Use the new sched_class::switching_from() method to dequeue delayed
tasks before switching to another class.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Tejun Heo <tj@kernel.org>
---
kernel/sched/core.c | 12 ++++++++----
kernel/sched/ext.c | 6 ------
kernel/sched/fair.c | 7 +++++++
kernel/sched/syscalls.c | 3 ---
4 files changed, 15 insertions(+), 13 deletions(-)
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -7368,9 +7368,6 @@ void rt_mutex_setprio(struct task_struct
queue_flag &= ~(DEQUEUE_SAVE | DEQUEUE_MOVE);
}
- if (prev_class != next_class && p->se.sched_delayed)
- dequeue_task(rq, p, DEQUEUE_SLEEP | DEQUEUE_DELAYED | DEQUEUE_NOCLOCK);
-
scoped_guard (sched_change, p, queue_flag) {
/*
* Boosting condition are:
@@ -10845,8 +10842,15 @@ struct sched_change_ctx *sched_change_be
if (WARN_ON_ONCE(flags & (DEQUEUE_SAVE | DEQUEUE_MOVE)))
flags &= ~(DEQUEUE_SAVE | DEQUEUE_MOVE);
- if (p->sched_class->switching_from)
+ 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));
p->sched_class->switching_from(rq, p);
+ }
}
*ctx = (struct sched_change_ctx){
--- a/kernel/sched/ext.c
+++ b/kernel/sched/ext.c
@@ -3924,9 +3924,6 @@ static void scx_disable_workfn(struct kt
queue_flags &= ~(DEQUEUE_SAVE | DEQUEUE_MOVE);
}
- if (old_class != new_class && p->se.sched_delayed)
- dequeue_task(task_rq(p), p, DEQUEUE_SLEEP | DEQUEUE_DELAYED | DEQUEUE_NOCLOCK);
-
scoped_guard (sched_change, p, queue_flags) {
p->sched_class = new_class;
}
@@ -4677,9 +4674,6 @@ static int scx_enable(struct sched_ext_o
queue_flags &= ~(DEQUEUE_SAVE | DEQUEUE_MOVE);
}
- if (old_class != new_class && p->se.sched_delayed)
- dequeue_task(task_rq(p), p, DEQUEUE_SLEEP | DEQUEUE_DELAYED | DEQUEUE_NOCLOCK);
-
scoped_guard (sched_change, p, queue_flags) {
p->scx.slice = SCX_SLICE_DFL;
p->sched_class = new_class;
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -13237,6 +13237,12 @@ static void attach_task_cfs_rq(struct ta
attach_entity_cfs_rq(se);
}
+static void switching_from_fair(struct rq *rq, struct task_struct *p)
+{
+ if (p->se.sched_delayed)
+ dequeue_task(rq, p, DEQUEUE_SLEEP | DEQUEUE_DELAYED | DEQUEUE_NOCLOCK);
+}
+
static void switched_from_fair(struct rq *rq, struct task_struct *p)
{
detach_task_cfs_rq(p);
@@ -13638,6 +13644,7 @@ DEFINE_SCHED_CLASS(fair) = {
.reweight_task = reweight_task_fair,
.prio_changed = prio_changed_fair,
+ .switching_from = switching_from_fair,
.switched_from = switched_from_fair,
.switched_to = switched_to_fair,
--- a/kernel/sched/syscalls.c
+++ b/kernel/sched/syscalls.c
@@ -689,9 +689,6 @@ int __sched_setscheduler(struct task_str
queue_flags &= ~(DEQUEUE_SAVE | DEQUEUE_MOVE);
}
- if (prev_class != next_class && p->se.sched_delayed)
- dequeue_task(rq, p, DEQUEUE_SLEEP | DEQUEUE_DELAYED | DEQUEUE_NOCLOCK);
-
scoped_guard (sched_change, p, queue_flags) {
if (!(attr->sched_flags & SCHED_FLAG_KEEP_PARAMS)) {On Mon, 6 Oct 2025 at 12:45, Peter Zijlstra <peterz@infradead.org> wrote:
>
> Use the new sched_class::switching_from() method to dequeue delayed
> tasks before switching to another class.
