[Intel-gfx] [PATCH v3 0/2] Handle unsupported configuration with IF-ID

[Daniel Vetter]

On Wed, Jul 19, 2017 at 08:22:03AM +0000, Lankhorst, Maarten wrote:
> Mahesh Kumar schreef op di 18-07-2017 om 18:12 [+0530]:
> > Hi Daniel,
> > 
> > On Monday 17 July 2017 12:56 PM, Daniel Vetter wrote:
> > > On Fri, Jun 30, 2017 at 05:40:58PM +0530, Mahesh Kumar wrote:
> > > > Gen9+ Interlace fetch mode doesn't support few plane
> > > > configurations & pipe scaling.
> > > >  - Y-tile
> > > >  - 90/270 rotation
> > > >  - pipe/plane scaling
> > > >  - 420 planar formats
> > > 
> > > Do we have igt testcases that try to exercise all this? When fixing
> > > bugs,
> > > pls make sure you don't fix the bugs, but also make sure we have
> > > solid
> > > coverage. Given that this escaped for years, it's very likely our
> > > coverage
> > > is _really_ bad and needs to be improve a lot for testing
> > > interlaced modes
> > > ...
> >  We have testdisplay with -y option to test Y-tiling (90/270 rotation
> > need Y-tiling so those are also covered).
> > But AFAIK we don't have any testcase to verify scaling with Interlace
> > mode & other combinations.
> > Should we extend existing IGT test for scaling/rotation/tiling/pixel-
> > format to include Interlace mode as well, or should we have a
> > separate Interlace mode specific IGT
> > which will exercise all combinations with all Interlace modes?
> 
> We need separate tests. testdisplay is nice for testing, but it's not
> automated. Ideally something that runs on all supported platforms,
> regardless of having an actual interlaced display connected. We do
> allow mode override with igt_output_override_mode, which could be used
> for this.

+1, testdisplay is not fully automated validation. We should probably move
it over to tools, really doesn't belong into the igt testcase folder.
-Daniel

> 
> > As IF-ID mode doesn't support all the above combination but PF-ID
> > mode does support them from GEN9(scaling Y-tile 90/270 rotation etc).
> > So I submitted a patch to always enable PF-ID mode for Interlace
> > https://patchwork.freedesktop.org/patch/160275/
> > But Ville suggested not to always enable PF-ID mode instead control
> > fetching mode based on property.
> > Currently there is no open source user for this property. What will
> > you suggest here?
> > 
> > Thanks,
> > -Mahesh
> > > Thanks, Daniel
> > > 
> > > > Changes since V2:
> > > >  - Address review comments from ville
> > > > 
> > > > Mahesh Kumar (2):
> > > >   drm/i915/skl+: Check for supported plane configuration in
> > > > Interlace
> > > >     mode
> > > >   drm/i915/skl+: Scaling not supported in IF-ID Interlace mode
> > > > 
> > > >  drivers/gpu/drm/i915/intel_atomic_plane.c | 15 +++++++++++++++
> > > >  drivers/gpu/drm/i915/intel_display.c      | 15 +++++++++++++++
> > > >  2 files changed, 30 insertions(+)
> > > > 

-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch