[LUNI] device identifiers in kernel error messages

[Mark Martin]

I was wondering if someone out there might be able and willing to help
me with a question regarding how the kernel names devices in messages
that it logs through syslog.  I've been receiving I/O errors and I'm
trying to identify the source of the errors.  Specifically, I am
receiving many errors such as

Nov  4 04:02:20 ergo kernel:  I/O error: dev 08:11, sector 25165872
Nov  4 04:02:20 ergo kernel: EXT2-fs error (device sd(8,17)): 
ext2_read_inode: unable to read inode block - inode=1569795, block=3145734

Am I correct in thinking that 08:11 in the first message refers to the
device with major number 8 and minor number 11?  Similarly, does the
device name in the second message sd(8,17) refer to major number 8 and
minor number 17?  Major number 8, refering to a SCSI hard drive, makes
sense in both cases and minor number 17, narrowing it down to /dev/sdb1,
makes sense in the second case.  But minor number 11, which refers to
/dev/sda11, doesn't make any sense for my system because the only
partitions on sda are partitions 1, 2, and 3.

I would very much appreciate help with interpreting these messages.  If
one of the error messages does not refer to major and minor numbers, a
reference where I could read about the naming convention used would also
be appreciated.  I looked through the kernel sources and documentation
and a little in LDP but maybe I didn't look in exactly the right place.

Many thanks,

Mark

P.S. Does anybody know if these kinds of SCSI problems are characteristic of 
kernel 2.4.8?  It might not be a SCSI-only problem either.  We've seen the 
same types of errors for ide drives with kernel 2.4.8.  Of course, it might 
be that we just have a couple of bad disks.