From: Morten Rasmussen <morten.rasmussen@arm.com>
To: mingo@kernel.org, peterz@infradead.org
Cc: pjt@google.com, arjan@linux.intel.com, rjw@sisk.pl,
dirk.j.brandewie@intel.com, vincent.guittot@linaro.org,
alex.shi@linaro.org, preeti@linux.vnet.ibm.com, efault@gmx.de,
corbet@lwn.net, tglx@linutronix.de, catalin.marinas@arm.com,
morten.rasmussen@arm.com, linux-kernel@vger.kernel.org,
linaro-kernel@lists.linaro.org
Subject: [RFC][PATCH 7/7] sched: power: Let the power driver choose the best wake-up cpu
Date: Fri, 11 Oct 2013 18:19:17 +0100 [thread overview]
Message-ID: <1381511957-29776-8-git-send-email-morten.rasmussen@arm.com> (raw)
In-Reply-To: <1381511957-29776-1-git-send-email-morten.rasmussen@arm.com>
Adds best_wake_cpu() to the power driver interface allowing the power
driver to influence which cpu will be woken up when additional cpus
are needed. Currently only a place holder until idle interface is in
place.
Signed-off-by: Morten Rasmussen <morten.rasmussen@arm.com>
---
drivers/cpufreq/cpufreq_schedpower.c | 11 +++++++++++
include/linux/sched/power.h | 2 ++
kernel/sched/fair.c | 5 ++++-
kernel/sched/power.c | 9 +++++++++
kernel/sched/sched.h | 6 ++++++
5 files changed, 32 insertions(+), 1 deletion(-)
diff --git a/drivers/cpufreq/cpufreq_schedpower.c b/drivers/cpufreq/cpufreq_schedpower.c
index 5952c79..62847ca 100644
--- a/drivers/cpufreq/cpufreq_schedpower.c
+++ b/drivers/cpufreq/cpufreq_schedpower.c
@@ -187,6 +187,16 @@ int pdriver_go_slower(int cpu, int hint)
return 1;
}
+int pdriver_best_wake_cpu(void)
+{
+ /*
+ * Place holder. Should ideally determine the best cpu to wake up
+ * when more cpus are needed. This may be based on coupled C-states,
+ * power domains, and other platform specific factors.
+ */
+ return nr_cpu_ids;
+}
+
void pdriver_late_callback(int cpu)
{
struct cpufreq_schedpower_cpuinfo *pcpu_info;
@@ -202,6 +212,7 @@ static struct power_driver pdriver = {
.at_max_capacity = pdriver_at_max_capacity,
.go_faster = pdriver_go_faster,
.go_slower = pdriver_go_slower,
+ .best_wake_cpu = pdriver_best_wake_cpu,
.late_callback = pdriver_late_callback,
};
diff --git a/include/linux/sched/power.h b/include/linux/sched/power.h
index aba5f47..5d76f44 100644
--- a/include/linux/sched/power.h
+++ b/include/linux/sched/power.h
@@ -25,6 +25,8 @@ struct power_driver {
int (*go_faster) (int cpu, int hint);
/* Decrease cpu capacity hint */
int (*go_slower) (int cpu, int hint);
+ /* Best cpu to wake up */
+ int (*best_wake_cpu) (void);
/* Scheduler call-back without rq lock held and with irq enabled */
void (*late_callback) (int cpu);
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index ee87e65..d32fb97 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -5458,7 +5458,10 @@ static struct {
static inline int find_new_ilb(int call_cpu)
{
- int ilb = cpumask_first(nohz.idle_cpus_mask);
+ int ilb = best_wake_cpu();
+
+ if (ilb >= nr_cpu_ids)
+ ilb = cpumask_first(nohz.idle_cpus_mask);
if (ilb < nr_cpu_ids && idle_cpu(ilb))
return ilb;
diff --git a/kernel/sched/power.c b/kernel/sched/power.c
index f70b563..e1357d2 100644
--- a/kernel/sched/power.c
+++ b/kernel/sched/power.c
@@ -11,6 +11,7 @@
*/
#include <linux/kernel.h>
+#include <linux/cpumask.h>
#include <linux/sched/power.h>
