Re: [PATCH v4 3/7] iio: frequency: adf41513: driver implementation

["Nuno Sá" <noname.nuno@gmail.com>]

On Wed, 2026-01-21 at 11:52 +0200, Andy Shevchenko wrote:
> On Wed, Jan 21, 2026 at 09:41:25AM +0000, Rodrigo Alencar wrote:
> > On 26/01/20 03:38PM, Andy Shevchenko wrote:
> > > On Tue, Jan 20, 2026 at 01:07:49PM +0000, Rodrigo Alencar wrote:
> > > > On 26/01/20 01:24PM, Andy Shevchenko wrote:
> > > > > On Tue, Jan 20, 2026 at 12:43 PM Rodrigo Alencar
> > > > > <455.rodrigo.alencar@gmail.com> wrote:
> > > > > > On 26/01/19 07:07PM, Andy Shevchenko wrote:
> > > > > > > On Mon, Jan 19, 2026 at 04:37:09PM +0000, Rodrigo Alencar wrote:
> > > > > > > > On 26/01/19 03:42PM, Andy Shevchenko wrote:
> > > > > > > > > On Mon, Jan 19, 2026 at 11:21:59AM +0000, Rodrigo Alencar wrote:
> > > > > > > > > > On 26/01/19 09:31AM, Andy Shevchenko wrote:
> > > > > > > > > > > On Fri, Jan 16, 2026 at 02:32:22PM +0000, Rodrigo Alencar via B4 Relay wrote:
> 
> ...
> 
> > > > > > > > > > The current implementation is kind of a stripped version of
> > > > > > > > > > __iio_str_to_fixpoint(). Would you prefer something like this, then?:
> > > > > > > > > 
> > > > > > > > > Do they have most of the parts in common? If so, why can't we use
> > > > > > > > > __iio_str_to_fixpoint() directly? Or why can't we slightly refactor
> > > > > > > > > that to give us the results we need here?
> > > > > > > > 
> > > > > > > > __iio_str_to_fixpoint() only parses "int" chunks, adf41513_parse_uhz
> > > > > > > > was modified to accomodate the u64 parsing removing unnecessary stuff.
> > > > > > > 
> > > > > > > But why? The fractional part most likely will be kept int (it's up to 10⁻⁹).
> > > > > > > The integer can be bigger than 10⁹?
> > > > > > 
> > > > > > Correct, integer part of the frequency value goes up to 26.5 GHz
> > > > > > (uint_max is approx 4.3 GHz). Also, with the dual modulus, the PLL can
> > > > > > achieve micro Hz resolution.
> > > > > 
> > > > > µHz is not a problem since it's up to nHz.
> > > > > So, the difference so far is the integer part that can be 64-bit.
> > > > > Again, can we factor out something to be used for this and for the
> > > > > __iio_str_to_fixpoint() cases?
> > > > 
> > > > I am not sure what you are suggesting,
> > > 
> > > To make changes to reuse the code.
> > > 
> > > > but I am avoiding changes to iio core at this point.
> > > 
> > > Why?
> > 
> > I understood that core changes would require more than one user
> > supporting the change.
> 
> At least one. And we have tons of them as the callers of
> __iio_str_to_fixpoint() are not going to disappear. Basically it's a surgery in
> the middle of the existing chain of APIs. To me one user is enough justification
> for such a surgery. For the newly introduced API (imagine __iio_str_to_fixpoint()
> as an example) it's indeed one user not enough.
> 
> > > > If any other user needs similar behavior,
> > > > I'd say we would need to have __iio_str_to_fixpoint() implementation
> > > > modified, so to create a version of iio_str_to_fixpoint() that handles
> > > > long long variables. Possibly consuming simple_strtoull instead of
> > > > doing the manual parsing.
> > > 
> > > That's the problem here. With Yet Another Cool Parser this all becomes
> > > unmaintainable very soon
> > 
> > Considering that the need for a new parser for 64-bit parts is only driven
> > by this specific PLL driver, I wonder how things become that unmaintainable.
> 
> Because there is a duplication of the code (to some extent) and if we found
> a bug in the one implementation it will be hard to fix (or even remeber) about
> the other.

Agreed!

> 
> > > (basically as you said when new comer needs a third
> > > variant of it). This is not good. Instead better to create (amend, expand)
> > > existing test cases, split out a foundation API that parses 64-bit parts
> > > (maybe even for fractional as well, dunno) and evolve a needed (sub)API
> > > from it.
> > 
> > I don't disagree with you though, I suppose I will need a green light to
> > move on with this?
> 
> Fine with me, let's gather opinions of David, Nuno, Jonathan, and others.


Fine with me.

- Nuno Sá