#define DEF_FREQUENCY_UP_THRESHOLD (80)
#define MICRO_FREQUENCY_DOWN_DIFFERENTIAL (3)
#define MICRO_FREQUENCY_UP_THRESHOLD (95)
+#define MICRO_FREQUENCY_MIN_SAMPLE_RATE (10000)
#define MIN_FREQUENCY_UP_THRESHOLD (11)
#define MAX_FREQUENCY_UP_THRESHOLD (100)
* this governor will not work.
* All times here are in uS.
*/
-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)
-/* Above MIN_SAMPLING_RATE will vanish with its sysfs file soon
- * Define the minimal settable sampling rate to the greater of:
- * - "HW transition latency" * 100 (same as default sampling / 10)
- * - MIN_STAT_SAMPLING_RATE
- * To avoid that userspace shoots itself.
-*/
-static unsigned int minimum_sampling_rate(void)
-{
- return max(def_sampling_rate / 10, MIN_STAT_SAMPLING_RATE);
-}
-/* This will also vanish soon with removing sampling_rate_max */
-#define MAX_SAMPLING_RATE (500 * def_sampling_rate)
+static unsigned int min_sampling_rate;
+
#define LATENCY_MULTIPLIER (1000)
+#define MIN_LATENCY_MULTIPLIER (100)
#define TRANSITION_LATENCY_LIMIT (10 * 1000 * 1000)
static void do_dbs_timer(struct work_struct *work);
+static int cpufreq_governor_dbs(struct cpufreq_policy *policy,
+ unsigned int event);
+
+#ifndef CONFIG_CPU_FREQ_DEFAULT_GOV_ONDEMAND
+static
+#endif
+struct cpufreq_governor cpufreq_gov_ondemand = {
+ .name = "ondemand",
+ .governor = cpufreq_governor_dbs,
+ .max_transition_latency = TRANSITION_LATENCY_LIMIT,
+ .owner = THIS_MODULE,
+};
/* Sampling types */
enum {DBS_NORMAL_SAMPLE, DBS_SUB_SAMPLE};
unsigned int freq_lo_jiffies;
unsigned int freq_hi_jiffies;
int cpu;
- unsigned int enable:1,
- sample_type:1;
+ unsigned int sample_type:1;
+ /*
+ * percpu mutex that serializes governor limit change with
+ * do_dbs_timer invocation. We do not want do_dbs_timer to run
+ * when user is changing the governor or limits.
+ */
+ struct mutex timer_mutex;
};
-static DEFINE_PER_CPU(struct cpu_dbs_info_s, cpu_dbs_info);
+static DEFINE_PER_CPU(struct cpu_dbs_info_s, od_cpu_dbs_info);
static unsigned int dbs_enable; /* number of CPUs using this policy */
/*
- * DEADLOCK ALERT! There is a ordering requirement between cpu_hotplug
- * lock and dbs_mutex. cpu_hotplug lock should always be held before
- * dbs_mutex. If any function that can potentially take cpu_hotplug lock
- * (like __cpufreq_driver_target()) is being called with dbs_mutex taken, then
- * cpu_hotplug lock should be taken before that. Note that cpu_hotplug lock
- * is recursive for the same process. -Venki
- * DEADLOCK ALERT! (2) : do_dbs_timer() must not take the dbs_mutex, because it
- * would deadlock with cancel_delayed_work_sync(), which is needed for proper
- * raceless workqueue teardown.
+ * dbs_mutex protects data in dbs_tuners_ins from concurrent changes on
+ * different CPUs. It protects dbs_enable in governor start/stop.
