tcp: remove poll() flakes with FastOpen
authorEric Dumazet <edumazet@google.com>
Tue, 18 Apr 2017 16:45:52 +0000 (09:45 -0700)
committerDavid S. Miller <davem@davemloft.net>
Thu, 20 Apr 2017 19:42:11 +0000 (15:42 -0400)
commit0f9fa831aecfc297b7b45d4f046759bcefcf87f0
tree24e08556cbd89b183cfb02743e268f61346e9997
parent3d4762639dd36a5f0f433f0c9d82e9743dc21a33
tcp: remove poll() flakes with FastOpen

When using TCP FastOpen for an active session, we send one wakeup event
from tcp_finish_connect(), right before the data eventually contained in
the received SYNACK is queued to sk->sk_receive_queue.

This means that depending on machine load or luck, poll() users
might receive POLLOUT events instead of POLLIN|POLLOUT

To fix this, we need to move the call to sk->sk_state_change()
after the (optional) call to tcp_rcv_fastopen_synack()

Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
net/ipv4/tcp_input.c