FS#79831 - [mesa] gnome-shell core dumps at startup on Intel gfx

Attached to Project: Arch Linux
Opened by Mark Blakeney (bulletmark) - Sunday, 01 October 2023, 23:17 GMT
Last edited by Buggy McBugFace (bugbot) - Saturday, 25 November 2023, 20:20 GMT
Task Type Bug Report
Category Packages: Extra
Status Closed
Assigned To Jan Alexander Steffens (heftig)
Laurent Carlier (lordheavy)
Felix Yan (felixonmars)
Architecture All
Severity Low
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 2
Private No

Details

Since the update:

2023-10-01 13:38:57 mesa 1:23.1.8-1 -> 1:23.2.1-1

gnome-shell core dumps at start up which causes gnome to disable all shell extensions. gnome-shell then is started again and runs ok but extensions have to enabled manually.

gnome-shell back-trace is:

(gdb) bt
#0 0x0000000000000000 in ()
#1 0x00007fc21a19e196 in end_query() () at ../mesa-23.2.1/src/mesa/main/queryobj.c:260
#2 0x00007fc23148f5e0 in cogl_gl_create_timestamp_query (context=0x559ad1fe75f0) at ../mutter/cogl/cogl/driver/gl/cogl-util-gl.c:520
#3 0x00007fc231507b41 in cogl_onscreen_egl_swap_buffers_with_damage (onscreen=0x559ad2082f50, rectangles=0x7fff6a1d8450, n_rectangles=0, info=0x559ad5df1cb0, user_data=<optimized out>) at ../mutter/cogl/cogl/winsys/cogl-onscreen-egl.c:319
#4 0x00007fc2313a682e in meta_onscreen_native_swap_buffers_with_damage (onscreen=0x559ad2082f50, rectangles=0x7fff6a1d8450, n_rectangles=0, frame_info=0x559ad5df1cb0, user_data=0x559ad2e52060) at ../mutter/src/backends/native/meta-onscreen-native.c:1092
#5 0x00007fc2314f9376 in cogl_onscreen_swap_buffers_with_damage
(onscreen=onscreen@entry=0x559ad2082f50, rectangles=rectangles@entry=0x7fff6a1d8450, n_rectangles=n_rectangles@entry=0, info=info@entry=0x559ad5df1cb0, user_data=user_data@entry=0x559ad2e52060) at ../mutter/cogl/cogl/cogl-onscreen.c:341
#6 0x00007fc231295c55 in swap_framebuffer
(stage_window=stage_window@entry=0x559ad2064760, stage_view=stage_view@entry=0x559ad207f350, swap_region=swap_region@entry=0x559ad5c3fdc0, swap_with_damage=swap_with_damage@entry=0, frame=frame@entry=0x559ad2e52060)
at ../mutter/src/backends/meta-stage-impl.c:324
#7 0x00007fc231296977 in meta_stage_impl_redraw_view_primary (frame=0x559ad2e52060, stage_view=0x559ad207f350, stage_impl=0x559ad2064760) at ../mutter/src/backends/meta-stage-impl.c:717
#8 meta_stage_impl_redraw_view (stage_window=0x559ad2064760, stage_view=0x559ad207f350, frame=0x559ad2e52060) at ../mutter/src/backends/meta-stage-impl.c:791
#9 0x00007fc2313af768 in meta_stage_native_redraw_view (stage_window=<optimized out>, view=0x559ad207f350, frame=0x559ad2e52060) at ../mutter/src/backends/native/meta-stage-native.c:144
#10 0x00007fc2316eaa9b in _clutter_stage_window_redraw_view (frame=0x559ad2e52060, view=0x559ad207f350, window=0x559ad2064760) at ../mutter/clutter/clutter/clutter-stage-window.c:95
#11 handle_frame_clock_frame (frame_clock=0x559ad1e83820, frame=0x559ad2e52060, user_data=0x559ad207f350) at ../mutter/clutter/clutter/clutter-stage-view.c:1267
#12 0x00007fc2316a9865 in clutter_frame_clock_dispatch (time_us=<optimized out>, frame_clock=0x559ad1e83820) at ../mutter/clutter/clutter/clutter-frame-clock.c:798
#13 frame_clock_source_dispatch (source=<optimized out>, callback=<optimized out>, user_data=<optimized out>) at ../mutter/clutter/clutter/clutter-frame-clock.c:846
#14 0x00007fc231934f19 in g_main_dispatch (context=0x559ad1cfd990) at ../glib/glib/gmain.c:3476
#15 0x00007fc2319932b7 in g_main_context_dispatch_unlocked (context=0x559ad1cfd990) at ../glib/glib/gmain.c:4284
#16 g_main_context_iterate_unlocked.isra.0 (context=0x559ad1cfd990, block=block@entry=1, dispatch=dispatch@entry=1, self=<optimized out>) at ../glib/glib/gmain.c:4349
#17 0x00007fc231935b47 in g_main_loop_run (loop=0x559ad3b9f0d0) at ../glib/glib/gmain.c:4551
#18 0x00007fc2312cc99b in meta_context_run_main_loop (context=context@entry=0x559ad1cfb8c0, error=error@entry=0x7fff6a1d89c0) at ../mutter/src/core/meta-context.c:482
#19 0x0000559ad1b4a526 in main (argc=<optimized out>, argv=<optimized out>) at ../gnome-shell/src/main.c:683

Downgrading mesa to the previous version fixes the issue. This problem is repeatable.
This task depends upon

Closed by  Buggy McBugFace (bugbot)
Saturday, 25 November 2023, 20:20 GMT
Reason for closing:  Moved
Additional comments about closing:  https://gitlab.archlinux.org/archlinux/p ackaging/packages/mesa/issues/2
Comment by Mark Blakeney (bulletmark) - Sunday, 01 October 2023, 23:29 GMT Comment by Maxime de Roucy (akira86) - Monday, 02 October 2023, 08:54 GMT
This issue also cause wayland session not to be accessible (gnome automatically fallback to X11).

