Merge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/klassert/ipsec-next

Steffen Klassert says:

====================
pull request (net-next): ipsec-next 2014-09-25

1) Remove useless hash_resize_mutex in xfrm_hash_resize().
   This mutex is used only there, but xfrm_hash_resize()
   can't be called concurrently at all. From Ying Xue.

2) Extend policy hashing to prefixed policies based on
   prefix lenght thresholds. From Christophe Gouault.

3) Make the policy hash table thresholds configurable
   via netlink. From Christophe Gouault.

4) Remove the maximum authentication length for AH.
   This was needed to limit stack usage. We switched
   already to allocate space, so no need to keep the
   limit. From Herbert Xu.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

10 files changed, 303 insertions, 38 deletions

diff --git a/include/net/ah.h b/include/net/ah.h
index ca95b98969d..4e2dfa474a7 100644
--- a/include/net/ah.h
+++ b/include/net/ah.h
@@ -3,9 +3,6 @@
 
 #include <linux/skbuff.h>
 
-/* This is the maximum truncated ICV length that we know of. */
-#define MAX_AH_AUTH_LEN	64
-
 struct crypto_ahash;
 
 struct ah_data {
diff --git a/include/net/netns/xfrm.h b/include/net/netns/xfrm.h
index 3492434baf8..9da798256f0 100644
--- a/include/net/netns/xfrm.h
+++ b/include/net/netns/xfrm.h
@@ -13,6 +13,19 @@ struct ctl_table_header;
 struct xfrm_policy_hash {
 	struct hlist_head	*table;
 	unsigned int		hmask;
+	u8			dbits4;
+	u8			sbits4;
+	u8			dbits6;
+	u8			sbits6;
+};
+
+struct xfrm_policy_hthresh {
+	struct work_struct	work;
+	seqlock_t		lock;
+	u8			lbits4;
+	u8			rbits4;
+	u8			lbits6;
+	u8			rbits6;
 };
 
 struct netns_xfrm {
@@ -41,6 +54,7 @@ struct netns_xfrm {
 	struct xfrm_policy_hash	policy_bydst[XFRM_POLICY_MAX * 2];
 	unsigned int		policy_count[XFRM_POLICY_MAX * 2];
 	struct work_struct	policy_hash_work;
+	struct xfrm_policy_hthresh policy_hthresh;
 
 
 	struct sock		*nlsk;
diff --git a/include/net/xfrm.h b/include/net/xfrm.h
index 721e9c3b11b..dc4865e90fe 100644
--- a/include/net/xfrm.h
+++ b/include/net/xfrm.h
@@ -1591,6 +1591,7 @@ struct xfrm_policy *xfrm_policy_bysel_ctx(struct net *net, u32 mark,
 struct xfrm_policy *xfrm_policy_byid(struct net *net, u32 mark, u8, int dir,
 				     u32 id, int delete, int *err);
 int xfrm_policy_flush(struct net *net, u8 type, bool task_valid);
+void xfrm_policy_hash_rebuild(struct net *net);
 u32 xfrm_get_acqseq(void);
 int verify_spi_info(u8 proto, u32 min, u32 max);
 int xfrm_alloc_spi(struct xfrm_state *x, u32 minspi, u32 maxspi);
diff --git a/include/uapi/linux/xfrm.h b/include/uapi/linux/xfrm.h
index 25e5dd916ba..02d5125a5ee 100644
--- a/include/uapi/linux/xfrm.h
+++ b/include/uapi/linux/xfrm.h
@@ -328,6 +328,8 @@ enum xfrm_spdattr_type_t {
 	XFRMA_SPD_UNSPEC,
 	XFRMA_SPD_INFO,
 	XFRMA_SPD_HINFO,
+	XFRMA_SPD_IPV4_HTHRESH,
+	XFRMA_SPD_IPV6_HTHRESH,
 	__XFRMA_SPD_MAX
 
