#include <linux/jiffies.h>
#include <linux/kernel_stat.h>
#include <linux/percpu.h>
+#include <linux/mutex.h>
/*
* dbs is used in this file as a shortform for demandbased switching
*/
#define DEF_FREQUENCY_UP_THRESHOLD (80)
-#define MIN_FREQUENCY_UP_THRESHOLD (0)
+#define MIN_FREQUENCY_UP_THRESHOLD (11)
#define MAX_FREQUENCY_UP_THRESHOLD (100)
-#define DEF_FREQUENCY_DOWN_THRESHOLD (20)
-#define MIN_FREQUENCY_DOWN_THRESHOLD (0)
-#define MAX_FREQUENCY_DOWN_THRESHOLD (100)
-
-/*
- * The polling frequency of this governor depends on the capability of
+/*
+ * The polling frequency of this governor depends on the capability of
* the processor. Default polling frequency is 1000 times the transition
- * latency of the processor. The governor will work on any processor with
- * transition latency <= 10mS, using appropriate sampling
+ * latency of the processor. The governor will work on any processor with
+ * transition latency <= 10mS, using appropriate sampling
* rate.
* For CPUs with transition latency > 10mS (mostly drivers with CPUFREQ_ETERNAL)
* this governor will not work.
* All times here are in uS.
*/
-static unsigned int def_sampling_rate;
-#define MIN_SAMPLING_RATE (def_sampling_rate / 2)
+static unsigned int def_sampling_rate;
+#define MIN_SAMPLING_RATE_RATIO (2)
+/* for correct statistics, we need at least 10 ticks between each measure */
+#define MIN_STAT_SAMPLING_RATE (MIN_SAMPLING_RATE_RATIO * jiffies_to_usecs(10))
+#define MIN_SAMPLING_RATE (def_sampling_rate / MIN_SAMPLING_RATE_RATIO)
#define MAX_SAMPLING_RATE (500 * def_sampling_rate)
#define DEF_SAMPLING_RATE_LATENCY_MULTIPLIER (1000)
-#define DEF_SAMPLING_DOWN_FACTOR (10)
+#define DEF_SAMPLING_DOWN_FACTOR (1)
+#define MAX_SAMPLING_DOWN_FACTOR (10)
#define TRANSITION_LATENCY_LIMIT (10 * 1000)
-#define sampling_rate_in_HZ(x) (((x * HZ) < (1000 * 1000))?1:((x * HZ) / (1000 * 1000)))
static void do_dbs_timer(void *data);
struct cpu_dbs_info_s {
- struct cpufreq_policy *cur_policy;
- unsigned int prev_cpu_idle_up;
- unsigned int prev_cpu_idle_down;
- unsigned int enable;
+ struct cpufreq_policy *cur_policy;
+ unsigned int prev_cpu_idle_up;
+ unsigned int prev_cpu_idle_down;
+ unsigned int enable;
};
static DEFINE_PER_CPU(struct cpu_dbs_info_s, cpu_dbs_info);
static unsigned int dbs_enable; /* number of CPUs using this policy */
-static DECLARE_MUTEX (dbs_sem);
+static DEFINE_MUTEX (dbs_mutex);
static DECLARE_WORK (dbs_work, do_dbs_timer, NULL);
+static struct workqueue_struct *dbs_workq;
+
struct dbs_tuners {
- unsigned int sampling_rate;
- unsigned int sampling_down_factor;
- unsigned int up_threshold;
- unsigned int down_threshold;
+ unsigned int sampling_rate;
+ unsigned int sampling_down_factor;
+ unsigned int up_threshold;
+ unsigned int ignore_nice;
};
static struct dbs_tuners dbs_tuners_ins = {
- .up_threshold = DEF_FREQUENCY_UP_THRESHOLD,
- .down_threshold = DEF_FREQUENCY_DOWN_THRESHOLD,
- .sampling_down_factor = DEF_SAMPLING_DOWN_FACTOR,
+ .up_threshold = DEF_FREQUENCY_UP_THRESHOLD,
+ .sampling_down_factor = DEF_SAMPLING_DOWN_FACTOR,
+ .ignore_nice = 0,
};
+static inline unsigned int get_cpu_idle_time(unsigned int cpu)
+{
+ return kstat_cpu(cpu).cpustat.idle +
+ kstat_cpu(cpu).cpustat.iowait +
+ ( dbs_tuners_ins.ignore_nice ?
