[50/93] netfilter: nf_conntrack: fix hash resizing with namespaces

[Greg KH <gregkh@suse.de>]

2.6.32-stable review patch.  If anyone has any objections, please let us know.

------------------

From: Patrick McHardy <kaber@trash.net>

commit d696c7bdaa55e2208e56c6f98e6bc1599f34286d upstream.

As noticed by Jon Masters <jonathan@jonmasters.org>, the conntrack hash
size is global and not per namespace, but modifiable at runtime through
/sys/module/nf_conntrack/hashsize. Changing the hash size will only
resize the hash in the current namespace however, so other namespaces
will use an invalid hash size. This can cause crashes when enlarging
the hashsize, or false negative lookups when shrinking it.

Move the hash size into the per-namespace data and only use the global
hash size to initialize the per-namespace value when instanciating a
new namespace. Additionally restrict hash resizing to init_net for
now as other namespaces are not handled currently.

Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

---
 include/net/netns/conntrack.h                         |    1 
 include/net/netns/ipv4.h                              |    1 
 net/ipv4/netfilter/nf_conntrack_l3proto_ipv4.c        |    2 
 net/ipv4/netfilter/nf_conntrack_l3proto_ipv4_compat.c |    4 -
 net/ipv4/netfilter/nf_nat_core.c                      |   22 +++----
 net/netfilter/nf_conntrack_core.c                     |   54 +++++++++---------
 net/netfilter/nf_conntrack_expect.c                   |    2 
 net/netfilter/nf_conntrack_helper.c                   |    2 
 net/netfilter/nf_conntrack_netlink.c                  |    2 
 net/netfilter/nf_conntrack_standalone.c               |    7 +-
 10 files changed, 50 insertions(+), 47 deletions(-)

--- a/include/net/netns/conntrack.h
+++ b/include/net/netns/conntrack.h
@@ -11,6 +11,7 @@ struct nf_conntrack_ecache;
 struct netns_ct {
 	atomic_t		count;
 	unsigned int		expect_count;
+	unsigned int		htable_size;
 	struct kmem_cache	*nf_conntrack_cachep;
 	struct hlist_nulls_head	*hash;
 	struct hlist_head	*expect_hash;
--- a/include/net/netns/ipv4.h
+++ b/include/net/netns/ipv4.h
@@ -40,6 +40,7 @@ struct netns_ipv4 {
 	struct xt_table		*iptable_security;
 	struct xt_table		*nat_table;
 	struct hlist_head	*nat_bysource;
+	unsigned int		nat_htable_size;
 	int			nat_vmalloced;
 #endif
 
--- a/net/ipv4/netfilter/nf_conntrack_l3proto_ipv4.c
+++ b/net/ipv4/netfilter/nf_conntrack_l3proto_ipv4.c
@@ -213,7 +213,7 @@ static ctl_table ip_ct_sysctl_table[] =
 	{
 		.ctl_name	= NET_IPV4_NF_CONNTRACK_BUCKETS,
 		.procname	= "ip_conntrack_buckets",
-		.data		= &nf_conntrack_htable_size,
+		.data		= &init_net.ct.htable_size,
 		.maxlen		= sizeof(unsigned int),
 		.mode		= 0444,
 		.proc_handler	= proc_dointvec,
--- a/net/ipv4/netfilter/nf_conntrack_l3proto_ipv4_compat.c
+++ b/net/ipv4/netfilter/nf_conntrack_l3proto_ipv4_compat.c
@@ -32,7 +32,7 @@ static struct hlist_nulls_node *ct_get_f
 	struct hlist_nulls_node *n;
 
 	for (st->bucket = 0;
-	     st->bucket < nf_conntrack_htable_size;
+	     st->bucket < net->ct.htable_size;
 	     st->bucket++) {
 		n = rcu_dereference(net->ct.hash[st->bucket].first);
 		if (!is_a_nulls(n))
@@ -50,7 +50,7 @@ static struct hlist_nulls_node *ct_get_n
 	head = rcu_dereference(head->next);
 	while (is_a_nulls(head)) {
 		if (likely(get_nulls_value(head) == st->bucket)) {
-			if (++st->bucket >= nf_conntrack_htable_size)
+			if (++st->bucket >= net->ct.htable_size)
 				return NULL;
 		}
 		head = rcu_dereference(net->ct.hash[st->bucket].first);
--- a/net/ipv4/netfilter/nf_nat_core.c
+++ b/net/ipv4/netfilter/nf_nat_core.c
@@ -35,9 +35,6 @@ static DEFINE_SPINLOCK(nf_nat_lock);
 
 static struct nf_conntrack_l3proto *l3proto __read_mostly;
 
