Re: [PATCH v5 1/2] iio: magnetometer: add support for Infineon TLV493D 3D Magentic sensor

[Jonathan Cameron <jic23@kernel.org>]

On Fri, 29 Aug 2025 08:23:42 +0530
Dixit Parmar <dixitparmar19@gmail.com> wrote:

> The Infineon TLV493D is a Low-Power 3D Magnetic Sensor. The Sensor
> applications includes joysticks, control elements (white goods,
> multifunction knops), or electric meters (anti tampering) and any
> other application that requires accurate angular measurements at
> low power consumptions.
> 
> The Sensor is configured over I2C, and as part of Sensor measurement
> data it provides 3-Axis magnetic fields and temperature core measurement.
> 
> The driver supports raw value read and buffered input via external trigger
> to allow streaming values with the same sensing timestamp.
> 
> While the sensor has an interrupt pin multiplexed with an I2C SCL pin.
> But for bus configurations interrupt(INT) is not recommended, unless timing
> constraints between I2C data transfers and interrupt pulses are monitored
> and aligned.
> 
> The Sensor's I2C register map and mode information is described in product
> User Manual [1].
> 
> Datasheet: https://www.infineon.com/assets/row/public/documents/24/49/infineon-tlv493d-a1b6-datasheet-en.pdf
> Link: https://www.mouser.com/pdfDocs/Infineon-TLV493D-A1B6_3DMagnetic-UserManual-v01_03-EN.pdf [1]
> Signed-off-by: Dixit Parmar <dixitparmar19@gmail.com>

Hi Dixit,

A few trivial formatting related things inline.  I'd just have tweaked these
whilst applying but we need to have a v6 for the dt-binding/maintainers stuff
anyway so please tidy these up as well.

Thanks,

Jonathan


> diff --git a/drivers/iio/magnetometer/Makefile b/drivers/iio/magnetometer/Makefile
> index 9297723a97d8..dfe970fcacb8 100644
> --- a/drivers/iio/magnetometer/Makefile
> +++ b/drivers/iio/magnetometer/Makefile
> @@ -23,6 +23,8 @@ st_magn-$(CONFIG_IIO_BUFFER) += st_magn_buffer.o
>  obj-$(CONFIG_IIO_ST_MAGN_I2C_3AXIS) += st_magn_i2c.o
>  obj-$(CONFIG_IIO_ST_MAGN_SPI_3AXIS) += st_magn_spi.o
>  
> +obj-$(CONFIG_INFINEON_TLV493D)		+= tlv493d.o
> +
>  obj-$(CONFIG_SENSORS_HMC5843)		+= hmc5843_core.o
>  obj-$(CONFIG_SENSORS_HMC5843_I2C)	+= hmc5843_i2c.o
>  obj-$(CONFIG_SENSORS_HMC5843_SPI)	+= hmc5843_spi.o
> diff --git a/drivers/iio/magnetometer/tlv493d.c b/drivers/iio/magnetometer/tlv493d.c
> new file mode 100644
> index 000000000000..b723eaac1d9e
> --- /dev/null
> +++ b/drivers/iio/magnetometer/tlv493d.c
> @@ -0,0 +1,533 @@
> +// SPDX-License-Identifier: GPL-2.0-only
> +/**

This isn't kernel doc format. Headers descriptions like this normally aren't
so /* not /** is appropriate. I haven't checked but I'd expect the kernel-doc
script to have warned about that if run on this file

> + * Driver for the Infineon TLV493D Low-Power 3D Magnetic Sensor
> + *
> + * Copyright (C) 2025 Dixit Parmar <dixitparmar19@gmail.com>
> + */
> +/*
> + * Different mode has different measurement sampling time, this time is
> + * used in deriving the sleep and timeout while reading the data from
> + * sensor in polling.
> + * Power-down mode: No measurement.
> + * Fast mode: Freq:3.3 KHz. Measurement time:305 usec.
> + * Low-power mode: Freq:100 Hz. Measurement time:10 msec.
> + * Ultra low-power mode: Freq:10 Hz. Measurement time:100 msec.
> + * Master controlled mode: Freq:3.3 Khz. Measurement time:305 usec.
> + */
> +static const u32 tlv493d_sample_rate_us[] = {
> +	[TLV493D_OP_MODE_POWERDOWN] = 0,
> +	[TLV493D_OP_MODE_FAST] = 305,
> +	[TLV493D_OP_MODE_LOWPOWER] = 10 * USEC_PER_MSEC,
> +	[TLV493D_OP_MODE_ULTRA_LOWPOWER] = 100 * USEC_PER_MSEC,
> +	[TLV493D_OP_MODE_MASTERCONTROLLED] = 305,
> +};
> +
> +static int tlv493d_write_all_regs(struct tlv493d_data *data)
> +{
> +	int ret;
> +	struct device *dev = &data->client->dev;
> +
> +	ret = i2c_master_send(data->client, data->wr_regs, ARRAY_SIZE(data->wr_regs));
> +	if (ret < 0) {
> +		dev_err(dev, "i2c write registers failed, error: %d\n", ret);
> +		return ret;
> +	}
> +
> +	return 0;
> +}


