Add WL_EXIT_ON_PM_DEATH pseudo-event.

Users of the WaitEventSet and WaitLatch() APIs can now choose between
asking for WL_POSTMASTER_DEATH and then handling it explicitly, or asking
for WL_EXIT_ON_PM_DEATH to trigger immediate exit on postmaster death.
This reduces code duplication, since almost all callers want the latter.

Repair all code that was previously ignoring postmaster death completely,
or requesting the event but ignoring it, or requesting the event but then
doing an unconditional PostmasterIsAlive() call every time through its
event loop (which is an expensive syscall on platforms for which we don't
have USE_POSTMASTER_DEATH_SIGNAL support).

Assert that callers of WaitLatchXXX() under the postmaster remember to
ask for either WL_POSTMASTER_DEATH or WL_EXIT_ON_PM_DEATH, to prevent
future bugs.

The only process that doesn't handle postmaster death is syslogger.  It
waits until all backends holding the write end of the syslog pipe
(including the postmaster) have closed it by exiting, to be sure to
capture any parting messages.  By using the WaitEventSet API directly
it avoids the new assertion, and as a by-product it may be slightly
more efficient on platforms that have epoll().

Author: Thomas Munro
Reviewed-by: Kyotaro Horiguchi, Heikki Linnakangas, Tom Lane
Discussion: https://postgr.es/m/CAEepm%3D1TCviRykkUb69ppWLr_V697rzd1j3eZsRMmbXvETfqbQ%40mail.gmail.com,
            https://postgr.es/m/CAEepm=2LqHzizbe7muD7-2yHUbTOoF7Q+qkSD5Q41kuhttRTwA@mail.gmail.com

src/backend/access/transam/parallel.c

@@ -692,13 +692,9 @@ WaitForParallelWorkersToAttach(ParallelContext *pcxt)
 				 * just end up waiting for the same worker again.
 				 */
 				rc = WaitLatch(MyLatch,
-							   WL_LATCH_SET | WL_POSTMASTER_DEATH,
+							   WL_LATCH_SET | WL_EXIT_ON_PM_DEATH,
 							   -1, WAIT_EVENT_BGWORKER_STARTUP);
 
-				/* emergency bailout if postmaster has died */
-				if (rc & WL_POSTMASTER_DEATH)
-					proc_exit(1);
-
 				if (rc & WL_LATCH_SET)
 					ResetLatch(MyLatch);
 			}
@@ -815,8 +811,8 @@ WaitForParallelWorkersToFinish(ParallelContext *pcxt)
 			}
 		}
 
-		WaitLatch(MyLatch, WL_LATCH_SET, -1,
-				  WAIT_EVENT_PARALLEL_FINISH);
+		(void) WaitLatch(MyLatch, WL_LATCH_SET | WL_EXIT_ON_PM_DEATH, -1,
+						 WAIT_EVENT_PARALLEL_FINISH);
 		ResetLatch(MyLatch);
 	}
 

src/backend/access/transam/xlog.c

@@ -6189,10 +6189,10 @@ recoveryApplyDelay(XLogReaderState *record)
 		elog(DEBUG2, "recovery apply delay %ld seconds, %d milliseconds",
 			 secs, microsecs / 1000);
 
-		WaitLatch(&XLogCtl->recoveryWakeupLatch,
-				  WL_LATCH_SET | WL_TIMEOUT | WL_POSTMASTER_DEATH,
-				  secs * 1000L + microsecs / 1000,
-				  WAIT_EVENT_RECOVERY_APPLY_DELAY);
+		(void) WaitLatch(&XLogCtl->recoveryWakeupLatch,
+						 WL_LATCH_SET | WL_TIMEOUT | WL_EXIT_ON_PM_DEATH,
+						 secs * 1000L + microsecs / 1000,
+						 WAIT_EVENT_RECOVERY_APPLY_DELAY);
 	}
 	return true;
 }
@@ -12093,9 +12093,11 @@ WaitForWALToBecomeAvailable(XLogRecPtr RecPtr, bool randAccess,
 						wait_time = wal_retrieve_retry_interval -
 							(secs * 1000 + usecs / 1000);
 
-						WaitLatch(&XLogCtl->recoveryWakeupLatch,
-								  WL_LATCH_SET | WL_TIMEOUT | WL_POSTMASTER_DEATH,
-								  wait_time, WAIT_EVENT_RECOVERY_WAL_STREAM);
+						(void) WaitLatch(&XLogCtl->recoveryWakeupLatch,
+										 WL_LATCH_SET | WL_TIMEOUT |
+										 WL_EXIT_ON_PM_DEATH,
+										 wait_time,
+										 WAIT_EVENT_RECOVERY_WAL_STREAM);
 						ResetLatch(&XLogCtl->recoveryWakeupLatch);
 						now = GetCurrentTimestamp();
 					}
@@ -12269,9 +12271,10 @@ WaitForWALToBecomeAvailable(XLogRecPtr RecPtr, bool randAccess,
 					 * Wait for more WAL to arrive. Time out after 5 seconds
 					 * to react to a trigger file promptly.
 					 */
-					WaitLatch(&XLogCtl->recoveryWakeupLatch,
-							  WL_LATCH_SET | WL_TIMEOUT | WL_POSTMASTER_DEATH,
-							  5000L, WAIT_EVENT_RECOVERY_WAL_ALL);
+					(void) WaitLatch(&XLogCtl->recoveryWakeupLatch,
+									 WL_LATCH_SET | WL_TIMEOUT |
+									 WL_EXIT_ON_PM_DEATH,
+									 5000L, WAIT_EVENT_RECOVERY_WAL_ALL);
 					ResetLatch(&XLogCtl->recoveryWakeupLatch);
 					break;
 				}

src/backend/access/transam/xlogfuncs.c

@@ -764,10 +764,10 @@ pg_promote(PG_FUNCTION_ARGS)
 
