[tip: sched/core] sched: Cleanup the sched_change NOCLOCK usage

["tip-bot2 for Peter Zijlstra" <tip-bot2@linutronix.de>]

The following commit has been merged into the sched/core branch of tip:

Commit-ID:     d4c64207b88a60dd15a38c790bb73c0b6f9a8c40
Gitweb:        https://git.kernel.org/tip/d4c64207b88a60dd15a38c790bb73c0b6f9a8c40
Author:        Peter Zijlstra <peterz@infradead.org>
AuthorDate:    Thu, 11 Sep 2025 12:09:19 +02:00
Committer:     Peter Zijlstra <peterz@infradead.org>
CommitterDate: Thu, 16 Oct 2025 11:13:54 +02:00

sched: Cleanup the sched_change NOCLOCK usage

Teach the sched_change pattern how to do update_rq_clock(); this
allows for some simplifications / cleanups.

Suggested-by: K Prateek Nayak <kprateek.nayak@amd.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Juri Lelli <juri.lelli@redhat.com>
Acked-by: Tejun Heo <tj@kernel.org>
Acked-by: Vincent Guittot <vincent.guittot@linaro.org>
---
 kernel/sched/core.c     | 33 +++++++++++----------------------
 kernel/sched/ext.c      |  4 +---
 kernel/sched/syscalls.c |  8 ++------
 3 files changed, 14 insertions(+), 31 deletions(-)

diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index e715147..3d5659f 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -2346,10 +2346,8 @@ static void migrate_disable_switch(struct rq *rq, struct task_struct *p)
 	if (p->cpus_ptr != &p->cpus_mask)
 		return;
 
-	scoped_guard (task_rq_lock, p) {
-		update_rq_clock(scope.rq);
+	scoped_guard (task_rq_lock, p)
 		do_set_cpus_allowed(p, &ac);
-	}
 }
 
 void ___migrate_enable(void)
@@ -2666,9 +2664,7 @@ void set_cpus_allowed_common(struct task_struct *p, struct affinity_context *ctx
 static void
 do_set_cpus_allowed(struct task_struct *p, struct affinity_context *ctx)
 {
-	u32 flags = DEQUEUE_SAVE | DEQUEUE_NOCLOCK;
-
-	scoped_guard (sched_change, p, flags) {
+	scoped_guard (sched_change, p, DEQUEUE_SAVE) {
 		p->sched_class->set_cpus_allowed(p, ctx);
 		mm_set_cpus_allowed(p->mm, ctx->new_mask);
 	}
@@ -2690,10 +2686,8 @@ void set_cpus_allowed_force(struct task_struct *p, const struct cpumask *new_mas
 		struct rcu_head rcu;
 	};
 
-	scoped_guard (__task_rq_lock, p) {
-		update_rq_clock(scope.rq);
+	scoped_guard (__task_rq_lock, p)
 		do_set_cpus_allowed(p, &ac);
-	}
 
 	/*
 	 * Because this is called with p->pi_lock held, it is not possible
@@ -9108,16 +9102,13 @@ static void sched_change_group(struct task_struct *tsk)
  */
 void sched_move_task(struct task_struct *tsk, bool for_autogroup)
 {
-	unsigned int queue_flags =
-		DEQUEUE_SAVE | DEQUEUE_MOVE | DEQUEUE_NOCLOCK;
+	unsigned int queue_flags = DEQUEUE_SAVE | DEQUEUE_MOVE;
 	bool resched = false;
 	struct rq *rq;
 
 	CLASS(task_rq_lock, rq_guard)(tsk);
 	rq = rq_guard.rq;
 
-	update_rq_clock(rq);
-
 	scoped_guard (sched_change, tsk, queue_flags) {
 		sched_change_group(tsk);
 		if (!for_autogroup)
@@ -10792,16 +10783,14 @@ struct sched_change_ctx *sched_change_begin(struct task_struct *p, unsigned int 
 
 	lockdep_assert_rq_held(rq);
 
+	if (!(flags & DEQUEUE_NOCLOCK)) {
+		update_rq_clock(rq);
+		flags |= DEQUEUE_NOCLOCK;
+	}
+
 	if (flags & DEQUEUE_CLASS) {
-		if (p->sched_class->switching_from) {
-			/*
-			 * switching_from_fair() assumes CLASS implies NOCLOCK;
-			 * fixing this assumption would mean switching_from()
-			 * would need to be able to change flags.
-			 */
-			WARN_ON(!(flags & DEQUEUE_NOCLOCK));
+		if (p->sched_class->switching_from)
 			p->sched_class->switching_from(rq, p);
-		}
 	}
 
 	*ctx = (struct sched_change_ctx){
@@ -10840,7 +10829,7 @@ void sched_change_end(struct sched_change_ctx *ctx)
 		p->sched_class->switching_to(rq, p);
 
 	if (ctx->queued)
-		enqueue_task(rq, p, ctx->flags | ENQUEUE_NOCLOCK);
+		enqueue_task(rq, p, ctx->flags);
 	if (ctx->running)
 		set_next_task(rq, p);
 
diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c
index ad371b6..5717042 100644
--- a/kernel/sched/ext.c
+++ b/kernel/sched/ext.c
@@ -4654,7 +4654,7 @@ static int scx_enable(struct sched_ext_ops *ops, struct bpf_link *link)
 	percpu_down_write(&scx_fork_rwsem);
 	scx_task_iter_start(&sti);
 	while ((p = scx_task_iter_next_locked(&sti))) {
-		unsigned int queue_flags = DEQUEUE_SAVE | DEQUEUE_MOVE | DEQUEUE_NOCLOCK;
+		unsigned int queue_flags = DEQUEUE_SAVE | DEQUEUE_MOVE;
 		const struct sched_class *old_class = p->sched_class;
 		const struct sched_class *new_class =
 			__setscheduler_class(p->policy, p->prio);
@@ -4662,8 +4662,6 @@ static int scx_enable(struct sched_ext_ops *ops, struct bpf_link *link)
 		if (!tryget_task_struct(p))
 			continue;
 
-		update_rq_clock(task_rq(p));
-
 		if (old_class != new_class)
 			queue_flags |= DEQUEUE_CLASS;
 
diff --git a/kernel/sched/syscalls.c b/kernel/sched/syscalls.c
index 20af564..8f0f603 100644
--- a/kernel/sched/syscalls.c
+++ b/kernel/sched/syscalls.c
@@ -64,7 +64,6 @@ static int effective_prio(struct task_struct *p)
 
 void set_user_nice(struct task_struct *p, long nice)
 {
-	struct rq *rq;
 	int old_prio;
 
 	if (task_nice(p) == nice || nice < MIN_NICE || nice > MAX_NICE)
@@ -73,10 +72,7 @@ void set_user_nice(struct task_struct *p, long nice)
 	 * We have to be careful, if called from sys_setpriority(),
 	 * the task might be in the middle of scheduling on another CPU.
 	 */
-	CLASS(task_rq_lock, rq_guard)(p);
-	rq = rq_guard.rq;
-
-	update_rq_clock(rq);
+	guard(task_rq_lock)(p);
 
 	/*
 	 * The RT priorities are set via sched_setscheduler(), but we still
@@ -89,7 +85,7 @@ void set_user_nice(struct task_struct *p, long nice)
 		return;
 	}
 
-	scoped_guard (sched_change, p, DEQUEUE_SAVE | DEQUEUE_NOCLOCK) {
+	scoped_guard (sched_change, p, DEQUEUE_SAVE) {
 		p->static_prio = NICE_TO_PRIO(nice);
 		set_load_weight(p, true);
 		old_prio = p->prio;