Merge branch 'master' into upstream

9 files changed, 1786 insertions, 391 deletions

diff --git a/net/xfrm/Kconfig b/net/xfrm/Kconfig
index 0c1c04322baf..0faab6332586 100644
--- a/net/xfrm/Kconfig
+++ b/net/xfrm/Kconfig
@@ -6,14 +6,24 @@ config XFRM
        depends on NET
 
 config XFRM_USER
-        tristate "IPsec user configuration interface"
+        tristate "Transformation user configuration interface"
         depends on INET && XFRM
         ---help---
-          Support for IPsec user configuration interface used
-          by native Linux tools.
+          Support for Transformation(XFRM) user configuration interface
+          like IPsec used by native Linux tools.
 
           If unsure, say Y.
 
+config XFRM_SUB_POLICY
+        bool "Transformation sub policy support (EXPERIMENTAL)"
+        depends on XFRM && EXPERIMENTAL
+        ---help---
+          Support sub policy for developers. By using sub policy with main
+          one, two policies can be applied to the same packet at once.
+          Policy which lives shorter time in kernel should be a sub.
+
+          If unsure, say N.
+
 config NET_KEY
         tristate "PF_KEY sockets"
         select XFRM
diff --git a/net/xfrm/Makefile b/net/xfrm/Makefile
index 693aac1aa833..de3c1a625a46 100644
--- a/net/xfrm/Makefile
+++ b/net/xfrm/Makefile
@@ -2,6 +2,7 @@
 # Makefile for the XFRM subsystem.
 #
 
-obj-$(CONFIG_XFRM) := xfrm_policy.o xfrm_state.o xfrm_input.o xfrm_algo.o
+obj-$(CONFIG_XFRM) := xfrm_policy.o xfrm_state.o xfrm_hash.o \
+                      xfrm_input.o xfrm_algo.o
 obj-$(CONFIG_XFRM_USER) += xfrm_user.o
 
diff --git a/net/xfrm/xfrm_algo.c b/net/xfrm/xfrm_algo.c
index 04e1aea58bc9..5a0dbeb6bbe8 100644
--- a/net/xfrm/xfrm_algo.c
+++ b/net/xfrm/xfrm_algo.c
@@ -30,7 +30,8 @@
  */
 static struct xfrm_algo_desc aalg_list[] = {
 {
-        .name = "digest_null",
+        .name = "hmac(digest_null)",
+        .compat = "digest_null",
         
         .uinfo = {
                 .auth = {
@@ -47,7 +48,8 @@ static struct xfrm_algo_desc aalg_list[] = {
         }
 },
 {
-        .name = "md5",
+        .name = "hmac(md5)",
+        .compat = "md5",
 
         .uinfo = {
                 .auth = {
@@ -64,7 +66,8 @@ static struct xfrm_algo_desc aalg_list[] = {
         }
 },
 {
-        .name = "sha1",
+        .name = "hmac(sha1)",
+        .compat = "sha1",
 
         .uinfo = {
                 .auth = {
@@ -81,7 +84,8 @@ static struct xfrm_algo_desc aalg_list[] = {
         }
 },
 {
-        .name = "sha256",
+        .name = "hmac(sha256)",
+        .compat = "sha256",
 
         .uinfo = {
                 .auth = {
@@ -98,7 +102,8 @@ static struct xfrm_algo_desc aalg_list[] = {
         }
 },
 {
-        .name = "ripemd160",
+        .name = "hmac(ripemd160)",
+        .compat = "ripemd160",
 
         .uinfo = {
                 .auth = {
@@ -118,7 +123,8 @@ static struct xfrm_algo_desc aalg_list[] = {
 
 static struct xfrm_algo_desc ealg_list[] = {
 {
-        .name = "cipher_null",
+        .name = "ecb(cipher_null)",
+        .compat = "cipher_null",
         
         .uinfo = {
                 .encr = {
@@ -135,7 +141,8 @@ static struct xfrm_algo_desc ealg_list[] = {
         }
 },
 {
-        .name = "des",
+        .name = "cbc(des)",
+        .compat = "des",
 
         .uinfo = {
                 .encr = {
@@ -152,7 +159,8 @@ static struct xfrm_algo_desc ealg_list[] = {
         }
 },
 {
-        .name = "des3_ede",
+        .name = "cbc(des3_ede)",
+        .compat = "des3_ede",
 
         .uinfo = {
                 .encr = {
@@ -169,7 +177,8 @@ static struct xfrm_algo_desc ealg_list[] = {
         }
 },
 {
-        .name = "cast128",
+        .name = "cbc(cast128)",
+        .compat = "cast128",
 
         .uinfo = {
                 .encr = {
@@ -186,7 +195,8 @@ static struct xfrm_algo_desc ealg_list[] = {
         }
 },
 {
-        .name = "blowfish",
+        .name = "cbc(blowfish)",
+        .compat = "blowfish",
 
         .uinfo = {
                 .encr = {
@@ -203,7 +213,8 @@ static struct xfrm_algo_desc ealg_list[] = {
         }
 },
 {
-        .name = "aes",
+        .name = "cbc(aes)",
+        .compat = "aes",
 
         .uinfo = {
                 .encr = {
@@ -220,7 +231,8 @@ static struct xfrm_algo_desc ealg_list[] = {
         }
 },
 {
-        .name = "serpent",
+        .name = "cbc(serpent)",
+        .compat = "serpent",
 
         .uinfo = {
                 .encr = {
@@ -237,7 +249,8 @@ static struct xfrm_algo_desc ealg_list[] = {
         }
 },
 {
-        .name = "twofish",
+        .name = "cbc(twofish)",
+        .compat = "twofish",
                  
         .uinfo = {
                 .encr = {
@@ -350,8 +363,8 @@ struct xfrm_algo_desc *xfrm_calg_get_byid(int alg_id)
 EXPORT_SYMBOL_GPL(xfrm_calg_get_byid);
 
 static struct xfrm_algo_desc *xfrm_get_byname(struct xfrm_algo_desc *list,
-                                              int entries, char *name,
-                                              int probe)
+                                              int entries, u32 type, u32 mask,
+                                              char *name, int probe)
 {
         int i, status;
 
@@ -359,7 +372,8 @@ static struct xfrm_algo_desc *xfrm_get_byname(struct xfrm_algo_desc *list,
                 return NULL;
 
         for (i = 0; i < entries; i++) {
-                if (strcmp(name, list[i].name))
+                if (strcmp(name, list[i].name) &&
+                    (!list[i].compat || strcmp(name, list[i].compat)))
                         continue;
 
                 if (list[i].available)
@@ -368,7 +382,7 @@ static struct xfrm_algo_desc *xfrm_get_byname(struct xfrm_algo_desc *list,
                 if (!probe)
                         break;
 
-                status = crypto_alg_available(name, 0);
+                status = crypto_has_alg(name, type, mask | CRYPTO_ALG_ASYNC);
                 if (!status)
                         break;
 
@@ -380,19 +394,25 @@ static struct xfrm_algo_desc *xfrm_get_byname(struct xfrm_algo_desc *list,
 
 struct xfrm_algo_desc *xfrm_aalg_get_byname(char *name, int probe)
 {
-        return xfrm_get_byname(aalg_list, aalg_entries(), name, probe);
+        return xfrm_get_byname(aalg_list, aalg_entries(),
+                               CRYPTO_ALG_TYPE_HASH, CRYPTO_ALG_TYPE_HASH_MASK,
+                               name, probe);
 }
 EXPORT_SYMBOL_GPL(xfrm_aalg_get_byname);
 
 struct xfrm_algo_desc *xfrm_ealg_get_byname(char *name, int probe)
 {
-        return xfrm_get_byname(ealg_list, ealg_entries(), name, probe);
+        return xfrm_get_byname(ealg_list, ealg_entries(),
+                               CRYPTO_ALG_TYPE_BLKCIPHER, CRYPTO_ALG_TYPE_MASK,
+                               name, probe);
 }
 EXPORT_SYMBOL_GPL(xfrm_ealg_get_byname);
 
 struct xfrm_algo_desc *xfrm_calg_get_byname(char *name, int probe)
 {
-        return xfrm_get_byname(calg_list, calg_entries(), name, probe);
+        return xfrm_get_byname(calg_list, calg_entries(),
+                               CRYPTO_ALG_TYPE_COMPRESS, CRYPTO_ALG_TYPE_MASK,
+                               name, probe);
 }
 EXPORT_SYMBOL_GPL(xfrm_calg_get_byname);
 
@@ -427,19 +447,22 @@ void xfrm_probe_algs(void)
         BUG_ON(in_softirq());
 
         for (i = 0; i < aalg_entries(); i++) {
-                status = crypto_alg_available(aalg_list[i].name, 0);
+                status = crypto_has_hash(aalg_list[i].name, 0,
+                                         CRYPTO_ALG_ASYNC);
                 if (aalg_list[i].available != status)
                         aalg_list[i].available = status;
         }
         
         for (i = 0; i < ealg_entries(); i++) {
-                status = crypto_alg_available(ealg_list[i].name, 0);
+                status = crypto_has_blkcipher(ealg_list[i].name, 0,
+                                              CRYPTO_ALG_ASYNC);
                 if (ealg_list[i].available != status)
                         ealg_list[i].available = status;
         }
         
         for (i = 0; i < calg_entries(); i++) {
-                status = crypto_alg_available(calg_list[i].name, 0);
+                status = crypto_has_comp(calg_list[i].name, 0,
+                                         CRYPTO_ALG_ASYNC);
                 if (calg_list[i].available != status)
                         calg_list[i].available = status;
         }
@@ -471,11 +494,12 @@ EXPORT_SYMBOL_GPL(xfrm_count_enc_supported);
 
 /* Move to common area: it is shared with AH. */
 
-void skb_icv_walk(const struct sk_buff *skb, struct crypto_tfm *tfm,
-                  int offset, int len, icv_update_fn_t icv_update)
+int skb_icv_walk(const struct sk_buff *skb, struct hash_desc *desc,
+                 int offset, int len, icv_update_fn_t icv_update)
 {
         int start = skb_headlen(skb);
         int i, copy = start - offset;
+        int err;
         struct scatterlist sg;
 
         /* Checksum header. */
@@ -487,10 +511,12 @@ void skb_icv_walk(const struct sk_buff *skb, struct crypto_tfm *tfm,
                 sg.offset = (unsigned long)(skb->data + offset) % PAGE_SIZE;
                 sg.length = copy;
                 
-                icv_update(tfm, &sg, 1);
+                err = icv_update(desc, &sg, copy);
+                if (unlikely(err))
+                        return err;
                 
                 if ((len -= copy) == 0)
-                        return;
+                        return 0;
                 offset += copy;
         }
 
@@ -510,10 +536,12 @@ void skb_icv_walk(const struct sk_buff *skb, struct crypto_tfm *tfm,
                         sg.offset = frag->page_offset + offset-start;
                         sg.length = copy;
                         
-                        icv_update(tfm, &sg, 1);
+                        err = icv_update(desc, &sg, copy);
+                        if (unlikely(err))
+                                return err;
 
                         if (!(len -= copy))
-                                return;
+                                return 0;
                         offset += copy;
                 }
                 start = end;
@@ -531,15 +559,19 @@ void skb_icv_walk(const struct sk_buff *skb, struct crypto_tfm *tfm,
                         if ((copy = end - offset) > 0) {
                                 if (copy > len)
                                         copy = len;
-                                skb_icv_walk(list, tfm, offset-start, copy, icv_update);
+                                err = skb_icv_walk(list, desc, offset-start,
+                                                   copy, icv_update);
+                                if (unlikely(err))
+                                        return err;
                                 if ((len -= copy) == 0)
-                                        return;
+                                        return 0;
                                 offset += copy;
                         }
                         start = end;
                 }
         }
         BUG_ON(len);
+        return 0;
 }
 EXPORT_SYMBOL_GPL(skb_icv_walk);
 
diff --git a/net/xfrm/xfrm_hash.c b/net/xfrm/xfrm_hash.c
new file mode 100644
index 000000000000..37643bb8768a
--- /dev/null
+++ b/net/xfrm/xfrm_hash.c
@@ -0,0 +1,41 @@
+/* xfrm_hash.c: Common hash table code.
+ *
+ * Copyright (C) 2006 David S. Miller (davem@davemloft.net)
+ */
+
+#include <linux/kernel.h>
+#include <linux/mm.h>
+#include <linux/bootmem.h>
+#include <linux/vmalloc.h>
+#include <linux/slab.h>
+#include <linux/xfrm.h>
+
+#include "xfrm_hash.h"
+
+struct hlist_head *xfrm_hash_alloc(unsigned int sz)
+{
+        struct hlist_head *n;
+
+        if (sz <= PAGE_SIZE)
+                n = kmalloc(sz, GFP_KERNEL);
+        else if (hashdist)
+                n = __vmalloc(sz, GFP_KERNEL, PAGE_KERNEL);
+        else
+                n = (struct hlist_head *)
+                        __get_free_pages(GFP_KERNEL, get_order(sz));
+
+        if (n)
+                memset(n, 0, sz);
+
+        return n;
+}
+
+void xfrm_hash_free(struct hlist_head *n, unsigned int sz)
+{
+        if (sz <= PAGE_SIZE)
+                kfree(n);
+        else if (hashdist)
+                vfree(n);
+        else
+                free_pages((unsigned long)n, get_order(sz));
+}
diff --git a/net/xfrm/xfrm_hash.h b/net/xfrm/xfrm_hash.h
new file mode 100644
index 000000000000..d3abb0b7dc62
--- /dev/null
+++ b/net/xfrm/xfrm_hash.h
@@ -0,0 +1,128 @@
+#ifndef _XFRM_HASH_H
+#define _XFRM_HASH_H
+
+#include <linux/xfrm.h>
+#include <linux/socket.h>
+
+static inline unsigned int __xfrm4_addr_hash(xfrm_address_t *addr)
+{
+        return ntohl(addr->a4);
+}
+
+static inline unsigned int __xfrm6_addr_hash(xfrm_address_t *addr)
+{
+        return ntohl(addr->a6[2] ^ addr->a6[3]);
+}
+
+static inline unsigned int __xfrm4_daddr_saddr_hash(xfrm_address_t *daddr, xfrm_address_t *saddr)
+{
+        return ntohl(daddr->a4 ^ saddr->a4);
+}
+
+static inline unsigned int __xfrm6_daddr_saddr_hash(xfrm_address_t *daddr, xfrm_address_t *saddr)
+{
+        return ntohl(daddr->a6[2] ^ daddr->a6[3] ^
+                     saddr->a6[2] ^ saddr->a6[3]);
+}
+
+static inline unsigned int __xfrm_dst_hash(xfrm_address_t *daddr, xfrm_address_t *saddr,
+                                           u32 reqid, unsigned short family,
+                                           unsigned int hmask)
+{
+        unsigned int h = family ^ reqid;
+        switch (family) {
+        case AF_INET:
+                h ^= __xfrm4_daddr_saddr_hash(daddr, saddr);
+                break;
+        case AF_INET6:
+                h ^= __xfrm6_daddr_saddr_hash(daddr, saddr);
+                break;
+        }
+        return (h ^ (h >> 16)) & hmask;
+}
+
+static inline unsigned __xfrm_src_hash(xfrm_address_t *saddr,
+                                       unsigned short family,
+                                       unsigned int hmask)
+{
+        unsigned int h = family;
+        switch (family) {
+        case AF_INET:
+                h ^= __xfrm4_addr_hash(saddr);
+                break;
+        case AF_INET6:
+                h ^= __xfrm6_addr_hash(saddr);
+                break;
+        };
+        return (h ^ (h >> 16)) & hmask;
+}
+
+static inline unsigned int
+__xfrm_spi_hash(xfrm_address_t *daddr, u32 spi, u8 proto, unsigned short family,
+                unsigned int hmask)
+{
+        unsigned int h = spi ^ proto;
+        switch (family) {
+        case AF_INET:
+                h ^= __xfrm4_addr_hash(daddr);
+                break;
+        case AF_INET6:
+                h ^= __xfrm6_addr_hash(daddr);
+                break;
+        }
+        return (h ^ (h >> 10) ^ (h >> 20)) & hmask;
+}
+
+static inline unsigned int __idx_hash(u32 index, unsigned int hmask)
+{
+        return (index ^ (index >> 8)) & hmask;
+}
+
+static inline unsigned int __sel_hash(struct xfrm_selector *sel, unsigned short family, unsigned int hmask)
+{
+        xfrm_address_t *daddr = &sel->daddr;
+        xfrm_address_t *saddr = &sel->saddr;
+        unsigned int h = 0;
+
+        switch (family) {
+        case AF_INET:
+                if (sel->prefixlen_d != 32 ||
+                    sel->prefixlen_s != 32)
+                        return hmask + 1;
+
+                h = __xfrm4_daddr_saddr_hash(daddr, saddr);
+                break;
+
+        case AF_INET6:
+                if (sel->prefixlen_d != 128 ||
+                    sel->prefixlen_s != 128)
+                        return hmask + 1;
+
+                h = __xfrm6_daddr_saddr_hash(daddr, saddr);
+                break;
+        };
+        h ^= (h >> 16);
+        return h & hmask;
+}
+
+static inline unsigned int __addr_hash(xfrm_address_t *daddr, xfrm_address_t *saddr, unsigned short family, unsigned int hmask)
+{
+        unsigned int h = 0;
+
+        switch (family) {
+        case AF_INET:
+                h = __xfrm4_daddr_saddr_hash(daddr, saddr);
+                break;
+
+        case AF_INET6:
+                h = __xfrm6_daddr_saddr_hash(daddr, saddr);
+                break;
+        };
+        h ^= (h >> 16);
+        return h & hmask;
+}
+
+extern struct hlist_head *xfrm_hash_alloc(unsigned int sz);
+extern void xfrm_hash_free(struct hlist_head *n, unsigned int sz);
+
+#endif /* _XFRM_HASH_H */
diff --git a/net/xfrm/xfrm_input.c b/net/xfrm/xfrm_input.c
index 891a6090cc09..dfc90bb1cf1f 100644
--- a/net/xfrm/xfrm_input.c
+++ b/net/xfrm/xfrm_input.c
@@ -82,8 +82,6 @@ void __init xfrm_input_init(void)
 {
         secpath_cachep = kmem_cache_create("secpath_cache",
                                            sizeof(struct sec_path),
-                                           0, SLAB_HWCACHE_ALIGN,
+                                           0, SLAB_HWCACHE_ALIGN|SLAB_PANIC,
                                            NULL, NULL);
-        if (!secpath_cachep)
-                panic("XFRM: failed to allocate secpath_cache\n");
 }
diff --git a/net/xfrm/xfrm_policy.c b/net/xfrm/xfrm_policy.c
index 3da67ca2c3ce..b6e2e79d7261 100644
--- a/net/xfrm/xfrm_policy.c
+++ b/net/xfrm/xfrm_policy.c
@@ -22,16 +22,19 @@
 #include <linux/netdevice.h>
 #include <linux/netfilter.h>
 #include <linux/module.h>
+#include <linux/cache.h>
 #include <net/xfrm.h>
 #include <net/ip.h>
 
