linux-pm.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Chanwoo Choi <cw00.choi@samsung.com>
To: "Leonard Crestez" <leonard.crestez@nxp.com>,
	"MyungJoo Ham" <myungjoo.ham@samsung.com>,
	"Kyungmin Park" <kyungmin.park@samsung.com>,
	"Artur Świgoń" <a.swigon@partner.samsung.com>
Cc: Saravana Kannan <saravanak@google.com>,
	Krzysztof Kozlowski <krzk@kernel.org>,
	Alexandre Bailon <abailon@baylibre.com>,
	Georgi Djakov <georgi.djakov@linaro.org>,
	Jacky Bai <ping.bai@nxp.com>,
	Viresh Kumar <viresh.kumar@linaro.org>,
	linux-pm@vger.kernel.org, linux-arm-kernel@lists.infradead.org,
	"cpgs (cpgs@samsung.com)" <cpgs@samsung.com>
Subject: Re: [PATCH v3 1/2] PM / devfreq: Add dev_pm_qos support
Date: Wed, 21 Aug 2019 10:44:26 +0900	[thread overview]
Message-ID: <6134bb9e-9a16-b432-c191-c91f93844319@samsung.com> (raw)
In-Reply-To: <3b93af7e61a573ea2a123c353255645b5ad2a805.1566314535.git.leonard.crestez@nxp.com>

Hi,

On 19. 8. 21. 오전 12:24, Leonard Crestez wrote:
> Add dev_pm_qos notifies to devfreq core in order to support frequency
> limits via the dev_pm_qos_add_request.
> 
> Unlike the rest of devfreq the dev_pm_qos frequency is measured in Khz,
> this is consistent with current dev_pm_qos usage for cpufreq and
> allows frequencies above 2Ghz (pm_qos expresses limits as s32).
> 
> Signed-off-by: Leonard Crestez <leonard.crestez@nxp.com>
> ---
>  drivers/devfreq/devfreq.c | 95 ++++++++++++++++++++++++++++++++++++---
>  include/linux/devfreq.h   |  5 +++
>  2 files changed, 95 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/devfreq/devfreq.c b/drivers/devfreq/devfreq.c
> index 784c08e4f931..58deffa52a37 100644
> --- a/drivers/devfreq/devfreq.c
> +++ b/drivers/devfreq/devfreq.c
> @@ -22,10 +22,11 @@
>  #include <linux/platform_device.h>
>  #include <linux/list.h>
>  #include <linux/printk.h>
>  #include <linux/hrtimer.h>
>  #include <linux/of.h>
> +#include <linux/pm_qos.h>
>  #include "governor.h"
>  
>  #define CREATE_TRACE_POINTS
>  #include <trace/events/devfreq.h>
>  
> @@ -96,10 +97,30 @@ static unsigned long find_available_max_freq(struct devfreq *devfreq)
>  		dev_pm_opp_put(opp);
>  
>  	return max_freq;
>  }
>  
> +static unsigned long get_effective_min_freq(struct devfreq *devfreq)

I'm not sure that 'effective' expression is correct.
From this function, the devfreq can get the final target frequency.

I think that we need to use the more correct expression
to give the meaning of function as following:

get_min_freq
get_max_freq

or

get_biggest_min_freq
get_biggest_max_freq

or

others if there are more proper function name.


> +{
> +	lockdep_assert_held(&devfreq->lock);
> +
> +	return max3(devfreq->scaling_min_freq, devfreq->min_freq,
> +		1000 * (unsigned long)dev_pm_qos_read_value(

I prefer to use 'KHZ' defintion instead of 1000.
The constant definition is more easy to inform
the correct meaning of constant.

> +				devfreq->dev.parent,
> +				DEV_PM_QOS_MIN_FREQUENCY));
> +}
> +
> +static unsigned long get_effective_max_freq(struct devfreq *devfreq)

ditto.

> +{
> +	lockdep_assert_held(&devfreq->lock);
> +
> +	return min3(devfreq->scaling_max_freq, devfreq->max_freq,
> +		1000 * (unsigned long)dev_pm_qos_read_value(

ditto.

