Re: [PATCH v4 6/6] sched/fair: Remove double_lock_balance() from load_balance()

[Peter Zijlstra <peterz@infradead.org>]

[-- Attachment #1: Type: text/plain, Size: 8532 bytes --]



Changed quite a bit around, _should_ be more or less the same end result
I suppose. Only compile tested so far.

---
Subject: sched/fair: Remove double_lock_balance() from load_balance()
From: Kirill Tkhai <ktkhai@parallels.com>
Date: Wed, 6 Aug 2014 12:07:04 +0400

Avoid double_rq_lock() and use ONRQ_MIGRATING for load_balance(). The
advantage is (obviously) not holding two 'rq->lock's at the same time
and thereby increasing parallelism.

Further note that if there was no task to migrate we will not have
acquired the second rq->lock at all.

The important point to note is that because we acquire dst->lock
immediately after releasing src->lock the potential wait time of
task_rq_lock() callers on ONRQ_MIGRATING is not longer than it would
have been in the double rq lock scenario.

Signed-off-by: Kirill Tkhai <ktkhai@parallels.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1407312424.8424.48.camel@tkhai
---
 kernel/sched/fair.c |  151 ++++++++++++++++++++++++++++++++++------------------
 1 file changed, 99 insertions(+), 52 deletions(-)

--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -4706,7 +4706,7 @@ static void check_preempt_wakeup(struct
 		return;
 
 	/*
-	 * This is possible from callers such as move_task(), in which we
+	 * This is possible from callers such as attach_tasks(), in which we
 	 * unconditionally check_prempt_curr() after an enqueue (which may have
 	 * lead to a throttle).  This both saves work and prevents false
 	 * next-buddy nomination below.
@@ -5114,21 +5114,10 @@ struct lb_env {
 	unsigned int		loop_max;
 
 	enum fbq_type		fbq_type;
+	struct list_head	tasks;
 };
 
 /*
- * move_task - move a task from one runqueue to another runqueue.
- * Both runqueues must be locked.
- */
-static void move_task(struct task_struct *p, struct lb_env *env)
-{
-	deactivate_task(env->src_rq, p, 0);
-	set_task_cpu(p, env->dst_cpu);
-	activate_task(env->dst_rq, p, 0);
-	check_preempt_curr(env->dst_rq, p, 0);
-}
-
-/*
  * Is this task likely cache-hot:
  */
 static int task_hot(struct task_struct *p, struct lb_env *env)
@@ -5343,6 +5332,18 @@ int can_migrate_task(struct task_struct
 }
 
 /*
+ * detach_task() -- detach the task for the migration specified in env
+ */
+static void detach_task(struct task_struct *p, struct lb_env *env)
+{
+	lockdep_assert_held(&env->src_rq->lock);
+
+	deactivate_task(env->src_rq, p, 0);
+	p->on_rq = ONRQ_MIGRATING;
+	set_task_cpu(p, env->dst_cpu);
+}
+
+/*
  * detach_one_task() -- tries to dequeue exactly one task from env->src_rq, as
  * part of active balancing operations within "domain".
  *
@@ -5358,15 +5359,13 @@ static struct task_struct *detach_one_ta
 		if (!can_migrate_task(p, env))
 			continue;
 
-		deactivate_task(env->src_rq, p, 0);
-		p->on_rq = ONRQ_MIGRATING;
-		set_task_cpu(p, env->dst_cpu);
+		detach_task(p, env);
 
 		/*
 		 * Right now, this is only the second place where
-		 * lb_gained[env->idle] is updated (other is move_tasks)
+		 * lb_gained[env->idle] is updated (other is detach_tasks)
 		 * so we can safely collect stats here rather than
-		 * inside move_tasks().
+		 * inside detach_tasks().
 		 */
 		schedstat_inc(env->sd, lb_gained[env->idle]);
 		return p;
@@ -5374,35 +5373,22 @@ static struct task_struct *detach_one_ta
 	return NULL;
 }
 
-/*
- * attach_one_task() -- attaches the task returned from detach_one_task() to
- * its new rq.
- */
-static void attach_one_task(struct rq *rq, struct task_struct *p)
-{
-	raw_spin_lock(&rq->lock);
-	BUG_ON(task_rq(p) != rq);
-	p->on_rq = ONRQ_QUEUED;
-	activate_task(rq, p, 0);
-	check_preempt_curr(rq, p, 0);
-	raw_spin_unlock(&rq->lock);
-}
-
 static const unsigned int sched_nr_migrate_break = 32;
 
 /*
- * move_tasks tries to move up to imbalance weighted load from busiest to
- * this_rq, as part of a balancing operation within domain "sd".
- * Returns 1 if successful and 0 otherwise.
+ * detach_tasks() -- tries to detach up to imbalance weighted load from
+ * busiest_rq, as part of a balancing operation within domain "sd".
  *
- * Called with both runqueues locked.
+ * Returns number of detached tasks if successful and 0 otherwise.
  */
-static int move_tasks(struct lb_env *env)
+static int detach_tasks(struct lb_env *env)
 {
 	struct list_head *tasks = &env->src_rq->cfs_tasks;
 	struct task_struct *p;
 	unsigned long load;
-	int pulled = 0;
+	int detached = 0;
+
+	lockdep_assert_held(&env->src_rq->lock);
 
 	if (env->imbalance <= 0)
 		return 0;
@@ -5433,14 +5419,16 @@ static int move_tasks(struct lb_env *env
 		if ((load / 2) > env->imbalance)
 			goto next;
 
