amd64-generic-tot-chromium-pfq-informational failing since April 6 |
|||||||||
Issue descriptionThe amd64-generic-tot-chromium-pfq-informational has been failing since last Friday and I can't find an issue filed for it. The VMTests are all failing with a chrome crash. Here is a link to the first failure:https://pantheon.corp.google.com/storage/browser/chromeos-image-archive/amd64-generic-tot-chromium-pfq-informational/R67-10552.0.0-b2451231/vm_test_results_1/pfq_suite/test_harness/all/SimpleTestUpdateAndVerify/2_autotest_tests/results-01-login_OwnershipApi/login_OwnershipApi/sysinfo/var/log_diff/chrome/ [719:719:0406/134751.673092:FATAL:window_tree_host_manager.cc(234)] Check failed: display::kInvalidDisplayId != primary_display_id (-1 vs. -1) I was able to reproduce this locally following the simple chrome VM steps (thanks Achuith!): https://docs.google.com/document/d/1KIVpg0eXR1gEMZuV4Shm8dH1VV8ev4Ysi541AEephMU/edit# I am attempting a bisect now.
,
Apr 12 2018
Confirmed that #548850 does not have the FATAL crash and #548901 does. Continuing bisect.
,
Apr 12 2018
This is the cause of the failures: commit fc0f49d77d09b465611339be74f687ccffb0ec5e Author: Miguel Casas <mcasas@chromium.org> Date: Fri Apr 6 18:13:37 2018 +0000 Edid: make a class out of the edid parsing functions ... Bug: 821393 Reviewed-on: https://chromium-review.googlesource.com/970085 +mcasas@, +dnicoara@ Unfortunately reverting this will also require reverting dependent change https://chromium-review.googlesource.com/1001572
,
Apr 12 2018
I deployed an unstripped binary and got a stack: [15092:15092:0412/113326.726210:FATAL:window_tree_host_manager.cc(234)] Check failed: display::kInvalidDisplayId != primary_display_id (-1 vs. -1) #0 0x572d2fb88eac base::debug::StackTrace::StackTrace() #1 0x572d2fba0f80 logging::LogMessage::~LogMessage() #2 0x572d31bf404e ash::WindowTreeHostManager::CreatePrimaryHost() #3 0x572d31c1a6f0 ash::Shell::Init() #4 0x572d31c19fe0 ash::Shell::CreateInstance() #5 0x572d33624661 ash::WindowManager::CreateShell() #6 0x572d336248f8 ash::WindowManager::OnWmConnected() #7 0x572d2df6ba4d ui::mojom::WindowManagerStubDispatch::Accept() #8 0x572d3008c277 mojo::InterfaceEndpointClient::HandleValidatedMessage() #9 0x572d30090c4a mojo::internal::MultiplexRouter::ProcessIncomingMessage() #10 0x572d300903f8 mojo::internal::MultiplexRouter::Accept() #11 0x572d3008ae40 mojo::Connector::ReadSingleMessage() #12 0x572d3008ac9f mojo::Connector::WaitForIncomingMessage() #13 0x572d30ed29df aura::WindowTreeClient::WaitForInitialDisplays() #14 0x572d32147f2f AshShellInit::AshShellInit() #15 0x572d31f8a55d ChromeBrowserMainExtraPartsAsh::PreProfileInit() #16 0x572d2f892d4a ChromeBrowserMainParts::PreProfileInit() #17 0x572d2f89405b ChromeBrowserMainPartsLinux::PreProfileInit() #18 0x572d2ec3b2fc chromeos::ChromeBrowserMainPartsChromeos::PreProfileInit() #19 0x572d2f8920b4 ChromeBrowserMainParts::PreMainMessageLoopRunImpl() #20 0x572d2f891937 ChromeBrowserMainParts::PreMainMessageLoopRun() #21 0x572d2ec3ad90 chromeos::ChromeBrowserMainPartsChromeos::PreMainMessageLoopRun() #22 0x572d2e2d525e content::BrowserMainLoop::PreMainMessageLoopRun() #23 0x572d2e65ac37 content::StartupTaskRunner::RunAllTasksNow() #24 0x572d2e2d3c12 content::BrowserMainLoop::CreateStartupTasks() #25 0x572d2e2d7b00 content::BrowserMainRunnerImpl::Initialize() #26 0x572d2e2d15dc content::BrowserMain() #27 0x572d2f87f2c3 content::ContentMainRunnerImpl::Run() #28 0x572d2f888bd6 service_manager::Main() #29 0x572d2f87dea1 content::ContentMain() #30 0x572d2d8dab1c ChromeMain #31 0x7fc36cf81736 __libc_start_main #32 0x572d2d8da939 _start
,
Apr 12 2018
I guess the mentioned r548858 (code now: [1]) breaks the CHECK [2,3] int64_t display_id = ConnectorIndex(device_index, info->index()); and now it reads int64_t display_id = -1; Let me whip up a CL to fix it. [1] https://chromium-review.googlesource.com/c/chromium/src/+/970085/23/ui/ozone/platform/drm/common/drm_util.cc [2] https://cs.chromium.org/chromium/src/ash/display/window_tree_host_manager.cc?q=window_tree_host_manager&sq=package:chromium&dr=C&l=234 [3] https://cs.chromium.org/chromium/src/ash/display/window_tree_host_manager.cc?q=window_tree_host_manager&sq=package:chromium&dr=C&l=265
,
Apr 12 2018
BTW, this is also (unsurprisingly) failing in the PFQ: amd64-generic-chromium-pfq
,
Apr 12 2018
crrev.com/c/1010517 contains the fix. Thanks for staying vigilant and sorry for the noise!
