Subject: Re: SCSI network
To: eeh@netbsd.org, Ignatios Souvatzis <ignatios@cs.uni-bonn.de>
From: Ignatios Souvatzis <ignatios@cs.uni-bonn.de>
List: tech-kern
Date: 10/05/1999 17:18:41
On Tue, Oct 05, 1999 at 08:15:15AM -0700, Eduardo E. Horvath wrote:
> On Tue, 5 Oct 1999, Ignatios Souvatzis wrote:
> 
> > On Tue, Oct 05, 1999 at 12:33:18PM +0930, Brett Lymn wrote:
> > > 
> > > FWIW I will add my A$0.02 that this can be done.  We had a project at
> > > work that chained together a collection of MVME167 boards via their
> > > scsi ports.  The OS was an in house developed multiprocessor,
> > > multitasking message passing system which used the scsi bus as the
> > > processor interconnect.  Unfortunately, I did not have anything to do
> > > with the scsi hackery but the guys managed to get more than 8
> > > processors on the scsi bus - perhaps they used the LUN address as well.
> > 
> > Or they didn't do SCSI bus arbitration, but something else. If the mvme167
> > scsi port is low-enough-level programmable, this should be doable, but would
> > NOT help us.
> 
> Differential or wide will give 16 targets.

Differential narrow gives 16 targets???

Besides: I don't believe that wide scsi was available in the deep past for 
MVME167.
	-is

-- 
 * Progress (n.): The process through which Usenet has evolved from
   smart people in front of dumb terminals to dumb people in front of
   smart terminals.  -- obs@burnout.demon.co.uk (obscurity)