Revert "Get rid of the dedicated latch for signaling the startup process".

Revert ac22929a26, as well as the followup fix 113d3591b8. Because it broke
the assumption that the startup process waiting for the recovery conflict
on buffer pin should be waken up only by buffer unpin or the timeout enabled
in ResolveRecoveryConflictWithBufferPin(). It caused, for example,
SIGHUP signal handler or walreceiver process to wake that startup process
up unnecessarily frequently.

Additionally, add the comments about why that dedicated latch that
the reverted patch tried to get rid of should not be removed.

Thanks to Kyotaro Horiguchi for the discussion.

Author: Fujii Masao
Discussion: https://postgr.es/m/d8c0c608-021b-3c73-fffd-3240829ee986@oss.nttdata.com

1 files changed, 25 insertions, 15 deletions

diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 8dd225c2e14..b1e5d2dbffd 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -681,8 +681,18 @@ typedef struct XLogCtlData
 	 * recoveryWakeupLatch is used to wake up the startup process to continue
 	 * WAL replay, if it is waiting for WAL to arrive or failover trigger file
 	 * to appear.
+	 *
+	 * Note that the startup process also uses another latch, its procLatch,
+	 * to wait for recovery conflict. If we get rid of recoveryWakeupLatch for
+	 * signaling the startup process in favor of using its procLatch, which
+	 * comports better with possible generic signal handlers using that latch.
+	 * But we should not do that because the startup process doesn't assume
+	 * that it's waken up by walreceiver process or SIGHUP signal handler
+	 * while it's waiting for recovery conflict. The separate latches,
+	 * recoveryWakeupLatch and procLatch, should be used for inter-process
+	 * communication for WAL replay and recovery conflict, respectively.
 	 */
-	Latch		*recoveryWakeupLatch;
+	Latch		recoveryWakeupLatch;
 
 	/*
 	 * During recovery, we keep a copy of the latest checkpoint record here.
@@ -5186,6 +5196,7 @@ XLOGShmemInit(void)
 	SpinLockInit(&XLogCtl->Insert.insertpos_lck);
 	SpinLockInit(&XLogCtl->info_lck);
 	SpinLockInit(&XLogCtl->ulsn_lck);
+	InitSharedLatch(&XLogCtl->recoveryWakeupLatch);
 }
 
 /*
@@ -6121,7 +6132,7 @@ recoveryApplyDelay(XLogReaderState *record)
 
 	while (true)
 	{
-		ResetLatch(MyLatch);
+		ResetLatch(&XLogCtl->recoveryWakeupLatch);
 
 		/* might change the trigger file's location */
 		HandleStartupProcInterrupts();
@@ -6140,7 +6151,7 @@ recoveryApplyDelay(XLogReaderState *record)
 
 		elog(DEBUG2, "recovery apply delay %ld milliseconds", msecs);
 
-		(void) WaitLatch(MyLatch,
+		(void) WaitLatch(&XLogCtl->recoveryWakeupLatch,
 						 WL_LATCH_SET | WL_TIMEOUT | WL_EXIT_ON_PM_DEATH,
 						 msecs,
 						 WAIT_EVENT_RECOVERY_APPLY_DELAY);
@@ -6469,11 +6480,11 @@ StartupXLOG(void)
 	}
 
 	/*
-	 * Advertise our latch that other processes can use to wake us up
-	 * if we're going to sleep during recovery.
+	 * Take ownership of the wakeup latch if we're going to sleep during
+	 * recovery.
 	 */
 	if (ArchiveRecoveryRequested)
-		XLogCtl->recoveryWakeupLatch = &MyProc->procLatch;
+		OwnLatch(&XLogCtl->recoveryWakeupLatch);
 
 	/* Set up XLOG reader facility */
 	MemSet(&private, 0, sizeof(XLogPageReadPrivate));
@@ -7484,11 +7495,11 @@ StartupXLOG(void)
 		ResetUnloggedRelations(UNLOGGED_RELATION_INIT);
 
 	/*
-	 * We don't need the latch anymore. It's not strictly necessary to reset
-	 * it to NULL, but let's do it for the sake of tidiness.
+	 * We don't need the latch anymore. It's not strictly necessary to disown
+	 * it, but let's do it for the sake of tidiness.
 	 */
 	if (ArchiveRecoveryRequested)
-		XLogCtl->recoveryWakeupLatch = NULL;
+		DisownLatch(&XLogCtl->recoveryWakeupLatch);
 
 	/*
 	 * We are now done reading the xlog from stream. Turn off streaming
@@ -12300,12 +12311,12 @@ WaitForWALToBecomeAvailable(XLogRecPtr RecPtr, bool randAccess,
 						wait_time = wal_retrieve_retry_interval -
 							TimestampDifferenceMilliseconds(last_fail_time, now);
 
-						(void) WaitLatch(MyLatch,
+						(void) WaitLatch(&XLogCtl->recoveryWakeupLatch,
 										 WL_LATCH_SET | WL_TIMEOUT |
 										 WL_EXIT_ON_PM_DEATH,
 										 wait_time,
 										 WAIT_EVENT_RECOVERY_RETRIEVE_RETRY_INTERVAL);
-						ResetLatch(MyLatch);
+						ResetLatch(&XLogCtl->recoveryWakeupLatch);
 						now = GetCurrentTimestamp();
 
 						/* Handle interrupt signals of startup process */
@@ -12559,11 +12570,11 @@ WaitForWALToBecomeAvailable(XLogRecPtr RecPtr, bool randAccess,
 					 * to react to a trigger file promptly and to check if the
 					 * WAL receiver is still active.
 					 */
-					(void) WaitLatch(MyLatch,
+					(void) WaitLatch(&XLogCtl->recoveryWakeupLatch,
 									 WL_LATCH_SET | WL_TIMEOUT |
 									 WL_EXIT_ON_PM_DEATH,
 									 5000L, WAIT_EVENT_RECOVERY_WAL_STREAM);
-					ResetLatch(MyLatch);
+					ResetLatch(&XLogCtl->recoveryWakeupLatch);
 					break;
 				}
 
@@ -12735,8 +12746,7 @@ CheckPromoteSignal(void)
 void
 WakeupRecovery(void)
 {
-	if (XLogCtl->recoveryWakeupLatch)
-		SetLatch(XLogCtl->recoveryWakeupLatch);
+	SetLatch(&XLogCtl->recoveryWakeupLatch);
 }
 
 /*