GNU bug report logs - #52786
[aarch64] elogind 246.10 fails to start on ROCK64, breaking sway

Previous Next

Package: guix;

Reported by: "pelzflorian (Florian Pelz)" <pelzflorian <at> pelzflorian.de>

Date: Sat, 25 Dec 2021 08:42:01 UTC

Severity: normal

Done: "pelzflorian (Florian Pelz)" <pelzflorian <at> pelzflorian.de>

Bug is archived. No further changes may be made.

To add a comment to this bug, you must first unarchive it, by sending
a message to control AT debbugs.gnu.org, with unarchive 52786 in the body.
You can then email your comments to 52786 AT debbugs.gnu.org in the normal way.

Toggle the display of automated, internal messages from the tracker.

View this report as an mbox folder, status mbox, maintainer mbox


Report forwarded to bug-guix <at> gnu.org:
bug#52786; Package guix. (Sat, 25 Dec 2021 08:42:01 GMT) Full text and rfc822 format available.

Acknowledgement sent to "pelzflorian (Florian Pelz)" <pelzflorian <at> pelzflorian.de>:
New bug report received and forwarded. Copy sent to bug-guix <at> gnu.org. (Sat, 25 Dec 2021 08:42:01 GMT) Full text and rfc822 format available.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: "pelzflorian (Florian Pelz)" <pelzflorian <at> pelzflorian.de>
To: bug-guix <at> gnu.org
Subject: [aarch64] elogind 246.10 fails to start on ROCK64, breaking sway
Date: Sat, 25 Dec 2021 09:41:09 +0100
[Message part 1 (text/plain, inline)]
On a ROCK64 system-on-chip, herd start elogind fails.  This is guix
commit 87624540b486d710749ad00ef5aa427a9e5c1d0c after the
core-updates-frozen merge.  This breaks sway for me, which I start
from the Linux console.

But if I

    Revert "gnu: elogind: Update to 246.10."
    
    This reverts commit a0bf66ab9432675f86095d9a2533d7b80f58a668.

then it is working again.

(Actually on reconfigure in order to build python-dbusmock I also need
to add #:tests? #f to python-dbusmock.)

I shall try out if updating to

commit 488f1c589df00e802163af534294d93372e5c025
    services: dbus: Wait 1 minute for elogind to get ready.

helps, but the failure is immediate, so that is probably unrelated.
If it does not help, I will bisect the elogind versions.

Attached is my config.  (Note that the setup is unusual; to boot, I
copy all files referenced in /boot/extlinux/extlinux.conf from
/gnu/store to /boot/gnu/store.)

Regards,
Florian
[rock64-config.scm (application/vnd.lotus-screencam, attachment)]

Information forwarded to bug-guix <at> gnu.org:
bug#52786; Package guix. (Tue, 28 Dec 2021 15:36:02 GMT) Full text and rfc822 format available.

Message #8 received at 52786 <at> debbugs.gnu.org (full text, mbox):

From: "pelzflorian (Florian Pelz)" <pelzflorian <at> pelzflorian.de>
To: 52786 <at> debbugs.gnu.org
Subject: Re: bug#52786: [aarch64] elogind 246.10 fails to start on ROCK64,
 breaking sway
Date: Tue, 28 Dec 2021 16:35:23 +0100
elogind commit 7db52c01ed07f543f8272ea9a726cb542e771595 is the first
elogind version that does not launch, but it is too entangled to
simply revert.  I will take another look tomorrow.

On Sat, Dec 25, 2021 at 09:41:09AM +0100, pelzflorian (Florian Pelz) wrote:
> I shall try out if updating to
> 
> commit 488f1c589df00e802163af534294d93372e5c025
>     services: dbus: Wait 1 minute for elogind to get ready.
> 
> helps, but the failure is immediate, so that is probably unrelated.

I have not yet tried because the check phase of the guix package had failed.

Regards,
Florian






Information forwarded to bug-guix <at> gnu.org:
bug#52786; Package guix. (Sat, 01 Jan 2022 11:48:02 GMT) Full text and rfc822 format available.

Message #11 received at 52786 <at> debbugs.gnu.org (full text, mbox):

From: "pelzflorian (Florian Pelz)" <pelzflorian <at> pelzflorian.de>
To: 52786 <at> debbugs.gnu.org
Subject: Re: bug#52786: [aarch64] elogind 246.10 fails to start on ROCK64,
 breaking sway
Date: Sat, 1 Jan 2022 12:47:04 +0100
[Message part 1 (text/plain, inline)]
On Tue, Dec 28, 2021 at 04:35:23PM +0100, pelzflorian (Florian Pelz) wrote:
> elogind commit 7db52c01ed07f543f8272ea9a726cb542e771595 is the first
> elogind version that does not launch, but it is too entangled to
> simply revert.  I will take another look tomorrow.

This was a dead end.  With 7db52c01ed07f543f8272ea9a726cb542e771595,
elogind did actually launch but shepherd did not recognize it.  The
attached patch was required, but it does not make a difference for
starting sway.  (Neither is it an alternative fix for
<https://issues.guix.gnu.org/issue/52051>.)

Elogind’s entire v246-stable branch does not run sway anymore on
ROCK64.  I don’t think bisecting is a good way to find the cause.

