Merge branch 'john/erts/efile-nif'

OTP-14256
OTP-14797

38 files changed, 5772 insertions, 9909 deletions

diff --git a/erts/doc/src/erl_nif.xml b/erts/doc/src/erl_nif.xml
index ef3cdb89e9..23f0c66429 100644
--- a/erts/doc/src/erl_nif.xml
+++ b/erts/doc/src/erl_nif.xml
@@ -367,6 +367,8 @@ return term;</code>
             <c>enif_ioq_deq()</c></seealso></item>
             <item><seealso marker="#enif_ioq_peek">
             <c>enif_ioq_peek()</c></seealso></item>
+            <item><seealso marker="#enif_ioq_peek_head">
+            <c>enif_ioq_peek_head()</c></seealso></item>
             <item><seealso marker="#enif_inspect_iovec">
             <c>enif_inspect_iovec()</c></seealso></item>
             <item><seealso marker="#enif_free_iovec">
@@ -1683,8 +1685,7 @@ enif_inspect_iovec(env, max_elements, term, &amp;tail, &amp;iovec);
       <fsummary>Peek inside the IO Queue</fsummary>
       <desc>
         <p>Get the I/O queue as a pointer to an array of <c>SysIOVec</c>s.
-          It also returns the number of elements in <c>iovlen</c>.
-          This is the only way to get data out of the queue.</p>
+           It also returns the number of elements in <c>iovlen</c>.</p>
         <p>Nothing is removed from the queue by this function, that must be done
           with <seealso marker="#enif_ioq_deq"><c>enif_ioq_deq</c></seealso>.</p>
          <p>The returned array is suitable to use with the Unix system
@@ -1693,6 +1694,21 @@ enif_inspect_iovec(env, max_elements, term, &amp;tail, &amp;iovec);
     </func>
 
     <func>
+      <name><ret>int</ret>
+      <nametext>enif_ioq_peek_head(ErlNifEnv *env, ErlNifIOQueue *q, size_t *size, ERL_NIF_TERM *bin_term)</nametext></name>
+      <fsummary>Peek the head of the IO Queue.</fsummary>
+      <desc>
+        <p>Get the head of the IO Queue as a binary term.</p>
+        <p>If <c>size</c> is not <c>NULL</c>, the size of the head is placed
+           there.</p>
+        <p>Nothing is removed from the queue by this function, that must be done
+           with <seealso marker="#enif_ioq_deq"><c>enif_ioq_deq</c></seealso>.</p>
+        <p>Returns <c>true</c> on success, or <c>false</c> if the queue is
+           empty.</p>
+      </desc>
+    </func>
+
+    <func>
       <name><ret>size_t</ret>
       <nametext>enif_ioq_size(ErlNifIOQueue *q)</nametext></name>
       <fsummary>Get the current size of the IO Queue</fsummary>
diff --git a/erts/emulator/Makefile.in b/erts/emulator/Makefile.in
index fdaf253fe5..15b65f1c71 100644
--- a/erts/emulator/Makefile.in
+++ b/erts/emulator/Makefile.in
@@ -634,6 +634,7 @@ GENERATE += $(TTF_DIR)/driver_tab.c
 PRELOAD_BEAM =		$(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_buffer.beam \
 			$(ERL_TOP)/erts/preloaded/ebin/prim_eval.beam \
 			$(ERL_TOP)/erts/preloaded/ebin/prim_inet.beam \
 			$(ERL_TOP)/erts/preloaded/ebin/prim_file.beam \
@@ -785,6 +786,9 @@ $(OBJDIR)/%.o: drivers/$(ERLANG_OSTYPE)/%.c
 $(OBJDIR)/%.o: nifs/common/%.c
 	$(V_CC) $(CFLAGS) -DLIBSCTP=$(LIBSCTP) $(INCLUDES) -Inifs/common -Inifs/$(ERLANG_OSTYPE) -c $< -o $@
 
+$(OBJDIR)/%.o: nifs/$(ERLANG_OSTYPE)/%.c
+	$(V_CC) $(CFLAGS) $(INCLUDES) -Inifs/common -Inifs/$(ERLANG_OSTYPE) -I../etc/$(ERLANG_OSTYPE) -c $< -o $@
+
 # ----------------------------------------------------------------------
 # Specials
 #
@@ -873,17 +877,17 @@ RUN_OBJS += \
 LTTNG_OBJS = $(OBJDIR)/erlang_lttng.o
 NIF_OBJS = \
 	   $(OBJDIR)/erl_tracer_nif.o \
+	   $(OBJDIR)/prim_buffer_nif.o \
+	   $(OBJDIR)/prim_file_nif.o \
 	   $(OBJDIR)/zlib_nif.o
 
 ifeq ($(TARGET),win32)
 DRV_OBJS = \
 	$(OBJDIR)/registry_drv.o \
-	$(OBJDIR)/efile_drv.o \
 	$(OBJDIR)/inet_drv.o \
 	$(OBJDIR)/ram_file_drv.o \
 	$(OBJDIR)/ttsl_drv.o
 OS_OBJS = \
-	$(OBJDIR)/win_efile.o \
 	$(OBJDIR)/win_con.o \
 	$(OBJDIR)/dll_sys.o \
 	$(OBJDIR)/driver_tab.o \
@@ -892,7 +896,8 @@ OS_OBJS = \
 	$(OBJDIR)/sys_time.o \
 	$(OBJDIR)/sys_interrupt.o \
 	$(OBJDIR)/sys_env.o \
-	$(OBJDIR)/dosmap.o
+	$(OBJDIR)/dosmap.o \
+	$(OBJDIR)/win_prim_file.o
 
 else
 OS_OBJS = \
@@ -900,14 +905,13 @@ OS_OBJS = \
 	$(OBJDIR)/sys_drivers.o \
 	$(OBJDIR)/sys_uds.o \
 	$(OBJDIR)/driver_tab.o \
-	$(OBJDIR)/unix_efile.o \
+	$(OBJDIR)/elib_memmove.o \
 	$(OBJDIR)/gzio.o \
-	$(OBJDIR)/elib_memmove.o
+	$(OBJDIR)/unix_prim_file.o
 
 	OS_OBJS += $(OBJDIR)/sys_float.o \
                    $(OBJDIR)/sys_time.o
 DRV_OBJS = \
-	$(OBJDIR)/efile_drv.o \
 	$(OBJDIR)/inet_drv.o \
 	$(OBJDIR)/ram_file_drv.o \
 	$(OBJDIR)/ttsl_drv.o
@@ -1135,6 +1139,7 @@ BEAM_SRC=$(wildcard beam/*.c)
 DRV_COMMON_SRC=$(wildcard drivers/common/*.c)
 DRV_OSTYPE_SRC=$(wildcard drivers/$(ERLANG_OSTYPE)/*.c)
 NIF_COMMON_SRC=$(wildcard nifs/common/*.c)
+NIF_OSTYPE_SRC=$(wildcard nifs/$(ERLANG_OSTYPE)/*.c)
 ALL_SYS_SRC=$(wildcard sys/$(ERLANG_OSTYPE)/*.c) $(wildcard sys/common/*.c)
 # We use $(shell ls) here instead of wildcard as $(wildcard ) resolved at
 # loadtime of the makefile and at that time these files are not generated yet.
@@ -1147,7 +1152,10 @@ ifeq ($(TARGET),win32)
 
 #DEP_CC=$(EMU_CC)
 DEP_CC=$(CC)
-DEP_FLAGS=-MM  $(subst -O2,,$(CFLAGS)) $(INCLUDES) -I../etc/win32 -Idrivers/common -Idrivers/$(ERLANG_OSTYPE)
+DEP_FLAGS=-MM  $(subst -O2,,$(CFLAGS)) $(INCLUDES) -I../etc/win32 \
+	-Idrivers/common -Idrivers/$(ERLANG_OSTYPE) \
+	-Inifs/common -Inifs/$(ERLANG_OSTYPE)
+
 # ifeq (@MIXED_CYGWIN_VC@,yes)
 # VC++ used for compiling. If __GNUC__ is defined we will include
 # other headers then when compiling which will result in faulty
@@ -1167,7 +1175,9 @@ MG_FLAG=-MG
 endif
 
 DEP_CC=$(CC)
-DEP_FLAGS=-MM $(MG_FLAG) $(CFLAGS) $(INCLUDES) -Inifs/common -Idrivers/common -Idrivers/$(ERLANG_OSTYPE)
+DEP_FLAGS=-MM $(MG_FLAG) $(CFLAGS) $(INCLUDES) \
+	-Idrivers/common -Idrivers/$(ERLANG_OSTYPE) \
+	-Inifs/common -Inifs/$(ERLANG_OSTYPE)
 SYS_SRC=$(ALL_SYS_SRC)
 endif
 
@@ -1198,6 +1208,8 @@ $(TTF_DIR)/depend.mk: $(TTF_DIR)/GENERATED $(PRELOAD_SRC)
 		| $(SED_DEPEND) >> $(TTF_DIR)/depend.mk
 	$(V_at)$(DEP_CC) $(DEP_FLAGS) $(NIF_COMMON_SRC) \
 		| $(SED_DEPEND) >> $(TTF_DIR)/depend.mk
+	$(V_at)$(DEP_CC) $(DEP_FLAGS) -I../etc/$(ERLANG_OSTYPE) $(NIF_OSTYPE_SRC) \
+		| $(SED_DEPEND) >> $(TTF_DIR)/depend.mk
 	$(V_at)$(DEP_CC) $(DEP_FLAGS) $(SYS_SRC) \
 		| $(SED_DEPEND) >> $(TTF_DIR)/depend.mk
 	$(V_at)$(DEP_CC) $(DEP_FLAGS) $(TARGET_SRC) \
diff --git a/erts/emulator/beam/erl_lock_check.c b/erts/emulator/beam/erl_lock_check.c
index 4cdef0200f..64950fc252 100644
--- a/erts/emulator/beam/erl_lock_check.c
+++ b/erts/emulator/beam/erl_lock_check.c
@@ -86,6 +86,10 @@ static erts_lc_lock_order_t erts_lock_order[] = {
     {	"hipe_mfait_lock",			NULL			},
 #endif
     {	"nodes_monitors",			NULL			},
+    {	"meta_name_tab",	         	"address"		},
+    {	"db_tab",				"address"		},
+    {	"db_tab_fix",				"address"		},
+    {	"db_hash_slot",				"address"		},
     {	"resource_monitors",			"address"	        },
     {   "driver_list",                          NULL                    },
     {	"proc_link",				"pid"			},
@@ -95,12 +99,8 @@ static erts_lc_lock_order_t erts_lock_order[] = {
     {	"dist_entry_links",			"address"		},
     {   "code_write_permission",                NULL                    },
     {	"purge_state",		      		NULL			},
-    {	"meta_name_tab",	         	"address"		},
-    {	"db_tab",				"address"		},
     {	"proc_status",				"pid"			},
     {	"proc_trace",				"pid"			},
-    {	"db_tab_fix",				"address"		},
-    {	"db_hash_slot",				"address"		},
     {	"node_table",				NULL			},
     {	"dist_table",				NULL			},
     {	"sys_tracers",				NULL			},
@@ -109,7 +109,6 @@ static erts_lc_lock_order_t erts_lock_order[] = {
     {	"fun_tab",				NULL			},
     {	"environ",				NULL			},
     {	"release_literal_areas",		NULL			},
-    {	"efile_drv",				"address"		},
     {	"drv_ev_state_grow",			NULL,   		},
     {	"drv_ev_state",				"address"		},
     {	"safe_hash",				"address"		},
@@ -170,9 +169,6 @@ static erts_lc_lock_order_t erts_lock_order[] = {
     {   "save_ops_lock",                        NULL                    },
 #endif
 #endif
-#ifdef	USE_VM_PROBES
-    {   "efile_drv dtrace mutex",               NULL                    },
-#endif
     {	"mtrace_buf",				NULL			},
     {	"os_monotonic_time",			NULL			},
     {	"erts_alloc_hard_debug",		NULL			},
diff --git a/erts/emulator/beam/erl_nif.c b/erts/emulator/beam/erl_nif.c
index dbcc894ac9..abb490cf9c 100644
--- a/erts/emulator/beam/erl_nif.c
+++ b/erts/emulator/beam/erl_nif.c
@@ -3605,6 +3605,55 @@ int enif_ioq_deq(ErlNifIOQueue *q, size_t elems, size_t *size)
     return 1;
 }
 
+int enif_ioq_peek_head(ErlNifEnv *env, ErlNifIOQueue *q, size_t *size, ERL_NIF_TERM *bin_term) {
+    SysIOVec *iov_entry;
+    Binary *ref_bin;
+
+    if (q->size == 0) {
+        return 0;
+    }
+
+    ASSERT(q->b_head != q->b_tail && q->v_head != q->v_tail);
+
+    ref_bin = &q->b_head[0]->nif;
+    iov_entry = &q->v_head[0];
+
+    if (size != NULL) {
+        *size = iov_entry->iov_len;
+    }
+
+    if (iov_entry->iov_len > ERL_ONHEAP_BIN_LIMIT) {
+        ProcBin *pb = (ProcBin*)alloc_heap(env, PROC_BIN_SIZE);
+
+        pb->thing_word = HEADER_PROC_BIN;
+        pb->next = MSO(env->proc).first;
+        pb->val = ref_bin;
+        pb->flags = 0;
+
+        ASSERT((byte*)iov_entry->iov_base >= (byte*)ref_bin->orig_bytes);
+        ASSERT(iov_entry->iov_len <= ref_bin->orig_size);
+
+        pb->bytes = (byte*)iov_entry->iov_base;
+        pb->size = iov_entry->iov_len;
+
+        MSO(env->proc).first = (struct erl_off_heap_header*) pb;
+        OH_OVERHEAD(&(MSO(env->proc)), pb->size / sizeof(Eterm));
+
+        erts_refc_inc(&ref_bin->intern.refc, 2);
+        *bin_term = make_binary(pb);
+    } else {
+        ErlHeapBin* hb = (ErlHeapBin*)alloc_heap(env, heap_bin_size(iov_entry->iov_len));
+
+        hb->thing_word = header_heap_bin(iov_entry->iov_len);
+        hb->size = iov_entry->iov_len;
+
+        sys_memcpy(hb->data, iov_entry->iov_base, iov_entry->iov_len);
+        *bin_term = make_binary(hb);
+    }
+
+    return 1;
+}
+
 SysIOVec *enif_ioq_peek(ErlNifIOQueue *q, int *iovlen)
 {
     return erts_ioq_peekq(q, iovlen);
diff --git a/erts/emulator/beam/erl_nif.h b/erts/emulator/beam/erl_nif.h
index 7fb447e4a8..053f7673c4 100644
--- a/erts/emulator/beam/erl_nif.h
+++ b/erts/emulator/beam/erl_nif.h
@@ -52,9 +52,10 @@
 ** 2.11: 19.0 enif_snprintf 
 ** 2.12: 20.0 add enif_select, enif_open_resource_type_x
 ** 2.13: 20.1 add enif_ioq
+** 2.14: 21.0 add enif_ioq_peek_head
 */
 #define ERL_NIF_MAJOR_VERSION 2
-#define ERL_NIF_MINOR_VERSION 13
+#define ERL_NIF_MINOR_VERSION 14
 
 /*
  * The emulator will refuse to load a nif-lib with a major version
diff --git a/erts/emulator/beam/erl_nif_api_funcs.h b/erts/emulator/beam/erl_nif_api_funcs.h
index 9e573307d8..3750fd9b68 100644
--- a/erts/emulator/beam/erl_nif_api_funcs.h
+++ b/erts/emulator/beam/erl_nif_api_funcs.h
@@ -198,6 +198,7 @@ ERL_NIF_API_FUNC_DECL(SysIOVec*,enif_ioq_peek,(ErlNifIOQueue *q, int *iovlen));
 ERL_NIF_API_FUNC_DECL(int,enif_inspect_iovec,(ErlNifEnv *env, size_t max_length, ERL_NIF_TERM iovec_term, ERL_NIF_TERM *tail, ErlNifIOVec **iovec));
 ERL_NIF_API_FUNC_DECL(void,enif_free_iovec,(ErlNifIOVec *iov));
 
+ERL_NIF_API_FUNC_DECL(int,enif_ioq_peek_head,(ErlNifEnv *env, ErlNifIOQueue *q, size_t *size, ERL_NIF_TERM *head));
 
 /*
 ** ADD NEW ENTRIES HERE (before this comment) !!!
@@ -373,6 +374,7 @@ ERL_NIF_API_FUNC_DECL(void,enif_free_iovec,(ErlNifIOVec *iov));
 #  define enif_ioq_peek ERL_NIF_API_FUNC_MACRO(enif_ioq_peek)
 #  define enif_inspect_iovec ERL_NIF_API_FUNC_MACRO(enif_inspect_iovec)
 #  define enif_free_iovec ERL_NIF_API_FUNC_MACRO(enif_free_iovec)
+#  define enif_ioq_peek_head ERL_NIF_API_FUNC_MACRO(enif_ioq_peek_head)
 
 /*
 ** ADD NEW ENTRIES HERE (before this comment)
diff --git a/erts/emulator/beam/erlang_dtrace.d b/erts/emulator/beam/erlang_dtrace.d
index 237889e0f5..d0b10e0306 100644
--- a/erts/emulator/beam/erlang_dtrace.d
+++ b/erts/emulator/beam/erlang_dtrace.d
@@ -634,72 +634,6 @@ provider erlang {
      */
     probe aio_pool__get(char *, int);
 
-    /* Probes for efile_drv.c */
-
-    /**
-     * Entry into the efile_drv.c file I/O driver
-     *
-     * For a list of command numbers used by this driver, see the section
-     * "Guide to efile_drv.c probe arguments" in ../../../HOWTO/DTRACE.md.
-     * That section also contains explanation of the various integer and
-     * string arguments that may be present when any particular probe fires.
-     *
-     * NOTE: Not all Linux platforms (using SystemTap) can support
-     *       arguments beyond arg9.
-     *
-     *
-     * TODO: Adding the port string, args[10], is a pain.  Making that
-     *       port string available to all the other efile_drv.c probes
-     *       will be more pain.  Is the pain worth it?  If yes, then
-     *       add them everywhere else and grit our teeth.  If no, then
-     *       rip it out.
-     *
-     * @param thread-id number of the scheduler Pthread                   arg0
-     * @param tag number: {thread-id, tag} uniquely names a driver operation
-     * @param user-tag string                                             arg2
-     * @param command number                                              arg3
-     * @param string argument 1                                           arg4
-     * @param string argument 2                                           arg5
-     * @param integer argument 1                                          arg6
-     * @param integer argument 2                                          arg7
-     * @param integer argument 3                                          arg8
-     * @param integer argument 4                                          arg9
-     * @param port the port ID of the busy port                       args[10]
-     */
-    probe efile_drv__entry(int, int, char *, int, char *, char *,
-                           int64_t, int64_t, int64_t, int64_t, char *);
-
-    /**
-     * Entry into the driver's internal work function.  Computation here
-     * is performed by a async worker pool Pthread.
-     *
-     * @param thread-id number
-     * @param tag number
-     * @param command number
-     */
-    probe efile_drv__int_entry(int, int, int);
-
-    /**
-     * Return from the driver's internal work function.
-     *
-     * @param thread-id number
-     * @param tag number
-     * @param command number
-     */
-    probe efile_drv__int_return(int, int, int);
-
-    /**
-     * Return from the efile_drv.c file I/O driver
-     *
-     * @param thread-id number                                            arg0
-     * @param tag number                                                  arg1
-     * @param user-tag string                                             arg2
-     * @param command number                                              arg3
-     * @param Success? 1 is success, 0 is failure                         arg4
-     * @param If failure, the errno of the error.                         arg5
-     */
-    probe efile_drv__return(int, int, char *, int, int, int);
-
 
