> My diskless NetBSD box no longer panics, it gets stuck because it thinks
> that the SCO box which gives it its IP address has portmap & NFS.
> 
> Can you configure a diskless NetBSD box so that it gets it's IP
> address from one machine and then NFS mounts a different machine?

that sounds odd.  unless i'm mistaken (and the code is in front of me,
and i used to use a three-host boot environment for many of the same
reasons...):

	NetBSD rarps to find out it's name
	NetBSD then broadcasts a 'pmapproc_callit' (indirect) message,
		to get bootparam data.
	the response contains info about the NFS server, etc.
	That info is used to determine what machine to mount.

i.e. you _should_ be able to set the bootparam info to whatever's
appropriate for the NFS root/swap files, and it should all work.

Note that the NFS boot code has significantly changed since 1.0; the
above description is for the -current sources.  in 1.0, i'm pretty
that it bootparam'd the machine that replied to the rarp request, so
you had to be running a bootparamd on that machine.  That's fixed in
the current sources.

For a while, i was using -current sources to boot a system
such that it was rarp'd from an OSF/1 machine, was bootparam'd from a
4.4BSD machine, and finally mounted root from a NetBSD machine.
(the reason for it was a bit odd; the point was/is that it should/does
work for me with current sources.  8-)



chris