[deliverable/linux.git] / net / netfilter / core.c

/* netfilter.c: look after the filters for various protocols.
 * Heavily influenced by the old firewall.c by David Bonn and Alan Cox.
 *
 * Thanks to Rob `CmdrTaco' Malda for not influencing this code in any
 * way.
 *
 * Rusty Russell (C)2000 -- This code is GPL.
 */
#include <linux/kernel.h>
#include <linux/netfilter.h>
#include <net/protocol.h>
#include <linux/init.h>
#include <linux/skbuff.h>
#include <linux/wait.h>
#include <linux/module.h>
#include <linux/interrupt.h>
#include <linux/if.h>
#include <linux/netdevice.h>
#include <linux/inetdevice.h>
#include <linux/proc_fs.h>
#include <linux/mutex.h>
#include <linux/slab.h>
#include <net/net_namespace.h>
#include <net/sock.h>

#include "nf_internals.h"

static DEFINE_MUTEX(afinfo_mutex);

const struct nf_afinfo __rcu *nf_afinfo[NFPROTO_NUMPROTO] __read_mostly;
EXPORT_SYMBOL(nf_afinfo);

int nf_register_afinfo(const struct nf_afinfo *afinfo)
{
	int err;

	err = mutex_lock_interruptible(&afinfo_mutex);
	if (err < 0)
		return err;
	RCU_INIT_POINTER(nf_afinfo[afinfo->family], afinfo);
	mutex_unlock(&afinfo_mutex);
	return 0;
}
EXPORT_SYMBOL_GPL(nf_register_afinfo);

void nf_unregister_afinfo(const struct nf_afinfo *afinfo)
{
	mutex_lock(&afinfo_mutex);
	RCU_INIT_POINTER(nf_afinfo[afinfo->family], NULL);
	mutex_unlock(&afinfo_mutex);
	synchronize_rcu();
}
EXPORT_SYMBOL_GPL(nf_unregister_afinfo);

struct list_head nf_hooks[NFPROTO_NUMPROTO][NF_MAX_HOOKS] __read_mostly;
EXPORT_SYMBOL(nf_hooks);

#if defined(CONFIG_JUMP_LABEL)
struct static_key nf_hooks_needed[NFPROTO_NUMPROTO][NF_MAX_HOOKS];
EXPORT_SYMBOL(nf_hooks_needed);
#endif

static DEFINE_MUTEX(nf_hook_mutex);

int nf_register_hook(struct nf_hook_ops *reg)
{
	struct nf_hook_ops *elem;
	int err;

	err = mutex_lock_interruptible(&nf_hook_mutex);
	if (err < 0)
		return err;
	list_for_each_entry(elem, &nf_hooks[reg->pf][reg->hooknum], list) {
		if (reg->priority < elem->priority)
			break;
	}
	list_add_rcu(&reg->list, elem->list.prev);
	mutex_unlock(&nf_hook_mutex);
#if defined(CONFIG_JUMP_LABEL)
	static_key_slow_inc(&nf_hooks_needed[reg->pf][reg->hooknum]);
#endif
	return 0;
}
EXPORT_SYMBOL(nf_register_hook);

void nf_unregister_hook(struct nf_hook_ops *reg)
{
	mutex_lock(&nf_hook_mutex);
	list_del_rcu(&reg->list);
	mutex_unlock(&nf_hook_mutex);
#if defined(CONFIG_JUMP_LABEL)
	static_key_slow_dec(&nf_hooks_needed[reg->pf][reg->hooknum]);
#endif
	synchronize_net();
}
EXPORT_SYMBOL(nf_unregister_hook);

int nf_register_hooks(struct nf_hook_ops *reg, unsigned int n)
{
	unsigned int i;
	int err = 0;

	for (i = 0; i < n; i++) {
		err = nf_register_hook(&reg[i]);
		if (err)
			goto err;
	}
	return err;

err:
	if (i > 0)
		nf_unregister_hooks(reg, i);
	return err;
}
EXPORT_SYMBOL(nf_register_hooks);

void nf_unregister_hooks(struct nf_hook_ops *reg, unsigned int n)
{
	while (n-- > 0)
		nf_unregister_hook(&reg[n]);
}
EXPORT_SYMBOL(nf_unregister_hooks);

unsigned int nf_iterate(struct list_head *head,
			struct sk_buff *skb,
			unsigned int hook,
			const struct net_device *indev,
			const struct net_device *outdev,
			struct nf_hook_ops **elemp,
			int (*okfn)(struct sk_buff *),
			int hook_thresh)
{
	unsigned int verdict;

