vol_id uuid detection regression

[Max Bowsher]

Binary package hint: udev

I've found that Jaunty's udev makes an incorrect /dev/disk/by-uuid/ symlink for one of my ext3 partitions, where Intrepid's version got the right uuid. On investigation, it turned out that the partition in question contained residue of an NTFS bootsector (rather surprisingly, since the partition in question was a manufacturer's preload of *Linux*!)

I'm aware there have been multiple bugs on this general topic previously closed as "Won't Fix", however in this case I believe there's a fairly obvious and fixable bug in the code: In my scenario, the ntfs prober did *not* detect the presence of an ntfs filesystem (i.e. vol_id --probe-all returned *only* ext3). However, if you examine the source of the ntfs prober, it overwrites the uuid field of the volume_id object long before it actually decides there's a valid filesystem there - this resulted in vol_id returning the rather bizarre combination of type=ext3, but a uuid populated by the ntfs prober.

I think the probers should only write into the result object *after* they've completed all their checks and decided that they are reporting the presence of their filesystem type.

ProblemType: Bug
Architecture: amd64
DistroRelease: Ubuntu 9.04
Package: udev 139-1
ProcEnviron:
 LC_COLLATE=C
 PATH=(custom, user)
 LANG=en_GB.UTF-8
 SHELL=/bin/bash
SourcePackage: udev
Uname: Linux 2.6.28-8-generic x86_64

[Kay Sievers (kaysievers)]

Hope this fixes it:
  http://git.kernel.org/?p=linux/hotplug/udev.git;a=commitdiff;h=1597517c9effc275b8b89c8722c808777c17173f
Thanks!

[Scott James Remnant (Canonical) (canonical-scott)]

Could you test that fix and let us know whether or not it helps?

[Max Bowsher (maxb)]

Verified fix works, attaching debdiff. (I may have been overly verbose in the changelog, feel free to trim.)

Not subscribing ubuntu-main-sponsors for now, since if Scott already has an eye on this bug, there's no need to request additional attention?

(I tried to do it via a bzr merge request as wiki.ubuntu.com/SponsorshipProcess asks, but I couldn't figure out how to build from the packaging bzr. There are files in the bzr that aren't in the archive's orig.tar.gz or diff, and vice versa.)

[Scott James Remnant (Canonical) (canonical-scott)]

On Thu, 2009-03-05 at 19:22 +0000, Max Bowsher wrote:

> Verified fix works, attaching debdiff. (I may have been overly verbose
> in the changelog, feel free to trim.)
>
> Not subscribing ubuntu-main-sponsors for now, since if Scott already has
> an eye on this bug, there's no need to request additional attention?
>
I generally prefer to only upload actual udev releases, rather than try
and pull patches out of GIT.

I'm sure there'll be one before the release anyway (with the planned
blkid migration), but if not, Kay will make one especially for us ;P
(won't you Kay :p)

Scott
--
Scott James Remnant
<email address hidden>

[Kay Sievers (kaysievers)]

Sure, any time. No problem, just let me know.

[Max Bowsher (maxb)]

Scott: Now that we're past the last Alpha and into Beta preparation, is it time to ask for that upstream release?

[Scott James Remnant (Canonical) (canonical-scott)]

On Thu, 2009-03-12 at 01:55 +0000, Max Bowsher wrote:

> Scott: Now that we're past the last Alpha and into Beta preparation, is
> it time to ask for that upstream release?
>
We're not past the last alpha yet.

Scott
--
Scott James Remnant
<email address hidden>

[Max Bowsher (maxb)]

Ok, Alpha 6 isn't completely released yet, but we're past its point on JauntyReleaseSchedule, and so I assume it's just pending final QA completion, and the window for getting bugfixes like this into it is well and truly closed.

[Scott James Remnant (Canonical) (canonical-scott)]

On Thu, 2009-03-12 at 13:43 +0000, Max Bowsher wrote:

> Ok, Alpha 6 isn't completely released yet, but we're past its point on
> JauntyReleaseSchedule, and so I assume it's just pending final QA
> completion, and the window for getting bugfixes like this into it is
> well and truly closed.
>
We use soft-freezes in Ubuntu.

In other words, the only thing stopping me uploading a new udev - and
causing the CD images to need to undergo another rebuild, and another
round of testing, etc. - is my own restraint.

Scott
--
Scott James Remnant
<email address hidden>

[Launchpad Janitor (janitor)]

This bug was fixed in the package udev - 140-1

---------------
udev (140-1) jaunty; urgency=low

  * New upstream release:
    - includes changes from 139-2
    - udevadm settle can accept a range of events to watch for.
    - vol_id: ntfs uuid setting fix. LP: #337015.

 -- Scott James Remnant <email address hidden> Fri, 13 Mar 2009 12:40:16 +0000