2019-08-08 15:45:51

by Himanshu Jha

Subject: [PATCH] iio: chemical: bme680: Fix pressure and humidity ABI mismatch

Standard ABI for reporting pressure is kilopascal and for
relative humidity it is millipercent.

What: /sys/bus/iio/devices/iio:deviceX/in_pressureY_input
What: /sys/bus/iio/devices/iio:deviceX/in_pressure_input
KernelVersion: 3.8
Contact: [email protected]
Description:
Scaled pressure measurement from channel Y, in kilopascal.

What: /sys/bus/iio/devices/iio:deviceX/in_humidityrelative_input
KernelVersion: 3.14
Contact: [email protected]
Description:
Scaled humidity measurement in milli percent.

Currently pressure is reported in hectopascal(hPa) and relative humidity
in percent. Hence fix this ABI mismatch conforming to the standard ABI.

Fixes: 1b3bd8592780 ("iio: chemical: Add support for Bosch BME680 sensor")
Signed-off-by: Himanshu Jha <[email protected]>
---

While cleaning this mess I wonder about the gas channel and there
exists no `in_resistance_input` in standard ABI :-(

We only have:

What: /sys/bus/iio/devices/iio:deviceX/in_resistance_raw
What: /sys/bus/iio/devices/iio:deviceX/in_resistanceX_raw
What: /sys/bus/iio/devices/iio:deviceX/out_resistance_raw
What: /sys/bus/iio/devices/iio:deviceX/out_resistanceX_raw
KernelVersion: 4.3
Contact: [email protected]
Description:
Raw (unscaled no offset etc.) resistance reading that can be processed
into an ohm value.

The sensor outputs processed value which is reported as is.

So, does it need a new ABI ?

drivers/iio/chemical/bme680_core.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/iio/chemical/bme680_core.c b/drivers/iio/chemical/bme680_core.c
index ccde4c65ff93..28cc96d1e3c8 100644
--- a/drivers/iio/chemical/bme680_core.c
+++ b/drivers/iio/chemical/bme680_core.c
@@ -670,7 +670,7 @@ static int bme680_read_press(struct bme680_data *data,
}

*val = bme680_compensate_press(data, adc_press);
- *val2 = 100;
+ *val2 = 1000;
return IIO_VAL_FRACTIONAL;
}

@@ -704,7 +704,7 @@ static int bme680_read_humid(struct bme680_data *data,
comp_humidity = bme680_compensate_humid(data, adc_humidity);

*val = comp_humidity;
- *val2 = 1000;
+ *val2 = 1000000;
return IIO_VAL_FRACTIONAL;
}

--
2.17.1

2019-08-11 08:33:41

by Jonathan Cameron

Subject: Re: [PATCH] iio: chemical: bme680: Fix pressure and humidity ABI mismatch

On Thu, 8 Aug 2019 21:13:50 +0530
Himanshu Jha <[email protected]> wrote:

> Standard ABI for reporting pressure is kilopascal and for
> relative humidity it is millipercent.
>
> What: /sys/bus/iio/devices/iio:deviceX/in_pressureY_input
> What: /sys/bus/iio/devices/iio:deviceX/in_pressure_input
> KernelVersion: 3.8
> Contact: [email protected]
> Description:
> Scaled pressure measurement from channel Y, in kilopascal.
>
> What: /sys/bus/iio/devices/iio:deviceX/in_humidityrelative_input
> KernelVersion: 3.14
> Contact: [email protected]
> Description:
> Scaled humidity measurement in milli percent.
>
> Currently pressure is reported in hectopascal(hPa) and relative humidity
> in percent. Hence fix this ABI mismatch conforming to the standard ABI.
>
> Fixes: 1b3bd8592780 ("iio: chemical: Add support for Bosch BME680 sensor")
> Signed-off-by: Himanshu Jha <[email protected]>

Hopefully any users will cope with their scripts getting broken
by this.

I'm going to let this one sit for a little longer to give others time
to take a look.

Give me a poke if it looks like I've lost it down the back of the sofa
in a week or two.

Thanks,

Jonathan


> ---
>
> While cleaning this mess I wonder about the gas channel and there
> exists no `in_resistance_input` in standard ABI :-(
>
> We only have:
>
> What: /sys/bus/iio/devices/iio:deviceX/in_resistance_raw
> What: /sys/bus/iio/devices/iio:deviceX/in_resistanceX_raw
> What: /sys/bus/iio/devices/iio:deviceX/out_resistance_raw
> What: /sys/bus/iio/devices/iio:deviceX/out_resistanceX_raw
> KernelVersion: 4.3
> Contact: [email protected]
> Description:
> Raw (unscaled no offset etc.) resistance reading that can be processed
> into an ohm value.
>
> The sensor outputs processed value which is reported as is.
>
> So, does it need a new ABI ?

New documentation as the ABI is clearly already there. Good for completeness
but that description for _raw makes it obvious what the units will be etc so
hopefully we don't have any disagreement between drivers.


>
> drivers/iio/chemical/bme680_core.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/iio/chemical/bme680_core.c b/drivers/iio/chemical/bme680_core.c
> index ccde4c65ff93..28cc96d1e3c8 100644
> --- a/drivers/iio/chemical/bme680_core.c
> +++ b/drivers/iio/chemical/bme680_core.c
> @@ -670,7 +670,7 @@ static int bme680_read_press(struct bme680_data *data,
> }
>
> *val = bme680_compensate_press(data, adc_press);
> - *val2 = 100;
> + *val2 = 1000;
> return IIO_VAL_FRACTIONAL;
> }
>
> @@ -704,7 +704,7 @@ static int bme680_read_humid(struct bme680_data *data,
> comp_humidity = bme680_compensate_humid(data, adc_humidity);
>
> *val = comp_humidity;
> - *val2 = 1000;
> + *val2 = 1000000;
> return IIO_VAL_FRACTIONAL;
> }
>