2009-08-21 20:28:31 +08:00
|
|
|
/*
|
|
|
|
* 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>
|
|
|
|
|
2016-06-14 00:44:26 +08:00
|
|
|
#include "rds_single_path.h"
|
2009-08-21 20:28:31 +08:00
|
|
|
#include "rds.h"
|
|
|
|
#include "tcp.h"
|
|
|
|
|
|
|
|
static void rds_tcp_cork(struct socket *sock, int val)
|
|
|
|
{
|
2017-03-19 09:20:27 +08:00
|
|
|
kernel_setsockopt(sock, SOL_TCP, TCP_CORK, (void *)&val, sizeof(val));
|
2009-08-21 20:28:31 +08:00
|
|
|
}
|
|
|
|
|
2016-07-01 07:11:10 +08:00
|
|
|
void rds_tcp_xmit_path_prepare(struct rds_conn_path *cp)
|
2009-08-21 20:28:31 +08:00
|
|
|
{
|
2016-07-01 07:11:10 +08:00
|
|
|
struct rds_tcp_connection *tc = cp->cp_transport_data;
|
2009-08-21 20:28:31 +08:00
|
|
|
|
|
|
|
rds_tcp_cork(tc->t_sock, 1);
|
|
|
|
}
|
|
|
|
|
2016-07-01 07:11:10 +08:00
|
|
|
void rds_tcp_xmit_path_complete(struct rds_conn_path *cp)
|
2009-08-21 20:28:31 +08:00
|
|
|
{
|
2016-07-01 07:11:10 +08:00
|
|
|
struct rds_tcp_connection *tc = cp->cp_transport_data;
|
2009-08-21 20:28:31 +08:00
|
|
|
|
|
|
|
rds_tcp_cork(tc->t_sock, 0);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* the core send_sem serializes this with other xmit and shutdown */
|
2010-10-19 16:08:33 +08:00
|
|
|
static int rds_tcp_sendmsg(struct socket *sock, void *data, unsigned int len)
|
2009-08-21 20:28:31 +08:00
|
|
|
{
|
|
|
|
struct kvec vec = {
|
2016-06-18 23:46:31 +08:00
|
|
|
.iov_base = data,
|
|
|
|
.iov_len = len,
|
|
|
|
};
|
|
|
|
struct msghdr msg = {
|
|
|
|
.msg_flags = MSG_DONTWAIT | MSG_NOSIGNAL,
|
2009-08-21 20:28:31 +08:00
|
|
|
};
|
|
|
|
|
|
|
|
return kernel_sendmsg(sock, &msg, &vec, 1, vec.iov_len);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* the core send_sem serializes this with other xmit and shutdown */
|
|
|
|
int rds_tcp_xmit(struct rds_connection *conn, struct rds_message *rm,
|
2016-06-18 23:46:31 +08:00
|
|
|
unsigned int hdr_off, unsigned int sg, unsigned int off)
|
2009-08-21 20:28:31 +08:00
|
|
|
{
|
2016-07-14 18:51:03 +08:00
|
|
|
struct rds_conn_path *cp = rm->m_inc.i_conn_path;
|
|
|
|
struct rds_tcp_connection *tc = cp->cp_transport_data;
|
2009-08-21 20:28:31 +08:00
|
|
|
int done = 0;
|
|
|
|
int ret = 0;
|
2015-10-01 04:54:09 +08:00
|
|
|
int more;
|
2009-08-21 20:28:31 +08:00
|
|
|
|
|
|
|
if (hdr_off == 0) {
|
|
|
|
/*
|
|
|
|
* m_ack_seq is set to the sequence number of the last byte of
|
|
|
|
* header and data. see rds_tcp_is_acked().