	/*
	 * The caller must not block between calls to this
	 * function because of risk of continuing from deleted element.
	 */
	list_for_each_entry_continue_rcu((*elemp), head, list) {
		if (hook_thresh > (*elemp)->priority)
			continue;

		/* Optimization: we don't need to hold module
		   reference here, since function can't sleep. --RR */
repeat:
		verdict = (*elemp)->hook(hook, skb, indev, outdev, okfn);
		if (verdict != NF_ACCEPT) {
#ifdef CONFIG_NETFILTER_DEBUG
			if (unlikely((verdict & NF_VERDICT_MASK)
							> NF_MAX_VERDICT)) {
				NFDEBUG("Evil return from %p(%u).\n",
					(*elemp)->hook, hook);
				continue;
			}
#endif
			if (verdict != NF_REPEAT)
				return verdict;
			goto repeat;
		}
	}
	return NF_ACCEPT;
}


/* Returns 1 if okfn() needs to be executed by the caller,
 * -EPERM for NF_DROP, 0 otherwise. */
int nf_hook_slow(u_int8_t pf, unsigned int hook, struct sk_buff *skb,
		 struct net_device *indev,
		 struct net_device *outdev,
		 int (*okfn)(struct sk_buff *),
		 int hook_thresh)
{
	struct nf_hook_ops *elem;
	unsigned int verdict;
	int ret = 0;

	/* We may already have this, but read-locks nest anyway */
	rcu_read_lock();

	elem = list_entry_rcu(&nf_hooks[pf][hook], struct nf_hook_ops, list);
next_hook:
	verdict = nf_iterate(&nf_hooks[pf][hook], skb, hook, indev,
			     outdev, &elem, okfn, hook_thresh);
	if (verdict == NF_ACCEPT || verdict == NF_STOP) {
		ret = 1;
	} else if ((verdict & NF_VERDICT_MASK) == NF_DROP) {
		kfree_skb(skb);
		ret = NF_DROP_GETERR(verdict);
		if (ret == 0)
			ret = -EPERM;
	} else if ((verdict & NF_VERDICT_MASK) == NF_QUEUE) {
		int err = nf_queue(skb, elem, pf, hook, indev, outdev, okfn,
						verdict >> NF_VERDICT_QBITS);
		if (err < 0) {
			if (err == -ECANCELED)
				goto next_hook;
			if (err == -ESRCH &&
			   (verdict & NF_VERDICT_FLAG_QUEUE_BYPASS))
				goto next_hook;
			kfree_skb(skb);
		}
	}
	rcu_read_unlock();
	return ret;
}
EXPORT_SYMBOL(nf_hook_slow);


int skb_make_writable(struct sk_buff *skb, unsigned int writable_len)
{
	if (writable_len > skb->len)
		return 0;

	/* Not exclusive use of packet?  Must copy. */
	if (!skb_cloned(skb)) {
		if (writable_len <= skb_headlen(skb))
			return 1;
	} else if (skb_clone_writable(skb, writable_len))
		return 1;

	if (writable_len <= skb_headlen(skb))
		writable_len = 0;
	else
		writable_len -= skb_headlen(skb);

	return !!__pskb_pull_tail(skb, writable_len);
}
EXPORT_SYMBOL(skb_make_writable);

