[mirror_ubuntu-artful-kernel.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
 *
 * based on ...
 *
 * Kernel module to match connection tracking information.
 * GPL (C) 1999  Rusty Russell (rusty@rustcorp.com.au).
 */
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
#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/rbtree.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>

#define CONNLIMIT_SLOTS		256U

#ifdef CONFIG_LOCKDEP
#define CONNLIMIT_LOCK_SLOTS	8U
#else
#define CONNLIMIT_LOCK_SLOTS	256U
#endif

#define CONNLIMIT_GC_MAX_NODES	8

/* we will save the tuples of all connections we care about */
struct xt_connlimit_conn {
	struct hlist_node		node;
	struct nf_conntrack_tuple	tuple;
	union nf_inet_addr		addr;
};

struct xt_connlimit_rb {
	struct rb_node node;
	struct hlist_head hhead; /* connections/hosts in same subnet */
	union nf_inet_addr addr; /* search key */
};

static spinlock_t xt_connlimit_locks[CONNLIMIT_LOCK_SLOTS] __cacheline_aligned_in_smp;

struct xt_connlimit_data {
	struct rb_root climit_root4[CONNLIMIT_SLOTS];
	struct rb_root climit_root6[CONNLIMIT_SLOTS];
};

static u_int32_t connlimit_rnd __read_mostly;
static struct kmem_cache *connlimit_rb_cachep __read_mostly;
static struct kmem_cache *connlimit_conn_cachep __read_mostly;

static inline unsigned int connlimit_iphash(__be32 addr)
{
	return jhash_1word((__force __u32)addr,
			    connlimit_rnd) % CONNLIMIT_SLOTS;
}

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) % CONNLIMIT_SLOTS;
}

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 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 ntohl(addr->ip & mask->ip) -
		       ntohl(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));
	}
}

static bool add_hlist(struct hlist_head *head,
		      const struct nf_conntrack_tuple *tuple,
		      const union nf_inet_addr *addr)
{
	struct xt_connlimit_conn *conn;

	conn = kmem_cache_alloc(connlimit_conn_cachep, GFP_ATOMIC);
	if (conn == NULL)
		return false;
	conn->tuple = *tuple;
	conn->addr = *addr;
	hlist_add_head(&conn->node, head);
	return true;
}

static unsigned int check_hlist(struct net *net,
				struct hlist_head *head,
				const struct nf_conntrack_tuple *tuple,
				const struct nf_conntrack_zone *zone,
				bool *addit)
{
	const struct nf_conntrack_tuple_hash *found;
	struct xt_connlimit_conn *conn;
	struct hlist_node *n;
	struct nf_conn *found_ct;
	unsigned int length = 0;

	*addit = true;
	rcu_read_lock();

	/* check the saved connections */
	hlist_for_each_entry_safe(conn, n, head, node) {
		found = nf_conntrack_find_get(net, zone, &conn->tuple);
		if (found == NULL) {
			hlist_del(&conn->node);
			kmem_cache_free(connlimit_conn_cachep, conn);
			continue;
		}

		found_ct = nf_ct_tuplehash_to_ctrack(found);

		if (nf_ct_tuple_equal(&conn->tuple, tuple)) {
			/*
			 * 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;
		} else if (already_closed(found_ct)) {
			/*
			 * we do not care about connections which are
			 * closed already -> ditch it
			 */
			nf_ct_put(found_ct);
			hlist_del(&conn->node);
			kmem_cache_free(connlimit_conn_cachep, conn);
			continue;
		}

		nf_ct_put(found_ct);
		length++;
	}

	rcu_read_unlock();

	return length;
}

static void tree_nodes_free(struct rb_root *root,
			    struct xt_connlimit_rb *gc_nodes[],
			    unsigned int gc_count)
{
	struct xt_connlimit_rb *rbconn;

	while (gc_count) {
		rbconn = gc_nodes[--gc_count];
		rb_erase(&rbconn->node, root);
		kmem_cache_free(connlimit_rb_cachep, rbconn);
	}
}

static unsigned int
count_tree(struct net *net, struct rb_root *root,
	   const struct nf_conntrack_tuple *tuple,
	   const union nf_inet_addr *addr, const union nf_inet_addr *mask,
	   u8 family, const struct nf_conntrack_zone *zone)
{
	struct xt_connlimit_rb *gc_nodes[CONNLIMIT_GC_MAX_NODES];
	struct rb_node **rbnode, *parent;
	struct xt_connlimit_rb *rbconn;
	struct xt_connlimit_conn *conn;
	unsigned int gc_count;
	bool no_gc = false;