+#include "xfrm_hash.h"
+
 DEFINE_MUTEX(xfrm_cfg_mutex);
 EXPORT_SYMBOL(xfrm_cfg_mutex);
 
 static DEFINE_RWLOCK(xfrm_policy_lock);
 
-struct xfrm_policy *xfrm_policy_list[XFRM_POLICY_MAX*2];
-EXPORT_SYMBOL(xfrm_policy_list);
+unsigned int xfrm_policy_count[XFRM_POLICY_MAX*2];
+EXPORT_SYMBOL(xfrm_policy_count);
 
 static DEFINE_RWLOCK(xfrm_policy_afinfo_lock);
 static struct xfrm_policy_afinfo *xfrm_policy_afinfo[NPROTO];
@@ -39,8 +42,7 @@ static struct xfrm_policy_afinfo *xfrm_policy_afinfo[NPROTO];
 static kmem_cache_t *xfrm_dst_cache __read_mostly;
 
 static struct work_struct xfrm_policy_gc_work;
-static struct list_head xfrm_policy_gc_list =
-        LIST_HEAD_INIT(xfrm_policy_gc_list);
+static HLIST_HEAD(xfrm_policy_gc_list);
 static DEFINE_SPINLOCK(xfrm_policy_gc_lock);
 
 static struct xfrm_policy_afinfo *xfrm_policy_get_afinfo(unsigned short family);
@@ -310,8 +312,10 @@ struct xfrm_policy *xfrm_policy_alloc(gfp_t gfp)
         policy = kzalloc(sizeof(struct xfrm_policy), gfp);
 
         if (policy) {
-                atomic_set(&policy->refcnt, 1);
+                INIT_HLIST_NODE(&policy->bydst);
+                INIT_HLIST_NODE(&policy->byidx);
                 rwlock_init(&policy->lock);
+                atomic_set(&policy->refcnt, 1);
                 init_timer(&policy->timer);
                 policy->timer.data = (unsigned long)policy;
                 policy->timer.function = xfrm_policy_timer;
@@ -357,17 +361,16 @@ static void xfrm_policy_gc_kill(struct xfrm_policy *policy)
 static void xfrm_policy_gc_task(void *data)
 {
         struct xfrm_policy *policy;
-        struct list_head *entry, *tmp;
-        struct list_head gc_list = LIST_HEAD_INIT(gc_list);
+        struct hlist_node *entry, *tmp;
+        struct hlist_head gc_list;
 
         spin_lock_bh(&xfrm_policy_gc_lock);
-        list_splice_init(&xfrm_policy_gc_list, &gc_list);
+        gc_list.first = xfrm_policy_gc_list.first;
+        INIT_HLIST_HEAD(&xfrm_policy_gc_list);
         spin_unlock_bh(&xfrm_policy_gc_lock);
 
-        list_for_each_safe(entry, tmp, &gc_list) {
-                policy = list_entry(entry, struct xfrm_policy, list);
+        hlist_for_each_entry_safe(policy, entry, tmp, &gc_list, bydst)
                 xfrm_policy_gc_kill(policy);
-        }
 }
 
 /* Rule must be locked. Release descentant resources, announce
@@ -389,70 +392,275 @@ static void xfrm_policy_kill(struct xfrm_policy *policy)
         }
 
         spin_lock(&xfrm_policy_gc_lock);
-        list_add(&policy->list, &xfrm_policy_gc_list);
+        hlist_add_head(&policy->bydst, &xfrm_policy_gc_list);
         spin_unlock(&xfrm_policy_gc_lock);
 
         schedule_work(&xfrm_policy_gc_work);
 }
 
+struct xfrm_policy_hash {
+        struct hlist_head       *table;
+        unsigned int            hmask;
+};
+
+static struct hlist_head xfrm_policy_inexact[XFRM_POLICY_MAX*2];
+static struct xfrm_policy_hash xfrm_policy_bydst[XFRM_POLICY_MAX*2] __read_mostly;
+static struct hlist_head *xfrm_policy_byidx __read_mostly;
+static unsigned int xfrm_idx_hmask __read_mostly;
+static unsigned int xfrm_policy_hashmax __read_mostly = 1 * 1024 * 1024;
+
+static inline unsigned int idx_hash(u32 index)
+{
+        return __idx_hash(index, xfrm_idx_hmask);
+}
+
+static struct hlist_head *policy_hash_bysel(struct xfrm_selector *sel, unsigned short family, int dir)
+{
+        unsigned int hmask = xfrm_policy_bydst[dir].hmask;
+        unsigned int hash = __sel_hash(sel, family, hmask);
+
+        return (hash == hmask + 1 ?
+                &xfrm_policy_inexact[dir] :
+                xfrm_policy_bydst[dir].table + hash);
+}
+
+static struct hlist_head *policy_hash_direct(xfrm_address_t *daddr, xfrm_address_t *saddr, unsigned short family, int dir)
+{
+        unsigned int hmask = xfrm_policy_bydst[dir].hmask;
+        unsigned int hash = __addr_hash(daddr, saddr, family, hmask);
+
+        return xfrm_policy_bydst[dir].table + hash;
+}
+
+static void xfrm_dst_hash_transfer(struct hlist_head *list,
+                                   struct hlist_head *ndsttable,
+                                   unsigned int nhashmask)
+{
+        struct hlist_node *entry, *tmp;
+        struct xfrm_policy *pol;
+
+        hlist_for_each_entry_safe(pol, entry, tmp, list, bydst) {
+                unsigned int h;
+
+                h = __addr_hash(&pol->selector.daddr, &pol->selector.saddr,
+                                pol->family, nhashmask);
+                hlist_add_head(&pol->bydst, ndsttable+h);
+        }
+}
+
+static void xfrm_idx_hash_transfer(struct hlist_head *list,
+                                   struct hlist_head *nidxtable,
+                                   unsigned int nhashmask)
+{
+        struct hlist_node *entry, *tmp;
+        struct xfrm_policy *pol;
+
+        hlist_for_each_entry_safe(pol, entry, tmp, list, byidx) {
+                unsigned int h;
+
+                h = __idx_hash(pol->index, nhashmask);
+                hlist_add_head(&pol->byidx, nidxtable+h);
+        }
+}
+
+static unsigned long xfrm_new_hash_mask(unsigned int old_hmask)
+{
+        return ((old_hmask + 1) << 1) - 1;
+}
+
+static void xfrm_bydst_resize(int dir)
+{
+        unsigned int hmask = xfrm_policy_bydst[dir].hmask;
+        unsigned int nhashmask = xfrm_new_hash_mask(hmask);
+        unsigned int nsize = (nhashmask + 1) * sizeof(struct hlist_head);
+        struct hlist_head *odst = xfrm_policy_bydst[dir].table;
+        struct hlist_head *ndst = xfrm_hash_alloc(nsize);
+        int i;
+
+        if (!ndst)
+                return;
+
+        write_lock_bh(&xfrm_policy_lock);
+
+        for (i = hmask; i >= 0; i--)
+                xfrm_dst_hash_transfer(odst + i, ndst, nhashmask);
+
+        xfrm_policy_bydst[dir].table = ndst;
+        xfrm_policy_bydst[dir].hmask = nhashmask;
+
+        write_unlock_bh(&xfrm_policy_lock);
+
+        xfrm_hash_free(odst, (hmask + 1) * sizeof(struct hlist_head));
+}
+
+static void xfrm_byidx_resize(int total)
+{
+        unsigned int hmask = xfrm_idx_hmask;
+        unsigned int nhashmask = xfrm_new_hash_mask(hmask);
+        unsigned int nsize = (nhashmask + 1) * sizeof(struct hlist_head);
+        struct hlist_head *oidx = xfrm_policy_byidx;
+        struct hlist_head *nidx = xfrm_hash_alloc(nsize);
+        int i;
+
+        if (!nidx)
+                return;
+
+        write_lock_bh(&xfrm_policy_lock);
+
+        for (i = hmask; i >= 0; i--)
+                xfrm_idx_hash_transfer(oidx + i, nidx, nhashmask);
+
+        xfrm_policy_byidx = nidx;
+        xfrm_idx_hmask = nhashmask;
+
+        write_unlock_bh(&xfrm_policy_lock);
+
+        xfrm_hash_free(oidx, (hmask + 1) * sizeof(struct hlist_head));
+}
+
+static inline int xfrm_bydst_should_resize(int dir, int *total)
+{
+        unsigned int cnt = xfrm_policy_count[dir];
+        unsigned int hmask = xfrm_policy_bydst[dir].hmask;
+
+        if (total)
+                *total += cnt;
+
+        if ((hmask + 1) < xfrm_policy_hashmax &&
+            cnt > hmask)
+                return 1;
+
+        return 0;
+}
+
+static inline int xfrm_byidx_should_resize(int total)
+{
+        unsigned int hmask = xfrm_idx_hmask;
+
+        if ((hmask + 1) < xfrm_policy_hashmax &&
+            total > hmask)
+                return 1;
+
+        return 0;
+}
+
+static DEFINE_MUTEX(hash_resize_mutex);
+
+static void xfrm_hash_resize(void *__unused)
+{
+        int dir, total;
+
+        mutex_lock(&hash_resize_mutex);
+
+        total = 0;
+        for (dir = 0; dir < XFRM_POLICY_MAX * 2; dir++) {
+                if (xfrm_bydst_should_resize(dir, &total))
+                        xfrm_bydst_resize(dir);
+        }
+        if (xfrm_byidx_should_resize(total))
+                xfrm_byidx_resize(total);
+
+        mutex_unlock(&hash_resize_mutex);
+}
+
+static DECLARE_WORK(xfrm_hash_work, xfrm_hash_resize, NULL);
+
 /* Generate new index... KAME seems to generate them ordered by cost
  * of an absolute inpredictability of ordering of rules. This will not pass. */
-static u32 xfrm_gen_index(int dir)
+static u32 xfrm_gen_index(u8 type, int dir)
 {
-        u32 idx;
-        struct xfrm_policy *p;
         static u32 idx_generator;
 
         for (;;) {
+                struct hlist_node *entry;
+                struct hlist_head *list;
+                struct xfrm_policy *p;
+                u32 idx;
+                int found;
+
                 idx = (idx_generator | dir);
                 idx_generator += 8;
                 if (idx == 0)
                         idx = 8;
-                for (p = xfrm_policy_list[dir]; p; p = p->next) {
-                        if (p->index == idx)
+                list = xfrm_policy_byidx + idx_hash(idx);
+                found = 0;
+                hlist_for_each_entry(p, entry, list, byidx) {
+                        if (p->index == idx) {
+                                found = 1;
                                 break;
+                        }
                 }
-                if (!p)
+                if (!found)
                         return idx;
         }
 }
 
+static inline int selector_cmp(struct xfrm_selector *s1, struct xfrm_selector *s2)
+{
+        u32 *p1 = (u32 *) s1;
+        u32 *p2 = (u32 *) s2;
+        int len = sizeof(struct xfrm_selector) / sizeof(u32);
+        int i;
+
+        for (i = 0; i < len; i++) {
+                if (p1[i] != p2[i])
+                        return 1;
+        }
+
+        return 0;
+}
+
 int xfrm_policy_insert(int dir, struct xfrm_policy *policy, int excl)
 {
-        struct xfrm_policy *pol, **p;
-        struct xfrm_policy *delpol = NULL;
-        struct xfrm_policy **newpos = NULL;
+        struct xfrm_policy *pol;
+        struct xfrm_policy *delpol;
+        struct hlist_head *chain;
+        struct hlist_node *entry, *newpos, *last;
         struct dst_entry *gc_list;
 
         write_lock_bh(&xfrm_policy_lock);
-        for (p = &xfrm_policy_list[dir]; (pol=*p)!=NULL;) {
-                if (!delpol && memcmp(&policy->selector, &pol->selector, sizeof(pol->selector)) == 0 &&
+        chain = policy_hash_bysel(&policy->selector, policy->family, dir);
+        delpol = NULL;
+        newpos = NULL;
+        last = NULL;
+        hlist_for_each_entry(pol, entry, chain, bydst) {
+                if (!delpol &&
+                    pol->type == policy->type &&
+                    !selector_cmp(&pol->selector, &policy->selector) &&
                     xfrm_sec_ctx_match(pol->security, policy->security)) {
                         if (excl) {
                                 write_unlock_bh(&xfrm_policy_lock);
                                 return -EEXIST;
                         }
-                        *p = pol->next;
                         delpol = pol;
                         if (policy->priority > pol->priority)
                                 continue;
                 } else if (policy->priority >= pol->priority) {
-                        p = &pol->next;
+                        last = &pol->bydst;
                         continue;
                 }
                 if (!newpos)
-                        newpos = p;
+                        newpos = &pol->bydst;
                 if (delpol)
                         break;
-                p = &pol->next;
+                last = &pol->bydst;
         }
+        if (!newpos)
+                newpos = last;
         if (newpos)
-                p = newpos;
+                hlist_add_after(newpos, &policy->bydst);
+        else
+                hlist_add_head(&policy->bydst, chain);
         xfrm_pol_hold(policy);
-        policy->next = *p;
-        *p = policy;
+        xfrm_policy_count[dir]++;
         atomic_inc(&flow_cache_genid);
-        policy->index = delpol ? delpol->index : xfrm_gen_index(dir);
+        if (delpol) {
+                hlist_del(&delpol->bydst);
+                hlist_del(&delpol->byidx);
+                xfrm_policy_count[dir]--;
+        }
+        policy->index = delpol ? delpol->index : xfrm_gen_index(policy->type, dir);
+        hlist_add_head(&policy->byidx, xfrm_policy_byidx+idx_hash(policy->index));
         policy->curlft.add_time = (unsigned long)xtime.tv_sec;
         policy->curlft.use_time = 0;
         if (!mod_timer(&policy->timer, jiffies + HZ))
@@ -461,10 +669,13 @@ int xfrm_policy_insert(int dir, struct xfrm_policy *policy, int excl)
 
         if (delpol)
                 xfrm_policy_kill(delpol);
+        else if (xfrm_bydst_should_resize(dir, NULL))
+                schedule_work(&xfrm_hash_work);
 
         read_lock_bh(&xfrm_policy_lock);
         gc_list = NULL;
-        for (policy = policy->next; policy; policy = policy->next) {
+        entry = &policy->bydst;
+        hlist_for_each_entry_continue(policy, entry, bydst) {
                 struct dst_entry *dst;
 
                 write_lock(&policy->lock);
@@ -493,87 +704,146 @@ int xfrm_policy_insert(int dir, struct xfrm_policy *policy, int excl)
 }
 EXPORT_SYMBOL(xfrm_policy_insert);
 
-struct xfrm_policy *xfrm_policy_bysel_ctx(int dir, struct xfrm_selector *sel,
+struct xfrm_policy *xfrm_policy_bysel_ctx(u8 type, int dir,
+                                          struct xfrm_selector *sel,
                                           struct xfrm_sec_ctx *ctx, int delete)
 {
-        struct xfrm_policy *pol, **p;
+        struct xfrm_policy *pol, *ret;
+        struct hlist_head *chain;
+        struct hlist_node *entry;
 
         write_lock_bh(&xfrm_policy_lock);
-        for (p = &xfrm_policy_list[dir]; (pol=*p)!=NULL; p = &pol->next) {
-                if ((memcmp(sel, &pol->selector, sizeof(*sel)) == 0) &&
-                    (xfrm_sec_ctx_match(ctx, pol->security))) {
+        chain = policy_hash_bysel(sel, sel->family, dir);
+        ret = NULL;
+        hlist_for_each_entry(pol, entry, chain, bydst) {
+                if (pol->type == type &&
+                    !selector_cmp(sel, &pol->selector) &&
+                    xfrm_sec_ctx_match(ctx, pol->security)) {
                         xfrm_pol_hold(pol);
-                        if (delete)
-                                *p = pol->next;
+                        if (delete) {
+                                hlist_del(&pol->bydst);
+                                hlist_del(&pol->byidx);
+                                xfrm_policy_count[dir]--;
+                        }
+                        ret = pol;
                         break;
                 }
         }
         write_unlock_bh(&xfrm_policy_lock);
 
-        if (pol && delete) {
+        if (ret && delete) {
                 atomic_inc(&flow_cache_genid);
-                xfrm_policy_kill(pol);
+                xfrm_policy_kill(ret);
         }
-        return pol;
+        return ret;
 }
 EXPORT_SYMBOL(xfrm_policy_bysel_ctx);
 
