Re: [PATCH v2 8/8] iio: adc: rzt2h: Add support for RZ/V2H(P) SoC

[Lad, Prabhakar]

Hi Andy,

Thank you for the review.


On Tue, Oct 13, 2026 at 4:16 PM Andy Shevchenko
<andriy.shevchenko@xxxxxxxxx> wrote:
>
> On Fri, Oct 09, 2026 at 07:41:22PM +0100, Prabhakar wrote:
>
> > Add support for the ADC found on the Renesas RZ/V2H(P) SoC, which shares
> > the RZ/T2H ADC IP but differs in channel count, sampling, calibration and
> > resolution support.
> >
> > The RZ/V2H(P) ADC provides 8 channels, has a fixed conversion time of
> > 20 ADCLK cycles, does not support calibration, and supports selectable
> > 12-bit and 8-bit conversion resolution.
>
> ...
>
> > +static int rzt2h_adc_write_scale(struct iio_dev *indio_dev,
> > +                              struct iio_chan_spec const *chan,
> > +                              int val, int val2)
> > +{
> > +     struct rzt2h_adc *adc = iio_priv(indio_dev);
> > +     unsigned int i;
> > +     u64 scale;
> > +     u32 rem;
> > +
> > +     if (!(adc->chip_info->features & RZT2H_ADC_FEATURE_RES_SELECT))
> > +             return -EINVAL;
> > +
> > +     for (i = 0; i < chan->num_ext_scan_type; i++) {
> > +             scale = (u64)RZT2H_ADC_VREF_mV * NANO >>
> > +                     chan->ext_scan_type[i].realbits;
> > +             if (div_u64_rem(scale, NANO, &rem) == val && rem == val2)
> > +                     break;
>
> Hmm... AFAICT we can use multiplication as the result of the division is just
> used to compare.
>
Agreed, I will drop it.

> > +     }
> > +     if (i == chan->num_ext_scan_type)
> > +             return -EINVAL;
> > +
> > +     /* The scan type changes with the resolution. */
> > +     IIO_DEV_ACQUIRE_DIRECT_MODE(indio_dev, claim);
> > +     if (IIO_DEV_ACQUIRE_FAILED(claim))
> > +             return -EBUSY;
> > +
> > +     adc->resolution = i;
> > +
> > +     return 0;
> > +}
>
> ...
>
> > +     ret = reset_control_assert(adc->rstc);
> > +     if (ret)
> > +             pwrseq_enable(adc->pwrseq);
> >
> > -     return reset_control_assert(adc->rstc);
> > +     return ret;
>
> Hmm... If it was added in the previous patches, add with unfolded check and
> explain in the comments that the code done this way to avoid churn in the
> future changes.
>
Ok while adding reset support I will check the return value to avoid this churn.

Cheers,
Prabhakar