]> git.karo-electronics.de Git - karo-tx-linux.git/commitdiff
netfilter: ipset: hash:net,iface type introduced
authorJozsef Kadlecsik <kadlec@blackhole.kfki.hu>
Thu, 16 Jun 2011 17:00:48 +0000 (19:00 +0200)
committerPatrick McHardy <kaber@trash.net>
Thu, 16 Jun 2011 17:00:48 +0000 (19:00 +0200)
The hash:net,iface type makes possible to store network address and
interface name pairs in a set. It's mostly suitable for egress
and ingress filtering. Examples:

        # ipset create test hash:net,iface
        # ipset add test 192.168.0.0/16,eth0
        # ipset add test 192.168.0.0/24,eth1

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

index 8955165f7dbb2feb1819d2ba3b167131e531be1e..e40917334d0472833c8846b19000e8b2866f1529 100644 (file)
@@ -105,6 +105,7 @@ enum {
        IPSET_ATTR_IP2,
        IPSET_ATTR_CIDR2,
        IPSET_ATTR_IP2_TO,
+       IPSET_ATTR_IFACE,
        __IPSET_ATTR_ADT_MAX,
 };
 #define IPSET_ATTR_ADT_MAX     (__IPSET_ATTR_ADT_MAX - 1)
@@ -153,6 +154,8 @@ enum ipset_cmd_flags {
 enum ipset_cadt_flags {
        IPSET_FLAG_BIT_BEFORE   = 0,
        IPSET_FLAG_BEFORE       = (1 << IPSET_FLAG_BIT_BEFORE),
+       IPSET_FLAG_BIT_PHYSDEV  = 1,
+       IPSET_FLAG_PHYSDEV      = (1 << IPSET_FLAG_BIT_PHYSDEV),
 };
 
 /* Commands with settype-specific attributes */
@@ -212,6 +215,8 @@ enum ip_set_feature {
        IPSET_TYPE_IP2 = (1 << IPSET_TYPE_IP2_FLAG),
        IPSET_TYPE_NAME_FLAG = 4,
        IPSET_TYPE_NAME = (1 << IPSET_TYPE_NAME_FLAG),
+       IPSET_TYPE_IFACE_FLAG = 5,
+       IPSET_TYPE_IFACE = (1 << IPSET_TYPE_IFACE_FLAG),
        /* Strictly speaking not a feature, but a flag for dumping:
         * this settype must be dumped last */
        IPSET_DUMP_LAST_FLAG = 7,
index 1c977e143885712f726a45dc500394c73d970dc3..8a0999a8baa11e1cee3f7d9173dc7ea5d799c823 100644 (file)
@@ -63,6 +63,9 @@ struct ip_set_hash {
 #ifdef IP_SET_HASH_WITH_NETMASK
        u8 netmask;             /* netmask value for subnets to store */
 #endif
+#ifdef IP_SET_HASH_WITH_RBTREE
+       struct rb_root rbtree;
+#endif
 #ifdef IP_SET_HASH_WITH_NETS
        struct ip_set_hash_nets nets[0]; /* book-keeping of prefixes */
 #endif
@@ -200,6 +203,9 @@ ip_set_hash_destroy(struct ip_set *set)
                del_timer_sync(&h->gc);
 
        ahash_destroy(h->table);
+#ifdef IP_SET_HASH_WITH_RBTREE
+       rbtree_destroy(&h->rbtree);
+#endif
        kfree(h);
 
        set->data = NULL;
index 2c5b348eb3a8a038726d4a7f689926ca89aad439..ba36c283d8372dfb3c890994c57c6bb0d46266b0 100644 (file)
@@ -109,6 +109,16 @@ config IP_SET_HASH_NETPORT
 
          To compile it as a module, choose M here.  If unsure, say N.
 
+config IP_SET_HASH_NETIFACE
+       tristate "hash:net,iface set support"
+       depends on IP_SET
+       help
+         This option adds the hash:net,iface set type support, by which
+         one can store IPv4/IPv6 network address/prefix and
+         interface name pairs as elements in a set.
+
+         To compile it as a module, choose M here.  If unsure, say N.
