1632755 - Crash in [@ nsRefreshDriver::Tick]

Status: RESOLVED FIXED

(Core :: Graphics: WebGPU, defect, P2)

Description

[Pascal Chevrel:pascalc]

This bug is for crash report bp-9c64052c-c782-4fd7-baea-f42460200424.

Top 10 frames of crashing thread:

0 xul.dll nsRefreshDriver::Tick layout/base/nsRefreshDriver.cpp:2264
1 xul.dll mozilla::RefreshDriverTimer::TickRefreshDrivers layout/base/nsRefreshDriver.cpp:351
2 xul.dll mozilla::RefreshDriverTimer::Tick layout/base/nsRefreshDriver.cpp:368
3 xul.dll mozilla::VsyncRefreshDriverTimer::RefreshDriverVsyncObserver::TickRefreshDriver layout/base/nsRefreshDriver.cpp:746
4 xul.dll mozilla::VsyncRefreshDriverTimer::RefreshDriverVsyncObserver::NotifyVsync layout/base/nsRefreshDriver.cpp:645
5 xul.dll mozilla::layout::VsyncChild::RecvNotify layout/ipc/VsyncChild.cpp:55
6 xul.dll mozilla::layout::PVsyncChild::OnMessageReceived ipc/ipdl/PVsyncChild.cpp:187
7 xul.dll mozilla::ipc::PBackgroundChild::OnMessageReceived ipc/ipdl/PBackgroundChild.cpp:5970
8 xul.dll mozilla::ipc::MessageChannel::DispatchMessage ipc/glue/MessageChannel.cpp:2110
9 xul.dll mozilla::ipc::MessageChannel::MessageTask::Run ipc/glue/MessageChannel.cpp:1989

We went from a couple of crashes a day to 50 crashes since yesterday. Most of the crashes are in build id 20200423145559 and some user comments in crash reports say that they tried the demos on https://hacks.mozilla.org/2020/04/experimental-webgpu-in-firefox/

Maybe this is not a regression but more attention given by yesterday's blog post to an existing crashing code path.

Comment 1

[Mats Palmgren (inactive)]

Probably better if GFX triage this if it's possibly related to webgpu in any way.

Comment 2

[Dzmitry Malyshau [:kvark]]

Yeah, I've seen this myself once. A bit confusing about how this relates to WebGPU, given that nothing in the call stack suggests that. Will look into this anyway.

Comment 3

[BugBot [:suhaib / :marco/ :calixte]]

Changing the priority to p2 as the bug is tracked by a release manager for the current nightly.
See What Do You Triage for more information

Comment 4

[Dzmitry Malyshau [:kvark]]

Attachment: Unlink CanvasContext from the refresh driver

Comment 5

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/7d6135dac2f2
Unlink CanvasContext from the refresh driver r=aosmond

Comment 6

[Oana Pop-Rus]

https://hg.mozilla.org/mozilla-central/rev/7d6135dac2f2

Comment 7

[Vlad Lucaci, Thunderbird QA (:vlucaci)]

Is there anything actionable by QA? if yes, could you please provide STR?
Thank you in advance!

Comment 8

[Pascal Chevrel:pascalc]

Nothing to do for QA, we don't have crashes on Beta with this signature.