 /*
  * The set of probes called by the erlang tracer nif backend. In order
diff --git a/erts/emulator/drivers/common/efile_drv.c b/erts/emulator/drivers/common/efile_drv.c
deleted file mode 100644
index 4e1d2f0d7f..0000000000
--- a/erts/emulator/drivers/common/efile_drv.c
+++ /dev/null
@@ -1,4295 +0,0 @@
-/*
- * %CopyrightBegin%
- *
- * Copyright Ericsson AB 1996-2017. 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.
- *
- * This file is generic, and does the work of decoding the commands
- * and encoding the responses.  System-specific functions are found in
- * the unix_efile.c and win_efile.c files.
- */
-
-/* Operations */
-
-#define FILE_OPEN		 1 /* Essential for startup */
-#define FILE_READ		 2
-#define FILE_LSEEK		 3
-#define FILE_WRITE		 4
-#define FILE_FSTAT		 5 /* Essential for startup */
-#define FILE_PWD                 6 /* Essential for startup */
-#define FILE_READDIR             7 /* Essential for startup */
-#define FILE_CHDIR               8
-#define FILE_FSYNC               9
-#define FILE_MKDIR              10
-#define FILE_DELETE             11
-#define FILE_RENAME             12
-#define FILE_RMDIR              13
-#define FILE_TRUNCATE           14
-#define FILE_READ_FILE          15 /* Essential for startup */
-#define FILE_WRITE_INFO		16
-#define FILE_LSTAT            	19
-#define FILE_READLINK        	20
-#define FILE_LINK             	21
-#define FILE_SYMLINK          	22
-#define FILE_CLOSE		23
-#define FILE_PWRITEV		24
-#define FILE_PREADV		25
-#define FILE_SETOPT		26
-#define FILE_IPREAD             27
-#define FILE_ALTNAME            28
-#define FILE_READ_LINE          29
-#define FILE_FDATASYNC          30
-#define FILE_FADVISE            31
-#define FILE_SENDFILE           32
-#define FILE_FALLOCATE          33
-#define FILE_CLOSE_ON_PORT_EXIT 34
-/* Return codes */
-
-#define FILE_RESP_OK         0
-#define FILE_RESP_ERROR      1
-#define FILE_RESP_DATA       2
-#define FILE_RESP_NUMBER     3
-#define FILE_RESP_INFO       4
-#define FILE_RESP_NUMERR     5
-#define FILE_RESP_LDATA      6
-#define FILE_RESP_N2DATA     7
-#define FILE_RESP_EOF        8
-#define FILE_RESP_FNAME      9
-#define FILE_RESP_ALL_DATA  10
-#define FILE_RESP_LFNAME    11
-
-/* Options */
-
-#define FILE_OPT_DELAYED_WRITE 0
-#define FILE_OPT_READ_AHEAD    1
-
-/* IPREAD variants */
-
-#define IPREAD_S32BU_P32BU 0
-
-/* Limits */
-
-#define FILE_SEGMENT_READ  (256*1024)
-#define FILE_SEGMENT_WRITE (256*1024)
-
-/* Internal */
-
-/* Set to 1 to test having read_ahead implicitly for read_line */ 
-#define ALWAYS_READ_LINE_AHEAD 0
-
-
-/* Must not be possible to get from malloc()! */
-#define FILE_FD_INVALID ((Sint)(-1))
-
-#ifdef HAVE_CONFIG_H
-#  include "config.h"
-#endif
-
-#include <ctype.h>
-#include <sys/types.h>
-#include <stdlib.h>
-
-/* Need (NON)BLOCKING macros for sendfile */
-#ifndef WANT_NONBLOCKING
-#define WANT_NONBLOCKING
-#endif
-
-#include "sys.h"
-
-#include "erl_driver.h"
-#include "erl_efile.h"
-#include "erl_threads.h"
-#include "gzio.h"
-#include "dtrace-wrapper.h" 
-
-
-static ErlDrvSysInfo sys_info;
-
-/* For explanation of this var, see comment for same var in erl_async.c */
-static unsigned gcc_optimizer_hack = 0;
-
-#ifdef  USE_VM_PROBES
-
-#define DTRACE_EFILE_BUFSIZ 128
-
-#define DTRACE_INVOKE_SETUP(op) \
-    do { DTRACE3(efile_drv_int_entry, d->sched_i1, d->sched_i2, op); } while (0)
-#define DTRACE_INVOKE_SETUP_BY_NAME(op) \
-    struct t_data *d = (struct t_data *) data ; \
-    DTRACE_INVOKE_SETUP(op)
-#define DTRACE_INVOKE_RETURN(op) \
-    do { DTRACE3(efile_drv_int_return, d->sched_i1, d->sched_i2, \
-                 op); } while (0) ; gcc_optimizer_hack++ ;
-
-/* Assign human-friendlier id numbers to scheduler & I/O worker threads */
-int             dt_driver_idnum = 0;
-int             dt_driver_io_worker_base = 5000;
-erts_mtx_t      dt_driver_mutex;
-pthread_key_t   dt_driver_key;
-
-typedef struct {
-    int         thread_num;
-    Uint64      tag;
-} dt_private;
-
-dt_private *get_dt_private(int);
-#else  /* USE_VM_PROBES */
-#define DTRACE_INVOKE_SETUP(op)            do {} while (0)
-#define DTRACE_INVOKE_SETUP_BY_NAME(op)    do {} while (0)
-#define DTRACE_INVOKE_RETURN(op)           do {} while (0)
-#endif  /* USE_VM_PROBES */
-
-/* #define TRACE 1 */
-#ifdef TRACE
-#    define TRACE_C(c) do { putchar(c); fflush(stdout); } while (0)
-#    define TRACE_S(s) do { fputs((s), stdout); fflush(stdout); } while (0)
-#    define TRACE_F(args) do { printf args ;fflush(stdout); } while (0)
-#else
-#    define TRACE_C(c) ((void)(0))
-#    define TRACE_S(s) ((void)(0))
-#    define TRACE_F(args) ((void)(0))
-#endif
-
-
-#define THRDS_AVAILABLE (sys_info.async_threads > 0)
-#ifdef HARDDEBUG /* HARDDEBUG in io.c is expected too */
-#define TRACE_DRIVER fprintf(stderr, "Efile: ")
-#else
-#define TRACE_DRIVER
-#endif
-#define MUTEX_INIT(m, p) do { IF_THRDS { TRACE_DRIVER; (m = driver_pdl_create(p)); } } while (0)
-#define MUTEX_LOCK(m)    do { IF_THRDS { TRACE_DRIVER; driver_pdl_lock(m);   } } while (0)
-#define MUTEX_UNLOCK(m)  do { IF_THRDS { TRACE_DRIVER; driver_pdl_unlock(m); } } while (0)
-#define IF_THRDS if (THRDS_AVAILABLE)
-
-
-#define SENDFILE_FLGS_USE_THREADS (1 << 0)
-/**
- * On DARWIN sendfile can deadlock with close if called in
- * different threads. So until Apple fixes so that sendfile
- * is not buggy we disable usage of the async pool for
- * DARWIN. The testcase t_sendfile_crashduring reproduces
- * this error when using +A 10 and enabling SENDFILE_FLGS_USE_THREADS.
- */
-#if defined(__APPLE__) && defined(__MACH__)
-#define USE_THRDS_FOR_SENDFILE(DATA) 0
-#else
-#define USE_THRDS_FOR_SENDFILE(DATA) (DATA->flags & SENDFILE_FLGS_USE_THREADS)
-#endif /* defined(__APPLE__) && defined(__MACH__) */
-
-
-
-#if 0
-/* Experimental, for forcing all file operations to use the same thread. */
-   static unsigned file_fixed_key = 1;
-#  define KEY(desc) (&file_fixed_key)
-#else
-#  define KEY(desc) (&(desc)->key)
-#endif
-
-#ifndef MAX
-#  define MAX(x, y) (((x) > (y)) ? (x) : (y))
-#endif
-
-#ifdef FILENAMES_16BIT
-#ifdef USE_VM_PROBES
-#error 16bit characters in filenames and dtrace in combination is not supported.
-#endif
-#  define FILENAME_BYTELEN(Str) filename_len_16bit(Str)
-#  define FILENAME_COPY(To,From) filename_cpy_16bit((To),(From)) 
-#  define FILENAME_CHARSIZE 2
-
-   static int filename_len_16bit(char *str) 
-   {
-       char *p = str;
-       while(*p != '\0' || p[1] != '\0') {
-	   p += 2;
-       }
-       return (p - str);
-   }
-
-   static void filename_cpy_16bit(char *to, char *from) 
-   {
-       while(*from != '\0' || from[1] != '\0') {
-	   *to++ = *from++;
-	   *to++ = *from++;
-       }
-       *to++ = *from++;
-       *to++ = *from++;
-   }
-
-#else
-#  define FILENAME_BYTELEN(Str) strlen(Str)
-#  define FILENAME_COPY(To,From) strcpy(To,From) 
-#  define FILENAME_CHARSIZE 1
-#endif
-
-#if     (MAXPATHLEN+1)*FILENAME_CHARSIZE+1 > BUFSIZ
-#  define    RESBUFSIZE  ((MAXPATHLEN+1)*FILENAME_CHARSIZE+1)
-#else
-#  define    RESBUFSIZE  BUFSIZ
-#endif
-
-#define READDIR_CHUNKS (5)
-
-
-
-#if ALWAYS_READ_LINE_AHEAD
-#define DEFAULT_LINEBUF_SIZE 2048
-#else
-#define DEFAULT_LINEBUF_SIZE 512 /* Small, it's usually discarded anyway */ 
-#endif
-
-typedef unsigned char uchar;
-
-static ErlDrvData file_start(ErlDrvPort port, char* command);
-static int file_init(void);
-static void file_stop(ErlDrvData);
-static void file_output(ErlDrvData, char* buf, ErlDrvSizeT len);
-static ErlDrvSSizeT file_control(ErlDrvData, unsigned int command,
-				 char* buf, ErlDrvSizeT len,
-				 char **rbuf, ErlDrvSizeT rlen);
-static void file_timeout(ErlDrvData);
-static void file_outputv(ErlDrvData, ErlIOVec*);
-static void file_async_ready(ErlDrvData, ErlDrvThreadData);
-static void file_flush(ErlDrvData);
-
-#ifdef HAVE_SENDFILE
-static void file_ready_output(ErlDrvData data, ErlDrvEvent event);
-static void file_stop_select(ErlDrvEvent event, void* _);
-#endif /* HAVE_SENDFILE */
-
-
-enum e_timer {timer_idle, timer_again, timer_write};
-#ifdef HAVE_SENDFILE
-enum e_sendfile {sending, not_sending};
-#define SENDFILE_USE_THREADS (1 << 0)
-#endif /* HAVE_SENDFILE */
-
-struct t_data;
-
-typedef struct {
-    SWord           fd;
-    ErlDrvPort      port;
-    unsigned int    key;      /* Async queue key */
-    unsigned        flags;    /* Original flags from FILE_OPEN. */
-    void          (*invoke)(void *);
-    struct t_data  *d;
-    void          (*free)(void *);
-    struct t_data  *cq_head;  /* Queue of incoming commands */
-    struct t_data  *cq_tail;  /* -""- */
-    enum e_timer    timer_state;
-#ifdef HAVE_SENDFILE
-    enum e_sendfile sendfile_state;
-#endif /* HAVE_SENDFILE */
-    size_t          read_bufsize;
-    ErlDrvBinary   *read_binp;
-    size_t          read_offset;
-    size_t          read_size;
-    size_t          write_bufsize;
-    unsigned long   write_delay;
-    int             write_error;
-    Efile_error     write_errInfo;
-    ErlDrvPDL       q_mtx;    /* Mutex for the driver queue, known by the emulator. Also used for
-				 mutual exclusion when accessing field(s) below. */
-    size_t          write_buffered;
-#ifdef USE_VM_PROBES
-    int             idnum;      /* Unique ID # for this driver thread/desc */
-    char            port_str[DTRACE_TERM_BUF_SIZE];
-#endif
-} file_descriptor;
-
-
-static int reply_error(file_descriptor*, Efile_error* errInfo);
-
-struct erl_drv_entry efile_driver_entry = {
-    file_init,
-    file_start,
-    file_stop,
-    file_output,
-    NULL,
-#ifdef HAVE_SENDFILE
-    file_ready_output,
-#else
-    NULL,
-#endif /* HAVE_SENDFILE */
-    "efile",
-    NULL,
-    NULL,
-    file_control,
-    file_timeout,
-    file_outputv,
-    file_async_ready,
-    file_flush,
-    NULL,
-    NULL,
-    ERL_DRV_EXTENDED_MARKER,
-    ERL_DRV_EXTENDED_MAJOR_VERSION,
-    ERL_DRV_EXTENDED_MINOR_VERSION,
-    ERL_DRV_FLAG_USE_PORT_LOCKING,
-    NULL,
-    NULL,
-#ifdef HAVE_SENDFILE
-    file_stop_select
-#else
-    NULL
-#endif /* HAVE_SENDFILE */
-};
-
-
-
-static int thread_short_circuit;
-
-#define DRIVER_ASYNC(level, desc, f_invoke, data, f_free) \
-if (thread_short_circuit >= (level)) { \
-    (*(f_invoke))(data); \
-    file_async_ready((ErlDrvData)(desc), (data)); \
-} else { \
-    driver_async((desc)->port, KEY(desc), (f_invoke), (data), (f_free)); \
-}
-
-
-
-struct t_pbuf_spec {
-    Sint64 offset;
-    size_t size;
-};
-
-struct t_pwritev {
-    ErlDrvPort         port;
-    ErlDrvPDL          q_mtx;
-    size_t             size;
-    unsigned           cnt;
-    unsigned           n;
-    struct t_pbuf_spec specs[1];
-};
-
-struct t_preadv {
-    ErlIOVec eiov;
-    unsigned n;
-    unsigned cnt;
-    size_t   size;
-    Sint64   offsets[1];
-};
-
-#define READDIR_BUFSIZE (8*1024)*READDIR_CHUNKS
-#if READDIR_BUFSIZE < (1 + (2 + MAXPATHLEN)*FILENAME_CHARSIZE*READDIR_CHUNKS)
-#  undef READDIR_BUFSIZE
-#  define READDIR_BUFSIZE (1 + (2 + MAXPATHLEN)*FILENAME_CHARSIZE*READDIR_CHUNKS)
-#endif
-
-struct t_readdir_buf {
-     struct t_readdir_buf *next;
-     size_t n;
-     char buf[READDIR_BUFSIZE];
-};
-
-struct t_data
-{
-    struct t_data *next;
-    int            command;
-    int            level;
-    void         (*invoke)(void *);
-    void         (*free)(void *);
-    void           *data_to_free; /* used by FILE_CLOSE_ON_PORT_EXIT only */
-    int            again;
-    int            reply;
-#ifdef  USE_VM_PROBES
-    int               sched_i1;
-    Uint64            sched_i2;
-    char              sched_utag[DTRACE_EFILE_BUFSIZ+1];
-#endif
-    int            result_ok;
-    Efile_error    errInfo;
-    int            flags;
-    SWord          fd;
-    int            is_fd_unused;
-    /**/
-    Efile_info        info;
-    EFILE_DIR_HANDLE  dir_handle; /* Handle to open directory. */
-    ErlDrvBinary     *bin;
-    int               drive;
-    size_t            n;
-    /*off_t             offset;*/
-    /*size_t            bytesRead; Bytes read from the file. */
-    /**/
-    union {
-	struct {
-	    Sint64 offset;
-	    int    origin;
-	    Sint64 location;
-	} lseek;
-	struct {
-	    ErlDrvPort    port;
-	    ErlDrvPDL     q_mtx;
-	    size_t        size;
-	    size_t        reply_size;
-	} writev;
-	struct t_pwritev pwritev;
-	struct t_preadv  preadv;
-	struct {
-	    ErlDrvBinary *binp;
-	    size_t        bin_offset;
-	    size_t        bin_size;
-	    size_t        size;
-	} read;
-	struct {
-	    ErlDrvBinary *binp; /* in - out */
-	    size_t        read_offset; /* in - out */
-	    size_t        read_size; /* in - out */
-	    size_t        nl_pos; /* out */
-	    short         nl_skip; /* out, 0 or 1 */
-#if !ALWAYS_READ_LINE_AHEAD
-	    short         read_ahead; /* in, bool */
-#endif
-	} read_line;
-	struct {
-	    ErlDrvBinary *binp;
-	    int           size;
-	    int           offset;
-	} read_file;
-	struct {
-	    struct t_readdir_buf *first_buf;
-	    struct t_readdir_buf *last_buf;
-	} read_dir;
-	struct {
-	    Sint64 offset;
-	    Sint64 length;
-	    int advise;
-	} fadvise;
-#ifdef HAVE_SENDFILE
-	struct {
-	    ErlDrvPort port;
-	    ErlDrvPDL q_mtx;
-	    int out_fd;
-	    off_t offset;
-	    Uint64 nbytes;
-	    Uint64 written;
-	} sendfile;
-#endif /* HAVE_SENDFILE */
-	struct {
-	    Sint64 offset;
-	    Sint64 length;
-	} fallocate;
-    } c;
-    char b[1];
-};
-
-#define EF_ALLOC(S)		driver_alloc((S))
-#define EF_REALLOC(P, S)	driver_realloc((P), (S))
-#define EF_SAFE_ALLOC(S)	ef_safe_alloc((S))
-#define EF_SAFE_REALLOC(P, S)	ef_safe_realloc((P), (S))
-#define EF_FREE(P)		do { if((P)) driver_free((P)); } while(0)
-
-static void *ef_safe_alloc(Uint s)
-{
-    void *p = EF_ALLOC(s);
-    if (!p) erts_exit(ERTS_ERROR_EXIT, "efile drv: Can't allocate %lu bytes of memory\n", (unsigned long)s);
-    return p;
-}
-
-/*********************************************************************
- * ErlIOVec manipulation functions.
- */
-
-/* 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) efile_ev_get_char(ev, p ,pp, qp)
-static int
-efile_ev_get_char(ErlIOVec *ev, char *p, size_t *pp, size_t *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 += 1;
-	else {
-	    *qp += 1;
-	    *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) efile_ev_get_uint32(ev, p, pp, qp)
-static int
-efile_ev_get_uint32(ErlIOVec *ev, Uint32 *p, size_t *pp, size_t *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 += 4;
-	else {
-	    *qp += 1;
-	    *pp = 0;
-	}
-	return !0;
-    }
-    return 0;
-}
-
-/* Uint64 EV_UINT64(ErlIOVec *ev, int p, int q)*/
-#define EV_UINT64(ev, p, q)						\
-    ((Uint64) ((unsigned char *)(ev)->iov[q].iov_base)[p])
-
-/* int EV_GET_UINT64(ErlIOVec *ev, Uint64 *p, int *pp, int *qp) */
-#define EV_GET_UINT64(ev, p, pp, qp) efile_ev_get_uint64(ev, p, pp, qp)
-static int
-efile_ev_get_uint64(ErlIOVec *ev, Uint64 *p, size_t *pp, size_t *qp) {
-    if (*pp + 8 <= ev->iov[*qp].iov_len) {
-	*p = (EV_UINT64(ev, *pp, *qp) << 56)
-	    | (EV_UINT64(ev, *pp + 1, *qp) << 48)
-	    | (EV_UINT64(ev, *pp + 2, *qp) << 40)
-	    | (EV_UINT64(ev, *pp + 3, *qp) << 32)
-	    | (EV_UINT64(ev, *pp + 4, *qp) << 24)
-	    | (EV_UINT64(ev, *pp + 5, *qp) << 16)
-	    | (EV_UINT64(ev, *pp + 6, *qp) << 8)
-	    | (EV_UINT64(ev, *pp + 7, *qp));
-	if (*pp + 8 < ev->iov[*qp].iov_len)
-	    *pp += 8;
-	else {
-	    *qp += 1;
-	    *pp = 0;
-	}
-	return !0;
-    }
-    return 0;
-}
-
-/* int EV_GET_SINT64(ErlIOVec *ev, Uint64 *p, int *pp, int *qp) */
-#define EV_GET_SINT64(ev, p, pp, qp) efile_ev_get_sint64(ev, p, pp, qp)
-static int
-efile_ev_get_sint64(ErlIOVec *ev, Sint64 *p, size_t *pp, size_t *qp) {
-    Uint64 *tmp = (Uint64*)p;
-    return EV_GET_UINT64(ev, tmp, pp, qp);
-}
-
-#if 0
-
-static void ev_clear(ErlIOVec *ev) {
-    ASSERT(ev);
-    ev->size = 0;
-    ev->vsize = 0;
-    ev->iov = NULL;
-    ev->binv = NULL;
-}
-
-/* Assumes that ->iov and ->binv were allocated with sys_alloc().
- */
-static void ev_free(ErlIOVec *ev) {
-    if (! ev) {
-	return;
-    }
-    if (ev->vsize > 0) {
-	int i;
-	ASSERT(ev->iov);
-	ASSERT(ev->binv);
-	for (i = 0; i < ev->vsize; i++) {
-	    if (ev->binv[i]) {
-		driver_free_binary(ev->binv[i]);
-	    }
-	}
-	EF_FREE(ev->iov);
-	EF_FREE(ev->binv);
-    }
-}
-
-/* Copy the contents from source to dest.
- * Data in binaries is not copied, just the pointers; 
- * and refc is incremented.
- */
-static ErlIOVec *ev_copy(ErlIOVec *dest, ErlIOVec *source) {
-    int *ip;
-    ASSERT(dest);
-    ASSERT(source);
-    if (source->vsize == 0) {
-	/* Empty source */
-	ev_clear(dest);
-	return dest;
-    }
-    /* Allocate ->iov and ->binv */
-    dest->iov = EF_ALLOC(sizeof(*dest->iov) * source->vsize);
-    if (! dest->iov) {
-	return NULL;
-    }
-    dest->binv = EF_ALLOC(sizeof(*dest->binv) * source->vsize);
-    if (! dest->binv) {
-	EF_FREE(dest->iov);
-	return NULL;
-    }
-    dest->size = source->size;
-    /* Copy one vector element at the time. 
-     * Use *ip as an alias for dest->vsize to improve readabiliy.
-     * Keep dest consistent in every iteration by using 
-     * dest->vsize==*ip as loop variable.
-     */
-    for (ip = &dest->vsize, *ip = 0;  *ip < source->vsize;  (*ip)++) {
-	if (source->iov[*ip].iov_len == 0) {
-	    /* Empty vector element */
-	    dest->iov[*ip].iov_len = 0;
-	    dest->iov[*ip].iov_base = NULL;
-	    dest->binv[*ip] = NULL;
-	} else {
-	    /* Non empty vector element */
-	    if (source->binv[*ip]) {
-		/* Contents in binary - copy pointers and increment refc */
-		dest->iov[*ip] = source->iov[*ip];
-		dest->binv[*ip] = source->binv[*ip];
-		driver_binary_inc_refc(source->binv[*ip]);
-	    } else {
-		/* Contents not in binary - allocate new binary and copy data */
-		if (! (dest->binv[*ip] = 
-		       driver_alloc_binary(source->iov[*ip].iov_len))) {
-		    goto failed;
-		}
-		sys_memcpy(dest->binv[*ip]->orig_bytes,
-			   source->iov[*ip].iov_base,
-			   source->iov[*ip].iov_len);
-		dest->iov[*ip].iov_base = dest->binv[*ip]->orig_bytes;
-		dest->iov[*ip].iov_len = source->iov[*ip].iov_len;
-	    }
-	}
-    }
-    return dest;
- failed:
-    ev_free(dest);
-    return NULL;
-}
-
-#endif
-
-
-
-/*********************************************************************
- * Command queue functions
- */
-
-static void cq_enq(file_descriptor *desc, struct t_data *d) {
-    ASSERT(d);
-    if (desc->cq_head) {
-	ASSERT(desc->cq_tail);
-	ASSERT(!desc->cq_tail->next);
-	desc->cq_tail = desc->cq_tail->next = d;
-    } else {
-	ASSERT(desc->cq_tail == NULL);
-	desc->cq_head = desc->cq_tail = d;
-    }
-    d->next = NULL;
-}
-
-static struct t_data *cq_deq(file_descriptor *desc) {
-    struct t_data *d = desc->cq_head;
-    ASSERT(d || (!d && !desc->cq_tail));
-    if (d) {
-	ASSERT(!d->next || (d->next && desc->cq_tail != d));
-	if ((desc->cq_head = d->next) == NULL) {
-	    ASSERT(desc->cq_tail == d);
-	    desc->cq_tail = NULL;
-	}
-    }	
-    return d;
-}
-
-
-/*********************************************************************
- * Driver entry point -> init
- */
-static int 
-file_init(void)
-{
-    char buf[21]; /* enough to hold any 64-bit integer */
-    size_t bufsz = sizeof(buf);
-    thread_short_circuit = (erl_drv_getenv("ERL_EFILE_THREAD_SHORT_CIRCUIT",
-					   buf,
-					   &bufsz) == 0
-			    ? atoi(buf)
-			    : 0);
-    driver_system_info(&sys_info, sizeof(ErlDrvSysInfo));
-
-    /* run initiation of efile_driver if needed */
-    efile_init();
-
-#ifdef  USE_VM_PROBES
-    erts_mtx_init(&dt_driver_mutex, "efile_drv dtrace mutex", NIL,
-        ERTS_LOCK_FLAGS_PROPERTY_STATIC | ERTS_LOCK_FLAGS_CATEGORY_IO);
-    pthread_key_create(&dt_driver_key, NULL);
-#endif  /* USE_VM_PROBES */
-
-    return 0;
-}
-
-
-/*********************************************************************
- * Driver entry point -> start
- */
-static ErlDrvData 
-file_start(ErlDrvPort port, char* command) 
-
-{
-    file_descriptor* desc;
-
-    if ((desc = (file_descriptor*) EF_ALLOC(sizeof(file_descriptor)))
-	== NULL) {
-	errno = ENOMEM;
-	return ERL_DRV_ERROR_ERRNO;
-    }
-    desc->fd = FILE_FD_INVALID;
-    desc->port = port;
-    desc->key = driver_async_port_key(port);
-    desc->flags = 0;
-    desc->invoke = NULL;
-    desc->d = NULL;
-    desc->free = NULL;
-    desc->cq_head = NULL;
-    desc->cq_tail = NULL;
-    desc->timer_state = timer_idle;
-#ifdef HAVE_SENDFILE
-    desc->sendfile_state = not_sending;
-#endif
-    desc->read_bufsize = 0;
-    desc->read_binp = NULL;
-    desc->read_offset = 0;
-    desc->read_size = 0;
-    desc->write_delay = 0L;
-    desc->write_bufsize = 0;
-    desc->write_error = 0;
-    MUTEX_INIT(desc->q_mtx, port); /* Refc is one, referenced by emulator now */
-    desc->write_buffered = 0;
-#ifdef  USE_VM_PROBES
-    dtrace_drvport_str(port, desc->port_str);
-    get_dt_private(0);           /* throw away return value */
-#endif  /* USE_VM_PROBES */
-    return (ErlDrvData) desc;
-}
-
-static void do_close(int flags, SWord fd) {
-    if (flags & EFILE_COMPRESSED) {
-	erts_gzclose((ErtsGzFile)(fd));
-    } else {
-	efile_closefile((int) fd);
-    }
-}
-
-static void invoke_close(void *data)
-{
-    struct t_data *d = (struct t_data *) data;
-    DTRACE_INVOKE_SETUP(FILE_CLOSE);
-    d->again = 0;
-    do_close(d->flags, d->fd);
-    DTRACE_INVOKE_RETURN(FILE_CLOSE);
-}
-
-static void free_data(void *data)
-{
-    struct t_data *d = (struct t_data *) data;
-
-    switch (d->command) {
-    case FILE_OPEN:
-        if (d->is_fd_unused && d->fd != FILE_FD_INVALID) {
-            /* This is OK to do in scheduler thread because there can be no async op
-               ongoing for this fd here, as we exited during async open.
-               Ideally, this close should happen in an async thread too, but that would
-               require a substantial rewrite, as we are here because of a dead port and
-               cannot schedule async jobs for that port any more... */
-            do_close(d->flags, d->fd);
-        }
-        break;
-    case FILE_CLOSE_ON_PORT_EXIT:
-        EF_FREE(d->data_to_free);
-        break;
-    }
-
-    EF_FREE(data);
-}
-
-
-/*
- * Sends back an error reply to Erlang.
- */
-
-static void reply_posix_error(file_descriptor *desc, int posix_errno) {
-    char response[256];		/* Response buffer. */
-    char* s;
-    char* t;
-    
-    /*
-     * Contents of buffer sent back:
-     *
-     * +-----------------------------------------+
-     * | FILE_RESP_ERROR | Posix error id string |
-     * +-----------------------------------------+
-     */
-
-    TRACE_C('E');
-
-    response[0] = FILE_RESP_ERROR;
-    for (s = erl_errno_id(posix_errno), t = response+1; *s; s++, t++)
-	*t = tolower(*s);
-    driver_output2(desc->port, response, t-response, NULL, 0);
-}
-
-static void reply_Uint_posix_error(file_descriptor *desc, Uint num, 
-				   int posix_errno) {
-    char response[256];		/* Response buffer. */
-    char* s;
-    char* t;
-    
-    /*
-     * Contents of buffer sent back:
-     *
-     * +----------------------------------------------------------------------+
-     * | FILE_RESP_NUMERR | 64-bit number (big-endian) | Posix error id string |
-     * +----------------------------------------------------------------------+
-     */
-
-    TRACE_C('N');
-
-    response[0] = FILE_RESP_NUMERR;
-#if SIZEOF_VOID_P == 4
-    put_int32(0, response+1);
-#else
-    put_int32(num>>32, response+1);
-#endif
-    put_int32((Uint32)num, response+1+4);
-    for (s = erl_errno_id(posix_errno), t = response+1+4+4; *s; s++, t++)
-	*t = tolower(*s);
-    driver_output2(desc->port, response, t-response, NULL, 0);
-}
-
-#ifdef HAVE_SENDFILE
-static void reply_string_error(file_descriptor *desc, char* str) {
-    char response[256];		/* Response buffer. */
-    char* s;
-    char* t;
-
-    response[0] = FILE_RESP_ERROR;
-    for (s = str, t = response+1; *s; s++, t++)
-	*t = tolower(*s);
-    driver_output2(desc->port, response, t-response, NULL, 0);
-}
-#endif
-
-static int reply_error(file_descriptor *desc, 
-		       Efile_error *errInfo) /* The error codes. */
-{
-    reply_posix_error(desc, errInfo->posix_errno);
-    return 0;
-}
-
-static int reply_Uint_error(file_descriptor *desc, Uint num, 
-			    Efile_error *errInfo) /* The error codes. */
-{
-    reply_Uint_posix_error(desc, num, errInfo->posix_errno);
-    return 0;
-}
-
-static int reply_ok(file_descriptor *desc) {
-    char c = FILE_RESP_OK;
-
-    driver_output2(desc->port, &c, 1, NULL, 0);
-    return 0;
-}
-
-static int reply(file_descriptor *desc, int ok, Efile_error *errInfo) {
-    if (!ok) {
-	reply_error(desc, errInfo);
-    } else {
-	TRACE_C('K');
-	reply_ok(desc);
-    }
-    return 0;
-}
-
-static int reply_Uint(file_descriptor *desc, Uint result) {
-    char tmp[1+4+4];
-
-    /*
-     * Contents of buffer sent back:
-     *
-     * +-----------------------------------------------+
-     * | FILE_RESP_NUMBER | 64-bit number (big-endian) |
-     * +-----------------------------------------------+
-     */
-
-    TRACE_C('R');
-
-    tmp[0] = FILE_RESP_NUMBER;
-#if SIZEOF_VOID_P == 4
-    put_int32(0, tmp+1);
-#else
-    put_int32(result>>32, tmp+1);
-#endif
-    put_int32((Uint32)result, tmp+1+4);
-    driver_output2(desc->port, tmp, sizeof(tmp), NULL, 0);
-    return 0;
-}
-
-static int reply_Sint64(file_descriptor *desc, Sint64 result) {
-    char tmp[1+4+4];
-
-    /*
-     * Contents of buffer sent back:
-     *
-     * +-----------------------------------------------+
-     * | FILE_RESP_NUMBER | 64-bit number (big-endian) |
-     * +-----------------------------------------------+
-     */
-
-    TRACE_C('R');
-
-    tmp[0] = FILE_RESP_NUMBER;
-    put_int64(result, tmp+1);
-    driver_output2(desc->port, tmp, sizeof(tmp), NULL, 0);
-    return 0;
-}
-
-#if 0
-static void reply_again(file_descriptor *desc) {
-    char tmp[1];
-    tmp[0] = FILE_RESP_AGAIN;
-    driver_output2(desc->port, tmp, sizeof(tmp), NULL, 0);
-}
-#endif
-
-static void reply_ev(file_descriptor *desc, char response, ErlIOVec *ev) {
-    char tmp[1];
-    /* Data arriving at the Erlang process:
-     * [Response, Binary0, Binary1, .... | BinaryN-1]
-     */
-    tmp[0] = response;
-    driver_outputv(desc->port, tmp, sizeof(tmp), ev, 0);
-}
-
-static void reply_data(file_descriptor *desc, 
-		       ErlDrvBinary *binp, size_t offset, size_t len) {
-    char header[1+4+4];
-    /* Data arriving at the Erlang process:
-     * [?FILE_RESP_DATA, 64-bit length (big-endian) | Data]
-     */
-    header[0] = FILE_RESP_DATA;
-#if SIZEOF_SIZE_T == 4
-    put_int32(0, header+1);
-#else
-    put_int32(len>>32, header+1);
-#endif
-    put_int32((Uint32)len, header+1+4);
-    driver_output_binary(desc->port, header, sizeof(header),
-			 binp, offset, len);
-}
-
-static void reply_buf(file_descriptor *desc, char *buf, size_t len) {
-    char header[1+4+4];
-    /* Data arriving at the Erlang process:
-     * [?FILE_RESP_DATA, 64-bit length (big-endian) | Data]
-     */
-    header[0] = FILE_RESP_DATA;
-#if SIZEOF_SIZE_T == 4
-    put_int32(0, header+1);
-#else
-    put_int32(len>>32, header+1);
-#endif
-    put_int32((Uint32)len, header+1+4);
-    driver_output2(desc->port, header, sizeof(header), buf, len);
-}
-
-static int reply_eof(file_descriptor *desc) {
-    char c = FILE_RESP_EOF;
-
-    driver_output2(desc->port, &c, 1, NULL, 0);
-    return 0;
-}
- 
-static void invoke_name(void *data, int (*f)(Efile_error *, char *))
-{
-    struct t_data *d = (struct t_data *) data;
-    char *name = (char *) d->b;
-
-    d->again = 0;
-    d->result_ok = (*f)(&d->errInfo, name);
-}
-
-static void invoke_mkdir(void *data)
-{
-    DTRACE_INVOKE_SETUP_BY_NAME(FILE_MKDIR);
-    invoke_name(data, efile_mkdir);
-    DTRACE_INVOKE_RETURN(FILE_MKDIR);
-}
-
-static void invoke_rmdir(void *data)
-{
-    DTRACE_INVOKE_SETUP_BY_NAME(FILE_RMDIR);
-    invoke_name(data, efile_rmdir);
-    DTRACE_INVOKE_RETURN(FILE_RMDIR);
-}
-
-static void invoke_delete_file(void *data)
-{
-    DTRACE_INVOKE_SETUP_BY_NAME(FILE_DELETE);
-    invoke_name(data, efile_delete_file);
-    DTRACE_INVOKE_RETURN(FILE_DELETE);
-}
-
-static void invoke_chdir(void *data)
-{
-    DTRACE_INVOKE_SETUP_BY_NAME(FILE_CHDIR);
-    invoke_name(data, efile_chdir);
-    DTRACE_INVOKE_RETURN(FILE_CHDIR);
-}
-
-static void invoke_fdatasync(void *data)
-{
-    struct t_data *d = (struct t_data *) data;
-    int fd = (int) d->fd;
-    DTRACE_INVOKE_SETUP(FILE_FDATASYNC);
-
-    d->again = 0;
-    d->result_ok = efile_fdatasync(&d->errInfo, fd);
-    DTRACE_INVOKE_RETURN(FILE_FDATASYNC);
-}
-
-static void invoke_fsync(void *data)
-{
-    struct t_data *d = (struct t_data *) data;
-    int fd = (int) d->fd;
-    DTRACE_INVOKE_SETUP(FILE_FSYNC);
-
-    d->again = 0;
-    d->result_ok = efile_fsync(&d->errInfo, fd);
-    DTRACE_INVOKE_RETURN(FILE_FSYNC);
-}
-
-static void invoke_truncate(void *data)
-{
-    struct t_data *d = (struct t_data *) data;
-    int fd = (int) d->fd;
-    DTRACE_INVOKE_SETUP(FILE_TRUNCATE);
-
-    d->again = 0;
-    d->result_ok = efile_truncate_file(&d->errInfo, &fd, d->flags);
-    DTRACE_INVOKE_RETURN(FILE_TRUNCATE);
-}
-
-static void invoke_read(void *data)
-{
-    struct t_data *d = (struct t_data *) data;
-    int status, segment;
-    size_t size, read_size;
-    DTRACE_INVOKE_SETUP(FILE_READ);
-
-    segment = d->again && d->c.read.bin_size >= 2*FILE_SEGMENT_READ;
-    if (segment) {
-	size = FILE_SEGMENT_READ;
-    } else {
-	size = d->c.read.bin_size;
-    }
-    read_size = size;
-    if (d->flags & EFILE_COMPRESSED) {
-	read_size = erts_gzread((ErtsGzFile)d->fd,
-				d->c.read.binp->orig_bytes + d->c.read.bin_offset,
-				size);
-	status = (read_size != (size_t) -1);
-	if (!status) {
-	    d->errInfo.posix_errno = EIO;
-	}
-    } else {
-	status = efile_read(&d->errInfo, d->flags, (int) d->fd,
-			    d->c.read.binp->orig_bytes + d->c.read.bin_offset,
-			    size,
-			    &read_size);
-    }
-    if ( (d->result_ok = status)) {
-	ASSERT(read_size <= size);
-	d->c.read.bin_offset += read_size;
-	if (read_size < size || !segment) {
-	    d->c.read.bin_size = 0;
-	    d->again = 0;
-	} else {
-	    d->c.read.bin_size -= read_size;
-	}
-    } else {
-	d->again = 0;
-    }
-    DTRACE_INVOKE_RETURN(FILE_READ);
-}
-
-static void free_read(void *data)
-{
-    struct t_data *d = (struct t_data *) data;
-
-    driver_free_binary(d->c.read.binp);
-    EF_FREE(d);
-}
-
-static void invoke_read_line(void *data)
-{
-    struct t_data *d = (struct t_data *) data;
-    int status;
-    size_t read_size = 0;
-    int local_loop = (d->again == 0);
-    DTRACE_INVOKE_SETUP(FILE_READ_LINE);
-
-    do {
-	size_t size = (d->c.read_line.binp)->orig_size - 
-	    d->c.read_line.read_offset - d->c.read_line.read_size;
-	if (size == 0) {
-	    /* Need more place */
-	    ErlDrvSizeT need = (d->c.read_line.read_size >= DEFAULT_LINEBUF_SIZE) ?
-		d->c.read_line.read_size + DEFAULT_LINEBUF_SIZE : DEFAULT_LINEBUF_SIZE;
-	    ErlDrvBinary   *newbin;
-#if !ALWAYS_READ_LINE_AHEAD
-	    /* Use read_ahead size if need does not exceed it */
-	    if (need < (d->c.read_line.binp)->orig_size && 
-		d->c.read_line.read_ahead)
-	      need = (d->c.read_line.binp)->orig_size;
-#endif
-	    newbin = driver_alloc_binary(need);
-	    if (newbin == NULL) {
-		d->result_ok = 0;
-		d->errInfo.posix_errno = ENOMEM;
-		d->again = 0;
-		break;
-	    }
-	    memcpy(newbin->orig_bytes, (d->c.read_line.binp)->orig_bytes + d->c.read_line.read_offset,  
-		   d->c.read_line.read_size);
-	    driver_free_binary(d->c.read_line.binp);
-	    d->c.read_line.binp = newbin;
-	    d->c.read_line.read_offset = 0;
-	    size = need - d->c.read_line.read_size;
-	}
-	if (d->flags & EFILE_COMPRESSED) {
-	    read_size = erts_gzread((ErtsGzFile)d->fd,
-				    d->c.read_line.binp->orig_bytes + 
-				    d->c.read_line.read_offset + d->c.read_line.read_size,
-				    size);
-	    status = (read_size != (size_t) -1);
-	    if (!status) {
-		d->errInfo.posix_errno = EIO;
-	    }
-	} else {
-	    status = efile_read(&d->errInfo, d->flags, (int) d->fd,
-				d->c.read_line.binp->orig_bytes + 
-				d->c.read_line.read_offset + d->c.read_line.read_size,
-				size,
-				&read_size);
-	}
-	if ( (d->result_ok = status)) {
-	    void *nl_ptr = memchr((d->c.read_line.binp)->orig_bytes + 
-				  d->c.read_line.read_offset + d->c.read_line.read_size,'\n',read_size);
-	    ASSERT(read_size <= size);
-	    d->c.read_line.read_size += read_size;
-	    if (nl_ptr != NULL) {
-		/* If found, we're done */
-		d->c.read_line.nl_pos = ((char *) nl_ptr) - 
-		    ((char *) ((d->c.read_line.binp)->orig_bytes)) + 1;
-		if (d->c.read_line.nl_pos > 1 &&
-		    *(((char *) nl_ptr) - 1) == '\r') {
-		    --d->c.read_line.nl_pos;
-		    *(((char *) nl_ptr) - 1) = '\n';
-		    d->c.read_line.nl_skip = 1;
-		} else {
-		    d->c.read_line.nl_skip = 0;
-		}
-		d->again = 0;
-#if !ALWAYS_READ_LINE_AHEAD
-		if (!(d->c.read_line.read_ahead)) {
-		    /* Ouch! Undo buffering... */
-		    size_t too_much = d->c.read_line.read_size - d->c.read_line.nl_skip - 
-			(d->c.read_line.nl_pos - d->c.read_line.read_offset);
-		    d->c.read_line.read_size -= too_much;
-		    ASSERT(d->c.read_line.read_size >= 0);
-		    if (d->flags & EFILE_COMPRESSED) {
-			Sint64 location = erts_gzseek((ErtsGzFile)d->fd,
-						      -((Sint64) too_much), EFILE_SEEK_CUR);
-			if (location == -1) {
-			    d->result_ok = 0;
-			    d->errInfo.posix_errno = errno;
-			}
-		    } else {
-			Sint64 location;
-			d->result_ok = efile_seek(&d->errInfo, (int) d->fd, 
-						-((Sint64) too_much), EFILE_SEEK_CUR,
-						&location);
-		    }
-		}
-#endif
-		break;
-	    } else if (read_size == 0) {
-		d->c.read_line.nl_pos = 
-		    d->c.read_line.read_offset + d->c.read_line.read_size;
-		d->c.read_line.nl_skip = 0;
-		d->again = 0;
-		break;
-	    }
-	} else {
-	    d->again = 0;
-	    break;
-	}
-    } while (local_loop);
-    DTRACE_INVOKE_RETURN(FILE_READ_LINE);
-}
-
-static void free_read_line(void *data)
-{
-    struct t_data *d = (struct t_data *) data;
-
-    driver_free_binary(d->c.read_line.binp);
-    EF_FREE(d);
-}
-
-void read_file_zero_size(struct t_data* d);
-#define ZERO_FILE_CHUNK (64 * 1024)
-
-/* [ERL-327] Some special files like /proc/... have reported size 0 */
-void read_file_zero_size(struct t_data* d) {
-    size_t total_read_size = 0;
-    size_t allocated_size = ZERO_FILE_CHUNK; /* allocd in invoke_read_file */
-    for (;;) {
-        size_t read_result;
-
-        /* Read until we hit EOF (read less than FILE_SEGMENT_READ) */
-        d->result_ok = efile_read(&d->errInfo,
-                                  EFILE_MODE_READ,
-                                  (int) d->fd,
-                                  (d->c.read_file.binp->orig_bytes +
-                                   total_read_size),
-                                  ZERO_FILE_CHUNK,
-                                  &read_result);
-        if (!d->result_ok) {
-            break;
-        }
-
-        total_read_size += read_result;
-        d->c.read_file.offset += read_result;
-        if (read_result < ZERO_FILE_CHUNK) {
-            break;
-        }
-
-        /* Grow before the next read call */
-        allocated_size = total_read_size + ZERO_FILE_CHUNK;
-        d->c.read_file.binp = driver_realloc_binary(d->c.read_file.binp,
-                                                    allocated_size);
-    }
-
-    /* Finalize the memory usage. Hopefully it was read fully on the first
-     * go, so the binary allocation overhead becomes:
-     * alloc ZERO_FILE_CHUNK (64kb) -> realloc real_size */
-    if (allocated_size != total_read_size) {
-        d->c.read_file.binp = driver_realloc_binary(d->c.read_file.binp,
-                                                    total_read_size);
-    }
-    d->again = 0;
-}
-
-static void invoke_read_file(void *data)
-{
-    struct t_data *d = (struct t_data *) data;
-    size_t read_size;
-    int chop;
-    DTRACE_INVOKE_SETUP(FILE_READ_FILE);
-    
-    if (! d->c.read_file.binp) { /* First invocation only */
-	int fd;
-	Sint64 size;
-	
-	if (! (d->result_ok = 
-	       efile_openfile(&d->errInfo, d->b, 
-			      EFILE_MODE_READ, &fd, &size))) {
-	    goto done;
-	}
-	d->fd = fd;
-	d->c.read_file.size = (int) size;
-
-        /* For zero sized files allocate a reasonable chunk to attempt reading
-         * anyway. Note: This will eat ZERO_FILE_CHUNK bytes for any 0 file
-         * and free them immediately after (if the file was empty). */
-        ERTS_ASSERT(size >= 0);
-        d->c.read_file.binp = driver_alloc_binary(size != 0 ? (size_t)size
-                                                            : ZERO_FILE_CHUNK);
-
-	if (size < 0 || size != d->c.read_file.size || !d->c.read_file.binp) {
-	    d->result_ok = 0;
-	    d->errInfo.posix_errno = ENOMEM;
-	    goto close;
-	}
-	d->c.read_file.offset = 0;
-    }
-    /* Invariant: d->c.read_file.size >= d->c.read_file.offset */
-    
-    if (d->c.read_file.size == 0) {
-        read_file_zero_size(d);
-        goto close;
-    }
-
-    read_size = (size_t) (d->c.read_file.size - d->c.read_file.offset);
-    if (! read_size) goto close;
-    chop = d->again && read_size >= FILE_SEGMENT_READ*2;
-    if (chop) read_size = FILE_SEGMENT_READ;
-    d->result_ok = 
-	efile_read(&d->errInfo, 
-		   EFILE_MODE_READ, 
-		   (int) d->fd, 
-		   d->c.read_file.binp->orig_bytes + d->c.read_file.offset,
-		   read_size, 
-		   &read_size);
-    if (d->result_ok) {
-	d->c.read_file.offset += read_size;
-	if (chop) goto chop_done; /* again */
-    }
- close:
-    efile_closefile((int) d->fd);
- done:
-    d->again = 0;
- chop_done:
-    DTRACE_INVOKE_RETURN(FILE_READ_FILE);
-}
-
-static void free_read_file(void *data)
-{
-    struct t_data *d = (struct t_data *) data;
-
-    if (d->c.read_file.binp) driver_free_binary(d->c.read_file.binp);
-    EF_FREE(d);
-}
-
-
-
-static void invoke_preadv(void *data)
-{
-    struct t_data   *d = (struct t_data *) data;
-    struct t_preadv *c = &d->c.preadv;
-    ErlIOVec        *ev = &c->eiov;
-    size_t           bytes_read_so_far = 0;
-    unsigned char   *p = (unsigned char *)ev->iov[0].iov_base + 4+4+8*c->cnt;
-    DTRACE_INVOKE_SETUP(FILE_PREADV);
-
-    while (c->cnt < c->n) {
-	size_t read_size = ev->iov[1 + c->cnt].iov_len - c->size;
-	size_t bytes_read = 0;
-	int chop = d->again 
-	    && bytes_read_so_far + read_size >= 2*FILE_SEGMENT_READ;
-	if (chop) {
-	    ASSERT(bytes_read_so_far < FILE_SEGMENT_READ);
-	    read_size = FILE_SEGMENT_READ + FILE_SEGMENT_READ/2
-		- bytes_read_so_far;
-	}
-	if ( (d->result_ok 
-	      = efile_pread(&d->errInfo, 
-			    (int) d->fd,
-			    c->offsets[c->cnt] + c->size,
-			    ((char *)ev->iov[1 + c->cnt].iov_base) + c->size,
-			    read_size,
-			    &bytes_read))) {
-	    bytes_read_so_far += bytes_read;
-	    if (chop && bytes_read == read_size) {
-		c->size += bytes_read;
-		goto done;
-	    }
-	    ASSERT(bytes_read <= read_size);
-	    ev->iov[1 + c->cnt].iov_len = bytes_read + c->size;
-	    ev->size += bytes_read + c->size;
-	    put_int64(bytes_read + c->size, p); p += 8;
-	    c->size = 0;
-	    c->cnt++;
-	    if (d->again 
-		&& bytes_read_so_far >= FILE_SEGMENT_READ
-		&& c->cnt < c->n) {
-		goto done;
-	    }
-	} else {
-	    /* In case of a read error, ev->size will not be correct,
-	     * which does not matter since no read data is returned
-	     * to Erlang.
-	     */
-	    break;
-	}
-    }					
-    d->again = 0;
- done:
-    DTRACE_INVOKE_RETURN(FILE_PREADV);
-}
-
-static void free_preadv(void *data) {
-    struct t_data *d = data;
-    int            i;
-    ErlIOVec      *ev = &d->c.preadv.eiov;
-    
-    for(i = 0; i < ev->vsize; i++) {
-	driver_free_binary(ev->binv[i]);
-    }
-    EF_FREE(d);
-}
-
-static void invoke_ipread(void *data)
-{
-    struct t_data   *d = data;
-    struct t_preadv *c = &d->c.preadv;
-    ErlIOVec        *ev = &c->eiov;
-    size_t bytes_read = 0;
-    char buf[2*sizeof(Uint32)];
-    Uint32 offset, size;
-    DTRACE_INVOKE_SETUP(FILE_IPREAD);
-    
-    /* Read indirection header */
-    if (! efile_pread(&d->errInfo, (int) d->fd, c->offsets[0], 
-		      buf, sizeof(buf), &bytes_read)) {
-	goto error;
-    }
-    if (bytes_read != sizeof(buf)) goto done; /* eof */
-    size = get_int32(buf);
-    offset = get_int32(buf+4);
-    if (size > c->size) goto done; /* eof */
-    c->n = 1;
-    c->cnt = 0;
-    c->size = 0;
-    c->offsets[0] = offset;
-    if (! (ev->binv[0] = driver_alloc_binary(3*8))) {
-	d->errInfo.posix_errno = ENOMEM;
-	goto error;
-    }
-    ev->vsize = 1;
-    ev->iov[0].iov_len = 3*8;
-    ev->iov[0].iov_base = ev->binv[0]->orig_bytes;
-    ev->size = ev->iov[0].iov_len;
-    put_int64(offset, ev->iov[0].iov_base);
-    put_int64(size, ((char *)ev->iov[0].iov_base) + 2*8);
-    if (size == 0) {
-	put_int64(size, ((char *)ev->iov[0].iov_base) + 8);
-	goto done;
-    }
-    if (! (ev->binv[1] = driver_alloc_binary(size))) {
-	d->errInfo.posix_errno = ENOMEM;
-	goto error;
-    }
-    ev->vsize = 2;
-    ev->iov[1].iov_len = size;
-    ev->iov[1].iov_base = ev->binv[1]->orig_bytes;
-    /* Read data block */
-    d->invoke = invoke_preadv;
-    invoke_preadv(data);
-    DTRACE_INVOKE_RETURN(FILE_IPREAD);
-    return;
- error:
-    d->result_ok = 0;
-    d->again = 0;
-    DTRACE_INVOKE_RETURN(FILE_IPREAD);
-    return;
- done:
-    d->result_ok = !0;
-    d->again = 0;
-    DTRACE_INVOKE_RETURN(FILE_IPREAD);
-}
-
-/* invoke_writev and invoke_pwritev are the only thread functions that
- * access non-thread data i.e the port queue and a mutex in the port
- * structure that is used to lock the port queue.
- *
- * The port will normally not be terminated until the port queue is
- * empty, but if the port is killed, i.e., exit(Port, kill) is called,
- * it will terminate regardless of the port queue state. When the
- * port is invalid driver_peekq() returns NULL and set the size to -1,
- * and driver_sizeq() returns -1.
- */
-
-static void invoke_writev(void *data) {
-    struct t_data *d = (struct t_data *) data;
-    SysIOVec      *iov0;
-    SysIOVec      *iov;
-    int            iovlen;
-    int            iovcnt;
-    size_t         size;
-    size_t         p;
-    int            segment;
-    DTRACE_INVOKE_SETUP(FILE_WRITE);
-
-    segment = d->again && d->c.writev.size >= 2*FILE_SEGMENT_WRITE;
-    if (segment) {
-	size = FILE_SEGMENT_WRITE;
-    } else {
-	size = d->c.writev.size;
-    }
-
-    /* Copy the io vector to avoid locking the port que while writing,
-     * also, both we and efile_writev might/will change the SysIOVec
-     * when segmenting or due to partial write and we do not want to
-     * tamper with the actual queue that we get from driver_peekq
-     */
-    MUTEX_LOCK(d->c.writev.q_mtx); /* Lock before accessing the port queue */
-    iov0 = driver_peekq(d->c.writev.port, &iovlen);
-
-    /* Calculate iovcnt */
-    for (p = 0, iovcnt = 0;
-	 p < size && iovcnt < iovlen;
-	 p += iov0[iovcnt++].iov_len)
-	;
-    iov = EF_SAFE_ALLOC(sizeof(SysIOVec)*iovcnt);
-    memcpy(iov,iov0,iovcnt*sizeof(SysIOVec));
-    MUTEX_UNLOCK(d->c.writev.q_mtx);
-    /* Let go of lock until we deque from original vector */
-
-    if (iovlen > 0) {
-	ASSERT(iov[iovcnt-1].iov_len > p - size);
-	iov[iovcnt-1].iov_len -= p - size;
-	if (d->flags & EFILE_COMPRESSED) {
-	    int i, status = 1;
-	    for (i = 0; i < iovcnt; i++) {
-		if (iov[i].iov_base && iov[i].iov_len > 0) {
-		    /* Just in case, I do not know what gzwrite does
-		     * with errno.
-		     */
-		    errno = EINVAL; 
-		    status = erts_gzwrite((ErtsGzFile)d->fd,
-					  iov[i].iov_base,
-					  iov[i].iov_len) == iov[i].iov_len;
-		    if (! status) {
-			d->errInfo.posix_errno =
-			    d->errInfo.os_errno = errno; /* XXX Correct? */
-			break;
-		    }
-		}
-	    }
-	    d->result_ok = status;
-	} else {
-	    d->result_ok = efile_writev(&d->errInfo, 
-					d->flags, (int) d->fd,
-					iov, iovcnt);
-	}
-    } else if (iovlen == 0) {
-	d->result_ok = 1;
-    }
-    else { /* Port has terminated */
-	d->result_ok = 0;
-	d->errInfo.posix_errno = d->errInfo.os_errno = EINVAL;
-    }
-    EF_FREE(iov);
-
-    if (! d->result_ok) {
-	d->again = 0;
-	MUTEX_LOCK(d->c.writev.q_mtx);
-	driver_deq(d->c.writev.port, d->c.writev.size);
-	MUTEX_UNLOCK(d->c.writev.q_mtx);
-    } else {
-	if (! segment) {
-	    d->again = 0;
-	}
-	d->c.writev.size -= size;
-	TRACE_F(("w%lu", (unsigned long)size));
-	MUTEX_LOCK(d->c.writev.q_mtx);
-	driver_deq(d->c.writev.port, size);
-	MUTEX_UNLOCK(d->c.writev.q_mtx);
-    }
-
-
-    DTRACE_INVOKE_RETURN(FILE_WRITE);
-}
-
-static void invoke_pwd(void *data)
-{
-    struct t_data *d = (struct t_data *) data;
-    DTRACE_INVOKE_SETUP(FILE_PWD);
-
-    d->again = 0;
-    d->result_ok = efile_getdcwd(&d->errInfo,d->drive, d->b+1,
-				 RESBUFSIZE-1);
-    DTRACE_INVOKE_RETURN(FILE_PWD);
-}
-
-static void invoke_readlink(void *data)
-{
-    struct t_data *d = (struct t_data *) data;
-    char resbuf[RESBUFSIZE];	/* Result buffer. */
-    DTRACE_INVOKE_SETUP(FILE_READLINK);
-
-    d->again = 0;
-    d->result_ok = efile_readlink(&d->errInfo, d->b, resbuf+1,
-				  RESBUFSIZE-1);
-    if (d->result_ok != 0)
-	FILENAME_COPY((char *) d->b + 1, resbuf+1);
-    DTRACE_INVOKE_RETURN(FILE_READLINK);
-}
-
-static void invoke_altname(void *data)
-{
-    struct t_data *d = (struct t_data *) data;
-    char resbuf[RESBUFSIZE];	/* Result buffer. */
-    DTRACE_INVOKE_SETUP(FILE_ALTNAME);
-
-    d->again = 0;
-    d->result_ok = efile_altname(&d->errInfo, d->b, resbuf+1,
-				  RESBUFSIZE-1);
-    if (d->result_ok != 0)
-	FILENAME_COPY((char *) d->b + 1, resbuf+1);
-    DTRACE_INVOKE_RETURN(FILE_ALTNAME);
-}
-
-static void invoke_pwritev(void *data) {
-    struct t_data* const d = (struct t_data *) data;
-    struct t_pwritev * const c = &d->c.pwritev;
-    SysIOVec         *iov0;
-    SysIOVec         *iov;
-    int               iovlen;
-    int               iovcnt;
-    size_t            p;
-    int               segment;
-    size_t            size, write_size, written;
-    DTRACE_INVOKE_SETUP(FILE_PWRITEV);
-
-    segment = d->again && c->size >= 2*FILE_SEGMENT_WRITE;
-    if (segment) {
-	size = FILE_SEGMENT_WRITE;
-    } else {
-	size = c->size;
-    }
-    d->result_ok = !0;
-    p = 0;
-    /* Lock the queue just for a while, we don't want it locked during write */
-    MUTEX_LOCK(c->q_mtx);
-    iov0 = driver_peekq(c->port, &iovlen);
-    iov = EF_SAFE_ALLOC(sizeof(SysIOVec)*iovlen);
-    memcpy(iov,iov0,sizeof(SysIOVec)*iovlen);
-    MUTEX_UNLOCK(c->q_mtx);
-
-    if (iovlen < 0)
-	goto error; /* Port terminated */
-    for (iovcnt = 0, written = 0;
-	 c->cnt < c->n && iovcnt < iovlen && written < size;
-	 c->cnt++) {
-	int chop;
-	write_size = c->specs[c->cnt].size;
-	if (iov[iovcnt].iov_len - p < write_size) {
-	    goto error;
-	}
-	chop = segment && written + write_size >= 2*FILE_SEGMENT_WRITE;
-	if (chop) {
-	    ASSERT(written < FILE_SEGMENT_WRITE);
-	    write_size = FILE_SEGMENT_WRITE + FILE_SEGMENT_WRITE/2 
-		- written;
-	}
-	d->result_ok = efile_pwrite(&d->errInfo, (int) d->fd,
-				    (char *)(iov[iovcnt].iov_base) + p,
-				    write_size,
-				    c->specs[c->cnt].offset);
-	if (! d->result_ok) {
-	    d->again = 0;
-	    goto deq_error;
-	}
-	written += write_size; 
-	c->size -= write_size;
-	if (chop) { 
-	    c->specs[c->cnt].offset += write_size;
-	    c->specs[c->cnt].size -= write_size;
-	    /* Schedule out (d->again != 0) */
-	    break;
-	}
-	/* Move forward in buffer */
-	p += write_size;
-	ASSERT(iov[iovcnt].iov_len >= p);
-	if (iov[iovcnt].iov_len == p) {
-	    /* Move to next iov[], we trust that it is not a 
-	     * zero length vector, and thereby depend on that
-	     * such are not queued.
-	     */
-	    iovcnt++; p = 0;
-	}
-    }
-    if (! segment) {
-	if (c->cnt != c->n) {
-	    /* Mismatch between number of 
-	     * pos/size specs vs number of queued buffers .
-	     */
-	error:
-	    d->errInfo.posix_errno = EINVAL;
-	    d->result_ok = 0;
-	    d->again = 0;
-	deq_error:
-	    MUTEX_LOCK(c->q_mtx);
-	    driver_deq(c->port, c->size);
-	    MUTEX_UNLOCK(c->q_mtx);
-
-	    goto done;
-	} else {
-	    ASSERT(written == size);
-	    d->again = 0;
-	}
-    } else {
-      ASSERT(written >= FILE_SEGMENT_WRITE);
-    }
-      
-    MUTEX_LOCK(c->q_mtx);
-    driver_deq(c->port, written);
-    MUTEX_UNLOCK(c->q_mtx);
- done:
-    EF_FREE(iov); /* Free our copy of the vector, nothing to restore */
-    
-    DTRACE_INVOKE_RETURN(FILE_PWRITEV);
-}
-
-static void invoke_flstat(void *data)
-{
-    struct t_data *d = (struct t_data *) data;
-
-    DTRACE3(efile_drv_int_entry, d->sched_i1, d->sched_i2,
-            d->command == FILE_LSTAT ? FILE_LSTAT : FILE_FSTAT);
-    d->again = 0;
-    d->result_ok = efile_fileinfo(&d->errInfo, &d->info,
-				  d->b, d->command == FILE_LSTAT);
-    DTRACE3(efile_drv_int_entry, d->sched_i1, d->sched_i2,
-            d->command == FILE_LSTAT ? FILE_LSTAT : FILE_FSTAT);
-    gcc_optimizer_hack++;
-}
-
-static void invoke_link(void *data)
-{
-    struct t_data *d = (struct t_data *) data;
-    char *name = d->b;
-    char *new_name;
-    DTRACE_INVOKE_SETUP(FILE_LINK);
-
-    d->again = 0;
-    new_name = name+FILENAME_BYTELEN(name)+FILENAME_CHARSIZE;
-    d->result_ok = efile_link(&d->errInfo, name, new_name);
-    DTRACE_INVOKE_RETURN(FILE_LINK);
-}
-
-static void invoke_symlink(void *data)
-{
-    struct t_data *d = (struct t_data *) data;
-    char *name = d->b;
-    char *new_name;
-    DTRACE_INVOKE_SETUP(FILE_SYMLINK);
-
-    d->again = 0;
-    new_name = name+FILENAME_BYTELEN(name)+FILENAME_CHARSIZE;
-    d->result_ok = efile_symlink(&d->errInfo, name, new_name);
-    DTRACE_INVOKE_RETURN(FILE_SYMLINK);
-}
-
-static void invoke_rename(void *data)
-{
-    struct t_data *d = (struct t_data *) data;
-    char *name = d->b;
-    char *new_name;
-    DTRACE_INVOKE_SETUP(FILE_RENAME);
-
-    d->again = 0;
-    new_name = name+FILENAME_BYTELEN(name)+FILENAME_CHARSIZE;
-    d->result_ok = efile_rename(&d->errInfo, name, new_name);
-    DTRACE_INVOKE_RETURN(FILE_RENAME);
-}
-
-static void invoke_write_info(void *data)
-{
-    struct t_data *d = (struct t_data *) data;
-    DTRACE_INVOKE_SETUP(FILE_WRITE_INFO);
-
-    d->again = 0;
-    d->result_ok = efile_write_info(&d->errInfo, &d->info, d->b);
-    DTRACE_INVOKE_RETURN(FILE_WRITE_INFO);
-}
-
-static void invoke_lseek(void *data)
-{
-    struct t_data *d = (struct t_data *) data;
-    int status;
-    DTRACE_INVOKE_SETUP(FILE_LSEEK);
-
-    d->again = 0;
-    if (d->flags & EFILE_COMPRESSED) {
-	int offset = (int) d->c.lseek.offset;
-	
-	if (offset != d->c.lseek.offset) {
-	    d->errInfo.posix_errno = EINVAL;
-	    status = 0;
-	} else {
-	    d->c.lseek.location = erts_gzseek((ErtsGzFile)d->fd,
-					      offset, d->c.lseek.origin);
-	    if (d->c.lseek.location == -1) {
-		d->errInfo.posix_errno = errno;
-		status = 0;
-	    } else {
-		status = 1;
-	    }
-	}
-    } else {
-	status = efile_seek(&d->errInfo, (int) d->fd, 
-			    d->c.lseek.offset, d->c.lseek.origin,
-			    &d->c.lseek.location);
-    }
-    d->result_ok = status;
-    DTRACE_INVOKE_RETURN(FILE_LSEEK);
-}
-
-static void invoke_readdir(void *data)
-{
-    struct t_data *d = (struct t_data *) data;
-    char *p = NULL;
-    size_t file_bs;
-    size_t n = 0, total = 0;
-    struct t_readdir_buf *b = NULL;
-    int res = 0;
-    DTRACE_INVOKE_SETUP(FILE_READDIR);
-
-    d->again = 0;
-    d->errInfo.posix_errno = 0;
-
-    do {
-	total   = READDIR_BUFSIZE;
-	n       = 1;
-	b       = EF_SAFE_ALLOC(sizeof(struct t_readdir_buf));
-	b->next = NULL;
-	
-	if (d->c.read_dir.last_buf) {
-	    d->c.read_dir.last_buf->next = b;
-	} else {
-	    d->c.read_dir.first_buf = b;
-	}
-	d->c.read_dir.last_buf = b;
-
-	p       = &b->buf[0];
-	p[0]    = FILE_RESP_LFNAME;
-	file_bs = READDIR_BUFSIZE - n;
-
-	do {
-	    res = efile_readdir(&d->errInfo, d->b, &d->dir_handle, p + n + 2, &file_bs);
-
-	    if (res) {
-		put_int16((Uint16)file_bs, p + n);
-		n += 2 + file_bs;
-		file_bs = READDIR_BUFSIZE - n;
-	    }
-	} while( res && ((total - n - 2) >= MAXPATHLEN*FILENAME_CHARSIZE));
-
-	b->n = n;
-    } while(res);
-
-    d->result_ok = (d->errInfo.posix_errno == 0);
-    DTRACE_INVOKE_RETURN(FILE_READDIR);
-}
-
-static void invoke_open(void *data)
-{
-    struct t_data *d = (struct t_data *) data;
-    int status = 1;		/* Status of open call. */
-    DTRACE_INVOKE_SETUP(FILE_OPEN);
-
-    d->again = 0;
-    if ((d->flags & EFILE_COMPRESSED) == 0) {
-	int fd;
-	status = efile_openfile(&d->errInfo, d->b, d->flags, &fd, NULL);
-	d->fd = fd;
-    } else {
-	char* mode = NULL;
-
-	if (((d->flags & (EFILE_MODE_READ_WRITE)) == EFILE_MODE_READ_WRITE) ||
-	    (d->flags & EFILE_MODE_APPEND)) {
-	    status = 0;
-	    d->errInfo.posix_errno = EINVAL;
-	} else {
-	    status = efile_may_openfile(&d->errInfo, d->b);
-	    if (status || (d->errInfo.posix_errno != EISDIR)) {
-		mode = (d->flags & EFILE_MODE_READ) ? "rb" : "wb";
-		d->fd = (SWord) erts_gzopen(d->b, mode);
-		if ((ErtsGzFile)d->fd) {
-		    status = 1;
-		} else {
-		    if (errno == 0) {
-			errno = ENOMEM;
-		    }
-		    d->errInfo.posix_errno = errno;
-		    status = 0;
-		}
-	    }
-	}
-    }
-
-    d->result_ok = status;
-    if (!status) {
-        d->fd = FILE_FD_INVALID;
-    }
-    DTRACE_INVOKE_RETURN(FILE_OPEN);
-}
-
-static void invoke_fadvise(void *data)
-{
-    struct t_data *d = (struct t_data *) data;
-    int fd = (int) d->fd;
-    off_t offset = (off_t) d->c.fadvise.offset;
-    off_t length = (off_t) d->c.fadvise.length;
-    int advise = (int) d->c.fadvise.advise;
-    DTRACE_INVOKE_SETUP(FILE_FADVISE);
-
-    d->again = 0;
-    d->result_ok = efile_fadvise(&d->errInfo, fd, offset, length, advise);
-    DTRACE_INVOKE_RETURN(FILE_FADVISE);
-}
-
-#ifdef HAVE_SENDFILE
-static void invoke_sendfile(void *data)
-{
-    struct t_data *d = (struct t_data *)data;
-    int fd = d->fd;
-    int out_fd = (int)d->c.sendfile.out_fd;
-    Uint64 nbytes = d->c.sendfile.nbytes;
-    int result = 0;
-    d->again = 0;
-
-    result = efile_sendfile(&d->errInfo, fd, out_fd, &d->c.sendfile.offset, &nbytes, NULL);
-
-    d->c.sendfile.written += nbytes;
-
-    if (result == 1 || (result == 0 && USE_THRDS_FOR_SENDFILE(d))) {
-      d->result_ok = 0;
-    } else if (result == 0 && (d->errInfo.posix_errno == EAGAIN
-				 || d->errInfo.posix_errno == EINTR)) {
-      if ((d->c.sendfile.nbytes - nbytes) != 0) {
-	d->result_ok = 1;
-	if (d->c.sendfile.nbytes != 0)
-	  d->c.sendfile.nbytes -= nbytes;
-      } else if (nbytes == 0 && d->c.sendfile.nbytes == 0) {
-	d->result_ok = 1;
-      } else
-	d->result_ok = 0;
-    } else {
-	d->result_ok = -1;
-    }
-}
-
-static void free_sendfile(void *data) {
-    struct t_data *d = (struct t_data *)data;
-    if (USE_THRDS_FOR_SENDFILE(d)) {
-	SET_NONBLOCKING(d->c.sendfile.out_fd);
-    } else {
-	MUTEX_LOCK(d->c.sendfile.q_mtx);
-	driver_deq(d->c.sendfile.port,1);
-	MUTEX_UNLOCK(d->c.sendfile.q_mtx);
-	driver_select(d->c.sendfile.port, (ErlDrvEvent)(long)d->c.sendfile.out_fd,
-                      ERL_DRV_USE_NO_CALLBACK|ERL_DRV_WRITE, 0);
-    }
-    EF_FREE(data);
-}
-
-static void file_ready_output(ErlDrvData data, ErlDrvEvent event)
-{
-    file_descriptor* fd = (file_descriptor*) data;
-
-    switch (fd->d->command) {
-    case FILE_SENDFILE:
-	driver_select(fd->d->c.sendfile.port, event,
-		      (int)ERL_DRV_WRITE,(int) 0);
-	invoke_sendfile((void *)fd->d);
-	file_async_ready(data, (ErlDrvThreadData)fd->d);
-	break;
-    default:
-	break;
-    }
-}
-
-static void file_stop_select(ErlDrvEvent event, void* _)
-{
-
-}
-
-static int flush_sendfile(file_descriptor *desc,void *_) {
-    if (desc->sendfile_state == sending) {
-	desc->d->result_ok = -1;
-	desc->d->errInfo.posix_errno = ECONNABORTED;
-	file_async_ready((ErlDrvData)desc,(ErlDrvThreadData)desc->d);
-    }
-    return 1;
-}
-#endif /* HAVE_SENDFILE */
-
-
-static void invoke_fallocate(void *data)
-{
-    struct t_data *d = (struct t_data *) data;
-    int fd = (int) d->fd;
-    Sint64 offset = d->c.fallocate.offset;
-    Sint64 length = d->c.fallocate.length;
-
-    d->again = 0;
-    d->result_ok = efile_fallocate(&d->errInfo, fd, offset, length);
-}
-
-static void free_readdir(void *data)
-{
-    struct t_data *d = (struct t_data *) data;
-    struct t_readdir_buf *b1 = d->c.read_dir.first_buf;
-
-    while (b1) {
-	struct t_readdir_buf *b2 = b1;
-	b1 = b1->next;
-	EF_FREE(b2);
-    }
-    EF_FREE(d);
-}
-
-
-
-static void try_free_read_bin(file_descriptor *desc) {
-    if ((desc->read_size == 0)
-	&& (desc->read_offset >= desc->read_binp->orig_size)) {
-	ASSERT(desc->read_offset == desc->read_binp->orig_size);
-	driver_free_binary(desc->read_binp);
-	desc->read_binp = NULL;
-	desc->read_offset = 0;
-	desc->read_size = 0;
-    }
-}
-
-
-
-static int try_again(file_descriptor *desc, struct t_data *d) {
-    if (! d->again)
-	return 0;
-    if (desc->timer_state != timer_idle) {
-	driver_cancel_timer(desc->port);
-    }
-    desc->timer_state = timer_again;
-    desc->invoke = d->invoke;
-    desc->d = d;
-    desc->free = d->free;
-    driver_set_timer(desc->port, 0L);
-    return !0;
-}
-
-
-
-static void cq_execute(file_descriptor *desc) {
-    struct t_data *d;
-    register void *void_ptr; /* Soft cast variable */
-    if (desc->timer_state == timer_again)
-	return;
-#ifdef HAVE_SENDFILE
-    if (desc->sendfile_state == sending)
-	return;
-#endif
-    if (! (d = cq_deq(desc)))
-	return;
-    TRACE_F(("x%i", (int) d->command));
-    d->again = sys_info.async_threads == 0;
-    DRIVER_ASYNC(d->level, desc, d->invoke, void_ptr=d, d->free);
-}
-
-static struct t_data *async_write(file_descriptor *desc, int *errp,
-		       int reply, Uint32 reply_size
-#ifdef USE_VM_PROBES
-		       ,Sint64 *dt_i1, Sint64 *dt_i2, Sint64 *dt_i3
-#endif
-) {
-    struct t_data *d;
-    if (! (d = EF_ALLOC(sizeof(struct t_data) - 1))) {
-	if (errp) *errp = ENOMEM;
-	return NULL;
-    }
-    TRACE_F(("w%lu", (unsigned long)desc->write_buffered));
-    d->command = FILE_WRITE;
-    d->fd = desc->fd;
-    d->flags = desc->flags;
-    d->c.writev.port = desc->port;
-    d->c.writev.q_mtx = desc->q_mtx;
-    d->c.writev.size = desc->write_buffered;
-#ifdef USE_VM_PROBES
-    if (dt_i1 != NULL) {
-        *dt_i1 = d->fd;
-        *dt_i2 = d->flags;
-        *dt_i3 = d->c.writev.size;
-    }
-#endif
-    d->reply = reply;
-    d->c.writev.reply_size = reply_size;
-    d->invoke = invoke_writev;
-    d->free = free_data;
-    d->level = 1;
-    cq_enq(desc, d);
-    desc->write_buffered = 0;
-    return d;
-}
-
-static int flush_write(file_descriptor *desc, int *errp
-#ifdef USE_VM_PROBES
-                       , dt_private *dt_priv, char *dt_utag
-#endif
-) {
-    int    result = 0;
-#ifdef USE_VM_PROBES
-    Sint64 dt_i1 = 0, dt_i2 = 0, dt_i3 = 0;
-#endif
-    struct t_data *d = NULL;
-
-    MUTEX_LOCK(desc->q_mtx);
-    if (desc->write_buffered > 0) {
-	if ((d = async_write(desc, errp, 0, 0
-#ifdef USE_VM_PROBES
-			     ,&dt_i1, &dt_i2, &dt_i3
-#endif
-			     )) == NULL) {
-            result = -1;
-        }
-    }
-    MUTEX_UNLOCK(desc->q_mtx);
-#ifdef USE_VM_PROBES
-    if (d != NULL) {
-        d->sched_i1 = dt_priv->thread_num;
-        d->sched_i2 = dt_priv->tag;
-        d->sched_utag[0] = '\0';
-        if (dt_utag != NULL) {
-            if (dt_utag[0] == '\0') {
-                dt_utag = NULL;
-            } else {
-                strncpy(d->sched_utag, dt_utag, sizeof(d->sched_utag) - 1);
-                d->sched_utag[sizeof(d->sched_utag) - 1] = '\0';
-            }
-        }
-        DTRACE11(efile_drv_entry, dt_priv->thread_num, dt_priv->tag++,
-                 dt_utag, FILE_WRITE,
-                 NULL, NULL, dt_i1, dt_i2, dt_i3, 0, desc->port_str);
-    }
-#endif /* USE_VM_PROBES */
-    return result;
-}
-
-static int check_write_error(file_descriptor *desc, int *errp) {
-    if (desc->write_error) {
-	if (errp) *errp = desc->write_errInfo.posix_errno;
-	desc->write_error = 0;
-	return -1;
-    }
-    return 0;
-}
-
-static int flush_write_check_error(file_descriptor *desc, int *errp
-#ifdef USE_VM_PROBES
-                                   , dt_private *dt_priv, char *dt_utag
-#endif
-				   ) {
-    int r;
-    if ( (r = flush_write(desc, errp
-#ifdef USE_VM_PROBES
-			  , dt_priv, dt_utag
-#endif
-			  )) != 0) {
-	check_write_error(desc, NULL);
-	return r;
-    } else {
-	return check_write_error(desc, errp);
-    }
-}
-
-static struct t_data *async_lseek(file_descriptor *desc, int *errp, int reply,
-				  Sint64 offset, int origin
-#ifdef USE_VM_PROBES
-				  , Sint64 *dt_i1, Sint64 *dt_i2, Sint64 *dt_i3
-#endif
-				  ) {
-    struct t_data *d;
-    if (! (d = EF_ALLOC(sizeof(struct t_data)))) {
-	*errp = ENOMEM;
-	return NULL;
-    }
-    d->flags = desc->flags;
-    d->fd = desc->fd;
-    d->command = FILE_LSEEK;
-    d->reply = reply;
-    d->c.lseek.offset = offset;
-    d->c.lseek.origin = origin;
-#ifdef USE_VM_PROBES
-    if (dt_i1 != NULL) {
-        *dt_i1 = d->fd;
-        *dt_i2 = d->c.lseek.offset;
-        *dt_i3 = d->c.lseek.origin;
-    }
-#endif
-    d->invoke = invoke_lseek;
-    d->free = free_data;
-    d->level = 1;
-    cq_enq(desc, d);
-    return d;
-}
-
-static void flush_read(file_descriptor *desc) {
-    desc->read_offset = 0;
-    desc->read_size = 0;
-    if (desc->read_binp) {
-	driver_free_binary(desc->read_binp);
-	desc->read_binp = NULL;
-    }
-}
-
-static int lseek_flush_read(file_descriptor *desc, int *errp
-#ifdef USE_VM_PROBES
-			    ,dt_private *dt_priv, char *dt_utag
-#endif
-			    ) {
-    int r = 0;
-    size_t read_size = desc->read_size;
-#ifdef USE_VM_PROBES
-    Sint64 dt_i1 = 0, dt_i2 = 0, dt_i3 = 0;
-#endif
-    struct t_data *d;
-
-    flush_read(desc);
-    if (read_size != 0) {
-	if ((d = async_lseek(desc, errp, 0,
-                             -((ssize_t)read_size), EFILE_SEEK_CUR
-#ifdef USE_VM_PROBES
-			     , &dt_i1, &dt_i2, &dt_i3
-#endif
-			     )) == NULL) {
-            r = -1;
-        } else {
-#ifdef USE_VM_PROBES
-            d->sched_i1 = dt_priv->thread_num;
-            d->sched_i2 = dt_priv->tag;
-            d->sched_utag[0] = '\0';
-            if (dt_utag != NULL) {
-                if (dt_utag[0] == '\0') {
-                    dt_utag = NULL;
-                } else {
-                    strncpy(d->sched_utag, dt_utag, sizeof(d->sched_utag) - 1);
-                    d->sched_utag[sizeof(d->sched_utag) - 1] = '\0';
-                }
-            }
-            DTRACE11(efile_drv_entry, dt_priv->thread_num, dt_priv->tag++,
-                     dt_utag, FILE_LSEEK,
-                     NULL, NULL, dt_i1, dt_i2, dt_i3, 0, desc->port_str);
-#endif /* USE_VM_PROBES */
-        }
-    }
-    return r;
-}
-
-
-/*********************************************************************
- * Driver entry point -> stop
- * The close has to be scheduled on async thread, so that currently active
- * async operation does not suddenly have the ground disappearing under their feet...
- */
-static void 
-file_stop(ErlDrvData e)
-{
-    file_descriptor* desc = (file_descriptor*)e;
-
-    TRACE_C('p');
-
-    IF_THRDS {
-	flush_read(desc);
-	if (desc->fd != FILE_FD_INVALID) {
-	    struct t_data *d = EF_SAFE_ALLOC(sizeof(struct t_data));
-	    d->command = FILE_CLOSE_ON_PORT_EXIT;
-	    d->reply = !0;
-	    d->fd = desc->fd;
-	    d->flags = desc->flags;
-	    d->invoke = invoke_close;
-	    d->free = free_data;
-	    d->level = 2;
-	    d->data_to_free = (void *) desc;
-	    cq_enq(desc, d);
-	    desc->fd = FILE_FD_INVALID;
-	    desc->flags = 0;
-	    cq_execute(desc);
-	} else {
-	    EF_FREE(desc);
-	}
-    } else {
-	if (desc->fd != FILE_FD_INVALID) {
-	    do_close(desc->flags, desc->fd);
-	    desc->fd = FILE_FD_INVALID;
-	    desc->flags = 0;
-	}
-	if (desc->read_binp) {
-	    driver_free_binary(desc->read_binp);
-	}
-	EF_FREE(desc);
-    }
-}
-
-/*********************************************************************
- * Driver entry point -> ready_async
- */
-static void 
-file_async_ready(ErlDrvData e, ErlDrvThreadData data)
-{
-    file_descriptor *desc = (file_descriptor*)e;
-    struct t_data *d = (struct t_data *) data;
-    char header[5];		/* result code + count */
-    char resbuf[RESBUFSIZE];	/* Result buffer. */
-#ifdef  USE_VM_PROBES
-    int sched_i1 = d->sched_i1, sched_i2 = d->sched_i2, command = d->command,
-        result_ok = d->result_ok,
-        posix_errno = d->result_ok ? 0 : d->errInfo.posix_errno;
-    DTRACE_CHARBUF(sched_utag, DTRACE_EFILE_BUFSIZ+1);
-
-    sched_utag[0] = '\0';
-    if (DTRACE_ENABLED(efile_drv_return)) {
-        strncpy(sched_utag, d->sched_utag, DTRACE_EFILE_BUFSIZ);
-        sched_utag[DTRACE_EFILE_BUFSIZ] = '\0';
-    }
-#endif  /* USE_VM_PROBES */
-
-    TRACE_C('r');
-
-    if (try_again(desc, d)) {
-        /* DTRACE TODO: what kind of probe makes sense here? */
-	return;
-    }
-
-    switch (d->command)
-    {
-    case FILE_READ:
-	if (!d->result_ok) {
-	    reply_error(desc, &d->errInfo);
-	} else {
-	    size_t available_bytes = 
-		d->c.read.bin_offset + d->c.read.bin_size - desc->read_offset;
-	    if (available_bytes < d->c.read.size) {
-		d->c.read.size = available_bytes;
-	    }
-	    TRACE_C('D');
-	    reply_data(desc, d->c.read.binp, 
-		       desc->read_offset, d->c.read.size);
-	    desc->read_offset += d->c.read.size;
-	    desc->read_size = 
-		d->c.read.bin_offset + d->c.read.bin_size - desc->read_offset;
-	    try_free_read_bin(desc);
-	}
-	free_read(data);
-	break;
-      case FILE_READ_LINE:
-	  /* The read_line structure differs from the read structure.
-	     The data->read_offset and d->c.read_line.read_offset are copies, as are
-	     data->read_size and d->c.read_line.read_size
-             The read_line function does not kniow in advance how large the binary has to be,
-	     why new allocation (but not reallocation of the old binary, for obvious reasons) 
-	     may happen in the worker thread. */
-	if (!d->result_ok) {
-	    reply_error(desc, &d->errInfo);
-	} else {
-	    size_t len = d->c.read_line.nl_pos - d->c.read_line.read_offset;
-	    TRACE_C('L');
-	    reply_data(desc, d->c.read_line.binp, 
-		       d->c.read_line.read_offset, len);
-	    desc->read_offset = d->c.read_line.read_offset + d->c.read_line.nl_skip + len;
-	    desc->read_size = 
-		d->c.read_line.read_size - d->c.read_line.nl_skip - len;
-	    if (desc->read_binp != d->c.read_line.binp) { /* New binary allocated */
-		driver_free_binary(desc->read_binp);
-		desc->read_binp =  d->c.read_line.binp;
-		driver_binary_inc_refc(desc->read_binp);
-	    }
-#if !ALWAYS_READ_LINE_AHEAD
-	    ASSERT(desc->read_bufsize > 0 || desc->read_size == 0);
-	    if (desc->read_bufsize == 0) {
-		desc->read_offset = desc->read_binp->orig_size; /* triggers cleanup */
-	    }
-#endif
-	    try_free_read_bin(desc);
-	}
-	free_read_line(data);
-	break;
-      case FILE_READ_FILE:
-	if (!d->result_ok)
-	    reply_error(desc, &d->errInfo);
-	else {
-	    header[0] = FILE_RESP_ALL_DATA;
-	    TRACE_C('R');
-	    driver_output_binary(desc->port, header, 1,
-				 d->c.read_file.binp,
-				 0, d->c.read_file.offset);
-	}
-	free_read_file(data);
-	break;
-      case FILE_WRITE:
-	  if (d->reply) {
-	      if (! d->result_ok) {
-		  reply_error(desc, &d->errInfo);
-	      } else {
-		  reply_Uint(desc, d->c.writev.reply_size);
-	      }
-	  } else {
-	      if (! d->result_ok) {
-		  desc->write_error = !0;
-		  desc->write_errInfo = d->errInfo;
-	      }
-	  }
-	  free_data(data);
-	  break;
-      case FILE_LSEEK:
-	  if (d->reply) {
-	      if (d->result_ok)
-		  reply_Sint64(desc, d->c.lseek.location);
-	      else
-		  reply_error(desc, &d->errInfo);
-	  }
-	  free_data(data);
-	  break;
-      case FILE_MKDIR:
-      case FILE_RMDIR:
-      case FILE_CHDIR:
-      case FILE_DELETE:
-      case FILE_FDATASYNC:
-      case FILE_FSYNC:
-      case FILE_TRUNCATE:
-      case FILE_LINK:
-      case FILE_SYMLINK:
-      case FILE_RENAME:
-      case FILE_WRITE_INFO:
-      case FILE_FADVISE:
-      case FILE_FALLOCATE:
-	reply(desc, d->result_ok, &d->errInfo);
-	free_data(data);
-	break;
-      case FILE_ALTNAME:
-      case FILE_PWD:
-      case FILE_READLINK:
-        {
-	    int length;
-	    char *resbuf = d->b;
-
-	    if (!d->result_ok)
-		reply_error(desc, &d->errInfo);
-	    else {
-		resbuf[0] = FILE_RESP_FNAME;
-		length = 1+FILENAME_BYTELEN((char*) resbuf+1);
-		TRACE_C('R');
-		driver_output2(desc->port, resbuf, 1, resbuf+1, length-1);
-	    }
-	    free_data(data);
-	    break;
-	}
-      case FILE_OPEN:
-	if (!d->result_ok) {
-	    reply_error(desc, &d->errInfo);
-	} else {
-	    ASSERT(d->is_fd_unused);
-	    desc->fd = d->fd;
-	    desc->flags = d->flags;
-	    d->is_fd_unused = 0;
-	    reply_Uint(desc, d->fd);
-	}
-	free_data(data);
-	break;
-      case FILE_FSTAT:
-      case FILE_LSTAT:
-        {
-	    if (d->result_ok) {
-		resbuf[0] = FILE_RESP_INFO;
-
-		put_int32(d->info.size_high,         &resbuf[1 + ( 0 * 4)]);
-		put_int32(d->info.size_low,          &resbuf[1 + ( 1 * 4)]);
-		put_int32(d->info.type,              &resbuf[1 + ( 2 * 4)]);
-
-		/* Note 64 bit indexing in resbuf here */
-		put_int64(d->info.accessTime,        &resbuf[1 + ( 3 * 4)]);
-		put_int64(d->info.modifyTime,        &resbuf[1 + ( 5 * 4)]);
-		put_int64(d->info.cTime,             &resbuf[1 + ( 7 * 4)]);
-
-		put_int32(d->info.mode,              &resbuf[1 + ( 9 * 4)]);
-		put_int32(d->info.links,             &resbuf[1 + (10 * 4)]);
-		put_int32(d->info.major_device,      &resbuf[1 + (11 * 4)]);
-		put_int32(d->info.minor_device,      &resbuf[1 + (12 * 4)]);
-		put_int32(d->info.inode,             &resbuf[1 + (13 * 4)]);
-		put_int32(d->info.uid,               &resbuf[1 + (14 * 4)]);
-		put_int32(d->info.gid,               &resbuf[1 + (15 * 4)]);
-		put_int32(d->info.access,            &resbuf[1 + (16 * 4)]);
-
-#define RESULT_SIZE (1 + (17 * 4))
-		TRACE_C('R');
-		driver_output2(desc->port, resbuf, RESULT_SIZE, NULL, 0);
-#undef RESULT_SIZE
-	    } else
-		reply_error(desc, &d->errInfo);
-	}
-	free_data(data);
-	break;
-      case FILE_READDIR:
-	if (!d->result_ok) {
-	    reply_error(desc, &d->errInfo);
-	} else {
-	    struct t_readdir_buf *b1 = d->c.read_dir.first_buf;
-	    char   op = FILE_RESP_LFNAME;
-
-	    TRACE_C('R');
-	    ASSERT(b1);
-
-	    while (b1) {
-		struct t_readdir_buf *b2 = b1;
-		char *p = &b1->buf[0];
-		driver_output2(desc->port, p, 1, p + 1, b1->n - 1);
-		b1 = b1->next;
-		EF_FREE(b2);
-	    }
-	    driver_output2(desc->port, &op, 1, NULL, 0);
-    
-	    d->c.read_dir.first_buf = NULL;
-	    d->c.read_dir.last_buf = NULL;
-	}
-	free_readdir(data);
-	break;
-      case FILE_CLOSE:
-	  if (d->reply) {
-	      TRACE_C('K');
-	      reply_ok(desc);
-#ifdef USE_VM_PROBES
-              result_ok = 1;
-#endif
-	  }
-	  free_data(data);
-	  break;
-      case FILE_PWRITEV:
-	  if (!d->result_ok) {
-	      reply_Uint_error(desc, d->c.pwritev.cnt, &d->errInfo);
-	  } else {
-	      reply_Uint(desc, d->c.pwritev.n);
-	  }
-	  free_data(data);
-	  break;
-      case FILE_PREADV:
-	  if (!d->result_ok) {
-	      reply_error(desc, &d->errInfo);
-	  } else {
-	      reply_ev(desc, FILE_RESP_LDATA, &d->c.preadv.eiov);
-	  }
-	  free_preadv(data);
-	  break;
-      case FILE_IPREAD:
-	  if (!d->result_ok) {
-	      reply_error(desc, &d->errInfo);
-	  } else if (!d->c.preadv.eiov.vsize) {
-	      reply_eof(desc);
-	  } else {
-	      reply_ev(desc, FILE_RESP_N2DATA, &d->c.preadv.eiov);
-	  }
-	  free_preadv(data);
-	  break;
-#ifdef HAVE_SENDFILE
-      case FILE_SENDFILE:
-	  if (d->result_ok == -1) {
-	      if (d->errInfo.posix_errno == ECONNRESET ||
-		  d->errInfo.posix_errno == ENOTCONN ||
-		  d->errInfo.posix_errno == EPIPE)
-		  reply_string_error(desc,"closed");
-	      else
-		  reply_error(desc, &d->errInfo);
-	      desc->sendfile_state = not_sending;
-	      free_sendfile(data);
-	  } else if (d->result_ok == 0) {
-	      reply_Sint64(desc, d->c.sendfile.written);
-	      desc->sendfile_state = not_sending;
-	      free_sendfile(data);
-	  } else if (d->result_ok == 1) { /* If we are using select to send the rest of the data */
-	      desc->sendfile_state = sending;
-	      desc->d = d;
-	      driver_select(desc->port, (ErlDrvEvent)(long)d->c.sendfile.out_fd,
-			    ERL_DRV_USE|ERL_DRV_WRITE, 1);
-	  }
-	  break;
-#endif
-      case FILE_CLOSE_ON_PORT_EXIT:
-	  /* See file_stop. However this is never invoked after the port is killed. */
-	  free_data(data);
-	  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,
-		  command, result_ok, posix_errno);
-	  return;
-      default:
-	abort();
-    }
-    DTRACE6(efile_drv_return, sched_i1, sched_i2, sched_utag,
-            command, result_ok, posix_errno);
-    if (desc->write_buffered != 0 && desc->timer_state == timer_idle ) {
-	desc->timer_state = timer_write;
-	driver_set_timer(desc->port, desc->write_delay);
-    }
-    cq_execute(desc);
-
-}
-
-
-/*********************************************************************
- * Driver entry point -> output
- */
-static void 
-file_output(ErlDrvData e, char* buf, ErlDrvSizeT count)
-{
-    file_descriptor* desc = (file_descriptor*)e;
-    Efile_error errInfo;	/* The error codes for the last operation. */
-    Sint fd;			/* The file descriptor for this port, if any,
-				 * -1 if none.
-				 */
-    char* name;			/* Points to the filename in buf. */
-    int command;
-    struct t_data *d = NULL;
-#ifdef  USE_VM_PROBES
-    char *dt_utag = NULL;
-    char *dt_s1 = NULL, *dt_s2 = NULL;
-    Sint64 dt_i1 = 0;
-    Sint64 dt_i2 = 0;
-    Sint64 dt_i3 = 0;
-    Sint64 dt_i4 = 0;
-    dt_private *dt_priv = get_dt_private(0);
-#endif  /* USE_VM_PROBES */
-
-    TRACE_C('o');
-
-    fd  = desc->fd;
-    name = buf+1;
-    command = *(uchar*)buf++;
-
-    switch(command) {
-
-    case FILE_MKDIR:
-    {
-	d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1 + FILENAME_BYTELEN(name) + FILENAME_CHARSIZE);
-	
-	FILENAME_COPY(d->b, name);
-#ifdef USE_VM_PROBES
-	dt_s1 = d->b;
-	dt_utag = name + FILENAME_BYTELEN(name) + FILENAME_CHARSIZE;
-#endif
-	d->command = command;
-	d->invoke = invoke_mkdir;
-	d->free = free_data;
-	d->level = 2;
-	goto done;
-    }
-    case FILE_RMDIR:
-    {
-	d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1 + FILENAME_BYTELEN(name) + FILENAME_CHARSIZE);
-	
-	FILENAME_COPY(d->b, name);
-#ifdef USE_VM_PROBES
-	dt_s1 = d->b;
-	dt_utag = name + FILENAME_BYTELEN(name) + FILENAME_CHARSIZE;
-#endif
-	d->command = command;
-	d->invoke = invoke_rmdir;
-	d->free = free_data;
-	d->level = 2;
-	goto done;
-    }
-    case FILE_DELETE:
-    {
-	d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1 + FILENAME_BYTELEN(name) + FILENAME_CHARSIZE);
-	
-	FILENAME_COPY(d->b, name);
-#ifdef USE_VM_PROBES
-	dt_s1 = d->b;
-	dt_utag = name + FILENAME_BYTELEN(name) + FILENAME_CHARSIZE;
-#endif
-	d->command = command;
-	d->invoke = invoke_delete_file;
-	d->free = free_data;
-	d->level = 2;
-	goto done;
-    }
-    case FILE_RENAME:
-	{
-	    char* new_name;
-	    int namelen = FILENAME_BYTELEN(name)+FILENAME_CHARSIZE;
-	    new_name = name+namelen;
-	    d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1
-			      + namelen
-			      + FILENAME_BYTELEN(new_name) + FILENAME_CHARSIZE);
-	
-	    FILENAME_COPY(d->b, name);
-	    FILENAME_COPY(d->b + namelen, new_name);
-#ifdef USE_VM_PROBES
-	    dt_s1 = d->b;
-	    dt_s2 = d->b + namelen;
-	    dt_utag = buf + namelen + FILENAME_BYTELEN(new_name) + FILENAME_CHARSIZE;
-#endif
-	    d->flags = desc->flags;
-	    d->fd = fd;
-	    d->command = command;
-	    d->invoke = invoke_rename;
-	    d->free = free_data;
-	    d->level = 2;
-	    goto done;
-	}
-    case FILE_CHDIR:
-    {
-	d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1 + FILENAME_BYTELEN(name) + FILENAME_CHARSIZE);
-	
-	FILENAME_COPY(d->b, name);
-#ifdef USE_VM_PROBES
-	dt_s1 = d->b;
-	dt_utag = name + FILENAME_BYTELEN(name) + FILENAME_CHARSIZE;
-#endif
-	d->command = command;
-	d->invoke = invoke_chdir;
-	d->free = free_data;
-	d->level = 2;
-	goto done;
-    }
-    case FILE_PWD:
-        {
-	    d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1 + RESBUFSIZE + 1);
-	
-	    d->drive = *(uchar*)buf;
-#ifdef USE_VM_PROBES
-	    dt_utag = buf + 1;
-#endif
-	    d->command = command;
-	    d->invoke = invoke_pwd;
-	    d->free = free_data;
-	    d->level = 2;
-	    goto done;
-	}
-
-    case FILE_READDIR: 
-	if (sys_info.async_threads > 0)
-	{
-	    d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1 + FILENAME_BYTELEN(name) + 
-			      FILENAME_CHARSIZE);
-	
-	    FILENAME_COPY(d->b, name);
-#ifdef USE_VM_PROBES
-	    dt_s1 = d->b;
-	    dt_utag = name + FILENAME_BYTELEN(name) + FILENAME_CHARSIZE;
-#endif
-	    d->dir_handle = NULL;
-	    d->command = command;
-	    d->invoke = invoke_readdir;
-	    d->free = free_readdir;
-	    d->level = 2;
-	    d->c.read_dir.first_buf = NULL;
-	    d->c.read_dir.last_buf = NULL;
-	    goto done;
-	}
-	else   
-	{
-	    size_t resbufsize;
-	    size_t n = 0, total = 0;
-	    int res = 0;
-	    char resbuf[READDIR_BUFSIZE];
-
-	    EFILE_DIR_HANDLE dir_handle; /* Handle to open directory. */
-
-	    total               = READDIR_BUFSIZE;
-	    errInfo.posix_errno = 0;
-	    dir_handle          = NULL;
-	    resbuf[0]           = FILE_RESP_LFNAME;
-
-#ifdef USE_VM_PROBES
-	    dt_s1 = name;
-	    dt_utag = name + FILENAME_BYTELEN(name) + FILENAME_CHARSIZE;
-#endif
-	    /* Fill the buffer with multiple directory listings before sending it to the
-	     * receiving process. READDIR_CHUNKS is minimum number of files sent to the
-	     * receiver.
-	     * Format for each driver_output2:
-	     * ------------------------------------
-	     * | Type   | Len     | Filename  | ...
-	     * | 1 byte | 2 bytes | Len bytes | ...
-	     * ------------------------------------
-	     */
-
-	    do {
-		n = 1;
-		resbufsize = READDIR_BUFSIZE - n;
-
-		do {
-		    res = efile_readdir(&errInfo, name, &dir_handle, resbuf + n + 2, &resbufsize);
-
-		    if (res) {
-			put_int16((Uint16)resbufsize, resbuf + n);
-			n += 2 + resbufsize;
-			resbufsize = READDIR_BUFSIZE - n;
-		    }
-		} while( res && ((total - n - 2) >= MAXPATHLEN*FILENAME_CHARSIZE));
-
-		if (n > 1) {
-		    driver_output2(desc->port, resbuf, 1, resbuf + 1, n - 1);
-		}
-	    } while(res);
-
-	    if (errInfo.posix_errno != 0) {
-		reply_error(desc, &errInfo);
-		return;
-	    }
-#ifdef USE_VM_PROBES
-	    if (dt_utag != NULL && dt_utag[0] == '\0') {
-                dt_utag = NULL;
-            } 
-
-	    DTRACE11(efile_drv_entry, dt_priv->thread_num, dt_priv->tag,
-		     dt_utag, command, name, dt_s2,
-		     dt_i1, dt_i2, dt_i3, dt_i4, desc->port_str);
-	    DTRACE6(efile_drv_return, dt_priv->thread_num, dt_priv->tag++, 
-		    dt_utag, command, 1, 0);
-#endif
-	    TRACE_C('R');
-	    driver_output2(desc->port, resbuf, 1, NULL, 0);
-	    return;
-	}
-    case FILE_OPEN:
-	{
-	    d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1 + FILENAME_BYTELEN(buf+4) + 
-			      FILENAME_CHARSIZE);
-	
-	    d->flags = get_int32((uchar*)buf);
-	    name = buf+4;
-	    FILENAME_COPY(d->b, name);
-#ifdef USE_VM_PROBES
-	    dt_i1 = d->flags;
-	    dt_s1 = d->b;
-	    dt_utag = name + FILENAME_BYTELEN(d->b) + FILENAME_CHARSIZE;
-#endif
-	    d->command = command;
-	    d->invoke = invoke_open;
-	    d->free = free_data;
-	    d->level = 2;
-	    d->is_fd_unused = 1;
-	    goto done;
-	}
-
-    case FILE_FDATASYNC:
-	{
-	    d = EF_SAFE_ALLOC(sizeof(struct t_data));
-	    
-	    d->fd = fd;
-#ifdef USE_VM_PROBES
-	    dt_utag = name;
-	    dt_i1 = fd;
-#endif
-	    d->command = command;
-	    d->invoke = invoke_fdatasync;
-	    d->free = free_data;
-	    d->level = 2;
-	    goto done;
-	}
-
-    case FILE_FSYNC:
-	{
-	    d = EF_SAFE_ALLOC(sizeof(struct t_data));
-	    
-	    d->fd = fd;
-#ifdef USE_VM_PROBES
-	    dt_utag = name;
-	    dt_i1 = fd;
-#endif
-	    d->command = command;
-	    d->invoke = invoke_fsync;
-	    d->free = free_data;
-	    d->level = 2;
-	    goto done;
-	}
-
-
-    case FILE_FSTAT: 
-    case FILE_LSTAT:
-	{
-	    d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1 + FILENAME_BYTELEN(name) + 
-			      FILENAME_CHARSIZE);
-	    
-	    FILENAME_COPY(d->b, name);
-	    d->fd = fd;
-#ifdef USE_VM_PROBES
-	    dt_utag = name + FILENAME_BYTELEN(d->b) + FILENAME_CHARSIZE;
-	    if (command == FILE_LSTAT) {
-		dt_s1 = d->b;
-	    } else {
-		dt_i1 = fd;
-	    }
-#endif
-	    d->command = command;
-	    d->invoke = invoke_flstat;
-	    d->free = free_data;
-	    d->level = 2;
-	    goto done;
-	}
-	
-    case FILE_TRUNCATE:
-        {
-	    d = EF_SAFE_ALLOC(sizeof(struct t_data));
-	    
-	    d->flags = desc->flags;
-	    d->fd = fd;
-#ifdef USE_VM_PROBES
-	    dt_utag = name;
-	    dt_i1 = fd;
-	    dt_i2 = d->flags;
-#endif
-	    d->command = command;
-	    d->invoke = invoke_truncate;
-	    d->free = free_data;
-	    d->level = 2;
-	    goto done;
-	}
-
-    case FILE_WRITE_INFO:
-	{
-	    d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1
-			      + FILENAME_BYTELEN(buf + 9*4) + FILENAME_CHARSIZE);
-	    
-	    d->info.mode       = get_int32(buf + 0 * 4);
-	    d->info.uid        = get_int32(buf + 1 * 4);
-	    d->info.gid        = get_int32(buf + 2 * 4);
-	    d->info.accessTime = get_int64(buf + 3 * 4);
-	    d->info.modifyTime = get_int64(buf + 5 * 4);
-	    d->info.cTime      = get_int64(buf + 7 * 4);
-
-	    FILENAME_COPY(d->b, buf + 9*4);
-#ifdef USE_VM_PROBES
-	    dt_i1              = d->info.mode;
-	    dt_i2              = d->info.uid;
-	    dt_i3              = d->info.gid;
-	    dt_s1 = d->b;
-	    dt_utag = buf + 9 * 4 + FILENAME_BYTELEN(d->b) + FILENAME_CHARSIZE;
-#endif
-	    d->command = command;
-	    d->invoke = invoke_write_info;
-	    d->free = free_data;
-	    d->level = 2;
-	    goto done;
-	}
-
-    case FILE_READLINK:
-	{
-	    d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1 + 
-			      MAX(RESBUFSIZE, (FILENAME_BYTELEN(name) +  
-					       FILENAME_CHARSIZE))  + 1);
-	    FILENAME_COPY(d->b, name);
-#ifdef USE_VM_PROBES
-	    dt_s1 = d->b;
-	    dt_utag = name + FILENAME_BYTELEN(d->b) + FILENAME_CHARSIZE;
-#endif
-	    d->command = command;
-	    d->invoke = invoke_readlink;
-	    d->free = free_data;
-	    d->level = 2;
-	    goto done;
-	}
-
-    case FILE_ALTNAME:
-	{
-	     d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1 + 
-			       MAX(RESBUFSIZE, (FILENAME_BYTELEN(name) +  
-						FILENAME_CHARSIZE))  + 1);
-	    FILENAME_COPY(d->b, name);
-#ifdef USE_VM_PROBES
-	    dt_s1 = d->b;
-	    dt_utag = name + FILENAME_BYTELEN(d->b) + FILENAME_CHARSIZE;
-#endif
-	    d->command = command;
-	    d->invoke = invoke_altname;
-	    d->free = free_data;
-	    d->level = 2;
-	    goto done;
-	}
-
-
-    case FILE_LINK:
-	{
-	    char* new_name;
-	    int namelen = FILENAME_BYTELEN(name) + FILENAME_CHARSIZE;
-
-	    new_name = name+namelen;
-	    d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1
-			      + namelen
-			      + FILENAME_BYTELEN(new_name) + FILENAME_CHARSIZE);
-	
-	    FILENAME_COPY(d->b, name);
-	    FILENAME_COPY(d->b + namelen, new_name);
-#ifdef USE_VM_PROBES
-	    dt_s1 = d->b;
-	    dt_s2 = d->b + namelen;
-	    dt_utag = buf + namelen + FILENAME_BYTELEN(dt_s2) + FILENAME_CHARSIZE;
-#endif
-	    d->flags = desc->flags;
-	    d->fd = fd;
-	    d->command = command;
-	    d->invoke = invoke_link;
-	    d->free = free_data;
-	    d->level = 2;
-	    goto done;
-	}
-
-    case FILE_SYMLINK:
-	{
-	    char* new_name;
-	    int namelen = FILENAME_BYTELEN(name) + FILENAME_CHARSIZE;
-
-	    new_name = name+namelen;
-	    d = EF_SAFE_ALLOC(sizeof(struct t_data) - 1
-			      + namelen
-			      + FILENAME_BYTELEN(new_name) + FILENAME_CHARSIZE);
-	
-	    FILENAME_COPY(d->b, name);
-	    FILENAME_COPY(d->b + namelen, new_name);
-#ifdef USE_VM_PROBES
-	    dt_s1 = d->b;
-	    dt_s2 = d->b + namelen;
-	    dt_utag = buf + namelen + FILENAME_BYTELEN(dt_s2) + FILENAME_CHARSIZE;
-#endif
-	    d->flags = desc->flags;
-	    d->fd = fd;
-	    d->command = command;
-	    d->invoke = invoke_symlink;
-	    d->free = free_data;
-	    d->level = 2;
-	    goto done;
-	}
-
-    case FILE_FADVISE:
-    {
-        d = EF_SAFE_ALLOC(sizeof(struct t_data));
-
-        d->fd = fd;
-        d->command = command;
-        d->invoke = invoke_fadvise;
-        d->free = free_data;
-        d->level = 2;
-        d->c.fadvise.offset = get_int64((uchar*) buf);
-        d->c.fadvise.length = get_int64(((uchar*) buf) + sizeof(Sint64));
-        d->c.fadvise.advise = get_int32(((uchar*) buf) + 2 * sizeof(Sint64));
-#ifdef USE_VM_PROBES
-        dt_i1 = d->fd;
-        dt_i2 = d->c.fadvise.offset;
-        dt_i3 = d->c.fadvise.length;
-        dt_i4 = d->c.fadvise.advise;
-        dt_utag = buf + 3 * sizeof(Sint64);
-#endif
-        goto done;
-    }
-
-    case FILE_FALLOCATE:
-    {
-        d = EF_SAFE_ALLOC(sizeof(struct t_data));
-
-        d->fd = fd;
-        d->command = command;
-        d->invoke = invoke_fallocate;
-        d->free = free_data;
-        d->level = 2;
-        d->c.fallocate.offset = get_int64((uchar*) buf);
-        d->c.fallocate.length = get_int64(((uchar*) buf) + sizeof(Sint64));
-        goto done;
-    }
-
-    }
-
-    /*
-     * Ignore anything else -- let the caller hang.
-     */
-     
-    return;
-
- done:
-    if (d) {
-#ifdef USE_VM_PROBES
-	d->sched_i1 = dt_priv->thread_num;
-	d->sched_i2 = dt_priv->tag;
-	d->sched_utag[0] = '\0';
-	if (dt_utag != NULL) {
-	    if (dt_utag[0] == '\0') {
-		dt_utag = NULL;
-	    } else {
-		strncpy(d->sched_utag, dt_utag, sizeof(d->sched_utag) - 1);
-		d->sched_utag[sizeof(d->sched_utag) - 1] = '\0';
-	    }
-	}
-	DTRACE11(efile_drv_entry, dt_priv->thread_num, dt_priv->tag++,
-		 dt_utag, command, dt_s1, dt_s2,
-		 dt_i1, dt_i2, dt_i3, dt_i4, desc->port_str);
-#endif
-	cq_enq(desc, d);
-    }
-}
-
-/*********************************************************************
- * Driver entry point -> flush
- */
-static void 
-file_flush(ErlDrvData e) {
-    file_descriptor *desc = (file_descriptor *)e;
-#ifdef DEBUG
-    int r;
-#endif
-#ifdef  USE_VM_PROBES
-    dt_private *dt_priv = get_dt_private(dt_driver_io_worker_base);
-#endif
-
-    TRACE_C('f');
-
-#ifdef HAVE_SENDFILE
-    flush_sendfile(desc, NULL);
-#endif
-
-#ifdef DEBUG
-    r = 
-#endif
-         flush_write(desc, NULL
-#ifdef USE_VM_PROBES
-		     , dt_priv, (desc->d == NULL) ? NULL : desc->d->sched_utag
-#endif
-		     );
-    /* Only possible reason for bad return value is ENOMEM, and 
-     * there is nobody to tell...
-     */
-#ifdef DEBUG
-    ASSERT(r == 0); 
-#endif
-    cq_execute(desc);
-}
-
-
-
-/*********************************************************************
- * Driver entry point -> control
- * Only debug functionality...
- */
-static ErlDrvSSizeT
-file_control(ErlDrvData e, unsigned int command, 
-	     char* buf, ErlDrvSizeT len, char **rbuf, ErlDrvSizeT rlen) {
-    file_descriptor *desc = (file_descriptor *)e;
-    switch (command) {
-    case 'K' :
-	if (rlen < 4) {
-	    *rbuf = EF_ALLOC(4);
-	}
-	(*rbuf)[0] = ((desc->key) >> 24) & 0xFF;
-	(*rbuf)[1] = ((desc->key) >> 16) & 0xFF;
-	(*rbuf)[2] = ((desc->key) >> 8) & 0xFF;
-	(*rbuf)[3] = (desc->key) & 0xFF;
-	return 4;
-    default:
-	return 0;
-    }
-}
-
-/*********************************************************************
- * Driver entry point -> timeout
- */
-static void 
-file_timeout(ErlDrvData e) {
-    file_descriptor *desc = (file_descriptor *)e;
-    enum e_timer timer_state = desc->timer_state;
-#ifdef  USE_VM_PROBES
-    dt_private *dt_priv = get_dt_private(dt_driver_io_worker_base);
-#endif
-
-    TRACE_C('t');
-
-    desc->timer_state = timer_idle;
-    switch (timer_state) {
-    case timer_idle:
-	ASSERT(0);
-	break;
-    case timer_again:
-	ASSERT(desc->invoke);
-	ASSERT(desc->free);
-	driver_async(desc->port, KEY(desc), desc->invoke, desc->d, desc->free);
-	break;
-    case timer_write: {
-#ifdef DEBUG
-	int r = 
-#endif
-	         flush_write(desc, NULL
-#ifdef USE_VM_PROBES
-			     , dt_priv, (desc->d == NULL) ? NULL : desc->d->sched_utag
-#endif
-			     );
-	/* Only possible reason for bad return value is ENOMEM, and 
-	 * there is nobody to tell...
-	 */
-	ASSERT(r == 0); 
-	cq_execute(desc);
-    } break;
-    } /* case */
-}
-
-
-
-/*********************************************************************
- * Driver entry point -> outputv
- */
-static void 
-file_outputv(ErlDrvData e, ErlIOVec *ev) {
-    file_descriptor* desc = (file_descriptor*)e;
-    char command;
-    size_t p, q;
-    int err;
-    struct t_data *d = NULL;
-#ifdef USE_VM_PROBES
-    Sint64 dt_i1 = 0, dt_i2 = 0, dt_i3 = 0;
-    Sint64 dt_i4 = 0;
-    char *dt_utag = NULL;
-    char *dt_s1 = NULL;
-    dt_private *dt_priv = get_dt_private(dt_driver_io_worker_base);
-#endif
-
-    TRACE_C('v');
-
-    p = 0; q = 1;
-    if (! EV_GET_CHAR(ev, &command, &p, &q)) {
-	/* Empty command */
-	reply_posix_error(desc, EINVAL);
-	goto done;
-    }
-    /* 'command' contains the decoded command number,
-     * 'p' and 'q' point out the next byte in the command:
-     * ((char *)ev->iov[q].iov_base) + p;
-     */
-    
-    TRACE_F(("%i", (int) command));
-
-    switch (command) {
-
-    case FILE_CLOSE: {
-#ifdef USE_VM_PROBES
-	dt_utag = EV_CHAR_P(ev, p, q);
-#endif
-	flush_read(desc);
-	if (flush_write_check_error(desc, &err
-#ifdef USE_VM_PROBES
-				    , dt_priv, dt_utag
-#endif
-				    ) < 0) {
-	    reply_posix_error(desc, err);
-	    goto done;
-	}
-	if (desc->fd != FILE_FD_INVALID) {
-	    if (! (d = EF_ALLOC(sizeof(struct t_data)))) {
-		reply_posix_error(desc, ENOMEM);
-	    } else {
-		d->command = command;
-		d->reply = !0;
-		d->fd = desc->fd;
-		d->flags = desc->flags;
-#ifdef USE_VM_PROBES
-		dt_i1 = d->fd;
-		dt_i2 = d->flags;
-#endif
-		d->invoke = invoke_close;
-		d->free = free_data;
-		d->level = 2;
-		cq_enq(desc, d);
-		desc->fd = FILE_FD_INVALID;
-		desc->flags = 0;
-	    }
-	} else {
-	    reply_posix_error(desc, EBADF);
-	}
-    } goto done;
-
-    case FILE_READ: {
-	Uint32 sizeH, sizeL;
-	size_t size, alloc_size;
-
-	if (!EV_GET_UINT32(ev, &sizeH, &p, &q)
-	    || !EV_GET_UINT32(ev, &sizeL, &p, &q)) {
-	    /* Wrong buffer length to contain the read count */
-	    reply_posix_error(desc, EINVAL);
-	    goto done;
-	}
-#ifdef USE_VM_PROBES
-	dt_utag = EV_CHAR_P(ev, p, q);
-#endif
-	if (flush_write_check_error(desc, &err
-#ifdef USE_VM_PROBES
-				    , dt_priv, dt_utag
-#endif
-				    ) < 0) {
-	    reply_posix_error(desc, err);
-	    goto done;
-	}
-#if ALWAYS_READ_LINE_AHEAD
-	if (desc->read_bufsize == 0 && desc->read_binp != NULL && desc->read_size > 0) {
-	    /* We have allocated a buffer for line mode but should not really have a 
-	       read-ahead buffer... */
-	    if (lseek_flush_read(desc, &err
-#ifdef USE_VM_PROBES
-				 , dt_priv, dt_utag
-#endif
-				 ) < 0) {
-		reply_posix_error(desc, err);
-		goto done;
-	    }
-	}
-#endif
-#if SIZEOF_SIZE_T == 4
-	if (sizeH != 0) {
-	    reply_posix_error(desc, EINVAL);
-	    goto done;
-	}
-	size = sizeL;
-#else
-	size = ((size_t)sizeH << 32) | sizeL;
-#endif
-	if ((desc->fd == FILE_FD_INVALID)
-	    || (! (desc->flags & EFILE_MODE_READ)) ) {
-	    reply_posix_error(desc, EBADF);
-	    goto done;
-	}
-	if (size == 0) {
-	    reply_buf(desc, &command, 0);
-	    goto done;
-	}
-	if (desc->read_size >= size) {
-	    /* We already have all data */
-	    TRACE_C('D');
-	    reply_data(desc, desc->read_binp, desc->read_offset, size);
-	    desc->read_offset += size;
-	    desc->read_size -= size;
-	    try_free_read_bin(desc);
-	    goto done;
-	}
-	/* We may have some of the data 
-	 */
-	/* Justification for the following strange formula:
-	 * If the read request is for such a large block as more than 
-	 * half the buffer size it may lead to a lot of unnecessary copying, 
-	 * since the tail of the old buffer is copied to the head of the
-	 * new, and if the tail is almost half the buffer it is a lot
-	 * to copy. Therefore allocate the exact amount needed in 
-	 * this case, giving no lingering tail. */
-	alloc_size = 
-	    size > (desc->read_bufsize>>1) ? 
-	    size : desc->read_bufsize;
-	if (! desc->read_binp) {
-	    /* Need to allocate a new binary for the result */
-	    if (! (desc->read_binp = driver_alloc_binary(alloc_size))) {
-		reply_posix_error(desc, ENOMEM);
-		goto done;
-	    }
-	} else {
-	    /* We already have a buffer */
-	    if (desc->read_binp->orig_size - desc->read_offset < size) {
-		/* Need to allocate a new binary for the result */
-		ErlDrvBinary *binp;
-		if (! (binp = driver_alloc_binary(alloc_size))) {
-		    reply_posix_error(desc, ENOMEM);
-		    goto done;
-		}
-		/* Move data we already have to the new binary */
-		sys_memcpy(binp->orig_bytes, 
-			   desc->read_binp->orig_bytes + desc->read_offset,
-			   desc->read_size);
-		driver_free_binary(desc->read_binp);
-		desc->read_offset = 0;
-		desc->read_binp = binp;
-	    }
-	} 
-	if (! (d = EF_ALLOC(sizeof(struct t_data)))) {
-	    reply_posix_error(desc, ENOMEM);
-	    goto done;
-	}
-	d->command = command;
-	d->reply = !0;
-	d->fd = desc->fd;
-	d->flags = desc->flags;
-	d->c.read.binp = desc->read_binp;
-	d->c.read.bin_offset = desc->read_offset + desc->read_size;
-	d->c.read.bin_size = desc->read_binp->orig_size - d->c.read.bin_offset;
-	d->c.read.size = size;
-#ifdef USE_VM_PROBES
-	dt_i1 = d->fd;
-	dt_i2 = d->flags;
-	dt_i3 = d->c.read.size;
-#endif
-	driver_binary_inc_refc(d->c.read.binp);
-	d->invoke = invoke_read;
-	d->free = free_read;
-	d->level = 1;
-	cq_enq(desc, d);
-    } goto done; /* case FILE_READ: */
-
-    case FILE_READ_LINE: {
-	/*
-	 * Icky little creature... We do mostly as ordinary file read, but with a few differences.
-	 * 1) We have to scan for proper newline sequence if there is a buffer already, we cannot know 
-	 *    in advance if the buffer contains a whole line without scanning.
-	 * 2) We do not know how large the buffer needs to be in advance. We give a default buffer,
-	 *    but the worker may need to allocate a new one. Freeing the old and rereferencing a newly 
-	 *    allocated binary + dealing with offsets and lengts are done in file_async ready
-	 *    for this OP.
-	 */
-#ifdef USE_VM_PROBES
-	dt_utag = EV_CHAR_P(ev, p, q);
-#endif
-	if (flush_write_check_error(desc, &err
-#ifdef USE_VM_PROBES
-				    , dt_priv, dt_utag
-#endif
-				    ) < 0) {
-	    reply_posix_error(desc, err);
-	    goto done;
-	}
-	if (ev->size != 1
-#ifdef USE_VM_PROBES
-	    + FILENAME_BYTELEN(dt_utag) + FILENAME_CHARSIZE
-#endif
-	    ) {
-	    /* Wrong command length */
-	    reply_posix_error(desc, EINVAL);
-	    goto done;
-	}
-	if ((desc->fd == FILE_FD_INVALID)
-	    || (! (desc->flags & EFILE_MODE_READ)) ) {
-	    reply_posix_error(desc, EBADF);
-	    goto done;
-	}
-	if (desc->read_size > 0) {
-	    /* look for '\n' in what we'we already got */
-	    void *nl_ptr = memchr(desc->read_binp->orig_bytes + desc->read_offset,'\n',desc->read_size);
-	    if (nl_ptr != NULL) {
-		/* If found, we're done */
-		int skip = 0;
-		size_t size = ((char *) nl_ptr) - 
-		    ((char *) (desc->read_binp->orig_bytes + desc->read_offset)) + 1;
-		if (size > 1 &&
-		    *(((char *) nl_ptr) - 1) == '\r') {
-		    *(((char *) nl_ptr) - 1) = '\n';		    
-		    skip = 1;
-		    --size;
-		} 
-		reply_data(desc, desc->read_binp, desc->read_offset, size);
-		desc->read_offset += (size + skip);
-		desc->read_size -= (size + skip);
-		try_free_read_bin(desc);
-		goto done;
-	    }
-	}
-	/* Now, it's up to the thread to work out the need for more buffers and such, it's
-	   no use doing it in this thread as we do not have the information required anyway. 
-	   Even a NULL buffer could be handled by the thread, but code is simplified by us 
-	   allocating it */
-	if (! desc->read_binp) {
-	    int alloc_size = (desc->read_bufsize > DEFAULT_LINEBUF_SIZE) ? desc->read_bufsize : 
-		DEFAULT_LINEBUF_SIZE;
-	    /* Allocate a new binary for the result */
-	    if (! (desc->read_binp = driver_alloc_binary(alloc_size))) {
-		reply_posix_error(desc, ENOMEM);
-		goto done;
-	    }
-	}	
-	if (! (d = EF_ALLOC(sizeof(struct t_data)))) {
-	    reply_posix_error(desc, ENOMEM);
-	    goto done;
-	}
-
-	d->command = command;
-	d->reply = !0;
-	d->fd = desc->fd;
-	d->flags = desc->flags;
-	d->c.read_line.binp = desc->read_binp;
-	d->c.read_line.read_offset = desc->read_offset;
-	d->c.read_line.read_size = desc->read_size;
-#ifdef USE_VM_PROBES
-	dt_i1 = d->fd;
-	dt_i2 = d->flags;
-	dt_i3 = d->c.read_line.read_offset;
-#endif
-#if !ALWAYS_READ_LINE_AHEAD
-	d->c.read_line.read_ahead = (desc->read_bufsize > 0);
-#ifdef USE_VM_PROBES
-	dt_i4 = d->c.read_line.read_ahead;
-#endif
-#endif 
-	driver_binary_inc_refc(d->c.read.binp);
-	d->invoke = invoke_read_line;
-	d->free = free_read_line;
-	d->level = 1;
-	cq_enq(desc, d);
-    } goto done;
-    case FILE_WRITE: { /* Dtrace: The dtrace user tag is not last in message, 
-			  but follows the message tag directly. 
-			  This is handled specially in prim_file.erl */
-	ErlDrvSizeT skip = 1;
-	ErlDrvSizeT size = ev->size - skip;
-
-#ifdef USE_VM_PROBES
-	dt_utag = EV_CHAR_P(ev, p, q);
-	skip += FILENAME_BYTELEN(dt_utag) + FILENAME_CHARSIZE;
-	size = ev->size - skip;
-#endif
-	if (lseek_flush_read(desc, &err
-#ifdef USE_VM_PROBES
-			     , dt_priv, dt_utag
-#endif
-			     ) < 0) {
-	    reply_posix_error(desc, err);
-	    goto done;
-	}
-	if (! (desc->flags & EFILE_MODE_WRITE)) {
-	    reply_posix_error(desc, EBADF);
-	    goto done;
-	}
-	if (size == 0) {
-	    reply_Uint(desc, size);
-	    goto done;
-	}
-	MUTEX_LOCK(desc->q_mtx);
-	if (driver_enqv(desc->port, ev, skip)) {
-	    MUTEX_UNLOCK(desc->q_mtx);
-	    reply_posix_error(desc, ENOMEM);
-	    goto done;
-	}
-	desc->write_buffered += size;
-	if (desc->write_buffered < desc->write_bufsize) {
-	    MUTEX_UNLOCK(desc->q_mtx);
-	    reply_Uint(desc, size);
-	    if (desc->timer_state == timer_idle) {
-		desc->timer_state = timer_write;
-		driver_set_timer(desc->port, desc->write_delay);
-	    }
-	} else {
-	    if ((d = async_write(desc, &err, !0, size
-#ifdef USE_VM_PROBES
-				 , &dt_i1, &dt_i2, &dt_i3
-#endif
-				 )) == NULL) {
-		MUTEX_UNLOCK(desc->q_mtx);
-		reply_posix_error(desc, err);
-		goto done;
-	    } else {
-		MUTEX_UNLOCK(desc->q_mtx);
-	    }
-	}
-    } goto done; /* case FILE_WRITE */
-
-    case FILE_PWRITEV: { /* Dtrace: The dtrace user tag is not last in message, 
-			   but follows the message tag directly. 
-			   This is handled specially in prim_file.erl */
-	Uint32 i, j, n; 
-	size_t total;
-#ifdef USE_VM_PROBES
-	char dt_tmp;
-	int dt_utag_bytes = 1;
-
-	dt_utag = EV_CHAR_P(ev, p, q);
-	/* This will work for UTF-8, but not for UTF-16 - extra reminder here */
-#ifdef FILENAMES_16BIT 
-#error 16bit characters in filenames and dtrace in combination is not supported.
-#endif
-	while (EV_GET_CHAR(ev, &dt_tmp, &p, &q) && dt_tmp != '\0') {
-	    dt_utag_bytes++;
-	}
-#endif
-	if (ev->size < 1+4
-#ifdef USE_VM_PROBES
-	    + dt_utag_bytes
-#endif
-	    || !EV_GET_UINT32(ev, &n, &p, &q)) {
-	    /* Buffer too short to contain even the number of pos/size specs */
-	    reply_Uint_posix_error(desc, 0, EINVAL);
-	    goto done;
-	}
-	if (lseek_flush_read(desc, &err
-#ifdef USE_VM_PROBES
-			     , dt_priv, dt_utag
-#endif
-			     ) < 0) {
-	    reply_Uint_posix_error(desc, 0, err);
-	    goto done;
-	}
-	if (flush_write_check_error(desc, &err
-#ifdef USE_VM_PROBES
-				    , dt_priv, dt_utag
-#endif
-				    ) < 0) {
-	    reply_Uint_posix_error(desc, 0, err);
-	    goto done;
-	}
-	if (n == 0) {
-	    /* Trivial case - nothing to write */
-	    if (ev->size != 1+4) {
-		reply_posix_error(desc, err);
-	    } else {
-		reply_Uint(desc, 0);
-	    }
-	    goto done;
-	}
-	if (ev->size < 1+4+8*(2*n)
-#ifdef USE_VM_PROBES
-	    + dt_utag_bytes
-#endif
-	    ) {
-	    /* Buffer too short to contain even the pos/size specs */
-	    reply_Uint_posix_error(desc, 0, EINVAL);
-	    goto done;
-	}
-	d = EF_ALLOC(sizeof(struct t_data) 
-		     + (n * sizeof(struct t_pbuf_spec)));
-	if (! d) {
-	    reply_Uint_posix_error(desc, 0, ENOMEM);
-	    goto done;
-	}
-	d->command = command;
-	d->reply = !0;
-	d->fd = desc->fd;
-	d->flags = desc->flags;
-#ifdef USE_VM_PROBES
-	dt_i1 = d->fd;
-	dt_i2 = d->flags;
-#endif
-	d->c.pwritev.port = desc->port;
-	d->c.pwritev.q_mtx = desc->q_mtx;
-	d->c.pwritev.n = n;
-	d->c.pwritev.cnt = 0;
-	total = 0;
-	j = 0;
-	/* Create pos/size specs in the thread data structure
-	 * for all non-zero size binaries. Calculate total size.
-	 */
-	for(i = 0; i < n; i++) {
-	    Uint32 sizeH, sizeL;
-	    size_t size;
-	    if (   !EV_GET_SINT64(ev, &d->c.pwritev.specs[i].offset, &p, &q)
-		|| !EV_GET_UINT32(ev, &sizeH, &p, &q)
-		|| !EV_GET_UINT32(ev, &sizeL, &p, &q)) {
-		/* Misalignment in buffer */
-		reply_Uint_posix_error(desc, 0, EINVAL);
-		EF_FREE(d);
-		goto done;
-	    }
-#if SIZEOF_SIZE_T == 4
-	    if (sizeH != 0) {
-		reply_Uint_posix_error(desc, 0, EINVAL);
-		EF_FREE(d);
-		goto done;
-	    }
-	    size = sizeL;
-#else
-	    size = ((size_t)sizeH<<32) | sizeL;
-#endif
-	    if (size > 0) {
-		total += size;
-		d->c.pwritev.specs[j].size = size;
-		j++;
-	    }
-	}
-	d->c.pwritev.size = total;
-#ifdef USE_VM_PROBES
-	dt_i3 = d->c.pwritev.size;
-#endif
-	if (j == 0) {
-	    /* Trivial case - nothing to write */
-	    EF_FREE(d);
-	    reply_Uint(desc, 0);
-	} else {
-	    ErlDrvSizeT skip = 1 + 4 + 8 * (2*n) 
-#ifdef USE_VM_PROBES
-		+ dt_utag_bytes
-#endif
-		;
-	    if (skip + total != ev->size) {
-		/* Actual amount of data does not match 
-		 * total of all pos/size specs
-		 */
-		EF_FREE(d);
-		reply_Uint_posix_error(desc, 0, EINVAL);
-	    } else {
-		/* Enqueue the data */
-		MUTEX_LOCK(desc->q_mtx);
-		driver_enqv(desc->port, ev, skip);
-		MUTEX_UNLOCK(desc->q_mtx);
-		/* Execute the command */
-		d->invoke = invoke_pwritev;
-		d->free = free_data;
-		d->level = 1;
-		cq_enq(desc, d);
-	    }
-	}
-    } goto done; /* case FILE_PWRITEV: */
-
-    case FILE_PREADV: { /* Dtrace: The dtrace user tag is not last in message, 
-			   but follows the message tag directly. 
-			   This is handled specially in prim_file.erl */
-	register void * void_ptr;
-	Uint32 i, n;
-	ErlIOVec *res_ev;
-#ifdef USE_VM_PROBES
-	char dt_tmp;
-	int dt_utag_bytes = 1;
-	/* This will work for UTF-8, but not for UTF-16 - extra reminder here */
-#ifdef FILENAMES_16BIT 
-#error 16bit characters in filenames and dtrace in combination is not supported.
-#endif
-	dt_utag = EV_CHAR_P(ev, p, q);
-	while (EV_GET_CHAR(ev, &dt_tmp, &p, &q) && dt_tmp != '\0') {
-	    dt_utag_bytes++;
-	}
-#endif
-	if (lseek_flush_read(desc, &err
-#ifdef USE_VM_PROBES
-			     , dt_priv, dt_utag
-#endif
-			     ) < 0) {
-	    reply_posix_error(desc, err);
-	    goto done;
-	}
-	if (flush_write_check_error(desc, &err
-#ifdef USE_VM_PROBES
-				    , dt_priv, dt_utag
-#endif
-				    ) < 0) {
-	    reply_posix_error(desc, err);
-	    goto done;
-	}
-	if (ev->size < 1+8
-#ifdef USE_VM_PROBES
-	    + dt_utag_bytes
-#endif
-	    || !EV_GET_UINT32(ev, &n, &p, &q)
-	    || !EV_GET_UINT32(ev, &n, &p, &q)) {
-	    /* Buffer too short to contain even the number of pos/size specs */
-	    reply_posix_error(desc, EINVAL);
-	    goto done;
-	}
-	if (ev->size < 1+8+8*(2*n)
-#ifdef USE_VM_PROBES
-	    + dt_utag_bytes
-#endif
-	    ) {
-	    /* Buffer wrong length to contain the pos/size specs */
-	    reply_posix_error(desc, EINVAL);
-	    goto done;
-	}
-	/* Create the thread data structure with the contained ErlIOVec 
-	 * and corresponding binaries for the response 
-	 */
-	d = EF_ALLOC(sizeof(*d)
-		     + (n * sizeof(*d->c.preadv.offsets))
-		     + ((1+n) * (sizeof(*res_ev->iov)
-				 + sizeof(*res_ev->binv))));
-	if (! d) {
-	    reply_posix_error(desc, ENOMEM);
-	    goto done;
-	}
-	d->command = command;
-	d->reply = !0;
-	d->fd = desc->fd;
-	d->flags = desc->flags;
-#ifdef USE_VM_PROBES
-	dt_i1 = d->fd;
-	dt_i2 = d->flags;
-#endif
-	d->c.preadv.n = n;
-	d->c.preadv.cnt = 0;
-	d->c.preadv.size = 0;
-	res_ev = &d->c.preadv.eiov;
-	/* XXX possible alignment problems here for weird machines */
-	res_ev->vsize = 1+d->c.preadv.n;
-	res_ev->iov = void_ptr = &d->c.preadv.offsets[d->c.preadv.n];
-	res_ev->binv = void_ptr = &res_ev->iov[res_ev->vsize];
-	/* Read in the pos/size specs and allocate binaries for the results */
-	for (i = 1; i < 1+n; i++) {
-	    Uint32 sizeH, sizeL;
-	    size_t size;
-	    if (   !EV_GET_SINT64(ev, &d->c.preadv.offsets[i-1], &p, &q)
-		|| !EV_GET_UINT32(ev, &sizeH, &p, &q)
-		|| !EV_GET_UINT32(ev, &sizeL, &p, &q)) {
-		reply_posix_error(desc, EINVAL);
-		break;
-	    }
-#if SIZEOF_SIZE_T == 4
-	    if (sizeH != 0) {
-		reply_posix_error(desc, EINVAL);
-		break;
-	    }
-	    size = sizeL;
-#else
-	    size = ((size_t)sizeH<<32) | sizeL;
-#endif
-#ifdef USE_VM_PROBES
-	    dt_i3 += size;
-#endif
-	    if (! (res_ev->binv[i] = driver_alloc_binary(size))) {
-		reply_posix_error(desc, ENOMEM);
-		break;
-	    } else {
-		res_ev->iov[i].iov_len  = size;
-		res_ev->iov[i].iov_base = res_ev->binv[i]->orig_bytes;
-	    }
-	}
-	if (i < 1+n) {
-	    for (i--; i > 0; i--) {
-		driver_free_binary(res_ev->binv[i]);
-	    }
-	    EF_FREE(d);
-	    goto done;
-	}
-	/* Allocate the header binary (index 0) */
-	res_ev->binv[0] = driver_alloc_binary(4+4+8*n);
-	if (! res_ev->binv[0]) {
-	    reply_posix_error(desc, ENOMEM);
-	    for (i = 1; i < 1+n; i++) {
-		driver_free_binary(res_ev->binv[i]);
-	    }
-	    EF_FREE(d);
-	    goto done;
-	}
-	res_ev->iov[0].iov_len = 4+4+8*n;
-	res_ev->iov[0].iov_base = res_ev->binv[0]->orig_bytes;
-	/* Fill in the number of buffers in the header */
-	put_int32(0, res_ev->iov[0].iov_base);
-	put_int32(n, (char *)(res_ev->iov[0].iov_base) + 4);
-	/**/
-	res_ev->size = res_ev->iov[0].iov_len;
-	if (n == 0) {
-	    /* Trivial case - nothing to read */
-	    reply_ev(desc, FILE_RESP_LDATA, res_ev);
-	    free_preadv(d);
-	    goto done;
-	} else {
-	    d->invoke = invoke_preadv;
-	    d->free = free_preadv;
-	    d->level = 1;
-	    cq_enq(desc, d);
-	}
-    } goto done; /* case FILE_PREADV: */
-
-    case FILE_LSEEK: {
-	Sint64 offset;		/* Offset for seek */
-	Uint32 origin;		/* Origin of seek. */
-
-	if (ev->size < 1+8+4
-	    || !EV_GET_SINT64(ev, &offset, &p, &q)
-	    || !EV_GET_UINT32(ev, &origin, &p, &q)) {
-	    /* Wrong length of buffer to contain offset and origin */
-	    reply_posix_error(desc, EINVAL);
-	    goto done;
-	}
-#ifdef USE_VM_PROBES
-	dt_utag = EV_CHAR_P(ev, p, q);
-#endif
-	if (lseek_flush_read(desc, &err
-#ifdef USE_VM_PROBES
-			     , dt_priv, dt_utag
-#endif
-			     ) < 0) {
-	    reply_posix_error(desc, err);
-	    goto done;
-	}
-	if (flush_write_check_error(desc, &err
-#ifdef USE_VM_PROBES
-				    , dt_priv, dt_utag
-#endif
-				    ) < 0) {
-	    reply_posix_error(desc, err);
-	    goto done;
-	}
-	if ((d = async_lseek(desc, &err, !0, offset, origin
-#ifdef USE_VM_PROBES
-			     , &dt_i1, &dt_i2, &dt_i3
-#endif
-			     )) == NULL) {
-	    reply_posix_error(desc, err);
-	    goto done;
-	}
-    } goto done;
-
-    case FILE_READ_FILE: {
-	char *filename;
-	if (ev->size < 1+1) {
-	    /* Buffer contains empty name */
-	    reply_posix_error(desc, ENOENT);
-	    goto done;
-	}
-#ifndef USE_VM_PROBES
-	/* In the dtrace case, the iov has an extra element, the dtrace utag - we will need 
-	   another test to see that
-	   the filename is in a single buffer: */
-	if (ev->size-1 != ev->iov[q].iov_len-p) {
-	    /* Name not in one single buffer */
-	    reply_posix_error(desc, EINVAL);
-	    goto done;
-	}
-#else
-	if (((byte *)ev->iov[q].iov_base)[ev->iov[q].iov_len-1] != '\0') {
-	    /* Name not in one single buffer */
-	    reply_posix_error(desc, EINVAL);
-	    goto done;
-	}	
-#endif
-	filename = EV_CHAR_P(ev, p, q);
-	d = EF_ALLOC(sizeof(struct t_data) -1 + FILENAME_BYTELEN(filename) + FILENAME_CHARSIZE);
-	if (! d) {
-	    reply_posix_error(desc, ENOMEM);
-	    goto done;
-	}
-	d->command = command;
-	d->reply = !0;
-	/* Copy name */
-	FILENAME_COPY(d->b, filename);
-#ifdef USE_VM_PROBES
-	{
-	    char dt_tmp;
-
-	    /* This will work for UTF-8, but not for UTF-16 - extra reminder here */
-#ifdef FILENAMES_16BIT 
-#error 16bit characters in filenames and dtrace in combination is not supported.
-#endif
-	    while (EV_GET_CHAR(ev, &dt_tmp, &p, &q) && dt_tmp != '\0') 
-		;
-	    dt_s1 = d->b;
-	    dt_utag = EV_CHAR_P(ev, p, q);
-	}
-#endif
-	d->c.read_file.binp = NULL;
-	d->invoke = invoke_read_file;
-	d->free = free_read_file;
-	d->level = 2;
-	cq_enq(desc, d);
-    } goto done;
-
-    case FILE_IPREAD: {
-	/* This operation cheets by using invoke_preadv() and free_preadv()
-	 * plus its own invoke_ipread. Therefore the result format is 
-	 * a bit awkward - the header binary contains one extra 64 bit
-	 * field that invoke_preadv() fortunately ignores,
-	 * and the first 64 bit field does not contain the number of 
-	 * data binaries which invoke_preadv() also ignores.
-	 */
-	register void * void_ptr;
-	char mode;
-	Sint64 hdr_offset;
-	Uint32 max_size;
-	ErlIOVec *res_ev;
-	int vsize;
-	if (! EV_GET_CHAR(ev, &mode, &p, &q)) {
-	    /* Empty command */
-	    reply_posix_error(desc, EINVAL);
-	    goto done;
-	}
-	if (mode != IPREAD_S32BU_P32BU) {
-	    reply_posix_error(desc, EINVAL);
-	    goto done;
-	}
-	if (ev->size < 1+1+8+4
-	    || !EV_GET_SINT64(ev, &hdr_offset, &p, &q)
-	    || !EV_GET_UINT32(ev, &max_size, &p, &q)) {
-	    /* Buffer too short to contain 
-	     * the header offset and max size spec */
-	    reply_posix_error(desc, EINVAL);
-	    goto done;
-	}
-#ifdef USE_VM_PROBES
-	dt_utag = EV_CHAR_P(ev, p, q);
-#endif
-	if (lseek_flush_read(desc, &err
-#ifdef USE_VM_PROBES
-			     , dt_priv, dt_utag
-#endif
-			     ) < 0) {
-	    reply_posix_error(desc, err);
-	    goto done;
-	}
-	if (flush_write_check_error(desc, &err
-#ifdef USE_VM_PROBES
-				    , dt_priv, dt_utag
-#endif
-				    ) < 0) {
-	    reply_posix_error(desc, err);
-	    goto done;
-	}
-	/* Create the thread data structure with the contained ErlIOVec 
-	 * and corresponding binaries for the response 
-	 */
-	vsize = 2;
-	d = EF_ALLOC(sizeof(*d) + 
-		     vsize*(sizeof(*res_ev->iov) + sizeof(*res_ev->binv)));
-	if (! d) {
-	    reply_posix_error(desc, ENOMEM);
-	    goto done;
-	}
-	d->command = command;
-	d->reply = !0;
-	d->fd = desc->fd;
-	d->flags = desc->flags;
-	d->c.preadv.offsets[0] = hdr_offset;
-	d->c.preadv.size = max_size;
-#ifdef USE_VM_PROBES
-	dt_i1 = d->fd;
-	dt_i2 = d->flags;
-	dt_i3 = d->c.preadv.offsets[0];
-	dt_i4 = d->c.preadv.size;
-#endif
-	res_ev = &d->c.preadv.eiov;
-	/* XXX possible alignment problems here for weird machines */
-	res_ev->iov = void_ptr = d + 1;
-	res_ev->binv = void_ptr = res_ev->iov + vsize;
-	res_ev->size = 0;
-	res_ev->vsize = 0;
-	d->invoke = invoke_ipread;
-	d->free = free_preadv;
-	d->level = 1;
-	cq_enq(desc, d);
-    } goto done; /* case FILE_IPREAD: */
-
-    case FILE_SETOPT: {
-	char opt;
-
-	if (ev->size < 1+1
-	    || !EV_GET_CHAR(ev, &opt, &p, &q)) {
-	    /* Buffer too short to contain even the option type */
-	    reply_posix_error(desc, EINVAL);
-	    goto done;
-	}
-#ifdef USE_VM_PROBES
-	dt_i1 = opt;
-	dt_utag = EV_CHAR_P(ev, p, q);
-#endif
-	switch (opt) {
-	case FILE_OPT_DELAYED_WRITE: {
-	    Uint32 sizeH, sizeL, delayH, delayL;
-	    if (ev->size != 1+1+4*sizeof(Uint32)
-#ifdef USE_VM_PROBES
-		+ FILENAME_BYTELEN(dt_utag) + FILENAME_CHARSIZE
-#endif
-		|| !EV_GET_UINT32(ev, &sizeH, &p, &q)
-		|| !EV_GET_UINT32(ev, &sizeL, &p, &q)
-		|| !EV_GET_UINT32(ev, &delayH, &p, &q)
-		|| !EV_GET_UINT32(ev, &delayL, &p, &q)) {
-		/* Buffer has wrong length to contain the option values */
-		reply_posix_error(desc, EINVAL);
-		goto done;
-	    }
-#if SIZEOF_SIZE_T == 4
-	    if (sizeH != 0) {
-		reply_posix_error(desc, EINVAL);
-		goto done;
-	    }
-	    desc->write_bufsize = sizeL;
-#else
-	    desc->write_bufsize = ((size_t)sizeH << 32) | sizeL;
-#endif
-#if SIZEOF_LONG == 4
-	    if (delayH != 0) {
-		reply_posix_error(desc, EINVAL);
-		goto done;
-	    }
-	    desc->write_delay = delayL;
-#else
-	    desc->write_delay = ((unsigned long)delayH << 32) | delayL;
-#endif
-#ifdef USE_VM_PROBES
-	    dt_i2 = desc->write_delay;
-#endif
-	    TRACE_C('K');
-	    reply_ok(desc);
-	} goto done;
-	case FILE_OPT_READ_AHEAD: {
-	    Uint32 sizeH, sizeL;
-	    if (ev->size != 1+1+2*sizeof(Uint32)
-#ifdef USE_VM_PROBES
-		+ FILENAME_BYTELEN(dt_utag)+FILENAME_CHARSIZE
-#endif
-		|| !EV_GET_UINT32(ev, &sizeH, &p, &q)
-		|| !EV_GET_UINT32(ev, &sizeL, &p, &q)) {
-		/* Buffer has wrong length to contain the option values */
-		reply_posix_error(desc, EINVAL);
-		goto done;
-	    }
-#if SIZEOF_SIZE_T == 4
-	    if (sizeH != 0) {
-		reply_posix_error(desc, EINVAL);
-		goto done;
-	    }
-	    desc->read_bufsize = sizeL;
-#else
-	    desc->read_bufsize = ((size_t)sizeH << 32) | sizeL;
-#endif
-#ifdef USE_VM_PROBES
-	    dt_i2 = desc->read_bufsize;
-#endif
-	    TRACE_C('K');
-	    reply_ok(desc);
-	} goto done;
-	default:
-	    reply_posix_error(desc, EINVAL);
-	    goto done;
-	} /* case FILE_OPT_DELAYED_WRITE: */
-    } ASSERT(0); goto done; /* case FILE_SETOPT: */
-
-    case FILE_SENDFILE: {
-
-#ifdef HAVE_SENDFILE
-        struct t_data *d;
-	Uint32 out_fd, offsetH, offsetL, hd_len, tl_len;
-	Uint64 nbytes;
-	char flags;
-
-	if (ev->size < 1 + 7 * sizeof(Uint32) + sizeof(char)
-		|| !EV_GET_UINT32(ev, &out_fd, &p, &q)
-		|| !EV_GET_CHAR(ev, &flags, &p, &q)
-		|| !EV_GET_UINT32(ev, &offsetH, &p, &q)
-		|| !EV_GET_UINT32(ev, &offsetL, &p, &q)
-		|| !EV_GET_UINT64(ev, &nbytes, &p, &q)
-		|| !EV_GET_UINT32(ev, &hd_len, &p, &q)
-		|| !EV_GET_UINT32(ev, &tl_len, &p, &q)) {
-	    /* Buffer has wrong length to contain all the needed values */
-	    reply_posix_error(desc, EINVAL);
-	    goto done;
-	}
-
-	if (hd_len != 0 || tl_len != 0) {
-	    /* We do not allow header, trailers */
-	    reply_posix_error(desc, EINVAL);
-	    goto done;
-	}
-
-	
-	if (flags & SENDFILE_FLGS_USE_THREADS && !THRDS_AVAILABLE) {
-	    /* We do not allow use_threads flag on a system where
-	       no threads are available. */
-	    reply_posix_error(desc, EINVAL);
-	    goto done;
-	}
-
-	d = EF_SAFE_ALLOC(sizeof(struct t_data));
-	d->fd = desc->fd;
-	d->command = command;
-	d->invoke = invoke_sendfile;
-	d->free = free_sendfile;
-	d->flags = flags;
-	d->level = 2;
-
-	d->c.sendfile.out_fd = (int) out_fd;
-	d->c.sendfile.written = 0;
-	d->c.sendfile.port = desc->port;
-	d->c.sendfile.q_mtx = desc->q_mtx;
-
-    #if SIZEOF_OFF_T == 4
-	if (offsetH != 0) {
-	    reply_posix_error(desc, EINVAL);
-	    goto done;
-	}
-	d->c.sendfile.offset = (off_t) offsetL;
-    #else
-	d->c.sendfile.offset = ((off_t) offsetH << 32) | offsetL;
-    #endif
-
-	d->c.sendfile.nbytes = nbytes;
-
-	if (USE_THRDS_FOR_SENDFILE(d)) {
-	    SET_BLOCKING(d->c.sendfile.out_fd);
-	} else {
-	    /**
-	     * Write a place holder to queue in order to force file_flush
-	     * to be called before the driver is closed.
-	     */
-	    char tmp[1] = "";
-	    MUTEX_LOCK(d->c.sendfile.q_mtx);
-	    if (driver_enq(d->c.sendfile.port, tmp, 1)) {
-	        MUTEX_UNLOCK(d->c.sendfile.q_mtx);
-		reply_posix_error(desc, ENOMEM);
-		goto done;
-	    }
-	    MUTEX_UNLOCK(d->c.sendfile.q_mtx);
-	}
-
-	cq_enq(desc, d);
-#else
-	reply_posix_error(desc, ENOTSUP);
-#endif
-	goto done;
-        } /* case FILE_SENDFILE: */
-
-    } /* switch(command) */
-
-    if (lseek_flush_read(desc, &err
-#ifdef USE_VM_PROBES
-			 , dt_priv, dt_utag
-#endif
-			 ) < 0) {
-	reply_posix_error(desc, err);
-	goto done;
-    }
-    if (flush_write_check_error(desc, &err
-#ifdef USE_VM_PROBES
-				, dt_priv, dt_utag
-#endif
-				) < 0) {
-	reply_posix_error(desc, err);
-	goto done;
-    } else {
-	/* Flatten buffer and send it to file_output(desc, buf, len) */
-	int len = ev->size;
-	char *buf = EF_ALLOC(len);
-	if (! buf) {
-	    reply_posix_error(desc, ENOMEM);
-	    goto done;
-	}
-	driver_vec_to_buf(ev, buf, len);
-	file_output((ErlDrvData) desc, buf, len);
-	EF_FREE(buf);
-	goto done;
-    }
-
- done:
-    if (d != NULL) {
-#ifdef USE_VM_PROBES
-	/*
-	 * If d == NULL, then either:
-	 *    1). There was an error of some sort, or
-	 *    2). The command given to us is actually implemented
-	 *	  by file_output() instead.
-	 *
-	 * Case #1 is probably a TODO item, perhaps?
-	 * Case #2 we definitely don't want to activate a probe.
-	 */
-	d->sched_i1 = dt_priv->thread_num;
-	d->sched_i2 = dt_priv->tag;
-	d->sched_utag[0] = '\0';
-	if (dt_utag != NULL) {
-	    if (dt_utag[0] == '\0') {
-                dt_utag = NULL;
-            } else {
-		strncpy(d->sched_utag, dt_utag, sizeof(d->sched_utag) - 1);
-		d->sched_utag[sizeof(d->sched_utag) - 1] = '\0';
-	    }
-	}
-	DTRACE11(efile_drv_entry, dt_priv->thread_num, dt_priv->tag++,
-		 dt_utag, command, dt_s1, NULL, dt_i1, dt_i2, dt_i3, dt_i4,
-		 desc->port_str);
-#endif
-    }
-    cq_execute(desc);
-}
-
-#ifdef  USE_VM_PROBES
-dt_private *
-get_dt_private(int base)
-{
-    dt_private *dt_priv = (dt_private *) pthread_getspecific(dt_driver_key);
-
-    if (dt_priv == NULL) {
-	dt_priv = EF_SAFE_ALLOC(sizeof(dt_private));
-	erts_mtx_lock(&dt_driver_mutex);
-	dt_priv->thread_num = (base + dt_driver_idnum++);
-	erts_mtx_unlock(&dt_driver_mutex);
-	dt_priv->tag = 0;
-	pthread_setspecific(dt_driver_key, dt_priv);
-    }
-    return dt_priv;
-}
-#endif  /* USE_VM_PROBES */
diff --git a/erts/emulator/drivers/common/erl_efile.h b/erts/emulator/drivers/common/erl_efile.h
deleted file mode 100644
index b7f063b4f2..0000000000
--- a/erts/emulator/drivers/common/erl_efile.h
+++ /dev/null
@@ -1,176 +0,0 @@
-/*
- * %CopyrightBegin%
- *
- * Copyright Ericsson AB 1997-2016. 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%
- */
-/*
- * Defines the interfaces between the generic efile driver and its
- * operating-system dependent helpers.
- */
-
-#include "sys.h"
-#include "erl_driver.h"
-
-/*
- * Open modes for efile_openfile().
- */
-#define EFILE_MODE_READ 1
-#define EFILE_MODE_WRITE 2	/* Implies truncating file when used alone. */
-#define EFILE_MODE_READ_WRITE 	3
-#define EFILE_MODE_APPEND	4
-#define EFILE_COMPRESSED 	8
-#define EFILE_MODE_EXCL        16
-#define EFILE_NO_TRUNCATE      32 /* Special for reopening on VxWorks */
-#define EFILE_MODE_SYNC        64
-
-/*
- * Seek modes for efile_seek().
- */
-#define	EFILE_SEEK_SET	0
-#define	EFILE_SEEK_CUR	1
-#define	EFILE_SEEK_END	2
-
-/*
- * File types returned by efile_fileinfo().
- */
-#define FT_DEVICE 	1
-#define FT_DIRECTORY 	2
-#define FT_REGULAR	3
-#define FT_SYMLINK	4
-#define FT_OTHER	5
-
-/*
- * Access attributes returned by efile_fileinfo() (the bits can be ORed
- * together).
- */
-#define FA_NONE    	0
-#define FA_WRITE 	1
-#define FA_READ		2
-
-/* Some OS'es (i.e. Windows) has filenames in wide charaqcters. That requires special handling */
-/* Note that we do *not* honor alignment in the communication to the OS specific driver, */
-/* which is not a problem on x86, but might be on other platforms. The OS specific efile */
-/* implementation is expected to align if needed */
-#ifdef __WIN32__
-#define FILENAMES_16BIT 1
-#endif
-
-/* We use sendfilev if it exist on solaris */
-#if !defined(HAVE_SENDFILE) && defined(HAVE_SENDFILEV)
-#define HAVE_SENDFILE
-#endif
-
-/*
- * An handle to an open directory.  To be cast to the correct type
- * in the system-dependent directory functions.
- */
-
-typedef struct _Efile_Dir_Handle* EFILE_DIR_HANDLE;
-
-/*
- * Error information from the last call.
- */  
-typedef struct _Efile_error {
-    int posix_errno;		/* Posix error number, as in <errno.h>. */
-    int os_errno;		/* Os-dependent error number (not used). */
-} Efile_error;
-
-/*
- * Describes what is returned by file:file_info/1.
- */
-
-typedef struct _Efile_info {
-    Uint32 size_low;		/* Size of file, lower 32 bits.. */
-    Uint32 size_high;		/* Size of file, higher 32 bits. */
-    Uint32 type;		/* Type of file -- one of FT_*. */
-    Uint32 access;		/* Access to file -- one of FA_*. */
-    Uint32 mode;		/* Access permissions -- bit field. */
-    Uint32 links;		/* Number of links to file. */
-    Uint32 major_device;	/* Major device or file system. */
-    Uint32 minor_device;	/* Minor device (for devices). */
-    Uint32 inode;		/* Inode number. */
-    Uint32 uid;			/* User id of owner. */
-    Uint32 gid;			/* Group id of owner. */
-    Sint64 accessTime;		/* Last time the file was accessed. */
-    Sint64 modifyTime;		/* Last time the file was modified. */
-    Sint64 cTime;		/* Creation time (Windows) or last
-				 * inode change (Unix).
-				 */
-} Efile_info;
-
-
-#ifdef HAVE_SENDFILE
-/*
- * Describes the structure of headers/trailers for sendfile
- */
-struct t_sendfile_hdtl {
-    SysIOVec *headers;
-    int hdr_cnt;
-    SysIOVec *trailers;
-    int trl_cnt;
-};
-#endif /* HAVE_SENDFILE */
-
-/*
- * Functions.
- */
-int efile_init(void);
-int efile_mkdir(Efile_error* errInfo, char* name);
-int efile_rmdir(Efile_error* errInfo, char* name);
-int efile_delete_file(Efile_error* errInfo, char* name);
-int efile_rename(Efile_error* errInfo, char* src, char* dst);
-int efile_chdir(Efile_error* errInfo, char* name);
-int efile_getdcwd(Efile_error* errInfo, int drive,
-		  char* buffer, size_t size);
-int efile_readdir(Efile_error* errInfo, char* name, 
-		  EFILE_DIR_HANDLE* dir_handle,
-		  char* buffer, size_t *size);
-int efile_openfile(Efile_error* errInfo, char* name, int flags,
-		   int* pfd, Sint64* pSize);
-void efile_closefile(int fd);
-int efile_fdatasync(Efile_error* errInfo, int fd);
-int efile_fsync(Efile_error* errInfo, int fd);
-int efile_fileinfo(Efile_error* errInfo, Efile_info* pInfo,
-		   char *name, int info_for_link);
-int efile_write_info(Efile_error* errInfo, Efile_info* pInfo, char *name);
-int efile_write(Efile_error* errInfo, int flags, int fd, 
-		char* buf, size_t count);
-int efile_writev(Efile_error* errInfo, int flags, int fd, 
-		 SysIOVec* iov, int iovcnt);
-int efile_read(Efile_error* errInfo, int flags, int fd, 
-	       char* buf, size_t count, size_t* pBytesRead);
-int efile_seek(Efile_error* errInfo, int fd, 
-	       Sint64 offset, int origin, Sint64* new_location);
-int efile_truncate_file(Efile_error* errInfo, int *fd, int flags);
-int efile_pwrite(Efile_error* errInfo, int fd, 
-		 char* buf, size_t count, Sint64 offset);
-int efile_pread(Efile_error* errInfo, int fd, 
-		Sint64 offset, char* buf, size_t count, size_t* pBytesRead);
-int efile_readlink(Efile_error* errInfo, char *name, 
-		   char* buffer, size_t size);
-int efile_altname(Efile_error* errInfo, char *name, 
-		  char* buffer, size_t size);
-int efile_link(Efile_error* errInfo, char* old, char* new);
-int efile_symlink(Efile_error* errInfo, char* old, char* new);
-int efile_may_openfile(Efile_error* errInfo, char *name);
-int efile_fadvise(Efile_error* errInfo, int fd, Sint64 offset, Sint64 length,
-		  int advise);
-#ifdef HAVE_SENDFILE
-int efile_sendfile(Efile_error* errInfo, int in_fd, int out_fd,
-		      off_t *offset, Uint64 *nbytes, struct t_sendfile_hdtl *hdtl);
-#endif /* HAVE_SENDFILE */
-int efile_fallocate(Efile_error* errInfo, int fd, Sint64 offset, Sint64 length);
diff --git a/erts/emulator/drivers/common/gzio.c b/erts/emulator/drivers/common/gzio.c
index f60c781894..86c3b07cea 100644
--- a/erts/emulator/drivers/common/gzio.c
+++ b/erts/emulator/drivers/common/gzio.c
@@ -19,726 +19,16 @@
 #include <unistd.h>
 #endif
 #include <ctype.h>
+
 #include "erl_driver.h"
-#include "erl_efile.h"
 #include "sys.h"
 
