Re: [PATCH 2/2] iio: light: opt3001: Add Support for opt3004 light sensor

[Jonathan Cameron <jic23@kernel.org>]

On Thu, 26 Dec 2024 06:14:59 +0000
Hardevsinh Palaniya <hardevsinh.palaniya@siliconsignals.io> wrote:

> Hi Andy
> 
> > On Wed, Dec 25, 2024 at 09:56:36AM +0000, Hardevsinh Palaniya wrote:  
> > > > On Tue, Dec 24, 2024 at 11:43:16AM +0530, Hardevsinh Palaniya wrote:  
> >  
> 
> ...
> 
> > > > > Add Support for OPT3004 Digital ambient light sensor (ALS) with
> > > > > increased angular IR rejection  
> > > >
> > > > Missing period here.  
> >  
> > > > > The OPT3004 sensor shares the same functionality and scale range as
> > > > > the OPT3001. This Adds the compatible string for OPT3004, enabling
> > > > > the driver to support it without any functional changes.
> > > > >
> > > > > Datasheet: https://www.ti.com/lit/gpn/opt3004  
> > > >  
> > > > >  
> > > >
> > > > This blank line is not needed.  
> > 
> > You left two above comments unanswered while Acking the rest, it's a bit confusing.
> > Are you agree on them or not?  
> 
> Apologies for overlooking those comments. They seemed straightforward, so I
> assumed your review was accurate, and I planned to address them directly in the
> next version without explicitly responding.
> 
> Regarding the second comment:
> The blank line was added to differentiate between the commit message and the
> SoB tag. Are you sure it should be removed?
> 
> ...
> 
> > > > >  static const struct of_device_id opt3001_of_match[] = {
> > > > >       { .compatible = "ti,opt3001", .data = &opt3001_chip_information },
> > > > >       { .compatible = "ti,opt3002", .data = &opt3002_chip_information },
> > > > > +     { .compatible = "ti,opt3004", .data = &opt3001_chip_information },
> > > > >       { }
> > > > >  };  
> > > >
> > > > I'm always puzzled why do we need a new compatible for the existing driver
> > > > data? Is this hardware has an additional feature that driver does not (yet)
> > > > implement?  
> > >
> > > OPT3001 and OPT3004 sensors are functionally identical, and there are no
> > > additional features in the OPT3004 that require separate handling in the driver.
> > >
> > > The new compatible string for the OPT3004 is being added, which will allow the
> > > driver to recognize and support this sensor in the same way it handles the OPT3001.  
> > But why? I understand if you put two compatible strings into the DT to make it
> > explicit in case of the future developments of the driver, but new compatible
> > in the driver makes only sense when you have either quirk(s) or feature(s) that
> > are different to the existing code. Since you haven't added either, what's the
> > point?  
> 
> Understood.
> 
> I also found a similar case with the ADXL346, which is identical to the ADXL345.
> In the mainline kernel, a compatible string was added as a fallback in the bindings
> but was not added to the driver itself.

There is a small difference. The ADXL346 at least has a different ID from the ADXL345.
The driver may sanity check that. It shouldn't reject an ID it doesn't recognise but
it may well print a message as sometimes this indicates a wrong DT.

The ADXL346 also has additional registers, so whilst it is backwards compatible with
the ADXL345 additional features may be support in future that need the distinction
to be in DT.

Jonathan


Jonathan

> 
> Thanks for the insight.
> 
> In the next version, I will drop this patch and only submit the bindings for the OPT3004.
> using the fallback mechanism.
> 
> Best Regards,
> Hardev