[igt-dev] [PATCH 2/2] kms_addfb_basic: Check for AMDGPU and MSM bad pitch.

Arkadiusz Hiler arek at hiler.eu
Fri Jul 9 12:32:09 UTC 2021


On Wed, Jul 07, 2021 at 12:09:14PM -0400, Mark Yacoub wrote:
> This is odd, idrk.
> 
> On Tue, Jul 6, 2021 at 9:41 AM Rodrigo Siqueira
> <Rodrigo.Siqueira at amd.com> wrote:
> >
> > On 07/01, Mark Yacoub wrote:
> > > From: Mark Yacoub <markyacoub at google.com>
> > >
> > > [Why]
> > > AMDGPU and MSM devices have a known errno when it fails coorectly for
> > > bad pitches.
> > >
> > > [How]
> > > Assert `errno` to `EINVAL` for both AMDGPU and MSM as well.
> > >
> > > Signed-off-by: Mark Yacoub <markyacoub at chromium.org>
> > > ---
> > >  tests/kms_addfb_basic.c | 2 +-
> > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > >
> > > diff --git a/tests/kms_addfb_basic.c b/tests/kms_addfb_basic.c
> > > index 85c14be5..91fb6ac9 100644
> > > --- a/tests/kms_addfb_basic.c
> > > +++ b/tests/kms_addfb_basic.c
> > > @@ -310,7 +310,7 @@ static void pitch_tests(int fd)
> > >                       f.pitches[0] = bad_pitches[i];
> > >                       igt_assert_eq(igt_ioctl(fd, DRM_IOCTL_MODE_ADDFB2, &f), -1);
> > >                       igt_assert(errno != 0);
> > > -                     if (is_i915_device(fd)) {
> > > +                     if (is_i915_device(fd) || is_amdgpu_device(fd) || is_msm_device(fd)) {
> > >                               igt_assert_eq(errno, EINVAL);
> > >                       } else if (is_nouveau_device(fd)) {
> > >                               if (bad_pitches[i] > 4 * 1024)
> > > --
> > > 2.32.0.93.g670b81a890-goog
> > >
> >
> > Reviewed-by: Rodrigo Siqueira <Rodrigo.Siqueira at amd.com>
> >
> > Btw, from the CI log, we have:
> >
> > igt at kms_addfb_basic@invalid-smem-bo-on-discrete:
> >   shard-apl: NOTRUN -> FAIL
> >   shard-tglb: NOTRUN -> FAIL
> >   shard-glk: NOTRUN -> FAIL
> >   shard-kbl: NOTRUN -> FAIL
> >   shard-snb: NOTRUN -> FAIL
> >   shard-iclb: NOTRUN -> FAIL
> >
> > igt at kms_big_fb@yf-tiled-max-hw-stride-64bpp-rotate-0-async-flip:
> >   shard-iclb: NOTRUN -> SKIP +1 similar issue
> >
> > I don't know why your patch makes something that was not running start
> > to run. Any theory?

The CI is not perfect and the test were just not run previously due to
unrelated issues. I've merged your changes as this is not related to
them.

As of the failing tests - those are new and were introduced recently by
5bb62cb290fa ("tests/kms_addfb_basic: Add invalid buffer object test for
discrete") and seem to fail reliably:

https://intel-gfx-ci.01.org/tree/drm-tip/shards-all.html?testfilter=invalid-smem-bo-on-discrete

I've CCed the author of the tests so they can have a look and fix the
tests

-- 
Cheers,
Arek


More information about the igt-dev mailing list