ipv6: do not delete previously existing ECMP routes if add fails
authorMichal Kubeček <mkubecek@suse.cz>
Mon, 18 May 2015 18:53:55 +0000 (20:53 +0200)
committerDavid S. Miller <davem@davemloft.net>
Wed, 20 May 2015 16:02:25 +0000 (12:02 -0400)
If adding a nexthop of an IPv6 multipath route fails, comment in
ip6_route_multipath() says we are going to delete all nexthops already
added. However, current implementation deletes even the routes it
hasn't even tried to add yet. For example, running

  ip route add 1234:5678::/64 \
      nexthop via fe80::aa dev dummy1 \
      nexthop via fe80::bb dev dummy1 \
      nexthop via fe80::cc dev dummy1

twice results in removing all routes first command added.

Limit the second (delete) run to nexthops that succeeded in the first
(add) run.

Fixes: 51ebd3181572 ("ipv6: add support of equal cost multipath (ECMP)")
Signed-off-by: Michal Kubecek <mkubecek@suse.cz>
Acked-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
net/ipv6/route.c

index d3588885f09705278b745c43b86563c85a0d3dea..3821a3517478c01739015f3075d3991040937e0b 100644 (file)
@@ -2504,9 +2504,9 @@ static int ip6_route_multipath(struct fib6_config *cfg, int add)
        int attrlen;
        int err = 0, last_err = 0;
 
+       remaining = cfg->fc_mp_len;
 beginning:
        rtnh = (struct rtnexthop *)cfg->fc_mp;
-       remaining = cfg->fc_mp_len;
 
        /* Parse a Multipath Entry */
        while (rtnh_ok(rtnh, remaining)) {
@@ -2536,6 +2536,7 @@ beginning:
                                 * next hops that have been already added.
                                 */
                                add = 0;
+                               remaining = cfg->fc_mp_len - remaining;
                                goto beginning;
                        }
                }