From: Morten Rasmussen <morten.rasmussen@arm.com>
To: peterz@infradead.org, mingo@redhat.com
Cc: dietmar.eggemann@arm.com, pjt@google.com, bsegall@google.com,
vincent.guittot@linaro.org, nicolas.pitre@linaro.org,
mturquette@linaro.org, rjw@rjwysocki.net,
linux-kernel@vger.kernel.org,
Morten Rasmussen <morten.rasmussen@arm.com>
Subject: [PATCH 5/7] sched: Implement usage tracking
Date: Mon, 22 Sep 2014 17:24:05 +0100 [thread overview]
Message-ID: <1411403047-32010-6-git-send-email-morten.rasmussen@arm.com> (raw)
In-Reply-To: <1411403047-32010-1-git-send-email-morten.rasmussen@arm.com>
With the framework for runnable tracking now fully in place, per-entity
usage tracking is a simple and low-overhead addition.
This is a rebased and significantly cut down version of a patch
originally authored by Paul Turner <pjt@google.com>.
cc: Paul Turner <pjt@google.com>
cc: Ben Segall <bsegall@google.com>
Signed-off-by: Morten Rasmussen <morten.rasmussen@arm.com>
---
include/linux/sched.h | 1 +
kernel/sched/debug.c | 1 +
kernel/sched/fair.c | 16 +++++++++++++---
3 files changed, 15 insertions(+), 3 deletions(-)
diff --git a/include/linux/sched.h b/include/linux/sched.h
index 18f5262..0bcd8a7 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -1080,6 +1080,7 @@ struct sched_avg {
u64 last_runnable_update;
s64 decay_count;
unsigned long load_avg_contrib;
+ u32 usage_avg_sum;
};
#ifdef CONFIG_SCHEDSTATS
diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c
index c7fe1ea0..ed5a9ce 100644
--- a/kernel/sched/debug.c
+++ b/kernel/sched/debug.c
@@ -95,6 +95,7 @@ static void print_cfs_group_stats(struct seq_file *m, int cpu, struct task_group
#ifdef CONFIG_SMP
P(se->avg.runnable_avg_sum);
P(se->avg.runnable_avg_period);
+ P(se->avg.usage_avg_sum);
P(se->avg.load_avg_contrib);
P(se->avg.decay_count);
#endif
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 52abb3e..d8a8c83 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -2299,7 +2299,8 @@ unsigned long arch_scale_load_capacity(int cpu);
*/
static __always_inline int __update_entity_runnable_avg(u64 now, int cpu,
struct sched_avg *sa,
- int runnable)
+ int runnable,
+ int running)
{
u64 delta, periods;
u32 runnable_contrib;
@@ -2341,6 +2342,8 @@ static __always_inline int __update_entity_runnable_avg(u64 now, int cpu,
if (runnable)
sa->runnable_avg_sum += (delta_w * scale_cap)
>> SCHED_CAPACITY_SHIFT;
+ if (running)
+ sa->usage_avg_sum += delta_w;
sa->runnable_avg_period += delta_w;
delta -= delta_w;
@@ -2353,6 +2356,7 @@ static __always_inline int __update_entity_runnable_avg(u64 now, int cpu,
periods + 1);
sa->runnable_avg_period = decay_load(sa->runnable_avg_period,
periods + 1);
+ sa->usage_avg_sum = decay_load(sa->usage_avg_sum, periods + 1);
/* Efficiently calculate \sum (1..n_period) 1024*y^i */
runnable_contrib = __compute_runnable_contrib(periods);
@@ -2360,6 +2364,8 @@ static __always_inline int __update_entity_runnable_avg(u64 now, int cpu,
if (runnable)
sa->runnable_avg_sum += (runnable_contrib * scale_cap)
>> SCHED_CAPACITY_SHIFT;
+ if (running)
+ sa->usage_avg_sum += runnable_contrib;
sa->runnable_avg_period += runnable_contrib;
}
@@ -2367,6 +2373,8 @@ static __always_inline int __update_entity_runnable_avg(u64 now, int cpu,
if (runnable)
sa->runnable_avg_sum += (delta * scale_cap)
>> SCHED_CAPACITY_SHIFT;
+ if (running)
+ sa->usage_avg_sum += delta;
sa->runnable_avg_period += delta;
return decayed;
@@ -2473,7 +2481,7 @@ static inline void __update_group_entity_contrib(struct sched_entity *se)
static inline void update_rq_runnable_avg(struct rq *rq, int runnable)
