netfilter: ipset: hash:net,iface fixed to handle overlapping nets behind different interfaces

If overlapping networks with different interfaces was added to
the set, the type did not handle it properly. Example

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

Now, if a packet was sent from 192.168.0.0/24,eth0, the type returned
a match.

In the patch the algorithm is fixed in order to correctly handle
overlapping networks.

Limitation: the same network cannot be stored with more than 64 different
interfaces in a single set.

Signed-off-by: Jozsef Kadlecsik <kadlec@blackhole.kfki.hu>
Signed-off-by: Patrick McHardy <kaber@trash.net>

8 files changed, 117 insertions, 51 deletions

diff --git a/include/linux/netfilter/ipset/ip_set_ahash.h b/include/linux/netfilter/ipset/ip_set_ahash.h
index 42b7d25a1b2e..1e7f7594cd02 100644
--- a/include/linux/netfilter/ipset/ip_set_ahash.h
+++ b/include/linux/netfilter/ipset/ip_set_ahash.h
@@ -28,7 +28,32 @@
 /* Number of elements to store in an initial array block */
 #define AHASH_INIT_SIZE			4
 /* Max number of elements to store in an array block */
-#define AHASH_MAX_SIZE			(3*4)
+#define AHASH_MAX_SIZE			(3*AHASH_INIT_SIZE)
+
+/* Max number of elements can be tuned */
+#ifdef IP_SET_HASH_WITH_MULTI
+#define AHASH_MAX(h)			((h)->ahash_max)
+
+static inline u8
+tune_ahash_max(u8 curr, u32 multi)
+{
+	u32 n;
+
+	if (multi < curr)
+		return curr;
+
+	n = curr + AHASH_INIT_SIZE;
+	/* Currently, at listing one hash bucket must fit into a message.
+	 * Therefore we have a hard limit here.
+	 */
+	return n > curr && n <= 64 ? n : curr;
+}
+#define TUNE_AHASH_MAX(h, multi)	\
+	((h)->ahash_max = tune_ahash_max((h)->ahash_max, multi))
+#else
+#define AHASH_MAX(h)			AHASH_MAX_SIZE
+#define TUNE_AHASH_MAX(h, multi)
+#endif
 
 /* A hash bucket */
 struct hbucket {
@@ -60,6 +85,9 @@ struct ip_set_hash {
 	u32 timeout;		/* timeout value, if enabled */
 	struct timer_list gc;	/* garbage collection when timeout enabled */
 	struct type_pf_next next; /* temporary storage for uadd */
+#ifdef IP_SET_HASH_WITH_MULTI
+	u8 ahash_max;		/* max elements in an array block */
+#endif
 #ifdef IP_SET_HASH_WITH_NETMASK
 	u8 netmask;		/* netmask value for subnets to store */
 #endif
@@ -279,12 +307,13 @@ ip_set_hash_destroy(struct ip_set *set)
 /* Add an element to the hash table when resizing the set:
  * we spare the maintenance of the internal counters. */
 static int
-type_pf_elem_add(struct hbucket *n, const struct type_pf_elem *value)
+type_pf_elem_add(struct hbucket *n, const struct type_pf_elem *value,
+		 u8 ahash_max)
 {
 	if (n->pos >= n->size) {
 		void *tmp;
 
-		if (n->size >= AHASH_MAX_SIZE)
+		if (n->size >= ahash_max)
 			/* Trigger rehashing */
 			return -EAGAIN;
 
@@ -339,7 +368,7 @@ retry:
 		for (j = 0; j < n->pos; j++) {
 			data = ahash_data(n, j);
 			m = hbucket(t, HKEY(data, h->initval, htable_bits));
-			ret = type_pf_elem_add(m, data);
+			ret = type_pf_elem_add(m, data, AHASH_MAX(h));
 			if (ret < 0) {
 				read_unlock_bh(&set->lock);
 				ahash_destroy(t);
@@ -376,7 +405,7 @@ type_pf_add(struct ip_set *set, void *value, u32 timeout, u32 flags)
 	const struct type_pf_elem *d = value;
 	struct hbucket *n;
 	int i, ret = 0;
-	u32 key;
+	u32 key, multi = 0;
 
