[PATCH v3 0/2] scpi-sensors: Fix SCP sensor readings scale

[PATCH v3 0/2] scpi-sensors: Fix SCP sensor readings scale

[Carlo Caione]

From: Carlo Caione <carlo@endlessm.com>

The implementation details for SCPI seems to suggest that the sensor
readings must be reported by SCP using a well defined scale
(millidegree Celsius for temperature, millivolts for voltage,
milliamperes for current, microwatts for power and microjoules for
energy).

This is also important for the interaction with other subsystems: for
example both the thermal sub-system and the hwmon sysfs interface expect
the temperature expressed in millidegree Celsius.

Unfortunately since this behaviour is dependent on the firmware
implementation there are cases where the sensor readings are reported
using a different scale. For example in the Amlogic SoCs the
temperature is reported in degree and not millidegree Celsius.

In this patchset we introduce a new DT property `scpi,sensors-scale` that is
used by the scpi-hwmon driver to convert the sensor readings to the expected
scale.

v2:
 - Added data validation (scale != 0)
 - s/unsigned int/u32/ for the array
 - Switched to array initialiser notation
 - Introduced scpi_scale_reading()

v3:
 - scpi_scale_reading marked as static

Carlo Caione (2):
  Documentation: bindings: Introduce scpi,sensors-scale
  hwmon: (scpi) Fix the scale of SCP sensor readings

 Documentation/devicetree/bindings/arm/arm,scpi.txt | 18 ++++++++++++
 drivers/hwmon/scpi-hwmon.c                         | 34 ++++++++++++++++++++++
 2 files changed, 52 insertions(+)

-- 
2.9.3

[PATCH v3 1/2] Documentation: bindings: Introduce scpi,sensors-scale

[Carlo Caione]

From: Carlo Caione <carlo@endlessm.com>

Document the new property `scpi,sensors-scale` used by the hwmon-scpi
driver to convert the sensor readings to the correct / expected scale.

Signed-off-by: Carlo Caione <carlo@endlessm.com>
---
 Documentation/devicetree/bindings/arm/arm,scpi.txt | 18 ++++++++++++++++++
 1 file changed, 18 insertions(+)

diff --git a/Documentation/devicetree/bindings/arm/arm,scpi.txt b/Documentation/devicetree/bindings/arm/arm,scpi.txt
index 4018319..35c621c 100644
--- a/Documentation/devicetree/bindings/arm/arm,scpi.txt
+++ b/Documentation/devicetree/bindings/arm/arm,scpi.txt
@@ -84,6 +84,24 @@ Required properties:
 			 as used by the firmware. Refer to  platform details
 			 for your implementation for the IDs to use.
 
+Optional properties:
+- scpi,sensors-scale: array of 5 elements defining the unit scale used by the
+                      SCP for the 5 different sensor classes (temperature,
+                      voltage, current, power, energy) such that 1 == unit,
+                      1000 == milli, 1000000 == micro, ...
+
+                      Example:
+                         scpi,sensors-scale = <1 1000 1000 1000000 1000000>
+                      In this case the values reported by SCP are:
+                      -> temperature in degree Celsius
+                      -> voltage in millivolts
+                      -> current in milliamperes
+                      -> power in microwatts
+                      -> energy in microjoules
+
+                      The default value is:
+                         scpi,sensors-scale = <1000 1000 1000 1000000 1000000>
+
 Power domain bindings for the power domains based on SCPI Message Protocol
 ------------------------------------------------------------
 
-- 
2.9.3

Re: [PATCH v3 1/2] Documentation: bindings: Introduce scpi,sensors-scale

[Sudeep Holla]

On 09/03/17 10:17, Carlo Caione wrote:
> From: Carlo Caione <carlo@endlessm.com>
> 
> Document the new property `scpi,sensors-scale` used by the hwmon-scpi
> driver to convert the sensor readings to the correct / expected scale.
> 

I am fine with DT bindings if DT maintainers agree with having one but I
would like to add couple of points to get their feedback:

1. ARM recognized the drawbacks of SCPI which started informally with
   ARM Ltd platforms and got adopted by few vendors. It's now being
   worked on and a much better and scalable replacement protocol is
   being discussed with vendors now. So future enhancement to this SCPI
   protocol is unlikely.

2. AmLogic has diverted from base protocol in many ways and most of them
   are handled with just compatible so far without any need for
   additional bindings. Can we avoid it ?

So based on these, I prefer not to add more bindings to handle such
deviations but make it just AmLogic specific.

