1 ===================================
2 Generic Thermal Sysfs driver How To
3 ===================================
5 Written by Sujith Thomas <sujith.thomas@intel.com>, Zhang Rui <rui.zhang@intel.com>
7 Updated: 2 January 2008
9 Copyright (c) 2008 Intel Corporation
15 The generic thermal sysfs provides a set of interfaces for thermal zone
16 devices (sensors) and thermal cooling devices (fan, processor...) to register
17 with the thermal management solution and to be a part of it.
19 This how-to focuses on enabling new thermal zone and cooling devices to
20 participate in thermal management.
21 This solution is platform independent and any type of thermal zone devices
22 and cooling devices should be able to make use of the infrastructure.
24 The main task of the thermal sysfs driver is to expose thermal zone attributes
25 as well as cooling device attributes to the user space.
26 An intelligent thermal management application can make decisions based on
27 inputs from thermal zone attributes (the current temperature and trip point
28 temperature) and throttle appropriate devices.
30 - `[0-*]` denotes any positive number starting from 0
31 - `[1-*]` denotes any positive number starting from 1
33 1. thermal sysfs driver interface functions
34 ===========================================
36 1.1 thermal zone device interface
37 ---------------------------------
41 struct thermal_zone_device
42 *thermal_zone_device_register(char *type,
43 int trips, int mask, void *devdata,
44 struct thermal_zone_device_ops *ops,
45 const struct thermal_zone_params *tzp,
46 int passive_delay, int polling_delay))
48 This interface function adds a new thermal zone device (sensor) to
49 /sys/class/thermal folder as `thermal_zone[0-*]`. It tries to bind all the
50 thermal cooling devices registered at the same time.
53 the thermal zone type.
55 the total number of trip points this thermal zone supports.
57 Bit string: If 'n'th bit is set, then trip point 'n' is writeable.
61 thermal zone device call-backs.
64 bind the thermal zone device with a thermal cooling device.
66 unbind the thermal zone device with a thermal cooling device.
68 get the current temperature of the thermal zone.
70 set the trip points window. Whenever the current temperature
71 is updated, the trip points immediately below and above the
72 current temperature are found.
74 get the current mode (enabled/disabled) of the thermal zone.
76 - "enabled" means the kernel thermal management is
78 - "disabled" will prevent kernel thermal driver action
79 upon trip points so that user applications can take
80 charge of thermal management.
82 set the mode (enabled/disabled) of the thermal zone.
84 get the type of certain trip point.
86 get the temperature above which the certain trip point
89 set the emulation temperature which helps in debugging
90 different threshold temperature points.
92 thermal zone platform parameters.
94 number of milliseconds to wait between polls when
95 performing passive cooling.
97 number of milliseconds to wait between polls when checking
98 whether trip points have been crossed (0 for interrupt driven systems).
102 void thermal_zone_device_unregister(struct thermal_zone_device *tz)
104 This interface function removes the thermal zone device.
105 It deletes the corresponding entry from /sys/class/thermal folder and
106 unbinds all the thermal cooling devices it uses.
110 struct thermal_zone_device
111 *thermal_zone_of_sensor_register(struct device *dev, int sensor_id,
113 const struct thermal_zone_of_device_ops *ops)
115 This interface adds a new sensor to a DT thermal zone.
116 This function will search the list of thermal zones described in
117 device tree and look for the zone that refer to the sensor device
118 pointed by dev->of_node as temperature providers. For the zone
119 pointing to the sensor node, the sensor will be added to the DT
122 The parameters for this interface are:
125 Device node of sensor containing valid node pointer in
128 a sensor identifier, in case the sensor IP has more
131 a private pointer (owned by the caller) that will be
132 passed back, when a temperature reading is needed.
134 `struct thermal_zone_of_device_ops *`.
136 ============== =======================================
137 get_temp a pointer to a function that reads the
138 sensor temperature. This is mandatory
139 callback provided by sensor driver.
140 set_trips a pointer to a function that sets a
141 temperature window. When this window is
142 left the driver must inform the thermal
143 core via thermal_zone_device_update.
144 get_trend a pointer to a function that reads the
145 sensor temperature trend.
146 set_emul_temp a pointer to a function that sets
147 sensor emulated temperature.
148 ============== =======================================
150 The thermal zone temperature is provided by the get_temp() function
151 pointer of thermal_zone_of_device_ops. When called, it will
152 have the private pointer @data back.
154 It returns error pointer if fails otherwise valid thermal zone device
155 handle. Caller should check the return handle with IS_ERR() for finding
156 whether success or not.
