[net-next-2.6.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 <net/sock.h>

#include "nf_internals.h"

static DEFINE_MUTEX(afinfo_mutex);

struct nf_afinfo *nf_afinfo[NPROTO] __read_mostly;
EXPORT_SYMBOL(nf_afinfo);

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

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

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

/* In this code, we can be waiting indefinitely for userspace to
 * service a packet if a hook returns NF_QUEUE.  We could keep a count
 * of skbuffs queued for userspace, and not deregister a hook unless
 * this is zero, but that sucks.  Now, we simply check when the
 * packets come back: if the hook is gone, the packet is discarded. */
struct list_head nf_hooks[NPROTO][NF_MAX_HOOKS] __read_mostly;
EXPORT_SYMBOL(nf_hooks);
static DEFINE_MUTEX(nf_hook_mutex);

int nf_register_hook(struct nf_hook_ops *reg)
{
	struct list_head *i;
	int err;

	err = mutex_lock_interruptible(&nf_hook_mutex);
	if (err < 0)
		return err;
	list_for_each(i, &nf_hooks[reg->pf][reg->hooknum]) {
		if (reg->priority < ((struct nf_hook_ops *)i)->priority)
			break;
	}
	list_add_rcu(&reg->list, i->prev);
	mutex_unlock(&nf_hook_mutex);
	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);

	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)
{
	unsigned int i;

	for (i = 0; i < n; i++)
		nf_unregister_hook(&reg[i]);
}
EXPORT_SYMBOL(nf_unregister_hooks);

unsigned int nf_iterate(struct list_head *head,
			struct sk_buff **skb,
			int hook,
			const struct net_device *indev,
			const struct net_device *outdev,
			struct list_head **i,
			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_continue_rcu(*i, head) {
		struct nf_hook_ops *elem = (struct nf_hook_ops *)*i;

		if (hook_thresh > elem->priority)
			continue;

		/* Optimization: we don't need to hold module
		   reference here, since function can't sleep. --RR */
		verdict = elem->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",
					elem->hook, hook);
				continue;
			}
#endif
			if (verdict != NF_REPEAT)
				return verdict;
			*i = (*i)->prev;
		}
	}
	return NF_ACCEPT;
}


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

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

	elem = &nf_hooks[pf][hook];
next_hook:
	verdict = nf_iterate(&nf_hooks[pf][hook], pskb, hook, indev,
			     outdev, &elem, okfn, hook_thresh);
	if (verdict == NF_ACCEPT || verdict == NF_STOP) {
		ret = 1;
		goto unlock;
	} else if (verdict == NF_DROP) {
		kfree_skb(*pskb);
		ret = -EPERM;
	} else if ((verdict & NF_VERDICT_MASK)  == NF_QUEUE) {
		NFDEBUG("nf_hook: Verdict = QUEUE.\n");
		if (!nf_queue(*pskb, elem, pf, hook, indev, outdev, okfn,
			      verdict >> NF_VERDICT_BITS))
			goto next_hook;
	}
unlock:
	rcu_read_unlock();
	return ret;
}
EXPORT_SYMBOL(nf_hook_slow);


int skb_make_writable(struct sk_buff **pskb, unsigned int writable_len)
{
	struct sk_buff *nskb;

	if (writable_len > (*pskb)->len)
		return 0;

	/* Not exclusive use of packet?  Must copy. */
	if (skb_shared(*pskb) || skb_cloned(*pskb))
		goto copy_skb;

	return pskb_may_pull(*pskb, writable_len);

copy_skb:
	nskb = skb_copy(*pskb, GFP_ATOMIC);
	if (!nskb)
		return 0;
	BUG_ON(skb_is_nonlinear(nskb));

	/* Rest of kernel will get very unhappy if we pass it a
	   suddenly-orphaned skbuff */
	if ((*pskb)->sk)
		skb_set_owner_w(nskb, (*pskb)->sk);
	kfree_skb(*pskb);
	*pskb = nskb;
	return 1;
}
EXPORT_SYMBOL(skb_make_writable);

void nf_proto_csum_replace4(__sum16 *sum, struct sk_buff *skb,
			    __be32 from, __be32 to, int pseudohdr)
{
	__be32 diff[] = { ~from, to };
	if (skb->ip_summed != CHECKSUM_PARTIAL) {
		*sum = csum_fold(csum_partial((char *)diff, sizeof(diff),
				~csum_unfold(*sum)));
		if (skb->ip_summed == CHECKSUM_COMPLETE && pseudohdr)
			skb->csum = ~csum_partial((char *)diff, sizeof(diff),
						~skb->csum);
	} else if (pseudohdr)
		*sum = ~csum_fold(csum_partial((char *)diff, sizeof(diff),
				csum_unfold(*sum)));
}
EXPORT_SYMBOL(nf_proto_csum_replace4);