-- 
Regards,
Sudeep

Re: [PATCH v3 1/2] Documentation: bindings: Introduce scpi,sensors-scale

[Guenter Roeck]

On Wed, Mar 15, 2017 at 06:20:33PM +0000, Sudeep Holla wrote:
> 
> 
> On 09/03/17 10:17, Carlo Caione wrote:
> > From: Carlo Caione <carlo@endlessm.com>
> > 
> > Document the new property `scpi,sensors-scale` used by the hwmon-scpi
> > driver to convert the sensor readings to the correct / expected scale.
> > 
> 
> I am fine with DT bindings if DT maintainers agree with having one but I
> would like to add couple of points to get their feedback:
> 
> 1. ARM recognized the drawbacks of SCPI which started informally with
>    ARM Ltd platforms and got adopted by few vendors. It's now being
>    worked on and a much better and scalable replacement protocol is
>    being discussed with vendors now. So future enhancement to this SCPI
>    protocol is unlikely.
> 
> 2. AmLogic has diverted from base protocol in many ways and most of them
>    are handled with just compatible so far without any need for
>    additional bindings. Can we avoid it ?
> 
> So based on these, I prefer not to add more bindings to handle such
> deviations but make it just AmLogic specific.
> 

I like the idea.

Guenter

Re: [PATCH v3 1/2] Documentation: bindings: Introduce scpi,sensors-scale

[Rob Herring]

On Wed, Mar 15, 2017 at 06:20:33PM +0000, Sudeep Holla wrote:
> 
> 
> On 09/03/17 10:17, Carlo Caione wrote:
> > From: Carlo Caione <carlo-6IF/jdPJHihWk0Htik3J/w@public.gmane.org>
> > 
> > Document the new property `scpi,sensors-scale` used by the hwmon-scpi
> > driver to convert the sensor readings to the correct / expected scale.
> > 
> 
> I am fine with DT bindings if DT maintainers agree with having one but I
> would like to add couple of points to get their feedback:
> 
> 1. ARM recognized the drawbacks of SCPI which started informally with
>    ARM Ltd platforms and got adopted by few vendors. It's now being
>    worked on and a much better and scalable replacement protocol is
>    being discussed with vendors now. So future enhancement to this SCPI
>    protocol is unlikely.
> 
> 2. AmLogic has diverted from base protocol in many ways and most of them
>    are handled with just compatible so far without any need for
>    additional bindings. Can we avoid it ?
> 
> So based on these, I prefer not to add more bindings to handle such
> deviations but make it just AmLogic specific.

Agreed.

Rob
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH v3 1/2] Documentation: bindings: Introduce scpi, sensors-scale

[Martin Blumenstingl]

On Wed, Mar 15, 2017 at 7:20 PM, Sudeep Holla <sudeep.holla@arm.com> wrote:
>
>
> On 09/03/17 10:17, Carlo Caione wrote:
>> From: Carlo Caione <carlo@endlessm.com>
>>
>> Document the new property `scpi,sensors-scale` used by the hwmon-scpi
>> driver to convert the sensor readings to the correct / expected scale.
>>
>
> I am fine with DT bindings if DT maintainers agree with having one but I
> would like to add couple of points to get their feedback:
>
> 1. ARM recognized the drawbacks of SCPI which started informally with
>    ARM Ltd platforms and got adopted by few vendors. It's now being
>    worked on and a much better and scalable replacement protocol is
>    being discussed with vendors now. So future enhancement to this SCPI
>    protocol is unlikely.
>
> 2. AmLogic has diverted from base protocol in many ways and most of them
>    are handled with just compatible so far without any need for
>    additional bindings. Can we avoid it ?
>
> So based on these, I prefer not to add more bindings to handle such
> deviations but make it just AmLogic specific.
could you please share your idea where this "Amlogic specific
implementation" fits best?
- in firmware/arm_scpi to keep a contract (for example: millicelsius)
to everyone who reads a temperature sensor for example
- or in hwmon/scpi-hwmon


Regards,
Martin

Re: [PATCH v3 1/2] Documentation: bindings: Introduce scpi,sensors-scale

[Guenter Roeck]

On Wed, Mar 15, 2017 at 09:56:09PM +0100, Martin Blumenstingl wrote:
> On Wed, Mar 15, 2017 at 7:20 PM, Sudeep Holla <sudeep.holla@arm.com> wrote:
> >
> >
> > On 09/03/17 10:17, Carlo Caione wrote:
> >> From: Carlo Caione <carlo@endlessm.com>
> >>
> >> Document the new property `scpi,sensors-scale` used by the hwmon-scpi
> >> driver to convert the sensor readings to the correct / expected scale.
> >>
> >
> > I am fine with DT bindings if DT maintainers agree with having one but I
> > would like to add couple of points to get their feedback:
> >
> > 1. ARM recognized the drawbacks of SCPI which started informally with
> >    ARM Ltd platforms and got adopted by few vendors. It's now being
> >    worked on and a much better and scalable replacement protocol is
> >    being discussed with vendors now. So future enhancement to this SCPI
> >    protocol is unlikely.
> >
> > 2. AmLogic has diverted from base protocol in many ways and most of them
> >    are handled with just compatible so far without any need for
> >    additional bindings. Can we avoid it ?
> >
> > So based on these, I prefer not to add more bindings to handle such
> > deviations but make it just AmLogic specific.
> could you please share your idea where this "Amlogic specific
> implementation" fits best?
> - in firmware/arm_scpi to keep a contract (for example: millicelsius)
> to everyone who reads a temperature sensor for example
> - or in hwmon/scpi-hwmon
> 

So far the assumption was that it would be in hwmon. I don't mind for it
to be elsewhere, but I would think that the presence or non-presence
of a devicetree property should not affect or change the implementation.

