Re: ✗ i915.CI.BAT: failure for tests/kms_display_modes: Fit modes within MST bandwidth constraints (rev7)

[Imre Deak <imre.deak@intel.com>]

On Mon, May 12, 2025 at 09:45:00AM +0300, Jani Saarinen wrote:
> Hi,
>
> [...]
>
> > tests/kms_display_modes: Fit modes within MST bandwidth constraints (rev7)
> > 
> > URL:
> > 
> > https://patchwork.freedesktop.org/series/146653/
> > 
> > State:
> > 
> > failure
> > 
> > Details:
> > 
> > https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_13109/index.html
> > 
> > [...]
> > 
> > Possible regressions
> > 
> > 
> > *	igt@i915_module_load@load:
> > 
> > 	*	bat-mtlp-9: PASS -> ABORT
> >             <https://intel-gfx-ci.01.org/tree/drm-tip/IGT_8361/bat-mtlp-9/igt@i915_module_load@load.html>
> >             <https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_13109/bat-mtlp-9/igt@i915_module_load@load.html>
> 
> Looks like this https://gitlab.freedesktop.org/drm/i915/kernel/-/issues/13494 , @Deak, Imre should this abort? 

This issue affects only the first modeset after booting, so I don't
think the whole test run should be aborted.

I suppose the abort happend due to the
<4> [41.290737] i915 0000:00:02.0: [drm] pipe_off wait timed out

WARN which doesn't always happen vs. the other related error messages
earlier starting with
<3> [40.411150] i915 0000:00:02.0: [drm] *ERROR* Timeout waiting for DDI BUF D to get active

which do not cause an abort (filtered out as a known issue).

A previous instance with the same WARN is in
drmtip_1838/bat-mtlp-9/24/dmesg.log

which was not aborted.

--Imre