Revert "vxlan: move encapsulation warning"

This reverts commit 546c044.

Nothing prevents user from sending frames to "external" VxLAN devices.
In fact kernel itself may generate icmp chatter.

This is fine, such frames should be dropped.

The point of the "missing encapsulation" warning was that
frames with missing encap should not make it into vxlan_xmit_one().
And vxlan_xmit() drops them cleanly, so let it just do that.

Without this revert the warning is triggered by the udp_tunnel_nic.sh
test, but the minimal repro is:

$ ip link add vxlan0 type vxlan \
     	      	     group 239.1.1.1 \
		     dev lo \
		     dstport 1234 \
		     external
$ ip li set dev vxlan0 up

[  419.165981] vxlan0: Missing encapsulation instructions
[  419.166551] WARNING: CPU: 0 PID: 1041 at drivers/net/vxlan.c:2889 vxlan_xmit+0x15c0/0x1fc0 [vxlan]

Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/vxlan.c

@@ -2651,6 +2651,11 @@ static void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 			udp_sum = !(flags & VXLAN_F_UDP_ZERO_CSUM6_TX);
 		label = vxlan->cfg.label;
 	} else {
+		if (!info) {
+			WARN_ONCE(1, "%s: Missing encapsulation instructions\n",
+				  dev->name);
+			goto drop;
+		}
 		remote_ip.sa.sa_family = ip_tunnel_info_af(info);
 		if (remote_ip.sa.sa_family == AF_INET) {
 			remote_ip.sin.sin_addr.s_addr = info->key.u.ipv4.dst;
@@ -2885,10 +2890,6 @@ static netdev_tx_t vxlan_xmit(struct sk_buff *skb, struct net_device *dev)
 		    info->mode & IP_TUNNEL_INFO_TX) {
 			vni = tunnel_id_to_key32(info->key.tun_id);
 		} else {
-			if (!info)
-				WARN_ONCE(1, "%s: Missing encapsulation instructions\n",
-					  dev->name);
-
 			if (info && info->mode & IP_TUNNEL_INFO_TX)
 				vxlan_xmit_one(skb, dev, vni, NULL, false);
 			else