From: Frederic Weisbecker <frederic@kernel.org>
To: LKML <linux-kernel@vger.kernel.org>
Cc: Frederic Weisbecker <frederic@kernel.org>,
Marco Crivellari <marco.crivellari@suse.com>,
Michal Hocko <mhocko@suse.com>,
Peter Zijlstra <peterz@infradead.org>, Tejun Heo <tj@kernel.org>,
Thomas Gleixner <tglx@linutronix.de>,
Waiman Long <longman@redhat.com>
Subject: [PATCH 30/33] kthread: Add API to update preferred affinity on kthread runtime
Date: Fri, 29 Aug 2025 17:48:11 +0200 [thread overview]
Message-ID: <20250829154814.47015-31-frederic@kernel.org> (raw)
In-Reply-To: <20250829154814.47015-1-frederic@kernel.org>
Kthreads can apply for a preferred affinity upon creation but they have
no means to update that preferred affinity after the first wake up.
kthread_affine_preferred() is optimized by assuming the kthread
is sleeping while applying the allowed cpumask.
Therefore introduce a new API to further update the preferred affinity.
It will be used by IRQ kthreads.
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
---
include/linux/kthread.h | 1 +
kernel/kthread.c | 55 +++++++++++++++++++++++++++++++++++------
2 files changed, 48 insertions(+), 8 deletions(-)
diff --git a/include/linux/kthread.h b/include/linux/kthread.h
index c92c1149ee6e..a06cae7f2c55 100644
--- a/include/linux/kthread.h
+++ b/include/linux/kthread.h
@@ -86,6 +86,7 @@ void free_kthread_struct(struct task_struct *k);
void kthread_bind(struct task_struct *k, unsigned int cpu);
void kthread_bind_mask(struct task_struct *k, const struct cpumask *mask);
int kthread_affine_preferred(struct task_struct *p, const struct cpumask *mask);
+int kthread_affine_preferred_update(struct task_struct *p, const struct cpumask *mask);
int kthread_stop(struct task_struct *k);
int kthread_stop_put(struct task_struct *k);
bool kthread_should_stop(void);
diff --git a/kernel/kthread.c b/kernel/kthread.c
index d36bdfbd004e..f3397cf7542a 100644
--- a/kernel/kthread.c
+++ b/kernel/kthread.c
@@ -322,17 +322,16 @@ EXPORT_SYMBOL_GPL(kthread_parkme);
void __noreturn kthread_exit(long result)
{
struct kthread *kthread = to_kthread(current);
+ struct cpumask *to_free = NULL;
kthread->result = result;
- if (!list_empty(&kthread->affinity_node)) {
- mutex_lock(&kthread_affinity_lock);
- list_del(&kthread->affinity_node);
- mutex_unlock(&kthread_affinity_lock);
- if (kthread->preferred_affinity) {
- kfree(kthread->preferred_affinity);
- kthread->preferred_affinity = NULL;
- }
+ scoped_guard(mutex, &kthread_affinity_lock) {
+ if (!list_empty(&kthread->affinity_node))
+ list_del_init(&kthread->affinity_node);
+ to_free = kthread->preferred_affinity;
+ kthread->preferred_affinity = NULL;
}
+ kfree(to_free);
do_exit(0);
}
EXPORT_SYMBOL(kthread_exit);
@@ -900,6 +899,46 @@ int kthread_affine_preferred(struct task_struct *p, const struct cpumask *mask)
}
EXPORT_SYMBOL_GPL(kthread_affine_preferred);
+/**
+ * kthread_affine_preferred_update - update a kthread's preferred affinity
+ * @p: thread created by kthread_create().
+ * @cpumask: new mask of CPUs (might not be online, must be possible) for @k
+ * to run on.
+ *
+ * Update the cpumask of the desired kthread's affinity that was passed by
+ * a previous call to kthread_affine_preferred(). This can be called either
+ * before or after the first wakeup of the kthread.
+ *
+ * Returns 0 if the affinity has been applied.
