nf_conntrack_core.c

linux 内核源代码

			  int do_acct)
{
	int event = 0;

	NF_CT_ASSERT(ct->timeout.data == (unsigned long)ct);
	NF_CT_ASSERT(skb);

	write_lock_bh(&nf_conntrack_lock);

	/* Only update if this is not a fixed timeout */
	if (test_bit(IPS_FIXED_TIMEOUT_BIT, &ct->status)) {
		write_unlock_bh(&nf_conntrack_lock);
		return;
	}

	/* If not in hash table, timer will not be active yet */
	if (!nf_ct_is_confirmed(ct)) {
		ct->timeout.expires = extra_jiffies;
		event = IPCT_REFRESH;
	} else {
		unsigned long newtime = jiffies + extra_jiffies;

		/* Only update the timeout if the new timeout is at least
		   HZ jiffies from the old timeout. Need del_timer for race
		   avoidance (may already be dying). */
		if (newtime - ct->timeout.expires >= HZ
		    && del_timer(&ct->timeout)) {
			ct->timeout.expires = newtime;
			add_timer(&ct->timeout);
			event = IPCT_REFRESH;
		}
	}

#ifdef CONFIG_NF_CT_ACCT
	if (do_acct) {
		ct->counters[CTINFO2DIR(ctinfo)].packets++;
		ct->counters[CTINFO2DIR(ctinfo)].bytes +=
			skb->len - skb_network_offset(skb);

		if ((ct->counters[CTINFO2DIR(ctinfo)].packets & 0x80000000)
		    || (ct->counters[CTINFO2DIR(ctinfo)].bytes & 0x80000000))
			event |= IPCT_COUNTER_FILLING;
	}
#endif

	write_unlock_bh(&nf_conntrack_lock);

	/* must be unlocked when calling event cache */
	if (event)
		nf_conntrack_event_cache(event, skb);
}
EXPORT_SYMBOL_GPL(__nf_ct_refresh_acct);

#if defined(CONFIG_NF_CT_NETLINK) || defined(CONFIG_NF_CT_NETLINK_MODULE)

#include <linux/netfilter/nfnetlink.h>
#include <linux/netfilter/nfnetlink_conntrack.h>
#include <linux/mutex.h>

/* Generic function for tcp/udp/sctp/dccp and alike. This needs to be
 * in ip_conntrack_core, since we don't want the protocols to autoload
 * or depend on ctnetlink */
int nf_ct_port_tuple_to_nlattr(struct sk_buff *skb,
			       const struct nf_conntrack_tuple *tuple)
{
	NLA_PUT(skb, CTA_PROTO_SRC_PORT, sizeof(u_int16_t),
		&tuple->src.u.tcp.port);
	NLA_PUT(skb, CTA_PROTO_DST_PORT, sizeof(u_int16_t),
		&tuple->dst.u.tcp.port);
	return 0;

nla_put_failure:
	return -1;
}
EXPORT_SYMBOL_GPL(nf_ct_port_tuple_to_nlattr);

const struct nla_policy nf_ct_port_nla_policy[CTA_PROTO_MAX+1] = {
	[CTA_PROTO_SRC_PORT]  = { .type = NLA_U16 },
	[CTA_PROTO_DST_PORT]  = { .type = NLA_U16 },
};
EXPORT_SYMBOL_GPL(nf_ct_port_nla_policy);

int nf_ct_port_nlattr_to_tuple(struct nlattr *tb[],
			       struct nf_conntrack_tuple *t)
{
	if (!tb[CTA_PROTO_SRC_PORT] || !tb[CTA_PROTO_DST_PORT])
		return -EINVAL;

	t->src.u.tcp.port = *(__be16 *)nla_data(tb[CTA_PROTO_SRC_PORT]);
	t->dst.u.tcp.port = *(__be16 *)nla_data(tb[CTA_PROTO_DST_PORT]);

	return 0;
}
EXPORT_SYMBOL_GPL(nf_ct_port_nlattr_to_tuple);
#endif

/* Used by ipt_REJECT and ip6t_REJECT. */
void __nf_conntrack_attach(struct sk_buff *nskb, struct sk_buff *skb)
{
	struct nf_conn *ct;
	enum ip_conntrack_info ctinfo;

	/* This ICMP is in reverse direction to the packet which caused it */
	ct = nf_ct_get(skb, &ctinfo);
	if (CTINFO2DIR(ctinfo) == IP_CT_DIR_ORIGINAL)
		ctinfo = IP_CT_RELATED + IP_CT_IS_REPLY;
	else
		ctinfo = IP_CT_RELATED;

