Subject: Re: generic kernel SCSI tape configuration
To: Donn Cave <donn@u.washington.edu>
From: Luke Mewburn <lukem@telstra.com.au>
List: port-pmax
Date: 07/12/1995 15:33:06
> | (taken from sys/arch/pmax/conf/GENERIC.pmax)
> | 
> | master		sii0	at nexus0 csr ?
> |   [...disks... at targets 0-4...]
> | tape		tz0	at sii0 drive 5 slave 0
> | tape		tz1	at sii0 drive 6 slave 0
> | master		asc0	at nexus0 csr ?

> I'm probably confused about something here - I don't have a DECstation
> NetBSD up for reference.

> I assume these are SCSI tape devices, since they seem to be the answer
> to a question about an Exabyte EXB8200, but on my NetBSD system the
> SCSI tape device is "st0" etc.  Does this vary from one architecture
> to another, within NetBSD?

I think you've found a rare exception to the rule. netbsd.pmax also
refers to 'rz' (instead of 'sd'). I suspect that it makes it easier
for ULTRIX types. The 'rz' naming scheme may also be a hangover
from the DECstation monitor program (it may only recognize 'rz' and
'tz', and not 'sd' and 'st' respectively.). This is all a guess..


> Is there a reason for wiring the tapes in at SCSI 5 and 6?  My config
> file says "st* at scsibus? target ? lun ? # scsi tapes" and leaves it
> at that.

Last I heard, netbsd.pmax still uses config.old, which doesn't support
the '*' notion. Again, the hardwiring could be a hangover from ULTRIX
habits (tz0 == scsi id 5, rz4 == cdrom == id 4, etc.)

BTW: the netbsd.pmax freak in the office (simonb@telstra.com.au)
just told me that he's reasonably sure that tape drives don't work
under netbsd.pmax.

-- 
Luke Mewburn <lukem@telstra.com.au>

    Don't steal. The Government (and the Banks) hate competition.  - (anon)