[Intel-gfx] [PATCH 13/15] drm/i915: Stop setting wraparound seqno on initialisation

Chris Wilson chris at chris-wilson.co.uk
Fri Dec 4 01:51:15 PST 2015

On Fri, Dec 04, 2015 at 10:36:45AM +0100, Daniel Vetter wrote:
> On Tue, Dec 01, 2015 at 04:57:11PM +0000, Dave Gordon wrote:
> > On 29/11/15 08:48, Chris Wilson wrote:
> > >We have testcases to ensure that seqno wraparound works fine, so we can
> > >forgo forcing everyone to encounter seqno wraparound during early
> > >uptime. seqno wraparound incurs a full GPU stall so not forcing it
> > >will eliminate one jitter from the early system.
> > >
> > >Advancing the global next_seqno after a GPU reset is equally pointless.
> > >
> > >Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
> > 
> > The real bug is that seqno wraparound forces GPU idle, can we fix that
> > instead?
> hw semaphores will fall over if we don't, but I guess we can forgo the
> idle if there's no hw semaphores enabled. And yeah I'd like to keep this
> little trick to exercise wrap-around more often too.

But we have a test case that you can run as often as you like as part of
CI setup even, to set the wraparound to happen at a psuedo-random location
during the testing.

Chris Wilson, Intel Open Source Technology Centre

More information about the Intel-gfx mailing list