> > >Speaking for cdrecord, the change not really needed as no writer uses > > >something differnt from lun 0. > > > > > Does this mean that if a real SCSI burner is used with a LUN of >0 it > > doesn't work with your software, or some part of the standard says it > > must be zero, or you have never seen one jumperd for non-zero, or ??? > > I would call it (close to) impossible to jumper the LUN of such a SCSI > device at all! ;-) > > One can imagine burners w/ disc change mechanism,
One should rather imagine e.g. FCAL tower enclosure for multiple off-the-shelf units. Host can address individual units by LUN, but it (addressing) is done at lower communication layer, not in command data block. I mean if one would project SCSI-3 on ISO 7 layer networking model and put command block at transport layer, LUN addressing would end up at network, if not link layer. As for disc changing units. MMC specification provides for this through MECHANISM STATUS(BDh) and "slotted" LOAD/UNLOAD MEDIA(A6h) commands. Media is changed explicitly by issueing latter command. A.
