From: Steve Sistare <steven.sistare@oracle.com>
To: mingo@redhat.com, peterz@infradead.org
Cc: subhra.mazumdar@oracle.com, dhaval.giani@oracle.com,
daniel.m.jordan@oracle.com, pavel.tatashin@microsoft.com,
matt@codeblueprint.co.uk, umgwanakikbuti@gmail.com,
riel@redhat.com, jbacik@fb.com, juri.lelli@redhat.com,
valentin.schneider@arm.com, vincent.guittot@linaro.org,
quentin.perret@arm.com, steven.sistare@oracle.com,
linux-kernel@vger.kernel.org
Subject: [PATCH v3 04/10] sched/fair: Dynamically update cfs_overload_cpus
Date: Fri, 9 Nov 2018 04:50:34 -0800 [thread overview]
Message-ID: <1541767840-93588-5-git-send-email-steven.sistare@oracle.com> (raw)
In-Reply-To: <1541767840-93588-1-git-send-email-steven.sistare@oracle.com>
An overloaded CPU has more than 1 runnable task. When a CFS task wakes
on a CPU, if h_nr_running transitions from 1 to more, then set the CPU in
the cfs_overload_cpus bitmap. When a CFS task sleeps, if h_nr_running
transitions from 2 to less, then clear the CPU in cfs_overload_cpus.
Signed-off-by: Steve Sistare <steven.sistare@oracle.com>
---
kernel/sched/fair.c | 52 ++++++++++++++++++++++++++++++++++++++++++++++++----
1 file changed, 48 insertions(+), 4 deletions(-)
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index ee271bb..9031d39 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -23,6 +23,7 @@
#include "sched.h"
#include <trace/events/sched.h>
+#include <linux/sparsemask.h>
/*
* Targeted preemption latency for CPU-bound tasks:
@@ -3724,6 +3725,28 @@ static inline void update_misfit_status(struct task_struct *p, struct rq *rq)
rq->misfit_task_load = task_h_load(p);
}
+static void overload_clear(struct rq *rq)
+{
+ struct sparsemask *overload_cpus;
+
+ rcu_read_lock();
+ overload_cpus = rcu_dereference(rq->cfs_overload_cpus);
+ if (overload_cpus)
+ sparsemask_clear_elem(rq->cpu, overload_cpus);
+ rcu_read_unlock();
+}
+
+static void overload_set(struct rq *rq)
+{
+ struct sparsemask *overload_cpus;
+
+ rcu_read_lock();
+ overload_cpus = rcu_dereference(rq->cfs_overload_cpus);
+ if (overload_cpus)
+ sparsemask_set_elem(rq->cpu, overload_cpus);
+ rcu_read_unlock();
+}
+
#else /* CONFIG_SMP */
#define UPDATE_TG 0x0
@@ -3747,6 +3770,9 @@ static inline int idle_balance(struct rq *rq, struct rq_flags *rf)
return 0;
}
+static inline void overload_clear(struct rq *rq) {}
+static inline void overload_set(struct rq *rq) {}
+
static inline void
util_est_enqueue(struct cfs_rq *cfs_rq, struct task_struct *p) {}
@@ -4441,6 +4467,7 @@ static int tg_throttle_down(struct task_group *tg, void *data)
static void throttle_cfs_rq(struct cfs_rq *cfs_rq)
{
struct rq *rq = rq_of(cfs_rq);
+ unsigned int prev_nr = rq->cfs.h_nr_running;
struct cfs_bandwidth *cfs_b = tg_cfs_bandwidth(cfs_rq->tg);
struct sched_entity *se;
long task_delta, dequeue = 1;
@@ -4468,8 +4495,12 @@ static void throttle_cfs_rq(struct cfs_rq *cfs_rq)
dequeue = 0;
}
- if (!se)
+ if (!se) {
sub_nr_running(rq, task_delta);
+ if (prev_nr >= 2 && prev_nr - task_delta < 2)
+ overload_clear(rq);
+
+ }
cfs_rq->throttled = 1;
cfs_rq->throttled_clock = rq_clock(rq);
@@ -4499,6 +4530,7 @@ static void throttle_cfs_rq(struct cfs_rq *cfs_rq)
void unthrottle_cfs_rq(struct cfs_rq *cfs_rq)
{
struct rq *rq = rq_of(cfs_rq);
+ unsigned int prev_nr = rq->cfs.h_nr_running;
struct cfs_bandwidth *cfs_b = tg_cfs_bandwidth(cfs_rq->tg);
struct sched_entity *se;
int enqueue = 1;
@@ -4535,8 +4567,11 @@ void unthrottle_cfs_rq(struct cfs_rq *cfs_rq)
break;
}
- if (!se)
+ if (!se) {
add_nr_running(rq, task_delta);
+ if (prev_nr < 2 && prev_nr + task_delta >= 2)
+ overload_set(rq);
+ }
/* Determine whether we need to wake up potentially idle CPU: */
