Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > linux.kernel > #1277468 > unrolled thread
| Started by | Javi Merino <javi.merino@arm.com> |
|---|---|
| First post | 2015-11-25 16:20 +0100 |
| Last post | 2015-11-25 19:50 +0100 |
| Articles | 4 — 2 participants |
Back to article view | Back to linux.kernel
[PATCH v3 0/4] Hierarchical thermal zones Javi Merino <javi.merino@arm.com> - 2015-11-25 16:20 +0100
[PATCH v3 2/4] devicetree: bindings: let thermal-sensor point to other thermal zones Javi Merino <javi.merino@arm.com> - 2015-11-25 16:20 +0100
Re: [PATCH v3 2/4] devicetree: bindings: let thermal-sensor point to other thermal zones Mark Rutland <mark.rutland@arm.com> - 2015-11-25 19:00 +0100
Re: [PATCH v3 2/4] devicetree: bindings: let thermal-sensor point to other thermal zones Javi Merino <javi.merino@arm.com> - 2015-11-25 19:50 +0100
| From | Javi Merino <javi.merino@arm.com> |
|---|---|
| Date | 2015-11-25 16:20 +0100 |
| Subject | [PATCH v3 0/4] Hierarchical thermal zones |
| Message-ID | <qyK1r-7ZU-9@gated-at.bofh.it> |
This series adds the ability to create a hierarchy of thermal zones.
Thermal zones created via platform code or device tree can be set up
to calculate their temperature as the maximum or weighted average of
all its underlying thermal zones. This came up from discussions
during LPC.
The first patch adds the basic support to thermal core. Patch 2
extends the devicetree bindings to cope with a hierarchy of thermal
zones. Patch 3 adds device tree support. The last patch exports the
hierarchy to sysfs, adding knobs to change the aggregation function
and adjust the weights of thermal zones.
Changes since v2:
- The aggregation function can be maximum or weighted average
- Separated the update of the devicetree binding and the of-thermal
into two separate patches
Javi Merino (4):
thermal: Add support for hierarchical thermal zones
devicetree: bindings: let thermal-sensor point to other thermal zones
thermal: of: parse stacked thermal zones from device tree
thermal: show the sub thermal zones in sysfs
.../devicetree/bindings/thermal/thermal.txt | 154 +++++++++-
Documentation/thermal/sysfs-api.txt | 72 +++++
drivers/thermal/of-thermal.c | 100 ++++++
drivers/thermal/thermal_core.c | 335 ++++++++++++++++++++-
include/linux/thermal.h | 44 ++-
5 files changed, 696 insertions(+), 9 deletions(-)
--
1.9.1
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[toc] | [next] | [standalone]
| From | Javi Merino <javi.merino@arm.com> |
|---|---|
| Date | 2015-11-25 16:20 +0100 |
| Subject | [PATCH v3 2/4] devicetree: bindings: let thermal-sensor point to other thermal zones |
| Message-ID | <qyK1s-7ZU-31@gated-at.bofh.it> |
| In reply to | #1277468 |
The thermal-sensor property of the thermal zone node accepts phandles to
thermal sensors. However, thermal zones can be created as an
aggregation of other thermal zones. Extend the thermal-sensors property
to allow phandles to other thermal zones. This patch also adds an
example that showcases how a board thermal zone can be created from the
aggregation of the cpu, gpu and lcd thermal zones.
Cc: Zhang Rui <rui.zhang@intel.com>
Cc: Eduardo Valentin <edubezval@gmail.com>
Cc: Rob Herring <robh+dt@kernel.org>
Cc: Pawel Moll <pawel.moll@arm.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Ian Campbell <ijc+devicetree@hellion.org.uk>
Cc: Kumar Gala <galak@codeaurora.org>
Cc: devicetree@vger.kernel.org
Signed-off-by: Javi Merino <javi.merino@arm.com>
---
Notes:
Hi devicetree,
Is it ok to extend the definition of the thermal-sensors property like
this? IOW are phandles strongly typed?