#if defined(CONFIG_NF_CONNTRACK) || defined(CONFIG_NF_CONNTRACK_MODULE)
/* 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 *);
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 *);
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);
#endif /* CONFIG_NF_CONNTRACK */

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

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

#ifdef CONFIG_PROC_FS
	proc_net_netfilter = proc_mkdir("netfilter", proc_net);
	if (!proc_net_netfilter)
		panic("cannot create netfilter proc entry");
#endif

	if (netfilter_queue_init() < 0)
		panic("cannot initialize nf_queue");
	if (netfilter_log_init() < 0)
		panic("cannot initialize nf_log");
}
Commit	Line	Data
601e68e1	1	/* netfilter.c: look after the filters for various protocols.
f6ebe77f HW	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.
f6ebe77f	8	*/
f6ebe77f HW	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>
d486dd1f	21	#include <linux/mutex.h>
f6ebe77f HW	22	#include <net/sock.h>
	23
	24	#include "nf_internals.h"
	25
d486dd1f	26	static DEFINE_MUTEX(afinfo_mutex);
bce8032e	27
e2b7606c	28	struct nf_afinfo *nf_afinfo[NPROTO] __read_mostly;
bce8032e PM	29	EXPORT_SYMBOL(nf_afinfo);
	30
	31	int nf_register_afinfo(struct nf_afinfo *afinfo)
	32	{
d486dd1f PM	33	int err;
	34
	35	err = mutex_lock_interruptible(&afinfo_mutex);
	36	if (err < 0)
	37	return err;
bce8032e	38	rcu_assign_pointer(nf_afinfo[afinfo->family], afinfo);
d486dd1f	39	mutex_unlock(&afinfo_mutex);
bce8032e PM	40	return 0;
	41	}
	42	EXPORT_SYMBOL_GPL(nf_register_afinfo);
	43
	44	void nf_unregister_afinfo(struct nf_afinfo *afinfo)
	45	{
d486dd1f	46	mutex_lock(&afinfo_mutex);
bce8032e	47	rcu_assign_pointer(nf_afinfo[afinfo->family], NULL);
d486dd1f	48	mutex_unlock(&afinfo_mutex);
bce8032e PM	49	synchronize_rcu();
	50	}
	51	EXPORT_SYMBOL_GPL(nf_unregister_afinfo);
	52
f6ebe77f HW	53	/* In this code, we can be waiting indefinitely for userspace to
	54	* service a packet if a hook returns NF_QUEUE. We could keep a count
	55	* of skbuffs queued for userspace, and not deregister a hook unless
	56	* this is zero, but that sucks. Now, we simply check when the
	57	* packets come back: if the hook is gone, the packet is discarded. */
e2b7606c	58	struct list_head nf_hooks[NPROTO][NF_MAX_HOOKS] __read_mostly;
f6ebe77f	59	EXPORT_SYMBOL(nf_hooks);
fd706d69	60	static DEFINE_MUTEX(nf_hook_mutex);
f6ebe77f HW	61
	62	int nf_register_hook(struct nf_hook_ops *reg)
	63	{
	64	struct list_head *i;
fd706d69	65	int err;
f6ebe77f	66
fd706d69 PM	67	err = mutex_lock_interruptible(&nf_hook_mutex);
	68	if (err < 0)
	69	return err;
f6ebe77f HW	70	list_for_each(i, &nf_hooks[reg->pf][reg->hooknum]) {
	71	if (reg->priority < ((struct nf_hook_ops *)i)->priority)
	72	break;
	73	}
	74	list_add_rcu(&reg->list, i->prev);
fd706d69	75	mutex_unlock(&nf_hook_mutex);
f6ebe77f HW	76	return 0;
	77	}
	78	EXPORT_SYMBOL(nf_register_hook);
	79
	80	void nf_unregister_hook(struct nf_hook_ops *reg)
	81	{
fd706d69	82	mutex_lock(&nf_hook_mutex);
f6ebe77f	83	list_del_rcu(&reg->list);
fd706d69	84	mutex_unlock(&nf_hook_mutex);
f6ebe77f HW	85
	86	synchronize_net();
	87	}
	88	EXPORT_SYMBOL(nf_unregister_hook);