*/
static DEFINE_MUTEX(dbs_mutex);
idle_time = cputime64_sub(cur_wall_time, busy_time);
if (wall)
- *wall = cur_wall_time;
+ *wall = (cputime64_t)jiffies_to_usecs(cur_wall_time);
- return idle_time;
+ return (cputime64_t)jiffies_to_usecs(idle_time);
}
static inline cputime64_t get_cpu_idle_time(unsigned int cpu, cputime64_t *wall)
unsigned int freq_hi, freq_lo;
unsigned int index = 0;
unsigned int jiffies_total, jiffies_hi, jiffies_lo;
- struct cpu_dbs_info_s *dbs_info = &per_cpu(cpu_dbs_info, policy->cpu);
+ struct cpu_dbs_info_s *dbs_info = &per_cpu(od_cpu_dbs_info,
+ policy->cpu);
if (!dbs_info->freq_table) {
dbs_info->freq_lo = 0;
return freq_hi;
}
+static void ondemand_powersave_bias_init_cpu(int cpu)
+{
+ struct cpu_dbs_info_s *dbs_info = &per_cpu(od_cpu_dbs_info, cpu);
+ dbs_info->freq_table = cpufreq_frequency_get_table(cpu);
+ dbs_info->freq_lo = 0;
+}
+
static void ondemand_powersave_bias_init(void)
{
int i;
for_each_online_cpu(i) {
- struct cpu_dbs_info_s *dbs_info = &per_cpu(cpu_dbs_info, i);
- dbs_info->freq_table = cpufreq_frequency_get_table(i);
- dbs_info->freq_lo = 0;
+ ondemand_powersave_bias_init_cpu(i);
}
}
/************************** sysfs interface ************************/
-static ssize_t show_sampling_rate_max(struct cpufreq_policy *policy, char *buf)
-{
- static int print_once;
- if (!print_once) {
- printk(KERN_INFO "CPUFREQ: ondemand sampling_rate_max "
- "sysfs file is deprecated - used by: %s\n",
- current->comm);
- print_once = 1;
- }
- return sprintf(buf, "%u\n", MAX_SAMPLING_RATE);
+static ssize_t show_sampling_rate_max(struct kobject *kobj,
+ struct attribute *attr, char *buf)
+{
+ printk_once(KERN_INFO "CPUFREQ: ondemand sampling_rate_max "
+ "sysfs file is deprecated - used by: %s\n", current->comm);
+ return sprintf(buf, "%u\n", -1U);
}
-static ssize_t show_sampling_rate_min(struct cpufreq_policy *policy, char *buf)
+static ssize_t show_sampling_rate_min(struct kobject *kobj,
+ struct attribute *attr, char *buf)
{
- static int print_once;
-
- if (!print_once) {
- printk(KERN_INFO "CPUFREQ: ondemand sampling_rate_min "
- "sysfs file is deprecated - used by: %s\n",
- current->comm);
- print_once = 1;
- }
- return sprintf(buf, "%u\n", MIN_SAMPLING_RATE);
+ return sprintf(buf, "%u\n", min_sampling_rate);
}
#define define_one_ro(_name) \
-static struct freq_attr _name = \
+static struct global_attr _name = \
__ATTR(_name, 0444, show_##_name, NULL)
define_one_ro(sampling_rate_max);
/* cpufreq_ondemand Governor Tunables */
#define show_one(file_name, object) \
static ssize_t show_##file_name \
-(struct cpufreq_policy *unused, char *buf) \
+(struct kobject *kobj, struct attribute *attr, char *buf) \
{ \
return sprintf(buf, "%u\n", dbs_tuners_ins.object); \
}
show_one(ignore_nice_load, ignore_nice);
show_one(powersave_bias, powersave_bias);
-static ssize_t store_sampling_rate(struct cpufreq_policy *unused,
- const char *buf, size_t count)
+/*** delete after deprecation time ***/
+
+#define DEPRECATION_MSG(file_name) \
