tcp: bind(0) remove the SO_REUSEADDR restriction when ephemeral ports are exhausted.

Commit aacd9289af8b82f5fb01bcdd53d0e3406d1333c7 ("tcp: bind() use stronger
condition for bind_conflict") introduced a restriction to forbid to bind
SO_REUSEADDR enabled sockets to the same (addr, port) tuple in order to
assign ports dispersedly so that we can connect to the same remote host.

The change results in accelerating port depletion so that we fail to bind
sockets to the same local port even if we want to connect to the different
remote hosts.

You can reproduce this issue by following instructions below.

  1. # sysctl -w net.ipv4.ip_local_port_range="32768 32768"
  2. set SO_REUSEADDR to two sockets.
  3. bind two sockets to (localhost, 0) and the latter fails.

Therefore, when ephemeral ports are exhausted, bind(0) should fallback to
the legacy behaviour to enable the SO_REUSEADDR option and make it possible
to connect to different remote (addr, port) tuples.

This patch allows us to bind SO_REUSEADDR enabled sockets to the same
(addr, port) only when net.ipv4.ip_autobind_reuse is set 1 and all
ephemeral ports are exhausted. This also allows connect() and listen() to
share ports in the following way and may break some applications. So the
ip_autobind_reuse is 0 by default and disables the feature.

  1. setsockopt(sk1, SO_REUSEADDR)
  2. setsockopt(sk2, SO_REUSEADDR)
  3. bind(sk1, saddr, 0)
  4. bind(sk2, saddr, 0)
  5. connect(sk1, daddr)
  6. listen(sk2)

If it is set 1, we can fully utilize the 4-tuples, but we should use
IP_BIND_ADDRESS_NO_PORT for bind()+connect() as possible.

The notable thing is that if all sockets bound to the same port have
both SO_REUSEADDR and SO_REUSEPORT enabled, we can bind sockets to an
ephemeral port and also do listen().

Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.co.jp>
Signed-off-by: David S. Miller <davem@davemloft.net>

4 files changed, 28 insertions, 1 deletions

diff --git a/Documentation/networking/ip-sysctl.txt b/Documentation/networking/ip-sysctl.txt
index 5f53faff4e25..ee961d322d93 100644
--- a/Documentation/networking/ip-sysctl.txt
+++ b/Documentation/networking/ip-sysctl.txt
@@ -958,6 +958,15 @@ ip_nonlocal_bind - BOOLEAN
 	which can be quite useful - but may break some applications.
 	Default: 0
 
+ip_autobind_reuse - BOOLEAN
+	By default, bind() does not select the ports automatically even if
+	the new socket and all sockets bound to the port have SO_REUSEADDR.
+	ip_autobind_reuse allows bind() to reuse the port and this is useful
+	when you use bind()+connect(), but may break some applications.
+	The preferred solution is to use IP_BIND_ADDRESS_NO_PORT and this
+	option should only be set by experts.
+	Default: 0
+
 ip_dynaddr - BOOLEAN
 	If set non-zero, enables support for dynamic addresses.
 	If set to a non-zero value larger than 1, a kernel log
diff --git a/include/net/netns/ipv4.h b/include/net/netns/ipv4.h
index 08b98414d94e..154b8f01499b 100644
--- a/include/net/netns/ipv4.h
+++ b/include/net/netns/ipv4.h
@@ -101,6 +101,7 @@ struct netns_ipv4 {
 	int sysctl_ip_fwd_use_pmtu;
 	int sysctl_ip_fwd_update_priority;
 	int sysctl_ip_nonlocal_bind;
+	int sysctl_ip_autobind_reuse;
 	/* Shall we try to damage output packets if routing dev changes? */
 	int sysctl_ip_dynaddr;
 	int sysctl_ip_early_demux;
diff --git a/net/ipv4/inet_connection_sock.c b/net/ipv4/inet_connection_sock.c
index 2e9549f49a82..497366b631f3 100644
--- a/net/ipv4/inet_connection_sock.c
+++ b/net/ipv4/inet_connection_sock.c
@@ -174,12 +174,14 @@ inet_csk_find_open_port(struct sock *sk, struct inet_bind_bucket **tb_ret, int *
 	int port = 0;
 	struct inet_bind_hashbucket *head;
 	struct net *net = sock_net(sk);
+	bool relax = false;
 	int i, low, high, attempt_half;
 	struct inet_bind_bucket *tb;
 	u32 remaining, offset;
 	int l3mdev;
 
 	l3mdev = inet_sk_bound_l3mdev(sk);
+ports_exhausted:
 	attempt_half = (sk->sk_reuse == SK_CAN_REUSE) ? 1 : 0;
 other_half_scan:
 	inet_get_local_port_range(net, &low, &high);
@@ -217,7 +219,7 @@ other_parity_scan:
 		inet_bind_bucket_for_each(tb, &head->chain)
 			if (net_eq(ib_net(tb), net) && tb->l3mdev == l3mdev &&
 			    tb->port == port) {
-				if (!inet_csk_bind_conflict(sk, tb, false, false))
+				if (!inet_csk_bind_conflict(sk, tb, relax, false))
 					goto success;
 				goto next_port;
 			}
@@ -237,6 +239,12 @@ next_port:
 		attempt_half = 2;
 		goto other_half_scan;
 	}
+
+	if (net->ipv4.sysctl_ip_autobind_reuse && !relax) {
+		/* We still have a chance to connect to different destinations */
+		relax = true;
+		goto ports_exhausted;
+	}
 	return NULL;
 success:
 	*port_ret = port;
diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c
index d9531b4b33f2..81b267e990a1 100644
--- a/net/ipv4/sysctl_net_ipv4.c
+++ b/net/ipv4/sysctl_net_ipv4.c
@@ -764,6 +764,15 @@ static struct ctl_table ipv4_net_table[] = {
 		.proc_handler	= proc_dointvec
 	},
 	{
+		.procname	= "ip_autobind_reuse",
+		.data		= &init_net.ipv4.sysctl_ip_autobind_reuse,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec_minmax,
+		.extra1         = SYSCTL_ZERO,
+		.extra2         = SYSCTL_ONE,
+	},
+	{
 		.procname	= "fwmark_reflect",
 		.data		= &init_net.ipv4.sysctl_fwmark_reflect,
 		.maxlen		= sizeof(int),