 #define XFRMA_SPD_MAX (__XFRMA_SPD_MAX - 1)
@@ -347,6 +349,11 @@ struct xfrmu_spdhinfo {
 	__u32 spdhmcnt;
 };
 
+struct xfrmu_spdhthresh {
+	__u8 lbits;
+	__u8 rbits;
+};
+
 struct xfrm_usersa_info {
 	struct xfrm_selector		sel;
 	struct xfrm_id			id;
diff --git a/net/ipv4/ah4.c b/net/ipv4/ah4.c
index a2afa89513a..ac9a32ec3ee 100644
--- a/net/ipv4/ah4.c
+++ b/net/ipv4/ah4.c
@@ -505,8 +505,6 @@ static int ah_init_state(struct xfrm_state *x)
 	ahp->icv_full_len = aalg_desc->uinfo.auth.icv_fullbits/8;
 	ahp->icv_trunc_len = x->aalg->alg_trunc_len/8;
 
-	BUG_ON(ahp->icv_trunc_len > MAX_AH_AUTH_LEN);
-
 	if (x->props.flags & XFRM_STATE_ALIGN4)
 		x->props.header_len = XFRM_ALIGN4(sizeof(struct ip_auth_hdr) +
 						  ahp->icv_trunc_len);
diff --git a/net/ipv6/ah6.c b/net/ipv6/ah6.c
index fcffd4e522c..6d16eb0e0c7 100644
--- a/net/ipv6/ah6.c
+++ b/net/ipv6/ah6.c
@@ -713,8 +713,6 @@ static int ah6_init_state(struct xfrm_state *x)
 	ahp->icv_full_len = aalg_desc->uinfo.auth.icv_fullbits/8;
 	ahp->icv_trunc_len = x->aalg->alg_trunc_len/8;
 
-	BUG_ON(ahp->icv_trunc_len > MAX_AH_AUTH_LEN);
-
 	x->props.header_len = XFRM_ALIGN8(sizeof(struct ip_auth_hdr) +
 					  ahp->icv_trunc_len);
 	switch (x->props.mode) {
diff --git a/net/xfrm/xfrm_hash.h b/net/xfrm/xfrm_hash.h
index 0622d319e1f..666c5ffe929 100644
--- a/net/xfrm/xfrm_hash.h
+++ b/net/xfrm/xfrm_hash.h
@@ -3,6 +3,7 @@
 
 #include <linux/xfrm.h>
 #include <linux/socket.h>
+#include <linux/jhash.h>
 
 static inline unsigned int __xfrm4_addr_hash(const xfrm_address_t *addr)
 {
@@ -28,6 +29,58 @@ static inline unsigned int __xfrm6_daddr_saddr_hash(const xfrm_address_t *daddr,
 		     saddr->a6[2] ^ saddr->a6[3]);
 }
 
+static inline u32 __bits2mask32(__u8 bits)
+{
+	u32 mask32 = 0xffffffff;
+
+	if (bits == 0)
+		mask32 = 0;
+	else if (bits < 32)
+		mask32 <<= (32 - bits);
+
+	return mask32;
+}
+
+static inline unsigned int __xfrm4_dpref_spref_hash(const xfrm_address_t *daddr,
+						    const xfrm_address_t *saddr,
+						    __u8 dbits,
+						    __u8 sbits)
+{
+	return jhash_2words(ntohl(daddr->a4) & __bits2mask32(dbits),
+			    ntohl(saddr->a4) & __bits2mask32(sbits),
+			    0);
+}
+
+static inline unsigned int __xfrm6_pref_hash(const xfrm_address_t *addr,
+					     __u8 prefixlen)
+{
+	int pdw;
+	int pbi;
+	u32 initval = 0;
+
+	pdw = prefixlen >> 5;     /* num of whole u32 in prefix */
+	pbi = prefixlen &  0x1f;  /* num of bits in incomplete u32 in prefix */
+
+	if (pbi) {
+		__be32 mask;
+
+		mask = htonl((0xffffffff) << (32 - pbi));
+
+		initval = (__force u32)(addr->a6[pdw] & mask);
+	}
+
+	return jhash2((__force u32 *)addr->a6, pdw, initval);
+}
+
+static inline unsigned int __xfrm6_dpref_spref_hash(const xfrm_address_t *daddr,
+						    const xfrm_address_t *saddr,
+						    __u8 dbits,
+						    __u8 sbits)
+{
+	return __xfrm6_pref_hash(daddr, dbits) ^
+	       __xfrm6_pref_hash(saddr, sbits);
+}
+
 static inline unsigned int __xfrm_dst_hash(const xfrm_address_t *daddr,
 					   const xfrm_address_t *saddr,
 					   u32 reqid, unsigned short family,
@@ -84,7 +137,8 @@ static inline unsigned int __idx_hash(u32 index, unsigned int hmask)
 }
 
