Bug#1072840: Acknowledgement (sddm: Sddm crashes regularly)
Hello Aurélien, Am Mon, Jun 30, 2025 at 12:18:03AM +0200 schrieb Aurélien COUDERC: > Le samedi 21 juin 2025, 18:41:59 heure d’été d’Europe centrale Helge > Kreutzmann a écrit : > > Hello Aurélien, > > Dear Helge, > > > did you have a chance to test this? This is really anyoing. > > > > I just discovered that it is also the other way round: > > run (as root) > > systemctl restart sddm > > → Console 7 is closed and programms running there are killed. > > Yes, I could reproduce this. Thanks for checking. > It is due to SDDM declaring a [email protected] in its service > file. See [1] and [2]. Ok, the I just have to find out how to properly edit the file, /usr is not simply an edito away. > I seem to remember that in the distant past this would have been required > because display managers had to run on tty7 not to step over user sessions or > for managing some interaction with plymouth or something… > I’m not sure this still makes much sense today, and you will probably be fine > removing that Conflicts relation with tty7 from the unit. Feel free to give > it a try and report back. > > I’d say it’s a pretty niche use case since the default install has 6 ttys > enabled by default and I’m not sure many people are both using 7 or more ttys > and using sddm. No idea either. > And I don’t feel confident changing this in the Debian package so close to > the release. I’m not sure if there are any interactions with other components > we could break this way. But we can certainly do the change early in the > forky cycle. Great. Maybe a short mention in the release notes could be done? Greetings Helge -- Dr. Helge Kreutzmann [email protected] Dipl.-Phys. http://www.helgefjell.de/debian.php 64bit GNU powered gpg signed mail preferred Help keep free software "libre": http://www.ffii.de/ signature.asc Description: PGP signature
Bug#1072840: Acknowledgement (sddm: Sddm crashes regularly)
Le samedi 21 juin 2025, 18:41:59 heure d’été d’Europe centrale Helge Kreutzmann a écrit : > Hello Aurélien, Dear Helge, > did you have a chance to test this? This is really anyoing. > > I just discovered that it is also the other way round: > run (as root) > systemctl restart sddm > → Console 7 is closed and programms running there are killed. Yes, I could reproduce this. It is due to SDDM declaring a [email protected] in its service file. See [1] and [2]. I seem to remember that in the distant past this would have been required because display managers had to run on tty7 not to step over user sessions or for managing some interaction with plymouth or something… I’m not sure this still makes much sense today, and you will probably be fine removing that Conflicts relation with tty7 from the unit. Feel free to give it a try and report back. I’d say it’s a pretty niche use case since the default install has 6 ttys enabled by default and I’m not sure many people are both using 7 or more ttys and using sddm. And I don’t feel confident changing this in the Debian package so close to the release. I’m not sure if there are any interactions with other components we could break this way. But we can certainly do the change early in the forky cycle. [1] https://salsa.debian.org/qt-kde-team/3rdparty/sddm/-/blob/master/debian/sddm.service?ref_type=heads#L5 [2] /usr/lib/systemd/system/sddm.service Happy hacking, -- Aurélien
Bug#1072840: Acknowledgement (sddm: Sddm crashes regularly)
Hello Aurélien,
did you have a chance to test this? This is really anyoing.
I just discovered that it is also the other way round:
run (as root)
systemctl restart sddm
→ Console 7 is closed and programms running there are killed.
Again, all other (of my 12) consoles are unaffected.
Greetings
Helge
Am Wed, Aug 07, 2024 at 03:47:37PM + schrieb Helge Kreutzmann:
> Hello Aurélien,
> the following is 100% reproducible:
>
> (1. Boot machine )
> 2. Log into Console 7 (it only happens with 7)
> 3. Log out again
>
> → Sddm is gone
>
> For all other consoles it works (I tried 2…6 and 8…12).
>
> Can you try this as well?
>
> Am Sun, Jun 23, 2024 at 09:39:40AM +0200 schrieb Aurélien COUDERC:
> > Le 23 juin 2024 09:28:41 GMT+02:00, Helge Kreutzmann
> > a écrit :
> > >Hello Sddm maintainers,
> > >I analysed this annyoing bug further. The following pattern seems to
> > >hold always:
> >
> > can you see anything in the kernel or system logs (dmesg or journalctl) ?
>
> 2024-08-07T17:24:58.765735+02:00 twentytwo sddm[1522]: Signal received:
> SIGTERM
> 2024-08-07T17:24:58.765803+02:00 twentytwo systemd[1]: Stopping sddm.service
> - Simple Desktop Display Manager...
> 2024-08-07T17:24:58.766008+02:00 twentytwo sddm-helper[3452]: Signal
> received: SIGTERM
> 2024-08-07T17:24:58.766946+02:00 twentytwo sddm-helper[3452]: [PAM] Closing
> session
> 2024-08-07T17:24:58.770046+02:00 twentytwo sddm-helper[3452]: [PAM] Ended.
