bpf: update skb->protocol in bpf_skb_net_grow
authorWillem de Bruijn <willemb@google.com>
Tue, 23 Apr 2019 18:43:48 +0000 (14:43 -0400)
committerDaniel Borkmann <daniel@iogearbox.net>
Tue, 23 Apr 2019 23:32:26 +0000 (01:32 +0200)
Some tunnels, like sit, change the network protocol of packet.
If so, update skb->protocol to match the new type.

Signed-off-by: Willem de Bruijn <willemb@google.com>
Reviewed-by: Alan Maguire <alan.maguire@oracle.com>
Acked-by: Yonghong Song <yhs@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
net/core/filter.c

index edb3a7c..2f88baf 100644 (file)
@@ -3081,6 +3081,14 @@ static int bpf_skb_net_grow(struct sk_buff *skb, u32 off, u32 len_diff,
 
                        skb_set_transport_header(skb, mac_len + nh_len);
                }
+
+               /* Match skb->protocol to new outer l3 protocol */
+               if (skb->protocol == htons(ETH_P_IP) &&
+                   flags & BPF_F_ADJ_ROOM_ENCAP_L3_IPV6)
+                       skb->protocol = htons(ETH_P_IPV6);
+               else if (skb->protocol == htons(ETH_P_IPV6) &&
+                        flags & BPF_F_ADJ_ROOM_ENCAP_L3_IPV4)
+                       skb->protocol = htons(ETH_P_IP);
        }
 
        if (skb_is_gso(skb)) {