Provide pg_preadv() and pg_pwritev().

Provide synchronous vectored file I/O routines.  These map to preadv()
and pwritev(), with fallback implementations for systems that don't have
them.  Also provide a wrapper pg_pwritev_with_retry() that automatically
retries on short writes.

Reviewed-by: Tom Lane <tgl@sss.pgh.pa.us>
Reviewed-by: Andres Freund <andres@anarazel.de>
Discussion: https://postgr.es/m/CA%2BhUKGJA%2Bu-220VONeoREBXJ9P3S94Y7J%2BkqCnTYmahvZJwM%3Dg%40mail.gmail.com

configure

@@ -13061,7 +13061,7 @@ $as_echo "#define HAVE_STDBOOL_H 1" >>confdefs.h
 fi
 
 
-for ac_header in atomic.h copyfile.h execinfo.h getopt.h ifaddrs.h langinfo.h mbarrier.h poll.h sys/epoll.h sys/event.h sys/ipc.h sys/prctl.h sys/procctl.h sys/pstat.h sys/resource.h sys/select.h sys/sem.h sys/shm.h sys/sockio.h sys/tas.h sys/un.h termios.h ucred.h wctype.h
+for ac_header in atomic.h copyfile.h execinfo.h getopt.h ifaddrs.h langinfo.h mbarrier.h poll.h sys/epoll.h sys/event.h sys/ipc.h sys/prctl.h sys/procctl.h sys/pstat.h sys/resource.h sys/select.h sys/sem.h sys/shm.h sys/sockio.h sys/tas.h sys/uio.h sys/un.h termios.h ucred.h wctype.h
 do :
   as_ac_Header=`$as_echo "ac_cv_header_$ac_header" | $as_tr_sh`
 ac_fn_c_check_header_mongrel "$LINENO" "$ac_header" "$as_ac_Header" "$ac_includes_default"
@@ -15155,7 +15155,7 @@ fi
 LIBS_including_readline="$LIBS"
 LIBS=`echo "$LIBS" | sed -e 's/-ledit//g' -e 's/-lreadline//g'`
 
-for ac_func in backtrace_symbols clock_gettime copyfile fdatasync getifaddrs getpeerucred getrlimit kqueue mbstowcs_l memset_s poll posix_fallocate ppoll pstat pthread_is_threaded_np readlink setproctitle setproctitle_fast setsid shm_open strchrnul strsignal symlink sync_file_range uselocale wcstombs_l
+for ac_func in backtrace_symbols clock_gettime copyfile fdatasync getifaddrs getpeerucred getrlimit kqueue mbstowcs_l memset_s poll posix_fallocate ppoll pread preadv pstat pthread_is_threaded_np pwrite pwritev readlink readv setproctitle setproctitle_fast setsid shm_open strchrnul strsignal symlink sync_file_range uselocale wcstombs_l writev
 do :
   as_ac_var=`$as_echo "ac_cv_func_$ac_func" | $as_tr_sh`
 ac_fn_c_check_func "$LINENO" "$ac_func" "$as_ac_var"
@@ -15832,32 +15832,6 @@ esac
 
 fi
 
-ac_fn_c_check_func "$LINENO" "pread" "ac_cv_func_pread"
-if test "x$ac_cv_func_pread" = xyes; then :
-  $as_echo "#define HAVE_PREAD 1" >>confdefs.h
-
-else
-  case " $LIBOBJS " in
-  *" pread.$ac_objext "* ) ;;
-  *) LIBOBJS="$LIBOBJS pread.$ac_objext"
- ;;
-esac
-
-fi
-
-ac_fn_c_check_func "$LINENO" "pwrite" "ac_cv_func_pwrite"
-if test "x$ac_cv_func_pwrite" = xyes; then :
-  $as_echo "#define HAVE_PWRITE 1" >>confdefs.h
-
-else
-  case " $LIBOBJS " in
-  *" pwrite.$ac_objext "* ) ;;
-  *) LIBOBJS="$LIBOBJS pwrite.$ac_objext"
- ;;
-esac
-
-fi
-
 ac_fn_c_check_func "$LINENO" "random" "ac_cv_func_random"
 if test "x$ac_cv_func_random" = xyes; then :
   $as_echo "#define HAVE_RANDOM 1" >>confdefs.h

configure.ac

@@ -1331,6 +1331,7 @@ AC_CHECK_HEADERS(m4_normalize([
 	sys/shm.h
 	sys/sockio.h
 	sys/tas.h
+	sys/uio.h
 	sys/un.h
 	termios.h
 	ucred.h
@@ -1660,9 +1661,14 @@ AC_CHECK_FUNCS(m4_normalize([
 	poll
 	posix_fallocate
 	ppoll
+	pread
+	preadv
 	pstat
 	pthread_is_threaded_np
+	pwrite
+	pwritev
 	readlink
+	readv
 	setproctitle
 	setproctitle_fast
 	setsid
@@ -1673,6 +1679,7 @@ AC_CHECK_FUNCS(m4_normalize([
 	sync_file_range
 	uselocale
 	wcstombs_l
+	writev
 ]))
 