 static inline unsigned int __sel_hash(const struct xfrm_selector *sel,
-				      unsigned short family, unsigned int hmask)
+				      unsigned short family, unsigned int hmask,
+				      u8 dbits, u8 sbits)
 {
 	const xfrm_address_t *daddr = &sel->daddr;
 	const xfrm_address_t *saddr = &sel->saddr;
@@ -92,19 +146,19 @@ static inline unsigned int __sel_hash(const struct xfrm_selector *sel,
 
 	switch (family) {
 	case AF_INET:
-		if (sel->prefixlen_d != 32 ||
-		    sel->prefixlen_s != 32)
+		if (sel->prefixlen_d < dbits ||
+		    sel->prefixlen_s < sbits)
 			return hmask + 1;
 
-		h = __xfrm4_daddr_saddr_hash(daddr, saddr);
+		h = __xfrm4_dpref_spref_hash(daddr, saddr, dbits, sbits);
 		break;
 
 	case AF_INET6:
-		if (sel->prefixlen_d != 128 ||
-		    sel->prefixlen_s != 128)
+		if (sel->prefixlen_d < dbits ||
+		    sel->prefixlen_s < sbits)
 			return hmask + 1;
 
-		h = __xfrm6_daddr_saddr_hash(daddr, saddr);
+		h = __xfrm6_dpref_spref_hash(daddr, saddr, dbits, sbits);
 		break;
 	}
 	h ^= (h >> 16);
@@ -113,17 +167,19 @@ static inline unsigned int __sel_hash(const struct xfrm_selector *sel,
 
 static inline unsigned int __addr_hash(const xfrm_address_t *daddr,
 				       const xfrm_address_t *saddr,
-				       unsigned short family, unsigned int hmask)
+				       unsigned short family,
+				       unsigned int hmask,
+				       u8 dbits, u8 sbits)
 {
 	unsigned int h = 0;
 
 	switch (family) {
 	case AF_INET:
-		h = __xfrm4_daddr_saddr_hash(daddr, saddr);
+		h = __xfrm4_dpref_spref_hash(daddr, saddr, dbits, sbits);
 		break;
 
 	case AF_INET6:
-		h = __xfrm6_daddr_saddr_hash(daddr, saddr);
+		h = __xfrm6_dpref_spref_hash(daddr, saddr, dbits, sbits);
 		break;
 	}
 	h ^= (h >> 16);
diff --git a/net/xfrm/xfrm_policy.c b/net/xfrm/xfrm_policy.c
index fdde51f4271..f623dca6ce3 100644
--- a/net/xfrm/xfrm_policy.c
+++ b/net/xfrm/xfrm_policy.c
@@ -349,12 +349,39 @@ static inline unsigned int idx_hash(struct net *net, u32 index)
 	return __idx_hash(index, net->xfrm.policy_idx_hmask);
 }
 
