Subject: Re: tun driver (was: PPP and multicast bits)
To: None <mouse@Collatz.McRCIM.McGill.EDU>
From: Niklas Hallqvist <>
List: tech-net
Date: 01/30/1996 13:33:53
>>>>> "der Mouse" == der Mouse <mouse@Collatz.McRCIM.McGill.EDU> writes:

der Mouse> I've got code closer to what I want that already works on
der Mouse> SunOS and with minor tweaks on my NeXT.  What's people's
der Mouse> opinion, would it be better for me to port my existing code
der Mouse> to NetBSD, or (try to) bash on if_tun.c to do what I want?
der Mouse> Any chance the results of either effort would end up in the
der Mouse> tree?

I might be clueless, but to me the restriction of tun to be
POINTTOPOINT seems overly restrictive.  I'd say hack on if_tun.c, if
it won't get into the tree, I'm interested in getting the code.

BTW, this reminds me of something I want: filtering packets through a
user-level process.  By filtering I don't mean blocking, but actually
changing the contents of packets without changing the IP header.
My initial reaction was have the packets routed to the tunnel driver,
but after thinking for two more seconds I realized that this would
mean I'd get a loop.  Hmm, come to think about it, maybe a source route
will help me here.  Otherwise, have someone experimented with routing
dependent on source interface as well as destination IP?


Niklas Hallqvist       Phone: +46-(0)31-40 75 00  Home: +46-(0)31-41 93 95
Applitron Datasystem   Fax:   +46-(0)31-83 39 50  Home: +46-(0)31-41 93 96
Molndalsvagen 95       Email:     GSM:  +46-(0)70-714 10 35
S-412 63  GOTEBORG     WWW:   Here
Sweden		       IRC:   niklas (#NetBSD)    ICB:  niklas (netbsd)