gnome-shell crashes on Intel Ice Lake with SIGSEGV at NULL from end_query() from cogl_gl_create_timestamp_query() from cogl_onscreen_egl_swap_buffers_with_damage()

Bug #2037497 reported by Anas
30
This bug affects 3 people
Affects Status Importance Assigned to Milestone
Mesa
New
Unknown
mesa (Fedora)
Fix Released
High
mesa (Ubuntu)
Fix Released
High
Timo Aaltonen

Bug Description

crash after update
wayland session is not accessible anymore

ProblemType: Crash
DistroRelease: Ubuntu 23.10
Package: gnome-shell 45.0-1ubuntu1
ProcVersionSignature: Ubuntu 6.5.0-5.5-generic 6.5.0
Uname: Linux 6.5.0-5-generic x86_64
NonfreeKernelModules: nvidia_modeset nvidia
ApportVersion: 2.27.0-0ubuntu2
Architecture: amd64
CasperMD5CheckResult: pass
CurrentDesktop: ubuntu:GNOME
Date: Wed Sep 27 05:52:26 2023
DisplayManager: gdm3
ExecutablePath: /usr/bin/gnome-shell
InstallationDate: Installed on 2023-09-26 (0 days ago)
InstallationMedia: Ubuntu 23.10 "Mantic Minotaur" - Beta amd64 (20230924)
ProcCmdline: /usr/bin/gnome-shell
ProcEnviron:
 LANG=fr_FR.UTF-8
 PATH=(custom, no user)
 SHELL=/bin/bash
 XDG_RUNTIME_DIR=<set>
RelatedPackageVersions: mutter-common 45.0-2ubuntu1
Signal: 11
SourcePackage: gnome-shell
StacktraceTop:
 ?? ()
 ?? () from /usr/lib/x86_64-linux-gnu/dri/iris_dri.so
 ?? () from /usr/lib/x86_64-linux-gnu/mutter-13/libmutter-cogl-13.so.0
 ?? () from /usr/lib/x86_64-linux-gnu/mutter-13/libmutter-cogl-13.so.0
 ?? () from /lib/x86_64-linux-gnu/libmutter-13.so.0
Title: gnome-shell crashed with SIGSEGV
UpgradeStatus: No upgrade log present (probably fresh install)
UserGroups: adm cdrom dip libvirt lpadmin plugdev sudo users
separator:

Revision history for this message
In , filbar (filbar-redhat-bugs) wrote :

After last update I have problems with GDM and gnome-shell. When GDM starts id hangs with error:

zář 13 11:40:40 fedora systemd[1]: Starting gdm.service - GNOME Display Manager...
zář 13 11:40:40 fedora systemd[1]: Started gdm.service - GNOME Display Manager.
zář 13 11:40:59 2001-1ae9-2e8-fa00-5260-9dab-358b-fd26.ip6.tmcz.cz gdm[928]: Gdm: GdmDisplay: Session never registered, failing
zář 13 11:40:59 2001-1ae9-2e8-fa00-5260-9dab-358b-fd26.ip6.tmcz.cz gdm[928]: Gdm: Child process -1064 was already dead.
zář 13 11:40:59 2001-1ae9-2e8-fa00-5260-9dab-358b-fd26.ip6.tmcz.cz gdm[928]: Gdm: GdmDisplay: Session never registered, failing
zář 13 11:40:59 2001-1ae9-2e8-fa00-5260-9dab-358b-fd26.ip6.tmcz.cz gdm[928]: Gdm: Child process -1064 was already dead

After downgrade of gnome-shell all starts working

Reproducible: Always

Steps to Reproduce:
1. Update fedora
2. Reboot
3. gdm hangs
Actual Results:
GDM hankgs

Expected Results:
GDM login screen

Revision history for this message
In , filbar (filbar-redhat-bugs) wrote :

Created attachment 1989131
Debug info

Revision history for this message
In , filbar (filbar-redhat-bugs) wrote :

I log i found

abrt-notification[1836]: Process 1141 (gnome-shell) crashed in cogl_gl_create_timestamp_query()

This works in gnome-shell beta - after upgrate to rc version I have this error-

Revision history for this message
In , kparal (kparal-redhat-bugs) wrote :

Filip, please attach a full system journal of the failed boot. `journalctl -b > journal.txt` for the current boot, or `journalctl -b -1 > journal.txt` for the previous boot. Thanks. Also please attach `lspci` output. Thanks.

