[systemd-devel] [PATCH] systemd-udevd.service: restore mount propagation

[Martin Pitt]

Lennart Poettering [2015-01-27 23:33 +0100]:
> > That sounds good indeed! I can sit down at qemu tomorrow and simulate
> > some CD insertions/removals, and come up with an udev rule for this.
> 
> That would be great. It would really just be a matter of setting
> SYSTEMD_READY=0 for block devices where media sense suggests that no
> media is in the drive. Probably a one line fix...

Turns out that there's no proper way to simulate that under QEMU --
that has no "eject" button, and the "eject" monitor command blocks
as long as the OS is still accessing it (i. e. it's mounted).

I borrowed the laptop from a friend which has a CD drive (I can do
this until Saturday while I'm in Belgium, FTR), and confirm the stale
mounts. I tried adding these rules to 60-cdrom_id.rules after the
IMPORT line:

  # remove a medium/eject CD: disable corresponding .mount units
  ENV{DISK_MEDIA_CHANGE}=="?*", ENV{ID_CDROM_MEDIA}!="?*", ENV{SYSTEMD_READY}="0"
  # insert a medium; undo the previous rule
  ENV{DISK_MEDIA_CHANGE}=="?*", ENV{ID_CDROM_MEDIA}=="?*", ENV{SYSTEMD_READY}="1"

These essentially make SYSTEMD_READY the negation of ID_CDROM_MEDIA,
but as the latter isn't actually 0 but disappears, and we can't
"negate" in udev without resorting to (much slower) shell hacks, we
need two rules.

In udevadm I see that this has the intended effect -- as soon as I
eject the CD, /dev/sr0 gets ENV{SYSTEMD_READY}="0". But there's still
something missing, as merely adding this property doesn't yet tell
systemd to stop the unit -- media-ubuntu-5ML.mount is still "active"
after that. systemd.device(5) also only documents the other transition
from 0 to 1, but not from 1 to 0. So do we need to teach systemd to listen
to 1 → 0 transitions too for stopping?

Martin

-- 
Martin Pitt                        | http://www.piware.de
Ubuntu Developer (www.ubuntu.com)  | Debian Developer  (www.debian.org)