#define NF_CONNTRACK_VERSION "0.5.0"
-#if 0
-#define DEBUGP printk
-#else
-#define DEBUGP(format, args...)
-#endif
-
DEFINE_RWLOCK(nf_conntrack_lock);
EXPORT_SYMBOL_GPL(nf_conntrack_lock);
int nf_conntrack_max __read_mostly;
EXPORT_SYMBOL_GPL(nf_conntrack_max);
-struct list_head *nf_conntrack_hash __read_mostly;
+struct hlist_head *nf_conntrack_hash __read_mostly;
EXPORT_SYMBOL_GPL(nf_conntrack_hash);
struct nf_conn nf_conntrack_untracked __read_mostly;
EXPORT_SYMBOL_GPL(nf_conntrack_untracked);
unsigned int nf_ct_log_invalid __read_mostly;
-LIST_HEAD(unconfirmed);
+HLIST_HEAD(unconfirmed);
static int nf_conntrack_vmalloc __read_mostly;
static struct kmem_cache *nf_conntrack_cachep __read_mostly;
static unsigned int nf_conntrack_next_id;
}
EXPORT_SYMBOL_GPL(nf_ct_get_tuple);
+int nf_ct_get_tuplepr(const struct sk_buff *skb,
+ unsigned int nhoff,
+ u_int16_t l3num,
+ struct nf_conntrack_tuple *tuple)
+{
+ struct nf_conntrack_l3proto *l3proto;
+ struct nf_conntrack_l4proto *l4proto;
+ unsigned int protoff;
+ u_int8_t protonum;
+ int ret;
+
+ rcu_read_lock();
+
+ l3proto = __nf_ct_l3proto_find(l3num);
+ ret = l3proto->get_l4proto(skb, nhoff, &protoff, &protonum);
+ if (ret != NF_ACCEPT) {
+ rcu_read_unlock();
+ return 0;
+ }
+
+ l4proto = __nf_ct_l4proto_find(l3num, protonum);
+
+ ret = nf_ct_get_tuple(skb, nhoff, protoff, l3num, protonum, tuple,
+ l3proto, l4proto);
+
+ rcu_read_unlock();
+ return ret;
+}
+EXPORT_SYMBOL_GPL(nf_ct_get_tuplepr);
+
int
nf_ct_invert_tuple(struct nf_conntrack_tuple *inverse,
const struct nf_conntrack_tuple *orig,
static void
clean_from_lists(struct nf_conn *ct)
{
- DEBUGP("clean_from_lists(%p)\n", ct);
- list_del(&ct->tuplehash[IP_CT_DIR_ORIGINAL].list);
- list_del(&ct->tuplehash[IP_CT_DIR_REPLY].list);
+ pr_debug("clean_from_lists(%p)\n", ct);
+ hlist_del(&ct->tuplehash[IP_CT_DIR_ORIGINAL].hnode);
+ hlist_del(&ct->tuplehash[IP_CT_DIR_REPLY].hnode);
/* Destroy all pending expectations */
nf_ct_remove_expectations(ct);
struct nf_conn *ct = (struct nf_conn *)nfct;
struct nf_conntrack_l4proto *l4proto;
- DEBUGP("destroy_conntrack(%p)\n", ct);
+ pr_debug("destroy_conntrack(%p)\n", ct);
NF_CT_ASSERT(atomic_read(&nfct->use) == 0);
NF_CT_ASSERT(!timer_pending(&ct->timeout));
/* We overload first tuple to link into unconfirmed list. */
if (!nf_ct_is_confirmed(ct)) {
- BUG_ON(list_empty(&ct->tuplehash[IP_CT_DIR_ORIGINAL].list));
- list_del(&ct->tuplehash[IP_CT_DIR_ORIGINAL].list);
+ BUG_ON(hlist_unhashed(&ct->tuplehash[IP_CT_DIR_ORIGINAL].hnode));
+ hlist_del(&ct->tuplehash[IP_CT_DIR_ORIGINAL].hnode);
}
NF_CT_STAT_INC(delete);
if (ct->master)
nf_ct_put(ct->master);
- DEBUGP("destroy_conntrack: returning ct=%p to slab\n", ct);
+ pr_debug("destroy_conntrack: returning ct=%p to slab\n", ct);
nf_conntrack_free(ct);
}
const struct nf_conn *ignored_conntrack)
{
struct nf_conntrack_tuple_hash *h;
+ struct hlist_node *n;
unsigned int hash = hash_conntrack(tuple);
- list_for_each_entry(h, &nf_conntrack_hash[hash], list) {
+ hlist_for_each_entry(h, n, &nf_conntrack_hash[hash], hnode) {
if (nf_ct_tuplehash_to_ctrack(h) != ignored_conntrack &&
nf_ct_tuple_equal(tuple, &h->tuple)) {
NF_CT_STAT_INC(found);
/* Find a connection corresponding to a tuple. */