I don't know if it's exactly the same issue because I already revert to 1:23.1.8-1 so I can't backtrack the dump…
coredumpctl debug gives me:
```
% sudo coredumpctl debug 876

[Thread debugging using libthread_db enabled]
Using host libthread_db library "/usr/lib/libthread_db.so.1".
Core was generated by `/usr/bin/gnome-shell'.
Program terminated with signal SIGSEGV, Segmentation fault.
#0 0x0000000000000000 in ?? ()
[Current thread is 1 (Thread 0x7f74b3b07d80 (LWP 876))]
(gdb) bt
#0 0x0000000000000000 in ()
#1 0x00007f74a8d9e196 in () at /usr/lib/dri/iris_dri.so
#2 0x000055ffed378e10 in ()
#3 0x00007ffdb4b2fef0 in ()
#4 0x000055ffeed7f750 in ()
#5 0x000055ffeed7f690 in ()
#6 0x00007f74b7f409e7 in () at /usr/lib/mutter-12/libmutter-cogl-12.so.0
#7 0x00007f74b7da682e in () at /usr/lib/libmutter-12.so.0
(gdb) quit
```
Comment by Mark Blakeney (bulletmark) - Monday, 02 October 2023, 12:34 GMT
@akira86, do `sudo -i coredumpctl debug 876` and say yes to debuginfod. Just put back the offending mesa version though before you do that. No need to reboot. Put the good version back after you get the back-trace. Although you will find it is the same as I report.
Comment by Maxime de Roucy (akira86) - Monday, 02 October 2023, 12:50 GMT
@bulletmark with `sudo -i` then `coredumpctl debug 876` and debuginfod :

```
(gdb) bt
#0 0x0000000000000000 in ()
#1 0x00007f74a8d9e196 in () at /usr/lib/dri/iris_dri.so
#2 0x000055ffed378e10 in ()
#3 0x00007ffdb4b2fef0 in ()
#4 0x000055ffeed7f750 in ()
#5 0x000055ffeed7f690 in ()
#6 0x00007f74b7f409e7 in batch_and_call (can_batch_callback=<optimized out>, batch_callback=<optimized out>, data=<optimized out>, n_entries=<optimized out>, entries=<optimized out>) at ../mutter/cogl/cogl/cogl-journal.c:260
#7 batch_and_call (data=0x7ffdb4b2fe00, batch_callback=0x7f74b7f40730 <_cogl_journal_flush_dither_and_entries>, can_batch_callback=<optimized out>, n_entries=<optimized out>, entries=0x55ffecffab38) at ../mutter/cogl/cogl/cogl-journal.c:229
#8 _cogl_journal_flush_viewport_and_entries (batch_start=0x55ffecffab38, batch_len=<optimized out>, data=0x7ffdb4b2fe00) at ../mutter/cogl/cogl/cogl-journal.c:1095
#9 0x00007f74b7da682e in meta_onscreen_native_swap_buffers_with_damage (onscreen=0x7ffdb4b2ffe0 [GInputStream], rectangles=0x600000041000001f, n_rectangles=268435495, frame_info=0x2000020c60000003, user_data=0x203650601000002d) at ../mutter/src/backends/native/meta-onscreen-native.c:1092
```
Comment by Toolybird (Toolybird) - Tuesday, 03 October 2023, 06:16 GMT
The backtraces look incomplete. Please review [1] and especially [2].

This doesn't appear to be an Arch packaging issue... let's wait to see what comes out of the linked mesa issue.

[1] https://wiki.archlinux.org/title/Debugging/Getting_traces
[2] https://blogs.gnome.org/mcatanzaro/2021/09/18/creating-quality-backtraces-for-crash-reports/
Comment by Kasper Seweryn (wvffle) - Wednesday, 04 October 2023, 06:34 GMT
I will just point to mine issue in the mutter repo with anotherstack trace: https://gitlab.gnome.org/GNOME/mutter/-/issues/3066#note_1859618
Comment by Kasper Seweryn (wvffle) - Wednesday, 04 October 2023, 18:59 GMT
I will just point to mine issue in the mutter repo with anotherstack trace: https://gitlab.gnome.org/GNOME/mutter/-/issues/3066#note_1859618
Comment by Toolybird (Toolybird) - Friday, 06 October 2023, 07:11 GMT
> let's wait to see what comes out of the linked mesa issue

It seems a fix has been found.
Comment by q rty (q234rty) - Friday, 06 October 2023, 18:00 GMT
Since the commit https://gitlab.freedesktop.org/mesa/mesa/-/commit/9590bce3e249a34665b2c42b20bfdbdc7f32147f is already upstream and seems to fix the issue, can we pick the patch ourselves?

Loading...