[net-next-2.6.git] / net / netfilter / xt_connlimit.c

/*
 * netfilter module to limit the number of parallel tcp
 * connections per IP address.
 *   (c) 2000 Gerd Knorr <kraxel@bytesex.org>
 *   Nov 2002: Martin Bene <martin.bene@icomedias.com>:
 *		only ignore TIME_WAIT or gone connections
 *   (C) CC Computer Consultants GmbH, 2007
 *   Contact: <jengelh@computergmbh.de>
 *
 * based on ...
 *
 * Kernel module to match connection tracking information.
 * GPL (C) 1999  Rusty Russell (rusty@rustcorp.com.au).
 */
#include <linux/in.h>
#include <linux/in6.h>
#include <linux/ip.h>
#include <linux/ipv6.h>
#include <linux/jhash.h>
#include <linux/list.h>
#include <linux/module.h>
#include <linux/random.h>
#include <linux/skbuff.h>
#include <linux/spinlock.h>
#include <linux/netfilter/nf_conntrack_tcp.h>
#include <linux/netfilter/x_tables.h>
#include <linux/netfilter/xt_connlimit.h>
#include <net/netfilter/nf_conntrack.h>
#include <net/netfilter/nf_conntrack_core.h>
#include <net/netfilter/nf_conntrack_tuple.h>
#include <net/netfilter/nf_conntrack_zones.h>

/* we will save the tuples of all connections we care about */
struct xt_connlimit_conn {
	struct list_head list;
	struct nf_conntrack_tuple tuple;
};

struct xt_connlimit_data {
	struct list_head iphash[256];
	spinlock_t lock;
};

static u_int32_t connlimit_rnd __read_mostly;
static bool connlimit_rnd_inited __read_mostly;

static inline unsigned int connlimit_iphash(__be32 addr)
{
	return jhash_1word((__force __u32)addr, connlimit_rnd) & 0xFF;
}

static inline unsigned int
connlimit_iphash6(const union nf_inet_addr *addr,
                  const union nf_inet_addr *mask)
{
	union nf_inet_addr res;
	unsigned int i;

	for (i = 0; i < ARRAY_SIZE(addr->ip6); ++i)
		res.ip6[i] = addr->ip6[i] & mask->ip6[i];

	return jhash2((u32 *)res.ip6, ARRAY_SIZE(res.ip6), connlimit_rnd) & 0xFF;
}

static inline bool already_closed(const struct nf_conn *conn)
{
	if (nf_ct_protonum(conn) == IPPROTO_TCP)
		return conn->proto.tcp.state == TCP_CONNTRACK_TIME_WAIT ||
		       conn->proto.tcp.state == TCP_CONNTRACK_CLOSE;
	else
		return 0;
}

static inline unsigned int
same_source_net(const union nf_inet_addr *addr,
		const union nf_inet_addr *mask,
		const union nf_inet_addr *u3, u_int8_t family)
{
	if (family == NFPROTO_IPV4) {
		return (addr->ip & mask->ip) == (u3->ip & mask->ip);
	} else {
		union nf_inet_addr lh, rh;
		unsigned int i;

		for (i = 0; i < ARRAY_SIZE(addr->ip6); ++i) {
			lh.ip6[i] = addr->ip6[i] & mask->ip6[i];
			rh.ip6[i] = u3->ip6[i] & mask->ip6[i];
		}

		return memcmp(&lh.ip6, &rh.ip6, sizeof(lh.ip6)) == 0;
	}
}

static int count_them(struct net *net,
		      struct xt_connlimit_data *data,
		      const struct nf_conntrack_tuple *tuple,
		      const union nf_inet_addr *addr,
		      const union nf_inet_addr *mask,
		      u_int8_t family)
{
	const struct nf_conntrack_tuple_hash *found;
	struct xt_connlimit_conn *conn;
	struct xt_connlimit_conn *tmp;
	struct nf_conn *found_ct;
	struct list_head *hash;
	bool addit = true;
	int matches = 0;

	if (family == NFPROTO_IPV6)
		hash = &data->iphash[connlimit_iphash6(addr, mask)];
	else
		hash = &data->iphash[connlimit_iphash(addr->ip & mask->ip)];

	rcu_read_lock();