> +static int tlv493d_get_measurements(struct tlv493d_data *data, s16 *x, s16 *y,
> +				    s16 *z, s16 *t)
> +{
> +	u8 buff[7] = {};
> +	int err, ret;
> +	struct device *dev = &data->client->dev;
> +	u32 sleep_us = tlv493d_sample_rate_us[data->mode];
> +
> +	guard(mutex)(&data->lock);
> +
> +	ret = pm_runtime_resume_and_get(dev);
> +	if (ret < 0)
> +		return ret;
> +
> +	/*
> +	 * Poll until data is valid,

. rather than , given next line is a new sentence.

> +	 * For a valid data TLV493D_TEMP_CHANNEL bit of TLV493D_RD_REG_TEMP
> +	 * should be set to 0. The sampling time depends on the sensor mode.
> +	 * Poll 3x the time of the sampling time.
> +	 */
> +	ret = read_poll_timeout(i2c_master_recv, err,
> +			err || !FIELD_GET(TLV493D_TEMP_CHANNEL,
> +			buff[TLV493D_RD_REG_TEMP]),

If you are going to split the FIELD_GET between parameters you must
align the second line of parameters after the (  otherwise this is
unnecessarily hard to read.

	ret = read_poll_timeout(i2c_master_recv, err,
		err || !FIELD_GET(TLV493D_TEMP_CHANNEL,	buff[TLV493D_RD_REG_TEMP]),
		sleep_us, 3 * sleep_us, false, data->client, buff,
		ARRAY_SIZE(buff));

Is a better way to format this.  Once we aren't aligning with the opening
bracket of read_poll_timeout because of the very long line, a single tab of
indent of the second line is enough.  Don't worry about going a little over 80 chars
here as it really hurts readability to split that FIELD_PREP() up.


> +	if (ret) {
> +		dev_err(dev, "i2c read poll timeout, error:%d\n", ret);
> +		goto out_put_autosuspend;
> +	}
> +	if (err < 0) {
> +		dev_err(dev, "i2c read data failed, error:%d\n", err);
> +		ret = err;
> +		goto out_put_autosuspend;
> +	}
> +
> +	*x = tlv493d_get_channel_data(buff, TLV493D_AXIS_X);
> +	*y = tlv493d_get_channel_data(buff, TLV493D_AXIS_Y);
> +	*z = tlv493d_get_channel_data(buff, TLV493D_AXIS_Z);
> +	*t = tlv493d_get_channel_data(buff, TLV493D_TEMPERATURE);
> +
> +out_put_autosuspend:
> +	pm_runtime_put_autosuspend(dev);
> +	return ret;
> +}


> +static int tlv493d_read_raw(struct iio_dev *indio_dev,
> +			const struct iio_chan_spec *chan, int *val,
> +			int *val2, long mask)
> +{
> +	struct tlv493d_data *data = iio_priv(indio_dev);
> +	s16 x, y, z, t;
> +	int ret;
> +
> +	switch (mask) {
> +	case IIO_CHAN_INFO_RAW:
> +		ret = tlv493d_get_measurements(data, &x, &y, &z, &t);
> +		if (ret)
> +			return ret;
> +
> +		/* Return raw values for requested channel */

Not return in the C meaning of it, and that comment doesn't
really add anything, so I'd just drop it.

> +		switch (chan->address) {
> +		case TLV493D_AXIS_X:
> +			*val = x;
> +			return IIO_VAL_INT;
> +		case TLV493D_AXIS_Y:
> +			*val = y;
> +			return IIO_VAL_INT;
> +		case TLV493D_AXIS_Z:
> +			*val = z;
> +			return IIO_VAL_INT;
> +		case TLV493D_TEMPERATURE:
> +			*val = t;
> +			return IIO_VAL_INT;
> +		default:
> +			return -EINVAL;
> +		}


> +static int tlv493d_probe(struct i2c_client *client)
> +{

> +	indio_dev->info = &tlv493d_info;
> +	indio_dev->modes = INDIO_DIRECT_MODE;
> +	indio_dev->name = client->name;
> +	indio_dev->channels = tlv493d_channels;
> +	indio_dev->num_channels = ARRAY_SIZE(tlv493d_channels);
> +	indio_dev->available_scan_masks = tlv493d_scan_masks;
> +
> +	ret = devm_iio_triggered_buffer_setup(dev, indio_dev,
> +				iio_pollfunc_store_time,
> +				tlv493d_trigger_handler,
> +				NULL);

If you align later lines of parameters with the start of the 1st
paramater, e.g. after the ( then this is still under 80 chars.

Given that is generally the preferred style where possible (in IIO
anyway) please do that here.

Jonathan