[BUG] udev doesn't run blkid on cdroms with one session

[BUG] udev doesn't run blkid on cdroms with one session

[Maxim Levitsky]

A line in 60-persistent-storage.rules states:

KERNEL="sr*", ENV{ID_CDROM_MEDIA}="?*", IMPORT{program}="/sbin/blkid -o udev -p -u noraid -O $env{ID_CDROM_MEDIA_SESSION_LAST_OFFSET} $tempnode"

It seems to assume that ID_CDROM_MEDIA_SESSION_LAST_OFFSET is always
set, since otherwise blkid fails:

maxim@maxim-laptop:~$ /sbin/blkid -o udev -p -u noraid -O  /dev/sr0
The low-probe option requires a device
maxim@maxim-laptop:~$ /sbin/blkid -o udev -p -u noraid   /dev/sr0
ID_FS_LABELÍROM
ID_FS_LABEL_ENCÍROM\x20\x20\x20\x20\x20\x20\x20\x20\x20\x20\x20
ID_FS_VERSION=Joliet\x20Extension
ID_FS_TYPE=iso9660
ID_FS_USAGE=filesystem
maxim@maxim-laptop:~$ /sbin/blkid -o udev -p -u noraid -O 0  /dev/sr0
ID_FS_LABELÍROM
ID_FS_LABEL_ENCÍROM\x20\x20\x20\x20\x20\x20\x20\x20\x20\x20\x20
ID_FS_VERSION=Joliet\x20Extension
ID_FS_TYPE=iso9660
ID_FS_USAGE=filesystem
maxim@maxim-laptop:~$ 


Maybe this is a change in blkid
(btw, you notice obvious bug, the label is 'padded' with spaces, I will
report that to util-linux package maintainers)


Now, ether the rule has to be split in two, or cdrom_id changed.

Currently in cdrom_id:

	if (cd_media_session_last_offset > 0)
		printf("ID_CDROM_MEDIA_SESSION_LAST_OFFSET=%llu\n", cd_media_session_last_offset);



Now, it is impossible to use any disks, since without blkid environment,
devicekit doesn't recognize the cdrom at all.


Best regards,
Maxim Levitsky

Re: [BUG] udev doesn't run blkid on cdroms with one session

[Kay Sievers]

On Sun, Sep 20, 2009 at 03:49, Maxim Levitsky <maximlevitsky@gmail.com> wrote:
> A line in 60-persistent-storage.rules states:
>
> KERNEL="sr*", ENV{ID_CDROM_MEDIA}="?*", IMPORT{program}="/sbin/blkid -o udev -p -u noraid -O $env{ID_CDROM_MEDIA_SESSION_LAST_OFFSET} $tempnode"
>
> It seems to assume that ID_CDROM_MEDIA_SESSION_LAST_OFFSET is always
> set, since otherwise blkid fails:
>
> maxim@maxim-laptop:~$ /sbin/blkid -o udev -p -u noraid -O  /dev/sr0
> The low-probe option requires a device

In vol_id it was an optional parameter, and worked fine with an empty
option, at least when long options are used:
   { "offset", optional_argument, NULL, 'o' },

We could also make cdrom_id print the "0", that might be the simplest.

> maxim@maxim-laptop:~$ /sbin/blkid -o udev -p -u noraid   /dev/sr0
> ID_FS_LABELÍROM
> ID_FS_LABEL_ENCÍROM\x20\x20\x20\x20\x20\x20\x20\x20\x20\x20\x20
> ID_FS_VERSION=Joliet\x20Extension
> ID_FS_TYPE=iso9660
> ID_FS_USAGE=filesystem
>
> Maybe this is a change in blkid
> (btw, you notice obvious bug, the label is 'padded' with spaces, I will
> report that to util-linux package maintainers)

Maybe this is expected in the *_ENC version of the string, I'm not sure.

Kay

Re: [BUG] udev doesn't run blkid on cdroms with one session

[Maxim Levitsky]

On Sun, 2009-09-20 at 06:00 +0200, Kay Sievers wrote: 
> On Sun, Sep 20, 2009 at 03:49, Maxim Levitsky <maximlevitsky@gmail.com> wrote:
> > A line in 60-persistent-storage.rules states:
> >
> > KERNEL="sr*", ENV{ID_CDROM_MEDIA}="?*", IMPORT{program}="/sbin/blkid -o udev -p -u noraid -O $env{ID_CDROM_MEDIA_SESSION_LAST_OFFSET} $tempnode"
> >
> > It seems to assume that ID_CDROM_MEDIA_SESSION_LAST_OFFSET is always
> > set, since otherwise blkid fails:
> >
> > maxim@maxim-laptop:~$ /sbin/blkid -o udev -p -u noraid -O  /dev/sr0
> > The low-probe option requires a device
> 
> In vol_id it was an optional parameter, and worked fine with an empty
> option, at least when long options are used:
>    { "offset", optional_argument, NULL, 'o' },
> 
> We could also make cdrom_id print the "0", that might be the simplest.

Indeed,
I have reported this bug at
https://bugs.launchpad.net/ubuntu/+source/udev/+bug/431055

Martin Pitt, has commited now a fix for this problem by splitting the
rule in two according to value existence of
ID_CDROM_MEDIA_SESSION_LAST_OFFSET



> 
> > maxim@maxim-laptop:~$ /sbin/blkid -o udev -p -u noraid   /dev/sr0
> > ID_FS_LABELÍROM
> > ID_FS_LABEL_ENCÍROM\x20\x20\x20\x20\x20\x20\x20\x20\x20\x20\x20
> > ID_FS_VERSION=Joliet\x20Extension
> > ID_FS_TYPE=iso9660
> > ID_FS_USAGE=filesystem
> >
> > Maybe this is a change in blkid
> > (btw, you notice obvious bug, the label is 'padded' with spaces, I will
> > report that to util-linux package maintainers)
> 
> Maybe this is expected in the *_ENC version of the string, I'm not sure.
> 

Do you know where to report the util-linux bugs?

There was a fork, util-linux-ng, but ubuntu seems to use util-linux
naming, but it seems to be based on -ng.

As usual, filling a lanchpad bug didn't really help.

I think that _ENC means that non printable chars are printed as \xNNN,
and indeed, here I see the long row of spaces after the volume name.

I am sure this is trivial to fix, but I want to know where to report
that (launchpad sadly doesn't count...)


Best regards,
Maxim Levitsky