Use rt_ifp as intended.

During s2k15 we fixed this ugly 20+ years loopback hack of having a
rt_ifp different than rt_ifa->ifa_ifp.

ok millert@, bluhm@

diff --git a/sys/netinet6/nd6.c b/sys/netinet6/nd6.c
index de25010..0083b44 100644 (file)
--- a/sys/netinet6/nd6.c
+++ b/sys/netinet6/nd6.c
@@ -1,4 +1,4 @@
-/*     $OpenBSD: nd6.c,v 1.155 2015/10/01 09:10:22 mpi Exp $   */
+/*     $OpenBSD: nd6.c,v 1.156 2015/10/22 10:27:22 mpi Exp $   */
 /*     $KAME: nd6.c,v 1.280 2002/06/08 19:52:07 itojun Exp $   */
 
 /*
@@ -689,15 +689,10 @@ nd6_lookup(struct in6_addr *addr6, int create, struct ifnet *ifp,
         * route from a parent route that has the L flag (e.g. the default
         * route to a p2p interface) may have the flag, too, while the
         * destination is not actually a neighbor.
-        * XXX: we can't use rt->rt_ifp to check for the interface, since
-        *      it might be the loopback interface if the entry is for our
-        *      own address on a non-loopback interface. Instead, we should
-        *      use rt->rt_ifa->ifa_ifp, which would specify the REAL
-        *      interface.
         */
        if ((rt->rt_flags & RTF_GATEWAY) || (rt->rt_flags & RTF_LLINFO) == 0 ||
            rt->rt_gateway->sa_family != AF_LINK || rt->rt_llinfo == NULL ||
-           (ifp && rt->rt_ifa->ifa_ifp != ifp)) {
+           (ifp != NULL && rt->rt_ifp != ifp)) {
                if (create) {
                        char addr[INET6_ADDRSTRLEN];
                        nd6log((LOG_DEBUG, "%s: failed to lookup %s (if=%s)\n",