summaryrefslogtreecommitdiffstats
path: root/net/ipv6/netfilter/ip6table_mangle.c
diff options
context:
space:
mode:
authorTimothy Pearson <tpearson@raptorengineering.com>2017-08-23 14:45:25 -0500
committerTimothy Pearson <tpearson@raptorengineering.com>2017-08-23 14:45:25 -0500
commitfcbb27b0ec6dcbc5a5108cb8fb19eae64593d204 (patch)
tree22962a4387943edc841c72a4e636a068c66d58fd /net/ipv6/netfilter/ip6table_mangle.c
downloadast2050-linux-kernel-fcbb27b0ec6dcbc5a5108cb8fb19eae64593d204.zip
ast2050-linux-kernel-fcbb27b0ec6dcbc5a5108cb8fb19eae64593d204.tar.gz
Initial import of modified Linux 2.6.28 tree
Original upstream URL: git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable.git | branch linux-2.6.28.y
Diffstat (limited to 'net/ipv6/netfilter/ip6table_mangle.c')
-rw-r--r--net/ipv6/netfilter/ip6table_mangle.c221
1 files changed, 221 insertions, 0 deletions
diff --git a/net/ipv6/netfilter/ip6table_mangle.c b/net/ipv6/netfilter/ip6table_mangle.c
new file mode 100644
index 0000000..d0b31b2
--- /dev/null
+++ b/net/ipv6/netfilter/ip6table_mangle.c
@@ -0,0 +1,221 @@
+/*
+ * IPv6 packet mangling table, a port of the IPv4 mangle table to IPv6
+ *
+ * Copyright (C) 2000-2001 by Harald Welte <laforge@gnumonks.org>
+ * Copyright (C) 2000-2004 Netfilter Core Team <coreteam@netfilter.org>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#include <linux/module.h>
+#include <linux/netfilter_ipv6/ip6_tables.h>
+
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("Netfilter Core Team <coreteam@netfilter.org>");
+MODULE_DESCRIPTION("ip6tables mangle table");
+
+#define MANGLE_VALID_HOOKS ((1 << NF_INET_PRE_ROUTING) | \
+ (1 << NF_INET_LOCAL_IN) | \
+ (1 << NF_INET_FORWARD) | \
+ (1 << NF_INET_LOCAL_OUT) | \
+ (1 << NF_INET_POST_ROUTING))
+
+static struct
+{
+ struct ip6t_replace repl;
+ struct ip6t_standard entries[5];
+ struct ip6t_error term;
+} initial_table __net_initdata = {
+ .repl = {
+ .name = "mangle",
+ .valid_hooks = MANGLE_VALID_HOOKS,
+ .num_entries = 6,
+ .size = sizeof(struct ip6t_standard) * 5 + sizeof(struct ip6t_error),
+ .hook_entry = {
+ [NF_INET_PRE_ROUTING] = 0,
+ [NF_INET_LOCAL_IN] = sizeof(struct ip6t_standard),
+ [NF_INET_FORWARD] = sizeof(struct ip6t_standard) * 2,
+ [NF_INET_LOCAL_OUT] = sizeof(struct ip6t_standard) * 3,
+ [NF_INET_POST_ROUTING] = sizeof(struct ip6t_standard) * 4,
+ },
+ .underflow = {
+ [NF_INET_PRE_ROUTING] = 0,
+ [NF_INET_LOCAL_IN] = sizeof(struct ip6t_standard),
+ [NF_INET_FORWARD] = sizeof(struct ip6t_standard) * 2,
+ [NF_INET_LOCAL_OUT] = sizeof(struct ip6t_standard) * 3,
+ [NF_INET_POST_ROUTING] = sizeof(struct ip6t_standard) * 4,
+ },
+ },
+ .entries = {
+ IP6T_STANDARD_INIT(NF_ACCEPT), /* PRE_ROUTING */
+ IP6T_STANDARD_INIT(NF_ACCEPT), /* LOCAL_IN */
+ IP6T_STANDARD_INIT(NF_ACCEPT), /* FORWARD */
+ IP6T_STANDARD_INIT(NF_ACCEPT), /* LOCAL_OUT */
+ IP6T_STANDARD_INIT(NF_ACCEPT), /* POST_ROUTING */
+ },
+ .term = IP6T_ERROR_INIT, /* ERROR */
+};
+
+static struct xt_table packet_mangler = {
+ .name = "mangle",
+ .valid_hooks = MANGLE_VALID_HOOKS,
