Hi Eduardo,
On Fri, Sep 27, 2013 at 04:13:10AM +0100, Eduardo Valentin wrote:
> This patch introduces a device tree bindings for
> describing the hardware thermal behavior and limits.
> Also a parser to read and interpret the data and feed
> it in the thermal framework is presented.
>
> This patch introduces a thermal data parser for device
> tree. The parsed data is used to build thermal zones
> and thermal binding parameters. The output data
> can then be used to deploy thermal policies.
>
> This patch adds also documentation regarding this
> API and how to define tree nodes to use
> this infrastructure.
>
> Note that, in order to be able to have control
> on the sensor registration on the DT thermal zone,
> it was required to allow changing the thermal zone
> .get_temp callback. For this reason, this patch
> also removes the 'const' modifier from the .ops
> field of thermal zone devices.
>
> Cc: Zhang Rui <rui.zhang@xxxxxxxxx>
> Cc: linux-pm@xxxxxxxxxxxxxxx
> Cc: linux-kernel@xxxxxxxxxxxxxxx
> Signed-off-by: Eduardo Valentin <eduardo.valentin@xxxxxx>
> ---
>
> Hello folks,
>
> For those interested on this work, here is a short changelog from v6:
> - Reviewed binding documentation and added type and size on all properties
> - Described 'thermal-zones' node in binding documentation
> - Using now cooling state terminology instead of cooling level
> - Renamed milliCelsius to millicelsius
> - Renamed cooling-attachments to 'cooling-maps'
> - Renamed sensor to thermal sensor
> - Renamed #sensor-cells to #thermal-sensor-cells
> - #thermal-sensor-cells now is allowed to be 0
> - Renamed 'sensors' property to 'thermal-sensors'
> - Changed trip type property to be now string and documented possible values
> - Described better the cooling properties
> - Now parses 'contribution' instead of 'usage'
> - Renamed cdev to cooling device
> - Renamed 'passive-delay' property to 'polling-delay-passive'
> - Fixed a couple of error messages to match the property name
> - Fixed a binding issue while using cpufreq as module
> - Reworked thermal_of_build_thermal_zone function so that it frees
> requested memory if something goes wrong and checks for sub-nodes
> with zero children.
>
> ---
> .../devicetree/bindings/thermal/thermal.txt | 537 +++++++++++++
> drivers/thermal/Kconfig | 13 +
> drivers/thermal/Makefile | 1 +
> drivers/thermal/of-thermal.c | 845 +++++++++++++++++++++
> drivers/thermal/thermal_core.c | 9 +-
> drivers/thermal/thermal_core.h | 9 +
> include/dt-bindings/thermal/thermal.h | 27 +
> include/linux/thermal.h | 28 +-
> 8 files changed, 1466 insertions(+), 3 deletions(-)
> create mode 100644 Documentation/devicetree/bindings/thermal/thermal.txt
> create mode 100644 drivers/thermal/of-thermal.c
> create mode 100644 include/dt-bindings/thermal/thermal.h
>
> diff --git a/Documentation/devicetree/bindings/thermal/thermal.txt b/Documentation/devicetree/bindings/thermal/thermal.txt
> new file mode 100644
> index 0000000..3dbc017
> --- /dev/null
> +++ b/Documentation/devicetree/bindings/thermal/thermal.txt
> @@ -0,0 +1,537 @@
> +* Thermal Framework Device Tree descriptor
> +
> +Generic binding to provide a way of defining hardware thermal
> +structure using device tree. A thermal structure includes thermal
> +zones and their components, such as trip points, polling intervals,
> +sensors and cooling devices binding descriptors.
> +
> +The target of device tree thermal descriptors is to describe only
> +the hardware thermal aspects, not how the system must control or which
> +algorithm or policy must be taken in place.
> +
> +There are five types of nodes involved to describe thermal bindings:
> +- sensors: used to describe the device source of temperature sensing;
> +- cooling devices: used to describe devices source of power dissipation control;
> +- trip points: used to describe points in temperature domain defined to
> +make the system aware of hardware limits;
> +- cooling maps: used to describe links between trip points and
> +cooling devices;
> +- thermal zones: used to describe thermal data within the hardware;
> +
> +It follows a description of each type of these device tree nodes.