-/* Calculated at init based on memory size */
-static unsigned int nf_nat_htable_size __read_mostly;
-
 #define MAX_IP_NAT_PROTO 256
 static const struct nf_nat_protocol *nf_nat_protos[MAX_IP_NAT_PROTO]
 						__read_mostly;
@@ -72,7 +69,7 @@ EXPORT_SYMBOL_GPL(nf_nat_proto_put);
 
 /* We keep an extra hash for each conntrack, for fast searching. */
 static inline unsigned int
-hash_by_src(const struct nf_conntrack_tuple *tuple)
+hash_by_src(const struct net *net, const struct nf_conntrack_tuple *tuple)
 {
 	unsigned int hash;
 
@@ -80,7 +77,7 @@ hash_by_src(const struct nf_conntrack_tu
 	hash = jhash_3words((__force u32)tuple->src.u3.ip,
 			    (__force u32)tuple->src.u.all,
 			    tuple->dst.protonum, 0);
-	return ((u64)hash * nf_nat_htable_size) >> 32;
+	return ((u64)hash * net->ipv4.nat_htable_size) >> 32;
 }
 
 /* Is this tuple already taken? (not by us) */
@@ -147,7 +144,7 @@ find_appropriate_src(struct net *net,
 		     struct nf_conntrack_tuple *result,
 		     const struct nf_nat_range *range)
 {
-	unsigned int h = hash_by_src(tuple);
+	unsigned int h = hash_by_src(net, tuple);
 	const struct nf_conn_nat *nat;
 	const struct nf_conn *ct;
 	const struct hlist_node *n;
@@ -330,7 +327,7 @@ nf_nat_setup_info(struct nf_conn *ct,
 	if (have_to_hash) {
 		unsigned int srchash;
 
-		srchash = hash_by_src(&ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple);
+		srchash = hash_by_src(net, &ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple);
 		spin_lock_bh(&nf_nat_lock);
 		/* nf_conntrack_alter_reply might re-allocate exntension aera */
 		nat = nfct_nat(ct);
@@ -679,8 +676,10 @@ nfnetlink_parse_nat_setup(struct nf_conn
 
 static int __net_init nf_nat_net_init(struct net *net)
 {
-	net->ipv4.nat_bysource = nf_ct_alloc_hashtable(&nf_nat_htable_size,
-						      &net->ipv4.nat_vmalloced, 0);
+	/* Leave them the same for the moment. */
+	net->ipv4.nat_htable_size = net->ct.htable_size;
+	net->ipv4.nat_bysource = nf_ct_alloc_hashtable(&net->ipv4.nat_htable_size,
+						       &net->ipv4.nat_vmalloced, 0);
 	if (!net->ipv4.nat_bysource)
 		return -ENOMEM;
 	return 0;
@@ -703,7 +702,7 @@ static void __net_exit nf_nat_net_exit(s
 	nf_ct_iterate_cleanup(net, &clean_nat, NULL);
 	synchronize_rcu();
 	nf_ct_free_hashtable(net->ipv4.nat_bysource, net->ipv4.nat_vmalloced,
-			     nf_nat_htable_size);
+			     net->ipv4.nat_htable_size);
 }
 
 static struct pernet_operations nf_nat_net_ops = {
@@ -724,9 +723,6 @@ static int __init nf_nat_init(void)
 		return ret;
 	}
 
-	/* Leave them the same for the moment. */
-	nf_nat_htable_size = nf_conntrack_htable_size;
-
 	ret = register_pernet_subsys(&nf_nat_net_ops);
 	if (ret < 0)
 		goto cleanup_extend;
--- a/net/netfilter/nf_conntrack_core.c
+++ b/net/netfilter/nf_conntrack_core.c
@@ -30,6 +30,7 @@
 #include <linux/netdevice.h>
 #include <linux/socket.h>
 #include <linux/mm.h>
+#include <linux/nsproxy.h>
 #include <linux/rculist_nulls.h>
 
