[v6,1/6] thermal: add generic cpufreq cooling implementation

Commit Message

Amit Kachhap Aug. 16, 2012, 11:41 a.m. UTC

This patchset introduces a new generic cooling device based on cpufreq
that can be used on non-ACPI platforms.  As a proof of concept, we have
drivers for the following platforms using this mechanism now:

 * Samsung Exynos (Exynos4 and Exynos5) in the current patchset.
 * Freescale i.MX (git://git.linaro.org/people/amitdanielk/linux.git imx6q_thermal)

There is a small change in cpufreq cooling registration APIs, so a minor
change is needed for Freescale platforms.

Brief Description:

1) The generic cooling devices code is placed inside driver/thermal/*
   as placing inside acpi folder will need un-necessary enabling of acpi
   code.  This code is architecture independent.

2) This patchset adds generic cpu cooling low level implementation
   through frequency clipping.  In future, other cpu related cooling
   devices may be added here.  An ACPI version of this already exists
   (drivers/acpi/processor_thermal.c) .But this will be useful for
   platforms like ARM using the generic thermal interface along with the
   generic cpu cooling devices.  The cooling device registration API's
   return cooling device pointers which can be easily binded with the
   thermal zone trip points.  The important APIs exposed are,

   a) struct thermal_cooling_device *cpufreq_cooling_register(
        struct cpumask *clip_cpus)
   b) void cpufreq_cooling_unregister(struct thermal_cooling_device *cdev)

3) Samsung exynos platform thermal implementation is done using the
   generic cpu cooling APIs and the new trip type.  The temperature sensor
   driver present in the hwmon folder(registered as hwmon driver) is moved
   to thermal folder and registered as a thermal driver.

A simple data/control flow diagrams is shown below,

Core Linux thermal <----->  Exynos thermal interface <----- Temperature Sensor
          |                             |
         \|/                            |
  Cpufreq cooling device <---------------

TODO:
*Will send the DT enablement patches later after the driver is merged.

This patch:

Add support for generic cpu thermal cooling low level implementations
using frequency scaling up/down based on the registration parameters.
Different cpu related cooling devices can be registered by the user and
the binding of these cooling devices to the corresponding trip points can
be easily done as the registration APIs return the cooling device pointer.
The user of these APIs are responsible for passing clipping frequency .
The drivers can also register to recieve notification about any cooling
action called.

[[e-mail]: fix comment layout]
Signed-off-by: Amit Daniel Kachhap <[e-mail]>
Cc: Guenter Roeck <[e-mail]>
Cc: SangWook Ju <[e-mail]>
Cc: Durgadoss <[e-mail]>
Cc: Len Brown <[e-mail]>
Cc: Jean Delvare <[e-mail]>
Cc: Kyungmin Park <[e-mail]>
Cc: Kukjin Kim <[e-mail]>
Signed-off-by: Andrew Morton <[e-mail]>
Signed-off-by: Amit Daniel Kachhap <[e-mail]>
---
 Documentation/thermal/cpu-cooling-api.txt |   52 +++
 drivers/thermal/Kconfig                   |   11 +
 drivers/thermal/Makefile                  |    1 +
 drivers/thermal/cpu_cooling.c             |  512 +++++++++++++++++++++++++++++
 include/linux/cpu_cooling.h               |   79 +++++
 5 files changed, 655 insertions(+), 0 deletions(-)
 create mode 100644 Documentation/thermal/cpu-cooling-api.txt
 create mode 100644 drivers/thermal/cpu_cooling.c
 create mode 100644 include/linux/cpu_cooling.h

Comments

Eduardo Valentin Aug. 16, 2012, 2:52 p.m. UTC | #1

Amit,

Thanks again for keeping this up.

On Thu, Aug 16, 2012 at 2:41 PM, Amit Daniel Kachhap
<[e-mail]> wrote:
> This patchset introduces a new generic cooling device based on cpufreq
> that can be used on non-ACPI platforms.  As a proof of concept, we have
> drivers for the following platforms using this mechanism now:
>
>  * Samsung Exynos (Exynos4 and Exynos5) in the current patchset.
>  * Freescale i.MX (git://git.linaro.org/people/amitdanielk/linux.git imx6q_thermal)

FYI, the OMAP code is under drivers/staging/omap-thermal/. The file
omap-thermal-common.c is the one which is using your generic cooling
device. But it needs to be updated accordingly to the API change you
mention.

>
> There is a small change in cpufreq cooling registration APIs, so a minor
> change is needed for Freescale platforms.
>
> Brief Description:
>
> 1) The generic cooling devices code is placed inside driver/thermal/*
>    as placing inside acpi folder will need un-necessary enabling of acpi
>    code.  This code is architecture independent.
>
> 2) This patchset adds generic cpu cooling low level implementation
>    through frequency clipping.  In future, other cpu related cooling
>    devices may be added here.  An ACPI version of this already exists
>    (drivers/acpi/processor_thermal.c) .But this will be useful for
>    platforms like ARM using the generic thermal interface along with the
>    generic cpu cooling devices.  The cooling device registration API's
>    return cooling device pointers which can be easily binded with the
>    thermal zone trip points.  The important APIs exposed are,
>
>    a) struct thermal_cooling_device *cpufreq_cooling_register(
>         struct cpumask *clip_cpus)
>    b) void cpufreq_cooling_unregister(struct thermal_cooling_device *cdev)
>
> 3) Samsung exynos platform thermal implementation is done using the
>    generic cpu cooling APIs and the new trip type.  The temperature sensor
>    driver present in the hwmon folder(registered as hwmon driver) is moved
>    to thermal folder and registered as a thermal driver.
>
> A simple data/control flow diagrams is shown below,
>
> Core Linux thermal <----->  Exynos thermal interface <----- Temperature Sensor
>           |                             |
>          \|/                            |
>   Cpufreq cooling device <---------------
>
> TODO:
> *Will send the DT enablement patches later after the driver is merged.
>
> This patch:
>
> Add support for generic cpu thermal cooling low level implementations
> using frequency scaling up/down based on the registration parameters.
> Different cpu related cooling devices can be registered by the user and
> the binding of these cooling devices to the corresponding trip points can
> be easily done as the registration APIs return the cooling device pointer.
> The user of these APIs are responsible for passing clipping frequency .
> The drivers can also register to recieve notification about any cooling
> action called.
>
> [[e-mail]: fix comment layout]
> Signed-off-by: Amit Daniel Kachhap <[e-mail]>
> Cc: Guenter Roeck <[e-mail]>
> Cc: SangWook Ju <[e-mail]>
> Cc: Durgadoss <[e-mail]>
> Cc: Len Brown <[e-mail]>
> Cc: Jean Delvare <[e-mail]>
> Cc: Kyungmin Park <[e-mail]>
> Cc: Kukjin Kim <[e-mail]>
> Signed-off-by: Andrew Morton <[e-mail]>
> Signed-off-by: Amit Daniel Kachhap <[e-mail]>


Reviewed-by: Eduardo Valentin <[e-mail]>


> ---
>  Documentation/thermal/cpu-cooling-api.txt |   52 +++
>  drivers/thermal/Kconfig                   |   11 +
>  drivers/thermal/Makefile                  |    1 +
>  drivers/thermal/cpu_cooling.c             |  512 +++++++++++++++++++++++++++++
>  include/linux/cpu_cooling.h               |   79 +++++
>  5 files changed, 655 insertions(+), 0 deletions(-)
>  create mode 100644 Documentation/thermal/cpu-cooling-api.txt
>  create mode 100644 drivers/thermal/cpu_cooling.c
>  create mode 100644 include/linux/cpu_cooling.h
>
> diff --git a/Documentation/thermal/cpu-cooling-api.txt b/Documentation/thermal/cpu-cooling-api.txt
> new file mode 100644
> index 0000000..a1f2a6b
> --- /dev/null
> +++ b/Documentation/thermal/cpu-cooling-api.txt
> @@ -0,0 +1,52 @@
> +CPU cooling APIs How To
> +===================================
> +
> +Written by Amit Daniel Kachhap <[e-mail]>
> +
> +Updated: 12 May 2012
> +
> +Copyright (c)  2012 Samsung Electronics Co., Ltd(http://www.samsung.com)
> +
> +0. Introduction
> +
> +The generic cpu cooling(freq clipping) provides registration/unregistration APIs
> +to the caller. The binding of the cooling devices to the trip point is left for
> +the user. The registration APIs returns the cooling device pointer.
> +
> +1. cpu cooling APIs
> +
> +1.1 cpufreq registration/unregistration APIs
> +1.1.1 struct thermal_cooling_device *cpufreq_cooling_register(
> +       struct cpumask *clip_cpus)
> +
> +    This interface function registers the cpufreq cooling device with the name
> +    "thermal-cpufreq-%x". This api can support multiple instances of cpufreq
> +    cooling devices.
> +
> +   clip_cpus: cpumask of cpus where the frequency constraints will happen.
> +
> +1.1.2 void cpufreq_cooling_unregister(struct thermal_cooling_device *cdev)
> +
> +    This interface function unregisters the "thermal-cpufreq-%x" cooling device.
> +
> +    cdev: Cooling device pointer which has to be unregistered.
> +
> +
> +1.2 CPU cooling action notifier register/unregister interface
> +1.2.1 int cputherm_register_notifier(struct notifier_block *nb,
> +       unsigned int list)
> +
> +    This interface registers a driver with cpu cooling layer. The driver will
> +    be notified when any cpu cooling action is called.
> +
> +    nb: notifier function to register
> +    list: CPUFREQ_COOLING_START or CPUFREQ_COOLING_STOP
> +
> +1.2.2 int cputherm_unregister_notifier(struct notifier_block *nb,
> +       unsigned int list)
> +
> +    This interface registers a driver with cpu cooling layer. The driver will
> +    be notified when any cpu cooling action is called.
> +
> +    nb: notifier function to register
> +    list: CPUFREQ_COOLING_START or CPUFREQ_COOLING_STOP
> diff --git a/drivers/thermal/Kconfig b/drivers/thermal/Kconfig
> index 7dd8c34..996003b 100644
> --- a/drivers/thermal/Kconfig
> +++ b/drivers/thermal/Kconfig
> @@ -19,6 +19,17 @@ config THERMAL_HWMON
>         depends on HWMON=y || HWMON=THERMAL
>         default y
>
> +config CPU_THERMAL
> +       bool "generic cpu cooling support"
> +       depends on THERMAL && CPU_FREQ
> +       help
> +         This implements the generic cpu cooling mechanism through frequency
> +         reduction, cpu hotplug and any other ways of reducing temperature. An
> +         ACPI version of this already exists(drivers/acpi/processor_thermal.c).
> +         This will be useful for platforms using the generic thermal interface
> +         and not the ACPI interface.
> +         If you want this support, you should say Y here.
> +
>  config SPEAR_THERMAL
>         bool "SPEAr thermal sensor driver"
>         depends on THERMAL
> diff --git a/drivers/thermal/Makefile b/drivers/thermal/Makefile
> index fd9369a..aae59ad 100644
> --- a/drivers/thermal/Makefile
> +++ b/drivers/thermal/Makefile
> @@ -3,5 +3,6 @@
>  #
>
>  obj-$(CONFIG_THERMAL)          += thermal_sys.o
> +obj-$(CONFIG_CPU_THERMAL)              += cpu_cooling.o
>  obj-$(CONFIG_SPEAR_THERMAL)            += spear_thermal.o
>  obj-$(CONFIG_RCAR_THERMAL)     += rcar_thermal.o
> diff --git a/drivers/thermal/cpu_cooling.c b/drivers/thermal/cpu_cooling.c
> new file mode 100644
> index 0000000..c42e557
> --- /dev/null
> +++ b/drivers/thermal/cpu_cooling.c
> @@ -0,0 +1,512 @@
> +/*
> + *  linux/drivers/thermal/cpu_cooling.c
> + *
> + *  Copyright (C) 2012 Samsung Electronics Co., Ltd(http://www.samsung.com)
> + *  Copyright (C) 2012  Amit Daniel <[e-mail]>
> + *
> + * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> + *  This program is free software; you can redistribute it and/or modify
> + *  it under the terms of the GNU General Public License as published by
> + *  the Free Software Foundation; version 2 of the License.
> + *
> + *  This program is distributed in the hope that it will be useful, but
> + *  WITHOUT ANY WARRANTY; without even the implied warranty of
> + *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
> + *  General Public License for more details.
> + *
> + *  You should have received a copy of the GNU General Public License along
> + *  with this program; if not, write to the Free Software Foundation, Inc.,
> + *  59 Temple Place, Suite 330, Boston, MA 02111-1307 USA.
> + *
> + * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> + */
> +#include <linux/kernel.h>
> +#include <linux/module.h>
> +#include <linux/thermal.h>
> +#include <linux/platform_device.h>
> +#include <linux/cpufreq.h>
> +#include <linux/err.h>
> +#include <linux/slab.h>
> +#include <linux/cpu.h>
> +#include <linux/cpu_cooling.h>
> +
> +/**
> + * struct cpufreq_cooling_device
> + * @id: unique integer value corresponding to each cpufreq_cooling_device
> + *     registered.
> + * @cool_dev: thermal_cooling_device pointer to keep track of the the
> + *     egistered cooling device.
> + * @cpufreq_state: integer value representing the current state of cpufreq
> + *     cooling devices.
> + * @cpufreq_val: integer value representing the absolute value of the clipped
> + *     frequency.
> + * @allowed_cpus: all the cpus involved for this cpufreq_cooling_device.
> + * @node: list_head to link all cpufreq_cooling_device together.
> + *
> + * This structure is required for keeping information of each
> + * cpufreq_cooling_device registered as a list whose head is represented by
> + * cooling_cpufreq_list. In order to prevent corruption of this list a
> + * mutex lock cooling_cpufreq_lock is used.
> + */
> +struct cpufreq_cooling_device {
> +       int id;
> +       struct thermal_cooling_device *cool_dev;
> +       unsigned int cpufreq_state;
> +       unsigned int cpufreq_val;
> +       struct cpumask allowed_cpus;
> +       struct list_head node;
> +};
> +static LIST_HEAD(cooling_cpufreq_list);
> +static DEFINE_IDR(cpufreq_idr);
> +
> +static struct mutex cooling_cpufreq_lock;
> +
> +/* notify_table passes value to the CPUFREQ_ADJUST callback function. */
> +#define NOTIFY_INVALID NULL
> +struct cpufreq_cooling_device *notify_device;
> +
> +/* Head of the blocking notifier chain to inform about frequency clamping */
> +static BLOCKING_NOTIFIER_HEAD(cputherm_state_notifier_list);
> +
> +/**
> + * get_idr - function to get a unique id.
> + * @idr: struct idr * handle used to create a id.
> + * @id: int * value generated by this function.
> + */
> +static int get_idr(struct idr *idr, int *id)
> +{
> +       int err;
> +again:
> +       if (unlikely(idr_pre_get(idr, GFP_KERNEL) == 0))
> +               return -ENOMEM;
> +
> +       mutex_lock(&cooling_cpufreq_lock);
> +       err = idr_get_new(idr, NULL, id);
> +       mutex_unlock(&cooling_cpufreq_lock);
> +
> +       if (unlikely(err == -EAGAIN))
> +               goto again;
> +       else if (unlikely(err))
> +               return err;
> +
> +       *id = *id & MAX_ID_MASK;
> +       return 0;
> +}
> +
> +/**
> + * release_idr - function to free the unique id.
> + * @idr: struct idr * handle used for creating the id.
> + * @id: int value representing the unique id.
> + */
> +static void release_idr(struct idr *idr, int id)
> +{
> +       mutex_lock(&cooling_cpufreq_lock);
> +       idr_remove(idr, id);
> +       mutex_unlock(&cooling_cpufreq_lock);
> +}
> +
> +/**
> + * cputherm_register_notifier - Register a notifier with cpu cooling interface.
> + * @nb:        struct notifier_block * with callback info.
> + * @list: integer value for which notification is needed. possible values are
> + *     CPUFREQ_COOLING_START and CPUFREQ_COOLING_STOP.
> + *
> + * This exported function registers a driver with cpu cooling layer. The driver
> + * will be notified when any cpu cooling action is called.
> + */
> +int cputherm_register_notifier(struct notifier_block *nb, unsigned int list)
> +{
> +       int ret = 0;
> +
> +       switch (list) {
> +       case CPUFREQ_COOLING_START:
> +       case CPUFREQ_COOLING_STOP:
> +               ret = blocking_notifier_chain_register(
> +                               &cputherm_state_notifier_list, nb);
> +               break;
> +       default:
> +               ret = -EINVAL;
> +       }
> +       return ret;
> +}
> +EXPORT_SYMBOL(cputherm_register_notifier);
> +
> +/**
> + * cputherm_unregister_notifier - Un-register a notifier.
> + * @nb:        struct notifier_block * with callback info.
> + * @list: integer value for which notification is needed. values possible are
> + *     CPUFREQ_COOLING_START or CPUFREQ_COOLING_STOP.
> + *
> + * This exported function un-registers a driver with cpu cooling layer.
> + */
> +int cputherm_unregister_notifier(struct notifier_block *nb, unsigned int list)
> +{
> +       int ret = 0;
> +
> +       switch (list) {
> +       case CPUFREQ_COOLING_START:
> +       case CPUFREQ_COOLING_STOP:
> +               ret = blocking_notifier_chain_unregister(
> +                               &cputherm_state_notifier_list, nb);
> +               break;
> +       default:
> +               ret = -EINVAL;
> +       }
> +       return ret;
> +}
> +EXPORT_SYMBOL(cputherm_unregister_notifier);
> +
> +/* Below code defines functions to be used for cpufreq as cooling device */
> +
> +/**
> + * is_cpufreq_valid - function to check if a cpu has frequency transition policy.
> + * @cpu: cpu for which check is needed.
> + */
> +static int is_cpufreq_valid(int cpu)
> +{
> +       struct cpufreq_policy policy;
> +       return !cpufreq_get_policy(&policy, cpu);
> +}
> +
> +/**
> + * get_cpu_frequency - get the absolute value of frequency from level.
> + * @cpu: cpu for which frequency is fetched.
> + * @level: level of frequency of the CPU
> + *     e.g level=1 --> 1st MAX FREQ, LEVEL=2 ---> 2nd MAX FREQ, .... etc
> + */
> +static unsigned int get_cpu_frequency(unsigned int cpu, unsigned long level)


I like this. With previous implemention I had to write a similar
function, but with output of the required freq  clip table. Now that
you have the freq walk here it simplifies the zone setup code.