> +
> +Note: This binding is a working in progress.
What does this mean? Is the binding expected to change in incompatible
ways?
I see we have similar wording at the top of the clock bindings, which
should probably go -- it's ABI now...
> +
> +* Thermal sensor devices
> +
> +Thermal sensor devices are nodes providing temperature sensing capabilities on
> +thermal zones. Typical devices are I2C ADC converters and bandgaps. Theses are
> +nodes providing temperature data to thermal zones. Thermal sensor devices may
> +control one or more internal sensors.
> +
> +Required property:
> +- #thermal-sensor-cells: Used to provide sensor device specific information
> + Type: unsigned while referring to it. Typically 0, on thermal sensor
> + Size: one cell nodes with only one sensor, and at least 1 on nodes
> + with several internal sensors, in order
> + to identify uniquely the sensor instances within
> + the IC. See thermal zone binding for more details
> + on how consumers refer to sensor devices.
> +
> +* Cooling device nodes
> +
> +Cooling devices are nodes providing control on power dissipation. There
> +are essentially two ways to provide control on power dissipation. First
> +is by means of regulating device performance, which is known as passive
> +cooling. A typical passive cooling is a CPU that has dynamic voltage and
> +frequency scaling, and uses lower frequencies as cooling states.
> +Second is by means of activating devices in order to remove
> +the dissipated heat, which is known as active cooling, e.g. regulating
> +fan speeds. In both cases, cooling devices shall have a way to determine
> +the state of cooling in which the device is.
> +
> +Required property:
> +- cooling-min-state: An integer indicating the smallest
> + Type: unsigned cooling state accepted. Typically 0.
> + Size: one cell
> +
> +- cooling-max-state: An integer indicating the largest
> + Type: unsigned cooling state accepted.
> + Size: one cell
> +
> +- #cooling-cells: Used to provide cooling device specific information
> + Type: unsigned while referring to it. Must be at least 2, in order
> + Size: one cell to specify minimum and maximum cooling state used
> + in the reference. The first cell is the minimum
> + cooling state requested and the second cell is
> + the maximum cooling state requested in the reference.
> + See Cooling device maps section below for more details
> + on how consumers refer to cooling devices.
> +
> +* Trip points
> +
> +The trip node is a node to describe a point in the temperature domain
> +in which the system takes an action. This node describes just the point,
> +not the action.
> +
> +Required properties:
> +- temperature: An integer indicating the trip temperature level,
> + Type: signed in millicelsius.
> + Size: one cell
> +
> +- hysteresis: a (low) hysteresis value on 'temperature'. This is a
> + Type: unsigned relative value, in millicelsius.
> + Size: one cell
> +
> +- type: a string containing the trip type. Supported values are:
Odd spacing above.
> + "active": A trip point to enable active cooling
> + "passive": A trip point to enable passive cooling
> + "hot": A trip point to notify emergency
> + "critical": Hardware not reliable.
> + Type: string
> +
> +There are also string constants defined at
> +include/dt-bindings/thermal/thermal.h.
> +
> +* Cooling device maps
> +
> +The cooling device maps node is a node to describe how cooling devices
> +get assigned to trip points of the zone. The cooling devices are expected
> +to be loaded in the target system.
> +
> +Required properties:
> +- cooling-device: A phandle of a cooling device with its specifier,
> + Type: phandle referring to which cooling device is used in this
More odd spacing. There's a fair amount later that would be nice to fix
up too.
> + binding. The required specifiers are: the minimum
> + cooling state and the maximum cooling level used
> + in this map.
There seems to be some confusion as to the word "specifier". Typically,
a specifier is all the cells after the phandle described by the #*-cells
property for the node pointed to. So here there is one
cooling-specifier, which describes the min and max values, and possibly
some other values.
> +- trip: A phandle of a trip point node within the same thermal
> + Type: phandle zone.
> +
> +Optional property:
> +- contribution: The cooling contribution to the thermal zone of the
> + Type: unsigned referred cooling device at the referred trip point.