|
|
|
|
*/
|
|
|
|
tc->t_last_sent_nxt = rds_tcp_snd_nxt(tc);
|
|
|
|
rm->m_ack_seq = tc->t_last_sent_nxt +
|
|
|
|
sizeof(struct rds_header) +
|
|
|
|
be32_to_cpu(rm->m_inc.i_hdr.h_len) - 1;
|
2014-03-18 01:06:10 +08:00
|
|
|
smp_mb__before_atomic();
|
2009-08-21 20:28:31 +08:00
|
|
|
set_bit(RDS_MSG_HAS_ACK_SEQ, &rm->m_flags);
|
|
|
|
tc->t_last_expected_una = rm->m_ack_seq + 1;
|
|
|
|
|
2016-11-17 05:29:48 +08:00
|
|
|
if (test_bit(RDS_MSG_RETRANSMITTED, &rm->m_flags))
|
|
|
|
rm->m_inc.i_hdr.h_flags |= RDS_FLAG_RETRANSMITTED;
|
|
|
|
|
2009-08-21 20:28:31 +08:00
|
|
|
rdsdebug("rm %p tcp nxt %u ack_seq %llu\n",
|
|
|
|
rm, rds_tcp_snd_nxt(tc),
|
|
|
|
(unsigned long long)rm->m_ack_seq);
|
|
|
|
}
|
|
|
|
|
|
|
|
if (hdr_off < sizeof(struct rds_header)) {
|
|
|
|
/* see rds_tcp_write_space() */
|
|
|
|
set_bit(SOCK_NOSPACE, &tc->t_sock->sk->sk_socket->flags);
|
|
|
|
|
|
|
|
ret = rds_tcp_sendmsg(tc->t_sock,
|
|
|
|
(void *)&rm->m_inc.i_hdr + hdr_off,
|
|
|
|
sizeof(rm->m_inc.i_hdr) - hdr_off);
|
|
|
|
if (ret < 0)
|
|
|
|
goto out;
|
|
|
|
done += ret;
|
|
|
|
if (hdr_off + done != sizeof(struct rds_header))
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
2015-10-01 04:54:09 +08:00
|
|
|
more = rm->data.op_nents > 1 ? (MSG_MORE | MSG_SENDPAGE_NOTLAST) : 0;
|
2010-01-28 10:04:18 +08:00
|
|
|
while (sg < rm->data.op_nents) {
|
2015-10-01 04:54:09 +08:00
|
|
|
int flags = MSG_DONTWAIT | MSG_NOSIGNAL | more;
|
|
|
|
|
2009-08-21 20:28:31 +08:00
|
|
|
ret = tc->t_sock->ops->sendpage(tc->t_sock,
|
2010-01-28 10:04:18 +08:00
|
|
|
sg_page(&rm->data.op_sg[sg]),
|
|
|
|
rm->data.op_sg[sg].offset + off,
|
|
|
|
rm->data.op_sg[sg].length - off,
|
2015-10-01 04:54:09 +08:00
|
|
|
flags);
|
2010-01-28 10:04:18 +08:00
|
|
|
rdsdebug("tcp sendpage %p:%u:%u ret %d\n", (void *)sg_page(&rm->data.op_sg[sg]),
|
|
|
|
rm->data.op_sg[sg].offset + off, rm->data.op_sg[sg].length - off,
|
2009-08-21 20:28:31 +08:00
|
|
|
ret);
|
|
|
|
if (ret <= 0)
|
|
|
|
break;
|
|
|
|
|
|
|
|
off += ret;
|
|
|
|
done += ret;
|
2010-01-28 10:04:18 +08:00
|
|
|
if (off == rm->data.op_sg[sg].length) {
|
2009-08-21 20:28:31 +08:00
|
|
|
off = 0;
|
|
|
|
sg++;
|
|
|
|
}
|
2015-10-01 04:54:09 +08:00
|
|
|
if (sg == rm->data.op_nents - 1)
|
|
|
|
more = 0;
|
2009-08-21 20:28:31 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
out:
|
|
|
|
if (ret <= 0) {
|
|
|
|
/* write_space will hit after EAGAIN, all else fatal */
|
|
|
|
if (ret == -EAGAIN) {
|
|
|
|
rds_tcp_stats_inc(s_tcp_sndbuf_full);
|
|
|
|
ret = 0;
|
|
|
|
} else {
|
2016-07-14 18:51:03 +08:00
|
|
|
/* No need to disconnect/reconnect if path_drop
|
|
|
|
* has already been triggered, because, e.g., of
|
|
|
|
* an incoming RST.
