Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
Bug#1038611: lightdm: Lightdm fails to start X after upgrade to 1.32.0
151 views
Skip to first unread message
Adilson dos Santos Dantas
Jun 18, 2023, 9:10:06 PM6/18/23
to
Package: lightdm
Version: 1.32.0-2
Severity: grave
Justification: renders package unusable
Dear Maintainer,
After upgrading lightdm to the latest upstream version (1.32.0-2) fails to start X session leading only to console mode.
Downgrading to 1.26.0-8, from testing, fixes this bug.
-- System Information:
Debian Release: trixie/sid
APT prefers unstable
APT policy: (500, 'unstable')
Architecture: amd64 (x86_64)
Foreign Architectures: i386
Kernel: Linux 6.3.8 (SMP w/8 CPU threads; PREEMPT)
Kernel taint flags: TAINT_PROPRIETARY_MODULE, TAINT_OOT_MODULE
Locale: LANG=pt_BR.UTF-8, LC_CTYPE=pt_BR.UTF-8 (charmap=UTF-8), LANGUAGE=pt_BR:p
t:en
Shell: /bin/sh linked to /usr/bin/dash
Init: systemd (via /run/systemd/system)
Versions of packages lightdm depends on:
ii adduser 3.134
ii dbus 1.14.8-1
ii debconf [debconf-2.0] 1.5.82
ii libaudit1 1:3.0.9-1
ii libc6 2.36-9
ii libgcrypt20 1.10.2-2
ii libglib2.0-0 2.74.6-2
ii libpam-systemd [logind] 252.11-1
ii libpam0g 1.5.2-6
ii libxcb1 1.15-1
ii libxdmcp6 1:1.1.2-3
ii lightdm-gtk-greeter [lightdm-greeter] 2.0.8-3
Versions of packages lightdm recommends:
ii xserver-xorg 1:7.7+23
Versions of packages lightdm suggests:
ii accountsservice 22.08.8-6
ii upower 0.99.20-2
ii xserver-xephyr 2:21.1.7-3
-- debconf information:
* shared/default-x-display-manager: lightdm
lightdm/daemon_name: /usr/sbin/lightdm
Carlos Laviola
Jun 20, 2023, 7:10:04 AM6/20/23
to
Package: lightdm
Followup-For: Bug #1038611
Followup-For: Bug #1038611
Can't reproduce, starts up just fine for me.
Could you perhaps include the logs from `/var/log/lightdm/´?
-- System Information:
Debian Release: trixie/sid
APT prefers stable-security
APT policy: (500, 'stable-security'), (500, 'proposed-updates'), (500, 'unstable')
Architecture: amd64 (x86_64)
Kernel: Linux 6.3.0-1-amd64 (SMP w/4 CPU threads; PREEMPT)
Locale: LANG=pt_BR.UTF-8, LC_CTYPE=pt_BR.UTF-8 (charmap=UTF-8), LANGUAGE=pt_BR:en
Shell: /bin/sh linked to /usr/bin/dash
Init: systemd (via /run/systemd/system)
LSM: AppArmor: enabled
Versions of packages lightdm depends on:
ii adduser 3.134
ii dbus 1.14.8-1
ii debconf [debconf-2.0] 1.5.82
ii libaudit1 1:3.0.9-1
ii libc6 2.36-9
ii libgcrypt20 1.10.2-2
ii libglib2.0-0 2.74.6-2
ii libpam-systemd [logind] 252.11-1
ii libpam0g 1.5.2-6
ii libxcb1 1.15-1
ii libxdmcp6 1:1.1.2-3
ii lightdm-gtk-greeter [lightdm-greeter] 2.0.8-3
Versions of packages lightdm recommends:
ii xserver-xorg 1:7.7+23
Versions of packages lightdm suggests:
ii accountsservice 22.08.8-6
ii upower 0.99.20-2
ii xserver-xephyr 2:21.1.7-3
-- debconf information:
* shared/default-x-display-manager: gdm3
lightdm/daemon_name: /usr/sbin/lightdm
Yves-Alexis Perez
Jun 21, 2023, 3:00:05 AM6/21/23
to
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256
On Wed, 2023-06-21 at 01:59 +0000, solneman33 wrote:
> I downgraded to xkb-data=2.35.1-1 lightdm=1.26.0-8 from testing repo and
> reinstalled xserver-xorg and xinit. That resolved the issue for me on both
> machines.