 # These typically are compiler builtins, for which AC_CHECK_FUNCS fails.
@@ -1733,8 +1740,6 @@ AC_REPLACE_FUNCS(m4_normalize([
 	inet_aton
 	link
 	mkdtemp
-	pread
-	pwrite
 	random
 	srandom
 	strlcat

src/include/pg_config.h.in

@@ -412,6 +412,9 @@
 /* Define to 1 if you have the `pread' function. */
 #undef HAVE_PREAD
 
+/* Define to 1 if you have the `preadv' function. */
+#undef HAVE_PREADV
+
 /* Define to 1 if you have the `pstat' function. */
 #undef HAVE_PSTAT
 
@@ -430,6 +433,9 @@
 /* Define to 1 if you have the `pwrite' function. */
 #undef HAVE_PWRITE
 
+/* Define to 1 if you have the `pwritev' function. */
+#undef HAVE_PWRITEV
+
 /* Define to 1 if you have the `random' function. */
 #undef HAVE_RANDOM
 
@@ -445,6 +451,9 @@
 /* Define to 1 if you have the `readlink' function. */
 #undef HAVE_READLINK
 
+/* Define to 1 if you have the `readv' function. */
+#undef HAVE_READV
+
 /* Define to 1 if you have the global variable
    'rl_completion_append_character'. */
 #undef HAVE_RL_COMPLETION_APPEND_CHARACTER
@@ -629,6 +638,9 @@
 /* Define to 1 if you have the <sys/ucred.h> header file. */
 #undef HAVE_SYS_UCRED_H
 
+/* Define to 1 if you have the <sys/uio.h> header file. */
+#undef HAVE_SYS_UIO_H
+
 /* Define to 1 if you have the <sys/un.h> header file. */
 #undef HAVE_SYS_UN_H
 
@@ -683,6 +695,9 @@
 /* Define to 1 if you have the <winldap.h> header file. */
 #undef HAVE_WINLDAP_H
 
+/* Define to 1 if you have the `writev' function. */
+#undef HAVE_WRITEV
+
 /* Define to 1 if you have the `X509_get_signature_nid' function. */
 #undef HAVE_X509_GET_SIGNATURE_NID
 

src/include/port.h

@@ -431,6 +431,8 @@ extern ssize_t pg_pread(int fd, void *buf, size_t nbyte, off_t offset);
 extern ssize_t pg_pwrite(int fd, const void *buf, size_t nbyte, off_t offset);
 #endif
 
+/* For pg_pwritev() and pg_preadv(), see port/pg_iovec.h. */
+
 #if !HAVE_DECL_STRLCAT
 extern size_t strlcat(char *dst, const char *src, size_t siz);
 #endif

src/include/port/pg_iovec.h

@@ -0,0 +1,59 @@
+/*-------------------------------------------------------------------------
+ *
+ * pg_iovec.h
+ *	  Header for the vectored I/O functions in src/port/p{read,write}.c.
+ *
+ * Portions Copyright (c) 1996-2020, PostgreSQL Global Development Group
+ * Portions Copyright (c) 1994, Regents of the University of California
+ *
+ * src/include/port/pg_iovec.h
+ *
+ *-------------------------------------------------------------------------
+ */
+#ifndef PG_IOVEC_H
+#define PG_IOVEC_H
+
+#include <limits.h>
+
+#ifdef HAVE_SYS_UIO_H
+#include <sys/uio.h>
+#endif
+
+/* If <sys/uio.h> is missing, define our own POSIX-compatible iovec struct. */
+#ifndef HAVE_SYS_UIO_H
+struct iovec
+{
+	void	   *iov_base;
+	size_t		iov_len;
+};
+#endif
+
+/*
+ * If <limits.h> didn't define IOV_MAX, define our own.  POSIX requires at
+ * least 16.
+ */
+#ifndef IOV_MAX
+#define IOV_MAX 16
+#endif
+
+/* Define a reasonable maximum that is safe to use on the stack. */
+#define PG_IOV_MAX Min(IOV_MAX, 32)
+
+#ifdef HAVE_PREADV
+#define pg_preadv preadv
+#else
+extern ssize_t pg_preadv(int fd, const struct iovec *iov, int iovcnt, off_t offset);
+#endif
+
+#ifdef HAVE_PWRITEV
+#define pg_pwritev pwritev
+#else
+extern ssize_t pg_pwritev(int fd, const struct iovec *iov, int iovcnt, off_t offset);
+#endif
+
+extern ssize_t pg_pwritev_with_retry(int fd,
+									 const struct iovec *iov,
+									 int iovcnt,
+									 off_t offset);
+
+#endif							/* PG_IOVEC_H */

