Stephen Jones wrote:

> Booting starts as normal.  The NCRs are probed and the disks on Channel
> A show up, however when the Channel B is probed it comes back with
> 'probe(scsibus2:0:0): command timeout'

per /dev/pci/ncr.c

**      Most NCR/SYMBIOS boards are delivered with a 40 Mhz clock.
**      53C860 and 53C875 rev. 1 support fast20 transfers but
**      do not have a clock doubler and so are provided with a
**      80 MHz clock. All other fast20 boards incorporate a doubler
**      and so should be delivered with a 40 MHz clock.
**      The future fast40 chips (895/895) use a 40 Mhz base clock
**      and provide a clock quadrupler (160 Mhz). The code below
**      tries to deal as cleverly as possible with all this stuff.

Also, via the intraserver.com webpage found out the whys of 
40Mhz versus 80Mhz.  I'm wondering if NetBSD is actually guessing
wrong on the 4280UE (rev A02).  In their docu they state that
HDR1 can be jumpered to double the clock so that Digital UNIX's
generic kernel can install on devices attached to it.  However, 
this rev A02 (which was in an NT server) has an HDR1 on the PCB,
but there are no posts and no traces.

I'm not exactly sure if ncr.c is losing with this board's revision or
if its another issue.

I went through 3 years of port-alpha archives and didn't see a mention
of this.  

smj