>
> I've never reported a bug before, my apologies if this is incorrect
> procedure.
Well, can you check by reinstalling lightdm 1.32 but making sure you have
xserver-xorg?
Regards,
- --
Yves-Alexis
-----BEGIN PGP SIGNATURE-----
iQEzBAEBCAAdFiEE8vi34Qgfo83x35gF3rYcyPpXRFsFAmSSnMIACgkQ3rYcyPpX
RFs1Hwf9HgLiSwVwW4gMRS2ajXiUXmTRcgNe/jF5libmZllhhKXrMGlq2W/Whf7q
9oIjZ8BYq2x5M3y/VVCH9vYihxFRT6D9Fpg2yw9ojFmJDt0shJuYf2ErJ/2Ymbpy
n5JO/1MXtEKbczhmu1ngc23Z9LzOxeuEiZwGJWZoFmhQDlkitD6uOWYByYA4LdA1
W+OsL7tVdE6rUMX9WPJ75mcIgp5+U8mC05Y/pH5+m40/55ZvD1mJotkt4ME6V6f7
fKggPqz8rVZukvXEiTp4oW5CGy/C4zqC5nWiPF8hY8rhZrwpEo9/VFBneEi5Ouze
vzSQCUzeQvxMbA1VdXB4EL03t3rt2w==
=Fg2s
-----END PGP SIGNATURE-----
Hash: SHA256
On Wed, 2023-06-21 at 01:59 +0000, solneman33 wrote:
> I downgraded to xkb-data=2.35.1-1 lightdm=1.26.0-8 from testing repo and
> reinstalled xserver-xorg and xinit. That resolved the issue for me on both
> machines.
>
> I've never reported a bug before, my apologies if this is incorrect
> procedure.
Well, can you check by reinstalling lightdm 1.32 but making sure you have
xserver-xorg?
Regards,
- --
Yves-Alexis
-----BEGIN PGP SIGNATURE-----
iQEzBAEBCAAdFiEE8vi34Qgfo83x35gF3rYcyPpXRFsFAmSSnMIACgkQ3rYcyPpX
RFs1Hwf9HgLiSwVwW4gMRS2ajXiUXmTRcgNe/jF5libmZllhhKXrMGlq2W/Whf7q
9oIjZ8BYq2x5M3y/VVCH9vYihxFRT6D9Fpg2yw9ojFmJDt0shJuYf2ErJ/2Ymbpy
n5JO/1MXtEKbczhmu1ngc23Z9LzOxeuEiZwGJWZoFmhQDlkitD6uOWYByYA4LdA1
W+OsL7tVdE6rUMX9WPJ75mcIgp5+U8mC05Y/pH5+m40/55ZvD1mJotkt4ME6V6f7
fKggPqz8rVZukvXEiTp4oW5CGy/C4zqC5nWiPF8hY8rhZrwpEo9/VFBneEi5Ouze
vzSQCUzeQvxMbA1VdXB4EL03t3rt2w==
=Fg2s
-----END PGP SIGNATURE-----
solneman33
Jun 25, 2023, 2:10:05 PM6/25/23
to
Yes, I was able to upgrade lightdm to 1.32 after I reinstalled xserver-xorg. I don't think it was lightdm causing the issue itself, it was due to xserver-xorg and xinit being removed during the a system upgrade. Once I had those re-installed everything is working great again.
Thanks for the follow-up.
Sent with Proton Mail secure email.
------- Original Message -------
Thanks for the follow-up.
Sent with Proton Mail secure email.
------- Original Message -------
On Wednesday, June 21st, 2023 at 12:46 AM, Yves-Alexis Perez <cor...@debian.org> wrote:
> On Wed, 2023-06-21 at 01:59 +0000, solneman33 wrote:
>
> > I downgraded to xkb-data=2.35.1-1 lightdm=1.26.0-8 from testing repo and
> > reinstalled xserver-xorg and xinit. That resolved the issue for me on both
> > machines.
> >
> > I've never reported a bug before, my apologies if this is incorrect
> > procedure.
>
>
> Well, can you check by reinstalling lightdm 1.32 but making sure you have
> xserver-xorg?
