Re: if_bridge and IPv6?

From: Bruce A. Mah (bmah_at_freebsd.org)
Date: 08/30/05

  • Next message: Bart Van Kerckhove: "Re: Feature requests / inquiries." 
    To: Andrew Thompson <thompsa@freebsd.org>
Date: Mon, 29 Aug 2005 22:24:55 -0700

    If memory serves me right, Andrew Thompson wrote:
    > On Mon, Aug 29, 2005 at 06:19:42PM -0700, Bruce A. Mah wrote:

    > > Bridging of IPv6 packets going between two hosts on either side of the
    > > bridge appears to work just fine (to the point where I can do ping6 and
    > > ssh across the bridge).
    >
    > Good to hear something works :)

    Well, just for the record (maybe I mentioned this before) ARP and
    unicast IPv4 work as advertised. So yeah, something works. :-)

    > > > PS. Food for thought: Should bridge devices have IFF_MULTICAST set?
    > > > Mine doesn't.
    > > >
    > > > hornet# ifconfig bridge0
    > > > bridge0: flags=41<UP,RUNNING> mtu 1500
    > > > inet (foo) netmask 0xffffff00
    > > > inet6 (bar) prefixlen 64
    > > > ether ac:de:48:cd:e2:32
    > > > priority 32768 hellotime 2 fwddelay 15 maxage 20
    > > > member: sis1 flags=3<LEARNING,DISCOVER>
    > > > member: sis0 flags=3<LEARNING,DISCOVER>
    > >
    > > This looks more and more suspicious to me. IPv6 ND requires multicast
    > > to work and there are several explicit checks for IFF_MULTICAST in the
    > > IPv6 output path and neighbor discovery code.
    >
    > I think you are right, i'll look into it.

    Thanks! If I had some hacking time right now, my first attempt would be
    to stick this...

            ifp->if_flags |= IFF_MULTICAST;

    ...somewhere in bridge_clone_create(). Given how little I actually have
    worked with FreeBSD's network drivers, this would really be the "Hail
    Mary" approach and I wasn't feeling quite up to it this evening... :-p

    Cheers,

    Bruce.

  • Next message: Bart Van Kerckhove: "Re: Feature requests / inquiries."
    • Quantcast