[Server-devel] IPv6 on schoolserver.laptop.org

[Daniel Jared Dominguez]

(moving this over to server-devel)

While trying to figure out which of the two it could be, I think I found
the problem (at least with the current setup). Notice the output of
ifconfig (irrelevant parts removed):

[root at schoolserver ~]# ifconfig
eth0      Link encap:Ethernet  HWaddr 00:40:63:E7:E4:22  
          inet addr:18.85.46.33  Bcast:18.85.46.255  Mask:255.255.255.0
          inet6 addr: fe80::240:63ff:fee7:e422/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1

eth1      Link encap:Ethernet  HWaddr 00:19:5B:68:A6:64  
          inet addr:172.18.0.1  Bcast:172.18.15.255  Mask:255.255.240.0
          UP BROADCAST MULTICAST  MTU:1500  Metric:1

eth2      Link encap:Ethernet  HWaddr 00:17:C4:00:00:01  
          inet6 addr: 2001:4830:2446:ff00:217:c4ff:fe00:1/64
Scope:Global
          inet6 addr: fe80::217:c4ff:fe00:1/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1

msh0      Link encap:Ethernet  HWaddr 00:17:C4:00:00:01  
          inet addr:172.18.16.1  Bcast:172.18.31.255  Mask:255.255.240.0
          inet6 addr: fe80::217:c4ff:fe00:1/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1

msh0:0    Link encap:Ethernet  HWaddr 00:17:C4:00:00:01  
          inet addr:172.31.255.254  Bcast:172.31.255.255
Mask:255.255.0.0
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1

I believe only eth0 is plugged in, but eth2 has the address
2001:4830:2446:ff00:217:c4ff:fe00:1. Obviously eth2 shouldn't be
pingable. schoolserver.laptop.org has an AAAA record for
2001:4830:2446:ff10::1, which isn't assigned to eth0.

Note that I don't remember what the old setup was, but it might be
easiest for both the wired and mesh side to have the address
2001:4830:2446:ff10::1 once we re-enable the mesh setup on schoolserver.
(It is okay for this address to be an anycast address since it is for a
router, at least according to the most recent IPv6 RFCs.)

--Jared

On Sat, 2007-06-16 at 16:21 -0400, Chris Ball wrote:
> I can recompile a kernel for F7 with the libertas driver and the fix in,
> once you know what it is.  I had a quick look at the web pages; the bug
> seems to have been bisected to:
> 
>    commit f11e6659ce9058928d73ff440f9b40a818d628ab
>       Author: David S. Miller <davem at sunset.davemloft.net>
>       Date:   Sat Mar 24 20:36:25 2007 -0700
>       [IPV6]: Fix routing round-robin locking.
> 
> but it's less clear what the fix is.  One possibility is:
> 
>    commit 7ebba6d14f8d63cad583bf1cc0330b601d5a8171
>       Author: YOSHIFUJI Hideaki <yoshfuji at linux-ipv6.org>
>       Date:   Tue May 29 01:13:24 2007 -0700
>       [PATCH] IPV6 ROUTE: No longer handle ::/0 specially.