On Fri, Aug 18, 2000 at 05:26:20PM -0400, Andrew Brown wrote:
> 
> it'd then have to be handling some arp stuff in the link layer which
> is probably not ideal.  arp being an upper layer and all...

Well, but at the point it knows it is passing the information to ARP,
it can check the sender address is the same in both packets without having
to handle the ARP state machine.

OTOH, if we were to pass the upper layer header to the ARP subsystem,
we would need an additional point to tell it where the sender address
is. (at least, ARP knows the length. Actually, this shouldn't be trusted,
either.)

So, maybe we should pass something like:

- pointer to hardware sender address
- hardware sender address length

Hm... the latter can be read from ifp->somethingIaddedbackthen.

Regards,
	Ignatios