+ kstat_cpu(cpu).cpustat.nice :
+ 0);
+}
+
/************************** sysfs interface ************************/
static ssize_t show_sampling_rate_max(struct cpufreq_policy *policy, char *buf)
{
return sprintf (buf, "%u\n", MIN_SAMPLING_RATE);
}
-#define define_one_ro(_name) \
-static struct freq_attr _name = \
+#define define_one_ro(_name) \
+static struct freq_attr _name = \
__ATTR(_name, 0444, show_##_name, NULL)
define_one_ro(sampling_rate_max);
show_one(sampling_rate, sampling_rate);
show_one(sampling_down_factor, sampling_down_factor);
show_one(up_threshold, up_threshold);
-show_one(down_threshold, down_threshold);
+show_one(ignore_nice_load, ignore_nice);
-static ssize_t store_sampling_down_factor(struct cpufreq_policy *unused,
+static ssize_t store_sampling_down_factor(struct cpufreq_policy *unused,
const char *buf, size_t count)
{
unsigned int input;
if (ret != 1 )
return -EINVAL;
- down(&dbs_sem);
+ if (input > MAX_SAMPLING_DOWN_FACTOR || input < 1)
+ return -EINVAL;
+
+ mutex_lock(&dbs_mutex);
dbs_tuners_ins.sampling_down_factor = input;
- up(&dbs_sem);
+ mutex_unlock(&dbs_mutex);
return count;
}
-static ssize_t store_sampling_rate(struct cpufreq_policy *unused,
+static ssize_t store_sampling_rate(struct cpufreq_policy *unused,
const char *buf, size_t count)
{
unsigned int input;
int ret;
ret = sscanf (buf, "%u", &input);
- down(&dbs_sem);
+ mutex_lock(&dbs_mutex);
if (ret != 1 || input > MAX_SAMPLING_RATE || input < MIN_SAMPLING_RATE) {
- up(&dbs_sem);
+ mutex_unlock(&dbs_mutex);
return -EINVAL;
}
dbs_tuners_ins.sampling_rate = input;
- up(&dbs_sem);
+ mutex_unlock(&dbs_mutex);
return count;
}
-static ssize_t store_up_threshold(struct cpufreq_policy *unused,
+static ssize_t store_up_threshold(struct cpufreq_policy *unused,
const char *buf, size_t count)
{
unsigned int input;
int ret;
ret = sscanf (buf, "%u", &input);
- down(&dbs_sem);
- if (ret != 1 || input > MAX_FREQUENCY_UP_THRESHOLD ||
- input < MIN_FREQUENCY_UP_THRESHOLD ||
- input <= dbs_tuners_ins.down_threshold) {
- up(&dbs_sem);
+ mutex_lock(&dbs_mutex);
+ if (ret != 1 || input > MAX_FREQUENCY_UP_THRESHOLD ||
+ input < MIN_FREQUENCY_UP_THRESHOLD) {
+ mutex_unlock(&dbs_mutex);
return -EINVAL;
}
dbs_tuners_ins.up_threshold = input;
- up(&dbs_sem);
+ mutex_unlock(&dbs_mutex);
return count;
}
-static ssize_t store_down_threshold(struct cpufreq_policy *unused,
+static ssize_t store_ignore_nice_load(struct cpufreq_policy *policy,
const char *buf, size_t count)
{
unsigned int input;
int ret;
- ret = sscanf (buf, "%u", &input);
- down(&dbs_sem);
- if (ret != 1 || input > MAX_FREQUENCY_DOWN_THRESHOLD ||
- input < MIN_FREQUENCY_DOWN_THRESHOLD ||
- input >= dbs_tuners_ins.up_threshold) {
- up(&dbs_sem);
+ unsigned int j;
+
+ ret = sscanf (buf, "%u", &input);
+ if ( ret != 1 )
return -EINVAL;
+
+ if ( input > 1 )
+ input = 1;
+
+ mutex_lock(&dbs_mutex);
+ if ( input == dbs_tuners_ins.ignore_nice ) { /* nothing to do */
+ mutex_unlock(&dbs_mutex);
+ return count;
}
+ dbs_tuners_ins.ignore_nice = input;
- dbs_tuners_ins.down_threshold = input;
- up(&dbs_sem);
+ /* we need to re-evaluate prev_cpu_idle_up and prev_cpu_idle_down */
+ for_each_online_cpu(j) {
+ struct cpu_dbs_info_s *j_dbs_info;
+ j_dbs_info = &per_cpu(cpu_dbs_info, j);
+ j_dbs_info->prev_cpu_idle_up = get_cpu_idle_time(j);