 		CHECK_FOR_INTERRUPTS();
 
-		WaitLatch(MyLatch,
-				  WL_LATCH_SET | WL_TIMEOUT | WL_POSTMASTER_DEATH,
-				  1000L / WAITS_PER_SECOND,
-				  WAIT_EVENT_PROMOTE);
+		(void) WaitLatch(MyLatch,
+						 WL_LATCH_SET | WL_TIMEOUT | WL_POSTMASTER_DEATH,
+						 1000L / WAITS_PER_SECOND,
+						 WAIT_EVENT_PROMOTE);
 	}
 
 	ereport(WARNING,

src/backend/executor/nodeGather.c

@@ -383,7 +383,8 @@ gather_readnext(GatherState *gatherstate)
 				return NULL;
 
 			/* Nothing to do except wait for developments. */
-			WaitLatch(MyLatch, WL_LATCH_SET, 0, WAIT_EVENT_EXECUTE_GATHER);
+			(void) WaitLatch(MyLatch, WL_LATCH_SET | WL_EXIT_ON_PM_DEATH, 0,
+							 WAIT_EVENT_EXECUTE_GATHER);
 			ResetLatch(MyLatch);
 			nvisited = 0;
 		}

src/backend/libpq/be-secure-openssl.c

@@ -410,8 +410,8 @@ aloop:
 				else
 					waitfor = WL_SOCKET_WRITEABLE;
 
-				WaitLatchOrSocket(MyLatch, waitfor, port->sock, 0,
-								  WAIT_EVENT_SSL_OPEN_SERVER);
+				(void) WaitLatchOrSocket(MyLatch, waitfor, port->sock, 0,
+										 WAIT_EVENT_SSL_OPEN_SERVER);
 				goto aloop;
 			case SSL_ERROR_SYSCALL:
 				if (r < 0)

src/backend/libpq/pqmq.c

@@ -168,8 +168,8 @@ mq_putmessage(char msgtype, const char *s, size_t len)
 		if (result != SHM_MQ_WOULD_BLOCK)
 			break;
 
-		WaitLatch(MyLatch, WL_LATCH_SET, 0,
-				  WAIT_EVENT_MQ_PUT_MESSAGE);
+		(void) WaitLatch(MyLatch, WL_LATCH_SET | WL_EXIT_ON_PM_DEATH, 0,
+						 WAIT_EVENT_MQ_PUT_MESSAGE);
 		ResetLatch(MyLatch);
 		CHECK_FOR_INTERRUPTS();
 	}

src/backend/postmaster/autovacuum.c

@@ -628,7 +628,6 @@ AutoVacLauncherMain(int argc, char *argv[])
 		struct timeval nap;
 		TimestampTz current_time = 0;
 		bool		can_launch;
-		int			rc;
 
 		/*
 		 * This loop is a bit different from the normal use of WaitLatch,
@@ -644,23 +643,16 @@ AutoVacLauncherMain(int argc, char *argv[])
 		 * Wait until naptime expires or we get some type of signal (all the
 		 * signal handlers will wake us by calling SetLatch).
 		 */
-		rc = WaitLatch(MyLatch,
-					   WL_LATCH_SET | WL_TIMEOUT | WL_POSTMASTER_DEATH,
-					   (nap.tv_sec * 1000L) + (nap.tv_usec / 1000L),
-					   WAIT_EVENT_AUTOVACUUM_MAIN);
+		(void) WaitLatch(MyLatch,
+						 WL_LATCH_SET | WL_TIMEOUT | WL_EXIT_ON_PM_DEATH,
+						 (nap.tv_sec * 1000L) + (nap.tv_usec / 1000L),
+						 WAIT_EVENT_AUTOVACUUM_MAIN);
 
 		ResetLatch(MyLatch);
 
 		/* Process sinval catchup interrupts that happened while sleeping */
 		ProcessCatchupInterrupt();
 
-		/*
-		 * Emergency bailout if postmaster has died.  This is to avoid the
-		 * necessity for manual cleanup of all postmaster children.
-		 */
-		if (rc & WL_POSTMASTER_DEATH)
-			proc_exit(1);
-
 		/* the normal shutdown case */
 		if (got_SIGTERM)
 			break;

src/backend/postmaster/bgwriter.c

@@ -335,7 +335,7 @@ BackgroundWriterMain(void)
 		 * normal operation.
 		 */
 		rc = WaitLatch(MyLatch,
-					   WL_LATCH_SET | WL_TIMEOUT | WL_POSTMASTER_DEATH,
+					   WL_LATCH_SET | WL_TIMEOUT | WL_EXIT_ON_PM_DEATH,
 					   BgWriterDelay /* ms */ , WAIT_EVENT_BGWRITER_MAIN);
 
 		/*
@@ -362,20 +362,13 @@ BackgroundWriterMain(void)
 			StrategyNotifyBgWriter(MyProc->pgprocno);
 			/* Sleep ... */
 			rc = WaitLatch(MyLatch,
-						   WL_LATCH_SET | WL_TIMEOUT | WL_POSTMASTER_DEATH,
+						   WL_LATCH_SET | WL_TIMEOUT | WL_EXIT_ON_PM_DEATH,
 						   BgWriterDelay * HIBERNATE_FACTOR,
 						   WAIT_EVENT_BGWRITER_HIBERNATE);
 			/* Reset the notification request in case we timed out */
 			StrategyNotifyBgWriter(-1);
 		}
 
-		/*
-		 * Emergency bailout if postmaster has died.  This is to avoid the
-		 * necessity for manual cleanup of all postmaster children.
-		 */
-		if (rc & WL_POSTMASTER_DEATH)
-			exit(1);
-
 		prev_hibernate = can_hibernate;
 	}
 }

src/backend/postmaster/checkpointer.c

@@ -340,7 +340,6 @@ CheckpointerMain(void)
 		pg_time_t	now;
 		int			elapsed_secs;
 		int			cur_timeout;
-		int			rc;
 