> 2024-08-07T17:24:58.771092+02:00 twentytwo sddm[1522]: Auth: sddm-helper
> exited with 255
> 2024-08-07T17:24:58.771147+02:00 twentytwo sddm[1522]: Socket server
> stopping...
> 2024-08-07T17:24:58.771187+02:00 twentytwo sddm[1522]: Socket server stopped.
> 2024-08-07T17:24:58.771218+02:00 twentytwo sddm[1522]: Display server
> stopping...
> 2024-08-07T17:24:58.825643+02:00 twentytwo sddm[1522]: Display server stopped.
> 2024-08-07T17:24:58.825738+02:00 twentytwo sddm[1522]: Running display stop
> script ("/usr/share/sddm/scripts/Xstop")
> 2024-08-07T17:24:58.828508+02:00 twentytwo systemd[1]: sddm.service:
> Deactivated successfully.
> 2024-08-07T17:24:58.828736+02:00 twentytwo systemd[1]: Stopped sddm.service -
> Simple Desktop Display Manager.
> 2024-08-07T17:24:58.828895+02:00 twentytwo systemd[1]: sddm.service: Consumed
> 578ms CPU time, 160.9M memory peak.
>
> Or 2nd try just afterwards (with some more context):
> 2024-08-07T17:25:21.985606+02:00 twentytwo systemd[1]: [email protected]:
> Deactivated successfully.
> 2024-08-07T17:25:22.018143+02:00 twentytwo systemd[1]:
> run-credentials-getty\x40tty7.service.mount: Deactivated successfully.
> 2024-08-07T17:25:22.019192+02:00 twentytwo systemd[1]: [email protected]:
> Scheduled restart job, restart counter is at 1.
> 2024-08-07T17:25:22.020508+02:00 twentytwo sddm[3717]: Signal received:
> SIGTERM
> 2024-08-07T17:25:22.020553+02:00 twentytwo systemd[1]: Stopping sddm.service
> - Simple Desktop Display Manager...
> 2024-08-07T17:25:22.020781+02:00 twentytwo sddm-helper[3826]: Signal
> received: SIGTERM
> 2024-08-07T17:25:22.021592+02:00 twentytwo sddm-helper[3826]: [PAM] Closing
> session
> 2024-08-07T17:25:22.023305+02:00 twentytwo sddm-helper[3826]: [PAM] Ended.
> 2024-08-07T17:25:22.025129+02:00 twentytwo sddm[3717]: Auth: sddm-helper
> exited with 255
> 2024-08-07T17:25:22.025175+02:00 twentytwo sddm[3717]: Socket server
> stopping...
> 2024-08-07T17:25:22.025212+02:00 twentytwo sddm[3717]: Socket server stopped.
> 2024-08-07T17:25:22.025241+02:00 twentytwo sddm[3717]: Display server
> stopping...
> 2024-08-07T17:25:22.080872+02:00 twentytwo sddm[3717]: Display server stopped.
> 2024-08-07T17:25:22.080989+02:00 twentytwo sddm[3717]: Running display stop
> script ("/usr/share/sddm/scripts/Xstop")
> 2024-08-07T17:25:22.083619+02:00 twentytwo systemd[1]: sddm.service:
> Deactivated successfully.
> 2024-08-07T17:25:22.083858+02:00 twentytwo systemd[1]: Stopped sddm.service -
> Simple Desktop Display Manager.
> 2024-08-07T17:25:22.085111+02:00 twentytwo systemd[1]: Started
> [email protected] - Getty on tty7.
> 2024-08-07T17:25:25.325658+02:00 twentytwo systemd[1]: Stopping
> [email protected] - Getty on tty7...
> 2024-08-07T17:25:25.325953+02:00 twentytwo systemd[1]: [email protected]:
> Deactivated successfully.
> 2024-08-07T17:25:25.326075+02:00 twentytwo systemd[1]: Stopped
> [email protected] - Getty on tty7.
> 2024-08-07T17:25:25.328211+02:00 twentytwo systemd[1]:
> run-credentials-getty\x40tty7.service.mount: Deactivated successfully.
>
> > Without logs it will be difficult to understand what is going on. But at
> > first sight I would lean towards a graphics driver bug more than an SDDM
> > bug.
>
> I really don't think so.
>
> > Maybe testing with different display managers (gdm, lightdm) could also
> > help narrowing down the issue ?
>
> Yes, I still need to do this.
>
> Greetings
>
> Helge
> --
> Dr. Helge Kreutzmann [email protected]
Bug#1072840: Acknowledgement (sddm: Sddm crashes regularly)
Hello Aurélien,
the following is 100% reproducible:
(1. Boot machine )
2. Log into Console 7 (it only happens with 7)
3. Log out again
→ Sddm is gone
For all other consoles it works (I tried 2…6 and 8…12).
