netfilter: xtables: "set" match and "SET" target support
authorJozsef Kadlecsik <kadlec@blackhole.kfki.hu>
Tue, 1 Feb 2011 14:56:00 +0000 (15:56 +0100)
committerPatrick McHardy <kaber@trash.net>
Tue, 1 Feb 2011 14:56:00 +0000 (15:56 +0100)
The patch adds the combined module of the "SET" target and "set" match
to netfilter. Both the previous and the current revisions are supported.

Signed-off-by: Jozsef Kadlecsik <kadlec@blackhole.kfki.hu>
Signed-off-by: Patrick McHardy <kaber@trash.net>
include/linux/netfilter/xt_set.h [new file with mode: 0644]
net/netfilter/Kconfig
net/netfilter/Makefile
net/netfilter/xt_set.c [new file with mode: 0644]

diff --git a/include/linux/netfilter/xt_set.h b/include/linux/netfilter/xt_set.h
new file mode 100644 (file)
index 0000000..69b2bd1
--- /dev/null
@@ -0,0 +1,55 @@
+#ifndef _XT_SET_H
+#define _XT_SET_H
+
+#include <linux/netfilter/ipset/ip_set.h>
+
+/* Revision 0 interface: backward compatible with netfilter/iptables */
+
+/*
+ * Option flags for kernel operations (xt_set_info_v0)
+ */
+#define IPSET_SRC              0x01    /* Source match/add */
+#define IPSET_DST              0x02    /* Destination match/add */
+#define IPSET_MATCH_INV                0x04    /* Inverse matching */
+
+struct xt_set_info_v0 {
+       ip_set_id_t index;
+       union {
+               __u32 flags[IPSET_DIM_MAX + 1];
+               struct {
+                       __u32 __flags[IPSET_DIM_MAX];
+                       __u8 dim;
+                       __u8 flags;
+               } compat;
+       } u;
+};
+
+/* match and target infos */
+struct xt_set_info_match_v0 {
+       struct xt_set_info_v0 match_set;
+};
+
+struct xt_set_info_target_v0 {
+       struct xt_set_info_v0 add_set;
+       struct xt_set_info_v0 del_set;
+};
+
+/* Revision 1: current interface to netfilter/iptables */
+
+struct xt_set_info {
+       ip_set_id_t index;
+       __u8 dim;
+       __u8 flags;
+};
+
+/* match and target infos */
+struct xt_set_info_match {
+       struct xt_set_info match_set;
+};
+
+struct xt_set_info_target {
+       struct xt_set_info add_set;
+       struct xt_set_info del_set;
+};
+
+#endif /*_XT_SET_H*/
index 351abf8ace1392eba6a64eb33751233510ec052c..06fa9e4e45c7deb8a5bee99c7bb4fd7571cfbfcb 100644 (file)
@@ -352,6 +352,18 @@ config NETFILTER_XT_CONNMARK
        ctmark), similarly to the packet mark (nfmark). Using this
        target and match, you can set and match on this mark.
 
+config NETFILTER_XT_SET
+       tristate 'set target and match support'
+       depends on IP_SET
+       depends on NETFILTER_ADVANCED
+       help
+         This option adds the "SET" target and "set" match.
+
+         Using this target and match, you can add/delete and match
+         elements in the sets created by ipset(8).
+
+         To compile it as a module, choose M here.  If unsure, say N.
+
 # alphabetically ordered list of targets
 
 comment "Xtables targets"
index 510b586ccb7f41fe70e0ea349ed5ee2c5f43345d..1148643559cba0c418414d8b8177dac80983f93e 100644 (file)
@@ -46,6 +46,7 @@ obj-$(CONFIG_NETFILTER_XTABLES) += x_tables.o xt_tcpudp.o
 # combos
 obj-$(CONFIG_NETFILTER_XT_MARK) += xt_mark.o
 obj-$(CONFIG_NETFILTER_XT_CONNMARK) += xt_connmark.o
+obj-$(CONFIG_NETFILTER_XT_SET) += xt_set.o
 
 # targets
 obj-$(CONFIG_NETFILTER_XT_TARGET_AUDIT) += xt_AUDIT.o
diff --git a/net/netfilter/xt_set.c b/net/netfilter/xt_set.c
new file mode 100644 (file)
index 0000000..061d48c
--- /dev/null
@@ -0,0 +1,359 @@
+/* Copyright (C) 2000-2002 Joakim Axelsson <gozem@linux.nu>
+ *                         Patrick Schaaf <bof@bof.de>
+ *                         Martin Josefsson <gandalf@wlug.westbo.se>
+ * Copyright (C) 2003-2011 Jozsef Kadlecsik <kadlec@blackhole.kfki.hu>
+ *
+ * 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.
