net/ipv6: onlink nexthop checks should default to main table
Because of differences in how ipv4 and ipv6 handle fib lookups,
verification of nexthops with onlink flag need to default to the main
table rather than the local table used by IPv4. As it stands an
address within a connected route on device 1 can be used with
onlink on device 2. Updating the table properly rejects the route
due to the egress device mismatch.
Update the extack message as well to show it could be a device
mismatch for the nexthop spec.
Fixes: fc1e64e1092f ("net/ipv6: Add support for onlink flag")
Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
diff --git a/net/ipv6/route.c b/net/ipv6/route.c
index 69c43d2..9dcfadd 100644
--- a/net/ipv6/route.c
+++ b/net/ipv6/route.c
@@ -2479,7 +2479,7 @@ static int ip6_route_check_nh_onlink(struct net *net,
struct net_device *dev,
struct netlink_ext_ack *extack)
{
- u32 tbid = l3mdev_fib_table(dev) ? : RT_TABLE_LOCAL;
+ u32 tbid = l3mdev_fib_table(dev) ? : RT_TABLE_MAIN;
const struct in6_addr *gw_addr = &cfg->fc_gateway;
u32 flags = RTF_LOCAL | RTF_ANYCAST | RTF_REJECT;
struct rt6_info *grt;
@@ -2490,7 +2490,8 @@ static int ip6_route_check_nh_onlink(struct net *net,
if (grt) {
if (!grt->dst.error &&
(grt->rt6i_flags & flags || dev != grt->dst.dev)) {
- NL_SET_ERR_MSG(extack, "Nexthop has invalid gateway");
+ NL_SET_ERR_MSG(extack,
+ "Nexthop has invalid gateway or device mismatch");
err = -EINVAL;
}