[NETFILTER]: x_tables: add connlimit match

ipt_connlimit has been sitting in POM-NG for a long time.
Here is a new shiny xt_connlimit with:

 * xtables'ified
 * will request the layer3 module
   (previously it hotdropped every packet when it was not loaded)
 * fixed: there was a deadlock in case of an OOM condition
 * support for any layer4 protocol (e.g. UDP/SCTP)
 * using jhash, as suggested by Eric Dumazet
 * ipv6 support

Signed-off-by: Jan Engelhardt <jengelh@gmx.de>
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

4 files changed, 338 insertions, 0 deletions

diff --git a/include/linux/netfilter/xt_connlimit.h b/include/linux/netfilter/xt_connlimit.h
new file mode 100644
index 000000000000..90ae8b474cb8
--- /dev/null
+++ b/include/linux/netfilter/xt_connlimit.h
@@ -0,0 +1,17 @@
+#ifndef _XT_CONNLIMIT_H
+#define _XT_CONNLIMIT_H
+
+struct xt_connlimit_data;
+
+struct xt_connlimit_info {
+        union {
+                u_int32_t v4_mask;
+                u_int32_t v6_mask[4];
+        };
+        unsigned int limit, inverse;
+
+        /* this needs to be at the end */
+        struct xt_connlimit_data *data __attribute__((aligned(8)));
+};
+
+#endif /* _XT_CONNLIMIT_H */
diff --git a/net/netfilter/Kconfig b/net/netfilter/Kconfig
index df5e8dab871d..9415b9a5dba6 100644
--- a/net/netfilter/Kconfig
+++ b/net/netfilter/Kconfig
@@ -423,6 +423,13 @@ config NETFILTER_XT_MATCH_CONNBYTES
           If you want to compile it as a module, say M here and read
           <file:Documentation/kbuild/modules.txt>.  If unsure, say `N'.
 