160 void thermal_zone_of_sensor_unregister(struct device *dev,
161 struct thermal_zone_device *tzd)
163 This interface unregisters a sensor from a DT thermal zone which was
164 successfully added by interface thermal_zone_of_sensor_register().
165 This function removes the sensor callbacks and private data from the
166 thermal zone device registered with thermal_zone_of_sensor_register()
167 interface. It will also silent the zone by remove the .get_temp() and
168 get_trend() thermal zone device callbacks.
172 struct thermal_zone_device
173 *devm_thermal_zone_of_sensor_register(struct device *dev,
176 const struct thermal_zone_of_device_ops *ops)
178 This interface is resource managed version of
179 thermal_zone_of_sensor_register().
181 All details of thermal_zone_of_sensor_register() described in
182 section 1.1.3 is applicable here.
184 The benefit of using this interface to register sensor is that it
185 is not require to explicitly call thermal_zone_of_sensor_unregister()
186 in error path or during driver unbinding as this is done by driver
191 void devm_thermal_zone_of_sensor_unregister(struct device *dev,
192 struct thermal_zone_device *tzd)
194 This interface is resource managed version of
195 thermal_zone_of_sensor_unregister().
196 All details of thermal_zone_of_sensor_unregister() described in
197 section 1.1.4 is applicable here.
198 Normally this function will not need to be called and the resource
199 management code will ensure that the resource is freed.
203 int thermal_zone_get_slope(struct thermal_zone_device *tz)
205 This interface is used to read the slope attribute value
206 for the thermal zone device, which might be useful for platform
207 drivers for temperature calculations.
211 int thermal_zone_get_offset(struct thermal_zone_device *tz)
213 This interface is used to read the offset attribute value
214 for the thermal zone device, which might be useful for platform
215 drivers for temperature calculations.
217 1.2 thermal cooling device interface
218 ------------------------------------
223 struct thermal_cooling_device
224 *thermal_cooling_device_register(char *name,
225 void *devdata, struct thermal_cooling_device_ops *)
227 This interface function adds a new thermal cooling device (fan/processor/...)
228 to /sys/class/thermal/ folder as `cooling_device[0-*]`. It tries to bind itself
229 to all the thermal zone devices registered at the same time.
232 the cooling device name.
236 thermal cooling devices call-backs.
239 get the Maximum throttle state of the cooling device.
241 get the Currently requested throttle state of the
244 set the Current throttle state of the cooling device.
248 void thermal_cooling_device_unregister(struct thermal_cooling_device *cdev)
250 This interface function removes the thermal cooling device.
251 It deletes the corresponding entry from /sys/class/thermal folder and
252 unbinds itself from all the thermal zone devices using it.
254 1.3 interface for binding a thermal zone device with a thermal cooling device
255 -----------------------------------------------------------------------------
259 int thermal_zone_bind_cooling_device(struct thermal_zone_device *tz,
260 int trip, struct thermal_cooling_device *cdev,
261 unsigned long upper, unsigned long lower, unsigned int weight);
263 This interface function binds a thermal cooling device to a particular trip
264 point of a thermal zone device.
266 This function is usually called in the thermal zone device .bind callback.
269 the thermal zone device
271 thermal cooling device
273 indicates which trip point in this thermal zone the cooling device
276 the Maximum cooling state for this trip point.
277 THERMAL_NO_LIMIT means no upper limit,
278 and the cooling device can be in max_state.
280 the Minimum cooling state can be used for this trip point.
281 THERMAL_NO_LIMIT means no lower limit,
282 and the cooling device can be in cooling state 0.
284 the influence of this cooling device in this thermal
285 zone. See 1.4.1 below for more information.
289 int thermal_zone_unbind_cooling_device(struct thermal_zone_device *tz,
290 int trip, struct thermal_cooling_device *cdev);
292 This interface function unbinds a thermal cooling device from a particular
293 trip point of a thermal zone device. This function is usually called in
294 the thermal zone device .unbind callback.
297 the thermal zone device
299 thermal cooling device
301 indicates which trip point in this thermal zone the cooling device
304 1.4 Thermal Zone Parameters
305 ---------------------------
309 struct thermal_bind_params
311 This structure defines the following parameters that are used to bind
312 a zone with a cooling device for a particular trip point.
315 The cooling device pointer
317 The 'influence' of a particular cooling device on this
318 zone. This is relative to the rest of the cooling
319 devices. For example, if all cooling devices have a
320 weight of 1, then they all contribute the same. You can
321 use percentages if you want, but it's not mandatory. A
322 weight of 0 means that this cooling device doesn't
323 contribute to the cooling of this zone unless all cooling
324 devices have a weight of 0. If all weights are 0, then
325 they all contribute the same.
