I failed to reproduce it after running live_hangcheck in a loop for couple hours. And the same preceded with live_contexts just in case.

The failure is so far localized to one particular machine and happened two times only? Given this should we move the priority down from highest?

What I did hit *once* in all these run is probably a rather worrying "ring head not parked" error.

Seen according ti cibuglog on 
fi-skl-iommu
fi-skl-guc
fi-skl-6700k2

(In reply to Jani Saarinen from comment #2)
> Seen according ti cibuglog on 
> fi-skl-iommu
> fi-skl-guc
> fi-skl-6700k2

iommu / guc are unrelated bugs.

There is simply no information here to do any diagnosis. ReadyForDev not.

Last seen 1 week, 3 days / 145 runs ago

Used to happen several time a day, now hasn't happened for 12 days. I think we should close it.

igt@drv_selftest@* are not run, so no results.
Tvrtko, any idea why these tests are not run? Are these tests are not supposed to run? If so, I can close this bug.

(In reply to Lakshmi from comment #6)
> igt@drv_selftest@* are not run, so no results.
> Tvrtko, any idea why these tests are not run? Are these tests are not
> supposed to run? If so, I can close this bug.

they got renamed to igt@i915_selftest@ ;)

So how do we track this in CI bug logger? I still see no failure related to this bug and I have no idea how to look for the same failure in the renamed tests.

(In reply to Francesco Balestrieri from comment #8)
> So how do we track this in CI bug logger? I still see no failure related to
> this bug and I have no idea how to look for the same failure in the renamed
> tests.

Francesco, filter has been updated according to the new name in CI bug log. So, its the same bug in CI bug log where you can look at failures with the new name.

As commented above, this is only happening on fi-skl-iommu and fi-skl-guc (the latter not in a month and a half). Lowering priority.

If this is now restricted to -iommu and -guc hosts, they have their own tracking bugs.

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

{- SKL: igt@i915_selftest@live_hangcheck - incomplete -}
{+ SKL KBL: igt@i915_selftest@live_hangcheck - incomplete +}

New failures caught by the filter:

* https://intel-gfx-ci.01.org/tree/drm-tip/IGT_4795/fi-kbl-8809g/igt@i915_selftest@live_hangcheck.html

(In reply to CI Bug Log from comment #12)
> A CI Bug Log filter associated to this bug has been updated:
> 
> {- SKL: igt@i915_selftest@live_hangcheck - incomplete -}
> {+ SKL KBL: igt@i915_selftest@live_hangcheck - incomplete +}
> 
> New failures caught by the filter:
> 
> *
> https://intel-gfx-ci.01.org/tree/drm-tip/IGT_4795/fi-kbl-8809g/
> igt@i915_selftest@live_hangcheck.html

Another one that dies in the same type of way...

More reset improvements, though this looks ancient by comparison.

This issue is now happening on fi-skl-iommu but not on other platforms.

https://intel-gfx-ci.01.org/tree/drm-tip/IGT_4978/fi-skl-iommu/igt@i915_selftest@live_hangcheck.html

(In reply to Lakshmi from comment #15)
> This issue is now happening on fi-skl-iommu but not on other platforms.
> 
> https://intel-gfx-ci.01.org/tree/drm-tip/IGT_4978/fi-skl-iommu/
> igt@i915_selftest@live_hangcheck.html

The iommu use after free is nothing to do with this and is bug 108602.