The sway errors are:

00:00:00.013 [ERROR] [wlr] [libseat] [libseat/backend/seatd.c:78] Could not connect to socket /run/seatd.sock: No such file or directory
00:00:00.031 [ERROR] [wlr] [libseat] [libseat/backend/logind.c:267] Could not activate session: Action org.freedesktop.login1.chvt is not registered
00:00:00.031 [ERROR] [wlr] [libseat] [libseat/libseat.c:73] No backend was able to open a seat
00:00:00.031 [ERROR] [wlr] [backend/session/session.c:84] Unable to create seat: Function not implemented
00:00:00.031 [ERROR] [wlr] [backend/session/session.c:218] Failed to load session backend
00:00:00.031 [ERROR] [wlr] [backend/backend.c:353] Failed to start a DRM session
00:00:00.031 [ERROR] [sway/server.c:53] Unable to create backend

The first line seems harmless and is also printed on old elogind
before sway starts.

I will try to actually understand what is happening.

Regards,
Florian
[TEST-elogind-Watch-PID-file.patch (text/plain, attachment)]

Information forwarded to bug-guix <at> gnu.org:
bug#52786; Package guix. (Mon, 10 Jan 2022 17:17:01 GMT) Full text and rfc822 format available.

Message #14 received at 52786 <at> debbugs.gnu.org (full text, mbox):

From: "pelzflorian (Florian Pelz)" <pelzflorian <at> pelzflorian.de>
To: 52786 <at> debbugs.gnu.org
Subject: Re: bug#52786: [aarch64] elogind 246.10 fails to start on ROCK64,
 breaking sway
Date: Mon, 10 Jan 2022 18:16:03 +0100
[Message part 1 (text/plain, inline)]
Reverting elogind commit 150d7b05074f16db70a2872765edbb39066d80af
makes sway start again.  See attached patch.  We could graft elogind
or put it on core-updates, but I have not yet properly tested nor
understood.  I will test and try understanding now.

Regards,
Florian
[gnu-elogind-Revert-changes-to-logind-waiting-for-pol.patch (text/plain, attachment)]

Reply sent to "pelzflorian (Florian Pelz)" <pelzflorian <at> pelzflorian.de>:
You have taken responsibility. (Wed, 22 Jun 2022 09:00:03 GMT) Full text and rfc822 format available.

Notification sent to "pelzflorian (Florian Pelz)" <pelzflorian <at> pelzflorian.de>:
bug acknowledged by developer. (Wed, 22 Jun 2022 09:00:03 GMT) Full text and rfc822 format available.

Message #19 received at 52786-done <at> debbugs.gnu.org (full text, mbox):

From: "pelzflorian (Florian Pelz)" <pelzflorian <at> pelzflorian.de>
To: 52786-done <at> debbugs.gnu.org
Subject: Re: bug#52786: [aarch64] elogind 246.10 fails to start on ROCK64,
 breaking sway
Date: Wed, 22 Jun 2022 10:59:50 +0200
On Mon, Jan 10, 2022 at 06:16:03PM +0100, pelzflorian (Florian Pelz) wrote:
> Reverting elogind commit 150d7b05074f16db70a2872765edbb39066d80af
> makes sway start again.  See attached patch.  We could graft elogind
> or put it on core-updates, but I have not yet properly tested nor
> understood.  I will test and try understanding now.

Closing, because:

* muradm added seatd-service-type and using (service
  seatd-service-type) in lieu of (elogind-service) works.  It looks
  like a better (?) workaround than making elogind work.

* I have not understood and do not plan to debug further.  Maybe that
  I can no longer start sway by typing sway in a virtual terminal is
  even a deliberate change in logind, but then again I don’t see it on
  x86_64.

Regards,
Florian




Information forwarded to bug-guix <at> gnu.org:
bug#52786; Package guix. (Wed, 22 Jun 2022 12:27:02 GMT) Full text and rfc822 format available.

Message #22 received at 52786-done <at> debbugs.gnu.org (full text, mbox):

From: Josselin Poiret <dev <at> jpoiret.xyz>
To: "pelzflorian (Florian Pelz)" <pelzflorian <at> pelzflorian.de>,
 52786-done <at> debbugs.gnu.org
Subject: Re: bug#52786: [aarch64] elogind 246.10 fails to start on ROCK64,
 breaking sway
Date: Wed, 22 Jun 2022 14:26:38 +0200
Hello,
"pelzflorian (Florian Pelz)" <pelzflorian <at> pelzflorian.de> writes:
> * muradm added seatd-service-type and using (service
>   seatd-service-type) in lieu of (elogind-service) works.  It looks
>   like a better (?) workaround than making elogind work.

Just FTR, seatd is a light alternative to the seat management part of
elogind, but does not replace all of its functionality!  There are a
couple of things that rely on (e)logind-specific functionality, such as
polkit (so by extension NetworkManager or UDisks for example).  It's a
good alternative, but better be aware of its shortcomings.  For a
minimal Sway-based desktop it'll work well though :)

Best,
-- 
Josselin Poiret




bug archived. Request was from Debbugs Internal Request <help-debbugs <at> gnu.org> to internal_control <at> debbugs.gnu.org. (Thu, 21 Jul 2022 11:24:10 GMT) Full text and rfc822 format available.

This bug report was last modified 1 year and 252 days ago.

Previous Next


GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.