>
> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
> Acked-by: Tejun Heo <tj@kernel.org>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
> ---
> kernel/sched/core.c | 12 ++++++++----
> kernel/sched/ext.c | 6 ------
> kernel/sched/fair.c | 7 +++++++
> kernel/sched/syscalls.c | 3 ---
> 4 files changed, 15 insertions(+), 13 deletions(-)
>
> --- a/kernel/sched/core.c
> +++ b/kernel/sched/core.c
> @@ -7368,9 +7368,6 @@ void rt_mutex_setprio(struct task_struct
> queue_flag &= ~(DEQUEUE_SAVE | DEQUEUE_MOVE);
> }
>
> - if (prev_class != next_class && p->se.sched_delayed)
> - dequeue_task(rq, p, DEQUEUE_SLEEP | DEQUEUE_DELAYED | DEQUEUE_NOCLOCK);
> -
> scoped_guard (sched_change, p, queue_flag) {
> /*
> * Boosting condition are:
> @@ -10845,8 +10842,15 @@ struct sched_change_ctx *sched_change_be
> if (WARN_ON_ONCE(flags & (DEQUEUE_SAVE | DEQUEUE_MOVE)))
> flags &= ~(DEQUEUE_SAVE | DEQUEUE_MOVE);
>
> - if (p->sched_class->switching_from)
> + 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));
> p->sched_class->switching_from(rq, p);
> + }
> }
>
> *ctx = (struct sched_change_ctx){
> --- a/kernel/sched/ext.c
> +++ b/kernel/sched/ext.c
> @@ -3924,9 +3924,6 @@ static void scx_disable_workfn(struct kt
> queue_flags &= ~(DEQUEUE_SAVE | DEQUEUE_MOVE);
> }
>
> - if (old_class != new_class && p->se.sched_delayed)
> - dequeue_task(task_rq(p), p, DEQUEUE_SLEEP | DEQUEUE_DELAYED | DEQUEUE_NOCLOCK);
> -
> scoped_guard (sched_change, p, queue_flags) {
> p->sched_class = new_class;
> }
> @@ -4677,9 +4674,6 @@ static int scx_enable(struct sched_ext_o
> queue_flags &= ~(DEQUEUE_SAVE | DEQUEUE_MOVE);
> }
>
> - if (old_class != new_class && p->se.sched_delayed)
> - dequeue_task(task_rq(p), p, DEQUEUE_SLEEP | DEQUEUE_DELAYED | DEQUEUE_NOCLOCK);
> -
> scoped_guard (sched_change, p, queue_flags) {
> p->scx.slice = SCX_SLICE_DFL;
> p->sched_class = new_class;
> --- a/kernel/sched/fair.c
> +++ b/kernel/sched/fair.c
> @@ -13237,6 +13237,12 @@ static void attach_task_cfs_rq(struct ta
> attach_entity_cfs_rq(se);
> }
>
> +static void switching_from_fair(struct rq *rq, struct task_struct *p)
> +{
> + if (p->se.sched_delayed)
> + dequeue_task(rq, p, DEQUEUE_SLEEP | DEQUEUE_DELAYED | DEQUEUE_NOCLOCK);
> +}
> +
> static void switched_from_fair(struct rq *rq, struct task_struct *p)
> {
> detach_task_cfs_rq(p);
> @@ -13638,6 +13644,7 @@ DEFINE_SCHED_CLASS(fair) = {
>
> .reweight_task = reweight_task_fair,
> .prio_changed = prio_changed_fair,
> + .switching_from = switching_from_fair,
> .switched_from = switched_from_fair,
> .switched_to = switched_to_fair,
>
> --- a/kernel/sched/syscalls.c
> +++ b/kernel/sched/syscalls.c
> @@ -689,9 +689,6 @@ int __sched_setscheduler(struct task_str
> queue_flags &= ~(DEQUEUE_SAVE | DEQUEUE_MOVE);
> }
>
> - if (prev_class != next_class && p->se.sched_delayed)
> - dequeue_task(rq, p, DEQUEUE_SLEEP | DEQUEUE_DELAYED | DEQUEUE_NOCLOCK);
> -
> scoped_guard (sched_change, p, queue_flags) {
>
> if (!(attr->sched_flags & SCHED_FLAG_KEEP_PARAMS)) {
>
>
The following commit has been merged into the sched/core branch of tip:
Commit-ID: 1ae5f5dfe5adc64a90b1b0ab5bd9bd7c9d140c28
Gitweb: https://git.kernel.org/tip/1ae5f5dfe5adc64a90b1b0ab5bd9bd7c9d140c28
Author: Peter Zijlstra <peterz@infradead.org>
AuthorDate: Wed, 30 Oct 2024 15:47:46 +01:00
Committer: Peter Zijlstra <peterz@infradead.org>
CommitterDate: Thu, 16 Oct 2025 11:13:51 +02:00
