OpenVZ Forum


Home » Mailing lists » Devel » [PATCH 2/5] netns netfilter: per-netns IPv6 FILTER, MANGLE, RAW
[PATCH 2/5] netns netfilter: per-netns IPv6 FILTER, MANGLE, RAW [message #26465] Thu, 24 January 2008 12:27 Go to previous message
Alexey Dobriyan is currently offline  Alexey Dobriyan
Messages: 195
Registered: August 2006
Senior Member
Now it's possible to list and manipulate per-netns ip6tables rules.
Filtering decisions are based on init_net's table so far.

P.S.: remove init_net check in inet6_create() to see the effect

Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
---

 include/net/netns/ipv6.h             |    6 +++++
 net/ipv6/netfilter/ip6table_filter.c |   40 +++++++++++++++++++++++++----------
 net/ipv6/netfilter/ip6table_mangle.c |   40 +++++++++++++++++++++++++----------
 net/ipv6/netfilter/ip6table_raw.c    |   38 ++++++++++++++++++++++++---------
 4 files changed, 92 insertions(+), 32 deletions(-)

--- a/include/net/netns/ipv6.h
+++ b/include/net/netns/ipv6.h
@@ -31,5 +31,10 @@ struct netns_ipv6 {
 	struct ipv6_devconf	*devconf_all;
 	struct ipv6_devconf	*devconf_dflt;
 	struct netns_frags	frags;
+#ifdef CONFIG_NETFILTER
+	struct xt_table		*ip6table_filter;
+	struct xt_table		*ip6table_mangle;
+	struct xt_table		*ip6table_raw;
+#endif
 };
 #endif
--- a/net/ipv6/netfilter/ip6table_filter.c
+++ b/net/ipv6/netfilter/ip6table_filter.c
@@ -26,7 +26,7 @@ static struct
 	struct ip6t_replace repl;
 	struct ip6t_standard entries[3];
 	struct ip6t_error term;
-} initial_table __initdata = {
+} initial_table __net_initdata = {
 	.repl = {
 		.name = "filter",
 		.valid_hooks = FILTER_VALID_HOOKS,
@@ -51,14 +51,13 @@ static struct
 	.term = IP6T_ERROR_INIT,		/* ERROR */
 };
 
-static struct xt_table __packet_filter = {
+static struct xt_table packet_filter = {
 	.name		= "filter",
 	.valid_hooks	= FILTER_VALID_HOOKS,
 	.lock		= RW_LOCK_UNLOCKED,
 	.me		= THIS_MODULE,
 	.af		= AF_INET6,
 };
-static struct xt_table *packet_filter;
 
 /* The work comes in here from netfilter.c. */
 static unsigned int
@@ -68,7 +67,7 @@ ip6t_hook(unsigned int hook,
 	 const struct net_device *out,
 	 int (*okfn)(struct sk_buff *))
 {
-	return ip6t_do_table(skb, hook, in, out, packet_filter);
+	return ip6t_do_table(skb, hook, in, out, init_net.ipv6.ip6table_filter);
 }
 
 static unsigned int
@@ -88,7 +87,7 @@ ip6t_local_out_hook(unsigned int hook,
 	}
 #endif
 
-	return ip6t_do_table(skb, hook, in, out, packet_filter);
+	return ip6t_do_table(skb, hook, in, out, init_net.ipv6.ip6table_filter);
 }
 
 static struct nf_hook_ops ip6t_ops[] __read_mostly = {
@@ -119,6 +118,26 @@ static struct nf_hook_ops ip6t_ops[] __read_mostly = {
 static int forward = NF_ACCEPT;
 module_param(forward, bool, 0000);
 
+static int __net_init ip6table_filter_net_init(struct net *net)
+{
+	/* Register table */
+	net->ipv6.ip6table_filter =
+		ip6t_register_table(net, &packet_filter, &initial_table.repl);
+	if (IS_ERR(net->ipv6.ip6table_filter))
+		return PTR_ERR(net->ipv6.ip6table_filter);
+	return 0;
+}
+
+static void __net_exit ip6table_filter_net_exit(struct net *net)
+{
+	ip6t_unregister_table(net->ipv6.ip6table_filter);
+}
+
+static struct pernet_operations ip6table_filter_net_ops = {
+	.init = ip6table_filter_net_init,
+	.exit = ip6table_filter_net_exit,
+};
+
 static int __init ip6table_filter_init(void)
 {
 	int ret;
@@ -131,10 +150,9 @@ static int __init ip6table_filter_init(void)
 	/* Entry 1 is the FORWARD hook */
 	initial_table.entries[1].target.verdict = -forward - 1;
 
