[libvirt] [PATCH 2/2 RFC] Purge marked callbacks before dispatching events
I don't have any justification for this except an empirical one: with
this patch the code from Bug 1363628 doesn't crash after "leaking".
I currently don't have properly working valgrind, but I'm working on it
and it might shed some light into this (although it might also not
happen due to the slowdown).
However in the meantime I attempted some analysis and I got even more
confused than before, I guess. With this patch the code doesn't crash,
even though virObjectEventStateQueueDispatch() properly skips callbacks
marked as deleted. What I feel is even more weird, that if I duplicate
the purgatory function (instead of moving it), it crashes again, and I
feel like even more often.
Weirdly-fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1363628
Signed-off-by: Martin Kletzander <mkletzan(a)redhat.com>
---
src/conf/object_event.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/src/conf/object_event.c b/src/conf/object_event.c
index e5af4be68a7e..4066b4673b42 100644
--- a/src/conf/object_event.c
+++ b/src/conf/object_event.c
@@ -821,13 +821,13 @@ virObjectEventStateFlush(virObjectEventStatePtr state)
if (state->timer != -1)
virEventUpdateTimeout(state->timer, -1);
+ /* Purge any deleted callbacks */
+ virObjectEventCallbackListPurgeMarked(state->callbacks);
+
virObjectEventStateQueueDispatch(state,
&tempQueue,
state->callbacks);
- /* Purge any deleted callbacks */
- virObjectEventCallbackListPurgeMarked(state->callbacks);
-
state->isDispatching = false;
virObjectEventStateUnlock(state);
}
--
2.10.1