-#ifdef __WIN32__
-#ifndef HAVE_CONFLICTING_FREAD_DECLARATION
-#define HAVE_CONFLICTING_FREAD_DECLARATION
-#endif
-#define FILENAMES_16BIT 1
-#endif
-
-#ifdef STDC
-#  define zstrerror(errnum) strerror(errnum)
-#else
-#  define zstrerror(errnum) ""
-#endif
-
 #include "gzio_zutil.h"
 #include "erl_zlib.h"
 #include "gzio.h"
 
-/********struct internal_state {int dummy;}; / * for buggy compilers */
-
-#define Z_BUFSIZE 4096
-
-#define ALLOC(size) driver_alloc(size)
-#define TRYFREE(p) {if (p) driver_free(p);}
-
 static int gz_magic[2] = {0x1f, 0x8b}; /* gzip magic header */
 
-/* gzip flag byte */
-#define ASCII_FLAG   0x01 /* bit 0 set: file probably ascii text */
-#define HEAD_CRC     0x02 /* bit 1 set: header CRC present */
-#define EXTRA_FIELD  0x04 /* bit 2 set: extra field present */
-#define ORIG_NAME    0x08 /* bit 3 set: original file name present */
-#define COMMENT      0x10 /* bit 4 set: file comment present */
-#define RESERVED     0xE0 /* bits 5..7: reserved */
-
-typedef struct gz_stream {
-    z_stream stream;
-    int      z_err;   /* error code for last stream operation */
-    int      z_eof;   /* set if end of input file */
-#ifdef UNIX
-    int      file;    /* .gz file descriptor */
-#else
-    FILE     *file;   /* .gz file */
-#endif
-    Byte     *inbuf;  /* input buffer */
-    Byte     *outbuf; /* output buffer */
-    uLong    crc;     /* crc32 of uncompressed data */
-    char     *msg;    /* error message */
-    char     *path;   /* path name for debugging only */
-    int      transparent; /* 1 if input file is not a .gz file */
-    char     mode;    /* 'w' or 'r' */
-    int      position; /* Position (for seek) */
-    int (*destroy)(struct gz_stream*); /* Function to destroy
-					    *  this structure. */
-} gz_stream;
-
-local ErtsGzFile gz_open      (const char *path, const char *mode);
-local int    get_byte     (gz_stream *s);
-local void   check_header (gz_stream *s);
-local int    destroy      (gz_stream *s);
-local uLong  getLong      (gz_stream *s);
-
-#ifdef UNIX
-/*
- * In Solaris 8 and earlier, fopen() and its friends cannot handle 
- * file descriptors larger than 255. Therefore, we use read()/write()
- * on all Unix systems.
- */
-# define ERTS_GZWRITE(File, Buf, Count) write((File), (Buf), (Count))
-# define ERTS_GZREAD(File, Buf, Count) read((File), (Buf), (Count))
-#else
-/*
- * On all other operating systems, using fopen(), fread()/fwrite(), since
- * there is not guaranteed to exist any read()/write() (not part of
- * ANSI/ISO-C).
- */
-# define ERTS_GZWRITE(File, Buf, Count) fwrite((Buf), 1, (Count), (File))
-# define ERTS_GZREAD(File, Buf, Count) fread((Buf), 1, (Count), (File))
-#endif
-
-/*
- * Ripped from efile_drv.c
- */
-
-#ifdef FILENAMES_16BIT
-#  define FILENAME_BYTELEN(Str) filename_len_16bit(Str)
-#  define FILENAME_COPY(To,From) filename_cpy_16bit((To),(From)) 
-#  define FILENAME_CHARSIZE 2
-
-   static int filename_len_16bit(const char *str) 
-   {
-       const char *p = str;
-       while(*p != '\0' || p[1] != '\0') {
-	   p += 2;
-       }
-       return (p - str);
-   }
-
-   static void filename_cpy_16bit(char *to, const char *from) 
-   {
-       while(*from != '\0' || from[1] != '\0') {
-	   *to++ = *from++;
-	   *to++ = *from++;
-       }
-       *to++ = *from++;
-       *to++ = *from++;
-   }
-
-#else
-#  define FILENAME_BYTELEN(Str) strlen(Str)
-#  define FILENAME_COPY(To,From) strcpy(To,From) 
-#  define FILENAME_CHARSIZE 1
-#endif
-
-/* ===========================================================================
-     Opens a gzip (.gz) file for reading or writing. The mode parameter
-   is as in fopen ("rb" or "wb"). The file is given either by file descriptor
-   or path name (if fd == -1).
-     gz_open return NULL if the file could not be opened or if there was
-   insufficient memory to allocate the (de)compression state; errno
-   can be checked to distinguish the two cases (if errno is zero, the
-   zlib error is Z_MEM_ERROR).
-*/
-local ErtsGzFile gz_open (path, mode)
-    const char *path;
-    const char *mode;
-{
-    int err;
-    int level = Z_DEFAULT_COMPRESSION; /* compression level */
-    char *p = (char*)mode;
-    gz_stream *s;
-    char fmode[80]; /* copy of mode, without the compression level */
-    char *m = fmode;
-
-    if (!path || !mode) return Z_NULL;
-
-    s = (gz_stream *)ALLOC(sizeof(gz_stream));
-    if (!s) return Z_NULL;
-
-    erl_zlib_alloc_init(&s->stream);
-    s->stream.next_in = s->inbuf = Z_NULL;
-    s->stream.next_out = s->outbuf = Z_NULL;
-    s->stream.avail_in = s->stream.avail_out = 0;
-#ifdef UNIX
-    s->file = -1;
-#else
-    s->file = NULL;
-#endif
-    s->z_err = Z_OK;
-    s->z_eof = 0;
-    s->crc = crc32(0L, Z_NULL, 0);
-    s->msg = NULL;
-    s->transparent = 0;
-    s->position = 0;
-    s->destroy = destroy;
-
-    s->path = (char*)ALLOC(FILENAME_BYTELEN(path)+FILENAME_CHARSIZE);
-    if (s->path == NULL) {
-        return s->destroy(s), (ErtsGzFile)Z_NULL;
-    }
-    FILENAME_COPY(s->path, path); /* do this early for debugging */
-
-    s->mode = '\0';
-    do {
-        if (*p == 'r')
-	    s->mode = 'r';
-	if (*p == 'w' || *p == 'a')
-	    s->mode = 'w';
-	if (isdigit((int)*p)) {
-	    level = *p - '0';
-	} else {
-	    *m++ = *p;		/* Copy the mode */
-	}
-    } while (*p++ && m < fmode + sizeof(fmode) - 1);
-    *m = '\0';
-    if (s->mode == '\0')
-	return s->destroy(s), (ErtsGzFile)Z_NULL;
-    
-    if (s->mode == 'w') {
-        err = deflateInit2(&(s->stream), level,
-                           Z_DEFLATED, MAX_WBITS+16, DEF_MEM_LEVEL, 0);
-        /* windowBits is passed < 0 to suppress zlib header */
-
-        s->stream.next_out = s->outbuf = (Byte*)ALLOC(Z_BUFSIZE);
-
-        if (err != Z_OK || s->outbuf == Z_NULL) {
-            return s->destroy(s), (ErtsGzFile)Z_NULL;
-        }
-    } else {
-	/*
-	 * It is tempting to use the built-in support in zlib
-	 * for handling GZIP headers, but unfortunately it
-	 * cannot handle multiple GZIP headers (which occur when
-	 * several GZIP files have been concatenated).
-	 */
-
-	err = inflateInit2(&(s->stream), -MAX_WBITS);
-        s->stream.next_in  = s->inbuf = (Byte*)ALLOC(Z_BUFSIZE);
-
-        if (err != Z_OK || s->inbuf == Z_NULL) {
-            return s->destroy(s), (ErtsGzFile)Z_NULL;
-        }
-    }
-    s->stream.avail_out = Z_BUFSIZE;
-
-    errno = 0;
-#if defined(FILENAMES_16BIT)
-    {
-	FILE* efile_wfopen(const WCHAR* name, const WCHAR* mode);
-	WCHAR wfmode[80];
-	int i = 0;
-	int j;
-	for(j = 0; fmode[j] != '\0'; ++j) {
-	    wfmode[i++] = (WCHAR) fmode[j];
-	}
-	wfmode[i++] = L'\0';
-	s->file = efile_wfopen((WCHAR *)path, wfmode);
-	if (s->file == NULL) {
-	    return s->destroy(s), (ErtsGzFile)Z_NULL;
-	}
-    }
-#elif defined(UNIX)
-    if (s->mode == 'r') {
-	s->file = open(path, O_RDONLY);
-    } else {
-	s->file = open(path, O_WRONLY | O_CREAT | O_TRUNC, 0666);
-    }
-    if (s->file == -1) {
-        return s->destroy(s), (ErtsGzFile)Z_NULL;
-    }
-#else
-    s->file = fopen(path, fmode);
-    if (s->file == NULL) {
-        return s->destroy(s), (ErtsGzFile)Z_NULL;
-    }
-#endif
-    if (s->mode == 'r') {
-	check_header(s); /* skip the .gz header */
-    }
-    return (ErtsGzFile)s;
-}
-
-/* ===========================================================================
-   Rewind a gzfile back to the beginning.
-*/
-
-local int gz_rewind (gz_stream *s)
-{
-    TRYFREE(s->msg);
-
-#ifdef UNIX
-    lseek(s->file, 0L, SEEK_SET);
-#else
-    fseek(s->file, 0L, SEEK_SET);
-#endif
-    inflateReset(&(s->stream));
-    s->stream.next_in = Z_NULL;
-    s->stream.next_out = Z_NULL;
-    s->stream.avail_in = s->stream.avail_out = 0;
-    s->z_err = Z_OK;
-    s->z_eof = 0;
-    s->crc = crc32(0L, Z_NULL, 0);
-    s->msg = NULL;
-    s->position = 0;
-    s->stream.next_in = s->inbuf;
-
-    s->stream.avail_out = Z_BUFSIZE;
-
-    check_header(s);		/* skip the .gz header */
-    return 1;
-}
-
-/* ===========================================================================
-     Opens a gzip (.gz) file for reading or writing.
-*/
-ErtsGzFile erts_gzopen (path, mode)
-    const char *path;
-    const char *mode;
-{
-    return gz_open (path, mode);
-}
-
-
-/* ===========================================================================
-     Read a byte from a gz_stream; update next_in and avail_in. Return EOF
-   for end of file.
-   IN assertion: the stream s has been successfully opened for reading.
-*/
-local int get_byte(s)
-    gz_stream *s;
-{
-    if (s->z_eof) return EOF;
-    if (s->stream.avail_in == 0) {
-#ifdef UNIX
-	ssize_t res;
-	errno = 0;
-	res = ERTS_GZREAD(s->file, s->inbuf, Z_BUFSIZE);
-	if (res == 0) {
-	    s->stream.avail_in = 0;
-	    s->z_eof = 1;
-	    return EOF;
-	} else if (res < 0) {
-	    s->stream.avail_in = 0;
-	    s->z_eof = 1;
-	    s->z_err = Z_ERRNO;
-	    return EOF;
-	} else {
-	    s->stream.avail_in = (uInt) res;
-	}
-#else
-	errno = 0;
-	s->stream.avail_in = ERTS_GZREAD(s->file, s->inbuf, Z_BUFSIZE);
-	if (s->stream.avail_in == 0) {
-	    s->z_eof = 1;
-	    if (s->file && ferror(s->file))
-		s->z_err = Z_ERRNO;
-	    return EOF;
-	}
-#endif
-	s->stream.next_in = s->inbuf;
-    }
-    s->stream.avail_in--;
-    return *(s->stream.next_in)++;
-}
-
-/* ===========================================================================
-      Check the gzip header of a gz_stream opened for reading. Set the stream
-    mode to transparent if the gzip magic header is not present; set s->err
-    to Z_DATA_ERROR if the magic header is present but the rest of the header
-    is incorrect.
-    IN assertion: the stream s has already been created sucessfully;
-       s->stream.avail_in is zero for the first time, but may be non-zero
-       for concatenated .gz files.
-*/
-local void check_header(s)
-    gz_stream *s;
-{
-    int method; /* method byte */
-    int flags;  /* flags byte */
-    uInt len;
-    int c;
-
-    /* Check the gzip magic header */
-    for (len = 0; len < 2; len++) {
-	c = get_byte(s);
-	if (c != gz_magic[len]) {
-	    if (len != 0) s->stream.avail_in++, s->stream.next_in--;
-	    if (c != EOF) {
-		s->stream.avail_in++, s->stream.next_in--;
-		s->transparent = 1;
-	    }
-	    s->z_err = s->stream.avail_in != 0 ? Z_OK : Z_STREAM_END;
-	    return;
-	}
-    }
-    method = get_byte(s);
-    flags = get_byte(s);
-    if (method != Z_DEFLATED || (flags & RESERVED) != 0) {
-	s->z_err = Z_DATA_ERROR;
-	return;
-    }
-
-    /* Discard time, xflags and OS code: */
-    for (len = 0; len < 6; len++) (void)get_byte(s);
-
-    if ((flags & EXTRA_FIELD) != 0) { /* skip the extra field */
-	len  =  (uInt)get_byte(s);
-	len += ((uInt)get_byte(s))<<8;
-	/* len is garbage if EOF but the loop below will quit anyway */
-	while (len-- != 0 && get_byte(s) != EOF) ;
-    }
-    if ((flags & ORIG_NAME) != 0) { /* skip the original file name */
-	while ((c = get_byte(s)) != 0 && c != EOF) ;
-    }
-    if ((flags & COMMENT) != 0) {   /* skip the .gz file comment */
-	while ((c = get_byte(s)) != 0 && c != EOF) ;
-    }
-    if ((flags & HEAD_CRC) != 0) {  /* skip the header crc */
-	for (len = 0; len < 2; len++) (void)get_byte(s);
-    }
-    s->z_err = s->z_eof ? Z_DATA_ERROR : Z_OK;
-}
-
- /* ===========================================================================
- * Cleanup then free the given gz_stream. Return a zlib error code.
-   Try freeing in the reverse order of allocations.
- */
-local int destroy (s)
-    gz_stream *s;
-{
-    int err = Z_OK;
-
-    if (!s) return Z_STREAM_ERROR;
-
-    TRYFREE(s->msg);
-
-    if (s->stream.state != NULL) {
-       if (s->mode == 'w') {
-           err = deflateEnd(&(s->stream));
-       } else if (s->mode == 'r') {
-           err = inflateEnd(&(s->stream));
-       }
-    }
-#ifdef UNIX
-    if (s->file != -1 && close(s->file)) {
-        err = Z_ERRNO;
-    }
-#else
-    if (s->file != NULL && fclose(s->file)) {
-        err = Z_ERRNO;
-    }
-#endif
-    if (s->z_err < 0) err = s->z_err;
-
-    TRYFREE(s->inbuf);
-    TRYFREE(s->outbuf);
-    TRYFREE(s->path);
-    TRYFREE(s);
-    return err;
-}
-
-/* ===========================================================================
-     Reads the given number of uncompressed bytes from the compressed file.
-   gzread returns the number of bytes actually read (0 for end of file).
-*/
-int
-erts_gzread(ErtsGzFile file, voidp buf, unsigned len)
-{
-    gz_stream *s = (gz_stream*)file;
-    Bytef *start = buf; /* starting point for crc computation */
-    Byte  *next_out; /* == stream.next_out but not forced far (for MSDOS) */
-
-    if (s == NULL || s->mode != 'r') return Z_STREAM_ERROR;
-
-    if (s->z_err == Z_DATA_ERROR || s->z_err == Z_ERRNO) return -1;
-    if (s->z_err == Z_STREAM_END) return 0;  /* EOF */
-
-    s->stream.next_out = next_out = buf;
-    s->stream.avail_out = len;
-
-    while (s->stream.avail_out != 0) {
-
-	if (s->transparent) {
-	    /* Copy first the lookahead bytes: */
-	    uInt n = s->stream.avail_in;
-	    if (n > s->stream.avail_out) n = s->stream.avail_out;
-	    if (n > 0) {
-		zmemcpy(s->stream.next_out, s->stream.next_in, n);
-		next_out += n;
-		s->stream.next_out = next_out;
-		s->stream.next_in   += n;
-		s->stream.avail_out -= n;
-		s->stream.avail_in  -= n;
-	    }
-	    if (s->stream.avail_out > 0) {
-		s->stream.avail_out -= ERTS_GZREAD(s->file, next_out,
-						   s->stream.avail_out);
-	    }
-	    len -= s->stream.avail_out;
-	    s->stream.total_in  += (uLong)len;
-	    s->stream.total_out += (uLong)len;
-            if (len == 0) s->z_eof = 1;
-	    s->position += (int)len;
-	    return (int)len;
-	}
-        if (s->stream.avail_in == 0 && !s->z_eof) {
-#ifdef UNIX
-	    ssize_t res;
-	    errno = 0;
-	    res = ERTS_GZREAD(s->file, s->inbuf, Z_BUFSIZE);
-	    if (res == 0) {
-		s->stream.avail_in = 0;
-		s->z_eof = 1;
-		return EOF;
-	    } else if (res < 0) {
-		s->stream.avail_in = 0;
-		s->z_eof = 1;
-		s->z_err = Z_ERRNO;
-		return EOF;
-	    } else {
-		s->stream.avail_in = (uInt) res;
-	    }
-#else
-	    errno = 0;
-            s->stream.avail_in = ERTS_GZREAD(s->file, s->inbuf, Z_BUFSIZE);
-            if (s->stream.avail_in == 0) {
-                s->z_eof = 1;
-		if (s->file && ferror(s->file)) {
-		    s->z_err = Z_ERRNO;
-		    break;
-		}
-            }
-#endif
-            s->stream.next_in = s->inbuf;
-        }
-        s->z_err = inflate(&(s->stream), Z_NO_FLUSH);
-
-	if (s->z_err == Z_STREAM_END) {
-	    /* Check CRC and original size */
-	    s->crc = crc32(s->crc, start, (uInt)(s->stream.next_out - start));
-	    start = s->stream.next_out;
-
-	    if (getLong(s) != s->crc) {
-		s->z_err = Z_DATA_ERROR;
-	    } else {
-	        (void)getLong(s);
-                /* The uncompressed length returned by above getlong() may
-                 * be different from s->stream.total_out) in case of
-		 * concatenated .gz files. Check for such files:
-		 */
-		check_header(s);
-		if (s->z_err == Z_OK) {
-		    uLong total_in = s->stream.total_in;
-		    uLong total_out = s->stream.total_out;
-
-		    inflateReset(&(s->stream));
-		    s->stream.total_in = total_in;
-		    s->stream.total_out = total_out;
-		    s->crc = crc32(0L, Z_NULL, 0);
-		}
-	    }
-	}
-	if (s->z_err != Z_OK || s->z_eof) break;
-    }
-    s->crc = crc32(s->crc, start, (uInt)(s->stream.next_out - start));
-
-    s->position += (int)(len - s->stream.avail_out);
-
-    return (int)(len - s->stream.avail_out);
-}
-
-/* ===========================================================================
-     Writes the given number of uncompressed bytes into the compressed file.
-   gzwrite returns the number of bytes actually written (0 in case of error).
-*/
-int
-erts_gzwrite(ErtsGzFile file, voidp buf, unsigned len)
-{
-    gz_stream *s = (gz_stream*)file;
-
-    if (s == NULL || s->mode != 'w') return Z_STREAM_ERROR;
-
-    s->stream.next_in = buf;
-    s->stream.avail_in = len;
-
-    while (s->stream.avail_in != 0) {
-
-        if (s->stream.avail_out == 0) {
-
-            s->stream.next_out = s->outbuf;
-            if (ERTS_GZWRITE(s->file, s->outbuf, Z_BUFSIZE) != Z_BUFSIZE) {
-                s->z_err = Z_ERRNO;
-                break;
-            }
-            s->stream.avail_out = Z_BUFSIZE;
-        }
-        s->z_err = deflate(&(s->stream), Z_NO_FLUSH);
-        if (s->z_err != Z_OK) break;
-    }
-    s->position += (int)(len - s->stream.avail_in);
-    return (int)(len - s->stream.avail_in);
-}
-
-/*
- * For use by Erlang file driver.
- *
- * XXX Limitations:
- *  - SEEK_END is not allowed (length of file is not known).
- *  - When writing, only forward seek is supported.
- */
-
-int
-erts_gzseek(ErtsGzFile file, int offset, int whence)
-{
-    int pos;
-    gz_stream* s = (gz_stream *) file;
-
-    switch (whence) {
-    case EFILE_SEEK_SET: whence = SEEK_SET; break;
-    case EFILE_SEEK_CUR: whence = SEEK_CUR; break;
-    case EFILE_SEEK_END: whence = SEEK_END; break;
-    default:
-	errno = EINVAL;
-	return -1;
-    }
-
-    if (s == NULL) {
-	errno = EINVAL;
-	return -1;
-    }
-    if (s->z_err == Z_DATA_ERROR || s->z_err == Z_ERRNO) {
-	errno = EIO;
-	return -1;
-    }
-
-    switch (whence) {
-    case SEEK_SET: pos = offset; break;
-    case SEEK_CUR: pos = s->position+offset; break;
-    case SEEK_END: 
-    default:
-      errno = EINVAL; return -1;
-    }
-
-    if (pos == s->position) {
-	return pos;
-    }
-
-    if (pos < s->position) {
-	if (s->mode == 'w') {
-	    errno = EINVAL;
-	    return -1;
-	}
-	gz_rewind(s);
-    }
-
-    while (s->position < pos) {
-	char buf[512];
-	int n;
-	int save_pos = s->position;
-
-	n = pos - s->position;
-	if (n > sizeof(buf))
-	    n = sizeof(buf);
-
-	if (s->mode == 'r') {
-	    erts_gzread(file, buf, n);
-	} else {
-	    memset(buf, '\0', n);
-	    erts_gzwrite(file, buf, n);
-	}
-	if (save_pos == s->position) break;
-    }
-
-    return s->position;
-}
-
-/* ===========================================================================
-     Flushes all pending output into the compressed file. The parameter
-   flush is as in the deflate() function.
-     gzflush should be called only when strictly necessary because it can
-   degrade compression.
-*/
-int
-erts_gzflush(ErtsGzFile file, int flush)
-{
-    uInt len;
-    int done = 0;
-    gz_stream *s = (gz_stream*)file;
-
-    if (s == NULL || s->mode != 'w') return Z_STREAM_ERROR;
-
-    s->stream.avail_in = 0; /* should be zero already anyway */
-
-    for (;;) {
-        len = Z_BUFSIZE - s->stream.avail_out;
-
-        if (len != 0) {
-            if ((uInt)ERTS_GZWRITE(s->file, s->outbuf, len) != len) {
-                s->z_err = Z_ERRNO;
-                return Z_ERRNO;
-            }
-            s->stream.next_out = s->outbuf;
-            s->stream.avail_out = Z_BUFSIZE;
-        }
-        if (done) break;
-        s->z_err = deflate(&(s->stream), flush);
-
-        /* deflate has finished flushing only when it hasn't used up
-         * all the available space in the output buffer: 
-         */
-        done = (s->stream.avail_out != 0 || s->z_err == Z_STREAM_END);
- 
-        if (s->z_err != Z_OK && s->z_err != Z_STREAM_END) break;
-    }
-#ifndef UNIX
-    fflush(s->file);
-#endif
-    return  s->z_err == Z_STREAM_END ? Z_OK : s->z_err;
-}
-
-/* ===========================================================================
-   Reads a long in LSB order from the given gz_stream. Sets 
-*/
-local uLong getLong (s)
-    gz_stream *s;
-{
-    uLong x = (uLong)get_byte(s);
-    int c;
-
-    x += ((uLong)get_byte(s))<<8;
-    x += ((uLong)get_byte(s))<<16;
-    c = get_byte(s);
-    if (c == EOF) s->z_err = Z_DATA_ERROR;
-    x += ((uLong)c)<<24;
-    return x;
-}
-
-/* ===========================================================================
-     Flushes all pending output if necessary, closes the compressed file
-   and deallocates all the (de)compression state.
-*/
-int
-erts_gzclose(ErtsGzFile file)
-{
-    int err;
-    gz_stream *s = (gz_stream*)file;
-
-    if (s == NULL) return Z_STREAM_ERROR;
-
-    if (s->mode == 'w') {
-        err = erts_gzflush (file, Z_FINISH);
-        if (err != Z_OK) return s->destroy(s);
-    }
-    return s->destroy(s);
-}
-
-
 /* ===========================================================================
    Uncompresses the buffer given and returns a pointer to a binary.
    If the buffer was not compressed with gzip, the buffer contents
diff --git a/erts/emulator/drivers/common/gzio.h b/erts/emulator/drivers/common/gzio.h
index ee0ebe7bd8..e331b5208b 100644
--- a/erts/emulator/drivers/common/gzio.h
+++ b/erts/emulator/drivers/common/gzio.h
@@ -20,13 +20,5 @@
 
 #include "zlib.h"
 
-typedef struct erts_gzFile* ErtsGzFile;
-
-ErtsGzFile erts_gzopen (const char *path, const char *mode);
-int erts_gzread(ErtsGzFile file, voidp buf, unsigned len);
-int erts_gzwrite(ErtsGzFile file, voidp buf, unsigned len);
-int erts_gzseek(ErtsGzFile, int, int);
-int erts_gzflush(ErtsGzFile file, int flush);
-int erts_gzclose(ErtsGzFile file);
 ErlDrvBinary* erts_gzinflate_buffer(char*, uLong);
 ErlDrvBinary* erts_gzdeflate_buffer(char*, uLong);
diff --git a/erts/emulator/drivers/common/inet_drv.c b/erts/emulator/drivers/common/inet_drv.c
index 95d61fcc5d..4294fb4f46 100644
--- a/erts/emulator/drivers/common/inet_drv.c
+++ b/erts/emulator/drivers/common/inet_drv.c
@@ -63,6 +63,20 @@
 #include <sys/un.h>
 #endif
 
+#ifdef HAVE_SENDFILE
+#if defined(__linux__) || (defined(__sun) && defined(__SVR4))
+    #include <sys/sendfile.h>
+#elif defined(__FreeBSD__) || defined(__DragonFly__)
+    /* Need to define __BSD_VISIBLE in order to expose prototype of sendfile */
+    #define __BSD_VISIBLE 1
+    #include <sys/socket.h>
+#endif
+#endif
+
+#if defined(__APPLE__) && defined(__MACH__) && !defined(__DARWIN__)
+    #define __DARWIN__ 1
+#endif
+
 /* All platforms fail on malloc errors. */
 #define FATAL_MALLOC
 
@@ -701,6 +715,7 @@ static size_t my_strnlen(const char *s, size_t maxlen)
 #define TCP_REQ_RECV           42
 #define TCP_REQ_UNRECV         43
 #define TCP_REQ_SHUTDOWN       44
+#define TCP_REQ_SENDFILE       45
 /* UDP and SCTP requests */
 #define PACKET_REQ_RECV        60 /* Common for UDP and SCTP         */
 /* #define SCTP_REQ_LISTEN       61 MERGED Different from TCP; not for UDP */
@@ -723,6 +738,7 @@ static size_t my_strnlen(const char *s, size_t maxlen)
 #define TCP_ADDF_DELAYED_ECONNRESET 128 /* An ECONNRESET error occurred on send or shutdown */
 #define TCP_ADDF_SHUTDOWN_WR_DONE 256 /* A shutdown(sock, SHUT_WR) or SHUT_RDWR was made */
 #define TCP_ADDF_LINGER_ZERO 	  512 /* Discard driver queue on port close */
+#define TCP_ADDF_SENDFILE         1024 /* Send from an fd instead of the driver queue */
 
 /* *_REQ_* replies */
 #define INET_REP_ERROR       0
@@ -1235,6 +1251,21 @@ typedef struct {
     inet_async_multi_op *multi_first;/* NULL == no multi-accept-queue, op is in ordinary queue */
     inet_async_multi_op *multi_last;
     MultiTimerData *mtd;        /* Timer structures for multiple accept */
+#ifdef HAVE_SENDFILE
+    struct {
+        ErlDrvSizeT ioq_skip;   /* The number of bytes in the queue at the time
+                                 * sendfile was issued, which must be sent
+                                 * before issuing the sendfile call itself. */
+        int dup_file_fd;        /* The file handle to send from; this is
+                                 * duplicated when sendfile is issued to
+                                 * reduce (but not eliminate) the impact of a
+                                 * nasty race, so we have to remember to close
+                                 * it. */
+        Uint64 bytes_sent;
+        Uint64 offset;
+        Uint64 length;
+    } sendfile;
+#endif
 } tcp_descriptor;
 
 /* send function */
@@ -1245,6 +1276,8 @@ static int tcp_deliver(tcp_descriptor* desc, int len);
 
 static int tcp_shutdown_error(tcp_descriptor* desc, int err);
 
+static int tcp_inet_sendfile(tcp_descriptor* desc);
+
 static int tcp_inet_output(tcp_descriptor* desc, HANDLE event);
 static int tcp_inet_input(tcp_descriptor* desc, HANDLE event);
 
@@ -1329,6 +1362,9 @@ static ErlDrvTermData am_ipv6_v6only;
 static ErlDrvTermData am_netns;
 static ErlDrvTermData am_bind_to_device;
 #endif
+#ifdef HAVE_SENDFILE
+static ErlDrvTermData am_sendfile;
+#endif
 
 static char str_eafnosupport[] = "eafnosupport";
 static char str_einval[] = "einval";
@@ -3875,6 +3911,10 @@ static int inet_init()
     INIT_ATOM(https);
     INIT_ATOM(scheme);
 
+#ifdef HAVE_SENDFILE
+    INIT_ATOM(sendfile);
+#endif
+
     /* add TCP, UDP and SCTP drivers */
     add_driver_entry(&tcp_inet_driver_entry);
 #ifdef HAVE_UDP
@@ -9270,6 +9310,13 @@ static void tcp_inet_stop(ErlDrvData e)
  * will be freed through tcp_inet_stop later on. */
 static void tcp_desc_close(tcp_descriptor* desc)
 {
+#ifdef HAVE_SENDFILE
+    if(desc->tcp_add_flags & TCP_ADDF_SENDFILE) {
+        desc->tcp_add_flags &= ~TCP_ADDF_SENDFILE;
+        close(desc->sendfile.dup_file_fd);
+    }
+#endif
+
     tcp_clear_input(desc);
     tcp_clear_output(desc);
 
@@ -9608,6 +9655,60 @@ static ErlDrvSSizeT tcp_inet_ctl(ErlDrvData e, unsigned int cmd,
 	    return ctl_reply(INET_REP_OK, NULL, 0, rbuf, rsize);
 	}
     }
+
+    case TCP_REQ_SENDFILE: {
+#ifdef HAVE_SENDFILE
+        const ErlDrvSizeT required_len =
+            sizeof(desc->sendfile.dup_file_fd) +
+            sizeof(Uint64) * 2;
+
+        int raw_file_fd;
+
+        DEBUGF(("tcp_inet_ctl(%ld): SENDFILE\r\n", (long)desc->inet.port));
+
+        if (len != required_len) {
+            return ctl_error(EINVAL, rbuf, rsize);
+        } else if (!IS_CONNECTED(INETP(desc))) {
+            return ctl_error(ENOTCONN, rbuf, rsize);
+        }
+
+        sys_memcpy(&raw_file_fd, buf, sizeof(raw_file_fd));
+        buf += sizeof(raw_file_fd);
+
+        desc->sendfile.dup_file_fd = dup(raw_file_fd);
+
+        if(desc->sendfile.dup_file_fd == -1) {
+            return ctl_error(errno, rbuf, rsize);
+        }
+
+        desc->sendfile.offset = get_int64(buf);
+        buf += sizeof(Uint64);
+
+        desc->sendfile.length = get_int64(buf);
+        buf += sizeof(Uint64);
+
+        ASSERT(desc->sendfile.offset >= 0);
+        ASSERT(desc->sendfile.length >= 0);
+
+        desc->sendfile.ioq_skip = driver_sizeq(desc->inet.port);
+        desc->sendfile.bytes_sent = 0;
+
+        desc->inet.caller = driver_caller(desc->inet.port);
+        desc->tcp_add_flags |= TCP_ADDF_SENDFILE;
+
+        /* See if we can finish sending without selecting & rescheduling. */
+        tcp_inet_sendfile(desc);
+
+        if(desc->sendfile.length > 0) {
+            sock_select(INETP(desc), FD_WRITE, 1);
+        }
+
+        return ctl_reply(INET_REP_OK, NULL, 0, rbuf, rsize);
+#else
+        return ctl_error(ENOTSUP, rbuf, rsize);
+#endif
+    }
+
     default:
 	DEBUGF(("tcp_inet_ctl(%ld): %u\r\n", (long)desc->inet.port, cmd)); 
 	return inet_ctl(INETP(desc), cmd, buf, len, rbuf, rsize);
@@ -9747,12 +9848,27 @@ static void tcp_inet_commandv(ErlDrvData e, ErlIOVec* ev)
 static void tcp_inet_flush(ErlDrvData e)
 {
     tcp_descriptor* desc = (tcp_descriptor*)e;
-    if (!(desc->inet.event_mask & FD_WRITE)) {
-	/* Discard send queue to avoid hanging port (OTP-7615) */
-	tcp_clear_output(desc);
+    int discard_output;
+
+    /* Discard send queue to avoid hanging port (OTP-7615) */
+    discard_output = !(desc->inet.event_mask & FD_WRITE);
+
+    discard_output |= desc->tcp_add_flags & TCP_ADDF_LINGER_ZERO;
+
+#ifdef HAVE_SENDFILE
+    /* The old file driver aborted when it was stopped during sendfile, so
+     * we'll clear the flag and discard all output. */
+    if(desc->tcp_add_flags & TCP_ADDF_SENDFILE) {
+        desc->tcp_add_flags &= ~TCP_ADDF_SENDFILE;
+        close(desc->sendfile.dup_file_fd);
+
+        discard_output = 1;
+    }
+#endif
+
+    if (discard_output) {
+        tcp_clear_output(desc);
     }
-    if (desc->tcp_add_flags & TCP_ADDF_LINGER_ZERO)
-	tcp_clear_output(desc);
 }
 
 static void tcp_inet_process_exit(ErlDrvData e, ErlDrvMonitor *monitorp) 
@@ -10647,7 +10763,9 @@ static int tcp_sendv(tcp_descriptor* desc, ErlIOVec* ev)
 	ev->size += h_len;
     }
 
-    if ((sz = driver_sizeq(ix)) > 0) {
+    sz = driver_sizeq(ix);
+
+    if ((desc->tcp_add_flags & TCP_ADDF_SENDFILE) || sz > 0) {
 	driver_enqv(ix, ev, 0);
 	if (sz+ev->size >= desc->high) {
 	    DEBUGF(("tcp_sendv(%ld): s=%d, sender forced busy\r\n",
@@ -10741,8 +10859,9 @@ static int tcp_send(tcp_descriptor* desc, char* ptr, ErlDrvSizeT len)
 
     inet_output_count(INETP(desc), len+h_len);
 
+    sz = driver_sizeq(ix);
 
-    if ((sz = driver_sizeq(ix)) > 0) {
+    if ((desc->tcp_add_flags & TCP_ADDF_SENDFILE) || sz > 0) {
 	if (h_len > 0)
 	    driver_enq(ix, buf, h_len);
 	driver_enq(ix, ptr, len);
@@ -10832,6 +10951,246 @@ static void tcp_inet_drv_input(ErlDrvData data, ErlDrvEvent event)
     (void)tcp_inet_input((tcp_descriptor*)data, (HANDLE)event);
 }
 