 		/* Clear any already-pending wakeups */
 		ResetLatch(MyLatch);
@@ -541,17 +540,10 @@ CheckpointerMain(void)
 			cur_timeout = Min(cur_timeout, XLogArchiveTimeout - elapsed_secs);
 		}
 
-		rc = WaitLatch(MyLatch,
-					   WL_LATCH_SET | WL_TIMEOUT | WL_POSTMASTER_DEATH,
-					   cur_timeout * 1000L /* convert to ms */ ,
-					   WAIT_EVENT_CHECKPOINTER_MAIN);
-
-		/*
-		 * Emergency bailout if postmaster has died.  This is to avoid the
-		 * necessity for manual cleanup of all postmaster children.
-		 */
-		if (rc & WL_POSTMASTER_DEATH)
-			exit(1);
+		(void) WaitLatch(MyLatch,
+						 WL_LATCH_SET | WL_TIMEOUT | WL_EXIT_ON_PM_DEATH,
+						 cur_timeout * 1000L /* convert to ms */ ,
+						 WAIT_EVENT_CHECKPOINTER_MAIN);
 	}
 }
 

src/backend/postmaster/pgarch.c

@@ -390,6 +390,8 @@ pgarch_MainLoop(void)
 							   WAIT_EVENT_ARCHIVER_MAIN);
 				if (rc & WL_TIMEOUT)
 					wakened = true;
+				if (rc & WL_POSTMASTER_DEATH)
+					time_to_stop = true;
 			}
 			else
 				wakened = true;
@@ -400,7 +402,7 @@ pgarch_MainLoop(void)
 		 * or after completing one more archiving cycle after receiving
 		 * SIGUSR2.
 		 */
-	} while (PostmasterIsAlive() && !time_to_stop);
+	} while (!time_to_stop);
 }
 
 /*

src/backend/postmaster/syslogger.c

@@ -168,6 +168,7 @@ SysLoggerMain(int argc, char *argv[])
 	char	   *currentLogFilename;
 	int			currentLogRotationAge;
 	pg_time_t	now;
+	WaitEventSet *wes;
 
 	now = MyStartTime;
 
@@ -294,13 +295,29 @@ SysLoggerMain(int argc, char *argv[])
 	 */
 	whereToSendOutput = DestNone;
 
+	/*
+	 * Set up a reusable WaitEventSet object we'll use to wait for our latch,
+	 * and (except on Windows) our socket.
+	 *
+	 * Unlike all other postmaster child processes, we'll ignore postmaster
+	 * death because we want to collect final log output from all backends and
+	 * then exit last.  We'll do that by running until we see EOF on the
+	 * syslog pipe, which implies that all other backends have exited
+	 * (including the postmaster).
+	 */
+	wes = CreateWaitEventSet(CurrentMemoryContext, 2);
+	AddWaitEventToSet(wes, WL_LATCH_SET, PGINVALID_SOCKET, MyLatch, NULL);
+#ifndef WIN32
+	AddWaitEventToSet(wes, WL_SOCKET_READABLE, syslogPipe[0], NULL, NULL);
+#endif
+
 	/* main worker loop */
 	for (;;)
 	{
 		bool		time_based_rotation = false;
 		int			size_rotation_for = 0;
 		long		cur_timeout;
-		int			cur_flags;
+		WaitEvent	event;
 
 #ifndef WIN32
 		int			rc;
@@ -436,25 +453,18 @@ SysLoggerMain(int argc, char *argv[])
 			}
 			else
 				cur_timeout = 0;
-			cur_flags = WL_TIMEOUT;
 		}
 		else
-		{
 			cur_timeout = -1L;
-			cur_flags = 0;
-		}
 
 		/*
 		 * Sleep until there's something to do
 		 */
 #ifndef WIN32
-		rc = WaitLatchOrSocket(MyLatch,
-							   WL_LATCH_SET | WL_SOCKET_READABLE | cur_flags,
-							   syslogPipe[0],
-							   cur_timeout,
-							   WAIT_EVENT_SYSLOGGER_MAIN);
+		rc = WaitEventSetWait(wes, cur_timeout, &event, 1,
+							  WAIT_EVENT_SYSLOGGER_MAIN);
 
-		if (rc & WL_SOCKET_READABLE)
+		if (rc == 1 && event.events == WL_SOCKET_READABLE)
 		{
 			int			bytesRead;
 
@@ -501,10 +511,8 @@ SysLoggerMain(int argc, char *argv[])
 		 */
 		LeaveCriticalSection(&sysloggerSection);
 
-		(void) WaitLatch(MyLatch,
-						 WL_LATCH_SET | cur_flags,
-						 cur_timeout,
-						 WAIT_EVENT_SYSLOGGER_MAIN);
+		(void) WaitEventSetWait(wes, cur_timeout, &event, 1,
+								WAIT_EVENT_SYSLOGGER_MAIN);
 
 		EnterCriticalSection(&sysloggerSection);
 #endif							/* WIN32 */

src/backend/postmaster/walwriter.c

@@ -223,7 +223,6 @@ WalWriterMain(void)
 	for (;;)
 	{
 		long		cur_timeout;
-		int			rc;
 
 		/*
 		 * Advertise whether we might hibernate in this cycle.  We do this
@@ -276,17 +275,10 @@ WalWriterMain(void)
 		else
 			cur_timeout = WalWriterDelay * HIBERNATE_FACTOR;
 