 restart:
	gc_count = 0;
	parent = NULL;
	rbnode = &(root->rb_node);
	while (*rbnode) {
		int diff;
		bool addit;

		rbconn = rb_entry(*rbnode, struct xt_connlimit_rb, node);

		parent = *rbnode;
		diff = same_source_net(addr, mask, &rbconn->addr, family);
		if (diff < 0) {
			rbnode = &((*rbnode)->rb_left);
		} else if (diff > 0) {
			rbnode = &((*rbnode)->rb_right);
		} else {
			/* same source network -> be counted! */
			unsigned int count;
			count = check_hlist(net, &rbconn->hhead, tuple, zone, &addit);

			tree_nodes_free(root, gc_nodes, gc_count);
			if (!addit)
				return count;

			if (!add_hlist(&rbconn->hhead, tuple, addr))
				return 0; /* hotdrop */

			return count + 1;
		}

		if (no_gc || gc_count >= ARRAY_SIZE(gc_nodes))
			continue;

		/* only used for GC on hhead, retval and 'addit' ignored */
		check_hlist(net, &rbconn->hhead, tuple, zone, &addit);
		if (hlist_empty(&rbconn->hhead))
			gc_nodes[gc_count++] = rbconn;
	}

	if (gc_count) {
		no_gc = true;
		tree_nodes_free(root, gc_nodes, gc_count);
		/* tree_node_free before new allocation permits
		 * allocator to re-use newly free'd object.
		 *
		 * This is a rare event; in most cases we will find
		 * existing node to re-use. (or gc_count is 0).
		 */
		goto restart;
	}

	/* no match, need to insert new node */
	rbconn = kmem_cache_alloc(connlimit_rb_cachep, GFP_ATOMIC);
	if (rbconn == NULL)
		return 0;

	conn = kmem_cache_alloc(connlimit_conn_cachep, GFP_ATOMIC);
	if (conn == NULL) {
		kmem_cache_free(connlimit_rb_cachep, rbconn);
		return 0;
	}

	conn->tuple = *tuple;
	conn->addr = *addr;
	rbconn->addr = *addr;

	INIT_HLIST_HEAD(&rbconn->hhead);
	hlist_add_head(&conn->node, &rbconn->hhead);

	rb_link_node(&rbconn->node, parent, rbnode);
	rb_insert_color(&rbconn->node, root);
	return 1;
}

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_zone *zone)
{
	struct rb_root *root;
	int count;
	u32 hash;

	if (family == NFPROTO_IPV6) {
		hash = connlimit_iphash6(addr, mask);
		root = &data->climit_root6[hash];
	} else {
		hash = connlimit_iphash(addr->ip & mask->ip);
		root = &data->climit_root4[hash];
	}

	spin_lock_bh(&xt_connlimit_locks[hash % CONNLIMIT_LOCK_SLOTS]);

	count = count_tree(net, root, tuple, addr, mask, family, zone);

	spin_unlock_bh(&xt_connlimit_locks[hash % CONNLIMIT_LOCK_SLOTS]);

	return count;
}

static bool
connlimit_mt(const struct sk_buff *skb, struct xt_action_param *par)
{
	struct net *net = xt_net(par);
	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;
	const struct nf_conntrack_zone *zone = &nf_ct_zone_dflt;
	enum ip_conntrack_info ctinfo;
	const struct nf_conn *ct;
	unsigned int connections;

	ct = nf_ct_get(skb, &ctinfo);
	if (ct != NULL) {
		tuple_ptr = &ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple;
		zone = nf_ct_zone(ct);
	} else if (!nf_ct_get_tuplepr(skb, skb_network_offset(skb),
				      xt_family(par), net, &tuple)) {
		goto hotdrop;
	}

	if (xt_family(par) == NFPROTO_IPV6) {
		const struct ipv6hdr *iph = ipv6_hdr(skb);
		memcpy(&addr.ip6, (info->flags & XT_CONNLIMIT_DADDR) ?
		       &iph->daddr : &iph->saddr, sizeof(addr.ip6));
	} else {
		const struct iphdr *iph = ip_hdr(skb);
		addr.ip = (info->flags & XT_CONNLIMIT_DADDR) ?
			  iph->daddr : iph->saddr;
	}

	connections = count_them(net, info->data, tuple_ptr, &addr,
	                         &info->mask, xt_family(par), zone);
	if (connections == 0)
		/* kmalloc failed, drop it entirely */
		goto hotdrop;

	return (connections > info->limit) ^
	       !!(info->flags & XT_CONNLIMIT_INVERT);