+ */
+
+/* Kernel module which implements the set match and SET target
+ * for netfilter/iptables. */
+
+#include <linux/module.h>
+#include <linux/skbuff.h>
+#include <linux/version.h>
+
+#include <linux/netfilter/x_tables.h>
+#include <linux/netfilter/xt_set.h>
+
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("Jozsef Kadlecsik <kadlec@blackhole.kfki.hu>");
+MODULE_DESCRIPTION("Xtables: IP set match and target module");
+MODULE_ALIAS("xt_SET");
+MODULE_ALIAS("ipt_set");
+MODULE_ALIAS("ip6t_set");
+MODULE_ALIAS("ipt_SET");
+MODULE_ALIAS("ip6t_SET");
+
+static inline int
+match_set(ip_set_id_t index, const struct sk_buff *skb,
+         u8 pf, u8 dim, u8 flags, int inv)
+{
+       if (ip_set_test(index, skb, pf, dim, flags))
+               inv = !inv;
+       return inv;
+}
+
+/* Revision 0 interface: backward compatible with netfilter/iptables */
+
+static bool
+set_match_v0(const struct sk_buff *skb, struct xt_action_param *par)
+{
+       const struct xt_set_info_match_v0 *info = par->matchinfo;
+
+       return match_set(info->match_set.index, skb, par->family,
+                        info->match_set.u.compat.dim,
+                        info->match_set.u.compat.flags,
+                        info->match_set.u.compat.flags & IPSET_INV_MATCH);
+}
+
+static void
+compat_flags(struct xt_set_info_v0 *info)
+{
+       u_int8_t i;
+
+       /* Fill out compatibility data according to enum ip_set_kopt */
+       info->u.compat.dim = IPSET_DIM_ZERO;
+       if (info->u.flags[0] & IPSET_MATCH_INV)
+               info->u.compat.flags |= IPSET_INV_MATCH;
+       for (i = 0; i < IPSET_DIM_MAX-1 && info->u.flags[i]; i++) {
+               info->u.compat.dim++;
+               if (info->u.flags[i] & IPSET_SRC)
+                       info->u.compat.flags |= (1<<info->u.compat.dim);
+       }
+}
+
+static int
+set_match_v0_checkentry(const struct xt_mtchk_param *par)
+{
+       struct xt_set_info_match_v0 *info = par->matchinfo;
+       ip_set_id_t index;
+
+       index = ip_set_nfnl_get_byindex(info->match_set.index);
+
+       if (index == IPSET_INVALID_ID) {
+               pr_warning("Cannot find set indentified by id %u to match\n",
+                          info->match_set.index);
+               return -ENOENT;
+       }
+       if (info->match_set.u.flags[IPSET_DIM_MAX-1] != 0) {
+               pr_warning("Protocol error: set match dimension "
+                          "is over the limit!\n");
+               return -ERANGE;
+       }
+
+       /* Fill out compatibility data */
+       compat_flags(&info->match_set);
+
+       return 0;
+}
+
+static void
+set_match_v0_destroy(const struct xt_mtdtor_param *par)
+{
+       struct xt_set_info_match_v0 *info = par->matchinfo;
+
+       ip_set_nfnl_put(info->match_set.index);
+}
+
+static unsigned int
+set_target_v0(struct sk_buff *skb, const struct xt_action_param *par)
+{
+       const struct xt_set_info_target_v0 *info = par->targinfo;
+
+       if (info->add_set.index != IPSET_INVALID_ID)
+               ip_set_add(info->add_set.index, skb, par->family,
+                          info->add_set.u.compat.dim,
+                          info->add_set.u.compat.flags);
+       if (info->del_set.index != IPSET_INVALID_ID)
+               ip_set_del(info->del_set.index, skb, par->family,
+                          info->del_set.u.compat.dim,
+                          info->del_set.u.compat.flags);
+
+       return XT_CONTINUE;
+}
+
+static int
+set_target_v0_checkentry(const struct xt_tgchk_param *par)
+{
+       struct xt_set_info_target_v0 *info = par->targinfo;
+       ip_set_id_t index;
+
+       if (info->add_set.index != IPSET_INVALID_ID) {
+               index = ip_set_nfnl_get_byindex(info->add_set.index);
+               if (index == IPSET_INVALID_ID) {
+                       pr_warning("Cannot find add_set index %u as target\n",
+                                  info->add_set.index);
+                       return -ENOENT;
+               }
+       }
+
