			   int exit_status,
-			   PROCESS pid)
-{
-    Port *prt;
-    ErtsSysReportExit *report_exit;
-
-    prt = erts_drvport2port(port_num);
-    if (prt != ERTS_INVALID_ERL_DRV_PORT) {
-       prt->os_pid = pid;
-    }
-
-    /* READ */
-    if (read_write & DO_READ) {
-       EfsStatus res = efs_examine_fd(ifd, FLIB_FD_HANDLE,
-				      &driver_data[ifd].handle, 0);
-       if (res != EFS_SUCCESS)
-	 ramlog_printf("%s:%d: efs_examine_fd(%d) failed with %d\n",
-		       __FILE__,__LINE__,ifd,errno);
-       driver_data[ifd].ifd = ifd;
-       driver_data[ifd].packet_bytes = packet_bytes;
-       driver_data[ifd].port_num = port_num;
-       driver_data[ifd].pid = pid;
-
-       /* async read struct */
-       memset(&driver_data[ifd].aiocb, 0, sizeof(struct aiocb));
-       driver_data[ifd].aiocb.aio_buf = driver_alloc(AIO_PIPE_SIZE);
-       driver_data[ifd].aiocb.aio_fildes = ifd;
-       driver_data[ifd].aiocb.aio_nbytes = (packet_bytes?packet_bytes:AIO_PIPE_SIZE);
-       driver_data[ifd].alive = 1;
-       driver_data[ifd].status = 0;
-       driver_data[ifd].input_event =
-          erl_drv_ose_event_alloc(FM_READ_PTR_REPLY,
-                driver_data[ifd].handle, resolve_signal,
-                &driver_data[ifd].ifd);
-
-       /* READ & WRITE */
-       if (read_write & DO_WRITE) {
-          driver_data[ifd].ofd = ofd;
-          efs_examine_fd(ofd, FLIB_FD_HANDLE, &driver_data[ofd].handle, 0);
-
-          driver_data[ifd].output_event =
-             erl_drv_ose_event_alloc(FM_WRITE_PTR_REPLY,
-                   driver_data[ofd].handle, resolve_signal,
-                   &driver_data[ofd].ofd);
-          driver_data[ofd].pid = pid;
-          if (ifd != ofd) {
-             driver_data[ofd] = driver_data[ifd];
-             driver_data[ofd].aiocb.aio_buf = NULL;
-           }
-       }
-       else { /* READ ONLY */
-          driver_data[ifd].ofd = -1;
-       }
-
-       /* enable input event */
-       (void) driver_select(port_num, driver_data[ifd].input_event,
-			     (ERL_DRV_READ | ERL_DRV_USE), 1);
-
-       if (aio_read(&driver_data[ifd].aiocb))
-	 ramlog_printf("%s:%d: aio_read(%d) failed with %d\n",
-		       __FILE__,__LINE__,ifd,errno);
-    }
-    else { /* WRITE ONLY */
-       efs_examine_fd(ofd, FLIB_FD_HANDLE, &driver_data[ofd].handle, 0);
-       driver_data[ofd].packet_bytes = packet_bytes;
-       driver_data[ofd].port_num = port_num;
-       driver_data[ofd].ofd = ofd;
-       driver_data[ofd].pid = pid;
-       driver_data[ofd].alive = 1;
-       driver_data[ofd].status = 0;
-       driver_data[ofd].output_event =
-          erl_drv_ose_event_alloc(FM_WRITE_PTR_REPLY, driver_data[ofd].handle,
-				    resolve_signal, &driver_data[ofd].ofd);
-       driver_data[ofd].input_event = driver_data[ofd].output_event;
-    }
-
-    /* this is used for spawned load modules, and is needed
-     * to properly uninstall them */
-    if (exit_status) {
-       struct PmProgramInfo *info;
-       int install_handle_size;
-       union SIGNAL *sig;
-       PmStatus pm_status;
-       report_exit = erts_alloc(ERTS_ALC_T_PRT_REP_EXIT,
-				 sizeof(ErtsSysReportExit));
-       report_exit->next = report_exit_list;
-       report_exit->port = erts_drvport2id(port_num);
-       report_exit->pid = pid;
-       report_exit->ifd = (read_write & DO_READ) ? ifd : -1;
-       report_exit->ofd = (read_write & DO_WRITE) ? ofd : -1;
-       report_exit_list = report_exit;
-       report_exit->attach_event =
-          erl_drv_ose_event_alloc(ERTS_SIGNAL_OSE_DRV_ATTACH, pid,
-				resolve_signal, &driver_data[ifd].ifd);
-
-       /* setup ifd and ofd report exit */
-       driver_data[ifd].report_exit = report_exit;
-       driver_data[ofd].report_exit = report_exit;
-
-       pm_status = ose_pm_program_info(pid, &info);
-       DEBUG_CHECK_RES(pm_status, PM_SUCCESS);
-
-       install_handle_size = strlen(info->install_handle)+1;
-       driver_data[ifd].install_handle = driver_alloc(install_handle_size);
-       strcpy(driver_data[ifd].install_handle,
-             info->install_handle);
-
-       free_buf((union SIGNAL **)&info);
-
-       sig = alloc(sizeof(struct async), ERTS_SIGNAL_OSE_DRV_ATTACH);
-       sig->async.target = pid;
-       send(&sig, get_signal_proxy_pid());
-
-       /* this event will trigger when we receive an attach signal
-        * from the recently dead load module */
-       (void)driver_select(port_num,report_exit->attach_event, DO_READ, 1);
-    }
-    else {
-       report_exit = NULL;
-    }
-
-    /* the return value is the pointer to the driver_data struct we created
-     * in this function, it will be used in the drivers input
-     * and output functions */
-    return (ErlDrvData)((!(read_write & DO_READ) && read_write & DO_WRITE)
-          ? &driver_data[ofd]
-          : &driver_data[ifd]);
-}
-
-static int spawn_init()
-{
-    int i;
-
-    driver_data = (struct driver_data *)
-      erts_alloc(ERTS_ALC_T_DRV_TAB, max_files * sizeof(struct driver_data));
-    erts_smp_atomic_add_nob(&sys_misc_mem_sz,
-			    max_files * sizeof(struct driver_data));
-
-    for (i = 0; i < max_files; i++)
-        driver_data[i].pid = -1;
-
-   return 1;
-}
-
-static void
-init_fd_data(int fd, ErlDrvPort port_num)
-{
-    fd_data[fd].buf = NULL;
-    fd_data[fd].cpos = NULL;
-    fd_data[fd].remain = 0;
-    fd_data[fd].sz = 0;
-    fd_data[fd].psz = 0;
-}
-
-/* FIXME write a decent text on pipes on ose */
-static ErlDrvData
-spawn_start(ErlDrvPort port_num, char* name, SysDriverOpts* opts)
-{
-    int ifd[2];
-    int ofd[2];
-    static uint32_t ticker = 1;
-    PmStatus pm_status;
-    OSDOMAIN domain = PM_NEW_DOMAIN;
-    PROCESS progpid, mainbid, mainpid;
-    char *handle = NULL;
-    struct PmProgramInfo *info;
-    char *args = NULL;
-    char *tmp_handle;
-    ErlDrvData res = (ErlDrvData)-1;
-    int handle_size;
-    char *ptr;
-
-   
-    args = driver_alloc(strlen(name)+1);
-    strcpy(args, name);
-    /* We need to handle name in three parts 
-     * - install handle (must be unique)
-     * - install binary (needed for ose_pm_install_load_module())
-     * - full path (as argument to the spawned applications env.var
-     */
-
-    /* full path including arguments */
-    args = driver_alloc(strlen(name)+1);
-    strcpy(args, name);
-
-    /* handle path */
-    tmp_handle = strrchr(name, '/');
-    if (tmp_handle == NULL) {
-       tmp_handle = name;
-    }
-    else {
-       tmp_handle++;
-    }
-
-    /* handle args */
-    ptr = strchr(tmp_handle, ' ');
-    if (ptr != NULL) {
-       *ptr = '\0';
-       handle_size = ptr - tmp_handle;
-    }
-    else {
-       handle_size = strlen(name)+1;
-    }
-
-    /* make room for ticker */
-    handle_size += (ticker<10)?3:((ticker<100)?4:5);
-    handle = driver_alloc(handle_size);
-    
-    do {
-       snprintf(handle, handle_size, "%s_%d", tmp_handle, ticker);
-       pm_status = ose_pm_install_load_module(0, "ELF", name, handle,
-                                              0, 0, NULL);
-       ticker++;
-    } while (pm_status == PM_EINSTALL_HANDLE_ALREADY_INSTALLED);
-
-    if (pm_status != PM_SUCCESS) {
-       errno = ENOSYS; /* FIXME add comment */
-       return ERL_DRV_ERROR_ERRNO; 
-    }
-
-    /* Create Program */
-    pm_status = ose_pm_create_program(&domain, handle, 0, 0,
-                                      NULL, &progpid, &mainbid);
-    DEBUG_CHECK_RES(pm_status, PM_SUCCESS);
-
-    /* Get the mainpid from the newly created program */
-    pm_status = ose_pm_program_info(progpid, &info);
-    DEBUG_CHECK_RES(pm_status, PM_SUCCESS);
-
-    mainpid = info->main_process;
-    free_buf ((union SIGNAL **)&info);
-
-    /* pipevm needs to be started
-     * pipe will return 0 if success, -1 if not,
-     * errno will be set */
-    if (pipe(ifd) != 0 || pipe(ofd) != 0) {
-       DEBUG_CHECK_RES(0, -1);
-       ASSERT(0);
-    }
-
-    /* setup driver data */
-    res = set_driver_data(port_num, ofd[0], ifd[1], opts->packet_bytes,
-              opts->read_write, 1 /* opts->exit_status */, progpid);
-
-    /* init the fd_data array for read/write */
-    init_fd_data(ofd[0], port_num);
-    init_fd_data(ifd[1], port_num);
-
-    /* setup additional configurations
-     * for the spawned applications environment */
-    if (args != NULL) {
-       set_env(progpid, "ARGV", args);
-    }
-    set_env(mainbid, "EFS_RESOLVE_TMO", 0);
-    set_spawn_fd(ifd[0], 0, mainpid);
-    set_spawn_fd(ofd[1], 1, mainpid);
-    set_spawn_fd(ofd[1], 2, mainpid);
-
-    /* start the spawned program */
-    pm_status = ose_pm_start_program(mainbid);
-    DEBUG_CHECK_RES(pm_status, PM_SUCCESS);
-
-    /* close unused fd's */
-    close(ifd[0]);
-    close(ofd[1]);
-
-    if (handle) {
-       driver_free(handle);
-    }
-
-    return (ErlDrvData)res;
-}
-
-#define FD_DEF_HEIGHT 24
-#define FD_DEF_WIDTH 80
-/* Control op */
-#define FD_CTRL_OP_GET_WINSIZE 100
-
-static int fd_get_window_size(int fd, Uint32 *width, Uint32 *height)
-{
-#ifdef TIOCGWINSZ
-    struct winsize ws;
-    if (ioctl(fd,TIOCGWINSZ,&ws) == 0) {
-	*width = (Uint32) ws.ws_col;
-	*height = (Uint32) ws.ws_row;
-	return 0;
-    }
-#endif
-    return -1;
-}
-
-static ErlDrvSSizeT fd_control(ErlDrvData drv_data,
-			       unsigned int command,
-			       char *buf, ErlDrvSizeT len,
-			       char **rbuf, ErlDrvSizeT rlen)
-{
-   struct driver_data *data = (struct driver_data *)drv_data;
-    char resbuff[2*sizeof(Uint32)];
-    switch (command) {
-    case FD_CTRL_OP_GET_WINSIZE:
-	{
-	    Uint32 w,h;
-	    if (fd_get_window_size(data->ifd,&w,&h))
-		return 0;
-	    memcpy(resbuff,&w,sizeof(Uint32));
-	    memcpy(resbuff+sizeof(Uint32),&h,sizeof(Uint32));
-	}
-	break;
-    default:
-	return 0;
-    }
-    if (rlen < 2*sizeof(Uint32)) {
-	*rbuf = driver_alloc(2*sizeof(Uint32));
-    }
-    memcpy(*rbuf,resbuff,2*sizeof(Uint32));
-    return 2*sizeof(Uint32);
-}
-
-static ErlDrvData fd_start(ErlDrvPort port_num, char* name,
-			   SysDriverOpts* opts)
-{
-    ErlDrvData res;
-
-    CHLD_STAT_LOCK;
-    if (opts->read_write & DO_READ) {
-	init_fd_data(opts->ifd, port_num);
-    }
-    if (opts->read_write & DO_WRITE) {
-	init_fd_data(opts->ofd, port_num);
-    }
-    res = set_driver_data(port_num, opts->ifd, opts->ofd,
-				      opts->packet_bytes,
-				      opts->read_write, 0, -1);
-    CHLD_STAT_UNLOCK;
-    return res;
-}
-
-static void clear_fd_data(int fd)
-{
-    if (fd_data[fd].sz > 0) {
-	erts_free(ERTS_ALC_T_FD_ENTRY_BUF, (void *) fd_data[fd].buf);
-	ASSERT(erts_smp_atomic_read_nob(&sys_misc_mem_sz) >= fd_data[fd].sz);
-	erts_smp_atomic_add_nob(&sys_misc_mem_sz, -1*fd_data[fd].sz);
-    }
-    fd_data[fd].buf = NULL;
-    fd_data[fd].sz = 0;
-    fd_data[fd].remain = 0;
-    fd_data[fd].cpos = NULL;
-    fd_data[fd].psz = 0;
-}
-
-static void nbio_stop_fd(ErlDrvPort prt, ErlDrvEvent ev)
-{
-    int *fd;
-    driver_select(prt,ev,DO_READ|DO_WRITE,0);
-    erl_drv_ose_event_fetch(ev, NULL, NULL, (void **)&fd);
-    clear_fd_data(*fd);
-    SET_BLOCKING(*fd);
-}
-
-static void fd_stop(ErlDrvData drv_data)  /* Does not close the fds */
-{
-   struct driver_data *data = (struct driver_data *)drv_data;
-
-   if (data->ofd != -1) {
-      if (data->ifd != data->ofd) { /* read and write */
-         nbio_stop_fd(data->port_num, data->input_event);
-         nbio_stop_fd(data->port_num, data->output_event);
-      }
-      else { /* write only */
-         nbio_stop_fd(data->port_num, data->output_event);
-      }
-   }
-   else { /* read only */
-      nbio_stop_fd(data->port_num, data->input_event);
-   }
-}
-
-
-static void erl_stop(ErlDrvData drv_data)
-{
-   struct driver_data *data = (struct driver_data *)drv_data;
-
-   CHLD_STAT_LOCK;
-   data->pid = -1;
-   CHLD_STAT_UNLOCK;
-
-   if (data->ofd != -1) {
-      if (data->ifd != data->ofd) { /* read and write */
-         nbio_stop_fd(data->port_num, data->input_event);
-         nbio_stop_fd(data->port_num, data->output_event);
-      }
-      else { /* write only */
-         nbio_stop_fd(data->port_num, data->output_event);
-      }
-   }
-   else { /* read only */
-      nbio_stop_fd(data->port_num, data->input_event);
-   }
-   close(data->ifd);
-   close(data->ofd);
-}
-
-/* The parameter e is a pointer to the driver_data structure
- * related to the fd to be used as output */
-static void output(ErlDrvData drv_data, char* buf, ErlDrvSizeT len)
-{
-    ErlDrvSizeT sz;
-    char lb[4];
-    char* lbp;
-    struct driver_data *data = (struct driver_data *)drv_data;
-
-    if (((data->packet_bytes == 2) &&
-             (len > 0xffff)) || (data->packet_bytes == 1 && len > 0xff)) {
-	driver_failure_posix(data->port_num, EINVAL);
-	return; /* -1; */
-    }
-    put_int32(len, lb);
-    lbp = lb + (4-(data->packet_bytes));
-
-    if ((sz = driver_sizeq(data->port_num)) > 0) {
-       if (data->packet_bytes != 0) {
-          driver_enq(data->port_num, lbp, data->packet_bytes);
-       }
-       driver_enq(data->port_num, buf, len);
-
-       if (sz + len + data->packet_bytes >= (1 << 13))
-	    set_busy_port(data->port_num, 1);
-    }
-    else {
-       char *pbbuf;
-       if (data->packet_bytes != 0) {
-          pbbuf = malloc(len + data->packet_bytes);
-          int i;
-          for (i = 0; i < data->packet_bytes; i++) {
-             *pbbuf++ = *lbp++;
-          }
-          strncpy(pbbuf, buf, len);
-          pbbuf -= data->packet_bytes;
-       }
-       driver_select(data->port_num, data->output_event,
-		      ERL_DRV_WRITE|ERL_DRV_USE, 1);
-       WRITE_AIO(data->ofd, 
-             (data->packet_bytes ? len+data->packet_bytes : len), 
-             (data->packet_bytes ? pbbuf : buf));
-       if (data->packet_bytes != 0) free(pbbuf);
-    }
-    return; /* 0; */
-}
-
-/* This function is being run when we in recieve
- * either a read of 0 bytes, or the attach signal from a dying
- * spawned load module */
-static int port_inp_failure(ErlDrvPort port_num, ErlDrvEvent ready_fd, int res)
-				/* Result: 0 (eof) or -1 (error) */
-{
-    int *fd;
-    SIGSELECT sig_no;
-    ASSERT(res <= 0);
-
-    erl_drv_ose_event_fetch(ready_fd,&sig_no, NULL, (void **)&fd);
-    /* As we need to handle two signals, we do this in two steps */
-    if (driver_data[*fd].alive) {
-       report_exit_status(driver_data[*fd].report_exit, 0); /* status? */
-    }
-    else {
-       driver_select(port_num,ready_fd,DO_READ|DO_WRITE,0);
-       clear_fd_data(*fd);
-       driver_report_exit(driver_data[*fd].port_num, driver_data[*fd].status);
-       /* As we do not really know if the spawn has crashed or exited nicely
-        * we do not check the result status of the following call.. FIXME
-        * can we handle this in a better way? */
-       ose_pm_uninstall_load_module(driver_data[*fd].install_handle);
-       driver_free(driver_data[*fd].install_handle);
-       driver_free((void *)driver_data[*fd].aiocb.aio_buf);
-
-       close(*fd);
-    }
-
-    return 0;
-}
-
-/* The parameter e is a pointer to the driver_data structure
- * related to the fd to be used as output.
- * ready_fd is the event that triggered this call to ready_input */
-static void ready_input(ErlDrvData drv_data, ErlDrvEvent ready_fd)
-{
-    int res;
-    Uint h;
-    char *buf;
-    union SIGNAL *sig;
-    struct driver_data *data = (struct driver_data *)drv_data;
-
-    sig = erl_drv_ose_get_signal(ready_fd);
-    ASSERT(sig);
-
-
-   while (sig) {
-      /* If we've recieved an attach signal, we need to handle
-       * it in port_inp_failure */
-      if (sig->sig_no == ERTS_SIGNAL_OSE_DRV_ATTACH) {
-         port_inp_failure(data->port_num, ready_fd, 0);
-       }
-       else {
-          res = sig->fm_read_reply.actual;
-          if (res == 0) {
-             port_inp_failure(data->port_num, ready_fd, res);
-             break;
-          }
-
-          if (data->packet_bytes == 0) {
-             if (res < 0) {
-                if ((errno != EINTR) && (errno != ERRNO_BLOCK)) {
-                   port_inp_failure(data->port_num, ready_fd, res);
-                }
-             }
-             else if (res == 0) {
-                /* read of 0 bytes, eof, otherside of pipe is assumed dead */
-                port_inp_failure(data->port_num, ready_fd, res);
-                break;
-             }
-             else {
-                buf = driver_alloc(res);
-                memcpy(buf, (void *)data->aiocb.aio_buf, res);
-                driver_select(data->port_num, data->output_event,
-                      ERL_DRV_WRITE|ERL_DRV_USE, 1);
-                driver_output(data->port_num, (char*) buf, res);
-                driver_free(buf);
-             }
-                /* clear the previous read */
-                memset(data->aiocb.aio_buf, 0, res);
-
-                /* issue a new read */
-                DISPATCH_AIO(sig);
-                aio_read(&data->aiocb);
-          }
-          else if (data->packet_bytes && fd_data[data->ifd].remain > 0) {
-             /* we've read a partial package, or a header */
-
-             if (res == fd_data[data->ifd].remain) { /* we are done! */
-                char *buf = data->aiocb.aio_buf;
-                int i;
-
-                /* do we have anything buffered? */
-                if (fd_data[data->ifd].buf != NULL) {
-                   memcpy(fd_data[data->ifd].buf + fd_data[data->ifd].sz,
-                          buf, res);
-                   buf = fd_data[data->ifd].buf;
-                }
-
-                fd_data[data->ifd].sz += res;
-                driver_output(data->port_num, buf, (fd_data[data->ifd].sz>0?fd_data[data->ifd].sz:res));
-                clear_fd_data(data->ifd);
-
-                /* clear the previous read */
-                memset(data->aiocb.aio_buf, 0, res);
-             
-                /* issue a new read */
-                DISPATCH_AIO(sig);
-                data->aiocb.aio_nbytes = data->packet_bytes;
-
-                if (data->aiocb.aio_buf == NULL) {
-                   port_inp_failure(data->port_num, ready_fd, -1);
-                }
-                aio_read(&data->aiocb);
-             }
-             else if(res < fd_data[data->ifd].remain) { /* received part of a package */
-                if (fd_data[data->ifd].sz == 0) {
-
-                   fd_data[data->ifd].sz += res;
-                   memcpy(fd_data[data->ifd].buf, data->aiocb.aio_buf, res);
-                   fd_data[data->ifd].remain -= res;
-                }
-                else {
-                   memcpy(fd_data[data->ifd].buf + fd_data[data->ifd].sz,
-                          data->aiocb.aio_buf, res);
-                   fd_data[data->ifd].sz += res;
-                   fd_data[data->ifd].remain -= res;
-                }
-                /* clear the previous read */
-                memset(data->aiocb.aio_buf, 0, res);
-
-                /* issue a new read */
-                DISPATCH_AIO(sig);
-                data->aiocb.aio_nbytes = fd_data[data->ifd].remain;
-
-                if (data->aiocb.aio_buf == NULL) {
-                    port_inp_failure(data->port_num, ready_fd, -1);
-                }
-                aio_read(&data->aiocb);
-             }
-          }
-          else if (data->packet_bytes && fd_data[data->ifd].remain == 0) { /* we've recieved a header */
-
-             /* analyze the header FIXME  */
-             switch (data->packet_bytes) {
-                case 1: h = get_int8(data->aiocb.aio_buf);  break;
-                case 2: h = get_int16(data->aiocb.aio_buf); break;
-                case 4: h = get_int32(data->aiocb.aio_buf); break;
-             }
-
-             fd_data[data->ifd].buf = erts_alloc_fnf(ERTS_ALC_T_FD_ENTRY_BUF, h + data->packet_bytes);
-             fd_data[data->ifd].remain = ((h + data->packet_bytes) - res);
-
-             /* clear the previous read */
-             memset(data->aiocb.aio_buf, 0, data->packet_bytes);
-
-             /* issue a new read */
-             DISPATCH_AIO(sig);
-             data->aiocb.aio_nbytes = h;
-
-             if (data->aiocb.aio_buf == NULL) {
-                port_inp_failure(data->port_num, ready_fd, -1);
-             }
-             aio_read(&data->aiocb);
-          }
-       }
-       sig = erl_drv_ose_get_signal(ready_fd);
-    }
-}
-
-
-/* The parameter e is a pointer to the driver_data structure
- * related to the fd to be used as output.
- * ready_fd is the event that triggered this call to ready_input */
-static void ready_output(ErlDrvData drv_data, ErlDrvEvent ready_fd)
-{
-   SysIOVec *iov;
-   int vlen;
-   int res;
-   union SIGNAL *sig;
-   struct driver_data *data = (struct driver_data *)drv_data;
-
-   sig = erl_drv_ose_get_signal(ready_fd);
-   ASSERT(sig);
-
-   while (sig != NULL) {
-      if (sig->fm_write_reply.actual <= 0) {
-         int status;
-
-         status = efs_status_to_errno(sig->fm_write_reply.status);
-         driver_select(data->port_num, ready_fd, ERL_DRV_WRITE, 0);
-         DISPATCH_AIO(sig);
-         FREE_AIO(sig->fm_write_reply.buffer);
-
-         driver_failure_posix(data->port_num, status);
-      }
-      else { /* written bytes > 0 */
-          iov = driver_peekq(data->port_num, &vlen);
-          if (vlen > 0) {
-             DISPATCH_AIO(sig);
-             FREE_AIO(sig->fm_write_reply.buffer);
-             res = driver_deq(data->port_num, iov[0].iov_len);
-             if (res > 0) { 
-                iov = driver_peekq(data->port_num, &vlen);
-                WRITE_AIO(data->ofd, iov[0].iov_len, iov[0].iov_base);
-             }
-         }
-         else if (vlen == 0) {
-            DISPATCH_AIO(sig);
-            FREE_AIO(sig->fm_write_reply.buffer);
-         }
-
-      }
-      sig = erl_drv_ose_get_signal(ready_fd);
-   }
-}
-
-static void stop_select(ErlDrvEvent ready_fd, void* _)
-{
-   int *fd;
-   erl_drv_ose_event_fetch(ready_fd, NULL, NULL, (void **)&fd);
-   erl_drv_ose_event_free(ready_fd);
-   close(*fd);
-}
-
-
-void erts_do_break_handling(void)
-{
-    struct termios temp_mode;
-    int saved = 0;
-
-    /*
-     * Most functions that do_break() calls are intentionally not thread safe;
-     * therefore, make sure that all threads but this one are blocked before
-     * proceeding!
-     */
-    erts_smp_thr_progress_block();
-
-    /* during break we revert to initial settings */
-    /* this is done differently for oldshell */
-    if (using_oldshell && !replace_intr) {
-      SET_BLOCKING(1);
-    }
-    else if (isatty(0)) {
-      tcgetattr(0,&temp_mode);
-      tcsetattr(0,TCSANOW,&initial_tty_mode);
-      saved = 1;
-    }
-
-    /* call the break handling function, reset the flag */
-    do_break();
-
-    fflush(stdout);
-
-    /* after break we go back to saved settings */
-    if (using_oldshell && !replace_intr) {
-      SET_NONBLOCKING(1);
-    }
-    else if (saved) {
-      tcsetattr(0,TCSANOW,&temp_mode);
-    }
-
-    erts_smp_thr_progress_unblock();
-}
-
-static pid_t
-getpid(void)
-{
-   return get_bid(current_process());
-}
-
-int getpagesize(void)
-{
-   return 1024;
-}
-
-
-/* Fills in the systems representation of the jam/beam process identifier.
-** The Pid is put in STRING representation in the supplied buffer,
-** no interpretatione of this should be done by the rest of the
-** emulator. The buffer should be at least 21 bytes long.
-*/
-void sys_get_pid(char *buffer, size_t buffer_size){
-    pid_t p = getpid();
-    /* Assume the pid is scalar and can rest in an unsigned long... */
-    erts_snprintf(buffer, buffer_size, "%lu",(unsigned long) p);
-}
-
-int
-erts_sys_putenv_raw(char *key, char *value) {
-    return erts_sys_putenv(key, value);
-}
-int
-erts_sys_putenv(char *key, char *value)
-{
-    int res;
-
-    erts_smp_rwmtx_rwlock(&environ_rwmtx);
-    res = set_env(get_bid(current_process()), key,
-		  value);
-    erts_smp_rwmtx_rwunlock(&environ_rwmtx);
-    return res;
-}
-
-
-int
-erts_sys_unsetenv(char *key)
-{
-    int res;
-
-    erts_smp_rwmtx_rwlock(&environ_rwmtx);
-    res = set_env(get_bid(current_process()),key,NULL);
-    erts_smp_rwmtx_rwunlock(&environ_rwmtx);
-
-    return res;
-}
-
-int
-erts_sys_getenv__(char *key, char *value, size_t *size)
-{
-    int res;
-    char *orig_value = get_env(get_bid(current_process()), key);
-    if (!orig_value)
-	res = -1;
-    else {
-	size_t len = sys_strlen(orig_value);
-	if (len >= *size) {
-	    *size = len + 1;
-	    res = 1;
-	}
-	else {
-	    *size = len;
-	    sys_memcpy((void *) value, (void *) orig_value, len+1);
-	    res = 0;
-	}
-	free_buf((union SIGNAL **)&orig_value);
-    }
-    return res;
-}
-
-int
-erts_sys_getenv_raw(char *key, char *value, size_t *size) {
-    return erts_sys_getenv(key, value, size);
-}
-
-/*
- * erts_sys_getenv
- * returns:
- *  -1, if environment key is not set with a value
- *   0, if environment key is set and value fits into buffer res
- *   1, if environment key is set but does not fit into buffer res
- *      res is set with the needed buffer res value
- */
-
-int
-erts_sys_getenv(char *key, char *value, size_t *size)
-{
-    int res;
-    erts_smp_rwmtx_rlock(&environ_rwmtx);
-    res = erts_sys_getenv__(key, value, size);
-    erts_smp_rwmtx_runlock(&environ_rwmtx);
-    return res;
-}
-
-void
-sys_init_io(void)
-{
-    fd_data = (struct fd_data *)
-	erts_alloc(ERTS_ALC_T_FD_TAB, max_files * sizeof(struct fd_data));
-    erts_smp_atomic_add_nob(&sys_misc_mem_sz,
-			    max_files * sizeof(struct fd_data));
-}
-
-extern const char pre_loaded_code[];
-extern Preload pre_loaded[];
-
-void erts_sys_alloc_init(void)
-{
-}
-
-void *erts_sys_alloc(ErtsAlcType_t t, void *x, Uint sz)
-{
-    void *res = malloc((size_t) sz);
-#if HAVE_ERTS_MSEG
-    if (!res) {
-	erts_mseg_clear_cache();
-	return malloc((size_t) sz);
-    }
-#endif
-    return res;
-}
-
-void *erts_sys_realloc(ErtsAlcType_t t, void *x, void *p, Uint sz)
-{
-    void *res = realloc(p, (size_t) sz);
-#if HAVE_ERTS_MSEG
-    if (!res) {
-	erts_mseg_clear_cache();
-	return realloc(p, (size_t) sz);
-    }
-#endif
-    return res;
-}
-
-void erts_sys_free(ErtsAlcType_t t, void *x, void *p)
-{
-    free(p);
-}
-
-/* Return a pointer to a vector of names of preloaded modules */
-
-Preload*
-sys_preloaded(void)
-{
-    return pre_loaded;
-}
-
-/* Return a pointer to preloaded code for module "module" */
-unsigned char*
-sys_preload_begin(Preload* p)
-{
-    return p->code;
-}
-
-/* Clean up if allocated */
-void sys_preload_end(Preload* p)
-{
-    /* Nothing */
-}
-
-/* Read a key from console (?) */
-
-int sys_get_key(fd)
-int fd;
-{
-    int c;
-    unsigned char rbuf[64];
-
-    fflush(stdout);		/* Flush query ??? */
-
-    if ((c = read(fd,rbuf,64)) <= 0) {
-      return c;
-    }
-
-    return rbuf[0];
-}
-
-
-#ifdef DEBUG
-
-extern int erts_initialized;
-void
-erl_assert_error(const char* expr, const char* func,
-		 const char* file, int line)
-{
-    fflush(stdout);
-    fprintf(stderr, "%s:%d:%s() Assertion failed: %s\n",
-	    file, line, func, expr);
-    fflush(stderr);
-    ramlog_printf("%s:%d:%s() Assertion failed: %s\n",
-		  file, line, func, expr);
-
-    abort();
-}
-
-void
-erl_debug(char* fmt, ...)
-{
-    char sbuf[1024];		/* Temporary buffer. */
-    va_list va;
-
-    if (debug_log) {
-	va_start(va, fmt);
-	vsprintf(sbuf, fmt, va);
-	va_end(va);
-	fprintf(stderr, "%s", sbuf);
-    }
-}
-
-#endif /* DEBUG */
-
-static ERTS_INLINE void
-report_exit_status(ErtsSysReportExit *rep, int status)
-{
-   if (rep->ifd >= 0) {
-      driver_data[rep->ifd].alive = 0;
-      driver_data[rep->ifd].status = status;
-   }
-   if (rep->ofd >= 0) {
-      driver_data[rep->ofd].alive = 0;
-      driver_data[rep->ofd].status = status;
-   }
-
-   erts_free(ERTS_ALC_T_PRT_REP_EXIT, rep);
-}
-
-#define ERTS_REPORT_EXIT_STATUS report_exit_status
-
-/*
- * Called from schedule() when it runs out of runnable processes,
- * or when Erlang code has performed INPUT_REDUCTIONS reduction
- * steps. runnable == 0 iff there are no runnable Erlang processes.
- */
-void
-erl_sys_schedule(int runnable)
-{
-    ASSERT(get_fsem(current_process()) == 0);
-#ifdef ERTS_SMP
-    ASSERT(erts_get_scheduler_data()->no == 1);
-    ERTS_CHK_IO(!runnable);
-#else
-    ERTS_CHK_IO( 1 );
-#endif
-    ASSERT(get_fsem(current_process()) == 0);
-    ERTS_SMP_LC_ASSERT(!erts_thr_progress_is_blocking());
-}
-
-
-#ifdef ERTS_SMP
-
-void
-erts_sys_main_thread(void)
-{
-    erts_thread_disable_fpe();
-
-    /* Become signal receiver thread... */
-#ifdef ERTS_ENABLE_LOCK_CHECK
-    erts_lc_set_thread_name("signal_receiver");
-#endif
-
-    while (1) {
-       static const SIGSELECT sigsel[] = {0};
-       union SIGNAL *msg = receive(sigsel);
-
-       fprintf(stderr,"Main thread got message %d from 0x%x!!\r\n",
-               msg->sig_no, sender(&msg));
-       free_buf(&msg);
-    }
-}
-
-#endif /* ERTS_SMP */
-
-void
-erl_sys_args(int* argc, char** argv)
-{
-    int i, j;
-
-    erts_smp_rwmtx_init(&environ_rwmtx, "environ");
-
-    init_check_io();
-
-    /* Handled arguments have been marked with NULL. Slide arguments
-       not handled towards the beginning of argv. */
-    for (i = 0, j = 0; i < *argc; i++) {
-	if (argv[i])
-	    argv[j++] = argv[i];
-    }
-    *argc = j;
-
-}
diff --git a/erts/emulator/sys/ose/sys_float.c b/erts/emulator/sys/ose/sys_float.c
deleted file mode 100644
index 3d9abc6bd1..0000000000
--- a/erts/emulator/sys/ose/sys_float.c
+++ /dev/null
@@ -1,845 +0,0 @@
-/*
- * %CopyrightBegin%
- *
- * Copyright Ericsson AB 2001-2013. All Rights Reserved.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- *
- * %CopyrightEnd%
- */
-
-#ifdef HAVE_CONFIG_H
-#  include "config.h"
-#endif
-
-#include "sys.h"
-#include "global.h"
-#include "erl_process.h"
-
-
-#ifdef NO_FPE_SIGNALS
-
-void
-erts_sys_init_float(void)
-{
-# ifdef SIGFPE
-    sys_sigset(SIGFPE, SIG_IGN); /* Ignore so we can test for NaN and Inf */
-# endif
-}
-
-#else  /* !NO_FPE_SIGNALS */
-
-#ifdef ERTS_SMP
-static erts_tsd_key_t fpe_key;
-
-/* once-only initialisation early in the main thread (via erts_sys_init_float()) */
-static void erts_init_fp_exception(void)
-{
-    /* XXX: the wrappers prevent using a pthread destructor to
-       deallocate the key's value; so when/where do we do that? */
-    erts_tsd_key_create(&fpe_key);
-}
-
-void erts_thread_init_fp_exception(void)
-{
-    unsigned long *fpe = erts_alloc(ERTS_ALC_T_FP_EXCEPTION, sizeof(*fpe));
-    *fpe = 0L;
-    erts_tsd_set(fpe_key, fpe);
-}
-
-static ERTS_INLINE volatile unsigned long *erts_thread_get_fp_exception(void)
-{
-    return (volatile unsigned long*)erts_tsd_get(fpe_key);
-}
-#else /* !SMP */
-#define erts_init_fp_exception()	/*empty*/
-static volatile unsigned long fp_exception;
-#define erts_thread_get_fp_exception()	(&fp_exception)
-#endif /* SMP */
-
-volatile unsigned long *erts_get_current_fp_exception(void)
-{
-    Process *c_p;
-
-    c_p = erts_get_current_process();
-    if (c_p)
-	return &c_p->fp_exception;
-    return erts_thread_get_fp_exception();
-}
-
-static void set_current_fp_exception(unsigned long pc)
-{
-    volatile unsigned long *fpexnp = erts_get_current_fp_exception();
-    ASSERT(fpexnp != NULL);
-    *fpexnp = pc;
-}
-
-void erts_fp_check_init_error(volatile unsigned long *fpexnp)
-{
-    char buf[64];
-    snprintf(buf, sizeof buf, "ERTS_FP_CHECK_INIT at %p: detected unhandled FPE at %p\r\n",
-	     __builtin_return_address(0), (void*)*fpexnp);
-    if (write(2, buf, strlen(buf)) <= 0)
-	erl_exit(ERTS_ABORT_EXIT, "%s", buf);
-    *fpexnp = 0;
-#if defined(__i386__) || defined(__x86_64__)
-    erts_restore_fpu();
-#endif
-}
-
-/* Is there no standard identifier for Darwin/MacOSX ? */
-#if defined(__APPLE__) && defined(__MACH__) && !defined(__DARWIN__)
-#define __DARWIN__ 1
-#endif
-
-#if (defined(__i386__) || defined(__x86_64__)) && defined(__GNUC__)
-
-static void unmask_x87(void)
-{
-    unsigned short cw;
-
-    __asm__ __volatile__("fstcw %0" : "=m"(cw));
-    cw &= ~(0x01|0x04|0x08);   /* unmask IM, ZM, OM */
-    __asm__ __volatile__("fldcw %0" : : "m"(cw));
-}
-
-/* mask x87 FPE, return true if the previous state was unmasked */
-static int mask_x87(void)
-{
-    unsigned short cw;
-    int unmasked;
-
-    __asm__ __volatile__("fstcw %0" : "=m"(cw));
-    unmasked = (cw & (0x01|0x04|0x08)) == 0;
-    /* or just set cw = 0x37f */
-    cw |= (0x01|0x04|0x08); /* mask IM, ZM, OM */
-    __asm__ __volatile__("fldcw %0" : : "m"(cw));
-    return unmasked;
-}
-
-static void unmask_sse2(void)
-{
-    unsigned int mxcsr;
-
-    __asm__ __volatile__("stmxcsr %0" : "=m"(mxcsr));
-    mxcsr &= ~(0x003F|0x0680); /* clear exn flags, unmask OM, ZM, IM (not PM, UM, DM) */
-    __asm__ __volatile__("ldmxcsr %0" : : "m"(mxcsr));
-}
-
-/* mask SSE2 FPE, return true if the previous state was unmasked */
-static int mask_sse2(void)
-{
-    unsigned int mxcsr;
-    int unmasked;
-
-    __asm__ __volatile__("stmxcsr %0" : "=m"(mxcsr));
-    unmasked = (mxcsr & 0x0680) == 0;
-    /* or just set mxcsr = 0x1f80 */
-    mxcsr &= ~0x003F; /* clear exn flags */
-    mxcsr |=  0x0680; /* mask OM, ZM, IM (not PM, UM, DM) */
-    __asm__ __volatile__("ldmxcsr %0" : : "m"(mxcsr));
-    return unmasked;
-}
-
-#if defined(__x86_64__)
-
-static inline int cpu_has_sse2(void) { return 1; }
-
-#else /* !__x86_64__ */
-
-/*
- * Check if an x86-32 processor has SSE2.
- */
-static unsigned int xor_eflags(unsigned int mask)
-{
-    unsigned int eax, edx;
-
-    eax = mask;			/* eax = mask */
-    __asm__("pushfl\n\t"
-	    "popl %0\n\t"	/* edx = original EFLAGS */
-	    "xorl %0, %1\n\t"	/* eax = mask ^ EFLAGS */
-	    "pushl %1\n\t"
-	    "popfl\n\t"		/* new EFLAGS = mask ^ original EFLAGS */
-	    "pushfl\n\t"
-	    "popl %1\n\t"	/* eax = new EFLAGS */
-	    "xorl %0, %1\n\t"	/* eax = new EFLAGS ^ old EFLAGS */
-	    "pushl %0\n\t"
-	    "popfl"		/* restore original EFLAGS */
-	    : "=d"(edx), "=a"(eax)
-	    : "1"(eax));
-    return eax;
-}
-
-static __inline__ unsigned int cpuid_eax(unsigned int op)
-{
-    unsigned int eax, save_ebx;
-
-    /* In PIC mode i386 reserves EBX. So we must save
-       and restore it ourselves to not upset gcc. */
-    __asm__(
-	"movl %%ebx, %1\n\t"
-	"cpuid\n\t"
-	"movl %1, %%ebx"
-	: "=a"(eax), "=m"(save_ebx)
-	: "0"(op)
-	: "cx", "dx");
-    return eax;
-}
-
-static __inline__ unsigned int cpuid_edx(unsigned int op)
-{
-    unsigned int eax, edx, save_ebx;
-
-    /* In PIC mode i386 reserves EBX. So we must save
-       and restore it ourselves to not upset gcc. */
-    __asm__(
-	"movl %%ebx, %2\n\t"
-	"cpuid\n\t"
-	"movl %2, %%ebx"
-	: "=a"(eax), "=d"(edx), "=m"(save_ebx)
-	: "0"(op)
-	: "cx");
-    return edx;
-}
-
-/* The AC bit, bit #18, is a new bit introduced in the EFLAGS
- * register on the Intel486 processor to generate alignment
- * faults. This bit cannot be set on the Intel386 processor.
- */
-static __inline__ int is_386(void)
-{
-    return ((xor_eflags(1<<18) >> 18) & 1) == 0;
-}
-
-/* Newer x86 processors have a CPUID instruction, as indicated by
- * the ID bit (#21) in EFLAGS being modifiable.
- */
-static __inline__ int has_CPUID(void)
-{
-    return (xor_eflags(1<<21) >> 21) & 1;
-}
-
-static int cpu_has_sse2(void)
-{
-    unsigned int maxlev, features;
-    static int has_sse2 = -1;
-
-    if (has_sse2 >= 0)
-	return has_sse2;
-    has_sse2 = 0;
-
-    if (is_386())
-	return 0;
-    if (!has_CPUID())
-	return 0;
-    maxlev = cpuid_eax(0);
-    /* Intel A-step Pentium had a preliminary version of CPUID.
-       It also didn't have SSE2. */
-    if ((maxlev & 0xFFFFFF00) == 0x0500)
-	return 0;
-    /* If max level is zero then CPUID cannot report any features. */
-    if (maxlev == 0)
-	return 0;
-    features = cpuid_edx(1);
-    has_sse2 = (features & (1 << 26)) != 0;
-
-    return has_sse2;
-}
-#endif /* !__x86_64__ */
-
-static void unmask_fpe(void)
-{
-    __asm__ __volatile__("fnclex");
-    unmask_x87();
-    if (cpu_has_sse2())
-	unmask_sse2();
-}
-
-static void unmask_fpe_conditional(int unmasked)
-{
-    if (unmasked)
-	unmask_fpe();
-}
-
-/* mask x86 FPE, return true if the previous state was unmasked */
-static int mask_fpe(void)
-{
-    int unmasked;
-
-    unmasked = mask_x87();
-    if (cpu_has_sse2())
-	unmasked |= mask_sse2();
-    return unmasked;
-}
-
-void erts_restore_fpu(void)
-{
-    __asm__ __volatile__("fninit");
-    unmask_x87();
-    if (cpu_has_sse2())
-	unmask_sse2();
-}
-
-#elif defined(__sparc__) && defined(__linux__)
-
-#if defined(__arch64__)
-#define LDX "ldx"
-#define STX "stx"
-#else
-#define LDX "ld"
-#define STX "st"
-#endif
-
-static void unmask_fpe(void)
-{
-    unsigned long fsr;
-
-    __asm__(STX " %%fsr, %0" : "=m"(fsr));
-    fsr &= ~(0x1FUL << 23);	/* clear FSR[TEM] field */
-    fsr |= (0x1AUL << 23);	/* enable NV, OF, DZ exceptions */
-    __asm__ __volatile__(LDX " %0, %%fsr" : : "m"(fsr));
-}
-
-static void unmask_fpe_conditional(int unmasked)
-{
-    if (unmasked)
-	unmask_fpe();
-}
-
-/* mask SPARC FPE, return true if the previous state was unmasked */
-static int mask_fpe(void)
-{
-    unsigned long fsr;
-    int unmasked;
-
-    __asm__(STX " %%fsr, %0" : "=m"(fsr));
-    unmasked = ((fsr >> 23) & 0x1A) == 0x1A;
-    fsr &= ~(0x1FUL << 23);	/* clear FSR[TEM] field */
-    __asm__ __volatile__(LDX " %0, %%fsr" : : "m"(fsr));
-    return unmasked;
-}
-
-#elif (defined(__powerpc__) && defined(__linux__)) || (defined(__ppc__) && defined(__DARWIN__))
-
-#if defined(__linux__)
-#include <sys/prctl.h>
-
-static void set_fpexc_precise(void)
-{
-    if (prctl(PR_SET_FPEXC, PR_FP_EXC_PRECISE) < 0) {
-	perror("PR_SET_FPEXC");
-	exit(1);
-    }
-}
-
-#elif defined(__DARWIN__)
-
-#include <mach/mach.h>
-#include <pthread.h>
-
-/*
- * FE0 FE1	MSR bits
- *  0   0	floating-point exceptions disabled
- *  0   1	floating-point imprecise nonrecoverable
- *  1   0	floating-point imprecise recoverable
- *  1   1	floating-point precise mode
- *
- * Apparently:
- * - Darwin 5.5 (MacOS X <= 10.1) starts with FE0 == FE1 == 0,
- *   and resets FE0 and FE1 to 0 after each SIGFPE.
- * - Darwin 6.0 (MacOS X 10.2) starts with FE0 == FE1 == 1,
- *   and does not reset FE0 or FE1 after a SIGFPE.
- */
-#define FE0_MASK	(1<<11)
-#define FE1_MASK	(1<<8)
-
-/* a thread cannot get or set its own MSR bits */
-static void *fpu_fpe_enable(void *arg)
-{
-    thread_t t = *(thread_t*)arg;
-    struct ppc_thread_state state;
-    unsigned int state_size = PPC_THREAD_STATE_COUNT;
-
-    if (thread_get_state(t, PPC_THREAD_STATE, (natural_t*)&state, &state_size) != KERN_SUCCESS) {
-	perror("thread_get_state");
-	exit(1);
-    }
-    if ((state.srr1 & (FE1_MASK|FE0_MASK)) != (FE1_MASK|FE0_MASK)) {
-#if 1
-	/* This would also have to be performed in the SIGFPE handler
-	   to work around the MSR reset older Darwin releases do. */
-	state.srr1 |= (FE1_MASK|FE0_MASK);
-	thread_set_state(t, PPC_THREAD_STATE, (natural_t*)&state, state_size);
-#else
-	fprintf(stderr, "srr1 == 0x%08x, your Darwin is too old\n", state.srr1);
-	exit(1);
-#endif
-    }
-    return NULL; /* Ok, we appear to be on Darwin 6.0 or later */
-}
-
-static void set_fpexc_precise(void)
-{
-    thread_t self = mach_thread_self();
-    pthread_t enabler;
-
-    if (pthread_create(&enabler, NULL, fpu_fpe_enable, &self)) {
-	perror("pthread_create");
-    } else if (pthread_join(enabler, NULL)) {
-	perror("pthread_join");
-    }
-}
-
-#endif
-
-static void set_fpscr(unsigned int fpscr)
-{
-    union {
-	double d;
-	unsigned int fpscr[2];
-    } u;
-
-    u.fpscr[0] = 0xFFF80000;
-    u.fpscr[1] = fpscr;
-    __asm__ __volatile__("mtfsf 255,%0" : : "f"(u.d));
-}
-
-static unsigned int get_fpscr(void)
-{
-    union {
-	double d;
-	unsigned int fpscr[2];
-    } u;
-
-    __asm__("mffs %0" : "=f"(u.d));
-    return u.fpscr[1];
-}
-
-static void unmask_fpe(void)
-{
-    set_fpexc_precise();
-    set_fpscr(0x80|0x40|0x10);	/* VE, OE, ZE; not UE or XE */
-}
-
-static void unmask_fpe_conditional(int unmasked)
-{
-    if (unmasked)
-	unmask_fpe();
-}
-
-/* mask PowerPC FPE, return true if the previous state was unmasked */
-static int mask_fpe(void)
-{
-    int unmasked;
-
-    unmasked = (get_fpscr() & (0x80|0x40|0x10)) == (0x80|0x40|0x10);
-    set_fpscr(0x00);
-    return unmasked;
-}
-
-#else
-
-static void unmask_fpe(void)
-{
-    fpsetmask(FP_X_INV | FP_X_OFL | FP_X_DZ);
-}
-
-static void unmask_fpe_conditional(int unmasked)
-{
-    if (unmasked)
-	unmask_fpe();
-}
-
-/* mask IEEE FPE, return true if previous state was unmasked */
-static int mask_fpe(void)
-{
-    const fp_except unmasked_mask = FP_X_INV | FP_X_OFL | FP_X_DZ;
-    fp_except old_mask;
-
-    old_mask = fpsetmask(0);
-    return (old_mask & unmasked_mask) == unmasked_mask;
-}
-
-#endif
-
-#if (defined(__linux__) && (defined(__i386__) || defined(__x86_64__) || defined(__sparc__) || defined(__powerpc__))) || (defined(__DARWIN__) && (defined(__i386__) || defined(__x86_64__) || defined(__ppc__))) || (defined(__FreeBSD__) && (defined(__x86_64__) || defined(__i386__))) || ((defined(__NetBSD__) || defined(__OpenBSD__)) && defined(__x86_64__)) || (defined(__sun__) && defined(__x86_64__))
-
-#if defined(__linux__) && defined(__i386__)
-#if !defined(X86_FXSR_MAGIC)
-#define X86_FXSR_MAGIC 0x0000
-#endif
-#elif defined(__FreeBSD__) && defined(__x86_64__)
-#include <sys/types.h>
-#include <machine/fpu.h>
-#elif defined(__FreeBSD__) && defined(__i386__)
-#include <sys/types.h>
-#include <machine/npx.h>
-#elif defined(__DARWIN__)
-#include <machine/signal.h>
-#elif defined(__OpenBSD__) && defined(__x86_64__)
-#include <sys/types.h>
-#include <machine/fpu.h>
-#endif
-#if !(defined(__OpenBSD__) && defined(__x86_64__))
-#include <ucontext.h>
-#endif
-#include <string.h>
-
-#if defined(__linux__) && defined(__x86_64__)
-#define mc_pc(mc)	((mc)->gregs[REG_RIP])
-#elif defined(__linux__) && defined(__i386__)
-#define mc_pc(mc)	((mc)->gregs[REG_EIP])
-#elif defined(__DARWIN__) && defined(__i386__)
-#ifdef DARWIN_MODERN_MCONTEXT
-#define mc_pc(mc)	((mc)->__ss.__eip)
-#else
-#define mc_pc(mc)	((mc)->ss.eip)
-#endif
-#elif defined(__DARWIN__) && defined(__x86_64__)
-#ifdef DARWIN_MODERN_MCONTEXT
-#define mc_pc(mc)	((mc)->__ss.__rip)
-#else
-#define mc_pc(mc)	((mc)->ss.rip)
-#endif
-#elif defined(__FreeBSD__) && defined(__x86_64__)
-#define mc_pc(mc)	((mc)->mc_rip)
-#elif defined(__FreeBSD__) && defined(__i386__)
-#define mc_pc(mc)	((mc)->mc_eip)
-#elif defined(__NetBSD__) && defined(__x86_64__)
-#define mc_pc(mc)	((mc)->__gregs[_REG_RIP])
-#elif defined(__NetBSD__) && defined(__i386__)
-#define mc_pc(mc)	((mc)->__gregs[_REG_EIP])
-#elif defined(__OpenBSD__) && defined(__x86_64__)
-#define mc_pc(mc)	((mc)->sc_rip)
-#elif defined(__sun__) && defined(__x86_64__)
-#define mc_pc(mc)	((mc)->gregs[REG_RIP])
-#endif
-
-static void fpe_sig_action(int sig, siginfo_t *si, void *puc)
-{
-    ucontext_t *uc = puc;
-    unsigned long pc;
-
-#if defined(__linux__)
-#if defined(__x86_64__)
-    mcontext_t *mc = &uc->uc_mcontext;
-    fpregset_t fpstate = mc->fpregs;
-    pc = mc_pc(mc);
-    /* A failed SSE2 instruction will restart. To avoid
-       looping we mask SSE2 exceptions now and unmask them
-       again later in erts_check_fpe()/erts_restore_fpu().
-       On RISCs we update PC to skip the failed instruction,
-       but the ever increasing complexity of the x86 instruction
-       set encoding makes that a poor solution here. */
-    fpstate->mxcsr = 0x1F80;
-    fpstate->swd &= ~0xFF;
-#elif defined(__i386__)
-    mcontext_t *mc = &uc->uc_mcontext;
-    fpregset_t fpstate = mc->fpregs;
-    pc = mc_pc(mc);
-    if ((fpstate->status >> 16) == X86_FXSR_MAGIC)
-	((struct _fpstate*)fpstate)->mxcsr = 0x1F80;
-    fpstate->sw &= ~0xFF;
-#elif defined(__sparc__) && defined(__arch64__)
-    /* on SPARC the 3rd parameter points to a sigcontext not a ucontext */
-    struct sigcontext *sc = (struct sigcontext*)puc;
-    pc = sc->sigc_regs.tpc;
-    sc->sigc_regs.tpc = sc->sigc_regs.tnpc;
-    sc->sigc_regs.tnpc += 4;
-#elif defined(__sparc__)
-    /* on SPARC the 3rd parameter points to a sigcontext not a ucontext */
-    struct sigcontext *sc = (struct sigcontext*)puc;
-    pc = sc->si_regs.pc;
-    sc->si_regs.pc = sc->si_regs.npc;
-    sc->si_regs.npc = (unsigned long)sc->si_regs.npc + 4;
-#elif defined(__powerpc__)
-#if defined(__powerpc64__)
-    mcontext_t *mc = &uc->uc_mcontext;
-    unsigned long *regs = &mc->gp_regs[0];
-#else
-    mcontext_t *mc = uc->uc_mcontext.uc_regs;
-    unsigned long *regs = &mc->gregs[0];
-#endif
-    pc = regs[PT_NIP];
-    regs[PT_NIP] += 4;
-    regs[PT_FPSCR] = 0x80|0x40|0x10;	/* VE, OE, ZE; not UE or XE */
-#endif
-#elif defined(__DARWIN__) && (defined(__i386__) || defined(__x86_64__))
-#ifdef DARWIN_MODERN_MCONTEXT
-    mcontext_t mc = uc->uc_mcontext;
-    pc = mc_pc(mc);
-    mc->__fs.__fpu_mxcsr = 0x1F80;
-    *(unsigned short *)&mc->__fs.__fpu_fsw &= ~0xFF;
-#else
-    mcontext_t mc = uc->uc_mcontext;
-    pc = mc_pc(mc);
-    mc->fs.fpu_mxcsr = 0x1F80;
-    *(unsigned short *)&mc->fs.fpu_fsw &= ~0xFF;
-#endif /* DARWIN_MODERN_MCONTEXT */
-#elif defined(__DARWIN__) && defined(__ppc__)
-    mcontext_t mc = uc->uc_mcontext;
-    pc = mc->ss.srr0;
-    mc->ss.srr0 += 4;
-    mc->fs.fpscr = 0x80|0x40|0x10;
-#elif defined(__FreeBSD__) && defined(__x86_64__)
-    mcontext_t *mc = &uc->uc_mcontext;
-    struct savefpu *savefpu = (struct savefpu*)&mc->mc_fpstate;
-    struct envxmm *envxmm = &savefpu->sv_env;
-    pc = mc_pc(mc);
-    envxmm->en_mxcsr = 0x1F80;
-    envxmm->en_sw &= ~0xFF;
-#elif defined(__FreeBSD__) && defined(__i386__)
-    mcontext_t *mc = &uc->uc_mcontext;
-    union savefpu *savefpu = (union savefpu*)&mc->mc_fpstate;
-    pc = mc_pc(mc);
-    if (mc->mc_fpformat == _MC_FPFMT_XMM) {
-	struct envxmm *envxmm = &savefpu->sv_xmm.sv_env;
-	envxmm->en_mxcsr = 0x1F80;
-	envxmm->en_sw &= ~0xFF;
-    } else {
-	struct env87 *env87 = &savefpu->sv_87.sv_env;
-	env87->en_sw &= ~0xFF;
-    }
-#elif defined(__NetBSD__) && defined(__x86_64__)
-    mcontext_t *mc = &uc->uc_mcontext;
-    struct fxsave64 *fxsave = (struct fxsave64 *)&mc->__fpregs;
-    pc = mc_pc(mc);
-    fxsave->fx_mxcsr = 0x1F80;
-    fxsave->fx_fsw &= ~0xFF;
-#elif defined(__NetBSD__) && defined(__i386__)
-    mcontext_t *mc = &uc->uc_mcontext;
-    pc = mc_pc(mc);
-    if (uc->uc_flags & _UC_FXSAVE) {
-	struct envxmm *envxmm = (struct envxmm *)&mc->__fpregs;
-	envxmm->en_mxcsr = 0x1F80;
-	envxmm->en_sw &= ~0xFF;
-    } else {
-	struct env87 *env87 = (struct env87 *)&mc->__fpregs;
-	env87->en_sw &= ~0xFF;
-    }
-#elif defined(__OpenBSD__) && defined(__x86_64__)
-    struct fxsave64 *fxsave = uc->sc_fpstate;
-    pc = mc_pc(uc);
-    fxsave->fx_mxcsr = 0x1F80;
-    fxsave->fx_fsw &= ~0xFF;
-#elif defined(__sun__) && defined(__x86_64__)
-    mcontext_t *mc = &uc->uc_mcontext;
-    struct fpchip_state *fpstate = &mc->fpregs.fp_reg_set.fpchip_state;
-    pc = mc_pc(mc);
-    fpstate->mxcsr = 0x1F80;
-    fpstate->sw &= ~0xFF;
-#endif
-#if 0
-    {
-	char buf[64];
-	snprintf(buf, sizeof buf, "%s: FPE at %p\r\n", __FUNCTION__, (void*)pc);
-	write(2, buf, strlen(buf));
-    }
-#endif
-    set_current_fp_exception(pc);
-}
-
-static void erts_thread_catch_fp_exceptions(void)
-{
-    struct sigaction act;
-    memset(&act, 0, sizeof act);
-    act.sa_sigaction = fpe_sig_action;
-    act.sa_flags = SA_SIGINFO;
-    sigaction(SIGFPE, &act, NULL);
-    unmask_fpe();
-}
-
-#else  /* !((__linux__ && (__i386__ || __x86_64__ || __powerpc__)) || (__DARWIN__ && (__i386__ || __x86_64__ || __ppc__))) */
-
-static void fpe_sig_handler(int sig)
-{
-    set_current_fp_exception(1); /* XXX: convert to sigaction so we can get the trap PC */
-}
-
-static void erts_thread_catch_fp_exceptions(void)
-{
-    sys_sigset(SIGFPE, fpe_sig_handler);
-    unmask_fpe();
-}
-
-#endif /* (__linux__ && (__i386__ || __x86_64__ || __powerpc__)) || (__DARWIN__ && (__i386__ || __x86_64__ || __ppc__))) */
-
-/* once-only initialisation early in the main thread */
-void erts_sys_init_float(void)
-{
-    erts_init_fp_exception();
-    erts_thread_catch_fp_exceptions();
-    erts_printf_block_fpe = erts_sys_block_fpe;
-    erts_printf_unblock_fpe = erts_sys_unblock_fpe;
-}
-
-#endif /* NO_FPE_SIGNALS */
-
-void erts_thread_init_float(void)
-{
-#ifdef ERTS_SMP
-    /* This allows Erlang schedulers to leave Erlang-process context
-       and still have working FP exceptions. XXX: is this needed? */
-    erts_thread_init_fp_exception();
-#endif
-
-#ifndef NO_FPE_SIGNALS
-    /* NOTE:
-     *  erts_thread_disable_fpe() is called in all threads at
-     *  creation. We at least need to call unmask_fpe()
-     */
-#if defined(__DARWIN__) || defined(__FreeBSD__)
-    /* Darwin (7.9.0) does not appear to propagate FP exception settings
-       to a new thread from its parent. So if we want FP exceptions, we
-       must manually re-enable them in each new thread.
-       FreeBSD 6.1 appears to suffer from a similar issue. */
-    erts_thread_catch_fp_exceptions();
-#else
-    unmask_fpe();
-#endif
-
-#endif
-}
-
-void erts_thread_disable_fpe(void)
-{
-#if !defined(NO_FPE_SIGNALS)
-    (void)mask_fpe();
-#endif
-}
-
-#if !defined(NO_FPE_SIGNALS)
-int erts_sys_block_fpe(void)
-{
-    return mask_fpe();
-}
-
-void erts_sys_unblock_fpe(int unmasked)
-{
-    unmask_fpe_conditional(unmasked);
-}
-#endif
-
-/* The following check is incorporated from the Vee machine */
-
-#define ISDIGIT(d) ((d) >= '0' && (d) <= '9')
-
-/*
- ** Convert a double to ascii format 0.dddde[+|-]ddd
- ** return number of characters converted or -1 if error.
- **
- ** These two functions should maybe use localeconv() to pick up
- ** the current radix character, but since it is uncertain how
- ** expensive such a system call is, and since no-one has heard
- ** of other radix characters than '.' and ',' an ad-hoc
- ** low execution time solution is used instead.
- */
-
-int
-sys_double_to_chars_ext(double fp, char *buffer, size_t buffer_size, size_t decimals)
-{
-    char *s = buffer;
-
-    if (erts_snprintf(buffer, buffer_size, "%.*e", decimals, fp) >= buffer_size)
-        return -1;
-    /* Search upto decimal point */
-    if (*s == '+' || *s == '-') s++;
-    while (ISDIGIT(*s)) s++;
-    if (*s == ',') *s++ = '.'; /* Replace ',' with '.' */
-    /* Scan to end of string */
-    while (*s) s++;
-    return s-buffer; /* i.e strlen(buffer) */
-}
-
-/* Float conversion */
-
-int
-sys_chars_to_double(char* buf, double* fp)
-{
-#ifndef NO_FPE_SIGNALS
-    volatile unsigned long *fpexnp = erts_get_current_fp_exception();
-#endif
-    char *s = buf, *t, *dp;
-
-    /* Robert says that something like this is what he really wanted:
-     * (The [.,] radix test is NOT what Robert wanted - it was added later)
-     *
-     * 7 == sscanf(Tbuf, "%[+-]%[0-9][.,]%[0-9]%[eE]%[+-]%[0-9]%s", ....);
-     * if (*s2 == 0 || *s3 == 0 || *s4 == 0 || *s6 == 0 || *s7)
-     *   break;
-     */
-
-    /* Scan string to check syntax. */
-    if (*s == '+' || *s == '-') s++;
-    if (!ISDIGIT(*s))		/* Leading digits. */
-      return -1;
-    while (ISDIGIT(*s)) s++;
-    if (*s != '.' && *s != ',')	/* Decimal part. */
-      return -1;
-    dp = s++;			/* Remember decimal point pos just in case */
-    if (!ISDIGIT(*s))
-      return -1;
-    while (ISDIGIT(*s)) s++;
-    if (*s == 'e' || *s == 'E') {
-	/* There is an exponent. */
-	s++;
-	if (*s == '+' || *s == '-') s++;
-	if (!ISDIGIT(*s))
-	  return -1;
-	while (ISDIGIT(*s)) s++;
-    }
-    if (*s)			/* That should be it */
-      return -1;
-
-#ifdef NO_FPE_SIGNALS
-    errno = 0;
-#endif
-    __ERTS_FP_CHECK_INIT(fpexnp);
-    *fp = strtod(buf, &t);
-    __ERTS_FP_ERROR_THOROUGH(fpexnp, *fp, return -1);
-    if (t != s) {		/* Whole string not scanned */
-	/* Try again with other radix char */
-	*dp = (*dp == '.') ? ',' : '.';
-	errno = 0;
-	__ERTS_FP_CHECK_INIT(fpexnp);
-	*fp = strtod(buf, &t);
-	__ERTS_FP_ERROR_THOROUGH(fpexnp, *fp, return -1);
-    }
-
-#ifdef NO_FPE_SIGNALS
-    if (errno == ERANGE) {
-	if (*fp == HUGE_VAL || *fp == -HUGE_VAL) {
-	    /* overflow, should give error */
-	    return -1;
-	} else if (t == s && *fp == 0.0) {
-	    /* This should give 0.0 - OTP-7178 */
-	    errno = 0;
-
-	} else if (*fp == 0.0) {
-	    return -1;
-	}
-    }
-#endif
-    return 0;
-}
-
-int
-matherr(struct exception *exc)
-{
-#if !defined(NO_FPE_SIGNALS)
-    volatile unsigned long *fpexnp = erts_get_current_fp_exception();
-    if (fpexnp != NULL)
-	*fpexnp = (unsigned long)__builtin_return_address(0);
-#endif
-    return 1;
-}
diff --git a/erts/emulator/sys/ose/sys_time.c b/erts/emulator/sys/ose/sys_time.c
deleted file mode 100644
index 5dac75956a..0000000000
--- a/erts/emulator/sys/ose/sys_time.c
+++ /dev/null
@@ -1,57 +0,0 @@
-/*
- * %CopyrightBegin%
- *
- * Copyright Ericsson AB 2005-2009. All Rights Reserved.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- *
- * %CopyrightEnd%
- */
-
-#ifdef HAVE_CONFIG_H
-#  include "config.h"
-#endif
-
-#include "sys.h"
-#include "global.h"
-
-/******************* Routines for time measurement *********************/
-
-int erts_ticks_per_sec = 0; /* Will be SYS_CLK_TCK in erl_unix_sys.h */
-
-int sys_init_time(void)
-{
-  return SYS_CLOCK_RESOLUTION;
-}
-
-clock_t sys_times(SysTimes *now) {
-  now->tms_utime = now->tms_stime = now->tms_cutime =  now->tms_cstime = 0;
-  return 0;
-}
-
-static OSTICK last_tick_count = 0;
-static SysHrTime wrap = 0;
-static OSTICK us_per_tick;
-
-void sys_init_hrtime() {
-  us_per_tick = system_tick();
-}
-
-SysHrTime sys_gethrtime() {
-  OSTICK ticks = get_ticks();
-  if (ticks < (SysHrTime) last_tick_count) {
-    wrap += 1ULL << 32;
-  }
-  last_tick_count = ticks;
-  return ((((SysHrTime) ticks) + wrap) * 1000*us_per_tick);
-}
diff --git a/erts/emulator/sys/unix/erl_child_setup.c b/erts/emulator/sys/unix/erl_child_setup.c
index a3c5c20641..4e61530cf1 100644
--- a/erts/emulator/sys/unix/erl_child_setup.c
+++ b/erts/emulator/sys/unix/erl_child_setup.c
@@ -1,7 +1,7 @@
 /*
  * %CopyrightBegin%
  * 
- * Copyright Ericsson AB 2002-2009. All Rights Reserved.
+ * Copyright Ericsson AB 2002-2015. All Rights Reserved.
  * 
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
@@ -19,109 +19,233 @@
  */
 
 /*
- * After a vfork() (or fork()) the child exec()s to this program which
- * sets up the child and exec()s to the user program (see spawn_start()
- * in sys.c and ticket OTP-4389).
+ * This program is started at erts startup and all fork's that
+ * have to be done are done in here. This is done for a couple
+ * of reasons:
+ *  - Allow usage of fork without a memory explosion.
+ *  -- we do not want to use vfork, as it blocks the VM
+ *     until the execv is done, and if the program that
+ *     is to be executed is on an NFS that is unavailable,
+ *     the execv can block for a very long time.
+ *  -- we cannot do fork inside the VM as that would temporarily
+ *     duplicate the memory usage of the VM per parallel exec.
+ *
+ * Some implementation notes:
+ *  - A single Unix Domain Socket is setup in between the VM and
+ *    this program. Over that UDS the file descriptors that should
+ *    be used to talk to the child program are sent.
+ *    The actual command to execute, together with options and the
+ *    environment, is sent over the pipe represented by the
+ *    file descriptors mentioned above. We don't send the
+ *    command over the UDS as that would increase the likely hood
+ *    that it's buffer would be full.
+ *
+ *  - Since it is this program that execv's, it has to take care of
+ *    all the SIGCHLD signals that the child programs generate. The
+ *    signals are received and the pid+exit reason is sent as data
+ *    on the UDS to the VM. The VM is then able to map the pid to the
+ *    port of the child program that just exited and deliver the status
+ *    code if requested.
  */
 
 #ifdef HAVE_CONFIG_H
 #  include "config.h"
 #endif
 
-#define NEED_CHILD_SETUP_DEFINES
-#include "sys.h"
-#include "erl_misc_utils.h"
+#include <stdlib.h>
+#include <stdio.h>
+#include <sys/wait.h>
 
-#ifdef SIG_SIGSET		/* Old SysV */
-void sys_sigrelease(int sig)
+#define WANT_NONBLOCKING
+
+#include "erl_driver.h"
+#include "sys_uds.h"
+#include "hash.h"
+#include "erl_child_setup.h"
+
+#define SET_CLOEXEC(fd) fcntl(fd, F_SETFD, fcntl(fd, F_GETFD) | FD_CLOEXEC)
+
+#if defined(__ANDROID__)
+#define SHELL "/system/bin/sh"
+#else
+#define SHELL "/bin/sh"
+#endif /* __ANDROID__ */
+
+//#define HARD_DEBUG
+#ifdef HARD_DEBUG
+#define DEBUG_PRINT(fmt, ...) fprintf(stderr, fmt "\r\n", ##__VA_ARGS__)
+#else
+#define DEBUG_PRINT(fmt, ...)
+#endif
+
+#define ABORT(fmt, ...) do {                                            \
+        fprintf(stderr, "erl_child_setup: " fmt "\r\n", ##__VA_ARGS__); \
+        abort();                                                        \
+    } while(0)
+
+#ifdef DEBUG
+void
+erl_assert_error(const char* expr, const char* func, const char* file, int line)
 {
-    sigrelse(sig);
+    fflush(stdout);
+    fprintf(stderr, "%s:%d:%s() Assertion failed: %s\n",
+            file, line, func, expr);
+    fflush(stderr);
+    abort();
 }
-#else /* !SIG_SIGSET */
-#ifdef SIG_SIGNAL		/* Old BSD */
-sys_sigrelease(int sig)
+#endif
+
+void sys_sigblock(int sig)
 {
-    sigsetmask(sigblock(0) & ~sigmask(sig));
+    sigset_t mask;
+
+    sigemptyset(&mask);
+    sigaddset(&mask, sig);
+    sigprocmask(SIG_BLOCK, &mask, (sigset_t *)NULL);
 }
-#else /* !SIG_SIGNAL */	/* The True Way - POSIX!:-) */
+
 void sys_sigrelease(int sig)
 {
     sigset_t mask;
-
     sigemptyset(&mask);
     sigaddset(&mask, sig);
     sigprocmask(SIG_UNBLOCK, &mask, (sigset_t *)NULL);
 }
-#endif /* !SIG_SIGNAL */
-#endif /* !SIG_SIGSET */
-
-#if defined(__ANDROID__)
-static int system_properties_fd(void);
-#endif /* __ANDROID__ */
 
-#if defined(__ANDROID__)
-#define SHELL "/system/bin/sh"
-#else
-#define SHELL "/bin/sh"
-#endif /* __ANDROID__ */
+static void add_os_pid_to_port_id_mapping(Eterm, pid_t);
+static Eterm get_port_id(pid_t);
+static int forker_hash_init(void);
 
+static int max_files = -1;
+static int sigchld_pipe[2];
 
-int
-main(int argc, char *argv[])
+static int
+start_new_child(int pipes[])
 {
-    int i, from, to;
-    int erts_spawn_executable = 0;
+    int size, res, i, pos = 0;
+    char *buff, *o_buff;
+
+    char *cmd, *wd, **new_environ, **args = NULL;
+
+    Sint cnt, flags;
 
-    /* OBSERVE!
-     * Keep child setup after fork() (implemented in sys.c) up to date
-     * if changes are made here.
-     */
+    /* only child executes here */
 
-    if (argc != CS_ARGV_NO_OF_ARGS) {
-	if (argc < CS_ARGV_NO_OF_ARGS) {
-	    return 1;
-	} else {
-	    erts_spawn_executable = 1;
-	}
+    do {
+        res = read(pipes[0], (char*)&size, sizeof(size));
+    } while(res < 0 && (errno == EINTR || errno == ERRNO_BLOCK));
+
+    if (res <= 0) {
+        goto child_error;
     }
 
-    if (strcmp("false", argv[CS_ARGV_UNBIND_IX]) != 0)
-	if (erts_unbind_from_cpu_str(argv[CS_ARGV_UNBIND_IX]) != 0)
-	    return 1;
+    buff = malloc(size);
+
+    DEBUG_PRINT("size = %d", size);
+
+    do {
+        if ((res = read(pipes[0], buff + pos, size - pos)) < 0) {
+            if (errno == ERRNO_BLOCK || errno == EINTR)
+                continue;
+            goto child_error;
+        }
+        if (res == 0) {
+            errno = EPIPE;
+            goto child_error;
+        }
+        pos += res;
+    } while(size - pos != 0);
+
+    o_buff = buff;
+
+    flags = get_int32(buff);
+    buff += sizeof(Sint32);
 
-    for (i = 0; i < CS_ARGV_NO_OF_DUP2_OPS; i++) {
-	if (argv[CS_ARGV_DUP2_OP_IX(i)][0] == '-'
-	    && argv[CS_ARGV_DUP2_OP_IX(i)][1] == '\0')
-	    break;
-	if (sscanf(argv[CS_ARGV_DUP2_OP_IX(i)], "%d:%d", &from, &to) != 2)
-	    return 1;
-	if (dup2(from, to) < 0)
-	    return 1;
+    DEBUG_PRINT("flags = %d", flags);
+
+    cmd = buff;
+    buff += strlen(buff) + 1;
+    if (*buff == '\0') {
+        wd = NULL;
+    } else {
+        wd = buff;
+        buff += strlen(buff) + 1;
     }
+    buff++;
 
-    if (sscanf(argv[CS_ARGV_FD_CR_IX], "%d:%d", &from, &to) != 2)
-	return 1;
+    DEBUG_PRINT("wd = %s", wd);
 
-#if defined(HAVE_CLOSEFROM)
-    closefrom(from);
-#elif defined(__ANDROID__)
-    if (from <= to) {
-	int spfd = system_properties_fd();
-	for (i = from; i <= to; i++) {
-	    if (i != spfd) {
-		(void) close(i);
-            }
-	}
+    cnt = get_int32(buff);
+    buff += sizeof(Sint32);
+    new_environ = malloc(sizeof(char*)*(cnt + 1));
+
+    DEBUG_PRINT("env_len = %ld", cnt);
+    for (i = 0; i < cnt; i++, buff++) {
+        new_environ[i] = buff;
+        while(*buff != '\0') buff++;
     }
-#else  /* !__ANDROID__ */
-    for (i = from; i <= to; i++) {
-	(void) close(i);
+    new_environ[cnt] = NULL;
+
+    if (o_buff + size != buff) {
+        /* This is a spawn executable call */
+        cnt = get_int32(buff);
+        buff += sizeof(Sint32);
+        args = malloc(sizeof(char*)*(cnt + 1));
+        for (i = 0; i < cnt; i++, buff++) {
+            args[i] = buff;
+            while(*buff != '\0') buff++;
+        }
+        args[cnt] = NULL;
     }
-#endif /* HAVE_CLOSEFROM */
 
-    if (!(argv[CS_ARGV_WD_IX][0] == '.' && argv[CS_ARGV_WD_IX][1] == '\0')
-	&& chdir(argv[CS_ARGV_WD_IX]) < 0)
-	return 1;
+    if (o_buff + size != buff) {
+        errno = EINVAL;
+        goto child_error;
+    }
+
+    DEBUG_PRINT("read ack");
+    do {
+        ErtsSysForkerProto proto;
+        res = read(pipes[0], &proto, sizeof(proto));
+        if (res > 0) {
+            ASSERT(proto.action == ErtsSysForkerProtoAction_Ack);
+            ASSERT(res == sizeof(proto));
+        }
+    } while(res < 0 && (errno == EINTR || errno == ERRNO_BLOCK));
+    if (res < 1) {
+        errno = EPIPE;
+        goto child_error;
+    }
+
+    DEBUG_PRINT("Do that forking business: '%s'\n",cmd);
+
+    /* When the dup2'ing below is done, only
+       fd's 0, 1, 2 and maybe 3, 4 should survive the
+       exec. All other fds (i.e. the unix domain sockets
+       and stray pipe ends) should have CLOEXEC set on them
+       so they will be closed when the exec happens */
+    if (flags & FORKER_FLAG_USE_STDIO) {
+        /* stdin for process */
+        if (flags & FORKER_FLAG_DO_WRITE &&
+            dup2(pipes[0], 0) < 0)
+            goto child_error;
+        /* stdout for process */
+        if (flags & FORKER_FLAG_DO_READ &&
+            dup2(pipes[1], 1) < 0)
+            goto child_error;
+    }
+    else {	/* XXX will fail if pipes[0] == 4 (unlikely..) */
+        if (flags & FORKER_FLAG_DO_READ && dup2(pipes[1], 4) < 0)
+            goto child_error;
+        if (flags & FORKER_FLAG_DO_WRITE && dup2(pipes[0], 3) < 0)
+            goto child_error;
+    }
+
+    if (dup2(pipes[2], 2) < 0)
+        goto child_error;
+
+    if (wd && chdir(wd) < 0)
+        goto child_error;
 
 #if defined(USE_SETPGRP_NOARGS)		/* SysV */
     (void) setpgrp();
@@ -131,34 +255,301 @@ main(int argc, char *argv[])
     (void) setsid();
 #endif
 
+    close(pipes[0]);
+    close(pipes[1]);
+    close(pipes[2]);
+
     sys_sigrelease(SIGCHLD);
-    sys_sigrelease(SIGINT);
-    sys_sigrelease(SIGUSR1);
-
-    if (erts_spawn_executable) {
-	if (argv[CS_ARGV_NO_OF_ARGS + 1] == NULL) {
-	    execl(argv[CS_ARGV_NO_OF_ARGS],argv[CS_ARGV_NO_OF_ARGS],
-		  (char *) NULL);
-	} else {
-	    execv(argv[CS_ARGV_NO_OF_ARGS],&(argv[CS_ARGV_NO_OF_ARGS + 1]));
-	}
+
+    if (args) {
+        /* spawn_executable */
+        execve(cmd, args, new_environ);
     } else {
-        execl(SHELL, "sh", "-c", argv[CS_ARGV_CMD_IX], (char *) NULL);
+        execle(SHELL, "sh", "-c", cmd, (char *) NULL, new_environ);
     }
-    return 1;
+child_error:
+    DEBUG_PRINT("exec error: %d\r\n",errno);
+    _exit(128 + errno);
+}
+
+
+/*
+ * [OTP-3906]
+ * Solaris signal management gets confused when threads are used and a
+ * lot of child processes dies. The confusion results in that SIGCHLD
+ * signals aren't delivered to the emulator which in turn results in
+ * a lot of defunct processes in the system.
+ *
+ * The problem seems to appear when a signal is frequently
+ * blocked/unblocked at the same time as the signal is frequently
+ * propagated. The child waiter thread is a workaround for this problem.
+ * The SIGCHLD signal is always blocked (in all threads), and the child
+ * waiter thread fetches the signal by a call to sigwait(). See
+ * child_waiter().
+ *
+ * This should be a non-issue since the fork:ing was moved outside of
+ * the emulator into erl_child_setup. I'm leaving the comment here
+ * for posterity. */
+
+static void handle_sigchld(int sig) {
+    int buff[2], res;
+
+    sys_sigblock(SIGCHLD);
+
+    while ((buff[0] = waitpid((pid_t)(-1), buff+1, WNOHANG)) > 0) {
+        do {
+            res = write(sigchld_pipe[1], buff, sizeof(buff));
+        } while (res < 0 && errno == EINTR);
+        if (res <= 0)
+            ABORT("Failed to write to sigchld_pipe (%d): %d (%d)", sigchld_pipe[1], res, errno);
+        DEBUG_PRINT("Reap child %d (%d)", buff[0], buff[1]);
+    }
+
+    sys_sigrelease(SIGCHLD);
 }
 
 #if defined(__ANDROID__)
 static int system_properties_fd(void)
 {
-    int fd;
+    static int fd = -2;
     char *env;
 
+    if (fd != -2) return fd;
     env = getenv("ANDROID_PROPERTY_WORKSPACE");
     if (!env) {
+        fd = -1;
         return -1;
     }
     fd = atoi(env);
     return fd;
 }
 #endif /* __ANDROID__ */
+
+int
+main(int argc, char *argv[])
+{
+    /* This fd should be open from beam */
+    int uds_fd = 3, max_fd = 3;
+#ifndef HAVE_CLOSEFROM
+    int i;
+#endif
+    struct sigaction sa;
+
+    if (argc < 1 || sscanf(argv[1],"%d",&max_files) != 1) {
+        ABORT("Invalid arguments to child_setup");
+    }
+
+/* We close all fds except the uds from beam.
+   All other fds from now on will have the
+   CLOEXEC flags set on them. This means that we
+   only have to close a very limited number of fds
+   after we fork before the exec. */
+#if defined(HAVE_CLOSEFROM)
+    closefrom(4);
+#else
+    for (i = 4; i < max_files; i++)
+#if defined(__ANDROID__)
+        if (i != system_properties_fd())
+#endif
+        (void) close(i);
+#endif
+
+    if (pipe(sigchld_pipe) < 0) {
+        ABORT("Failed to setup sigchld pipe (%d)", errno);
+    }
+
+    SET_CLOEXEC(sigchld_pipe[0]);
+    SET_CLOEXEC(sigchld_pipe[1]);
+
+    max_fd = max_fd < sigchld_pipe[0] ? sigchld_pipe[0] : max_fd;
+
+    sa.sa_handler = &handle_sigchld;
+    sigemptyset(&sa.sa_mask);
+    sa.sa_flags = SA_RESTART | SA_NOCLDSTOP;
+    if (sigaction(SIGCHLD, &sa, 0) == -1) {
+        perror(0);
+        exit(1);
+    }
+
+    forker_hash_init();
+
+    SET_CLOEXEC(uds_fd);
+
+    DEBUG_PRINT("Starting forker %d", max_files);
+
+    while (1) {
+        fd_set read_fds;
+        int res;
+        FD_ZERO(&read_fds);
+        FD_SET(uds_fd, &read_fds);
+        FD_SET(sigchld_pipe[0], &read_fds);
+        DEBUG_PRINT("child_setup selecting on %d, %d (%d)",
+                uds_fd, sigchld_pipe[0], max_fd);
+        res = select(max_fd+1, &read_fds, NULL, NULL, NULL);
+
+        if (res < 0) {
+            if (errno == EINTR) continue;
+            ABORT("Select failed: %d (%d)",res, errno);
+        }
+
+        if (FD_ISSET(uds_fd, &read_fds)) {
+            int pipes[3], res, os_pid;
+            ErtsSysForkerProto proto;
+            errno = 0;
+            if ((res = sys_uds_read(uds_fd, (char*)&proto, sizeof(proto),
+                                    pipes, 3, MSG_DONTWAIT)) < 0) {
+                if (errno == EINTR)
+                    continue;
+                DEBUG_PRINT("erl_child_setup failed to read from uds: %d, %d", res, errno);
+                _exit(0);
+            }
+
+            if (res == 0) {
+                DEBUG_PRINT("uds was closed!");
+                _exit(0);
+            }
+            /* Since we use unix domain sockets and send the entire data in
+               one go we *should* get the entire payload at once. */
+            ASSERT(res == sizeof(proto));
+            ASSERT(proto.action == ErtsSysForkerProtoAction_Start);
+
+            sys_sigblock(SIGCHLD);
+
+            errno = 0;
+
+            os_pid = fork();
+            if (os_pid == 0)
+                start_new_child(pipes);
+
+            add_os_pid_to_port_id_mapping(proto.u.start.port_id, os_pid);
+
+            /* We write an ack here, but expect the reply on
+               the pipes[0] inside the fork */
+            proto.action = ErtsSysForkerProtoAction_Go;
+            proto.u.go.os_pid = os_pid;
+            proto.u.go.error_number = errno;
+            while (write(pipes[1], &proto, sizeof(proto)) < 0 && errno == EINTR)
+                ; /* remove gcc warning */
+
+#ifdef FORKER_PROTO_START_ACK
+            proto.action = ErtsSysForkerProtoAction_StartAck;
+            while (write(uds_fd, &proto, sizeof(proto)) < 0 && errno == EINTR)
+                ; /* remove gcc warning */
+#endif
+
+            sys_sigrelease(SIGCHLD);
+            close(pipes[0]);
+            close(pipes[1]);
+            close(pipes[2]);
+        }
+
+        if (FD_ISSET(sigchld_pipe[0], &read_fds)) {
+            int ibuff[2];
+            ErtsSysForkerProto proto;
+            res = read(sigchld_pipe[0], ibuff, sizeof(ibuff));
+            if (res <= 0) {
+                if (errno == EINTR)
+                    continue;
+                ABORT("Failed to read from sigchld pipe: %d (%d)", res, errno);
+            }
+
+            proto.u.sigchld.port_id = get_port_id((pid_t)(ibuff[0]));
+
+            if (proto.u.sigchld.port_id == THE_NON_VALUE)
+                continue; /* exit status report not requested */
+
+            proto.action = ErtsSysForkerProtoAction_SigChld;
+            proto.u.sigchld.error_number = ibuff[1];
+            DEBUG_PRINT("send %s to %d", buff, uds_fd);
+            if (write(uds_fd, &proto, sizeof(proto)) < 0) {
+                if (errno == EINTR)
+                    continue;
+                /* The uds was close, which most likely means that the VM
+                   has exited. This will be detected when we try to read
+                   from the uds_fd. */
+                DEBUG_PRINT("Failed to write to uds: %d (%d)", uds_fd, errno);
+            }
+        }
+    }
+    return 1;
+}
+
+typedef struct exit_status {
+    HashBucket hb;
+    pid_t os_pid;
+    Eterm port_id;
+} ErtsSysExitStatus;
+
+static Hash *forker_hash;
+
+static void add_os_pid_to_port_id_mapping(Eterm port_id, pid_t os_pid)
+{
+    if (port_id != THE_NON_VALUE) {
+        /* exit status report requested */
+        ErtsSysExitStatus es;
+        es.os_pid = os_pid;
+        es.port_id = port_id;
+        hash_put(forker_hash, &es);
+    }
+}
+
+static Eterm get_port_id(pid_t os_pid)
+{
+    ErtsSysExitStatus est, *es;
+    Eterm port_id;
+    est.os_pid = os_pid;
+    es = hash_remove(forker_hash, &est);
+    if (!es) return THE_NON_VALUE;
+    port_id = es->port_id;
+    free(es);
+    return port_id;
+}
+
+static int fcmp(void *a, void *b)
+{
+    ErtsSysExitStatus *sa = a;
+    ErtsSysExitStatus *sb = b;
+    return !(sa->os_pid == sb->os_pid);
+}
+
+static HashValue fhash(void *e)
+{
+    ErtsSysExitStatus *se = e;
+    Uint32 val = se->os_pid;
+    val = (val+0x7ed55d16) + (val<<12);
+    val = (val^0xc761c23c) ^ (val>>19);
+    val = (val+0x165667b1) + (val<<5);
+    val = (val+0xd3a2646c) ^ (val<<9);
+    val = (val+0xfd7046c5) + (val<<3);
+    val = (val^0xb55a4f09) ^ (val>>16);
+    return val;
+}
+
+static void *falloc(void *e)
+{
+    ErtsSysExitStatus *se = e;
+    ErtsSysExitStatus *ne = malloc(sizeof(ErtsSysExitStatus));
+    ne->os_pid = se->os_pid;
+    ne->port_id = se->port_id;
+    return ne;
+}
+
+static void *meta_alloc(int type, size_t size) { return malloc(size); }
+static void meta_free(int type, void *p)       { free(p); }
+
+static int forker_hash_init(void)
+{
+    HashFunctions forker_hash_functions;
+    forker_hash_functions.hash = fhash;
+    forker_hash_functions.cmp = fcmp;
+    forker_hash_functions.alloc = falloc;
+    forker_hash_functions.free = free;
+    forker_hash_functions.meta_alloc = meta_alloc;
+    forker_hash_functions.meta_free  = meta_free;
+    forker_hash_functions.meta_print = NULL;
+
+    forker_hash = hash_new(0, "forker_hash",
+                           16, forker_hash_functions);
+
+    return 1;
+}
diff --git a/erts/emulator/sys/unix/erl_child_setup.h b/erts/emulator/sys/unix/erl_child_setup.h
new file mode 100644
index 0000000000..a28b136bfc
--- /dev/null
+++ b/erts/emulator/sys/unix/erl_child_setup.h
@@ -0,0 +1,77 @@
+/*
+ * %CopyrightBegin%
+ * 
+ * Copyright Ericsson AB 2015-2015. All Rights Reserved.
+ * 
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ * 
+ * %CopyrightEnd%
+ *
+ * This file defines the interface inbetween erts and child_setup.
+ */
+
+#ifndef _ERL_UNIX_FORKER_H
+#define _ERL_UNIX_FORKER_H
+
+#include "sys.h"
+
+#ifdef __FreeBSD__
+/* The freebsd sendmsg man page explicitly states that
+   you should not close fds before they are known
+   to have reached the other side, so this Ack protects
+   against that. */
+#define FORKER_PROTO_START_ACK 1
+#endif
+
+#define FORKER_ARGV_NO_OF_ARGS  3
+#define FORKER_ARGV_PROGNAME_IX	0    /* Program name                          */
+#define FORKER_ARGV_MAX_FILES	1    /* max_files                             */
+
+#define FORKER_FLAG_USE_STDIO   (1 << 0)    /* dup the pipe to stdin/stderr   */
+#define FORKER_FLAG_EXIT_STATUS (1 << 1)    /* send the exit status to parent */
+#define FORKER_FLAG_DO_READ     (1 << 2)    /* dup write fd */
+#define FORKER_FLAG_DO_WRITE    (1 << 3)    /* dup read  fd  */
+
+#if SIZEOF_VOID_P == SIZEOF_LONG
+typedef unsigned long ErtsSysPortId;
+#elif SIZEOF_VOID_P == SIZEOF_INT
+typedef unsigned int ErtsSysPortId;
+#elif SIZEOF_VOID_P == SIZEOF_LONG_LONG
+typedef unsigned long long ErtsSysPortId;
+#endif
+
+typedef struct ErtsSysForkerProto_ {
+    enum {
+        ErtsSysForkerProtoAction_Start,
+        ErtsSysForkerProtoAction_StartAck,
+        ErtsSysForkerProtoAction_Go,
+        ErtsSysForkerProtoAction_SigChld,
+        ErtsSysForkerProtoAction_Ack
+    } action;
+    union {
+        struct {
+            ErtsSysPortId port_id;
+            int fds[3];
+        } start;
+        struct {
+            pid_t os_pid;
+            int error_number;
+        } go;
+        struct {
+            ErtsSysPortId port_id;
+            int error_number;
+        } sigchld;
+    } u;
+} ErtsSysForkerProto;
+
+#endif /* #ifndef _ERL_UNIX_FORKER_H */
diff --git a/erts/emulator/sys/unix/erl_unix_sys.h b/erts/emulator/sys/unix/erl_unix_sys.h
index 8d4e98bf3a..0352ee1b3c 100644
--- a/erts/emulator/sys/unix/erl_unix_sys.h
+++ b/erts/emulator/sys/unix/erl_unix_sys.h
@@ -30,9 +30,7 @@
 #include <limits.h>
 #include <stdlib.h>
 #include <string.h>
-#ifndef QNX
 #include <memory.h>
-#endif
 
 #if defined(__sun__) && defined(__SVR4) && !defined(__EXTENSIONS__)
 #   define __EXTENSIONS__
@@ -92,11 +90,6 @@
 #include <ieeefp.h>
 #endif
 
-#ifdef QNX
-#include <process.h>
-#include <sys/qnx_glob.h>
-#endif
-
 #include <pwd.h>
 
 #ifndef HZ
@@ -136,13 +129,6 @@
 #  define ERTS_POLL_NEED_ASYNC_INTERRUPT_SUPPORT
 #endif
 
-#ifndef ENABLE_CHILD_WAITER_THREAD
-#  ifdef ERTS_SMP
-#    define ERTS_SMP_SCHEDULERS_NEED_TO_CHECK_CHILDREN
-void erts_check_children(void);
-#  endif
-#endif
-
 typedef void *GETENV_STATE;
 
 /*
@@ -310,7 +296,6 @@ typedef void (*SIGFUNC)(int);
 extern SIGFUNC sys_signal(int, SIGFUNC);
 extern void sys_sigrelease(int);
 extern void sys_sigblock(int);
-extern void sys_stop_cat(void);
 
 /*
  * Handling of floating point exceptions.
@@ -425,19 +410,6 @@ void erts_sys_unblock_fpe(int);
 #define ERTS_FP_ERROR_THOROUGH(p, f, A)	__ERTS_FP_ERROR_THOROUGH(&(p)->fp_exception, f, A)
 
 
-#ifdef NEED_CHILD_SETUP_DEFINES
-/* The child setup argv[] */
-#define CS_ARGV_PROGNAME_IX	0		/* Program name		*/
-#define CS_ARGV_UNBIND_IX	1		/* Unbind from cpu	*/
-#define CS_ARGV_WD_IX		2		/* Working directory	*/
-#define CS_ARGV_CMD_IX		3		/* Command		*/
-#define CS_ARGV_FD_CR_IX	4		/* Fd close range	*/
-#define CS_ARGV_DUP2_OP_IX(N)	((N) + 5)	/* dup2 operations	*/
-
-#define CS_ARGV_NO_OF_DUP2_OPS	3		/* Number of dup2 ops	*/
-#define CS_ARGV_NO_OF_ARGS	8		/* Number of arguments	*/
-#endif /* #ifdef NEED_CHILD_SETUP_DEFINES */
-
 /* Threads */
 #ifdef USE_THREADS
 extern int init_async(int);
diff --git a/erts/emulator/sys/unix/sys.c b/erts/emulator/sys/unix/sys.c
index b036b20b7b..2ad5f3b4d5 100644
--- a/erts/emulator/sys/unix/sys.c
+++ b/erts/emulator/sys/unix/sys.c
@@ -49,7 +49,6 @@
 #include <sys/ioctl.h>
 #endif
 
-#define NEED_CHILD_SETUP_DEFINES
 #define ERTS_WANT_BREAK_HANDLING
 #define ERTS_WANT_GOT_SIGUSR1
 #define WANT_NONBLOCKING    /* must define this to pull in defs from sys.h */
@@ -67,7 +66,7 @@
 #include "erl_mseg.h"
 
 extern char **environ;
-static erts_smp_rwmtx_t environ_rwmtx;
+erts_smp_rwmtx_t environ_rwmtx;
 
 #define MAX_VSIZE 16		/* Max number of entries allowed in an I/O
 				 * vector sock_sendv().
@@ -76,89 +75,12 @@ static erts_smp_rwmtx_t environ_rwmtx;
  * Don't need global.h, but bif_table.h (included by bif.h),
  * won't compile otherwise
  */
-#include "global.h" 
+#include "global.h"
 #include "bif.h"
 
-#include "erl_sys_driver.h"
 #include "erl_check_io.h"
 #include "erl_cpu_topology.h"
 
