Regression window: not a new crash, but the frequency changed sharply after the
2026-10-28 update.

The GTK build has not changed. libgtk-4-1 4.22.4+ds-0ubuntu0.1 was installed on
2026-06-16 and is still current, and the identical fault (segfault at 3c ...
in libgtk-4.so.1.2200.4[5a33b1,...]) appears in this machine's journal in July
(gnome-system-monitor), September (ptyxis x2, nautilus x2) and November
(ghostty x5). So the latent bug predates the recent cluster.

Crash frequency per boot, cross-referenced with the mutter and kernel that were
actually running:

  mutter 50.1-0ubuntu2.2 / kernel 7.0.0-31      36 resumes, 4 crashes
  mutter 50.1-0ubuntu2.4 / kernel 7.0.0-34/38   84 resumes, 0 crashes
  mutter 50.1-0ubuntu2.5 / kernel 7.0.0-42       6 resumes, 5 crashes

Package history:

  2026-06-16 07:27:47  upgrade libgtk-4-1   4.22.2+ds-1ubuntu1 -> 
4.22.4+ds-0ubuntu0.1
  2026-10-28 04:07:04  upgrade libmutter-18-0  50.1-0ubuntu2.4 -> 
50.1-0ubuntu2.5
  2026-10-28 04:07:07  install linux-image-7.0.0-42-generic

Both changes landed on 2026-10-28 and first took effect at the 2026-10-31
reboot. Everything before that reboot, including 84 consecutive resumes, was
crash-free.

mutter 50.1-0ubuntu2.5 changelog:

  - Fix alt-tab with sloppy/mouse focus-mode. When using mouse/sloppy focus,
    the focus should stick to a window not under the pointer when activating
    using e.g. alt-tab. (LP: #2152847)
  - Fix stylus input not working after lifting the stylus in fullscreen
    (LP: #2169577)
  - Fix drag-and-drop across windows with stylus input (LP: #2169504)
  - Fix moving maximized window to another monitor by keyboard shortcuts
    (LP: #2147214)

Three of the four touch seat focus handling and tablet/stylus input, which is
precisely the code that decides which client is focused, and therefore whether
seat->keyboard_focus is populated, at the moment the tablet pad re-registers on
resume.

The non-suspend crash on 2026-11-02 08:36:41 also falls inside this window: the
USB hub the tablet is attached to re-enumerated and the pad re-registered while
the session was live.

Proposed experiment to separate the two variables: kernel 7.0.0-38 is still
installed alongside 7.0.0-42. Booting 7.0.0-38 with mutter 2.5 and suspending
several times would show whether this is the kernel's resume timing or mutter's
focus handling. At 5 crashes in 6 resumes a handful of cycles should be
conclusive.

Caveat: this is correlation from a single machine, not proof. The 2.2 row shows
the window was reachable before the 2026-10-28 update as well; what changed is
how often it is hit.

-- 
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/2172827

Title:
  gtk4 4.22.4: SIGSEGV in gdk_surface_handle_event() on a tablet-pad
  event carrying no surface (GDK_PAD_GROUP_MODE) after USB re-
  enumeration

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/gtk4/+bug/2172827/+subscriptions


-- 
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to