+#ifdef HAVE_SENDFILE
+static int tcp_sendfile_completed(tcp_descriptor* desc) {
+    ErlDrvTermData spec[LOAD_PORT_CNT + LOAD_TUPLE_CNT * 2 +
+        LOAD_ATOM_CNT * 2 + LOAD_UINT_CNT * 2];
+    Uint32 sent_low, sent_high;
+    int i;
+
+    desc->tcp_add_flags &= ~TCP_ADDF_SENDFILE;
+    close(desc->sendfile.dup_file_fd);
+
+    /* While we flushed the output queue prior to sending the file, we've
+     * deferred clearing busy status until now as there's no point in doing so
+     * while we still have a file to send.
+     *
+     * The watermark is checked since more data may have been added while we
+     * were sending the file. */
+
+    if (driver_sizeq(desc->inet.port) <= desc->low) {
+        if (IS_BUSY(INETP(desc))) {
+            desc->inet.caller = desc->inet.busy_caller;
+            desc->inet.state &= ~INET_F_BUSY;
+
+            set_busy_port(desc->inet.port, 0);
+
+            /* if we have a timer then cancel and send ok to client */
+            if (desc->busy_on_send) {
+                driver_cancel_timer(desc->inet.port);
+                desc->busy_on_send = 0;
+            }
+
+            inet_reply_ok(INETP(desc));
+        }
+    }
+
+    if (driver_sizeq(desc->inet.port) == 0) {
+        sock_select(INETP(desc), FD_WRITE, 0);
+        send_empty_out_q_msgs(INETP(desc));
+
+        if (desc->tcp_add_flags & TCP_ADDF_PENDING_SHUTDOWN) {
+            tcp_shutdown_async(desc);
+        }
+    }
+
+    sent_low = ((Uint64)desc->sendfile.bytes_sent >> 0) & 0xFFFFFFFF;
+    sent_high = ((Uint64)desc->sendfile.bytes_sent >> 32) & 0xFFFFFFFF;
+
+    i = LOAD_ATOM(spec, 0, am_sendfile);
+    i = LOAD_PORT(spec, i, desc->inet.dport);
+    i = LOAD_ATOM(spec, i, am_ok);
+    i = LOAD_UINT(spec, i, sent_low);
+    i = LOAD_UINT(spec, i, sent_high);
+    i = LOAD_TUPLE(spec, i, 3);
+    i = LOAD_TUPLE(spec, i, 3);
+
+    ASSERT(i == sizeof(spec)/sizeof(*spec));
+
+    return erl_drv_output_term(desc->inet.dport, spec, i);
+}
+
+static int tcp_sendfile_aborted(tcp_descriptor* desc, int socket_error) {
+    ErlDrvTermData spec[LOAD_PORT_CNT + LOAD_TUPLE_CNT * 2 + LOAD_ATOM_CNT * 3];
+    int i;
+
+    /* We don't clean up sendfile state here, as that's done in tcp_desc_close
+     * following normal error handling. All we do here is report the failure. */
+
+    i = LOAD_ATOM(spec, 0, am_sendfile);
+    i = LOAD_PORT(spec, i, desc->inet.dport);
+    i = LOAD_ATOM(spec, i, am_error);
+
+    switch (socket_error) {
+    case ECONNRESET:
+    case ENOTCONN:
+    case EPIPE:
+        i = LOAD_ATOM(spec, i, am_closed);
+        break;
+    default:
+        i = LOAD_ATOM(spec, i, error_atom(socket_error));
+    }
+
+    i = LOAD_TUPLE(spec, i, 2);
+    i = LOAD_TUPLE(spec, i, 3);
+
+    ASSERT(i == sizeof(spec)/sizeof(*spec));
+
+    return erl_drv_output_term(desc->inet.dport, spec, i);
+}
+
+static int tcp_inet_sendfile(tcp_descriptor* desc) {
+    ErlDrvPort ix = desc->inet.port;
+    int result = 0;
+    ssize_t n;
+
+    DEBUGF(("tcp_inet_sendfile(%ld) {s=%d\r\n", (long)ix, desc->inet.s));
+
+    /* If there was any data in the queue by the time sendfile was issued,
+     * we'll need to skip it first. Note that we don't clear busy status until
+     * we're finished sending the file. */
+    while (desc->sendfile.ioq_skip > 0) {
+        ssize_t bytes_to_send;
+        SysIOVec* iov;
+        int vsize;
+
+        ASSERT(driver_sizeq(ix) >= desc->sendfile.ioq_skip);
+
+        if ((iov = driver_peekq(ix, &vsize)) == NULL) {
+            ERTS_INTERNAL_ERROR("ioq empty when sendfile.ioq_skip > 0");
+        }
+
+        bytes_to_send = MIN(desc->sendfile.ioq_skip, iov[0].iov_len);
+        n = sock_send(desc->inet.s, iov[0].iov_base, bytes_to_send, 0);
+
+        if (!IS_SOCKET_ERROR(n)) {
+            desc->sendfile.ioq_skip -= n;
+            driver_deq(ix, n);
+        } else if (sock_errno() == ERRNO_BLOCK) {
+#ifdef __WIN32__
+            desc->inet.send_would_block = 1;
+#endif
+            goto done;
+        } else if (sock_errno() != EINTR) {
+            goto socket_error;
+        }
+    }
+
+    while (desc->sendfile.length > 0) {
+        /* For some reason the maximum ssize_t cannot be used as the max size.
+         * 1GB seems to work on all platforms */
+        const Sint64 SENDFILE_CHUNK_SIZE = ((1UL << 30) - 1);
+
+        ssize_t bytes_to_send = MIN(SENDFILE_CHUNK_SIZE, desc->sendfile.length);
+        off_t offset = desc->sendfile.offset;
+
+#if defined(__linux__)
+        n = sendfile(desc->inet.s, desc->sendfile.dup_file_fd, &offset,
+            bytes_to_send);
+#elif defined(__FreeBSD__) || defined(__DragonFly__) || defined(__DARWIN__)
+        {
+            off_t bytes_sent;
+            int error;
+
+    #if defined(__DARWIN__)
+            bytes_sent = bytes_to_send;
+
+            error = sendfile(desc->sendfile.dup_file_fd, desc->inet.s, offset,
+                &bytes_sent, NULL, 0);
+            n = bytes_sent;
+    #else
+            error = sendfile(desc->sendfile.dup_file_fd, desc->inet.s, offset,
+                bytes_to_send, NULL, &bytes_sent, 0);
+            n = bytes_sent;
+    #endif
+
+            if(error < 0) {
+                /* EAGAIN/EINTR report partial success by setting bytes_sent,
+                 * so we have to skip error handling if nonzero, and skip EOF
+                 * handling if zero, as it's possible that we didn't manage to
+                 * send anything at all before being interrupted by a
+                 * signal. */
+                if((errno != EAGAIN && errno != EINTR) || bytes_sent == 0) {
+                    n = -1;
+                }
+            }
+        }
+#elif defined(__sun) && defined(__SVR4) && defined(HAVE_SENDFILEV)
+        {
+            sendfilevec_t sfvec[1];
+            size_t bytes_sent;
+            ssize_t error;
+
+            sfvec[0].sfv_fd = desc->sendfile.dup_file_fd;
+            sfvec[0].sfv_len = bytes_to_send;
+            sfvec[0].sfv_off = offset;
+            sfvec[0].sfv_flag = 0;
+
+            error = sendfilev(desc->inet.s, sfvec, 1, &bytes_sent);
+            n = bytes_sent;
+
+            if(error < 0) {
+                if(errno == EINVAL) {
+                    /* On some solaris versions (I've seen it on SunOS 5.10),
+                     * using a sfv_len larger than the filesize will result in
+                     * a (-1 && errno == EINVAL). We translate this to a
+                     * successful send of the data.*/
+                } else {
+                    /* EAGAIN/EINTR behavior is identical to *BSD. */
+                    if((errno != EAGAIN && errno != EINTR) || bytes_sent == 0) {
+                        n = -1;
+                    }
+                }
+            }
+        }
+#else
+        #error "Unsupported sendfile syscall; update configure test."
+#endif
+
+        if (n > 0) {
+            desc->sendfile.bytes_sent += n;
+            desc->sendfile.offset += n;
+            desc->sendfile.length -= n;
+        } else if (n == 0) {
+            /* EOF. */
+            desc->sendfile.length = 0;
+            break;
+        } else if (IS_SOCKET_ERROR(n) && sock_errno() != EINTR) {
+            if (sock_errno() != ERRNO_BLOCK) {
+                goto socket_error;
+            }
+
+#ifdef __WIN32__
+            desc->inet.send_would_block = 1;
+#endif
+            break;
+        }
+    }
+
+    if (desc->sendfile.length == 0) {
+        tcp_sendfile_completed(desc);
+    }
+
+    goto done;
+
+socket_error: {
+        int socket_errno = sock_errno();
+
+        DEBUGF(("tcp_inet_sendfile(%ld): send errno = %d (errno %d)\r\n",
+            (long)desc->inet.port, socket_errno, errno));
+
+        result = tcp_send_error(desc, socket_errno);
+        tcp_sendfile_aborted(desc, socket_errno);
+
+        goto done;
+    }
+
+done:
+    DEBUGF(("tcp_inet_sendfile(%ld) }\r\n", (long)desc->inet.port));
+    return result;
+}
+#endif /* HAVE_SENDFILE */
+
 /* socket ready for ouput:
 ** 1. INET_STATE_CONNECTING => non block connect ?
 ** 2. INET_STATE_CONNECTED  => write output
@@ -10892,7 +11251,14 @@ static int tcp_inet_output(tcp_descriptor* desc, HANDLE event)
 	async_ok(INETP(desc));
     }
     else if (IS_CONNECTED(INETP(desc))) {
-	for (;;) {
+
+#ifdef HAVE_SENDFILE
+        if(desc->tcp_add_flags & TCP_ADDF_SENDFILE) {
+            return tcp_inet_sendfile(desc);
+        }
+#endif
+
+        for (;;) {
 	    int vsize;
 	    ssize_t n;
 	    SysIOVec* iov;
diff --git a/erts/emulator/drivers/unix/unix_efile.c b/erts/emulator/drivers/unix/unix_efile.c
deleted file mode 100644
index 33e4d75ef7..0000000000
--- a/erts/emulator/drivers/unix/unix_efile.c
+++ /dev/null
@@ -1,1102 +0,0 @@
-/*
- * %CopyrightBegin%
- *
- * Copyright Ericsson AB 1997-2017. 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 Unix.
- */
-#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"
-#include <utime.h>
-#ifdef HAVE_UNISTD_H
-#include <unistd.h>
-#endif
-#ifdef HAVE_SYS_UIO_H
-#include <sys/types.h>
-#include <sys/uio.h>
-#if defined(HAVE_SENDFILE) && (defined(__FreeBSD__) || defined(__DragonFly__))
-/* Need to define __BSD_VISIBLE in order to expose prototype of sendfile */
-#define __BSD_VISIBLE 1
-#include <sys/socket.h>
-#endif
-#endif
-#if defined(HAVE_SENDFILE) && (defined(__linux__) || (defined(__sun) && defined(__SVR4)))
-#include <sys/sendfile.h>
-#endif
-
-#if defined(__APPLE__) && defined(__MACH__) && !defined(__DARWIN__)
-#define __DARWIN__ 1
-#endif
-
-#if defined(__DARWIN__) || defined(HAVE_LINUX_FALLOC_H) || defined(HAVE_POSIX_FALLOCATE)
-#include <fcntl.h>
-#endif
-
-#ifdef HAVE_LINUX_FALLOC_H
-#include <linux/falloc.h>
-#endif
-
-#ifdef SUNOS4
-#  define getcwd(buf, size) getwd(buf)
-#endif
-
-/* 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) (S_ISDIR((st).st_mode))
-#define ISREG(st) (S_ISREG((st).st_mode))
-#define ISDEV(st) (S_ISCHR((st).st_mode) || S_ISBLK((st).st_mode))
-#define ISLNK(st) (S_ISLNK((st).st_mode))
-#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')))
-
-static int check_error(int result, Efile_error* errInfo);
-
-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() {
-   return 1;
-}
-
-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
-    return check_error(mkdir(name, DIR_MODE), 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) {
-	int saved_errno = errno;
-	struct stat file_stat;
-	struct stat cwd_stat;
-
-	/*
-	 *  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 = EINVAL;
-	}
-	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. */
-{
-    if (unlink(name) == 0) {
-	return 1;
-    }
-    if (errno == EISDIR) {	/* Linux sets the wrong error code. */
-	errno = EPERM;
-    }
-    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. */
-{
-    if (rename(src, dst) == 0) {
-	return 1;
-    }
-    if (errno == ENOTEMPTY) {
-	errno = EEXIST;
-    }
-#if defined (sparc)
-    /*
-     * SunOS 4.1.4 reports overwriting a non-empty directory with a
-     * directory as EINVAL instead of EEXIST (first rule out the correct
-     * EINVAL result code for moving a directory into itself).  Must be
-     * conditionally compiled because realpath() is only defined on SunOS.
-     */
-
-    if (errno == EINVAL) {
-	char srcPath[MAXPATHLEN], dstPath[MAXPATHLEN];
-	DIR *dirPtr;
-	struct dirent *dirEntPtr;
-
-#ifdef PURIFY
-	memset(srcPath, '\0', sizeof(srcPath));
-	memset(dstPath, '\0', sizeof(dstPath));
-#endif
-
-	if ((realpath(src, srcPath) != NULL)
-		&& (realpath(dst, dstPath) != NULL)
-		&& (strncmp(srcPath, dstPath, strlen(srcPath)) != 0)) {
-	    dirPtr = opendir(dst);
-	    if (dirPtr != NULL) {
-		while ((dirEntPtr = readdir(dirPtr)) != NULL) {
-		    if ((strcmp(dirEntPtr->d_name, ".") != 0) &&
-			    (strcmp(dirEntPtr->d_name, "..") != 0)) {
-			errno = EEXIST;
-			closedir(dirPtr);
-			return check_error(-1, errInfo);
-		    }
-		}
-		closedir(dirPtr);
-	    }
-	}
-	errno = EINVAL;
-    }
-#endif	/* sparc */
-
-    if (strcmp(src, "/") == 0) {
-	/*
-	 * Alpha reports renaming / as EBUSY and Linux reports it as EACCES,
-	 * instead of EINVAL.
-	 */
-	 
-	errno = EINVAL;
-    }
-
-    /*
-     * DEC Alpha OSF1 V3.0 returns EACCES when attempting to move a
-     * file across filesystems and the parent directory of that file is
-     * not writable.  Most other systems return EXDEV.  Does nothing to
-     * correct this behavior.
-     */
-
-    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);
-	
-#ifdef SIMSPARCSOLARIS
-	/* We get "host:" prepended to the dirname - remove!. */
-	{
-	  int i = 0;
-	  int j = 0;
-	  while ((buffer[i] != ':') && (buffer[i] != '\0')) i++;
-	  if (buffer[i] == ':') {
-	    i++;
-	    while ((buffer[j++] = buffer[i++]) != '\0');
-	  }
-	}
-#endif	      
-	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. */
-
-    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;
-    }
-    if (flags & EFILE_MODE_SYNC) {
-#ifdef O_SYNC
-	mode |= O_SYNC;
-#else
-	errno = ENOTSUP;
-	return check_error(-1, errInfo);
-#endif
-    }
-
-#ifdef HAVE_FSTAT
-    while (((fd = open(name, mode, FILE_MODE)) < 0) && (errno == EINTR));
-    if (!check_error(fd, errInfo)) return 0;
-#endif
-
-    if (
-#ifdef HAVE_FSTAT
-        fstat(fd, &statbuf) < 0
-#else
-        stat(name, &statbuf) < 0
-#endif
-        ) {
-        /* statbuf is undefined: if the caller depends on it,
-           i.e. invoke_read_file(), fail the call immediately */
-        if (pSize && flags == EFILE_MODE_READ) {
-            check_error(-1, errInfo);
-#ifdef HAVE_FSTAT
-            efile_closefile(fd);
-#endif
-            return 0;
-        }
-    }
-    else if (! ISREG(statbuf)) {
-        struct stat nullstatbuf;
-	/*
-	 * For UNIX only, here is some ugly code to allow
-	 * /dev/null to be opened as a file.
-	 */
-        if ( (stat("/dev/null", &nullstatbuf) < 0)
-             || (statbuf.st_ino != nullstatbuf.st_ino)
-             || (statbuf.st_dev != nullstatbuf.st_dev) ) {
-#ifdef HAVE_FSTAT
-            efile_closefile(fd);
-#endif
-	    errno = EISDIR;
-	    return check_error(-1, errInfo);
-        }
-    }
-
-#ifndef HAVE_FSTAT
-    while (((fd = open(name, mode, FILE_MODE)) < 0) && (errno == EINTR));
-    if (!check_error(fd, errInfo)) return 0;
-#endif
-
-    *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)
-{
-    close(fd);
-}
-
-int
-efile_fdatasync(Efile_error *errInfo, /* Where to return error codes. */
-	    int fd)               /* File descriptor for file to sync data. */
-{
-#if defined(HAVE_FDATASYNC) && !defined(__DARWIN__)
-    return check_error(fdatasync(fd), errInfo);
-#else
-    return efile_fsync(errInfo, fd);
-#endif
-}
-
-int
-efile_fsync(Efile_error *errInfo, /* Where to return error codes. */
-	    int fd)               /* File descriptor for file to sync. */
-{
-#ifdef NO_FSYNC
-  undefined fsync /* XXX: Really? */
-#else
-#if defined(__DARWIN__) && defined(F_FULLFSYNC)
-    return check_error(fcntl(fd, F_FULLFSYNC), errInfo);
-#else
-    return check_error(fsync(fd), errInfo);
-#endif /* __DARWIN__ */
-#endif /* NO_FSYNC */
-}
-
-int
-efile_fileinfo(Efile_error* errInfo, Efile_info* pInfo, 
-	       char* name, int info_for_link)
-{
-    struct stat statbuf;	/* Information about the file */
-    int result;
-
-    if (info_for_link) {
-	result = lstat(name, &statbuf);
-    } else {
-	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   = (Sint64)statbuf.st_atime;
-    pInfo->modifyTime   = (Sint64)statbuf.st_mtime;
-    pInfo->cTime        = (Sint64)statbuf.st_ctime;
-
-    pInfo->mode         = statbuf.st_mode;
-    pInfo->links        = statbuf.st_nlink;
-    pInfo->major_device = statbuf.st_dev;
-    pInfo->minor_device = statbuf.st_rdev;
-    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)
-{
-    struct utimbuf tval;
-
-    /*
-     * 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 (chown(name, pInfo->uid, pInfo->gid) && errno != EPERM) {
-	return check_error(-1, errInfo);
-    }
-
-    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);
-	    }
-	}
-    }
-
-    tval.actime  = (time_t)pInfo->accessTime;
-    tval.modtime = (time_t)pInfo->modifyTime;
-
-    return check_error(utime(name, &tval), errInfo);
-}
-
-
-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. */
-
-    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 */
-{
-    int cnt = 0;                     /* Buffers so far written */
-
-    ASSERT(iovcnt >= 0);
-    
-    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 */
-#ifdef HAVE_WRITEV
-	    int b = iovcnt - cnt;        /* Buffers to write */
-	    /* Use as many buffers as MAXIOV allows */
-	    if (b > MAXIOV)
-		b = MAXIOV;
-	    if (b > 1) {
-		do {
-		    w = writev(fd, &iov[cnt], b);
-		} while (w < 0 && errno == EINTR);
-		if (w < 0 && errno == EINVAL) {
-		    goto single_write;
-		}
-	    } else
-	    single_write:
-		/* Degenerated io vector - use regular write */
-#endif
-		{
-		    do {
-			size_t iov_len = iov[cnt].iov_len;
-			size_t limit = 1024*1024*1024; /* 1GB */
-			if (iov_len > limit) {
-			    iov_len = limit;
-			}
-			w = write(fd, iov[cnt].iov_base, iov_len);
-		    } while (w < 0 && errno == EINTR);
-		    ASSERT(w <= iov[cnt].iov_len ||
-			   (w == -1 && errno != EINTR));
-		}
-	    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 = ((char *)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;
-
-    for (;;) {
-	if ((n = read(fd, buf, count)) >= 0)
-	    break;
-	else if (errno != EINTR)
-	    return check_error(-1, errInfo);
-    }
-    *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. */
-{
-#if defined(HAVE_PREAD) && defined(HAVE_PWRITE)
-    ssize_t n;
-    off_t off = (off_t) offset;
-    if (off != offset) {
-	errno = EINVAL;
-	return check_error(-1, errInfo);
-    }
-    for (;;) {
-	if ((n = pread(fd, buf, count, offset)) >= 0)
-	    break;
-	else if (errno != EINTR)
-	    return check_error(-1, errInfo);
-    }
-    *pBytesRead = (size_t) n;
-    return 1;
-#else
-    {
-	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;
-	}
-    }
-#endif
-}
-
-
-
-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 */
-{ 
-#if defined(HAVE_PREAD) && defined(HAVE_PWRITE)
-    ssize_t written;		    /* Bytes written in last operation. */
-    off_t off = (off_t) offset;
-    if (off != offset) {
-	errno = EINVAL;
-	return check_error(-1, errInfo);
-    }
-    
-    while (count > 0) {
-	if ((written = pwrite(fd, buf, count, offset)) < 0) {
-	    if (errno != EINTR)
-		return check_error(-1, errInfo);
-	    else
-		written = 0;
-	}
-	ASSERT(written <= count);
-	buf += written;
-	count -= written;
-	offset += written;
-    }
-    return 1;
-#else  /* For unix systems that don't support pread() and pwrite() */    
-    {
-	int res = efile_seek(errInfo, fd, offset, EFILE_SEEK_SET, NULL);
-	
-	if (res) {
-	    return efile_write(errInfo, 0, fd, buf, count);
-	} else {
-	    return res;
-	}
-    }
-#endif
-}
-
-
-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;
-
-    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);
-    }
-    off = (off_t) offset;
-    if (off != offset) {
-	errno = EINVAL;
-	return check_error(-1, errInfo);
-    }
-    
-    errno = 0;
-    result = lseek(fd, off, origin);
-
-    /*
-     * Note that the man page for lseek (on SunOs 5) says:
-     * 
-     * "if fildes is a remote file  descriptor  and  offset  is
-     * negative,  lseek()  returns  the  file pointer even if it is
-     * negative."
-     */
-
-    if (result < 0 && errno == 0)
-	errno = EINVAL;
-    if (result < 0)
-	return check_error(-1, errInfo);
-    if (new_location) {
-	*new_location = result;
-    }
-    return 1;
-}
-
-
-int
-efile_truncate_file(Efile_error* errInfo, int *fd, int flags)
-{
-#ifndef NO_FTRUNCATE
-    off_t offset;
-
-    return check_error((offset = lseek(*fd, 0, 1)) >= 0 &&
-		       ftruncate(*fd, offset) == 0 ? 1 : -1,
-		       errInfo);
-#else
-    return 1;
-#endif
-}
-
-int
-efile_readlink(Efile_error* errInfo, char* name, char* buffer, size_t size)
-{
-    int len;
-    ASSERT(size > 0);
-    len = readlink(name, buffer, size-1);
-    if (len == -1) {
-	return check_error(-1, errInfo);
-    }
-    buffer[len] = '\0';
-    return 1;
-}
-
-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)
-{
-    return check_error(link(old, new), errInfo);
-}
-
-int
-efile_symlink(Efile_error* errInfo, char* old, char* new)
-{
-    return check_error(symlink(old, new), errInfo);
-}
-
-int
-efile_fadvise(Efile_error* errInfo, int fd, Sint64 offset,
-		  Sint64 length, int advise)
-{
-#ifdef HAVE_POSIX_FADVISE
-    return check_error(posix_fadvise(fd, offset, length, advise), errInfo);
-#else
-    return check_error(0, errInfo);
-#endif
-}
-
-#ifdef HAVE_SENDFILE
-/* For some reason the maximum size_t cannot be used as the max size
-   3GB seems to work on all platforms */
-#define SENDFILE_CHUNK_SIZE ((1UL << 30) -1)
-
-/*
- * sendfile: The implementation of the sendfile system call varies
- * a lot on different *nix platforms so to make the api similar in all
- * we have to emulate some things in linux and play with variables on
- * bsd/darwin.
- *
- * All of the calls will split a command which tries to send more than
- * SENDFILE_CHUNK_SIZE of data at once.
- *
- * On platforms where *nbytes of 0 does not mean the entire file, this is
- * simulated.
- *
- * It could be possible to implement header/trailer in sendfile. Though
- * you would have to emulate it in linux and on BSD/Darwin some complex
- * calculations have to be made when using a non blocking socket to figure
- * out how much of the header/file/trailer was sent in each command.
- *
- * The semantics of the API is this:
- * Return value: 1 if all data was sent and the function does not need to
- * be called again. 0 if an error occures OR if there is more data which
- * has to be sent (EAGAIN or EINTR will be set appropriately)
- *
- * The amount of data written in a call is returned through nbytes.
- *
- */
-
-int
-efile_sendfile(Efile_error* errInfo, int in_fd, int out_fd,
-	       off_t *offset, Uint64 *nbytes, struct t_sendfile_hdtl* hdtl)
-{
-    Uint64 written = 0;
-#if defined(__linux__)
-    ssize_t retval;
-    do {
-      /* check if *nbytes is 0 or greater than chunk size */
-      if (*nbytes == 0 || *nbytes > SENDFILE_CHUNK_SIZE)
-	retval = sendfile(out_fd, in_fd, offset, SENDFILE_CHUNK_SIZE);
-      else
-	retval = sendfile(out_fd, in_fd, offset, *nbytes);
-      if (retval > 0) {
-	written += retval;
-	*nbytes -= retval;
-      }
-    } while (retval == SENDFILE_CHUNK_SIZE);
-    if (written != 0) {
-      /* -1 is not returned by the linux API so we have to simulate it */
-      retval = -1;
-      errno = EAGAIN;
-    }
-#elif defined(__sun) && defined(__SVR4) && defined(HAVE_SENDFILEV)
-    ssize_t retval;
-    size_t len;
-    sendfilevec_t fdrec;
-    fdrec.sfv_fd = in_fd;
-    fdrec.sfv_flag = 0;
-    do {
-      fdrec.sfv_off = *offset;
-      len = 0;
-      /* check if *nbytes is 0 or greater than chunk size */
-      if (*nbytes == 0 || *nbytes > SENDFILE_CHUNK_SIZE)
-	fdrec.sfv_len = SENDFILE_CHUNK_SIZE;
-      else
-	fdrec.sfv_len = *nbytes;
-
-      retval = sendfilev(out_fd, &fdrec, 1, &len);
-
-      if (retval == -1 && errno == EINVAL) {
-          /* On some solaris versions (I've seen it on SunOS 5.10),
-             using a sfv_len larger then a filesize will result in
-             a -1 && errno == EINVAL return. We translate this so
-             a successful send of the data.*/
-          retval = len;
-      }
-
-      if (retval != -1 || errno == EAGAIN || errno == EINTR) {
-        *offset += len;
-	*nbytes -= len;
-	written += len;
-      }
-    } while (len == SENDFILE_CHUNK_SIZE);
-#elif defined(__DARWIN__)
-    int retval;
-    off_t len;
-    do {
-      /* check if *nbytes is 0 or greater than chunk size */
-      if(*nbytes > SENDFILE_CHUNK_SIZE)
-	len = SENDFILE_CHUNK_SIZE;
-      else
-	len = *nbytes;
-      retval = sendfile(in_fd, out_fd, *offset, &len, NULL, 0);
-      if (retval != -1 || errno == EAGAIN || errno == EINTR) {
-        *offset += len;
-	*nbytes -= len;
-	written += len;
-      }
-    } while (len == SENDFILE_CHUNK_SIZE);
-#elif defined(__FreeBSD__) || defined(__DragonFly__)
-    off_t len;
-    int retval;
-    do {
-      if (*nbytes > SENDFILE_CHUNK_SIZE)
-	retval = sendfile(in_fd, out_fd, *offset, SENDFILE_CHUNK_SIZE,
-			  NULL, &len, 0);
-      else
-	retval = sendfile(in_fd, out_fd, *offset, *nbytes, NULL, &len, 0);
-      if (retval != -1 || errno == EAGAIN || errno == EINTR) {
-	*offset += len;
-	*nbytes -= len;
-	written += len;
-      }
-    } while(len == SENDFILE_CHUNK_SIZE);
-#endif
-    *nbytes = written;
-    return check_error(retval, errInfo);
-}
-#endif /* HAVE_SENDFILE */
-
-#ifdef HAVE_POSIX_FALLOCATE
-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;
-}
-#endif /* HAVE_POSIX_FALLOCATE */
-
-int
-efile_fallocate(Efile_error* errInfo, int fd, Sint64 offset, Sint64 length)
-{
-#if defined HAVE_FALLOCATE
-    /* Linux specific, more efficient than posix_fallocate. */
-    int ret;
-
-    do {
-        ret = fallocate(fd, FALLOC_FL_KEEP_SIZE, (off_t) offset, (off_t) length);
-    } while (ret != 0 && errno == EINTR);
-
-#if defined HAVE_POSIX_FALLOCATE
-    /* Fallback to posix_fallocate if available. */
-    if (ret != 0) {
-        ret = call_posix_fallocate(fd, offset, length);
-    }
-#endif
-
-    return check_error(ret, errInfo);
-#elif defined F_PREALLOCATE
-    /* Mac OS X specific, equivalent to posix_fallocate. */
-    int ret;
-    fstore_t fs;
-
-    memset(&fs, 0, sizeof(fs));
-    fs.fst_flags = F_ALLOCATECONTIG;
-    fs.fst_posmode = F_VOLPOSMODE;
-    fs.fst_offset = (off_t) offset;
-    fs.fst_length = (off_t) length;
-
-    ret = fcntl(fd, F_PREALLOCATE, &fs);
-
-    if (-1 == ret) {
-        fs.fst_flags = F_ALLOCATEALL;
-        ret = fcntl(fd, F_PREALLOCATE, &fs);
-
-#if defined HAVE_POSIX_FALLOCATE
-        /* Fallback to posix_fallocate if available. */
-        if (-1 == ret) {
-            ret = call_posix_fallocate(fd, offset, length);
-        }
-#endif
-    }
-
-    return check_error(ret, errInfo);
-#elif defined HAVE_POSIX_FALLOCATE
-    /* Other Unixes, use posix_fallocate if available. */
-    return check_error(call_posix_fallocate(fd, offset, length), errInfo);
-#else
-    errno = ENOTSUP;
-    return check_error(-1, errInfo);
-#endif
-}
diff --git a/erts/emulator/drivers/win32/win_efile.c b/erts/emulator/drivers/win32/win_efile.c
deleted file mode 100644
index 2d366b5833..0000000000
--- a/erts/emulator/drivers/win32/win_efile.c
+++ /dev/null
@@ -1,2058 +0,0 @@
-/*
- * %CopyrightBegin%
- *
- * Copyright Ericsson AB 1997-2016. 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%
- */
-/*
- * Purpose: Provides file and directory operations for Windows.
- */
-
-#include <windows.h>
-#ifdef HAVE_CONFIG_H
-#  include "config.h"
-#endif
-#include "sys.h"
-#include <ctype.h>
-#include <wchar.h>
-#include "erl_efile.h"
-
-#define DBG_TRACE_MASK 0
-/* 1 = file name ops
- * 2 = file descr ops
- * 4 = errors
- * 8 = path name conversion
- */
-#if !DBG_TRACE_MASK
-#  define DBG_TRACE(M,S)
-#  define DBG_TRACE1(M,FMT,A)
-#  define DBG_TRACE2(M,FMT,A,B)
-#else
-#  define DBG_TRACE(M,S) do { if ((M)&DBG_TRACE_MASK) fwprintf(stderr, L"DBG_TRACE %d: %s\r\n", __LINE__, (WCHAR*)(S)); }while(0)
-#  define DBG_TRACE1(M,FMT,A) do { if ((M)&DBG_TRACE_MASK) fwprintf(stderr, L"DBG_TRACE %d: " L##FMT L"\r\n", __LINE__, (A)); }while(0)
-#  define DBG_TRACE2(M,FMT,A,B) do { if ((M)&DBG_TRACE_MASK) fwprintf(stderr, L"DBG_TRACE %d: " L##FMT L"\r\n", __LINE__, (A), (B)); }while(0)
-#endif
-
-/*
- * Microsoft-specific function to map a WIN32 error code to a Posix errno.
- */
-
-#define ISSLASH(a)  ((a) == L'\\' || (a) == L'/')
-#define ISDIR(st) (((st).st_mode&S_IFMT) == S_IFDIR)
-#define ISREG(st) (((st).st_mode&S_IFMT) == S_IFREG)
-
-#define IS_DOT_OR_DOTDOT(s) \
-    ((s)[0] == L'.' && ((s)[1] == L'\0' || ((s)[1] == L'.' && (s)[2] == L'\0')))
-
-#define FILE_SHARE_FLAGS (FILE_SHARE_READ | FILE_SHARE_WRITE | FILE_SHARE_DELETE)
-
-#ifndef INVALID_FILE_ATTRIBUTES
-#define INVALID_FILE_ATTRIBUTES ((DWORD) 0xFFFFFFFF)
-#endif
-
-#define TICKS_PER_SECOND (10000000ULL)
-#define EPOCH_DIFFERENCE (11644473600LL)
-
-#define FILETIME_TO_EPOCH(epoch, ft) \
-    do { \
-	ULARGE_INTEGER ull; \
-	ull.LowPart  = (ft).dwLowDateTime; \
-	ull.HighPart = (ft).dwHighDateTime; \
-	(epoch) = ((ull.QuadPart / TICKS_PER_SECOND) - EPOCH_DIFFERENCE); \
-    } while(0)
-
-#define EPOCH_TO_FILETIME(ft, epoch) \
-    do { \
-	ULARGE_INTEGER ull; \
-	ull.QuadPart = (((epoch) + EPOCH_DIFFERENCE) * TICKS_PER_SECOND); \
-	(ft).dwLowDateTime = ull.LowPart; \
-	(ft).dwHighDateTime = ull.HighPart; \
-    } while(0)
-
-
-static int check_error(int result, Efile_error* errInfo);
-static int set_error(Efile_error* errInfo);
-static int set_os_errno(Efile_error* errInfo, DWORD os_errno);
-static int is_root_unc_name(const WCHAR *path);
-static int extract_root(WCHAR *name);
-static unsigned short dos_to_posix_mode(int attr, const WCHAR *name);
-
-
-struct wpath_tmp_buffer {
-    struct wpath_tmp_buffer* next;
-    WCHAR buffer[1];
-};
-
-typedef struct {
-    Efile_error* errInfo;
-    struct wpath_tmp_buffer* buf_list;
-}Efile_call_state;
-
-static void call_state_init(Efile_call_state* state, Efile_error* errInfo)
-{
-    state->errInfo = errInfo;
-    state->buf_list = NULL;
-}
-static WCHAR* wpath_tmp_alloc(Efile_call_state* state, size_t len)
-{
-    size_t sz = offsetof(struct wpath_tmp_buffer, buffer)
-	+ (len+1)*sizeof(WCHAR);
-    struct wpath_tmp_buffer* p = driver_alloc(sz);
-    p->next = state->buf_list;
-    state->buf_list = p;
-    return p->buffer;
-}
-static void call_state_free(Efile_call_state* state)
-{
-    while(state->buf_list) {
-	struct wpath_tmp_buffer* next = state->buf_list->next;
-	driver_free(state->buf_list);
-	state->buf_list = next;
-    }
-}
-static WCHAR* get_cwd_wpath_tmp(Efile_call_state* state)
-{
-    WCHAR dummy;
-    DWORD size = GetCurrentDirectoryW(0, &dummy);
-    WCHAR* ret = NULL;
-
-    if (size) {
-	ret = wpath_tmp_alloc(state, size);
-	if (!GetCurrentDirectoryW(size, ret)) {
-	    ret = NULL;
-	}
-    }
-    return ret;
-}
-static WCHAR* get_full_wpath_tmp(Efile_call_state* state,
-                                 const WCHAR* file,
-				 WCHAR** file_part,
-				 DWORD extra)
-{
-    WCHAR dummy;
-    DWORD size = GetFullPathNameW(file, 0, &dummy, NULL);
-    WCHAR* ret = NULL;
-
-    if (size) {
-	int ok;
-	ret = wpath_tmp_alloc(state, size + extra);
-	if (file_part) {
-	    ok = (GetFullPathNameW(file, size, ret, file_part) != 0);
-	}
-	else {
-	    ok = (_wfullpath(ret, file, size) != NULL);
-	}
-	if (!ok) {
-	    ret = NULL;
-	}
-    }
-    return ret;
-}
-
-static void ensure_wpath_max(Efile_call_state* state, WCHAR** pathp, size_t max);
-static int do_rmdir(Efile_call_state*, char* name);
-static int do_rename(Efile_call_state*, char* src, char* dst);
-static int do_readdir(Efile_call_state*, char* name, EFILE_DIR_HANDLE*, char* buffer, size_t *size);
-static int do_fileinfo(Efile_call_state*, Efile_info*, char* orig_name, int info_for_link);
-static char* do_readlink(Efile_call_state*, char* name, char* buffer, size_t size);
-static int do_altname(Efile_call_state*, char* orig_name, char* buffer, size_t size);
-
-
-static int errno_map(DWORD last_error) {
-
-    switch (last_error) {
-    case ERROR_SUCCESS:
-	return 0;
-    case ERROR_INVALID_FUNCTION:
-    case ERROR_INVALID_DATA:
-    case ERROR_INVALID_PARAMETER:
-    case ERROR_INVALID_TARGET_HANDLE:
-    case ERROR_INVALID_CATEGORY:
-    case ERROR_NEGATIVE_SEEK:
-	return EINVAL;
-    case ERROR_DIR_NOT_EMPTY:
-	return EEXIST;
-    case ERROR_BAD_FORMAT:
-	return ENOEXEC;
-    case ERROR_PATH_NOT_FOUND:
-    case ERROR_FILE_NOT_FOUND:
-    case ERROR_NO_MORE_FILES:
-	return ENOENT;
-    case ERROR_TOO_MANY_OPEN_FILES:
-	return EMFILE;
-    case ERROR_ACCESS_DENIED:
-    case ERROR_INVALID_ACCESS:
-    case ERROR_CURRENT_DIRECTORY:
-    case ERROR_SHARING_VIOLATION:
-    case ERROR_LOCK_VIOLATION:
-    case ERROR_INVALID_PASSWORD:
-    case ERROR_DRIVE_LOCKED:
-	return EACCES;
-    case ERROR_INVALID_HANDLE:
-	return EBADF;
-    case ERROR_NOT_ENOUGH_MEMORY:
-    case ERROR_OUTOFMEMORY:
-    case ERROR_OUT_OF_STRUCTURES:
-	return ENOMEM;
-    case ERROR_INVALID_DRIVE:
-    case ERROR_BAD_UNIT:
-    case ERROR_NOT_READY:
-    case ERROR_REM_NOT_LIST:
-    case ERROR_DUP_NAME:
-    case ERROR_BAD_NETPATH:
-    case ERROR_NETWORK_BUSY:
-    case ERROR_DEV_NOT_EXIST:
-    case ERROR_BAD_NET_NAME:
-	return ENXIO;
-    case ERROR_NOT_SAME_DEVICE:
-	return EXDEV;
-    case ERROR_WRITE_PROTECT:
-	return EROFS;
-    case ERROR_BAD_LENGTH:
-    case ERROR_BUFFER_OVERFLOW:
-	return E2BIG;
-    case ERROR_SEEK:
-    case ERROR_SECTOR_NOT_FOUND:
-	return ESPIPE;
-    case ERROR_NOT_DOS_DISK:
-	return ENODEV;
-    case ERROR_GEN_FAILURE:
-	return ENODEV;
-    case ERROR_SHARING_BUFFER_EXCEEDED:
-    case ERROR_NO_MORE_SEARCH_HANDLES:
-	return EMFILE;
-    case ERROR_HANDLE_EOF:
-    case ERROR_BROKEN_PIPE:
-	return EPIPE;
-    case ERROR_HANDLE_DISK_FULL:
-    case ERROR_DISK_FULL:
-	return ENOSPC;
-    case ERROR_NOT_SUPPORTED:
-	return ENOTSUP;
-    case ERROR_FILE_EXISTS:
-    case ERROR_ALREADY_EXISTS:
-    case ERROR_CANNOT_MAKE:
-	return EEXIST;
-    case ERROR_ALREADY_ASSIGNED:
-	return EBUSY;
-    case ERROR_NO_PROC_SLOTS:
-	return EAGAIN;
-    case ERROR_CANT_RESOLVE_FILENAME:
-	return EMLINK;
-    case ERROR_PRIVILEGE_NOT_HELD:
-	return EPERM;
-    case ERROR_ARENA_TRASHED:
-    case ERROR_INVALID_BLOCK:
-    case ERROR_BAD_ENVIRONMENT:
-    case ERROR_BAD_COMMAND:
-    case ERROR_CRC:
-    case ERROR_OUT_OF_PAPER:
-    case ERROR_READ_FAULT:
-    case ERROR_WRITE_FAULT:
-    case ERROR_WRONG_DISK:
-    case ERROR_NET_WRITE_FAULT:
-	return EIO;
-    default: /* not to do with files I expect. */
-	return EIO;
-    }	
-}
-
-static int
-check_error(int result, Efile_error* errInfo)
-{
-    if (result < 0) {
-	errInfo->posix_errno = errno;
-	errInfo->os_errno = GetLastError();
-	DBG_TRACE2(4, "ERROR os_error=%d errno=%d @@@@@@@@@@@@@@@@@@@@@@@@@@@@",
-	        errInfo->os_errno, errInfo->posix_errno);
-	return 0;
-    }
-    return 1;
-}
-
-static void
-save_last_error(Efile_error* errInfo)
-{
-    errInfo->posix_errno = errno;
-    errInfo->os_errno = GetLastError();
-    DBG_TRACE2(4, "ERROR os_error=%d errno=%d $$$$$$$$$$$$$$$$$$$$$$$$$$$$$",
-	errInfo->os_errno, errInfo->posix_errno);
-}
-
-
-/*
- * Fills the provided error information structure with information
- * with the error code given by GetLastError() and its corresponding
- * Posix error number.
- *
- * Returns 0.
- */
-
-static int
-set_error(Efile_error* errInfo)
-{
-    set_os_errno(errInfo, GetLastError());
-    return 0;
-}
-
-
-static int
-set_os_errno(Efile_error* errInfo, DWORD os_errno)
-{
-    errInfo->os_errno = os_errno;
-    errInfo->posix_errno = errno_map(os_errno);
-    DBG_TRACE2(4, "ERROR os_error=%d errno=%d ############################",
-	    errInfo->os_errno, errInfo->posix_errno);
-    return 0;
-}
-
-int
-efile_init() {
-   return 1;
-}
-
-/*
- * A writev with Unix semantics, but with Windows arguments 
- */
-static int 
-win_writev(Efile_error* errInfo,
-	   HANDLE fd,                  /* handle to file */
-	   FILE_SEGMENT_ELEMENT iov[], /* array of buffer pointers */
-	   DWORD *size)                /* number of bytes to write */
-{
-    OVERLAPPED ov;
-    ov.Offset = 0L;
-    ov.OffsetHigh = 0L;
-    ov.hEvent = CreateEvent(NULL, TRUE, FALSE, NULL);
-    if (ov.hEvent == NULL)
-	return set_error(errInfo);
-    if (! write_file_gather(fd, iov, *size, NULL, &ov))
-	return set_error(errInfo);
-    if (WaitForSingleObject(ov.hEvent, INFINITE) != WAIT_OBJECT_0)
-	return set_error(errInfo);
-    if (! GetOverlappedResult(fd, &ov, size, FALSE))
-	return set_error(errInfo);
-    return 1;
-}
-
-
-/* Check '*pathp' and convert it if needed to something that windows will accept.
- * Typically use UNC path with \\?\ prefix if absolute path is longer than 260.
- */
-static void ensure_wpath(Efile_call_state* state, WCHAR** pathp)
-{
-    ensure_wpath_max(state, pathp, MAX_PATH);
-}
-
-static void ensure_wpath_max(Efile_call_state* state, WCHAR** pathp, size_t max)
-{
-    WCHAR* path = *pathp;
-    WCHAR* p;
-    size_t len = wcslen(path);
-    int unc_fixup = 0;
-
-    if (path[0] == 0) {
-	DBG_TRACE(8, L"Let empty path pass through");
-	return;
-    }
-
-    DBG_TRACE1(8,"IN: %s", path);
-
-    if (path[1] == L':' && ISSLASH(path[2])) { /* absolute path */
-	if (len >= max) {
-	    WCHAR *src, *dst;
-
-	    *pathp = wpath_tmp_alloc(state, 4+len+1);
-	    dst = *pathp;
-	    wcscpy(dst, L"\\\\?\\");
-	    for (src=path,dst+=4; *src; src++) {
-		if (*src == L'/') {
-		    if (dst[-1] != L'\\') {
-			*dst++ =  L'\\';
-		    }
-		    /*else ignore redundant slashes */
-		}
-		else
-		    *dst++ = *src;
-	    }
-	    *dst = 0;
-	    unc_fixup = 1;
-	}
-    }
-    else if (!(ISSLASH(path[0]) && ISSLASH(path[1]))) { /* relative path */
-	DWORD cwdLen = GetCurrentDirectoryW(0, NULL);
-	DWORD absLen = cwdLen + 1 + len;
-	if (absLen >= max) {
-	    WCHAR *fullPath = wpath_tmp_alloc(state, 4+4+absLen);
-	    DWORD fullLen;
-
-	    fullLen = GetFullPathNameW(path, 4 + absLen, fullPath+4, NULL);
-	    if (fullLen >= 4+absLen) {
-		*pathp = path;
-		DBG_TRACE2(8,"ensure_wpath FAILED absLen=%u %s", (int)absLen, path);
-		return;
-	    }
-	    /* GetFullPathNameW can return paths longer than MAX_PATH without the \\?\ prefix.
-	     * At least seen on Windows 7. Go figure...
-	     */
-	    if (fullLen >= max && wcsncmp(fullPath+4, L"\\\\?\\", 4) != 0) {
-		wcsncpy(fullPath, L"\\\\?\\", 4);
-		*pathp = fullPath;
-	    }
-	    else {
-		*pathp = fullPath + 4;
-	    }
-	}
-    }
-
-    if (unc_fixup) {
-	WCHAR* endp;
-
-	p = *pathp;
-	len = wcslen(p);
-	endp = p + len;
-	if (len > 4) {
-	    p += 4;
-	    while (*p) {
-		if (p[0] == L'\\' && p[1] == L'.') {
-		    if (p[2] == L'\\' || !p[2]) { /* single dot */
-			wmemmove(p, p+2, (&endp[1] - &p[2]));
-			endp -= 2;
-		    }
-		    else if (p[2] == L'.' && (p[3] == L'\\' || !p[3])) { /* double dot */
-			WCHAR* r;
-			for (r=p-1; *r == L'\\'; --r)
-			    /*skip redundant slashes*/;
-			for (; *r != L'\\'; --r)
-			    /*find start of prev directory*/;
-			if (r < *pathp + 6)
-			    break;
-			wmemmove(r, p+3, (&endp[1] - &p[3]));
-			p = r;
-		    }
-		    else p += 3;
-		}
-		else ++p;
-	    }
-	}
-    }
-    DBG_TRACE1(8,"OUT: %s", *pathp);
-}
-
-int
-efile_mkdir(Efile_error* errInfo,	/* Where to return error codes. */
-	    char* name)			/* Name of directory to create. */
-{
-    Efile_call_state state;
-    WCHAR* wname = (WCHAR*)name;
-    int ret;
-
-    DBG_TRACE(1, name);
-    call_state_init(&state, errInfo);
-    ensure_wpath_max(&state, &wname, 248); /* Yes, 248 limit for normal paths */
-
-    ret = (int) CreateDirectoryW(wname, NULL);
-    if (!ret)
-	set_error(errInfo);
-
-    call_state_free(&state);
-    return ret;
-}
-
-int
-efile_rmdir(Efile_error* errInfo,	/* Where to return error codes. */
-	    char* name)			/* Name of directory to delete. */
-{
-    Efile_call_state state;
-    int ret;
-
-    DBG_TRACE(1, name);
-    call_state_init(&state, errInfo);
-    ret = do_rmdir(&state, name);
-    call_state_free(&state);
-    return ret;
-}
-
-static int do_rmdir(Efile_call_state* state, char* name)
-{
-    OSVERSIONINFO os;
-    DWORD attr;
-    WCHAR *wname = (WCHAR *) name;
-    WCHAR *buffer = NULL;
-
-    ensure_wpath(state, &wname);
-
-    if (RemoveDirectoryW(wname) != FALSE) {
-	return 1;
-    }
-    errno = errno_map(GetLastError());
-    if (errno == EACCES) {
-	attr = GetFileAttributesW(wname);
-	if (attr != (DWORD) -1) {
-	    if ((attr & FILE_ATTRIBUTE_DIRECTORY) == 0) {
-		/* 
-		 * Windows 95 reports calling RemoveDirectory on a file as an 
-		 * EACCES, not an ENOTDIR.
-		 */
-		
-		errno = ENOTDIR;
-		goto end;
-	    }
-
-	    /* 
-	     * Windows 95 reports removing a non-empty directory as
-	     * an EACCES, not an EEXIST.  If the directory is not empty,
-	     * change errno so caller knows what's going on.
-	     */
-
-	    os.dwOSVersionInfoSize = sizeof(os);
-	    GetVersionEx(&os);
-	    if (os.dwPlatformId == VER_PLATFORM_WIN32_WINDOWS) {
-		HANDLE handle;
-		WIN32_FIND_DATAW data;
-		int len = wcslen(wname);
-
-		buffer = wpath_tmp_alloc(state, len + 4);
-		wcscpy(buffer, wname);
-		if (buffer[0] && buffer[len-1] != L'\\' && buffer[len-1] != L'/') {
-		    wcscat(buffer, L"\\");
-		}
-		wcscat(buffer, L"*.*");
-		handle = FindFirstFileW(buffer, &data);
-		if (handle != INVALID_HANDLE_VALUE) {
-		    while (1) {
-			if ((wcscmp(data.cFileName, L".") != 0)
-				&& (wcscmp(data.cFileName, L"..") != 0)) {
-			    /*
-			     * Found something in this directory.
-			     */
-
-			    errno = EEXIST;
-			    break;
-			}
-			if (FindNextFileW(handle, &data) == FALSE) {
-			    break;
-			}
-		    }
-		    FindClose(handle);
-		}
-	    }
-	}
-    }
-
-    if (errno == ENOTEMPTY) {
-	/* 
-	 * Posix allows both EEXIST or ENOTEMPTY, but we'll always
-	 * return EEXIST to allow easy matching in Erlang code.
-	 */
-
-	errno = EEXIST;
-    }
-
- end:
-    save_last_error(state->errInfo);
-    return 0;
-}
-
-int
-efile_delete_file(Efile_error* errInfo,		/* Where to return error codes. */
-		  char* name)			/* Name of file to delete. */
-{
-    Efile_call_state state;
-    int ret;
-    DBG_TRACE(1, name);
-    call_state_init(&state, errInfo);
-    ret = do_delete_file(&state, name);
-    call_state_free(&state);
-    return ret;
-}
-
-static int do_delete_file(Efile_call_state* state, char* name)
-{
-    DWORD attr;
-    WCHAR *wname = (WCHAR *) name;
-
-    ensure_wpath(state, &wname);
-
-    if (DeleteFileW(wname) != FALSE) {
-	return 1;
-    }
-
-    errno = errno_map(GetLastError());
-    if (errno == EACCES) {
-        attr = GetFileAttributesW(wname);
-	if (attr != (DWORD) -1) {
-	    if (attr & FILE_ATTRIBUTE_DIRECTORY) {
-		/*
-		 * Windows NT reports removing a directory as EACCES instead
-		 * of EPERM.
-		 */
-
-		errno = EPERM;
-	    }
-	}
-    } else if (errno == ENOENT) {
-        attr = GetFileAttributesW(wname);
-	if (attr != (DWORD) -1) {
-	    if (attr & FILE_ATTRIBUTE_DIRECTORY) {
-	    	/*
-		 * Windows 95 reports removing a directory as ENOENT instead 
-		 * of EPERM.
-		 */
-
-		errno = EPERM;
-	    }
-	}
-    } else if (errno == EINVAL) {
-	/*
-	 * Windows NT reports removing a char device as EINVAL instead of
-	 * EACCES.
-	 */
-	
-	errno = EACCES;
-    }
-
-    return check_error(-1, state->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.  
- *
- *	Some possible error codes:
- *
- *	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, char* src, char* dst)
-{
-    Efile_call_state state;
-    int ret;
-    DBG_TRACE(1, src);
-    call_state_init(&state, errInfo);
-    ret = do_rename(&state, src, dst);
-    call_state_free(&state);
-    return ret;
-}
-
-static int
-do_rename(Efile_call_state* state,
-	  char* src,			/* Original name. */
-	  char* dst)			/* New name. */
-{
-    DWORD srcAttr, dstAttr;
-    WCHAR *wsrc = (WCHAR *) src;
-    WCHAR *wdst = (WCHAR *) dst;
-
-    ensure_wpath(state, &wsrc);
-    ensure_wpath(state, &wdst);
-
-    if (MoveFileW(wsrc, wdst) != FALSE) {
-	return 1;
-    }
-
-    errno = errno_map(GetLastError());
-    srcAttr = GetFileAttributesW(wsrc);
-    dstAttr = GetFileAttributesW(wdst);
-    if (srcAttr == (DWORD) -1) {
-	srcAttr = 0;
-    }
-    if (dstAttr == (DWORD) -1) {
-	dstAttr = 0;
-    }
-
-    if (errno == EBADF) {
-	errno = EACCES;
-	return check_error(-1, state->errInfo);
-    }
-    if (errno == EACCES) {
-	decode:
-	if (srcAttr & FILE_ATTRIBUTE_DIRECTORY) {
-	    WCHAR *srcPath, *dstPath;
-	    WCHAR *srcRest, *dstRest;
-	    int size;
-
-	    srcPath = get_full_wpath_tmp(state, wsrc, &srcRest, 0);
-	    if (!srcPath) {
-		save_last_error(state->errInfo);
-		return 0;
-	    }
-
-	    dstPath = get_full_wpath_tmp(state, wdst, &dstRest, 0);
-	    if (!dstPath) {
-		save_last_error(state->errInfo);
-		return 0;
-	    }
-
-	    if (srcRest == NULL) {
-		srcRest = srcPath + wcslen(srcPath);
-	    }
-	    if (_wcsnicmp(srcPath, dstPath, srcRest - srcPath) == 0) {
-		/*
-		 * Trying to move a directory into itself.
-		 */
-
-		errno = EINVAL;
-	    }
-	    if (extract_root(srcPath)) {
-		/*
-		 * Attempt to move a root directory.  Never allowed.
-		 */
-		errno = EINVAL;
-	    }
-
-	    (void) extract_root(dstPath);
-	    if (dstPath[0] == L'\0') {
-		/*
-		 * The filename was invalid.  (Don't know why,
-		 * but play it safe.)
-		 */
-		errno = EINVAL;
-	    }
-	    if (_wcsicmp(srcPath, dstPath) != 0) {
-		/*
-		 * If src is a directory and dst filesystem != src
-		 * filesystem, errno should be EXDEV.  It is very
-		 * important to get this behavior, so that the caller
-		 * can respond to a cross filesystem rename by
-		 * simulating it with copy and delete.  The MoveFile
-		 * system call already handles the case of moving a
-		 * *file* between filesystems.
-		 */
-
-		errno = EXDEV;
-	    }
-	}
-
-	/*
-	 * Other types of access failure is that dst is a read-only
-	 * filesystem, that an open file referred to src or dest, or that
-	 * src or dest specified the current working directory on the
-	 * current filesystem.  EACCES is returned for those cases.
-	 */
-
-    } else if (errno == EEXIST) {
-	/*
-	 * Reports EEXIST any time the target already exists.  If it makes
-	 * sense, remove the old file and try renaming again.
-	 */
-
-	if (srcAttr & FILE_ATTRIBUTE_DIRECTORY) {
-	    if (dstAttr & FILE_ATTRIBUTE_DIRECTORY) {
-		/*
-		 * Overwrite empty dst directory with src directory.  The
-		 * following call will remove an empty directory.  If it
-		 * fails, it's because it wasn't empty.
-		 */
-
-		if (RemoveDirectoryW(wdst)) {
-		    /*
-		     * Now that that empty directory is gone, we can try
-		     * renaming again.  If that fails, we'll put this empty
-		     * directory back, for completeness.
-		     */
-
-		    if (MoveFileW(wsrc, wdst) != FALSE) {
-			return 1;
-		    }
-
-		    /*
-		     * Some new error has occurred.  Don't know what it
-		     * could be, but report this one.
-		     */
-
-		    errno = errno_map(GetLastError());
-		    CreateDirectoryW(wdst, NULL);
-		    SetFileAttributesW(wdst, dstAttr);
-		    if (errno == EACCES) {
-			/*
-			 * Decode the EACCES to a more meaningful error.
-			 */
-
-			goto decode;
-		    }
-		}
-	    } else {	/* (dstAttr & FILE_ATTRIBUTE_DIRECTORY) == 0 */
-		errno = ENOTDIR;
-	    }
-	} else {    /* (srcAttr & FILE_ATTRIBUTE_DIRECTORY) == 0 */
-	    if (dstAttr & FILE_ATTRIBUTE_DIRECTORY) {
-		errno = EISDIR;
-	    } else {
-		/*
-		 * Overwrite existing file by:
-		 * 
-		 * 1. Rename existing file to temp name.
-		 * 2. Rename old file to new name.
-		 * 3. If success, delete temp file.  If failure,
-		 *    put temp file back to old name.
-		 */
-
-		WCHAR *tempName;
-		int result;
-		WCHAR *rest;
-		
-		tempName = get_full_wpath_tmp(state, wdst, &rest, 14);
-		if (!tempName || !rest) {
-		    save_last_error(state->errInfo);
-		    return 0;
-		}
-
-		*rest = L'\0';
-		result = -1;
-		if (GetTempFileNameW(tempName, L"erlr", 0, tempName) != 0) {
-		    /*
-		     * Strictly speaking, need the following DeleteFile and
-		     * MoveFile to be joined as an atomic operation so no
-		     * other app comes along in the meantime and creates the
-		     * same temp file.
-		     */
-		     
-		    DeleteFileW(tempName);
-		    if (MoveFileW(wdst, tempName) != FALSE) {
-			if (MoveFileW(wsrc, wdst) != FALSE) {
-			    SetFileAttributesW(tempName, FILE_ATTRIBUTE_NORMAL);
-			    DeleteFileW(tempName);
-			    return 1;
-			} else {
-			    DeleteFileW(wdst);
-			    MoveFileW(tempName, wdst);
-			}
-		    } 
-
-		    /*
-		     * Can't backup dst file or move src file.  Return that
-		     * error.  Could happen if an open file refers to dst.
-		     */
-
-		    errno = errno_map(GetLastError());
-		    if (errno == EACCES) {
-			/*
-			 * Decode the EACCES to a more meaningful error.
-			 */
-			goto decode;
-		    }
-		}
-		return result;
-	    }
-	}
-    }
-    return check_error(-1, state->errInfo);
-}
-
-int
-efile_chdir(Efile_error* errInfo,	/* Where to return error codes. */
-	    char* name)			/* Name of directory to make current. */
-{

-    /* We don't even try to handle long paths here

-     * as current working directory is always limited to MAX_PATH

-     * even if we use UNC paths and SetCurrentDirectoryW()

-     */

-    int success = check_error(_wchdir((WCHAR *) name), errInfo);

-    if (!success && errInfo->posix_errno == EINVAL)