 	if (h->elements >= h->maxelem)
 		return -IPSET_ERR_HASH_FULL;
@@ -386,12 +415,12 @@ type_pf_add(struct ip_set *set, void *value, u32 timeout, u32 flags)
 	key = HKEY(value, h->initval, t->htable_bits);
 	n = hbucket(t, key);
 	for (i = 0; i < n->pos; i++)
-		if (type_pf_data_equal(ahash_data(n, i), d)) {
+		if (type_pf_data_equal(ahash_data(n, i), d, &multi)) {
 			ret = -IPSET_ERR_EXIST;
 			goto out;
 		}
-
-	ret = type_pf_elem_add(n, value);
+	TUNE_AHASH_MAX(h, multi);
+	ret = type_pf_elem_add(n, value, AHASH_MAX(h));
 	if (ret != 0) {
 		if (ret == -EAGAIN)
 			type_pf_data_next(h, d);
@@ -419,13 +448,13 @@ type_pf_del(struct ip_set *set, void *value, u32 timeout, u32 flags)
 	struct hbucket *n;
 	int i;
 	struct type_pf_elem *data;
-	u32 key;
+	u32 key, multi = 0;
 
 	key = HKEY(value, h->initval, t->htable_bits);
 	n = hbucket(t, key);
 	for (i = 0; i < n->pos; i++) {
 		data = ahash_data(n, i);
-		if (!type_pf_data_equal(data, d))
+		if (!type_pf_data_equal(data, d, &multi))
 			continue;
 		if (i != n->pos - 1)
 			/* Not last one */
@@ -466,17 +495,17 @@ type_pf_test_cidrs(struct ip_set *set, struct type_pf_elem *d, u32 timeout)
 	struct hbucket *n;
 	const struct type_pf_elem *data;
 	int i, j = 0;
-	u32 key;
+	u32 key, multi = 0;
 	u8 host_mask = SET_HOST_MASK(set->family);
 
 	pr_debug("test by nets\n");
-	for (; j < host_mask && h->nets[j].cidr; j++) {
+	for (; j < host_mask && h->nets[j].cidr && !multi; j++) {
 		type_pf_data_netmask(d, h->nets[j].cidr);
 		key = HKEY(d, h->initval, t->htable_bits);
 		n = hbucket(t, key);
 		for (i = 0; i < n->pos; i++) {
 			data = ahash_data(n, i);
-			if (type_pf_data_equal(data, d))
+			if (type_pf_data_equal(data, d, &multi))
 				return 1;
 		}
 	}
@@ -494,7 +523,7 @@ type_pf_test(struct ip_set *set, void *value, u32 timeout, u32 flags)
 	struct hbucket *n;
 	const struct type_pf_elem *data;
 	int i;
-	u32 key;
+	u32 key, multi = 0;
 
 #ifdef IP_SET_HASH_WITH_NETS
 	/* If we test an IP address and not a network address,
@@ -507,7 +536,7 @@ type_pf_test(struct ip_set *set, void *value, u32 timeout, u32 flags)
 	n = hbucket(t, key);
 	for (i = 0; i < n->pos; i++) {
 		data = ahash_data(n, i);
-		if (type_pf_data_equal(data, d))
+		if (type_pf_data_equal(data, d, &multi))
 			return 1;
 	}
 	return 0;
@@ -664,14 +693,14 @@ type_pf_data_timeout_set(struct type_pf_elem *data, u32 timeout)
 
 static int
 type_pf_elem_tadd(struct hbucket *n, const struct type_pf_elem *value,
-		  u32 timeout)
+		  u8 ahash_max, u32 timeout)
 {
 	struct type_pf_elem *data;
 
 	if (n->pos >= n->size) {
 		void *tmp;
 
-		if (n->size >= AHASH_MAX_SIZE)
+		if (n->size >= ahash_max)
 			/* Trigger rehashing */
 			return -EAGAIN;
 
