[Intel-gfx] ✗ Fi.CI.IGT: failure for series starting with [01/21] drm/i915/gtt: Downgrade Baytrail back to aliasing-ppgtt

[Chris Wilson]

Quoting Patchwork (2019-08-31 00:33:37)
>   Here are the unknown changes that may have been introduced in Patchwork_14231_full:
> 
> ### IGT changes ###
> 
> #### Possible regressions ####
> 
>   * igt at gem_busy@close-race:
>     - shard-snb:          [PASS][1] -> [FAIL][2]
>    [1]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_6806/shard-snb6/igt@gem_busy@close-race.html
>    [2]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_14231/shard-snb6/igt@gem_busy@close-race.html

Hmm, I may have pared away one MI flush too many, or it was really fond
of the extra forced MI_SET_CONTEXT.

>   * igt at gem_softpin@softpin:
>     - shard-skl:          [PASS][3] -> [TIMEOUT][4]
>    [3]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_6806/shard-skl5/igt@gem_softpin@softpin.html
>    [4]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_14231/shard-skl4/igt@gem_softpin@softpin.html
> 
>   * igt at gem_tiled_swapping@non-threaded:
>     - shard-hsw:          [PASS][5] -> [DMESG-FAIL][6]
>    [5]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_6806/shard-hsw6/igt@gem_tiled_swapping@non-threaded.html
>    [6]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_14231/shard-hsw8/igt@gem_tiled_swapping@non-threaded.html
>     - shard-kbl:          [PASS][7] -> [DMESG-FAIL][8]
>    [7]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_6806/shard-kbl6/igt@gem_tiled_swapping@non-threaded.html
>    [8]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_14231/shard-kbl2/igt@gem_tiled_swapping@non-threaded.html

Not great, looks like the oomkiller is even more trigger happy.

> ### Piglit changes ###
> 
> #### Possible regressions ####
> 
>   * spec at arb_gpu_shader5@texturegather at fs-rg-0-uint-cube (NEW):
>     - pig-hsw-4770r:      NOTRUN -> [CRASH][9] +122 similar issues
>    [9]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_14231/pig-hsw-4770r/spec@arb_gpu_shader5@texturegather@fs-rg-0-uint-cube.html

That looked to be a bad errno to vmf_fault conversion.

On the whole, that's only a little smoke.
-Chris