Bug 110829

Putting the bug on GUC because it recently got updated and so far the only machine with this new issue are fi-kbl-guc.

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

### New filters associated

* GUC: igt@i915_pm_rpm@basic-pci-d3-state - fail - Failed assertion: wait_for_suspended()
  - https://intel-gfx-ci.01.org/tree/drm-tip/IGT_5028/fi-kbl-guc/igt@i915_pm_rpm@basic-pci-d3-state.html

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

{- GUC: igt@i915_pm_rpm@basic-pci-d3-state - fail - Failed assertion: wait_for_suspended() -}
{+ GUC: igt@i915_pm_rpm@basic-pci-d3-state|igt@i915_pm_rpm@system-suspend-devices - fail - Failed assertion: wait_for_suspended() +}

New failures caught by the filter:

  * https://intel-gfx-ci.01.org/tree/drm-tip/drmtip_306/fi-kbl-guc/igt@i915_pm_rpm@system-suspend-devices.html

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

{- GUC: igt@i915_pm_rpm@basic-pci-d3-state|igt@i915_pm_rpm@system-suspend-devices - fail - Failed assertion: wait_for_suspended() -}
{+ GUC: igt@i915_pm_rpm@* - fail - Failed assertion: wait_for_suspended() +}

New failures caught by the filter:

  * https://intel-gfx-ci.01.org/tree/drm-tip/drmtip_308/fi-kbl-guc/igt@i915_pm_rpm@pm-caching.html

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

{- GUC: igt@i915_pm_rpm@* - fail - Failed assertion: wait_for_suspended() -}
{+ GUC: igt@i915_pm_rpm@* - fail - Failed assertion: wait_for_suspended() +}

New failures caught by the filter:

  * https://intel-gfx-ci.01.org/tree/drm-tip/drmtip_312/fi-kbl-guc/igt@i915_pm_rpm@debugfs-forcewake-user.html

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

{- GUC: igt@i915_pm_rpm@* - fail - Failed assertion: wait_for_suspended() -}
{+ GUC: igt@i915_pm_rpm@* - fail - Failed assertion: wait_for_suspended() +}

New failures caught by the filter:

  * https://intel-gfx-ci.01.org/tree/drm-tip/drmtip_316/fi-kbl-guc/igt@i915_pm_rpm@gem-idle.html

Looks like a display related issue, possibly on the target board. 
intel_hdmi_detect() is called by the drm subsystem when the device is suspended, and then an HDMI interrupt wakes it back up. There are several HDMI-related logs in dmesg about invalid settings.

(In reply to Jon Ewins from comment #7)
> Looks like a display related issue, possibly on the target board. 
> intel_hdmi_detect() is called by the drm subsystem when the device is
> suspended, and then an HDMI interrupt wakes it back up. There are several
> HDMI-related logs in dmesg about invalid settings.

Thanks for looking into this. However, could you follow the normal bug assessment process so as a priority could be set? Right now, the only thing you have done is saying it wasn't likely due to the GuC.

(In reply to Martin Peres from comment #8)
> (In reply to Jon Ewins from comment #7)
> > Looks like a display related issue, possibly on the target board. 
> > intel_hdmi_detect() is called by the drm subsystem when the device is
> > suspended, and then an HDMI interrupt wakes it back up. There are several
> > HDMI-related logs in dmesg about invalid settings.
> 
> Thanks for looking into this. However, could you follow the normal bug
> assessment process so as a priority could be set? Right now, the only thing
> you have done is saying it wasn't likely due to the GuC.

Hey Martin/John, I'll do the bug assessment process on this. Thanks.

Issue: http://gfx-ci.fi.intel.com/cibuglog-ng/issue/1496

There are several of the 'gt@i915_pm_rpm' tests failing ('tests/i915/i915_pm_rpm.c') on fi-kbl-guc starting with IGT_5028 (2 months, 2 weeks old) all the way to present drmtip_345 (3 days, 20 hours old).

All of the tests are failing while wait_for_suspended(). As Jon pointed out, there a lots of HDMI errors in the logs and since this appears to be only happening on one machine ('fi-kbl-guc') so this machine should be looked into. Either way, this looks to be a display issue.

This appears to be 100% reproducible on 'fi-kbl-guc' Setting priority to medium as issue appears isolated to a single machine.

Efforts underway to test guc usage disabled to identify if this is not related to GuC and more specific to the display configuration of this device.

I created some patches to IGT to only run the failing test and to i915 to force GuC off. I had some issues learning how to get the patches to run together. I finally got it working but then the host fi-kbl-guc skipped the test. I submitted the patches to trybot again and I'm waiting for the results.

There really seems to be an issue with 'fi-kbl-guc' as every trybot run excludes this machine. Give that, and the HDMI issues, this really seems to be a machine/display specific issue. 

Lakshmina: Can you please assign this to the display group as I'm not sure who that would be.

-- 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/704.