1521930 - Crash in vcruntime140.dll | mozilla::gfx::VRSystemManagerExternal::PullState

Status: RESOLVED FIXED

(Core :: WebVR, defect)

Description

[Marcia Knous [:marcia]]

This bug is for crash report bp-5bfb5157-47d5-4649-9a4a-3eff80190115.

Seen while looking at crash stats - crashes started using 20190115103851: https://bit.ly/2U7NleU

Possible regression range based on Build ID: https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=edca8877b0505cd1c31beaf6d907ca32e022aa52&tochange=2bf2c209f520c75405b70ed5a5a12c2a8938dffe

Possibly Bug 1516554? ni on :daoshengmu

Top 10 frames of crashing thread:

0 vcruntime140.dll vcruntime140.dll@0xca27 
1 xul.dll bool mozilla::gfx::VRSystemManagerExternal::PullState gfx/vr/gfxVRExternal.cpp:802
2 xul.dll mozilla::gfx::impl::VRDisplayExternal::SubmitFrame gfx/vr/gfxVRExternal.cpp:276
3 xul.dll void mozilla::gfx::VRDisplayHost::SubmitFrameInternal gfx/vr/VRDisplayHost.cpp:288
4 xul.dll nsresult mozilla::detail::RunnableMethodImpl<mozilla::gfx::VRDisplayHost*, void  xpcom/threads/nsThreadUtils.h:1158
5 xul.dll nsThread::ProcessNextEvent xpcom/threads/nsThread.cpp:1167
6 xul.dll NS_ProcessNextEvent xpcom/threads/nsThreadUtils.cpp:468
7 xul.dll mozilla::ipc::MessagePumpForNonMainThreads::Run ipc/glue/MessagePump.cpp:303
8 xul.dll MessageLoop::RunHandler ipc/chromium/src/base/message_loop.cc:308
9 xul.dll static void nsThread::ThreadFunc xpcom/threads/nsThread.cpp:450

Comment 1

[Daosheng Mu[:daoshengmu]]

It can be reproduced by Bug 1521183.

Comment 2

[Andrew Overholt [:overholt]]

I was told we're aiming for bug 1476092 to ride with 67 so updating the flags here.

Comment 3

[Pascal Chevrel:pascalc]

Kip, given comment #2, can we get this bug prioritized? Thanks

Comment 4

[:kip (Kearwood Gilbert)]

I'll take and investigate this now.

Comment 5

[Pascal Chevrel:pascalc]

Kip, did you make any progress on this bug? Thanks

Comment 6

[:kip (Kearwood Gilbert)]

I'm spending today on stability fixes + secure bugs. This is at the top of my queue. I'll follow up shortly.

Comment 7

[:kip (Kearwood Gilbert)]

I have reproduced this call stack by killing the VR process. This suggests that Daosheng's fix for Bug 1523923 may also fix this issue.

Comment 8

[:kip (Kearwood Gilbert)]

I was able to verify Daosheng's fix for Bug 1523923 with the test case I used to replicate this bug. I was able to kill the VR process and recover, requiring only to refresh the WebVR page to start a new VR process and session.

Comment 9

[Neha Kochar [:neha]]

Kearwood, now that Bug 1523923 is resolved, does the fix suffice for this bug too?

Comment 10

[:kip (Kearwood Gilbert)]

(In reply to Neha Kochar [:neha] from comment #9)

Kearwood, now that Bug 1523923 is resolved, does the fix suffice for this bug too?

Yes, I expect it to correct this issue as well.