+ printk_once(KERN_INFO "CPUFREQ: Per core ondemand sysfs " \
+ "interface is deprecated - " #file_name "\n");
+
+#define show_one_old(file_name) \
+static ssize_t show_##file_name##_old \
+(struct cpufreq_policy *unused, char *buf) \
+{ \
+ printk_once(KERN_INFO "CPUFREQ: Per core ondemand sysfs " \
+ "interface is deprecated - " #file_name "\n"); \
+ return show_##file_name(NULL, NULL, buf); \
+}
+show_one_old(sampling_rate);
+show_one_old(up_threshold);
+show_one_old(ignore_nice_load);
+show_one_old(powersave_bias);
+show_one_old(sampling_rate_min);
+show_one_old(sampling_rate_max);
+
+#define define_one_ro_old(object, _name) \
+static struct freq_attr object = \
+__ATTR(_name, 0444, show_##_name##_old, NULL)
+
+define_one_ro_old(sampling_rate_min_old, sampling_rate_min);
+define_one_ro_old(sampling_rate_max_old, sampling_rate_max);
+
+/*** delete after deprecation time ***/
+
+static ssize_t store_sampling_rate(struct kobject *a, struct attribute *b,
+ const char *buf, size_t count)
{
unsigned int input;
int ret;
ret = sscanf(buf, "%u", &input);
+ if (ret != 1)
+ return -EINVAL;
mutex_lock(&dbs_mutex);
- if (ret != 1) {
- mutex_unlock(&dbs_mutex);
- return -EINVAL;
- }
- dbs_tuners_ins.sampling_rate = max(input, minimum_sampling_rate());
+ dbs_tuners_ins.sampling_rate = max(input, min_sampling_rate);
mutex_unlock(&dbs_mutex);
return count;
}
-static ssize_t store_up_threshold(struct cpufreq_policy *unused,
- const char *buf, size_t count)
+static ssize_t store_up_threshold(struct kobject *a, struct attribute *b,
+ const char *buf, size_t count)
{
unsigned int input;
int ret;
ret = sscanf(buf, "%u", &input);
- mutex_lock(&dbs_mutex);
if (ret != 1 || input > MAX_FREQUENCY_UP_THRESHOLD ||
input < MIN_FREQUENCY_UP_THRESHOLD) {
- mutex_unlock(&dbs_mutex);
return -EINVAL;
}
+ mutex_lock(&dbs_mutex);
dbs_tuners_ins.up_threshold = input;
mutex_unlock(&dbs_mutex);
return count;
}
-static ssize_t store_ignore_nice_load(struct cpufreq_policy *policy,
- const char *buf, size_t count)
+static ssize_t store_ignore_nice_load(struct kobject *a, struct attribute *b,
+ const char *buf, size_t count)
{
unsigned int input;
int ret;
/* we need to re-evaluate prev_cpu_idle */
for_each_online_cpu(j) {
struct cpu_dbs_info_s *dbs_info;
- dbs_info = &per_cpu(cpu_dbs_info, j);
+ dbs_info = &per_cpu(od_cpu_dbs_info, j);
dbs_info->prev_cpu_idle = get_cpu_idle_time(j,
&dbs_info->prev_cpu_wall);
if (dbs_tuners_ins.ignore_nice)
return count;
}
-static ssize_t store_powersave_bias(struct cpufreq_policy *unused,
- const char *buf, size_t count)
+static ssize_t store_powersave_bias(struct kobject *a, struct attribute *b,
+ const char *buf, size_t count)
{
unsigned int input;
int ret;
}
#define define_one_rw(_name) \
-static struct freq_attr _name = \
+static struct global_attr _name = \
__ATTR(_name, 0644, show_##_name, store_##_name)
define_one_rw(sampling_rate);
.name = "ondemand",
};
+/*** delete after deprecation time ***/
+
+#define write_one_old(file_name) \
+static ssize_t store_##file_name##_old \
