When disconnecting from a VPN (in my case VPNC), NetworkManager flushes
all addresses from the physical device. Tracing it showed it doing:
ip route del default
ip addr flush dev eth1
ip route flush dev eth1
ip neigh flush all
This of course also blows away all the IPv6 addresses, including the
link-local address required for stateless autoconf. The attached patch
against the 0.7.0 svn3138 snapshot used in Fedora 8 simply re-adds the
IPv6 link-local address after taking down the VPN connection. This
allows IPv6 autoconf to do its magic and restore connectivity.
This problem is also reported & unresolved in the Fedora BZ
https://bugzilla.redhat.com/show_bug.cgi?id=233757
Regards,
Dan.
--
|=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=|
|=- Perl modules: http://search.cpan.org/~danberr/ -=|
|=- Projects: http://freshmeat.net/~danielpb/ -=|
|=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=|
diff -rup NetworkManager-0.7.0.orig/src/vpn-manager/nm-vpn-connection.c
NetworkManager-0.7.0.new/src/vpn-manager/nm-vpn-connection.c
--- NetworkManager-0.7.0.orig/src/vpn-manager/nm-vpn-connection.c
2007-11-14 11:00:59.000000000 -0500
+++ NetworkManager-0.7.0.new/src/vpn-manager/nm-vpn-connection.c
2007-12-15 13:31:19.000000000 -0500
@@ -727,6 +727,7 @@ connection_state_changed (NMVPNConnectio
/* Reset routes, nameservers, and domains of the
currently active device */
nm_system_device_set_from_ip4_config (priv->parent_dev);
+ nm_system_device_add_ip6_link_address
(priv->parent_dev);
}
if (priv->banner) {
_______________________________________________
NetworkManager-list mailing list
[email protected]
http://mail.gnome.org/mailman/listinfo/networkmanager-list
