New issue
Advanced search Search tips

Issue 831378 link

Starred by 1 user

Issue metadata

Status: Available
Owner: ----
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 2
Type: Bug



Sign in to add a comment

site_per_process_webkit_layout_tests and webkit_layout_tests failed intermittently on linux_chromium_rel_ng for the same patchset

Project Member Reported by oksamyt@chromium.org, Apr 10 2018

Issue description

site_per_process_webkit_layout_tests and webkit_layout_tests failed intermittently on linux_chromium_rel_ng for the same patchset:

https://ci.chromium.org/p/chromium/builders/luci.chromium.try/linux_chromium_rel_ng/66357 - 1 unexpected failure for site_per_process_webkit_layout_tests with patch, 4 unexpected failures without patch
https://ci.chromium.org/p/chromium/builders/luci.chromium.try/linux_chromium_rel_ng/66464 - 7 unexpected failures for webkit_layout_tests with patch
https://ci.chromium.org/p/chromium/builders/luci.chromium.try/linux_chromium_rel_ng/66630 - all tests passed


 
Description: Show this description
Components: UI>Input>VirtualKeyboard Blink>Compositing
Summary: site_per_process_webkit_layout_tests and webkit_layout_tests failed intermittently on linux_chromium_rel_ng for the same patchset (was: images/image-hover-display-alt.html flaking CQ tryjobs)

Comment 4 by enne@chromium.org, Apr 11 2018

Cc: schenney@chromium.org vmp...@chromium.org
Labels: -Pri-3 Pri-2
Maybe a dupe of  issue 828602 , although the failures there are not the same compositing-related tests and that bug only lists site per process and not both.  So it might be something different.

I'm seeing very similar failures in https://chromium-review.googlesource.com/c/chromium/src/+/998682 tryjobs.  Similar tests are failing without patch.

Comment 5 by enne@chromium.org, Apr 11 2018

Cc: enne@chromium.org
Status: Available

Comment 7 by enne@chromium.org, Apr 11 2018

I looked through a bunch of these, and this has been happening for quite some time.  Here's more examples (non-exhaustive):
https://ci.chromium.org/p/chromium/builders/luci.chromium.ci/Linux%20Tests/69016
https://ci.chromium.org/p/chromium/builders/luci.chromium.ci/Linux%20Tests/68965
https://ci.chromium.org/p/chromium/builders/luci.chromium.ci/Linux%20Tests/68959
https://ci.chromium.org/p/chromium/builders/luci.chromium.ci/Linux%20Tests/68944
https://ci.chromium.org/p/chromium/builders/luci.chromium.ci/Linux%20Tests/68941
https://ci.chromium.org/p/chromium/builders/luci.chromium.ci/Linux%20Tests/68940
https://ci.chromium.org/p/chromium/builders/luci.chromium.ci/Linux%20Tests/68939
https://ci.chromium.org/p/chromium/builders/luci.chromium.ci/Linux%20Tests/68911
https://ci.chromium.org/p/chromium/builders/luci.chromium.ci/Linux%20Tests/68910
https://ci.chromium.org/p/chromium/builders/luci.chromium.ci/Linux%20Tests/68907
https://ci.chromium.org/p/chromium/builders/luci.chromium.ci/Linux%20Tests/68889
https://ci.chromium.org/p/chromium/builders/luci.chromium.ci/Linux%20Tests/68867
https://ci.chromium.org/p/chromium/builders/luci.chromium.ci/Linux%20Tests/68856
https://ci.chromium.org/p/chromium/builders/luci.chromium.ci/Linux%20Tests/68855
https://ci.chromium.org/p/chromium/builders/luci.chromium.ci/Linux%20Tests/68854
https://ci.chromium.org/p/chromium/builders/luci.chromium.ci/Linux%20Tests/68852
https://ci.chromium.org/p/chromium/builders/luci.chromium.ci/Linux%20Tests/68851
https://ci.chromium.org/p/chromium/builders/luci.chromium.ci/Linux%20Tests/68844
https://ci.chromium.org/p/chromium/builders/luci.chromium.ci/Linux%20Tests/68843
https://ci.chromium.org/p/chromium/builders/luci.chromium.ci/Linux%20Tests/68836
https://ci.chromium.org/p/chromium/builders/luci.chromium.ci/Linux%20Tests/68834
https://ci.chromium.org/p/chromium/builders/luci.chromium.ci/Linux%20Tests/68831
# maybe this one, although it's a 6s timeout vs the above that are also timing out at 30s.
https://ci.chromium.org/p/chromium/builders/luci.chromium.ci/Linux%20Tests/68712

And then logdog ran out of old logs.  It looks like there's just an intermittent set of what I'd consider "gpu heavy" tests (filters, webgl, etc) that time out.