#if IS_ENABLED(CONFIG_NF_CONNTRACK)
/* This does not belong here, but locally generated errors need it if connection
   tracking in use: without this, connection may not be in hash table, and hence
   manufactured ICMP or RST packets will not be associated with it. */
void (*ip_ct_attach)(struct sk_buff *, struct sk_buff *) __rcu __read_mostly;
EXPORT_SYMBOL(ip_ct_attach);

void nf_ct_attach(struct sk_buff *new, struct sk_buff *skb)
{
	void (*attach)(struct sk_buff *, struct sk_buff *);

	if (skb->nfct) {
		rcu_read_lock();
		attach = rcu_dereference(ip_ct_attach);
		if (attach)
			attach(new, skb);
		rcu_read_unlock();
	}
}
EXPORT_SYMBOL(nf_ct_attach);

void (*nf_ct_destroy)(struct nf_conntrack *) __rcu __read_mostly;
EXPORT_SYMBOL(nf_ct_destroy);

void nf_conntrack_destroy(struct nf_conntrack *nfct)
{
	void (*destroy)(struct nf_conntrack *);

	rcu_read_lock();
	destroy = rcu_dereference(nf_ct_destroy);
	BUG_ON(destroy == NULL);
	destroy(nfct);
	rcu_read_unlock();
}
EXPORT_SYMBOL(nf_conntrack_destroy);

struct nfq_ct_hook __rcu *nfq_ct_hook __read_mostly;
EXPORT_SYMBOL_GPL(nfq_ct_hook);

struct nfq_ct_nat_hook __rcu *nfq_ct_nat_hook __read_mostly;
EXPORT_SYMBOL_GPL(nfq_ct_nat_hook);

#endif /* CONFIG_NF_CONNTRACK */

#ifdef CONFIG_NF_NAT_NEEDED
void (*nf_nat_decode_session_hook)(struct sk_buff *, struct flowi *);
EXPORT_SYMBOL(nf_nat_decode_session_hook);
#endif

#ifdef CONFIG_PROC_FS
struct proc_dir_entry *proc_net_netfilter;
EXPORT_SYMBOL(proc_net_netfilter);
#endif

static int __net_init netfilter_net_init(struct net *net)
{
#ifdef CONFIG_PROC_FS
	net->nf.proc_netfilter = proc_net_mkdir(net, "netfilter",
						net->proc_net);
	if (net_eq(net, &init_net)) {
		if (!net->nf.proc_netfilter)
			return -ENOMEM;
		else
			proc_net_netfilter = net->nf.proc_netfilter;
	} else if (!net->nf.proc_netfilter) {
		pr_err("cannot create netfilter proc entry");
		return -ENOMEM;
	}
#endif
	return 0;
}

static void __net_exit netfilter_net_exit(struct net *net)
{
	remove_proc_entry("netfilter", net->proc_net);
}

static struct pernet_operations netfilter_net_ops = {
	.init = netfilter_net_init,
	.exit = netfilter_net_exit,
};

void __init netfilter_init(void)
{
	int i, h;
	for (i = 0; i < ARRAY_SIZE(nf_hooks); i++) {
		for (h = 0; h < NF_MAX_HOOKS; h++)
			INIT_LIST_HEAD(&nf_hooks[i][h]);
	}

	if (register_pernet_subsys(&netfilter_net_ops) < 0)
		panic("cannot create netfilter proc entry");