{
__update_entity_runnable_avg(rq_clock_task(rq), rq->cpu, &rq->avg,
- runnable);
+ runnable, runnable);
__update_tg_runnable_avg(&rq->avg, &rq->cfs);
}
#else /* CONFIG_FAIR_GROUP_SCHED */
@@ -2539,7 +2547,8 @@ static inline void update_entity_load_avg(struct sched_entity *se,
else
now = cfs_rq_clock_task(group_cfs_rq(se));
- if (!__update_entity_runnable_avg(now, cpu, &se->avg, se->on_rq))
+ if (!__update_entity_runnable_avg(now, cpu, &se->avg, se->on_rq,
+ cfs_rq->curr == se))
return;
contrib_delta = __update_entity_load_avg_contrib(se);
@@ -2980,6 +2989,7 @@ set_next_entity(struct cfs_rq *cfs_rq, struct sched_entity *se)
*/
update_stats_wait_end(cfs_rq, se);
__dequeue_entity(cfs_rq, se);
+ update_entity_load_avg(se, 1);
}
update_stats_curr_start(cfs_rq, se);
--
1.7.9.5
next prev parent reply other threads:[~2014-09-22 16:24 UTC|newest]
Thread overview: 40+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-09-22 16:24 [PATCH 0/7] sched: Scale-invariant per-entity load-tracking Morten Rasmussen
2014-09-22 16:24 ` [PATCH 1/7] sched: Introduce scale-invariant load tracking Morten Rasmussen
2014-09-25 13:48 ` Vincent Guittot
2014-09-25 17:23 ` Morten Rasmussen
2014-09-26 7:36 ` Vincent Guittot
2014-09-26 9:38 ` Morten Rasmussen
2014-10-02 20:34 ` Peter Zijlstra
2014-10-08 11:00 ` Morten Rasmussen
2014-10-08 11:21 ` Vincent Guittot
2014-10-08 13:53 ` Morten Rasmussen
2014-10-08 14:08 ` Vincent Guittot
2014-10-08 14:16 ` Morten Rasmussen
2014-10-08 11:38 ` Vincent Guittot
2014-10-08 14:05 ` Morten Rasmussen
2014-10-10 9:07 ` Peter Zijlstra
2014-10-08 0:50 ` Yuyang Du
2014-10-08 12:54 ` Dietmar Eggemann
2014-10-10 9:16 ` Peter Zijlstra
2014-10-10 9:14 ` Peter Zijlstra
2014-09-22 16:24 ` [PATCH 2/7] cpufreq: Architecture specific callback for frequency changes Morten Rasmussen
2014-10-08 6:07 ` Mike Turquette
2014-10-08 6:26 ` [PATCH RFC 0/2] introduce capacity_ops to CFS Mike Turquette
2014-10-08 6:26 ` [PATCH RFC 1/2] sched: cfs: introduce capacity_ops Mike Turquette
2014-10-08 8:37 ` Peter Zijlstra
[not found] ` <20141008232836.4379.3339@quantum>
2014-10-09 9:00 ` Peter Zijlstra
[not found] ` <20141009173433.4379.58492@quantum>
2014-10-09 19:00 ` Peter Zijlstra
2014-10-08 6:26 ` [PATCH RFC 2/2] cpufreq: arm_big_little: provide cpu capacity Mike Turquette
2014-10-08 15:48 ` Morten Rasmussen
[not found] ` <20141008223732.4379.78047@quantum>
2014-10-09 9:02 ` Peter Zijlstra
[not found] ` <20141009172513.4379.56718@quantum>
2014-10-09 17:38 ` Peter Zijlstra
2014-09-22 16:24 ` [PATCH 3/7] arm: Frequency invariant scheduler load-tracking support Morten Rasmussen
2014-09-22 16:24 ` [PATCH 4/7] arm: Micro-architecture invariant load tracking support Morten Rasmussen
2014-09-22 16:24 ` Morten Rasmussen [this message]
2014-09-22 16:24 ` [PATCH 6/7] sched: Make sched entity usage tracking scale-invariant Morten Rasmussen
2014-09-22 17:13 ` bsegall
2014-09-23 13:35 ` Morten Rasmussen
2014-10-02 21:04 ` Peter Zijlstra
2014-09-22 16:24 ` [PATCH 7/7] sched: Track sched_entity usage contributions Morten Rasmussen
2014-09-22 17:09 ` bsegall
2014-09-23 13:59 ` Morten Rasmussen
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=1411403047-32010-6-git-send-email-morten.rasmussen@arm.com \
--to=morten.rasmussen@arm.com \
--cc=bsegall@google.com \
--cc=dietmar.eggemann@arm.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@redhat.com \
--cc=mturquette@linaro.org \
--cc=nicolas.pitre@linaro.org \
--cc=peterz@infradead.org \
--cc=pjt@google.com \
--cc=rjw@rjwysocki.net \
--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®