Comment 17 for bug 561585

Martin Pitt (pitti) wrote :

Thanks so far! Now, some questions based on my evaluation of the recent data:

To assert that it's actually udev/linux which is wrong, in all those cases you guys actually had a blank CD media in the drive, right? (Because the straces say the opposite sometimes)

jocko, if you are sure that http://launchpadlibrarian.net/43952779/cdrom_id_151-5.trace is really from -5 (and not from -8), this would prove that the question whether it works or not does not really depend on the particular udev version, but rather that installing a package triggers the change.

Can you try if the cdrom_id output changes if you do

  sudo apt-get install --reinstall udev

? The critical piece what to watch out for is if the cdrom_id /dev/sr0 output contains ID_CDROM_MEDIA_STATE=blank; if it does, things should work.

Also, since you both confirmed that downgrading to -5 works, does it keep working after a few reboots and unloading/reloading the media? Conversely, does -8 consistently fail even after a few reboots and media reloads? The straces seem to indicate that it sometimes works and sometimes not, with the same udev version, so we need to identify a pattern here.

Finally, I just uploaded a new udev version (-9) which fixes a bug in cdrom_id. That bug was pretty much the opposite problem, so I don't expect that it changes the behaviour for this bug here. But since this is odd enough, testing -9 can't hurt.

Thanks!