[ovs-dev] Deferring ofproto_class::destruct vs. ovs-appctl exit
petrm at mellanox.com
Tue Oct 4 21:34:44 UTC 2016
Petr Machata <petrm at mellanox.com> writes:
> We don't actually see a crash, but an AddressSanitizer citation. As I'm
> now trying to reproduce on vanilla branch-2.5, I actually can't either
> (using the NULL trick). All the rule_delete calls end up being
> scheduled before the backer close. In our code, they come later.
> Let me look into it some more.
I had a chance to look at this some more on Monday. I think the problem
is that our main thread is taking a long time to quiesce, and as a
result the deferred work is only done later on. We actually see
warnings about the thread taking unreasonably long to quiesce.
More information about the dev