Re: ✗ Xe.CI.Full: failure for LMTT invalidation (rev4)

[Michal Wajdeczko]

On 12.07.2025 09:40, Patchwork wrote:
> == Series Details ==
> 
> Series: LMTT invalidation (rev4)
> URL   : https://patchwork.freedesktop.org/series/151092/
> State : failure
> 
> == Summary ==
> 
> CI Bug Log - changes from xe-3403-9f1c50aa2358b47480008da45c785d88b0b27ba3_FULL -> xe-pw-151092v4_FULL
> ====================================================
> 
> Summary
> -------
> 
>   **FAILURE**
> 
>   Serious unknown changes coming with xe-pw-151092v4_FULL absolutely need to be
>   verified manually.
>   
>   If you think the reported changes have nothing to do with the changes
>   introduced in xe-pw-151092v4_FULL, please notify your bug team (I915-ci-infra at lists.freedesktop.org) to allow them
>   to document this new failure mode, which will reduce false positives in CI.
> 
>   
> 
> Participating hosts (4 -> 4)
> ------------------------------
> 
>   No changes in participating hosts
> 
> Possible new issues
> -------------------
> 
>   Here are the unknown changes that may have been introduced in xe-pw-151092v4_FULL:
> 
> ### IGT changes ###
> 
> #### Possible regressions ####
> 
>   * igt at core_getversion@basic:
>     - shard-adlp:         [PASS][1] -> [FAIL][2]
>    [1]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-3403-9f1c50aa2358b47480008da45c785d88b0b27ba3/shard-adlp-3/igt@core_getversion@basic.html
>    [2]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-151092v4/shard-adlp-4/igt@core_getversion@basic.html

looks like a test or CI glitch, it didn't survive long error injection
test that was run before and was killed, impacting all later tests...

<5>[  470.985109] [IGT] Per-test timeout exceeded. Killing the current
test with SIGQUIT.
	
(core_getversion:3939) CRITICAL: Test assertion failure function
__igt_unique____real_main84, file ../tests/core_getversion.c:91:
(core_getversion:3939) CRITICAL: Failed assertion: fd >= 0
(core_getversion:3939) CRITICAL: Last errno: 2, No such file or directory
(core_getversion:3939) CRITICAL: file descriptor fd failed
Test core_getversion failed.

> 
>   * igt at xe_exec_reset@parallel-gt-reset:
>     - shard-adlp:         [PASS][3] -> [DMESG-WARN][4]
>    [3]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-3403-9f1c50aa2358b47480008da45c785d88b0b27ba3/shard-adlp-3/igt@xe_exec_reset@parallel-gt-reset.html
>    [4]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-151092v4/shard-adlp-4/igt@xe_exec_reset@parallel-gt-reset.html

actually, the PASS case is interesting, since there is no reset at all:

<7>[  433.143255] [IGT] xe_exec_reset: executing
<7>[  433.146001] [IGT] xe_exec_reset: starting subtest parallel-gt-reset
<7>[  433.146280] [IGT] xe_exec_reset: finished subtest
parallel-gt-reset, SUCCESS
<7>[  433.146850] xe 0000:00:02.0: [drm:intel_power_well_enable [xe]]
enabling DC_off
<7>[  433.146971] xe 0000:00:02.0: [drm:gen9_set_dc_state.part.0 [xe]]
Setting DC state from 02 to 00
<7>[  433.161367] xe 0000:00:02.0: [drm:drm_client_dev_restore] fbdev: ret=0
<7>[  433.161647] [IGT] xe_exec_reset: exiting, ret=0

while there is reset sequence in DMESG-WARN case, which is expected, as
this is what the test name suggests at least

and it looks similar what is happening sporadically on BMG [1] even
without this series [2]

[1]
https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-151092v4/shard-bmg-4/igt@xe_exec_reset@parallel-gt-reset.html

[2]
https://intel-gfx-ci.01.org/tree/intel-xe/xe-3402-63c7eef609dde585843b5aa6b8f078c211a556fb/shard-bmg-7/igt@xe_exec_reset@parallel-gt-reset.html