+/* calculate policy hash thresholds */
+static void __get_hash_thresh(struct net *net,
+			      unsigned short family, int dir,
+			      u8 *dbits, u8 *sbits)
+{
+	switch (family) {
+	case AF_INET:
+		*dbits = net->xfrm.policy_bydst[dir].dbits4;
+		*sbits = net->xfrm.policy_bydst[dir].sbits4;
+		break;
+
+	case AF_INET6:
+		*dbits = net->xfrm.policy_bydst[dir].dbits6;
+		*sbits = net->xfrm.policy_bydst[dir].sbits6;
+		break;
+
+	default:
+		*dbits = 0;
+		*sbits = 0;
+	}
+}
+
 static struct hlist_head *policy_hash_bysel(struct net *net,
 					    const struct xfrm_selector *sel,
 					    unsigned short family, int dir)
 {
 	unsigned int hmask = net->xfrm.policy_bydst[dir].hmask;
-	unsigned int hash = __sel_hash(sel, family, hmask);
+	unsigned int hash;
+	u8 dbits;
+	u8 sbits;
+
+	__get_hash_thresh(net, family, dir, &dbits, &sbits);
+	hash = __sel_hash(sel, family, hmask, dbits, sbits);
 
 	return (hash == hmask + 1 ?
 		&net->xfrm.policy_inexact[dir] :
@@ -367,25 +394,35 @@ static struct hlist_head *policy_hash_direct(struct net *net,
 					     unsigned short family, int dir)
 {
 	unsigned int hmask = net->xfrm.policy_bydst[dir].hmask;
-	unsigned int hash = __addr_hash(daddr, saddr, family, hmask);
+	unsigned int hash;
+	u8 dbits;
+	u8 sbits;
+
+	__get_hash_thresh(net, family, dir, &dbits, &sbits);
+	hash = __addr_hash(daddr, saddr, family, hmask, dbits, sbits);
 
 	return net->xfrm.policy_bydst[dir].table + hash;
 }
 
-static void xfrm_dst_hash_transfer(struct hlist_head *list,
+static void xfrm_dst_hash_transfer(struct net *net,
+				   struct hlist_head *list,
 				   struct hlist_head *ndsttable,
-				   unsigned int nhashmask)
+				   unsigned int nhashmask,
+				   int dir)
 {
 	struct hlist_node *tmp, *entry0 = NULL;
 	struct xfrm_policy *pol;
 	unsigned int h0 = 0;
+	u8 dbits;
+	u8 sbits;
 
 redo:
 	hlist_for_each_entry_safe(pol, tmp, list, bydst) {
 		unsigned int h;
 
+		__get_hash_thresh(net, pol->family, dir, &dbits, &sbits);
 		h = __addr_hash(&pol->selector.daddr, &pol->selector.saddr,
-				pol->family, nhashmask);
+				pol->family, nhashmask, dbits, sbits);
 		if (!entry0) {
 			hlist_del(&pol->bydst);
 			hlist_add_head(&pol->bydst, ndsttable+h);
@@ -439,7 +476,7 @@ static void xfrm_bydst_resize(struct net *net, int dir)
 	write_lock_bh(&net->xfrm.xfrm_policy_lock);
 
 	for (i = hmask; i >= 0; i--)
-		xfrm_dst_hash_transfer(odst + i, ndst, nhashmask);
+		xfrm_dst_hash_transfer(net, odst + i, ndst, nhashmask, dir);
 
 	net->xfrm.policy_bydst[dir].table = ndst;
 	net->xfrm.policy_bydst[dir].hmask = nhashmask;
@@ -534,6 +571,86 @@ static void xfrm_hash_resize(struct work_struct *work)
 	mutex_unlock(&hash_resize_mutex);
 }
 
