[PATCH 3.12 042/116] sched: Allow balance callbacks for check_class_changed()
From: Jiri Slaby
Date: Fri Mar 04 2016 - 04:39:56 EST
From: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
commit 4c9a4bc89a9cca8128bce67d6bc8870d6b7ee0b2 upstream.
In order to remove dropping rq->lock from the
switched_{to,from}()/prio_changed() sched_class methods, run the
balance callbacks after it.
We need to remove dropping rq->lock because its buggy,
suppose using sched_setattr()/sched_setscheduler() to change a running
task from FIFO to OTHER.
By the time we get to switched_from_rt() the task is already enqueued
on the cfs runqueues. If switched_from_rt() does pull_rt_task() and
drops rq->lock, load-balancing can come in and move our task @p to
another rq.
The subsequent switched_to_fair() still assumes @p is on @rq and bad
things will happen.
By using balance callbacks we delay the load-balancing operations
{rt,dl}x{push,pull} until we've done all the important work and the
task is fully set up.
Furthermore, the balance callbacks do not know about @p, therefore
they cannot get confused like this.
Reported-by: Mike Galbraith <umgwanakikbuti@xxxxxxxxx>
Signed-off-by: Peter Zijlstra (Intel) <peterz@xxxxxxxxxxxxx>
Cc: ktkhai@xxxxxxxxxxxxx
Cc: rostedt@xxxxxxxxxxx
Cc: juri.lelli@xxxxxxxxx
Cc: pang.xunlei@xxxxxxxxxx
Cc: oleg@xxxxxxxxxx
Cc: wanpeng.li@xxxxxxxxxxxxxxx
Link: http://lkml.kernel.org/r/20150611124742.615343911@xxxxxxxxxxxxx
Signed-off-by: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
Signed-off-by: Byungchul Park <byungchul.park@xxxxxxx>
Signed-off-by: Jiri Slaby <jslaby@xxxxxxx>
---
kernel/sched/core.c | 24 +++++++++++++++++++++++-
1 file changed, 23 insertions(+), 1 deletion(-)
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 7bf52708993e..9b2394fffc2f 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -941,6 +941,13 @@ inline int task_curr(const struct task_struct *p)
return cpu_curr(task_cpu(p)) == p;
}
+/*
+ * switched_from, switched_to and prio_changed must _NOT_ drop rq->lock,
+ * use the balance_callback list if you want balancing.
+ *
+ * this means any call to check_class_changed() must be followed by a call to
+ * balance_callback().
+ */
static inline void check_class_changed(struct rq *rq, struct task_struct *p,
const struct sched_class *prev_class,
int oldprio)
@@ -1325,8 +1332,12 @@ ttwu_do_wakeup(struct rq *rq, struct task_struct *p, int wake_flags)
p->state = TASK_RUNNING;
#ifdef CONFIG_SMP
- if (p->sched_class->task_woken)
+ if (p->sched_class->task_woken) {
+ /*
+ * XXX can drop rq->lock; most likely ok.
+ */
p->sched_class->task_woken(rq, p);
+ }
if (rq->idle_stamp) {
u64 delta = rq_clock(rq) - rq->idle_stamp;
@@ -3087,7 +3098,11 @@ void rt_mutex_setprio(struct task_struct *p, int prio)
check_class_changed(rq, p, prev_class, oldprio);
out_unlock:
+ preempt_disable(); /* avoid rq from going away on us */
__task_rq_unlock(rq);
+
+ balance_callback(rq);
+ preempt_enable();
}
#endif
void set_user_nice(struct task_struct *p, long nice)
@@ -3442,10 +3457,17 @@ recheck:
enqueue_task(rq, p, 0);
check_class_changed(rq, p, prev_class, oldprio);
+ preempt_disable(); /* avoid rq from going away on us */
task_rq_unlock(rq, p, &flags);
rt_mutex_adjust_pi(p);
+ /*
+ * Run balance callbacks after we've adjusted the PI chain.
+ */
+ balance_callback(rq);
+ preempt_enable();
+
return 0;
}
--
2.7.2