gnome session is logged out after resuming from suspend.

odror · 2026-03-26 20:41:05

Hi

In the last 3 days I am having a strange issue. when I resume from suspend I am logged out of the gnome session. I switched from gdm to sddm. It did not make a difference. base on the log file the issue happens during resume.

I am getting this error in journalctl

gnome-shell[7619]: (EE) failed to write to Xwayland fd: Broken pipe

Last edited by odror (2026-03-28 19:05:16)

SRG · 2026-04-04 12:40:40

I have the same symptoms, since ~one week :
- i suspend my (up to date, non-LTS) arch desktop system with the regular "systemctl suspend" command (as before)
- it goes properly into sleep mode
- when resumed, i'm logged out from my previous session

I use GNOME and GDM, nothing special / fancy and nothing modified recently.

I also see these errors

avril 04 09:18:11 moon systemd[1]: Started Process Core Dump (PID 3536211/UID 0).
avril 04 09:18:13 moon systemd[1]: Starting Realtime RGB LED Driver for Linux...
avril 04 09:18:13 moon systemd[1]: Started Realtime RGB LED Driver for Linux.
avril 04 09:18:13 moon systemd[1]: eruption-hotplug-helper.service: Deactivated successfully.
avril 04 09:18:13 moon eruption[3536215]: ** EXPERIMENTAL FEATURES are ENABLED, this may expose serious bugs! **
avril 04 09:18:13 moon kernel: input: Eruption Virtual Keyboard as /devices/virtual/input/input501
avril 04 09:18:13 moon kernel: input: Eruption Virtual Mouse as /devices/virtual/input/input502
avril 04 09:18:13 moon systemd-logind[754]: Watching system buttons on /dev/input/event5 (Eruption Virtual Keyboard)
avril 04 09:18:13 moon systemd-coredump[3536212]: [?] Process 2073096 (gnome-shell) of user 1000 dumped core.
                                                  
                                                  Stack trace of thread 2073096:
                                                  #0  0x00007ff7d778cce1 n/a (libmutter-17.so.0 + 0x18cce1)
                                                  #1  0x00007ff7d778d3d6 n/a (libmutter-17.so.0 + 0x18d3d6)
                                                  #2  0x00007ff7d83ad97c g_closure_invoke (libgobject-2.0.so.0 + 0x1197c)
                                                  #3  0x00007ff7d83ccdd7 n/a (libgobject-2.0.so.0 + 0x30dd7)
                                                  #4  0x00007ff7d83ce84f n/a (libgobject-2.0.so.0 + 0x3284f)
                                                  #5  0x00007ff7d83ceac9 g_signal_emit_valist (libgobject-2.0.so.0 + 0x32ac9)
                                                  #6  0x00007ff7d83ceb84 g_signal_emit (libgobject-2.0.so.0 + 0x32b84)
                                                  #7  0x00007ff7d76adb6d n/a (libmutter-17.so.0 + 0xadb6d)
                                                  #8  0x00007ff7d76adff1 meta_monitor_manager_rebuild (libmutter-17.so.0 + 0xadff1)
                                                  #9  0x00007ff7d77d31fa n/a (libmutter-17.so.0 + 0x1d31fa)
                                                  #10 0x00007ff7d76a252c meta_monitor_manager_ensure_configured (libmutter-17.so.0 + 0xa252c)
                                                  #11 0x00007ff7d77d2be0 n/a (libmutter-17.so.0 + 0x1d2be0)
                                                  #12 0x00007ff7d83ad97c g_closure_invoke (libgobject-2.0.so.0 + 0x1197c)
                                                  #13 0x00007ff7d83cc7eb n/a (libgobject-2.0.so.0 + 0x307eb)
                                                  #14 0x00007ff7d83ce84f n/a (libgobject-2.0.so.0 + 0x3284f)
                                                  #15 0x00007ff7d83ceac9 g_signal_emit_valist (libgobject-2.0.so.0 + 0x32ac9)
                                                  #16 0x00007ff7d83ceb84 g_signal_emit (libgobject-2.0.so.0 + 0x32b84)
                                                  #17 0x00007ff7d77ca2e2 n/a (libmutter-17.so.0 + 0x1ca2e2)
                                                  #18 0x00007ff7d7671b8c n/a (libmutter-17.so.0 + 0x71b8c)
                                                  #19 0x00007ff7d83ad97c g_closure_invoke (libgobject-2.0.so.0 + 0x1197c)
                                                  #20 0x00007ff7d83cc7eb n/a (libgobject-2.0.so.0 + 0x307eb)
                                                  #21 0x00007ff7d83ce84f n/a (libgobject-2.0.so.0 + 0x3284f)
                                                  #22 0x00007ff7d83ceac9 g_signal_emit_valist (libgobject-2.0.so.0 + 0x32ac9)
                                                  #23 0x00007ff7d83ceb84 g_signal_emit (libgobject-2.0.so.0 + 0x32b84)
                                                  #24 0x00007ff7d83ba306 n/a (libgobject-2.0.so.0 + 0x1e306)

