Skip to contentSkip to conversation

cpufreq: Fix ondemand governor powersave_bias execution time misuse

PATCH In review Youquan Song 4 participants 7 messages Nov 5, 2010 – Mar 2, 2011 CPU frequency scaling framework

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 cpufreq: Fix ondemand governor powersave_bias execution time misuse

Ondemand governor use powersave_bias tunable to do aggressive power save by decrease CPU average frequency. The average frequency achieve by adjust low and high frequency's execution time proportion during one sample time interval.

Current kernel, only the high frequency executes in time proportion, but the low frequency wrongly execute in one whole sample time interval when powersave_bias is set.

The patch fix it by set low frequency execution time to it deserved.

---

drivers/cpufreq/cpufreq_ondemand.c +2 -0
index c631f27..01d84fa 100644
@@ -666,6 +666,8 @@ static void do_dbs_timer(struct work_struct *work)
 	} else {
 		__cpufreq_driver_target(dbs_info->cur_policy,
 			dbs_info->freq_lo, CPUFREQ_RELATION_H);
+		if (dbs_info->sample_type == DBS_SUB_SAMPLE)
+			delay = dbs_info->freq_lo_jiffies;
 	}
 	queue_delayed_work_on(cpu, kondemand_wq, &dbs_info->work, delay);
 	mutex_unlock(&dbs_info->timer_mutex);
#1 lore raw
Quoted 4 lines · Youquan Song

Ondemand governor use powersave_bias tunable to do aggressive power save by decrease CPU average frequency. The average frequency achieve by adjust low and high frequency's execution time proportion during one sample time interval.

Quoted 8 lines

Current kernel, only the high frequency executes in time proportion, but the low frequency wrongly execute in one whole sample time interval when powersave_bias is set.

The patch fix it by set low frequency execution time to it deserved.

---

while your patch looks correct, I think the whole feature is wonky and likely ought to be removed...........

#2 lore raw
Quoted 9 lines · Youquan Song
Quoted 8 lines

Ondemand governor use powersave_bias tunable to do aggressive power save by decrease CPU average frequency. The average frequency achieve by adjust low and high frequency's execution time proportion during one sample time interval.

Current kernel, only the high frequency executes in time proportion, but the low frequency wrongly execute in one whole sample time interval when powersave_bias is set.

Quoted 8 lines
Quoted 4 lines

The patch fix it by set low frequency execution time to it deserved.

---

while your patch looks correct, I think the whole feature is wonky and likely ought to be removed...........

powersave_bias should be a useful feature, which give user an opportunity to set the CPU to work at wider average frequency accord to real
workload requirement, not just CPU frequency specific at P0,P1,...Pn.
At the same time, it meets end user power comsumpation requirement in some degree.
Unfortunately, it does not work for a long time.

Thanks
-Youquan

#3 lore raw
Quoted 14 lines · Youquan Song
Quoted 14 lines · Youquan Song
Quoted 12 lines

Ondemand governor use powersave_bias tunable to do aggressive power save by decrease CPU average frequency. The average frequency achieve by adjust low and high frequency's execution time proportion during one sample time interval.

Current kernel, only the high frequency executes in time proportion, but the low frequency wrongly execute in one whole sample time interval when powersave_bias is set.

The patch fix it by set low frequency execution time to it deserved.

---

Quoted 8 lines
Quoted 2 lines

while your patch looks correct, I think the whole feature is wonky and likely ought to be removed...........

powersave_bias should be a useful feature, which give user an opportunity to set the CPU to work at wider average frequency accord to real
workload requirement, not just CPU frequency specific at P0,P1,...Pn.
At the same time, it meets end user power comsumpation requirement in some degree.
Unfortunately, it does not work for a long time.

in all measurements I've done... the intermediate states (between P1 and Pn) seem to not be power efficient.....

