[Intel-gfx] ✗ Fi.CI.BAT: failure for drm/i915: Skip object locking around a no-op set-domain ioctl

[Chris Wilson]

Quoting Chris Wilson (2019-03-21 09:11:30)
> Quoting Patchwork (2019-03-21 09:04:59)
> > == Series Details ==
> > 
> > Series: drm/i915: Skip object locking around a no-op set-domain ioctl
> > URL   : https://patchwork.freedesktop.org/series/58325/
> > State : failure
> > 
> > == Summary ==
> > 
> > CI Bug Log - changes from CI_DRM_5785 -> Patchwork_12542
> > ====================================================
> > 
> > Summary
> > -------
> > 
> >   **FAILURE**
> > 
> >   Serious unknown changes coming with Patchwork_12542 absolutely need to be
> >   verified manually.
> >   
> >   If you think the reported changes have nothing to do with the changes
> >   introduced in Patchwork_12542, 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/58325/revisions/1/mbox/
> > 
> > Possible new issues
> > -------------------
> > 
> >   Here are the unknown changes that may have been introduced in Patchwork_12542:
> > 
> > ### IGT changes ###
> > 
> > #### Possible regressions ####
> > 
> >   * igt at gem_exec_reloc@basic-cpu-active:
> >     - fi-bsw-kefka:       PASS -> FAIL +7
> >     - fi-hsw-4770r:       PASS -> FAIL +3
> >     - fi-cfl-8109u:       PASS -> FAIL +3
> >     - fi-byt-clapper:     PASS -> FAIL +7
> >     - fi-ivb-3770:        PASS -> FAIL +3
> >     - fi-icl-u3:          PASS -> FAIL +3
> >     - fi-skl-gvtdvm:      PASS -> FAIL +2
> 
> Intriguing, now that might explain some of the random failure we've
> seen here over the years, as we unexpectedly take the slow path.

Sadly, not these are the active variant so the inaccurate test
shortcircuited the required read synchronisation.

And I hope the lack of the read clflush will explain the !llc..
-Chris