> +{
> +       int ret = 0, i = 0;
> +       unsigned long level_index;
> +       bool descend = false;
> +       struct cpufreq_frequency_table *table =
> +                                       cpufreq_frequency_get_table(cpu);
> +       if (!table)
> +               return ret;
> +
> +       while (table[i].frequency != CPUFREQ_TABLE_END) {
> +               if (table[i].frequency == CPUFREQ_ENTRY_INVALID)
> +                       continue;
> +
> +               /*check if table in ascending or descending order*/
> +               if ((table[i + 1].frequency != CPUFREQ_TABLE_END) &&
> +                       (table[i + 1].frequency < table[i].frequency)
> +                       && !descend) {
> +                       descend = true;
> +               }
> +
> +               /*return if level matched and table in descending order*/
> +               if (descend && i == level)
> +                       return table[i].frequency;
> +               i++;
> +       }
> +       i--;
> +
> +       if (level > i || descend)
> +               return ret;
> +       level_index = i - level;
> +
> +       /*Scan the table in reverse order and match the level*/
> +       while (i >= 0) {
> +               if (table[i].frequency == CPUFREQ_ENTRY_INVALID)
> +                       continue;
> +               /*return if level matched*/
> +               if (i == level_index)
> +                       return table[i].frequency;
> +               i--;
> +       }
> +       return ret;
> +}
> +
> +/**
> + * cpufreq_apply_cooling - function to apply frequency clipping.
> + * @cpufreq_device: cpufreq_cooling_device pointer containing frequency
> + *     clipping data.
> + * @cooling_state: value of the cooling state.
> + */
> +static int cpufreq_apply_cooling(struct cpufreq_cooling_device *cpufreq_device,
> +                               unsigned long cooling_state)
> +{
> +       unsigned int event, cpuid, clip_freq;
> +       struct cpumask *maskPtr = &cpufreq_device->allowed_cpus;
> +       unsigned int cpu = cpumask_any(maskPtr);
> +
> +
> +       /* Check if the old cooling action is same as new cooling action */
> +       if (cpufreq_device->cpufreq_state == cooling_state)
> +               return 0;
> +
> +       clip_freq = get_cpu_frequency(cpu, cooling_state);
> +       if (!clip_freq)
> +               return -EINVAL;
> +
> +       cpufreq_device->cpufreq_state = cooling_state;
> +       cpufreq_device->cpufreq_val = clip_freq;
> +       notify_device = cpufreq_device;
> +
> +       if (cooling_state != 0)
> +               event = CPUFREQ_COOLING_START;
> +       else
> +               event = CPUFREQ_COOLING_STOP;
> +
> +       blocking_notifier_call_chain(&cputherm_state_notifier_list,
> +                                               event, &clip_freq);
> +
> +       for_each_cpu(cpuid, maskPtr) {
> +               if (is_cpufreq_valid(cpuid))
> +                       cpufreq_update_policy(cpuid);
> +       }
> +
> +       notify_device = NOTIFY_INVALID;
> +
> +       return 0;
> +}
> +
> +/**
> + * cpufreq_thermal_notifier - notifier callback for cpufreq policy change.
> + * @nb:        struct notifier_block * with callback info.
> + * @event: value showing cpufreq event for which this function invoked.
> + * @data: callback-specific data
> + */
> +static int cpufreq_thermal_notifier(struct notifier_block *nb,
> +                                       unsigned long event, void *data)
> +{
> +       struct cpufreq_policy *policy = data;
> +       unsigned long max_freq = 0;
> +
> +       if (event != CPUFREQ_ADJUST || notify_device == NOTIFY_INVALID)
> +               return 0;
> +
> +       if (cpumask_test_cpu(policy->cpu, &notify_device->allowed_cpus))
> +               max_freq = notify_device->cpufreq_val;
> +
> +       /* Never exceed user_policy.max*/
> +       if (max_freq > policy->user_policy.max)
> +               max_freq = policy->user_policy.max;
> +
> +       if (policy->max != max_freq)
> +               cpufreq_verify_within_limits(policy, 0, max_freq);
> +
> +       return 0;
> +}
> +
> +/*
> + * cpufreq cooling device callback functions are defined below
> + */
> +
> +/**
> + * cpufreq_get_max_state - callback function to get the max cooling state.
> + * @cdev: thermal cooling device pointer.
> + * @state: fill this variable with the max cooling state.
> + */
> +static int cpufreq_get_max_state(struct thermal_cooling_device *cdev,
> +                                unsigned long *state)
> +{
> +       int ret = -EINVAL, i = 0;
> +       struct cpufreq_cooling_device *cpufreq_device;
> +       struct cpumask *maskPtr;
> +       unsigned int cpu;
> +       struct cpufreq_frequency_table *table;
> +
> +       mutex_lock(&cooling_cpufreq_lock);
> +       list_for_each_entry(cpufreq_device, &cooling_cpufreq_list, node) {
> +               if (cpufreq_device && cpufreq_device->cool_dev == cdev)
> +                       break;
> +       }
> +       if (cpufreq_device == NULL)
> +               goto return_get_max_state;
> +
> +       maskPtr = &cpufreq_device->allowed_cpus;
> +       cpu = cpumask_any(maskPtr);
> +       table = cpufreq_frequency_get_table(cpu);
> +       if (!table) {
> +               *state = 0;
> +               ret = 0;
> +               goto return_get_max_state;
> +       }
> +
> +       while (table[i].frequency != CPUFREQ_TABLE_END) {
> +               if (table[i].frequency == CPUFREQ_ENTRY_INVALID)
> +                       continue;
> +               i++;
> +       }
> +       if (i > 0) {
> +               *state = --i;
> +               ret = 0;
> +       }
> +
> +return_get_max_state:
> +       mutex_unlock(&cooling_cpufreq_lock);
> +       return ret;
> +}
> +
> +/**
> + * cpufreq_get_cur_state - callback function to get the current cooling state.
> + * @cdev: thermal cooling device pointer.
> + * @state: fill this variable with the current cooling state.
> + */
> +static int cpufreq_get_cur_state(struct thermal_cooling_device *cdev,
> +                                unsigned long *state)
> +{
> +       int ret = -EINVAL;
> +       struct cpufreq_cooling_device *cpufreq_device;
> +
> +       mutex_lock(&cooling_cpufreq_lock);
> +       list_for_each_entry(cpufreq_device, &cooling_cpufreq_list, node) {
> +               if (cpufreq_device && cpufreq_device->cool_dev == cdev) {
> +                       *state = cpufreq_device->cpufreq_state;
> +                       ret = 0;
> +                       break;
> +               }
> +       }
> +       mutex_unlock(&cooling_cpufreq_lock);
> +
> +       return ret;
> +}
> +
> +/**
> + * cpufreq_set_cur_state - callback function to set the current cooling state.
> + * @cdev: thermal cooling device pointer.
> + * @state: set this variable to the current cooling state.
> + */
> +static int cpufreq_set_cur_state(struct thermal_cooling_device *cdev,
> +                                unsigned long state)
> +{
> +       int ret = -EINVAL;
> +       struct cpufreq_cooling_device *cpufreq_device;
> +
> +       mutex_lock(&cooling_cpufreq_lock);
> +       list_for_each_entry(cpufreq_device, &cooling_cpufreq_list, node) {
> +               if (cpufreq_device && cpufreq_device->cool_dev == cdev) {
> +                       ret = 0;
> +                       break;
> +               }
> +       }
> +       if (!ret)
> +               ret = cpufreq_apply_cooling(cpufreq_device, state);
> +
> +       mutex_unlock(&cooling_cpufreq_lock);
> +
> +       return ret;
> +}
> +
> +/* Bind cpufreq callbacks to thermal cooling device ops */
> +static struct thermal_cooling_device_ops const cpufreq_cooling_ops = {
> +       .get_max_state = cpufreq_get_max_state,
> +       .get_cur_state = cpufreq_get_cur_state,
> +       .set_cur_state = cpufreq_set_cur_state,
> +};
> +
> +/* Notifier for cpufreq policy change */
> +static struct notifier_block thermal_cpufreq_notifier_block = {
> +       .notifier_call = cpufreq_thermal_notifier,
> +};
> +
> +/**
> + * cpufreq_cooling_register - function to create cpufreq cooling device.
> + * @clip_cpus: cpumask of cpus where the frequency constraints will happen.
> + */
> +struct thermal_cooling_device *cpufreq_cooling_register(
> +       struct cpumask *clip_cpus)
> +{
> +       struct thermal_cooling_device *cool_dev;
> +       struct cpufreq_cooling_device *cpufreq_dev = NULL;
> +       unsigned int cpufreq_dev_count = 0, min = 0, max = 0;
> +       char dev_name[THERMAL_NAME_LENGTH];
> +       int ret = 0, id = 0, i;
> +       struct cpufreq_policy policy;
> +
> +       list_for_each_entry(cpufreq_dev, &cooling_cpufreq_list, node)
> +               cpufreq_dev_count++;
> +
> +       /*Verify that all the clip cpus have same freq_min, freq_max limit*/
> +       for_each_cpu(i, clip_cpus) {
> +               /*continue if cpufreq policy not found and not return error*/
> +               if (!cpufreq_get_policy(&policy, i))
> +                       continue;
> +               if (min == 0 && max == 0) {
> +                       min = policy.cpuinfo.min_freq;
> +                       max = policy.cpuinfo.max_freq;
> +               } else {
> +                       if (min != policy.cpuinfo.min_freq ||
> +                               max != policy.cpuinfo.max_freq)
> +                               return ERR_PTR(-EINVAL);
> +}
> +       }
> +       cpufreq_dev = kzalloc(sizeof(struct cpufreq_cooling_device),
> +                       GFP_KERNEL);
> +       if (!cpufreq_dev)
> +               return ERR_PTR(-ENOMEM);
> +
> +       cpumask_copy(&cpufreq_dev->allowed_cpus, clip_cpus);
> +
> +       if (cpufreq_dev_count == 0)
> +               mutex_init(&cooling_cpufreq_lock);
> +
> +       ret = get_idr(&cpufreq_idr, &cpufreq_dev->id);
> +       if (ret) {
> +               kfree(cpufreq_dev);
> +               return ERR_PTR(-EINVAL);
> +       }
> +
> +       sprintf(dev_name, "thermal-cpufreq-%d", cpufreq_dev->id);
> +
> +       cool_dev = thermal_cooling_device_register(dev_name, cpufreq_dev,
> +                                               &cpufreq_cooling_ops);
> +       if (!cool_dev) {
> +               release_idr(&cpufreq_idr, cpufreq_dev->id);
> +               kfree(cpufreq_dev);
> +               return ERR_PTR(-EINVAL);
> +       }
> +       cpufreq_dev->id = id;
> +       cpufreq_dev->cool_dev = cool_dev;
> +       cpufreq_dev->cpufreq_state = 0;
> +       mutex_lock(&cooling_cpufreq_lock);
> +       list_add_tail(&cpufreq_dev->node, &cooling_cpufreq_list);
> +
> +       /* Register the notifier for first cpufreq cooling device */
> +       if (cpufreq_dev_count == 0)
> +               cpufreq_register_notifier(&thermal_cpufreq_notifier_block,
> +                                               CPUFREQ_POLICY_NOTIFIER);
> +
> +       mutex_unlock(&cooling_cpufreq_lock);
> +       return cool_dev;
> +}
> +EXPORT_SYMBOL(cpufreq_cooling_register);
> +
> +/**
> + * cpufreq_cooling_unregister - function to remove cpufreq cooling device.
> + * @cdev: thermal cooling device pointer.
> + */
> +void cpufreq_cooling_unregister(struct thermal_cooling_device *cdev)
> +{
> +       struct cpufreq_cooling_device *cpufreq_dev = NULL;
> +       unsigned int cpufreq_dev_count = 0;
> +
> +       mutex_lock(&cooling_cpufreq_lock);
> +       list_for_each_entry(cpufreq_dev, &cooling_cpufreq_list, node) {
> +               if (cpufreq_dev && cpufreq_dev->cool_dev == cdev)
> +                       break;
> +               cpufreq_dev_count++;
> +       }
> +
> +       if (!cpufreq_dev || cpufreq_dev->cool_dev != cdev) {
> +               mutex_unlock(&cooling_cpufreq_lock);
> +               return;
> +       }
> +
> +       list_del(&cpufreq_dev->node);
> +
> +       /* Unregister the notifier for the last cpufreq cooling device */
> +       if (cpufreq_dev_count == 1) {
> +               cpufreq_unregister_notifier(&thermal_cpufreq_notifier_block,
> +                                       CPUFREQ_POLICY_NOTIFIER);
> +       }
> +       mutex_unlock(&cooling_cpufreq_lock);
> +       thermal_cooling_device_unregister(cpufreq_dev->cool_dev);
> +       release_idr(&cpufreq_idr, cpufreq_dev->id);
> +       if (cpufreq_dev_count == 1)
> +               mutex_destroy(&cooling_cpufreq_lock);
> +       kfree(cpufreq_dev);
> +}
> +EXPORT_SYMBOL(cpufreq_cooling_unregister);
> diff --git a/include/linux/cpu_cooling.h b/include/linux/cpu_cooling.h
> new file mode 100644
> index 0000000..555b917
> --- /dev/null
> +++ b/include/linux/cpu_cooling.h
> @@ -0,0 +1,79 @@
> +/*
> + *  linux/include/linux/cpu_cooling.h
> + *
> + *  Copyright (C) 2012 Samsung Electronics Co., Ltd(http://www.samsung.com)
> + *  Copyright (C) 2012  Amit Daniel <[e-mail]>
> + *
> + * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> + *  This program is free software; you can redistribute it and/or modify
> + *  it under the terms of the GNU General Public License as published by
> + *  the Free Software Foundation; version 2 of the License.
> + *
> + *  This program is distributed in the hope that it will be useful, but
> + *  WITHOUT ANY WARRANTY; without even the implied warranty of
> + *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
> + *  General Public License for more details.
> + *
> + *  You should have received a copy of the GNU General Public License along
> + *  with this program; if not, write to the Free Software Foundation, Inc.,
> + *  59 Temple Place, Suite 330, Boston, MA 02111-1307 USA.
> + *
> + * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> + */
> +
> +#ifndef __CPU_COOLING_H__
> +#define __CPU_COOLING_H__
> +
> +#include <linux/thermal.h>
> +
> +#define CPUFREQ_COOLING_START          0
> +#define CPUFREQ_COOLING_STOP           1
> +
> +/**
> + * cputherm_register_notifier - Register a notifier with cpu cooling interface.
> + * @nb:        struct notifier_block * with callback info.
> + * @list: integer value for which notification is needed. possible values are
> + *     CPUFREQ_COOLING_START and CPUFREQ_COOLING_STOP.
> + *
> + * This exported function registers a driver with cpu cooling layer. The driver
> + * will be notified when any cpu cooling action is called.
> + */
> +int cputherm_register_notifier(struct notifier_block *nb, unsigned int list);
> +
> +/**
> + * cputherm_unregister_notifier - Un-register a notifier.
> + * @nb:        struct notifier_block * with callback info.
> + * @list: integer value for which notification is needed. values possible are
> + *     CPUFREQ_COOLING_START and CPUFREQ_COOLING_STOP.
> + *
> + * This exported function un-registers a driver with cpu cooling layer.
> + */
> +int cputherm_unregister_notifier(struct notifier_block *nb, unsigned int list);
> +
> +#ifdef CONFIG_CPU_THERMAL
> +/**
> + * cpufreq_cooling_register - function to create cpufreq cooling device.
> + * @clip_cpus: cpumask of cpus where the frequency constraints will happen
> + */
> +struct thermal_cooling_device *cpufreq_cooling_register(
> +               struct cpumask *clip_cpus);
> +
> +/**
> + * cpufreq_cooling_unregister - function to remove cpufreq cooling device.
> + * @cdev: thermal cooling device pointer.
> + */
> +void cpufreq_cooling_unregister(struct thermal_cooling_device *cdev);
> +#else /* !CONFIG_CPU_THERMAL */
> +static inline struct thermal_cooling_device *cpufreq_cooling_register(
> +       struct cpumask *clip_cpus)
> +{
> +       return NULL;
> +}
> +static inline void cpufreq_cooling_unregister(
> +               struct thermal_cooling_device *cdev)
> +{
> +       return;
> +}
> +#endif /* CONFIG_CPU_THERMAL */
> +
> +#endif /* __CPU_COOLING_H__ */
> --
> 1.7.1
>

Amit Kachhap Aug. 17, 2012, 3:43 a.m. UTC | #2

On 16 August 2012 20:22, Valentin, Eduardo <[e-mail]> wrote:
>
> Amit,
>
> Thanks again for keeping this up.
>
> On Thu, Aug 16, 2012 at 2:41 PM, Amit Daniel Kachhap
> <[e-mail]> wrote:
> > This patchset introduces a new generic cooling device based on cpufreq
> > that can be used on non-ACPI platforms.  As a proof of concept, we have
> > drivers for the following platforms using this mechanism now:
> >
> >  * Samsung Exynos (Exynos4 and Exynos5) in the current patchset.
> >  * Freescale i.MX (git://git.linaro.org/people/amitdanielk/linux.git
> > imx6q_thermal)
>
> FYI, the OMAP code is under drivers/staging/omap-thermal/. The file
> omap-thermal-common.c is the one which is using your generic cooling
> device. But it needs to be updated accordingly to the API change you
> mention.
>
> >
> > There is a small change in cpufreq cooling registration APIs, so a minor
> > change is needed for Freescale platforms.
> >
> > Brief Description:
> >
> > 1) The generic cooling devices code is placed inside driver/thermal/*
> >    as placing inside acpi folder will need un-necessary enabling of acpi
> >    code.  This code is architecture independent.
> >
> > 2) This patchset adds generic cpu cooling low level implementation
> >    through frequency clipping.  In future, other cpu related cooling
> >    devices may be added here.  An ACPI version of this already exists
> >    (drivers/acpi/processor_thermal.c) .But this will be useful for
> >    platforms like ARM using the generic thermal interface along with the
> >    generic cpu cooling devices.  The cooling device registration API's
> >    return cooling device pointers which can be easily binded with the
> >    thermal zone trip points.  The important APIs exposed are,
> >
> >    a) struct thermal_cooling_device *cpufreq_cooling_register(
> >         struct cpumask *clip_cpus)
> >    b) void cpufreq_cooling_unregister(struct thermal_cooling_device
> > *cdev)
> >
> > 3) Samsung exynos platform thermal implementation is done using the
> >    generic cpu cooling APIs and the new trip type.  The temperature
> > sensor
> >    driver present in the hwmon folder(registered as hwmon driver) is
> > moved
> >    to thermal folder and registered as a thermal driver.
> >
> > A simple data/control flow diagrams is shown below,
> >
> > Core Linux thermal <----->  Exynos thermal interface <----- Temperature
> > Sensor
> >           |                             |
> >          \|/                            |
> >   Cpufreq cooling device <---------------
> >
> > TODO:
> > *Will send the DT enablement patches later after the driver is merged.
> >
> > This patch:
> >
> > Add support for generic cpu thermal cooling low level implementations
> > using frequency scaling up/down based on the registration parameters.
> > Different cpu related cooling devices can be registered by the user and
> > the binding of these cooling devices to the corresponding trip points
> > can
> > be easily done as the registration APIs return the cooling device
> > pointer.
> > The user of these APIs are responsible for passing clipping frequency .
> > The drivers can also register to recieve notification about any cooling
> > action called.
> >
> > [[e-mail]: fix comment layout]
> > Signed-off-by: Amit Daniel Kachhap <[e-mail]>
> > Cc: Guenter Roeck <[e-mail]>
> > Cc: SangWook Ju <[e-mail]>
> > Cc: Durgadoss <[e-mail]>
> > Cc: Len Brown <[e-mail]>
> > Cc: Jean Delvare <[e-mail]>
> > Cc: Kyungmin Park <[e-mail]>
> > Cc: Kukjin Kim <[e-mail]>
> > Signed-off-by: Andrew Morton <[e-mail]>
> > Signed-off-by: Amit Daniel Kachhap <[e-mail]>
>
>
> Reviewed-by: Eduardo Valentin <[e-mail]>

Thanks Eduardo .