(...)
avril 04 09:18:13 moon systemd[1]: systemd-coredump@7-24577-3536211_3536212-0.service: Deactivated successfully.
avril 04 09:18:13 moon systemd[1]: systemd-coredump@7-24577-3536211_3536212-0.service: Consumed 1.146s CPU time over 1.221s wall clock time, 746.1M memory peak.
avril 04 09:18:13 moon seafile-applet[2073269]: Error reading events from display: Relais brisé (pipe)
avril 04 09:18:13 moon thunar-volman[3401575]: Error reading events from display: Relais brisé (pipe)
avril 04 09:18:13 moon gnome-shell[2073171]: (EE) failed to write to Xwayland fd: Broken pipe

seth · 2026-04-04 14:46:24

Looks related to https://gitlab.gnome.org/GNOME/mutter/-/issues/4369
You can use https://wiki.archlinux.org/title/Debuginfod for a better backtrace

Whether @odror has a similar issue needs to be seen, the broken pipe is basically just a "FYI, the wayland compositor (gnome-shell) crashed"

odror · 2026-04-04 16:59:10

My issue is not related to thunderbolt. About a week ago there was an update that caused it. I noticed some changes in /lib/modprobe.d and /etc/modprobe.d. It is most likely like nvidia, mutter issue. Unrelated and a a few days later time my gpu (rtx 4070ti) broke. I have a temporary card the quadro m4000. Because of that I had to downgrade the nvidia driver to the one in the aur. I am still haveing the issue. I do not how to debug it or fix it.

SRG · 2026-04-04 18:49:54

I also don't have anything being Thunderbolt related (i have a NVIDIA T600 card with 4 DP outputs, connected to 3 HDMI monitors)

seth · 2026-04-04 19:38:56

This has nothing to do w/ thunderbolt specifically, the gist is that the output unregisters on DPMS and that knocks mutter off (null-sized array or null arithmetics)
Do you have a better backtrace?

SRG · 2026-04-04 19:47:43

In dmesg there is also one segfault regarding libmutter :

[sam. 4 avril 20:45:00 2026] OOM killer enabled.
[sam. 4 avril 20:45:00 2026] Restarting tasks: Starting
[sam. 4 avril 20:45:00 2026] Restarting tasks: Done
[sam. 4 avril 20:45:00 2026] efivarfs: resyncing variable state
[sam. 4 avril 20:45:00 2026] efivarfs: finished resyncing variable state
[sam. 4 avril 20:45:00 2026] random: crng reseeded on system resumption
[sam. 4 avril 20:45:00 2026] ata3: SATA link down (SStatus 4 SControl 300)
[sam. 4 avril 20:45:00 2026] ata1: SATA link down (SStatus 4 SControl 300)
[sam. 4 avril 20:45:00 2026] ata6: SATA link down (SStatus 4 SControl 300)
[sam. 4 avril 20:45:00 2026] ata5: SATA link down (SStatus 4 SControl 300)
[sam. 4 avril 20:45:00 2026] ata2: SATA link down (SStatus 4 SControl 300)
[sam. 4 avril 20:45:00 2026] ata4: SATA link down (SStatus 4 SControl 300)
[sam. 4 avril 20:45:01 2026] PM: suspend exit
[sam. 4 avril 20:45:01 2026] rfkill: input handler disabled
[sam. 4 avril 20:45:02 2026] gnome-shell[3540721]: segfault at 1c ip 00007f04a738cce1 sp 00007ffe2fcbab50 error 4 in libmutter-17.so.0.0.0[18cce1,7f04a723c000+1dc000] likely on CPU 4 (core 4, socket 0)
[sam. 4 avril 20:45:02 2026] Code: 83 ec 68 48 63 05 2f 76 14 00 89 4d cc 4c 89 45 90 4c 8b 6c 02 38 ff 15 ed f9 13 00 48 63 0d 16 76 14 00 49 8b 7f 58 89 45 b0 <41> 8b 45 1c 48 01 d9 89 45 c0 41 8b 45 20 48 8b 51 28 89 45 b8 41
[sam. 4 avril 20:45:03 2026] input: Eruption Virtual Keyboard as /devices/virtual/input/input503
[sam. 4 avril 20:45:03 2026] input: Eruption Virtual Mouse as /devices/virtual/input/input504
[sam. 4 avril 20:45:03 2026] rfkill: input handler enabled