-struct xfrm_policy *xfrm_policy_byid(int dir, u32 id, int delete)
+struct xfrm_policy *xfrm_policy_byid(u8 type, int dir, u32 id, int delete)
 {
-        struct xfrm_policy *pol, **p;
+        struct xfrm_policy *pol, *ret;
+        struct hlist_head *chain;
+        struct hlist_node *entry;
 
         write_lock_bh(&xfrm_policy_lock);
-        for (p = &xfrm_policy_list[dir]; (pol=*p)!=NULL; p = &pol->next) {
-                if (pol->index == id) {
+        chain = xfrm_policy_byidx + idx_hash(id);
+        ret = NULL;
+        hlist_for_each_entry(pol, entry, chain, byidx) {
+                if (pol->type == type && pol->index == id) {
                         xfrm_pol_hold(pol);
-                        if (delete)
-                                *p = pol->next;
+                        if (delete) {
+                                hlist_del(&pol->bydst);
+                                hlist_del(&pol->byidx);
+                                xfrm_policy_count[dir]--;
+                        }
+                        ret = pol;
                         break;
                 }
         }
         write_unlock_bh(&xfrm_policy_lock);
 
-        if (pol && delete) {
+        if (ret && delete) {
                 atomic_inc(&flow_cache_genid);
-                xfrm_policy_kill(pol);
+                xfrm_policy_kill(ret);
         }
-        return pol;
+        return ret;
 }
 EXPORT_SYMBOL(xfrm_policy_byid);
 
-void xfrm_policy_flush(void)
+void xfrm_policy_flush(u8 type)
 {
-        struct xfrm_policy *xp;
         int dir;
 
         write_lock_bh(&xfrm_policy_lock);
         for (dir = 0; dir < XFRM_POLICY_MAX; dir++) {
-                while ((xp = xfrm_policy_list[dir]) != NULL) {
-                        xfrm_policy_list[dir] = xp->next;
+                struct xfrm_policy *pol;
+                struct hlist_node *entry;
+                int i;
+
+        again1:
+                hlist_for_each_entry(pol, entry,
+                                     &xfrm_policy_inexact[dir], bydst) {
+                        if (pol->type != type)
+                                continue;
+                        hlist_del(&pol->bydst);
+                        hlist_del(&pol->byidx);
                         write_unlock_bh(&xfrm_policy_lock);
 
-                        xfrm_policy_kill(xp);
+                        xfrm_policy_kill(pol);
 
                         write_lock_bh(&xfrm_policy_lock);
+                        goto again1;
                 }
+
+                for (i = xfrm_policy_bydst[dir].hmask; i >= 0; i--) {
+        again2:
+                        hlist_for_each_entry(pol, entry,
+                                             xfrm_policy_bydst[dir].table + i,
+                                             bydst) {
+                                if (pol->type != type)
+                                        continue;
+                                hlist_del(&pol->bydst);
+                                hlist_del(&pol->byidx);
+                                write_unlock_bh(&xfrm_policy_lock);
+
+                                xfrm_policy_kill(pol);
+
+                                write_lock_bh(&xfrm_policy_lock);
+                                goto again2;
+                        }
+                }
+
+                xfrm_policy_count[dir] = 0;
         }
         atomic_inc(&flow_cache_genid);
         write_unlock_bh(&xfrm_policy_lock);
 }
 EXPORT_SYMBOL(xfrm_policy_flush);
 
-int xfrm_policy_walk(int (*func)(struct xfrm_policy *, int, int, void*),
+int xfrm_policy_walk(u8 type, int (*func)(struct xfrm_policy *, int, int, void*),
                      void *data)
 {
-        struct xfrm_policy *xp;
-        int dir;
-        int count = 0;
-        int error = 0;
+        struct xfrm_policy *pol;
+        struct hlist_node *entry;
+        int dir, count, error;
 
         read_lock_bh(&xfrm_policy_lock);
+        count = 0;
         for (dir = 0; dir < 2*XFRM_POLICY_MAX; dir++) {
-                for (xp = xfrm_policy_list[dir]; xp; xp = xp->next)
-                        count++;
+                struct hlist_head *table = xfrm_policy_bydst[dir].table;
+                int i;
+
+                hlist_for_each_entry(pol, entry,
+                                     &xfrm_policy_inexact[dir], bydst) {
+                        if (pol->type == type)
+                                count++;
+                }
+                for (i = xfrm_policy_bydst[dir].hmask; i >= 0; i--) {
+                        hlist_for_each_entry(pol, entry, table + i, bydst) {
+                                if (pol->type == type)
+                                        count++;
+                        }
+                }
         }
 
         if (count == 0) {
@@ -582,13 +852,28 @@ int xfrm_policy_walk(int (*func)(struct xfrm_policy *, int, int, void*),
         }
 
         for (dir = 0; dir < 2*XFRM_POLICY_MAX; dir++) {
-                for (xp = xfrm_policy_list[dir]; xp; xp = xp->next) {
-                        error = func(xp, dir%XFRM_POLICY_MAX, --count, data);
+                struct hlist_head *table = xfrm_policy_bydst[dir].table;
+                int i;
+
+                hlist_for_each_entry(pol, entry,
+                                     &xfrm_policy_inexact[dir], bydst) {
+                        if (pol->type != type)
+                                continue;
+                        error = func(pol, dir % XFRM_POLICY_MAX, --count, data);
                         if (error)
                                 goto out;
                 }
+                for (i = xfrm_policy_bydst[dir].hmask; i >= 0; i--) {
+                        hlist_for_each_entry(pol, entry, table + i, bydst) {
+                                if (pol->type != type)
+                                        continue;
+                                error = func(pol, dir % XFRM_POLICY_MAX, --count, data);
+                                if (error)
+                                        goto out;
+                        }
+                }
         }
-
+        error = 0;
 out:
         read_unlock_bh(&xfrm_policy_lock);
         return error;
@@ -597,29 +882,79 @@ EXPORT_SYMBOL(xfrm_policy_walk);
 
 /* Find policy to apply to this flow. */
 
-static void xfrm_policy_lookup(struct flowi *fl, u32 sk_sid, u16 family, u8 dir,
-                               void **objp, atomic_t **obj_refp)
+static int xfrm_policy_match(struct xfrm_policy *pol, struct flowi *fl,
+                             u8 type, u16 family, int dir)
 {
-        struct xfrm_policy *pol;
+        struct xfrm_selector *sel = &pol->selector;
+        int match;
 
-        read_lock_bh(&xfrm_policy_lock);
-        for (pol = xfrm_policy_list[dir]; pol; pol = pol->next) {
-                struct xfrm_selector *sel = &pol->selector;
-                int match;
+        if (pol->family != family ||
+            pol->type != type)
+                return 0;
 
-                if (pol->family != family)
-                        continue;
+        match = xfrm_selector_match(sel, fl, family);
+        if (match) {
+                if (!security_xfrm_policy_lookup(pol, fl->secid, dir))
+                        return 1;
+        }
 
-                match = xfrm_selector_match(sel, fl, family);
+        return 0;
+}
 
-                if (match) {
-                        if (!security_xfrm_policy_lookup(pol, sk_sid, dir)) {
-                                xfrm_pol_hold(pol);
-                                break;
-                        }
+static struct xfrm_policy *xfrm_policy_lookup_bytype(u8 type, struct flowi *fl,
+                                                     u16 family, u8 dir)
+{
+        struct xfrm_policy *pol, *ret;
+        xfrm_address_t *daddr, *saddr;
+        struct hlist_node *entry;
+        struct hlist_head *chain;
+        u32 priority = ~0U;
+
+        daddr = xfrm_flowi_daddr(fl, family);
+        saddr = xfrm_flowi_saddr(fl, family);
+        if (unlikely(!daddr || !saddr))
+                return NULL;
+
+        read_lock_bh(&xfrm_policy_lock);
+        chain = policy_hash_direct(daddr, saddr, family, dir);
+        ret = NULL;
+        hlist_for_each_entry(pol, entry, chain, bydst) {
+                if (xfrm_policy_match(pol, fl, type, family, dir)) {
+                        ret = pol;
+                        priority = ret->priority;
+                        break;
+                }
+        }
+        chain = &xfrm_policy_inexact[dir];
+        hlist_for_each_entry(pol, entry, chain, bydst) {
+                if (xfrm_policy_match(pol, fl, type, family, dir) &&
+                    pol->priority < priority) {
+                        ret = pol;
+                        break;
                 }
         }
+        if (ret)
+                xfrm_pol_hold(ret);
         read_unlock_bh(&xfrm_policy_lock);
+
+        return ret;
+}
+
+static void xfrm_policy_lookup(struct flowi *fl, u16 family, u8 dir,
+                               void **objp, atomic_t **obj_refp)
+{
+        struct xfrm_policy *pol;
+
+#ifdef CONFIG_XFRM_SUB_POLICY
+        pol = xfrm_policy_lookup_bytype(XFRM_POLICY_TYPE_SUB, fl, family, dir);
+        if (pol)
+                goto end;
+#endif
+        pol = xfrm_policy_lookup_bytype(XFRM_POLICY_TYPE_MAIN, fl, family, dir);
+
+#ifdef CONFIG_XFRM_SUB_POLICY
+end:
+#endif
         if ((*objp = (void *) pol) != NULL)
                 *obj_refp = &pol->refcnt;
 }
@@ -641,7 +976,7 @@ static inline int policy_to_flow_dir(int dir)
         };
 }
 
-static struct xfrm_policy *xfrm_sk_policy_lookup(struct sock *sk, int dir, struct flowi *fl, u32 sk_sid)
+static struct xfrm_policy *xfrm_sk_policy_lookup(struct sock *sk, int dir, struct flowi *fl)
 {
         struct xfrm_policy *pol;
 
@@ -652,7 +987,7 @@ static struct xfrm_policy *xfrm_sk_policy_lookup(struct sock *sk, int dir, struc
                 int err = 0;
 
                 if (match)
-                  err = security_xfrm_policy_lookup(pol, sk_sid, policy_to_flow_dir(dir));
+                  err = security_xfrm_policy_lookup(pol, fl->secid, policy_to_flow_dir(dir));
 
                 if (match && !err)
                         xfrm_pol_hold(pol);
@@ -665,24 +1000,29 @@ static struct xfrm_policy *xfrm_sk_policy_lookup(struct sock *sk, int dir, struc
 
 static void __xfrm_policy_link(struct xfrm_policy *pol, int dir)
 {
-        pol->next = xfrm_policy_list[dir];
-        xfrm_policy_list[dir] = pol;
+        struct hlist_head *chain = policy_hash_bysel(&pol->selector,
+                                                     pol->family, dir);
+
+        hlist_add_head(&pol->bydst, chain);
+        hlist_add_head(&pol->byidx, xfrm_policy_byidx+idx_hash(pol->index));
+        xfrm_policy_count[dir]++;
         xfrm_pol_hold(pol);
+
+        if (xfrm_bydst_should_resize(dir, NULL))
+                schedule_work(&xfrm_hash_work);
 }
 
 static struct xfrm_policy *__xfrm_policy_unlink(struct xfrm_policy *pol,
                                                 int dir)
 {
-        struct xfrm_policy **polp;
+        if (hlist_unhashed(&pol->bydst))
+                return NULL;
 
-        for (polp = &xfrm_policy_list[dir];
-             *polp != NULL; polp = &(*polp)->next) {
-                if (*polp == pol) {
-                        *polp = pol->next;
-                        return pol;
-                }
-        }
-        return NULL;
+        hlist_del(&pol->bydst);
+        hlist_del(&pol->byidx);
+        xfrm_policy_count[dir]--;
+
+        return pol;
 }
 
 int xfrm_policy_delete(struct xfrm_policy *pol, int dir)
@@ -704,12 +1044,17 @@ int xfrm_sk_policy_insert(struct sock *sk, int dir, struct xfrm_policy *pol)
 {
         struct xfrm_policy *old_pol;
 
+#ifdef CONFIG_XFRM_SUB_POLICY
+        if (pol && pol->type != XFRM_POLICY_TYPE_MAIN)
+                return -EINVAL;
+#endif
+
         write_lock_bh(&xfrm_policy_lock);
         old_pol = sk->sk_policy[dir];
         sk->sk_policy[dir] = pol;
         if (pol) {
                 pol->curlft.add_time = (unsigned long)xtime.tv_sec;
-                pol->index = xfrm_gen_index(XFRM_POLICY_MAX+dir);
+                pol->index = xfrm_gen_index(pol->type, XFRM_POLICY_MAX+dir);
                 __xfrm_policy_link(pol, XFRM_POLICY_MAX+dir);
         }
         if (old_pol)
@@ -738,6 +1083,7 @@ static struct xfrm_policy *clone_policy(struct xfrm_policy *old, int dir)
                 newp->flags = old->flags;
                 newp->xfrm_nr = old->xfrm_nr;
                 newp->index = old->index;
+                newp->type = old->type;
                 memcpy(newp->xfrm_vec, old->xfrm_vec,
                        newp->xfrm_nr*sizeof(struct xfrm_tmpl));
                 write_lock_bh(&xfrm_policy_lock);
@@ -761,17 +1107,32 @@ int __xfrm_sk_clone_policy(struct sock *sk)
         return 0;
 }
 
+static int
+xfrm_get_saddr(xfrm_address_t *local, xfrm_address_t *remote,
+               unsigned short family)
+{
+        int err;
+        struct xfrm_policy_afinfo *afinfo = xfrm_policy_get_afinfo(family);
+
+        if (unlikely(afinfo == NULL))
+                return -EINVAL;
+        err = afinfo->get_saddr(local, remote);
+        xfrm_policy_put_afinfo(afinfo);
+        return err;
+}
+
 /* Resolve list of templates for the flow, given policy. */
 
 static int
-xfrm_tmpl_resolve(struct xfrm_policy *policy, struct flowi *fl,
-                  struct xfrm_state **xfrm,
-                  unsigned short family)
+xfrm_tmpl_resolve_one(struct xfrm_policy *policy, struct flowi *fl,
+                      struct xfrm_state **xfrm,
+                      unsigned short family)
 {
         int nx;
         int i, error;
         xfrm_address_t *daddr = xfrm_flowi_daddr(fl, family);
         xfrm_address_t *saddr = xfrm_flowi_saddr(fl, family);
+        xfrm_address_t tmp;
 
         for (nx=0, i = 0; i < policy->xfrm_nr; i++) {
                 struct xfrm_state *x;
@@ -779,9 +1140,15 @@ xfrm_tmpl_resolve(struct xfrm_policy *policy, struct flowi *fl,
                 xfrm_address_t *local  = saddr;
                 struct xfrm_tmpl *tmpl = &policy->xfrm_vec[i];
 
-                if (tmpl->mode) {
+                if (tmpl->mode == XFRM_MODE_TUNNEL) {
                         remote = &tmpl->id.daddr;
                         local = &tmpl->saddr;
+                        if (xfrm_addr_any(local, family)) {
+                                error = xfrm_get_saddr(&tmp, remote, family);
+                                if (error)
+                                        goto fail;
+                                local = &tmp;
+                        }
                 }
 
                 x = xfrm_state_find(remote, local, fl, tmpl, policy, &error, family);
@@ -809,6 +1176,45 @@ fail:
         return error;
 }
 
+static int
+xfrm_tmpl_resolve(struct xfrm_policy **pols, int npols, struct flowi *fl,
+                  struct xfrm_state **xfrm,
+                  unsigned short family)
+{
+        struct xfrm_state *tp[XFRM_MAX_DEPTH];
+        struct xfrm_state **tpp = (npols > 1) ? tp : xfrm;
+        int cnx = 0;
+        int error;
+        int ret;
+        int i;
+
+        for (i = 0; i < npols; i++) {
+                if (cnx + pols[i]->xfrm_nr >= XFRM_MAX_DEPTH) {
+                        error = -ENOBUFS;
+                        goto fail;
+                }
+
+                ret = xfrm_tmpl_resolve_one(pols[i], fl, &tpp[cnx], family);
+                if (ret < 0) {
+                        error = ret;
+                        goto fail;
+                } else
+                        cnx += ret;
+        }
+
+        /* found states are sorted for outbound processing */
+        if (npols > 1)
+                xfrm_state_sort(xfrm, tpp, cnx, family);
+
+        return cnx;
+
+ fail:
+        for (cnx--; cnx>=0; cnx--)
+                xfrm_state_put(tpp[cnx]);
+        return error;
+
+}
+
 /* Check that the bundle accepts the flow and its components are
  * still valid.
  */
@@ -855,6 +1261,11 @@ int xfrm_lookup(struct dst_entry **dst_p, struct flowi *fl,
                 struct sock *sk, int flags)
 {
         struct xfrm_policy *policy;
+        struct xfrm_policy *pols[XFRM_POLICY_TYPE_MAX];
+        int npols;
+        int pol_dead;
+        int xfrm_nr;
+        int pi;
         struct xfrm_state *xfrm[XFRM_MAX_DEPTH];
         struct dst_entry *dst, *dst_orig = *dst_p;
         int nx = 0;
@@ -862,19 +1273,26 @@ int xfrm_lookup(struct dst_entry **dst_p, struct flowi *fl,
         u32 genid;
         u16 family;
         u8 dir = policy_to_flow_dir(XFRM_POLICY_OUT);
-        u32 sk_sid = security_sk_sid(sk, fl, dir);
+
 restart:
         genid = atomic_read(&flow_cache_genid);
         policy = NULL;
+        for (pi = 0; pi < ARRAY_SIZE(pols); pi++)
+                pols[pi] = NULL;
+        npols = 0;
+        pol_dead = 0;
+        xfrm_nr = 0;
+
         if (sk && sk->sk_policy[1])
-                policy = xfrm_sk_policy_lookup(sk, XFRM_POLICY_OUT, fl, sk_sid);
+                policy = xfrm_sk_policy_lookup(sk, XFRM_POLICY_OUT, fl);
 
         if (!policy) {
                 /* To accelerate a bit...  */
-                if ((dst_orig->flags & DST_NOXFRM) || !xfrm_policy_list[XFRM_POLICY_OUT])
+                if ((dst_orig->flags & DST_NOXFRM) ||
+                    !xfrm_policy_count[XFRM_POLICY_OUT])
                         return 0;
 
