Create and use wait events for read, write, and fsync operations.

Previous commits, notably 53be0b1add and
6f3bd98ebf, made it possible to see from
pg_stat_activity when a backend was stuck waiting for another backend,
but it's also fairly common for a backend to be stuck waiting for an
I/O.  Add wait events for those operations, too.

Rushabh Lathia, with further hacking by me.  Reviewed and tested by
Michael Paquier, Amit Kapila, Rajkumar Raghuwanshi, and Rahila Syed.

Discussion: http://postgr.es/m/CAGPqQf0LsYHXREPAZqYGVkDqHSyjf=KsD=k0GTVPAuzyThh-VQ@mail.gmail.com

src/backend/storage/file/buffile.c

@@ -37,6 +37,7 @@
 #include "postgres.h"
 
 #include "executor/instrument.h"
+#include "pgstat.h"
 #include "storage/fd.h"
 #include "storage/buffile.h"
 #include "storage/buf_internals.h"
@@ -254,7 +255,10 @@ BufFileLoadBuffer(BufFile *file)
 	/*
 	 * Read whatever we can get, up to a full bufferload.
 	 */
-	file->nbytes = FileRead(thisfile, file->buffer, sizeof(file->buffer));
+	file->nbytes = FileRead(thisfile,
+							file->buffer,
+							sizeof(file->buffer),
+							WAIT_EVENT_BUFFILE_READ);
 	if (file->nbytes < 0)
 		file->nbytes = 0;
 	file->offsets[file->curFile] += file->nbytes;
@@ -317,7 +321,10 @@ BufFileDumpBuffer(BufFile *file)
 				return;			/* seek failed, give up */
 			file->offsets[file->curFile] = file->curOffset;
 		}
-		bytestowrite = FileWrite(thisfile, file->buffer + wpos, bytestowrite);
+		bytestowrite = FileWrite(thisfile,
+								 file->buffer + wpos,
+								 bytestowrite,
+								 WAIT_EVENT_BUFFILE_WRITE);
 		if (bytestowrite <= 0)
 			return;				/* failed to write */
 		file->offsets[file->curFile] += bytestowrite;

src/backend/storage/file/copydir.c

@@ -25,7 +25,7 @@
 #include "storage/copydir.h"
 #include "storage/fd.h"
 #include "miscadmin.h"
-
+#include "pgstat.h"
 
 /*
  * copydir: copy a directory
@@ -169,7 +169,9 @@ copy_file(char *fromfile, char *tofile)
 		/* If we got a cancel signal during the copy of the file, quit */
 		CHECK_FOR_INTERRUPTS();
 
+		pgstat_report_wait_start(WAIT_EVENT_COPY_FILE_READ);
 		nbytes = read(srcfd, buffer, COPY_BUF_SIZE);
+		pgstat_report_wait_end();
 		if (nbytes < 0)
 			ereport(ERROR,
 					(errcode_for_file_access(),
@@ -177,8 +179,10 @@ copy_file(char *fromfile, char *tofile)
 		if (nbytes == 0)
 			break;
 		errno = 0;
+		pgstat_report_wait_start(WAIT_EVENT_COPY_FILE_WRITE);
 		if ((int) write(dstfd, buffer, nbytes) != nbytes)
 		{
+			pgstat_report_wait_end();
 			/* if write didn't set errno, assume problem is no disk space */
 			if (errno == 0)
 				errno = ENOSPC;
@@ -186,6 +190,7 @@ copy_file(char *fromfile, char *tofile)
 					(errcode_for_file_access(),
 					 errmsg("could not write to file \"%s\": %m", tofile)));
 		}
+		pgstat_report_wait_end();
 
 		/*
 		 * We fsync the files later but first flush them to avoid spamming the

src/backend/storage/file/fd.c

@@ -1550,7 +1550,7 @@ FileClose(File file)
  * to read into.
  */
 int
-FilePrefetch(File file, off_t offset, int amount)
+FilePrefetch(File file, off_t offset, int amount, uint32 wait_event_info)
 {
 #if defined(USE_POSIX_FADVISE) && defined(POSIX_FADV_WILLNEED)
 	int			returnCode;
@@ -1565,8 +1565,10 @@ FilePrefetch(File file, off_t offset, int amount)
 	if (returnCode < 0)
 		return returnCode;
 
+	pgstat_report_wait_start(wait_event_info);
 	returnCode = posix_fadvise(VfdCache[file].fd, offset, amount,
 							   POSIX_FADV_WILLNEED);
+	pgstat_report_wait_end();
 
 	return returnCode;
 #else
@@ -1576,7 +1578,7 @@ FilePrefetch(File file, off_t offset, int amount)
 }
 
 void
-FileWriteback(File file, off_t offset, off_t nbytes)
+FileWriteback(File file, off_t offset, off_t nbytes, uint32 wait_event_info)
 {
 	int			returnCode;
 
@@ -1597,11 +1599,13 @@ FileWriteback(File file, off_t offset, off_t nbytes)
 	if (returnCode < 0)
 		return;
 
