Use SA_RESTART for all signals, including SIGALRM.

The exclusion of SIGALRM dates back to Berkeley days, when Postgres used
SIGALRM in only one very short stretch of code.  Nowadays, allowing it to
interrupt kernel calls doesn't seem like a very good idea, since its use
for statement_timeout means SIGALRM could occur anyplace in the code, and
there are far too many call sites where we aren't prepared to deal with
EINTR failures.  When third-party code is taken into consideration, it
seems impossible that we ever could be fully EINTR-proof, so better to
use SA_RESTART always and deal with the implications of that.  One such
implication is that we should not assume pg_usleep() will be terminated
early by a signal.  Therefore, long sleeps should probably be replaced
by WaitLatch operations where practical.

Back-patch to 9.3 so we can get some beta testing on this change.

src/backend/port/sysv_sema.c

@@ -405,8 +405,8 @@ PGSemaphoreLock(PGSemaphore sema, bool interruptOK)
 	 * it's necessary for cancel/die interrupts to be serviced directly by the
 	 * signal handler.	On these platforms the behavior is really the same
 	 * whether the signal arrives just before the semop() begins, or while it
-	 * is waiting.	The loop on EINTR is thus important only for other types
-	 * of interrupts.
+	 * is waiting.	The loop on EINTR is thus important only for platforms
+	 * without SA_RESTART.
 	 */
 	do
 	{

src/port/pqsignal.c

@@ -60,9 +60,7 @@ pqsignal(int signo, pqsigfunc func)
 
 	act.sa_handler = func;
 	sigemptyset(&act.sa_mask);
-	act.sa_flags = 0;
-	if (signo != SIGALRM)
-		act.sa_flags |= SA_RESTART;
+	act.sa_flags = SA_RESTART;
 #ifdef SA_NOCLDSTOP
 	if (signo == SIGCHLD)
 		act.sa_flags |= SA_NOCLDSTOP;