Re: [PATCH v3 2/2] iio: adc: add Axiado SARADC driver

[Petar Stepanovic]

On 6/28/2026 1:07 AM, David Lechner wrote:
>> +#define AX_SARADC_CH(_index, _id)                                       \
>> +     {                                                               \
>> +             .type = IIO_VOLTAGE,                                    \
>> +             .indexed = 1,                                           \
>> +             .channel = (_index),                                    \
>> +             .info_mask_separate = BIT(IIO_CHAN_INFO_RAW),           \
>> +             .info_mask_shared_by_type = BIT(IIO_CHAN_INFO_SCALE),   \
>> +             .datasheet_name = (_id),                                \
> This could probably be:
>
>                 .datasheet_name = "adc" #_index,
>
> and avoid the need for _id.

Thanks for the review, David.
Yes, that makes sense. I will update this and remove the extra _id
argument.

>> +     }
>> +
>> +static const struct iio_chan_spec axiado_saradc_iio_channels[] = {
>> +     AX_SARADC_CH(0, "adc0"),   AX_SARADC_CH(1, "adc1"),
>> +     AX_SARADC_CH(2, "adc2"),   AX_SARADC_CH(3, "adc3"),
>> +     AX_SARADC_CH(4, "adc4"),   AX_SARADC_CH(5, "adc5"),
>> +     AX_SARADC_CH(6, "adc6"),   AX_SARADC_CH(7, "adc7"),
>> +     AX_SARADC_CH(8, "adc8"),   AX_SARADC_CH(9, "adc9"),
>> +     AX_SARADC_CH(10, "adc10"), AX_SARADC_CH(11, "adc11"),
>> +     AX_SARADC_CH(12, "adc12"), AX_SARADC_CH(13, "adc13"),
>> +     AX_SARADC_CH(14, "adc14"), AX_SARADC_CH(15, "adc15"),
> Two columns looks a bit odd.

I will also reformat the channel table to one entry per line.

>> +};
>> +
>> +static void axiado_saradc_disable(void *data)
>> +{
>> +     struct axiado_saradc *info = data;
>> +
>> +     writel(AX_SARADC_GLOBAL_CTRL_PD, info->regs + AX_SARADC_GLOBAL_CTRL_REG);
> People usual make read and write wrappers or use regmap to avoid having
> to write `info->regs + AX_SARADC_GLOBAL_CTRL_REG` so many times.

My understanding is that simple read/write wrappers are not always
preferred unless they provide additional value. Would switching the
driver to regmap be acceptable here to avoid repeating the base address
calculation?

Regards,
Petar