Revision history for this message
In , james (james-redhat-bugs-1) wrote :

Seen here after upgrade to F39 on a machine with i3-1005G1 graphics.

kernel-6.5.4-300.fc39.x86_64
mesa-dri-drivers-23.2.0~rc3-1.fc39.x86_64
gdm-45.0.1-1.fc39.x86_64
mutter-45.0-2.fc39.x86_64
gnome-shell-45.0-1.fc39.x86_64

My journal, lspci and backtrace attached below.

Revision history for this message
In , james (james-redhat-bugs-1) wrote :

Created attachment 1990365
output from journalctl -b

Revision history for this message
In , james (james-redhat-bugs-1) wrote :

Created attachment 1990366
output from lspci

Revision history for this message
In , james (james-redhat-bugs-1) wrote :

Created attachment 1990367
backtrace

Revision history for this message
In , kparal (kparal-redhat-bugs) wrote :

Thanks, James. Could you please file a ticket at https://gitlab.gnome.org/GNOME/gnome-shell/-/issues , attach or link the files you attached here, and paste the resulting upstream ticket URL here?

Revision history for this message
Anas (anas-blinux) wrote :
information type: Private → Public
Revision history for this message
Apport retracing service (apport) wrote :

StacktraceTop:
 ?? ()
 end_query (ctx=0x559358438740, q=0x55935ba85cd0) at ../src/mesa/main/queryobj.c:260
 cogl_gl_create_timestamp_query (context=0x5593584860b0) at ../cogl/cogl/driver/gl/cogl-util-gl.c:520
 cogl_onscreen_egl_swap_buffers_with_damage (onscreen=0x559357fce9e0, rectangles=0x7ffcc5bd48d0, n_rectangles=0, info=0x55935c10ad30, user_data=<optimized out>) at ../cogl/cogl/winsys/cogl-onscreen-egl.c:319
 meta_onscreen_native_swap_buffers_with_damage (onscreen=0x559357fce9e0, rectangles=0x7ffcc5bd48d0, n_rectangles=0, frame_info=0x55935c10ad30, user_data=0x55935afcdd30) at ../src/backends/native/meta-onscreen-native.c:1126

Revision history for this message
Apport retracing service (apport) wrote : Stacktrace.txt
Revision history for this message
Apport retracing service (apport) wrote : StacktraceSource.txt
Revision history for this message
Apport retracing service (apport) wrote : ThreadStacktrace.txt
Changed in gnome-shell (Ubuntu):
importance: Undecided → Medium
tags: removed: need-amd64-retrace
Revision history for this message
Daniel van Vugt (vanvugt) wrote : Re: gnome-shell crashed with SIGSEGV in ?? (iris_dri.so) from end_query() from cogl_gl_create_timestamp_query() from cogl_onscreen_egl_swap_buffers_with_damage()
summary: - gnome-shell crashed with SIGSEGV
+ gnome-shell crashed with SIGSEGV in ?? (iris_dri.so) from end_query()
+ from cogl_gl_create_timestamp_query() from
+ cogl_onscreen_egl_swap_buffers_with_damage()
tags: added: regression regression-release rls-mm-incoming
affects: gnome-shell (Ubuntu) → mesa (Ubuntu)
tags: removed: regression-release
Changed in mesa (Ubuntu):
importance: Medium → Critical
Changed in mutter (Ubuntu):
importance: Undecided → Critical
tags: added: regression-release
Revision history for this message
In , james (james-redhat-bugs-1) wrote :

(In reply to Kamil Páral from comment #8)
> Thanks, James. Could you please file a ticket at
> https://gitlab.gnome.org/GNOME/gnome-shell/-/issues , attach or link the
> files you attached here, and paste the resulting upstream ticket URL here?

https://gitlab.gnome.org/GNOME/gnome-shell/-/issues/7053

Revision history for this message
In , james (james-redhat-bugs-1) wrote :
Revision history for this message
In , james (james-redhat-bugs-1) wrote :
Revision history for this message
In , jeischma (jeischma-redhat-bugs) wrote :

I'm experiencing the same crashes. GNOME does start up in the end (maybe because I have autologin?), but the boot is slow due to the crashes and GNOME fallbacks to X11 with no multi-finger gestures not working etc.

I have a 2020 Dell XPS 13 with Intel Iris Plus GPU.

Revision history for this message
In , blockerbugs (blockerbugs-redhat-bugs) wrote :

Proposed as a Blocker for 39-final by Fedora user eischmann using the blocker tracking app because:

 This bug severely impacts F39 Workstation experience on machines with Intel Iris Plus GPU. GNOME Shell/GDM doesn't start up at all or fallbacks to X11 with degraded experience.