Can you try this as well?
Am Sun, Jun 23, 2024 at 09:39:40AM +0200 schrieb Aurélien COUDERC:
> Le 23 juin 2024 09:28:41 GMT+02:00, Helge Kreutzmann a
> écrit :
> >Hello Sddm maintainers,
> >I analysed this annyoing bug further. The following pattern seems to
> >hold always:
>
> can you see anything in the kernel or system logs (dmesg or journalctl) ?
2024-08-07T17:24:58.765735+02:00 twentytwo sddm[1522]: Signal received: SIGTERM
2024-08-07T17:24:58.765803+02:00 twentytwo systemd[1]: Stopping sddm.service -
Simple Desktop Display Manager...
2024-08-07T17:24:58.766008+02:00 twentytwo sddm-helper[3452]: Signal received:
SIGTERM
2024-08-07T17:24:58.766946+02:00 twentytwo sddm-helper[3452]: [PAM] Closing
session
2024-08-07T17:24:58.770046+02:00 twentytwo sddm-helper[3452]: [PAM] Ended.
2024-08-07T17:24:58.771092+02:00 twentytwo sddm[1522]: Auth: sddm-helper exited
with 255
2024-08-07T17:24:58.771147+02:00 twentytwo sddm[1522]: Socket server stopping...
2024-08-07T17:24:58.771187+02:00 twentytwo sddm[1522]: Socket server stopped.
2024-08-07T17:24:58.771218+02:00 twentytwo sddm[1522]: Display server
stopping...
2024-08-07T17:24:58.825643+02:00 twentytwo sddm[1522]: Display server stopped.
2024-08-07T17:24:58.825738+02:00 twentytwo sddm[1522]: Running display stop
script ("/usr/share/sddm/scripts/Xstop")
2024-08-07T17:24:58.828508+02:00 twentytwo systemd[1]: sddm.service:
Deactivated successfully.
2024-08-07T17:24:58.828736+02:00 twentytwo systemd[1]: Stopped sddm.service -
Simple Desktop Display Manager.
2024-08-07T17:24:58.828895+02:00 twentytwo systemd[1]: sddm.service: Consumed
578ms CPU time, 160.9M memory peak.
Or 2nd try just afterwards (with some more context):
2024-08-07T17:25:21.985606+02:00 twentytwo systemd[1]: [email protected]:
Deactivated successfully.
2024-08-07T17:25:22.018143+02:00 twentytwo systemd[1]:
run-credentials-getty\x40tty7.service.mount: Deactivated successfully.
2024-08-07T17:25:22.019192+02:00 twentytwo systemd[1]: [email protected]:
Scheduled restart job, restart counter is at 1.
2024-08-07T17:25:22.020508+02:00 twentytwo sddm[3717]: Signal received: SIGTERM
2024-08-07T17:25:22.020553+02:00 twentytwo systemd[1]: Stopping sddm.service -
Simple Desktop Display Manager...
2024-08-07T17:25:22.020781+02:00 twentytwo sddm-helper[3826]: Signal received:
SIGTERM
2024-08-07T17:25:22.021592+02:00 twentytwo sddm-helper[3826]: [PAM] Closing
session
2024-08-07T17:25:22.023305+02:00 twentytwo sddm-helper[3826]: [PAM] Ended.
2024-08-07T17:25:22.025129+02:00 twentytwo sddm[3717]: Auth: sddm-helper exited
with 255
2024-08-07T17:25:22.025175+02:00 twentytwo sddm[3717]: Socket server stopping...
2024-08-07T17:25:22.025212+02:00 twentytwo sddm[3717]: Socket server stopped.
2024-08-07T17:25:22.025241+02:00 twentytwo sddm[3717]: Display server
stopping...
2024-08-07T17:25:22.080872+02:00 twentytwo sddm[3717]: Display server stopped.
2024-08-07T17:25:22.080989+02:00 twentytwo sddm[3717]: Running display stop
script ("/usr/share/sddm/scripts/Xstop")
2024-08-07T17:25:22.083619+02:00 twentytwo systemd[1]: sddm.service:
Deactivated successfully.
2024-08-07T17:25:22.083858+02:00 twentytwo systemd[1]: Stopped sddm.service -
Simple Desktop Display Manager.
2024-08-07T17:25:22.085111+02:00 twentytwo systemd[1]: Started
[email protected] - Getty on tty7.
2024-08-07T17:25:25.325658+02:00 twentytwo systemd[1]: Stopping
[email protected] - Getty on tty7...
2024-08-07T17:25:25.325953+02:00 twentytwo systemd[1]: [email protected]:
Deactivated successfully.