	/* check the saved connections */
	list_for_each_entry_safe(conn, tmp, hash, list) {
		found    = nf_conntrack_find_get(net, NF_CT_DEFAULT_ZONE,
						 &conn->tuple);
		found_ct = NULL;

		if (found != NULL)
			found_ct = nf_ct_tuplehash_to_ctrack(found);

		if (found_ct != NULL &&
		    nf_ct_tuple_equal(&conn->tuple, tuple) &&
		    !already_closed(found_ct))
			/*
			 * Just to be sure we have it only once in the list.
			 * We should not see tuples twice unless someone hooks
			 * this into a table without "-p tcp --syn".
			 */
			addit = false;

		if (found == NULL) {
			/* this one is gone */
			list_del(&conn->list);
			kfree(conn);
			continue;
		}

		if (already_closed(found_ct)) {
			/*
			 * we do not care about connections which are
			 * closed already -> ditch it
			 */
			nf_ct_put(found_ct);
			list_del(&conn->list);
			kfree(conn);
			continue;
		}

		if (same_source_net(addr, mask, &conn->tuple.src.u3, family))
			/* same source network -> be counted! */
			++matches;
		nf_ct_put(found_ct);
	}

	rcu_read_unlock();

	if (addit) {
		/* save the new connection in our list */
		conn = kzalloc(sizeof(*conn), GFP_ATOMIC);
		if (conn == NULL)
			return -ENOMEM;
		conn->tuple = *tuple;
		list_add(&conn->list, hash);
		++matches;
	}

	return matches;
}

static bool
connlimit_mt(const struct sk_buff *skb, const struct xt_match_param *par)
{
	struct net *net = dev_net(par->in ? par->in : par->out);
	const struct xt_connlimit_info *info = par->matchinfo;
	union nf_inet_addr addr;
	struct nf_conntrack_tuple tuple;
	const struct nf_conntrack_tuple *tuple_ptr = &tuple;
	enum ip_conntrack_info ctinfo;
	const struct nf_conn *ct;
	int connections;

	ct = nf_ct_get(skb, &ctinfo);
	if (ct != NULL)
		tuple_ptr = &ct->tuplehash[0].tuple;
	else if (!nf_ct_get_tuplepr(skb, skb_network_offset(skb),
				    par->family, &tuple))
		goto hotdrop;

	if (par->family == NFPROTO_IPV6) {
		const struct ipv6hdr *iph = ipv6_hdr(skb);
		memcpy(&addr.ip6, &iph->saddr, sizeof(iph->saddr));
	} else {
		const struct iphdr *iph = ip_hdr(skb);
		addr.ip = iph->saddr;
	}

	spin_lock_bh(&info->data->lock);
	connections = count_them(net, info->data, tuple_ptr, &addr,
	                         &info->mask, par->family);
	spin_unlock_bh(&info->data->lock);

	if (connections < 0) {
		/* kmalloc failed, drop it entirely */
		*par->hotdrop = true;
		return false;
	}

	return (connections > info->limit) ^ info->inverse;

 hotdrop:
	*par->hotdrop = true;
	return false;
}

static bool connlimit_mt_check(const struct xt_mtchk_param *par)
{
	struct xt_connlimit_info *info = par->matchinfo;
	unsigned int i;

	if (unlikely(!connlimit_rnd_inited)) {
		get_random_bytes(&connlimit_rnd, sizeof(connlimit_rnd));
		connlimit_rnd_inited = true;
	}
	if (nf_ct_l3proto_try_module_get(par->family) < 0) {
		printk(KERN_WARNING "cannot load conntrack support for "
		       "address family %u\n", par->family);
		return false;
	}