Revision history for this message
In , awilliam (awilliam-redhat-bugs) wrote :

Can reporters please test downgrading packages to figure out what actually fixes this? I'd suggest this order:

1. Downgrade mesa to https://koji.fedoraproject.org/koji/buildinfo?buildID=2274339
2. If that doesn't fix it, downgrade mutter and gnome-shell to https://koji.fedoraproject.org/koji/buildinfo?buildID=2284675 and https://koji.fedoraproject.org/koji/buildinfo?buildID=2284676

if that doesn't fix it, report back and we'll try something else, I guess. :D

Instructions for downgrading: `dnf -y install koji`, then make a scratch dir somewhere (in your home dir or /var/tmp or something), and do `koji download-build --arch=x86_64 --arch=noarch (buildid)` , where (buildid) is the build ID from the URL, so 2274339 for mesa. Then do `dnf downgrade *.rpm`.

Revision history for this message
Daniel van Vugt (vanvugt) wrote :

It appears 'pipe' is NULL in the Mesa source at:

static void
end_query(struct gl_context *ctx, struct gl_query_object *q)
{
   struct st_context *st = st_context(ctx);
   struct pipe_context *pipe = ctx->pipe;
   bool ret = false;

   st_flush_bitmap_cache(st_context(ctx));

   if ((q->Target == GL_TIMESTAMP ||
        q->Target == GL_TIME_ELAPSED) &&
       !q->pq) {
      q->pq = pipe->create_query(pipe, PIPE_QUERY_TIMESTAMP, 0);
      q->type = PIPE_QUERY_TIMESTAMP;
   }

I can't reproduce the problem myself so I wonder are you using two GPUs? Do you have a monitor plugged into an Nvidia GPU?

Revision history for this message
Daniel van Vugt (vanvugt) wrote :
Changed in mesa (Ubuntu):
status: New → Confirmed
Changed in mutter (Ubuntu):
status: New → Confirmed
summary: - gnome-shell crashed with SIGSEGV in ?? (iris_dri.so) from end_query()
- from cogl_gl_create_timestamp_query() from
+ gnome-shell crashed with SIGSEGV at NULL from end_query() from
+ cogl_gl_create_timestamp_query() from
cogl_onscreen_egl_swap_buffers_with_damage()
Changed in mesa (Fedora):
importance: Unknown → High
status: Unknown → Confirmed
Changed in mesa:
status: Unknown → New
Changed in mutter:
status: Unknown → Fix Released
Revision history for this message
Anas (anas-blinux) wrote (last edit ): Re: gnome-shell crashed with SIGSEGV at NULL from end_query() from cogl_gl_create_timestamp_query() from cogl_onscreen_egl_swap_buffers_with_damage()

For my case i have two GPUs,
 * iGPU for both laptop's screen and a monitor through an HDMI.
 * Nvidia for AI / CUDA stuff

the same problem on Fedora 39 ...

Revision history for this message
In , james (james-redhat-bugs-1) wrote :

(In reply to Adam Williamson from comment #14)
> Can reporters please test downgrading packages to figure out what actually
> fixes this? I'd suggest this order:
>
> 1. Downgrade mesa to
> https://koji.fedoraproject.org/koji/buildinfo?buildID=2274339
> 2. If that doesn't fix it, downgrade mutter and gnome-shell to
> https://koji.fedoraproject.org/koji/buildinfo?buildID=2284675 and
> https://koji.fedoraproject.org/koji/buildinfo?buildID=2284676
>
> if that doesn't fix it, report back and we'll try something else, I guess. :D
>
> Instructions for downgrading: `dnf -y install koji`, then make a scratch dir
> somewhere (in your home dir or /var/tmp or something), and do `koji
> download-build --arch=x86_64 --arch=noarch (buildid)` , where (buildid) is
> the build ID from the URL, so 2274339 for mesa. Then do `dnf downgrade
> *.rpm`.

None of this worked, nor mesa-23.2.0~rc2-1.fc39.

I was able to get it working again by downgrading to mesa-23.1.5-1.fc39 (build 2268808). Will try to bisect.

Revision history for this message
In , awilliam (awilliam-redhat-bugs) wrote :

aha, thanks!

Revision history for this message
In , kparal (kparal-redhat-bugs) wrote :
Revision history for this message
Daniel van Vugt (vanvugt) wrote : Re: gnome-shell crashed with SIGSEGV at NULL from end_query() from cogl_gl_create_timestamp_query() from cogl_onscreen_egl_swap_buffers_with_damage()

Dropped severity because this bug is proving elusive. Most of us are never seeing it, even when we try with similar hardware.