-	/* POSIXification of errno */

-	errInfo->posix_errno = ENOENT;
-    return success;
-}
-
-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. */
-{
-    WCHAR *wbuffer = (WCHAR *) buffer;
-    size_t wbuffer_size = size / 2; 
-    DBG_TRACE(1, L"#getdcwd#");
-    if (_wgetdcwd(drive, wbuffer, wbuffer_size) == NULL) {
-	return check_error(-1, errInfo);
-    }
-    DBG_TRACE1(8, "getdcwd OS=%s", wbuffer);
-    if (wcsncmp(wbuffer, L"\\\\?\\", 4) == 0) {
-	wmemmove(wbuffer, wbuffer+4, wcslen(wbuffer+4)+1);
-    }
-    for ( ; *wbuffer; wbuffer++) 
-	if (*wbuffer == L'\\')
-	    *wbuffer = L'/';
-    DBG_TRACE1(8, "getdcwd ERLANG=%s", (WCHAR*)buffer);
-    return 1;
-}
-
-int
-efile_readdir(Efile_error* errInfo, char* name, EFILE_DIR_HANDLE* dir_handle,
-	      char* buffer, size_t *size)
-{
-    Efile_call_state state;
-    int ret;
-    DBG_TRACE(dir_handle?2:1, name);
-    call_state_init(&state, errInfo);
-    ret = do_readdir(&state, name, dir_handle, buffer, size);
-    call_state_free(&state);
-    return ret;
-}
-
-static int do_readdir(Efile_call_state* state,
-                      char* name,                   /* Name of directory to list */
-		      EFILE_DIR_HANDLE* dir_handle, /* Handle of opened directory or NULL */
-		      char* buffer,                 /* Buffer to put one filename in */
-		      size_t *size)                 /* in-out size of buffer/size of filename excluding zero
-					               termination in bytes*/
-{
-    HANDLE dir;			/* Handle to directory. */
-    WIN32_FIND_DATAW findData;	/* Data found by FindFirstFile() or FindNext(). */
-    /* Alignment is not honored, this works on x86 because of alignment fixup by processor.
-       Not perfect, but faster than alinging by hand (really) */
-    WCHAR *wbuffer = (WCHAR *) buffer;
-
-    /*
-     * First time we must setup everything.
-     */
-
-    if (*dir_handle == NULL) {
-	WCHAR *wname = (WCHAR *) name;
-	WCHAR* wildcard;
-	int length;
-	WCHAR* s;
-
-	ensure_wpath_max(state, &wname, MAX_PATH-2);
-	length = wcslen(wname);
-
-	wildcard = wpath_tmp_alloc(state, length+3);
-
-	wcscpy(wildcard, wname);
-	s = wildcard+length-1;
-	if (*s != L'/' && *s != L'\\')
-	    *++s = L'\\';
-	*++s = L'*';
-	*++s = L'\0';
-	DEBUGF(("Reading %ws\n", wildcard));
-	dir = FindFirstFileW(wildcard, &findData);
-	if (dir == INVALID_HANDLE_VALUE) {
-	    set_error(state->errInfo);
-	    return 0;
-	}
-	*dir_handle = (EFILE_DIR_HANDLE) dir;
-
-	if (!IS_DOT_OR_DOTDOT(findData.cFileName)) {
-	    wcscpy(wbuffer, findData.cFileName);
-	    *size = wcslen(wbuffer)*2;
-	    return 1;
-	}
-    }
-
-    /*
-     * Retrieve the name of the next file using the directory handle.
-     */
-
-    dir = (HANDLE) *dir_handle;
-
-    for (;;) {
-	if (FindNextFileW(dir, &findData)) {
-	    if (IS_DOT_OR_DOTDOT(findData.cFileName))
-		continue;
-	    wcscpy(wbuffer, findData.cFileName);
-	    *size = wcslen(wbuffer)*2;
-	    return 1;
-	}
-
-	if (GetLastError() == ERROR_NO_MORE_FILES) {
-	    state->errInfo->posix_errno = state->errInfo->os_errno = 0;
-	}
-	else {
-	    set_error(state->errInfo);
-	}
-	FindClose(dir);
-	return 0;
-    }
-}
-
-int
-efile_openfile(Efile_error* errInfo, char* name, int flags, int* pfd, Sint64* pSize)
-{
-    Efile_call_state state;
-    int ret;
-    DBG_TRACE1(1, "openfile(%s)", name);
-    call_state_init(&state, errInfo);
-    ret = do_openfile(&state, name, flags, pfd, pSize);
-    call_state_free(&state);
-    return ret;
-}
-
-static
-int do_openfile(Efile_call_state* state,        /* Where to return error codes. */
-	        char* name,			/* Name of directory to open. */
-	        int flags,			/* Flags to use for opening. */
-	        int* pfd,			/* Where to store the file descriptor. */
-	        Sint64* pSize)			/* Where to store the size of the file. */
-{
-    Efile_error* errInfo = state->errInfo;
-    BY_HANDLE_FILE_INFORMATION fileInfo; /* File information from a handle. */
-    HANDLE fd;			/* Handle to open file. */
-    DWORD access;		/* Access mode: GENERIC_READ, GENERIC_WRITE. */
-    DWORD crFlags;
-    DWORD flagsAndAttrs = FILE_ATTRIBUTE_NORMAL;
-    WCHAR *wname = (WCHAR *) name;
-
-    switch (flags & (EFILE_MODE_READ|EFILE_MODE_WRITE)) {
-    case EFILE_MODE_READ:
-	access = GENERIC_READ;
-	crFlags = OPEN_EXISTING;
-	break;
-    case EFILE_MODE_WRITE:
-	access = GENERIC_WRITE;
-	crFlags = CREATE_ALWAYS;
-	break;
-    case EFILE_MODE_READ_WRITE:
-	access = GENERIC_READ|GENERIC_WRITE;
-	crFlags = OPEN_ALWAYS;
-	break;
-    default:
-	errno = EINVAL;
-	check_error(-1, errInfo);
-	return 0;
-    }
-
-    if (flags & EFILE_MODE_SYNC) {
-        flagsAndAttrs = FILE_FLAG_WRITE_THROUGH;
-    }
-
-    if (flags & EFILE_MODE_APPEND) {
-	crFlags = OPEN_ALWAYS;
-    }
-    if (flags & EFILE_MODE_EXCL) {
-	crFlags = CREATE_NEW;
-    }
-    ensure_wpath(state, &wname);
-    fd = CreateFileW(wname, access,
-		    FILE_SHARE_FLAGS,
-		    NULL, crFlags, flagsAndAttrs, NULL);
-
-    /*
-     * Check for errors.
-     */
-
-    if (fd == INVALID_HANDLE_VALUE) {
-	DWORD attr;
-
-	set_error(errInfo);
-
-	/*
-	 * If the error is EACESS, the reason could be that we tried to
-	 * open a directory.  In that case, we'll change the error code
-	 * to EISDIR.
-	 */
-	if (errInfo->posix_errno && 
-	    (attr = GetFileAttributesW(wname)) != INVALID_FILE_ATTRIBUTES && 
-	    (attr & FILE_ATTRIBUTE_DIRECTORY)) {
-	    errInfo->posix_errno = EISDIR;
-	}
-	return 0;
-    }
-
-    /*
-     * Get and return the length of the open file.
-     */
-       
-    if (!GetFileInformationByHandle(fd, &fileInfo))
-	return set_error(errInfo);
-    *pfd = (int) fd;
-    if (pSize) {
-	*pSize = (Sint64)
-	    (((Uint64)fileInfo.nFileSizeHigh << 32) | 
-	     (Uint64)fileInfo.nFileSizeLow);
-    }
-    return 1;
-}
-
-int 
-efile_may_openfile(Efile_error* errInfo, char *name)
-{
-    Efile_call_state state;
-    WCHAR *wname = (WCHAR *) name;
-    DWORD attr;
-    int ret;
-
-    DBG_TRACE(1, name);
-    call_state_init(&state, errInfo);
-    ensure_wpath(&state, &wname);
-    if ((attr = GetFileAttributesW(wname)) == INVALID_FILE_ATTRIBUTES) {
-	errno = ENOENT;
-	ret = check_error(-1, errInfo);
-    }
-    else if (attr & FILE_ATTRIBUTE_DIRECTORY) {
-	errno = EISDIR;
-	ret = check_error(-1, errInfo);
-    }
-    else ret = 1;
-
-    call_state_free(&state);
-    return ret;
-}
-
-void
-efile_closefile(fd)
-int fd;				/* File descriptor for file to close. */
-{
-    DBG_TRACE(2, L"");
-    CloseHandle((HANDLE) fd);
-}
-
-FILE* efile_wfopen(const WCHAR* name, const WCHAR* mode)
-{
-    Efile_call_state state;
-    Efile_error dummy;
-    FILE* f;
-    call_state_init(&state, &dummy);
-    ensure_wpath(&state, (WCHAR**)&name);
-    f = _wfopen(name, mode);
-    call_state_free(&state);
-    return f;
-}
-
-int
-efile_fdatasync(errInfo, fd)
-Efile_error* errInfo;		/* Where to return error codes. */
-int fd;				/* File descriptor for file to sync. */
-{
-    DBG_TRACE(2, L"");
-    /* Not available in Windows, just call regular fsync */
-    return efile_fsync(errInfo, fd);
-}
-
-int
-efile_fsync(errInfo, fd)
-Efile_error* errInfo;		/* Where to return error codes. */
-int fd;				/* File descriptor for file to sync. */
-{
-    DBG_TRACE(2, L"");
-    if (!FlushFileBuffers((HANDLE) fd)) {
-	return check_error(-1, errInfo);
-    }
-    return 1;
-}
-
-int
-efile_fileinfo(Efile_error* errInfo, Efile_info* pInfo,
-	       char* orig_name, int info_for_link)
-{
-    Efile_call_state state;
-    int ret;
-    DBG_TRACE(1, L"");
-    call_state_init(&state, errInfo);
-    ret = do_fileinfo(&state, pInfo, orig_name, info_for_link);
-    call_state_free(&state);
-    return ret;
-}
-
-static int
-do_fileinfo(Efile_call_state* state, Efile_info* pInfo,
-	    char* orig_name, int info_for_link)
-{
-    Efile_error* errInfo = state->errInfo;
-    HANDLE findhandle;		/* Handle returned by FindFirstFile(). */
-    WIN32_FIND_DATAW findbuf;	/* Data return by FindFirstFile(). */
-    WCHAR* name = NULL;
-    WCHAR* win_path;
-    int name_len;
-    int drive;			/* Drive for filename (1 = A:, 2 = B: etc). */
-    WCHAR *worig_name = (WCHAR *) orig_name;
-
-    ensure_wpath(state, &worig_name);
-    /* Don't allow wildcards to be interpreted by system */
-
-
-    /*
-     * Move the name to a buffer and make sure to remove a trailing
-     * slash, because it causes FindFirstFile() to fail on Win95.
-     */
-
-    name_len = wcslen(worig_name);
-
-    name = wpath_tmp_alloc(state, name_len+1);
-    wcscpy(name, worig_name);
-    if (name_len > 2 && ISSLASH(name[name_len-1]) &&
-	name[name_len-2] != L':') {
-	name[name_len-1] = L'\0';
-    }
-
-    win_path = name;
-    if (wcsncmp(name, L"\\\\?\\", 4) == 0) {
-	win_path += 4;
-    }
-
-    if (wcspbrk(win_path, L"?*")) {
-    enoent:
-	errInfo->posix_errno = ENOENT;
-	errInfo->os_errno = ERROR_FILE_NOT_FOUND;
-        return 0;
-    }
-
-    /* Try to get disk from name.  If none, get current disk.  */
-
-    if (win_path[1] != L':') {
-	WCHAR* cwd_path = get_cwd_wpath_tmp(state);
-        drive = 0;
-	if (cwd_path[1] == L':') {
-	    drive = towlower(cwd_path[0]) - L'a' + 1;
-	}
-    } else if (*win_path && win_path[2] == L'\0') {
-	/*
-	 * X: and nothing more is an error.
-	 */
-	errInfo->posix_errno = ENOENT;
-	errInfo->os_errno = ERROR_FILE_NOT_FOUND;
-	return 0;
-    } else {
-        drive = towlower(*win_path) - L'a' + 1;
-    }
-
-    findhandle = FindFirstFileW(name, &findbuf);
-    if (findhandle == INVALID_HANDLE_VALUE) {
-	WCHAR* path = NULL;
-
-        if (!(wcspbrk(name, L"./\\") &&
-	      (path = get_full_wpath_tmp(state, name, NULL, 0)) &&
-	      /* root dir. ('C:\') or UNC root dir. ('\\server\share\') */
-	      ((wcslen(path) == 3) || is_root_unc_name(path)) &&
-	      (GetDriveTypeW(path) > 1)   ) ) {
-
-	    errInfo->posix_errno = ENOENT;
-	    errInfo->os_errno = ERROR_FILE_NOT_FOUND;
-	    return 0;
-	}
-
-        /*
-         * Root directories (such as C:\ or \\server\share\ are fabricated.
-         */
-	
-        findbuf.dwFileAttributes = FILE_ATTRIBUTE_DIRECTORY;
-        findbuf.nFileSizeHigh = 0;
-        findbuf.nFileSizeLow = 0;
-        findbuf.cFileName[0] = L'\0';
-
-	pInfo->links = 1;
-	pInfo->cTime = pInfo->accessTime = pInfo->modifyTime = 0;
-    } else {
-	SYSTEMTIME SystemTime;
-        FILETIME LocalFTime;
-
-	/*first check if we are a symlink */
-	if (!info_for_link && (findbuf.dwFileAttributes &
-			       FILE_ATTRIBUTE_REPARSE_POINT)){
-	    /*
-	     * given that we know this is a symlink,
-	     we should be able to find its target */
-	    WCHAR* target_name = (WCHAR*) do_readlink(state, (char *) name, NULL, 0);
-	    if (target_name) {
-		FindClose(findhandle);
-		return do_fileinfo(state, pInfo,
-				   (char *) target_name, info_for_link);
-	    }
-	}
-
-	/* number of links: */
-	{
-	    HANDLE handle;	/* Handle returned by CreateFile() */
-	    BY_HANDLE_FILE_INFORMATION fileInfo; /* from  CreateFile() */
-
-            /* We initialise nNumberOfLinks as GetFileInformationByHandle
-               does not always initialise this field */
-            fileInfo.nNumberOfLinks = 1;
-	    if (handle = CreateFileW(name, GENERIC_READ, FILE_SHARE_FLAGS, NULL,
-				    OPEN_EXISTING, 0, NULL)) {
-		GetFileInformationByHandle(handle, &fileInfo);
-		pInfo->links = fileInfo.nNumberOfLinks;
-		CloseHandle(handle);
-	    } else {
-		pInfo->links = 1;
-	    }	
-	}
-
-        FILETIME_TO_EPOCH(pInfo->modifyTime, findbuf.ftLastWriteTime);
-
-        if (findbuf.ftLastAccessTime.dwLowDateTime == 0 &&
-	    findbuf.ftLastAccessTime.dwHighDateTime == 0) {
-	    pInfo->accessTime = pInfo->modifyTime;
-	} else {
-	    FILETIME_TO_EPOCH(pInfo->accessTime, findbuf.ftLastAccessTime);
-	}
-
-        if (findbuf.ftCreationTime.dwLowDateTime == 0 &&
-	    findbuf.ftCreationTime.dwHighDateTime == 0) {
-	    pInfo->cTime = pInfo->modifyTime;
-	} else {
-	    FILETIME_TO_EPOCH(pInfo->cTime ,findbuf.ftCreationTime);
-	}
-        FindClose(findhandle);
-    }
-
-    pInfo->size_low = findbuf.nFileSizeLow;
-    pInfo->size_high = findbuf.nFileSizeHigh;
-	
-    if (info_for_link && (findbuf.dwFileAttributes &
-			  FILE_ATTRIBUTE_REPARSE_POINT))
-	pInfo->type = FT_SYMLINK;
-    else if (findbuf.dwFileAttributes & FILE_ATTRIBUTE_DIRECTORY)
-	pInfo->type = FT_DIRECTORY;
-    else
-	pInfo->type = FT_REGULAR;
-
-    if (findbuf.dwFileAttributes & FILE_ATTRIBUTE_READONLY)
-	pInfo->access = FA_READ;
-    else
-	pInfo->access = FA_READ|FA_WRITE;
-
-    pInfo->mode = dos_to_posix_mode(findbuf.dwFileAttributes, name);
-    pInfo->major_device = drive;
-    pInfo->minor_device = 0;
-    pInfo->inode = 0;
-    pInfo->uid = 0;
-    pInfo->gid = 0;
-    
-    return 1;
-}
-
-int
-efile_write_info(Efile_error* errInfo,
-		 Efile_info* pInfo,
-		 char* name)
-{
-    Efile_call_state state;
-    int ret;
-    call_state_init(&state, errInfo);
-    ret = do_write_info(&state, pInfo, name);
-    call_state_free(&state);
-    return ret;
-}
-
-static int
-do_write_info(Efile_call_state* state,
-              Efile_info* pInfo,
-	      char* name)
-{
-    Efile_error* errInfo = state->errInfo;
-    SYSTEMTIME timebuf;
-    FILETIME ModifyFileTime;
-    FILETIME AccessFileTime;
-    FILETIME CreationFileTime;
-    HANDLE fd;
-    DWORD attr;
-    DWORD tempAttr;
-    WCHAR *wname = (WCHAR *) name;
-
-    DBG_TRACE(1, name);
-
-    ensure_wpath(state, &wname);
-
-    /*
-     * Get the attributes for the file.
-     */
-
-    tempAttr = attr = GetFileAttributesW(wname);
-    if (attr == 0xffffffff) {
-	return set_error(errInfo);
-    }
-    if (pInfo->mode != -1) {
-	if (pInfo->mode & _S_IWRITE) {
-	    /* clear read only bit */
-	    attr &= ~FILE_ATTRIBUTE_READONLY;
-	} else {
-	    /* set read only bit */
-	    attr |= FILE_ATTRIBUTE_READONLY;
-	}
-    }
-
-    /*
-     * Construct all file times.
-     */
-
-    EPOCH_TO_FILETIME(ModifyFileTime,   pInfo->modifyTime);
-    EPOCH_TO_FILETIME(AccessFileTime,   pInfo->accessTime);
-    EPOCH_TO_FILETIME(CreationFileTime, pInfo->cTime);
-
-    /*
-     * If necessary, set the file times.
-     */
-
-    /*
-     * If the has read only access, we must temporarily turn on
-     * write access (this is necessary for native filesystems,
-     * but not for NFS filesystems).
-     */
-
-    if (tempAttr & FILE_ATTRIBUTE_READONLY) {
-	tempAttr &= ~FILE_ATTRIBUTE_READONLY;
-	if (!SetFileAttributesW(wname, tempAttr)) {
-	    return set_error(errInfo);
-	}
-    }
-
-    fd = CreateFileW(wname, GENERIC_READ|GENERIC_WRITE,
-	    FILE_SHARE_FLAGS,
-	    NULL, OPEN_EXISTING, FILE_ATTRIBUTE_NORMAL, NULL);
-    if (fd != INVALID_HANDLE_VALUE) {
-	BOOL result = SetFileTime(fd, &CreationFileTime, &AccessFileTime, &ModifyFileTime);
-	if (!result) {
-	    return set_error(errInfo);
-	}
-	CloseHandle(fd);
-    }
-
-    /*
-     * If the file doesn't have the correct attributes, set them now.
-     * (It could have been done before setting the file times, above).
-     */
-
-    if (tempAttr != attr) {
-	if (!SetFileAttributesW(wname, attr)) {
-	    return set_error(errInfo);
-	}
-    }
-    return 1;
-}
-
-
-int
-efile_pwrite(errInfo, fd, buf, count, offset)
-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;
-    DBG_TRACE(2, L"");
-    res = efile_seek(errInfo, fd, offset, EFILE_SEEK_SET, NULL);
-    if (res) {
-	return efile_write(errInfo, EFILE_MODE_WRITE, fd, buf, count);
-    } else {
-	return res;
-    }
-}
-
-/* position and read/write as a single atomic op */
-int
-efile_pread(errInfo, fd, offset, buf, count, pBytesRead)
-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;
-    DBG_TRACE(2, L"");
-    res = efile_seek(errInfo, fd, offset, EFILE_SEEK_SET, NULL);
-    if (res) {
-	return efile_read(errInfo, EFILE_MODE_READ, fd, buf, count, pBytesRead);
-    } else {
-	return res;
-    }
-}
-
-
-
-int
-efile_write(errInfo, flags, fd, buf, count)
-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. */
-{
-    DWORD written;		/* Bytes written in last operation. */
-    OVERLAPPED overlapped;
-    OVERLAPPED* pOverlapped = NULL;
-
-    DBG_TRACE(2, L"");
-    if (flags & EFILE_MODE_APPEND) {
-	memset(&overlapped, 0, sizeof(overlapped));
-	overlapped.Offset = 0xffffffff;
-	overlapped.OffsetHigh = 0xffffffff;
-	pOverlapped = &overlapped;
-    }
-    while (count > 0) {
-	if (!WriteFile((HANDLE) fd, buf, count, &written, pOverlapped))
-	    return set_error(errInfo);
-	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 are unchanged 
-				      * after the call */
-	     int iovcnt)             /* Number of structs in vector */
-{
-    int cnt;                         /* Buffers so far written */
-    OVERLAPPED overlapped;
-    OVERLAPPED* pOverlapped = NULL;
-
-    DBG_TRACE(2, L"");
-    ASSERT(iovcnt >= 0);
-    
-    if (flags & EFILE_MODE_APPEND) {
-	memset(&overlapped, 0, sizeof(overlapped));
-	overlapped.Offset = 0xffffffff;
-	overlapped.OffsetHigh = 0xffffffff;
-	pOverlapped = &overlapped;
-    }
-    for (cnt = 0; cnt < iovcnt; cnt++) {
-	if (iov[cnt].iov_base && iov[cnt].iov_len > 0) {
-	    /* Non-empty buffer */
-	    int p;                   /* Position in buffer */
-	    int w = iov[cnt].iov_len;/* Bytes written in this call */
-	    for (p = 0; p < iov[cnt].iov_len; p += w) {
-		if (!WriteFile((HANDLE) fd, 
-			       iov[cnt].iov_base + p, 
-			       iov[cnt].iov_len - p, 
-			       &w, 
-			       pOverlapped))
-		    return set_error(errInfo);
-	    }
-	}
-    }	
-    return 1;
-}
-
-int
-efile_read(errInfo, flags, fd, buf, count, pBytesRead)
-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. */
-{
-    DWORD nbytes = 0;
-
-    DBG_TRACE(2, L"");
-    if (!ReadFile((HANDLE) fd, buf, count, &nbytes, NULL))
-	return set_error(errInfo);
-
-    *pBytesRead = nbytes;
-    return 1;
-}
-
-int
-efile_seek(errInfo, fd, offset, origin, new_location)
-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. */
-{
-    LARGE_INTEGER off, new_loc;
-    
-    DBG_TRACE(2, L"");
-    switch (origin) {
-    case EFILE_SEEK_SET: origin = FILE_BEGIN; break;
-    case EFILE_SEEK_CUR: origin = FILE_CURRENT; break;
-    case EFILE_SEEK_END: origin = FILE_END; break;
-    default:
-	errno = EINVAL;
-	check_error(-1, errInfo);
-	break;
-    }
-    
-    off.QuadPart = offset;
-    if (! SetFilePointerEx((HANDLE) fd, off,
-	new_location ? &new_loc : NULL, origin)) {
-	return set_error(errInfo);
-    }
-    if (new_location) {
-	*new_location = new_loc.QuadPart;
-	DEBUGF(("efile_seek(offset=%ld, origin=%d) -> %ld\n", 
-		(long) offset, origin, (long) *new_location));
-    } else {
-	DEBUGF(("efile_seek(offset=%ld, origin=%d)\n", (long) offset, origin));
-    }
-    return 1;
-}
-
-int
-efile_truncate_file(errInfo, fd, flags)
-Efile_error* errInfo;		/* Where to return error codes. */
-int *fd;				/* File descriptor for file to truncate. */
-int flags;
-{
-    DBG_TRACE(2, L"");
-    if (!SetEndOfFile((HANDLE) (*fd)))
-	return set_error(errInfo);
-    return 1;
-}
-
-
-/*
- * is_root_unc_name - returns TRUE if the argument is a UNC name specifying
- *      a root share.  That is, if it is of the form \\server\share\.
- *      This routine will also return true if the argument is of the
- *      form \\server\share (no trailing slash) but Win32 currently
- *      does not like that form.
- *
- *      Forward slashes ('/') may be used instead of backslashes ('\').
- */
-
-static int
-is_root_unc_name(const WCHAR *path)
-{
-    /*
-     * If a root UNC name, path will start with 2 (but not 3) slashes
-     */
-
-    if ((wcslen(path) >= 5) /* minimum string is "//x/y" */
-	&& ISSLASH(path[0]) && ISSLASH(path[1]))
-    {
-        const WCHAR *p = path + 2;
-
-        /*
-         * find the slash between the server name and share name
-         */
-        while ( * ++ p )
-            if ( ISSLASH(*p) )
-                break ;
-
-        if ( *p && p[1] )
-        {
-            /*
-             * is there a further slash?
-             */
-            while ( * ++ p )
-                if ( ISSLASH(*p) )
-                    break ;
-
-            /*
-             * just final slash (or no final slash)
-             */
-            if ( !*p || !p[1])
-                return 1;
-        }
-    }
-
-    return 0 ;
-}
-
-/*
- * Extracts the root part of an absolute filename (by modifying the string
- * pointed to by the name argument).  The name can start
- * with either a driver letter (for example, C:\), or a UNC name
- * (for example, \\guinness\bjorn).
- *
- * If the name is invalid, the buffer will be modified to point to
- * an empty string.
- *
- * Returns: 1 if the name consists of just the root part, 0 if
- * 	    the name was longer.
- */
-
-static int
-extract_root(WCHAR* name)
-{
-    int len = wcslen(name);
-
-    if (iswalpha(name[0]) && name[1] == L':' && ISSLASH(name[2])) {
-	WCHAR c = name[3];
-	name[3] = L'\0';
-	return c == L'\0';
-    } else if (len < 5 || !ISSLASH(name[0]) || !ISSLASH(name[1])) {
-	goto error;
-    } else {			/* Try to find the end of the UNC name. */
-	WCHAR* p;
-	WCHAR c;
-
-        /*
-         * Find the slash between the server name and share name.
-         */
-
-	for (p = name + 2; *p; p++)
-            if (ISSLASH(*p))
-                break;
-	if (*p == L'\0')
-	    goto error;
-
-	/*
-	 * Find the slash after the share name.
-	 */
-
-	for (p++; *p; p++)
-            if (ISSLASH(*p))
-                break;
-	c = *p;
-	*p = L'\0';
-	return c == L'\0' || p[1] == L'\0';
-    }
-
- error:
-    *name = L'\0';
-    return 1;
-}
-
-static unsigned short
-dos_to_posix_mode(int attr, const WCHAR *name)
-{
-    register unsigned short uxmode;
-    unsigned dosmode;
-    register const WCHAR *p;
-
-    dosmode = attr & 0xff;
-    if ((p = name)[1] == L':')
-	p += 2;
-
-    /* check to see if this is a directory - note we must make a special
-     * check for the root, which DOS thinks is not a directory
-     */
-
-    uxmode = (unsigned short)
-	(((ISSLASH(*p) && !p[1]) || (dosmode & FILE_ATTRIBUTE_DIRECTORY) ||
-	       *p == L'\0') ? _S_IFDIR|_S_IEXEC : _S_IFREG);
-
-    /* If attribute byte does not have read-only bit, it is read-write */
-
-    uxmode |= (dosmode & FILE_ATTRIBUTE_READONLY) ?
-	_S_IREAD : (_S_IREAD|_S_IWRITE);
-
-    /* see if file appears to be executable - check extension of name */
-
-    if (p = wcsrchr(name, L'.')) {
-        if (!_wcsicmp(p, L".exe") ||
-	    !_wcsicmp(p, L".cmd") ||
-	    !_wcsicmp(p, L".bat") ||
-	    !_wcsicmp(p, L".com"))
-            uxmode |= _S_IEXEC;
-    }
-
-    /* propagate user read/write/execute bits to group/other fields */
-
-    uxmode |= (uxmode & 0700) >> 3;
-    uxmode |= (uxmode & 0700) >> 6;
-
-    return uxmode;
-}
-
-
-int
-efile_readlink(Efile_error* errInfo, char* name, char* buffer, size_t size)
-{
-    Efile_call_state state;
-    int ret;
-    DBG_TRACE(1, name);
-    call_state_init(&state, errInfo);
-    ret = !!do_readlink(&state, name, buffer, size);
-    call_state_free(&state);
-    return ret;
-}
-
-/* If buffer==0, return buffer allocated by wpath_tmp_allocate
-*/
-static char*
-do_readlink(Efile_call_state* state, char* name, char* buffer, size_t size)
-{
-    /*
-     * load dll and see if we have CreateSymbolicLink at runtime:
-     * (Vista only)
-     */
-    HINSTANCE hModule = NULL;
-    WCHAR *wname = (WCHAR *) name;
-    WCHAR *wbuffer = (WCHAR *) buffer;
-    DWORD wsize = size / sizeof(WCHAR);
-    char* ret = NULL;
-
-    if ((hModule = LoadLibrary("kernel32.dll")) != NULL) {
-	typedef DWORD (WINAPI * GETFINALPATHNAMEBYHANDLEPTR)(
-							     HANDLE hFile,
-							     LPCWSTR lpFilePath,
-							     DWORD cchFilePath,
-							     DWORD dwFlags);
-
-	GETFINALPATHNAMEBYHANDLEPTR pGetFinalPathNameByHandle =
-	    (GETFINALPATHNAMEBYHANDLEPTR)GetProcAddress(hModule, "GetFinalPathNameByHandleW");
-
-	if (pGetFinalPathNameByHandle != NULL) {
-	    DWORD fileAttributes;
-	    ensure_wpath(state, &wname);
-	    /* first check if file is a symlink; {error, einval} otherwise */
-	    fileAttributes = GetFileAttributesW(wname);
-	    if ((fileAttributes & FILE_ATTRIBUTE_REPARSE_POINT)) {
-		DWORD success = 0;
-		HANDLE h = CreateFileW(wname, GENERIC_READ, FILE_SHARE_FLAGS, NULL, OPEN_EXISTING, FILE_FLAG_BACKUP_SEMANTICS, NULL);
-		int len;
-		if(h != INVALID_HANDLE_VALUE) {
-		    if (!wbuffer) { /* dynamic allocation */
-			WCHAR dummy;
-			wsize = pGetFinalPathNameByHandle(h, &dummy, 0, 0);
-			if (wsize) {
-			    wbuffer = wpath_tmp_alloc(state, wsize);
-			}
-		    }
-		    if (wbuffer
-			&& (success = pGetFinalPathNameByHandle(h, wbuffer, wsize, 0))
-			&& success < wsize) {
-			WCHAR* wp;
-
-			/* GetFinalPathNameByHandle prepends path with "\\?\": */
-			len = wcslen(wbuffer);
-			wmemmove(wbuffer,wbuffer+4,len-3);
-			if (len - 4 >= 2 && wbuffer[1] == L':' && wbuffer[0] >= L'A' &&
-			    wbuffer[0] <= L'Z') {
-			    wbuffer[0] = wbuffer[0] + L'a' - L'A';
-			}
-			
-			for (wp=wbuffer ; *wp; wp++)
-			    if (*wp == L'\\')
-				*wp = L'/';
-		    }
-		    CloseHandle(h);
-		}
-		if (success) {
-		    ret = (char*) wbuffer;
-		} else {
-		    set_error(state->errInfo);
-		}
-	    } else {
-		errno = EINVAL;
-		save_last_error(state->errInfo);
-	    }
-	    goto done;
-	}
-    }
-    errno = ENOTSUP;
-    save_last_error(state->errInfo);
-
-done:
-    if (hModule)
-	FreeLibrary(hModule);
-    return ret;
-}
-
-
-int
-efile_altname(Efile_error* errInfo, char* orig_name, char* buffer, size_t size)
-{
-    Efile_call_state state;
-    int ret;
-    DBG_TRACE(1, orig_name);
-    call_state_init(&state, errInfo);
-    ret = do_altname(&state, orig_name, buffer, size);
-    call_state_free(&state);
-    return ret;
-}
-
-static int
-do_altname(Efile_call_state* state, char* orig_name, char* buffer, size_t size)
-{
-    WIN32_FIND_DATAW wfd;
-    HANDLE fh;
-    WCHAR* name;
-    int name_len;
-    WCHAR* full_path = NULL;
-    WCHAR *worig_name = (WCHAR *) orig_name;
-    WCHAR *wbuffer = (WCHAR *) buffer;
-    int drive;			/* Drive for filename (1 = A:, 2 = B: etc). */
-
-    /* Don't allow wildcards to be interpreted by system */
-
-    if (wcspbrk(worig_name, L"?*")) {
-    enoent:
-	state->errInfo->posix_errno = ENOENT;
-	state->errInfo->os_errno = ERROR_FILE_NOT_FOUND;
-        return 0;
-    }
-
-    /*
-     * Move the name to a buffer and make sure to remove a trailing
-     * slash, because it causes FindFirstFile() to fail on Win95.
-     */
-    ensure_wpath(state, &worig_name);
-    name_len = wcslen(worig_name);
-
-    name = wpath_tmp_alloc(state, name_len + 1);
-    wcscpy(name, worig_name);
-    if (name_len > 2 && ISSLASH(name[name_len-1]) &&
-	name[name_len-2] != L':') {
-	name[name_len-1] = L'\0';
-    }
-    
-    /* Try to get disk from name.  If none, get current disk.  */
-
-    if (name[1] != L':') {
-	WCHAR* cwd_path = get_cwd_wpath_tmp(state);
-        drive = 0;
-        if (cwd_path[1] == L':') {
-	    drive = towlower(cwd_path[0]) - L'a' + 1;
-	}
-    } else if (*name && name[2] == L'\0') {
-	/*
-	 * X: and nothing more is an error.
-	 */
-	goto enoent;
-    } else {
-        drive = towlower(*name) - L'a' + 1;
-    }
-    fh = FindFirstFileW(name,&wfd);
-    if (fh == INVALID_HANDLE_VALUE) {
-	DWORD fff_error = GetLastError();
-        if (!(wcspbrk(name, L"./\\") &&
-	      (full_path = get_full_wpath_tmp(state, name, NULL, 0)) &&
-	      /* root dir. ('C:\') or UNC root dir. ('\\server\share\') */
-	      ((wcslen(full_path) == 3) || is_root_unc_name(full_path)) &&
-	      (GetDriveTypeW(full_path) > 1)   ) ) {
-
-	    set_os_errno(state->errInfo, fff_error);
-	    return 0;
-	}
-        /*
-         * Root directories (such as C:\ or \\server\share\ are fabricated.
-         */
-	wcscpy(wbuffer,name);
-	return 1;
-    }
-	
-    wcscpy(wbuffer,wfd.cAlternateFileName);
-    if (!*wbuffer) {
-	wcscpy(wbuffer,wfd.cFileName);
-    }
-    FindClose(fh);
-    return 1;
-}
-
-
-int
-efile_link(Efile_error* errInfo, char* old, char* new)
-{
-    Efile_call_state state;
-    WCHAR *wold = (WCHAR *) old;
-    WCHAR *wnew = (WCHAR *) new;
-    int ret;
-    DBG_TRACE(1, old);
-    call_state_init(&state, errInfo);
-    ensure_wpath(&state, &wold);
-    ensure_wpath(&state, &wnew);
-    if(!CreateHardLinkW(wnew, wold, NULL)) {
-	ret = set_error(errInfo);
-    }
-    else ret =1;
-    call_state_free(&state);
-    return ret;
-}
-
-int
-efile_symlink(Efile_error* errInfo, char* old, char* new)
-{
-    Efile_call_state state;
-    int ret;
-    DBG_TRACE2(1, "symlink(%s <- %s)", old, new);
-    call_state_init(&state, errInfo);
-    ret = do_symlink(&state, old, new);
-    call_state_free(&state);
-    return ret;
-}
-
-static int
-do_symlink(Efile_call_state* state, char* old, char* new)
-{
-    /*
-     * Load dll and see if we have CreateSymbolicLink at runtime:
-     * (Vista only)
-     */
-    HINSTANCE hModule = NULL;
-    WCHAR *wold = (WCHAR *) old;
-    WCHAR *wnew = (WCHAR *) new;
-
-    DBG_TRACE(1, old);
-    if ((hModule = LoadLibrary("kernel32.dll")) != NULL) {
-	typedef BOOLEAN (WINAPI  * CREATESYMBOLICLINKFUNCPTR) (
-	     LPCWSTR lpSymlinkFileName,
-	     LPCWSTR lpTargetFileName,
-	     DWORD dwFlags);
-
-	CREATESYMBOLICLINKFUNCPTR pCreateSymbolicLink =
-	    (CREATESYMBOLICLINKFUNCPTR) GetProcAddress(hModule,
-						       "CreateSymbolicLinkW");
-	/* A for MBCS, W for UNICODE... char* above implies 'W'! */
-	if (pCreateSymbolicLink != NULL) {
-	  ensure_wpath(state, &wold);
-	  ensure_wpath(state, &wnew);
-	  {
-	    DWORD attr = GetFileAttributesW(wold);
-	    int flag = (attr != INVALID_FILE_ATTRIBUTES &&
-			attr & FILE_ATTRIBUTE_DIRECTORY) ? 1 : 0;
-	    /*  SYMBOLIC_LINK_FLAG_DIRECTORY = 1 */
-	    BOOLEAN success = pCreateSymbolicLink(wnew, wold, flag);
-	    FreeLibrary(hModule);
-
-	    if (success) {
-		return 1;
-	    } else {
-		return set_error(state->errInfo);
-	    }
-	  }
-	} else
-	    FreeLibrary(hModule);
-    }
-    errno = ENOTSUP;
-    return check_error(-1, state->errInfo);
-}
-
-int
-efile_fadvise(Efile_error* errInfo, int fd, Sint64 offset,
-	    Sint64 length, int advise)
-{
-    DBG_TRACE(2, L"");
-    /* posix_fadvise is not available on Windows, do nothing */
-    errno = ERROR_SUCCESS;
-    return check_error(0, errInfo);
-}
-
-int
-efile_fallocate(Efile_error* errInfo, int fd, Sint64 offset, Sint64 length)
-{
-    DBG_TRACE(2, L"");
-    /* No file preallocation method available in Windows. */
-    errno = errno_map(ERROR_NOT_SUPPORTED);
-    SetLastError(ERROR_NOT_SUPPORTED);
-
-    return check_error(-1, errInfo);
-}
diff --git a/erts/emulator/nifs/common/prim_buffer_nif.c b/erts/emulator/nifs/common/prim_buffer_nif.c
new file mode 100644
index 0000000000..a8ef5fc355
--- /dev/null
+++ b/erts/emulator/nifs/common/prim_buffer_nif.c
@@ -0,0 +1,512 @@
+/*
+ * %CopyrightBegin%
+ *
+ * Copyright Ericsson 2017. 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%
+ */
+
+#define STATIC_ERLANG_NIF 1
+
+#include "erl_nif.h"
+#include "config.h"
+#include "sys.h"
+
+#ifdef VALGRIND
+#  include <valgrind/memcheck.h>
+#endif
+
+#define ACCUMULATOR_SIZE (2 << 10)
+
+#define FIND_NIF_RESCHEDULE_SIZE (1 << 20)
+
+/* NIF interface declarations */
+static int load(ErlNifEnv *env, void** priv_data, ERL_NIF_TERM load_info);
+static int upgrade(ErlNifEnv *env, void** priv_data, void** old_priv_data, ERL_NIF_TERM load_info);
+static void unload(ErlNifEnv *env, void* priv_data);
+
+static ErlNifResourceType *rtype_buffer;
+
+static ERL_NIF_TERM am_ok;
+static ERL_NIF_TERM am_error;
+
+static ERL_NIF_TERM am_lock_order_violation;
+
+static ERL_NIF_TERM am_acquired;
+static ERL_NIF_TERM am_busy;
+
+static ERL_NIF_TERM am_continue;
+
+static ERL_NIF_TERM am_out_of_memory;
+static ERL_NIF_TERM am_not_found;
+
+typedef struct {
+#ifdef DEBUG
+    erts_atomic32_t concurrent_users;
+#endif
+
+    ErlNifBinary accumulator;
+    size_t accumulated_bytes;
+    int accumulator_present;
+
+    ErlNifIOQueue *queue;
+
+    erts_atomic32_t external_lock;
+} buffer_data_t;
+
+static ERL_NIF_TERM new_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]);
+
+static ERL_NIF_TERM peek_head_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]);
+static ERL_NIF_TERM skip_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]);
+static ERL_NIF_TERM size_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]);
+
+static ERL_NIF_TERM write_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]);
+static ERL_NIF_TERM copying_read_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]);
+
+static ERL_NIF_TERM find_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]);
+
+static ERL_NIF_TERM trylock_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]);
+static ERL_NIF_TERM unlock_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]);
+
+static ErlNifFunc nif_funcs[] = {
+    {"new", 0, new_nif},
+    {"size", 1, size_nif},
+    {"peek_head", 1, peek_head_nif},
+    {"copying_read", 2, copying_read_nif},
+    {"write", 2, write_nif},
+    {"skip", 2, skip_nif},
+    {"find_byte_index", 2, find_nif},
+    {"try_lock", 1, trylock_nif},
+    {"unlock", 1, unlock_nif},
+};
+
+ERL_NIF_INIT(prim_buffer, nif_funcs, load, NULL, upgrade, unload)
+
+static void gc_buffer(ErlNifEnv *env, void* data);
+
+static int load(ErlNifEnv *env, void** priv_data, ERL_NIF_TERM load_info)
+{
+    am_ok = enif_make_atom(env, "ok");
+    am_error = enif_make_atom(env, "error");
+
+    am_lock_order_violation = enif_make_atom(env, "lock_order_violation");
+    am_acquired = enif_make_atom(env, "acquired");
+    am_busy = enif_make_atom(env, "busy");
+
+    am_continue = enif_make_atom(env, "continue");
+
+    am_out_of_memory = enif_make_atom(env, "out_of_memory");
+    am_not_found = enif_make_atom(env, "not_found");
+
+    rtype_buffer = enif_open_resource_type(env, NULL, "gc_buffer", gc_buffer,
+        ERL_NIF_RT_CREATE, NULL);
+
+    *priv_data = NULL;
+
+    return 0;
+}
+
+static void unload(ErlNifEnv *env, void* priv_data)
+{
+
+}
+
+static int upgrade(ErlNifEnv *env, void** priv_data, void** old_priv_data, ERL_NIF_TERM load_info)
+{
+    if(*old_priv_data != NULL) {
+        return -1; /* Don't know how to do that */
+    }
+
+    if(*priv_data != NULL) {
+        return -1; /* Don't know how to do that */
+    }
+
+    if(load(env, priv_data, load_info)) {
+        return -1;
+    }
+
+    return 0;
+}
+
+static void gc_buffer(ErlNifEnv *env, void* data) {
+    buffer_data_t *buffer = (buffer_data_t*)data;
+
+    if(buffer->accumulator_present) {
+        enif_release_binary(&buffer->accumulator);
+    }
+
+    enif_ioq_destroy(buffer->queue);
+}
+
+static int get_buffer_data(ErlNifEnv *env, ERL_NIF_TERM opaque, buffer_data_t **buffer) {
+    return enif_get_resource(env, opaque, rtype_buffer, (void **)buffer);
+}
+
+/* Copies a number of bytes from the head of the iovec, skipping "vec_skip"
+ * vector elements followed by "byte_skip" bytes on the target vector. */
+static void copy_from_iovec(SysIOVec *iovec, int vec_len, int vec_skip,
+        size_t byte_skip, size_t size, char *data) {
+
+    size_t bytes_copied, skip_offset;
+    int vec_index;
+
+    skip_offset = byte_skip;
+    vec_index = vec_skip;
+    bytes_copied = 0;
+
+    while(bytes_copied < size) {
+        size_t block_size, copy_size;
+        char *block_start;
+
+        ASSERT(vec_index < vec_len);
+
+        block_start = (char*)iovec[vec_index].iov_base;
+        block_size = iovec[vec_index].iov_len;
+
+        copy_size = MIN(size - bytes_copied, block_size - skip_offset);
+        sys_memcpy(&data[bytes_copied], &block_start[skip_offset], copy_size);
+
+        bytes_copied += copy_size;
+        skip_offset = 0;
+
+        vec_index++;
+    }
+}
+
+/* Convenience function for copy_from_iovec over queues. */
+static void copy_from_queue(ErlNifIOQueue *queue, int queue_skip,
+    size_t byte_skip, size_t size, char *data) {
+
+    SysIOVec *queued_data;
+    int queue_length;
+
+    queued_data = enif_ioq_peek(queue, &queue_length);
+    ASSERT(queue_skip < queue_length);
+
+    copy_from_iovec(queued_data, queue_length, queue_skip, byte_skip, size, data);
+}
+
+static int enqueue_write_accumulator(buffer_data_t *buffer) {
+    ASSERT(!buffer->accumulator_present ^ (buffer->accumulated_bytes > 0));
+
+    if(buffer->accumulator_present && buffer->accumulated_bytes > 0) {
+        if(!enif_realloc_binary(&buffer->accumulator, buffer->accumulated_bytes)) {
+            return 0;
+        } else if(!enif_ioq_enq_binary(buffer->queue, &buffer->accumulator, 0)) {
+            return 0;
+        }
+
+        /* The queue owns the accumulator now. */
+        buffer->accumulator_present = 0;
+        buffer->accumulated_bytes = 0;
+    }
+
+    return 1;
+}
+
+static int combine_small_writes(buffer_data_t *buffer, ErlNifIOVec *iovec) {
+    ASSERT(!buffer->accumulator_present ^ (buffer->accumulated_bytes > 0));
+
+    if(buffer->accumulated_bytes + iovec->size >= ACCUMULATOR_SIZE) {
+        if(iovec->size >= (ACCUMULATOR_SIZE / 2)) {
+            return 0;
+        }
+
+        if(!enqueue_write_accumulator(buffer)) {
+            return 0;
+        }
+    }
+
+    if(!buffer->accumulator_present) {
+        if(!enif_alloc_binary(ACCUMULATOR_SIZE, &buffer->accumulator)) {
+            return 0;
+        }
+
+        buffer->accumulator_present = 1;
+    }
+
+    copy_from_iovec(iovec->iov, iovec->iovcnt, 0, 0, iovec->size,
+            (char*)&buffer->accumulator.data[buffer->accumulated_bytes]);
+    buffer->accumulated_bytes += iovec->size;
+
+    return 1;
+}
+
+/* *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** */
+
+static ERL_NIF_TERM new_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    buffer_data_t *buffer;
+    ERL_NIF_TERM result;
+
+    buffer = (buffer_data_t*)enif_alloc_resource(rtype_buffer, sizeof(buffer_data_t));
+    buffer->queue = enif_ioq_create(ERL_NIF_IOQ_NORMAL);
+
+    if(buffer->queue != NULL) {
+#ifdef DEBUG
+        erts_atomic32_init_nob(&buffer->concurrent_users, 0);
+#endif
+        erts_atomic32_init_nob(&buffer->external_lock, 0);
+
+        buffer->accumulator_present = 0;
+        buffer->accumulated_bytes = 0;
+
+        result = enif_make_resource(env, buffer);
+    } else {
+        result = enif_raise_exception(env, am_out_of_memory);
+    }
+
+    enif_release_resource(buffer);
+
+    return result;
+}
+
+static ERL_NIF_TERM size_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    buffer_data_t *buffer;
+
+    size_t total_size;
+
+    if(argc != 1 || !get_buffer_data(env, argv[0], &buffer)) {
+        return enif_make_badarg(env);
+    }
+
+    ASSERT(erts_atomic32_inc_read_acqb(&buffer->concurrent_users) == 1);
+
+    total_size = enif_ioq_size(buffer->queue);
+
+    if(buffer->accumulator_present) {
+        total_size += buffer->accumulated_bytes;
+    } else {
+        ASSERT(buffer->accumulated_bytes == 0);
+    }
+
+    ASSERT(erts_atomic32_dec_read_relb(&buffer->concurrent_users) == 0);
+
+    return enif_make_uint64(env, total_size);
+}
+
+static ERL_NIF_TERM copying_read_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    buffer_data_t *buffer;
+
+    ERL_NIF_TERM result;
+    unsigned char *data;
+    Uint64 block_size;
+
+    if(argc != 2 || !get_buffer_data(env, argv[0], &buffer)
+                 || !enif_get_uint64(env, argv[1], &block_size)) {
+        return enif_make_badarg(env);
+    }
+
+    ASSERT(erts_atomic32_inc_read_acqb(&buffer->concurrent_users) == 1);
+
+    if(!enqueue_write_accumulator(buffer)) {
+        return enif_raise_exception(env, am_out_of_memory);
+    }
+
+    if(enif_ioq_size(buffer->queue) < block_size) {
+        return enif_make_badarg(env);
+    }
+
+    data = enif_make_new_binary(env, block_size, &result);
+
+    if(block_size > 0) {
+        copy_from_queue(buffer->queue, 0, 0, block_size, (char*)data);
+        enif_ioq_deq(buffer->queue, block_size, NULL);
+    }
+
+    ASSERT(erts_atomic32_dec_read_relb(&buffer->concurrent_users) == 0);
+
+    return result;
+}
+
+static ERL_NIF_TERM write_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    buffer_data_t *buffer;
+
+    ErlNifIOVec vec, *iovec = &vec;
+    ERL_NIF_TERM tail;
+
+    if(argc != 2 || !get_buffer_data(env, argv[0], &buffer)
+                 || !enif_inspect_iovec(env, 64, argv[1], &tail, &iovec)) {
+        return enif_make_badarg(env);
+    }
+
+    ASSERT(erts_atomic32_inc_read_acqb(&buffer->concurrent_users) == 1);
+
+    if(!combine_small_writes(buffer, iovec)) {
+        if(!enqueue_write_accumulator(buffer) || !enif_ioq_enqv(buffer->queue, iovec, 0)) {
+            return enif_raise_exception(env, am_out_of_memory);
+        }
+    }
+
+    ASSERT(erts_atomic32_dec_read_relb(&buffer->concurrent_users) == 0);
+
+    if(!enif_is_empty_list(env, tail)) {
+        const ERL_NIF_TERM new_argv[2] = {argv[0], tail};
+
+        return enif_schedule_nif(env, "write", 0, &write_nif, argc, new_argv);
+    }
+
+    return am_ok;
+}
+
+static ERL_NIF_TERM peek_head_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    buffer_data_t *buffer;
+
+    ERL_NIF_TERM result;
+
+    if(argc != 1 || !get_buffer_data(env, argv[0], &buffer)) {
+        return enif_make_badarg(env);
+    }
+
+    ASSERT(erts_atomic32_inc_read_acqb(&buffer->concurrent_users) == 1);
+
+    if(!enqueue_write_accumulator(buffer)) {
+        return enif_raise_exception(env, am_out_of_memory);
+    }
+
+    if(!enif_ioq_peek_head(env, buffer->queue, NULL, &result)) {
+        return enif_make_badarg(env);
+    }
+
+    ASSERT(erts_atomic32_dec_read_relb(&buffer->concurrent_users) == 0);
+
+    return result;
+}
+
+static ERL_NIF_TERM skip_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    buffer_data_t *buffer;
+
+    Uint64 block_size;
+
+    if(argc != 2 || !get_buffer_data(env, argv[0], &buffer)
+                 || !enif_get_uint64(env, argv[1], &block_size)) {
+        return enif_make_badarg(env);
+    }
+
+    ASSERT(erts_atomic32_inc_read_acqb(&buffer->concurrent_users) == 1);
+
+    if(!enqueue_write_accumulator(buffer)) {
+        return enif_raise_exception(env, am_out_of_memory);
+    } else if(enif_ioq_size(buffer->queue) < block_size) {
+        return enif_make_badarg(env);
+    }
+
+    enif_ioq_deq(buffer->queue, block_size, NULL);
+
+    ASSERT(erts_atomic32_dec_read_relb(&buffer->concurrent_users) == 0);
+
+    return am_ok;
+}
+
+static ERL_NIF_TERM find_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    buffer_data_t *buffer;
+
+    int queue_length, queue_index;
+    SysIOVec *queued_data;
+    size_t queue_size;
+
+    size_t search_offset;
+    int needle;
+
+    if(argc != 2 || !get_buffer_data(env, argv[0], &buffer)
+                 || !enif_get_int(env, argv[1], &needle)) {
+        return enif_make_badarg(env);
+    }
+
+    ASSERT(erts_atomic32_inc_read_acqb(&buffer->concurrent_users) == 1);
+
+    if(!enqueue_write_accumulator(buffer)) {
+        return enif_raise_exception(env, am_out_of_memory);
+    } else if(needle < 0 || needle > 255) {
+        return enif_make_badarg(env);
+    }
+
+    queued_data = enif_ioq_peek(buffer->queue, &queue_length);
+    queue_size = enif_ioq_size(buffer->queue);
+    queue_index = 0;
+
+    search_offset = 0;
+
+    if(queue_size > (FIND_NIF_RESCHEDULE_SIZE / 100)) {
+        if(enif_thread_type() == ERL_NIF_THR_NORMAL_SCHEDULER) {
+            int timeslice_percent;
+
+            if(queue_size >= FIND_NIF_RESCHEDULE_SIZE) {
+                ASSERT(erts_atomic32_dec_read_relb(&buffer->concurrent_users) == 0);
+
+                return enif_schedule_nif(env, "find",
+                    ERL_NIF_DIRTY_JOB_CPU_BOUND, &find_nif, argc, argv);
+            }
+
+            timeslice_percent = (queue_size * 100) / FIND_NIF_RESCHEDULE_SIZE;
+            enif_consume_timeslice(env, timeslice_percent);
+        }
+    }
+
+    while(queue_index < queue_length) {
+        char *needle_address;
+        char *block_start;
+        size_t block_size;
+
+        block_start = queued_data[queue_index].iov_base;
+        block_size = queued_data[queue_index].iov_len;
+
+        needle_address = memchr(block_start, needle, block_size);
+
+        if(needle_address != NULL) {
+            size_t result = search_offset + (needle_address - block_start);
+
+            ASSERT(erts_atomic32_dec_read_relb(&buffer->concurrent_users) == 0);
+
+            return enif_make_tuple2(env, am_ok, enif_make_uint64(env, result));
+        }
+
+        search_offset += block_size;
+        queue_index++;
+    }
+
+    ASSERT(erts_atomic32_dec_read_relb(&buffer->concurrent_users) == 0);
+
+    return am_not_found;
+}
+
+/* */
+
+static ERL_NIF_TERM trylock_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    buffer_data_t *buffer;
+
+    if(argc != 1 || !get_buffer_data(env, argv[0], &buffer)) {
+        return enif_make_badarg(env);
+    }
+
+    if(erts_atomic32_cmpxchg_acqb(&buffer->external_lock, 1, 0) == 0) {
+        return am_acquired;
+    }
+
+    return am_busy;
+}
+
+static ERL_NIF_TERM unlock_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    buffer_data_t *buffer;
+
+    if(argc != 1 || !get_buffer_data(env, argv[0], &buffer)) {
+        return enif_make_badarg(env);
+    }
+
+    if(erts_atomic32_cmpxchg_relb(&buffer->external_lock, 0, 1) == 0) {
+        return enif_raise_exception(env, am_lock_order_violation);
+    }
+
+    return am_ok;
+}
diff --git a/erts/emulator/nifs/common/prim_file_nif.c b/erts/emulator/nifs/common/prim_file_nif.c
new file mode 100644
index 0000000000..6874f41d75
--- /dev/null
+++ b/erts/emulator/nifs/common/prim_file_nif.c
@@ -0,0 +1,1237 @@
+/*
+ * %CopyrightBegin%
+ *
+ * Copyright Ericsson 2017. 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%
+ */
+
+#define STATIC_ERLANG_NIF 1
+
+#include "erl_nif.h"
+#include "config.h"
+#include "sys.h"
+
+#ifdef VALGRIND
+#  include <valgrind/memcheck.h>
+#endif
+
+#include "erl_driver.h"
+#include "prim_file_nif.h"
+
+/* NIF interface declarations */
+static int load(ErlNifEnv *env, void** priv_data, ERL_NIF_TERM load_info);
+static int upgrade(ErlNifEnv *env, void** priv_data, void** old_priv_data, ERL_NIF_TERM load_info);
+static void unload(ErlNifEnv *env, void* priv_data);
+
+static ErlNifResourceType *efile_resource_type;
+
+static ERL_NIF_TERM am_ok;
+static ERL_NIF_TERM am_error;
+static ERL_NIF_TERM am_continue;
+
+static ERL_NIF_TERM am_file_info;
+
+/* File modes */
+static ERL_NIF_TERM am_read;
+static ERL_NIF_TERM am_write;
+static ERL_NIF_TERM am_exclusive;
+static ERL_NIF_TERM am_append;
+static ERL_NIF_TERM am_sync;
+static ERL_NIF_TERM am_skip_type_check;
+
+/* enum efile_access_t; read and write are defined above.*/
+static ERL_NIF_TERM am_read_write;
+static ERL_NIF_TERM am_none;
+
+/* enum efile_advise_t */
+static ERL_NIF_TERM am_normal;
+static ERL_NIF_TERM am_random;
+static ERL_NIF_TERM am_sequential;
+static ERL_NIF_TERM am_will_need;
+static ERL_NIF_TERM am_dont_need;
+static ERL_NIF_TERM am_no_reuse;
+
+/* enum efile_filetype_t */
+static ERL_NIF_TERM am_device;
+static ERL_NIF_TERM am_directory;
+static ERL_NIF_TERM am_regular;
+static ERL_NIF_TERM am_symlink;
+static ERL_NIF_TERM am_other;
+
+/* enum efile_seek_t, 'eof' marker. */
+static ERL_NIF_TERM am_bof;
+static ERL_NIF_TERM am_cur;
+static ERL_NIF_TERM am_eof;
+
+static ERL_NIF_TERM read_info_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]);
+static ERL_NIF_TERM set_permissions_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]);
+static ERL_NIF_TERM set_owner_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]);
+static ERL_NIF_TERM set_time_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]);
+
+static ERL_NIF_TERM read_link_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]);
+static ERL_NIF_TERM list_dir_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]);
+
+static ERL_NIF_TERM make_hard_link_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]);
+static ERL_NIF_TERM make_soft_link_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]);
+static ERL_NIF_TERM rename_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]);
+static ERL_NIF_TERM make_dir_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]);
+static ERL_NIF_TERM del_file_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]);
+static ERL_NIF_TERM del_dir_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]);
+static ERL_NIF_TERM get_device_cwd_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]);
+static ERL_NIF_TERM get_cwd_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]);
+static ERL_NIF_TERM set_cwd_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]);
+
+static ERL_NIF_TERM read_file_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]);
+
+static ERL_NIF_TERM get_handle_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]);
+static ERL_NIF_TERM altname_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]);
+
+static ERL_NIF_TERM open_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]);
+
+/* All file handle operations are passed through a wrapper that handles state
+ * transitions, marking it as busy during the course of the operation, and
+ * closing on completion if the owner died in the middle of an operation.
+ *
+ * This is pretty ugly but required as there's no way to tell when it's safe to
+ * asynchronously close a file; the event could have fired just before landing
+ * in a system call which will fail with EBADF at best or alias a newly opened
+ * fd at worst.
+ *
+ * The old driver got away with enqueueing the close operation on the same
+ * async queue as all of its other operations, but since dirty schedulers use a
+ * single global queue there's no natural way to schedule an asynchronous close
+ * "behind" other operations.
+ *
+ * The states may transition as follows:
+ *
+ * IDLE ->
+ *      BUSY (file_handle_wrapper) |
+ *      CLOSED (owner_death_callback)
+ *
+ * BUSY ->
+ *      IDLE (file_handle_wrapper)
+ *      CLOSED (close_nif_impl)
+ *      CLOSE_PENDING (owner_death_callback)
+ *
+ * CLOSE_PENDING ->
+ *      CLOSED (file_handle_wrapper)
+ */
+
+typedef ERL_NIF_TERM (*file_op_impl_t)(efile_data_t *d, ErlNifEnv *env,
+    int argc, const ERL_NIF_TERM argv[]);
+static ERL_NIF_TERM file_handle_wrapper(file_op_impl_t operation, ErlNifEnv *env,
+    int argc, const ERL_NIF_TERM argv[]);
+
+#define WRAP_FILE_HANDLE_EXPORT(name) \
+    static ERL_NIF_TERM name ## _impl (efile_data_t *d, ErlNifEnv *env, \
+        int argc, const ERL_NIF_TERM argv[]);\
+    static ERL_NIF_TERM name(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) { \
+        return file_handle_wrapper( name ## _impl , env, argc, argv); \
+    }
+
+WRAP_FILE_HANDLE_EXPORT(close_nif)
+WRAP_FILE_HANDLE_EXPORT(read_nif)
+WRAP_FILE_HANDLE_EXPORT(write_nif)
+WRAP_FILE_HANDLE_EXPORT(pread_nif)
+WRAP_FILE_HANDLE_EXPORT(pwrite_nif)
+WRAP_FILE_HANDLE_EXPORT(seek_nif)
+WRAP_FILE_HANDLE_EXPORT(sync_nif)
+WRAP_FILE_HANDLE_EXPORT(truncate_nif)
+WRAP_FILE_HANDLE_EXPORT(allocate_nif)
+WRAP_FILE_HANDLE_EXPORT(advise_nif)
+WRAP_FILE_HANDLE_EXPORT(get_handle_nif)
+WRAP_FILE_HANDLE_EXPORT(ipread_s32bu_p32bu_nif)
+
+static ErlNifFunc nif_funcs[] = {
+    /* File handle ops */
+    {"open_nif", 2, open_nif, ERL_NIF_DIRTY_JOB_IO_BOUND},
+    {"close_nif", 1, close_nif, ERL_NIF_DIRTY_JOB_IO_BOUND},
+    {"read_nif", 2, read_nif, ERL_NIF_DIRTY_JOB_IO_BOUND},
+    {"write_nif", 2, write_nif, ERL_NIF_DIRTY_JOB_IO_BOUND},
+    {"pread_nif", 3, pread_nif, ERL_NIF_DIRTY_JOB_IO_BOUND},
+    {"pwrite_nif", 3, pwrite_nif, ERL_NIF_DIRTY_JOB_IO_BOUND},
+    {"seek_nif", 3, seek_nif, ERL_NIF_DIRTY_JOB_IO_BOUND},
+    {"sync_nif", 2, sync_nif, ERL_NIF_DIRTY_JOB_IO_BOUND},
+    {"truncate_nif", 1, truncate_nif, ERL_NIF_DIRTY_JOB_IO_BOUND},
+    {"allocate_nif", 3, allocate_nif, ERL_NIF_DIRTY_JOB_IO_BOUND},
+    {"advise_nif", 4, advise_nif, ERL_NIF_DIRTY_JOB_IO_BOUND},
+
+    /* Filesystem ops */
+    {"make_hard_link_nif", 2, make_hard_link_nif, ERL_NIF_DIRTY_JOB_IO_BOUND},
+    {"make_soft_link_nif", 2, make_soft_link_nif, ERL_NIF_DIRTY_JOB_IO_BOUND},
+    {"rename_nif", 2, rename_nif, ERL_NIF_DIRTY_JOB_IO_BOUND},
+    {"read_info_nif", 2, read_info_nif, ERL_NIF_DIRTY_JOB_IO_BOUND},
+    {"set_permissions_nif", 2, set_permissions_nif, ERL_NIF_DIRTY_JOB_IO_BOUND},
+    {"set_owner_nif", 3, set_owner_nif, ERL_NIF_DIRTY_JOB_IO_BOUND},
+    {"set_time_nif", 4, set_time_nif, ERL_NIF_DIRTY_JOB_IO_BOUND},
+    {"read_link_nif", 1, read_link_nif, ERL_NIF_DIRTY_JOB_IO_BOUND},
+    {"list_dir_nif", 1, list_dir_nif, ERL_NIF_DIRTY_JOB_IO_BOUND},
+    {"make_dir_nif", 1, make_dir_nif, ERL_NIF_DIRTY_JOB_IO_BOUND},
+    {"del_file_nif", 1, del_file_nif, ERL_NIF_DIRTY_JOB_IO_BOUND},
+    {"del_dir_nif", 1, del_dir_nif, ERL_NIF_DIRTY_JOB_IO_BOUND},
+    {"get_device_cwd_nif", 1, get_device_cwd_nif, ERL_NIF_DIRTY_JOB_IO_BOUND},
+    {"set_cwd_nif", 1, set_cwd_nif, ERL_NIF_DIRTY_JOB_IO_BOUND},
+    {"get_cwd_nif", 0, get_cwd_nif, ERL_NIF_DIRTY_JOB_IO_BOUND},
+
+    /* These operations are equivalent to chained calls of other operations,
+     * but have been moved down to avoid excessive rescheduling. */
+    {"ipread_s32bu_p32bu_nif", 3, ipread_s32bu_p32bu_nif, ERL_NIF_DIRTY_JOB_IO_BOUND},
+    {"read_file_nif", 1, read_file_nif, ERL_NIF_DIRTY_JOB_IO_BOUND},
+
+    /* Internal ops. */
+    {"get_handle_nif", 1, get_handle_nif},
+    {"altname_nif", 1, altname_nif, ERL_NIF_DIRTY_JOB_IO_BOUND},
+};
+
+ERL_NIF_INIT(prim_file, nif_funcs, load, NULL, upgrade, unload)
+
+static void owner_death_callback(ErlNifEnv* env, void* obj, ErlNifPid* pid, ErlNifMonitor* mon);
+static void gc_callback(ErlNifEnv *env, void* data);
+
+static int load(ErlNifEnv *env, void** priv_data, ERL_NIF_TERM load_info)
+{
+    ErlNifResourceTypeInit callbacks;
+
+    am_ok = enif_make_atom(env, "ok");
+    am_error = enif_make_atom(env, "error");
+    am_continue = enif_make_atom(env, "continue");
+
+    am_read = enif_make_atom(env, "read");
+    am_write = enif_make_atom(env, "write");
+    am_exclusive = enif_make_atom(env, "exclusive");
+    am_append = enif_make_atom(env, "append");
+    am_sync = enif_make_atom(env, "sync");
+    am_skip_type_check = enif_make_atom(env, "skip_type_check");
+
+    am_read_write = enif_make_atom(env, "read_write");
+    am_none = enif_make_atom(env, "none");
+
+    am_normal = enif_make_atom(env, "normal");
+    am_random = enif_make_atom(env, "random");
+    am_sequential = enif_make_atom(env, "sequential");
+    am_will_need = enif_make_atom(env, "will_need");
+    am_dont_need = enif_make_atom(env, "dont_need");
+    am_no_reuse = enif_make_atom(env, "no_reuse");
+
+    am_device = enif_make_atom(env, "device");
+    am_directory = enif_make_atom(env, "directory");
+    am_regular = enif_make_atom(env, "regular");
+    am_symlink = enif_make_atom(env, "symlink");
+    am_other = enif_make_atom(env, "other");
+
+    am_file_info = enif_make_atom(env, "file_info");
+
+    am_bof = enif_make_atom(env, "bof");
+    am_cur = enif_make_atom(env, "cur");
+    am_eof = enif_make_atom(env, "eof");
+
+    callbacks.down = owner_death_callback;
+    callbacks.dtor = gc_callback;
+    callbacks.stop = NULL;
+
+    efile_resource_type = enif_open_resource_type_x(env, "efile", &callbacks,
+        ERL_NIF_RT_CREATE, NULL);
+
+    *priv_data = NULL;
+
+    return 0;
+}
+
+static void unload(ErlNifEnv *env, void* priv_data)
+{
+
+}
+
+static int upgrade(ErlNifEnv *env, void** priv_data, void** old_priv_data, ERL_NIF_TERM load_info)
+{
+    if(*old_priv_data != NULL) {
+        return -1; /* Don't know how to do that */
+    }
+    if(*priv_data != NULL) {
+        return -1; /* Don't know how to do that */
+    }
+    if(load(env, priv_data, load_info)) {
+        return -1;
+    }
+    return 0;
+}
+
+static ERL_NIF_TERM posix_error_to_tuple(ErlNifEnv *env, posix_errno_t posix_errno) {
+    ERL_NIF_TERM error = enif_make_atom(env, erl_errno_id(posix_errno));
+    return enif_make_tuple2(env, am_error, error);
+}
+
+static int get_file_data(ErlNifEnv *env, ERL_NIF_TERM opaque, efile_data_t **d) {
+    return enif_get_resource(env, opaque, efile_resource_type, (void **)d);
+}
+
+static ERL_NIF_TERM file_handle_wrapper(file_op_impl_t operation, ErlNifEnv *env,
+        int argc, const ERL_NIF_TERM argv[]) {
+
+    efile_data_t *d;
+
+    enum efile_state_t previous_state;
+    ERL_NIF_TERM result;
+
+    if(argc < 1 || !get_file_data(env, argv[0], &d)) {
+        return enif_make_badarg(env);
+    }
+
+    previous_state = erts_atomic32_cmpxchg_acqb(&d->state,
+        EFILE_STATE_BUSY, EFILE_STATE_IDLE);
+
+    if(previous_state == EFILE_STATE_IDLE) {
+        result = operation(d, env, argc - 1, &argv[1]);
+
+        previous_state = erts_atomic32_cmpxchg_relb(&d->state,
+            EFILE_STATE_IDLE, EFILE_STATE_BUSY);
+
+        ASSERT(previous_state != EFILE_STATE_IDLE);
+
+        if(previous_state == EFILE_STATE_CLOSE_PENDING) {
+            /* This is the only point where a change from CLOSE_PENDING is
+             * possible, and we're running synchronously, so we can't race with
+             * anything else here. */
+            erts_atomic32_set_acqb(&d->state, EFILE_STATE_CLOSED);
+            efile_close(d);
+        }
+    } else {
+        /* CLOSE_PENDING should be impossible at this point since it requires
+         * a transition from BUSY; the only valid state here is CLOSED. */
+        ASSERT(previous_state == EFILE_STATE_CLOSED);
+
+        result = posix_error_to_tuple(env, EINVAL);
+    }
+
+    return result;
+}
+
+static void owner_death_callback(ErlNifEnv* env, void* obj, ErlNifPid* pid, ErlNifMonitor* mon) {
+    efile_data_t *d = (efile_data_t*)obj;
+
+    (void)env;
+    (void)pid;
+    (void)mon;
+
+    for(;;) {
+        enum efile_state_t previous_state;
+
+        previous_state = erts_atomic32_cmpxchg_acqb(&d->state,
+            EFILE_STATE_CLOSED, EFILE_STATE_IDLE);
+
+        switch(previous_state) {
+        case EFILE_STATE_IDLE:
+            efile_close(d);
+            return;
+        case EFILE_STATE_CLOSE_PENDING:
+        case EFILE_STATE_CLOSED:
+            /* We're either already closed or managed to mark ourselves for
+             * closure in the previous iteration. */
+            return;
+        case EFILE_STATE_BUSY:
+            /* Schedule ourselves to be closed once the current operation
+             * finishes, retrying the [IDLE -> CLOSED] transition in case we
+             * narrowly passed the [BUSY -> IDLE] one. */
+            erts_atomic32_cmpxchg_nob(&d->state,
+                EFILE_STATE_CLOSE_PENDING, EFILE_STATE_BUSY);
+            break;
+        }
+    }
+}
+
+static void gc_callback(ErlNifEnv *env, void* data) {
+    efile_data_t *d = (efile_data_t*)data;
+
+    enum efile_state_t previous_state;
+
+    (void)env;
+
+    previous_state = erts_atomic32_cmpxchg_acqb(&d->state,
+        EFILE_STATE_CLOSED, EFILE_STATE_IDLE);
+
+    ASSERT(previous_state != EFILE_STATE_CLOSE_PENDING &&
+        previous_state != EFILE_STATE_BUSY);
+
+    if(previous_state == EFILE_STATE_IDLE) {
+        efile_close(d);
+    }
+}
+
+static ERL_NIF_TERM efile_filetype_to_atom(enum efile_filetype_t type) {
+    switch(type) {
+        case EFILE_FILETYPE_DEVICE: return am_device;
+        case EFILE_FILETYPE_DIRECTORY: return am_directory;
+        case EFILE_FILETYPE_REGULAR: return am_regular;
+        case EFILE_FILETYPE_SYMLINK: return am_symlink;
+        case EFILE_FILETYPE_OTHER: return am_other;
+    }
+
+    return am_other;
+}
+
+static ERL_NIF_TERM efile_access_to_atom(enum efile_access_t type) {
+    if(type & EFILE_ACCESS_READ && !(type & EFILE_ACCESS_WRITE)) {
+        return am_read;
+    } else if(type & EFILE_ACCESS_WRITE && !(type & EFILE_ACCESS_READ)) {
+        return am_write;
+    } else if(type & EFILE_ACCESS_READ_WRITE) {
+        return am_read_write;
+    }
+
+    return am_none;
+}
+
+static enum efile_modes_t efile_translate_modelist(ErlNifEnv *env, ERL_NIF_TERM list) {
+    enum efile_modes_t modes;
+    ERL_NIF_TERM head, tail;
+
+    modes = 0;
+
+    while(enif_get_list_cell(env, list, &head, &tail)) {
+        if(enif_is_identical(head, am_read)) {
+            modes |= EFILE_MODE_READ;
+        } else if(enif_is_identical(head, am_write)) {
+            modes |= EFILE_MODE_WRITE;
+        } else if(enif_is_identical(head, am_exclusive)) {
+            modes |= EFILE_MODE_EXCLUSIVE;
+        } else if(enif_is_identical(head, am_append)) {
+            modes |= EFILE_MODE_APPEND;
+        } else if(enif_is_identical(head, am_sync)) {
+            modes |= EFILE_MODE_SYNC;
+        } else if(enif_is_identical(head, am_skip_type_check)) {
+            modes |= EFILE_MODE_SKIP_TYPE_CHECK;
+        } else {
+            /* Modes like 'raw', 'ram', 'delayed_writes' etc are handled
+             * further up the chain. */
+        }
+
+        list = tail;
+    }
+
+    if(modes & (EFILE_MODE_APPEND | EFILE_MODE_EXCLUSIVE)) {
+        /* 'append' and 'exclusive' are documented as "open for writing." */
+        modes |= EFILE_MODE_WRITE;
+    } else if(!(modes & EFILE_MODE_READ_WRITE)) {
+        /* Defaulting to read if !(W|R) is undocumented, but specifically
+         * tested against in file_SUITE. */
+        modes |= EFILE_MODE_READ;
+    }
+
+    return modes;
+}
+
+static ERL_NIF_TERM open_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    posix_errno_t posix_errno;
+    efile_data_t *d;
+
+    ErlNifPid controlling_process;
+    enum efile_modes_t modes;
+    ERL_NIF_TERM result;
+    efile_path_t path;
+
+    if(argc != 2 || !enif_is_list(env, argv[1])) {
+        return enif_make_badarg(env);
+    }
+
+    modes = efile_translate_modelist(env, argv[1]);
+
+    if((posix_errno = efile_marshal_path(env, argv[0], &path))) {
+        return posix_error_to_tuple(env, posix_errno);
+    } else if((posix_errno = efile_open(&path, modes, efile_resource_type, &d))) {
+        return posix_error_to_tuple(env, posix_errno);
+    }
+
+    result = enif_make_resource(env, d);
+    enif_release_resource(d);
+
+    enif_self(env, &controlling_process);
+
+    if(enif_monitor_process(env, d, &controlling_process, &d->monitor)) {
+        return posix_error_to_tuple(env, EINVAL);
+    }
+
+    return enif_make_tuple2(env, am_ok, result);
+}
+
+static ERL_NIF_TERM close_nif_impl(efile_data_t *d, ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    enum efile_state_t previous_state;
+
+    if(argc != 0) {
+        return enif_make_badarg(env);
+    }
+
+    previous_state = erts_atomic32_cmpxchg_acqb(&d->state,
+        EFILE_STATE_CLOSED, EFILE_STATE_BUSY);
+
+    ASSERT(previous_state == EFILE_STATE_CLOSE_PENDING ||
+        previous_state == EFILE_STATE_BUSY);
+
+    if(previous_state == EFILE_STATE_BUSY) {
+        enif_demonitor_process(env, d, &d->monitor);
+
+        if(!efile_close(d)) {
+            return posix_error_to_tuple(env, d->posix_errno);
+        }
+    }
+
+    return am_ok;
+}
+
+static ERL_NIF_TERM read_nif_impl(efile_data_t *d, ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    Sint64 bytes_read, block_size;
+    SysIOVec read_vec[1];
+    ErlNifBinary result;
+
+    if(argc != 1 || !enif_is_number(env, argv[0])) {
+        return enif_make_badarg(env);
+    }
+
+    if(!enif_get_int64(env, argv[0], &block_size) || block_size < 0) {
+        return posix_error_to_tuple(env, EINVAL);
+    }
+
+    if(!enif_alloc_binary(block_size, &result)) {
+        return posix_error_to_tuple(env, ENOMEM);
+    }
+
+    read_vec[0].iov_base = result.data;
+    read_vec[0].iov_len = result.size;
+
+    bytes_read = efile_readv(d, read_vec, 1);
+    ASSERT(bytes_read <= block_size);
+
+    if(bytes_read < 0) {
+        return posix_error_to_tuple(env, d->posix_errno);
+    } else if(bytes_read == 0) {
+        enif_release_binary(&result);
+        return am_eof;
+    }
+
+    if(bytes_read < block_size && !enif_realloc_binary(&result, bytes_read)) {
+        ERTS_INTERNAL_ERROR("Failed to shrink read result.");
+    }
+
+    return enif_make_tuple2(env, am_ok, enif_make_binary(env, &result));
+}
+
+static ERL_NIF_TERM write_nif_impl(efile_data_t *d, ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    ErlNifIOVec vec, *input = &vec;
+    Sint64 bytes_written;
+    ERL_NIF_TERM tail;
+
+    if(argc != 1 || !enif_inspect_iovec(env, 64, argv[0], &tail, &input)) {
+        return enif_make_badarg(env);
+    }
+
+    bytes_written = efile_writev(d, input->iov, input->iovcnt);
+
+    if(bytes_written < 0) {
+        return posix_error_to_tuple(env, d->posix_errno);
+    }
+
+    if(!enif_is_empty_list(env, tail)) {
+        ASSERT(bytes_written > 0);
+        return enif_make_tuple2(env, am_continue, tail);
+    }
+
+    return am_ok;
+}
+
+static ERL_NIF_TERM pread_nif_impl(efile_data_t *d, ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    Sint64 bytes_read, block_size, offset;
+    SysIOVec read_vec[1];
+    ErlNifBinary result;
+
+    if(argc != 2 || !enif_is_number(env, argv[0])
+                 || !enif_is_number(env, argv[1])) {
+        return enif_make_badarg(env);
+    }
+
+    if(!enif_get_int64(env, argv[0], &offset) ||
+       !enif_get_int64(env, argv[1], &block_size) ||
+       (offset < 0 || block_size < 0)) {
+        return posix_error_to_tuple(env, EINVAL);
+    }
+
+    if(!enif_alloc_binary(block_size, &result)) {
+        return posix_error_to_tuple(env, ENOMEM);
+    }
+
+    read_vec[0].iov_base = result.data;
+    read_vec[0].iov_len = result.size;
+
+    bytes_read = efile_preadv(d, offset, read_vec, 1);
+
+    if(bytes_read < 0) {
+        return posix_error_to_tuple(env, d->posix_errno);
+    } else if(bytes_read == 0) {
+        enif_release_binary(&result);
+        return am_eof;
+    }
+
+    if(bytes_read < block_size && !enif_realloc_binary(&result, bytes_read)) {
+        ERTS_INTERNAL_ERROR("Failed to shrink pread result.");
+    }
+
+    return enif_make_tuple2(env, am_ok, enif_make_binary(env, &result));
+}
+
+static ERL_NIF_TERM pwrite_nif_impl(efile_data_t *d, ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    ErlNifIOVec vec, *input = &vec;
+    Sint64 bytes_written, offset;
+    ERL_NIF_TERM tail;
+
+    if(argc != 2 || !enif_is_number(env, argv[0])
+                 || !enif_inspect_iovec(env, 64, argv[1], &tail, &input)) {
+        return enif_make_badarg(env);
+    }
+
+    if(!enif_get_int64(env, argv[0], &offset) || offset < 0) {
+        return posix_error_to_tuple(env, EINVAL);
+    }
+
+    bytes_written = efile_pwritev(d, offset, input->iov, input->iovcnt);
+
+    if(bytes_written < 0) {
+        return posix_error_to_tuple(env, d->posix_errno);
+    }
+
+    if(!enif_is_empty_list(env, tail)) {
+        ASSERT(bytes_written > 0);
+        return enif_make_tuple3(env, am_continue,
+            enif_make_int64(env, bytes_written), tail);
+    }
+
+    return am_ok;
+}
+
+static ERL_NIF_TERM seek_nif_impl(efile_data_t *d, ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    Sint64 new_position, offset;
+    enum efile_seek_t seek;
+
+    if(argc != 2 || !enif_get_int64(env, argv[1], &offset)) {
+        return enif_make_badarg(env);
+    }
+
+    if(enif_is_identical(argv[0], am_bof)) {
+        seek = EFILE_SEEK_BOF;
+    } else if(enif_is_identical(argv[0], am_cur)) {
+        seek = EFILE_SEEK_CUR;
+    } else if(enif_is_identical(argv[0], am_eof)) {
+        seek = EFILE_SEEK_EOF;
+    } else {
+        return enif_make_badarg(env);
+    }
+
+    if(!efile_seek(d, seek, offset, &new_position)) {
+        return posix_error_to_tuple(env, d->posix_errno);
+    }
+
+    return enif_make_tuple2(env, am_ok, enif_make_uint64(env, new_position));
+}
+
+static ERL_NIF_TERM sync_nif_impl(efile_data_t *d, ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    int data_only;
+
+    if(argc != 1 || !enif_get_int(env, argv[0], &data_only)) {
+        return enif_make_badarg(env);
+    }
+
+    if(!efile_sync(d, data_only)) {
+        return posix_error_to_tuple(env, d->posix_errno);
+    }
+
+    return am_ok;
+}
+
+static ERL_NIF_TERM truncate_nif_impl(efile_data_t *d, ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    if(argc != 0) {
+        return enif_make_badarg(env);
+    }
+
+    if(!efile_truncate(d)) {
+        return posix_error_to_tuple(env, d->posix_errno);
+    }
+
+    return am_ok;
+}
+
+static ERL_NIF_TERM allocate_nif_impl(efile_data_t *d, ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    Sint64 offset, length;
+
+    if(argc != 2 || !enif_is_number(env, argv[0])
+                 || !enif_is_number(env, argv[1])) {
+        return enif_make_badarg(env);
+    }
+
+    if(!enif_get_int64(env, argv[0], &offset) ||
+       !enif_get_int64(env, argv[1], &length) ||
+       (offset < 0 || length < 0)) {
+        return posix_error_to_tuple(env, EINVAL);
+    }
+
+    if(!efile_allocate(d, offset, length)) {
+        return posix_error_to_tuple(env, d->posix_errno);
+    }
+
+    return am_ok;
+}
+
+static ERL_NIF_TERM advise_nif_impl(efile_data_t *d, ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    enum efile_advise_t advise;
+    Sint64 offset, length;
+
+    if(argc != 3 || !enif_is_number(env, argv[0])
+                 || !enif_is_number(env, argv[1])) {
+        return enif_make_badarg(env);
+    }
+
+    if(!enif_get_int64(env, argv[0], &offset) ||
+       !enif_get_int64(env, argv[1], &length) ||
+       (offset < 0 || length < 0)) {
+        return posix_error_to_tuple(env, EINVAL);
+    }
+
+    if(enif_is_identical(argv[2], am_normal)) {
+        advise = EFILE_ADVISE_NORMAL;
+    } else if(enif_is_identical(argv[2], am_random)) {
+        advise = EFILE_ADVISE_RANDOM;
+    } else if(enif_is_identical(argv[2], am_sequential)) {
+        advise = EFILE_ADVISE_SEQUENTIAL;
+    } else if(enif_is_identical(argv[2], am_will_need)) {
+        advise = EFILE_ADVISE_WILL_NEED;
+    } else if(enif_is_identical(argv[2], am_dont_need)) {
+        advise = EFILE_ADVISE_DONT_NEED;
+    } else if(enif_is_identical(argv[2], am_no_reuse)) {
+        advise = EFILE_ADVISE_NO_REUSE;
+    } else {
+        /* The tests check for EINVAL instead of badarg. Sigh. */
+        return posix_error_to_tuple(env, EINVAL);
+    }
+
+    if(!efile_advise(d, offset, length, advise)) {
+        return posix_error_to_tuple(env, d->posix_errno);
+    }
+
+    return am_ok;
+}
+
+/* This undocumented function reads a pointer and then reads the data block
+ * described by said pointer. It was reverse-engineered from the old
+ * implementation so while all tests pass it may not be entirely correct. Our
+ * current understanding is as follows:
+ *
+ * Pointer layout:
+ *
+ *     <<Size:1/integer-unit:32, Offset:1/integer-unit:32>>
+ *
+ * Where Offset is the -absolute- address to the data block.
+ *
+ * *) If we fail to read the pointer block in its entirety, we return eof.
+ * *) If the provided max_payload_size is larger than Size, we return eof.
+ * *) If we fail to read any data whatsoever at Offset, we return
+ *    {ok, {Size, Offset, eof}}
+ * *) Otherwise, we return {ok, {Size, Offset, Data}}. Note that the size
+ *    of Data may be smaller than Size if we encounter EOF before we could
+ *    read the entire block.
+ *
+ * On errors we'll return {error, posix()} regardless of whether they
+ * happened before or after reading the pointer block. */
+static ERL_NIF_TERM ipread_s32bu_p32bu_nif_impl(efile_data_t *d, ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    Sint64 payload_offset, payload_size;
+
+    SysIOVec read_vec[1];
+    Sint64 bytes_read;
+
+    ErlNifBinary payload;
+
+    if(argc != 2 || !enif_is_number(env, argv[0])
+                 || !enif_is_number(env, argv[1])) {
+        return enif_make_badarg(env);
+    }
+
+    {
+        Sint64 max_payload_size, pointer_offset;
+        unsigned char pointer_block[8];
+
+        if(!enif_get_int64(env, argv[0], &pointer_offset) ||
+           !enif_get_int64(env, argv[1], &max_payload_size) ||
+           (pointer_offset < 0 || max_payload_size >= 1u << 31)) {
+            return posix_error_to_tuple(env, EINVAL);
+        }
+
+        read_vec[0].iov_base = pointer_block;
+        read_vec[0].iov_len = sizeof(pointer_block);
+
+        bytes_read = efile_preadv(d, pointer_offset, read_vec, 1);
+
+        if(bytes_read < 0) {
+            return posix_error_to_tuple(env, d->posix_errno);
+        } else if(bytes_read < sizeof(pointer_block)) {
+            return am_eof;
+        }
+
+        payload_size = (Uint32)get_int32(&pointer_block[0]);
+        payload_offset = (Uint32)get_int32(&pointer_block[4]);
+
+        if(payload_size > max_payload_size) {
+            return am_eof;
+        }
+    }
+
+    if(!enif_alloc_binary(payload_size, &payload)) {
+        return posix_error_to_tuple(env, ENOMEM);
+    }
+
+    read_vec[0].iov_base = payload.data;
+    read_vec[0].iov_len = payload.size;
+
+    bytes_read = efile_preadv(d, payload_offset, read_vec, 1);
+
+    if(bytes_read < 0) {
+        return posix_error_to_tuple(env, d->posix_errno);
+    } else if(bytes_read == 0) {
+        enif_release_binary(&payload);
+
+        return enif_make_tuple2(env, am_ok,
+                enif_make_tuple3(env,
+                    enif_make_uint(env, payload_size),
+                    enif_make_uint(env, payload_offset),
+                    am_eof));
+    }
+
+    if(bytes_read < payload.size && !enif_realloc_binary(&payload, bytes_read)) {
+        ERTS_INTERNAL_ERROR("Failed to shrink ipread payload.");
+    }
+
+    return enif_make_tuple2(env, am_ok,
+        enif_make_tuple3(env,
+            enif_make_uint(env, payload_size),
+            enif_make_uint(env, payload_offset),
+            enif_make_binary(env, &payload)));
+}
+
+static ERL_NIF_TERM get_handle_nif_impl(efile_data_t *d, ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    if(argc != 0) {
+        return enif_make_badarg(env);
+    }
+
+    return efile_get_handle(env, d);
+}
+
+static ERL_NIF_TERM read_info_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    posix_errno_t posix_errno;
+
+    efile_fileinfo_t info = {0};
+    efile_path_t path;
+    int follow_links;
+
+    if(argc != 2 || !enif_get_int(env, argv[1], &follow_links)) {
+        return enif_make_badarg(env);
+    }
+
+    if((posix_errno = efile_marshal_path(env, argv[0], &path))) {
+        return posix_error_to_tuple(env, posix_errno);
+    } else if((posix_errno = efile_read_info(&path, follow_links, &info))) {
+        return posix_error_to_tuple(env, posix_errno);
+    }
+
+    /* #file_info as declared in file.hrl */
+    return enif_make_tuple(env, 14,
+        am_file_info,
+        enif_make_uint64(env, info.size),
+        efile_filetype_to_atom(info.type),
+        efile_access_to_atom(info.access),
+        enif_make_int64(env, MAX(EFILE_MIN_FILETIME, info.a_time)),
+        enif_make_int64(env, MAX(EFILE_MIN_FILETIME, info.m_time)),
+        enif_make_int64(env, MAX(EFILE_MIN_FILETIME, info.c_time)),
+        enif_make_uint(env, info.mode),
+        enif_make_uint(env, info.links),
+        enif_make_uint(env, info.major_device),
+        enif_make_uint(env, info.minor_device),
+        enif_make_uint(env, info.inode),
+        enif_make_uint(env, info.uid),
+        enif_make_uint(env, info.gid)
+    );
+}
+
+static ERL_NIF_TERM set_permissions_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    posix_errno_t posix_errno;
+
+    efile_path_t path;
+    Uint32 permissions;
+
+    if(argc != 2 || !enif_get_uint(env, argv[1], &permissions)) {
+        return enif_make_badarg(env);
+    }
+
+    if((posix_errno = efile_marshal_path(env, argv[0], &path))) {
+        return posix_error_to_tuple(env, posix_errno);
+    } else if((posix_errno = efile_set_permissions(&path, permissions))) {
+        return posix_error_to_tuple(env, posix_errno);
+    }
+
+    return am_ok;
+}
+
+static ERL_NIF_TERM set_owner_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    posix_errno_t posix_errno;
+
+    efile_path_t path;
+    Uint32 uid, gid;
+
+    if(argc != 3 || !enif_get_uint(env, argv[1], &uid)
+                 || !enif_get_uint(env, argv[2], &gid)) {
+        return enif_make_badarg(env);
+    }
+
+    if((posix_errno = efile_marshal_path(env, argv[0], &path))) {
+        return posix_error_to_tuple(env, posix_errno);
+    } else if((posix_errno = efile_set_owner(&path, uid, gid))) {
+        return posix_error_to_tuple(env, posix_errno);
+    }
+
+    return am_ok;
+}
+
+static ERL_NIF_TERM set_time_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    posix_errno_t posix_errno;
+
+    Sint64 accessed, modified, created;
+    efile_path_t path;
+
+    if(argc != 4 || !enif_get_int64(env, argv[1], &accessed)
+                 || !enif_get_int64(env, argv[2], &modified)
+                 || !enif_get_int64(env, argv[3], &created)) {
+        return enif_make_badarg(env);
+    }
+
+    if((posix_errno = efile_marshal_path(env, argv[0], &path))) {
+        return posix_error_to_tuple(env, posix_errno);
+    } else if((posix_errno = efile_set_time(&path, accessed, modified, created))) {
+        return posix_error_to_tuple(env, posix_errno);
+    }
+
+    return am_ok;
+}
+
+static ERL_NIF_TERM read_link_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    posix_errno_t posix_errno;
+
+    efile_path_t path;
+    ERL_NIF_TERM result;
+
+    if(argc != 1) {
+        return enif_make_badarg(env);
+    }
+
+    if((posix_errno = efile_marshal_path(env, argv[0], &path))) {
+        return posix_error_to_tuple(env, posix_errno);
+    } else if((posix_errno = efile_read_link(env, &path, &result))) {
+        return posix_error_to_tuple(env, posix_errno);
+    }
+
+    return enif_make_tuple2(env, am_ok, result);
+}
+
+static ERL_NIF_TERM list_dir_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    posix_errno_t posix_errno;
+
+    efile_path_t path;
+    ERL_NIF_TERM result;
+
+    if(argc != 1) {
+        return enif_make_badarg(env);
+    }
+
+    if((posix_errno = efile_marshal_path(env, argv[0], &path))) {
+        return posix_error_to_tuple(env, posix_errno);
+    } else if((posix_errno = efile_list_dir(env, &path, &result))) {
+        return posix_error_to_tuple(env, posix_errno);
+    }
+
+    return enif_make_tuple2(env, am_ok, result);
+}
+
+static ERL_NIF_TERM rename_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    posix_errno_t posix_errno;
+
+    efile_path_t existing_path, new_path;
+
+    if(argc != 2) {
+        return enif_make_badarg(env);
+    }
+
+    if((posix_errno = efile_marshal_path(env, argv[0], &existing_path))) {
+        return posix_error_to_tuple(env, posix_errno);
+    } else if((posix_errno = efile_marshal_path(env, argv[1], &new_path))) {
+        return posix_error_to_tuple(env, posix_errno);
+    } else if((posix_errno = efile_rename(&existing_path, &new_path))) {
+        return posix_error_to_tuple(env, posix_errno);
+    }
+
+    return am_ok;
+}
+
+static ERL_NIF_TERM make_hard_link_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    posix_errno_t posix_errno;
+
+    efile_path_t existing_path, new_path;
+
+    if(argc != 2) {
+        return enif_make_badarg(env);
+    }
+
+    if((posix_errno = efile_marshal_path(env, argv[0], &existing_path))) {
+        return posix_error_to_tuple(env, posix_errno);
+    } else if((posix_errno = efile_marshal_path(env, argv[1], &new_path))) {
+        return posix_error_to_tuple(env, posix_errno);
+    } else if((posix_errno = efile_make_hard_link(&existing_path, &new_path))) {
+        return posix_error_to_tuple(env, posix_errno);
+    }
+
+    return am_ok;
+}
+
+static ERL_NIF_TERM make_soft_link_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    posix_errno_t posix_errno;
+
+    efile_path_t existing_path, new_path;
+
+    if(argc != 2) {
+        return enif_make_badarg(env);
+    }
+
+    if((posix_errno = efile_marshal_path(env, argv[0], &existing_path))) {
+        return posix_error_to_tuple(env, posix_errno);
+    } else if((posix_errno = efile_marshal_path(env, argv[1], &new_path))) {
+        return posix_error_to_tuple(env, posix_errno);
+    } else if((posix_errno = efile_make_soft_link(&existing_path, &new_path))) {
+        return posix_error_to_tuple(env, posix_errno);
+    }
+
+    return am_ok;
+}
+
+static ERL_NIF_TERM make_dir_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    posix_errno_t posix_errno;
+
+    efile_path_t path;
+
+    if(argc != 1) {
+        return enif_make_badarg(env);
+    }
+
+    if((posix_errno = efile_marshal_path(env, argv[0], &path))) {
+        return posix_error_to_tuple(env, posix_errno);
+    } else if((posix_errno = efile_make_dir(&path))) {
+        return posix_error_to_tuple(env, posix_errno);
+    }
+
+    return am_ok;
+}
+
+static ERL_NIF_TERM del_file_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    posix_errno_t posix_errno;
+
+    efile_path_t path;
+
+    if(argc != 1) {
+        return enif_make_badarg(env);
+    }
+
+    if((posix_errno = efile_marshal_path(env, argv[0], &path))) {
+        return posix_error_to_tuple(env, posix_errno);
+    } else if((posix_errno = efile_del_file(&path))) {
+        return posix_error_to_tuple(env, posix_errno);
+    }
+
+    return am_ok;
+}
+
+static ERL_NIF_TERM del_dir_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    posix_errno_t posix_errno;
+
+    efile_path_t path;
+
+    if(argc != 1) {
+        return enif_make_badarg(env);
+    }
+
+    if((posix_errno = efile_marshal_path(env, argv[0], &path))) {
+        return posix_error_to_tuple(env, posix_errno);
+    } else if((posix_errno = efile_del_dir(&path))) {
+        return posix_error_to_tuple(env, posix_errno);
+    }
+
+    return am_ok;
+}
+
+static ERL_NIF_TERM get_device_cwd_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    posix_errno_t posix_errno;
+
+    ERL_NIF_TERM result;
+    int device_index;
+
+    if(argc != 1 || !enif_get_int(env, argv[0], &device_index)) {
+        return enif_make_badarg(env);
+    }
+
