1300972 - Crash in mozilla::layers::CanvasClientSharedSurface::UpdateRenderer

Status: RESOLVED WORKSFORME

(Core :: Graphics, defect)

Description

[Kan-Ru Chen [:kanru] (UTC+9)]

This bug was filed from the Socorro interface and is 
report bp-d1489a13-5e98-4e32-8c2a-2d3f82160906.
=============================================================

This is a low volume crash from a single installation. I suspect it's a regression introduced by bug 1284324.

ni Kip in case it's VR related.

Comment 1

[:kip (Kearwood Gilbert)]

(In reply to Kan-Ru Chen [:kanru] (UTC+8) from comment #0)
> This bug was filed from the Socorro interface and is 
> report bp-d1489a13-5e98-4e32-8c2a-2d3f82160906.
> =============================================================
> 
> This is a low volume crash from a single installation. I suspect it's a
> regression introduced by bug 1284324.
> 
> ni Kip in case it's VR related.
Thanks :kanru,

This would indeed affect only VR users.  I have reproduced this while tracking down Bug 1296446, which may be related.

I'll take this bug as well.

Comment 2

[:kip (Kearwood Gilbert)]

When running debug builds, I can see that the BlitFrameBuffer call in GLScreenBuffer::AssureBlitted() is failing.  I have not yet confirmed if this only happens only on the first frame, but allowing execution to continue results in correct rendering in the HMD and in the mirrored texture.

The callstack in this case matches the one in this report, but deeper below the UpdateRenderer function.  This difference may be due to optimization in non-debug builds.

Perhaps if I can fix the failure reproducible in the debug builds, this will correct the core cause of this crash.

Comment 3

[Daosheng Mu[:daoshengmu]]

(In reply to :kip (Kearwood Gilbert) from comment #2)
> When running debug builds, I can see that the BlitFrameBuffer call in
> GLScreenBuffer::AssureBlitted() is failing.  I have not yet confirmed if
> this only happens only on the first frame, but allowing execution to
> continue results in correct rendering in the HMD and in the mirrored texture.
> 
> The callstack in this case matches the one in this report, but deeper below
> the UpdateRenderer function.  This difference may be due to optimization in
> non-debug builds.
> 
> Perhaps if I can fix the failure reproducible in the debug builds, this will
> correct the core cause of this crash.

Kip, I was investigating this issue as well. Please take a look at Bug 1326376 comment 17, probably, it would bring some ideas.

Comment 4

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

Closing because no crashes reported for 12 weeks.