1 files changed, 57 insertions, 46 deletions

diff --git a/net/rds/tcp_send.c b/net/rds/tcp_send.c
index 81cf5a4c5e40..7d284ac7e81a 100644
--- a/net/rds/tcp_send.c
+++ b/net/rds/tcp_send.c
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2006 Oracle.  All rights reserved.
+ * Copyright (c) 2006, 2017 Oracle and/or its affiliates. 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
@@ -34,53 +34,46 @@
 #include <linux/in.h>
 #include <net/tcp.h>
 
+#include "rds_single_path.h"
 #include "rds.h"
 #include "tcp.h"
 
-static void rds_tcp_cork(struct socket *sock, int val)
+void rds_tcp_xmit_path_prepare(struct rds_conn_path *cp)
 {
-	mm_segment_t oldfs;
+	struct rds_tcp_connection *tc = cp->cp_transport_data;
 
-	oldfs = get_fs();
-	set_fs(KERNEL_DS);
-	sock->ops->setsockopt(sock, SOL_TCP, TCP_CORK, (char __user *)&val,
-			      sizeof(val));
-	set_fs(oldfs);
+	tcp_sock_set_cork(tc->t_sock->sk, true);
 }
 
-void rds_tcp_xmit_prepare(struct rds_connection *conn)
+void rds_tcp_xmit_path_complete(struct rds_conn_path *cp)
 {
-	struct rds_tcp_connection *tc = conn->c_transport_data;
+	struct rds_tcp_connection *tc = cp->cp_transport_data;
 
-	rds_tcp_cork(tc->t_sock, 1);
-}
-
-void rds_tcp_xmit_complete(struct rds_connection *conn)
-{
-	struct rds_tcp_connection *tc = conn->c_transport_data;
-
-	rds_tcp_cork(tc->t_sock, 0);
+	tcp_sock_set_cork(tc->t_sock->sk, false);
 }
 
 /* the core send_sem serializes this with other xmit and shutdown */
 static int rds_tcp_sendmsg(struct socket *sock, void *data, unsigned int len)
 {
 	struct kvec vec = {
-                .iov_base = data,
-                .iov_len = len,
+		.iov_base = data,
+		.iov_len = len,
+	};
+	struct msghdr msg = {
+		.msg_flags = MSG_DONTWAIT | MSG_NOSIGNAL,
 	};
-        struct msghdr msg = {
-                .msg_flags = MSG_DONTWAIT | MSG_NOSIGNAL,
-        };
 
 	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,
-	         unsigned int hdr_off, unsigned int sg, unsigned int off)
+		 unsigned int hdr_off, unsigned int sg, unsigned int off)
 {
-	struct rds_tcp_connection *tc = conn->c_transport_data;
+	struct rds_conn_path *cp = rm->m_inc.i_conn_path;
+	struct rds_tcp_connection *tc = cp->cp_transport_data;
+	struct msghdr msg = {};
+	struct bio_vec bvec;
 	int done = 0;
 	int ret = 0;
 
@@ -89,16 +82,19 @@ int rds_tcp_xmit(struct rds_connection *conn, struct rds_message *rm,
 		 * 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);
+		tc->t_last_sent_nxt = rds_tcp_write_seq(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;
-		smp_mb__before_clear_bit();
+		smp_mb__before_atomic();
 		set_bit(RDS_MSG_HAS_ACK_SEQ, &rm->m_flags);
 		tc->t_last_expected_una = rm->m_ack_seq + 1;
 
+		if (test_bit(RDS_MSG_RETRANSMITTED, &rm->m_flags))
+			rm->m_inc.i_hdr.h_flags |= RDS_FLAG_RETRANSMITTED;
+
 		rdsdebug("rm %p tcp nxt %u ack_seq %llu\n",
-			 rm, rds_tcp_snd_nxt(tc),
+			 rm, rds_tcp_write_seq(tc),
 			 (unsigned long long)rm->m_ack_seq);
 	}
 
@@ -117,11 +113,16 @@ int rds_tcp_xmit(struct rds_connection *conn, struct rds_message *rm,
 	}
 
 	while (sg < rm->data.op_nents) {
-		ret = tc->t_sock->ops->sendpage(tc->t_sock,
-						sg_page(&rm->data.op_sg[sg]),
-						rm->data.op_sg[sg].offset + off,
-						rm->data.op_sg[sg].length - off,
-						MSG_DONTWAIT|MSG_NOSIGNAL);
+		msg.msg_flags = MSG_SPLICE_PAGES | MSG_DONTWAIT | MSG_NOSIGNAL;
+		if (sg + 1 < rm->data.op_nents)
+			msg.msg_flags |= MSG_MORE;
+
+		bvec_set_page(&bvec, sg_page(&rm->data.op_sg[sg]),
+			      rm->data.op_sg[sg].length - off,
+			      rm->data.op_sg[sg].offset + off);
+		iov_iter_bvec(&msg.msg_iter, ITER_SOURCE, &bvec, 1,
+			      rm->data.op_sg[sg].length - off);
+		ret = sock_sendmsg(tc->t_sock, &msg);
 		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,
 			 ret);
@@ -143,10 +144,17 @@ out:
 			rds_tcp_stats_inc(s_tcp_sndbuf_full);
 			ret = 0;
 		} else {
-			printk(KERN_WARNING "RDS/tcp: send to %pI4 "
-			       "returned %d, disconnecting and reconnecting\n",
-			       &conn->c_faddr, ret);
-			rds_conn_drop(conn);
+			/* 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 %pI6c on cp [%d]"
+					"returned %d, "
+					"disconnecting and reconnecting\n",
+					&conn->c_faddr, cp->cp_index, ret);
+				rds_conn_path_drop(cp, false);
+			}
 		}
 	}
 	if (done == 0)
@@ -171,30 +179,33 @@ static int rds_tcp_is_acked(struct rds_message *rm, uint64_t ack)
 void rds_tcp_write_space(struct sock *sk)
 {
 	void (*write_space)(struct sock *sk);
-	struct rds_connection *conn;
+	struct rds_conn_path *cp;
 	struct rds_tcp_connection *tc;
 
-	read_lock(&sk->sk_callback_lock);
-	conn = sk->sk_user_data;
-	if (!conn) {
+	read_lock_bh(&sk->sk_callback_lock);
+	cp = sk->sk_user_data;
+	if (!cp) {
 		write_space = sk->sk_write_space;
 		goto out;
 	}
 
-	tc = conn->c_transport_data;
+	tc = cp->cp_transport_data;
 	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);
-	rds_send_drop_acked(conn, rds_tcp_snd_una(tc), rds_tcp_is_acked);
+	rds_send_path_drop_acked(cp, rds_tcp_snd_una(tc), rds_tcp_is_acked);
 
-        if ((atomic_read(&sk->sk_wmem_alloc) << 1) <= sk->sk_sndbuf)
-		queue_delayed_work(rds_wq, &conn->c_send_w, 0);
+	rcu_read_lock();
+	if ((refcount_read(&sk->sk_wmem_alloc) << 1) <= sk->sk_sndbuf &&
+	    !rds_destroy_pending(cp->cp_conn))
+		queue_delayed_work(rds_wq, &cp->cp_send_w, 0);
+	rcu_read_unlock();
 
 out:
-	read_unlock(&sk->sk_callback_lock);
+	read_unlock_bh(&sk->sk_callback_lock);
 
 	/*
 	 * write_space is only called when data leaves tcp's send queue if