+config NETFILTER_XT_MATCH_CONNLIMIT
+        tristate '"connlimit" match support"'
+        depends on NETFILTER_XTABLES
+        ---help---
+          This match allows you to match against the number of parallel
+          connections to a server per client IP address (or address block).
+
 config NETFILTER_XT_MATCH_CONNMARK
         tristate  '"connmark" connection mark match support'
         depends on NETFILTER_XTABLES
diff --git a/net/netfilter/Makefile b/net/netfilter/Makefile
index 58b4245a1723..3e4a16aeb04e 100644
--- a/net/netfilter/Makefile
+++ b/net/netfilter/Makefile
@@ -52,6 +52,7 @@ obj-$(CONFIG_NETFILTER_XT_TARGET_CONNSECMARK) += xt_CONNSECMARK.o
 # matches
 obj-$(CONFIG_NETFILTER_XT_MATCH_COMMENT) += xt_comment.o
 obj-$(CONFIG_NETFILTER_XT_MATCH_CONNBYTES) += xt_connbytes.o
+obj-$(CONFIG_NETFILTER_XT_MATCH_CONNLIMIT) += xt_connlimit.o
 obj-$(CONFIG_NETFILTER_XT_MATCH_CONNMARK) += xt_connmark.o
 obj-$(CONFIG_NETFILTER_XT_MATCH_CONNTRACK) += xt_conntrack.o
 obj-$(CONFIG_NETFILTER_XT_MATCH_DCCP) += xt_dccp.o
diff --git a/net/netfilter/xt_connlimit.c b/net/netfilter/xt_connlimit.c
new file mode 100644
index 000000000000..3335dd5be962
--- /dev/null
+++ b/net/netfilter/xt_connlimit.c
@@ -0,0 +1,313 @@
+/*
+ * netfilter module to limit the number of parallel tcp
+ * connections per IP address.
+ *   (c) 2000 Gerd Knorr <kraxel@bytesex.org>
+ *   Nov 2002: Martin Bene <martin.bene@icomedias.com>:
+ *              only ignore TIME_WAIT or gone connections
+ *   Copyright © Jan Engelhardt <jengelh@gmx.de>, 2007
+ *
+ * based on ...
+ *
+ * Kernel module to match connection tracking information.
+ * GPL (C) 1999  Rusty Russell (rusty@rustcorp.com.au).
+ */
+#include <linux/in.h>
+#include <linux/in6.h>
+#include <linux/ip.h>
+#include <linux/ipv6.h>
+#include <linux/jhash.h>
+#include <linux/list.h>
+#include <linux/module.h>
+#include <linux/random.h>
+#include <linux/skbuff.h>
+#include <linux/spinlock.h>
+#include <linux/netfilter/nf_conntrack_tcp.h>
+#include <linux/netfilter/x_tables.h>
+#include <linux/netfilter/xt_connlimit.h>
+#include <net/netfilter/nf_conntrack.h>
+#include <net/netfilter/nf_conntrack_core.h>
+#include <net/netfilter/nf_conntrack_tuple.h>
+
+/* we will save the tuples of all connections we care about */
+struct xt_connlimit_conn {
+        struct list_head list;
+        struct nf_conntrack_tuple tuple;
+};
+
+struct xt_connlimit_data {
+        struct list_head iphash[256];
+        spinlock_t lock;
+};
+
+static u_int32_t connlimit_rnd;
+static bool connlimit_rnd_inited;
+
+static inline unsigned int connlimit_iphash(u_int32_t addr)
+{
+        if (unlikely(!connlimit_rnd_inited)) {
+                get_random_bytes(&connlimit_rnd, sizeof(connlimit_rnd));
+                connlimit_rnd_inited = true;
+        }
+        return jhash_1word(addr, connlimit_rnd) & 0xFF;
+}
+
+static inline unsigned int
+connlimit_iphash6(const union nf_conntrack_address *addr,
+                  const union nf_conntrack_address *mask)
+{
+        union nf_conntrack_address res;
+        unsigned int i;
+
+        if (unlikely(!connlimit_rnd_inited)) {
+                get_random_bytes(&connlimit_rnd, sizeof(connlimit_rnd));
+                connlimit_rnd_inited = true;
+        }
+
+        for (i = 0; i < ARRAY_SIZE(addr->ip6); ++i)
+                res.ip6[i] = addr->ip6[i] & mask->ip6[i];
+
+        return jhash2(res.ip6, ARRAY_SIZE(res.ip6), connlimit_rnd) & 0xFF;
+}
+
+static inline bool already_closed(const struct nf_conn *conn)
+{
+        u_int16_t proto = conn->tuplehash[0].tuple.dst.protonum;
+
+        if (proto == IPPROTO_TCP)
+                return conn->proto.tcp.state == TCP_CONNTRACK_TIME_WAIT;
+        else
+                return 0;
+}
+
+static inline unsigned int
+same_source_net(const union nf_conntrack_address *addr,
+                const union nf_conntrack_address *mask,
+                const union nf_conntrack_address *u3, unsigned int family)
+{
+        if (family == AF_INET) {
+                return (addr->ip & mask->ip) == (u3->ip & mask->ip);
+        } else {
+                union nf_conntrack_address lh, rh;
+                unsigned int i;
+
+                for (i = 0; i < ARRAY_SIZE(addr->ip6); ++i) {
+                        lh.ip6[i] = addr->ip6[i] & mask->ip6[i];
+                        rh.ip6[i] = u3->ip6[i] & mask->ip6[i];
+                }
+
+                return memcmp(&lh.ip6, &rh.ip6, sizeof(lh.ip6)) == 0;
+        }
+}
+
+static int count_them(struct xt_connlimit_data *data,
+                      const struct nf_conntrack_tuple *tuple,
+                      const union nf_conntrack_address *addr,
+                      const union nf_conntrack_address *mask,
+                      const struct xt_match *match)
+{
+        struct nf_conntrack_tuple_hash *found;
+        struct xt_connlimit_conn *conn;
+        struct xt_connlimit_conn *tmp;
+        struct nf_conn *found_ct;
+        struct list_head *hash;
+        bool addit = true;
+        int matches = 0;
+
+
+        if (match->family == AF_INET6)
+                hash = &data->iphash[connlimit_iphash6(addr, mask)];
+        else
+                hash = &data->iphash[connlimit_iphash(addr->ip & mask->ip)];
+
+        read_lock_bh(&nf_conntrack_lock);
+
+        /* check the saved connections */
+        list_for_each_entry_safe(conn, tmp, hash, list) {
+                found    = __nf_conntrack_find(&conn->tuple, NULL);
+                found_ct = NULL;
+
+                if (found != NULL)
+                        found_ct = nf_ct_tuplehash_to_ctrack(found);
+
+                if (found_ct != NULL &&
+                    nf_ct_tuple_equal(&conn->tuple, tuple) &&
+                    !already_closed(found_ct))
+                        /*
+                         * Just to be sure we have it only once in the list.
+                         * We should not see tuples twice unless someone hooks
+                         * this into a table without "-p tcp --syn".
+                         */
