[PATCH] netfilter: Fix ip_ct_selective_cleanup(), and rename ip_ct_iterate_cleanup()

Several places use ip_ct_selective_cleanup() as a general iterator, which it
was not intended for (it takes a const ip_conntrack *).  So rename it, and
make it take a non-const argument.

Also, it missed unconfirmed connections, which aren't in the hash table.  This
introduces a potential problem for users which expect to iterate all
connections (such as the helper deletion code).  So keep a linked list of
unconfirmed connections as well.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

include/linux/netfilter_ipv4/ip_conntrack.h

@@ -283,10 +283,10 @@ extern int ip_ct_no_defrag;
 struct sk_buff *
 ip_ct_gather_frags(struct sk_buff *skb);
 
-/* Delete all conntracks which match. */
+/* Iterate over all conntracks: if iter returns true, it's deleted. */
 extern void
-ip_ct_selective_cleanup(int (*kill)(const struct ip_conntrack *i, void *data),
-			void *data);
+ip_ct_iterate_cleanup(int (*iter)(struct ip_conntrack *i, void *data),
+		      void *data);
 
 /* It's confirmed if it is, or has been in the hash table. */
 static inline int is_confirmed(struct ip_conntrack *ct)

net/ipv4/netfilter/ip_conntrack_core.c

@@ -74,6 +74,7 @@ static kmem_cache_t *ip_conntrack_cachep;
 static kmem_cache_t *ip_conntrack_expect_cachep;
 struct ip_conntrack ip_conntrack_untracked;
 unsigned int ip_ct_log_invalid;
+static LIST_HEAD(unconfirmed);
 
 DEFINE_PER_CPU(struct ip_conntrack_stat, ip_conntrack_stat);
 
@@ -302,6 +303,12 @@ destroy_conntrack(struct nf_conntrack *nfct)
 	if (ct->expecting)
 		remove_expectations(ct, 1);
 
+	/* We overload first tuple to link into unconfirmed list. */
+	if (!is_confirmed(ct)) {
+		BUG_ON(list_empty(&ct->tuplehash[IP_CT_DIR_ORIGINAL].list));
+		list_del(&ct->tuplehash[IP_CT_DIR_ORIGINAL].list);
+	}
+
 	/* Delete our master expectation */
 	if (ct->master) {
 		if (ct->master->expectant) {
@@ -412,6 +419,7 @@ __ip_conntrack_confirm(struct sk_buff *skb)
 	DEBUGP("Confirming conntrack %p\n", ct);
 
 	WRITE_LOCK(&ip_conntrack_lock);
+
 	/* See if there's one in the list already, including reverse:
            NAT could have grabbed it without realizing, since we're
            not in the hash.  If there is, we lost race. */
@@ -423,6 +431,9 @@ __ip_conntrack_confirm(struct sk_buff *skb)
 			  conntrack_tuple_cmp,
 			  struct ip_conntrack_tuple_hash *,
 			  &ct->tuplehash[IP_CT_DIR_REPLY].tuple, NULL)) {
+		/* Remove from unconfirmed list */
+		list_del(&ct->tuplehash[IP_CT_DIR_ORIGINAL].list);
+
 		list_prepend(&ip_conntrack_hash[hash],
 			     &ct->tuplehash[IP_CT_DIR_ORIGINAL]);
 		list_prepend(&ip_conntrack_hash[repl_hash],
@@ -603,6 +614,10 @@ init_conntrack(const struct ip_conntrack_tuple *tuple,
 
 		/* this is a braindead... --pablo */
 		atomic_inc(&ip_conntrack_count);
+
+		/* Overload tuple linked list to put us in unconfirmed list. */
+		list_add(&conntrack->tuplehash[IP_CT_DIR_ORIGINAL].list,
+			 &unconfirmed);
 		WRITE_UNLOCK(&ip_conntrack_lock);
 
 		if (expected->expectfn)
@@ -617,7 +632,11 @@ init_conntrack(const struct ip_conntrack_tuple *tuple,
 		CONNTRACK_STAT_INC(new);
 	}
 
