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>
diff --git a/net/key/af_key.c b/net/key/af_key.c
index 3fa811c..9a039ac 100644
--- a/net/key/af_key.c
+++ b/net/key/af_key.c
@@ -1785,7 +1785,9 @@
static void pfkey_dump_sa_done(struct pfkey_sock *pfk)
{
- xfrm_state_walk_done(&pfk->dump.u.state);
+ struct net *net = sock_net(&pfk->sk);
+
+ xfrm_state_walk_done(&pfk->dump.u.state, net);
}
static int pfkey_dump(struct sock *sk, struct sk_buff *skb, const struct sadb_msg *hdr, void * const *ext_hdrs)
@@ -1861,7 +1863,7 @@
reqid = IPSEC_MANUAL_REQID_MAX+1;
xfrm_policy_walk_init(&walk, XFRM_POLICY_TYPE_MAIN);
rc = xfrm_policy_walk(net, &walk, check_reqid, (void*)&reqid);
- xfrm_policy_walk_done(&walk);
+ xfrm_policy_walk_done(&walk, net);
if (rc != -EEXIST)
return reqid;
} while (reqid != start);
@@ -2660,7 +2662,9 @@
static void pfkey_dump_sp_done(struct pfkey_sock *pfk)
{
- xfrm_policy_walk_done(&pfk->dump.u.policy);
+ struct net *net = sock_net((struct sock *)pfk);
+
+ xfrm_policy_walk_done(&pfk->dump.u.policy, net);
}
static int pfkey_spddump(struct sock *sk, struct sk_buff *skb, const struct sadb_msg *hdr, void * const *ext_hdrs)
@@ -3570,6 +3574,7 @@
struct sk_buff *skb = NULL;
struct sadb_msg *hdr = NULL;
int err;
+ struct net *net = sock_net(sk);
err = -EOPNOTSUPP;
if (msg->msg_flags & MSG_OOB)
@@ -3592,9 +3597,9 @@
if (!hdr)
goto out;
- mutex_lock(&xfrm_cfg_mutex);
+ mutex_lock(&net->xfrm.xfrm_cfg_mutex);
err = pfkey_process(sk, skb, hdr);
- mutex_unlock(&xfrm_cfg_mutex);
+ mutex_unlock(&net->xfrm.xfrm_cfg_mutex);
out:
if (err && hdr && pfkey_error(hdr, err, sk) == 0)