>
>
> > ---
> >  Documentation/thermal/cpu-cooling-api.txt |   52 +++
> >  drivers/thermal/Kconfig                   |   11 +
> >  drivers/thermal/Makefile                  |    1 +
> >  drivers/thermal/cpu_cooling.c             |  512
> > +++++++++++++++++++++++++++++
> >  include/linux/cpu_cooling.h               |   79 +++++
> >  5 files changed, 655 insertions(+), 0 deletions(-)
> >  create mode 100644 Documentation/thermal/cpu-cooling-api.txt
> >  create mode 100644 drivers/thermal/cpu_cooling.c
> >  create mode 100644 include/linux/cpu_cooling.h
> >
> > diff --git a/Documentation/thermal/cpu-cooling-api.txt
> > b/Documentation/thermal/cpu-cooling-api.txt
> > new file mode 100644
> > index 0000000..a1f2a6b
> > --- /dev/null
> > +++ b/Documentation/thermal/cpu-cooling-api.txt
> > @@ -0,0 +1,52 @@
> > +CPU cooling APIs How To
> > +===================================
> > +
> > +Written by Amit Daniel Kachhap <[e-mail]>
> > +
> > +Updated: 12 May 2012
> > +
> > +Copyright (c)  2012 Samsung Electronics Co.,
> > Ltd(http://www.samsung.com)
> > +
> > +0. Introduction
> > +
> > +The generic cpu cooling(freq clipping) provides
> > registration/unregistration APIs
> > +to the caller. The binding of the cooling devices to the trip point is
> > left for
> > +the user. The registration APIs returns the cooling device pointer.
> > +
> > +1. cpu cooling APIs
> > +
> > +1.1 cpufreq registration/unregistration APIs
> > +1.1.1 struct thermal_cooling_device *cpufreq_cooling_register(
> > +       struct cpumask *clip_cpus)
> > +
> > +    This interface function registers the cpufreq cooling device with
> > the name
> > +    "thermal-cpufreq-%x". This api can support multiple instances of
> > cpufreq
> > +    cooling devices.
> > +
> > +   clip_cpus: cpumask of cpus where the frequency constraints will
> > happen.
> > +
> > +1.1.2 void cpufreq_cooling_unregister(struct thermal_cooling_device
> > *cdev)
> > +
> > +    This interface function unregisters the "thermal-cpufreq-%x"
> > cooling device.
> > +
> > +    cdev: Cooling device pointer which has to be unregistered.
> > +
> > +
> > +1.2 CPU cooling action notifier register/unregister interface
> > +1.2.1 int cputherm_register_notifier(struct notifier_block *nb,
> > +       unsigned int list)
> > +
> > +    This interface registers a driver with cpu cooling layer. The
> > driver will
> > +    be notified when any cpu cooling action is called.
> > +
> > +    nb: notifier function to register
> > +    list: CPUFREQ_COOLING_START or CPUFREQ_COOLING_STOP
> > +
> > +1.2.2 int cputherm_unregister_notifier(struct notifier_block *nb,
> > +       unsigned int list)
> > +
> > +    This interface registers a driver with cpu cooling layer. The
> > driver will
> > +    be notified when any cpu cooling action is called.
> > +
> > +    nb: notifier function to register
> > +    list: CPUFREQ_COOLING_START or CPUFREQ_COOLING_STOP
> > diff --git a/drivers/thermal/Kconfig b/drivers/thermal/Kconfig
> > index 7dd8c34..996003b 100644
> > --- a/drivers/thermal/Kconfig
> > +++ b/drivers/thermal/Kconfig
> > @@ -19,6 +19,17 @@ config THERMAL_HWMON
> >         depends on HWMON=y || HWMON=THERMAL
> >         default y
> >
> > +config CPU_THERMAL
> > +       bool "generic cpu cooling support"
> > +       depends on THERMAL && CPU_FREQ
> > +       help
> > +         This implements the generic cpu cooling mechanism through
> > frequency
> > +         reduction, cpu hotplug and any other ways of reducing
> > temperature. An
> > +         ACPI version of this already
> > exists(drivers/acpi/processor_thermal.c).
> > +         This will be useful for platforms using the generic thermal
> > interface
> > +         and not the ACPI interface.
> > +         If you want this support, you should say Y here.
> > +
> >  config SPEAR_THERMAL
> >         bool "SPEAr thermal sensor driver"
> >         depends on THERMAL
> > diff --git a/drivers/thermal/Makefile b/drivers/thermal/Makefile
> > index fd9369a..aae59ad 100644
> > --- a/drivers/thermal/Makefile
> > +++ b/drivers/thermal/Makefile
> > @@ -3,5 +3,6 @@
> >  #
> >
> >  obj-$(CONFIG_THERMAL)          += thermal_sys.o
> > +obj-$(CONFIG_CPU_THERMAL)              += cpu_cooling.o
> >  obj-$(CONFIG_SPEAR_THERMAL)            += spear_thermal.o
> >  obj-$(CONFIG_RCAR_THERMAL)     += rcar_thermal.o
> > diff --git a/drivers/thermal/cpu_cooling.c
> > b/drivers/thermal/cpu_cooling.c
> > new file mode 100644
> > index 0000000..c42e557
> > --- /dev/null
> > +++ b/drivers/thermal/cpu_cooling.c
> > @@ -0,0 +1,512 @@
> > +/*
> > + *  linux/drivers/thermal/cpu_cooling.c
> > + *
> > + *  Copyright (C) 2012 Samsung Electronics Co.,
> > Ltd(http://www.samsung.com)
> > + *  Copyright (C) 2012  Amit Daniel <[e-mail]>
> > + *
> > + *
> > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > + *  This program is free software; you can redistribute it and/or
> > modify
> > + *  it under the terms of the GNU General Public License as published
> > by
> > + *  the Free Software Foundation; version 2 of the License.
> > + *
> > + *  This program is distributed in the hope that it will be useful, but
> > + *  WITHOUT ANY WARRANTY; without even the implied warranty of
> > + *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
> > + *  General Public License for more details.
> > + *
> > + *  You should have received a copy of the GNU General Public License
> > along
> > + *  with this program; if not, write to the Free Software Foundation,
> > Inc.,
> > + *  59 Temple Place, Suite 330, Boston, MA 02111-1307 USA.
> > + *
> > + *
> > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > + */
> > +#include <linux/kernel.h>
> > +#include <linux/module.h>
> > +#include <linux/thermal.h>
> > +#include <linux/platform_device.h>
> > +#include <linux/cpufreq.h>
> > +#include <linux/err.h>
> > +#include <linux/slab.h>
> > +#include <linux/cpu.h>
> > +#include <linux/cpu_cooling.h>
> > +
> > +/**
> > + * struct cpufreq_cooling_device
> > + * @id: unique integer value corresponding to each
> > cpufreq_cooling_device
> > + *     registered.
> > + * @cool_dev: thermal_cooling_device pointer to keep track of the the
> > + *     egistered cooling device.
> > + * @cpufreq_state: integer value representing the current state of
> > cpufreq
> > + *     cooling devices.
> > + * @cpufreq_val: integer value representing the absolute value of the
> > clipped
> > + *     frequency.
> > + * @allowed_cpus: all the cpus involved for this
> > cpufreq_cooling_device.
> > + * @node: list_head to link all cpufreq_cooling_device together.
> > + *
> > + * This structure is required for keeping information of each
> > + * cpufreq_cooling_device registered as a list whose head is
> > represented by
> > + * cooling_cpufreq_list. In order to prevent corruption of this list a
> > + * mutex lock cooling_cpufreq_lock is used.
> > + */
> > +struct cpufreq_cooling_device {
> > +       int id;
> > +       struct thermal_cooling_device *cool_dev;
> > +       unsigned int cpufreq_state;
> > +       unsigned int cpufreq_val;
> > +       struct cpumask allowed_cpus;
> > +       struct list_head node;
> > +};
> > +static LIST_HEAD(cooling_cpufreq_list);
> > +static DEFINE_IDR(cpufreq_idr);
> > +
> > +static struct mutex cooling_cpufreq_lock;
> > +
> > +/* notify_table passes value to the CPUFREQ_ADJUST callback function.
> > */
> > +#define NOTIFY_INVALID NULL
> > +struct cpufreq_cooling_device *notify_device;
> > +
> > +/* Head of the blocking notifier chain to inform about frequency
> > clamping */
> > +static BLOCKING_NOTIFIER_HEAD(cputherm_state_notifier_list);
> > +
> > +/**
> > + * get_idr - function to get a unique id.
> > + * @idr: struct idr * handle used to create a id.
> > + * @id: int * value generated by this function.
> > + */
> > +static int get_idr(struct idr *idr, int *id)
> > +{
> > +       int err;
> > +again:
> > +       if (unlikely(idr_pre_get(idr, GFP_KERNEL) == 0))
> > +               return -ENOMEM;
> > +
> > +       mutex_lock(&cooling_cpufreq_lock);
> > +       err = idr_get_new(idr, NULL, id);
> > +       mutex_unlock(&cooling_cpufreq_lock);
> > +
> > +       if (unlikely(err == -EAGAIN))
> > +               goto again;
> > +       else if (unlikely(err))
> > +               return err;
> > +
> > +       *id = *id & MAX_ID_MASK;
> > +       return 0;
> > +}
> > +
> > +/**
> > + * release_idr - function to free the unique id.
> > + * @idr: struct idr * handle used for creating the id.
> > + * @id: int value representing the unique id.
> > + */
> > +static void release_idr(struct idr *idr, int id)
> > +{
> > +       mutex_lock(&cooling_cpufreq_lock);
> > +       idr_remove(idr, id);
> > +       mutex_unlock(&cooling_cpufreq_lock);
> > +}
> > +
> > +/**
> > + * cputherm_register_notifier - Register a notifier with cpu cooling
> > interface.
> > + * @nb:        struct notifier_block * with callback info.
> > + * @list: integer value for which notification is needed. possible
> > values are
> > + *     CPUFREQ_COOLING_START and CPUFREQ_COOLING_STOP.
> > + *
> > + * This exported function registers a driver with cpu cooling layer.
> > The driver
> > + * will be notified when any cpu cooling action is called.
> > + */
> > +int cputherm_register_notifier(struct notifier_block *nb, unsigned int
> > list)
> > +{
> > +       int ret = 0;
> > +
> > +       switch (list) {
> > +       case CPUFREQ_COOLING_START:
> > +       case CPUFREQ_COOLING_STOP:
> > +               ret = blocking_notifier_chain_register(
> > +                               &cputherm_state_notifier_list, nb);
> > +               break;
> > +       default:
> > +               ret = -EINVAL;
> > +       }
> > +       return ret;
> > +}
> > +EXPORT_SYMBOL(cputherm_register_notifier);
> > +
> > +/**
> > + * cputherm_unregister_notifier - Un-register a notifier.
> > + * @nb:        struct notifier_block * with callback info.
> > + * @list: integer value for which notification is needed. values
> > possible are
> > + *     CPUFREQ_COOLING_START or CPUFREQ_COOLING_STOP.
> > + *
> > + * This exported function un-registers a driver with cpu cooling layer.
> > + */
> > +int cputherm_unregister_notifier(struct notifier_block *nb, unsigned
> > int list)
> > +{
> > +       int ret = 0;
> > +
> > +       switch (list) {
> > +       case CPUFREQ_COOLING_START:
> > +       case CPUFREQ_COOLING_STOP:
> > +               ret = blocking_notifier_chain_unregister(
> > +                               &cputherm_state_notifier_list, nb);
> > +               break;
> > +       default:
> > +               ret = -EINVAL;
> > +       }
> > +       return ret;
> > +}
> > +EXPORT_SYMBOL(cputherm_unregister_notifier);
> > +
> > +/* Below code defines functions to be used for cpufreq as cooling
> > device */
> > +
> > +/**
> > + * is_cpufreq_valid - function to check if a cpu has frequency
> > transition policy.
> > + * @cpu: cpu for which check is needed.
> > + */
> > +static int is_cpufreq_valid(int cpu)
> > +{
> > +       struct cpufreq_policy policy;
> > +       return !cpufreq_get_policy(&policy, cpu);
> > +}
> > +
> > +/**
> > + * get_cpu_frequency - get the absolute value of frequency from level.
> > + * @cpu: cpu for which frequency is fetched.
> > + * @level: level of frequency of the CPU
> > + *     e.g level=1 --> 1st MAX FREQ, LEVEL=2 ---> 2nd MAX FREQ, ....
> > etc
> > + */
> > +static unsigned int get_cpu_frequency(unsigned int cpu, unsigned long
> > level)
>
>
> I like this. With previous implemention I had to write a similar
> function, but with output of the required freq  clip table. Now that
> you have the freq walk here it simplifies the zone setup code.
>
>
> > +{
> > +       int ret = 0, i = 0;
> > +       unsigned long level_index;
> > +       bool descend = false;
> > +       struct cpufreq_frequency_table *table =
> > +
> > cpufreq_frequency_get_table(cpu);
> > +       if (!table)
> > +               return ret;
> > +
> > +       while (table[i].frequency != CPUFREQ_TABLE_END) {
> > +               if (table[i].frequency == CPUFREQ_ENTRY_INVALID)
> > +                       continue;
> > +
> > +               /*check if table in ascending or descending order*/
> > +               if ((table[i + 1].frequency != CPUFREQ_TABLE_END) &&
> > +                       (table[i + 1].frequency < table[i].frequency)
> > +                       && !descend) {
> > +                       descend = true;
> > +               }
> > +
> > +               /*return if level matched and table in descending
> > order*/
> > +               if (descend && i == level)
> > +                       return table[i].frequency;
> > +               i++;
> > +       }
> > +       i--;
> > +
> > +       if (level > i || descend)
> > +               return ret;
> > +       level_index = i - level;
> > +
> > +       /*Scan the table in reverse order and match the level*/
> > +       while (i >= 0) {
> > +               if (table[i].frequency == CPUFREQ_ENTRY_INVALID)
> > +                       continue;
> > +               /*return if level matched*/
> > +               if (i == level_index)
> > +                       return table[i].frequency;
> > +               i--;
> > +       }
> > +       return ret;
> > +}
> > +
> > +/**
> > + * cpufreq_apply_cooling - function to apply frequency clipping.
> > + * @cpufreq_device: cpufreq_cooling_device pointer containing frequency
> > + *     clipping data.
> > + * @cooling_state: value of the cooling state.
> > + */
> > +static int cpufreq_apply_cooling(struct cpufreq_cooling_device
> > *cpufreq_device,
> > +                               unsigned long cooling_state)
> > +{
> > +       unsigned int event, cpuid, clip_freq;
> > +       struct cpumask *maskPtr = &cpufreq_device->allowed_cpus;
> > +       unsigned int cpu = cpumask_any(maskPtr);
> > +
> > +
> > +       /* Check if the old cooling action is same as new cooling action
> > */
> > +       if (cpufreq_device->cpufreq_state == cooling_state)
> > +               return 0;
> > +
> > +       clip_freq = get_cpu_frequency(cpu, cooling_state);
> > +       if (!clip_freq)
> > +               return -EINVAL;
> > +
> > +       cpufreq_device->cpufreq_state = cooling_state;
> > +       cpufreq_device->cpufreq_val = clip_freq;
> > +       notify_device = cpufreq_device;
> > +
> > +       if (cooling_state != 0)
> > +               event = CPUFREQ_COOLING_START;
> > +       else
> > +               event = CPUFREQ_COOLING_STOP;
> > +
> > +       blocking_notifier_call_chain(&cputherm_state_notifier_list,
> > +                                               event, &clip_freq);
> > +
> > +       for_each_cpu(cpuid, maskPtr) {
> > +               if (is_cpufreq_valid(cpuid))
> > +                       cpufreq_update_policy(cpuid);
> > +       }
> > +
> > +       notify_device = NOTIFY_INVALID;
> > +
> > +       return 0;
> > +}
> > +
> > +/**
> > + * cpufreq_thermal_notifier - notifier callback for cpufreq policy
> > change.
> > + * @nb:        struct notifier_block * with callback info.
> > + * @event: value showing cpufreq event for which this function invoked.
> > + * @data: callback-specific data
> > + */
> > +static int cpufreq_thermal_notifier(struct notifier_block *nb,
> > +                                       unsigned long event, void *data)
> > +{
> > +       struct cpufreq_policy *policy = data;
> > +       unsigned long max_freq = 0;
> > +
> > +       if (event != CPUFREQ_ADJUST || notify_device == NOTIFY_INVALID)
> > +               return 0;
> > +
> > +       if (cpumask_test_cpu(policy->cpu, &notify_device->allowed_cpus))
> > +               max_freq = notify_device->cpufreq_val;
> > +
> > +       /* Never exceed user_policy.max*/
> > +       if (max_freq > policy->user_policy.max)
> > +               max_freq = policy->user_policy.max;
> > +
> > +       if (policy->max != max_freq)
> > +               cpufreq_verify_within_limits(policy, 0, max_freq);
> > +
> > +       return 0;
> > +}
> > +
> > +/*
> > + * cpufreq cooling device callback functions are defined below
> > + */
> > +
> > +/**
> > + * cpufreq_get_max_state - callback function to get the max cooling
> > state.
> > + * @cdev: thermal cooling device pointer.
> > + * @state: fill this variable with the max cooling state.
> > + */
> > +static int cpufreq_get_max_state(struct thermal_cooling_device *cdev,
> > +                                unsigned long *state)
> > +{
> > +       int ret = -EINVAL, i = 0;
> > +       struct cpufreq_cooling_device *cpufreq_device;
> > +       struct cpumask *maskPtr;
> > +       unsigned int cpu;
> > +       struct cpufreq_frequency_table *table;
> > +
> > +       mutex_lock(&cooling_cpufreq_lock);
> > +       list_for_each_entry(cpufreq_device, &cooling_cpufreq_list, node)
> > {
> > +               if (cpufreq_device && cpufreq_device->cool_dev == cdev)
> > +                       break;
> > +       }
> > +       if (cpufreq_device == NULL)
> > +               goto return_get_max_state;
> > +
> > +       maskPtr = &cpufreq_device->allowed_cpus;
> > +       cpu = cpumask_any(maskPtr);
> > +       table = cpufreq_frequency_get_table(cpu);
> > +       if (!table) {
> > +               *state = 0;
> > +               ret = 0;
> > +               goto return_get_max_state;
> > +       }
> > +
> > +       while (table[i].frequency != CPUFREQ_TABLE_END) {
> > +               if (table[i].frequency == CPUFREQ_ENTRY_INVALID)
> > +                       continue;
> > +               i++;
> > +       }
> > +       if (i > 0) {
> > +               *state = --i;
> > +               ret = 0;
> > +       }
> > +
> > +return_get_max_state:
> > +       mutex_unlock(&cooling_cpufreq_lock);
> > +       return ret;
> > +}
> > +
> > +/**
> > + * cpufreq_get_cur_state - callback function to get the current cooling
> > state.
> > + * @cdev: thermal cooling device pointer.
> > + * @state: fill this variable with the current cooling state.
> > + */
> > +static int cpufreq_get_cur_state(struct thermal_cooling_device *cdev,
> > +                                unsigned long *state)
> > +{
> > +       int ret = -EINVAL;
> > +       struct cpufreq_cooling_device *cpufreq_device;
> > +
> > +       mutex_lock(&cooling_cpufreq_lock);
> > +       list_for_each_entry(cpufreq_device, &cooling_cpufreq_list, node)
> > {
> > +               if (cpufreq_device && cpufreq_device->cool_dev == cdev)
> > {
> > +                       *state = cpufreq_device->cpufreq_state;
> > +                       ret = 0;
> > +                       break;
> > +               }
> > +       }
> > +       mutex_unlock(&cooling_cpufreq_lock);
> > +
> > +       return ret;
> > +}
> > +
> > +/**
> > + * cpufreq_set_cur_state - callback function to set the current cooling
> > state.
> > + * @cdev: thermal cooling device pointer.
> > + * @state: set this variable to the current cooling state.
> > + */
> > +static int cpufreq_set_cur_state(struct thermal_cooling_device *cdev,
> > +                                unsigned long state)
> > +{
> > +       int ret = -EINVAL;
> > +       struct cpufreq_cooling_device *cpufreq_device;
> > +
> > +       mutex_lock(&cooling_cpufreq_lock);
> > +       list_for_each_entry(cpufreq_device, &cooling_cpufreq_list, node)
> > {
> > +               if (cpufreq_device && cpufreq_device->cool_dev == cdev)
> > {
> > +                       ret = 0;
> > +                       break;
> > +               }
> > +       }
> > +       if (!ret)
> > +               ret = cpufreq_apply_cooling(cpufreq_device, state);
> > +
> > +       mutex_unlock(&cooling_cpufreq_lock);
> > +
> > +       return ret;
> > +}
> > +
> > +/* Bind cpufreq callbacks to thermal cooling device ops */
> > +static struct thermal_cooling_device_ops const cpufreq_cooling_ops = {
> > +       .get_max_state = cpufreq_get_max_state,
> > +       .get_cur_state = cpufreq_get_cur_state,
> > +       .set_cur_state = cpufreq_set_cur_state,
> > +};
> > +
> > +/* Notifier for cpufreq policy change */
> > +static struct notifier_block thermal_cpufreq_notifier_block = {
> > +       .notifier_call = cpufreq_thermal_notifier,
> > +};
> > +
> > +/**
> > + * cpufreq_cooling_register - function to create cpufreq cooling
> > device.
> > + * @clip_cpus: cpumask of cpus where the frequency constraints will
> > happen.
> > + */
> > +struct thermal_cooling_device *cpufreq_cooling_register(
> > +       struct cpumask *clip_cpus)
> > +{
> > +       struct thermal_cooling_device *cool_dev;
> > +       struct cpufreq_cooling_device *cpufreq_dev = NULL;
> > +       unsigned int cpufreq_dev_count = 0, min = 0, max = 0;
> > +       char dev_name[THERMAL_NAME_LENGTH];
> > +       int ret = 0, id = 0, i;
> > +       struct cpufreq_policy policy;
> > +
> > +       list_for_each_entry(cpufreq_dev, &cooling_cpufreq_list, node)
> > +               cpufreq_dev_count++;
> > +
> > +       /*Verify that all the clip cpus have same freq_min, freq_max
> > limit*/
> > +       for_each_cpu(i, clip_cpus) {
> > +               /*continue if cpufreq policy not found and not return
> > error*/
> > +               if (!cpufreq_get_policy(&policy, i))
> > +                       continue;
> > +               if (min == 0 && max == 0) {
> > +                       min = policy.cpuinfo.min_freq;
> > +                       max = policy.cpuinfo.max_freq;
> > +               } else {
> > +                       if (min != policy.cpuinfo.min_freq ||
> > +                               max != policy.cpuinfo.max_freq)
> > +                               return ERR_PTR(-EINVAL);
> > +}
> > +       }
> > +       cpufreq_dev = kzalloc(sizeof(struct cpufreq_cooling_device),
> > +                       GFP_KERNEL);
> > +       if (!cpufreq_dev)
> > +               return ERR_PTR(-ENOMEM);
> > +
> > +       cpumask_copy(&cpufreq_dev->allowed_cpus, clip_cpus);
> > +
> > +       if (cpufreq_dev_count == 0)
> > +               mutex_init(&cooling_cpufreq_lock);
> > +
> > +       ret = get_idr(&cpufreq_idr, &cpufreq_dev->id);
> > +       if (ret) {
> > +               kfree(cpufreq_dev);
> > +               return ERR_PTR(-EINVAL);
> > +       }
> > +
> > +       sprintf(dev_name, "thermal-cpufreq-%d", cpufreq_dev->id);
> > +
> > +       cool_dev = thermal_cooling_device_register(dev_name,
> > cpufreq_dev,
> > +                                               &cpufreq_cooling_ops);
> > +       if (!cool_dev) {
> > +               release_idr(&cpufreq_idr, cpufreq_dev->id);
> > +               kfree(cpufreq_dev);
> > +               return ERR_PTR(-EINVAL);
> > +       }
> > +       cpufreq_dev->id = id;
> > +       cpufreq_dev->cool_dev = cool_dev;
> > +       cpufreq_dev->cpufreq_state = 0;
> > +       mutex_lock(&cooling_cpufreq_lock);
> > +       list_add_tail(&cpufreq_dev->node, &cooling_cpufreq_list);
> > +
> > +       /* Register the notifier for first cpufreq cooling device */
> > +       if (cpufreq_dev_count == 0)
> > +
> > cpufreq_register_notifier(&thermal_cpufreq_notifier_block,
> > +
> > CPUFREQ_POLICY_NOTIFIER);
> > +
> > +       mutex_unlock(&cooling_cpufreq_lock);
> > +       return cool_dev;
> > +}
> > +EXPORT_SYMBOL(cpufreq_cooling_register);
> > +
> > +/**
> > + * cpufreq_cooling_unregister - function to remove cpufreq cooling
> > device.
> > + * @cdev: thermal cooling device pointer.
> > + */
> > +void cpufreq_cooling_unregister(struct thermal_cooling_device *cdev)
> > +{
> > +       struct cpufreq_cooling_device *cpufreq_dev = NULL;
> > +       unsigned int cpufreq_dev_count = 0;
> > +
> > +       mutex_lock(&cooling_cpufreq_lock);
> > +       list_for_each_entry(cpufreq_dev, &cooling_cpufreq_list, node) {
> > +               if (cpufreq_dev && cpufreq_dev->cool_dev == cdev)
> > +                       break;
> > +               cpufreq_dev_count++;
> > +       }
> > +
> > +       if (!cpufreq_dev || cpufreq_dev->cool_dev != cdev) {
> > +               mutex_unlock(&cooling_cpufreq_lock);
> > +               return;
> > +       }
> > +
> > +       list_del(&cpufreq_dev->node);
> > +
> > +       /* Unregister the notifier for the last cpufreq cooling device
> > */
> > +       if (cpufreq_dev_count == 1) {
> > +
> > cpufreq_unregister_notifier(&thermal_cpufreq_notifier_block,
> > +                                       CPUFREQ_POLICY_NOTIFIER);
> > +       }
> > +       mutex_unlock(&cooling_cpufreq_lock);
> > +       thermal_cooling_device_unregister(cpufreq_dev->cool_dev);
> > +       release_idr(&cpufreq_idr, cpufreq_dev->id);
> > +       if (cpufreq_dev_count == 1)
> > +               mutex_destroy(&cooling_cpufreq_lock);
> > +       kfree(cpufreq_dev);
> > +}
> > +EXPORT_SYMBOL(cpufreq_cooling_unregister);
> > diff --git a/include/linux/cpu_cooling.h b/include/linux/cpu_cooling.h
> > new file mode 100644
> > index 0000000..555b917
> > --- /dev/null
> > +++ b/include/linux/cpu_cooling.h
> > @@ -0,0 +1,79 @@
> > +/*
> > + *  linux/include/linux/cpu_cooling.h
> > + *
> > + *  Copyright (C) 2012 Samsung Electronics Co.,
> > Ltd(http://www.samsung.com)
> > + *  Copyright (C) 2012  Amit Daniel <[e-mail]>
> > + *
> > + *
> > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > + *  This program is free software; you can redistribute it and/or
> > modify
> > + *  it under the terms of the GNU General Public License as published
> > by
> > + *  the Free Software Foundation; version 2 of the License.
> > + *
> > + *  This program is distributed in the hope that it will be useful, but
> > + *  WITHOUT ANY WARRANTY; without even the implied warranty of
> > + *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
> > + *  General Public License for more details.
> > + *
> > + *  You should have received a copy of the GNU General Public License
> > along
> > + *  with this program; if not, write to the Free Software Foundation,
> > Inc.,
> > + *  59 Temple Place, Suite 330, Boston, MA 02111-1307 USA.
> > + *
> > + *
> > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > + */
> > +
> > +#ifndef __CPU_COOLING_H__
> > +#define __CPU_COOLING_H__
> > +
> > +#include <linux/thermal.h>
> > +
> > +#define CPUFREQ_COOLING_START          0
> > +#define CPUFREQ_COOLING_STOP           1
> > +
> > +/**
> > + * cputherm_register_notifier - Register a notifier with cpu cooling
> > interface.
> > + * @nb:        struct notifier_block * with callback info.
> > + * @list: integer value for which notification is needed. possible
> > values are
> > + *     CPUFREQ_COOLING_START and CPUFREQ_COOLING_STOP.
> > + *
> > + * This exported function registers a driver with cpu cooling layer.
> > The driver
> > + * will be notified when any cpu cooling action is called.
> > + */
> > +int cputherm_register_notifier(struct notifier_block *nb, unsigned int
> > list);
> > +
> > +/**
> > + * cputherm_unregister_notifier - Un-register a notifier.
> > + * @nb:        struct notifier_block * with callback info.
> > + * @list: integer value for which notification is needed. values
> > possible are
> > + *     CPUFREQ_COOLING_START and CPUFREQ_COOLING_STOP.
> > + *
> > + * This exported function un-registers a driver with cpu cooling layer.
> > + */
> > +int cputherm_unregister_notifier(struct notifier_block *nb, unsigned
> > int list);
> > +
> > +#ifdef CONFIG_CPU_THERMAL
> > +/**
> > + * cpufreq_cooling_register - function to create cpufreq cooling
> > device.
> > + * @clip_cpus: cpumask of cpus where the frequency constraints will
> > happen
> > + */
> > +struct thermal_cooling_device *cpufreq_cooling_register(
> > +               struct cpumask *clip_cpus);
> > +
> > +/**
> > + * cpufreq_cooling_unregister - function to remove cpufreq cooling
> > device.
> > + * @cdev: thermal cooling device pointer.
> > + */
> > +void cpufreq_cooling_unregister(struct thermal_cooling_device *cdev);
> > +#else /* !CONFIG_CPU_THERMAL */
> > +static inline struct thermal_cooling_device *cpufreq_cooling_register(
> > +       struct cpumask *clip_cpus)
> > +{
> > +       return NULL;
> > +}
> > +static inline void cpufreq_cooling_unregister(
> > +               struct thermal_cooling_device *cdev)
> > +{
> > +       return;
> > +}
> > +#endif /* CONFIG_CPU_THERMAL */
> > +
> > +#endif /* __CPU_COOLING_H__ */
> > --
> > 1.7.1
> >
>
>
>
> --
>
> Eduardo Valentin
--
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
the body of a message to [e-mail]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Zhang Rui Aug. 17, 2012, 7:24 a.m. UTC | #3

On ?, 2012-08-16 at 17:11 +0530, Amit Daniel Kachhap wrote:
> This patchset introduces a new generic cooling device based on cpufreq
> that can be used on non-ACPI platforms.  As a proof of concept, we have
> drivers for the following platforms using this mechanism now:
> 
>  * Samsung Exynos (Exynos4 and Exynos5) in the current patchset.
>  * Freescale i.MX (git://git.linaro.org/people/amitdanielk/linux.git imx6q_thermal)
> 
> There is a small change in cpufreq cooling registration APIs, so a minor
> change is needed for Freescale platforms.
> 
> Brief Description:
> 
> 1) The generic cooling devices code is placed inside driver/thermal/*
>    as placing inside acpi folder will need un-necessary enabling of acpi
>    code.  This code is architecture independent.
> 
> 2) This patchset adds generic cpu cooling low level implementation
>    through frequency clipping.  In future, other cpu related cooling
>    devices may be added here.  An ACPI version of this already exists
>    (drivers/acpi/processor_thermal.c) .But this will be useful for
>    platforms like ARM using the generic thermal interface along with the
>    generic cpu cooling devices.  The cooling device registration API's
>    return cooling device pointers which can be easily binded with the
>    thermal zone trip points.  The important APIs exposed are,
> 
>    a) struct thermal_cooling_device *cpufreq_cooling_register(
>         struct cpumask *clip_cpus)
>    b) void cpufreq_cooling_unregister(struct thermal_cooling_device *cdev)
> 
> 3) Samsung exynos platform thermal implementation is done using the
>    generic cpu cooling APIs and the new trip type.  The temperature sensor
>    driver present in the hwmon folder(registered as hwmon driver) is moved
>    to thermal folder and registered as a thermal driver.
> 
> A simple data/control flow diagrams is shown below,
> 
> Core Linux thermal <----->  Exynos thermal interface <----- Temperature Sensor
>           |                             |
>          \|/                            |
>   Cpufreq cooling device <---------------
> 
> TODO:
> *Will send the DT enablement patches later after the driver is merged.
> 
> This patch:
> 
> Add support for generic cpu thermal cooling low level implementations
> using frequency scaling up/down based on the registration parameters.
> Different cpu related cooling devices can be registered by the user and
> the binding of these cooling devices to the corresponding trip points can
> be easily done as the registration APIs return the cooling device pointer.
> The user of these APIs are responsible for passing clipping frequency .
> The drivers can also register to recieve notification about any cooling
> action called.
> 
> [[e-mail]: fix comment layout]
> Signed-off-by: Amit Daniel Kachhap <[e-mail]>
> Cc: Guenter Roeck <[e-mail]>
> Cc: SangWook Ju <[e-mail]>
> Cc: Durgadoss <[e-mail]>
> Cc: Len Brown <[e-mail]>
> Cc: Jean Delvare <[e-mail]>
> Cc: Kyungmin Park <[e-mail]>
> Cc: Kukjin Kim <[e-mail]>
> Signed-off-by: Andrew Morton <[e-mail]>
> Signed-off-by: Amit Daniel Kachhap <[e-mail]>
> ---
>  Documentation/thermal/cpu-cooling-api.txt |   52 +++
>  drivers/thermal/Kconfig                   |   11 +
>  drivers/thermal/Makefile                  |    1 +
>  drivers/thermal/cpu_cooling.c             |  512 +++++++++++++++++++++++++++++
>  include/linux/cpu_cooling.h               |   79 +++++
>  5 files changed, 655 insertions(+), 0 deletions(-)
>  create mode 100644 Documentation/thermal/cpu-cooling-api.txt
>  create mode 100644 drivers/thermal/cpu_cooling.c
>  create mode 100644 include/linux/cpu_cooling.h
> 
> diff --git a/Documentation/thermal/cpu-cooling-api.txt b/Documentation/thermal/cpu-cooling-api.txt
> new file mode 100644
> index 0000000..a1f2a6b
> --- /dev/null
> +++ b/Documentation/thermal/cpu-cooling-api.txt
> @@ -0,0 +1,52 @@
> +CPU cooling APIs How To
> +===================================
> +
> +Written by Amit Daniel Kachhap <[e-mail]>
> +
> +Updated: 12 May 2012
> +
> +Copyright (c)  2012 Samsung Electronics Co., Ltd(http://www.samsung.com)
> +
> +0. Introduction
> +
> +The generic cpu cooling(freq clipping) provides registration/unregistration APIs
> +to the caller. The binding of the cooling devices to the trip point is left for
> +the user. The registration APIs returns the cooling device pointer.
> +
> +1. cpu cooling APIs
> +
> +1.1 cpufreq registration/unregistration APIs
> +1.1.1 struct thermal_cooling_device *cpufreq_cooling_register(
> +	struct cpumask *clip_cpus)
> +
> +    This interface function registers the cpufreq cooling device with the name
> +    "thermal-cpufreq-%x". This api can support multiple instances of cpufreq
> +    cooling devices.
> +
> +   clip_cpus: cpumask of cpus where the frequency constraints will happen.
> +
> +1.1.2 void cpufreq_cooling_unregister(struct thermal_cooling_device *cdev)
> +
> +    This interface function unregisters the "thermal-cpufreq-%x" cooling device.
> +
> +    cdev: Cooling device pointer which has to be unregistered.
> +
> +
> +1.2 CPU cooling action notifier register/unregister interface
> +1.2.1 int cputherm_register_notifier(struct notifier_block *nb,
> +	unsigned int list)
> +
> +    This interface registers a driver with cpu cooling layer. The driver will
> +    be notified when any cpu cooling action is called.
> +
> +    nb: notifier function to register
> +    list: CPUFREQ_COOLING_START or CPUFREQ_COOLING_STOP
> +
> +1.2.2 int cputherm_unregister_notifier(struct notifier_block *nb,
> +	unsigned int list)
> +
> +    This interface registers a driver with cpu cooling layer. The driver will
> +    be notified when any cpu cooling action is called.
> +
> +    nb: notifier function to register
> +    list: CPUFREQ_COOLING_START or CPUFREQ_COOLING_STOP

what are these two APIs used for?
I did not see they are used in your patch set, do I miss something?

