ipv6: blackhole_netdev needs snmp6 counters

Whenever rt6_uncached_list_flush_dev() swaps rt->rt6_idev
to the blackhole device, parts of IPv6 stack might still need
to increment one SNMP counter.

Root cause, patch from Ido, changelog from Eric :)

This bug suggests that we need to audit rt->rt6_idev usages
and make sure they are properly using RCU protection.

Fixes: e5f80fc ("ipv6: give an IPv6 dev to blackhole_netdev")
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: syzbot <syzkaller@googlegroups.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net/ipv6/addrconf.c

@@ -400,16 +400,16 @@ static struct inet6_dev *ipv6_add_dev(struct net_device *dev)
 	/* We refer to the device */
 	dev_hold_track(dev, &ndev->dev_tracker, GFP_KERNEL);
 
-	if (dev != blackhole_netdev) {
-		if (snmp6_alloc_dev(ndev) < 0) {
-			netdev_dbg(dev, "%s: cannot allocate memory for statistics\n",
-				   __func__);
-			neigh_parms_release(&nd_tbl, ndev->nd_parms);
-			dev_put_track(dev, &ndev->dev_tracker);
-			kfree(ndev);
-			return ERR_PTR(err);
-		}
+	if (snmp6_alloc_dev(ndev) < 0) {
+		netdev_dbg(dev, "%s: cannot allocate memory for statistics\n",
+			   __func__);
+		neigh_parms_release(&nd_tbl, ndev->nd_parms);
+		dev_put_track(dev, &ndev->dev_tracker);
+		kfree(ndev);
+		return ERR_PTR(err);
+	}
 
+	if (dev != blackhole_netdev) {
 		if (snmp6_register_dev(ndev) < 0) {
 			netdev_dbg(dev, "%s: cannot create /proc/net/dev_snmp6/%s\n",
 				   __func__, dev->name);