	89
972d1cb1 PM	90	int nf_register_hooks(struct nf_hook_ops *reg, unsigned int n)
	91	{
	92	unsigned int i;
	93	int err = 0;
	94
	95	for (i = 0; i < n; i++) {
	96	err = nf_register_hook(&reg[i]);
	97	if (err)
	98	goto err;
	99	}
	100	return err;
	101
	102	err:
	103	if (i > 0)
	104	nf_unregister_hooks(reg, i);
	105	return err;
	106	}
	107	EXPORT_SYMBOL(nf_register_hooks);
	108
	109	void nf_unregister_hooks(struct nf_hook_ops *reg, unsigned int n)
	110	{
	111	unsigned int i;
	112
	113	for (i = 0; i < n; i++)
	114	nf_unregister_hook(&reg[i]);
	115	}
	116	EXPORT_SYMBOL(nf_unregister_hooks);
	117
f6ebe77f HW	118	unsigned int nf_iterate(struct list_head *head,
	119	struct sk_buff **skb,
	120	int hook,
	121	const struct net_device *indev,
	122	const struct net_device *outdev,
	123	struct list_head **i,
	124	int (okfn)(struct sk_buff ),
	125	int hook_thresh)
	126	{
	127	unsigned int verdict;
	128
	129	/*
	130	* The caller must not block between calls to this
	131	* function because of risk of continuing from deleted element.
	132	*/
	133	list_for_each_continue_rcu(*i, head) {
	134	struct nf_hook_ops elem = (struct nf_hook_ops )*i;
	135
	136	if (hook_thresh > elem->priority)
	137	continue;
	138
	139	/* Optimization: we don't need to hold module
601e68e1	140	reference here, since function can't sleep. --RR */
f6ebe77f HW	141	verdict = elem->hook(hook, skb, indev, outdev, okfn);
	142	if (verdict != NF_ACCEPT) {
	143	#ifdef CONFIG_NETFILTER_DEBUG
	144	if (unlikely((verdict & NF_VERDICT_MASK)
	145	> NF_MAX_VERDICT)) {
	146	NFDEBUG("Evil return from %p(%u).\n",
601e68e1	147	elem->hook, hook);
f6ebe77f HW	148	continue;
	149	}
	150	#endif
	151	if (verdict != NF_REPEAT)
	152	return verdict;
	153	i = (i)->prev;
	154	}
	155	}
	156	return NF_ACCEPT;
	157	}
	158
	159
	160	/* Returns 1 if okfn() needs to be executed by the caller,
	161	* -EPERM for NF_DROP, 0 otherwise. */
	162	int nf_hook_slow(int pf, unsigned int hook, struct sk_buff **pskb,
	163	struct net_device *indev,
	164	struct net_device *outdev,
	165	int (okfn)(struct sk_buff ),
	166	int hook_thresh)
	167	{
	168	struct list_head *elem;
	169	unsigned int verdict;
	170	int ret = 0;
	171
	172	/* We may already have this, but read-locks nest anyway */
	173	rcu_read_lock();
	174
	175	elem = &nf_hooks[pf][hook];
	176	next_hook:
	177	verdict = nf_iterate(&nf_hooks[pf][hook], pskb, hook, indev,
	178	outdev, &elem, okfn, hook_thresh);
	179	if (verdict == NF_ACCEPT \|\| verdict == NF_STOP) {
	180	ret = 1;
	181	goto unlock;
	182	} else if (verdict == NF_DROP) {
	183	kfree_skb(*pskb);
	184	ret = -EPERM;
	185	} else if ((verdict & NF_VERDICT_MASK) == NF_QUEUE) {
	186	NFDEBUG("nf_hook: Verdict = QUEUE.\n");
394f545d	187	if (!nf_queue(*pskb, elem, pf, hook, indev, outdev, okfn,
f6ebe77f HW	188	verdict >> NF_VERDICT_BITS))
	189	goto next_hook;
	190	}
	191	unlock:
	192	rcu_read_unlock();
	193	return ret;
	194	}
	195	EXPORT_SYMBOL(nf_hook_slow);
	196
	197
	198	int skb_make_writable(struct sk_buff **pskb, unsigned int writable_len)
	199	{
	200	struct sk_buff *nskb;
	201
	202	if (writable_len > (*pskb)->len)
	203	return 0;
	204
	205	/* Not exclusive use of packet? Must copy. */