> +				devfreq->dev.parent,
> +				DEV_PM_QOS_MAX_FREQUENCY));
> +}
> +
>  /**
>   * devfreq_get_freq_level() - Lookup freq_table for the frequency
>   * @devfreq:	the devfreq instance
>   * @freq:	the target frequency
>   */
> @@ -356,12 +377,12 @@ int update_devfreq(struct devfreq *devfreq)
>  	 *
>  	 * List from the highest priority
>  	 * max_freq
>  	 * min_freq
>  	 */
> -	max_freq = min(devfreq->scaling_max_freq, devfreq->max_freq);
> -	min_freq = max(devfreq->scaling_min_freq, devfreq->min_freq);
> +	max_freq = get_effective_max_freq(devfreq);
> +	min_freq = get_effective_min_freq(devfreq);
>  
>  	if (freq < min_freq) {
>  		freq = min_freq;
>  		flags &= ~DEVFREQ_FLAG_LEAST_UPPER_BOUND; /* Use GLB */
>  	}
> @@ -570,10 +591,37 @@ static int devfreq_notifier_call(struct notifier_block *nb, unsigned long type,
>  	mutex_unlock(&devfreq->lock);
>  
>  	return ret;
>  }
>  
> +static int devfreq_qos_notifier_call(struct devfreq *devfreq)
> +{
> +	int ret;
> +
> +	mutex_lock(&devfreq->lock);
> +	ret = update_devfreq(devfreq);
> +	mutex_unlock(&devfreq->lock);
> +
> +	return ret;
> +}
> +
> +static int devfreq_qos_min_notifier_call(struct notifier_block *nb,
> +					 unsigned long val, void *ptr)
> +{
> +	struct devfreq *devfreq = container_of(nb, struct devfreq, nb_min);
> +
> +	return devfreq_qos_notifier_call(devfreq);
> +}
> +
> +static int devfreq_qos_max_notifier_call(struct notifier_block *nb,
> +					 unsigned long val, void *ptr)
> +{
> +	struct devfreq *devfreq = container_of(nb, struct devfreq, nb_max);
> +
> +	return devfreq_qos_notifier_call(devfreq);
> +}

Although the all functions has not the function description in devfreq.c,
You need to add the function description for newly added functions.

> +
>  /**
>   * devfreq_dev_release() - Callback for struct device to release the device.
>   * @dev:	the devfreq device
>   *
>   * Remove devfreq from the list and release its resources.
> @@ -592,10 +640,14 @@ static void devfreq_dev_release(struct device *dev)
>  	mutex_unlock(&devfreq_list_lock);
>  
>  	if (devfreq->profile->exit)
>  		devfreq->profile->exit(devfreq->dev.parent);
>  
> +	dev_pm_qos_remove_notifier(devfreq->dev.parent, &devfreq->nb_max,
> +			DEV_PM_QOS_MAX_FREQUENCY);
> +	dev_pm_qos_remove_notifier(devfreq->dev.parent, &devfreq->nb_min,
> +			DEV_PM_QOS_MIN_FREQUENCY);
>  	mutex_destroy(&devfreq->lock);
>  	kfree(devfreq);
>  }
>  
>  /**
> @@ -636,21 +688,40 @@ struct devfreq *devfreq_add_device(struct device *dev,
>  		err = -ENOMEM;
>  		goto err_out;
>  	}
>  
>  	mutex_init(&devfreq->lock);
> -	mutex_lock(&devfreq->lock);

Basically, I think that it is safe to lock when touch
the variable of the devfreq.

it is not proper way for the dev_pm_qos because 
it breaks the existing locking reason of devfreq's variables. 