+    if((posix_errno = efile_get_device_cwd(env, device_index, &result))) {
+        return posix_error_to_tuple(env, posix_errno);
+    }
+
+    return enif_make_tuple2(env, am_ok, result);
+}
+
+static ERL_NIF_TERM get_cwd_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    posix_errno_t posix_errno;
+    ERL_NIF_TERM result;
+
+    if(argc != 0) {
+        return enif_make_badarg(env);
+    }
+
+    if((posix_errno = efile_get_cwd(env, &result))) {
+        return posix_error_to_tuple(env, posix_errno);
+    }
+
+    return enif_make_tuple2(env, am_ok, result);
+}
+
+static ERL_NIF_TERM set_cwd_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    posix_errno_t posix_errno;
+
+    efile_path_t path;
+
+    if(argc != 1) {
+        return enif_make_badarg(env);
+    }
+
+    if((posix_errno = efile_marshal_path(env, argv[0], &path))) {
+        return posix_error_to_tuple(env, posix_errno);
+    } else if((posix_errno = efile_set_cwd(&path))) {
+        return posix_error_to_tuple(env, posix_errno);
+    }
+
+    return am_ok;
+}
+
+/** @brief Reads an entire file into \c result, stopping after \c size bytes or
+ * EOF. It will read until EOF if size is 0. */
+static posix_errno_t read_file(efile_data_t *d, size_t size, ErlNifBinary *result) {
+    size_t initial_buffer_size;
+    ssize_t bytes_read;
+
+    if(size == 0) {
+        initial_buffer_size = 16 << 10;
+    } else {
+        initial_buffer_size = size;
+    }
+
+    if(!enif_alloc_binary(initial_buffer_size, result)) {
+        return ENOMEM;
+    }
+
+    bytes_read = 0;
+
+    for(;;) {
+        ssize_t block_bytes_read;
+        SysIOVec read_vec[1];
+
+        read_vec[0].iov_base = result->data + bytes_read;
+        read_vec[0].iov_len = result->size - bytes_read;
+
+        block_bytes_read = efile_readv(d, read_vec, 1);
+
+        if(block_bytes_read < 0) {
+            enif_release_binary(result);
+            return d->posix_errno;
+        }
+
+        bytes_read += block_bytes_read;
+
+        if(block_bytes_read < (result->size - bytes_read)) {
+            /* EOF */
+            break;
+        } else if(bytes_read == size) {
+            break;
+        }
+
+        if(!enif_realloc_binary(result, bytes_read * 2)) {
+            enif_release_binary(result);
+            return ENOMEM;
+        }
+    }
+
+    /* The file may have shrunk since we queried its size, so we have to do
+     * this even when the size is known. */
+    if(bytes_read < result->size && !enif_realloc_binary(result, bytes_read)) {
+        ERTS_INTERNAL_ERROR("Failed to shrink read_file result.");
+    }
+
+    return 0;
+}
+
+static ERL_NIF_TERM read_file_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    posix_errno_t posix_errno;
+
+    efile_fileinfo_t info = {0};
+    efile_path_t path;
+    efile_data_t *d;
+
+    ErlNifBinary result;
+
+    if(argc != 1) {
+        return enif_make_badarg(env);
+    }
+
+    if((posix_errno = efile_marshal_path(env, argv[0], &path))) {
+        return posix_error_to_tuple(env, posix_errno);
+    } else if((posix_errno = efile_read_info(&path, 1, &info))) {
+        return posix_error_to_tuple(env, posix_errno);
+    } else if((posix_errno = efile_open(&path, EFILE_MODE_READ, efile_resource_type, &d))) {
+        return posix_error_to_tuple(env, posix_errno);
+    }
+
+    posix_errno = read_file(d, info.size, &result);
+    enif_release_resource(d);
+
+    if(posix_errno) {
+        return posix_error_to_tuple(env, posix_errno);
+    }
+
+    return enif_make_tuple2(env, am_ok, enif_make_binary(env, &result));
+}
+
+static ERL_NIF_TERM altname_nif(ErlNifEnv *env, int argc, const ERL_NIF_TERM argv[]) {
+    posix_errno_t posix_errno;
+
+    efile_path_t path;
+    ERL_NIF_TERM result;
+
+    if(argc != 1) {
+        return enif_make_badarg(env);
+    }
+
+    if((posix_errno = efile_marshal_path(env, argv[0], &path))) {
+        return posix_error_to_tuple(env, posix_errno);
+    } else if((posix_errno = efile_altname(env, &path, &result))) {
+        return posix_error_to_tuple(env, posix_errno);
+    }
+
+    return enif_make_tuple2(env, am_ok, result);
+}
diff --git a/erts/emulator/nifs/common/prim_file_nif.h b/erts/emulator/nifs/common/prim_file_nif.h
new file mode 100644
index 0000000000..cc9bc8f5c3
--- /dev/null
+++ b/erts/emulator/nifs/common/prim_file_nif.h
@@ -0,0 +1,240 @@
+/*
+ * %CopyrightBegin%
+ *
+ * Copyright Ericsson 2017. 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%
+ */
+
+typedef int posix_errno_t;
+
+enum efile_modes_t {
+    EFILE_MODE_READ = (1 << 0),
+    EFILE_MODE_WRITE = (1 << 1), /* Implies truncating file when used alone. */
+    EFILE_MODE_APPEND = (1 << 2),
+    EFILE_MODE_EXCLUSIVE = (1 << 3),
+    EFILE_MODE_SYNC = (1 << 4),
+
+    EFILE_MODE_SKIP_TYPE_CHECK = (1 << 5), /* Special for device files on Unix. */
+    EFILE_MODE_NO_TRUNCATE = (1 << 6), /* Special for reopening on VxWorks. */
+
+    EFILE_MODE_READ_WRITE = EFILE_MODE_READ | EFILE_MODE_WRITE
+};
+
+enum efile_access_t {
+    EFILE_ACCESS_NONE = 0,
+    EFILE_ACCESS_READ = 1,
+    EFILE_ACCESS_WRITE = 2,
+    EFILE_ACCESS_READ_WRITE = EFILE_ACCESS_READ | EFILE_ACCESS_WRITE
+};
+
+enum efile_seek_t {
+    EFILE_SEEK_BOF,
+    EFILE_SEEK_CUR,
+    EFILE_SEEK_EOF
+};
+
+enum efile_filetype_t {
+    EFILE_FILETYPE_DEVICE,
+    EFILE_FILETYPE_DIRECTORY,
+    EFILE_FILETYPE_REGULAR,
+    EFILE_FILETYPE_SYMLINK,
+    EFILE_FILETYPE_OTHER
+};
+
+enum efile_advise_t {
+    EFILE_ADVISE_NORMAL,
+    EFILE_ADVISE_RANDOM,
+    EFILE_ADVISE_SEQUENTIAL,
+    EFILE_ADVISE_WILL_NEED,
+    EFILE_ADVISE_DONT_NEED,
+    EFILE_ADVISE_NO_REUSE
+};
+
+enum efile_state_t {
+    EFILE_STATE_IDLE = 0,
+    EFILE_STATE_BUSY = 1,
+    EFILE_STATE_CLOSE_PENDING = 2,
+    EFILE_STATE_CLOSED = 3
+};
+
+typedef struct {
+    Sint64 size;            /* Size of file */
+    Uint32 type;            /* Type of file -- one of EFILE_FILETYPE_*. */
+    Uint32 access;          /* Access to file -- one of EFILE_ACCESS_*. */
+    Uint32 mode;            /* Access permissions -- bit field. */
+    Uint32 links;           /* Number of links to file. */
+    Uint32 major_device;    /* Major device or file system. */
+    Uint32 minor_device;    /* Minor device (for devices). */
+    Uint32 inode;           /* Inode number. */
+    Uint32 uid;             /* User id of owner. */
+    Uint32 gid;             /* Group id of owner. */
+    Sint64 a_time;          /* Last time the file was accessed. */
+    Sint64 m_time;          /* Last time the file was modified. */
+    Sint64 c_time;          /* Windows: creation time, Unix: last inode
+                             * change. */
+} efile_fileinfo_t;
+
+/* The smallest value that can be converted freely between universal, local,
+ * and POSIX time, as required by read_file_info/2. Corresponds to
+ * {{1902,1,1},{0,0,0}} */
+#define EFILE_MIN_FILETIME -2145916800
+
+/* Initializes an efile_data_t; must be used in efile_open on success. */
+#define EFILE_INIT_RESOURCE(__d, __modes) do { \
+        erts_atomic32_init_acqb(&(__d)->state, EFILE_STATE_IDLE); \
+        (__d)->posix_errno = 0; \
+        (__d)->modes = __modes; \
+    } while(0)
+
+typedef struct {
+    erts_atomic32_t state;
+
+    posix_errno_t posix_errno;
+    enum efile_modes_t modes;
+
+    ErlNifMonitor monitor;
+} efile_data_t;
+
+typedef ErlNifBinary efile_path_t;
+
+/* @brief Translates the given "raw name" into the format expected by the APIs
+ * used by the underlying implementation. The result is transient and does not
+ * need to be released.
+ *
+ * This may change the structure of the path and its results should never be
+ * passed on to the user. Refer to the OS-specific implementation for details.
+ *
+ * @param path The term to translate; it must have been encoded with
+ * prim_file:internal_native2name for compatibility reasons. */
+posix_errno_t efile_marshal_path(ErlNifEnv *env, ERL_NIF_TERM path, efile_path_t *result);
+
+/* @brief Returns the underlying handle as an implementation-defined term.
+ *
+ * This is an internal function intended to support tests and tricky
+ * operations like sendfile(2). */
+ERL_NIF_TERM efile_get_handle(ErlNifEnv *env, efile_data_t *d);
+
+/* @brief Read until EOF or the given iovec has been filled.
+ *
+ * @return -1 on failure, or the number of bytes read on success. The return
+ * value will be 0 if no bytes could be read before EOF or the end of the
+ * iovec. */
+Sint64 efile_readv(efile_data_t *d, SysIOVec *iov, int iovlen);
+
+/* @brief Write the entirety of the given iovec.
+ *
+ * @return -1 on failure, or the number of bytes written on success. "Partial"
+ * failures will be reported with -1 and not the number of bytes we managed to
+ * write to disk before the failure. */
+Sint64 efile_writev(efile_data_t *d, SysIOVec *iov, int iovlen);
+
+/* @brief As \c efile_readv, but starting from a file offset. */
+Sint64 efile_preadv(efile_data_t *d, Sint64 offset, SysIOVec *iov, int iovlen);
+
+/* @brief As \c efile_writev, but starting from a file offset. */
+Sint64 efile_pwritev(efile_data_t *d, Sint64 offset, SysIOVec *iov, int iovlen);
+
+int efile_seek(efile_data_t *d, enum efile_seek_t seek, Sint64 offset, Sint64 *new_position);
+
+int efile_sync(efile_data_t *d, int data_only);
+
+int efile_advise(efile_data_t *d, Sint64 offset, Sint64 length, enum efile_advise_t advise);
+int efile_allocate(efile_data_t *d, Sint64 offset, Sint64 length);
+int efile_truncate(efile_data_t *d);
+
+posix_errno_t efile_open(const efile_path_t *path, enum efile_modes_t modes,
+        ErlNifResourceType *nif_type, efile_data_t **d);
+
+/** @brief Closes a file. The file must have entered the CLOSED state prior to
+ * calling this to prevent double close. */
+int efile_close(efile_data_t *d);
+
+/* **** **** **** **** **** **** **** **** **** **** **** **** **** **** **** */
+
+posix_errno_t efile_read_info(const efile_path_t *path, int follow_link, efile_fileinfo_t *result);
+
+/** @brief Sets the file times to the given values. Refer to efile_fileinfo_t
+ * for a description of each. */
+posix_errno_t efile_set_time(const efile_path_t *path, Sint64 a_time, Sint64 m_time, Sint64 c_time);
+
+/** @brief On Unix, this sets the file permissions according to the docs for
+ * file:write_file_info/2. On Windows it uses the "owner write permission" flag
+ * to toggle whether the file is read-only or not. */
+posix_errno_t efile_set_permissions(const efile_path_t *path, Uint32 permissions);
+
+/** @brief On Unix, this will set the owner/group to the given values. It will
+ * do nothing on other platforms. */
+posix_errno_t efile_set_owner(const efile_path_t *path, Uint32 owner, Uint32 group);
+
+/** @brief Resolves the final path of the given link. */
+posix_errno_t efile_read_link(ErlNifEnv *env, const efile_path_t *path, ERL_NIF_TERM *result);
+
+/** @brief Lists the contents of the given directory.
+ * @param result [out] A list of all the directory/file names contained in the
+ * given directory. */
+posix_errno_t efile_list_dir(ErlNifEnv *env, const efile_path_t *path, ERL_NIF_TERM *result);
+
+/** @brief Changes the name of an existing file or directory, from old_path
+ * to new_path.
+ *
+ * If old_path and new_path refer to the same file or directory, it does
+ * nothing and returns success. Otherwise if new_path already exists, it will
+ * be deleted and replaced by src subject to the following conditions:
+ *
+ *     If old_path is a directory, new_path may be an empty directory.
+ *     If old_path is a file, new_path may be a file.
+ *
+ * Neither of these are guaranteed to be atomic. In any other situation where
+ * new_path already exists, the rename will fail.
+ *
+ * Some possible error codes:
+ *
+ * - EACCES:     Either paths or one of their parent directories can't be read
+ *               and/or written.
+ * - EEXIST:     new_path is a non-empty directory.
+ * - EINVAL:     old_path is a root directory or new_path is a subdirectory
+ *               of new_path.
+ * - EISDIR:     new_path is a directory, but old_path is not.
+ * - ENOTDIR:    old_path is a directory, but new_path is not.
+ * - ENOENT:     old_path doesn't exist, or either path is "".
+ * - EXDEV:      The paths are on different filesystems.
+ *
+ * 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. */
+posix_errno_t efile_rename(const efile_path_t *old_path, const efile_path_t *new_path);
+
+posix_errno_t efile_make_hard_link(const efile_path_t *existing_path, const efile_path_t *new_path);
+posix_errno_t efile_make_soft_link(const efile_path_t *existing_path, const efile_path_t *new_path);
+posix_errno_t efile_make_dir(const efile_path_t *path);
+
+posix_errno_t efile_del_file(const efile_path_t *path);
+posix_errno_t efile_del_dir(const efile_path_t *path);
+
+posix_errno_t efile_get_cwd(ErlNifEnv *env, ERL_NIF_TERM *result);
+posix_errno_t efile_set_cwd(const efile_path_t *path);
+
+/** @brief A Windows-specific function for returning the working directory of a
+ * given device.
+ *
+ * @param device_index The drive index; 1 for A, 2 for B, etc.
+ * @param result [out] The working directory of the given device
+ */
+posix_errno_t efile_get_device_cwd(ErlNifEnv *env, int device_index, ERL_NIF_TERM *result);
+
+/** @brief A Windows-specific function for returning the 8.3-name of a given
+ * file or directory. */
+posix_errno_t efile_altname(ErlNifEnv *env, const efile_path_t *path, ERL_NIF_TERM *result);
diff --git a/erts/emulator/nifs/unix/unix_prim_file.c b/erts/emulator/nifs/unix/unix_prim_file.c
new file mode 100644
index 0000000000..57c8ef62e1
--- /dev/null
+++ b/erts/emulator/nifs/unix/unix_prim_file.c
@@ -0,0 +1,957 @@
+/*
+ * %CopyrightBegin%
+ *
+ * Copyright Ericsson 2017. 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 "erl_nif.h"
+#include "config.h"
+#include "sys.h"
+
+#ifdef VALGRIND
+#  include <valgrind/memcheck.h>
+#endif
+
+#include "prim_file_nif.h"
+
+#if defined(__APPLE__) && defined(__MACH__) && !defined(__DARWIN__)
+#define __DARWIN__ 1
+#endif
+
+#if defined(__DARWIN__) || defined(HAVE_LINUX_FALLOC_H) || defined(HAVE_POSIX_FALLOCATE)
+#include <fcntl.h>
+#endif
+
+#ifdef HAVE_LINUX_FALLOC_H
+#include <linux/falloc.h>
+#endif
+
+#include <utime.h>
+
+/* Macros for testing file types. */
+#ifdef NO_UMASK
+#define FILE_MODE 0644
+#define DIR_MODE  0755
+#else
+#define FILE_MODE 0666
+#define DIR_MODE  0777
+#endif
+
+/* Old platforms might not have IOV_MAX defined. */
+#if !defined(IOV_MAX) && defined(UIO_MAXIOV)
+#define IOV_MAX UIO_MAXIOV
+#elif !defined(IOV_MAX)
+#define IOV_MAX 16
+#endif
+
+typedef struct {
+    efile_data_t common;
+    int fd;
+} efile_unix_t;
+
+static int has_invalid_null_termination(const ErlNifBinary *path) {
+    const char *null_pos, *end_pos;
+
+    null_pos = memchr(path->data, '\0', path->size);
+    end_pos = (const char*)&path->data[path->size] - 1;
+
+    if(null_pos == NULL) {
+        return 1;
+    }
+
+    /* prim_file:internal_name2native sometimes feeds us data that is "doubly"
+     * NUL-terminated, so we'll accept any number of trailing NULs so long as
+     * they aren't interrupted by anything else. */
+    while(null_pos < end_pos && (*null_pos) == '\0') {
+        null_pos++;
+    }
+
+    return null_pos != end_pos;
+}
+
+posix_errno_t efile_marshal_path(ErlNifEnv *env, ERL_NIF_TERM path, efile_path_t *result) {
+    if(!enif_inspect_binary(env, path, result)) {
+        return EINVAL;
+    }
+
+    if(has_invalid_null_termination(result)) {
+        return EINVAL;
+    }
+
+    return 0;
+}
+
+ERL_NIF_TERM efile_get_handle(ErlNifEnv *env, efile_data_t *d) {
+    efile_unix_t *u = (efile_unix_t*)d;
+
+    ERL_NIF_TERM result;
+    unsigned char *bits;
+
+    bits = enif_make_new_binary(env, sizeof(u->fd), &result);
+    memcpy(bits, &u->fd, sizeof(u->fd));
+
+    return result;
+}
+
+static int open_file_type_check(const efile_path_t *path, int fd) {
+    struct stat file_info;
+    int error;
+
+#ifndef HAVE_FSTAT
+    error = stat((const char*)path->data, &file_info);
+    (void)fd;
+#else
+    error = fstat(fd, &file_info);
+    (void)path;
+#endif
+
+    if(error < 0) {
+        /* If we failed to stat assume success and let the next call handle the
+         * error. The old driver checked whether the file was to be used
+         * immediately in a read within the call, but the new implementation
+         * never does that. */
+         return 1;
+    } else {
+        /* The old driver tolerated opening /dev/null despite the "no devices"
+         * limitation. It provided no explanation for this but we still need
+         * to match the behavior. We're checking through stat(2) instead of
+         * comparing the name to account for links. */
+        struct stat null_device_info;
+        int is_dev_null;
+
+        is_dev_null = (stat("/dev/null", &null_device_info) == 0);
+        is_dev_null &= (file_info.st_ino == null_device_info.st_ino);
+        is_dev_null &= (file_info.st_dev == null_device_info.st_dev);
+
+        if(is_dev_null) {
+            return 1;
+        }
+    }
+
+    if(!S_ISREG(file_info.st_mode)) {
+        return 0;
+    }
+
+    return 1;
+}
+
+posix_errno_t efile_open(const efile_path_t *path, enum efile_modes_t modes,
+        ErlNifResourceType *nif_type, efile_data_t **d) {
+
+    int flags, fd;
+
+    flags = 0;
+
+    if(modes & EFILE_MODE_READ && !(modes & EFILE_MODE_WRITE)) {
+        flags |= O_RDONLY;
+    } else if(modes & EFILE_MODE_WRITE && !(modes & EFILE_MODE_READ)) {
+        if(!(modes & EFILE_MODE_NO_TRUNCATE)) {
+            flags |= O_TRUNC;
+        }
+
+        flags |= O_WRONLY | O_CREAT;
+    } else if(modes & EFILE_MODE_READ_WRITE) {
+        flags |= O_RDWR | O_CREAT;
+    } else {
+        return EINVAL;
+    }
+
+    if(modes & EFILE_MODE_APPEND) {
+        flags &= ~O_TRUNC;
+        flags |= O_APPEND;
+    }
+
+    if(modes & EFILE_MODE_EXCLUSIVE) {
+        flags |= O_EXCL;
+    }
+
+    if(modes & EFILE_MODE_SYNC) {
+#ifndef O_SYNC
+        return ENOTSUP;
+#else
+        flags |= O_SYNC;
+#endif
+    }
+
+    do {
+        fd = open((const char*)path->data, flags, FILE_MODE);
+    } while(fd == -1 && errno == EINTR);
+
+    if(fd != -1) {
+        efile_unix_t *u;
+
+        if(!(modes & EFILE_MODE_SKIP_TYPE_CHECK) && !open_file_type_check(path, fd)) {
+            close(fd);
+
+            /* This is blatantly incorrect, but we're documented as returning
+             * this for everything that isn't a file. */
+            return EISDIR;
+        }
+
+        u = (efile_unix_t*)enif_alloc_resource(nif_type, sizeof(efile_unix_t));
+        u->fd = fd;
+
+        EFILE_INIT_RESOURCE(&u->common, modes);
+        (*d) = &u->common;
+
+        return 0;
+    }
+
+    (*d) = NULL;
+    return errno;
+}
+
+int efile_close(efile_data_t *d) {
+    efile_unix_t *u = (efile_unix_t*)d;
+    int fd;
+
+    ASSERT(erts_atomic32_read_nob(&d->state) == EFILE_STATE_CLOSED);
+    ASSERT(u->fd != -1);
+
+    fd = u->fd;
+    u->fd = -1;
+
+    /* close(2) either always closes (*BSD, Linux) or leaves the fd in an
+     * undefined state (POSIX 2008, Solaris), so we must not retry on EINTR. */
+
+    if(close(fd) < 0) {
+        u->common.posix_errno = errno;
+        return 0;
+    }
+
+    return 1;
+}
+
+static void shift_iov(SysIOVec **iov, int *iovlen, ssize_t shift) {
+    SysIOVec *head_vec = (*iov);
+
+    ASSERT(shift >= 0);
+
+    while(shift > 0) {
+        ASSERT(head_vec < &(*iov)[*iovlen]);
+
+        if(shift < head_vec->iov_len) {
+            head_vec->iov_base = (char*)head_vec->iov_base + shift;
+            head_vec->iov_len -= shift;
+            break;
+        } else {
+            shift -= head_vec->iov_len;
+            head_vec++;
+        }
+    }
+
+    (*iovlen) -= head_vec - (*iov);
+    (*iov) = head_vec;
+}
+
+Sint64 efile_readv(efile_data_t *d, SysIOVec *iov, int iovlen) {
+    efile_unix_t *u = (efile_unix_t*)d;
+
+    Sint64 bytes_read;
+    ssize_t result;
+
+    bytes_read = 0;
+
+    do {
+        int use_fallback = 0;
+
+        if(iovlen < 1) {
+            result = 0;
+            break;
+        }
+
+        /* writev(2) implies readv(2) */
+#ifdef HAVE_WRITEV
+        result = readv(u->fd, iov, MIN(IOV_MAX, iovlen));
+
+        /* Fall back to using read(2) if readv(2) reports that the combined
+         * size of iov is greater than SSIZE_T_MAX. */
+        use_fallback = (result < 0 && errno == EINVAL);
+#else
+        use_fallback = 1;
+#endif
+
+        if(use_fallback) {
+            result = read(u->fd, iov->iov_base, iov->iov_len);
+        }
+
+        if(result > 0) {
+            shift_iov(&iov, &iovlen, result);
+            bytes_read += result;
+        }
+    } while(result > 0 || (result < 0 && errno == EINTR));
+
+    u->common.posix_errno = errno;
+
+    if(result == 0 && bytes_read > 0) {
+        return bytes_read;
+    }
+
+    return result;
+}
+
+Sint64 efile_writev(efile_data_t *d, SysIOVec *iov, int iovlen) {
+    efile_unix_t *u = (efile_unix_t*)d;
+
+    Sint64 bytes_written;
+    ssize_t result;
+
+    bytes_written = 0;
+
+    do {
+        int use_fallback = 0;
+
+        if(iovlen < 1) {
+            result = 0;
+            break;
+        }
+
+#ifdef HAVE_WRITEV
+        result = writev(u->fd, iov, MIN(IOV_MAX, iovlen));
+
+        /* Fall back to using write(2) if writev(2) reports that the combined
+         * size of iov is greater than SSIZE_T_MAX. */
+        use_fallback = (result < 0 && errno == EINVAL);
+#else
+        use_fallback = 1;
+#endif
+
+        if(use_fallback) {
+            result = write(u->fd, iov->iov_base, iov->iov_len);
+        }
+
+        if(result > 0) {
+            shift_iov(&iov, &iovlen, result);
+            bytes_written += result;
+        }
+    } while(result > 0 || (result < 0 && errno == EINTR));
+
+    u->common.posix_errno = errno;
+
+    if(result == 0 && bytes_written > 0) {
+        return bytes_written;
+    }
+
+    return result;
+}
+
+Sint64 efile_preadv(efile_data_t *d, Sint64 offset, SysIOVec *iov, int iovlen) {
+    efile_unix_t *u = (efile_unix_t*)d;
+
+    Uint64 bytes_read;
+    Sint64 result;
+
+#if !defined(HAVE_PREADV) && !defined(HAVE_PREAD)
+    /* This function is documented as leaving the file position undefined, but
+     * the old driver always reset it so there's probably code in the wild that
+     * relies on this behavior. */
+    off_t original_position = lseek(u->fd, 0, SEEK_CUR);
+
+    if(original_position < 0 || lseek(u->fd, offset, SEEK_SET) < 0) {
+        u->common.posix_errno = errno;
+        return -1;
+    }
+#endif
+
+    bytes_read = 0;
+
+    do {
+        if(iovlen < 1) {
+            result = 0;
+            break;
+        }
+
+#if defined(HAVE_PREADV)
+        result = preadv(u->fd, iov, MIN(IOV_MAX, iovlen), offset);
+#elif defined(HAVE_PREAD)
+        result = pread(u->fd, iov->iov_base, iov->iov_len, offset);
+#else
+        result = read(u->fd, iov->iov_base, iov->iov_len);
+#endif
+
+        if(result > 0) {
+            shift_iov(&iov, &iovlen, result);
+            bytes_read += result;
+            offset += result;
+        }
+    } while(result > 0 || (result < 0 && errno == EINTR));
+
+    u->common.posix_errno = errno;
+
+#if !defined(HAVE_PREADV) && !defined(HAVE_PREAD)
+    if(result >= 0) {
+        if(lseek(u->fd, original_position, SEEK_SET) < 0) {
+            u->common.posix_errno = errno;
+            return -1;
+        }
+    }
+#endif
+
+    if(result == 0 && bytes_read > 0) {
+        return bytes_read;
+    }
+
+    return result;
+}
+
+Sint64 efile_pwritev(efile_data_t *d, Sint64 offset, SysIOVec *iov, int iovlen) {
+    efile_unix_t *u = (efile_unix_t*)d;
+
+    Sint64 bytes_written;
+    ssize_t result;
+
+#if !defined(HAVE_PWRITEV) && !defined(HAVE_PWRITE)
+    off_t original_position = lseek(u->fd, 0, SEEK_CUR);
+
+    if(original_position < 0 || lseek(u->fd, offset, SEEK_SET) < 0) {
+        u->common.posix_errno = errno;
+        return -1;
+    }
+#endif
+
+    bytes_written = 0;
+
+    do {
+        if(iovlen < 1) {
+            result = 0;
+            break;
+        }
+
+#if defined(HAVE_PWRITEV)
+        result = pwritev(u->fd, iov, MIN(IOV_MAX, iovlen), offset);
+#elif defined(HAVE_PWRITE)
+        result = pwrite(u->fd, iov->iov_base, iov->iov_len, offset);
+#else
+        result = write(u->fd, iov->iov_base, iov->iov_len);
+#endif
+
+        if(result > 0) {
+            shift_iov(&iov, &iovlen, result);
+            bytes_written += result;
+            offset += result;
+        }
+    } while(result > 0 || (result < 0 && errno == EINTR));
+
+    u->common.posix_errno = errno;
+
+#if !defined(HAVE_PWRITEV) && !defined(HAVE_PWRITE)
+    if(result >= 0) {
+        if(lseek(u->fd, original_position, SEEK_SET) < 0) {
+            u->common.posix_errno = errno;
+            return -1;
+        }
+    }
+#endif
+
+    if(result == 0 && bytes_written > 0) {
+        return bytes_written;
+    }
+
+    return result;
+}
+
+int efile_seek(efile_data_t *d, enum efile_seek_t seek, Sint64 offset, Sint64 *new_position) {
+    efile_unix_t *u = (efile_unix_t*)d;
+    off_t result;
+    int whence;
+
+    switch(seek) {
+        case EFILE_SEEK_BOF: whence = SEEK_SET; break;
+        case EFILE_SEEK_CUR: whence = SEEK_CUR; break;
+        case EFILE_SEEK_EOF: whence = SEEK_END; break;
+        default: ERTS_INTERNAL_ERROR("Invalid seek parameter");
+    }
+
+    result = lseek(u->fd, offset, whence);
+
+    /*
+     * The man page for lseek (on SunOs 5) says:
+     *
+     * "if fildes is a remote file descriptor and offset is negative, lseek()
+     * returns the file pointer even if it is negative."
+     */
+    if(result < 0 && errno == 0) {
+        errno = EINVAL;
+    }
+
+    if(result < 0) {
+        u->common.posix_errno = errno;
+        return 0;
+    }
+
+    (*new_position) = result;
+
+    return 1;
+}
+
+int efile_sync(efile_data_t *d, int data_only) {
+    efile_unix_t *u = (efile_unix_t*)d;
+
+#if defined(HAVE_FDATASYNC) && !defined(__DARWIN__)
+    if(data_only) {
+        if(fdatasync(u->fd) < 0) {
+            u->common.posix_errno = errno;
+            return 0;
+        }
+
+        return 1;
+    }
+#endif
+
+#if defined(__DARWIN__) && defined(F_FULLFSYNC)
+    if(fcntl(u->fd, F_FULLFSYNC) < 0) {
+#else
+    if(fsync(u->fd) < 0) {
+#endif
+        u->common.posix_errno = errno;
+        return 0;
+    }
+
+    return 1;
+}
+
+int efile_advise(efile_data_t *d, Sint64 offset, Sint64 length, enum efile_advise_t advise) {
+    efile_unix_t *u = (efile_unix_t*)d;
+#ifdef HAVE_POSIX_FADVISE
+    int p_advise;
+
+    switch(advise) {
+        case EFILE_ADVISE_NORMAL: p_advise = POSIX_FADV_NORMAL; break;
+        case EFILE_ADVISE_RANDOM: p_advise = POSIX_FADV_RANDOM; break;
+        case EFILE_ADVISE_SEQUENTIAL: p_advise = POSIX_FADV_SEQUENTIAL; break;
+        case EFILE_ADVISE_WILL_NEED: p_advise = POSIX_FADV_WILLNEED; break;
+        case EFILE_ADVISE_DONT_NEED: p_advise = POSIX_FADV_DONTNEED; break;
+        case EFILE_ADVISE_NO_REUSE: p_advise = POSIX_FADV_NOREUSE; break;
+        default:
+            u->common.posix_errno = EINVAL;
+            return 0;
+    }
+
+    if(posix_fadvise(u->fd, offset, length, p_advise) < 0) {
+        u->common.posix_errno = errno;
+        return 0;
+    }
+
+    return 1;
+#else
+    /* We'll pretend to support this syscall as it's only a recommendation even
+     * on systems that do support it. */
+    return 1;
+#endif
+}
+
+int efile_allocate(efile_data_t *d, Sint64 offset, Sint64 length) {
+    efile_unix_t *u = (efile_unix_t*)d;
+    int ret = -1;
+
+    /* We prefer OS-specific methods, but fall back to posix_fallocate on
+     * failure. It's unclear whether this has any practical benefit on
+     * modern systems, but the old driver did it. */
+
+#if defined(HAVE_FALLOCATE)
+    /* Linux-specific */
+    do {
+        ret = fallocate(u->fd, FALLOC_FL_KEEP_SIZE, offset, length);
+    } while(ret < 0 && errno == EINTR);
+#elif defined(F_PREALLOCATE)
+    /* Mac-specific */
+    fstore_t fs = {};
+
+    fs.fst_flags = F_ALLOCATECONTIG;
+    fs.fst_posmode = F_VOLPOSMODE;
+    fs.fst_offset = offset;
+    fs.fst_length = length;
+
+    ret = fcntl(u->fd, F_PREALLOCATE, &fs);
+    if(ret < 0) {
+        fs.fst_flags = F_ALLOCATEALL;
+        ret = fcntl(u->fd, F_PREALLOCATE, &fs);
+    }
+#elif !defined(HAVE_POSIX_FALLOCATE)
+    u->common.posix_errno = ENOTSUP;
+    return 0;
+#endif
+
+#ifdef HAVE_POSIX_FALLOCATE
+    if(ret < 0) {
+        do {
+            ret = posix_fallocate(u->fd, offset, length);
+
+            /* 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. */
+            if (ret > 0) {
+                errno = ret;
+                ret = -1;
+            }
+        } while(ret < 0 && errno == EINTR);
+    }
+#endif
+
+    if(ret < 0) {
+        u->common.posix_errno = errno;
+        return 0;
+    }
+
+    return 1;
+}
+
+int efile_truncate(efile_data_t *d) {
+    efile_unix_t *u = (efile_unix_t*)d;
+    off_t offset;
+
+    offset = lseek(u->fd, 0, SEEK_CUR);
+
+    if(offset < 0) {
+        u->common.posix_errno = errno;
+        return 0;
+    }
+
+    if(ftruncate(u->fd, offset) < 0) {
+        u->common.posix_errno = errno;
+        return 0;
+    }
+
+    return 1;
+}
+
+posix_errno_t efile_read_info(const efile_path_t *path, int follow_links, efile_fileinfo_t *result) {
+    struct stat data;
+
+    if(follow_links) {
+        if(stat((const char*)path->data, &data) < 0) {
+            return errno;
+        }
+    } else {
+        if(lstat((const char*)path->data, &data) < 0) {
+            return errno;
+        }
+    }
+
+    if(S_ISCHR(data.st_mode) || S_ISBLK(data.st_mode)) {
+        result->type = EFILE_FILETYPE_DEVICE;
+    } else if(S_ISDIR(data.st_mode)) {
+        result->type = EFILE_FILETYPE_DIRECTORY;
+    } else if(S_ISREG(data.st_mode)) {
+        result->type = EFILE_FILETYPE_REGULAR;
+    } else if(S_ISLNK(data.st_mode)) {
+        result->type = EFILE_FILETYPE_SYMLINK;
+    } else {
+        result->type = EFILE_FILETYPE_OTHER;
+    }
+
+    result->a_time = (Sint64)data.st_atime;
+    result->m_time = (Sint64)data.st_mtime;
+    result->c_time = (Sint64)data.st_ctime;
+    result->size = data.st_size;
+
+    result->major_device = data.st_dev;
+    result->minor_device = data.st_rdev;
+    result->links = data.st_nlink;
+    result->inode = data.st_ino;
+    result->mode = data.st_mode;
+    result->uid = data.st_uid;
+    result->gid = data.st_gid;
+
+#ifndef NO_ACCESS
+    result->access = EFILE_ACCESS_NONE;
+
+    if(access((const char*)path->data, R_OK) == 0) {
+        result->access |= EFILE_ACCESS_READ;
+    }
+    if(access((const char*)path->data, W_OK) == 0) {
+        result->access |= EFILE_ACCESS_WRITE;
+    }
+#else
+    /* Just look at read/write access for owner. */
+    result->access = ((data.st_mode >> 6) & 07) >> 1;
+#endif
+
+    return 0;
+}
+
+posix_errno_t efile_set_permissions(const efile_path_t *path, Uint32 permissions) {
+    const mode_t MUTABLE_MODES = (S_ISUID | S_ISGID | S_IRWXU | S_IRWXG | S_IRWXO);
+    mode_t new_modes = permissions & MUTABLE_MODES;
+
+    if(chmod((const char*)path->data, new_modes) < 0) {
+        new_modes &= ~(S_ISUID | S_ISGID);
+
+        if (chmod((const char*)path->data, new_modes) < 0) {
+            return errno;
+        }
+    }
+
+    return 0;
+}
+
+posix_errno_t efile_set_owner(const efile_path_t *path, Uint32 owner, Uint32 group) {
+    if(chown((const char*)path->data, owner, group) < 0) {
+        return errno;
+    }
+
+    return 0;
+}
+
+posix_errno_t efile_set_time(const efile_path_t *path, Sint64 a_time, Sint64 m_time, Sint64 c_time) {
+    struct utimbuf tval;
+
+    tval.actime = (time_t)a_time;
+    tval.modtime = (time_t)m_time;
+
+    (void)c_time;
+
+    if(utime((const char*)path->data, &tval) < 0) {
+        return errno;
+    }
+
+    return 0;
+}
+
+posix_errno_t efile_read_link(ErlNifEnv *env, const efile_path_t *path, ERL_NIF_TERM *result) {
+    ErlNifBinary result_bin;
+
+    if(!enif_alloc_binary(256, &result_bin)) {
+        return ENOMEM;
+    }
+
+    for(;;) {
+        ssize_t bytes_copied;
+
+        bytes_copied = readlink((const char*)path->data, (char*)result_bin.data,
+            result_bin.size);
+
+        if(bytes_copied <= 0) {
+            posix_errno_t saved_errno = errno;
+            enif_release_binary(&result_bin);
+            return saved_errno;
+        } else if(bytes_copied < result_bin.size) {
+            if(!enif_realloc_binary(&result_bin, bytes_copied)) {
+                enif_release_binary(&result_bin);
+                return ENOMEM;
+            }
+
+            (*result) = enif_make_binary(env, &result_bin);
+
+            return 0;
+        }
+
+        /* The result didn't fit into the buffer, so we'll try again with a
+         * larger one. */
+
+        if(!enif_realloc_binary(&result_bin, result_bin.size * 2)) {
+            enif_release_binary(&result_bin);
+            return ENOMEM;
+        }
+    }
+}
+
+static int is_ignored_name(int name_length, const char *name) {
+    if(name_length == 1 && name[0] == '.') {
+        return 1;
+    } else if(name_length == 2 && memcmp(name, "..", 2) == 0) {
+        return 1;
+    }
+
+    return 0;
+}
+
+posix_errno_t efile_list_dir(ErlNifEnv *env, const efile_path_t *path, ERL_NIF_TERM *result) {
+    ERL_NIF_TERM list_head;
+
+    struct dirent *dir_entry;
+    DIR *dir_stream;
+
+    dir_stream = opendir((const char*)path->data);
+    if(dir_stream == NULL) {
+        posix_errno_t saved_errno = errno;
+        *result = enif_make_list(env, 0);
+        return saved_errno;
+    }
+
+    list_head = enif_make_list(env, 0);
+    dir_entry = readdir(dir_stream);
+
+    while(dir_entry != NULL) {
+        int name_length = strlen(dir_entry->d_name);
+
+        if(!is_ignored_name(name_length, dir_entry->d_name)) {
+            unsigned char *name_bytes;
+            ERL_NIF_TERM name_term;
+
+            name_bytes = enif_make_new_binary(env, name_length, &name_term);
+            sys_memcpy(name_bytes, dir_entry->d_name, name_length);
+
+            list_head = enif_make_list_cell(env, name_term, list_head);
+        }
+
+        dir_entry = readdir(dir_stream);
+    }
+
+    (*result) = list_head;
+    closedir(dir_stream);
+
+    return 0;
+}
+
+posix_errno_t efile_rename(const efile_path_t *old_path, const efile_path_t *new_path) {
+    if(rename((const char*)old_path->data, (const char*)new_path->data) < 0) {
+        if(errno == ENOTEMPTY) {
+            return EEXIST;
+        }
+
+        if(strcmp((const char*)old_path->data, "/") == 0) {
+            /* Alpha reports renaming / as EBUSY and Linux reports it as EACCES
+             * instead of EINVAL.*/
+             return EINVAL;
+        }
+
+        return errno;
+    }
+
+    return 0;
+}
+
+posix_errno_t efile_make_hard_link(const efile_path_t *existing_path, const efile_path_t *new_path) {
+    if(link((const char*)existing_path->data, (const char*)new_path->data) < 0) {
+        return errno;
+    }
+
+    return 0;
+}
+
+posix_errno_t efile_make_soft_link(const efile_path_t *existing_path, const efile_path_t *new_path) {
+    if(symlink((const char*)existing_path->data, (const char*)new_path->data) < 0) {
+        return errno;
+    }
+
+    return 0;
+}
+
+posix_errno_t efile_make_dir(const efile_path_t *path) {
+#ifdef NO_MKDIR_MODE
+    if(mkdir((const char*)path->data) < 0) {
+#else
+    if(mkdir((const char*)path->data, DIR_MODE) < 0) {
+#endif
+        return errno;
+    }
+
+    return 0;
+}
+
+posix_errno_t efile_del_file(const efile_path_t *path) {
+    if(unlink((const char*)path->data) < 0) {
+        /* Linux sets the wrong error code. */
+        if(errno == EISDIR) {
+            return EPERM;
+        }
+
+        return errno;
+    }
+
+    return 0;
+}
+
+posix_errno_t efile_del_dir(const efile_path_t *path) {
+    if(rmdir((const char*)path->data) < 0) {
+        posix_errno_t saved_errno = errno;
+
+        if(saved_errno == ENOTEMPTY) {
+            saved_errno = EEXIST;
+        }
+
+        /* The error code might be wrong if we're trying to delete the current
+         * directory. */
+        if(saved_errno == EEXIST) {
+            struct stat path_stat, cwd_stat;
+            int has_stat;
+
+            has_stat = (stat((const char*)path->data, &path_stat) == 0);
+            has_stat &= (stat(".", &cwd_stat) == 0);
+
+            if(has_stat && path_stat.st_ino == cwd_stat.st_ino) {
+                if(path_stat.st_dev == cwd_stat.st_dev) {
+                    return EINVAL;
+                }
+            }
+        }
+
+        return saved_errno;
+    }
+
+    return 0;
+}
+
+posix_errno_t efile_set_cwd(const efile_path_t *path) {
+    if(chdir((const char*)path->data) < 0) {
+        return errno;
+    }
+
+    return 0;
+}
+
+posix_errno_t efile_get_device_cwd(ErlNifEnv *env, int device_index, ERL_NIF_TERM *result) {
+    (void)device_index;
+    (void)result;
+    (void)env;
+
+    return ENOTSUP;
+}
+
+posix_errno_t efile_get_cwd(ErlNifEnv *env, ERL_NIF_TERM *result) {
+    ErlNifBinary result_bin;
+    size_t bytes_copied;
+
+    if(!enif_alloc_binary(256, &result_bin)) {
+        return ENOMEM;
+    }
+
+    while(getcwd((char*)result_bin.data, result_bin.size) == NULL) {
+        posix_errno_t saved_errno = errno;
+
+        if(saved_errno != ERANGE) {
+            enif_release_binary(&result_bin);
+            return saved_errno;
+        } else {
+            if(!enif_realloc_binary(&result_bin, result_bin.size * 2)) {
+                enif_release_binary(&result_bin);
+                return ENOMEM;
+            }
+        }
+    }
+
+    /* getcwd(2) guarantees null-termination. */
+    bytes_copied = strlen((const char*)result_bin.data);
+
+    if(!enif_realloc_binary(&result_bin, bytes_copied)) {
+        enif_release_binary(&result_bin);
+        return ENOMEM;
+    }
+
+    (*result) = enif_make_binary(env, &result_bin);
+
+    return 0;
+}
+
+posix_errno_t efile_altname(ErlNifEnv *env, const efile_path_t *path, ERL_NIF_TERM *result) {
+    (void)path;
+    (void)result;
+
+    return ENOTSUP;
+}
diff --git a/erts/emulator/nifs/win32/win_prim_file.c b/erts/emulator/nifs/win32/win_prim_file.c
new file mode 100644
index 0000000000..9f993f1d24
--- /dev/null
+++ b/erts/emulator/nifs/win32/win_prim_file.c
@@ -0,0 +1,1427 @@
+/*
+ * %CopyrightBegin%
+ *
+ * Copyright Ericsson 2017. 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 "erl_nif.h"
+#include "config.h"
+#include "sys.h"
+
+#include "prim_file_nif.h"
+
+#include <windows.h>
+#include <strsafe.h>
+#include <wchar.h>
+
+#define IS_SLASH(a)  ((a) == L'\\' || (a) == L'/')
+
+#define FILE_SHARE_FLAGS (FILE_SHARE_READ | FILE_SHARE_WRITE | FILE_SHARE_DELETE)
+
+#define LP_PREFIX L"\\\\?\\"
+#define LP_PREFIX_SIZE (sizeof(LP_PREFIX) - sizeof(WCHAR))
+#define LP_PREFIX_LENGTH (LP_PREFIX_SIZE / sizeof(WCHAR))
+
+#define PATH_LENGTH(path) (path->size / sizeof(WCHAR) - 1)
+
+#define ASSERT_PATH_FORMAT(path) \
+    do { \
+        ASSERT(PATH_LENGTH(path) >= 4 && \
+            !memcmp(path->data, LP_PREFIX, LP_PREFIX_SIZE)); \
+        ASSERT(PATH_LENGTH(path) == wcslen((WCHAR*)path->data)); \
+    } while(0)
+
+#define TICKS_PER_SECOND (10000000ULL)
+#define EPOCH_DIFFERENCE (11644473600LL)
+
+#define FILETIME_TO_EPOCH(epoch, ft) \
+    do { \
+        ULARGE_INTEGER ull; \
+        ull.LowPart  = (ft).dwLowDateTime; \
+        ull.HighPart = (ft).dwHighDateTime; \
+        (epoch) = ((ull.QuadPart / TICKS_PER_SECOND) - EPOCH_DIFFERENCE); \
+    } while(0)
+
+#define EPOCH_TO_FILETIME(ft, epoch) \
+    do { \
+        ULARGE_INTEGER ull; \
+        ull.QuadPart = (((epoch) + EPOCH_DIFFERENCE) * TICKS_PER_SECOND); \
+        (ft).dwLowDateTime = ull.LowPart; \
+        (ft).dwHighDateTime = ull.HighPart; \
+    } while(0)
+
+typedef struct {
+    efile_data_t common;
+    HANDLE handle;
+} efile_win_t;
+
+static int windows_to_posix_errno(DWORD last_error);
+
+static int has_invalid_null_termination(const ErlNifBinary *path) {
+    const WCHAR *null_pos, *end_pos;
+
+    null_pos = wmemchr((const WCHAR*)path->data, L'\0', path->size);
+    end_pos = (const WCHAR*)&path->data[path->size] - 1;
+
+    if(null_pos == NULL) {
+        return 1;
+    }
+
+    /* prim_file:internal_name2native sometimes feeds us data that is "doubly"
+     * NUL-terminated, so we'll accept any number of trailing NULs so long as
+     * they aren't interrupted by anything else. */
+    while(null_pos < end_pos && (*null_pos) == L'\0') {
+        null_pos++;
+    }
+
+    return null_pos != end_pos;
+}
+
+static posix_errno_t get_full_path(ErlNifEnv *env, WCHAR *input, efile_path_t *result) {
+    DWORD maximum_length, actual_length;
+    int add_long_prefix;
+
+    maximum_length = GetFullPathNameW(input, 0, NULL, NULL);
+    add_long_prefix = 0;
+
+    if(maximum_length == 0) {
+        /* POSIX doesn't have the concept of a "path error" in the same way
+         * Windows does, so we'll return ENOENT since that's what most POSIX
+         * APIs would return if they were fed such garbage. */
+        return ENOENT;
+    }
+
+    maximum_length += LP_PREFIX_LENGTH;
+
+    if(!enif_alloc_binary(maximum_length * sizeof(WCHAR), result)) {
+        return ENOMEM;
+    }
+
+    actual_length = GetFullPathNameW(input, maximum_length, (WCHAR*)result->data, NULL);
+
+    if(actual_length < maximum_length) {
+        int has_long_path_prefix;
+        WCHAR *path_start;
+
+        /* Make sure we have a long-path prefix; GetFullPathNameW only adds one
+         * if the path is relative. */
+        has_long_path_prefix = actual_length >= LP_PREFIX_LENGTH &&
+            !sys_memcmp(result->data, LP_PREFIX, LP_PREFIX_SIZE);
+
+        if(!has_long_path_prefix) {
+            sys_memmove(result->data + LP_PREFIX_SIZE, result->data,
+                (actual_length + 1) * sizeof(WCHAR));
+            sys_memcpy(result->data, LP_PREFIX, LP_PREFIX_SIZE);
+            actual_length += LP_PREFIX_LENGTH;
+        }
+
+        path_start = (WCHAR*)result->data;
+
+        /* We're removing trailing slashes since quite a few APIs refuse to
+         * work with them, and none require them. We only check the last
+         * character since GetFullPathNameW folds slashes together. */
+        if(IS_SLASH(path_start[actual_length - 1])) {
+            if(path_start[actual_length - 2] != L':') {
+                path_start[actual_length - 1] = L'\0';
+                actual_length--;
+            }
+        }
+
+        if(!enif_realloc_binary(result, (actual_length + 1) * sizeof(WCHAR))) {
+            enif_release_binary(result);
+            return ENOMEM;
+        }
+
+        enif_make_binary(env, result);
+        return 0;
+    }
+
+    /* We may end up here if the current directory changes to something longer
+     * between/during GetFullPathName. There's nothing sensible we can do about
+     * this. */
+
+    enif_release_binary(result);
+
+    return EINVAL;
+}
+
+posix_errno_t efile_marshal_path(ErlNifEnv *env, ERL_NIF_TERM path, efile_path_t *result) {
+    ErlNifBinary raw_path;
+
+    if(!enif_inspect_binary(env, path, &raw_path)) {
+        return EINVAL;
+    } else if(raw_path.size % sizeof(WCHAR)) {
+        return EINVAL;
+    }
+
+    if(has_invalid_null_termination(&raw_path)) {
+        return EINVAL;
+    }
+
+    return get_full_path(env, (WCHAR*)raw_path.data, result);
+}
+
+ERL_NIF_TERM efile_get_handle(ErlNifEnv *env, efile_data_t *d) {
+    efile_win_t *w = (efile_win_t*)d;
+
+    ERL_NIF_TERM result;
+    unsigned char *bits;
+
+    bits = enif_make_new_binary(env, sizeof(w->handle), &result);
+    memcpy(bits, &w->handle, sizeof(w->handle));
+
+    return result;
+}
+
+/** @brief Converts a native path to the preferred form in "erlang space,"
+ * without path-prefixes, forward-slashes, or NUL terminators. */
+static int normalize_path_result(ErlNifBinary *path) {
+    WCHAR *path_iterator, *path_start, *path_end;
+    int length;
+
+    path_start = (WCHAR*)path->data;
+    length = wcslen(path_start);
+
+    ASSERT(length < path->size / sizeof(WCHAR));
+
+    /* Get rid of the long-path prefix, if present. */
+    if(length >= LP_PREFIX_LENGTH) {
+        if(!sys_memcmp(path_start, LP_PREFIX, LP_PREFIX_SIZE)) {
+            length -= LP_PREFIX_LENGTH;
+
+            sys_memmove(path_start, &path_start[LP_PREFIX_LENGTH],
+                length * sizeof(WCHAR));
+        }
+    }
+
+    path_end = &path_start[length];
+    path_iterator = path_start;
+
+    /* Convert drive letters to lowercase, if present. */
+    if(length >= 2 && path_start[1] == L':') {
+        WCHAR drive_letter = path_start[0];
+
+        if(drive_letter >= L'A' && drive_letter <= L'Z') {
+            path_start[0] = drive_letter - L'A' + L'a';
+        }
+    }
+
+    while(path_iterator < path_end) {
+        if(*path_iterator == L'\\') {
+            *path_iterator = L'/';
+        }
+
+        path_iterator++;
+    }
+
+    /* Truncate the result to its actual length; we don't want to include the
+     * NUL terminator. */
+    return enif_realloc_binary(path, length * sizeof(WCHAR));
+}
+
+/* @brief Checks whether all the given attributes are set on the object at the
+ * given path. Note that it assumes false on errors. */
+static int has_file_attributes(const efile_path_t *path, DWORD mask) {
+    DWORD attributes = GetFileAttributesW((WCHAR*)path->data);
+
+    if(attributes == INVALID_FILE_ATTRIBUTES) {
+        return 0;
+    }
+
+    return !!((attributes & mask) == mask);
+}
+
+static int is_ignored_name(int name_length, const WCHAR *name) {
+    if(name_length == 1 && name[0] == L'.') {
+        return 1;
+    } else if(name_length == 2 && !sys_memcmp(name, L"..", 2 * sizeof(WCHAR))) {
+        return 1;
+    }
+
+    return 0;
+}
+
+static int get_drive_number(const efile_path_t *path) {
+    const WCHAR *path_start;
+    int length;
+
+    ASSERT_PATH_FORMAT(path);
+
+    path_start = (WCHAR*)path->data + LP_PREFIX_LENGTH;
+    length = PATH_LENGTH(path) - LP_PREFIX_LENGTH;
+
+    if(length >= 2 && path_start[1] == L':') {
+        WCHAR drive_letter = path_start[0];
+
+        if(drive_letter >= L'A' && drive_letter <= L'Z') {
+            return drive_letter - L'A' + 1;
+        } else if(drive_letter >= L'a' && drive_letter <= L'z') {
+            return drive_letter - L'a' + 1;
+        }
+    }
+
+    return -1;
+}
+
+/* @brief Checks whether two *paths* are on the same mount point; they don't
+ * have to refer to existing or accessible files/directories. */
+static int has_same_mount_point(const efile_path_t *path_a, const efile_path_t *path_b) {
+    WCHAR *mount_a, *mount_b;
+    int result = 0;
+
+    mount_a = enif_alloc(path_a->size);
+    mount_b = enif_alloc(path_b->size);
+
+    if(mount_a != NULL && mount_b != NULL) {
+        int length_a, length_b;
+
+        length_a = PATH_LENGTH(path_a);
+        length_b = PATH_LENGTH(path_b);
+
+        if(GetVolumePathNameW((WCHAR*)path_a->data, mount_a, length_a)) {
+            ASSERT(wcslen(mount_a) <= length_a);
+
+            if(GetVolumePathNameW((WCHAR*)path_b->data, mount_b, length_b)) {
+                ASSERT(wcslen(mount_b) <= length_b);
+
+                result = !_wcsicmp(mount_a, mount_b);
+            }
+        }
+    }
+
+    if(mount_b != NULL) {
+        enif_free(mount_b);
+    }
+
+    if(mount_a != NULL) {
+        enif_free(mount_a);
+    }
+
+    return result;
+}
+
+/* Mirrors the PathIsRootW function of the shell API, but doesn't choke on
+ * paths longer than MAX_PATH. */
+static int is_path_root(const efile_path_t *path) {
+    const WCHAR *path_start, *path_end;
+    int length;
+
+    ASSERT_PATH_FORMAT(path);
+
+    path_start = (WCHAR*)path->data + LP_PREFIX_LENGTH;
+    length = PATH_LENGTH(path) - LP_PREFIX_LENGTH;
+
+    path_end = &path_start[length];
+
+    if(length == 1) {
+        /* A single \ refers to the root of the current working directory. */
+        return IS_SLASH(path_start[0]);
+    } else if(length == 3 && iswalpha(path_start[0]) && path_start[1] == L':') {
+        /* Drive letter. */
+        return IS_SLASH(path_start[2]);
+    } else if(length >= 4) {
+        /* Check whether we're a UNC root, eg. \\server, \\server\share */
+        const WCHAR *path_iterator;
+
+        if(!IS_SLASH(path_start[0]) || !IS_SLASH(path_start[1])) {
+            return 0;
+        }
+
+        path_iterator = path_start + 2;
+
+        /* Slide to the slash between the server and share names, if present. */
+        while(path_iterator < path_end && !IS_SLASH(*path_iterator)) {
+            path_iterator++;
+        }
+
+        /* Slide past the end of the string, stopping at the first slash we
+         * encounter. */
+        do {
+            path_iterator++;
+        }  while(path_iterator < path_end && !IS_SLASH(*path_iterator));
+
+        /* If we're past the end of the string and it didnt't end with a slash,
+         * then we're a root path. */
+        return path_iterator >= path_end && !IS_SLASH(path_start[length - 1]);
+    }
+
+    return 0;
+}
+
+posix_errno_t efile_open(const efile_path_t *path, enum efile_modes_t modes,
+        ErlNifResourceType *nif_type, efile_data_t **d) {
+
+    DWORD attributes, access_flags, open_mode;
+    HANDLE handle;
+
+    ASSERT_PATH_FORMAT(path);
+
+    access_flags = 0;
+    open_mode = 0;
+
+    if(modes & EFILE_MODE_READ && !(modes & EFILE_MODE_WRITE)) {
+        access_flags = GENERIC_READ;
+        open_mode = OPEN_EXISTING;
+    } else if(modes & EFILE_MODE_WRITE && !(modes & EFILE_MODE_READ)) {
+        access_flags = GENERIC_WRITE;
+        open_mode = CREATE_ALWAYS;
+    } else if(modes & EFILE_MODE_READ_WRITE) {
+        access_flags = GENERIC_READ | GENERIC_WRITE;
+        open_mode = OPEN_ALWAYS;
+    } else {
+        return EINVAL;
+    }
+
+    if(modes & EFILE_MODE_APPEND) {
+        access_flags |= FILE_APPEND_DATA;
+        open_mode = OPEN_ALWAYS;
+    }
+
+    if(modes & EFILE_MODE_EXCLUSIVE) {
+        open_mode = CREATE_NEW;
+    }
+
+    if(modes & EFILE_MODE_SYNC) {
+        attributes = FILE_FLAG_WRITE_THROUGH;
+    } else {
+        attributes = FILE_ATTRIBUTE_NORMAL;
+    }
+
+    handle = CreateFileW((WCHAR*)path->data, access_flags,
+        FILE_SHARE_FLAGS, NULL, open_mode, attributes, NULL);
+
+    if(handle != INVALID_HANDLE_VALUE) {
+        efile_win_t *w;
+
+        w = (efile_win_t*)enif_alloc_resource(nif_type, sizeof(efile_win_t));
+        w->handle = handle;
+
+        EFILE_INIT_RESOURCE(&w->common, modes);
+        (*d) = &w->common;
+
+        return 0;
+    } else {
+        DWORD last_error = GetLastError();
+
+        /* Rewrite all failures on directories to EISDIR to match the old
+         * driver. */
+        if(has_file_attributes(path, FILE_ATTRIBUTE_DIRECTORY)) {
+            return EISDIR;
+        }
+
+        return windows_to_posix_errno(last_error);
+    }
+}
+
+int efile_close(efile_data_t *d) {
+    efile_win_t *w = (efile_win_t*)d;
+    HANDLE handle;
+
+    ASSERT(erts_atomic32_read_nob(&d->state) == EFILE_STATE_CLOSED);
+    ASSERT(w->handle != INVALID_HANDLE_VALUE);
+
+    handle = w->handle;
+    w->handle = INVALID_HANDLE_VALUE;
+
+    if(!CloseHandle(handle)) {
+        w->common.posix_errno = windows_to_posix_errno(GetLastError());
+        return 0;
+    }
+
+    return 1;
+}
+
+static void shift_overlapped(OVERLAPPED *overlapped, DWORD shift) {
+    LARGE_INTEGER offset;
+
+    ASSERT(shift >= 0);
+
+    offset.HighPart = overlapped->OffsetHigh;
+    offset.LowPart = overlapped->Offset;
+
+    /* ~(Uint64)0 is a magic value ("append to end of file") which needs to be
+     * preserved. Other positions resulting in overflow would have errored out
+     * just prior to this point. */
+    if(offset.QuadPart != ERTS_UINT64_MAX) {
+        offset.QuadPart += shift;
+    }
+
+    /* All unused fields must be zeroed for the next call. */
+    sys_memset(overlapped, 0, sizeof(*overlapped));
+    overlapped->OffsetHigh = offset.HighPart;
+    overlapped->Offset = offset.LowPart;
+}
+
+static void shift_iov(SysIOVec **iov, int *iovlen, DWORD shift) {
+    SysIOVec *head_vec = (*iov);
+
+    ASSERT(shift >= 0);
+
+    while(shift > 0) {
+        ASSERT(head_vec < &(*iov)[*iovlen]);
+
+        if(shift < head_vec->iov_len) {
+            head_vec->iov_base = (char*)head_vec->iov_base + shift;
+            head_vec->iov_len -= shift;
+            break;
+        } else {
+            shift -= head_vec->iov_len;
+            head_vec++;
+        }
+    }
+
+    (*iovlen) -= head_vec - (*iov);
+    (*iov) = head_vec;
+}
+
+typedef BOOL (WINAPI *io_op_t)(HANDLE, LPVOID, DWORD, LPDWORD, LPOVERLAPPED);
+
+static Sint64 internal_sync_io(efile_win_t *w, io_op_t operation,
+        SysIOVec *iov, int iovlen, OVERLAPPED *overlapped) {
+
+    Sint64 bytes_processed = 0;
+
+    for(;;) {
+        DWORD block_bytes_processed, last_error;
+        BOOL succeeded;
+
+        if(iovlen < 1) {
+            return bytes_processed;
+        }
+
+        succeeded = operation(w->handle, iov->iov_base, iov->iov_len,
+            &block_bytes_processed, overlapped);
+        last_error = GetLastError();
+
+        if(!succeeded && (last_error != ERROR_HANDLE_EOF)) {
+            w->common.posix_errno = windows_to_posix_errno(last_error);
+            return -1;
+        } else if(block_bytes_processed == 0) {
+            /* EOF */
+            return bytes_processed;
+        }
+
+        if(overlapped != NULL) {
+            shift_overlapped(overlapped, block_bytes_processed);
+        }
+
+        shift_iov(&iov, &iovlen, block_bytes_processed);
+
+        bytes_processed += block_bytes_processed;
+    }
+}
+
+Sint64 efile_readv(efile_data_t *d, SysIOVec *iov, int iovlen) {
+    efile_win_t *w = (efile_win_t*)d;
+
+    return internal_sync_io(w, ReadFile, iov, iovlen, NULL);
+}
+
+Sint64 efile_writev(efile_data_t *d, SysIOVec *iov, int iovlen) {
+    efile_win_t *w = (efile_win_t*)d;
+
+    OVERLAPPED __overlapped, *overlapped;
+    Uint64 bytes_written;
+
+    if(w->common.modes & EFILE_MODE_APPEND) {
+        overlapped = &__overlapped;
+
+        sys_memset(overlapped, 0, sizeof(*overlapped));
+        overlapped->OffsetHigh = 0xFFFFFFFF;
+        overlapped->Offset = 0xFFFFFFFF;
+    } else {
+        overlapped = NULL;
+    }
+
+    return internal_sync_io(w, WriteFile, iov, iovlen, overlapped);
+}
+
+Sint64 efile_preadv(efile_data_t *d, Sint64 offset, SysIOVec *iov, int iovlen) {
+    efile_win_t *w = (efile_win_t*)d;
+
+    OVERLAPPED overlapped;
+
+    sys_memset(&overlapped, 0, sizeof(overlapped));
+    overlapped.OffsetHigh = (offset >> 32) & 0xFFFFFFFF;
+    overlapped.Offset = offset & 0xFFFFFFFF;
+
+    return internal_sync_io(w, ReadFile, iov, iovlen, &overlapped);
+}
+
+Sint64 efile_pwritev(efile_data_t *d, Sint64 offset, SysIOVec *iov, int iovlen) {
+    efile_win_t *w = (efile_win_t*)d;
+
+    OVERLAPPED overlapped;
+
+    sys_memset(&overlapped, 0, sizeof(overlapped));
+    overlapped.OffsetHigh = (offset >> 32) & 0xFFFFFFFF;
+    overlapped.Offset = offset & 0xFFFFFFFF;
+
+    return internal_sync_io(w, WriteFile, iov, iovlen, &overlapped);
+}
+
+int efile_seek(efile_data_t *d, enum efile_seek_t seek, Sint64 offset, Sint64 *new_position) {
+    efile_win_t *w = (efile_win_t*)d;
+
+    LARGE_INTEGER large_offset, large_new_position;
+    DWORD whence;
+
+    switch(seek) {
+        case EFILE_SEEK_BOF: whence = FILE_BEGIN; break;
+        case EFILE_SEEK_CUR: whence = FILE_CURRENT; break;
+        case EFILE_SEEK_EOF: whence = FILE_END; break;
+        default: ERTS_INTERNAL_ERROR("Invalid seek parameter");
+    }
+
+    large_offset.QuadPart = offset;
+
+    if(!SetFilePointerEx(w->handle, large_offset, &large_new_position, whence)) {
+        w->common.posix_errno = windows_to_posix_errno(GetLastError());
+        return 0;
+    }
+
+    (*new_position) = large_new_position.QuadPart;
+
+    return 1;
+}
+
+int efile_sync(efile_data_t *d, int data_only) {
+    efile_win_t *w = (efile_win_t*)d;
+
+    /* Windows doesn't support data-only syncing. */
+    (void)data_only;
+
+    if(!FlushFileBuffers(w->handle)) {
+        w->common.posix_errno = windows_to_posix_errno(GetLastError());
+        return 0;
+    }
+
+    return 1;
+}
+
+int efile_advise(efile_data_t *d, Sint64 offset, Sint64 length, enum efile_advise_t advise) {
+    /* Windows doesn't support this, but we'll pretend it does since the call
+     * is only a recommendation even on systems that do support it. */
+
+    (void)d;
+    (void)offset;
+    (void)length;
+    (void)advise;
+
+    return 1;
+}
+
+int efile_allocate(efile_data_t *d, Sint64 offset, Sint64 length) {
+    efile_win_t *w = (efile_win_t*)d;
+
+    (void)d;
+    (void)offset;
+    (void)length;
+
+    w->common.posix_errno = ENOTSUP;
+
+    return 0;
+}
+
+int efile_truncate(efile_data_t *d) {
+    efile_win_t *w = (efile_win_t*)d;
+
+    if(!SetEndOfFile(w->handle)) {
+        w->common.posix_errno = windows_to_posix_errno(GetLastError());
+        return 0;
+    }
+
+    return 1;
+}
+
+static int is_executable_file(const efile_path_t *path) {
+    /* We're using the file extension in order to be quirks-compliant with the
+     * old driver, which never bothered to check the actual permissions. We
+     * could easily do so now (cf. GetNamedSecurityInfo) but the execute
+     * permission is only relevant for files that are started with the default
+     * loader, and batch files run just fine with read permission alone. */
+
+    int length = PATH_LENGTH(path);
+
+    if(length >= 4) {
+        const WCHAR *last_four = &((WCHAR*)path->data)[length - 4];
+
+        if (!_wcsicmp(last_four, L".exe") ||
+            !_wcsicmp(last_four, L".cmd") ||
+            !_wcsicmp(last_four, L".bat") ||
+            !_wcsicmp(last_four, L".com")) {
+            return 1;
+        }
+    }
+
+    return 0;
+}
+
+posix_errno_t efile_read_info(const efile_path_t *path, int follow_links, efile_fileinfo_t *result) {
+    BY_HANDLE_FILE_INFORMATION native_file_info;
+    DWORD attributes;
+
+    sys_memset(&native_file_info, 0, sizeof(native_file_info));
+
+    attributes = GetFileAttributesW((WCHAR*)path->data);
+
+    if(attributes == INVALID_FILE_ATTRIBUTES) {
+        DWORD last_error = GetLastError();
+
+        /* Querying a network share root fails with ERROR_BAD_NETPATH, so we'll
+         * fake it as a directory just like local roots. */
+        if(!is_path_root(path) || last_error != ERROR_BAD_NETPATH) {
+            return windows_to_posix_errno(last_error);
+        }
+
+        attributes = FILE_ATTRIBUTE_DIRECTORY;
+    } else if(is_path_root(path)) {
+        /* Local (or mounted) roots can be queried with GetFileAttributesW but
+         * lack support for GetFileInformationByHandle, so we'll skip that
+         * part. */
+    } else {
+        HANDLE handle;
+
+        if(follow_links && (attributes & FILE_ATTRIBUTE_REPARSE_POINT)) {
+            posix_errno_t posix_errno;
+            efile_path_t resolved_path;
+
+            posix_errno = internal_read_link(path, &resolved_path);
+
+            if(posix_errno != 0) {
+                return posix_errno;
+            }
+
+            return efile_read_info(&resolved_path, 0, result);
+        }
+
+        handle = CreateFileW((const WCHAR*)path->data, GENERIC_READ,
+            FILE_SHARE_FLAGS, NULL, OPEN_EXISTING, FILE_FLAG_BACKUP_SEMANTICS,
+            NULL);
+
+        /* The old driver never cared whether this succeeded. */
+        if(handle != INVALID_HANDLE_VALUE) {
+            GetFileInformationByHandle(handle, &native_file_info);
+            CloseHandle(handle);
+        }
+
+        FILETIME_TO_EPOCH(result->m_time, native_file_info.ftLastWriteTime);
+        FILETIME_TO_EPOCH(result->a_time, native_file_info.ftLastAccessTime);
+        FILETIME_TO_EPOCH(result->c_time, native_file_info.ftCreationTime);
+
+        if(result->m_time == -EPOCH_DIFFERENCE) {
+            /* Default to 1970 just like the old driver. */
+            result->m_time = 0;
+        }
+
+        if(result->a_time == -EPOCH_DIFFERENCE) {
+            result->a_time = result->m_time;
+        }
+
+        if(result->c_time == -EPOCH_DIFFERENCE) {
+            result->c_time = result->m_time;
+        }
+    }
+
+    if(attributes & FILE_ATTRIBUTE_REPARSE_POINT) {
+        result->type = EFILE_FILETYPE_SYMLINK;
+        /* This should be _S_IFLNK, but the old driver always set
+         * non-directories to _S_IFREG. */
+        result->mode |= _S_IFREG;
+    } else if(attributes & FILE_ATTRIBUTE_DIRECTORY) {
+        result->type = EFILE_FILETYPE_DIRECTORY;
+        result->mode |= _S_IFDIR | _S_IEXEC;
+    } else {
+        if(is_executable_file(path)) {
+            result->mode |= _S_IEXEC;
+        }
+
+        result->type = EFILE_FILETYPE_REGULAR;
+        result->mode |= _S_IFREG;
+    }
+
+    if(!(attributes & FILE_ATTRIBUTE_READONLY)) {
+        result->access = EFILE_ACCESS_READ | EFILE_ACCESS_WRITE;
+        result->mode |= _S_IREAD | _S_IWRITE;
+    } else {
+        result->access = EFILE_ACCESS_READ;
+        result->mode |= _S_IREAD;
+    }
+
+    /* Propagate user mode-bits to group/other fields */
+    result->mode |= (result->mode & 0700) >> 3;
+    result->mode |= (result->mode & 0700) >> 6;
+
+    result->size =
+        ((Uint64)native_file_info.nFileSizeHigh << 32ull) |
+        (Uint64)native_file_info.nFileSizeLow;
+
+    result->links = MAX(1, native_file_info.nNumberOfLinks);
+
+    result->major_device = get_drive_number(path);
+    result->minor_device = 0;
+    result->inode = 0;
+    result->uid = 0;
+    result->gid = 0;
+
+    return 0;
+}
+
+posix_errno_t efile_set_permissions(const efile_path_t *path, Uint32 permissions) {
+    DWORD attributes = GetFileAttributesW((WCHAR*)path->data);
+
+    if(attributes == INVALID_FILE_ATTRIBUTES) {
+        return windows_to_posix_errno(GetLastError());
+    }
+
+    if(permissions & _S_IWRITE) {
+        attributes &= ~FILE_ATTRIBUTE_READONLY;
+    } else {
+        attributes |= FILE_ATTRIBUTE_READONLY;
+    }
+
+    if(SetFileAttributesW((WCHAR*)path->data, attributes)) {
+        return 0;
+    }
+
+    return windows_to_posix_errno(GetLastError());
+}
+
+posix_errno_t efile_set_owner(const efile_path_t *path, Uint32 owner, Uint32 group) {
+    (void)path;
+    (void)owner;
+    (void)group;
+
+    return 0;
+}
+
+posix_errno_t efile_set_time(const efile_path_t *path, Sint64 a_time, Sint64 m_time, Sint64 c_time) {
+    FILETIME accessed, modified, created;
+    DWORD last_error, attributes;
+    HANDLE handle;
+
+    attributes = GetFileAttributesW((WCHAR*)path->data);
+
+    if(attributes == INVALID_FILE_ATTRIBUTES) {
+        return windows_to_posix_errno(GetLastError());
+    }
+
+    /* If the file is read-only, we have to make it temporarily writable while
+     * setting new metadata. */
+    if(attributes & FILE_ATTRIBUTE_READONLY) {
+        DWORD without_readonly = attributes & ~FILE_ATTRIBUTE_READONLY;
+
+        if(!SetFileAttributesW((WCHAR*)path->data, without_readonly)) {
+            return windows_to_posix_errno(GetLastError());
+        }
+    }
+
+    EPOCH_TO_FILETIME(modified, m_time);
+    EPOCH_TO_FILETIME(accessed, a_time);
+    EPOCH_TO_FILETIME(created, c_time);
+
+    handle = CreateFileW((WCHAR*)path->data, GENERIC_READ | GENERIC_WRITE,
+        FILE_SHARE_FLAGS, NULL, OPEN_EXISTING, FILE_FLAG_BACKUP_SEMANTICS, NULL);
+    last_error = GetLastError();
+
+    if(handle != INVALID_HANDLE_VALUE) {
+        if(SetFileTime(handle, &created, &accessed, &modified)) {
+            last_error = ERROR_SUCCESS;
+        } else {
+            last_error = GetLastError();
+        }
+
+        CloseHandle(handle);
+    }
+
+    if(attributes & FILE_ATTRIBUTE_READONLY) {
+        SetFileAttributesW((WCHAR*)path->data, attributes);
+    }
+
+    return windows_to_posix_errno(last_error);
+}
+
+static posix_errno_t internal_read_link(const efile_path_t *path, efile_path_t *result) {
+    DWORD required_length, actual_length;
+    HANDLE link_handle;
+    DWORD last_error;
+
+    link_handle = CreateFileW((WCHAR*)path->data, GENERIC_READ,
+        FILE_SHARE_FLAGS, NULL, OPEN_EXISTING, FILE_FLAG_BACKUP_SEMANTICS, NULL);
+    last_error = GetLastError();
+
+    if(link_handle == INVALID_HANDLE_VALUE) {
+        return windows_to_posix_errno(last_error);
+    }
+
+    required_length = GetFinalPathNameByHandleW(link_handle, NULL, 0, 0);
+    last_error = GetLastError();
+
+    if(required_length <= 0) {
+        CloseHandle(link_handle);
+        return windows_to_posix_errno(last_error);
+    }
+
+    /* Unlike many other path functions (eg. GetFullPathNameW), this one
+     * includes the NUL terminator in its required length. */
+    if(!enif_alloc_binary(required_length * sizeof(WCHAR), result)) {
+        CloseHandle(link_handle);
+        return ENOMEM;
+    }
+
+    actual_length = GetFinalPathNameByHandleW(link_handle,
+        (WCHAR*)result->data, required_length, 0);
+    last_error = GetLastError();
+
+    CloseHandle(link_handle);
+
+    if(actual_length == 0 || actual_length >= required_length) {
+        enif_release_binary(result);
+        return windows_to_posix_errno(last_error);
+    }
+
+    /* GetFinalPathNameByHandle always prepends with "\\?\" and NUL-terminates,
+     * so we never have to touch-up the resulting path. */
+
+    ASSERT_PATH_FORMAT(result);
+
+    return 0;
+}
+
+posix_errno_t efile_read_link(ErlNifEnv *env, const efile_path_t *path, ERL_NIF_TERM *result) {
+    posix_errno_t posix_errno;
+    ErlNifBinary result_bin;
+    DWORD attributes;
+
+    ASSERT_PATH_FORMAT(path);
+
+    attributes = GetFileAttributesW((WCHAR*)path->data);
+
+    if(attributes == INVALID_FILE_ATTRIBUTES) {
+        return windows_to_posix_errno(GetLastError());
+    } else if(!(attributes & FILE_ATTRIBUTE_REPARSE_POINT)) {
+        return EINVAL;
+    }
+
+    posix_errno = internal_read_link(path, &result_bin);
+
+    if(posix_errno == 0) {
+        if(!normalize_path_result(&result_bin)) {
+            enif_release_binary(&result_bin);
+            return ENOMEM;
+        }
+
+        (*result) = enif_make_binary(env, &result_bin);
+    }
+
+    return posix_errno;
+}
+
+posix_errno_t efile_list_dir(ErlNifEnv *env, const efile_path_t *path, ERL_NIF_TERM *result) {
+    ERL_NIF_TERM list_head;
+    WIN32_FIND_DATAW data;
+    HANDLE search_handle;
+    WCHAR *search_path;
+    DWORD last_error;
+
+    ASSERT_PATH_FORMAT(path);
+
+    search_path = enif_alloc(path->size + 2 * sizeof(WCHAR));
+
+    if(search_path == NULL) {
+        return ENOMEM;
+    }
+
+    sys_memcpy(search_path, path->data, path->size);
+    search_path[PATH_LENGTH(path) + 0] = L'\\';
+    search_path[PATH_LENGTH(path) + 1] = L'*';
+    search_path[PATH_LENGTH(path) + 2] = L'\0';
+
+    search_handle = FindFirstFileW(search_path, &data);
+    last_error = GetLastError();
+
+    enif_free(search_path);
+
+    if(search_handle == INVALID_HANDLE_VALUE) {
+        return windows_to_posix_errno(last_error);
+    }
+
+    list_head = enif_make_list(env, 0);
+
+    do {
+        int name_length = wcslen(data.cFileName);
+
+        if(!is_ignored_name(name_length, data.cFileName)) {
+            unsigned char *name_bytes;
+            ERL_NIF_TERM name_term;
+            size_t name_size;
+
+            name_size = name_length * sizeof(WCHAR);
+
+            name_bytes = enif_make_new_binary(env, name_size, &name_term);
+            sys_memcpy(name_bytes, data.cFileName, name_size);
+
+            list_head = enif_make_list_cell(env, name_term, list_head);
+        }
+    } while(FindNextFileW(search_handle, &data));
+
+    FindClose(search_handle);
+    (*result) = list_head;
+
+    return 0;
+}
+
+posix_errno_t efile_rename(const efile_path_t *old_path, const efile_path_t *new_path) {
+    BOOL old_is_directory, new_is_directory;
+    DWORD move_flags, last_error;
+
+    ASSERT_PATH_FORMAT(old_path);
+    ASSERT_PATH_FORMAT(new_path);
+
+    move_flags = MOVEFILE_COPY_ALLOWED | MOVEFILE_WRITE_THROUGH;
+
+    if(MoveFileExW((WCHAR*)old_path->data, (WCHAR*)new_path->data, move_flags)) {
+        return 0;
+    }
+
+    last_error = GetLastError();
+
+    old_is_directory = has_file_attributes(old_path, FILE_ATTRIBUTE_DIRECTORY);
+    new_is_directory = has_file_attributes(new_path, FILE_ATTRIBUTE_DIRECTORY);
+
+    switch(last_error) {
+    case ERROR_SHARING_VIOLATION:
+    case ERROR_ACCESS_DENIED:
+        if(old_is_directory) {
+            BOOL moved_into_itself;
+
+            moved_into_itself = (old_path->size <= new_path->size) &&
+                !_wcsnicmp((WCHAR*)old_path->data, (WCHAR*)new_path->data,
+                    PATH_LENGTH(old_path));
+
+            if(moved_into_itself) {
+                return EINVAL;
+            } else if(is_path_root(old_path)) {
+                return EINVAL;
+            }
+
+            /* Renaming a directory across volumes needs to be rewritten as
+             * EXDEV so that the caller can respond by simulating it with
+             * copy/delete operations.
+             *
+             * Files are handled through MOVEFILE_COPY_ALLOWED. */
+            if(!has_same_mount_point(old_path, new_path)) {
+                return EXDEV;
+            }
+        }
+        break;
+    case ERROR_PATH_NOT_FOUND:
+    case ERROR_FILE_NOT_FOUND:
+        return ENOENT;
+    case ERROR_ALREADY_EXISTS:
+    case ERROR_FILE_EXISTS:
+        if(old_is_directory && !new_is_directory) {
+            return ENOTDIR;
+        } else if(!old_is_directory && new_is_directory) {
+            return EISDIR;
+        } else if(old_is_directory && new_is_directory) {
+            /* This will fail if the destination isn't empty. */
+            if(RemoveDirectoryW((WCHAR*)new_path->data)) {
+                return efile_rename(old_path, new_path);
+            }
+
+            return EEXIST;
+        } else if(!old_is_directory && !new_is_directory) {
+            /* This is pretty iffy; the public documentation says that the
+             * operation may EACCES on some systems when either file is open,
+             * which gives us room to use MOVEFILE_REPLACE_EXISTING and be done
+             * with it, but the old implementation simulated Unix semantics and
+             * there's a lot of code that relies on that.
+             *
+             * The simulation renames the destination to a scratch name to get
+             * around the fact that it's impossible to open (and by extension
+             * rename) a file that's been deleted while open. It has a few
+             * drawbacks though;
+             *
+             * 1) It's not atomic as there's a small window where there's no
+             *    file at all on the destination path.
+             * 2) It will confuse applications that subscribe to folder
+             *    changes.
+             * 3) It will fail if we lack general permission to write in the
+             *    same folder. */
+
+            WCHAR *swap_path = enif_alloc(new_path->size + sizeof(WCHAR) * 64);
+
+            if(swap_path == NULL) {
+                return ENOMEM;
+            } else {
+                static LONGLONG unique_counter = 0;
+                WCHAR *swap_path_end;
+
+                /* We swap in the same folder as the destination to be
+                 * reasonably sure that it's on the same volume. Note that
+                 * we're avoiding GetTempFileNameW as it will fail on long
+                 * paths. */
+
+                sys_memcpy(swap_path, (WCHAR*)new_path->data, new_path->size);
+                swap_path_end = swap_path + PATH_LENGTH(new_path);
+
+                while(!IS_SLASH(*swap_path_end)) {
+                    ASSERT(swap_path_end > swap_path);
+                    swap_path_end--;
+                }
+
+                StringCchPrintfW(&swap_path_end[1], 64, L"erl-%lx-%llx.tmp",
+                    GetCurrentProcessId(), unique_counter);
+                InterlockedIncrement64(&unique_counter);
+            }
+
+            if(MoveFileExW((WCHAR*)new_path->data, swap_path, MOVEFILE_REPLACE_EXISTING)) {
+                if(MoveFileExW((WCHAR*)old_path->data, (WCHAR*)new_path->data, move_flags)) {
+                    last_error = ERROR_SUCCESS;
+                    DeleteFileW(swap_path);
+                } else {
+                    last_error = GetLastError();
+                    MoveFileW(swap_path, (WCHAR*)new_path->data);
+                }
+            } else {
+                last_error = GetLastError();
+                DeleteFileW(swap_path);
+            }
+
+            enif_free(swap_path);
+
+            return windows_to_posix_errno(last_error);
+        }
+
+        return EEXIST;
+    }
+
+    return windows_to_posix_errno(last_error);
+}
+
+posix_errno_t efile_make_hard_link(const efile_path_t *existing_path, const efile_path_t *new_path) {
+    ASSERT_PATH_FORMAT(existing_path);
+    ASSERT_PATH_FORMAT(new_path);
+
+    if(!CreateHardLinkW((WCHAR*)new_path->data, (WCHAR*)existing_path->data, NULL)) {
+        return windows_to_posix_errno(GetLastError());
+    }
+
+    return 0;
+}
+
+posix_errno_t efile_make_soft_link(const efile_path_t *existing_path, const efile_path_t *new_path) {
+    DWORD link_flags;
+
+    ASSERT_PATH_FORMAT(existing_path);
+    ASSERT_PATH_FORMAT(new_path);
+
+    if(has_file_attributes(existing_path, FILE_ATTRIBUTE_DIRECTORY)) {
+        link_flags = SYMBOLIC_LINK_FLAG_DIRECTORY;
+    } else {
+        link_flags = 0;
+    }
+
+    if(!CreateSymbolicLinkW((WCHAR*)new_path->data, (WCHAR*)existing_path->data, link_flags)) {
+        return windows_to_posix_errno(GetLastError());
+    }
+
+    return 0;
+}
+
+posix_errno_t efile_make_dir(const efile_path_t *path) {
+    ASSERT_PATH_FORMAT(path);
+
+    if(!CreateDirectoryW((WCHAR*)path->data, NULL)) {
+        return windows_to_posix_errno(GetLastError());
+    }
+
+    return 0;
+}
+
+posix_errno_t efile_del_file(const efile_path_t *path) {
+    ASSERT_PATH_FORMAT(path);
+
+    if(!DeleteFileW((WCHAR*)path->data)) {
+        DWORD last_error = GetLastError();
+
+        switch(last_error) {
+        case ERROR_INVALID_NAME:
+            /* Attempted to delete a device or similar. */
+            return EACCES;
+        case ERROR_ACCESS_DENIED:
+            /* Windows NT reports removing a directory as EACCES instead of
+             * EPERM. */
+            if(has_file_attributes(path, FILE_ATTRIBUTE_DIRECTORY)) {
+                return EPERM;
+            }
+            break;
+        }
+
+        return windows_to_posix_errno(last_error);
+    }
+
+    return 0;
+}
+
+posix_errno_t efile_del_dir(const efile_path_t *path) {
+    ASSERT_PATH_FORMAT(path);
+
+    if(!RemoveDirectoryW((WCHAR*)path->data)) {
+        DWORD last_error = GetLastError();
+
+        if(last_error == ERROR_DIRECTORY) {
+            return ENOTDIR;
+        }
+
+        return windows_to_posix_errno(last_error);
+    }
+
+    return 0;
+}
+
+posix_errno_t efile_set_cwd(const efile_path_t *path) {
+    const WCHAR *path_start;
+
+    ASSERT_PATH_FORMAT(path);
+
+    /* We have to use _wchdir since that's the only function that updates the
+     * per-drive working directory, but it naively assumes that all paths
+     * starting with \\ are UNC paths, so we have to skip the \\?\-prefix. */
+    path_start = (WCHAR*)path->data + LP_PREFIX_LENGTH;
+
+    if(_wchdir(path_start)) {
+        return windows_to_posix_errno(GetLastError());
+    }
+
+    return 0;
+}
+
+static int is_valid_drive(int device_index) {
+    WCHAR drive_path[4] = {L'?', L':', L'\\', L'\0'};
+
+    if(device_index == 0) {
+        /* Default drive; always valid. */
+        return 1;
+    } else if(device_index > (L'Z' - L'A' + 1)) {
+        return 0;
+    }
+
+    drive_path[0] = device_index + L'A' - 1;
+
+    switch(GetDriveTypeW(drive_path)) {
+    case DRIVE_NO_ROOT_DIR:
+    case DRIVE_UNKNOWN:
+        return 0;
+    }
+
+    return 1;
+}
+
+posix_errno_t efile_get_device_cwd(ErlNifEnv *env, int device_index, ERL_NIF_TERM *result) {
+    ErlNifBinary result_bin;
+
+    /* _wgetdcwd might crash the entire emulator on debug builds since the CRT
+     * invalid parameter handler asserts if passed a non-existent drive (Or
+     * simply one that has been unmounted), so we check it ourselves to avoid
+     * that. */
+    if(!is_valid_drive(device_index)) {
+        return EACCES;
+    }
+
+    if(!enif_alloc_binary(MAX_PATH * sizeof(WCHAR), &result_bin)) {
+        return ENOMEM;
+    }
+
+    if(_wgetdcwd(device_index, (WCHAR*)result_bin.data, MAX_PATH) == NULL) {
+        enif_release_binary(&result_bin);
+        return EACCES;
+    }
+
+    if(!normalize_path_result(&result_bin)) {
+        enif_release_binary(&result_bin);
+        return ENOMEM;
+    }
+
+    (*result) = enif_make_binary(env, &result_bin);
+
+    return 0;
+}
+
+posix_errno_t efile_get_cwd(ErlNifEnv *env, ERL_NIF_TERM *result) {
+    return efile_get_device_cwd(env, 0, result);
+}
+
+posix_errno_t efile_altname(ErlNifEnv *env, const efile_path_t *path, ERL_NIF_TERM *result) {
+    ErlNifBinary result_bin;
+
+    ASSERT_PATH_FORMAT(path);
+
+    if(is_path_root(path)) {
+        /* Root paths can't be queried so we'll just return them as they are. */
+        if(!enif_alloc_binary(path->size, &result_bin)) {
+            return ENOMEM;
+        }
+
+        sys_memcpy(result_bin.data, path->data, path->size);
+    } else {
+        WIN32_FIND_DATAW data;
+        HANDLE handle;
+
+        WCHAR *name_buffer;
+        int name_length;
+
+        /* Reject path wildcards. */
+        if(wcspbrk(&((const WCHAR*)path->data)[4], L"?*")) {
+            return ENOENT;
+        }
+
+        handle = FindFirstFileW((const WCHAR*)path->data, &data);
+
+        if(handle == INVALID_HANDLE_VALUE) {
+            return windows_to_posix_errno(GetLastError());
+        }
+
+        FindClose(handle);
+
+        name_length = wcslen(data.cAlternateFileName);
+
+        if(name_length > 0) {
+            name_buffer = data.cAlternateFileName;
+        } else {
+            name_length = wcslen(data.cFileName);
+            name_buffer = data.cFileName;
+        }
+
+        /* Include NUL-terminator; it will be removed after normalization. */
+        name_length += 1;
+
+        if(!enif_alloc_binary(name_length * sizeof(WCHAR), &result_bin)) {
+            return ENOMEM;
+        }
+
+        sys_memcpy(result_bin.data, name_buffer, name_length * sizeof(WCHAR));
+    }
+
+    if(!normalize_path_result(&result_bin)) {
+        enif_release_binary(&result_bin);
+        return ENOMEM;
+    }
+
+    (*result) = enif_make_binary(env, &result_bin);
+
+    return 0;
+}
+
+static int windows_to_posix_errno(DWORD last_error) {
+    switch(last_error) {
+    case ERROR_SUCCESS:
+        return 0;
+    case ERROR_INVALID_FUNCTION:
+    case ERROR_INVALID_DATA:
+    case ERROR_INVALID_PARAMETER:
+    case ERROR_INVALID_TARGET_HANDLE:
+    case ERROR_INVALID_CATEGORY:
+    case ERROR_NEGATIVE_SEEK:
+        return EINVAL;
+    case ERROR_DIR_NOT_EMPTY:
+        return EEXIST;
+    case ERROR_BAD_FORMAT:
+        return ENOEXEC;
+    case ERROR_PATH_NOT_FOUND:
+    case ERROR_FILE_NOT_FOUND:
+    case ERROR_NO_MORE_FILES:
+        return ENOENT;
+    case ERROR_TOO_MANY_OPEN_FILES:
+        return EMFILE;
+    case ERROR_ACCESS_DENIED:
+    case ERROR_INVALID_ACCESS:
+    case ERROR_CURRENT_DIRECTORY:
+    case ERROR_SHARING_VIOLATION:
+    case ERROR_LOCK_VIOLATION:
+    case ERROR_INVALID_PASSWORD:
+    case ERROR_DRIVE_LOCKED:
+        return EACCES;
+    case ERROR_INVALID_HANDLE:
+        return EBADF;
+    case ERROR_NOT_ENOUGH_MEMORY:
+    case ERROR_OUTOFMEMORY:
+    case ERROR_OUT_OF_STRUCTURES:
+        return ENOMEM;
+    case ERROR_INVALID_DRIVE:
+    case ERROR_BAD_UNIT:
+    case ERROR_NOT_READY:
+    case ERROR_REM_NOT_LIST:
+    case ERROR_DUP_NAME:
+    case ERROR_BAD_NETPATH:
+    case ERROR_NETWORK_BUSY:
+    case ERROR_DEV_NOT_EXIST:
+    case ERROR_BAD_NET_NAME:
+        return ENXIO;
+    case ERROR_NOT_SAME_DEVICE:
+        return EXDEV;
+    case ERROR_WRITE_PROTECT:
+        return EROFS;
+    case ERROR_BAD_LENGTH:
+    case ERROR_BUFFER_OVERFLOW:
+        return E2BIG;
+    case ERROR_SEEK:
+    case ERROR_SECTOR_NOT_FOUND:
+        return ESPIPE;
+    case ERROR_NOT_DOS_DISK:
+        return ENODEV;
+    case ERROR_GEN_FAILURE:
+        return ENODEV;
+    case ERROR_SHARING_BUFFER_EXCEEDED:
+    case ERROR_NO_MORE_SEARCH_HANDLES:
+        return EMFILE;
+    case ERROR_HANDLE_EOF:
+    case ERROR_BROKEN_PIPE:
+        return EPIPE;
+    case ERROR_HANDLE_DISK_FULL:
+    case ERROR_DISK_FULL:
+        return ENOSPC;
+    case ERROR_NOT_SUPPORTED:
+        return ENOTSUP;
+    case ERROR_FILE_EXISTS:
+    case ERROR_ALREADY_EXISTS:
+    case ERROR_CANNOT_MAKE:
+        return EEXIST;
+    case ERROR_ALREADY_ASSIGNED:
+        return EBUSY;
+    case ERROR_NO_PROC_SLOTS:
+        return EAGAIN;
+    case ERROR_CANT_RESOLVE_FILENAME:
+        return EMLINK;
+    case ERROR_PRIVILEGE_NOT_HELD:
+        return EPERM;
+    case ERROR_ARENA_TRASHED:
+    case ERROR_INVALID_BLOCK:
+    case ERROR_BAD_ENVIRONMENT:
+    case ERROR_BAD_COMMAND:
+    case ERROR_CRC:
+    case ERROR_OUT_OF_PAPER:
+    case ERROR_READ_FAULT:
+    case ERROR_WRITE_FAULT:
+    case ERROR_WRONG_DISK:
+    case ERROR_NET_WRITE_FAULT:
+        return EIO;
+    default: /* not to do with files I expect. */
+        return EIO;
+    }
+}
diff --git a/erts/emulator/test/ddll_SUITE.erl b/erts/emulator/test/ddll_SUITE.erl
index 031b05790d..4998fc08be 100644
--- a/erts/emulator/test/ddll_SUITE.erl
+++ b/erts/emulator/test/ddll_SUITE.erl
@@ -775,7 +775,7 @@ errors(Config) when is_list(Config) ->
     {error, bad_driver_name} = erl_ddll:load_driver(Path, wrongname_drv),
 