.../devicetree/bindings/thermal/thermal.txt | 154 ++++++++++++++++++++-
1 file changed, 151 insertions(+), 3 deletions(-)
diff --git a/Documentation/devicetree/bindings/thermal/thermal.txt b/Documentation/devicetree/bindings/thermal/thermal.txt
index 41b817f7b670..52b7e9ae3b4d 100644
--- a/Documentation/devicetree/bindings/thermal/thermal.txt
+++ b/Documentation/devicetree/bindings/thermal/thermal.txt
@@ -145,9 +145,12 @@ Required properties:
Size: one cell
- thermal-sensors: A list of thermal sensor phandles and sensor specifier
- Type: list of used while monitoring the thermal zone.
- phandles + sensor
- specifier
+ Type: list of used while monitoring the thermal zone. The phandles
+ phandles + sensor can point to thermal sensors or other thermal zone
+ specifier nodes. If it points to other thermal zone
+ nodes you should omit the sensor specifier
+ and set #thermal-sensor-cells to 0 for the
+ thermal zone.
- trips: A sub-node which is a container of only trip point nodes
Type: sub-node required to describe the thermal zone.
@@ -603,3 +606,148 @@ thermal-zones {
The above example is a mix of previous examples, a sensor IP with several internal
sensors used to monitor different zones, one of them is composed by several sensors and
with different cooling devices.
+
+(e) Board thermal with stacked thermal zones
+
+Instead of setting up one thermal zone combining multiple thermal
+zones and multiple trip points for each cooling device, we can create
+a hierarchy of thermal zones.
+
+#include <dt-bindings/thermal/thermal.h>
+
+&i2c1 {
+ ...
+ /*
+ * An IC with several temperature sensor.
+ */
+ adc_dummy: sensor@0x50 {
+ ...
+ #thermal-sensor-cells = <1>; /* sensor internal ID */
+ };
+};
+
+thermal-zones {
+
+ cpu_thermal: cpu_thermal {
+ polling-delay-passive = <1000>; /* milliseconds */
+ polling-delay = <2500>; /* milliseconds */
+
+ sustainable-power = <2500>;
+
+ thermal-sensors = <&adc_dummy 0>
+
+ trips {
+ cpu_trip: cpu-trip {
+ temperature = <60000>; /* millicelsius */
+ hysteresis = <2000>; /* millicelsius */
+ type = "passive";
+ };
+ };
+
+ cooling-maps {
+ map0 {
+ trip = <&cpu_trip>;
+ cooling-device = <&cpu0 0 2>;
+ };
+ };
+ };
+
+ gpu_thermal: gpu_thermal {
+ polling-delay-passive = <1000>; /* milliseconds */
+ polling-delay = <2500>; /* milliseconds */
+
+ sustainable-power = <2500>;
+
+ thermal-sensors = <&adc_dummy 2>
+
+ trips {
+ gpu_trip: gpu-trip {
+ temperature = <55000>; /* millicelsius */
+ hysteresis = <2000>; /* millicelsius */
+ type = "passive";
+ }
+ };
+
+ cooling-maps {
+ map0 {
+ trip = <&gpu_trip>;
+ cooling-device = <&gpu0 0 2>;
+ };
+ };
+ };
+
+ lcd_thermal: lcd_thermal {
+ polling-delay-passive = <1000>; /* milliseconds */
+ polling-delay = <2500>; /* milliseconds */
+
+ sustainable-power = <2500>;
+
+ thermal-sensors = <&adc_dummy 1>
+
+ trips {
+ lcd_trip: lcp-trip {
+ temperature = <53000>; /* millicelsius */
+ hysteresis = <2000>; /* millicelsius */
+ type = "passive";
+ };
+ };
+
+ cooling-maps {
+ map0 {
+ trip = <&lcd_trip>;
+ cooling-device = <&lcd0 5 10>;
+ };
+ };
+ };
+
+ board_thermal: board-thermal {
+ polling-delay-passive = <1000>; /* milliseconds */
+ polling-delay = <2500>; /* milliseconds */
+
+ thermal-sensors = <&cpu_thermal &gpu_thermal &lcd_thermal>
+
+ sustainable-power = <2500>;
+
+ trips {
+ warm_trip: warm-trip {
+ temperature = <62000>; /* millicelsius */
+ hysteresis = <2000>; /* millicelsius */
+ type = "passive";
+ };
+ crit_trip: crit-trip {
+ temperature = <68000>; /* millicelsius */
+ hysteresis = <2000>; /* millicelsius */
+ type = "critical";
+ };
+ };
+
+ cooling-maps {
+ map0 {
+ trip = <&warm_trip>;
+ cooling-device = <&cpu0 2 2>;
+ contribution = <55>;
+ };
+ map1 {
+ trip = <&warm_trip>;
+ cooling-device = <&gpu0 2 2>;
+ contribution = <20>;
+ };
+ map2 {
+ trip = <&lcd_trip>;
+ cooling-device = <&lcd0 7 10>;
+ contribution = <15>;
+ };
+ };
+ };
+};
+
+The above example is a different take at example (d). We create one
+thermal zone per sensor: cpu_thermal, gpu_thermal and lcd_thermal.