-end:	atomic_inc(&ip_conntrack_count);
+end:
+	/* Overload tuple linked list to put us in unconfirmed list. */
+	list_add(&conntrack->tuplehash[IP_CT_DIR_ORIGINAL].list, &unconfirmed);
+
+	atomic_inc(&ip_conntrack_count);
 	WRITE_UNLOCK(&ip_conntrack_lock);
 
 ret:	return &conntrack->tuplehash[IP_CT_DIR_ORIGINAL];
@@ -1067,6 +1086,7 @@ void ip_conntrack_helper_unregister(struct ip_conntrack_helper *me)
 	LIST_DELETE(&helpers, me);
 
 	/* Get rid of expecteds, set helpers to NULL. */
+	LIST_FIND_W(&unconfirmed, unhelp, struct ip_conntrack_tuple_hash*, me);
 	for (i = 0; i < ip_conntrack_htable_size; i++)
 		LIST_FIND_W(&ip_conntrack_hash[i], unhelp,
 			    struct ip_conntrack_tuple_hash *, me);
@@ -1179,40 +1199,44 @@ static void ip_conntrack_attach(struct sk_buff *nskb, struct sk_buff *skb)
 }
 
 static inline int
-do_kill(const struct ip_conntrack_tuple_hash *i,
-	int (*kill)(const struct ip_conntrack *i, void *data),
+do_iter(const struct ip_conntrack_tuple_hash *i,
+	int (*iter)(struct ip_conntrack *i, void *data),
 	void *data)
 {
-	return kill(i->ctrack, data);
+	return iter(i->ctrack, data);
 }
 
 /* Bring out ya dead! */
 static struct ip_conntrack_tuple_hash *
-get_next_corpse(int (*kill)(const struct ip_conntrack *i, void *data),
+get_next_corpse(int (*iter)(struct ip_conntrack *i, void *data),
 		void *data, unsigned int *bucket)
 {
 	struct ip_conntrack_tuple_hash *h = NULL;
 
-	READ_LOCK(&ip_conntrack_lock);
-	for (; !h && *bucket < ip_conntrack_htable_size; (*bucket)++) {
-		h = LIST_FIND(&ip_conntrack_hash[*bucket], do_kill,
-			      struct ip_conntrack_tuple_hash *, kill, data);
+	WRITE_LOCK(&ip_conntrack_lock);
+	for (; *bucket < ip_conntrack_htable_size; (*bucket)++) {
+		h = LIST_FIND_W(&ip_conntrack_hash[*bucket], do_iter,
+				struct ip_conntrack_tuple_hash *, iter, data);
+		if (h)
+			break;
 	}
+	if (!h)
+		h = LIST_FIND_W(&unconfirmed, do_iter,
+				struct ip_conntrack_tuple_hash *, iter, data);
 	if (h)
 		atomic_inc(&h->ctrack->ct_general.use);
-	READ_UNLOCK(&ip_conntrack_lock);
+	WRITE_UNLOCK(&ip_conntrack_lock);
 
 	return h;
 }
 
 void
-ip_ct_selective_cleanup(int (*kill)(const struct ip_conntrack *i, void *data),
-			void *data)
+ip_ct_iterate_cleanup(int (*iter)(struct ip_conntrack *i, void *), void *data)
 {
 	struct ip_conntrack_tuple_hash *h;
 	unsigned int bucket = 0;
 
-	while ((h = get_next_corpse(kill, data, &bucket)) != NULL) {
+	while ((h = get_next_corpse(iter, data, &bucket)) != NULL) {
 		/* Time to push up daises... */
 		if (del_timer(&h->ctrack->timeout))
 			death_by_timeout((unsigned long)h->ctrack);
@@ -1283,7 +1307,7 @@ static struct nf_sockopt_ops so_getorigdst = {
 	.get		= &getorigdst,
 };
 
-static int kill_all(const struct ip_conntrack *i, void *data)
+static int kill_all(struct ip_conntrack *i, void *data)
 {
 	return 1;
 }
@@ -1299,7 +1323,7 @@ void ip_conntrack_cleanup(void)
 	synchronize_net();
  
  i_see_dead_people:
-	ip_ct_selective_cleanup(kill_all, NULL);
+	ip_ct_iterate_cleanup(kill_all, NULL);
 	if (atomic_read(&ip_conntrack_count) != 0) {
 		schedule();
 		goto i_see_dead_people;

net/ipv4/netfilter/ip_conntrack_standalone.c

@@ -48,7 +48,7 @@ MODULE_LICENSE("GPL");
 extern atomic_t ip_conntrack_count;
 DECLARE_PER_CPU(struct ip_conntrack_stat, ip_conntrack_stat);
 
