On Fri, 8 Oct 1999, Thor Lancelot Simon wrote:
> BSD basically doesn't *have* an abstract layer for link-layer
> processing and when you start to look at things like PPPoE or VLANs or
> EtherChannel it really shows.

Also once could do multipath routing, MAC address takeover etc.

I'm interested in this from the FreeBSD POV and have been thinking that a
solution that abstracts the current interfaces away from the hardware
driver would be the way to go.

IPv4, IPv6, Appletalk, IPX, ISO ...
----------------------------------------------------
MAC Framing Layer	ethernet_II, 802.3, HDLC ...
Network Interface	eth, tok, fddi, sppp ...
Hardware Interface	de, fxp, xl ...

One would bind hardware interfaces to a network interface before use.  The
ability to get/set/add/delete hardware interface MAC addresses would be
needed in some case.

ifconfig could be split to provide 'ifconfig' and 'hwifconfig' or extended
to know the difference but I suspect that it would be better just to split
out the relevent bits to 'hwifconfig' to avoid all the bloat.

If we can come up with interfaces, structures and semantics that make
sense then I don't see why this solution couldn't be adopted by both
NetBSD and FreeBSD.  Its sorely needed.

-- 
| Matthew N. Dodd  | '78 Datsun 280Z | '75 Volvo 164E | FreeBSD/NetBSD  |
| winter@jurai.net |       2 x '84 Volvo 245DL        | ix86,sparc,pmax |
| http://www.jurai.net/~winter | This Space For Rent  | ISO8802.5 4ever |