+Each of which has its own trip point for each own cooling device. We
+then define a board_thermal thermal zone that is a combination of all
+the other thermal zones. If the board hits its warm_trip, then all
+cooling devices are throttled.
+
+This example illustrates how we can throttle each device individually
+if its too hot and at the same time have some control over the whole
+system.
--
1.9.1
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[toc] | [prev] | [next] | [standalone]
| From | Mark Rutland <mark.rutland@arm.com> |
|---|---|
| Date | 2015-11-25 19:00 +0100 |
| Subject | Re: [PATCH v3 2/4] devicetree: bindings: let thermal-sensor point to other thermal zones |
| Message-ID | <qyMwj-12v-27@gated-at.bofh.it> |
| In reply to | #1277471 |
On Wed, Nov 25, 2015 at 03:09:44PM +0000, Javi Merino wrote:
> The thermal-sensor property of the thermal zone node accepts phandles to
> thermal sensors. However, thermal zones can be created as an
> aggregation of other thermal zones. Extend the thermal-sensors property
> to allow phandles to other thermal zones. This patch also adds an
> example that showcases how a board thermal zone can be created from the
> aggregation of the cpu, gpu and lcd thermal zones.
>
> Cc: Zhang Rui <rui.zhang@intel.com>
> Cc: Eduardo Valentin <edubezval@gmail.com>
> Cc: Rob Herring <robh+dt@kernel.org>
> Cc: Pawel Moll <pawel.moll@arm.com>
> Cc: Mark Rutland <mark.rutland@arm.com>
> Cc: Ian Campbell <ijc+devicetree@hellion.org.uk>
> Cc: Kumar Gala <galak@codeaurora.org>
> Cc: devicetree@vger.kernel.org
> Signed-off-by: Javi Merino <javi.merino@arm.com>
> ---
>
> Notes:
> Hi devicetree,
>
> Is it ok to extend the definition of the thermal-sensors property like
> this? IOW are phandles strongly typed?
I think it's OK so long as each thermal zone has #theremal-sensor-cells
set explicitly, if used as a sensor, and we can agree on the semantics
of what it means for a thermal zone to be a sensor.
I don't really follow why you need the zone to be a sensor, and can't
simply refer to the sensor from two zones. Are you trying to imply an
ordering of trip points (e.g. that the sub-zones' trips should be taken
into account first)?
>
> .../devicetree/bindings/thermal/thermal.txt | 154 ++++++++++++++++++++-
> 1 file changed, 151 insertions(+), 3 deletions(-)
>
> diff --git a/Documentation/devicetree/bindings/thermal/thermal.txt b/Documentation/devicetree/bindings/thermal/thermal.txt
> index 41b817f7b670..52b7e9ae3b4d 100644
> --- a/Documentation/devicetree/bindings/thermal/thermal.txt
> +++ b/Documentation/devicetree/bindings/thermal/thermal.txt
> @@ -145,9 +145,12 @@ Required properties:
> Size: one cell
>
> - thermal-sensors: A list of thermal sensor phandles and sensor specifier
> - Type: list of used while monitoring the thermal zone.
> - phandles + sensor
> - specifier
> + Type: list of used while monitoring the thermal zone. The phandles
> + phandles + sensor can point to thermal sensors or other thermal zone
> + specifier nodes. If it points to other thermal zone
> + nodes you should omit the sensor specifier
> + and set #thermal-sensor-cells to 0 for the
> + thermal zone.
The example misses #thermal-sensor-cells = <0> for each of the zones.
Can a zone normal have multiple sensors? If so, what is the aggregate
value if a zone is used as a sensor? Max? Min? Scaled by contribution
somehow?
