This commit was manufactured by cvs2svn to create branch 'vserver'.
[linux-2.6.git] / net / ipv4 / netfilter / nf_nat_rule.c
diff --git a/net/ipv4/netfilter/nf_nat_rule.c b/net/ipv4/netfilter/nf_nat_rule.c
new file mode 100644 (file)
index 0000000..b868ee0
--- /dev/null
@@ -0,0 +1,343 @@
+/* (C) 1999-2001 Paul `Rusty' Russell
+ * (C) 2002-2006 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.
+ */
+
+/* Everything about the rules for NAT. */
+#include <linux/types.h>
+#include <linux/ip.h>
+#include <linux/netfilter.h>
+#include <linux/netfilter_ipv4.h>
+#include <linux/module.h>
+#include <linux/kmod.h>
+#include <linux/skbuff.h>
+#include <linux/proc_fs.h>
+#include <net/checksum.h>
+#include <net/route.h>
+#include <linux/bitops.h>
+
+#include <linux/netfilter_ipv4/ip_tables.h>
+#include <net/netfilter/nf_nat.h>
+#include <net/netfilter/nf_nat_core.h>
+#include <net/netfilter/nf_nat_rule.h>
+
+#if 0
+#define DEBUGP printk
+#else
+#define DEBUGP(format, args...)
+#endif
+
+#define NAT_VALID_HOOKS ((1<<NF_IP_PRE_ROUTING) | (1<<NF_IP_POST_ROUTING) | (1<<NF_IP_LOCAL_OUT))
+
+static struct
+{
+       struct ipt_replace repl;
+       struct ipt_standard entries[3];
+       struct ipt_error term;
+} nat_initial_table __initdata = {
+       .repl = {
+               .name = "nat",
+               .valid_hooks = NAT_VALID_HOOKS,
+               .num_entries = 4,
+               .size = sizeof(struct ipt_standard) * 3 + sizeof(struct ipt_error),
+               .hook_entry = {
+                       [NF_IP_PRE_ROUTING] = 0,
+                       [NF_IP_POST_ROUTING] = sizeof(struct ipt_standard),
+                       [NF_IP_LOCAL_OUT] = sizeof(struct ipt_standard) * 2 },
+               .underflow = {
+                       [NF_IP_PRE_ROUTING] = 0,
+                       [NF_IP_POST_ROUTING] = sizeof(struct ipt_standard),
+                       [NF_IP_LOCAL_OUT] = sizeof(struct ipt_standard) * 2 },
+       },
+       .entries = {
+               /* PRE_ROUTING */
+               {
+                       .entry = {
+                               .target_offset = sizeof(struct ipt_entry),
+                               .next_offset = sizeof(struct ipt_standard),
+                       },
+                       .target = {
+                               .target = {
+                                       .u = {
+                                               .target_size = IPT_ALIGN(sizeof(struct ipt_standard_target)),
+                                       },
+                               },
+                               .verdict = -NF_ACCEPT - 1,
+                       },
+               },
+               /* POST_ROUTING */
+               {
+                       .entry = {
+                               .target_offset = sizeof(struct ipt_entry),
+                               .next_offset = sizeof(struct ipt_standard),
+                       },
+                       .target = {
+                               .target = {
+                                       .u = {
+                                               .target_size = IPT_ALIGN(sizeof(struct ipt_standard_target)),
+                                       },
+                               },
+                               .verdict = -NF_ACCEPT - 1,
+                       },
+               },
+               /* LOCAL_OUT */
+               {
+                       .entry = {
+                               .target_offset = sizeof(struct ipt_entry),
+                               .next_offset = sizeof(struct ipt_standard),
+                       },
+                       .target = {
+                               .target = {
+                                       .u = {
+                                               .target_size = IPT_ALIGN(sizeof(struct ipt_standard_target)),
+                                       },
+                               },
+                               .verdict = -NF_ACCEPT - 1,
+                       },
+               },
+       },
+       /* ERROR */
+       .term = {
+               .entry = {
+                       .target_offset = sizeof(struct ipt_entry),
+                       .next_offset = sizeof(struct ipt_error),
+               },
+               .target = {
+                       .target = {
+                               .u = {
+                                       .user = {
+                                               .target_size = IPT_ALIGN(sizeof(struct ipt_error_target)),