struct nf_conntrack_tuple_hash *
-nf_conntrack_find_get(const struct nf_conntrack_tuple *tuple,
- const struct nf_conn *ignored_conntrack)
+nf_conntrack_find_get(const struct nf_conntrack_tuple *tuple)
{
struct nf_conntrack_tuple_hash *h;
read_lock_bh(&nf_conntrack_lock);
- h = __nf_conntrack_find(tuple, ignored_conntrack);
+ h = __nf_conntrack_find(tuple, NULL);
if (h)
atomic_inc(&nf_ct_tuplehash_to_ctrack(h)->ct_general.use);
read_unlock_bh(&nf_conntrack_lock);
unsigned int repl_hash)
{
ct->id = ++nf_conntrack_next_id;
- list_add(&ct->tuplehash[IP_CT_DIR_ORIGINAL].list,
- &nf_conntrack_hash[hash]);
- list_add(&ct->tuplehash[IP_CT_DIR_REPLY].list,
- &nf_conntrack_hash[repl_hash]);
+ hlist_add_head(&ct->tuplehash[IP_CT_DIR_ORIGINAL].hnode,
+ &nf_conntrack_hash[hash]);
+ hlist_add_head(&ct->tuplehash[IP_CT_DIR_REPLY].hnode,
+ &nf_conntrack_hash[repl_hash]);
}
void nf_conntrack_hash_insert(struct nf_conn *ct)
struct nf_conntrack_tuple_hash *h;
struct nf_conn *ct;
struct nf_conn_help *help;
+ struct hlist_node *n;
enum ip_conntrack_info ctinfo;
ct = nf_ct_get(*pskb, &ctinfo);
/* No external references means noone else could have
confirmed us. */
NF_CT_ASSERT(!nf_ct_is_confirmed(ct));
- DEBUGP("Confirming conntrack %p\n", ct);
+ pr_debug("Confirming conntrack %p\n", ct);
write_lock_bh(&nf_conntrack_lock);
/* See if there's one in the list already, including reverse:
NAT could have grabbed it without realizing, since we're
not in the hash. If there is, we lost race. */
- list_for_each_entry(h, &nf_conntrack_hash[hash], list)
+ hlist_for_each_entry(h, n, &nf_conntrack_hash[hash], hnode)
if (nf_ct_tuple_equal(&ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple,
&h->tuple))
goto out;
- list_for_each_entry(h, &nf_conntrack_hash[repl_hash], list)
+ hlist_for_each_entry(h, n, &nf_conntrack_hash[repl_hash], hnode)
if (nf_ct_tuple_equal(&ct->tuplehash[IP_CT_DIR_REPLY].tuple,
&h->tuple))
goto out;
/* Remove from unconfirmed list */
- list_del(&ct->tuplehash[IP_CT_DIR_ORIGINAL].list);
+ hlist_del(&ct->tuplehash[IP_CT_DIR_ORIGINAL].hnode);
__nf_conntrack_hash_insert(ct, hash, repl_hash);
/* Timer relative to confirmation time, not original
}
EXPORT_SYMBOL_GPL(nf_conntrack_tuple_taken);
+#define NF_CT_EVICTION_RANGE 8
+
/* There's a small race here where we may free a just-assured
connection. Too bad: we're in trouble anyway. */
-static int early_drop(struct list_head *chain)
+static int early_drop(unsigned int hash)
{
- /* Traverse backwards: gives us oldest, which is roughly LRU */
+ /* Use oldest entry, which is roughly LRU */
struct nf_conntrack_tuple_hash *h;
struct nf_conn *ct = NULL, *tmp;
+ struct hlist_node *n;
+ unsigned int i, cnt = 0;
int dropped = 0;
read_lock_bh(&nf_conntrack_lock);
- list_for_each_entry_reverse(h, chain, list) {
- tmp = nf_ct_tuplehash_to_ctrack(h);
- if (!test_bit(IPS_ASSURED_BIT, &tmp->status)) {
- ct = tmp;
- atomic_inc(&ct->ct_general.use);
- break;
+ for (i = 0; i < nf_conntrack_htable_size; i++) {
+ hlist_for_each_entry(h, n, &nf_conntrack_hash[hash], hnode) {
+ tmp = nf_ct_tuplehash_to_ctrack(h);
+ if (!test_bit(IPS_ASSURED_BIT, &tmp->status))
+ ct = tmp;
+ cnt++;
}
+ if (ct || cnt >= NF_CT_EVICTION_RANGE)
+ break;