src/port/Makefile

@@ -53,6 +53,8 @@ OBJS = \
 	pgstrcasecmp.o \
 	pgstrsignal.o \
 	pqsignal.o \
+	pread.o \
+	pwrite.o \
 	qsort.o \
 	qsort_arg.o \
 	quotes.o \

src/port/pread.c

@@ -1,7 +1,7 @@
 /*-------------------------------------------------------------------------
  *
  * pread.c
- *	  Implementation of pread(2) for platforms that lack one.
+ *	  Implementation of pread[v](2) for platforms that lack one.
  *
  * Portions Copyright (c) 1996-2021, PostgreSQL Global Development Group
  *
@@ -9,7 +9,8 @@
  *	  src/port/pread.c
  *
  * Note that this implementation changes the current file position, unlike
- * the POSIX function, so we use the name pg_pread().
+ * the POSIX function, so we use the name pg_pread().  Likewise for the
+ * iovec version.
  *
  *-------------------------------------------------------------------------
  */
@@ -23,6 +24,9 @@
 #include <unistd.h>
 #endif
 
+#include "port/pg_iovec.h"
+
+#ifndef HAVE_PREAD
 ssize_t
 pg_pread(int fd, void *buf, size_t size, off_t offset)
 {
@@ -56,3 +60,38 @@ pg_pread(int fd, void *buf, size_t size, off_t offset)
 	return read(fd, buf, size);
 #endif
 }
+#endif
+
+#ifndef HAVE_PREADV
+ssize_t
+pg_preadv(int fd, const struct iovec *iov, int iovcnt, off_t offset)
+{
+#ifdef HAVE_READV
+	if (iovcnt == 1)
+		return pg_pread(fd, iov[0].iov_base, iov[0].iov_len, offset);
+	if (lseek(fd, offset, SEEK_SET) < 0)
+		return -1;
+	return readv(fd, iov, iovcnt);
+#else
+	ssize_t 	sum = 0;
+	ssize_t 	part;
+
+	for (int i = 0; i < iovcnt; ++i)
+	{
+		part = pg_pread(fd, iov[i].iov_base, iov[i].iov_len, offset);
+		if (part < 0)
+		{
+			if (i == 0)
+				return -1;
+			else
+				return sum;
+		}
+		sum += part;
+		offset += part;
+		if (part < iov[i].iov_len)
+			return sum;
+	}
+	return sum;
+#endif
+}
+#endif

src/port/pwrite.c

@@ -1,7 +1,7 @@
 /*-------------------------------------------------------------------------
  *
  * pwrite.c
- *	  Implementation of pwrite(2) for platforms that lack one.
+ *	  Implementation of pwrite[v](2) for platforms that lack one.
  *
  * Portions Copyright (c) 1996-2021, PostgreSQL Global Development Group
  *
@@ -9,7 +9,8 @@
  *	  src/port/pwrite.c
  *
  * Note that this implementation changes the current file position, unlike
- * the POSIX function, so we use the name pg_pwrite().
+ * the POSIX function, so we use the name pg_pwrite().  Likewise for the
+ * iovec version.
  *
  *-------------------------------------------------------------------------
  */
@@ -23,6 +24,9 @@
 #include <unistd.h>
 #endif
 
