1627816 - Picture cache does not invalidate when clipped content is hidden

Status: VERIFIED FIXED

(Core :: Graphics: WebRender, defect, P2)

Description

[Miko Mynttinen]

Attachment: mask-dynamic-visibility-1.html

The attached testcase dynamically changes the visibility of the masked content.
When picture caching is enabled, the hidden element remains on the screen until repaint is triggered.

Comment 1

[Miko Mynttinen]

Attachment: clip-dynamic-visibility-1.html

This also happens with simple clipping.

Comment 2

[Miko Mynttinen]

Regression range: https://hg.mozilla.org/integration/autoland/pushloghtml?fromchange=1330041fa8973ac2a56f074a7c1eee9bb1cff151&tochange=93ccc760c4dd69f63d6ac171ace357b0dd31412a
Regressed by bug 1599965 or bug 1599656, possibly both.

Comment 3

[Glenn Watson [:gw]]

Attachment: Bug 1627816 - Fix skipping composites when tile surface changes.

This fixes a case where the backing surface of a picture cache tile
changes, but was still being considered a no-op frame, which skips
the composite as an optimization if nothing has changed.

In this case, the tile surface changes from a rasterized texture
to a solid color that doesn't require a picture cache texture.

The patch ensures that the composite surface descriptors that are
used to detect no-op frame compositions include the backing
surface for each of the tiles that make up this virtual surface.

Comment 4

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/5b784c47235c
Fix skipping composites when tile surface changes. r=nical

Comment 5

[Narcis Beleuzu [:NarcisB]]

https://hg.mozilla.org/mozilla-central/rev/5b784c47235c

Comment 6

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

The patch landed in nightly and beta is affected.
:gw, is this bug important enough to require an uplift?
If not please set status_beta to wontfix.

For more information, please visit auto_nag documentation.

Comment 7

[Glenn Watson [:gw]]

I don't think it's likely to be hit in real web content, so we probably don't need to uplift this to beta.

Miko, do you think we need to uplift the fix for this to beta?

Comment 8

[Miko Mynttinen]

(In reply to Glenn Watson [:gw] from comment #7)

I don't think it's likely to be hit in real web content, so we probably don't need to uplift this to beta.

Miko, do you think we need to uplift the fix for this to beta?

I would expect the overflow clipping case to be somewhat common, but since most user interactions trigger a repaint, it probably won't be noticeable.

Comment 9

[Ryan VanderMeulen [:RyanVM]]

Sounds like a wontfix in that case.

Comment 10

[Alexandru Trif, Desktop QA [:atrif]]

Reproduced the issue using Firefox 77.0a1 (20200407214402) on Windows 10x64 and the attached test cases from comment 0 and comment 1.
The issue is verified fixed with Firefox 77.0b7 (20200515125749) on Windows 10x64.