,
Apr 12 2018
Thanks for the quick fix! Confirmed locally that the CL fixes this. Will close this once the informational builder picks up the fix.
,
Apr 12 2018
,
Apr 13 2018
Issue 831637 has been merged into this issue.
,
Apr 13 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/8f8853c2b80cc058545b60afd14f32b45dc69e9a commit 8f8853c2b80cc058545b60afd14f32b45dc69e9a Author: Miguel Casas <mcasas@chromium.org> Date: Fri Apr 13 03:40:12 2018 drm_util: correct initial value of |display_id| This CL corrects the initialization value of |display_id| from -1 (which is bad) to what it used to have before the CL reviewed in https://chromium-review.googlesource.com/c/chromium/src/+/970085/23/ui/ozone/platform/drm/common/drm_util.cc#b385 The -1 it breaks (at least) amd64-generic-tot-chromium-pfq-informational (see bug). TBR=dnicoara@chromium.org since it's a trivial fix and has been verified. Bug: 832136 Change-Id: I3c1db741979f20877483880b87f2d052b2a02cf9 Reviewed-on: https://chromium-review.googlesource.com/1010517 Commit-Queue: Miguel Casas <mcasas@chromium.org> Reviewed-by: Daniel Nicoara <dnicoara@chromium.org> Reviewed-by: Steven Bennetts <stevenjb@chromium.org> Cr-Commit-Position: refs/heads/master@{#550495} [modify] https://crrev.com/8f8853c2b80cc058545b60afd14f32b45dc69e9a/ui/ozone/platform/drm/common/drm_util.cc
,
Apr 13 2018
Issue 832010 has been merged into this issue.
,
Apr 13 2018
amd64-generic-tot-chromium-pfq-informational is green. amd64-generic-pfq was built at 3396.0 -> master@550002 so it does not have the fix (which appears taken a long time to go through the CQ, see issue 832355 for more info). Resolving as fixed since the informational builder is green at least.
,
Apr 14 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/bff9e2df3f45399aa207d9e380da1716e2435fe8 commit bff9e2df3f45399aa207d9e380da1716e2435fe8 Author: Alexander Alekseev <alemate@chromium.org> Date: Sat Apr 14 02:42:32 2018 drm_util: correct initial value of |display_id| This CL corrects the initialization value of |display_id| from -1 (which is bad) to what it used to have before the CL reviewed in https://chromium-review.googlesource.com/c/chromium/src/+/970085/23/ui/ozone/platform/drm/common/drm_util.cc#b385 The -1 it breaks (at least) amd64-generic-tot-chromium-pfq-informational (see bug). TBR=dnicoara@chromium.org since it's a trivial fix and has been verified., mcasas@chromium.org (cherry picked from commit 8f8853c2b80cc058545b60afd14f32b45dc69e9a) Bug: 832136 Change-Id: I3c1db741979f20877483880b87f2d052b2a02cf9 Reviewed-on: https://chromium-review.googlesource.com/1010517 Commit-Queue: Miguel Casas <mcasas@chromium.org> Reviewed-by: Daniel Nicoara <dnicoara@chromium.org> Reviewed-by: Steven Bennetts <stevenjb@chromium.org> Cr-Original-Commit-Position: refs/heads/master@{#550495} Reviewed-on: https://chromium-review.googlesource.com/1013386 Reviewed-by: Alexander Alekseev <alemate@chromium.org> Cr-Commit-Position: refs/branch-heads/3396@{#5} Cr-Branched-From: 9ef2aa869bc7bc0c089e255d698cca6e47d6b038-refs/heads/master@{#550428} [modify] https://crrev.com/bff9e2df3f45399aa207d9e380da1716e2435fe8/ui/ozone/platform/drm/common/drm_util.cc
,
Apr 17 2018
Issue 825425 has been merged into this issue.
,
Apr 17 2018
Issue 833563 has been merged into this issue.
,
Apr 17 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/8f8853c2b80cc058545b60afd14f32b45dc69e9a commit 8f8853c2b80cc058545b60afd14f32b45dc69e9a Author: Miguel Casas <mcasas@chromium.org> Date: Fri Apr 13 03:40:12 2018 drm_util: correct initial value of |display_id| This CL corrects the initialization value of |display_id| from -1 (which is bad) to what it used to have before the CL reviewed in https://chromium-review.googlesource.com/c/chromium/src/+/970085/23/ui/ozone/platform/drm/common/drm_util.cc#b385 The -1 it breaks (at least) amd64-generic-tot-chromium-pfq-informational (see bug). TBR=dnicoara@chromium.org since it's a trivial fix and has been verified. Bug: 832136 Change-Id: I3c1db741979f20877483880b87f2d052b2a02cf9 Reviewed-on: https://chromium-review.googlesource.com/1010517 Commit-Queue: Miguel Casas <mcasas@chromium.org> Reviewed-by: Daniel Nicoara <dnicoara@chromium.org> Reviewed-by: Steven Bennetts <stevenjb@chromium.org> Cr-Commit-Position: refs/heads/master@{#550495} [modify] https://crrev.com/8f8853c2b80cc058545b60afd14f32b45dc69e9a/ui/ozone/platform/drm/common/drm_util.cc |
|||||||||
►
Sign in to add a comment |
|||||||||
Comment 1 by steve...@chromium.org
, Apr 12 2018Bisect range from first failing builder is: master@{#548850} -> master@{#548901}