>
> Regards,
> --
> Yves-Alexis
> On Wed, 2023-06-21 at 01:59 +0000, solneman33 wrote:
>
> > I downgraded to xkb-data=2.35.1-1 lightdm=1.26.0-8 from testing repo and
> > reinstalled xserver-xorg and xinit. That resolved the issue for me on both
> > machines.
> >
> > I've never reported a bug before, my apologies if this is incorrect
> > procedure.
>
>
> Well, can you check by reinstalling lightdm 1.32 but making sure you have
> xserver-xorg?
>
> Regards,
> --
Steev Klimaszewski
Aug 8, 2023, 2:30:05 AM8/8/23
to
Hi Yves,
I'm running into a very similar issue as the original submitter,
however, when I'm running into it, I am *not* using the nVidia binary
driver, but I am using a custom 5.15.44 kernel for a Raspberry Pi.
What I found after digging for quite a while, was that, yes,
downgrading to 1.26 would start Xorg, and upgrading to 1.32 would
cause it to not start Xorg. After diffing the contents between 1.26
and 1.32, it seems that the option "logind-check-graphical" has
changed from the default of false, to a default of true.
Simply adding in
logind-check-graphical=false
under the [LightDM] heading in /etc/lightdm/lightdm.conf shows it as
starting again. This happened for me on both Pi3 and Pi4, armhf and
arm64. I'm not entirely sure why this is the case, and the kernel
hasn't changed on these devices since 2022-07-03 when we last built
the kernel for them. Perhaps the original submitter could also see if
changing that option works for them with the nVidia binary driver?
I'm running into a very similar issue as the original submitter,
however, when I'm running into it, I am *not* using the nVidia binary
driver, but I am using a custom 5.15.44 kernel for a Raspberry Pi.
What I found after digging for quite a while, was that, yes,
downgrading to 1.26 would start Xorg, and upgrading to 1.32 would
cause it to not start Xorg. After diffing the contents between 1.26
and 1.32, it seems that the option "logind-check-graphical" has
changed from the default of false, to a default of true.
Simply adding in
logind-check-graphical=false
under the [LightDM] heading in /etc/lightdm/lightdm.conf shows it as
starting again. This happened for me on both Pi3 and Pi4, armhf and
arm64. I'm not entirely sure why this is the case, and the kernel
hasn't changed on these devices since 2022-07-03 when we last built
the kernel for them. Perhaps the original submitter could also see if
changing that option works for them with the nVidia binary driver?
Adilson dos Santos Dantas
Sep 23, 2023, 9:30:05 PM9/23/23
to
I tested this solution from Steev Klimaszewski and it worked.
So we have a workaround for this problem.
Maybe it can be used with sddm. But I need to see where the configuration is to do that.
Best regards,
Yves-Alexis Perez
Oct 1, 2023, 10:30:06 AM10/1/23
to
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256
Hash: SHA256
so it might be linked to https://github.com/canonical/lightdm/issues/263
Looking at a bug linked from above
(https://github.com/canonical/lightdm/issues/165) it looks like the default
was changed in order to fix a race condition or something.
The documentation says:
# logind-check-graphical = True to on start seats that are marked as graphical
by logind
Could you check the lightdm.log and check if you have messages about seats
beeing added and whether it's graphical or not (you can add logs here). With
both value for the logind-check-graphical option.
It looks to me that there's an issue deeper in the stacks (in the NVIDIA stuff
or in the RPi graphical stuff) and maybe the seats arent't marked as graphical
or something. So it's ok to tune the option locally as a workaround, but I'm
not sure about reverting it globally.
Regards,
- --
Yves-Alexis
-----BEGIN PGP SIGNATURE-----
RFtkzAgAvSJE1yDFRCZrdI/1zGTW/SWH2KUXuxpYw8b+LrwcvkLBVwSzQhkxKkS4
rd8VUjRRXVcaPXTrPJxeKqObAAYN2iUhiFCKdYAYUxdvlIPWxOkQEf8CeLm/AG6f
rCaHMmQNZY5SFkTCQ5AGUzH38IAp3a4Sdn3E+x1xVMsiYGn6h5I/z0eDcx5135mP
omuBRUYZGnoTfsApetBOQCK7pMzUJX1QRxdaiMjLZCUEsKjwoJc/6ZaLSHB4goYQ
AXAYcrc4jOhYfv6KFqbaxEBWxR/gbdG8+YBh2u8a44KEniJgXl+T4FEKfTA1poxc
muyLJuBzqpHqGyfOvZS73TjWPVZcHw==
=Lrpk
-----END PGP SIGNATURE-----
Yves-Alexis Perez
Oct 2, 2023, 8:30:05 AM10/2/23
to
On Sun, Oct 01, 2023 at 08:58:32PM -0300, Adilson dos Santos Dantas wrote:
> Hi.