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

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

	net_get_random_once(&connlimit_rnd, sizeof(connlimit_rnd));

	ret = nf_ct_netns_get(par->net, par->family);
	if (ret < 0) {
		pr_info("cannot load conntrack support for "
			"address family %u\n", par->family);
		return ret;
	}

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

	for (i = 0; i < ARRAY_SIZE(info->data->climit_root4); ++i)
		info->data->climit_root4[i] = RB_ROOT;
	for (i = 0; i < ARRAY_SIZE(info->data->climit_root6); ++i)
		info->data->climit_root6[i] = RB_ROOT;

	return 0;
}

static void destroy_tree(struct rb_root *r)
{
	struct xt_connlimit_conn *conn;
	struct xt_connlimit_rb *rbconn;
	struct hlist_node *n;
	struct rb_node *node;

	while ((node = rb_first(r)) != NULL) {
		rbconn = rb_entry(node, struct xt_connlimit_rb, node);

		rb_erase(node, r);

		hlist_for_each_entry_safe(conn, n, &rbconn->hhead, node)
			kmem_cache_free(connlimit_conn_cachep, conn);

		kmem_cache_free(connlimit_rb_cachep, rbconn);
	}
}

static void connlimit_mt_destroy(const struct xt_mtdtor_param *par)
{
	const struct xt_connlimit_info *info = par->matchinfo;
	unsigned int i;

	nf_ct_netns_put(par->net, par->family);

	for (i = 0; i < ARRAY_SIZE(info->data->climit_root4); ++i)
		destroy_tree(&info->data->climit_root4[i]);
	for (i = 0; i < ARRAY_SIZE(info->data->climit_root6); ++i)
		destroy_tree(&info->data->climit_root6[i]);

	kfree(info->data);
}

static struct xt_match connlimit_mt_reg __read_mostly = {
	.name       = "connlimit",
	.revision   = 1,
	.family     = NFPROTO_UNSPEC,
	.checkentry = connlimit_mt_check,
	.match      = connlimit_mt,
	.matchsize  = sizeof(struct xt_connlimit_info),
	.usersize   = offsetof(struct xt_connlimit_info, data),
	.destroy    = connlimit_mt_destroy,
	.me         = THIS_MODULE,
};

static int __init connlimit_mt_init(void)
{
	int ret, i;

	BUILD_BUG_ON(CONNLIMIT_LOCK_SLOTS > CONNLIMIT_SLOTS);
	BUILD_BUG_ON((CONNLIMIT_SLOTS % CONNLIMIT_LOCK_SLOTS) != 0);

	for (i = 0; i < CONNLIMIT_LOCK_SLOTS; ++i)
		spin_lock_init(&xt_connlimit_locks[i]);

	connlimit_conn_cachep = kmem_cache_create("xt_connlimit_conn",
					   sizeof(struct xt_connlimit_conn),
					   0, 0, NULL);
	if (!connlimit_conn_cachep)
		return -ENOMEM;

	connlimit_rb_cachep = kmem_cache_create("xt_connlimit_rb",
					   sizeof(struct xt_connlimit_rb),
					   0, 0, NULL);
	if (!connlimit_rb_cachep) {
		kmem_cache_destroy(connlimit_conn_cachep);
		return -ENOMEM;
	}
	ret = xt_register_match(&connlimit_mt_reg);
	if (ret != 0) {
		kmem_cache_destroy(connlimit_conn_cachep);
		kmem_cache_destroy(connlimit_rb_cachep);
	}
	return ret;
}

