[Intel-gfx] ✗ Fi.CI.BAT: failure for igt: add timeline test cases (rev2)

[Chris Wilson]

Quoting zhoucm1 (2018-12-07 10:36:37)
> 
> 
> On 2018年12月07日 18:29, Chris Wilson wrote:
> > Quoting Patchwork (2018-12-07 10:27:46)
> >> == Series Details ==
> >>
> >> Series: igt: add timeline test cases (rev2)
> >> URL   : https://patchwork.freedesktop.org/series/53743/
> >> State : failure
> >>
> >> == Summary ==
> >>
> >> CI Bug Log - changes from CI_DRM_5281 -> IGTPW_2133
> >> ====================================================
> >>
> >> Summary
> >> -------
> >>
> >>    **FAILURE**
> >>
> >>    Serious unknown changes coming with IGTPW_2133 absolutely need to be
> >>    verified manually.
> >>    
> >>    If you think the reported changes have nothing to do with the changes
> >>    introduced in IGTPW_2133, 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/53743/revisions/2/mbox/
> >>
> >> Possible new issues
> >> -------------------
> >>
> >>    Here are the unknown changes that may have been introduced in IGTPW_2133:
> >>
> >> ### IGT changes ###
> >>
> >> #### Possible regressions ####
> >>
> >>    * igt at amdgpu/amd_basic at userptr:
> >>      - fi-kbl-8809g:       PASS -> DMESG-WARN
> > What fortuitous timing! Maybe you would like to take a stab at the
> > use-after-free in amdgpu's mmu_notifier.
> It's totally new test cases and should be nothing with others, Do you 
> mean I need to resend it after sometime?

It's an unrelated failure that has been plaguing us ever since we added a
kbl-g to our farm. But as it is a use-after-free with core mm coupling,
it's quite serious.
-Chris