[TCP]: Allow choosing TCP congestion control via sockopt.

Allow using setsockopt to set TCP congestion control to use on a per
socket basis.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

4 files changed, 76 insertions, 5 deletions

diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index f3dbc8dc1263..882436da9a3a 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -1927,6 +1927,25 @@ int tcp_setsockopt(struct sock *sk, int level, int optname, char __user *optval,
 		return tp->af_specific->setsockopt(sk, level, optname,
 						   optval, optlen);
 
+	/* This is a string value all the others are int's */
+	if (optname == TCP_CONGESTION) {
+		char name[TCP_CA_NAME_MAX];
+
+		if (optlen < 1)
+			return -EINVAL;
+
+		val = strncpy_from_user(name, optval,
+					min(TCP_CA_NAME_MAX-1, optlen));
+		if (val < 0)
+			return -EFAULT;
+		name[val] = 0;
+
+		lock_sock(sk);
+		err = tcp_set_congestion_control(tp, name);
+		release_sock(sk);
+		return err;
+	}
+
 	if (optlen < sizeof(int))
 		return -EINVAL;
 
@@ -2211,6 +2230,16 @@ int tcp_getsockopt(struct sock *sk, int level, int optname, char __user *optval,
 	case TCP_QUICKACK:
 		val = !tp->ack.pingpong;
 		break;
+
+	case TCP_CONGESTION:
+		if (get_user(len, optlen))
+			return -EFAULT;
+		len = min_t(unsigned int, len, TCP_CA_NAME_MAX);
+		if (put_user(len, optlen))
+			return -EFAULT;
+		if (copy_to_user(optval, tp->ca_ops->name, len))
+			return -EFAULT;
+		return 0;
 	default:
 		return -ENOPROTOOPT;
 	};
@@ -2224,7 +2253,7 @@ int tcp_getsockopt(struct sock *sk, int level, int optname, char __user *optval,
 
 
 extern void __skb_cb_too_small_for_tcp(int, int);
-extern void tcpdiag_init(void);
+extern struct tcp_congestion_ops tcp_reno;
 
 static __initdata unsigned long thash_entries;
 static int __init set_thash_entries(char *str)
diff --git a/net/ipv4/tcp_cong.c b/net/ipv4/tcp_cong.c
index 665394a63ae4..4970d10a7785 100644
--- a/net/ipv4/tcp_cong.c
+++ b/net/ipv4/tcp_cong.c
@@ -21,7 +21,7 @@ static struct tcp_congestion_ops *tcp_ca_find(const char *name)
 {
 	struct tcp_congestion_ops *e;
 
-	list_for_each_entry(e, &tcp_cong_list, list) {
+	list_for_each_entry_rcu(e, &tcp_cong_list, list) {
 		if (strcmp(e->name, name) == 0)
 			return e;
 	}
@@ -77,6 +77,9 @@ void tcp_init_congestion_control(struct tcp_sock *tp)
 {
 	struct tcp_congestion_ops *ca;
 
+	if (tp->ca_ops != &tcp_init_congestion_ops)
+		return;
+
 	rcu_read_lock();
 	list_for_each_entry_rcu(ca, &tcp_cong_list, list) {
 		if (try_module_get(ca->owner)) {
@@ -139,6 +142,34 @@ void tcp_get_default_congestion_control(char *name)
 	rcu_read_unlock();
 }
 
+/* Change congestion control for socket */
+int tcp_set_congestion_control(struct tcp_sock *tp, const char *name)
+{
+	struct tcp_congestion_ops *ca;
+	int err = 0;
+
+	rcu_read_lock();
+	ca = tcp_ca_find(name);
+	if (ca == tp->ca_ops)
+		goto out;
+
+	if (!ca)
+		err = -ENOENT;
+
+	else if (!try_module_get(ca->owner))
+		err = -EBUSY;
+
+	else {
+		tcp_cleanup_congestion_control(tp);
+		tp->ca_ops = ca;
+		if (tp->ca_ops->init)
+			tp->ca_ops->init(tp);
+	}
+ out:
+	rcu_read_unlock();
+	return err;
+}
+
 /*
  * TCP Reno congestion control
  * This is special case used for fallback as well.
@@ -192,4 +223,15 @@ struct tcp_congestion_ops tcp_reno = {
 	.min_cwnd	= tcp_reno_min_cwnd,
 };
 
-EXPORT_SYMBOL_GPL(tcp_reno);
+/* Initial congestion control used (until SYN)
+ * really reno under another name so we can tell difference
+ * during tcp_set_default_congestion_control
+ */
+struct tcp_congestion_ops tcp_init_congestion_ops  = {
+	.name		= "",
+	.owner		= THIS_MODULE,
+	.ssthresh	= tcp_reno_ssthresh,
+	.cong_avoid	= tcp_reno_cong_avoid,
+	.min_cwnd	= tcp_reno_min_cwnd,
+};
+EXPORT_SYMBOL_GPL(tcp_init_congestion_ops);
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 9122814c13ad..ebf112347a97 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -2048,7 +2048,7 @@ static int tcp_v4_init_sock(struct sock *sk)
 	tp->mss_cache_std = tp->mss_cache = 536;
 
 	tp->reordering = sysctl_tcp_reordering;
-	tp->ca_ops = &tcp_reno;
+	tp->ca_ops = &tcp_init_congestion_ops;
 
 	sk->sk_state = TCP_CLOSE;
 
diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
index fce56039b0e9..9dac7fdf4726 100644
--- a/net/ipv6/tcp_ipv6.c
+++ b/net/ipv6/tcp_ipv6.c
@@ -2025,7 +2025,7 @@ static int tcp_v6_init_sock(struct sock *sk)
 	sk->sk_state = TCP_CLOSE;
 
 	tp->af_specific = &ipv6_specific;
-	tp->ca_ops = &tcp_reno;
+	tp->ca_ops = &tcp_init_congestion_ops;
 	sk->sk_write_space = sk_stream_write_space;
 	sock_set_flag(sk, SOCK_USE_WRITE_QUEUE);