> diff --git a/drivers/thermal/Kconfig b/drivers/thermal/Kconfig
> index 7dd8c34..996003b 100644
> --- a/drivers/thermal/Kconfig
> +++ b/drivers/thermal/Kconfig
> @@ -19,6 +19,17 @@ config THERMAL_HWMON
>  	depends on HWMON=y || HWMON=THERMAL
>  	default y
>  
> +config CPU_THERMAL
> +	bool "generic cpu cooling support"
> +	depends on THERMAL && CPU_FREQ
> +	help
> +	  This implements the generic cpu cooling mechanism through frequency
> +	  reduction, cpu hotplug and any other ways of reducing temperature. An
> +	  ACPI version of this already exists(drivers/acpi/processor_thermal.c).
> +	  This will be useful for platforms using the generic thermal interface
> +	  and not the ACPI interface.
> +	  If you want this support, you should say Y here.
> +
>  config SPEAR_THERMAL
>  	bool "SPEAr thermal sensor driver"
>  	depends on THERMAL
> diff --git a/drivers/thermal/Makefile b/drivers/thermal/Makefile
> index fd9369a..aae59ad 100644
> --- a/drivers/thermal/Makefile
> +++ b/drivers/thermal/Makefile
> @@ -3,5 +3,6 @@
>  #
>  
>  obj-$(CONFIG_THERMAL)		+= thermal_sys.o
> +obj-$(CONFIG_CPU_THERMAL)		+= cpu_cooling.o
>  obj-$(CONFIG_SPEAR_THERMAL)		+= spear_thermal.o
>  obj-$(CONFIG_RCAR_THERMAL)	+= rcar_thermal.o
> diff --git a/drivers/thermal/cpu_cooling.c b/drivers/thermal/cpu_cooling.c
> new file mode 100644
> index 0000000..c42e557
> --- /dev/null
> +++ b/drivers/thermal/cpu_cooling.c
> @@ -0,0 +1,512 @@
> +/*
> + *  linux/drivers/thermal/cpu_cooling.c
> + *
> + *  Copyright (C) 2012	Samsung Electronics Co., Ltd(http://www.samsung.com)
> + *  Copyright (C) 2012  Amit Daniel <[e-mail]>
> + *
> + * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> + *  This program is free software; you can redistribute it and/or modify
> + *  it under the terms of the GNU General Public License as published by
> + *  the Free Software Foundation; version 2 of the License.
> + *
> + *  This program is distributed in the hope that it will be useful, but
> + *  WITHOUT ANY WARRANTY; without even the implied warranty of
> + *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
> + *  General Public License for more details.
> + *
> + *  You should have received a copy of the GNU General Public License along
> + *  with this program; if not, write to the Free Software Foundation, Inc.,
> + *  59 Temple Place, Suite 330, Boston, MA 02111-1307 USA.
> + *
> + * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> + */
> +#include <linux/kernel.h>
> +#include <linux/module.h>
> +#include <linux/thermal.h>
> +#include <linux/platform_device.h>
> +#include <linux/cpufreq.h>
> +#include <linux/err.h>
> +#include <linux/slab.h>
> +#include <linux/cpu.h>
> +#include <linux/cpu_cooling.h>
> +
> +/**
> + * struct cpufreq_cooling_device
> + * @id: unique integer value corresponding to each cpufreq_cooling_device
> + *	registered.
> + * @cool_dev: thermal_cooling_device pointer to keep track of the the
> + *	egistered cooling device.
> + * @cpufreq_state: integer value representing the current state of cpufreq
> + *	cooling	devices.
> + * @cpufreq_val: integer value representing the absolute value of the clipped
> + *	frequency.
> + * @allowed_cpus: all the cpus involved for this cpufreq_cooling_device.
> + * @node: list_head to link all cpufreq_cooling_device together.
> + *
> + * This structure is required for keeping information of each
> + * cpufreq_cooling_device registered as a list whose head is represented by
> + * cooling_cpufreq_list. In order to prevent corruption of this list a
> + * mutex lock cooling_cpufreq_lock is used.
> + */
> +struct cpufreq_cooling_device {
> +	int id;
> +	struct thermal_cooling_device *cool_dev;
> +	unsigned int cpufreq_state;
> +	unsigned int cpufreq_val;
> +	struct cpumask allowed_cpus;
> +	struct list_head node;
> +};
> +static LIST_HEAD(cooling_cpufreq_list);
> +static DEFINE_IDR(cpufreq_idr);
> +
> +static struct mutex cooling_cpufreq_lock;
> +
> +/* notify_table passes value to the CPUFREQ_ADJUST callback function. */
> +#define NOTIFY_INVALID NULL
> +struct cpufreq_cooling_device *notify_device;
> +
> +/* Head of the blocking notifier chain to inform about frequency clamping */
> +static BLOCKING_NOTIFIER_HEAD(cputherm_state_notifier_list);
> +
> +/**
> + * get_idr - function to get a unique id.
> + * @idr: struct idr * handle used to create a id.
> + * @id: int * value generated by this function.
> + */
> +static int get_idr(struct idr *idr, int *id)
> +{
> +	int err;
> +again:
> +	if (unlikely(idr_pre_get(idr, GFP_KERNEL) == 0))
> +		return -ENOMEM;
> +
> +	mutex_lock(&cooling_cpufreq_lock);
> +	err = idr_get_new(idr, NULL, id);
> +	mutex_unlock(&cooling_cpufreq_lock);
> +
> +	if (unlikely(err == -EAGAIN))
> +		goto again;
> +	else if (unlikely(err))
> +		return err;
> +
> +	*id = *id & MAX_ID_MASK;
> +	return 0;
> +}
> +
> +/**
> + * release_idr - function to free the unique id.
> + * @idr: struct idr * handle used for creating the id.
> + * @id: int value representing the unique id.
> + */
> +static void release_idr(struct idr *idr, int id)
> +{
> +	mutex_lock(&cooling_cpufreq_lock);
> +	idr_remove(idr, id);
> +	mutex_unlock(&cooling_cpufreq_lock);
> +}
> +
> +/**
> + * cputherm_register_notifier - Register a notifier with cpu cooling interface.
> + * @nb:	struct notifier_block * with callback info.
> + * @list: integer value for which notification is needed. possible values are
> + *	CPUFREQ_COOLING_START and CPUFREQ_COOLING_STOP.
> + *
> + * This exported function registers a driver with cpu cooling layer. The driver
> + * will be notified when any cpu cooling action is called.
> + */
> +int cputherm_register_notifier(struct notifier_block *nb, unsigned int list)
> +{
> +	int ret = 0;
> +
> +	switch (list) {
> +	case CPUFREQ_COOLING_START:
> +	case CPUFREQ_COOLING_STOP:
> +		ret = blocking_notifier_chain_register(
> +				&cputherm_state_notifier_list, nb);
> +		break;
> +	default:
> +		ret = -EINVAL;
> +	}
> +	return ret;
> +}
> +EXPORT_SYMBOL(cputherm_register_notifier);
> +
> +/**
> + * cputherm_unregister_notifier - Un-register a notifier.
> + * @nb:	struct notifier_block * with callback info.
> + * @list: integer value for which notification is needed. values possible are
> + *	CPUFREQ_COOLING_START or CPUFREQ_COOLING_STOP.
> + *
> + * This exported function un-registers a driver with cpu cooling layer.
> + */
> +int cputherm_unregister_notifier(struct notifier_block *nb, unsigned int list)
> +{
> +	int ret = 0;
> +
> +	switch (list) {
> +	case CPUFREQ_COOLING_START:
> +	case CPUFREQ_COOLING_STOP:
> +		ret = blocking_notifier_chain_unregister(
> +				&cputherm_state_notifier_list, nb);
> +		break;
> +	default:
> +		ret = -EINVAL;
> +	}
> +	return ret;
> +}
> +EXPORT_SYMBOL(cputherm_unregister_notifier);
> +
> +/* Below code defines functions to be used for cpufreq as cooling device */
> +
> +/**
> + * is_cpufreq_valid - function to check if a cpu has frequency transition policy.
> + * @cpu: cpu for which check is needed.
> + */
> +static int is_cpufreq_valid(int cpu)
> +{
> +	struct cpufreq_policy policy;
> +	return !cpufreq_get_policy(&policy, cpu);
> +}
> +
> +/**
> + * get_cpu_frequency - get the absolute value of frequency from level.
> + * @cpu: cpu for which frequency is fetched.
> + * @level: level of frequency of the CPU
> + *	e.g level=1 --> 1st MAX FREQ, LEVEL=2 ---> 2nd MAX FREQ, .... etc
> + */
> +static unsigned int get_cpu_frequency(unsigned int cpu, unsigned long level)
> +{
> +	int ret = 0, i = 0;
> +	unsigned long level_index;
> +	bool descend = false;
> +	struct cpufreq_frequency_table *table =
> +					cpufreq_frequency_get_table(cpu);
> +	if (!table)
> +		return ret;
> +
> +	while (table[i].frequency != CPUFREQ_TABLE_END) {
> +		if (table[i].frequency == CPUFREQ_ENTRY_INVALID)
> +			continue;
> +
> +		/*check if table in ascending or descending order*/
> +		if ((table[i + 1].frequency != CPUFREQ_TABLE_END) &&
> +			(table[i + 1].frequency < table[i].frequency)
> +			&& !descend) {
> +			descend = true;
> +		}
> +
> +		/*return if level matched and table in descending order*/
> +		if (descend && i == level)
> +			return table[i].frequency;
> +		i++;
> +	}
> +	i--;
> +
> +	if (level > i || descend)
> +		return ret;
> +	level_index = i - level;
> +
> +	/*Scan the table in reverse order and match the level*/
> +	while (i >= 0) {
> +		if (table[i].frequency == CPUFREQ_ENTRY_INVALID)
> +			continue;
> +		/*return if level matched*/
> +		if (i == level_index)
> +			return table[i].frequency;
> +		i--;
> +	}
> +	return ret;
> +}
> +
> +/**
> + * cpufreq_apply_cooling - function to apply frequency clipping.
> + * @cpufreq_device: cpufreq_cooling_device pointer containing frequency
> + *	clipping data.
> + * @cooling_state: value of the cooling state.
> + */
> +static int cpufreq_apply_cooling(struct cpufreq_cooling_device *cpufreq_device,
> +				unsigned long cooling_state)
> +{
> +	unsigned int event, cpuid, clip_freq;
> +	struct cpumask *maskPtr = &cpufreq_device->allowed_cpus;
> +	unsigned int cpu = cpumask_any(maskPtr);
> +
> +
> +	/* Check if the old cooling action is same as new cooling action */
> +	if (cpufreq_device->cpufreq_state == cooling_state)
> +		return 0;
> +
> +	clip_freq = get_cpu_frequency(cpu, cooling_state);
> +	if (!clip_freq)
> +		return -EINVAL;
> +
> +	cpufreq_device->cpufreq_state = cooling_state;
> +	cpufreq_device->cpufreq_val = clip_freq;
> +	notify_device = cpufreq_device;
> +
> +	if (cooling_state != 0)
> +		event = CPUFREQ_COOLING_START;
> +	else
> +		event = CPUFREQ_COOLING_STOP;
> +
> +	blocking_notifier_call_chain(&cputherm_state_notifier_list,
> +						event, &clip_freq);
> +
> +	for_each_cpu(cpuid, maskPtr) {
> +		if (is_cpufreq_valid(cpuid))
> +			cpufreq_update_policy(cpuid);
> +	}
> +
> +	notify_device = NOTIFY_INVALID;
> +
> +	return 0;
> +}
> +
> +/**
> + * cpufreq_thermal_notifier - notifier callback for cpufreq policy change.
> + * @nb:	struct notifier_block * with callback info.
> + * @event: value showing cpufreq event for which this function invoked.
> + * @data: callback-specific data
> + */
> +static int cpufreq_thermal_notifier(struct notifier_block *nb,
> +					unsigned long event, void *data)
> +{
> +	struct cpufreq_policy *policy = data;
> +	unsigned long max_freq = 0;
> +
> +	if (event != CPUFREQ_ADJUST || notify_device == NOTIFY_INVALID)
> +		return 0;
> +
> +	if (cpumask_test_cpu(policy->cpu, &notify_device->allowed_cpus))
> +		max_freq = notify_device->cpufreq_val;
> +
> +	/* Never exceed user_policy.max*/
> +	if (max_freq > policy->user_policy.max)
> +		max_freq = policy->user_policy.max;
> +
> +	if (policy->max != max_freq)
> +		cpufreq_verify_within_limits(policy, 0, max_freq);
> +
> +	return 0;
> +}
> +
> +/*
> + * cpufreq cooling device callback functions are defined below
> + */
> +
> +/**
> + * cpufreq_get_max_state - callback function to get the max cooling state.
> + * @cdev: thermal cooling device pointer.
> + * @state: fill this variable with the max cooling state.
> + */
> +static int cpufreq_get_max_state(struct thermal_cooling_device *cdev,
> +				 unsigned long *state)
> +{
> +	int ret = -EINVAL, i = 0;
> +	struct cpufreq_cooling_device *cpufreq_device;
> +	struct cpumask *maskPtr;
> +	unsigned int cpu;
> +	struct cpufreq_frequency_table *table;
> +
> +	mutex_lock(&cooling_cpufreq_lock);
> +	list_for_each_entry(cpufreq_device, &cooling_cpufreq_list, node) {
> +		if (cpufreq_device && cpufreq_device->cool_dev == cdev)
> +			break;
> +	}
> +	if (cpufreq_device == NULL)
> +		goto return_get_max_state;
> +
> +	maskPtr = &cpufreq_device->allowed_cpus;
> +	cpu = cpumask_any(maskPtr);
> +	table = cpufreq_frequency_get_table(cpu);
> +	if (!table) {
> +		*state = 0;
> +		ret = 0;
> +		goto return_get_max_state;
> +	}
> +
> +	while (table[i].frequency != CPUFREQ_TABLE_END) {
> +		if (table[i].frequency == CPUFREQ_ENTRY_INVALID)
> +			continue;
> +		i++;
> +	}
> +	if (i > 0) {
> +		*state = --i;
> +		ret = 0;
> +	}
> +
> +return_get_max_state:
> +	mutex_unlock(&cooling_cpufreq_lock);
> +	return ret;
> +}
> +
> +/**
> + * cpufreq_get_cur_state - callback function to get the current cooling state.
> + * @cdev: thermal cooling device pointer.
> + * @state: fill this variable with the current cooling state.
> + */
> +static int cpufreq_get_cur_state(struct thermal_cooling_device *cdev,
> +				 unsigned long *state)
> +{
> +	int ret = -EINVAL;
> +	struct cpufreq_cooling_device *cpufreq_device;
> +
> +	mutex_lock(&cooling_cpufreq_lock);
> +	list_for_each_entry(cpufreq_device, &cooling_cpufreq_list, node) {
> +		if (cpufreq_device && cpufreq_device->cool_dev == cdev) {
> +			*state = cpufreq_device->cpufreq_state;
> +			ret = 0;
> +			break;
> +		}
> +	}
> +	mutex_unlock(&cooling_cpufreq_lock);
> +

as cpufreq may be changed in other places, e.g. via sysfs I/F, we should
use the current cpu frequency to get the REAL cooling state, rather than
using a cached value.

thanks,
rui


--
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
the body of a message to [e-mail]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Amit Kachhap Aug. 17, 2012, 7:58 a.m. UTC | #4

On 17 August 2012 12:54, Zhang Rui <[e-mail]> wrote:
> On ?, 2012-08-16 at 17:11 +0530, Amit Daniel Kachhap wrote:
>> This patchset introduces a new generic cooling device based on cpufreq
>> that can be used on non-ACPI platforms.  As a proof of concept, we have
>> drivers for the following platforms using this mechanism now:
>>
>>  * Samsung Exynos (Exynos4 and Exynos5) in the current patchset.
>>  * Freescale i.MX (git://git.linaro.org/people/amitdanielk/linux.git imx6q_thermal)
>>
>> There is a small change in cpufreq cooling registration APIs, so a minor
>> change is needed for Freescale platforms.
>>
>> Brief Description:
>>
>> 1) The generic cooling devices code is placed inside driver/thermal/*
>>    as placing inside acpi folder will need un-necessary enabling of acpi
>>    code.  This code is architecture independent.
>>
>> 2) This patchset adds generic cpu cooling low level implementation
>>    through frequency clipping.  In future, other cpu related cooling
>>    devices may be added here.  An ACPI version of this already exists
>>    (drivers/acpi/processor_thermal.c) .But this will be useful for
>>    platforms like ARM using the generic thermal interface along with the
>>    generic cpu cooling devices.  The cooling device registration API's
>>    return cooling device pointers which can be easily binded with the
>>    thermal zone trip points.  The important APIs exposed are,
>>
>>    a) struct thermal_cooling_device *cpufreq_cooling_register(
>>         struct cpumask *clip_cpus)
>>    b) void cpufreq_cooling_unregister(struct thermal_cooling_device *cdev)
>>
>> 3) Samsung exynos platform thermal implementation is done using the
>>    generic cpu cooling APIs and the new trip type.  The temperature sensor
>>    driver present in the hwmon folder(registered as hwmon driver) is moved
>>    to thermal folder and registered as a thermal driver.
>>
>> A simple data/control flow diagrams is shown below,
>>
>> Core Linux thermal <----->  Exynos thermal interface <----- Temperature Sensor
>>           |                             |
>>          \|/                            |
>>   Cpufreq cooling device <---------------
>>
>> TODO:
>> *Will send the DT enablement patches later after the driver is merged.
>>
>> This patch:
>>
>> Add support for generic cpu thermal cooling low level implementations
>> using frequency scaling up/down based on the registration parameters.
>> Different cpu related cooling devices can be registered by the user and
>> the binding of these cooling devices to the corresponding trip points can
>> be easily done as the registration APIs return the cooling device pointer.
>> The user of these APIs are responsible for passing clipping frequency .
>> The drivers can also register to recieve notification about any cooling
>> action called.
>>
>> [[e-mail]: fix comment layout]
>> Signed-off-by: Amit Daniel Kachhap <[e-mail]>
>> Cc: Guenter Roeck <[e-mail]>
>> Cc: SangWook Ju <[e-mail]>
>> Cc: Durgadoss <[e-mail]>
>> Cc: Len Brown <[e-mail]>
>> Cc: Jean Delvare <[e-mail]>
>> Cc: Kyungmin Park <[e-mail]>
>> Cc: Kukjin Kim <[e-mail]>
>> Signed-off-by: Andrew Morton <[e-mail]>
>> Signed-off-by: Amit Daniel Kachhap <[e-mail]>
>> ---
>>  Documentation/thermal/cpu-cooling-api.txt |   52 +++
>>  drivers/thermal/Kconfig                   |   11 +
>>  drivers/thermal/Makefile                  |    1 +
>>  drivers/thermal/cpu_cooling.c             |  512 +++++++++++++++++++++++++++++
>>  include/linux/cpu_cooling.h               |   79 +++++
>>  5 files changed, 655 insertions(+), 0 deletions(-)
>>  create mode 100644 Documentation/thermal/cpu-cooling-api.txt
>>  create mode 100644 drivers/thermal/cpu_cooling.c
>>  create mode 100644 include/linux/cpu_cooling.h
>>
>> diff --git a/Documentation/thermal/cpu-cooling-api.txt b/Documentation/thermal/cpu-cooling-api.txt
>> new file mode 100644
>> index 0000000..a1f2a6b
>> --- /dev/null
>> +++ b/Documentation/thermal/cpu-cooling-api.txt
>> @@ -0,0 +1,52 @@
>> +CPU cooling APIs How To
>> +===================================
>> +
>> +Written by Amit Daniel Kachhap <[e-mail]>
>> +
>> +Updated: 12 May 2012
>> +
>> +Copyright (c)  2012 Samsung Electronics Co., Ltd(http://www.samsung.com)
>> +
>> +0. Introduction
>> +
>> +The generic cpu cooling(freq clipping) provides registration/unregistration APIs
>> +to the caller. The binding of the cooling devices to the trip point is left for
>> +the user. The registration APIs returns the cooling device pointer.
>> +
>> +1. cpu cooling APIs
>> +
>> +1.1 cpufreq registration/unregistration APIs
>> +1.1.1 struct thermal_cooling_device *cpufreq_cooling_register(
>> +     struct cpumask *clip_cpus)
>> +
>> +    This interface function registers the cpufreq cooling device with the name
>> +    "thermal-cpufreq-%x". This api can support multiple instances of cpufreq
>> +    cooling devices.
>> +
>> +   clip_cpus: cpumask of cpus where the frequency constraints will happen.
>> +
>> +1.1.2 void cpufreq_cooling_unregister(struct thermal_cooling_device *cdev)
>> +
>> +    This interface function unregisters the "thermal-cpufreq-%x" cooling device.
>> +
>> +    cdev: Cooling device pointer which has to be unregistered.
>> +
>> +
>> +1.2 CPU cooling action notifier register/unregister interface
>> +1.2.1 int cputherm_register_notifier(struct notifier_block *nb,
>> +     unsigned int list)
>> +
>> +    This interface registers a driver with cpu cooling layer. The driver will
>> +    be notified when any cpu cooling action is called.
>> +
>> +    nb: notifier function to register
>> +    list: CPUFREQ_COOLING_START or CPUFREQ_COOLING_STOP
>> +
>> +1.2.2 int cputherm_unregister_notifier(struct notifier_block *nb,
>> +     unsigned int list)
>> +
>> +    This interface registers a driver with cpu cooling layer. The driver will
>> +    be notified when any cpu cooling action is called.
>> +
>> +    nb: notifier function to register
>> +    list: CPUFREQ_COOLING_START or CPUFREQ_COOLING_STOP
>
> what are these two APIs used for?
> I did not see they are used in your patch set, do I miss something?
No currently they are not used by my patches. I added them on request
from Eduardo and others
>
>> diff --git a/drivers/thermal/Kconfig b/drivers/thermal/Kconfig
>> index 7dd8c34..996003b 100644
>> --- a/drivers/thermal/Kconfig
>> +++ b/drivers/thermal/Kconfig
>> @@ -19,6 +19,17 @@ config THERMAL_HWMON
>>       depends on HWMON=y || HWMON=THERMAL
>>       default y
>>
>> +config CPU_THERMAL
>> +     bool "generic cpu cooling support"
>> +     depends on THERMAL && CPU_FREQ
>> +     help
>> +       This implements the generic cpu cooling mechanism through frequency
>> +       reduction, cpu hotplug and any other ways of reducing temperature. An
>> +       ACPI version of this already exists(drivers/acpi/processor_thermal.c).
>> +       This will be useful for platforms using the generic thermal interface
>> +       and not the ACPI interface.
>> +       If you want this support, you should say Y here.
>> +
>>  config SPEAR_THERMAL
>>       bool "SPEAr thermal sensor driver"
>>       depends on THERMAL
>> diff --git a/drivers/thermal/Makefile b/drivers/thermal/Makefile
>> index fd9369a..aae59ad 100644
>> --- a/drivers/thermal/Makefile
>> +++ b/drivers/thermal/Makefile
>> @@ -3,5 +3,6 @@
>>  #
>>
>>  obj-$(CONFIG_THERMAL)                += thermal_sys.o
>> +obj-$(CONFIG_CPU_THERMAL)            += cpu_cooling.o
>>  obj-$(CONFIG_SPEAR_THERMAL)          += spear_thermal.o
>>  obj-$(CONFIG_RCAR_THERMAL)   += rcar_thermal.o
>> diff --git a/drivers/thermal/cpu_cooling.c b/drivers/thermal/cpu_cooling.c
>> new file mode 100644
>> index 0000000..c42e557
>> --- /dev/null
>> +++ b/drivers/thermal/cpu_cooling.c
>> @@ -0,0 +1,512 @@
>> +/*
>> + *  linux/drivers/thermal/cpu_cooling.c
>> + *
>> + *  Copyright (C) 2012       Samsung Electronics Co., Ltd(http://www.samsung.com)
>> + *  Copyright (C) 2012  Amit Daniel <[e-mail]>
>> + *
>> + * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>> + *  This program is free software; you can redistribute it and/or modify
>> + *  it under the terms of the GNU General Public License as published by
>> + *  the Free Software Foundation; version 2 of the License.
>> + *
>> + *  This program is distributed in the hope that it will be useful, but
>> + *  WITHOUT ANY WARRANTY; without even the implied warranty of
>> + *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
>> + *  General Public License for more details.
>> + *
>> + *  You should have received a copy of the GNU General Public License along
>> + *  with this program; if not, write to the Free Software Foundation, Inc.,
>> + *  59 Temple Place, Suite 330, Boston, MA 02111-1307 USA.
>> + *
>> + * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>> + */
>> +#include <linux/kernel.h>
>> +#include <linux/module.h>
>> +#include <linux/thermal.h>
>> +#include <linux/platform_device.h>
>> +#include <linux/cpufreq.h>
>> +#include <linux/err.h>
>> +#include <linux/slab.h>
>> +#include <linux/cpu.h>
>> +#include <linux/cpu_cooling.h>
>> +
>> +/**
>> + * struct cpufreq_cooling_device
>> + * @id: unique integer value corresponding to each cpufreq_cooling_device
>> + *   registered.
>> + * @cool_dev: thermal_cooling_device pointer to keep track of the the
>> + *   egistered cooling device.
>> + * @cpufreq_state: integer value representing the current state of cpufreq
>> + *   cooling devices.
>> + * @cpufreq_val: integer value representing the absolute value of the clipped
>> + *   frequency.
>> + * @allowed_cpus: all the cpus involved for this cpufreq_cooling_device.
>> + * @node: list_head to link all cpufreq_cooling_device together.
>> + *
>> + * This structure is required for keeping information of each
>> + * cpufreq_cooling_device registered as a list whose head is represented by
>> + * cooling_cpufreq_list. In order to prevent corruption of this list a
>> + * mutex lock cooling_cpufreq_lock is used.
>> + */
>> +struct cpufreq_cooling_device {
>> +     int id;
>> +     struct thermal_cooling_device *cool_dev;
>> +     unsigned int cpufreq_state;
>> +     unsigned int cpufreq_val;
>> +     struct cpumask allowed_cpus;
>> +     struct list_head node;
>> +};
>> +static LIST_HEAD(cooling_cpufreq_list);
>> +static DEFINE_IDR(cpufreq_idr);
>> +
>> +static struct mutex cooling_cpufreq_lock;
>> +
>> +/* notify_table passes value to the CPUFREQ_ADJUST callback function. */
>> +#define NOTIFY_INVALID NULL
>> +struct cpufreq_cooling_device *notify_device;
>> +
>> +/* Head of the blocking notifier chain to inform about frequency clamping */
>> +static BLOCKING_NOTIFIER_HEAD(cputherm_state_notifier_list);
>> +
>> +/**
>> + * get_idr - function to get a unique id.
>> + * @idr: struct idr * handle used to create a id.
>> + * @id: int * value generated by this function.
>> + */
>> +static int get_idr(struct idr *idr, int *id)
>> +{
>> +     int err;
>> +again:
>> +     if (unlikely(idr_pre_get(idr, GFP_KERNEL) == 0))
>> +             return -ENOMEM;
>> +
>> +     mutex_lock(&cooling_cpufreq_lock);
>> +     err = idr_get_new(idr, NULL, id);
>> +     mutex_unlock(&cooling_cpufreq_lock);
>> +
>> +     if (unlikely(err == -EAGAIN))
>> +             goto again;
>> +     else if (unlikely(err))
>> +             return err;
>> +
>> +     *id = *id & MAX_ID_MASK;
>> +     return 0;
>> +}
>> +
>> +/**
>> + * release_idr - function to free the unique id.
>> + * @idr: struct idr * handle used for creating the id.
>> + * @id: int value representing the unique id.
>> + */
>> +static void release_idr(struct idr *idr, int id)
>> +{
>> +     mutex_lock(&cooling_cpufreq_lock);
>> +     idr_remove(idr, id);
>> +     mutex_unlock(&cooling_cpufreq_lock);
>> +}
>> +
>> +/**
>> + * cputherm_register_notifier - Register a notifier with cpu cooling interface.
>> + * @nb:      struct notifier_block * with callback info.
>> + * @list: integer value for which notification is needed. possible values are
>> + *   CPUFREQ_COOLING_START and CPUFREQ_COOLING_STOP.
>> + *
>> + * This exported function registers a driver with cpu cooling layer. The driver
>> + * will be notified when any cpu cooling action is called.
>> + */
>> +int cputherm_register_notifier(struct notifier_block *nb, unsigned int list)
>> +{
>> +     int ret = 0;
>> +
>> +     switch (list) {
>> +     case CPUFREQ_COOLING_START:
>> +     case CPUFREQ_COOLING_STOP:
>> +             ret = blocking_notifier_chain_register(
>> +                             &cputherm_state_notifier_list, nb);
>> +             break;
>> +     default:
>> +             ret = -EINVAL;
>> +     }
>> +     return ret;
>> +}
>> +EXPORT_SYMBOL(cputherm_register_notifier);
>> +
>> +/**
>> + * cputherm_unregister_notifier - Un-register a notifier.
>> + * @nb:      struct notifier_block * with callback info.
>> + * @list: integer value for which notification is needed. values possible are
>> + *   CPUFREQ_COOLING_START or CPUFREQ_COOLING_STOP.
>> + *
>> + * This exported function un-registers a driver with cpu cooling layer.
>> + */
>> +int cputherm_unregister_notifier(struct notifier_block *nb, unsigned int list)
>> +{
>> +     int ret = 0;
>> +
>> +     switch (list) {
>> +     case CPUFREQ_COOLING_START:
>> +     case CPUFREQ_COOLING_STOP:
>> +             ret = blocking_notifier_chain_unregister(
>> +                             &cputherm_state_notifier_list, nb);
>> +             break;
>> +     default:
>> +             ret = -EINVAL;
>> +     }
>> +     return ret;
>> +}
>> +EXPORT_SYMBOL(cputherm_unregister_notifier);
>> +
>> +/* Below code defines functions to be used for cpufreq as cooling device */
>> +
>> +/**
>> + * is_cpufreq_valid - function to check if a cpu has frequency transition policy.
>> + * @cpu: cpu for which check is needed.
>> + */
>> +static int is_cpufreq_valid(int cpu)
>> +{
>> +     struct cpufreq_policy policy;
>> +     return !cpufreq_get_policy(&policy, cpu);
>> +}
>> +
>> +/**
>> + * get_cpu_frequency - get the absolute value of frequency from level.
>> + * @cpu: cpu for which frequency is fetched.
>> + * @level: level of frequency of the CPU
>> + *   e.g level=1 --> 1st MAX FREQ, LEVEL=2 ---> 2nd MAX FREQ, .... etc
>> + */
>> +static unsigned int get_cpu_frequency(unsigned int cpu, unsigned long level)
>> +{
>> +     int ret = 0, i = 0;
>> +     unsigned long level_index;
>> +     bool descend = false;
>> +     struct cpufreq_frequency_table *table =
>> +                                     cpufreq_frequency_get_table(cpu);
>> +     if (!table)
>> +             return ret;
>> +
>> +     while (table[i].frequency != CPUFREQ_TABLE_END) {
>> +             if (table[i].frequency == CPUFREQ_ENTRY_INVALID)
>> +                     continue;
>> +
>> +             /*check if table in ascending or descending order*/
>> +             if ((table[i + 1].frequency != CPUFREQ_TABLE_END) &&
>> +                     (table[i + 1].frequency < table[i].frequency)
>> +                     && !descend) {
>> +                     descend = true;
>> +             }
>> +
>> +             /*return if level matched and table in descending order*/
>> +             if (descend && i == level)
>> +                     return table[i].frequency;
>> +             i++;
>> +     }
>> +     i--;
>> +
>> +     if (level > i || descend)
>> +             return ret;
>> +     level_index = i - level;
>> +
>> +     /*Scan the table in reverse order and match the level*/
>> +     while (i >= 0) {
>> +             if (table[i].frequency == CPUFREQ_ENTRY_INVALID)
>> +                     continue;
>> +             /*return if level matched*/
>> +             if (i == level_index)
>> +                     return table[i].frequency;
>> +             i--;
>> +     }
>> +     return ret;
>> +}
>> +
>> +/**
>> + * cpufreq_apply_cooling - function to apply frequency clipping.
>> + * @cpufreq_device: cpufreq_cooling_device pointer containing frequency
>> + *   clipping data.
>> + * @cooling_state: value of the cooling state.
>> + */
>> +static int cpufreq_apply_cooling(struct cpufreq_cooling_device *cpufreq_device,
>> +                             unsigned long cooling_state)
>> +{
>> +     unsigned int event, cpuid, clip_freq;
>> +     struct cpumask *maskPtr = &cpufreq_device->allowed_cpus;
>> +     unsigned int cpu = cpumask_any(maskPtr);
>> +
>> +
>> +     /* Check if the old cooling action is same as new cooling action */
>> +     if (cpufreq_device->cpufreq_state == cooling_state)
>> +             return 0;
>> +
>> +     clip_freq = get_cpu_frequency(cpu, cooling_state);
>> +     if (!clip_freq)
>> +             return -EINVAL;
>> +
>> +     cpufreq_device->cpufreq_state = cooling_state;
>> +     cpufreq_device->cpufreq_val = clip_freq;
>> +     notify_device = cpufreq_device;
>> +
>> +     if (cooling_state != 0)
>> +             event = CPUFREQ_COOLING_START;
>> +     else
>> +             event = CPUFREQ_COOLING_STOP;
>> +
>> +     blocking_notifier_call_chain(&cputherm_state_notifier_list,
>> +                                             event, &clip_freq);
>> +
>> +     for_each_cpu(cpuid, maskPtr) {
>> +             if (is_cpufreq_valid(cpuid))
>> +                     cpufreq_update_policy(cpuid);
>> +     }
>> +
>> +     notify_device = NOTIFY_INVALID;
>> +
>> +     return 0;
>> +}
>> +
>> +/**
>> + * cpufreq_thermal_notifier - notifier callback for cpufreq policy change.
>> + * @nb:      struct notifier_block * with callback info.
>> + * @event: value showing cpufreq event for which this function invoked.
>> + * @data: callback-specific data
>> + */
>> +static int cpufreq_thermal_notifier(struct notifier_block *nb,
>> +                                     unsigned long event, void *data)
>> +{
>> +     struct cpufreq_policy *policy = data;
>> +     unsigned long max_freq = 0;
>> +
>> +     if (event != CPUFREQ_ADJUST || notify_device == NOTIFY_INVALID)
>> +             return 0;
>> +
>> +     if (cpumask_test_cpu(policy->cpu, &notify_device->allowed_cpus))
>> +             max_freq = notify_device->cpufreq_val;
>> +
>> +     /* Never exceed user_policy.max*/
>> +     if (max_freq > policy->user_policy.max)
>> +             max_freq = policy->user_policy.max;
>> +
>> +     if (policy->max != max_freq)
>> +             cpufreq_verify_within_limits(policy, 0, max_freq);
>> +
>> +     return 0;
>> +}
>> +
>> +/*
>> + * cpufreq cooling device callback functions are defined below
>> + */
>> +
>> +/**
>> + * cpufreq_get_max_state - callback function to get the max cooling state.
>> + * @cdev: thermal cooling device pointer.
>> + * @state: fill this variable with the max cooling state.
>> + */
>> +static int cpufreq_get_max_state(struct thermal_cooling_device *cdev,
>> +                              unsigned long *state)
>> +{
>> +     int ret = -EINVAL, i = 0;
>> +     struct cpufreq_cooling_device *cpufreq_device;
>> +     struct cpumask *maskPtr;
>> +     unsigned int cpu;
>> +     struct cpufreq_frequency_table *table;
>> +
>> +     mutex_lock(&cooling_cpufreq_lock);
>> +     list_for_each_entry(cpufreq_device, &cooling_cpufreq_list, node) {
>> +             if (cpufreq_device && cpufreq_device->cool_dev == cdev)
>> +                     break;
>> +     }
>> +     if (cpufreq_device == NULL)
>> +             goto return_get_max_state;
>> +
>> +     maskPtr = &cpufreq_device->allowed_cpus;
>> +     cpu = cpumask_any(maskPtr);
>> +     table = cpufreq_frequency_get_table(cpu);
>> +     if (!table) {
>> +             *state = 0;
>> +             ret = 0;
>> +             goto return_get_max_state;
>> +     }
>> +
>> +     while (table[i].frequency != CPUFREQ_TABLE_END) {
>> +             if (table[i].frequency == CPUFREQ_ENTRY_INVALID)
>> +                     continue;
>> +             i++;
>> +     }
>> +     if (i > 0) {
>> +             *state = --i;
>> +             ret = 0;
>> +     }
>> +
>> +return_get_max_state:
>> +     mutex_unlock(&cooling_cpufreq_lock);
>> +     return ret;
>> +}
>> +
>> +/**
>> + * cpufreq_get_cur_state - callback function to get the current cooling state.
>> + * @cdev: thermal cooling device pointer.
>> + * @state: fill this variable with the current cooling state.
>> + */
>> +static int cpufreq_get_cur_state(struct thermal_cooling_device *cdev,
>> +                              unsigned long *state)
>> +{
>> +     int ret = -EINVAL;
>> +     struct cpufreq_cooling_device *cpufreq_device;
>> +
>> +     mutex_lock(&cooling_cpufreq_lock);
>> +     list_for_each_entry(cpufreq_device, &cooling_cpufreq_list, node) {
>> +             if (cpufreq_device && cpufreq_device->cool_dev == cdev) {
>> +                     *state = cpufreq_device->cpufreq_state;
>> +                     ret = 0;
>> +                     break;
>> +             }
>> +     }
>> +     mutex_unlock(&cooling_cpufreq_lock);
>> +
>
> as cpufreq may be changed in other places, e.g. via sysfs I/F, we should
> use the current cpu frequency to get the REAL cooling state, rather than
> using a cached value.