+ j_dbs_info->prev_cpu_idle_down = j_dbs_info->prev_cpu_idle_up;
+ }
+ mutex_unlock(&dbs_mutex);
return count;
}
define_one_rw(sampling_rate);
define_one_rw(sampling_down_factor);
define_one_rw(up_threshold);
-define_one_rw(down_threshold);
+define_one_rw(ignore_nice_load);
static struct attribute * dbs_attributes[] = {
&sampling_rate_max.attr,
&sampling_rate.attr,
&sampling_down_factor.attr,
&up_threshold.attr,
- &down_threshold.attr,
+ &ignore_nice_load.attr,
NULL
};
static void dbs_check_cpu(int cpu)
{
- unsigned int idle_ticks, up_idle_ticks, down_idle_ticks;
- unsigned int total_idle_ticks;
- unsigned int freq_down_step;
+ unsigned int idle_ticks, up_idle_ticks, total_ticks;
+ unsigned int freq_next;
unsigned int freq_down_sampling_rate;
static int down_skip[NR_CPUS];
struct cpu_dbs_info_s *this_dbs_info;
return;
policy = this_dbs_info->cur_policy;
- /*
- * The default safe range is 20% to 80%
- * Every sampling_rate, we check
- * - If current idle time is less than 20%, then we try to
- * increase frequency
- * Every sampling_rate*sampling_down_factor, we check
- * - If current idle time is more than 80%, then we try to
- * decrease frequency
+ /*
+ * Every sampling_rate, we check, if current idle time is less
+ * than 20% (default), then we try to increase frequency
+ * Every sampling_rate*sampling_down_factor, we look for a the lowest
+ * frequency which can sustain the load while keeping idle time over
+ * 30%. If such a frequency exist, we try to decrease to this frequency.
*
- * Any frequency increase takes it to the maximum frequency.
- * Frequency reduction happens at minimum steps of
- * 5% of max_frequency
+ * Any frequency increase takes it to the maximum frequency.
+ * Frequency reduction happens at minimum steps of
+ * 5% (default) of current frequency
*/
/* Check for frequency increase */
- total_idle_ticks = kstat_cpu(cpu).cpustat.idle +
- kstat_cpu(cpu).cpustat.iowait;
- idle_ticks = total_idle_ticks -
- this_dbs_info->prev_cpu_idle_up;
- this_dbs_info->prev_cpu_idle_up = total_idle_ticks;
-
-
+ idle_ticks = UINT_MAX;
for_each_cpu_mask(j, policy->cpus) {
- unsigned int tmp_idle_ticks;
+ unsigned int tmp_idle_ticks, total_idle_ticks;
struct cpu_dbs_info_s *j_dbs_info;
- if (j == cpu)
- continue;
-
j_dbs_info = &per_cpu(cpu_dbs_info, j);
- /* Check for frequency increase */
- total_idle_ticks = kstat_cpu(j).cpustat.idle +
- kstat_cpu(j).cpustat.iowait;
+ total_idle_ticks = get_cpu_idle_time(j);
tmp_idle_ticks = total_idle_ticks -
j_dbs_info->prev_cpu_idle_up;
j_dbs_info->prev_cpu_idle_up = total_idle_ticks;
/* Scale idle ticks by 100 and compare with up and down ticks */
idle_ticks *= 100;
up_idle_ticks = (100 - dbs_tuners_ins.up_threshold) *
- sampling_rate_in_HZ(dbs_tuners_ins.sampling_rate);
+ usecs_to_jiffies(dbs_tuners_ins.sampling_rate);
if (idle_ticks < up_idle_ticks) {
- __cpufreq_driver_target(policy, policy->max,
- CPUFREQ_RELATION_H);
down_skip[cpu] = 0;
- this_dbs_info->prev_cpu_idle_down = total_idle_ticks;
+ for_each_cpu_mask(j, policy->cpus) {
+ struct cpu_dbs_info_s *j_dbs_info;
+
+ j_dbs_info = &per_cpu(cpu_dbs_info, j);
+ j_dbs_info->prev_cpu_idle_down =
+ j_dbs_info->prev_cpu_idle_up;
+ }
+ /* if we are already at full speed then break out early */
+ if (policy->cur == policy->max)
+ return;
+
