* sysdeps/unix/sysv/linux/kernel-features.h: Define

	__ASSUME_COMPLETE_READV_WRITEV.
	* sysdeps/unix/sysv/linux/readv.c: No need for userlevel fallback
	with modern kernels.
	* sysdeps/unix/sysv/linux/writev.c: Likewise.

	* sysdeps/posix/readv.c: Since read is a cancellation point we have
	to free a possible malloced buffer in case of cancellation.
	* sysdeps/posix/writev.c: Likewise for write.

c2009-04-01  Ulrich Drepper  <drepper@redhat.com>

6 files changed, 79 insertions, 53 deletions

diff --git a/ChangeLog b/ChangeLog
index 08e0dd3c8f..cae3791ac7 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,15 @@
+2009-04-03  Ulrich Drepper  <drepper@redhat.com>
+
+	* sysdeps/unix/sysv/linux/kernel-features.h: Define
+	__ASSUME_COMPLETE_READV_WRITEV.
+	* sysdeps/unix/sysv/linux/readv.c: No need for userlevel fallback
+	with modern kernels.
+	* sysdeps/unix/sysv/linux/writev.c: Likewise.
+
+	* sysdeps/posix/readv.c: Since read is a cancellation point we have
+	to free a possible malloced buffer in case of cancellation.
+	* sysdeps/posix/writev.c: Likewise for write.
+
 2009-04-02  Ulrich Drepper  <drepper@redhat.com>
 
 	* sysdeps/unix/sysv/linux/bits/socket.h: Add missing protocol numbers.
@@ -29,7 +41,7 @@
 	(R_SPARC_NUM): Update.
 	From Dave Miller <davem@davemloft.net>.
 
-2009-04-01  Ulrich Drepper  <drepper@redhat.com>
+c2009-04-01  Ulrich Drepper  <drepper@redhat.com>
 
 	* sysdeps/unix/sysv/linux/sys/eventfd.h (EFD_SEMAPHORE): Define.
 
diff --git a/sysdeps/posix/readv.c b/sysdeps/posix/readv.c
index f0e78e6662..50bcc91315 100644
--- a/sysdeps/posix/readv.c
+++ b/sysdeps/posix/readv.c
@@ -1,4 +1,4 @@
-/* Copyright (C) 1991, 1992, 1996, 1997, 2002 Free Software Foundation, Inc.
+/* Copyright (C) 1991,1992,1996,1997,2002,2009 Free Software Foundation, Inc.
    This file is part of the GNU C Library.
 
    The GNU C Library is free software; you can redistribute it and/or
@@ -25,24 +25,24 @@
 #include <sys/uio.h>
 #include <errno.h>
 
+
+static void
+ifree (char **ptrp)
+{
+  free (*ptrp);
+}
+
 /* Read data from file descriptor FD, and put the result in the
-   buffers described by VECTOR, which is a vector of COUNT `struct iovec's.
+   buffers described by VECTOR, which is a vector of COUNT 'struct iovec's.
    The buffers are filled in the order specified.
-   Operates just like `read' (see <unistd.h>) except that data are
+   Operates just like 'read' (see <unistd.h>) except that data are
    put in VECTOR instead of a contiguous buffer.  */
 ssize_t
 __libc_readv (int fd, const struct iovec *vector, int count)
 {
-  char *buffer;
-  char *buffer_start;
-  size_t bytes;
-  ssize_t bytes_read;
-  int i;
-  bool use_malloc = false;
-
   /* Find the total number of bytes to be read.  */
-  bytes = 0;
-  for (i = 0; i < count; ++i)
+  size_t bytes = 0;
+  for (int i = 0; i < count; ++i)
     {
       /* Check for ssize_t overflow.  */
       if (SSIZE_MAX - bytes < vector[i].iov_len)
@@ -57,28 +57,25 @@ __libc_readv (int fd, const struct iovec *vector, int count)
      use alloca since it's faster and does not require synchronization
      with other threads.  But we cannot if the amount of memory
      required is too large.  */
+  char *buffer;
+  char *malloced_buffer __attribute__ ((__cleanup__ (ifree))) = NULL;
   if (__libc_use_alloca (bytes))
     buffer = (char *) __alloca (bytes);
   else
     {
-      buffer = (char *) malloc (bytes);
+      malloced_buffer = buffer = (char *) malloc (bytes);
       if (buffer == NULL)
-	/* XXX I don't know whether it is acceptable to try reading
-	   the data in chunks.  Probably not so we just fail here.  */
 	return -1;
-
-      use_malloc = true;
     }
 