+static void xfrm_hash_rebuild(struct work_struct *work)
+{
+	struct net *net = container_of(work, struct net,
+				       xfrm.policy_hthresh.work);
+	unsigned int hmask;
+	struct xfrm_policy *pol;
+	struct xfrm_policy *policy;
+	struct hlist_head *chain;
+	struct hlist_head *odst;
+	struct hlist_node *newpos;
+	int i;
+	int dir;
+	unsigned seq;
+	u8 lbits4, rbits4, lbits6, rbits6;
+
+	mutex_lock(&hash_resize_mutex);
+
+	/* read selector prefixlen thresholds */
+	do {
+		seq = read_seqbegin(&net->xfrm.policy_hthresh.lock);
+
+		lbits4 = net->xfrm.policy_hthresh.lbits4;
+		rbits4 = net->xfrm.policy_hthresh.rbits4;
+		lbits6 = net->xfrm.policy_hthresh.lbits6;
+		rbits6 = net->xfrm.policy_hthresh.rbits6;
+	} while (read_seqretry(&net->xfrm.policy_hthresh.lock, seq));
+
+	write_lock_bh(&net->xfrm.xfrm_policy_lock);
+
+	/* reset the bydst and inexact table in all directions */
+	for (dir = 0; dir < XFRM_POLICY_MAX * 2; dir++) {
+		INIT_HLIST_HEAD(&net->xfrm.policy_inexact[dir]);
+		hmask = net->xfrm.policy_bydst[dir].hmask;
+		odst = net->xfrm.policy_bydst[dir].table;
+		for (i = hmask; i >= 0; i--)
+			INIT_HLIST_HEAD(odst + i);
+		if ((dir & XFRM_POLICY_MASK) == XFRM_POLICY_OUT) {
+			/* dir out => dst = remote, src = local */
+			net->xfrm.policy_bydst[dir].dbits4 = rbits4;
+			net->xfrm.policy_bydst[dir].sbits4 = lbits4;
+			net->xfrm.policy_bydst[dir].dbits6 = rbits6;
+			net->xfrm.policy_bydst[dir].sbits6 = lbits6;
+		} else {
+			/* dir in/fwd => dst = local, src = remote */
+			net->xfrm.policy_bydst[dir].dbits4 = lbits4;
+			net->xfrm.policy_bydst[dir].sbits4 = rbits4;
+			net->xfrm.policy_bydst[dir].dbits6 = lbits6;
+			net->xfrm.policy_bydst[dir].sbits6 = rbits6;
+		}
+	}
+
+	/* re-insert all policies by order of creation */
+	list_for_each_entry_reverse(policy, &net->xfrm.policy_all, walk.all) {
+		newpos = NULL;
+		chain = policy_hash_bysel(net, &policy->selector,
+					  policy->family,
+					  xfrm_policy_id2dir(policy->index));
+		hlist_for_each_entry(pol, chain, bydst) {
+			if (policy->priority >= pol->priority)
+				newpos = &pol->bydst;
+			else
+				break;
+		}
+		if (newpos)
+			hlist_add_behind(&policy->bydst, newpos);
+		else
+			hlist_add_head(&policy->bydst, chain);
+	}
+
+	write_unlock_bh(&net->xfrm.xfrm_policy_lock);
+
+	mutex_unlock(&hash_resize_mutex);
+}
+
+void xfrm_policy_hash_rebuild(struct net *net)
+{
+	schedule_work(&net->xfrm.policy_hthresh.work);
+}
+EXPORT_SYMBOL(xfrm_policy_hash_rebuild);
+
 /* 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(struct net *net, int dir, u32 index)
@@ -2862,10 +2979,21 @@ static int __net_init xfrm_policy_init(struct net *net)
 		if (!htab->table)
 			goto out_bydst;
 		htab->hmask = hmask;
+		htab->dbits4 = 32;
+		htab->sbits4 = 32;
+		htab->dbits6 = 128;
+		htab->sbits6 = 128;
 	}
+	net->xfrm.policy_hthresh.lbits4 = 32;
+	net->xfrm.policy_hthresh.rbits4 = 32;
+	net->xfrm.policy_hthresh.lbits6 = 128;
+	net->xfrm.policy_hthresh.rbits6 = 128;
+
+	seqlock_init(&net->xfrm.policy_hthresh.lock);
 
