nfs3 vs. nfs2 in RH6

[Michael Jinks <mjinks bsd uchicago edu>]

It all started when I decided to help put some sort of *n*x on our NeXT
slabs.  The Linux->NeXT port isn't terribly far along, and the developer
group that is working on it does not appear to be terribly active, so
when we ran into problems we couldn't get much help fixing, we decided
to take a crack with some other OS, at least for now.

The NetBSD port appears to be further along, so now we're trying that. 
As with Linux, NetBSD doesn't have drivers for the NeXT SCSI controller,
so we have to net boot.

Cutting to the chase, right now the place where we appear to be hung up
is: The boot loader goes to the NeXT box, the kernel downloads, and then
we get an NFS timeout.  According to the (very helpful, BTW) guys on the
NetBSD-NeXT port mailing list, this may be due to an inconsistency in
RedHat 6.0 between mountd and the kernel's NFS implementation.

The way they describe it, RH6's mountd identifies as nfs3, but the OS
only implements nfs2.  This is never a problem when mounting volumes
between Linux boxes because nobody ever actually tries to use nfs3.  But
apparently the NetBSD kernel (though not the bootloader) will try to use
nfs3 if it thinks it's available.

Hence our problem.

Does anybody know: Can I either get nfs3 working for real on my RedHat
6.0 boot server, or can I get a version of mountd that only advertises
as nfs2, or is there some other course we can take?  Or, do the NetBSD
guys not really know what they're talking about? ;)

I'll welcome any suggestions that get me closer to some sort of real
*n*x on these NeXen...

Thanks,
-m