-		rc = WaitLatch(MyLatch,
-					   WL_LATCH_SET | WL_TIMEOUT | WL_POSTMASTER_DEATH,
-					   cur_timeout,
-					   WAIT_EVENT_WAL_WRITER_MAIN);
-
-		/*
-		 * Emergency bailout if postmaster has died.  This is to avoid the
-		 * necessity for manual cleanup of all postmaster children.
-		 */
-		if (rc & WL_POSTMASTER_DEATH)
-			exit(1);
+		(void) WaitLatch(MyLatch,
+						 WL_LATCH_SET | WL_TIMEOUT | WL_EXIT_ON_PM_DEATH,
+						 cur_timeout,
+						 WAIT_EVENT_WAL_WRITER_MAIN);
 	}
 }
 

src/backend/replication/basebackup.c

@@ -1686,7 +1686,7 @@ throttle(size_t increment)
 		 * the maximum time to sleep. Thus the cast to long is safe.
 		 */
 		wait_result = WaitLatch(MyLatch,
-								WL_LATCH_SET | WL_TIMEOUT | WL_POSTMASTER_DEATH,
+								WL_LATCH_SET | WL_TIMEOUT | WL_EXIT_ON_PM_DEATH,
 								(long) (sleep / 1000),
 								WAIT_EVENT_BASE_BACKUP_THROTTLE);
 

src/backend/replication/libpqwalreceiver/libpqwalreceiver.c

@@ -186,16 +186,11 @@ libpqrcv_connect(const char *conninfo, bool logical, const char *appname,
 			io_flag = WL_SOCKET_WRITEABLE;
 
 		rc = WaitLatchOrSocket(MyLatch,
-							   WL_POSTMASTER_DEATH |
-							   WL_LATCH_SET | io_flag,
+							   WL_EXIT_ON_PM_DEATH | WL_LATCH_SET | io_flag,
 							   PQsocket(conn->streamConn),
 							   0,
 							   WAIT_EVENT_LIBPQWALRECEIVER_CONNECT);
 
-		/* Emergency bailout? */
-		if (rc & WL_POSTMASTER_DEATH)
-			exit(1);
-
 		/* Interrupted? */
 		if (rc & WL_LATCH_SET)
 		{
@@ -610,16 +605,12 @@ libpqrcv_PQexec(PGconn *streamConn, const char *query)
 			 * replication connection.
 			 */
 			rc = WaitLatchOrSocket(MyLatch,
-								   WL_POSTMASTER_DEATH | WL_SOCKET_READABLE |
+								   WL_EXIT_ON_PM_DEATH | WL_SOCKET_READABLE |
 								   WL_LATCH_SET,
 								   PQsocket(streamConn),
 								   0,
 								   WAIT_EVENT_LIBPQWALRECEIVER_RECEIVE);
 
-			/* Emergency bailout? */
-			if (rc & WL_POSTMASTER_DEATH)
-				exit(1);
-
 			/* Interrupted? */
 			if (rc & WL_LATCH_SET)
 			{

src/backend/replication/logical/launcher.c

@@ -221,13 +221,9 @@ WaitForReplicationWorkerAttach(LogicalRepWorker *worker,
 		 * about the worker attach.  But we don't expect to have to wait long.
 		 */
 		rc = WaitLatch(MyLatch,
-					   WL_LATCH_SET | WL_TIMEOUT | WL_POSTMASTER_DEATH,
+					   WL_LATCH_SET | WL_TIMEOUT | WL_EXIT_ON_PM_DEATH,
 					   10L, WAIT_EVENT_BGWORKER_STARTUP);
 
-		/* emergency bailout if postmaster has died */
-		if (rc & WL_POSTMASTER_DEATH)
-			proc_exit(1);
-
 		if (rc & WL_LATCH_SET)
 		{
 			ResetLatch(MyLatch);
@@ -498,13 +494,9 @@ logicalrep_worker_stop(Oid subid, Oid relid)
 
 		/* Wait a bit --- we don't expect to have to wait long. */
 		rc = WaitLatch(MyLatch,
-					   WL_LATCH_SET | WL_TIMEOUT | WL_POSTMASTER_DEATH,
+					   WL_LATCH_SET | WL_TIMEOUT | WL_EXIT_ON_PM_DEATH,
 					   10L, WAIT_EVENT_BGWORKER_STARTUP);
 
-		/* emergency bailout if postmaster has died */
-		if (rc & WL_POSTMASTER_DEATH)
-			proc_exit(1);
-
 		if (rc & WL_LATCH_SET)
 		{
 			ResetLatch(MyLatch);
@@ -546,13 +538,9 @@ logicalrep_worker_stop(Oid subid, Oid relid)
 
 		/* Wait a bit --- we don't expect to have to wait long. */
 		rc = WaitLatch(MyLatch,
-					   WL_LATCH_SET | WL_TIMEOUT | WL_POSTMASTER_DEATH,
+					   WL_LATCH_SET | WL_TIMEOUT | WL_EXIT_ON_PM_DEATH,
 					   10L, WAIT_EVENT_BGWORKER_SHUTDOWN);
 
-		/* emergency bailout if postmaster has died */
-		if (rc & WL_POSTMASTER_DEATH)
-			proc_exit(1);
-
 		if (rc & WL_LATCH_SET)
 		{
 			ResetLatch(MyLatch);
@@ -1072,14 +1060,10 @@ ApplyLauncherMain(Datum main_arg)
 
 		/* Wait for more work. */
 		rc = WaitLatch(MyLatch,
-					   WL_LATCH_SET | WL_TIMEOUT | WL_POSTMASTER_DEATH,
+					   WL_LATCH_SET | WL_TIMEOUT | WL_EXIT_ON_PM_DEATH,
 					   wait_time,
 					   WAIT_EVENT_LOGICAL_LAUNCHER_MAIN);
 