Yes agreed , I will repost with your suggestion.

Thanks,
Amit
>
> thanks,
> rui
>
>
--
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
the body of a message to [e-mail]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Eduardo Valentin Aug. 17, 2012, 8:56 a.m. UTC | #5

Hello,

On Fri, Aug 17, 2012 at 10:58 AM, Amit Kachhap <[e-mail]> wrote:
> On 17 August 2012 12:54, Zhang Rui <[e-mail]> wrote:
>> On ?, 2012-08-16 at 17:11 +0530, Amit Daniel Kachhap wrote:
>>> This patchset introduces a new generic cooling device based on cpufreq
>>> that can be used on non-ACPI platforms.  As a proof of concept, we have
>>> drivers for the following platforms using this mechanism now:
>>>
>>>  * Samsung Exynos (Exynos4 and Exynos5) in the current patchset.
>>>  * Freescale i.MX (git://git.linaro.org/people/amitdanielk/linux.git imx6q_thermal)
>>>
>>> There is a small change in cpufreq cooling registration APIs, so a minor
>>> change is needed for Freescale platforms.
>>>
>>> Brief Description:
>>>
>>> 1) The generic cooling devices code is placed inside driver/thermal/*
>>>    as placing inside acpi folder will need un-necessary enabling of acpi
>>>    code.  This code is architecture independent.
>>>
>>> 2) This patchset adds generic cpu cooling low level implementation
>>>    through frequency clipping.  In future, other cpu related cooling
>>>    devices may be added here.  An ACPI version of this already exists
>>>    (drivers/acpi/processor_thermal.c) .But this will be useful for
>>>    platforms like ARM using the generic thermal interface along with the
>>>    generic cpu cooling devices.  The cooling device registration API's
>>>    return cooling device pointers which can be easily binded with the
>>>    thermal zone trip points.  The important APIs exposed are,
>>>
>>>    a) struct thermal_cooling_device *cpufreq_cooling_register(
>>>         struct cpumask *clip_cpus)
>>>    b) void cpufreq_cooling_unregister(struct thermal_cooling_device *cdev)
>>>
>>> 3) Samsung exynos platform thermal implementation is done using the
>>>    generic cpu cooling APIs and the new trip type.  The temperature sensor
>>>    driver present in the hwmon folder(registered as hwmon driver) is moved
>>>    to thermal folder and registered as a thermal driver.
>>>
>>> A simple data/control flow diagrams is shown below,
>>>
>>> Core Linux thermal <----->  Exynos thermal interface <----- Temperature Sensor
>>>           |                             |
>>>          \|/                            |
>>>   Cpufreq cooling device <---------------
>>>
>>> TODO:
>>> *Will send the DT enablement patches later after the driver is merged.
>>>
>>> This patch:
>>>
>>> Add support for generic cpu thermal cooling low level implementations
>>> using frequency scaling up/down based on the registration parameters.
>>> Different cpu related cooling devices can be registered by the user and
>>> the binding of these cooling devices to the corresponding trip points can
>>> be easily done as the registration APIs return the cooling device pointer.
>>> The user of these APIs are responsible for passing clipping frequency .
>>> The drivers can also register to recieve notification about any cooling
>>> action called.
>>>
>>> [[e-mail]: fix comment layout]
>>> Signed-off-by: Amit Daniel Kachhap <[e-mail]>
>>> Cc: Guenter Roeck <[e-mail]>
>>> Cc: SangWook Ju <[e-mail]>
>>> Cc: Durgadoss <[e-mail]>
>>> Cc: Len Brown <[e-mail]>
>>> Cc: Jean Delvare <[e-mail]>
>>> Cc: Kyungmin Park <[e-mail]>
>>> Cc: Kukjin Kim <[e-mail]>
>>> Signed-off-by: Andrew Morton <[e-mail]>
>>> Signed-off-by: Amit Daniel Kachhap <[e-mail]>
>>> ---
>>>  Documentation/thermal/cpu-cooling-api.txt |   52 +++
>>>  drivers/thermal/Kconfig                   |   11 +
>>>  drivers/thermal/Makefile                  |    1 +
>>>  drivers/thermal/cpu_cooling.c             |  512 +++++++++++++++++++++++++++++
>>>  include/linux/cpu_cooling.h               |   79 +++++
>>>  5 files changed, 655 insertions(+), 0 deletions(-)
>>>  create mode 100644 Documentation/thermal/cpu-cooling-api.txt
>>>  create mode 100644 drivers/thermal/cpu_cooling.c
>>>  create mode 100644 include/linux/cpu_cooling.h
>>>
>>> diff --git a/Documentation/thermal/cpu-cooling-api.txt b/Documentation/thermal/cpu-cooling-api.txt
>>> new file mode 100644
>>> index 0000000..a1f2a6b
>>> --- /dev/null
>>> +++ b/Documentation/thermal/cpu-cooling-api.txt
>>> @@ -0,0 +1,52 @@
>>> +CPU cooling APIs How To
>>> +===================================
>>> +
>>> +Written by Amit Daniel Kachhap <[e-mail]>
>>> +
>>> +Updated: 12 May 2012
>>> +
>>> +Copyright (c)  2012 Samsung Electronics Co., Ltd(http://www.samsung.com)
>>> +
>>> +0. Introduction
>>> +
>>> +The generic cpu cooling(freq clipping) provides registration/unregistration APIs
>>> +to the caller. The binding of the cooling devices to the trip point is left for
>>> +the user. The registration APIs returns the cooling device pointer.
>>> +
>>> +1. cpu cooling APIs
>>> +
>>> +1.1 cpufreq registration/unregistration APIs
>>> +1.1.1 struct thermal_cooling_device *cpufreq_cooling_register(
>>> +     struct cpumask *clip_cpus)
>>> +
>>> +    This interface function registers the cpufreq cooling device with the name
>>> +    "thermal-cpufreq-%x". This api can support multiple instances of cpufreq
>>> +    cooling devices.
>>> +
>>> +   clip_cpus: cpumask of cpus where the frequency constraints will happen.
>>> +
>>> +1.1.2 void cpufreq_cooling_unregister(struct thermal_cooling_device *cdev)
>>> +
>>> +    This interface function unregisters the "thermal-cpufreq-%x" cooling device.
>>> +
>>> +    cdev: Cooling device pointer which has to be unregistered.
>>> +
>>> +
>>> +1.2 CPU cooling action notifier register/unregister interface
>>> +1.2.1 int cputherm_register_notifier(struct notifier_block *nb,
>>> +     unsigned int list)
>>> +
>>> +    This interface registers a driver with cpu cooling layer. The driver will
>>> +    be notified when any cpu cooling action is called.
>>> +
>>> +    nb: notifier function to register
>>> +    list: CPUFREQ_COOLING_START or CPUFREQ_COOLING_STOP
>>> +
>>> +1.2.2 int cputherm_unregister_notifier(struct notifier_block *nb,
>>> +     unsigned int list)
>>> +
>>> +    This interface registers a driver with cpu cooling layer. The driver will
>>> +    be notified when any cpu cooling action is called.
>>> +
>>> +    nb: notifier function to register
>>> +    list: CPUFREQ_COOLING_START or CPUFREQ_COOLING_STOP
>>
>> what are these two APIs used for?
>> I did not see they are used in your patch set, do I miss something?
> No currently they are not used by my patches. I added them on request
> from Eduardo and others

Yeah, this was a suggestion as we didn't really know how the FW part
would evolve by that time.

The reasoning is to allow any interested user (in kernel) to be
notified when max frequency changes. Actually, the use case behind
this is to allow such users to perform some handshaking or stop their
activities or even change some paramenters, in case the max frequency
would change. Ideally it would be possible to nack the cooling
transition. But that is yet a wider discussion. So far we don't have
users for this.

