erts: Create forker process for spawn driver

Instead of forking from the beam process, we create a separate
process in which all forks are done. This has several advantages:

1) performance:
  * don't have to close all fd's in the world
  * fork only has to copy stuff from a small process
  * work is done in a completely seperate process
  * a 3x performance increase has been measured,
    can be made even greater (10x) if we cache the
    environment in child setup

2) stability
  * the exec is done in another process than beam, which means that
    if the file that we exec to is on an nfs that is not available
    right now we will not block a scheduler until the nfs returns.

3) simplicity
  * don't have to deal with SIGCHLD in the erts

Unfortunately, this solution also implies some badness.

1) There will always be a seperate process running together with
   beam on unix. This could be confusing and undesirable.
2) We have to transfer the entire environment to child_setup
   for each command.

OTP-13088

1 files changed, 12 insertions, 37 deletions

diff --git a/erts/configure.in b/erts/configure.in
index 96ad58f941..32d8bca98c 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)
 
@@ -1291,11 +1290,7 @@ 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
@@ -1305,16 +1300,6 @@ else
 		else
 			AC_MSG_RESULT(no)
 		fi
-		if test "x$ETHR_THR_LIB_BASE_TYPE" != "xposix_nptl"; then
-		    # Child waiter thread cannot be enabled
-		    disable_child_waiter_thread=yes
-		    enable_child_waiter_thread=no
-		fi
-		;;
-	    win32)
-		# Child waiter thread cannot be enabled
-		disable_child_waiter_thread=yes
-		enable_child_waiter_thread=no
 		;;
 	    *)
 		;;
@@ -1334,24 +1319,6 @@ else
 	    esac
 	done
 	EMU_THR_DEFS=$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)
@@ -1483,19 +1450,27 @@ 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...