+                                               .name = IPT_ERROR_TARGET,
+                                       },
+                               },
+                       },
+                       .errorname = "ERROR",
+               },
+       }
+};
+
+static struct ipt_table nat_table = {
+       .name           = "nat",
+       .valid_hooks    = NAT_VALID_HOOKS,
+       .lock           = RW_LOCK_UNLOCKED,
+       .me             = THIS_MODULE,
+       .af             = AF_INET,
+};
+
+/* Source NAT */
+static unsigned int ipt_snat_target(struct sk_buff **pskb,
+                                   const struct net_device *in,
+                                   const struct net_device *out,
+                                   unsigned int hooknum,
+                                   const struct xt_target *target,
+                                   const void *targinfo)
+{
+       struct nf_conn *ct;
+       enum ip_conntrack_info ctinfo;
+       const struct nf_nat_multi_range_compat *mr = targinfo;
+
+       NF_CT_ASSERT(hooknum == NF_IP_POST_ROUTING);
+
+       ct = nf_ct_get(*pskb, &ctinfo);
+
+       /* Connection must be valid and new. */
+       NF_CT_ASSERT(ct && (ctinfo == IP_CT_NEW || ctinfo == IP_CT_RELATED ||
+                           ctinfo == IP_CT_RELATED + IP_CT_IS_REPLY));
+       NF_CT_ASSERT(out);
+
+       return nf_nat_setup_info(ct, &mr->range[0], hooknum);
+}
+
+/* Before 2.6.11 we did implicit source NAT if required. Warn about change. */
+static void warn_if_extra_mangle(__be32 dstip, __be32 srcip)
+{
+       static int warned = 0;
+       struct flowi fl = { .nl_u = { .ip4_u = { .daddr = dstip } } };
+       struct rtable *rt;
+
+       if (ip_route_output_key(&rt, &fl) != 0)
+               return;
+
+       if (rt->rt_src != srcip && !warned) {
+               printk("NAT: no longer support implicit source local NAT\n");
+               printk("NAT: packet src %u.%u.%u.%u -> dst %u.%u.%u.%u\n",
+                      NIPQUAD(srcip), NIPQUAD(dstip));
+               warned = 1;
+       }
+       ip_rt_put(rt);
+}
+
+static unsigned int ipt_dnat_target(struct sk_buff **pskb,
+                                   const struct net_device *in,
+                                   const struct net_device *out,
+                                   unsigned int hooknum,
+                                   const struct xt_target *target,
+                                   const void *targinfo)
+{
+       struct nf_conn *ct;
+       enum ip_conntrack_info ctinfo;
+       const struct nf_nat_multi_range_compat *mr = targinfo;
+
+       NF_CT_ASSERT(hooknum == NF_IP_PRE_ROUTING ||
+                    hooknum == NF_IP_LOCAL_OUT);
+
+       ct = nf_ct_get(*pskb, &ctinfo);
+
+       /* Connection must be valid and new. */
+       NF_CT_ASSERT(ct && (ctinfo == IP_CT_NEW || ctinfo == IP_CT_RELATED));
+
+       if (hooknum == NF_IP_LOCAL_OUT &&
+           mr->range[0].flags & IP_NAT_RANGE_MAP_IPS)
+               warn_if_extra_mangle((*pskb)->nh.iph->daddr,
+                                    mr->range[0].min_ip);
+
+       return nf_nat_setup_info(ct, &mr->range[0], hooknum);
+}
+
+static int ipt_snat_checkentry(const char *tablename,
+                              const void *entry,
+                              const struct xt_target *target,
+                              void *targinfo,
+                              unsigned int hook_mask)
+{
+       struct nf_nat_multi_range_compat *mr = targinfo;
+
+       /* Must be a valid range */
+       if (mr->rangesize != 1) {
+               printk("SNAT: multiple ranges no longer supported\n");
+               return 0;
+       }
+       return 1;
+}
+
+static int ipt_dnat_checkentry(const char *tablename,
+                              const void *entry,
+                              const struct xt_target *target,
+                              void *targinfo,
+                              unsigned int hook_mask)
+{
+       struct nf_nat_multi_range_compat *mr = targinfo;
+
+       /* Must be a valid range */
+       if (mr->rangesize != 1) {
+               printk("DNAT: multiple ranges no longer supported\n");
+               return 0;
+       }
+       return 1;
+}
+
+inline unsigned int
+alloc_null_binding(struct nf_conn *ct,
+                  struct nf_nat_info *info,
+                  unsigned int hooknum)
+{
+       /* Force range to this IP; let proto decide mapping for
+          per-proto parts (hence not IP_NAT_RANGE_PROTO_SPECIFIED).