if (rq->curr == rq->idle && rq->cfs.nr_running)
@@ -5082,6 +5117,7 @@ static inline void hrtick_update(struct rq *rq)
{
struct cfs_rq *cfs_rq;
struct sched_entity *se = &p->se;
+ unsigned int prev_nr = rq->cfs.h_nr_running;
/*
* The code below (indirectly) updates schedutil which looks at
@@ -5129,8 +5165,12 @@ static inline void hrtick_update(struct rq *rq)
update_cfs_group(se);
}
- if (!se)
+ if (!se) {
add_nr_running(rq, 1);
+ if (prev_nr == 1)
+ overload_set(rq);
+
+ }
hrtick_update(rq);
}
@@ -5147,6 +5187,7 @@ static void dequeue_task_fair(struct rq *rq, struct task_struct *p, int flags)
struct cfs_rq *cfs_rq;
struct sched_entity *se = &p->se;
int task_sleep = flags & DEQUEUE_SLEEP;
+ unsigned int prev_nr = rq->cfs.h_nr_running;
for_each_sched_entity(se) {
cfs_rq = cfs_rq_of(se);
@@ -5188,8 +5229,11 @@ static void dequeue_task_fair(struct rq *rq, struct task_struct *p, int flags)
update_cfs_group(se);
}
- if (!se)
+ if (!se) {
sub_nr_running(rq, 1);
+ if (prev_nr == 2)
+ overload_clear(rq);
+ }
util_est_dequeue(&rq->cfs, p, task_sleep);
hrtick_update(rq);
--
1.8.3.1
next prev parent reply other threads:[~2018-11-09 13:00 UTC|newest]
Thread overview: 30+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-11-09 12:50 [PATCH v3 00/10] steal tasks to improve CPU utilization Steve Sistare
2018-11-09 12:50 ` [PATCH v3 01/10] sched: Provide sparsemask, a reduced contention bitmap Steve Sistare
2018-11-27 15:16 ` Steven Sistare
2018-11-28 1:19 ` Omar Sandoval
2018-12-06 16:07 ` Steven Sistare
2018-12-06 18:19 ` Omar Sandoval
2018-11-09 12:50 ` [PATCH v3 02/10] sched/topology: Provide hooks to allocate data shared per LLC Steve Sistare
2018-11-09 12:50 ` [PATCH v3 03/10] sched/topology: Provide cfs_overload_cpus bitmap Steve Sistare
2018-11-09 17:38 ` Valentin Schneider
2018-11-19 17:32 ` Steven Sistare
2018-11-20 12:52 ` Valentin Schneider
2018-11-12 16:42 ` Valentin Schneider
2018-11-19 17:33 ` Steven Sistare
2018-11-20 12:42 ` Valentin Schneider
2018-11-26 19:06 ` Steven Sistare
2018-12-03 16:56 ` Valentin Schneider
2018-12-06 16:40 ` Steven Sistare
2018-12-06 17:28 ` Valentin Schneider
2018-11-09 12:50 ` Steve Sistare [this message]
2018-11-09 12:50 ` [PATCH v3 05/10] sched/fair: Hoist idle_stamp up from idle_balance Steve Sistare
2018-11-09 19:07 ` Valentin Schneider
2018-11-19 17:31 ` Steven Sistare
2018-11-20 10:24 ` Valentin Schneider
2018-11-09 12:50 ` [PATCH v3 06/10] sched/fair: Generalize the detach_task interface Steve Sistare
2018-11-09 12:50 ` [PATCH v3 07/10] sched/fair: Provide can_migrate_task_llc Steve Sistare
2018-11-09 12:50 ` [PATCH v3 08/10] sched/fair: Steal work from an overloaded CPU when CPU goes idle Steve Sistare
2018-11-09 12:50 ` [PATCH v3 09/10] sched/fair: disable stealing if too many NUMA nodes Steve Sistare
2018-11-09 12:50 ` [PATCH v3 10/10] sched/fair: Provide idle search schedstats Steve Sistare
2018-11-10 17:08 ` kbuild test robot
2018-11-09 15:02 ` hackbench run scripts Steven Sistare
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=1541767840-93588-5-git-send-email-steven.sistare@oracle.com \
--to=steven.sistare@oracle.com \
--cc=daniel.m.jordan@oracle.com \
--cc=dhaval.giani@oracle.com \
--cc=jbacik@fb.com \
--cc=juri.lelli@redhat.com \
--cc=linux-kernel@vger.kernel.org \
--cc=matt@codeblueprint.co.uk \
--cc=mingo@redhat.com \
--cc=pavel.tatashin@microsoft.com \
--cc=peterz@infradead.org \
--cc=quentin.perret@arm.com \
--cc=riel@redhat.com \
--cc=subhra.mazumdar@oracle.com \
--cc=umgwanakikbuti@gmail.com \
--cc=valentin.schneider@arm.com \
--cc=vincent.guittot@linaro.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;
as well as URLs for NNTP newsgroup(s).