+       if (info->del_set.index != IPSET_INVALID_ID) {
+               index = ip_set_nfnl_get_byindex(info->del_set.index);
+               if (index == IPSET_INVALID_ID) {
+                       pr_warning("Cannot find del_set index %u as target\n",
+                                  info->del_set.index);
+                       return -ENOENT;
+               }
+       }
+       if (info->add_set.u.flags[IPSET_DIM_MAX-1] != 0 ||
+           info->del_set.u.flags[IPSET_DIM_MAX-1] != 0) {
+               pr_warning("Protocol error: SET target dimension "
+                          "is over the limit!\n");
+               return -ERANGE;
+       }
+
+       /* Fill out compatibility data */
+       compat_flags(&info->add_set);
+       compat_flags(&info->del_set);
+
+       return 0;
+}
+
+static void
+set_target_v0_destroy(const struct xt_tgdtor_param *par)
+{
+       const struct xt_set_info_target_v0 *info = par->targinfo;
+
+       if (info->add_set.index != IPSET_INVALID_ID)
+               ip_set_nfnl_put(info->add_set.index);
+       if (info->del_set.index != IPSET_INVALID_ID)
+               ip_set_nfnl_put(info->del_set.index);
+}
+
+/* Revision 1: current interface to netfilter/iptables */
+
+static bool
+set_match(const struct sk_buff *skb, struct xt_action_param *par)
+{
+       const struct xt_set_info_match *info = par->matchinfo;
+
+       return match_set(info->match_set.index, skb, par->family,
+                        info->match_set.dim,
+                        info->match_set.flags,
+                        info->match_set.flags & IPSET_INV_MATCH);
+}
+
+static int
+set_match_checkentry(const struct xt_mtchk_param *par)
+{
+       struct xt_set_info_match *info = par->matchinfo;
+       ip_set_id_t index;
+
+       index = ip_set_nfnl_get_byindex(info->match_set.index);
+
+       if (index == IPSET_INVALID_ID) {
+               pr_warning("Cannot find set indentified by id %u to match\n",
+                          info->match_set.index);
+               return -ENOENT;
+       }
+       if (info->match_set.dim > IPSET_DIM_MAX) {
+               pr_warning("Protocol error: set match dimension "
+                          "is over the limit!\n");
+               return -ERANGE;
+       }
+
+       return 0;
+}
+
+static void
+set_match_destroy(const struct xt_mtdtor_param *par)
+{
+       struct xt_set_info_match *info = par->matchinfo;
+
+       ip_set_nfnl_put(info->match_set.index);
+}
+
+static unsigned int
+set_target(struct sk_buff *skb, const struct xt_action_param *par)
+{
+       const struct xt_set_info_target *info = par->targinfo;
+
+       if (info->add_set.index != IPSET_INVALID_ID)
+               ip_set_add(info->add_set.index,
+                          skb, par->family,
+                          info->add_set.dim,
+                          info->add_set.flags);
+       if (info->del_set.index != IPSET_INVALID_ID)
+               ip_set_del(info->del_set.index,
+                          skb, par->family,
+                          info->add_set.dim,
+                          info->del_set.flags);
+
+       return XT_CONTINUE;
+}
+
+static int
+set_target_checkentry(const struct xt_tgchk_param *par)
+{
+       const struct xt_set_info_target *info = par->targinfo;
+       ip_set_id_t index;
+
+       if (info->add_set.index != IPSET_INVALID_ID) {
+               index = ip_set_nfnl_get_byindex(info->add_set.index);
+               if (index == IPSET_INVALID_ID) {
+                       pr_warning("Cannot find add_set index %u as target\n",
+                                  info->add_set.index);
+                       return -ENOENT;
+               }
+       }
+
+       if (info->del_set.index != IPSET_INVALID_ID) {
+               index = ip_set_nfnl_get_byindex(info->del_set.index);
+               if (index == IPSET_INVALID_ID) {
+                       pr_warning("Cannot find del_set index %u as target\n",
+                                  info->del_set.index);
+                       return -ENOENT;