+	pgstat_report_wait_start(wait_event_info);
 	pg_flush_data(VfdCache[file].fd, offset, nbytes);
+	pgstat_report_wait_end();
 }
 
 int
-FileRead(File file, char *buffer, int amount)
+FileRead(File file, char *buffer, int amount, uint32 wait_event_info)
 {
 	int			returnCode;
 	Vfd		   *vfdP;
@@ -1620,7 +1624,9 @@ FileRead(File file, char *buffer, int amount)
 	vfdP = &VfdCache[file];
 
 retry:
+	pgstat_report_wait_start(wait_event_info);
 	returnCode = read(vfdP->fd, buffer, amount);
+	pgstat_report_wait_end();
 
 	if (returnCode >= 0)
 	{
@@ -1663,7 +1669,7 @@ retry:
 }
 
 int
-FileWrite(File file, char *buffer, int amount)
+FileWrite(File file, char *buffer, int amount, uint32 wait_event_info)
 {
 	int			returnCode;
 	Vfd		   *vfdP;
@@ -1721,7 +1727,9 @@ FileWrite(File file, char *buffer, int amount)
 
 retry:
 	errno = 0;
+	pgstat_report_wait_start(wait_event_info);
 	returnCode = write(vfdP->fd, buffer, amount);
+	pgstat_report_wait_end();
 
 	/* if write didn't set errno, assume problem is no disk space */
 	if (returnCode != amount && errno == 0)
@@ -1782,7 +1790,7 @@ retry:
 }
 
 int
-FileSync(File file)
+FileSync(File file, uint32 wait_event_info)
 {
 	int			returnCode;
 
@@ -1795,7 +1803,11 @@ FileSync(File file)
 	if (returnCode < 0)
 		return returnCode;
 
-	return pg_fsync(VfdCache[file].fd);
+	pgstat_report_wait_start(wait_event_info);
+	returnCode = pg_fsync(VfdCache[file].fd);
+	pgstat_report_wait_end();
+
+	return returnCode;
 }
 
 off_t
@@ -1887,7 +1899,7 @@ FileTell(File file)
 #endif
 
 int
-FileTruncate(File file, off_t offset)
+FileTruncate(File file, off_t offset, uint32 wait_event_info)
 {
 	int			returnCode;
 
@@ -1900,7 +1912,9 @@ FileTruncate(File file, off_t offset)
 	if (returnCode < 0)
 		return returnCode;
 
+	pgstat_report_wait_start(wait_event_info);
 	returnCode = ftruncate(VfdCache[file].fd, offset);
+	pgstat_report_wait_end();
 
 	if (returnCode == 0 && VfdCache[file].fileSize > offset)
 	{

src/backend/storage/ipc/dsm_impl.c

@@ -60,6 +60,7 @@
 #ifdef HAVE_SYS_SHM_H
 #include <sys/shm.h>
 #endif
+#include "pgstat.h"
 
 #include "portability/mem.h"
 #include "storage/dsm_impl.h"
@@ -911,10 +912,12 @@ dsm_impl_mmap(dsm_op op, dsm_handle handle, Size request_size,
 
 			if (goal > ZBUFFER_SIZE)
 				goal = ZBUFFER_SIZE;
+			pgstat_report_wait_start(WAIT_EVENT_DSM_FILL_ZERO_WRITE);
 			if (write(fd, zbuffer, goal) == goal)
 				remaining -= goal;
 			else
 				success = false;
+			pgstat_report_wait_end();
 		}
 
 		if (!success)

src/backend/storage/smgr/md.c

@@ -28,6 +28,7 @@
 #include "miscadmin.h"
 #include "access/xlog.h"
 #include "catalog/catalog.h"
+#include "pgstat.h"
 #include "portability/instr_time.h"
 #include "postmaster/bgwriter.h"
 #include "storage/fd.h"
@@ -536,7 +537,7 @@ mdextend(SMgrRelation reln, ForkNumber forknum, BlockNumber blocknum,
 				 errmsg("could not seek to block %u in file \"%s\": %m",
 						blocknum, FilePathName(v->mdfd_vfd))));
 
-	if ((nbytes = FileWrite(v->mdfd_vfd, buffer, BLCKSZ)) != BLCKSZ)
+	if ((nbytes = FileWrite(v->mdfd_vfd, buffer, BLCKSZ, WAIT_EVENT_DATA_FILE_EXTEND)) != BLCKSZ)
 	{
 		if (nbytes < 0)
 			ereport(ERROR,
@@ -667,7 +668,7 @@ mdprefetch(SMgrRelation reln, ForkNumber forknum, BlockNumber blocknum)
 
 	Assert(seekpos < (off_t) BLCKSZ * RELSEG_SIZE);
 
-	(void) FilePrefetch(v->mdfd_vfd, seekpos, BLCKSZ);
+	(void) FilePrefetch(v->mdfd_vfd, seekpos, BLCKSZ, WAIT_EVENT_DATA_FILE_PREFETCH);
 #endif   /* USE_PREFETCH */
 }
 
