On Tue, Jan 12, 2016 at 12:27:05PM +0200, Mika Kuoppala wrote:
> > for_each_ring(ring, dev_priv, i) {
> > - seqno[i] = ring->get_seqno(ring);
> > acthd[i] = intel_ring_get_active_head(ring);
> > + seqno[i] = ring->get_seqno(ring);
>
> Oh! Perhaps I am overly optimistic but did you just show how to solve
> the 'hangcheck elapsed <random> ring idle..' coherency issue
> in hangcheck?
No. There are two causes, one is that we genuinely inspect the seqno
before it is written in the interrupt bottom-half and the other is
indeed the race you keep mentioning between the hangcheck looking at
waiters and the waiter setting itself up.
-Chris
--
Chris Wilson, Intel Open Source Technology Centre
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