>>
>>> diff --git a/drivers/thermal/Kconfig b/drivers/thermal/Kconfig
>>> index 7dd8c34..996003b 100644
>>> --- a/drivers/thermal/Kconfig
>>> +++ b/drivers/thermal/Kconfig
>>> @@ -19,6 +19,17 @@ config THERMAL_HWMON
>>>       depends on HWMON=y || HWMON=THERMAL
>>>       default y
>>>
>>> +config CPU_THERMAL
>>> +     bool "generic cpu cooling support"
>>> +     depends on THERMAL && CPU_FREQ
>>> +     help
>>> +       This implements the generic cpu cooling mechanism through frequency
>>> +       reduction, cpu hotplug and any other ways of reducing temperature. An
>>> +       ACPI version of this already exists(drivers/acpi/processor_thermal.c).
>>> +       This will be useful for platforms using the generic thermal interface
>>> +       and not the ACPI interface.
>>> +       If you want this support, you should say Y here.
>>> +
>>>  config SPEAR_THERMAL
>>>       bool "SPEAr thermal sensor driver"
>>>       depends on THERMAL
>>> diff --git a/drivers/thermal/Makefile b/drivers/thermal/Makefile
>>> index fd9369a..aae59ad 100644
>>> --- a/drivers/thermal/Makefile
>>> +++ b/drivers/thermal/Makefile
>>> @@ -3,5 +3,6 @@
>>>  #
>>>
>>>  obj-$(CONFIG_THERMAL)                += thermal_sys.o
>>> +obj-$(CONFIG_CPU_THERMAL)            += cpu_cooling.o
>>>  obj-$(CONFIG_SPEAR_THERMAL)          += spear_thermal.o
>>>  obj-$(CONFIG_RCAR_THERMAL)   += rcar_thermal.o
>>> diff --git a/drivers/thermal/cpu_cooling.c b/drivers/thermal/cpu_cooling.c
>>> new file mode 100644
>>> index 0000000..c42e557
>>> --- /dev/null
>>> +++ b/drivers/thermal/cpu_cooling.c
>>> @@ -0,0 +1,512 @@
>>> +/*
>>> + *  linux/drivers/thermal/cpu_cooling.c
>>> + *
>>> + *  Copyright (C) 2012       Samsung Electronics Co., Ltd(http://www.samsung.com)
>>> + *  Copyright (C) 2012  Amit Daniel <[e-mail]>
>>> + *
>>> + * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>>> + *  This program is free software; you can redistribute it and/or modify
>>> + *  it under the terms of the GNU General Public License as published by
>>> + *  the Free Software Foundation; version 2 of the License.
>>> + *
>>> + *  This program is distributed in the hope that it will be useful, but
>>> + *  WITHOUT ANY WARRANTY; without even the implied warranty of
>>> + *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
>>> + *  General Public License for more details.
>>> + *
>>> + *  You should have received a copy of the GNU General Public License along
>>> + *  with this program; if not, write to the Free Software Foundation, Inc.,
>>> + *  59 Temple Place, Suite 330, Boston, MA 02111-1307 USA.
>>> + *
>>> + * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>>> + */
>>> +#include <linux/kernel.h>
>>> +#include <linux/module.h>
>>> +#include <linux/thermal.h>
>>> +#include <linux/platform_device.h>
>>> +#include <linux/cpufreq.h>
>>> +#include <linux/err.h>
>>> +#include <linux/slab.h>
>>> +#include <linux/cpu.h>
>>> +#include <linux/cpu_cooling.h>
>>> +
>>> +/**
>>> + * struct cpufreq_cooling_device
>>> + * @id: unique integer value corresponding to each cpufreq_cooling_device
>>> + *   registered.
>>> + * @cool_dev: thermal_cooling_device pointer to keep track of the the
>>> + *   egistered cooling device.
>>> + * @cpufreq_state: integer value representing the current state of cpufreq
>>> + *   cooling devices.
>>> + * @cpufreq_val: integer value representing the absolute value of the clipped
>>> + *   frequency.
>>> + * @allowed_cpus: all the cpus involved for this cpufreq_cooling_device.
>>> + * @node: list_head to link all cpufreq_cooling_device together.
>>> + *
>>> + * This structure is required for keeping information of each
>>> + * cpufreq_cooling_device registered as a list whose head is represented by
>>> + * cooling_cpufreq_list. In order to prevent corruption of this list a
>>> + * mutex lock cooling_cpufreq_lock is used.
>>> + */
>>> +struct cpufreq_cooling_device {
>>> +     int id;
>>> +     struct thermal_cooling_device *cool_dev;
>>> +     unsigned int cpufreq_state;
>>> +     unsigned int cpufreq_val;
>>> +     struct cpumask allowed_cpus;
>>> +     struct list_head node;
>>> +};
>>> +static LIST_HEAD(cooling_cpufreq_list);
>>> +static DEFINE_IDR(cpufreq_idr);
>>> +
>>> +static struct mutex cooling_cpufreq_lock;
>>> +
>>> +/* notify_table passes value to the CPUFREQ_ADJUST callback function. */
>>> +#define NOTIFY_INVALID NULL
>>> +struct cpufreq_cooling_device *notify_device;
>>> +
>>> +/* Head of the blocking notifier chain to inform about frequency clamping */
>>> +static BLOCKING_NOTIFIER_HEAD(cputherm_state_notifier_list);
>>> +
>>> +/**
>>> + * get_idr - function to get a unique id.
>>> + * @idr: struct idr * handle used to create a id.
>>> + * @id: int * value generated by this function.
>>> + */
>>> +static int get_idr(struct idr *idr, int *id)
>>> +{
>>> +     int err;
>>> +again:
>>> +     if (unlikely(idr_pre_get(idr, GFP_KERNEL) == 0))
>>> +             return -ENOMEM;
>>> +
>>> +     mutex_lock(&cooling_cpufreq_lock);
>>> +     err = idr_get_new(idr, NULL, id);
>>> +     mutex_unlock(&cooling_cpufreq_lock);
>>> +
>>> +     if (unlikely(err == -EAGAIN))
>>> +             goto again;
>>> +     else if (unlikely(err))
>>> +             return err;
>>> +
>>> +     *id = *id & MAX_ID_MASK;
>>> +     return 0;
>>> +}
>>> +
>>> +/**
>>> + * release_idr - function to free the unique id.
>>> + * @idr: struct idr * handle used for creating the id.
>>> + * @id: int value representing the unique id.
>>> + */
>>> +static void release_idr(struct idr *idr, int id)
>>> +{
>>> +     mutex_lock(&cooling_cpufreq_lock);
>>> +     idr_remove(idr, id);
>>> +     mutex_unlock(&cooling_cpufreq_lock);
>>> +}
>>> +
>>> +/**
>>> + * cputherm_register_notifier - Register a notifier with cpu cooling interface.
>>> + * @nb:      struct notifier_block * with callback info.
>>> + * @list: integer value for which notification is needed. possible values are
>>> + *   CPUFREQ_COOLING_START and CPUFREQ_COOLING_STOP.
>>> + *
>>> + * This exported function registers a driver with cpu cooling layer. The driver
>>> + * will be notified when any cpu cooling action is called.
>>> + */
>>> +int cputherm_register_notifier(struct notifier_block *nb, unsigned int list)
>>> +{
>>> +     int ret = 0;
>>> +
>>> +     switch (list) {
>>> +     case CPUFREQ_COOLING_START:
>>> +     case CPUFREQ_COOLING_STOP:
>>> +             ret = blocking_notifier_chain_register(
>>> +                             &cputherm_state_notifier_list, nb);
>>> +             break;
>>> +     default:
>>> +             ret = -EINVAL;
>>> +     }
>>> +     return ret;
>>> +}
>>> +EXPORT_SYMBOL(cputherm_register_notifier);
>>> +
>>> +/**
>>> + * cputherm_unregister_notifier - Un-register a notifier.
>>> + * @nb:      struct notifier_block * with callback info.
>>> + * @list: integer value for which notification is needed. values possible are
>>> + *   CPUFREQ_COOLING_START or CPUFREQ_COOLING_STOP.
>>> + *
>>> + * This exported function un-registers a driver with cpu cooling layer.
>>> + */
>>> +int cputherm_unregister_notifier(struct notifier_block *nb, unsigned int list)
>>> +{
>>> +     int ret = 0;
>>> +
>>> +     switch (list) {
>>> +     case CPUFREQ_COOLING_START:
>>> +     case CPUFREQ_COOLING_STOP:
>>> +             ret = blocking_notifier_chain_unregister(
>>> +                             &cputherm_state_notifier_list, nb);
>>> +             break;
>>> +     default:
>>> +             ret = -EINVAL;
>>> +     }
>>> +     return ret;
>>> +}
>>> +EXPORT_SYMBOL(cputherm_unregister_notifier);
>>> +
>>> +/* Below code defines functions to be used for cpufreq as cooling device */
>>> +
>>> +/**
>>> + * is_cpufreq_valid - function to check if a cpu has frequency transition policy.
>>> + * @cpu: cpu for which check is needed.
>>> + */
>>> +static int is_cpufreq_valid(int cpu)
>>> +{
>>> +     struct cpufreq_policy policy;
>>> +     return !cpufreq_get_policy(&policy, cpu);
>>> +}
>>> +
>>> +/**
>>> + * get_cpu_frequency - get the absolute value of frequency from level.
>>> + * @cpu: cpu for which frequency is fetched.
>>> + * @level: level of frequency of the CPU
>>> + *   e.g level=1 --> 1st MAX FREQ, LEVEL=2 ---> 2nd MAX FREQ, .... etc
>>> + */
>>> +static unsigned int get_cpu_frequency(unsigned int cpu, unsigned long level)
>>> +{
>>> +     int ret = 0, i = 0;
>>> +     unsigned long level_index;
>>> +     bool descend = false;
>>> +     struct cpufreq_frequency_table *table =
>>> +                                     cpufreq_frequency_get_table(cpu);
>>> +     if (!table)
>>> +             return ret;
>>> +
>>> +     while (table[i].frequency != CPUFREQ_TABLE_END) {
>>> +             if (table[i].frequency == CPUFREQ_ENTRY_INVALID)
>>> +                     continue;
>>> +
>>> +             /*check if table in ascending or descending order*/
>>> +             if ((table[i + 1].frequency != CPUFREQ_TABLE_END) &&
>>> +                     (table[i + 1].frequency < table[i].frequency)
>>> +                     && !descend) {
>>> +                     descend = true;
>>> +             }
>>> +
>>> +             /*return if level matched and table in descending order*/
>>> +             if (descend && i == level)
>>> +                     return table[i].frequency;
>>> +             i++;
>>> +     }
>>> +     i--;
>>> +
>>> +     if (level > i || descend)
>>> +             return ret;
>>> +     level_index = i - level;
>>> +
>>> +     /*Scan the table in reverse order and match the level*/
>>> +     while (i >= 0) {
>>> +             if (table[i].frequency == CPUFREQ_ENTRY_INVALID)
>>> +                     continue;
>>> +             /*return if level matched*/
>>> +             if (i == level_index)
>>> +                     return table[i].frequency;
>>> +             i--;
>>> +     }
>>> +     return ret;
>>> +}
>>> +
>>> +/**
>>> + * cpufreq_apply_cooling - function to apply frequency clipping.
>>> + * @cpufreq_device: cpufreq_cooling_device pointer containing frequency
>>> + *   clipping data.
>>> + * @cooling_state: value of the cooling state.
>>> + */
>>> +static int cpufreq_apply_cooling(struct cpufreq_cooling_device *cpufreq_device,
>>> +                             unsigned long cooling_state)
>>> +{
>>> +     unsigned int event, cpuid, clip_freq;
>>> +     struct cpumask *maskPtr = &cpufreq_device->allowed_cpus;
>>> +     unsigned int cpu = cpumask_any(maskPtr);
>>> +
>>> +
>>> +     /* Check if the old cooling action is same as new cooling action */
>>> +     if (cpufreq_device->cpufreq_state == cooling_state)
>>> +             return 0;
>>> +
>>> +     clip_freq = get_cpu_frequency(cpu, cooling_state);
>>> +     if (!clip_freq)
>>> +             return -EINVAL;
>>> +
>>> +     cpufreq_device->cpufreq_state = cooling_state;
>>> +     cpufreq_device->cpufreq_val = clip_freq;
>>> +     notify_device = cpufreq_device;
>>> +
>>> +     if (cooling_state != 0)
>>> +             event = CPUFREQ_COOLING_START;
>>> +     else
>>> +             event = CPUFREQ_COOLING_STOP;
>>> +
>>> +     blocking_notifier_call_chain(&cputherm_state_notifier_list,
>>> +                                             event, &clip_freq);
>>> +
>>> +     for_each_cpu(cpuid, maskPtr) {
>>> +             if (is_cpufreq_valid(cpuid))
>>> +                     cpufreq_update_policy(cpuid);
>>> +     }
>>> +
>>> +     notify_device = NOTIFY_INVALID;
>>> +
>>> +     return 0;
>>> +}
>>> +
>>> +/**
>>> + * cpufreq_thermal_notifier - notifier callback for cpufreq policy change.
>>> + * @nb:      struct notifier_block * with callback info.
>>> + * @event: value showing cpufreq event for which this function invoked.
>>> + * @data: callback-specific data
>>> + */
>>> +static int cpufreq_thermal_notifier(struct notifier_block *nb,
>>> +                                     unsigned long event, void *data)
>>> +{
>>> +     struct cpufreq_policy *policy = data;
>>> +     unsigned long max_freq = 0;
>>> +
>>> +     if (event != CPUFREQ_ADJUST || notify_device == NOTIFY_INVALID)
>>> +             return 0;
>>> +
>>> +     if (cpumask_test_cpu(policy->cpu, &notify_device->allowed_cpus))
>>> +             max_freq = notify_device->cpufreq_val;
>>> +
>>> +     /* Never exceed user_policy.max*/
>>> +     if (max_freq > policy->user_policy.max)
>>> +             max_freq = policy->user_policy.max;
>>> +
>>> +     if (policy->max != max_freq)
>>> +             cpufreq_verify_within_limits(policy, 0, max_freq);
>>> +
>>> +     return 0;
>>> +}
>>> +
>>> +/*
>>> + * cpufreq cooling device callback functions are defined below
>>> + */
>>> +
>>> +/**
>>> + * cpufreq_get_max_state - callback function to get the max cooling state.
>>> + * @cdev: thermal cooling device pointer.
>>> + * @state: fill this variable with the max cooling state.
>>> + */
>>> +static int cpufreq_get_max_state(struct thermal_cooling_device *cdev,
>>> +                              unsigned long *state)
>>> +{
>>> +     int ret = -EINVAL, i = 0;
>>> +     struct cpufreq_cooling_device *cpufreq_device;
>>> +     struct cpumask *maskPtr;
>>> +     unsigned int cpu;
>>> +     struct cpufreq_frequency_table *table;
>>> +
>>> +     mutex_lock(&cooling_cpufreq_lock);
>>> +     list_for_each_entry(cpufreq_device, &cooling_cpufreq_list, node) {
>>> +             if (cpufreq_device && cpufreq_device->cool_dev == cdev)
>>> +                     break;
>>> +     }
>>> +     if (cpufreq_device == NULL)
>>> +             goto return_get_max_state;
>>> +
>>> +     maskPtr = &cpufreq_device->allowed_cpus;
>>> +     cpu = cpumask_any(maskPtr);
>>> +     table = cpufreq_frequency_get_table(cpu);
>>> +     if (!table) {
>>> +             *state = 0;
>>> +             ret = 0;
>>> +             goto return_get_max_state;
>>> +     }
>>> +
>>> +     while (table[i].frequency != CPUFREQ_TABLE_END) {
>>> +             if (table[i].frequency == CPUFREQ_ENTRY_INVALID)
>>> +                     continue;
>>> +             i++;
>>> +     }
>>> +     if (i > 0) {
>>> +             *state = --i;
>>> +             ret = 0;
>>> +     }
>>> +
>>> +return_get_max_state:
>>> +     mutex_unlock(&cooling_cpufreq_lock);
>>> +     return ret;
>>> +}
>>> +
>>> +/**
>>> + * cpufreq_get_cur_state - callback function to get the current cooling state.
>>> + * @cdev: thermal cooling device pointer.
>>> + * @state: fill this variable with the current cooling state.
>>> + */
>>> +static int cpufreq_get_cur_state(struct thermal_cooling_device *cdev,
>>> +                              unsigned long *state)
>>> +{
>>> +     int ret = -EINVAL;
>>> +     struct cpufreq_cooling_device *cpufreq_device;
>>> +
>>> +     mutex_lock(&cooling_cpufreq_lock);
>>> +     list_for_each_entry(cpufreq_device, &cooling_cpufreq_list, node) {
>>> +             if (cpufreq_device && cpufreq_device->cool_dev == cdev) {
>>> +                     *state = cpufreq_device->cpufreq_state;
>>> +                     ret = 0;
>>> +                     break;
>>> +             }
>>> +     }
>>> +     mutex_unlock(&cooling_cpufreq_lock);
>>> +
>>
>> as cpufreq may be changed in other places, e.g. via sysfs I/F, we should
>> use the current cpu frequency to get the REAL cooling state, rather than
>> using a cached value.
>
> Yes agreed , I will repost with your suggestion.
>
> Thanks,
> Amit
>>
>> thanks,
>> rui
>>
>>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
> the body of a message to [e-mail]
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Zhang Rui Aug. 20, 2012, 2:16 a.m. UTC | #6

On ?, 2012-08-17 at 11:56 +0300, Valentin, Eduardo wrote:
> Hello,

> >>> +
> >>> +
> >>> +1.2 CPU cooling action notifier register/unregister interface
> >>> +1.2.1 int cputherm_register_notifier(struct notifier_block *nb,
> >>> +     unsigned int list)
> >>> +
> >>> +    This interface registers a driver with cpu cooling layer. The driver will
> >>> +    be notified when any cpu cooling action is called.
> >>> +
> >>> +    nb: notifier function to register
> >>> +    list: CPUFREQ_COOLING_START or CPUFREQ_COOLING_STOP
> >>> +
> >>> +1.2.2 int cputherm_unregister_notifier(struct notifier_block *nb,
> >>> +     unsigned int list)
> >>> +
> >>> +    This interface registers a driver with cpu cooling layer. The driver will
> >>> +    be notified when any cpu cooling action is called.
> >>> +
> >>> +    nb: notifier function to register
> >>> +    list: CPUFREQ_COOLING_START or CPUFREQ_COOLING_STOP
> >>
> >> what are these two APIs used for?
> >> I did not see they are used in your patch set, do I miss something?
> > No currently they are not used by my patches. I added them on request
> > from Eduardo and others
> 
> Yeah, this was a suggestion as we didn't really know how the FW part
> would evolve by that time.
> 
> The reasoning is to allow any interested user (in kernel) to be
> notified when max frequency changes.

in this case, the cooling device should be updated.
Say all the target of the thermal_instances of a cooling devices is 0,
which means they want the cpu to run at maximum frequency, when the max
frequency changes, we should set the processor to the new max frequency
as well.
Using notification is not a good idea as user can not handle this
without interacting with the thermal framework.

IMO, we should introduce a new API to handle this, rather than just
notifications to users.

>  Actually, the use case behind
> this is to allow such users to perform some handshaking or stop their
> activities or even change some paramenters, in case the max frequency
> would change.

It is the cooling device driver that notice this change first, and it
should invoke something like thermal_cooling_device_update/rebind() to
tell the thermal framework this change.

>  Ideally it would be possible to nack the cooling
> transition. But that is yet a wider discussion. So far we don't have
> users for this.
> 
yep.
I thought about this before, but I'd prefer to do this when there is a
real user. Or else, we are kind of over-designing here.
how about removing this piece of code for now?

thanks,
rui

> >>
> >>> diff --git a/drivers/thermal/Kconfig b/drivers/thermal/Kconfig
> >>> index 7dd8c34..996003b 100644
> >>> --- a/drivers/thermal/Kconfig
> >>> +++ b/drivers/thermal/Kconfig
> >>> @@ -19,6 +19,17 @@ config THERMAL_HWMON
> >>>       depends on HWMON=y || HWMON=THERMAL
> >>>       default y
> >>>
> >>> +config CPU_THERMAL
> >>> +     bool "generic cpu cooling support"
> >>> +     depends on THERMAL && CPU_FREQ
> >>> +     help
> >>> +       This implements the generic cpu cooling mechanism through frequency
> >>> +       reduction, cpu hotplug and any other ways of reducing temperature. An
> >>> +       ACPI version of this already exists(drivers/acpi/processor_thermal.c).
> >>> +       This will be useful for platforms using the generic thermal interface
> >>> +       and not the ACPI interface.
> >>> +       If you want this support, you should say Y here.
> >>> +
> >>>  config SPEAR_THERMAL
> >>>       bool "SPEAr thermal sensor driver"
> >>>       depends on THERMAL
> >>> diff --git a/drivers/thermal/Makefile b/drivers/thermal/Makefile
> >>> index fd9369a..aae59ad 100644
> >>> --- a/drivers/thermal/Makefile
> >>> +++ b/drivers/thermal/Makefile
> >>> @@ -3,5 +3,6 @@
> >>>  #
> >>>
> >>>  obj-$(CONFIG_THERMAL)                += thermal_sys.o
> >>> +obj-$(CONFIG_CPU_THERMAL)            += cpu_cooling.o
> >>>  obj-$(CONFIG_SPEAR_THERMAL)          += spear_thermal.o
> >>>  obj-$(CONFIG_RCAR_THERMAL)   += rcar_thermal.o
> >>> diff --git a/drivers/thermal/cpu_cooling.c b/drivers/thermal/cpu_cooling.c
> >>> new file mode 100644
> >>> index 0000000..c42e557
> >>> --- /dev/null
> >>> +++ b/drivers/thermal/cpu_cooling.c
> >>> @@ -0,0 +1,512 @@
> >>> +/*
> >>> + *  linux/drivers/thermal/cpu_cooling.c
> >>> + *
> >>> + *  Copyright (C) 2012       Samsung Electronics Co., Ltd(http://www.samsung.com)
> >>> + *  Copyright (C) 2012  Amit Daniel <[e-mail]>
> >>> + *
> >>> + * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> >>> + *  This program is free software; you can redistribute it and/or modify
> >>> + *  it under the terms of the GNU General Public License as published by
> >>> + *  the Free Software Foundation; version 2 of the License.
> >>> + *
> >>> + *  This program is distributed in the hope that it will be useful, but
> >>> + *  WITHOUT ANY WARRANTY; without even the implied warranty of
> >>> + *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
> >>> + *  General Public License for more details.
> >>> + *
> >>> + *  You should have received a copy of the GNU General Public License along
> >>> + *  with this program; if not, write to the Free Software Foundation, Inc.,
> >>> + *  59 Temple Place, Suite 330, Boston, MA 02111-1307 USA.
> >>> + *
> >>> + * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> >>> + */
> >>> +#include <linux/kernel.h>
> >>> +#include <linux/module.h>
> >>> +#include <linux/thermal.h>
> >>> +#include <linux/platform_device.h>
> >>> +#include <linux/cpufreq.h>
> >>> +#include <linux/err.h>
> >>> +#include <linux/slab.h>
> >>> +#include <linux/cpu.h>
> >>> +#include <linux/cpu_cooling.h>
> >>> +
> >>> +/**
> >>> + * struct cpufreq_cooling_device
> >>> + * @id: unique integer value corresponding to each cpufreq_cooling_device
> >>> + *   registered.
> >>> + * @cool_dev: thermal_cooling_device pointer to keep track of the the
> >>> + *   egistered cooling device.
> >>> + * @cpufreq_state: integer value representing the current state of cpufreq
> >>> + *   cooling devices.
> >>> + * @cpufreq_val: integer value representing the absolute value of the clipped
> >>> + *   frequency.
> >>> + * @allowed_cpus: all the cpus involved for this cpufreq_cooling_device.
> >>> + * @node: list_head to link all cpufreq_cooling_device together.
> >>> + *
> >>> + * This structure is required for keeping information of each
> >>> + * cpufreq_cooling_device registered as a list whose head is represented by
> >>> + * cooling_cpufreq_list. In order to prevent corruption of this list a
> >>> + * mutex lock cooling_cpufreq_lock is used.
> >>> + */
> >>> +struct cpufreq_cooling_device {
> >>> +     int id;
> >>> +     struct thermal_cooling_device *cool_dev;
> >>> +     unsigned int cpufreq_state;
> >>> +     unsigned int cpufreq_val;
> >>> +     struct cpumask allowed_cpus;
> >>> +     struct list_head node;
> >>> +};
> >>> +static LIST_HEAD(cooling_cpufreq_list);
> >>> +static DEFINE_IDR(cpufreq_idr);
> >>> +
> >>> +static struct mutex cooling_cpufreq_lock;
> >>> +
> >>> +/* notify_table passes value to the CPUFREQ_ADJUST callback function. */
> >>> +#define NOTIFY_INVALID NULL
> >>> +struct cpufreq_cooling_device *notify_device;
> >>> +
> >>> +/* Head of the blocking notifier chain to inform about frequency clamping */
> >>> +static BLOCKING_NOTIFIER_HEAD(cputherm_state_notifier_list);
> >>> +
> >>> +/**
> >>> + * get_idr - function to get a unique id.
> >>> + * @idr: struct idr * handle used to create a id.
> >>> + * @id: int * value generated by this function.
> >>> + */
> >>> +static int get_idr(struct idr *idr, int *id)
> >>> +{
> >>> +     int err;
> >>> +again:
> >>> +     if (unlikely(idr_pre_get(idr, GFP_KERNEL) == 0))
> >>> +             return -ENOMEM;
> >>> +
> >>> +     mutex_lock(&cooling_cpufreq_lock);
> >>> +     err = idr_get_new(idr, NULL, id);
> >>> +     mutex_unlock(&cooling_cpufreq_lock);
> >>> +
> >>> +     if (unlikely(err == -EAGAIN))
> >>> +             goto again;
> >>> +     else if (unlikely(err))
> >>> +             return err;
> >>> +
> >>> +     *id = *id & MAX_ID_MASK;
> >>> +     return 0;
> >>> +}
> >>> +
> >>> +/**
> >>> + * release_idr - function to free the unique id.
> >>> + * @idr: struct idr * handle used for creating the id.
> >>> + * @id: int value representing the unique id.
> >>> + */
> >>> +static void release_idr(struct idr *idr, int id)
> >>> +{
> >>> +     mutex_lock(&cooling_cpufreq_lock);
> >>> +     idr_remove(idr, id);
> >>> +     mutex_unlock(&cooling_cpufreq_lock);
> >>> +}
> >>> +
> >>> +/**
> >>> + * cputherm_register_notifier - Register a notifier with cpu cooling interface.
> >>> + * @nb:      struct notifier_block * with callback info.
> >>> + * @list: integer value for which notification is needed. possible values are
> >>> + *   CPUFREQ_COOLING_START and CPUFREQ_COOLING_STOP.
> >>> + *
> >>> + * This exported function registers a driver with cpu cooling layer. The driver
> >>> + * will be notified when any cpu cooling action is called.
> >>> + */
> >>> +int cputherm_register_notifier(struct notifier_block *nb, unsigned int list)
> >>> +{
> >>> +     int ret = 0;
> >>> +
> >>> +     switch (list) {
> >>> +     case CPUFREQ_COOLING_START:
> >>> +     case CPUFREQ_COOLING_STOP:
> >>> +             ret = blocking_notifier_chain_register(
> >>> +                             &cputherm_state_notifier_list, nb);
> >>> +             break;
> >>> +     default:
> >>> +             ret = -EINVAL;
> >>> +     }
> >>> +     return ret;
> >>> +}
> >>> +EXPORT_SYMBOL(cputherm_register_notifier);
> >>> +
> >>> +/**
> >>> + * cputherm_unregister_notifier - Un-register a notifier.
> >>> + * @nb:      struct notifier_block * with callback info.
> >>> + * @list: integer value for which notification is needed. values possible are
> >>> + *   CPUFREQ_COOLING_START or CPUFREQ_COOLING_STOP.
> >>> + *
> >>> + * This exported function un-registers a driver with cpu cooling layer.
> >>> + */
> >>> +int cputherm_unregister_notifier(struct notifier_block *nb, unsigned int list)
> >>> +{
> >>> +     int ret = 0;
> >>> +
> >>> +     switch (list) {
> >>> +     case CPUFREQ_COOLING_START:
> >>> +     case CPUFREQ_COOLING_STOP:
> >>> +             ret = blocking_notifier_chain_unregister(
> >>> +                             &cputherm_state_notifier_list, nb);
> >>> +             break;
> >>> +     default:
> >>> +             ret = -EINVAL;
> >>> +     }
> >>> +     return ret;
> >>> +}
> >>> +EXPORT_SYMBOL(cputherm_unregister_notifier);
> >>> +
> >>> +/* Below code defines functions to be used for cpufreq as cooling device */
> >>> +
> >>> +/**
> >>> + * is_cpufreq_valid - function to check if a cpu has frequency transition policy.
> >>> + * @cpu: cpu for which check is needed.
> >>> + */
> >>> +static int is_cpufreq_valid(int cpu)
> >>> +{
> >>> +     struct cpufreq_policy policy;
> >>> +     return !cpufreq_get_policy(&policy, cpu);
> >>> +}
> >>> +
> >>> +/**
> >>> + * get_cpu_frequency - get the absolute value of frequency from level.
> >>> + * @cpu: cpu for which frequency is fetched.
> >>> + * @level: level of frequency of the CPU
> >>> + *   e.g level=1 --> 1st MAX FREQ, LEVEL=2 ---> 2nd MAX FREQ, .... etc
> >>> + */
> >>> +static unsigned int get_cpu_frequency(unsigned int cpu, unsigned long level)
> >>> +{
> >>> +     int ret = 0, i = 0;
> >>> +     unsigned long level_index;
> >>> +     bool descend = false;
> >>> +     struct cpufreq_frequency_table *table =
> >>> +                                     cpufreq_frequency_get_table(cpu);
> >>> +     if (!table)
> >>> +             return ret;
> >>> +
> >>> +     while (table[i].frequency != CPUFREQ_TABLE_END) {
> >>> +             if (table[i].frequency == CPUFREQ_ENTRY_INVALID)
> >>> +                     continue;
> >>> +
> >>> +             /*check if table in ascending or descending order*/
> >>> +             if ((table[i + 1].frequency != CPUFREQ_TABLE_END) &&
> >>> +                     (table[i + 1].frequency < table[i].frequency)
> >>> +                     && !descend) {
> >>> +                     descend = true;
> >>> +             }
> >>> +
> >>> +             /*return if level matched and table in descending order*/
> >>> +             if (descend && i == level)
> >>> +                     return table[i].frequency;
> >>> +             i++;
> >>> +     }
> >>> +     i--;
> >>> +
> >>> +     if (level > i || descend)
> >>> +             return ret;
> >>> +     level_index = i - level;
> >>> +
> >>> +     /*Scan the table in reverse order and match the level*/
> >>> +     while (i >= 0) {
> >>> +             if (table[i].frequency == CPUFREQ_ENTRY_INVALID)
> >>> +                     continue;
> >>> +             /*return if level matched*/
> >>> +             if (i == level_index)
> >>> +                     return table[i].frequency;
> >>> +             i--;
> >>> +     }
> >>> +     return ret;
> >>> +}
> >>> +
> >>> +/**
> >>> + * cpufreq_apply_cooling - function to apply frequency clipping.
> >>> + * @cpufreq_device: cpufreq_cooling_device pointer containing frequency
> >>> + *   clipping data.
> >>> + * @cooling_state: value of the cooling state.
> >>> + */
> >>> +static int cpufreq_apply_cooling(struct cpufreq_cooling_device *cpufreq_device,
> >>> +                             unsigned long cooling_state)
> >>> +{
> >>> +     unsigned int event, cpuid, clip_freq;
> >>> +     struct cpumask *maskPtr = &cpufreq_device->allowed_cpus;
> >>> +     unsigned int cpu = cpumask_any(maskPtr);
> >>> +
> >>> +
> >>> +     /* Check if the old cooling action is same as new cooling action */
> >>> +     if (cpufreq_device->cpufreq_state == cooling_state)
> >>> +             return 0;
> >>> +
> >>> +     clip_freq = get_cpu_frequency(cpu, cooling_state);
> >>> +     if (!clip_freq)
> >>> +             return -EINVAL;
> >>> +
> >>> +     cpufreq_device->cpufreq_state = cooling_state;
> >>> +     cpufreq_device->cpufreq_val = clip_freq;
> >>> +     notify_device = cpufreq_device;
> >>> +
> >>> +     if (cooling_state != 0)
> >>> +             event = CPUFREQ_COOLING_START;
> >>> +     else
> >>> +             event = CPUFREQ_COOLING_STOP;
> >>> +
> >>> +     blocking_notifier_call_chain(&cputherm_state_notifier_list,
> >>> +                                             event, &clip_freq);
> >>> +
> >>> +     for_each_cpu(cpuid, maskPtr) {
> >>> +             if (is_cpufreq_valid(cpuid))
> >>> +                     cpufreq_update_policy(cpuid);
> >>> +     }
> >>> +
> >>> +     notify_device = NOTIFY_INVALID;
> >>> +
> >>> +     return 0;
> >>> +}
> >>> +
> >>> +/**
> >>> + * cpufreq_thermal_notifier - notifier callback for cpufreq policy change.
> >>> + * @nb:      struct notifier_block * with callback info.
> >>> + * @event: value showing cpufreq event for which this function invoked.
> >>> + * @data: callback-specific data
> >>> + */
> >>> +static int cpufreq_thermal_notifier(struct notifier_block *nb,
> >>> +                                     unsigned long event, void *data)
> >>> +{
> >>> +     struct cpufreq_policy *policy = data;
> >>> +     unsigned long max_freq = 0;
> >>> +
> >>> +     if (event != CPUFREQ_ADJUST || notify_device == NOTIFY_INVALID)
> >>> +             return 0;
> >>> +
> >>> +     if (cpumask_test_cpu(policy->cpu, &notify_device->allowed_cpus))
> >>> +             max_freq = notify_device->cpufreq_val;
> >>> +
> >>> +     /* Never exceed user_policy.max*/
> >>> +     if (max_freq > policy->user_policy.max)
> >>> +             max_freq = policy->user_policy.max;
> >>> +
> >>> +     if (policy->max != max_freq)
> >>> +             cpufreq_verify_within_limits(policy, 0, max_freq);
> >>> +
> >>> +     return 0;
> >>> +}
> >>> +
> >>> +/*
> >>> + * cpufreq cooling device callback functions are defined below
> >>> + */
> >>> +
> >>> +/**
> >>> + * cpufreq_get_max_state - callback function to get the max cooling state.
> >>> + * @cdev: thermal cooling device pointer.
> >>> + * @state: fill this variable with the max cooling state.
> >>> + */
> >>> +static int cpufreq_get_max_state(struct thermal_cooling_device *cdev,
> >>> +                              unsigned long *state)
> >>> +{
> >>> +     int ret = -EINVAL, i = 0;
> >>> +     struct cpufreq_cooling_device *cpufreq_device;
> >>> +     struct cpumask *maskPtr;
> >>> +     unsigned int cpu;
> >>> +     struct cpufreq_frequency_table *table;
> >>> +
> >>> +     mutex_lock(&cooling_cpufreq_lock);
> >>> +     list_for_each_entry(cpufreq_device, &cooling_cpufreq_list, node) {
> >>> +             if (cpufreq_device && cpufreq_device->cool_dev == cdev)
> >>> +                     break;
> >>> +     }
> >>> +     if (cpufreq_device == NULL)
> >>> +             goto return_get_max_state;
> >>> +
> >>> +     maskPtr = &cpufreq_device->allowed_cpus;
> >>> +     cpu = cpumask_any(maskPtr);
> >>> +     table = cpufreq_frequency_get_table(cpu);
> >>> +     if (!table) {
> >>> +             *state = 0;
> >>> +             ret = 0;
> >>> +             goto return_get_max_state;
> >>> +     }
> >>> +
> >>> +     while (table[i].frequency != CPUFREQ_TABLE_END) {
> >>> +             if (table[i].frequency == CPUFREQ_ENTRY_INVALID)
> >>> +                     continue;
> >>> +             i++;
> >>> +     }
> >>> +     if (i > 0) {
> >>> +             *state = --i;
> >>> +             ret = 0;
> >>> +     }
> >>> +
> >>> +return_get_max_state:
> >>> +     mutex_unlock(&cooling_cpufreq_lock);
> >>> +     return ret;
> >>> +}
> >>> +
> >>> +/**
> >>> + * cpufreq_get_cur_state - callback function to get the current cooling state.
> >>> + * @cdev: thermal cooling device pointer.
> >>> + * @state: fill this variable with the current cooling state.
> >>> + */
> >>> +static int cpufreq_get_cur_state(struct thermal_cooling_device *cdev,
> >>> +                              unsigned long *state)
> >>> +{
> >>> +     int ret = -EINVAL;
> >>> +     struct cpufreq_cooling_device *cpufreq_device;
> >>> +
> >>> +     mutex_lock(&cooling_cpufreq_lock);
> >>> +     list_for_each_entry(cpufreq_device, &cooling_cpufreq_list, node) {
> >>> +             if (cpufreq_device && cpufreq_device->cool_dev == cdev) {
> >>> +                     *state = cpufreq_device->cpufreq_state;
> >>> +                     ret = 0;
> >>> +                     break;
> >>> +             }
> >>> +     }
> >>> +     mutex_unlock(&cooling_cpufreq_lock);
> >>> +
> >>
> >> as cpufreq may be changed in other places, e.g. via sysfs I/F, we should
> >> use the current cpu frequency to get the REAL cooling state, rather than
> >> using a cached value.
> >
> > Yes agreed , I will repost with your suggestion.
> >
> > Thanks,
> > Amit
> >>
> >> thanks,
> >> rui
> >>
> >>
> > --
> > To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
> > the body of a message to [e-mail]
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 
> 
> 


