This seems to be a regression introduced in drmtip_53, bumping the priority. See also https://bugs.freedesktop.org/show_bug.cgi?id=106723

https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_4262/fi-glk-j4005/igt@kms_flip@basic-flip-vs-wf_vblank.html

(kms_flip:3486) CRITICAL: Test assertion failure function check_state, file ../tests/kms_flip.c:530:
(kms_flip:3486) CRITICAL: Failed assertion: fabs((elapsed - expected) / expected) <= 0.005
(kms_flip:3486) CRITICAL: Last errno: 25, Inappropriate ioctl for device
(kms_flip:3486) CRITICAL: inconsistent flip ts/seq: last 405.138609/757, current 405.348146/769: elapsed=209537.0us expected=200009.2us
Subtest basic-flip-vs-wf_vblank failed.

https://intel-gfx-ci.01.org/tree/drm-tip/drmtip_56/fi-glk-j4005/igt@kms_flip@flip-vs-absolute-wf_vblank-interruptible.html

(kms_flip:1451) CRITICAL: Test assertion failure function check_state, file ../tests/kms_flip.c:530:
(kms_flip:1451) CRITICAL: Failed assertion: fabs((elapsed - expected) / expected) <= 0.005
(kms_flip:1451) CRITICAL: Last errno: 4, Interrupted system call
(kms_flip:1451) CRITICAL: inconsistent flip ts/seq: last 370.659160/6855, current 370.868326/6867: elapsed=209166.0us expected=200013.0us
Subtest flip-vs-absolute-wf_vblank-interruptible failed.

This issue wasn't seen since 1 month 1 week. Before that it used to fail very frequently. Looking at the history, this issue needs to be kept under observation at least for the next 1 month.

A CI Bug Log filter associated to this bug has been updated:

{- GLK: igt@kms_flip@*flip* - fail - Failed assertion: fabs((elapsed - expected) / expected) <= 0.005 -}
{+ GLK ICL: igt@kms_flip@*flip* - fail - Failed assertion: fabs((elapsed - expected) / expected) <= 0.005 +}

New failures caught by the filter:

* https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_5386/shard-iclb3/igt@kms_flip@flip-vs-absolute-wf_vblank.html
* https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_5386/shard-iclb3/igt@kms_flip@plain-flip-fb-recreate.html

CI_DRM_5386_full (2 months, 2 weeks / 1511 runs ago). Dropping priority to Medium.

From Mika:
The test expects that we flip in a certain time frame. For yet unknown reasons, the flip may take longer than we anticipate. I suspect that user doesn't realize if flip took one or two milliseconds more time. Maybe we could relax a requirement a bit?

Now last seen 3 months ago, close?

(In reply to Jani Saarinen from comment #8)
> Now last seen 3 months ago, close?

Used to happen every other run in average, now not seen since drmtip_135 (128 runs ago). It however was seen once on ICL, 3 months ago. Closing now!

The CI Bug Log issue associated to this bug has been archived.

New failures matching the above filters will not be associated to this bug anymore.