@@ -776,7 +805,7 @@ retry:
 		for (j = 0; j < n->pos; j++) {
 			data = ahash_tdata(n, j);
 			m = hbucket(t, HKEY(data, h->initval, htable_bits));
-			ret = type_pf_elem_tadd(m, data,
+			ret = type_pf_elem_tadd(m, data, AHASH_MAX(h),
 						type_pf_data_timeout(data));
 			if (ret < 0) {
 				read_unlock_bh(&set->lock);
@@ -807,9 +836,9 @@ type_pf_tadd(struct ip_set *set, void *value, u32 timeout, u32 flags)
 	const struct type_pf_elem *d = value;
 	struct hbucket *n;
 	struct type_pf_elem *data;
-	int ret = 0, i, j = AHASH_MAX_SIZE + 1;
+	int ret = 0, i, j = AHASH_MAX(h) + 1;
 	bool flag_exist = flags & IPSET_FLAG_EXIST;
-	u32 key;
+	u32 key, multi = 0;
 
 	if (h->elements >= h->maxelem)
 		/* FIXME: when set is full, we slow down here */
@@ -823,18 +852,18 @@ type_pf_tadd(struct ip_set *set, void *value, u32 timeout, u32 flags)
 	n = hbucket(t, key);
 	for (i = 0; i < n->pos; i++) {
 		data = ahash_tdata(n, i);
-		if (type_pf_data_equal(data, d)) {
+		if (type_pf_data_equal(data, d, &multi)) {
 			if (type_pf_data_expired(data) || flag_exist)
 				j = i;
 			else {
 				ret = -IPSET_ERR_EXIST;
 				goto out;
 			}
-		} else if (j == AHASH_MAX_SIZE + 1 &&
+		} else if (j == AHASH_MAX(h) + 1 &&
 			   type_pf_data_expired(data))
 			j = i;
 	}
-	if (j != AHASH_MAX_SIZE + 1) {
+	if (j != AHASH_MAX(h) + 1) {
 		data = ahash_tdata(n, j);
 #ifdef IP_SET_HASH_WITH_NETS
 		del_cidr(h, data->cidr, HOST_MASK);
@@ -844,7 +873,8 @@ type_pf_tadd(struct ip_set *set, void *value, u32 timeout, u32 flags)
 		type_pf_data_timeout_set(data, timeout);
 		goto out;
 	}
-	ret = type_pf_elem_tadd(n, d, timeout);
+	TUNE_AHASH_MAX(h, multi);
+	ret = type_pf_elem_tadd(n, d, AHASH_MAX(h), timeout);
 	if (ret != 0) {
 		if (ret == -EAGAIN)
 			type_pf_data_next(h, d);
@@ -869,13 +899,13 @@ type_pf_tdel(struct ip_set *set, void *value, u32 timeout, u32 flags)
 	struct hbucket *n;
 	int i;
 	struct type_pf_elem *data;
-	u32 key;
+	u32 key, multi = 0;
 
 	key = HKEY(value, h->initval, t->htable_bits);
 	n = hbucket(t, key);
 	for (i = 0; i < n->pos; i++) {
 		data = ahash_tdata(n, i);
-		if (!type_pf_data_equal(data, d))
+		if (!type_pf_data_equal(data, d, &multi))
 			continue;
 		if (type_pf_data_expired(data))
 			return -IPSET_ERR_EXIST;
@@ -915,16 +945,16 @@ type_pf_ttest_cidrs(struct ip_set *set, struct type_pf_elem *d, u32 timeout)
 	struct type_pf_elem *data;
 	struct hbucket *n;
 	int i, j = 0;
-	u32 key;
+	u32 key, multi = 0;
 	u8 host_mask = SET_HOST_MASK(set->family);
 
-	for (; j < host_mask && h->nets[j].cidr; j++) {
+	for (; j < host_mask && h->nets[j].cidr && !multi; j++) {
 		type_pf_data_netmask(d, h->nets[j].cidr);
 		key = HKEY(d, h->initval, t->htable_bits);
 		n = hbucket(t, key);
 		for (i = 0; i < n->pos; i++) {
 			data = ahash_tdata(n, i);
-			if (type_pf_data_equal(data, d))
+			if (type_pf_data_equal(data, d, &multi))
 				return !type_pf_data_expired(data);
 		}
 	}
@@ -940,7 +970,7 @@ type_pf_ttest(struct ip_set *set, void *value, u32 timeout, u32 flags)
 	struct type_pf_elem *data, *d = value;
 	struct hbucket *n;
 	int i;
-	u32 key;
+	u32 key, multi = 0;
 