-		/* emergency bailout if postmaster has died */
-		if (rc & WL_POSTMASTER_DEATH)
-			proc_exit(1);
-
 		if (rc & WL_LATCH_SET)
 		{
 			ResetLatch(MyLatch);

src/backend/replication/logical/tablesync.c

@@ -159,7 +159,6 @@ finish_sync_worker(void)
 static bool
 wait_for_relation_state_change(Oid relid, char expected_state)
 {
-	int			rc;
 	char		state;
 
 	for (;;)
@@ -192,13 +191,9 @@ wait_for_relation_state_change(Oid relid, char expected_state)
 		if (!worker)
 			return false;
 
-		rc = WaitLatch(MyLatch,
-					   WL_LATCH_SET | WL_TIMEOUT | WL_POSTMASTER_DEATH,
-					   1000L, WAIT_EVENT_LOGICAL_SYNC_STATE_CHANGE);
-
-		/* emergency bailout if postmaster has died */
-		if (rc & WL_POSTMASTER_DEATH)
-			proc_exit(1);
+		(void) WaitLatch(MyLatch,
+						 WL_LATCH_SET | WL_TIMEOUT | WL_EXIT_ON_PM_DEATH,
+						 1000L, WAIT_EVENT_LOGICAL_SYNC_STATE_CHANGE);
 
 		ResetLatch(MyLatch);
 	}
@@ -250,13 +245,9 @@ wait_for_worker_state_change(char expected_state)
 		 * but use a timeout in case it dies without sending one.
 		 */
 		rc = WaitLatch(MyLatch,
-					   WL_LATCH_SET | WL_TIMEOUT | WL_POSTMASTER_DEATH,
+					   WL_LATCH_SET | WL_TIMEOUT | WL_EXIT_ON_PM_DEATH,
 					   1000L, WAIT_EVENT_LOGICAL_SYNC_STATE_CHANGE);
 
-		/* emergency bailout if postmaster has died */
-		if (rc & WL_POSTMASTER_DEATH)
-			proc_exit(1);
-
 		if (rc & WL_LATCH_SET)
 			ResetLatch(MyLatch);
 	}
@@ -593,7 +584,6 @@ copy_read_data(void *outbuf, int minread, int maxread)
 	while (maxread > 0 && bytesread < minread)
 	{
 		pgsocket	fd = PGINVALID_SOCKET;
-		int			rc;
 		int			len;
 		char	   *buf = NULL;
 
@@ -632,14 +622,10 @@ copy_read_data(void *outbuf, int minread, int maxread)
 		/*
 		 * Wait for more data or latch.
 		 */
-		rc = WaitLatchOrSocket(MyLatch,
-							   WL_SOCKET_READABLE | WL_LATCH_SET |
-							   WL_TIMEOUT | WL_POSTMASTER_DEATH,
-							   fd, 1000L, WAIT_EVENT_LOGICAL_SYNC_DATA);
-
-		/* Emergency bailout if postmaster has died */
-		if (rc & WL_POSTMASTER_DEATH)
-			proc_exit(1);
+		(void) WaitLatchOrSocket(MyLatch,
+								 WL_SOCKET_READABLE | WL_LATCH_SET |
+								 WL_TIMEOUT | WL_EXIT_ON_PM_DEATH,
+								 fd, 1000L, WAIT_EVENT_LOGICAL_SYNC_DATA);
 
 		ResetLatch(MyLatch);
 	}

src/backend/replication/logical/worker.c

@@ -1264,14 +1264,10 @@ LogicalRepApplyLoop(XLogRecPtr last_received)
 
 		rc = WaitLatchOrSocket(MyLatch,
 							   WL_SOCKET_READABLE | WL_LATCH_SET |
-							   WL_TIMEOUT | WL_POSTMASTER_DEATH,
+							   WL_TIMEOUT | WL_EXIT_ON_PM_DEATH,
 							   fd, wait_time,
 							   WAIT_EVENT_LOGICAL_APPLY_MAIN);
 
-		/* Emergency bailout if postmaster has died */
-		if (rc & WL_POSTMASTER_DEATH)
-			proc_exit(1);
-
 		if (rc & WL_LATCH_SET)
 		{
 			ResetLatch(MyLatch);

src/backend/replication/syncrep.c

@@ -214,6 +214,8 @@ SyncRepWaitForLSN(XLogRecPtr lsn, bool commit)
 	 */
 	for (;;)
 	{
+		int			rc;
+
 		/* Must reset the latch before testing state. */
 		ResetLatch(MyLatch);
 
@@ -266,25 +268,25 @@ SyncRepWaitForLSN(XLogRecPtr lsn, bool commit)
 			break;
 		}
 
+		/*
+		 * Wait on latch.  Any condition that should wake us up will set the
+		 * latch, so no need for timeout.
+		 */
+		rc = WaitLatch(MyLatch, WL_LATCH_SET | WL_POSTMASTER_DEATH, -1,
+					   WAIT_EVENT_SYNC_REP);
+
 		/*
 		 * If the postmaster dies, we'll probably never get an
 		 * acknowledgment, because all the wal sender processes will exit. So
 		 * just bail out.
 		 */
-		if (!PostmasterIsAlive())
+		if (rc & WL_POSTMASTER_DEATH)
 		{
 			ProcDiePending = true;
 			whereToSendOutput = DestNone;
 			SyncRepCancelWait();
 			break;
 		}
-
-		/*
-		 * Wait on latch.  Any condition that should wake us up will set the
-		 * latch, so no need for timeout.
-		 */
-		WaitLatch(MyLatch, WL_LATCH_SET | WL_POSTMASTER_DEATH, -1,
-				  WAIT_EVENT_SYNC_REP);
 	}
 