+ hash = (hash + 1) % nf_conntrack_htable_size;
}
+ if (ct)
+ atomic_inc(&ct->ct_general.use);
read_unlock_bh(&nf_conntrack_lock);
if (!ct)
if (nf_conntrack_max
&& atomic_read(&nf_conntrack_count) > nf_conntrack_max) {
unsigned int hash = hash_conntrack(orig);
- /* Try dropping from this hash chain. */
- if (!early_drop(&nf_conntrack_hash[hash])) {
+ if (!early_drop(hash)) {
atomic_dec(&nf_conntrack_count);
if (net_ratelimit())
printk(KERN_WARNING
conntrack = kmem_cache_zalloc(nf_conntrack_cachep, GFP_ATOMIC);
if (conntrack == NULL) {
- DEBUGP("nf_conntrack_alloc: Can't alloc conntrack.\n");
+ pr_debug("nf_conntrack_alloc: Can't alloc conntrack.\n");
atomic_dec(&nf_conntrack_count);
return ERR_PTR(-ENOMEM);
}
struct nf_conntrack_expect *exp;
if (!nf_ct_invert_tuple(&repl_tuple, tuple, l3proto, l4proto)) {
- DEBUGP("Can't invert tuple.\n");
+ pr_debug("Can't invert tuple.\n");
return NULL;
}
conntrack = nf_conntrack_alloc(tuple, &repl_tuple);
if (conntrack == NULL || IS_ERR(conntrack)) {
- DEBUGP("Can't allocate conntrack.\n");
+ pr_debug("Can't allocate conntrack.\n");
return (struct nf_conntrack_tuple_hash *)conntrack;
}
if (!l4proto->new(conntrack, skb, dataoff)) {
nf_conntrack_free(conntrack);
- DEBUGP("init conntrack: can't track with proto module\n");
+ pr_debug("init conntrack: can't track with proto module\n");
return NULL;
}
write_lock_bh(&nf_conntrack_lock);
- exp = find_expectation(tuple);
+ exp = nf_ct_find_expectation(tuple);
if (exp) {
- DEBUGP("conntrack: expectation arrives ct=%p exp=%p\n",
- conntrack, exp);
+ pr_debug("conntrack: expectation arrives ct=%p exp=%p\n",
+ conntrack, exp);
/* Welcome, Mr. Bond. We've been expecting you... */
__set_bit(IPS_EXPECTED_BIT, &conntrack->status);
conntrack->master = exp->master;
if (exp->helper) {
- help = nf_ct_ext_add(conntrack, NF_CT_EXT_HELPER,
- GFP_ATOMIC);
+ help = nf_ct_helper_ext_add(conntrack, GFP_ATOMIC);
if (help)
rcu_assign_pointer(help->helper, exp->helper);
- else
- DEBUGP("failed to add helper extension area");
}
#ifdef CONFIG_NF_CONNTRACK_MARK
helper = __nf_ct_helper_find(&repl_tuple);
if (helper) {
- help = nf_ct_ext_add(conntrack, NF_CT_EXT_HELPER,
- GFP_ATOMIC);
+ help = nf_ct_helper_ext_add(conntrack, GFP_ATOMIC);
if (help)
- /* not in hash table yet, so not strictly
- necessary */
rcu_assign_pointer(help->helper, helper);
- else
- DEBUGP("failed to add helper extension area");
}
NF_CT_STAT_INC(new);
}
/* Overload tuple linked list to put us in unconfirmed list. */
- list_add(&conntrack->tuplehash[IP_CT_DIR_ORIGINAL].list, &unconfirmed);
+ hlist_add_head(&conntrack->tuplehash[IP_CT_DIR_ORIGINAL].hnode,
+ &unconfirmed);
write_unlock_bh(&nf_conntrack_lock);
if (exp) {
if (exp->expectfn)
exp->expectfn(conntrack, exp);
- nf_conntrack_expect_put(exp);
+ nf_ct_expect_put(exp);
}
return &conntrack->tuplehash[IP_CT_DIR_ORIGINAL];
if (!nf_ct_get_tuple(skb, skb_network_offset(skb),
dataoff, l3num, protonum, &tuple, l3proto,
l4proto)) {
- DEBUGP("resolve_normal_ct: Can't get tuple\n");
+ pr_debug("resolve_normal_ct: Can't get tuple\n");
return NULL;
}
/* look for tuple match */
- h = nf_conntrack_find_get(&tuple, NULL);
+ h = nf_conntrack_find_get(&tuple);
if (!h) {