-#ifndef DISABLE_VFORK
-#define DISABLE_VFORK 0
-#endif
-
-#if defined IOV_MAX
-#define MAXIOV IOV_MAX
-#elif defined UIO_MAXIOV
-#define MAXIOV UIO_MAXIOV
-#else
-#define MAXIOV 16
-#endif
-
-#ifdef USE_THREADS
-#  ifdef ENABLE_CHILD_WAITER_THREAD
-#    define CHLDWTHR ENABLE_CHILD_WAITER_THREAD
-#  else
-#    define CHLDWTHR 0
-#  endif
-#  define FDBLOCK 1
-#else
-#  define CHLDWTHR 0
-#  define FDBLOCK 0
-#endif
-/*
- * [OTP-3906]
- * Solaris signal management gets confused when threads are used and a
- * lot of child processes dies. The confusion results in that SIGCHLD
- * signals aren't delivered to the emulator which in turn results in
- * a lot of defunct processes in the system.
- *
- * The problem seems to appear when a signal is frequently
- * blocked/unblocked at the same time as the signal is frequently
- * propagated. The child waiter thread is a workaround for this problem.
- * The SIGCHLD signal is always blocked (in all threads), and the child
- * waiter thread fetches the signal by a call to sigwait(). See
- * child_waiter().
- */
-
-typedef struct ErtsSysReportExit_ ErtsSysReportExit;
-struct ErtsSysReportExit_ {
-    ErtsSysReportExit *next;
-    Eterm port;
-    int pid;
-    int ifd;
-    int ofd;
-#if CHLDWTHR && !defined(ERTS_SMP)
-    int status;
-#endif
-};
-
-/* Used by the fd driver iff the fd could not be set to non-blocking */
-typedef struct ErtsSysBlocking_ {
-    ErlDrvPDL pdl;
-    int res;
-    int err;
-    unsigned int pkey;
-} ErtsSysBlocking;
-
-
-/* This data is shared by these drivers - initialized by spawn_init() */
-static struct driver_data {
-    ErlDrvPort port_num;
-    int ofd, packet_bytes;
-    ErtsSysReportExit *report_exit;
-    int pid;
-    int alive;
-    int status;
-    int terminating;
-    ErtsSysBlocking *blocking;
-} *driver_data;			/* indexed by fd */
-
-static ErtsSysReportExit *report_exit_list;
-#if CHLDWTHR && !defined(ERTS_SMP)
-static ErtsSysReportExit *report_exit_transit_list;
-#endif
-
 extern int  driver_interrupt(int, int);
 extern void do_break(void);
 
@@ -170,33 +92,6 @@ extern void erts_sys_init_float(void);
 
 extern void erl_crash_dump(char* file, int line, char* fmt, ...);
 
-#define DIR_SEPARATOR_CHAR    '/'
-
-#if defined(__ANDROID__)
-#define SHELL "/system/bin/sh"
-#else
-#define SHELL "/bin/sh"
-#endif /* __ANDROID__ */
-
-
-#if defined(DEBUG)
-#define ERL_BUILD_TYPE_MARKER ".debug"
-#elif defined(PURIFY)
-#define ERL_BUILD_TYPE_MARKER ".purify"
-#elif defined(QUANTIFY)
-#define ERL_BUILD_TYPE_MARKER ".quantify"
-#elif defined(PURECOV)
-#define ERL_BUILD_TYPE_MARKER ".purecov"
-#elif defined(VALGRIND)
-#define ERL_BUILD_TYPE_MARKER ".valgrind"
-#else /* opt */
-#define ERL_BUILD_TYPE_MARKER
-#endif
-
-#define CHILD_SETUP_PROG_NAME	"child_setup" ERL_BUILD_TYPE_MARKER
-#if !DISABLE_VFORK
-static char *child_setup_prog;
-#endif
 
 #ifdef DEBUG
 static int debug_log = 0;
@@ -220,7 +115,7 @@ static volatile int have_prepared_crash_dump;
   (have_prepared_crash_dump++)
 #endif
 
-static erts_smp_atomic_t sys_misc_mem_sz;
+erts_smp_atomic_t sys_misc_mem_sz;
 
 #if defined(ERTS_SMP)
 static void smp_sig_notify(char c);
@@ -233,46 +128,6 @@ static int sig_suspend_fds[2] = {-1, -1};
 
 jmp_buf erts_sys_sigsegv_jmp;
 
-#if CHLDWTHR || defined(ERTS_SMP)
-erts_mtx_t chld_stat_mtx;
-#endif
-#if CHLDWTHR
-static erts_tid_t child_waiter_tid;
-/* chld_stat_mtx is used to protect against concurrent accesses
-   of the driver_data fields pid, alive, and status. */
-erts_cnd_t chld_stat_cnd;
-static long children_alive;
-#define CHLD_STAT_LOCK		erts_mtx_lock(&chld_stat_mtx)
-#define CHLD_STAT_UNLOCK	erts_mtx_unlock(&chld_stat_mtx)
-#define CHLD_STAT_WAIT		erts_cnd_wait(&chld_stat_cnd, &chld_stat_mtx)
-#define CHLD_STAT_SIGNAL	erts_cnd_signal(&chld_stat_cnd)
-#elif defined(ERTS_SMP) /* ------------------------------------------------- */
-#define CHLD_STAT_LOCK		erts_mtx_lock(&chld_stat_mtx)
-#define CHLD_STAT_UNLOCK	erts_mtx_unlock(&chld_stat_mtx)
-
-#else /* ------------------------------------------------------------------- */
-#define CHLD_STAT_LOCK
-#define CHLD_STAT_UNLOCK
-static volatile int children_died;
-#endif
-
-
-static struct fd_data {
-    char  pbuf[4];   /* hold partial packet bytes */
-    int   psz;       /* size of pbuf */
-    char  *buf;
-    char  *cpos;
-    int   sz;
-    int   remain;  /* for input on fd */
-} *fd_data;			/* indexed by fd */
-
-/* static FUNCTION(int, write_fill, (int, char*, int)); unused? */
-static void note_child_death(int, int);
-
-#if CHLDWTHR
-static void* child_waiter(void *);
-#endif
-
 static int crashdump_companion_cube_fd = -1;
 
 /********************* General functions ****************************/
@@ -453,9 +308,10 @@ MALLOC_USE_HASH(1);
 #ifdef USE_THREADS
 
 #ifdef ERTS_THR_HAVE_SIG_FUNCS
+
 /*
  * Child thread inherits parents signal mask at creation. In order to
- * guarantee that the main thread will receive all SIGINT, SIGCHLD, and
+ * guarantee that the main thread will receive all SIGINT, and
  * SIGUSR1 signals sent to the process, we block these signals in the
  * parent thread when creating a new thread.
  */
@@ -551,14 +407,11 @@ erts_sys_pre_init(void)
 #ifdef ERTS_THR_HAVE_SIG_FUNCS
     sigemptyset(&thr_create_sigmask);
     sigaddset(&thr_create_sigmask, SIGINT);   /* block interrupt */
-    sigaddset(&thr_create_sigmask, SIGCHLD);  /* block child signals */
     sigaddset(&thr_create_sigmask, SIGUSR1);  /* block user defined signal */
 #endif
 
     erts_thr_init(&eid);
 
-    report_exit_list = NULL;
-
 #ifdef ERTS_ENABLE_LOCK_COUNT
     erts_lcnt_init();
 #endif
@@ -569,17 +422,6 @@ erts_sys_pre_init(void)
 
 #ifdef USE_THREADS
 
-#if CHLDWTHR || defined(ERTS_SMP)
-    erts_mtx_init(&chld_stat_mtx, "child_status");
-#endif
-#if CHLDWTHR
-#ifndef ERTS_SMP
-    report_exit_transit_list = NULL;
-#endif
-    erts_cnd_init(&chld_stat_cnd);
-    children_alive = 0;
-#endif
-
 #ifdef ERTS_SMP
     erts_smp_atomic32_init_nob(&erts_break_requested, 0);
     erts_smp_atomic32_init_nob(&erts_got_sigusr1, 0);
@@ -589,9 +431,6 @@ erts_sys_pre_init(void)
     erts_got_sigusr1 = 0;
     have_prepared_crash_dump = 0;
 #endif
-#if !CHLDWTHR && !defined(ERTS_SMP)
-    children_died = 0;
-#endif
 
 #endif /* USE_THREADS */
 
@@ -628,39 +467,6 @@ erts_sys_pre_init(void)
 void
 erl_sys_init(void)
 {
-#if !DISABLE_VFORK
- {
-    int res;
-    char bindir[MAXPATHLEN];
-    size_t bindirsz = sizeof(bindir);
-    Uint csp_path_sz;
-
-    res = erts_sys_getenv_raw("BINDIR", bindir, &bindirsz);
-    if (res != 0) {
-	if (res < 0)
-	    erl_exit(-1,
-		     "Environment variable BINDIR is not set\n");
-	if (res > 0)
-	    erl_exit(-1,
-		     "Value of environment variable BINDIR is too large\n");
-    }
-    if (bindir[0] != DIR_SEPARATOR_CHAR)
-	erl_exit(-1,
-		 "Environment variable BINDIR does not contain an"
-		 " absolute path\n");
-    csp_path_sz = (strlen(bindir)
-		   + 1 /* DIR_SEPARATOR_CHAR */
-		   + sizeof(CHILD_SETUP_PROG_NAME)
-		   + 1);
-    child_setup_prog = erts_alloc(ERTS_ALC_T_CS_PROG_PATH, csp_path_sz);
-    erts_smp_atomic_add_nob(&sys_misc_mem_sz, csp_path_sz);
-    erts_snprintf(child_setup_prog, csp_path_sz,
-            "%s%c%s",
-            bindir,
-            DIR_SEPARATOR_CHAR,
-            CHILD_SETUP_PROG_NAME);
- }
-#endif
 
 #ifdef USE_SETLINEBUF
     setlinebuf(stdout);
@@ -978,43 +784,6 @@ int sys_max_files(void)
    return(max_files);
 }
 
-static void block_signals(void)
-{
-#if !CHLDWTHR
-   sys_sigblock(SIGCHLD);
-#endif
-#ifndef ERTS_SMP
-   sys_sigblock(SIGINT);
-#ifndef ETHR_UNUSABLE_SIGUSRX
-   sys_sigblock(SIGUSR1);
-#endif /* #ifndef ETHR_UNUSABLE_SIGUSRX */
-#endif /* #ifndef ERTS_SMP */
-
-#if defined(ERTS_SMP) && !defined(ETHR_UNUSABLE_SIGUSRX)
-   sys_sigblock(ERTS_SYS_SUSPEND_SIGNAL);
-#endif
-
-}
-
-static void unblock_signals(void)
-{
-    /* Update erl_child_setup.c if changed */
-#if !CHLDWTHR
-   sys_sigrelease(SIGCHLD);
-#endif
-#ifndef ERTS_SMP
-   sys_sigrelease(SIGINT);
-#ifndef ETHR_UNUSABLE_SIGUSRX
-   sys_sigrelease(SIGUSR1);
-#endif /* #ifndef ETHR_UNUSABLE_SIGUSRX */
-#endif /* #ifndef ERTS_SMP */
-
-#if defined(ERTS_SMP) && !defined(ETHR_UNUSABLE_SIGUSRX)
-   sys_sigrelease(ERTS_SYS_SUSPEND_SIGNAL);
-#endif
-
-}
-
 /************************** OS info *******************************/
 
 /* Used by erlang:info/1. */
@@ -1102,1494 +871,6 @@ void fini_getenv_state(GETENV_STATE *state)
    erts_smp_rwmtx_runlock(&environ_rwmtx);
 }
 
-
-/************************** Port I/O *******************************/
-
-
-
-/* I. Common stuff */
-
-/*
- * Decreasing the size of it below 16384 is not allowed.
- */
-
-/* II. The spawn/fd/vanilla drivers */
-
-#define ERTS_SYS_READ_BUF_SZ (64*1024)
-
-/* Driver interfaces */
-static ErlDrvData spawn_start(ErlDrvPort, char*, SysDriverOpts*);
-static ErlDrvData fd_start(ErlDrvPort, char*, SysDriverOpts*);
-#if FDBLOCK
-static void fd_async(void *);
-static void fd_ready_async(ErlDrvData drv_data, ErlDrvThreadData thread_data);
-#endif
-static ErlDrvSSizeT fd_control(ErlDrvData, unsigned int, char *, ErlDrvSizeT,
-			       char **, ErlDrvSizeT);
-static ErlDrvData vanilla_start(ErlDrvPort, char*, SysDriverOpts*);
-static int spawn_init(void);
-static void fd_stop(ErlDrvData);
-static void fd_flush(ErlDrvData);
-static void stop(ErlDrvData);
-static void ready_input(ErlDrvData, ErlDrvEvent);
-static void ready_output(ErlDrvData, ErlDrvEvent);
-static void output(ErlDrvData, char*, ErlDrvSizeT);
-static void outputv(ErlDrvData, ErlIOVec*);
-static void stop_select(ErlDrvEvent, void*);
-
-struct erl_drv_entry spawn_driver_entry = {
-    spawn_init,
-    spawn_start,
-    stop,
-    output,
-    ready_input,
-    ready_output,
-    "spawn",
-    NULL,
-    NULL,
-    NULL,
-    NULL,
-    NULL,
-    NULL,
-    NULL,
-    NULL,
-    NULL,
-    ERL_DRV_EXTENDED_MARKER,
-    ERL_DRV_EXTENDED_MAJOR_VERSION,
-    ERL_DRV_EXTENDED_MINOR_VERSION,
-    ERL_DRV_FLAG_USE_PORT_LOCKING,
-    NULL, NULL,
-    stop_select
-};
-struct erl_drv_entry fd_driver_entry = {
-    NULL,
-    fd_start,
-    fd_stop,
-    output,
-    ready_input,
-    ready_output, 
-    "fd",
-    NULL,
-    NULL,
-    fd_control,
-    NULL,
-    outputv,
-#if FDBLOCK
-    fd_ready_async, /* ready_async */
-#else
-    NULL,
-#endif
-    fd_flush, /* flush */
-    NULL, /* call */
-    NULL, /* event */
-    ERL_DRV_EXTENDED_MARKER,
-    ERL_DRV_EXTENDED_MAJOR_VERSION,
-    ERL_DRV_EXTENDED_MINOR_VERSION,
-    0, /* ERL_DRV_FLAGs */
-    NULL, /* handle2 */
-    NULL, /* process_exit */
-    stop_select
-};
-struct erl_drv_entry vanilla_driver_entry = {
-    NULL,
-    vanilla_start,
-    stop,
-    output,
-    ready_input,
-    ready_output,
-    "vanilla",
-    NULL,
-    NULL,
-    NULL,
-    NULL,
-    NULL,
-    NULL,
-    NULL, /* flush */
-    NULL, /* call */
-    NULL, /* event */
-    ERL_DRV_EXTENDED_MARKER,
-    ERL_DRV_EXTENDED_MAJOR_VERSION,
-    ERL_DRV_EXTENDED_MINOR_VERSION,
-    0, /* ERL_DRV_FLAGs */
-    NULL, /* handle2 */
-    NULL, /* process_exit */
-    stop_select
-};
-
-/* Handle SIGCHLD signals. */
-#if (defined(SIG_SIGSET) || defined(SIG_SIGNAL))
-static RETSIGTYPE onchld(void)
-#else
-static RETSIGTYPE onchld(int signum)
-#endif
-{
-#if CHLDWTHR
-    ASSERT(0); /* We should *never* catch a SIGCHLD signal */
-#elif defined(ERTS_SMP)
-    smp_sig_notify('C');
-#else
-    children_died = 1;
-    ERTS_CHK_IO_AS_INTR(); /* Make sure we don't sleep in poll */
-#endif
-}
-
-static int set_blocking_data(struct driver_data *dd) {
-
-    dd->blocking = erts_alloc(ERTS_ALC_T_SYS_BLOCKING, sizeof(ErtsSysBlocking));
-
-    erts_smp_atomic_add_nob(&sys_misc_mem_sz, sizeof(ErtsSysBlocking));
-
-    dd->blocking->pdl = driver_pdl_create(dd->port_num);
-    dd->blocking->res = 0;
-    dd->blocking->err = 0;
-    dd->blocking->pkey = driver_async_port_key(dd->port_num);
-
-    return 1;
-}
-
-static int set_driver_data(ErlDrvPort port_num,
-			   int ifd,
-			   int ofd,
-			   int packet_bytes,
-			   int read_write,
-			   int exit_status,
-			   int pid,
-                           int is_blocking)
-{
-    Port *prt;
-    ErtsSysReportExit *report_exit;
-
-    if (!exit_status)
-	report_exit = NULL;
-    else {
-	report_exit = erts_alloc(ERTS_ALC_T_PRT_REP_EXIT,
-				 sizeof(ErtsSysReportExit));
-	report_exit->next = report_exit_list;
-	report_exit->port = erts_drvport2id(port_num);
-	report_exit->pid = pid;
-	report_exit->ifd = read_write & DO_READ ? ifd : -1;
-	report_exit->ofd = read_write & DO_WRITE ? ofd : -1;
-#if CHLDWTHR && !defined(ERTS_SMP)
-	report_exit->status = 0;
-#endif
-	report_exit_list = report_exit;
-    }
-
-    prt = erts_drvport2port(port_num);
-    if (prt != ERTS_INVALID_ERL_DRV_PORT)
-	prt->os_pid = pid;
-
-    if (read_write & DO_READ) {
-	driver_data[ifd].packet_bytes = packet_bytes;
-	driver_data[ifd].port_num = port_num;
-	driver_data[ifd].report_exit = report_exit;
-	driver_data[ifd].pid = pid;
-	driver_data[ifd].alive = 1;
-	driver_data[ifd].status = 0;
-        driver_data[ifd].terminating = 0;
-        driver_data[ifd].blocking = NULL;
-	if (read_write & DO_WRITE) {
-	    driver_data[ifd].ofd = ofd;
-            if (is_blocking && FDBLOCK)
-                if (!set_blocking_data(driver_data+ifd))
-                    return -1;
-	    if (ifd != ofd)
-		driver_data[ofd] = driver_data[ifd];  /* structure copy */
-	} else {		/* DO_READ only */
-	    driver_data[ifd].ofd = -1;
-	}
-	(void) driver_select(port_num, ifd, (ERL_DRV_READ|ERL_DRV_USE), 1);
-	return(ifd);
-    } else {			/* DO_WRITE only */
-	driver_data[ofd].packet_bytes = packet_bytes;
-	driver_data[ofd].port_num = port_num;
-	driver_data[ofd].report_exit = report_exit;
-	driver_data[ofd].ofd = ofd;
-	driver_data[ofd].pid = pid;
-	driver_data[ofd].alive = 1;
-	driver_data[ofd].status = 0;
-        driver_data[ofd].terminating = 0;
-        driver_data[ofd].blocking = NULL;
-        if (is_blocking && FDBLOCK)
-            if (!set_blocking_data(driver_data+ofd))
-                return -1;
-	return(ofd);
-    }
-}
-
-static int spawn_init()
-{
-   int i;
-#if CHLDWTHR
-   erts_thr_opts_t thr_opts = ERTS_THR_OPTS_DEFAULT_INITER;
-
-   thr_opts.detached = 0;
-   thr_opts.suggested_stack_size = 0; /* Smallest possible */
-   thr_opts.name = "child_waiter";
-#endif
-
-   sys_signal(SIGPIPE, SIG_IGN); /* Ignore - we'll handle the write failure */
-   driver_data = (struct driver_data *)
-       erts_alloc(ERTS_ALC_T_DRV_TAB, max_files * sizeof(struct driver_data));
-   erts_smp_atomic_add_nob(&sys_misc_mem_sz,
-			   max_files * sizeof(struct driver_data));
-
-   for (i = 0; i < max_files; i++)
-      driver_data[i].pid = -1;
-
-#if CHLDWTHR
-   sys_sigblock(SIGCHLD);
-#endif
-
-   sys_signal(SIGCHLD, onchld); /* Reap children */
-
-#if CHLDWTHR
-   erts_thr_create(&child_waiter_tid, child_waiter, NULL, &thr_opts);
-#endif
-
-   return 1;
-}
-
-static void close_pipes(int ifd[2], int ofd[2], int read_write)
-{
-    if (read_write & DO_READ) {
-	(void) close(ifd[0]);
-	(void) close(ifd[1]);
-    }
-    if (read_write & DO_WRITE) {
-	(void) close(ofd[0]);
-	(void) close(ofd[1]);
-    }
-}
-
-static void init_fd_data(int fd, ErlDrvPort port_num)
-{
-    fd_data[fd].buf = NULL;
-    fd_data[fd].cpos = NULL;
-    fd_data[fd].remain = 0;
-    fd_data[fd].sz = 0;
-    fd_data[fd].psz = 0;
-}
-
-static char **build_unix_environment(char *block)
-{
-    int i;
-    int j;
-    int len;
-    char *cp;
-    char **cpp;
-    char** old_env;
-    
-    ERTS_SMP_LC_ASSERT(erts_smp_lc_rwmtx_is_rlocked(&environ_rwmtx));
-
-    cp = block;
-    len = 0;
-    while (*cp != '\0') {
-	cp += strlen(cp) + 1;
-	len++;
-    }
-    old_env = environ;
-    while (*old_env++ != NULL) {
-	len++;
-    }
-    
-    cpp = (char **) erts_alloc_fnf(ERTS_ALC_T_ENVIRONMENT,
-				   sizeof(char *) * (len+1));
-    if (cpp == NULL) {
-	return NULL;
-    }
-
-    cp = block;
-    len = 0;
-    while (*cp != '\0') {
-	cpp[len] = cp;
-	cp += strlen(cp) + 1;
-	len++;
-    }
-    
-    i = len;
-    for (old_env = environ; *old_env; old_env++) {
-	char* old = *old_env;
-
-	for (j = 0; j < len; j++) {
-	    char *s, *t;
-
-	    s = cpp[j];
-	    t = old;
-	    while (*s == *t && *s != '=') {
-		s++, t++;
-	    }
-	    if (*s == '=' && *t == '=') {
-		break;
-	    }
-	}
-
-	if (j == len) {		/* New version not found */
-	    cpp[len++] = old;
-	}
-    }
-
-    for (j = 0; j < i; ) {
-        size_t last = strlen(cpp[j])-1;
-	if (cpp[j][last] == '=' && strchr(cpp[j], '=') == cpp[j]+last) {
-	    cpp[j] = cpp[--len];
-	    if (len < i) {
-		i--;
-	    } else {
-		j++;
-	    }
-	}
-	else {
-	    j++;
-	}
-    }
-
-    cpp[len] = NULL;
-    return cpp;
-}
-
-/*
-  [arndt] In most Unix systems, including Solaris 2.5, 'fork' allocates memory
-  in swap space for the child of a 'fork', whereas 'vfork' does not do this.
-  The natural call to use here is therefore 'vfork'. Due to a bug in
-  'vfork' in Solaris 2.5 (apparently fixed in 2.6), using 'vfork'
-  can be dangerous in what seems to be these circumstances:
-      If the child code under a vfork sets the signal action to SIG_DFL
-      (or SIG_IGN)
-      for any signal which was previously set to a signal handler, the
-      state of the parent is clobbered, so that the later arrival of
-      such a signal yields a sigsegv in the parent. If the signal was
-      not set to a signal handler, but ignored, all seems to work.
-  If you change the forking code below, beware of this.
- */
-
-static ErlDrvData spawn_start(ErlDrvPort port_num, char* name, SysDriverOpts* opts)
-{
-#define CMD_LINE_PREFIX_STR "exec "
-#define CMD_LINE_PREFIX_STR_SZ (sizeof(CMD_LINE_PREFIX_STR) - 1)
-
-    int ifd[2], ofd[2], len, pid, i;
-    char **volatile new_environ; /* volatile since a vfork() then cannot
-				    cause 'new_environ' to be clobbered
-				    in the parent process. */
-    int saved_errno;
-    long res;
-    char *cmd_line;
-#ifndef QNX
-    int unbind;
-#endif
-#if !DISABLE_VFORK
-    int no_vfork;
-    size_t no_vfork_sz = sizeof(no_vfork);
-
-    no_vfork = (erts_sys_getenv_raw("ERL_NO_VFORK",
-				    (char *) &no_vfork,
-				    &no_vfork_sz) >= 0);
-#endif
-
-    switch (opts->read_write) {
-    case DO_READ:
-	if (pipe(ifd) < 0)
-	    return ERL_DRV_ERROR_ERRNO;
-	if (ifd[0] >= max_files) {
-	    close_pipes(ifd, ofd, opts->read_write);
-	    errno = EMFILE;
-	    return ERL_DRV_ERROR_ERRNO;
-	}
-	ofd[1] = -1;		/* keep purify happy */
-	break;
-    case DO_WRITE:
-	if (pipe(ofd) < 0) return ERL_DRV_ERROR_ERRNO;
-	if (ofd[1] >= max_files) {
-	    close_pipes(ifd, ofd, opts->read_write);
-	    errno = EMFILE;
-	    return ERL_DRV_ERROR_ERRNO;
-	}
-	ifd[0] = -1;		/* keep purify happy */
-	break;
-    case DO_READ|DO_WRITE:
-	if (pipe(ifd) < 0) return ERL_DRV_ERROR_ERRNO;
-	errno = EMFILE;		/* default for next two conditions */
-	if (ifd[0] >= max_files || pipe(ofd) < 0) {
-	    close_pipes(ifd, ofd, DO_READ);
-	    return ERL_DRV_ERROR_ERRNO;
-	}
-	if (ofd[1] >= max_files) {
-	    close_pipes(ifd, ofd, opts->read_write);
-	    errno = EMFILE;
-	    return ERL_DRV_ERROR_ERRNO;
-	}
-	break;
-    default:
-	ASSERT(0);
-	return ERL_DRV_ERROR_GENERAL;
-    }
-
-    if (opts->spawn_type == ERTS_SPAWN_EXECUTABLE) {
-	/* started with spawn_executable, not with spawn */
-	len = strlen(name);
-	cmd_line = (char *) erts_alloc_fnf(ERTS_ALC_T_TMP, len + 1);
-	if (!cmd_line) {
-	    close_pipes(ifd, ofd, opts->read_write);
-	    errno = ENOMEM;
-	    return ERL_DRV_ERROR_ERRNO;
-	}
-	memcpy((void *) cmd_line,(void *) name, len);
-	cmd_line[len] = '\0';
-	if (access(cmd_line,X_OK) != 0) {
-	    int save_errno = errno;
-	    erts_free(ERTS_ALC_T_TMP, cmd_line);
-	    errno = save_errno;
-	    return ERL_DRV_ERROR_ERRNO;
-	}
-    } else {
-	/* make the string suitable for giving to "sh" */
-	len = strlen(name);
-	cmd_line = (char *) erts_alloc_fnf(ERTS_ALC_T_TMP,
-					   CMD_LINE_PREFIX_STR_SZ + len + 1);
-	if (!cmd_line) {
-	    close_pipes(ifd, ofd, opts->read_write);
-	    errno = ENOMEM;
-	    return ERL_DRV_ERROR_ERRNO;
-	}
-	memcpy((void *) cmd_line,
-	       (void *) CMD_LINE_PREFIX_STR,
-	       CMD_LINE_PREFIX_STR_SZ);
-	memcpy((void *) (cmd_line + CMD_LINE_PREFIX_STR_SZ), (void *) name, len);
-	cmd_line[CMD_LINE_PREFIX_STR_SZ + len] = '\0';
-    }
-
-    erts_smp_rwmtx_rlock(&environ_rwmtx);
-
-    if (opts->envir == NULL) {
-	new_environ = environ;
-    } else if ((new_environ = build_unix_environment(opts->envir)) == NULL) {
-	erts_smp_rwmtx_runlock(&environ_rwmtx);
-	erts_free(ERTS_ALC_T_TMP, (void *) cmd_line);
-	errno = ENOMEM;
-	return ERL_DRV_ERROR_ERRNO;
-    } 
-
-#ifndef QNX
-    /* Block child from SIGINT and SIGUSR1. Must be before fork()
-       to be safe. */
-    block_signals();
-
-    CHLD_STAT_LOCK;
-
-    unbind = erts_sched_bind_atfork_prepare();
-
-#if !DISABLE_VFORK
-    /* See fork/vfork discussion before this function. */
-    if (no_vfork) {
-#endif
-
-	DEBUGF(("Using fork\n"));
-	pid = fork();
-
-	if (pid == 0) {
-	    /* The child! Setup child... */
-
-	    if (erts_sched_bind_atfork_child(unbind) != 0)
-		goto child_error;
-
-	    /* OBSERVE!
-	     * Keep child setup after vfork() (implemented below and in
-	     * erl_child_setup.c) up to date if changes are made here.
-	     */
-
-	    if (opts->use_stdio) {
-		if (opts->read_write & DO_READ) {
-		    /* stdout for process */
-		    if (dup2(ifd[1], 1) < 0)
-			goto child_error;
-		    if(opts->redir_stderr)
-			/* stderr for process */
-			if (dup2(ifd[1], 2) < 0)
-			    goto child_error;
-		}
-		if (opts->read_write & DO_WRITE)
-		    /* stdin for process */
-		    if (dup2(ofd[0], 0) < 0)
-			goto child_error;
-	    }
-	    else {	/* XXX will fail if ofd[0] == 4 (unlikely..) */
-		if (opts->read_write & DO_READ)
-		    if (dup2(ifd[1], 4) < 0)
-			goto child_error;
-		if (opts->read_write & DO_WRITE)
-		    if (dup2(ofd[0], 3) < 0)
-			goto child_error;
-	    }
-
-#if defined(HAVE_CLOSEFROM)
-	    closefrom(opts->use_stdio ? 3 : 5);
-#else
-	    for (i = opts->use_stdio ? 3 : 5; i < max_files; i++)
-		(void) close(i);
-#endif
-
-	    if (opts->wd && chdir(opts->wd) < 0)
-		goto child_error;
-
-#if defined(USE_SETPGRP_NOARGS)		/* SysV */
-	    (void) setpgrp();
-#elif defined(USE_SETPGRP)		/* BSD */
-	    (void) setpgrp(0, getpid());
-#else					/* POSIX */
-	    (void) setsid();
-#endif
-	    
-	    unblock_signals();
-
-	    if (opts->spawn_type == ERTS_SPAWN_EXECUTABLE) {
-		if (opts->argv == NULL) {
-		    execle(cmd_line,cmd_line,(char *) NULL, new_environ);
-		} else {
-		    if (opts->argv[0] == erts_default_arg0) {
-			opts->argv[0] = cmd_line;
-		    }
-		    execve(cmd_line, opts->argv, new_environ);
-		    if (opts->argv[0] == cmd_line) {
-			opts->argv[0] = erts_default_arg0;
-		    }
-		}
-	    } else {
-		execle(SHELL, "sh", "-c", cmd_line, (char *) NULL, new_environ);
-	    }
-	child_error:
-	    _exit(1);
-	}
-#if !DISABLE_VFORK
-    }
-#define ENOUGH_BYTES (44)
-    else { /* Use vfork() */
-	char **cs_argv= erts_alloc(ERTS_ALC_T_TMP,(CS_ARGV_NO_OF_ARGS + 1)*
-				   sizeof(char *));
-	char fd_close_range[ENOUGH_BYTES];                  /* 44 bytes are enough to  */
-	char dup2_op[CS_ARGV_NO_OF_DUP2_OPS][ENOUGH_BYTES]; /* hold any "%d:%d" string */
-                                                            /* on a 64-bit machine.    */
-
-	/* Setup argv[] for the child setup program (implemented in
-	   erl_child_setup.c) */
-	i = 0;
-	if (opts->use_stdio) {
-	    if (opts->read_write & DO_READ){
-		/* stdout for process */
-		erts_snprintf(&dup2_op[i++][0], ENOUGH_BYTES, "%d:%d", ifd[1], 1);
-		if(opts->redir_stderr)
-		    /* stderr for process */
-		    erts_snprintf(&dup2_op[i++][0], ENOUGH_BYTES, "%d:%d", ifd[1], 2);
-	    }
-	    if (opts->read_write & DO_WRITE)
-		/* stdin for process */
-		erts_snprintf(&dup2_op[i++][0], ENOUGH_BYTES, "%d:%d", ofd[0], 0);
-	} else {	/* XXX will fail if ofd[0] == 4 (unlikely..) */
-	    if (opts->read_write & DO_READ)
-		erts_snprintf(&dup2_op[i++][0], ENOUGH_BYTES, "%d:%d", ifd[1], 4);
-	    if (opts->read_write & DO_WRITE)
-		erts_snprintf(&dup2_op[i++][0], ENOUGH_BYTES, "%d:%d", ofd[0], 3);
-	}
-	for (; i < CS_ARGV_NO_OF_DUP2_OPS; i++)
-	    strcpy(&dup2_op[i][0], "-");
-	erts_snprintf(fd_close_range, ENOUGH_BYTES, "%d:%d", opts->use_stdio ? 3 : 5, max_files-1);
-
-	cs_argv[CS_ARGV_PROGNAME_IX] = child_setup_prog;
-	cs_argv[CS_ARGV_WD_IX] = opts->wd ? opts->wd : ".";
-	cs_argv[CS_ARGV_UNBIND_IX] = erts_sched_bind_atvfork_child(unbind);
-	cs_argv[CS_ARGV_FD_CR_IX] = fd_close_range;
-	for (i = 0; i < CS_ARGV_NO_OF_DUP2_OPS; i++)
-	    cs_argv[CS_ARGV_DUP2_OP_IX(i)] = &dup2_op[i][0];
-
-	if (opts->spawn_type == ERTS_SPAWN_EXECUTABLE) {
-	    int num = 0;
-	    int j = 0;
-	    if (opts->argv != NULL) {
-		for(; opts->argv[num] != NULL; ++num)
-		    ;
-	    }
-	    cs_argv = erts_realloc(ERTS_ALC_T_TMP,cs_argv, (CS_ARGV_NO_OF_ARGS + 1 + num + 1) * sizeof(char *));
-	    cs_argv[CS_ARGV_CMD_IX] = "-";
-	    cs_argv[CS_ARGV_NO_OF_ARGS] = cmd_line;
-	    if (opts->argv != NULL) {
-		for (;opts->argv[j] != NULL; ++j) {
-		    if (opts->argv[j] == erts_default_arg0) {
-			cs_argv[CS_ARGV_NO_OF_ARGS + 1 + j] = cmd_line;
-		    } else {
-			cs_argv[CS_ARGV_NO_OF_ARGS + 1 + j] = opts->argv[j];
-		    }
-		}
-	    }
-	    cs_argv[CS_ARGV_NO_OF_ARGS + 1 + j] = NULL;
-	} else {
-	    cs_argv[CS_ARGV_CMD_IX] = cmd_line; /* Command */
-	    cs_argv[CS_ARGV_NO_OF_ARGS] = NULL;
-	}
-	DEBUGF(("Using vfork\n"));
-	pid = vfork();
-
-	if (pid == 0) {
-	    /* The child! */
-
-	    /* Observe!
-	     * OTP-4389: The child setup program (implemented in
-	     * erl_child_setup.c) will perform the necessary setup of the
-	     * child before it execs to the user program. This because
-	     * vfork() only allow an *immediate* execve() or _exit() in the
-	     * child.
-	     */
-	    execve(child_setup_prog, cs_argv, new_environ);
-	    _exit(1);
-	}
-	erts_free(ERTS_ALC_T_TMP,cs_argv);
-    }
-#undef ENOUGH_BYTES
-#endif
-
-    erts_sched_bind_atfork_parent(unbind);
-
-    if (pid == -1) {
-        saved_errno = errno;
-	CHLD_STAT_UNLOCK;
-	erts_smp_rwmtx_runlock(&environ_rwmtx);
-	erts_free(ERTS_ALC_T_TMP, (void *) cmd_line);
-        unblock_signals();
-        close_pipes(ifd, ofd, opts->read_write);
-	errno = saved_errno;
-	return ERL_DRV_ERROR_ERRNO;
-    }
-#else /* QNX */
-    if (opts->use_stdio) {
-	if (opts->read_write & DO_READ)
-	    qnx_spawn_options.iov[1] = ifd[1];  /* stdout for process */
-	if (opts->read_write & DO_WRITE)
-	    qnx_spawn_options.iov[0] = ofd[0];  /* stdin for process */
-	} 
-    else {
-	if (opts->read_write & DO_READ)
-	    qnx_spawn_options.iov[4] = ifd[1];
-	if (opts->read_write & DO_WRITE)
-	    qnx_spawn_options.iov[3] = ofd[0];
-    }
-    /* Close fds on exec */
-    for (i = 3; i < max_files; i++)
-	fcntl(i, F_SETFD, 1);
-
-    qnx_spawn_options.flags = _SPAWN_SETSID;
-    if ((pid = spawnl(P_NOWAIT, SHELL, SHELL, "-c", cmd_line, 
-                      (char *) 0)) < 0) {
-	erts_free(ERTS_ALC_T_TMP, (void *) cmd_line);
-        reset_qnx_spawn();
-	erts_smp_rwmtx_runlock(&environ_rwmtx);
-	close_pipes(ifd, ofd, opts->read_write);
-	return ERL_DRV_ERROR_GENERAL;
-    }
-    reset_qnx_spawn();
-#endif /* QNX */
-
-    erts_free(ERTS_ALC_T_TMP, (void *) cmd_line);
-
-    if (new_environ != environ)
-	erts_free(ERTS_ALC_T_ENVIRONMENT, (void *) new_environ);
-
-    if (opts->read_write & DO_READ) 
-	(void) close(ifd[1]);
-    if (opts->read_write & DO_WRITE)
-	(void) close(ofd[0]);
-	
-    if (opts->read_write & DO_READ) {
-	SET_NONBLOCKING(ifd[0]);
-	init_fd_data(ifd[0], port_num);
-    }
-    if (opts->read_write & DO_WRITE) {
-	SET_NONBLOCKING(ofd[1]);
-        init_fd_data(ofd[1], port_num);
-    }
-
-    res = set_driver_data(port_num, ifd[0], ofd[1], opts->packet_bytes,
-			  opts->read_write, opts->exit_status, pid, 0);
-    /* Don't unblock SIGCHLD until now, since the call above must
-       first complete putting away the info about our new subprocess. */
-    unblock_signals();
-
-#if CHLDWTHR
-    ASSERT(children_alive >= 0);
-
-    if (!(children_alive++))
-	CHLD_STAT_SIGNAL; /* Wake up child waiter thread if no children
-			     was alive before we fork()ed ... */
-#endif
-    /* Don't unlock chld_stat_mtx until now of the same reason as above */
-    CHLD_STAT_UNLOCK;
-
-    erts_smp_rwmtx_runlock(&environ_rwmtx);
-
-    return (ErlDrvData)res;
-#undef CMD_LINE_PREFIX_STR
-#undef CMD_LINE_PREFIX_STR_SZ
-}
-
-#ifdef QNX
-static reset_qnx_spawn()
-{
-    int i;
-
-    /* Reset qnx_spawn_options */
-    qnx_spawn_options.flags = 0; 
-    qnx_spawn_options.iov[0] = 0xff;
-    qnx_spawn_options.iov[1] = 0xff;
-    qnx_spawn_options.iov[2] = 0xff;
-    qnx_spawn_options.iov[3] = 0xff;
-}
-#endif
-
-#define FD_DEF_HEIGHT 24
-#define FD_DEF_WIDTH 80
-/* Control op */
-#define FD_CTRL_OP_GET_WINSIZE 100
-
-static int fd_get_window_size(int fd, Uint32 *width, Uint32 *height)
-{
-#ifdef TIOCGWINSZ 
-    struct winsize ws;
-    if (ioctl(fd,TIOCGWINSZ,&ws) == 0) {
-	*width = (Uint32) ws.ws_col;
-	*height = (Uint32) ws.ws_row;
-	return 0;
-    }
-#endif
-    return -1;
-}
-
-static ErlDrvSSizeT fd_control(ErlDrvData drv_data,
-			       unsigned int command,
-			       char *buf, ErlDrvSizeT len,
-			       char **rbuf, ErlDrvSizeT rlen)
-{
-    int fd = (int)(long)drv_data;
-    char resbuff[2*sizeof(Uint32)];
-    switch (command) {
-    case FD_CTRL_OP_GET_WINSIZE:
-	{
-	    Uint32 w,h;
-	    if (fd_get_window_size(fd,&w,&h)) 
-		return 0;
-	    memcpy(resbuff,&w,sizeof(Uint32));
-	    memcpy(resbuff+sizeof(Uint32),&h,sizeof(Uint32));
-	}
-	break;
-    default:
-	return 0;
-    }
-    if (rlen < 2*sizeof(Uint32)) {
-	*rbuf = driver_alloc(2*sizeof(Uint32));
-    }
-    memcpy(*rbuf,resbuff,2*sizeof(Uint32));
-    return 2*sizeof(Uint32);
-}
-
-static ErlDrvData fd_start(ErlDrvPort port_num, char* name,
-			   SysDriverOpts* opts)
-{
-    ErlDrvData res;
-    int non_blocking = 0;
-
-    if (((opts->read_write & DO_READ) && opts->ifd >= max_files) ||
-	((opts->read_write & DO_WRITE) && opts->ofd >= max_files))
-	return ERL_DRV_ERROR_GENERAL;
-
-    /*
-     * Historical:
-     *
-     * "Note about nonblocking I/O.
-     *
-     * At least on Solaris, setting the write end of a TTY to nonblocking,
-     * will set the input end to nonblocking as well (and vice-versa).
-     * If erl is run in a pipeline like this:  cat | erl
-     * the input end of the TTY will be the standard input of cat.
-     * And cat is not prepared to handle nonblocking I/O."
-     *
-     * Actually, the reason for this is not that the tty itself gets set
-     * in non-blocking mode, but that the "input end" (cat's stdin) and
-     * the "output end" (erlang's stdout) are typically the "same" file
-     * descriptor, dup()'ed from a single fd by one of this process'
-     * ancestors.
-     *
-     * The workaround for this problem used to be a rather bad kludge,
-     * interposing an extra process ("internal cat") between erlang's
-     * stdout and the original stdout, allowing erlang to set its stdout
-     * in non-blocking mode without affecting the stdin of the preceding
-     * process in the pipeline - and being a kludge, it caused all kinds
-     * of weird problems.
-     *
-     * So, this is the current logic:
-     *
-     * The only reason to set non-blocking mode on the output fd at all is
-     * if it's something that can cause a write() to block, of course,
-     * i.e. primarily if it points to a tty, socket, pipe, or fifo. 
-     *
-     * If we don't set non-blocking mode when we "should" have, and output
-     * becomes blocked, the entire runtime system will be suspended - this
-     * is normally bad of course, and can happen fairly "easily" - e.g. user
-     * hits ^S on tty - but doesn't necessarily happen.
-     * 
-     * If we do set non-blocking mode when we "shouldn't" have, the runtime
-     * system will end up seeing EOF on the input fd (due to the preceding
-     * process dying), which typically will cause the entire runtime system
-     * to terminate immediately (due to whatever erlang process is seeing
-     * the EOF taking it as a signal to halt the system). This is *very* bad.
-     * 
-     * I.e. we should take a conservative approach, and only set non-
-     * blocking mode when we a) need to, and b) are reasonably certain
-     * that it won't be a problem. And as in the example above, the problem
-     * occurs when input fd and output fd point to different "things".
-     *
-     * However, determining that they are not just the same "type" of
-     * "thing", but actually the same instance of that type of thing, is
-     * unreasonably complex in many/most cases.
-     *
-     * Also, with pipes, sockets, and fifos it's far from obvious that the
-     * user *wants* non-blocking output: If you're running erlang inside
-     * some complex pipeline, you're probably not running a real-time system
-     * that must never stop, but rather *want* it to suspend if the output
-     * channel is "full".
-     *
-     * So, the bottom line: We will only set the output fd non-blocking if
-     * it points to a tty, and either a) the input fd also points to a tty,
-     * or b) we can make sure that setting the output fd non-blocking
-     * doesn't interfere with someone else's input, via a somewhat milder
-     * kludge than the above.
-     *
-     * Also keep in mind that while this code is almost exclusively run as
-     * a result of an erlang open_port({fd,0,1}, ...), that isn't the only
-     * case - it can be called with any old pre-existing file descriptors,
-     * the relations between which (if they're even two) we can only guess
-     * at - still, we try our best...
-     *
-     * Added note OTP 18: Some systems seem to use stdout/stderr to log data
-     * using unix pipes, so we cannot allow the system to block on a write.
-     * Therefore we use an async thread to write the data to fd's that could
-     * not be set to non-blocking. When no async threads are available we
-     * fall back on the old behaviour.
-     *
-     * Also the guarantee about what is delivered to the OS has changed.
-     * Pre 18 the fd driver did no flushing of data before terminating.
-     * Now it does. This is because we want to be able to guarantee that things
-     * such as escripts and friends really have outputted all data before
-     * terminating. This could potentially block the termination of the system
-     * for a very long time, but if the user wants to terminate fast she should
-     * use erlang:halt with flush=false.
-     */
-
-    if (opts->read_write & DO_READ) {
-	init_fd_data(opts->ifd, port_num);
-    }
-    if (opts->read_write & DO_WRITE) {
-	init_fd_data(opts->ofd, port_num);
-
-	/* If we don't have a read end, all bets are off - no non-blocking. */
-	if (opts->read_write & DO_READ) {
-
-	    if (isatty(opts->ofd)) { /* output fd is a tty:-) */
-
-		if (isatty(opts->ifd)) { /* input fd is also a tty */
-
-		    /* To really do this "right", we should also check that
-		       input and output fd point to the *same* tty - but
-		       this seems like overkill; ttyname() isn't for free,
-		       and this is a very common case - and it's hard to
-		       imagine a scenario where setting non-blocking mode
-		       here would cause problems - go ahead and do it. */
-
-                    non_blocking = 1;
-		    SET_NONBLOCKING(opts->ofd);
-
-		} else {	/* output fd is a tty, input fd isn't */
-
-		    /* This is a "problem case", but also common (see the
-		       example above) - i.e. it makes sense to try a bit
-		       harder before giving up on non-blocking mode: Try to
-		       re-open the tty that the output fd points to, and if
-		       successful replace the original one with the "new" fd
-		       obtained this way, and set *that* one in non-blocking
-		       mode. (Yes, this is a kludge.)
-
-		       However, re-opening the tty may fail in a couple of
-		       (unusual) cases:
-
-		       1) The name of the tty (or an equivalent one, i.e.
-			  same major/minor number) can't be found, because
-			  it actually lives somewhere other than /dev (or
-			  wherever ttyname() looks for it), and isn't
-			  equivalent to any of those that do live in the
-			  "standard" place - this should be *very* unusual.
-
-		       2) Permissions on the tty don't allow us to open it -
-			  it's perfectly possible to have an fd open to an
-			  object whose permissions wouldn't allow us to open
-			  it. This is not as unusual as it sounds, one case
-			  is if the user has su'ed to someone else (not
-			  root) - we have a read/write fd open to the tty
-			  (because it has been inherited all the way down
-			  here), but we have neither read nor write
-			  permission for the tty.
-
-		       In these cases, we finally give up, and don't set the
-		       output fd in non-blocking mode. */
-
-		    char *tty;
-		    int nfd;
-
-		    if ((tty = ttyname(opts->ofd)) != NULL &&
-			(nfd = open(tty, O_WRONLY)) != -1) {
-			dup2(nfd, opts->ofd);
-			close(nfd);
-                        non_blocking = 1;
-			SET_NONBLOCKING(opts->ofd);
-		    }
-		}
-	    }
-	}
-    }
-    CHLD_STAT_LOCK;
-    res = (ErlDrvData)(long)set_driver_data(port_num, opts->ifd, opts->ofd,
-                                            opts->packet_bytes,
-                                            opts->read_write, 0, -1,
-                                            !non_blocking);
-    CHLD_STAT_UNLOCK;
-    return res;
-}
-
-static void clear_fd_data(int fd) 
-{
-    if (fd_data[fd].sz > 0) {
-	erts_free(ERTS_ALC_T_FD_ENTRY_BUF, (void *) fd_data[fd].buf);
-	ASSERT(erts_smp_atomic_read_nob(&sys_misc_mem_sz) >= fd_data[fd].sz);
-	erts_smp_atomic_add_nob(&sys_misc_mem_sz, -1*fd_data[fd].sz);
-    }
-    fd_data[fd].buf = NULL;
-    fd_data[fd].sz = 0;
-    fd_data[fd].remain = 0;
-    fd_data[fd].cpos = NULL;
-    fd_data[fd].psz = 0;
-}
-
-static void nbio_stop_fd(ErlDrvPort prt, int fd)
-{
-    driver_select(prt,fd,DO_READ|DO_WRITE,0);
-    clear_fd_data(fd);
-    SET_BLOCKING(fd);
-}
-
-static void fd_stop(ErlDrvData ev)  /* Does not close the fds */
-{
-    int ofd;
-    int fd = (int)(long)ev;
-    ErlDrvPort prt = driver_data[fd].port_num;
-    
-#if FDBLOCK
-    if (driver_data[fd].blocking) {
-        erts_free(ERTS_ALC_T_SYS_BLOCKING,driver_data[fd].blocking);
-        driver_data[fd].blocking = NULL;
-        erts_smp_atomic_add_nob(&sys_misc_mem_sz, -1*sizeof(ErtsSysBlocking));
-    }
-#endif
-
-    nbio_stop_fd(prt, fd);
-    ofd = driver_data[fd].ofd;
-    if (ofd != fd && ofd != -1)
-	nbio_stop_fd(prt, ofd);
-}
-
-static void fd_flush(ErlDrvData fd)
-{
-    if (!driver_data[(int)(long)fd].terminating)
-        driver_data[(int)(long)fd].terminating = 1;
-}
-
-static ErlDrvData vanilla_start(ErlDrvPort port_num, char* name,
-				SysDriverOpts* opts)
-{
-    int flags, fd;
-    ErlDrvData res;
-
-    flags = (opts->read_write == DO_READ ? O_RDONLY :
-	     opts->read_write == DO_WRITE ? O_WRONLY|O_CREAT|O_TRUNC :
-	     O_RDWR|O_CREAT);
-    if ((fd = open(name, flags, 0666)) < 0)
-	return ERL_DRV_ERROR_GENERAL;
-    if (fd >= max_files) {
-	close(fd);
-	return ERL_DRV_ERROR_GENERAL;
-    }
-    SET_NONBLOCKING(fd);
-    init_fd_data(fd, port_num);
-
-    CHLD_STAT_LOCK;
-    res = (ErlDrvData)(long)set_driver_data(port_num, fd, fd,
-                                            opts->packet_bytes,
-                                            opts->read_write, 0, -1, 0);
-    CHLD_STAT_UNLOCK;
-    return res;
-}
-
-/* Note that driver_data[fd].ifd == fd if the port was opened for reading, */
-/* otherwise (i.e. write only) driver_data[fd].ofd = fd.  */
-
-static void stop(ErlDrvData fd)
-{
-    ErlDrvPort prt;
-    int ofd;
-
-    prt = driver_data[(int)(long)fd].port_num;
-    nbio_stop_fd(prt, (int)(long)fd);
-
-    ofd = driver_data[(int)(long)fd].ofd;
-    if (ofd != (int)(long)fd && (int)(long)ofd != -1)
-	nbio_stop_fd(prt, ofd);
-    else
-	ofd = -1;
-
-    CHLD_STAT_LOCK;
-
-    /* Mark as unused. */
-    driver_data[(int)(long)fd].pid = -1;
-
-    CHLD_STAT_UNLOCK;
-
-    /* SMP note: Close has to be last thing done (open file descriptors work
-       as locks on driver_data[] entries) */
-    driver_select(prt, (int)(long)fd, ERL_DRV_USE, 0);  /* close(fd); */
-    if (ofd >= 0) {
-	driver_select(prt, (int)(long)ofd, ERL_DRV_USE, 0);  /* close(ofd); */
-    }
-}
-
-/* used by fd_driver */
-static void outputv(ErlDrvData e, ErlIOVec* ev)
-{
-    int fd = (int)(long)e;
-    ErlDrvPort ix = driver_data[fd].port_num;
-    int pb = driver_data[fd].packet_bytes;
-    int ofd = driver_data[fd].ofd;
-    ssize_t n;
-    ErlDrvSizeT sz;
-    char lb[4];
-    char* lbp;
-    ErlDrvSizeT len = ev->size;
-
-    /* (len > ((unsigned long)-1 >> (4-pb)*8)) */
-    /*    if (pb >= 0 && (len & (((ErlDrvSizeT)1 << (pb*8))) - 1) != len) {*/
-    if (((pb == 2) && (len > 0xffff)) || (pb == 1 && len > 0xff)) {
-	driver_failure_posix(ix, EINVAL);
-	return; /* -1; */
-    }
-    /* Handles 0 <= pb <= 4 only */
-    put_int32((Uint32) len, lb);
-    lbp = lb + (4-pb);
-
-    ev->iov[0].iov_base = lbp;
-    ev->iov[0].iov_len = pb;
-    ev->size += pb;
-
-    if (driver_data[fd].blocking && FDBLOCK)
-        driver_pdl_lock(driver_data[fd].blocking->pdl);
-
-    if ((sz = driver_sizeq(ix)) > 0) {
-	driver_enqv(ix, ev, 0);
-
-        if (driver_data[fd].blocking && FDBLOCK)
-            driver_pdl_unlock(driver_data[fd].blocking->pdl);
-
-	if (sz + ev->size >= (1 << 13))
-	    set_busy_port(ix, 1);
-    }
-    else if (!driver_data[fd].blocking || !FDBLOCK) {
-        /* We try to write directly if the fd in non-blocking */
-	int vsize = ev->vsize > MAX_VSIZE ? MAX_VSIZE : ev->vsize;
-
-	n = writev(ofd, (const void *) (ev->iov), vsize);
-	if (n == ev->size)
-	    return; /* 0;*/
-	if (n < 0) {
-	    if ((errno != EINTR) && (errno != ERRNO_BLOCK)) {
-		driver_failure_posix(ix, errno);
-		return; /* -1;*/
-	    }
-	    n = 0;
-	}
-	driver_enqv(ix, ev, n);  /* n is the skip value */
-	driver_select(ix, ofd, ERL_DRV_WRITE|ERL_DRV_USE, 1);
-    }
-#if FDBLOCK
-    else {
-        if (ev->size != 0) {
-            driver_enqv(ix, ev, 0);
-            driver_pdl_unlock(driver_data[fd].blocking->pdl);
-            driver_async(ix, &driver_data[fd].blocking->pkey,
-                         fd_async, driver_data+fd, NULL);
-        } else {
-            driver_pdl_unlock(driver_data[fd].blocking->pdl);
-        }
-    }
-#endif
-    /* return 0;*/
-}
-
-/* Used by spawn_driver and vanilla driver */
-static void output(ErlDrvData e, char* buf, ErlDrvSizeT len)
-{
-    int fd = (int)(long)e;
-    ErlDrvPort ix = driver_data[fd].port_num;
-    int pb = driver_data[fd].packet_bytes;
-    int ofd = driver_data[fd].ofd;
-    ssize_t n;
-    ErlDrvSizeT sz;
-    char lb[4];
-    char* lbp;
-    struct iovec iv[2];
-
-    /* (len > ((unsigned long)-1 >> (4-pb)*8)) */
-    if (((pb == 2) && (len > 0xffff)) || (pb == 1 && len > 0xff)) {
-	driver_failure_posix(ix, EINVAL);
-	return; /* -1; */
-    }
-    put_int32(len, lb);
-    lbp = lb + (4-pb);
-
-    if ((sz = driver_sizeq(ix)) > 0) {
-	driver_enq(ix, lbp, pb);
-	driver_enq(ix, buf, len);
-	if (sz + len + pb >= (1 << 13))
-	    set_busy_port(ix, 1);
-    }
-    else {
-	iv[0].iov_base = lbp;
-	iv[0].iov_len = pb;  /* should work for pb=0 */
-	iv[1].iov_base = buf;
-	iv[1].iov_len = len;
-	n = writev(ofd, iv, 2);
-	if (n == pb+len)
-	    return; /* 0; */
-	if (n < 0) {
-	    if ((errno != EINTR) && (errno != ERRNO_BLOCK)) {
-		driver_failure_posix(ix, errno);
-		return; /* -1; */
-	    }
-	    n = 0;
-	}
-	if (n < pb) {
-	    driver_enq(ix, lbp+n, pb-n);
-	    driver_enq(ix, buf, len);
-	}
-	else {
-	    n -= pb;
-	    driver_enq(ix, buf+n, len-n);
-	}
-	driver_select(ix, ofd, ERL_DRV_WRITE|ERL_DRV_USE, 1);
-    }
-    return; /* 0; */
-}
-
-static int port_inp_failure(ErlDrvPort port_num, int ready_fd, int res)
-				/* Result: 0 (eof) or -1 (error) */
-{
-    int err = errno;
-
-    ASSERT(res <= 0);
-    (void) driver_select(port_num, ready_fd, ERL_DRV_READ|ERL_DRV_WRITE, 0); 
-    clear_fd_data(ready_fd);
-
-    if (driver_data[ready_fd].blocking && FDBLOCK) {
-        driver_pdl_lock(driver_data[ready_fd].blocking->pdl);
-        if (driver_sizeq(driver_data[ready_fd].port_num) > 0) {
-            driver_pdl_unlock(driver_data[ready_fd].blocking->pdl);
-            /* We have stuff in the output queue, so we just
-               set the state to terminating and wait for fd_async_ready
-               to terminate the port */
-            if (res == 0)
-                driver_data[ready_fd].terminating = 2;
-            else
-                driver_data[ready_fd].terminating = -err;
-            return 0;
-        }
-        driver_pdl_unlock(driver_data[ready_fd].blocking->pdl);
-    }
-
-    if (res == 0) {
-	if (driver_data[ready_fd].report_exit) {
-	    CHLD_STAT_LOCK;
-
-	    if (driver_data[ready_fd].alive) {
-		/*
-		 * We have eof and want to report exit status, but the process
-		 * hasn't exited yet. When it does report_exit_status() will
-		 * driver_select() this fd which will make sure that we get
-		 * back here with driver_data[ready_fd].alive == 0 and
-		 * driver_data[ready_fd].status set.
-		 */
-		CHLD_STAT_UNLOCK;
-		return 0;
-	    }
-	    else {
-		int status = driver_data[ready_fd].status;
-		CHLD_STAT_UNLOCK;
-
-		/* We need not be prepared for stopped/continued processes. */
-		if (WIFSIGNALED(status))
-		    status = 128 + WTERMSIG(status);
-		else
-		    status = WEXITSTATUS(status);
-
-		driver_report_exit(driver_data[ready_fd].port_num, status);
-	    }
-       }
-       driver_failure_eof(port_num);
-    } else {
-	driver_failure_posix(port_num, err);
-    }
-    return 0;
-}
-
-/* fd is the drv_data that is returned from the */
-/* initial start routine                        */
-/* ready_fd is the descriptor that is ready to read */
-
-static void ready_input(ErlDrvData e, ErlDrvEvent ready_fd)
-{
-    int fd = (int)(long)e;
-    ErlDrvPort port_num;
-    int packet_bytes;
-    int res;
-    Uint h;
-
-    port_num = driver_data[fd].port_num;
-    packet_bytes = driver_data[fd].packet_bytes;
-
-
-    if (packet_bytes == 0) {
-	byte *read_buf = (byte *) erts_alloc(ERTS_ALC_T_SYS_READ_BUF,
-					     ERTS_SYS_READ_BUF_SZ);
-	res = read(ready_fd, read_buf, ERTS_SYS_READ_BUF_SZ);
-	if (res < 0) {
-	    if ((errno != EINTR) && (errno != ERRNO_BLOCK))
-		port_inp_failure(port_num, ready_fd, res);
-	}
-	else if (res == 0)
-	    port_inp_failure(port_num, ready_fd, res);
-	else 
-	    driver_output(port_num, (char*) read_buf, res);
-	erts_free(ERTS_ALC_T_SYS_READ_BUF, (void *) read_buf);
-    }
-    else if (fd_data[ready_fd].remain > 0) { /* We try to read the remainder */
-	/* space is allocated in buf */
-	res = read(ready_fd, fd_data[ready_fd].cpos, 
-		   fd_data[ready_fd].remain);
-	if (res < 0) {
-	    if ((errno != EINTR) && (errno != ERRNO_BLOCK))
-		port_inp_failure(port_num, ready_fd, res);
-	}
-	else if (res == 0) {
-	    port_inp_failure(port_num, ready_fd, res);
-	}
-	else if (res == fd_data[ready_fd].remain) { /* we're done  */
-	    driver_output(port_num, fd_data[ready_fd].buf, 
-			  fd_data[ready_fd].sz);
-	    clear_fd_data(ready_fd);
-	}
-	else { /*  if (res < fd_data[ready_fd].remain) */
-	    fd_data[ready_fd].cpos += res;
-	    fd_data[ready_fd].remain -= res;
-	}
-    }
-    else if (fd_data[ready_fd].remain == 0) { /* clean fd */
-	byte *read_buf = (byte *) erts_alloc(ERTS_ALC_T_SYS_READ_BUF,
-					     ERTS_SYS_READ_BUF_SZ);
-	/* We make one read attempt and see what happens */
-	res = read(ready_fd, read_buf, ERTS_SYS_READ_BUF_SZ);
-	if (res < 0) {  
-	    if ((errno != EINTR) && (errno != ERRNO_BLOCK))
-		port_inp_failure(port_num, ready_fd, res);
-	}
-	else if (res == 0) {     	/* eof */
-	    port_inp_failure(port_num, ready_fd, res);
-	} 
-	else if (res < packet_bytes - fd_data[ready_fd].psz) { 
-	    memcpy(fd_data[ready_fd].pbuf+fd_data[ready_fd].psz,
-		   read_buf, res);
-	    fd_data[ready_fd].psz += res;
-	}
-	else  { /* if (res >= packet_bytes) */
-	    unsigned char* cpos = read_buf;
-	    int bytes_left = res;
-
-	    while (1) {
-		int psz = fd_data[ready_fd].psz;
-		char* pbp = fd_data[ready_fd].pbuf + psz;
-
-		while(bytes_left && (psz < packet_bytes)) {
-		    *pbp++ = *cpos++;
-		    bytes_left--;
-		    psz++;
-		}
-		
-		if (psz < packet_bytes) {
-		    fd_data[ready_fd].psz = psz;
-		    break;
-		}
-		fd_data[ready_fd].psz = 0;
-
-		switch (packet_bytes) {
-		case 1: h = get_int8(fd_data[ready_fd].pbuf);  break;
-		case 2: h = get_int16(fd_data[ready_fd].pbuf); break;
-		case 4: h = get_int32(fd_data[ready_fd].pbuf); break;
-		default: ASSERT(0); return; /* -1; */
-		}
-
-		if (h <= (bytes_left)) {
-		    driver_output(port_num, (char*) cpos, h);
-		    cpos += h;
-		    bytes_left -= h;
-		    continue;
-		}
-		else {		/* The last message we got was split */
-		        char *buf = erts_alloc_fnf(ERTS_ALC_T_FD_ENTRY_BUF, h);
-		    if (!buf) {
-			errno = ENOMEM;
-			port_inp_failure(port_num, ready_fd, -1);
-		    }
-		    else {
-			erts_smp_atomic_add_nob(&sys_misc_mem_sz, h);
-			sys_memcpy(buf, cpos, bytes_left);
-			fd_data[ready_fd].buf = buf;
-			fd_data[ready_fd].sz = h;
-			fd_data[ready_fd].remain = h - bytes_left;
-			fd_data[ready_fd].cpos = buf + bytes_left;
-		    }
-		    break;
-		}
-	    }
-	}
-	erts_free(ERTS_ALC_T_SYS_READ_BUF, (void *) read_buf);
-    }
-}
-
-
-/* fd is the drv_data that is returned from the */
-/* initial start routine                        */
-/* ready_fd is the descriptor that is ready to read */
-
-static void ready_output(ErlDrvData e, ErlDrvEvent ready_fd)
-{
-    int fd = (int)(long)e;
-    ErlDrvPort ix = driver_data[fd].port_num;
-    int n;
-    struct iovec* iv;
-    int vsize;
-    
-
-    if ((iv = (struct iovec*) driver_peekq(ix, &vsize)) == NULL) {
-	driver_select(ix, ready_fd, ERL_DRV_WRITE, 0);
-        if (driver_data[fd].terminating)
-            driver_failure_atom(driver_data[fd].port_num,"normal");
-	return; /* 0; */
-    }
-    vsize = vsize > MAX_VSIZE ? MAX_VSIZE : vsize;
-    if ((n = writev(ready_fd, iv, vsize)) > 0) {
-	if (driver_deq(ix, n) == 0)
-	    set_busy_port(ix, 0);
-    }
-    else if (n < 0) {
-	if (errno == ERRNO_BLOCK || errno == EINTR)
-	    return; /* 0; */
-	else {
-	    int res = errno;
-	    driver_select(ix, ready_fd, ERL_DRV_WRITE, 0);
-	    driver_failure_posix(ix, res);
-	    return; /* -1; */
-	}
-    }
-    return; /* 0; */
-}
-
-static void stop_select(ErlDrvEvent fd, void* _)
-{
-    close((int)fd);
-}
-
-#if FDBLOCK
-
-static void
-fd_async(void *async_data)
-{
-    int res;
-    struct driver_data *dd = (struct driver_data*)async_data;
-    SysIOVec      *iov0;
-    SysIOVec      *iov;
-    int            iovlen;
-    int            err;
-    /* much of this code is stolen from efile_drv:invoke_writev */
-    driver_pdl_lock(dd->blocking->pdl);
-    iov0 = driver_peekq(dd->port_num, &iovlen);
-    iovlen = iovlen < MAXIOV ? iovlen : MAXIOV;
-    iov = erts_alloc_fnf(ERTS_ALC_T_SYS_WRITE_BUF,
-                         sizeof(SysIOVec)*iovlen);
-    if (!iov) {
-        res = -1;
-        err = ENOMEM;
-        driver_pdl_unlock(dd->blocking->pdl);
-    } else {
-        memcpy(iov,iov0,iovlen*sizeof(SysIOVec));
-        driver_pdl_unlock(dd->blocking->pdl);
-
-        res = writev(dd->ofd, iov, iovlen);
-        err = errno;
-
-        erts_free(ERTS_ALC_T_SYS_WRITE_BUF, iov);
-    }
-    dd->blocking->res = res;
-    dd->blocking->err = err;
-}
-
-void fd_ready_async(ErlDrvData drv_data,
-                    ErlDrvThreadData thread_data) {
-    struct driver_data *dd = (struct driver_data *)thread_data;
-    ErlDrvPort port_num = dd->port_num;
-
-    ASSERT(dd->blocking);
-    ASSERT(dd == (driver_data + (int)(long)drv_data));
-
-    if (dd->blocking->res > 0) {
-        driver_pdl_lock(dd->blocking->pdl);
-        if (driver_deq(port_num, dd->blocking->res) == 0) {
-            driver_pdl_unlock(dd->blocking->pdl);
-            set_busy_port(port_num, 0);
-            if (dd->terminating) {
-                /* The port is has been ordered to terminate
-                   from either fd_flush or port_inp_failure */
-                if (dd->terminating == 1)
-                    driver_failure_atom(port_num, "normal");
-                else if (dd->terminating == 2)
-                    driver_failure_eof(port_num);
-                else if (dd->terminating < 0)
-                    driver_failure_posix(port_num, -dd->terminating);
-                return; /* -1; */
-            }
-        } else {
-            driver_pdl_unlock(dd->blocking->pdl);
-            /* still data left to write in queue */
-            driver_async(port_num, &dd->blocking->pkey, fd_async, dd, NULL);
-            return /* 0; */;
-        }
-    } else if (dd->blocking->res < 0) {
-        driver_failure_posix(port_num, dd->blocking->err);
-        return; /* -1; */
-    }
-    return; /* 0; */
-}
-
-#endif
-
 void erts_do_break_handling(void)
 {
     struct termios temp_mode;
@@ -2730,10 +1011,6 @@ erts_sys_unsetenv(char *key)
 void
 sys_init_io(void)
 {
-    fd_data = (struct fd_data *)
-	erts_alloc(ERTS_ALC_T_FD_TAB, max_files * sizeof(struct fd_data));
-    erts_smp_atomic_add_nob(&sys_misc_mem_sz,
-			    max_files * sizeof(struct fd_data));
 }
 
 #if (0) /* unused? */
@@ -2927,179 +1204,6 @@ erl_debug(char* fmt, ...)
 
 #endif /* DEBUG */
 
-static ERTS_INLINE void
-report_exit_status(ErtsSysReportExit *rep, int status)
-{
-    Port *pp;
-#ifdef ERTS_SMP
-    CHLD_STAT_UNLOCK;
-    pp = erts_thr_id2port_sflgs(rep->port,
-				ERTS_PORT_SFLGS_INVALID_DRIVER_LOOKUP);
-    CHLD_STAT_LOCK;
-#else
-    pp = erts_id2port_sflgs(rep->port,
-			    NULL,
-			    0,
-			    ERTS_PORT_SFLGS_INVALID_DRIVER_LOOKUP);
-#endif
-    if (pp) {
-	if (rep->ifd >= 0) {
-	    driver_data[rep->ifd].alive = 0;
-	    driver_data[rep->ifd].status = status;
-	    (void) driver_select(ERTS_Port2ErlDrvPort(pp),
-				 rep->ifd,
-				 (ERL_DRV_READ|ERL_DRV_USE),
-				 1);
-	}
-	if (rep->ofd >= 0) {
-	    driver_data[rep->ofd].alive = 0;
-	    driver_data[rep->ofd].status = status;
-	    (void) driver_select(ERTS_Port2ErlDrvPort(pp),
-				 rep->ofd,
-				 (ERL_DRV_WRITE|ERL_DRV_USE),
-				 1);
-	}
-#ifdef ERTS_SMP
-	erts_thr_port_release(pp);
-#else
-	erts_port_release(pp);
-#endif
-    }
-    erts_free(ERTS_ALC_T_PRT_REP_EXIT, rep);
-}
-
-#if !CHLDWTHR  /* ---------------------------------------------------------- */
-
-#define ERTS_REPORT_EXIT_STATUS report_exit_status
-
-static int check_children(void)
-{
-    int res = 0;
-    int pid;
-    int status;
-
-#ifndef ERTS_SMP
-    if (children_died)
-#endif
-    {
-	sys_sigblock(SIGCHLD);
-	CHLD_STAT_LOCK;
-	while ((pid = waitpid(-1, &status, WNOHANG)) > 0)
-	    note_child_death(pid, status);
-#ifndef ERTS_SMP
-	children_died = 0;
-#endif
-	CHLD_STAT_UNLOCK;
-	sys_sigrelease(SIGCHLD);
-	res = 1;
-    }
-    return res;
-}
-
-#ifdef ERTS_SMP
-
-void
-erts_check_children(void)
-{
-    (void) check_children();
-}
-
-#endif
-
-#elif CHLDWTHR && defined(ERTS_SMP) /* ------------------------------------- */
-
-#define ERTS_REPORT_EXIT_STATUS report_exit_status
-
-#define check_children() (0)
-
-
-#else /* CHLDWTHR && !defined(ERTS_SMP) ------------------------------------ */
-
-#define ERTS_REPORT_EXIT_STATUS initiate_report_exit_status
-
-static ERTS_INLINE void
-initiate_report_exit_status(ErtsSysReportExit *rep, int status)
-{
-    rep->next = report_exit_transit_list;
-    rep->status = status;
-    report_exit_transit_list = rep;
-    erts_sys_schedule_interrupt(1);
-}
-
-static int check_children(void)
-{
-    int res;
-    ErtsSysReportExit *rep;
-    CHLD_STAT_LOCK;
-    rep = report_exit_transit_list;
-    res = rep != NULL;
-    while (rep) {
-	ErtsSysReportExit *curr_rep = rep;
-	rep = rep->next;
-	report_exit_status(curr_rep, curr_rep->status);
-    }
-    report_exit_transit_list = NULL;
-    CHLD_STAT_UNLOCK;
-    return res;
-}
-
-#endif /* ------------------------------------------------------------------ */
-
-static void note_child_death(int pid, int status)
-{
-    ErtsSysReportExit **repp = &report_exit_list;
-    ErtsSysReportExit *rep = report_exit_list;
-
-    while (rep) {
-	if (pid == rep->pid) {
-	    *repp = rep->next;
-	    ERTS_REPORT_EXIT_STATUS(rep, status);
-	    break;
-	}
-	repp = &rep->next;
-	rep = rep->next;
-    }
-}
-
-#if CHLDWTHR
-
-static void *
-child_waiter(void *unused)
-{
-  int pid;
-  int status;
-
-#ifdef ERTS_ENABLE_LOCK_CHECK
-  erts_lc_set_thread_name("child waiter");
-#endif
-
-  while(1) {
-#ifdef DEBUG
-      int waitpid_errno;
-#endif
-      pid = waitpid(-1, &status, 0);
-#ifdef DEBUG
-      waitpid_errno = errno;
-#endif
-      CHLD_STAT_LOCK;
-      if (pid < 0) {
-	  ASSERT(waitpid_errno == ECHILD);
-      }
-      else {
-	  children_alive--;
-	  ASSERT(children_alive >= 0);
-	  note_child_death(pid, status);
-      }
-      while (!children_alive)
-	  CHLD_STAT_WAIT; /* Wait for children to wait on... :) */
-      CHLD_STAT_UNLOCK;
-  }
-
-  return NULL;
-}
-
-#endif
-
 /*
  * Called from schedule() when it runs out of runnable processes,
  * or when Erlang code has performed INPUT_REDUCTIONS reduction
@@ -3108,13 +1212,8 @@ child_waiter(void *unused)
 void
 erl_sys_schedule(int runnable)
 {
-#ifdef ERTS_SMP
     ERTS_CHK_IO(!runnable);
-#else
-    ERTS_CHK_IO(runnable ? 0 : !check_children());
-#endif
     ERTS_SMP_LC_ASSERT(!erts_thr_progress_is_blocking());
-    (void) check_children();
 }
 