+ */
+int kthread_affine_preferred_update(struct task_struct *p,
+ const struct cpumask *mask)
+{
+ struct kthread *kthread = to_kthread(p);
+ cpumask_var_t affinity;
+ int ret = 0;
+
+ if (!zalloc_cpumask_var(&affinity, GFP_KERNEL))
+ return -ENOMEM;
+
+ scoped_guard(mutex, &kthread_affinity_lock) {
+ if (WARN_ON_ONCE(!kthread->preferred_affinity ||
+ list_empty(&kthread->affinity_node))) {
+ ret = -EINVAL;
+ goto out;
+ }
+
+ cpumask_copy(kthread->preferred_affinity, mask);
+ kthread_fetch_affinity(kthread, affinity);
+ set_cpus_allowed_ptr(p, affinity);
+ }
+out:
+ free_cpumask_var(affinity);
+
+ return ret;
+}
+EXPORT_SYMBOL_GPL(kthread_affine_preferred_update);
+
static int kthreads_update_affinity(bool force)
{
cpumask_var_t affinity;
--
2.51.0
next prev parent reply other threads:[~2025-08-29 15:49 UTC|newest]
Thread overview: 64+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-08-29 15:47 [PATCH 00/33 v2] cpuset/isolation: Honour kthreads preferred affinity Frederic Weisbecker
2025-08-29 15:47 ` [PATCH 01/33] sched/isolation: Remove housekeeping static key Frederic Weisbecker
2025-08-29 21:34 ` Waiman Long
2025-09-18 12:04 ` Frederic Weisbecker
2025-09-01 10:26 ` Peter Zijlstra
2025-09-18 13:18 ` Frederic Weisbecker
2025-09-11 20:57 ` Phil Auld
2025-08-29 15:47 ` [PATCH 02/33] PCI: Protect against concurrent change of housekeeping cpumask Frederic Weisbecker
[not found] ` <458c5db8-0c31-4c02-9c41-b7eca851d04a@redhat.com>
2025-09-18 14:00 ` Frederic Weisbecker
2025-09-22 21:51 ` Waiman Long
2025-09-23 9:07 ` Frederic Weisbecker
2025-08-29 15:47 ` [PATCH 03/33] cpu: Revert "cpu/hotplug: Prevent self deadlock on CPU hot-unplug" Frederic Weisbecker
2025-08-29 15:47 ` [PATCH 04/33] memcg: Prepare to protect against concurrent isolated cpuset change Frederic Weisbecker
2025-08-29 15:47 ` [PATCH 05/33] mm: vmstat: " Frederic Weisbecker
2025-08-29 15:47 ` [PATCH 06/33] sched/isolation: Save boot defined domain flags Frederic Weisbecker
2025-09-11 21:02 ` Phil Auld
2025-08-29 15:47 ` [PATCH 07/33] cpuset: Convert boot_hk_cpus to use HK_TYPE_DOMAIN_BOOT Frederic Weisbecker
2025-09-11 21:03 ` Phil Auld
2025-08-29 15:47 ` [PATCH 08/33] driver core: cpu: Convert /sys/devices/system/cpu/isolated " Frederic Weisbecker
2025-08-29 15:47 ` [PATCH 09/33] net: Keep ignoring isolated cpuset change Frederic Weisbecker
2025-08-29 15:47 ` [PATCH 10/33] block: Protect against concurrent " Frederic Weisbecker
2025-08-29 15:47 ` [PATCH 11/33] cpu: Provide lockdep check for CPU hotplug lock write-held Frederic Weisbecker
2025-08-29 15:47 ` [PATCH 12/33] cpuset: Provide lockdep check for cpuset lock held Frederic Weisbecker
2025-08-29 15:47 ` [PATCH 13/33] sched/isolation: Convert housekeeping cpumasks to rcu pointers Frederic Weisbecker
2025-08-29 15:47 ` [PATCH 14/33] cpuset: Update HK_TYPE_DOMAIN cpumask from cpuset Frederic Weisbecker
2025-09-01 0:40 ` Waiman Long
2025-09-22 14:57 ` Frederic Weisbecker
2025-08-29 15:47 ` [PATCH 15/33] sched/isolation: Flush memcg workqueues on cpuset isolated partition change Frederic Weisbecker