> + Size: one cell The contribution is a ratio of the sum
> + of all cooling contributions within a thermal zone.
> +
> +Note: Using the THERMAL_NO_LIMIT (-1L) constant in the cooling-device phandle
> +limit specifier means:
> +(i) - minimum state allowed for minimum cooling level used in the reference.
> +(ii) - maximum state allowed for maximum cooling level used in the reference.
> +Refer to include/dt-bindings/thermal/thermal.h for definition of this constant.
How does that work with an unsigned value? Does the code check for
0xffffffff, or compare against -1?
> +
> +* Thermal zone nodes
> +
> +The thermal zone node is the node containing all the required info
> +for describing a thermal zone, including its cooling device bindings. The
> +thermal zone node must contain, apart from its own properties, one node containing
> +trip nodes and one node containing all the zone cooling maps.
> +
> +Required properties:
> +- polling-delay-passive: The maximum number of milliseconds to wait
> + Type: unsigned between polls when performing passive cooling.
> + Size: one cell
Is this when performing passive cooling, or any passive or above (i.e.
any cooling whatsoever)?
> +
> +- polling-delay: The maximum number of milliseconds to wait between polls
> + Type: unsigned when checking this thermal zone.
> + Size: one cell
> +
> +- thermal-sensors: A list of sensor phandles and sensor specifier
> + Type: list of phandles used while monitoring the thermal zone.
The type and the description don't agree here. The type description can
probably be omitted as the description describes the format of the
property.
> +
> +- trips: A sub-node which is a container of only trip point nodes
> + Type: sub-node required to describe the thermal zone.
> +
> +- cooling-maps: A sub-node which is a container of only cooling device
> + Type: sub-node map nodes, used to describe the relation between trips
> + and cooling devices.
> +
> +Optional property:
> +- thermal-sensors-names: List of thermal sensor name strings sorted
> + Type: list of strings in the same order as the thermal-sensors
> + property.
Was there a reason for adding this?
There was some confusion over my request for cooling-specifier and
thermal-sensor-specifier terminology last time [1]. I was simply asking
for terminology, not -names properties.
> +
> +- coefficients: An array of integers (one signed cell) containing
> + Type: array coefficients to compose a linear relation between
> + Elem size: one cell the sensors described in the thermal-sensors property.
> + Elem type: signed Coefficients defaults to 1, in case this property
> + is not specified. A simple linear polynomial is used:
> + Z = c0 * x0 + c1 + x1 + ... + c(n-1) * x(n-1) + cn.
> +
> + The coefficients are ordered and they match with sensors
> + by means of sensor ID. Additional coefficients are
> + interpreted as constant offsets.
> +
> +Note: The delay properties are bound to the maximum dT/dt (temperature
> +derivative over time) in two situations for a thermal zone:
> +(i) - when active cooling is activated (polling-delay-passive); and
Ah, so any coolnig implies the passive polling delay. That makes teh
naming confusing (active cooling means we're polling for the passive
delay). Sorry for the poor suggestion.
How about having polling-delay-cooled instead. That better describes
when it applies.
> +(ii) - when the zone just needs to be monitored (polling-delay).
> +The maximum dT/dt is highly bound to hardware power consumption and dissipation
> +capability.
It would be nice to have another line break here.
> +The delays are chosen to account for said max dT/dt, such that a device
s/are/should be/
> +does not cross several trip boundaries unexpectedly between polls. Choosing
> +the right polling delays shall avoid having the device in temperature ranges
> +that may damage the silicon structures and reduce silicon lifetime.
> +
> +* The thermal-zones node
> +
> +The "thermal-zones" node is a container for all thermal zone nodes. It shall
> +contain only sub-nodes describing thermal zones as in the section
> +"Thermal zone nodes".
I assume thermal-zones appears under /, rather than under a subnode? It
might be worth mentioning.
> +
> +* Examples
> +
> +Below are several examples on how to use thermal data descriptors
> +using device tree bindings:
> +
> +(a) - CPU thermal zone
> +
> +The CPU thermal zone example below describes how to setup one thermal zone
> +using one single sensor as temperature source and many cooling devices and
> +power dissipation control sources.
