netfilter: nf_conntrack_defrag: check socket type before touching nodefrag flag
authorJiri Olsa <jolsa@redhat.com>
Tue, 21 Sep 2010 21:17:34 +0000 (21:17 +0000)
committerDavid S. Miller <davem@davemloft.net>
Wed, 22 Sep 2010 20:13:34 +0000 (13:13 -0700)
we need to check proper socket type within ipv4_conntrack_defrag
function before referencing the nodefrag flag.

For example the tun driver receive path produces skbs with
AF_UNSPEC socket type, and so current code is causing unwanted
fragmented packets going out.

Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
net/ipv4/netfilter/nf_defrag_ipv4.c

index eab8de32f200af74828bbf0fb3187ff578451e9b..f3a9b42b16c620a7f569e00039beea13bc75bc3c 100644 (file)
@@ -66,9 +66,11 @@ static unsigned int ipv4_conntrack_defrag(unsigned int hooknum,
                                          const struct net_device *out,
                                          int (*okfn)(struct sk_buff *))
 {
+       struct sock *sk = skb->sk;
        struct inet_sock *inet = inet_sk(skb->sk);
 
-       if (inet && inet->nodefrag)
+       if (sk && (sk->sk_family == PF_INET) &&
+           inet->nodefrag)
                return NF_ACCEPT;
 
 #if defined(CONFIG_NF_CONNTRACK) || defined(CONFIG_NF_CONNTRACK_MODULE)