[Intel-gfx] ✗ Fi.CI.IGT: failure for drm/i915/tgl: Fix Combo PHY DPLL fractional divider for 38.4MHz ref clock (rev3)

[Vudum, Lakshminarayana]

Thanks for the pointers Imre. I have re-reported the results.

Lakshmi.
-----Original Message-----
From: Imre Deak <imre.deak at intel.com> 
Sent: Saturday, October 3, 2020 6:49 AM
To: intel-gfx at lists.freedesktop.org; Vudum, Lakshminarayana <lakshminarayana.vudum at intel.com>; Sarvela, Tomi P <tomi.p.sarvela at intel.com>
Subject: Re: ✗ Fi.CI.IGT: failure for drm/i915/tgl: Fix Combo PHY DPLL fractional divider for 38.4MHz ref clock (rev3)

Hi Lakshmi, Tomi,

On Sat, Oct 03, 2020 at 03:56:00AM +0000, Patchwork wrote:
> == Series Details ==
> 
> Series: drm/i915/tgl: Fix Combo PHY DPLL fractional divider for 38.4MHz ref clock (rev3)
> URL   : https://patchwork.freedesktop.org/series/82173/
> State : failure
> 
> == Summary ==
> 
> CI Bug Log - changes from CI_DRM_9093_full -> Patchwork_18620_full 
> ====================================================
> 
> Summary
> -------
> 
>   **FAILURE**
> 
>   Serious unknown changes coming with Patchwork_18620_full absolutely need to be
>   verified manually.
>   
>   If you think the reported changes have nothing to do with the changes
>   introduced in Patchwork_18620_full, please notify your bug team to allow them
>   to document this new failure mode, which will reduce false positives in CI.
> 
>   
> 
> Possible new issues
> -------------------
> 
>   Here are the unknown changes that may have been introduced in Patchwork_18620_full:
> 
> ### IGT changes ###
> 
> #### Possible regressions ####
> 
>   * igt at gem_userptr_blits@unsync-unmap-cycles:
>     - shard-skl:          [PASS][1] -> [TIMEOUT][2]
>    [1]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_9093/shard-skl6/igt@gem_userptr_blits@unsync-unmap-cycles.html
>    [2]: 
> https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_18620/shard-skl4/ig
> t at gem_userptr_blits@unsync-unmap-cycles.html

This looks like
https://gitlab.freedesktop.org/drm/intel/-/issues/2424
still happening at
https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_9090/shard-skl5/igt@gem_userptr_blits@sync-unmap-cycles.html

Could you update the filter to include sync-unmap-cycles as well?


>   * igt at kms_flip@flip-vs-suspend at c-edp1:
>     - shard-iclb:         [PASS][3] -> [INCOMPLETE][4]
>    [3]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_9093/shard-iclb3/igt@kms_flip@flip-vs-suspend@c-edp1.html
>    [4]: 
> https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_18620/shard-iclb3/i
> gt at kms_flip@flip-vs-suspend at c-edp1.html

Looks like shard-iclb3 has a file system corruption, or just broken storage device/host:
<3>[  240.012780] blk_update_request: I/O error, dev sda, sector 76863120 op 0x1:(WRITE) flags 0x800 phys_seg 10 prio class 0

This has been consistently happening now for a while on the shard-iclb3 machine leading to the same hang, see for instance https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_9063/shard-iclb3/pstore7-1601302770_Panic_1.txt

Similar sporadic filesystem errors have been also happening on shard-iclb4.

I haven't seen any recent I/O errors on any of the other 7 shard-icl machines.

shard-iclb3/4 would probably need reinstall/new storage device/new host.

--Imre