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

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	7	* (C) CC Computer Consultants GmbH, 2007
370786f9 JE	8	*
	9	* based on ...
	10	*
	11	* Kernel module to match connection tracking information.
	12	* GPL (C) 1999 Rusty Russell (rusty@rustcorp.com.au).
	13	*/
8bee4bad	14	#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
370786f9 JE	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	21	#include <linux/list.h>
7d084877	22	#include <linux/rbtree.h>
370786f9 JE	23	#include <linux/module.h>
	24	#include <linux/random.h>
	25	#include <linux/skbuff.h>
	26	#include <linux/spinlock.h>
	27	#include <linux/netfilter/nf_conntrack_tcp.h>
	28	#include <linux/netfilter/x_tables.h>
	29	#include <linux/netfilter/xt_connlimit.h>
	30	#include <net/netfilter/nf_conntrack.h>
	31	#include <net/netfilter/nf_conntrack_core.h>
	32	#include <net/netfilter/nf_conntrack_tuple.h>
5d0aa2cc	33	#include <net/netfilter/nf_conntrack_zones.h>
370786f9	34
e00b437b FW	35	#define CONNLIMIT_SLOTS 256U
	36
	37	#ifdef CONFIG_LOCKDEP
	38	#define CONNLIMIT_LOCK_SLOTS 8U
	39	#else
	40	#define CONNLIMIT_LOCK_SLOTS 256U
	41	#endif
	42
7d084877	43	#define CONNLIMIT_GC_MAX_NODES 8
1442e750	44
370786f9 JE	45	/* we will save the tuples of all connections we care about */
370786f9 JE	46	struct xt_connlimit_conn {
3e0d5149	47	struct hlist_node node;
8183e3a8 CG	48	struct nf_conntrack_tuple tuple;
8183e3a8 CG	49	union nf_inet_addr addr;
370786f9 JE	50	};
370786f9 JE	51
7d084877 FW	52	struct xt_connlimit_rb {
	53	struct rb_node node;
	54	struct hlist_head hhead; /* connections/hosts in same subnet */
	55	union nf_inet_addr addr; /* search key */
	56	};
	57
e00b437b FW	58	static spinlock_t xt_connlimit_locks[CONNLIMIT_LOCK_SLOTS] __cacheline_aligned_in_smp;
e00b437b FW	59
370786f9	60	struct xt_connlimit_data {
7d084877 FW	61	struct rb_root climit_root4[CONNLIMIT_SLOTS];
7d084877 FW	62	struct rb_root climit_root6[CONNLIMIT_SLOTS];
370786f9 JE	63	};
370786f9 JE	64
294188ae	65	static u_int32_t connlimit_rnd __read_mostly;
7d084877	66	static struct kmem_cache *connlimit_rb_cachep __read_mostly;
14e1a977	67	static struct kmem_cache *connlimit_conn_cachep __read_mostly;
370786f9	68
a34c4589	69	static inline unsigned int connlimit_iphash(__be32 addr)
370786f9	70	{
1442e750 FW	71	return jhash_1word((__force __u32)addr,
1442e750 FW	72	connlimit_rnd) % CONNLIMIT_SLOTS;
370786f9 JE	73	}
	74
	75	static inline unsigned int
643a2c15 JE	76	connlimit_iphash6(const union nf_inet_addr *addr,
643a2c15 JE	77	const union nf_inet_addr *mask)
370786f9	78	{
643a2c15	79	union nf_inet_addr res;
370786f9 JE	80	unsigned int i;
370786f9 JE	81
370786f9 JE	82	for (i = 0; i < ARRAY_SIZE(addr->ip6); ++i)
	83	res.ip6[i] = addr->ip6[i] & mask->ip6[i];
	84
1442e750 FW	85	return jhash2((u32 *)res.ip6, ARRAY_SIZE(res.ip6),
1442e750 FW	86	connlimit_rnd) % CONNLIMIT_SLOTS;
370786f9 JE	87	}
	88
	89	static inline bool already_closed(const struct nf_conn *conn)
	90	{
5e8fbe2a	91	if (nf_ct_protonum(conn) == IPPROTO_TCP)
d2ee3f2c DW	92	return conn->proto.tcp.state == TCP_CONNTRACK_TIME_WAIT \|\|
d2ee3f2c DW	93	conn->proto.tcp.state == TCP_CONNTRACK_CLOSE;
370786f9 JE	94	else
	95	return 0;
	96	}
	97
50e0e9b1	98	static int
643a2c15 JE	99	same_source_net(const union nf_inet_addr *addr,
