Re: drm-ci: flaky tests for i915 driver testing

[Vignesh Raman <vignesh.raman@xxxxxxxxxxxxx>]

Hi Rodrigo,

On 11/09/24 01:05, Rodrigo Vivi wrote:
> On Tue, Sep 10, 2024 at 07:32:56AM +0530, Vignesh Raman wrote:
> > Hi Maintainers,
> >
> > On 12/07/24 11:22, Vignesh Raman wrote:
> > > Hi Maintainers,
> > >
> > > On 28/05/24 12:07, Vignesh Raman wrote:
> > > > Hi Maintainers,
> > > >
> > > > There are some flaky tests reported for i915 driver testing in
> > > > drm-ci for the below boards.
> > > >
> > > > *)
> > > > # Board Name: asus-C523NA-A20057-coral
> > > > # IGT Version: 1.28-g0df7b9b97
> > > > # Linux Version: 6.9.0-rc7
> > > > # Failure Rate: 50
> > > > kms_fb_coherency@memset-crc
> > > >
> > > > *)
> > > > # Board Name: asus-C436FA-Flip-hatch
> > > > # IGT Version: 1.28-g0df7b9b97
> > > > # Linux Version: 6.9.0-rc7
> > > > # Failure Rate: 50
> > > > kms_plane_alpha_blend@constant-alpha-min
> > > >
> > > > *)
> > > > # Board Name: hp-x360-12b-ca0010nr-n4020-octopus
> > > > # IGT Version: 1.28-g0df7b9b97
> > > > # Linux Version: 6.9.0-rc7
> > > > # Failure Rate: 50
> > > > core_hotunplug@unplug-rescan
> > > > kms_fb_coherency@memset-crc
> > > >
> > > > *)
> > > > # Board Name: hp-x360-14-G1-sona
> > > > # IGT Version: 1.28-g0df7b9b97
> > > > # Linux Version: 6.9.0-rc7
> > > > # Failure Rate: 50
> > > > prime_busy@hang
> > > >
> > > > *)
> > > > # Board Name: dell-latitude-5400-8665U-sarien
> > > > # IGT Version: 1.28-g0df7b9b97
> > > > # Linux Version: 6.9.0-rc7
> > > > # Failure Rate: 50
> > > > kms_pm_rpm@modeset-lpsp-stress
> > > >
> > > > *)
> > > > # Board Name: asus-C433TA-AJ0005-rammus
> > > > # IGT Version: 1.28-g0df7b9b97
> > > > # Linux Version: 6.9.0-rc7
> > > > # Failure Rate: 50
> > > > i915_hangman@engine-engine-error
> > > > i915_hangman@gt-engine-hang
> > > > kms_async_flips@crc
> > > > kms_universal_plane@cursor-fb-leak
> > > >
> > > > Will add these tests in,
> > > > drivers/gpu/drm/ci/xfails/i915-amly-flakes.txt
> > > > drivers/gpu/drm/ci/xfails/i915-apl-flakes.txt
> > > > drivers/gpu/drm/ci/xfails/i915-cml-flakes.txt
> > > > drivers/gpu/drm/ci/xfails/i915-glk-flakes.txt
> > > > drivers/gpu/drm/ci/xfails/i915-kbl-flakes.txt
> > > > drivers/gpu/drm/ci/xfails/i915-whl-flakes.txt
> > > >
> > > > https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/tree/Documentation/gpu/automated_testing.rst#n70
> > > >
> > > > Please could you have a look at these test results and let us know
> > > > if you need more information. Thank you.
> > >
> > > There are some flaky tests reported for i915 driver testing in drm-ci
> > > with the recent IGT uprev in drm-ci (https://gitlab.freedesktop.org/drm/misc/kernel/-/commit/0e7f4e6a20d550252c4f355d5a303b1d9c8ff052)
> > >
> > > *)
> > > # Board Name: asus-C433TA-AJ0005-rammus
> > > # Bug Report: https://lore.kernel.org/intel-gfx/af4ca4df-a3ef-4943-bdbf-4c3af2c333af@xxxxxxxxxxxxx/T/#u
> > > # Failure Rate: 50
> > > # IGT Version: 1.28-gf13702b8e
> > > # Linux Version: 6.10.0-rc5
> > > kms_sysfs_edid_timing
> > > i915_hangman@engine-engine-hang
> > > kms_pm_rpm@modeset-lpsp-stress
> > >
> > > *)
> > > # Board Name: asus-C436FA-Flip-hatch
> > > # Bug Report: https://lore.kernel.org/intel-gfx/af4ca4df-a3ef-4943-bdbf-4c3af2c333af@xxxxxxxxxxxxx/T/#u
> > > # Failure Rate: 50
> > > # IGT Version: 1.28-gf13702b8e
> > > # Linux Version: 6.10.0-rc5
> > > kms_atomic_transition@plane-all-modeset-transition-internal-panels
> > >
> > > The expectation files have been updated with these tests,
> > > https://gitlab.freedesktop.org/drm/misc/kernel/-/blob/drm-misc-next/drivers/gpu/drm/ci/xfails/i915-amly-flakes.txt
> > >
> > > https://gitlab.freedesktop.org/drm/misc/kernel/-/blob/drm-misc-next/drivers/gpu/drm/ci/xfails/i915-cml-flakes.txt
> >
> > There are some flaky tests reported for i915 driver testing in drm-ci with
> > the recent IGT uprev in drm-ci.
> >
> > *)
> > # Board Name: asus-C433TA-AJ0005-rammus
> > # Failure Rate: 50
> > # IGT Version: 1.28-ga73311079
> > # Linux Version: 6.11.0-rc2
> > kms_pm_rpm@drm-resources-equal
> >
> > *)
> > # Board Name: asus-C523NA-A20057-coral
> > # Failure Rate: 100
> > # IGT Version: 1.28-ga73311079
> > # Linux Version: 6.11.0-rc2
> > kms_universal_plane@cursor-fb-leak
> >
> > *)
> > # Board Name: asus-C436FA-Flip-hatch
> > # Failure Rate: 100
> > # IGT Version: 1.28-ga73311079
> > # Linux Version: 6.11.0-rc2
> > kms_plane_alpha_blend@constant-alpha-min
> >
> > # Board Name: asus-C436FA-Flip-hatch
> > # Failure Rate: 50
> > # IGT Version: 1.28-ga73311079
> > # Linux Version: 6.11.0-rc2
> > kms_async_flips@crc
> >
> > The expectation files have been updated with these tests,
> > https://gitlab.freedesktop.org/drm/misc/kernel/-/blob/drm-misc-next/drivers/gpu/drm/ci/xfails/i915-amly-flakes.txt
> >
> > https://gitlab.freedesktop.org/drm/misc/kernel/-/blob/drm-misc-next/drivers/gpu/drm/ci/xfails/i915-apl-flakes.txt
> >
> > https://gitlab.freedesktop.org/drm/misc/kernel/-/blob/drm-misc-next/drivers/gpu/drm/ci/xfails/i915-cml-flakes.txt
> >
> > Please could you have a look at these test results and let us know if you
> > need more information. Thank you.
>
> Hi Vignesh,
>
> Where are exactly the logs? Any gitlab/issue reported for each case?

