On Sat, 31 Aug 1996 16:28:50 +0000 (GMT) 
 Tom I Helbekkmo <tih@Hamartun.Priv.NO> wrote:

 > My thanks to all who responded!
 > 
 > The problem with the Pioneer DRM-610 turns out to be that it's old
 > enough (it's from 1990) to need the LUN present in the SCSI command
 > itself.  Thanks to Michael L. Hitch, who explained this to me, I've
 > now got a local hack in the SCSI driver so that it does this for me
 > for this particular drive.
 > 
 > I'll look into making a proper "quirk" for this, and send-pr it.

Hrm... I'm currently working on some changes to fix the problem with our 
SCSI subsystem not handling targets > 7 (we have some wide controllers at 
work that will need this to work).  In the case of wide SCSI, I'm under 
the impression that the max LUN can be 63 (I need to look it up to be 
sure).

Now, it's probably the case that a wide device isn't going to need the 
LUN to be stashed in the cdb (indeed, it wouldn't fit).  So, what I was 
thinking of doing is having the MI layer stash the LUN in the cdb in the 
case of the scsibus's "maxlun" property being 7 ... some of our SCSI 
controller drivers do this already ... it should be done in the MI layer, 
if done at all... and it sounds like it's necessary ... (Gee, how many 
devices are in the quirk table unfairly? :-)

SCSI-2 doesn't require this, but it is allowed for compatibility with 
legacy devices, IIRC.

So, would anyone object to having the MI layer stash the LUN in the cdb 
before calling the controller entry point (when the bus's "maxlun" 
property allows it to fit)?

 -- save the ancient forests - http://www.bayarea.net/~thorpej/forest/ -- 
Jason R. Thorpe                                       thorpej@nas.nasa.gov
NASA Ames Research Center                               Home: 408.866.1912
NAS: M/S 258-6                                          Work: 415.604.0935
Moffett Field, CA 94035                                Pager: 415.428.6939