   /* Read the data.  */
-  bytes_read = __read (fd, buffer, bytes);
+  ssize_t bytes_read = __read (fd, buffer, bytes);
   if (bytes_read <= 0)
     return -1;
 
   /* Copy the data from BUFFER into the memory specified by VECTOR.  */
   bytes = bytes_read;
-  buffer_start = buffer;
-  for (i = 0; i < count; ++i)
+  for (int i = 0; i < count; ++i)
     {
       size_t copy = MIN (vector[i].iov_len, bytes);
 
@@ -90,9 +87,6 @@ __libc_readv (int fd, const struct iovec *vector, int count)
 	break;
     }
 
-  if (use_malloc)
-    free (buffer_start);
-
   return bytes_read;
 }
 #ifndef __libc_readv
diff --git a/sysdeps/posix/writev.c b/sysdeps/posix/writev.c
index a347cc2eac..203cf9cebd 100644
--- a/sysdeps/posix/writev.c
+++ b/sysdeps/posix/writev.c
@@ -1,4 +1,4 @@
-/* Copyright (C) 1991, 1992, 1996, 1997, 2002 Free Software Foundation, Inc.
+/* Copyright (C) 1991,1992,1996,1997,2002,2009 Free Software Foundation, Inc.
    This file is part of the GNU C Library.
 
    The GNU C Library is free software; you can redistribute it and/or
@@ -25,24 +25,25 @@
 #include <sys/uio.h>
 #include <errno.h>
 
+
+static void
+ifree (char **ptrp)
+{
+  free (*ptrp);
+}
+
+
 /* Write data pointed by the buffers described by VECTOR, which
-   is a vector of COUNT `struct iovec's, to file descriptor FD.
+   is a vector of COUNT 'struct iovec's, to file descriptor FD.
    The data is written in the order specified.
-   Operates just like `write' (see <unistd.h>) except that the data
+   Operates just like 'write' (see <unistd.h>) except that the data
    are taken from VECTOR instead of a contiguous buffer.  */
 ssize_t
 __libc_writev (int fd, const struct iovec *vector, int count)
 {
-  char *buffer;
-  register char *bp;
-  size_t bytes, to_copy;
-  ssize_t bytes_written;
-  int i;
-  bool use_malloc = false;
-
   /* Find the total number of bytes to be written.  */
-  bytes = 0;
-  for (i = 0; i < count; ++i)
+  size_t bytes = 0;
+  for (int i = 0; i < count; ++i)
     {
       /* Check for ssize_t overflow.  */
       if (SSIZE_MAX - bytes < vector[i].iov_len)
@@ -57,23 +58,23 @@ __libc_writev (int fd, const struct iovec *vector, int count)
      use alloca since it's faster and does not require synchronization
      with other threads.  But we cannot if the amount of memory
      required is too large.  */
+  char *buffer;
+  char *malloced_buffer __attribute__ ((__cleanup__ (ifree))) = NULL;
   if (__libc_use_alloca (bytes))
     buffer = (char *) __alloca (bytes);
   else
     {
-      buffer = (char *) malloc (bytes);
+      malloced_buffer = buffer = (char *) malloc (bytes);
       if (buffer == NULL)
 	/* XXX I don't know whether it is acceptable to try writing
 	   the data in chunks.  Probably not so we just fail here.  */
 	return -1;
-
-      use_malloc = true;
     }
 
   /* Copy the data into BUFFER.  */
-  to_copy = bytes;
-  bp = buffer;
-  for (i = 0; i < count; ++i)
+  size_t to_copy = bytes;
+  char *bp = buffer;
+  for (int i = 0; i < count; ++i)
     {
       size_t copy = MIN (vector[i].iov_len, to_copy);
 
@@ -84,10 +85,7 @@ __libc_writev (int fd, const struct iovec *vector, int count)
 	break;
     }
 
-  bytes_written = __write (fd, buffer, bytes);
-
-  if (use_malloc)
-    free (buffer);
+  ssize_t bytes_written = __write (fd, buffer, bytes);
 
   return bytes_written;
 }
diff --git a/sysdeps/unix/sysv/linux/kernel-features.h b/sysdeps/unix/sysv/linux/kernel-features.h
index 29afe27711..9053df1789 100644
--- a/sysdeps/unix/sysv/linux/kernel-features.h
+++ b/sysdeps/unix/sysv/linux/kernel-features.h
@@ -462,6 +462,13 @@
 # define __ASSUME_SET_ROBUST_LIST	1
 #endif
 