 
@@ -3142,10 +1241,6 @@ smp_sig_notify(char c)
 static void *
 signal_dispatcher_thread_func(void *unused)
 {
-#if !CHLDWTHR
-    int initialized = 0;
-    int notify_check_children = 0;
-#endif
 #ifdef ERTS_ENABLE_LOCK_CHECK
     erts_lc_set_thread_name("signal_dispatcher");
 #endif
@@ -3183,19 +1278,7 @@ signal_dispatcher_thread_func(void *unused)
 	     */
 	    switch (buf[i]) {
 	    case 0: /* Emulator initialized */
-#if !CHLDWTHR
-		initialized = 1;
-		if (!notify_check_children)
-#endif
-		    break;
-#if !CHLDWTHR
-	    case 'C': /* SIGCHLD */
-		if (initialized)
-		    erts_smp_notify_check_children_needed();
-		else
-		    notify_check_children = 1;
-		break;
-#endif
+                break;
 	    case 'I': /* SIGINT */
 		break_requested();
 		break;
diff --git a/erts/emulator/sys/unix/sys_drivers.c b/erts/emulator/sys/unix/sys_drivers.c
new file mode 100644
index 0000000000..2a7cd91265
--- /dev/null
+++ b/erts/emulator/sys/unix/sys_drivers.c
@@ -0,0 +1,1862 @@
+/*
+ * %CopyrightBegin%
+ *
+ * Copyright Ericsson AB 1996-2014. All Rights Reserved.
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ *
+ * %CopyrightEnd%
+ */
+
+#ifdef HAVE_CONFIG_H
+#  include "config.h"
+#endif
+
+#ifdef ISC32
+#define _POSIX_SOURCE
+#define _XOPEN_SOURCE
+#endif
+
+#include <sys/times.h>		/* ! */
+#include <time.h>
+#include <signal.h>
+#include <sys/wait.h>
+#include <sys/uio.h>
+#include <termios.h>
+#include <ctype.h>
+#include <sys/utsname.h>
+#include <sys/select.h>
+#include <arpa/inet.h>
+
+#ifdef ISC32
+#include <sys/bsdtypes.h>
+#endif
+
+#include <termios.h>
+#ifdef HAVE_FCNTL_H
+#include <fcntl.h>
+#endif
+#ifdef HAVE_SYS_IOCTL_H
+#include <sys/ioctl.h>
+#endif
+
+#define WANT_NONBLOCKING    /* must define this to pull in defs from sys.h */
+#include "sys.h"
+
+#ifdef USE_THREADS
+#include "erl_threads.h"
+#endif
+
+extern char **environ;
+extern erts_smp_rwmtx_t environ_rwmtx;
+
+extern erts_smp_atomic_t sys_misc_mem_sz;
+
+static Eterm forker_port;
+
+#define MAX_VSIZE 16		/* Max number of entries allowed in an I/O
+				 * vector sock_sendv().
+				 */
+/*
+ * Don't need global.h, but erl_cpu_topology.h won't compile otherwise
+ */
+#include "global.h"
+#include "erl_cpu_topology.h"
+
+#include "erl_sys_driver.h"
+#include "sys_uds.h"
+
+#include "erl_child_setup.h"
+
+#if defined IOV_MAX
+#define MAXIOV IOV_MAX
+#elif defined UIO_MAXIOV
+#define MAXIOV UIO_MAXIOV
+#else
+#define MAXIOV 16
+#endif
+
+#ifdef USE_THREADS
+#  define FDBLOCK 1
+#else
+#  define FDBLOCK 0
+#endif
+
+/* Used by the fd driver iff the fd could not be set to non-blocking */
+typedef struct ErtsSysBlocking_ {
+    ErlDrvPDL pdl;
+    int res;
+    int err;
+    unsigned int pkey;
+} ErtsSysBlocking;
+
+typedef struct fd_data {
+    int   fd;
+    char  pbuf[4];   /* hold partial packet bytes */
+    int   psz;       /* size of pbuf */
+    char  *buf;
+    char  *cpos;
+    int   sz;
+    int   remain;  /* for input on fd */
+} ErtsSysFdData;
+
+typedef struct driver_data {
+    ErlDrvPort port_num;
+    ErtsSysFdData *ofd;
+    ErtsSysFdData *ifd;
+    int packet_bytes;
+    int pid;
+    int alive;
+    int status;
+    int terminating;
+    ErtsSysBlocking *blocking;
+} ErtsSysDriverData;
+
+#define DIR_SEPARATOR_CHAR    '/'
+
+#if defined(__ANDROID__)
+#define SHELL "/system/bin/sh"
+#else
+#define SHELL "/bin/sh"
+#endif /* __ANDROID__ */
+
+#if defined(DEBUG)
+#define ERL_BUILD_TYPE_MARKER ".debug"
+#elif defined(PURIFY)
+#define ERL_BUILD_TYPE_MARKER ".purify"
+#elif defined(QUANTIFY)
+#define ERL_BUILD_TYPE_MARKER ".quantify"
+#elif defined(PURECOV)
+#define ERL_BUILD_TYPE_MARKER ".purecov"
+#elif defined(VALGRIND)
+#define ERL_BUILD_TYPE_MARKER ".valgrind"
+#else /* opt */
+#define ERL_BUILD_TYPE_MARKER
+#endif
+
+#ifdef DEBUG
+#define close(fd) do { int res = close(fd); ASSERT(res > -1); } while(0)
+#endif
+
+#define CHILD_SETUP_PROG_NAME	"erl_child_setup" ERL_BUILD_TYPE_MARKER
+
+// #define HARD_DEBUG
+#ifdef HARD_DEBUG
+#define driver_select(port_num, fd, flags, onoff)                       \
+    do {                                                                \
+        if (((flags) & ERL_DRV_READ) && onoff)                          \
+            fprintf(stderr,"%010d %p: read select %d\r\n", __LINE__, port_num, (int)fd); \
+        if (((flags) & ERL_DRV_WRITE) && onoff)                         \
+            fprintf(stderr,"%010d %p: writ select %d\r\n", __LINE__, port_num, (int)fd); \
+        if (((flags) & ERL_DRV_READ) && !onoff)                          \
+            fprintf(stderr,"%010d %p: read unsele %d\r\n", __LINE__, port_num, (int)fd); \
+        if (((flags) & ERL_DRV_WRITE) && !onoff)                         \
+            fprintf(stderr,"%010d %p: writ unsele %d\r\n", __LINE__, port_num, (int)fd); \
+        driver_select_nkp(port_num, fd, flags, onoff);                  \
+    } while(0)
+#endif
+
+/*
+ * Decreasing the size of it below 16384 is not allowed.
+ */
+
+#define ERTS_SYS_READ_BUF_SZ (64*1024)
+
+/* I. Initialization */
+
+void
+erl_sys_late_init(void)
+{
+    SysDriverOpts opts;
+#ifdef ERTS_SMP
+    Port *port;
+#endif
+
+    sys_signal(SIGPIPE, SIG_IGN); /* Ignore - we'll handle the write failure */
+
+    opts.packet_bytes = 0;
+    opts.use_stdio = 1;
+    opts.redir_stderr = 0;
+    opts.read_write = 0;
+    opts.hide_window = 0;
+    opts.wd = NULL;
+    opts.envir = NULL;
+    opts.exit_status = 0;
+    opts.overlapped_io = 0;
+    opts.spawn_type = ERTS_SPAWN_ANY;
+    opts.argv = NULL;
+    opts.parallelism = erts_port_parallelism;
+
+#ifdef ERTS_SMP
+    port =
+#endif
+        erts_open_driver(&forker_driver, make_internal_pid(0), "forker", &opts, NULL, NULL);
+#ifdef ERTS_SMP
+    erts_mtx_unlock(port->lock);
+#endif
+}
+
+/* II. Prototypes */
+
+/* II.I Spawn prototypes */
+static ErlDrvData spawn_start(ErlDrvPort, char*, SysDriverOpts*);
+static ErlDrvSSizeT spawn_control(ErlDrvData, unsigned int, char *,
+                                  ErlDrvSizeT, char **, ErlDrvSizeT);
+
+/* II.II Vanilla prototypes */
+static ErlDrvData vanilla_start(ErlDrvPort, char*, SysDriverOpts*);
+
+
+/* II.III FD prototypes */
+static ErlDrvData fd_start(ErlDrvPort, char*, SysDriverOpts*);
+#if FDBLOCK
+static void fd_async(void *);
+static void fd_ready_async(ErlDrvData drv_data, ErlDrvThreadData thread_data);
+#endif
+static ErlDrvSSizeT fd_control(ErlDrvData, unsigned int, char *, ErlDrvSizeT,
+			       char **, ErlDrvSizeT);
+static void fd_stop(ErlDrvData);
+static void fd_flush(ErlDrvData);
+
+/* II.IV Common prototypes */
+static void stop(ErlDrvData);
+static void ready_input(ErlDrvData, ErlDrvEvent);
+static void ready_output(ErlDrvData, ErlDrvEvent);
+static void output(ErlDrvData, char*, ErlDrvSizeT);
+static void outputv(ErlDrvData, ErlIOVec*);
+static void stop_select(ErlDrvEvent, void*);
+
+/* II.V Forker prototypes */
+static ErlDrvData forker_start(ErlDrvPort, char*, SysDriverOpts*);
+static void forker_stop(ErlDrvData);
+static void forker_ready_input(ErlDrvData, ErlDrvEvent);
+static void forker_ready_output(ErlDrvData, ErlDrvEvent);
+static ErlDrvSSizeT forker_control(ErlDrvData, unsigned int, char *,
+                                   ErlDrvSizeT, char **, ErlDrvSizeT);
+
+/* III Driver entries */
+
+/* III.I The spawn driver */
+struct erl_drv_entry spawn_driver_entry = {
+    NULL,
+    spawn_start,
+    stop,
+    output,
+    ready_input,
+    ready_output,
+    "spawn",
+    NULL,
+    NULL,
+    spawn_control,
+    NULL,
+    NULL,
+    NULL,
+    NULL,
+    NULL,
+    NULL,
+    ERL_DRV_EXTENDED_MARKER,
+    ERL_DRV_EXTENDED_MAJOR_VERSION,
+    ERL_DRV_EXTENDED_MINOR_VERSION,
+    ERL_DRV_FLAG_USE_PORT_LOCKING | ERL_DRV_FLAG_USE_INIT_ACK,
+    NULL, NULL,
+    stop_select
+};
+
+/* III.II The fd driver */
+struct erl_drv_entry fd_driver_entry = {
+    NULL,
+    fd_start,
+    fd_stop,
+    output,
+    ready_input,
+    ready_output,
+    "fd",
+    NULL,
+    NULL,
+    fd_control,
+    NULL,
+    outputv,
+#if FDBLOCK
+    fd_ready_async, /* ready_async */
+#else
+    NULL,
+#endif
+    fd_flush, /* flush */
+    NULL, /* call */
+    NULL, /* event */
+    ERL_DRV_EXTENDED_MARKER,
+    ERL_DRV_EXTENDED_MAJOR_VERSION,
+    ERL_DRV_EXTENDED_MINOR_VERSION,
+    0, /* ERL_DRV_FLAGs */
+    NULL, /* handle2 */
+    NULL, /* process_exit */
+    stop_select
+};
+
+/* III.III The vanilla driver */
+struct erl_drv_entry vanilla_driver_entry = {
+    NULL,
+    vanilla_start,
+    stop,
+    output,
+    ready_input,
+    ready_output,
+    "vanilla",
+    NULL,
+    NULL,
+    NULL,
+    NULL,
+    NULL,
+    NULL,
+    NULL, /* flush */
+    NULL, /* call */
+    NULL, /* event */
+    ERL_DRV_EXTENDED_MARKER,
+    ERL_DRV_EXTENDED_MAJOR_VERSION,
+    ERL_DRV_EXTENDED_MINOR_VERSION,
+    0, /* ERL_DRV_FLAGs */
+    NULL, /* handle2 */
+    NULL, /* process_exit */
+    stop_select
+};
+
+/* III.III The forker driver */
+struct erl_drv_entry forker_driver_entry = {
+    NULL,
+    forker_start,
+    forker_stop,
+    NULL,
+    forker_ready_input,
+    forker_ready_output,
+    "spawn_forker",
+    NULL,
+    NULL,
+    forker_control,
+    NULL,
+    NULL,
+    NULL,
+    NULL,
+    NULL,
+    NULL,
+    ERL_DRV_EXTENDED_MARKER,
+    ERL_DRV_EXTENDED_MAJOR_VERSION,
+    ERL_DRV_EXTENDED_MINOR_VERSION,
+    0,
+    NULL, NULL,
+    stop_select
+};
+
+/* Untility functions */
+
+static int set_blocking_data(ErtsSysDriverData *dd) {
+
+    dd->blocking = erts_alloc(ERTS_ALC_T_SYS_BLOCKING, sizeof(ErtsSysBlocking));
+
+    erts_smp_atomic_add_nob(&sys_misc_mem_sz, sizeof(ErtsSysBlocking));
+
+    dd->blocking->pdl = driver_pdl_create(dd->port_num);
+    dd->blocking->res = 0;
+    dd->blocking->err = 0;
+    dd->blocking->pkey = driver_async_port_key(dd->port_num);
+
+    return 1;
+}
+
+static void init_fd_data(ErtsSysFdData *fd_data, int fd)
+{
+    fd_data->fd = fd;
+    fd_data->buf = NULL;
+    fd_data->cpos = NULL;
+    fd_data->remain = 0;
+    fd_data->sz = 0;
+    fd_data->psz = 0;
+}
+
+static ErtsSysDriverData *
+create_driver_data(ErlDrvPort port_num,
+                   int ifd,
+                   int ofd,
+                   int packet_bytes,
+                   int read_write,
+                   int exit_status,
+                   int pid,
+                   int is_blocking)
+{
+    Port *prt;
+    ErtsSysDriverData *driver_data;
+    char *data;
+    int size = sizeof(ErtsSysDriverData);
+
+    if (read_write & DO_READ)
+        size += sizeof(ErtsSysFdData);
+
+    if ((read_write & DO_WRITE) &&
+        ((ifd != ofd || ofd == -1) || !(read_write & DO_READ)))
+        size += sizeof(ErtsSysFdData);
+
+    data = erts_alloc(ERTS_ALC_T_DRV_TAB,size);
+    erts_smp_atomic_add_nob(&sys_misc_mem_sz, size);
+
+    driver_data = (ErtsSysDriverData*)data;
+    data += sizeof(*driver_data);
+
+    prt = erts_drvport2port(port_num);
+    if (prt != ERTS_INVALID_ERL_DRV_PORT)
+	prt->os_pid = pid;
+
+    driver_data->packet_bytes = packet_bytes;
+    driver_data->port_num = port_num;
+    driver_data->pid = pid;
+    driver_data->alive = exit_status ? 1 : 0;
+    driver_data->status = 0;
+    driver_data->terminating = 0;
+    driver_data->blocking = NULL;
+
+    if (read_write & DO_READ) {
+        driver_data->ifd = (ErtsSysFdData*)data;
+        data += sizeof(*driver_data->ifd);
+        init_fd_data(driver_data->ifd, ifd);
+        driver_select(port_num, ifd, (ERL_DRV_READ|ERL_DRV_USE), 1);
+    } else {
+        driver_data->ifd = NULL;
+    }
+
+    if (read_write & DO_WRITE) {
+        if (ofd != -1 && ifd == ofd && read_write & DO_READ) {
+            /* This is for when ifd and ofd are the same fd */
+            driver_data->ofd = driver_data->ifd;
+        } else {
+            driver_data->ofd = (ErtsSysFdData*)data;
+            data += sizeof(*driver_data->ofd);
+            init_fd_data(driver_data->ofd, ofd);
+        }
+        if (is_blocking && FDBLOCK)
+            if (!set_blocking_data(driver_data)) {
+                erts_free(ERTS_ALC_T_DRV_TAB, driver_data);
+                return NULL;
+            }
+    } else {
+        driver_data->ofd = NULL;
+    }
+
+    return driver_data;
+}
+
+/* Spawn driver */
+
+static void close_pipes(int ifd[2], int ofd[2])
+{
+    close(ifd[0]);
+    close(ifd[1]);
+    close(ofd[0]);
+    close(ofd[1]);
+}
+
+static char **build_unix_environment(char *block)
+{
+    int i;
+    int j;
+    int len;
+    char *cp;
+    char **cpp;
+    char** old_env;
+    
+    ERTS_SMP_LC_ASSERT(erts_smp_lc_rwmtx_is_rlocked(&environ_rwmtx));
+
+    cp = block;
+    len = 0;
+    while (*cp != '\0') {
+	cp += strlen(cp) + 1;
+	len++;
+    }
+    old_env = environ;
+    while (*old_env++ != NULL) {
+	len++;
+    }
+    
+    cpp = (char **) erts_alloc_fnf(ERTS_ALC_T_ENVIRONMENT,
+				   sizeof(char *) * (len+1));
+    if (cpp == NULL) {
+	return NULL;
+    }
+
+    cp = block;
+    len = 0;
+    while (*cp != '\0') {
+	cpp[len] = cp;
+	cp += strlen(cp) + 1;
+	len++;
+    }
+    
+    i = len;
+    for (old_env = environ; *old_env; old_env++) {
+	char* old = *old_env;
+
+	for (j = 0; j < len; j++) {
+	    char *s, *t;
+
+	    /* check if cpp[j] equals old
+	       before the = sign,
+	       i.e.
+	       "TMPDIR=/tmp/" */
+	    s = cpp[j];
+	    t = old;
+	    while (*s == *t && *s != '=') {
+		s++, t++;
+	    }
+	    if (*s == '=' && *t == '=') {
+		break;
+	    }
+	}
+
+	if (j == len) {		/* New version not found */
+	    cpp[len++] = old;
+	}
+    }
+
+    for (j = 0; j < i; ) {
+        size_t last = strlen(cpp[j])-1;
+	if (cpp[j][last] == '=' && strchr(cpp[j], '=') == cpp[j]+last) {
+	    cpp[j] = cpp[--len];
+	    if (len < i) {
+		i--;
+	    } else {
+		j++;
+	    }
+	}
+	else {
+	    j++;
+	}
+    }
+
+    cpp[len] = NULL;
+    return cpp;
+}
+
+static ErlDrvData spawn_start(ErlDrvPort port_num, char* name,
+                              SysDriverOpts* opts)
+{
+#define CMD_LINE_PREFIX_STR "exec "
+#define CMD_LINE_PREFIX_STR_SZ (sizeof(CMD_LINE_PREFIX_STR) - 1)
+
+    int len;
+    char **new_environ;
+    ErtsSysDriverData *dd;
+    char *cmd_line;
+    char wd_buff[MAXPATHLEN+1];
+    char *wd;
+    int ifd[2], ofd[2], stderrfd;
+
+    if (pipe(ifd) < 0) return ERL_DRV_ERROR_ERRNO;
+    errno = EMFILE;		/* default for next three conditions */
+    if (ifd[0] >= sys_max_files() || pipe(ofd) < 0) {
+        close(ifd[0]);
+        close(ifd[1]);
+        return ERL_DRV_ERROR_ERRNO;
+    }
+    if (ofd[1] >= sys_max_files()) {
+        close_pipes(ifd, ofd);
+        errno = EMFILE;
+        return ERL_DRV_ERROR_ERRNO;
+    }
+
+    SET_NONBLOCKING(ifd[0]);
+    SET_NONBLOCKING(ofd[1]);
+
+    stderrfd = opts->redir_stderr ? ifd[1] : dup(2);
+
+    if (stderrfd >= sys_max_files() || stderrfd < 0) {
+        close_pipes(ifd, ofd);
+        if (stderrfd > -1)
+            close(stderrfd);
+        return ERL_DRV_ERROR_ERRNO;
+    }
+
+    if (opts->spawn_type == ERTS_SPAWN_EXECUTABLE) {
+	/* started with spawn_executable, not with spawn */
+	len = strlen(name);
+	cmd_line = (char *) erts_alloc_fnf(ERTS_ALC_T_TMP, len + 1);
+	if (!cmd_line) {
+            close_pipes(ifd, ofd);
+	    errno = ENOMEM;
+	    return ERL_DRV_ERROR_ERRNO;
+	}
+	memcpy((void *) cmd_line,(void *) name, len);
+	cmd_line[len] = '\0';
+	len = len + 1;
+	if (access(cmd_line,X_OK) != 0) {
+	    int save_errno = errno;
+	    erts_free(ERTS_ALC_T_TMP, cmd_line);
+            close_pipes(ifd, ofd);
+	    errno = save_errno;
+	    return ERL_DRV_ERROR_ERRNO;
+	}
+    } else {
+	/* make the string suitable for giving to "sh" */
+	len = strlen(name);
+	cmd_line = (char *) erts_alloc_fnf(ERTS_ALC_T_TMP,
+					   CMD_LINE_PREFIX_STR_SZ + len + 1);
+	if (!cmd_line) {
+            close_pipes(ifd, ofd);
+	    errno = ENOMEM;
+	    return ERL_DRV_ERROR_ERRNO;
+	}
+	memcpy((void *) cmd_line,
+	       (void *) CMD_LINE_PREFIX_STR,
+	       CMD_LINE_PREFIX_STR_SZ);
+	memcpy((void *) (cmd_line + CMD_LINE_PREFIX_STR_SZ), (void *) name, len);
+	cmd_line[CMD_LINE_PREFIX_STR_SZ + len] = '\0';
+	len = CMD_LINE_PREFIX_STR_SZ + len + 1;
+    }
+
+    erts_smp_rwmtx_rlock(&environ_rwmtx);
+
+    if (opts->envir == NULL) {
+	new_environ = environ;
+    } else if ((new_environ = build_unix_environment(opts->envir)) == NULL) {
+	erts_smp_rwmtx_runlock(&environ_rwmtx);
+        close_pipes(ifd, ofd);
+	erts_free(ERTS_ALC_T_TMP, (void *) cmd_line);
+	errno = ENOMEM;
+	return ERL_DRV_ERROR_ERRNO;
+    }
+
+    if (opts->wd == NULL) {
+        if ((wd = getcwd(wd_buff, MAXPATHLEN+1)) == NULL) {
+            /* on some OSs this call opens a fd in the
+               background which means that this can
+               return EMFILE */
+            int err = errno;
+            close_pipes(ifd, ofd);
+            erts_free(ERTS_ALC_T_TMP, (void *) cmd_line);
+            if (new_environ != environ)
+                erts_free(ERTS_ALC_T_ENVIRONMENT, (void *) new_environ);
+            erts_smp_rwmtx_runlock(&environ_rwmtx);
+            errno = err;
+            return ERL_DRV_ERROR_ERRNO;
+        }
+    } else {
+        wd = opts->wd;
+    }
+
+    {
+        struct iovec *io_vector;
+        int iov_len = 5;
+        char nullbuff[] = "\0";
+        int j, i = 0, res;
+        Sint32 buffsz = 0, env_len = 0, argv_len = 0,
+            flags = (opts->use_stdio ? FORKER_FLAG_USE_STDIO : 0)
+            | (opts->exit_status ? FORKER_FLAG_EXIT_STATUS : 0)
+            | (opts->read_write & DO_READ ? FORKER_FLAG_DO_READ : 0)
+            | (opts->read_write & DO_WRITE ? FORKER_FLAG_DO_WRITE : 0);
+
+        /* count number of elements in environment */
+        while(new_environ[env_len] != NULL)
+            env_len++;
+        iov_len += 1 + env_len; /* num envs including size int */
+
+        /* count number of element in argument list */
+        if (opts->spawn_type == ERTS_SPAWN_EXECUTABLE) {
+            if (opts->argv != NULL) {
+                while(opts->argv[argv_len] != NULL)
+                    argv_len++;
+            } else {
+                argv_len++;
+            }
+            iov_len += 1 + argv_len; /* num argvs including size int */
+        }
+
+        io_vector = erts_alloc_fnf(ERTS_ALC_T_TMP, sizeof(struct iovec) * iov_len);
+
+        if (!io_vector) {
+            close_pipes(ifd, ofd);
+            erts_smp_rwmtx_runlock(&environ_rwmtx);
+            erts_free(ERTS_ALC_T_TMP, (void *) cmd_line);
+            if (new_environ != environ)
+                erts_free(ERTS_ALC_T_ENVIRONMENT, (void *) new_environ);
+            errno = ENOMEM;
+            return ERL_DRV_ERROR_ERRNO;
+        }
+
+        io_vector[i].iov_base = (void*)&buffsz;
+        io_vector[i++].iov_len = sizeof(buffsz);
+
+        io_vector[i].iov_base = (void*)&flags;
+        flags = htonl(flags);
+        io_vector[i++].iov_len = sizeof(flags);
+        buffsz += sizeof(flags);
+
+        io_vector[i].iov_base = cmd_line;
+        io_vector[i++].iov_len = len;
+        buffsz += len;
+
+        io_vector[i].iov_base = wd;
+        io_vector[i].iov_len = strlen(io_vector[i].iov_base) + 1;
+        buffsz += io_vector[i++].iov_len;
+
+        io_vector[i].iov_base = nullbuff;
+        io_vector[i++].iov_len = 1;
+        buffsz += io_vector[i-1].iov_len;
+
+        io_vector[i].iov_base = (void*)&env_len;
+        env_len = htonl(env_len);
+        io_vector[i++].iov_len = sizeof(env_len);
+        buffsz += io_vector[i-1].iov_len;
+
+        for (j = 0; new_environ[j] != NULL; j++) {
+            io_vector[i].iov_base = new_environ[j];
+            io_vector[i++].iov_len = strlen(new_environ[j]) + 1;
+            buffsz += io_vector[i-1].iov_len;
+        }
+
+        /* only append arguments if this was a spawn_executable */
+        if (opts->spawn_type == ERTS_SPAWN_EXECUTABLE) {
+
+            io_vector[i].iov_base = (void*)&argv_len;
+            argv_len = htonl(argv_len);
+            io_vector[i++].iov_len = sizeof(argv_len);
+            buffsz += io_vector[i-1].iov_len;
+
+            if (opts->argv) {
+                /* If there are arguments we copy in the references to
+                   them into the iov */
+                for (j = 0; opts->argv[j]; j++) {
+                    if (opts->argv[j] == erts_default_arg0)
+                        io_vector[i].iov_base = cmd_line;
+                    else
+                        io_vector[i].iov_base = opts->argv[j];
+                    io_vector[i].iov_len = strlen(io_vector[i].iov_base) + 1;
+                    buffsz += io_vector[i++].iov_len;
+                }
+            } else {
+                io_vector[i].iov_base = cmd_line;
+                io_vector[i].iov_len = strlen(io_vector[i].iov_base) + 1;
+                buffsz += io_vector[i++].iov_len;
+            }
+        }
+
+        /* we send the request to do the fork */
+        if ((res = writev(ofd[1], io_vector, iov_len > MAXIOV ? MAXIOV : iov_len)) < 0) {
+            if (errno == ERRNO_BLOCK) {
+                res = 0;
+            } else {
+                int err = errno;
+                close_pipes(ifd, ofd);
+                erts_free(ERTS_ALC_T_TMP, io_vector);
+                if (new_environ != environ)
+                    erts_free(ERTS_ALC_T_ENVIRONMENT, (void *) new_environ);
+                erts_smp_rwmtx_runlock(&environ_rwmtx);
+                erts_free(ERTS_ALC_T_TMP, (void *) cmd_line);
+                errno = err;
+                return ERL_DRV_ERROR_ERRNO;
+            }
+        }
+
+        if (res < buffsz) {
+            /* we only wrote part of the command payload. Enqueue the rest. */
+            for (i = 0; i < iov_len; i++) {
+                driver_enq(port_num, io_vector[i].iov_base, io_vector[i].iov_len);
+            }
+            driver_deq(port_num, res);
+            driver_select(port_num, ofd[1], ERL_DRV_WRITE|ERL_DRV_USE, 1);
+        }
+
+        erts_free(ERTS_ALC_T_TMP, io_vector);
+    }
+
+    erts_free(ERTS_ALC_T_TMP, (void *) cmd_line);
+
+    if (new_environ != environ)
+	erts_free(ERTS_ALC_T_ENVIRONMENT, (void *) new_environ);
+
+    erts_smp_rwmtx_runlock(&environ_rwmtx);
+
+    dd = create_driver_data(port_num, ifd[0], ofd[1], opts->packet_bytes,
+                             DO_WRITE | DO_READ, opts->exit_status,
+                             0, 0);
+
+    {
+        /* send ofd[0] + ifd[1] + stderrfd to forker port */
+        ErtsSysForkerProto *proto =
+            erts_alloc(ERTS_ALC_T_DRV_CTRL_DATA,
+                       sizeof(ErtsSysForkerProto));
+        memset(proto, 0, sizeof(ErtsSysForkerProto));
+        proto->action = ErtsSysForkerProtoAction_Start;
+        proto->u.start.fds[0] = ofd[0];
+        proto->u.start.fds[1] = ifd[1];
+        proto->u.start.fds[2] = stderrfd;
+        proto->u.start.port_id = opts->exit_status ? erts_drvport2id(port_num) : THE_NON_VALUE;
+        if (erl_drv_port_control(forker_port, 'S', (char*)proto, sizeof(*proto))) {
+            /* The forker port has been killed, we close both fd's which will
+               make open_port throw an epipe error */
+            close(ofd[0]);
+            close(ifd[1]);
+        }
+    }
+
+    /* we set these fds to negative to mark if
+       they should be closed after the handshake */
+    if (!(opts->read_write & DO_READ))
+        dd->ifd->fd *= -1;
+
+    if (!(opts->read_write & DO_WRITE))
+        dd->ofd->fd *= -1;
+
+    return (ErlDrvData)dd;
+#undef CMD_LINE_PREFIX_STR
+#undef CMD_LINE_PREFIX_STR_SZ
+}
+
+static ErlDrvSSizeT spawn_control(ErlDrvData e, unsigned int cmd, char *buf,
+                                  ErlDrvSizeT len, char **rbuf, ErlDrvSizeT rlen)
+{
+    ErtsSysDriverData *dd = (ErtsSysDriverData*)e;
+    ErtsSysForkerProto *proto = (ErtsSysForkerProto *)buf;
+
+    ASSERT(len == sizeof(*proto));
+    ASSERT(proto->action == ErtsSysForkerProtoAction_SigChld);
+
+    dd->status = proto->u.sigchld.error_number;
+    dd->alive = -1;
+
+    if (dd->ifd)
+        driver_select(dd->port_num, abs(dd->ifd->fd), ERL_DRV_READ | ERL_DRV_USE, 1);
+
+    if (dd->ofd)
+        driver_select(dd->port_num, abs(dd->ofd->fd), ERL_DRV_WRITE | ERL_DRV_USE, 1);
+
+    return 0;
+}
+
+#define FD_DEF_HEIGHT 24
+#define FD_DEF_WIDTH 80
+/* Control op */
+#define FD_CTRL_OP_GET_WINSIZE 100
+
+static int fd_get_window_size(int fd, Uint32 *width, Uint32 *height)
+{
+#ifdef TIOCGWINSZ 
+    struct winsize ws;
+    if (ioctl(fd,TIOCGWINSZ,&ws) == 0) {
+	*width = (Uint32) ws.ws_col;
+	*height = (Uint32) ws.ws_row;
+	return 0;
+    }
+#endif
+    return -1;
+}
+
+static ErlDrvSSizeT fd_control(ErlDrvData drv_data,
+			       unsigned int command,
+			       char *buf, ErlDrvSizeT len,
+			       char **rbuf, ErlDrvSizeT rlen)
+{
+    int fd = (int)(long)drv_data;
+    char resbuff[2*sizeof(Uint32)];
+    switch (command) {
+    case FD_CTRL_OP_GET_WINSIZE:
+	{
+	    Uint32 w,h;
+	    if (fd_get_window_size(fd,&w,&h)) 
+		return 0;
+	    memcpy(resbuff,&w,sizeof(Uint32));
+	    memcpy(resbuff+sizeof(Uint32),&h,sizeof(Uint32));
+	}
+	break;
+    default:
+	return 0;
+    }
+    if (rlen < 2*sizeof(Uint32)) {
+	*rbuf = driver_alloc(2*sizeof(Uint32));
+    }
+    memcpy(*rbuf,resbuff,2*sizeof(Uint32));
+    return 2*sizeof(Uint32);
+}
+
+static ErlDrvData fd_start(ErlDrvPort port_num, char* name,
+			   SysDriverOpts* opts)
+{
+    int non_blocking = 0;
+
+    if (((opts->read_write & DO_READ) && opts->ifd >= sys_max_files()) ||
+	((opts->read_write & DO_WRITE) && opts->ofd >= sys_max_files()))
+	return ERL_DRV_ERROR_GENERAL;
+
+    /*
+     * Historical:
+     *
+     * "Note about nonblocking I/O.
+     *
+     * At least on Solaris, setting the write end of a TTY to nonblocking,
+     * will set the input end to nonblocking as well (and vice-versa).
+     * If erl is run in a pipeline like this:  cat | erl
+     * the input end of the TTY will be the standard input of cat.
+     * And cat is not prepared to handle nonblocking I/O."
+     *
+     * Actually, the reason for this is not that the tty itself gets set
+     * in non-blocking mode, but that the "input end" (cat's stdin) and
+     * the "output end" (erlang's stdout) are typically the "same" file
+     * descriptor, dup()'ed from a single fd by one of this process'
+     * ancestors.
+     *
+     * The workaround for this problem used to be a rather bad kludge,
+     * interposing an extra process ("internal cat") between erlang's
+     * stdout and the original stdout, allowing erlang to set its stdout
+     * in non-blocking mode without affecting the stdin of the preceding
+     * process in the pipeline - and being a kludge, it caused all kinds
+     * of weird problems.
+     *
+     * So, this is the current logic:
+     *
+     * The only reason to set non-blocking mode on the output fd at all is
+     * if it's something that can cause a write() to block, of course,
+     * i.e. primarily if it points to a tty, socket, pipe, or fifo. 
+     *
+     * If we don't set non-blocking mode when we "should" have, and output
+     * becomes blocked, the entire runtime system will be suspended - this
+     * is normally bad of course, and can happen fairly "easily" - e.g. user
+     * hits ^S on tty - but doesn't necessarily happen.
+     * 
+     * If we do set non-blocking mode when we "shouldn't" have, the runtime
+     * system will end up seeing EOF on the input fd (due to the preceding
+     * process dying), which typically will cause the entire runtime system
+     * to terminate immediately (due to whatever erlang process is seeing
+     * the EOF taking it as a signal to halt the system). This is *very* bad.
+     * 
+     * I.e. we should take a conservative approach, and only set non-
+     * blocking mode when we a) need to, and b) are reasonably certain
+     * that it won't be a problem. And as in the example above, the problem
+     * occurs when input fd and output fd point to different "things".
+     *
+     * However, determining that they are not just the same "type" of
+     * "thing", but actually the same instance of that type of thing, is
+     * unreasonably complex in many/most cases.
+     *
+     * Also, with pipes, sockets, and fifos it's far from obvious that the
+     * user *wants* non-blocking output: If you're running erlang inside
+     * some complex pipeline, you're probably not running a real-time system
+     * that must never stop, but rather *want* it to suspend if the output
+     * channel is "full".
+     *
+     * So, the bottom line: We will only set the output fd non-blocking if
+     * it points to a tty, and either a) the input fd also points to a tty,
+     * or b) we can make sure that setting the output fd non-blocking
+     * doesn't interfere with someone else's input, via a somewhat milder
+     * kludge than the above.
+     *
+     * Also keep in mind that while this code is almost exclusively run as
+     * a result of an erlang open_port({fd,0,1}, ...), that isn't the only
+     * case - it can be called with any old pre-existing file descriptors,
+     * the relations between which (if they're even two) we can only guess
+     * at - still, we try our best...
+     *
+     * Added note OTP 18: Some systems seem to use stdout/stderr to log data
+     * using unix pipes, so we cannot allow the system to block on a write.
+     * Therefore we use an async thread to write the data to fd's that could
+     * not be set to non-blocking. When no async threads are available we
+     * fall back on the old behaviour.
+     *
+     * Also the guarantee about what is delivered to the OS has changed.
+     * Pre 18 the fd driver did no flushing of data before terminating.
+     * Now it does. This is because we want to be able to guarantee that things
+     * such as escripts and friends really have outputted all data before
+     * terminating. This could potentially block the termination of the system
+     * for a very long time, but if the user wants to terminate fast she should
+     * use erlang:halt with flush=false.
+     */
+
+    /* Try to figure out if we can use non-blocking writes */
+    if (opts->read_write & DO_WRITE) {
+
+	/* If we don't have a read end, all bets are off - no non-blocking. */
+	if (opts->read_write & DO_READ) {
+
+	    if (isatty(opts->ofd)) { /* output fd is a tty:-) */
+
+		if (isatty(opts->ifd)) { /* input fd is also a tty */
+
+		    /* To really do this "right", we should also check that
+		       input and output fd point to the *same* tty - but
+		       this seems like overkill; ttyname() isn't for free,
+		       and this is a very common case - and it's hard to
+		       imagine a scenario where setting non-blocking mode
+		       here would cause problems - go ahead and do it. */
+
+                    non_blocking = 1;
+		    SET_NONBLOCKING(opts->ofd);
+
+		} else {	/* output fd is a tty, input fd isn't */
+
+		    /* This is a "problem case", but also common (see the
+		       example above) - i.e. it makes sense to try a bit
+		       harder before giving up on non-blocking mode: Try to
+		       re-open the tty that the output fd points to, and if
+		       successful replace the original one with the "new" fd
+		       obtained this way, and set *that* one in non-blocking
+		       mode. (Yes, this is a kludge.)
+
+		       However, re-opening the tty may fail in a couple of
+		       (unusual) cases:
+
+		       1) The name of the tty (or an equivalent one, i.e.
+			  same major/minor number) can't be found, because
+			  it actually lives somewhere other than /dev (or
+			  wherever ttyname() looks for it), and isn't
+			  equivalent to any of those that do live in the
+			  "standard" place - this should be *very* unusual.
+
+		       2) Permissions on the tty don't allow us to open it -
+			  it's perfectly possible to have an fd open to an
+			  object whose permissions wouldn't allow us to open
+			  it. This is not as unusual as it sounds, one case
+			  is if the user has su'ed to someone else (not
+			  root) - we have a read/write fd open to the tty
+			  (because it has been inherited all the way down
+			  here), but we have neither read nor write
+			  permission for the tty.
+
+		       In these cases, we finally give up, and don't set the
+		       output fd in non-blocking mode. */
+
+		    char *tty;
+		    int nfd;
+
+		    if ((tty = ttyname(opts->ofd)) != NULL &&
+			(nfd = open(tty, O_WRONLY)) != -1) {
+			dup2(nfd, opts->ofd);
+			close(nfd);
+                        non_blocking = 1;
+			SET_NONBLOCKING(opts->ofd);
+		    }
+		}
+	    }
+	}
+    }
+    return (ErlDrvData)create_driver_data(port_num, opts->ifd, opts->ofd,
+                                          opts->packet_bytes,
+                                          opts->read_write, 0, -1,
+                                          !non_blocking);
+}
+
+static void clear_fd_data(ErtsSysFdData *fdd)
+{
+    if (fdd->sz > 0) {
+	erts_free(ERTS_ALC_T_FD_ENTRY_BUF, (void *) fdd->buf);
+	ASSERT(erts_smp_atomic_read_nob(&sys_misc_mem_sz) >= fdd->sz);
+	erts_smp_atomic_add_nob(&sys_misc_mem_sz, -1*fdd->sz);
+    }
+    fdd->buf = NULL;
+    fdd->sz = 0;
+    fdd->remain = 0;
+    fdd->cpos = NULL;
+    fdd->psz = 0;
+}
+
+static void nbio_stop_fd(ErlDrvPort prt, ErtsSysFdData *fdd)
+{
+    driver_select(prt, abs(fdd->fd), DO_READ|DO_WRITE, 0);
+    clear_fd_data(fdd);
+    SET_BLOCKING(abs(fdd->fd));
+
+}
+
+static void fd_stop(ErlDrvData ev)  /* Does not close the fds */
+{
+    ErtsSysDriverData* dd = (ErtsSysDriverData*)ev;
+    ErlDrvPort prt = dd->port_num;
+    int sz = sizeof(ErtsSysDriverData);
+
+#if FDBLOCK
+    if (dd->blocking) {
+        erts_free(ERTS_ALC_T_SYS_BLOCKING, dd->blocking);
+        dd->blocking = NULL;
+        sz += sizeof(ErtsSysBlocking);
+    }
+#endif
+
+    if (dd->ifd) {
+        sz += sizeof(ErtsSysFdData);
+        nbio_stop_fd(prt, dd->ifd);
+    }
+    if (dd->ofd && dd->ofd != dd->ifd) {
+        sz += sizeof(ErtsSysFdData);
+        nbio_stop_fd(prt, dd->ofd);
+    }
+
+     erts_free(ERTS_ALC_T_DRV_TAB, dd);
+     erts_smp_atomic_add_nob(&sys_misc_mem_sz, -sz);
+}
+
+static void fd_flush(ErlDrvData ev)
+{
+    ErtsSysDriverData* dd = (ErtsSysDriverData*)ev;
+    if (!dd->terminating)
+        dd->terminating = 1;
+}
+
+static ErlDrvData vanilla_start(ErlDrvPort port_num, char* name,
+				SysDriverOpts* opts)
+{
+    int flags, fd;
+    ErlDrvData res;
+
+    flags = (opts->read_write == DO_READ ? O_RDONLY :
+	     opts->read_write == DO_WRITE ? O_WRONLY|O_CREAT|O_TRUNC :
+	     O_RDWR|O_CREAT);
+    if ((fd = open(name, flags, 0666)) < 0)
+	return ERL_DRV_ERROR_GENERAL;
+    if (fd >= sys_max_files()) {
+	close(fd);
+	return ERL_DRV_ERROR_GENERAL;
+    }
+    SET_NONBLOCKING(fd);
+
+    res = (ErlDrvData)(long)create_driver_data(port_num, fd, fd,
+                                               opts->packet_bytes,
+                                               opts->read_write, 0, -1, 0);
+    return res;
+}
+
+/* Note that driver_data[fd].ifd == fd if the port was opened for reading, */
+/* otherwise (i.e. write only) driver_data[fd].ofd = fd.  */
+
+static void stop(ErlDrvData ev)
+{
+    ErtsSysDriverData* dd = (ErtsSysDriverData*)ev;
+    ErlDrvPort prt = dd->port_num;
+
+    if (dd->ifd) {
+        nbio_stop_fd(prt, dd->ifd);
+        driver_select(prt, abs(dd->ifd->fd), ERL_DRV_USE, 0);  /* close(ifd); */
+    }
+
+    if (dd->ofd && dd->ofd != dd->ifd) {
+	nbio_stop_fd(prt, dd->ofd);
+	driver_select(prt, abs(dd->ofd->fd), ERL_DRV_USE, 0);  /* close(ofd); */
+    }
+
+    erts_free(ERTS_ALC_T_DRV_TAB, dd);
+}
+
+/* used by fd_driver */
+static void outputv(ErlDrvData e, ErlIOVec* ev)
+{
+    ErtsSysDriverData *dd = (ErtsSysDriverData*)e;
+    ErlDrvPort ix = dd->port_num;
+    int pb = dd->packet_bytes;
+    int ofd = dd->ofd ? dd->ofd->fd : -1;
+    ssize_t n;
+    ErlDrvSizeT sz;
+    char lb[4];
+    char* lbp;
+    ErlDrvSizeT len = ev->size;
+
+    /* (len > ((unsigned long)-1 >> (4-pb)*8)) */
+    /*    if (pb >= 0 && (len & (((ErlDrvSizeT)1 << (pb*8))) - 1) != len) {*/
+    if (((pb == 2) && (len > 0xffff)) || (pb == 1 && len > 0xff)) {
+	driver_failure_posix(ix, EINVAL);
+	return; /* -1; */
+    }
+    /* Handles 0 <= pb <= 4 only */
+    put_int32((Uint32) len, lb);
+    lbp = lb + (4-pb);
+
+    ev->iov[0].iov_base = lbp;
+    ev->iov[0].iov_len = pb;
+    ev->size += pb;
+
+    if (dd->blocking && FDBLOCK)
+        driver_pdl_lock(dd->blocking->pdl);
+
+    if ((sz = driver_sizeq(ix)) > 0) {
+	driver_enqv(ix, ev, 0);
+
+        if (dd->blocking && FDBLOCK)
+            driver_pdl_unlock(dd->blocking->pdl);
+
+	if (sz + ev->size >= (1 << 13))
+	    set_busy_port(ix, 1);
+    }
+    else if (!dd->blocking || !FDBLOCK) {
+        /* We try to write directly if the fd in non-blocking */
+	int vsize = ev->vsize > MAX_VSIZE ? MAX_VSIZE : ev->vsize;
+
+	n = writev(ofd, (const void *) (ev->iov), vsize);
+	if (n == ev->size)
+	    return; /* 0;*/
+	if (n < 0) {
+	    if ((errno != EINTR) && (errno != ERRNO_BLOCK)) {
+		driver_failure_posix(ix, errno);
+		return; /* -1;*/
+	    }
+	    n = 0;
+	}
+	driver_enqv(ix, ev, n);  /* n is the skip value */
+	driver_select(ix, ofd, ERL_DRV_WRITE|ERL_DRV_USE, 1);
+    }
+#if FDBLOCK
+    else {
+        if (ev->size != 0) {
+            driver_enqv(ix, ev, 0);
+            driver_pdl_unlock(dd->blocking->pdl);
+            driver_async(ix, &dd->blocking->pkey,
+                         fd_async, dd, NULL);
+        } else {
+            driver_pdl_unlock(dd->blocking->pdl);
+        }
+    }
+#endif
+    /* return 0;*/
+}
+
+/* Used by spawn_driver and vanilla driver */
+static void output(ErlDrvData e, char* buf, ErlDrvSizeT len)
+{
+    ErtsSysDriverData *dd = (ErtsSysDriverData*)e;
+    ErlDrvPort ix = dd->port_num;
+    int pb = dd->packet_bytes;
+    int ofd = dd->ofd ? dd->ofd->fd : -1;
+    ssize_t n;
+    ErlDrvSizeT sz;
+    char lb[4];
+    char* lbp;
+    struct iovec iv[2];
+
+    /* (len > ((unsigned long)-1 >> (4-pb)*8)) */
+    if (((pb == 2) && (len > 0xffff))
+        || (pb == 1 && len > 0xff)
+        || dd->pid == 0 /* Attempt at output before port is ready */) {
+	driver_failure_posix(ix, EINVAL);
+	return; /* -1; */
+    }
+    put_int32(len, lb);
+    lbp = lb + (4-pb);
+
+    if ((sz = driver_sizeq(ix)) > 0) {
+	driver_enq(ix, lbp, pb);
+	driver_enq(ix, buf, len);
+	if (sz + len + pb >= (1 << 13))
+	    set_busy_port(ix, 1);
+    }
+    else {
+	iv[0].iov_base = lbp;
+	iv[0].iov_len = pb;  /* should work for pb=0 */
+	iv[1].iov_base = buf;
+	iv[1].iov_len = len;
+	n = writev(ofd, iv, 2);
+	if (n == pb+len)
+	    return; /* 0; */
+	if (n < 0) {
+	    if ((errno != EINTR) && (errno != ERRNO_BLOCK)) {
+		driver_failure_posix(ix, errno);
+		return; /* -1; */
+	    }
+	    n = 0;
+	}
+	if (n < pb) {
+	    driver_enq(ix, lbp+n, pb-n);
+	    driver_enq(ix, buf, len);
+	}
+	else {
+	    n -= pb;
+	    driver_enq(ix, buf+n, len-n);
+	}
+	driver_select(ix, ofd, ERL_DRV_WRITE|ERL_DRV_USE, 1);
+    }
+    return; /* 0; */
+}
+
+static int port_inp_failure(ErtsSysDriverData *dd, int res)
+				/* Result: 0 (eof) or -1 (error) */
+{
+    int err = errno;
+
+    ASSERT(res <= 0);
+    if (dd->ifd) {
+        driver_select(dd->port_num, dd->ifd->fd, ERL_DRV_READ|ERL_DRV_WRITE, 0);
+        clear_fd_data(dd->ifd);
+    }
+
+    if (dd->blocking && FDBLOCK) {
+        driver_pdl_lock(dd->blocking->pdl);
+        if (driver_sizeq(dd->port_num) > 0) {
+            driver_pdl_unlock(dd->blocking->pdl);
+            /* We have stuff in the output queue, so we just
+               set the state to terminating and wait for fd_async_ready
+               to terminate the port */
+            if (res == 0)
+                dd->terminating = 2;
+            else
+                dd->terminating = -err;
+            return 0;
+        }
+        driver_pdl_unlock(dd->blocking->pdl);
+    }
+
+    if (res == 0) {
+        if (dd->alive == 1) {
+            /*
+             * We have eof and want to report exit status, but the process
+             * hasn't exited yet. When it does ready_input will
+             * driver_select() this fd which will make sure that we get
+             * back here with dd->alive == -1 and dd->status set.
+             */
+            return 0;
+        }
+        else if (dd->alive == -1) {
+            int status = dd->status;
+
+            /* We need not be prepared for stopped/continued processes. */
+            if (WIFSIGNALED(status))
+                status = 128 + WTERMSIG(status);
+            else
+                status = WEXITSTATUS(status);
+            driver_report_exit(dd->port_num, status);
+        }
+       driver_failure_eof(dd->port_num);
+    } else if (dd->ifd) {
+        erl_drv_init_ack(dd->port_num, ERL_DRV_ERROR_ERRNO);
+    } else {
+	driver_failure_posix(dd->port_num, err);
+    }
+    return 0;
+}
+
+/* fd is the drv_data that is returned from the */
+/* initial start routine                        */
+/* ready_fd is the descriptor that is ready to read */
+
+static void ready_input(ErlDrvData e, ErlDrvEvent ready_fd)
+{
+    ErtsSysDriverData *dd = (ErtsSysDriverData*)e;
+    ErlDrvPort port_num;
+    int packet_bytes;
+    int res;
+    Uint h;
+
+    port_num = dd->port_num;
+    packet_bytes = dd->packet_bytes;
+
+    ASSERT(abs(dd->ifd->fd) == ready_fd);
+
+    if (dd->pid == 0) {
+        /* the pid is sent from erl_child_setup. spawn driver only. */
+        ErtsSysForkerProto proto;
+        int res;
+
+        if((res = read(ready_fd, &proto, sizeof(proto))) <= 0) {
+            /* hmm, child setup seems to have closed the pipe too early...
+               we close the port as there is not much else we can do */
+            if (res < 0 && errno == ERRNO_BLOCK)
+                return;
+            driver_select(port_num, ready_fd, ERL_DRV_READ, 0);
+            if (res == 0)
+                errno = EPIPE;
+            port_inp_failure(dd, -1);
+            return;
+        }
+
+        ASSERT(proto.action == ErtsSysForkerProtoAction_Go);
+        dd->pid = proto.u.go.os_pid;
+
+        if (dd->pid == -1) {
+            /* Setup failed! The only reason why this should happen is if
+               the fork fails. */
+            errno = proto.u.go.error_number;
+            port_inp_failure(dd, -1);
+            return;
+        }
+
+        proto.action = ErtsSysForkerProtoAction_Ack;
+
+        if (driver_sizeq(port_num) > 0) {
+            driver_enq(port_num, (char*)&proto, sizeof(proto));
+            } else {
+                if (write(abs(dd->ofd->fd), &proto, sizeof(proto)) < 0)
+                    if (errno == ERRNO_BLOCK || errno == EINTR)
+                        driver_enq(port_num, (char*)&proto, sizeof(proto));
+                /* do nothing on failure here. If the ofd is broken, then
+                   the ifd will probably also be broken and trigger
+                   a port_inp_failure */
+            }
+
+            if (dd->ifd->fd < 0) {
+                driver_select(port_num, abs(dd->ifd->fd), ERL_DRV_READ|ERL_DRV_USE, 0);
+                erts_smp_atomic_add_nob(&sys_misc_mem_sz, -sizeof(ErtsSysFdData));
+                dd->ifd = NULL;
+            }
+
+            if (dd->ofd->fd < 0  || driver_sizeq(port_num) > 0)
+                /* we select in order to close fd or write to queue,
+                   child setup will close this fd if fd < 0 */
+                driver_select(port_num, abs(dd->ofd->fd), ERL_DRV_WRITE|ERL_DRV_USE, 1);
+
+            erl_drv_set_os_pid(port_num, dd->pid);
+            erl_drv_init_ack(port_num, e);
+            return;
+    }
+
+    if (packet_bytes == 0) {
+	byte *read_buf = (byte *) erts_alloc(ERTS_ALC_T_SYS_READ_BUF,
+					     ERTS_SYS_READ_BUF_SZ);
+	res = read(ready_fd, read_buf, ERTS_SYS_READ_BUF_SZ);
+	if (res < 0) {
+	    if ((errno != EINTR) && (errno != ERRNO_BLOCK))
+		port_inp_failure(dd, res);
+	}
+	else if (res == 0)
+	    port_inp_failure(dd, res);
+	else
+	    driver_output(port_num, (char*) read_buf, res);
+	erts_free(ERTS_ALC_T_SYS_READ_BUF, (void *) read_buf);
+    }
+    else if (dd->ifd->remain > 0) { /* We try to read the remainder */
+	/* space is allocated in buf */
+	res = read(ready_fd, dd->ifd->cpos,
+		   dd->ifd->remain);
+	if (res < 0) {
+	    if ((errno != EINTR) && (errno != ERRNO_BLOCK))
+		port_inp_failure(dd, res);
+	}
+	else if (res == 0) {
+	    port_inp_failure(dd, res);
+	}
+	else if (res == dd->ifd->remain) { /* we're done  */
+	    driver_output(port_num, dd->ifd->buf,
+			  dd->ifd->sz);
+	    clear_fd_data(dd->ifd);
+	}
+	else { /*  if (res < dd->ifd->remain) */
+	    dd->ifd->cpos += res;
+	    dd->ifd->remain -= res;
+	}
+    }
+    else if (dd->ifd->remain == 0) { /* clean fd */
+	byte *read_buf = (byte *) erts_alloc(ERTS_ALC_T_SYS_READ_BUF,
+					     ERTS_SYS_READ_BUF_SZ);
+	/* We make one read attempt and see what happens */
+	res = read(ready_fd, read_buf, ERTS_SYS_READ_BUF_SZ);
+	if (res < 0) {
+	    if ((errno != EINTR) && (errno != ERRNO_BLOCK))
+		port_inp_failure(dd, res);
+	}
+	else if (res == 0) {     	/* eof */
+	    port_inp_failure(dd, res);
+	}
+	else if (res < packet_bytes - dd->ifd->psz) {
+	    memcpy(dd->ifd->pbuf+dd->ifd->psz,
+		   read_buf, res);
+	    dd->ifd->psz += res;
+	}
+	else  { /* if (res >= packet_bytes) */
+	    unsigned char* cpos = read_buf;
+	    int bytes_left = res;
+
+	    while (1) {
+		int psz = dd->ifd->psz;
+		char* pbp = dd->ifd->pbuf + psz;
+
+		while(bytes_left && (psz < packet_bytes)) {
+		    *pbp++ = *cpos++;
+		    bytes_left--;
+		    psz++;
+		}
+
+		if (psz < packet_bytes) {
+		    dd->ifd->psz = psz;
+		    break;
+		}
+		dd->ifd->psz = 0;
+
+		switch (packet_bytes) {
+		case 1: h = get_int8(dd->ifd->pbuf);  break;
+		case 2: h = get_int16(dd->ifd->pbuf); break;
+		case 4: h = get_int32(dd->ifd->pbuf); break;
+		default: ASSERT(0); return; /* -1; */
+		}
+
+		if (h <= (bytes_left)) {
+		    driver_output(port_num, (char*) cpos, h);
+		    cpos += h;
+		    bytes_left -= h;
+		    continue;
+		}
+		else {		/* The last message we got was split */
+		        char *buf = erts_alloc_fnf(ERTS_ALC_T_FD_ENTRY_BUF, h);
+		    if (!buf) {
+			errno = ENOMEM;
+			port_inp_failure(dd, -1);
+		    }
+		    else {
+			erts_smp_atomic_add_nob(&sys_misc_mem_sz, h);
+			sys_memcpy(buf, cpos, bytes_left);
+			dd->ifd->buf = buf;
+			dd->ifd->sz = h;
+			dd->ifd->remain = h - bytes_left;
+			dd->ifd->cpos = buf + bytes_left;
+		    }
+		    break;
+		}
+	    }
+	}
+	erts_free(ERTS_ALC_T_SYS_READ_BUF, (void *) read_buf);
+    }
+}
+
+
+/* fd is the drv_data that is returned from the */
+/* initial start routine                        */
+/* ready_fd is the descriptor that is ready to read */
+
+static void ready_output(ErlDrvData e, ErlDrvEvent ready_fd)
+{
+    ErtsSysDriverData *dd = (ErtsSysDriverData*)e;
+    ErlDrvPort ix = dd->port_num;
+    int n;
+    struct iovec* iv;
+    int vsize;
+
+    if ((iv = (struct iovec*) driver_peekq(ix, &vsize)) == NULL) {
+	driver_select(ix, ready_fd, ERL_DRV_WRITE, 0);
+        if (dd->pid > 0 && dd->ofd->fd < 0) {
+            /* The port was opened with 'in' option, which means we
+               should close the output fd as soon as the command has
+               been sent. */
+            driver_select(ix, ready_fd, ERL_DRV_WRITE|ERL_DRV_USE, 0);
+            erts_smp_atomic_add_nob(&sys_misc_mem_sz, -sizeof(ErtsSysFdData));
+            dd->ofd = NULL;
+        }
+        if (dd->terminating)
+            driver_failure_atom(dd->port_num,"normal");
+	return; /* 0; */
+    }
+    vsize = vsize > MAX_VSIZE ? MAX_VSIZE : vsize;
+    if ((n = writev(ready_fd, iv, vsize)) > 0) {
+	if (driver_deq(ix, n) == 0)
+	    set_busy_port(ix, 0);
+    }
+    else if (n < 0) {
+	if (errno == ERRNO_BLOCK || errno == EINTR)
+	    return; /* 0; */
+	else {
+	    int res = errno;
+	    driver_select(ix, ready_fd, ERL_DRV_WRITE, 0);
+	    driver_failure_posix(ix, res);
+	    return; /* -1; */
+	}
+    }
+    return; /* 0; */
+}
+
+static void stop_select(ErlDrvEvent fd, void* _)
+{
+    close((int)fd);
+}
+
+#if FDBLOCK
+
+static void
+fd_async(void *async_data)
+{
+    int res;
+    ErtsSysDriverData *dd = (ErtsSysDriverData *)async_data;
+    SysIOVec      *iov0;
+    SysIOVec      *iov;
+    int            iovlen;
+    int            err = 0;
+    /* much of this code is stolen from efile_drv:invoke_writev */
+    driver_pdl_lock(dd->blocking->pdl);
+    iov0 = driver_peekq(dd->port_num, &iovlen);
+    iovlen = iovlen < MAXIOV ? iovlen : MAXIOV;
+    iov = erts_alloc_fnf(ERTS_ALC_T_SYS_WRITE_BUF,
+                         sizeof(SysIOVec)*iovlen);
+    if (!iov) {
+        res = -1;
+        err = ENOMEM;
+        driver_pdl_unlock(dd->blocking->pdl);
+    } else {
+        memcpy(iov,iov0,iovlen*sizeof(SysIOVec));
+        driver_pdl_unlock(dd->blocking->pdl);
+
+        do {
+            res = writev(dd->ofd->fd, iov, iovlen);
+        } while (res < 0 && errno == EINTR);
+        if (res < 0)
+            err = errno;
+        err = errno;
+
+        erts_free(ERTS_ALC_T_SYS_WRITE_BUF, iov);
+    }
+    dd->blocking->res = res;
+    dd->blocking->err = err;
+}
+
+void fd_ready_async(ErlDrvData drv_data,
+                    ErlDrvThreadData thread_data) {
+    ErtsSysDriverData *dd = (ErtsSysDriverData *)thread_data;
+    ErlDrvPort port_num = dd->port_num;
+
+    ASSERT(dd->blocking);
+
+    if (dd->blocking->res > 0) {
+        driver_pdl_lock(dd->blocking->pdl);
+        if (driver_deq(port_num, dd->blocking->res) == 0) {
+            driver_pdl_unlock(dd->blocking->pdl);
+            set_busy_port(port_num, 0);
+            if (dd->terminating) {
+                /* The port is has been ordered to terminate
+                   from either fd_flush or port_inp_failure */
+                if (dd->terminating == 1)
+                    driver_failure_atom(port_num, "normal");
+                else if (dd->terminating == 2)
+                    driver_failure_eof(port_num);
+                else if (dd->terminating < 0)
+                    driver_failure_posix(port_num, -dd->terminating);
+                return; /* -1; */
+            }
+        } else {
+            driver_pdl_unlock(dd->blocking->pdl);
+            /* still data left to write in queue */
+            driver_async(port_num, &dd->blocking->pkey, fd_async, dd, NULL);
+            return /* 0; */;
+        }
+    } else if (dd->blocking->res < 0) {
+        if (dd->blocking->err == ERRNO_BLOCK) {
+            set_busy_port(port_num, 1);
+            /* still data left to write in queue */
+            driver_async(port_num, &dd->blocking->pkey, fd_async, dd, NULL);
+        } else
+            driver_failure_posix(port_num, dd->blocking->err);
+        return; /* -1; */
+    }
+    return; /* 0; */
+}
+
+#endif
+
+/* Forker driver */
+
+static int forker_fd;
+
+static ErlDrvData forker_start(ErlDrvPort port_num, char* name,
+                               SysDriverOpts* opts)
+{
+
+    int i;
+    int fds[2];
+    int res, unbind;
+    char bindir[MAXPATHLEN];
+    size_t bindirsz = sizeof(bindir);
+    Uint csp_path_sz;
+    char *child_setup_prog;
+
+    forker_port = erts_drvport2id(port_num);
+
+    res = erts_sys_getenv_raw("BINDIR", bindir, &bindirsz);
+    if (res != 0) {
+        if (res < 0)
+            erl_exit(-1,
+                     "Environment variable BINDIR is not set\n");
+        if (res > 0)
+            erl_exit(-1,
+                     "Value of environment variable BINDIR is too large\n");
+    }
+    if (bindir[0] != DIR_SEPARATOR_CHAR)
+        erl_exit(-1,
+                 "Environment variable BINDIR does not contain an"
+                 " absolute path\n");
+    csp_path_sz = (strlen(bindir)
+                   + 1 /* DIR_SEPARATOR_CHAR */
+                   + sizeof(CHILD_SETUP_PROG_NAME)
+                   + 1);
+    child_setup_prog = erts_alloc(ERTS_ALC_T_CS_PROG_PATH, csp_path_sz);
+    erts_snprintf(child_setup_prog, csp_path_sz,
+                  "%s%c%s",
+                  bindir,
+                  DIR_SEPARATOR_CHAR,
+                  CHILD_SETUP_PROG_NAME);
+    if (socketpair(AF_UNIX, SOCK_STREAM, 0, fds) < 0) {
+        erl_exit(ERTS_ABORT_EXIT,
+                 "Could not open unix domain socket in spawn_init: %d\n",
+                 errno);
+    }
+
+    forker_fd = fds[0];
+
+    unbind = erts_sched_bind_atfork_prepare();
+
+    i = fork();
+
+    if (i == 0) {
+        /* The child */
+        char *cs_argv[FORKER_ARGV_NO_OF_ARGS] =
+            {CHILD_SETUP_PROG_NAME, NULL, NULL};
+        char buff[128];
+
+        erts_sched_bind_atfork_child(unbind);
+
+        snprintf(buff, 128, "%d", sys_max_files());
+        cs_argv[FORKER_ARGV_MAX_FILES] = buff;
+
+        /* We preallocate fd 3 for the uds fd */
+        if (fds[1] != 3) {
+            dup2(fds[1], 3);
+        }
+
+#if defined(USE_SETPGRP_NOARGS)		/* SysV */
+    (void) setpgrp();
+#elif defined(USE_SETPGRP)		/* BSD */
+    (void) setpgrp(0, getpid());
+#else					/* POSIX */
+    (void) setsid();
+#endif
+
+        execv(child_setup_prog, cs_argv);
+        _exit(1);
+    }
+
+    erts_sched_bind_atfork_parent(unbind);
+
+    erts_free(ERTS_ALC_T_CS_PROG_PATH, child_setup_prog);
+
+    close(fds[1]);
+
+    SET_NONBLOCKING(forker_fd);
+
+    driver_select(port_num, forker_fd, ERL_DRV_READ|ERL_DRV_USE, 1);
+
+    return (ErlDrvData)port_num;
+}
+
+static void forker_stop(ErlDrvData e)
+{
+    /* we probably should do something here,
+       the port has been closed by the user. */
+}
+
+static void forker_ready_input(ErlDrvData e, ErlDrvEvent fd)
+{
+    int res;
+    ErtsSysForkerProto *proto;
+
+    proto = erts_alloc(ERTS_ALC_T_DRV_CTRL_DATA, sizeof(*proto));
+
+    if ((res = read(fd, proto, sizeof(*proto))) < 0) {
+        if (errno == ERRNO_BLOCK)
+            return;
+        erl_exit(ERTS_DUMP_EXIT, "Failed to read from erl_child_setup: %d\n", errno);
+    }
+
+    if (res == 0)
+        erl_exit(ERTS_DUMP_EXIT, "erl_child_setup closed\n");
+
+    ASSERT(res == sizeof(*proto));
+
+#ifdef FORKER_PROTO_START_ACK
+    if (proto->action == ErtsSysForkerProtoAction_StartAck) {
+        /* Ideally we would like to not have to ack each Start
+           command being sent over the uds, but it would seem
+           that some operating systems (only observed on FreeBSD)
+           throw away data on the uds when the socket becomes full,
+           so we have to.
+        */
+        ErlDrvPort port_num = (ErlDrvPort)e;
+        int vlen;
+        SysIOVec *iov = driver_peekq(port_num, &vlen);
+        ErtsSysForkerProto *proto = (ErtsSysForkerProto *)iov[0].iov_base;
+
+        close(proto->u.start.fds[0]);
+        close(proto->u.start.fds[1]);
+        if (proto->u.start.fds[1] != proto->u.start.fds[2])
+            close(proto->u.start.fds[2]);
+
+        driver_deq(port_num, sizeof(*proto));
+
+        if (driver_sizeq(port_num) > 0)
+            driver_select(port_num, forker_fd, ERL_DRV_WRITE|ERL_DRV_USE, 1);
+    } else
+#endif
+    {
+        ASSERT(proto->action == ErtsSysForkerProtoAction_SigChld);
+
+        /* ideally this would be a port_command call, but as command is
+           already used by the spawn_driver, we use control instead.
+           Note that when using erl_drv_port_control it is an asynchronous
+           control. */
+        erl_drv_port_control(proto->u.sigchld.port_id, 'S',
+                             (char*)proto, sizeof(*proto));
+    }
+
+}
+
+static void forker_ready_output(ErlDrvData e, ErlDrvEvent fd)
+{
+    ErlDrvPort port_num = (ErlDrvPort)e;
+
+#ifndef FORKER_PROTO_START_ACK
+    while (driver_sizeq(port_num) > 0) {
+#endif
+        int vlen;
+        SysIOVec *iov = driver_peekq(port_num, &vlen);
+        ErtsSysForkerProto *proto = (ErtsSysForkerProto *)iov[0].iov_base;
+        ASSERT(iov[0].iov_len >= (sizeof(*proto)));
+        if (sys_uds_write(forker_fd, (char*)proto, sizeof(*proto),
+                          proto->u.start.fds, 3, 0) < 0) {
+            if (errno == ERRNO_BLOCK)
+                return;
+            erl_exit(ERTS_DUMP_EXIT, "Failed to write to erl_child_setup: %d\n", errno);
+        }
+#ifndef FORKER_PROTO_START_ACK
+        close(proto->u.start.fds[0]);
+        close(proto->u.start.fds[1]);
+        if (proto->u.start.fds[1] != proto->u.start.fds[2])
+            close(proto->u.start.fds[2]);
+        driver_deq(port_num, sizeof(*proto));
+    }
+#endif
+
+    driver_select(port_num, forker_fd, ERL_DRV_WRITE, 0);
+}
+
+static ErlDrvSSizeT forker_control(ErlDrvData e, unsigned int cmd, char *buf,
+                                   ErlDrvSizeT len, char **rbuf, ErlDrvSizeT rlen)
+{
+    ErtsSysForkerProto *proto = (ErtsSysForkerProto *)buf;
+    ErlDrvPort port_num = (ErlDrvPort)e;
+    int res;
+
+    driver_enq(port_num, buf, len);
+    if (driver_sizeq(port_num) > sizeof(*proto)) {
+        return 0;
+    }
+
+    if ((res = sys_uds_write(forker_fd, (char*)proto, sizeof(*proto),
+                             proto->u.start.fds, 3, 0)) < 0) {
+        if (errno == ERRNO_BLOCK) {
+            driver_select(port_num, forker_fd, ERL_DRV_WRITE|ERL_DRV_USE, 1);
+            return 0;
+        }
+        erl_exit(ERTS_DUMP_EXIT, "Failed to write to erl_child_setup: %d\n", errno);
+    }
+
+#ifndef FORKER_PROTO_START_ACK
+    ASSERT(res == sizeof(*proto));
+    close(proto->u.start.fds[0]);
+    close(proto->u.start.fds[1]);
+    if (proto->u.start.fds[1] != proto->u.start.fds[2])
+        close(proto->u.start.fds[2]);
+    driver_deq(port_num, sizeof(*proto));
+#endif
+
+    return 0;
+}
diff --git a/erts/emulator/sys/unix/sys_uds.c b/erts/emulator/sys/unix/sys_uds.c
new file mode 100644
index 0000000000..015d0346a1
--- /dev/null
+++ b/erts/emulator/sys/unix/sys_uds.c
@@ -0,0 +1,155 @@
+/*
+ * %CopyrightBegin%
+ * 
+ * Copyright Ericsson AB 2002-2009. All Rights Reserved.
+ * 
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ * 
+ * %CopyrightEnd%
+ */
+
+#include "sys_uds.h"
+
+int
+sys_uds_readv(int fd, struct iovec *iov, size_t iov_len,
+             int *fds, int fd_count, int flags) {
+    struct msghdr msg;
+    struct cmsghdr *cmsg = NULL;
+    char ancillary_buff[256] = {0};
+    int res, i = 0;
+
+    /* setup a place to fill in message contents */
+    memset(&msg, 0, sizeof(struct msghdr));
+    msg.msg_iov = iov;
+    msg.msg_iovlen = iov_len;
+
+    /* provide space for the ancillary data */
+    msg.msg_control = ancillary_buff;
+    msg.msg_controllen = sizeof(ancillary_buff);
+
+    if((res = recvmsg(fd, &msg, flags)) < 0) {
+#if defined(__APPLE__) && defined(__MACH__) && !defined(__DARWIN__)
+        /* When some OS X versions run out of fd's
+           they give EMSGSIZE instead of EMFILE.
+           We remap this as we want the correct
+           error to appear for the user */
+        if (errno == EMSGSIZE)
+            errno = EMFILE;
+#endif
+        return res;
+    }
+
+    if((msg.msg_flags & MSG_CTRUNC) == MSG_CTRUNC)
+    {
+        /* We assume that we have given enough space for any header
+           that are sent to us. So the only remaining reason to get
+           this flag set is if the caller has run out of file descriptors.
+        */
+        errno = EMFILE;
+        return -1;
+    }
+
+    for (cmsg = CMSG_FIRSTHDR(&msg); cmsg; cmsg = CMSG_NXTHDR(&msg, cmsg) ) {
+        if ((cmsg->cmsg_level == SOL_SOCKET) &&
+            (cmsg->cmsg_type == SCM_RIGHTS)) {
+                int *cmsg_data = (int *)CMSG_DATA(cmsg);
+                while ((char*)cmsg_data < (char*)cmsg + cmsg->cmsg_len) {
+                    if (i < fd_count) {
+                        fds[i++] = *cmsg_data++;
+                    } else {
+                        /* for some strange reason, we have received more FD's
+                           than we wanted... close them if we are not running
+                           debug. */
+                        if(i >= fd_count) abort();
+                        close(*cmsg_data++);
+                    }
+                }
+            }
+    }
+
+    return res;
+}
+
+int
+sys_uds_read(int fd, char *buff, size_t len,
+             int *fds, int fd_count, int flags) {
+    struct iovec iov;
+    iov.iov_base = buff;
+    iov.iov_len = len;
+    return sys_uds_readv(fd, &iov, 1, fds, fd_count, flags);
+}
+
+
+int
+sys_uds_writev(int fd, struct iovec *iov, size_t iov_len,
+               int *fds, int fd_count, int flags) {
+
+    struct msghdr msg;
+    struct cmsghdr *cmsg = NULL;
+    int res, i;
+
+    /* initialize socket message */
+    memset(&msg, 0, sizeof(struct msghdr));
+
+    /* We flatten the iov if it is too long */
+    if (iov_len > MAXIOV) {
+        int size = 0;
+        char *buff;
+        for (i = 0; i < iov_len; i++)
+            size += iov[i].iov_len;
+        buff = malloc(size);
+
+        for (i = 0; i < iov_len; i++) {
+            memcpy(buff, iov[i].iov_base, iov[i].iov_len);
+            buff += iov[i].iov_len;
+        }
+
+        iov[0].iov_base = buff - size;
+        iov[0].iov_len = size;
+        msg.msg_iov = iov;
+        msg.msg_iovlen = 1;
+    } else {
+        msg.msg_iov = iov;
+        msg.msg_iovlen = iov_len;
+    }
+
+    /* initialize the ancillary data */
+    msg.msg_control = calloc(1, CMSG_SPACE(sizeof(int) * fd_count));
+    msg.msg_controllen = CMSG_SPACE(sizeof(int) * fd_count);
+
+    /* copy the fd array into the ancillary data */
+    cmsg = CMSG_FIRSTHDR(&msg);
+    if(!cmsg) abort();
+    cmsg->cmsg_level = SOL_SOCKET;
+    cmsg->cmsg_type = SCM_RIGHTS;
+    cmsg->cmsg_len = CMSG_LEN(sizeof(int) * fd_count);
+    memcpy(CMSG_DATA(cmsg), fds, sizeof(int) * fd_count);
+
+    res = sendmsg(fd, &msg, flags);
+
+    if (iov_len > MAXIOV)
+        free(iov[0].iov_base);
+
+    free(msg.msg_control);
+
+    return res;
+}
+
+int
+sys_uds_write(int fd, char *buff, size_t len,
+              int *fds, int fd_count, int flags) {
+    struct iovec iov;
+    iov.iov_base = buff;
+    iov.iov_len = len;
+    return sys_uds_writev(fd, &iov, 1, fds, fd_count, flags);
+}
diff --git a/erts/emulator/sys/unix/sys_uds.h b/erts/emulator/sys/unix/sys_uds.h
new file mode 100644
index 0000000000..844a2804d8
--- /dev/null
+++ b/erts/emulator/sys/unix/sys_uds.h
@@ -0,0 +1,57 @@
+/*
+ * %CopyrightBegin%
+ *
+ * Copyright Ericsson AB 2002-2009. All Rights Reserved.
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ *
+ * %CopyrightEnd%
+ */
+
+#ifndef _ERL_UNIX_UDS_H
+#define _ERL_UNIX_UDS_H
+
+#ifdef HAVE_CONFIG_H
+#  include "config.h"
+#endif
+
+#if defined(__sun__) && !defined(_XOPEN_SOURCE)
+#define _XOPEN_SOURCE 500
+#endif
+
+#include <limits.h>
+
+#include <sys/types.h>
+#include <sys/socket.h>
+#include <sys/uio.h>
+
+#if defined IOV_MAX
+#define MAXIOV IOV_MAX
+#elif defined UIO_MAXIOV
+#define MAXIOV UIO_MAXIOV
+#else
+#define MAXIOV 16
+#endif
+
+#include "sys.h"
+
+int sys_uds_readv(int fd, struct iovec *iov, size_t iov_len,
+                  int *fds, int fd_count, int flags);
+int sys_uds_read(int fd, char *buff, size_t len,
+                 int *fds, int fd_count, int flags);
+int sys_uds_writev(int fd, struct iovec *iov, size_t iov_len,
+                   int *fds, int fd_count, int flags);
+int sys_uds_write(int fd, char *buff, size_t len,
+                  int *fds, int fd_count, int flags);
+
+#endif /* #ifndef _ERL_UNIX_UDS_H */
diff --git a/erts/emulator/sys/win32/erl_win32_sys_ddll.c b/erts/emulator/sys/win32/erl_win32_sys_ddll.c
index 9a5557e93d..7c24a77e31 100644
--- a/erts/emulator/sys/win32/erl_win32_sys_ddll.c
+++ b/erts/emulator/sys/win32/erl_win32_sys_ddll.c
@@ -52,7 +52,8 @@ void erl_sys_ddll_init(void) {
 #define ERL_NIF_API_FUNC_DECL(RET,NAME,ARGS) nif_callbacks.NAME = NAME
 #include "erl_nif_api_funcs.h"
 #undef ERL_NIF_API_FUNC_DECL
-   
+    nif_callbacks.erts_alc_test = erts_alc_test;
+ 
     return;
 }
 
diff --git a/erts/emulator/sys/win32/erl_win_dyn_driver.h b/erts/emulator/sys/win32/erl_win_dyn_driver.h
index 5e62320be4..baac7c903e 100644
--- a/erts/emulator/sys/win32/erl_win_dyn_driver.h
+++ b/erts/emulator/sys/win32/erl_win_dyn_driver.h
@@ -145,8 +145,8 @@ WDD_TYPEDEF(ErlDrvTid, erl_drv_thread_self, (void));
 WDD_TYPEDEF(int, erl_drv_equal_tids, (ErlDrvTid tid1, ErlDrvTid tid2));
 WDD_TYPEDEF(void, erl_drv_thread_exit, (void *resp));
 WDD_TYPEDEF(int, erl_drv_thread_join, (ErlDrvTid, void **respp));
-WDD_TYPEDEF(int, erl_drv_putenv, (char *key, char *value));
-WDD_TYPEDEF(int, erl_drv_getenv, (char *key, char *value, size_t *value_size));
+WDD_TYPEDEF(int, erl_drv_putenv, (const char *key, char *value));
+WDD_TYPEDEF(int, erl_drv_getenv, (const char *key, char *value, size_t *value_size));
 
 typedef struct {
     WDD_FTYPE(null_func) *null_func;
diff --git a/erts/emulator/sys/win32/sys.c b/erts/emulator/sys/win32/sys.c
index fce76db28f..76ce25916a 100644
--- a/erts/emulator/sys/win32/sys.c
+++ b/erts/emulator/sys/win32/sys.c
@@ -1334,10 +1334,8 @@ spawn_start(ErlDrvPort port_num, char* utf8_name, SysDriverOpts* opts)
 	retval = set_driver_data(dp, hFromChild, hToChild, opts->read_write,
 				 opts->exit_status);
 	if (retval != ERL_DRV_ERROR_GENERAL && retval != ERL_DRV_ERROR_ERRNO) {
-	    Port *prt = erts_drvport2port(port_num);
-		/* We assume that this cannot generate a negative number */
-	    ASSERT(prt != ERTS_INVALID_ERL_DRV_PORT);
-	    prt->os_pid = (SWord) pid;
+            /* We assume that this cannot generate a negative number */
+            erl_drv_set_os_pid(port_num, pid);
 	}
     }
     
@@ -1528,8 +1526,8 @@ create_child_process
 	 * Parse out the program name from the command line (it can be quoted and
 	 * contain spaces).
 	 */
-	newcmdline = (wchar_t *) erts_alloc(ERTS_ALC_T_TMP, 2048*sizeof(wchar_t));
 	cmdlength = parse_command(origcmd);
+	newcmdline = (wchar_t *) erts_alloc(ERTS_ALC_T_TMP, (MAX_PATH+wcslen(origcmd)-cmdlength)*sizeof(wchar_t));
 	thecommand = (wchar_t *) erts_alloc(ERTS_ALC_T_TMP, (cmdlength+1)*sizeof(wchar_t));
 	wcsncpy(thecommand, origcmd, cmdlength);
 	thecommand[cmdlength] = L'\0';
@@ -3273,6 +3271,12 @@ void erl_sys_init(void)
 }
 
 void
+erl_sys_late_init(void)
+{
+    /* do nothing */
+}
+
+void
 erts_sys_schedule_interrupt(int set)
 {
     erts_check_io_interrupt(set);
diff --git a/erts/emulator/test/Makefile b/erts/emulator/test/Makefile
index 77614d455c..8cc47937b7 100644
--- a/erts/emulator/test/Makefile
+++ b/erts/emulator/test/Makefile
@@ -79,6 +79,7 @@ MODULES= \
 	node_container_SUITE \
 	nofrag_SUITE \
 	num_bif_SUITE \
+	message_queue_data_SUITE \
 	op_SUITE \
 	port_SUITE \
 	port_bif_SUITE \
diff --git a/erts/emulator/test/alloc_SUITE.erl b/erts/emulator/test/alloc_SUITE.erl
index 7c7ddde5d4..516bc873a5 100644
--- a/erts/emulator/test/alloc_SUITE.erl
+++ b/erts/emulator/test/alloc_SUITE.erl
@@ -31,7 +31,8 @@
 	 rbtree/1,
 	 mseg_clear_cache/1,
 	 erts_mmap/1,
-	 cpool/1]).
+	 cpool/1,
+	 migration/1]).
 
 -export([init_per_testcase/2, end_per_testcase/2]).
 
@@ -43,7 +44,7 @@ suite() -> [{ct_hooks,[ts_install_cth]}].
 
 all() -> 
     [basic, coalesce, threads, realloc_copy, bucket_index,
-     bucket_mask, rbtree, mseg_clear_cache, erts_mmap, cpool].
+     bucket_mask, rbtree, mseg_clear_cache, erts_mmap, cpool, migration].
 
 groups() -> 
     [].
@@ -64,7 +65,7 @@ end_per_group(_GroupName, Config) ->
 
 init_per_testcase(Case, Config) when is_list(Config) ->
     Dog = ?t:timetrap(?t:seconds(?DEFAULT_TIMETRAP_SECS)),
-    [{watchdog, Dog},{testcase, Case}|Config].
+    [{watchdog, Dog}, {testcase, Case}, {debug,false} | Config].
 
 end_per_testcase(_Case, Config) when is_list(Config) ->
     Dog = ?config(watchdog, Config),
@@ -112,14 +113,19 @@ cpool(suite) -> [];
 cpool(doc) ->   [];
 cpool(Cfg) -> ?line drv_case(Cfg).
 
+migration(Cfg) ->
+    case erlang:system_info(smp_support) of
+	true ->
+	    drv_case(Cfg, concurrent, "+MZe true");
+	false ->
+	    {skipped, "No smp"}
+    end.
+
 erts_mmap(Config) when is_list(Config) ->
-    case {?t:os_type(), is_halfword_vm()} of
-	{{unix, _}, false} ->
+    case ?t:os_type() of
+	{unix, _} ->
 	    [erts_mmap_do(Config, SCO, SCRPM, SCRFSD)
 	     || SCO <-[true,false], SCRFSD <-[1234,0], SCRPM <- [true,false]];
-
-	{_,true} ->
-	    {skipped, "No supercarrier support on halfword vm"};
 	{SkipOs,_} ->
 	    ?line {skipped,
 		   lists:flatten(["Not run on "
@@ -176,18 +182,17 @@ erts_mmap_do(Config, SCO, SCRPM, SCRFSD) ->
 %%                                                                        %%
 
 drv_case(Config) ->
-    drv_case(Config, "").
+    drv_case(Config, one_shot, "").
 
-drv_case(Config, Command) when is_list(Config),
-			       is_list(Command) ->
+drv_case(Config, Mode, NodeOpts) when is_list(Config) ->
     case ?t:os_type() of
 	{Family, _} when Family == unix; Family == win32 ->
-	    ?line {ok, Node} = start_node(Config),
+	    ?line {ok, Node} = start_node(Config, NodeOpts),
 	    ?line Self = self(),
 	    ?line Ref = make_ref(),
 	    ?line spawn_link(Node,
 			     fun () ->
-				     Res = run_drv_case(Config, Command),
+				     Res = run_drv_case(Config, Mode),
 				     Self ! {Ref, Res}
 			     end),
 	    ?line Result = receive {Ref, Rslt} -> Rslt end,
@@ -199,49 +204,172 @@ drv_case(Config, Command) when is_list(Config),
 				  | io_lib:format("~p",[SkipOs])])}
     end.
 
-run_drv_case(Config, Command) ->
-    ?line DataDir = ?config(data_dir,Config),
-    ?line CaseName = ?config(testcase,Config),
-    case erl_ddll:load_driver(DataDir, CaseName) of
-	ok -> ok;
-	{error, Error} ->
-	    io:format("~s\n", [erl_ddll:format_error(Error)]),
-	    ?line ?t:fail()
+run_drv_case(Config, Mode) ->
+    DataDir = ?config(data_dir,Config),
+    CaseName = ?config(testcase,Config),
+    File = filename:join(DataDir, CaseName),
+    {ok,CaseName,Bin} = compile:file(File, [binary,return_errors]),
+    {module,CaseName} = erlang:load_module(CaseName,Bin),
+    print_stats(CaseName),
+    ok = CaseName:init(File),
+
+    SlaveState = slave_init(CaseName),
+    case Mode of
+	one_shot ->
+	    Result = one_shot(CaseName);
+
+	concurrent ->
+	    Result = concurrent(CaseName)
     end,
-    ?line Port = open_port({spawn, atom_to_list(CaseName)}, []),
-    ?line true = is_port(Port),
-    ?line Port ! {self(), {command, Command}},
-    ?line Result = receive_drv_result(Port, CaseName),
-    ?line Port ! {self(), close},
-    ?line receive 
-	      {Port, closed} ->
-		  ok
-	  end,
-    ?line ok = erl_ddll:unload_driver(CaseName),
-    ?line Result.
-
-receive_drv_result(Port, CaseName) ->
-    ?line receive
-	      {print, Port, CaseName, Str} ->
-		  ?line ?t:format("~s", [Str]),
-		  ?line receive_drv_result(Port, CaseName);
-	      {'EXIT', Port, Error} ->
-		  ?line ?t:fail(Error);
-	      {'EXIT', error, Error} ->
-		  ?line ?t:fail(Error);
-	      {failed, Port, CaseName, Comment} ->
-		  ?line ?t:fail(Comment);
-	      {skipped, Port, CaseName, Comment} ->
-		  ?line {skipped, Comment};
-	      {succeeded, Port, CaseName, ""} ->
-		  ?line succeeded;
-	      {succeeded, Port, CaseName, Comment} ->
-		  ?line {comment, Comment}
-	  end.
-
-start_node(Config) ->
-    start_node(Config, []).
+
+    wait_for_memory_deallocations(),
+    print_stats(CaseName),
+
+    true = erlang:delete_module(CaseName),
+    slave_end(SlaveState),
+    Result.
+
+slave_init(migration) ->
+    A0 = case application:start(sasl) of
+	     ok -> [sasl];
+	     _ -> []
+	 end,
+    case application:start(os_mon) of
+	ok -> [os_mon|A0];
+	_ -> A0
+    end;
+slave_init(_) -> [].
+
+slave_end(Apps) ->
+    lists:foreach(fun (A) -> application:stop(A) end, Apps).
+
+wait_for_memory_deallocations() ->
+    try
+        erts_debug:set_internal_state(wait, deallocations)
+    catch
+        error:undef ->
+            erts_debug:set_internal_state(available_internal_state, true),
+            wait_for_memory_deallocations()
+    end.
+
+print_stats(migration) ->
+    {Btot,Ctot} = lists:foldl(fun({instance,Inr,Istats}, {Bacc,Cacc}) ->
+				      {mbcs,MBCS} = lists:keyfind(mbcs, 1, Istats),
+				      Btup = lists:keyfind(blocks, 1, MBCS),
+				      Ctup = lists:keyfind(carriers, 1, MBCS),
+				      io:format("{instance,~p,~p,~p}\n", [Inr, Btup, Ctup]),
+				      {tuple_add(Bacc,Btup),tuple_add(Cacc,Ctup)};
+				 (_, Acc) -> Acc
+			      end,
+			      {{blocks,0,0,0},{carriers,0,0,0}},
+			      erlang:system_info({allocator,test_alloc})),
+
+    io:format("Number of blocks  : ~p\n", [Btot]),
+    io:format("Number of carriers: ~p\n", [Ctot]);
+
+print_stats(_) -> ok.
+
+tuple_add(T1, T2) ->
+    list_to_tuple(lists:zipwith(fun(E1,E2) when is_number(E1), is_number(E2) ->
+					 E1 + E2;
+				    (A,A) ->
+					 A
+				 end,
+				 tuple_to_list(T1), tuple_to_list(T2))).
+
+
+one_shot(CaseName) ->
+    State = CaseName:start({1, 0, erlang:system_info(build_type)}),
+    Result0 = CaseName:run(State),
+    false = (Result0 =:= continue),
+    Result1 = handle_result(State, Result0),
+    CaseName:stop(State),
+    Result1.
+
+
+many_shot(CaseName, I, Mem) ->
+    State = CaseName:start({I, Mem, erlang:system_info(build_type)}),
+    Result1 = repeat_while(fun() ->
+				   Result0 = CaseName:run(State),
+				   handle_result(State, Result0)
+			   end,
+			   10*1000, I),
+    CaseName:stop(State),
+    flush_log(),
+    Result1.
+
+concurrent(CaseName) ->
+    NSched = erlang:system_info(schedulers),
+    Mem = (free_memory() * 3) div 4,
+    PRs = lists:map(fun(I) -> spawn_opt(fun() ->
+						many_shot(CaseName, I,
+							  Mem div NSched)
+					end,
+				       [monitor, {scheduler,I}])
+		    end,
+		    lists:seq(1, NSched)),
+    lists:foreach(fun({Pid,Ref}) ->
+			  receive {'DOWN', Ref, process, Pid, Reason} ->
+				  Reason
+			  end
+		  end,
+		  PRs),
+    ok.
+
+repeat_while(Fun, Timeout, I) ->
+    TRef = erlang:start_timer(Timeout, self(), timeout),
+    R = repeat_while_loop(Fun, TRef, I),
+    erlang:cancel_timer(TRef, [{async,true},{info,false}]),
+    R.
+
+repeat_while_loop(Fun, TRef, I) ->
+    receive
+	{timeout, TRef, timeout} ->
+	    io:format("~p: Timeout, enough is enough.",[I]),
+	    succeeded
+    after 0 ->
+	    %%io:format("~p calls fun\n", [self()]),
+	    case Fun() of
+		continue -> repeat_while_loop(Fun, TRef, I);
+		R -> R
+	    end
+    end.
+
+flush_log() ->
+    receive
+	{print, Str} ->
+	    ?t:format("~s", [Str]),
+	    flush_log()
+    after 0 ->
+	    ok
+    end.
+
+handle_result(_State, Result0) ->
+    flush_log(),
+    case Result0 of
+	{'EXIT', Error} ->
+	    ?line ?t:fail(Error);
+	{'EXIT', error, Error} ->
+	    ?line ?t:fail(Error);
+	{failed, Comment} ->
+	    ?line ?t:fail(Comment);
+	{skipped, Comment} ->
+	    ?line {skipped, Comment};
+	{succeeded, ""} ->
+	    ?line succeeded;
+	{succeeded, Comment} ->
+	    ?line {comment, Comment};
+	continue ->
+	    continue
+    end.
+
 start_node(Config, Opts) when is_list(Config), is_list(Opts) ->
+    case ?config(debug,Config) of
+	true -> {ok, node()};
+	_ -> start_node_1(Config, Opts)
+    end.
+
+start_node_1(Config, Opts) ->
     Pa = filename:dirname(code:which(?MODULE)),
     Name = list_to_atom(atom_to_list(?MODULE)
 			++ "-"
@@ -252,12 +380,27 @@ start_node(Config, Opts) when is_list(Config), is_list(Opts) ->
 			++ integer_to_list(erlang:unique_integer([positive]))),
     ?t:start_node(Name, slave, [{args, Opts++" -pa "++Pa}]).
 
+stop_node(Node) when Node =:= node() -> ok;
 stop_node(Node) ->
     ?t:stop_node(Node).
 
-is_halfword_vm() ->
-    case {erlang:system_info({wordsize, internal}),
-	  erlang:system_info({wordsize, external})} of
-	{4, 8} -> true;
-	{WS, WS} -> false
+free_memory() ->
+    %% Free memory in MB.
+    try
+	SMD = memsup:get_system_memory_data(),
+	{value, {free_memory, Free}} = lists:keysearch(free_memory, 1, SMD),
+	TotFree = (Free +
+		   case lists:keysearch(cached_memory, 1, SMD) of
+		       {value, {cached_memory, Cached}} -> Cached;
+		       false -> 0
+		   end +
+		   case lists:keysearch(buffered_memory, 1, SMD) of
+		       {value, {buffered_memory, Buffed}} -> Buffed;
+		       false -> 0
+		   end),
+	TotFree div (1024*1024)
+    catch
+	error : undef ->
+	    ?t:fail({"os_mon not built"})
     end.
+
diff --git a/erts/emulator/test/alloc_SUITE_data/Makefile.src b/erts/emulator/test/alloc_SUITE_data/Makefile.src
index a441fe946b..e31de54e1b 100644
--- a/erts/emulator/test/alloc_SUITE_data/Makefile.src
+++ b/erts/emulator/test/alloc_SUITE_data/Makefile.src
@@ -25,7 +25,8 @@ TEST_DRVS =	basic@dll@		\
 		bucket_mask@dll@	\
 		rbtree@dll@		\
 		mseg_clear_cache@dll@	\
-		cpool@dll@
+		cpool@dll@		\
+		migration@dll@
 
 CC = @CC@
 LD = @LD@
diff --git a/erts/emulator/test/alloc_SUITE_data/allocator_test.h b/erts/emulator/test/alloc_SUITE_data/allocator_test.h
index 1d6b2f4907..97ee58cdad 100644
--- a/erts/emulator/test/alloc_SUITE_data/allocator_test.h
+++ b/erts/emulator/test/alloc_SUITE_data/allocator_test.h
@@ -20,9 +20,20 @@
 #ifndef ALLOCATOR_TEST_H__
 #define ALLOCATOR_TEST_H__
 
-typedef ErlDrvUInt Ulong;
+#if SIZEOF_VOID_P == SIZEOF_INT
+typedef unsigned int Ulong;
+#elif SIZEOF_VOID_P == SIZEOF_LONG
+typedef unsigned long Ulong;
+#elif SIZEOF_VOID_P == SIZEOF_LONG_LONG
+typedef unsigned long long Ulong;
+#else
+# error No pointer sized integer type found ???
+#endif
 
-#ifndef __WIN32__
+#ifdef __WIN32__
+typedef Ulong erts_alc_test_Fn(Ulong, Ulong, Ulong, Ulong);
+#  define erts_alc_test ((erts_alc_test_Fn*)WinDynNifCallbacks.erts_alc_test)
+#else
 Ulong erts_alc_test(Ulong, Ulong, Ulong, Ulong);
 #endif
 
@@ -85,6 +96,7 @@ typedef void* erts_cond;
 #define CPOOL_DELETE(A,B)	((Carrier_t *)	ALC_TEST2(0x022, (A), (B)))
 #define CPOOL_IS_EMPTY(A)	((int)		ALC_TEST1(0x023, (A)))
 #define CPOOL_IS_IN_POOL(A,B)	((int)		ALC_TEST2(0x024, (A), (B)))
+#define UMEM2BLK_TEST(P)	((Block_t*)	ALC_TEST1(0x025, (P)))
 
 /* From erl_goodfit_alloc.c */
 #define BKT_IX(A, S)		((Ulong)	ALC_TEST2(0x100, (A), (S)))
@@ -142,5 +154,9 @@ typedef void* erts_cond;
 #define THR_JOIN(T)		((void)		ALC_TEST1(0xf11, (T)))
 #define THR_EXIT(R)		((void)		ALC_TEST1(0xf12, (R)))
 #define IS_SMP_ENABLED		((int)		ALC_TEST0(0xf13))
+#define ALLOC_TEST(S)		((void*)	ALC_TEST1(0xf14, (S)))
+#define FREE_TEST(P)		((void)		ALC_TEST1(0xf15, (P)))
+#define SET_TEST_MBC_USER_HEADER(SZ,CMBC,DMBC) ((int)ALC_TEST3(0xf16, (SZ), (CMBC), (DMBC)))
+#define GET_TEST_MBC_SIZE()     ((int)          ALC_TEST0(0xf17))
 
 #endif
diff --git a/erts/emulator/test/alloc_SUITE_data/basic.c b/erts/emulator/test/alloc_SUITE_data/basic.c
index 323a24a11f..debb3d7ebe 100644
--- a/erts/emulator/test/alloc_SUITE_data/basic.c
+++ b/erts/emulator/test/alloc_SUITE_data/basic.c
@@ -60,3 +60,6 @@ testcase_cleanup(TestCaseState_t *tcs)
     if (tcs->extra)
 	STOP_ALC((Allctr_t *) tcs->extra);
 }
+
+ERL_NIF_INIT(basic, testcase_nif_funcs, testcase_nif_init,
+	     NULL, NULL, NULL);
diff --git a/erts/emulator/test/alloc_SUITE_data/basic.erl b/erts/emulator/test/alloc_SUITE_data/basic.erl
new file mode 100644
index 0000000000..a018fd5582
--- /dev/null
+++ b/erts/emulator/test/alloc_SUITE_data/basic.erl
@@ -0,0 +1,10 @@
+-module(basic).
+
+-export([init/1, start/1, run/1, stop/1]).
+
+init(File) ->
+    ok = erlang:load_nif(File, 0).
+
+start(_) -> erlang:nif_error(not_loaded).
+run(_)  -> erlang:nif_error(not_loaded).
+stop(_) -> erlang:nif_error(not_loaded).
diff --git a/erts/emulator/test/alloc_SUITE_data/bucket_index.c b/erts/emulator/test/alloc_SUITE_data/bucket_index.c
index c13f229049..45cb53fbf7 100644
--- a/erts/emulator/test/alloc_SUITE_data/bucket_index.c
+++ b/erts/emulator/test/alloc_SUITE_data/bucket_index.c
@@ -113,3 +113,5 @@ test_it(TestCaseState_t *tcs, unsigned sbct)
 		    sbct ? sbct_buf : "default");
 }
 
+ERL_NIF_INIT(bucket_index, testcase_nif_funcs, testcase_nif_init,
+	     NULL, NULL, NULL);
diff --git a/erts/emulator/test/alloc_SUITE_data/bucket_index.erl b/erts/emulator/test/alloc_SUITE_data/bucket_index.erl
new file mode 100644
index 0000000000..c54f54e2f5
--- /dev/null
+++ b/erts/emulator/test/alloc_SUITE_data/bucket_index.erl
@@ -0,0 +1,10 @@
+-module(bucket_index).
+
+-export([init/1, start/1, run/1, stop/1]).
+
+init(File) ->
+    ok = erlang:load_nif(File, 0).
+
+start(_) -> erlang:nif_error(not_loaded).
+run(_)  -> erlang:nif_error(not_loaded).
+stop(_) -> erlang:nif_error(not_loaded).
diff --git a/erts/emulator/test/alloc_SUITE_data/bucket_mask.c b/erts/emulator/test/alloc_SUITE_data/bucket_mask.c
index 8d6166771e..c94c265f4e 100644
--- a/erts/emulator/test/alloc_SUITE_data/bucket_mask.c
+++ b/erts/emulator/test/alloc_SUITE_data/bucket_mask.c
@@ -52,7 +52,7 @@ testcase_run(TestCaseState_t *tcs)
     typedef struct linked_block {
 	struct linked_block* next;
     }Linked;
-    Linked* link;
+    Linked* link = NULL;
     Linked* fence_list;
     Linked* pad_list;
     void* tmp;
@@ -183,3 +183,5 @@ testcase_run(TestCaseState_t *tcs)
     tcs->extra = NULL;
 }
 
+ERL_NIF_INIT(bucket_mask, testcase_nif_funcs, testcase_nif_init,
+	     NULL, NULL, NULL);
diff --git a/erts/emulator/test/alloc_SUITE_data/bucket_mask.erl b/erts/emulator/test/alloc_SUITE_data/bucket_mask.erl
new file mode 100644
index 0000000000..589a50e1fa
--- /dev/null
+++ b/erts/emulator/test/alloc_SUITE_data/bucket_mask.erl
@@ -0,0 +1,10 @@
+-module(bucket_mask).
+
+-export([init/1, start/1, run/1, stop/1]).
+
+init(File) ->
+    ok = erlang:load_nif(File, 0).
+
+start(_) -> erlang:nif_error(not_loaded).
+run(_)  -> erlang:nif_error(not_loaded).
+stop(_) -> erlang:nif_error(not_loaded).
diff --git a/erts/emulator/test/alloc_SUITE_data/coalesce.c b/erts/emulator/test/alloc_SUITE_data/coalesce.c
index 0a5e0c5b0e..7791409a34 100644
--- a/erts/emulator/test/alloc_SUITE_data/coalesce.c
+++ b/erts/emulator/test/alloc_SUITE_data/coalesce.c
@@ -317,3 +317,6 @@ testcase_cleanup(TestCaseState_t *tcs)
     if (tcs->extra)
 	STOP_ALC((Allctr_t *) tcs->extra);
 }
+
+ERL_NIF_INIT(coalesce, testcase_nif_funcs, testcase_nif_init,
+	     NULL, NULL, NULL);
diff --git a/erts/emulator/test/alloc_SUITE_data/coalesce.erl b/erts/emulator/test/alloc_SUITE_data/coalesce.erl
new file mode 100644
index 0000000000..453c726c4e
--- /dev/null
+++ b/erts/emulator/test/alloc_SUITE_data/coalesce.erl
@@ -0,0 +1,10 @@
+-module(coalesce).
+
+-export([init/1, start/1, run/1, stop/1]).
+
+init(File) ->
+    ok = erlang:load_nif(File, 0).
+
+start(_) -> erlang:nif_error(not_loaded).
+run(_)  -> erlang:nif_error(not_loaded).
+stop(_) -> erlang:nif_error(not_loaded).
diff --git a/erts/emulator/test/alloc_SUITE_data/cpool.c b/erts/emulator/test/alloc_SUITE_data/cpool.c
index 75c2bc13ae..73026cc758 100644
--- a/erts/emulator/test/alloc_SUITE_data/cpool.c
+++ b/erts/emulator/test/alloc_SUITE_data/cpool.c
@@ -86,13 +86,13 @@ thread_func(void *arg)
     for (i = 0; i < (TEST_NO_CARRIERS_PER_THREAD+TEST_CARRIERS_OFFSET); i++) {
 	int d;
 	if (i < TEST_NO_CARRIERS_PER_THREAD) {
-	    CPOOL_INSERT(alloc, crr[i]);
+	    (void) CPOOL_INSERT(alloc, crr[i]);
 	    if ((i & 0x7) == 0)
 		FATAL_ASSERT(CPOOL_IS_IN_POOL(alloc, crr[i]));
 	}
 	d = i-TEST_CARRIERS_OFFSET;
 	if (d >= 0) {
-	    CPOOL_DELETE(alloc, crr[d]);
+	    (void) CPOOL_DELETE(alloc, crr[d]);
 	    if ((d & 0x7) == 0)
 		FATAL_ASSERT(!CPOOL_IS_IN_POOL(alloc, crr[d]));
 	}
@@ -129,7 +129,7 @@ testcase_run(TestCaseState_t *tcs)
 	for (c = 0; c < TEST_NO_CARRIERS_PER_THREAD; c++) {
 	    Carrier_t *crr = (Carrier_t *) p;
 	    p += zcrr_sz;
-	    ZERO_CRR_INIT(alloc, crr);
+	    (void) ZERO_CRR_INIT(alloc, crr);
 	    threads[t].crr[c] = crr;
 	}
     }
@@ -156,3 +156,6 @@ testcase_run(TestCaseState_t *tcs)
 
     ASSERT(tcs, no_threads == TEST_NO_THREADS);
 }
+
+ERL_NIF_INIT(cpool, testcase_nif_funcs, testcase_nif_init,
+	     NULL, NULL, NULL);
diff --git a/erts/emulator/test/alloc_SUITE_data/cpool.erl b/erts/emulator/test/alloc_SUITE_data/cpool.erl
new file mode 100644
index 0000000000..89053471fa
--- /dev/null
+++ b/erts/emulator/test/alloc_SUITE_data/cpool.erl
@@ -0,0 +1,10 @@
+-module(cpool).
+
+-export([init/1, start/1, run/1, stop/1]).
+
+init(File) ->
+    ok = erlang:load_nif(File, 0).
+
+start(_) -> erlang:nif_error(not_loaded).
+run(_)  -> erlang:nif_error(not_loaded).
+stop(_) -> erlang:nif_error(not_loaded).
diff --git a/erts/emulator/test/alloc_SUITE_data/migration.c b/erts/emulator/test/alloc_SUITE_data/migration.c
new file mode 100644
index 0000000000..b006360043
--- /dev/null
+++ b/erts/emulator/test/alloc_SUITE_data/migration.c
@@ -0,0 +1,343 @@
+/*
+ * %CopyrightBegin%
+ *
+ * Copyright Ericsson AB 2014. All Rights Reserved.
+ *
+ * The contents of this file are subject to the Erlang Public License,
+ * Version 1.1, (the "License"); you may not use this file except in
+ * compliance with the License. You should have received a copy of the
+ * Erlang Public License along with this software. If not, it can be
+ * retrieved online at http://www.erlang.org/.
+ *
+ * Software distributed under the License is distributed on an "AS IS"
+ * basis, WITHOUT WARRANTY OF ANY KIND, either express or implied. See
+ * the License for the specific language governing rights and limitations
+ * under the License.
+ *
+ * %CopyrightEnd%
+ */
+
+/*
+ * Test the carrier migration logic
+ */
+
+#ifndef __WIN32__
+#include <sys/types.h>
+#include <unistd.h>
+#include <errno.h>
+#endif
+#include <stdio.h>
+#include <stdlib.h>
+#include <stdarg.h>
+#include <string.h>
+#include "testcase_driver.h"
+#include "allocator_test.h"
+
+#define FATAL_ASSERT(A)						\
+    ((void) ((A)						\
+	     ? 1						\
+	     : (fatal_assert_failed(#A,				\
+				    (char *) __FILE__,		\
+				    __LINE__),			\
+		0)))
+
+static void
+fatal_assert_failed(char* expr, char* file, int line)
+{
+    fflush(stdout);
+    fprintf(stderr, "%s:%d: Assertion failed: %s\n",
+	    file, line, expr);
+    fflush(stderr);
+    abort();
+}
+
+
+char *
+testcase_name(void)
+{
+    return "migration";
+}
+
+/* Turns out random_r() is a nonstandard glibc extension.
+#define HAVE_RANDOM_R
+*/
+#ifdef HAVE_RANDOM_R
+
+typedef struct { struct random_data rnd; char rndbuf[32]; } MyRandState;
+
+static void myrand_init(MyRandState* mrs, unsigned int seed)
+{
+    int res;
+    memset(&mrs->rnd, 0, sizeof(mrs->rnd));
+    res = initstate_r(seed, mrs->rndbuf, sizeof(mrs->rndbuf), &mrs->rnd);
+    FATAL_ASSERT(res == 0);
+}
+
+static int myrand(MyRandState* mrs)
+{
+    int32_t x;
+    int res = random_r(&mrs->rnd, &x);
+    FATAL_ASSERT(res == 0);
+    return (int)x;
+}
+
+#else /* !HAVE_RANDOM_R */
+
+typedef unsigned int MyRandState;
+
+static void myrand_init(MyRandState* mrs, unsigned int seed)
+{
+    *mrs = seed;
+}
+
+static int myrand(MyRandState* mrs)
+{
+    /* Taken from rand(3) man page.
+     * Modified to return a full 31-bit value by using low half of *mrs as well.
+     */
+    *mrs = (*mrs) * 1103515245 + 12345;
+    return (int) (((*mrs >> 16) | (*mrs << 16)) & ~(1 << 31));
+}
+
+#endif /* !HAVE_RANDOM_R */
+
+#define MAX_BLOCK_PER_THR 200
+#define BLOCKS_PER_MBC 10
+#define MAX_ROUNDS 10000
+
+typedef struct MyBlock_ {
+    struct MyBlock_* next;
+    struct MyBlock_** prevp;
+} MyBlock;
+
+typedef struct {
+    MyBlock* blockv[MAX_BLOCK_PER_THR];
+    MyRandState rand_state;
+    enum { GROWING, SHRINKING, CLEANUP, DONE } phase;
+    int nblocks;
+    int goal_nblocks;
+    int round;
+    int nr_of_migrations;
+    int nr_of_carriers;
+    int max_blocks_in_mbc;
+    int block_size;
+    int max_nblocks;
+} MigrationState;
+
+typedef struct {
+    ErlNifMutex* mtx;
+    int nblocks;
+    MyBlock* first;
+    MigrationState* employer;
+} MyCrrInfo;
+
+
+static int crr_info_offset = -1;
+static void (*orig_create_mbc_fn)(Allctr_t *allctr, Carrier_t *carrier);
+static void (*orig_destroying_mbc_fn)(Allctr_t *allctr, Carrier_t *carrier);
+
+static void my_creating_mbc(Allctr_t *allctr, Carrier_t *carrier)
+{
+    MyCrrInfo* mci = (MyCrrInfo*) ((char*)carrier + crr_info_offset);
+    if (orig_create_mbc_fn)
+	orig_create_mbc_fn(allctr, carrier);
+
+    mci->mtx = enif_mutex_create("alloc_SUITE.migration");
+    mci->nblocks = 0;
+    mci->first = NULL;
+    mci->employer = NULL;
+}
+
+static void my_destroying_mbc(Allctr_t *allctr, Carrier_t *carrier)
+{
+    MyCrrInfo* mci = (MyCrrInfo*) ((char*)carrier + crr_info_offset);
+
+    FATAL_ASSERT(mci->nblocks == 0);
+    FATAL_ASSERT(mci->first == NULL);
+    enif_mutex_destroy(mci->mtx);
+
+    if (orig_destroying_mbc_fn)
+	orig_destroying_mbc_fn(allctr, carrier);
+}
+
+static int migration_init(ErlNifEnv* env, void** priv_data, ERL_NIF_TERM load_info)
+{
+    void* creating_mbc_arg = (void*)my_creating_mbc;
+    void* destroying_mbc_arg = (void*)my_destroying_mbc;
+
+    if (testcase_nif_init(env, priv_data, load_info))
+	return -1;
+
+    crr_info_offset = SET_TEST_MBC_USER_HEADER(sizeof(MyCrrInfo),
+					       &creating_mbc_arg,
+					       &destroying_mbc_arg);
+    FATAL_ASSERT(crr_info_offset >= 0);
+    orig_create_mbc_fn = creating_mbc_arg;
+    orig_destroying_mbc_fn = destroying_mbc_arg;
+
+    return 0;
+}
+
+static void add_block(MyBlock* p, MigrationState* state)
+{
+    MyCrrInfo* mci = (MyCrrInfo*)((char*)BLK_TO_MBC(UMEM2BLK_TEST(p)) + crr_info_offset);
+
+    enif_mutex_lock(mci->mtx);
+    if (++mci->nblocks > state->max_blocks_in_mbc)
+	state->max_blocks_in_mbc = mci->nblocks;
+    p->next = mci->first;
+    p->prevp = &mci->first;
+    mci->first = p;
+    if (p->next)
+	p->next->prevp = &p->next;
+    if (mci->employer != state) {
+	if (!mci->employer) {
+	    FATAL_ASSERT(mci->nblocks == 1);
+	    state->nr_of_carriers++;
+	}
+	else {
+	    state->nr_of_migrations++;
+	}
+	mci->employer = state;
+    }
+    enif_mutex_unlock(mci->mtx);
+}
+
+static void remove_block(MyBlock* p)
+{
+    MyCrrInfo* mci = (MyCrrInfo*)((char*)BLK_TO_MBC(UMEM2BLK_TEST(p)) + crr_info_offset);
+
+    enif_mutex_lock(mci->mtx);
+    mci->nblocks--;
+    if (p->next)
+	p->next->prevp = p->prevp;
+    *p->prevp = p->next;
+    enif_mutex_unlock(mci->mtx);
+}
+
+static int rand_int(MigrationState* state, int low, int high)
+{
+    int x;
+    FATAL_ASSERT(high >= low);
+    x = myrand(&state->rand_state);
+    return low + (x % (high+1-low));
+}
+
+
+static void do_cleanup(TestCaseState_t *tcs, MigrationState* state)
+{
+    if (state->nblocks == 0) {
+	state->phase = DONE;
+	testcase_printf(tcs, "%d: Done %d rounds", tcs->thr_nr, state->round);
+	testcase_printf(tcs, "%d: Cleanup all blocks", tcs->thr_nr);
+	testcase_printf(tcs, "%d: Empty carriers detected = %d", tcs->thr_nr,
+			state->nr_of_carriers);
+	testcase_printf(tcs, "%d: Migrations detected     = %d", tcs->thr_nr,
+			state->nr_of_migrations);
+	testcase_printf(tcs, "%d: Max blocks in carrier   = %d", tcs->thr_nr,
+			state->max_blocks_in_mbc);
+    }
+    else {
+	state->nblocks--;
+	if (state->blockv[state->nblocks]) {
+	    remove_block(state->blockv[state->nblocks]);
+	    FREE_TEST(state->blockv[state->nblocks]);
+	}
+    }
+}
+
+
+void
+testcase_run(TestCaseState_t *tcs)
+{
+    MigrationState* state = (MigrationState*) tcs->extra;
+
+    if (!tcs->extra) {
+	if (!IS_SMP_ENABLED)
+	    testcase_skipped(tcs, "No SMP support");
+
+	tcs->extra = enif_alloc(sizeof(MigrationState));
+	state = (MigrationState*) tcs->extra;
+	memset(state->blockv, 0, sizeof(state->blockv));
+	myrand_init(&state->rand_state, tcs->thr_nr);
+	state->phase = GROWING;
+	state->nblocks = 0;
+	state->round = 0;
+	state->nr_of_migrations = 0;
+	state->nr_of_carriers = 0;
+	state->max_blocks_in_mbc = 0;
+	state->block_size = GET_TEST_MBC_SIZE() / (BLOCKS_PER_MBC+1);
+	if (MAX_BLOCK_PER_THR * state->block_size < tcs->free_mem) {
+	    state->max_nblocks = MAX_BLOCK_PER_THR;
+	} else {
+	    state->max_nblocks = tcs->free_mem / state->block_size;
+	}
+	state->goal_nblocks = rand_int(state, 1, state->max_nblocks);
+    }
+
+    switch (state->phase) {
+    case GROWING: {
+	MyBlock* p;
+	FATAL_ASSERT(!state->blockv[state->nblocks]);
+	p = ALLOC_TEST(rand_int(state, state->block_size/2, state->block_size));
+	FATAL_ASSERT(p);
+	add_block(p, state);
+	state->blockv[state->nblocks] = p;
+	if (++state->nblocks >= state->goal_nblocks) {
+	    /*testcase_printf(tcs, "%d: Grown to %d blocks", tcs->thr_nr, state->nblocks);*/
+	    state->phase = SHRINKING;
+	    state->goal_nblocks = rand_int(state, 0, state->goal_nblocks-1);
+	}
+	else
+	    FATAL_ASSERT(!state->blockv[state->nblocks]);
+	break;
+    }
+    case SHRINKING: {
+	int ix = rand_int(state, 0, state->nblocks-1);
+	FATAL_ASSERT(state->blockv[ix]);
+	remove_block(state->blockv[ix]);
+	FREE_TEST(state->blockv[ix]);
+	state->blockv[ix] = state->blockv[--state->nblocks];
+	state->blockv[state->nblocks] = NULL;
+
+	if (state->nblocks <= state->goal_nblocks) {
+	    /*testcase_printf(tcs, "%d: Shrunk to %d blocks", tcs->thr_nr, state->nblocks);*/
+	    if (++state->round >= MAX_ROUNDS) {
+		state->phase = CLEANUP;
+	    } else {
+		state->phase = GROWING;
+		state->goal_nblocks = rand_int(state, state->goal_nblocks+1, state->max_nblocks);
+	    }
+	}
+	break;
+    }
+    case CLEANUP:
+	do_cleanup(tcs, state);
+	break;
+
+    default:
+	FATAL_ASSERT(!"Invalid phase");
+    }
+
+    if (state->phase == DONE) {
+    }
+    else {
+	testcase_continue(tcs);
+    }
+}
+
+void
+testcase_cleanup(TestCaseState_t *tcs)
+{
+    MigrationState* state = (MigrationState*) tcs->extra;
+
+    while (state->phase != DONE)
+	do_cleanup(tcs, state);
+
+    enif_free(tcs->extra);
+    tcs->extra = NULL;
+}
+
+
+ERL_NIF_INIT(migration, testcase_nif_funcs, migration_init,
+	     NULL, NULL, NULL);
diff --git a/erts/emulator/test/alloc_SUITE_data/migration.erl b/erts/emulator/test/alloc_SUITE_data/migration.erl
new file mode 100644
index 0000000000..440a99becd
--- /dev/null
+++ b/erts/emulator/test/alloc_SUITE_data/migration.erl
@@ -0,0 +1,10 @@
+-module(migration).
+
+-export([init/1, start/1, run/1, stop/1]).
+
+init(File) ->
+    ok = erlang:load_nif(File, 0).
+
+start(_) -> erlang:nif_error(not_loaded).
+run(_)  -> erlang:nif_error(not_loaded).
+stop(_) -> erlang:nif_error(not_loaded).
diff --git a/erts/emulator/test/alloc_SUITE_data/mseg_clear_cache.c b/erts/emulator/test/alloc_SUITE_data/mseg_clear_cache.c
index 9c03f3a331..e5df3d647f 100644
--- a/erts/emulator/test/alloc_SUITE_data/mseg_clear_cache.c
+++ b/erts/emulator/test/alloc_SUITE_data/mseg_clear_cache.c
@@ -101,3 +101,6 @@ testcase_cleanup(TestCaseState_t *tcs)
 	tcs->extra = NULL;
     }
 }
+
+ERL_NIF_INIT(mseg_clear_cache, testcase_nif_funcs, testcase_nif_init,
+	     NULL, NULL, NULL);
diff --git a/erts/emulator/test/alloc_SUITE_data/mseg_clear_cache.erl b/erts/emulator/test/alloc_SUITE_data/mseg_clear_cache.erl
new file mode 100644
index 0000000000..befd6c2e8e
--- /dev/null
+++ b/erts/emulator/test/alloc_SUITE_data/mseg_clear_cache.erl
@@ -0,0 +1,10 @@
+-module(mseg_clear_cache).
+
+-export([init/1, start/1, run/1, stop/1]).
+
+init(File) ->
+    ok = erlang:load_nif(File, 0).
+
+start(_) -> erlang:nif_error(not_loaded).
+run(_)  -> erlang:nif_error(not_loaded).
+stop(_) -> erlang:nif_error(not_loaded).
diff --git a/erts/emulator/test/alloc_SUITE_data/rbtree.c b/erts/emulator/test/alloc_SUITE_data/rbtree.c
index 8d4d5535a8..38bbbdf90c 100644
--- a/erts/emulator/test/alloc_SUITE_data/rbtree.c
+++ b/erts/emulator/test/alloc_SUITE_data/rbtree.c
@@ -20,7 +20,7 @@
 #include "testcase_driver.h"
 #include "allocator_test.h"
 