     %% We assume that there is a statically linked driver named "ddll":
-    {error, linked_in_driver} = erl_ddll:unload_driver(efile),
+    {error, linked_in_driver} = erl_ddll:unload_driver(ram_file_drv),
     {error, not_loaded} = erl_ddll:unload_driver("__pucko_driver__"),
 
     case os:type() of
diff --git a/erts/emulator/test/dirty_bif_SUITE.erl b/erts/emulator/test/dirty_bif_SUITE.erl
index 981ec4d48d..c8f0cbf42d 100644
--- a/erts/emulator/test/dirty_bif_SUITE.erl
+++ b/erts/emulator/test/dirty_bif_SUITE.erl
@@ -230,6 +230,11 @@ dirty_scheduler_exit(Config) when is_list(Config) ->
     {ok, Node} = start_node(Config, "+SDio 1"),
     [ok] = mcall(Node,
                  [fun() ->
+                          %% Perform a dry run to ensure that all required code
+                          %% is loaded. Otherwise the test will fail since code
+                          %% loading is done through dirty IO and it won't make
+                          %% any progress during this test.
+                          _DryRun = test_dirty_scheduler_exit(),
 			  Start = erlang:monotonic_time(millisecond),
                           ok = test_dirty_scheduler_exit(),
 			  End = erlang:monotonic_time(millisecond),
@@ -246,23 +251,22 @@ test_dse(0,Pids) ->
     timer:sleep(100),
     kill_dse(Pids,[]);
 test_dse(N,Pids) ->
-    Pid = spawn_link(fun () -> erts_debug:dirty_io(wait, 5000) end),
+    Pid = spawn_link(fun () -> erts_debug:dirty_io(wait, 1000) end),
     test_dse(N-1,[Pid|Pids]).
 
 kill_dse([],Killed) ->
-    wait_dse(Killed);
+    wait_dse(Killed, ok);
 kill_dse([Pid|Pids],AlreadyKilled) ->
     exit(Pid,kill),
     kill_dse(Pids,[Pid|AlreadyKilled]).
 
-wait_dse([]) ->
-    ok;
-wait_dse([Pid|Pids]) ->
+wait_dse([], Result) ->
+    Result;
+wait_dse([Pid|Pids], Result) ->
     receive
-        {'EXIT',Pid,Reason} ->
-	    killed = Reason
-    end,
-    wait_dse(Pids).
+        {'EXIT', Pid, killed} -> wait_dse(Pids, Result);
+        {'EXIT', Pid, _Other} -> wait_dse(Pids, failed)
+    end.
 
 dirty_call_while_terminated(Config) when is_list(Config) ->
     Me = self(),
diff --git a/erts/emulator/test/dirty_nif_SUITE.erl b/erts/emulator/test/dirty_nif_SUITE.erl
index 13806fd5c4..d36113c3e3 100644
--- a/erts/emulator/test/dirty_nif_SUITE.erl
+++ b/erts/emulator/test/dirty_nif_SUITE.erl
@@ -151,6 +151,11 @@ dirty_scheduler_exit(Config) when is_list(Config) ->
     [ok] = mcall(Node,
                  [fun() ->
                           ok = erlang:load_nif(NifLib, []),
+                          %% Perform a dry run to ensure that all required code
+                          %% is loaded. Otherwise the test will fail since code
+                          %% loading is done through dirty IO and it won't make
+                          %% any progress during this test.
+                          _DryRun = test_dirty_scheduler_exit(),
 			  Start = erlang:monotonic_time(millisecond),
                           ok = test_dirty_scheduler_exit(),
 			  End = erlang:monotonic_time(millisecond),
@@ -171,19 +176,18 @@ test_dse(N,Pids) ->
     test_dse(N-1,[Pid|Pids]).
 
 kill_dse([],Killed) ->
-    wait_dse(Killed);
+    wait_dse(Killed, ok);
 kill_dse([Pid|Pids],AlreadyKilled) ->
     exit(Pid,kill),
     kill_dse(Pids,[Pid|AlreadyKilled]).
 
-wait_dse([]) ->
-    ok;
-wait_dse([Pid|Pids]) ->
+wait_dse([], Result) ->
+    Result;
+wait_dse([Pid|Pids], Result) ->
     receive
-        {'EXIT',Pid,Reason} ->
-	    killed = Reason
-    end,
-    wait_dse(Pids).
+        {'EXIT', Pid, killed} -> wait_dse(Pids, Result);
+        {'EXIT', Pid, _Other} -> wait_dse(Pids, failed)
+    end.
 
 dirty_call_while_terminated(Config) when is_list(Config) ->
     Me = self(),
@@ -287,9 +291,9 @@ access_dirty_heap(Dirty, RGL, N, R) ->
 %% dirty NIF where the main lock is needed for that access do not get
 %% blocked. Each test passes its pid to dirty_sleeper, which sends a
 %% 'ready' message when it's running on a dirty scheduler and just before
-%% it starts a 6 second sleep. When it receives the message, it verifies
+%% it starts a 2 second sleep. When it receives the message, it verifies
 %% that access to the dirty process is as it expects.  After the dirty
-%% process finishes its 6 second sleep but before it returns from the dirty
+%% process finishes its 2 second sleep but before it returns from the dirty
 %% scheduler, it sends a 'done' message. If the tester already received
 %% that message, the test fails because it means attempting to access the
 %% dirty process waited for that process to return to a regular scheduler,
@@ -353,7 +357,7 @@ dirty_process_trace(Config) when is_list(Config) ->
 			      error(missing_trace_return_message)
 		      end
 	      after
-		  6500 ->
+		  2500 ->
 		      error(missing_done_message)
 	      end,
 	      ok
@@ -380,7 +384,7 @@ code_purge(Config) when is_list(Config) ->
     Start = erlang:monotonic_time(),
     {Pid1, Mon1} = spawn_monitor(fun () ->
 				       dirty_code_test:func(fun () ->
-								    %% Sleep for 6 seconds
+								    %% Sleep for 2 seconds
 								    %% in dirty nif...
 								    dirty_sleeper()
 							    end)
@@ -388,7 +392,7 @@ code_purge(Config) when is_list(Config) ->
     {module, dirty_code_test} = erlang:load_module(dirty_code_test, Bin),
     {Pid2, Mon2} = spawn_monitor(fun () ->
 				       dirty_code_test:func(fun () ->
-								    %% Sleep for 6 seconds
+								    %% Sleep for 2 seconds
 								    %% in dirty nif...
 								    dirty_sleeper()
 							    end)
@@ -490,7 +494,7 @@ test_dirty_process_access(Start, Test, Finish) ->
 			 ok
 		 end
 	 after
-	     3000 ->
+	     1000 ->
 		 error(timeout)
 	 end,
     ok = Finish(NifPid).
diff --git a/erts/emulator/test/dirty_nif_SUITE_data/dirty_nif_SUITE.c b/erts/emulator/test/dirty_nif_SUITE_data/dirty_nif_SUITE.c
index 0321b9898f..2a8b999307 100644
--- a/erts/emulator/test/dirty_nif_SUITE_data/dirty_nif_SUITE.c
+++ b/erts/emulator/test/dirty_nif_SUITE_data/dirty_nif_SUITE.c
@@ -217,9 +217,9 @@ dirty_sleeper(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[])
     }
 
 #ifdef __WIN32__
-    Sleep(6000);
+    Sleep(2000);
 #else
-    sleep(6);
+    sleep(2);
 #endif
 
     if (argc == 1) {
diff --git a/erts/emulator/test/efile_SUITE.erl b/erts/emulator/test/efile_SUITE.erl
index 08d5597d78..16d581a567 100644
--- a/erts/emulator/test/efile_SUITE.erl
+++ b/erts/emulator/test/efile_SUITE.erl
@@ -19,12 +19,9 @@
 
 -module(efile_SUITE).
 -export([all/0, suite/0]).
--export([async_dist/1,
-         iter_max_files/1,
-         proc_zero_sized_files/1
-        ]).
+-export([iter_max_files/1, proc_zero_sized_files/1]).
 
--export([do_iter_max_files/2, do_async_dist/1]).
+-export([do_iter_max_files/2]).
 
 -include_lib("common_test/include/ct.hrl").
 -include_lib("stdlib/include/assert.hrl").
@@ -32,84 +29,7 @@
 suite() -> [{ct_hooks,[ts_install_cth]}].
 
 all() -> 
-    [iter_max_files, async_dist, proc_zero_sized_files].
-
-do_async_dist(Dir) ->
-    X = 100,
-    AT = erlang:system_info(thread_pool_size),
-    Keys = file_keys(Dir,AT*X,[],[]),
-    Tab = ets:new(x,[ordered_set]),
-    [ ets:insert(Tab,{N,0}) || N <- lists:seq(0,AT-1) ],
-    [ ets:update_counter(Tab,(N rem AT),1) || N <- Keys ],
-    Res = [ V || {_,V} <- ets:tab2list(Tab) ],
-    ets:delete(Tab),
-    {Res, sdev(Res)/X}.
-
-sdev(List) ->
-    Len = length(List),
-    Mean = lists:sum(List)/Len,
-    math:sqrt(lists:sum([ (X - Mean) * (X - Mean) || X <- List ]) / Len).
-
-file_keys(_,0,FdList,FnList) ->
-    [ file:close(FD) || FD <- FdList ],
-    [ file:delete(FN) || FN <- FnList ],
-    [];
-file_keys(Dir,Num,FdList,FnList) ->
-    Name = "dummy"++integer_to_list(Num),
-    FN = filename:join([Dir,Name]),
-    case file:open(FN,[write,raw]) of
-        {ok,FD} ->
-            {file_descriptor,prim_file,{Port,_}} = FD,
-            <<X:32/integer-big>> = 
-            iolist_to_binary(erlang:port_control(Port,$K,[])),
-            [X | file_keys(Dir,Num-1,[FD|FdList],[FN|FnList])];
-        {error,_} ->
-            % Try freeing up FD's if there are any
-            case FdList of 
-                [] ->
-                    exit({cannot_open_file,FN});
-                _ ->
-                    [ file:close(FD) || FD <- FdList ],
-                    [ file:delete(F) || F <- FnList ],
-                    file_keys(Dir,Num,[],[])
-            end
-    end.
-
-%% Check that the distribution of files over async threads is fair
-async_dist(Config) when is_list(Config) ->
-    DataDir = proplists:get_value(data_dir,Config),
-    Dir = filename:dirname(code:which(?MODULE)),
-    AsyncSizes = [7,10,100,255,256,64,63,65],
-    Max = 0.5,
-
-    lists:foreach(fun(Size) ->
-                          {ok,Node} = 
-                          test_server:start_node
-                          (test_iter_max_files,slave,
-                           [{args,
-                             "+A "++integer_to_list(Size)++
-                             " -pa " ++ Dir}]),
-                          {Distr,SD} = rpc:call(Node,?MODULE,do_async_dist,
-                                                [DataDir]),
-                          test_server:stop_node(Node),
-                          if
-                              SD > Max ->
-                                  io:format("Bad async queue distribution for "
-                                            "~p async threads:~n"
-                                            "    Standard deviation is ~p~n"
-                                            "    Key distribution:~n    ~lp~n",
-                                            [Size,SD,Distr]),
-                                  exit({bad_async_dist,Size,SD,Distr});
-                              true ->
-                                  io:format("OK async queue distribution for "
-                                            "~p async threads:~n"
-                                            "    Standard deviation is ~p~n"
-                                            "    Key distribution:~n    ~lp~n",
-                                            [Size,SD,Distr]),
-                                  ok
-                          end
-                  end, AsyncSizes),
-    ok.
+    [iter_max_files, proc_zero_sized_files].
 
 %%
 %% Open as many files as possible. Do this several times and check 
@@ -117,6 +37,12 @@ async_dist(Config) when is_list(Config) ->
 %%
 
 iter_max_files(Config) when is_list(Config) ->
+    case os:type() of
+        {win32, _} -> {skip, "Windows lacks a hard limit on file handles"};
+        _ -> iter_max_files_1(Config)
+    end.
+
+iter_max_files_1(Config) ->
     DataDir = proplists:get_value(data_dir,Config),
     TestFile = filename:join(DataDir, "existing_file"),
     N = 10,
diff --git a/erts/emulator/test/nif_SUITE.erl b/erts/emulator/test/nif_SUITE.erl
index bec2291867..2afeb7e4cf 100644
--- a/erts/emulator/test/nif_SUITE.erl
+++ b/erts/emulator/test/nif_SUITE.erl
@@ -3037,14 +3037,17 @@ nif_ioq(Config) ->
          {enqbraw,a},
          {enqbraw,a, 5},
          {peek,   a},
+         {peek_head,  a},
          {deq,    a, 42},
 
          %% Test enqv
          {enqv,   a, 2, 100},
+         {peek_head,  a},
          {deq,    a, all},
 
          %% This skips all elements but one in the iolist
          {enqv,   a, 5, iolist_size(nif_ioq_payload(5)) - 1},
+         {peek_head,  a},
          {peek,   a},
 
          %% Test to enqueue a bunch of refc binaries
@@ -3074,9 +3077,13 @@ nif_ioq(Config) ->
 
     Q = ioq_nif(create),
 
+    false = ioq_nif(peek_head, Q),
+
     {'EXIT', {badarg, _}} = (catch ioq_nif(deq, Q, 1)),
     {'EXIT', {badarg, _}} = (catch ioq_nif(enqv, Q, 1, 1234)),
 
+    false = ioq_nif(peek_head, Q),
+
     {'EXIT', {badarg, _}} = (catch ioq_nif(enqv, Q, [atom_in_list], 0)),
     {'EXIT', {badarg, _}} = (catch ioq_nif(enqv, Q, [make_ref()], 0)),
     {'EXIT', {badarg, _}} = (catch ioq_nif(enqv, Q, [256], 0)),
@@ -3085,6 +3092,8 @@ nif_ioq(Config) ->
     {'EXIT', {badarg, _}} = (catch ioq_nif(enqv, Q, [1 bsl 64], 0)),
     {'EXIT', {badarg, _}} = (catch ioq_nif(enqv, Q, [{tuple}], 0)),
 
+    false = ioq_nif(peek_head, Q),
+
     {'EXIT', {badarg, _}} = (catch ioq_nif(inspect,  [atom_in_list], use_stack)),
     {'EXIT', {badarg, _}} = (catch ioq_nif(inspect,  [make_ref()], no_stack)),
     {'EXIT', {badarg, _}} = (catch ioq_nif(inspect,  [256], use_stack)),
@@ -3145,6 +3154,20 @@ nif_ioq_run([{peek, Name} = H|T], State) ->
 
     true = iolist_to_binary(B) == iolist_to_binary(Data),
     nif_ioq_run(T, State);
+nif_ioq_run([{peek_head, Name} = H|T], State) ->
+    #{ q := IOQ, b := B } = maps:get(Name, State),
+    RefData = iolist_to_binary(B),
+
+    ct:log("~p", [H]),
+
+    {true, QueueHead} = ioq_nif(peek_head, IOQ),
+    true = byte_size(QueueHead) > 0,
+
+    {RefHead, _Tail} = split_binary(RefData, byte_size(QueueHead)),
+
+    true = QueueHead =:= RefHead,
+
+    nif_ioq_run(T, State);
 nif_ioq_run([{deq, Name, all}|T], State) ->
     #{ q := IOQ, b := B } = maps:get(Name, State),
     Size = ioq_nif(size, IOQ),
diff --git a/erts/emulator/test/nif_SUITE_data/nif_SUITE.c b/erts/emulator/test/nif_SUITE_data/nif_SUITE.c
index 79560a38aa..fa9ae1015c 100644
--- a/erts/emulator/test/nif_SUITE_data/nif_SUITE.c
+++ b/erts/emulator/test/nif_SUITE_data/nif_SUITE.c
@@ -3403,6 +3403,15 @@ static ERL_NIF_TERM ioq(ErlNifEnv* env, int argc, const ERL_NIF_TERM argv[])
 		return enif_make_badarg(env);
 	    else
 		return enif_make_atom(env, "true");
+        } else if (enif_is_identical(argv[0], enif_make_atom(env, "peek_head"))) {
+            ERL_NIF_TERM head_term;
+
+            if(enif_ioq_peek_head(env, ioq->q, NULL, &head_term)) {
+                return enif_make_tuple2(env,
+                    enif_make_atom(env, "true"), head_term);
+            }
+
+            return enif_make_atom(env, "false");
         } else if (enif_is_identical(argv[0], enif_make_atom(env, "peek"))) {
             int iovlen, num, i, off = 0;
             SysIOVec *iov = enif_ioq_peek(ioq->q, &iovlen);
diff --git a/erts/preloaded/ebin/erl_prim_loader.beam b/erts/preloaded/ebin/erl_prim_loader.beam
index 198032c18d..2d86dd3843 100644
--- a/erts/preloaded/ebin/erl_prim_loader.beam
+++ b/erts/preloaded/ebin/erl_prim_loader.beam
diff --git a/erts/preloaded/ebin/init.beam b/erts/preloaded/ebin/init.beam
index fb4fc67148..a2dd41b435 100644
--- a/erts/preloaded/ebin/init.beam
+++ b/erts/preloaded/ebin/init.beam
diff --git a/erts/preloaded/ebin/prim_buffer.beam b/erts/preloaded/ebin/prim_buffer.beam
new file mode 100644
index 0000000000..06d9276247
--- /dev/null
+++ b/erts/preloaded/ebin/prim_buffer.beam
diff --git a/erts/preloaded/ebin/prim_file.beam b/erts/preloaded/ebin/prim_file.beam
index eb326ac4b6..902b0945c6 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 32b104ae95..a42f45c180 100644
--- a/erts/preloaded/ebin/prim_inet.beam
+++ b/erts/preloaded/ebin/prim_inet.beam
diff --git a/erts/preloaded/src/Makefile b/erts/preloaded/src/Makefile
index edb9f35258..a2872082f2 100644
--- a/erts/preloaded/src/Makefile
+++ b/erts/preloaded/src/Makefile
@@ -36,6 +36,7 @@ include $(ERL_TOP)/lib/kernel/vsn.mk
 PRE_LOADED_ERL_MODULES = \
 	erl_prim_loader \
 	init \
+	prim_buffer \
 	prim_file \
 	prim_inet \
 	zlib \
@@ -116,7 +117,7 @@ prim_eval.beam: prim_eval.S prim_eval.abstr
 
 # Include dependencies -- list below added by PaN
 $(EBIN)/erl_prim_loader.beam: $(KERNEL_SRC)/inet_boot.hrl $(KERNEL_INCLUDE)/file.hrl
-$(EBIN)/prim_file.beam: $(KERNEL_INCLUDE)/file.hrl
+$(EBIN)/prim_file.beam: $(KERNEL_SRC)/file_int.hrl $(KERNEL_INCLUDE)/file.hrl
 $(EBIN)/prim_inet.beam: $(KERNEL_SRC)/inet_int.hrl $(KERNEL_INCLUDE)/inet_sctp.hrl
 $(EBIN)/prim_zip.beam: zip_internal.hrl $(KERNEL_INCLUDE)/file.hrl $(STDLIB_INCLUDE)/zip.hrl
 $(EBIN)/init.erl: $(KERNEL_INCLUDE)/file.hrl
diff --git a/erts/preloaded/src/erl_prim_loader.erl b/erts/preloaded/src/erl_prim_loader.erl
index 1d09aeded9..d16fe5a0bc 100644
--- a/erts/preloaded/src/erl_prim_loader.erl
+++ b/erts/preloaded/src/erl_prim_loader.erl
@@ -151,9 +151,8 @@ start_inet(Parent) ->
     loop(State, Parent, []).
 
 start_efile(Parent) ->
-    {ok, Port} = prim_file:start(),
     %% Check that we started in a valid directory.
-    case prim_file:get_cwd(Port) of
+    case prim_file:get_cwd() of
 	{error, _} ->
 	    %% At this point in the startup, we have no error_logger at all.
 	    Report = "Invalid current directory or invalid filename "
@@ -165,7 +164,7 @@ start_efile(Parent) ->
     end,
     PS = prim_init(),
     State = #state {loader = efile,
-                    data = Port,
+                    data = noport,
                     timeout = ?EFILE_IDLE_TIMEOUT,
                     prim_state = PS},
     loop(State, Parent, []).
@@ -401,12 +400,12 @@ handle_get_cwd(State = #state{loader = inet}, Drive) ->
     ?SAFE2(inet_get_cwd(State, Drive), State).
     
 handle_stop(State = #state{loader = efile}) ->
-    efile_stop_port(State);
+    State;
 handle_stop(State = #state{loader = inet}) ->
     inet_stop_port(State).
 
-handle_exit(State = #state{loader = efile}, Who, Reason) ->
-    efile_exit_port(State, Who, Reason);
+handle_exit(State = #state{loader = efile}, _Who, _Reason) ->
+    State;
 handle_exit(State = #state{loader = inet}, Who, Reason) ->
     inet_exit_port(State, Who, Reason).
 
@@ -475,15 +474,6 @@ efile_get_cwd(#state{prim_state = PS} = State, Drive) ->
     {Res, PS2} = prim_get_cwd(PS, Drive),
     {Res, State#state{prim_state = PS2}}.
 
-efile_stop_port(#state{data=Port}=State) ->
-    prim_file:close(Port),
-    State#state{data=noport}.
-
-efile_exit_port(State, Port, Reason) when State#state.data =:= Port ->
-    exit({port_died,Reason});
-efile_exit_port(State, _Port, _Reason) ->
-    State.
-
 efile_timeout_handler(State, _Parent) ->
     prim_purge_cache(),
     State.
diff --git a/erts/preloaded/src/erts.app.src b/erts/preloaded/src/erts.app.src
index e2ab8686d2..338f168158 100644
--- a/erts/preloaded/src/erts.app.src
+++ b/erts/preloaded/src/erts.app.src
@@ -28,6 +28,7 @@
 		init,
 		otp_ring0,
 		erts_code_purger,
+		prim_buffer,
 		prim_eval,
 		prim_file,
 		prim_inet,
diff --git a/erts/preloaded/src/init.erl b/erts/preloaded/src/init.erl
index 34a9f6b8b9..679a2241d2 100644
--- a/erts/preloaded/src/init.erl
+++ b/erts/preloaded/src/init.erl
@@ -200,10 +200,11 @@ boot(BootArgs) ->
     register(init, self()),
     process_flag(trap_exit, true),
 
-    %% Load the zlib nif
+    %% Load the static nifs
     zlib:on_load(),
-    %% Load the tracer nif
     erl_tracer:on_load(),
+    prim_buffer:on_load(),
+    prim_file:on_load(),
 
     {Start0,Flags,Args} = parse_boot_args(BootArgs),
     %% We don't get to profile parsing of BootArgs
diff --git a/erts/preloaded/src/prim_buffer.erl b/erts/preloaded/src/prim_buffer.erl
new file mode 100644
index 0000000000..e0d35a6792
--- /dev/null
+++ b/erts/preloaded/src/prim_buffer.erl
@@ -0,0 +1,140 @@
+%%
+%% %CopyrightBegin%
+%%
+%% Copyright Ericsson AB 2017. 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(prim_buffer).
+
+-export([on_load/0]).
+
+%% This is a mutable binary buffer that helps break out buffering logic from
+%% NIFs/drivers, which is often the only thing that prevents the C code from
+%% being reduced to bare system call wrappers.
+%%
+%% All operations in this file are thread-unsafe and risk crashing the emulator
+%% if you're not careful.
+
+-export([new/0, size/1, wipe/1, read/2, read_iovec/2, write/2, skip/2]).
+
+-export([find_byte_index/2]).
+
+-export([try_lock/1, unlock/1]).
+
+-type prim_buffer() :: term().
+
+%% Controls when to copy rather than extract sub-binaries from the buffer,
+%% reducing the risk of small reads keeping a large binary alive.
+-define(COPYING_READ_LIMIT, 512).
+
+%% Reads that fit into heap binaries are always copied since the cost of
+%% peeking binaries that short is largely equivalent to copying.
+-define(ERL_ONHEAP_BIN_LIMIT, 64).
+
+on_load() ->
+    case erlang:load_nif(atom_to_list(?MODULE), 0) of
+        ok -> ok
+    end.
+
+-spec new() -> prim_buffer().
+new() ->
+    erlang:nif_error(undef).
+
+-spec size(Buffer :: prim_buffer()) -> non_neg_integer().
+size(_Buffer) ->
+    erlang:nif_error(undef).
+
+%% Reads data as a binary from the front of the buffer. This will almost always
+%% result in copying so it should be avoided unless you absolutely must have a
+%% binary.
+-spec read(Buffer :: prim_buffer(), Size :: non_neg_integer()) -> binary().
+read(Buffer, Size) when Size =< ?ERL_ONHEAP_BIN_LIMIT ->
+    copying_read(Buffer, Size);
+read(Buffer, Size) when Size > ?ERL_ONHEAP_BIN_LIMIT ->
+    iolist_to_binary(read_iovec(Buffer, Size)).
+
+%% Reads data as an erlang:iovec() binary from the front of the buffer,
+%% avoiding copying if reasonable.
+-spec read_iovec(Buffer, Size) -> IOVec when
+      Buffer :: prim_buffer(),
+      Size :: non_neg_integer(),
+      IOVec :: erlang:iovec().
+read_iovec(Buffer, Size) when Size =< ?ERL_ONHEAP_BIN_LIMIT ->
+    [copying_read(Buffer, Size)];
+read_iovec(Buffer, Size) when Size > ?ERL_ONHEAP_BIN_LIMIT ->
+    Head = peek_head(Buffer),
+    HeadSize = byte_size(Head),
+    if
+        (HeadSize - Size) > ?COPYING_READ_LIMIT, Size =< ?COPYING_READ_LIMIT ->
+            [copying_read(Buffer, Size)];
+        HeadSize > Size ->
+            skip(Buffer, Size),
+            {First, _Rest} = split_binary(Head, Size),
+            [First];
+        HeadSize < Size ->
+            skip(Buffer, HeadSize),
+            [Head | read_iovec(Buffer, Size - HeadSize)];
+        HeadSize =:= Size ->
+            skip(Buffer, Size),
+            [Head]
+    end.
+
+%% Writes an erlang:iovec() to the back of the buffer.
+-spec write(Buffer :: prim_buffer(), IOVec :: erlang:iovec()) -> ok.
+write(_Buffer, _IOVec) ->
+    erlang:nif_error(undef).
+
+%% Removes data from the front of the buffer without reading it.
+-spec skip(Buffer :: prim_buffer(), Size :: non_neg_integer()) -> ok.
+skip(_Buffer, _Size) ->
+    erlang:nif_error(undef).
+
+-spec wipe(Buffer :: prim_buffer()) -> ok.
+wipe(Buffer) ->
+    skip(Buffer, prim_buffer:size(Buffer)).
+
+%% Finds the start-index of the first occurence of Needle, for implementing
+%% read_line and similar.
+-spec find_byte_index(Buffer, Needle) -> Result when
+      Buffer :: prim_buffer(),
+      Needle :: non_neg_integer(),
+      Result :: {ok, non_neg_integer()} |
+                not_found.
+find_byte_index(_Buffer, _Needle) ->
+    erlang:nif_error(undef).
+
+%% Attempts to take a unique lock on the buffer. Failure handling is left to
+%% the user.
+-spec try_lock(Buffer :: prim_buffer()) -> acquired | busy.
+try_lock(_Buffer) ->
+    erlang:nif_error(undef).
+
+-spec unlock(Buffer :: prim_buffer()) -> ok.
+unlock(_Buffer) ->
+    erlang:nif_error(undef).
+
+%% Unexported helper functions:
+
+%% Reads data from the front of the buffer, returning a copy of the data to
+%% avoid holding references.
+-spec copying_read(Buffer :: prim_buffer(), Size :: non_neg_integer()) -> binary().
+copying_read(_Buffer, _Size) ->
+    erlang:nif_error(undef).
+
+%% Returns the binary at the front of the buffer without modifying the buffer.
+-spec peek_head(Buffer :: prim_buffer()) -> binary().
+peek_head(_Buffer) ->
+    erlang:nif_error(undef).
diff --git a/erts/preloaded/src/prim_file.erl b/erts/preloaded/src/prim_file.erl
index ab5359ebbc..35042a7c72 100644
--- a/erts/preloaded/src/prim_file.erl
+++ b/erts/preloaded/src/prim_file.erl
@@ -1,7 +1,7 @@
 %%
 %% %CopyrightBegin%
 %%
-%% Copyright Ericsson AB 2000-2016. All Rights Reserved.
+%% Copyright Ericsson AB 2000-2017. 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.
@@ -19,147 +19,45 @@
 %%
 -module(prim_file).
 
-%% Interface module to the file driver.
+-export([on_load/0]).
 
+-export([open/2, close/1,
+         sync/1, datasync/1, truncate/1, advise/4, allocate/3,
+         read_line/1, read/2, write/2, position/2,
+         pread/2, pread/3, pwrite/2, pwrite/3]).
 
+%% OTP internal.
+-export([ipread_s32bu_p32bu/3, sendfile/8, altname/1, get_handle/1]).
 
-%%% Interface towards a single file's contents. Uses ?FD_DRV.
+-export([read_file/1, write_file/2]).
 
-%% Generic file contents operations
--export([open/2, close/1, datasync/1, sync/1, advise/4, position/2, truncate/1,
-	 write/2, pwrite/2, pwrite/3, read/2, read_line/1, pread/2, pread/3,
-	 copy/3, sendfile/8, allocate/3]).
+-export([read_link/1, read_link_all/1,
+         read_link_info/1, read_link_info/2,
+         read_file_info/1, read_file_info/2,
+         write_file_info/2, write_file_info/3]).
 
-%% Specialized file operations
--export([open/1, open/3]).
--export([read_file/1, read_file/2, write_file/2]).
--export([ipread_s32bu_p32bu/3]).
+-export([list_dir/1, list_dir_all/1]).
 
+-export([get_cwd/0, get_cwd/1, set_cwd/1,
+         delete/1, rename/2,
+         make_dir/1, del_dir/1,
+         make_link/2, make_symlink/2]).
 
-
-%%% Interface towards file system and metadata. Uses ?DRV.
-
-%% Takes an optional port (opens a ?DRV port per default) as first argument.
--export([get_cwd/0, get_cwd/1, get_cwd/2, 
-	 set_cwd/1, set_cwd/2,
-	 delete/1, delete/2, 
-	 rename/2, rename/3, 
-	 make_dir/1, make_dir/2,
-	 del_dir/1, del_dir/2,
-	 read_file_info/1, read_file_info/2, read_file_info/3,
-	 altname/1, altname/2,
-	 write_file_info/2, write_file_info/3, write_file_info/4,
-	 make_link/2, make_link/3,
-	 make_symlink/2, make_symlink/3,
-	 read_link/1, read_link/2, read_link_all/1, read_link_all/2,
-	 read_link_info/1, read_link_info/2, read_link_info/3,
-	 list_dir/1, list_dir/2, list_dir_all/1, list_dir_all/2]).
-%% How to start and stop the ?DRV port.
--export([start/0, stop/1]).
-
-%% Debug exports
--export([open_int/4, open_mode/1, open_mode/4]).
-
-%%%-----------------------------------------------------------------
-%%% Includes and defines
-
--include("file.hrl").
-
--define(DRV,    "efile").
--define(FD_DRV, "efile").
-
+-define(MIN_READLINE_SIZE, 256).
 -define(LARGEFILESIZE, (1 bsl 63)).
 
-%% Driver commands
--define(FILE_OPEN,             1).
--define(FILE_READ,             2).
--define(FILE_LSEEK,            3).
--define(FILE_WRITE,            4).
--define(FILE_FSTAT,            5).
--define(FILE_PWD,              6).
--define(FILE_READDIR,          7).
--define(FILE_CHDIR,            8).
--define(FILE_FSYNC,            9).
--define(FILE_MKDIR,            10).
--define(FILE_DELETE,           11).
--define(FILE_RENAME,           12).
--define(FILE_RMDIR,            13).
--define(FILE_TRUNCATE,         14).
--define(FILE_READ_FILE,        15).
--define(FILE_WRITE_INFO,       16).
--define(FILE_LSTAT,            19).
--define(FILE_READLINK,         20).
--define(FILE_LINK,             21).
--define(FILE_SYMLINK,          22).
--define(FILE_CLOSE,            23).
--define(FILE_PWRITEV,          24).
--define(FILE_PREADV,           25).
--define(FILE_SETOPT,           26).
--define(FILE_IPREAD,           27).
--define(FILE_ALTNAME,          28).
--define(FILE_READ_LINE,        29).
--define(FILE_FDATASYNC,        30).
--define(FILE_ADVISE,           31).
--define(FILE_SENDFILE,         32).
--define(FILE_ALLOCATE,         33).
-
-%% Driver responses
--define(FILE_RESP_OK,          0).
--define(FILE_RESP_ERROR,       1).
--define(FILE_RESP_DATA,        2).
--define(FILE_RESP_NUMBER,      3).
--define(FILE_RESP_INFO,        4).
--define(FILE_RESP_NUMERR,      5).
--define(FILE_RESP_LDATA,       6).
--define(FILE_RESP_N2DATA,      7).
--define(FILE_RESP_EOF,         8).
--define(FILE_RESP_FNAME,       9).
--define(FILE_RESP_ALL_DATA,   10).
--define(FILE_RESP_LFNAME,     11).
-
-%% Open modes for the driver's open function.
--define(EFILE_MODE_READ,       1).
--define(EFILE_MODE_WRITE,      2).
--define(EFILE_MODE_READ_WRITE, 3).  
--define(EFILE_MODE_APPEND,     4).
--define(EFILE_COMPRESSED,      8).
--define(EFILE_MODE_EXCL,       16).
-%% Note: bit 5 (32) is used internally for VxWorks
--define(EFILE_MODE_SYNC,       64).
-
-%% Use this mask to get just the mode bits to be passed to the driver.
--define(EFILE_MODE_MASK, 127).
-
-%% Seek modes for the driver's seek function.
--define(EFILE_SEEK_SET, 0).
--define(EFILE_SEEK_CUR, 1).
--define(EFILE_SEEK_END, 2).
-
-%% Options
--define(FILE_OPT_DELAYED_WRITE, 0).
--define(FILE_OPT_READ_AHEAD,    1).
-
-%% IPREAD variants
--define(IPREAD_S32BU_P32BU, 0).
-
-%% POSIX file advises
--define(POSIX_FADV_NORMAL,     0).
--define(POSIX_FADV_RANDOM,     1).
--define(POSIX_FADV_SEQUENTIAL, 2).
--define(POSIX_FADV_WILLNEED,   3).
--define(POSIX_FADV_DONTNEED,   4).
--define(POSIX_FADV_NOREUSE,    5).
-
-%% Sendfile flags
--define(EFILE_SENDFILE_USE_THREADS, 1).
+-export([copy/3]).
 
+-include("file_int.hrl").
+
+-type prim_file_ref() :: term().
 
 %%% BIFs
 
 -export([internal_name2native/1,
          internal_native2name/1,
          internal_normalize_utf8/1,
-	 is_translatable/1]).
+         is_translatable/1]).
 
 -type prim_file_name() :: string() | unicode:unicode_binary().
 -type prim_file_name_error() :: 'error' | 'ignore' | 'warning'.
@@ -185,1316 +83,733 @@ internal_normalize_utf8(_) ->
 is_translatable(_) ->
     erlang:nif_error(undefined).
 
-%%% End of BIFs
-
-%%%-----------------------------------------------------------------
-%%% Functions operating on a file through a handle. ?FD_DRV.
-%%%
-%%% Generic file contents operations.
-%%%
-%%% Supposed to be called by applications through module file.
-
-
-%% Opens a file using the driver port Port. Returns {error, Reason}
-%% | {ok, FileDescriptor}
-open(Port, File, ModeList) when is_port(Port), 
-                                (is_list(File) orelse is_binary(File)), 
-                                is_list(ModeList) ->
-    case open_mode(ModeList) of
-	{Mode, _Portopts, _Setopts} ->
-	    open_int(Port, File, Mode, []);
-	Reason ->
-	    {error, Reason}
-    end;
-open(_,_,_) ->
-    {error, badarg}.
-
-%% Opens a file. Returns {error, Reason} | {ok, FileDescriptor}.
-open(File, ModeList) when (is_list(File) orelse is_binary(File)), 
-			  is_list(ModeList) ->
-    case open_mode(ModeList) of
-	{Mode, Portopts, Setopts} ->
-	    open_int({?FD_DRV, Portopts},File, Mode, Setopts);
-	Reason ->
-	    {error, Reason}
-    end;
-open(_, _) ->
-    {error, badarg}.
+%%
 
-%% Opens a port that can be used for open/3 or read_file/2.
-%% Returns {ok, Port} | {error, Reason}.
-open(Portopts) when is_list(Portopts) ->
-    drv_open(?FD_DRV, [binary|Portopts]);
-open(_) ->
-    {error, badarg}.
+%% Returns {error, Reason} | {ok, BytesCopied}
+copy(#file_descriptor{module = ?MODULE} = Source,
+     #file_descriptor{module = ?MODULE} = Dest,
+     Length)
+  when is_integer(Length), Length >= 0;
+       is_atom(Length) ->
+    %% XXX Should be moved down to the driver for optimization.
+    file:copy_opened(Source, Dest, Length).
 
-open_int({Driver, Portopts}, File, Mode, Setopts) ->
-    case drv_open(Driver, Portopts) of
-	{ok, Port} ->
-	    open_int(Port, File, Mode, Setopts);
-	{error, _} = Error ->
-	    Error
-    end;
-open_int(Port, File, Mode, Setopts) ->
-    M = Mode band ?EFILE_MODE_MASK,
-    case drv_command(Port, [<<?FILE_OPEN, M:32>>, pathname(File)]) of
-	{ok, Number} ->
-	    open_int_setopts(Port, Number, Setopts);
-	Error ->
-	    drv_close(Port),
-	    Error
-    end.
+on_load() ->
+    ok = erlang:load_nif(atom_to_list(?MODULE), 0).
 
-open_int_setopts(Port, Number, []) ->
-    {ok, #file_descriptor{module = ?MODULE, data = {Port, Number}}};    
-open_int_setopts(Port, Number, [Cmd | Tail]) ->
-    case drv_command(Port, Cmd) of
-	ok ->
-	    open_int_setopts(Port, Number, Tail);
-	Error ->
-	    drv_close(Port),
-	    Error
+open(Name, Modes) ->
+    %% The try/catch pattern seen here is used throughout the file to adhere to
+    %% the public file interface, which has leaked through for ages because of
+    %% "raw files."
+    try open_nif(encode_path(Name), Modes) of
+        {ok, Ref} -> {ok, make_fd(Ref, Modes)};
+        {error, Reason} -> {error, Reason}
+    catch
+        error:badarg -> {error, badarg}
     end.
 
+make_fd(FRef, Modes) ->
+    #file_descriptor{module = ?MODULE, data = build_fd_data(FRef, Modes) }.
 
-
-%% Returns ok.
-
-close(#file_descriptor{module = ?MODULE, data = {Port, _}}) ->
-    case drv_command(Port, <<?FILE_CLOSE>>) of
-	ok ->
-	    drv_close(Port);
-	Error ->
-	    Error
-    end;
-%% Closes a port opened with open/1.
-close(Port) when is_port(Port) ->
-    drv_close(Port).
-
--define(ADVISE(Offs, Len, Adv),
-	<<?FILE_ADVISE, Offs:64/signed, Len:64/signed,
-	  Adv:32/signed>>).
-
-%% Returns {error, Reason} | ok.
-advise(#file_descriptor{module = ?MODULE, data = {Port, _}},
-       Offset, Length, Advise) ->
-    case Advise of
-	normal ->
-	    Cmd = ?ADVISE(Offset, Length, ?POSIX_FADV_NORMAL),
-	    drv_command(Port, Cmd);
-	random ->
-	    Cmd = ?ADVISE(Offset, Length, ?POSIX_FADV_RANDOM),
-	    drv_command(Port, Cmd);
-	sequential ->
-	    Cmd = ?ADVISE(Offset, Length, ?POSIX_FADV_SEQUENTIAL),
-	    drv_command(Port, Cmd);
-	will_need ->
-	    Cmd = ?ADVISE(Offset, Length, ?POSIX_FADV_WILLNEED),
-	    drv_command(Port, Cmd);
-	dont_need ->
-	    Cmd = ?ADVISE(Offset, Length, ?POSIX_FADV_DONTNEED),
-	    drv_command(Port, Cmd);
-	no_reuse ->
-	    Cmd = ?ADVISE(Offset, Length, ?POSIX_FADV_NOREUSE),
-	    drv_command(Port, Cmd);
-	_ ->
-	    {error, einval}
+close(Fd) ->
+    try
+        #{ handle := FRef } = get_fd_data(Fd),
+        close_nif(FRef)
+    catch
+        error:badarg -> {error, badarg}
     end.
 
-%% Returns {error, Reason} | ok.
-allocate(#file_descriptor{module = ?MODULE, data = {Port, _}}, Offset, Length) ->
-    Cmd = <<?FILE_ALLOCATE, Offset:64/signed, Length:64/signed>>,
-    drv_command(Port, Cmd).
-
-%% Returns {error, Reason} | ok.
-write(#file_descriptor{module = ?MODULE, data = {Port, _}}, Bytes) ->
-    case drv_command_nt(Port, [?FILE_WRITE,erlang:dt_prepend_vm_tag_data(Bytes)],undefined) of
-	{ok, _Size} ->
-	    ok;
-	Error ->
-	    Error
+read(Fd, Size) ->
+    try
+        #{ handle := FRef,
+           r_ahead_size := RASz,
+           r_buffer := RBuf } = get_fd_data(Fd),
+        read_1(FRef, RBuf, prim_buffer:size(RBuf), RASz, Size)
+    catch
+        error:badarg -> {error, badarg}
     end.
 
