This patch resolves a deadlock between the serialize lock (in
exec/main.c) and the timer lock (in exec/timer.c). I observed this
deadlock happening fairly quickly on a cluster using the EVT service
from OpenAIS. (OpenAIS 1.1.4, Corosync 1.4.1)
In prioritized_timer_thread(), it was grabbing:
1) timer lock
2) serialize lock
In another thread, you have:
1) grab the serialize lock in deliver_fn() of exec/main.c
2) grab the timer lock in corosync_timer_add_duration().
The patch just swaps the locking order in the timer thread.
Signed-off-by: Russell Bryant <[email protected]>
---
exec/timer.c | 4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/exec/timer.c b/exec/timer.c
index 69f9a95..02ca51b 100644
--- a/exec/timer.c
+++ b/exec/timer.c
@@ -131,13 +131,13 @@ static void *prioritized_timer_thread (void *data)
if (fds < 0) {
return NULL;
}
- pthread_mutex_lock (&timer_mutex);
timer_serialize_lock_fn ();
+ pthread_mutex_lock (&timer_mutex);
timerlist_expire (&timers_timerlist);
- timer_serialize_unlock_fn ();
pthread_mutex_unlock (&timer_mutex);
+ timer_serialize_unlock_fn ();
}
}
--
1.7.6
_______________________________________________
Openais mailing list
[email protected]
https://lists.linux-foundation.org/mailman/listinfo/openais