	/* Attach to new skbuff, and increment count */
	nskb->nfct = &ct->ct_general;
	nskb->nfctinfo = ctinfo;
	nf_conntrack_get(nskb->nfct);
}
EXPORT_SYMBOL_GPL(__nf_conntrack_attach);

static inline int
do_iter(const struct nf_conntrack_tuple_hash *i,
	int (*iter)(struct nf_conn *i, void *data),
	void *data)
{
	return iter(nf_ct_tuplehash_to_ctrack(i), data);
}

/* Bring out ya dead! */
static struct nf_conn *
get_next_corpse(int (*iter)(struct nf_conn *i, void *data),
		void *data, unsigned int *bucket)
{
	struct nf_conntrack_tuple_hash *h;
	struct nf_conn *ct;
	struct hlist_node *n;

	write_lock_bh(&nf_conntrack_lock);
	for (; *bucket < nf_conntrack_htable_size; (*bucket)++) {
		hlist_for_each_entry(h, n, &nf_conntrack_hash[*bucket], hnode) {
			ct = nf_ct_tuplehash_to_ctrack(h);
			if (iter(ct, data))
				goto found;
		}
	}
	hlist_for_each_entry(h, n, &unconfirmed, hnode) {
		ct = nf_ct_tuplehash_to_ctrack(h);
		if (iter(ct, data))
			set_bit(IPS_DYING_BIT, &ct->status);
	}
	write_unlock_bh(&nf_conntrack_lock);
	return NULL;
found:
	atomic_inc(&ct->ct_general.use);
	write_unlock_bh(&nf_conntrack_lock);
	return ct;
}

void
nf_ct_iterate_cleanup(int (*iter)(struct nf_conn *i, void *data), void *data)
{
	struct nf_conn *ct;
	unsigned int bucket = 0;

	while ((ct = get_next_corpse(iter, data, &bucket)) != NULL) {
		/* Time to push up daises... */
		if (del_timer(&ct->timeout))
			death_by_timeout((unsigned long)ct);
		/* ... else the timer will get him soon. */

		nf_ct_put(ct);
	}
}
EXPORT_SYMBOL_GPL(nf_ct_iterate_cleanup);

static int kill_all(struct nf_conn *i, void *data)
{
	return 1;
}

void nf_ct_free_hashtable(struct hlist_head *hash, int vmalloced, int size)
{
	if (vmalloced)
		vfree(hash);
	else
		free_pages((unsigned long)hash,
			   get_order(sizeof(struct hlist_head) * size));
}
EXPORT_SYMBOL_GPL(nf_ct_free_hashtable);

void nf_conntrack_flush(void)
{
	nf_ct_iterate_cleanup(kill_all, NULL);
}
EXPORT_SYMBOL_GPL(nf_conntrack_flush);

/* Mishearing the voices in his head, our hero wonders how he's
   supposed to kill the mall. */
void nf_conntrack_cleanup(void)
{
	rcu_assign_pointer(ip_ct_attach, NULL);

	/* This makes sure all current packets have passed through
	   netfilter framework.  Roll on, two-stage module
	   delete... */
	synchronize_net();

	nf_ct_event_cache_flush();
 i_see_dead_people:
	nf_conntrack_flush();
	if (atomic_read(&nf_conntrack_count) != 0) {
		schedule();
		goto i_see_dead_people;
	}
	/* wait until all references to nf_conntrack_untracked are dropped */
	while (atomic_read(&nf_conntrack_untracked.ct_general.use) > 1)
		schedule();

	rcu_assign_pointer(nf_ct_destroy, NULL);

	kmem_cache_destroy(nf_conntrack_cachep);
	nf_ct_free_hashtable(nf_conntrack_hash, nf_conntrack_vmalloc,
			     nf_conntrack_htable_size);

	nf_conntrack_proto_fini();
	nf_conntrack_helper_fini();
	nf_conntrack_expect_fini();
}

struct hlist_head *nf_ct_alloc_hashtable(int *sizep, int *vmalloced)
{
	struct hlist_head *hash;
	unsigned int size, i;

	*vmalloced = 0;

	size = *sizep = roundup(*sizep, PAGE_SIZE / sizeof(struct hlist_head));
	hash = (void*)__get_free_pages(GFP_KERNEL|__GFP_NOWARN,
				       get_order(sizeof(struct hlist_head)
						 * size));
	if (!hash) {
		*vmalloced = 1;
		printk(KERN_WARNING "nf_conntrack: falling back to vmalloc.\n");
		hash = vmalloc(sizeof(struct hlist_head) * size);
	}

	if (hash)
		for (i = 0; i < size; i++)
			INIT_HLIST_HEAD(&hash[i]);

	return hash;
}
EXPORT_SYMBOL_GPL(nf_ct_alloc_hashtable);

int nf_conntrack_set_hashsize(const char *val, struct kernel_param *kp)
{
	int i, bucket, hashsize, vmalloced;
	int old_vmalloced, old_size;
	int rnd;
	struct hlist_head *hash, *old_hash;
	struct nf_conntrack_tuple_hash *h;