+
 config IP_SET_LIST_SET
        tristate "list:set set support"
        depends on IP_SET
index 5adbdab67bd2ff24374cdafda05905a788048f43..6e965ecd5444696b4dc7e09bb98b1d09420ce9ab 100644 (file)
@@ -19,6 +19,7 @@ obj-$(CONFIG_IP_SET_HASH_IPPORTIP) += ip_set_hash_ipportip.o
 obj-$(CONFIG_IP_SET_HASH_IPPORTNET) += ip_set_hash_ipportnet.o
 obj-$(CONFIG_IP_SET_HASH_NET) += ip_set_hash_net.o
 obj-$(CONFIG_IP_SET_HASH_NETPORT) += ip_set_hash_netport.o
+obj-$(CONFIG_IP_SET_HASH_NETIFACE) += ip_set_hash_netiface.o
 
 # list types
 obj-$(CONFIG_IP_SET_LIST_SET) += ip_set_list_set.o
diff --git a/net/netfilter/ipset/ip_set_hash_netiface.c b/net/netfilter/ipset/ip_set_hash_netiface.c
new file mode 100644 (file)
index 0000000..51e5df1
--- /dev/null
@@ -0,0 +1,762 @@
+/* Copyright (C) 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 implementing an IP set type: the hash:net,iface type */
+
+#include <linux/jhash.h>
+#include <linux/module.h>
+#include <linux/ip.h>
+#include <linux/skbuff.h>
+#include <linux/errno.h>
+#include <linux/random.h>
+#include <linux/rbtree.h>
+#include <net/ip.h>
+#include <net/ipv6.h>
+#include <net/netlink.h>
+
+#include <linux/netfilter.h>
+#include <linux/netfilter/ipset/pfxlen.h>
+#include <linux/netfilter/ipset/ip_set.h>
+#include <linux/netfilter/ipset/ip_set_timeout.h>
+#include <linux/netfilter/ipset/ip_set_hash.h>
+
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("Jozsef Kadlecsik <kadlec@blackhole.kfki.hu>");
+MODULE_DESCRIPTION("hash:net,iface type of IP sets");
+MODULE_ALIAS("ip_set_hash:net,iface");
+
+/* Interface name rbtree */
+
+struct iface_node {
+       struct rb_node node;
+       char iface[IFNAMSIZ];
+};
+
+#define iface_data(n)  (rb_entry(n, struct iface_node, node)->iface)
+
+static inline long
+ifname_compare(const char *_a, const char *_b)
+{
+       const long *a = (const long *)_a;
+       const long *b = (const long *)_b;
+
+       BUILD_BUG_ON(IFNAMSIZ > 4 * sizeof(unsigned long));
+       if (a[0] != b[0])
+               return a[0] - b[0];
+       if (IFNAMSIZ > sizeof(long)) {
+               if (a[1] != b[1])
+                       return a[1] - b[1];
+       }
+       if (IFNAMSIZ > 2 * sizeof(long)) {
+               if (a[2] != b[2])
+                       return a[2] - b[2];
+       }
+       if (IFNAMSIZ > 3 * sizeof(long)) {
+               if (a[3] != b[3])
+                       return a[3] - b[3];
+       }
+       return 0;
+}
+
+static void
+rbtree_destroy(struct rb_root *root)
+{
+       struct rb_node *p, *n = root->rb_node;
+       struct iface_node *node;
+
+       /* Non-recursive destroy, like in ext3 */
+       while (n) {
+               if (n->rb_left) {
+                       n = n->rb_left;
+                       continue;
+               }
+               if (n->rb_right) {
+                       n = n->rb_right;
+                       continue;
+               }
+               p = rb_parent(n);
+               node = rb_entry(n, struct iface_node, node);
+               if (!p)
+                       *root = RB_ROOT;
+               else if (p->rb_left == n)
+                       p->rb_left = NULL;
+               else if (p->rb_right == n)
+                       p->rb_right = NULL;
+
+               kfree(node);
+               n = p;
+       }
+}
+
+static int
+iface_test(struct rb_root *root, const char **iface)
+{
+       struct rb_node *n = root->rb_node;
+
+       while (n) {
+               const char *d = iface_data(n);