 	INIT_LIST_HEAD(&net->xfrm.policy_all);
 	INIT_WORK(&net->xfrm.policy_hash_work, xfrm_hash_resize);
+	INIT_WORK(&net->xfrm.policy_hthresh.work, xfrm_hash_rebuild);
 	if (net_eq(net, &init_net))
 		register_netdevice_notifier(&xfrm_dev_notifier);
 	return 0;
diff --git a/net/xfrm/xfrm_state.c b/net/xfrm/xfrm_state.c
index 0ab54134bb4..de971b6d38c 100644
--- a/net/xfrm/xfrm_state.c
+++ b/net/xfrm/xfrm_state.c
@@ -97,8 +97,6 @@ static unsigned long xfrm_hash_new_size(unsigned int state_hmask)
 	return ((state_hmask + 1) << 1) * sizeof(struct hlist_head);
 }
 
-static DEFINE_MUTEX(hash_resize_mutex);
-
 static void xfrm_hash_resize(struct work_struct *work)
 {
 	struct net *net = container_of(work, struct net, xfrm.state_hash_work);
@@ -107,22 +105,20 @@ static void xfrm_hash_resize(struct work_struct *work)
 	unsigned int nhashmask, ohashmask;
 	int i;
 
-	mutex_lock(&hash_resize_mutex);
-
 	nsize = xfrm_hash_new_size(net->xfrm.state_hmask);
 	ndst = xfrm_hash_alloc(nsize);
 	if (!ndst)
-		goto out_unlock;
+		return;
 	nsrc = xfrm_hash_alloc(nsize);
 	if (!nsrc) {
 		xfrm_hash_free(ndst, nsize);
-		goto out_unlock;
+		return;
 	}
 	nspi = xfrm_hash_alloc(nsize);
 	if (!nspi) {
 		xfrm_hash_free(ndst, nsize);
 		xfrm_hash_free(nsrc, nsize);
-		goto out_unlock;
+		return;
 	}
 
 	spin_lock_bh(&net->xfrm.xfrm_state_lock);
@@ -148,9 +144,6 @@ static void xfrm_hash_resize(struct work_struct *work)
 	xfrm_hash_free(odst, osize);
 	xfrm_hash_free(osrc, osize);
 	xfrm_hash_free(ospi, osize);
-
-out_unlock:
-	mutex_unlock(&hash_resize_mutex);
 }
 
 static DEFINE_SPINLOCK(xfrm_state_afinfo_lock);
diff --git a/net/xfrm/xfrm_user.c b/net/xfrm/xfrm_user.c
index d4db6ebb089..e812e988c11 100644
--- a/net/xfrm/xfrm_user.c
+++ b/net/xfrm/xfrm_user.c
@@ -333,8 +333,7 @@ static int attach_auth_trunc(struct xfrm_algo_auth **algpp, u8 *props,
 	algo = xfrm_aalg_get_byname(ualg->alg_name, 1);
 	if (!algo)
 		return -ENOSYS;
-	if ((ualg->alg_trunc_len / 8) > MAX_AH_AUTH_LEN ||
-	    ualg->alg_trunc_len > algo->uinfo.auth.icv_fullbits)
+	if (ualg->alg_trunc_len > algo->uinfo.auth.icv_fullbits)
 		return -EINVAL;
 	*props = algo->desc.sadb_alg_id;
 