> +
> +#include <dt-bindings/thermal/thermal.h>
> +
> +cpus {
> + cpu0: cpu@0 {
> + ...
> + cooling-min-state = <0>;
> + cooling-max-state = <3>;
> + #cooling-cells = <2>; /* min followed by max */
> + };
> + ...
> +};
If we're going to give an example of cpu nodes with cooling-cells, we
should have a binding for cpu nodes explaining what the abstract cooling
properties mean for them.
What _specifically_ do each of the values 0,1,2,3 mean here? Are they
indexes of OPPs to use (which aren't shown in the example)? Are they
specific to a particular CPU (the compatible string for which is not
shown in the example)? Are they generic and applicable to all CPUs
somehow?
> +
> +&i2c1 {
> + ...
Having a comment inline describing this fictional sensor would be
useful, something like:
/*
* A simple fan controlller. Supports 10 speeds of operation
* (represented as 0-9).
*/
I didn't notice the block below until already having tried to comprehend
the example. Inline comments would help in that regard.
> + fan0: fan@0x48 {
> + ...
> + cooling-min-state = <0>;
> + cooling-max-state = <9>;
> + #cooling-cells = <2>; /* min followed by max */
> + };
> +};
> +
Similarly, it would be nice to have something here like:
/*
* A simple IC with a single bandgap temperature sensor.
*/
> +bandgap0: bandgap@0x0000ED00 {
> + ...
> + #thermal-sensor-cells = <0>;
> +};
> +
> +cpu-thermal: cpu-thermal {
This should be under a thermal-zones node. Similarly for the other
examples, they should have all relevant context.
> + polling-delay-passive = <250>; /* milliseconds */
> + polling-delay = <1000>; /* milliseconds */
> +
The comment below is no longer relevant.
> + /* sensor ID */
> + thermal-sensors = <&bandgap0>;
> +
> + trips {
> + cpu-alert0: cpu-alert {
> + temperature = <90000>; /* millicelsius */
> + hysteresis = <2000>; /* millicelsius */
> + type = THERMAL_TRIP_ACTIVE;
This got changed to a string in the binding, but is a CPP value here.
Similarly in many other places below.
> + };
> + cpu-alert1: cpu-alert {
> + temperature = <100000>; /* millicelsius */
> + hysteresis = <2000>; /* millicelsius */
> + type = THERMAL_TRIP_PASSIVE;
> + };
> + cpu-crit: cpu-crit {
> + temperature = <125000>; /* millicelsius */
> + hysteresis = <2000>; /* millicelsius */
> + type = THERMAL_TRIP_CRITICAL;
> + };
> + };
> +
> + cooling-maps {
> + map0 {
> + trip = <&cpu-alert0>;
> + cooling-device = <&fan0 THERMAL_NO_LIMITS 4>;
> + };
> + map1 {
> + trip = <&cpu-alert1>;
> + cooling-device = <&fan0 5 THERMAL_NO_LIMITS>;
> + };
> + map2 {
> + trip = <&cpu-alert1>;
> + cooling-device =
> + <&cpu0 THERMAL_NO_LIMITS THERMAL_NO_LIMITS>;
> + };
> + };
> +};
> +
> +In the example above, the ADC sensor at address 0x0000ED00 is used to monitor
> +the zone 'cpu-thermal' using its the sensor 0. The fan0, a fan device controlled
> +via I2C bus 1, at adress 0x48, which has ten different cooling states.
This is a little difficult to understand. How about:
In the example above, the ADC sensor (bandgap0) at address 0x0000ED00 is
used to monitor the zone 'cpu-thermal' using its sole sensor. A fan
device (fan0) is controlled via I2C bus 1, at address 0x48, and has ten
different cooling states 0-9.
> +It is used to remove the heat out of the thermal zone 'cpu-thermal' using its
> +cooling states from its minimum to 4, when it reaches trip point 'cpu-alert0'
> +at 90C, as an example of active cooling. The same cooling device is used at
> +'cpu-alert1', but from 5 to its maximum state. The cpu@0 device is also
> +linked to the same thermal zone, 'cpu-thermal', as a passive cooling device,
> +using all its cooling states at trip point 'cpu-alert1',
> +which is a trip point at 100C.
