Bug 107401

Summary:	[CI][SHARDS] igt@kms_draw_crc@draw-method-xrgb8888-pwrite-ytiled - fail - Failed assertion: !mismatch
Product:	DRI	Reporter:	Martin Peres <martin.peres>
Component:	DRM/Intel	Assignee:	Intel GFX Bugs mailing list <intel-gfx-bugs>
Status:	CLOSED WORKSFORME	QA Contact:	Intel GFX Bugs mailing list <intel-gfx-bugs>
Severity:	normal
Priority:	highest	CC:	intel-gfx-bugs
Version:	XOrg git
Hardware:	Other
OS:	All
Whiteboard:	ReadyForDev
i915 platform:	GLK	i915 features:	display/Other

Description Martin Peres 2018-07-27 11:30:04 UTC

https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_4549/shard-glk4/igt@kms_draw_crc@draw-method-xrgb8888-pwrite-ytiled.html

(kms_draw_crc:1336) igt_debugfs-CRITICAL: Test assertion failure function igt_assert_crc_equal, file ../lib/igt_debugfs.c:370:
(kms_draw_crc:1336) igt_debugfs-CRITICAL: Failed assertion: !mismatch
Subtest draw-method-xrgb8888-pwrite-ytiled failed.

Comment 1 Martin Peres 2018-07-27 11:32:22 UTC

Setting to highest due to the regression introduced by Chris with this patch series: https://patchwork.freedesktop.org/series/47279/

BAT did not pass because Tomi did not suppress cnl-psr before adding it back to CI, but Chris should have reported the problem to us and hit re-run.

Comment 2 Chris Wilson 2018-07-27 11:39:50 UTC

Don't you mean CI_DRM_4548 instead (or an IGT update at that time)?
https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_4548/shard-glk1/igt@kms_draw_crc@draw-method-xrgb8888-pwrite-ytiled.html

Comment 3 Martin Peres 2018-07-27 11:46:14 UTC

Appologies Chris, the issue was introduced in https://intel-gfx-ci.01.org/tree/drm-tip/IGT_4576/git-log-oneline.log, which looks unrelated...

The mystery thickens...

Comment 4 Chris Wilson 2018-08-08 21:02:36 UTC

We seem to be building quite a collection of bugs where tests only fail for a given shardlist test ordering. It would be cool if we could apply delta debugging (e.g. https://github.com/sambayless/linedd/blob/master/linedd) to reduce the test list to a minimal one; hopefully identifying the culprit.

Comment 5 Chris Wilson 2018-08-10 16:20:53 UTC

I believe this was an artifact of the powermanagement shenanigans follow drv_module_reload.

Comment 6 Martin Peres 2018-09-03 11:45:24 UTC

It did not happen since CI_DRM_4603_full (1 month ago). So seems like it is indeed fixed.

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.