@@ -964,7 +963,9 @@ static inline size_t xfrm_spdinfo_msgsize(void)
 {
 	return NLMSG_ALIGN(4)
 	       + nla_total_size(sizeof(struct xfrmu_spdinfo))
-	       + nla_total_size(sizeof(struct xfrmu_spdhinfo));
+	       + nla_total_size(sizeof(struct xfrmu_spdhinfo))
+	       + nla_total_size(sizeof(struct xfrmu_spdhthresh))
+	       + nla_total_size(sizeof(struct xfrmu_spdhthresh));
 }
 
 static int build_spdinfo(struct sk_buff *skb, struct net *net,
@@ -973,9 +974,11 @@ static int build_spdinfo(struct sk_buff *skb, struct net *net,
 	struct xfrmk_spdinfo si;
 	struct xfrmu_spdinfo spc;
 	struct xfrmu_spdhinfo sph;
+	struct xfrmu_spdhthresh spt4, spt6;
 	struct nlmsghdr *nlh;
 	int err;
 	u32 *f;
+	unsigned lseq;
 
 	nlh = nlmsg_put(skb, portid, seq, XFRM_MSG_NEWSPDINFO, sizeof(u32), 0);
 	if (nlh == NULL) /* shouldn't really happen ... */
@@ -993,9 +996,22 @@ static int build_spdinfo(struct sk_buff *skb, struct net *net,
 	sph.spdhcnt = si.spdhcnt;
 	sph.spdhmcnt = si.spdhmcnt;
 
+	do {
+		lseq = read_seqbegin(&net->xfrm.policy_hthresh.lock);
+
+		spt4.lbits = net->xfrm.policy_hthresh.lbits4;
+		spt4.rbits = net->xfrm.policy_hthresh.rbits4;
+		spt6.lbits = net->xfrm.policy_hthresh.lbits6;
+		spt6.rbits = net->xfrm.policy_hthresh.rbits6;
+	} while (read_seqretry(&net->xfrm.policy_hthresh.lock, lseq));
+
 	err = nla_put(skb, XFRMA_SPD_INFO, sizeof(spc), &spc);
 	if (!err)
 		err = nla_put(skb, XFRMA_SPD_HINFO, sizeof(sph), &sph);
+	if (!err)
+		err = nla_put(skb, XFRMA_SPD_IPV4_HTHRESH, sizeof(spt4), &spt4);
+	if (!err)
+		err = nla_put(skb, XFRMA_SPD_IPV6_HTHRESH, sizeof(spt6), &spt6);
 	if (err) {
 		nlmsg_cancel(skb, nlh);
 		return err;
@@ -1004,6 +1020,51 @@ static int build_spdinfo(struct sk_buff *skb, struct net *net,
 	return nlmsg_end(skb, nlh);
 }
 
+static int xfrm_set_spdinfo(struct sk_buff *skb, struct nlmsghdr *nlh,
+			    struct nlattr **attrs)
+{
+	struct net *net = sock_net(skb->sk);
+	struct xfrmu_spdhthresh *thresh4 = NULL;
+	struct xfrmu_spdhthresh *thresh6 = NULL;
+
+	/* selector prefixlen thresholds to hash policies */
+	if (attrs[XFRMA_SPD_IPV4_HTHRESH]) {
+		struct nlattr *rta = attrs[XFRMA_SPD_IPV4_HTHRESH];
+
+		if (nla_len(rta) < sizeof(*thresh4))
+			return -EINVAL;
+		thresh4 = nla_data(rta);
+		if (thresh4->lbits > 32 || thresh4->rbits > 32)
+			return -EINVAL;
+	}
+	if (attrs[XFRMA_SPD_IPV6_HTHRESH]) {
+		struct nlattr *rta = attrs[XFRMA_SPD_IPV6_HTHRESH];
+
+		if (nla_len(rta) < sizeof(*thresh6))
+			return -EINVAL;
+		thresh6 = nla_data(rta);
+		if (thresh6->lbits > 128 || thresh6->rbits > 128)
+			return -EINVAL;
+	}
+
+	if (thresh4 || thresh6) {
+		write_seqlock(&net->xfrm.policy_hthresh.lock);
+		if (thresh4) {
+			net->xfrm.policy_hthresh.lbits4 = thresh4->lbits;
+			net->xfrm.policy_hthresh.rbits4 = thresh4->rbits;
+		}
+		if (thresh6) {
+			net->xfrm.policy_hthresh.lbits6 = thresh6->lbits;
+			net->xfrm.policy_hthresh.rbits6 = thresh6->rbits;
+		}
+		write_sequnlock(&net->xfrm.policy_hthresh.lock);
+
+		xfrm_policy_hash_rebuild(net);
+	}
+
+	return 0;
+}
+
 static int xfrm_get_spdinfo(struct sk_buff *skb, struct nlmsghdr *nlh,
 		struct nlattr **attrs)
 {
@@ -2274,6 +2335,7 @@ static const int xfrm_msg_min[XFRM_NR_MSGTYPES] = {
 	[XFRM_MSG_REPORT      - XFRM_MSG_BASE] = XMSGSIZE(xfrm_user_report),
 	[XFRM_MSG_MIGRATE     - XFRM_MSG_BASE] = XMSGSIZE(xfrm_userpolicy_id),
 	[XFRM_MSG_GETSADINFO  - XFRM_MSG_BASE] = sizeof(u32),
+	[XFRM_MSG_NEWSPDINFO  - XFRM_MSG_BASE] = sizeof(u32),
 	[XFRM_MSG_GETSPDINFO  - XFRM_MSG_BASE] = sizeof(u32),
 };
 
