Bug 110384

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

### New filters associated

* ICL: igt@kms_flip@* - incomplete - system hang/timeout
  - https://intel-gfx-ci.01.org/tree/drm-tip/drmtip_248/fi-icl-u2/igt@kms_flip@2x-blocking-absolute-wf_vblank.html
  - https://intel-gfx-ci.01.org/tree/drm-tip/drmtip_250/fi-icl-u2/igt@kms_flip@flip-vs-modeset-vs-hang-interruptible.html
  - https://intel-gfx-ci.01.org/tree/drm-tip/drmtip_250/fi-icl-u2/igt@kms_flip@2x-dpms-vs-vblank-race-interruptible.html
  - https://intel-gfx-ci.01.org/tree/drm-tip/drmtip_250/fi-icl-u3/igt@kms_flip@bo-too-big-interruptible.html
  - https://intel-gfx-ci.01.org/tree/drm-tip/drmtip_250/fi-icl-u3/igt@kms_flip@2x-flip-vs-modeset-vs-hang-interruptible.html
  - https://intel-gfx-ci.01.org/tree/drm-tip/drmtip_250/fi-icl-y/igt@kms_flip@bo-too-big-interruptible.html
  - https://intel-gfx-ci.01.org/tree/drm-tip/drmtip_251/fi-icl-u2/igt@kms_flip@flip-vs-modeset-vs-hang-interruptible.html
  - https://intel-gfx-ci.01.org/tree/drm-tip/drmtip_251/fi-icl-u3/igt@kms_flip@2x-flip-vs-modeset-vs-hang-interruptible.html

Also seen on shards:
https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_5922/shard-iclb7/igt@kms_flip@flip-vs-modeset-vs-hang-interruptible.html

Bumping the priority!

I looked through all the pstore panics (there's only 3 I've found). We died because watchdog timeout and then reboot. There's one more pstore where we get a really pretty lockdep splat fireworks afterwards, but that's standard issue locking troubles from panic paths. Not a real thing.

Looking at logs, the test simply stops.

I have no idea at all. Would be interesting to see where else this happens, both in terms of tests or in terms of other platforms.

Ok so a large pile of these are the Jenkins SSH connector going upside down. Tomi says he fixed those.

But there's a bunch in here which look like serious issues, e.g.

https://intel-gfx-ci.01.org/tree/drm-tip/drmtip_250/fi-icl-u2/igt@kms_flip@flip-vs-modeset-vs-hang-interruptible.html

- khungtaskd fires
- unfortunately all the useful data from the hung task checker in pstore is drowned out in the ftrace dump (I'm trying to create a patch for this)
- plus lockdep splat (already wrote a patch for that, need to convince myself it's correct)
- plus panic when we probably shouldn't (working with Tomi to fix the Kconfig)

hung task on icl: not good news at all.

Would be good to know whether this also happens on other platforms, but that needs someone with access to the CI results filesystem to grep for khungtaskd - cibuglog ins't good enough.

kms_flip tests various cases regarding flipping. These flip cases are against plain flip, vblank, fencing, dpms, modesetting, panning, hanging, race conditions etc. 

Failure on flipping could cause flickering image.

#assessment

-- GitLab Migration Automatic Message --

This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity.

You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/drm/intel/issues/261.