	/* init private data */
	info->data = kmalloc(sizeof(struct xt_connlimit_data), GFP_KERNEL);
	if (info->data == NULL) {
		nf_ct_l3proto_module_put(par->family);
		return false;
	}

	spin_lock_init(&info->data->lock);
	for (i = 0; i < ARRAY_SIZE(info->data->iphash); ++i)
		INIT_LIST_HEAD(&info->data->iphash[i]);

	return true;
}

static void connlimit_mt_destroy(const struct xt_mtdtor_param *par)
{
	const struct xt_connlimit_info *info = par->matchinfo;
	struct xt_connlimit_conn *conn;
	struct xt_connlimit_conn *tmp;
	struct list_head *hash = info->data->iphash;
	unsigned int i;

	nf_ct_l3proto_module_put(par->family);

	for (i = 0; i < ARRAY_SIZE(info->data->iphash); ++i) {
		list_for_each_entry_safe(conn, tmp, &hash[i], list) {
			list_del(&conn->list);
			kfree(conn);
		}
	}

	kfree(info->data);
}

static struct xt_match connlimit_mt_reg __read_mostly = {
	.name       = "connlimit",
	.revision   = 0,
	.family     = NFPROTO_UNSPEC,
	.checkentry = connlimit_mt_check,
	.match      = connlimit_mt,
	.matchsize  = sizeof(struct xt_connlimit_info),
	.destroy    = connlimit_mt_destroy,
	.me         = THIS_MODULE,
};

static int __init connlimit_mt_init(void)
{
	return xt_register_match(&connlimit_mt_reg);
}

static void __exit connlimit_mt_exit(void)
{
	xt_unregister_match(&connlimit_mt_reg);
}