 #ifdef IP_SET_HASH_WITH_NETS
 	if (d->cidr == SET_HOST_MASK(set->family))
@@ -950,7 +980,7 @@ type_pf_ttest(struct ip_set *set, void *value, u32 timeout, u32 flags)
 	n = hbucket(t, key);
 	for (i = 0; i < n->pos; i++) {
 		data = ahash_tdata(n, i);
-		if (type_pf_data_equal(data, d))
+		if (type_pf_data_equal(data, d, &multi))
 			return !type_pf_data_expired(data);
 	}
 	return 0;
diff --git a/net/netfilter/ipset/ip_set_hash_ip.c b/net/netfilter/ipset/ip_set_hash_ip.c
index fa80bb9b9c81..f2d576e6b769 100644
--- a/net/netfilter/ipset/ip_set_hash_ip.c
+++ b/net/netfilter/ipset/ip_set_hash_ip.c
@@ -53,7 +53,8 @@ struct hash_ip4_telem {
 
 static inline bool
 hash_ip4_data_equal(const struct hash_ip4_elem *ip1,
-		    const struct hash_ip4_elem *ip2)
+		    const struct hash_ip4_elem *ip2,
+		    u32 *multi)
 {
 	return ip1->ip == ip2->ip;
 }
@@ -225,7 +226,8 @@ struct hash_ip6_telem {
 
 static inline bool
 hash_ip6_data_equal(const struct hash_ip6_elem *ip1,
-		    const struct hash_ip6_elem *ip2)
+		    const struct hash_ip6_elem *ip2,
+		    u32 *multi)
 {
 	return ipv6_addr_cmp(&ip1->ip.in6, &ip2->ip.in6) == 0;
 }
diff --git a/net/netfilter/ipset/ip_set_hash_ipport.c b/net/netfilter/ipset/ip_set_hash_ipport.c
index bbf51b67b170..6ee10f5d59bd 100644
--- a/net/netfilter/ipset/ip_set_hash_ipport.c
+++ b/net/netfilter/ipset/ip_set_hash_ipport.c
@@ -60,7 +60,8 @@ struct hash_ipport4_telem {
 
 static inline bool
 hash_ipport4_data_equal(const struct hash_ipport4_elem *ip1,
-			const struct hash_ipport4_elem *ip2)
+			const struct hash_ipport4_elem *ip2,
+			u32 *multi)
 {
 	return ip1->ip == ip2->ip &&
 	       ip1->port == ip2->port &&
@@ -276,7 +277,8 @@ struct hash_ipport6_telem {
 
 static inline bool
 hash_ipport6_data_equal(const struct hash_ipport6_elem *ip1,
-			const struct hash_ipport6_elem *ip2)
+			const struct hash_ipport6_elem *ip2,
+			u32 *multi)
 {
 	return ipv6_addr_cmp(&ip1->ip.in6, &ip2->ip.in6) == 0 &&
 	       ip1->port == ip2->port &&
diff --git a/net/netfilter/ipset/ip_set_hash_ipportip.c b/net/netfilter/ipset/ip_set_hash_ipportip.c
index 96525f529a54..fb90e344e907 100644
--- a/net/netfilter/ipset/ip_set_hash_ipportip.c
+++ b/net/netfilter/ipset/ip_set_hash_ipportip.c
@@ -62,7 +62,8 @@ struct hash_ipportip4_telem {
 
 static inline bool
 hash_ipportip4_data_equal(const struct hash_ipportip4_elem *ip1,
-			  const struct hash_ipportip4_elem *ip2)
+			  const struct hash_ipportip4_elem *ip2,
+			  u32 *multi)
 {
 	return ip1->ip == ip2->ip &&
 	       ip1->ip2 == ip2->ip2 &&
@@ -286,7 +287,8 @@ struct hash_ipportip6_telem {
 
