MemoryPeakDetectorTest.StartStopQuickly sometimes fails on Fuchsia |
|||||
Issue description[00011.809] 02443.02591> [ RUN ] MemoryPeakDetectorTest.StartStopQuickly [00011.809] 02443.02591> ../../base/trace_event/memory_peak_detector_unittest.cc:377: Failure [00011.809] 02443.02591> Expected: (time_ms) >= (kNumPolls * kConfigNoCallbacks.polling_interval_ms), actual: 19.987 vs 20 [00011.809] 02443.02591> [ FAILED ] MemoryPeakDetectorTest.StartStopQuickly (21 ms) Other tests (e.g. bug 734130 and bug 734216 ) seem to suffer from the scheduler, so this may be racy.
,
Jun 19 2017
https://chromium-review.googlesource.com/540081 makes this fail reliably on Linux. So probably similar problem as described in https://bugs.chromium.org/p/chromium/issues/detail?id=734216#c2.
,
Jun 21 2017
As "thanks" for bug 734216 don't suppose this looks the same to you, does it? :) (Mostly based on the similar test name.)
,
Jun 21 2017
I was just looking at this ;) The issue seems just the same to me (as you mentioned in #2) but since PollFastMemoryTotal doesn't take a input pram we can't play the same trick as we did to fix bug 734216 :( One option is to accept that we may or may not get the calls from the for loop an relax the constraints on the test, something like: https://chromium-review.googlesource.com/c/543043/1/base/trace_event/memory_peak_detector_unittest.cc I'll show that to primiano and see how they feel, maybe there is a clever way I'm missing.
,
Jun 22 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/09c033a460c8d09ccb0d470a10d6e54256f85feb commit 09c033a460c8d09ccb0d470a10d6e54256f85feb Author: Hector Dearman <hjd@google.com> Date: Thu Jun 22 11:29:50 2017 memory-infra: Fix race in MemoryPeakDetectorTest.StartStopQuickly Bug: 734218 Change-Id: Id3c304265f42ffe694f240d8ea3754e76bbfe4b8 Reviewed-on: https://chromium-review.googlesource.com/544818 Commit-Queue: Primiano Tucci <primiano@chromium.org> Reviewed-by: Primiano Tucci <primiano@chromium.org> Cr-Commit-Position: refs/heads/master@{#481505} [modify] https://crrev.com/09c033a460c8d09ccb0d470a10d6e54256f85feb/base/trace_event/memory_peak_detector_unittest.cc
,
Jun 22 2017
Tried a slightly different fix that primiano suggested. I tested this on top of https://chromium-review.googlesource.com/540081 and it seems to fix it. I'm going to close but please reopen if you see more flakes. :)
,
Jun 22 2017
Thanks :)
,
Jun 22 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/fa8116176f4c3fd9a86083a2b08147fdc749b078 commit fa8116176f4c3fd9a86083a2b08147fdc749b078 Author: Scott Graham <scottmg@chromium.org> Date: Thu Jun 22 16:44:36 2017 fuchsia: remove suppression of MemoryPeakDetectorTest.StartStopQuickly Fixed by https://chromium-review.googlesource.com/544818. Bug: 734218 , 706592 Change-Id: I3359266830902964bd884293e5f418529c49c318 Reviewed-on: https://chromium-review.googlesource.com/544518 Reviewed-by: Nico Weber <thakis@chromium.org> Commit-Queue: Scott Graham <scottmg@chromium.org> Cr-Commit-Position: refs/heads/master@{#481562} [modify] https://crrev.com/fa8116176f4c3fd9a86083a2b08147fdc749b078/testing/buildbot/filters/fuchsia.base_unittests.filter
,
Jun 29 2017
|
|||||
►
Sign in to add a comment |
|||||
Comment 1 by scottmg@chromium.org
, Jun 16 2017