[Mesa-dev] [PATCH] i965/gen7_urb: Re-emit PUSH_CONSTANT_ALLOC on GLK

Nanley Chery nanleychery at gmail.com
Mon Aug 27 22:40:12 UTC 2018


On Mon, Aug 27, 2018 at 03:25:37PM -0700, Kenneth Graunke wrote:
> On Monday, August 27, 2018 11:03:33 AM PDT Nanley Chery wrote:
> > On Fri, Aug 24, 2018 at 05:46:44PM -0700, Nanley Chery wrote:
> > > According to internal docs, some gen9 platforms have a pixel shader push
> > > constant synchronization issue. Although not listed among said
> > > platforms, this issue seems to be present on the GeminiLake 2x6's we've
> > > tested.
> > > 
> > > We consider the available workarounds to be too detrimental on
> > > performance. Instead, we mitigate the issue by applying part of one of
> > > the workarounds. Re-emit PUSH_CONSTANT_ALLOC at the top of every batch
> > > (as suggested by Ken).
> > > 
> > > Fixes ext_framebuffer_multisample-accuracy piglit test failures with the
> > > following options:
> > > * 6 depth_draw small depthstencil
> > > * 8 stencil_draw small depthstencil
> > > * 6 stencil_draw small depthstencil
> > > * 8 depth_resolve small
> > > * 6 stencil_resolve small depthstencil
> > > * 4 stencil_draw small depthstencil
> > > * 16 stencil_draw small depthstencil
> > > * 16 depth_draw small depthstencil
> > > * 2 stencil_resolve small depthstencil
> > > * 6 stencil_draw small
> > > * all_samples stencil_draw small
> > > * 2 depth_draw small depthstencil
> > > * all_samples depth_draw small depthstencil
> > > * all_samples stencil_resolve small
> > > * 4 depth_draw small depthstencil
> > > * all_samples depth_draw small
> > > * all_samples stencil_draw small depthstencil
> > > * 4 stencil_resolve small depthstencil
> > > * 4 depth_resolve small depthstencil
> > > * all_samples stencil_resolve small depthstencil
> > > 
> > > Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=106865
> > > Cc: <mesa-stable at lists.freedesktop.org>
> > > ---
> > >  src/mesa/drivers/dri/i965/gen7_urb.c | 23 +++++++++++++++++++++++
> > >  1 file changed, 23 insertions(+)
> > 
> > Ping?
> 
> So...I believe those tests are still intermittent on other platforms.
> 
> And...these platforms aren't listed as having the bug that you're trying
> to work around here.
> 
> Which makes me wonder if a GLK2x6 hack is really the right thing to do.

Right, it'd be good to know if this helps any other platforms. I'll see
if we could add a more generic patch to CI to see if stability improves
for other platforms.


More information about the mesa-dev mailing list