soreuseport: fix NULL ptr dereference SO_REUSEPORT after bind

Marc Dionne discovered a NULL pointer dereference when setting
SO_REUSEPORT on a socket after it is bound.
This patch removes the assumption that at least one socket in the
reuseport group is bound with the SO_REUSEPORT option before other
bind calls occur.

Fixes: e32ea7e ("soreuseport: fast reuseport UDP socket selection")
Reported-by: Marc Dionne <marc.c.dionne@gmail.com>
Signed-off-by: Craig Gallek <kraig@google.com>
Tested-by: Marc Dionne <marc.dionne@auristor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

include/net/sock_reuseport.h

@@ -16,7 +16,7 @@ struct sock_reuseport {
 };
 
 extern int reuseport_alloc(struct sock *sk);
-extern int reuseport_add_sock(struct sock *sk, const struct sock *sk2);
+extern int reuseport_add_sock(struct sock *sk, struct sock *sk2);
 extern void reuseport_detach_sock(struct sock *sk);
 extern struct sock *reuseport_select_sock(struct sock *sk,
 					  u32 hash,

net/core/sock_reuseport.c

@@ -93,10 +93,17 @@ static struct sock_reuseport *reuseport_grow(struct sock_reuseport *reuse)
  *  @sk2: Socket belonging to the existing reuseport group.
  *  May return ENOMEM and not add socket to group under memory pressure.
  */
-int reuseport_add_sock(struct sock *sk, const struct sock *sk2)
+int reuseport_add_sock(struct sock *sk, struct sock *sk2)
 {
 	struct sock_reuseport *reuse;
 
+	if (!rcu_access_pointer(sk2->sk_reuseport_cb)) {
+		int err = reuseport_alloc(sk2);
+
+		if (err)
+			return err;
+	}
+
 	spin_lock_bh(&reuseport_lock);
 	reuse = rcu_dereference_protected(sk2->sk_reuseport_cb,
 					  lockdep_is_held(&reuseport_lock)),