TomatoUSB

Hi, just another bug report, not a complaint ;-)

I'm using 6to4 anycast for IPv6 which works quite good, but I discovered a bug:
My IPv4 address changes every 24 hours (forced PPPoE disconnect from provider). When this happens I get issued a new IPv6 address which is also correctly send to the clients, but the router itself still shows the old IPv6 address in the gui, and keeps it on br0. So for example this is currently on br0:

br0 Link encap:Ethernet HWaddr XX:XX:XX:XX:XX:XX
inet addr:192.168.2.1 Bcast:192.168.2.255 Mask:255.255.255.0
inet6 addr: 2002:5de5:abe1:1::1/64 Scope:Global
inet6 addr: 2002:579f:9e0a:1::1/64 Scope:Global
…

The first inet6 addr (2002:5de5:abe1:1::1/64) is the current one, the second (2002:579f:9e0a:1::1/64) is the one from yesterday. If I then disconnect and reconnect the PPPoE connection another inet6 address is added to br0 and the old ones continue to stick around. So I guess, over a few days, I might collect a nice amount of IPv6 addresses on br0 ;-).

Also it takes a few minutes for the clients to discard the old IPv6 prefix, so for a few minutes the clients get confused when trying to connect somewhere over IPv6. I can live with this, but it would be nice (if it's possible at all) if the router would revoke the old IPv6 prefix when the wan connections goes down.

Thanks! Indeed, the old address is not getting removed from the lan iface - I'll fix that.

As for the few minutes delay, the radvd is currently set to advertise the prefix with 2 ("preferred") to 5 (max) mins lifetime for anycast tunnels (if you use latest git sources). I'm not sure if should be made event shorter - but maybe…