Use signals for postmaster death on Linux.

Linux provides a way to ask for a signal when your parent process dies.
Use that to make PostmasterIsAlive() very cheap.

Based on a suggestion from Andres Freund.

Author: Thomas Munro, Heikki Linnakangas
Reviewed-By: Michael Paquier
Discussion: https://postgr.es/m/7261eb39-0369-f2f4-1bb5-62f3b6083b5e%40iki.fi
Discussion: https://postgr.es/m/20180411002643.6buofht4ranhei7k%40alap3.anarazel.de

3 files changed, 119 insertions, 16 deletions

diff --git a/src/backend/storage/ipc/latch.c b/src/backend/storage/ipc/latch.c
index e6706f7fb80..f6dda9cc9ac 100644
--- a/src/backend/storage/ipc/latch.c
+++ b/src/backend/storage/ipc/latch.c
@@ -1112,7 +1112,7 @@ WaitEventSetWaitBlock(WaitEventSet *set, int cur_timeout,
 			 * WL_POSTMASTER_DEATH event would be painful. Re-checking doesn't
 			 * cost much.
 			 */
-			if (!PostmasterIsAlive())
+			if (!PostmasterIsAliveInternal())
 			{
 				occurred_events->fd = PGINVALID_SOCKET;
 				occurred_events->events = WL_POSTMASTER_DEATH;
@@ -1230,7 +1230,7 @@ WaitEventSetWaitBlock(WaitEventSet *set, int cur_timeout,
 			 * WL_POSTMASTER_DEATH event would be painful. Re-checking doesn't
 			 * cost much.
 			 */
-			if (!PostmasterIsAlive())
+			if (!PostmasterIsAliveInternal())
 			{
 				occurred_events->fd = PGINVALID_SOCKET;
 				occurred_events->events = WL_POSTMASTER_DEATH;
@@ -1390,7 +1390,7 @@ WaitEventSetWaitBlock(WaitEventSet *set, int cur_timeout,
 		 * even though there is no known reason to think that the event could
 		 * be falsely set on Windows.
 		 */
-		if (!PostmasterIsAlive())
+		if (!PostmasterIsAliveInternal())
 		{
 			occurred_events->fd = PGINVALID_SOCKET;
 			occurred_events->events = WL_POSTMASTER_DEATH;
diff --git a/src/backend/storage/ipc/pmsignal.c b/src/backend/storage/ipc/pmsignal.c
index be61858fc67..ebf027306f7 100644
--- a/src/backend/storage/ipc/pmsignal.c
+++ b/src/backend/storage/ipc/pmsignal.c
@@ -17,6 +17,10 @@
 #include <signal.h>
 #include <unistd.h>
 
+#ifdef HAVE_SYS_PRCTL_H
+#include <sys/prctl.h>
+#endif
+
 #include "miscadmin.h"
 #include "postmaster/postmaster.h"
 #include "replication/walsender.h"
@@ -71,6 +75,35 @@ struct PMSignalData
 
 NON_EXEC_STATIC volatile PMSignalData *PMSignalState = NULL;
 
+/*
+ * Signal handler to be notified if postmaster dies.
+ */
+#ifdef USE_POSTMASTER_DEATH_SIGNAL
+volatile sig_atomic_t postmaster_possibly_dead = false;
+
+static void
+postmaster_death_handler(int signo)
+{
+	postmaster_possibly_dead = true;
+}
+
+/*
+ * The available signals depend on the OS.  SIGUSR1 and SIGUSR2 are already
+ * used for other things, so choose another one.
+ *
+ * Currently, we assume that we can always find a signal to use.  That
+ * seems like a reasonable assumption for all platforms that are modern
+ * enough to have a parent-death signaling mechanism.
+ */
+#if defined(SIGINFO)
+#define POSTMASTER_DEATH_SIGNAL SIGINFO
+#elif defined(SIGPWR)
+#define POSTMASTER_DEATH_SIGNAL SIGPWR
+#else
+#error "cannot find a signal to use for postmaster death"
+#endif
+
+#endif							/* USE_POSTMASTER_DEATH_SIGNAL */
 
 /*
  * PMSignalShmemSize
@@ -266,28 +299,94 @@ MarkPostmasterChildInactive(void)
 