+ __cpufreq_driver_target(policy, policy->max,
+ CPUFREQ_RELATION_H);
return;
}
if (down_skip[cpu] < dbs_tuners_ins.sampling_down_factor)
return;
- total_idle_ticks = kstat_cpu(cpu).cpustat.idle +
- kstat_cpu(cpu).cpustat.iowait;
- idle_ticks = total_idle_ticks -
- this_dbs_info->prev_cpu_idle_down;
- this_dbs_info->prev_cpu_idle_down = total_idle_ticks;
-
+ idle_ticks = UINT_MAX;
for_each_cpu_mask(j, policy->cpus) {
- unsigned int tmp_idle_ticks;
+ unsigned int tmp_idle_ticks, total_idle_ticks;
struct cpu_dbs_info_s *j_dbs_info;
- if (j == cpu)
- continue;
-
j_dbs_info = &per_cpu(cpu_dbs_info, j);
- /* Check for frequency increase */
- total_idle_ticks = kstat_cpu(j).cpustat.idle +
- kstat_cpu(j).cpustat.iowait;
+ /* Check for frequency decrease */
+ total_idle_ticks = j_dbs_info->prev_cpu_idle_up;
tmp_idle_ticks = total_idle_ticks -
j_dbs_info->prev_cpu_idle_down;
j_dbs_info->prev_cpu_idle_down = total_idle_ticks;
idle_ticks = tmp_idle_ticks;
}
- /* Scale idle ticks by 100 and compare with up and down ticks */
- idle_ticks *= 100;
down_skip[cpu] = 0;
+ /* if we cannot reduce the frequency anymore, break out early */
+ if (policy->cur == policy->min)
+ return;
+ /* Compute how many ticks there are between two measurements */
freq_down_sampling_rate = dbs_tuners_ins.sampling_rate *
dbs_tuners_ins.sampling_down_factor;
- down_idle_ticks = (100 - dbs_tuners_ins.down_threshold) *
- sampling_rate_in_HZ(freq_down_sampling_rate);
+ total_ticks = usecs_to_jiffies(freq_down_sampling_rate);
- if (idle_ticks > down_idle_ticks ) {
- freq_down_step = (5 * policy->max) / 100;
+ /*
+ * The optimal frequency is the frequency that is the lowest that
+ * can support the current CPU usage without triggering the up
+ * policy. To be safe, we focus 10 points under the threshold.
+ */
+ freq_next = ((total_ticks - idle_ticks) * 100) / total_ticks;
+ freq_next = (freq_next * policy->cur) /
+ (dbs_tuners_ins.up_threshold - 10);
- /* max freq cannot be less than 100. But who knows.... */
- if (unlikely(freq_down_step == 0))
- freq_down_step = 5;
+ if (freq_next < policy->min)
+ freq_next = policy->min;
- __cpufreq_driver_target(policy,
- policy->cur - freq_down_step,
- CPUFREQ_RELATION_H);
- return;
- }
+ if (freq_next <= ((policy->cur * 95) / 100))
+ __cpufreq_driver_target(policy, freq_next, CPUFREQ_RELATION_L);
}
static void do_dbs_timer(void *data)
-{
+{
int i;
- down(&dbs_sem);
- for (i = 0; i < NR_CPUS; i++)
- if (cpu_online(i))
- dbs_check_cpu(i);
- schedule_delayed_work(&dbs_work,
- sampling_rate_in_HZ(dbs_tuners_ins.sampling_rate));
- up(&dbs_sem);
-}
+ mutex_lock(&dbs_mutex);
+ for_each_online_cpu(i)
+ dbs_check_cpu(i);
+ queue_delayed_work(dbs_workq, &dbs_work,
+ usecs_to_jiffies(dbs_tuners_ins.sampling_rate));
+ mutex_unlock(&dbs_mutex);
+}
static inline void dbs_timer_init(void)
{
INIT_WORK(&dbs_work, do_dbs_timer, NULL);
- schedule_delayed_work(&dbs_work,
- sampling_rate_in_HZ(dbs_tuners_ins.sampling_rate));
+ if (!dbs_workq)
+ dbs_workq = create_singlethread_workqueue("ondemand");
+ if (!dbs_workq) {
+ printk(KERN_ERR "ondemand: Cannot initialize kernel thread\n");
+ return;
+ }
+ queue_delayed_work(dbs_workq, &dbs_work,
+ usecs_to_jiffies(dbs_tuners_ins.sampling_rate));
return;
}
static inline void dbs_timer_exit(void)
{
- cancel_delayed_work(&dbs_work);
- return;