	/* On boot, we can set this without any fancy locking. */
	if (!nf_conntrack_htable_size)
		return param_set_uint(val, kp);

	hashsize = simple_strtol(val, NULL, 0);
	if (!hashsize)
		return -EINVAL;

	hash = nf_ct_alloc_hashtable(&hashsize, &vmalloced);
	if (!hash)
		return -ENOMEM;

	/* We have to rehahs for the new table anyway, so we also can
	 * use a newrandom seed */
	get_random_bytes(&rnd, 4);

	write_lock_bh(&nf_conntrack_lock);
	for (i = 0; i < nf_conntrack_htable_size; i++) {
		while (!hlist_empty(&nf_conntrack_hash[i])) {
			h = hlist_entry(nf_conntrack_hash[i].first,
					struct nf_conntrack_tuple_hash, hnode);
			hlist_del(&h->hnode);
			bucket = __hash_conntrack(&h->tuple, hashsize, rnd);
			hlist_add_head(&h->hnode, &hash[bucket]);
		}
	}
	old_size = nf_conntrack_htable_size;
	old_vmalloced = nf_conntrack_vmalloc;
	old_hash = nf_conntrack_hash;

	nf_conntrack_htable_size = hashsize;
	nf_conntrack_vmalloc = vmalloced;
	nf_conntrack_hash = hash;
	nf_conntrack_hash_rnd = rnd;
	write_unlock_bh(&nf_conntrack_lock);

	nf_ct_free_hashtable(old_hash, old_vmalloced, old_size);
	return 0;
}
EXPORT_SYMBOL_GPL(nf_conntrack_set_hashsize);

module_param_call(hashsize, nf_conntrack_set_hashsize, param_get_uint,
		  &nf_conntrack_htable_size, 0600);

int __init nf_conntrack_init(void)
{
	int max_factor = 8;
	int ret;

	/* Idea from tcp.c: use 1/16384 of memory.  On i386: 32MB
	 * machine has 512 buckets. >= 1GB machines have 16384 buckets. */
	if (!nf_conntrack_htable_size) {
		nf_conntrack_htable_size
			= (((num_physpages << PAGE_SHIFT) / 16384)
			   / sizeof(struct hlist_head));
		if (num_physpages > (1024 * 1024 * 1024 / PAGE_SIZE))
			nf_conntrack_htable_size = 16384;
		if (nf_conntrack_htable_size < 32)
			nf_conntrack_htable_size = 32;

		/* Use a max. factor of four by default to get the same max as
		 * with the old struct list_heads. When a table size is given
		 * we use the old value of 8 to avoid reducing the max.
		 * entries. */
		max_factor = 4;
	}
	nf_conntrack_hash = nf_ct_alloc_hashtable(&nf_conntrack_htable_size,
						  &nf_conntrack_vmalloc);
	if (!nf_conntrack_hash) {
		printk(KERN_ERR "Unable to create nf_conntrack_hash\n");
		goto err_out;
	}

	nf_conntrack_max = max_factor * nf_conntrack_htable_size;

	printk("nf_conntrack version %s (%u buckets, %d max)\n",
	       NF_CONNTRACK_VERSION, nf_conntrack_htable_size,
	       nf_conntrack_max);

	nf_conntrack_cachep = kmem_cache_create("nf_conntrack",
						sizeof(struct nf_conn),
						0, 0, NULL);
	if (!nf_conntrack_cachep) {
		printk(KERN_ERR "Unable to create nf_conn slab cache\n");
		goto err_free_hash;
	}

	ret = nf_conntrack_proto_init();
	if (ret < 0)
		goto err_free_conntrack_slab;

	ret = nf_conntrack_expect_init();
	if (ret < 0)
		goto out_fini_proto;

	ret = nf_conntrack_helper_init();
	if (ret < 0)
		goto out_fini_expect;

	/* For use by REJECT target */
	rcu_assign_pointer(ip_ct_attach, __nf_conntrack_attach);
	rcu_assign_pointer(nf_ct_destroy, destroy_conntrack);

	/* Set up fake conntrack:
	    - to never be deleted, not in any hashes */
	atomic_set(&nf_conntrack_untracked.ct_general.use, 1);
	/*  - and look it like as a confirmed connection */
	set_bit(IPS_CONFIRMED_BIT, &nf_conntrack_untracked.status);

	return ret;

out_fini_expect:
	nf_conntrack_expect_fini();
out_fini_proto:
	nf_conntrack_proto_fini();
err_free_conntrack_slab:
	kmem_cache_destroy(nf_conntrack_cachep);
err_free_hash:
	nf_ct_free_hashtable(nf_conntrack_hash, nf_conntrack_vmalloc,
			     nf_conntrack_htable_size);
err_out:
	return -ENOMEM;
}