 #include <net/netfilter/nf_conntrack.h>
@@ -84,9 +85,10 @@ static u_int32_t __hash_conntrack(const
 	return ((u64)h * size) >> 32;
 }
 
-static inline u_int32_t hash_conntrack(const struct nf_conntrack_tuple *tuple)
+static inline u_int32_t hash_conntrack(const struct net *net,
+				       const struct nf_conntrack_tuple *tuple)
 {
-	return __hash_conntrack(tuple, nf_conntrack_htable_size,
+	return __hash_conntrack(tuple, net->ct.htable_size,
 				nf_conntrack_hash_rnd);
 }
 
@@ -294,7 +296,7 @@ __nf_conntrack_find(struct net *net, con
 {
 	struct nf_conntrack_tuple_hash *h;
 	struct hlist_nulls_node *n;
-	unsigned int hash = hash_conntrack(tuple);
+	unsigned int hash = hash_conntrack(net, tuple);
 
 	/* Disable BHs the entire time since we normally need to disable them
 	 * at least once for the stats anyway.
@@ -364,10 +366,11 @@ static void __nf_conntrack_hash_insert(s
 
 void nf_conntrack_hash_insert(struct nf_conn *ct)
 {
+	struct net *net = nf_ct_net(ct);
 	unsigned int hash, repl_hash;
 
-	hash = hash_conntrack(&ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple);
-	repl_hash = hash_conntrack(&ct->tuplehash[IP_CT_DIR_REPLY].tuple);
+	hash = hash_conntrack(net, &ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple);
+	repl_hash = hash_conntrack(net, &ct->tuplehash[IP_CT_DIR_REPLY].tuple);
 
 	__nf_conntrack_hash_insert(ct, hash, repl_hash);
 }
@@ -395,8 +398,8 @@ __nf_conntrack_confirm(struct sk_buff *s
 	if (CTINFO2DIR(ctinfo) != IP_CT_DIR_ORIGINAL)
 		return NF_ACCEPT;
 
-	hash = hash_conntrack(&ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple);
-	repl_hash = hash_conntrack(&ct->tuplehash[IP_CT_DIR_REPLY].tuple);
+	hash = hash_conntrack(net, &ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple);
+	repl_hash = hash_conntrack(net, &ct->tuplehash[IP_CT_DIR_REPLY].tuple);
 
 	/* We're not in hash table, and we refuse to set up related
 	   connections for unconfirmed conns.  But packet copies and
@@ -466,7 +469,7 @@ nf_conntrack_tuple_taken(const struct nf
 	struct net *net = nf_ct_net(ignored_conntrack);
 	struct nf_conntrack_tuple_hash *h;
 	struct hlist_nulls_node *n;
-	unsigned int hash = hash_conntrack(tuple);
+	unsigned int hash = hash_conntrack(net, tuple);
 
 	/* Disable BHs the entire time since we need to disable them at
 	 * least once for the stats anyway.
@@ -501,7 +504,7 @@ static noinline int early_drop(struct ne
 	int dropped = 0;
 
 	rcu_read_lock();
-	for (i = 0; i < nf_conntrack_htable_size; i++) {
+	for (i = 0; i < net->ct.htable_size; i++) {
 		hlist_nulls_for_each_entry_rcu(h, n, &net->ct.hash[hash],
 					 hnnode) {
 			tmp = nf_ct_tuplehash_to_ctrack(h);
@@ -515,7 +518,8 @@ static noinline int early_drop(struct ne
 			ct = NULL;
 		if (ct || cnt >= NF_CT_EVICTION_RANGE)
 			break;
-		hash = (hash + 1) % nf_conntrack_htable_size;
+
+		hash = (hash + 1) % net->ct.htable_size;
 	}
 	rcu_read_unlock();
 