 static inline bool
 hash_ipportip6_data_equal(const struct hash_ipportip6_elem *ip1,
-			  const struct hash_ipportip6_elem *ip2)
+			  const struct hash_ipportip6_elem *ip2,
+			  u32 *multi)
 {
 	return ipv6_addr_cmp(&ip1->ip.in6, &ip2->ip.in6) == 0 &&
 	       ipv6_addr_cmp(&ip1->ip2.in6, &ip2->ip2.in6) == 0 &&
diff --git a/net/netfilter/ipset/ip_set_hash_ipportnet.c b/net/netfilter/ipset/ip_set_hash_ipportnet.c
index d2d6ab89f087..deb3e3dfa5fc 100644
--- a/net/netfilter/ipset/ip_set_hash_ipportnet.c
+++ b/net/netfilter/ipset/ip_set_hash_ipportnet.c
@@ -62,7 +62,8 @@ struct hash_ipportnet4_telem {
 
 static inline bool
 hash_ipportnet4_data_equal(const struct hash_ipportnet4_elem *ip1,
-			   const struct hash_ipportnet4_elem *ip2)
+			   const struct hash_ipportnet4_elem *ip2,
+			   u32 *multi)
 {
 	return ip1->ip == ip2->ip &&
 	       ip1->ip2 == ip2->ip2 &&
@@ -335,7 +336,8 @@ struct hash_ipportnet6_telem {
 
 static inline bool
 hash_ipportnet6_data_equal(const struct hash_ipportnet6_elem *ip1,
-			   const struct hash_ipportnet6_elem *ip2)
+			   const struct hash_ipportnet6_elem *ip2,
+			   u32 *multi)
 {
 	return ipv6_addr_cmp(&ip1->ip.in6, &ip2->ip.in6) == 0 &&
 	       ipv6_addr_cmp(&ip1->ip2.in6, &ip2->ip2.in6) == 0 &&
diff --git a/net/netfilter/ipset/ip_set_hash_net.c b/net/netfilter/ipset/ip_set_hash_net.c
index 2d4b1f48e8c9..60d016541c58 100644
--- a/net/netfilter/ipset/ip_set_hash_net.c
+++ b/net/netfilter/ipset/ip_set_hash_net.c
@@ -58,7 +58,8 @@ struct hash_net4_telem {
 
 static inline bool
 hash_net4_data_equal(const struct hash_net4_elem *ip1,
-		    const struct hash_net4_elem *ip2)
+		     const struct hash_net4_elem *ip2,
+		     u32 *multi)
 {
 	return ip1->ip == ip2->ip && ip1->cidr == ip2->cidr;
 }
@@ -249,7 +250,8 @@ struct hash_net6_telem {
 
 static inline bool
 hash_net6_data_equal(const struct hash_net6_elem *ip1,
-		     const struct hash_net6_elem *ip2)
+		     const struct hash_net6_elem *ip2,
+		     u32 *multi)
 {
 	return ipv6_addr_cmp(&ip1->ip.in6, &ip2->ip.in6) == 0 &&
 	       ip1->cidr == ip2->cidr;
diff --git a/net/netfilter/ipset/ip_set_hash_netiface.c b/net/netfilter/ipset/ip_set_hash_netiface.c
index 3d6c53b6211a..e13095deb50d 100644
--- a/net/netfilter/ipset/ip_set_hash_netiface.c
+++ b/net/netfilter/ipset/ip_set_hash_netiface.c
@@ -99,7 +99,7 @@ iface_test(struct rb_root *root, const char **iface)
 
 	while (n) {
 		const char *d = iface_data(n);
-		int res = ifname_compare(*iface, d);
+		long res = ifname_compare(*iface, d);
 
 		if (res < 0)
 			n = n->rb_left;
@@ -121,7 +121,7 @@ iface_add(struct rb_root *root, const char **iface)
 
 	while (*n) {
 		char *ifname = iface_data(*n);
-		int res = ifname_compare(*iface, ifname);
+		long res = ifname_compare(*iface, ifname);
 
 		p = *n;
 		if (res < 0)
@@ -159,31 +159,42 @@ hash_netiface_same_set(const struct ip_set *a, const struct ip_set *b);
 
 /* The type variant functions: IPv4 */
 
+struct hash_netiface4_elem_hashed {
+	__be32 ip;
+	u8 physdev;
+	u8 cidr;
+	u16 padding;
+};
+
+#define HKEY_DATALEN	sizeof(struct hash_netiface4_elem_hashed)
+
 /* Member elements without timeout */
 struct hash_netiface4_elem {
 	__be32 ip;
-	const char *iface;
 	u8 physdev;
 	u8 cidr;
 	u16 padding;
+	const char *iface;
 };
 
