Re: [PATCH v8 10/11] iio: adc: ad7768-1: add filter type and oversampling ratio attributes

[Jonathan Cameron <jic23@kernel.org>]

On Mon, 19 May 2025 19:31:30 -0300
Jonathan Santos <jonath4nns@gmail.com> wrote:

> On 05/19, David Lechner wrote:
> > On 5/19/25 12:13 PM, Jonathan Santos wrote:  
> > > On 05/15, Jonathan Santos wrote:  
> > 
> > ...  
> > >> +
> > >> +/* Decimation Rate range for each filter type */
> > >> +static const int ad7768_dec_rate_range[][3] = {
> > >> +	[AD7768_FILTER_SINC5] = { 8, 8, 1024 },
> > >> +	[AD7768_FILTER_SINC3] = { 32, 32, 163840 },
> > >> +	[AD7768_FILTER_WIDEBAND] = { 32, 32, 1024 },
> > >> +	[AD7768_FILTER_SINC3_REJ60] = { 32, 32, 163840 },
> > >> +};
> > >> +  
> > > 
> > > Since we're still discussing some points — is the `step` in 
> > > `[min step max]` for the IIO range additive or multiplicative? It is not 
> > > clear on documentation, maybe on purpose or I have missed something.
> > > 
> > > Here, decimation/OSR doubles from 8 or 32 for SINC5/WIDEBAND, and is a 
> > > multiple of 32 for SINC3. So I'm still unsure how to represent this to be
> > > clear to the user.  

Ah.  Indeed. It is intended to be additive. Good catch.

> > 
> > Sounds to me like sinc5/wideband should be lists instead of ranges.
> > It is only 6 values.  
> 
> So can we mix lists and ranges for the same attribute? I see that this
> is not usual.

Should be fine to mix. Userspace should never be assuming one or the other anyway.
I guess maybe some pretty tool might change how it presents the data depending
on what it sees in the first entry?  Do we know of any tool this would trip up?

Jonathan

> 
> > 
> >