Thanks,
Guenter

[PATCH v3 2/2] hwmon: (scpi) Fix the scale of SCP sensor readings

[Carlo Caione]

From: Carlo Caione <carlo@endlessm.com>

The implementation details for SCPI seems to suggest that the sensor
readings must be reported by SCP using a well defined scale
(millidegree Celsius for temperature, millivolts for voltage,
milliamperes for current, microwatts for power and microjoules for
energy).

This is also important for the interaction with other subsystems: for
example both the thermal sub-system and the hwmon sysfs interface expect
the temperature expressed in millidegree Celsius.

Unfortunately since this behaviour is dependent on the firmware
implementation there are cases where the sensor readings are reported
using a different scale. For example in the Amlogic SoCs the
temperature is reported in degree and not millidegree Celsius.

To take into account this discrepancy and fixup the values reported by
SCP a new DT property `scpi,sensors-scale' is introduced and used in
this patch by the scpi-hwmon driver to convert the sensor readings to
the expected scale.

Acked-by: Punit Agrawal <punit.agrawal@arm.com>
Signed-off-by: Carlo Caione <carlo@endlessm.com>
---
 drivers/hwmon/scpi-hwmon.c | 34 ++++++++++++++++++++++++++++++++++
 1 file changed, 34 insertions(+)

diff --git a/drivers/hwmon/scpi-hwmon.c b/drivers/hwmon/scpi-hwmon.c
index 094f948..4a4b8a4 100644
--- a/drivers/hwmon/scpi-hwmon.c
+++ b/drivers/hwmon/scpi-hwmon.c
@@ -23,6 +23,7 @@
 #include <linux/thermal.h>
 
 struct sensor_data {
+	unsigned int scale;
 	struct scpi_sensor_info info;
 	struct device_attribute dev_attr_input;
 	struct device_attribute dev_attr_label;
@@ -44,6 +45,22 @@ struct scpi_sensors {
 	const struct attribute_group *groups[2];
 };
 
+static const u32 scpi_scale[] = {
+	[TEMPERATURE]	= 1000,		/* (millicelsius)	*/
+	[VOLTAGE]	= 1000,		/* (millivolts)		*/
+	[CURRENT]	= 1000,		/* (milliamperes)	*/
+	[POWER]		= 1000000,	/* (microwatts)		*/
+	[ENERGY]	= 1000000,	/* (microjoules)	*/
+};
+
+static void scpi_scale_reading(u64 *value, struct sensor_data *sensor)
+{
+	if (scpi_scale[sensor->info.class] != sensor->scale) {
+		*value *= scpi_scale[sensor->info.class];
+		do_div(*value, sensor->scale);
+	}
+}
+
 static int scpi_read_temp(void *dev, int *temp)
 {
 	struct scpi_thermal_zone *zone = dev;
@@ -57,6 +74,8 @@ static int scpi_read_temp(void *dev, int *temp)
 	if (ret)
 		return ret;
 
+	scpi_scale_reading(&value, sensor);
+
 	*temp = value;
 	return 0;
 }
@@ -77,6 +96,8 @@ scpi_show_sensor(struct device *dev, struct device_attribute *attr, char *buf)
 	if (ret)
 		return ret;
 
+	scpi_scale_reading(&value, sensor);
+
 	return sprintf(buf, "%llu\n", value);
 }
 
@@ -97,6 +118,7 @@ static struct thermal_zone_of_device_ops scpi_sensor_ops = {
 static int scpi_hwmon_probe(struct platform_device *pdev)
 {
 	u16 nr_sensors, i;
+	u32 scale[] = { 1000, 1000, 1000, 1000000, 1000000 };
 	int num_temp = 0, num_volt = 0, num_current = 0, num_power = 0;
 	int num_energy = 0;
 	struct scpi_ops *scpi_ops;
@@ -131,6 +153,16 @@ static int scpi_hwmon_probe(struct platform_device *pdev)
 
 	scpi_sensors->scpi_ops = scpi_ops;
 
+	of_property_read_u32_array(dev->of_node, "scpi,sensors-scale",
+				   scale, ARRAY_SIZE(scale));
+
+	for (i = 0; i < ARRAY_SIZE(scale); i++) {
+		if (!scale[i]) {
+			dev_err(dev, "%s: scale cannot be zero (%d)\n", __func__, i);
+			return -EINVAL;
+		}
+	}
+
 	for (i = 0, idx = 0; i < nr_sensors; i++) {
 		struct sensor_data *sensor = &scpi_sensors->data[idx];
 
@@ -178,6 +210,8 @@ static int scpi_hwmon_probe(struct platform_device *pdev)
 			continue;
 		}
 
+		sensor->scale = scale[sensor->info.class];
+
 		sensor->dev_attr_input.attr.mode = S_IRUGO;
 		sensor->dev_attr_input.show = scpi_show_sensor;
 		sensor->dev_attr_input.attr.name = sensor->input;
-- 
2.9.3