>
> Here are the logs with and without the "logind-check-graphical=false"
> option.
>
> With this opinion, a new seat is added and no seat is added when this
> option is commented.
Thanks for the log. I'm unsure if there's no seat at all or if the seat
is marked as "non graphical" in logind.
In both cases and before logging in as an user, so when lightdm is
started and a greeter is displayed (with =false) or you get the black
screen (with =true), could you check the loginctl output (as root)?
I'm not too sure where to look, but at least:
- loginctl
- loginctl list-seats
- loginctl show-seat seat0
Regards,
--
Yves-Alexis
> Hi.
>
> Here are the logs with and without the "logind-check-graphical=false"
> option.
>
> With this opinion, a new seat is added and no seat is added when this
> option is commented.
Thanks for the log. I'm unsure if there's no seat at all or if the seat
is marked as "non graphical" in logind.
In both cases and before logging in as an user, so when lightdm is
started and a greeter is displayed (with =false) or you get the black
screen (with =true), could you check the loginctl output (as root)?
I'm not too sure where to look, but at least:
- loginctl
- loginctl list-seats
- loginctl show-seat seat0
Regards,
--
Yves-Alexis
Adilson dos Santos Dantas
Oct 2, 2023, 9:00:05 AM10/2/23
to
Here are the results of 'loginctl show-seat seat0' from both 'login-check-grafical' options.
logind-check-graphical=true
loginctl show-seat seat0
Id=seat0
CanTTY=yes
CanGraphical=no
Sessions=
IdleHint=yes
IdleSinceHint=0
IdleSinceHintMonotonic=0
logind-check-graphical=false
loginctl show-seat seat0
Id=seat0
ActiveSession=c1
CanTTY=yes
CanGraphical=no
Sessions=c1
IdleHint=no
IdleSinceHint=0
IdleSinceHintMonotonic=0
loginctl show-seat seat0
Id=seat0
CanTTY=yes
CanGraphical=no
Sessions=
IdleHint=yes
IdleSinceHint=0
IdleSinceHintMonotonic=0
logind-check-graphical=false
loginctl show-seat seat0
Id=seat0
ActiveSession=c1
CanTTY=yes
CanGraphical=no
Sessions=c1
IdleHint=no
IdleSinceHint=0
IdleSinceHintMonotonic=0
Best regards,
Adilson
Yves-Alexis Perez
Oct 2, 2023, 9:40:04 AM10/2/23
to
On Mon, Oct 02, 2023 at 09:47:38AM -0300, Adilson dos Santos Dantas wrote:
> Here are the results of 'loginctl show-seat seat0' from both
> 'login-check-grafical' options.
>
> logind-check-graphical=true
> loginctl show-seat seat0
> Id=seat0
> CanTTY=yes
> CanGraphical=no
> Sessions=
> IdleHint=yes
> IdleSinceHint=0
> IdleSinceHintMonotonic=0
>
> logind-check-graphical=false
> loginctl show-seat seat0
> Id=seat0
> ActiveSession=c1
> CanTTY=yes
> CanGraphical=no
> Sessions=c1
> IdleHint=no
> IdleSinceHint=0
> IdleSinceHintMonotonic=0
>
Thanks, so that confirm the issue on your installation: the seat is
> Here are the results of 'loginctl show-seat seat0' from both
> 'login-check-grafical' options.
>
> logind-check-graphical=true
> loginctl show-seat seat0
> Id=seat0
> CanTTY=yes
> CanGraphical=no
> Sessions=
> IdleHint=yes
> IdleSinceHint=0
> IdleSinceHintMonotonic=0
>
> logind-check-graphical=false
> loginctl show-seat seat0
> Id=seat0
> ActiveSession=c1
> CanTTY=yes
> CanGraphical=no
> Sessions=c1
> IdleHint=no
> IdleSinceHint=0
> IdleSinceHintMonotonic=0
>
marked as non graphical for some reason (likely related to the graphics
drivers but honestly I'm not too sure).
At that point I don't think it's a problem in LightDM and I don't really
think it's a good idea to generalize the workaround (and divert from
upstream).
