On Tue, 18 Mar 1997 08:43:17 +0100 (CET) 
 "Elmar Kolkman" <kolkmae@la1.apd.dec.com> wrote:

 > Hm. Does this account for diskless booting too ?
 > Because, there is no new SYS_NBOOT.

Oh, I totally forgot to mention... the uboot.lif is the "unified boot"
program.  copy that to mdec/rbootd/SYS_UBOOT and configure your rbootd
to use SYS_UBOOT.  The old PBOOT and NBOOT are gone.

 > I used both, with the same result (netbsd-gen AND netbsd-dl) : After the HP
 > got the kernel, it didn't do anything... I had a tcpdump running on the

Yes, this was because of a bug in the way the kernel load address was
computed in the boot program.  It has been fixed for the snapshot.

 > 	1) wrong bootblock. With a new miniroot and a unliffed uboot I could
 > 		try booting from tape again, or with a new SYS_NBOOT.

The HP ROM only recognizes LIF files... SYS_UBOOT, rdboot, etc. are all
LIF'ified files.

 > 	2) the new kernel doesn't support the DCM card as console device.

So, to follow up on this, there seem to be serious bugs with the DCM
as console... Scott Reynolds has ready access to hardware, and I talk
with him on a very regular basis, and he's working on the problems.
(Thanks a million, Scott!)  I have a DCM, but since it's the "reasonable"
serial port for my PPP connection (which must be up 24/7), I can't use
it for testing.

Jason R. Thorpe <thorpej@netbsd.org>
NetBSD/hp300 portmeister