On Sat, 13 Apr 2024 18:11:50 +0300
Alisa-Dariana Roman <alisadariana@xxxxxxxxx> wrote:
> AINCOM should actually be a supply. If present and it has a non-zero
> voltage, the pseudo-differential channels are configured as single-ended
> with an offset. Otherwise, they are configured as differential channels
> between AINx and AINCOM pins.
>
> Signed-off-by: Alisa-Dariana Roman <alisa.roman@xxxxxxxxxx>
Hi Alisa,
Main comment here is that you can't modify the channel arrays in place
because there may be more that one instance of the hardware.
My preference would be to pick between static const iio_chan_spec
arrays depending on whether the aincom is provided or not.
Jonathan
> ---
> drivers/iio/adc/ad7192.c | 53 +++++++++++++++++++++++++++++++++++++---
> 1 file changed, 49 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/iio/adc/ad7192.c b/drivers/iio/adc/ad7192.c
> index ac737221beae..a9eb4fab39ca 100644
> --- a/drivers/iio/adc/ad7192.c
> +++ b/drivers/iio/adc/ad7192.c
> @@ -175,7 +175,7 @@ enum {
> struct ad7192_chip_info {
> unsigned int chip_id;
> const char *name;
> - const struct iio_chan_spec *channels;
> + struct iio_chan_spec *channels;
> u8 num_channels;
> const struct iio_info *info;
> };
> @@ -186,6 +186,7 @@ struct ad7192_state {
> struct regulator *vref;
> struct clk *mclk;
> u16 int_vref_mv;
> + u16 aincom_mv;
> u32 fclk;
> u32 mode;
> u32 conf;
> @@ -745,6 +746,9 @@ static int ad7192_read_raw(struct iio_dev *indio_dev,
> /* Kelvin to Celsius */
> if (chan->type == IIO_TEMP)
> *val -= 273 * ad7192_get_temp_scale(unipolar);
> + else if (st->aincom_mv && chan->channel2 == -1)
> + *val += DIV_ROUND_CLOSEST_ULL((u64)st->aincom_mv * 1000000000,
> + st->scale_avail[gain][1]);
> return IIO_VAL_INT;
> case IIO_CHAN_INFO_SAMP_FREQ:
> *val = DIV_ROUND_CLOSEST(ad7192_get_f_adc(st), 1024);
> @@ -982,7 +986,7 @@ static const struct iio_info ad7195_info = {
> #define AD7193_CHANNEL(_si, _channel1, _address) \
> AD7193_DIFF_CHANNEL(_si, _channel1, -1, _address)
>
> -static const struct iio_chan_spec ad7192_channels[] = {
> +static struct iio_chan_spec ad7192_channels[] = {
> AD719x_DIFF_CHANNEL(0, 1, 2, AD7192_CH_AIN1P_AIN2M),
> AD719x_DIFF_CHANNEL(1, 3, 4, AD7192_CH_AIN3P_AIN4M),
> AD719x_TEMP_CHANNEL(2, AD7192_CH_TEMP),
> @@ -994,7 +998,7 @@ static const struct iio_chan_spec ad7192_channels[] = {
> IIO_CHAN_SOFT_TIMESTAMP(8),
> };
>
> -static const struct iio_chan_spec ad7193_channels[] = {
> +static struct iio_chan_spec ad7193_channels[] = {
No to this.
Imagine you have 2 devices wired in parallel to the same host
but only one of them is using vincomm. Which setting you
get will depend on which order the probe happens in.
You need to take a copy of the channels array or have enough
variants that you can pick between static const arrays depending
on the setting. If that works here that is the preferred path to
take. Have ad7193_channels + ad7193_channels_with_vincomm or
something like that.
> AD7193_DIFF_CHANNEL(0, 1, 2, AD7193_CH_AIN1P_AIN2M),
> AD7193_DIFF_CHANNEL(1, 3, 4, AD7193_CH_AIN3P_AIN4M),
> AD7193_DIFF_CHANNEL(2, 5, 6, AD7193_CH_AIN5P_AIN6M),
> @@ -1048,11 +1052,27 @@ static void ad7192_reg_disable(void *reg)
> regulator_disable(reg);
> }
>
> +static int ad7192_config_channels(struct ad7192_state *st)
> +{
> + struct iio_chan_spec *channels = st->chip_info->channels;
> + int i;
> +
> + if (!st->aincom_mv)
> + for (i = 0; i < st->chip_info->num_channels; i++)
> + if (channels[i].channel2 == -1) {
> + channels[i].differential = 1;
> + channels[i].channel2 = 0;
> + }
> +
> + return 0;
> +}
> +
> static int ad7192_probe(struct spi_device *spi)
> {
> struct ad7192_state *st;
> struct iio_dev *indio_dev;
> - int ret;
> + struct regulator *aincom;
> + int ret = 0;
>
> if (!spi->irq) {
> dev_err(&spi->dev, "no IRQ?\n");
> @@ -1067,6 +1087,26 @@ static int ad7192_probe(struct spi_device *spi)
>
> mutex_init(&st->lock);
>
> + aincom = devm_regulator_get_optional(&spi->dev, "aincom");
> + if (!IS_ERR(aincom)) {
> + ret = regulator_enable(aincom);
> + if (ret) {
> + dev_err(&spi->dev, "Failed to enable specified AINCOM supply\n");
> + return ret;
> + }
> +
> + ret = devm_add_action_or_reset(&spi->dev, ad7192_reg_disable, aincom);
> + if (ret)
> + return ret;
> +
> + ret = regulator_get_voltage(aincom);
> + if (ret < 0)
> + return dev_err_probe(&spi->dev, ret,
> + "Device tree error, AINCOM voltage undefined\n");
> + }
> +
> + st->aincom_mv = ret / 1000;
I think I'd make the two paths more explicit
ret = regulator_get_voltage(aincom);
if (ret < 0)
return dev_err_probe(&spi->dev, ret,
"Device tree error, AINCOM voltage undefined\n");
st->aincom_mv = ret / 1000;
} else {
st->aincom_mv = 0;
}
or rely on default for st->aincom_mv already being zero and don't bother adding
the else, just move aincom_mv setting into the path with the voltage being read
and not the other path.
Setting it unconditionally by dividing 0 / 1000 is correct of course, but
not particularly obvious!
> +
> st->avdd = devm_regulator_get(&spi->dev, "avdd");
> if (IS_ERR(st->avdd))
> return PTR_ERR(st->avdd);
> @@ -1113,6 +1153,11 @@ static int ad7192_probe(struct spi_device *spi)
> st->int_vref_mv = ret / 1000;
>
> st->chip_info = spi_get_device_match_data(spi);
> +
> + ret = ad7192_config_channels(st);
> + if (ret)
> + return ret;
> +
> indio_dev->name = st->chip_info->name;
> indio_dev->modes = INDIO_DIRECT_MODE;
> indio_dev->channels = st->chip_info->channels;