327 This is a bit mask that gives the binding relation between
328 this thermal zone and cdev, for a particular trip point.
329 If nth bit is set, then the cdev and thermal zone are bound
332 This is an array of cooling state limits. Must have
333 exactly 2 * thermal_zone.number_of_trip_points. It is an
334 array consisting of tuples <lower-state upper-state> of
335 state limits. Each trip will be associated with one state
336 limit tuple when binding. A NULL pointer means
337 <THERMAL_NO_LIMITS THERMAL_NO_LIMITS> on all trips.
338 These limits are used when binding a cdev to a trip point.
340 This call back returns success(0) if the 'tz and cdev' need to
341 be bound, as per platform data.
345 struct thermal_zone_params
347 This structure defines the platform level parameters for a thermal zone.
348 This data, for each thermal zone should come from the platform layer.
349 This is an optional feature where some platforms can choose not to
353 Name of the thermal governor used for this zone
355 a boolean to indicate if the thermal to hwmon sysfs interface
356 is required. when no_hwmon == false, a hwmon sysfs interface
357 will be created. when no_hwmon == true, nothing will be done.
358 In case the thermal_zone_params is NULL, the hwmon interface
359 will be created (for backward compatibility).
361 Number of thermal_bind_params entries for this zone
363 thermal_bind_params entries
365 2. sysfs attributes structure
366 =============================
374 Thermal sysfs attributes will be represented under /sys/class/thermal.
375 Hwmon sysfs I/F extension is also available under /sys/class/hwmon
376 if hwmon is compiled in or built as a module.
378 Thermal zone device sys I/F, created once it's registered::
380 /sys/class/thermal/thermal_zone[0-*]:
381 |---type: Type of the thermal zone
382 |---temp: Current temperature
383 |---mode: Working mode of the thermal zone
384 |---policy: Thermal governor used for this zone
385 |---available_policies: Available thermal governors for this zone
386 |---trip_point_[0-*]_temp: Trip point temperature
387 |---trip_point_[0-*]_type: Trip point type
388 |---trip_point_[0-*]_hyst: Hysteresis value for this trip point
389 |---emul_temp: Emulated temperature set node
390 |---sustainable_power: Sustainable dissipatable power
391 |---k_po: Proportional term during temperature overshoot
392 |---k_pu: Proportional term during temperature undershoot
393 |---k_i: PID's integral term in the power allocator gov
394 |---k_d: PID's derivative term in the power allocator
395 |---integral_cutoff: Offset above which errors are accumulated
396 |---slope: Slope constant applied as linear extrapolation
397 |---offset: Offset constant applied as linear extrapolation
399 Thermal cooling device sys I/F, created once it's registered::
401 /sys/class/thermal/cooling_device[0-*]:
402 |---type: Type of the cooling device(processor/fan/...)
403 |---max_state: Maximum cooling state of the cooling device
404 |---cur_state: Current cooling state of the cooling device
405 |---stats: Directory containing cooling device's statistics
406 |---stats/reset: Writing any value resets the statistics
407 |---stats/time_in_state_ms: Time (msec) spent in various cooling states
408 |---stats/total_trans: Total number of times cooling state is changed
409 |---stats/trans_table: Cooing state transition table
412 Then next two dynamic attributes are created/removed in pairs. They represent
413 the relationship between a thermal zone and its associated cooling device.
414 They are created/removed for each successful execution of
415 thermal_zone_bind_cooling_device/thermal_zone_unbind_cooling_device.
419 /sys/class/thermal/thermal_zone[0-*]:
420 |---cdev[0-*]: [0-*]th cooling device in current thermal zone
421 |---cdev[0-*]_trip_point: Trip point that cdev[0-*] is associated with
422 |---cdev[0-*]_weight: Influence of the cooling device in
425 Besides the thermal zone device sysfs I/F and cooling device sysfs I/F,
426 the generic thermal driver also creates a hwmon sysfs I/F for each _type_
427 of thermal zone device. E.g. the generic thermal driver registers one hwmon
428 class device and build the associated hwmon sysfs I/F for all the registered
433 /sys/class/hwmon/hwmon[0-*]:
434 |---name: The type of the thermal zone devices
435 |---temp[1-*]_input: The current temperature of thermal zone [1-*]
436 |---temp[1-*]_critical: The critical trip point of thermal zone [1-*]
438 Please read Documentation/hwmon/sysfs-interface.rst for additional information.
440 Thermal zone attributes
441 -----------------------
444 Strings which represent the thermal zone type.
445 This is given by thermal zone driver as part of registration.
446 E.g: "acpitz" indicates it's an ACPI thermal device.
447 In order to keep it consistent with hwmon sys attribute; this should
448 be a short, lowercase string, not containing spaces nor dashes.
