1306689 - Firefox49 showing graphics artifacts with HWA enabled

Status: RESOLVED WONTFIX

(Core :: Graphics, defect, P1)

Description

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

+++ This bug was initially created as a clone of Bug #1304360 +++

We didn't really fix bug 1304360 - we just disabled acceleration.  There are indications that we're running into (our?) bug with D3D9 (compositor), and we should follow up.

See the original bug, but we are hitting "CompositorD3D9::CreateRenderTargetFromSource: Failed to update texture" error in CompositorD3D9::CreateTexture with D3DERR_INVALIDCALL error from StretchRect().

There may be more information we should be able to get with a diagnostic patch in CompositorD3D9::CreateTexture.  Let's get a build and ask the reporter in bug 1304360 to run it and see if we can get more info.  I imagine parameters to StretchRect may be useful.

Comment 1

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

George, let's get the build and see if we can reproduce locally.

Comment 2

[Justin [:JW_SoftvisionQA]]

setting flags per comment 0

Comment 3

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

This got backed out of 50 (pending 49 uplift) in bug 1306465, so we have to figure it out for 51 or back it out of there as well.

Comment 4

[Liz Henry (:lizzard) (relman/hg->git project)]

George, any luck here? We still have a week and a half till 51 moves to beta. 
Milan, do want to leave this on for beta 51?

Comment 5

[George Wright (:gw280) (needinfo me!)]

It looks like we're putting this in nightly and 51. I'll prep the patches now.

See https://bugzilla.mozilla.org/show_bug.cgi?id=1306465#c26

Comment 6

[George Wright (:gw280) (needinfo me!)]

Sorry, by "this" i mean the patch to remove fallback from bug 1306465

Comment 7

[Ryan VanderMeulen [:RyanVM]]

Is there still something to do here (especially since we dropped support for XP, which was our only D3D9 consumer)?

Comment 8

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

Nothing to see here, keep moving along... :)  So, yes, we're not going to bother chasing down the potential issue and destabilizing other code, and make that change in a 52 ESR, so we're just going to close this bug as wontfix.