New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 607393 link

Starred by 1 user

Issue metadata

Status: Duplicate
Merged: issue 606021
Owner:
Closed: May 2016
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Windows
Pri: 2
Type: Bug



Sign in to add a comment

WebglConformance.deqp_functional_gles3_texturefiltering_cube_combinations_01 failing/flaky in Windows/x64 release ANGLE bot

Project Member Reported by jmad...@chromium.org, Apr 28 2016

Issue description

Doesn't seem to happen on the main/FYI GPU waterfalls. Unsure why.

Example failure:

https://build.chromium.org/p/tryserver.chromium.angle/builders/win_angle_x64_rel_ng/builds/1159

Only seems to happen in release/x64, but is happening often in try jobs:
https://build.chromium.org/p/tryserver.chromium.angle/builders/win_angle_x64_rel_ng

Error text:

Init testcase: filtering.cube_combinations.linear_linear_mirror_mirror
Test0
Coordinates: -1.25,-1.2 -> 1.2,1.25
Face -X
Face +X
Face -Y
Face +Y
Face -Z
filtering.cube_combinations.linear_linear_mirror_mirror: Cannot remap cube coordinates
FAIL Cannot remap cube coordinates
Error: Cannot remap cube coordinates

Assigning to current GPU sherrif ccameron but we might need to add a suppression as it's causing issue for the ANGLE CQ.
 
Odd. I'd like to understand more about why this is failing.
Me too. I tried running it a bunch locally with Release/x64 ToT ANGLE and couldn't repro.
Okay, then I'll just go with a blacklist.
Hmm, there are conflicts with the suppression from  issue 606074 .

Comment 5 by kbr@chromium.org, May 2 2016

Status: Assigned (was: Untriaged)
Chris, do you need help untangling suppressions for these failures on various platforms? The suppressions in  Issue 606074  should be Mac-specific and should not collide with expectations on Windows.

Project Member

Comment 6 by bugdroid1@chromium.org, May 10 2016

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/3368e697ee669e6d315fd868918c55cdb731d0c0

commit 3368e697ee669e6d315fd868918c55cdb731d0c0
Author: jmadill <jmadill@chromium.org>
Date: Tue May 10 21:41:32 2016

Suppress gles3_texturefiltering_cube_combinations on Windows/rel

This will help fix the ANGLE trybots.

BUG= 607393 
TBR=kbr@chromium.org,zmo@chromium.org
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel

Review-Url: https://codereview.chromium.org/1967753002
Cr-Commit-Position: refs/heads/master@{#392732}

[modify] https://crrev.com/3368e697ee669e6d315fd868918c55cdb731d0c0/content/test/gpu/gpu_tests/webgl2_conformance_expectations.py

Labels: -Pri-3 Pri-2
Owner: jmad...@chromium.org
Looks like the bots are not retrying these flaky failures despite the CL in #6.. Ken can you help me figure out what's going on here? We really should fix this.

Comment 8 by kbr@chromium.org, May 11 2016

Jamie, the suppressions in src/content/test/gpu/gpu_tests/webgl2_conformance_expectations.py are wrong -- 'deqp/functional/gles3/texturefiltering/cube_combinations_01' should be 'deqp/functional/gles3/texturefiltering/cube_combinations_01.html' and similarly for _05. I'm surprised that the presubmit checks didn't catch this (please talk with cwallez@), but could you please fix them?

Thanks for pointing that out... I'll try to fix the presubmit with Corentin's help.
Project Member

Comment 10 by bugdroid1@chromium.org, May 13 2016

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/c88d76425b8d15a6f569c04557ad292312f67517

commit c88d76425b8d15a6f569c04557ad292312f67517
Author: cwallez <cwallez@chromium.org>
Date: Fri May 13 16:51:19 2016

WebGL expectations: check the path for Flaky expectations too

BUG= 607393 
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel

Review-Url: https://codereview.chromium.org/1973053002
Cr-Commit-Position: refs/heads/master@{#393551}

[modify] https://crrev.com/c88d76425b8d15a6f569c04557ad292312f67517/content/test/gpu/gpu_tests/webgl2_conformance_expectations.py
[modify] https://crrev.com/c88d76425b8d15a6f569c04557ad292312f67517/content/test/gpu/gpu_tests/webgl_conformance_expectations.py

These flaky tests don't seem to be picking up this expectation. It's hard to tell from the test output what expectation is actually selected for these tests. Landing a change now to try and mark them as "Fail" to try and get the bots green.
Project Member

Comment 12 by bugdroid1@chromium.org, May 17 2016

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/5973a8b5408df9e0004622fde802c5abe8ac08d6

commit 5973a8b5408df9e0004622fde802c5abe8ac08d6
Author: jmadill <jmadill@chromium.org>
Date: Tue May 17 20:42:08 2016

webgl: Try changing cube_combinations expectations to Fail.

This test is still failing on the GPU FYI Windows x64 bots. Not sure
why they aren't picking up the Flaky expectation.

BUG= 607393 
TBR=cwallez@chromium.org, kbr@chromium.org, zmo@chromium.org
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel

Review-Url: https://codereview.chromium.org/1990433002
Cr-Commit-Position: refs/heads/master@{#394221}

[modify] https://crrev.com/5973a8b5408df9e0004622fde802c5abe8ac08d6/content/test/gpu/gpu_tests/webgl2_conformance_expectations.py

Comment 13 by kbr@chromium.org, May 17 2016

Sorry about that. That's distressing. As an aside, the harness here is being rewritten so it should be a lot simpler in the future to figure out why flaky expectations might not be taking effect.

Mergedinto: 606021
Status: Duplicate (was: Assigned)
Seems to have cleared up with the fix in  issue 606021 .

Sign in to add a comment