Otherwise, journalctl (longer extract) : https://pastebin.com/E8X8qLc2

seth · 2026-04-04 19:56:05

https://wiki.archlinux.org/title/Core_d … _core_dump
https://wiki.archlinux.org/title/Debuginfod will produce a more speaking backtrace.

SRG · 2026-04-04 20:04:29

The coredumpctl output : https://pastebin.com/1Vv9sdxe

seth · 2026-04-04 20:06:12

w/o using debuginfod this doesn't provide any additional information on the problem.

odror · 2026-04-04 20:09:58

my system was analyzed by google gimini. It determined that I am missing " /sys/module/nvidia/parameters" specifically I am missing the parameters directory. It claims that there is some kind of race condition, where the Nvidia driver is not loaded properly. I do not know that if it is correct. I tried everything that was suggested, but it did not work. Just to remind you, shortly after the appearance of this bug, I had to replace my Nvidia card from rtx4070ti to quadro m4000 and I had to downgrade the drive to the propriety driver from the AUR. So this can be a complication issue in my system, but it does not explain why I am missing the parameters subdrirectory.

SRG · 2026-04-04 20:22:44

Well, on my side, i have 0 ideas about how to use debuginfod, and the arch wiki page is not understandable at all (for me).

seth · 2026-04-04 20:33:02

my system was analyzed by google gimini … but it did not work.

Quelle surprise
This most likely has nothing to do w/ any graphics driver implementation - it's what your output does during/after the suspend and a bug in gnome when handling monitors.
----
Edit, sorry - you've actually not shown a backtrace at all, so maybe we start w/ that to see whether SRG is just hijacking the topic.
Please post your complete system journal for the boot after a crash

sudo journalctl -b | nc termbin.com 9999

---

i have 0 ideas about how to use debuginfod

You'll be asked whether you want to use debuginfod when inspecting the backtrace and then don't enter "n" (I'm pretty sure it defaults to "y")

This is going nowhere…

Last edited by seth (2026-04-04 20:39:14)

odror · 2026-04-04 20:54:37

I posted the journalctl. The missing parameters subdirectory is not an issue, since I have arch linux on a different computer also without this subdirectory, but properly working suspend. I want to mention that my nvidia driver is the one from the AUR for old nvidia cards (580.142)

SRG · 2026-04-04 20:55:39

Well, i had the same symptoms, having started to appear from the same date, so it seemed better to join this thread.
But ok, i'll let ordor provide his traces, no problem, anyway i still don't understand what to do around debuginfod.

seth · 2026-04-04 20:58:00

I posted the journalctl.

Where? Afaict only SRG posted journals and coredumps?