-%% Returns ok | {error, {WrittenCount, Reason}}
-pwrite(#file_descriptor{module = ?MODULE, data = {Port, _}}, L)
-  when is_list(L) ->
-    pwrite_int(Port, L, 0, [], []).
-
-pwrite_int(_, [], 0, [], []) ->
-    ok;
-pwrite_int(Port, [], N, Spec, Data) ->
-    Header = list_to_binary([?FILE_PWRITEV, erlang:dt_prepend_vm_tag_data(<<N:32>>) | reverse(Spec)]),
-    case drv_command_nt(Port, [Header | reverse(Data)], undefined) of
-	{ok, _Size} ->
-	    ok;
-	Error ->
-	    Error
-    end;
-pwrite_int(Port, [{Offs, Bytes} | T], N, Spec, Data)
-  when is_integer(Offs) ->
-    if
-	-(?LARGEFILESIZE) =< Offs, Offs < ?LARGEFILESIZE ->
-	    pwrite_int(Port, T, N, Spec, Data, Offs, Bytes);
-	true ->
-	    {error, einval}
+-spec read_1(FRef, RBuf, RBufSz, RASz, RSz) -> Result when
+      FRef :: prim_file_ref(),
+      RBuf :: term(),
+      RBufSz :: non_neg_integer(),
+      RASz :: non_neg_integer(),
+      RSz :: non_neg_integer(),
+      Result :: eof | {ok, binary()} | {error, Reason :: atom()}.
+read_1(_FRef, RBuf, RBufSz, _RASz, RSz) when RBufSz >= RSz ->
+    {ok, prim_buffer:read(RBuf, RSz)};
+read_1(FRef, RBuf, RBufSz, RASz, RSz) when RBufSz > 0 ->
+    Buffered = prim_buffer:read(RBuf, RBufSz),
+    case read_1(FRef, RBuf, 0, RASz, RSz - RBufSz) of
+        {ok, Data} ->
+            {ok, <<Buffered/binary, Data/binary>>};
+        eof ->
+            {ok, Buffered};
+        {error, Reason} ->
+            {error, Reason}
     end;
-pwrite_int(_, [_|_], _N, _Spec, _Data) ->
-    {error, badarg}.
+read_1(FRef, RBuf, RBufSz, RASz, RSz) when RBufSz =:= 0 ->
+    case read_nif(FRef, RASz + RSz) of
+        {ok, Data} when byte_size(Data) > RSz ->
+            {First, Rest} = split_binary(Data, RSz),
+            prim_buffer:write(RBuf, [Rest]),
+            {ok, First};
+        {ok, Data} when byte_size(Data) =< RSz ->
+            {ok, Data};
+        eof ->
+            eof;
+        {error, Reason} ->
+            {error, Reason}
+    end.
 
-pwrite_int(Port, T, N, Spec, Data, Offs, Bin)
-  when is_binary(Bin) ->
-    Size = byte_size(Bin),
-    pwrite_int(Port, T, N+1, 
-	       [<<Offs:64/signed, Size:64>> | Spec], 
-	       [Bin | Data]);
-pwrite_int(Port, T, N, Spec, Data, Offs, Bytes) ->
-    try list_to_binary(Bytes) of
-	Bin ->
-	    pwrite_int(Port, T, N, Spec, Data, Offs, Bin)
+read_line(Fd) ->
+    try
+        #{ handle := FRef,
+           r_ahead_size := RASz,
+           r_buffer := RBuf } = get_fd_data(Fd),
+        SearchResult = prim_buffer:find_byte_index(RBuf, $\n),
+        LineSize = max(?MIN_READLINE_SIZE, RASz),
+        read_line_1(FRef, RBuf, SearchResult, LineSize)
     catch
-	error:Reason ->
-	    {error, Reason}
+        error:badarg -> {error, badarg}
     end.
 
-
-
-%% Returns {error, Reason} | ok.
-pwrite(#file_descriptor{module = ?MODULE, data = {Port, _}}, Offs, Bytes) 
-  when is_integer(Offs) ->
-    if
-	-(?LARGEFILESIZE) =< Offs, Offs < ?LARGEFILESIZE ->
-	    case pwrite_int(Port, [], 0, [], [], Offs, Bytes) of
-		{error, {_, Reason}} ->
-		    {error, Reason};
-		Result ->
-		    Result
-	    end;
-	true ->
-	    {error, einval}
+-spec read_line_1(FRef, RBuf, SearchResult, LineSize) -> Result when
+      FRef :: prim_file_ref(),
+      RBuf :: term(),
+      SearchResult :: not_found | {ok, non_neg_integer()},
+      LineSize :: non_neg_integer(),
+      Result :: eof | {ok, binary()} | {error, Reason :: atom()}.
+read_line_1(FRef, RBuf, not_found, LineSize) ->
+    case read_nif(FRef, LineSize) of
+        {ok, Data} ->
+            prim_buffer:write(RBuf, [Data]),
+            SearchResult = prim_buffer:find_byte_index(RBuf, $\n),
+            read_line_1(FRef, RBuf, SearchResult, LineSize);
+        eof ->
+            case prim_buffer:size(RBuf) of
+                Size when Size > 0 -> {ok, prim_buffer:read(RBuf, Size)};
+                Size when Size =:= 0 -> eof
+            end;
+        {error, Reason} ->
+            {error, Reason}
     end;
-pwrite(#file_descriptor{module = ?MODULE}, _, _) ->
-    {error, badarg}.
-
+read_line_1(_FRef, RBuf, {ok, LFIndex}, _LineSize) ->
+    %% Translate CRLF into just LF, completely ignoring which encoding is used.
+    CRIndex = (LFIndex - 1),
+    case prim_buffer:read(RBuf, LFIndex + 1) of
+        <<Line:CRIndex/binary, "\r\n">> -> {ok, <<Line/binary, "\n">>};
+        Line -> {ok, Line}
+    end.
 
-%% Returns {error, Reason} | ok.
-datasync(#file_descriptor{module = ?MODULE, data = {Port, _}}) ->
-    drv_command(Port, [?FILE_FDATASYNC]).
-
-%% Returns {error, Reason} | ok.
-sync(#file_descriptor{module = ?MODULE, data = {Port, _}}) ->
-    drv_command(Port, [?FILE_FSYNC]).
-
-%% Returns {ok, Data} | eof | {error, Reason}.
-read_line(#file_descriptor{module = ?MODULE, data = {Port, _}}) ->
-    case drv_command(Port, <<?FILE_READ_LINE>>) of
-	{ok, {0, _Data}} ->
-	    eof;
-	{ok, {_Size, Data}} ->
-	    {ok, Data};
-	{error, enomem} ->
-	    erlang:garbage_collect(),
-	    case drv_command(Port, <<?FILE_READ_LINE>>) of
-		{ok, {0, _Data}} ->
-		    eof;
-		{ok, {_Size, Data}} ->
-		    {ok, Data};
-		Other ->
-		    Other
-	    end;
-	Error ->
-	    Error
+write(Fd, IOData) ->
+    try
+        #{ handle := FRef } = get_fd_data(Fd),
+        reset_write_position(Fd),
+        write_1(FRef, erlang:iolist_to_iovec(IOData))
+    catch
+        error:badarg -> {error, badarg}
     end.
-	
-%% Returns {ok, Data} | eof | {error, Reason}.
-read(#file_descriptor{module = ?MODULE, data = {Port, _}}, Size)
-  when is_integer(Size), 0 =< Size ->
-    if
-	Size < ?LARGEFILESIZE ->
-	    case drv_command(Port, <<?FILE_READ, Size:64>>) of
-		{ok, {0, _Data}} when Size =/= 0 ->
-		    eof;
-		{ok, {_Size, Data}} ->
-		    {ok, Data};
-		{error, enomem} ->
-		    %% Garbage collecting here might help if
-		    %% the current processes have some old binaries left.
-		    erlang:garbage_collect(),
-		    case drv_command(Port, <<?FILE_READ, Size:64>>) of
-			{ok, {0, _Data}} when Size =/= 0 ->
-			    eof;
-			{ok, {_Size, Data}} ->
-			    {ok, Data};
-			Other ->
-			    Other
-		    end;
-		Error ->
-		    Error
-	    end;
-	true ->
-	    {error, einval}
+write_1(FRef, IOVec) ->
+    case write_nif(FRef, IOVec) of
+        {continue, Remainder} ->
+            write_1(FRef, Remainder);
+        ok ->
+            ok;
+        {error, Reason} ->
+            {error, Reason}
     end.
 
-%% Returns {ok, [Data|eof, ...]} | {error, Reason}
-pread(#file_descriptor{module = ?MODULE, data = {Port, _}}, L)
-  when is_list(L) ->
-    pread_int(Port, L, 0, []).
-
-pread_int(_, [], 0, []) ->
-    {ok, []};
-pread_int(Port, [], N, Spec) ->
-    drv_command_nt(Port, [?FILE_PREADV, erlang:dt_prepend_vm_tag_data(<<0:32, N:32>>) | reverse(Spec)],undefined);
-pread_int(Port, [{Offs, Size} | T], N, Spec)
-  when is_integer(Offs), is_integer(Size), 0 =< Size ->
-    if
-	-(?LARGEFILESIZE) =< Offs, Offs < ?LARGEFILESIZE,
-	Size < ?LARGEFILESIZE ->
-	    pread_int(Port, T, N+1, [<<Offs:64/signed, Size:64>> | Spec]);
-	true ->
-	    {error, einval}
-    end;
-pread_int(_, [_|_], _N, _Spec) ->
-    {error, badarg}.
-
-
-
-%% Returns {ok, Data} | eof | {error, Reason}.
-pread(#file_descriptor{module = ?MODULE, data = {Port, _}}, Offs, Size) 
-  when is_integer(Offs), is_integer(Size), 0 =< Size ->
-    if
-	-(?LARGEFILESIZE) =< Offs, Offs < ?LARGEFILESIZE,
-	Size < ?LARGEFILESIZE ->
-	    case drv_command_nt(Port, 
-				[?FILE_PREADV, erlang:dt_prepend_vm_tag_data(<<0:32, 1:32,
-									     Offs:64/signed, Size:64>>)], undefined) of
-		{ok, [eof]} ->
-		    eof;
-		{ok, [Data]} ->
-		    {ok, Data};
-		Error ->
-		    Error
-	    end;
-	true ->
-	    {error, einval}
-    end;
-pread(#file_descriptor{module = ?MODULE, data = {_, _}}, _, _) ->
-    {error, badarg}.
-
-
-
-%% Returns {ok, Position} | {error, Reason}.
-position(#file_descriptor{module = ?MODULE, data = {Port, _}}, At) ->
-    case lseek_position(At) of
-	{Offs, Whence}
-	when -(?LARGEFILESIZE) =< Offs, Offs < ?LARGEFILESIZE ->
-	    drv_command(Port, <<?FILE_LSEEK, Offs:64/signed, Whence:32>>);
-	{_, _} ->
-	    {error, einval};
-	Reason ->
-	    {error, Reason}
+truncate(Fd) ->
+    try
+        #{ handle := FRef } = get_fd_data(Fd),
+        reset_write_position(Fd),
+        truncate_nif(FRef)
+    catch
+        error:badarg -> {error, badarg}
     end.
 
-%% Returns {error, Reason} | ok.
-truncate(#file_descriptor{module = ?MODULE, data = {Port, _}}) ->
-    drv_command(Port, <<?FILE_TRUNCATE>>).
-
-
-
-%% Returns {error, Reason} | {ok, BytesCopied}
-copy(#file_descriptor{module = ?MODULE} = Source,
-     #file_descriptor{module = ?MODULE} = Dest,
-     Length)
-  when is_integer(Length), Length >= 0;
-       is_atom(Length) ->
-    %% XXX Should be moved down to the driver for optimization.
-    file:copy_opened(Source, Dest, Length).
-
-
-
-ipread_s32bu_p32bu(#file_descriptor{module = ?MODULE,
-				    data = {_, _}} = Handle,
-		   Offs,
-		   Infinity) when is_atom(Infinity) ->
-    ipread_s32bu_p32bu(Handle, Offs, (1 bsl 31)-1);
-ipread_s32bu_p32bu(#file_descriptor{module = ?MODULE, data = {Port, _}},
-		   Offs,
-		   MaxSize)
-  when is_integer(Offs), is_integer(MaxSize) ->
-    if
-	-(?LARGEFILESIZE) =< Offs, Offs < ?LARGEFILESIZE,
-	0 =< MaxSize, MaxSize < (1 bsl 31) ->
-	    drv_command(Port, <<?FILE_IPREAD, ?IPREAD_S32BU_P32BU,
-			       Offs:64, MaxSize:32>>);
-	true ->
-	    {error, einval}
-    end;
-ipread_s32bu_p32bu(#file_descriptor{module = ?MODULE, data = {_, _}},
-		   _Offs,
-		   _MaxSize) ->
-    {error, badarg}.
+advise(Fd, Offset, Length, Advise) ->
+    try
+        #{ handle := FRef } = get_fd_data(Fd),
+        advise_nif(FRef, Offset, Length, Advise)
+    catch
+        error:badarg -> {error, badarg}
+    end.
 
+allocate(Fd, Offset, Length) ->
+    try
+        #{ handle := FRef } = get_fd_data(Fd),
+        allocate_nif(FRef, Offset, Length)
+    catch
+        error:badarg -> {error, badarg}
+    end.
 
+sync(Fd) ->
+    try
+        #{ handle := FRef } = get_fd_data(Fd),
+        sync_nif(FRef, 0)
+    catch
+        error:badarg -> {error, badarg}
+    end.
 
-%% Returns {ok, Contents} | {error, Reason}
-read_file(File) when (is_list(File) orelse is_binary(File)) ->
-    case drv_open(?FD_DRV, [binary]) of
-	{ok, Port} ->
-	    Result = read_file(Port, File),
-	    close(Port),
-	    Result;
-	{error, _} = Error ->
-	    Error
-    end;
-read_file(_) ->
-    {error, badarg}.
+datasync(Fd) ->
+    try
+        #{ handle := FRef } = get_fd_data(Fd),
+        sync_nif(FRef, 1)
+    catch
+        error:badarg -> {error, badarg}
+    end.
 
-%% Takes a Port opened with open/1.
-read_file(Port, File) when is_port(Port),
-			   (is_list(File) orelse is_binary(File)) ->
-    Cmd = [?FILE_READ_FILE | pathname(File)],
-    case drv_command(Port, Cmd) of
-	{error, enomem} ->
-	    %% It could possibly help to do a 
-	    %% garbage collection here, 
-	    %% if the file server has some references
-	    %% to binaries read earlier.
-	    erlang:garbage_collect(),
-	    drv_command(Port, Cmd);
-	Result ->
-	    Result
+position(Fd, {cur, Offset}) ->
+    try
+        %% Adjust our current position according to how much we've read ahead.
+        #{ r_buffer := RBuf } = get_fd_data(Fd),
+        position_1(Fd, cur, Offset - prim_buffer:size(RBuf))
+    catch
+        error:badarg -> {error, badarg}
     end;
-read_file(_,_) ->
-    {error, badarg}.
-
-    
-
-%% Returns {error, Reason} | ok.
-write_file(File, Bin) when (is_list(File) orelse is_binary(File)) ->
-    case open(File, [binary, write]) of
-	{ok, Handle} ->
-	    Result = write(Handle, Bin),
-	    close(Handle),
-	    Result;
-	Error ->
-	    Error
+position(Fd, {Mark, Offset}) ->
+    try
+        position_1(Fd, Mark, Offset)
+    catch
+        error:badarg -> {error, badarg}
     end;
-write_file(_, _) -> 
-    {error, badarg}.
+position(Fd, cur) -> position(Fd, {cur, 0});
+position(Fd, bof) -> position(Fd, {bof, 0});
+position(Fd, eof) -> position(Fd, {eof, 0});
+position(Fd, Offset) -> position(Fd, {bof, Offset}).
 
+position_1(Fd, Mark, Offset) ->
+    #{ handle := FRef, r_buffer := RBuf } = get_fd_data(Fd),
+    prim_buffer:wipe(RBuf),
+    seek_nif(FRef, Mark, Offset).
 
-%% Returns {error, Reason} | {ok, BytesCopied}
-%sendfile(_,_,_,_,_,_,_,_,_,_) ->
-%    {error, enotsup};
-sendfile(#file_descriptor{module = ?MODULE, data = {Port, _}},
-	 Dest, Offset, Bytes, _ChunkSize, Headers, Trailers,
-	 Flags) ->
-    case erlang:port_get_data(Dest) of
-	Data when Data == inet_tcp; Data == inet6_tcp ->
-	    ok = inet:lock_socket(Dest,true),
-	    {ok, DestFD} = prim_inet:getfd(Dest),
-	    IntFlags = translate_sendfile_flags(Flags),
-	    try drv_command(Port, [<<?FILE_SENDFILE, DestFD:32,
-				     IntFlags:8,
-				     Offset:64/unsigned,
-				     Bytes:64/unsigned,
-				     (iolist_size(Headers)):32/unsigned,
-				     (iolist_size(Trailers)):32/unsigned>>,
-				   Headers,Trailers])
-	    after
-		ok = inet:lock_socket(Dest,false)
-	    end;
-	_Else ->
-	    {error,badarg}
+pread(Fd, Offset, Size) ->
+    try
+        #{ handle := FRef } = get_fd_data(Fd),
+        pread_nif(FRef, Offset, Size)
+    catch
+        error:badarg -> {error, badarg}
     end.
 
-translate_sendfile_flags([{use_threads,true}|T]) ->
-    ?EFILE_SENDFILE_USE_THREADS bor translate_sendfile_flags(T);
-translate_sendfile_flags([_|T]) ->
-    translate_sendfile_flags(T);
-translate_sendfile_flags([]) ->
-    0.
-
-
-%%%-----------------------------------------------------------------
-%%% Functions operating on files without handle to the file. ?DRV.
-%%%
-%%% Supposed to be called by applications through module file.
-
-
-
-%% Returns {ok, Port}, the Port should be used as first argument in all
-%% the following functions. Returns {error, Reason} upon failure.
-start() ->
-    drv_open(?DRV, [binary]).
-
-stop(Port) when is_port(Port) ->
-    try erlang:port_close(Port) of
-	_ ->
-	    ok
+pread(Fd, LocNums) ->
+    try
+        #{ handle := FRef } = get_fd_data(Fd),
+        pread_list(FRef, LocNums, [])
     catch
-	_:_ ->
-	    ok
+        error:badarg -> {error, badarg}
     end.
 
+-spec pread_list(FRef, LocNums, ResultList) -> Result when
+      FRef :: prim_file_ref(),
+      LocNums :: list({Offset :: non_neg_integer(),
+                       Size :: non_neg_integer()}),
+      ResultList :: list(eof | binary()),
+      Result :: {ok, ResultList} | {error, Reason :: atom()}.
+pread_list(_FRef, [], ResultList) ->
+    {ok, reverse_list(ResultList)};
+pread_list(FRef, [{Offset, Size} | Rest], ResultList) ->
+    case pread_nif(FRef, Offset, Size) of
+        {ok, Data} ->
+            pread_list(FRef, Rest, [Data | ResultList]);
+        eof ->
+            pread_list(FRef, Rest, [eof | ResultList]);
+        {error, Reason} ->
+            {error, Reason}
+    end.
 
+pwrite(Fd, Offset, IOData) ->
+    try
+        #{ handle := FRef, r_buffer := RBuf } = get_fd_data(Fd),
+        prim_buffer:wipe(RBuf),
+        pwrite_plain(FRef, Offset, erlang:iolist_to_iovec(IOData))
+    catch
+        error:badarg -> {error, badarg}
+    end.
+pwrite_plain(FRef, Offset, IOVec) ->
+    case pwrite_nif(FRef, Offset, IOVec) of
+        {continue, BytesWritten, Remainder} ->
+            pwrite_plain(FRef, Offset + BytesWritten, Remainder);
+        ok ->
+            ok;
+        {error, Reason} ->
+            {error, Reason}
+    end.
 
-%%% The following functions take an optional Port as first argument.
-%%% If the port is not supplied, a temporary one is opened and then
-%%% closed after the request has been performed.
-
-
-
-%% get_cwd/{0,1,2}
-
-get_cwd() ->
-    get_cwd_int(0).
-
-get_cwd(Port) when is_port(Port) ->
-    get_cwd_int(Port, 0);
-get_cwd([]) ->
-    get_cwd_int(0);
-get_cwd([Letter, $: | _]) when $a =< Letter, Letter =< $z ->
-    get_cwd_int(Letter - $a + 1);
-get_cwd([Letter, $: | _]) when $A =< Letter, Letter =< $Z ->
-    get_cwd_int(Letter - $A + 1);
-get_cwd([_|_]) ->
-    {error, einval};
-get_cwd(_) ->
-    {error, badarg}.
-
-get_cwd(Port, []) when is_port(Port) ->
-    get_cwd_int(Port, 0);
-get_cwd(Port, [Letter, $: | _])
-  when is_port(Port), $a =< Letter, Letter =< $z ->
-    get_cwd_int(Port, Letter - $a + 1);
-get_cwd(Port, [Letter, $: | _])
-  when is_port(Port), $A =< Letter, Letter =< $Z ->
-    get_cwd_int(Port, Letter - $A + 1);
-get_cwd(Port, [_|_]) when is_port(Port) ->
-    {error, einval};
-get_cwd(_, _) ->
-    {error, badarg}.
-
-get_cwd_int(Drive) ->
-    get_cwd_int({?DRV, [binary]}, Drive).
-
-get_cwd_int(Port, Drive) ->
-    drv_command(Port, <<?FILE_PWD, Drive>>,
-		fun handle_fname_response/1).
-
-
-
-%% set_cwd/{1,2}
+pwrite(Fd, LocBytes) ->
+    try
+        #{ handle := FRef, r_buffer := RBuf } = get_fd_data(Fd),
+        prim_buffer:wipe(RBuf),
+        pwrite_list(FRef, LocBytes, 0)
+    catch
+        error:badarg -> {error, badarg}
+    end.
 
-set_cwd(Dir) ->
-    set_cwd_int({?DRV, [binary]}, Dir).
+-spec pwrite_list(FRef, LocBytes, Successes) -> Result when
+      FRef :: prim_file_ref(),
+      LocBytes :: list({Offset :: non_neg_integer(),
+                        IOData :: iodata()}),
+      Successes :: non_neg_integer(),
+      Result :: ok | {error, {Successes, Reason :: atom()}}.
+pwrite_list(_FRef, [], _Successes) ->
+    ok;
+pwrite_list(FRef, [{Offset, IOData} | Rest], Successes) ->
+    case pwrite_plain(FRef, Offset, erlang:iolist_to_iovec(IOData)) of
+        {error, Reason} -> {error, {Successes, Reason}};
+        ok -> pwrite_list(FRef, Rest, Successes + 1)
+    end.
 
-set_cwd(Port, Dir) when is_port(Port) ->
-    set_cwd_int(Port, Dir).
+sendfile(Fd, Socket, Offset, Bytes, _ChunkSize, [], [], _Flags) ->
+    %% There's a very nasty race in here; if we die just prior to duplicating
+    %% the handle down in the sendfile call, it might get reused by something
+    %% entirely different and we'll leak unknown data to the socket until it
+    %% dies soon after.
+    %%
+    %% This bug was inherited from the old driver, except it was vulnerable to
+    %% the bug at any point and not just during setup.
+    %%
+    %% We'll have to live with this until we have a way to unambiguously
+    %% transfer things between drivers or NIFs. Current ideas all fall afoul
+    %% of the Two Generals problem.
+    try
+        advise(Fd, Offset, Bytes, sequential),
+        prim_inet:sendfile(Socket, get_handle(Fd), Offset, Bytes)
+    catch
+        error:badarg -> {error, badarg}
+    end;
+sendfile(_Fd, _Socket, _Offset, _Bytes, _ChunkSize, _Headers, _Trailers, _Flags) ->
+    {error, enotsup}.
 
-set_cwd_int(Port, Dir) when is_binary(Dir) ->
-    case prim_file:is_translatable(Dir) of
-	false ->
-	    {error, no_translation};
-	true ->
-	    drv_command(Port, [?FILE_CHDIR, pathname(Dir)])
+%% Undocumented internal function that reads a data block with indirection.
+%%
+%% This is only used once in DETS and can easily be emulated with pread/2, but
+%% it's pretty performance-sensitive so we've implemented it down in the NIF to
+%% avoid excessive rescheduling.
+-spec ipread_s32bu_p32bu(Fd, Offset, MaxSize) -> Result when
+      Fd :: #file_descriptor{},
+      Offset :: non_neg_integer(),
+      MaxSize :: non_neg_integer() | infinity,
+      Result :: {ok, Size :: non_neg_integer(),
+                     Pointer :: non_neg_integer(),
+                     Data :: iodata() | eof} |
+                eof |
+                {error, Reason :: atom()}.
+ipread_s32bu_p32bu(Fd, Offset, Infinity) when is_atom(Infinity) ->
+    ipread_s32bu_p32bu(Fd, Offset, (1 bsl 31) - 1);
+ipread_s32bu_p32bu(Fd, Offset, MaxSize)
+        when is_integer(Offset), is_integer(MaxSize) ->
+    try
+        #{ handle := FRef } = get_fd_data(Fd),
+        ipread_s32bu_p32bu_nif(FRef, Offset, MaxSize)
+    catch
+        error:badarg -> {error, badarg}
     end;
-set_cwd_int(Port, Dir) when is_list(Dir) ->
-    drv_command(Port, [?FILE_CHDIR, pathname(Dir)]);
-set_cwd_int(_, _) ->
+ipread_s32bu_p32bu(_Fd, _Offset, _MaxSize) ->
     {error, badarg}.
-
-
-
-%% delete/{1,2}
-
-delete(File) ->
-    delete_int({?DRV, [binary]}, File).
-
-delete(Port, File) when is_port(Port) ->
-    delete_int(Port, File).
-
-delete_int(Port, File) ->
-    drv_command(Port, [?FILE_DELETE, pathname(File)]).
-
-
-
-%% rename/{2,3}
-
-rename(From, To) ->
-    rename_int({?DRV, [binary]}, From, To).
-
-rename(Port, From, To) when is_port(Port) ->
-    rename_int(Port, From, To).
-
-rename_int(Port, From, To) ->
-    drv_command(Port, [?FILE_RENAME, pathname(From), pathname(To)]).
-
-
-
-%% make_dir/{1,2}
-
-make_dir(Dir) ->
-    make_dir_int({?DRV, [binary]}, Dir).
-
-make_dir(Port, Dir) when is_port(Port) ->
-    make_dir_int(Port, Dir).
-
-make_dir_int(Port, Dir) ->
-    drv_command(Port, [?FILE_MKDIR, pathname(Dir)]).
-
-
-
-%% del_dir/{1,2}
-
-del_dir(Dir) ->
-    del_dir_int({?DRV, [binary]}, Dir).
-
-del_dir(Port, Dir) when is_port(Port) ->
-    del_dir_int(Port, Dir).
-
-del_dir_int(Port, Dir) ->
-    drv_command(Port, [?FILE_RMDIR, pathname(Dir)]).
-
-
-
-%% read_file_info/{1,2,3}
-
-read_file_info(File) ->
-    read_file_info_int({?DRV, [binary]}, File, local).
-
-read_file_info(Port, File) when is_port(Port) ->
-    read_file_info_int(Port, File, local);
-read_file_info(File, Opts) ->
-    read_file_info_int({?DRV, [binary]}, File, plgv(time, Opts, local)).
-
-read_file_info(Port, File, Opts) when is_port(Port) ->
-    read_file_info_int(Port, File, plgv(time, Opts, local)).
-
-read_file_info_int(Port, File, TimeType) ->
+ipread_s32bu_p32bu_nif(_FRef, _Offset, _MaxSize) ->
+    erlang:nif_error(undef).
+
+%% Returns the binary representation of the underlying handle, for use in
+%% tricky operations like sendfile/8.
+-spec get_handle(Fd) -> Result when
+      Fd :: #file_descriptor{},
+      Result :: binary() | {error, Reason :: atom()}.
+get_handle(Fd) ->
     try
-	case drv_command(Port, [?FILE_FSTAT, pathname(File)]) of
-	    {ok, FI} -> {ok, FI#file_info{
-			ctime = from_seconds(FI#file_info.ctime, TimeType),
-			mtime = from_seconds(FI#file_info.mtime, TimeType),
-			atime = from_seconds(FI#file_info.atime, TimeType)
-		    }};
-	    Error    -> Error
-	end
+        #{ handle := FRef } = get_fd_data(Fd),
+        get_handle_nif(FRef)
     catch
-	error:_ -> {error, badarg}
+        error:badarg -> {error, badarg}
     end.
 
+%% Resets the write head to the position the user believes we're at, which may
+%% not be the same as the real one when read caching is in effect.
+reset_write_position(Fd) ->
+    #{ r_buffer := RBuf } = Fd#file_descriptor.data,
+    case prim_buffer:size(RBuf) of
+        Size when Size > 0 -> position(Fd, cur);
+        Size when Size =:= 0 -> ok
+    end.
 
-%% altname/{1,2}
-
-altname(File) ->
-    altname_int({?DRV, [binary]}, File).
-
-altname(Port, File) when is_port(Port) ->
-    altname_int(Port, File).
-
-altname_int(Port, File) ->
-    drv_command(Port, [?FILE_ALTNAME, pathname(File)],
-		fun handle_fname_response/1).
-
-%% write_file_info/{2,3,4}
-
-write_file_info(File, Info) ->
-    write_file_info_int({?DRV, [binary]}, File, Info, local).
-
-write_file_info(Port, File, Info) when is_port(Port) ->
-    write_file_info_int(Port, File, Info, local);
-write_file_info(File, Info, Opts) ->
-    write_file_info_int({?DRV, [binary]}, File, Info, plgv(time, Opts, local)).
-
-write_file_info(Port, File, Info, Opts) when is_port(Port) ->
-    write_file_info_int(Port, File, Info, plgv(time, Opts, local)).
+get_fd_data(#file_descriptor{ data = Data }) ->
+    #{ owner := Owner } = Data,
+    case self() of
+        Owner -> Data;
+        _ -> error(not_on_controlling_process)
+    end.
 
-write_file_info_int(Port, File, 
-		    #file_info{mode=Mode, 
-			       uid=Uid, 
-			       gid=Gid,
-			       atime=Atime0, 
-			       mtime=Mtime0, 
-			       ctime=Ctime0},
-		       TimeType) ->
+build_fd_data(FRef, Modes) ->
+    Defaults =
+        #{ owner => self(),
+           handle => FRef,
+           r_ahead_size => 0,
+           r_buffer => prim_buffer:new() },
+    fill_fd_option_map(Modes, Defaults).
+
+fill_fd_option_map([], Map) ->
+    Map;
+
+fill_fd_option_map([read_ahead | Modes], Map) ->
+    fill_fd_option_map([{read_ahead, 64 bsl 10} | Modes], Map);
+fill_fd_option_map([{read_ahead, Size} | Modes], Map) ->
+    fill_fd_option_map(Modes, Map#{ r_ahead_size => Size });
+
+fill_fd_option_map([_Ignored | Modes], Map) ->
+    fill_fd_option_map(Modes, Map).
+
+open_nif(_Name, _Modes) ->
+    erlang:nif_error(undef).
+close_nif(_FileRef) ->
+    erlang:nif_error(undef).
+read_nif(_FileRef, _Size) ->
+    erlang:nif_error(undef).
+write_nif(_FileRef, _IOVec) ->
+    erlang:nif_error(undef).
+pread_nif(_FileRef, _Offset, _Size) ->
+    erlang:nif_error(undef).
+pwrite_nif(_FileRef, _Offset, _IOVec) ->
+    erlang:nif_error(undef).
+seek_nif(_FileRef, _Mark, _Offset) ->
+    erlang:nif_error(undef).
+sync_nif(_FileRef, _DataOnly) ->
+    erlang:nif_error(undef).
+advise_nif(_FileRef, _Offset, _Length, _Advise) ->
+    erlang:nif_error(undef).
+allocate_nif(_FileRef, _Offset, _Length) ->
+    erlang:nif_error(undef).
+truncate_nif(_FileRef) ->
+    erlang:nif_error(undef).
+get_handle_nif(_FileRef) ->
+    erlang:nif_error(undef).
 
-    % Atime and/or Mtime might be undefined
-    %  - use localtime() for atime, if atime is undefined
-    %  - use atime as mtime if mtime is undefined
-    %  - use mtime as ctime if ctime is undefined
+%%
+%% Quality-of-life helpers
+%%
 
+read_file(Filename) ->
+    %% We're doing this operation in the NIF to avoid excessive rescheduling.
     try
-	Atime = file_info_validate_atime(Atime0, TimeType),
-	Mtime = file_info_validate_mtime(Mtime0, Atime),
-	Ctime = file_info_validate_ctime(Ctime0, Mtime),
-
-	drv_command(Port, [?FILE_WRITE_INFO, 
-		int_to_int32bytes(Mode), 
-		int_to_int32bytes(Uid), 
-		int_to_int32bytes(Gid),
-		int_to_int64bytes(to_seconds(Atime, TimeType)), 
-		int_to_int64bytes(to_seconds(Mtime, TimeType)), 
-		int_to_int64bytes(to_seconds(Ctime, TimeType)),
-		pathname(File)])
+        read_file_nif(encode_path(Filename))
     catch
-	error:_ -> {error, badarg}
+        error:badarg -> {error, badarg}
+    end.
+read_file_nif(_Filename) ->
+    erlang:nif_error(undef).
+
+write_file(Filename, Bytes) ->
+    write_file(Filename, Bytes, []).
+write_file(Filename, Bytes, Modes) ->
+    case open(Filename, [write, binary | Modes]) of
+        {ok, Fd} ->
+            Result = write(Fd, Bytes),
+            close(Fd),
+            Result;
+        {error, Reason} ->
+            {error, Reason}
     end.
 
+%%
+%% Filesystem operations
+%%
 
-file_info_validate_atime(Atime, _) when Atime =/= undefined -> Atime;
-file_info_validate_atime(undefined, local)     -> erlang:localtime();
-file_info_validate_atime(undefined, universal) -> erlang:universaltime();
-file_info_validate_atime(undefined, posix)     -> erlang:universaltime_to_posixtime(erlang:universaltime()).
-
-file_info_validate_mtime(undefined, Atime) -> Atime;
-file_info_validate_mtime(Mtime, _)         -> Mtime.
-
-file_info_validate_ctime(undefined, Mtime) -> Mtime;
-file_info_validate_ctime(Ctime, _)         -> Ctime.
-
-%% make_link/{2,3}
-
-make_link(Old, New) ->
-    make_link_int({?DRV, [binary]}, Old, New).
-
-make_link(Port, Old, New) when is_port(Port) ->
-    make_link_int(Port, Old, New).
-
-make_link_int(Port, Old, New) ->
-    drv_command(Port, [?FILE_LINK, pathname(Old), pathname(New)]).
-
-
-
-%% make_symlink/{2,3}
-
-make_symlink(Old, New) ->
-    make_symlink_int({?DRV, [binary]}, Old, New).
-
-make_symlink(Port, Old, New) when is_port(Port) ->
-    make_symlink_int(Port, Old, New).
-
-make_symlink_int(Port, Old, New) ->
-    drv_command(Port, [?FILE_SYMLINK, pathname(Old), pathname(New)]).
-
-
-
-%% read_link/{2,3}
-
-read_link(Link) ->
-    read_link_int({?DRV, [binary]}, Link).
-
-read_link(Port, Link) when is_port(Port) ->
-    read_link_int(Port, Link).
-
-read_link_int(Port, Link) ->
-    drv_command(Port, [?FILE_READLINK, pathname(Link)],
-		fun handle_fname_response/1).
-
-%% read_link_all/{2,3}
-
-read_link_all(Link) ->
-    read_link_all_int({?DRV, [binary]}, Link).
-
-read_link_all(Port, Link) when is_port(Port) ->
-    read_link_all_int(Port, Link).
-
-read_link_all_int(Port, Link) ->
-    drv_command(Port, [?FILE_READLINK, pathname(Link)],
-		fun handle_fname_response_all/1).
-
-
-
-%% read_link_info/{2,3}
-
-read_link_info(Link) ->
-    read_link_info_int({?DRV, [binary]}, Link, local).
-
-read_link_info(Port, Link) when is_port(Port) ->
-    read_link_info_int(Port, Link, local);
+read_link(Name) -> read_link_1(Name, false).
+read_link_all(Name) -> read_link_1(Name, true).
 
-read_link_info(Link, Opts) ->
-    read_link_info_int({?DRV, [binary]}, Link, plgv(time, Opts, local)).
+read_link_1(Name, AcceptRawNames) ->
+    try read_link_nif(encode_path(Name)) of
+        {ok, RawName} -> translate_raw_name(RawName, AcceptRawNames);
+        {error, Reason} -> {error, Reason}
+    catch
+        error:badarg -> {error, badarg}
+    end.
 
-read_link_info(Port, Link, Opts) when is_port(Port) ->
-    read_link_info_int(Port, Link, plgv(time, Opts, local)).
+translate_raw_name(RawName, SilentFailure) ->
+    case decode_path(RawName) of
+        Converted when is_list(Converted) -> {ok, Converted};
+        {error, _Reason} when SilentFailure =:= false -> {error, einval};
+        {error, _Reason} when SilentFailure =:= true -> {ok, RawName}
+    end.
 
+list_dir(Name) -> list_dir_1(Name, true).
+list_dir_all(Name) -> list_dir_1(Name, false).
 
-read_link_info_int(Port, Link, TimeType) ->
-    try
-	case drv_command(Port, [?FILE_LSTAT, pathname(Link)]) of
-	    {ok, FI} -> {ok, FI#file_info{
-			ctime = from_seconds(FI#file_info.ctime, TimeType),
-			mtime = from_seconds(FI#file_info.mtime, TimeType),
-			atime = from_seconds(FI#file_info.atime, TimeType)
-		    }};
-	    Error    -> Error
-	end
+list_dir_1(Name, SkipInvalid) ->
+    try list_dir_nif(encode_path(Name)) of
+        {ok, RawNames} -> list_dir_convert(RawNames, SkipInvalid, []);
+        {error, Reason} -> {error, Reason}
     catch
-	error:_ -> {error, badarg}
+        error:badarg -> {error, badarg}
     end.
 
-%% list_dir/{1,2}
-
-list_dir(Dir) ->
-    list_dir_int({?DRV, [binary]}, Dir).
-
-list_dir(Port, Dir) when is_port(Port) ->
-    list_dir_int(Port, Dir).
-
-list_dir_int(Port, Dir) ->
-    drv_command(Port, [?FILE_READDIR, pathname(Dir)],
-		fun(P) ->
-			case list_dir_response(P, []) of
-			    {ok, RawNames} ->
-				try
-				    {ok, list_dir_convert(RawNames)}
-				catch
-				    throw:Reason ->
-					Reason
-				end;
-			    Error ->
-				Error
-			end
-		end).
-
-list_dir_all(Dir) ->
-    list_dir_all_int({?DRV, [binary]}, Dir).
-
-list_dir_all(Port, Dir) when is_port(Port) ->
-    list_dir_all_int(Port, Dir).
-
-list_dir_all_int(Port, Dir) ->
-    drv_command(Port, [?FILE_READDIR, pathname(Dir)],
-		fun(P) ->
-			case list_dir_response(P, []) of
-			    {ok, RawNames} ->
-				{ok, list_dir_convert_all(RawNames)};
-			    Error ->
-				Error
-			end
-		end).
-
-list_dir_response(Port, Acc0) ->
-    case drv_get_response(Port) of
-	{lfname, []} ->
-	    {ok, Acc0};
-	{lfname, Names} ->
-	    Acc = [Name || <<L:16,Name:L/binary>> <= Names] ++ Acc0,
-	    list_dir_response(Port, Acc);
-	Error ->
-	    Error
+list_dir_convert([], _SkipInvalid, Result) ->
+    {ok, Result};
+list_dir_convert([RawName | Rest], SkipInvalid, Result) ->
+    case decode_path(RawName) of
+        Converted when is_list(Converted) ->
+            list_dir_convert(Rest, SkipInvalid, [Converted | Result]);
+        {error, _} when SkipInvalid =:= false ->
+            list_dir_convert(Rest, SkipInvalid, [RawName | Result]);
+
+        %% If the filename cannot be converted, return error or ignore with
+        %% optional error logger warning depending on +fn{u|a}{i|e|w} emulator
+        %% switches.
+        {error, ignore} ->
+            list_dir_convert(Rest, SkipInvalid, Result);
+        {error, warning} ->
+            error_logger:warning_msg(
+                "Non-unicode filename ~p ignored\n", [RawName]),
+            list_dir_convert(Rest, SkipInvalid, Result);
+        {error, _} ->
+            {error, {no_translation, RawName}}
     end.
 
-list_dir_convert([Name|Names]) ->
-    %% If the filename cannot be converted, return error or ignore
-    %% with optional error logger warning, depending on +fn{u|a}{i|e|w}
-    %% emulator switches.
-    case prim_file:internal_native2name(Name) of
-	{error, warning} ->
-	    error_logger:warning_msg("Non-unicode filename ~p ignored\n",
-				     [Name]),
-	    list_dir_convert(Names);
-	{error, ignore} ->
-	    list_dir_convert(Names);
-	{error, error} ->
-	    throw({error, {no_translation, Name}});
-	Converted when is_list(Converted) ->
-	    [Converted|list_dir_convert(Names)]
-    end;
-list_dir_convert([]) -> [].
-
-list_dir_convert_all([Name|Names]) ->
-    %% If the filename cannot be converted, retain the filename as
-    %% a binary.
-    case prim_file:internal_native2name(Name) of
-	{error, _} ->
-	    [Name|list_dir_convert_all(Names)];
-	Converted when is_list(Converted) ->
-	    [Converted|list_dir_convert_all(Names)]
-    end;
-list_dir_convert_all([]) -> [].
-
-%%%-----------------------------------------------------------------
-%%% Functions to communicate with the driver
-
-handle_fname_response(Port) ->
-    case drv_get_response(Port) of
-	{fname, Name} ->
-	    case prim_file:internal_native2name(Name) of
-		{error, warning} ->
-		    error_logger:warning_msg("Non-unicode filename ~p "
-					     "ignored when reading link\n",
-					     [Name]),
-		    {error, einval};
-		{error, _} ->
-		    {error, einval};
-		Converted when is_list(Converted) ->
-		    {ok, Converted}
-	    end;
-	Error ->
-	    Error
+read_file_info(Filename) ->
+    read_info_1(Filename, 1, local).
+read_file_info(Filename, Opts) ->
+    read_info_1(Filename, 1, proplist_get_value(time, Opts, local)).
+
+read_link_info(Name) ->
+    read_info_1(Name, 0, local).
+read_link_info(Name, Opts) ->
+    read_info_1(Name, 0, proplist_get_value(time, Opts, local)).
+
+read_info_1(Name, FollowLinks, TimeType) ->
+    try read_info_nif(encode_path(Name), FollowLinks) of
+        {error, Reason} -> {error, Reason};
+        FileInfo ->
+            CTime = from_posix_seconds(FileInfo#file_info.ctime, TimeType),
+            MTime = from_posix_seconds(FileInfo#file_info.mtime, TimeType),
+            ATime = from_posix_seconds(FileInfo#file_info.atime, TimeType),
+            {ok, FileInfo#file_info{ ctime = CTime, mtime = MTime, atime = ATime }}
+    catch
+        error:badarg -> {error, badarg}
     end.
 
-handle_fname_response_all(Port) ->
-    case drv_get_response(Port) of
-	{fname, Name} ->
-	    case prim_file:internal_native2name(Name) of
-		{error, _} ->
-		    {ok, Name};
-		Converted when is_list(Converted) ->
-		    {ok, Converted}
-	    end;
-	Error ->
-	    Error
+write_file_info(Filename, Info) ->
+    write_file_info_1(Filename, Info, local).
+write_file_info(Filename, Info, Opts) ->
+    write_file_info_1(Filename, Info, proplist_get_value(time, Opts, local)).
+
+write_file_info_1(Filename, Info, TimeType) ->
+    #file_info{ mode = Modes,
+                uid = Uid,
+                gid = Gid,
+                atime = ATime0,
+                mtime = MTime0,
+                ctime = CTime0} = Info,
+    try
+        % ATime and/or MTime might be undefined
+        %  - use localtime() for atime, if atime is undefined
+        %  - use atime as mtime if mtime is undefined
+        %  - use mtime as ctime if ctime is undefined
+        ATime = file_info_convert_atime(ATime0, TimeType),
+        MTime = file_info_convert_mtime(MTime0, ATime, TimeType),
+        CTime = file_info_convert_ctime(CTime0, MTime, TimeType),
+        EncodedName = encode_path(Filename),
+
+        %% This is a bit ugly but we need to handle partial failures the same
+        %% way the old driver did.
+        throw_on_error(set_owner(EncodedName, Uid, Gid)),
+        throw_on_error(set_permissions(EncodedName, Modes)),
+        throw_on_error(set_time(EncodedName, ATime, MTime, CTime))
+    catch
+        throw:Reason -> {error, Reason};
+        error:_ -> {error, badarg}
     end.
 
-%% Opens a driver port and converts any problems into {error, emfile}.
-%% Returns {ok, Port} when successful.
+set_owner(_EncodedName, undefined, undefined) ->
+    ok;
+set_owner(EncodedName, Uid, Gid) ->
+    set_owner_nif(EncodedName, Uid, Gid).
+set_owner_nif(_Path, _Uid, _Gid) ->
+    erlang:nif_error(undef).
 
-drv_open(Driver, Portopts) ->
-    try erlang:open_port({spawn_driver, Driver}, Portopts) of
-	Port ->
-	    {ok, Port}
+set_permissions(_EncodedName, undefined) ->
+    ok;
+set_permissions(EncodedName, Permissions) ->
+    set_permissions_nif(EncodedName, Permissions).
+set_permissions_nif(_Path, _Permissions) ->
+    erlang:nif_error(undef).
+
+set_time(EncodedName, ATime, MTime, CTime) ->
+    set_time_nif(EncodedName, ATime, MTime, CTime).
+set_time_nif(_Path, _ATime, _MTime, _CTime) ->
+    erlang:nif_error(undef).
+
+throw_on_error(ok) -> ok;
+throw_on_error({error, enotsup}) -> ok;
+throw_on_error({error, Reason}) -> throw(Reason).
+
+file_info_convert_atime(ATime, TimeType) when ATime =/= undefined ->
+    to_posix_seconds(ATime, TimeType);
+file_info_convert_atime(undefined, local) ->
+    to_posix_seconds(erlang:localtime(), local);
+file_info_convert_atime(undefined, universal) ->
+    to_posix_seconds(erlang:universaltime(), universal);
+file_info_convert_atime(undefined, posix) ->
+    erlang:universaltime_to_posixtime(erlang:universaltime()).
+
+file_info_convert_mtime(undefined, ATime, _TimeType) ->
+    ATime;
+file_info_convert_mtime(MTime, _ATime, TimeType) ->
+    to_posix_seconds(MTime, TimeType).
+
+file_info_convert_ctime(undefined, MTime, _TimeType) ->
+    MTime;
+file_info_convert_ctime(CTime, _MTime, TimeType) ->
+    to_posix_seconds(CTime, TimeType).
+
+%% This is only relevant on Windows, so we assume that format to simplify the
+%% internals.
+get_cwd([Letter, $:]) when Letter >= $A, Letter =< $Z ->
+    get_dcwd(Letter - $A + 1);
+get_cwd([Letter, $:]) when Letter >= $a, Letter =< $z ->
+    get_dcwd(Letter - $a + 1);
+get_cwd([_|_]) ->
+    {error, einval};
+get_cwd(_) ->
+    {error, badarg}.
+get_dcwd(Index) ->
+    try get_device_cwd_nif(Index) of
+        {ok, RawPath} -> {ok, decode_path(RawPath)};
+        {error, Reason} -> {error, Reason}
     catch
-	error:Reason ->
-	    {error, Reason}
+        error:badarg -> {error, badarg}
     end.
 
-
-
-%% Closes a port in a safe way. Returns ok.
-
-drv_close(Port) ->
-    Save = erlang:dt_spread_tag(false),
+get_cwd() ->
+    try get_cwd_nif() of
+        {ok, RawPath} -> {ok, decode_path(RawPath)};
+        {error, Reason} -> {error, Reason}
+    catch
+        error:badarg -> {error, badarg}
+    end.
+set_cwd(Path) ->
     try
-	try erlang:port_close(Port) catch error:_ -> ok end,
-	receive %% Ugly workaround in case the caller==owner traps exits
-	    {'EXIT', Port, _Reason} -> 
-		ok
-	after 0 -> 
-		ok
-	end
-    after
-	erlang:dt_restore_tag(Save)
+        case is_path_translatable(Path) of
+            true -> set_cwd_nif(encode_path(Path));
+            false -> {error, no_translation}
+        end
+    catch
+        error:badarg -> {error, badarg}
     end.
 
-
-
-%% Issues a command to a port and gets the response.
-%% If Port is {Driver, Portopts} a port is first opened and 
-%% then closed after the result has been received.
-%% Returns {ok, Result} or {error, Reason}.
-
-drv_command(Port, Command) ->
-    drv_command(Port, Command, undefined).
-
-drv_command(Port, Command, R) when is_binary(Command) ->
-    drv_command(Port, Command, true, R);
-drv_command(Port, Command, R) ->
-    try erlang:iolist_size(Command) of
-	_ ->
-	    drv_command(Port, Command, true, R)
+delete(Path) ->
+    try
+        del_file_nif(encode_path(Path))
     catch
-	error:Reason ->
-	    {error, Reason}
+        error:badarg -> {error, badarg}
     end.
 
-drv_command(Port, Command, Validated, R) when is_port(Port) ->
-    Save = erlang:dt_spread_tag(false),
-    try erlang:port_command(Port, erlang:dt_append_vm_tag_data(Command)) of
-	true ->
-	    drv_get_response(Port, R)
+rename(Source, Destination) ->
+    try
+        rename_nif(encode_path(Source), encode_path(Destination))
     catch
-	%% If the Command is valid, knowing that the port is a port,
-	%% a badarg error must mean it is a dead port, that is:
-	%% a currently invalid filehandle, -> einval, not badarg.
-	error:badarg when Validated ->
-	    {error, einval};
-	error:badarg ->
-	    try erlang:iolist_size(Command) of
-		_ -> % Valid
-		    {error, einval}
-	    catch
-		error:_ ->
-		    {error, badarg}
-	    end;
-	error:Reason ->
-	    {error, Reason}
-    after
-	erlang:dt_restore_tag(Save)
-    end;
-drv_command({Driver, Portopts}, Command, Validated, R) ->
-    case drv_open(Driver, Portopts) of
-	{ok, Port} ->
-	    Result = drv_command(Port, Command, Validated, R),
-	    drv_close(Port),
-	    Result;
-	Error ->
-	    Error
+        error:badarg -> {error, badarg}
     end.
-drv_command_nt(Port, Command, R) when is_port(Port) ->
-    Save = erlang:dt_spread_tag(false),
-    try erlang:port_command(Port, Command) of
-	true ->
-	    drv_get_response(Port, R)
+make_dir(Path) ->
+    try
+        make_dir_nif(encode_path(Path))
     catch
-	error:badarg ->
-	    try erlang:iolist_size(Command) of
-		_ -> % Valid
-		    {error, einval}
-	    catch
-		error:_ ->
-		    {error, badarg}
-	    end;
-	error:Reason ->
-	    {error, Reason}
-    after
-	erlang:dt_restore_tag(Save)
+        error:badarg -> {error, badarg}
     end.
-
-
-    
-%% Receives the response from a driver port.
-%% Returns: {ok, ListOrBinary}|{error, Reason}
-
-drv_get_response(Port, undefined) ->
-    drv_get_response(Port);
-drv_get_response(Port, Fun) when is_function(Fun, 1) ->
-    Fun(Port).
-
-drv_get_response(Port) ->
-    erlang:bump_reductions(100),
-    receive
-	{Port, {data, [Response|Rest] = Data}} ->
-	    try translate_response(Response, Rest)
-	    catch
-		error:Reason ->
-		    {error, {bad_response_from_port, Data, 
-			     {Reason, erlang:get_stacktrace()}}}
-	    end;
-	{'EXIT', Port, Reason} ->
-	    {error, {port_died, Reason}}
+del_dir(Path) ->
+    try
+        del_dir_nif(encode_path(Path))
+    catch
+        error:badarg -> {error, badarg}
     end.
-
-
-%%%-----------------------------------------------------------------
-%%% Utility functions.
-
-%% Converts a list of mode atoms into a mode word for the driver.
-%% Returns {Mode, Portopts, Setopts} where Portopts is a list of 
-%% options for erlang:open_port/2 and Setopts is a list of 
-%% setopt commands to send to the port, or error Reason upon failure.
-
-open_mode(List) when is_list(List) ->
-    case open_mode(List, 0, [], []) of
-	{Mode, Portopts, Setopts} when Mode band 
-			  (?EFILE_MODE_READ bor ?EFILE_MODE_WRITE) 
-			  =:= 0 ->
-	    {Mode bor ?EFILE_MODE_READ, Portopts, Setopts};
-	Other ->
-	    Other
+make_link(Existing, New) ->
+    try
+        make_hard_link_nif(encode_path(Existing), encode_path(New))
+    catch
+        error:badarg -> {error, badarg}
     end.
-
-open_mode([raw|Rest], Mode, Portopts, Setopts) ->
-    open_mode(Rest, Mode, Portopts, Setopts);
-open_mode([read|Rest], Mode, Portopts, Setopts) ->
-    open_mode(Rest, Mode bor ?EFILE_MODE_READ, Portopts, Setopts);
-open_mode([write|Rest], Mode, Portopts, Setopts) ->
-    open_mode(Rest, Mode bor ?EFILE_MODE_WRITE, Portopts, Setopts);
-open_mode([binary|Rest], Mode, Portopts, Setopts) ->
-    open_mode(Rest, Mode, [binary | Portopts], Setopts);
-open_mode([compressed|Rest], Mode, Portopts, Setopts) ->
-    open_mode(Rest, Mode bor ?EFILE_COMPRESSED, Portopts, Setopts);
-open_mode([append|Rest], Mode, Portopts, Setopts) ->
-    open_mode(Rest, Mode bor ?EFILE_MODE_APPEND bor ?EFILE_MODE_WRITE, 
-	      Portopts, Setopts);
-open_mode([exclusive|Rest], Mode, Portopts, Setopts) ->
-    open_mode(Rest, Mode bor ?EFILE_MODE_EXCL, Portopts, Setopts);
-open_mode([sync|Rest], Mode, Portopts, Setopts) ->
-    open_mode(Rest, Mode bor ?EFILE_MODE_SYNC, Portopts, Setopts);
-open_mode([delayed_write|Rest], Mode, Portopts, Setopts) ->
-    open_mode([{delayed_write, 64*1024, 2000}|Rest], Mode,
-	      Portopts, Setopts);
-open_mode([{delayed_write, Size, Delay}|Rest], Mode, Portopts, Setopts) 
-  when is_integer(Size), 0 =< Size, is_integer(Delay), 0 =< Delay ->
-    if
-	Size < ?LARGEFILESIZE, Delay < 1 bsl 64 ->
-	    open_mode(Rest, Mode, Portopts, 
-		      [<<?FILE_SETOPT, ?FILE_OPT_DELAYED_WRITE,
-			Size:64, Delay:64>> 
-		       | Setopts]);
-	true ->
-	    einval
-    end;
-open_mode([read_ahead|Rest], Mode, Portopts, Setopts) ->
-    open_mode([{read_ahead, 64*1024}|Rest], Mode, Portopts, Setopts);
-open_mode([{read_ahead, Size}|Rest], Mode, Portopts, Setopts)
-  when is_integer(Size), 0 =< Size ->
-    if
-	Size < ?LARGEFILESIZE ->
-	    open_mode(Rest, Mode, Portopts,
-		      [<<?FILE_SETOPT, ?FILE_OPT_READ_AHEAD,
-			Size:64>> | Setopts]);
-	true ->
-	    einval
-    end;
-open_mode([], Mode, Portopts, Setopts) ->
-    {Mode, reverse(Portopts), reverse(Setopts)};
-open_mode(_, _Mode, _Portopts, _Setopts) ->
-    badarg.
-
-
-
-%% Converts a position tuple {bof, X} | {cur, X} | {eof, X} into
-%% {Offset, OriginCode} for the driver.
-%% Returns badarg upon failure.
-
-lseek_position(Pos)
-  when is_integer(Pos) ->
-    lseek_position({bof, Pos});
-lseek_position(bof) ->
-    lseek_position({bof, 0});
-lseek_position(cur) ->
-    lseek_position({cur, 0});
-lseek_position(eof) ->
-    lseek_position({eof, 0});
-lseek_position({bof, Offset})
-  when is_integer(Offset) ->
-    {Offset, ?EFILE_SEEK_SET};
-lseek_position({cur, Offset})
-  when is_integer(Offset) ->
-    {Offset, ?EFILE_SEEK_CUR};
-lseek_position({eof, Offset})
-  when is_integer(Offset) ->
-    {Offset, ?EFILE_SEEK_END};
-lseek_position(_) ->
-    badarg.
-
-
-
-%% 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) ->
-    {error, list_to_atom(List)};
-translate_response(?FILE_RESP_NUMBER, List) ->
-    {N, []} = get_uint64(List),
-    {ok, N};
-translate_response(?FILE_RESP_DATA, List) ->
-    {_N, _Data} = ND = get_uint64(List),
-    {ok, ND};
-translate_response(?FILE_RESP_INFO, List) when is_list(List) ->
-    {ok, transform_info(List)};
-translate_response(?FILE_RESP_NUMERR, L0) ->
-    {N, L1} = get_uint64(L0),
-    {error, {N, list_to_atom(L1)}};
-translate_response(?FILE_RESP_LDATA, List) ->
-    {ok, transform_ldata(List)};
-translate_response(?FILE_RESP_N2DATA, 
-		   <<Offset:64, 0:64, Size:64>>) ->
-    {ok, {Size, Offset, eof}};
-translate_response(?FILE_RESP_N2DATA, 
-		   [<<Offset:64, 0:64, Size:64>> | <<>>]) ->
-    {ok, {Size, Offset, eof}};
-translate_response(?FILE_RESP_N2DATA = X, 
-		   [<<_:64, 0:64, _:64>> | _] = Data) ->
-    {error, {bad_response_from_port, [X | Data]}};
-translate_response(?FILE_RESP_N2DATA = X, 
-		   [<<_:64, _:64, _:64>> | <<>>] = Data) ->
-    {error, {bad_response_from_port, [X | Data]}};
-translate_response(?FILE_RESP_N2DATA, 
-		   [<<Offset:64, _ReadSize:64, Size:64>> | D]) ->
-    {ok, {Size, Offset, D}};
-translate_response(?FILE_RESP_N2DATA = X, L0) when is_list(L0) ->
-    {Offset, L1}    = get_uint64(L0),
-    {ReadSize, L2}  = get_uint64(L1),
-    {Size, L3}      = get_uint64(L2),
-    case {ReadSize, L3} of
-	{0, []} ->
-	    {ok, {Size, Offset, eof}};
-	{0, _} ->
-	    {error, {bad_response_from_port, [X | L0]}};
-	{_, []} ->
-	    {error, {bad_response_from_port, [X | L0]}};
-	_ ->
-	    {ok, {Size, Offset, L3}}
-    end;
-translate_response(?FILE_RESP_EOF, []) ->
-    eof;
-translate_response(?FILE_RESP_FNAME, Data) ->
-    {fname, Data};
-translate_response(?FILE_RESP_LFNAME, Data) ->
-    {lfname, Data};
-translate_response(?FILE_RESP_ALL_DATA, Data) ->
-    {ok, Data};
-translate_response(X, Data) ->
-    {error, {bad_response_from_port, [X | Data]}}.
-
-transform_info([
-    Hsize1, Hsize2, Hsize3, Hsize4, 
-    Lsize1, Lsize2, Lsize3, Lsize4,
-    Type1,  Type2,  Type3,  Type4,
-    Atime1, Atime2, Atime3, Atime4, Atime5, Atime6, Atime7, Atime8,
-    Mtime1, Mtime2, Mtime3, Mtime4, Mtime5, Mtime6, Mtime7, Mtime8,
-    Ctime1, Ctime2, Ctime3, Ctime4, Ctime5, Ctime6, Ctime7, Ctime8,
-    Mode1,  Mode2,  Mode3,  Mode4,
-    Links1, Links2, Links3, Links4,
-    Major1, Major2, Major3, Major4,
-    Minor1, Minor2, Minor3, Minor4,
-    Inode1, Inode2, Inode3, Inode4,
-    Uid1,   Uid2,   Uid3,   Uid4,
-    Gid1,   Gid2,   Gid3,   Gid4,
-    Access1,Access2,Access3,Access4]) ->
-    #file_info {
-		size   = uint32(Hsize1,Hsize2,Hsize3,Hsize4)*16#100000000 + uint32(Lsize1,Lsize2,Lsize3,Lsize4),
-		type   = file_type(uint32(Type1,Type2,Type3,Type4)),
-		access = file_access(uint32(Access1,Access2,Access3,Access4)),
-		atime  = sint64(Atime1, Atime2, Atime3, Atime4, Atime5, Atime6, Atime7, Atime8),
-		mtime  = sint64(Mtime1, Mtime2, Mtime3, Mtime4, Mtime5, Mtime6, Mtime7, Mtime8),
-		ctime  = sint64(Ctime1, Ctime2, Ctime3, Ctime4, Ctime5, Ctime6, Ctime7, Ctime8),
-		mode   = uint32(Mode1,Mode2,Mode3,Mode4),
-		links  = uint32(Links1,Links2,Links3,Links4),
-		major_device = uint32(Major1,Major2,Major3,Major4),
-		minor_device = uint32(Minor1,Minor2,Minor3,Minor4),
-		inode = uint32(Inode1,Inode2,Inode3,Inode4),
-		uid   = uint32(Uid1,Uid2,Uid3,Uid4),
-		gid   = uint32(Gid1,Gid2,Gid3,Gid4)
-	    }.
-    
-    
-file_type(1) -> device;
-file_type(2) -> directory;
-file_type(3) -> regular;
-file_type(4) -> symlink;
-file_type(_) -> other.
-
-file_access(0) -> none;   
-file_access(1) -> write;
-file_access(2) -> read;
-file_access(3) -> read_write.
-
-int_to_int32bytes(Int) when is_integer(Int) ->
-    <<Int:32>>;
-int_to_int32bytes(undefined) ->
-    <<-1:32>>.
-
-int_to_int64bytes(Int) when is_integer(Int) ->
-    <<Int:64/signed>>.
-
-
-sint64(I1,I2,I3,I4,I5,I6,I7,I8) when I1 > 127 ->
-    ((I1 bsl 56) bor (I2 bsl 48) bor (I3 bsl 40) bor (I4 bsl 32) bor
-	(I5 bsl 24) bor (I6 bsl 16) bor (I7 bsl  8) bor I8) - (1 bsl 64);
-sint64(I1,I2,I3,I4,I5,I6,I7,I8) ->
-    ((I1 bsl 56) bor (I2 bsl 48) bor (I3 bsl 40) bor (I4 bsl 32) bor
-	(I5 bsl 24) bor (I6 bsl 16) bor (I7 bsl  8) bor I8).
-
-
-uint32(X1,X2,X3,X4) ->
-    (X1 bsl 24) bor (X2 bsl 16) bor (X3 bsl 8) bor X4.
-
-get_uint64(L0) ->
-    {X1, L1} = get_uint32(L0),
-    {X2, L2} = get_uint32(L1),
-    {(X1 bsl 32) bor X2, L2}.
-
-get_uint32([X1,X2,X3,X4|List]) ->
-    {(((((X1 bsl 8) bor X2) bsl 8) bor X3) bsl 8) bor X4, List}.
-
-
-%% Binary mode
-transform_ldata(<<0:32, 0:32>>) ->
-    [];
-transform_ldata([<<0:32, N:32, Sizes/binary>> | Datas]) ->
-    transform_ldata(N, Sizes, Datas, []);
-%% List mode
-transform_ldata([_,_,_,_,_,_,_,_|_] = L0) ->
-    {0, L1} = get_uint32(L0),
-    {N, L2} = get_uint32(L1),
-    transform_ldata(N, L2, []).
-
-%% List mode
-transform_ldata(0, List, Sizes) ->
-    transform_ldata(0, List, reverse(Sizes), []);
-transform_ldata(N, L0, Sizes) ->
-    {Size, L1} = get_uint64(L0),
-    transform_ldata(N-1, L1, [Size | Sizes]).
-
-%% Binary mode
-transform_ldata(1, <<0:64>>, <<>>, R) ->
-    reverse(R, [eof]);
-transform_ldata(1, <<Size:64>>, Data, R) 
-  when byte_size(Data) =:= Size ->
-    reverse(R, [Data]);
-transform_ldata(N, <<0:64, Sizes/binary>>, [<<>> | Datas], R) ->
-    transform_ldata(N-1, Sizes, Datas, [eof | R]);
-transform_ldata(N, <<Size:64, Sizes/binary>>, [Data | Datas], R) 
-  when byte_size(Data) =:= Size ->
-    transform_ldata(N-1, Sizes, Datas, [Data | R]);
-%% List mode
-transform_ldata(0, [], [], R) ->
-    reverse(R);
-transform_ldata(0, List, [0 | Sizes], R) ->
-    transform_ldata(0, List, Sizes, [eof | R]);
-transform_ldata(0, List, [Size | Sizes], R) ->
-    {Front, Rear} = lists_split(List, Size),
-    transform_ldata(0, Rear, Sizes, [Front | R]).
-
-lists_split(List, 0) when is_list(List) ->
-    {[], List};
-lists_split(List, N) when is_list(List), is_integer(N), N < 0 ->
-    erlang:error(badarg, [List, N]);
-lists_split(List, N) when is_list(List), is_integer(N) ->
-    case lists_split(List, N, []) of
-	premature_end_of_list ->
-	    erlang:error(badarg, [List, N]);
-	Result ->
-	    Result
+make_symlink(Existing, New) ->
+    try
+        make_soft_link_nif(encode_path(Existing), encode_path(New))
+    catch
+        error:badarg -> {error, badarg}
     end.
 
-lists_split(List, 0, Rev) ->
-    {reverse(Rev), List};
-lists_split([], _, _) ->
-    premature_end_of_list;
-lists_split([Hd | Tl], N, Rev) ->
-    lists_split(Tl, N-1, [Hd | Rev]).
-
-%% We KNOW that lists:reverse/2 is a BIF.
-
-reverse(X) -> lists:reverse(X, []).
-reverse(L, T) -> lists:reverse(L, T).
+altname(Path) ->
+    try altname_nif(encode_path(Path)) of
+        {ok, RawPath} -> {ok, decode_path(RawPath)};
+        Other -> Other
+    catch
+        error:badarg -> {error, badarg}
+    end.
 
-% Will add zero termination too
-% The 'EXIT' tuple from a bad argument will eventually generate an error
-% in list_to_binary, which is caught and generates the {error,badarg} return
-pathname(File) ->
-    (catch prim_file:internal_name2native(File)).
+list_dir_nif(_Path) ->
+    erlang:nif_error(undef).
+read_link_nif(_Path) ->
+    erlang:nif_error(undef).
+read_info_nif(_Path, _FollowLinks) ->
+    erlang:nif_error(undef).
+make_hard_link_nif(_Existing, _New) ->
+    erlang:nif_error(undef).
+make_soft_link_nif(_Existing, _New) ->
+    erlang:nif_error(undef).
+rename_nif(_Source, _Destination) ->
+    erlang:nif_error(undef).
+make_dir_nif(_Path) ->
+    erlang:nif_error(undef).
+del_file_nif(_Path) ->
+    erlang:nif_error(undef).
+del_dir_nif(_Path) ->
+    erlang:nif_error(undef).
+get_device_cwd_nif(_DevicePath) ->
+    erlang:nif_error(undef).
+set_cwd_nif(_Path) ->
+    erlang:nif_error(undef).
+get_cwd_nif() ->
+    erlang:nif_error(undef).
+altname_nif(_Path) ->
+    erlang:nif_error(undef).
 
+%%
+%% General helper functions.
+%%
 
-%% proplist:get_value/3
-plgv(K, [{K, V}|_], _) -> V;
-plgv(K, [_|KVs], D)    -> plgv(K, KVs, D);
-plgv(_, [], D)         -> D.
+%% We know for certain that lists:reverse/2 is a BIF, so it's safe to use it
+%% even though this module is preloaded.
+reverse_list(List) -> lists:reverse(List).
+
+proplist_get_value(_Key, [], Default) ->
+    Default;
+proplist_get_value(Key, [{Key, Value} | _Rest], _Default) ->
+    Value;
+proplist_get_value(Key, [Key | _Rest], _Default) ->
+    true;
+proplist_get_value(Key, [_Other | Rest], Default) ->
+    proplist_get_value(Key, Rest, Default).
+
+encode_path(Path) ->
+    prim_file:internal_name2native(Path).
+decode_path(NativePath) when is_binary(NativePath) ->
+    prim_file:internal_native2name(NativePath).
+
+is_path_translatable(Path) when is_list(Path) ->
+    true;
+is_path_translatable(Path) ->
+    prim_file:is_translatable(Path).
 
-%%
 %% We don't actually want this here
+%%
 %% We want to use posix time in all prim but erl_prim_loader makes that tricky
 %% It is probably needed to redo the whole erl_prim_loader
 
-from_seconds(Seconds, posix) when is_integer(Seconds) ->
+from_posix_seconds(Seconds, posix) when is_integer(Seconds) ->
     Seconds;
-from_seconds(Seconds, universal) when is_integer(Seconds) ->
+from_posix_seconds(Seconds, universal) when is_integer(Seconds) ->
     erlang:posixtime_to_universaltime(Seconds);
-from_seconds(Seconds, local) when is_integer(Seconds) ->
+from_posix_seconds(Seconds, local) when is_integer(Seconds) ->
     erlang:universaltime_to_localtime(erlang:posixtime_to_universaltime(Seconds)).
 
-to_seconds(Seconds, posix) when is_integer(Seconds) ->
+to_posix_seconds(Seconds, posix) when is_integer(Seconds) ->
     Seconds;
-to_seconds({_,_} = Datetime, universal) ->
+to_posix_seconds({_,_} = Datetime, universal) ->
     erlang:universaltime_to_posixtime(Datetime);
-to_seconds({_,_} = Datetime, local) ->
+to_posix_seconds({_,_} = Datetime, local) ->
     erlang:universaltime_to_posixtime(erlang:localtime_to_universaltime(Datetime)).
diff --git a/erts/preloaded/src/prim_inet.erl b/erts/preloaded/src/prim_inet.erl
index 017a706a8b..2a3605260d 100644
--- a/erts/preloaded/src/prim_inet.erl
+++ b/erts/preloaded/src/prim_inet.erl
@@ -31,7 +31,7 @@
 -export([connect/3, connect/4, async_connect/4]).
 -export([accept/1, accept/2, async_accept/2]).
 -export([shutdown/2]).
--export([send/2, send/3, sendto/4, sendmsg/3]).
+-export([send/2, send/3, sendto/4, sendmsg/3, sendfile/4]).
 -export([recv/2, recv/3, async_recv/3]).
 -export([unrecv/2]).
 -export([recvfrom/2, recvfrom/3]).
@@ -500,6 +500,55 @@ sendmsg(S, #sctp_sndrcvinfo{}=SRI, Data) when is_port(S) ->
 
 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 %%
+%% SENDFILE(outsock(), Fd, Offset, Length) -> {ok,BytesSent} | {error, Reason}
+%%
+%% send Length data bytes from a file handle, to a socket, starting at Offset
+%%
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+%% "sendfile" is for TCP:
+
+sendfile(S, FileHandle, Offset, Length)
+        when not is_port(S);
+             not is_binary(FileHandle);
+             not is_integer(Offset);
+             not is_integer(Length) ->
+    {error, badarg};
+sendfile(S, FileHandle, Offset, Length) ->
+    case erlang:port_info(S, connected) of
+        {connected, Pid} when Pid =:= self() ->
+            sendfile_1(S, FileHandle, Offset, Length);
+        {connected, Pid} when Pid =/= self() ->
+            {error, not_owner};
+        _Other ->
+            {error, einval}
+    end.
+
+sendfile_1(S, FileHandle, Offset, 0) ->
+    sendfile_1(S, FileHandle, Offset, (1 bsl 63) - 1);
+sendfile_1(_S, _FileHandle, Offset, Length) when
+         Offset < 0; Offset > ((1 bsl 63) - 1);
+         Length < 0; Length > ((1 bsl 63) - 1) ->
+    {error, einval};
+sendfile_1(S, FileHandle, Offset, Length) ->
+    Args = [FileHandle,
+            ?int64(Offset),
+            ?int64(Length)],
+    case ctl_cmd(S, ?TCP_REQ_SENDFILE, Args) of
+        {ok, []} ->
+            receive
+                {sendfile, S, {ok, SentLow, SentHigh}} ->
+                    {ok, SentLow bor (SentHigh bsl 32)};
+                {sendfile, S, {error, Reason}} ->
+                    {error, Reason};
+                {'EXIT', S, _Reason} ->
+                    {error, closed}
+            end;
+        {error, Reason} ->
+            {error, Reason}
+    end.
+
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+%%
 %% RECV(insock(), Length, [Timeout]) -> {ok,Data} | {error, Reason}
 %%
 %% receive Length data bytes from a socket