--
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
the body of a message to [e-mail]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Eduardo Valentin Aug. 20, 2012, 9:36 a.m. UTC | #7

Hello Rui,

On Mon, Aug 20, 2012 at 5:16 AM, Zhang Rui <[e-mail]> wrote:
> On ?, 2012-08-17 at 11:56 +0300, Valentin, Eduardo wrote:
>> Hello,
>
>> >>> +
>> >>> +
>> >>> +1.2 CPU cooling action notifier register/unregister interface
>> >>> +1.2.1 int cputherm_register_notifier(struct notifier_block *nb,
>> >>> +     unsigned int list)
>> >>> +
>> >>> +    This interface registers a driver with cpu cooling layer. The driver will
>> >>> +    be notified when any cpu cooling action is called.
>> >>> +
>> >>> +    nb: notifier function to register
>> >>> +    list: CPUFREQ_COOLING_START or CPUFREQ_COOLING_STOP
>> >>> +
>> >>> +1.2.2 int cputherm_unregister_notifier(struct notifier_block *nb,
>> >>> +     unsigned int list)
>> >>> +
>> >>> +    This interface registers a driver with cpu cooling layer. The driver will
>> >>> +    be notified when any cpu cooling action is called.
>> >>> +
>> >>> +    nb: notifier function to register
>> >>> +    list: CPUFREQ_COOLING_START or CPUFREQ_COOLING_STOP
>> >>
>> >> what are these two APIs used for?
>> >> I did not see they are used in your patch set, do I miss something?
>> > No currently they are not used by my patches. I added them on request
>> > from Eduardo and others
>>
>> Yeah, this was a suggestion as we didn't really know how the FW part
>> would evolve by that time.
>>
>> The reasoning is to allow any interested user (in kernel) to be
>> notified when max frequency changes.
>
> in this case, the cooling device should be updated.
> Say all the target of the thermal_instances of a cooling devices is 0,
> which means they want the cpu to run at maximum frequency, when the max
> frequency changes, we should set the processor to the new max frequency
> as well.
> Using notification is not a good idea as user can not handle this
> without interacting with the thermal framework.
>
> IMO, we should introduce a new API to handle this, rather than just
> notifications to users.

Agreed. The context is actually much wider than the CPU cooling. In
fact, cooling co-processors is actually where things gets a bit
complicated. The idea behind this type of handshaking is to allow the
affected subsystem to change their actual setup when max performance
is not allowed anymore.

>
>>  Actually, the use case behind
>> this is to allow such users to perform some handshaking or stop their
>> activities or even change some paramenters, in case the max frequency
>> would change.
>
> It is the cooling device driver that notice this change first, and it
> should invoke something like thermal_cooling_device_update/rebind() to
> tell the thermal framework this change.
>

Yeah. Ideally, the framework needs to be aware of cooling device state
change. Today, as we have pretty much a broadcast/unidirectional type
of messaging, the framework/policy always assumes the cooling devices
will be in sync with what it is dictated by the policy.

>>  Ideally it would be possible to nack the cooling
>> transition. But that is yet a wider discussion. So far we don't have
>> users for this.
>>
> yep.
> I thought about this before, but I'd prefer to do this when there is a
> real user. Or else, we are kind of over-designing here.
> how about removing this piece of code for now?


Agreed. I second you that this problem is a much wider issue and needs
improvement in the FW itself and how the cooling devices are designed.

>
> thanks,
> rui
>
>> >>
>> >>> diff --git a/drivers/thermal/Kconfig b/drivers/thermal/Kconfig
>> >>> index 7dd8c34..996003b 100644
>> >>> --- a/drivers/thermal/Kconfig
>> >>> +++ b/drivers/thermal/Kconfig
>> >>> @@ -19,6 +19,17 @@ config THERMAL_HWMON
>> >>>       depends on HWMON=y || HWMON=THERMAL
>> >>>       default y
>> >>>
>> >>> +config CPU_THERMAL
>> >>> +     bool "generic cpu cooling support"
>> >>> +     depends on THERMAL && CPU_FREQ
>> >>> +     help
>> >>> +       This implements the generic cpu cooling mechanism through frequency
>> >>> +       reduction, cpu hotplug and any other ways of reducing temperature. An
>> >>> +       ACPI version of this already exists(drivers/acpi/processor_thermal.c).
>> >>> +       This will be useful for platforms using the generic thermal interface
>> >>> +       and not the ACPI interface.
>> >>> +       If you want this support, you should say Y here.
>> >>> +
>> >>>  config SPEAR_THERMAL
>> >>>       bool "SPEAr thermal sensor driver"
>> >>>       depends on THERMAL
>> >>> diff --git a/drivers/thermal/Makefile b/drivers/thermal/Makefile
>> >>> index fd9369a..aae59ad 100644
>> >>> --- a/drivers/thermal/Makefile
>> >>> +++ b/drivers/thermal/Makefile
>> >>> @@ -3,5 +3,6 @@
>> >>>  #
>> >>>
>> >>>  obj-$(CONFIG_THERMAL)                += thermal_sys.o
>> >>> +obj-$(CONFIG_CPU_THERMAL)            += cpu_cooling.o
>> >>>  obj-$(CONFIG_SPEAR_THERMAL)          += spear_thermal.o
>> >>>  obj-$(CONFIG_RCAR_THERMAL)   += rcar_thermal.o
>> >>> diff --git a/drivers/thermal/cpu_cooling.c b/drivers/thermal/cpu_cooling.c
>> >>> new file mode 100644
>> >>> index 0000000..c42e557
>> >>> --- /dev/null
>> >>> +++ b/drivers/thermal/cpu_cooling.c
>> >>> @@ -0,0 +1,512 @@
>> >>> +/*
>> >>> + *  linux/drivers/thermal/cpu_cooling.c
>> >>> + *
>> >>> + *  Copyright (C) 2012       Samsung Electronics Co., Ltd(http://www.samsung.com)
>> >>> + *  Copyright (C) 2012  Amit Daniel <[e-mail]>
>> >>> + *
>> >>> + * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>> >>> + *  This program is free software; you can redistribute it and/or modify
>> >>> + *  it under the terms of the GNU General Public License as published by
>> >>> + *  the Free Software Foundation; version 2 of the License.
>> >>> + *
>> >>> + *  This program is distributed in the hope that it will be useful, but
>> >>> + *  WITHOUT ANY WARRANTY; without even the implied warranty of
>> >>> + *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
>> >>> + *  General Public License for more details.
>> >>> + *
>> >>> + *  You should have received a copy of the GNU General Public License along
>> >>> + *  with this program; if not, write to the Free Software Foundation, Inc.,
>> >>> + *  59 Temple Place, Suite 330, Boston, MA 02111-1307 USA.
>> >>> + *
>> >>> + * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>> >>> + */
>> >>> +#include <linux/kernel.h>
>> >>> +#include <linux/module.h>
>> >>> +#include <linux/thermal.h>
>> >>> +#include <linux/platform_device.h>
>> >>> +#include <linux/cpufreq.h>
>> >>> +#include <linux/err.h>
>> >>> +#include <linux/slab.h>
>> >>> +#include <linux/cpu.h>
>> >>> +#include <linux/cpu_cooling.h>
>> >>> +
>> >>> +/**
>> >>> + * struct cpufreq_cooling_device
>> >>> + * @id: unique integer value corresponding to each cpufreq_cooling_device
>> >>> + *   registered.
>> >>> + * @cool_dev: thermal_cooling_device pointer to keep track of the the
>> >>> + *   egistered cooling device.
>> >>> + * @cpufreq_state: integer value representing the current state of cpufreq
>> >>> + *   cooling devices.
>> >>> + * @cpufreq_val: integer value representing the absolute value of the clipped
>> >>> + *   frequency.
>> >>> + * @allowed_cpus: all the cpus involved for this cpufreq_cooling_device.
>> >>> + * @node: list_head to link all cpufreq_cooling_device together.
>> >>> + *
>> >>> + * This structure is required for keeping information of each
>> >>> + * cpufreq_cooling_device registered as a list whose head is represented by
>> >>> + * cooling_cpufreq_list. In order to prevent corruption of this list a
>> >>> + * mutex lock cooling_cpufreq_lock is used.
>> >>> + */
>> >>> +struct cpufreq_cooling_device {
>> >>> +     int id;
>> >>> +     struct thermal_cooling_device *cool_dev;
>> >>> +     unsigned int cpufreq_state;
>> >>> +     unsigned int cpufreq_val;
>> >>> +     struct cpumask allowed_cpus;
>> >>> +     struct list_head node;
>> >>> +};
>> >>> +static LIST_HEAD(cooling_cpufreq_list);
>> >>> +static DEFINE_IDR(cpufreq_idr);
>> >>> +
>> >>> +static struct mutex cooling_cpufreq_lock;
>> >>> +
>> >>> +/* notify_table passes value to the CPUFREQ_ADJUST callback function. */
>> >>> +#define NOTIFY_INVALID NULL
>> >>> +struct cpufreq_cooling_device *notify_device;
>> >>> +
>> >>> +/* Head of the blocking notifier chain to inform about frequency clamping */
>> >>> +static BLOCKING_NOTIFIER_HEAD(cputherm_state_notifier_list);
>> >>> +
>> >>> +/**
>> >>> + * get_idr - function to get a unique id.
>> >>> + * @idr: struct idr * handle used to create a id.
>> >>> + * @id: int * value generated by this function.
>> >>> + */
>> >>> +static int get_idr(struct idr *idr, int *id)
>> >>> +{
>> >>> +     int err;
>> >>> +again:
>> >>> +     if (unlikely(idr_pre_get(idr, GFP_KERNEL) == 0))
>> >>> +             return -ENOMEM;
>> >>> +
>> >>> +     mutex_lock(&cooling_cpufreq_lock);
>> >>> +     err = idr_get_new(idr, NULL, id);
>> >>> +     mutex_unlock(&cooling_cpufreq_lock);
>> >>> +
>> >>> +     if (unlikely(err == -EAGAIN))
>> >>> +             goto again;
>> >>> +     else if (unlikely(err))
>> >>> +             return err;
>> >>> +
>> >>> +     *id = *id & MAX_ID_MASK;
>> >>> +     return 0;
>> >>> +}
>> >>> +
>> >>> +/**
>> >>> + * release_idr - function to free the unique id.
>> >>> + * @idr: struct idr * handle used for creating the id.
>> >>> + * @id: int value representing the unique id.
>> >>> + */
>> >>> +static void release_idr(struct idr *idr, int id)
>> >>> +{
>> >>> +     mutex_lock(&cooling_cpufreq_lock);
>> >>> +     idr_remove(idr, id);
>> >>> +     mutex_unlock(&cooling_cpufreq_lock);
>> >>> +}
>> >>> +
>> >>> +/**
>> >>> + * cputherm_register_notifier - Register a notifier with cpu cooling interface.
>> >>> + * @nb:      struct notifier_block * with callback info.
>> >>> + * @list: integer value for which notification is needed. possible values are
>> >>> + *   CPUFREQ_COOLING_START and CPUFREQ_COOLING_STOP.
>> >>> + *
>> >>> + * This exported function registers a driver with cpu cooling layer. The driver
>> >>> + * will be notified when any cpu cooling action is called.
>> >>> + */
>> >>> +int cputherm_register_notifier(struct notifier_block *nb, unsigned int list)
>> >>> +{
>> >>> +     int ret = 0;
>> >>> +
>> >>> +     switch (list) {
>> >>> +     case CPUFREQ_COOLING_START:
>> >>> +     case CPUFREQ_COOLING_STOP:
>> >>> +             ret = blocking_notifier_chain_register(
>> >>> +                             &cputherm_state_notifier_list, nb);
>> >>> +             break;
>> >>> +     default:
>> >>> +             ret = -EINVAL;
>> >>> +     }
>> >>> +     return ret;
>> >>> +}
>> >>> +EXPORT_SYMBOL(cputherm_register_notifier);
>> >>> +
>> >>> +/**
>> >>> + * cputherm_unregister_notifier - Un-register a notifier.
>> >>> + * @nb:      struct notifier_block * with callback info.
>> >>> + * @list: integer value for which notification is needed. values possible are
>> >>> + *   CPUFREQ_COOLING_START or CPUFREQ_COOLING_STOP.
>> >>> + *
>> >>> + * This exported function un-registers a driver with cpu cooling layer.
>> >>> + */
>> >>> +int cputherm_unregister_notifier(struct notifier_block *nb, unsigned int list)
>> >>> +{
>> >>> +     int ret = 0;
>> >>> +
>> >>> +     switch (list) {
>> >>> +     case CPUFREQ_COOLING_START:
>> >>> +     case CPUFREQ_COOLING_STOP:
>> >>> +             ret = blocking_notifier_chain_unregister(
>> >>> +                             &cputherm_state_notifier_list, nb);
>> >>> +             break;
>> >>> +     default:
>> >>> +             ret = -EINVAL;
>> >>> +     }
>> >>> +     return ret;
>> >>> +}
>> >>> +EXPORT_SYMBOL(cputherm_unregister_notifier);
>> >>> +
>> >>> +/* Below code defines functions to be used for cpufreq as cooling device */
>> >>> +
>> >>> +/**
>> >>> + * is_cpufreq_valid - function to check if a cpu has frequency transition policy.
>> >>> + * @cpu: cpu for which check is needed.
>> >>> + */
>> >>> +static int is_cpufreq_valid(int cpu)
>> >>> +{
>> >>> +     struct cpufreq_policy policy;
>> >>> +     return !cpufreq_get_policy(&policy, cpu);
>> >>> +}
>> >>> +
>> >>> +/**
>> >>> + * get_cpu_frequency - get the absolute value of frequency from level.
>> >>> + * @cpu: cpu for which frequency is fetched.
>> >>> + * @level: level of frequency of the CPU
>> >>> + *   e.g level=1 --> 1st MAX FREQ, LEVEL=2 ---> 2nd MAX FREQ, .... etc
>> >>> + */
>> >>> +static unsigned int get_cpu_frequency(unsigned int cpu, unsigned long level)
>> >>> +{
>> >>> +     int ret = 0, i = 0;
>> >>> +     unsigned long level_index;
>> >>> +     bool descend = false;
>> >>> +     struct cpufreq_frequency_table *table =
>> >>> +                                     cpufreq_frequency_get_table(cpu);
>> >>> +     if (!table)
>> >>> +             return ret;
>> >>> +
>> >>> +     while (table[i].frequency != CPUFREQ_TABLE_END) {
>> >>> +             if (table[i].frequency == CPUFREQ_ENTRY_INVALID)
>> >>> +                     continue;
>> >>> +
>> >>> +             /*check if table in ascending or descending order*/
>> >>> +             if ((table[i + 1].frequency != CPUFREQ_TABLE_END) &&
>> >>> +                     (table[i + 1].frequency < table[i].frequency)
>> >>> +                     && !descend) {
>> >>> +                     descend = true;
>> >>> +             }
>> >>> +
>> >>> +             /*return if level matched and table in descending order*/
>> >>> +             if (descend && i == level)
>> >>> +                     return table[i].frequency;
>> >>> +             i++;
>> >>> +     }
>> >>> +     i--;
>> >>> +
>> >>> +     if (level > i || descend)
>> >>> +             return ret;
>> >>> +     level_index = i - level;
>> >>> +
>> >>> +     /*Scan the table in reverse order and match the level*/
>> >>> +     while (i >= 0) {
>> >>> +             if (table[i].frequency == CPUFREQ_ENTRY_INVALID)
>> >>> +                     continue;
>> >>> +             /*return if level matched*/
>> >>> +             if (i == level_index)
>> >>> +                     return table[i].frequency;
>> >>> +             i--;
>> >>> +     }
>> >>> +     return ret;
>> >>> +}
>> >>> +
>> >>> +/**
>> >>> + * cpufreq_apply_cooling - function to apply frequency clipping.
>> >>> + * @cpufreq_device: cpufreq_cooling_device pointer containing frequency
>> >>> + *   clipping data.
>> >>> + * @cooling_state: value of the cooling state.
>> >>> + */
>> >>> +static int cpufreq_apply_cooling(struct cpufreq_cooling_device *cpufreq_device,
>> >>> +                             unsigned long cooling_state)
>> >>> +{
>> >>> +     unsigned int event, cpuid, clip_freq;
>> >>> +     struct cpumask *maskPtr = &cpufreq_device->allowed_cpus;
>> >>> +     unsigned int cpu = cpumask_any(maskPtr);
>> >>> +
>> >>> +
>> >>> +     /* Check if the old cooling action is same as new cooling action */
>> >>> +     if (cpufreq_device->cpufreq_state == cooling_state)
>> >>> +             return 0;
>> >>> +
>> >>> +     clip_freq = get_cpu_frequency(cpu, cooling_state);
>> >>> +     if (!clip_freq)
>> >>> +             return -EINVAL;
>> >>> +
>> >>> +     cpufreq_device->cpufreq_state = cooling_state;
>> >>> +     cpufreq_device->cpufreq_val = clip_freq;
>> >>> +     notify_device = cpufreq_device;
>> >>> +
>> >>> +     if (cooling_state != 0)
>> >>> +             event = CPUFREQ_COOLING_START;
>> >>> +     else
>> >>> +             event = CPUFREQ_COOLING_STOP;
>> >>> +
>> >>> +     blocking_notifier_call_chain(&cputherm_state_notifier_list,
>> >>> +                                             event, &clip_freq);
>> >>> +
>> >>> +     for_each_cpu(cpuid, maskPtr) {
>> >>> +             if (is_cpufreq_valid(cpuid))
>> >>> +                     cpufreq_update_policy(cpuid);
>> >>> +     }
>> >>> +
>> >>> +     notify_device = NOTIFY_INVALID;
>> >>> +
>> >>> +     return 0;
>> >>> +}
>> >>> +
>> >>> +/**
>> >>> + * cpufreq_thermal_notifier - notifier callback for cpufreq policy change.
>> >>> + * @nb:      struct notifier_block * with callback info.
>> >>> + * @event: value showing cpufreq event for which this function invoked.
>> >>> + * @data: callback-specific data
>> >>> + */
>> >>> +static int cpufreq_thermal_notifier(struct notifier_block *nb,
>> >>> +                                     unsigned long event, void *data)
>> >>> +{
>> >>> +     struct cpufreq_policy *policy = data;
>> >>> +     unsigned long max_freq = 0;
>> >>> +
>> >>> +     if (event != CPUFREQ_ADJUST || notify_device == NOTIFY_INVALID)
>> >>> +             return 0;
>> >>> +
>> >>> +     if (cpumask_test_cpu(policy->cpu, &notify_device->allowed_cpus))
>> >>> +             max_freq = notify_device->cpufreq_val;
>> >>> +
>> >>> +     /* Never exceed user_policy.max*/
>> >>> +     if (max_freq > policy->user_policy.max)
>> >>> +             max_freq = policy->user_policy.max;
>> >>> +
>> >>> +     if (policy->max != max_freq)
>> >>> +             cpufreq_verify_within_limits(policy, 0, max_freq);
>> >>> +
>> >>> +     return 0;
>> >>> +}
>> >>> +
>> >>> +/*
>> >>> + * cpufreq cooling device callback functions are defined below
>> >>> + */
>> >>> +
>> >>> +/**
>> >>> + * cpufreq_get_max_state - callback function to get the max cooling state.
>> >>> + * @cdev: thermal cooling device pointer.
>> >>> + * @state: fill this variable with the max cooling state.
>> >>> + */
>> >>> +static int cpufreq_get_max_state(struct thermal_cooling_device *cdev,
>> >>> +                              unsigned long *state)
>> >>> +{
>> >>> +     int ret = -EINVAL, i = 0;
>> >>> +     struct cpufreq_cooling_device *cpufreq_device;
>> >>> +     struct cpumask *maskPtr;
>> >>> +     unsigned int cpu;
>> >>> +     struct cpufreq_frequency_table *table;
>> >>> +
>> >>> +     mutex_lock(&cooling_cpufreq_lock);
>> >>> +     list_for_each_entry(cpufreq_device, &cooling_cpufreq_list, node) {
>> >>> +             if (cpufreq_device && cpufreq_device->cool_dev == cdev)
>> >>> +                     break;
>> >>> +     }
>> >>> +     if (cpufreq_device == NULL)
>> >>> +             goto return_get_max_state;
>> >>> +
>> >>> +     maskPtr = &cpufreq_device->allowed_cpus;
>> >>> +     cpu = cpumask_any(maskPtr);
>> >>> +     table = cpufreq_frequency_get_table(cpu);
>> >>> +     if (!table) {
>> >>> +             *state = 0;
>> >>> +             ret = 0;
>> >>> +             goto return_get_max_state;
>> >>> +     }
>> >>> +
>> >>> +     while (table[i].frequency != CPUFREQ_TABLE_END) {
>> >>> +             if (table[i].frequency == CPUFREQ_ENTRY_INVALID)
>> >>> +                     continue;
>> >>> +             i++;
>> >>> +     }
>> >>> +     if (i > 0) {
>> >>> +             *state = --i;
>> >>> +             ret = 0;
>> >>> +     }
>> >>> +
>> >>> +return_get_max_state:
>> >>> +     mutex_unlock(&cooling_cpufreq_lock);
>> >>> +     return ret;
>> >>> +}
>> >>> +
>> >>> +/**
>> >>> + * cpufreq_get_cur_state - callback function to get the current cooling state.
>> >>> + * @cdev: thermal cooling device pointer.
>> >>> + * @state: fill this variable with the current cooling state.
>> >>> + */
>> >>> +static int cpufreq_get_cur_state(struct thermal_cooling_device *cdev,
>> >>> +                              unsigned long *state)
>> >>> +{
>> >>> +     int ret = -EINVAL;
>> >>> +     struct cpufreq_cooling_device *cpufreq_device;
>> >>> +
>> >>> +     mutex_lock(&cooling_cpufreq_lock);
>> >>> +     list_for_each_entry(cpufreq_device, &cooling_cpufreq_list, node) {
>> >>> +             if (cpufreq_device && cpufreq_device->cool_dev == cdev) {
>> >>> +                     *state = cpufreq_device->cpufreq_state;
>> >>> +                     ret = 0;
>> >>> +                     break;
>> >>> +             }
>> >>> +     }
>> >>> +     mutex_unlock(&cooling_cpufreq_lock);
>> >>> +
>> >>
>> >> as cpufreq may be changed in other places, e.g. via sysfs I/F, we should
>> >> use the current cpu frequency to get the REAL cooling state, rather than
>> >> using a cached value.
>> >
>> > Yes agreed , I will repost with your suggestion.
>> >
>> > Thanks,
>> > Amit
>> >>
>> >> thanks,
>> >> rui
>> >>
>> >>
>> > --
>> > To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
>> > the body of a message to [e-mail]
>> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
>>
>>
>>
>
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
> the body of a message to [e-mail]
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Patch

diff --git a/Documentation/thermal/cpu-cooling-api.txt b/Documentation/thermal/cpu-cooling-api.txt
new file mode 100644
index 0000000..a1f2a6b
--- /dev/null
+++ b/Documentation/thermal/cpu-cooling-api.txt
@@ -0,0 +1,52 @@ 
+CPU cooling APIs How To
+===================================
+
+Written by Amit Daniel Kachhap <[e-mail]>
+
+Updated: 12 May 2012
+
+Copyright (c)  2012 Samsung Electronics Co., Ltd(http://www.samsung.com)
+
+0. Introduction
+
+The generic cpu cooling(freq clipping) provides registration/unregistration APIs
+to the caller. The binding of the cooling devices to the trip point is left for
+the user. The registration APIs returns the cooling device pointer.
+
+1. cpu cooling APIs
+
+1.1 cpufreq registration/unregistration APIs
+1.1.1 struct thermal_cooling_device *cpufreq_cooling_register(
+	struct cpumask *clip_cpus)
+
+    This interface function registers the cpufreq cooling device with the name
+    "thermal-cpufreq-%x". This api can support multiple instances of cpufreq
+    cooling devices.
+
+   clip_cpus: cpumask of cpus where the frequency constraints will happen.
+
+1.1.2 void cpufreq_cooling_unregister(struct thermal_cooling_device *cdev)
+
+    This interface function unregisters the "thermal-cpufreq-%x" cooling device.
+
+    cdev: Cooling device pointer which has to be unregistered.
+
+
+1.2 CPU cooling action notifier register/unregister interface
+1.2.1 int cputherm_register_notifier(struct notifier_block *nb,
+	unsigned int list)
+
+    This interface registers a driver with cpu cooling layer. The driver will
+    be notified when any cpu cooling action is called.
+
+    nb: notifier function to register
+    list: CPUFREQ_COOLING_START or CPUFREQ_COOLING_STOP
+
+1.2.2 int cputherm_unregister_notifier(struct notifier_block *nb,
+	unsigned int list)
+
+    This interface registers a driver with cpu cooling layer. The driver will
+    be notified when any cpu cooling action is called.
+
+    nb: notifier function to register
+    list: CPUFREQ_COOLING_START or CPUFREQ_COOLING_STOP
diff --git a/drivers/thermal/Kconfig b/drivers/thermal/Kconfig
index 7dd8c34..996003b 100644
--- a/drivers/thermal/Kconfig
+++ b/drivers/thermal/Kconfig
@@ -19,6 +19,17 @@  config THERMAL_HWMON
 	depends on HWMON=y || HWMON=THERMAL
 	default y
 
