[PATCH v2] sched/core: Reduce cost of sched_move_task when config autogroup

[wuchi]

Some sched_move_task calls are useless because that
task_struct->sched_task_group maybe not changed (equals task_group
of cpu_cgroup) when system enable autogroup. So do some checks in
sched_move_task.

sched_move_task eg:
task A belongs to cpu_cgroup0 and autogroup0, it will always belong
to cpu_cgroup0 when do_exit. So there is no need to do {de|en}queue.
The call graph is as follow.

do_exit
  sched_autogroup_exit_task
    sched_move_task
      dequeue_task
        sched_change_group
	  A.sched_task_group = sched_get_task_group (=cpu_cgroup0)
      enqueue_task

Performance results:
===========================
1. env
        cpu: bogomips=4600.00
     kernel: 6.3.0-rc3
 cpu_cgroup: 6:cpu,cpuacct:/user.slice

2. cmds
do_exit script:
```
for i in {0..10000}; do
    sleep 0 &
    done
wait
```
Run the above script, then use the following bpftrace cmd to get
the cost of sched_move_task:

bpftrace -e 'k:sched_move_task { @ts[tid] = nsecs; }
             kr:sched_move_task /@ts[tid]/
                { @ns += nsecs - @ts[tid]; delete(@ts[tid]); }'

3. cost time(ns):
 without patch: 43528033
 with    patch: 18541416
          diff:-24986617  -57.4%

As the result show, the patch will save 57.4% in the scenario.

Signed-off-by: wuchi <wuchi.zero@xxxxxxxxx>
Suggested-by: Peter Zijlstra (Intel) <peterz@xxxxxxxxxxxxx>
---
 kernel/sched/core.c | 30 ++++++++++++++++++++++++++++--
 1 file changed, 28 insertions(+), 2 deletions(-)

diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index a380f34789a2..1e7d6a8c3455 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -10330,7 +10330,7 @@ void sched_release_group(struct task_group *tg)
 	spin_unlock_irqrestore(&task_group_lock, flags);
 }
 
-static void sched_change_group(struct task_struct *tsk)
+static struct task_group *sched_get_task_group(struct task_struct *tsk)
 {
 	struct task_group *tg;
 
@@ -10342,7 +10342,28 @@ static void sched_change_group(struct task_struct *tsk)
 	tg = container_of(task_css_check(tsk, cpu_cgrp_id, true),
 			  struct task_group, css);
 	tg = autogroup_task_group(tsk, tg);
-	tsk->sched_task_group = tg;
+
+	return tg;
+}
+
+static bool sched_task_group_changed(struct task_struct *tsk)
+{
+	/*
+	 * Some sched_move_task calls are useless because that
+	 * task_struct->sched_task_group maybe not changed (equals
+	 * task_group of cpu_cgroup) when system enable autogroup.
+	 * So do some checks in sched_move_task.
+	 */
+#ifdef CONFIG_SCHED_AUTOGROUP
+	return sched_get_task_group(tsk) != tsk->sched_task_group;
+#else
+	return true;
+#endif /* CONFIG_SCHED_AUTOGROUP */
+}
+
+static void sched_change_group(struct task_struct *tsk)
+{
+	tsk->sched_task_group = sched_get_task_group(tsk);
 
 #ifdef CONFIG_FAIR_GROUP_SCHED
 	if (tsk->sched_class->task_change_group)
@@ -10367,6 +10388,10 @@ void sched_move_task(struct task_struct *tsk)
 	struct rq *rq;
 
 	rq = task_rq_lock(tsk, &rf);
+
+	if (!sched_task_group_changed(tsk))
+		goto unlock;
+
 	update_rq_clock(rq);
 
 	running = task_current(rq, tsk);
@@ -10391,6 +10416,7 @@ void sched_move_task(struct task_struct *tsk)
 		resched_curr(rq);
 	}
 
+unlock:
 	task_rq_unlock(rq, tsk, &rf);
 }
 
-- 
2.20.1