Well, I just received another changer device, and it also fails to
work.  :(

This one is a Sony CDL-1100.  It at least claims to be compliant with
the SCSI2 changer spec, which is a good sign.  On startup, I get this:

Sep 29 14:15:00 hideo /netbsd: ch0 at scsibus0 targ 4 lun 0: <SONY, CDL1100, 1.11> SCSI2 8/changer removable
Sep 29 14:15:01 hideo /netbsd: ch0(ahc0:4:0):  Check Condition on CDB: 0x1a 08 1d 00 18 00
Sep 29 14:15:01 hideo /netbsd:     SENSE KEY:  Illegal Request
Sep 29 14:15:01 hideo /netbsd:      ASC/ASCQ:  Illegal Field in CDB
Sep 29 14:15:01 hideo /netbsd: 
Sep 29 14:15:01 hideo /netbsd: ch0: could not sense element address page
Sep 29 14:15:01 hideo /netbsd: ch0: offline
Sep 29 14:15:01 hideo /netbsd: probe(ahc0:4:1):  Check Condition on CDB: 0x12 20 00 00 2c 00
Sep 29 14:15:01 hideo /netbsd:     SENSE KEY:  Illegal Request
Sep 29 14:15:01 hideo /netbsd:      ASC/ASCQ:  Logical Unit Not Supported
Sep 29 14:15:01 hideo /netbsd: 

Not knowing what a CDB is, I'm uncertain as to what field in it could
be extraneous.  It does seem odd, though, that it probes all eight
LUNs (I redacted the other 6 probes; they're like the probe of 0:4:1),
when the changer doesn't support multiple LUNs.  I'm going to try the
trick (which didn't work) that I tried with my last non-working
changer device, which is to hack out the bits that appear to disable
block descriptors.  We'll see if that works, but it's just a hack, and
I don't particularly like it.

If anybody else has any ideas on this, I'd be greatly appreciative.

Chris

-- 
-----------------------------------------------------cjones@math.montana.edu
Chris Jones                                          cjones@honors.montana.edu
           Mad scientist at large                    cjones@nervana.montana.edu
"Is this going to be a stand-up programming session, sir, or another bug hunt?"