-	/* Register table */
-	packet_filter = ip6t_register_table(&init_net, &__packet_filter, &initial_table.repl);
-	if (IS_ERR(packet_filter))
-		return PTR_ERR(packet_filter);
+	ret = register_pernet_subsys(&ip6table_filter_net_ops);
+	if (ret < 0)
+		return ret;
 
 	/* Register hooks */
 	ret = nf_register_hooks(ip6t_ops, ARRAY_SIZE(ip6t_ops));
@@ -144,14 +162,14 @@ static int __init ip6table_filter_init(void)
 	return ret;
 
  cleanup_table:
-	ip6t_unregister_table(packet_filter);
+	unregister_pernet_subsys(&ip6table_filter_net_ops);
 	return ret;
 }
 
 static void __exit ip6table_filter_fini(void)
 {
 	nf_unregister_hooks(ip6t_ops, ARRAY_SIZE(ip6t_ops));
-	ip6t_unregister_table(packet_filter);
+	unregister_pernet_subsys(&ip6table_filter_net_ops);
 }
 
 module_init(ip6table_filter_init);
--- a/net/ipv6/netfilter/ip6table_mangle.c
+++ b/net/ipv6/netfilter/ip6table_mangle.c
@@ -26,7 +26,7 @@ static struct
 	struct ip6t_replace repl;
 	struct ip6t_standard entries[5];
 	struct ip6t_error term;
-} initial_table __initdata = {
+} initial_table __net_initdata = {
 	.repl = {
 		.name = "mangle",
 		.valid_hooks = MANGLE_VALID_HOOKS,
@@ -57,14 +57,13 @@ static struct
 	.term = IP6T_ERROR_INIT,		/* ERROR */
 };
 
-static struct xt_table __packet_mangler = {
+static struct xt_table packet_mangler = {
 	.name		= "mangle",
 	.valid_hooks	= MANGLE_VALID_HOOKS,
 	.lock		= RW_LOCK_UNLOCKED,
 	.me		= THIS_MODULE,
 	.af		= AF_INET6,
 };
-static struct xt_table *packet_mangler;
 
 /* The work comes in here from netfilter.c. */
 static unsigned int
@@ -74,7 +73,7 @@ ip6t_route_hook(unsigned int hook,
 	 const struct net_device *out,
 	 int (*okfn)(struct sk_buff *))
 {
-	return ip6t_do_table(skb, hook, in, out, packet_mangler);
+	return ip6t_do_table(skb, hook, in, out, init_net.ipv6.ip6table_mangle);
 }
 
 static unsigned int
@@ -109,7 +108,7 @@ ip6t_local_hook(unsigned int hook,
 	/* flowlabel and prio (includes version, which shouldn't change either */
 	flowlabel = *((u_int32_t *)ipv6_hdr(skb));
 
-	ret = ip6t_do_table(skb, hook, in, out, packet_mangler);
+	ret = ip6t_do_table(skb, hook, in, out, init_net.ipv6.ip6table_mangle);
 
 	if (ret != NF_DROP && ret != NF_STOLEN
 		&& (memcmp(&ipv6_hdr(skb)->saddr, &saddr, sizeof(saddr))
@@ -159,14 +158,33 @@ static struct nf_hook_ops ip6t_ops[] __read_mostly = {
 	},
 };
 
+static int __net_init ip6table_mangle_net_init(struct net *net)
+{
+	/* Register table */
+	net->ipv6.ip6table_mangle =
+		ip6t_register_table(net, &packet_mangler, &initial_table.repl);
+	if (IS_ERR(net->ipv6.ip6table_mangle))
+		return PTR_ERR(net->ipv6.ip6table_mangle);
+	return 0;
+}
+
+static void __net_exit ip6table_mangle_net_exit(struct net *net)
+{
+	ip6t_unregister_table(net->ipv6.ip6table_mangle);
+}
+
+static struct pernet_operations ip6table_mangle_net_ops = {
+	.init = ip6table_mangle_net_init,
+	.exit = ip6table_mangle_net_exit,
+};
+
 static int __init ip6table_mangle_init(void)
 {
 	int ret;
 
-	/* Register table */
-	packet_mangler = ip6t_register_table(&init_net, &__packet_mangler, &initial_table.repl);
-	if (IS_ERR(packet_mangler))
-		return PTR_ERR(packet_mangler);
+	ret = register_pernet_subsys(&ip6table_mangle_net_ops);
+	if (ret < 0)
+		return ret;
 