h = init_conntrack(&tuple, l3proto, l4proto, skb, dataoff);
if (!h)
} else {
/* Once we've had two way comms, always ESTABLISHED. */
if (test_bit(IPS_SEEN_REPLY_BIT, &ct->status)) {
- DEBUGP("nf_conntrack_in: normal packet for %p\n", ct);
+ pr_debug("nf_conntrack_in: normal packet for %p\n", ct);
*ctinfo = IP_CT_ESTABLISHED;
} else if (test_bit(IPS_EXPECTED_BIT, &ct->status)) {
- DEBUGP("nf_conntrack_in: related packet for %p\n", ct);
+ pr_debug("nf_conntrack_in: related packet for %p\n",
+ ct);
*ctinfo = IP_CT_RELATED;
} else {
- DEBUGP("nf_conntrack_in: new packet for %p\n", ct);
+ pr_debug("nf_conntrack_in: new packet for %p\n", ct);
*ctinfo = IP_CT_NEW;
}
*set_reply = 0;
/* rcu_read_lock()ed by nf_hook_slow */
l3proto = __nf_ct_l3proto_find((u_int16_t)pf);
-
- if ((ret = l3proto->prepare(pskb, hooknum, &dataoff, &protonum)) <= 0) {
- DEBUGP("not prepared to track yet or error occured\n");
+ ret = l3proto->get_l4proto(*pskb, skb_network_offset(*pskb),
+ &dataoff, &protonum);
+ if (ret <= 0) {
+ pr_debug("not prepared to track yet or error occured\n");
+ NF_CT_STAT_INC_ATOMIC(error);
+ NF_CT_STAT_INC_ATOMIC(invalid);
return -ret;
}
if (ret < 0) {
/* Invalid: inverse of the return code tells
* the netfilter core what to do */
- DEBUGP("nf_conntrack_in: Can't track with proto module\n");
+ pr_debug("nf_conntrack_in: Can't track with proto module\n");
nf_conntrack_put((*pskb)->nfct);
(*pskb)->nfct = NULL;
NF_CT_STAT_INC_ATOMIC(invalid);
/* Should be unconfirmed, so not in hash table yet */
NF_CT_ASSERT(!nf_ct_is_confirmed(ct));
- DEBUGP("Altering reply tuple of %p to ", ct);
+ pr_debug("Altering reply tuple of %p to ", ct);
NF_CT_DUMP_TUPLE(newreply);
ct->tuplehash[IP_CT_DIR_REPLY].tuple = *newreply;
}
if (help == NULL) {
- help = nf_ct_ext_add(ct, NF_CT_EXT_HELPER, GFP_ATOMIC);
- if (help == NULL) {
- DEBUGP("failed to add helper extension area");
+ help = nf_ct_helper_ext_add(ct, GFP_ATOMIC);
+ if (help == NULL)
goto out;
- }
} else {
memset(&help->help, 0, sizeof(help->help));
}
{
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)++) {
- list_for_each_entry(h, &nf_conntrack_hash[*bucket], list) {
+ hlist_for_each_entry(h, n, &nf_conntrack_hash[*bucket], hnode) {
ct = nf_ct_tuplehash_to_ctrack(h);
if (iter(ct, data))
goto found;
}
}
- list_for_each_entry(h, &unconfirmed, list) {
+ 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);
return 1;
}
-static void free_conntrack_hash(struct list_head *hash, int vmalloced, int size)
+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 list_head) * size));
+ get_order(sizeof(struct hlist_head) * size));
}
+EXPORT_SYMBOL_GPL(nf_ct_free_hashtable);
void nf_conntrack_flush(void)
{
rcu_assign_pointer(nf_ct_destroy, NULL);
kmem_cache_destroy(nf_conntrack_cachep);
- kmem_cache_destroy(nf_conntrack_expect_cachep);
- free_conntrack_hash(nf_conntrack_hash, nf_conntrack_vmalloc,
- nf_conntrack_htable_size);
+ 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();
}
-static struct list_head *alloc_hashtable(int size, int *vmalloced)
+struct hlist_head *nf_ct_alloc_hashtable(int *sizep, int *vmalloced)
{
- struct list_head *hash;
- unsigned int i;
+ 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,
- get_order(sizeof(struct list_head)