+(struct cpufreq_policy *unused, const char *buf, size_t count) \
+{ \
+ printk_once(KERN_INFO "CPUFREQ: Per core ondemand sysfs " \
+ "interface is deprecated - " #file_name "\n"); \
+ return store_##file_name(NULL, NULL, buf, count); \
+}
+write_one_old(sampling_rate);
+write_one_old(up_threshold);
+write_one_old(ignore_nice_load);
+write_one_old(powersave_bias);
+
+#define define_one_rw_old(object, _name) \
+static struct freq_attr object = \
+__ATTR(_name, 0644, show_##_name##_old, store_##_name##_old)
+
+define_one_rw_old(sampling_rate_old, sampling_rate);
+define_one_rw_old(up_threshold_old, up_threshold);
+define_one_rw_old(ignore_nice_load_old, ignore_nice_load);
+define_one_rw_old(powersave_bias_old, powersave_bias);
+
+static struct attribute *dbs_attributes_old[] = {
+ &sampling_rate_max_old.attr,
+ &sampling_rate_min_old.attr,
+ &sampling_rate_old.attr,
+ &up_threshold_old.attr,
+ &ignore_nice_load_old.attr,
+ &powersave_bias_old.attr,
+ NULL
+};
+
+static struct attribute_group dbs_attr_group_old = {
+ .attrs = dbs_attributes_old,
+ .name = "ondemand",
+};
+
+/*** delete after deprecation time ***/
+
/************************** sysfs end ************************/
static void dbs_check_cpu(struct cpu_dbs_info_s *this_dbs_info)
struct cpufreq_policy *policy;
unsigned int j;
- if (!this_dbs_info->enable)
- return;
-
this_dbs_info->freq_lo = 0;
policy = this_dbs_info->cur_policy;
unsigned int load, load_freq;
int freq_avg;
- j_dbs_info = &per_cpu(cpu_dbs_info, j);
+ j_dbs_info = &per_cpu(od_cpu_dbs_info, j);
cur_idle_time = get_cpu_idle_time(j, &cur_wall_time);
int delay = usecs_to_jiffies(dbs_tuners_ins.sampling_rate);
delay -= jiffies % delay;
-
- if (lock_policy_rwsem_write(cpu) < 0)
- return;
-
- if (!dbs_info->enable) {
- unlock_policy_rwsem_write(cpu);
- return;
- }
+ mutex_lock(&dbs_info->timer_mutex);
/* Common NORMAL_SAMPLE setup */
dbs_info->sample_type = DBS_NORMAL_SAMPLE;
dbs_info->freq_lo, CPUFREQ_RELATION_H);
}
queue_delayed_work_on(cpu, kondemand_wq, &dbs_info->work, delay);
- unlock_policy_rwsem_write(cpu);
+ mutex_unlock(&dbs_info->timer_mutex);
}
static inline void dbs_timer_init(struct cpu_dbs_info_s *dbs_info)
int delay = usecs_to_jiffies(dbs_tuners_ins.sampling_rate);
delay -= jiffies % delay;
- dbs_info->enable = 1;
- ondemand_powersave_bias_init();
dbs_info->sample_type = DBS_NORMAL_SAMPLE;
INIT_DELAYED_WORK_DEFERRABLE(&dbs_info->work, do_dbs_timer);
queue_delayed_work_on(dbs_info->cpu, kondemand_wq, &dbs_info->work,
static inline void dbs_timer_exit(struct cpu_dbs_info_s *dbs_info)
{
- dbs_info->enable = 0;
cancel_delayed_work_sync(&dbs_info->work);
}
unsigned int j;
int rc;
- this_dbs_info = &per_cpu(cpu_dbs_info, cpu);
+ this_dbs_info = &per_cpu(od_cpu_dbs_info, cpu);
switch (event) {
case CPUFREQ_GOV_START:
if ((!cpu_online(cpu)) || (!policy->cur))
return -EINVAL;
- if (this_dbs_info->enable) /* Already enabled */
- break;
-
mutex_lock(&dbs_mutex);
- dbs_enable++;
- rc = sysfs_create_group(&policy->kobj, &dbs_attr_group);