@@ -716,7 +717,7 @@ mdwriteback(SMgrRelation reln, ForkNumber forknum,
 
 		seekpos = (off_t) BLCKSZ *(blocknum % ((BlockNumber) RELSEG_SIZE));
 
-		FileWriteback(v->mdfd_vfd, seekpos, (off_t) BLCKSZ * nflush);
+		FileWriteback(v->mdfd_vfd, seekpos, (off_t) BLCKSZ * nflush, WAIT_EVENT_DATA_FILE_FLUSH);
 
 		nblocks -= nflush;
 		blocknum += nflush;
@@ -753,7 +754,7 @@ mdread(SMgrRelation reln, ForkNumber forknum, BlockNumber blocknum,
 				 errmsg("could not seek to block %u in file \"%s\": %m",
 						blocknum, FilePathName(v->mdfd_vfd))));
 
-	nbytes = FileRead(v->mdfd_vfd, buffer, BLCKSZ);
+	nbytes = FileRead(v->mdfd_vfd, buffer, BLCKSZ, WAIT_EVENT_DATA_FILE_READ);
 
 	TRACE_POSTGRESQL_SMGR_MD_READ_DONE(forknum, blocknum,
 									   reln->smgr_rnode.node.spcNode,
@@ -829,7 +830,7 @@ mdwrite(SMgrRelation reln, ForkNumber forknum, BlockNumber blocknum,
 				 errmsg("could not seek to block %u in file \"%s\": %m",
 						blocknum, FilePathName(v->mdfd_vfd))));
 
-	nbytes = FileWrite(v->mdfd_vfd, buffer, BLCKSZ);
+	nbytes = FileWrite(v->mdfd_vfd, buffer, BLCKSZ, WAIT_EVENT_DATA_FILE_WRITE);
 
 	TRACE_POSTGRESQL_SMGR_MD_WRITE_DONE(forknum, blocknum,
 										reln->smgr_rnode.node.spcNode,
@@ -967,7 +968,7 @@ mdtruncate(SMgrRelation reln, ForkNumber forknum, BlockNumber nblocks)
 			 * This segment is no longer active. We truncate the file, but do
 			 * not delete it, for reasons explained in the header comments.
 			 */
-			if (FileTruncate(v->mdfd_vfd, 0) < 0)
+			if (FileTruncate(v->mdfd_vfd, 0, WAIT_EVENT_DATA_FILE_TRUNCATE) < 0)
 				ereport(ERROR,
 						(errcode_for_file_access(),
 						 errmsg("could not truncate file \"%s\": %m",
@@ -993,7 +994,7 @@ mdtruncate(SMgrRelation reln, ForkNumber forknum, BlockNumber nblocks)
 			 */
 			BlockNumber lastsegblocks = nblocks - priorblocks;
 
-			if (FileTruncate(v->mdfd_vfd, (off_t) lastsegblocks * BLCKSZ) < 0)
+			if (FileTruncate(v->mdfd_vfd, (off_t) lastsegblocks * BLCKSZ, WAIT_EVENT_DATA_FILE_TRUNCATE) < 0)
 				ereport(ERROR,
 						(errcode_for_file_access(),
 					errmsg("could not truncate file \"%s\" to %u blocks: %m",
@@ -1037,7 +1038,7 @@ mdimmedsync(SMgrRelation reln, ForkNumber forknum)
 	{
 		MdfdVec    *v = &reln->md_seg_fds[forknum][segno - 1];
 
-		if (FileSync(v->mdfd_vfd) < 0)
+		if (FileSync(v->mdfd_vfd, WAIT_EVENT_DATA_FILE_IMMEDIATE_SYNC) < 0)
 			ereport(ERROR,
 					(errcode_for_file_access(),
 					 errmsg("could not fsync file \"%s\": %m",
@@ -1232,7 +1233,7 @@ mdsync(void)
 					INSTR_TIME_SET_CURRENT(sync_start);
 
 					if (seg != NULL &&
-						FileSync(seg->mdfd_vfd) >= 0)
+						FileSync(seg->mdfd_vfd, WAIT_EVENT_DATA_FILE_SYNC) >= 0)
 					{
 						/* Success; update statistics about sync timing */
 						INSTR_TIME_SET_CURRENT(sync_end);
@@ -1443,7 +1444,7 @@ register_dirty_segment(SMgrRelation reln, ForkNumber forknum, MdfdVec *seg)
 		ereport(DEBUG1,
 				(errmsg("could not forward fsync request because request queue is full")));
 
-		if (FileSync(seg->mdfd_vfd) < 0)
+		if (FileSync(seg->mdfd_vfd, WAIT_EVENT_DATA_FILE_SYNC) < 0)
 			ereport(ERROR,
 					(errcode_for_file_access(),
 					 errmsg("could not fsync file \"%s\": %m",