[igt-dev] ✗ Fi.CI.IGT: failure for i915/perf_pmu: Compare semaphore and busy measurements (rev2)

[Petri Latvala]

On Thu, Sep 26, 2019 at 09:20:23AM +0100, Tvrtko Ursulin wrote:
> 
> On 26/09/2019 01:13, Patchwork wrote:
> > == Series Details ==
> > 
> > Series: i915/perf_pmu: Compare semaphore and busy measurements (rev2)
> > URL   : https://patchwork.freedesktop.org/series/67183/
> > State : failure
> > 
> > == Summary ==
> > 
> > CI Bug Log - changes from CI_DRM_6954_full -> IGTPW_3498_full
> > ====================================================
> > 
> > Summary
> > -------
> > 
> >    **FAILURE**
> > 
> >    Serious unknown changes coming with IGTPW_3498_full absolutely need to be
> >    verified manually.
> >    If you think the reported changes have nothing to do with the changes
> >    introduced in IGTPW_3498_full, please notify your bug team to allow them
> >    to document this new failure mode, which will reduce false positives in CI.
> > 
> >    External URL: https://patchwork.freedesktop.org/api/1.0/series/67183/revisions/2/mbox/
> > 
> > Possible new issues
> > -------------------
> > 
> >    Here are the unknown changes that may have been introduced in IGTPW_3498_full:
> > 
> > ### IGT changes ###
> > 
> > #### Possible regressions ####
> > 
> >    * igt at i915_pm_sseu@full-enable:
> >      - shard-glk:          [PASS][1] -> [FAIL][2]
> >     [1]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_6954/shard-glk1/igt@i915_pm_sseu@full-enable.html
> >     [2]: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3498/shard-glk5/igt@i915_pm_sseu@full-enable.html
> > 
> > New tests
> > ---------
> > 
> >    New tests have been introduced between CI_DRM_6954_full and IGTPW_3498_full:
> > 
> > ### New IGT tests (6) ###
> > 
> >    * igt at perf_pmu@semaphore-busy-bcs0:
> >      - Statuses : 4 pass(s) 2 skip(s)
> >      - Exec time: [0.0, 6.18] s
> > 
> >    * igt at perf_pmu@semaphore-busy-rcs0:
> >      - Statuses : 4 pass(s) 2 skip(s)
> >      - Exec time: [0.0, 6.21] s
> > 
> >    * igt at perf_pmu@semaphore-busy-vcs0:
> >      - Statuses : 4 pass(s) 2 skip(s)
> >      - Exec time: [0.0, 6.17] s
> > 
> >    * igt at perf_pmu@semaphore-busy-vcs1:
> >      - Statuses : 2 pass(s)
> >      - Exec time: [6.16, 6.18] s
> > 
> >    * igt at perf_pmu@semaphore-busy-vcs2:
> >      - Statuses :
> >      - Exec time: [None] s
> 
> Is this odd?

No such test semaphore-busy-vcs2. It exists only when you
--list-subtests, when executing IGT now uses engine discovery and
doesn't really have that subtest.


> 
> Also..
> 
> [snip]
> 
> > 
> > == Logs ==
> > 
> > For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3498/
> 
> ... I don't see the new tests in HTML?


https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3498/shards-all.html?testfilter=perf_pmu@semaphore-busy

You have to select "shards all" from the view menu up top, the default
view is "issues only".


-- 
Petri Latvala