1338639 - Crash in mozilla::layers::SyncObjectD3D11::Init

Status: RESOLVED FIXED

(Core :: Graphics: Layers, defect, P3)

Description

[Andrew McCreight [:mccr8] (PTO until 12-2)]

This bug was filed from the Socorro interface and is 
report bp-7e461dfa-d513-46f4-9e16-1c8bc2170210.
=============================================================

Bug 1322741 landed a patch for this issue, but I'm still seeing crashes with this signature on recent builds of Aurora and Nightly.

Comment 1

[Andrew McCreight [:mccr8] (PTO until 12-2)]

Can you please take a look Vincent? Thanks. Feel free to dupe this to the other bug and reopen that one if you want. I don't know if this is a separate issue or what.

Comment 2

[Peter Chang[:pchang]]

There was a gfx critical note listed in the following, I believe we failed to create D3D11 texture and caused crash.

[C0][GFX1-]: Failed to OpenSharedResource for SyncObjectD3D11: 0x80070057 (t=14.7052) 

The 0x80070057(E_INVALIDARG) means one or more arguments are not valid'.

Comment 3

[Nicholas Nethercote [inactive]]

22 crashes with this signature in Aurora 20170210004018.

Comment 4

[Vincent Liu[:vliu]]

It seems this issue is unlikely the root cause on Bug 1322741, which was SyncObjectD3D11 missing the ID3D11Device. I saw some discussions on Bug 1160157 Comment 110 might similar to this one. In that discussion, there was link got the same error return number. Please see [1] I posted from Bug 1160157 Comment 110.

[1]: https://crash-stats.mozilla.com/report/index/aaa9e577-4ddb-40b3-84fc-5f5c82170202#tab-metadata 


[C8][GFX1]: Failed to D3D11 OpenSharedResource for frame finalization: 0x80070057 (t=863.299)

Comment 5

[Kevin Chen[:kechen] (UTC + 8)]

Attachment: Handle the OpenSharedResource failure when device rest happened.

Before Bug 1319557, we had an early return to handling failure of OpenSharedResource and preventing accessing null mD3D11Texture when device reset happened[1]. I think we should add it back.

David, could you help me to review it ? Thank you.

[1] https://dxr.mozilla.org/mozilla-release/rev/c8c53bdd499ba5c6d08960bf003e648f4b2efd77/gfx/layers/d3d11/TextureD3D11.cpp#1201

Comment 6

[Kevin Chen[:kechen] (UTC + 8)]

Try result looks good.
https://treeherder.mozilla.org/#/jobs?repo=try&revision=729c5ffb420fae7794d9cc29f949530173fc8579

Comment 7

[Kevin Chen[:kechen] (UTC + 8)]

I will uplift this to Firefox 53 after the change is landed into Nightly.

Comment 8

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/422bcd4dddd7
Handle the OpenSharedResource failure in SyncObjectD3D11::Init when device rest happened. r=dvander

Comment 9

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

https://hg.mozilla.org/mozilla-central/rev/422bcd4dddd7

Comment 10

[Ryan VanderMeulen [:RyanVM]]

Please request Aurora approval on this when you get a chance.

Comment 11

[Kevin Chen[:kechen] (UTC + 8)]

Comment on attachment 8837077 [details] [diff] [review]
Handle the OpenSharedResource failure when device rest happened.

Approval Request Comment
[Feature/Bug causing the regression]:
 Bug 1319557
[User impact if declined]:
 The crash will still happens in aurora
[Is this code covered by automated tests?]:
 No.
[Has the fix been verified in Nightly?]:
 Yes.
[Needs manual test from QE? If yes, steps to reproduce]: 
 No.
[List of other uplifts needed for the feature/fix]:
 None.
[Is the change risky?]:
 Low risk.
[Why is the change risky/not risky?]:
 A simple code flow change and the return condition exists in the versions before 53.
[String changes made/needed]:
 None.

Comment 12

[Gerry Chang [:gchang]]

Comment on attachment 8837077 [details] [diff] [review]
Handle the OpenSharedResource failure when device rest happened.

Fix a crash. Aurora53+.

Comment 13

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/releases/mozilla-aurora/rev/9599f3612a60

Comment 14

[Andrei Vaida [:avaida]]

Setting qe-verify- based on Kevin's assessment on manual testing needs (see Comment 11).