This seems really bad and really frequent.  I am surprised that it's been going on for a week, but maybe because it's a slightly different set of tests every time, nobody has pinned down a root cause.
Another failing build: https://ci.chromium.org/p/chromium/builders/luci.chromium.try/linux_chromium_rel_ng/84965

Swarming logs: https://chromium-swarm.appspot.com/task?id=3d40d42984ca1810&refresh=10&show_raw=1

15:01:19.542 20887 worker/4 virtual/layout_ng_experimental/fast/pagination/auto-height.html crashed, (stderr lines):
15:01:19.543 20887   [1:1:0503/150117.991366:FATAL:ng_box_fragment_painter.cc(328)] Check failed: false. "Type: '2' Size: '0x64' Offset: '0,0' Placed: '1'"
15:01:19.543 20887   #0 0x0000037a6dac base::debug::StackTrace::StackTrace()
15:01:19.543 20887   #1 0x000003718d9b logging::LogMessage::~LogMessage()
15:01:19.543 20887   #2 0x000005fc358f blink::NGBoxFragmentPainter::PaintBlockChildren()
15:01:19.543 20887   #3 0x000005fc268e blink::NGBoxFragmentPainter::PaintObject()
15:01:19.543 20887   #4 0x000005fc14fd blink::NGBoxFragmentPainter::PaintWithAdjustedOffset()
15:01:19.543 20887   #5 0x000005fc131e blink::NGBoxFragmentPainter::Paint()
15:01:19.543 20887   #6 0x000005fc3484 blink::NGBoxFragmentPainter::PaintBlockChildren()
15:01:19.543 20887   #7 0x000005fc268e blink::NGBoxFragmentPainter::PaintObject()
15:01:19.543 20887   #8 0x000005fc14fd blink::NGBoxFragmentPainter::PaintWithAdjustedOffset()
15:01:19.543 20887   #9 0x000005fc131e blink::NGBoxFragmentPainter::Paint()
15:01:19.543 20887   #10 0x000005fc0909 blink::NGBlockFlowPainter::Paint()
15:01:19.543 20887   #11 0x000005e1c7d4 blink::LayoutNGMixin<>::Paint()
15:01:19.543 20887   #12 0x000005f6ca04 blink::BlockPainter::PaintChildren()
15:01:19.543 20887   #13 0x000005c82ca5 blink::LayoutBlock::PaintChildren()
15:01:19.543 20887   #14 0x000005f6dc8e blink::BlockPainter::PaintContents()
15:01:19.543 20887   #15 0x000005f6ed5a blink::BlockFlowPainter::PaintContents()
15:01:19.543 20887   #16 0x000005f6d9d4 blink::BlockPainter::PaintObject()
15:01:19.543 20887   #17 0x000005c82cc5 blink::LayoutBlock::PaintObject()
15:01:19.543 20887   #18 0x000005f6c051 blink::BlockPainter::Paint()
15:01:19.543 20887   #19 0x000005c82c85 blink::LayoutBlock::Paint()
15:01:19.543 20887   #20 0x000005f6ca04 blink::BlockPainter::PaintChildren()
15:01:19.543 20887   #21 0x000005c82ca5 blink::LayoutBlock::PaintChildren()
15:01:19.543 20887   #22 0x000005f6dc8e blink::BlockPainter::PaintContents()
15:01:19.543 20887   #23 0x000005f6ed5a blink::BlockFlowPainter::PaintContents()
15:01:19.543 20887   #24 0x000005f6d9d4 blink::BlockPainter::PaintObject()
15:01:19.543 20887   #25 0x000005c82cc5 blink::LayoutBlock::PaintObject()
15:01:19.543 20887   #26 0x000005f6c051 blink::BlockPainter::Paint()
15:01:19.543 20887   #27 0x000005c82c85 blink::LayoutBlock::Paint()
15:01:19.543 20887   #28 0x000005ff39a2 blink::PaintLayerPainter::PaintFragmentWithPhase()
15:01:19.543 20887   #29 0x000005ff3bcc blink::PaintLayerPainter::PaintForegroundForFragmentsWithPhase()
15:01:19.543 20887   #30 0x000005ff1061 blink::PaintLayerPainter::PaintForegroundForFragments()
