1828763 - mutter closes wayland connection due to communication error: "wl_surface@91: error 2: Buffer size (1173x128) must be an integer multiple of the buffer_scale (2)."

Status: RESOLVED DUPLICATE of bug 1803016

(Core :: Widget: Gtk, defect)

Description

[Jeroen van Leusen]

Attachment: MOZ_LOG="WidgetWayland:5,timestamp,sync"

User Agent: Mozilla/5.0 (X11; Linux x86_64; rv:109.0) Gecko/20100101 Firefox/112.0

Steps to reproduce:

Start firefox as a wayland session on gnome 44.
Just browse the web.

Actual results:

It randomly shuts down the the following output in the journald:

Apr 18 16:26:26 cetoddle .gnome-shell-wr[1526]: WL: error in client communication (pid 16220)
Apr 18 16:26:26 cetoddle .firefox-wrappe[16220]: Error reading events from display: Connection reset by peer

In addition I've ran into the bug again with MOZ_LOG="WidgetWayland:5,timestamp,sync" and this was the final line:

2023-04-18 20:51:26.651159 UTC - [Parent 77578: Renderer]: D/WidgetWayland WindowSurfaceWaylandMB::Commit [7fe696fd3800] flush failed

Full log is attached.

As a side note while trying to reproduce the error I noticed I had to only use a single firefox window, otherwise the log output would just end without anything unusual being logged.

Expected results:

I'd like it to not crash. This makes using private tabs quite difficult since all state is lost when this bug occurs.

Comment 1

[Jeroen van Leusen]

So I checked about:crashes on a hunch and I found the following crash was reported today:
https://crash-stats.mozilla.org/report/index/0cdc5fe2-0e6a-4386-ad26-053de0230418

It looks like the stated reason (resolution not a multiple of the buffer scale) also applies to the final lines of the attached moz_log output, so these crashes are probably happening for the same reason.

Comment 2

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

The Bugbug bot thinks this bug should belong to the 'Core::Graphics: WebRender' component, and is moving the bug to that component. Please correct in case you think the bot is wrong.

Comment 3

[Darkspirit]

Thanks for the report! Please check if this bug has been fixed in https://nightly.mozilla.org.
(bug 1820293 has been a new regression in 112.)

Comment 4

[Martin Stránský [:stransky] (ni? me)]

How-to test latest nightly:
https://fedoraproject.org/wiki/How_to_debug_Firefox_problems#Testing_Mozilla_binaries
Thanks.

Comment 5

[Jeroen van Leusen]

(In reply to Martin Stránský [:stransky] (ni? me) from comment #4)

How-to test latest nightly:
https://fedoraproject.org/wiki/How_to_debug_Firefox_problems#Testing_Mozilla_binaries
Thanks.

Used the nixpkgs-mozilla overlay to run firefox "114.0a1 (2023-04-18) (64-bit)"

No crashes after 5 hours of use, so it looks like it has been fixed.

Comment 6

[Martin Stránský [:stransky] (ni? me)]

Good, Thanks.