Merge branch 'inet_csk_get_port-and-soreusport-fixes'

Tom Herbert says:

====================
inet: Fixes for inet_csk_get_port and soreusport

This patch set fixes a couple of issues I noticed while debugging our
softlockup issue in inet_csk_get_port.

- Don't allow jump into port scan in inet_csk_get_port if function
  was called with non-zero port number (looking up explicit port
  number).
- When inet_csk_get_port is called with zero port number (ie. perform
  scan) an reuseport is set on the socket, don't match sockets that
  also have reuseport set. The intent from the user should be
  to get a new port number and then explictly bind other
  sockets to that number using soreuseport.

Tested:

Ran first patch on production workload with no ill effect.

For second patch, ran a little listener application and first
demonstrated that unbound sockets with soreuseport can indeed
be bound to unrelated soreuseport sockets.
====================

Signed-off-by: David S. Miller <[email protected]>

include/net/inet6_connection_sock.h

@@ -22,7 +22,8 @@ struct sock;
 struct sockaddr;
 
 int inet6_csk_bind_conflict(const struct sock *sk,
-			    const struct inet_bind_bucket *tb, bool relax);
+			    const struct inet_bind_bucket *tb, bool relax,
+			    bool soreuseport_ok);
 
 struct dst_entry *inet6_csk_route_req(const struct sock *sk, struct flowi6 *fl6,
 				      const struct request_sock *req, u8 proto);

include/net/inet_connection_sock.h

@@ -63,7 +63,8 @@ struct inet_connection_sock_af_ops {
 #endif
 	void	    (*addr2sockaddr)(struct sock *sk, struct sockaddr *);
 	int	    (*bind_conflict)(const struct sock *sk,
-				     const struct inet_bind_bucket *tb, bool relax);
+				     const struct inet_bind_bucket *tb,
+				     bool relax, bool soreuseport_ok);
 	void	    (*mtu_reduced)(struct sock *sk);
 };
 
@@ -261,7 +262,8 @@ inet_csk_rto_backoff(const struct inet_connection_sock *icsk,
 struct sock *inet_csk_accept(struct sock *sk, int flags, int *err);
 
 int inet_csk_bind_conflict(const struct sock *sk,
-			   const struct inet_bind_bucket *tb, bool relax);
+			   const struct inet_bind_bucket *tb, bool relax,
+			   bool soreuseport_ok);
 int inet_csk_get_port(struct sock *sk, unsigned short snum);
 
 struct dst_entry *inet_csk_route_req(const struct sock *sk, struct flowi4 *fl4,

net/ipv4/inet_connection_sock.c

@@ -45,11 +45,12 @@ void inet_get_local_port_range(struct net *net, int *low, int *high)
 EXPORT_SYMBOL(inet_get_local_port_range);
 
 int inet_csk_bind_conflict(const struct sock *sk,
-			   const struct inet_bind_bucket *tb, bool relax)
+			   const struct inet_bind_bucket *tb, bool relax,
+			   bool reuseport_ok)
 {
 	struct sock *sk2;
-	int reuse = sk->sk_reuse;
-	int reuseport = sk->sk_reuseport;
+	bool reuse = sk->sk_reuse;
+	bool reuseport = !!sk->sk_reuseport && reuseport_ok;
 	kuid_t uid = sock_i_uid((struct sock *)sk);
 
 	/*
@@ -105,6 +106,7 @@ int inet_csk_get_port(struct sock *sk, unsigned short snum)
 	struct inet_bind_bucket *tb;
 	kuid_t uid = sock_i_uid(sk);
 	u32 remaining, offset;
+	bool reuseport_ok = !!snum;
 
 	if (port) {
 have_port:
@@ -165,7 +167,8 @@ int inet_csk_get_port(struct sock *sk, unsigned short snum)
 					smallest_size = tb->num_owners;
 					smallest_port = port;
 				}
-				if (!inet_csk(sk)->icsk_af_ops->bind_conflict(sk, tb, false))
+				if (!inet_csk(sk)->icsk_af_ops->bind_conflict(sk, tb, false,
+									      reuseport_ok))
 					goto tb_found;
 				goto next_port;
 			}
@@ -206,13 +209,14 @@ int inet_csk_get_port(struct sock *sk, unsigned short snum)
 		      sk->sk_reuseport && uid_eq(tb->fastuid, uid))) &&
 		    smallest_size == -1)
 			goto success;
-		if (inet_csk(sk)->icsk_af_ops->bind_conflict(sk, tb, true)) {
+		if (inet_csk(sk)->icsk_af_ops->bind_conflict(sk, tb, true,
+							     reuseport_ok)) {
 			if ((reuse ||
 			     (tb->fastreuseport > 0 &&
 			      sk->sk_reuseport &&
 			      !rcu_access_pointer(sk->sk_reuseport_cb) &&
 			      uid_eq(tb->fastuid, uid))) &&
-			    smallest_size != -1 && --attempts >= 0) {
+			    !snum && smallest_size != -1 && --attempts >= 0) {
 				spin_unlock_bh(&head->lock);
 				goto again;
 			}

net/ipv6/inet6_connection_sock.c

@@ -29,11 +29,12 @@
 #include <net/sock_reuseport.h>
 
 int inet6_csk_bind_conflict(const struct sock *sk,
-			    const struct inet_bind_bucket *tb, bool relax)
+			    const struct inet_bind_bucket *tb, bool relax,
+			    bool reuseport_ok)
 {
 	const struct sock *sk2;
-	int reuse = sk->sk_reuse;
-	int reuseport = sk->sk_reuseport;
+	bool reuse = !!sk->sk_reuse;
+	bool reuseport = !!sk->sk_reuseport && reuseport_ok;
 	kuid_t uid = sock_i_uid((struct sock *)sk);
 
 	/* We must walk the whole port owner list in this case. -DaveM */