-		move_task(p, env);
-		pulled++;
+		detach_task(p, env);
+		list_add(&p->se.group_node, &env->tasks);
+
+		detached++;
 		env->imbalance -= load;
 
 #ifdef CONFIG_PREEMPT
 		/*
 		 * NEWIDLE balancing is a source of latency, so preemptible
-		 * kernels will stop after the first task is pulled to minimize
+		 * kernels will stop after the first task is detached to minimize
 		 * the critical section.
 		 */
 		if (env->idle == CPU_NEWLY_IDLE)
@@ -5460,13 +5448,58 @@ static int move_tasks(struct lb_env *env
 	}
 
 	/*
-	 * Right now, this is one of only two places move_task() is called,
-	 * so we can safely collect move_task() stats here rather than
-	 * inside move_task().
+	 * Right now, this is one of only two places we collect this stat
+	 * so we can safely collect detach_one_task() stats here rather
+	 * than inside detach_one_task().
 	 */
-	schedstat_add(env->sd, lb_gained[env->idle], pulled);
+	schedstat_add(env->sd, lb_gained[env->idle], detached);
 
-	return pulled;
+	return detached;
+}
+
+/*
+ * attach_task() -- attach the task detached by detach_task() to its new rq.
+ */
+static void attach_task(struct rq *rq, struct task_struct *p)
+{
+	lockdep_assert_held(&rq->lock);
+
+	BUG_ON(task_rq(p) != rq);
+	p->on_rq = ONRQ_QUEUED;
+	activate_task(rq, p, 0);
+	check_preempt_curr(rq, p, 0);
+}
+
+/*
+ * attach_one_task() -- attaches the task returned from detach_one_task() to
+ * its new rq.
+ */
+static void attach_one_task(struct rq *rq, struct task_struct *p)
+{
+	raw_spin_lock(&rq->lock);
+	attach_task(rq, p);
+	raw_spin_unlock(&rq->lock);
+}
+
+/*
+ * attach_tasks() -- attaches all tasks detached by detach_tasks() to their
+ * new rq.
+ */
+static void attach_tasks(struct lb_env *env)
+{
+	struct list_head *tasks = &env->tasks;
+	struct task_struct *p;
+
+	raw_spin_lock(&env->dst_rq->lock);
+
+	while (!list_empty(tasks)) {
+		p = list_first_entry(tasks, struct task_struct, se.group_node);
+		list_del_init(&p->se.group_node);
+
+		attach_task(env->dst_rq, p);
+	}
+
+	raw_spin_unlock(&env->dst_rq->lock);
 }
 
 #ifdef CONFIG_FAIR_GROUP_SCHED
@@ -6600,6 +6633,7 @@ static int load_balance(int this_cpu, st
 		.loop_break	= sched_nr_migrate_break,
 		.cpus		= cpus,
 		.fbq_type	= all,
+		.tasks		= LIST_HEAD_INIT(env.tasks),
 	};
 
 	/*
@@ -6649,16 +6683,29 @@ static int load_balance(int this_cpu, st
 		env.loop_max  = min(sysctl_sched_nr_migrate, busiest->nr_running);
 
 more_balance:
-		local_irq_save(flags);
-		double_rq_lock(env.dst_rq, busiest);
+		raw_spin_lock_irqsave(&busiest->lock, flags);
 
 		/*
 		 * cur_ld_moved - load moved in current iteration
 		 * ld_moved     - cumulative load moved across iterations
 		 */
-		cur_ld_moved = move_tasks(&env);
-		ld_moved += cur_ld_moved;
-		double_rq_unlock(env.dst_rq, busiest);
+		cur_ld_moved = detach_tasks(&env);
+
+		/*
+		 * We've detached some tasks from busiest_rq. Every
+		 * task is masked "ONRQ_MIGRATED", so we can safely
+		 * unlock busiest->lock, and we are able to be sure
+		 * that nobody can manipulate the tasks in parallel.
+		 * See task_rq_lock() family for the details.
+		 */
+
+		raw_spin_unlock(&busiest->lock);
+
+		if (cur_ld_moved) {
+			attach_tasks(&env);
+			ld_moved += cur_ld_moved;
+		}
+
 		local_irq_restore(flags);
 
 		/*
@@ -6794,7 +6841,7 @@ static int load_balance(int this_cpu, st
 		 * If we've begun active balancing, start to back off. This
 		 * case may not be covered by the all_pinned logic if there
 		 * is only 1 task on the busy runqueue (because we don't call
-		 * move_tasks).
+		 * detach_tasks).
 		 */
 		if (sd->balance_interval < sd->max_interval)
 			sd->balance_interval *= 2;

[-- Attachment #2: Type: application/pgp-signature, Size: 836 bytes --]