Changed in mesa (Ubuntu):
importance: Critical → High
Changed in mutter (Ubuntu):
importance: Critical → High
Revision history for this message
In , ndegraef (ndegraef-redhat-bugs) wrote :

*** Bug 2241332 has been marked as a duplicate of this bug. ***

Changed in mesa (Ubuntu):
milestone: none → ubuntu-23.10
Changed in mutter (Ubuntu):
status: Confirmed → Invalid
Tim Holmes-Mitra (timhm)
tags: added: rls-mm-notfixing
removed: rls-mm-incoming
Revision history for this message
In , awilliam (awilliam-redhat-bugs) wrote :

Some upstream discussion indicates https://gitlab.freedesktop.org/mesa/mesa/-/commit/9590bce3e249a34665b2c42b20bfdbdc7f32147f may fix this. Setting POST, will do a scratch build for folks to test shortly.

Revision history for this message
In , awilliam (awilliam-redhat-bugs) wrote :

Scratch build is done for x86_64: https://koji.fedoraproject.org/koji/taskinfo?taskID=107101328

can folks test that and see if it helps? Thanks!

Revision history for this message
In , james (james-redhat-bugs-1) wrote :

(In reply to Adam Williamson from comment #20)
> Scratch build is done for x86_64:
> https://koji.fedoraproject.org/koji/taskinfo?taskID=107101328
>
> can folks test that and see if it helps? Thanks!

Works after updating to the packages in that Koji task.

Revision history for this message
In , updates (updates-redhat-bugs) wrote :

FEDORA-2023-86e10b6cae has been submitted as an update to Fedora 39. https://bodhi.fedoraproject.org/updates/FEDORA-2023-86e10b6cae

Revision history for this message
In , awilliam (awilliam-redhat-bugs) wrote :

Thanks James, I've sent an official update as you can see.

Revision history for this message
Daniel van Vugt (vanvugt) wrote : Re: gnome-shell crashed with SIGSEGV at NULL from end_query() from cogl_gl_create_timestamp_query() from cogl_onscreen_egl_swap_buffers_with_damage()
Changed in mesa (Ubuntu):
milestone: ubuntu-23.10 → mantic-updates
tags: added: fixed-in-mesa-23.3 fixed-upstream
Changed in mesa (Ubuntu):
status: Confirmed → Triaged
summary: - gnome-shell crashed with SIGSEGV at NULL from end_query() from
- cogl_gl_create_timestamp_query() from
+ gnome-shell crashes on Intel Ice Lake with SIGSEGV at NULL from
+ end_query() from cogl_gl_create_timestamp_query() from
cogl_onscreen_egl_swap_buffers_with_damage()
Revision history for this message
In , updates (updates-redhat-bugs) wrote :

FEDORA-2023-86e10b6cae has been pushed to the Fedora 39 testing repository.
Soon you'll be able to install the update with the following command:
`sudo dnf upgrade --enablerepo=updates-testing --refresh --advisory=FEDORA-2023-86e10b6cae`
You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-2023-86e10b6cae

See also https://fedoraproject.org/wiki/QA:Updates_Testing for more information on how to test updates.

no longer affects: mutter (Ubuntu)
no longer affects: mutter
Revision history for this message
Timo Aaltonen (tjaalton) wrote :

I'll add that patch to mantic now, as it would take some time to get it via a bugfix release

Changed in mesa (Ubuntu):
assignee: nobody → Timo Aaltonen (tjaalton)
Changed in mesa (Ubuntu):
milestone: mantic-updates → ubuntu-23.10
Changed in mesa (Fedora):
status: Confirmed → Fix Committed
Revision history for this message
Launchpad Janitor (janitor) wrote :

This bug was fixed in the package mesa - 23.2.1-1ubuntu3

---------------
mesa (23.2.1-1ubuntu3) mantic; urgency=medium

  * patches: Fix a function name collision causing crashes. (LP:
    #2037497)

 -- Timo Aaltonen <email address hidden> Fri, 06 Oct 2023 11:58:09 +0300

Changed in mesa (Ubuntu):
status: Triaged → Fix Released
Revision history for this message
In , updates (updates-redhat-bugs) wrote :

FEDORA-2023-86e10b6cae has been pushed to the Fedora 39 stable repository.
If problem still persists, please make note of it in this bug report.

Changed in mesa (Fedora):
status: Fix Committed → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Duplicates of this bug

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.