module_init(connlimit_mt_init);
module_exit(connlimit_mt_exit);
MODULE_AUTHOR("Jan Engelhardt <jengelh@medozas.de>");
MODULE_DESCRIPTION("Xtables: Number of connections matching");
MODULE_LICENSE("GPL");
MODULE_ALIAS("ipt_connlimit");
MODULE_ALIAS("ip6t_connlimit");
Commit	Line	Data
370786f9 JE	1	/*
	2	* netfilter module to limit the number of parallel tcp
	3	* connections per IP address.
	4	* (c) 2000 Gerd Knorr <kraxel@bytesex.org>
	5	* Nov 2002: Martin Bene <martin.bene@icomedias.com>:
	6	* only ignore TIME_WAIT or gone connections
ba5dc275 JE	7	* (C) CC Computer Consultants GmbH, 2007
ba5dc275 JE	8	* Contact: <jengelh@computergmbh.de>
370786f9 JE	9	*
	10	* based on ...
	11	*
	12	* Kernel module to match connection tracking information.
	13	* GPL (C) 1999 Rusty Russell (rusty@rustcorp.com.au).
	14	*/
	15	#include <linux/in.h>
	16	#include <linux/in6.h>
	17	#include <linux/ip.h>
	18	#include <linux/ipv6.h>
	19	#include <linux/jhash.h>
	20	#include <linux/list.h>
	21	#include <linux/module.h>
	22	#include <linux/random.h>
	23	#include <linux/skbuff.h>
	24	#include <linux/spinlock.h>
	25	#include <linux/netfilter/nf_conntrack_tcp.h>
	26	#include <linux/netfilter/x_tables.h>
	27	#include <linux/netfilter/xt_connlimit.h>
	28	#include <net/netfilter/nf_conntrack.h>
	29	#include <net/netfilter/nf_conntrack_core.h>
	30	#include <net/netfilter/nf_conntrack_tuple.h>
5d0aa2cc	31	#include <net/netfilter/nf_conntrack_zones.h>
370786f9 JE	32
	33	/* we will save the tuples of all connections we care about */
	34	struct xt_connlimit_conn {
	35	struct list_head list;
	36	struct nf_conntrack_tuple tuple;
	37	};
	38
	39	struct xt_connlimit_data {
	40	struct list_head iphash[256];
	41	spinlock_t lock;
	42	};
	43
294188ae JE	44	static u_int32_t connlimit_rnd __read_mostly;
294188ae JE	45	static bool connlimit_rnd_inited __read_mostly;
370786f9	46
a34c4589	47	static inline unsigned int connlimit_iphash(__be32 addr)
370786f9	48	{
a34c4589	49	return jhash_1word((__force __u32)addr, connlimit_rnd) & 0xFF;
370786f9 JE	50	}
	51
	52	static inline unsigned int
643a2c15 JE	53	connlimit_iphash6(const union nf_inet_addr *addr,
643a2c15 JE	54	const union nf_inet_addr *mask)
370786f9	55	{
643a2c15	56	union nf_inet_addr res;
370786f9 JE	57	unsigned int i;
370786f9 JE	58
370786f9 JE	59	for (i = 0; i < ARRAY_SIZE(addr->ip6); ++i)
	60	res.ip6[i] = addr->ip6[i] & mask->ip6[i];
	61
a34c4589	62	return jhash2((u32 *)res.ip6, ARRAY_SIZE(res.ip6), connlimit_rnd) & 0xFF;
370786f9 JE	63	}
	64
	65	static inline bool already_closed(const struct nf_conn *conn)
	66	{
5e8fbe2a	67	if (nf_ct_protonum(conn) == IPPROTO_TCP)
d2ee3f2c DW	68	return conn->proto.tcp.state == TCP_CONNTRACK_TIME_WAIT \|\|
d2ee3f2c DW	69	conn->proto.tcp.state == TCP_CONNTRACK_CLOSE;
370786f9 JE	70	else
	71	return 0;
	72	}
	73
	74	static inline unsigned int
643a2c15 JE	75	same_source_net(const union nf_inet_addr *addr,
643a2c15 JE	76	const union nf_inet_addr *mask,
76108cea	77	const union nf_inet_addr *u3, u_int8_t family)
370786f9	78	{
ee999d8b	79	if (family == NFPROTO_IPV4) {
370786f9 JE	80	return (addr->ip & mask->ip) == (u3->ip & mask->ip);
370786f9 JE	81	} else {
643a2c15	82	union nf_inet_addr lh, rh;
370786f9 JE	83	unsigned int i;
	84
	85	for (i = 0; i < ARRAY_SIZE(addr->ip6); ++i) {
	86	lh.ip6[i] = addr->ip6[i] & mask->ip6[i];