+                        addit = false;
+
+                if (found == NULL) {
+                        /* this one is gone */
+                        list_del(&conn->list);
+                        kfree(conn);
+                        continue;
+                }
+
+                if (already_closed(found_ct)) {
+                        /*
+                         * we do not care about connections which are
+                         * closed already -> ditch it
+                         */
+                        list_del(&conn->list);
+                        kfree(conn);
+                        continue;
+                }
+
+                if (same_source_net(addr, mask, &conn->tuple.src.u3,
+                    match->family))
+                        /* same source network -> be counted! */
+                        ++matches;
+        }
+
+        read_unlock_bh(&nf_conntrack_lock);
+
+        if (addit) {
+                /* save the new connection in our list */
+                conn = kzalloc(sizeof(*conn), GFP_ATOMIC);
+                if (conn == NULL)
+                        return -ENOMEM;
+                conn->tuple = *tuple;
+                list_add(&conn->list, hash);
+                ++matches;
+        }
+
+        return matches;
+}
+
+static bool connlimit_match(const struct sk_buff *skb,
+                            const struct net_device *in,
+                            const struct net_device *out,
+                            const struct xt_match *match,
+                            const void *matchinfo, int offset,
+                            unsigned int protoff, bool *hotdrop)
+{
+        const struct xt_connlimit_info *info = matchinfo;
+        union nf_conntrack_address addr, mask;
+        struct nf_conntrack_tuple tuple;
+        const struct nf_conntrack_tuple *tuple_ptr = &tuple;
+        enum ip_conntrack_info ctinfo;
+        const struct nf_conn *ct;
+        int connections;
+
+        ct = nf_ct_get(skb, &ctinfo);
+        if (ct != NULL)
+                tuple_ptr = &ct->tuplehash[0].tuple;
+        else if (!nf_ct_get_tuplepr(skb, skb_network_offset(skb),
+                                    match->family, &tuple))
+                goto hotdrop;
+
+        if (match->family == AF_INET6) {
+                const struct ipv6hdr *iph = ipv6_hdr(skb);
+                memcpy(&addr.ip6, &iph->saddr, sizeof(iph->saddr));
+                memcpy(&mask.ip6, info->v6_mask, sizeof(info->v6_mask));
+        } else {
+                const struct iphdr *iph = ip_hdr(skb);
+                addr.ip = iph->saddr;
+                mask.ip = info->v4_mask;
+        }
+
+        spin_lock_bh(&info->data->lock);
+        connections = count_them(info->data, tuple_ptr, &addr, &mask, match);
+        spin_unlock_bh(&info->data->lock);
+
+        if (connections < 0) {
+                /* kmalloc failed, drop it entirely */
+                *hotdrop = true;
+                return false;
+        }
+
+        return (connections > info->limit) ^ info->inverse;
+
+ hotdrop:
+        *hotdrop = true;
+        return false;
+}
+
+static bool connlimit_check(const char *tablename, const void *ip,
+                            const struct xt_match *match, void *matchinfo,
+                            unsigned int hook_mask)
+{
+        struct xt_connlimit_info *info = matchinfo;
+        unsigned int i;
+
+        if (nf_ct_l3proto_try_module_get(match->family) < 0) {
+                printk(KERN_WARNING "cannot load conntrack support for "
+                       "address family %u\n", match->family);
+                return false;
+        }
+
+        /* init private data */
+        info->data = kmalloc(sizeof(struct xt_connlimit_data), GFP_KERNEL);
+        if (info->data == NULL) {
+                nf_ct_l3proto_module_put(match->family);
+                return false;
+        }
+
+        spin_lock_init(&info->data->lock);
+        for (i = 0; i < ARRAY_SIZE(info->data->iphash); ++i)
+                INIT_LIST_HEAD(&info->data->iphash[i]);
+
+        return true;
+}
+
+static void connlimit_destroy(const struct xt_match *match, void *matchinfo)
+{
+        struct xt_connlimit_info *info = matchinfo;
+        struct xt_connlimit_conn *conn;
+        struct xt_connlimit_conn *tmp;
+        struct list_head *hash = info->data->iphash;
+        unsigned int i;
+
+        nf_ct_l3proto_module_put(match->family);
+
+        for (i = 0; i < ARRAY_SIZE(info->data->iphash); ++i) {
+                list_for_each_entry_safe(conn, tmp, &hash[i], list) {
+                        list_del(&conn->list);
+                        kfree(conn);
+                }
+        }
+
+        kfree(info->data);
+}
+
+static struct xt_match connlimit_reg[] __read_mostly = {
+        {
+                .name       = "connlimit",
+                .family     = AF_INET,
+                .checkentry = connlimit_check,
+                .match      = connlimit_match,
+                .matchsize  = sizeof(struct xt_connlimit_info),
+                .destroy    = connlimit_destroy,
+                .me         = THIS_MODULE,
+        },
+        {
+                .name       = "connlimit",
+                .family     = AF_INET6,
+                .checkentry = connlimit_check,
+                .match      = connlimit_match,
+                .matchsize  = sizeof(struct xt_connlimit_info),
+                .destroy    = connlimit_destroy,
+                .me         = THIS_MODULE,
+        },
+};
+
+static int __init xt_connlimit_init(void)
+{
+        return xt_register_matches(connlimit_reg, ARRAY_SIZE(connlimit_reg));
+}
+
+static void __exit xt_connlimit_exit(void)
+{
+        xt_unregister_matches(connlimit_reg, ARRAY_SIZE(connlimit_reg));
+}
+
+module_init(xt_connlimit_init);
+module_exit(xt_connlimit_exit);
+MODULE_AUTHOR("Jan Engelhardt <jengelh@gmx.de>");
+MODULE_DESCRIPTION("netfilter xt_connlimit match module");
+MODULE_LICENSE("GPL");
+MODULE_ALIAS("ipt_connlimit");
+MODULE_ALIAS("ip6t_connlimit");