|
|
|
|
*/
|
|
|
|
if (rds_conn_path_up(cp)) {
|
|
|
|
pr_warn("RDS/tcp: send to %pI4 on cp [%d]"
|
|
|
|
"returned %d, "
|
|
|
|
"disconnecting and reconnecting\n",
|
|
|
|
&conn->c_faddr, cp->cp_index, ret);
|
2017-07-17 07:43:46 +08:00
|
|
|
rds_conn_path_drop(cp, false);
|
2016-07-14 18:51:03 +08:00
|
|
|
}
|
2009-08-21 20:28:31 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
if (done == 0)
|
|
|
|
done = ret;
|
|
|
|
return done;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* rm->m_ack_seq is set to the tcp sequence number that corresponds to the
|
|
|
|
* last byte of the message, including the header. This means that the
|
|
|
|
* entire message has been received if rm->m_ack_seq is "before" the next
|
|
|
|
* unacked byte of the TCP sequence space. We have to do very careful
|
|
|
|
* wrapping 32bit comparisons here.
|
|
|
|
*/
|
|
|
|
static int rds_tcp_is_acked(struct rds_message *rm, uint64_t ack)
|
|
|
|
{
|
|
|
|
if (!test_bit(RDS_MSG_HAS_ACK_SEQ, &rm->m_flags))
|
|
|
|
return 0;
|
|
|
|
return (__s32)((u32)rm->m_ack_seq - (u32)ack) < 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
void rds_tcp_write_space(struct sock *sk)
|
|
|
|
{
|
|
|
|
void (*write_space)(struct sock *sk);
|
2016-07-01 07:11:14 +08:00
|
|
|
struct rds_conn_path *cp;
|
2009-08-21 20:28:31 +08:00
|
|
|
struct rds_tcp_connection *tc;
|
|
|
|
|
2016-05-18 08:44:08 +08:00
|
|
|
read_lock_bh(&sk->sk_callback_lock);
|
2016-07-01 07:11:14 +08:00
|
|
|
cp = sk->sk_user_data;
|
|
|
|
if (!cp) {
|
2009-08-21 20:28:31 +08:00
|
|
|
write_space = sk->sk_write_space;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
2016-07-01 07:11:14 +08:00
|
|
|
tc = cp->cp_transport_data;
|
2009-08-21 20:28:31 +08:00
|
|
|
rdsdebug("write_space for tc %p\n", tc);
|
|
|
|
write_space = tc->t_orig_write_space;
|
|
|
|
rds_tcp_stats_inc(s_tcp_write_space_calls);
|
|
|
|
|
|
|
|
rdsdebug("tcp una %u\n", rds_tcp_snd_una(tc));
|
|
|
|
tc->t_last_seen_una = rds_tcp_snd_una(tc);
|
2016-07-01 07:11:14 +08:00
|
|
|
rds_send_path_drop_acked(cp, rds_tcp_snd_una(tc), rds_tcp_is_acked);
|
2009-08-21 20:28:31 +08:00
|
|
|
|
2017-06-30 18:08:00 +08:00
|
|
|
if ((refcount_read(&sk->sk_wmem_alloc) << 1) <= sk->sk_sndbuf)
|
2016-07-01 07:11:14 +08:00
|
|
|
queue_delayed_work(rds_wq, &cp->cp_send_w, 0);
|
2010-03-11 21:49:58 +08:00
|
|
|
|
2009-08-21 20:28:31 +08:00
|
|
|
out:
|
2016-05-18 08:44:08 +08:00
|
|
|
read_unlock_bh(&sk->sk_callback_lock);
|
2009-08-21 20:28:31 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* write_space is only called when data leaves tcp's send queue if
|
|
|
|
* SOCK_NOSPACE is set. We set SOCK_NOSPACE every time we put
|
|
|
|
* data in tcp's send queue because we use write_space to parse the
|
|
|
|
* sequence numbers and notice that rds messages have been fully
|
|
|
|
* received.
|
|
|
|
*
|
|
|
|
* tcp's write_space clears SOCK_NOSPACE if the send queue has more
|
|
|
|
* than a certain amount of space. So we need to set it again *after*
|
|
|
|
* we call tcp's write_space or else we might only get called on the
|
|
|
|
* first of a series of incoming tcp acks.
|
|
|
|
*/
|
|
|
|
write_space(sk);
|
|
|
|
|
|
|
|
if (sk->sk_socket)
|
|
|
|
set_bit(SOCK_NOSPACE, &sk->sk_socket->flags);
|
|
|
|
}
|