	87	rh.ip6[i] = u3->ip6[i] & mask->ip6[i];
	88	}
	89
	90	return memcmp(&lh.ip6, &rh.ip6, sizeof(lh.ip6)) == 0;
	91	}
	92	}
	93
83fc8102 AD	94	static int count_them(struct net *net,
83fc8102 AD	95	struct xt_connlimit_data *data,
370786f9	96	const struct nf_conntrack_tuple *tuple,
643a2c15 JE	97	const union nf_inet_addr *addr,
643a2c15 JE	98	const union nf_inet_addr *mask,
539054a8	99	u_int8_t family)
370786f9	100	{
3cf93c96	101	const struct nf_conntrack_tuple_hash *found;
370786f9 JE	102	struct xt_connlimit_conn *conn;
370786f9 JE	103	struct xt_connlimit_conn *tmp;
ea781f19	104	struct nf_conn *found_ct;
370786f9 JE	105	struct list_head *hash;
	106	bool addit = true;
	107	int matches = 0;
	108
539054a8	109	if (family == NFPROTO_IPV6)
370786f9 JE	110	hash = &data->iphash[connlimit_iphash6(addr, mask)];
	111	else
	112	hash = &data->iphash[connlimit_iphash(addr->ip & mask->ip)];
	113
76507f69	114	rcu_read_lock();
370786f9 JE	115
	116	/* check the saved connections */
	117	list_for_each_entry_safe(conn, tmp, hash, list) {
5d0aa2cc PM	118	found = nf_conntrack_find_get(net, NF_CT_DEFAULT_ZONE,
5d0aa2cc PM	119	&conn->tuple);
370786f9 JE	120	found_ct = NULL;
	121
	122	if (found != NULL)
	123	found_ct = nf_ct_tuplehash_to_ctrack(found);
	124
	125	if (found_ct != NULL &&
	126	nf_ct_tuple_equal(&conn->tuple, tuple) &&
	127	!already_closed(found_ct))
	128	/*
	129	* Just to be sure we have it only once in the list.
	130	* We should not see tuples twice unless someone hooks
	131	* this into a table without "-p tcp --syn".
	132	*/
	133	addit = false;
	134
	135	if (found == NULL) {
	136	/* this one is gone */
	137	list_del(&conn->list);
	138	kfree(conn);
	139	continue;
	140	}
	141
	142	if (already_closed(found_ct)) {
	143	/*
	144	* we do not care about connections which are
	145	* closed already -> ditch it
	146	*/
ea781f19	147	nf_ct_put(found_ct);
370786f9 JE	148	list_del(&conn->list);
	149	kfree(conn);
	150	continue;
	151	}
	152
539054a8	153	if (same_source_net(addr, mask, &conn->tuple.src.u3, family))
370786f9 JE	154	/* same source network -> be counted! */
370786f9 JE	155	++matches;
ea781f19	156	nf_ct_put(found_ct);
370786f9 JE	157	}
370786f9 JE	158
76507f69	159	rcu_read_unlock();
370786f9 JE	160
	161	if (addit) {
	162	/* save the new connection in our list */
	163	conn = kzalloc(sizeof(*conn), GFP_ATOMIC);
	164	if (conn == NULL)
	165	return -ENOMEM;
	166	conn->tuple = *tuple;
	167	list_add(&conn->list, hash);
	168	++matches;
	169	}
	170
	171	return matches;
	172	}
	173
d3c5ee6d	174	static bool
f7108a20	175	connlimit_mt(const struct sk_buff skb, const struct xt_match_param par)
370786f9	176	{
83fc8102	177	struct net *net = dev_net(par->in ? par->in : par->out);
f7108a20	178	const struct xt_connlimit_info *info = par->matchinfo;
22c2d8bc	179	union nf_inet_addr addr;
370786f9 JE	180	struct nf_conntrack_tuple tuple;
	181	const struct nf_conntrack_tuple *tuple_ptr = &tuple;
	182	enum ip_conntrack_info ctinfo;
	183	const struct nf_conn *ct;
	184	int connections;
	185
	186	ct = nf_ct_get(skb, &ctinfo);
	187	if (ct != NULL)
	188	tuple_ptr = &ct->tuplehash[0].tuple;
	189	else if (!nf_ct_get_tuplepr(skb, skb_network_offset(skb),
92f3b2b1	190	par->family, &tuple))
370786f9 JE	191	goto hotdrop;
370786f9 JE	192
92f3b2b1	193	if (par->family == NFPROTO_IPV6) {
370786f9 JE	194	const struct ipv6hdr *iph = ipv6_hdr(skb);