643a2c15 JE	100	const union nf_inet_addr *mask,
76108cea	101	const union nf_inet_addr *u3, u_int8_t family)
370786f9	102	{
ee999d8b	103	if (family == NFPROTO_IPV4) {
50e0e9b1 FW	104	return ntohl(addr->ip & mask->ip) -
50e0e9b1 FW	105	ntohl(u3->ip & mask->ip);
370786f9	106	} else {
643a2c15	107	union nf_inet_addr lh, rh;
370786f9 JE	108	unsigned int i;
	109
	110	for (i = 0; i < ARRAY_SIZE(addr->ip6); ++i) {
	111	lh.ip6[i] = addr->ip6[i] & mask->ip6[i];
	112	rh.ip6[i] = u3->ip6[i] & mask->ip6[i];
	113	}
	114
50e0e9b1	115	return memcmp(&lh.ip6, &rh.ip6, sizeof(lh.ip6));
370786f9 JE	116	}
	117	}
	118
7d084877 FW	119	static bool add_hlist(struct hlist_head *head,
	120	const struct nf_conntrack_tuple *tuple,
	121	const union nf_inet_addr *addr)
	122	{
	123	struct xt_connlimit_conn *conn;
	124
	125	conn = kmem_cache_alloc(connlimit_conn_cachep, GFP_ATOMIC);
	126	if (conn == NULL)
	127	return false;
	128	conn->tuple = *tuple;
	129	conn->addr = *addr;
	130	hlist_add_head(&conn->node, head);
	131	return true;
	132	}
	133
	134	static unsigned int check_hlist(struct net *net,
	135	struct hlist_head *head,
	136	const struct nf_conntrack_tuple *tuple,
308ac914	137	const struct nf_conntrack_zone *zone,
7d084877	138	bool *addit)
370786f9	139	{
3cf93c96	140	const struct nf_conntrack_tuple_hash *found;
370786f9	141	struct xt_connlimit_conn *conn;
b67bfe0d	142	struct hlist_node *n;
ea781f19	143	struct nf_conn *found_ct;
7d084877	144	unsigned int length = 0;
370786f9	145
7d084877	146	*addit = true;
76507f69	147	rcu_read_lock();
370786f9 JE	148
370786f9 JE	149	/* check the saved connections */
15cfd528	150	hlist_for_each_entry_safe(conn, n, head, node) {
e59ea3df	151	found = nf_conntrack_find_get(net, zone, &conn->tuple);
d9ec4f1e FW	152	if (found == NULL) {
d9ec4f1e FW	153	hlist_del(&conn->node);
14e1a977	154	kmem_cache_free(connlimit_conn_cachep, conn);
d9ec4f1e FW	155	continue;
d9ec4f1e FW	156	}
370786f9	157
d9ec4f1e	158	found_ct = nf_ct_tuplehash_to_ctrack(found);
370786f9	159
d9ec4f1e	160	if (nf_ct_tuple_equal(&conn->tuple, tuple)) {
370786f9 JE	161	/*
	162	* Just to be sure we have it only once in the list.
	163	* We should not see tuples twice unless someone hooks
	164	* this into a table without "-p tcp --syn".
	165	*/
3bcc5fdf	166	*addit = false;
d9ec4f1e	167	} else if (already_closed(found_ct)) {
370786f9 JE	168	/*
	169	* we do not care about connections which are
	170	* closed already -> ditch it
	171	*/
ea781f19	172	nf_ct_put(found_ct);
3e0d5149	173	hlist_del(&conn->node);
14e1a977	174	kmem_cache_free(connlimit_conn_cachep, conn);
370786f9 JE	175	continue;
	176	}
	177
ea781f19	178	nf_ct_put(found_ct);
7d084877	179	length++;
370786f9 JE	180	}
370786f9 JE	181
76507f69	182	rcu_read_unlock();
370786f9	183
7d084877	184	return length;
370786f9 JE	185	}
370786f9 JE	186
7d084877 FW	187	static void tree_nodes_free(struct rb_root *root,
	188	struct xt_connlimit_rb *gc_nodes[],
	189	unsigned int gc_count)
	190	{
	191	struct xt_connlimit_rb *rbconn;
	192
	193	while (gc_count) {
	194	rbconn = gc_nodes[--gc_count];
	195	rb_erase(&rbconn->node, root);
	196	kmem_cache_free(connlimit_rb_cachep, rbconn);
	197	}
	198	}
	199
	200	static unsigned int
	201	count_tree(struct net net, struct rb_root root,
	202	const struct nf_conntrack_tuple *tuple,