+config CPU_THERMAL
+	bool "generic cpu cooling support"
+	depends on THERMAL && CPU_FREQ
+	help
+	  This implements the generic cpu cooling mechanism through frequency
+	  reduction, cpu hotplug and any other ways of reducing temperature. An
+	  ACPI version of this already exists(drivers/acpi/processor_thermal.c).
+	  This will be useful for platforms using the generic thermal interface
+	  and not the ACPI interface.
+	  If you want this support, you should say Y here.
+
 config SPEAR_THERMAL
 	bool "SPEAr thermal sensor driver"
 	depends on THERMAL
diff --git a/drivers/thermal/Makefile b/drivers/thermal/Makefile
index fd9369a..aae59ad 100644
--- a/drivers/thermal/Makefile
+++ b/drivers/thermal/Makefile
@@ -3,5 +3,6 @@ 
 #
 
 obj-$(CONFIG_THERMAL)		+= thermal_sys.o
+obj-$(CONFIG_CPU_THERMAL)		+= cpu_cooling.o
 obj-$(CONFIG_SPEAR_THERMAL)		+= spear_thermal.o
 obj-$(CONFIG_RCAR_THERMAL)	+= rcar_thermal.o
diff --git a/drivers/thermal/cpu_cooling.c b/drivers/thermal/cpu_cooling.c
new file mode 100644
index 0000000..c42e557
--- /dev/null
+++ b/drivers/thermal/cpu_cooling.c
@@ -0,0 +1,512 @@ 
+/*
+ *  linux/drivers/thermal/cpu_cooling.c
+ *
+ *  Copyright (C) 2012	Samsung Electronics Co., Ltd(http://www.samsung.com)
+ *  Copyright (C) 2012  Amit Daniel <[e-mail]>
+ *
+ * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License as published by
+ *  the Free Software Foundation; version 2 of the License.
+ *
+ *  This program is distributed in the hope that it will be useful, but
+ *  WITHOUT ANY WARRANTY; without even the implied warranty of
+ *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ *  General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License along
+ *  with this program; if not, write to the Free Software Foundation, Inc.,
+ *  59 Temple Place, Suite 330, Boston, MA 02111-1307 USA.
+ *
+ * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+ */
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/thermal.h>
+#include <linux/platform_device.h>
+#include <linux/cpufreq.h>
+#include <linux/err.h>
+#include <linux/slab.h>
+#include <linux/cpu.h>
+#include <linux/cpu_cooling.h>
+
+/**
+ * struct cpufreq_cooling_device
+ * @id: unique integer value corresponding to each cpufreq_cooling_device
+ *	registered.
+ * @cool_dev: thermal_cooling_device pointer to keep track of the the
+ *	egistered cooling device.
+ * @cpufreq_state: integer value representing the current state of cpufreq
+ *	cooling	devices.
+ * @cpufreq_val: integer value representing the absolute value of the clipped
+ *	frequency.
+ * @allowed_cpus: all the cpus involved for this cpufreq_cooling_device.
+ * @node: list_head to link all cpufreq_cooling_device together.
+ *
+ * This structure is required for keeping information of each
+ * cpufreq_cooling_device registered as a list whose head is represented by
+ * cooling_cpufreq_list. In order to prevent corruption of this list a
+ * mutex lock cooling_cpufreq_lock is used.
+ */
+struct cpufreq_cooling_device {
+	int id;
+	struct thermal_cooling_device *cool_dev;
+	unsigned int cpufreq_state;
+	unsigned int cpufreq_val;
+	struct cpumask allowed_cpus;
+	struct list_head node;
+};
+static LIST_HEAD(cooling_cpufreq_list);
+static DEFINE_IDR(cpufreq_idr);
+
+static struct mutex cooling_cpufreq_lock;
+
+/* notify_table passes value to the CPUFREQ_ADJUST callback function. */
+#define NOTIFY_INVALID NULL
+struct cpufreq_cooling_device *notify_device;
+
+/* Head of the blocking notifier chain to inform about frequency clamping */
+static BLOCKING_NOTIFIER_HEAD(cputherm_state_notifier_list);
+
+/**
+ * get_idr - function to get a unique id.
+ * @idr: struct idr * handle used to create a id.
+ * @id: int * value generated by this function.
+ */
+static int get_idr(struct idr *idr, int *id)
+{
+	int err;
+again:
+	if (unlikely(idr_pre_get(idr, GFP_KERNEL) == 0))
+		return -ENOMEM;
+
+	mutex_lock(&cooling_cpufreq_lock);
+	err = idr_get_new(idr, NULL, id);
+	mutex_unlock(&cooling_cpufreq_lock);
+
+	if (unlikely(err == -EAGAIN))
+		goto again;
+	else if (unlikely(err))
+		return err;
+
+	*id = *id & MAX_ID_MASK;
+	return 0;
+}
+
+/**
+ * release_idr - function to free the unique id.
+ * @idr: struct idr * handle used for creating the id.
+ * @id: int value representing the unique id.
+ */
+static void release_idr(struct idr *idr, int id)
+{
+	mutex_lock(&cooling_cpufreq_lock);
+	idr_remove(idr, id);
+	mutex_unlock(&cooling_cpufreq_lock);
+}
+
+/**
+ * cputherm_register_notifier - Register a notifier with cpu cooling interface.
+ * @nb:	struct notifier_block * with callback info.
+ * @list: integer value for which notification is needed. possible values are
+ *	CPUFREQ_COOLING_START and CPUFREQ_COOLING_STOP.
+ *
+ * This exported function registers a driver with cpu cooling layer. The driver
+ * will be notified when any cpu cooling action is called.
+ */
+int cputherm_register_notifier(struct notifier_block *nb, unsigned int list)
+{
+	int ret = 0;
+
+	switch (list) {
+	case CPUFREQ_COOLING_START:
+	case CPUFREQ_COOLING_STOP:
+		ret = blocking_notifier_chain_register(
+				&cputherm_state_notifier_list, nb);
+		break;
+	default:
+		ret = -EINVAL;
+	}
+	return ret;
+}
+EXPORT_SYMBOL(cputherm_register_notifier);
+
+/**
+ * cputherm_unregister_notifier - Un-register a notifier.
+ * @nb:	struct notifier_block * with callback info.
+ * @list: integer value for which notification is needed. values possible are
+ *	CPUFREQ_COOLING_START or CPUFREQ_COOLING_STOP.
+ *
+ * This exported function un-registers a driver with cpu cooling layer.
+ */
+int cputherm_unregister_notifier(struct notifier_block *nb, unsigned int list)
+{
+	int ret = 0;
+
+	switch (list) {
+	case CPUFREQ_COOLING_START:
+	case CPUFREQ_COOLING_STOP:
+		ret = blocking_notifier_chain_unregister(
+				&cputherm_state_notifier_list, nb);
+		break;
+	default:
+		ret = -EINVAL;
+	}
+	return ret;
+}
+EXPORT_SYMBOL(cputherm_unregister_notifier);
+
+/* Below code defines functions to be used for cpufreq as cooling device */
+
+/**
+ * is_cpufreq_valid - function to check if a cpu has frequency transition policy.
+ * @cpu: cpu for which check is needed.
+ */
+static int is_cpufreq_valid(int cpu)
+{
+	struct cpufreq_policy policy;
+	return !cpufreq_get_policy(&policy, cpu);
+}
+
+/**
+ * get_cpu_frequency - get the absolute value of frequency from level.
+ * @cpu: cpu for which frequency is fetched.
+ * @level: level of frequency of the CPU
+ *	e.g level=1 --> 1st MAX FREQ, LEVEL=2 ---> 2nd MAX FREQ, .... etc
+ */
+static unsigned int get_cpu_frequency(unsigned int cpu, unsigned long level)
+{
+	int ret = 0, i = 0;
+	unsigned long level_index;
+	bool descend = false;
+	struct cpufreq_frequency_table *table =
+					cpufreq_frequency_get_table(cpu);
+	if (!table)
+		return ret;
+
+	while (table[i].frequency != CPUFREQ_TABLE_END) {
+		if (table[i].frequency == CPUFREQ_ENTRY_INVALID)
+			continue;
+
+		/*check if table in ascending or descending order*/
+		if ((table[i + 1].frequency != CPUFREQ_TABLE_END) &&
+			(table[i + 1].frequency < table[i].frequency)
+			&& !descend) {
+			descend = true;
+		}
+
+		/*return if level matched and table in descending order*/
+		if (descend && i == level)
+			return table[i].frequency;
+		i++;
+	}
+	i--;
+
+	if (level > i || descend)
+		return ret;
+	level_index = i - level;
+
+	/*Scan the table in reverse order and match the level*/
+	while (i >= 0) {
+		if (table[i].frequency == CPUFREQ_ENTRY_INVALID)
+			continue;
+		/*return if level matched*/
+		if (i == level_index)
+			return table[i].frequency;
+		i--;
+	}
+	return ret;
+}
+
+/**
+ * cpufreq_apply_cooling - function to apply frequency clipping.
+ * @cpufreq_device: cpufreq_cooling_device pointer containing frequency
+ *	clipping data.
+ * @cooling_state: value of the cooling state.
+ */
+static int cpufreq_apply_cooling(struct cpufreq_cooling_device *cpufreq_device,
+				unsigned long cooling_state)
+{
+	unsigned int event, cpuid, clip_freq;
+	struct cpumask *maskPtr = &cpufreq_device->allowed_cpus;
+	unsigned int cpu = cpumask_any(maskPtr);
+
+
+	/* Check if the old cooling action is same as new cooling action */
+	if (cpufreq_device->cpufreq_state == cooling_state)
+		return 0;
+
+	clip_freq = get_cpu_frequency(cpu, cooling_state);
+	if (!clip_freq)
+		return -EINVAL;
+
+	cpufreq_device->cpufreq_state = cooling_state;
+	cpufreq_device->cpufreq_val = clip_freq;
+	notify_device = cpufreq_device;
+
+	if (cooling_state != 0)
+		event = CPUFREQ_COOLING_START;
+	else
+		event = CPUFREQ_COOLING_STOP;
+
+	blocking_notifier_call_chain(&cputherm_state_notifier_list,
+						event, &clip_freq);
+
+	for_each_cpu(cpuid, maskPtr) {
+		if (is_cpufreq_valid(cpuid))
+			cpufreq_update_policy(cpuid);
+	}
+
+	notify_device = NOTIFY_INVALID;
+
+	return 0;
+}
+
+/**
+ * cpufreq_thermal_notifier - notifier callback for cpufreq policy change.
+ * @nb:	struct notifier_block * with callback info.
+ * @event: value showing cpufreq event for which this function invoked.
+ * @data: callback-specific data
+ */
+static int cpufreq_thermal_notifier(struct notifier_block *nb,
+					unsigned long event, void *data)
+{
+	struct cpufreq_policy *policy = data;
+	unsigned long max_freq = 0;
+
+	if (event != CPUFREQ_ADJUST || notify_device == NOTIFY_INVALID)
+		return 0;
+
+	if (cpumask_test_cpu(policy->cpu, &notify_device->allowed_cpus))
+		max_freq = notify_device->cpufreq_val;
+
+	/* Never exceed user_policy.max*/
+	if (max_freq > policy->user_policy.max)
+		max_freq = policy->user_policy.max;
+
+	if (policy->max != max_freq)
+		cpufreq_verify_within_limits(policy, 0, max_freq);
+
+	return 0;
+}
+
+/*
+ * cpufreq cooling device callback functions are defined below
+ */
+
+/**
+ * cpufreq_get_max_state - callback function to get the max cooling state.
+ * @cdev: thermal cooling device pointer.
+ * @state: fill this variable with the max cooling state.
+ */
+static int cpufreq_get_max_state(struct thermal_cooling_device *cdev,
+				 unsigned long *state)
+{
+	int ret = -EINVAL, i = 0;
+	struct cpufreq_cooling_device *cpufreq_device;
+	struct cpumask *maskPtr;
+	unsigned int cpu;
+	struct cpufreq_frequency_table *table;
+
+	mutex_lock(&cooling_cpufreq_lock);
+	list_for_each_entry(cpufreq_device, &cooling_cpufreq_list, node) {
+		if (cpufreq_device && cpufreq_device->cool_dev == cdev)
+			break;
+	}
+	if (cpufreq_device == NULL)
+		goto return_get_max_state;
+
+	maskPtr = &cpufreq_device->allowed_cpus;
+	cpu = cpumask_any(maskPtr);
+	table = cpufreq_frequency_get_table(cpu);
+	if (!table) {
+		*state = 0;
+		ret = 0;
+		goto return_get_max_state;
+	}
+
+	while (table[i].frequency != CPUFREQ_TABLE_END) {
+		if (table[i].frequency == CPUFREQ_ENTRY_INVALID)
+			continue;
+		i++;
+	}
+	if (i > 0) {
+		*state = --i;
+		ret = 0;
+	}
+
+return_get_max_state:
+	mutex_unlock(&cooling_cpufreq_lock);
+	return ret;
+}
+
+/**
+ * cpufreq_get_cur_state - callback function to get the current cooling state.
+ * @cdev: thermal cooling device pointer.
+ * @state: fill this variable with the current cooling state.
+ */
+static int cpufreq_get_cur_state(struct thermal_cooling_device *cdev,
+				 unsigned long *state)
+{
+	int ret = -EINVAL;
+	struct cpufreq_cooling_device *cpufreq_device;
+
+	mutex_lock(&cooling_cpufreq_lock);
+	list_for_each_entry(cpufreq_device, &cooling_cpufreq_list, node) {
+		if (cpufreq_device && cpufreq_device->cool_dev == cdev) {
+			*state = cpufreq_device->cpufreq_state;
+			ret = 0;
+			break;
+		}
+	}
+	mutex_unlock(&cooling_cpufreq_lock);
+
+	return ret;
+}
+
+/**
+ * cpufreq_set_cur_state - callback function to set the current cooling state.
+ * @cdev: thermal cooling device pointer.
+ * @state: set this variable to the current cooling state.
+ */
+static int cpufreq_set_cur_state(struct thermal_cooling_device *cdev,
+				 unsigned long state)
+{
+	int ret = -EINVAL;
+	struct cpufreq_cooling_device *cpufreq_device;
+
+	mutex_lock(&cooling_cpufreq_lock);
+	list_for_each_entry(cpufreq_device, &cooling_cpufreq_list, node) {
+		if (cpufreq_device && cpufreq_device->cool_dev == cdev) {
+			ret = 0;
+			break;
+		}
+	}
+	if (!ret)
+		ret = cpufreq_apply_cooling(cpufreq_device, state);
+
+	mutex_unlock(&cooling_cpufreq_lock);
+
+	return ret;
+}
+
+/* Bind cpufreq callbacks to thermal cooling device ops */
+static struct thermal_cooling_device_ops const cpufreq_cooling_ops = {
+	.get_max_state = cpufreq_get_max_state,
+	.get_cur_state = cpufreq_get_cur_state,
+	.set_cur_state = cpufreq_set_cur_state,
+};
+
+/* Notifier for cpufreq policy change */
+static struct notifier_block thermal_cpufreq_notifier_block = {
+	.notifier_call = cpufreq_thermal_notifier,
+};
+
+/**
+ * cpufreq_cooling_register - function to create cpufreq cooling device.
+ * @clip_cpus: cpumask of cpus where the frequency constraints will happen.
+ */
+struct thermal_cooling_device *cpufreq_cooling_register(
+	struct cpumask *clip_cpus)
+{
+	struct thermal_cooling_device *cool_dev;
+	struct cpufreq_cooling_device *cpufreq_dev = NULL;
+	unsigned int cpufreq_dev_count = 0, min = 0, max = 0;
+	char dev_name[THERMAL_NAME_LENGTH];
+	int ret = 0, id = 0, i;
+	struct cpufreq_policy policy;
+
+	list_for_each_entry(cpufreq_dev, &cooling_cpufreq_list, node)
+		cpufreq_dev_count++;
+
+	/*Verify that all the clip cpus have same freq_min, freq_max limit*/
+	for_each_cpu(i, clip_cpus) {
+		/*continue if cpufreq policy not found and not return error*/
+		if (!cpufreq_get_policy(&policy, i))
+			continue;
+		if (min == 0 && max == 0) {
+			min = policy.cpuinfo.min_freq;
+			max = policy.cpuinfo.max_freq;
+		} else {
+			if (min != policy.cpuinfo.min_freq ||
+				max != policy.cpuinfo.max_freq)
+				return ERR_PTR(-EINVAL);
+}
+	}
+	cpufreq_dev = kzalloc(sizeof(struct cpufreq_cooling_device),
+			GFP_KERNEL);
+	if (!cpufreq_dev)
+		return ERR_PTR(-ENOMEM);
+
+	cpumask_copy(&cpufreq_dev->allowed_cpus, clip_cpus);
+
+	if (cpufreq_dev_count == 0)
+		mutex_init(&cooling_cpufreq_lock);
+
+	ret = get_idr(&cpufreq_idr, &cpufreq_dev->id);
+	if (ret) {
+		kfree(cpufreq_dev);
+		return ERR_PTR(-EINVAL);
+	}
+
+	sprintf(dev_name, "thermal-cpufreq-%d", cpufreq_dev->id);
+
+	cool_dev = thermal_cooling_device_register(dev_name, cpufreq_dev,
+						&cpufreq_cooling_ops);
+	if (!cool_dev) {
+		release_idr(&cpufreq_idr, cpufreq_dev->id);
+		kfree(cpufreq_dev);
+		return ERR_PTR(-EINVAL);
+	}
+	cpufreq_dev->id = id;
+	cpufreq_dev->cool_dev = cool_dev;
+	cpufreq_dev->cpufreq_state = 0;
+	mutex_lock(&cooling_cpufreq_lock);
+	list_add_tail(&cpufreq_dev->node, &cooling_cpufreq_list);
+
+	/* Register the notifier for first cpufreq cooling device */
+	if (cpufreq_dev_count == 0)
+		cpufreq_register_notifier(&thermal_cpufreq_notifier_block,
+						CPUFREQ_POLICY_NOTIFIER);
+
+	mutex_unlock(&cooling_cpufreq_lock);
+	return cool_dev;
+}
+EXPORT_SYMBOL(cpufreq_cooling_register);
+
+/**
+ * cpufreq_cooling_unregister - function to remove cpufreq cooling device.
+ * @cdev: thermal cooling device pointer.
+ */
+void cpufreq_cooling_unregister(struct thermal_cooling_device *cdev)
+{
+	struct cpufreq_cooling_device *cpufreq_dev = NULL;
+	unsigned int cpufreq_dev_count = 0;
+
+	mutex_lock(&cooling_cpufreq_lock);
+	list_for_each_entry(cpufreq_dev, &cooling_cpufreq_list, node) {
+		if (cpufreq_dev && cpufreq_dev->cool_dev == cdev)
+			break;
+		cpufreq_dev_count++;
+	}
+
+	if (!cpufreq_dev || cpufreq_dev->cool_dev != cdev) {
+		mutex_unlock(&cooling_cpufreq_lock);
+		return;
+	}
+
+	list_del(&cpufreq_dev->node);
+
+	/* Unregister the notifier for the last cpufreq cooling device */
+	if (cpufreq_dev_count == 1) {
+		cpufreq_unregister_notifier(&thermal_cpufreq_notifier_block,
+					CPUFREQ_POLICY_NOTIFIER);
+	}
+	mutex_unlock(&cooling_cpufreq_lock);
+	thermal_cooling_device_unregister(cpufreq_dev->cool_dev);
+	release_idr(&cpufreq_idr, cpufreq_dev->id);
+	if (cpufreq_dev_count == 1)
+		mutex_destroy(&cooling_cpufreq_lock);
+	kfree(cpufreq_dev);
+}
+EXPORT_SYMBOL(cpufreq_cooling_unregister);
diff --git a/include/linux/cpu_cooling.h b/include/linux/cpu_cooling.h
new file mode 100644
index 0000000..555b917
--- /dev/null
+++ b/include/linux/cpu_cooling.h
@@ -0,0 +1,79 @@ 
+/*
+ *  linux/include/linux/cpu_cooling.h
+ *
+ *  Copyright (C) 2012	Samsung Electronics Co., Ltd(http://www.samsung.com)
+ *  Copyright (C) 2012  Amit Daniel <[e-mail]>
+ *
+ * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License as published by
+ *  the Free Software Foundation; version 2 of the License.
+ *
+ *  This program is distributed in the hope that it will be useful, but
+ *  WITHOUT ANY WARRANTY; without even the implied warranty of
+ *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ *  General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License along
+ *  with this program; if not, write to the Free Software Foundation, Inc.,
+ *  59 Temple Place, Suite 330, Boston, MA 02111-1307 USA.
+ *
+ * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+ */
+
+#ifndef __CPU_COOLING_H__
+#define __CPU_COOLING_H__
+
+#include <linux/thermal.h>
+
+#define CPUFREQ_COOLING_START		0
+#define CPUFREQ_COOLING_STOP		1
+
+/**
+ * cputherm_register_notifier - Register a notifier with cpu cooling interface.
+ * @nb:	struct notifier_block * with callback info.
+ * @list: integer value for which notification is needed. possible values are
+ *	CPUFREQ_COOLING_START and CPUFREQ_COOLING_STOP.
+ *
+ * This exported function registers a driver with cpu cooling layer. The driver
+ * will be notified when any cpu cooling action is called.
+ */
+int cputherm_register_notifier(struct notifier_block *nb, unsigned int list);
+
+/**
+ * cputherm_unregister_notifier - Un-register a notifier.
+ * @nb:	struct notifier_block * with callback info.
+ * @list: integer value for which notification is needed. values possible are
+ *	CPUFREQ_COOLING_START and CPUFREQ_COOLING_STOP.
+ *
+ * This exported function un-registers a driver with cpu cooling layer.
+ */
+int cputherm_unregister_notifier(struct notifier_block *nb, unsigned int list);
+
+#ifdef CONFIG_CPU_THERMAL
+/**
+ * cpufreq_cooling_register - function to create cpufreq cooling device.
+ * @clip_cpus: cpumask of cpus where the frequency constraints will happen
+ */
+struct thermal_cooling_device *cpufreq_cooling_register(
+		struct cpumask *clip_cpus);
+
+/**
+ * cpufreq_cooling_unregister - function to remove cpufreq cooling device.
+ * @cdev: thermal cooling device pointer.
+ */
+void cpufreq_cooling_unregister(struct thermal_cooling_device *cdev);
+#else /* !CONFIG_CPU_THERMAL */
+static inline struct thermal_cooling_device *cpufreq_cooling_register(
+	struct cpumask *clip_cpus)
+{
+	return NULL;
+}
+static inline void cpufreq_cooling_unregister(
+		struct thermal_cooling_device *cdev)
+{
+	return;
+}
+#endif	/* CONFIG_CPU_THERMAL */
+
+#endif /* __CPU_COOLING_H__ */