 
 /*
- * PostmasterIsAlive - check whether postmaster process is still alive
+ * PostmasterIsAliveInternal - check whether postmaster process is still alive
+ *
+ * This is the slow path of PostmasterIsAlive(), where the caller has already
+ * checked 'postmaster_possibly_dead'.  (On platforms that don't support
+ * a signal for parent death, PostmasterIsAlive() is just an alias for this.)
  */
 bool
-PostmasterIsAlive(void)
+PostmasterIsAliveInternal(void)
 {
-#ifndef WIN32
-	char		c;
-	ssize_t		rc;
+#ifdef USE_POSTMASTER_DEATH_SIGNAL
+	/*
+	 * Reset the flag before checking, so that we don't miss a signal if
+	 * postmaster dies right after the check.  If postmaster was indeed dead,
+	 * we'll re-arm it before returning to caller.
+	 */
+	postmaster_possibly_dead = false;
+#endif
 
-	rc = read(postmaster_alive_fds[POSTMASTER_FD_WATCH], &c, 1);
-	if (rc < 0)
+#ifndef WIN32
 	{
-		if (errno == EAGAIN || errno == EWOULDBLOCK)
+		char		c;
+		ssize_t		rc;
+
+		rc = read(postmaster_alive_fds[POSTMASTER_FD_WATCH], &c, 1);
+
+		/*
+		 * In the usual case, the postmaster is still alive, and there is no
+		 * data in the pipe.
+		 */
+		if (rc < 0 && (errno == EAGAIN || errno == EWOULDBLOCK))
 			return true;
 		else
-			elog(FATAL, "read on postmaster death monitoring pipe failed: %m");
+		{
+			/*
+			 * Postmaster is dead, or something went wrong with the read()
+			 * call.
+			 */
+
+#ifdef USE_POSTMASTER_DEATH_SIGNAL
+			postmaster_possibly_dead = true;
+#endif
+
+			if (rc < 0)
+				elog(FATAL, "read on postmaster death monitoring pipe failed: %m");
+			else if (rc > 0)
+				elog(FATAL, "unexpected data in postmaster death monitoring pipe");
+
+			return false;
+		}
 	}
-	else if (rc > 0)
-		elog(FATAL, "unexpected data in postmaster death monitoring pipe");
 
-	return false;
 #else							/* WIN32 */
-	return (WaitForSingleObject(PostmasterHandle, 0) == WAIT_TIMEOUT);
+	if (WaitForSingleObject(PostmasterHandle, 0) == WAIT_TIMEOUT)
+		return true;
+	else
+	{
+#ifdef USE_POSTMASTER_DEATH_SIGNAL
+		postmaster_possibly_dead = true;
+#endif
+		return false;
+	}
 #endif							/* WIN32 */
 }
+
+/*
+ * PostmasterDeathSignalInit - request signal on postmaster death if possible
+ */
+void
+PostmasterDeathSignalInit(void)
+{
+#ifdef USE_POSTMASTER_DEATH_SIGNAL
+	int 		signum = POSTMASTER_DEATH_SIGNAL;
+
+	/* Register our signal handler. */
+	pqsignal(signum, postmaster_death_handler);
+
+	/* Request a signal on parent exit. */
+#if defined(PR_SET_PDEATHSIG)
+	if (prctl(PR_SET_PDEATHSIG, signum) < 0)
+		elog(ERROR, "could not request parent death signal: %m");
+#else
+#error "USE_POSTMASTER_DEATH_SIGNAL set, but there is no mechanism to request the signal"
+#endif
+
+	/*
+	 * Just in case the parent was gone already and we missed it, we'd better
+	 * check the slow way on the first call.
+	 */
+	postmaster_possibly_dead = true;
+#endif							/* USE_POSTMASTER_DEATH_SIGNAL */
+}
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index 03b28c3604a..4bb28938c27 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -43,6 +43,7 @@
 #include "storage/ipc.h"
 #include "storage/latch.h"
 #include "storage/pg_shmem.h"
+#include "storage/pmsignal.h"
 #include "storage/proc.h"
 #include "storage/procarray.h"
 #include "utils/builtins.h"
@@ -304,6 +305,9 @@ InitPostmasterChild(void)
 	if (setsid() < 0)
 		elog(FATAL, "setsid() failed: %m");
 #endif
+
+	/* Request a signal if the postmaster dies, if possible. */
+	PostmasterDeathSignalInit();
 }
 
 /*