15:01:19.543 20887   #31 0x000005feeb59 blink::PaintLayerPainter::PaintLayerContents()
15:01:19.556 20887   #32 0x000005fedd30 blink::PaintLayerPainter::PaintLayerContentsCompositingAllPhases()
15:01:19.556 20887   #33 0x000005fed616 blink::PaintLayerPainter::Paint()
15:01:19.556 20887   #34 0x000005ff0bb7 blink::PaintLayerPainter::PaintChildren()
15:01:19.556 20887   #35 0x000005feebab blink::PaintLayerPainter::PaintLayerContents()
15:01:19.556 20887   #36 0x000005f8b25b blink::CompositedLayerMapping::DoPaintTask()
15:01:19.556 20887   #37 0x000005f8c0ce blink::CompositedLayerMapping::PaintContents()
15:01:19.556 20887   #38 0x00000510b47b blink::GraphicsLayer::PaintWithoutCommit()
15:01:19.556 20887   #39 0x00000510ac27 blink::GraphicsLayer::Paint()
15:01:19.556 20887   #40 0x00000510a9a0 blink::GraphicsLayer::PaintRecursivelyInternal()
15:01:19.556 20887   #41 0x00000510aa6c blink::GraphicsLayer::PaintRecursivelyInternal()
15:01:19.556 20887   #42 0x00000510aa6c blink::GraphicsLayer::PaintRecursivelyInternal()
15:01:19.556 20887   #43 0x00000510a6d8 blink::GraphicsLayer::PaintRecursively()
15:01:19.556 20887   #44 0x0000058a5c77 blink::LocalFrameView::PaintTree()
15:01:19.556 20887   #45 0x0000058a33f0 blink::LocalFrameView::UpdateLifecyclePhasesInternal()
15:01:19.556 20887   #46 0x0000058a2e37 blink::LocalFrameView::UpdateAllLifecyclePhases()
15:01:19.556 20887   #47 0x000005f3609e blink::PageAnimator::UpdateAllLifecyclePhases()
15:01:19.556 20887   #48 0x0000057c4178 blink::WebViewImpl::UpdateLifecycle()
15:01:19.556 20887   #49 0x000006d087c8 blink::WebViewFrameWidget::UpdateLifecycle()
15:01:19.556 20887   #50 0x0000036dc447 content::BlinkTestRunner::TestFinished()
15:01:19.556 20887   #51 0x000007245274 test_runner::TestRunner::WorkQueue::ProcessWorkSoon()
15:01:19.556 20887   #52 0x000007247cc0 test_runner::TestRunner::LocationChangeDone()
15:01:19.556 20887   #53 0x000007247bfa test_runner::TestRunner::tryToClearTopLoadingFrame()
15:01:19.556 20887   #54 0x000005eda597 blink::ProgressTracker::ProgressCompleted()
15:01:19.556 20887   #55 0x000005ebf0b7 blink::FrameLoader::DidFinishNavigation()
15:01:19.556 20887   #56 0x000005566543 blink::Document::CheckCompleted()
15:01:19.556 20887   #57 0x000005ebed5c blink::FrameLoader::FinishedParsing()
15:01:19.557 20887   #58 0x000005572c8f blink::Document::FinishedParsing()
15:01:19.557 20887   #59 0x00000599d23a blink::HTMLConstructionSite::FinishedParsing()
15:01:19.557 20887   #60 0x0000059fd6b4 blink::HTMLTreeBuilder::Finished()
15:01:19.557 20887   #61 0x00000598a676 blink::HTMLDocumentParser::end()

Comment 9 by shend@chromium.org, Jun 27 2018

Components: -UI>Input>VirtualKeyboard
I don't think virtual keyboard affects anything in webkit_layout_tests.
If I apply this patch (which has been reverted)
https://chromium-review.googlesource.com/c/chromium/src/+/1100830
... and executes the test:
virtual/layout_ng_experimental/fast/multicol/paged-becomes-multicol-auto-height.html
... I can reproduce the stack trace in comment #8

The CL causes the navigation to start loading its data slightly earlier. Maybe it is enough to make thing to be scheduled in a different order.

Sign in to add a comment