452 Current temperature as reported by thermal zone (sensor).
453 Unit: millidegree Celsius
457 One of the predefined values in [enabled, disabled].
458 This file gives information about the algorithm that is currently
459 managing the thermal zone. It can be either default kernel based
460 algorithm or user space application.
463 enable Kernel Thermal management.
465 Preventing kernel thermal zone driver actions upon
466 trip points so that user application can take full
467 charge of the thermal management.
472 One of the various thermal governors used for a particular zone.
477 Available thermal governors which can be used for a particular zone.
481 `trip_point_[0-*]_temp`
482 The temperature above which trip point will be fired.
484 Unit: millidegree Celsius
488 `trip_point_[0-*]_type`
489 Strings which indicate the type of the trip point.
491 E.g. it can be one of critical, hot, passive, `active[0-*]` for ACPI
496 `trip_point_[0-*]_hyst`
497 The hysteresis value for a trip point, represented as an integer
502 Sysfs link to the thermal cooling device node where the sys I/F
503 for cooling device throttling control represents.
507 `cdev[0-*]_trip_point`
508 The trip point in this thermal zone which `cdev[0-*]` is associated
509 with; -1 means the cooling device is not associated with any trip
515 The influence of `cdev[0-*]` in this thermal zone. This value
516 is relative to the rest of cooling devices in the thermal
517 zone. For example, if a cooling device has a weight double
518 than that of other, it's twice as effective in cooling the
524 Attribute is only present for zones in which the passive cooling
525 policy is not supported by native thermal driver. Default is zero
526 and can be set to a temperature (in millidegrees) to enable a
527 passive trip point for the zone. Activation is done by polling with
528 an interval of 1 second.
530 Unit: millidegrees Celsius
532 Valid values: 0 (disabled) or greater than 1000
537 Interface to set the emulated temperature method in thermal zone
538 (sensor). After setting this temperature, the thermal zone may pass
539 this temperature to platform emulation function if registered or
540 cache it locally. This is useful in debugging different temperature
541 threshold and its associated cooling action. This is write only node
542 and writing 0 on this node should disable emulation.
543 Unit: millidegree Celsius
548 Be careful while enabling this option on production systems,
549 because userland can easily disable the thermal policy by simply
550 flooding this sysfs node with low temperature values.
553 An estimate of the sustained power that can be dissipated by
554 the thermal zone. Used by the power allocator governor. For
555 more information see Documentation/driver-api/thermal/power_allocator.rst
562 The proportional term of the power allocator governor's PID
563 controller during temperature overshoot. Temperature overshoot
564 is when the current temperature is above the "desired
565 temperature" trip point. For more information see
566 Documentation/driver-api/thermal/power_allocator.rst
571 The proportional term of the power allocator governor's PID
572 controller during temperature undershoot. Temperature undershoot
573 is when the current temperature is below the "desired
574 temperature" trip point. For more information see
575 Documentation/driver-api/thermal/power_allocator.rst
580 The integral term of the power allocator governor's PID
581 controller. This term allows the PID controller to compensate
582 for long term drift. For more information see
583 Documentation/driver-api/thermal/power_allocator.rst
588 The derivative term of the power allocator governor's PID
589 controller. For more information see
590 Documentation/driver-api/thermal/power_allocator.rst
595 Temperature offset from the desired temperature trip point
596 above which the integral term of the power allocator
597 governor's PID controller starts accumulating errors. For
598 example, if integral_cutoff is 0, then the integral term only
599 accumulates error when temperature is above the desired
600 temperature trip point. For more information see
601 Documentation/driver-api/thermal/power_allocator.rst
603 Unit: millidegree Celsius
608 The slope constant used in a linear extrapolation model
609 to determine a hotspot temperature based off the sensor's
610 raw readings. It is up to the device driver to determine
611 the usage of these values.
616 The offset constant used in a linear extrapolation model
617 to determine a hotspot temperature based off the sensor's
618 raw readings. It is up to the device driver to determine
619 the usage of these values.
623 Cooling device attributes
624 -------------------------
627 String which represents the type of device, e.g:
629 - for generic ACPI: should be "Fan", "Processor" or "LCD"
630 - for memory controller device on intel_menlow platform:
631 should be "Memory controller".
