mptcp: fix race in msk status update

Currently subflow_finish_connect() changes unconditionally
any msk socket status other than TCP_ESTABLISHED.

If an unblocking connect() races with close(), we can end-up
triggering:

IPv4: Attempt to release TCP socket in state 1 00000000e32b8b7e

when the msk socket is disposed.

Be sure to enter the established status only from SYN_SENT.

Fixes: c3c123d16c0e ("net: mptcp: don't hang in mptcp_sendmsg() after TCP fallback")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

diff --git a/net/mptcp/subflow.c b/net/mptcp/subflow.c
index fabd06f2ff455e76cf281c4e6fc2d41b144ac2d9..71256f03707fe6063c99220cc11fc29b6cb7354d 100644 (file)
--- a/net/mptcp/subflow.c
+++ b/net/mptcp/subflow.c
@@ -225,7 +225,7 @@ static void subflow_finish_connect(struct sock *sk, const struct sk_buff *skb)
 
        subflow->icsk_af_ops->sk_rx_dst_set(sk, skb);
 
-       if (inet_sk_state_load(parent) != TCP_ESTABLISHED) {
+       if (inet_sk_state_load(parent) == TCP_SYN_SENT) {
                inet_sk_state_store(parent, TCP_ESTABLISHED);
                parent->sk_state_change(parent);
        }