This patch introduces the SCHED_ENERGY_FREQ sched feature, which is implemented using jump labels when SCHED_DEBUG is defined. It is statically set to false when SCHED_DEBUG is not defined and thus disabled by default.

Signed-off-by: Michael Turquette mturquette@linaro.org --- kernel/sched/fair.c | 5 +++++ kernel/sched/features.h | 6 ++++++ 2 files changed, 11 insertions(+)

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 46855d0..75aec8d 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -4207,6 +4207,11 @@ static inline void hrtick_update(struct rq *rq) } #endif

+static inline bool sched_energy_freq(void) +{ + return sched_feat(SCHED_ENERGY_FREQ); +} + /* * The enqueue_task method is called before nr_running is * increased. Here we update the fair scheduling stats and diff --git a/kernel/sched/features.h b/kernel/sched/features.h index 91e33cd..77381cf 100644 --- a/kernel/sched/features.h +++ b/kernel/sched/features.h @@ -96,3 +96,9 @@ SCHED_FEAT(NUMA_FAVOUR_HIGHER, true) */ SCHED_FEAT(NUMA_RESIST_LOWER, false) #endif + +/* + * Scheduler-driven CPU frequency selection aimed to save energy based on + * load tracking + */ +SCHED_FEAT(SCHED_ENERGY_FREQ, false) -- 1.9.1

Scheduler-driven cpu frequency selection is desirable as part of the on-going effort to make the scheduler better aware of energy consumption. No piece of the Linux kernel has a better view of the factors that affect a cpu frequency selection policy than the scheduler[0], and this patch is an attempt to get that discussion going again.

This patch implements a cpufreq governor that directly accesses scheduler statistics, in particular the pelt data from cfs via the get_cpu_usage() function.

Put plainly, this governor selects the lowest cpu frequency that will prevent a runqueue from being over-utilized (until we hit the highest frequency of course). This is accomplished by requesting a frequency that matches the current capacity utilization, plus a margin.

Unlike the previous posting from 2014[1] this governor implements a "follow the utilization" method, where utilization is defined as the frequency-invariant product of cfs.utilization_load_avg and cpu_capacity_orig.

This governor is event-driven. There is no polling loop to check cpu idle time nor any other method which is unsynchronized with the scheduler. The entry points for this policy are in fair.c: enqueue_task_fair, dequeue_task_fair and task_tick_fair.

This policy is implemented using the cpufreq governor interface for two main reasons:

1) re-using the cpufreq machine drivers without using the governor interface is hard.

2) using the cpufreq interface allows us to switch between the scheduler-driven policy and legacy cpufreq governors such as ondemand at run-time. This is very useful for comparative testing and tuning.

Finally, it is worth mentioning that this approach neglects all scheduling classes except for cfs. It is possible to add support for deadline and other other classes here, but I also wonder if a multi-governor approach would be a more maintainable solution, where the cpufreq core aggregates the constraints set by multiple governors. Supporting such an approach in the cpufreq core would also allow for peripheral devices to place constraint on cpu frequency without having to hack such behavior in at the governor level.

Thanks to Juri Lelli juri.lelli@arm.com for doing a good bit of testing, bug fixing and contributing towards the design. I've included his sign-off with his permission after stealing lots of his code.

[0] http://article.gmane.org/gmane.linux.kernel/1499836 [1] https://lkml.org/lkml/2014/10/22/22

Signed-off-by: Michael Turquette mturquette@linaro.org Signed-off-by: Juri Lelli juri.lelli@arm.com --- For those that are very curious, there were recently two previous postings of these patches to the public eas-dev mailing list. Of interest in those threads is the discussion around using a utilization threadhold versus purely matching frequency to capacity utilization versus using a margin (as this series does):

https://lists.linaro.org/pipermail/eas-dev/2015-April/000074.html https://lists.linaro.org/pipermail/eas-dev/2015-April/000115.html

drivers/cpufreq/Kconfig | 24 ++++ include/linux/cpufreq.h | 3 + kernel/sched/Makefile | 1 + kernel/sched/cpufreq_cfs.c | 311 +++++++++++++++++++++++++++++++++++++++++++++ kernel/sched/fair.c | 11 ++ kernel/sched/sched.h | 6 + 6 files changed, 356 insertions(+) create mode 100644 kernel/sched/cpufreq_cfs.c

diff --git a/drivers/cpufreq/Kconfig b/drivers/cpufreq/Kconfig index a171fef..83d51b4 100644 --- a/drivers/cpufreq/Kconfig +++ b/drivers/cpufreq/Kconfig @@ -102,6 +102,15 @@ config CPU_FREQ_DEFAULT_GOV_CONSERVATIVE Be aware that not all cpufreq drivers support the conservative governor. If unsure have a look at the help section of the driver. Fallback governor will be the performance governor. + +config CPU_FREQ_DEFAULT_GOV_CFS + bool "cfs" + select CPU_FREQ_GOV_CFS + select CPU_FREQ_GOV_PERFORMANCE + help + Use the CPUfreq governor 'cfs' as default. This scales + cpu frequency from the scheduler as per-entity load tracking + statistics are updated. endchoice

config CPU_FREQ_GOV_PERFORMANCE @@ -183,6 +192,21 @@ config CPU_FREQ_GOV_CONSERVATIVE

If in doubt, say N.

