提交 9335f047 编写于 作者: A Alexey Dobriyan 提交者: David S. Miller

[NETFILTER]: ip_tables: per-netns FILTER, MANGLE, RAW

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: NAlexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
上级 34bd137b
......@@ -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
......@@ -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);
......
......@@ -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);
......
......@@ -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);
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册