Skip to contentSkip to conversation

Updated autofs-4.1.3-signal-race-fix.patch

PATCH In review Chris Feist 2 participants 2 messages Oct 18, 2004 – Oct 19, 2004 Kernel automounter

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 Updated autofs-4.1.3-signal-race-fix.patch

Ian,

I've attached an updated patch from Jeff Moyer's earlier signal race patch which allows the SIGCHLD signal to be received in the middle of st_expire() before ap.state can be set to ST_EXPIRE. This causes the signal handling function for SIG_CHLD to never run the alarm() function which causes the SIGALRM signal to never be sent (or received) by automount, which causes directories w/ a timeout to never be unmounted.

Thanks,
Chris

daemon/automount.c renamed from daemon/automount.c.signal-race-fix +3 -2
@@ -833,10 +833,10 @@
 	alarm(0);
 
 	/* Prevent any new mounts */
-	ap.state = ST_SHUTDOWN_PENDING;
-
 	sigprocmask(SIG_SETMASK, &lock_sigs, NULL);
 
+	ap.state = ST_SHUTDOWN_PENDING;
+
 	/* Unmount everything */
 	exp = expire_proc(1);
 
@@ -904,6 +904,7 @@
 		return 1;
 
 	case EXP_STARTED:
+		sigprocmask(SIG_SETMASK, &lock_sigs, NULL);
 		ap.state = ST_EXPIRE;
 		sigprocmask(SIG_SETMASK, &ready_sigs, NULL);
 		return 0;

_______________________________________________
autofs mailing list
[email protected]
http://linux.kernel.org/mailman/listinfo/autofs

maintainer
#1 lore raw
Quoted 9 lines · Chris Feist

Ian,

I've attached an updated patch from Jeff Moyer's earlier signal race patch which allows the SIGCHLD signal to be received in the middle of st_expire() before ap.state can be set to ST_EXPIRE. This causes the signal handling function for SIG_CHLD to never run the alarm() function which causes the SIGALRM signal to never be sent (or received) by automount, which causes directories w/ a timeout to never be unmounted.

Cool. Thanks.

I'll check and add it to the list of patches I'm collecting.

Ian