+ if (dbs_workq)
+ cancel_rearming_delayed_workqueue(dbs_workq, &dbs_work);
}
static int cpufreq_governor_dbs(struct cpufreq_policy *policy,
switch (event) {
case CPUFREQ_GOV_START:
- if ((!cpu_online(cpu)) ||
+ if ((!cpu_online(cpu)) ||
(!policy->cur))
return -EINVAL;
if (policy->cpuinfo.transition_latency >
- (TRANSITION_LATENCY_LIMIT * 1000))
+ (TRANSITION_LATENCY_LIMIT * 1000)) {
+ printk(KERN_WARNING "ondemand governor failed to load "
+ "due to too long transition latency\n");
return -EINVAL;
+ }
if (this_dbs_info->enable) /* Already enabled */
break;
-
- down(&dbs_sem);
+
+ mutex_lock(&dbs_mutex);
for_each_cpu_mask(j, policy->cpus) {
struct cpu_dbs_info_s *j_dbs_info;
j_dbs_info = &per_cpu(cpu_dbs_info, j);
j_dbs_info->cur_policy = policy;
-
- j_dbs_info->prev_cpu_idle_up =
- kstat_cpu(j).cpustat.idle +
- kstat_cpu(j).cpustat.iowait;
- j_dbs_info->prev_cpu_idle_down =
- kstat_cpu(j).cpustat.idle +
- kstat_cpu(j).cpustat.iowait;
+
+ j_dbs_info->prev_cpu_idle_up = get_cpu_idle_time(j);
+ j_dbs_info->prev_cpu_idle_down
+ = j_dbs_info->prev_cpu_idle_up;
}
this_dbs_info->enable = 1;
sysfs_create_group(&policy->kobj, &dbs_attr_group);
if (dbs_enable == 1) {
unsigned int latency;
/* policy latency is in nS. Convert it to uS first */
+ latency = policy->cpuinfo.transition_latency / 1000;
+ if (latency == 0)
+ latency = 1;
- latency = policy->cpuinfo.transition_latency;
- if (latency < 1000)
- latency = 1000;
-
- def_sampling_rate = (latency / 1000) *
+ def_sampling_rate = latency *
DEF_SAMPLING_RATE_LATENCY_MULTIPLIER;
- dbs_tuners_ins.sampling_rate = def_sampling_rate;
+ if (def_sampling_rate < MIN_STAT_SAMPLING_RATE)
+ def_sampling_rate = MIN_STAT_SAMPLING_RATE;
+
+ dbs_tuners_ins.sampling_rate = def_sampling_rate;
dbs_timer_init();
}
-
- up(&dbs_sem);
+
+ mutex_unlock(&dbs_mutex);
break;
case CPUFREQ_GOV_STOP:
- down(&dbs_sem);
+ mutex_lock(&dbs_mutex);
this_dbs_info->enable = 0;
sysfs_remove_group(&policy->kobj, &dbs_attr_group);
dbs_enable--;
* Stop the timerschedule work, when this governor
* is used for first time
*/
- if (dbs_enable == 0)
+ if (dbs_enable == 0)
dbs_timer_exit();
-
- up(&dbs_sem);
+
+ mutex_unlock(&dbs_mutex);
break;
case CPUFREQ_GOV_LIMITS:
- down(&dbs_sem);
+ mutex_lock(&dbs_mutex);
if (policy->max < this_dbs_info->cur_policy->cur)
__cpufreq_driver_target(
this_dbs_info->cur_policy,
- policy->max, CPUFREQ_RELATION_H);
+ policy->max, CPUFREQ_RELATION_H);
else if (policy->min > this_dbs_info->cur_policy->cur)
__cpufreq_driver_target(
this_dbs_info->cur_policy,
- policy->min, CPUFREQ_RELATION_L);
- up(&dbs_sem);
+ policy->min, CPUFREQ_RELATION_L);
+ mutex_unlock(&dbs_mutex);
break;
}
return 0;
}
-struct cpufreq_governor cpufreq_gov_dbs = {
+static struct cpufreq_governor cpufreq_gov_dbs = {
.name = "ondemand",
.governor = cpufreq_governor_dbs,
.owner = THIS_MODULE,
};
-EXPORT_SYMBOL(cpufreq_gov_dbs);
static int __init cpufreq_gov_dbs_init(void)
{
static void __exit cpufreq_gov_dbs_exit(void)
{
- /* Make sure that the scheduled work is indeed not running */
- flush_scheduled_work();
+ /* Make sure that the scheduled work is indeed not running.
+ Assumes the timer has been cancelled first. */
+ if (dbs_workq) {
+ flush_workqueue(dbs_workq);
+ destroy_workqueue(dbs_workq);
+ }
cpufreq_unregister_governor(&cpufreq_gov_dbs);
}