Conntrack leak with raw sockets

In the event a raw socket is created for sending purposes only, the creator never bothers to check the socket's receive queue. But we continue to add skbs to its queue until it fills up.

Unfortunately, if ip_conntrack is loaded on the box, each skb we add to the queue potentially holds a reference to a conntrack. If the user attempts to unload ip_conntrack, we will spin around forever since the queued skbs are pinned.

This behaviour can be witnessed in Fedora distributions which use NetworkManager. Arguably there should be an option to create a 'sending-only' socket which won't suffer from this problem, but in the interim I think the cleanest solution is to drop the conntrack reference before adding the skb to the socket's queue. The below patch does just that.

This fixes Netfilter bugzilla #91 and Redhat bugzilla #112630.

Phil

diff -ru linux-orig/net/packet/af_packet.c linux-new/net/packet/af_packet.c
@@ -274,6 +274,10 @@
 	dst_release(skb->dst);
 	skb->dst = NULL;
 
+	/* drop conntrack reference */
+	nf_conntrack_put(skb->nfct);
+	skb->nfct = NULL;
+
 	spkt = (struct sockaddr_pkt*)skb->cb;
 
 	skb_push(skb, skb->data-skb->mac.raw);

@@ -517,6 +521,10 @@
 	dst_release(skb->dst);
 	skb->dst = NULL;
 
+	/* drop conntrack reference */
+	nf_conntrack_put(skb->nfct);
+	skb->nfct = NULL;
+
 	spin_lock(&sk->sk_receive_queue.lock);
 	po->stats.tp_packets++;
 	__skb_queue_tail(&sk->sk_receive_queue, skb);

Great work tracking this down. But I fear the problem will come back haunt us with this patch. The are more places where a packet can be queued indefinitely, for example stopped qdiscs. IMO the best fix is to drop the conntrack reference once the packet leaves IP, so we don't have to make any assumptions about what will happen to the packet - this would be in ip_finish_output2(). Could you send a patch that does this? While you're at it, you could also remove this part from ip_conntrack_standalone:

#if !defined(CONFIG_IP_NF_NAT) && !defined(CONFIG_IP_NF_NAT_MODULE)
        /* Previously seen (loopback)?  Ignore.  Do this before
           fragment check. */
        if ((*pskb)->nfct)
                return NF_ACCEPT;

#endif

Regards
Patrick

Agreed.

BTW, please use nf_reset() instead of open coding this.

Cheers,

Visit Openswan at http://www.openswan.org/
Email: Herbert Xu ~{PmV>HI~} <[email protected]>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt

Something like the attached?

Phil

diff -ru linux-orig/net/ipv4/ip_output.c linux-new/net/ipv4/ip_output.c
@@ -195,6 +195,8 @@
 	nf_debug_ip_finish_output2(skb);
 #endif /*CONFIG_NETFILTER_DEBUG*/
 
+	nf_reset(skb);
+
 	if (hh) {
 		int hh_alen;
 

diff -ru linux-orig/net/ipv4/netfilter/ip_conntrack_standalone.c linux-new/net/ipv4/netfilter/ip_conntrack_standalone.c
@@ -423,13 +423,6 @@
 				        const struct net_device *out,
 				        int (*okfn)(struct sk_buff *))
 {
-#if !defined(CONFIG_IP_NF_NAT) && !defined(CONFIG_IP_NF_NAT_MODULE)
-	/* Previously seen (loopback)?  Ignore.  Do this before
-           fragment check. */
-	if ((*pskb)->nfct)
-		return NF_ACCEPT;
-#endif
-
 	/* Gather fragments. */
 	if ((*pskb)->nh.iph->frag_off & htons(IP_MF|IP_OFFSET)) {
 		*pskb = ip_ct_gather_frags(*pskb,

Perfect, thanks. I'll push it to Dave with the next batch of netfilter patches.

Regards
Patrick