af_unix: Annotate data-race of sk->sk_state in unix_accept().

Once sk->sk_state is changed to TCP_LISTEN, it never changes.

unix_accept() takes the advantage and reads sk->sk_state without
holding unix_state_lock().

Let's use READ_ONCE() there.

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Signed-off-by: Kuniyuki Iwashima <[email protected]>
Signed-off-by: Paolo Abeni <[email protected]>

diff --git a/net/unix/af_unix.c b/net/unix/af_unix.c
index 84552826530df0cd0c0fea6e74ab977fd057fc1d..4763c26ae480a1ca06b6aeb478c9713d6f2dd660 100644 (file)
--- a/net/unix/af_unix.c
+++ b/net/unix/af_unix.c
@@ -1710,7 +1710,7 @@ static int unix_accept(struct socket *sock, struct socket *newsock,
                goto out;
 
        arg->err = -EINVAL;
-       if (sk->sk_state != TCP_LISTEN)
+       if (READ_ONCE(sk->sk_state) != TCP_LISTEN)
                goto out;
 
        /* If socket state is TCP_LISTEN it cannot change (for now...),