(usagi-users 02440) stuck with linux 2.4.21

[Miek Gieben <miekg@xxxxxxxxx>]

Hello,

i'm running linux 2.4.21 without usagi (as there is no patch yet). And i'm
trying to set up ipv6 and although al lot of stuff seems to work, I cannot
get my internal network to speak ip6. I'm really stuck here.

This is the setup. My workstation (elektron) sits behind a firewall (kasteel).
This firewall tunnels to an ipv6 gateway. This is up and running. On the
firewall ipv6 in compiled into the kernel on my workstation ipv6 is a module.

On the firewall radvd runs. I have 1:7b8:206:1003::/64 confgured for use 
on my network. eth0 is the interface to the intranet for both elektron and
kasteel.

elektron has ip address: 2001:7b8:206:1003:280:48ff:feb3:26eb
kasteel  has ip address: 2001:7b8:206:1003::1

Now the following happens. First I try to ping kasteel from elektron. Which
works:
$ ping6 2001:7b8:206:1003::1
PING 2001:7b8:206:1003::1(2001:7b8:206:1003::1) from
2001:7b8:206:1003:280:48ff:feb3:26eb : 56 data bytes
64 bytes from 2001:7b8:206:1003::1: icmp_seq=1 ttl=64 time=0.139 ms

Lets list the ip6 arp table:
$ ip -6 ne
2001:7b8:206:1003::1 dev eth0 lladdr 00:02:44:52:09:30 nud reachable
fe80::202:44ff:fe52:930 dev eth0 lladdr 00:02:44:52:09:30 nud stale

ok, the ip for kasteel is in the cache.

Now I see a radvd announcement comming by:
Router advertisement from fe80::202:44ff:fe52:930 (hoplimit 255)
Received by interface eth0
        # Note: {Min,Max}RtrAdvInterval cannot be obtained with radvdump
        AdvCurHopLimit: 64
        AdvManagedFlag: off
        AdvOtherConfigFlag: off
        AdvHomeAgentFlag: off
        AdvReachableTime: 0
        AdvRetransTimer: 0
        Prefix 2001:7b8:206:1003::/64
                AdvValidLifetime: 2592000
                AdvPreferredLifetime: 604800
                AdvOnLink: on
                AdvAutonomous: on
                AdvRouterAddr: on
        AdvSourceLLAddress: 00 02 44 52 09 30

Ok, now lets ping again:

$ ping6 2001:7b8:206:1003::1
PING 2001:7b8:206:1003::1(2001:7b8:206:1003::1) 56 data bytes

--- kopje ping statistics ---
2 packets transmitted, 0 received, 100% loss, time 1015ms

Listing the arp table again tells me why:
2001:7b8:206:1003::1 dev eth0  nud failed
fe80::202:44ff:fe52:930 dev eth0 lladdr 00:02:44:52:09:30 nud stale

indeed failed for 2001:7b8:206:1003::1. But why?

Come somebody shine some light onto this? Will the USAGI patch solve this?

Thanks,

grtz  Miek


--
:wq!