	206	if (skb_shared(pskb) \|\| skb_cloned(pskb))
	207	goto copy_skb;
	208
	209	return pskb_may_pull(*pskb, writable_len);
	210
	211	copy_skb:
	212	nskb = skb_copy(*pskb, GFP_ATOMIC);
	213	if (!nskb)
	214	return 0;
	215	BUG_ON(skb_is_nonlinear(nskb));
	216
	217	/* Rest of kernel will get very unhappy if we pass it a
	218	suddenly-orphaned skbuff */
	219	if ((*pskb)->sk)
	220	skb_set_owner_w(nskb, (*pskb)->sk);
	221	kfree_skb(*pskb);
	222	*pskb = nskb;
	223	return 1;
	224	}
	225	EXPORT_SYMBOL(skb_make_writable);
	226
43bc0ca7 AV	227	void nf_proto_csum_replace4(__sum16 sum, struct sk_buff skb,
43bc0ca7 AV	228	__be32 from, __be32 to, int pseudohdr)
4cf411de	229	{
43bc0ca7	230	__be32 diff[] = { ~from, to };
4cf411de	231	if (skb->ip_summed != CHECKSUM_PARTIAL) {
43bc0ca7 AV	232	sum = csum_fold(csum_partial((char )diff, sizeof(diff),
43bc0ca7 AV	233	~csum_unfold(*sum)));
4cf411de	234	if (skb->ip_summed == CHECKSUM_COMPLETE && pseudohdr)
43bc0ca7 AV	235	skb->csum = ~csum_partial((char *)diff, sizeof(diff),
43bc0ca7 AV	236	~skb->csum);
4cf411de	237	} else if (pseudohdr)
43bc0ca7 AV	238	sum = ~csum_fold(csum_partial((char )diff, sizeof(diff),
43bc0ca7 AV	239	csum_unfold(*sum)));
4cf411de	240	}
43bc0ca7	241	EXPORT_SYMBOL(nf_proto_csum_replace4);
f6ebe77f	242
5f79e0f9	243	#if defined(CONFIG_NF_CONNTRACK) \|\| defined(CONFIG_NF_CONNTRACK_MODULE)
f6ebe77f HW	244	/* This does not belong here, but locally generated errors need it if connection
	245	tracking in use: without this, connection may not be in hash table, and hence
	246	manufactured ICMP or RST packets will not be associated with it. */
	247	void (ip_ct_attach)(struct sk_buff , struct sk_buff *);
	248	EXPORT_SYMBOL(ip_ct_attach);
	249
	250	void nf_ct_attach(struct sk_buff new, struct sk_buff skb)
	251	{
	252	void (attach)(struct sk_buff , struct sk_buff *);
	253
c3a47ab3 PM	254	if (skb->nfct) {
	255	rcu_read_lock();
	256	attach = rcu_dereference(ip_ct_attach);
	257	if (attach)
	258	attach(new, skb);
	259	rcu_read_unlock();
f6ebe77f HW	260	}
	261	}
	262	EXPORT_SYMBOL(nf_ct_attach);
de6e05c4 YK	263
	264	void (nf_ct_destroy)(struct nf_conntrack );
	265	EXPORT_SYMBOL(nf_ct_destroy);
	266
	267	void nf_conntrack_destroy(struct nf_conntrack *nfct)
	268	{
	269	void (destroy)(struct nf_conntrack );
	270
	271	rcu_read_lock();
	272	destroy = rcu_dereference(nf_ct_destroy);
	273	BUG_ON(destroy == NULL);
	274	destroy(nfct);
	275	rcu_read_unlock();
	276	}
	277	EXPORT_SYMBOL(nf_conntrack_destroy);
	278	#endif /* CONFIG_NF_CONNTRACK */
f6ebe77f HW	279
	280	#ifdef CONFIG_PROC_FS
	281	struct proc_dir_entry *proc_net_netfilter;
	282	EXPORT_SYMBOL(proc_net_netfilter);
	283	#endif
	284
	285	void __init netfilter_init(void)
	286	{
	287	int i, h;
	288	for (i = 0; i < NPROTO; i++) {
	289	for (h = 0; h < NF_MAX_HOOKS; h++)
	290	INIT_LIST_HEAD(&nf_hooks[i][h]);
	291	}
	292
	293	#ifdef CONFIG_PROC_FS
	294	proc_net_netfilter = proc_mkdir("netfilter", proc_net);
	295	if (!proc_net_netfilter)
	296	panic("cannot create netfilter proc entry");
	297	#endif
	298
	299	if (netfilter_queue_init() < 0)
	300	panic("cannot initialize nf_queue");
	301	if (netfilter_log_init() < 0)
	302	panic("cannot initialize nf_log");
	303	}