	if (netfilter_log_init() < 0)
		panic("cannot initialize nf_log");
}
Commit	Line	Data
	1	/* netfilter.c: look after the filters for various protocols.
	2	* Heavily influenced by the old firewall.c by David Bonn and Alan Cox.
	3	*
	4	* Thanks to Rob `CmdrTaco' Malda for not influencing this code in any
	5	* way.
	6	*
	7	* Rusty Russell (C)2000 -- This code is GPL.
	8	*/
	9	#include <linux/kernel.h>
	10	#include <linux/netfilter.h>
	11	#include <net/protocol.h>
	12	#include <linux/init.h>
	13	#include <linux/skbuff.h>
	14	#include <linux/wait.h>
	15	#include <linux/module.h>
	16	#include <linux/interrupt.h>
	17	#include <linux/if.h>
	18	#include <linux/netdevice.h>
	19	#include <linux/inetdevice.h>
	20	#include <linux/proc_fs.h>
	21	#include <linux/mutex.h>
	22	#include <linux/slab.h>
	23	#include <net/net_namespace.h>
	24	#include <net/sock.h>
	25
	26	#include "nf_internals.h"
	27
	28	static DEFINE_MUTEX(afinfo_mutex);
	29
	30	const struct nf_afinfo __rcu *nf_afinfo[NFPROTO_NUMPROTO] __read_mostly;
	31	EXPORT_SYMBOL(nf_afinfo);
	32
	33	int nf_register_afinfo(const struct nf_afinfo *afinfo)
	34	{
	35	int err;
	36
	37	err = mutex_lock_interruptible(&afinfo_mutex);
	38	if (err < 0)
	39	return err;
	40	RCU_INIT_POINTER(nf_afinfo[afinfo->family], afinfo);
	41	mutex_unlock(&afinfo_mutex);
	42	return 0;
	43	}
	44	EXPORT_SYMBOL_GPL(nf_register_afinfo);
	45
	46	void nf_unregister_afinfo(const struct nf_afinfo *afinfo)
	47	{
	48	mutex_lock(&afinfo_mutex);
	49	RCU_INIT_POINTER(nf_afinfo[afinfo->family], NULL);
	50	mutex_unlock(&afinfo_mutex);
	51	synchronize_rcu();
	52	}
	53	EXPORT_SYMBOL_GPL(nf_unregister_afinfo);
	54
	55	struct list_head nf_hooks[NFPROTO_NUMPROTO][NF_MAX_HOOKS] __read_mostly;
	56	EXPORT_SYMBOL(nf_hooks);
	57
	58	#if defined(CONFIG_JUMP_LABEL)
	59	struct static_key nf_hooks_needed[NFPROTO_NUMPROTO][NF_MAX_HOOKS];
	60	EXPORT_SYMBOL(nf_hooks_needed);
	61	#endif
	62
	63	static DEFINE_MUTEX(nf_hook_mutex);
	64
	65	int nf_register_hook(struct nf_hook_ops *reg)
	66	{
	67	struct nf_hook_ops *elem;
	68	int err;
	69
	70	err = mutex_lock_interruptible(&nf_hook_mutex);
	71	if (err < 0)
	72	return err;
	73	list_for_each_entry(elem, &nf_hooks[reg->pf][reg->hooknum], list) {
	74	if (reg->priority < elem->priority)
	75	break;
	76	}
	77	list_add_rcu(&reg->list, elem->list.prev);
	78	mutex_unlock(&nf_hook_mutex);
	79	#if defined(CONFIG_JUMP_LABEL)
	80	static_key_slow_inc(&nf_hooks_needed[reg->pf][reg->hooknum]);
	81	#endif
	82	return 0;
	83	}
	84	EXPORT_SYMBOL(nf_register_hook);
	85
	86	void nf_unregister_hook(struct nf_hook_ops *reg)
	87	{
	88	mutex_lock(&nf_hook_mutex);
	89	list_del_rcu(&reg->list);
	90	mutex_unlock(&nf_hook_mutex);
	91	#if defined(CONFIG_JUMP_LABEL)
	92	static_key_slow_dec(&nf_hooks_needed[reg->pf][reg->hooknum]);
	93	#endif
	94	synchronize_net();
	95	}
	96	EXPORT_SYMBOL(nf_unregister_hook);
	97
	98	int nf_register_hooks(struct nf_hook_ops *reg, unsigned int n)
	99	{
	100	unsigned int i;
	101	int err = 0;
	102
	103	for (i = 0; i < n; i++) {
	104	err = nf_register_hook(&reg[i]);
	105	if (err)
	106	goto err;
	107	}
	108	return err;
	109
	110	err:
	111	if (i > 0)
	112	nf_unregister_hooks(reg, i);
	113	return err;
	114	}
	115	EXPORT_SYMBOL(nf_register_hooks);
	116
	117	void nf_unregister_hooks(struct nf_hook_ops *reg, unsigned int n)
	118	{
	119	while (n-- > 0)
	120	nf_unregister_hook(&reg[n]);
	121	}
	122	EXPORT_SYMBOL(nf_unregister_hooks);
	123
	124	unsigned int nf_iterate(struct list_head *head,
	125	struct sk_buff *skb,
	126	unsigned int hook,
	127	const struct net_device *indev,
	128	const struct net_device *outdev,
	129	struct nf_hook_ops **elemp,
	130	int (okfn)(struct sk_buff ),
	131	int hook_thresh)
	132	{
	133	unsigned int verdict;
	134
	135	/*
	136	* The caller must not block between calls to this
	137	* function because of risk of continuing from deleted element.
	138	*/
	139	list_for_each_entry_continue_rcu((*elemp), head, list) {
	140	if (hook_thresh > (*elemp)->priority)
	141	continue;
	142
	143	/* Optimization: we don't need to hold module
	144	reference here, since function can't sleep. --RR */
	145	repeat:
	146	verdict = (*elemp)->hook(hook, skb, indev, outdev, okfn);
	147	if (verdict != NF_ACCEPT) {
	148	#ifdef CONFIG_NETFILTER_DEBUG
	149	if (unlikely((verdict & NF_VERDICT_MASK)
	150	> NF_MAX_VERDICT)) {
	151	NFDEBUG("Evil return from %p(%u).\n",
	152	(*elemp)->hook, hook);
	153	continue;
	154	}
	155	#endif
	156	if (verdict != NF_REPEAT)
	157	return verdict;
	158	goto repeat;
	159	}
	160	}
	161	return NF_ACCEPT;
	162	}
	163
	164
	165	/* Returns 1 if okfn() needs to be executed by the caller,
	166	* -EPERM for NF_DROP, 0 otherwise. */
	167	int nf_hook_slow(u_int8_t pf, unsigned int hook, struct sk_buff *skb,
	168	struct net_device *indev,
	169	struct net_device *outdev,
	170	int (okfn)(struct sk_buff ),
	171	int hook_thresh)
	172	{
	173	struct nf_hook_ops *elem;
	174	unsigned int verdict;
	175	int ret = 0;
	176
	177	/* We may already have this, but read-locks nest anyway */
	178	rcu_read_lock();
	179
	180	elem = list_entry_rcu(&nf_hooks[pf][hook], struct nf_hook_ops, list);
	181	next_hook:
	182	verdict = nf_iterate(&nf_hooks[pf][hook], skb, hook, indev,
	183	outdev, &elem, okfn, hook_thresh);
	184	if (verdict == NF_ACCEPT \|\| verdict == NF_STOP) {
	185	ret = 1;
	186	} else if ((verdict & NF_VERDICT_MASK) == NF_DROP) {
	187	kfree_skb(skb);
	188	ret = NF_DROP_GETERR(verdict);
	189	if (ret == 0)
	190	ret = -EPERM;
	191	} else if ((verdict & NF_VERDICT_MASK) == NF_QUEUE) {
	192	int err = nf_queue(skb, elem, pf, hook, indev, outdev, okfn,
	193	verdict >> NF_VERDICT_QBITS);
	194	if (err < 0) {
	195	if (err == -ECANCELED)
	196	goto next_hook;
	197	if (err == -ESRCH &&
	198	(verdict & NF_VERDICT_FLAG_QUEUE_BYPASS))
	199	goto next_hook;
	200	kfree_skb(skb);
	201	}
	202	}
	203	rcu_read_unlock();
	204	return ret;
	205	}
	206	EXPORT_SYMBOL(nf_hook_slow);
	207
	208
	209	int skb_make_writable(struct sk_buff *skb, unsigned int writable_len)
	210	{
	211	if (writable_len > skb->len)
	212	return 0;
	213
	214	/* Not exclusive use of packet? Must copy. */
	215	if (!skb_cloned(skb)) {
	216	if (writable_len <= skb_headlen(skb))
	217	return 1;
	218	} else if (skb_clone_writable(skb, writable_len))