+ 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 list_head) * size);
+ hash = vmalloc(sizeof(struct hlist_head) * size);
}
if (hash)
for (i = 0; i < size; i++)
- INIT_LIST_HEAD(&hash[i]);
+ INIT_HLIST_HEAD(&hash[i]);
return hash;
}
+EXPORT_SYMBOL_GPL(nf_ct_alloc_hashtable);
int set_hashsize(const char *val, struct kernel_param *kp)
{
int i, bucket, hashsize, vmalloced;
int old_vmalloced, old_size;
int rnd;
- struct list_head *hash, *old_hash;
+ struct hlist_head *hash, *old_hash;
struct nf_conntrack_tuple_hash *h;
/* On boot, we can set this without any fancy locking. */
if (!hashsize)
return -EINVAL;
- hash = alloc_hashtable(hashsize, &vmalloced);
+ hash = nf_ct_alloc_hashtable(&hashsize, &vmalloced);
if (!hash)
return -ENOMEM;
write_lock_bh(&nf_conntrack_lock);
for (i = 0; i < nf_conntrack_htable_size; i++) {
- while (!list_empty(&nf_conntrack_hash[i])) {
- h = list_entry(nf_conntrack_hash[i].next,
- struct nf_conntrack_tuple_hash, list);
- list_del(&h->list);
+ 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);
- list_add_tail(&h->list, &hash[bucket]);
+ hlist_add_head(&h->hnode, &hash[bucket]);
}
}
old_size = nf_conntrack_htable_size;
nf_conntrack_hash_rnd = rnd;
write_unlock_bh(&nf_conntrack_lock);
- free_conntrack_hash(old_hash, old_vmalloced, old_size);
+ nf_ct_free_hashtable(old_hash, old_vmalloced, old_size);
return 0;
}
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 256 buckets. >= 1GB machines have 8192 buckets. */
+ * 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 list_head));
+ / sizeof(struct hlist_head));
if (num_physpages > (1024 * 1024 * 1024 / PAGE_SIZE))
- nf_conntrack_htable_size = 8192;
- if (nf_conntrack_htable_size < 16)
- nf_conntrack_htable_size = 16;
+ 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_max = 8 * 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_hash = alloc_hashtable(nf_conntrack_htable_size,
- &nf_conntrack_vmalloc);
+ 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, NULL);
+ 0, 0, NULL);
if (!nf_conntrack_cachep) {
printk(KERN_ERR "Unable to create nf_conn slab cache\n");
goto err_free_hash;
}
- nf_conntrack_expect_cachep = kmem_cache_create("nf_conntrack_expect",
- sizeof(struct nf_conntrack_expect),
- 0, 0, NULL, NULL);
- if (!nf_conntrack_expect_cachep) {
- printk(KERN_ERR "Unable to create nf_expect slab cache\n");
+ ret = nf_conntrack_proto_init();
+ if (ret < 0)
goto err_free_conntrack_slab;
- }
- ret = nf_conntrack_proto_init();
+ ret = nf_conntrack_expect_init();
if (ret < 0)
- goto out_free_expect_slab;
+ goto out_fini_proto;
ret = nf_conntrack_helper_init();
if (ret < 0)
- goto out_fini_proto;
+ goto out_fini_expect;
/* For use by REJECT target */
rcu_assign_pointer(ip_ct_attach, __nf_conntrack_attach);
return ret;
+out_fini_expect:
+ nf_conntrack_expect_fini();
out_fini_proto:
nf_conntrack_proto_fini();
-out_free_expect_slab:
- kmem_cache_destroy(nf_conntrack_expect_cachep);
err_free_conntrack_slab:
kmem_cache_destroy(nf_conntrack_cachep);
err_free_hash:
- free_conntrack_hash(nf_conntrack_hash, nf_conntrack_vmalloc,
- nf_conntrack_htable_size);
+ nf_ct_free_hashtable(nf_conntrack_hash, nf_conntrack_vmalloc,
+ nf_conntrack_htable_size);
err_out:
return -ENOMEM;
}