-                policy = flow_cache_lookup(fl, sk_sid, dst_orig->ops->family,
+                policy = flow_cache_lookup(fl, dst_orig->ops->family,
                                            dir, xfrm_policy_lookup);
         }
 
@@ -883,6 +1301,9 @@ restart:
 
         family = dst_orig->ops->family;
         policy->curlft.use_time = (unsigned long)xtime.tv_sec;
+        pols[0] = policy;
+        npols ++;
+        xfrm_nr += pols[0]->xfrm_nr;
 
         switch (policy->action) {
         case XFRM_POLICY_BLOCK:
@@ -891,11 +1312,13 @@ restart:
                 goto error;
 
         case XFRM_POLICY_ALLOW:
+#ifndef CONFIG_XFRM_SUB_POLICY
                 if (policy->xfrm_nr == 0) {
                         /* Flow passes not transformed. */
                         xfrm_pol_put(policy);
                         return 0;
                 }
+#endif
 
                 /* Try to find matching bundle.
                  *
@@ -911,7 +1334,36 @@ restart:
                 if (dst)
                         break;
 
-                nx = xfrm_tmpl_resolve(policy, fl, xfrm, family);
+#ifdef CONFIG_XFRM_SUB_POLICY
+                if (pols[0]->type != XFRM_POLICY_TYPE_MAIN) {
+                        pols[1] = xfrm_policy_lookup_bytype(XFRM_POLICY_TYPE_MAIN,
+                                                            fl, family,
+                                                            XFRM_POLICY_OUT);
+                        if (pols[1]) {
+                                if (pols[1]->action == XFRM_POLICY_BLOCK) {
+                                        err = -EPERM;
+                                        goto error;
+                                }
+                                npols ++;
+                                xfrm_nr += pols[1]->xfrm_nr;
+                        }
+                }
+
+                /*
+                 * Because neither flowi nor bundle information knows about
+                 * transformation template size. On more than one policy usage
+                 * we can realize whether all of them is bypass or not after
+                 * they are searched. See above not-transformed bypass
+                 * is surrounded by non-sub policy configuration, too.
+                 */
+                if (xfrm_nr == 0) {
+                        /* Flow passes not transformed. */
+                        xfrm_pols_put(pols, npols);
+                        return 0;
+                }
+
+#endif
+                nx = xfrm_tmpl_resolve(pols, npols, fl, xfrm, family);
 
                 if (unlikely(nx<0)) {
                         err = nx;
@@ -924,7 +1376,7 @@ restart:
                                 set_current_state(TASK_RUNNING);
                                 remove_wait_queue(&km_waitq, &wait);
 
-                                nx = xfrm_tmpl_resolve(policy, fl, xfrm, family);
+                                nx = xfrm_tmpl_resolve(pols, npols, fl, xfrm, family);
 
                                 if (nx == -EAGAIN && signal_pending(current)) {
                                         err = -ERESTART;
@@ -932,7 +1384,7 @@ restart:
                                 }
                                 if (nx == -EAGAIN ||
                                     genid != atomic_read(&flow_cache_genid)) {
-                                        xfrm_pol_put(policy);
+                                        xfrm_pols_put(pols, npols);
                                         goto restart;
                                 }
                                 err = nx;
@@ -942,7 +1394,7 @@ restart:
                 }
                 if (nx == 0) {
                         /* Flow passes not transformed. */
-                        xfrm_pol_put(policy);
+                        xfrm_pols_put(pols, npols);
                         return 0;
                 }
 
@@ -956,8 +1408,14 @@ restart:
                         goto error;
                 }
 
+                for (pi = 0; pi < npols; pi++) {
+                        read_lock_bh(&pols[pi]->lock);
+                        pol_dead |= pols[pi]->dead;
+                        read_unlock_bh(&pols[pi]->lock);
+                }
+
                 write_lock_bh(&policy->lock);
-                if (unlikely(policy->dead || stale_bundle(dst))) {
+                if (unlikely(pol_dead || stale_bundle(dst))) {
                         /* Wow! While we worked on resolving, this
                          * policy has gone. Retry. It is not paranoia,
                          * we just cannot enlist new bundle to dead object.
@@ -977,17 +1435,34 @@ restart:
         }
         *dst_p = dst;
         dst_release(dst_orig);
-        xfrm_pol_put(policy);
+        xfrm_pols_put(pols, npols);
         return 0;
 
 error:
         dst_release(dst_orig);
-        xfrm_pol_put(policy);
+        xfrm_pols_put(pols, npols);
         *dst_p = NULL;
         return err;
 }
 EXPORT_SYMBOL(xfrm_lookup);
 
+static inline int
+xfrm_secpath_reject(int idx, struct sk_buff *skb, struct flowi *fl)
+{
+        struct xfrm_state *x;
+        int err;
+
+        if (!skb->sp || idx < 0 || idx >= skb->sp->len)
+                return 0;
+        x = skb->sp->xvec[idx];
+        if (!x->type->reject)
+                return 0;
+        xfrm_state_hold(x);
+        err = x->type->reject(x, skb, fl);
+        xfrm_state_put(x);
+        return err;
+}
+
 /* When skb is transformed back to its "native" form, we have to
  * check policy restrictions. At the moment we make this in maximally
  * stupid way. Shame on me. :-) Of course, connected sockets must
@@ -1004,10 +1479,19 @@ xfrm_state_ok(struct xfrm_tmpl *tmpl, struct xfrm_state *x,
                 (x->id.spi == tmpl->id.spi || !tmpl->id.spi) &&
                 (x->props.reqid == tmpl->reqid || !tmpl->reqid) &&
                 x->props.mode == tmpl->mode &&
-                (tmpl->aalgos & (1<<x->props.aalgo)) &&
-                !(x->props.mode && xfrm_state_addr_cmp(tmpl, x, family));
+                ((tmpl->aalgos & (1<<x->props.aalgo)) ||
+                 !(xfrm_id_proto_match(tmpl->id.proto, IPSEC_PROTO_ANY))) &&
+                !(x->props.mode != XFRM_MODE_TRANSPORT &&
+                  xfrm_state_addr_cmp(tmpl, x, family));
 }
 
+/*
+ * 0 or more than 0 is returned when validation is succeeded (either bypass
+ * because of optional transport mode, or next index of the mathced secpath
+ * state with the template.
+ * -1 is returned when no matching template is found.
+ * Otherwise "-2 - errored_index" is returned.
+ */
 static inline int
 xfrm_policy_ok(struct xfrm_tmpl *tmpl, struct sec_path *sp, int start,
                unsigned short family)
@@ -1015,15 +1499,18 @@ xfrm_policy_ok(struct xfrm_tmpl *tmpl, struct sec_path *sp, int start,
         int idx = start;
 
         if (tmpl->optional) {
-                if (!tmpl->mode)
+                if (tmpl->mode == XFRM_MODE_TRANSPORT)
                         return start;
         } else
                 start = -1;
         for (; idx < sp->len; idx++) {
                 if (xfrm_state_ok(tmpl, sp->xvec[idx], family))
                         return ++idx;
-                if (sp->xvec[idx]->props.mode)
+                if (sp->xvec[idx]->props.mode != XFRM_MODE_TRANSPORT) {
+                        if (start == -1)
+                                start = -2-idx;
                         break;
+                }
         }
         return start;
 }
@@ -1032,21 +1519,25 @@ int
 xfrm_decode_session(struct sk_buff *skb, struct flowi *fl, unsigned short family)
 {
         struct xfrm_policy_afinfo *afinfo = xfrm_policy_get_afinfo(family);
+        int err;
 
         if (unlikely(afinfo == NULL))
                 return -EAFNOSUPPORT;
 
         afinfo->decode_session(skb, fl);
+        err = security_xfrm_decode_session(skb, &fl->secid);
         xfrm_policy_put_afinfo(afinfo);
-        return 0;
+        return err;
 }
 EXPORT_SYMBOL(xfrm_decode_session);
 
-static inline int secpath_has_tunnel(struct sec_path *sp, int k)
+static inline int secpath_has_nontransport(struct sec_path *sp, int k, int *idxp)
 {
         for (; k < sp->len; k++) {
-                if (sp->xvec[k]->props.mode)
+                if (sp->xvec[k]->props.mode != XFRM_MODE_TRANSPORT) {
+                        *idxp = k;
                         return 1;
+                }
         }
 
         return 0;
@@ -1056,16 +1547,18 @@ int __xfrm_policy_check(struct sock *sk, int dir, struct sk_buff *skb,
                         unsigned short family)
 {
         struct xfrm_policy *pol;
+        struct xfrm_policy *pols[XFRM_POLICY_TYPE_MAX];
+        int npols = 0;
+        int xfrm_nr;
+        int pi;
         struct flowi fl;
         u8 fl_dir = policy_to_flow_dir(dir);
-        u32 sk_sid;
+        int xerr_idx = -1;
 
         if (xfrm_decode_session(skb, &fl, family) < 0)
                 return 0;
         nf_nat_decode_session(skb, &fl, family);
 
-        sk_sid = security_sk_sid(sk, &fl, fl_dir);
-
         /* First, check used SA against their selectors. */
         if (skb->sp) {
                 int i;
@@ -1079,46 +1572,90 @@ int __xfrm_policy_check(struct sock *sk, int dir, struct sk_buff *skb,
 
         pol = NULL;
         if (sk && sk->sk_policy[dir])
-                pol = xfrm_sk_policy_lookup(sk, dir, &fl, sk_sid);
+                pol = xfrm_sk_policy_lookup(sk, dir, &fl);
 
         if (!pol)
-                pol = flow_cache_lookup(&fl, sk_sid, family, fl_dir,
+                pol = flow_cache_lookup(&fl, family, fl_dir,
                                         xfrm_policy_lookup);
 
-        if (!pol)
-                return !skb->sp || !secpath_has_tunnel(skb->sp, 0);
+        if (!pol) {
+                if (skb->sp && secpath_has_nontransport(skb->sp, 0, &xerr_idx)) {
+                        xfrm_secpath_reject(xerr_idx, skb, &fl);
+                        return 0;
+                }
+                return 1;
+        }
 
         pol->curlft.use_time = (unsigned long)xtime.tv_sec;
 
+        pols[0] = pol;
+        npols ++;
+#ifdef CONFIG_XFRM_SUB_POLICY
+        if (pols[0]->type != XFRM_POLICY_TYPE_MAIN) {
+                pols[1] = xfrm_policy_lookup_bytype(XFRM_POLICY_TYPE_MAIN,
+                                                    &fl, family,
+                                                    XFRM_POLICY_IN);
+                if (pols[1]) {
+                        pols[1]->curlft.use_time = (unsigned long)xtime.tv_sec;
+                        npols ++;
+                }
+        }
+#endif
+
         if (pol->action == XFRM_POLICY_ALLOW) {
                 struct sec_path *sp;
                 static struct sec_path dummy;
+                struct xfrm_tmpl *tp[XFRM_MAX_DEPTH];
+                struct xfrm_tmpl *stp[XFRM_MAX_DEPTH];
+                struct xfrm_tmpl **tpp = tp;
+                int ti = 0;
                 int i, k;
 
                 if ((sp = skb->sp) == NULL)
                         sp = &dummy;
 
+                for (pi = 0; pi < npols; pi++) {
+                        if (pols[pi] != pol &&
+                            pols[pi]->action != XFRM_POLICY_ALLOW)
+                                goto reject;
+                        if (ti + pols[pi]->xfrm_nr >= XFRM_MAX_DEPTH)
+                                goto reject_error;
+                        for (i = 0; i < pols[pi]->xfrm_nr; i++)
+                                tpp[ti++] = &pols[pi]->xfrm_vec[i];
+                }
+                xfrm_nr = ti;
+                if (npols > 1) {
+                        xfrm_tmpl_sort(stp, tpp, xfrm_nr, family);
+                        tpp = stp;
+                }
+
                 /* For each tunnel xfrm, find the first matching tmpl.
                  * For each tmpl before that, find corresponding xfrm.
                  * Order is _important_. Later we will implement
                  * some barriers, but at the moment barriers
                  * are implied between each two transformations.
                  */
-                for (i = pol->xfrm_nr-1, k = 0; i >= 0; i--) {
-                        k = xfrm_policy_ok(pol->xfrm_vec+i, sp, k, family);
-                        if (k < 0)
+                for (i = xfrm_nr-1, k = 0; i >= 0; i--) {
+                        k = xfrm_policy_ok(tpp[i], sp, k, family);
+                        if (k < 0) {
+                                if (k < -1)
+                                        /* "-2 - errored_index" returned */
+                                        xerr_idx = -(2+k);
                                 goto reject;
+                        }
                 }
 
-                if (secpath_has_tunnel(sp, k))
+                if (secpath_has_nontransport(sp, k, &xerr_idx))
                         goto reject;
 
-                xfrm_pol_put(pol);
+                xfrm_pols_put(pols, npols);
                 return 1;
         }
 
 reject:
-        xfrm_pol_put(pol);
+        xfrm_secpath_reject(xerr_idx, skb, &fl);
+reject_error:
+        xfrm_pols_put(pols, npols);
         return 0;
 }
 EXPORT_SYMBOL(__xfrm_policy_check);
@@ -1166,7 +1703,7 @@ static struct dst_entry *xfrm_dst_check(struct dst_entry *dst, u32 cookie)
 
 static int stale_bundle(struct dst_entry *dst)
 {
-        return !xfrm_bundle_ok((struct xfrm_dst *)dst, NULL, AF_UNSPEC);
+        return !xfrm_bundle_ok((struct xfrm_dst *)dst, NULL, AF_UNSPEC, 0);
 }
 
 void xfrm_dst_ifdown(struct dst_entry *dst, struct net_device *dev)
@@ -1196,33 +1733,50 @@ static struct dst_entry *xfrm_negative_advice(struct dst_entry *dst)
         return dst;
 }
 
+static void prune_one_bundle(struct xfrm_policy *pol, int (*func)(struct dst_entry *), struct dst_entry **gc_list_p)
+{
+        struct dst_entry *dst, **dstp;
+
+        write_lock(&pol->lock);
+        dstp = &pol->bundles;
+        while ((dst=*dstp) != NULL) {
+                if (func(dst)) {
+                        *dstp = dst->next;
+                        dst->next = *gc_list_p;
+                        *gc_list_p = dst;
+                } else {
+                        dstp = &dst->next;
+                }
+        }
+        write_unlock(&pol->lock);
+}
+
 static void xfrm_prune_bundles(int (*func)(struct dst_entry *))
 {
-        int i;
-        struct xfrm_policy *pol;
-        struct dst_entry *dst, **dstp, *gc_list = NULL;
+        struct dst_entry *gc_list = NULL;
+        int dir;
 
         read_lock_bh(&xfrm_policy_lock);
-        for (i=0; i<2*XFRM_POLICY_MAX; i++) {
-                for (pol = xfrm_policy_list[i]; pol; pol = pol->next) {
-                        write_lock(&pol->lock);
-                        dstp = &pol->bundles;
-                        while ((dst=*dstp) != NULL) {
-                                if (func(dst)) {
-                                        *dstp = dst->next;
-                                        dst->next = gc_list;
-                                        gc_list = dst;
-                                } else {
-                                        dstp = &dst->next;
-                                }
-                        }
-                        write_unlock(&pol->lock);
+        for (dir = 0; dir < XFRM_POLICY_MAX * 2; dir++) {
+                struct xfrm_policy *pol;
+                struct hlist_node *entry;
+                struct hlist_head *table;
+                int i;
+
+                hlist_for_each_entry(pol, entry,
+                                     &xfrm_policy_inexact[dir], bydst)
+                        prune_one_bundle(pol, func, &gc_list);
+
+                table = xfrm_policy_bydst[dir].table;
+                for (i = xfrm_policy_bydst[dir].hmask; i >= 0; i--) {
+                        hlist_for_each_entry(pol, entry, table + i, bydst)
+                                prune_one_bundle(pol, func, &gc_list);
                 }
         }
         read_unlock_bh(&xfrm_policy_lock);
 
         while (gc_list) {
-                dst = gc_list;
+                struct dst_entry *dst = gc_list;
                 gc_list = dst->next;
                 dst_free(dst);
         }
@@ -1238,22 +1792,12 @@ static void __xfrm_garbage_collect(void)
         xfrm_prune_bundles(unused_bundle);
 }
 
-int xfrm_flush_bundles(void)
+static int xfrm_flush_bundles(void)
 {
         xfrm_prune_bundles(stale_bundle);
         return 0;
 }
 
-static int always_true(struct dst_entry *dst)
-{
-        return 1;
-}
-
-void xfrm_flush_all_bundles(void)
-{
-        xfrm_prune_bundles(always_true);
-}
-
 void xfrm_init_pmtu(struct dst_entry *dst)
 {
         do {
@@ -1281,7 +1825,7 @@ EXPORT_SYMBOL(xfrm_init_pmtu);
  * still valid.
  */
 
