From: K Prateek Nayak <kprateek.nayak@amd.com>
To: Ingo Molnar <mingo@redhat.com>,
Peter Zijlstra <peterz@infradead.org>,
Juri Lelli <juri.lelli@redhat.com>,
Vincent Guittot <vincent.guittot@linaro.org>,
<linux-kernel@vger.kernel.org>
Cc: Dietmar Eggemann <dietmar.eggemann@arm.com>,
Steven Rostedt <rostedt@goodmis.org>,
Ben Segall <bsegall@google.com>, Mel Gorman <mgorman@suse.de>,
Valentin Schneider <vschneid@redhat.com>,
Chen Yu <yu.c.chen@intel.com>,
Shrikanth Hegde <sshegde@linux.ibm.com>,
"Gautham R. Shenoy" <gautham.shenoy@amd.com>,
K Prateek Nayak <kprateek.nayak@amd.com>
Subject: [PATCH v3 8/8] sched/fair: Simplify SIS_UTIL handling in select_idle_cpu()
Date: Tue, 20 Jan 2026 11:32:46 +0000 [thread overview]
Message-ID: <20260120113246.27987-9-kprateek.nayak@amd.com> (raw)
In-Reply-To: <20260120113246.27987-1-kprateek.nayak@amd.com>
Use the "sd_llc" passed to select_idle_cpu() to obtain the
"sd_llc_shared" instead of dereferencing the per-CPU variable.
Since "sd->shared" is always reclaimed at the same time as "sd" via
call_rcu() and update_top_cache_domain() always ensures a valid
"sd->shared" assignment when "sd_llc" is present, "sd_llc->shared" can
always be dereferenced without needing an additional check.
While at it move the cpumask_and() operation after the SIS_UTIL bailout
check to avoid unnecessarily computing the cpumask.
Signed-off-by: K Prateek Nayak <kprateek.nayak@amd.com>
---
Changelog rfc v2..v3:
o No changes. to the diff. Added more details on directly dereferencing
"sd->shared" without a NULL check in the commit message.
---
kernel/sched/fair.c | 19 ++++++++-----------
1 file changed, 8 insertions(+), 11 deletions(-)
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index c308c0700a7f..b4ae9444d32f 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -7629,21 +7629,18 @@ static int select_idle_cpu(struct task_struct *p, struct sched_domain *sd, bool
{
struct cpumask *cpus = this_cpu_cpumask_var_ptr(select_rq_mask);
int i, cpu, idle_cpu = -1, nr = INT_MAX;
- struct sched_domain_shared *sd_share;
-
- cpumask_and(cpus, sched_domain_span(sd), p->cpus_ptr);
if (sched_feat(SIS_UTIL)) {
- sd_share = rcu_dereference_all(per_cpu(sd_llc_shared, target));
- if (sd_share) {
- /* because !--nr is the condition to stop scan */
- nr = READ_ONCE(sd_share->nr_idle_scan) + 1;
- /* overloaded LLC is unlikely to have idle cpu/core */
- if (nr == 1)
- return -1;
- }
+ /* because !--nr is the condition to stop scan */
+ nr = READ_ONCE(sd->shared->nr_idle_scan) + 1;
+ /* overloaded LLC is unlikely to have idle cpu/core */
+ if (nr == 1)
+ return -1;
}
+ if (!cpumask_and(cpus, sched_domain_span(sd), p->cpus_ptr))
+ return -1;
+
if (static_branch_unlikely(&sched_cluster_active)) {
struct sched_group *sg = sd->groups;
--
2.34.1
next prev parent reply other threads:[~2026-01-20 11:35 UTC|newest]
Thread overview: 36+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-01-20 11:32 [PATCH v3 0/8] sched/topology: Optimize sd->shared allocation K Prateek Nayak
2026-01-20 11:32 ` [PATCH v3 1/8] sched/topology: Compute sd_weight considering cpuset partitions K Prateek Nayak
2026-01-21 14:45 ` Chen, Yu C
2026-01-21 15:42 ` Shrikanth Hegde
2026-01-22 2:51 ` K Prateek Nayak
2026-02-05 16:53 ` Valentin Schneider
2026-01-20 11:32 ` [PATCH v3 2/8] sched/topology: Allocate per-CPU sched_domain_shared in s_data K Prateek Nayak
2026-01-21 15:17 ` Chen, Yu C
2026-02-05 16:53 ` Valentin Schneider
2026-01-20 11:32 ` [PATCH v3 3/8] sched/topology: Switch to assigning "sd->shared" from s_data K Prateek Nayak
2026-01-21 15:26 ` Chen, Yu C
2026-01-22 2:49 ` K Prateek Nayak
2026-01-22 8:12 ` Shrikanth Hegde
2026-01-22 8:36 ` K Prateek Nayak
2026-01-23 4:08 ` Shrikanth Hegde
2026-01-23 4:53 ` K Prateek Nayak
2026-02-05 16:53 ` Valentin Schneider
2026-02-06 5:20 ` K Prateek Nayak
2026-02-06 9:38 ` Valentin Schneider
2026-02-14 3:04 ` Chen, Yu C
2026-02-16 3:50 ` K Prateek Nayak
2026-02-14 2:59 ` Chen, Yu C
2026-01-20 11:32 ` [PATCH v3 4/8] sched/topology: Remove sched_domain_shared allocation with sd_data K Prateek Nayak
2026-02-05 16:53 ` Valentin Schneider
2026-01-20 11:32 ` [PATCH v3 5/8] sched/core: Check for rcu_read_lock_any_held() in idle_get_state() K Prateek Nayak
2026-01-20 11:32 ` [PATCH v3 6/8] sched/fair: Remove superfluous rcu_read_lock() in the wakeup path K Prateek Nayak
2026-01-20 11:32 ` [PATCH v3 7/8] sched/fair: Simplify the entry condition for update_idle_cpu_scan() K Prateek Nayak
2026-02-14 15:41 ` Chen, Yu C
2026-01-20 11:32 ` K Prateek Nayak [this message]
2026-01-23 6:06 ` [PATCH v3 8/8] sched/fair: Simplify SIS_UTIL handling in select_idle_cpu() Shrikanth Hegde
2026-01-23 6:27 ` K Prateek Nayak
2026-01-23 7:14 ` Shrikanth Hegde
2026-02-14 15:56 ` Chen, Yu C
2026-01-21 16:16 ` [PATCH v3 0/8] sched/topology: Optimize sd->shared allocation Peter Zijlstra
2026-01-22 2:56 ` K Prateek Nayak
2026-01-23 9:54 ` Peter Zijlstra
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=20260120113246.27987-9-kprateek.nayak@amd.com \
--to=kprateek.nayak@amd.com \
--cc=bsegall@google.com \
--cc=dietmar.eggemann@arm.com \
--cc=gautham.shenoy@amd.com \
--cc=juri.lelli@redhat.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mgorman@suse.de \
--cc=mingo@redhat.com \
--cc=peterz@infradead.org \
--cc=rostedt@goodmis.org \
--cc=sshegde@linux.ibm.com \
--cc=vincent.guittot@linaro.org \
--cc=vschneid@redhat.com \
--cc=yu.c.chen@intel.com \
/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
all inboxes | Powered by JetHome®