Loop calling CallNamedPipe() several times in case it fails,

since it can be transient failures, causing kill() to not
properly send signals.

Original patch from Steve Marshall, modified by me.

1 files changed, 25 insertions, 14 deletions

diff --git a/src/port/kill.c b/src/port/kill.c
index 0b6ee0d2ed..5a2faaa751 100644
--- a/src/port/kill.c
+++ b/src/port/kill.c
@@ -25,6 +25,7 @@ pgkill(int pid, int sig)
 	BYTE		sigData = sig;
 	BYTE		sigRet = 0;
 	DWORD		bytes;
+	int			pipe_tries;
 
 	/* we allow signal 0 here, but it will be ignored in pg_queue_signal */
 	if (sig >= PG_SIGNAL_COUNT || sig < 0)
@@ -39,23 +40,33 @@ pgkill(int pid, int sig)
 		return -1;
 	}
 	snprintf(pipename, sizeof(pipename), "\\\\.\\pipe\\pgsignal_%u", pid);
-	if (!CallNamedPipe(pipename, &sigData, 1, &sigRet, 1, &bytes, 1000))
-	{
-		if (GetLastError() == ERROR_FILE_NOT_FOUND)
-			errno = ESRCH;
-		else if (GetLastError() == ERROR_ACCESS_DENIED)
-			errno = EPERM;
-		else
-			errno = EINVAL;
-		return -1;
-	}
-	if (bytes != 1 || sigRet != sig)
+
+	/*
+	 * Writing data to the named pipe can fail for transient reasons.
+	 * Therefore, it is useful to retry if it fails.  The maximum number of
+	 * calls to make was empirically determined from a 90-hour notification
+	 * stress test.
+	 */
+	for (pipe_tries = 0; pipe_tries < 3; pipe_tries++)
 	{
-		errno = ESRCH;
-		return -1;
+		if (CallNamedPipe(pipename, &sigData, 1, &sigRet, 1, &bytes, 1000))
+		{
+			if (bytes != 1 || sigRet != sig)
+			{
+				errno = ESRCH;
+				return -1;
+			}
+			return 0;
+		}
 	}
 
-	return 0;
+	if (GetLastError() == ERROR_FILE_NOT_FOUND)
+		errno = ESRCH;
+	else if (GetLastError() == ERROR_ACCESS_DENIED)
+		errno = EPERM;
+	else
+		errno = EINVAL;
+	return -1;
 }
 
 #endif