Thanks,
Mark.
>
> - trips: A sub-node which is a container of only trip point nodes
> Type: sub-node required to describe the thermal zone.
> @@ -603,3 +606,148 @@ thermal-zones {
> The above example is a mix of previous examples, a sensor IP with several internal
> sensors used to monitor different zones, one of them is composed by several sensors and
> with different cooling devices.
> +
> +(e) Board thermal with stacked thermal zones
> +
> +Instead of setting up one thermal zone combining multiple thermal
> +zones and multiple trip points for each cooling device, we can create
> +a hierarchy of thermal zones.
> +
> +#include <dt-bindings/thermal/thermal.h>
> +
> +&i2c1 {
> + ...
> + /*
> + * An IC with several temperature sensor.
> + */
> + adc_dummy: sensor@0x50 {
> + ...
> + #thermal-sensor-cells = <1>; /* sensor internal ID */
> + };
> +};
> +
> +thermal-zones {
> +
> + cpu_thermal: cpu_thermal {
> + polling-delay-passive = <1000>; /* milliseconds */
> + polling-delay = <2500>; /* milliseconds */
> +
> + sustainable-power = <2500>;
> +
> + thermal-sensors = <&adc_dummy 0>
> +
> + trips {
> + cpu_trip: cpu-trip {
> + temperature = <60000>; /* millicelsius */
> + hysteresis = <2000>; /* millicelsius */
> + type = "passive";
> + };
> + };
> +
> + cooling-maps {
> + map0 {
> + trip = <&cpu_trip>;
> + cooling-device = <&cpu0 0 2>;
> + };
> + };
> + };
> +
> + gpu_thermal: gpu_thermal {
> + polling-delay-passive = <1000>; /* milliseconds */
> + polling-delay = <2500>; /* milliseconds */
> +
> + sustainable-power = <2500>;
> +
> + thermal-sensors = <&adc_dummy 2>
> +
> + trips {
> + gpu_trip: gpu-trip {
> + temperature = <55000>; /* millicelsius */
> + hysteresis = <2000>; /* millicelsius */
> + type = "passive";
> + }
> + };
> +
> + cooling-maps {
> + map0 {
> + trip = <&gpu_trip>;
> + cooling-device = <&gpu0 0 2>;
> + };
> + };
> + };
> +
> + lcd_thermal: lcd_thermal {
> + polling-delay-passive = <1000>; /* milliseconds */
> + polling-delay = <2500>; /* milliseconds */
> +
> + sustainable-power = <2500>;
> +
> + thermal-sensors = <&adc_dummy 1>
> +
> + trips {
> + lcd_trip: lcp-trip {
> + temperature = <53000>; /* millicelsius */
> + hysteresis = <2000>; /* millicelsius */
> + type = "passive";
> + };
> + };
> +
> + cooling-maps {
> + map0 {
> + trip = <&lcd_trip>;
> + cooling-device = <&lcd0 5 10>;
> + };
> + };
> + };
> +
> + board_thermal: board-thermal {
> + polling-delay-passive = <1000>; /* milliseconds */
> + polling-delay = <2500>; /* milliseconds */
> +
> + thermal-sensors = <&cpu_thermal &gpu_thermal &lcd_thermal>
> +
> + sustainable-power = <2500>;
> +
> + trips {
> + warm_trip: warm-trip {
> + temperature = <62000>; /* millicelsius */
> + hysteresis = <2000>; /* millicelsius */
> + type = "passive";
> + };
> + crit_trip: crit-trip {
> + temperature = <68000>; /* millicelsius */
> + hysteresis = <2000>; /* millicelsius */
> + type = "critical";
> + };
> + };
> +
> + cooling-maps {
> + map0 {
> + trip = <&warm_trip>;
> + cooling-device = <&cpu0 2 2>;
> + contribution = <55>;
> + };
> + map1 {
> + trip = <&warm_trip>;
> + cooling-device = <&gpu0 2 2>;
> + contribution = <20>;
> + };
> + map2 {
> + trip = <&lcd_trip>;
> + cooling-device = <&lcd0 7 10>;
> + contribution = <15>;
> + };
> + };
> + };
> +};
> +
> +The above example is a different take at example (d). We create one
> +thermal zone per sensor: cpu_thermal, gpu_thermal and lcd_thermal.