> +
> +(b) - IC with several internal sensors
> +
> +The example below describes how to deploy several thermal zones based off a
> +single sensor IC, assuming it has several internal sensors. This is a common
> +case on SoC designs with several internal IPs that may need different thermal
> +requirements, and thus may have their own sensor to monitor or detect internal
> +hotspots in their silicon.
> +
> +#include <dt-bindings/thermal/thermal.h>
> +
> +bandgap0: bandgap@0x0000ED00 {
> + ...
> + #thermal-sensor-cells = <1>;
> +};
> +
> +cpu-thermal: cpu-thermal {
> + polling-delay-passive = <250>; /* milliseconds */
> + polling-delay = <1000>; /* milliseconds */
> +
> + /* sensor ID */
> + thermal-sensors = <&bandgap0 0>;
> +
> + trips {
> + /* each zone within the SoC may have its own trips */
> + cpu-alert: cpu-alert {
> + temperature = <100000>; /* millicelsius */
> + hysteresis = <2000>; /* millicelsius */
> + type = THERMAL_TRIP_PASSIVE;
> + };
> + cpu-crit: cpu-crit {
> + temperature = <125000>; /* millicelsius */
> + hysteresis = <2000>; /* millicelsius */
> + type = THERMAL_TRIP_CRITICAL;
> + };
> + };
> +
> + cooling-maps {
> + /* each zone within the SoC may have its own cooling */
> + ...
> + };
> +};
> +
> +gpu-thermal: gpu-thermal {
> + polling-delay-passive = <120>; /* milliseconds */
> + polling-delay = <1000>; /* milliseconds */
> +
> + /* sensor ID */
> + thermal-sensors = <&bandgap0 1>;
> +
> + trips {
> + /* each zone within the SoC may have its own trips */
> + gpu-alert: gpu-alert {
> + temperature = <90000>; /* millicelsius */
> + hysteresis = <2000>; /* millicelsius */
> + type = THERMAL_TRIP_PASSIVE;
> + };
> + gpu-crit: gpu-crit {
> + temperature = <105000>; /* millicelsius */
> + hysteresis = <2000>; /* millicelsius */
> + type = THERMAL_TRIP_CRITICAL;
> + };
> + };
> +
> + cooling-maps {
> + /* each zone within the SoC may have its own cooling */
> + ...
> + };
> +};
> +
> +dsp-thermal: dsp-thermal {
> + polling-delay-passive = <50>; /* milliseconds */
> + polling-delay = <1000>; /* milliseconds */
> +
> + /* sensor ID */
> + thermal-sensors = <&bandgap0 2>;
> +
> + trips {
> + /* each zone within the SoC may have its own trips */
> + dsp-alert: gpu-alert {
> + temperature = <90000>; /* millicelsius */
> + hysteresis = <2000>; /* millicelsius */
> + type = THERMAL_TRIP_PASSIVE;
> + };
> + dsp-crit: gpu-crit {
> + temperature = <135000>; /* millicelsius */
> + hysteresis = <2000>; /* millicelsius */
> + type = THERMAL_TRIP_CRITICAL;
> + };
> + };
> +
> + cooling-maps {
> + /* each zone within the SoC may have its own cooling */
> + ...
> + };
> +};
> +
> +In the example above there is one bandgap IC which has the capability to
> +monitor three sensors. The hardware has been designed so that sensors are
> +placed on different places in the DIE to monitor different temperature
> +hotspots: one for CPU thermal zone, one for GPU thermal zone and the
> +other to monitor a DSP thermal zone.
> +
> +Thus, there is a need to assign each sensor provided by the bandgap IC
> +to different thermal zones. This is achieved by means of using the
> +#thermal-sensor-cells property and using the first specifier as sensor ID.
> +In the example, then, bandgap.sensor0 is used to monitor CPU thermal zone,
The reference to bandgap.sensor0 is confusing. How about saying
<&bandgap 0> instead? It makes it far easier to see correspondence
between the text and the dt. Similarly for the other instances below.