I am getting this error in journalctl

is insignificant.

seth · 2026-04-04 21:01:47

@SRG from teh wiki

gdb automatically tries to download debug files from servers specified in the DEBUGINFOD_URLS environment variable, which is a string of space separated URLs.

echo $DEBUGINFOD_URLS

is supposed to be https://debuginfod.archlinux.org so when you follow https://wiki.archlinux.org/title/Core_d … _core_dump you'll be asked whether you want to fetch the debug symbols and then you just say "y"
That's all, the rest happens automagically. gdb will fetch the relevant debug symbols, apply them to the coredump and you get an extensive dump that will confirm or reject whether this is any similar to https://gitlab.gnome.org/GNOME/mutter/- … te_2703085

SRG · 2026-04-05 08:59:07

I still have no ideas what i'm doing :
- i launched "coredumpctl debug <ID>"
- there is no questions about fetching debug symbols, or in fact, any other questions (i'm directly landing on a prompt), so obviously maybe the first step if wrong, i don't know (otherwise it says "No debugging symbols found in /usr/bin/gnome-shell" somewhere at the beginning, maybe this is why the question is skipped, but i have also no idea what to do about that)
- at this point if i type "bt" in the console i get this stack : https://pastebin.com/02CPD83q (but it does not seem to be more helpful)

odror · 2026-04-05 16:36:21

I have an update. I do not have the issue with the plasma desktop or cosmic desktop. It is a gnome issue. I rolled back all the updates related to gnome from the last week. It did not resolve the gnome issue. In fact I have zfs. I rolled back my system to the time before the issue. It also did not resolve the problem.

seth · 2026-04-05 18:00:48

@SRG,

echo $DEBUGINFOD_URLS

https://wiki.archlinux.org/title/Debugi … stallation

libelf, a dependency of gdb, ships with /etc/profile.d/debuginfod.sh and /etc/profile.d/debuginfod.csh scripts that set the variable on login, so there is no need to install additional packages. The scripts parse .urls files in /etc/debuginfod/ and set the environment variable to https://debuginfod.archlinux.org by default.

@odror
Do you have any intentions at all to provide a journal or relevant backtrace for the crash?

odror · 2026-04-05 18:13:26

seth wrote:

@SRG,
echo $DEBUGINFOD_URLS
@odror
Do you have any intentions at all to provide a journal or relevant backtrace for the crash?

I thought that I reported it on termbin.com 9999

seth · 2026-04-05 18:27:37

You'll have to post the link you get in return

Arch Linux

#1 2026-03-26 20:41:05

gnome session is logged out after resuming from suspend.

#2 2026-04-04 12:40:40

Re: gnome session is logged out after resuming from suspend.

#3 2026-04-04 14:46:24

Re: gnome session is logged out after resuming from suspend.

#4 2026-04-04 16:59:10

Re: gnome session is logged out after resuming from suspend.

#5 2026-04-04 18:49:54

Re: gnome session is logged out after resuming from suspend.

#6 2026-04-04 19:38:56

Re: gnome session is logged out after resuming from suspend.

#7 2026-04-04 19:47:43

Re: gnome session is logged out after resuming from suspend.

#8 2026-04-04 19:56:05

Re: gnome session is logged out after resuming from suspend.

#9 2026-04-04 20:04:29

Re: gnome session is logged out after resuming from suspend.

#10 2026-04-04 20:06:12

Re: gnome session is logged out after resuming from suspend.

#11 2026-04-04 20:09:58

Re: gnome session is logged out after resuming from suspend.

#12 2026-04-04 20:22:44

Re: gnome session is logged out after resuming from suspend.

#13 2026-04-04 20:33:02

Re: gnome session is logged out after resuming from suspend.

#14 2026-04-04 20:54:37

Re: gnome session is logged out after resuming from suspend.

#15 2026-04-04 20:55:39

Re: gnome session is logged out after resuming from suspend.

#16 2026-04-04 20:58:00

Re: gnome session is logged out after resuming from suspend.

#17 2026-04-04 21:01:47

Re: gnome session is logged out after resuming from suspend.

#18 2026-04-05 08:59:07

Re: gnome session is logged out after resuming from suspend.

#19 2026-04-05 16:36:21

Re: gnome session is logged out after resuming from suspend.

#20 2026-04-05 18:00:48

Re: gnome session is logged out after resuming from suspend.

#21 2026-04-05 18:13:26

Re: gnome session is logged out after resuming from suspend.

#22 2026-04-05 18:27:37

Re: gnome session is logged out after resuming from suspend.

Board footer