-int xfrm_bundle_ok(struct xfrm_dst *first, struct flowi *fl, int family)
+int xfrm_bundle_ok(struct xfrm_dst *first, struct flowi *fl, int family, int strict)
 {
         struct dst_entry *dst = &first->u.dst;
         struct xfrm_dst *last;
@@ -1298,8 +1842,16 @@ int xfrm_bundle_ok(struct xfrm_dst *first, struct flowi *fl, int family)
 
                 if (fl && !xfrm_selector_match(&dst->xfrm->sel, fl, family))
                         return 0;
+                if (fl && !security_xfrm_flow_state_match(fl, dst->xfrm))
+                        return 0;
                 if (dst->xfrm->km.state != XFRM_STATE_VALID)
                         return 0;
+                if (xdst->genid != dst->xfrm->genid)
+                        return 0;
+
+                if (strict && fl && dst->xfrm->props.mode != XFRM_MODE_TUNNEL &&
+                    !xfrm_state_addr_flow_check(dst->xfrm, fl, family))
+                        return 0;
 
                 mtu = dst_mtu(dst->child);
                 if (xdst->child_mtu_cached != mtu) {
@@ -1448,12 +2000,33 @@ static struct notifier_block xfrm_dev_notifier = {
 
 static void __init xfrm_policy_init(void)
 {
+        unsigned int hmask, sz;
+        int dir;
+
         xfrm_dst_cache = kmem_cache_create("xfrm_dst_cache",
                                            sizeof(struct xfrm_dst),
-                                           0, SLAB_HWCACHE_ALIGN,
+                                           0, SLAB_HWCACHE_ALIGN|SLAB_PANIC,
                                            NULL, NULL);
-        if (!xfrm_dst_cache)
-                panic("XFRM: failed to allocate xfrm_dst_cache\n");
+
+        hmask = 8 - 1;
+        sz = (hmask+1) * sizeof(struct hlist_head);
+
+        xfrm_policy_byidx = xfrm_hash_alloc(sz);
+        xfrm_idx_hmask = hmask;
+        if (!xfrm_policy_byidx)
+                panic("XFRM: failed to allocate byidx hash\n");
+
+        for (dir = 0; dir < XFRM_POLICY_MAX * 2; dir++) {
+                struct xfrm_policy_hash *htab;
+
+                INIT_HLIST_HEAD(&xfrm_policy_inexact[dir]);
+
+                htab = &xfrm_policy_bydst[dir];
+                htab->table = xfrm_hash_alloc(sz);
+                htab->hmask = hmask;
+                if (!htab->table)
+                        panic("XFRM: failed to allocate bydst hash\n");
+        }
 
         INIT_WORK(&xfrm_policy_gc_work, xfrm_policy_gc_task, NULL);
         register_netdevice_notifier(&xfrm_dev_notifier);
diff --git a/net/xfrm/xfrm_state.c b/net/xfrm/xfrm_state.c
index 0021aad5db43..9f63edd39346 100644
--- a/net/xfrm/xfrm_state.c
+++ b/net/xfrm/xfrm_state.c
@@ -18,8 +18,11 @@
 #include <linux/pfkeyv2.h>
 #include <linux/ipsec.h>
 #include <linux/module.h>
+#include <linux/cache.h>
 #include <asm/uaccess.h>
 
+#include "xfrm_hash.h"
+
 struct sock *xfrm_nl;
 EXPORT_SYMBOL(xfrm_nl);
 
@@ -32,7 +35,7 @@ EXPORT_SYMBOL(sysctl_xfrm_aevent_rseqth);
 /* Each xfrm_state may be linked to two tables:
 
    1. Hash table by (spi,daddr,ah/esp) to find SA by SPI. (input,ctl)
-   2. Hash table by daddr to find what SAs exist for given
+   2. Hash table by (daddr,family,reqid) to find what SAs exist for given
       destination/tunnel endpoint. (output)
  */
 
@@ -44,8 +47,123 @@ static DEFINE_SPINLOCK(xfrm_state_lock);
  * Main use is finding SA after policy selected tunnel or transport mode.
  * Also, it can be used by ah/esp icmp error handler to find offending SA.
  */
-static struct list_head xfrm_state_bydst[XFRM_DST_HSIZE];
-static struct list_head xfrm_state_byspi[XFRM_DST_HSIZE];
+static struct hlist_head *xfrm_state_bydst __read_mostly;
+static struct hlist_head *xfrm_state_bysrc __read_mostly;
+static struct hlist_head *xfrm_state_byspi __read_mostly;
+static unsigned int xfrm_state_hmask __read_mostly;
+static unsigned int xfrm_state_hashmax __read_mostly = 1 * 1024 * 1024;
+static unsigned int xfrm_state_num;
+static unsigned int xfrm_state_genid;
+
+static inline unsigned int xfrm_dst_hash(xfrm_address_t *daddr,
+                                         xfrm_address_t *saddr,
+                                         u32 reqid,
+                                         unsigned short family)
+{
+        return __xfrm_dst_hash(daddr, saddr, reqid, family, xfrm_state_hmask);
+}
+
+static inline unsigned int xfrm_src_hash(xfrm_address_t *addr,
+                                         unsigned short family)
+{
+        return __xfrm_src_hash(addr, family, xfrm_state_hmask);
+}
+
+static inline unsigned int
+xfrm_spi_hash(xfrm_address_t *daddr, u32 spi, u8 proto, unsigned short family)
+{
+        return __xfrm_spi_hash(daddr, spi, proto, family, xfrm_state_hmask);
+}
+
+static void xfrm_hash_transfer(struct hlist_head *list,
+                               struct hlist_head *ndsttable,
+                               struct hlist_head *nsrctable,
+                               struct hlist_head *nspitable,
+                               unsigned int nhashmask)
+{
+        struct hlist_node *entry, *tmp;
+        struct xfrm_state *x;
+
+        hlist_for_each_entry_safe(x, entry, tmp, list, bydst) {
+                unsigned int h;
+
+                h = __xfrm_dst_hash(&x->id.daddr, &x->props.saddr,
+                                    x->props.reqid, x->props.family,
+                                    nhashmask);
+                hlist_add_head(&x->bydst, ndsttable+h);
+
+                h = __xfrm_src_hash(&x->props.saddr, x->props.family,
+                                    nhashmask);
+                hlist_add_head(&x->bysrc, nsrctable+h);
+
+                h = __xfrm_spi_hash(&x->id.daddr, x->id.spi, x->id.proto,
+                                    x->props.family, nhashmask);
+                hlist_add_head(&x->byspi, nspitable+h);
+        }
+}
+
+static unsigned long xfrm_hash_new_size(void)
+{
+        return ((xfrm_state_hmask + 1) << 1) *
+                sizeof(struct hlist_head);
+}
+
+static DEFINE_MUTEX(hash_resize_mutex);
+
+static void xfrm_hash_resize(void *__unused)
+{
+        struct hlist_head *ndst, *nsrc, *nspi, *odst, *osrc, *ospi;
+        unsigned long nsize, osize;
+        unsigned int nhashmask, ohashmask;
+        int i;
+
+        mutex_lock(&hash_resize_mutex);
+
+        nsize = xfrm_hash_new_size();
+        ndst = xfrm_hash_alloc(nsize);
+        if (!ndst)
+                goto out_unlock;
+        nsrc = xfrm_hash_alloc(nsize);
+        if (!nsrc) {
+                xfrm_hash_free(ndst, nsize);
+                goto out_unlock;
+        }
+        nspi = xfrm_hash_alloc(nsize);
+        if (!nspi) {
+                xfrm_hash_free(ndst, nsize);
+                xfrm_hash_free(nsrc, nsize);
+                goto out_unlock;
+        }
+
+        spin_lock_bh(&xfrm_state_lock);
+
+        nhashmask = (nsize / sizeof(struct hlist_head)) - 1U;
+        for (i = xfrm_state_hmask; i >= 0; i--)
+                xfrm_hash_transfer(xfrm_state_bydst+i, ndst, nsrc, nspi,
+                                   nhashmask);
+
+        odst = xfrm_state_bydst;
+        osrc = xfrm_state_bysrc;
+        ospi = xfrm_state_byspi;
+        ohashmask = xfrm_state_hmask;
+
+        xfrm_state_bydst = ndst;
+        xfrm_state_bysrc = nsrc;
+        xfrm_state_byspi = nspi;
+        xfrm_state_hmask = nhashmask;
+
+        spin_unlock_bh(&xfrm_state_lock);
+
+        osize = (ohashmask + 1) * sizeof(struct hlist_head);
+        xfrm_hash_free(odst, osize);
+        xfrm_hash_free(osrc, osize);
+        xfrm_hash_free(ospi, osize);
+
+out_unlock:
+        mutex_unlock(&hash_resize_mutex);
+}
+
+static DECLARE_WORK(xfrm_hash_work, xfrm_hash_resize, NULL);
 
 DECLARE_WAIT_QUEUE_HEAD(km_waitq);
 EXPORT_SYMBOL(km_waitq);
@@ -54,11 +172,9 @@ static DEFINE_RWLOCK(xfrm_state_afinfo_lock);
 static struct xfrm_state_afinfo *xfrm_state_afinfo[NPROTO];
 
 static struct work_struct xfrm_state_gc_work;
-static struct list_head xfrm_state_gc_list = LIST_HEAD_INIT(xfrm_state_gc_list);
+static HLIST_HEAD(xfrm_state_gc_list);
 static DEFINE_SPINLOCK(xfrm_state_gc_lock);
 
-static int xfrm_state_gc_flush_bundles;
-
 int __xfrm_state_delete(struct xfrm_state *x);
 
 static struct xfrm_state_afinfo *xfrm_state_get_afinfo(unsigned short family);
@@ -69,14 +185,13 @@ void km_state_expired(struct xfrm_state *x, int hard, u32 pid);
 
 static void xfrm_state_gc_destroy(struct xfrm_state *x)
 {
-        if (del_timer(&x->timer))
-                BUG();
-        if (del_timer(&x->rtimer))
-                BUG();
+        del_timer_sync(&x->timer);
+        del_timer_sync(&x->rtimer);
         kfree(x->aalg);
         kfree(x->ealg);
         kfree(x->calg);
         kfree(x->encap);
+        kfree(x->coaddr);
         if (x->mode)
                 xfrm_put_mode(x->mode);
         if (x->type) {
@@ -90,22 +205,17 @@ static void xfrm_state_gc_destroy(struct xfrm_state *x)
 static void xfrm_state_gc_task(void *data)
 {
         struct xfrm_state *x;
-        struct list_head *entry, *tmp;
-        struct list_head gc_list = LIST_HEAD_INIT(gc_list);
-
-        if (xfrm_state_gc_flush_bundles) {
-                xfrm_state_gc_flush_bundles = 0;
-                xfrm_flush_bundles();
-        }
+        struct hlist_node *entry, *tmp;
+        struct hlist_head gc_list;
 
         spin_lock_bh(&xfrm_state_gc_lock);
-        list_splice_init(&xfrm_state_gc_list, &gc_list);
+        gc_list.first = xfrm_state_gc_list.first;
+        INIT_HLIST_HEAD(&xfrm_state_gc_list);
         spin_unlock_bh(&xfrm_state_gc_lock);
 
-        list_for_each_safe(entry, tmp, &gc_list) {
-                x = list_entry(entry, struct xfrm_state, bydst);
+        hlist_for_each_entry_safe(x, entry, tmp, &gc_list, bydst)
                 xfrm_state_gc_destroy(x);
-        }
+
         wake_up(&km_waitq);
 }
 
@@ -168,9 +278,9 @@ static void xfrm_timer_handler(unsigned long data)
         if (warn)
                 km_state_expired(x, 0, 0);
 resched:
-        if (next != LONG_MAX &&
-            !mod_timer(&x->timer, jiffies + make_jiffies(next)))
-                xfrm_state_hold(x);
+        if (next != LONG_MAX)
+                mod_timer(&x->timer, jiffies + make_jiffies(next));
+
         goto out;
 
 expired:
@@ -185,7 +295,6 @@ expired:
 
 out:
         spin_unlock(&x->lock);
-        xfrm_state_put(x);
 }
 
 static void xfrm_replay_timer_handler(unsigned long data);
@@ -199,8 +308,9 @@ struct xfrm_state *xfrm_state_alloc(void)
         if (x) {
                 atomic_set(&x->refcnt, 1);
                 atomic_set(&x->tunnel_users, 0);
-                INIT_LIST_HEAD(&x->bydst);
-                INIT_LIST_HEAD(&x->byspi);
+                INIT_HLIST_NODE(&x->bydst);
+                INIT_HLIST_NODE(&x->bysrc);
+                INIT_HLIST_NODE(&x->byspi);
                 init_timer(&x->timer);
                 x->timer.function = xfrm_timer_handler;
                 x->timer.data     = (unsigned long)x;
@@ -225,7 +335,7 @@ void __xfrm_state_destroy(struct xfrm_state *x)
         BUG_TRAP(x->km.state == XFRM_STATE_DEAD);
 
         spin_lock_bh(&xfrm_state_gc_lock);
-        list_add(&x->bydst, &xfrm_state_gc_list);
+        hlist_add_head(&x->bydst, &xfrm_state_gc_list);
         spin_unlock_bh(&xfrm_state_gc_lock);
         schedule_work(&xfrm_state_gc_work);
 }
@@ -238,27 +348,12 @@ int __xfrm_state_delete(struct xfrm_state *x)
         if (x->km.state != XFRM_STATE_DEAD) {
                 x->km.state = XFRM_STATE_DEAD;
                 spin_lock(&xfrm_state_lock);
-                list_del(&x->bydst);
-                __xfrm_state_put(x);
-                if (x->id.spi) {
-                        list_del(&x->byspi);
-                        __xfrm_state_put(x);
-                }
+                hlist_del(&x->bydst);
+                hlist_del(&x->bysrc);
+                if (x->id.spi)
+                        hlist_del(&x->byspi);
+                xfrm_state_num--;
                 spin_unlock(&xfrm_state_lock);
-                if (del_timer(&x->timer))
-                        __xfrm_state_put(x);
-                if (del_timer(&x->rtimer))
-                        __xfrm_state_put(x);
-
-                /* The number two in this test is the reference
-                 * mentioned in the comment below plus the reference
-                 * our caller holds.  A larger value means that
-                 * there are DSTs attached to this xfrm_state.
-                 */
-                if (atomic_read(&x->refcnt) > 2) {
-                        xfrm_state_gc_flush_bundles = 1;
-                        schedule_work(&xfrm_state_gc_work);
-                }
 
                 /* All xfrm_state objects are created by xfrm_state_alloc.
                  * The xfrm_state_alloc call gives a reference, and that
@@ -287,14 +382,15 @@ EXPORT_SYMBOL(xfrm_state_delete);
 void xfrm_state_flush(u8 proto)
 {
         int i;
-        struct xfrm_state *x;
 
         spin_lock_bh(&xfrm_state_lock);
-        for (i = 0; i < XFRM_DST_HSIZE; i++) {
+        for (i = 0; i <= xfrm_state_hmask; i++) {
+                struct hlist_node *entry;
+                struct xfrm_state *x;
 restart:
-                list_for_each_entry(x, xfrm_state_bydst+i, bydst) {
+                hlist_for_each_entry(x, entry, xfrm_state_bydst+i, bydst) {
                         if (!xfrm_state_kern(x) &&
-                            (proto == IPSEC_PROTO_ANY || x->id.proto == proto)) {
+                            xfrm_id_proto_match(x->id.proto, proto)) {
                                 xfrm_state_hold(x);
                                 spin_unlock_bh(&xfrm_state_lock);
 
@@ -325,29 +421,103 @@ xfrm_init_tempsel(struct xfrm_state *x, struct flowi *fl,
         return 0;
 }
 
+static struct xfrm_state *__xfrm_state_lookup(xfrm_address_t *daddr, u32 spi, u8 proto, unsigned short family)
+{
+        unsigned int h = xfrm_spi_hash(daddr, spi, proto, family);
+        struct xfrm_state *x;
+        struct hlist_node *entry;
+
+        hlist_for_each_entry(x, entry, xfrm_state_byspi+h, byspi) {
+                if (x->props.family != family ||
+                    x->id.spi       != spi ||
+                    x->id.proto     != proto)
+                        continue;
+
+                switch (family) {
+                case AF_INET:
+                        if (x->id.daddr.a4 != daddr->a4)
+                                continue;
+                        break;
+                case AF_INET6:
+                        if (!ipv6_addr_equal((struct in6_addr *)daddr,
+                                             (struct in6_addr *)
+                                             x->id.daddr.a6))
+                                continue;
+                        break;
+                };
+
+                xfrm_state_hold(x);
+                return x;
+        }
+
+        return NULL;
+}
+
+static struct xfrm_state *__xfrm_state_lookup_byaddr(xfrm_address_t *daddr, xfrm_address_t *saddr, u8 proto, unsigned short family)
+{
+        unsigned int h = xfrm_src_hash(saddr, family);
+        struct xfrm_state *x;
+        struct hlist_node *entry;
+
+        hlist_for_each_entry(x, entry, xfrm_state_bysrc+h, bysrc) {
+                if (x->props.family != family ||
+                    x->id.proto     != proto)
+                        continue;
+
+                switch (family) {
+                case AF_INET:
+                        if (x->id.daddr.a4 != daddr->a4 ||
+                            x->props.saddr.a4 != saddr->a4)
+                                continue;
+                        break;
+                case AF_INET6:
+                        if (!ipv6_addr_equal((struct in6_addr *)daddr,
+                                             (struct in6_addr *)
+                                             x->id.daddr.a6) ||
+                            !ipv6_addr_equal((struct in6_addr *)saddr,
+                                             (struct in6_addr *)
+                                             x->props.saddr.a6))
+                                continue;
+                        break;
+                };
+
+                xfrm_state_hold(x);
+                return x;
+        }
+
+        return NULL;
+}
+
+static inline struct xfrm_state *
+__xfrm_state_locate(struct xfrm_state *x, int use_spi, int family)
+{
+        if (use_spi)
+                return __xfrm_state_lookup(&x->id.daddr, x->id.spi,
+                                           x->id.proto, family);
+        else
+                return __xfrm_state_lookup_byaddr(&x->id.daddr,
+                                                  &x->props.saddr,
+                                                  x->id.proto, family);
+}
+
 struct xfrm_state *
 xfrm_state_find(xfrm_address_t *daddr, xfrm_address_t *saddr, 
                 struct flowi *fl, struct xfrm_tmpl *tmpl,
                 struct xfrm_policy *pol, int *err,
                 unsigned short family)
 {
-        unsigned h = xfrm_dst_hash(daddr, family);
+        unsigned int h = xfrm_dst_hash(daddr, saddr, tmpl->reqid, family);
+        struct hlist_node *entry;
         struct xfrm_state *x, *x0;
         int acquire_in_progress = 0;
         int error = 0;
         struct xfrm_state *best = NULL;
-        struct xfrm_state_afinfo *afinfo;
         
