Re: [PATCH 2/4] iio: adc: Add support for the Renesas RZ/N1 ADC

[Herve Codina]

Hi Geert,

On Fri, 17 Oct 2025 09:40:42 +0200
Geert Uytterhoeven <geert@xxxxxxxxxxxxxx> wrote:

> Hi Hervé,
> 
> On Fri, 17 Oct 2025 at 09:37, Herve Codina <herve.codina@xxxxxxxxxxx> wrote:
> > Wolfram Sang <wsa+renesas@xxxxxxxxxxxxxxxxxxxx> wrote:  
> > > On Wed, Oct 15, 2025 at 04:28:14PM +0200, Herve Codina (Schneider Electric) wrote:  
> > > > +static void rzn1_adc_vc_setup_conversion(struct rzn1_adc *rzn1_adc, u32 ch,
> > > > +                                    int adc1_ch, int adc2_ch)
> > > > +{
> > > > +   u32 vc = 0;
> > > > +
> > > > +   if (adc1_ch != -1)
> > > > +           vc |= RZN1_ADC_VC_ADC1_ENABLE | RZN1_ADC_VC_ADC1_CHANNEL_SEL(adc1_ch);
> > > > +
> > > > +   if (adc2_ch != -1)
> > > > +           vc |= RZN1_ADC_VC_ADC2_ENABLE | RZN1_ADC_VC_ADC2_CHANNEL_SEL(adc2_ch);  
> > >
> > > Are you open to either use an errno (maybe EACCES) or define something
> > > custom (maybe RZN1_ADC_NO_CHANNEL) instead of hardcoded -1? I think I
> > > like the latter a tad more.  
> >
> > I prefer RZN1_ADC_NO_CHANNEL too instead of an error code and I will use
> > that instead of -1 in the next iteration.  
> 
> Or just -ENODEV or -ENOENT, and change the checks above to
> "if (adc1_ch >= 0)"?

I prefer to avoid mixing channel numbers with error code.

The specific RZN1_ADC_NO_CHANNEL value looks good to me meaning "No channel
to use".

Best regards,
Hervé