tls: improve lockless access safety of tls_err_abort()

Most protos' poll() methods insert a memory barrier between
writes to sk_err and sk_error_report(). This dates back to
commit a4d258036e ("tcp: Fix race in tcp_poll").

I guess we should do the same thing in TLS, tcp_poll() does
not hold the socket lock.

Fixes: 3c4d755915 ("tls: kernel TLS support")
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net/tls/tls_strp.c

@@ -20,7 +20,9 @@ static void tls_strp_abort_strp(struct tls_strparser *strp, int err)
 	strp->stopped = 1;
 
 	/* Report an error on the lower socket */
-	strp->sk->sk_err = -err;
+	WRITE_ONCE(strp->sk->sk_err, -err);
+	/* Paired with smp_rmb() in tcp_poll() */
+	smp_wmb();
 	sk_error_report(strp->sk);
 }
 

net/tls/tls_sw.c

@@ -70,7 +70,9 @@ noinline void tls_err_abort(struct sock *sk, int err)
 {
 	WARN_ON_ONCE(err >= 0);
 	/* sk->sk_err should contain a positive error code. */
-	sk->sk_err = -err;
+	WRITE_ONCE(sk->sk_err, -err);
+	/* Paired with smp_rmb() in tcp_poll() */
+	smp_wmb();
 	sk_error_report(sk);
 }