+               int res = ifname_compare(*iface, d);
+               
+               if (res < 0)
+                       n = n->rb_left;
+               else if (res > 0)
+                       n = n->rb_right;
+               else {
+                       *iface = d;
+                       return 1;
+               }
+       }
+       return 0;
+}
+
+static int
+iface_add(struct rb_root *root, const char **iface)
+{
+       struct rb_node **n = &(root->rb_node), *p = NULL;
+       struct iface_node *d;
+       
+       while (*n) {
+               char *ifname = iface_data(*n);
+               int res = ifname_compare(*iface, ifname);
+
+               p = *n;
+               if (res < 0)
+                       n = &((*n)->rb_left);
+               else if (res > 0)
+                       n = &((*n)->rb_right);
+               else {
+                       *iface = ifname;
+                       return 0;
+               }
+       }
+
+       d = kzalloc(sizeof(*d), GFP_ATOMIC);
+       if (!d)
+               return -ENOMEM;
+       strcpy(d->iface, *iface);
+
+       rb_link_node(&d->node, p, n);
+       rb_insert_color(&d->node, root);
+
+       *iface = d->iface;
+       return 0;
+}
+
+/* Type specific function prefix */
+#define TYPE           hash_netiface
+
+static bool
+hash_netiface_same_set(const struct ip_set *a, const struct ip_set *b);
+
+#define hash_netiface4_same_set        hash_netiface_same_set
+#define hash_netiface6_same_set        hash_netiface_same_set
+
+#define STREQ(a, b)    (strcmp(a, b) == 0)
+
+/* The type variant functions: IPv4 */
+
+/* Member elements without timeout */
+struct hash_netiface4_elem {
+       __be32 ip;
+       const char *iface;
+       u8 physdev;
+       u8 cidr;
+       u16 padding;
+};
+
+/* Member elements with timeout support */
+struct hash_netiface4_telem {
+       __be32 ip;
+       const char *iface;
+       u8 physdev;
+       u8 cidr;
+       u16 padding;
+       unsigned long timeout;
+};
+
+static inline bool
+hash_netiface4_data_equal(const struct hash_netiface4_elem *ip1,
+                         const struct hash_netiface4_elem *ip2)
+{
+       return ip1->ip == ip2->ip &&
+              ip1->cidr == ip2->cidr &&
+              ip1->physdev == ip2->physdev &&
+              ip1->iface == ip2->iface;
+}
+
+static inline bool
+hash_netiface4_data_isnull(const struct hash_netiface4_elem *elem)
+{
+       return elem->cidr == 0;
+}
+
+static inline void
+hash_netiface4_data_copy(struct hash_netiface4_elem *dst,
+                        const struct hash_netiface4_elem *src) {
+       dst->ip = src->ip;
+       dst->cidr = src->cidr;
+       dst->physdev = src->physdev;
+       dst->iface = src->iface;
+}
+
+static inline void
+hash_netiface4_data_netmask(struct hash_netiface4_elem *elem, u8 cidr)
+{
+       elem->ip &= ip_set_netmask(cidr);
+       elem->cidr = cidr;
+}
+
+static inline void
+hash_netiface4_data_zero_out(struct hash_netiface4_elem *elem)
+{
+       elem->cidr = 0;
+}
+
+static bool
+hash_netiface4_data_list(struct sk_buff *skb,
+                        const struct hash_netiface4_elem *data)
+{
+       u32 flags = data->physdev ? IPSET_FLAG_PHYSDEV : 0;
+
+       NLA_PUT_IPADDR4(skb, IPSET_ATTR_IP, data->ip);
+       NLA_PUT_U8(skb, IPSET_ATTR_CIDR, data->cidr);
+       NLA_PUT_STRING(skb, IPSET_ATTR_IFACE, data->iface);
+       if (flags)
+               NLA_PUT_NET32(skb, IPSET_ATTR_CADT_FLAGS, flags);
+       return 0;
+
+nla_put_failure:
+       return 1;
+}
+
+static bool
+hash_netiface4_data_tlist(struct sk_buff *skb,
+                         const struct hash_netiface4_elem *data)
+{
+       const struct hash_netiface4_telem *tdata =
+               (const struct hash_netiface4_telem *)data;
+       u32 flags = data->physdev ? IPSET_FLAG_PHYSDEV : 0;
+
+       NLA_PUT_IPADDR4(skb, IPSET_ATTR_IP, data->ip);
+       NLA_PUT_U8(skb, IPSET_ATTR_CIDR, data->cidr);
+       NLA_PUT_STRING(skb, IPSET_ATTR_IFACE, data->iface);
+       if (flags)
+               NLA_PUT_NET32(skb, IPSET_ATTR_CADT_FLAGS, flags);
+       NLA_PUT_NET32(skb, IPSET_ATTR_TIMEOUT,
+                     htonl(ip_set_timeout_get(tdata->timeout)));
+
+       return 0;
+
+nla_put_failure:
+       return 1;
+}
+
+#define IP_SET_HASH_WITH_NETS
+#define IP_SET_HASH_WITH_RBTREE
+
+#define PF             4
+#define HOST_MASK      32
+#include <linux/netfilter/ipset/ip_set_ahash.h>
+
+static inline void
+hash_netiface4_data_next(struct ip_set_hash *h,
+                        const struct hash_netiface4_elem *d)
+{
+       h->next.ip = ntohl(d->ip);
+}
+
+static int
+hash_netiface4_kadt(struct ip_set *set, const struct sk_buff *skb,
+                   const struct xt_action_param *par,
+                   enum ipset_adt adt, const struct ip_set_adt_opt *opt)
+{
+       struct ip_set_hash *h = set->data;
+       ipset_adtfn adtfn = set->variant->adt[adt];
+       struct hash_netiface4_elem data = {
+               .cidr = h->nets[0].cidr ? h->nets[0].cidr : HOST_MASK
+       };
+       int ret;
+
+       if (data.cidr == 0)
+               return -EINVAL;
+       if (adt == IPSET_TEST)
+               data.cidr = HOST_MASK;
+
+       ip4addrptr(skb, opt->flags & IPSET_DIM_ONE_SRC, &data.ip);
+       data.ip &= ip_set_netmask(data.cidr);
+
+#define IFACE(dir)     (par->dir ? par->dir->name : NULL)
+#define PHYSDEV(dir)   (nf_bridge->dir ? nf_bridge->dir->name : NULL)
+#define SRCDIR         (opt->flags & IPSET_DIM_TWO_SRC)
+
+       if (opt->cmdflags & IPSET_FLAG_PHYSDEV) {
+#ifdef CONFIG_BRIDGE_NETFILTER
+               const struct nf_bridge_info *nf_bridge = skb->nf_bridge;
+               
+               if (!nf_bridge)
+                       return -EINVAL;
+               data.iface = SRCDIR ? PHYSDEV(physindev): PHYSDEV(physoutdev);
+               data.physdev = 1;
+#else
+               data.iface = NULL;
+#endif
+       } else
+               data.iface = SRCDIR ? IFACE(in) : IFACE(out);
+
+       if (!data.iface)
+               return -EINVAL;
+       ret = iface_test(&h->rbtree, &data.iface);
+       if (adt == IPSET_ADD) {
+               if (!ret) {
+                       ret = iface_add(&h->rbtree, &data.iface);
+                       if (ret)
+                               return ret;
+               }
+       } else if (!ret)
+               return ret;
+
+       return adtfn(set, &data, opt_timeout(opt, h), opt->cmdflags);
+}
+
+static int
+hash_netiface4_uadt(struct ip_set *set, struct nlattr *tb[],
+                   enum ipset_adt adt, u32 *lineno, u32 flags, bool retried)
+{
+       struct ip_set_hash *h = set->data;
+       ipset_adtfn adtfn = set->variant->adt[adt];
+       struct hash_netiface4_elem data = { .cidr = HOST_MASK };
+       u32 ip = 0, ip_to, last;
+       u32 timeout = h->timeout;
+       char iface[IFNAMSIZ] = {};
+       int ret;
+
+       if (unlikely(!tb[IPSET_ATTR_IP] ||
+                    !tb[IPSET_ATTR_IFACE] ||
