netfilter: ctnetlink: fix refcount leak on table dump

[ Upstream commit de788b2 ]

There is a reference count leak in ctnetlink_dump_table():
      if (res < 0) {
                nf_conntrack_get(&ct->ct_general); // HERE
                cb->args[1] = (unsigned long)ct;
                ...

While its very unlikely, its possible that ct == last.
If this happens, then the refcount of ct was already incremented.
This 2nd increment is never undone.

This prevents the conntrack object from being released, which in turn
keeps prevents cnet->count from dropping back to 0.

This will then block the netns dismantle (or conntrack rmmod) as
nf_conntrack_cleanup_net_list() will wait forever.

This can be reproduced by running conntrack_resize.sh selftest in a loop.
It takes ~20 minutes for me on a preemptible kernel on average before
I see a runaway kworker spinning in nf_conntrack_cleanup_net_list.

One fix would to change this to:
        if (res < 0) {
		if (ct != last)
	                nf_conntrack_get(&ct->ct_general);

But this reference counting isn't needed in the first place.
We can just store a cookie value instead.

A followup patch will do the same for ctnetlink_exp_dump_table,
it looks to me as if this has the same problem and like
ctnetlink_dump_table, we only need a 'skip hint', not the actual
object so we can apply the same cookie strategy there as well.

Fixes: d205dc4 ("[NETFILTER]: ctnetlink: fix deadlock in table dumping")
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

Author: Florian Westphal

net/netfilter/nf_conntrack_netlink.c

@@ -860,8 +860,6 @@ ctnetlink_conntrack_event(unsigned int events, const struct nf_ct_event *item)
 
 static int ctnetlink_done(struct netlink_callback *cb)
 {
-	if (cb->args[1])
-		nf_ct_put((struct nf_conn *)cb->args[1]);
 	kfree(cb->data);
 	return 0;
 }
@@ -1184,19 +1182,26 @@ static int ctnetlink_filter_match(struct nf_conn *ct, void *data)
 	return 0;
 }
 
+static unsigned long ctnetlink_get_id(const struct nf_conn *ct)
+{
+	unsigned long id = nf_ct_get_id(ct);
+
+	return id ? id : 1;
+}
+
 static int
 ctnetlink_dump_table(struct sk_buff *skb, struct netlink_callback *cb)
 {
 	unsigned int flags = cb->data ? NLM_F_DUMP_FILTERED : 0;
 	struct net *net = sock_net(skb->sk);
-	struct nf_conn *ct, *last;
+	unsigned long last_id = cb->args[1];
 	struct nf_conntrack_tuple_hash *h;
 	struct hlist_nulls_node *n;
 	struct nf_conn *nf_ct_evict[8];
+	struct nf_conn *ct;
 	int res, i;
 	spinlock_t *lockp;
 
-	last = (struct nf_conn *)cb->args[1];
 	i = 0;
 
 	local_bh_disable();
@@ -1233,7 +1238,7 @@ ctnetlink_dump_table(struct sk_buff *skb, struct netlink_callback *cb)
 				continue;
 
 			if (cb->args[1]) {
-				if (ct != last)
+				if (ctnetlink_get_id(ct) != last_id)
 					continue;
 				cb->args[1] = 0;
 			}
@@ -1246,8 +1251,7 @@ ctnetlink_dump_table(struct sk_buff *skb, struct netlink_callback *cb)
 					    NFNL_MSG_TYPE(cb->nlh->nlmsg_type),
 					    ct, true, flags);
 			if (res < 0) {
-				nf_conntrack_get(&ct->ct_general);
-				cb->args[1] = (unsigned long)ct;
+				cb->args[1] = ctnetlink_get_id(ct);
 				spin_unlock(lockp);
 				goto out;
 			}
@@ -1260,12 +1264,10 @@ ctnetlink_dump_table(struct sk_buff *skb, struct netlink_callback *cb)
 	}
 out:
 	local_bh_enable();
-	if (last) {
+	if (last_id) {
 		/* nf ct hash resize happened, now clear the leftover. */
-		if ((struct nf_conn *)cb->args[1] == last)
+		if (cb->args[1] == last_id)
 			cb->args[1] = 0;
-
-		nf_ct_put(last);
 	}
 
 	while (i) {