> David (no sarcasm here): huh? I must be misunderstanding you, because
> that seems bizarre.  How on earth do you get two CPUs running in the
> upper-half context of the same process "easily" and "with greater
> benefit", starting from where NetBSD is now?

Because I am thinking of server data flows - where there are multiple
applications generating ther data.  The amount of processing required
for a TCP transmit should be fairly limited.  You only need to write-lock
the per-connection data to update the sequence numbers, and the MAC code
to grab a slot in the ring.  Usually a single PCI access is enough to
kick the hardware.

Oh I have written ethernet drivers for SMP systems....  Albeit ones
where saturating a 100M lan was the target.

	David

-- 
David Laight: david@l8s.co.uk