636 The maximum permissible cooling state of this cooling device.
641 The current cooling state of this cooling device.
642 The value can any integer numbers between 0 and max_state:
644 - cur_state == 0 means no cooling
645 - cur_state == max_state means the maximum cooling.
650 Writing any value resets the cooling device's statistics.
653 stats/time_in_state_ms:
654 The amount of time spent by the cooling device in various cooling
655 states. The output will have "<state> <time>" pair in each line, which
656 will mean this cooling device spent <time> msec of time at <state>.
657 Output will have one line for each of the supported states.
662 A single positive value showing the total number of times the state of a
663 cooling device is changed.
668 This gives fine grained information about all the cooling state
669 transitions. The cat output here is a two dimensional matrix, where an
670 entry <i,j> (row i, column j) represents the number of transitions from
671 State_i to State_j. If the transition table is bigger than PAGE_SIZE,
672 reading this will return an -EFBIG error.
675 3. A simple implementation
676 ==========================
678 ACPI thermal zone may support multiple trip points like critical, hot,
679 passive, active. If an ACPI thermal zone supports critical, passive,
680 active[0] and active[1] at the same time, it may register itself as a
681 thermal_zone_device (thermal_zone1) with 4 trip points in all.
682 It has one processor and one fan, which are both registered as
683 thermal_cooling_device. Both are considered to have the same
684 effectiveness in cooling the thermal zone.
686 If the processor is listed in _PSL method, and the fan is listed in _AL0
687 method, the sys I/F structure will be built like this::
694 |---policy: step_wise
695 |---available_policies: step_wise fair_share
696 |---trip_point_0_temp: 100000
697 |---trip_point_0_type: critical
698 |---trip_point_1_temp: 80000
699 |---trip_point_1_type: passive
700 |---trip_point_2_temp: 70000
701 |---trip_point_2_type: active0
702 |---trip_point_3_temp: 60000
703 |---trip_point_3_type: active1
704 |---cdev0: --->/sys/class/thermal/cooling_device0
705 |---cdev0_trip_point: 1 /* cdev0 can be used for passive */
706 |---cdev0_weight: 1024
707 |---cdev1: --->/sys/class/thermal/cooling_device3
708 |---cdev1_trip_point: 2 /* cdev1 can be used for active[0]*/
709 |---cdev1_weight: 1024
724 |---temp1_input: 37000
725 |---temp1_crit: 100000
727 4. Export Symbol APIs
728 =====================
733 This function returns the trend of a thermal zone, i.e the rate of change
734 of temperature of the thermal zone. Ideally, the thermal sensor drivers
735 are supposed to implement the callback. If they don't, the thermal
736 framework calculated the trend by comparing the previous and the current
739 4.2. get_thermal_instance
740 -------------------------
742 This function returns the thermal_instance corresponding to a given
743 {thermal_zone, cooling_device, trip_point} combination. Returns NULL
744 if such an instance does not exist.
746 4.3. thermal_notify_framework
747 -----------------------------
749 This function handles the trip events from sensor drivers. It starts
750 throttling the cooling devices according to the policy configured.
751 For CRITICAL and HOT trip points, this notifies the respective drivers,
752 and does actual throttling for other trip points i.e ACTIVE and PASSIVE.
753 The throttling policy is based on the configured platform data; if no
754 platform data is provided, this uses the step_wise throttling policy.
756 4.4. thermal_cdev_update
757 ------------------------
759 This function serves as an arbitrator to set the state of a cooling
760 device. It sets the cooling device to the deepest cooling state if
763 5. thermal_emergency_poweroff
764 =============================
766 On an event of critical trip temperature crossing. Thermal framework
767 allows the system to shutdown gracefully by calling orderly_poweroff().
768 In the event of a failure of orderly_poweroff() to shut down the system
769 we are in danger of keeping the system alive at undesirably high
770 temperatures. To mitigate this high risk scenario we program a work
771 queue to fire after a pre-determined number of seconds to start
772 an emergency shutdown of the device using the kernel_power_off()
773 function. In case kernel_power_off() fails then finally
774 emergency_restart() is called in the worst case.
776 The delay should be carefully profiled so as to give adequate time for
777 orderly_poweroff(). In case of failure of an orderly_poweroff() the
778 emergency poweroff kicks in after the delay has elapsed and shuts down
781 If set to 0 emergency poweroff will not be supported. So a carefully
782 profiled non-zero positive value is a must for emergerncy poweroff to be