+ rc = sysfs_create_group(&policy->kobj, &dbs_attr_group_old);
if (rc) {
- dbs_enable--;
mutex_unlock(&dbs_mutex);
return rc;
}
+ dbs_enable++;
for_each_cpu(j, policy->cpus) {
struct cpu_dbs_info_s *j_dbs_info;
- j_dbs_info = &per_cpu(cpu_dbs_info, j);
+ j_dbs_info = &per_cpu(od_cpu_dbs_info, j);
j_dbs_info->cur_policy = policy;
j_dbs_info->prev_cpu_idle = get_cpu_idle_time(j,
}
}
this_dbs_info->cpu = cpu;
+ ondemand_powersave_bias_init_cpu(cpu);
/*
* Start the timerschedule work, when this governor
* is used for first time
*/
if (dbs_enable == 1) {
unsigned int latency;
+
+ rc = sysfs_create_group(cpufreq_global_kobject,
+ &dbs_attr_group);
+ if (rc) {
+ mutex_unlock(&dbs_mutex);
+ return rc;
+ }
+
/* policy latency is in nS. Convert it to uS first */
latency = policy->cpuinfo.transition_latency / 1000;
if (latency == 0)
latency = 1;
-
- def_sampling_rate =
- max(latency * LATENCY_MULTIPLIER,
- MIN_STAT_SAMPLING_RATE);
-
- dbs_tuners_ins.sampling_rate = def_sampling_rate;
+ /* Bring kernel and HW constraints together */
+ min_sampling_rate = max(min_sampling_rate,
+ MIN_LATENCY_MULTIPLIER * latency);
+ dbs_tuners_ins.sampling_rate =
+ max(min_sampling_rate,
+ latency * LATENCY_MULTIPLIER);
}
- dbs_timer_init(this_dbs_info);
-
mutex_unlock(&dbs_mutex);
+
+ mutex_init(&this_dbs_info->timer_mutex);
+ dbs_timer_init(this_dbs_info);
break;
case CPUFREQ_GOV_STOP:
- mutex_lock(&dbs_mutex);
dbs_timer_exit(this_dbs_info);
- sysfs_remove_group(&policy->kobj, &dbs_attr_group);
+
+ mutex_lock(&dbs_mutex);
+ sysfs_remove_group(&policy->kobj, &dbs_attr_group_old);
+ mutex_destroy(&this_dbs_info->timer_mutex);
dbs_enable--;
mutex_unlock(&dbs_mutex);
+ if (!dbs_enable)
+ sysfs_remove_group(cpufreq_global_kobject,
+ &dbs_attr_group);
break;
case CPUFREQ_GOV_LIMITS:
- mutex_lock(&dbs_mutex);
+ mutex_lock(&this_dbs_info->timer_mutex);
if (policy->max < this_dbs_info->cur_policy->cur)
__cpufreq_driver_target(this_dbs_info->cur_policy,
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);
- mutex_unlock(&dbs_mutex);
+ mutex_unlock(&this_dbs_info->timer_mutex);
break;
}
return 0;
}
-#ifndef CONFIG_CPU_FREQ_DEFAULT_GOV_ONDEMAND
-static
-#endif
-struct cpufreq_governor cpufreq_gov_ondemand = {
- .name = "ondemand",
- .governor = cpufreq_governor_dbs,
- .max_transition_latency = TRANSITION_LATENCY_LIMIT,
- .owner = THIS_MODULE,
-};
-
static int __init cpufreq_gov_dbs_init(void)
{
int err;
dbs_tuners_ins.up_threshold = MICRO_FREQUENCY_UP_THRESHOLD;
dbs_tuners_ins.down_differential =
MICRO_FREQUENCY_DOWN_DIFFERENTIAL;
+ /*
+ * In no_hz/micro accounting case we set the minimum frequency
+ * not depending on HZ, but fixed (very low). The deferred
+ * timer might skip some samples if idle/sleeping as needed.
+ */
+ min_sampling_rate = MICRO_FREQUENCY_MIN_SAMPLE_RATE;
+ } else {
+ /* For correct statistics, we need 10 ticks for each measure */
+ min_sampling_rate =
+ MIN_SAMPLING_RATE_RATIO * jiffies_to_usecs(10);
}
kondemand_wq = create_workqueue("kondemand");