Bug#1029388: linux-image-6.1.0-2-amd64: system does not boot up when swap partion is a file

[Salvatore Bonaccorso <carnil@debian.org>]

Control: reassign -1 dkms
Control: forcemerge 1029063 -1

Hi,

On Sun, Jan 22, 2023 at 05:17:39PM +0100, Salvatore Bonaccorso wrote:
> Control: severity -1 important
> Control: tags -1 + moreinfo
> 
> On Sun, Jan 22, 2023 at 11:12:37AM +0100, magnus wrote:
> > Package: src:linux
> > Version: 6.1.7-1
> > Severity: critical
> > Justification: breaks the whole system
> > X-Debbugs-Cc: magnus.ekdahl@gmail.com
> > 
> > Dear Maintainer,
> > 
> > *** Reporter, please consider answering these questions, where appropriate ***
> > 
> >    * What led up to the situation?
> > 
> > apt dist-upgrade from an older version of debian unstable
> > 
> >    * What exactly did you do (or not do) that was effective (or
> >      ineffective)?
> > 
> > I got some errors (unfortunatly not saved) from upgrade process. I.e. the
> > script failed ***swap problem***, I assumed that meant the installation did not
> > finish and the old kernel would boot up.
> > 
> >    * What was the outcome of this action?
> > 
> > After bootup I got kenel panic. A working workaround was to:
> > 
> > 1. boot into live debian (from the debian installer via usb)
> > 2. chroot into standard /
> > 3. comment out swapfile from /etc/fstab (from chroot)
> > 4. run apt upgrade again (this time it completed)
> > 5. reboot into standard /
> > 
> >    * What outcome did you expect instead?
> > 
> > I expect that the kernel-package:
> > 
> > 1: can handle if fstab conain an entry like: /swapfile swap swap defaults 0 0
> > 2: if the kenel installation script breaks, I expect the old kenerl to boot up.
> > If this is currently not technically possible I would expect a clear warning
> > that the system will be in a broken state if I reboot now.
> 
> Can you please try to get the respecive installation errors still from
> the log files, you might get sucessful from /var/log/apt/term.log .
> There was a similar issue reported earlier this week, but was not
> found to be reproducible. So if you can get to this information that
> would be helpful.

Got off-bug a reply from Magnus with the install log and it turns out
that it is actually the dkms issue in #1029063. Re-assigning and
merging the two bugs accordingly.

Regards,
Salvatore