Re: [PATCH v2 2/2] iio: adc: ad9467: support write/read offset

["Nuno Sá" <noname.nuno@gmail.com>]

On Tue, 2025-12-02 at 17:01 +0200, Tomas Melin wrote:
> 
> 
> On 02/12/2025 16:11, Andy Shevchenko wrote:
> > On Tue, Dec 02, 2025 at 12:53:09PM +0000, Tomas Melin wrote:
> > > Support configuring output calibration value. Among the devices
> > > currently supported by this driver, this setting is specific to
> > > ad9434. The offset can be used to calibrate the output against
> > > a known input. The register is called offset, but the procedure
> > > is best mapped internally with calibbias operation.
> > 
> > ...
> > 
> > >  static const struct iio_chan_spec ad9434_channels[] = {
> > > -	AD9467_CHAN(0, BIT(IIO_CHAN_INFO_SCALE), 0, 12, 's'),
> > > +	{
> > > +		.type = IIO_VOLTAGE,
> > > +		.indexed = 1,
> > > +		.channel = 0,
> > > +		.info_mask_shared_by_type =
> > > +		BIT(IIO_CHAN_INFO_SCALE) |
> > > +		BIT(IIO_CHAN_INFO_SAMP_FREQ) |
> > > +		BIT(IIO_CHAN_INFO_CALIBBIAS),
> > 
> > Wrong indentation.
> 
> Can you please provide example of your preferred indentation for this
> particular case? This is used in several places around the code and
> seemed like one of the more readable.
> 
> > 
> > > +		.info_mask_shared_by_type_available =
> > > +		BIT(IIO_CHAN_INFO_SCALE) |
> > > +		BIT(IIO_CHAN_INFO_CALIBBIAS),
> > 
> > Ditto.
> > 
> > > +		.scan_index = 0,
> > > +		.scan_type = {
> > > +			.sign = 's',
> > > +			.realbits = 12,
> > > +			.storagebits = 16,
> > > +		},
> > > +	},
> > >  };
> > 
> > I'm not sure about macro-less approach here, I think that we want more
> > consistency and hence before doing this change probably we want to clean up
> > the existing macro, then split it to two, and add another one here based on
> > the low-level, which was split in the previous clean up.
> 
> As mentioned, this is only needed for a single channel, and since it is
> different than the other, it needs to be separated. Do You think we
> actually need another macro for this?
> 
> > 
> > ...
> > 
> > > +	return ad9467_spi_write(st, AN877_ADC_REG_TRANSFER,
> > > +				AN877_ADC_TRANSFER_SYNC);
> > 
> > I would make it one line, despite on being 85 characters long.
> > But it's up to you and maintainers.
> I would like to not fight against checkpatch here.
> 
> > 

AFAIK, Jonathan policy is that 80 column limit is still the preferred limit unless readability is
hurt. So I would say the line break here is up to the IIO policy.

- Nuno Sá