There is no gitlab issue created. It is a good idea to create a gitlab 
issue in https://gitlab.freedesktop.org/drm/i915/kernel/-/issues/ (for 
i915 driver) and https://gitlab.freedesktop.org/drm/misc/kernel/-/issues 
for other drivers for each flake test.

Please let me know if this works.

> Getting this for instance:
>
> # Board Name: asus-C436FA-Flip-hatch
> # Failure Rate: 100
> # IGT Version: 1.28-ga73311079
> # Linux Version: 6.11.0-rc2
> kms_plane_alpha_blend@constant-alpha-min
>
> Looking to drm-tip CI:
> https://intel-gfx-ci.01.org/tree/drm-tip/shards-all.html?testfilter=constant-alpha-min
>
> it doesn't look like a 100% failure there.

Yes, so it is reported as flake.

Logs are in,
https://gitlab.freedesktop.org/vigneshraman/linux/-/jobs/63394178
https://gitlab.freedesktop.org/vigneshraman/linux/-/jobs/63340372

The artifacts are available for only 4 weeks.

From the artifacts,
IGT-Version: 1.28-ga73311079 (x86_64) (Linux: 6.11.0-rc5-g5d3429a7e9aa 
x86_64)
Using IGT_SRANDOM=1725947780 for randomisation
Opened device: /dev/dri/card0
Starting subtest: constant-alpha-min
Starting dynamic subtest: pipe-A-eDP-1
Testing plane 0
Testing plane 1
Dynamic subtest pipe-A-eDP-1: SUCCESS (1.094s)
Starting dynamic subtest: pipe-C-eDP-1
Testing plane 0
Testing plane 1
Stack trace:
  #0 ../lib/igt_core.c:2051 __igt_fail_assert()
  #1 [igt_assert_crc_equal+0x9b]
  #2 ../tests/kms_plane_alpha_blend.c:376 constant_alpha_min()
  #3 ../tests/kms_plane_alpha_blend.c:568 run_subtests()
  #4 ../tests/kms_plane_alpha_blend.c:752 __igt_unique____real_main730()
  #5 ../tests/kms_plane_alpha_blend.c:730 main()
  #6 [__libc_init_first+0x8a]
  #7 [__libc_start_main+0x85]
  #8 [_start+0x21]
Dynamic subtest pipe-C-eDP-1: FAIL (1.334s)
Subtest constant-alpha-min: FAIL (2.568s)


> Also, nothing special on 6.11.0-rc2 that could trigger that I'm afraid.
>
> So, it would be good to have real individual bug reports [1] for each case
> and some help on bisecting the commit level as well.

Agree. Will create individual gitlab issues.

> So, with more information we might get more attention to help these cases.
> Otherwise I believe this list will just grow without a proper plan to move
> back to the basic test list.
>
> Thanks,
> Rodrigo.
>
> [1] - https://drm.pages.freedesktop.org/intel-docs/how-to-file-i915-bugs.html

Regards,
Vignesh