[NETFILTER]: ip_tables: per-netns FILTER, MANGLE, RAW
authorAlexey Dobriyan <adobriyan@sw.ru>
Thu, 31 Jan 2008 12:03:23 +0000 (04:03 -0800)
committerDavid S. Miller <davem@davemloft.net>
Fri, 1 Feb 2008 03:27:38 +0000 (19:27 -0800)
Now, iptables show and configure different set of rules in different
netnss'. Filtering decisions are still made by consulting only
init_net's set.

Changes are identical except naming so no splitting.

P.S.: one need to remove init_net checks in nf_sockopt.c and inet_create()
      to see the effect.

Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
include/net/netns/ipv4.h
net/ipv4/netfilter/iptable_filter.c
net/ipv4/netfilter/iptable_mangle.c
net/ipv4/netfilter/iptable_raw.c

index 15a0b05..aeb0c3b 100644 (file)
@@ -27,5 +27,10 @@ struct netns_ipv4 {
        struct sock             *fibnl;
 
        struct netns_frags      frags;
+#ifdef CONFIG_NETFILTER
+       struct xt_table         *iptable_filter;
+       struct xt_table         *iptable_mangle;
+       struct xt_table         *iptable_raw;
+#endif
 };
 #endif
index 3b43ca0..69f3d7e 100644 (file)
@@ -28,7 +28,7 @@ static struct
        struct ipt_replace repl;
        struct ipt_standard entries[3];
        struct ipt_error term;
-} initial_table __initdata = {
+} initial_table __net_initdata = {
        .repl = {
                .name = "filter",
                .valid_hooks = FILTER_VALID_HOOKS,
@@ -53,14 +53,13 @@ static struct
        .term = IPT_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_INET,
 };
-static struct xt_table *packet_filter;
 
 /* The work comes in here from netfilter.c. */
 static unsigned int
@@ -70,7 +69,7 @@ ipt_hook(unsigned int hook,
         const struct net_device *out,
         int (*okfn)(struct sk_buff *))
 {
-       return ipt_do_table(skb, hook, in, out, packet_filter);
+       return ipt_do_table(skb, hook, in, out, init_net.ipv4.iptable_filter);
 }
 
 static unsigned int
@@ -89,7 +88,7 @@ ipt_local_out_hook(unsigned int hook,
                return NF_ACCEPT;
        }
 
-       return ipt_do_table(skb, hook, in, out, packet_filter);
+       return ipt_do_table(skb, hook, in, out, init_net.ipv4.iptable_filter);
 }
 
 static struct nf_hook_ops ipt_ops[] __read_mostly = {
@@ -120,6 +119,26 @@ static struct nf_hook_ops ipt_ops[] __read_mostly = {
 static int forward = NF_ACCEPT;
 module_param(forward, bool, 0000);
 
+static int __net_init iptable_filter_net_init(struct net *net)
+{
+       /* Register table */
+       net->ipv4.iptable_filter =
+               ipt_register_table(net, &packet_filter, &initial_table.repl);
+       if (IS_ERR(net->ipv4.iptable_filter))
+               return PTR_ERR(net->ipv4.iptable_filter);
+       return 0;
+}
+
+static void __net_exit iptable_filter_net_exit(struct net *net)
+{
+       ipt_unregister_table(net->ipv4.iptable_filter);
+}
+
+static struct pernet_operations iptable_filter_net_ops = {
+       .init = iptable_filter_net_init,
+       .exit = iptable_filter_net_exit,
+};
+
 static int __init iptable_filter_init(void)
 {
        int ret;
@@ -132,11 +151,9 @@ static int __init iptable_filter_init(void)
        /* Entry 1 is the FORWARD hook */
        initial_table.entries[1].target.verdict = -forward - 1;
 
-       /* Register table */
-       packet_filter = ipt_register_table(&init_net, &__packet_filter,
-                                          &initial_table.repl);
-       if (IS_ERR(packet_filter))
-               return PTR_ERR(packet_filter);
+       ret = register_pernet_subsys(&iptable_filter_net_ops);
+       if (ret < 0)
+               return ret;
 
        /* Register hooks */
        ret = nf_register_hooks(ipt_ops, ARRAY_SIZE(ipt_ops));
@@ -146,14 +163,14 @@ static int __init iptable_filter_init(void)
        return ret;
 
  cleanup_table:
-       ipt_unregister_table(packet_filter);
+       unregister_pernet_subsys(&iptable_filter_net_ops);
        return ret;
 }
 
 static void __exit iptable_filter_fini(void)
 {
        nf_unregister_hooks(ipt_ops, ARRAY_SIZE(ipt_ops));
-       ipt_unregister_table(packet_filter);
+       unregister_pernet_subsys(&iptable_filter_net_ops);
 }
 
 module_init(iptable_filter_init);
index 292f2ed..c55a210 100644 (file)
@@ -33,7 +33,7 @@ static struct
        struct ipt_replace repl;
        struct ipt_standard entries[5];
        struct ipt_error term;
-} initial_table __initdata = {
+} initial_table __net_initdata = {
        .repl = {
                .name = "mangle",
                .valid_hooks = MANGLE_VALID_HOOKS,
@@ -64,14 +64,13 @@ static struct
        .term = IPT_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_INET,
 };
-static struct xt_table *packet_mangler;
 
 /* The work comes in here from netfilter.c. */
 static unsigned int
@@ -81,7 +80,7 @@ ipt_route_hook(unsigned int hook,
         const struct net_device *out,
         int (*okfn)(struct sk_buff *))
 {
-       return ipt_do_table(skb, hook, in, out, packet_mangler);
+       return ipt_do_table(skb, hook, in, out, init_net.ipv4.iptable_mangle);
 }
 
 static unsigned int
@@ -113,7 +112,7 @@ ipt_local_hook(unsigned int hook,
        daddr = iph->daddr;
        tos = iph->tos;
 