@@ -549,7 +553,7 @@ struct nf_conn *nf_conntrack_alloc(struc
 
 	if (nf_conntrack_max &&
 	    unlikely(atomic_read(&net->ct.count) > nf_conntrack_max)) {
-		unsigned int hash = hash_conntrack(orig);
+		unsigned int hash = hash_conntrack(net, orig);
 		if (!early_drop(net, hash)) {
 			atomic_dec(&net->ct.count);
 			if (net_ratelimit())
@@ -1006,7 +1010,7 @@ get_next_corpse(struct net *net, int (*i
 	struct hlist_nulls_node *n;
 
 	spin_lock_bh(&nf_conntrack_lock);
-	for (; *bucket < nf_conntrack_htable_size; (*bucket)++) {
+	for (; *bucket < net->ct.htable_size; (*bucket)++) {
 		hlist_nulls_for_each_entry(h, n, &net->ct.hash[*bucket], hnnode) {
 			ct = nf_ct_tuplehash_to_ctrack(h);
 			if (iter(ct, data))
@@ -1124,7 +1128,7 @@ static void nf_conntrack_cleanup_net(str
 	}
 
 	nf_ct_free_hashtable(net->ct.hash, net->ct.hash_vmalloc,
-			     nf_conntrack_htable_size);
+			     net->ct.htable_size);
 	nf_conntrack_ecache_fini(net);
 	nf_conntrack_acct_fini(net);
 	nf_conntrack_expect_fini(net);
@@ -1184,10 +1188,12 @@ int nf_conntrack_set_hashsize(const char
 {
 	int i, bucket, vmalloced, old_vmalloced;
 	unsigned int hashsize, old_size;
-	int rnd;
 	struct hlist_nulls_head *hash, *old_hash;
 	struct nf_conntrack_tuple_hash *h;
 
+	if (current->nsproxy->net_ns != &init_net)
+		return -EOPNOTSUPP;
+
 	/* On boot, we can set this without any fancy locking. */
 	if (!nf_conntrack_htable_size)
 		return param_set_uint(val, kp);
@@ -1200,33 +1206,29 @@ int nf_conntrack_set_hashsize(const char
 	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, sizeof(rnd));
-
 	/* Lookups in the old hash might happen in parallel, which means we
 	 * might get false negatives during connection lookup. New connections
 	 * created because of a false negative won't make it into the hash
 	 * though since that required taking the lock.
 	 */
 	spin_lock_bh(&nf_conntrack_lock);
-	for (i = 0; i < nf_conntrack_htable_size; i++) {
+	for (i = 0; i < init_net.ct.htable_size; i++) {
 		while (!hlist_nulls_empty(&init_net.ct.hash[i])) {
 			h = hlist_nulls_entry(init_net.ct.hash[i].first,
 					struct nf_conntrack_tuple_hash, hnnode);
 			hlist_nulls_del_rcu(&h->hnnode);
-			bucket = __hash_conntrack(&h->tuple, hashsize, rnd);
+			bucket = __hash_conntrack(&h->tuple, hashsize,
+						  nf_conntrack_hash_rnd);
 			hlist_nulls_add_head_rcu(&h->hnnode, &hash[bucket]);
 		}
 	}
-	old_size = nf_conntrack_htable_size;
+	old_size = init_net.ct.htable_size;
 	old_vmalloced = init_net.ct.hash_vmalloc;
 	old_hash = init_net.ct.hash;
 
-	nf_conntrack_htable_size = hashsize;
+	init_net.ct.htable_size = nf_conntrack_htable_size = hashsize;
 	init_net.ct.hash_vmalloc = vmalloced;
 	init_net.ct.hash = hash;
-	nf_conntrack_hash_rnd = rnd;
 	spin_unlock_bh(&nf_conntrack_lock);
 