 	/*

src/backend/replication/walreceiver.c

@@ -503,7 +503,7 @@ WalReceiverMain(void)
 				 */
 				Assert(wait_fd != PGINVALID_SOCKET);
 				rc = WaitLatchOrSocket(walrcv->latch,
-									   WL_POSTMASTER_DEATH | WL_SOCKET_READABLE |
+									   WL_EXIT_ON_PM_DEATH | WL_SOCKET_READABLE |
 									   WL_TIMEOUT | WL_LATCH_SET,
 									   wait_fd,
 									   NAPTIME_PER_CYCLE,
@@ -524,15 +524,6 @@ WalReceiverMain(void)
 						XLogWalRcvSendReply(true, false);
 					}
 				}
-				if (rc & WL_POSTMASTER_DEATH)
-				{
-					/*
-					 * Emergency bailout if postmaster has died.  This is to
-					 * avoid the necessity for manual cleanup of all
-					 * postmaster children.
-					 */
-					exit(1);
-				}
 				if (rc & WL_TIMEOUT)
 				{
 					/*
@@ -673,13 +664,6 @@ WalRcvWaitForStartPosition(XLogRecPtr *startpoint, TimeLineID *startpointTLI)
 	{
 		ResetLatch(walrcv->latch);
 
-		/*
-		 * Emergency bailout if postmaster has died.  This is to avoid the
-		 * necessity for manual cleanup of all postmaster children.
-		 */
-		if (!PostmasterIsAlive())
-			exit(1);
-
 		ProcessWalRcvInterrupts();
 
 		SpinLockAcquire(&walrcv->mutex);
@@ -706,8 +690,8 @@ WalRcvWaitForStartPosition(XLogRecPtr *startpoint, TimeLineID *startpointTLI)
 		}
 		SpinLockRelease(&walrcv->mutex);
 
-		WaitLatch(walrcv->latch, WL_LATCH_SET | WL_POSTMASTER_DEATH, 0,
-				  WAIT_EVENT_WAL_RECEIVER_WAIT_START);
+		(void) WaitLatch(walrcv->latch, WL_LATCH_SET | WL_EXIT_ON_PM_DEATH, 0,
+						 WAIT_EVENT_WAL_RECEIVER_WAIT_START);
 	}
 
 	if (update_process_title)

src/backend/replication/walsender.c

@@ -1218,20 +1218,13 @@ WalSndWriteData(LogicalDecodingContext *ctx, XLogRecPtr lsn, TransactionId xid,
 
 		sleeptime = WalSndComputeSleeptime(GetCurrentTimestamp());
 
-		wakeEvents = WL_LATCH_SET | WL_POSTMASTER_DEATH |
+		wakeEvents = WL_LATCH_SET | WL_EXIT_ON_PM_DEATH |
 			WL_SOCKET_WRITEABLE | WL_SOCKET_READABLE | WL_TIMEOUT;
 
 		/* Sleep until something happens or we time out */
-		WaitLatchOrSocket(MyLatch, wakeEvents,
-						  MyProcPort->sock, sleeptime,
-						  WAIT_EVENT_WAL_SENDER_WRITE_DATA);
-
-		/*
-		 * Emergency bailout if postmaster has died.  This is to avoid the
-		 * necessity for manual cleanup of all postmaster children.
-		 */
-		if (!PostmasterIsAlive())
-			exit(1);
+		(void) WaitLatchOrSocket(MyLatch, wakeEvents,
+								 MyProcPort->sock, sleeptime,
+								 WAIT_EVENT_WAL_SENDER_WRITE_DATA);
 
 		/* Clear any already-pending wakeups */
 		ResetLatch(MyLatch);
@@ -1312,13 +1305,6 @@ WalSndWaitForWal(XLogRecPtr loc)
 	{
 		long		sleeptime;
 
-		/*
-		 * Emergency bailout if postmaster has died.  This is to avoid the
-		 * necessity for manual cleanup of all postmaster children.
-		 */
-		if (!PostmasterIsAlive())
-			exit(1);
-
 		/* Clear any already-pending wakeups */
 		ResetLatch(MyLatch);
 
@@ -1410,15 +1396,15 @@ WalSndWaitForWal(XLogRecPtr loc)
 		 */
 		sleeptime = WalSndComputeSleeptime(GetCurrentTimestamp());
 
-		wakeEvents = WL_LATCH_SET | WL_POSTMASTER_DEATH |
-			WL_SOCKET_READABLE | WL_TIMEOUT;
+		wakeEvents = WL_LATCH_SET | WL_EXIT_ON_PM_DEATH |
+					 WL_SOCKET_READABLE | WL_TIMEOUT;
 
 		if (pq_is_send_pending())
 			wakeEvents |= WL_SOCKET_WRITEABLE;
 
-		WaitLatchOrSocket(MyLatch, wakeEvents,
-						  MyProcPort->sock, sleeptime,
-						  WAIT_EVENT_WAL_SENDER_WAIT_WAL);
+		(void) WaitLatchOrSocket(MyLatch, wakeEvents,
+								 MyProcPort->sock, sleeptime,
+								 WAIT_EVENT_WAL_SENDER_WAIT_WAL);
 	}
 
 	/* reactivate latch so WalSndLoop knows to continue */
@@ -2126,13 +2112,6 @@ WalSndLoop(WalSndSendDataCallback send_data)
 	 */
 	for (;;)
 	{
-		/*
-		 * Emergency bailout if postmaster has died.  This is to avoid the
-		 * necessity for manual cleanup of all postmaster children.
-		 */
-		if (!PostmasterIsAlive())
-			exit(1);
-
 		/* Clear any already-pending wakeups */
 		ResetLatch(MyLatch);
 
@@ -2222,8 +2201,8 @@ WalSndLoop(WalSndSendDataCallback send_data)
 			long		sleeptime;
 			int			wakeEvents;
 
-			wakeEvents = WL_LATCH_SET | WL_POSTMASTER_DEATH | WL_TIMEOUT |
-				WL_SOCKET_READABLE;
+			wakeEvents = WL_LATCH_SET | WL_EXIT_ON_PM_DEATH | WL_TIMEOUT |
+						 WL_SOCKET_READABLE;
 
