[PATCH] iio: accel: dmard09: Implement IIO_CHAN_INFO_SCALE

[PATCH] iio: accel: dmard09: Implement IIO_CHAN_INFO_SCALE

[Mert Seftali]

From: Mert Seftali <mertsftl@gmail.com>

in_accel_scale has returned -EINVAL ever since the driver was added.

The channels advertise scale via info_mask_shared_by_type, so the IIO
core exposes in_accel_scale, but dmard09_read_raw() only handles
IIO_CHAN_INFO_RAW; a SCALE read falls through to 'default: return
-EINVAL':

    $ cat .../iio:deviceX/in_accel_scale
    cat: in_accel_scale: Invalid argument

leaving userspace with raw counts it cannot convert to m/s^2.

The driver was written from a vendor source [1] without a datasheet, and
the scale was declared but never implemented. The vendor source carries
the sensitivity: its conversion is

    acc = raw * GRAVITY_EARTH_1000 / sensitivity   (then / 1000 -> m/s^2)

with sensitivity = 32 and GRAVITY_EARTH_1000 = 9807 ("about
(9.80665)*1000"), i.e. 32 counts correspond to 1 g.

That sensitivity applies to the value this driver already reports as raw.
The vendor processes each 16-bit sample as 'data >>= 3; data &= 0x1ff'
plus a sign-extend of bit 8 - a signed 9-bit quantity from register bits
[11:3] - and dmard09_read_raw()'s 'accel <<= 4; accel >>= 7' yields the
same value. It is self-consistent: 256 counts / 32 = 8 g full scale,
matching the part's +/-8g range.

Implement the scale derived from that sensitivity using standard gravity:

    scale = 9.80665 / 32 = 0.3064578125 m/s^2 per LSB

[1] https://github.com/minstrelsy/mediatek/blob/1f49d8c87b839651bc89afc870277e8e0f2e2d55/custom/common/kernel/accelerometer/dmard09/dmard09.c

Fixes: a4fa6509dda4 ("iio: accel: add support for the Domintech DMARD09 3-axis accelerometer")
Signed-off-by: Mert Seftali <mertsftl@gmail.com>
---
 drivers/iio/accel/dmard09.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/drivers/iio/accel/dmard09.c b/drivers/iio/accel/dmard09.c
index fe35a1270786..a61943120d07 100644
--- a/drivers/iio/accel/dmard09.c
+++ b/drivers/iio/accel/dmard09.c
@@ -27,6 +27,9 @@
 #define DMARD09_AXIS_Y_OFFSET ((DMARD09_AXIS_Y + 1) * 2)
 #define DMARD09_AXIS_Z_OFFSET ((DMARD09_AXIS_Z + 1) * 2)
 
+/* Sensitivity is 32 LSB/g; scale = 9.80665 / 32 m/s^2 per LSB. */
+#define DMARD09_SCALE_NANO	306457813
+
 struct dmard09_data {
 	struct i2c_client *client;
 };
@@ -79,6 +82,10 @@ static int dmard09_read_raw(struct iio_dev *indio_dev,
 		*val = accel;
 
 		return IIO_VAL_INT;
+	case IIO_CHAN_INFO_SCALE:
+		*val = 0;
+		*val2 = DMARD09_SCALE_NANO;
+		return IIO_VAL_INT_PLUS_NANO;
 	default:
 		return -EINVAL;
 	}
-- 
2.55.0

Re: [PATCH] iio: accel: dmard09: Implement IIO_CHAN_INFO_SCALE

[Andy Shevchenko]

On Fri, Jul 03, 2026 at 03:07:50PM +0200, Mert Seftali wrote:

> in_accel_scale has returned -EINVAL ever since the driver was added.

You mean "Reading the in_accel_scale attribute ends up in returning..." ?

> The channels advertise scale via info_mask_shared_by_type, so the IIO
> core exposes in_accel_scale, but dmard09_read_raw() only handles
> IIO_CHAN_INFO_RAW;

> a SCALE read falls through to 'default: return
> -EINVAL':

This part is repetition of the first paragraph. Drop one of them.

>     $ cat .../iio:deviceX/in_accel_scale
>     cat: in_accel_scale: Invalid argument
> 
> leaving userspace with raw counts it cannot convert to m/s^2.
> 
> The driver was written from a vendor source [1] without a datasheet, and
> the scale was declared but never implemented. The vendor source carries
> the sensitivity: its conversion is
> 
>     acc = raw * GRAVITY_EARTH_1000 / sensitivity   (then / 1000 -> m/s^2)
> 
> with sensitivity = 32 and GRAVITY_EARTH_1000 = 9807 ("about
> (9.80665)*1000"), i.e. 32 counts correspond to 1 g.
> 
> That sensitivity applies to the value this driver already reports as raw.
> The vendor processes each 16-bit sample as 'data >>= 3; data &= 0x1ff'

Try to write it more human and less programmish. Something like

"...16-bit sample as a signed 9-bit..." In other words, drop those technical
details, people can check the code by the link.

> plus a sign-extend of bit 8 - a signed 9-bit quantity from register bits
> [11:3] - and dmard09_read_raw()'s 'accel <<= 4; accel >>= 7' yields the

"...and in dmard09_read_raw() preparation yields the..."

> same value. It is self-consistent: 256 counts / 32 = 8 g full scale,
> matching the part's +/-8g range.
> 
> Implement the scale derived from that sensitivity using standard gravity:
> 
>     scale = 9.80665 / 32 = 0.3064578125 m/s^2 per LSB

> [1] https://github.com/minstrelsy/mediatek/blob/1f49d8c87b839651bc89afc870277e8e0f2e2d55/custom/common/kernel/accelerometer/dmard09/dmard09.c
> 

Make this a Link tag

Link: $URL [1]

> Fixes: a4fa6509dda4 ("iio: accel: add support for the Domintech DMARD09 3-axis accelerometer")
> Signed-off-by: Mert Seftali <mertsftl@gmail.com>

...

> +/* Sensitivity is 32 LSB/g; scale = 9.80665 / 32 m/s^2 per LSB. */
> +#define DMARD09_SCALE_NANO	306457813

I'm not sure how it's derived. Can we use more of calculations here instead of
precalculated value?

-- 
With Best Regards,
Andy Shevchenko

Re: [PATCH] iio: accel: dmard09: Implement IIO_CHAN_INFO_SCALE

[Joshua Crofts]

On Fri,  3 Jul 2026 15:07:50 +0200
Mert Seftali <mertsftl@gmail.com> wrote:
>  struct dmard09_data {
>  	struct i2c_client *client;
>  };
> @@ -79,6 +82,10 @@ static int dmard09_read_raw(struct iio_dev *indio_dev,
>  		*val = accel;
>  
>  		return IIO_VAL_INT;
> +	case IIO_CHAN_INFO_SCALE:
> +		*val = 0;
> +		*val2 = DMARD09_SCALE_NANO;

Add a blank line here - same as the previous case.

> +		return IIO_VAL_INT_PLUS_NANO;
>  	default:
>  		return -EINVAL;
>  	}



-- 
Kind regards

CJD