@@ -2056,6 +2056,19 @@ int ip_route_use_hint(struct sk_buff *skb, __be32 daddr, __be32 saddr,
return err;
}
+/* get device for dst_alloc with local routes */
+static struct net_device *ip_rt_get_dev(struct net *net,
+ const struct fib_result *res)
+{
+ struct fib_nh_common *nhc = res->fi ? res->nhc : NULL;
+ struct net_device *dev = NULL;
+
+ if (nhc)
+ dev = l3mdev_master_dev_rcu(nhc->nhc_dev);
+
+ return dev ? : net->loopback_dev;
+}
+
/*
* NOTE. We drop all the packets that has local source
* addresses, because every properly looped back packet
@@ -2212,7 +2225,7 @@ out: return err;
}
}
- rth = rt_dst_alloc(l3mdev_master_dev_rcu(dev) ? : net->loopback_dev,
+ rth = rt_dst_alloc(ip_rt_get_dev(net, res),
flags | RTCF_LOCAL, res->type,
IN_DEV_ORCONF(in_dev, NOPOLICY), false);
if (!rth)
@@ -1384,12 +1384,37 @@ ipv4_rt_replace()
ipv4_rt_replace_mpath
}
+# checks that cached input route on VRF port is deleted
+# when VRF is deleted
+ipv4_local_rt_cache()
+{
+ run_cmd "ip addr add 10.0.0.1/32 dev lo"
+ run_cmd "ip netns add test-ns"
+ run_cmd "ip link add veth-outside type veth peer name veth-inside"
+ run_cmd "ip link add vrf-100 type vrf table 1100"
+ run_cmd "ip link set veth-outside master vrf-100"
+ run_cmd "ip link set veth-inside netns test-ns"
+ run_cmd "ip link set veth-outside up"
+ run_cmd "ip link set vrf-100 up"
+ run_cmd "ip route add 10.1.1.1/32 dev veth-outside table 1100"
+ run_cmd "ip netns exec test-ns ip link set veth-inside up"
+ run_cmd "ip netns exec test-ns ip addr add 10.1.1.1/32 dev veth-inside"
+ run_cmd "ip netns exec test-ns ip route add 10.0.0.1/32 dev veth-inside"
+ run_cmd "ip netns exec test-ns ip route add default via 10.0.0.1"
+ run_cmd "ip netns exec test-ns ping 10.0.0.1 -c 1 -i 1"
+ run_cmd "ip link delete vrf-100"
+
+ # if we do not hang test is a success
+ log_test $? 0 "Cached route removed from VRF port device"
+}
+
ipv4_route_test()
{
route_setup
ipv4_rt_add
ipv4_rt_replace
+ ipv4_local_rt_cache
route_cleanup
}
Oliver reported a use case where deleting a VRF device can hang waiting for the refcnt to drop to 0. The root cause is that the dst is allocated against the VRF device but cached on the loopback device. The use case (added to the selftests) has an implicit VRF crossing due to the ordering of the FIB rules (lookup local is before the l3mdev rule, but the problem occurs even if the FIB rules are re-ordered with local after l3mdev because the VRF table does not have a default route to terminate the lookup). The end result is is that the FIB lookup returns the loopback device as the nexthop, but the ingress device is in a VRF. The mismatch causes the dst alloc against the VRF device but then cached on the loopback. The fix is to bring the trick used for IPv6 (see ip6_rt_get_dev_rcu): pick the dst alloc device based the fib lookup result but with checks that the result has a nexthop device (e.g., not an unreachable or prohibit entry). Fixes: f5a0aab84b74 ("net: ipv4: dst for local input routes should use l3mdev if relevant") Reported-by: Oliver Herms <oliver.peter.herms@gmail.com> Signed-off-by: David Ahern <dsahern@kernel.org> --- net/ipv4/route.c | 15 +++++++++++++- tools/testing/selftests/net/fib_tests.sh | 25 ++++++++++++++++++++++++ 2 files changed, 39 insertions(+), 1 deletion(-)