DCHECK fails in picture_layer_impl.cc |
|||||
Issue descriptionIn testing a feature I've been loading various sites from httparchive and this one hits a DCHECK. Wondered if it was useful to debug: http://www.ussailing.org/ [1:5:0707/104219.488231:FATAL:picture_layer_impl.cc(1245)] Check failed: draw_transform.IsScaleOrTranslation(). #0 0x7ffa7aa256c7 base::debug::StackTrace::StackTrace() #1 0x7ffa7aa4c9f1 logging::LogMessage::~LogMessage() #2 0x7ffa772196c7 cc::PictureLayerImpl::CalculateRasterTranslation() #3 0x7ffa772175fd cc::PictureLayerImpl::AddTilingsForRasterScale() #4 0x7ffa772163de cc::PictureLayerImpl::UpdateTiles() #5 0x7ffa7731251f cc::LayerTreeImpl::UpdateDrawProperties() #6 0x7ffa772f95bb cc::LayerTreeHostImpl::UpdateSyncTreeAfterCommitOrImplSideInvalidation() #7 0x7ffa772f935d cc::LayerTreeHostImpl::CommitComplete() #8 0x7ffa77337d52 cc::ProxyImpl::ScheduledActionCommit() #9 0x7ffa772a53f9 cc::Scheduler::ProcessScheduledActions() #10 0x7ffa772a5d09 cc::Scheduler::NotifyReadyToCommit() #11 0x7ffa7733491c cc::ProxyImpl::NotifyReadyToCommitOnImpl() #12 0x7ffa7aa11f85 _ZNO4base8CallbackIFvvELNS_8internal8CopyModeE0ELNS2_10RepeatModeE0EE3RunEv #13 0x7ffa7aa25f0b base::debug::TaskAnnotator::RunTask() #14 0x7ffa726f4c73 blink::scheduler::TaskQueueManager::ProcessTaskFromWorkQueue() #15 0x7ffa726f22c1 blink::scheduler::TaskQueueManager::DoWork() #16 0x7ffa726f71f2 _ZN4base8internal7InvokerINS0_9BindStateIMN5blink9scheduler16TaskQueueManagerEFvbEJNS_7WeakPtrIS5_EEbEEEFvvEE3RunEPNS0_13BindStateBaseE #17 0x7ffa7aa11f85 _ZNO4base8CallbackIFvvELNS_8internal8CopyModeE0ELNS2_10RepeatModeE0EE3RunEv #18 0x7ffa7aa25f0b base::debug::TaskAnnotator::RunTask() #19 0x7ffa7aa5a00d base::MessageLoop::RunTask() #20 0x7ffa7aa5a352 base::MessageLoop::DeferOrRunPendingTask() #21 0x7ffa7aa5a73f base::MessageLoop::DoWork() #22 0x7ffa7aa5c1d9 base::MessagePumpDefault::Run() #23 0x7ffa7aa59bcf base::MessageLoop::Run() #24 0x7ffa7aa8e3b7 base::RunLoop::Run() #25 0x7ffa7aacb3ec base::Thread::Run() #26 0x7ffa7aacb98f base::Thread::ThreadMain() #27 0x7ffa7aac251c base::(anonymous namespace)::ThreadFunc() #28 0x7ffa7aba1184 start_thread #29 0x7ffa6faebffd clone
,
Jul 7 2017
Another site: http://www.successtrading.ru/
,
Jul 7 2017
,
Jul 21 2017
,
Sep 14
I'm leaving the team, thus re-assigning. This is very similar to issue 778440. We checked that there is potential animation to screen matrix, and there is no rotation/perspective component in the matrix before enabling transformed rasterization. Nevertheless we ended up with a rotated matrix on impl-thread. This DCHECK itself should be harmless, but it scratches because it implies something unknown to us is happening in cc.
,
Sep 17
Unfortu
,
Sep 17
nately, neither of these pages repro. I'm going to remove the DCHECK for now as I don't think anybody is going to have time to investigate this low priority mystery. It's too bad it didn't get looked at earlier.
,
Sep 19
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/8f7b4562043922f6235ef3c1662adacac4425ccb commit 8f7b4562043922f6235ef3c1662adacac4425ccb Author: Adrienne Walker <enne@chromium.org> Date: Wed Sep 19 18:17:51 2018 cc: Remove DCHECKs from PLI::CalculateRasterTranslation These DCHECKs are sometimes being hit in practice, which says that they should not be DCHECKs. Because these are low priority mysteries, remove the DCHECKs and just add an early out to avoid performance cliffs. Bug: 778440, 740113 Cq-Include-Trybots: luci.chromium.try:android_optional_gpu_tests_rel;master.tryserver.blink:linux_trusty_blink_rel Change-Id: I36d1af371127d4d29be608e1351d53a2b547b25c Reviewed-on: https://chromium-review.googlesource.com/1228981 Reviewed-by: danakj <danakj@chromium.org> Commit-Queue: enne <enne@chromium.org> Cr-Commit-Position: refs/heads/master@{#592480} [modify] https://crrev.com/8f7b4562043922f6235ef3c1662adacac4425ccb/cc/layers/picture_layer_impl.cc |
|||||
►
Sign in to add a comment |
|||||
Comment 1 by dtapu...@chromium.org
, Jul 7 2017