 	/* Register hooks */
 	ret = nf_register_hooks(ip6t_ops, ARRAY_SIZE(ip6t_ops));
@@ -176,14 +194,14 @@ static int __init ip6table_mangle_init(void)
 	return ret;
 
  cleanup_table:
-	ip6t_unregister_table(packet_mangler);
+	unregister_pernet_subsys(&ip6table_mangle_net_ops);
 	return ret;
 }
 
 static void __exit ip6table_mangle_fini(void)
 {
 	nf_unregister_hooks(ip6t_ops, ARRAY_SIZE(ip6t_ops));
-	ip6t_unregister_table(packet_mangler);
+	unregister_pernet_subsys(&ip6table_mangle_net_ops);
 }
 
 module_init(ip6table_mangle_init);
--- a/net/ipv6/netfilter/ip6table_raw.c
+++ b/net/ipv6/netfilter/ip6table_raw.c
@@ -13,7 +13,7 @@ static struct
 	struct ip6t_replace repl;
 	struct ip6t_standard entries[2];
 	struct ip6t_error term;
-} initial_table __initdata = {
+} initial_table __net_initdata = {
 	.repl = {
 		.name = "raw",
 		.valid_hooks = RAW_VALID_HOOKS,
@@ -35,14 +35,13 @@ static struct
 	.term = IP6T_ERROR_INIT,		/* ERROR */
 };
 
-static struct xt_table __packet_raw = {
+static struct xt_table packet_raw = {
 	.name = "raw",
 	.valid_hooks = RAW_VALID_HOOKS,
 	.lock = RW_LOCK_UNLOCKED,
 	.me = THIS_MODULE,
 	.af = AF_INET6,
 };
-static struct xt_table *packet_raw;
 
 /* The work comes in here from netfilter.c. */
 static unsigned int
@@ -52,7 +51,7 @@ ip6t_hook(unsigned int hook,
 	 const struct net_device *out,
 	 int (*okfn)(struct sk_buff *))
 {
-	return ip6t_do_table(skb, hook, in, out, packet_raw);
+	return ip6t_do_table(skb, hook, in, out, init_net.ipv6.ip6table_raw);
 }
 
 static struct nf_hook_ops ip6t_ops[] __read_mostly = {
@@ -72,14 +71,33 @@ static struct nf_hook_ops ip6t_ops[] __read_mostly = {
 	},
 };
 
+static int __net_init ip6table_raw_net_init(struct net *net)
+{
+	/* Register table */
+	net->ipv6.ip6table_raw =
+		ip6t_register_table(net, &packet_raw, &initial_table.repl);
+	if (IS_ERR(net->ipv6.ip6table_raw))
+		return PTR_ERR(net->ipv6.ip6table_raw);
+	return 0;
+}
+
+static void __net_exit ip6table_raw_net_exit(struct net *net)
+{
+	ip6t_unregister_table(net->ipv6.ip6table_raw);
+}
+
+static struct pernet_operations ip6table_raw_net_ops = {
+	.init = ip6table_raw_net_init,
+	.exit = ip6table_raw_net_exit,
+};
+
 static int __init ip6table_raw_init(void)
 {
 	int ret;
 
-	/* Register table */
-	packet_raw = ip6t_register_table(&init_net, &__packet_raw, &initial_table.repl);
-	if (IS_ERR(packet_raw))
-		return PTR_ERR(packet_raw);
+	ret = register_pernet_subsys(&ip6table_raw_net_ops);
+	if (ret < 0)
+		return ret;
 
 	/* Register hooks */
 	ret = nf_register_hooks(ip6t_ops, ARRAY_SIZE(ip6t_ops));
@@ -89,14 +107,14 @@ static int __init ip6table_raw_init(void)
 	return ret;
 
  cleanup_table:
-	ip6t_unregister_table(packet_raw);
+	unregister_pernet_subsys(&ip6table_raw_net_ops);
 	return ret;
 }
 
 static void __exit ip6table_raw_fini(void)
 {
 	nf_unregister_hooks(ip6t_ops, ARRAY_SIZE(ip6t_ops));
-	ip6t_unregister_table(packet_raw);
+	unregister_pernet_subsys(&ip6table_raw_net_ops);
 }
 
 module_init(ip6table_raw_init);
 
Read Message
Read Message
Previous Topic: [PATCH net-2.6.25][IPV6]: Introduce the INET6_TW_MATCH() macro.
Next Topic: [PATCH 3/5] netns netfilter: per-netns arp_tables
Goto Forum:
  


Current Time: Tue Jul 23 01:26:33 GMT 2024

Total time taken to generate the page: 0.02340 seconds