> I would appreciate it if some of you kind souls out there would look
> at the code, and either confirm my hypothesis, or if I'm wrong, point
> out a better way.

Your hypothesis is correct at least for this case...  I don't have the
scsi spec in front of me, so some of this might be suspect.  When a
command comes into the scsi code, it's sent via scsi_gen.  If that
fails, a request sense command is sent via scsi_group0.  I don't think
that the lun is consistently in the top three bits of the second command
byte.  In fact, the comments in scsi_all.h lead me to believe that this
is an obsolete method of addressing the luns.  For test unit ready,
though, which is what's being sent here, we can add this here.  My
guess is that the drives that were misbehaving were older drives...
My suggested kludge is to use your fix, but only in the case of a
test unit ready command.

> BTW, another discovery that I made is that the code in serattach()
> (mac68k/dev/ser.c) that initializes the serial ports kills the serial
> boot echo.

Oops.  We'll have to look at that... ;-)

-allen

-- 
Allen Briggs         Find the cost of freedom / Buried in the ground
- end                Mother Earth will swallow you / Lay your body down
      killing -    
allen.briggs@vt.edu          ***** MacBSD == NetBSD/Mac *****

------------------------------------------------------------------------------