-        afinfo = xfrm_state_get_afinfo(family);
-        if (afinfo == NULL) {
-                *err = -EAFNOSUPPORT;
-                return NULL;
-        }
-
         spin_lock_bh(&xfrm_state_lock);
-        list_for_each_entry(x, xfrm_state_bydst+h, bydst) {
+        hlist_for_each_entry(x, entry, xfrm_state_bydst+h, bydst) {
                 if (x->props.family == family &&
                     x->props.reqid == tmpl->reqid &&
+                    !(x->props.flags & XFRM_STATE_WILDRECV) &&
                     xfrm_state_addr_check(x, daddr, saddr, family) &&
                     tmpl->mode == x->props.mode &&
                     tmpl->id.proto == x->id.proto &&
@@ -367,7 +537,7 @@ xfrm_state_find(xfrm_address_t *daddr, xfrm_address_t *saddr,
                          */
                         if (x->km.state == XFRM_STATE_VALID) {
                                 if (!xfrm_selector_match(&x->sel, fl, family) ||
-                                    !xfrm_sec_ctx_match(pol->security, x->security))
+                                    !security_xfrm_state_pol_flow_match(x, pol, fl))
                                         continue;
                                 if (!best ||
                                     best->km.dying > x->km.dying ||
@@ -379,7 +549,7 @@ xfrm_state_find(xfrm_address_t *daddr, xfrm_address_t *saddr,
                         } else if (x->km.state == XFRM_STATE_ERROR ||
                                    x->km.state == XFRM_STATE_EXPIRED) {
                                 if (xfrm_selector_match(&x->sel, fl, family) &&
-                                    xfrm_sec_ctx_match(pol->security, x->security))
+                                    security_xfrm_state_pol_flow_match(x, pol, fl))
                                         error = -ESRCH;
                         }
                 }
@@ -388,8 +558,8 @@ xfrm_state_find(xfrm_address_t *daddr, xfrm_address_t *saddr,
         x = best;
         if (!x && !error && !acquire_in_progress) {
                 if (tmpl->id.spi &&
-                    (x0 = afinfo->state_lookup(daddr, tmpl->id.spi,
-                                               tmpl->id.proto)) != NULL) {
+                    (x0 = __xfrm_state_lookup(daddr, tmpl->id.spi,
+                                              tmpl->id.proto, family)) != NULL) {
                         xfrm_state_put(x0);
                         error = -EEXIST;
                         goto out;
@@ -403,17 +573,24 @@ xfrm_state_find(xfrm_address_t *daddr, xfrm_address_t *saddr,
                  * to current session. */
                 xfrm_init_tempsel(x, fl, tmpl, daddr, saddr, family);
 
+                error = security_xfrm_state_alloc_acquire(x, pol->security, fl->secid);
+                if (error) {
+                        x->km.state = XFRM_STATE_DEAD;
+                        xfrm_state_put(x);
+                        x = NULL;
+                        goto out;
+                }
+
                 if (km_query(x, tmpl, pol) == 0) {
                         x->km.state = XFRM_STATE_ACQ;
-                        list_add_tail(&x->bydst, xfrm_state_bydst+h);
-                        xfrm_state_hold(x);
+                        hlist_add_head(&x->bydst, xfrm_state_bydst+h);
+                        h = xfrm_src_hash(saddr, family);
+                        hlist_add_head(&x->bysrc, xfrm_state_bysrc+h);
                         if (x->id.spi) {
                                 h = xfrm_spi_hash(&x->id.daddr, x->id.spi, x->id.proto, family);
-                                list_add(&x->byspi, xfrm_state_byspi+h);
-                                xfrm_state_hold(x);
+                                hlist_add_head(&x->byspi, xfrm_state_byspi+h);
                         }
                         x->lft.hard_add_expires_seconds = XFRM_ACQ_EXPIRES;
-                        xfrm_state_hold(x);
                         x->timer.expires = jiffies + XFRM_ACQ_EXPIRES*HZ;
                         add_timer(&x->timer);
                 } else {
@@ -429,59 +606,167 @@ out:
         else
                 *err = acquire_in_progress ? -EAGAIN : error;
         spin_unlock_bh(&xfrm_state_lock);
-        xfrm_state_put_afinfo(afinfo);
         return x;
 }
 
 static void __xfrm_state_insert(struct xfrm_state *x)
 {
-        unsigned h = xfrm_dst_hash(&x->id.daddr, x->props.family);
+        unsigned int h;
 
-        list_add(&x->bydst, xfrm_state_bydst+h);
-        xfrm_state_hold(x);
+        x->genid = ++xfrm_state_genid;
 
-        h = xfrm_spi_hash(&x->id.daddr, x->id.spi, x->id.proto, x->props.family);
+        h = xfrm_dst_hash(&x->id.daddr, &x->props.saddr,
+                          x->props.reqid, x->props.family);
+        hlist_add_head(&x->bydst, xfrm_state_bydst+h);
 
-        list_add(&x->byspi, xfrm_state_byspi+h);
-        xfrm_state_hold(x);
+        h = xfrm_src_hash(&x->props.saddr, x->props.family);
+        hlist_add_head(&x->bysrc, xfrm_state_bysrc+h);
 
-        if (!mod_timer(&x->timer, jiffies + HZ))
-                xfrm_state_hold(x);
+        if (xfrm_id_proto_match(x->id.proto, IPSEC_PROTO_ANY)) {
+                h = xfrm_spi_hash(&x->id.daddr, x->id.spi, x->id.proto,
+                                  x->props.family);
 
-        if (x->replay_maxage &&
-            !mod_timer(&x->rtimer, jiffies + x->replay_maxage))
-                xfrm_state_hold(x);
+                hlist_add_head(&x->byspi, xfrm_state_byspi+h);
+        }
+
+        mod_timer(&x->timer, jiffies + HZ);
+        if (x->replay_maxage)
+                mod_timer(&x->rtimer, jiffies + x->replay_maxage);
 
         wake_up(&km_waitq);
+
+        xfrm_state_num++;
+
+        if (x->bydst.next != NULL &&
+            (xfrm_state_hmask + 1) < xfrm_state_hashmax &&
+            xfrm_state_num > xfrm_state_hmask)
+                schedule_work(&xfrm_hash_work);
+}
+
+/* xfrm_state_lock is held */
+static void __xfrm_state_bump_genids(struct xfrm_state *xnew)
+{
+        unsigned short family = xnew->props.family;
+        u32 reqid = xnew->props.reqid;
+        struct xfrm_state *x;
+        struct hlist_node *entry;
+        unsigned int h;
+
+        h = xfrm_dst_hash(&xnew->id.daddr, &xnew->props.saddr, reqid, family);
+        hlist_for_each_entry(x, entry, xfrm_state_bydst+h, bydst) {
+                if (x->props.family     == family &&
+                    x->props.reqid      == reqid &&
+                    !xfrm_addr_cmp(&x->id.daddr, &xnew->id.daddr, family) &&
+                    !xfrm_addr_cmp(&x->props.saddr, &xnew->props.saddr, family))
+                        x->genid = xfrm_state_genid;
+        }
 }
 
 void xfrm_state_insert(struct xfrm_state *x)
 {
         spin_lock_bh(&xfrm_state_lock);
+        __xfrm_state_bump_genids(x);
         __xfrm_state_insert(x);
         spin_unlock_bh(&xfrm_state_lock);
-
-        xfrm_flush_all_bundles();
 }
 EXPORT_SYMBOL(xfrm_state_insert);
 
+/* xfrm_state_lock is held */
+static struct xfrm_state *__find_acq_core(unsigned short family, u8 mode, u32 reqid, u8 proto, xfrm_address_t *daddr, xfrm_address_t *saddr, int create)
+{
+        unsigned int h = xfrm_dst_hash(daddr, saddr, reqid, family);
+        struct hlist_node *entry;
+        struct xfrm_state *x;
+
+        hlist_for_each_entry(x, entry, xfrm_state_bydst+h, bydst) {
+                if (x->props.reqid  != reqid ||
+                    x->props.mode   != mode ||
+                    x->props.family != family ||
+                    x->km.state     != XFRM_STATE_ACQ ||
+                    x->id.spi       != 0)
+                        continue;
+
+                switch (family) {
+                case AF_INET:
+                        if (x->id.daddr.a4    != daddr->a4 ||
+                            x->props.saddr.a4 != saddr->a4)
+                                continue;
+                        break;
+                case AF_INET6:
+                        if (!ipv6_addr_equal((struct in6_addr *)x->id.daddr.a6,
+                                             (struct in6_addr *)daddr) ||
+                            !ipv6_addr_equal((struct in6_addr *)
+                                             x->props.saddr.a6,
+                                             (struct in6_addr *)saddr))
+                                continue;
+                        break;
+                };
+
+                xfrm_state_hold(x);
+                return x;
+        }
+
+        if (!create)
+                return NULL;
+
+        x = xfrm_state_alloc();
+        if (likely(x)) {
+                switch (family) {
+                case AF_INET:
+                        x->sel.daddr.a4 = daddr->a4;
+                        x->sel.saddr.a4 = saddr->a4;
+                        x->sel.prefixlen_d = 32;
+                        x->sel.prefixlen_s = 32;
+                        x->props.saddr.a4 = saddr->a4;
+                        x->id.daddr.a4 = daddr->a4;
+                        break;
+
+                case AF_INET6:
+                        ipv6_addr_copy((struct in6_addr *)x->sel.daddr.a6,
+                                       (struct in6_addr *)daddr);
+                        ipv6_addr_copy((struct in6_addr *)x->sel.saddr.a6,
+                                       (struct in6_addr *)saddr);
+                        x->sel.prefixlen_d = 128;
+                        x->sel.prefixlen_s = 128;
+                        ipv6_addr_copy((struct in6_addr *)x->props.saddr.a6,
+                                       (struct in6_addr *)saddr);
+                        ipv6_addr_copy((struct in6_addr *)x->id.daddr.a6,
+                                       (struct in6_addr *)daddr);
+                        break;
+                };
+
+                x->km.state = XFRM_STATE_ACQ;
+                x->id.proto = proto;
+                x->props.family = family;
+                x->props.mode = mode;
+                x->props.reqid = reqid;
+                x->lft.hard_add_expires_seconds = XFRM_ACQ_EXPIRES;
+                xfrm_state_hold(x);
+                x->timer.expires = jiffies + XFRM_ACQ_EXPIRES*HZ;
+                add_timer(&x->timer);
+                hlist_add_head(&x->bydst, xfrm_state_bydst+h);
+                h = xfrm_src_hash(saddr, family);
+                hlist_add_head(&x->bysrc, xfrm_state_bysrc+h);
+                wake_up(&km_waitq);
+        }
+
+        return x;
+}
+
 static struct xfrm_state *__xfrm_find_acq_byseq(u32 seq);
 
 int xfrm_state_add(struct xfrm_state *x)
 {
-        struct xfrm_state_afinfo *afinfo;
         struct xfrm_state *x1;
         int family;
         int err;
+        int use_spi = xfrm_id_proto_match(x->id.proto, IPSEC_PROTO_ANY);
 
         family = x->props.family;
-        afinfo = xfrm_state_get_afinfo(family);
-        if (unlikely(afinfo == NULL))
-                return -EAFNOSUPPORT;
 
         spin_lock_bh(&xfrm_state_lock);
 
-        x1 = afinfo->state_lookup(&x->id.daddr, x->id.spi, x->id.proto);
+        x1 = __xfrm_state_locate(x, use_spi, family);
         if (x1) {
                 xfrm_state_put(x1);
                 x1 = NULL;
@@ -489,7 +774,7 @@ int xfrm_state_add(struct xfrm_state *x)
                 goto out;
         }
 
-        if (x->km.seq) {
+        if (use_spi && x->km.seq) {
                 x1 = __xfrm_find_acq_byseq(x->km.seq);
                 if (x1 && xfrm_addr_cmp(&x1->id.daddr, &x->id.daddr, family)) {
                         xfrm_state_put(x1);
@@ -497,20 +782,17 @@ int xfrm_state_add(struct xfrm_state *x)
                 }
         }
 
-        if (!x1)
-                x1 = afinfo->find_acq(
-                        x->props.mode, x->props.reqid, x->id.proto,
-                        &x->id.daddr, &x->props.saddr, 0);
+        if (use_spi && !x1)
+                x1 = __find_acq_core(family, x->props.mode, x->props.reqid,
+                                     x->id.proto,
+                                     &x->id.daddr, &x->props.saddr, 0);
 
+        __xfrm_state_bump_genids(x);
         __xfrm_state_insert(x);
         err = 0;
 
 out:
         spin_unlock_bh(&xfrm_state_lock);
-        xfrm_state_put_afinfo(afinfo);
-
-        if (!err)
-                xfrm_flush_all_bundles();
 
         if (x1) {
                 xfrm_state_delete(x1);
@@ -523,16 +805,12 @@ EXPORT_SYMBOL(xfrm_state_add);
 
 int xfrm_state_update(struct xfrm_state *x)
 {
-        struct xfrm_state_afinfo *afinfo;
         struct xfrm_state *x1;
         int err;
-
-        afinfo = xfrm_state_get_afinfo(x->props.family);
-        if (unlikely(afinfo == NULL))
-                return -EAFNOSUPPORT;
+        int use_spi = xfrm_id_proto_match(x->id.proto, IPSEC_PROTO_ANY);
 
         spin_lock_bh(&xfrm_state_lock);
-        x1 = afinfo->state_lookup(&x->id.daddr, x->id.spi, x->id.proto);
+        x1 = __xfrm_state_locate(x, use_spi, x->props.family);
 
         err = -ESRCH;
         if (!x1)
@@ -552,7 +830,6 @@ int xfrm_state_update(struct xfrm_state *x)
 
 out:
         spin_unlock_bh(&xfrm_state_lock);
-        xfrm_state_put_afinfo(afinfo);
 
         if (err)
                 return err;
@@ -568,11 +845,15 @@ out:
         if (likely(x1->km.state == XFRM_STATE_VALID)) {
                 if (x->encap && x1->encap)
                         memcpy(x1->encap, x->encap, sizeof(*x1->encap));
+                if (x->coaddr && x1->coaddr) {
+                        memcpy(x1->coaddr, x->coaddr, sizeof(*x1->coaddr));
+                }
+                if (!use_spi && memcmp(&x1->sel, &x->sel, sizeof(x1->sel)))
+                        memcpy(&x1->sel, &x->sel, sizeof(x1->sel));
                 memcpy(&x1->lft, &x->lft, sizeof(x1->lft));
                 x1->km.dying = 0;
 
-                if (!mod_timer(&x1->timer, jiffies + HZ))
-                        xfrm_state_hold(x1);
+                mod_timer(&x1->timer, jiffies + HZ);
                 if (x1->curlft.use_time)
                         xfrm_state_check_expire(x1);
 
@@ -597,8 +878,7 @@ int xfrm_state_check_expire(struct xfrm_state *x)
         if (x->curlft.bytes >= x->lft.hard_byte_limit ||
             x->curlft.packets >= x->lft.hard_packet_limit) {
                 x->km.state = XFRM_STATE_EXPIRED;
-                if (!mod_timer(&x->timer, jiffies))
-                        xfrm_state_hold(x);
+                mod_timer(&x->timer, jiffies);
                 return -EINVAL;
         }
 