+/* Pessimistically assume that 2.6.18 introduced real handling of
+   large numbers of requests to readv and writev and that we don't
+   need a fallback.  It likely worked for much longer.  */
+#if __LINUX_KERNEL_VERSION >= 0x020612
+# define __ASSUME_COMPLETE_READV_WRITEV	1
+#endif
+
 /* Support for PI futexes was added in 2.6.18.  */
 #if __LINUX_KERNEL_VERSION >= 0x020612
 # define __ASSUME_FUTEX_LOCK_PI	1
diff --git a/sysdeps/unix/sysv/linux/readv.c b/sysdeps/unix/sysv/linux/readv.c
index 250c00a075..d12fa49373 100644
--- a/sysdeps/unix/sysv/linux/readv.c
+++ b/sysdeps/unix/sysv/linux/readv.c
@@ -1,5 +1,5 @@
 /* readv supports all Linux kernels >= 2.0.
-   Copyright (C) 1997,1998,2000,2002,2003 Free Software Foundation, Inc.
+   Copyright (C) 1997,1998,2000,2002,2003,2009 Free Software Foundation, Inc.
    This file is part of the GNU C Library.
 
    The GNU C Library is free software; you can redistribute it and/or
@@ -25,6 +25,7 @@
 #include <sysdep-cancel.h>
 #include <sys/syscall.h>
 #include <bp-checks.h>
+#include <kernel-features.h>
 
 static ssize_t __atomic_readv_replacement (int, __const struct iovec *,
 					   int) internal_function;
@@ -45,10 +46,14 @@ do_readv (int fd, const struct iovec *vector, int count)
 
   bytes_read = INLINE_SYSCALL (readv, 3, fd, CHECK_N (vector, count), count);
 
+#ifdef __ASSUME_COMPLETE_READV_WRITEV
+  return bytes_read;
+#else
   if (bytes_read >= 0 || errno != EINVAL || count <= UIO_FASTIOV)
     return bytes_read;
 
   return __atomic_readv_replacement (fd, vector, count);
+#endif
 }
 
 
@@ -72,5 +77,7 @@ __libc_readv (fd, vector, count)
 strong_alias (__libc_readv, __readv)
 weak_alias (__libc_readv, readv)
 
-#define __libc_readv static internal_function __atomic_readv_replacement
-#include <sysdeps/posix/readv.c>
+#ifndef __ASSUME_COMPLETE_READV_WRITEV
+# define __libc_readv static internal_function __atomic_readv_replacement
+# include <sysdeps/posix/readv.c>
+#endif
diff --git a/sysdeps/unix/sysv/linux/writev.c b/sysdeps/unix/sysv/linux/writev.c
index 05978665fa..398a841bb3 100644
--- a/sysdeps/unix/sysv/linux/writev.c
+++ b/sysdeps/unix/sysv/linux/writev.c
@@ -1,5 +1,5 @@
 /* writev supports all Linux kernels >= 2.0.
-   Copyright (C) 1997, 1998, 2000, 2002, 2003 Free Software Foundation, Inc.
+   Copyright (C) 1997,1998,2000,2002,2003,2009 Free Software Foundation, Inc.
    This file is part of the GNU C Library.
 
    The GNU C Library is free software; you can redistribute it and/or
@@ -25,6 +25,7 @@
 #include <sysdep-cancel.h>
 #include <sys/syscall.h>
 #include <bp-checks.h>
+#include <kernel-features.h>
 
 static ssize_t __atomic_writev_replacement (int, const struct iovec *,
 					    int) internal_function;
@@ -43,12 +44,17 @@ do_writev (int fd, const struct iovec *vector, int count)
 {
   ssize_t bytes_written;
 
-  bytes_written = INLINE_SYSCALL (writev, 3, fd, CHECK_N (vector, count), count);
+  bytes_written = INLINE_SYSCALL (writev, 3, fd, CHECK_N (vector, count),
+				  count);
 
+#ifdef __ASSUME_COMPLETE_READV_WRITEV
+  return bytes_written;
+#else
   if (bytes_written >= 0 || errno != EINVAL || count <= UIO_FASTIOV)
     return bytes_written;
 
   return __atomic_writev_replacement (fd, vector, count);
+#endif
 }
 
 ssize_t
@@ -71,5 +77,7 @@ __libc_writev (fd, vector, count)
 strong_alias (__libc_writev, __writev)
 weak_alias (__libc_writev, writev)
 
-#define __libc_writev static internal_function __atomic_writev_replacement
-#include <sysdeps/posix/writev.c>
+#ifndef __ASSUME_COMPLETE_READV_WRITEV
+# define __libc_writev static internal_function __atomic_writev_replacement
+# include <sysdeps/posix/writev.c>
+#endif