2038906 - Disable WebGL overlay when RenderDXGITextureHost::GetDCompositionTexture() is slow

Status: NEW

(Core :: Graphics: WebRender, enhancement)

Description

[Sotaro Ikeda [:sotaro]]

It is for preventing Bug 2036365.

Comment 1

[Sotaro Ikeda [:sotaro]]

Attachment: Bug 2038906 - Disable WebGL overlay when RenderDXGITextureHost::GetDCompositionTexture() is slow

Comment 2

[Pulsebot]

Pushed by [e-mail]:
https://github.com/mozilla-firefox/firefox/commit/aeb5cb535896
https://hg.mozilla.org/integration/autoland/rev/cf8b1b0f4d53
Disable WebGL overlay when RenderDXGITextureHost::GetDCompositionTexture() is slow r=gfx-reviewers,gw

Comment 3

[Jeff Muizelaar [:jrmuizel]]

Does Chrome have something similar?

Comment 4

[Sotaro Ikeda [:sotaro]]

(In reply to Jeff Muizelaar [:jrmuizel] from comment #3)

Does Chrome have something similar?

chromium allocates IDCompositionTexture in D3DImageBacking. It is triggered by D3DImageBackingFactory::CreateSharedImage().

chromium does not have a code to disable overlay like D299978. chromium seems not have a problem like Bug 2036365. Since D3DImageBacking seems always recycled.

Bug 2036655 is going to recycle DXGITextureHostD3D11 for WebGPU and WebGL.

Comment 5

[Cosmin Sabou [:CosminS]]

https://hg.mozilla.org/mozilla-central/rev/cf8b1b0f4d53

Comment 6

[Jeff Muizelaar [:jrmuizel]]

Is the intent to remove this code after bug 2036655 lands?

Comment 7

[Sotaro Ikeda [:sotaro]]

(In reply to Jeff Muizelaar [:jrmuizel] from comment #6)

Is the intent to remove this code after bug 2036655 lands?

:jrmuizel, do you think is it better to remove it? I thought to keep it.

Comment 8

[Jeff Muizelaar [:jrmuizel]]

My inclination would be to remove it. Changing behaviour based on timing can add nondeterminism that can make problems harder to debug. e.g. most people will be hitting the main path most of the time but sometimes it will be different.

Comment 9

[Sotaro Ikeda [:sotaro]]

OK, I am going to remove it after bug 2036655 lands.

Comment 10

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

The bug has a release status flag that shows some version of Firefox is affected, thus it will be considered confirmed.