-static int kill_proto(const struct ip_conntrack *i, void *data)
+static int kill_proto(struct ip_conntrack *i, void *data)
 {
 	return (i->tuplehash[IP_CT_DIR_ORIGINAL].tuple.dst.protonum == 
 			*((u_int8_t *) data));
@@ -861,7 +861,7 @@ void ip_conntrack_protocol_unregister(struct ip_conntrack_protocol *proto)
 	synchronize_net();
 
 	/* Remove all contrack entries for this protocol */
-	ip_ct_selective_cleanup(kill_proto, &proto->proto);
+	ip_ct_iterate_cleanup(kill_proto, &proto->proto);
 }
 
 static int __init init(void)
@@ -892,7 +892,7 @@ EXPORT_SYMBOL(ip_conntrack_destroyed);
 EXPORT_SYMBOL(need_ip_conntrack);
 EXPORT_SYMBOL(ip_conntrack_helper_register);
 EXPORT_SYMBOL(ip_conntrack_helper_unregister);
-EXPORT_SYMBOL(ip_ct_selective_cleanup);
+EXPORT_SYMBOL(ip_ct_iterate_cleanup);
 EXPORT_SYMBOL(ip_ct_refresh_acct);
 EXPORT_SYMBOL(ip_ct_protos);
 EXPORT_SYMBOL(ip_ct_find_proto);

net/ipv4/netfilter/ip_nat_core.c

@@ -1006,16 +1006,16 @@ int __init ip_nat_init(void)
 }
 
 /* Clear NAT section of all conntracks, in case we're loaded again. */
-static int clean_nat(const struct ip_conntrack *i, void *data)
+static int clean_nat(struct ip_conntrack *i, void *data)
 {
-	memset((void *)&i->nat, 0, sizeof(i->nat));
+	memset(&i->nat, 0, sizeof(i->nat));
 	return 0;
 }
 
 /* Not __exit: called from ip_nat_standalone.c:init_or_cleanup() --RR */
 void ip_nat_cleanup(void)
 {
-	ip_ct_selective_cleanup(&clean_nat, NULL);
+	ip_ct_iterate_cleanup(&clean_nat, NULL);
 	ip_conntrack_destroyed = NULL;
 	vfree(bysource);
 }

net/ipv4/netfilter/ip_nat_helper.c

@@ -444,7 +444,7 @@ ip_nat_find_helper(const struct ip_conntrack_tuple *tuple)
 }
 
 static int
-kill_helper(const struct ip_conntrack *i, void *helper)
+kill_helper(struct ip_conntrack *i, void *helper)
 {
 	int ret;
 
@@ -474,5 +474,5 @@ void ip_nat_helper_unregister(struct ip_nat_helper *me)
 	   forces admins to gen fake RSTs or bounce box, either of
 	   which is just a long-winded way of making things
 	   worse. --RR */
-	ip_ct_selective_cleanup(kill_helper, me);
+	ip_ct_iterate_cleanup(kill_helper, me);
 }

net/ipv4/netfilter/ipt_MASQUERADE.c

@@ -118,7 +118,7 @@ masquerade_target(struct sk_buff **pskb,
 }
 
 static inline int
-device_cmp(const struct ip_conntrack *i, void *ifindex)
+device_cmp(struct ip_conntrack *i, void *ifindex)
 {
 	int ret;
 
@@ -141,7 +141,7 @@ static int masq_device_event(struct notifier_block *this,
 		   and forget them. */
 		IP_NF_ASSERT(dev->ifindex != 0);
 
-		ip_ct_selective_cleanup(device_cmp, (void *)(long)dev->ifindex);
+		ip_ct_iterate_cleanup(device_cmp, (void *)(long)dev->ifindex);
 	}
 
 	return NOTIFY_DONE;
@@ -159,7 +159,7 @@ static int masq_inet_event(struct notifier_block *this,
 		   and forget them. */
 		IP_NF_ASSERT(dev->ifindex != 0);
 
-		ip_ct_selective_cleanup(device_cmp, (void *)(long)dev->ifindex);
+		ip_ct_iterate_cleanup(device_cmp, (void *)(long)dev->ifindex);
 	}
 
 	return NOTIFY_DONE;