vxlan: Don't assume linear buffers in error handler

In commit c3a43b9f ("vxlan: ICMP error lookup handler") I wrongly
assumed buffers from icmp_socket_deliver() would be linear. This is not
the case: icmp_socket_deliver() only guarantees we have 8 bytes of linear
data.

Eric fixed this same issue for fou and fou6 in commits 26fc181e
("fou, fou6: do not assume linear skbs") and 5355ed63 ("fou, fou6:
avoid uninit-value in gue_err() and gue6_err()").

Use pskb_may_pull() instead of checking skb->len, and take into account
the fact we later access the VXLAN header with udp_hdr(), so we also
need to sum skb_transport_header() here.
Reported-by: Guillaume Nault <gnault@redhat.com>
Fixes: c3a43b9f ("vxlan: ICMP error lookup handler")
Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/vxlan.c

@@ -1766,7 +1766,7 @@ static int vxlan_err_lookup(struct sock *sk, struct sk_buff *skb)
 	struct vxlanhdr *hdr;
 	__be32 vni;
 
-	if (skb->len < VXLAN_HLEN)
+	if (!pskb_may_pull(skb, skb_transport_offset(skb) + VXLAN_HLEN))
 		return -EINVAL;
 
 	hdr = vxlan_hdr(skb);