+config CPU_FREQ_GOV_CFS + tristate "'cfs' cpufreq governor" + depends on CPU_FREQ + select CPU_FREQ_GOV_COMMON + help + 'cfs' - this governor scales cpu frequency from the + scheduler as a function of cpu capacity utilization. It does + not evaluate utilization on a periodic basis (as ondemand + does) but instead is invoked from the completely fair + scheduler when updating per-entity load tracking statistics. + Latency to respond to changes in load is improved over polling + governors due to its event-driven design. + + If in doubt, say N. + comment "CPU frequency scaling drivers"

config CPUFREQ_DT diff --git a/include/linux/cpufreq.h b/include/linux/cpufreq.h index 2ee4888..62e8152 100644 --- a/include/linux/cpufreq.h +++ b/include/linux/cpufreq.h @@ -485,6 +485,9 @@ extern struct cpufreq_governor cpufreq_gov_ondemand; #elif defined(CONFIG_CPU_FREQ_DEFAULT_GOV_CONSERVATIVE) extern struct cpufreq_governor cpufreq_gov_conservative; #define CPUFREQ_DEFAULT_GOVERNOR (&cpufreq_gov_conservative) +#elif defined(CONFIG_CPU_FREQ_DEFAULT_GOV_CAP_GOV) +extern struct cpufreq_governor cpufreq_gov_cap_gov; +#define CPUFREQ_DEFAULT_GOVERNOR (&cpufreq_gov_cap_gov) #endif