-#define NO_BLOCKS 100000
+int NO_BLOCKS;
 
 #define RIGHT_VISITED (1 << 0)
 #define LEFT_VISITED (1 << 1)
@@ -265,9 +265,10 @@ check_tree(TestCaseState_t *tcs, Allctr_t *alc, Ulong size)
     ASSERT(tcs, curr_blacks == 0);
     ASSERT(tcs, i == -1);
 
+    /*
     testcase_printf(tcs, "Red-Black Tree OK! Max depth = %d; "
 		    "Black depth = %d\n", max_i+1, blacks < 0 ? 0 : blacks);
-
+    */
     return res;
 
 }
@@ -468,6 +469,12 @@ testcase_run(TestCaseState_t *tcs)
     Allctr_t *a;
     rbtree_test_data *td;
 
+    NO_BLOCKS = 100*1000;
+    if (enif_is_identical(tcs->build_type,
+                          enif_make_atom(tcs->curr_env,"valgrind"))) {
+        NO_BLOCKS /= 10;
+    }
+
     /* Best fit... */
 
     testcase_printf(tcs, "Setup...\n");
@@ -577,3 +584,6 @@ testcase_run(TestCaseState_t *tcs)
     testcase_printf(tcs, "aoffcaobf test succeeded!\n");
 
 }
+
+ERL_NIF_INIT(rbtree, testcase_nif_funcs, testcase_nif_init,
+	     NULL, NULL, NULL);
diff --git a/erts/emulator/test/alloc_SUITE_data/rbtree.erl b/erts/emulator/test/alloc_SUITE_data/rbtree.erl
new file mode 100644
index 0000000000..f5b7120ff2
--- /dev/null
+++ b/erts/emulator/test/alloc_SUITE_data/rbtree.erl
@@ -0,0 +1,10 @@
+-module(rbtree).
+
+-export([init/1, start/1, run/1, stop/1]).
+
+init(File) ->
+    ok = erlang:load_nif(File, 0).
+
+start(_) -> erlang:nif_error(not_loaded).
+run(_)  -> erlang:nif_error(not_loaded).
+stop(_) -> erlang:nif_error(not_loaded).
diff --git a/erts/emulator/test/alloc_SUITE_data/realloc_copy.c b/erts/emulator/test/alloc_SUITE_data/realloc_copy.c
index e405f06225..c4147eb00d 100644
--- a/erts/emulator/test/alloc_SUITE_data/realloc_copy.c
+++ b/erts/emulator/test/alloc_SUITE_data/realloc_copy.c
@@ -278,3 +278,5 @@ testcase_cleanup(TestCaseState_t *tcs)
 	STOP_ALC((Allctr_t *) tcs->extra);
 }
 
+ERL_NIF_INIT(realloc_copy, testcase_nif_funcs, testcase_nif_init,
+	     NULL, NULL, NULL);
diff --git a/erts/emulator/test/alloc_SUITE_data/realloc_copy.erl b/erts/emulator/test/alloc_SUITE_data/realloc_copy.erl
new file mode 100644
index 0000000000..cc6617bf64
--- /dev/null
+++ b/erts/emulator/test/alloc_SUITE_data/realloc_copy.erl
@@ -0,0 +1,10 @@
+-module(realloc_copy).
+
+-export([init/1, start/1, run/1, stop/1]).
+
+init(File) ->
+    ok = erlang:load_nif(File, 0).
+
+start(_) -> erlang:nif_error(not_loaded).
+run(_)  -> erlang:nif_error(not_loaded).
+stop(_) -> erlang:nif_error(not_loaded).
diff --git a/erts/emulator/test/alloc_SUITE_data/testcase_driver.c b/erts/emulator/test/alloc_SUITE_data/testcase_driver.c
index bc674c56b7..7dcca544e5 100644
--- a/erts/emulator/test/alloc_SUITE_data/testcase_driver.c
+++ b/erts/emulator/test/alloc_SUITE_data/testcase_driver.c
@@ -23,141 +23,147 @@
 #include <stdarg.h>
 #include <setjmp.h>
 #include <string.h>
+#include <limits.h>
 
 #ifdef __WIN32__
-#undef HAVE_VSNPRINTF
-#define HAVE_VSNPRINTF 1
-#define vsnprintf _vsnprintf
+static void my_vsnprintf(char *outBuf, size_t size, const char *format, va_list ap)
+{
+    _vsnprintf(outBuf, size, format, ap);
+    outBuf[size-1] = 0;  /* be sure string is terminated */
+}
+#elif defined(HAVE_VSNPRINTF)
+#  define my_vsnprintf(B,S,F,A) (void)vsnprintf(B,S,F,A)
+#else
+#  warning Using unsafe 'vsprintf' without buffer overflow protection
+#  define my_vsnprintf(B,S,F,A) (void)vsprintf(B,F,A)
 #endif
 
-#ifndef HAVE_VSNPRINTF
-#define HAVE_VSNPRINTF 0
-#endif
+static void my_snprintf(char *outBuf, size_t size, const char *format, ...)
+{
+    va_list ap;
+    va_start(ap, format);
+    my_vsnprintf(outBuf, size, format, ap);
+    va_end(ap);
+}
 
 #define COMMENT_BUF_SZ 4096
 
 #define TESTCASE_FAILED		0
 #define TESTCASE_SKIPPED	1
 #define TESTCASE_SUCCEEDED	2
+#define TESTCASE_CONTINUE	3
 
 typedef struct {
     TestCaseState_t visible;
-    ErlDrvPort port;
-    ErlDrvTermData  port_id;
     int result;
-    jmp_buf done_jmp_buf;
+    jmp_buf* done_jmp_buf;
     char *comment;
     char comment_buf[COMMENT_BUF_SZ];
 } InternalTestCaseState_t;
 
-ErlDrvData testcase_drv_start(ErlDrvPort port, char *command);
-void testcase_drv_stop(ErlDrvData drv_data);
-void testcase_drv_run(ErlDrvData drv_data, char *buf, ErlDrvSizeT len);
-
-static ErlDrvEntry testcase_drv_entry = { 
-    NULL,
-    testcase_drv_start,
-    testcase_drv_stop,
-    testcase_drv_run,
-    NULL,
-    NULL,
-    "testcase_drv",
-    NULL,
-    NULL,
-    NULL,
-    NULL,
-    NULL,
-    NULL,
-    NULL,
-    NULL,
-    NULL,
-    ERL_DRV_EXTENDED_MARKER,
-    ERL_DRV_EXTENDED_MAJOR_VERSION,
-    ERL_DRV_EXTENDED_MINOR_VERSION,
-    0,
-    NULL,
-    NULL,
-    NULL
+ERL_NIF_TERM testcase_nif_start(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]);
+ERL_NIF_TERM testcase_nif_stop(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]);
+ERL_NIF_TERM testcase_nif_run(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[]);
+
+ErlNifFunc testcase_nif_funcs[] =
+{
+    {"start", 1, testcase_nif_start},
+    {"run", 1, testcase_nif_run},
+    {"stop", 1, testcase_nif_stop}
 };
 
+static ErlNifResourceType* testcase_rt;
+static ERL_NIF_TERM print_atom;
 
-DRIVER_INIT(testcase_drv)
+int testcase_nif_init(ErlNifEnv* env, void** priv_data, ERL_NIF_TERM load_info)
 {
-    testcase_drv_entry.driver_name = testcase_name();
-    return &testcase_drv_entry;
+    testcase_rt = enif_open_resource_type(env, NULL, "testcase_rt", NULL,
+					  ERL_NIF_RT_CREATE, NULL);
+
+    print_atom = enif_make_atom(env, "print");
+    return 0;
 }
 
-ErlDrvData
-testcase_drv_start(ErlDrvPort port, char *command)
-{
+ERL_NIF_TERM
+testcase_nif_start(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[])
+{ /* (ThrNr, FreeMeg, BuildType) */
+    ERL_NIF_TERM ret;
     InternalTestCaseState_t *itcs = (InternalTestCaseState_t *)
-	driver_alloc(sizeof(InternalTestCaseState_t));
-    if (!itcs) {
-	return ERL_DRV_ERROR_GENERAL;
+             enif_alloc_resource(testcase_rt, sizeof(InternalTestCaseState_t));
+    int free_megabyte;
+    const int max_megabyte = INT_MAX / (1024*1024);
+    const ERL_NIF_TERM* tpl;
+    int tpl_arity;
+
+    if (!itcs
+        || !enif_get_tuple(env, argv[0], &tpl_arity, &tpl)
+        || tpl_arity != 3
+	|| !enif_get_int(env, tpl[0], &itcs->visible.thr_nr)
+	|| !enif_get_int(env, tpl[1], &free_megabyte)) {
+	enif_make_badarg(env);
     }
-
+    itcs->visible.free_mem = (free_megabyte < max_megabyte ?
+                              free_megabyte : max_megabyte) * (1024*1024);
     itcs->visible.testcase_name = testcase_name();
+    itcs->visible.build_type = tpl[2];
     itcs->visible.extra = NULL;
-    itcs->port = port;
-    itcs->port_id = driver_mk_port(port);
     itcs->result = TESTCASE_FAILED;
     itcs->comment = "";
 
-    return (ErlDrvData) itcs;
+    ret = enif_make_resource(env, itcs);
+    enif_release_resource(itcs);
+    return ret;
 }
 
-void
-testcase_drv_stop(ErlDrvData drv_data)
+ERL_NIF_TERM
+testcase_nif_stop(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[])
 {
-    testcase_cleanup((TestCaseState_t *) drv_data);
-    driver_free((void *) drv_data);
+    InternalTestCaseState_t *itcs;
+    if (!enif_get_resource(env, argv[0], testcase_rt, (void**)&itcs))
+	return enif_make_badarg(env);
+    testcase_cleanup(&itcs->visible);
+    return enif_make_atom(env,"ok");
 }
 
-void
-testcase_drv_run(ErlDrvData drv_data, char *buf, ErlDrvSizeT len)
+ERL_NIF_TERM
+testcase_nif_run(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[])
 {
-    InternalTestCaseState_t *itcs = (InternalTestCaseState_t *) drv_data;
-    ErlDrvTermData result_atom;
-    ErlDrvTermData msg[12];
+    InternalTestCaseState_t *itcs;
+    const char* result_atom;
+    jmp_buf the_jmp_buf;
+
+    if (!enif_get_resource(env, argv[0], testcase_rt, (void**)&itcs))
+	return enif_make_badarg(env);
 
-    itcs->visible.command = buf;
-    itcs->visible.command_len = len;
+    itcs->visible.curr_env = env;
 
-    if (setjmp(itcs->done_jmp_buf) == 0) {
-	testcase_run((TestCaseState_t *) itcs);
+    /* For some unknown reason, first call to setjmp crashes on win64
+     * when jmp_buf is allocated as part of the resource. But it works when
+     * allocated on stack. It used to work when this was a driver.
+     */
+    itcs->done_jmp_buf = &the_jmp_buf;
+
+    if (setjmp(the_jmp_buf) == 0) {
+	testcase_run(&itcs->visible);
 	itcs->result = TESTCASE_SUCCEEDED;
     }
 
     switch (itcs->result) {
-    case TESTCASE_SUCCEEDED:
-	result_atom = driver_mk_atom("succeeded");
-	break;
-    case TESTCASE_SKIPPED:
-	result_atom = driver_mk_atom("skipped");
-	break;
-    case TESTCASE_FAILED:
+    case TESTCASE_CONTINUE:
+	return enif_make_atom(env, "continue");
+
+    case TESTCASE_SUCCEEDED: result_atom = "succeeded"; break;
+    case TESTCASE_SKIPPED: result_atom = "skipped"; break;
+    case TESTCASE_FAILED: result_atom = "failed"; break;
     default:
-	result_atom = driver_mk_atom("failed");
-	break;
+        result_atom = "failed";
+        my_snprintf(itcs->comment_buf, sizeof(itcs->comment_buf),
+		    "Unexpected test result code %d.", itcs->result);
+        itcs->comment = itcs->comment_buf;
     }
 
-    msg[0] = ERL_DRV_ATOM;
-    msg[1] = (ErlDrvTermData) result_atom;
-
-    msg[2] = ERL_DRV_PORT;
-    msg[3] = itcs->port_id;
-
-    msg[4] = ERL_DRV_ATOM;
-    msg[5] = driver_mk_atom(itcs->visible.testcase_name);
- 
-    msg[6] = ERL_DRV_STRING;
-    msg[7] = (ErlDrvTermData) itcs->comment;
-    msg[8] = (ErlDrvTermData) strlen(itcs->comment);
-
-    msg[9] = ERL_DRV_TUPLE;
-    msg[10] = (ErlDrvTermData) 4;
-
-    erl_drv_output_term(itcs->port_id, msg, 11);
+    return enif_make_tuple2(env, enif_make_atom(env, result_atom),
+			    enif_make_string(env, itcs->comment, ERL_NIF_LATIN1));
 }
 
 int
@@ -172,34 +178,22 @@ testcase_assertion_failed(TestCaseState_t *tcs,
 void
 testcase_printf(TestCaseState_t *tcs, char *frmt, ...)
 {
-    InternalTestCaseState_t *itcs = (InternalTestCaseState_t *) tcs;
-    ErlDrvTermData msg[12];
+    InternalTestCaseState_t* itcs = (InternalTestCaseState_t*)tcs;
+    ErlNifPid pid;
+    ErlNifEnv* msg_env = enif_alloc_env();
+    ERL_NIF_TERM msg;
     va_list va;
     va_start(va, frmt);
-#if HAVE_VSNPRINTF
-    vsnprintf(itcs->comment_buf, COMMENT_BUF_SZ, frmt, va);
-#else
-    vsprintf(itcs->comment_buf, frmt, va);
-#endif
+    my_vsnprintf(itcs->comment_buf, COMMENT_BUF_SZ, frmt, va);
     va_end(va);
 
-    msg[0] = ERL_DRV_ATOM;
-    msg[1] = (ErlDrvTermData) driver_mk_atom("print");
+    msg = enif_make_tuple2(msg_env, print_atom,
+		    enif_make_string(msg_env, itcs->comment_buf, ERL_NIF_LATIN1));
 
-    msg[2] = ERL_DRV_PORT;
-    msg[3] = itcs->port_id;
+    enif_send(itcs->visible.curr_env, enif_self(itcs->visible.curr_env, &pid),
+	      msg_env, msg);
 
-    msg[4] = ERL_DRV_ATOM;
-    msg[5] = driver_mk_atom(itcs->visible.testcase_name);
-
-    msg[6] = ERL_DRV_STRING;
-    msg[7] = (ErlDrvTermData) itcs->comment_buf;
-    msg[8] = (ErlDrvTermData) strlen(itcs->comment_buf);
-
-    msg[9] = ERL_DRV_TUPLE;
-    msg[10] = (ErlDrvTermData) 4;
-
-    erl_drv_output_term(itcs->port_id, msg, 11);
+    enif_free_env(msg_env);
 }
 
 
@@ -208,17 +202,13 @@ void testcase_succeeded(TestCaseState_t *tcs, char *frmt, ...)
     InternalTestCaseState_t *itcs = (InternalTestCaseState_t *) tcs;
     va_list va;
     va_start(va, frmt);
-#if HAVE_VSNPRINTF
-    vsnprintf(itcs->comment_buf, COMMENT_BUF_SZ, frmt, va);
-#else
-    vsprintf(itcs->comment_buf, frmt, va);
-#endif
+    my_vsnprintf(itcs->comment_buf, COMMENT_BUF_SZ, frmt, va);
     va_end(va);
 
     itcs->result = TESTCASE_SUCCEEDED;
     itcs->comment = itcs->comment_buf;
 
-    longjmp(itcs->done_jmp_buf, 1);
+    longjmp(*itcs->done_jmp_buf, 1);
 }
 
 void testcase_skipped(TestCaseState_t *tcs, char *frmt, ...)
@@ -226,17 +216,20 @@ void testcase_skipped(TestCaseState_t *tcs, char *frmt, ...)
     InternalTestCaseState_t *itcs = (InternalTestCaseState_t *) tcs;
     va_list va;
     va_start(va, frmt);
-#if HAVE_VSNPRINTF
-    vsnprintf(itcs->comment_buf, COMMENT_BUF_SZ, frmt, va);
-#else
-    vsprintf(itcs->comment_buf, frmt, va);
-#endif
+    my_vsnprintf(itcs->comment_buf, COMMENT_BUF_SZ, frmt, va);
     va_end(va);
 
     itcs->result = TESTCASE_SKIPPED;
     itcs->comment = itcs->comment_buf;
 
-    longjmp(itcs->done_jmp_buf, 1);
+    longjmp(*itcs->done_jmp_buf, 1);
+}
+
+void testcase_continue(TestCaseState_t *tcs)
+{
+    InternalTestCaseState_t *itcs = (InternalTestCaseState_t *) tcs;
+    itcs->result = TESTCASE_CONTINUE;
+    longjmp(*itcs->done_jmp_buf, 1);
 }
 
 void testcase_failed(TestCaseState_t *tcs, char *frmt, ...)
@@ -246,37 +239,33 @@ void testcase_failed(TestCaseState_t *tcs, char *frmt, ...)
     size_t bufsz = sizeof(buf);
     va_list va;
     va_start(va, frmt);
-#if HAVE_VSNPRINTF
-    vsnprintf(itcs->comment_buf, COMMENT_BUF_SZ, frmt, va);
-#else
-    vsprintf(itcs->comment_buf, frmt, va);
-#endif
+    my_vsnprintf(itcs->comment_buf, COMMENT_BUF_SZ, frmt, va);
     va_end(va);
 
     itcs->result = TESTCASE_FAILED;
     itcs->comment = itcs->comment_buf;
 
-    if (erl_drv_getenv("ERL_ABORT_ON_FAILURE", buf, &bufsz) == 0
+    if (enif_getenv("ERL_ABORT_ON_FAILURE", buf, &bufsz) == 0
 	&& strcmp("true", buf) == 0) {
 	fprintf(stderr, "Testcase \"%s\" failed: %s\n",
 		itcs->visible.testcase_name, itcs->comment);
 	abort();
     }
 
-    longjmp(itcs->done_jmp_buf, 1);
+    longjmp(*itcs->done_jmp_buf, 1);
 }
 
 void *testcase_alloc(size_t size)
 {
-    return driver_alloc(size);
+    return enif_alloc(size);
 }
 
 void *testcase_realloc(void *ptr, size_t size)
 {
-    return driver_realloc(ptr, size);
+    return enif_realloc(ptr, size);
 }
 
 void testcase_free(void *ptr)
 {
-    driver_free(ptr);
+    enif_free(ptr);
 }
diff --git a/erts/emulator/test/alloc_SUITE_data/testcase_driver.h b/erts/emulator/test/alloc_SUITE_data/testcase_driver.h
index 5d17eaec64..f0ca91bd06 100644
--- a/erts/emulator/test/alloc_SUITE_data/testcase_driver.h
+++ b/erts/emulator/test/alloc_SUITE_data/testcase_driver.h
@@ -20,13 +20,15 @@
 #ifndef TESTCASE_DRIVER_H__
 #define TESTCASE_DRIVER_H__
 
-#include "erl_driver.h"
+#include "erl_nif.h"
 #include <stdlib.h>
 
 typedef struct {
+    ErlNifEnv* curr_env;
     char *testcase_name;
-    char *command;
-    int command_len;
+    int thr_nr;
+    int free_mem; /* in bytes */
+    ERL_NIF_TERM build_type;  /* opt, debug, valgrind, ... */
     void *extra;
 } TestCaseState_t;
 
@@ -34,9 +36,11 @@ typedef struct {
   ((void) ((B) ? 1 : testcase_assertion_failed((TCS), __FILE__, __LINE__, #B)))
 
 
+int testcase_nif_init(ErlNifEnv* env, void** priv_data, ERL_NIF_TERM load_info);
 void testcase_printf(TestCaseState_t *tcs, char *frmt, ...);
 void testcase_succeeded(TestCaseState_t *tcs, char *frmt, ...);
 void testcase_skipped(TestCaseState_t *tcs, char *frmt, ...);
+void testcase_continue(TestCaseState_t *tcs);
 void testcase_failed(TestCaseState_t *tcs, char *frmt, ...);
 int testcase_assertion_failed(TestCaseState_t *tcs, char *file, int line,
 			      char *assertion);
@@ -45,8 +49,11 @@ void *testcase_realloc(void *ptr, size_t size);
 void testcase_free(void *ptr);
 
 
+/* Implemented by testcase: */
 char *testcase_name(void);
 void testcase_run(TestCaseState_t *tcs);
 void testcase_cleanup(TestCaseState_t *tcs);
 
-#endif
+extern ErlNifFunc testcase_nif_funcs[3];
+
+#endif /* TESTCASE_DRIVER_H__ */
diff --git a/erts/emulator/test/alloc_SUITE_data/threads.c b/erts/emulator/test/alloc_SUITE_data/threads.c
index edad24ee6b..2f5f841e3d 100644
--- a/erts/emulator/test/alloc_SUITE_data/threads.c
+++ b/erts/emulator/test/alloc_SUITE_data/threads.c
@@ -86,7 +86,7 @@ static void fail(int t_no, char *frmt, ...)
 
     tc_failed = 1;
 
-    if (erl_drv_getenv("ERL_ABORT_ON_FAILURE", buf, &bufsz) == 0
+    if (enif_getenv("ERL_ABORT_ON_FAILURE", buf, &bufsz) == 0
 	&& strcmp("true", buf) == 0) {
 	fprintf(stderr, "Testcase \"%s\" failed: %s\n",
 		testcase_name(), err_buf);
@@ -187,7 +187,6 @@ testcase_run(TestCaseState_t *tcs)
 
     for(i = 1; i <= NO_OF_THREADS; i++) {
 	char *alc;
-	int res;
 
 	threads[i].arg.no_ops_per_bl = NO_OF_OPS_PER_BL;
 
@@ -397,7 +396,7 @@ alloc_op(int t_no, Allctr_t *a, block *bp, int id, int clean_up)
 	bp->p = (unsigned char *) ALLOC(a, bp->s);
 	if(!bp->p)
 	    fail(t_no, "ALLOC(%lu) failed [id=%d])\n", bp->s, id);
-	memset((void *) bp->p, id, (size_t) bp->s);
+	memset((void *) bp->p, (unsigned char)id, (size_t) bp->s);
     }
     else {
 	unsigned char *p = (unsigned char *) REALLOC(a, bp->p, bp->as[bp->i]);
@@ -407,7 +406,7 @@ alloc_op(int t_no, Allctr_t *a, block *bp, int id, int clean_up)
 
 	if(bp->s < bp->as[bp->i]) {
 	    CHECK_BLOCK_DATA(t_no, p, bp->s, id);
-	    memset((void *) p, id, (size_t) bp->as[bp->i]);
+	    memset((void *) p, (unsigned char)id, (size_t) bp->as[bp->i]);
 	}
 	else
 	    CHECK_BLOCK_DATA(t_no, p, bp->as[bp->i], id);
@@ -446,3 +445,6 @@ thread_func(void *arg)
     exit_thread(td->t_no, 1);
     return NULL;
 }
+
+ERL_NIF_INIT(threads, testcase_nif_funcs, testcase_nif_init,
+	     NULL, NULL, NULL);
diff --git a/erts/emulator/test/alloc_SUITE_data/threads.erl b/erts/emulator/test/alloc_SUITE_data/threads.erl
new file mode 100644
index 0000000000..a7b4965f5e
--- /dev/null
+++ b/erts/emulator/test/alloc_SUITE_data/threads.erl
@@ -0,0 +1,10 @@
+-module(threads).
+
+-export([init/1, start/1, run/1, stop/1]).
+
+init(File) ->
+    ok = erlang:load_nif(File, 0).
+
+start(_) -> erlang:nif_error(not_loaded).
+run(_)  -> erlang:nif_error(not_loaded).
+stop(_) -> erlang:nif_error(not_loaded).
diff --git a/erts/emulator/test/bif_SUITE.erl b/erts/emulator/test/bif_SUITE.erl
index d6a771e7b9..ebc4db53c4 100644
--- a/erts/emulator/test/bif_SUITE.erl
+++ b/erts/emulator/test/bif_SUITE.erl
@@ -32,7 +32,8 @@
 	 specs/1,improper_bif_stubs/1,auto_imports/1,
 	 t_list_to_existing_atom/1,os_env/1,otp_7526/1,
 	 binary_to_atom/1,binary_to_existing_atom/1,
-	 atom_to_binary/1,min_max/1, erlang_halt/1]).
+	 atom_to_binary/1,min_max/1, erlang_halt/1,
+	 is_builtin/1]).
 
 suite() -> [{ct_hooks,[ts_install_cth]}].
 
@@ -42,7 +43,7 @@ all() ->
      t_list_to_existing_atom, os_env, otp_7526,
      display,
      atom_to_binary, binary_to_atom, binary_to_existing_atom,
-     min_max, erlang_halt].
+     min_max, erlang_halt, is_builtin].
 
 groups() -> 
     [].
@@ -716,6 +717,22 @@ wait_until_stable_size(File,PrevSz) ->
             wait_until_stable_size(File,NewSz)
     end.
 
+is_builtin(_Config) ->
+    Exp0 = [{M,F,A} || {M,_} <- code:all_loaded(),
+		       {F,A} <- M:module_info(exports)],
+    Exp = ordsets:from_list(Exp0),
+
+    %% erlang:apply/3 is considered to be built-in, but is not
+    %% implemented as other BIFs.
+
+    Builtins0 = [{erlang,apply,3}|erlang:system_info(snifs)],
+    Builtins = ordsets:from_list(Builtins0),
+    NotBuiltin = ordsets:subtract(Exp, Builtins),
+    _ = [true = erlang:is_builtin(M, F, A) || {M,F,A} <- Builtins],
+    _ = [false = erlang:is_builtin(M, F, A) || {M,F,A} <- NotBuiltin],
+
+    ok.
+
 
 %% Helpers
     
diff --git a/erts/emulator/test/binary_SUITE.erl b/erts/emulator/test/binary_SUITE.erl
index 96ba2f64d4..f8f71efecc 100644
--- a/erts/emulator/test/binary_SUITE.erl
+++ b/erts/emulator/test/binary_SUITE.erl
@@ -521,30 +521,29 @@ external_size_1(Term, Size0, Limit) when Size0 < Limit ->
 external_size_1(_, _, _) -> ok.
 
 t_iolist_size(Config) when is_list(Config) ->
-    ?line Seed = {erlang:monotonic_time(),
-		  erlang:time_offset(),
-		  erlang:unique_integer([positive])},
-    ?line io:format("Seed: ~p", [Seed]),
-    ?line random:seed(Seed),
-    ?line Base = <<0:(1 bsl 20)/unit:8>>,
-    ?line Powers = [1 bsl N || N <- lists:seq(2, 37)],
-    ?line Sizes0 = [[N - random:uniform(N div 2),
-		     lists:seq(N-2, N+2),
-		     N+N div 2,
-		     N + random:uniform(N div 2)] ||
-		       N <- Powers],
+    _ = rand:uniform(),				%Seed generator
+    io:format("Seed: ~p", [rand:export_seed()]),
+
+    Base = <<0:(1 bsl 20)/unit:8>>,
+    Powers = [1 bsl N || N <- lists:seq(2, 37)],
+    Sizes0 = [[N - rand:uniform(N div 2),
+	       lists:seq(N-2, N+2),
+	       N+N div 2,
+	       N + rand:uniform(N div 2)] ||
+		 N <- Powers],
+
     %% Test sizes around 1^32 more thoroughly.
     FourGigs = 1 bsl 32,
-    ?line Sizes1 = [FourGigs+N || N <- lists:seq(-8, 40)] ++ Sizes0,
-    ?line Sizes2 = lists:flatten(Sizes1),
-    ?line Sizes = lists:usort(Sizes2),
+    Sizes1 = [FourGigs+N || N <- lists:seq(-8, 40)] ++ Sizes0,
+    Sizes2 = lists:flatten(Sizes1),
+    Sizes = lists:usort(Sizes2),
     io:format("~p sizes:", [length(Sizes)]),
     io:format("~p\n", [Sizes]),
-    ?line [Sz = iolist_size(build_iolist(Sz, Base)) || Sz <- Sizes],
+    _ = [Sz = iolist_size(build_iolist(Sz, Base)) || Sz <- Sizes],
     ok.
 
 build_iolist(N, Base) when N < 16 ->
-    case random:uniform(3) of
+    case rand:uniform(3) of
 	1 ->
 	    <<Bin:N/binary,_/binary>> = Base,
 	    Bin;
@@ -552,7 +551,7 @@ build_iolist(N, Base) when N < 16 ->
 	    lists:seq(1, N)
     end;
 build_iolist(N, Base) when N =< byte_size(Base) ->
-    case random:uniform(3) of
+    case rand:uniform(3) of
 	1 ->
 	    <<Bin:N/binary,_/binary>> = Base,
 	    Bin;
@@ -570,7 +569,7 @@ build_iolist(N, Base) when N =< byte_size(Base) ->
 	    end
     end;
 build_iolist(N0, Base) ->
-    Small = random:uniform(15),
+    Small = rand:uniform(15),
     Seq = lists:seq(1, Small),
     N = N0 - Small,
     case N rem 2 of
@@ -1604,7 +1603,7 @@ bit_sized_binary(Bin0) ->
 
 unaligned_sub_bin(Bin, 0) -> Bin;
 unaligned_sub_bin(Bin0, Offs) ->
-    F = random:uniform(256),
+    F = rand:uniform(256),
     Roffs = 8-Offs,
     Bin1 = <<F:Offs,Bin0/binary,F:Roffs>>,
     Sz = size(Bin0),
diff --git a/erts/emulator/test/bs_bincomp_SUITE.erl b/erts/emulator/test/bs_bincomp_SUITE.erl
index dcd13c19df..8836fe40ae 100644
--- a/erts/emulator/test/bs_bincomp_SUITE.erl
+++ b/erts/emulator/test/bs_bincomp_SUITE.erl
@@ -131,7 +131,7 @@ tracing(Config) when is_list(Config) ->
 
 random_binary() ->
     Seq = [1,2,3,4,5,6,7,8,9,10],
-    << <<($a + random:uniform($z - $a)):8>> || _ <- Seq >>.
+    << <<($a + rand:uniform($z - $a)):8>> || _ <- Seq >>.
 
 random_binaries(N) when N > 0 ->
     random_binary(),
diff --git a/erts/emulator/test/code_SUITE.erl b/erts/emulator/test/code_SUITE.erl
index 9f318a38be..1acc4538fb 100644
--- a/erts/emulator/test/code_SUITE.erl
+++ b/erts/emulator/test/code_SUITE.erl
@@ -26,7 +26,8 @@
 	 t_check_process_code_ets/1,
 	 external_fun/1,get_chunk/1,module_md5/1,make_stub/1,
 	 make_stub_many_funs/1,constant_pools/1,constant_refc_binaries/1,
-	 false_dependency/1,coverage/1,fun_confusion/1]).
+	 false_dependency/1,coverage/1,fun_confusion/1,
+         t_copy_literals/1]).
 
 -define(line_trace, 1).
 -include_lib("test_server/include/test_server.hrl").
@@ -38,7 +39,7 @@ all() ->
      t_check_process_code_ets, t_check_old_code, external_fun, get_chunk,
      module_md5, make_stub, make_stub_many_funs,
      constant_pools, constant_refc_binaries, false_dependency,
-     coverage, fun_confusion].
+     coverage, fun_confusion, t_copy_literals].
 
 groups() -> 
     [].
@@ -753,6 +754,80 @@ compile_load(Mod, Src, Ver) ->
     {module,Mod} = code:load_binary(Mod, "fun_confusion.beam", Code1),
     ok.
 
+
+t_copy_literals(Config) when is_list(Config) ->
+    %% Compile the the literals module.
+    Data = ?config(data_dir, Config),
+    File = filename:join(Data, "literals"),
+    {ok,literals,Code} = compile:file(File, [report,binary]),
+    {module,literals} = erlang:load_module(literals, Code),
+
+    N   = 30,
+    Me  = self(),
+    %% reload literals code every 567 ms
+    Rel = spawn_link(fun() -> reloader(literals,Code,567) end),
+    %% add new literal msgs to the loop every 789 ms
+    Sat = spawn_link(fun() -> saturate(Me,789) end),
+    %% run for 10s
+    _   = spawn_link(fun() -> receive after 10000 -> Me ! done end end),
+    ok  = chase_msg(N, Me),
+    %% cleanup
+    Rel ! done,
+    Sat ! done,
+    ok  = flush(),
+    ok.
+
+
+chase_msg(0, Pid) ->
+    chase_loop(Pid);
+chase_msg(N, Master) ->
+    Pid = spawn_link(fun() -> chase_msg(N - 1,Master) end),
+    chase_loop(Pid).
+
+chase_loop(Pid) ->
+    receive
+        done ->
+            Pid ! done,
+            ok;
+        {_From,Msg} ->
+            Pid ! {self(), Msg},
+            ok = traverse(Msg),
+            chase_loop(Pid)
+    end.
+
+saturate(Pid,Time) ->
+    Es = [msg1,msg2,msg3,msg4,msg5],
+    Msg = [literals:E()||E <- Es],
+    Pid ! {self(), Msg},
+    receive
+        done -> ok
+    after Time ->
+              saturate(Pid,Time)
+    end.
+
+traverse([]) -> ok;
+traverse([H|T]) ->
+    ok = traverse(H),
+    traverse(T);
+traverse(T) when is_tuple(T) -> ok;
+traverse(B) when is_binary(B) -> ok;
+traverse(I) when is_integer(I) -> ok;
+traverse(#{ 1 := V1, b := V2 }) ->
+    ok = traverse(V1),
+    ok = traverse(V2),
+    ok.
+
+
+reloader(Mod,Code,Time) ->
+    receive
+        done -> ok
+    after Time ->
+              code:purge(Mod),
+              {module,Mod} = erlang:load_module(Mod, Code),
+              reloader(Mod,Code,Time)
+    end.
+
+
 %% Utilities.
 
 make_sub_binary(Bin) when is_binary(Bin) ->
@@ -775,4 +850,7 @@ bit_sized_binary(Bin0) ->
     BitSize = 8*size(Bin) + 1,
     Bin.
 
+flush() ->
+    receive _ -> flush() after 0 -> ok end.
+
 id(I) -> I.
diff --git a/erts/emulator/test/code_SUITE_data/literals.erl b/erts/emulator/test/code_SUITE_data/literals.erl
index 9802d9d3f9..a36bfe09dd 100644
--- a/erts/emulator/test/code_SUITE_data/literals.erl
+++ b/erts/emulator/test/code_SUITE_data/literals.erl
@@ -20,6 +20,7 @@
 
 -module(literals).
 -export([a/0,b/0,huge_bignum/0,binary/0,unused_binaries/0,bits/0]).
+-export([msg1/0,msg2/0,msg3/0,msg4/0,msg5/0]).
 
 a() ->
     {a,42.0,[7,38877938333399637266518333334747]}.
@@ -101,3 +102,9 @@ unused_binaries() ->
 
 bits() ->
     {bits,<<42:13,?MB_1>>}.
+
+msg1() -> "halloj".
+msg2() -> {"hello","world"}.
+msg3() -> <<"halloj">>.
+msg4() -> #{ 1=> "hello", b => "world"}.
+msg5() -> {1,2,3,4,5,6}.
diff --git a/erts/emulator/test/decode_packet_SUITE.erl b/erts/emulator/test/decode_packet_SUITE.erl
index 6a5ca20ac3..65ae94d0dc 100644
--- a/erts/emulator/test/decode_packet_SUITE.erl
+++ b/erts/emulator/test/decode_packet_SUITE.erl
@@ -53,11 +53,8 @@ end_per_group(_GroupName, Config) ->
 
 
 init_per_testcase(Func, Config) when is_atom(Func), is_list(Config) ->
-    Seed = {S1,S2,S3} = {erlang:monotonic_time(),
-			 erlang:time_offset(),
-			 erlang:unique_integer()},
-    random:seed(S1,S2,S3),
-    io:format("*** SEED: ~p ***\n", [Seed]),
+    rand:seed(exsplus),
+    io:format("*** SEED: ~p ***\n", [rand:export_seed()]),
     Dog=?t:timetrap(?t:minutes(1)),
     [{watchdog, Dog}|Config].
 
@@ -136,7 +133,7 @@ pack(Type,Body,Rest,BitOffs) ->
     {Packet,Unpacked} = pack(Type,Body),
 
     %% Make Bin a sub-bin with an arbitrary bitoffset within Orig
-    Prefix = random:uniform(1 bsl BitOffs) - 1,
+    Prefix = rand:uniform(1 bsl BitOffs) - 1,
     Orig = <<Prefix:BitOffs,Packet/binary,Rest/bits>>,
     <<_:BitOffs,Bin/bits>> = Orig,
     {Bin,Unpacked,Orig}.
@@ -151,13 +148,13 @@ pack(4,Bin) ->
     Psz = byte_size(Bin),
     {<<Psz:32,Bin/binary>>, Bin};
 pack(asn1,Bin) ->
-    Ident = case random:uniform(3) of
+    Ident = case rand:uniform(3) of
 		1 -> <<17>>;
 		2 -> <<16#1f,16#81,17>>;
 		3 -> <<16#1f,16#81,16#80,16#80,17>>
 	    end,
     Psz = byte_size(Bin),
-    Length = case random:uniform(4) of
+    Length = case rand:uniform(4) of
 		 1 when Psz < 128 -> 
 		     <<Psz:8>>;
 		 R when R=<2 andalso Psz < 16#10000 ->
@@ -177,42 +174,42 @@ pack(sunrm,Bin) ->
     {Res,Res};
 pack(cdr,Bin) ->
     GIOP = <<"GIOP">>,
-    Major = random:uniform(256) - 1,
-    Minor = random:uniform(256) - 1,
-    MType = random:uniform(256) - 1,
+    Major = rand:uniform(256) - 1,
+    Minor = rand:uniform(256) - 1,
+    MType = rand:uniform(256) - 1,
     Psz = byte_size(Bin),
-    Res = case random:uniform(2) of 
+    Res = case rand:uniform(2) of
 	      1 -> <<GIOP/binary,Major:8,Minor:8,0:8,MType:8,Psz:32/big,Bin/binary>>;
 	      2 -> <<GIOP/binary,Major:8,Minor:8,1:8,MType:8,Psz:32/little,Bin/binary>>
 	  end,
     {Res,Res};
 pack(fcgi,Bin) ->
     Ver = 1,
-    Type = random:uniform(256) - 1,
-    Id = random:uniform(65536) - 1,
-    PaddSz = random:uniform(16) - 1,    
+    Type = rand:uniform(256) - 1,
+    Id = rand:uniform(65536) - 1,
+    PaddSz = rand:uniform(16) - 1,
     Psz = byte_size(Bin),
-    Reserv = random:uniform(256) - 1,
+    Reserv = rand:uniform(256) - 1,
     Padd = case PaddSz of
 	       0 -> <<>>;
-	       _ -> list_to_binary([random:uniform(256)-1
+	       _ -> list_to_binary([rand:uniform(256)-1
 				    || _<- lists:seq(1,PaddSz)])
 	   end,
     Res = <<Ver:8,Type:8,Id:16,Psz:16/big,PaddSz:8,Reserv:8,Bin/binary>>,
     {<<Res/binary,Padd/binary>>, Res};
 pack(tpkt,Bin) ->
     Ver = 3,
-    Reserv = random:uniform(256) - 1,
+    Reserv = rand:uniform(256) - 1,
     Size = byte_size(Bin) + 4,
     Res = <<Ver:8,Reserv:8,Size:16,Bin/binary>>,
     {Res, Res};
 pack(ssl_tls,Bin) ->
-    Content = case (random:uniform(256) - 1) of
+    Content = case (rand:uniform(256) - 1) of
 		  C when C<128 -> C;
 		  _ -> v2hello
 	      end,
-    Major = random:uniform(256) - 1,
-    Minor = random:uniform(256) - 1,
+    Major = rand:uniform(256) - 1,
+    Minor = rand:uniform(256) - 1,
     pack_ssl(Content,Major,Minor,Bin).
 
 pack_ssl(Content, Major, Minor, Body) ->
@@ -371,10 +368,10 @@ http_do({Bin,[{_Line,PL,PB}|Tail]}, Type) ->
     ?line {ok, PB, Rest} = decode_pkt(http_with_bin(Type),Bin),
 
     %% Same tests again but as SubBin
-    PreLen = random:uniform(64),
-    Prefix = random:uniform(1 bsl PreLen) - 1,
-    SufLen = random:uniform(64),
-    Suffix = random:uniform(1 bsl SufLen) - 1,
+    PreLen = rand:uniform(64),
+    Prefix = rand:uniform(1 bsl PreLen) - 1,
+    SufLen = rand:uniform(64),
+    Suffix = rand:uniform(1 bsl SufLen) - 1,
     Orig = <<Prefix:PreLen, Bin/bits, Suffix:SufLen>>,
     BinLen = bit_size(Bin),
     <<_:PreLen, SubBin:BinLen/bits, _/bits>> = Orig, % Make SubBin
diff --git a/erts/emulator/test/distribution_SUITE_data/run.erl b/erts/emulator/test/distribution_SUITE_data/run.erl
index f5169e160c..d5ed139369 100644
--- a/erts/emulator/test/distribution_SUITE_data/run.erl
+++ b/erts/emulator/test/distribution_SUITE_data/run.erl
@@ -30,16 +30,19 @@ from(H, [_ | T]) -> from(H, T);
 from(H, []) -> [].
 
 start() ->
-    net_kernel:start([fideridum,shortnames]),
-    {ok, Node} = slave:start(host(), heppel),
-    P = spawn(Node, a, b, []),
-    B1 = term_to_binary(P),
-    N1 = node(P),
-    ok = net_kernel:stop(),
-    N2 = node(P),
-    io:format("~w~n", [N1 == N2]),
+    Result = do_it(),
+
+    %% Do GCs and node_and_dist_references
+    %% in an attempt to crash the VM (without OTP-13076 fix)
+    lists:foreach(fun(P) -> erlang:garbage_collect(P) end,
+		  processes()),
+    erts_debug:set_internal_state(available_internal_state, true),
+    erts_debug:get_internal_state(node_and_dist_references),
+
+    io:format("~w~n", [Result]),
+
     if
-	N1 == N2 -> 
+	Result ->
 	    init:stop();
 	true ->
 	    %% Make sure that the io:format/2 output is really written
@@ -47,3 +50,29 @@ start() ->
 	    erlang:yield(),
 	    init:stop()
     end.
+
+
+do_it() ->
+    {ok, _} = net_kernel:start([fideridum,shortnames]),
+    {ok, Node} = slave:start(host(), heppel),
+    P = spawn(Node, net_kernel, stop, []),
+    B1 = term_to_binary(P),
+    N1 = node(P),
+    ok = net_kernel:stop(),
+    N2 = node(P),
+
+    %% OTP-13076
+    %% Restart distribution with same node name as previous remote node
+    %% Repeat to wrap around creation
+    Result = lists:foldl(fun(_, Acc) ->
+				 timer:sleep(2),  % give net_kernel:stop() time to take effect :-(
+				 {ok, _} = net_kernel:start([heppel,shortnames]),
+				 N3 = node(P),
+				 ok = net_kernel:stop(),
+				 N4 = node(P),
+				 Acc and (N3 =:= N1) and (N4 =:= N1)
+			 end,
+			 (N2 =:= N1),
+			 lists:seq(1,3)),
+
+    Result.
diff --git a/erts/emulator/test/driver_SUITE.erl b/erts/emulator/test/driver_SUITE.erl
index 4211c49848..ce55fe3c52 100644
--- a/erts/emulator/test/driver_SUITE.erl
+++ b/erts/emulator/test/driver_SUITE.erl
@@ -224,7 +224,7 @@ outputv_errors_1(Term) ->
     port_close(Port).
 
 build_iolist(N, Base) when N < 16 ->
-    case random:uniform(3) of
+    case rand:uniform(3) of
 	1 ->
 	    <<Bin:N/binary,_/binary>> = Base,
 	    Bin;
@@ -232,7 +232,7 @@ build_iolist(N, Base) when N < 16 ->
 	    lists:seq(1, N)
     end;
 build_iolist(N, Base) when N =< byte_size(Base) ->
-    case random:uniform(3) of
+    case rand:uniform(3) of
 	1 ->
 	    <<Bin:N/binary,_/binary>> = Base,
 	    Bin;
@@ -250,7 +250,7 @@ build_iolist(N, Base) when N =< byte_size(Base) ->
 	    end
     end;
 build_iolist(N0, Base) ->
-    Small = random:uniform(15),
+    Small = rand:uniform(15),
     Seq = lists:seq(1, Small),
     N = N0 - Small,
     case N rem 2 of
@@ -1877,10 +1877,7 @@ mseg_alloc_cached_segments() ->
     mseg_alloc_cached_segments(mseg_inst_info(0)).
 
 mseg_alloc_cached_segments(MsegAllocInfo) ->
-    MemName = case is_halfword_vm() of
-	true -> "high memory";
-	false -> "all memory"
-    end,
+    MemName = "all memory",
     ?line [{memkind,DrvMem}]
 	= lists:filter(fun(E) -> case E of
 				    {memkind, [{name, MemName} | _]} -> true;
@@ -1899,13 +1896,6 @@ mseg_inst_info(I) ->
 			  erlang:system_info({allocator,mseg_alloc})),
     Value.
 
-is_halfword_vm() ->
-    case {erlang:system_info({wordsize, internal}),
-	  erlang:system_info({wordsize, external})} of
-	{4, 8} -> true;
-	{WS, WS} -> false
-    end.
-
 driver_alloc_sbct() ->
     {_, _, _, As} = erlang:system_info(allocator),
     case lists:keysearch(driver_alloc, 1, As) of
@@ -2405,13 +2395,35 @@ z_test(Config) when is_list(Config) ->
 
 check_io_debug() ->
     get_stable_check_io_info(),
-    {NoErrorFds, NoUsedFds, NoDrvSelStructs, NoDrvEvStructs}
+    {NoErrorFds, NoUsedFds, NoDrvSelStructs, NoDrvEvStructs} = CheckIoDebug
 	= erts_debug:get_internal_state(check_io_debug),
+    HasGetHost = has_gethost(),
+    ct:log("check_io_debug: ~p~n"
+           "HasGetHost: ~p",[CheckIoDebug, HasGetHost]),
     0 = NoErrorFds,
-    NoUsedFds = NoDrvSelStructs,
+    if
+        NoUsedFds == NoDrvSelStructs ->
+            ok;
+        HasGetHost andalso (NoUsedFds == (NoDrvSelStructs - 1)) ->
+            %% If the inet_gethost port is alive, we may have
+            %% one extra used fd that is not selected on
+            ok
+    end,
     0 = NoDrvEvStructs,
     ok.
 
+has_gethost() ->
+    has_gethost(erlang:ports()).
+has_gethost([P|T]) ->
+    case erlang:port_info(P, name) of
+        {name,"inet_gethost"++_} ->
+            true;
+        _ ->
+            has_gethost(T)
+    end;
+has_gethost([]) ->
+    false.
+
 %flush_msgs() ->
 %    receive
 %	M ->
@@ -2512,14 +2524,7 @@ random_char() ->
     uniform(256) - 1.
 
 uniform(N) ->
-    case get(random_seed) of
-	undefined ->
-	    {X, Y, Z} = time(),
-	    random:seed(X, Y, Z);
-	_ ->
-	    ok
-    end,
-    random:uniform(N).
+    rand:uniform(N).
 
 erl_millisecs() ->
     erl_millisecs(erlang:monotonic_time()).
diff --git a/erts/emulator/test/emulator.spec.ose b/erts/emulator/test/emulator.spec.ose
deleted file mode 100644
index 9f494609d9..0000000000
--- a/erts/emulator/test/emulator.spec.ose
+++ /dev/null
@@ -1,2 +0,0 @@
-{topcase, {dir, "../emulator_test"}}.
-{skip, {obsolete_SUITE, "Not on ose"}}.
diff --git a/erts/emulator/test/erts_debug_SUITE.erl b/erts/emulator/test/erts_debug_SUITE.erl
index 3dd77eb920..bbba829501 100644
--- a/erts/emulator/test/erts_debug_SUITE.erl
+++ b/erts/emulator/test/erts_debug_SUITE.erl
@@ -24,13 +24,13 @@
 -export([all/0, suite/0,groups/0,init_per_suite/1, end_per_suite/1, 
 	 init_per_group/2,end_per_group/2,
 	 init_per_testcase/2,end_per_testcase/2,
-	 test_size/1,flat_size_big/1,df/1,
+	 test_size/1,flat_size_big/1,df/1,term_type/1,
 	 instructions/1]).
 
 suite() -> [{ct_hooks,[ts_install_cth]}].
 
 all() -> 
-    [test_size, flat_size_big, df, instructions].
+    [test_size, flat_size_big, df, instructions, term_type].
 
 groups() -> 
     [].
@@ -138,24 +138,90 @@ flat_size_big_1(Term, Size0, Limit) when Size0 < Limit ->
     end;
 flat_size_big_1(_, _, _) -> ok.
 
+
+term_type(Config) when is_list(Config) ->
+    Ts = [{fixnum, 1},
+          {fixnum, -1},
+          {bignum, 1 bsl 300},
+          {bignum, -(1 bsl 300)},
+          {hfloat, 0.0},
+          {hfloat, 0.0/-1},
+          {hfloat, 1.0/(1 bsl 302)},
+          {hfloat, 1.0*(1 bsl 302)},
+          {hfloat, -1.0/(1 bsl 302)},
+          {hfloat, -1.0*(1 bsl 302)},
+          {hfloat, 3.1416},
+          {hfloat, 1.0e18},
+          {hfloat, -3.1416},
+          {hfloat, -1.0e18},
+
+          {heap_binary, <<1,2,3>>},
+          {refc_binary, <<0:(8*80)>>},
+          {sub_binary,  <<5:7>>},
+
+          {flatmap, #{ a => 1}},
+          {hashmap, maps:from_list([{I,I}||I <- lists:seq(1,76)])},
+
+          {list, [1,2,3]},
+          {nil, []},
+          {tuple, {1,2,3}},
+          {tuple, {}},
+
+          {export, fun lists:sort/1},
+          {'fun', fun() -> ok end},
+          {pid, self()},
+          {atom, atom}],
+    lists:foreach(fun({E,Val}) ->
+                          R = erts_internal:term_type(Val),
+                          io:format("expecting term type ~w, got ~w (~p)~n", [E,R,Val]),
+                          E = R
+                  end, Ts),
+    ok.
+
+
 df(Config) when is_list(Config) ->
-    ?line P0 = pps(),
-    ?line PrivDir = ?config(priv_dir, Config),
-    ?line ok = file:set_cwd(PrivDir),
-    ?line erts_debug:df(?MODULE),
-    ?line Beam = filename:join(PrivDir, ?MODULE_STRING++".dis"),
-    ?line {ok,Bin} = file:read_file(Beam),
-    ?line ok = io:put_chars(binary_to_list(Bin)),
-    ?line ok = file:delete(Beam),
-    ?line true = (P0 == pps()),    
+    P0 = pps(),
+    PrivDir = ?config(priv_dir, Config),
+    ok = file:set_cwd(PrivDir),
+
+    AllLoaded = [M || {M,_} <- code:all_loaded()],
+    {Pid,Ref} = spawn_monitor(fun() -> df_smoke(AllLoaded) end),
+    receive
+	{'DOWN',Ref,process,Pid,Status} ->
+	    normal = Status
+    after 20*1000 ->
+	    %% Not finished (i.e. a slow computer). Stop now.
+	    Pid ! stop,
+	    receive
+		{'DOWN',Ref,process,Pid,Status} ->
+		    normal = Status,
+		    io:format("...")
+	    end
+    end,
+    io:nl(),
+    _ = [_ = file:delete(atom_to_list(M) ++ ".dis") ||
+	    M <- AllLoaded],
+
+    true = (P0 == pps()),
     ok.
 
+df_smoke([M|Ms]) ->
+    io:format("~p", [M]),
+    erts_debug:df(M),
+    receive
+	stop ->
+	    ok
+    after 0 ->
+	    df_smoke(Ms)
+    end;
+df_smoke([]) -> ok.
+
 pps() ->
     {erlang:ports()}.
 
 instructions(Config) when is_list(Config) ->
-    ?line Is = erts_debug:instructions(),
-    ?line _ = [list_to_atom(I) || I <- Is],
+    Is = erts_debug:instructions(),
+    _ = [list_to_atom(I) || I <- Is],
     ok.
 
 id(I) ->
diff --git a/erts/emulator/test/evil_SUITE.erl b/erts/emulator/test/evil_SUITE.erl
index 484d2a8bf5..d28e4d9596 100644
--- a/erts/emulator/test/evil_SUITE.erl
+++ b/erts/emulator/test/evil_SUITE.erl
@@ -382,10 +382,10 @@ my_appender_1(N, T0) ->
     my_appender_1(N-1, T).
     
 seed() ->
-    random:seed(3172, 9815, 20129).
+    rand:seed(exsplus, {3172,9815,20129}).
 
 rnd_term() ->
-    U0 = random:uniform(),
+    U0 = rand:uniform(),
     B = <<U0/float>>,
     {U0,U0 * 2.5 + 3.14,[U0*2.3,B]}.
 
diff --git a/erts/emulator/test/float_SUITE.erl b/erts/emulator/test/float_SUITE.erl
index c1a76b8af4..8826026f88 100644
--- a/erts/emulator/test/float_SUITE.erl
+++ b/erts/emulator/test/float_SUITE.erl
@@ -26,9 +26,11 @@
 	 init_per_group/2,end_per_group/2,
 	 init_per_testcase/2,end_per_testcase/2,
 	 fpe/1,fp_drv/1,fp_drv_thread/1,denormalized/1,match/1,
+         t_mul_add_ops/1,
 	 bad_float_unpack/1, write/1, cmp_zero/1, cmp_integer/1, cmp_bignum/1]).
 -export([otp_7178/1]).
 -export([hidden_inf/1]).
+-export([arith/1]).
 
 
 init_per_testcase(Func, Config) when is_atom(Func), is_list(Config) ->
@@ -45,6 +47,7 @@ all() ->
     [fpe, fp_drv, fp_drv_thread, otp_7178, denormalized,
      match, bad_float_unpack, write, {group, comparison}
      ,hidden_inf
+     ,arith, t_mul_add_ops
     ].
 
 groups() -> 
@@ -328,3 +331,124 @@ hidden_inf_1(A, B, Zero, Huge) ->
     {'EXIT',{badarith,_}} = (catch (B * (Huge + Huge))),
     {'EXIT',{badarith,_}} = (catch (B / (-Huge - Huge))),
     {'EXIT',{badarith,_}} = (catch (B * (-Huge - Huge))).
+
+%% Improve code coverage in our different arithmetic functions
+%% and make sure they yield consistent results.
+arith(_Config) ->
+    _TAG_IMMED1_SIZE = 4,
+
+    <<FLOAT_MAX/float>> = <<0:1, 16#7fe:11, -1:52>>,
+    <<FLOAT_MIN/float>> = <<0:1, 0:11, 1:52>>,
+    <<FloatNegZero/float>> = <<1:1, 0:11, 0:52>>,
+
+    WORD_BITS = erlang:system_info(wordsize) * 8,
+    SMALL_BITS = (WORD_BITS - _TAG_IMMED1_SIZE),
+    SMALL_MAX = (1 bsl (SMALL_BITS-1)) - 1,
+    SMALL_MIN = -(1 bsl (SMALL_BITS-1)),
+    BIG1_MAX = (1 bsl WORD_BITS) - 1,
+    BIG2_MAX = (1 bsl (WORD_BITS*2)) - 1,
+
+    fixnum = erts_internal:term_type(SMALL_MAX),
+    fixnum = erts_internal:term_type(SMALL_MIN),
+    bignum = erts_internal:term_type(SMALL_MAX + 1),
+    bignum = erts_internal:term_type(SMALL_MIN - 1),
+
+    L = [0, 0.0, FloatNegZero, 1, 1.0, 17, 17.0, 0.17,
+	 FLOAT_MIN, FLOAT_MAX,
+	 SMALL_MAX, SMALL_MAX+1,
+	 SMALL_MIN, SMALL_MIN-1,
+	 BIG1_MAX, BIG1_MAX+1,
+	 BIG2_MAX, BIG2_MAX+1,
+	 trunc(FLOAT_MAX), trunc(FLOAT_MAX)+1, trunc(FLOAT_MAX)*2,
+
+	 immed_badarg,
+	 "list badarg",
+	 {"boxed badarg"}
+	],
+
+    foreach_pair(fun(A,B) -> do_bin_ops(A,B) end, L).
+
+foreach_pair(F, L) ->
+    lists:foreach(
+      fun(A) -> lists:foreach(fun(B) -> F(A,B) end, L) end,
+      L).
+
+do_bin_ops(A, B) ->
+    Fun = fun(Op) ->
+		  Op(A,B),
+		  is_number(A) andalso Op(-A,B),
+		  is_number(B) andalso Op(A,-B),
+		  is_number(A) andalso is_number(B) andalso Op(-A,-B)
+	  end,
+    lists:foreach(Fun,
+		  [fun op_add/2, fun op_sub/2, fun op_mul/2, fun op_div/2]).
+
+op_add(A, B) ->
+    Info = [A,B],
+    R = unify(catch A + B, Info),
+    R = unify(my_apply(erlang,'+',[A,B]), Info),
+    case R of
+	_ when A + B =:= element(1,R) -> ok;
+	{{'EXIT',badarith}, Info} -> ok
+    end.
+
+op_sub(A, B) ->
+    Info = [A,B],
+    R = unify(catch A - B, Info),
+    R = unify(my_apply(erlang,'-',[A,B]), Info),
+    case R of
+	_ when A - B =:= element(1,R) -> ok;
+	{{'EXIT',badarith}, Info} -> ok
+    end.
+
+op_mul(A, B) ->
+    Info = [A,B],
+    R = unify(catch A * B, Info),
+    R = unify(my_apply(erlang,'*',[A,B]), Info),
+    case R of
+	_ when A * B =:= element(1,R) -> ok;
+	{{'EXIT',badarith}, Info} -> ok
+    end.
+
+op_div(A, B) ->
+    Info = [A,B],
+    R = unify(catch A / B, Info),
+    R = unify(my_apply(erlang,'/',[A,B]), Info),
+    case R of
+	_ when A / B =:= element(1,R) -> ok;
+	{{'EXIT',badarith}, Info} -> ok
+    end.
+
+my_apply(M, F, A) ->
+    catch apply(id(M), id(F), A).
+
+% Unify exceptions be removing stack traces.
+% and add argument info to make it easer to debug failed matches.
+unify({'EXIT',{Reason,_Stack}}, Info) ->
+    {{'EXIT', Reason}, Info};
+unify(Other, Info) ->
+    {Other, Info}.
+
+
+-define(epsilon, 1.0e-20).
+check_epsilon(R,Val) ->
+    if erlang:abs(R-Val) < ?epsilon -> ok;
+       true -> ?t:fail({R,Val})
+    end.
+
+t_mul_add_ops(Config) when is_list(Config) ->
+    check_epsilon(op_mul_add(1, 2.0, 1.0, 0.0), 1.0),
+    check_epsilon(op_mul_add(2, 2.0, 1.0, 0.0), 3.0),
+    check_epsilon(op_mul_add(3, 2.0, 1.0, 0.0), 7.0),
+    check_epsilon(op_mul_add(4, 2.0, 1.0, 0.0), 15.0),
+    check_epsilon(op_mul_add(5, 2.0, 1.0, 0.0), 31.0),
+    check_epsilon(op_mul_add(6, 2.0, 1.0, 0.0), 63.0),
+    check_epsilon(op_mul_add(6, 2.0, 1.3, 0.0), 81.9),
+    check_epsilon(op_mul_add(6, 2.03, 1.3, 0.0), 87.06260151458997),
+    ok.
+
+
+op_mul_add(0, _, _, R) -> R;
+op_mul_add(N, A, B, R) when is_float(A), is_float(B), is_float(R) ->
+    op_mul_add(N - 1, A, B, R * A + B).
+
diff --git a/erts/emulator/test/float_SUITE_data/fp_drv.c b/erts/emulator/test/float_SUITE_data/fp_drv.c
index 5919dd8e2f..a91d622040 100644
--- a/erts/emulator/test/float_SUITE_data/fp_drv.c
+++ b/erts/emulator/test/float_SUITE_data/fp_drv.c
@@ -18,6 +18,7 @@
  */
 
 #if defined(DEBUG) || 0
+#  include <stdio.h>
 #  define PRINTF(X) printf X
 #else
 #  define PRINTF(X)
diff --git a/erts/emulator/test/hash_SUITE.erl b/erts/emulator/test/hash_SUITE.erl
index 2ea49467b8..1b2acf48e1 100644
--- a/erts/emulator/test/hash_SUITE.erl
+++ b/erts/emulator/test/hash_SUITE.erl
@@ -223,11 +223,10 @@ basic_test() ->
 
 
 range_test() ->
-    random:seed(),
     F = fun(From,From,_FF) ->
 		ok;
 	   (From,To,FF) ->
-		R = random:uniform(16#FFFFFFFFFFFFFFFF),
+		R = rand:uniform(16#FFFFFFFFFFFFFFFF),
 		X = erlang:phash(R, From),
 		Y = erlang:phash(R, 16#100000000) - 1,
 		Z = (Y rem From) + 1,
@@ -265,14 +264,13 @@ spread_test(N) ->
 
 
 cmp_test(N) ->
-    % No need to save seed, the error indicates what number caused it.
-    random:seed(),
     do_cmp_hashes(N,8).
+
 do_cmp_hashes(0,_) ->
     ok;
 do_cmp_hashes(N,Steps) ->
-    R0 = random:uniform(1 bsl Steps - 1) + random:uniform(16#FFFFFFFF),
-    R = case random:uniform(2) of
+    R0 = rand:uniform(1 bsl Steps - 1) + rand:uniform(16#FFFFFFFF),
+    R = case rand:uniform(2) of
 	    1 ->
 		R0;
 	    _ ->
diff --git a/erts/emulator/test/map_SUITE.erl b/erts/emulator/test/map_SUITE.erl
index 1a89101916..5e9814be60 100644
--- a/erts/emulator/test/map_SUITE.erl
+++ b/erts/emulator/test/map_SUITE.erl
@@ -58,6 +58,7 @@
 	%% erlang
 	t_erlang_hash/1,
 	t_map_encode_decode/1,
+	t_gc_rare_map_overflow/1,
 
 	%% non specific BIF related
 	t_bif_build_and_check/1,
@@ -121,6 +122,7 @@ all() -> [
 
 	%% erlang
 	t_erlang_hash, t_map_encode_decode,
+	t_gc_rare_map_overflow,
 	t_map_size, t_is_map,
 
 	%% non specific BIF related
@@ -1509,11 +1511,8 @@ t_map_equal(Config) when is_list(Config) ->
 
 
 t_map_compare(Config) when is_list(Config) ->
-    Seed = {erlang:monotonic_time(),
-	    erlang:time_offset(),
-	    erlang:unique_integer()},
-    io:format("seed = ~p\n", [Seed]),
-    random:seed(Seed),
+    rand:seed(exsplus),
+    io:format("seed = ~p\n", [rand:export_seed()]),
     repeat(100, fun(_) -> float_int_compare() end, []),
     repeat(100, fun(_) -> recursive_compare() end, []),
     ok.
@@ -1531,7 +1530,7 @@ float_int_compare() ->
 
 numeric_keys(N) ->
     lists:foldl(fun(_,Acc) ->
-			Int = random:uniform(N*4) - N*2,
+			Int = rand:uniform(N*4) - N*2,
 			Float = float(Int),
 			[Int, Float, Float * 0.99, Float * 1.01 | Acc]
 		end,
@@ -1562,7 +1561,7 @@ do_compare([Gen1, Gen2]) ->
 
     %% Change one key from int to float (or vice versa) and check compare
     ML1 = maps:to_list(M1),
-    {K1,V1} = lists:nth(random:uniform(length(ML1)), ML1),
+    {K1,V1} = lists:nth(rand:uniform(length(ML1)), ML1),
     case K1 of
 	I when is_integer(I) ->
 	    case maps:find(float(I),M1) of
@@ -1653,9 +1652,9 @@ cmp_others(T1, T2, _) ->
 
 map_gen(Pairs, Size) ->
     {_,L} = lists:foldl(fun(_, {Keys, Acc}) ->
-				KI = random:uniform(size(Keys)),
+				KI = rand:uniform(size(Keys)),
 				K = element(KI,Keys),
-				KV = element(random:uniform(size(K)), K),
+				KV = element(rand:uniform(size(K)), K),
 				{erlang:delete_element(KI,Keys), [KV | Acc]}
 			end,
 			{Pairs, []},
@@ -1695,15 +1694,15 @@ term_gen_recursive(Leafs, Flags, Depth) ->
     MaxDepth = 10,
     Rnd = case {Flags, Depth} of
 	      {_, MaxDepth} -> % Only leafs
-		  random:uniform(size(Leafs)) + 3;
+		  rand:uniform(size(Leafs)) + 3;
 	      {0, 0} ->        % Only containers
-		  random:uniform(3);
+		  rand:uniform(3);
 	      {0,_} ->         % Anything
-		  random:uniform(size(Leafs)+3)
+		  rand:uniform(size(Leafs)+3)
 	  end,
     case Rnd of
 	1 -> % Make map
-	    Size = random:uniform(size(Leafs)),
+	    Size = rand:uniform(size(Leafs)),
 	    lists:foldl(fun(_, {Acc1,Acc2}) ->
 				{K1,K2} = term_gen_recursive(Leafs, Flags,
 							     Depth+1),
@@ -1718,7 +1717,7 @@ term_gen_recursive(Leafs, Flags, Depth) ->
 	    {Cdr1,Cdr2} = term_gen_recursive(Leafs, Flags, Depth+1),
 	    {[Car1 | Cdr1], [Car2 | Cdr2]};
 	3 -> % Make tuple
-	    Size = random:uniform(size(Leafs)),
+	    Size = rand:uniform(size(Leafs)),
 	    L = lists:map(fun(_) -> term_gen_recursive(Leafs, Flags, Depth+1) end,
 			  lists:seq(1,Size)),
 	    {L1, L2} = lists:unzip(L),
@@ -1727,7 +1726,7 @@ term_gen_recursive(Leafs, Flags, Depth) ->
 	N -> % Make leaf
 	    case element(N-3, Leafs) of
 		I when is_integer(I) ->
-		    case random:uniform(4) of
+		    case rand:uniform(4) of
 			1 -> {I, float(I)};
 			2 -> {float(I), I};
 			_ -> {I,I}
@@ -2181,7 +2180,9 @@ t_map_encode_decode(Config) when is_list(Config) ->
 	{<<>>, sc9}, {3.14158, sc10},
 	{[3.14158], sc11}, {more_atoms, sc12},
 	{{more_tuples}, sc13}, {self(), sc14},
-	{{},{}},{[],[]}
+	{{},{}},{[],[]},
+        {map_s, #{a=>a, 2=>b, 3=>c}},
+        {map_l, maps:from_list([{I,I}||I <- lists:seq(1,74)])}
     ],
     ok = map_encode_decode_and_match(Pairs,[],#{}),
 
@@ -2245,9 +2246,30 @@ t_map_encode_decode(Config) when is_list(Config) ->
 
     %% bad size (too small) .. should fail just truncate it .. weird.
     %% possibly change external format so truncated will be #{a:=1}
-    #{ a:=b } =
-	erlang:binary_to_term(<<131,116,0,0,0,1,100,0,1,97,100,0,1,98,97,1,97,1>>),
-
+    #{ a:=b } = erlang:binary_to_term(<<131,116,0,0,0,1,100,0,1,97,100,0,1,98,97,1,97,1>>),
+
+    %% specific fannerl (opensource app) binary_to_term error in 18.1
+
+    #{bias := {1,1,0},
+      bit_fail := 0,
+      connections := #{{2,9}   := _,
+                       {8,14}  := _,
+                       {2,12}  := _,
+                       {5,7}   := _,
+                       {11,16} := _,
+                       {11,15} := _},
+      layers := {5,7,3},
+      network_type := fann_nettype_layer,
+      num_input := 5,
+      num_layers := 3,
+      num_output := 3,
+      rprop_delta_max := _,
+      rprop_delta_min := _,
+      total_connections := 66,
+      total_neurons := 17,
+      train_error_function := fann_errorfunc_tanh,
+      train_stop_function := fann_stopfunc_mse,
+      training_algorithm := fann_train_rprop} = erlang:binary_to_term(fannerl()),
     ok.
 
 map_encode_decode_and_match([{K,V}|Pairs], EncodedPairs, M0) ->
@@ -2391,6 +2413,9 @@ check_keys_exist([K|Ks],M) ->
     check_keys_exist(Ks,M).
 
 t_bif_merge_and_check(Config) when is_list(Config) ->
+
+    io:format("rand:export_seed() -> ~p\n",[rand:export_seed()]),
+
     %% simple disjunct ones
     %% make sure all keys are unique
     Kss = [[a,b,c,d],
@@ -2438,8 +2463,49 @@ t_bif_merge_and_check(Config) when is_list(Config) ->
     M41 = maps:merge(M4,M1),
     ok  = check_key_values(KVs1 ++ [{d,5}] ++ KVs, M41),
 
+    [begin Ma = random_map(SzA, a),
+	   Mb = random_map(SzB, b),
+	   ok = merge_maps(Ma, Mb)
+     end || SzA <- [3,10,20,100,200,1000], SzB <- [3,10,20,100,200,1000]],
+
+    ok.
+
+% Generate random map with an average of Sz number of pairs: K -> {V,K}
+random_map(Sz, V) ->
+    random_map_insert(#{}, 0, V, Sz*2).
+
+random_map_insert(M0, K0, _, Sz) when K0 > Sz ->
+    M0;
+random_map_insert(M0, K0, V, Sz) ->
+    Key = K0 + rand:uniform(3),
+    random_map_insert(M0#{Key => {V,Key}}, Key, V, Sz).
+
+
+merge_maps(A, B) ->
+    AB = maps:merge(A, B),
+    %%io:format("A=~p\nB=~p\n",[A,B]),
+    maps_foreach(fun(K,VB) -> VB = maps:get(K, AB)
+		 end, B),
+    maps_foreach(fun(K,VA) ->
+			 case {maps:get(K, AB),maps:find(K, B)} of
+			     {VA, error} -> ok;
+			     {VB, {ok, VB}} -> ok
+			 end
+		 end, A),
+
+    maps_foreach(fun(K,V) ->
+			 case {maps:find(K, A),maps:find(K, B)} of
+			     {{ok, V}, error} -> ok;
+			     {error, {ok, V}} -> ok;
+			     {{ok,_}, {ok, V}} -> ok
+			 end
+		 end, AB),
     ok.
 
+maps_foreach(Fun, Map) ->
+    maps:fold(fun(K,V,_) -> Fun(K,V) end, void, Map).
+
+
 check_key_values([],_) -> ok;
 check_key_values([{K,V}|KVs],M) ->
     V = maps:get(K,M),
@@ -2529,7 +2595,7 @@ hashmap_balance(KeyFun) ->
     F = fun(I, {M0,Max0}) ->
 		Key = KeyFun(I),
 		M1 = M0#{Key => Key},
-		Max1 = case erts_internal:map_type(M1) of
+		Max1 = case erts_internal:term_type(M1) of
 			   hashmap ->
 			       Nodes = hashmap_nodes(M1),
 			       Avg = maps:size(M1) * 0.4,
@@ -2922,3 +2988,189 @@ do_badmap_17(Config) ->
 
 %% Use this function to avoid compile-time evaluation of an expression.
 id(I) -> I.
+
+
+%% OTP-13146
+%% Provoke major GC with a lot of "fat" maps on external format in msg queue
+%% causing heap fragments to be allocated.
+t_gc_rare_map_overflow(Config) ->
+    Pa = filename:dirname(code:which(?MODULE)),
+    {ok, Node} = test_server:start_node(gc_rare_map_overflow, slave, [{args, "-pa \""++Pa++"\""}]),
+    erts_debug:set_internal_state(available_internal_state, true),
+    try
+	Echo = spawn_link(Node, fun Loop() -> receive {From,Msg} -> From ! Msg
+					      end,
+					      Loop()
+				end),
+	FatMap = fatmap(34),
+	false = (flatmap =:= erts_internal:term_type(FatMap)),
+
+	t_gc_rare_map_overflow_do(Echo, FatMap, fun() -> erlang:garbage_collect() end),
+
+	%% Repeat test for minor gc:
+	t_gc_rare_map_overflow_do(Echo, FatMap, fun() -> minor_collect() end),
+
+	unlink(Echo),
+
+	%% Test fatmap in exit signal
+	Exiter = spawn_link(Node, fun Loop() -> receive {From,Msg} ->
+							"not_a_map" = Msg  % badmatch!
+						end,
+						Loop()
+				  end),
+	process_flag(trap_exit, true),
+	Exiter ! {self(), FatMap},
+	{'EXIT', Exiter, {{badmatch,FatMap}, _}} = receive M -> M end,
+	ok
+
+    after
+	process_flag(trap_exit, false),
+	erts_debug:set_internal_state(available_internal_state, false),
+	test_server:stop_node(Node)
+    end.
+
+t_gc_rare_map_overflow_do(Echo, FatMap, GcFun) ->
+    Master = self(),
+    true = receive M -> false after 0 -> true end,   % assert empty msg queue
+    Echo ! {Master, token},
+    repeat(1000, fun(_) -> Echo ! {Master, FatMap} end, void),
+
+    timer:sleep(100),                % Wait for maps to arrive in our msg queue
+    token = receive Tok -> Tok end,  % and provoke move from outer to inner msg queue
+
+    %% Do GC that will "overflow" and create heap frags due to all the fat maps
+    GcFun(),
+
+    %% Now check that all maps in msg queueu are intact
+    %% Will crash emulator in OTP-18.1
+    repeat(1000, fun(_) -> FatMap = receive FM -> FM end end, void),
+    ok.
+
+minor_collect() ->
+    Before = minor_gcs(),
+    erts_debug:set_internal_state(force_gc, self()),
+    erlang:yield(),
+    After = minor_gcs(),
+    io:format("minor_gcs: ~p -> ~p\n", [Before, After]).
+
+minor_gcs() ->
+    {garbage_collection, Info} = process_info(self(), garbage_collection),
+    {minor_gcs, GCS} = lists:keyfind(minor_gcs, 1, Info),
+    GCS.
+
+%% Generate a map with N (or N+1) keys that has an abnormal heap demand.
+%% Done by finding keys that collide in the first 32-bit hash.
+fatmap(N) ->
+    %%erts_debug:set_internal_state(available_internal_state, true),
+    Table = ets:new(void, [bag, private]),
+
+    Seed0 = rand:seed_s(exsplus, {4711, 3141592, 2718281}),
+    Seed1 = fatmap_populate(Table, Seed0, (1 bsl 16)),
+    Keys = fatmap_generate(Table, Seed1, N, []),
+    ets:delete(Table),
+    maps:from_list([{K,K} || K <- Keys]).
+
+fatmap_populate(_, Seed, 0) -> Seed;
+fatmap_populate(Table, Seed, N) ->
+    {I, NextSeed} = rand:uniform_s(1 bsl 48, Seed),
+    Hash = internal_hash(I),
+    ets:insert(Table, [{Hash, I}]),
+    fatmap_populate(Table, NextSeed, N-1).
+
+
+fatmap_generate(_, _, N, Acc) when N =< 0 ->
+    Acc;
+fatmap_generate(Table, Seed, N0, Acc0) ->    
+    {I, NextSeed} = rand:uniform_s(1 bsl 48, Seed),
+    Hash = internal_hash(I),
+    case ets:member(Table, Hash) of
+	true ->
+	    NewKeys = [I | ets:lookup_element(Table, Hash, 2)],
+	    Acc1 = lists:usort(Acc0 ++ NewKeys),
+	    N1 = N0 - (length(Acc1) - length(Acc0)),
+	    fatmap_generate(Table, NextSeed, N1, Acc1);
+	false ->
+	    fatmap_generate(Table, NextSeed, N0, Acc0)
+    end.
+
+internal_hash(Term) ->
+    erts_debug:get_internal_state({internal_hash, Term}).
+
+
+%% map external_format (fannerl).
+fannerl() ->
+    <<131,116,0,0,0,28,100,0,13,108,101,97,114,110,105,110,103,95,114,
+      97,116,101,70,63,230,102,102,96,0,0,0,100,0,17,108,101,97,114,110,105,110,
+      103,95,109,111,109,101,110,116,117,109,70,0,0,0,0,0,0,0,0,100,0,
+      18,116,114,97,105,110,105,110,103,95,97,108,103,111,114,105,116,104,109,100,0,
+      16,102,97,110,110,95,116,114,97,105,110,95,114,112,114,111,112,
+      100,0,17,109,101,97,110,95,115,113,117,97,114,101,95,101,114,114,111,114,70,
+      0,0,0,0,0,0,0,0,100,0,8,98,105,116,95,102,97,105,108,97,0,100,0,20,
+      116,114,97,105,110,95,101,114,114,111,114,95,102,117,110,99,116,105,111,
+      110,100,0,19,102,97,110,110,95,101,114,114,111,114,102,117,110,99,
+      95,116,97,110,104,100,0,9,110,117,109,95,105,110,112,117,116,97,5,100,0,10,110,
+      117,109,95,111,117,116,112,117,116,97,3,100,0,13,116,111,116,97,108,
+      95,110,101,117,114,111,110,115,97,17,100,0,17,116,111,116,97,108,95,99,111,110,
+      110,101,99,116,105,111,110,115,97,66,100,0,12,110,101,116,119,111,114,107,
+      95,116,121,112,101,100,0,18,102,97,110,110,95,110,101,116,116,121,112,101,
+      95,108,97,121,101,114,100,0,15,99,111,110,110,101,99,116,105,111,110,95,
+      114,97,116,101,70,63,240,0,0,0,0,0,0,100,0,10,110,117,109,95,108,97,121,101,
+      114,115,97,3,100,0,19,116,114,97,105,110,95,115,116,111,112,95,102,117,110,
+      99,116,105,111,110,100,0,17,102,97,110,110,95,115,116,111,112,102,117,110,
+      99,95,109,115,101,100,0,15,113,117,105,99,107,112,114,111,112,95,100,101,99,
+      97,121,70,191,26,54,226,224,0,0,0,100,0,12,113,117,105,99,107,112,114,
+      111,112,95,109,117,70,63,252,0,0,0,0,0,0,100,0,21,114,112,114,111,112,95,105,
+      110,99,114,101,97,115,101,95,102,97,99,116,111,114,70,63,243,51,51,
+      64,0,0,0,100,0,21,114,112,114,111,112,95,100,101,99,114,101,97,115,101,
+      95,102,97,99,116,111,114,70,63,224,0,0,0,0,0,0,100,0,15,114,112,114,111,112,
+      95,100,101,108,116,97,95,109,105,110,70,0,0,0,0,0,0,0,0,100,0,15,114,112,114,
+      111,112,95,100,101,108,116,97,95,109,97,120,70,64,73,0,0,0,0,0,0,100,0,
+      16,114,112,114,111,112,95,100,101,108,116,97,95,122,101,114,111,70,63,185,153,
+      153,160,0,0,0,100,0,26,115,97,114,112,114,111,112,95,119,101,105,103,
+      104,116,95,100,101,99,97,121,95,115,104,105,102,116,70,192,26,147,116,192,0,0,0,
+      100,0,35,115,97,114,112,114,111,112,95,115,116,101,112,95,101,114,
+      114,111,114,95,116,104,114,101,115,104,111,108,100,95,102,97,99,116,111,114,70,
+      63,185,153,153,160,0,0,0,100,0,24,115,97,114,112,114,111,112,95,115,
+      116,101,112,95,101,114,114,111,114,95,115,104,105,102,116,70,63,246,40,245,
+      192,0,0,0,100,0,19,115,97,114,112,114,111,112,95,116,101,109,112,101,114,
+      97,116,117,114,101,70,63,142,184,81,224,0,0,0,100,0,6,108,97,121,101,114,115,
+      104,3,97,5,97,7,97,3,100,0,4,98,105,97,115,104,3,97,1,97,1,97,0,100,0,11,
+      99,111,110,110,101,99,116,105,111,110,115,116,0,0,0,66,104,2,97,0,97,6,70,
+      191,179,51,44,64,0,0,0,104,2,97,1,97,6,70,63,178,130,90,32,0,0,0,104,2,97,2,
+      97,6,70,63,82,90,88,0,0,0,0,104,2,97,3,97,6,70,63,162,91,63,192,0,0,0,104,2,
+      97,4,97,6,70,191,151,70,169,0,0,0,0,104,2,97,5,97,6,70,191,117,52,222,0,0,0,
+      0,104,2,97,0,97,7,70,63,152,240,139,0,0,0,0,104,2,97,1,97,7,70,191,166,31,
+      187,160,0,0,0,104,2,97,2,97,7,70,191,150,70,63,0,0,0,0,104,2,97,3,97,7,70,
+      63,152,181,126,128,0,0,0,104,2,97,4,97,7,70,63,151,187,162,128,0,0,0,104,2,
+      97,5,97,7,70,191,143,161,101,0,0,0,0,104,2,97,0,97,8,70,191,153,102,36,128,0,
+      0,0,104,2,97,1,97,8,70,63,160,139,250,64,0,0,0,104,2,97,2,97,8,70,63,164,62,
+      196,64,0,0,0,104,2,97,3,97,8,70,191,178,78,209,192,0,0,0,104,2,97,4,97,8,70,
+      191,185,19,76,224,0,0,0,104,2,97,5,97,8,70,63,183,142,196,96,0,0,0,104,2,97,0,
+      97,9,70,63,150,104,248,0,0,0,0,104,2,97,1,97,9,70,191,164,4,100,224,0,0,0,
+      104,2,97,2,97,9,70,191,169,42,42,224,0,0,0,104,2,97,3,97,9,70,63,145,54,78,128,0,
+      0,0,104,2,97,4,97,9,70,63,126,243,134,0,0,0,0,104,2,97,5,97,9,70,63,177,
+      203,25,96,0,0,0,104,2,97,0,97,10,70,63,172,104,47,64,0,0,0,104,2,97,1,97,10,
+      70,63,161,242,193,64,0,0,0,104,2,97,2,97,10,70,63,175,208,241,192,0,0,0,104,2,
+      97,3,97,10,70,191,129,202,161,0,0,0,0,104,2,97,4,97,10,70,63,178,151,55,32,0,0,0,
+      104,2,97,5,97,10,70,63,137,155,94,0,0,0,0,104,2,97,0,97,11,70,191,179,
+      106,160,0,0,0,0,104,2,97,1,97,11,70,63,184,253,164,96,0,0,0,104,2,97,2,97,11,
+      70,191,143,30,157,0,0,0,0,104,2,97,3,97,11,70,63,153,225,140,128,0,0,0,104,
+      2,97,4,97,11,70,63,161,35,85,192,0,0,0,104,2,97,5,97,11,70,63,175,200,55,192,
+      0,0,0,104,2,97,0,97,12,70,191,180,116,132,96,0,0,0,104,2,97,1,97,12,70,191,
+      165,151,152,0,0,0,0,104,2,97,2,97,12,70,191,180,197,91,160,0,0,0,104,2,97,3,97,12,
+      70,191,91,30,160,0,0,0,0,104,2,97,4,97,12,70,63,180,251,45,32,0,0,0,
+      104,2,97,5,97,12,70,63,165,134,77,64,0,0,0,104,2,97,6,97,14,70,63,181,56,242,96,
+      0,0,0,104,2,97,7,97,14,70,191,165,239,234,224,0,0,0,104,2,97,8,97,14,
+      70,191,154,65,216,128,0,0,0,104,2,97,9,97,14,70,63,150,250,236,0,0,0,0,104,2,97,
+      10,97,14,70,191,141,105,108,0,0,0,0,104,2,97,11,97,14,70,191,152,40,
+      165,0,0,0,0,104,2,97,12,97,14,70,63,141,159,46,0,0,0,0,104,2,97,13,97,14,70,
+      191,183,172,137,32,0,0,0,104,2,97,6,97,15,70,63,163,26,123,192,0,0,0,104,
+      2,97,7,97,15,70,63,176,184,106,32,0,0,0,104,2,97,8,97,15,70,63,152,234,144,
+      0,0,0,0,104,2,97,9,97,15,70,191,172,58,70,160,0,0,0,104,2,97,10,97,15,70,
+      63,161,211,211,192,0,0,0,104,2,97,11,97,15,70,191,148,171,120,128,0,0,0,104,
+      2,97,12,97,15,70,63,180,117,214,224,0,0,0,104,2,97,13,97,15,70,191,104,
+      230,216,0,0,0,0,104,2,97,6,97,16,70,63,178,53,103,96,0,0,0,104,2,97,7,97,16,
+      70,63,170,230,232,64,0,0,0,104,2,97,8,97,16,70,191,183,45,100,192,0,0,0,
+      104,2,97,9,97,16,70,63,184,100,97,32,0,0,0,104,2,97,10,97,16,70,63,169,174,
+      254,64,0,0,0,104,2,97,11,97,16,70,191,119,121,234,0,0,0,0,104,2,97,12,97,
+      16,70,63,149,12,170,128,0,0,0,104,2,97,13,97,16,70,191,144,193,191,0,0,0,0>>.
diff --git a/erts/emulator/test/match_spec_SUITE.erl b/erts/emulator/test/match_spec_SUITE.erl
index c6cc414bba..3f986ca2ed 100644
--- a/erts/emulator/test/match_spec_SUITE.erl
+++ b/erts/emulator/test/match_spec_SUITE.erl
@@ -170,12 +170,11 @@ test_1(Config) when is_list(Config) ->
 	     [{['$1','$1'],[{is_atom, '$1'}],[kakalorum]}],
 	     [{call, {?MODULE, f2, [a, a]}}]),
 
-%    case tr0(fun() -> ?MODULE:f2(a, a) end,
-%	     {?MODULE, f2, 2},
-%	     [{['$1','$1'],[{is_atom, '$1'}],[{message, {process_dump}}]}]) of
-%	[{trace, _, call, {?MODULE, f2, [a, a]}, Bin}] ->
-%	    erlang:display(binary_to_list(Bin))
-%    end,
+    %% Verify that 'process_dump' can handle a matchstate on the stack.
+    tr(fun() -> fbinmatch(<<0>>, 0) end,
+       {?MODULE, f1, 1},
+       [{['_'],[],[{message, {process_dump}}]}],
+       [fun({trace, _, call, {?MODULE, f1, [0]}, _Bin}) -> true end]),
 
 % Error cases
     ?line errchk([{['$1','$1'],[{is_atom, '$1'}],[{banka, kanin}]}]),
@@ -999,14 +998,14 @@ tr(Fun, MFA, TraceFlags, Pat, PatFlags, Expected0) ->
 	       erlang:trace(P, true, TraceFlags),
 	       erlang:trace_pattern(MFA, Pat, PatFlags),
 	       lists:map(
-		 fun(X) -> 
-			 list_to_tuple([trace, P | tuple_to_list(X)])
+		 fun(X) when is_function(X,1) -> X;
+		    (X) -> list_to_tuple([trace, P | tuple_to_list(X)])
 		 end,
 		 Expected0)
        end).
 
 tr(RunFun, ControlFun) ->
-    P = spawn(?MODULE, runner, [self(), RunFun]),
+    P = spawn_link(?MODULE, runner, [self(), RunFun]),
     collect(P, ControlFun(P)).
 
 collect(P, TMs) ->
@@ -1025,18 +1024,33 @@ collect([]) ->
 collect([TM | TMs]) ->
     ?t:format(        "Expecting:      ~p~n", [TM]),
     receive
-	M ->
-	    case if element(1, M) == trace_ts ->
-			 list_to_tuple(lists:reverse(
-					 tl(lists:reverse(tuple_to_list(M)))));
-		    true -> M
-		 end of
-		TM ->
-		    ?t:format("Got:            ~p~n", [M]),
-		    collect(TMs);
-		_ ->
-		    ?t:format("Got unexpected: ~p~n", [M]),
-		    flush({got_unexpected,M})
+	M0 ->
+	    M = case element(1, M0) of
+		    trace_ts ->
+			list_to_tuple(lists:reverse(
+					tl(lists:reverse(tuple_to_list(M0)))));
+		    _ -> M0
+		end,
+	    case is_function(TM,1) of
+		true ->
+		    case (catch TM(M)) of
+			true ->
+			    ?t:format("Got:            ~p~n", [M]),
+			    collect(TMs);
+			_ ->
+			    ?t:format("Got unexpected: ~p~n", [M]),
+			    flush({got_unexpected,M})
+		    end;
+
+		false ->
+		    case M of
+			TM ->
+			    ?t:format("Got:            ~p~n", [M]),
+			    collect(TMs);
+			_ ->
+			    ?t:format("Got unexpected: ~p~n", [M]),
+			    flush({got_unexpected,M})
+		    end
 	    end
     end.
 
@@ -1116,6 +1130,10 @@ fn(X, Y) ->
 fn(X, Y, Z) ->
     [X, Y, Z].
 