	203	const union nf_inet_addr addr, const union nf_inet_addr mask,
308ac914	204	u8 family, const struct nf_conntrack_zone *zone)
3bcc5fdf	205	{
7d084877 FW	206	struct xt_connlimit_rb *gc_nodes[CONNLIMIT_GC_MAX_NODES];
	207	struct rb_node *rbnode, parent;
	208	struct xt_connlimit_rb *rbconn;
14e1a977	209	struct xt_connlimit_conn *conn;
7d084877 FW	210	unsigned int gc_count;
	211	bool no_gc = false;
	212
	213	restart:
	214	gc_count = 0;
	215	parent = NULL;
	216	rbnode = &(root->rb_node);
	217	while (*rbnode) {
	218	int diff;
	219	bool addit;
	220
4cc4b72c	221	rbconn = rb_entry(*rbnode, struct xt_connlimit_rb, node);
7d084877 FW	222
	223	parent = *rbnode;
	224	diff = same_source_net(addr, mask, &rbconn->addr, family);
	225	if (diff < 0) {
	226	rbnode = &((*rbnode)->rb_left);
	227	} else if (diff > 0) {
	228	rbnode = &((*rbnode)->rb_right);
	229	} else {
	230	/* same source network -> be counted! */
	231	unsigned int count;
e59ea3df	232	count = check_hlist(net, &rbconn->hhead, tuple, zone, &addit);
7d084877 FW	233
	234	tree_nodes_free(root, gc_nodes, gc_count);
	235	if (!addit)
	236	return count;
	237
	238	if (!add_hlist(&rbconn->hhead, tuple, addr))
	239	return 0; /* hotdrop */
	240
	241	return count + 1;
	242	}
	243
	244	if (no_gc \|\| gc_count >= ARRAY_SIZE(gc_nodes))
	245	continue;
	246
	247	/* only used for GC on hhead, retval and 'addit' ignored */
e59ea3df	248	check_hlist(net, &rbconn->hhead, tuple, zone, &addit);
7d084877 FW	249	if (hlist_empty(&rbconn->hhead))
	250	gc_nodes[gc_count++] = rbconn;
	251	}
	252
	253	if (gc_count) {
	254	no_gc = true;
	255	tree_nodes_free(root, gc_nodes, gc_count);
	256	/* tree_node_free before new allocation permits
	257	* allocator to re-use newly free'd object.
	258	*
	259	* This is a rare event; in most cases we will find
	260	* existing node to re-use. (or gc_count is 0).
	261	*/
	262	goto restart;
	263	}
	264
	265	/* no match, need to insert new node */
	266	rbconn = kmem_cache_alloc(connlimit_rb_cachep, GFP_ATOMIC);
	267	if (rbconn == NULL)
	268	return 0;
14e1a977 FW	269
14e1a977 FW	270	conn = kmem_cache_alloc(connlimit_conn_cachep, GFP_ATOMIC);
7d084877 FW	271	if (conn == NULL) {
	272	kmem_cache_free(connlimit_rb_cachep, rbconn);
	273	return 0;
	274	}
	275
3bcc5fdf FW	276	conn->tuple = *tuple;
3bcc5fdf FW	277	conn->addr = *addr;
7d084877 FW	278	rbconn->addr = *addr;
	279
	280	INIT_HLIST_HEAD(&rbconn->hhead);
	281	hlist_add_head(&conn->node, &rbconn->hhead);
	282
	283	rb_link_node(&rbconn->node, parent, rbnode);
	284	rb_insert_color(&rbconn->node, root);
	285	return 1;
3bcc5fdf FW	286	}
3bcc5fdf FW	287
15cfd528 FW	288	static int count_them(struct net *net,
	289	struct xt_connlimit_data *data,
	290	const struct nf_conntrack_tuple *tuple,
	291	const union nf_inet_addr *addr,
	292	const union nf_inet_addr *mask,
308ac914 DB	293	u_int8_t family,
308ac914 DB	294	const struct nf_conntrack_zone *zone)
15cfd528	295	{
7d084877	296	struct rb_root *root;
15cfd528 FW	297	int count;
	298	u32 hash;
	299
7d084877	300	if (family == NFPROTO_IPV6) {
15cfd528	301	hash = connlimit_iphash6(addr, mask);
7d084877 FW	302	root = &data->climit_root6[hash];
7d084877 FW	303	} else {
15cfd528	304	hash = connlimit_iphash(addr->ip & mask->ip);
7d084877 FW	305	root = &data->climit_root4[hash];
7d084877 FW	306	}
15cfd528	307