+                    !ip_set_optattr_netorder(tb, IPSET_ATTR_TIMEOUT) ||
+                    !ip_set_optattr_netorder(tb, IPSET_ATTR_CADT_FLAGS)))
+               return -IPSET_ERR_PROTOCOL;
+
+       if (tb[IPSET_ATTR_LINENO])
+               *lineno = nla_get_u32(tb[IPSET_ATTR_LINENO]);
+
+       ret = ip_set_get_hostipaddr4(tb[IPSET_ATTR_IP], &ip);
+       if (ret)
+               return ret;
+
+       if (tb[IPSET_ATTR_CIDR]) {
+               data.cidr = nla_get_u8(tb[IPSET_ATTR_CIDR]);
+               if (!data.cidr)
+                       return -IPSET_ERR_INVALID_CIDR; 
+       }
+
+       if (tb[IPSET_ATTR_TIMEOUT]) {
+               if (!with_timeout(h->timeout))
+                       return -IPSET_ERR_TIMEOUT;
+               timeout = ip_set_timeout_uget(tb[IPSET_ATTR_TIMEOUT]);
+       }
+
+       strcpy(iface, nla_data(tb[IPSET_ATTR_IFACE])); 
+       data.iface = iface;
+       ret = iface_test(&h->rbtree, &data.iface);
+       if (adt == IPSET_ADD) {
+               if (!ret) {
+                       ret = iface_add(&h->rbtree, &data.iface);
+                       if (ret)
+                               return ret;
+               }
+       } else if (!ret)
+               return ret;
+
+       if (tb[IPSET_ATTR_CADT_FLAGS]) {
+               u32 flags = ip_set_get_h32(tb[IPSET_ATTR_CADT_FLAGS]);
+               if (flags & IPSET_FLAG_PHYSDEV)
+                       data.physdev = 1;
+       }
+
+       if (adt == IPSET_TEST || !tb[IPSET_ATTR_IP_TO]) {
+               data.ip = htonl(ip & ip_set_hostmask(data.cidr));
+               ret = adtfn(set, &data, timeout, flags);
+               return ip_set_eexist(ret, flags) ? 0 : ret;
+       }
+
+       if (tb[IPSET_ATTR_IP_TO]) {
+               ret = ip_set_get_hostipaddr4(tb[IPSET_ATTR_IP_TO], &ip_to);
+               if (ret)
+                       return ret;
+               if (ip_to < ip)
+                       swap(ip, ip_to);
+               if (ip + UINT_MAX == ip_to)
+                       return -IPSET_ERR_HASH_RANGE;
+       } else {
+               ip_set_mask_from_to(ip, ip_to, data.cidr);
+       }
+
+       if (retried)
+               ip = h->next.ip;
+       while (!after(ip, ip_to)) {
+               data.ip = htonl(ip);
+               last = ip_set_range_to_cidr(ip, ip_to, &data.cidr);
+               ret = adtfn(set, &data, timeout, flags);
+
+               if (ret && !ip_set_eexist(ret, flags))
+                       return ret;
+               else
+                       ret = 0;
+               ip = last + 1;
+       }
+       return ret;
+}
+
+static bool
+hash_netiface_same_set(const struct ip_set *a, const struct ip_set *b)
+{
+       const struct ip_set_hash *x = a->data;
+       const struct ip_set_hash *y = b->data;
+
+       /* Resizing changes htable_bits, so we ignore it */
+       return x->maxelem == y->maxelem &&
+              x->timeout == y->timeout;
+}
+
+/* The type variant functions: IPv6 */
+
+struct hash_netiface6_elem {
+       union nf_inet_addr ip;
+       const char *iface;
+       u8 physdev;
+       u8 cidr;
+       u16 padding;
+};
+
+struct hash_netiface6_telem {
+       union nf_inet_addr ip;
+       const char *iface;
+       u8 physdev;
+       u8 cidr;
+       u16 padding;
+       unsigned long timeout;
+};
+
+static inline bool
+hash_netiface6_data_equal(const struct hash_netiface6_elem *ip1,
+                         const struct hash_netiface6_elem *ip2)
+{
+       return ipv6_addr_cmp(&ip1->ip.in6, &ip2->ip.in6) == 0 &&