#4 lore raw
In reply to Youquan Song
Quoted 18 lines · On Fri, 5 Nov 2010 11:19:51 -0400 Youquan Song <[email protected]>…

Ondemand governor use powersave_bias tunable to do aggressive power save by decrease CPU average frequency. The average frequency achieve by adjust low and high frequency's execution time proportion during one sample time interval.

Current kernel, only the high frequency executes in time proportion, but the low frequency wrongly execute in one whole sample time interval when powersave_bias is set.

The patch fix it by set low frequency execution time to it deserved.

---

drivers/cpufreq/cpufreq_ondemand.c +0 -0
index c631f27..01d84fa 100644

drivers/cpufreq/cpufreq_ondemand.c static void do_dbs_timer(struct work_struct *work) · line 673

Quoted 8 lines
drivers/cpufreq/cpufreq_ondemand.c +2 -0
@@ -666,6 +666,8 @@ static void do_dbs_timer(struct work_struct *work)
 	} else {
 		__cpufreq_driver_target(dbs_info->cur_policy,
 			dbs_info->freq_lo, CPUFREQ_RELATION_H);
+		if (dbs_info->sample_type == DBS_SUB_SAMPLE)
+			delay = dbs_info->freq_lo_jiffies;
 	}
 	queue_delayed_work_on(cpu, kondemand_wq, &dbs_info->work, delay);
 	mutex_unlock(&dbs_info->timer_mutex);

Where are we with this patch? Still needed in curent mainline?

Thnks.

#5 lore raw
Quoted 2 lines

Where are we with this patch? Still needed in curent mainline?

Do need it by current mainline. Without the patch, the ondemand power_bias feature is broken and does not work at all. Please take it.

Thanks
-Youquan

#6 lore raw
Quoted 5 lines · Youquan Song
Quoted 2 lines

Where are we with this patch? Still needed in curent mainline?

Do need it by current mainline. Without the patch, the ondemand power_bias feature is broken and does not work at all. Please take it.

This patch from Vincent Guittot is also pending, which seems to achieve the same thing ?

commit 1f5a2969af7e0d040876c7ff25308c840eef18d4

Author: Vincent Guittot <[email protected]>
Date:   Mon Feb 7 17:14:25 2011 +0100

[CPUFREQ] calculate delay after dbs_check_cpu

calculate ondemand delay after dbs_check_cpu call because it can modify rate_mult value

use freq_lo_jiffies value for the sub sample period of powersave_bias mode

drivers/cpufreq/cpufreq_ondemand.c +11 -6
index c631f27..b4e1f98 100644
@@ -644,12 +644,7 @@ static void do_dbs_timer(struct work_struct *work)
 	unsigned int cpu = dbs_info->cpu;
 	int sample_type = dbs_info->sample_type;
 
-	/* We want all CPUs to do sampling nearly on same jiffy */
-	int delay = usecs_to_jiffies(dbs_tuners_ins.sampling_rate
-		* dbs_info->rate_mult);
-
-	if (num_online_cpus() > 1)
-		delay -= jiffies % delay;
+	int delay;
 
 	mutex_lock(&dbs_info->timer_mutex);
 
@@ -662,10 +657,20 @@ static void do_dbs_timer(struct work_struct *work)
 			/* Setup timer for SUB_SAMPLE */
 			dbs_info->sample_type = DBS_SUB_SAMPLE;
 			delay = dbs_info->freq_hi_jiffies;
+		} else {
+			/* We want all CPUs to do sampling nearly on
+			 * same jiffy
+			 */
+			delay = usecs_to_jiffies(dbs_tuners_ins.sampling_rate
+				* dbs_info->rate_mult);
+
+			if (num_online_cpus() > 1)
+				delay -= jiffies % delay;
 		}
 	} else {
 		__cpufreq_driver_target(dbs_info->cur_policy,
 			dbs_info->freq_lo, CPUFREQ_RELATION_H);
+		delay = dbs_info->freq_lo_jiffies;
 	}
 	queue_delayed_work_on(cpu, kondemand_wq, &dbs_info->work, delay);
 	mutex_unlock(&dbs_info->timer_mutex);