tcp_connect.c 4.3 KB
Newer Older
A
Andy Grover 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45
/*
 * Copyright (c) 2006 Oracle.  All rights reserved.
 *
 * This software is available to you under a choice of one of two
 * licenses.  You may choose to be licensed under the terms of the GNU
 * General Public License (GPL) Version 2, available from the file
 * COPYING in the main directory of this source tree, or the
 * OpenIB.org BSD license below:
 *
 *     Redistribution and use in source and binary forms, with or
 *     without modification, are permitted provided that the following
 *     conditions are met:
 *
 *      - Redistributions of source code must retain the above
 *        copyright notice, this list of conditions and the following
 *        disclaimer.
 *
 *      - Redistributions in binary form must reproduce the above
 *        copyright notice, this list of conditions and the following
 *        disclaimer in the documentation and/or other materials
 *        provided with the distribution.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
 * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
 * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
 * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
 * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
 * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
 * SOFTWARE.
 *
 */
#include <linux/kernel.h>
#include <linux/in.h>
#include <net/tcp.h>

#include "rds.h"
#include "tcp.h"

void rds_tcp_state_change(struct sock *sk)
{
	void (*state_change)(struct sock *sk);
	struct rds_connection *conn;
	struct rds_tcp_connection *tc;

Y
Ying Xue 已提交
46
	read_lock(&sk->sk_callback_lock);
A
Andy Grover 已提交
47
	conn = sk->sk_user_data;
48
	if (!conn) {
A
Andy Grover 已提交
49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64
		state_change = sk->sk_state_change;
		goto out;
	}
	tc = conn->c_transport_data;
	state_change = tc->t_orig_state_change;

	rdsdebug("sock %p state_change to %d\n", tc->t_sock, sk->sk_state);

	switch(sk->sk_state) {
		/* ignore connecting sockets as they make progress */
		case TCP_SYN_SENT:
		case TCP_SYN_RECV:
			break;
		case TCP_ESTABLISHED:
			rds_connect_complete(conn);
			break;
65
		case TCP_CLOSE_WAIT:
A
Andy Grover 已提交
66 67 68 69 70 71
		case TCP_CLOSE:
			rds_conn_drop(conn);
		default:
			break;
	}
out:
Y
Ying Xue 已提交
72
	read_unlock(&sk->sk_callback_lock);
A
Andy Grover 已提交
73 74 75 76 77 78 79 80 81
	state_change(sk);
}

int rds_tcp_conn_connect(struct rds_connection *conn)
{
	struct socket *sock = NULL;
	struct sockaddr_in src, dest;
	int ret;

82 83
	ret = sock_create_kern(rds_conn_net(conn), PF_INET,
			       SOCK_STREAM, IPPROTO_TCP, &sock);
A
Andy Grover 已提交
84 85 86 87 88 89 90 91 92 93 94
	if (ret < 0)
		goto out;

	rds_tcp_tune(sock);

	src.sin_family = AF_INET;
	src.sin_addr.s_addr = (__force u32)conn->c_laddr;
	src.sin_port = (__force u16)htons(0);

	ret = sock->ops->bind(sock, (struct sockaddr *)&src, sizeof(src));
	if (ret) {
95 96
		rdsdebug("bind failed with %d at address %pI4\n",
			 ret, &conn->c_laddr);
A
Andy Grover 已提交
97 98 99 100 101 102 103 104 105 106 107 108 109 110 111
		goto out;
	}

	dest.sin_family = AF_INET;
	dest.sin_addr.s_addr = (__force u32)conn->c_faddr;
	dest.sin_port = (__force u16)htons(RDS_TCP_PORT);

	/*
	 * once we call connect() we can start getting callbacks and they
	 * own the socket
	 */
	rds_tcp_set_callbacks(sock, conn);
	ret = sock->ops->connect(sock, (struct sockaddr *)&dest, sizeof(dest),
				 O_NONBLOCK);

112
	rdsdebug("connect to address %pI4 returned %d\n", &conn->c_faddr, ret);
A
Andy Grover 已提交
113 114
	if (ret == -EINPROGRESS)
		ret = 0;
115 116
	if (ret == 0) {
		rds_tcp_keepalive(sock);
117
		sock = NULL;
118
	} else {
119
		rds_tcp_restore_callbacks(sock, conn->c_transport_data);
120
	}
A
Andy Grover 已提交
121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150

out:
	if (sock)
		sock_release(sock);
	return ret;
}

/*
 * Before killing the tcp socket this needs to serialize with callbacks.  The
 * caller has already grabbed the sending sem so we're serialized with other
 * senders.
 *
 * TCP calls the callbacks with the sock lock so we hold it while we reset the
 * callbacks to those set by TCP.  Our callbacks won't execute again once we
 * hold the sock lock.
 */
void rds_tcp_conn_shutdown(struct rds_connection *conn)
{
	struct rds_tcp_connection *tc = conn->c_transport_data;
	struct socket *sock = tc->t_sock;

	rdsdebug("shutting down conn %p tc %p sock %p\n", conn, tc, sock);

	if (sock) {
		sock->ops->shutdown(sock, RCV_SHUTDOWN | SEND_SHUTDOWN);
		lock_sock(sock->sk);
		rds_tcp_restore_callbacks(sock, tc); /* tc->tc_sock = NULL */

		release_sock(sock->sk);
		sock_release(sock);
151
	}
A
Andy Grover 已提交
152 153 154 155 156 157 158 159

	if (tc->t_tinc) {
		rds_inc_put(&tc->t_tinc->ti_inc);
		tc->t_tinc = NULL;
	}
	tc->t_tinc_hdr_rem = sizeof(struct rds_header);
	tc->t_tinc_data_rem = 0;
}