 			/*
 			 * Use fresh timestamp, not last_processed, to reduce the chance
@@ -2235,9 +2214,9 @@ WalSndLoop(WalSndSendDataCallback send_data)
 				wakeEvents |= WL_SOCKET_WRITEABLE;
 
 			/* Sleep until something happens or we time out */
-			WaitLatchOrSocket(MyLatch, wakeEvents,
-							  MyProcPort->sock, sleeptime,
-							  WAIT_EVENT_WAL_SENDER_MAIN);
+			(void) WaitLatchOrSocket(MyLatch, wakeEvents,
+									 MyProcPort->sock, sleeptime,
+									 WAIT_EVENT_WAL_SENDER_MAIN);
 		}
 	}
 	return;

src/backend/storage/ipc/latch.c

@@ -48,6 +48,7 @@
 #include "port/atomics.h"
 #include "portability/instr_time.h"
 #include "postmaster/postmaster.h"
+#include "storage/ipc.h"
 #include "storage/latch.h"
 #include "storage/pmsignal.h"
 #include "storage/shmem.h"
@@ -92,6 +93,13 @@ struct WaitEventSet
 	Latch	   *latch;
 	int			latch_pos;
 
+	/*
+	 * WL_EXIT_ON_PM_DEATH is converted to WL_POSTMASTER_DEATH, but this flag
+	 * is set so that we'll exit immediately if postmaster death is detected,
+	 * instead of returning.
+	 */
+	bool		exit_on_postmaster_death;
+
 #if defined(WAIT_USE_EPOLL)
 	int			epoll_fd;
 	/* epoll_wait returns events in a user provided arrays, allocate once */
@@ -348,6 +356,11 @@ WaitLatch(volatile Latch *latch, int wakeEvents, long timeout,
  * to be reported as readable/writable/connected, so that the caller can deal
  * with the condition.
  *
+ * wakeEvents must include either WL_EXIT_ON_PM_DEATH for automatic exit
+ * if the postmaster dies or WL_POSTMASTER_DEATH for a flag set in the
+ * return value if the postmaster dies.  The latter is useful for rare cases
+ * where some behavior other than immediate exit is needed.
+ *
  * NB: These days this is just a wrapper around the WaitEventSet API. When
  * using a latch very frequently, consider creating a longer living
  * WaitEventSet instead; that's more efficient.
@@ -370,10 +383,19 @@ WaitLatchOrSocket(volatile Latch *latch, int wakeEvents, pgsocket sock,
 		AddWaitEventToSet(set, WL_LATCH_SET, PGINVALID_SOCKET,
 						  (Latch *) latch, NULL);
 
-	if (wakeEvents & WL_POSTMASTER_DEATH && IsUnderPostmaster)
+	/* Postmaster-managed callers must handle postmaster death somehow. */
+	Assert(!IsUnderPostmaster ||
+		   (wakeEvents & WL_EXIT_ON_PM_DEATH) ||
+		   (wakeEvents & WL_POSTMASTER_DEATH));
+
+	if ((wakeEvents & WL_POSTMASTER_DEATH) && IsUnderPostmaster)
 		AddWaitEventToSet(set, WL_POSTMASTER_DEATH, PGINVALID_SOCKET,
 						  NULL, NULL);
 
+	if ((wakeEvents & WL_EXIT_ON_PM_DEATH) && IsUnderPostmaster)
+		AddWaitEventToSet(set, WL_EXIT_ON_PM_DEATH, PGINVALID_SOCKET,
+						  NULL, NULL);
+
 	if (wakeEvents & WL_SOCKET_MASK)
 	{
 		int			ev;
@@ -562,6 +584,7 @@ CreateWaitEventSet(MemoryContext context, int nevents)
 
 	set->latch = NULL;
 	set->nevents_space = nevents;
+	set->exit_on_postmaster_death = false;
 
 #if defined(WAIT_USE_EPOLL)
 #ifdef EPOLL_CLOEXEC
@@ -646,6 +669,7 @@ FreeWaitEventSet(WaitEventSet *set)
  * - WL_SOCKET_CONNECTED: Wait for socket connection to be established,
  *	 can be combined with other WL_SOCKET_* events (on non-Windows
  *	 platforms, this is the same as WL_SOCKET_WRITEABLE)
+ * - WL_EXIT_ON_PM_DEATH: Exit immediately if the postmaster dies
  *
  * Returns the offset in WaitEventSet->events (starting from 0), which can be
  * used to modify previously added wait events using ModifyWaitEvent().
@@ -671,6 +695,12 @@ AddWaitEventToSet(WaitEventSet *set, uint32 events, pgsocket fd, Latch *latch,
 	/* not enough space */
 	Assert(set->nevents < set->nevents_space);
 
+	if (events == WL_EXIT_ON_PM_DEATH)
+	{
+		events = WL_POSTMASTER_DEATH;
+		set->exit_on_postmaster_death = true;
+	}
+
 	if (latch)
 	{
 		if (latch->owner_pid != MyProcPid)
@@ -1114,6 +1144,8 @@ WaitEventSetWaitBlock(WaitEventSet *set, int cur_timeout,
 			 */
 			if (!PostmasterIsAliveInternal())
 			{
+				if (set->exit_on_postmaster_death)
+					proc_exit(1);
 				occurred_events->fd = PGINVALID_SOCKET;
 				occurred_events->events = WL_POSTMASTER_DEATH;
 				occurred_events++;
@@ -1232,6 +1264,8 @@ WaitEventSetWaitBlock(WaitEventSet *set, int cur_timeout,
 			 */
 			if (!PostmasterIsAliveInternal())
 			{
+				if (set->exit_on_postmaster_death)
+					proc_exit(1);
 				occurred_events->fd = PGINVALID_SOCKET;
 				occurred_events->events = WL_POSTMASTER_DEATH;
 				occurred_events++;
@@ -1392,6 +1426,8 @@ WaitEventSetWaitBlock(WaitEventSet *set, int cur_timeout,
 		 */
 		if (!PostmasterIsAliveInternal())
 		{
+			if (set->exit_on_postmaster_death)
+				proc_exit(1);
 			occurred_events->fd = PGINVALID_SOCKET;
 			occurred_events->events = WL_POSTMASTER_DEATH;
 			occurred_events++;