+fbinmatch(<<Int, Rest/binary>>, Acc) ->
+    fbinmatch(Rest, [?MODULE:f1(Int) | Acc]);
+fbinmatch(<<>>, Acc) -> Acc.
+
 id(X) ->
     X.
 
diff --git a/erts/emulator/test/message_queue_data_SUITE.erl b/erts/emulator/test/message_queue_data_SUITE.erl
new file mode 100644
index 0000000000..11481409aa
--- /dev/null
+++ b/erts/emulator/test/message_queue_data_SUITE.erl
@@ -0,0 +1,239 @@
+%%
+%% %CopyrightBegin%
+%% 
+%% Copyright Ericsson AB 2014. All Rights Reserved.
+%% 
+%% Licensed under the Apache License, Version 2.0 (the "License");
+%% you may not use this file except in compliance with the License.
+%% You may obtain a copy of the License at
+%%
+%%     http://www.apache.org/licenses/LICENSE-2.0
+%%
+%% Unless required by applicable law or agreed to in writing, software
+%% distributed under the License is distributed on an "AS IS" BASIS,
+%% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+%% See the License for the specific language governing permissions and
+%% limitations under the License.
+%% 
+%% %CopyrightEnd%
+%%
+
+-module(message_queue_data_SUITE).
+
+-export([all/0, suite/0,groups/0,init_per_suite/1, end_per_suite/1, 
+	 init_per_group/2,end_per_group/2,
+	 init_per_testcase/2,end_per_testcase/2]).
+-export([basic/1, process_info_messages/1]).
+
+-export([basic_test/1]).
+
+-include_lib("test_server/include/test_server.hrl").
+
+init_per_testcase(Case, Config) ->
+    ?line Dog=test_server:timetrap(test_server:minutes(2)),
+    [{watchdog, Dog}, {testcase, Case}|Config].
+
+end_per_testcase(_, Config) ->
+    Dog=?config(watchdog, Config),
+    test_server:timetrap_cancel(Dog),
+    ok.
+
+suite() -> [{ct_hooks,[ts_install_cth]}].
+
+all() -> 
+    [basic, process_info_messages].
+
+groups() -> 
+    [].
+
+init_per_suite(Config) ->
+%%    erts_debug:set_internal_state(available_internal_state, true),
+    Config.
+
+end_per_suite(_Config) ->
+%%    erts_debug:set_internal_state(available_internal_state, false),
+    ok.
+
+init_per_group(_GroupName, Config) ->
+    Config.
+
+end_per_group(_GroupName, Config) ->
+    Config.
+
+%%
+%%
+%% Test cases
+%%
+%%
+
+basic(Config) when is_list(Config) ->
+
+    basic_test(erlang:system_info(message_queue_data)),
+
+    {ok, Node1} = start_node(Config, "+xmqd off_heap"),
+    ok = rpc:call(Node1, ?MODULE, basic_test, [off_heap]),
+    stop_node(Node1),
+
+    {ok, Node2} = start_node(Config, "+xmqd on_heap"),
+    ok = rpc:call(Node2, ?MODULE, basic_test, [on_heap]),
+    stop_node(Node2),
+
+    {ok, Node3} = start_node(Config, "+xmqd mixed"),
+    ok = rpc:call(Node3, ?MODULE, basic_test, [mixed]),
+    stop_node(Node3),
+
+    ok.
+
+is_valid_mqd_value(off_heap) ->
+    true;
+is_valid_mqd_value(on_heap) ->
+    true;
+is_valid_mqd_value(mixed) ->
+    true;
+is_valid_mqd_value(_) ->
+    false.
+
+
+basic_test(Default) ->
+
+    Default = erlang:system_info(message_queue_data),
+    true = is_valid_mqd_value(Default),
+
+    {message_queue_data, Default} = process_info(self(), message_queue_data),
+    Default = process_flag(message_queue_data, off_heap),
+    {message_queue_data, off_heap} = process_info(self(), message_queue_data),
+    off_heap = process_flag(message_queue_data, on_heap),
+    {message_queue_data, on_heap} = process_info(self(), message_queue_data),
+    on_heap = process_flag(message_queue_data, mixed),
+    {message_queue_data, mixed} = process_info(self(), message_queue_data),
+    mixed = process_flag(message_queue_data, Default),
+    {'EXIT', _} = (catch process_flag(message_queue_data, blupp)),
+
+    P1 = spawn_opt(fun () -> receive after infinity -> ok end end,
+		   [link]),
+    {message_queue_data, Default} = process_info(P1, message_queue_data),
+    unlink(P1),
+    exit(P1, bye),
+
+    P2 = spawn_opt(fun () -> receive after infinity -> ok end end,
+		   [link, {message_queue_data, off_heap}]),
+    {message_queue_data, off_heap} = process_info(P2, message_queue_data),
+    unlink(P2),
+    exit(P2, bye),
+
+    P3 = spawn_opt(fun () -> receive after infinity -> ok end end,
+		   [link, {message_queue_data, on_heap}]),
+    {message_queue_data, on_heap} = process_info(P3, message_queue_data),
+    unlink(P3),
+    exit(P3, bye),
+
+    P4 = spawn_opt(fun () -> receive after infinity -> ok end end,
+		   [link, {message_queue_data, mixed}]),
+    {message_queue_data, mixed} = process_info(P4, message_queue_data),
+    unlink(P4),
+    exit(P4, bye),
+
+    {'EXIT', _} = (catch spawn_opt(fun () -> receive after infinity -> ok end end,
+		   [link, {message_queue_data, blapp}])),
+
+    ok.
+
+process_info_messages(Config) when is_list(Config) ->
+    Tester = self(),
+    P1 = spawn_opt(fun () ->
+			   receive after 500 -> ok end,
+			   mixed = process_flag(message_queue_data, off_heap),
+			   Tester ! first,
+			   receive after 500 -> ok end,
+			   off_heap = process_flag(message_queue_data, on_heap),
+			   Tester ! second,
+			   receive after 500 -> ok end,
+			   on_heap = process_flag(message_queue_data, mixed),
+			   Tester ! third,
+			   receive after 500 -> ok end,
+			   mixed = process_flag(message_queue_data, off_heap),
+			   Tester ! fourth,
+			    
+			   receive after infinity -> ok end
+		   end,
+		   [link, {message_queue_data, mixed}]),
+    
+    P1 ! "A",
+    receive first -> ok end,
+    P1 ! "B",
+    receive second -> ok end,
+    P1 ! "C",
+    receive third -> ok end,
+    P1 ! "D",
+    receive fourth -> ok end,
+    P1 ! "E",
+
+    {messages, ["A", "B", "C", "D", "E"]} = process_info(P1, messages),
+
+    P2 = spawn_opt(fun () ->
+			   receive after 500 -> ok end,
+			   mixed = process_flag(message_queue_data, off_heap),
+			   Tester ! first,
+			   receive after 500 -> ok end,
+			   off_heap = process_flag(message_queue_data, on_heap),
+			   Tester ! second,
+			   receive after 500 -> ok end,
+			   on_heap = process_flag(message_queue_data, mixed),
+			   Tester ! third,
+			   receive after 500 -> ok end,
+			   mixed = process_flag(message_queue_data, off_heap),
+			   Tester ! fourth,
+			   receive after 500 -> ok end,
+
+			   Tester ! process_info(self(), messages),
+
+			   receive M1 -> M1 = "A" end,
+			   receive M2 -> M2 = "B" end,
+			   receive M3 -> M3 = "C" end,
+			   receive M4 -> M4 = "D" end,
+			   receive M5 -> M5 = "E" end,
+
+			   Tester ! self()
+		   end,
+		   [link, {message_queue_data, mixed}]),
+
+    P2 ! "A",
+    receive first -> ok end,
+    P2 ! "B",
+    receive second -> ok end,
+    P2 ! "C",
+    receive third -> ok end,
+    P2 ! "D",
+    receive fourth -> ok end,
+    P2 ! "E",
+
+    receive
+	Msg ->
+	    {messages, ["A", "B", "C", "D", "E"]} = Msg
+    end,
+    
+    receive P2 -> ok end,
+
+    ok.
+
+%%
+%%
+%% helpers
+%%
+%%
+
+start_node(Config) ->
+    start_node(Config, []).
+start_node(Config, Opts) when is_list(Config), is_list(Opts) ->
+    Pa = filename:dirname(code:which(?MODULE)),
+    Name = list_to_atom(atom_to_list(?MODULE)
+			++ "-"
+			++ atom_to_list(?config(testcase, Config))
+			++ "-"
+			++ integer_to_list(erlang:system_time(seconds))
+			++ "-"
+			++ integer_to_list(erlang:unique_integer([positive]))),
+    ?t:start_node(Name, slave, [{args, Opts++" -pa "++Pa}]).
+
+stop_node(Node) ->
+    ?t:stop_node(Node).
diff --git a/erts/emulator/test/nif_SUITE.erl b/erts/emulator/test/nif_SUITE.erl
index af2b955184..56b36d2626 100644
--- a/erts/emulator/test/nif_SUITE.erl
+++ b/erts/emulator/test/nif_SUITE.erl
@@ -1192,11 +1192,8 @@ send3(Config) when is_list(Config) ->
     %% Let a number of processes send random message blobs between each other
     %% using enif_send. Kill and spawn new ones randomly to keep a ~constant
     %% number of workers running.
-    Seed = {erlang:monotonic_time(),
-	    erlang:time_offset(),
-	    erlang:unique_integer()},
-    io:format("seed: ~p\n",[Seed]), 
-    random:seed(Seed),    
+    rand:seed(exsplus),
+    io:format("seed: ~p\n",[rand:export_seed()]),
     ets:new(nif_SUITE,[named_table,public]),
     ?line true = ets:insert(nif_SUITE,{send3,0,0,0,0}),
     timer:send_after(10000, timeout), % Run for 10 seconds
@@ -1229,7 +1226,7 @@ send3_controller(SpawnCnt0, Mons0, Pids0, Tick) ->
     after Tick -> 
         Max = 20,
         N = length(Pids0),
-        PidN = random:uniform(Max),
+        PidN = rand:uniform(Max),
         %%io:format("N=~p PidN=~p Pids0=~p\n", [N,PidN,Pids0]), 
         case PidN > N of
             true ->
@@ -1293,7 +1290,7 @@ send3_proc(Pids0, Counters={Rcv,SndOk,SndFail}, State0) ->
     end.
 
 send3_proc_send(Pids, {Rcv,SndOk,SndFail}, State0) ->
-    To = lists:nth(random:uniform(length(Pids)),Pids),
+    To = lists:nth(rand:uniform(length(Pids)),Pids),
     Blob = send3_make_blob(),
     State1 = send3_new_state(State0,Blob), 
     case send3_send(To, Blob) of
@@ -1305,12 +1302,12 @@ send3_proc_send(Pids, {Rcv,SndOk,SndFail}, State0) ->
 
 
 send3_make_blob() ->    
-    case random:uniform(20)-1 of
+    case rand:uniform(20)-1 of
         0 -> {term,[]};
         N ->
             MsgEnv = alloc_msgenv(), 
             repeat(N bsr 1,
-                   fun(_) -> grow_blob(MsgEnv,other_term(),random:uniform(1 bsl 20))
+                   fun(_) -> grow_blob(MsgEnv,other_term(),rand:uniform(1 bsl 20))
                    end, void),
             case (N band 1) of
                 0 -> {term,copy_blob(MsgEnv)};
@@ -1320,7 +1317,7 @@ send3_make_blob() ->
 
 send3_send(Pid, Msg) ->
     %% 90% enif_send and 10% normal bang
-    case random:uniform(10) of
+    case rand:uniform(10) of
         1 -> send3_send_bang(Pid,Msg);
         _ -> send3_send_nif(Pid,Msg)
     end.
@@ -1341,7 +1338,7 @@ send3_send_bang(Pid, {msgenv,MsgEnv}) ->
     true.
 
 send3_new_state(State, Blob) ->
-    case random:uniform(5+2) of
+    case rand:uniform(5+2) of
         N when N =< 5-> setelement(N, State, Blob);
         _ -> State  % Don't store blob
     end.
diff --git a/erts/emulator/test/nif_SUITE_data/nif_mod.c b/erts/emulator/test/nif_SUITE_data/nif_mod.c
index 9c78c0e04d..f7e729e2b6 100644
--- a/erts/emulator/test/nif_SUITE_data/nif_mod.c
+++ b/erts/emulator/test/nif_SUITE_data/nif_mod.c
@@ -240,7 +240,7 @@ static ERL_NIF_TERM lib_version(ErlNifEnv* env, int argc, const ERL_NIF_TERM arg
 static ERL_NIF_TERM get_priv_data_ptr(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[])
 {
     ADD_CALL("get_priv_data_ptr");
-    return enif_make_ulong(env, (unsigned long)priv_data(env));
+    return enif_make_uint64(env, (ErlNifUInt64)priv_data(env));
 }
 
 static ERL_NIF_TERM make_new_resource(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[])
diff --git a/erts/emulator/test/num_bif_SUITE.erl b/erts/emulator/test/num_bif_SUITE.erl
index f07f79b83d..90b6a36262 100644
--- a/erts/emulator/test/num_bif_SUITE.erl
+++ b/erts/emulator/test/num_bif_SUITE.erl
@@ -429,7 +429,7 @@ t_string_to_integer(Config) when is_list(Config) ->
 				       list_to_binary(Value))),
 			  {'EXIT', {badarg, _}} = 
 			      (catch erlang:list_to_integer(Value))
-		  end,["1.0"," 1"," -1",""]),
+		  end,["1.0"," 1"," -1","","+"]),
     
     % Custom base error cases
     lists:foreach(fun({Value,Base}) ->
diff --git a/erts/emulator/test/op_SUITE.erl b/erts/emulator/test/op_SUITE.erl
index 6eda78a57b..65a5a4c505 100644
--- a/erts/emulator/test/op_SUITE.erl
+++ b/erts/emulator/test/op_SUITE.erl
@@ -97,10 +97,11 @@ relop_simple(Config) when is_list(Config) ->
     lists:foreach(fun({A,B}) -> relop_simple_do(A,B) end,
 		  Combos),
 
-    repeat(fun() -> Size = random:uniform(100),
-		    Rnd1 = make_rand_term(Size),
-		    {Rnd2,0} = clone_and_mutate(Rnd1, random:uniform(Size)),
-		    relop_simple_do(Rnd1,Rnd2)
+    repeat(fun() ->
+		   Size = rand:uniform(100),
+		   Rnd1 = make_rand_term(Size),
+		   {Rnd2,0} = clone_and_mutate(Rnd1, rand:uniform(Size)),
+		   relop_simple_do(Rnd1,Rnd2)
 	   end,
 	   1000),
     ok.
@@ -158,7 +159,7 @@ cmp_emu(A,B) ->
 make_rand_term(1) ->
     make_rand_term_single();
 make_rand_term(Arity) ->
-    case random:uniform(3) of
+    case rand:uniform(3) of
 	1 ->
 	    make_rand_list(Arity);
 	2 ->
@@ -169,17 +170,17 @@ make_rand_term(Arity) ->
     end.
 
 make_rand_term_single() ->
-    Range = 1 bsl random:uniform(200),
-    case random:uniform(12) of
+    Range = 1 bsl rand:uniform(200),
+    case rand:uniform(12) of
 	1 -> random;
 	2 -> uniform;
-	3 -> random:uniform(Range) - (Range div 2);
-	4 -> Range * (random:uniform() - 0.5);
+	3 -> rand:uniform(Range) - (Range div 2);
+	4 -> Range * (rand:uniform() - 0.5);
 	5 -> 0;
 	6 -> 0.0;
 	7 -> make_ref();
 	8 -> self();
-	9 -> term_to_binary(random:uniform(Range));
+	9 -> term_to_binary(rand:uniform(Range));
 	10 -> fun(X) -> X*Range end; 
 	11 -> fun(X) -> X/Range end;
 	12 -> []
@@ -188,7 +189,7 @@ make_rand_term_single() ->
 make_rand_term_rand_size(1) ->
     {make_rand_term(1), 0};
 make_rand_term_rand_size(MaxArity) ->
-    Arity = random:uniform(MaxArity-1),
+    Arity = rand:uniform(MaxArity-1),
     {make_rand_term(Arity), MaxArity-Arity}.
 
 make_rand_list(0) -> [];
diff --git a/erts/emulator/test/port_SUITE.erl b/erts/emulator/test/port_SUITE.erl
index 3d0509a28c..ff75ee86d6 100644
--- a/erts/emulator/test/port_SUITE.erl
+++ b/erts/emulator/test/port_SUITE.erl
@@ -82,10 +82,11 @@
 	 mul_basic/1, mul_slow_writes/1,
 	 dying_port/1, port_program_with_path/1,
 	 open_input_file_port/1, open_output_file_port/1,
+         count_fds/1,
 	 iter_max_ports/1, eof/1, input_only/1, output_only/1,
 	 name1/1,
 	 t_binary/1, parallell/1, t_exit/1,
-	 env/1, bad_env/1, cd/1, exit_status/1,
+	 env/1, huge_env/1, bad_env/1, cd/1, exit_status/1,
 	 tps_16_bytes/1, tps_1K/1, line/1, stderr_to_stdout/1,
 	 otp_3906/1, otp_4389/1, win_massive/1, win_massive_client/1,
 	 mix_up_ports/1, otp_5112/1, otp_5119/1, otp_6224/1,
@@ -111,8 +112,8 @@ all() ->
      bad_packet, bad_port_messages, {group, options},
      {group, multiple_packets}, parallell, dying_port,
      port_program_with_path, open_input_file_port,
-     open_output_file_port, name1, env, bad_env, cd,
-     exit_status, iter_max_ports, t_exit, {group, tps}, line,
+     open_output_file_port, name1, env, huge_env, bad_env, cd,
+     exit_status, iter_max_ports, count_fds, t_exit, {group, tps}, line,
      stderr_to_stdout, otp_3906, otp_4389, win_massive,
      mix_up_ports, otp_5112, otp_5119,
      exit_status_multi_scheduling_block, ports, spawn_driver,
@@ -385,27 +386,33 @@ input_only(Config) when is_list(Config) ->
 output_only(Config) when is_list(Config) ->
     Dog = test_server:timetrap(test_server:seconds(100)),
     Dir = ?config(priv_dir, Config),
+
+    %% First we test that the port program gets the data
     Filename = filename:join(Dir, "output_only_stream"),
-    output_and_verify(Config, Filename, "-h0",
-          	    random_packet(35777, "echo")),
+    Data = random_packet(35777, "echo"),
+    output_and_verify(Config, ["-h0 -o", Filename], Data),
+    Wait_time = 500,
+    test_server:sleep(Wait_time),
+    {ok, Written} = file:read_file(Filename),
+    Data = binary_to_list(Written),
+
+    %% Then we test that any writes to stdout from
+    %% the port program is not sent to erlang
+    output_and_verify(Config, ["-h0"], Data),
+
     test_server:timetrap_cancel(Dog),
     ok.
 
-output_and_verify(Config, Filename, Options, Data) ->
+output_and_verify(Config, Options, Data) ->
     PortTest = port_test(Config),
-    Command = lists:concat([PortTest, " ",
-          		  Options, " -o", Filename]),
+    Command = lists:concat([PortTest, " " | Options]),
     Port = open_port({spawn, Command}, [out]),
     Port ! {self(), {command, Data}},
     Port ! {self(), close},
     receive
-        {Port, closed} -> ok
-    end,
-    Wait_time = 500,
-    test_server:sleep(Wait_time),
-    {ok, Written} = file:read_file(Filename),
-    Data = binary_to_list(Written),
-    ok.
+        {Port, closed} -> ok;
+        Msg -> ct:fail({received_unexpected_message, Msg})
+    end.
 
 %% Test that receiving several packages written in the same
 %% write operation works.
@@ -610,6 +617,38 @@ open_output_file_port(Config) when is_list(Config) ->
     test_server:timetrap_cancel(Dog),
     ok.
 
+%% Tests that all appropriate fd's have been closed in the port program
+count_fds(suite) -> [];
+count_fds(Config) when is_list(Config) ->
+    case os:type() of
+        {unix, _} ->
+            PrivDir = proplists:get_value(priv_dir, Config),
+            Filename = filename:join(PrivDir, "my_fd_counter"),
+
+            RunTest = fun(PortOpts) ->
+                              PortTest = port_test(Config),
+                              Command = lists:concat([PortTest, " -n -f -o", Filename]),
+                              Port = open_port({spawn, Command}, PortOpts),
+                              Port ! {self(), close},
+                              receive
+                                  {Port, closed} -> ok
+                              end,
+                              test_server:sleep(500),
+                              {ok, Written} = file:read_file(Filename),
+                              Written
+                      end,
+                    <<4:32/native>> = RunTest([out, nouse_stdio]),
+                    <<4:32/native>> = RunTest([in, nouse_stdio]),
+                    <<5:32/native>> = RunTest([in, out, nouse_stdio]),
+            <<3:32/native>> = RunTest([out, use_stdio]),
+            <<3:32/native>> = RunTest([in, use_stdio]),
+            <<3:32/native>> = RunTest([in, out, use_stdio]),
+            <<3:32/native>> = RunTest([in, out, use_stdio, stderr_to_stdout]),
+            <<3:32/native>> = RunTest([out, use_stdio, stderr_to_stdout]);
+        _ ->
+            {skip, "Skipped on windows"}
+    end.
+
 %%
 %% Open as many ports as possible. Do this several times and check
 %% that we get the same number of ports every time.
@@ -680,7 +719,16 @@ close_ports([]) ->
     ok.
 
 open_ports(Name, Settings) ->
-    test_server:sleep(5),
+    case os:type() of
+        {unix, freebsd} ->
+            %% FreeBsd has issues with sendmsg/recvmsg in fork
+            %% implementation and we therefor have to spawn
+            %% slower to make sure that we always hit the same
+            %% make roof.
+            test_server:sleep(10);
+        _ ->
+            test_server:sleep(5)
+    end,
     case catch open_port(Name, Settings) of
 	P when is_port(P) ->
 	    [P| open_ports(Name, Settings)];
@@ -923,6 +971,40 @@ try_bad_env(Env) ->
 	error:badarg -> ok
     end.
 
+%% Test that we can handle a very very large environment gracefully.
+huge_env(Config) when is_list(Config) ->
+    Vars = case os:type() of
+                {win32,_} -> 500;
+                _ ->
+                   %% We create a huge environment,
+                   %% 20000 variables is about 25MB
+                   %% which seems to be the limit on Linux.
+                   20000
+            end,
+    Env = [{[$a + I div (25*25*25*25) rem 25,
+              $a + I div (25*25*25) rem 25,
+              $a + I div (25*25) rem 25,
+              $a+I div 25 rem 25, $a+I rem 25],
+            lists:duplicate(100,$a+I rem 25)}
+           || I <- lists:seq(1,Vars)],
+    try erlang:open_port({spawn,"ls"},[exit_status, {env, Env}]) of
+        P ->
+            receive
+                {P, {exit_status,N}} = M ->
+                    %% We test that the exit status is an integer, this means
+                    %% that the child program has started. If we get an atom
+                    %% something went wrong in the driver which is not ok.
+                    ct:log("Got ~p",[M]),
+                    true = is_integer(N)
+            end
+    catch E:R ->
+            %% Have to catch the error here, as printing the stackdump
+            %% in the ct log is way to heavy for some test machines.
+            ct:fail("Open port failed ~p:~p",[E,R])
+    end.
+
+
+
 %% 'cd' option
 %% (Can perhaps be made smaller by calling the other utility functions
 %% in this module.)
@@ -1228,13 +1310,15 @@ otp_4389(Config)  when is_list(Config) ->
 							    {P,{exit_status,_}} ->
 								TCR ! {self(),ok};
 							    {'EXIT',_,{R2,_}} when R2 == emfile;
-										   R2 == eagain ->
+										   R2 == eagain;
+                                                                                   R2 == enomem ->
 								TCR ! {self(),ok};
 							    Err2 ->
 								TCR ! {self(),{msg,Err2}}
 							end;
 						    {'EXIT',{R1,_}} when R1 == emfile;
-									 R1 == eagain ->
+									 R1 == eagain;
+                                                                         R1 == enomem ->
 							TCR ! {self(),ok};
 						    Err1 ->
 							TCR ! {self(), {open_port,Err1}}
@@ -1840,10 +1924,12 @@ exit_status_msb_test(Config, SleepSecs) when is_list(Config) ->
 					      {Prt,
 					       erlang:system_info(scheduler_id)};
 					  {'EXIT', {Err, _}} when Err == eagain;
-								  Err == emfile ->
+								  Err == emfile;
+                                                                  Err == enomem ->
 					      noop;
 					  {'EXIT', Err} when Err == eagain;
-							     Err == emfile ->
+							     Err == emfile;
+                                                             Err == enomem ->
 					      noop;
 					  Error ->
 					      ?t:fail(Error)
@@ -2183,15 +2269,14 @@ random_char(Chars) ->
     lists:nth(uniform(length(Chars)), Chars).
 
 uniform(N) ->
-    case get(random_seed) of
-	undefined ->	    
-	    {X, Y, Z} = Seed = time(),
-	    io:format("Random seed = ~p\n",[Seed]),
-	    random:seed(X, Y, Z);
+    case rand:export_seed() of
+	undefined ->
+	    rand:seed(exsplus),
+	    io:format("Random seed = ~p\n", [rand:export_seed()]);
 	_ ->
 	    ok
     end,
-    random:uniform(N).
+    rand:uniform(N).
 
 fun_spawn(Fun) ->
     fun_spawn(Fun, []).
@@ -2331,7 +2416,7 @@ close_deaf_port(Config) when is_list(Config) ->
 close_deaf_port_1(200, _) ->
     ok;
 close_deaf_port_1(N, Cmd) ->
-    Timeout = integer_to_list(random:uniform(5*1000)),
+    Timeout = integer_to_list(rand:uniform(5*1000)),
     try open_port({spawn_executable,Cmd},[{args,[Timeout]}]) of
     	Port ->
 	    erlang:port_command(Port,"Hello, can you hear me!?!?"),
@@ -2372,7 +2457,7 @@ port_setget_data(Config) when is_list(Config) ->
     ok.
 
 port_setget_data_hammer(Port, HeapData, IsSet0, N) ->
-    Rand = random:uniform(3),
+    Rand = rand:uniform(3),
     IsSet1 = try case Rand of
 		     1 -> true = erlang:port_set_data(Port, atom), true;
 		     2 -> true = erlang:port_set_data(Port, HeapData), true;
diff --git a/erts/emulator/test/port_SUITE_data/port_test.c b/erts/emulator/test/port_SUITE_data/port_test.c
index 7abefab2e3..cc3ebdf0f8 100644
--- a/erts/emulator/test/port_SUITE_data/port_test.c
+++ b/erts/emulator/test/port_SUITE_data/port_test.c
@@ -13,6 +13,7 @@
 
 #ifndef __WIN32__
 #include <unistd.h>
+#include <limits.h>
 
 #include <sys/time.h>
 
@@ -48,6 +49,7 @@ typedef struct {
 				 * after reading the header for a packet
 				 * before reading the rest.
 				 */
+    int fd_count;               /* Count the number of open fds */
     int break_mode;		/* If set, this program will close standard
 				 * input, which should case broken pipe
 				 * error in the writer.
@@ -107,7 +109,7 @@ MAIN(argc, argv)
 int argc;
 char *argv[];
 {
-  int ret;
+  int ret, fd_count;
   if((port_data = (PORT_TEST_DATA *) malloc(sizeof(PORT_TEST_DATA))) == NULL) {
     fprintf(stderr, "Couldn't malloc for port_data");
     exit(1);
@@ -115,6 +117,7 @@ char *argv[];
   port_data->header_size = 0;
   port_data->io_buf_size = 0;	
   port_data->delay_mode = 0;	
+  port_data->fd_count = 0;
   port_data->break_mode = 0;	
   port_data->quit_mode = 0;	
   port_data->slow_writes = 0;	
@@ -144,6 +147,9 @@ char *argv[];
     case 'e': 
       port_data->fd_to_erl = 2;
       break;
+    case 'f':
+      port_data->fd_count = 1;
+      break;
     case 'h':			/* Header size for packets. */
       switch (argv[1][2]) {
       case '0': port_data->header_size = 0; break;
@@ -189,18 +195,31 @@ char *argv[];
     /* XXX Add error printout here */
   }
 
+  if (port_data->fd_count) {
+#ifdef __WIN32__
+      DWORD handles;
+      GetProcessHandleCount(GetCurrentProcess(), &handles);
+      fd_count = handles;
+#else
+      int i;
+      for (i = 0, fd_count = 0; i < 1024; i++)
+          if (fcntl(i, F_GETFD) >= 0) {
+              fd_count++;
+          }
+#endif
+  }
+
+  if (port_data->output_file)
+      replace_stdout(port_data->output_file);
+
+  if (port_data->fd_count)
+      reply(&fd_count, sizeof(fd_count));
+
   if (port_data->no_packet_loop){
       free(port_data);
       exit(0);
   }
 
-  /*
-   * If an output file was given, let it replace standard output.
-   */
-
-  if (port_data->output_file)
-    replace_stdout(port_data->output_file);
-
   ret = packet_loop();
   if(port_data->io_buf_size > 0)
       free(port_data->io_buf);
diff --git a/erts/emulator/test/port_bif_SUITE.erl b/erts/emulator/test/port_bif_SUITE.erl
index b65a22a528..981899b167 100644
--- a/erts/emulator/test/port_bif_SUITE.erl
+++ b/erts/emulator/test/port_bif_SUITE.erl
@@ -485,14 +485,7 @@ random_char(Chars) ->
     lists:nth(uniform(length(Chars)), Chars).
 
 uniform(N) ->
-    case get(random_seed) of
-	undefined ->
-	    {X, Y, Z} = time(),
-	    random:seed(X, Y, Z);
-	_ ->
-	    ok
-    end,
-    random:uniform(N).
+    rand:uniform(N).
 
 unaligned_sub_bin(Bin0) ->
     Bin1 = <<0:3,Bin0/binary,31:5>>,
diff --git a/erts/emulator/test/process_SUITE.erl b/erts/emulator/test/process_SUITE.erl
index 4c311e1f06..97aa5e573e 100644
--- a/erts/emulator/test/process_SUITE.erl
+++ b/erts/emulator/test/process_SUITE.erl
@@ -1478,7 +1478,15 @@ processes_this_tab(doc) ->
 processes_this_tab(suite) ->
     [];
 processes_this_tab(Config) when is_list(Config) ->
-    sys_mem_cond_run(1024, fun () -> chk_processes_bif_test_res(processes_bif_test()) end).
+    Mem = case {erlang:system_info(build_type),
+                erlang:system_info(allocator)} of
+              {lcnt, {_, _Vsn, [sys_alloc], _Opts}} ->
+                  %% When running +Mea min + lcnt we may need more memory
+                  1024 * 4;
+              _ ->
+                  1024
+          end,
+    sys_mem_cond_run(Mem, fun () -> chk_processes_bif_test_res(processes_bif_test()) end).
 
 chk_processes_bif_test_res(ok) -> ok;
 chk_processes_bif_test_res({comment, _} = Comment) -> Comment;
diff --git a/erts/emulator/test/random_iolist.erl b/erts/emulator/test/random_iolist.erl
index 9a0f034e72..6da7da04de 100644
--- a/erts/emulator/test/random_iolist.erl
+++ b/erts/emulator/test/random_iolist.erl
@@ -36,7 +36,7 @@ run2(Iter,Fun1,Fun2) ->
     compare2(Iter,Fun1,Fun2).
 
 random_byte() ->
-     random:uniform(256) - 1.
+     rand:uniform(256) - 1.
 
 random_list(0,Acc) ->
     Acc;
@@ -45,7 +45,7 @@ random_list(N,Acc) ->
 
 random_binary(N) ->
     B = list_to_binary(random_list(N,[])),
-    case {random:uniform(2),size(B)} of
+    case {rand:uniform(2),size(B)} of
 	{2,M} when M > 1 ->
 	    S = M-1,
 	    <<_:3,C:S/binary,_:5>> = B,
@@ -57,7 +57,7 @@ random_list(N) ->
     random_list(N,[]).
 
 front() ->
-    case random:uniform(10) of
+    case rand:uniform(10) of
 	10 ->
 	    false;
 	_ ->
@@ -65,7 +65,7 @@ front() ->
     end.
 
 any_type() ->
-    case random:uniform(10) of
+    case rand:uniform(10) of
 	1 ->
 	    list;
 	2 ->
@@ -77,7 +77,7 @@ any_type() ->
     end.
 
 tail_type() ->
-    case random:uniform(5) of
+    case rand:uniform(5) of
 	1 ->
 	    list;
 	2 ->
@@ -90,9 +90,9 @@ random_length(N) ->
     UpperLimit = 255,
     case N of
 	M when M > UpperLimit ->
-	    random:uniform(UpperLimit+1) - 1;
+	    rand:uniform(UpperLimit+1) - 1;
 	_ ->
-	    random:uniform(N+1) - 1
+	    rand:uniform(N+1) - 1
     end.
 
 random_iolist(0,Acc) ->
@@ -139,7 +139,7 @@ random_iolist(N) ->
     
 
 standard_seed() ->
-    random:seed(1201,855653,380975).
+    rand:seed(exsplus, {1201,855653,380975}).
 
 do_comp(List,F1,F2) ->
     X = F1(List),
diff --git a/erts/emulator/test/save_calls_SUITE.erl b/erts/emulator/test/save_calls_SUITE.erl
index 544d841f16..810bc07eed 100644
--- a/erts/emulator/test/save_calls_SUITE.erl
+++ b/erts/emulator/test/save_calls_SUITE.erl
@@ -189,7 +189,7 @@ is_local_function(_) ->
 
 % Number crunching for reds test.
 carmichaels_below(N) ->
-    random:seed(3172,9814,20125),
+    rand:seed(exsplus, {3172,9814,20125}),
     carmichaels_below(1,N).
 
 carmichaels_below(N,N2) when N >= N2 ->
@@ -219,7 +219,7 @@ expmod(Base,Exp,Mod) ->
     (Base * expmod(Base,Exp - 1,Mod)) rem Mod.
 
 uniform(N) ->
-    random:uniform(N-1).
+    rand:uniform(N-1).
 
 fermat(N) ->    
     R = uniform(N),
diff --git a/erts/emulator/test/statistics_SUITE.erl b/erts/emulator/test/statistics_SUITE.erl
index 56ecf4195a..53c9ba8715 100644
--- a/erts/emulator/test/statistics_SUITE.erl
+++ b/erts/emulator/test/statistics_SUITE.erl
@@ -32,7 +32,7 @@
 	 run_queue_one/1,
 	 scheduler_wall_time/1,
 	 reductions/1, reductions_big/1, garbage_collection/1, io/1,
-	 badarg/1]).
+	 badarg/1, run_queues_lengths_active_tasks/1]).
 
 %% Internal exports.
 
@@ -54,7 +54,8 @@ suite() -> [{ct_hooks,[ts_install_cth]}].
 all() -> 
     [{group, wall_clock}, {group, runtime}, reductions,
      reductions_big, {group, run_queue}, scheduler_wall_time,
-     garbage_collection, io, badarg].
+     garbage_collection, io, badarg,
+     run_queues_lengths_active_tasks].
 
 groups() -> 
     [{wall_clock, [],
@@ -409,3 +410,59 @@ badarg(Config) when is_list(Config) ->
     ?line case catch statistics(bad_atom) of
 	      {'EXIT', {badarg, _}} -> ok
 	  end.
+
+tok_loop() ->
+    tok_loop().
+
+run_queues_lengths_active_tasks(Config) ->
+    TokLoops = lists:map(fun (_) ->
+				 spawn_opt(fun () ->
+						   tok_loop()
+					   end,
+					   [link, {priority, low}])
+			 end,
+			 lists:seq(1,10)),
+
+    TRQLs0 = statistics(total_run_queue_lengths),
+    TATs0 = statistics(total_active_tasks),
+    true = is_integer(TRQLs0),
+    true = is_integer(TATs0),
+    true = TRQLs0 >= 0,
+    true = TATs0 >= 11,
+
+    NoScheds = erlang:system_info(schedulers),
+    RQLs0 = statistics(run_queue_lengths),
+    ATs0 = statistics(active_tasks),
+    NoScheds = length(RQLs0),
+    NoScheds = length(ATs0),
+    true = lists:sum(RQLs0) >= 0,
+    true = lists:sum(ATs0) >= 11,
+
+    SO = erlang:system_flag(schedulers_online, 1),
+
+    TRQLs1 = statistics(total_run_queue_lengths),
+    TATs1 = statistics(total_active_tasks),
+    true = TRQLs1 >= 10,
+    true = TATs1 >= 11,
+    NoScheds = erlang:system_info(schedulers),
+
+    RQLs1 = statistics(run_queue_lengths),
+    ATs1 = statistics(active_tasks),
+    NoScheds = length(RQLs1),
+    NoScheds = length(ATs1),
+    TRQLs2 = lists:sum(RQLs1),
+    TATs2 = lists:sum(ATs1),
+    true = TRQLs2 >= 10,
+    true = TATs2 >= 11,
+    [TRQLs2|_] = RQLs1,
+    [TATs2|_] = ATs1,
+
+    erlang:system_flag(schedulers_online, SO),
+
+    lists:foreach(fun (P) ->
+			  unlink(P),
+			  exit(P, bang)
+		  end,
+		  TokLoops),
+
+    ok.
diff --git a/erts/emulator/test/system_info_SUITE.erl b/erts/emulator/test/system_info_SUITE.erl
index bee42c07d9..51122e5d55 100644
--- a/erts/emulator/test/system_info_SUITE.erl
+++ b/erts/emulator/test/system_info_SUITE.erl
@@ -185,8 +185,6 @@ wordsize(Config) when is_list(Config) ->
 	    {comment, "True 32-bit emulator"};
 	{8,8} ->
 	    {comment, "True 64-bit emulator"};
-	{8,4} ->
-	    {comment, "Halfword 64-bit emulator"};
 	Other ->
 	    exit({unexpected_wordsizes,Other})
     end.
diff --git a/erts/emulator/test/system_profile_SUITE.erl b/erts/emulator/test/system_profile_SUITE.erl
index e4b6511d1f..0a0784337f 100644
--- a/erts/emulator/test/system_profile_SUITE.erl
+++ b/erts/emulator/test/system_profile_SUITE.erl
@@ -448,7 +448,7 @@ run_load(N, Pids) ->
     run_load(N - 1, [Pid | Pids]).
 
 list_load() -> 
-    ok = case math:sin(random:uniform(32451)) of
+    ok = case math:sin(rand:uniform(32451)) of
     	A when is_float(A) -> ok;
 	_ -> ok
     end,
diff --git a/erts/emulator/test/time_SUITE.erl b/erts/emulator/test/time_SUITE.erl
index 33076c7461..3bd28a6d20 100644
--- a/erts/emulator/test/time_SUITE.erl
+++ b/erts/emulator/test/time_SUITE.erl
@@ -69,7 +69,7 @@
 init_per_testcase(Func, Config) when is_atom(Func), is_list(Config) ->
     [{testcase, Func}|Config].
 
-end_per_testcase(_Func, Config) ->
+end_per_testcase(_Func, _Config) ->
     ok.
 
 suite() -> [{ct_hooks,[ts_install_cth]}].
@@ -742,25 +742,21 @@ chk_strc(Res0, Res1) ->
     ok.
 
 chk_random_values(FR, TR) ->
-%    case (FR rem TR == 0) orelse (TR rem FR == 0) of
-%	true ->
-	    io:format("rand values ~p -> ~p~n", [FR, TR]),
-	    random:seed(268438039, 268440479, 268439161),
-	    Values = lists:map(fun (_) -> random:uniform(1 bsl 65) - (1 bsl 64) end,
-			       lists:seq(1, 100000)),
-	    CheckFun = fun (V) ->
-			       CV = erlang:convert_time_unit(V, FR, TR),
-			       case {(FR*CV) div TR =< V,
-				     (FR*(CV+1)) div TR >= V} of
-				   {true, true} ->
-				       ok;
-				   Failure ->
-				       ?t:fail({Failure, CV, V, FR, TR})
-			       end
-		       end,
-	    lists:foreach(CheckFun, Values).%;
-%	false -> ok
-%    end.
+    io:format("rand values ~p -> ~p~n", [FR, TR]),
+    rand:seed(exsplus, {268438039,268440479,268439161}),
+    Values = lists:map(fun (_) -> rand:uniform(1 bsl 65) - (1 bsl 64) end,
+		       lists:seq(1, 100000)),
+    CheckFun = fun (V) ->
+		       CV = erlang:convert_time_unit(V, FR, TR),
+		       case {(FR*CV) div TR =< V,
+			     (FR*(CV+1)) div TR >= V} of
+			   {true, true} ->
+			       ok;
+			   Failure ->
+			       ?t:fail({Failure, CV, V, FR, TR})
+		       end
+	       end,
+    lists:foreach(CheckFun, Values).
 		       
 
 chk_values_per_value(_FromRes, _ToRes,
diff --git a/erts/emulator/test/trace_SUITE.erl b/erts/emulator/test/trace_SUITE.erl
index 6eae182e45..00b90e3c3d 100644
--- a/erts/emulator/test/trace_SUITE.erl
+++ b/erts/emulator/test/trace_SUITE.erl
@@ -933,7 +933,7 @@ suspend_exit(suite) ->
     [];
 suspend_exit(Config) when is_list(Config) ->
     ?line Dog = test_server:timetrap(test_server:minutes(2)),
-    ?line random:seed(4711,17,4711),
+    rand:seed(exsplus, {4711,17,4711}),
     ?line do_suspend_exit(5000),
     ?line test_server:timetrap_cancel(Dog),
     ?line ok.
@@ -941,7 +941,7 @@ suspend_exit(Config) when is_list(Config) ->
 do_suspend_exit(0) ->
     ?line ok;
 do_suspend_exit(N) ->
-    ?line Work = random:uniform(50),
+    Work = rand:uniform(50),
     ?line Parent = self(),
     ?line {Suspendee, Mon2}
 	= spawn_monitor(fun () ->
diff --git a/erts/emulator/test/trace_port_SUITE.erl b/erts/emulator/test/trace_port_SUITE.erl
index d6346f3af0..a77710205e 100644
--- a/erts/emulator/test/trace_port_SUITE.erl
+++ b/erts/emulator/test/trace_port_SUITE.erl
@@ -35,7 +35,8 @@
 	 fake_schedule_after_getting_linked/1,
 	 fake_schedule_after_getting_unlinked/1,
 	 gc/1,
-	 default_tracer/1]).
+	 default_tracer/1,
+	 tracer_port_crash/1]).
 
 -include_lib("test_server/include/test_server.hrl").
 
@@ -45,7 +46,7 @@ test_cases() ->
      fake_schedule_after_register,
      fake_schedule_after_getting_linked,
      fake_schedule_after_getting_unlinked, gc,
-     default_tracer].
+     default_tracer, tracer_port_crash].
 
 suite() -> [{ct_hooks,[ts_install_cth]}].
 
@@ -473,6 +474,42 @@ default_tracer(Config) when is_list(Config) ->
     ?line M = N,
     ok.
 
+tracer_port_crash(Config) when is_list(Config) ->
+    case test_server:is_native(?MODULE) orelse
+	test_server:is_native(lists) of
+	true -> 
+	    {skip,"Native code"};
+	false ->
+	    Tr = start_tracer(Config),
+	    Port = get(tracer_port),
+	    Tracee = spawn(fun () ->
+				   register(trace_port_linker, self()),
+				   link(Port),
+				   receive go -> ok end,
+				   lists:reverse([1,b,c]),
+				   receive die -> ok end
+			   end),
+	    Tr ! {unlink_tracer_port, self()},
+	    receive {unlinked_tracer_port, Tr} -> ok end,
+	    port_control(Port, $c, []), %% Make port commands crash tracer port...
+	    trace_func({lists,reverse,1}, []),
+	    trace_pid(Tracee, true, [call]),
+	    trace_info(Tracee, flags),
+	    trace_info(self(), tracer),
+	    Tracee ! go,
+	    receive after 1000 -> ok end,
+	    case whereis(trace_port_linker) of
+		undefined ->
+		    ok;
+		Id ->
+%		    erts_debug:set_internal_state(available_internal_state, true),
+%		    erts_debug:set_internal_state(abort, {trace_port_linker, Id})
+		    ?t:fail({trace_port_linker, Id})
+	    end,
+	    undefined = process_info(Tracee),
+	    ok
+    end.
+
 %%% Help functions.
 
 huge_data() -> huge_data(16384).
@@ -631,6 +668,10 @@ tracer_loop(RelayTo, Port) ->
 	{Port,{data,Msg}} ->
 	    RelayTo ! binary_to_term(Msg),
 	    tracer_loop(RelayTo, Port);
+	{unlink_tracer_port, From} ->
+	    unlink(Port),
+	    From ! {unlinked_tracer_port, self()},
+	    tracer_loop(RelayTo, Port);
 	Other ->
 	    exit({bad_message,Other})
     end.
diff --git a/erts/emulator/test/trace_port_SUITE_data/echo_drv.c b/erts/emulator/test/trace_port_SUITE_data/echo_drv.c
index a8d4ede4fe..e40b9193ea 100644
--- a/erts/emulator/test/trace_port_SUITE_data/echo_drv.c
+++ b/erts/emulator/test/trace_port_SUITE_data/echo_drv.c
@@ -1,5 +1,6 @@
 #include <stdio.h>
 #include "erl_driver.h"
+#include <errno.h>
 
 
 
@@ -14,6 +15,7 @@ enum e_heavy {
 typedef struct _erl_drv_data {
     ErlDrvPort   erlang_port;
     enum e_heavy heavy;
+    int crash;
 } EchoDrvData;
 
 static EchoDrvData echo_drv_data, *echo_drv_data_p;
@@ -78,6 +80,7 @@ static EchoDrvData *echo_drv_start(ErlDrvPort port, char *command)
     echo_drv_data_p = &echo_drv_data;
     echo_drv_data_p->erlang_port = port;
     echo_drv_data_p->heavy = heavy_off;
+    echo_drv_data_p->crash = 0;
     return echo_drv_data_p;
 }
 
@@ -87,6 +90,12 @@ static void echo_drv_stop(EchoDrvData *data_p) {
 
 static void echo_drv_output(ErlDrvData drv_data, char *buf, ErlDrvSizeT len) {
     EchoDrvData* data_p = (EchoDrvData *) drv_data;
+
+    if (data_p->crash) {
+	driver_failure_posix(data_p->erlang_port, EINTR);
+	return;
+    }
+
     driver_output(data_p->erlang_port, buf, len);
     switch (data_p->heavy) {
     case heavy_off:
@@ -100,6 +109,7 @@ static void echo_drv_output(ErlDrvData drv_data, char *buf, ErlDrvSizeT len) {
 	data_p->heavy = heavy_off;
 	break;
     }
+
 }
 
 static void echo_drv_finish() {
@@ -115,6 +125,8 @@ static ErlDrvSSizeT echo_drv_control(ErlDrvData drv_data,
     case 'h':
 	data_p->heavy = heavy_set;
 	break;
+    case 'c':
+	data_p->crash = 1;
     }
     return 0;
 }
diff --git a/erts/emulator/test/z_SUITE.erl b/erts/emulator/test/z_SUITE.erl
index f4d9030255..abc353fb01 100644
--- a/erts/emulator/test/z_SUITE.erl
+++ b/erts/emulator/test/z_SUITE.erl
@@ -249,6 +249,7 @@ pollset_size(Config) when is_list(Config) ->
     ?line io:format("Initial: ~p~nFinal: ~p~n", [InitChkIo, FinChkIo]),
     ?line InitPollsetSize = lists:keysearch(total_poll_set_size, 1, InitChkIo),
     ?line FinPollsetSize = lists:keysearch(total_poll_set_size, 1, FinChkIo),
+    HasGethost = case has_gethost() of true -> 1; _ -> 0 end,
     ?line case InitPollsetSize =:= FinPollsetSize of
 	      true ->
 		  case InitPollsetSize of
@@ -269,7 +270,7 @@ pollset_size(Config) when is_list(Config) ->
 		      = InitPollsetSize,
 		  ?line {value, {total_poll_set_size, FinSize}}
 		      = FinPollsetSize,
-		  ?line true = FinSize < InitSize,
+		  ?line true = FinSize < (InitSize + HasGethost),
 		  ?line true = 2 =< FinSize,
 		  ?line {comment,
 			 "Start pollset size: "
@@ -289,16 +290,39 @@ check_io_debug(Config) when is_list(Config) ->
 	  end.
 
 check_io_debug_test() ->
-    ?line erlang:display(get_check_io_info()),
-    ?line erts_debug:set_internal_state(available_internal_state, true),
-    ?line {NoErrorFds, NoUsedFds, NoDrvSelStructs, NoDrvEvStructs}
+    erlang:display(get_check_io_info()),
+    erts_debug:set_internal_state(available_internal_state, true),
+    {NoErrorFds, NoUsedFds, NoDrvSelStructs, NoDrvEvStructs} = CheckIoDebug
 	= erts_debug:get_internal_state(check_io_debug),
-    ?line erts_debug:set_internal_state(available_internal_state, false),
-    ?line 0 = NoErrorFds,
-    ?line NoUsedFds = NoDrvSelStructs,
+    erts_debug:set_internal_state(available_internal_state, false),
+    HasGetHost = has_gethost(),
+    ct:log("check_io_debug: ~p~n"
+           "HasGetHost: ~p",[CheckIoDebug, HasGetHost]),
+    0 = NoErrorFds,
+    if
+        NoUsedFds == NoDrvSelStructs ->
+            ok;
+        HasGetHost andalso (NoUsedFds == (NoDrvSelStructs - 1)) ->
+            %% If the inet_gethost port is alive, we may have
+            %% one extra used fd that is not selected on.
+            %% This happens when the initial setup of the
+            %% port returns an EAGAIN
+            ok
+    end,
     ?line 0 = NoDrvEvStructs,
     ?line ok.
 
+has_gethost() ->
+    has_gethost(erlang:ports()).
+has_gethost([P|T]) ->
+    case erlang:port_info(P, name) of
+        {name,"inet_gethost"++_} ->
+            true;
+        _ ->
+            has_gethost(T)
+    end;
+has_gethost([]) ->
+    false.
 
 
 %%
diff --git a/erts/emulator/utils/beam_makeops b/erts/emulator/utils/beam_makeops
index 9a8c3585e6..f805e7cc64 100755
--- a/erts/emulator/utils/beam_makeops
+++ b/erts/emulator/utils/beam_makeops
@@ -48,12 +48,17 @@ $pack_shift[4] = ['0', 'BEAM_LOOSE_SHIFT', # Only for 64 bit wordsize
 		  '(3*BEAM_LOOSE_SHIFT)'];
 
 $pack_mask[2]  = ['BEAM_LOOSE_MASK', $WHOLE_WORD];
-$pack_mask[3]  = ['BEAM_TIGHT_MASK', 'BEAM_TIGHT_MASK', 'BEAM_TIGHT_MASK'];
+$pack_mask[3]  = ['BEAM_TIGHT_MASK', 'BEAM_TIGHT_MASK', $WHOLE_WORD];
 $pack_mask[4]  = ['BEAM_LOOSE_MASK', # Only for 64 bit wordsize
 		  'BEAM_LOOSE_MASK',
 		  'BEAM_LOOSE_MASK',
 		  $WHOLE_WORD];
 
+# Mapping from packagable arguments to number of packed arguments per
+# word.  Initialized after the wordsize is known.
+
+my @args_per_word;
+
 # There are two types of instructions: generic and specific.
 # The generic instructions are those generated by the Beam compiler.
 # Corresponding to each generic instruction, there is generally a
@@ -176,11 +181,12 @@ sub define_type_bit {
     }
 
     # Composed types.
-    define_type_bit('d', $type_bit{'x'} | $type_bit{'y'} | $type_bit{'r'});
+    define_type_bit('d', $type_bit{'x'} | $type_bit{'y'});
     define_type_bit('c', $type_bit{'i'} | $type_bit{'a'} |
 		    $type_bit{'n'} | $type_bit{'q'});
     define_type_bit('s', $type_bit{'d'} | $type_bit{'i'} |
-		    $type_bit{'a'} | $type_bit{'n'});
+		    $type_bit{'a'} | $type_bit{'n'} |
+		    $type_bit{'q'});
     define_type_bit('j', $type_bit{'f'} | $type_bit{'p'});
 
     # Aliases (for matching purposes).
@@ -236,6 +242,20 @@ while (@ARGV && $ARGV[0] =~ /^-(.*)/) {
 }
 
 #
+# Initialize number of arguments per packed word.
+#
+
+$args_per_word[2] = 2;
+$args_per_word[3] = 3;
+$args_per_word[4] = 2;
+$args_per_word[5] = 3;
+$args_per_word[6] = 3;
+
+if ($wordsize == 64) {
+    $args_per_word[4] = 4;
+}
+
+#
 # Parse the input files.
 #
 
@@ -527,12 +547,16 @@ sub emulator_output {
 
 	    my(@bits) = (0) x ($max_spec_operands/2);
 	    my($i);
+	    my $involves_r = 0;
 	    for ($i = 0; $i < $max_spec_operands && defined $args[$i]; $i++) {
 		my $t = $args[$i];
-		if (defined $type_bit{$t}) {
-		    my $shift = $max_genop_types * ($i % 2);
-		    $bits[int($i/2)] |= $type_bit{$t} << $shift;
+		my $bits = $type_bit{$t};
+		if ($t eq 'r') {
+		    $bits |= $type_bit{'x'};
+		    $involves_r |= 1 << $i;
 		}
+		my $shift = $max_genop_types * ($i % 2);
+		$bits[int($i/2)] |= $bits << $shift;
 	    }
 
 	    printf "/* %3d */  ", $spec_opnum;
@@ -544,7 +568,7 @@ sub emulator_output {
 		$sep = ",";
 	    }
 	    $init .= "}";
-	    &init_item($print_name, $init, $size, $pack, $sign, 0);
+	    init_item($print_name, $init, $involves_r, $size, $pack, $sign, 0);
 	    $op_to_name[$spec_opnum] = $instr;
 	    $spec_opnum++;
 	}
@@ -601,15 +625,12 @@ sub emulator_output {
     print "#define MAX_GENERIC_OPCODE ", $num_file_opcodes-1, "\n";
     print "#define NUM_GENERIC_OPS ", scalar(@gen_opname), "\n";
     print "#define NUM_SPECIFIC_OPS ", scalar(@op_to_name), "\n";
+    print "#define SCRATCH_X_REG 1023\n";
     print "\n";
     print "#ifdef ARCH_64\n";
     print "#  define BEAM_WIDE_MASK 0xFFFFUL\n";
-    print "#  define BEAM_LOOSE_MASK 0x1FFFUL\n";
-    print "#if HALFWORD_HEAP\n";
-    print "#  define BEAM_TIGHT_MASK 0x1FFCUL\n";
-    print "#else\n";
-    print "#  define BEAM_TIGHT_MASK 0x1FF8UL\n";
-    print "#endif\n";
+    print "#  define BEAM_LOOSE_MASK 0xFFFFUL\n";
+    print "#  define BEAM_TIGHT_MASK 0xFFFFUL\n";
     print "#  define BEAM_WIDE_SHIFT 32\n";
     print "#  define BEAM_LOOSE_SHIFT 16\n";
     print "#  define BEAM_TIGHT_SHIFT 16\n";
@@ -734,7 +755,7 @@ sub init_item {
 	    print "${sep}NULL";
 	} elsif (/^\{/) {
 	    print "$sep$_";
-	} elsif (/^-?\d/) {
+	} elsif (/^-?\d+$/) {
 	    print "$sep$_";
 	} else {
 	    print "$sep\"$_\"";
@@ -896,6 +917,7 @@ sub basic_generator {
     my($var_decls) = '';
     my($gen_dest_arg) = 'StoreSimpleDest';
     my($i);
+    my($no_prefetch) = 0;
 
     # The following argument types should be included as macro arguments.
     my(%incl_arg) = ('c' => 1,
@@ -994,6 +1016,7 @@ sub basic_generator {
     #
 
     $flags =~ /-fail_action/ and do {
+	$no_prefetch = 1;
 	if (!defined $fail_type) {
 	    my($i);
 	    for ($i = 0; $i < @f_types; $i++) {
@@ -1040,6 +1063,12 @@ sub basic_generator {
 			 "I += $size + 1;",
 			 "goto $goto;",
 			 "}");
+	} elsif ($no_prefetch) {
+	    $code = join("\n",
+			 "{ $var_decls",
+			 $macro_code,
+			 "Next($size);",
+			 "}", "");
 	} else {
 	    $code = join("\n",
 			 "{ $var_decls",
@@ -1061,6 +1090,7 @@ sub do_pack {
     my($packable_args) = 0;
     my @is_packable;		# Packability (boolean) for each argument.
     my $wide_packing = 0;
+    my(@orig_args) = @args;
 
     #
     # Count the number of packable arguments.  If we encounter any 's' or 'd'
@@ -1081,6 +1111,18 @@ sub do_pack {
 	    }
 	} elsif ($arg =~ /^[sd]/) {
 	    return ('', '', @args);
+	} elsif ($arg =~ /^[scq]/ and $packable_args > 0) {
+	    # When packing, this operand will be picked up from the
+	    # code array, put onto the packing stack, and later put
+	    # back into a different location in the code. The problem
+	    # is that if this operand is a literal, the original
+	    # location in the code would have been remembered in a
+	    # literal patch.  For packing to work, we would have to
+	    # adjust the position in the literal patch. For the
+	    # moment, adding additional instructions to the packing
+	    # engine to handle this does not seem worth it, so we will
+	    # just turn off packing.
+	    return ('', '', @args);
 	} else {
 	    push @is_packable, 0;
 	}
@@ -1090,7 +1132,6 @@ sub do_pack {
     # Get out of here if too few or too many arguments.
     #
     return ('', '', @args) if $packable_args < 2;
-    &error("too many packable arguments") if $packable_args > 4;
 
     my($size) = 0;
     my($pack_prefix) = '';
@@ -1098,14 +1139,8 @@ sub do_pack {
 				# beginning).
     my($up) = '';		# Pack commands (storing back while
 				# moving forward).
-    my $args_per_word;
-    if ($packable_args < 4 or $wordsize == 64) {
-	$args_per_word = $packable_args;
-    } else {
-	# 4 packable argument, 32 bit wordsize. Need 2 words.
-	$args_per_word = 2;
-    }
 
+    my $args_per_word = $args_per_word[$packable_args];
     my @shift;
     my @mask;
     my @instr;
@@ -1299,6 +1334,8 @@ sub tr_parse_op {
 	foreach (split('', $type)) {
 	    &error("bad type in $op")
 		unless defined $type_bit{$_} or $type eq '*';
+	    $_ eq 'r' and
+		error("$op: 'r' is not allowed in transformations")
 	}
     }
 
@@ -1332,7 +1369,10 @@ sub tr_parse_op {
     }
 
     # Get an optional value. (In destination.)
+    $type_val = $type eq 'x' ? 1023 : 0;
     if (/^=(.*)/) {
+	error("value not allowed in source: $op")
+	    if $src;
 	$type_val = $1;
 	$_ = '';
     }
@@ -1351,11 +1391,6 @@ sub tr_parse_op {
 	    if $var && $type;
     }
 
-    # Test that source has no values.
-    if ($src) {
-	error("value not allowed in source: $op")
-	    if $type_val;
-    }
     ($var,$type,$type_val,$cond,$cond_val);
 }
 
diff --git a/erts/emulator/utils/make_alloc_types b/erts/emulator/utils/make_alloc_types
index 88f537ea09..925b9d5810 100755
--- a/erts/emulator/utils/make_alloc_types
+++ b/erts/emulator/utils/make_alloc_types
@@ -246,7 +246,7 @@ print DST "
 
 print DST "#define ERTS_ALC_C_MIN ($c_no)\n\n";
 
-foreach my $c (keys(%c_tab)) {
+foreach my $c (sort keys(%c_tab)) {
   push(@c_order, $c);
   set_number($c_tab{$c}, $c_no);
   print DST "#define ERTS_ALC_C_$c  ($c_no)\n";
diff --git a/erts/emulator/utils/make_version b/erts/emulator/utils/make_version
index 3461dc1637..37bdff181a 100755
--- a/erts/emulator/utils/make_version
+++ b/erts/emulator/utils/make_version
@@ -59,7 +59,11 @@ print FILE <<EOF;
 #define ERLANG_OTP_RELEASE "$release"
 #define ERLANG_OTP_VERSION "$otp_version"
 #define ERLANG_VERSION "$version"
-#define ERLANG_COMPILE_DATE "$time_str"
+#if ERTS_SAVED_COMPILE_TIME
+#  define ERLANG_COMPILE_DATE "$time_str"
+#else
+#  define ERLANG_COMPILE_DATE ""
+#endif
 #define ERLANG_ARCHITECTURE "$architecture"
 EOF
 
diff --git a/erts/emulator/utils/mkver.c b/erts/emulator/utils/mkver.c
index 96cd315a95..6641873712 100644
--- a/erts/emulator/utils/mkver.c
+++ b/erts/emulator/utils/mkver.c
@@ -35,8 +35,10 @@ int argc;
 char** argv;
 {
     FILE *file;
+#if ERTS_SAVED_COMPILE_TIME
     time_t now;
-    char *cnow;
+#endif
+    char *cnow = "";
 
     if (argc != 2) {
 	fprintf(stderr, "usage: mkver version\n");
@@ -48,9 +50,11 @@ char** argv;
 	exit(1);
     }
 
+#if ERTS_SAVED_COMPILE_TIME
     time(&now);
     cnow = ctime(&now);
     cnow[24] = '\0';		/* tidelipom */
+#endif
     fprintf(file, "/* This file was created by mkver -- don't modify.*/\n");
     fprintf(file, "#define ERLANG_VERSION \"%s\"\n", argv[1]);
     fprintf(file, "#define ERLANG_COMPILE_DATE \"%s\"\n", cnow);
diff --git a/erts/emulator/valgrind/suppress.halfword b/erts/emulator/valgrind/suppress.halfword
deleted file mode 100644
index 8fe448d897..0000000000
--- a/erts/emulator/valgrind/suppress.halfword
+++ /dev/null
@@ -1,56 +0,0 @@
-# Extra suppressions specific for the halfword emulator.
-
-# --- Suppress all offheap binaries ---
-# Valgrinds leak check does not recognize pointers that are stored
-# at unaligned addresses. In halfword emulator we store 64-bit pointers
-# to offheap data on 32-bit aligned heaps.
-# We solve this by suppressing allocation of all offheap structures
-# that are not referenced by other tables (ie binaries).
-
-{
-Halfword erts_bin_nrml_alloc
-Memcheck:Leak
-...
-fun:erts_bin_nrml_alloc
-...
-}
-
-{
-Halfword erts_bin_realloc
-Memcheck:Leak
-...
-fun:erts_bin_realloc
-...
-}
-
-{
-Halfword erts_bin_realloc_fnf
-Memcheck:Leak
-...
-fun:erts_bin_realloc_fnf
-...
-}
-
-{
-Halfword erts_bin_drv_alloc
-Memcheck:Leak
-...
-fun:erts_bin_drv_alloc
-...
-}
-
-{
-Halfword erts_bin_drv_alloc_fnf
-Memcheck:Leak
-...
-fun:erts_bin_drv_alloc_fnf
-...
-}
-
-{
-Halfword erts_create_magic_binary
-Memcheck:Leak
-...
-fun:erts_create_magic_binary
-...
-}
diff --git a/erts/epmd/src/Makefile.in b/erts/epmd/src/Makefile.in
index b6e3ba7762..1266be44cb 100644
--- a/erts/epmd/src/Makefile.in
+++ b/erts/epmd/src/Makefile.in
@@ -19,10 +19,6 @@
 #
 include $(ERL_TOP)/make/target.mk
 
-ifeq ($(findstring ose,$(TARGET)),ose)
-include $(ERL_TOP)/make/$(TARGET)/ose_lm.mk
-endif
-
 ifeq ($(TYPE),debug)
 PURIFY     =
 TYPEMARKER = .debug
@@ -32,21 +28,13 @@ else
 ifeq ($(TYPE),purify)
 PURIFY     = purify
 TYPEMARKER =
-ifeq ($(findstring ose,$(TARGET)),ose)
-  TYPE_FLAGS = -DPURIFY
-else
-  TYPE_FLAGS = -O2 -DPURIFY
-endif
+TYPE_FLAGS = -O2 -DPURIFY
 else
 
 override TYPE = opt
 PURIFY     =
 TYPEMARKER =
-ifeq ($(findstring ose,$(TARGET)),ose)
-  TYPE_FLAGS =
-else
-  TYPE_FLAGS = -O2
-endif
+TYPE_FLAGS = -O2
 endif
 endif
 
@@ -68,13 +56,9 @@ else
 ifeq ($(findstring vxworks,$(TARGET)),vxworks)
 ERTS_INTERNAL_LIBS=-L../../lib/internal/$(TARGET) -lerts_internal$(ERTS_LIB_TYPEMARKER) @ERTS_INTERNAL_X_LIBS@
 else
-ifeq ($(findstring ose,$(TARGET)),ose)
-ERTS_INTERNAL_LIBS=-L../../lib/internal/$(TARGET) -lerts_internal$(ERTS_LIB_TYPEMARKER) @ERTS_INTERNAL_X_LIBS@
-else
 ERTS_INTERNAL_LIBS=-L../../lib/internal/$(TARGET) -lerts_internal$(ERTS_LIB_TYPEMARKER) @ERTS_INTERNAL_X_LIBS@ -lm
 endif
 endif
-endif
 
 ERTS_LIB = $(ERL_TOP)/erts/lib_src/obj/$(TARGET)/$(TYPE)/MADE
 
@@ -82,11 +66,7 @@ CC      = @CC@
 WFLAGS  = @WFLAGS@
 CFLAGS  = @CFLAGS@ @DEFS@ $(TYPE_FLAGS) $(WFLAGS) $(ERTS_INCL)
 LD      = @LD@
-ifeq ($(findstring ose,$(TARGET)),ose)
-LIBS    = $(ERTS_INTERNAL_LIBS) @LIBS@
-else
 LIBS    = @LIBS@ @SYSTEMD_DAEMON_LIBS@ $(ERTS_INTERNAL_LIBS)
-endif
 LDFLAGS = @LDFLAGS@
 
 
@@ -135,25 +115,12 @@ clean:
 	rm -f *.o
 	rm -f *~ core
 
-ifeq ($(findstring ose,$(TARGET)),ose)
-$(OBJDIR)/ose_confd.o: $(OSE_CONFD)
-	$(V_CC) $(CFLAGS) -o $@ -c $<
-$(OBJDIR)/crt0_lm.o: $(CRT0_LM)
-	$(V_CC) $(CFLAGS) -o $@ -c $<
-OSE_LM_OBJS += $(OBJDIR)/ose_confd.o $(OBJDIR)/crt0_lm.o
-endif
-
 #
 # Objects & executables
 #
 
-ifeq ($(findstring ose,$(TARGET)),ose)
-$(BINDIR)/$(EPMD): $(EPMD_OBJS) $(ERTS_LIB) $(OSE_LM_OBJS)
-	$(call build-ose-load-module, $@, $(EPMD_OBJS) $(OSE_LM_OBJS), $(LIBS), $(EPMD_LMCONF))
-else
 $(BINDIR)/$(EPMD): $(EPMD_OBJS) $(ERTS_LIB)
 	$(ld_verbose)$(PURIFY) $(LD) $(LDFLAGS) -o $@ $(EPMD_OBJS) $(LIBS)
-endif
 
 $(OBJDIR)/%.o: %.c epmd.h epmd_int.h
 	$(V_CC) $(CFLAGS) $(EPMD_FLAGS) -o $@ -c $<
diff --git a/erts/epmd/src/epmd.c b/erts/epmd/src/epmd.c
index 132bda725c..21ce4e52b0 100644
--- a/erts/epmd/src/epmd.c
+++ b/erts/epmd/src/epmd.c
@@ -29,7 +29,7 @@
 #ifdef HAVE_STDLIB_H
 #  include <stdlib.h>
 #endif
-
+#include <time.h>
 /* forward declarations */
 
 static void usage(EpmdVars *);
@@ -397,7 +397,7 @@ static void run_daemon(EpmdVars *g)
 }
 #endif
 
-#if defined(VXWORKS) || defined(__OSE__)
+#if defined(VXWORKS)
 static void run_daemon(EpmdVars *g)
 {
     run(g);
diff --git a/erts/epmd/src/epmd_int.h b/erts/epmd/src/epmd_int.h
index 26100afc93..e222abb4b7 100644
--- a/erts/epmd/src/epmd_int.h
+++ b/erts/epmd/src/epmd_int.h
@@ -37,13 +37,6 @@
 #define DONT_USE_MAIN
 #endif
 
-#ifdef __OSE__
-#  define NO_DAEMON
-#  define NO_SYSLOG
-#  define NO_SYSCONF
-#  define NO_FCNTL
-#endif
-
 /* ************************************************************************ */
 /* Standard includes                                                        */
 
@@ -100,12 +93,7 @@
 #endif /* ! WIN32 */
 
 #include <ctype.h>
-
-#if !defined(__OSE__)
-#  include <signal.h>
-#endif
-
-
+#include <signal.h>
 #include <errno.h>
 
 #ifdef HAVE_SYSLOG_H
@@ -122,10 +110,6 @@
 
 #include <stdarg.h>
 
-#ifdef __OSE__
-#  include "sys/select.h"
-#endif
-
 #ifdef HAVE_SYSTEMD_DAEMON
 #  include <systemd/sd-daemon.h>
 #endif /* HAVE_SYSTEMD_DAEMON */
diff --git a/erts/epmd/src/epmd_srv.c b/erts/epmd/src/epmd_srv.c
index 8c8d7304f2..20ad0af829 100644
--- a/erts/epmd/src/epmd_srv.c
+++ b/erts/epmd/src/epmd_srv.c
@@ -30,11 +30,6 @@
 #  define INADDR_NONE 0xffffffff
 #endif
 
-#if defined(__OSE__)
-#  include "sys/ioctl.h"
-#  define sleep(x) delay(x*1000)
-#endif
-
 /*
  *  
  *  This server is a local name server for Erlang nodes. Erlang nodes can
@@ -315,7 +310,7 @@ void run(EpmdVars *g)
     }
 #endif /* HAVE_SYSTEMD_DAEMON */
 
-#if !defined(__WIN32__) && !defined(__OSE__)
+#if !defined(__WIN32__)
   /* We ignore the SIGPIPE signal that is raised when we call write
      twice on a socket closed by the other end. */
   signal(SIGPIPE, SIG_IGN);
@@ -700,9 +695,6 @@ static void do_request(g, fd, s, buf, bsize)
 	    put_int16(node->creation, wbuf+2);
 	}
   
-	if (g->delay_write)		/* Test of busy server */
-	  sleep(g->delay_write);
-
 	if (reply(g, fd, wbuf, 4) != 4)
 	  {
 	    dbg_tty_printf(g,1,"** failed to send ALIVE2_RESP for \"%s\"",
diff --git a/erts/etc/common/Makefile.in b/erts/etc/common/Makefile.in
index 8e55fa78c9..05d925f19f 100644
--- a/erts/etc/common/Makefile.in
+++ b/erts/etc/common/Makefile.in
@@ -21,10 +21,6 @@
 include $(ERL_TOP)/make/output.mk
 include $(ERL_TOP)/make/target.mk
 
-ifeq ($(findstring ose,$(TARGET)),ose)
-include $(ERL_TOP)/make/$(TARGET)/ose_lm.mk
-endif
-
 ERTS_LIB_TYPEMARKER=.$(TYPE)
 
 USING_MINGW=@MIXED_CYGWIN_MINGW@
@@ -85,18 +81,13 @@ EMUOSDIR  = $(ERL_TOP)/erts/emulator/@ERLANG_OSTYPE@
 SYSDIR    = $(ERL_TOP)/erts/emulator/sys/@ERLANG_OSTYPE@
 DRVDIR    = $(ERL_TOP)/erts/emulator/drivers/@ERLANG_OSTYPE@
 UXETC     = ../unix
-OSEETC	  = ../ose
 WINETC	  = ../win32
 
 ifeq ($(TARGET), win32)
 ETC       = $(WINETC)
 else
-ifeq ($(findstring ose,$(TARGET)),ose)
-ETC       = $(OSEETC)
-else
 ETC       = $(UXETC)
 endif
-endif
 
 ifeq ($(TARGET), win32)
 ERLEXEC = erlexec.dll
@@ -180,25 +171,6 @@ PORT_ENTRY_POINT=erl_port_entry
 ENTRY_LDFLAGS=-entry:$(PORT_ENTRY_POINT)
 
 else
-ifeq ($(findstring ose,$(TARGET)),ose)
-ENTRY_LDFLAGS=
-ENTRY_OBJ=
-ERLSRV_OBJECTS=
-MC_OUTPUTS=
-INET_GETHOST =
-INSTALL_EMBEDDED_PROGS = $(BINDIR)/run_erl_lm
-INSTALL_EMBEDDED_DATA =
-INSTALL_TOP = Install
-INSTALL_TOP_BIN =
-INSTALL_MISC =
-INSTALL_SRC =
-ERLEXECDIR = .
-INSTALL_LIBS =
-INSTALL_OBJS =
-INSTALL_INCLUDES =
-TEXTFILES = Install erl.src
-INSTALL_PROGS = $(INSTALL_EMBEDDED_PROGS)
-else # UNIX (!win32 && !ose)
 ENTRY_LDFLAGS=
 ENTRY_OBJ=
 ERLSRV_OBJECTS= 
@@ -223,7 +195,6 @@ INSTALL_PROGS =              \
 	$(BINDIR)/$(ERLEXEC) \
 	$(INSTALL_EMBEDDED_PROGS)
 endif
-endif
 
 .PHONY: etc
 etc:	$(ENTRY_OBJ) $(INSTALL_PROGS) $(INSTALL_LIBS) $(TEXTFILES) $(INSTALL_TOP_BIN)
@@ -269,8 +240,8 @@ endif
 	rm -f $(ERL_TOP)/erts/obj*/$(TARGET)/to_erl.o
 	rm -f $(ERL_TOP)/erts/obj*/$(TARGET)/dyn_erl.o
 	rm -f $(ERL_TOP)/erts/obj*/$(TARGET)/safe_string.o
-	rm -f $(ERL_TOP)/erts/obj*/$(TARGET)/run_erl_common.o
-	rm -f $(ERL_TOP)/erts/obj*/$(TARGET)/to_erl_common.o
+	rm -f $(ERL_TOP)/erts/obj*/$(TARGET)/run_erl.o
+	rm -f $(ERL_TOP)/erts/obj*/$(TARGET)/to_erl.o
 	rm -f $(ERL_TOP)/erts/obj*/$(TARGET)/typer.o
 	rm -f $(ERL_TOP)/erts/obj*/$(TARGET)/ct_run.o
 	rm -f $(ERL_TOP)/erts/obj*/$(TARGET)/vxcall.o
@@ -423,28 +394,24 @@ $(BINDIR)/inet_gethost@EXEEXT@: $(OBJDIR)/inet_gethost.o $(ENTRY_OBJ) $(ERTS_LIB
 	$(ld_verbose)$(PURIFY) $(LD) $(LDFLAGS) $(ENTRY_LDFLAGS) -o $@ $(OBJDIR)/inet_gethost.o $(ENTRY_OBJ) $(LIBS) $(ERTS_INTERNAL_LIBS)
 
 # run_erl
-$(BINDIR)/run_erl: $(OBJDIR)/safe_string.o $(OBJDIR)/run_erl.o $(OBJDIR)/run_erl_common.o
-	$(V_LD) $(LDFLAGS) -o $@ $^ $(LIBS)
-$(OBJDIR)/run_erl.o: $(ETC)/run_erl.c ../common/run_erl_common.h $(RC_GENERATED)
-	$(V_CC) $(CFLAGS) -I ../common/ -o $@ -c $(ETC)/run_erl.c
-$(OBJDIR)/run_erl_common.o: ../common/run_erl_common.c ../common/run_erl_common.h $(RC_GENERATED)
-	$(V_CC) $(CFLAGS) -o $@ -c $<
+$(BINDIR)/run_erl: $(OBJDIR)/safe_string.o $(OBJDIR)/run_erl.o
+	$(V_LD) $(LDFLAGS) -o $@ $(OBJDIR)/safe_string.o $(OBJDIR)/run_erl.o $(LIBS)
+$(OBJDIR)/run_erl.o: $(ETC)/run_erl.c $(RC_GENERATED)
+	$(V_CC) $(CFLAGS) -o $@ -c $(ETC)/run_erl.c
 
 # to_erl
-$(BINDIR)/to_erl: $(OBJDIR)/safe_string.o $(OBJDIR)/to_erl.o $(OBJDIR)/to_erl_common.o
-	$(V_LD) $(LDFLAGS) -o $@  $^
-$(OBJDIR)/to_erl.o: $(ETC)/to_erl.c ../common/safe_string.h $(RC_GENERATED)
-	$(V_CC) $(CFLAGS) -I ../common/ -o $@ -c $(ETC)/to_erl.c
-$(OBJDIR)/to_erl_common.o: ../common/to_erl_common.c ../common/to_erl_common.h $(RC_GENERATED)
-	$(V_CC) $(CFLAGS) -o $@ -c $<
+$(BINDIR)/to_erl: $(OBJDIR)/safe_string.o $(OBJDIR)/to_erl.o
+	$(V_LD) $(LDFLAGS) -o $@  $(OBJDIR)/safe_string.o $(OBJDIR)/to_erl.o
+$(OBJDIR)/to_erl.o: $(ETC)/to_erl.c $(RC_GENERATED)
+	$(V_CC) $(CFLAGS) -o $@ -c $(ETC)/to_erl.c
 
 # dyn_erl
 $(BINDIR)/dyn_erl: $(OBJDIR)/safe_string.o $(OBJDIR)/dyn_erl.o
 	$(V_LD) $(LDFLAGS) -o $@  $(OBJDIR)/safe_string.o $(OBJDIR)/dyn_erl.o
 $(OBJDIR)/dyn_erl.o: $(UXETC)/dyn_erl.c $(RC_GENERATED)
 	$(V_CC) $(CFLAGS) -o $@ -c $(UXETC)/dyn_erl.c
-$(OBJDIR)/safe_string.o: ../common/safe_string.c $(RC_GENERATED)
-	$(V_CC) $(CFLAGS) -o $@ -c ../common/safe_string.c
+$(OBJDIR)/safe_string.o: $(ETC)/safe_string.c $(RC_GENERATED)
+	$(V_CC) $(CFLAGS) -o $@ -c $(ETC)/safe_string.c
 
 ifneq ($(TARGET),win32)
 $(BINDIR)/$(ERLEXEC): $(OBJDIR)/$(ERLEXEC).o $(ERTS_LIB)
@@ -499,30 +466,6 @@ erl.src: $(UXETC)/erl.src.src ../../vsn.mk $(TARGET)/Makefile
             -e 's;%VSN%;$(VSN);' \
                   $(UXETC)/erl.src.src > erl.src
 
-#---------------------------------------------------------
-# OSE specific targets
-#---------------------------------------------------------
-ifeq ($(findstring ose,$(TARGET)),ose)
-$(OBJDIR)/ose_confd.o: $(OSE_CONFD)
-	$(V_CC) $(CFLAGS) -o $@ -c $<
-$(OBJDIR)/crt0_lm.o: $(CRT0_LM)
-	$(V_CC) $(CFLAGS) -o $@ -c $<
-OSE_LM_OBJS += $(OBJDIR)/ose_confd.o $(OBJDIR)/crt0_lm.o
-
-$(BINDIR)/run_erl_lm: $(OBJDIR)/run_erl_main.o $(OBJDIR)/safe_string.o $(OBJDIR)/run_erl.o $(OBJDIR)/run_erl_common.o $(OBJDIR)/to_erl_common.o $(OSE_LM_OBJS)
-	$(call build-ose-load-module, $@, $^, $(LIBS), $(RUN_ERL_LMCONF))
-
-
-$(OBJDIR)/run_erl_main.o: $(OSEETC)/run_erl_main.c $(OSEETC)/run_erl.h ../common/to_erl_common.h $(RC_GENERATED)
-	$(V_CC) $(CFLAGS) -I ../common/ -o $@ -c $(OSEETC)/run_erl_main.c
-
-endif
-
-#---------------------------------------------------------
-# End of ose specific targets.
-#---------------------------------------------------------
-
-
 # ----------------------------------------------------
 # Release Target
 # ---------------------------------------------------- 
@@ -537,11 +480,9 @@ endif
 	$(INSTALL_DIR) "$(RELEASE_PATH)/erts-$(VSN)/bin"
 ifneq ($(TARGET), win32)
 ifneq ($(findstring vxworks,$(TARGET)), vxworks)
-ifneq ($(findstring ose,$(TARGET)), ose)
 	$(INSTALL_SCRIPT) erl.src "$(RELEASE_PATH)/erts-$(VSN)/bin"
 endif
 endif
-endif
 ifneq ($(INSTALL_PROGS),)
 	$(INSTALL_PROGRAM) $(INSTALL_PROGS) "$(RELEASE_PATH)/erts-$(VSN)/bin"
 endif
diff --git a/erts/etc/common/erlexec.c b/erts/etc/common/erlexec.c
index cde0b25a2a..f21671e837 100644
--- a/erts/etc/common/erlexec.c
+++ b/erts/etc/common/erlexec.c
@@ -73,6 +73,7 @@ static const char plusM_au_allocs[]= {
     'R',	/* driver_alloc		*/
     'S',	/* sl_alloc		*/
     'T',	/* temp_alloc		*/
+    'Z',        /* test_alloc           */
     '\0'
 };
 
@@ -155,6 +156,12 @@ static char *plusr_val_switches[] = {
     NULL
 };
 
+/* +x arguments with values */
+static char *plusx_val_switches[] = {
+    "mqd",
+    NULL
+};
+
 /* +z arguments with values */
 static char *plusz_val_switches[] = {
     "dbbl",
@@ -718,7 +725,7 @@ int main(int argc, char **argv)
 		     * on itself here.  We'll avoid doing that.
 		     */
 		    if (strcmp(argv[i], "-make") == 0) {
-			add_args("-noshell", "-noinput", "-s", "make", "all", NULL);
+			add_args("-noshell", "-noinput", "-s", "make", "all_or_nothing", NULL);
 			add_Eargs("-B");
 			haltAfterwards = 1;
 			i = argc; /* Skip rest of command line */
@@ -975,6 +982,20 @@ int main(int argc, char **argv)
 		      add_Eargs(argv[i+1]);
 		      i++;
 		      break;
+		  case 'x':
+		      if (!is_one_of_strings(&argv[i][2], plusx_val_switches)) {
+			  goto the_default;
+		      } else {
+			  if (i+1 >= argc
+			      || argv[i+1][0] == '-'
+			      || argv[i+1][0] == '+')
+			      usage(argv[i]);
+			  argv[i][0] = '-';
+			  add_Eargs(argv[i]);
+			  add_Eargs(argv[i+1]);
+			  i++;
+		      }
+		      break;
 		  case 'z':
 		      if (!is_one_of_strings(&argv[i][2], plusz_val_switches)) {
 			  goto the_default;
@@ -1175,7 +1196,7 @@ usage_aux(void)
 	  "[+S NO_SCHEDULERS:NO_SCHEDULERS_ONLINE] "
 	  "[+SP PERCENTAGE_SCHEDULERS:PERCENTAGE_SCHEDULERS_ONLINE] "
 	  "[+T LEVEL] [+V] [+v] "
-	  "[+W<i|w|e>] [+z MISC_OPTION] [args ...]\n");
+	  "[+W<i|w|e>]  [+x DEFAULT_PROC_FLAGS] [+z MISC_OPTION] [args ...]\n");
   exit(1);
 }
 
