1061713 - crash in @0x0 | aticfx32.dll@0x10c99 or aticfx32.dll@0x10d09

Status: VERIFIED FIXED

(Core :: Graphics, defect)

Description

[Benoit Jacob [:bjacob] (mostly away)]

This bug was filed from the Socorro interface and is 
report bp-7c50c8b1-7ac8-4685-bd37-86d6f2140901.
=============================================================

This is the 38th topcrasher among sessions attempting D3D11 layers (prerelease channels on Windows).

Comment 1

[Benoit Jacob [:bjacob] (mostly away)]

Actually there are multiple signatures for the same crash --- putting it together, this is more like the 15th top crasher. Much higher still if we look at the population of ATI users.

Comment 2

[Milan Sreckovic [:milan] (needinfo for best results)]

Bas, do you think this is the "40% more D3D11 users" and we need to blocklist, or the recent changes?

Comment 3

[Bas Schouten (:bas.schouten)]

(In reply to Milan Sreckovic [:milan] from comment #2)
> Bas, do you think this is the "40% more D3D11 users" and we need to
> blocklist, or the recent changes?

If there's a corellation with older ATI driver versions, probably, but I don't know if that's true?

Comment 4

[Nicolas Silva [:nical]]

Attachment: (speculative fix) Don't crash if a texture is unlocked without being locked.

Hard to be sure, but this could be caused by allocation failing in the middle of some messy logic in ContentClient, the latter not properly handling it and trying to unlock a texture that wasn't locked/allocated. this patch makes sure we don't crash in this case in release builds (should definitely crash debug builds because it's bad to be in this state).

Comment 5

[Nicolas Silva [:nical]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/b5c3cd762cf2

Comment 6

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/b5c3cd762cf2

Comment 7

[[SV Manager] Florin Mezei, QA (:FlorinMezei)]

Socorro shows no more crashes with these signatures over the past 4 weeks.