I'll ping the logind/systemd people with a summary, in case they have an idea
here.
Regards,
--
Yves-Alexis Perez
Yves-Alexis Perez
Oct 2, 2023, 3:00:04 PM10/2/23
to
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256
Hi systemd/logind maintainers.
Hash: SHA256
I'm reaching out to you about a bug (#1038611) reported in LightDM which is
likely somewhere else in the stack (maybe in logind or maybe in Xorg or the
DDX).
With the update to 1.32 some people experienced an issue where LightDM
wouldn't start. At first it seemed linked to people using NVidia binary
driver, but we also had one user on a RaspberryPi 4.
Some investigation linked that to a option in lightdm.conf whose default value
changed with 1.32. That option is `login-check-graphical` and seems to check
if the seat (reported by logind) is marked as graphical or not.
Adilson checked and indeed on their installation seat0 is not marked as
graphical:
> logind-check-graphical=true
> loginctl show-seat seat0
> Id=seat0
> CanTTY=yes
> CanGraphical=no
> Sessions=
> IdleHint=yes
> IdleSinceHint=0
> IdleSinceHintMonotonic=0
graphical system or something, but I'm not too sure how to investigate more,
so I'm adding you to the loop in case you can shine some light.
At this point I don't think it's a bug in LightDM but since I'm not sure where
the bug is, let's not reassign just yet.
Regards,
- --
Yves-Alexis
-----BEGIN PGP SIGNATURE-----
RFuqdwgAkPorc8enxSgGWMBQI+kbX+RX76fQQSxgjN/UHdfkToHFcvZxjUi8uryv
J/t6cCUrH9hrs1lcxZDLcoHy5wclbhJ0f2TroUOvtrN5Hstb8dJsr4+uPKGs1Xnz
nxfCR3GBl52+6TgPRjqq8/f09L9tOcOmRAjOtpXsnvnAABNZXOx6QWs7gzZ/IVza
P4kq+UQBT4ECicySD7cEH0y7zpcEJaLFt7Blf2SrBHo79qMb5lRfheFsJ45lunLs
uJ/ATODBMgec9DFAGyzuUbKwbRnkKYSldVlLpgA9KNxzzagUlEDCkfoBXpFE01F2
60yyKJ3I9p0f8duCDKJpY4WneKsMGQ==
=Di8i
-----END PGP SIGNATURE-----
Yves-Alexis Perez
Oct 2, 2023, 3:00:04 PM10/2/23
to
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256
Hash: SHA256
On Mon, 2023-10-02 at 15:33 +0200, Yves-Alexis Perez wrote:
> At that point I don't think it's a problem in LightDM and I don't really
> think it's a good idea to generalize the workaround (and divert from
> upstream).
Just to be sure, can you test with:
> At that point I don't think it's a problem in LightDM and I don't really
> think it's a good idea to generalize the workaround (and divert from
> upstream).
- - LightDM 1.26
- - login-check-graphical=true
Regards,
- --
Yves-Alexis
-----BEGIN PGP SIGNATURE-----
RFv38AgA5qxC4JCT1rzmu8ibemyV21C+VQIY+W+TIXw/IOBnWOmZskFeBY7VXdvY
4+simhXUdijgULROQKSHmhefO+AXSBbLXU8CPM0n89O3FIr1Fjyraff/bj+NEZUZ
FoNeSBdYI8jQ0PK5BE9zXEWgseMrCDXy3w9YmoZcliIB/gvjr2cdoeKo/9V7+1lA
jTvILl5MoaHRywdLVi7F3VAKGrhVJzUJU+CzNFFG7gFEhfQIWb9vlZ9gfO+Os7bC
+jsK91J5+hg8tPWBUGnVNNv8JjQusMUy2XBw6m0SzuUl8UhuaCZ+wmYYklryd0Hq
h7YKEzgKNsM3Bo3kQMIqS8snTCt6/w==
=MFQU
-----END PGP SIGNATURE-----
Adilson dos Santos Dantas
Oct 2, 2023, 3:20:04 PM10/2/23
to
Em seg., 2 de out. de 2023 às 15:48, Yves-Alexis Perez <cor...@debian.org> escreveu:
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256
On Mon, 2023-10-02 at 15:33 +0200, Yves-Alexis Perez wrote:
> At that point I don't think it's a problem in LightDM and I don't really
> think it's a good idea to generalize the workaround (and divert from
> upstream).