> +Each of which has its own trip point for each own cooling device. We
> +then define a board_thermal thermal zone that is a combination of all
> +the other thermal zones. If the board hits its warm_trip, then all
> +cooling devices are throttled.
> +
> +This example illustrates how we can throttle each device individually
> +if its too hot and at the same time have some control over the whole
> +system.
> --
> 1.9.1
>
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[toc] | [prev] | [next] | [standalone]
| From | Javi Merino <javi.merino@arm.com> |
|---|---|
| Date | 2015-11-25 19:50 +0100 |
| Subject | Re: [PATCH v3 2/4] devicetree: bindings: let thermal-sensor point to other thermal zones |
| Message-ID | <qyNiF-1z2-13@gated-at.bofh.it> |
| In reply to | #1277664 |
On Wed, Nov 25, 2015 at 05:54:41PM +0000, Mark Rutland wrote: > On Wed, Nov 25, 2015 at 03:09:44PM +0000, Javi Merino wrote: > > The thermal-sensor property of the thermal zone node accepts phandles to > > thermal sensors. However, thermal zones can be created as an > > aggregation of other thermal zones. Extend the thermal-sensors property > > to allow phandles to other thermal zones. This patch also adds an > > example that showcases how a board thermal zone can be created from the > > aggregation of the cpu, gpu and lcd thermal zones. > > > > Cc: Zhang Rui <rui.zhang@intel.com> > > Cc: Eduardo Valentin <edubezval@gmail.com> > > Cc: Rob Herring <robh+dt@kernel.org> > > Cc: Pawel Moll <pawel.moll@arm.com> > > Cc: Mark Rutland <mark.rutland@arm.com> > > Cc: Ian Campbell <ijc+devicetree@hellion.org.uk> > > Cc: Kumar Gala <galak@codeaurora.org> > > Cc: devicetree@vger.kernel.org > > Signed-off-by: Javi Merino <javi.merino@arm.com> > > --- > > > > Notes: > > Hi devicetree, > > > > Is it ok to extend the definition of the thermal-sensors property like > > this? IOW are phandles strongly typed? > > I think it's OK so long as each thermal zone has #thermal-sensor-cells > set explicitly, if used as a sensor, and we can agree on the semantics > of what it means for a thermal zone to be a sensor. > > I don't really follow why you need the zone to be a sensor, and can't > simply refer to the sensor from two zones. Are you trying to imply an > ordering of trip points (e.g. that the sub-zones' trips should be taken > into account first)? No, it doesn't affect the ordering of trip points. This came out of a discussion at LPC. Currently thermal zones can only have on thermal sensor associated with them. After some discussion, Mike Turquette suggested that we could use an approach similar to what it's done with power domains and stack them. > > .../devicetree/bindings/thermal/thermal.txt | 154 ++++++++++++++++++++- > > 1 file changed, 151 insertions(+), 3 deletions(-) > > > > diff --git a/Documentation/devicetree/bindings/thermal/thermal.txt b/Documentation/devicetree/bindings/thermal/thermal.txt > > index 41b817f7b670..52b7e9ae3b4d 100644 > > --- a/Documentation/devicetree/bindings/thermal/thermal.txt > > +++ b/Documentation/devicetree/bindings/thermal/thermal.txt > > @@ -145,9 +145,12 @@ Required properties: > > Size: one cell > > > > - thermal-sensors: A list of thermal sensor phandles and sensor specifier > > - Type: list of used while monitoring the thermal zone. > > - phandles + sensor > > - specifier > > + Type: list of used while monitoring the thermal zone. The phandles > > + phandles + sensor can point to thermal sensors or other thermal zone > > + specifier nodes. If it points to other thermal zone > > + nodes you should omit the sensor specifier > > + and set #thermal-sensor-cells to 0 for the > > + thermal zone. > > The example misses #thermal-sensor-cells = <0> for each of the zones. You're right, I'll fix it for the next version > Can a zone normal have multiple sensors? If so, what is the aggregate > value if a zone is used as a sensor? Max? Min? Scaled by contribution > somehow? No, currently a thermal zone can only specify one sensor in its thermal-sensors property Cheers, Javi -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
[toc] | [prev] | [standalone]
Back to top | Article view | linux.kernel
csiph-web