+ .lock = __RW_LOCK_UNLOCKED(packet_mangler.lock),
+ .me = THIS_MODULE,
+ .af = AF_INET6,
+};
+
+/* The work comes in here from netfilter.c. */
+static unsigned int
+ip6t_in_hook(unsigned int hook,
+ struct sk_buff *skb,
+ const struct net_device *in,
+ const struct net_device *out,
+ int (*okfn)(struct sk_buff *))
+{
+ return ip6t_do_table(skb, hook, in, out,
+ dev_net(in)->ipv6.ip6table_mangle);
+}
+
+static unsigned int
+ip6t_post_routing_hook(unsigned int hook,
+ struct sk_buff *skb,
+ const struct net_device *in,
+ const struct net_device *out,
+ int (*okfn)(struct sk_buff *))
+{
+ return ip6t_do_table(skb, hook, in, out,
+ dev_net(out)->ipv6.ip6table_mangle);
+}
+
+static unsigned int
+ip6t_local_out_hook(unsigned int hook,
+ struct sk_buff *skb,
+ const struct net_device *in,
+ const struct net_device *out,
+ int (*okfn)(struct sk_buff *))
+{
+
+ unsigned int ret;
+ struct in6_addr saddr, daddr;
+ u_int8_t hop_limit;
+ u_int32_t flowlabel, mark;
+
+#if 0
+ /* root is playing with raw sockets. */
+ if (skb->len < sizeof(struct iphdr)
+ || ip_hdrlen(skb) < sizeof(struct iphdr)) {
+ if (net_ratelimit())
+ printk("ip6t_hook: happy cracking.\n");
+ return NF_ACCEPT;
+ }
+#endif
+
+ /* save source/dest address, mark, hoplimit, flowlabel, priority, */
+ memcpy(&saddr, &ipv6_hdr(skb)->saddr, sizeof(saddr));
+ memcpy(&daddr, &ipv6_hdr(skb)->daddr, sizeof(daddr));
+ mark = skb->mark;
+ hop_limit = ipv6_hdr(skb)->hop_limit;
+
+ /* 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,
+ dev_net(out)->ipv6.ip6table_mangle);
+
+ if (ret != NF_DROP && ret != NF_STOLEN
+ && (memcmp(&ipv6_hdr(skb)->saddr, &saddr, sizeof(saddr))
+ || memcmp(&ipv6_hdr(skb)->daddr, &daddr, sizeof(daddr))
+ || skb->mark != mark
+ || ipv6_hdr(skb)->hop_limit != hop_limit))
+ return ip6_route_me_harder(skb) == 0 ? ret : NF_DROP;
+
+ return ret;
+}
+
+static struct nf_hook_ops ip6t_ops[] __read_mostly = {
+ {
+ .hook = ip6t_in_hook,
+ .owner = THIS_MODULE,
+ .pf = PF_INET6,
+ .hooknum = NF_INET_PRE_ROUTING,
+ .priority = NF_IP6_PRI_MANGLE,
+ },
+ {
+ .hook = ip6t_in_hook,
+ .owner = THIS_MODULE,
+ .pf = PF_INET6,
+ .hooknum = NF_INET_LOCAL_IN,
+ .priority = NF_IP6_PRI_MANGLE,
+ },
+ {
+ .hook = ip6t_in_hook,
+ .owner = THIS_MODULE,
+ .pf = PF_INET6,
+ .hooknum = NF_INET_FORWARD,
+ .priority = NF_IP6_PRI_MANGLE,
+ },
+ {
+ .hook = ip6t_local_out_hook,
+ .owner = THIS_MODULE,
+ .pf = PF_INET6,
+ .hooknum = NF_INET_LOCAL_OUT,
+ .priority = NF_IP6_PRI_MANGLE,
+ },
+ {
+ .hook = ip6t_post_routing_hook,
+ .owner = THIS_MODULE,
+ .pf = PF_INET6,
+ .hooknum = NF_INET_POST_ROUTING,
+ .priority = NF_IP6_PRI_MANGLE,
+ },
+};
+
+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;
+
+ 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));
+ if (ret < 0)
+ goto cleanup_table;
+
+ return ret;
+
+ cleanup_table:
+ 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));
+ unregister_pernet_subsys(&ip6table_mangle_net_ops);
+}
+
+module_init(ip6table_mangle_init);
+module_exit(ip6table_mangle_fini);
OpenPOWER on IntegriCloud