net: sk_dst_cache RCUification

With latest CONFIG_PROVE_RCU stuff, I felt more comfortable to make this
work.

sk->sk_dst_cache is currently protected by a rwlock (sk_dst_lock)

This rwlock is readlocked for a very small amount of time, and dst
entries are already freed after RCU grace period. This calls for RCU
again :)

This patch converts sk_dst_lock to a spinlock, and use RCU for readers.

__sk_dst_get() is supposed to be called with rcu_read_lock() or if
socket locked by user, so use appropriate rcu_dereference_check()
condition (rcu_read_lock_held() || sock_owned_by_user(sk))

This patch avoids two atomic ops per tx packet on UDP connected sockets,
for example, and permits sk_dst_lock to be much less dirtied.

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

1 files changed, 13 insertions, 12 deletions

diff --git a/net/ipv6/ipv6_sockglue.c b/net/ipv6/ipv6_sockglue.c
index 33f60fc..1160400 100644
--- a/net/ipv6/ipv6_sockglue.c
+++ b/net/ipv6/ipv6_sockglue.c
@@ -114,9 +114,9 @@ struct ipv6_txoptions *ipv6_update_options(struct sock *sk,
 		}
 		opt = xchg(&inet6_sk(sk)->opt, opt);
 	} else {
-		write_lock(&sk->sk_dst_lock);
+		spin_lock(&sk->sk_dst_lock);
 		opt = xchg(&inet6_sk(sk)->opt, opt);
-		write_unlock(&sk->sk_dst_lock);
+		spin_unlock(&sk->sk_dst_lock);
 	}
 	sk_dst_reset(sk);
 
@@ -971,14 +971,13 @@ static int do_ipv6_getsockopt(struct sock *sk, int level, int optname,
 	case IPV6_MTU:
 	{
 		struct dst_entry *dst;
+
 		val = 0;
-		lock_sock(sk);
-		dst = sk_dst_get(sk);
-		if (dst) {
+		rcu_read_lock();
+		dst = __sk_dst_get(sk);
+		if (dst)
 			val = dst_mtu(dst);
-			dst_release(dst);
-		}
-		release_sock(sk);
+		rcu_read_unlock();
 		if (!val)
 			return -ENOTCONN;
 		break;
@@ -1066,12 +1065,14 @@ static int do_ipv6_getsockopt(struct sock *sk, int level, int optname,
 		else
 			val = np->mcast_hops;
 
-		dst = sk_dst_get(sk);
-		if (dst) {
-			if (val < 0)
+		if (val < 0) {
+			rcu_read_lock();
+			dst = __sk_dst_get(sk);
+			if (dst)
 				val = ip6_dst_hoplimit(dst);
-			dst_release(dst);
+			rcu_read_unlock();
 		}
+
 		if (val < 0)
 			val = sock_net(sk)->ipv6.devconf_all->hop_limit;
 		break;