2025-08-29 15:47 ` [PATCH 16/33] sched/isolation: Flush vmstat " Frederic Weisbecker
2025-08-29 15:47 ` [PATCH 17/33] cpuset: Propagate cpuset isolation update to workqueue through housekeeping Frederic Weisbecker
2025-09-01 2:51 ` Waiman Long
2025-09-22 15:10 ` Frederic Weisbecker
2025-08-29 15:47 ` [PATCH 18/33] cpuset: Remove cpuset_cpu_is_isolated() Frederic Weisbecker
2025-08-29 15:48 ` [PATCH 19/33] sched/isolation: Remove HK_TYPE_TICK test from cpu_is_isolated() Frederic Weisbecker
2025-09-02 14:28 ` Waiman Long
2025-09-02 15:48 ` Waiman Long
2025-09-22 15:20 ` Frederic Weisbecker
2025-09-22 15:19 ` Frederic Weisbecker
2025-09-22 21:59 ` Waiman Long
2025-09-23 9:11 ` Frederic Weisbecker
2025-08-29 15:48 ` [PATCH 20/33] PCI: Remove superfluous HK_TYPE_WQ check Frederic Weisbecker
2025-08-29 15:48 ` [PATCH 21/33] kthread: Refine naming of affinity related fields Frederic Weisbecker
2025-08-29 15:48 ` [PATCH 22/33] kthread: Include unbound kthreads in the managed affinity list Frederic Weisbecker
2025-08-29 15:48 ` [PATCH 23/33] kthread: Include kthreadd to " Frederic Weisbecker
2025-08-29 15:48 ` [PATCH 24/33] kthread: Rely on HK_TYPE_DOMAIN for preferred affinity management Frederic Weisbecker
2025-08-29 15:48 ` [PATCH 25/33] sched: Switch the fallback task allowed cpumask to HK_TYPE_DOMAIN Frederic Weisbecker
2025-08-29 15:48 ` [PATCH 26/33] cgroup/cpuset: Fail if isolated and nohz_full don't leave any housekeeping Frederic Weisbecker
2025-09-02 15:44 ` Waiman Long
2025-09-23 9:17 ` Frederic Weisbecker
2025-09-23 9:24 ` Gabriele Monaco
2025-08-29 15:48 ` [PATCH 27/33] sched/arm64: Move fallback task cpumask to HK_TYPE_DOMAIN Frederic Weisbecker
2025-09-02 16:43 ` Waiman Long
2025-09-23 9:43 ` Frederic Weisbecker
2025-08-29 15:48 ` [PATCH 28/33] kthread: Honour kthreads preferred affinity after cpuset changes Frederic Weisbecker
2025-08-29 15:48 ` [PATCH 29/33] kthread: Comment on the purpose and placement of kthread_affine_node() call Frederic Weisbecker
2025-08-29 15:48 ` Frederic Weisbecker [this message]
2025-08-29 15:48 ` [PATCH 31/33] kthread: Document kthread_affine_preferred() Frederic Weisbecker
2025-08-29 15:48 ` [RFC PATCH 32/33] genirq: Correctly handle preferred kthreads affinity Frederic Weisbecker
2025-08-29 15:48 ` [PATCH 33/33] doc: Add housekeeping documentation Frederic Weisbecker
2025-09-02 19:12 ` [PATCH 00/33 v2] cpuset/isolation: Honour kthreads preferred affinity Waiman Long
2025-09-23 9:48 ` Frederic Weisbecker
2025-10-13 20:31 [PATCH 00/33 v3] " Frederic Weisbecker
2025-10-13 20:31 ` [PATCH 30/33] kthread: Add API to update preferred affinity on kthread runtime Frederic Weisbecker
2025-10-14 12:35 ` Simon Horman
2025-11-05 17:26 ` Frederic Weisbecker
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20250829154814.47015-31-frederic@kernel.org \
--to=frederic@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=longman@redhat.com \
--cc=marco.crivellari@suse.com \
--cc=mhocko@suse.com \
--cc=peterz@infradead.org \
--cc=tglx@linutronix.de \
--cc=tj@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
Powered by JetHome