 	nf_ct_free_hashtable(old_hash, old_vmalloced, old_size);
@@ -1322,7 +1324,9 @@ static int nf_conntrack_init_net(struct
 		ret = -ENOMEM;
 		goto err_cache;
 	}
-	net->ct.hash = nf_ct_alloc_hashtable(&nf_conntrack_htable_size,
+
+	net->ct.htable_size = nf_conntrack_htable_size;
+	net->ct.hash = nf_ct_alloc_hashtable(&net->ct.htable_size,
 					     &net->ct.hash_vmalloc, 1);
 	if (!net->ct.hash) {
 		ret = -ENOMEM;
@@ -1347,7 +1351,7 @@ err_acct:
 	nf_conntrack_expect_fini(net);
 err_expect:
 	nf_ct_free_hashtable(net->ct.hash, net->ct.hash_vmalloc,
-			     nf_conntrack_htable_size);
+			     net->ct.htable_size);
 err_hash:
 	kmem_cache_destroy(net->ct.nf_conntrack_cachep);
 err_cache:
--- a/net/netfilter/nf_conntrack_expect.c
+++ b/net/netfilter/nf_conntrack_expect.c
@@ -577,7 +577,7 @@ int nf_conntrack_expect_init(struct net
 
 	if (net_eq(net, &init_net)) {
 		if (!nf_ct_expect_hsize) {
-			nf_ct_expect_hsize = nf_conntrack_htable_size / 256;
+			nf_ct_expect_hsize = net->ct.htable_size / 256;
 			if (!nf_ct_expect_hsize)
 				nf_ct_expect_hsize = 1;
 		}
--- a/net/netfilter/nf_conntrack_helper.c
+++ b/net/netfilter/nf_conntrack_helper.c
@@ -192,7 +192,7 @@ static void __nf_conntrack_helper_unregi
 	/* Get rid of expecteds, set helpers to NULL. */
 	hlist_nulls_for_each_entry(h, nn, &net->ct.unconfirmed, hnnode)
 		unhelp(h, me);
-	for (i = 0; i < nf_conntrack_htable_size; i++) {
+	for (i = 0; i < net->ct.htable_size; i++) {
 		hlist_nulls_for_each_entry(h, nn, &net->ct.hash[i], hnnode)
 			unhelp(h, me);
 	}
--- a/net/netfilter/nf_conntrack_netlink.c
+++ b/net/netfilter/nf_conntrack_netlink.c
@@ -594,7 +594,7 @@ ctnetlink_dump_table(struct sk_buff *skb
 
 	rcu_read_lock();
 	last = (struct nf_conn *)cb->args[1];
-	for (; cb->args[0] < nf_conntrack_htable_size; cb->args[0]++) {
+	for (; cb->args[0] < init_net.ct.htable_size; cb->args[0]++) {
 restart:
 		hlist_nulls_for_each_entry_rcu(h, n, &init_net.ct.hash[cb->args[0]],
 					 hnnode) {
--- a/net/netfilter/nf_conntrack_standalone.c
+++ b/net/netfilter/nf_conntrack_standalone.c
@@ -51,7 +51,7 @@ static struct hlist_nulls_node *ct_get_f
 	struct hlist_nulls_node *n;
 
 	for (st->bucket = 0;
-	     st->bucket < nf_conntrack_htable_size;
+	     st->bucket < net->ct.htable_size;
 	     st->bucket++) {
 		n = rcu_dereference(net->ct.hash[st->bucket].first);
 		if (!is_a_nulls(n))
@@ -69,7 +69,7 @@ static struct hlist_nulls_node *ct_get_n
 	head = rcu_dereference(head->next);
 	while (is_a_nulls(head)) {
 		if (likely(get_nulls_value(head) == st->bucket)) {
-			if (++st->bucket >= nf_conntrack_htable_size)
+			if (++st->bucket >= net->ct.htable_size)
 				return NULL;
 		}
 		head = rcu_dereference(net->ct.hash[st->bucket].first);
@@ -358,7 +358,7 @@ static ctl_table nf_ct_sysctl_table[] =
 	{
 		.ctl_name       = NET_NF_CONNTRACK_BUCKETS,
 		.procname       = "nf_conntrack_buckets",
-		.data           = &nf_conntrack_htable_size,
+		.data           = &init_net.ct.htable_size,
 		.maxlen         = sizeof(unsigned int),
 		.mode           = 0444,
 		.proc_handler   = proc_dointvec,
@@ -429,6 +429,7 @@ static int nf_conntrack_standalone_init_
 		goto out_kmemdup;
 
 	table[1].data = &net->ct.count;
+	table[2].data = &net->ct.htable_size;
 	table[3].data = &net->ct.sysctl_checksum;
 	table[4].data = &net->ct.sysctl_log_invalid;