Bug 102457

Summary:	[IGT] igt@kms_frontbuffer_tracking@basic blows up
Product:	DRI	Reporter:	Martin Peres <martin.peres>
Component:	IGT	Assignee:	Default DRI bug account <dri-devel>
Status:	CLOSED WORKSFORME	QA Contact:
Severity:	critical
Priority:	high	CC:	intel-gfx-bugs
Version:	XOrg git
Hardware:	Other
OS:	All
Whiteboard:	ReadyForDev
i915 platform:	ALL	i915 features:	display/Other

Description Martin Peres 2017-08-28 15:48:07 UTC

On CI_DRM_3011, the machine fi-glk-2a hit the following issue when running igt@kms_frontbuffer_tracking@basic:

(kms_frontbuffer_tracking:3506) intel-batchbuffer-CRITICAL: Test assertion failure function intel_batchbuffer_flush_on_ring, file intel_batchbuffer.c:184:
(kms_frontbuffer_tracking:3506) intel-batchbuffer-CRITICAL: Failed assertion: (drm_intel_gem_bo_context_exec(batch->bo, ctx, used, ring)) == 0
(kms_frontbuffer_tracking:3506) intel-batchbuffer-CRITICAL: Last errno: 5, Input/output error

This appears to be due to the test igt@gem_exec_suspend@basic-s3 htting the WARN_ON(wait_for_engine(engine, 50)) (https://bugs.freedesktop.org/show_bug.cgi?id=102456) which led to a jammed GEM. Many tests just skip after that, but this test apparently is relying on GEM without testing for its availability using igt_require_gem.

Full logs: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3011/fi-glk-2a/igt@kms_frontbuffer_tracking@basic.html

Comment 1 Jari Tahvanainen 2017-09-06 07:15:47 UTC

Changed component to igt since fixing needs to be done there...

Comment 2 Daniel Vetter 2017-09-08 06:21:20 UTC

Moving back.

You do _not_ have to call igt_require_gem on an i915-only testcase, that's for intel-specific subtests in generic tests.

Please never assign igt test failures to the igt component, that's just for tools and stuff like that. For testcases the default assumption must be that the kernel is broken.

Which is the case here.

Comment 3 Daniel Vetter 2017-09-08 06:21:44 UTC

Also changing the summary, Martin is wrong here.

Comment 4 Martin Peres 2017-09-08 07:16:11 UTC

(In reply to Daniel Vetter from comment #3)
> Also changing the summary, Martin is wrong here.

And this is why we always need to trust the actual developers ;)

Sorry that my guess was wrong!

Comment 5 Jari Tahvanainen 2017-09-13 08:42:34 UTC

Moving high as being sporadic

Comment 6 Joonas Lahtinen 2017-09-14 06:26:10 UTC

This needs to be reproduced by a developer. -EIO can result from a few places and none are obvious from the debug logs.

Comment 7 Chris Wilson 2017-09-14 09:09:41 UTC

(In reply to Martin Peres from comment #4)
> (In reply to Daniel Vetter from comment #3)
> > Also changing the summary, Martin is wrong here.
> 
> And this is why we always need to trust the actual developers ;)
> 
> Sorry that my guess was wrong!

No, you were spot on.

Comment 8 Marta Löfstedt 2017-10-16 11:22:52 UTC

This issue was filed against a machine that is no longer in BAT. The issue has never been reproduced on the current GLK machine in BAT.

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.