370786f9 JE	195	memcpy(&addr.ip6, &iph->saddr, sizeof(iph->saddr));
370786f9 JE	196	} else {
	197	const struct iphdr *iph = ip_hdr(skb);
	198	addr.ip = iph->saddr;
370786f9 JE	199	}
	200
	201	spin_lock_bh(&info->data->lock);
83fc8102	202	connections = count_them(net, info->data, tuple_ptr, &addr,
539054a8	203	&info->mask, par->family);
370786f9 JE	204	spin_unlock_bh(&info->data->lock);
	205
	206	if (connections < 0) {
	207	/* kmalloc failed, drop it entirely */
f7108a20	208	*par->hotdrop = true;
370786f9 JE	209	return false;
	210	}
	211
	212	return (connections > info->limit) ^ info->inverse;
	213
	214	hotdrop:
f7108a20	215	*par->hotdrop = true;
370786f9 JE	216	return false;
	217	}
	218
9b4fce7a	219	static bool connlimit_mt_check(const struct xt_mtchk_param *par)
370786f9	220	{
9b4fce7a	221	struct xt_connlimit_info *info = par->matchinfo;
370786f9 JE	222	unsigned int i;
370786f9 JE	223
294188ae JE	224	if (unlikely(!connlimit_rnd_inited)) {
	225	get_random_bytes(&connlimit_rnd, sizeof(connlimit_rnd));
	226	connlimit_rnd_inited = true;
	227	}
92f3b2b1	228	if (nf_ct_l3proto_try_module_get(par->family) < 0) {
370786f9	229	printk(KERN_WARNING "cannot load conntrack support for "
92f3b2b1	230	"address family %u\n", par->family);
370786f9 JE	231	return false;
	232	}
	233
	234	/* init private data */
	235	info->data = kmalloc(sizeof(struct xt_connlimit_data), GFP_KERNEL);
	236	if (info->data == NULL) {
92f3b2b1	237	nf_ct_l3proto_module_put(par->family);
370786f9 JE	238	return false;
	239	}
	240
	241	spin_lock_init(&info->data->lock);
	242	for (i = 0; i < ARRAY_SIZE(info->data->iphash); ++i)
	243	INIT_LIST_HEAD(&info->data->iphash[i]);
	244
	245	return true;
	246	}
	247
6be3d859	248	static void connlimit_mt_destroy(const struct xt_mtdtor_param *par)
370786f9	249	{
6be3d859	250	const struct xt_connlimit_info *info = par->matchinfo;
370786f9 JE	251	struct xt_connlimit_conn *conn;
	252	struct xt_connlimit_conn *tmp;
	253	struct list_head *hash = info->data->iphash;
	254	unsigned int i;
	255
92f3b2b1	256	nf_ct_l3proto_module_put(par->family);
370786f9 JE	257
	258	for (i = 0; i < ARRAY_SIZE(info->data->iphash); ++i) {
	259	list_for_each_entry_safe(conn, tmp, &hash[i], list) {
	260	list_del(&conn->list);
	261	kfree(conn);
	262	}
	263	}
	264
	265	kfree(info->data);
	266	}
	267
92f3b2b1 JE	268	static struct xt_match connlimit_mt_reg __read_mostly = {
	269	.name = "connlimit",
	270	.revision = 0,
	271	.family = NFPROTO_UNSPEC,
	272	.checkentry = connlimit_mt_check,
	273	.match = connlimit_mt,
	274	.matchsize = sizeof(struct xt_connlimit_info),
	275	.destroy = connlimit_mt_destroy,
	276	.me = THIS_MODULE,
370786f9 JE	277	};
370786f9 JE	278
d3c5ee6d	279	static int __init connlimit_mt_init(void)
370786f9	280	{
92f3b2b1	281	return xt_register_match(&connlimit_mt_reg);
370786f9 JE	282	}
370786f9 JE	283
d3c5ee6d	284	static void __exit connlimit_mt_exit(void)
370786f9	285	{
92f3b2b1	286	xt_unregister_match(&connlimit_mt_reg);
370786f9 JE	287	}
370786f9 JE	288
d3c5ee6d JE	289	module_init(connlimit_mt_init);
d3c5ee6d JE	290	module_exit(connlimit_mt_exit);
92f3b2b1	291	MODULE_AUTHOR("Jan Engelhardt <jengelh@medozas.de>");
2ae15b64	292	MODULE_DESCRIPTION("Xtables: Number of connections matching");
370786f9 JE	293	MODULE_LICENSE("GPL");
	294	MODULE_ALIAS("ipt_connlimit");
	295	MODULE_ALIAS("ip6t_connlimit");