+              ip1->cidr == ip2->cidr &&
+              ip1->physdev == ip2->physdev &&
+              ip1->iface == ip2->iface;
+}
+
+static inline bool
+hash_netiface6_data_isnull(const struct hash_netiface6_elem *elem)
+{
+       return elem->cidr == 0;
+}
+
+static inline void
+hash_netiface6_data_copy(struct hash_netiface6_elem *dst,
+                        const struct hash_netiface6_elem *src)
+{
+       memcpy(dst, src, sizeof(*dst));
+}
+
+static inline void
+hash_netiface6_data_zero_out(struct hash_netiface6_elem *elem)
+{
+}
+
+static inline void
+ip6_netmask(union nf_inet_addr *ip, u8 prefix)
+{
+       ip->ip6[0] &= ip_set_netmask6(prefix)[0];
+       ip->ip6[1] &= ip_set_netmask6(prefix)[1];
+       ip->ip6[2] &= ip_set_netmask6(prefix)[2];
+       ip->ip6[3] &= ip_set_netmask6(prefix)[3];
+}
+
+static inline void
+hash_netiface6_data_netmask(struct hash_netiface6_elem *elem, u8 cidr)
+{
+       ip6_netmask(&elem->ip, cidr);
+       elem->cidr = cidr;
+}
+
+static bool
+hash_netiface6_data_list(struct sk_buff *skb,
+                        const struct hash_netiface6_elem *data)
+{
+       u32 flags = data->physdev ? IPSET_FLAG_PHYSDEV : 0;
+
+       NLA_PUT_IPADDR6(skb, IPSET_ATTR_IP, &data->ip);
+       NLA_PUT_U8(skb, IPSET_ATTR_CIDR, data->cidr);
+       NLA_PUT_STRING(skb, IPSET_ATTR_IFACE, data->iface);
+       if (flags)
+               NLA_PUT_NET32(skb, IPSET_ATTR_CADT_FLAGS, flags);
+       return 0;
+
+nla_put_failure:
+       return 1;
+}
+
+static bool
+hash_netiface6_data_tlist(struct sk_buff *skb,
+                         const struct hash_netiface6_elem *data)
+{
+       const struct hash_netiface6_telem *e =
+               (const struct hash_netiface6_telem *)data;
+       u32 flags = data->physdev ? IPSET_FLAG_PHYSDEV : 0;
+
+       NLA_PUT_IPADDR6(skb, IPSET_ATTR_IP, &e->ip);
+       NLA_PUT_U8(skb, IPSET_ATTR_CIDR, data->cidr);
+       NLA_PUT_STRING(skb, IPSET_ATTR_IFACE, data->iface);
+       if (flags)
+               NLA_PUT_NET32(skb, IPSET_ATTR_CADT_FLAGS, flags);
+       NLA_PUT_NET32(skb, IPSET_ATTR_TIMEOUT,
+                     htonl(ip_set_timeout_get(e->timeout)));
+       return 0;
+
+nla_put_failure:
+       return 1;
+}
+
+#undef PF
+#undef HOST_MASK
+
+#define PF             6
+#define HOST_MASK      128
+#include <linux/netfilter/ipset/ip_set_ahash.h>
+
+static inline void
+hash_netiface6_data_next(struct ip_set_hash *h,
+                        const struct hash_netiface6_elem *d)
+{
+}
+
+static int
+hash_netiface6_kadt(struct ip_set *set, const struct sk_buff *skb,
+                   const struct xt_action_param *par,
+                   enum ipset_adt adt, const struct ip_set_adt_opt *opt)
+{
+       struct ip_set_hash *h = set->data;
+       ipset_adtfn adtfn = set->variant->adt[adt];
+       struct hash_netiface6_elem data = {
+               .cidr = h->nets[0].cidr ? h->nets[0].cidr : HOST_MASK
+       };
+       int ret;
+
+       if (data.cidr == 0)
+               return -EINVAL;
+       if (adt == IPSET_TEST)
+               data.cidr = HOST_MASK;
+
+       ip6addrptr(skb, opt->flags & IPSET_DIM_ONE_SRC, &data.ip.in6);
+       ip6_netmask(&data.ip, data.cidr);
+
+       if (opt->cmdflags & IPSET_FLAG_PHYSDEV) {
+#ifdef CONFIG_BRIDGE_NETFILTER
+               const struct nf_bridge_info *nf_bridge = skb->nf_bridge;
+               
+               if (!nf_bridge)
+                       return -EINVAL;
+               data.iface = SRCDIR ? PHYSDEV(physindev): PHYSDEV(physoutdev);
