Bug 110306

Summary:	Out of memory situations may cause bugs to be filed against the wrong testcase
Product:	DRI	Reporter:	Francesco Balestrieri <francesco.balestrieri>
Component:	DRM/Intel	Assignee:	Intel GFX Bugs mailing list <intel-gfx-bugs>
Status:	CLOSED FIXED	QA Contact:	Intel GFX Bugs mailing list <intel-gfx-bugs>
Severity:	normal
Priority:	medium	CC:	arkadiusz.hiler, intel-gfx-bugs, petri.latvala
Version:	unspecified
Hardware:	All
OS:	All
Whiteboard:	ReadyForDev
i915 platform:	ALL	i915 features:	CI Infra

Description Francesco Balestrieri 2019-04-02 11:39:24 UTC

Sometimes, a test that causes out-of-memory conditions will trigger the OOM killer, causing the abort of the test in question plus in some cases the ones that follow (if OOM killer is still "out for blood"). This causes the bug to be filed against the wrong test case, as it happened for example with Bug 110255.

Comment 1 Lakshmi 2019-07-19 10:49:16 UTC

Arek/Petri, How do we proceed with this bug? Any idea, what are the next steps?

Comment 2 Arek Hiler 2019-09-17 11:30:47 UTC

IIRC there were few things happening in this regard: Petri fine-tuning OOM from the igt_runner and Chris making the tests that stress the memory a bit less stressful.

Comment 3 Arek Hiler 2019-09-17 11:32:43 UTC

Right now it is on people to notice similar issues and keep on tuning things until we get this under control or figure a way of getting them classified as caused by OOM killer realiably.

Comment 4 Petri Latvala 2019-11-12 13:19:07 UTC

Closing this, haven't seen OOM cases in a while. Reopen if there's new sightings.

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.