[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/slab.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>
5a0e3ad6	20	#include <linux/slab.h>
370786f9 JE	21	#include <linux/list.h>
	22	#include <linux/module.h>
	23	#include <linux/random.h>
	24	#include <linux/skbuff.h>
	25	#include <linux/spinlock.h>
	26	#include <linux/netfilter/nf_conntrack_tcp.h>
	27	#include <linux/netfilter/x_tables.h>
	28	#include <linux/netfilter/xt_connlimit.h>
	29	#include <net/netfilter/nf_conntrack.h>
	30	#include <net/netfilter/nf_conntrack_core.h>
	31	#include <net/netfilter/nf_conntrack_tuple.h>
5d0aa2cc	32	#include <net/netfilter/nf_conntrack_zones.h>
370786f9 JE	33
	34	/* we will save the tuples of all connections we care about */
	35	struct xt_connlimit_conn {
	36	struct list_head list;
	37	struct nf_conntrack_tuple tuple;
	38	};
	39
	40	struct xt_connlimit_data {
	41	struct list_head iphash[256];
	42	spinlock_t lock;
	43	};
	44
294188ae JE	45	static u_int32_t connlimit_rnd __read_mostly;
294188ae JE	46	static bool connlimit_rnd_inited __read_mostly;
370786f9	47
a34c4589	48	static inline unsigned int connlimit_iphash(__be32 addr)
370786f9	49	{
a34c4589	50	return jhash_1word((__force __u32)addr, connlimit_rnd) & 0xFF;
370786f9 JE	51	}
	52
	53	static inline unsigned int
643a2c15 JE	54	connlimit_iphash6(const union nf_inet_addr *addr,
643a2c15 JE	55	const union nf_inet_addr *mask)
370786f9	56	{
643a2c15	57	union nf_inet_addr res;
370786f9 JE	58	unsigned int i;
370786f9 JE	59
370786f9 JE	60	for (i = 0; i < ARRAY_SIZE(addr->ip6); ++i)
	61	res.ip6[i] = addr->ip6[i] & mask->ip6[i];
	62
a34c4589	63	return jhash2((u32 *)res.ip6, ARRAY_SIZE(res.ip6), connlimit_rnd) & 0xFF;
370786f9 JE	64	}
	65
	66	static inline bool already_closed(const struct nf_conn *conn)
	67	{
5e8fbe2a	68	if (nf_ct_protonum(conn) == IPPROTO_TCP)
d2ee3f2c DW	69	return conn->proto.tcp.state == TCP_CONNTRACK_TIME_WAIT \|\|
d2ee3f2c DW	70	conn->proto.tcp.state == TCP_CONNTRACK_CLOSE;
370786f9 JE	71	else
	72	return 0;
	73	}
	74
	75	static inline unsigned int
643a2c15 JE	76	same_source_net(const union nf_inet_addr *addr,
643a2c15 JE	77	const union nf_inet_addr *mask,
76108cea	78	const union nf_inet_addr *u3, u_int8_t family)
370786f9	79	{
ee999d8b	80	if (family == NFPROTO_IPV4) {
370786f9 JE	81	return (addr->ip & mask->ip) == (u3->ip & mask->ip);
370786f9 JE	82	} else {
643a2c15	83	union nf_inet_addr lh, rh;
370786f9 JE	84	unsigned int i;
	85
	86	for (i = 0; i < ARRAY_SIZE(addr->ip6); ++i) {
	87	lh.ip6[i] = addr->ip6[i] & mask->ip6[i];
	88	rh.ip6[i] = u3->ip6[i] & mask->ip6[i];
	89	}
	90
	91	return memcmp(&lh.ip6, &rh.ip6, sizeof(lh.ip6)) == 0;
	92	}
	93	}
	94
83fc8102 AD	95	static int count_them(struct net *net,
83fc8102 AD	96	struct xt_connlimit_data *data,
370786f9	97	const struct nf_conntrack_tuple *tuple,
643a2c15 JE	98	const union nf_inet_addr *addr,
643a2c15 JE	99	const union nf_inet_addr *mask,
539054a8	100	u_int8_t family)
370786f9	101	{
3cf93c96	102	const struct nf_conntrack_tuple_hash *found;
370786f9 JE	103	struct xt_connlimit_conn *conn;