Just to be sure, can you test with:
- - LightDM 1.26
- - login-check-graphical=true
Ok. Running the command "loginctl show-seat seat0" with lightdm 1.26.0-8 and this option above:
Id=seat0
CanTTY=yes
CanGraphical=no
Sessions=
IdleHint=yes
IdleSinceHint=0
IdleSinceHintMonotonic=0
CanTTY=yes
CanGraphical=no
Sessions=
IdleHint=yes
IdleSinceHint=0
IdleSinceHintMonotonic=0
Regards,
- --
Yves-Alexis
-----BEGIN PGP SIGNATURE-----
iQEzBAEBCAAdFiEE8vi34Qgfo83x35gF3rYcyPpXRFsFAmUbEHAACgkQ3rYcyPpX
RFv38AgA5qxC4JCT1rzmu8ibemyV21C+VQIY+W+TIXw/IOBnWOmZskFeBY7VXdvY
4+simhXUdijgULROQKSHmhefO+AXSBbLXU8CPM0n89O3FIr1Fjyraff/bj+NEZUZ
FoNeSBdYI8jQ0PK5BE9zXEWgseMrCDXy3w9YmoZcliIB/gvjr2cdoeKo/9V7+1lA
jTvILl5MoaHRywdLVi7F3VAKGrhVJzUJU+CzNFFG7gFEhfQIWb9vlZ9gfO+Os7bC
+jsK91J5+hg8tPWBUGnVNNv8JjQusMUy2XBw6m0SzuUl8UhuaCZ+wmYYklryd0Hq
h7YKEzgKNsM3Bo3kQMIqS8snTCt6/w==
=MFQU
-----END PGP SIGNATURE-----
Yves-Alexis Perez
Oct 2, 2023, 3:40:06 PM10/2/23
to
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256
Hash: SHA256
On Mon, 2023-10-02 at 16:09 -0300, Adilson dos Santos Dantas wrote:
> Ok. Running the command "loginctl show-seat seat0" with lightdm 1.26.0-8 and
> this option above:
> Id=seat0
> CanTTY=yes
> CanGraphical=no
> Sessions=
> IdleHint=yes
> IdleSinceHint=0
> IdleSinceHintMonotonic=0
And does LightDM starts correctly with the greeter?
> Ok. Running the command "loginctl show-seat seat0" with lightdm 1.26.0-8 and
> this option above:
> Id=seat0
> CanTTY=yes
> CanGraphical=no
> Sessions=
> IdleHint=yes
> IdleSinceHint=0
> IdleSinceHintMonotonic=0
- --
Yves-Alexis
-----BEGIN PGP SIGNATURE-----
iQEzBAEBCAAdFiEE8vi34Qgfo83x35gF3rYcyPpXRFsFAmUbGlIACgkQ3rYcyPpX
Yves-Alexis
-----BEGIN PGP SIGNATURE-----
RFsO8QgAtbR7v03OgNuubCD1/IK8s/quSCltCg9ctR2vWLwK8sPIUkQjx6PFv7mV
nrrS/OpCU5O2ihj/7qdARhHEulLDsYwg6eosOvRccjoxYtwSdSkNR5pBU9WqWQKR
2kZqunQfcsOlMPcDbULy1QMR636f2U+oA836tUulvjBj1KS9nVoIL1yBjHdC78jY
ln1xv4x3nx+/KWbISeHfSVfNnxAXhnEH7jKyoLJ6MR6xwgon48GGpLuFqhgQpz4T
UbZ12PA0t9zC/YJn2MxFP7Z3HJlXTW5NNhzrBbDmT6F2vciWNdghePq1FGY7yMWV
xz7Us8AyE3flcZWNmHCesqr4L36O8w==
=U+h1
-----END PGP SIGNATURE-----
Adilson dos Santos Dantas
Oct 2, 2023, 7:50:04 PM10/2/23
to
Em seg., 2 de out. de 2023 às 16:30, Yves-Alexis Perez <cor...@debian.org> escreveu:
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256
On Mon, 2023-10-02 at 16:09 -0300, Adilson dos Santos Dantas wrote:
> Ok. Running the command "loginctl show-seat seat0" with lightdm 1.26.0-8 and
> this option above:
> Id=seat0
> CanTTY=yes
> CanGraphical=no
> Sessions=
> IdleHint=yes
> IdleSinceHint=0
> IdleSinceHintMonotonic=0
And does LightDM starts correctly with the greeter?