src/backend/storage/ipc/shm_mq.c

@@ -949,7 +949,8 @@ shm_mq_send_bytes(shm_mq_handle *mqh, Size nbytes, const void *data,
 			 * at top of loop, because setting an already-set latch is much
 			 * cheaper than setting one that has been reset.
 			 */
-			WaitLatch(MyLatch, WL_LATCH_SET, 0, WAIT_EVENT_MQ_SEND);
+			(void) WaitLatch(MyLatch, WL_LATCH_SET | WL_EXIT_ON_PM_DEATH, 0,
+							 WAIT_EVENT_MQ_SEND);
 
 			/* Reset the latch so we don't spin. */
 			ResetLatch(MyLatch);
@@ -1093,7 +1094,8 @@ shm_mq_receive_bytes(shm_mq_handle *mqh, Size bytes_needed, bool nowait,
 		 * loop, because setting an already-set latch is much cheaper than
 		 * setting one that has been reset.
 		 */
-		WaitLatch(MyLatch, WL_LATCH_SET, 0, WAIT_EVENT_MQ_RECEIVE);
+		(void) WaitLatch(MyLatch, WL_LATCH_SET | WL_EXIT_ON_PM_DEATH, 0,
+						 WAIT_EVENT_MQ_RECEIVE);
 
 		/* Reset the latch so we don't spin. */
 		ResetLatch(MyLatch);
@@ -1181,7 +1183,8 @@ shm_mq_wait_internal(shm_mq *mq, PGPROC **ptr, BackgroundWorkerHandle *handle)
 		}
 
 		/* Wait to be signalled. */
-		WaitLatch(MyLatch, WL_LATCH_SET, 0, WAIT_EVENT_MQ_INTERNAL);
+		(void) WaitLatch(MyLatch, WL_LATCH_SET | WL_EXIT_ON_PM_DEATH, 0,
+						 WAIT_EVENT_MQ_INTERNAL);
 
 		/* Reset the latch so we don't spin. */
 		ResetLatch(MyLatch);

src/backend/storage/lmgr/condition_variable.c

@@ -72,7 +72,7 @@ ConditionVariablePrepareToSleep(ConditionVariable *cv)
 		new_event_set = CreateWaitEventSet(TopMemoryContext, 2);
 		AddWaitEventToSet(new_event_set, WL_LATCH_SET, PGINVALID_SOCKET,
 						  MyLatch, NULL);
-		AddWaitEventToSet(new_event_set, WL_POSTMASTER_DEATH, PGINVALID_SOCKET,
+		AddWaitEventToSet(new_event_set, WL_EXIT_ON_PM_DEATH, PGINVALID_SOCKET,
 						  NULL, NULL);
 		/* Don't set cv_wait_event_set until we have a correct WES. */
 		cv_wait_event_set = new_event_set;
@@ -154,16 +154,8 @@ ConditionVariableSleep(ConditionVariable *cv, uint32 wait_event_info)
 		 * Wait for latch to be set.  (If we're awakened for some other
 		 * reason, the code below will cope anyway.)
 		 */
-		WaitEventSetWait(cv_wait_event_set, -1, &event, 1, wait_event_info);
-
-		if (event.events & WL_POSTMASTER_DEATH)
-		{
-			/*
-			 * Emergency bailout if postmaster has died.  This is to avoid the
-			 * necessity for manual cleanup of all postmaster children.
-			 */
-			exit(1);
-		}
+		(void) WaitEventSetWait(cv_wait_event_set, -1, &event, 1,
+								wait_event_info);
 
 		/* Reset latch before examining the state of the wait list. */
 		ResetLatch(MyLatch);

src/backend/storage/lmgr/proc.c

@@ -1270,8 +1270,8 @@ ProcSleep(LOCALLOCK *locallock, LockMethod lockMethodTable)
 		}
 		else
 		{
-			WaitLatch(MyLatch, WL_LATCH_SET, 0,
-					  PG_WAIT_LOCK | locallock->tag.lock.locktag_type);
+			(void) WaitLatch(MyLatch, WL_LATCH_SET | WL_EXIT_ON_PM_DEATH, 0,
+							 PG_WAIT_LOCK | locallock->tag.lock.locktag_type);
 			ResetLatch(MyLatch);
 			/* check for deadlocks first, as that's probably log-worthy */
 			if (got_deadlock_timeout)
@@ -1783,7 +1783,8 @@ CheckDeadLockAlert(void)
 void
 ProcWaitForSignal(uint32 wait_event_info)
 {
-	WaitLatch(MyLatch, WL_LATCH_SET, 0, wait_event_info);
+	(void) WaitLatch(MyLatch, WL_LATCH_SET | WL_EXIT_ON_PM_DEATH, 0,
+					 wait_event_info);
 	ResetLatch(MyLatch);
 	CHECK_FOR_INTERRUPTS();
 }

src/backend/utils/adt/misc.c

@@ -379,7 +379,7 @@ pg_sleep(PG_FUNCTION_ARGS)
 			break;
 
 		(void) WaitLatch(MyLatch,
-						 WL_LATCH_SET | WL_TIMEOUT,
+						 WL_LATCH_SET | WL_TIMEOUT | WL_EXIT_ON_PM_DEATH,
 						 delay_ms,
 						 WAIT_EVENT_PG_SLEEP);
 		ResetLatch(MyLatch);