 /* Member elements with timeout support */
 struct hash_netiface4_telem {
 	__be32 ip;
-	const char *iface;
 	u8 physdev;
 	u8 cidr;
 	u16 padding;
+	const char *iface;
 	unsigned long timeout;
 };
 
 static inline bool
 hash_netiface4_data_equal(const struct hash_netiface4_elem *ip1,
-			  const struct hash_netiface4_elem *ip2)
+			  const struct hash_netiface4_elem *ip2,
+			  u32 *multi)
 {
 	return ip1->ip == ip2->ip &&
 	       ip1->cidr == ip2->cidr &&
+	       (++*multi) &&
 	       ip1->physdev == ip2->physdev &&
 	       ip1->iface == ip2->iface;
 }
@@ -257,6 +268,7 @@ nla_put_failure:
 
 #define IP_SET_HASH_WITH_NETS
 #define IP_SET_HASH_WITH_RBTREE
+#define IP_SET_HASH_WITH_MULTI
 
 #define PF		4
 #define HOST_MASK	32
@@ -424,29 +436,40 @@ hash_netiface_same_set(const struct ip_set *a, const struct ip_set *b)
 
 /* The type variant functions: IPv6 */
 
+struct hash_netiface6_elem_hashed {
+	union nf_inet_addr ip;
+	u8 physdev;
+	u8 cidr;
+	u16 padding;
+};
+
+#define HKEY_DATALEN	sizeof(struct hash_netiface6_elem_hashed)
+
 struct hash_netiface6_elem {
 	union nf_inet_addr ip;
-	const char *iface;
 	u8 physdev;
 	u8 cidr;
 	u16 padding;
+	const char *iface;
 };
 
 struct hash_netiface6_telem {
 	union nf_inet_addr ip;
-	const char *iface;
 	u8 physdev;
 	u8 cidr;
 	u16 padding;
+	const char *iface;
 	unsigned long timeout;
 };
 
 static inline bool
 hash_netiface6_data_equal(const struct hash_netiface6_elem *ip1,
-			  const struct hash_netiface6_elem *ip2)
+			  const struct hash_netiface6_elem *ip2,
+			  u32 *multi)
 {
 	return ipv6_addr_cmp(&ip1->ip.in6, &ip2->ip.in6) == 0 &&
 	       ip1->cidr == ip2->cidr &&
+	       (++*multi) &&
 	       ip1->physdev == ip2->physdev &&
 	       ip1->iface == ip2->iface;
 }
@@ -681,6 +704,7 @@ hash_netiface_create(struct ip_set *set, struct nlattr *tb[], u32 flags)
 	h->maxelem = maxelem;
 	get_random_bytes(&h->initval, sizeof(h->initval));
 	h->timeout = IPSET_NO_TIMEOUT;
+	h->ahash_max = AHASH_MAX_SIZE;
 
 	hbits = htable_bits(hashsize);
 	h->table = ip_set_alloc(
diff --git a/net/netfilter/ipset/ip_set_hash_netport.c b/net/netfilter/ipset/ip_set_hash_netport.c
index fe203d12f56b..8f9de7207ec9 100644
--- a/net/netfilter/ipset/ip_set_hash_netport.c
+++ b/net/netfilter/ipset/ip_set_hash_netport.c
@@ -59,7 +59,8 @@ struct hash_netport4_telem {
 
 static inline bool
 hash_netport4_data_equal(const struct hash_netport4_elem *ip1,
-			 const struct hash_netport4_elem *ip2)
+			 const struct hash_netport4_elem *ip2,
+			 u32 *multi)
 {
 	return ip1->ip == ip2->ip &&
 	       ip1->port == ip2->port &&
@@ -300,7 +301,8 @@ struct hash_netport6_telem {
 
 static inline bool
 hash_netport6_data_equal(const struct hash_netport6_elem *ip1,
-			 const struct hash_netport6_elem *ip2)
+			 const struct hash_netport6_elem *ip2,
+			 u32 *multi)
 {
 	return ipv6_addr_cmp(&ip1->ip.in6, &ip2->ip.in6) == 0 &&
 	       ip1->port == ip2->port &&