2024-08-07T17:25:25.326075+02:00 twentytwo systemd[1]: Stopped
[email protected] - Getty on tty7.
2024-08-07T17:25:25.328211+02:00 twentytwo systemd[1]:
run-credentials-getty\x40tty7.service.mount: Deactivated successfully.
> Without logs it will be difficult to understand what is going on. But at
> first sight I would lean towards a graphics driver bug more than an SDDM bug.
I really don't think so.
> Maybe testing with different display managers (gdm, lightdm) could also help
> narrowing down the issue ?
Yes, I still need to do this.
Greetings
Helge
--
Dr. Helge Kreutzmann [email protected]
Dipl.-Phys. http://www.helgefjell.de/debian.php
64bit GNU powered gpg signed mail preferred
Help keep free software "libre": http://www.ffii.de/
signature.asc
Description: PGP signature
Bug#1072840: Acknowledgement (sddm: Sddm crashes regularly)
Hello Aurélien, Am Sun, Jun 23, 2024 at 09:39:40AM +0200 schrieb Aurélien COUDERC: > Le 23 juin 2024 09:28:41 GMT+02:00, Helge Kreutzmann a > écrit : > >Hello Sddm maintainers, > >I analysed this annyoing bug further. The following pattern seems to > >hold always: > > can you see anything in the kernel or system logs (dmesg or journalctl) ? Yes. In syslog I see, e.g.: 2024-06-15T22:01:03.517821+02:00 twentytwo sddm[1402833]: Auth: sddm-helper (--socket /tmp/sddm-auth-e05fef53-ea3e-46c0-bea9-88063f3db194 --id 2 --start /usr/bin/sddm-greeter --socket /tmp/sddm-:0-fDrYVa --theme /usr/share/sddm/themes/debian-theme --user sddm --greeter) crashed (exit code 1) 2024-06-15T22:01:03.517853+02:00 twentytwo sddm[1402833]: Error from greeter session: "Process crashed" 2024-06-15T22:01:03.517873+02:00 twentytwo sddm[1402833]: Auth: sddm-helper exited with 1 2024-06-15T22:01:03.517907+02:00 twentytwo sddm[1402833]: Greeter stopped. SDDM::Auth::HELPER_AUTH_ERROR I also attached the journalctl output from the last week (I filtered for sddm). > Without logs it will be difficult to understand what is going on. But at > first sight I would lean towards a graphics driver bug more than an SDDM bug. Is there some docs how to trace this down? For a normal programm, using the debug version running in gdb or stracing it I know, but for X (display managers/login managers)? > Maybe testing with different display managers (gdm, lightdm) could also help > narrowing down the issue ? I can try them. > Happy hacking, You too and thanks for the feedback Greetings Helge -- Dr. Helge Kreutzmann [email protected] Dipl.-Phys. http://www.helgefjell.de/debian.php 64bit GNU powered gpg signed mail preferred Help keep free software "libre": http://www.ffii.de/ sdlog.xz Description: application/xz signature.asc Description: PGP signature
Bug#1072840: Acknowledgement (sddm: Sddm crashes regularly)
Dear Helge, Le 23 juin 2024 09:28:41 GMT+02:00, Helge Kreutzmann a écrit : >Hello Sddm maintainers, >I analysed this annyoing bug further. The following pattern seems to >hold always: can you see anything in the kernel or system logs (dmesg or journalctl) ? Without logs it will be difficult to understand what is going on. But at first sight I would lean towards a graphics driver bug more than an SDDM bug. Maybe testing with different display managers (gdm, lightdm) could also help narrowing down the issue ? Happy hacking, -- Aurélien
Bug#1072840: Acknowledgement (sddm: Sddm crashes regularly)
Hello Sddm maintainers, I analysed this annyoing bug further. The following pattern seems to hold always: 1. Start X session (various window managers, doest not seem to be relevant) 2. Start several user consoles. 3. Work for some time, switching back and forth between the virtual consoles and X 4. Log out from one (or more) user consoles. 5. Sddm crashes, sometimes directly, sometimes I can go into X (and back to the consoles) one final time. Step 3 seems to be important, because I tried to omit it, i.e. use a freshly booted system and start and close several consoles without doing something there and sddm did *not* crash. (I switched back and forth between consoles and X serveral times). Could it be ssh related? Because I use ssh a lot (both in X and on the console) and I did not use this in my tests (I can do so next time). Is there anything I should watch out for, or any other programm which might interact with sddm during step 3, which I should try to trigger to make this bug step-by-step reproducible? Thanks! Greetings Helge -- Dr. Helge Kreutzmann [email protected] Dipl.-Phys. http://www.helgefjell.de/debian.php 64bit GNU powered gpg signed mail preferred Help keep free software "libre": http://www.ffii.de/ signature.asc Description: PGP signature
