Re: [PATCH 03/13] libceph: delete useless SOCK_CLOSED manipulations

[Alex Elder <elder@inktank.com>]

On 05/30/2012 02:34 PM, Alex Elder wrote:
> In con_close_socket(), SOCK_CLOSED is set in the connection state,
> then cleared again after shutting down the socket. Nothing between
> the setting and clearing of that bit will ever be affected by it,
> so there's no point in setting/clearing it at all. So don't.
>
> Signed-off-by: Alex Elder <elder@inktank.com>

I am retracting this proposed change.

I believe it's possible for the con->sock->ops->shutdown()
call to trigger a TCP_CLOSE socket state change event,
which means that there *is* something that can be affected
by that state bit being set.

					-Alex

> ---
> net/ceph/messenger.c | 2 --
> 1 files changed, 0 insertions(+), 2 deletions(-)
>
> diff --git a/net/ceph/messenger.c b/net/ceph/messenger.c
> index 07af994..fe3c2a1 100644
> --- a/net/ceph/messenger.c
> +++ b/net/ceph/messenger.c
> @@ -338,11 +338,9 @@ static int con_close_socket(struct ceph_connection
> *con)
> dout("con_close_socket on %p sock %p\n", con, con->sock);
> if (!con->sock)
> return 0;
> - set_bit(SOCK_CLOSED, &con->state);
> rc = con->sock->ops->shutdown(con->sock, SHUT_RDWR);
> sock_release(con->sock);
> con->sock = NULL;
> - clear_bit(SOCK_CLOSED, &con->state);
> return rc;
> }
>