1737727 - SW WR/Gnome X11/Nvidia: Back button enabled but appears disabled

Status: RESOLVED WORKSFORME

(Core :: Graphics: WebRender, defect)

Description

[Randell Jesup [:jesup] (needinfo me)]

I've seen this a few times: when I navigated to a google search result, the back button appeared disabled, but was actually enabled (and highlighted if you moved the mouse over it). Opening a new tab and returning to the tab with the issue caused it to be painted correctly.

Comment 1

[:Gijs (he/him)]

Because of how that stuff is implemented (attributes + CSS) this smells like a graphics bug to me; the CSS that makes it look disabled is here: https://searchfox.org/mozilla-central/rev/3b86063e6d46b2f130513c499343cd47773062b1/browser/themes/shared/toolbarbuttons.inc.css#48-53 and the CSS that makes the :hover feedback work is here: https://searchfox.org/mozilla-central/rev/3b86063e6d46b2f130513c499343cd47773062b1/browser/themes/shared/toolbarbuttons.inc.css#167-176 and both rely on the disabled attribute, which is either there or not. If it's there, you should get disabled styling and no hover, if it isn't, the inverse of that. So if you're getting any other combination, graphics is doing something weird.

Comment 2

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

The severity field is not set for this bug.
:jimm, could you have a look please?

For more information, please visit auto_nag documentation.

Comment 3

[Sotaro Ikeda [:sotaro]]

:gw, do you have any ideas what could cause the problem?

Comment 4

[Glenn Watson [:gw]]

No ideas from that info, sorry. I'd probably need to be able to reproduce it to diagnose what's going on.

I guess this is running on Linux, is that right?

Robert, are there any partial present preferences that could be changed to see if it might be related to something like that?

Comment 5

[Robert Mader [:rmader]]

Randell, can you post the content of your about:support here so we have an idea what kind of setup we're dealing with?

Comment 6

[Randell Jesup [:jesup] (needinfo me)]

Attachment: support

Comment 7

[Glenn Watson [:gw]]

It's interesting that you're getting software webrender.

Lee, any ideas what might cause this? Do we do anything like partial present in the config that Randell has?

Jeff, do you know why this set up is getting sw-wr, should it be getting hw-wr?

Comment 8

[Glenn Watson [:gw]]

Maybe the sw-wr is related to the Driver Vendor: nvidia/unknown field?

Comment 9

[Lee Salzman [:lsalzman]]

Seems like he is using X11_EGL. so maybe this is bug 1743051 or bug 1712969? See the suggestions from Darkspirit in bug 1712969 comment 14 about toggling prefs, which are:

gfx.webrender.allow-partial-present-buffer-age = false
gfx.webrender.max-partial-present-rects = 0

Comment 10

[Jeff Muizelaar [:jrmuizel]]

Randal, do you have layers.acceleration.disabled set to true?

Comment 11

[Darkspirit]

(In reply to Randell Jesup [:jesup] (needinfo me) from comment #6)

Compositing: WebRender (Software)

layers.acceleration.disabled: true

Does this bug really occur with this configuration (Software WebRender)
or do you have disabled hardware rendering to fix this bug or for other reasons?

Comment 12

[Randell Jesup [:jesup] (needinfo me)]

(In reply to Jeff Muizelaar [:jrmuizel] from comment #10)

Randal, do you have layers.acceleration.disabled set to true?

Yes

I don't remember when or why I set that; it wasn't anytime recently. It may have been a long time ago.

Comment 13

[Darkspirit]

I should check if bug 1743051 also occurs with SW WR. (Gtk, Nvidia X server driver)

How reproducible is this bug for you? Does it occur often?

Please set these prefs to check if they prevent the problem from occurring:

(Lee Salzman [:lsalzman] from comment #9)

gfx.webrender.allow-partial-present-buffer-age = false
gfx.webrender.max-partial-present-rects = 0

Comment 14

[Robert Mader [:rmader]]

Hi! Does the issue here still happen to you or can we close the bug? Thanks!