On Thu, 2018-05-10 at 06:12 +0000, Vallish Vaidyeshwara wrote:
> Server initiated socket close can corrupt connection state tracking
> table in conjunction with other network middle boxes. In situations
> like these, client connection hangs till connection state tracking
> table entries age out and get purged. Client reconnection with a new
> port in such a situation will avoid connection hang.
>
> Reviewed-by: Jacob Strauss <jsstraus@xxxxxxxxxx>
> Reviewed-by: Alakesh Haloi <alakeshh@xxxxxxxxxx>
> Signed-off-by: Vallish Vaidyeshwara <vallish@xxxxxxxxxx>
> ---
> net/sunrpc/xprtsock.c | 5 +++++
> 1 file changed, 5 insertions(+)
>
> diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c
> index 5bf75b3..d293c8d 100644
> --- a/net/sunrpc/xprtsock.c
> +++ b/net/sunrpc/xprtsock.c
> @@ -1629,6 +1629,8 @@ static void xs_tcp_state_change(struct sock
> *sk)
> /* The server initiated a shutdown of the socket */
> xprt->connect_cookie++;
> clear_bit(XPRT_CONNECTED, &xprt->state);
> + /* Server sent FIN, reconnect with a new port */
> + transport->srcport = 0;
> xs_tcp_force_close(xprt);
> /* fall through */
> case TCP_CLOSING:
> @@ -1650,6 +1652,9 @@ static void xs_tcp_state_change(struct sock
> *sk)
> &transport->sock_state))
> xprt_clear_connecting(xprt);
> clear_bit(XPRT_CLOSING, &xprt->state);
> + /* Server sent RST, reconnect with a new port */
> + if (sk->sk_err == ECONNRESET)
> + transport->srcport = 0;
> if (sk->sk_err)
> xprt_wake_pending_tasks(xprt, -sk->sk_err);
> /* Trigger the socket release */
NACK. This will utterly break NFSv2, NFSv3 and NFSv4.0 duplicate replay
cache semantics.
Cheers
Trond
--
Trond Myklebust
Linux NFS client maintainer, Hammerspace
trond.myklebust@xxxxxxxxxxxxxxx
��.n��������+%������w��{.n�����{��w��� jg��� �����ݢj����G���� ���j:+v���w�m������w�������h�����٥