> +bandgap.sensor1 is used to monitor GPU thermal zone and bandgap.sensor2
> +is used to monitor DSP thermal zone. Each zone may be uncorrelated,
> +having its own dT/dt requirements, trips and cooling maps.
> +
> +
> +(c) - Several sensors within one single thermal zone
> +
> +The example below illustrates how to use more than one sensor within
> +one thermal zone.
> +
> +#include <dt-bindings/thermal/thermal.h>
> +
> +&i2c1 {
> + ...
> + adc: sensor@0x49 {
> + ...
> + #thermal-sensor-cells = <0>;
> + };
> +};
> +
> +bandgap0: bandgap@0x0000ED00 {
> + ...
> + #thermal-sensor-cells = <0>;
> +};
> +
> +cpu-thermal: cpu-thermal {
> + polling-delay-passive = <250>; /* milliseconds */
> + polling-delay = <1000>; /* milliseconds */
> +
> + /* sensor ID */
> + thermal-sensors = <&bandgap0>,
> + <&adc>;
> + thermal-sensors-names = "cpu", "pcb north";
> +
> + /* hotspot = 100 * bandgap - 120 * adc + 484 */
> + coefficients = <100 -120 484>;
> +
> + trips {
> + ...
> + };
> +
> + cooling-maps {
> + ...
> + };
> +};
> +
> +In some cases, there is a need to use more than one sensor to extrapolate
> +a thermal hotspot in the silicon. The above example illustrate this situation.
s/illustrate/illustrates/
> +For instance, it may be the case that a sensor external to CPU IP may be place
s/place/placed/
> +close to CPU hotspot and together with internal CPU sensor, it is used
> +to determine the hotspot. The hyppotetical extrapolation rule would be:
s/hyppotetical/hypothetical/
[...]
> +static int of_thermal_set_trip_temp(struct thermal_zone_device *tz, int trip,
> + unsigned long temp)
> +{
> + struct __thermal_zone *data = tz->devdata;
> +
> + if (trip >= data->ntrips || trip < 0)
> + return -EDOM;
> +
> + /* thermal fw should take care of data->mask & (1 << trip) */
I mentioned this last time, and it's tripped me up again while reading:
It's rather too easy to read "fw" as "firmware" and get confused.
s/fw/framework/ please. :)
> + data->trips[trip].temperature = temp;
> +
> + return 0;
> +}
[...]
> +struct thermal_zone_device *
> +thermal_zone_of_sensor_register(struct device *dev, int sensor_id,
> + void *data, int (*get_temp)(void *, long *),
> + int (*get_trend)(void *, long *))
> +{
> + struct device_node *np, *child, *sensor_np;
> +
> + np = of_find_node_by_name(NULL, "thermal-zones");
> + if (!np)
> + return ERR_PTR(-ENODEV);
> +
> + if (!dev || !dev->of_node)
> + return ERR_PTR(-EINVAL);
> +
> + sensor_np = dev->of_node;
> +
> + for_each_child_of_node(np, child) {
> + struct of_phandle_args sensor_specs;
> + int ret, id;
> +
> + /* For now, thermal framework supports only 1 sensor per zone */
> + ret = of_parse_phandle_with_args(child, "thermal-sensors",
> + "#thermal-sensor-cells",
> + 0, &sensor_specs);
> + if (ret)
> + continue;
> +
> + if (sensor_specs.args_count < 1)
> + id = 0;
> + else
> + id = sensor_specs.args[0];
> +
For the moment it might be worth printing a warning if args_count > 1,
as thermal management might not function correctly.
> + if (sensor_specs.np == sensor_np && id == sensor_id) {
> + of_node_put(np);
> + return thermal_zone_of_add_sensor(child, sensor_np,
> + data,
> + get_temp,
> + get_trend);
> + }
> + }
> + of_node_put(np);
> +
> + return ERR_PTR(-ENODEV);
> +}
> +EXPORT_SYMBOL_GPL(thermal_zone_of_sensor_register);
Cheers,
Mark.
[1] https://lkml.org/lkml/2013/9/24/501
_______________________________________________
lm-sensors mailing list
lm-sensors@xxxxxxxxxxxxxx
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