>  	devfreq->dev.parent = dev;
>  	devfreq->dev.class = devfreq_class;
>  	devfreq->dev.release = devfreq_dev_release;
>  	devfreq->profile = profile;
>  	strncpy(devfreq->governor_name, governor_name, DEVFREQ_NAME_LEN);
>  	devfreq->previous_freq = profile->initial_freq;
>  	devfreq->last_status.current_frequency = profile->initial_freq;
>  	devfreq->data = data;
>  	devfreq->nb.notifier_call = devfreq_notifier_call;
>  
> +	/*
> +	 * notifier from pm_qos
> +	 *
> +	 * initialized outside of devfreq->lock to avoid circular warning
> +	 * between devfreq->lock and dev_pm_qos_mtx
> +	 */
> +	devfreq->nb_min.notifier_call = devfreq_qos_min_notifier_call;
> +	devfreq->nb_max.notifier_call = devfreq_qos_max_notifier_call;
> +
> +	err = dev_pm_qos_add_notifier(devfreq->dev.parent, &devfreq->nb_min,
> +				      DEV_PM_QOS_MIN_FREQUENCY);
> +	if (err)
> +		goto err_dev;
> +
> +	err = dev_pm_qos_add_notifier(devfreq->dev.parent, &devfreq->nb_max,
> +				      DEV_PM_QOS_MAX_FREQUENCY);
> +	if (err)
> +		goto err_dev;
> +
> +	mutex_lock(&devfreq->lock);
>  	if (!devfreq->profile->max_state && !devfreq->profile->freq_table) {
>  		mutex_unlock(&devfreq->lock);
>  		err = set_freq_table(devfreq);
>  		if (err < 0)
>  			goto err_dev;
> @@ -741,10 +812,14 @@ struct devfreq *devfreq_add_device(struct device *dev,
>  	mutex_unlock(&devfreq_list_lock);
>  err_devfreq:
>  	devfreq_remove_device(devfreq);
>  	devfreq = NULL;
>  err_dev:
> +	dev_pm_qos_remove_notifier(devfreq->dev.parent, &devfreq->nb_max,
> +				   DEV_PM_QOS_MAX_FREQUENCY);
> +	dev_pm_qos_remove_notifier(devfreq->dev.parent, &devfreq->nb_min,
> +				   DEV_PM_QOS_MIN_FREQUENCY);
>  	kfree(devfreq);
>  err_out:
>  	return ERR_PTR(err);
>  }
>  EXPORT_SYMBOL(devfreq_add_device);
> @@ -1312,12 +1387,17 @@ static ssize_t min_freq_store(struct device *dev, struct device_attribute *attr,
>  
>  static ssize_t min_freq_show(struct device *dev, struct device_attribute *attr,
>  			     char *buf)
>  {
>  	struct devfreq *df = to_devfreq(dev);
> +	ssize_t ret;
> +
> +	mutex_lock(&df->lock);
> +	ret = sprintf(buf, "%lu\n", get_effective_min_freq(df));
> +	mutex_unlock(&df->lock);
>  
> -	return sprintf(buf, "%lu\n", max(df->scaling_min_freq, df->min_freq));
> +	return ret;
>  }
>  
>  static ssize_t max_freq_store(struct device *dev, struct device_attribute *attr,
>  			      const char *buf, size_t count)
>  {
> @@ -1357,12 +1437,17 @@ static DEVICE_ATTR_RW(min_freq);
>  
>  static ssize_t max_freq_show(struct device *dev, struct device_attribute *attr,
>  			     char *buf)
>  {
>  	struct devfreq *df = to_devfreq(dev);
> +	ssize_t ret;
>  
> -	return sprintf(buf, "%lu\n", min(df->scaling_max_freq, df->max_freq));
> +	mutex_lock(&df->lock);
> +	ret = sprintf(buf, "%lu\n", get_effective_max_freq(df));
> +	mutex_unlock(&df->lock);
> +
> +	return ret;
>  }
>  static DEVICE_ATTR_RW(max_freq);
>  
>  static ssize_t available_frequencies_show(struct device *d,
>  					  struct device_attribute *attr,
> diff --git a/include/linux/devfreq.h b/include/linux/devfreq.h
> index 2bae9ed3c783..8b92ccbd1962 100644
> --- a/include/linux/devfreq.h
> +++ b/include/linux/devfreq.h
> @@ -134,10 +134,12 @@ struct devfreq_dev_profile {
>   * @total_trans:	Number of devfreq transitions
>   * @trans_table:	Statistics of devfreq transitions
>   * @time_in_state:	Statistics of devfreq states
>   * @last_stat_updated:	The last time stat updated
>   * @transition_notifier_list: list head of DEVFREQ_TRANSITION_NOTIFIER notifier
> + * @nb_min:		Notifier block for DEV_PM_QOS_MIN_FREQUENCY
> + * @nb_max:		Notifier block for DEV_PM_QOS_MAX_FREQUENCY
>   *
>   * This structure stores the devfreq information for a give device.
>   *
>   * Note that when a governor accesses entries in struct devfreq in its
>   * functions except for the context of callbacks defined in struct
> @@ -176,10 +178,13 @@ struct devfreq {
>  	unsigned int *trans_table;
>  	unsigned long *time_in_state;
>  	unsigned long last_stat_updated;
>  
>  	struct srcu_notifier_head transition_notifier_list;
> +
> +	struct notifier_block nb_min;
> +	struct notifier_block nb_max;
>  };
>  
>  struct devfreq_freqs {
>  	unsigned long old;
>  	unsigned long new;
> 


-- 
Best Regards,
Chanwoo Choi
Samsung Electronics

  reply	other threads:[~2019-08-21  1:40 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-08-20 15:23 [PATCH v3 0/2] PM / devfreq: Add dev_pm_qos support Leonard Crestez
2019-08-20 15:24 ` [PATCH v3 1/2] " Leonard Crestez
2019-08-21  1:44   ` Chanwoo Choi [this message]
2019-08-21 13:00     ` Leonard Crestez
2019-08-22 10:14       ` Chanwoo Choi
2019-08-20 15:24 ` [PATCH v3 2/2] PM / devfreq: Use dev_pm_qos for sysfs min/max_freq Leonard Crestez
2019-08-21  2:04   ` Chanwoo Choi
2019-08-21 13:03     ` Leonard Crestez
2019-08-22 10:07       ` Chanwoo Choi
2019-08-22 10:58         ` Leonard Crestez
2019-08-22 11:10           ` Chanwoo Choi
2019-08-22 12:24             ` Leonard Crestez
2019-08-23  8:40               ` Chanwoo Choi
2019-08-23 11:47                 ` Leonard Crestez

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=6134bb9e-9a16-b432-c191-c91f93844319@samsung.com \
    --to=cw00.choi@samsung.com \
    --cc=a.swigon@partner.samsung.com \
    --cc=abailon@baylibre.com \
    --cc=cpgs@samsung.com \
    --cc=georgi.djakov@linaro.org \
    --cc=krzk@kernel.org \
    --cc=kyungmin.park@samsung.com \
    --cc=leonard.crestez@nxp.com \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-pm@vger.kernel.org \
    --cc=myungjoo.ham@samsung.com \
    --cc=ping.bai@nxp.com \
    --cc=saravanak@google.com \
    --cc=viresh.kumar@linaro.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).