netfilter: xtables: don't hook tables by default

delay hook registration until the table is being requested inside a
namespace.

Historically, a particular table (iptables mangle, ip6tables filter, etc)
was registered on module load.

When netns support was added to iptables only the ip/ip6tables ruleset was
made namespace aware, not the actual hook points.

This means f.e. that when ipt_filter table/module is loaded on a system,
then each namespace on that system has an (empty) iptables filter ruleset.

In other words, if a namespace sends a packet, such skb is 'caught' by
netfilter machinery and fed to hooking points for that table (i.e. INPUT,
FORWARD, etc).

Thanks to Eric Biederman, hooks are no longer global, but per namespace.

This means that we can avoid allocation of empty ruleset in a namespace and
defer hook registration until we need the functionality.

We register a tables hook entry points ONLY in the initial namespace.
When an iptables get/setockopt is issued inside a given namespace, we check
if the table is found in the per-namespace list.

If not, we attempt to find it in the initial namespace, and, if found,
create an empty default table in the requesting namespace and register the
needed hooks.

Hook points are destroyed only once namespace is deleted, there is no
'usage count' (it makes no sense since there is no 'remove table' operation
in xtables api).

Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>

1 files changed, 23 insertions, 14 deletions

diff --git a/net/ipv6/netfilter/ip6table_raw.c b/net/ipv6/netfilter/ip6table_raw.c
index 5fac433da069..d4bc56443dc1 100644
--- a/net/ipv6/netfilter/ip6table_raw.c
+++ b/net/ipv6/netfilter/ip6table_raw.c
@@ -9,12 +9,15 @@
 
 #define RAW_VALID_HOOKS ((1 << NF_INET_PRE_ROUTING) | (1 << NF_INET_LOCAL_OUT))
 
+static int __net_init ip6table_raw_table_init(struct net *net);
+
 static const struct xt_table packet_raw = {
 	.name = "raw",
 	.valid_hooks = RAW_VALID_HOOKS,
 	.me = THIS_MODULE,
 	.af = NFPROTO_IPV6,
 	.priority = NF_IP6_PRI_RAW,
+	.table_init = ip6table_raw_table_init,
 };
 
 /* The work comes in here from netfilter.c. */
@@ -27,11 +30,14 @@ ip6table_raw_hook(void *priv, struct sk_buff *skb,
 
 static struct nf_hook_ops *rawtable_ops __read_mostly;
 
-static int __net_init ip6table_raw_net_init(struct net *net)
+static int __net_init ip6table_raw_table_init(struct net *net)
 {
 	struct ip6t_replace *repl;
 	int ret;
 
+	if (net->ipv6.ip6table_raw)
+		return 0;
+
 	repl = ip6t_alloc_initial_table(&packet_raw);
 	if (repl == NULL)
 		return -ENOMEM;
@@ -43,11 +49,13 @@ static int __net_init ip6table_raw_net_init(struct net *net)
 
 static void __net_exit ip6table_raw_net_exit(struct net *net)
 {
+	if (!net->ipv6.ip6table_raw)
+		return;
 	ip6t_unregister_table(net, net->ipv6.ip6table_raw, rawtable_ops);
+	net->ipv6.ip6table_raw = NULL;
 }
 
 static struct pernet_operations ip6table_raw_net_ops = {
-	.init = ip6table_raw_net_init,
 	.exit = ip6table_raw_net_exit,
 };
 
@@ -55,28 +63,29 @@ static int __init ip6table_raw_init(void)
 {
 	int ret;
 
+	/* Register hooks */
+	rawtable_ops = xt_hook_ops_alloc(&packet_raw, ip6table_raw_hook);
+	if (IS_ERR(rawtable_ops))
+		return PTR_ERR(rawtable_ops);
+
 	ret = register_pernet_subsys(&ip6table_raw_net_ops);
-	if (ret < 0)
+	if (ret < 0) {
+		kfree(rawtable_ops);
 		return ret;
-
-	/* Register hooks */
-	rawtable_ops = xt_hook_link(&packet_raw, ip6table_raw_hook);
-	if (IS_ERR(rawtable_ops)) {
-		ret = PTR_ERR(rawtable_ops);
-		goto cleanup_table;
 	}
 
-	return ret;
-
- cleanup_table:
-	unregister_pernet_subsys(&ip6table_raw_net_ops);
+	ret = ip6table_raw_table_init(&init_net);
+	if (ret) {
+		unregister_pernet_subsys(&ip6table_raw_net_ops);
+		kfree(rawtable_ops);
+	}
 	return ret;
 }
 
 static void __exit ip6table_raw_fini(void)
 {
-	xt_hook_unlink(&packet_raw, rawtable_ops);
 	unregister_pernet_subsys(&ip6table_raw_net_ops);
+	kfree(rawtable_ops);
 }
 
 module_init(ip6table_raw_init);