angle end2end failures on Linux FYI GPU TSAN |
|
Issue descriptionsince https://ci.chromium.org/p/chromium/builders/luci.chromium.ci/Linux%20FYI%20GPU%20TSAN%20Release/19861: [ FAILED ] EGLProgramCacheControlTest.NegativeAPI/ES2_OPENGL, where GetParam() = ES2_OPENGL [ FAILED ] EGLQueryContextTest.NotInitialized/ES3_OPENGL, where GetParam() = ES3_OPENGL It times out since that build, but it also started failing those two tests on that build. The timeout may be related to the failures? And it shouldn't time out because it went from 1 to 4 shards on that same build.
,
Sep 4
Apparently a failure can make gtest start running tests one at a time, and since angle_end2end has like a 15 second start-up time, this leads to timing out. In the best case scenario we pay the 15 seconds once then run hundreds of tests. If we pay it for every test we timeout.
,
Sep 5
I've reproduced this in a standalone ANGLE ToT build with the following: angle_end2end_tests --gtest_filter=EGLDebugTest.CorrectCallbackParameters/ES2_OPENGL:EGLProgramCacheControlTest.NegativeAPI/ES2_OPENGL It seems EGLDebugTest.CorrectCallbackParameters leaves an error callback in the state that can cause following tests to fail. It's very sensitive to the order in which tests run which explains why it appeared after shards were added to this bot.
,
Sep 7
The following revision refers to this bug: https://chromium.googlesource.com/angle/angle/+/3b9680534909944560e4a12e0cce6be5943cc3e8 commit 3b9680534909944560e4a12e0cce6be5943cc3e8 Author: Frank Henigman <fjhenigman@chromium.org> Date: Fri Sep 07 18:01:50 2018 Clean up state in EGLDebugTest. Reset the EGL error callback for the benefit of following tests. BUG=chromium:880495 Change-Id: Idd94fe032b89b8a0fe829cb4656e4204d9d71b53 Reviewed-on: https://chromium-review.googlesource.com/1208310 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Frank Henigman <fjhenigman@chromium.org> [modify] https://crrev.com/3b9680534909944560e4a12e0cce6be5943cc3e8/src/tests/egl_tests/EGLDebugTest.cpp
,
Sep 7
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/239c4db338b061101707273c59ff84c9827cac94 commit 239c4db338b061101707273c59ff84c9827cac94 Author: angle-chromium-autoroll <angle-chromium-autoroll@skia-buildbots.google.com.iam.gserviceaccount.com> Date: Fri Sep 07 23:03:30 2018 Roll src/third_party/angle d84b6737fa10..3b9680534909 (4 commits) https://chromium.googlesource.com/angle/angle.git/+log/d84b6737fa10..3b9680534909 git log d84b6737fa10..3b9680534909 --date=short --no-merges --format='%ad %ae %s' 2018-09-07 fjhenigman@chromium.org Clean up state in EGLDebugTest. 2018-09-07 jmadill@chromium.org Vulkan: Make ContextVk driver uniforms a dirty bit. 2018-09-07 jmadill@chromium.org Vulkan: Correct gl_FragCoord with default FBO. 2018-09-07 fjhenigman@chromium.org Fix ClearTest.RGB8WithMaskFramebuffer. Created with: gclient setdep -r src/third_party/angle@3b9680534909 The AutoRoll server is located here: https://autoroll.skia.org/r/angle-chromium-autoroll Documentation for the AutoRoller is here: https://skia.googlesource.com/buildbot/+/master/autoroll/README.md If the roll is causing failures, please contact the current sheriff, who should be CC'd on the roll, and stop the roller if necessary. CQ_INCLUDE_TRYBOTS=luci.chromium.try:android_optional_gpu_tests_rel;luci.chromium.try:linux_optional_gpu_tests_rel;luci.chromium.try:mac_optional_gpu_tests_rel;luci.chromium.try:win_optional_gpu_tests_rel BUG=chromium:880495 TBR=fjhenigman@chromium.org Change-Id: Ibc35a52b39df10cc9d6cab6500eff19204dca7aa Reviewed-on: https://chromium-review.googlesource.com/1213826 Reviewed-by: angle-chromium-autoroll <angle-chromium-autoroll@skia-buildbots.google.com.iam.gserviceaccount.com> Commit-Queue: angle-chromium-autoroll <angle-chromium-autoroll@skia-buildbots.google.com.iam.gserviceaccount.com> Cr-Commit-Position: refs/heads/master@{#589705} [modify] https://crrev.com/239c4db338b061101707273c59ff84c9827cac94/DEPS |
|
►
Sign in to add a comment |
|
Comment 1 by fjhenigman@chromium.org
, Sep 4Owner: fjhenigman@chromium.org
Status: Started (was: Untriaged)