-       ret = ipt_do_table(skb, hook, in, out, packet_mangler);
+       ret = ipt_do_table(skb, hook, in, out, init_net.ipv4.iptable_mangle);
        /* Reroute for ANY change. */
        if (ret != NF_DROP && ret != NF_STOLEN && ret != NF_QUEUE) {
                iph = ip_hdr(skb);
@@ -167,15 +166,33 @@ static struct nf_hook_ops ipt_ops[] __read_mostly = {
        },
 };
 
+static int __net_init iptable_mangle_net_init(struct net *net)
+{
+       /* Register table */
+       net->ipv4.iptable_mangle =
+               ipt_register_table(net, &packet_mangler, &initial_table.repl);
+       if (IS_ERR(net->ipv4.iptable_mangle))
+               return PTR_ERR(net->ipv4.iptable_mangle);
+       return 0;
+}
+
+static void __net_exit iptable_mangle_net_exit(struct net *net)
+{
+       ipt_unregister_table(net->ipv4.iptable_mangle);
+}
+
+static struct pernet_operations iptable_mangle_net_ops = {
+       .init = iptable_mangle_net_init,
+       .exit = iptable_mangle_net_exit,
+};
+
 static int __init iptable_mangle_init(void)
 {
        int ret;
 
-       /* Register table */
-       packet_mangler = ipt_register_table(&init_net, &__packet_mangler,
-                                           &initial_table.repl);
-       if (IS_ERR(packet_mangler))
-               return PTR_ERR(packet_mangler);
+       ret = register_pernet_subsys(&iptable_mangle_net_ops);
+       if (ret < 0)
+               return ret;
 
        /* Register hooks */
        ret = nf_register_hooks(ipt_ops, ARRAY_SIZE(ipt_ops));
@@ -185,14 +202,14 @@ static int __init iptable_mangle_init(void)
        return ret;
 
  cleanup_table:
-       ipt_unregister_table(packet_mangler);
+       unregister_pernet_subsys(&iptable_mangle_net_ops);
        return ret;
 }
 
 static void __exit iptable_mangle_fini(void)
 {
        nf_unregister_hooks(ipt_ops, ARRAY_SIZE(ipt_ops));
-       ipt_unregister_table(packet_mangler);
+       unregister_pernet_subsys(&iptable_mangle_net_ops);
 }
 
 module_init(iptable_mangle_init);
index dab863d..e41fe8c 100644 (file)
@@ -14,7 +14,7 @@ static struct
        struct ipt_replace repl;
        struct ipt_standard entries[2];
        struct ipt_error term;
-} initial_table __initdata = {
+} initial_table __net_initdata = {
        .repl = {
                .name = "raw",
                .valid_hooks = RAW_VALID_HOOKS,
@@ -36,14 +36,13 @@ static struct
        .term = IPT_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_INET,
 };
-static struct xt_table *packet_raw;
 
 /* The work comes in here from netfilter.c. */
 static unsigned int
@@ -53,7 +52,7 @@ ipt_hook(unsigned int hook,
         const struct net_device *out,
         int (*okfn)(struct sk_buff *))
 {
-       return ipt_do_table(skb, hook, in, out, packet_raw);
+       return ipt_do_table(skb, hook, in, out, init_net.ipv4.iptable_raw);
 }
 
 static unsigned int
@@ -71,7 +70,7 @@ ipt_local_hook(unsigned int hook,
                               "packet.\n");
                return NF_ACCEPT;
        }
-       return ipt_do_table(skb, hook, in, out, packet_raw);
+       return ipt_do_table(skb, hook, in, out, init_net.ipv4.iptable_raw);
 }
 
 /* 'raw' is the very first table. */
@@ -92,15 +91,33 @@ static struct nf_hook_ops ipt_ops[] __read_mostly = {
        },
 };
 
+static int __net_init iptable_raw_net_init(struct net *net)
+{
+       /* Register table */
+       net->ipv4.iptable_raw =
+               ipt_register_table(net, &packet_raw, &initial_table.repl);
+       if (IS_ERR(net->ipv4.iptable_raw))
+               return PTR_ERR(net->ipv4.iptable_raw);
+       return 0;
+}
+
+static void __net_exit iptable_raw_net_exit(struct net *net)
+{
+       ipt_unregister_table(net->ipv4.iptable_raw);
+}
+
+static struct pernet_operations iptable_raw_net_ops = {
+       .init = iptable_raw_net_init,
+       .exit = iptable_raw_net_exit,
+};
+
 static int __init iptable_raw_init(void)
 {
        int ret;
 
-       /* Register table */
-       packet_raw = ipt_register_table(&init_net, &__packet_raw,
-                                       &initial_table.repl);
-       if (IS_ERR(packet_raw))
-               return PTR_ERR(packet_raw);
+       ret = register_pernet_subsys(&iptable_raw_net_ops);
+       if (ret < 0)
+               return ret;
 
        /* Register hooks */
        ret = nf_register_hooks(ipt_ops, ARRAY_SIZE(ipt_ops));
@@ -110,14 +127,14 @@ static int __init iptable_raw_init(void)
        return ret;
 
  cleanup_table:
-       ipt_unregister_table(packet_raw);
+       unregister_pernet_subsys(&iptable_raw_net_ops);
        return ret;
 }
 
 static void __exit iptable_raw_fini(void)
 {
        nf_unregister_hooks(ipt_ops, ARRAY_SIZE(ipt_ops));
-       ipt_unregister_table(packet_raw);
+       unregister_pernet_subsys(&iptable_raw_net_ops);
 }
 
 module_init(iptable_raw_init);