No. LightDM does not start Xorg.
- --
Yves-Alexis
-----BEGIN PGP SIGNATURE-----
iQEzBAEBCAAdFiEE8vi34Qgfo83x35gF3rYcyPpXRFsFAmUbGlIACgkQ3rYcyPpX
RFsO8QgAtbR7v03OgNuubCD1/IK8s/quSCltCg9ctR2vWLwK8sPIUkQjx6PFv7mV
nrrS/OpCU5O2ihj/7qdARhHEulLDsYwg6eosOvRccjoxYtwSdSkNR5pBU9WqWQKR
2kZqunQfcsOlMPcDbULy1QMR636f2U+oA836tUulvjBj1KS9nVoIL1yBjHdC78jY
ln1xv4x3nx+/KWbISeHfSVfNnxAXhnEH7jKyoLJ6MR6xwgon48GGpLuFqhgQpz4T
UbZ12PA0t9zC/YJn2MxFP7Z3HJlXTW5NNhzrBbDmT6F2vciWNdghePq1FGY7yMWV
xz7Us8AyE3flcZWNmHCesqr4L36O8w==
=U+h1
-----END PGP SIGNATURE-----
Steev Klimaszewski
Oct 3, 2023, 1:00:05 AM10/3/23
to
Hi Yves,
> Em seg., 2 de out. de 2023 às 09:25, Yves-Alexis Perez <cor...@debian.org> escreveu:
>>
>> On Sun, Oct 01, 2023 at 08:58:32PM -0300, Adilson dos Santos Dantas wrote:
>> > Hi.
>> >
>> > Here are the logs with and without the "logind-check-graphical=false"
>> > option.
>> >
>> > With this opinion, a new seat is added and no seat is added when this
>> > option is commented.
>>
>> Thanks for the log. I'm unsure if there's no seat at all or if the seat
>> is marked as "non graphical" in logind.
>>
>> In both cases and before logging in as an user, so when lightdm is
>> started and a greeter is displayed (with =false) or you get the black
>> screen (with =true), could you check the loginctl output (as root)?
>>
>> I'm not too sure where to look, but at least:
>>
>> - loginctl
>> - loginctl list-seats
>> - loginctl show-seat seat0
>>
>> Regards,
>> --
>> Yves-Alexis
Unfortunately, I don't have access to a Raspberry Pi at the moment to
do this testing (a user was giving me remote access while they sat at
it), but I do have one on the way.
We (Kali) have no issue with changing the configuration option locally
on our Raspberry Pi images, but it would be nice to track down why
this is. From what I've heard, debian testing images do not show the
same issue, so it very well could be a fix in a newer kernel than what
we have available to use (we can't upgrade due to the wifi firmware we
need to use with our images), but I also plan to look into if the same
occurs with the Pi5 when it shows up.
--steev
> Em seg., 2 de out. de 2023 às 09:25, Yves-Alexis Perez <cor...@debian.org> escreveu:
>>
>> On Sun, Oct 01, 2023 at 08:58:32PM -0300, Adilson dos Santos Dantas wrote:
>> > Hi.
>> >
>> > Here are the logs with and without the "logind-check-graphical=false"
>> > option.
>> >
>> > With this opinion, a new seat is added and no seat is added when this
>> > option is commented.
>>
>> Thanks for the log. I'm unsure if there's no seat at all or if the seat
>> is marked as "non graphical" in logind.
>>
>> In both cases and before logging in as an user, so when lightdm is
>> started and a greeter is displayed (with =false) or you get the black
>> screen (with =true), could you check the loginctl output (as root)?
>>
>> I'm not too sure where to look, but at least:
>>
>> - loginctl
>> - loginctl list-seats
>> - loginctl show-seat seat0
>>
>> Regards,
>> --
>> Yves-Alexis
do this testing (a user was giving me remote access while they sat at
it), but I do have one on the way.
We (Kali) have no issue with changing the configuration option locally
on our Raspberry Pi images, but it would be nice to track down why
this is. From what I've heard, debian testing images do not show the
same issue, so it very well could be a fix in a newer kernel than what
we have available to use (we can't upgrade due to the wifi firmware we
need to use with our images), but I also plan to look into if the same
occurs with the Pi5 when it shows up.