diff --git a/erts/etc/common/heart.c b/erts/etc/common/heart.c
index 01ef840b5d..9571b83ffd 100644
--- a/erts/etc/common/heart.c
+++ b/erts/etc/common/heart.c
@@ -718,14 +718,12 @@ do_terminate(int erlin_fd, int reason) {
 		print_error("Would reboot. Terminating.");
 	    else {
 		kill_old_erlang();
-		/* suppress gcc warning with 'if' */
 		ret = system(command);
 		print_error("Executed \"%s\" -> %d. Terminating.",command, ret);
 	    }
 	    free_env_val(command);
 	} else {
 	    kill_old_erlang();
-	    /* suppress gcc warning with 'if' */
 	    ret = system((char*)&cmd[0]);
 	    print_error("Executed \"%s\" -> %d. Terminating.",cmd, ret);
 	}
diff --git a/erts/etc/common/run_erl_common.c b/erts/etc/common/run_erl_common.c
deleted file mode 100644
index c03d5e3ae2..0000000000
--- a/erts/etc/common/run_erl_common.c
+++ /dev/null
@@ -1,696 +0,0 @@
-/*
- * %CopyrightBegin%
- *
- * Copyright Ericsson AB 2014. All Rights Reserved.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- *
- * %CopyrightEnd%
- */
-#ifdef HAVE_CONFIG_H
-#  include "config.h"
-#endif
-
-#include <dirent.h>
-#include <errno.h>
-#include <fcntl.h>
-#include <stdarg.h>
-#include <stdio.h>
-#include <stdlib.h>
-#include <string.h>
-#include <sys/stat.h>
-#include <sys/types.h>
-#include <time.h>
-#include <unistd.h>
-
-#ifdef __ANDROID__
-#  include <termios.h>
-#endif
-
-#ifdef HAVE_SYSLOG_H
-#  include <syslog.h>
-#endif
-
-#ifdef HAVE_SYS_IOCTL_H
-#  include <sys/ioctl.h>
-#endif
-
-#ifdef __OSE__
-#  include "ramlog.h"
-#endif
-
-#include "run_erl_common.h"
-#include "safe_string.h"
-
-#define DEFAULT_LOG_GENERATIONS 5
-#define LOG_MAX_GENERATIONS     1000      /* No more than 1000 log files */
-#define LOG_MIN_GENERATIONS     2         /* At least two to switch between */
-#define DEFAULT_LOG_MAXSIZE     100000
-#define LOG_MIN_MAXSIZE         1000      /* Smallast value for changing log file */
-#define LOG_STUBNAME            "erlang.log."
-#define LOG_PERM                0664
-#define DEFAULT_LOG_ACTIVITY_MINUTES    5
-#define DEFAULT_LOG_ALIVE_MINUTES       15
-#define DEFAULT_LOG_ALIVE_FORMAT        "%a %b %e %T %Z %Y"
-#define ALIVE_BUFFSIZ                   1024
-
-#define STATUSFILENAME  "/run_erl.log"
-
-#define PIPE_STUBNAME   "erlang.pipe"
-#define PIPE_STUBLEN    strlen(PIPE_STUBNAME)
-#define PERM            (S_IWUSR | S_IRUSR | S_IWOTH | S_IROTH | S_IWGRP | S_IRGRP)
-
-/* OSE has defined O_SYNC but it is not recognized by open */
-#if !defined(O_SYNC) || defined(__OSE__)
-#undef O_SYNC
-#define O_SYNC 0
-#define USE_FSYNC 1
-#endif
-
-/* Global variable definitions
- * We need this complex way of handling global variables because of how
- * OSE works here. We want to make it possible to run the shell command
- * run_erl multiple times with different global variables without them
- * effecting eachother.
- */
-
-#define STATUSFILE           (RE_DATA->statusfile)
-#define LOG_DIR              (RE_DATA->log_dir)
-#define STDSTATUS            (RE_DATA->stdstatus)
-#define LOG_GENERATIONS      (RE_DATA->log_generations)
-#define LOG_MAXSIZE          (RE_DATA->log_maxsize)
-#define LOG_ACTIVITY_MINUTES (RE_DATA->log_activity_minutes)
-#define LOG_ALIVE_IN_GMT     (RE_DATA->log_alive_in_gmt)
-#define LOG_ALIVE_FORMAT     (RE_DATA->log_alive_format)
-#define RUN_DAEMON           (RE_DATA->run_daemon)
-#define LOG_ALIVE_MINUTES    (RE_DATA->log_alive_minutes)
-#define LOG_NUM              (RE_DATA->log_num)
-#define LFD                  (RE_DATA->lfd)
-#define PROTOCOL_VER         (RE_DATA->protocol_ver)
-
-struct run_erl_ {
-  /* constant config data */
-  char statusfile[FILENAME_BUFSIZ];
-  char log_dir[FILENAME_BUFSIZ];
-  FILE *stdstatus;
-  int log_generations;
-  int log_maxsize;
-  int log_activity_minutes;
-  int log_alive_in_gmt;
-  char log_alive_format[ALIVE_BUFFSIZ+1];
-  int run_daemon;
-  int log_alive_minutes;
-  /* Current log number and log fd */
-  int log_num;
-  int lfd;
-  unsigned protocol_ver;
-};
-
-typedef struct run_erl_ run_erl;
-
-#ifdef __OSE__
-static OSPPDKEY run_erl_pp_key;
-#define RE_DATA (*(run_erl**)ose_get_ppdata(run_erl_pp_key))
-#else
-static run_erl re;
-#define RE_DATA (&re)
-#endif
-
-/* prototypes */
-
-static int next_log(int log_num);
-static int prev_log(int log_num);
-static int find_next_log_num(void);
-static int open_log(int log_num, int flags);
-
-/*
- * getenv_int:
- */
-static char *getenv_int(const char *name) {
-#ifdef __OSE__
-   return get_env(get_bid(current_process()),name);
-#else
-   return getenv(name);
-#endif
-}
-
-/*
- * next_log:
- * Returns the index number that follows the given index number.
- * (Wrapping after log_generations)
- */
-static int next_log(int log_num) {
-  return log_num>=LOG_GENERATIONS?1:log_num+1;
-}
-
-/*
- * prev_log:
- * Returns the index number that precedes the given index number.
- * (Wrapping after log_generations)
- */
-static int prev_log(int log_num) {
-  return log_num<=1?LOG_GENERATIONS:log_num-1;
-}
-
-/*
- * find_next_log_num()
- * Searches through the log directory to check which logs that already
- * exist. It finds the "hole" in the sequence, and returns the index
- * number for the last log in the log sequence. If there is no hole, index
- * 1 is returned.
- */
-static int find_next_log_num(void) {
-  int i, next_gen, log_gen;
-  DIR *dirp;
-  struct dirent *direntp;
-  int log_exists[LOG_MAX_GENERATIONS+1];
-  int stub_len = strlen(LOG_STUBNAME);
-
-  /* Initialize exiting log table */
-
-  for(i=LOG_GENERATIONS; i>=0; i--)
-    log_exists[i] = 0;
-  dirp = opendir(LOG_DIR);
-  if(!dirp) {
-    ERRNO_ERR1(LOG_ERR,"Can't access log directory '%s'", LOG_DIR);
-    exit(1);
-  }
-
-  /* Check the directory for existing logs */
-
-  while((direntp=readdir(dirp)) != NULL) {
-    if(strncmp(direntp->d_name,LOG_STUBNAME,stub_len)==0) {
-      int num = atoi(direntp->d_name+stub_len);
-      if(num < 1 || num > LOG_GENERATIONS)
-	continue;
-      log_exists[num] = 1;
-    }
-  }
-  closedir(dirp);
-
-  /* Find out the next available log file number */
-
-  next_gen = 0;
-  for(i=LOG_GENERATIONS; i>=0; i--) {
-    if(log_exists[i])
-      if(next_gen)
-	break;
-      else
-	;
-    else
-      next_gen = i;
-  }
-
-  /* Find out the current log file number */
-
-  if(next_gen)
-    log_gen = prev_log(next_gen);
-  else
-    log_gen = 1;
-
-  return log_gen;
-} /* find_next_log_num() */
-
-static int open_log(int log_num, int flags)
-{
-  char buf[FILENAME_MAX];
-  time_t now;
-  struct tm *tmptr;
-  char log_buffer[ALIVE_BUFFSIZ+1];
-
-  /* Remove the next log (to keep a "hole" in the log sequence) */
-  sn_printf(buf, sizeof(buf), "%s/%s%d",
-	    LOG_DIR, LOG_STUBNAME, next_log(log_num));
-  unlink(buf);
-
-  /* Create or continue on the current log file */
-  sn_printf(buf, sizeof(buf), "%s/%s%d", LOG_DIR, LOG_STUBNAME, log_num);
-
-  LFD = sf_open(buf, flags, LOG_PERM);
-
-  if(LFD <0){
-      ERRNO_ERR1(LOG_ERR,"Can't open log file '%s'.", buf);
-    exit(1);
-  }
-
-  /* Write a LOGGING STARTED and time stamp into the log file */
-  time(&now);
-  if (LOG_ALIVE_IN_GMT) {
-      tmptr = gmtime(&now);
-  } else {
-      tmptr = localtime(&now);
-  }
-  if (!strftime(log_buffer, ALIVE_BUFFSIZ, LOG_ALIVE_FORMAT,
-		tmptr)) {
-      strn_cpy(log_buffer, sizeof(log_buffer),
-	      "(could not format time in 256 positions "
-	      "with current format string.)");
-  }
-  log_buffer[ALIVE_BUFFSIZ] = '\0';
-
-  sn_printf(buf, sizeof(buf), "\n=====\n===== LOGGING STARTED %s\n=====\n",
-	    log_buffer);
-  if (erts_run_erl_write_all(LFD, buf, strlen(buf)) < 0)
-      erts_run_erl_log_status("Error in writing to log.\n");
-
-#if USE_FSYNC
-  fsync(LFD);
-#endif
-
-  return LFD;
-}
-
-/* Instead of making sure basename exists, we do our own */
-char *simple_basename(char *path)
-{
-    char *ptr;
-    for (ptr = path; *ptr != '\0'; ++ptr) {
-	if (*ptr == '/') {
-	    path = ptr + 1;
-	}
-    }
-    return path;
-}
-
-ssize_t sf_read(int fd, void *buffer, size_t len) {
-    ssize_t n = 0;
-
-    do { n = read(fd, buffer, len); } while (n < 0 && errno == EINTR);
-
-    return n;
-}
-
-ssize_t sf_write(int fd, const void *buffer, size_t len) {
-    ssize_t n = 0;
-
-    do { n = write(fd, buffer, len); } while (n < 0 && errno == EINTR);
-
-    return n;
-}
-
-int sf_open(const char *path, int type, mode_t mode) {
-    int fd = 0;
-
-    do { fd = open(path, type, mode); } while(fd < 0 && errno == EINTR);
-
-    return fd;
-}
-
-int sf_close(int fd) {
-    int res = 0;
-
-    do { res = close(fd); } while(res < 0 && errno == EINTR);
-
-    return res;
-}
-
-/* Call write() until entire buffer has been written or error.
- * Return len or -1.
- */
-int erts_run_erl_write_all(int fd, const char* buf, int len)
-{
-    int left = len;
-    int written;
-    for (;;) {
-        do {
-	  written = write(fd,buf,left);
-	} while (written < 0 && errno == EINTR);
-	if (written == left) {
-	    return len;
-	}
-	if (written < 0) {
-	    return -1;
-	}
-	left -= written;
-	buf += written;
-    }
-    return written;
-}
-
-/* erts_run_erl_log_status()
- * Prints the arguments to a status file
- * Works like printf (see vfrpintf)
- */
-void erts_run_erl_log_status(const char *format,...)
-{
-  va_list args;
-  time_t now;
-
-  if (STDSTATUS == NULL)
-    STDSTATUS = fopen(STATUSFILE, "w");
-  if (STDSTATUS == NULL)
-    return;
-  now = time(NULL);
-  fprintf(STDSTATUS, "run_erl [%d] %s",
-#ifdef __OSE__
-	  (int)current_process(),
-#else
-	  (int)getpid(),
-#endif
-	  ctime(&now));
-  va_start(args, format);
-  vfprintf(STDSTATUS, format, args);
-  va_end(args);
-  fflush(STDSTATUS);
-  return;
-}
-
-/* Fetch the current log alive minutes */
-int erts_run_erl_log_alive_minutes() {
-  return LOG_ALIVE_MINUTES;
-}
-
-/* error_logf()
- * Prints the arguments to stderr or syslog
- * Works like printf (see vfprintf)
- */
-void erts_run_erl_log_error(int priority, int line, const char *format, ...)
-{
-    va_list args;
-    va_start(args, format);
-
-#ifdef HAVE_SYSLOG_H
-    if (RUN_DAEMON) {
-	vsyslog(priority,format,args);
-    }
-    else
-#endif
-#ifdef __OSE__
-    if (RUN_DAEMON) {
-      char *buff = malloc(sizeof(char)*1024);
-      vsnprintf(buff,1024,format, args);
-      ramlog_printf(buff);
-    }
-    else
-#endif
-    {
-	time_t now = time(NULL);
-	fprintf(stderr, "run_erl:%d [%d] %s", line,
-#ifdef __OSE__
-		(int)current_process(),
-#else
-		(int)getpid(),
-#endif
-		ctime(&now));
-	vfprintf(stderr, format, args);
-    }
-    va_end(args);
-}
-
-/* erts_run_erl_log_write()
- * Writes a message to lfd. If the current log file is full,
- * a new log file is opened.
- */
-int erts_run_erl_log_write(char* buf, size_t len)
-{
-  int size;
-  ssize_t res;
-  /* Decide if new logfile needed, and open if so */
-
-  size = lseek(LFD,0,SEEK_END);
-  if(size+len > LOG_MAXSIZE) {
-    int res;
-    do {
-      res = close(LFD);
-    } while (res < 0 && errno == EINTR);
-    LOG_NUM = next_log(LOG_NUM);
-    LFD = open_log(LOG_NUM, O_RDWR|O_CREAT|O_TRUNC|O_SYNC);
-  }
-
-  /* Write to log file */
-
-  if ((res = erts_run_erl_write_all(LFD, buf, len)) < 0) {
-    erts_run_erl_log_status("Error in writing to log.\n");
-  }
-
-#if USE_FSYNC
-  fsync(LFD);
-#endif
-  return res;
-}
-
-int erts_run_erl_log_activity(int timeout,time_t now,time_t last_activity) {
-  char log_alive_buffer[ALIVE_BUFFSIZ+1];
-  char buf[BUFSIZ];
-
-  if (timeout || now - last_activity > LOG_ACTIVITY_MINUTES*60) {
-    /* Either a time out: 15 minutes without action, */
-    /* or something is coming in right now, but it's a long time */
-    /* since last time, so let's write a time stamp this message */
-    struct tm *tmptr;
-    if (LOG_ALIVE_IN_GMT) {
-      tmptr = gmtime(&now);
-    } else {
-      tmptr = localtime(&now);
-    }
-    if (!strftime(log_alive_buffer, ALIVE_BUFFSIZ, LOG_ALIVE_FORMAT,
-		  tmptr)) {
-      strn_cpy(log_alive_buffer, sizeof(log_alive_buffer),
-	       "(could not format time in 256 positions "
-	       "with current format string.)");
-    }
-    log_alive_buffer[ALIVE_BUFFSIZ] = '\0';
-
-    sn_printf(buf, sizeof(buf), "\n===== %s%s\n",
-	      timeout?"ALIVE ":"", log_alive_buffer);
-    return erts_run_erl_log_write(buf, strlen(buf));
-  }
-  return 0;
-}
-
-int erts_run_erl_log_open() {
-
-  LOG_NUM = find_next_log_num();
-  LFD = open_log(LOG_NUM, O_RDWR|O_APPEND|O_CREAT|O_SYNC);
-  return 0;
-}
-
-int erts_run_erl_log_init(int daemon, char* logdir) {
-  char *p;
-
-#ifdef __OSE__
-  run_erl **re_pp;
-  if (!run_erl_pp_key)
-     ose_create_ppdata("run_erl_ppdata",&run_erl_pp_key);
-  re_pp = (run_erl **)ose_get_ppdata(run_erl_pp_key);
-  *re_pp = malloc(sizeof(run_erl));
-#endif
-
-  STDSTATUS = NULL;
-  LOG_GENERATIONS = DEFAULT_LOG_GENERATIONS;
-  LOG_MAXSIZE     = DEFAULT_LOG_MAXSIZE;
-  LOG_ACTIVITY_MINUTES = DEFAULT_LOG_ACTIVITY_MINUTES;
-  LOG_ALIVE_IN_GMT = 0;
-  RUN_DAEMON = 0;
-  LOG_ALIVE_MINUTES = DEFAULT_LOG_ALIVE_MINUTES;
-  LFD = 0;
-  PROTOCOL_VER = RUN_ERL_LO_VER; /* assume lowest to begin with */
-
-  /* Get values for LOG file handling from the environment */
-  if ((p = getenv_int("RUN_ERL_LOG_ALIVE_MINUTES"))) {
-      LOG_ALIVE_MINUTES = atoi(p);
-      if (!LOG_ALIVE_MINUTES) {
-	  ERROR1(LOG_ERR,"Minimum value for RUN_ERL_LOG_ALIVE_MINUTES is 1 "
-		 "(current value is %s)",p);
-      }
-      LOG_ACTIVITY_MINUTES = LOG_ALIVE_MINUTES / 3;
-      if (!LOG_ACTIVITY_MINUTES) {
-	  ++LOG_ACTIVITY_MINUTES;
-      }
-  }
-  if ((p = getenv_int(
-		   "RUN_ERL_LOG_ACTIVITY_MINUTES"))) {
-     LOG_ACTIVITY_MINUTES = atoi(p);
-      if (!LOG_ACTIVITY_MINUTES) {
-	  ERROR1(LOG_ERR,"Minimum value for RUN_ERL_LOG_ACTIVITY_MINUTES is 1 "
-		 "(current value is %s)",p);
-      }
-  }
-  if ((p = getenv_int("RUN_ERL_LOG_ALIVE_FORMAT"))) {
-      if (strlen(p) > ALIVE_BUFFSIZ) {
-	  ERROR1(LOG_ERR, "RUN_ERL_LOG_ALIVE_FORMAT can contain a maximum of "
-		 "%d characters", ALIVE_BUFFSIZ);
-      }
-      strn_cpy(LOG_ALIVE_FORMAT, sizeof(LOG_ALIVE_FORMAT), p);
-  } else {
-      strn_cpy(LOG_ALIVE_FORMAT, sizeof(LOG_ALIVE_FORMAT),
-	       DEFAULT_LOG_ALIVE_FORMAT);
-  }
-  if ((p = getenv_int("RUN_ERL_LOG_ALIVE_IN_UTC"))
-      && strcmp(p,"0")) {
-      ++LOG_ALIVE_IN_GMT;
-  }
-  if ((p = getenv_int("RUN_ERL_LOG_GENERATIONS"))) {
-    LOG_GENERATIONS = atoi(p);
-    if (LOG_GENERATIONS < LOG_MIN_GENERATIONS)
-      ERROR1(LOG_ERR,"Minimum RUN_ERL_LOG_GENERATIONS is %d",
-	     LOG_MIN_GENERATIONS);
-    if (LOG_GENERATIONS > LOG_MAX_GENERATIONS)
-      ERROR1(LOG_ERR,"Maximum RUN_ERL_LOG_GENERATIONS is %d",
-	     LOG_MAX_GENERATIONS);
-  }
-
-  if ((p = getenv_int("RUN_ERL_LOG_MAXSIZE"))) {
-    LOG_MAXSIZE = atoi(p);
-    if (LOG_MAXSIZE < LOG_MIN_MAXSIZE)
-      ERROR1(LOG_ERR,"Minimum RUN_ERL_LOG_MAXSIZE is %d", LOG_MIN_MAXSIZE);
-  }
-
-  RUN_DAEMON = daemon;
-
-  strn_cpy(LOG_DIR, sizeof(LOG_DIR), logdir);
-  strn_cpy(STATUSFILE, sizeof(STATUSFILE), LOG_DIR);
-  strn_cat(STATUSFILE, sizeof(STATUSFILE), STATUSFILENAME);
-
-  return 0;
-}
-
-/* create_fifo()
- * Creates a new fifo with the given name and permission.
- */
-static int create_fifo(char *name, int perm)
-{
-  if ((mkfifo(name, perm) < 0) && (errno != EEXIST))
-    return -1;
-  return 0;
-}
-
-/*
- * w- and r_pipename have to be pre-allocated of atleast FILENAME_MAX size
- */
-int erts_run_erl_open_fifo(char *pipename,char *w_pipename,char *r_pipename) {
-  int calculated_pipename = 0;
-  int highest_pipe_num = 0;
-  int fd;
-
-  /*
-   * Create FIFOs and open them
-   */
-
-  if(*pipename && pipename[strlen(pipename)-1] == '/') {
-    /* The user wishes us to find a unique pipe name in the specified */
-    /* directory */
-    DIR *dirp;
-    struct dirent *direntp;
-
-    calculated_pipename = 1;
-    dirp = opendir(pipename);
-    if(!dirp) {
-      ERRNO_ERR1(LOG_ERR,"Can't access pipe directory '%s'.", pipename);
-      return 1;
-    }
-
-    /* Check the directory for existing pipes */
-
-    while((direntp=readdir(dirp)) != NULL) {
-      if(strncmp(direntp->d_name,PIPE_STUBNAME,PIPE_STUBLEN)==0) {
-	int num = atoi(direntp->d_name+PIPE_STUBLEN+1);
-	if(num > highest_pipe_num)
-	  highest_pipe_num = num;
-      }
-    }
-    closedir(dirp);
-    strn_catf(pipename, BUFSIZ, "%s.%d",
-	      PIPE_STUBNAME, highest_pipe_num+1);
-  } /* if */
-
-  for(;;) {
-      /* write FIFO - is read FIFO for `to_erl' program */
-      strn_cpy(w_pipename, BUFSIZ, pipename);
-      strn_cat(w_pipename, BUFSIZ, ".r");
-      if (create_fifo(w_pipename, PERM) < 0) {
-	  ERRNO_ERR1(LOG_ERR,"Cannot create FIFO %s for writing.",
-		     w_pipename);
-	  return 1;
-      }
-
-      /* read FIFO - is write FIFO for `to_erl' program */
-      strn_cpy(r_pipename, BUFSIZ, pipename);
-      strn_cat(r_pipename, BUFSIZ, ".w");
-
-      /* Check that nobody is running run_erl already */
-      if ((fd = sf_open(r_pipename, O_WRONLY|DONT_BLOCK_PLEASE, 0)) >= 0) {
-	  /* Open as client succeeded -- run_erl is already running! */
-	  sf_close(fd);
-	  if (calculated_pipename) {
-	      ++highest_pipe_num;
-	      strn_catf(pipename, BUFSIZ, "%s.%d",
-			PIPE_STUBNAME, highest_pipe_num+1);
-	      continue;
-	  }
-	  ERROR1(LOG_ERR, "Erlang already running on pipe %s.\n", pipename);
-	  unlink(w_pipename);
-	  return 1;
-      }
-      if (create_fifo(r_pipename, PERM) < 0) {
-	  unlink(w_pipename);
-	  ERRNO_ERR1(LOG_ERR,"Cannot create FIFO %s for reading.",
-		     r_pipename);
-	  return 1;
-      }
-      break;
-  }
-  return 0;
-}
-
-/* Extract any control sequences that are ment only for run_erl
- * and should not be forwarded to the pty.
- */
-int erts_run_erl_extract_ctrl_seq(char* buf, int len, int mfd)
-{
-    static const char prefix[] = "\033_";
-    static const char suffix[] = "\033\\";
-    char* bufend = buf + len;
-    char* start = buf;
-    char* command;
-    char* end;
-
-    for (;;) {
-	start = find_str(start, bufend-start, prefix);
-	if (!start) break;
-
-	command = start + strlen(prefix);
-	end = find_str(command, bufend-command, suffix);
-	if (end) {
-	    unsigned col, row;
-	    if (sscanf(command,"version=%u", &PROTOCOL_VER)==1) {
-		/*fprintf(stderr,"to_erl v%u\n", protocol_ver);*/
-	    }
-	    else if (sscanf(command,"winsize=%u,%u", &col, &row)==2) {
-#ifdef TIOCSWINSZ
-	      struct winsize ws;
-	      ws.ws_col = col;
-	      ws.ws_row = row;
-	      if (ioctl(mfd, TIOCSWINSZ, &ws) < 0) {
-		ERRNO_ERR0(LOG_ERR,"Failed to set window size");
-	      }
-#endif
-	    }
-	    else {
-		ERROR2(LOG_ERR, "Ignoring unknown ctrl command '%.*s'\n",
-		       (int)(end-command), command);
-	    }
-
-	    /* Remove ctrl sequence from buf */
-	    end += strlen(suffix);
-	    memmove(start, end, bufend-end);
-	    bufend -= end - start;
-	}
-	else {
-	    ERROR2(LOG_ERR, "Missing suffix in ctrl sequence '%.*s'\n",
-		   (int)(bufend-start), start);
-	    break;
-	}
-    }
-    return bufend - buf;
-}
diff --git a/erts/etc/common/run_erl_common.h b/erts/etc/common/run_erl_common.h
deleted file mode 100644
index cecf7521f9..0000000000
--- a/erts/etc/common/run_erl_common.h
+++ /dev/null
@@ -1,97 +0,0 @@
-/*
- * %CopyrightBegin%
- *
- * Copyright Ericsson AB 2013. All Rights Reserved.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- *
- * %CopyrightEnd%
- */
-/*
- * Functions that are common to both OSE and unix implementations of run_erl
- */
-#ifndef ERL_RUN_ERL_LOG_H
-#define ERL_RUN_ERL_LOG_H
-
-#include <stdio.h>
-#include <time.h>
-#include <unistd.h>
-
-#include "run_erl_vsn.h"
-
-/* Log handling */
-int erts_run_erl_log_init(int run_daemon, char* logdir);
-int erts_run_erl_log_open(void);
-int erts_run_erl_log_close(void);
-int erts_run_erl_log_write(char *buff, size_t len);
-int erts_run_erl_log_activity(int timeout, time_t now, time_t last_activity);
-
-void erts_run_erl_log_status(const char *format,...);
-void erts_run_erl_log_error(int priority, int line, const char *format,...);
-
-int erts_run_erl_open_fifo(char *pipename,char *w_pipename,char *r_pipename);
-int erts_run_erl_log_alive_minutes(void);
-int erts_run_erl_extract_ctrl_seq(char* buf, int len, int mfd);
-
-/* File operations */
-ssize_t sf_read(int fd, void *buffer, size_t len);
-ssize_t sf_write(int fd, const void *buffer, size_t len);
-int sf_open(const char *path, int type, mode_t mode);
-int sf_close(int fd);
-int erts_run_erl_write_all(int fd, const char* buf, int len);
-char *simple_basename(char *path);
-
-#ifndef LOG_ERR
-#ifdef __OSE__
-#define LOG_ERR 0
-#else
-#define LOG_ERR NULL
-#endif
-#endif
-
-#define ERROR0(Prio,Format) erts_run_erl_log_error(Prio,__LINE__,Format"\n")
-#define ERROR1(Prio,Format,A1) erts_run_erl_log_error(Prio,__LINE__,Format"\n",A1)
-#define ERROR2(Prio,Format,A1,A2) erts_run_erl_log_error(Prio,__LINE__,Format"\n",A1,A2)
-
-#ifdef HAVE_STRERROR
-#    define ADD_ERRNO(Format) "errno=%d '%s'\n"Format"\n",errno,strerror(errno)
-#else
-#    define ADD_ERRNO(Format) "errno=%d\n"Format"\n",errno
-#endif
-#define ERRNO_ERR0(Prio,Format) erts_run_erl_log_error(Prio,__LINE__,ADD_ERRNO(Format))
-#define ERRNO_ERR1(Prio,Format,A1) erts_run_erl_log_error(Prio,__LINE__,ADD_ERRNO(Format),A1)
-#define ERRNO_ERR2(Prio,Format,A1,A2) erts_run_erl_log_error(Prio,__LINE__,ADD_ERRNO(Format),A1,A2)
-
-#define RUN_ERL_USAGE \
-  "%s (pipe_name|pipe_dir/) log_dir \"command [parameters ...]\"" \
-  "\n\nDESCRIPTION:\n"							\
-  "You may also set the environment variables RUN_ERL_LOG_GENERATIONS\n" \
-  "and RUN_ERL_LOG_MAXSIZE to the number of log files to use and the\n"	\
-  "size of the log file when to switch to the next log file\n"
-
-#ifndef FILENAME_MAX
-#define FILENAME_MAX 250
-#endif
-
-#define FILENAME_BUFSIZ FILENAME_MAX
-
-#ifdef O_NONBLOCK
-#  define DONT_BLOCK_PLEASE O_NONBLOCK
-#else
-#  define DONT_BLOCK_PLEASE O_NDELAY
-#  ifndef EAGAIN
-#    define EAGAIN -3898734
-#  endif
-#endif
-
-#endif
diff --git a/erts/etc/common/to_erl_common.c b/erts/etc/common/to_erl_common.c
deleted file mode 100644
index 8aa94ccfa4..0000000000
--- a/erts/etc/common/to_erl_common.c
+++ /dev/null
@@ -1,717 +0,0 @@
-/*
- * %CopyrightBegin%
- *
- * Copyright Ericsson AB 1996-2013. All Rights Reserved.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- *
- * %CopyrightEnd%
- */
-/*
- * Module: to_erl.c
- *
- * This module implements a process that opens two specified FIFOs, one
- * for reading and one for writing; reads from its stdin, and writes what
- * it has read to the write FIF0; reads from the read FIFO, and writes to
- * its stdout.
- *
-  ________                            _________
- |        |--<-- pipe.r (fifo1) --<--|         |
- | to_erl |                          | run_erl | (parent)
- |________|-->-- pipe.w (fifo2) -->--|_________|
-                                          ^ master pty
-                                          |
-                                          | slave pty
-                                      ____V____
-                                     |         |
-                                     |  "erl"  | (child)
-                                     |_________|
- */
-#ifdef HAVE_CONFIG_H
-#  include "config.h"
-#endif
-
-#include <sys/types.h>
-#include <sys/stat.h>
-#include <sys/time.h>
-#include <sys/types.h>
-#include <fcntl.h>
-#include <unistd.h>
-#include <stdio.h>
-#include <stdlib.h>
-#include <string.h>
-#include <dirent.h>
-#include <errno.h>
-
-#ifdef __OSE__
-#include <aio.h>
-#include "ose.h"
-#include "efs.h"
-#include "ose_spi/fm.sig"
-#else /* __UNIX__ */
-#include <termios.h>
-#include <signal.h>
-#endif
-
-#ifdef HAVE_SYS_IOCTL_H
-#  include <sys/ioctl.h>
-#endif
-
-#include "to_erl_common.h"
-#include "run_erl_vsn.h"
-#include "safe_string.h"   /* strn_cpy, strn_catf, sn_printf, etc. */
-
-#if defined(O_NONBLOCK)
-# define DONT_BLOCK_PLEASE O_NONBLOCK
-#else
-# define DONT_BLOCK_PLEASE O_NDELAY
-# if !defined(EAGAIN)
-#  define EAGAIN -3898734
-# endif
-#endif
-
-#ifdef HAVE_STRERROR
-#  define STRERROR(x) strerror(x)
-#else
-#  define STRERROR(x) ""
-#endif
-
-#define noDEBUG
-
-#ifdef __OSE__
-#define PIPE_DIR        "/pipe/"
-#else
-#define PIPE_DIR        "/tmp/"
-#endif
-#define PIPE_STUBNAME   "erlang.pipe"
-#define PIPE_STUBLEN    strlen(PIPE_STUBNAME)
-
-#ifdef DEBUG
-#define STATUS(s)  { fprintf(stderr, (s)); fflush(stderr); }
-#else
-#define STATUS(s)
-#endif
-
-#ifndef FILENAME_MAX
-#define FILENAME_MAX 250
-#endif
-
-static int tty_eof = 0;
-static int protocol_ver = RUN_ERL_LO_VER; /* assume lowest to begin with */
-
-static int write_all(int fd, const char* buf, int len);
-static int version_handshake(char* buf, int len, int wfd);
-
-
-#ifdef __OSE__
-
-#define SET_AIO(REQ,FD,SIZE,BUFF)					\
-  /* Make sure to clean data structure of previous request */		\
-  memset(&(REQ),0,sizeof(REQ));						\
-  (REQ).aio_fildes = FD;						\
-  (REQ).aio_offset = FM_POSITION_CURRENT;				\
-  (REQ).aio_nbytes = SIZE;						\
-  (REQ).aio_buf = BUFF;							\
-  (REQ).aio_sigevent.sigev_notify = SIGEV_NONE
-
-#define READ_AIO(REQ,FD,SIZE,BUFF)					\
-  SET_AIO(REQ,FD,SIZE,BUFF);						\
-  if (aio_read(&(REQ)) != 0)						\
-    fprintf(stderr,"aio_read of child_read_req(%d) failed"		\
-	    "with error %d\n",FD,errno)
-
-union SIGNAL {
-  SIGSELECT signo;
-  struct FmReadPtr fm_read_ptr;
-};
-
-#else /* __UNIX__ */
-static int recv_sig = 0;
-static struct termios tty_smode, tty_rmode;
-static int window_size_seq(char* buf, size_t bufsz);
-#ifdef DEBUG
-static void show_terminal_settings(struct termios *);
-#endif
-
-static void handle_ctrlc(int sig)
-{
-    /* Reinstall the handler, and signal break flag */
-    signal(SIGINT,handle_ctrlc);
-    recv_sig = SIGINT;
-}
-
-static void handle_sigwinch(int sig)
-{
-    recv_sig = SIGWINCH;
-}
-#endif
-
-static void usage(char *pname)
-{
-    fprintf(stderr, "Usage: ");
-    fprintf(stderr,TO_ERL_USAGE,pname);
-}
-
-int to_erl(int argc, char **argv)
-{
-    char  FIFO1[FILENAME_MAX], FIFO2[FILENAME_MAX];
-    int i, len, wfd, rfd;
-    char pipename[FILENAME_MAX];
-    int pipeIx = 1;
-    int force_lock = 0;
-    int got_some = 0;
-
-#ifdef __OSE__
-    struct aiocb stdin_read_req, pipe_read_req;
-    FmHandle stdin_fh, pipe_fh;
-    char *stdin_buf, *pipe_buf;
-    char *buf;
-    union SIGNAL *sig;
-#else /* __UNIX__ */
-    char buf[BUFSIZ];
-    fd_set readfds;
-#endif
-
-    if (argc >= 2 && argv[1][0]=='-') {
-	switch (argv[1][1]) {
-	case 'h':
-	    usage(argv[0]);
-	    exit(1);
-	case 'F':
-	    force_lock = 1;
-	    break;
-	default:
-	    fprintf(stderr,"Invalid option '%s'\n",argv[1]);
-	    exit(1);
-	}
-	pipeIx = 2;
-    }
-
-#ifdef DEBUG
-    fprintf(stderr, "%s: pid is : %d\n", argv[0],(int)
-#ifdef __OSE__
-	    current_process()
-#else /* __UNIX__ */
-	    getpid()
-#endif
-	    );
-#endif
-
-    strn_cpy(pipename, sizeof(pipename),
-	     (argv[pipeIx] ? argv[pipeIx] : PIPE_DIR));
-
-    if(*pipename && pipename[strlen(pipename)-1] == '/') {
-	/* The user wishes us to find a pipe name in the specified */
-	/* directory */
-	int highest_pipe_num = 0;
-	DIR *dirp;
-	struct dirent *direntp;
-
-	dirp = opendir(pipename);
-	if(!dirp) {
-	    fprintf(stderr, "Can't access pipe directory %s: %s\n", pipename, strerror(errno));
-	    exit(1);
-	}
-
-	/* Check the directory for existing pipes */
-
-	while((direntp=readdir(dirp)) != NULL) {
-	    if(strncmp(direntp->d_name,PIPE_STUBNAME,PIPE_STUBLEN)==0) {
-		int num = atoi(direntp->d_name+PIPE_STUBLEN+1);
-		if(num > highest_pipe_num)
-		    highest_pipe_num = num;
-	    }
-	}
-	closedir(dirp);
-	strn_catf(pipename, sizeof(pipename), (highest_pipe_num?"%s.%d":"%s"),
-		  PIPE_STUBNAME, highest_pipe_num);
-    } /* if */
-
-    /* read FIFO */
-    sn_printf(FIFO1,sizeof(FIFO1),"%s.r",pipename);
-    /* write FIFO */
-    sn_printf(FIFO2,sizeof(FIFO2),"%s.w",pipename);
-
-#ifndef __OSE__
-    /* Check that nobody is running to_erl on this pipe already */
-    if ((wfd = open (FIFO1, O_WRONLY|DONT_BLOCK_PLEASE, 0)) >= 0) {
-	/* Open as server succeeded -- to_erl is already running! */
-	close(wfd);
-	fprintf(stderr, "Another to_erl process already attached to pipe "
-			"%s.\n", pipename);
-	if (force_lock) {
-	    fprintf(stderr, "But we proceed anyway by force (-F).\n");
-	}
-	else {
-	    exit(1);
-	}
-    }
-#endif
-
-    if ((rfd = open (FIFO1, O_RDONLY|DONT_BLOCK_PLEASE, 0)) < 0) {
-#ifdef DEBUG
-	fprintf(stderr, "Could not open FIFO %s for reading.\n", FIFO1);
-#endif
-	fprintf(stderr, "No running Erlang on pipe %s: %s\n", pipename, strerror(errno));
-	exit(1);
-    }
-#ifdef DEBUG
-    fprintf(stderr, "to_erl: %s opened for reading\n", FIFO1);
-#endif
-
-    if ((wfd = open (FIFO2, O_WRONLY|DONT_BLOCK_PLEASE, 0)) < 0) {
-#ifdef DEBUG
-	fprintf(stderr, "Could not open FIFO %s for writing.\n", FIFO2);
-#endif
-	fprintf(stderr, "No running Erlang on pipe %s: %s\n", pipename, strerror(errno));
-	close(rfd);
-	exit(1);
-    }
-#ifdef DEBUG
-    fprintf(stderr, "to_erl: %s opened for writing\n", FIFO2);
-#endif
-
-#ifndef __OSE__
-    fprintf(stderr, "Attaching to %s (^D to exit)\n\n", pipename);
-#else
-    fprintf(stderr, "Attaching to %s (^C to exit)\n\n", pipename);
-#endif
-
-#ifndef __OSE__
-    /* Set break handler to our handler */
-    signal(SIGINT,handle_ctrlc);
-
-    /*
-     * Save the current state of the terminal, and set raw mode.
-     */
-    if (tcgetattr(0, &tty_rmode) , 0) {
-	fprintf(stderr, "Cannot get terminals current mode\n");
-	exit(-1);
-    }
-    tty_smode = tty_rmode;
-    tty_eof = '\004'; /* Ctrl+D to exit */
-#ifdef DEBUG
-    show_terminal_settings(&tty_rmode);
-#endif
-    tty_smode.c_iflag =
-	1*BRKINT |/*Signal interrupt on break.*/
-	    1*IGNPAR |/*Ignore characters with parity errors.*/
-		1*ISTRIP |/*Strip character.*/
-		    0;
-
-#if 0
-0*IGNBRK |/*Ignore break condition.*/
-0*PARMRK |/*Mark parity errors.*/
-0*INPCK  |/*Enable input parity check.*/
-0*INLCR  |/*Map NL to CR on input.*/
-0*IGNCR  |/*Ignore CR.*/
-0*ICRNL  |/*Map CR to NL on input.*/
-0*IUCLC  |/*Map upper-case to lower-case on input.*/
-0*IXON   |/*Enable start/stop output control.*/
-0*IXANY  |/*Enable any character to restart output.*/
-0*IXOFF  |/*Enable start/stop input control.*/
-0*IMAXBEL|/*Echo BEL on input line too long.*/
-#endif
-
-    tty_smode.c_oflag =
-	1*OPOST  |/*Post-process output.*/
-	    1*ONLCR  |/*Map NL to CR-NL on output.*/
-#ifdef XTABS
-		1*XTABS  |/*Expand tabs to spaces. (Linux)*/
-#endif
-#ifdef OXTABS
-		    1*OXTABS  |/*Expand tabs to spaces. (FreeBSD)*/
-#endif
-#ifdef NL0
-			1*NL0    |/*Select newline delays*/
-#endif
-#ifdef CR0
-			    1*CR0    |/*Select carriage-return delays*/
-#endif
-#ifdef TAB0
-				1*TAB0   |/*Select horizontal tab delays*/
-#endif
-#ifdef BS0
-				    1*BS0    |/*Select backspace delays*/
-#endif
-#ifdef VT0
-					1*VT0    |/*Select vertical tab delays*/
-#endif
-#ifdef FF0
-					    1*FF0    |/*Select form feed delays*/
-#endif
-											    0;
-
-#if 0
-0*OLCUC  |/*Map lower case to upper on output.*/
-0*OCRNL  |/*Map CR to NL on output.*/
-0*ONOCR  |/*No CR output at column 0.*/
-0*ONLRET |/*NL performs CR function.*/
-0*OFILL  |/*Use fill characters for delay.*/
-0*OFDEL  |/*Fill is DEL, else NULL.*/
-0*NL1    |
-0*CR1    |
-0*CR2    |
-0*CR3    |
-0*TAB1   |
-0*TAB2   |
-0*TAB3   |/*Expand tabs to spaces.*/
-0*BS1    |
-0*VT1    |
-0*FF1    |
-#endif
-
-    /* JALI: removed setting the tty_smode.c_cflag flags, since this is not */
-    /* advisable if this is a *real* terminal, such as the console. In fact */
-    /* this may hang the entire machine, deep, deep down (signalling break */
-    /* or toggling the abort switch doesn't help) */
-
-    tty_smode.c_lflag =
-									0;
-
-#if 0
-0*ISIG   |/*Enable signals.*/
-0*ICANON |/*Canonical input (erase and kill processing).*/
-0*XCASE  |/*Canonical upper/lower presentation.*/
-0*ECHO   |/*Enable echo.*/
-0*ECHOE  |/*Echo erase character as BS-SP-BS.*/
-0*ECHOK  |/*Echo NL after kill character.*/
-0*ECHONL |/*Echo NL.*/
-0*NOFLSH |/*Disable flush after interrupt or quit.*/
-0*TOSTOP |/*Send SIGTTOU for background output.*/
-0*ECHOCTL|/*Echo control characters as ^char, delete as ^?.*/
-0*ECHOPRT|/*Echo erase character as character erased.*/
-0*ECHOKE |/*BS-SP-BS erase entire line on line kill.*/
-0*FLUSHO |/*Output is being flushed.*/
-0*PENDIN |/*Retype pending input at next read or input character.*/
-0*IEXTEN |/*Enable extended (implementation-defined) functions.*/
-#endif
-
-    tty_smode.c_cc[VMIN]      =0;/* Note that VMIN is the same as VEOF! */
-    tty_smode.c_cc[VTIME]     =0;/* Note that VTIME is the same as VEOL! */
-    tty_smode.c_cc[VINTR]     =3;
-
-    tcsetattr(0, TCSADRAIN, &tty_smode);
-
-#ifdef DEBUG
-    show_terminal_settings(&tty_smode);
-#endif
-
-#endif /* !__OSE__ */
-    /*
-     * 	 "Write a ^L to the FIFO which causes the other end to redisplay
-     *    the input line."
-     * This does not seem to work as was intended in old comment above.
-     * However, this control character is now (R12B-3) used by run_erl
-     * to trigger the version handshaking between to_erl and run_erl
-     * at the start of every new to_erl-session.
-     */
-
-    if (write(wfd, "\014", 1) < 0) {
-	fprintf(stderr, "Error in writing ^L to FIFO.\n");
-    }
-
-#ifdef __OSE__
-    /* we have a tiny stack so we malloc the buffers */
-    stdin_buf = malloc(sizeof(char) * BUFSIZ);
-    pipe_buf = malloc(sizeof(char) * BUFSIZ);
-
-    efs_examine_fd(rfd,FLIB_FD_HANDLE,&pipe_fh);
-    efs_examine_fd(0,FLIB_FD_HANDLE,&stdin_fh);
-    READ_AIO(stdin_read_req,0,BUFSIZ,stdin_buf);
-    READ_AIO(pipe_read_req,rfd,BUFSIZ,pipe_buf);
-#endif
-
-    /*
-     * read and write
-     */
-    while (1) {
-#ifndef __OSE__
-	FD_ZERO(&readfds);
-	FD_SET(0, &readfds);
-	FD_SET(rfd, &readfds);
-	if (select(rfd + 1, &readfds, NULL, NULL, NULL) < 0) {
-	    if (recv_sig) {
-		FD_ZERO(&readfds);
-	    }
-	    else {
-		fprintf(stderr, "Error in select.\n");
-		break;
-	    }
-	}
-	len = 0;
-
-	/*
-	 * Read from terminal and write to FIFO
-         */
-	if (recv_sig) {
-	    switch (recv_sig) {
-	    case SIGINT:
-		fprintf(stderr, "[Break]\n\r");
-		buf[0] = '\003';
-		len = 1;
-		break;
-	    case SIGWINCH:
-		len = window_size_seq(buf,sizeof(buf));
-		break;
-	    default:
-		fprintf(stderr,"Unexpected signal: %u\n",recv_sig);
-	    }
-	    recv_sig = 0;
-	}
-	else
-#else /* __OSE__ */
-	SIGSELECT sigsel[] = {0};
-	sig = receive(sigsel);
-	len = 0;
-#endif
-#ifndef __OSE__
-	  if (FD_ISSET(0,&readfds)) {
-	    len = read(0, buf, sizeof(buf));
-#else /* __OSE__ */
-	  if (sig->signo == FM_READ_PTR_REPLY &&
-	      sig->fm_read_ptr.handle == stdin_fh) {
-	    len = sig->fm_read_ptr.status == EFS_SUCCESS ? sig->fm_read_ptr.actual : -1;
-	    buf = sig->fm_read_ptr.buffer;
-#endif
-	    if (len <= 0) {
-		close(rfd);
-		close(wfd);
-		if (len < 0) {
-		    fprintf(stderr, "Error in reading from stdin.\n");
-		} else {
-		    fprintf(stderr, "[EOF]\n\r");
-		}
-		break;
-	    }
-	    /* check if there is an eof character in input */
-	    for (i = 0; i < len-1 && buf[i] != tty_eof; i++);
-	    if (buf[i] == tty_eof) {
-		fprintf(stderr, "[Quit]\n\r");
-		break;
-	    }
-	}
-
-	if (len) {
-#ifdef DEBUG
-	    if(write(1, buf, len));
-#endif
-	    if (write_all(wfd, buf, len) != len) {
-		fprintf(stderr, "Error in writing to FIFO.\n");
-		close(rfd);
-		close(wfd);
-		break;
-	    }
-	    STATUS("\" OK\r\n");
-#ifdef __OSE__
-	    aio_dispatch(sig);
-	    READ_AIO(stdin_read_req, 0, BUFSIZ, stdin_buf);
-#endif
-	}
-
-	/*
-	 * Read from FIFO, write to terminal.
-	 */
-#ifndef __OSE__
-	if (FD_ISSET(rfd, &readfds)) {
-	    STATUS("FIFO read: ");
-	    len = read(rfd, buf, BUFSIZ);
-#else /* __OSE__ */
-        if (sig->signo == FM_READ_PTR_REPLY &&
-	    sig->fm_read_ptr.handle == pipe_fh) {
-	    len = sig->fm_read_ptr.status == EFS_SUCCESS ? sig->fm_read_ptr.actual : -1;
-	    buf = sig->fm_read_ptr.buffer;
-#endif
-	    if (len < 0 && errno == EAGAIN) {
-		/*
-		 * No data this time, but the writing end of the FIFO is still open.
-		 * Do nothing.
-		 */
-		;
-	    } else if (len <= 0) {
-		/*
-		 * Either an error or end of file. In either case, break out
-		 * of the loop.
-		 */
-		close(rfd);
-		close(wfd);
-		if (len < 0) {
-		    fprintf(stderr, "Error in reading from FIFO.\n");
-		} else
-		    fprintf(stderr, "[End]\n\r");
-		break;
-	    } else {
-		if (!got_some) {
-		    if ((len=version_handshake(buf,len,wfd)) < 0) {
-			close(rfd);
-			close(wfd);
-			break;
-		    }
-#ifndef __OSE__
-		    if (protocol_ver >= 1) {
-			/* Tell run_erl size of terminal window */
-			signal(SIGWINCH, handle_sigwinch);
-			raise(SIGWINCH);
-		    }
-#endif
-		    got_some = 1;
-		}
-
-		/*
-		 * We successfully read at least one character. Write what we got.
-		 */
-		STATUS("Terminal write: \"");
-		if (write_all(1, buf, len) != len) {
-		    fprintf(stderr, "Error in writing to terminal.\n");
-		    close(rfd);
-		    close(wfd);
-		    break;
-		}
-		STATUS("\" OK\r\n");
-#ifdef __OSE__
-		aio_dispatch(sig);
-		READ_AIO(pipe_read_req, rfd, BUFSIZ, pipe_buf);
-#endif
-	    }
-	}
-    }
-
-#ifndef __OSE__
-    /*
-     * Reset terminal characterstics
-     * XXX
-     */
-    tcsetattr(0, TCSADRAIN, &tty_rmode);
-#endif
-    return 0;
-}
-
-/* Call write() until entire buffer has been written or error.
- * Return len or -1.
- */
-static int write_all(int fd, const char* buf, int len)
-{
-    int left = len;
-    int written;
-    while (left) {
-	written = write(fd,buf,left);
-	if (written < 0) {
-	    return -1;
-	}
-	left -= written;
-	buf += written;
-    }
-    return len;
-}
-
-#ifndef __OSE__
-static int window_size_seq(char* buf, size_t bufsz)
-{
-#ifdef TIOCGWINSZ
-    struct winsize ws;
-    static const char prefix[] = "\033_";
-    static const char suffix[] = "\033\\";
-    /* This Esc sequence is called "Application Program Command"
-       and seems suitable to use for our own customized stuff. */
-
-    if (ioctl(STDIN_FILENO, TIOCGWINSZ, &ws) == 0) {
-	int len = sn_printf(buf, bufsz, "%swinsize=%u,%u%s",
-			    prefix, ws.ws_col, ws.ws_row, suffix);
-	return len;
-    }
-#endif /* TIOCGWINSZ */
-    return 0;
-}
-#endif /* !__OSE__ */
-
-/*   to_erl                     run_erl
- *     |                           |
- *     |---------- '\014' -------->| (session start)
- *     |                           |
- *     |<---- "[run_erl v1-0]" ----| (version interval)
- *     |                           |
- *     |--- Esc_"version=1"Esc\ -->| (common version)
- *     |                           |
- */
-static int version_handshake(char* buf, int len, int wfd)
-{
-    unsigned re_high=0, re_low;
-    char *end = find_str(buf,len,"]\n");
-
-    if (end && sscanf(buf,"[run_erl v%u-%u",&re_high,&re_low)==2) {
-	char wbuf[30];
-	int wlen;
-
-	if (re_low > RUN_ERL_HI_VER || re_high < RUN_ERL_LO_VER) {
-	    fprintf(stderr,"Incompatible versions: to_erl=v%u-%u run_erl=v%u-%u\n",
-		    RUN_ERL_HI_VER, RUN_ERL_LO_VER, re_high, re_low);
-	    return -1;
-	}
-	/* Choose highest common version */
-	protocol_ver = re_high < RUN_ERL_HI_VER ? re_high : RUN_ERL_HI_VER;
-
-	wlen = sn_printf(wbuf, sizeof(wbuf), "\033_version=%u\033\\",
-			 protocol_ver);
-	if (write_all(wfd, wbuf, wlen) < 0) {
-	    fprintf(stderr,"Failed to send version handshake\n");
-	    return -1;
-	}
-	end += 2;
-	len -= (end-buf);
-	memmove(buf,end,len);
-
-    }
-    else {  /* we assume old run_erl without version handshake */
-	protocol_ver = 0;
-    }
-
-    if (re_high != RUN_ERL_HI_VER) {
-	fprintf(stderr,"run_erl has different version, "
-		"using common protocol level %u\n", protocol_ver);
-    }
-
-    return len;
-}
-
-
-#if defined(DEBUG) && !defined(__OSE__)
-#define S(x)  ((x) > 0 ? 1 : 0)
-
-static void show_terminal_settings(struct termios *t)
-{
-  fprintf(stderr,"c_iflag:\n");
-  fprintf(stderr,"Signal interrupt on break:   BRKINT  %d\n", S(t->c_iflag & BRKINT));
-  fprintf(stderr,"Map CR to NL on input:       ICRNL   %d\n", S(t->c_iflag & ICRNL));
-  fprintf(stderr,"Ignore break condition:      IGNBRK  %d\n", S(t->c_iflag & IGNBRK));
-  fprintf(stderr,"Ignore CR:                   IGNCR   %d\n", S(t->c_iflag & IGNCR));
-  fprintf(stderr,"Ignore char with par. err's: IGNPAR  %d\n", S(t->c_iflag & IGNPAR));
-  fprintf(stderr,"Map NL to CR on input:       INLCR   %d\n", S(t->c_iflag & INLCR));
-  fprintf(stderr,"Enable input parity check:   INPCK   %d\n", S(t->c_iflag & INPCK));
-  fprintf(stderr,"Strip character              ISTRIP  %d\n", S(t->c_iflag & ISTRIP));
-  fprintf(stderr,"Enable start/stop input ctrl IXOFF   %d\n", S(t->c_iflag & IXOFF));
-  fprintf(stderr,"ditto output ctrl            IXON    %d\n", S(t->c_iflag & IXON));
-  fprintf(stderr,"Mark parity errors           PARMRK  %d\n", S(t->c_iflag & PARMRK));
-  fprintf(stderr,"\n");
-  fprintf(stderr,"c_oflag:\n");
-  fprintf(stderr,"Perform output processing    OPOST   %d\n", S(t->c_oflag & OPOST));
-  fprintf(stderr,"\n");
-  fprintf(stderr,"c_cflag:\n");
-  fprintf(stderr,"Ignore modem status lines    CLOCAL  %d\n", S(t->c_cflag & CLOCAL));
-  fprintf(stderr,"\n");
-  fprintf(stderr,"c_local:\n");
-  fprintf(stderr,"Enable echo                  ECHO    %d\n", S(t->c_lflag & ECHO));
-  fprintf(stderr,"\n");
-  fprintf(stderr,"c_cc:\n");
-  fprintf(stderr,"c_cc[VEOF]                           %d\n", t->c_cc[VEOF]);
-}
-#endif /* DEBUG && !__OSE__ */
diff --git a/erts/etc/common/to_erl_common.h b/erts/etc/common/to_erl_common.h
deleted file mode 100644
index a139da418f..0000000000
--- a/erts/etc/common/to_erl_common.h
+++ /dev/null
@@ -1,29 +0,0 @@
-/*
- * %CopyrightBegin%
- *
- * Copyright Ericsson AB 2013. All Rights Reserved.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- *
- * %CopyrightEnd%
- */
-#ifndef ERL_TO_ERL_H
-#define ERL_TO_ERL_H
-
-#define TO_ERL_USAGE "to_erl [-h|-F] %s\n"			\
-  "\t-h\tThis help text.\n"						\
-  "\t-f\tForce connection even though pipe is locked by other to_erl process."
-
-int to_erl(int argc, char **argv);
-
-#endif
diff --git a/erts/etc/ose/etc.lmconf b/erts/etc/ose/etc.lmconf
deleted file mode 100644
index b402b325b1..0000000000
--- a/erts/etc/ose/etc.lmconf
+++ /dev/null
@@ -1,20 +0,0 @@
-OSE_LM_STACK_SIZES=256,512,1024,2048,4096,8192,16384,65536
-OSE_LM_SIGNAL_SIZES=31,63,127,255,1023,4095,16383,65535
-OSE_LM_POOL_SIZE=0x200000
-OSE_LM_MAIN_NAME=main
-OSE_LM_MAIN_STACK_SIZE=0xF000
-OSE_LM_MAIN_PRIORITY=20
-## Has to be of a type that allows MAM
-OSE_LM_PROGRAM_TYPE=APP_RAM
-OSE_LM_DATA_INIT=YES
-OSE_LM_BSS_INIT=YES
-OSE_LM_EXEC_MODEL=SHARED
-HEAP_MAX_SIZE=1000000000
-HEAP_SMALL_BUF_INIT_SIZE=64000000
-HEAP_LARGE_BUF_THRESHOLD=16000000
-HEAP_LOCK_TYPE=2
-
-# Setting the environment variable EFS_RESOLVE_TMO on the block to 0.
-# This will eliminiate delays when trying to open files on not mounted
-# volumes.
-EFS_RESOLVE_TMO=0
diff --git a/erts/etc/ose/run_erl.c b/erts/etc/ose/run_erl.c
deleted file mode 100644
index 6775525297..0000000000
--- a/erts/etc/ose/run_erl.c
+++ /dev/null
@@ -1,664 +0,0 @@
-/*
- * %CopyrightBegin%
- *
- * Copyright Ericsson AB 2013. All Rights Reserved.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- *
- * %CopyrightEnd%
- */
-/*
- * Module: run_erl.c
- *
- */
-
-#ifdef HAVE_CONFIG_H
-#  include "config.h"
-#endif
-
-/* System includes */
-#include <aio.h>
-#include <errno.h>
-#include <dirent.h>
-#include <stdlib.h>
-#include <string.h>
-#include <stdio.h>
-#include <sys/stat.h>
-#include <unistd.h>
-
-/* OSE includes */
-#include "ose.h"
-#include "ose_spi/ose_spi.h"
-#include "efs.h"
-#include "pm.h"
-#include "ose_spi/fm.sig"
-
-/* erts includes */
-#include "run_erl.h"
-#include "run_erl_common.h"
-#include "safe_string.h"    /* sn_printf, strn_cpy, strn_cat, etc */
-
-typedef struct RunErlSetup_ {
-  SIGSELECT signo;
-  int run_daemon;
-  char *logdir;
-  char *command;
-  char *pipename;
-  char *blockname;
-} RunErlSetup;
-
-typedef struct ProgramState_ {
-  /* child process */
-  int ifd, ofd;
-  OSDOMAIN domain;
-  PROCESS progpid, mainbid;
-  struct PmProgramInfo *info;
-  /* to_erl */
-  char w_pipe[FILENAME_BUFSIZ],
-       r_pipe[FILENAME_BUFSIZ];
-} ProgramState;
-
-union SIGNAL {
-  SIGSELECT signo;
-  RunErlSetup setup;
-  struct FmReadPtr fm_read_ptr;
-  struct FmWritePtr fm_write_ptr;
-};
-
-static OSBOOLEAN hunt_in_block(char *block_name,
-			       char *process_name,
-			       PROCESS *pid);
-static int create_child_process(char *command_string, char *blockname,
-				ProgramState *state);
-
-
-static OSBOOLEAN hunt_in_block(char *block_name,
-			       char *process_name,
-			       PROCESS *pid) {
-  struct OS_pid_list *list;
-  PROCESS block_id = OSE_ILLEGAL_PROCESS;
-  int i;
-  char *name;
-
-  *pid = OSE_ILLEGAL_PROCESS;
-
-  list = get_bid_list(0);
-
-  if (!list)
-    return 0;
-
-  for (i = 0; i < list->count; i++) {
-
-    if (list->list[i] == get_bid(current_process()))
-      continue;
-
-    name = (char*)get_pid_info(list->list[i], OSE_PI_NAME);
-    if (name) {
-      if (strcmp(name,block_name) == 0) {
-	block_id = list->list[i];
-	free_buf((union SIGNAL**)&name);
-	break;
-      }
-      free_buf((union SIGNAL**)&name);
-    }
-  }
-
-  free_buf((union SIGNAL**)&list);
-
-  if (block_id == OSE_ILLEGAL_PROCESS)
-    return 0;
-
-  list = get_pid_list(block_id);
-
-  if (!list)
-    return 0;
-
-  for (i = 0; i < list->count; i++) {
-    name = (char*)get_pid_info(list->list[i], OSE_PI_NAME);
-    if (name) {
-      if (strcmp(name,process_name) == 0) {
-	*pid = list->list[i];
-	free_buf((union SIGNAL**)&name);
-	break;
-      }
-      free_buf((union SIGNAL**)&name);
-    }
-  }
-
-  free_buf((union SIGNAL**)&list);
-
-  if (*pid == OSE_ILLEGAL_PROCESS)
-    return 0;
-
-  return 1;
-
-}
-
-
-static int create_child_process(char *command_string,  char *blockname,
-				ProgramState *state) {
-  char *command = command_string;
-  char *argv;
-  int i = 0;
-  int ret_status;
-  PmStatus pm_status;
-  int tmp_io[2];
-  int fd_arr[3];
-  int ifd[2], ofd[2];
-  char *handle;
-  struct PmLoadModuleInfoReply *mod_info;
-
-  /* Parse out cmd and argv from the command string */
-  while (1) {
-    if (command[i] == ' ' || command[i] == '\0') {
-      if (command[i] == '\0')
-	argv = NULL;
-      else {
-	command[i] = '\0';
-	argv = command_string + i + 1;
-      }
-      break;
-    }
-    i++;
-  }
-
-  if (blockname)
-    handle = blockname;
-  else
-    handle = simple_basename(command);
-
-  if (ose_pm_load_module_info(handle,&mod_info) == PM_SUCCESS) {
-    /* Already installed */
-    free_buf((union SIGNAL**)&mod_info);
-  } else if ((pm_status = ose_pm_install_load_module(0,"ELF",command,handle,0,0,NULL))
-	     != PM_SUCCESS) {
-      ERROR1(LOG_ERR,"ose_pm_install_load_module failed - pmstatus: 0x%08x\n",
-	     pm_status);
-      return 0;
-  }
-
-  state->domain = PM_NEW_DOMAIN;
-
-  pm_status = ose_pm_create_program(&state->domain, handle, 0, 0 , NULL,
-				    &state->progpid, &state->mainbid);
-
-  if (pm_status != PM_SUCCESS) {
-    if (pm_status == PM_EINSTALL_HANDLE_IN_USE)
-      ERROR1(LOG_ERR,"ose_pm_create_program failed - "
-	     "install handle \"%s\" is in use. You can specify another "
-	     "install handle by using the -block option to run_erl.\n",handle);
-    else
-      ERROR1(LOG_ERR,"ose_pm_create_program failed - pmstatus: 0x%08x\n",
-	     pm_status);
-    return 0;
-  }
-
-  pm_status = ose_pm_program_info(state->progpid, &state->info);
-  /* FIXME don't forget to free this ((union SIGNAL **)&info) */
-  if (pm_status != PM_SUCCESS) {
-    ERROR1(LOG_ERR,"ose_pm_program_info failed - pmstatus: 0x%08x\n",
-	   pm_status);
-    return 0;
-  }
-
-  /* We only clone stdin+stdout, what about stderr? */
-
-  /* create pipes */
-  if (pipe(ifd) < 0) {
-    if (errno == ENOENT)
-      ERRNO_ERR0(LOG_ERR,"The /pipe file system is not available\n");
-    else
-      ERRNO_ERR0(LOG_ERR,"pipe ifd failed\n");
-    return 0;
-  }
-
-  if (pipe(ofd) < 0) {
-    ERRNO_ERR0(LOG_ERR,"pipe ofd failed\n");
-    return 0;
-  }
-
-  /* FIXME Lock? */
-
-  /* backup our stdin stdout */
-  if ((tmp_io[0] = dup(0)) < 0) {
-    ERRNO_ERR0(LOG_ERR,"dup 0 failed\n");
-    return 0;
-  }
-
-  if ((tmp_io[1] = dup(1)) < 0) {
-    ERRNO_ERR0(LOG_ERR,"dup 1 failed\n");
-    return 0;
-  }
-
-  /* set new pipe to fd 0,1 */
-  if (dup2(ifd[1], 1) < 0) {
-    ERRNO_ERR0(LOG_ERR,"dup2 1 failed\n");
-    return 0;
-  }
-
-  if (dup2(ofd[0], 0) < 0) {
-    ERRNO_ERR0(LOG_ERR,"dup2 0 failed\n");
-    return 0;
-  }
-
-  /* clone array to newly created */
-  fd_arr[0] = 2; /* Number of fd's */
-  fd_arr[1] = 0;
-  fd_arr[2] = 1;
-
-  if ((ret_status = efs_clone_array(state->info->main_process, fd_arr))
-      != EFS_SUCCESS) {
-    ERROR1(LOG_ERR,"efs_close_array filed, errcode: %d\n", ret_status);
-    return 0;
-  }
-
-  if (dup2(tmp_io[1], 1) < 0) {
-    ERRNO_ERR0(LOG_ERR,"restoring dup2 1 failed\n");
-    return 0;
-  }
-
-  if (dup2(tmp_io[0], 0) < 0) {
-    ERRNO_ERR0(LOG_ERR,"restoring dup2 1 failed\n");
-    return 0;
-  }
-
-  /* close loose-ends */
-  sf_close(tmp_io[0]);
-  sf_close(tmp_io[1]);
-  sf_close(ifd[1]);
-  sf_close(ofd[0]);
-  state->ifd = ifd[0];
-  state->ofd = ofd[1];
-
-  if (argv && set_env(state->progpid, "ARGV", argv)) {
-    ERRNO_ERR0(LOG_ERR,"something went wrong with set_env\n");
-  }
-
-  /*
-   * Start the program.
-   */
-  pm_status = ose_pm_start_program(state->progpid);
-  if (pm_status != PM_SUCCESS) {
-    ERROR1(LOG_ERR,"ose_pm_install_load_module failed - pmstatus: 0x%08x\n",
-	   pm_status);
-    return 0;
-  }
-
-  return 1;
-}
-
-#define SET_AIO(REQ,FD,SIZE,BUFF)					\
-  /* Make sure to clean data structure of previous request */		\
-  memset(&(REQ),0,sizeof(REQ));						\
-  (REQ).aio_fildes = FD;						\
-  (REQ).aio_offset = FM_POSITION_CURRENT;				\
-  (REQ).aio_nbytes = SIZE;						\
-  (REQ).aio_buf = BUFF;							\
-  (REQ).aio_sigevent.sigev_notify = SIGEV_NONE
-
-#define READ_AIO(REQ,FD,SIZE,BUFF) do {					\
-  SET_AIO(REQ,FD,SIZE,BUFF);						\
-  if (aio_read(&(REQ)) != 0)						\
-    ERRNO_ERR1(LOG_ERR,"aio_read of child_read_req(%d) failed\n",FD);	\
-  } while (0)
-
-#define WRITE_AIO(FD,SIZE,BUFF) do {					\
-    struct aiocb *write_req = malloc(sizeof(struct aiocb));		\
-    char *write_buff = malloc(sizeof(char)*SIZE);			\
-    memcpy(write_buff,BUFF,SIZE);					\
-    SET_AIO(*write_req,FD,SIZE,write_buff);				\
-    if (aio_write(write_req) != 0)					\
-      ERRNO_ERR1(LOG_ERR,"aio_write of write_req(%d) failed\n",FD);	\
-  } while(0)
-
-int pass_on(ProgramState *state);
-int pass_on(ProgramState *s) {
-  SIGSELECT sigsel[] = {0,FM_READ_PTR_REPLY};
-  union SIGNAL *sig;
-  char child_read_buff[BUFSIZ], pipe_read_buff[BUFSIZ];
-  struct aiocb child_read_req, pipe_read_req;
-  int rfd, wfd = 0;
-  FmHandle rfh, child_rfh;
-  int outstanding_writes = 0, got_some = 0, child_done = 0;
-
-  if ((rfd = sf_open(s->r_pipe, O_RDONLY, 0)) < 0) {
-    ERRNO_ERR1(LOG_ERR,"Could not open FIFO '%s' for reading.\n", s->r_pipe);
-    rfd = 0;
-    return 1;
-  }
-
-  attach(NULL,s->progpid);
-
-  /* Open the log file */
-  erts_run_erl_log_open();
-
-  efs_examine_fd(rfd,FLIB_FD_HANDLE,&rfh);
-  efs_examine_fd(s->ifd,FLIB_FD_HANDLE,&child_rfh);
-
-  READ_AIO(child_read_req,s->ifd,BUFSIZ,child_read_buff);
-  READ_AIO(pipe_read_req,rfd,BUFSIZ,pipe_read_buff);
-
-  while (1) {
-    time_t now,last_activity;
-
-    time(&last_activity);
-    sig = receive_w_tmo(erts_run_erl_log_alive_minutes()*60000,sigsel);
-
-    time(&now);
-
-    if (sig) {
-      erts_run_erl_log_activity(0,now,last_activity);
-    } else {
-      /* timeout */
-      erts_run_erl_log_activity(1,now,last_activity);
-      continue;
-    }
-
-    switch (sig->signo) {
-    case OS_ATTACH_SIG: {
-      if (rfd) { sf_close(rfd); rfd = 0; }
-      free_buf(&sig);
-      child_done = 1;
-      /* Make sure to to let all outstanding write request finish */
-      if (outstanding_writes)
-	break;
-      if (wfd) sf_close(wfd);
-      return 0;
-    }
-    case FM_WRITE_PTR_REPLY: {
-      if (sig->fm_write_ptr.status == EFS_SUCCESS) {
-	if (sig->fm_write_ptr.actual < sig->fm_write_ptr.requested) {
-	  WRITE_AIO(wfd, sig->fm_write_ptr.requested-sig->fm_write_ptr.actual,
-		    sig->fm_write_ptr.buffer+sig->fm_write_ptr.actual);
-	}
-      } else {
-	/* Assume to_erl has terminated. */
-	sf_close(wfd);
-	wfd = 0;
-      }
-      free((char*)sig->fm_write_ptr.buffer);
-      aio_dispatch(sig);
-      if ((--outstanding_writes == 0) && child_done) {
-	if (wfd) sf_close(wfd);
-	return 0;
-      }
-      break;
-    }
-    case FM_READ_PTR_REPLY: {
-      /* Child fd */
-      if (sig->fm_read_ptr.handle == child_rfh) {
-
-	/* Child terminated */
-	if (sig->fm_read_ptr.status != EFS_SUCCESS ||
-	    sig->fm_read_ptr.actual == 0) {
-
-	  if (rfd) { sf_close(rfd); rfd = 0; }
-
-	  if (sig->fm_read_ptr.status != EFS_SUCCESS) {
-	    ERROR0(LOG_ERR,"Erlang closed the connection.");
-	    aio_dispatch(sig);
-	    return 1;
-	  }
-
-	  /* child closed connection gracefully */
-	  aio_dispatch(sig);
-	  if (outstanding_writes) {
-	    child_done = 1;
-	    break;
-	  }
-
-	  if (wfd) sf_close(wfd);
-
-	  return 0;
-	} else {
-	  erts_run_erl_log_write(sig->fm_read_ptr.buffer,
-				 sig->fm_read_ptr.actual);
-	  if (wfd) {
-	    WRITE_AIO(wfd, sig->fm_read_ptr.actual, sig->fm_read_ptr.buffer);
-	    outstanding_writes++;
-	  }
-	  aio_dispatch(sig);
-	  READ_AIO(child_read_req, s->ifd,BUFSIZ, child_read_buff);
-	}
-      /* pipe fd */
-      } else if (sig->fm_read_ptr.handle == rfh) {
-	if (sig->fm_read_ptr.status != EFS_SUCCESS) {
-	  if(rfd) sf_close(rfd);
-	  if(wfd) sf_close(wfd);
-	  aio_dispatch(sig);
-	  ERRNO_ERR0(LOG_ERR,"Error in reading from FIFO.");
-	  return 1;
-	}
-	if (sig->fm_read_ptr.actual == 0) {
-	  /* to_erl closed its end of the pipe */
-	  aio_dispatch(sig);
-	  sf_close(rfd);
-	  rfd = sf_open(s->r_pipe,O_RDONLY|DONT_BLOCK_PLEASE, 0);
-	  if (rfd < 0) {
-	    ERRNO_ERR1(LOG_ERR,"Could not open FIFO '%s' for reading.",
-		       s->r_pipe);
-	    rfd = 0;
-	  } else {
-	    READ_AIO(pipe_read_req,rfd,BUFSIZ,pipe_read_buff);
-	  }
-	  got_some = 0; /* reset for next session */
-	} else {
-	  int len = sig->fm_read_ptr.actual;
-	  char *buffer = sig->fm_read_ptr.buffer;
-	  if (!wfd) {
-	    /* Try to open the write pipe to to_erl. Now that we got some data
-	     * from to_erl, to_erl should already be reading this pipe - open
-	     * should succeed. But in case of error, we just ignore it.
-	     */
-	    if ((wfd = sf_open(s->w_pipe, O_WRONLY|DONT_BLOCK_PLEASE, 0)) < 0) {
-	      erts_run_erl_log_status("Client expected on FIFO %s, "
-				      "but can't open (len=%d)\n",
-				      s->w_pipe, sig->fm_read_ptr.actual);
-	      sf_close(rfd);
-	      rfd = sf_open(s->r_pipe, O_RDONLY|DONT_BLOCK_PLEASE, 0);
-	      if (rfd < 0) {
-		ERRNO_ERR1(LOG_ERR,"Could not open FIFO '%s' for reading.",
-			   s->r_pipe);
-		return 1;
-	      }
-	      wfd = 0;
-	    } else {
-#ifdef DEBUG
-	      erts_run_erl_log_status("run_erl: %s opened for writing\n",
-				      s->w_pipe);
-#endif
-	    }
-	  }
-
-	  if (!got_some && wfd && buffer[0] == '\014') {
-	    char wbuf[30];
-	    int wlen = sn_printf(wbuf,sizeof(wbuf),"[run_erl v%u-%u]\n",
-				 RUN_ERL_HI_VER, RUN_ERL_LO_VER);
-	    /* For some reason this, the first write aio seems to
-	       not get an FM_WRITE_PTR_REPLY, so we do not do:
-	       outstanding_writes++;
-	    */
-	    WRITE_AIO(wfd, wlen, wbuf);
-	  }
-	  got_some = 1;
-
-	  /* Write the message */
-#ifdef DEBUG
-	  erts_run_erl_log_status("Pty master write; ");
-#endif
-	  len = erts_run_erl_extract_ctrl_seq(buffer,len, s->ofd);
-
-	  if (len > 0) {
-	    int wlen = erts_run_erl_write_all(s->ofd, buffer, len);
-	    if (wlen != len) {
-	      aio_dispatch(sig);
-	      ERRNO_ERR0(LOG_ERR,"Error in writing to terminal.");
-	      if(rfd) sf_close(rfd);
-	      if(wfd) sf_close(wfd);
-	      return 1;
-	    }
-	  }
-#ifdef DEBUG
-	  erts_run_erl_log_status("OK\n");
-#endif
-	  aio_dispatch(sig);
-	  READ_AIO(pipe_read_req,rfd,BUFSIZ,pipe_read_buff);
-	}
-	}
-      break;
-    }
-    default: {
-      free_buf(&sig);
-      break;
-    }
-    }
-  }
-}
-
-OS_PROCESS(run_erl_process) {
-  char *logdir, *command, *blockname;
-  SIGSELECT sigsel[] = {1,ERTS_SIGNAL_RUN_ERL_SETUP};
-  union SIGNAL *sig = receive(sigsel);
-  ProgramState state;
-  char pipename[FILENAME_BUFSIZ];
-
-  state.info = NULL;
-
-  logdir = strdup(sig->setup.logdir);
-  command = strdup(sig->setup.command);
-  strn_cpy(pipename,sizeof(pipename),sig->setup.pipename);
-
-  if (sig->setup.blockname)
-    blockname = strdup(sig->setup.blockname);
-  else
-    blockname = NULL;
-
-  erts_run_erl_log_init(sig->setup.run_daemon, logdir);
-
-  free_buf(&sig);
-
-  if (erts_run_erl_open_fifo(pipename,state.w_pipe,state.r_pipe))
-    kill_proc(current_process());
-
-  if (create_child_process(command,blockname,&state))
-    pass_on(&state);
-
-  free(logdir);
-  free(command);
-  if (blockname)
-    free(blockname);
-
-  if (state.info)
-    free_buf(((union SIGNAL**)&state.info));
-
-  sf_close(state.ifd);
-  sf_close(state.ofd);
-
-  unlink(state.w_pipe);
-  unlink(state.r_pipe);
-
-  kill_proc(current_process());
-}
-
-int run_erl(int argc,char **argv) {
-  char *pipename, *logdir, *command, *blockname = NULL;
-  int pipename_len, logdir_len, command_len, blockname_len = 0;
-  int i = 1, run_daemon = 0;
-  PROCESS pid;
-  SIGSELECT sigsel[] = {0};
-  union SIGNAL *sig;
-
-  if(argc < 4) {
-    fprintf(stderr,RUN_ERL_USAGE,"run_erl");
-    return 1;
-  }
-
-  while (1) {
-    if (argv[i][0] != '-')
-      break;
-    if (!strcmp(argv[i],"-daemon")) {
-      run_daemon = 1;
-      i++;
-      continue;
-    }
-    if (!strcmp(argv[i],"-block")) {
-      blockname = argv[i+1];
-      blockname_len = strlen(argv[i+1]) + 1;
-      i+=2;
-      continue;
-    }
-    fprintf(stderr,RUN_ERL_USAGE,"run_erl");
-    return 1;
-  }
-
-  pipename = argv[i++];
-  logdir = argv[i++];
-  command = argv[i++];
-
-  /* + 1 to include NULL at end */
-  logdir_len = strlen(logdir) + 1;
-  command_len = strlen(command) + 1;
-  pipename_len = strlen(pipename) + 1;
-
-  if (run_daemon) {
-    /* We request that the run_erl_process should be started from the
-       main process so that it does not die when the shell command
-       returns */
-    PROCESS main_pid;
-    hunt_in_block("run_erl","main",&main_pid);
-    sig = alloc(sizeof(*sig),ERTS_SIGNAL_RUN_ERL_DAEMON);
-    send(&sig,main_pid);
-    sig = receive(sigsel);
-    pid = sender(&sig);
-    free_buf(&sig);
-  } else {
-    pid = create_process(OS_BG_PROC,"run_erl_process",
-			 run_erl_process, 0x800,
-			 0, 0, 0, NULL, 0, 0);
-  }
-
-  sig = alloc(sizeof(RunErlSetup)+
-	      logdir_len+command_len+pipename_len+blockname_len,
-	      ERTS_SIGNAL_RUN_ERL_SETUP);
-  sig->setup.run_daemon = run_daemon;
-  sig->setup.logdir = ((char*)sig)+sizeof(RunErlSetup);
-  sig->setup.command = ((char*)sig)+sizeof(RunErlSetup)+logdir_len;
-  sig->setup.pipename = ((char*)sig)+sizeof(RunErlSetup)+logdir_len+command_len;
-  if (blockname)
-    sig->setup.blockname = ((char*)sig)+sizeof(RunErlSetup)+
-      logdir_len+command_len+pipename_len;
-  else
-    sig->setup.blockname = NULL;
-
-  strcpy(sig->setup.logdir,logdir);
-  strcpy(sig->setup.command,command);
-  strcpy(sig->setup.pipename,pipename);
-  if (blockname) strcpy(sig->setup.blockname,blockname);
-
-  send(&sig,pid);
-
-  if (run_daemon) {
-    /* We are a daemon, error msgs will be sent to ramlog */
-    start(pid);
-    return 1;
-  }
-
-  /* We are not daemon, error msgs will be sent to stderr and we block here */
-  efs_clone(pid);
-  start(pid);
-
-  attach(NULL,pid);
-  sig = receive(sigsel);
-
-  return 1;
-}
diff --git a/erts/etc/ose/run_erl.h b/erts/etc/ose/run_erl.h
deleted file mode 100644
index bdc8b6c355..0000000000
--- a/erts/etc/ose/run_erl.h
+++ /dev/null
@@ -1,30 +0,0 @@
-/*
- * %CopyrightBegin%
- *
- * Copyright Ericsson AB 2013. All Rights Reserved.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- *
- * %CopyrightEnd%
- */
-#ifndef ERL_RUN_ERL_H
-#define ERL_RUN_ERL_H
-
-#include "ose.h"
-
-#include "erts.sig"
-
-int run_erl(int argc, char **argv);
-OS_PROCESS(run_erl_process);
-
-#endif
diff --git a/erts/etc/ose/run_erl_main.c b/erts/etc/ose/run_erl_main.c
deleted file mode 100644
index 8895c773a1..0000000000
--- a/erts/etc/ose/run_erl_main.c
+++ /dev/null
@@ -1,80 +0,0 @@
-/*
- * %CopyrightBegin%
- *
- * Copyright Ericsson AB 2013. All Rights Reserved.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- *
- * %CopyrightEnd%
- */
-/*
- * Module: run_erl_main.c
- *
- * Container for load module that installs both run_erl and to_erl command.
- */
-
-#ifdef HAVE_CONFIG_H
-#  include "config.h"
-#endif
-
-#include <stdio.h>
-
-#include "ose.h"
-#include "shell.h"
-
-#include "run_erl_common.h"
-#include "run_erl.h"
-#include "to_erl_common.h"
-
-union SIGNAL {
-  SIGSELECT signo;
-};
-
-int main(int argc, char **argv)
-{
-
-  char run_erl_usage[320],
-    to_erl_usage[120];
-
-  (void)stdin;(void)stdout;(void)stderr;
-
-  sprintf(run_erl_usage,RUN_ERL_USAGE,"run_erl [-daemon] [-block blockname]");
-  sprintf(to_erl_usage,TO_ERL_USAGE,"pipename");
-
-  shell_add_cmd_attrs(
-    "run_erl",run_erl_usage,
-    "Redirect Erlang input and output streams",
-    run_erl,DEFAULT_PROC_TYPE,DEFAULT_PRIORITY,DEFAULT_STACK_SIZE);
-
-  shell_add_cmd_attrs(
-    "to_erl",to_erl_usage,
-    "Attach to redirected Erlang input and output streams",
-    to_erl,DEFAULT_PROC_TYPE,DEFAULT_PRIORITY,DEFAULT_STACK_SIZE);
-
-  while (1) {
-    static const SIGSELECT sigsel[] = {0};
-    union SIGNAL *sig = receive(sigsel);
-
-    if (sig->signo == ERTS_SIGNAL_RUN_ERL_DAEMON) {
-      PROCESS pid = create_process(OS_BG_PROC,"run_erl_daemon",
-				   run_erl_process, 0x800,
-				   0, 0, 0, NULL, 0, 0);
-      send_w_s(&sig,pid,sender(&sig));
-    } else {
-      printf("Got unexpected signal!");
-      free_buf(&sig);
-    }
-  }
-
-  return 1;
-}
diff --git a/erts/etc/unix/etp-commands.in b/erts/etc/unix/etp-commands.in
index 19d67de92f..7b554e71f2 100644
--- a/erts/etc/unix/etp-commands.in
+++ b/erts/etc/unix/etp-commands.in
@@ -774,7 +774,7 @@ define etp-pid-1
 #
   set $etp_pid_1 = (Eterm)($arg0)
   if ($etp_pid_1 & 0xF) == 0x3
-    if (etp_arch_bits == 64 && etp_halfword == 0)
+    if (etp_arch_bits == 64)
       if (etp_big_endian)
       	set $etp_pid_data = (unsigned) ((((Uint64) $etp_pid_1) >> 36) & 0x0fffffff)
       else
@@ -834,7 +834,7 @@ define etp-port-1
 #
   set $etp_port_1 = (Eterm)($arg0)
   if ($etp_port_1 & 0xF) == 0x7
-    if (etp_arch_bits == 64 && etp_halfword == 0)
+    if (etp_arch_bits == 64)
       if (etp_big_endian)
       	set $etp_port_data = (unsigned) ((((Uint64) $etp_port_1) >> 36) & 0x0fffffff)
       else
@@ -1577,7 +1577,7 @@ define etp-term-dump-pid
 #
   set $etp_pid_1 = (Eterm)($arg0)
   if ($etp_pid_1 & 0xF) == 0x3
-    if (etp_arch_bits == 64 && etp_halfword == 0)
+    if (etp_arch_bits == 64)
       if (etp_big_endian)
         set $etp_pid_data = (unsigned) ((((Uint64) $etp_pid_1) >> 36) & 0x0fffffff)
       else
@@ -1622,7 +1622,7 @@ end
 define etp-pid2pix-1
 # Args: Eterm
 #
-   if (etp_arch_bits == 64 && etp_halfword == 0)
+   if (etp_arch_bits == 64)
       if (etp_big_endian)
       	 set $etp_pix = (int) (((Uint64) $arg0) & 0x0fffffff)
       else
@@ -1965,7 +1965,7 @@ end
 define etp-port-id2pix-1
 # Args: Eterm
 #
-   if (etp_arch_bits == 64 && etp_halfword == 0)
+   if (etp_arch_bits == 64)
       if (etp_big_endian)
       	 set $etp_pix = (int) (((Uint64) $arg0) & 0x0fffffff)
       elser
@@ -2493,12 +2493,6 @@ define etp-system-info
     printf "Little\n"
   end
   printf "Word size: %d-bit\n", etp_arch_bits
-  printf "Halfword: "
-  if (etp_halfword)
-    printf "yes\n"
-  else
-    printf "no\n"
-  end
   printf "HiPE support: "
   if (etp_hipe)
     printf "yes\n"
diff --git a/erts/etc/unix/run_erl.c b/erts/etc/unix/run_erl.c
index c8414030ca..44efb975ba 100644
--- a/erts/etc/unix/run_erl.c
+++ b/erts/etc/unix/run_erl.c
@@ -1,7 +1,7 @@
 /*
  * %CopyrightBegin%
  * 
- * Copyright Ericsson AB 1996-2013. All Rights Reserved.
+ * Copyright Ericsson AB 1996-2015. All Rights Reserved.
  * 
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
@@ -41,13 +41,11 @@
 #ifdef HAVE_CONFIG_H
 #  include "config.h"
 #endif
-
 #ifdef HAVE_WORKING_POSIX_OPENPT
 #ifndef _XOPEN_SOURCE
 #define _XOPEN_SOURCE 600 
 #endif
 #endif
-
 #include <sys/types.h>
 #include <sys/wait.h>
 #include <sys/stat.h>
@@ -65,6 +63,11 @@
 #include <dirent.h>
 #include <termios.h>
 #include <time.h>
+
+#ifdef __ANDROID__
+#  include <termios.h>
+#endif
+
 #ifdef HAVE_SYSLOG_H
 #  include <syslog.h>
 #endif
@@ -84,25 +87,81 @@
 #  include <stropts.h>
 #endif
 
-#include "run_erl_common.h"
+#include "run_erl.h"
 #include "safe_string.h"    /* sn_printf, strn_cpy, strn_cat, etc */
 
+#ifdef O_NONBLOCK
+#  define DONT_BLOCK_PLEASE O_NONBLOCK
+#else
+#  define DONT_BLOCK_PLEASE O_NDELAY
+#  ifndef EAGAIN
+#    define EAGAIN -3898734
+#  endif
+#endif
+
+#define noDEBUG
+
+#define DEFAULT_LOG_GENERATIONS 5
+#define LOG_MAX_GENERATIONS     1000      /* No more than 1000 log files */
+#define LOG_MIN_GENERATIONS     2         /* At least two to switch between */
+#define DEFAULT_LOG_MAXSIZE     100000
+#define LOG_MIN_MAXSIZE         1000      /* Smallast value for changing log file */
+#define LOG_STUBNAME            "erlang.log."
+#define LOG_PERM                0664
+#define DEFAULT_LOG_ACTIVITY_MINUTES    5
+#define DEFAULT_LOG_ALIVE_MINUTES       15
+#define DEFAULT_LOG_ALIVE_FORMAT        "%a %b %e %T %Z %Y"
+#define ALIVE_BUFFSIZ                   256
+
+#define PERM            0600
+#define STATUSFILENAME  "/run_erl.log"
+#define PIPE_STUBNAME   "erlang.pipe"
+#define PIPE_STUBLEN    strlen(PIPE_STUBNAME)
+
+#ifndef FILENAME_MAX
+#define FILENAME_MAX 250
+#endif
+
+#ifndef O_SYNC
+#define O_SYNC 0
+#define USE_FSYNC 1
+#endif
+
 #define MAX(x,y)  ((x) > (y) ? (x) : (y))
 
+#define FILENAME_BUFSIZ FILENAME_MAX
+
 /* prototypes */
 static void usage(char *);
+static int create_fifo(char *name, int perm);
 static int open_pty_master(char **name, int *sfd);
 static int open_pty_slave(char *name);
 static void pass_on(pid_t);
 static void exec_shell(char **);
+static void status(const char *format,...);
+static void error_logf(int priority, int line, const char *format,...);
 static void catch_sigchild(int);
+static int next_log(int log_num);
+static int prev_log(int log_num);
+static int find_next_log_num(void);
+static int open_log(int log_num, int flags);
+static void write_to_log(int* lfd, int* log_num, char* buf, int len);
 static void daemon_init(void);
+static char *simple_basename(char *path);
 static void init_outbuf(void);
 static int outbuf_size(void);
 static void clear_outbuf(void);
 static char* outbuf_first(void);
 static void outbuf_delete(int bytes);
 static void outbuf_append(const char* bytes, int n);
+static int write_all(int fd, const char* buf, int len);
+static int extract_ctrl_seq(char* buf, int len);
+static void set_window_size(unsigned col, unsigned row);
+
+static ssize_t sf_write(int fd, const void *buffer, size_t len);
+static ssize_t sf_read(int fd, void *buffer, size_t len);
+static int sf_open(const char *path, int flags, mode_t mode);
+static int sf_close(int fd);
 
 #ifdef DEBUG
 static void show_terminal_settings(struct termios *t);
@@ -110,11 +169,20 @@ static void show_terminal_settings(struct termios *t);
 
 /* static data */
 static char fifo1[FILENAME_BUFSIZ], fifo2[FILENAME_BUFSIZ];
+static char statusfile[FILENAME_BUFSIZ];
+static char log_dir[FILENAME_BUFSIZ];
 static char pipename[FILENAME_BUFSIZ];
 static FILE *stdstatus = NULL;
+static int log_generations = DEFAULT_LOG_GENERATIONS;
+static int log_maxsize     = DEFAULT_LOG_MAXSIZE;
+static int log_alive_minutes = DEFAULT_LOG_ALIVE_MINUTES;
+static int log_activity_minutes = DEFAULT_LOG_ACTIVITY_MINUTES;
+static int log_alive_in_gmt = 0;
+static char log_alive_format[ALIVE_BUFFSIZ+1];
 static int run_daemon = 0;
 static char *program_name;
 static int mfd; /* master pty fd */
+static unsigned protocol_ver = RUN_ERL_LO_VER; /* assume lowest to begin with */
 
 /*
  * Output buffer.
@@ -145,13 +213,29 @@ static char* outbuf_in;
                                   LOG_PID|LOG_CONS|LOG_NOWAIT,LOG_USER)
 #endif
 
+#define ERROR0(Prio,Format) error_logf(Prio,__LINE__,Format"\n")
+#define ERROR1(Prio,Format,A1) error_logf(Prio,__LINE__,Format"\n",A1)
+#define ERROR2(Prio,Format,A1,A2) error_logf(Prio,__LINE__,Format"\n",A1,A2)
+
+#ifdef HAVE_STRERROR
+#    define ADD_ERRNO(Format) "errno=%d '%s'\n"Format"\n",errno,strerror(errno)
+#else
+#    define ADD_ERRNO(Format) "errno=%d\n"Format"\n",errno
+#endif
+#define ERRNO_ERR0(Prio,Format) error_logf(Prio,__LINE__,ADD_ERRNO(Format))
+#define ERRNO_ERR1(Prio,Format,A1) error_logf(Prio,__LINE__,ADD_ERRNO(Format),A1)
+
+
 int main(int argc, char **argv)
 {
   int childpid;
   int sfd = -1;
-  char *ptyslave=NULL;
+  int fd;
+  char *p, *ptyslave=NULL;
   int i = 1;
   int off_argv;
+  int calculated_pipename = 0;
+  int highest_pipe_num = 0;
 
   program_name = argv[0];
 
@@ -169,16 +253,122 @@ int main(int argc, char **argv)
 
   off_argv = i;
   strn_cpy(pipename, sizeof(pipename), argv[i++]);
-
-  erts_run_erl_log_init(run_daemon,argv[i]);
+  strn_cpy(log_dir, sizeof(log_dir), argv[i]);
+  strn_cpy(statusfile, sizeof(statusfile), log_dir);
+  strn_cat(statusfile, sizeof(statusfile), STATUSFILENAME);
 
 #ifdef DEBUG
-  erts_run_erl_log_status("%s: pid is : %d\n", argv[0], getpid());
+  status("%s: pid is : %d\n", argv[0], getpid());
 #endif
 
-  /* Open read and write fifo */
-  if (erts_run_erl_open_fifo(pipename,fifo1,fifo2))
-    exit(1);
+  /* Get values for LOG file handling from the environment */
+  if ((p = getenv("RUN_ERL_LOG_ALIVE_MINUTES"))) {
+      log_alive_minutes = atoi(p);
+      if (!log_alive_minutes) {
+	  ERROR1(LOG_ERR,"Minimum value for RUN_ERL_LOG_ALIVE_MINUTES is 1 "
+		 "(current value is %s)",p);
+      }
+      log_activity_minutes = log_alive_minutes / 3;
+      if (!log_activity_minutes) {
+	  ++log_activity_minutes;
+      }
+  }
+  if ((p = getenv("RUN_ERL_LOG_ACTIVITY_MINUTES"))) {
+     log_activity_minutes = atoi(p);
+      if (!log_activity_minutes) {
+	  ERROR1(LOG_ERR,"Minimum value for RUN_ERL_LOG_ACTIVITY_MINUTES is 1 "
+		 "(current value is %s)",p);
+      }
+  } 
+  if ((p = getenv("RUN_ERL_LOG_ALIVE_FORMAT"))) {
+      if (strlen(p) > ALIVE_BUFFSIZ) {
+	  ERROR1(LOG_ERR, "RUN_ERL_LOG_ALIVE_FORMAT can contain a maximum of "
+		 "%d characters", ALIVE_BUFFSIZ);
+      }
+      strn_cpy(log_alive_format, sizeof(log_alive_format), p);
+  } else {
+      strn_cpy(log_alive_format, sizeof(log_alive_format), DEFAULT_LOG_ALIVE_FORMAT);
+  }
+  if ((p = getenv("RUN_ERL_LOG_ALIVE_IN_UTC")) && strcmp(p,"0")) {
+      ++log_alive_in_gmt;
+  }
+  if ((p = getenv("RUN_ERL_LOG_GENERATIONS"))) {
+    log_generations = atoi(p);
+    if (log_generations < LOG_MIN_GENERATIONS)
+      ERROR1(LOG_ERR,"Minimum RUN_ERL_LOG_GENERATIONS is %d", LOG_MIN_GENERATIONS);
+    if (log_generations > LOG_MAX_GENERATIONS)
+      ERROR1(LOG_ERR,"Maximum RUN_ERL_LOG_GENERATIONS is %d", LOG_MAX_GENERATIONS);
+  }
+
+  if ((p = getenv("RUN_ERL_LOG_MAXSIZE"))) {
+    log_maxsize = atoi(p);
+    if (log_maxsize < LOG_MIN_MAXSIZE)
+      ERROR1(LOG_ERR,"Minimum RUN_ERL_LOG_MAXSIZE is %d", LOG_MIN_MAXSIZE);
+  }
+
+  /*
+   * Create FIFOs and open them 
+   */
+
+  if(*pipename && pipename[strlen(pipename)-1] == '/') {
+    /* The user wishes us to find a unique pipe name in the specified */
+    /* directory */
+    DIR *dirp;
+    struct dirent *direntp;
+
+    calculated_pipename = 1;
+    dirp = opendir(pipename);
+    if(!dirp) {
+      ERRNO_ERR1(LOG_ERR,"Can't access pipe directory '%s'.", pipename);
+      exit(1);
+    }
+
+    /* Check the directory for existing pipes */
+    
+    while((direntp=readdir(dirp)) != NULL) {
+      if(strncmp(direntp->d_name,PIPE_STUBNAME,PIPE_STUBLEN)==0) {
+	int num = atoi(direntp->d_name+PIPE_STUBLEN+1);
+	if(num > highest_pipe_num)
+	  highest_pipe_num = num;
+      }
+    }	
+    closedir(dirp);
+    strn_catf(pipename, sizeof(pipename), "%s.%d",
+	      PIPE_STUBNAME, highest_pipe_num+1);
+  } /* if */
+
+  for(;;) {
+      /* write FIFO - is read FIFO for `to_erl' program */
+      strn_cpy(fifo1, sizeof(fifo1), pipename);
+      strn_cat(fifo1, sizeof(fifo1), ".r");
+      if (create_fifo(fifo1, PERM) < 0) {
+	  ERRNO_ERR1(LOG_ERR,"Cannot create FIFO %s for writing.", fifo1);
+	  exit(1);
+      }
+      
+      /* read FIFO - is write FIFO for `to_erl' program */
+      strn_cpy(fifo2, sizeof(fifo2), pipename);
+      strn_cat(fifo2, sizeof(fifo2), ".w");
+      
+      /* Check that nobody is running run_erl already */
+      if ((fd = sf_open(fifo2, O_WRONLY|DONT_BLOCK_PLEASE, 0)) >= 0) {
+	  /* Open as client succeeded -- run_erl is already running! */
+	  sf_close(fd);
+	  if (calculated_pipename) {
+	      ++highest_pipe_num;
+	      strn_catf(pipename, sizeof(pipename), "%s.%d",
+			PIPE_STUBNAME, highest_pipe_num+1);
+	      continue;
+	  } 
+	  fprintf(stderr, "Erlang already running on pipe %s.\n", pipename);
+	  exit(1);
+      }
+      if (create_fifo(fifo2, PERM) < 0) { 
+	  ERRNO_ERR1(LOG_ERR,"Cannot create FIFO %s for reading.", fifo2);
+	  exit(1);
+      }
+      break;
+  }
 
   /*
    * Open master pseudo-terminal
@@ -250,7 +440,7 @@ int main(int argc, char **argv)
     sf_close(2);
 
     if (dup(sfd) != 0 || dup(sfd) != 1 || dup(sfd) != 2) {
-      erts_run_erl_log_status("Cannot dup\n");
+      status("Cannot dup\n");
     }
     sf_close(sfd);
     exec_shell(argv+off_argv); /* exec_shell expects argv[2] to be */
@@ -293,7 +483,9 @@ static void pass_on(pid_t childpid)
     struct timeval timeout;
     time_t last_activity;
     char buf[BUFSIZ];
-    int rfd, wfd=0;
+    char log_alive_buffer[ALIVE_BUFFSIZ+1];
+    int lognum;
+    int rfd, wfd=0, lfd=0;
     int maxfd;
     int ready;
     int got_some = 0; /* from to_erl */
@@ -308,12 +500,13 @@ static void pass_on(pid_t childpid)
     }
     
 #ifdef DEBUG
-    erts_run_erl_log_status("run_erl: %s opened for reading\n", fifo2);
+    status("run_erl: %s opened for reading\n", fifo2);
 #endif
     
     /* Open the log file */
     
-    erts_run_erl_log_open();
+    lognum = find_next_log_num();
+    lfd = open_log(lognum, O_RDWR|O_APPEND|O_CREAT|O_SYNC);
     
     /* Enter the work loop */
     
@@ -332,8 +525,7 @@ static void pass_on(pid_t childpid)
 	    writefds_ptr = &writefds;
 	}
 	time(&last_activity);
-	/* don't assume old BSD bug */
-	timeout.tv_sec  = erts_run_erl_log_alive_minutes()*60;
+	timeout.tv_sec  = log_alive_minutes*60; /* don't assume old BSD bug */
 	timeout.tv_usec = 0;
 	ready = select(maxfd + 1, &readfds, writefds_ptr, NULL, &timeout);
 	if (ready < 0) {
@@ -363,7 +555,28 @@ static void pass_on(pid_t childpid)
 
 	    /* Check how long time we've been inactive */
 	    time(&now);
-	    erts_run_erl_log_activity(!ready,now,last_activity);
+	    if(!ready || now - last_activity > log_activity_minutes*60) {
+		/* Either a time out: 15 minutes without action, */
+		/* or something is coming in right now, but it's a long time */
+		/* since last time, so let's write a time stamp this message */
+		struct tm *tmptr;
+		if (log_alive_in_gmt) {
+		    tmptr = gmtime(&now);
+		} else {
+		    tmptr = localtime(&now);
+		}
+		if (!strftime(log_alive_buffer, ALIVE_BUFFSIZ, log_alive_format,
+			      tmptr)) {
+		    strn_cpy(log_alive_buffer, sizeof(log_alive_buffer),
+			     "(could not format time in 256 positions "
+			     "with current format string.)");
+		}
+		log_alive_buffer[ALIVE_BUFFSIZ] = '\0';
+
+		sn_printf(buf, sizeof(buf), "\n===== %s%s\n", 
+			  ready?"":"ALIVE ", log_alive_buffer);
+		write_to_log(&lfd, &lognum, buf, strlen(buf));
+	    }
 	}
 
 	/*
@@ -398,7 +611,7 @@ static void pass_on(pid_t childpid)
 	 */
 	if (FD_ISSET(mfd, &readfds)) {
 #ifdef DEBUG
-	    erts_run_erl_log_status("Pty master read; ");
+	    status("Pty master read; ");
 #endif
 	    if ((len = sf_read(mfd, buf, BUFSIZ)) <= 0) {
 		sf_close(rfd);
@@ -416,7 +629,7 @@ static void pass_on(pid_t childpid)
 		exit(0);
 	    }
 
-	    erts_run_erl_log_write(buf, len);
+	    write_to_log(&lfd, &lognum, buf, len);
 
 	    /*
 	     * Save in the output queue.
@@ -432,7 +645,7 @@ static void pass_on(pid_t childpid)
 	 */
 	if (FD_ISSET(rfd, &readfds)) {
 #ifdef DEBUG
-	    erts_run_erl_log_status("FIFO read; ");
+	    status("FIFO read; ");
 #endif
 	    if ((len = sf_read(rfd, buf, BUFSIZ)) < 0) {
 		sf_close(rfd);
@@ -461,7 +674,7 @@ static void pass_on(pid_t childpid)
 		     * should succeed. But in case of error, we just ignore it.
 		     */
 		    if ((wfd = sf_open(fifo1, O_WRONLY|DONT_BLOCK_PLEASE, 0)) < 0) {
-			erts_run_erl_log_status("Client expected on FIFO %s, but can't open (len=%d)\n",
+			status("Client expected on FIFO %s, but can't open (len=%d)\n",
 			       fifo1, len);
 			sf_close(rfd);
 			rfd = sf_open(fifo2, O_RDONLY|DONT_BLOCK_PLEASE, 0);
@@ -473,7 +686,7 @@ static void pass_on(pid_t childpid)
 		    } 
 		    else {
 #ifdef DEBUG
-			erts_run_erl_log_status("run_erl: %s opened for writing\n", fifo1);
+			status("run_erl: %s opened for writing\n", fifo1);
 #endif
 		    }
 		}
@@ -489,15 +702,14 @@ static void pass_on(pid_t childpid)
 
 		/* Write the message */
 #ifdef DEBUG
-		erts_run_erl_log_status("Pty master write; ");
+		status("Pty master write; ");
 #endif
-		len = erts_run_erl_extract_ctrl_seq(buf, len, mfd);
+		len = extract_ctrl_seq(buf, len);
 
 		if(len==1 && buf[0] == '\003') {
 		    kill(childpid,SIGINT);
-		}
-		else if (len>0 && erts_run_erl_write_all(mfd, buf, len) != len)
-		  {
+		} 
+		else if (len>0 && write_all(mfd, buf, len) != len) {
 		    ERRNO_ERR0(LOG_ERR,"Error in writing to terminal.");
 		    sf_close(rfd);
 		    if(wfd) sf_close(wfd);
@@ -506,7 +718,7 @@ static void pass_on(pid_t childpid)
 		}
 	    }
 #ifdef DEBUG
-	    erts_run_erl_log_status("OK\n");
+	    status("OK\n");
 #endif
 	}
     }
@@ -516,6 +728,173 @@ static void catch_sigchild(int sig)
 {
 }
 
+/*
+ * next_log:
+ * Returns the index number that follows the given index number.
+ * (Wrapping after log_generations)
+ */
+static int next_log(int log_num) {
+  return log_num>=log_generations?1:log_num+1;
+}
+
+/*
+ * prev_log:
+ * Returns the index number that precedes the given index number.
+ * (Wrapping after log_generations)
+ */
+static int prev_log(int log_num) {
+  return log_num<=1?log_generations:log_num-1;
+}
+
+/*
+ * find_next_log_num()
+ * Searches through the log directory to check which logs that already
+ * exist. It finds the "hole" in the sequence, and returns the index
+ * number for the last log in the log sequence. If there is no hole, index
+ * 1 is returned.
+ */
+static int find_next_log_num(void) {
+  int i, next_gen, log_gen;
+  DIR *dirp;
+  struct dirent *direntp;
+  int log_exists[LOG_MAX_GENERATIONS+1];
+  int stub_len = strlen(LOG_STUBNAME);
+
+  /* Initialize exiting log table */
+
+  for(i=log_generations; i>=0; i--)
+    log_exists[i] = 0;
+  dirp = opendir(log_dir);
+  if(!dirp) {
+    ERRNO_ERR1(LOG_ERR,"Can't access log directory '%s'", log_dir);
+    exit(1);
+  }
+
+  /* Check the directory for existing logs */
+
+  while((direntp=readdir(dirp)) != NULL) {
+    if(strncmp(direntp->d_name,LOG_STUBNAME,stub_len)==0) {
+      int num = atoi(direntp->d_name+stub_len);
+      if(num < 1 || num > log_generations)
+	continue;
+      log_exists[num] = 1;
+    }
+  }	
+  closedir(dirp);
+
+  /* Find out the next available log file number */
+
+  next_gen = 0;
+  for(i=log_generations; i>=0; i--) {
+    if(log_exists[i])
+      if(next_gen)
+	break;
+      else 
+	;
+    else
+      next_gen = i;
+  }
+
+  /* Find out the current log file number */
+
+  if(next_gen)
+    log_gen = prev_log(next_gen);
+  else
+    log_gen = 1;
+
+  return log_gen;
+} /* find_next_log_num() */
+
+/* open_log()
+ * Opens a log file (with given index) for writing. Writing may be
+ * at the end or a trucnating write, according to flags.
+ * A LOGGING STARTED and time stamp message is inserted into the log file
+ */
+static int open_log(int log_num, int flags)
+{
+  char buf[FILENAME_MAX];
+  time_t now;
+  struct tm *tmptr;
+  char log_buffer[ALIVE_BUFFSIZ+1];
+  int lfd;
+
+  /* Remove the next log (to keep a "hole" in the log sequence) */
+  sn_printf(buf, sizeof(buf), "%s/%s%d",
+	    log_dir, LOG_STUBNAME, next_log(log_num));
+  unlink(buf);
+
+  /* Create or continue on the current log file */
+  sn_printf(buf, sizeof(buf), "%s/%s%d", log_dir, LOG_STUBNAME, log_num);
+  if((lfd = sf_open(buf, flags, LOG_PERM))<0){
+      ERRNO_ERR1(LOG_ERR,"Can't open log file '%s'.", buf);
+    exit(1);
+  }
+
+  /* Write a LOGGING STARTED and time stamp into the log file */
+  time(&now);
+  if (log_alive_in_gmt) {
+      tmptr = gmtime(&now);
+  } else {
+      tmptr = localtime(&now);
+  }
+  if (!strftime(log_buffer, ALIVE_BUFFSIZ, log_alive_format,
+		tmptr)) {
+      strn_cpy(log_buffer, sizeof(log_buffer),
+	      "(could not format time in 256 positions "
+	      "with current format string.)");
+  }
+  log_buffer[ALIVE_BUFFSIZ] = '\0';
+
+  sn_printf(buf, sizeof(buf), "\n=====\n===== LOGGING STARTED %s\n=====\n",
+	    log_buffer);
+  if (write_all(lfd, buf, strlen(buf)) < 0)
+      status("Error in writing to log.\n");
+
+#if USE_FSYNC
+  fsync(lfd);
+#endif
+
+  return lfd;
+}
+
+/* write_to_log()
+ * Writes a message to a log file. If the current log file is full,
+ * a new log file is opened.
+ */
+static void write_to_log(int* lfd, int* log_num, char* buf, int len)
+{
+  int size;
+
+  /* Decide if new logfile needed, and open if so */
+  
+  size = lseek(*lfd,0,SEEK_END);
+  if(size+len > log_maxsize) {
+    sf_close(*lfd);
+    *log_num = next_log(*log_num);
+    *lfd = open_log(*log_num, O_RDWR|O_CREAT|O_TRUNC|O_SYNC); 
+  }
+
+  /* Write to log file */
+
+  if (write_all(*lfd, buf, len) < 0) {
+    status("Error in writing to log.\n");
+  }
+
+#if USE_FSYNC
+  fsync(*lfd);
+#endif
+}
+
+/* create_fifo()
+ * Creates a new fifo with the given name and permission.
+ */
+static int create_fifo(char *name, int perm)
+{
+  if ((mkfifo(name, perm) < 0) && (errno != EEXIST))
+    return -1;
+  return 0;
+}
+
 
 /* open_pty_master()
  * Find a master device, open and return fd and slave device name.
@@ -712,9 +1091,9 @@ static void exec_shell(char **argv)
   else
     argv[0] = sh;
   argv[1] = "-c";
-  erts_run_erl_log_status("Args before exec of shell:\n");
+  status("Args before exec of shell:\n");
   for (vp = argv, i = 0; *vp; vp++, i++)
-    erts_run_erl_log_status("argv[%d] = %s\n", i, *vp);
+    status("argv[%d] = %s\n", i, *vp);
   if (stdstatus) {
       fclose(stdstatus);
   }
@@ -725,6 +1104,26 @@ static void exec_shell(char **argv)
   ERRNO_ERR0(LOG_ERR,"Could not execv");
 }
 
+/* status()
+ * Prints the arguments to a status file
+ * Works like printf (see vfrpintf)
+ */
+static void status(const char *format,...)
+{
+  va_list args;
+  time_t now;
+
+  if (stdstatus == NULL)
+    stdstatus = fopen(statusfile, "w");
+  if (stdstatus == NULL)
+    return;
+  now = time(NULL);
+  fprintf(stdstatus, "run_erl [%d] %s", (int)getpid(), ctime(&now));
+  va_start(args, format);
+  vfprintf(stdstatus, format, args);
+  va_end(args);
+  fflush(stdstatus);
+}
 
 static void daemon_init(void) 
      /* As R Stevens wants it, to a certain extent anyway... */ 
@@ -764,10 +1163,47 @@ static void daemon_init(void)
     run_daemon = 1;
 }
 
+/* error_logf()
+ * Prints the arguments to stderr or syslog
+ * Works like printf (see vfprintf)
+ */
+static void error_logf(int priority, int line, const char *format, ...)
+{
+    va_list args;
+    va_start(args, format);
+
+#ifdef HAVE_SYSLOG_H
+    if (run_daemon) {
+	vsyslog(priority,format,args);
+    }
+    else
+#endif
+    {
+	time_t now = time(NULL);
+	fprintf(stderr, "run_erl:%d [%d] %s", line, (int)getpid(), ctime(&now));
+	vfprintf(stderr, format, args);
+    }
+    va_end(args);
+}	
+
 static void usage(char *pname)
 {
-  fprintf(stderr, "Usage: ");
-  fprintf(stderr, RUN_ERL_USAGE, pname);
+  fprintf(stderr, "Usage: %s (pipe_name|pipe_dir/) log_dir \"command [parameters ...]\"\n", pname);
+  fprintf(stderr, "\nYou may also set the environment variables RUN_ERL_LOG_GENERATIONS\n");
+  fprintf(stderr, "and RUN_ERL_LOG_MAXSIZE to the number of log files to use and the\n");
+  fprintf(stderr, "size of the log file when to switch to the next log file\n");
+}
+
+/* Instead of making sure basename exists, we do our own */
+static char *simple_basename(char *path)
+{
+    char *ptr;
+    for (ptr = path; *ptr != '\0'; ++ptr) {
+	if (*ptr == '/') {
+	    path = ptr + 1;
+	}
+    }
+    return path;
 }
 
 static void init_outbuf(void)
@@ -838,6 +1274,114 @@ static void outbuf_append(const char* buf, int n)
     outbuf_in += n;
 }
 
