| Message ID | 09a94ff044ff6a6f7a5d953c3b1f3102c1dc50cf.1359653181.git.viresh.kumar@linaro.org (mailing list archive) |
|---|---|
| State | Accepted, archived |
| Headers | show |
On 31 January 2013 22:58, Viresh Kumar <viresh.kumar@linaro.org> wrote: > CPUFREQ_GOV_START/STOP are called only once for all policy->cpus and hence we > don't need to adapt cpufreq_governor_dbs() routine for multiple calls. > > So, this patch removes dbs_data->enable field entirely. And rearrange code a > bit. > > Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org> > --- > Hi Fabio, > > I have fixed all the pending issues, but haven't checked these patches. Can you > please add your tested-by (obviously after testing them) ? BTW, these are rebased over your patches and git log shows: af8a2e1 cpufreq: governors: Remove code redundancy between governors 09a94ff cpufreq: governors: Get rid of dbs_data->enable field faa8107 cpufreq: governors: implement generic policy_is_shared 88fe0a9 cpufreq: governors: clean timer init and exit code 13c18e8 cpufreq: governors: fix misuse of cdbs.cpu -- To unsubscribe from this list: send the line "unsubscribe linux-pm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Thu, Jan 31, 2013 at 10:58:01PM +0530, Viresh Kumar wrote: > CPUFREQ_GOV_START/STOP are called only once for all policy->cpus and hence we > don't need to adapt cpufreq_governor_dbs() routine for multiple calls. > > So, this patch removes dbs_data->enable field entirely. And rearrange code a > bit. > > Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org> > --- > Hi Fabio, > > I have fixed all the pending issues, but haven't checked these patches. Can you > please add your tested-by (obviously after testing them) ? > > Compile tested only. Hello Viresh, thanks for getting this done... looks much cleaner now! I tested both patches on my ux500 setup (dual Cortex-A9) and it seems to run correctly on both CPU load changes and CPU hotplug, so: Tested-by: Fabio Baltieri <fabio.baltieri@linaro.org> As a sidenote, I noticed just now that since: bc92bea cpufreq: Notify governors when cpus are hot-[un]plugged governor's sampling_rate gets reset to default every time you hotplug a CPU (the one you read/write on /sys/devices/system/cpu/cpufreq/ondemand/sampling_rate). If you need further tests, I'll be back on Monday. Thanks, Fabio
On 1 February 2013 00:14, Fabio Baltieri <fabio.baltieri@linaro.org> wrote: > Hello Viresh, thanks for getting this done... looks much cleaner now! > > I tested both patches on my ux500 setup (dual Cortex-A9) and it seems to > run correctly on both CPU load changes and CPU hotplug, so: > > Tested-by: Fabio Baltieri <fabio.baltieri@linaro.org> Thanks. > As a sidenote, I noticed just now that since: > > bc92bea cpufreq: Notify governors when cpus are hot-[un]plugged > > governor's sampling_rate gets reset to default every time you hotplug a > CPU (the one you read/write on > /sys/devices/system/cpu/cpufreq/ondemand/sampling_rate). > > If you need further tests, I'll be back on Monday. I will have a look. -- To unsubscribe from this list: send the line "unsubscribe linux-pm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 1 February 2013 09:22, Viresh Kumar <viresh.kumar@linaro.org> wrote: > On 1 February 2013 00:14, Fabio Baltieri <fabio.baltieri@linaro.org> wrote: >> As a sidenote, I noticed just now that since: >> >> bc92bea cpufreq: Notify governors when cpus are hot-[un]plugged >> >> governor's sampling_rate gets reset to default every time you hotplug a >> CPU (the one you read/write on >> /sys/devices/system/cpu/cpufreq/ondemand/sampling_rate). >> >> If you need further tests, I'll be back on Monday. > > I will have a look. Fixed with: https://lkml.org/lkml/2013/2/1/6 -- To unsubscribe from this list: send the line "unsubscribe linux-pm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/drivers/cpufreq/cpufreq_governor.c b/drivers/cpufreq/cpufreq_governor.c index 46f1c78..29d6a59 100644 --- a/drivers/cpufreq/cpufreq_governor.c +++ b/drivers/cpufreq/cpufreq_governor.c @@ -182,6 +182,8 @@ int cpufreq_governor_dbs(struct dbs_data *dbs_data, { struct od_cpu_dbs_info_s *od_dbs_info = NULL; struct cs_cpu_dbs_info_s *cs_dbs_info = NULL; + struct cs_ops *cs_ops = NULL; + struct od_ops *od_ops = NULL; struct od_dbs_tuners *od_tuners = dbs_data->tuners; struct cs_dbs_tuners *cs_tuners = dbs_data->tuners; struct cpu_dbs_common_info *cpu_cdbs; @@ -194,10 +196,12 @@ int cpufreq_governor_dbs(struct dbs_data *dbs_data, cs_dbs_info = dbs_data->get_cpu_dbs_info_s(cpu); sampling_rate = &cs_tuners->sampling_rate; ignore_nice = cs_tuners->ignore_nice; + cs_ops = dbs_data->gov_ops; } else { od_dbs_info = dbs_data->get_cpu_dbs_info_s(cpu); sampling_rate = &od_tuners->sampling_rate; ignore_nice = od_tuners->ignore_nice; + od_ops = dbs_data->gov_ops; } switch (event) { @@ -207,10 +211,9 @@ int cpufreq_governor_dbs(struct dbs_data *dbs_data, mutex_lock(&dbs_data->mutex); - dbs_data->enable++; for_each_cpu(j, policy->cpus) { - struct cpu_dbs_common_info *j_cdbs; - j_cdbs = dbs_data->get_cpu_cdbs(j); + struct cpu_dbs_common_info *j_cdbs = + dbs_data->get_cpu_cdbs(j); j_cdbs->cpu = j; j_cdbs->cur_policy = policy; @@ -225,13 +228,6 @@ int cpufreq_governor_dbs(struct dbs_data *dbs_data, dbs_data->gov_dbs_timer); } - /* - * Start the timerschedule work, when this governor is used for - * first time - */ - if (dbs_data->enable != 1) - goto second_time; - rc = sysfs_create_group(cpufreq_global_kobject, dbs_data->attr_group); if (rc) { @@ -249,17 +245,19 @@ int cpufreq_governor_dbs(struct dbs_data *dbs_data, * governor, thus we are bound to jiffes/HZ */ if (dbs_data->governor == GOV_CONSERVATIVE) { - struct cs_ops *ops = dbs_data->gov_ops; - - cpufreq_register_notifier(ops->notifier_block, + cs_dbs_info->down_skip = 0; + cs_dbs_info->enable = 1; + cs_dbs_info->requested_freq = policy->cur; + cpufreq_register_notifier(cs_ops->notifier_block, CPUFREQ_TRANSITION_NOTIFIER); dbs_data->min_sampling_rate = MIN_SAMPLING_RATE_RATIO * jiffies_to_usecs(10); } else { - struct od_ops *ops = dbs_data->gov_ops; - - od_tuners->io_is_busy = ops->io_busy(); + od_dbs_info->rate_mult = 1; + od_dbs_info->sample_type = OD_NORMAL_SAMPLE; + od_ops->powersave_bias_init_cpu(cpu); + od_tuners->io_is_busy = od_ops->io_busy(); } /* Bring kernel and HW constraints together */ @@ -267,18 +265,6 @@ int cpufreq_governor_dbs(struct dbs_data *dbs_data, MIN_LATENCY_MULTIPLIER * latency); *sampling_rate = max(dbs_data->min_sampling_rate, latency * LATENCY_MULTIPLIER); - -second_time: - if (dbs_data->governor == GOV_CONSERVATIVE) { - cs_dbs_info->down_skip = 0; - cs_dbs_info->enable = 1; - cs_dbs_info->requested_freq = policy->cur; - } else { - struct od_ops *ops = dbs_data->gov_ops; - od_dbs_info->rate_mult = 1; - od_dbs_info->sample_type = OD_NORMAL_SAMPLE; - ops->powersave_bias_init_cpu(cpu); - } mutex_unlock(&dbs_data->mutex); /* Initiate timer time stamp */ @@ -297,16 +283,12 @@ second_time: mutex_lock(&dbs_data->mutex); mutex_destroy(&cpu_cdbs->timer_mutex); - dbs_data->enable--; - if (!dbs_data->enable) { - struct cs_ops *ops = dbs_data->gov_ops; - - sysfs_remove_group(cpufreq_global_kobject, - dbs_data->attr_group); - if (dbs_data->governor == GOV_CONSERVATIVE) - cpufreq_unregister_notifier(ops->notifier_block, - CPUFREQ_TRANSITION_NOTIFIER); - } + + sysfs_remove_group(cpufreq_global_kobject, + dbs_data->attr_group); + if (dbs_data->governor == GOV_CONSERVATIVE) + cpufreq_unregister_notifier(cs_ops->notifier_block, + CPUFREQ_TRANSITION_NOTIFIER); mutex_unlock(&dbs_data->mutex); break; diff --git a/drivers/cpufreq/cpufreq_governor.h b/drivers/cpufreq/cpufreq_governor.h index b72e628..c19a16c 100644 --- a/drivers/cpufreq/cpufreq_governor.h +++ b/drivers/cpufreq/cpufreq_governor.h @@ -130,7 +130,6 @@ struct dbs_data { #define GOV_CONSERVATIVE 1 int governor; unsigned int min_sampling_rate; - unsigned int enable; /* number of CPUs using this policy */ struct attribute_group *attr_group; void *tuners;
CPUFREQ_GOV_START/STOP are called only once for all policy->cpus and hence we don't need to adapt cpufreq_governor_dbs() routine for multiple calls. So, this patch removes dbs_data->enable field entirely. And rearrange code a bit. Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org> --- Hi Fabio, I have fixed all the pending issues, but haven't checked these patches. Can you please add your tested-by (obviously after testing them) ? Compile tested only. drivers/cpufreq/cpufreq_governor.c | 58 +++++++++++++------------------------- drivers/cpufreq/cpufreq_governor.h | 1 - 2 files changed, 20 insertions(+), 39 deletions(-)