+               }
+       }
+       if (info->add_set.dim > IPSET_DIM_MAX ||
+           info->del_set.flags > IPSET_DIM_MAX) {
+               pr_warning("Protocol error: SET target dimension "
+                          "is over the limit!\n");
+               return -ERANGE;
+       }
+
+       return 0;
+}
+
+static void
+set_target_destroy(const struct xt_tgdtor_param *par)
+{
+       const struct xt_set_info_target *info = par->targinfo;
+
+       if (info->add_set.index != IPSET_INVALID_ID)
+               ip_set_nfnl_put(info->add_set.index);
+       if (info->del_set.index != IPSET_INVALID_ID)
+               ip_set_nfnl_put(info->del_set.index);
+}
+
+static struct xt_match set_matches[] __read_mostly = {
+       {
+               .name           = "set",
+               .family         = NFPROTO_IPV4,
+               .revision       = 0,
+               .match          = set_match_v0,
+               .matchsize      = sizeof(struct xt_set_info_match_v0),
+               .checkentry     = set_match_v0_checkentry,
+               .destroy        = set_match_v0_destroy,
+               .me             = THIS_MODULE
+       },
+       {
+               .name           = "set",
+               .family         = NFPROTO_IPV4,
+               .revision       = 1,
+               .match          = set_match,
+               .matchsize      = sizeof(struct xt_set_info_match),
+               .checkentry     = set_match_checkentry,
+               .destroy        = set_match_destroy,
+               .me             = THIS_MODULE
+       },
+       {
+               .name           = "set",
+               .family         = NFPROTO_IPV6,
+               .revision       = 1,
+               .match          = set_match,
+               .matchsize      = sizeof(struct xt_set_info_match),
+               .checkentry     = set_match_checkentry,
+               .destroy        = set_match_destroy,
+               .me             = THIS_MODULE
+       },
+};
+
+static struct xt_target set_targets[] __read_mostly = {
+       {
+               .name           = "SET",
+               .revision       = 0,
+               .family         = NFPROTO_IPV4,
+               .target         = set_target_v0,
+               .targetsize     = sizeof(struct xt_set_info_target_v0),
+               .checkentry     = set_target_v0_checkentry,
+               .destroy        = set_target_v0_destroy,
+               .me             = THIS_MODULE
+       },
+       {
+               .name           = "SET",
+               .revision       = 1,
+               .family         = NFPROTO_IPV4,
+               .target         = set_target,
+               .targetsize     = sizeof(struct xt_set_info_target),
+               .checkentry     = set_target_checkentry,
+               .destroy        = set_target_destroy,
+               .me             = THIS_MODULE
+       },
+       {
+               .name           = "SET",
+               .revision       = 1,
+               .family         = NFPROTO_IPV6,
+               .target         = set_target,
+               .targetsize     = sizeof(struct xt_set_info_target),
+               .checkentry     = set_target_checkentry,
+               .destroy        = set_target_destroy,
+               .me             = THIS_MODULE
+       },
+};
+
+static int __init xt_set_init(void)
+{
+       int ret = xt_register_matches(set_matches, ARRAY_SIZE(set_matches));
+
+       if (!ret) {
+               ret = xt_register_targets(set_targets,
+                                         ARRAY_SIZE(set_targets));
+               if (ret)
+                       xt_unregister_matches(set_matches,
+                                             ARRAY_SIZE(set_matches));
+       }
+       return ret;
+}
+
+static void __exit xt_set_fini(void)
+{
+       xt_unregister_matches(set_matches, ARRAY_SIZE(set_matches));
+       xt_unregister_targets(set_targets, ARRAY_SIZE(set_targets));
+}
+
+module_init(xt_set_init);
+module_exit(xt_set_fini);