vti: don't allow to add the same tunnel twice

Before the patch, it was possible to add two times the same tunnel:
ip l a vti1 type vti remote 10.16.0.121 local 10.16.0.249 key 41
ip l a vti2 type vti remote 10.16.0.121 local 10.16.0.249 key 41

It was possible, because ip_tunnel_newlink() calls ip_tunnel_find() with the
argument dev->type, which was set only later (when calling ndo_init handler
in register_netdevice()). Let's set this type in the setup handler, which is
called before newlink handler.

Introduced by commit b9959fd ("vti: switch to new ip tunnel code").

CC: Cong Wang <amwang@redhat.com>
CC: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net/ipv4/ip_vti.c

@@ -337,6 +337,7 @@ static const struct net_device_ops vti_netdev_ops = {
 static void vti_tunnel_setup(struct net_device *dev)
 {
 	dev->netdev_ops		= &vti_netdev_ops;
+	dev->type		= ARPHRD_TUNNEL;
 	ip_tunnel_setup(dev, vti_net_id);
 }
 
@@ -348,7 +349,6 @@ static int vti_tunnel_init(struct net_device *dev)
 	memcpy(dev->dev_addr, &iph->saddr, 4);
 	memcpy(dev->broadcast, &iph->daddr, 4);
 
-	dev->type		= ARPHRD_TUNNEL;
 	dev->hard_header_len	= LL_MAX_HEADER + sizeof(struct iphdr);
 	dev->mtu		= ETH_DATA_LEN;
 	dev->flags		= IFF_NOARP;