xfrm: Namespacify xfrm state/policy locks

By semantics, xfrm layer is fully name space aware,
so will the locks, e.g. xfrm_state/pocliy_lock.
Ensure exclusive access into state/policy link list
for different name space with one global lock is not
right in terms of semantics aspect at first place,
as they are indeed mutually independent with each
other, but also more seriously causes scalability
problem.

One practical scenario is on a Open Network Stack,
more than hundreds of lxc tenants acts as routers
within one host, a global xfrm_state/policy_lock
becomes the bottleneck. But onces those locks are
decoupled in a per-namespace fashion, locks contend
is just with in specific name space scope, without
causing additional SPD/SAD access delay for other
name space.

Also this patch improve scalability while as without
changing original xfrm behavior.

Signed-off-by: Fan Du <fan.du@windriver.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

1 files changed, 10 insertions, 4 deletions

diff --git a/net/xfrm/xfrm_user.c b/net/xfrm/xfrm_user.c
index 840cc8d9fbe4..16c84608e81e 100644
--- a/net/xfrm/xfrm_user.c
+++ b/net/xfrm/xfrm_user.c
@@ -877,7 +877,10 @@ static int dump_one_state(struct xfrm_state *x, int count, void *ptr)
 static int xfrm_dump_sa_done(struct netlink_callback *cb)
 {
 	struct xfrm_state_walk *walk = (struct xfrm_state_walk *) &cb->args[1];
-	xfrm_state_walk_done(walk);
+	struct sock *sk = cb->skb->sk;
+	struct net *net = sock_net(sk);
+
+	xfrm_state_walk_done(walk, net);
 	return 0;
 }
 
@@ -1555,8 +1558,9 @@ static int dump_one_policy(struct xfrm_policy *xp, int dir, int count, void *ptr
 static int xfrm_dump_policy_done(struct netlink_callback *cb)
 {
 	struct xfrm_policy_walk *walk = (struct xfrm_policy_walk *) &cb->args[1];
+	struct net *net = sock_net(cb->skb->sk);
 
-	xfrm_policy_walk_done(walk);
+	xfrm_policy_walk_done(walk, net);
 	return 0;
 }
 
@@ -2403,9 +2407,11 @@ static int xfrm_user_rcv_msg(struct sk_buff *skb, struct nlmsghdr *nlh)
 
 static void xfrm_netlink_rcv(struct sk_buff *skb)
 {
-	mutex_lock(&xfrm_cfg_mutex);
+	struct net *net = sock_net(skb->sk);
+
+	mutex_lock(&net->xfrm.xfrm_cfg_mutex);
 	netlink_rcv_skb(skb, &xfrm_user_rcv_msg);
-	mutex_unlock(&xfrm_cfg_mutex);
+	mutex_unlock(&net->xfrm.xfrm_cfg_mutex);
 }
 
 static inline size_t xfrm_expire_msgsize(void)