Staging
v0.7.0
v0.7.0
https://github.com/torvalds/linux
Tip revision: e783362eb54cd99b2cac8b3a9aeac942e6f6ac07 authored by Linus Torvalds on 23 January 2022, 08:12:53 UTC
Linux 5.17-rc1
Linux 5.17-rc1
Tip revision: e783362
nvidia,tegra124-soctherm.txt
Tegra124 SOCTHERM thermal management system
The SOCTHERM IP block contains thermal sensors, support for polled
or interrupt-based thermal monitoring, CPU and GPU throttling based
on temperature trip points, and handling external overcurrent
notifications. It is also used to manage emergency shutdown in an
overheating situation.
Required properties :
- compatible : For Tegra124, must contain "nvidia,tegra124-soctherm".
For Tegra132, must contain "nvidia,tegra132-soctherm".
For Tegra210, must contain "nvidia,tegra210-soctherm".
- reg : Should contain at least 2 entries for each entry in reg-names:
- SOCTHERM register set
- Tegra CAR register set: Required for Tegra124 and Tegra210.
- CCROC register set: Required for Tegra132.
- reg-names : Should contain at least 2 entries:
- soctherm-reg
- car-reg
- ccroc-reg
- interrupts : Defines the interrupt used by SOCTHERM
- clocks : Must contain an entry for each entry in clock-names.
See ../clocks/clock-bindings.txt for details.
- clock-names : Must include the following entries:
- tsensor
- soctherm
- resets : Must contain an entry for each entry in reset-names.
See ../reset/reset.txt for details.
- reset-names : Must include the following entries:
- soctherm
- #thermal-sensor-cells : Should be 1. For a description of this property, see
Documentation/devicetree/bindings/thermal/thermal-sensor.yaml.
See <dt-bindings/thermal/tegra124-soctherm.h> for a list of valid values
when referring to thermal sensors.
- throttle-cfgs: A sub-node which is a container of configuration for each
hardware throttle events. These events can be set as cooling devices.
* throttle events: Sub-nodes must be named as "light" or "heavy".
Properties:
- nvidia,priority: Each throttles has its own throttle settings, so the
SW need to set priorities for various throttle, the HW arbiter can select
the final throttle settings.
Bigger value indicates higher priority, In general, higher priority
translates to lower target frequency. SW needs to ensure that critical
thermal alarms are given higher priority, and ensure that there is
no race if priority of two vectors is set to the same value.
The range of this value is 1~100.
- nvidia,cpu-throt-percent: This property is for Tegra124 and Tegra210.
It is the throttling depth of pulse skippers, it's the percentage
throttling.
- nvidia,cpu-throt-level: This property is only for Tegra132, it is the
level of pulse skippers, which used to throttle clock frequencies. It
indicates cpu clock throttling depth, and the depth can be programmed.
Must set as following values:
TEGRA_SOCTHERM_THROT_LEVEL_LOW, TEGRA_SOCTHERM_THROT_LEVEL_MED
TEGRA_SOCTHERM_THROT_LEVEL_HIGH, TEGRA_SOCTHERM_THROT_LEVEL_NONE
- nvidia,gpu-throt-level: This property is for Tegra124 and Tegra210.
It is the level of pulse skippers, which used to throttle clock
frequencies. It indicates gpu clock throttling depth and can be
programmed to any of the following values which represent a throttling
percentage:
TEGRA_SOCTHERM_THROT_LEVEL_NONE (0%)
TEGRA_SOCTHERM_THROT_LEVEL_LOW (50%),
TEGRA_SOCTHERM_THROT_LEVEL_MED (75%),
TEGRA_SOCTHERM_THROT_LEVEL_HIGH (85%).
- #cooling-cells: Should be 1. This cooling device only support on/off state.
For a description of this property see:
Documentation/devicetree/bindings/thermal/thermal-cooling-devices.yaml
Optional properties: The following properties are T210 specific and
valid only for OCx throttle events.
- nvidia,count-threshold: Specifies the number of OC events that are
required for triggering an interrupt. Interrupts are not triggered if
the property is missing. A value of 0 will interrupt on every OC alarm.
- nvidia,polarity-active-low: Configures the polarity of the OC alaram
signal. If present, this means assert low, otherwise assert high.
- nvidia,alarm-filter: Number of clocks to filter event. When the filter
expires (which means the OC event has not occurred for a long time),
the counter is cleared and filter is rearmed. Default value is 0.
- nvidia,throttle-period-us: Specifies the number of uSec for which
throttling is engaged after the OC event is deasserted. Default value
is 0.
Optional properties:
- nvidia,thermtrips : When present, this property specifies the temperature at
which the soctherm hardware will assert the thermal trigger signal to the
Power Management IC, which can be configured to reset or shutdown the device.