370786f9 JE	104	struct xt_connlimit_conn *tmp;
ea781f19	105	struct nf_conn *found_ct;
370786f9 JE	106	struct list_head *hash;
	107	bool addit = true;
	108	int matches = 0;
	109
539054a8	110	if (family == NFPROTO_IPV6)
370786f9 JE	111	hash = &data->iphash[connlimit_iphash6(addr, mask)];
	112	else
	113	hash = &data->iphash[connlimit_iphash(addr->ip & mask->ip)];
	114
76507f69	115	rcu_read_lock();
370786f9 JE	116
	117	/* check the saved connections */
	118	list_for_each_entry_safe(conn, tmp, hash, list) {
5d0aa2cc PM	119	found = nf_conntrack_find_get(net, NF_CT_DEFAULT_ZONE,
5d0aa2cc PM	120	&conn->tuple);
370786f9 JE	121	found_ct = NULL;
	122
	123	if (found != NULL)
	124	found_ct = nf_ct_tuplehash_to_ctrack(found);
	125
	126	if (found_ct != NULL &&
	127	nf_ct_tuple_equal(&conn->tuple, tuple) &&
	128	!already_closed(found_ct))
	129	/*
	130	* Just to be sure we have it only once in the list.
	131	* We should not see tuples twice unless someone hooks
	132	* this into a table without "-p tcp --syn".
	133	*/
	134	addit = false;
	135
	136	if (found == NULL) {
	137	/* this one is gone */
	138	list_del(&conn->list);
	139	kfree(conn);
	140	continue;
	141	}
	142
	143	if (already_closed(found_ct)) {
	144	/*
	145	* we do not care about connections which are
	146	* closed already -> ditch it
	147	*/
ea781f19	148	nf_ct_put(found_ct);
370786f9 JE	149	list_del(&conn->list);
	150	kfree(conn);
	151	continue;
	152	}
	153
539054a8	154	if (same_source_net(addr, mask, &conn->tuple.src.u3, family))
370786f9 JE	155	/* same source network -> be counted! */
370786f9 JE	156	++matches;
ea781f19	157	nf_ct_put(found_ct);
370786f9 JE	158	}
370786f9 JE	159
76507f69	160	rcu_read_unlock();
370786f9 JE	161
	162	if (addit) {
	163	/* save the new connection in our list */
	164	conn = kzalloc(sizeof(*conn), GFP_ATOMIC);
	165	if (conn == NULL)
	166	return -ENOMEM;
	167	conn->tuple = *tuple;
	168	list_add(&conn->list, hash);
	169	++matches;
	170	}
	171
	172	return matches;
	173	}
	174
d3c5ee6d	175	static bool
f7108a20	176	connlimit_mt(const struct sk_buff skb, const struct xt_match_param par)
370786f9	177	{
83fc8102	178	struct net *net = dev_net(par->in ? par->in : par->out);
f7108a20	179	const struct xt_connlimit_info *info = par->matchinfo;
22c2d8bc	180	union nf_inet_addr addr;
370786f9 JE	181	struct nf_conntrack_tuple tuple;
	182	const struct nf_conntrack_tuple *tuple_ptr = &tuple;
	183	enum ip_conntrack_info ctinfo;
	184	const struct nf_conn *ct;
	185	int connections;
	186
	187	ct = nf_ct_get(skb, &ctinfo);
	188	if (ct != NULL)
	189	tuple_ptr = &ct->tuplehash[0].tuple;
	190	else if (!nf_ct_get_tuplepr(skb, skb_network_offset(skb),
92f3b2b1	191	par->family, &tuple))
370786f9 JE	192	goto hotdrop;
370786f9 JE	193
92f3b2b1	194	if (par->family == NFPROTO_IPV6) {
370786f9 JE	195	const struct ipv6hdr *iph = ipv6_hdr(skb);
370786f9 JE	196	memcpy(&addr.ip6, &iph->saddr, sizeof(iph->saddr));
370786f9 JE	197	} else {
	198	const struct iphdr *iph = ip_hdr(skb);
	199	addr.ip = iph->saddr;
370786f9 JE	200	}
	201
	202	spin_lock_bh(&info->data->lock);
83fc8102	203	connections = count_them(net, info->data, tuple_ptr, &addr,