	219	return 1;
	220
	221	if (writable_len <= skb_headlen(skb))
	222	writable_len = 0;
	223	else
	224	writable_len -= skb_headlen(skb);
	225
	226	return !!__pskb_pull_tail(skb, writable_len);
	227	}
	228	EXPORT_SYMBOL(skb_make_writable);
	229
	230	#if IS_ENABLED(CONFIG_NF_CONNTRACK)
	231	/* This does not belong here, but locally generated errors need it if connection
	232	tracking in use: without this, connection may not be in hash table, and hence
	233	manufactured ICMP or RST packets will not be associated with it. */
	234	void (ip_ct_attach)(struct sk_buff , struct sk_buff *) __rcu __read_mostly;
	235	EXPORT_SYMBOL(ip_ct_attach);
	236
	237	void nf_ct_attach(struct sk_buff new, struct sk_buff skb)
	238	{
	239	void (attach)(struct sk_buff , struct sk_buff *);
	240
	241	if (skb->nfct) {
	242	rcu_read_lock();
	243	attach = rcu_dereference(ip_ct_attach);
	244	if (attach)
	245	attach(new, skb);
	246	rcu_read_unlock();
	247	}
	248	}
	249	EXPORT_SYMBOL(nf_ct_attach);
	250
	251	void (nf_ct_destroy)(struct nf_conntrack ) __rcu __read_mostly;
	252	EXPORT_SYMBOL(nf_ct_destroy);
	253
	254	void nf_conntrack_destroy(struct nf_conntrack *nfct)
	255	{
	256	void (destroy)(struct nf_conntrack );
	257
	258	rcu_read_lock();
	259	destroy = rcu_dereference(nf_ct_destroy);
	260	BUG_ON(destroy == NULL);
	261	destroy(nfct);
	262	rcu_read_unlock();
	263	}
	264	EXPORT_SYMBOL(nf_conntrack_destroy);
	265
	266	struct nfq_ct_hook __rcu *nfq_ct_hook __read_mostly;
	267	EXPORT_SYMBOL_GPL(nfq_ct_hook);
	268
	269	struct nfq_ct_nat_hook __rcu *nfq_ct_nat_hook __read_mostly;
	270	EXPORT_SYMBOL_GPL(nfq_ct_nat_hook);
	271
	272	#endif /* CONFIG_NF_CONNTRACK */
	273
	274	#ifdef CONFIG_NF_NAT_NEEDED
	275	void (nf_nat_decode_session_hook)(struct sk_buff , struct flowi *);
	276	EXPORT_SYMBOL(nf_nat_decode_session_hook);
	277	#endif
	278
	279	#ifdef CONFIG_PROC_FS
	280	struct proc_dir_entry *proc_net_netfilter;
	281	EXPORT_SYMBOL(proc_net_netfilter);
	282	#endif
	283
	284	static int __net_init netfilter_net_init(struct net *net)
	285	{
	286	#ifdef CONFIG_PROC_FS
	287	net->nf.proc_netfilter = proc_net_mkdir(net, "netfilter",
	288	net->proc_net);
	289	if (net_eq(net, &init_net)) {
	290	if (!net->nf.proc_netfilter)
	291	return -ENOMEM;
	292	else
	293	proc_net_netfilter = net->nf.proc_netfilter;
	294	} else if (!net->nf.proc_netfilter) {
	295	pr_err("cannot create netfilter proc entry");
	296	return -ENOMEM;
	297	}
	298	#endif
	299	return 0;
	300	}
	301
	302	static void __net_exit netfilter_net_exit(struct net *net)
	303	{
	304	remove_proc_entry("netfilter", net->proc_net);
	305	}
	306
	307	static struct pernet_operations netfilter_net_ops = {
	308	.init = netfilter_net_init,
	309	.exit = netfilter_net_exit,
	310	};
	311
	312	void __init netfilter_init(void)
	313	{
	314	int i, h;
	315	for (i = 0; i < ARRAY_SIZE(nf_hooks); i++) {
	316	for (h = 0; h < NF_MAX_HOOKS; h++)
	317	INIT_LIST_HEAD(&nf_hooks[i][h]);
	318	}
	319
	320	if (register_pernet_subsys(&netfilter_net_ops) < 0)
	321	panic("cannot create netfilter proc entry");
	322
	323	if (netfilter_log_init() < 0)
	324	panic("cannot initialize nf_log");
	325	}