On Wednesday, November 28, 2012 11:51:20 AM Fabio Baltieri wrote:

Hello Rafael,

Hi,

On Tue, Nov 27, 2012 at 11:05:52PM +0100, Rafael J. Wysocki wrote:

quoted

quoted

 static inline void dbs_timer_init(struct dbs_data *dbs_data,
-		struct cpu_dbs_common_info *cdbs, unsigned int sampling_rate)
+				  struct cpu_dbs_common_info *cdbs,
+				  unsigned int sampling_rate,
+				  int cpu)
 {
 	int delay = delay_for_sampling_rate(sampling_rate);
+	struct cpu_dbs_common_info *cdbs_local = dbs_data->get_cpu_cdbs(cpu);
+	struct od_cpu_dbs_info_s *od_dbs_info;
+
+	cancel_delayed_work_sync(&cdbs_local->work);
+
+	if (dbs_data->governor == GOV_ONDEMAND) {
+		od_dbs_info = dbs_data->get_cpu_dbs_info_s(cpu);
+		od_dbs_info->sample_type = OD_NORMAL_SAMPLE;
+	}

The patch looks good in general except for the special case above.

Why exactly is it necessary?

Now that you point it out... it's not!  It was part of ondemand init and
moved in cpufreq_governor_dbs, I forgot to take it out the way.

Also, I think that cancel_delayed_work_sync can be removed too.

Should I send an updated version as soon as I get an ack for the other
patches in the series or do you want me to wait until 3.8-rc1?

Well, if it's not very urgent, I'd prefer it to wait a bit longer,
get some more testing and so on.

Thanks,
Rafael


-- 
I speak only for myself.
Rafael J. Wysocki, Intel Open Source Technology Center.