I logged out, so that I was back to the lightdm display manager.
I did a restart, but it took a long time due to a timeout for
xdg-document-portal.service. This was the first time I got this
issue.

I've attached the journalctl output, starting at the logout.

On screen, there were various kernel errors, but they actually
occurred just after booting: firmware that failed to load (this
has always happened) and an issue that seems to be due to
"Lockdown: Xorg: raw io port access is restricted; see man
kernel_lockdown.7" (which was occurring with the 6.7.12-amd64
kernel too).

I don't know whether this is related, but there's an upstream bug
about this timeout:

https://github.com/flatpak/xdg-desktop-portal/issues/999

Control: tags -1 upstream fixed-upstream
Control: forwarded -1 https://github.com/flatpak/xdg-desktop-portal/issues/999

According to

https://github.com/flatpak/xdg-desktop-portal/issues/999#issuecomment-5151219119

it got fixed upstream by

https://github.com/flatpak/xdg-desktop-portal/pull/1896

which is in version 1.21.1.

The above comment also says that there is another bug with the
*same journal signature* for portal ≥ 1.21.1, libfuse < 3.18.0,
but Debian/unstable is no longer concerned (with fuse3 3.18.1-1
on 21 Dec 2025).

So I suppose that this bug should be closed as fixed in 1.22.1+ds-1
(the first fixed version in unstable, which is also the current
version).

For the first time with 1.22.1+ds-1, I got a timing issue in my
last shutdown:

[...]
Sep 01 13:22:29 cventin systemd[1711]: Stopped xdg-desktop-portal.service - Portal service.
Sep 01 13:22:29 cventin systemd[1]: Started plymouth-reboot.service - Show Plymouth Reboot Screen.
Sep 01 13:22:29 cventin systemd[1]: plymouth-switch-root-initramfs.service - Tell Plymouth To Jump To initramfs skipped, no trigger condition checks were met.
Sep 01 13:22:29 cventin (sd-pam)[2650562]: pam_unix(systemd-user:session): session closed for user lightdm
Sep 01 13:22:29 cventin systemd-logind[883]: Removed session 3016.
Sep 01 13:22:29 cventin systemd[1]: user@118.service: Deactivated successfully.
Sep 01 13:22:29 cventin systemd[1]: Stopped user@118.service - User Manager for UID 118.
Sep 01 13:22:29 cventin systemd[1]: Stopping user-runtime-dir@118.service - User Runtime Directory /run/user/118...
Sep 01 13:22:29 cventin systemd[1]: run-user-118.mount: Deactivated successfully.
Sep 01 13:22:29 cventin systemd[1]: Unmounted run-user-118.mount - /run/user/118.
Sep 01 13:22:29 cventin systemd[1]: user-runtime-dir@118.service: Deactivated successfully.
Sep 01 13:22:29 cventin systemd[1]: Stopped user-runtime-dir@118.service - User Runtime Directory /run/user/118.
Sep 01 13:22:29 cventin systemd[1]: Removed slice user-118.slice - User Slice of UID 118.
Sep 01 13:22:29 cventin systemd[1]: user-118.slice: Consumed 3.492s CPU time over 9.820s wall clock time, 70.1M memory peak.
Sep 01 13:22:42 cventin systemd[1711]: Stopped xdg-document-portal.service - flatpak document portal service.
Sep 01 13:22:42 cventin systemd[1711]: Stopping dbus.service - D-Bus User Message Bus...
Sep 01 13:22:42 cventin systemd[1711]: Stopped dbus.service - D-Bus User Message Bus.
Sep 01 13:22:42 cventin systemd[1711]: Removed slice session.slice - User Core Session Slice.
Sep 01 13:22:42 cventin systemd[1711]: session.slice: Consumed 4.413s CPU time over 1w 4d 21h 43min 866ms wall clock time, 29.8M memory peak.
[...]

(13 seconds before the "Stopped xdg-document-portal.service - flatpak
document portal service."), but this is not a timeout. I don't see why
it takes so much time, but this isn't the same bug anyway.