539054a8	204	&info->mask, par->family);
370786f9 JE	205	spin_unlock_bh(&info->data->lock);
	206
	207	if (connections < 0) {
	208	/* kmalloc failed, drop it entirely */
f7108a20	209	*par->hotdrop = true;
370786f9 JE	210	return false;
	211	}
	212
	213	return (connections > info->limit) ^ info->inverse;
	214
	215	hotdrop:
f7108a20	216	*par->hotdrop = true;
370786f9 JE	217	return false;
	218	}
	219
9b4fce7a	220	static bool connlimit_mt_check(const struct xt_mtchk_param *par)
370786f9	221	{
9b4fce7a	222	struct xt_connlimit_info *info = par->matchinfo;
370786f9 JE	223	unsigned int i;
370786f9 JE	224
294188ae JE	225	if (unlikely(!connlimit_rnd_inited)) {
	226	get_random_bytes(&connlimit_rnd, sizeof(connlimit_rnd));
	227	connlimit_rnd_inited = true;
	228	}
92f3b2b1	229	if (nf_ct_l3proto_try_module_get(par->family) < 0) {
370786f9	230	printk(KERN_WARNING "cannot load conntrack support for "
92f3b2b1	231	"address family %u\n", par->family);
370786f9 JE	232	return false;
	233	}
	234
	235	/* init private data */
	236	info->data = kmalloc(sizeof(struct xt_connlimit_data), GFP_KERNEL);
	237	if (info->data == NULL) {
92f3b2b1	238	nf_ct_l3proto_module_put(par->family);
370786f9 JE	239	return false;
	240	}
	241
	242	spin_lock_init(&info->data->lock);
	243	for (i = 0; i < ARRAY_SIZE(info->data->iphash); ++i)
	244	INIT_LIST_HEAD(&info->data->iphash[i]);
	245
	246	return true;
	247	}
	248
6be3d859	249	static void connlimit_mt_destroy(const struct xt_mtdtor_param *par)
370786f9	250	{
6be3d859	251	const struct xt_connlimit_info *info = par->matchinfo;
370786f9 JE	252	struct xt_connlimit_conn *conn;
	253	struct xt_connlimit_conn *tmp;
	254	struct list_head *hash = info->data->iphash;
	255	unsigned int i;
	256
92f3b2b1	257	nf_ct_l3proto_module_put(par->family);
370786f9 JE	258
	259	for (i = 0; i < ARRAY_SIZE(info->data->iphash); ++i) {
	260	list_for_each_entry_safe(conn, tmp, &hash[i], list) {
	261	list_del(&conn->list);
	262	kfree(conn);
	263	}
	264	}
	265
	266	kfree(info->data);
	267	}
	268
92f3b2b1 JE	269	static struct xt_match connlimit_mt_reg __read_mostly = {
	270	.name = "connlimit",
	271	.revision = 0,
	272	.family = NFPROTO_UNSPEC,
	273	.checkentry = connlimit_mt_check,
	274	.match = connlimit_mt,
	275	.matchsize = sizeof(struct xt_connlimit_info),
	276	.destroy = connlimit_mt_destroy,
	277	.me = THIS_MODULE,
370786f9 JE	278	};
370786f9 JE	279
d3c5ee6d	280	static int __init connlimit_mt_init(void)
370786f9	281	{
92f3b2b1	282	return xt_register_match(&connlimit_mt_reg);
370786f9 JE	283	}
370786f9 JE	284
d3c5ee6d	285	static void __exit connlimit_mt_exit(void)
370786f9	286	{
92f3b2b1	287	xt_unregister_match(&connlimit_mt_reg);
370786f9 JE	288	}
370786f9 JE	289
d3c5ee6d JE	290	module_init(connlimit_mt_init);
d3c5ee6d JE	291	module_exit(connlimit_mt_exit);
92f3b2b1	292	MODULE_AUTHOR("Jan Engelhardt <jengelh@medozas.de>");
2ae15b64	293	MODULE_DESCRIPTION("Xtables: Number of connections matching");
370786f9 JE	294	MODULE_LICENSE("GPL");
	295	MODULE_ALIAS("ipt_connlimit");
	296	MODULE_ALIAS("ip6t_connlimit");