sched: Cleanup sched_delayed handling for class switches
Use the new sched_class::switching_from() method to dequeue delayed
tasks before switching to another class.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
Reviewed-by: Juri Lelli <juri.lelli@redhat.com>
Acked-by: Tejun Heo <tj@kernel.org>
---
kernel/sched/core.c | 12 ++++++++----
kernel/sched/ext.c | 6 ------
kernel/sched/fair.c | 7 +++++++
kernel/sched/syscalls.c | 3 ---
4 files changed, 15 insertions(+), 13 deletions(-)
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 4dbd206..bd2c551 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -7366,9 +7366,6 @@ void rt_mutex_setprio(struct task_struct *p, struct task_struct *pi_task)
if (prev_class != next_class)
queue_flag |= DEQUEUE_CLASS;
- if (prev_class != next_class && p->se.sched_delayed)
- dequeue_task(rq, p, DEQUEUE_SLEEP | DEQUEUE_DELAYED | DEQUEUE_NOCLOCK);
-
scoped_guard (sched_change, p, queue_flag) {
/*
* Boosting condition are:
@@ -10840,8 +10837,15 @@ struct sched_change_ctx *sched_change_begin(struct task_struct *p, unsigned int
lockdep_assert_rq_held(rq);
if (flags & DEQUEUE_CLASS) {
- if (p->sched_class->switching_from)
+ 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));
p->sched_class->switching_from(rq, p);
+ }
}
*ctx = (struct sched_change_ctx){
diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c
index a408c39..b0a1e2a 100644
--- a/kernel/sched/ext.c
+++ b/kernel/sched/ext.c
@@ -3922,9 +3922,6 @@ static void scx_disable_workfn(struct kthread_work *work)
if (old_class != new_class)
queue_flags |= DEQUEUE_CLASS;
- if (old_class != new_class && p->se.sched_delayed)
- dequeue_task(task_rq(p), p, DEQUEUE_SLEEP | DEQUEUE_DELAYED | DEQUEUE_NOCLOCK);
-
scoped_guard (sched_change, p, queue_flags) {
p->sched_class = new_class;
}
@@ -4673,9 +4670,6 @@ static int scx_enable(struct sched_ext_ops *ops, struct bpf_link *link)
if (old_class != new_class)
queue_flags |= DEQUEUE_CLASS;
- if (old_class != new_class && p->se.sched_delayed)
- dequeue_task(task_rq(p), p, DEQUEUE_SLEEP | DEQUEUE_DELAYED | DEQUEUE_NOCLOCK);
-
scoped_guard (sched_change, p, queue_flags) {
p->scx.slice = SCX_SLICE_DFL;
p->sched_class = new_class;
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index ac881df..6c462e4 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -13249,6 +13249,12 @@ static void attach_task_cfs_rq(struct task_struct *p)
attach_entity_cfs_rq(se);
}
+static void switching_from_fair(struct rq *rq, struct task_struct *p)
+{
+ if (p->se.sched_delayed)
+ dequeue_task(rq, p, DEQUEUE_SLEEP | DEQUEUE_DELAYED | DEQUEUE_NOCLOCK);
+}
+
static void switched_from_fair(struct rq *rq, struct task_struct *p)
{
detach_task_cfs_rq(p);
@@ -13650,6 +13656,7 @@ DEFINE_SCHED_CLASS(fair) = {
.reweight_task = reweight_task_fair,
.prio_changed = prio_changed_fair,
+ .switching_from = switching_from_fair,
.switched_from = switched_from_fair,
.switched_to = switched_to_fair,
diff --git a/kernel/sched/syscalls.c b/kernel/sched/syscalls.c
index bcef5c7..6583faf 100644
--- a/kernel/sched/syscalls.c
+++ b/kernel/sched/syscalls.c
@@ -687,9 +687,6 @@ change:
if (prev_class != next_class)
queue_flags |= DEQUEUE_CLASS;
- if (prev_class != next_class && p->se.sched_delayed)
- dequeue_task(rq, p, DEQUEUE_SLEEP | DEQUEUE_DELAYED | DEQUEUE_NOCLOCK);
-
scoped_guard (sched_change, p, queue_flags) {
if (!(attr->sched_flags & SCHED_FLAG_KEEP_PARAMS)) {© 2016 - 2026 Red Hat, Inc.