e00b437b	308	spin_lock_bh(&xt_connlimit_locks[hash % CONNLIMIT_LOCK_SLOTS]);
7d084877	309
e59ea3df	310	count = count_tree(net, root, tuple, addr, mask, family, zone);
7d084877	311
e00b437b	312	spin_unlock_bh(&xt_connlimit_locks[hash % CONNLIMIT_LOCK_SLOTS]);
15cfd528 FW	313
	314	return count;
	315	}
	316
d3c5ee6d	317	static bool
62fc8051	318	connlimit_mt(const struct sk_buff skb, struct xt_action_param par)
370786f9	319	{
613dbd95	320	struct net *net = xt_net(par);
f7108a20	321	const struct xt_connlimit_info *info = par->matchinfo;
22c2d8bc	322	union nf_inet_addr addr;
370786f9 JE	323	struct nf_conntrack_tuple tuple;
370786f9 JE	324	const struct nf_conntrack_tuple *tuple_ptr = &tuple;
308ac914	325	const struct nf_conntrack_zone *zone = &nf_ct_zone_dflt;
370786f9 JE	326	enum ip_conntrack_info ctinfo;
370786f9 JE	327	const struct nf_conn *ct;
7d084877	328	unsigned int connections;
370786f9 JE	329
370786f9 JE	330	ct = nf_ct_get(skb, &ctinfo);
e59ea3df	331	if (ct != NULL) {
8183e3a8	332	tuple_ptr = &ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple;
e59ea3df FW	333	zone = nf_ct_zone(ct);
e59ea3df FW	334	} else if (!nf_ct_get_tuplepr(skb, skb_network_offset(skb),
613dbd95	335	xt_family(par), net, &tuple)) {
370786f9	336	goto hotdrop;
e59ea3df	337	}
370786f9	338
613dbd95	339	if (xt_family(par) == NFPROTO_IPV6) {
370786f9	340	const struct ipv6hdr *iph = ipv6_hdr(skb);
cc4fc022 JE	341	memcpy(&addr.ip6, (info->flags & XT_CONNLIMIT_DADDR) ?
cc4fc022 JE	342	&iph->daddr : &iph->saddr, sizeof(addr.ip6));
370786f9 JE	343	} else {
370786f9 JE	344	const struct iphdr *iph = ip_hdr(skb);
cc4fc022 JE	345	addr.ip = (info->flags & XT_CONNLIMIT_DADDR) ?
cc4fc022 JE	346	iph->daddr : iph->saddr;
370786f9 JE	347	}
370786f9 JE	348
83fc8102	349	connections = count_them(net, info->data, tuple_ptr, &addr,
613dbd95	350	&info->mask, xt_family(par), zone);
7d084877	351	if (connections == 0)
370786f9	352	/* kmalloc failed, drop it entirely */
1cc34c30	353	goto hotdrop;
370786f9	354
cc4fc022 JE	355	return (connections > info->limit) ^
cc4fc022 JE	356	!!(info->flags & XT_CONNLIMIT_INVERT);
370786f9 JE	357
370786f9 JE	358	hotdrop:
b4ba2611	359	par->hotdrop = true;
370786f9 JE	360	return false;
	361	}
	362
b0f38452	363	static int connlimit_mt_check(const struct xt_mtchk_param *par)
370786f9	364	{
9b4fce7a	365	struct xt_connlimit_info *info = par->matchinfo;
370786f9	366	unsigned int i;
4a5a5c73	367	int ret;
370786f9	368
7bdc6624	369	net_get_random_once(&connlimit_rnd, sizeof(connlimit_rnd));
4656c4d6	370
ecb2421b	371	ret = nf_ct_netns_get(par->net, par->family);
4a5a5c73	372	if (ret < 0) {
8bee4bad JE	373	pr_info("cannot load conntrack support for "
8bee4bad JE	374	"address family %u\n", par->family);
4a5a5c73	375	return ret;
370786f9 JE	376	}
	377
	378	/* init private data */
	379	info->data = kmalloc(sizeof(struct xt_connlimit_data), GFP_KERNEL);
	380	if (info->data == NULL) {
ecb2421b	381	nf_ct_netns_put(par->net, par->family);
4a5a5c73	382	return -ENOMEM;
370786f9 JE	383	}
370786f9 JE	384
7d084877 FW	385	for (i = 0; i < ARRAY_SIZE(info->data->climit_root4); ++i)
	386	info->data->climit_root4[i] = RB_ROOT;
	387	for (i = 0; i < ARRAY_SIZE(info->data->climit_root6); ++i)
	388	info->data->climit_root6[i] = RB_ROOT;
370786f9	389
bd414ee6	390	return 0;
370786f9 JE	391	}
370786f9 JE	392