+               data.physdev = 1;
+#else
+               data.iface = NULL;
+#endif
+       } else
+               data.iface = SRCDIR ? IFACE(in) : IFACE(out);
+
+       if (!data.iface)
+               return -EINVAL;
+       ret = iface_test(&h->rbtree, &data.iface);
+       if (adt == IPSET_ADD) {
+               if (!ret) {
+                       ret = iface_add(&h->rbtree, &data.iface);
+                       if (ret)
+                               return ret;
+               }
+       } else if (!ret)
+               return ret;
+
+       return adtfn(set, &data, opt_timeout(opt, h), opt->cmdflags);
+}
+
+static int
+hash_netiface6_uadt(struct ip_set *set, struct nlattr *tb[],
+                  enum ipset_adt adt, u32 *lineno, u32 flags, bool retried)
+{
+       struct ip_set_hash *h = set->data;
+       ipset_adtfn adtfn = set->variant->adt[adt];
+       struct hash_netiface6_elem data = { .cidr = HOST_MASK };
+       u32 timeout = h->timeout;
+       char iface[IFNAMSIZ] = {};
+       int ret;
+
+       if (unlikely(!tb[IPSET_ATTR_IP] ||
+                    !tb[IPSET_ATTR_IFACE] ||
+                    !ip_set_optattr_netorder(tb, IPSET_ATTR_TIMEOUT) ||
+                    !ip_set_optattr_netorder(tb, IPSET_ATTR_CADT_FLAGS)))
+               return -IPSET_ERR_PROTOCOL;
+       if (unlikely(tb[IPSET_ATTR_IP_TO]))
+               return -IPSET_ERR_HASH_RANGE_UNSUPPORTED;
+
+       if (tb[IPSET_ATTR_LINENO])
+               *lineno = nla_get_u32(tb[IPSET_ATTR_LINENO]);
+
+       ret = ip_set_get_ipaddr6(tb[IPSET_ATTR_IP], &data.ip);
+       if (ret)
+               return ret;
+
+       if (tb[IPSET_ATTR_CIDR])
+               data.cidr = nla_get_u8(tb[IPSET_ATTR_CIDR]);
+       if (!data.cidr)
+               return -IPSET_ERR_INVALID_CIDR;
+       ip6_netmask(&data.ip, data.cidr);
+
+       if (tb[IPSET_ATTR_TIMEOUT]) {
+               if (!with_timeout(h->timeout))
+                       return -IPSET_ERR_TIMEOUT;
+               timeout = ip_set_timeout_uget(tb[IPSET_ATTR_TIMEOUT]);
+       }
+
+       strcpy(iface, nla_data(tb[IPSET_ATTR_IFACE])); 
+       data.iface = iface;
+       ret = iface_test(&h->rbtree, &data.iface);
+       if (adt == IPSET_ADD) {
+               if (!ret) {
+                       ret = iface_add(&h->rbtree, &data.iface);
+                       if (ret)
+                               return ret;
+               }
+       } else if (!ret)
+               return ret;
+
+       if (tb[IPSET_ATTR_CADT_FLAGS]) {
+               u32 flags = ip_set_get_h32(tb[IPSET_ATTR_CADT_FLAGS]);
+               if (flags & IPSET_FLAG_PHYSDEV)
+                       data.physdev = 1;
+       }
+
+       ret = adtfn(set, &data, timeout, flags);
+
+       return ip_set_eexist(ret, flags) ? 0 : ret;
+}
+
+/* Create hash:ip type of sets */
+
+static int
+hash_netiface_create(struct ip_set *set, struct nlattr *tb[], u32 flags)
+{
+       struct ip_set_hash *h;
+       u32 hashsize = IPSET_DEFAULT_HASHSIZE, maxelem = IPSET_DEFAULT_MAXELEM;
+       u8 hbits;
+
+       if (!(set->family == AF_INET || set->family == AF_INET6))
+               return -IPSET_ERR_INVALID_FAMILY;
+
+       if (unlikely(!ip_set_optattr_netorder(tb, IPSET_ATTR_HASHSIZE) ||
+                    !ip_set_optattr_netorder(tb, IPSET_ATTR_MAXELEM) ||
+                    !ip_set_optattr_netorder(tb, IPSET_ATTR_TIMEOUT)))
+               return -IPSET_ERR_PROTOCOL;
+
+       if (tb[IPSET_ATTR_HASHSIZE]) {
