Subject: Re: question: forcing IPV6 traffic down a IPV6/IPV4 tunnel
To: Robert Elz <kre@munnari.OZ.AU>
From: Jun-ichiro itojun Hagino <itojun@iijlab.net>
List: tech-net
Date: 03/11/2000 18:07:20
>  | 	If you are wondering about IPv6 ISP does: IIJ is exchanging route
>  | 	with customer (/48), with filtering for inbonud routes.
>This is fine if your provider does it.   My IPv6 provider does static
>routing only.   Fortunately I'm not using NetBSD (or KAME) as the end
>point of a tunnel, I'm using a router that does let me set a default
>route to point to an interface (a tunnel interface in this case).

	Are you okay with having tunnel's link-local address written in
	the config file?  If so, as long as you use tunnels, you may be
	able to do
	# route add -inet6 default <gif's link-local, local side>
	since our tunnel device (gif) is IFF_POINTOPOINT.
	We can make it robust against ethernet card swaps, by
	getting gif's link-local address by using
	# route add -inet6 default `ifconfig gif0 | sed -e whatever`.

	we may be able to try changing sys/net/rtsock.c:route_output.c()
	to honor ifp setting from route(8) on RTM_ADD.
	(it seems to me that RTM_CHANGE case tries to touch uninitialized
	pointer... needs cleanup here...)

	Do you think we should try convincing ISPs to exchange routes with
	downstream using routing protocols, rather than static route
	configuration?

itojun