> On Fri, 1 Oct 1999, der Mouse wrote:
> 
> > I'm contemplating using SCSI interfaces as a networking medium.  Does
> > anyone have any thoughts on where I should look for (a) convincing a
> > machine that it should speak on other than ID 7 or (b) convincing a
> > machine to respond as if it were a drive, ie, to respond to another
> > device on the bus attempting to contact it?  If necessary I'm prepared
> > to make the link asymmetric, with one of the hosts distinctly in
> > control of what happens, though I'd prefer to avoid that.
> 
> As others have mentioned, teaching the controller to not be ID 7 shouldn't
> be too hard. But teaching it to be a target is harder. :-)
> 
> One other note, from one of the AIX lists I was on way back when. AIX had
> the two-hosts-on-a-bus trick for quite a while. The recomended practice
> (at least on the list) was to make one host use ID 6, and the other ID 5.
> That way you didn't have anything at ID 7. Why? Because all of the POST
> (PowerOn SelfTest) code would set the SCSI controller up to be ID 7. Thus
> if you wanted one a 7 and one at 6, the one at 6 would actually be at 7
> while booting. With the normal operating #'s not at 7, things were fine as
> long as you didn't boot both machines at once.

Now here's a bit of an add-on for the ambitious.....

There *is* a protocol for negotiating parallel SCSI IDs with multiple
initiators- it's called SCAM (SCSI Configured Automagically- I kid you
not, it's even in various ANSI documents). It would be interesting to see
if this could be put into NetBSD for all these silly little bit-banger
older SCSI interfaces- that would, in fact, make a small cluster
interconnect more interesting than not. This is not particularly practical
for your Adaptec/NCR/Qlogic large f/w devices (although the Qlogic claims
to have some SCAM support), but for things like the ESP100 chip and
variants, it might be kind of cool...