Yu Watanabe

Unfortunately, I do not have any good way to gracefully solve the issue. With PR #24250, you can save/load the backlight level for amdbpu_bl0. But, at the same time, you...

@poettering Even if the reporter's laptop erroneously reports unnecessary backlight devices, the issue is potentially occurs on multi-GPU systems. E.g. a machine has multiple graphics cards and each has own...

> > Even if the reporter's laptop erroneously reports unnecessary backlight devices, the issue is potentially occurs on multi-GPU systems. E.g. a machine has multiple graphics cards and each has...

``` stdout: ### Testing systemd-repart --empty=create ### stderr: + set -o pipefail + [[ -e /dev/loop-control ]] + repart=/builddir/build/BUILD/systemd-251/redhat-linux-build/systemd-repart + udevadm=/builddir/build/BUILD/systemd-251/redhat-linux-build/udevadm + test -x /builddir/build/BUILD/systemd-251/redhat-linux-build/systemd-repart + test -x /builddir/build/BUILD/systemd-251/redhat-linux-build/udevadm +...

cc @Geass-LL and @jcg190701.

I changed the logic. Now, this is based on #23803.

(The failures in mkosi CIs should not be related.)

@jcg190701 Which variant?

@jcg190701 Sorry for late response. During running your test, do you see `(mount-monitor-dispatch) entered rate limit` or `(mount-monitor-dispatch) left rate limit` in journal?

@msekletar Now `mount_catchup()` is introduced to make anyway coldplug mounts with deserialized state. Also suggested assertions about `Mount.control_pid` are added. PTAL.