static struct power_driver *power_driver;
@@ -56,6 +57,14 @@ int go_slower(int cpu, int hint)
return power_driver->go_slower(cpu, hint);
}
+int best_wake_cpu(void)
+{
+ if (!power_driver)
+ return nr_cpu_ids;
+
+ return power_driver->best_wake_cpu();
+}
+
void power_late_callback(int cpu)
{
if (!power_driver)
diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index 88e7968..a368734 100644
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -1370,6 +1370,7 @@ extern void power_late_callback(int cpu);
extern int at_max_capacity(int cpu);
extern int go_faster(int cpu, int hint);
extern int go_slower(int cpu, int hint);
+extern int best_wake_cpu(void);
#else
static inline void power_late_callback(int cpu)
{
@@ -1389,4 +1390,9 @@ static inline int go_slower(int cpu, int hint)
{
return 0;
}
+
+static inline int best_wake_cpu(void)
+{
+ return nr_cpu_ids;
+}
#endif /* CONFIG_SCHED_POWER */
--
1.7.9.5
next prev parent reply other threads:[~2013-10-11 17:19 UTC|newest]
Thread overview: 29+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-10-11 17:19 [RFC][PATCH 0/7] Power-aware scheduling v2 Morten Rasmussen
2013-10-11 17:19 ` [RFC][PATCH 1/7] Initial power driver interface infrastructure Morten Rasmussen
2013-10-11 17:19 ` [RFC][PATCH 2/7] sched: power: Power driver late callback interface Morten Rasmussen
2013-10-14 13:42 ` Peter Zijlstra
2013-10-11 17:19 ` [RFC][PATCH 3/7] sched: power: go_faster/slower power driver hints Morten Rasmussen
2013-10-12 2:58 ` Michael wang
2013-10-14 12:42 ` Morten Rasmussen
2013-10-14 13:48 ` Peter Zijlstra
2013-10-14 15:55 ` Morten Rasmussen
2013-10-11 17:19 ` [RFC][PATCH 4/7] sched: power: Remove power capacity hints for kworker threads Morten Rasmussen
2013-10-14 13:33 ` Peter Zijlstra
2013-10-14 15:14 ` Arjan van de Ven
2013-10-17 16:40 ` Morten Rasmussen
2013-10-17 16:54 ` Peter Zijlstra
2013-10-17 17:18 ` Arjan van de Ven
2013-10-18 8:47 ` Morten Rasmussen
2013-10-18 13:43 ` Arjan van de Ven
2013-10-18 8:38 ` Morten Rasmussen
2013-10-14 16:10 ` Morten Rasmussen
2013-10-14 16:13 ` Arjan van de Ven
2013-10-14 17:19 ` Morten Rasmussen
2013-10-11 17:19 ` [RFC][PATCH 5/7] sched: power: Increase cpu capacity based on rq tracked load Morten Rasmussen
2013-10-11 17:19 ` [RFC][PATCH 6/7] sched: power: cpufreq: Initial schedpower cpufreq governor/power driver Morten Rasmussen
2013-10-11 17:19 ` Morten Rasmussen [this message]
2013-10-14 13:32 ` [RFC][PATCH 0/7] Power-aware scheduling v2 Peter Zijlstra
2013-10-14 17:15 ` Morten Rasmussen
2013-10-14 17:31 ` Peter Zijlstra
2013-10-15 17:05 ` Morten Rasmussen
2013-10-15 9:57 ` Preeti U Murthy
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=1381511957-29776-8-git-send-email-morten.rasmussen@arm.com \
--to=morten.rasmussen@arm.com \
--cc=alex.shi@linaro.org \
--cc=arjan@linux.intel.com \
--cc=catalin.marinas@arm.com \
--cc=corbet@lwn.net \
--cc=dirk.j.brandewie@intel.com \
--cc=efault@gmx.de \
--cc=linaro-kernel@lists.linaro.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@kernel.org \
--cc=peterz@infradead.org \
--cc=pjt@google.com \
--cc=preeti@linux.vnet.ibm.com \
--cc=rjw@sisk.pl \
--cc=tglx@linutronix.de \
--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
all inboxes | Powered by JetHome®