+#include "port/pg_iovec.h"
+
+#ifndef HAVE_PWRITE
 ssize_t
 pg_pwrite(int fd, const void *buf, size_t size, off_t offset)
 {
@@ -53,3 +57,102 @@ pg_pwrite(int fd, const void *buf, size_t size, off_t offset)
 	return write(fd, buf, size);
 #endif
 }
+#endif
+
+#ifndef HAVE_PWRITEV
+ssize_t
+pg_pwritev(int fd, const struct iovec *iov, int iovcnt, off_t offset)
+{
+#ifdef HAVE_WRITEV
+	if (iovcnt == 1)
+		return pg_pwrite(fd, iov[0].iov_base, iov[0].iov_len, offset);
+	if (lseek(fd, offset, SEEK_SET) < 0)
+		return -1;
+	return writev(fd, iov, iovcnt);
+#else
+	ssize_t 	sum = 0;
+	ssize_t 	part;
+
+	for (int i = 0; i < iovcnt; ++i)
+	{
+		part = pg_pwrite(fd, iov[i].iov_base, iov[i].iov_len, offset);
+		if (part < 0)
+		{
+			if (i == 0)
+				return -1;
+			else
+				return sum;
+		}
+		sum += part;
+		offset += part;
+		if (part < iov[i].iov_len)
+			return sum;
+	}
+	return sum;
+#endif
+}
+#endif
+
+/*
+ * A convenience wrapper for pg_pwritev() that retries on partial write.  If an
+ * error is returned, it is unspecified how much has been written.
+ */
+ssize_t
+pg_pwritev_with_retry(int fd, const struct iovec *iov, int iovcnt, off_t offset)
+{
+	struct iovec iov_copy[PG_IOV_MAX];
+	ssize_t		sum = 0;
+	ssize_t		part;
+
+	/* We'd better have space to make a copy, in case we need to retry. */
+	if (iovcnt > PG_IOV_MAX)
+	{
+		errno = EINVAL;
+		return -1;
+	}
+
+	for (;;)
+	{
+		/* Write as much as we can. */
+		part = pg_pwritev(fd, iov, iovcnt, offset);
+		if (part < 0)
+			return -1;
+
+#ifdef SIMULATE_SHORT_WRITE
+		part = Min(part, 4096);
+#endif
+
+		/* Count our progress. */
+		sum += part;
+		offset += part;
+
+		/* Step over iovecs that are done. */
+		while (iovcnt > 0 && iov->iov_len <= part)
+		{
+			part -= iov->iov_len;
+			++iov;
+			--iovcnt;
+		}
+
+		/* Are they all done? */
+		if (iovcnt == 0)
+		{
+			if (part > 0)
+				elog(ERROR, "unexpectedly wrote more than requested");
+			break;
+		}
+
+		/*
+		 * Move whatever's left to the front of our mutable copy and adjust the
+		 * leading iovec.
+		 */
+		Assert(iovcnt > 0);
+		memmove(iov_copy, iov, sizeof(*iov) * iovcnt);
+		Assert(iov->iov_len > part);
+		iov_copy[0].iov_base = (char *) iov_copy[0].iov_base + part;
+		iov_copy[0].iov_len -= part;
+		iov = iov_copy;
+	}
+
+	return sum;
+}

src/tools/msvc/Solution.pm

@@ -329,17 +329,20 @@ sub GenerateFiles
 		HAVE_PPC_LWARX_MUTEX_HINT   => undef,
 		HAVE_PPOLL                  => undef,
 		HAVE_PREAD                  => undef,
+		HAVE_PREADV                 => undef,
 		HAVE_PSTAT                  => undef,
 		HAVE_PS_STRINGS             => undef,
 		HAVE_PTHREAD                => undef,
 		HAVE_PTHREAD_IS_THREADED_NP => undef,
 		HAVE_PTHREAD_PRIO_INHERIT   => undef,
 		HAVE_PWRITE                 => undef,
+		HAVE_PWRITEV                => undef,
 		HAVE_RANDOM                 => undef,
 		HAVE_READLINE_H             => undef,
 		HAVE_READLINE_HISTORY_H     => undef,
 		HAVE_READLINE_READLINE_H    => undef,
 		HAVE_READLINK               => undef,
+		HAVE_READV                  => undef,
 		HAVE_RL_COMPLETION_APPEND_CHARACTER      => undef,
 		HAVE_RL_COMPLETION_MATCHES               => undef,
 		HAVE_RL_COMPLETION_SUPPRESS_QUOTE        => undef,
@@ -400,6 +403,7 @@ sub GenerateFiles
 		HAVE_SYS_TAS_H                           => undef,
 		HAVE_SYS_TYPES_H                         => 1,
 		HAVE_SYS_UCRED_H                         => undef,
+		HAVE_SYS_UIO_H                           => undef,
 		HAVE_SYS_UN_H                            => undef,
 		HAVE_TERMIOS_H                           => undef,
 		HAVE_TYPEOF                              => undef,
@@ -418,6 +422,7 @@ sub GenerateFiles
 		HAVE_WINLDAP_H                           => undef,
 		HAVE_WCSTOMBS_L                          => 1,
 		HAVE_WCTYPE_H                            => 1,
+		HAVE_WRITEV                              => undef,
 		HAVE_X509_GET_SIGNATURE_NID              => 1,
 		HAVE_X86_64_POPCNTQ                      => undef,
 		HAVE__BOOL                               => undef,