+/* Call write() until entire buffer has been written or error.
+ * Return len or -1.
+ */
+static int write_all(int fd, const char* buf, int len)
+{
+    int left = len;
+    int written;
+    for (;;) {
+	written = sf_write(fd,buf,left);
+	if (written == left) {
+	    return len;
+	}
+	if (written < 0) {
+	    return -1;
+	}
+	left -= written;
+	buf += written;
+    }
+}
+
+static ssize_t sf_read(int fd, void *buffer, size_t len) {
+    ssize_t n = 0;
+
+    do { n = read(fd, buffer, len); } while (n < 0 && errno == EINTR);
+
+    return n;
+}
+
+static ssize_t sf_write(int fd, const void *buffer, size_t len) {
+    ssize_t n = 0;
+
+    do { n = write(fd, buffer, len); } while (n < 0 && errno == EINTR);
+
+    return n;
+}
+
+static int sf_open(const char *path, int type, mode_t mode) {
+    int fd = 0;
+
+    do { fd = open(path, type, mode); } while(fd < 0 && errno == EINTR);
+
+    return fd;
+}
+static int sf_close(int fd) {
+    int res = 0;
+
+    do { res = close(fd); } while(fd < 0 && errno == EINTR);
+
+    return res;
+}
+/* Extract any control sequences that are ment only for run_erl
+ * and should not be forwarded to the pty.
+ */
+static int extract_ctrl_seq(char* buf, int len)
+{
+    static const char prefix[] = "\033_";
+    static const char suffix[] = "\033\\";
+    char* bufend = buf + len;
+    char* start = buf;
+    char* command;
+    char* end;
+    
+    for (;;) {
+	start = find_str(start, bufend-start, prefix);
+	if (!start) break;
+	
+	command = start + strlen(prefix);
+	end = find_str(command, bufend-command, suffix);
+	if (end) {
+	    unsigned col, row;
+	    if (sscanf(command,"version=%u", &protocol_ver)==1) {
+		/*fprintf(stderr,"to_erl v%u\n", protocol_ver);*/
+	    }
+	    else if (sscanf(command,"winsize=%u,%u", &col, &row)==2) {
+		set_window_size(col,row);
+	    }
+	    else {
+		ERROR2(LOG_ERR, "Ignoring unknown ctrl command '%.*s'\n",
+		       (int)(end-command), command);
+	    }
+	  
+	    /* Remove ctrl sequence from buf */
+	    end += strlen(suffix);
+	    memmove(start, end, bufend-end);
+	    bufend -= end - start;
+	}
+	else {
+	    ERROR2(LOG_ERR, "Missing suffix in ctrl sequence '%.*s'\n",
+		   (int)(bufend-start), start);
+	    break;
+	}
+    }
+    return bufend - buf;
+}
+
+static void set_window_size(unsigned col, unsigned row)
+{
+#ifdef TIOCSWINSZ	
+    struct winsize ws;
+    ws.ws_col = col;
+    ws.ws_row = row;
+    if (ioctl(mfd, TIOCSWINSZ, &ws) < 0) {
+	ERRNO_ERR0(LOG_ERR,"Failed to set window size");
+    }
+#endif
+}
+
+
 #ifdef DEBUG
 
 #define S(x)  ((x) > 0 ? 1 : 0)
diff --git a/erts/etc/common/run_erl_vsn.h b/erts/etc/unix/run_erl.h
index 2c3e67e81c..cc70a98e52 100644
--- a/erts/etc/common/run_erl_vsn.h
+++ b/erts/etc/unix/run_erl.h
@@ -1,8 +1,8 @@
 /*
  * %CopyrightBegin%
- *
+ * 
  * Copyright Ericsson AB 2008-2009. All Rights Reserved.
- *
+ * 
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
  * You may obtain a copy of the License at
@@ -14,7 +14,7 @@
  * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
  * See the License for the specific language governing permissions and
  * limitations under the License.
- *
+ * 
  * %CopyrightEnd%
  */
 
@@ -28,3 +28,4 @@
  * 0: Older, without version handshake
  * 1: R12B-3, version handshake + window size ctrl
  */
+
diff --git a/erts/etc/common/safe_string.c b/erts/etc/unix/safe_string.c
index cdcdbf16f0..a5c11d41d8 100644
--- a/erts/etc/common/safe_string.c
+++ b/erts/etc/unix/safe_string.c
@@ -1,8 +1,8 @@
 /*
  * %CopyrightBegin%
- *
+ * 
  * Copyright Ericsson AB 2008-2009. All Rights Reserved.
- *
+ * 
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
  * You may obtain a copy of the License at
@@ -14,12 +14,12 @@
  * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
  * See the License for the specific language governing permissions and
  * limitations under the License.
- *
+ * 
  * %CopyrightEnd%
  */
-/*
+/* 
  * Module: safe_string.c
- *
+ * 
  * This is a bunch of generic string operation
  * that are safe regarding buffer overflow.
  *
@@ -121,3 +121,4 @@ void* memmove(void *dest, const void *src, size_t n)
     return dest;
 }
 #endif /* HAVE_MEMMOVE */
+
diff --git a/erts/etc/common/safe_string.h b/erts/etc/unix/safe_string.h
index f9d2b2023a..5a471f10de 100644
--- a/erts/etc/common/safe_string.h
+++ b/erts/etc/unix/safe_string.h
@@ -1,8 +1,8 @@
 /*
  * %CopyrightBegin%
- *
+ * 
  * Copyright Ericsson AB 2008-2009. All Rights Reserved.
- *
+ * 
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
  * You may obtain a copy of the License at
@@ -14,12 +14,12 @@
  * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
  * See the License for the specific language governing permissions and
  * limitations under the License.
- *
+ * 
  * %CopyrightEnd%
  */
-/*
+/* 
  * Module: safe_string.h
- *
+ * 
  * This is an interface to a bunch of generic string operation
  * that are safe regarding buffer overflow.
  *
@@ -63,3 +63,4 @@ char* find_str(const char* haystack, int size, const char* needle);
 #ifndef HAVE_MEMMOVE
 void* memmove(void *dest, const void *src, size_t n);
 #endif
+
diff --git a/erts/etc/unix/to_erl.c b/erts/etc/unix/to_erl.c
index 82d3218964..0bd469727c 100644
--- a/erts/etc/unix/to_erl.c
+++ b/erts/etc/unix/to_erl.c
@@ -1,7 +1,7 @@
 /*
  * %CopyrightBegin%
  * 
- * Copyright Ericsson AB 1996-2013. All Rights Reserved.
+ * Copyright Ericsson AB 1996-2015. All Rights Reserved.
  * 
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
@@ -17,9 +17,592 @@
  * 
  * %CopyrightEnd%
  */
+/* 
+ * Module: to_erl.c
+ * 
+ * This module implements a process that opens two specified FIFOs, one
+ * for reading and one for writing; reads from its stdin, and writes what
+ * it has read to the write FIF0; reads from the read FIFO, and writes to
+ * its stdout.
+ *
+  ________                            _________ 
+ |        |--<-- pipe.r (fifo1) --<--|         |
+ | to_erl |                          | run_erl | (parent)
+ |________|-->-- pipe.w (fifo2) -->--|_________|
+                                          ^ master pty
+                                          |
+                                          | slave pty
+                                      ____V____ 
+                                     |         |
+                                     |  "erl"  | (child)
+                                     |_________|
+ */
+#ifdef HAVE_CONFIG_H
+#  include "config.h"
+#endif
+
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <sys/time.h>
+#include <sys/types.h>
+#include <fcntl.h>
+#include <unistd.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <termios.h>
+#include <dirent.h>
+#include <signal.h>
+#include <errno.h>
+#ifdef HAVE_SYS_IOCTL_H
+#  include <sys/ioctl.h>
+#endif
+
+#include "run_erl.h"
+#include "safe_string.h"   /* strn_cpy, strn_catf, sn_printf, etc. */
+
+#if defined(O_NONBLOCK)
+# define DONT_BLOCK_PLEASE O_NONBLOCK
+#else
+# define DONT_BLOCK_PLEASE O_NDELAY
+# if !defined(EAGAIN)
+#  define EAGAIN -3898734
+# endif
+#endif
+
+#ifdef HAVE_STRERROR
+#  define STRERROR(x) strerror(x)
+#else
+#  define STRERROR(x) ""
+#endif
+
+#define noDEBUG
+
+#define PIPE_DIR        "/tmp/"
+#define PIPE_STUBNAME   "erlang.pipe"
+#define PIPE_STUBLEN    strlen(PIPE_STUBNAME)
+
+#ifdef DEBUG
+#define STATUS(s)  { fprintf(stderr, (s)); fflush(stderr); }
+#else
+#define STATUS(s)
+#endif
+
+#ifndef FILENAME_MAX
+#define FILENAME_MAX 250
+#endif
+
+static struct termios tty_smode, tty_rmode;
+static int tty_eof = 0;
+static int recv_sig = 0;
+static int protocol_ver = RUN_ERL_LO_VER; /* assume lowest to begin with */
+
+static int write_all(int fd, const char* buf, int len);
+static int window_size_seq(char* buf, size_t bufsz);
+static int version_handshake(char* buf, int len, int wfd);
+#ifdef DEBUG
+static void show_terminal_settings(struct termios *);
+#endif
+
+static void handle_ctrlc(int sig)
+{
+    /* Reinstall the handler, and signal break flag */
+    signal(SIGINT,handle_ctrlc);
+    recv_sig = SIGINT;
+}  
+
+static void handle_sigwinch(int sig)
+{
+    recv_sig = SIGWINCH;
+}
+
+static void usage(char *pname)
+{
+    fprintf(stderr, "Usage: %s [-h|-F] [pipe_name|pipe_dir/]\n", pname);
+    fprintf(stderr, "\t-h\tThis help text.\n");
+    fprintf(stderr, "\t-F\tForce connection even though pipe is locked by other to_erl process.\n");
+}
+
+int main(int argc, char **argv)
+{
+    char  FIFO1[FILENAME_MAX], FIFO2[FILENAME_MAX];
+    int i, len, wfd, rfd;
+    fd_set readfds;
+    char buf[BUFSIZ];
+    char pipename[FILENAME_MAX];
+    int pipeIx = 1;
+    int force_lock = 0;
+    int got_some = 0;
+
+    if (argc >= 2 && argv[1][0]=='-') {
+	switch (argv[1][1]) {
+	case 'h':
+	    usage(argv[0]);
+	    exit(1);
+	case 'F':
+	    force_lock = 1;
+	    break;
+	default:
+	    fprintf(stderr,"Invalid option '%s'\n",argv[1]);
+	    exit(1);
+	}
+	pipeIx = 2;
+    }
+    
+#ifdef DEBUG
+    fprintf(stderr, "%s: pid is : %d\n", argv[0], (int)getpid());
+#endif
+    
+    strn_cpy(pipename, sizeof(pipename),
+	     (argv[pipeIx] ? argv[pipeIx] : PIPE_DIR));
+    
+    if(*pipename && pipename[strlen(pipename)-1] == '/') {
+	/* The user wishes us to find a pipe name in the specified */
+	/* directory */
+	int highest_pipe_num = 0;
+	DIR *dirp;
+	struct dirent *direntp;
+
+	dirp = opendir(pipename);
+	if(!dirp) {
+	    fprintf(stderr, "Can't access pipe directory %s: %s\n", pipename, strerror(errno));
+	    exit(1);
+	}
+
+	/* Check the directory for existing pipes */
+    
+	while((direntp=readdir(dirp)) != NULL) {
+	    if(strncmp(direntp->d_name,PIPE_STUBNAME,PIPE_STUBLEN)==0) {
+		int num = atoi(direntp->d_name+PIPE_STUBLEN+1);
+		if(num > highest_pipe_num)
+		    highest_pipe_num = num;
+	    }
+	}	
+	closedir(dirp);
+	strn_catf(pipename, sizeof(pipename), (highest_pipe_num?"%s.%d":"%s"),
+		  PIPE_STUBNAME, highest_pipe_num);
+    } /* if */
+
+    /* read FIFO */
+    sn_printf(FIFO1,sizeof(FIFO1),"%s.r",pipename);
+    /* write FIFO */
+    sn_printf(FIFO2,sizeof(FIFO2),"%s.w",pipename);
+
+    /* Check that nobody is running to_erl on this pipe already */
+    if ((wfd = open (FIFO1, O_WRONLY|DONT_BLOCK_PLEASE, 0)) >= 0) {
+	/* Open as server succeeded -- to_erl is already running! */
+	close(wfd);
+	fprintf(stderr, "Another to_erl process already attached to pipe "
+			"%s.\n", pipename);
+	if (force_lock) {
+	    fprintf(stderr, "But we proceed anyway by force (-F).\n");
+	} 
+	else {
+	    exit(1);
+	}
+    }
+
+    if ((rfd = open (FIFO1, O_RDONLY|DONT_BLOCK_PLEASE, 0)) < 0) {
+#ifdef DEBUG
+	fprintf(stderr, "Could not open FIFO %s for reading.\n", FIFO1);
+#endif
+	fprintf(stderr, "No running Erlang on pipe %s: %s\n", pipename, strerror(errno));
+	exit(1);
+    }
+#ifdef DEBUG
+    fprintf(stderr, "to_erl: %s opened for reading\n", FIFO1);
+#endif
+    
+    if ((wfd = open (FIFO2, O_WRONLY|DONT_BLOCK_PLEASE, 0)) < 0) {
+#ifdef DEBUG
+	fprintf(stderr, "Could not open FIFO %s for writing.\n", FIFO2);
+#endif
+	fprintf(stderr, "No running Erlang on pipe %s: %s\n", pipename, strerror(errno));
+	close(rfd);
+	exit(1);
+    }
+#ifdef DEBUG
+    fprintf(stderr, "to_erl: %s opened for writing\n", FIFO2);
+#endif
+    
+    fprintf(stderr, "Attaching to %s (^D to exit)\n\n", pipename);
+    
+    /* Set break handler to our handler */
+    signal(SIGINT,handle_ctrlc);
+
+    /* 
+     * Save the current state of the terminal, and set raw mode.
+     */
+    if (tcgetattr(0, &tty_rmode) , 0) {
+	fprintf(stderr, "Cannot get terminals current mode\n");
+	exit(-1);
+    }
+    tty_smode = tty_rmode;
+    tty_eof = '\004'; /* Ctrl+D to exit */
+#ifdef DEBUG
+    show_terminal_settings(&tty_rmode);
+#endif
+    tty_smode.c_iflag =
+	1*BRKINT |/*Signal interrupt on break.*/
+	    1*IGNPAR |/*Ignore characters with parity errors.*/
+		1*ISTRIP |/*Strip character.*/
+		    0;
+    
+#if 0
+0*IGNBRK |/*Ignore break condition.*/
+0*PARMRK |/*Mark parity errors.*/
+0*INPCK  |/*Enable input parity check.*/
+0*INLCR  |/*Map NL to CR on input.*/
+0*IGNCR  |/*Ignore CR.*/
+0*ICRNL  |/*Map CR to NL on input.*/
+0*IUCLC  |/*Map upper-case to lower-case on input.*/
+0*IXON   |/*Enable start/stop output control.*/
+0*IXANY  |/*Enable any character to restart output.*/
+0*IXOFF  |/*Enable start/stop input control.*/
+0*IMAXBEL|/*Echo BEL on input line too long.*/
+#endif
+						
+    tty_smode.c_oflag =
+	1*OPOST  |/*Post-process output.*/
+	    1*ONLCR  |/*Map NL to CR-NL on output.*/
+#ifdef XTABS
+		1*XTABS  |/*Expand tabs to spaces. (Linux)*/
+#endif
+#ifdef OXTABS
+		    1*OXTABS  |/*Expand tabs to spaces. (FreeBSD)*/
+#endif
+#ifdef NL0
+			1*NL0    |/*Select newline delays*/
+#endif
+#ifdef CR0
+			    1*CR0    |/*Select carriage-return delays*/
+#endif
+#ifdef TAB0
+				1*TAB0   |/*Select horizontal tab delays*/
+#endif
+#ifdef BS0
+				    1*BS0    |/*Select backspace delays*/
+#endif
+#ifdef VT0
+					1*VT0    |/*Select vertical tab delays*/
+#endif
+#ifdef FF0
+					    1*FF0    |/*Select form feed delays*/
+#endif
+											    0;
+    
+#if 0
+0*OLCUC  |/*Map lower case to upper on output.*/
+0*OCRNL  |/*Map CR to NL on output.*/
+0*ONOCR  |/*No CR output at column 0.*/
+0*ONLRET |/*NL performs CR function.*/
+0*OFILL  |/*Use fill characters for delay.*/
+0*OFDEL  |/*Fill is DEL, else NULL.*/
+0*NL1    |
+0*CR1    |
+0*CR2    |
+0*CR3    |
+0*TAB1   |
+0*TAB2   |
+0*TAB3   |/*Expand tabs to spaces.*/
+0*BS1    |
+0*VT1    |
+0*FF1    |
+#endif
+								    
+    /* JALI: removed setting the tty_smode.c_cflag flags, since this is not */
+    /* advisable if this is a *real* terminal, such as the console. In fact */
+    /* this may hang the entire machine, deep, deep down (signalling break */
+    /* or toggling the abort switch doesn't help) */
+    
+    tty_smode.c_lflag =
+									0;
+    
+#if 0
+0*ISIG   |/*Enable signals.*/
+0*ICANON |/*Canonical input (erase and kill processing).*/
+0*XCASE  |/*Canonical upper/lower presentation.*/
+0*ECHO   |/*Enable echo.*/
+0*ECHOE  |/*Echo erase character as BS-SP-BS.*/
+0*ECHOK  |/*Echo NL after kill character.*/
+0*ECHONL |/*Echo NL.*/
+0*NOFLSH |/*Disable flush after interrupt or quit.*/
+0*TOSTOP |/*Send SIGTTOU for background output.*/
+0*ECHOCTL|/*Echo control characters as ^char, delete as ^?.*/
+0*ECHOPRT|/*Echo erase character as character erased.*/
+0*ECHOKE |/*BS-SP-BS erase entire line on line kill.*/
+0*FLUSHO |/*Output is being flushed.*/
+0*PENDIN |/*Retype pending input at next read or input character.*/
+0*IEXTEN |/*Enable extended (implementation-defined) functions.*/
+#endif
+								
+    tty_smode.c_cc[VMIN]      =0;/* Note that VMIN is the same as VEOF! */
+    tty_smode.c_cc[VTIME]     =0;/* Note that VTIME is the same as VEOL! */
+    tty_smode.c_cc[VINTR]     =3;
+    
+    tcsetattr(0, TCSADRAIN, &tty_smode);
+    
+#ifdef DEBUG
+    show_terminal_settings(&tty_smode);
+#endif
+    /*
+     * 	 "Write a ^L to the FIFO which causes the other end to redisplay
+     *    the input line."
+     * This does not seem to work as was intended in old comment above.
+     * However, this control character is now (R12B-3) used by run_erl
+     * to trigger the version handshaking between to_erl and run_erl
+     * at the start of every new to_erl-session.
+     */
+
+    if (write(wfd, "\014", 1) < 0) {
+	fprintf(stderr, "Error in writing ^L to FIFO.\n");
+    }
+
+    /*
+     * read and write
+     */
+    while (1) {
+	FD_ZERO(&readfds);
+	FD_SET(0, &readfds);
+	FD_SET(rfd, &readfds);
+	if (select(rfd + 1, &readfds, NULL, NULL, NULL) < 0) {
+	    if (recv_sig) {
+		FD_ZERO(&readfds);
+	    }
+	    else {
+		fprintf(stderr, "Error in select.\n");
+		break;
+	    }
+	}
+	len = 0;
+
+	/*
+	 * Read from terminal and write to FIFO
+         */
+	if (recv_sig) {
+	    switch (recv_sig) {
+	    case SIGINT:
+		fprintf(stderr, "[Break]\n\r");
+		buf[0] = '\003';
+		len = 1;
+		break;
+	    case SIGWINCH:
+		len = window_size_seq(buf,sizeof(buf));
+		break;
+	    default:
+		fprintf(stderr,"Unexpected signal: %u\n",recv_sig);
+	    }
+	    recv_sig = 0;
+	}
+	else if (FD_ISSET(0, &readfds)) {
+	    len = read(0, buf, sizeof(buf));
+	    if (len <= 0) {
+		close(rfd);
+		close(wfd);
+		if (len < 0) {
+		    fprintf(stderr, "Error in reading from stdin.\n");
+		} else {
+		    fprintf(stderr, "[EOF]\n\r");
+		}
+		break;
+	    }
+	    /* check if there is an eof character in input */
+	    for (i = 0; i < len && buf[i] != tty_eof; i++);
+	    if (buf[i] == tty_eof) {
+		fprintf(stderr, "[Quit]\n\r");
+		break;
+	    }
+	}
+
+	if (len) {
+#ifdef DEBUG
+	    if(write(1, buf, len));
+#endif
+	    if (write_all(wfd, buf, len) != len) {
+		fprintf(stderr, "Error in writing to FIFO.\n");
+		close(rfd);
+		close(wfd);
+		break;
+	    }
+	    STATUS("\" OK\r\n");
+	}
+
+	/*
+	 * Read from FIFO, write to terminal.
+	 */
+	if (FD_ISSET(rfd, &readfds)) {
+	    STATUS("FIFO read: ");
+	    len = read(rfd, buf, BUFSIZ);
+	    if (len < 0 && errno == EAGAIN) {
+		/*
+		 * No data this time, but the writing end of the FIFO is still open.
+		 * Do nothing.
+		 */
+		;
+	    } else if (len <= 0) {
+		/*
+		 * Either an error or end of file. In either case, break out
+		 * of the loop.
+		 */
+		close(rfd);
+		close(wfd);
+		if (len < 0) {
+		    fprintf(stderr, "Error in reading from FIFO.\n");
+		} else
+		    fprintf(stderr, "[End]\n\r");
+		break;
+	    } else {
+		if (!got_some) {
+		    if ((len=version_handshake(buf,len,wfd)) < 0) {
+			close(rfd);
+			close(wfd);
+			break;
+		    }
+		    if (protocol_ver >= 1) {
+			/* Tell run_erl size of terminal window */
+			signal(SIGWINCH, handle_sigwinch);
+			raise(SIGWINCH);
+		    }
+		    got_some = 1;
+		}
+
+		/*
+		 * We successfully read at least one character. Write what we got.
+		 */
+		STATUS("Terminal write: \"");
+		if (write_all(1, buf, len) != len) {
+		    fprintf(stderr, "Error in writing to terminal.\n");
+		    close(rfd);
+		    close(wfd);
+		    break;
+		}
+		STATUS("\" OK\r\n");
+	    }
+	}
+    }
+
+    /* 
+     * Reset terminal characterstics 
+     * XXX
+     */
+    tcsetattr(0, TCSADRAIN, &tty_rmode);
+    return 0;
+}
+
+/* Call write() until entire buffer has been written or error.
+ * Return len or -1.
+ */
+static int write_all(int fd, const char* buf, int len)
+{
+    int left = len;
+    int written;
+    while (left) {
+	written = write(fd,buf,left);
+	if (written < 0) {
+	    return -1;
+	}
+	left -= written;
+	buf += written;
+    }
+    return len;
+}
+
+static int window_size_seq(char* buf, size_t bufsz)
+{
+#ifdef TIOCGWINSZ
+    struct winsize ws;
+    static const char prefix[] = "\033_";
+    static const char suffix[] = "\033\\";
+    /* This Esc sequence is called "Application Program Command"
+       and seems suitable to use for our own customized stuff. */
+
+    if (ioctl(STDIN_FILENO, TIOCGWINSZ, &ws) == 0) {
+	int len = sn_printf(buf, bufsz, "%swinsize=%u,%u%s",
+			    prefix, ws.ws_col, ws.ws_row, suffix);
+	return len;
+    }
+#endif /* TIOCGWINSZ */
+    return 0;
+}
+
+/*   to_erl                     run_erl
+ *     |                           |
+ *     |---------- '\014' -------->| (session start)
+ *     |                           |
+ *     |<---- "[run_erl v1-0]" ----| (version interval)
+ *     |                           |
+ *     |--- Esc_"version=1"Esc\ -->| (common version)
+ *     |                           |
+ */
+static int version_handshake(char* buf, int len, int wfd)
+{
+    unsigned re_high=0, re_low;
+    char *end = find_str(buf,len,"]\n");
+    
+    if (end && sscanf(buf,"[run_erl v%u-%u",&re_high,&re_low)==2) {
+	char wbuf[30];
+	int wlen;
+
+	if (re_low > RUN_ERL_HI_VER || re_high < RUN_ERL_LO_VER) {
+	    fprintf(stderr,"Incompatible versions: to_erl=v%u-%u run_erl=v%u-%u\n",
+		    RUN_ERL_HI_VER, RUN_ERL_LO_VER, re_high, re_low);
+	    return -1;
+	}
+	/* Choose highest common version */
+	protocol_ver = re_high < RUN_ERL_HI_VER ? re_high : RUN_ERL_HI_VER;
+
+	wlen = sn_printf(wbuf, sizeof(wbuf), "\033_version=%u\033\\",
+			 protocol_ver);
+	if (write_all(wfd, wbuf, wlen) < 0) {
+	    fprintf(stderr,"Failed to send version handshake\n");
+	    return -1;
+	}
+	end += 2;
+	len -= (end-buf);
+	memmove(buf,end,len);
+
+    }
+    else {  /* we assume old run_erl without version handshake */
+	protocol_ver = 0;
+    }
+
+    if (re_high != RUN_ERL_HI_VER) {
+	fprintf(stderr,"run_erl has different version, "
+		"using common protocol level %u\n", protocol_ver);
+    }
+
+    return len;
+}
+
 
-#include "to_erl_common.h"
+#ifdef DEBUG
+#define S(x)  ((x) > 0 ? 1 : 0)
 
-int main(int argc,char **argv) {
-  return to_erl(argc,argv);
+static void show_terminal_settings(struct termios *t)
+{
+  fprintf(stderr,"c_iflag:\n");
+  fprintf(stderr,"Signal interrupt on break:   BRKINT  %d\n", S(t->c_iflag & BRKINT));
+  fprintf(stderr,"Map CR to NL on input:       ICRNL   %d\n", S(t->c_iflag & ICRNL));
+  fprintf(stderr,"Ignore break condition:      IGNBRK  %d\n", S(t->c_iflag & IGNBRK));
+  fprintf(stderr,"Ignore CR:                   IGNCR   %d\n", S(t->c_iflag & IGNCR));
+  fprintf(stderr,"Ignore char with par. err's: IGNPAR  %d\n", S(t->c_iflag & IGNPAR));
+  fprintf(stderr,"Map NL to CR on input:       INLCR   %d\n", S(t->c_iflag & INLCR));
+  fprintf(stderr,"Enable input parity check:   INPCK   %d\n", S(t->c_iflag & INPCK));
+  fprintf(stderr,"Strip character              ISTRIP  %d\n", S(t->c_iflag & ISTRIP));
+  fprintf(stderr,"Enable start/stop input ctrl IXOFF   %d\n", S(t->c_iflag & IXOFF));
+  fprintf(stderr,"ditto output ctrl            IXON    %d\n", S(t->c_iflag & IXON));
+  fprintf(stderr,"Mark parity errors           PARMRK  %d\n", S(t->c_iflag & PARMRK));
+  fprintf(stderr,"\n");
+  fprintf(stderr,"c_oflag:\n");
+  fprintf(stderr,"Perform output processing    OPOST   %d\n", S(t->c_oflag & OPOST));
+  fprintf(stderr,"\n");
+  fprintf(stderr,"c_cflag:\n");
+  fprintf(stderr,"Ignore modem status lines    CLOCAL  %d\n", S(t->c_cflag & CLOCAL));
+  fprintf(stderr,"\n");
+  fprintf(stderr,"c_local:\n");
+  fprintf(stderr,"Enable echo                  ECHO    %d\n", S(t->c_lflag & ECHO));
+  fprintf(stderr,"\n");
+  fprintf(stderr,"c_cc:\n");
+  fprintf(stderr,"c_cc[VEOF]                           %d\n", t->c_cc[VEOF]);
 }
+#endif
diff --git a/erts/etc/win32/msys_tools/vc/cc.sh b/erts/etc/win32/msys_tools/vc/cc.sh
index ad05e5375b..ac89aac34e 100644
--- a/erts/etc/win32/msys_tools/vc/cc.sh
+++ b/erts/etc/win32/msys_tools/vc/cc.sh
@@ -242,7 +242,7 @@ for x in $SOURCES; do
     if [ $PREPROCESSING = true ]; then
 	output_flag="-E"
     else
-	output_flag="-c -Fo`cmd //C echo ${output_filename}`"
+	output_flag="-FS -c -Fo`cmd //C echo ${output_filename}`"
     fi
     params="$COMMON_CFLAGS $MD $DEBUG_FLAGS $OPTIMIZE_FLAGS \
 	    $CMD ${output_flag} $MPATH"
@@ -250,6 +250,8 @@ for x in $SOURCES; do
 	echo cc.sh "$SAVE" >>$CC_SH_DEBUG_LOG
 	echo cl.exe $params >>$CC_SH_DEBUG_LOG
     fi
+    # MSYS2 (currently) converts the paths wrong, avoid it
+    export MSYS2_ARG_CONV_EXCL=-FoC
     eval cl.exe $params >$MSG_FILE 2>$ERR_FILE
     RES=$?
     if test $PREPROCESSING = false; then
@@ -274,6 +276,7 @@ for x in $SOURCES; do
     fi
     rm -f $ERR_FILE $MSG_FILE
     if [ $RES != 0 ]; then
+	echo Failed: cl.exe $params
 	rm -rf $TMPOBJDIR
 	exit $RES
     fi
@@ -312,7 +315,10 @@ if [ $LINKING = true ]; then
 	    stdlib="-lLIBMTD";;
     esac
     # And finally call the next script to do the linking...
-    params="$out_spec $LINKCMD $stdlib" 
+    params="$out_spec $LINKCMD $stdlib"
+    if [ "X$CC_SH_DEBUG_LOG" != "X" ]; then
+	echo ld.sh $ACCUM_OBJECTS $params
+    fi
     eval ld.sh $ACCUM_OBJECTS $params
     RES=$?
 fi
diff --git a/erts/etc/win32/msys_tools/vc/emu_cc.sh b/erts/etc/win32/msys_tools/vc/emu_cc.sh
index 01f75b2468..10d59214ea 100644
--- a/erts/etc/win32/msys_tools/vc/emu_cc.sh
+++ b/erts/etc/win32/msys_tools/vc/emu_cc.sh
@@ -29,6 +29,7 @@ WTOOLDIR0=`win2msys_path.sh "$TOOLDIR"`
 WTOOLDIR=`cmd //C echo $WTOOLDIR0`
 # Do primitive 'make'
 newer_exe=`find $TOOLDIR -newer $COFFIX.c -name coffix.exe -print`
+export MSYS2_ARG_CONV_EXCL="-FeC"
 if [ -z $newer_exe ]; then
     echo recompiling $COFFIX.exe
     cl.exe -Fe${WTOOLDIR}/coffix.exe ${WTOOLDIR}/coffix.c
diff --git a/erts/etc/win32/msys_tools/vc/mc.sh b/erts/etc/win32/msys_tools/vc/mc.sh
index e9ea9ff9a9..14b5ebaa8f 100644
--- a/erts/etc/win32/msys_tools/vc/mc.sh
+++ b/erts/etc/win32/msys_tools/vc/mc.sh
@@ -80,9 +80,14 @@ if [ -n "$OUTPUT_DIRNAME" ]; then
 	exit $RES
     fi
 fi
+# MSYS2 (currently) converts the paths wrong, avoid it
+export MSYS2_ARG_CONV_EXCL=
 eval $MCC "$CMD"  >/tmp/mc.exe.${p}.1 2>/tmp/mc.exe.${p}.2
 RES=$?
-tail +2 /tmp/mc.exe.${p}.2 >&2
+if [ $RES != 0 ]; then
+    echo Failed: $MCC "$CMD" 
+fi
+tail -n +2 /tmp/mc.exe.${p}.2 >&2
 cat /tmp/mc.exe.${p}.1
 rm -f /tmp/mc.exe.${p}.2 /tmp/mc.exe.${p}.1
 exit $RES
diff --git a/erts/etc/win32/msys_tools/vc/rc.sh b/erts/etc/win32/msys_tools/vc/rc.sh
index 1b3b1c85bd..1f8ade17cb 100644
--- a/erts/etc/win32/msys_tools/vc/rc.sh
+++ b/erts/etc/win32/msys_tools/vc/rc.sh
@@ -79,9 +79,14 @@ if [ "X$RC_SH_DEBUG_LOG" != "X" ]; then
     echo rc.sh "$SAVE" >>$RC_SH_DEBUG_LOG
     echo rc.exe $CMD >>$RC_SH_DEBUG_LOG
 fi
+# MSYS2 (currently) converts the paths wrong, avoid it
+export MSYS2_ARG_CONV_EXCL=-Fo
 eval $RCC "$CMD"  >/tmp/rc.exe.${p}.1 2>/tmp/rc.exe.${p}.2
 RES=$?
-tail +2 /tmp/rc.exe.${p}.2 >&2
+if [ $RES != 0 ]; then
+    echo Failed: $RCC "$CMD" 
+fi
+tail -n +2 /tmp/rc.exe.${p}.2 >&2
 cat /tmp/rc.exe.${p}.1
 rm -f /tmp/rc.exe.${p}.2 /tmp/rc.exe.${p}.1
 exit $RES
diff --git a/erts/etc/win32/nsis/Makefile b/erts/etc/win32/nsis/Makefile
index 49d835170a..64f44ff86d 100644
--- a/erts/etc/win32/nsis/Makefile
+++ b/erts/etc/win32/nsis/Makefile
@@ -42,7 +42,13 @@ include $(ERL_TOP)/make/otp_release_targets.mk
 
 TARGET_DIR = $(RELEASE_PATH)
 
-ifeq ($(MSYSTEM),MINGW32)
+ifdef MSYSTEM
+ ifeq ($(MSYSTEM),$(filter $(MSYSTEM),MSYS MINGW32 MINGW64))
+  USEMSYS := true
+ endif
+endif
+
+ifeq ($(USEMSYS),true)
 
   MAKENSISFLAGS = //V2
   WTESTROOT=$(shell (msys2win_path.sh "$(RELEASE_PATH)"))
@@ -63,7 +69,7 @@ else
 endif
 
 REDIST_FILE=$(shell (sh ./find_redist.sh || echo ""))
-ifeq ($(MSYSTEM),MINGW32)
+ifeq ($(USEMSYS),true)
 	NICEREDISTFILE=$(shell (msys2win_path.sh -m "$(REDIST_FILE)" 2>/dev/null || echo ""))
 else
 	NICEREDISTFILE=$(shell (cygpath -d -m "$(REDIST_FILE)" 2>/dev/null || echo ""))
diff --git a/erts/etc/win32/nsis/find_redist.sh b/erts/etc/win32/nsis/find_redist.sh
index c0895c9dd5..03e92b21c7 100755
--- a/erts/etc/win32/nsis/find_redist.sh
+++ b/erts/etc/win32/nsis/find_redist.sh
@@ -164,7 +164,7 @@ fi
 
 #echo $BPATH_LIST
 for BP in $BPATH_LIST; do
-    for verdir in "sdk v2.0" "sdk v3.5" "v6.0A" "v7.0"  "v7.0A" "v7.1"; do
+    for verdir in "sdk v2.0" "sdk v3.5" "v6.0A" "v7.0"  "v7.0A" "v7.1" "VC redist 1033"; do
 	BPATH=$BP
 	fail=false
 	allow_fail=false
diff --git a/erts/include/erl_int_sizes_config.h.in b/erts/include/erl_int_sizes_config.h.in
index b18f5ebc00..88c74cdeff 100644
--- a/erts/include/erl_int_sizes_config.h.in
+++ b/erts/include/erl_int_sizes_config.h.in
@@ -35,6 +35,3 @@
 
 /* The size of a pointer. */
 #undef SIZEOF_VOID_P
-
-/* Define if building a halfword-heap 64bit emulator (needed for NIF's) */
-#undef HALFWORD_HEAP_EMULATOR
diff --git a/erts/include/internal/erl_printf_format.h b/erts/include/internal/erl_printf_format.h
index efd926be99..953022017a 100644
--- a/erts/include/internal/erl_printf_format.h
+++ b/erts/include/internal/erl_printf_format.h
@@ -44,7 +44,6 @@ typedef long long          ErlPfSWord;
 #error Found no appropriate type to use for 'Eterm', 'Uint' and 'Sint'
 #endif
 
-
 typedef int (*fmtfn_t)(void*, char*, size_t);
 
 extern int erts_printf_format(fmtfn_t, void*, char*, va_list);
@@ -57,17 +56,9 @@ extern int erts_printf_uword(fmtfn_t, void*, char, int, int, ErlPfUWord);
 extern int erts_printf_sword(fmtfn_t, void*, char, int, int, ErlPfSWord);
 extern int erts_printf_double(fmtfn_t, void *, char, int, int, double);
 
-#ifdef HALFWORD_HEAP_EMULATOR
-#  if SIZEOF_INT != 4
-#    error Unsupported integer size for HALFWORD_HEAP_EMULATOR
-#  endif
-typedef unsigned int ErlPfEterm;
-#else
 typedef ErlPfUWord ErlPfEterm;
-#endif
-
-extern int (*erts_printf_eterm_func)(fmtfn_t, void*, ErlPfEterm, long, ErlPfEterm*);
 
+extern int (*erts_printf_eterm_func)(fmtfn_t, void*, ErlPfEterm, long);
 
 #endif /* ERL_PRINTF_FORMAT_H__ */
 
diff --git a/erts/include/internal/ethr_mutex.h b/erts/include/internal/ethr_mutex.h
index f76c4262ca..b402a139f5 100644
--- a/erts/include/internal/ethr_mutex.h
+++ b/erts/include/internal/ethr_mutex.h
@@ -98,7 +98,7 @@ void LeaveCriticalSection(CRITICAL_SECTION *);
 #if 0
 #  define ETHR_MTX_Q_LOCK_SPINLOCK__
 #  define ETHR_MTX_QLOCK_TYPE__ ethr_spinlock_t
-#elif defined(ETHR_PTHREADS) || defined(ETHR_OSE_THREADS)
+#elif defined(ETHR_PTHREADS)
 #  define ETHR_MTX_Q_LOCK_PTHREAD_MUTEX__
 #  define ETHR_MTX_QLOCK_TYPE__ pthread_mutex_t
 #elif defined(ETHR_WIN32_THREADS)
@@ -211,7 +211,7 @@ struct ethr_cond_ {
 #endif
 };
 
-#elif (defined(ETHR_PTHREADS) || defined(ETHR_OSE_THREADS)) && !defined(ETHR_DBG_WIN_MTX_WITH_PTHREADS)
+#elif defined(ETHR_PTHREADS) && !defined(ETHR_DBG_WIN_MTX_WITH_PTHREADS)
 
 typedef struct ethr_mutex_ ethr_mutex;
 struct ethr_mutex_ {
@@ -355,7 +355,7 @@ void ethr_rwmutex_rwunlock(ethr_rwmutex *);
 
 #ifdef ETHR_MTX_HARD_DEBUG
 #define ETHR_MTX_HARD_ASSERT(A) \
-  ((void) ((A) ? 1 : ethr_assert_failed(__FILE__, __LINE__, __func__,#A)))
+  ((void) ((A) ? 1 : ethr_assert_failed(__FILE__, __LINE__, __func__, #A)))
 #else
 #define ETHR_MTX_HARD_ASSERT(A) ((void) 1)
 #endif
@@ -634,7 +634,7 @@ ETHR_INLINE_MTX_FUNC_NAME_(ethr_mutex_unlock)(ethr_mutex *mtx)
 
 #endif /* ETHR_TRY_INLINE_FUNCS */
 
-#elif (defined(ETHR_PTHREADS) || defined(ETHR_OSE_THREADS)) && !defined(ETHR_DBG_WIN_MTX_WITH_PTHREADS)
+#elif defined(ETHR_PTHREADS) && !defined(ETHR_DBG_WIN_MTX_WITH_PTHREADS)
 
 #if defined(ETHR_TRY_INLINE_FUNCS) || defined(ETHR_MUTEX_IMPL__)
 
diff --git a/erts/include/internal/ethread.h b/erts/include/internal/ethread.h
index f9c203e97c..4eeb7097f4 100644
--- a/erts/include/internal/ethread.h
+++ b/erts/include/internal/ethread.h
@@ -54,7 +54,8 @@
 #endif
 
 #if defined(ETHR_DEBUG) || !defined(ETHR_INLINE) || ETHR_XCHK \
-    || (defined(__GNUC__) && defined(ERTS_MIXED_CYGWIN_VC))
+    || (defined(__GNUC__) && defined(ERTS_MIXED_CYGWIN_VC)) \
+    || (defined(__GNUC__) && defined(ERTS_MIXED_MSYS_VC))
 #  undef ETHR_INLINE
 #  define ETHR_INLINE 
 #  undef ETHR_FORCE_INLINE
@@ -193,28 +194,6 @@ typedef DWORD ethr_tsd_key;
 
 #define ETHR_YIELD() (Sleep(0), 0)
 
-#elif defined(ETHR_OSE_THREADS)
-
-#include "ose.h"
-#undef NIL
-
-#if defined(ETHR_HAVE_PTHREAD_H)
-#include <pthread.h>
-#endif
-
-typedef struct {
-  PROCESS id;
-  unsigned int tsd_key_index;
-  void *res;
-} ethr_tid;
-
-typedef OSPPDKEY ethr_tsd_key;
-
-#undef ETHR_HAVE_ETHR_SIG_FUNCS
-
-/* Out own RW mutexes are probably faster, but use OSEs mutexes */
-#define ETHR_USE_OWN_RWMTX_IMPL__
-
 #else /* No supported thread lib found */
 
 #ifdef ETHR_NO_SUPP_THR_LIB_NOT_FATAL
@@ -382,19 +361,7 @@ extern ethr_runtime_t ethr_runtime__;
 
 #include "ethr_atomics.h" /* The atomics API */
 
-#if defined (ETHR_OSE_THREADS)
-static ETHR_INLINE void
-ose_yield(void)
-{
-    if (get_ptype(current_process()) == OS_PRI_PROC) {
-        set_pri(get_pri(current_process()));
-    } else {
-        delay(1);
-    }
-}
-#endif
-
-#if defined(__GNUC__) && !defined(ETHR_OSE_THREADS)
+#if defined(__GNUC__)
 #  ifndef ETHR_SPIN_BODY
 #    if defined(__i386__) || defined(__x86_64__)
 #      define ETHR_SPIN_BODY __asm__ __volatile__("rep;nop" : : : "memory")
@@ -410,20 +377,9 @@ ose_yield(void)
 #  ifndef ETHR_SPIN_BODY
 #    define ETHR_SPIN_BODY do {YieldProcessor();ETHR_COMPILER_BARRIER;} while(0)
 #  endif
-#elif defined(ETHR_OSE_THREADS)
-#  ifndef ETHR_SPIN_BODY
-#    define ETHR_SPIN_BODY ose_yield()
-#  else
-#    error "OSE should use ose_yield()"
-#  endif
 #endif
 
-#ifndef ETHR_OSE_THREADS
 #define ETHR_YIELD_AFTER_BUSY_LOOPS 50
-#else
-#define ETHR_YIELD_AFTER_BUSY_LOOPS 0
-#endif
-
 
 #ifndef ETHR_SPIN_BODY
 #  define ETHR_SPIN_BODY ETHR_COMPILER_BARRIER
@@ -446,18 +402,13 @@ ose_yield(void)
 #    else
 #      define ETHR_YIELD() (pthread_yield(), 0)
 #    endif
-#  elif defined(ETHR_OSE_THREADS)
-#    define ETHR_YIELD() (ose_yield(), 0)
 #  else
 #    define ETHR_YIELD() (ethr_compiler_barrier(), 0)
 #  endif
 #endif
 
-#if defined(VALGRIND) || defined(ETHR_OSE_THREADS)
-/* mutex as fallback for spinlock for VALGRIND and OSE.
-   OSE cannot use spinlocks as processes working on the
-   same execution unit have a tendency to deadlock.
- */
+#if defined(VALGRIND)
+/* mutex as fallback for spinlock for VALGRIND. */
 #  undef ETHR_HAVE_NATIVE_SPINLOCKS
 #  undef ETHR_HAVE_NATIVE_RWSPINLOCKS
 #else
@@ -504,16 +455,9 @@ typedef struct {
     int detached;			/* boolean (default false) */
     int suggested_stack_size;		/* kilo words (default sys dependent) */
     char *name;                         /* max 14 char long (default no-name) */
-#ifdef ETHR_OSE_THREADS
-    U32 coreNo;
-#endif
 } ethr_thr_opts;
 
-#if defined(ETHR_OSE_THREADS)
-#define ETHR_THR_OPTS_DEFAULT_INITER {0, -1, NULL, 0}
-#else
 #define ETHR_THR_OPTS_DEFAULT_INITER {0, -1, NULL}
-#endif
 
 #if !defined(ETHR_TRY_INLINE_FUNCS) || defined(ETHR_AUX_IMPL__)
 #  define ETHR_NEED_SPINLOCK_PROTOTYPES__
@@ -627,8 +571,6 @@ typedef struct ethr_ts_event_ ethr_ts_event; /* Needed by ethr_mutex.h */
 #  include "win/ethr_event.h"
 #elif defined(ETHR_PTHREADS)
 #  include "pthread/ethr_event.h"
-#elif defined(ETHR_OSE_THREADS)
-#  include "ose/ethr_event.h"
 #endif
 
 int ethr_set_main_thr_status(int, int);
@@ -718,37 +660,6 @@ ETHR_INLINE_FUNC_NAME_(ethr_leave_ts_event)(ethr_ts_event *tsep)
 
 #endif
 
-#elif  defined (ETHR_OSE_THREADS)
-
-#if defined(ETHR_TRY_INLINE_FUNCS) || defined(ETHREAD_IMPL__)
-
-extern ethr_tsd_key ethr_ts_event_key__;
-
-static ETHR_INLINE ethr_ts_event *
-ETHR_INLINE_FUNC_NAME_(ethr_get_ts_event)(void)
-{
-    ethr_ts_event *tsep = *(ethr_ts_event**)ose_get_ppdata(ethr_ts_event_key__);
-    if (!tsep) {
-	int res = ethr_get_tmp_ts_event__(&tsep);
-	if (res != 0)
-	    ETHR_FATAL_ERROR__(res);
-	ETHR_ASSERT(tsep);
-    }
-    return tsep;
-}
-
-static ETHR_INLINE void
-ETHR_INLINE_FUNC_NAME_(ethr_leave_ts_event)(ethr_ts_event *tsep)
-{
-    if (tsep->iflgs & ETHR_TS_EV_TMP) {
-	int res = ethr_free_ts_event__(tsep);
-	if (res != 0)
-	    ETHR_FATAL_ERROR__(res);
-    }
-}
-
-#endif
-
 #endif
 
 #include "ethr_mutex.h" /* Need atomic declarations and tse */
diff --git a/erts/include/internal/ose/ethr_event.h b/erts/include/internal/ose/ethr_event.h
deleted file mode 100644
index c18f30aa4a..0000000000
--- a/erts/include/internal/ose/ethr_event.h
+++ /dev/null
@@ -1,114 +0,0 @@
-/*
- * %CopyrightBegin%
- *
- * Copyright Ericsson AB 2009-2011. All Rights Reserved.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- *
- * %CopyrightEnd%
- */
-
-/*
- * Author: Rickard Green
- */
-
-//#define USE_PTHREAD_API
-
-#define ETHR_EVENT_OFF_WAITER__		-1L
-#define ETHR_EVENT_OFF__		1L
-#define ETHR_EVENT_ON__ 		0L
-
-#ifdef USE_PTHREAD_API
-
-typedef struct {
-    ethr_atomic32_t state;
-    pthread_mutex_t mtx;
-    pthread_cond_t cnd;
-} ethr_event;
-
-#if defined(ETHR_TRY_INLINE_FUNCS) || defined(ETHR_EVENT_IMPL__)
-
-static void ETHR_INLINE
-ETHR_INLINE_FUNC_NAME_(ethr_event_set)(ethr_event *e)
-{
-    ethr_sint32_t val;
-    val = ethr_atomic32_xchg_mb(&e->state, ETHR_EVENT_ON__);
-    if (val == ETHR_EVENT_OFF_WAITER__) {
-	int res = pthread_mutex_lock(&e->mtx);
-	if (res != 0)
-	    ETHR_FATAL_ERROR__(res);
-	res = pthread_cond_signal(&e->cnd);
-	if (res != 0)
-	    ETHR_FATAL_ERROR__(res);
-	res = pthread_mutex_unlock(&e->mtx);
-	if (res != 0)
-	    ETHR_FATAL_ERROR__(res);
-    }
-}
-
-static void ETHR_INLINE
-ETHR_INLINE_FUNC_NAME_(ethr_event_reset)(ethr_event *e)
-{
-    ethr_atomic32_set(&e->state, ETHR_EVENT_OFF__);
-    ETHR_MEMORY_BARRIER;
-}
-
-#endif
-
-#else
-
-typedef struct {
-    ethr_atomic32_t state;
-    PROCESS proc;
-} ethr_event;
-
-#if defined(ETHR_TRY_INLINE_FUNCS) || defined(ETHR_EVENT_IMPL__)
-
-static void ETHR_INLINE
-ETHR_INLINE_FUNC_NAME_(ethr_event_set)(ethr_event *e)
-{
-    ethr_sint32_t val = ethr_atomic32_xchg_mb(&e->state, ETHR_EVENT_ON__);
-    if (val == ETHR_EVENT_OFF_WAITER__) {
-#ifdef DEBUG
-      OSFSEMVAL fsem_val = get_fsem(e->proc);
-
-      /* There is a race in this assert.
-	 This is because the state is set before the wait call in wait__.
-	 We hope that a delay of 10 ms is enough */
-      if (fsem_val == 0)
-	delay(10);
-      ETHR_ASSERT(get_fsem(e->proc) == -1);
-#endif
-      signal_fsem(e->proc);
-    }
-}
-
-static void ETHR_INLINE
-ETHR_INLINE_FUNC_NAME_(ethr_event_reset)(ethr_event *e)
-{
-    ethr_atomic32_set(&e->state, ETHR_EVENT_OFF__);
-    ETHR_MEMORY_BARRIER;
-}
-
-#endif
-
-#endif
-
-int ethr_event_init(ethr_event *e);
-int ethr_event_destroy(ethr_event *e);
-int ethr_event_wait(ethr_event *e);
-int ethr_event_swait(ethr_event *e, int spincount);
-#if !defined(ETHR_TRY_INLINE_FUNCS) || defined(ETHR_EVENT_IMPL__)
-void ethr_event_set(ethr_event *e);
-void ethr_event_reset(ethr_event *e);
-#endif
diff --git a/erts/include/internal/pthread/ethr_event.h b/erts/include/internal/pthread/ethr_event.h
index 74cfa68e16..deb4b29686 100644
--- a/erts/include/internal/pthread/ethr_event.h
+++ b/erts/include/internal/pthread/ethr_event.h
@@ -83,12 +83,22 @@ ETHR_INLINE_FUNC_NAME_(ethr_event_reset)(ethr_event *e)
 #elif defined(ETHR_PTHREADS)
 /* --- Posix mutex/cond pipe/select implementation of events ---------------- */
 
+#if defined(__APPLE__) && defined(__MACH__) && !defined(__DARWIN__)
+#  define __DARWIN__ 1
+#endif
+
+#ifdef __DARWIN__
+typedef struct ethr_event_fdsets___ ethr_event_fdsets__;
+#endif
 
 typedef struct {
     ethr_atomic32_t state;
     pthread_mutex_t mtx;
     pthread_cond_t cnd;
     int fd[2];
+#ifdef __DARWIN__
+    ethr_event_fdsets__ *fdsets;
+#endif
 } ethr_event;
 
 #define ETHR_EVENT_OFF_WAITER_SELECT__	((ethr_sint32_t) -2)
@@ -148,6 +158,7 @@ ETHR_INLINE_FUNC_NAME_(ethr_event_reset)(ethr_event *e)
 #endif
 
 int ethr_event_init(ethr_event *e);
+int ethr_event_prepare_timed(ethr_event *e);
 int ethr_event_destroy(ethr_event *e);
 int ethr_event_wait(ethr_event *e);
 int ethr_event_swait(ethr_event *e, int spincount);
diff --git a/erts/include/internal/win/ethr_event.h b/erts/include/internal/win/ethr_event.h
index bf110e10f9..458565b9ea 100644
--- a/erts/include/internal/win/ethr_event.h
+++ b/erts/include/internal/win/ethr_event.h
@@ -56,6 +56,7 @@ ETHR_INLINE_FUNC_NAME_(ethr_event_reset)(ethr_event *e)
 #endif
 
 int ethr_event_init(ethr_event *e);
+int ethr_event_prepare_timed(ethr_event *e);
 int ethr_event_destroy(ethr_event *e);
 int ethr_event_wait(ethr_event *e);
 int ethr_event_swait(ethr_event *e, int spincount);
diff --git a/erts/lib_src/common/erl_misc_utils.c b/erts/lib_src/common/erl_misc_utils.c
index 1262c50718..053217304b 100644
--- a/erts/lib_src/common/erl_misc_utils.c
+++ b/erts/lib_src/common/erl_misc_utils.c
@@ -160,8 +160,6 @@ erts_milli_sleep(long ms)
     if (ms > 0) {
 #ifdef __WIN32__
 	Sleep((DWORD) ms);
-#elif defined(__OSE__)
-	delay(ms);
 #else
 	struct timeval tv;
 	tv.tv_sec = ms / 1000;
@@ -320,10 +318,6 @@ erts_cpu_info_update(erts_cpu_info_t *cpuinfo)
 	    online = 0;
 #endif
     }
-#elif defined(__OSE__)
-    online = ose_num_cpus();
-    configured = ose_num_cpus();
-    available = ose_num_cpus();
 #endif
 
     if (online > configured)
diff --git a/erts/lib_src/common/erl_printf_format.c b/erts/lib_src/common/erl_printf_format.c
index 307680505c..e7d5d4413e 100644
--- a/erts/lib_src/common/erl_printf_format.c
+++ b/erts/lib_src/common/erl_printf_format.c
@@ -78,15 +78,7 @@
 #endif
 
 #ifndef ERTS_SIZEOF_ETERM
-#  ifdef HALFWORD_HEAP_EMULATOR
-#    if SIZEOF_VOID_P == 8
-#      define ERTS_SIZEOF_ETERM 4
-#    else
-#      error "HALFWORD_HEAP_EMULATOR only allowed on 64-bit architecture"
-#    endif
-#  else
-#    define ERTS_SIZEOF_ETERM SIZEOF_VOID_P
-#  endif
+#define ERTS_SIZEOF_ETERM SIZEOF_VOID_P
 #endif
 
 #if defined(__GNUC__)
@@ -102,7 +94,7 @@
 #endif
 
 #define FMTC_d    0x0000
-#define FMTC_R    0x0001
+/*empty           0x0001 was RELATIVE */
 #define FMTC_o    0x0002
 #define FMTC_u    0x0003
 #define FMTC_x    0x0004
@@ -166,7 +158,7 @@ static char heX[] = "0123456789ABCDEF";
 #define SIGN(X) ((X) > 0 ? 1 : ((X) < 0 ? -1 : 0)) 
 #define USIGN(X) ((X) == 0 ? 0 : 1)
 
-int (*erts_printf_eterm_func)(fmtfn_t, void*, ErlPfEterm, long, ErlPfEterm*) = NULL;
+int (*erts_printf_eterm_func)(fmtfn_t, void*, ErlPfEterm, long) = NULL;
 
 static int
 noop_fn(void *vfp, char* buf, size_t len)
@@ -645,7 +637,6 @@ int erts_printf_format(fmtfn_t fn, void* arg, char* fmt, va_list ap)
 	    case 'p': ptr++; fmt |= FMTC_p; break;
 	    case 'n': ptr++; fmt |= FMTC_n; break;
 	    case 'T': ptr++; fmt |= FMTC_T; break;
-	    case 'R': ptr++; fmt |= FMTC_R; break;
 	    case '%':
 		FMT(fn,arg,ptr,1,count);
 		ptr++;
@@ -820,11 +811,9 @@ int erts_printf_format(fmtfn_t fn, void* arg, char* fmt, va_list ap)
 		default: *va_arg(ap,int*) = count; break;
 		}
 		break;
-	    case FMTC_T:    /* Eterm */
-	    case FMTC_R: {  /* Eterm, Eterm* base  (base ignored if !HALFWORD_HEAP) */
+	    case FMTC_T: {    /* Eterm */
 		long prec;
 		ErlPfEterm eterm;
-		ErlPfEterm* eterm_base;
 		
 		if (!erts_printf_eterm_func)
 		    return -EINVAL;
@@ -835,16 +824,14 @@ int erts_printf_format(fmtfn_t fn, void* arg, char* fmt, va_list ap)
 		else
 		    prec = (long) precision;
 		eterm = va_arg(ap, ErlPfEterm);
-		eterm_base = ((fmt & FMTC_MASK) == FMTC_R) ?
-		    va_arg(ap, ErlPfEterm*) : NULL;
 		if (width > 0 && !(fmt & FMTF_adj)) {
-		    res = (*erts_printf_eterm_func)(noop_fn, NULL, eterm, prec, eterm_base);
+		    res = (*erts_printf_eterm_func)(noop_fn, NULL, eterm, prec);
 		    if (res < 0)
 			return res;
 		    if (width > res)
 			BLANKS(fn, arg, width - res, count);
 		}
-		res = (*erts_printf_eterm_func)(fn, arg, eterm, prec, eterm_base);
+		res = (*erts_printf_eterm_func)(fn, arg, eterm, prec);
 		if (res < 0)
 		    return res;
 		count += res;
diff --git a/erts/lib_src/common/ethr_aux.c b/erts/lib_src/common/ethr_aux.c
index 0cbb1b2fb8..56fecf81b8 100644
--- a/erts/lib_src/common/ethr_aux.c
+++ b/erts/lib_src/common/ethr_aux.c
@@ -207,18 +207,7 @@ ethr_init_common__(ethr_init_data *id)
 
     ethr_min_stack_size__ = ETHR_B2KW(ethr_min_stack_size__);
 
-#ifdef __OSE__
-    /* For supervisor processes, OSE adds a number of bytes to the requested stack. With this
-     * addition, the resulting size must not exceed the largest available stack size. The number
-     * of bytes that will be added  is configured in the monolith and can therefore not be
-     * specified here. We simply assume that it is less than 0x1000. The available stack sizes
-     * are configured in the .lmconf file and the largest one is usually 65536 bytes.
-     * Consequently, the requested stack size is limited to 0xF000.
-     */
-    ethr_max_stack_size__ = 0xF000;
-#else
     ethr_max_stack_size__ = 32*1024*1024;
-#endif
 #if SIZEOF_VOID_P == 8
     ethr_max_stack_size__ *= 2;
 #endif
@@ -664,10 +653,6 @@ ETHR_IMPL_NORETURN__ ethr_fatal_error__(const char *file,
 int ethr_assert_failed(const char *file, int line, const char *func, char *a)
 {
     fprintf(stderr, "%s:%d: %s(): Assertion failed: %s\n", file, line, func, a);
-#ifdef __OSE__
-    ramlog_printf("%d: %s:%d: %s(): Assertion failed: %s\n",
-		  current_process(),file, line, func, a);
-#endif
     ethr_abort__();
     return 0;
 }
diff --git a/erts/lib_src/common/ethr_mutex.c b/erts/lib_src/common/ethr_mutex.c
index 72aa34ec1c..a596e6c31c 100644
--- a/erts/lib_src/common/ethr_mutex.c
+++ b/erts/lib_src/common/ethr_mutex.c
@@ -1250,7 +1250,7 @@ ethr_cond_wait(ethr_cond *cnd, ethr_mutex *mtx)
     return 0;
 }
 
-#elif (defined(ETHR_PTHREADS) || defined(ETHR_OSE_THREADS)) && !defined(ETHR_DBG_WIN_MTX_WITH_PTHREADS)
+#elif defined(ETHR_PTHREADS) && !defined(ETHR_DBG_WIN_MTX_WITH_PTHREADS)
 /* -- pthread mutex and condition variables -------------------------------- */
 
 int
diff --git a/erts/lib_src/ose/ethr_event.c b/erts/lib_src/ose/ethr_event.c
deleted file mode 100644
index 24ea191c4b..0000000000
--- a/erts/lib_src/ose/ethr_event.c
+++ /dev/null
@@ -1,220 +0,0 @@
-/*
- * %CopyrightBegin%
- *
- * Copyright Ericsson AB 2009-2010. All Rights Reserved.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- *
- * %CopyrightEnd%
- */
-
-/*
- * Author: Rickard Green
- */
-
-#define ETHR_INLINE_FUNC_NAME_(X) X ## __
-#define ETHR_EVENT_IMPL__
-
-#ifdef HAVE_CONFIG_H
-#include "config.h"
-#endif
-
-#include "ethread.h"
-
-#ifdef USE_PTHREAD_API
-
-int
-ethr_event_init(ethr_event *e)
-{
-    int res;
-    ethr_atomic32_init(&e->state, ETHR_EVENT_OFF__);
-    res = pthread_mutex_init(&e->mtx, NULL);
-    if (res != 0)
-	return res;
-    res = pthread_cond_init(&e->cnd, NULL);
-    if (res != 0) {
-	pthread_mutex_destroy(&e->mtx);
-	return res;
-    }
-    return 0;
-}
-
-int
-ethr_event_destroy(ethr_event *e)
-{
-    int res;
-    res = pthread_mutex_destroy(&e->mtx);
-    if (res != 0)
-	return res;
-    res = pthread_cond_destroy(&e->cnd);
-    if (res != 0)
-	return res;
-    return 0;
-}
-
-static ETHR_INLINE int
-wait__(ethr_event *e, int spincount)
-{
-    int sc = spincount;
-    ethr_sint32_t val;
-    int res, ulres;
-    int until_yield = ETHR_YIELD_AFTER_BUSY_LOOPS;
-
-    if (spincount < 0)
-	ETHR_FATAL_ERROR__(EINVAL);
-
-    while (1) {
-	val = ethr_atomic32_read(&e->state);
-	if (val == ETHR_EVENT_ON__)
-	    return 0;
-	if (sc == 0)
-	    break;
-	sc--;
-	ETHR_SPIN_BODY;
-	if (--until_yield == 0) {
-	    until_yield = ETHR_YIELD_AFTER_BUSY_LOOPS;
-	    res = ETHR_YIELD();
-	    if (res != 0)
-		ETHR_FATAL_ERROR__(res);
-	}
-    }
-
-    if (val != ETHR_EVENT_OFF_WAITER__) {
-	val = ethr_atomic32_cmpxchg(&e->state,
-				    ETHR_EVENT_OFF_WAITER__,
-				    ETHR_EVENT_OFF__);
-	if (val == ETHR_EVENT_ON__)
-	    return 0;
-	ETHR_ASSERT(val == ETHR_EVENT_OFF__);
-    }
-
-    ETHR_ASSERT(val == ETHR_EVENT_OFF_WAITER__
-		|| val == ETHR_EVENT_OFF__);
-
-    res = pthread_mutex_lock(&e->mtx);
-    if (res != 0)
-	ETHR_FATAL_ERROR__(res);
-
-    while (1) {
-
-	val = ethr_atomic32_read(&e->state);
-	if (val == ETHR_EVENT_ON__)
-	    break;
-
-	res = pthread_cond_wait(&e->cnd, &e->mtx);
-	if (res == EINTR)
-	    break;
-	if (res != 0)
-	    ETHR_FATAL_ERROR__(res);
-    }
-
-    ulres = pthread_mutex_unlock(&e->mtx);
-    if (ulres != 0)
-	ETHR_FATAL_ERROR__(ulres);
-
-    return res; /* 0 || EINTR */
-}
-
-#else
-/* --- OSE implementation of events ---------------------------- */
-
-#ifdef DEBUG
-union SIGNAL {
-  SIGSELECT signo;
-};
-#endif
-
-int
-ethr_event_init(ethr_event *e)
-{
-    ethr_atomic32_init(&e->state, ETHR_EVENT_OFF__);
-    e->proc = current_process();
-    return 0;
-}
-
-int
-ethr_event_destroy(ethr_event *e)
-{
-  return 0;
-}
-
-static ETHR_INLINE int
-wait__(ethr_event *e, int spincount)
-{
-    int sc = spincount;
-    int res;
-    int until_yield = ETHR_YIELD_AFTER_BUSY_LOOPS;
-
-    if (spincount < 0)
-	ETHR_FATAL_ERROR__(EINVAL);
-
-    ETHR_ASSERT(e->proc == current_process());
-    ETHR_ASSERT(get_fsem(current_process()) == 0);
-
-    while (1) {
-      ethr_sint32_t val;
-      while (1) {
-        val = ethr_atomic32_read(&e->state);
-	if (val == ETHR_EVENT_ON__)
-	  return 0;
-	if (sc == 0)
-	  break;
-	sc--;
-	ETHR_SPIN_BODY;
-	if (--until_yield == 0) {
-	  until_yield = ETHR_YIELD_AFTER_BUSY_LOOPS;
-	  res = ETHR_YIELD();
-	  if (res != 0)
-	    ETHR_FATAL_ERROR__(res);
-	}
-      }
-      if (val != ETHR_EVENT_OFF_WAITER__) {
-	val = ethr_atomic32_cmpxchg(&e->state,
-				    ETHR_EVENT_OFF_WAITER__,
-				    ETHR_EVENT_OFF__);
-	if (val == ETHR_EVENT_ON__)
-	  return 0;
-	ETHR_ASSERT(val == ETHR_EVENT_OFF__);
-      }
-
-      wait_fsem(1);
-
-      ETHR_ASSERT(get_fsem(current_process()) == 0);
-    }
-}
-
-#endif
-
-void
-ethr_event_reset(ethr_event *e)
-{
-    ethr_event_reset__(e);
-}
-
-void
-ethr_event_set(ethr_event *e)
-{
-    ethr_event_set__(e);
-}
-
-int
-ethr_event_wait(ethr_event *e)
-{
-    return wait__(e, 0);
-}
-
-int
-ethr_event_swait(ethr_event *e, int spincount)
-{
-    return wait__(e, spincount);
-}
diff --git a/erts/lib_src/ose/ethread.c b/erts/lib_src/ose/ethread.c
deleted file mode 100644
index dc16acdd08..0000000000
--- a/erts/lib_src/ose/ethread.c
+++ /dev/null
@@ -1,833 +0,0 @@
-/*
- * %CopyrightBegin%
- *
- * Copyright Ericsson AB 2010-2011. All Rights Reserved.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- *
- * %CopyrightEnd%
- */
-
-/*
- * Description: OSE implementation of the ethread library
- * Author: Lukas Larsson
- */
-
-
-#ifdef HAVE_CONFIG_H
-#include "config.h"
-#endif
-
-#include "stdio.h"
-#ifdef ETHR_TIME_WITH_SYS_TIME
-#  include "time.h"
-#  include "sys/time.h"
-#else
-#  ifdef ETHR_HAVE_SYS_TIME_H
-#    include "sys/time.h"
-#  else
-#    include "time.h"
-#  endif
-#endif
-#include "sys/types.h"
-#include "unistd.h"
-
-#include "limits.h"
-
-#define ETHR_INLINE_FUNC_NAME_(X) X ## __
-#define ETHREAD_IMPL__
-
-#include "ethread.h"
-#include "ethr_internal.h"
-
-#include "erl_printf.h"
-#include "efs.h"
-#include "ose.h"
-
-#include "ose_spi.h"
-
-#include "string.h"
-#include "ctype.h"
-#include "stdlib.h"
-
-#ifndef ETHR_HAVE_ETHREAD_DEFINES
-#error Missing configure defines
-#endif
-
-#define ETHR_INVALID_TID_ID -1
-
-#define DEFAULT_PRIO_NAME        "ERTS_ETHR_DEFAULT_PRIO"
-
-/* Set the define to 1 to get some logging */
-#if 0
-#include "ramlog.h"
-#define LOG(output) ramlog_printf output
-#else
-#define LOG(output)
-#endif
-
-static ethr_tid main_thr_tid;
-static const char* own_tid_key = "ethread_own_tid";
-ethr_tsd_key ethr_ts_event_key__;
-
-#define ETHREADWRAPDATASIG 1
-
-/* Init data sent to thr_wrapper() */
-typedef struct {
-    SIGSELECT sig_no;
-    ethr_ts_event *tse;
-    ethr_tid *tid;
-    ethr_sint32_t result;
-    void *(*thr_func)(void *);
-    void *arg;
-    void *prep_func_res;
-    const char *name;
-} ethr_thr_wrap_data__;
-
-union SIGNAL {
-  SIGSELECT sig_no;
-  ethr_thr_wrap_data__ data;
-};
-
-#define ETHR_GET_OWN_TID__	((ethr_tid *) get_envp(current_process(),\
-						       own_tid_key))
-
-/*
- * --------------------------------------------------------------------------
- * Static functions
- * --------------------------------------------------------------------------
- */
-
-/* Will retrive the instrinsic name by removing the 'prefix' and the
- * suffix from 'name'.
- * The 'prefix' is given as an inparameter. If NULL or an empty string no
- * prefix will be removed.
- * If 'strip_suffix' is 1 suffixes in the form of '_123' will be removed.
- * Will return a pointer to a newly allocated buffer containing the intrinsic
- * name in uppercase characters.
- * The caller must remember to free this buffer when no lnger needed.
- */
-static char *
-ethr_intrinsic_name(const char *name, const char *prefix, int strip_suffix)
-{
-   const char *start = name;
-   const char *end  = name + strlen(name);
-   char *intrinsic_name = NULL;
-   int i;
-
-   if (name == NULL) {
-      LOG(("ERTS - ethr_intrinsic_namNo input name.\n"));
-      return NULL;
-   }
-
-   /* take care of the prefix */
-   if ((prefix != NULL) && (*prefix != '\0')) {
-      const char *found = strstr(name, prefix);
-
-      if (found == name) {
-         /* found the prefix at the beginning */
-         start += strlen(prefix);
-      }
-   }
-
-   /* take care of the suffix */
-   if (strip_suffix) {
-      const char *suffix_start = strrchr(start, '_');
-
-      if (suffix_start != NULL) {
-         const char *ch;
-         int only_numbers = 1;
-
-         for (ch = suffix_start + 1; *ch != '\0'; ch++) {
-            if (strchr("0123456789", *ch) == NULL) {
-               only_numbers = 0;
-               break;
-            }
-         }
-
-         if (only_numbers) {
-            end = suffix_start;
-         }
-      }
-   }
-
-   intrinsic_name = malloc(end - start + 1);
-   for (i = 0; (start + i) < end; i++) {
-      intrinsic_name[i] = toupper(start[i]);
-   }
-   intrinsic_name[i] = '\0';
-
-   return intrinsic_name;
-}
-
-static char *
-ethr_get_amended_env(const char *name, const char *prefix, const char *suffix)
-{
-   unsigned len;
-   char *env_name = NULL;
-   char *env_value = NULL;
-
-   if (name == NULL) {
-      return NULL;
-   }
-
-   len = strlen(name);
-
-   if (prefix != NULL) {
-      len += strlen(prefix);
-   }
-
-   if (suffix != NULL) {
-      len += strlen(suffix);
-   }
-
-   env_name = malloc(len + 1);
-   sprintf(env_name, "%s%s%s", (prefix != NULL) ? prefix : "",
-                               name,
-                               (suffix != NULL) ? suffix : "");
-   env_value = get_env(get_bid(current_process()), env_name);
-
-   if (env_value == NULL) {
-      LOG(("ERTS - ethr_get_amended_env(): %s environment variable not present\n", env_name));
-   } else {
-      LOG(("ERTS - ethr_get_amended_env(): Found %s environment variable: %s.\n", env_name, env_value));
-   }
-   free(env_name);
-
-   return env_value;
-}
-
-/* Reads the environment variable derived from 'name' and interprets it as as an
- * OSE priority. If successfull it will update 'out_prio'.
- * Returns:  0 if successfull
- *          -1 orherwise.
- */
-static int
-ethr_get_prio(const char *name, OSPRIORITY *out_prio)
-{
-   int rc = -1;
-   char *intrinsic_name = NULL;
-   char *prio_env = NULL;
-   long prio;
-   char *endptr = NULL;
-
-   LOG(("ERTS - ethr_get_prio(): name: %s.\n", name));
-
-   intrinsic_name = ethr_intrinsic_name(name, NULL, 1);
-   LOG(("ERTS - ethr_get_prio(): Intrinsic name: %s.\n", intrinsic_name));
-
-   prio_env = ethr_get_amended_env(intrinsic_name, "ERTS_", "_PRIO");
-   if (prio_env == NULL) {
-      goto fini;
-   }
-
-   prio = efs_str_to_long(prio_env, (const char **)&endptr);
-   if (endptr != NULL) {
-      LOG(("ERTS - ethr_get_prio(): Environment varible for '%s' includes "
-           "non-numerical characters: '%s'.\n", intrinsic_name, prio_env));
-      goto fini;
-   }
-
-   if ((prio < 0) || (prio > 32)) {
-      LOG(("ERTS - ethr_get_prio(): prio for '%s' (%d) is out of bounds (0-32).\n",
-               intrinsic_name, prio));
-      goto fini;
-   }
-
-   /* Success */
-   *out_prio = (OSPRIORITY)prio;
-   rc = 0;
-
-fini:
-   if (intrinsic_name != NULL) {
-      free(intrinsic_name);
-   }
-   if (prio_env != NULL) {
-      free_buf((union SIGNAL **) &prio_env);
-   }
-
-   return rc;
-}
-
-static PROCESS blockId(void) {
-   static PROCESS bid = (PROCESS)0;
-
-   /* For now we only use the same block. */
-   /*   if (bid == 0) {
-      bid = create_block("Erlang-VM", 0, 0, 0, 0);
-   }
-   return bid; */
-   return 0;
-}
-
-static void thr_exit_cleanup(ethr_tid *tid, void *res)
-{
-
-     ETHR_ASSERT(tid == ETHR_GET_OWN_TID__);
-
-     tid->res = res;
-
-     ethr_run_exit_handlers__();
-     ethr_ts_event_destructor__((void *) ethr_get_tse__());
-}
-
-//static OS_PROCESS(thr_wrapper);
-static OS_PROCESS(thr_wrapper)
-{
-    ethr_tid my_tid;
-    ethr_sint32_t result;
-    void *res;
-    void *(*thr_func)(void *);
-    void *arg;
-    ethr_ts_event *tsep = NULL;
-
-#ifdef DEBUG
-    {
-       PROCESS pid = current_process();
-
-       const char *execMode;
-
-       PROCESS     bid      = get_bid(pid);
-
-       /* In the call below, 16 is a secret number provided by frbr that makes
-        * the function return current domain. */
-       OSADDRESS   domain   = get_pid_info(current_process(), 16);
-
-#ifdef HAVE_OSE_SPI_H
-       execMode = get_pid_info(pid, OSE_PI_SUPERVISOR)
-          ? "Supervisor"
-          : "User";
-#else
-       execMode = "unknown";
-#endif
-
-       fprintf(stderr,"[0x%x] New process. Bid:0x%x, domain:%d, exec mode:%s\n",
-               current_process(), bid, domain, execMode);
-    }
-#endif
-
-    {
-       SIGSELECT sigsel[] = {1,ETHREADWRAPDATASIG};
-       union SIGNAL *init_msg = receive(sigsel);
-
-       thr_func = init_msg->data.thr_func;
-       arg      = init_msg->data.arg;
-
-       result = (ethr_sint32_t) ethr_make_ts_event__(&tsep);
-
-       if (result == 0) {
-          tsep->iflgs |= ETHR_TS_EV_ETHREAD;
-          my_tid = *init_msg->data.tid;
-          set_envp(current_process(), own_tid_key, (OSADDRESS)&my_tid);
-          if (ethr_thr_child_func__)
-             ethr_thr_child_func__(init_msg->data.prep_func_res);
-       }
-
-       init_msg->data.result = result;
-
-       send(&init_msg,sender(&init_msg));
-    }
-
-    /* pthread mutex api says we have to do this */
-    signal_fsem(current_process());
-    ETHR_ASSERT(get_fsem(current_process()) == 0);
-
-    res = result == 0 ? (*thr_func)(arg) : NULL;
-
-    ethr_thr_exit(&res);
-}
-
-/* internal exports */
-
-int ethr_set_tse__(ethr_ts_event *tsep)
-{
-  return ethr_tsd_set(ethr_ts_event_key__,(void *) tsep);
-}
-
-ethr_ts_event *ethr_get_tse__(void)
-{
-  return (ethr_ts_event *) ethr_tsd_get(ethr_ts_event_key__);
-}
-
-#if defined(ETHR_PPC_RUNTIME_CONF__)
-
-static int
-ppc_init__(void)
-{
-    int pid;
-
-
-    ethr_runtime__.conf.have_lwsync = 0;
-
-    return 0;
-}
-
-#endif
-
-#if defined(ETHR_X86_RUNTIME_CONF__)
-
-void
-ethr_x86_cpuid__(int *eax, int *ebx, int *ecx, int *edx)
-{
-#if ETHR_SIZEOF_PTR == 4
-    int have_cpuid;
-    /*
-     * If it is possible to toggle eflags bit 21,
-     * we have the cpuid instruction.
-     */
-    __asm__ ("pushf\n\t"
-             "popl %%eax\n\t"
-             "movl %%eax, %%ecx\n\t"
-             "xorl $0x200000, %%eax\n\t"
-             "pushl %%eax\n\t"
-             "popf\n\t"
-             "pushf\n\t"
-             "popl %%eax\n\t"
-             "movl $0x0, %0\n\t"
-             "xorl %%ecx, %%eax\n\t"
-             "jz no_cpuid\n\t"
-	     "movl $0x1, %0\n\t"
-             "no_cpuid:\n\t"
-             : "=r"(have_cpuid)
-             :
-             : "%eax", "%ecx", "cc");
-    if (!have_cpuid) {
-	*eax = *ebx = *ecx = *edx = 0;
-	return;
-    }
-#endif
-#if ETHR_SIZEOF_PTR == 4 && defined(__PIC__) && __PIC__
-    /*
-     * When position independet code is used in 32-bit mode, the B register
-     * is used for storage of global offset table address, and we may not
-     * use it as input or output in an asm. We need to save and restore the
-     * B register explicitly (for some reason gcc doesn't provide this
-     * service to us).
-     */
-    __asm__ ("pushl %%ebx\n\t"
-	     "cpuid\n\t"
-	     "movl %%ebx, %1\n\t"
-	     "popl %%ebx\n\t"
-	     : "=a"(*eax), "=r"(*ebx), "=c"(*ecx), "=d"(*edx)
-	     : "0"(*eax)
-	     : "cc");
-#else
-    __asm__ ("cpuid\n\t"
-	     : "=a"(*eax), "=b"(*ebx), "=c"(*ecx), "=d"(*edx)
-	     : "0"(*eax)
-	     : "cc");
-#endif
-}
-
-#endif /* ETHR_X86_RUNTIME_CONF__ */
-
-/*
- * --------------------------------------------------------------------------
- * Exported functions
- * --------------------------------------------------------------------------
- */
-
-int
-ethr_init(ethr_init_data *id)
-{
-    int res;
-
-    if (!ethr_not_inited__)
-	return EINVAL;
-
-
-#if defined(ETHR_PPC_RUNTIME_CONF__)
-    res = ppc_init__();
-    if (res != 0)
-	goto error;
-#endif
-
-    res = ethr_init_common__(id);
-    if (res != 0)
-	goto error;
-
-    main_thr_tid.id = current_process();
-    main_thr_tid.tsd_key_index = 0;
-
-    set_envp(current_process(),own_tid_key,(OSADDRESS)&main_thr_tid);
-    signal_fsem(current_process());
-
-
-    ETHR_ASSERT(&main_thr_tid == ETHR_GET_OWN_TID__);
-
-    ethr_not_inited__ = 0;
-
-    ethr_tsd_key_create(&ethr_ts_event_key__,"ethread_tse");
-
-    return 0;
- error:
-    ethr_not_inited__ = 1;
-    return res;
-
-}
-
-int
-ethr_late_init(ethr_late_init_data *id)
-{
-    int res = ethr_late_init_common__(id);
-    if (res != 0)
-	return res;
-    ethr_not_completely_inited__ = 0;
-    return res;
-}
-
-int
-ethr_thr_create(ethr_tid *tid, void * (*func)(void *), void *arg,
-		ethr_thr_opts *opts)
-{
-    int res;
-    int use_stack_size = (opts && opts->suggested_stack_size >= 0
-			  ? opts->suggested_stack_size
-			  : 0x200 /* Use system default */);
-    OSPRIORITY use_prio;
-    char *use_name;
-    char default_thr_name[20];
-    static int no_of_thr = 0;
-    cpuid_t use_core;
-
-    union SIGNAL *init_msg;
-    SIGSELECT sigsel[] = {1,ETHREADWRAPDATASIG};
-    void *prep_func_res;
-
-
-    if (opts != NULL) {
-        LOG(("ERTS - ethr_thr_create(): opts supplied: name: %s, coreNo: %u.\n",
-                      opts->name, opts->coreNo));
-        use_name = opts->name;
-        use_core = opts->coreNo;
-        if (0 != ethr_get_prio(use_name, &use_prio)) {
-           if (0 != ethr_get_prio("DEFAULT", &use_prio)) {
-              use_prio = get_pri(current_process());
-              LOG(("ERTS - ethr_thr_create(): Using current process' prio: %d.\n", use_prio));
-           } else {
-              LOG(("ERTS - ethr_thr_create(): Using default prio: %d.\n", use_prio));
-           }
-        } else {
-           LOG(("ERTS - ethr_thr_create(): Using configured prio: %d.\n", use_prio));
-        }
-    } else {
-        LOG(("ERTS - ethr_thr_create(): opts not supplied. Using defaults.\n"));
-        no_of_thr++;
-        sprintf(default_thr_name, "ethread_%d", no_of_thr);
-        use_name = default_thr_name;
-        use_core = ose_cpu_id();
-
-        if (0 != ethr_get_prio("DEFAULT", &use_prio)) {
-           use_prio = get_pri(current_process());
-           LOG(("ERTS - ethr_thr_create(): Using current process' prio: %d.\n", use_prio));
-        }
-    }
-
-#ifdef ETHR_MODIFIED_DEFAULT_STACK_SIZE
-    if (use_stack_size < 0)
-	use_stack_size = ETHR_MODIFIED_DEFAULT_STACK_SIZE;
-#endif
-
-#if ETHR_XCHK
-    if (ethr_not_completely_inited__) {
-	ETHR_ASSERT(0);
-	return EACCES;
-    }
-    if (!tid || !func) {
-	ETHR_ASSERT(0);
-	return EINVAL;
-    }
-#endif
-
-    if (use_stack_size >= 0) {
-	size_t suggested_stack_size = (size_t) use_stack_size;
-	size_t stack_size;
-#ifdef ETHR_DEBUG
-	suggested_stack_size /= 2; /* Make sure we got margin */
-#endif
-#ifdef ETHR_STACK_GUARD_SIZE
-	/* The guard is at least on some platforms included in the stack size
-	   passed when creating threads */
-	suggested_stack_size += ETHR_B2KW(ETHR_STACK_GUARD_SIZE);
-#endif
-
-	if (suggested_stack_size < ethr_min_stack_size__)
-	    stack_size = ETHR_KW2B(ethr_min_stack_size__);
-	else if (suggested_stack_size > ethr_max_stack_size__)
-	    stack_size = ETHR_KW2B(ethr_max_stack_size__);
-	else
-	    stack_size = ETHR_PAGE_ALIGN(ETHR_KW2B(suggested_stack_size));
-	use_stack_size = stack_size;
-    }
-
-    init_msg = alloc(sizeof(ethr_thr_wrap_data__), ETHREADWRAPDATASIG);
-
-    /* Call prepare func if it exist */
-    if (ethr_thr_prepare_func__)
-	init_msg->data.prep_func_res = ethr_thr_prepare_func__();
-    else
-	init_msg->data.prep_func_res = NULL;
-
-    LOG(("ERTS - ethr_thr_create(): Process [0x%x] is creating '%s', coreNo = %u, prio:%u\n",
-                  current_process(), use_name, use_core, use_prio));
-
-    tid->id = create_process(OS_PRI_PROC, use_name, thr_wrapper,
-			     use_stack_size, use_prio, 0,
-                             get_bid(current_process()), NULL, 0, 0);
-      if (ose_bind_process(tid->id, use_core)) {
-         LOG(("ERTS - ethr_thr_create(): Bound pid 0x%x (%s) to core no %u.\n",
-              tid->id, use_name, use_core));
-      } else {
-            LOG(("ERTS - ethr_thr_create(): Failed binding pid 0x%x (%s) to core no %u.\n",
-                 tid->id, use_name, use_core));
-      }
-
-    /*FIXME!!! Normally this shouldn't be used in shared mode. Still there is
-     * a problem with stdin fd in fd_ processes which should be further
-     * investigated */
-    efs_clone(tid->id);
-
-    tid->tsd_key_index = 0;
-    tid->res = NULL;
-
-    init_msg->data.tse      = ethr_get_ts_event();
-    init_msg->data.thr_func = func;
-    init_msg->data.arg      = arg;
-    init_msg->data.tid      = tid;
-    init_msg->data.name     = opts->name;
-
-    send(&init_msg, tid->id);
-
-    start(tid->id);
-    init_msg = receive(sigsel);
-
-    res = init_msg->data.result;
-    prep_func_res = init_msg->data.prep_func_res;
-
-    free_buf(&init_msg);
-    /* Cleanup... */
-
-    if (ethr_thr_parent_func__)
-	ethr_thr_parent_func__(prep_func_res);
-
-    LOG(("ERTS - ethr_thr_create(): Exiting.\n"));
-    return res;
-}
-
-int
-ethr_thr_join(ethr_tid tid, void **res)
-{
-    SIGSELECT sigsel[] = {1,OS_ATTACH_SIG};
-#if ETHR_XCHK
-    if (ethr_not_inited__) {
-	ETHR_ASSERT(0);
-	return EACCES;
-    }
-#endif
-
-    if (tid.id == ETHR_INVALID_TID_ID)
-      return EINVAL;
-
-    attach(NULL,tid.id);
-    receive(sigsel);
-
-    if (res)
-      *res = tid.res;
-
-    return 0;
-}
-
-int
-ethr_thr_detach(ethr_tid tid)
-{
-#if ETHR_XCHK
-    if (ethr_not_inited__) {
-	ETHR_ASSERT(0);
-	return EACCES;
-    }
-#endif
-    return 0;
-}
-
-void
-ethr_thr_exit(void *res)
-{
-    ethr_tid *tid;
-#if ETHR_XCHK
-    if (ethr_not_inited__) {
-	ETHR_ASSERT(0);
-	return;
-    }
-#endif
-    tid = ETHR_GET_OWN_TID__;
-    if (!tid) {
-	ETHR_ASSERT(0);
-	kill_proc(current_process());
-    }
-    thr_exit_cleanup(tid, res);
-    /* Harakiri possible? */
-    kill_proc(current_process());
-}
-
-ethr_tid
-ethr_self(void)
-{
-    ethr_tid *tid;
-#if ETHR_XCHK
-    if (ethr_not_inited__) {
-      ethr_tid dummy_tid = {ETHR_INVALID_TID_ID, 0, NULL};
-	ETHR_ASSERT(0);
-	return dummy_tid;
-    }
-#endif
-    tid = ETHR_GET_OWN_TID__;
-    if (!tid) {
-	ethr_tid dummy_tid = {ETHR_INVALID_TID_ID, 0, NULL};
-	return dummy_tid;
-    }
-    return *tid;
-}
-
-int
-ethr_equal_tids(ethr_tid tid1, ethr_tid tid2)
-{
-    return tid1.id == tid2.id && tid1.id != ETHR_INVALID_TID_ID;
-}
-
-
-/*
- * Thread specific events
- */
-
-ethr_ts_event *
-ethr_get_ts_event(void)
-{
-    return ethr_get_ts_event__();
-}
-
-void
-ethr_leave_ts_event(ethr_ts_event *tsep)
-{
-    ethr_leave_ts_event__(tsep);
-}
-
-/*
- * Thread specific data
- */
-
-int
-ethr_tsd_key_create(ethr_tsd_key *keyp, char *keyname)
-{
-
-#if ETHR_XCHK
-    if (ethr_not_inited__) {
-	ETHR_ASSERT(0);
-	return EACCES;
-    }
-    if (!keyp) {
-	ETHR_ASSERT(0);
-	return EINVAL;
-    }
-#endif
-
-    ose_create_ppdata(keyname,keyp);
-
-    return 0;
-}
-
-int
-ethr_tsd_key_delete(ethr_tsd_key key)
-{
-#if ETHR_XCHK
-    if (ethr_not_inited__) {
-	ETHR_ASSERT(0);
-	return EACCES;
-    }
-#endif
-    /* Not possible to delete ppdata */
-
-    return 0;
-}
-
-int
-ethr_tsd_set(ethr_tsd_key key, void *value)
-{
-    void **ppdp;
-#if ETHR_XCHK
-    if (ethr_not_inited__) {
-	ETHR_ASSERT(0);
-	return EACCES;
-    }
-#endif
-    ppdp = (void **)ose_get_ppdata(key);
-    *ppdp = value;
-    return 0;
-}
-
-void *
-ethr_tsd_get(ethr_tsd_key key)
-{
-#if ETHR_XCHK
-    if (ethr_not_inited__) {
-	ETHR_ASSERT(0);
-	return NULL;
-    }
-#endif
-    return *(void**)ose_get_ppdata(key);
-}
-
-/*
- * Signal functions
- */
-
-#if ETHR_HAVE_ETHR_SIG_FUNCS
-
-int ethr_sigmask(int how, const sigset_t *set, sigset_t *oset)
-{
-#if ETHR_XCHK
-    if (ethr_not_inited__) {
-	ETHR_ASSERT(0);
-	return EACCES;
-    }
-    if (!set && !oset) {
-	ETHR_ASSERT(0);
-	return EINVAL;
-    }
-#endif
-  return pthread_sigmask(how, set, oset);
-}
-
-int ethr_sigwait(const sigset_t *set, int *sig)
-{
-#if ETHR_XCHK
-    if (ethr_not_inited__) {
-	ETHR_ASSERT(0);
-	return EACCES;
-    }
-    if (!set || !sig) {
-	ETHR_ASSERT(0);
-	return EINVAL;
-    }
-#endif
-    if (sigwait(set, sig) < 0)
-	return errno;
-    return 0;
-}
-
-#endif /* #if ETHR_HAVE_ETHR_SIG_FUNCS */
-
-ETHR_IMPL_NORETURN__
-ethr_abort__(void)
-{
-    abort();
-}
diff --git a/erts/lib_src/pthread/ethr_event.c b/erts/lib_src/pthread/ethr_event.c
index ba664236f6..0629b4dfcd 100644
--- a/erts/lib_src/pthread/ethr_event.c
+++ b/erts/lib_src/pthread/ethr_event.c
@@ -29,6 +29,13 @@
 #include "config.h"
 #endif
 
+#if defined(__APPLE__) && defined(__MACH__) && !defined(__DARWIN__)
+#  define __DARWIN__ 1
+#endif
+#ifdef __DARWIN__
+#  define _DARWIN_UNLIMITED_SELECT
+#endif
+
 #include "ethread.h"
 #undef ETHR_INCLUDE_MONOTONIC_CLOCK__
 #define ETHR_INCLUDE_MONOTONIC_CLOCK__
@@ -56,6 +63,12 @@ ethr_event_init(ethr_event *e)
 }
 
 int
+ethr_event_prepare_timed(ethr_event *e)
+{
+    return 0;
+}
+
+int
 ethr_event_destroy(ethr_event *e)
 {
     return 0;
@@ -171,6 +184,17 @@ return_event_on:
 #include <errno.h>
 #include <string.h>
 
+#ifdef __DARWIN__
+
+struct ethr_event_fdsets___ {
+    fd_set *rsetp;
+    fd_set *esetp;
+    size_t mem_size;
+    fd_mask mem[1];
+};
+
+#endif
+
 static void
 setup_nonblocking_pipe(ethr_event *e)
 {
@@ -188,6 +212,35 @@ setup_nonblocking_pipe(ethr_event *e)
     flgs = fcntl(e->fd[1], F_GETFL, 0);
     fcntl(e->fd[1], F_SETFL, flgs | O_NONBLOCK);
 
+
+#ifndef __DARWIN__
+    if (e->fd[0] >= FD_SETSIZE)
+	ETHR_FATAL_ERROR__(ENOTSUP);
+#else
+    {
+	int nmasks;
+	ethr_event_fdsets__ *fdsets;
+	size_t mem_size;
+
+	nmasks = (e->fd[0]+NFDBITS)/NFDBITS;
+	mem_size = 2*nmasks*sizeof(fd_mask);
+	if (mem_size < 2*sizeof(fd_set)) {
+	    mem_size = 2*sizeof(fd_set);
+	    nmasks = mem_size/(2*sizeof(fd_mask));
+	}
+
+	fdsets = malloc(sizeof(ethr_event_fdsets__)
+			+ mem_size
+			- sizeof(fd_mask));
+	if (!fdsets)
+	    ETHR_FATAL_ERROR__(ENOMEM);
+	fdsets->rsetp = (fd_set *) (char *) &fdsets->mem[0];
+	fdsets->esetp = (fd_set *) (char *) &fdsets->mem[nmasks];
+	fdsets->mem_size = mem_size;
+	e->fdsets = fdsets;
+    }
+#endif
+
     ETHR_MEMBAR(ETHR_StoreStore);
 }
 
@@ -252,6 +305,19 @@ ethr_event_init(ethr_event *e)
 
     e->fd[0] = e->fd[1] = ETHR_EVENT_INVALID_FD__;
 
+#ifdef __DARWIN__
+    e->fdsets = NULL;
+#endif
+
+    return 0;
+}
+
+int
+ethr_event_prepare_timed(ethr_event *e)
+{
+    if (e->fd[0] == ETHR_EVENT_INVALID_FD__)
+	setup_nonblocking_pipe(e);
+
     return 0;
 }
 
@@ -263,13 +329,14 @@ ethr_event_destroy(ethr_event *e)
 	close(e->fd[0]);
 	close(e->fd[1]);
     }
+#ifdef __DARWIN__
+    if (e->fdsets)
+	free(e->fdsets);
+#endif
     res = pthread_mutex_destroy(&e->mtx);
     if (res != 0)
 	return res;
-    res = pthread_cond_destroy(&e->cnd);
-    if (res != 0)
-	return res;
-    return 0;
+    return pthread_cond_destroy(&e->cnd);
 }
 
 static ETHR_INLINE int
@@ -403,8 +470,10 @@ wait__(ethr_event *e, int spincount, ethr_sint64_t timeout)
 	int fd;
 	int sres;
 	ssize_t rres;
-	fd_set rset;
-	fd_set eset;
+#ifndef __DARWIN__
+	fd_set rset, eset;
+#endif
+	fd_set *rsetp, *esetp;
 	struct timeval select_timeout;
 
 #ifdef ETHR_HAVE_ETHR_GET_MONOTONIC_TIME
@@ -457,12 +526,22 @@ wait__(ethr_event *e, int spincount, ethr_sint64_t timeout)
 	    ETHR_ASSERT(act == val);
 	}
 
+
+#ifdef __DARWIN__
+	rsetp = e->fdsets->rsetp;
+	esetp = e->fdsets->esetp;
+	memset((void *) &e->fdsets->mem[0], 0, e->fdsets->mem_size);
+#else
 	FD_ZERO(&rset);
-	FD_SET(fd, &rset);
 	FD_ZERO(&eset);
-	FD_SET(fd, &eset);
+	rsetp = &rset;
+	esetp = &eset;
+#endif
+
+	FD_SET(fd, rsetp);
+	FD_SET(fd, esetp);
 
-	sres = select(fd + 1, &rset, NULL, &eset, &select_timeout);
+	sres = select(fd + 1, rsetp, NULL, esetp, &select_timeout);
 	if (sres == 0)
 	    res = ETIMEDOUT;
 	else {
diff --git a/erts/lib_src/win/ethr_event.c b/erts/lib_src/win/ethr_event.c
index c88c8784a2..6951a216c5 100644
--- a/erts/lib_src/win/ethr_event.c
+++ b/erts/lib_src/win/ethr_event.c
@@ -47,6 +47,12 @@ ethr_event_init(ethr_event *e)
 }
 
 int
+ethr_event_prepare_timed(ethr_event *e)
+{
+    return 0;
+}
+
+int
 ethr_event_destroy(ethr_event *e)
 {
     BOOL res = CloseHandle(e->handle);
diff --git a/erts/preloaded/ebin/erl_prim_loader.beam b/erts/preloaded/ebin/erl_prim_loader.beam
index df12c6f8e0..f6057b45d4 100644
--- a/erts/preloaded/ebin/erl_prim_loader.beam
+++ b/erts/preloaded/ebin/erl_prim_loader.beam
diff --git a/erts/preloaded/ebin/erlang.beam b/erts/preloaded/ebin/erlang.beam
index 863a5e61ef..b6e38e4b5b 100644
--- a/erts/preloaded/ebin/erlang.beam
+++ b/erts/preloaded/ebin/erlang.beam
diff --git a/erts/preloaded/ebin/erts_code_purger.beam b/erts/preloaded/ebin/erts_code_purger.beam
new file mode 100644
index 0000000000..227d96d4c8
--- /dev/null
+++ b/erts/preloaded/ebin/erts_code_purger.beam
diff --git a/erts/preloaded/ebin/erts_internal.beam b/erts/preloaded/ebin/erts_internal.beam
index dc8c711e1a..d3d990519d 100644
--- a/erts/preloaded/ebin/erts_internal.beam
+++ b/erts/preloaded/ebin/erts_internal.beam
diff --git a/erts/preloaded/ebin/init.beam b/erts/preloaded/ebin/init.beam
index 851513b2e9..b6d1df7bbc 100644
--- a/erts/preloaded/ebin/init.beam
+++ b/erts/preloaded/ebin/init.beam
diff --git a/erts/preloaded/ebin/otp_ring0.beam b/erts/preloaded/ebin/otp_ring0.beam
index 33c112f4de..328520844d 100644
--- a/erts/preloaded/ebin/otp_ring0.beam
+++ b/erts/preloaded/ebin/otp_ring0.beam
diff --git a/erts/preloaded/ebin/prim_eval.beam b/erts/preloaded/ebin/prim_eval.beam
index ebca6e7eea..8d6c1927fd 100644
--- a/erts/preloaded/ebin/prim_eval.beam
+++ b/erts/preloaded/ebin/prim_eval.beam
diff --git a/erts/preloaded/ebin/prim_file.beam b/erts/preloaded/ebin/prim_file.beam
index e8817d183e..1221c513db 100644
--- a/erts/preloaded/ebin/prim_file.beam
+++ b/erts/preloaded/ebin/prim_file.beam
diff --git a/erts/preloaded/ebin/prim_inet.beam b/erts/preloaded/ebin/prim_inet.beam
index 5a188be3ba..fa617f1f51 100644
--- a/erts/preloaded/ebin/prim_inet.beam
+++ b/erts/preloaded/ebin/prim_inet.beam
diff --git a/erts/preloaded/ebin/prim_zip.beam b/erts/preloaded/ebin/prim_zip.beam
index 969239be98..83e1e49974 100644
--- a/erts/preloaded/ebin/prim_zip.beam
+++ b/erts/preloaded/ebin/prim_zip.beam
diff --git a/erts/preloaded/ebin/zlib.beam b/erts/preloaded/ebin/zlib.beam
index 281f668f8c..8f654e3abf 100644
--- a/erts/preloaded/ebin/zlib.beam
+++ b/erts/preloaded/ebin/zlib.beam
diff --git a/erts/preloaded/src/Makefile b/erts/preloaded/src/Makefile
index 52034a0881..31383dda83 100644
--- a/erts/preloaded/src/Makefile
+++ b/erts/preloaded/src/Makefile
@@ -41,6 +41,7 @@ PRE_LOADED_ERL_MODULES = \
 	zlib \
 	prim_zip \
 	otp_ring0 \
+	erts_code_purger \
 	erlang \
 	erts_internal
 
diff --git a/erts/preloaded/src/erl_prim_loader.erl b/erts/preloaded/src/erl_prim_loader.erl
index 9f6cba33bd..824ed3435d 100644
--- a/erts/preloaded/src/erl_prim_loader.erl
+++ b/erts/preloaded/src/erl_prim_loader.erl
@@ -42,7 +42,7 @@
 -include("inet_boot.hrl").
 
