[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
(usagi-users 03576) Problem with RFC4193 addresses
- To: usagi-users@xxxxxxxxxxxxxx
- Subject: (usagi-users 03576) Problem with RFC4193 addresses
- From: RÃmi Denis-Courmont <rdenis@xxxxxxxxxxxxxxxx>
- Date: Sun, 18 Dec 2005 14:28:56 +0100
- Organization: SimPhalempin.Com
- Reply-to: usagi-users@xxxxxxxxxxxxxx
- User-agent: KMail/1.8.3
Hello,
I've been trying to use Unique Local Addresses as specified in RFC4193,
but whether with official kernel 2.6.14, or with 2.6.14+USAGI and
2.6.15-rc5-git7, both of which seems to have improved source selection,
ULA seems to break :
Whenever I advertise a ULA prefix with radvd (on a separate Linux-based
IPv6 router), my host starts using the ULA as default source address
instead of its global unicast public address, even to contact external
nodes. As such, iy can no longer reach external IPv6 sites.
"ip -6 addr list" returns :
1: lo: <LOOPBACK,UP> mtu 16436
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
3: eth0: <BROADCAST,MULTICAST,UP> mtu 1500 qlen 1000
inet6 fd97:ded3:8f16:0:20d:60ff:fe38:6d16/64 scope global dynamic
valid_lft 2591997sec preferred_lft 604797sec
inet6 fec0::20d:60ff:fe38:6d16/64 scope site dynamic
valid_lft 2591997sec preferred_lft 604797sec
inet6 2002:527f:c007:1234:20d:60ff:fe38:6d16/64 scope global dynamic
valid_lft 297sec preferred_lft 117sec
inet6 fe80::20d:60ff:fe38:6d16/64 scope link
valid_lft forever preferred_lft forever
I would expect the kernel choose the address that has the longest
matching prefix rather than the ULA one, but it doesn't seem to do
that. Is there any plan on supporting RFC4193 ?
--
RÃmi Denis-Courmont
http://www.simphalempin.com/home/