/********************************************************************* diff --git a/kernel/sched/Makefile b/kernel/sched/Makefile index 46be870..466960d 100644 --- a/kernel/sched/Makefile +++ b/kernel/sched/Makefile @@ -19,3 +19,4 @@ obj-$(CONFIG_SCHED_AUTOGROUP) += auto_group.o obj-$(CONFIG_SCHEDSTATS) += stats.o obj-$(CONFIG_SCHED_DEBUG) += debug.o obj-$(CONFIG_CGROUP_CPUACCT) += cpuacct.o +obj-$(CONFIG_CPU_FREQ_GOV_CFS) += cpufreq_cfs.o diff --git a/kernel/sched/cpufreq_cfs.c b/kernel/sched/cpufreq_cfs.c new file mode 100644 index 0000000..49edc81 --- /dev/null +++ b/kernel/sched/cpufreq_cfs.c @@ -0,0 +1,311 @@ +/* + * Copyright (C) 2015 Michael Turquette mturquette@linaro.org + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + */ + +#include <linux/cpufreq.h> +#include <linux/module.h> +#include <linux/kthread.h> +#include <linux/percpu.h> +#include <linux/irq_work.h> + +#include "sched.h" + +#define MARGIN_PCT 125 /* taken from imbalance_pct = 125 */ +#define THROTTLE_NSEC 50000000 /* 50ms default */ + +/** + * gov_data - per-policy data internal to the governor + * @throttle: next throttling period expiry. Derived from throttle_nsec + * @throttle_nsec: throttle period length in nanoseconds + * @task: worker thread for dvfs transition that may block/sleep + * @irq_work: callback used to wake up worker thread + * + * struct gov_data is the per-policy cpufreq_cfs-specific data structure. A + * per-policy instance of it is created when the cpufreq_cfs governor receives + * the CPUFREQ_GOV_START condition and a pointer to it exists in the gov_data + * member of struct cpufreq_policy. + * + * Readers of this data must call down_read(policy->rwsem). Writers must + * call down_write(policy->rwsem). + */ +struct gov_data { + ktime_t throttle; + unsigned int throttle_nsec; + struct task_struct *task; + struct irq_work irq_work; + struct cpufreq_policy *policy; +}; + +/** + * cpufreq_cfs_select_freq - pick the next frequency for a cpu + * @policy: the cpufreq policy whose frequency may be changed + * + * cpufreq_cfs_select_freq selects a frequency based on pelt load statistics + * tracked by cfs. First it finds the most utilized cpu in the policy and then + * maps that utilization value onto a cpu frequency and returns it. + * + * Additionally, cpufreq_cfs_select_freq adds a margin to the cpu utilization value + * before converting it to a frequency. The margin is derived from MARGIN_PCT, + * which itself is inspired by imbalance_pct in cfs. This is needed to + * proactively increase frequency in the case of increasing load. + * + * This approach attempts to maintain headroom of 25% unutilized cpu capacity. + * A traditional way of doing this is to take 75% of the current capacity and + * check if current utilization crosses that threshold. The only problem with + * that approach is determining the next cpu frequency target if that threshold + * is crossed. + * + * Instead of using the 75% threshold, cpufreq_cfs_select_freq adds a 25% + * utilization margin to the utilization and converts that to a frequency. This + * removes conditional logic around checking thresholds and better supports + * drivers that use non-discretized frequency ranges (i.e. no pre-defined + * frequency tables or operating points). + * + * Returns frequency selected. + */ +static unsigned long cpufreq_cfs_select_freq(struct cpufreq_policy *policy) +{ + int cpu = 0; + struct gov_data *gd; + unsigned max_usage = 0, usage = 0; + + if (!policy->governor_data) + return 0; + + gd = policy->governor_data; + + /* + * get_cpu_usage is called without locking the runqueues. This is the + * same behavior used by find_busiest_cpu in load_balance. We are + * willing to accept occasionally stale data here in exchange for + * lockless behavior. + */ + for_each_cpu(cpu, policy->cpus) { + usage = get_cpu_usage(cpu); + if (usage > max_usage) + max_usage = usage; + } + + /* add margin to max_usage based on imbalance_pct */ + max_usage = max_usage * MARGIN_PCT / 100; + + cpu = cpumask_first(policy->cpus); + + /* freq is current utilization + 25% */ + return max_usage * policy->max / capacity_orig_of(cpu); +} + +/* + * we pass in struct cpufreq_policy. This is safe because changing out the + * policy requires a call to __cpufreq_governor(policy, CPUFREQ_GOV_STOP), + * which tears down all of the data structures and __cpufreq_governor(policy, + * CPUFREQ_GOV_START) will do a full rebuild, including this kthread with the + * new policy pointer + */ +static int cpufreq_cfs_thread(void *data) +{ + struct sched_param param; + struct cpufreq_policy *policy; + struct gov_data *gd; + unsigned long freq; + int ret; + + policy = (struct cpufreq_policy *) data; + if (!policy) { + pr_warn("%s: missing policy\n", __func__); + do_exit(-EINVAL); + } + + gd = policy->governor_data; + if (!gd) { + pr_warn("%s: missing governor data\n", __func__); + do_exit(-EINVAL); + } + + param.sched_priority = 50; + ret = sched_setscheduler_nocheck(gd->task, SCHED_FIFO, &param); + if (ret) { + pr_warn("%s: failed to set SCHED_FIFO\n", __func__); + do_exit(-EINVAL); + } else { + pr_debug("%s: kthread (%d) set to SCHED_FIFO\n", + __func__, gd->task->pid); + } + + ret = set_cpus_allowed_ptr(gd->task, policy->related_cpus); + if (ret) { + pr_warn("%s: failed to set allowed ptr\n", __func__); + do_exit(-EINVAL); + } + + /* main loop of the per-policy kthread */ + do { + set_current_state(TASK_INTERRUPTIBLE); + schedule(); + if (kthread_should_stop()) + break; + + /* avoid race with cpufreq_cfs_stop */ + if (!down_write_trylock(&policy->rwsem)) + continue; + + freq = cpufreq_cfs_select_freq(policy); + + ret = __cpufreq_driver_target(policy, freq, + CPUFREQ_RELATION_L); + if (ret) + pr_debug("%s: __cpufreq_driver_target returned %d\n", + __func__, ret); + + gd->throttle = ktime_add_ns(ktime_get(), gd->throttle_nsec); + up_write(&policy->rwsem); + } while (!kthread_should_stop()); + + do_exit(0); +} + +static void cpufreq_cfs_irq_work(struct irq_work *irq_work) +{ + struct gov_data *gd; + + gd = container_of(irq_work, struct gov_data, irq_work); + if (!gd) { + return; + } + + wake_up_process(gd->task); +} + +/** + * cpufreq_cfs_update_cpu - interface to scheduler for changing capacity values + * @cpu: cpu whose capacity utilization has recently changed + * + * cpufreq_cfs_udpate_cpu is an interface exposed to the scheduler so that the + * scheduler may inform the governor of updates to capacity utilization and + * make changes to cpu frequency. Currently this interface is designed around + * PELT values in CFS. It can be expanded to other scheduling classes in the + * future if needed. + * + * cpufreq_cfs_update_cpu raises an IPI. The irq_work handler for that IPI wakes up + * the thread that does the actual work, cpufreq_cfs_thread. + */ +void cpufreq_cfs_update_cpu(int cpu) +{ + struct cpufreq_policy *policy; + struct gov_data *gd; + + /* XXX put policy pointer in per-cpu data? */ + policy = cpufreq_cpu_get(cpu); + if (IS_ERR_OR_NULL(policy)) { + return; + } + + if (!policy->governor_data) { + goto out; + } + + gd = policy->governor_data; + + /* bail early if we are throttled */ + if (ktime_before(ktime_get(), gd->throttle)) { + goto out; + } + + irq_work_queue_on(&gd->irq_work, cpu); + +out: + cpufreq_cpu_put(policy); + return; +} + +static void cpufreq_cfs_start(struct cpufreq_policy *policy) +{ + struct gov_data *gd; + + /* prepare per-policy private data */ + gd = kzalloc(sizeof(*gd), GFP_KERNEL); + if (!gd) { + pr_debug("%s: failed to allocate private data\n", __func__); + return; + } + + /* + * Don't ask for freq changes at an higher rate than what + * the driver advertises as transition latency. + */ + gd->throttle_nsec = policy->cpuinfo.transition_latency ? + policy->cpuinfo.transition_latency : + THROTTLE_NSEC; + pr_debug("%s: throttle threshold = %u [ns]\n", + __func__, gd->throttle_nsec); + + /* init per-policy kthread */ + gd->task = kthread_run(cpufreq_cfs_thread, policy, "kcpufreq_cfs_task"); + if (IS_ERR_OR_NULL(gd->task)) + pr_err("%s: failed to create kcpufreq_cfs_task thread\n", __func__); + + init_irq_work(&gd->irq_work, cpufreq_cfs_irq_work); + policy->governor_data = gd; + gd->policy = policy; +} + +static void cpufreq_cfs_stop(struct cpufreq_policy *policy) +{ + struct gov_data *gd; + + gd = policy->governor_data; + kthread_stop(gd->task); + + policy->governor_data = NULL; + + /* FIXME replace with devm counterparts? */ + kfree(gd); +} + +static int cpufreq_cfs_setup(struct cpufreq_policy *policy, unsigned int event) +{ + switch (event) { + case CPUFREQ_GOV_START: + /* Start managing the frequency */ + cpufreq_cfs_start(policy); + return 0; + + case CPUFREQ_GOV_STOP: + cpufreq_cfs_stop(policy); + return 0; + + case CPUFREQ_GOV_LIMITS: /* unused */ + case CPUFREQ_GOV_POLICY_INIT: /* unused */ + case CPUFREQ_GOV_POLICY_EXIT: /* unused */ + break; + } + return 0; +} + +#ifndef CONFIG_CPU_FREQ_DEFAULT_GOV_SCHED_CFS +static +#endif +struct cpufreq_governor cpufreq_cfs = { + .name = "cfs", + .governor = cpufreq_cfs_setup, + .owner = THIS_MODULE, +}; + +static int __init cpufreq_cfs_init(void) +{ + return cpufreq_register_governor(&cpufreq_cfs); +} + +static void __exit cpufreq_cfs_exit(void) +{ + cpufreq_unregister_governor(&cpufreq_cfs); +} + +/* Try to make this the default governor */ +fs_initcall(cpufreq_cfs_init); + +MODULE_LICENSE("GPL"); diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 9e37d49..3cf3024 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -4257,6 +4257,10 @@ enqueue_task_fair(struct rq *rq, struct task_struct *p, int flags) update_rq_runnable_avg(rq, rq->nr_running); add_nr_running(rq, 1); } + + if(sched_energy_freq()) + cpufreq_cfs_update_cpu(cpu_of(rq)); + hrtick_update(rq); }