@@ -640,46 +920,93 @@ xfrm_state_lookup(xfrm_address_t *daddr, u32 spi, u8 proto,
                   unsigned short family)
 {
         struct xfrm_state *x;
-        struct xfrm_state_afinfo *afinfo = xfrm_state_get_afinfo(family);
-        if (!afinfo)
-                return NULL;
 
         spin_lock_bh(&xfrm_state_lock);
-        x = afinfo->state_lookup(daddr, spi, proto);
+        x = __xfrm_state_lookup(daddr, spi, proto, family);
         spin_unlock_bh(&xfrm_state_lock);
-        xfrm_state_put_afinfo(afinfo);
         return x;
 }
 EXPORT_SYMBOL(xfrm_state_lookup);
 
 struct xfrm_state *
+xfrm_state_lookup_byaddr(xfrm_address_t *daddr, xfrm_address_t *saddr,
+                         u8 proto, unsigned short family)
+{
+        struct xfrm_state *x;
+
+        spin_lock_bh(&xfrm_state_lock);
+        x = __xfrm_state_lookup_byaddr(daddr, saddr, proto, family);
+        spin_unlock_bh(&xfrm_state_lock);
+        return x;
+}
+EXPORT_SYMBOL(xfrm_state_lookup_byaddr);
+
+struct xfrm_state *
 xfrm_find_acq(u8 mode, u32 reqid, u8 proto, 
               xfrm_address_t *daddr, xfrm_address_t *saddr, 
               int create, unsigned short family)
 {
         struct xfrm_state *x;
+
+        spin_lock_bh(&xfrm_state_lock);
+        x = __find_acq_core(family, mode, reqid, proto, daddr, saddr, create);
+        spin_unlock_bh(&xfrm_state_lock);
+
+        return x;
+}
+EXPORT_SYMBOL(xfrm_find_acq);
+
+#ifdef CONFIG_XFRM_SUB_POLICY
+int
+xfrm_tmpl_sort(struct xfrm_tmpl **dst, struct xfrm_tmpl **src, int n,
+               unsigned short family)
+{
+        int err = 0;
         struct xfrm_state_afinfo *afinfo = xfrm_state_get_afinfo(family);
         if (!afinfo)
-                return NULL;
+                return -EAFNOSUPPORT;
 
         spin_lock_bh(&xfrm_state_lock);
-        x = afinfo->find_acq(mode, reqid, proto, daddr, saddr, create);
+        if (afinfo->tmpl_sort)
+                err = afinfo->tmpl_sort(dst, src, n);
         spin_unlock_bh(&xfrm_state_lock);
         xfrm_state_put_afinfo(afinfo);
-        return x;
+        return err;
 }
-EXPORT_SYMBOL(xfrm_find_acq);
+EXPORT_SYMBOL(xfrm_tmpl_sort);
+
+int
+xfrm_state_sort(struct xfrm_state **dst, struct xfrm_state **src, int n,
+                unsigned short family)
+{
+        int err = 0;
+        struct xfrm_state_afinfo *afinfo = xfrm_state_get_afinfo(family);
+        if (!afinfo)
+                return -EAFNOSUPPORT;
+
+        spin_lock_bh(&xfrm_state_lock);
+        if (afinfo->state_sort)
+                err = afinfo->state_sort(dst, src, n);
+        spin_unlock_bh(&xfrm_state_lock);
+        xfrm_state_put_afinfo(afinfo);
+        return err;
+}
+EXPORT_SYMBOL(xfrm_state_sort);
+#endif
 
 /* Silly enough, but I'm lazy to build resolution list */
 
 static struct xfrm_state *__xfrm_find_acq_byseq(u32 seq)
 {
         int i;
-        struct xfrm_state *x;
 
-        for (i = 0; i < XFRM_DST_HSIZE; i++) {
-                list_for_each_entry(x, xfrm_state_bydst+i, bydst) {
-                        if (x->km.seq == seq && x->km.state == XFRM_STATE_ACQ) {
+        for (i = 0; i <= xfrm_state_hmask; i++) {
+                struct hlist_node *entry;
+                struct xfrm_state *x;
+
+                hlist_for_each_entry(x, entry, xfrm_state_bydst+i, bydst) {
+                        if (x->km.seq == seq &&
+                            x->km.state == XFRM_STATE_ACQ) {
                                 xfrm_state_hold(x);
                                 return x;
                         }
@@ -715,7 +1042,7 @@ EXPORT_SYMBOL(xfrm_get_acqseq);
 void
 xfrm_alloc_spi(struct xfrm_state *x, u32 minspi, u32 maxspi)
 {
-        u32 h;
+        unsigned int h;
         struct xfrm_state *x0;
 
         if (x->id.spi)
@@ -745,8 +1072,7 @@ xfrm_alloc_spi(struct xfrm_state *x, u32 minspi, u32 maxspi)
         if (x->id.spi) {
                 spin_lock_bh(&xfrm_state_lock);
                 h = xfrm_spi_hash(&x->id.daddr, x->id.spi, x->id.proto, x->props.family);
-                list_add(&x->byspi, xfrm_state_byspi+h);
-                xfrm_state_hold(x);
+                hlist_add_head(&x->byspi, xfrm_state_byspi+h);
                 spin_unlock_bh(&xfrm_state_lock);
                 wake_up(&km_waitq);
         }
@@ -758,13 +1084,14 @@ int xfrm_state_walk(u8 proto, int (*func)(struct xfrm_state *, int, void*),
 {
         int i;
         struct xfrm_state *x;
+        struct hlist_node *entry;
         int count = 0;
         int err = 0;
 
         spin_lock_bh(&xfrm_state_lock);
-        for (i = 0; i < XFRM_DST_HSIZE; i++) {
-                list_for_each_entry(x, xfrm_state_bydst+i, bydst) {
-                        if (proto == IPSEC_PROTO_ANY || x->id.proto == proto)
+        for (i = 0; i <= xfrm_state_hmask; i++) {
+                hlist_for_each_entry(x, entry, xfrm_state_bydst+i, bydst) {
+                        if (xfrm_id_proto_match(x->id.proto, proto))
                                 count++;
                 }
         }
@@ -773,9 +1100,9 @@ int xfrm_state_walk(u8 proto, int (*func)(struct xfrm_state *, int, void*),
                 goto out;
         }
 
-        for (i = 0; i < XFRM_DST_HSIZE; i++) {
-                list_for_each_entry(x, xfrm_state_bydst+i, bydst) {
-                        if (proto != IPSEC_PROTO_ANY && x->id.proto != proto)
+        for (i = 0; i <= xfrm_state_hmask; i++) {
+                hlist_for_each_entry(x, entry, xfrm_state_bydst+i, bydst) {
+                        if (!xfrm_id_proto_match(x->id.proto, proto))
                                 continue;
                         err = func(x, --count, data);
                         if (err)
@@ -832,10 +1159,8 @@ void xfrm_replay_notify(struct xfrm_state *x, int event)
         km_state_notify(x, &c);
 
         if (x->replay_maxage &&
-            !mod_timer(&x->rtimer, jiffies + x->replay_maxage)) {
-                xfrm_state_hold(x);
+            !mod_timer(&x->rtimer, jiffies + x->replay_maxage))
                 x->xflags &= ~XFRM_TIME_DEFER;
-        }
 }
 EXPORT_SYMBOL(xfrm_replay_notify);
 
@@ -853,7 +1178,6 @@ static void xfrm_replay_timer_handler(unsigned long data)
         }
 
         spin_unlock(&x->lock);
-        xfrm_state_put(x);
 }
 
 int xfrm_replay_check(struct xfrm_state *x, u32 seq)
@@ -997,6 +1321,25 @@ void km_policy_expired(struct xfrm_policy *pol, int dir, int hard, u32 pid)
 }
 EXPORT_SYMBOL(km_policy_expired);
 
+int km_report(u8 proto, struct xfrm_selector *sel, xfrm_address_t *addr)
+{
+        int err = -EINVAL;
+        int ret;
+        struct xfrm_mgr *km;
+
+        read_lock(&xfrm_km_lock);
+        list_for_each_entry(km, &xfrm_km_list, list) {
+                if (km->report) {
+                        ret = km->report(proto, sel, addr);
+                        if (!ret)
+                                err = ret;
+                }
+        }
+        read_unlock(&xfrm_km_lock);
+        return err;
+}
+EXPORT_SYMBOL(km_report);
+
 int xfrm_user_policy(struct sock *sk, int optname, u8 __user *optval, int optlen)
 {
         int err;
@@ -1018,7 +1361,7 @@ int xfrm_user_policy(struct sock *sk, int optname, u8 __user *optval, int optlen
         err = -EINVAL;
         read_lock(&xfrm_km_lock);
         list_for_each_entry(km, &xfrm_km_list, list) {
-                pol = km->compile_policy(sk->sk_family, optname, data,
+                pol = km->compile_policy(sk, optname, data,
                                          optlen, &err);
                 if (err >= 0)
                         break;
@@ -1065,11 +1408,8 @@ int xfrm_state_register_afinfo(struct xfrm_state_afinfo *afinfo)
         write_lock_bh(&xfrm_state_afinfo_lock);
         if (unlikely(xfrm_state_afinfo[afinfo->family] != NULL))
                 err = -ENOBUFS;
-        else {
-                afinfo->state_bydst = xfrm_state_bydst;
-                afinfo->state_byspi = xfrm_state_byspi;
+        else
                 xfrm_state_afinfo[afinfo->family] = afinfo;
-        }
         write_unlock_bh(&xfrm_state_afinfo_lock);
         return err;
 }
@@ -1086,11 +1426,8 @@ int xfrm_state_unregister_afinfo(struct xfrm_state_afinfo *afinfo)
         if (likely(xfrm_state_afinfo[afinfo->family] != NULL)) {
                 if (unlikely(xfrm_state_afinfo[afinfo->family] != afinfo))
                         err = -EINVAL;
-                else {
+                else
                         xfrm_state_afinfo[afinfo->family] = NULL;
-                        afinfo->state_byspi = NULL;
-                        afinfo->state_bydst = NULL;
-                }
         }
         write_unlock_bh(&xfrm_state_afinfo_lock);
         return err;
@@ -1206,12 +1543,17 @@ EXPORT_SYMBOL(xfrm_init_state);
  
 void __init xfrm_state_init(void)
 {
-        int i;
+        unsigned int sz;
+
+        sz = sizeof(struct hlist_head) * 8;
+
+        xfrm_state_bydst = xfrm_hash_alloc(sz);
+        xfrm_state_bysrc = xfrm_hash_alloc(sz);
+        xfrm_state_byspi = xfrm_hash_alloc(sz);
+        if (!xfrm_state_bydst || !xfrm_state_bysrc || !xfrm_state_byspi)
+                panic("XFRM: Cannot allocate bydst/bysrc/byspi hashes.");
+        xfrm_state_hmask = ((sz / sizeof(struct hlist_head)) - 1);
 
-        for (i=0; i<XFRM_DST_HSIZE; i++) {
-                INIT_LIST_HEAD(&xfrm_state_bydst[i]);
-                INIT_LIST_HEAD(&xfrm_state_byspi[i]);
-        }
         INIT_WORK(&xfrm_state_gc_work, xfrm_state_gc_task, NULL);
 }
 
diff --git a/net/xfrm/xfrm_user.c b/net/xfrm/xfrm_user.c
index 3e6a722d072e..c59a78d2923a 100644
--- a/net/xfrm/xfrm_user.c
+++ b/net/xfrm/xfrm_user.c
@@ -10,6 +10,7 @@
  *
  */
 
+#include <linux/crypto.h>
 #include <linux/module.h>
 #include <linux/kernel.h>
 #include <linux/types.h>
@@ -27,6 +28,9 @@
 #include <net/xfrm.h>
 #include <net/netlink.h>
 #include <asm/uaccess.h>
+#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
+#include <linux/in6.h>
+#endif
 
 static int verify_one_alg(struct rtattr **xfrma, enum xfrm_attr_type_t type)
 {
@@ -86,6 +90,22 @@ static int verify_encap_tmpl(struct rtattr **xfrma)
         return 0;
 }
 
+static int verify_one_addr(struct rtattr **xfrma, enum xfrm_attr_type_t type,
+                           xfrm_address_t **addrp)
+{
+        struct rtattr *rt = xfrma[type - 1];
+
+        if (!rt)
+                return 0;
+
+        if ((rt->rta_len - sizeof(*rt)) < sizeof(**addrp))
+                return -EINVAL;
+
+        if (addrp)
+                *addrp = RTA_DATA(rt);
+
+        return 0;
+}
 
 static inline int verify_sec_ctx_len(struct rtattr **xfrma)
 {
@@ -156,6 +176,19 @@ static int verify_newsa_info(struct xfrm_usersa_info *p,
                         goto out;
                 break;
 
+#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
+        case IPPROTO_DSTOPTS:
+        case IPPROTO_ROUTING:
+                if (xfrma[XFRMA_ALG_COMP-1]     ||
+                    xfrma[XFRMA_ALG_AUTH-1]     ||
+                    xfrma[XFRMA_ALG_CRYPT-1]    ||
+                    xfrma[XFRMA_ENCAP-1]        ||
+                    xfrma[XFRMA_SEC_CTX-1]      ||
+                    !xfrma[XFRMA_COADDR-1])
+                        goto out;
+                break;
+#endif
+
         default:
                 goto out;
         };
@@ -170,11 +203,14 @@ static int verify_newsa_info(struct xfrm_usersa_info *p,
                 goto out;
         if ((err = verify_sec_ctx_len(xfrma)))
                 goto out;
+        if ((err = verify_one_addr(xfrma, XFRMA_COADDR, NULL)))
+                goto out;
 
         err = -EINVAL;
         switch (p->mode) {
-        case 0:
-        case 1:
+        case XFRM_MODE_TRANSPORT:
+        case XFRM_MODE_TUNNEL:
+        case XFRM_MODE_ROUTEOPTIMIZATION:
                 break;
 
         default:
@@ -212,6 +248,7 @@ static int attach_one_algo(struct xfrm_algo **algpp, u8 *props,
                 return -ENOMEM;
 
         memcpy(p, ualg, len);
+        strcpy(p->alg_name, algo->name);
         *algpp = p;
         return 0;
 }
@@ -258,6 +295,24 @@ static int attach_sec_ctx(struct xfrm_state *x, struct rtattr *u_arg)
         return security_xfrm_state_alloc(x, uctx);
 }
 
+static int attach_one_addr(xfrm_address_t **addrpp, struct rtattr *u_arg)
+{
+        struct rtattr *rta = u_arg;
+        xfrm_address_t *p, *uaddrp;
+
+        if (!rta)
+                return 0;
+
+        uaddrp = RTA_DATA(rta);
+        p = kmalloc(sizeof(*p), GFP_KERNEL);
+        if (!p)
+                return -ENOMEM;
+
+        memcpy(p, uaddrp, sizeof(*p));
+        *addrpp = p;
+        return 0;
+}
+
 static void copy_from_user_state(struct xfrm_state *x, struct xfrm_usersa_info *p)
 {
         memcpy(&x->id, &p->id, sizeof(x->id));
@@ -347,7 +402,8 @@ static struct xfrm_state *xfrm_state_construct(struct xfrm_usersa_info *p,
                 goto error;
         if ((err = attach_encap_tmpl(&x->encap, xfrma[XFRMA_ENCAP-1])))
                 goto error;
-
+        if ((err = attach_one_addr(&x->coaddr, xfrma[XFRMA_COADDR-1])))
+                goto error;
         err = xfrm_init_state(x);
         if (err)
                 goto error;
@@ -416,16 +472,48 @@ out:
         return err;
 }
 
+static struct xfrm_state *xfrm_user_state_lookup(struct xfrm_usersa_id *p,
+                                                 struct rtattr **xfrma,
+                                                 int *errp)
+{
+        struct xfrm_state *x = NULL;
+        int err;
+
+        if (xfrm_id_proto_match(p->proto, IPSEC_PROTO_ANY)) {
+                err = -ESRCH;
+                x = xfrm_state_lookup(&p->daddr, p->spi, p->proto, p->family);
+        } else {
+                xfrm_address_t *saddr = NULL;
+
+                err = verify_one_addr(xfrma, XFRMA_SRCADDR, &saddr);
+                if (err)
+                        goto out;
+
+                if (!saddr) {
+                        err = -EINVAL;
+                        goto out;
+                }
+
+                x = xfrm_state_lookup_byaddr(&p->daddr, saddr, p->proto,
+                                             p->family);
+        }
+
+ out:
+        if (!x && errp)
+                *errp = err;
+        return x;
+}
+
 static int xfrm_del_sa(struct sk_buff *skb, struct nlmsghdr *nlh, void **xfrma)
 {
         struct xfrm_state *x;
-        int err;
+        int err = -ESRCH;
         struct km_event c;
         struct xfrm_usersa_id *p = NLMSG_DATA(nlh);
 
-        x = xfrm_state_lookup(&p->daddr, p->spi, p->proto, p->family);
+        x = xfrm_user_state_lookup(p, (struct rtattr **)xfrma, &err);
         if (x == NULL)
-                return -ESRCH;
+                return err;
 
         if ((err = security_xfrm_state_delete(x)) != 0)
                 goto out;
@@ -519,6 +607,13 @@ static int dump_one_state(struct xfrm_state *x, int count, void *ptr)
                 uctx->ctx_len = x->security->ctx_len;
                 memcpy(uctx + 1, x->security->ctx_str, x->security->ctx_len);
         }
+
+        if (x->coaddr)
+                RTA_PUT(skb, XFRMA_COADDR, sizeof(*x->coaddr), x->coaddr);
+
+        if (x->lastused)
+                RTA_PUT(skb, XFRMA_LASTUSED, sizeof(x->lastused), &x->lastused);
+
         nlh->nlmsg_len = skb->tail - b;
 out:
         sp->this_idx++;
@@ -540,7 +635,7 @@ static int xfrm_dump_sa(struct sk_buff *skb, struct netlink_callback *cb)
         info.nlmsg_flags = NLM_F_MULTI;
         info.this_idx = 0;
         info.start_idx = cb->args[0];
-        (void) xfrm_state_walk(IPSEC_PROTO_ANY, dump_one_state, &info);
+        (void) xfrm_state_walk(0, dump_one_state, &info);
         cb->args[0] = info.this_idx;
 
         return skb->len;
@@ -576,10 +671,9 @@ static int xfrm_get_sa(struct sk_buff *skb, struct nlmsghdr *nlh, void **xfrma)
         struct xfrm_usersa_id *p = NLMSG_DATA(nlh);
         struct xfrm_state *x;
         struct sk_buff *resp_skb;
-        int err;
+        int err = -ESRCH;
 
-        x = xfrm_state_lookup(&p->daddr, p->spi, p->proto, p->family);
-        err = -ESRCH;
+        x = xfrm_user_state_lookup(p, (struct rtattr **)xfrma, &err);
         if (x == NULL)
                 goto out_noput;
 
@@ -692,6 +786,22 @@ static int verify_policy_dir(__u8 dir)
         return 0;
 }
 
