From: Peter Zijlstra <peterz@infradead.org>
Date: Wed, 1 Oct 2025 13:50:15 +0000 (+0200)
Subject: sched: Detect per-class runqueue changes
X-Git-Url: http://git.ipfire.org/gitweb.cgi?a=commitdiff_plain;h=1e900f415c6082cd4bcdae4c92515d21fb389473;p=thirdparty%2Flinux.git

sched: Detect per-class runqueue changes

Have enqueue/dequeue set a per-class bit in rq->queue_mask. This then
enables easy tracking of which runqueues are modified over a
lock-break.

Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Tejun Heo <tj@kernel.org>
---

diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index e2199e4db0dc4..9fc990ff68454 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -2089,6 +2089,7 @@ void enqueue_task(struct rq *rq, struct task_struct *p, int flags)
 	 */
 	uclamp_rq_inc(rq, p, flags);
 
+	rq->queue_mask |= p->sched_class->queue_mask;
 	p->sched_class->enqueue_task(rq, p, flags);
 
 	psi_enqueue(p, flags);
@@ -2121,6 +2122,7 @@ inline bool dequeue_task(struct rq *rq, struct task_struct *p, int flags)
 	 * and mark the task ->sched_delayed.
 	 */
 	uclamp_rq_dec(rq, p);
+	rq->queue_mask |= p->sched_class->queue_mask;
 	return p->sched_class->dequeue_task(rq, p, flags);
 }
 
diff --git a/kernel/sched/deadline.c b/kernel/sched/deadline.c
index 1f94994984038..83e6175d79f5f 100644
--- a/kernel/sched/deadline.c
+++ b/kernel/sched/deadline.c
@@ -3092,6 +3092,8 @@ static int task_is_throttled_dl(struct task_struct *p, int cpu)
 
 DEFINE_SCHED_CLASS(dl) = {
 
+	.queue_mask		= 8,
+
 	.enqueue_task		= enqueue_task_dl,
 	.dequeue_task		= dequeue_task_dl,
 	.yield_task		= yield_task_dl,
diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c
index 57170423c3d97..949c3a6e24d4b 100644
--- a/kernel/sched/ext.c
+++ b/kernel/sched/ext.c
@@ -3234,6 +3234,8 @@ static void scx_cgroup_unlock(void) {}
  *   their current sched_class. Call them directly from sched core instead.
  */
 DEFINE_SCHED_CLASS(ext) = {
+	.queue_mask		= 1,
+
 	.enqueue_task		= enqueue_task_scx,
 	.dequeue_task		= dequeue_task_scx,
 	.yield_task		= yield_task_scx,
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 77a713ecde9d1..23ac05cca4a49 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -12841,6 +12841,7 @@ static int sched_balance_newidle(struct rq *this_rq, struct rq_flags *rf)
 	}
 	rcu_read_unlock();
 
+	rq_modified_clear(this_rq);
 	raw_spin_rq_unlock(this_rq);
 
 	t0 = sched_clock_cpu(this_cpu);
@@ -12898,8 +12899,8 @@ static int sched_balance_newidle(struct rq *this_rq, struct rq_flags *rf)
 	if (this_rq->cfs.h_nr_queued && !pulled_task)
 		pulled_task = 1;
 
-	/* Is there a task of a high priority class? */
-	if (this_rq->nr_running != this_rq->cfs.h_nr_queued)
+	/* If a higher prio class was modified, restart the pick */
+	if (rq_modified_above(this_rq, &fair_sched_class))
 		pulled_task = -1;
 
 out:
@@ -13633,6 +13634,8 @@ static unsigned int get_rr_interval_fair(struct rq *rq, struct task_struct *task
  */
 DEFINE_SCHED_CLASS(fair) = {
 
+	.queue_mask		= 2,
+
 	.enqueue_task		= enqueue_task_fair,
 	.dequeue_task		= dequeue_task_fair,
 	.yield_task		= yield_task_fair,
diff --git a/kernel/sched/idle.c b/kernel/sched/idle.c
index dee6e019dcf81..055b0ddbcd54d 100644
--- a/kernel/sched/idle.c
+++ b/kernel/sched/idle.c
@@ -521,6 +521,8 @@ static void update_curr_idle(struct rq *rq)
  */
 DEFINE_SCHED_CLASS(idle) = {
 
+	.queue_mask		= 0,
+
 	/* no enqueue/yield_task for idle tasks */
 
 	/* dequeue is not valid, we print a debug message there: */
diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c
index c2347e485dc65..9bc828d59121a 100644
--- a/kernel/sched/rt.c
+++ b/kernel/sched/rt.c
@@ -2569,6 +2569,8 @@ static int task_is_throttled_rt(struct task_struct *p, int cpu)
 
 DEFINE_SCHED_CLASS(rt) = {
 
+	.queue_mask		= 4,
+
 	.enqueue_task		= enqueue_task_rt,
 	.dequeue_task		= dequeue_task_rt,
 	.yield_task		= yield_task_rt,
diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index e3d271013c8b0..f4a323007dced 100644
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -1118,6 +1118,8 @@ struct rq {
 	/* runqueue lock: */
 	raw_spinlock_t		__lock;
 
+	/* Per class runqueue modification mask; bits in class order. */
+	unsigned int		queue_mask;
 	unsigned int		nr_running;
 #ifdef CONFIG_NUMA_BALANCING
 	unsigned int		nr_numa_running;
@@ -2414,6 +2416,15 @@ struct sched_class {
 #ifdef CONFIG_UCLAMP_TASK
 	int uclamp_enabled;
 #endif
+	/*
+	 * idle:  0
+	 * ext:   1
+	 * fair:  2
+	 * rt:    4
+	 * dl:    8
+	 * stop: 16
+	 */
+	unsigned int queue_mask;
 
 	/*
 	 * move_queued_task/activate_task/enqueue_task: rq->lock
@@ -2571,6 +2582,20 @@ struct sched_class {
 #endif
 };
 
+/*
+ * Does not nest; only used around sched_class::pick_task() rq-lock-breaks.
+ */
+static inline void rq_modified_clear(struct rq *rq)
+{
+	rq->queue_mask = 0;
+}
+
+static inline bool rq_modified_above(struct rq *rq, const struct sched_class * class)
+{
+	unsigned int mask = class->queue_mask;
+	return rq->queue_mask & ~((mask << 1) - 1);
+}
+
 static inline void put_prev_task(struct rq *rq, struct task_struct *prev)
 {
 	WARN_ON_ONCE(rq->donor != prev);
diff --git a/kernel/sched/stop_task.c b/kernel/sched/stop_task.c
index 73aa8de190675..d98c453c9b4eb 100644
--- a/kernel/sched/stop_task.c
+++ b/kernel/sched/stop_task.c
@@ -98,6 +98,8 @@ static void update_curr_stop(struct rq *rq)
  */
 DEFINE_SCHED_CLASS(stop) = {
 
+	.queue_mask		= 16,
+
 	.enqueue_task		= enqueue_task_stop,
 	.dequeue_task		= dequeue_task_stop,
 	.yield_task		= yield_task_stop,