> Hmm...  Have you had tcpdump running for the entire boot process (i.e. so that 
> you don't miss anything)?  

Yes, I ran both for the entire boot sequence, and right after SYS_UBOOT
I could see nothing useful.  I also ran arpwatch on the host (I dont know
if that prints rarp requests or not, but still) and it detected nothing
also.

> IIRC, you don't get the error message:
> 
> No response for RARP request
> le(0,0,0,0): Unknown error: code 60
> boot: Unknown error: code 60

I get nothing like that, and that's why I think that the hp is not even
asking for an IP.  Even if the host's rarp daemon wasnt working, the
client should see it.  Here is the exact output:

>> NetBSD/hp300 Primary Boot, Revision 1.9
>> <scottr@polka, Fri Mar 13 18:11:29 CST 1998>
>> HP 9000/360 SPU
>> Enter "reset" to reset system.
Boot: [[[le0a:]netbsd][-a][-s][-d]] :-

And it hangs right there.
Actually, I think I was rong about the 340, it's a 360.  clerical error. 

> Yup.  The following does a good job of explaining it:
> 
> http://www.netbsd.org/Documentation/network/netboot/intro.hp300.html
> 
> Later...
> 
> Greg Oster

I read it.  So after SYS_UBOOT, it's supposed to do RARP.  Which it doesnt
seem to be doing.  Is it SYS_UBOOT which is telling the machine to look
for rarp or is that the boot-sequence of the machine itself?

-kannan