@@ -2308,10 +2370,17 @@ static const struct nla_policy xfrma_policy[XFRMA_MAX+1] = {
 	[XFRMA_ADDRESS_FILTER]	= { .len = sizeof(struct xfrm_address_filter) },
 };
 
+static const struct nla_policy xfrma_spd_policy[XFRMA_SPD_MAX+1] = {
+	[XFRMA_SPD_IPV4_HTHRESH] = { .len = sizeof(struct xfrmu_spdhthresh) },
+	[XFRMA_SPD_IPV6_HTHRESH] = { .len = sizeof(struct xfrmu_spdhthresh) },
+};
+
 static const struct xfrm_link {
 	int (*doit)(struct sk_buff *, struct nlmsghdr *, struct nlattr **);
 	int (*dump)(struct sk_buff *, struct netlink_callback *);
 	int (*done)(struct netlink_callback *);
+	const struct nla_policy *nla_pol;
+	int nla_max;
 } xfrm_dispatch[XFRM_NR_MSGTYPES] = {
 	[XFRM_MSG_NEWSA       - XFRM_MSG_BASE] = { .doit = xfrm_add_sa        },
 	[XFRM_MSG_DELSA       - XFRM_MSG_BASE] = { .doit = xfrm_del_sa        },
@@ -2335,6 +2404,9 @@ static const struct xfrm_link {
 	[XFRM_MSG_GETAE       - XFRM_MSG_BASE] = { .doit = xfrm_get_ae  },
 	[XFRM_MSG_MIGRATE     - XFRM_MSG_BASE] = { .doit = xfrm_do_migrate    },
 	[XFRM_MSG_GETSADINFO  - XFRM_MSG_BASE] = { .doit = xfrm_get_sadinfo   },
+	[XFRM_MSG_NEWSPDINFO  - XFRM_MSG_BASE] = { .doit = xfrm_set_spdinfo,
+						   .nla_pol = xfrma_spd_policy,
+						   .nla_max = XFRMA_SPD_MAX },
 	[XFRM_MSG_GETSPDINFO  - XFRM_MSG_BASE] = { .doit = xfrm_get_spdinfo   },
 };
 
@@ -2371,8 +2443,9 @@ static int xfrm_user_rcv_msg(struct sk_buff *skb, struct nlmsghdr *nlh)
 		}
 	}
 
-	err = nlmsg_parse(nlh, xfrm_msg_min[type], attrs, XFRMA_MAX,
-			  xfrma_policy);
+	err = nlmsg_parse(nlh, xfrm_msg_min[type], attrs,
+			  link->nla_max ? : XFRMA_MAX,
+			  link->nla_pol ? : xfrma_policy);
 	if (err < 0)
 		return err;