[igt-dev] ✗ Fi.CI.IGT: failure for tests/kms_cursor_crc: start crc only once per test (rev4)

[Juha-Pekka Heikkila]

On 28.11.2019 15.39, Petri Latvala wrote:
> On Thu, Nov 28, 2019 at 01:29:51PM +0000, Patchwork wrote:
>> == Series Details ==
>>
>> Series: tests/kms_cursor_crc: start crc only once per test (rev4)
>> URL   : https://patchwork.freedesktop.org/series/70061/
>> State : failure
>>
>> == Summary ==
>>
>> CI Bug Log - changes from CI_DRM_7431_full -> IGTPW_3768_full
>> ====================================================
>>
>> Summary
>> -------
>>
>>    **FAILURE**
>>
>>    Serious unknown changes coming with IGTPW_3768_full absolutely need to be
>>    verified manually.
>>    
>>    If you think the reported changes have nothing to do with the changes
>>    introduced in IGTPW_3768_full, please notify your bug team to allow them
>>    to document this new failure mode, which will reduce false positives in CI.
>>
>>    External URL: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3768/index.html
>>
>> Possible new issues
>> -------------------
>>
>>    Here are the unknown changes that may have been introduced in IGTPW_3768_full:
>>
>> ### IGT changes ###
>>
>> #### Possible regressions ####
>>
>>    * igt at i915_selftest@live_blt:
>>      - shard-hsw:          [PASS][1] -> [DMESG-FAIL][2]
>>     [1]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_7431/shard-hsw8/igt@i915_selftest@live_blt.html
>>     [2]: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3768/shard-hsw1/igt@i915_selftest@live_blt.html
>>
>>    * igt at kms_cursor_crc@pipe-c-cursor-128x42-sliding:
>>      - shard-iclb:         NOTRUN -> [DMESG-WARN][3]
>>     [3]: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3768/shard-iclb3/igt@kms_cursor_crc@pipe-c-cursor-128x42-sliding.html
> 
> 
> 
> <3> [544.041813] [drm:drm_crtc_add_crc_entry] *ERROR* Overflow of CRC buffer, userspace reads too slow.
> 

Ya, I noticed that too. There's those fb being painted during test and 
it seems those need to be rendered in setup and used from buffered 
versions. I'll look into it, it probably shave off some extra time from 
these tests -- though no hundreds of second to shave off anymore :D


/Juha-Pekka