+               hashsize = ip_set_get_h32(tb[IPSET_ATTR_HASHSIZE]);
+               if (hashsize < IPSET_MIMINAL_HASHSIZE)
+                       hashsize = IPSET_MIMINAL_HASHSIZE;
+       }
+
+       if (tb[IPSET_ATTR_MAXELEM])
+               maxelem = ip_set_get_h32(tb[IPSET_ATTR_MAXELEM]);
+
+       h = kzalloc(sizeof(*h)
+                   + sizeof(struct ip_set_hash_nets)
+                     * (set->family == AF_INET ? 32 : 128), GFP_KERNEL);
+       if (!h)
+               return -ENOMEM;
+
+       h->maxelem = maxelem;
+       get_random_bytes(&h->initval, sizeof(h->initval));
+       h->timeout = IPSET_NO_TIMEOUT;
+
+       hbits = htable_bits(hashsize);
+       h->table = ip_set_alloc(
+                       sizeof(struct htable)
+                       + jhash_size(hbits) * sizeof(struct hbucket));
+       if (!h->table) {
+               kfree(h);
+               return -ENOMEM;
+       }
+       h->table->htable_bits = hbits;
+       h->rbtree = RB_ROOT;
+
+       set->data = h;
+
+       if (tb[IPSET_ATTR_TIMEOUT]) {
+               h->timeout = ip_set_timeout_uget(tb[IPSET_ATTR_TIMEOUT]);
+
+               set->variant = set->family == AF_INET
+                       ? &hash_netiface4_tvariant : &hash_netiface6_tvariant;
+
+               if (set->family == AF_INET)
+                       hash_netiface4_gc_init(set);
+               else
+                       hash_netiface6_gc_init(set);
+       } else {
+               set->variant = set->family == AF_INET
+                       ? &hash_netiface4_variant : &hash_netiface6_variant;
+       }
+
+       pr_debug("create %s hashsize %u (%u) maxelem %u: %p(%p)\n",
+                set->name, jhash_size(h->table->htable_bits),
+                h->table->htable_bits, h->maxelem, set->data, h->table);
+
+       return 0;
+}
+
+static struct ip_set_type hash_netiface_type __read_mostly = {
+       .name           = "hash:net,iface",
+       .protocol       = IPSET_PROTOCOL,
+       .features       = IPSET_TYPE_IP | IPSET_TYPE_IFACE,
+       .dimension      = IPSET_DIM_TWO,
+       .family         = AF_UNSPEC,
+       .revision_min   = 0,
+       .create         = hash_netiface_create,
+       .create_policy  = {
+               [IPSET_ATTR_HASHSIZE]   = { .type = NLA_U32 },
+               [IPSET_ATTR_MAXELEM]    = { .type = NLA_U32 },
+               [IPSET_ATTR_PROBES]     = { .type = NLA_U8 },
+               [IPSET_ATTR_RESIZE]     = { .type = NLA_U8  },
+               [IPSET_ATTR_PROTO]      = { .type = NLA_U8 },
+               [IPSET_ATTR_TIMEOUT]    = { .type = NLA_U32 },
+       },
+       .adt_policy     = {
+               [IPSET_ATTR_IP]         = { .type = NLA_NESTED },
+               [IPSET_ATTR_IP_TO]      = { .type = NLA_NESTED },
+               [IPSET_ATTR_IFACE]      = { .type = NLA_NUL_STRING,
+                                           .len = IPSET_MAXNAMELEN - 1 },
+               [IPSET_ATTR_CADT_FLAGS] = { .type = NLA_U32 },
+               [IPSET_ATTR_CIDR]       = { .type = NLA_U8 },
+               [IPSET_ATTR_TIMEOUT]    = { .type = NLA_U32 },
+               [IPSET_ATTR_LINENO]     = { .type = NLA_U32 },
+       },
+       .me             = THIS_MODULE,
+};
+
+static int __init
+hash_netiface_init(void)
+{
+       return ip_set_type_register(&hash_netiface_type);
+}
+
+static void __exit
+hash_netiface_fini(void)
+{
+       ip_set_type_unregister(&hash_netiface_type);
+}
+
+module_init(hash_netiface_init);
+module_exit(hash_netiface_fini);