7d084877	393	static void destroy_tree(struct rb_root *r)
370786f9	394	{
370786f9	395	struct xt_connlimit_conn *conn;
7d084877	396	struct xt_connlimit_rb *rbconn;
b67bfe0d	397	struct hlist_node *n;
7d084877	398	struct rb_node *node;
370786f9	399
7d084877	400	while ((node = rb_first(r)) != NULL) {
4cc4b72c	401	rbconn = rb_entry(node, struct xt_connlimit_rb, node);
370786f9	402
7d084877 FW	403	rb_erase(node, r);
	404
	405	hlist_for_each_entry_safe(conn, n, &rbconn->hhead, node)
14e1a977	406	kmem_cache_free(connlimit_conn_cachep, conn);
7d084877 FW	407
7d084877 FW	408	kmem_cache_free(connlimit_rb_cachep, rbconn);
370786f9	409	}
7d084877 FW	410	}
	411
	412	static void connlimit_mt_destroy(const struct xt_mtdtor_param *par)
	413	{
	414	const struct xt_connlimit_info *info = par->matchinfo;
	415	unsigned int i;
	416
ecb2421b	417	nf_ct_netns_put(par->net, par->family);
7d084877 FW	418
	419	for (i = 0; i < ARRAY_SIZE(info->data->climit_root4); ++i)
	420	destroy_tree(&info->data->climit_root4[i]);
	421	for (i = 0; i < ARRAY_SIZE(info->data->climit_root6); ++i)
	422	destroy_tree(&info->data->climit_root6[i]);
370786f9 JE	423
	424	kfree(info->data);
	425	}
	426
68c07cb6 CW	427	static struct xt_match connlimit_mt_reg __read_mostly = {
	428	.name = "connlimit",
	429	.revision = 1,
	430	.family = NFPROTO_UNSPEC,
	431	.checkentry = connlimit_mt_check,
	432	.match = connlimit_mt,
	433	.matchsize = sizeof(struct xt_connlimit_info),
ec231890	434	.usersize = offsetof(struct xt_connlimit_info, data),
68c07cb6 CW	435	.destroy = connlimit_mt_destroy,
68c07cb6 CW	436	.me = THIS_MODULE,
370786f9 JE	437	};
370786f9 JE	438
d3c5ee6d	439	static int __init connlimit_mt_init(void)
370786f9	440	{
e00b437b	441	int ret, i;
1442e750 FW	442
	443	BUILD_BUG_ON(CONNLIMIT_LOCK_SLOTS > CONNLIMIT_SLOTS);
	444	BUILD_BUG_ON((CONNLIMIT_SLOTS % CONNLIMIT_LOCK_SLOTS) != 0);
	445
e00b437b FW	446	for (i = 0; i < CONNLIMIT_LOCK_SLOTS; ++i)
	447	spin_lock_init(&xt_connlimit_locks[i]);
	448
14e1a977 FW	449	connlimit_conn_cachep = kmem_cache_create("xt_connlimit_conn",
	450	sizeof(struct xt_connlimit_conn),
	451	0, 0, NULL);
	452	if (!connlimit_conn_cachep)
	453	return -ENOMEM;
	454
7d084877 FW	455	connlimit_rb_cachep = kmem_cache_create("xt_connlimit_rb",
	456	sizeof(struct xt_connlimit_rb),
	457	0, 0, NULL);
	458	if (!connlimit_rb_cachep) {
	459	kmem_cache_destroy(connlimit_conn_cachep);
	460	return -ENOMEM;
	461	}
14e1a977	462	ret = xt_register_match(&connlimit_mt_reg);
7d084877	463	if (ret != 0) {
14e1a977	464	kmem_cache_destroy(connlimit_conn_cachep);
7d084877 FW	465	kmem_cache_destroy(connlimit_rb_cachep);
7d084877 FW	466	}
14e1a977	467	return ret;
370786f9 JE	468	}
370786f9 JE	469
d3c5ee6d	470	static void __exit connlimit_mt_exit(void)
370786f9	471	{
68c07cb6	472	xt_unregister_match(&connlimit_mt_reg);
14e1a977	473	kmem_cache_destroy(connlimit_conn_cachep);
7d084877	474	kmem_cache_destroy(connlimit_rb_cachep);
370786f9 JE	475	}
370786f9 JE	476
d3c5ee6d JE	477	module_init(connlimit_mt_init);
d3c5ee6d JE	478	module_exit(connlimit_mt_exit);
92f3b2b1	479	MODULE_AUTHOR("Jan Engelhardt <jengelh@medozas.de>");
2ae15b64	480	MODULE_DESCRIPTION("Xtables: Number of connections matching");
370786f9 JE	481	MODULE_LICENSE("GPL");
	482	MODULE_ALIAS("ipt_connlimit");
	483	MODULE_ALIAS("ip6t_connlimit");