Comment # 15 on bug 1020034 from Stanislav Brabec

Michael Metz-Heimburg: Looking at the output of cdinfo in the comment 5, your
CD is a non-standard CD, out of any specs. Desktops don't have to handle such
strange things.

Standard CD-Extra has audio tracks first, then gap and then data track
containing from a specially formatted ISO image with shifted start block.

Here is an example:

goblin:~ # cdinfo /dev/sr0 
CD Info 1.1 | (c) 1996-98 Gerd Knorr & Heiko Ei���feldt
__________________________________
track list (1 - 11)
 nr: msf      lba      ctrl adr  type
  1: 00:02:00 (000000) 0x0 0x1 audio
  2: 04:03:03 (018078) 0x0 0x1 audio
  3: 10:04:31 (045181) 0x0 0x1 audio
  4: 14:58:25 (067225) 0x0 0x1 audio
  5: 19:33:57 (087882) 0x0 0x1 audio
  6: 24:48:40 (111490) 0x0 0x1 audio
  7: 30:17:33 (136158) 0x0 0x1 audio
  8: 34:24:15 (154665) 0x0 0x1 audio
  9: 39:30:16 (177616) 0x0 0x1 audio
 10: 43:08:27 (193977) 0x0 0x1 audio
 11: 48:58:72 (220272) 0x4 0x1 data 
170: 52:39:68 (236843) 0x4 0x1 data  (leadout)
__________________________________
what ioctl's report
get mcn     : 
disc status : unknown (failed?)
multisession: 220272 XA
audio status: invalid
__________________________________
try to find out what sort of CD this is
session #2 starts at track 11, offset 220272, isofs size 229170
iso9660: 447 MB size, label `ALGIDANCD1                      '
CD-ROM with both Macintosh HFS and iso9660 fs
iso9660: 447 MB size, label `ALGIDANCD1                      '
XA sectors   Multisession, offset = 220272   CD-Plus/Extra   


Wolfgang Bauer: Thing made me a bit confused.

There is a big difference between output of lsblk and blkid:

- blkid shows directly output generated by blkid.

- lsblk shows all devices and its mount status, i. e. something that is
dependent on a proper work of desktop system.


And there is one more source of confusion:
/dev/.blkid.tab

Tools can reuse information cached here and then make incorrect assumptions
about the version that broken/fixed the stuff.


=> I need output of blkid from different versions. There are problably some
differences in CD drives or CDs itself that affect

dmesg -C

# now insert the CD

rm /dev/.blkid.tab
LIBBLKID_DEBUG=65535 blkid /dev/sr0 >blkid.log 2>blkid-debug.log
dmesg -c >blkid-dmesg.log


On my system with my CD-Extra, blkid is consistently broken for CD-Extra, and
both XFCE and GNOME consistently working across versions.

blkid needs to be fixed in all cases, but I suspect that it will not be
sufficient for the reporter.


I see errors in the log while no errors while scanning both audio CD and
CD-Extra and none on data CD. I am on util-linux-2.29 (which should contain
your patch), and I still see blkid broken.

Your proposed patch may help, as libblkid can be confused by data tracks placed
after the initial data session. blkid probably identifies standard data CD,
then identifies more tracks/sessions. If the first track is a data track, next
track should be next session of a multi-session CD. But its evaluation fails.
=> error