+          Use reply in case it's already been mangled (eg local packet).
+       */
+       __be32 ip
+               = (HOOK2MANIP(hooknum) == IP_NAT_MANIP_SRC
+                  ? ct->tuplehash[IP_CT_DIR_REPLY].tuple.dst.u3.ip
+                  : ct->tuplehash[IP_CT_DIR_REPLY].tuple.src.u3.ip);
+       struct nf_nat_range range
+               = { IP_NAT_RANGE_MAP_IPS, ip, ip, { 0 }, { 0 } };
+
+       DEBUGP("Allocating NULL binding for %p (%u.%u.%u.%u)\n",
+              ct, NIPQUAD(ip));
+       return nf_nat_setup_info(ct, &range, hooknum);
+}
+
+unsigned int
+alloc_null_binding_confirmed(struct nf_conn *ct,
+                             struct nf_nat_info *info,
+                             unsigned int hooknum)
+{
+       __be32 ip
+               = (HOOK2MANIP(hooknum) == IP_NAT_MANIP_SRC
+                  ? ct->tuplehash[IP_CT_DIR_REPLY].tuple.dst.u3.ip
+                  : ct->tuplehash[IP_CT_DIR_REPLY].tuple.src.u3.ip);
+       u_int16_t all
+               = (HOOK2MANIP(hooknum) == IP_NAT_MANIP_SRC
+                  ? ct->tuplehash[IP_CT_DIR_REPLY].tuple.dst.u.all
+                  : ct->tuplehash[IP_CT_DIR_REPLY].tuple.src.u.all);
+       struct nf_nat_range range
+               = { IP_NAT_RANGE_MAP_IPS, ip, ip, { all }, { all } };
+
+       DEBUGP("Allocating NULL binding for confirmed %p (%u.%u.%u.%u)\n",
+              ct, NIPQUAD(ip));
+       return nf_nat_setup_info(ct, &range, hooknum);
+}
+
+int nf_nat_rule_find(struct sk_buff **pskb,
+                    unsigned int hooknum,
+                    const struct net_device *in,
+                    const struct net_device *out,
+                    struct nf_conn *ct,
+                    struct nf_nat_info *info)
+{
+       int ret;
+
+       ret = ipt_do_table(pskb, hooknum, in, out, &nat_table);
+
+       if (ret == NF_ACCEPT) {
+               if (!nf_nat_initialized(ct, HOOK2MANIP(hooknum)))
+                       /* NUL mapping */
+                       ret = alloc_null_binding(ct, info, hooknum);
+       }
+       return ret;
+}
+
+static struct ipt_target ipt_snat_reg = {
+       .name           = "SNAT",
+       .target         = ipt_snat_target,
+       .targetsize     = sizeof(struct nf_nat_multi_range_compat),
+       .table          = "nat",
+       .hooks          = 1 << NF_IP_POST_ROUTING,
+       .checkentry     = ipt_snat_checkentry,
+       .family         = AF_INET,
+};
+
+static struct xt_target ipt_dnat_reg = {
+       .name           = "DNAT",
+       .target         = ipt_dnat_target,
+       .targetsize     = sizeof(struct nf_nat_multi_range_compat),
+       .table          = "nat",
+       .hooks          = (1 << NF_IP_PRE_ROUTING) | (1 << NF_IP_LOCAL_OUT),
+       .checkentry     = ipt_dnat_checkentry,
+       .family         = AF_INET,
+};
+
+int __init nf_nat_rule_init(void)
+{
+       int ret;
+
+       ret = ipt_register_table(&nat_table, &nat_initial_table.repl);
+       if (ret != 0)
+               return ret;
+       ret = xt_register_target(&ipt_snat_reg);
+       if (ret != 0)
+               goto unregister_table;
+
+       ret = xt_register_target(&ipt_dnat_reg);
+       if (ret != 0)
+               goto unregister_snat;
+
+       return ret;
+
+ unregister_snat:
+       xt_unregister_target(&ipt_snat_reg);
+ unregister_table:
+       ipt_unregister_table(&nat_table);
+
+       return ret;
+}
+
+void nf_nat_rule_cleanup(void)
+{
+       xt_unregister_target(&ipt_dnat_reg);
+       xt_unregister_target(&ipt_snat_reg);
+       ipt_unregister_table(&nat_table);
+}