It is an array of pairs where each pair represents a tsensor id followed by a
temperature in milli Celcius. In the absence of this property the critical
trip point will be used for thermtrip temperature.
Note:
- the "critical" type trip points will be used to set the temperature at which
the SOC_THERM hardware will assert a thermal trigger if the "nvidia,thermtrips"
property is missing. When the thermtrips property is present, the breach of a
critical trip point is reported back to the thermal framework to implement
software shutdown.
- the "hot" type trip points will be set to SOC_THERM hardware as the throttle
temperature. Once the the temperature of this thermal zone is higher
than it, it will trigger the HW throttle event.
Example :
soctherm@700e2000 {
compatible = "nvidia,tegra124-soctherm";
reg = <0x0 0x700e2000 0x0 0x600 /* SOC_THERM reg_base */
0x0 0x60006000 0x0 0x400 /* CAR reg_base */
reg-names = "soctherm-reg", "car-reg";
interrupts = <GIC_SPI 48 IRQ_TYPE_LEVEL_HIGH>;
clocks = <&tegra_car TEGRA124_CLK_TSENSOR>,
<&tegra_car TEGRA124_CLK_SOC_THERM>;
clock-names = "tsensor", "soctherm";
resets = <&tegra_car 78>;
reset-names = "soctherm";
#thermal-sensor-cells = <1>;
nvidia,thermtrips = <TEGRA124_SOCTHERM_SENSOR_CPU 102500
TEGRA124_SOCTHERM_SENSOR_GPU 103000>;
throttle-cfgs {
/*
* When the "heavy" cooling device triggered,
* the HW will skip cpu clock's pulse in 85% depth,
* skip gpu clock's pulse in 85% level
*/
throttle_heavy: heavy {
nvidia,priority = <100>;
nvidia,cpu-throt-percent = <85>;
nvidia,gpu-throt-level = <TEGRA_SOCTHERM_THROT_LEVEL_HIGH>;
#cooling-cells = <1>;
};
/*
* When the "light" cooling device triggered,
* the HW will skip cpu clock's pulse in 50% depth,
* skip gpu clock's pulse in 50% level
*/
throttle_light: light {
nvidia,priority = <80>;
nvidia,cpu-throt-percent = <50>;
nvidia,gpu-throt-level = <TEGRA_SOCTHERM_THROT_LEVEL_LOW>;
#cooling-cells = <1>;
};
/*
* If these two devices are triggered in same time, the HW throttle
* arbiter will select the highest priority as the final throttle
* settings to skip cpu pulse.
*/
throttle_oc1: oc1 {
nvidia,priority = <50>;
nvidia,polarity-active-low;
nvidia,count-threshold = <100>;
nvidia,alarm-filter = <5100000>;
nvidia,throttle-period-us = <0>;
nvidia,cpu-throt-percent = <75>;
nvidia,gpu-throt-level =
<TEGRA_SOCTHERM_THROT_LEVEL_MED>;
};
};
};
Example: referring to Tegra132's "reg", "reg-names" and "throttle-cfgs" :
soctherm@700e2000 {
compatible = "nvidia,tegra132-soctherm";
reg = <0x0 0x700e2000 0x0 0x600 /* SOC_THERM reg_base */
0x0 0x70040000 0x0 0x200>; /* CCROC reg_base */;
reg-names = "soctherm-reg", "ccroc-reg";
throttle-cfgs {
/*
* When the "heavy" cooling device triggered,
* the HW will skip cpu clock's pulse in HIGH level
*/
throttle_heavy: heavy {
nvidia,priority = <100>;
nvidia,cpu-throt-level = <TEGRA_SOCTHERM_THROT_LEVEL_HIGH>;
#cooling-cells = <1>;
};
/*
* When the "light" cooling device triggered,
* the HW will skip cpu clock's pulse in MED level
*/
throttle_light: light {
nvidia,priority = <80>;
nvidia,cpu-throt-level = <TEGRA_SOCTHERM_THROT_LEVEL_MED>;
#cooling-cells = <1>;
};
/*
* If these two devices are triggered in same time, the HW throttle
* arbiter will select the highest priority as the final throttle
* settings to skip cpu pulse.
*/
};
};
Example: referring to thermal sensors :
thermal-zones {
cpu {
polling-delay-passive = <1000>;
polling-delay = <1000>;
thermal-sensors =
<&soctherm TEGRA124_SOCTHERM_SENSOR_CPU>;
trips {
cpu_shutdown_trip: shutdown-trip {
temperature = <102500>;
hysteresis = <1000>;
type = "critical";
};
cpu_throttle_trip: throttle-trip {
temperature = <100000>;
hysteresis = <1000>;
type = "hot";
};
};
cooling-maps {
map0 {
trip = <&cpu_throttle_trip>;
cooling-device = <&throttle_heavy 1 1>;
};
};
};
};