 %% Public
--export([start/3, set_path/1, get_path/0, get_file/1, get_files/2,
+-export([start/0, set_path/1, get_path/0, get_file/1,
          list_dir/1, read_file_info/1, read_link_info/1, get_cwd/0, get_cwd/1]).
 
 %% Used by erl_boot_server
@@ -50,30 +50,28 @@
          prim_read_file_info/3, prim_get_cwd/2]).
 
 %% Used by escript and code
--export([set_primary_archive/4, release_archives/0]).
+-export([set_primary_archive/4]).
+
+%% Used by test suites
+-export([purge_archive_cache/0]).
 
 -include_lib("kernel/include/file.hrl").
 
 -type host() :: atom().
 
 -record(prim_state, {debug :: boolean(),
-		     cache,
 		     primary_archive}).
 -type prim_state() :: #prim_state{}.
 
 -record(state, 
         {loader            :: 'efile' | 'inet',
          hosts = []        :: [host()], % hosts list (to boot from)
-         id,                      % not used any more?
          data              :: 'noport' | port(), % data port etc
-         timeout           :: timeout(),         % idle timeout
-	 %% Number of timeouts before archives are released
-	 n_timeouts        :: non_neg_integer(),
-         multi_get = false :: boolean(),
+         timeout           :: timeout(),	 % idle timeout
          prim_state        :: prim_state()}).    % state for efile code loader
 
--define(IDLE_TIMEOUT, 60000).  %% tear inet connection after 1 minutes
--define(N_TIMEOUTS, 6).        %% release efile archive after 6 minutes
+-define(EFILE_IDLE_TIMEOUT, (6*60*1000)).	%purge archives
+-define(INET_IDLE_TIMEOUT, (60*1000)). 		%tear down connection timeout
 
 %% Defines for inet as prim_loader
 -define(INET_FAMILY, inet).
@@ -103,26 +101,13 @@ debug(#prim_state{debug = Deb}, Term) ->
 %%% Interface Functions. 
 %%% --------------------------------------------------------
 
--spec start(Id, Loader, Hosts) ->
+-spec start() ->
 	    {'ok', Pid} | {'error', What} when
-      Id :: term(),
-      Loader :: atom() | string(),
-      Hosts :: Host | [Host],
-      Host :: host(),
       Pid :: pid(),
       What :: term().
-start(Id, Pgm, Hosts) when is_atom(Hosts) ->
-    start(Id, Pgm, [Hosts]);
-start(Id, Pgm0, Hosts) ->
-    Pgm = if
-              is_atom(Pgm0) ->
-                  atom_to_list(Pgm0);
-              true ->
-                  Pgm0
-          end,
+start() ->
     Self = self(),
-    Pid = spawn_link(fun() -> start_it(Pgm, Id, Self, Hosts) end),
-    register(erl_prim_loader, Pid),
+    Pid = spawn_link(fun() -> start_it(Self) end),
     receive
         {Pid,ok} ->
             {ok,Pid};
@@ -130,26 +115,39 @@ start(Id, Pgm0, Hosts) ->
             {error,Reason}
     end.
 
-%% Hosts must be a list of form ['1.2.3.4' ...]
-start_it("inet", Id, Pid, Hosts) ->
+start_it(Parent) ->
     process_flag(trap_exit, true),
-    ?dbg(inet, {Id,Pid,Hosts}),
+    register(erl_prim_loader, self()),
+    Loader = case init:get_argument(loader) of
+		 {ok,[[Loader0]]} ->
+		     Loader0;
+		 error ->
+		     "efile"
+	     end,
+    case Loader of
+	"efile" -> start_efile(Parent);
+	"inet" -> start_inet(Parent)
+    end.
+
+%% Hosts must be a list of form ['1.2.3.4' ...]
+start_inet(Parent) ->
+    Hosts = case init:get_argument(hosts) of
+		{ok,[Hosts0]} -> Hosts0;
+		_ -> []
+	    end,
     AL = ipv4_list(Hosts),
     ?dbg(addresses, AL),
     {ok,Tcp} = find_master(AL),
-    init_ack(Pid),
+    init_ack(Parent),
     PS = prim_init(),
     State = #state {loader = inet,
                     hosts = AL,
-                    id = Id,
                     data = Tcp,
-                    timeout = ?IDLE_TIMEOUT,
-		    n_timeouts = ?N_TIMEOUTS,
+                    timeout = ?INET_IDLE_TIMEOUT,
                     prim_state = PS},
-    loop(State, Pid, []);
+    loop(State, Parent, []).
 
-start_it("efile", Id, Pid, _Hosts) ->
-    process_flag(trap_exit, true),
+start_efile(Parent) ->
     {ok, Port} = prim_file:start(),
     %% Check that we started in a valid directory.
     case prim_file:get_cwd(Port) of
@@ -160,20 +158,14 @@ start_it("efile", Id, Pid, _Hosts) ->
 	    erlang:display(Report),
 	    exit({error, invalid_current_directory});
 	_ ->
-	    init_ack(Pid)
+	    init_ack(Parent)
     end,
-    MultiGet = case erlang:system_info(thread_pool_size) of
-                   0 -> false;
-                   _ -> true
-               end,
     PS = prim_init(),
     State = #state {loader = efile,
-                    id = Id,
                     data = Port,
-                    timeout = infinity,
-                    multi_get = MultiGet,
+                    timeout = ?EFILE_IDLE_TIMEOUT,
                     prim_state = PS},
-    loop(State, Pid, []).
+    loop(State, Parent, []).
 
 init_ack(Pid) ->
     Pid ! {self(),ok},
@@ -198,20 +190,6 @@ get_file(File) when is_atom(File) ->
 get_file(File) ->
     check_file_result(get_file, File, request({get_file,File})).
 
--spec get_files([{atom(), string()}],
-		fun((atom(),binary(),string()) -> 'ok' | {'error', atom()})) ->
-			'ok' | {'error', atom()}.
-get_files(ModFiles, Fun) ->
-    case request({get_files,{ModFiles,Fun}}) of
-        E = {error,_M} ->
-            E;
-        {error,Reason,M} ->
-            check_file_result(get_files, M, {error,Reason}),
-            {error,M};
-        ok ->
-            ok
-    end.
-
 -spec list_dir(Dir) -> {'ok', Filenames} | 'error' when
       Dir :: string(),
       Filenames :: [Filename :: string()].
@@ -250,10 +228,13 @@ set_primary_archive(File, ArchiveBin, FileInfo, ParserFun)
   when is_list(File), is_binary(ArchiveBin), is_record(FileInfo, file_info) ->
     request({set_primary_archive, File, ArchiveBin, FileInfo, ParserFun}).
 
--spec release_archives() -> 'ok' | {'error', _}.
+%% NOTE: Does not close the primary archive. Only closes all
+%% open zip files kept in the cache. Should be called before an archive
+%% file is to be removed (for example in the test suites).
 
-release_archives() ->
-    request(release_archives).
+-spec purge_archive_cache() -> 'ok' | {'error', _}.
+purge_archive_cache() ->
+    request(purge_archive_cache).
 
 request(Req) ->
     Loader = whereis(erl_prim_loader),
@@ -310,76 +291,59 @@ check_file_result(_, _, Other) ->
 %%% The main loop.
 %%% --------------------------------------------------------
 
-loop(State, Parent, Paths) ->
+loop(St0, Parent, Paths) ->
     receive
+	{Pid,{set_path,NewPaths}} when is_pid(Pid) ->
+	    Pid ! {self(),ok},
+	    loop(St0, Parent, to_strs(NewPaths));
         {Pid,Req} when is_pid(Pid) ->
-            %% erlang:display(Req),
-            {Resp,State2,Paths2} =
-                case Req of
-                    {set_path,NewPaths} ->
-                        {ok,State,to_strs(NewPaths)};
-                    {get_path,_} ->
-                        {{ok,Paths},State,Paths};
-                    {get_file,File} ->
-                        {Res,State1} = handle_get_file(State, Paths, File),
-                        {Res,State1,Paths};
-                    {get_files,{ModFiles,Fun}} ->
-                        {Res,State1} = handle_get_files(State, ModFiles, Paths, Fun),
-                        {Res,State1,Paths};
-                    {list_dir,Dir} ->
-                        {Res,State1} = handle_list_dir(State, Dir),
-                        {Res,State1,Paths};
-                    {read_file_info,File} ->
-                        {Res,State1} = handle_read_file_info(State, File),
-                        {Res,State1,Paths};
-                    {read_link_info,File} ->
-                        {Res,State1} = handle_read_link_info(State, File),
-                        {Res,State1,Paths};
-                    {get_cwd,[]} ->
-                        {Res,State1} = handle_get_cwd(State, []),
-                        {Res,State1,Paths};
-                    {get_cwd,[_]=Args} ->
-                        {Res,State1} = handle_get_cwd(State, Args),
-                        {Res,State1,Paths};
-                    {set_primary_archive,File,ArchiveBin,FileInfo,ParserFun} ->
-                        {Res,State1} =
-			    handle_set_primary_archive(State, File,
-						       ArchiveBin, FileInfo,
-						       ParserFun),
-                        {Res,State1,Paths};
-                    release_archives ->
-                        {Res,State1} = handle_release_archives(State),
-                        {Res,State1,Paths};
-                    _Other ->
-                        {ignore,State,Paths}
-                end,
-            if Resp =:= ignore -> ok;
-               true -> Pid ! {self(),Resp}, ok
-            end,
-            if 
-                is_record(State2, state) ->
-                    loop(State2, Parent, Paths2);
-                true ->
-                    exit({bad_state, Req, State2})          
+	    case handle_request(Req, Paths, St0) of
+		ignore ->
+		    ok;
+		{Resp,#state{}=St1} ->
+		    Pid ! {self(),Resp},
+                    loop(St1, Parent, Paths);
+		{_,State2,_} ->
+                    exit({bad_state,Req,State2})
             end;
         {'EXIT',Parent,W} ->
-            _State1 = handle_stop(State),
+            _ = handle_stop(St0),
             exit(W);
         {'EXIT',P,W} ->
-            State1 = handle_exit(State, P, W),
-            loop(State1, Parent, Paths);
+            St1 = handle_exit(St0, P, W),
+            loop(St1, Parent, Paths);
         _Message ->
-            loop(State, Parent, Paths)
-    after State#state.timeout ->
-            State1 = handle_timeout(State, Parent),
-            loop(State1, Parent, Paths)
+            loop(St0, Parent, Paths)
+    after St0#state.timeout ->
+            St1 = handle_timeout(St0, Parent),
+            loop(St1, Parent, Paths)
+    end.
+
+handle_request(Req, Paths, St0) ->
+    case Req of
+	{get_path,_} ->
+	    {{ok,Paths},St0};
+	{get_file,File} ->
+	    handle_get_file(St0, Paths, File);
+	{list_dir,Dir} ->
+	    handle_list_dir(St0, Dir);
+	{read_file_info,File} ->
+	    handle_read_file_info(St0, File);
+	{read_link_info,File} ->
+	    handle_read_link_info(St0, File);
+	{get_cwd,[]} ->
+	    handle_get_cwd(St0, []);
+	{get_cwd,[_]=Args} ->
+	    handle_get_cwd(St0, Args);
+	{set_primary_archive,File,ArchiveBin,FileInfo,ParserFun} ->
+	    handle_set_primary_archive(St0, File, ArchiveBin,
+				       FileInfo, ParserFun);
+	purge_archive_cache ->
+	    handle_purge_archive_cache(St0);
+	_ ->
+	    ignore
     end.
 
-handle_get_files(State = #state{multi_get = true}, ModFiles, Paths, Fun) ->
-    ?SAFE2(efile_multi_get_file_from_port(State, ModFiles, Paths, Fun), State);
-handle_get_files(State, _ModFiles, _Paths, _Fun) ->     % no multi get
-    {{error,no_multi_get},State}.
-    
 handle_get_file(State = #state{loader = efile}, Paths, File) ->
     ?SAFE2(efile_get_file_from_port(State, File, Paths), State);
 handle_get_file(State = #state{loader = inet}, Paths, File) ->
@@ -388,8 +352,9 @@ handle_get_file(State = #state{loader = inet}, Paths, File) ->
 handle_set_primary_archive(State= #state{loader = efile}, File, ArchiveBin, FileInfo, ParserFun) ->
     ?SAFE2(efile_set_primary_archive(State, File, ArchiveBin, FileInfo, ParserFun), State).
 
-handle_release_archives(State= #state{loader = efile}) ->
-    ?SAFE2(efile_release_archives(State), State).
+handle_purge_archive_cache(#state{loader = efile}=State) ->
+    prim_purge_cache(),
+    {ok,State}.
 
 handle_list_dir(State = #state{loader = efile}, Dir) ->
     ?SAFE2(efile_list_dir(State, Dir), State);
@@ -430,53 +395,6 @@ handle_timeout(State = #state{loader = inet}, Parent) ->
 %%% Functions which handle efile as prim_loader (default).
 %%% --------------------------------------------------------
 
-%%% Reading many files in parallel is an optimization. 
-%%% See also comment in init.erl.
-
-%% -> {ok,State} | {{error,Module},State} | {{error,Reason,Module},State}
-efile_multi_get_file_from_port(State, ModFiles, Paths, Fun) ->
-    Ref = make_ref(),
-    %% More than 200 processes is no gain.
-    Max = erlang:min(200, erlang:system_info(thread_pool_size)),
-    efile_multi_get_file_from_port2(ModFiles, 0, Max, State, Paths, Fun, Ref, ok).
-
-efile_multi_get_file_from_port2([MF | MFs], Out, Max, State, Paths, Fun, Ref, Ret) when Out < Max ->
-    Self = self(),
-    _Pid = spawn(fun() -> efile_par_get_file(Ref, State, MF, Paths, Self, Fun) end),
-    efile_multi_get_file_from_port2(MFs, Out+1, Max, State, Paths, Fun, Ref, Ret);
-efile_multi_get_file_from_port2(MFs, Out, Max, _State, Paths, Fun, Ref, Ret) when Out > 0 ->
-    receive 
-        {Ref, ok, State1} ->
-            efile_multi_get_file_from_port2(MFs, Out-1, Max, State1, Paths, Fun, Ref, Ret);
-        {Ref, {error,_Mod} = Error, State1} ->
-            efile_multi_get_file_from_port2(MFs, Out-1, Max, State1, Paths, Fun, Ref, Error);
-        {Ref, MF, {error,emfile,State1}} ->
-            %% Max can take negative values. Out cannot.
-            efile_multi_get_file_from_port2([MF | MFs], Out-1, Max-1, State1, Paths, Fun, Ref, Ret);
-        {Ref, {M,_F}, {error,Error,State1}} -> 
-            efile_multi_get_file_from_port2(MFs, Out-1, 0, State1, Paths, Fun, Ref, {error,Error,M})
-    end;
-efile_multi_get_file_from_port2(_MFs, 0, _Max, State, _Paths, _Fun, _Ref, Ret) ->
-    {Ret,State}.
-
-efile_par_get_file(Ref, State, {Mod,File} = MF, Paths, Pid, Fun) ->
-    %% One port for each file read in "parallel":
-    case prim_file:start() of
-        {ok, Port} ->
-            Port0 = State#state.data,
-            State1 = State#state{data = Port},
-            R = case efile_get_file_from_port(State1, File, Paths) of
-                    {{error,Reason},State2} -> 
-                        {Ref,MF,{error,Reason,State2}};
-                    {{ok,BinFile,Full},State2} -> 
-                        %% Fun(...) -> ok | {error,Mod}
-                        {Ref,Fun(Mod, BinFile, Full),State2#state{data=Port0}}
-                end,
-            prim_file:close(Port),
-            Pid ! R;
-        {error, Error} ->
-            Pid ! {Ref,MF,{error,Error,State}}
-    end.
 
 %% -> {{ok,BinFile,File},State} | {{error,Reason},State}
 efile_get_file_from_port(State, File, Paths) ->
@@ -521,10 +439,6 @@ efile_set_primary_archive(#state{prim_state = PS} = State, File,
 					  FileInfo, ParserFun),
     {Res,State#state{prim_state = PS2}}.
 
-efile_release_archives(#state{prim_state = PS} = State) ->
-    {Res, PS2} = prim_release_archives(PS),
-    {Res,State#state{prim_state = PS2}}.
-
 efile_list_dir(#state{prim_state = PS} = State, Dir) ->
     {Res, PS2} = prim_list_dir(PS, Dir),
     {Res, State#state{prim_state = PS2}}.
@@ -546,14 +460,9 @@ efile_exit_port(State, Port, Reason) when State#state.data =:= Port ->
 efile_exit_port(State, _Port, _Reason) ->
     State.
 
-efile_timeout_handler(#state{n_timeouts = N} = State, _Parent) ->
-    if
-	N =< 0 ->
-	    {_Res, State2} = efile_release_archives(State),
-	    State2#state{n_timeouts = ?N_TIMEOUTS};
-	true ->
-	    State#state{n_timeouts = N - 1}
-    end.
+efile_timeout_handler(State, _Parent) ->
+    prim_purge_cache(),
+    State.
 
 %%% --------------------------------------------------------
 %%% Functions which handle inet prim_loader
@@ -694,7 +603,7 @@ inet_get_file_from_port1(_File, [], State) ->
 inet_send_and_rcv(Msg, Tag, State) when State#state.data =:= noport ->
     {ok,Tcp} = find_master(State#state.hosts),     %% reconnect
     inet_send_and_rcv(Msg, Tag, State#state{data = Tcp,
-					    timeout = ?IDLE_TIMEOUT});
+					    timeout = ?INET_IDLE_TIMEOUT});
 inet_send_and_rcv(Msg, Tag, #state{data = Tcp, timeout = Timeout} = State) ->
     prim_inet:send(Tcp, term_to_binary(Msg)),
     receive
@@ -819,32 +728,19 @@ prim_init() ->
         end,
     cache_new(#prim_state{debug = Deb}).
 
-prim_release_archives(PS) ->
-    debug(PS, release_archives),
-    {Res, PS2} = prim_do_release_archives(PS, get(), []),
-    debug(PS2, {return, Res}),
-    {Res, PS2}.
-
-prim_do_release_archives(PS, [{ArchiveFile, DictVal} | KeyVals], Acc) ->
-    Res = 
-	case DictVal of
-	    {primary, _PrimZip, _FI, _ParserFun} ->
-		ok; % Keep primary archive
-	    {Cache, _FI} ->
-		debug(PS, {release, cache, ArchiveFile}),
-		erase(ArchiveFile),
-		clear_cache(ArchiveFile, Cache)
-	end,
-    case Res of
-	ok ->
-	    prim_do_release_archives(PS, KeyVals, Acc);
-	{error, Reason} ->
-	    prim_do_release_archives(PS, KeyVals, [{ArchiveFile, Reason} | Acc])
-    end;
-prim_do_release_archives(PS, [], []) ->
-    {ok, PS#prim_state{primary_archive = undefined}};
-prim_do_release_archives(PS, [], Errors) ->
-    {{error, Errors}, PS#prim_state{primary_archive = undefined}}.
+prim_purge_cache() ->
+    do_prim_purge_cache(get()).
+
+do_prim_purge_cache([{Key,Val}|T]) ->
+    case Val of
+	{Cache,_FI} ->
+	    catch clear_cache(Key, Cache);
+	_ ->
+	    ok
+    end,
+    do_prim_purge_cache(T);
+do_prim_purge_cache([]) ->
+    ok.
 
 prim_set_primary_archive(PS, undefined, undefined, undefined, _ParserFun) ->
     debug(PS, {set_primary_archive, clean}),
@@ -1287,70 +1183,62 @@ path_join([Path],Acc) ->
 path_join([Path|Paths],Acc) ->
     path_join(Paths,"/" ++ reverse(Path) ++ Acc).
 
-name_split(ArchiveFile, File0) ->
-    File = absname(File0),
-    do_name_split(ArchiveFile, File).
-    
-do_name_split(undefined, File) ->
+name_split(undefined, File) ->
     %% Ignore primary archive
-    case string_split(File, init:archive_extension(), []) of
+    RevExt = reverse(init:archive_extension()),
+    case archive_split(File, RevExt, []) of
         no_split ->
-            %% Plain file
             {file, File};
-        {split, _RevArchiveBase, RevArchiveFile, []} ->
-            %% Top dir in archive
-            ArchiveFile = reverse(RevArchiveFile),
-            {archive, ArchiveFile, []};
-        {split, _RevArchiveBase, RevArchiveFile, [$/ | FileInArchive]} ->
-            %% File in archive
-            ArchiveFile = reverse(RevArchiveFile),
-            {archive, ArchiveFile, FileInArchive};
-        {split, _RevArchiveBase, _RevArchiveFile, _FileInArchive} ->
-	    %% False match. Assume plain file
-            {file, File}
+	Archive ->
+	    Archive
     end;
-do_name_split(ArchiveFile, File) ->
+name_split(ArchiveFile, File0) ->
     %% Look first in primary archive
-    case string_match(real_path(File), ArchiveFile, []) of
+    File = absname(File0),
+    case string_match(real_path(File), ArchiveFile) of
         no_match ->
             %% Archive or plain file
-            do_name_split(undefined, File);
-        {match, _RevPrimArchiveFile, FileInArchive} ->
+            name_split(undefined, File);
+        {match, FileInArchive} ->
             %% Primary archive
 	    {archive, ArchiveFile, FileInArchive}
     end.
 
-string_match([Char | File], [Char | Archive], RevTop) ->
-    string_match(File, Archive, [Char | RevTop]);
-string_match([] = File, [], RevTop) ->
-    {match, RevTop, File};
-string_match([$/ | File], [], RevTop) ->
-    {match, RevTop, File};
-string_match(_File, _Archive, _RevTop) ->
+string_match([Char | File], [Char | Archive]) ->
+    string_match(File, Archive);
+string_match([] = File, []) ->
+    {match, File};
+string_match([$/ | File], []) ->
+    {match, File};
+string_match(_File, _Archive) ->
     no_match.
 
-string_split([Char | File], [Char | Ext] = FullExt, RevTop) ->
-    RevTop2 = [Char | RevTop],
-    string_split2(File, Ext, RevTop, RevTop2, File, FullExt, RevTop2);
-string_split([Char | File], Ext, RevTop) ->
-    string_split(File, Ext, [Char | RevTop]);
-string_split([], _Ext, _RevTop) ->
-    no_split.
-
-string_split2([Char | File], [Char | Ext], RevBase, RevTop, SaveFile, SaveExt, SaveTop) ->
-    string_split2(File, Ext, RevBase, [Char | RevTop], SaveFile, SaveExt, SaveTop);
-string_split2(File, [], RevBase, RevTop, _SaveFile, _SaveExt, _SaveTop) ->
-    {split, RevBase, RevTop, File};
-string_split2(_, _Ext, _RevBase, _RevTop, SaveFile, SaveExt, SaveTop) ->
-    string_split(SaveFile, SaveExt, SaveTop).
+archive_split("/"++File, RevExt, Acc) ->
+    case is_prefix(RevExt, Acc) of
+	false ->
+	    archive_split(File, RevExt, [$/|Acc]);
+	true ->
+	    ArchiveFile = absname(reverse(Acc)),
+	    {archive, ArchiveFile, File}
+    end;
+archive_split([H|T], RevExt, Acc) ->
+    archive_split(T, RevExt, [H|Acc]);
+archive_split([], RevExt, Acc) ->
+    case is_prefix(RevExt, Acc) of
+	false ->
+	    no_split;
+	true ->
+	    ArchiveFile = absname(reverse(Acc)),
+	    {archive, ArchiveFile, []}
+    end.
+
+is_prefix([H|T1], [H|T2]) -> is_prefix(T1, T2);
+is_prefix([_|_], _) -> false;
+is_prefix([], _ ) -> true.
 
 %% Parse list of ipv4 addresses 
 ipv4_list([H | T]) ->
-    IPV = if is_atom(H) -> ipv4_address(atom_to_list(H));
-             is_list(H) -> ipv4_address(H);
-             true -> {error,einal}
-          end,
-    case IPV of
+    case ipv4_address(H) of
         {ok,IP} -> [IP | ipv4_list(T)];
         _ -> ipv4_list(T)
     end;
@@ -1415,8 +1303,6 @@ absname_vr([Drive, $\: | NameRest], _) ->
 %% Assumes normalized name
 pathtype(Name) when is_list(Name) -> 
     case erlang:system_info(os_type) of
-	{ose, _}  ->
-	    unix_pathtype(Name);
 	{unix, _}  -> 
 	    unix_pathtype(Name);
 	{win32, _} ->
diff --git a/erts/preloaded/src/erlang.erl b/erts/preloaded/src/erlang.erl
index 291356c7b1..40d5aedd24 100644
--- a/erts/preloaded/src/erlang.erl
+++ b/erts/preloaded/src/erlang.erl
@@ -460,7 +460,7 @@ check_old_code(_Module) ->
       CheckResult :: boolean().
 check_process_code(Pid, Module) ->
     try
-	erlang:check_process_code(Pid, Module, [{allow_gc, true}])
+	erts_internal:check_process_code(Pid, Module, [{allow_gc, true}])
     catch
 	error:Error -> erlang:error(Error, [Pid, Module])
     end.
@@ -475,51 +475,11 @@ check_process_code(Pid, Module) ->
       CheckResult :: boolean() | aborted.
 check_process_code(Pid, Module, OptionList)  ->
     try
-	{Async, AllowGC} = get_cpc_opts(OptionList, sync, true),
-	case Async of
-	    {async, ReqId} ->
-		{priority, Prio} = erlang:process_info(erlang:self(),
-						       priority),
-		erts_internal:request_system_task(Pid,
-						  Prio,
-						  {check_process_code,
-						   ReqId,
-						   Module,
-						   AllowGC}),
-		async;
-	    sync ->
-		case Pid == erlang:self() of
-		    true ->
-			erts_internal:check_process_code(Module,
-							 [{allow_gc, AllowGC}]);
-		    false ->
-			{priority, Prio} = erlang:process_info(erlang:self(),
-							       priority),
-			ReqId = erlang:make_ref(),
-			erts_internal:request_system_task(Pid,
-							  Prio,
-							  {check_process_code,
-							   ReqId,
-							   Module,
-							   AllowGC}),
-			receive
-			    {check_process_code, ReqId, CheckResult} ->
-				CheckResult
-			end
-		end
-	end
+	erts_internal:check_process_code(Pid, Module, OptionList)
     catch
 	error:Error -> erlang:error(Error, [Pid, Module, OptionList])
     end.
 
-% gets async and allow_gc opts and verify valid option list
-get_cpc_opts([{async, _ReqId} = AsyncTuple | Options], _OldAsync, AllowGC) ->
-    get_cpc_opts(Options, AsyncTuple, AllowGC);
-get_cpc_opts([{allow_gc, AllowGC} | Options], Async, _OldAllowGC) ->
-    get_cpc_opts(Options, Async, AllowGC);
-get_cpc_opts([], Async, AllowGC) ->
-    {Async, AllowGC}.
-
 %% crc32/1
 -spec erlang:crc32(Data) -> non_neg_integer() when
       Data :: iodata().
@@ -1464,8 +1424,16 @@ processes() ->
 %% purge_module/1
 -spec purge_module(Module) -> true when
       Module :: atom().
-purge_module(_Module) ->
-    erlang:nif_error(undefined).
+purge_module(Module) when erlang:is_atom(Module) ->
+    case erts_code_purger:purge(Module) of
+	{false, _} ->
+	    erlang:error(badarg, [Module]);
+	{true, _} ->
+	    true
+    end;
+purge_module(Arg) ->
+    erlang:error(badarg, [Arg]).
+
 
 %% put/2
 -spec put(Key, Val) -> term() when
@@ -1974,6 +1942,7 @@ localtime_to_universaltime(_Localtime, _IsDst) ->
 %% CHECK! Why the strange very thorough specification of the error
 %% condition with disallowed arity in erl_bif_types?
 %% Not documented
+%% Shadowed by erl_bif_types: erlang:make_fun/3
 -spec erlang:make_fun(Module, Function, Arity) -> function() when
       Module :: atom(),
       Function :: atom(),
@@ -2026,12 +1995,21 @@ nodes(_Arg) ->
            | eof
 	   | {parallelism, Boolean :: boolean()}
 	   | hide.
-open_port(_PortName,_PortSettings) ->
-    erlang:nif_error(undefined).
+open_port(PortName, PortSettings) ->
+    case case erts_internal:open_port(PortName, PortSettings) of
+	     Ref when erlang:is_reference(Ref) -> receive {Ref, Res} -> Res end;
+	     Res -> Res
+	 end of
+	Port when erlang:is_port(Port) -> Port;
+	Error -> erlang:error(Error, [PortName, PortSettings])
+    end.
 
 -type priority_level() ::
       low | normal | high | max.
 
+-type message_queue_data() ::
+	off_heap | on_heap | mixed.
+
 -spec process_flag(trap_exit, Boolean) -> OldBoolean when
       Boolean :: boolean(),
       OldBoolean :: boolean();
@@ -2044,6 +2022,9 @@ open_port(_PortName,_PortSettings) ->
                   (min_bin_vheap_size, MinBinVHeapSize) -> OldMinBinVHeapSize when
       MinBinVHeapSize :: non_neg_integer(),
       OldMinBinVHeapSize :: non_neg_integer();
+                  (message_queue_data, MQD) -> OldMQD when
+      MQD :: message_queue_data(),
+      OldMQD :: message_queue_data();
                   (priority, Level) -> OldLevel when
       Level :: priority_level(),
       OldLevel :: priority_level();
@@ -2082,6 +2063,7 @@ process_flag(_Flag, _Value) ->
       min_bin_vheap_size |
       monitored_by |
       monitors |
+      message_queue_data |
       priority |
       reductions |
       registered_name |
@@ -2123,6 +2105,7 @@ process_flag(_Flag, _Value) ->
       {monitors,
        Monitors :: [{process, Pid :: pid() |
                                      {RegName :: atom(), Node :: node()}}]} |
+      {message_queue_data, MQD :: message_queue_data()} |
       {priority, Level :: priority_level()} |
       {reductions, Number :: non_neg_integer()} |
       {registered_name, Atom :: atom()} |
@@ -2204,7 +2187,9 @@ setelement(_Index, _Tuple1, _Value) ->
 spawn_opt(_Tuple) ->
    erlang:nif_error(undefined).
 
--spec statistics(context_switches) -> {ContextSwitches,0} when
+-spec statistics(active_tasks) -> [ActiveTasks] when
+      ActiveTasks :: non_neg_integer();
+		(context_switches) -> {ContextSwitches,0} when
       ContextSwitches :: non_neg_integer();
                 (exact_reductions) -> {Total_Exact_Reductions,
                                        Exact_Reductions_Since_Last_Call} when
@@ -2221,6 +2206,8 @@ spawn_opt(_Tuple) ->
       Total_Reductions :: non_neg_integer(),
       Reductions_Since_Last_Call :: non_neg_integer();
                 (run_queue) -> non_neg_integer();
+                (run_queue_lengths) -> [RunQueueLenght] when
+      RunQueueLenght :: non_neg_integer();
                 (runtime) -> {Total_Run_Time, Time_Since_Last_Call} when
       Total_Run_Time :: non_neg_integer(),
       Time_Since_Last_Call :: non_neg_integer();
@@ -2228,6 +2215,10 @@ spawn_opt(_Tuple) ->
       SchedulerId :: pos_integer(),
       ActiveTime  :: non_neg_integer(),
       TotalTime   :: non_neg_integer();
+		(total_active_tasks) -> ActiveTasks when
+      ActiveTasks :: non_neg_integer();
+                (total_run_queue_lengths) -> TotalRunQueueLenghts when
+      TotalRunQueueLenghts :: non_neg_integer();
                 (wall_clock) -> {Total_Wallclock_Time,
                                  Wallclock_Time_Since_Last_Call} when
       Total_Wallclock_Time :: non_neg_integer(),
@@ -2423,8 +2414,9 @@ tuple_to_list(_Tuple) ->
                                   MinBinVHeapSize :: pos_integer()};
          (modified_timing_level) -> integer() | undefined;
          (multi_scheduling) -> disabled | blocked | enabled;
-         (multi_scheduling_blockers) -> [PID :: pid()];
+         (multi_scheduling_blockers) -> [Pid :: pid()];
          (nif_version) -> string();
+         (message_queue_data) -> message_queue_data();
          (otp_release) -> string();
          (os_monotonic_time_source) -> [{atom(),term()}];
          (os_system_time_source) -> [{atom(),term()}];
@@ -2552,14 +2544,19 @@ spawn_monitor(M, F, A) when erlang:is_atom(M),
 spawn_monitor(M, F, A) ->
     erlang:error(badarg, [M,F,A]).
 
+
+-type spawn_opt_option() ::
+	link
+      | monitor
+      | {priority, Level :: priority_level()}
+      | {fullsweep_after, Number :: non_neg_integer()}
+      | {min_heap_size, Size :: non_neg_integer()}
+      | {min_bin_vheap_size, VSize :: non_neg_integer()}
+      | {message_queue_data, MQD :: message_queue_data()}.
+
 -spec spawn_opt(Fun, Options) -> pid() | {pid(), reference()} when
       Fun :: function(),
-      Options :: [Option],
-      Option :: link | monitor
-              | {priority, Level :: priority_level()}
-              | {fullsweep_after, Number :: non_neg_integer()}
-              | {min_heap_size, Size :: non_neg_integer()}
-              | {min_bin_vheap_size, VSize :: non_neg_integer()}.
+      Options :: [spawn_opt_option()].
 spawn_opt(F, O) when erlang:is_function(F) ->
     spawn_opt(erlang, apply, [F, []], O);
 spawn_opt({M,F}=MF, O) when erlang:is_atom(M), erlang:is_atom(F) ->
@@ -2572,12 +2569,7 @@ spawn_opt(F, O) ->
 -spec spawn_opt(Node, Fun, Options) -> pid() | {pid(), reference()} when
       Node :: node(),
       Fun :: function(),
-      Options :: [Option],
-      Option :: link | monitor
-              | {priority, Level :: priority_level()}
-              | {fullsweep_after, Number :: non_neg_integer()}
-              | {min_heap_size, Size :: non_neg_integer()}
-              | {min_bin_vheap_size, VSize :: non_neg_integer()}.
+      Options :: [spawn_opt_option()].
 spawn_opt(N, F, O) when N =:= erlang:node() ->
     spawn_opt(F, O);
 spawn_opt(N, F, O) when erlang:is_function(F) ->
@@ -2664,12 +2656,7 @@ spawn_link(N,M,F,A) ->
       Module :: module(),
       Function :: atom(),
       Args :: [term()],
-      Options :: [Option],
-      Option :: link | monitor
-              | {priority, Level :: priority_level()}
-              | {fullsweep_after, Number :: non_neg_integer()}
-              | {min_heap_size, Size :: non_neg_integer()}
-              | {min_bin_vheap_size, VSize :: non_neg_integer()}.
+      Options :: [spawn_opt_option()].
 spawn_opt(M, F, A, Opts) ->
     case catch erlang:spawn_opt({M,F,A,Opts}) of
 	{'EXIT',{Reason,_}} ->
@@ -2684,12 +2671,7 @@ spawn_opt(M, F, A, Opts) ->
       Module :: module(),
       Function :: atom(),
       Args :: [term()],
-      Options :: [Option],
-      Option :: link | monitor
-              | {priority, Level :: priority_level()}
-              | {fullsweep_after, Number :: non_neg_integer()}
-              | {min_heap_size, Size :: non_neg_integer()}
-              | {min_bin_vheap_size, VSize :: non_neg_integer()}.
+      Options :: [spawn_opt_option()].
 spawn_opt(N, M, F, A, O) when N =:= erlang:node(),
 			      erlang:is_atom(M), erlang:is_atom(F),
                               erlang:is_list(A), erlang:is_list(O) ->
diff --git a/erts/preloaded/src/erts.app.src b/erts/preloaded/src/erts.app.src
index 8442aaf7e8..e53b6e5bab 100644
--- a/erts/preloaded/src/erts.app.src
+++ b/erts/preloaded/src/erts.app.src
@@ -27,6 +27,7 @@
 		erts_internal,
 		init,
 		otp_ring0,
+		erts_code_purger,
 		prim_eval,
 		prim_file,
 		prim_inet,
diff --git a/erts/preloaded/src/erts_code_purger.erl b/erts/preloaded/src/erts_code_purger.erl
new file mode 100644
index 0000000000..a64860bec8
--- /dev/null
+++ b/erts/preloaded/src/erts_code_purger.erl
@@ -0,0 +1,299 @@
+%%
+%% %CopyrightBegin%
+%%
+%% Copyright Ericsson AB 2016. All Rights Reserved.
+%%
+%% Licensed under the Apache License, Version 2.0 (the "License");
+%% you may not use this file except in compliance with the License.
+%% You may obtain a copy of the License at
+%%
+%%     http://www.apache.org/licenses/LICENSE-2.0
+%%
+%% Unless required by applicable law or agreed to in writing, software
+%% distributed under the License is distributed on an "AS IS" BASIS,
+%% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+%% See the License for the specific language governing permissions and
+%% limitations under the License.
+%%
+%% %CopyrightEnd%
+%%
+-module(erts_code_purger).
+
+%% Purpose : Implement system process erts_code_purger
+%%           to handle code module purging.
+
+-export([start/0, purge/1, soft_purge/1]).
+
+-spec start() -> term().
+start() ->
+    register(erts_code_purger, self()),
+    process_flag(trap_exit, true),
+    loop().
+
+loop() ->
+    receive
+	{purge,Mod,From,Ref} when is_atom(Mod), is_pid(From) ->
+	    Res = do_purge(Mod),
+	    From ! {reply, purge, Res, Ref};
+
+	{soft_purge,Mod,From,Ref} when is_atom(Mod), is_pid(From) ->
+	    Res = do_soft_purge(Mod),
+	    From ! {reply, soft_purge, Res, Ref};
+
+	_Other -> ignore
+    end,
+    loop().
+
+
+%% purge(Module)
+%%  Kill all processes running code from *old* Module, and then purge the
+%%  module. Return {WasOld, DidKill}:
+%%  {false, false} there was no old module to purge
+%%  {true, false} module purged, no process killed
+%%  {true, true} module purged, at least one process killed
+
+purge(Mod) when is_atom(Mod) ->
+    Ref = make_ref(),
+    erts_code_purger ! {purge, Mod, self(), Ref},
+    receive
+	{reply, purge, Result, Ref} ->
+	    Result
+    end.
+
+
+do_purge(Mod) ->
+    case erts_internal:copy_literals(Mod, true) of
+        false ->
+            {false, false};
+        true ->
+            DidKill = check_proc_code(erlang:processes(), Mod, true),
+            true = erts_internal:copy_literals(Mod, false),
+	    WasPurged = erts_internal:purge_module(Mod),
+            {WasPurged, DidKill}
+    end.
+
+%% soft_purge(Module)
+%% Purge old code only if no procs remain that run old code.
+%% Return true in that case, false if procs remain (in this
+%% case old code is not purged)
+
+soft_purge(Mod) ->
+    Ref = make_ref(),
+    erts_code_purger ! {soft_purge, Mod, self(), Ref},
+    receive
+	{reply, soft_purge, Result, Ref} ->
+	    Result
+    end.
+
+
+do_soft_purge(Mod) ->
+    case erts_internal:copy_literals(Mod, true) of
+	false ->
+	    true;
+	true ->
+	    DoPurge = check_proc_code(erlang:processes(), Mod, false),
+	    true = erts_internal:copy_literals(Mod, false),
+	    case DoPurge of
+		false ->
+		    false;
+		true ->
+		    erts_internal:purge_module(Mod),
+		    true
+	    end
+    end.
+
+%%
+%% check_proc_code(Pids, Mod, Hard) - Send asynchronous
+%%   requests to all processes to perform a check_process_code
+%%   operation. Each process will check their own state and
+%%   reply with the result. If 'Hard' equals
+%%   - true, processes that refer 'Mod' will be killed. If
+%%     any processes were killed true is returned; otherwise,
+%%     false.
+%%   - false, and any processes refer 'Mod', false will
+%%     returned; otherwise, true.
+%%
+%%   Requests will be sent to all processes identified by
+%%   Pids at once, but without allowing GC to be performed.
+%%   Check process code operations that are aborted due to
+%%   GC need, will be restarted allowing GC. However, only
+%%   ?MAX_CPC_GC_PROCS outstanding operation allowing GC at
+%%   a time will be allowed. This in order not to blow up
+%%   memory wise.
+%%
+%%   We also only allow ?MAX_CPC_NO_OUTSTANDING_KILLS
+%%   outstanding kills. This both in order to avoid flooding
+%%   our message queue with 'DOWN' messages and limiting the
+%%   amount of memory used to keep references to all
+%%   outstanding kills.
+%%
+
+%% We maybe should allow more than two outstanding
+%% GC requests, but for now we play it safe...
+-define(MAX_CPC_GC_PROCS, 2).
+-define(MAX_CPC_NO_OUTSTANDING_KILLS, 10).
+
+-record(cpc_static, {hard, module, tag}).
+
+-record(cpc_kill, {outstanding = [],
+		   no_outstanding = 0,
+		   waiting = [],
+		   killed = false}).
+
+check_proc_code(Pids, Mod, Hard) ->
+    Tag = erlang:make_ref(),
+    CpcS = #cpc_static{hard = Hard,
+		       module = Mod,
+		       tag = Tag},
+    check_proc_code(CpcS, cpc_init(CpcS, Pids, 0), 0, [], #cpc_kill{}, true).
+
+check_proc_code(#cpc_static{hard = true}, 0, 0, [],
+		#cpc_kill{outstanding = [], waiting = [], killed = Killed},
+		true) ->
+    %% No outstanding requests. We did a hard check, so result is whether or
+    %% not we killed any processes...
+    Killed;
+check_proc_code(#cpc_static{hard = false}, 0, 0, [], _KillState, Success) ->
+    %% No outstanding requests and we did a soft check...
+    Success;
+check_proc_code(#cpc_static{hard = false, tag = Tag} = CpcS, NoReq0, NoGcReq0,
+		[], _KillState, false) ->
+    %% Failed soft check; just cleanup the remaining replies corresponding
+    %% to the requests we've sent...
+    {NoReq1, NoGcReq1} = receive
+			     {check_process_code, {Tag, _P, GC}, _Res} ->
+				 case GC of
+				     false -> {NoReq0-1, NoGcReq0};
+				     true -> {NoReq0, NoGcReq0-1}
+				 end
+			 end,
+    check_proc_code(CpcS, NoReq1, NoGcReq1, [], _KillState, false);
+check_proc_code(#cpc_static{tag = Tag} = CpcS, NoReq0, NoGcReq0, NeedGC0,
+		KillState0, Success) ->
+
+    %% Check if we should request a GC operation
+    {NoGcReq1, NeedGC1} = case NoGcReq0 < ?MAX_CPC_GC_PROCS of
+			      GcOpAllowed when GcOpAllowed == false;
+					       NeedGC0 == [] ->
+				  {NoGcReq0, NeedGC0};
+			      _ ->
+				  {NoGcReq0+1, cpc_request_gc(CpcS,NeedGC0)}
+			  end,
+
+    %% Wait for a cpc reply or 'DOWN' message
+    {NoReq1, NoGcReq2, Pid, Result, KillState1} = cpc_recv(Tag,
+							   NoReq0,
+							   NoGcReq1,
+							   KillState0),
+
+    %% Check the result of the reply
+    case Result of
+	aborted ->
+	    %% Operation aborted due to the need to GC in order to
+	    %% determine if the process is referring the module.
+	    %% Schedule the operation for restart allowing GC...
+	    check_proc_code(CpcS, NoReq1, NoGcReq2, [Pid|NeedGC1], KillState1,
+			    Success);
+	false ->
+	    %% Process not referring the module; done with this process...
+	    check_proc_code(CpcS, NoReq1, NoGcReq2, NeedGC1, KillState1,
+			    Success);
+	true ->
+	    %% Process referring the module...
+	    case CpcS#cpc_static.hard of
+		false ->
+		    %% ... and soft check. The whole operation failed so
+		    %% no point continuing; clean up and fail...
+		    check_proc_code(CpcS, NoReq1, NoGcReq2, [], KillState1,
+				    false);
+		true ->
+		    %% ... and hard check; schedule kill of it...
+		    check_proc_code(CpcS, NoReq1, NoGcReq2, NeedGC1,
+				    cpc_sched_kill(Pid, KillState1), Success)
+	    end;
+	'DOWN' ->
+	    %% Handled 'DOWN' message
+	    check_proc_code(CpcS, NoReq1, NoGcReq2, NeedGC1,
+			    KillState1, Success)
+    end.
+
+cpc_recv(Tag, NoReq, NoGcReq, #cpc_kill{outstanding = []} = KillState) ->
+    receive
+	{check_process_code, {Tag, Pid, GC}, Res} ->
+	    cpc_handle_cpc(NoReq, NoGcReq, GC, Pid, Res, KillState)
+    end;
+cpc_recv(Tag, NoReq, NoGcReq,
+	 #cpc_kill{outstanding = [R0, R1, R2, R3, R4 | _]} = KillState) ->
+    receive
+	{'DOWN', R, process, _, _} when R == R0;
+					R == R1;
+					R == R2;
+					R == R3;
+					R == R4 ->
+	    cpc_handle_down(NoReq, NoGcReq, R, KillState);
+	{check_process_code, {Tag, Pid, GC}, Res} ->
+	    cpc_handle_cpc(NoReq, NoGcReq, GC, Pid, Res, KillState)
+    end;
+cpc_recv(Tag, NoReq, NoGcReq, #cpc_kill{outstanding = [R|_]} = KillState) ->
+    receive
+	{'DOWN', R, process, _, _} ->
+	    cpc_handle_down(NoReq, NoGcReq, R, KillState);
+	{check_process_code, {Tag, Pid, GC}, Res} ->
+	    cpc_handle_cpc(NoReq, NoGcReq, GC, Pid, Res, KillState)
+    end.
+
+cpc_handle_down(NoReq, NoGcReq, R, #cpc_kill{outstanding = Rs,
+					     no_outstanding = N} = KillState) ->
+    {NoReq, NoGcReq, undefined, 'DOWN',
+     cpc_sched_kill_waiting(KillState#cpc_kill{outstanding = cpc_list_rm(R, Rs),
+					       no_outstanding = N-1})}.
+
+cpc_list_rm(R, [R|Rs]) ->
+    Rs;
+cpc_list_rm(R0, [R1|Rs]) ->
+    [R1|cpc_list_rm(R0, Rs)].
+
+cpc_handle_cpc(NoReq, NoGcReq, false, Pid, Res, KillState) ->
+    {NoReq-1, NoGcReq, Pid, Res, KillState};
+cpc_handle_cpc(NoReq, NoGcReq, true, Pid, Res, KillState) ->
+    {NoReq, NoGcReq-1, Pid, Res, KillState}.
+
+cpc_sched_kill_waiting(#cpc_kill{waiting = []} = KillState) ->
+    KillState;
+cpc_sched_kill_waiting(#cpc_kill{outstanding = Rs,
+				 no_outstanding = N,
+				 waiting = [P|Ps]} = KillState) ->
+    R = erlang:monitor(process, P),
+    exit(P, kill),
+    KillState#cpc_kill{outstanding = [R|Rs],
+		       no_outstanding = N+1,
+		       waiting = Ps,
+		       killed = true}.
+
+cpc_sched_kill(Pid, #cpc_kill{no_outstanding = N, waiting = Pids} = KillState)
+  when N >= ?MAX_CPC_NO_OUTSTANDING_KILLS ->
+    KillState#cpc_kill{waiting = [Pid|Pids]};
+cpc_sched_kill(Pid,
+	       #cpc_kill{outstanding = Rs, no_outstanding = N} = KillState) ->
+    R = erlang:monitor(process, Pid),
+    exit(Pid, kill),
+    KillState#cpc_kill{outstanding = [R|Rs],
+		       no_outstanding = N+1,
+		       killed = true}.
+
+cpc_request(#cpc_static{tag = Tag, module = Mod}, Pid, AllowGc) ->
+    erts_internal:check_process_code(Pid, Mod, [{async, {Tag, Pid, AllowGc}},
+						{allow_gc, AllowGc},
+						{copy_literals, true}]).
+
+cpc_request_gc(CpcS, [Pid|Pids]) ->
+    cpc_request(CpcS, Pid, true),
+    Pids.
+
+cpc_init(_CpcS, [], NoReqs) ->
+    NoReqs;
+cpc_init(CpcS, [Pid|Pids], NoReqs) ->
+    cpc_request(CpcS, Pid, false),
+    cpc_init(CpcS, Pids, NoReqs+1).
+
+% end of check_proc_code() implementation.
diff --git a/erts/preloaded/src/erts_internal.erl b/erts/preloaded/src/erts_internal.erl
index 7ed4efea4b..84dedab930 100644
--- a/erts/preloaded/src/erts_internal.erl
+++ b/erts/preloaded/src/erts_internal.erl
@@ -31,13 +31,15 @@
 
 -export([await_port_send_result/3]).
 -export([cmp_term/2]).
--export([map_to_tuple_keys/1, map_type/1, map_hashmap_children/1]).
--export([port_command/3, port_connect/2, port_close/1,
+-export([map_to_tuple_keys/1, term_type/1, map_hashmap_children/1]).
+-export([open_port/2, port_command/3, port_connect/2, port_close/1,
 	 port_control/3, port_call/3, port_info/1, port_info/2]).
 
 -export([request_system_task/3]).
 
--export([check_process_code/2]).
+-export([check_process_code/3]).
+-export([copy_literals/2]).
+-export([purge_module/1]).
 
 -export([flush_monitor_messages/3]).
 
@@ -47,6 +49,9 @@
 
 -export([is_system_process/1]).
 
+%% Auto import name clash
+-export([check_process_code/2]).
+
 %%
 %% Await result of send to port
 %%
@@ -88,6 +93,13 @@ gather_io_bytes(Ref, No, InAcc, OutAcc) ->
 %% Statically linked port NIFs
 %%
 
+-spec erts_internal:open_port(PortName, PortSettings) -> Result when
+      PortName :: tuple(),
+      PortSettings :: term(),
+      Result :: port() | reference() | atom().
+open_port(_PortName, _PortSettings) ->
+    erlang:nif_error(undefined).
+
 -spec erts_internal:port_command(Port, Data, OptionList) -> Result when
       Port :: port() | atom(),
       Data :: iodata(),
@@ -190,11 +202,80 @@ port_info(_Result, _Item) ->
 request_system_task(_Pid, _Prio, _Request) ->
     erlang:nif_error(undefined).
 
--spec check_process_code(Module, OptionList) -> boolean() when
+-define(ERTS_CPC_ALLOW_GC, (1 bsl 0)).
+-define(ERTS_CPC_COPY_LITERALS, (1 bsl 1)).
+
+-spec check_process_code(Module, Flags) -> boolean() when
+      Module :: module(),
+      Flags :: non_neg_integer().
+check_process_code(_Module, _Flags) ->
+    erlang:nif_error(undefined).
+
+-spec check_process_code(Pid, Module, OptionList) -> CheckResult | async when
+      Pid :: pid(),
+      Module :: module(),
+      RequestId :: term(),
+      Option :: {async, RequestId} | {allow_gc, boolean()} | {copy_literals, boolean()},
+      OptionList :: [Option],
+      CheckResult :: boolean() | aborted.
+check_process_code(Pid, Module, OptionList)  ->
+    {Async, Flags} = get_cpc_opts(OptionList, sync, ?ERTS_CPC_ALLOW_GC),
+    case Async of
+	{async, ReqId} ->
+	    {priority, Prio} = erlang:process_info(erlang:self(),
+						   priority),
+	    erts_internal:request_system_task(Pid,
+					      Prio,
+					      {check_process_code,
+					       ReqId,
+					       Module,
+					       Flags}),
+	    async;
+	sync ->
+	    case Pid == erlang:self() of
+		true ->
+		    erts_internal:check_process_code(Module, Flags);
+		false ->
+		    {priority, Prio} = erlang:process_info(erlang:self(),
+							   priority),
+		    ReqId = erlang:make_ref(),
+		    erts_internal:request_system_task(Pid,
+						      Prio,
+						      {check_process_code,
+						       ReqId,
+						       Module,
+						       Flags}),
+		    receive
+			{check_process_code, ReqId, CheckResult} ->
+			    CheckResult
+		    end
+	    end
+    end.
+
+% gets async and flag opts and verify valid option list
+get_cpc_opts([{async, _ReqId} = AsyncTuple | Options], _OldAsync, Flags) ->
+    get_cpc_opts(Options, AsyncTuple, Flags);
+get_cpc_opts([{allow_gc, AllowGC} | Options], Async, Flags) ->
+    get_cpc_opts(Options, Async, cpc_flags(Flags, ?ERTS_CPC_ALLOW_GC, AllowGC));
+get_cpc_opts([{copy_literals, CopyLit} | Options], Async, Flags) ->
+    get_cpc_opts(Options, Async, cpc_flags(Flags, ?ERTS_CPC_COPY_LITERALS, CopyLit));
+get_cpc_opts([], Async, Flags) ->
+    {Async, Flags}.
+
+cpc_flags(OldFlags, Bit, true) ->
+    OldFlags bor Bit;
+cpc_flags(OldFlags, Bit, false) ->
+    OldFlags band (bnot Bit).
+
+-spec copy_literals(Module,Bool) -> 'true' | 'false' | 'aborted' when
       Module :: module(),
-      Option :: {allow_gc, boolean()},
-      OptionList :: [Option].
-check_process_code(_Module, _OptionList) ->
+      Bool :: boolean().
+copy_literals(_Mod, _Bool) ->
+    erlang:nif_error(undefined).
+
+-spec purge_module(Module) -> boolean() when
+      Module :: module().
+purge_module(_Module) ->
     erlang:nif_error(undefined).
 
 %% term compare where integer() < float() = true
@@ -215,12 +296,18 @@ cmp_term(_A,_B) ->
 map_to_tuple_keys(_M) ->
     erlang:nif_error(undefined).
 
-%% return the internal map type
--spec map_type(M) -> Type when
-    M :: map(),
-    Type :: 'flatmap' | 'hashmap' | 'hashmap_node'.
-
-map_type(_M) ->
+%% return the internal term type
+-spec term_type(T) -> Type when
+    T :: term(),
+    Type :: 'flatmap' | 'hashmap' | 'hashmap_node'
+          | 'fixnum'  | 'bignum'  | 'hfloat'
+          | 'list' | 'tuple' | 'export' | 'fun'
+          | 'refc_binary' | 'heap_binary' | 'sub_binary'
+          | 'reference'   | 'external_reference'
+          | 'pid' | 'external_pid' | 'port' | 'external_port'
+          | 'atom' | 'catch' | 'nil'.
+
+term_type(_T) ->
     erlang:nif_error(undefined).
 
 %% return the internal hashmap sub-nodes from
diff --git a/erts/preloaded/src/init.erl b/erts/preloaded/src/init.erl
index c4e37b76f1..ed65c57c0d 100644
--- a/erts/preloaded/src/init.erl
+++ b/erts/preloaded/src/init.erl
@@ -23,7 +23,6 @@
 %% a local file or distributed from another erlang node.
 %% 
 %% Flags:
-%%        -id Identity   : identity of the system.
 %%        -boot File     : Absolute file name of the boot script.
 %%        -boot_var Var Value
 %%                       : $Var in the boot script is expanded to
@@ -75,6 +74,19 @@
 		subscribed = []}).
 -type state() :: #state{}.
 
+%% Data for eval_script/2.
+-record(es,
+	{init,
+	 debug,
+	 path,
+	 pa,
+	 pz,
+	 path_choice,
+	 prim_load,
+	 load_mode,
+	 vars
+	}).
+
 -define(ON_LOAD_HANDLER, init__boot__on_load_handler).
 
 debug(false, _) -> ok;
@@ -167,11 +179,9 @@ stop(Status) -> init ! {stop,{stop,Status}}, ok.
 boot(BootArgs) ->
     register(init, self()),
     process_flag(trap_exit, true),
-    start_on_load_handler_process(),
     {Start0,Flags,Args} = parse_boot_args(BootArgs),
     Start = map(fun prepare_run_args/1, Start0),
-    Flags0 = flags_to_atoms_again(Flags),
-    boot(Start,Flags0,Args).
+    boot(Start, Flags, Args).
 
 prepare_run_args({eval, [Expr]}) ->
     {eval,Expr};
@@ -203,16 +213,6 @@ map(_F, []) ->
 map(F, [X|Rest]) ->
     [F(X) | map(F, Rest)].
 
-flags_to_atoms_again([]) ->
-    [];
-flags_to_atoms_again([{F0,L0}|Rest]) ->
-    L = L0,
-    F = b2a(F0),
-    [{F,L}|flags_to_atoms_again(Rest)];
-flags_to_atoms_again([{F0}|Rest]) ->
-    F = b2a(F0),
-    [{F}|flags_to_atoms_again(Rest)].
-
 -spec code_path_choice() -> 'relaxed' | 'strict'.
 code_path_choice() ->
     case get_argument(code_path_choice) of
@@ -225,6 +225,7 @@ code_path_choice() ->
     end.
 
 boot(Start,Flags,Args) ->
+    start_on_load_handler_process(),
     BootPid = do_boot(Flags,Start),
     State = #state{flags = Flags,
 		   args = Args,
@@ -337,7 +338,7 @@ boot_loop(BootPid, State) ->
     end.
 
 ensure_loaded(Module, Loaded) ->
-    File = concat([Module,objfile_extension()]),
+    File = atom_to_list(Module) ++ objfile_extension(),
     case catch load_mod(Module,File) of
 	{ok, FullName} ->
 	    {{module, Module}, [{Module, FullName}|Loaded]};
@@ -451,9 +452,9 @@ do_handle_msg(Msg,State) ->
 %%% -------------------------------------------------
 
 make_permanent(Boot,Config,Flags0,State) ->
-    case set_flag('-boot',Boot,Flags0) of
+    case set_flag(boot, Boot, Flags0) of
 	{ok,Flags1} ->
-	    case set_flag('-config',Config,Flags1) of
+	    case set_flag(config, Config, Flags1) of
 		{ok,Flags} ->
 		    {ok,State#state{flags = Flags}};
 		Error ->
@@ -635,9 +636,9 @@ unload(_) ->
     do_unload(sub([heart|erlang:pre_loaded()],erlang:loaded())).
 
 do_unload([M|Mods]) ->
-    catch erlang:purge_module(M),
+    catch erts_internal:purge_module(M),
     catch erlang:delete_module(M),
-    catch erlang:purge_module(M),
+    catch erts_internal:purge_module(M),
     do_unload(Mods);
 do_unload([]) ->
     purge_all_hipe_refs(),
@@ -691,17 +692,15 @@ sleep(T) -> receive after T -> ok end.
 %%% The loader shall run for ever!
 %%% -------------------------------------------------
 
-start_prim_loader(Init,Id,Pgm,Nodes,Path,{Pa,Pz}) ->
-    case erl_prim_loader:start(Id,Pgm,Nodes) of
-	{ok,Pid} when Path =:= false ->
-	    InitPath = append(Pa,["."|Pz]),
-	    erl_prim_loader:set_path(InitPath),
-	    add_to_kernel(Init,Pid),
-	    Pid;
+start_prim_loader(Init, Path0, {Pa,Pz}) ->
+    Path = case Path0 of
+	       false -> Pa ++ ["."|Pz];
+	       _ -> Path0
+	   end,
+    case erl_prim_loader:start() of
 	{ok,Pid} ->
 	    erl_prim_loader:set_path(Path),
-	    add_to_kernel(Init,Pid),
-	    Pid;
+	    add_to_kernel(Init, Pid);
 	{error,Reason} ->
 	    erlang:display({"cannot start loader",Reason}),
 	    exit(Reason)
@@ -715,13 +714,6 @@ add_to_kernel(Init,Pid) ->
 	    ok
     end.
 
-prim_load_flags(Flags) ->
-    PortPgm = get_flag('-loader',Flags,<<"efile">>),
-    Hosts = get_flag_list('-hosts', Flags, []),
-    Id = get_flag('-id',Flags,none),
-    Path = get_flag_list('-path',Flags,false),
-    {PortPgm, Hosts, Id, Path}.
-
 %%% -------------------------------------------------
 %%% The boot process fetches a boot script and loads
 %%% all modules specified and starts spec. processes.
@@ -734,24 +726,23 @@ do_boot(Flags,Start) ->
 
 do_boot(Init,Flags,Start) ->
     process_flag(trap_exit,true),
-    {Pgm0,Nodes,Id,Path} = prim_load_flags(Flags),
-    Root = b2s(get_flag('-root',Flags)),
-    PathFls = path_flags(Flags),
-    Pgm = b2s(Pgm0),
-    _Pid = start_prim_loader(Init,b2a(Id),Pgm,bs2as(Nodes),
-			     bs2ss(Path),PathFls),
+    Root = get_root(Flags),
+    Path = get_flag_list(path, Flags, false),
+    {Pa,Pz} = PathFls = path_flags(Flags),
+    start_prim_loader(Init, bs2ss(Path), PathFls),
     BootFile = bootfile(Flags,Root),
     BootList = get_boot(BootFile,Root),
-    LoadMode = b2a(get_flag('-mode',Flags,false)),
-    Deb = b2a(get_flag('-init_debug',Flags,false)),
+    LoadMode = b2a(get_flag(mode, Flags, false)),
+    Deb = b2a(get_flag(init_debug, Flags, false)),
     catch ?ON_LOAD_HANDLER ! {init_debug_flag,Deb},
-    BootVars = get_flag_args('-boot_var',Flags),
-    ParallelLoad = 
-	(Pgm =:= "efile") and (erlang:system_info(thread_pool_size) > 0),
+    BootVars = get_boot_vars(Root, Flags),
 
     PathChoice = code_path_choice(),
-    eval_script(BootList,Init,PathFls,{Root,BootVars},Path,
-		{true,LoadMode,ParallelLoad},Deb,PathChoice),
+    Es = #es{init=Init,debug=Deb,path=Path,pa=Pa,pz=Pz,
+	     path_choice=PathChoice,
+	     prim_load=true,load_mode=LoadMode,
+	     vars=BootVars},
+    eval_script(BootList, Es),
 
     %% To help identifying Purify windows that pop up,
     %% print the node name into the Purify log.
@@ -759,21 +750,43 @@ do_boot(Init,Flags,Start) ->
 
     start_em(Start).
 
+get_root(Flags) ->
+    case get_argument(root, Flags) of
+	{ok,[[Root]]} ->
+	    Root;
+	_ ->
+	    exit(no_or_multiple_root_variables)
+    end.
+
+get_boot_vars(Root, Flags) ->
+    BootVars = get_boot_vars_1(#{}, Flags),
+    RootKey = <<"ROOT">>,
+    BootVars#{RootKey=>Root}.
+
+get_boot_vars_1(Vars, [{boot_var,[Key,Value]}|T]) ->
+    get_boot_vars_1(Vars#{Key=>Value}, T);
+get_boot_vars_1(_, [{boot_var,_}|_]) ->
+    exit(invalid_boot_var_argument);
+get_boot_vars_1(Vars, [_|T]) ->
+    get_boot_vars_1(Vars, T);
+get_boot_vars_1(Vars, []) ->
+    Vars.
+
 bootfile(Flags,Root) ->
-    b2s(get_flag('-boot',Flags,concat([Root,"/bin/start"]))).
+    b2s(get_flag(boot, Flags, Root++"/bin/start")).
 
 path_flags(Flags) ->
-    Pa = append(reverse(get_flag_args('-pa',Flags))),
-    Pz = append(get_flag_args('-pz',Flags)),
+    Pa = append(reverse(get_flag_args(pa, Flags))),
+    Pz = append(get_flag_args(pz, Flags)),
     {bs2ss(Pa),bs2ss(Pz)}.
 
 get_boot(BootFile0,Root) ->
-    BootFile = concat([BootFile0,".boot"]),
+    BootFile = BootFile0 ++ ".boot",
     case get_boot(BootFile) of
 	{ok, CmdList} ->
 	    CmdList;
 	not_found -> %% Check for default.
-	    BootF = concat([Root,"/bin/",BootFile]),
+	    BootF = Root ++ "/bin/" ++ BootFile,
 	    case get_boot(BootF)  of
 		{ok, CmdList} ->
 		    CmdList;
@@ -807,93 +820,67 @@ get_boot(BootFile) ->
 %% boot process hangs (we want to ensure syncronicity).
 %%
 
-eval_script([{progress,Info}|CfgL],Init,PathFs,Vars,P,Ph,Deb,PathChoice) ->
-    debug(Deb,{progress,Info}),
+eval_script([{progress,Info}=Progress|T], #es{debug=Deb}=Es) ->
+    debug(Deb, Progress),
     init ! {self(),progress,Info},
-    eval_script(CfgL,Init,PathFs,Vars,P,Ph,Deb,PathChoice);
-eval_script([{preLoaded,_}|CfgL],Init,PathFs,Vars,P,Ph,Deb,PathChoice) ->
-    eval_script(CfgL,Init,PathFs,Vars,P,Ph,Deb,PathChoice);
-eval_script([{path,Path}|CfgL],Init,{Pa,Pz},Vars,false,Ph,Deb,PathChoice) ->
+    eval_script(T, Es);
+eval_script([{preLoaded,_}|T], #es{}=Es) ->
+    eval_script(T, Es);
+eval_script([{path,Path}|T], #es{path=false,pa=Pa,pz=Pz,
+				 path_choice=PathChoice,
+				 vars=Vars}=Es) ->
     RealPath0 = make_path(Pa, Pz, Path, Vars),
     RealPath = patch_path(RealPath0, PathChoice),
     erl_prim_loader:set_path(RealPath),
-    eval_script(CfgL,Init,{Pa,Pz},Vars,false,Ph,Deb,PathChoice);
-eval_script([{path,_}|CfgL],Init,PathFs,Vars,P,Ph,Deb,PathChoice) ->
+    eval_script(T, Es);
+eval_script([{path,_}|T], #es{}=Es) ->
     %% Ignore, use the command line -path flag.
-    eval_script(CfgL,Init,PathFs,Vars,P,Ph,Deb,PathChoice);
-eval_script([{kernel_load_completed}|CfgL],Init,PathFs,Vars,P,{_,embedded,Par},Deb,PathChoice) ->
-    eval_script(CfgL,Init,PathFs,Vars,P,{true,embedded,Par},Deb,PathChoice);
-eval_script([{kernel_load_completed}|CfgL],Init,PathFs,Vars,P,{_,E,Par},Deb,PathChoice) ->
-    eval_script(CfgL,Init,PathFs,Vars,P,{false,E,Par},Deb,PathChoice);
-eval_script([{primLoad,Mods}|CfgL],Init,PathFs,Vars,P,{true,E,Par},Deb,PathChoice)
+    eval_script(T, Es);
+eval_script([{kernel_load_completed}|T], #es{load_mode=Mode}=Es0) ->
+    Es = case Mode of
+	     embedded -> Es0;
+	     _ -> Es0#es{prim_load=false}
+	 end,
+    eval_script(T, Es);
+eval_script([{primLoad,[Mod]}|T], #es{prim_load=true}=Es) ->
+    %% Common special case (loading of error_handler). Nothing
+    %% to gain by parallel loading.
+    File = atom_to_list(Mod) ++ objfile_extension(),
+    {ok,Full} = load_mod(Mod, File),
+    init ! {self(),loaded,{Mod,Full}}, % Tell init about loaded module
+    eval_script(T, Es);
+eval_script([{primLoad,Mods}|T], #es{init=Init,prim_load=PrimLoad}=Es)
   when is_list(Mods) ->
-    if 
-	Par =:= true ->
-	    par_load_modules(Mods,Init);
+    case PrimLoad of
 	true ->
-	    load_modules(Mods)
+	    load_modules(Mods, Init);
+	false ->
+	    %% Do not load now, code_server does that dynamically!
+	    ok
     end,
-    eval_script(CfgL,Init,PathFs,Vars,P,{true,E,Par},Deb,PathChoice);
-eval_script([{primLoad,_Mods}|CfgL],Init,PathFs,Vars,P,{false,E,Par},Deb,PathChoice) ->
-    %% Do not load now, code_server does that dynamically!
-    eval_script(CfgL,Init,PathFs,Vars,P,{false,E,Par},Deb,PathChoice);
-eval_script([{kernelProcess,Server,{Mod,Fun,Args}}|CfgL],Init,
-	    PathFs,Vars,P,Ph,Deb,PathChoice) ->
-    debug(Deb,{start,Server}),
-    start_in_kernel(Server,Mod,Fun,Args,Init),
-    eval_script(CfgL,Init,PathFs,Vars,P,Ph,Deb,PathChoice);
-eval_script([{apply,{Mod,Fun,Args}}|CfgL],Init,PathFs,Vars,P,Ph,Deb,PathChoice) ->
-    debug(Deb,{apply,{Mod,Fun,Args}}),
-    apply(Mod,Fun,Args),
-    eval_script(CfgL,Init,PathFs,Vars,P,Ph,Deb,PathChoice);
-eval_script([],_,_,_,_,_,_,_) ->
+    eval_script(T, Es);
+eval_script([{kernelProcess,Server,{Mod,Fun,Args}}|T],
+	    #es{init=Init,debug=Deb}=Es) ->
+    debug(Deb, {start,Server}),
+    start_in_kernel(Server, Mod, Fun, Args, Init),
+    eval_script(T, Es);
+eval_script([{apply,{Mod,Fun,Args}}=Apply|T], #es{debug=Deb}=Es) ->
+    debug(Deb, Apply),
+    apply(Mod, Fun, Args),
+    eval_script(T, Es);
+eval_script([], #es{}) ->
     ok;
-eval_script(What,_,_,_,_,_,_,_) ->
+eval_script(What, #es{}) ->
     exit({'unexpected command in bootfile',What}).
 
-load_modules([Mod|Mods]) ->
-    File = concat([Mod,objfile_extension()]),
+load_modules([Mod|Mods], Init) ->
+    File = atom_to_list(Mod) ++ objfile_extension(),
     {ok,Full} = load_mod(Mod,File),
-    init ! {self(),loaded,{Mod,Full}},  %% Tell init about loaded module
-    load_modules(Mods);
-load_modules([]) ->
+    Init ! {self(),loaded,{Mod,Full}},	%Tell init about loaded module
+    load_modules(Mods, Init);
+load_modules([], _) ->
     ok.
 
-%%% An optimization: erl_prim_loader gets the chance of loading many
-%%% files in parallel, using threads. This will reduce the seek times,
-%%% and loaded code can be processed while other threads are waiting
-%%% for the disk. The optimization is not tried unless the loader is
-%%% "efile" and there is a non-empty pool of threads.
-%%%
-%%% Many threads are needed to get a good result, so it would be
-%%% beneficial to load several applications in parallel. However,
-%%% measurements show that the file system handles one directory at a
-%%% time, regardless if parallel threads are created for files on
-%%% several directories (a guess: writing the meta information when
-%%% the file was last read ('mtime'), forces the file system to sync
-%%% between directories).
-
-par_load_modules(Mods,Init) ->
-    Ext = objfile_extension(),
-    ModFiles = [{Mod,concat([Mod,Ext])} || Mod <- Mods,
-					   not erlang:module_loaded(Mod)],
-    Self = self(),
-    Fun = fun(Mod, BinCode, FullName) ->
-		  case catch load_mod_code(Mod, BinCode, FullName) of
-		      {ok, _} ->
-			  Init ! {Self,loaded,{Mod,FullName}},
-			  ok;
-		      _EXIT ->
-			  {error, Mod}
-		  end
-	  end,
-    case erl_prim_loader:get_files(ModFiles, Fun) of
-	ok ->
-	    ok;
-	{error,Mod} ->
-	    exit({'cannot load',Mod,get_files})
-    end.
-
 make_path(Pa, Pz, Path, Vars) ->
     append([Pa,append([fix_path(Path,Vars),Pz])]).
 
@@ -908,34 +895,25 @@ fix_path([Path|Ps], Vars) ->
 fix_path(_, _) ->
     [].
 
-add_var("$ROOT/" ++ Path, {Root,_}) ->
-    concat([Root, "/", Path]);
-add_var([$$|Path0], {_,VarList}) ->
-    {Var,Path} = extract_var(Path0,[]),
-    Value = b2s(get_var_value(list_to_binary(Var),VarList)),
-    concat([Value, "/", Path]);
-add_var(Path, _)   ->
+add_var("$"++Path0, Vars) ->
+    {Var,Path} = extract_var(Path0, []),
+    Key = list_to_binary(Var),
+    case Vars of
+	#{Key:=Value0} ->
+	    Value = b2s(Value0),
+	    Value ++ "/" ++ Path;
+	_ ->
+	    Error0 = "cannot expand $" ++ Var ++ " in bootfile",
+	    Error = list_to_atom(Error0),
+	    exit(Error)
+    end;
+add_var(Path, _) ->
     Path.
 
 extract_var([$/|Path],Var) -> {reverse(Var),Path};
 extract_var([H|T],Var)     -> extract_var(T,[H|Var]);
 extract_var([],Var)        -> {reverse(Var),[]}.
 
-%% get_var_value(Var, [Vars]) where Vars == [atom()]
-get_var_value(Var,[Vars|VarList]) ->
-    case get_var_val(Var,Vars) of
-	{ok, Value} ->
-	    Value;
-	_ ->
-	    get_var_value(Var,VarList)
-    end;
-get_var_value(Var,[]) ->
-    exit(list_to_atom(concat(["cannot expand \$", Var, " in bootfile"]))).
-
-get_var_val(Var,[Var,Value|_]) -> {ok, Value};
-get_var_val(Var,[_,_|Vars])    -> get_var_val(Var,Vars);
-get_var_val(_,_)               -> false.
-
 patch_path(Dirs, strict) ->
     Dirs;
 patch_path(Dirs, relaxed) ->
@@ -1049,18 +1027,10 @@ start_it({eval,Bin}) ->
     {value, _Value, _Bs} = erl_eval:exprs(Expr, erl_eval:new_bindings()),
     ok;
 start_it([_|_]=MFA) ->
-    Ref = make_ref(),
-    case catch {Ref,case MFA of
-			[M]        -> M:start();
-			[M,F]      -> M:F();
-			[M,F|Args] -> M:F(Args)	% Args is a list
-		    end} of
-	{Ref,R} ->
-	    R;
-	{'EXIT',Reason} ->
-	    exit(Reason);
-	Other ->
-	    throw(Other)
+    case MFA of
+	[M]        -> M:start();
+	[M,F]      -> M:F();
+	[M,F|Args] -> M:F(Args)	% Args is a list
     end.
 
 %%
@@ -1102,7 +1072,7 @@ load_mod_code(Mod, BinCode, FullName) ->
 %% --------------------------------------------------------
 
 shutdown_timer(Flags) ->
-    case get_flag('-shutdown_time',Flags,infinity) of
+    case get_flag(shutdown_time, Flags, infinity) of
 	infinity ->
 	    self();
 	Time ->
@@ -1152,14 +1122,10 @@ parse_boot_args([B|Bs], Ss, Fs, As) ->
 	eval_arg ->
 	    {Expr,Rest} = get_args(Bs, []),
 	    parse_boot_args(Rest, [{eval, Expr}|Ss], Fs, As);
-	flag ->
+	{flag,A} ->
 	    {F,Rest} = get_args(Bs, []),
-	    Fl = case F of
-		     []   -> [B];
-		     FF   -> [B,FF]
-		 end,
-	    parse_boot_args(Rest, Ss,  
-			    [list_to_tuple(Fl)|Fs], As);
+	    Fl = {A,F},
+	    parse_boot_args(Rest, Ss, [Fl|Fs], As);
 	arg ->
 	    parse_boot_args(Bs, Ss, Fs, [B|As]);
 	end_args ->
@@ -1173,12 +1139,8 @@ check(<<"-s">>) -> start_arg;
 check(<<"-run">>) -> start_arg2;
 check(<<"-eval">>) -> eval_arg;
 check(<<"--">>) -> end_args;
-check(X) when is_binary(X) ->
-    case binary_to_list(X) of
-	[$-|_Rest] -> flag;
-	_Chars     -> arg			%Even empty atoms
-    end;
-check(_X) -> arg.				%This should never occur
+check(<<"-",Flag/binary>>) -> {flag,b2a(Flag)};
+check(_) -> arg.
 
 get_args([B|Bs], As) ->
     case check(B) of
@@ -1187,7 +1149,7 @@ get_args([B|Bs], As) ->
 	start_arg2 -> {reverse(As), [B|Bs]};
 	eval_arg -> {reverse(As), [B|Bs]};
 	end_args -> {reverse(As), Bs};
-	flag -> {reverse(As), [B|Bs]};
+	{flag,_} -> {reverse(As), [B|Bs]};
 	arg ->
 	    get_args(Bs, [B|As])
     end;
@@ -1199,44 +1161,28 @@ get_args([], As) -> {reverse(As),[]}.
 %%         atom() if a single arg was given.
 %%         list(atom()) if several args were given.
 %%
-get_flag(F,Flags,Default) ->
-    case catch get_flag(F,Flags) of
-	{'EXIT',_} ->
-	    Default;
-	Value ->
-	    Value
-    end.
-
-get_flag(F,Flags) ->
-    case search(F,Flags) of
-	{value,{F,[V]}} ->
+get_flag(F, Flags, Default) ->
+    case lists:keyfind(F, 1, Flags) of
+	{F,[]} ->
+	    true;
+	{F,[V]} ->
 	    V;
-	{value,{F,V}} ->
+	{F,V} ->
 	    V;
-	{value,{F}} -> % Flag given!
-	    true;
 	_ ->
-	    exit(list_to_atom(concat(["no ",F," flag"])))
+	    Default
     end.
 
 %%
 %% Internal get_flag function, with default value.
 %% Return: list(atom()) 
 %%
-get_flag_list(F,Flags,Default) ->
-    case catch get_flag_list(F,Flags) of
-	{'EXIT',_} ->
-	    Default;
-	Value ->
-	    Value
-    end.
-
-get_flag_list(F,Flags) ->
-    case search(F,Flags) of
-	{value,{F,V}} ->
+get_flag_list(F, Flags, Default) ->
+    case lists:keyfind(F, 1, Flags) of
+	{F,[_|_]=V} ->
 	    V;
 	_ ->
-	    exit(list_to_atom(concat(["no ",F," flag"])))
+	    Default
     end.
 
 %%
@@ -1246,21 +1192,15 @@ get_flag_list(F,Flags) ->
 %%
 get_flag_args(F,Flags) -> get_flag_args(F,Flags,[]).
 
-get_flag_args(F,[{F,V}|Flags],Acc) when is_list(V) ->
-    get_flag_args(F,Flags,[V|Acc]);
 get_flag_args(F,[{F,V}|Flags],Acc) ->
-    get_flag_args(F,Flags,[[V]|Acc]);
+    get_flag_args(F,Flags,[V|Acc]);
 get_flag_args(F,[_|Flags],Acc) ->
     get_flag_args(F,Flags,Acc);
 get_flag_args(_,[],Acc) ->
     reverse(Acc).
 
 get_arguments([{F,V}|Flags]) ->
-    [$-|Fl] = atom_to_list(F),
-    [{list_to_atom(Fl),to_strings(V)}|get_arguments(Flags)];
-get_arguments([{F}|Flags]) ->
-    [$-|Fl] = atom_to_list(F),
-    [{list_to_atom(Fl),[]}|get_arguments(Flags)];
+    [{F,to_strings(V)}|get_arguments(Flags)];
 get_arguments([]) ->
     [].
 
@@ -1268,44 +1208,26 @@ to_strings([H|T]) when is_atom(H) -> [atom_to_list(H)|to_strings(T)];
 to_strings([H|T]) when is_binary(H) -> [b2s(H)|to_strings(T)];
 to_strings([])    -> [].
 
-get_argument(Arg,Flags) ->
-    Args = get_arguments(Flags),
-    case get_argument1(Arg,Args) of
-	[] ->
-	    error;
-	Value ->
-	    {ok,Value}
+get_argument(Arg, Flags) ->
+    case get_argument1(Arg, Flags) of
+	[] -> error;
+	Value -> {ok,Value}
     end.
 
-get_argument1(Arg,[{Arg,V}|Args]) ->
-    [V|get_argument1(Arg,Args)];
-get_argument1(Arg,[_|Args]) ->
-    get_argument1(Arg,Args);
-get_argument1(_,[]) ->
+get_argument1(Arg, [{Arg,V}|Args]) ->
+    [to_strings(V)|get_argument1(Arg, Args)];
+get_argument1(Arg, [_|Args]) ->
+    get_argument1(Arg, Args);
+get_argument1(_, []) ->
     [].
 
 set_argument([{Flag,_}|Flags],Flag,Value) ->
     [{Flag,[Value]}|Flags];
-set_argument([{Flag}|Flags],Flag,Value) ->
-    [{Flag,[Value]}|Flags];
 set_argument([Item|Flags],Flag,Value) ->
     [Item|set_argument(Flags,Flag,Value)];
 set_argument([],Flag,Value) ->
     [{Flag,[Value]}].
 
-concat([A|T]) when is_atom(A) ->
-    atom_to_list(A) ++ concat(T);
-concat([C|T]) when is_integer(C), 0 =< C, C =< 255 ->
-    [C|concat(T)];
-concat([Bin|T]) when is_binary(Bin) ->
-    binary_to_list(Bin) ++ concat(T);
-concat([S|T]) ->
-    S ++ concat(T);
-concat([]) ->
-    [].
-
-append(L, Z) -> L ++ Z.
-
 append([E]) -> E;
 append([H|T]) ->
     H ++ append(T);
@@ -1320,13 +1242,6 @@ reverse([A, B]) ->
 reverse([A, B | L]) ->
     lists:reverse(L, [B, A]). % BIF
 			
-search(Key, [H|_T]) when is_tuple(H), element(1, H) =:= Key ->
-    {value, H};
-search(Key, [_|T]) ->
-    search(Key, T);
-search(_Key, []) ->
-    false.
-
 -spec objfile_extension() -> nonempty_string().
 objfile_extension() ->
     ".beam".
diff --git a/erts/preloaded/src/prim_file.erl b/erts/preloaded/src/prim_file.erl
index c87b2645ec..2eb1b1d408 100644
--- a/erts/preloaded/src/prim_file.erl
+++ b/erts/preloaded/src/prim_file.erl
@@ -1276,6 +1276,7 @@ lseek_position(_) ->
 %% Translates the response from the driver into 
 %% {ok, Result} or {error, Reason}.
 
+-dialyzer({no_improper_lists, translate_response/2}).
 translate_response(?FILE_RESP_OK, []) ->
     ok;
 translate_response(?FILE_RESP_ERROR, List) when is_list(List) ->
diff --git a/erts/preloaded/src/prim_inet.erl b/erts/preloaded/src/prim_inet.erl
index 4d04e1dacb..bd74831bb7 100644
--- a/erts/preloaded/src/prim_inet.erl
+++ b/erts/preloaded/src/prim_inet.erl
@@ -232,7 +232,7 @@ bindx(S, AddFlag, Addrs) ->
 %% if timeout is given:
 %%       timeout < 0  -> infinity
 %%                 0  -> immediate connect (mostly works for loopback)
-%%               > 0  -> wait for timout ms if not connected then 
+%%               > 0  -> wait for timeout ms if not connected then 
 %%                       return {error, timeout} 
 %%
 %% ASYNC_CONNECT(insock(), IP, Port, Timeout) -> {ok, S, Ref} | {error, Reason}
@@ -273,7 +273,7 @@ async_connect(S, IP, Port, Time) ->
 %% if timeout is given:
 %%       timeout < 0  -> infinity
 %%                 0  -> immediate accept (poll)
-%%               > 0  -> wait for timout ms for accept if no accept then 
+%%               > 0  -> wait for timeout ms for accept if no accept then 
 %%                       return {error, timeout}
 %%
 %% ASYNC_ACCEPT(insock(), Timeout)
@@ -1147,6 +1147,7 @@ enc_opt(packet_size)     -> ?INET_LOPT_PACKET_SIZE;
 enc_opt(read_packets)    -> ?INET_LOPT_READ_PACKETS;
 enc_opt(netns)           -> ?INET_LOPT_NETNS;
 enc_opt(show_econnreset) -> ?INET_LOPT_TCP_SHOW_ECONNRESET;
+enc_opt(line_delimiter)  -> ?INET_LOPT_LINE_DELIM;
 enc_opt(raw)             -> ?INET_OPT_RAW;
 % Names of SCTP opts:
 enc_opt(sctp_rtoinfo)	 	   -> ?SCTP_OPT_RTOINFO;
@@ -1205,6 +1206,7 @@ dec_opt(?INET_LOPT_PACKET_SIZE)      -> packet_size;
 dec_opt(?INET_LOPT_READ_PACKETS)     -> read_packets;
 dec_opt(?INET_LOPT_NETNS)           -> netns;
 dec_opt(?INET_LOPT_TCP_SHOW_ECONNRESET) -> show_econnreset;
+dec_opt(?INET_LOPT_LINE_DELIM)      -> line_delimiter;
 dec_opt(?INET_OPT_RAW)              -> raw;
 dec_opt(I) when is_integer(I)     -> undefined.
 
@@ -1287,6 +1289,7 @@ type_opt_1(packet) ->
 	   {httph_bin,?TCP_PB_HTTPH_BIN},
 	   {ssl, ?TCP_PB_SSL_TLS}, % obsolete
 	   {ssl_tls, ?TCP_PB_SSL_TLS}]};
+type_opt_1(line_delimiter)  -> int;
 type_opt_1(mode) ->
     {enum,[{list, ?INET_MODE_LIST},
 	   {binary, ?INET_MODE_BINARY}]};
diff --git a/erts/test/ethread_SUITE.erl b/erts/test/ethread_SUITE.erl
index 4a40dbb11e..388af66b23 100644
--- a/erts/test/ethread_SUITE.erl
+++ b/erts/test/ethread_SUITE.erl
@@ -123,38 +123,6 @@ try_lock_mutex(suite) ->
 try_lock_mutex(Config) ->
     run_case(Config, "try_lock_mutex", "").
 
-%% Remove dead code?
-
-% wd_dispatch(P) ->
-%     receive
-% 	bye ->
-% 	    ?line true = port_command(P, "-1 "),
-% 	    ?line bye;
-% 	L when is_list(L) ->
-% 	    ?line true = port_command(P, L),
-% 	    ?line wd_dispatch(P)
-%     end.
-% 
-% watchdog(Port) ->
-%     ?line process_flag(priority, max),
-%     ?line receive after 500 -> ok end,
-% 
-%     ?line random:seed(),
-%     ?line true = port_command(Port, "0 "),
-%     ?line lists:foreach(fun (T) ->
-% 				erlang:send_after(T,
-% 						  self(),
-% 						  integer_to_list(T)
-% 						  ++ " ")
-% 			end,
-% 			lists:usort(lists:map(fun (_) ->
-% 						      random:uniform(4500)+500
-% 					      end,
-% 					      lists:duplicate(50,0)))),
-%     ?line erlang:send_after(5100, self(), bye),
-% 
-%     wd_dispatch(Port).
-
 cond_wait(doc) ->
     ["Tests ethr_cond_wait with ethr_cond_signal and ethr_cond_broadcast."];
 cond_wait(suite) ->
diff --git a/erts/test/otp_SUITE.erl b/erts/test/otp_SUITE.erl
index 69a0d19719..87d620b180 100644
--- a/erts/test/otp_SUITE.erl
+++ b/erts/test/otp_SUITE.erl
@@ -290,7 +290,7 @@ call_to_deprecated(Config) when is_list(Config) ->
 call_to_size_1(Config) when is_list(Config) ->
     %% Applications that do not call erlang:size/1:
     Apps = [asn1,compiler,debugger,kernel,observer,parsetools,
-	    runtime_tools,stdlib,tools,webtool],
+	    runtime_tools,stdlib,tools],
     not_recommended_calls(Config, Apps, {erlang,size,1}).
 
 call_to_now_0(Config) when is_list(Config) ->
@@ -298,12 +298,14 @@ call_to_now_0(Config) when is_list(Config) ->
     Apps = [asn1,common_test,compiler,debugger,dialyzer,
 	    gs,kernel,mnesia,observer,parsetools,reltool,
 	    runtime_tools,sasl,stdlib,syntax_tools,
-	    test_server,tools,webtool],
+	    test_server,tools],
     not_recommended_calls(Config, Apps, {erlang,now,0}).
 
-not_recommended_calls(Config, Apps, MFA) ->
+not_recommended_calls(Config, Apps0, MFA) ->
     Server = ?config(xref_server, Config),
 
+    Apps = [App || App <- Apps0, is_present_application(App, Server)],
+
     Fs = [MFA],
 
     Q1 = io_lib:format("E || ~p : Fun", [Fs]),
@@ -337,11 +339,28 @@ not_recommended_calls(Config, Apps, MFA) ->
     end,
     case CallsToMFA of
 	[] -> 
-            ok;
+            SkippedApps = ordsets:subtract(ordsets:from_list(Apps0),
+                                           ordsets:from_list(Apps)),
+            case SkippedApps of
+                [] ->
+                    ok;
+                _ ->
+                    AppStrings = [atom_to_list(A) || A <- SkippedApps],
+                    Mess = io_lib:format("Application(s) not present: ~s\n",
+                                         [string:join(AppStrings, ", ")]),
+                    {comment, Mess}
+            end;
 	_ ->
 	    ?t:fail({length(CallsToMFA),calls_to_size_1})
     end.
 
+is_present_application(Name, Server) ->
+    Q = io_lib:format("~w : App", [Name]),
+    case xref:q(Server, lists:flatten(Q)) of
+        {ok,[Name]} -> true;
+        {error,_,_} -> false
+    end.
+
 strong_components(Config) when is_list(Config) ->
     Server = ?config(xref_server, Config),
     ?line {ok,Cs} = xref:q(Server, "components AE"),
diff --git a/erts/test/run_erl_SUITE.erl b/erts/test/run_erl_SUITE.erl
index 328477d870..6759d41a2b 100644
--- a/erts/test/run_erl_SUITE.erl
+++ b/erts/test/run_erl_SUITE.erl
@@ -141,12 +141,10 @@ heavier_1(Config) ->
 
     ?line ToErl = open_port({spawn,"to_erl "++Pipe}, []),
     io:format("ToErl = ~p\n", [ToErl]),
-    X = 1,
-    Y = 555,
-    Z = 42,
-    ?line random:seed(X, Y, Z),
-    SeedCmd = lists:flatten(io_lib:format("random:seed(~p, ~p, ~p). \r\n",
-					  [X,Y,Z])),
+    Seed = {1,555,42},
+    rand:seed(exsplus, Seed),
+    SeedCmd = lists:flatten(io_lib:format("rand:seed(exsplus, ~p). \r\n",
+					  [Seed])),
     ?line io:format("~p\n", [SeedCmd]),
     ?line erlang:port_command(ToErl, SeedCmd),
 
@@ -157,9 +155,9 @@ heavier_1(Config) ->
 	"F = fun(F,0) -> ok; "++
 	       "(F,N) -> " ++
 	           "io:format(\"\\\"~s\\\"~n\","++
-	                     "[[35|[random:uniform(25)+65 || " ++
+	                     "[[35|[rand:uniform(25)+65 || " ++
 	                     "_ <- lists:seq(1, "++
-	                                "random:uniform("++
+	                                "rand:uniform("++
                                              integer_to_list(MaxLen)++
                                         "))]]]), "++
 	           "F(F,N-1) "++
@@ -189,8 +187,8 @@ receive_all(Iter, ToErl, MaxLen) ->
 
 receive_all_1(0, _, _, _) -> ok;
 receive_all_1(Iter, Line, ToErl, MaxLen) ->
-    NumChars = random:uniform(MaxLen),
-    Pattern = [random:uniform(25)+65 || _ <- lists:seq(1, NumChars)],
+    NumChars = rand:uniform(MaxLen),
+    Pattern = [rand:uniform(25)+65 || _ <- lists:seq(1, NumChars)],
     receive_all_2(Iter, {NumChars,Pattern}, Line, ToErl, MaxLen).
     
 
diff --git a/erts/test/upgrade_SUITE.erl b/erts/test/upgrade_SUITE.erl
index 8a91cf5b7e..83cd2359d8 100644
--- a/erts/test/upgrade_SUITE.erl
+++ b/erts/test/upgrade_SUITE.erl
@@ -40,7 +40,7 @@
 %% - typer requires hipe (in the .app file)
 %% - erl_interface, jinterface support no upgrade
 -define(appup_exclude, 
-	[dialyzer,hipe,typer,erl_interface,jinterface]).
+	[dialyzer,hipe,typer,erl_interface,jinterface,ose]).
 
 init_per_suite(Config) ->
     %% Check that a real release is running, not e.g. cerl
diff --git a/erts/vsn.mk b/erts/vsn.mk
index 1012f5eafd..9e4248a668 100644
--- a/erts/vsn.mk
+++ b/erts/vsn.mk
@@ -18,7 +18,7 @@
 # %CopyrightEnd%
 # 
 
-VSN = 7.0
+VSN = 8.0
 
 # Port number 4365 in 4.2
 # Port number 4366 in 4.3