--steev
Yves-Alexis Perez
Oct 5, 2023, 5:20:04 AM10/5/23
to
On Thu, Oct 05, 2023 at 01:06:43AM +0200, Michael Biebl wrote:
> have you checked if the graphics device is properly tagged by
>
> /lib/udev/rules.d/71-seat.rules
>
> udevadm info /sys/class/drm/card0
> → E: TAGS=:seat:master-of-seat:uaccess:
>
Hi Michael,
thanks for the tip. I'm not the one experiencing the bug, I'm handling
this as LightDM maintainer. So I'm adding back the bug and the two
people who were experiencing it so they can test (especially Adilson).
Regards,
--
Yves-Alexis Perez
> have you checked if the graphics device is properly tagged by
>
> /lib/udev/rules.d/71-seat.rules
>
> udevadm info /sys/class/drm/card0
> → E: TAGS=:seat:master-of-seat:uaccess:
>
Hi Michael,
thanks for the tip. I'm not the one experiencing the bug, I'm handling
this as LightDM maintainer. So I'm adding back the bug and the two
people who were experiencing it so they can test (especially Adilson).
Regards,
--
Yves-Alexis Perez
Adilson dos Santos Dantas
Oct 5, 2023, 10:10:05 PM10/5/23
to
Hi.
I checked this command and I got this:
udevadm info /sys/class/drm/card0
Unknown device "/sys/class/drm/card0": No such device
Unknown device "/sys/class/drm/card0": No such device
So I tried to find out why this error and I discovered that the nvidia_drm module is not loaded.
After loading this module the udevadm got some information:
udevadm info /sys/class/drm/card0
P: /devices/pci0000:00/0000:00:01.0/0000:01:00.0/drm/card0
M: card0
R: 0
U: drm
T: drm_minor
D: c 226:0
N: dri/card0
L: 0
S: dri/by-path/pci-0000:01:00.0-card
E: DEVPATH=/devices/pci0000:00/0000:00:01.0/0000:01:00.0/drm/card0
E: SUBSYSTEM=drm
E: DEVNAME=/dev/dri/card0
E: DEVTYPE=drm_minor
E: MAJOR=226
E: MINOR=0
E: USEC_INITIALIZED=4363555
E: ID_PATH=pci-0000:01:00.0
E: ID_PATH_TAG=pci-0000_01_00_0
E: ID_FOR_SEAT=drm-pci-0000_01_00_0
E: DEVLINKS=/dev/dri/by-path/pci-0000:01:00.0-card
E: TAGS=:master-of-seat:uaccess:seat:
E: CURRENT_TAGS=:master-of-seat:uaccess:seat:
P: /devices/pci0000:00/0000:00:01.0/0000:01:00.0/drm/card0
M: card0
R: 0
U: drm
T: drm_minor
D: c 226:0
N: dri/card0
L: 0
S: dri/by-path/pci-0000:01:00.0-card
E: DEVPATH=/devices/pci0000:00/0000:00:01.0/0000:01:00.0/drm/card0
E: SUBSYSTEM=drm
E: DEVNAME=/dev/dri/card0
E: DEVTYPE=drm_minor
E: MAJOR=226
E: MINOR=0
E: USEC_INITIALIZED=4363555
E: ID_PATH=pci-0000:01:00.0
E: ID_PATH_TAG=pci-0000_01_00_0
E: ID_FOR_SEAT=drm-pci-0000_01_00_0
E: DEVLINKS=/dev/dri/by-path/pci-0000:01:00.0-card
E: TAGS=:master-of-seat:uaccess:seat:
E: CURRENT_TAGS=:master-of-seat:uaccess:seat:
But this module was not loaded at boot time and I found an old nvidia.conf at /etc/modprobe.d. Probably it was from an old nvidia package and it prevented this module from loading..
I removed this file and finally nvidia_drm was loaded in the next boot. It worked with lightdm and with sddm. With sddm it stopped working for more than one year until I removed nvidia.conf from modprobe.d.
Then, for me, the root cause is a nvidia_drm module not loaded due to a file in modprobe.d . I don't know if there is something similar with the RasperryPi 4.
Thanks for the help.
Best regards,
Adilson
--
To unsubscribe, send mail to 1038611-u...@bugs.debian.org.
0 new messages