Looks like in lib/netlink-socket.c, nl_dump_next() is calling
nl_dump_refill() while holding dump->mutex, and nl_dump_refill is calling
nl_sock_recv__() which ultimately stuck at recvmsg(sock->fd, &msg, wait ?
0:MSG_DONTWAIT).

Other threads are thus blocked at acquiring the mutex..


On Mon, Aug 18, 2014 at 11:35 AM, Ben Pfaff <[email protected]> wrote:

> On Mon, Aug 18, 2014 at 11:33:59AM -0400, Xu (Simon) Chen wrote:
> > I am running debian wheezy with 3.14.17 kernel, and openvswitch from
> trunk
> > (2.3.90). It seems that ovs-vsctl commands are all hanging, although it
> > actually worked in terms of adding/removing bridges and ports.
> >
> > In the log, vswitchd log, I see the following:
> > ovs_rcu(urcu3)|WARN|blocked xxx ms waiting for revalidator113 to quiesce
> >
> > Any ideas? Thanks a lot..
>
> I guess we need a backtrace across all threads.
>
_______________________________________________
discuss mailing list
[email protected]
http://openvswitch.org/mailman/listinfo/discuss

Reply via email to