rds: per-rm flush_wait waitq

This removes a global waitqueue used to wait for rds messages
and replaces it with a waitqueue inside the rds_message struct.

The global waitqueue turns into a global lock and significantly
bottlenecks operations on large machines.

Signed-off-by: Chris Mason <chris.mason@oracle.com>
diff --git a/net/rds/message.c b/net/rds/message.c
index 4cb1ed7..96e2bf7 100644
--- a/net/rds/message.c
+++ b/net/rds/message.c
@@ -35,8 +35,6 @@
 
 #include "rds.h"
 
-static DECLARE_WAIT_QUEUE_HEAD(rds_message_flush_waitq);
-
 static unsigned int	rds_exthdr_size[__RDS_EXTHDR_MAX] = {
 [RDS_EXTHDR_NONE]	= 0,
 [RDS_EXTHDR_VERSION]	= sizeof(struct rds_ext_header_version),
@@ -226,6 +224,7 @@
 	INIT_LIST_HEAD(&rm->m_sock_item);
 	INIT_LIST_HEAD(&rm->m_conn_item);
 	spin_lock_init(&rm->m_rs_lock);
+	init_waitqueue_head(&rm->m_flush_wait);
 
 out:
 	return rm;
@@ -399,14 +398,14 @@
  */
 void rds_message_wait(struct rds_message *rm)
 {
-	wait_event_interruptible(rds_message_flush_waitq,
+	wait_event_interruptible(rm->m_flush_wait,
 			!test_bit(RDS_MSG_MAPPED, &rm->m_flags));
 }
 
 void rds_message_unmapped(struct rds_message *rm)
 {
 	clear_bit(RDS_MSG_MAPPED, &rm->m_flags);
-	wake_up_interruptible(&rds_message_flush_waitq);
+	wake_up_interruptible(&rm->m_flush_wait);
 }
 EXPORT_SYMBOL_GPL(rds_message_unmapped);
 
diff --git a/net/rds/rds.h b/net/rds/rds.h
index b57cb50..c22bd7b 100644
--- a/net/rds/rds.h
+++ b/net/rds/rds.h
@@ -301,6 +301,8 @@
 	 *   -> rs->rs_lock
 	 */
 	spinlock_t		m_rs_lock;
+	wait_queue_head_t	m_flush_wait;
+
 	struct rds_sock		*m_rs;
 
 	/* cookie to send to remote, in rds header */