@@ -4318,6 +4322,10 @@ static void dequeue_task_fair(struct rq *rq, struct task_struct *p, int flags) sub_nr_running(rq, 1); update_rq_runnable_avg(rq, 1); } + + if(sched_energy_freq()) + cpufreq_cfs_update_cpu(cpu_of(rq)); + hrtick_update(rq); }

@@ -7789,6 +7797,9 @@ static void task_tick_fair(struct rq *rq, struct task_struct *curr, int queued) task_tick_numa(rq, curr);

update_rq_runnable_avg(rq, 1); + + if(sched_energy_freq()) + cpufreq_cfs_update_cpu(cpu_of(rq)); }

/* diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index 8bf35d3..457d98b 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -1429,6 +1429,12 @@ static inline int get_cpu_usage(int cpu) return (usage * capacity) >> SCHED_LOAD_SHIFT; }

+#ifdef CONFIG_CPU_FREQ_GOV_CFS +void cpufreq_cfs_update_cpu(int cpu); +#else +static inline void cpufreq_cfs_update_cpu(int cpu) {} +#endif + static inline void sched_rt_avg_update(struct rq *rq, u64 rt_delta) { rq->rt_avg += rt_delta * arch_scale_freq_capacity(NULL, cpu_of(rq)); -- 1.9.1