+static int verify_policy_type(__u8 type)
+{
+        switch (type) {
+        case XFRM_POLICY_TYPE_MAIN:
+#ifdef CONFIG_XFRM_SUB_POLICY
+        case XFRM_POLICY_TYPE_SUB:
+#endif
+                break;
+
+        default:
+                return -EINVAL;
+        };
+
+        return 0;
+}
+
 static int verify_newpolicy_info(struct xfrm_userpolicy_info *p)
 {
         switch (p->share) {
@@ -785,6 +895,29 @@ static int copy_from_user_tmpl(struct xfrm_policy *pol, struct rtattr **xfrma)
         return 0;
 }
 
+static int copy_from_user_policy_type(u8 *tp, struct rtattr **xfrma)
+{
+        struct rtattr *rt = xfrma[XFRMA_POLICY_TYPE-1];
+        struct xfrm_userpolicy_type *upt;
+        __u8 type = XFRM_POLICY_TYPE_MAIN;
+        int err;
+
+        if (rt) {
+                if (rt->rta_len < sizeof(*upt))
+                        return -EINVAL;
+
+                upt = RTA_DATA(rt);
+                type = upt->type;
+        }
+
+        err = verify_policy_type(type);
+        if (err)
+                return err;
+
+        *tp = type;
+        return 0;
+}
+
 static void copy_from_user_policy(struct xfrm_policy *xp, struct xfrm_userpolicy_info *p)
 {
         xp->priority = p->priority;
@@ -823,16 +956,20 @@ static struct xfrm_policy *xfrm_policy_construct(struct xfrm_userpolicy_info *p,
 
         copy_from_user_policy(xp, p);
 
+        err = copy_from_user_policy_type(&xp->type, xfrma);
+        if (err)
+                goto error;
+
         if (!(err = copy_from_user_tmpl(xp, xfrma)))
                 err = copy_from_user_sec_ctx(xp, xfrma);
-
-        if (err) {
-                *errp = err;
-                kfree(xp);
-                xp = NULL;
-        }
+        if (err)
+                goto error;
 
         return xp;
+ error:
+        *errp = err;
+        kfree(xp);
+        return NULL;
 }
 
 static int xfrm_add_policy(struct sk_buff *skb, struct nlmsghdr *nlh, void **xfrma)
@@ -909,27 +1046,63 @@ rtattr_failure:
         return -1;
 }
 
-static int copy_to_user_sec_ctx(struct xfrm_policy *xp, struct sk_buff *skb)
+static int copy_sec_ctx(struct xfrm_sec_ctx *s, struct sk_buff *skb)
 {
-        if (xp->security) {
-                int ctx_size = sizeof(struct xfrm_sec_ctx) +
-                                xp->security->ctx_len;
-                struct rtattr *rt = __RTA_PUT(skb, XFRMA_SEC_CTX, ctx_size);
-                struct xfrm_user_sec_ctx *uctx = RTA_DATA(rt);
+        int ctx_size = sizeof(struct xfrm_sec_ctx) + s->ctx_len;
+        struct rtattr *rt = __RTA_PUT(skb, XFRMA_SEC_CTX, ctx_size);
+        struct xfrm_user_sec_ctx *uctx = RTA_DATA(rt);
+
+        uctx->exttype = XFRMA_SEC_CTX;
+        uctx->len = ctx_size;
+        uctx->ctx_doi = s->ctx_doi;
+        uctx->ctx_alg = s->ctx_alg;
+        uctx->ctx_len = s->ctx_len;
+        memcpy(uctx + 1, s->ctx_str, s->ctx_len);
+        return 0;
 
-                uctx->exttype = XFRMA_SEC_CTX;
-                uctx->len = ctx_size;
-                uctx->ctx_doi = xp->security->ctx_doi;
-                uctx->ctx_alg = xp->security->ctx_alg;
-                uctx->ctx_len = xp->security->ctx_len;
-                memcpy(uctx + 1, xp->security->ctx_str, xp->security->ctx_len);
+ rtattr_failure:
+        return -1;
+}
+
+static inline int copy_to_user_state_sec_ctx(struct xfrm_state *x, struct sk_buff *skb)
+{
+        if (x->security) {
+                return copy_sec_ctx(x->security, skb);
         }
         return 0;
+}
 
- rtattr_failure:
+static inline int copy_to_user_sec_ctx(struct xfrm_policy *xp, struct sk_buff *skb)
+{
+        if (xp->security) {
+                return copy_sec_ctx(xp->security, skb);
+        }
+        return 0;
+}
+
+#ifdef CONFIG_XFRM_SUB_POLICY
+static int copy_to_user_policy_type(struct xfrm_policy *xp, struct sk_buff *skb)
+{
+        struct xfrm_userpolicy_type upt;
+
+        memset(&upt, 0, sizeof(upt));
+        upt.type = xp->type;
+
+        RTA_PUT(skb, XFRMA_POLICY_TYPE, sizeof(upt), &upt);
+
+        return 0;
+
+rtattr_failure:
         return -1;
 }
 
+#else
+static inline int copy_to_user_policy_type(struct xfrm_policy *xp, struct sk_buff *skb)
+{
+        return 0;
+}
+#endif
+
 static int dump_one_policy(struct xfrm_policy *xp, int dir, int count, void *ptr)
 {
         struct xfrm_dump_info *sp = ptr;
@@ -953,6 +1126,8 @@ static int dump_one_policy(struct xfrm_policy *xp, int dir, int count, void *ptr
                 goto nlmsg_failure;
         if (copy_to_user_sec_ctx(xp, skb))
                 goto nlmsg_failure;
+        if (copy_to_user_policy_type(xp, skb) < 0)
+                goto nlmsg_failure;
 
         nlh->nlmsg_len = skb->tail - b;
 out:
@@ -974,7 +1149,10 @@ static int xfrm_dump_policy(struct sk_buff *skb, struct netlink_callback *cb)
         info.nlmsg_flags = NLM_F_MULTI;
         info.this_idx = 0;
         info.start_idx = cb->args[0];
-        (void) xfrm_policy_walk(dump_one_policy, &info);
+        (void) xfrm_policy_walk(XFRM_POLICY_TYPE_MAIN, dump_one_policy, &info);
+#ifdef CONFIG_XFRM_SUB_POLICY
+        (void) xfrm_policy_walk(XFRM_POLICY_TYPE_SUB, dump_one_policy, &info);
+#endif
         cb->args[0] = info.this_idx;
 
         return skb->len;
@@ -1010,6 +1188,7 @@ static int xfrm_get_policy(struct sk_buff *skb, struct nlmsghdr *nlh, void **xfr
 {
         struct xfrm_policy *xp;
         struct xfrm_userpolicy_id *p;
+        __u8 type = XFRM_POLICY_TYPE_MAIN;
         int err;
         struct km_event c;
         int delete;
@@ -1017,12 +1196,16 @@ static int xfrm_get_policy(struct sk_buff *skb, struct nlmsghdr *nlh, void **xfr
         p = NLMSG_DATA(nlh);
         delete = nlh->nlmsg_type == XFRM_MSG_DELPOLICY;
 
+        err = copy_from_user_policy_type(&type, (struct rtattr **)xfrma);
+        if (err)
+                return err;
+
         err = verify_policy_dir(p->dir);
         if (err)
                 return err;
 
         if (p->index)
-                xp = xfrm_policy_byid(p->dir, p->index, delete);
+                xp = xfrm_policy_byid(type, p->dir, p->index, delete);
         else {
                 struct rtattr **rtattrs = (struct rtattr **)xfrma;
                 struct rtattr *rt = rtattrs[XFRMA_SEC_CTX-1];
@@ -1039,7 +1222,7 @@ static int xfrm_get_policy(struct sk_buff *skb, struct nlmsghdr *nlh, void **xfr
                         if ((err = security_xfrm_policy_alloc(&tmp, uctx)))
                                 return err;
                 }
-                xp = xfrm_policy_bysel_ctx(p->dir, &p->sel, tmp.security, delete);
+                xp = xfrm_policy_bysel_ctx(type, p->dir, &p->sel, tmp.security, delete);
                 security_xfrm_policy_free(&tmp);
         }
         if (xp == NULL)
@@ -1222,9 +1405,16 @@ out:
 
 static int xfrm_flush_policy(struct sk_buff *skb, struct nlmsghdr *nlh, void **xfrma)
 {
-struct km_event c;
+        struct km_event c;
+        __u8 type = XFRM_POLICY_TYPE_MAIN;
+        int err;
+
+        err = copy_from_user_policy_type(&type, (struct rtattr **)xfrma);
+        if (err)
+                return err;
 
-        xfrm_policy_flush();
+        xfrm_policy_flush(type);
+        c.data.type = type;
         c.event = nlh->nlmsg_type;
         c.seq = nlh->nlmsg_seq;
         c.pid = nlh->nlmsg_pid;
@@ -1237,10 +1427,15 @@ static int xfrm_add_pol_expire(struct sk_buff *skb, struct nlmsghdr *nlh, void *
         struct xfrm_policy *xp;
         struct xfrm_user_polexpire *up = NLMSG_DATA(nlh);
         struct xfrm_userpolicy_info *p = &up->pol;
+        __u8 type = XFRM_POLICY_TYPE_MAIN;
         int err = -ENOENT;
 
+        err = copy_from_user_policy_type(&type, (struct rtattr **)xfrma);
+        if (err)
+                return err;
+
         if (p->index)
-                xp = xfrm_policy_byid(p->dir, p->index, 0);
+                xp = xfrm_policy_byid(type, p->dir, p->index, 0);
         else {
                 struct rtattr **rtattrs = (struct rtattr **)xfrma;
                 struct rtattr *rt = rtattrs[XFRMA_SEC_CTX-1];
@@ -1257,7 +1452,7 @@ static int xfrm_add_pol_expire(struct sk_buff *skb, struct nlmsghdr *nlh, void *
                         if ((err = security_xfrm_policy_alloc(&tmp, uctx)))
                                 return err;
                 }
-                xp = xfrm_policy_bysel_ctx(p->dir, &p->sel, tmp.security, 0);
+                xp = xfrm_policy_bysel_ctx(type, p->dir, &p->sel, tmp.security, 0);
                 security_xfrm_policy_free(&tmp);
         }
 
@@ -1384,6 +1579,7 @@ static const int xfrm_msg_min[XFRM_NR_MSGTYPES] = {
         [XFRM_MSG_FLUSHPOLICY - XFRM_MSG_BASE] = NLMSG_LENGTH(0),
         [XFRM_MSG_NEWAE       - XFRM_MSG_BASE] = XMSGSIZE(xfrm_aevent_id),
         [XFRM_MSG_GETAE       - XFRM_MSG_BASE] = XMSGSIZE(xfrm_aevent_id),
+        [XFRM_MSG_REPORT      - XFRM_MSG_BASE] = XMSGSIZE(xfrm_user_report),
 };
 
 #undef XMSGSIZE
@@ -1708,7 +1904,9 @@ static int build_acquire(struct sk_buff *skb, struct xfrm_state *x,
 
         if (copy_to_user_tmpl(xp, skb) < 0)
                 goto nlmsg_failure;
-        if (copy_to_user_sec_ctx(xp, skb))
+        if (copy_to_user_state_sec_ctx(x, skb))
+                goto nlmsg_failure;
+        if (copy_to_user_policy_type(xp, skb) < 0)
                 goto nlmsg_failure;
 
         nlh->nlmsg_len = skb->tail - b;
@@ -1742,7 +1940,7 @@ static int xfrm_send_acquire(struct xfrm_state *x, struct xfrm_tmpl *xt,
 /* User gives us xfrm_user_policy_info followed by an array of 0
  * or more templates.
  */
-static struct xfrm_policy *xfrm_compile_policy(u16 family, int opt,
+static struct xfrm_policy *xfrm_compile_policy(struct sock *sk, int opt,
                                                u8 *data, int len, int *dir)
 {
         struct xfrm_userpolicy_info *p = (struct xfrm_userpolicy_info *)data;
@@ -1750,7 +1948,7 @@ static struct xfrm_policy *xfrm_compile_policy(u16 family, int opt,
         struct xfrm_policy *xp;
         int nr;
 
-        switch (family) {
+        switch (sk->sk_family) {
         case AF_INET:
                 if (opt != IP_XFRM_POLICY) {
                         *dir = -EOPNOTSUPP;
@@ -1790,8 +1988,18 @@ static struct xfrm_policy *xfrm_compile_policy(u16 family, int opt,
         }
 
         copy_from_user_policy(xp, p);
+        xp->type = XFRM_POLICY_TYPE_MAIN;
         copy_templates(xp, ut, nr);
 
+        if (!xp->security) {
+                int err = security_xfrm_sock_policy_alloc(xp, sk);
+                if (err) {
+                        kfree(xp);
+                        *dir = err;
+                        return NULL;
+                }
+        }
+
         *dir = p->dir;
 
         return xp;
@@ -1814,6 +2022,8 @@ static int build_polexpire(struct sk_buff *skb, struct xfrm_policy *xp,
                 goto nlmsg_failure;
         if (copy_to_user_sec_ctx(xp, skb))
                 goto nlmsg_failure;
+        if (copy_to_user_policy_type(xp, skb) < 0)
+                goto nlmsg_failure;
         upe->hard = !!hard;
 
         nlh->nlmsg_len = skb->tail - b;
@@ -1885,6 +2095,8 @@ static int xfrm_notify_policy(struct xfrm_policy *xp, int dir, struct km_event *
         copy_to_user_policy(xp, p, dir);
         if (copy_to_user_tmpl(xp, skb) < 0)
                 goto nlmsg_failure;
+        if (copy_to_user_policy_type(xp, skb) < 0)
+                goto nlmsg_failure;
 
         nlh->nlmsg_len = skb->tail - b;
 
@@ -1902,6 +2114,9 @@ static int xfrm_notify_policy_flush(struct km_event *c)
         struct nlmsghdr *nlh;
         struct sk_buff *skb;
         unsigned char *b;
+#ifdef CONFIG_XFRM_SUB_POLICY
+        struct xfrm_userpolicy_type upt;
+#endif
         int len = NLMSG_LENGTH(0);
 
         skb = alloc_skb(len, GFP_ATOMIC);
@@ -1911,6 +2126,13 @@ static int xfrm_notify_policy_flush(struct km_event *c)
 
 
         nlh = NLMSG_PUT(skb, c->pid, c->seq, XFRM_MSG_FLUSHPOLICY, 0);
+        nlh->nlmsg_flags = 0;
+
+#ifdef CONFIG_XFRM_SUB_POLICY
+        memset(&upt, 0, sizeof(upt));
+        upt.type = c->data.type;
+        RTA_PUT(skb, XFRMA_POLICY_TYPE, sizeof(upt), &upt);
+#endif
 
         nlh->nlmsg_len = skb->tail - b;
 
@@ -1918,6 +2140,9 @@ static int xfrm_notify_policy_flush(struct km_event *c)
         return netlink_broadcast(xfrm_nl, skb, 0, XFRMNLGRP_POLICY, GFP_ATOMIC);
 
 nlmsg_failure:
+#ifdef CONFIG_XFRM_SUB_POLICY
+rtattr_failure:
+#endif
         kfree_skb(skb);
         return -1;
 }
@@ -1942,19 +2167,64 @@ static int xfrm_send_policy_notify(struct xfrm_policy *xp, int dir, struct km_ev
 
 }
 
+static int build_report(struct sk_buff *skb, u8 proto,
+                        struct xfrm_selector *sel, xfrm_address_t *addr)
+{
+        struct xfrm_user_report *ur;
+        struct nlmsghdr *nlh;
+        unsigned char *b = skb->tail;
+
+        nlh = NLMSG_PUT(skb, 0, 0, XFRM_MSG_REPORT, sizeof(*ur));
+        ur = NLMSG_DATA(nlh);
+        nlh->nlmsg_flags = 0;
+
+        ur->proto = proto;
+        memcpy(&ur->sel, sel, sizeof(ur->sel));
+
+        if (addr)
+                RTA_PUT(skb, XFRMA_COADDR, sizeof(*addr), addr);
+
+        nlh->nlmsg_len = skb->tail - b;
+        return skb->len;
+
+nlmsg_failure:
+rtattr_failure:
+        skb_trim(skb, b - skb->data);
+        return -1;
+}
+
+static int xfrm_send_report(u8 proto, struct xfrm_selector *sel,
+                            xfrm_address_t *addr)
+{
+        struct sk_buff *skb;
+        size_t len;
+
+        len = NLMSG_ALIGN(NLMSG_LENGTH(sizeof(struct xfrm_user_report)));
+        skb = alloc_skb(len, GFP_ATOMIC);
+        if (skb == NULL)
+                return -ENOMEM;
+
+        if (build_report(skb, proto, sel, addr) < 0)
+                BUG();
+
+        NETLINK_CB(skb).dst_group = XFRMNLGRP_REPORT;
+        return netlink_broadcast(xfrm_nl, skb, 0, XFRMNLGRP_REPORT, GFP_ATOMIC);
+}
+
 static struct xfrm_mgr netlink_mgr = {
         .id             = "netlink",
         .notify         = xfrm_send_state_notify,
         .acquire        = xfrm_send_acquire,
         .compile_policy = xfrm_compile_policy,
         .notify_policy  = xfrm_send_policy_notify,
+        .report         = xfrm_send_report,
 };
 
 static int __init xfrm_user_init(void)
 {
         struct sock *nlsk;
 
-        printk(KERN_INFO "Initializing IPsec netlink socket\n");
+        printk(KERN_INFO "Initializing XFRM netlink socket\n");
 
         nlsk = netlink_kernel_create(NETLINK_XFRM, XFRMNLGRP_MAX,
                                      xfrm_netlink_rcv, THIS_MODULE);