Re: [PATCH 01/11] iio: adc: Update bindings for ADC7 name used on QCOM PMICs

[Jonathan Cameron <Jonathan.Cameron@Huawei.com>]

On Mon, 23 Oct 2023 11:35:43 +0530
Jishnu Prakash <quic_jprakash@quicinc.com> wrote:

> Hi Jonathan,
> 
> Sorry for the late reply, I could not get back earlier as I got occupied 
> with other work till now. I have addressed your comments inline.
> 
> On 7/8/2023 8:28 PM, Jonathan Cameron wrote:
> > On Sat, 8 Jul 2023 12:58:25 +0530
> > Jishnu Prakash <quic_jprakash@quicinc.com> wrote:
> >  
> >> The name used initially for this version of Qualcomm Technologies, Inc.
> >> PMIC ADC was ADC7, following the convention of calling the PMIC generation
> >> PMIC7. However, the names were later amended internally to ADC5 Gen2 and
> >> PMIC5 Gen2. In addition, the latest PMIC generation now is known as
> >> PMIC5 Gen3 with ADC5 Gen3 supported on it. With this addition, it makes more
> >> sense to correct the name for this version of ADCs to ADC5 Gen2 from ADC7.
> >> Since this affects ADC devices across some PMICs, update the names accordingly.
> >>
> >> In order to avoid breaking the existing implementations of ADC7, add
> >> support for ADC5 Gen2 first now and remove the ADC7 support in a later
> >> patch.
> >>
> >> Signed-off-by: Jishnu Prakash <quic_jprakash@quicinc.com>  
> > Hi Jishnu.
> >
> > Whilst I can appreciate why you've picked this particular approach to
> > deal with the renames I'm not sure it's the smoothest path - or the
> > easiest to review.
> >
> > If doing a single patch for the complete rename was too much, perhaps
> > doing one header (or if it makes sense set of headers)
> > at a time would be easier to read?  With a final patch doing the compatible
> > addition.  Maybe let's see what other reviewers think though.  
> 
> 
> I don't completely understand what you mean here - but first let me 
> briefly recap what I was trying to do here.
> 
> In patches 1-5 of this series, I intended to update all existing support 
> for ADC7 by renaming it to ADC5 Gen2 to match the correct name used 
> internally. In addition, since I am adding support for ADC5 Gen3 in 
> patches 6 and 7, I thought it would make sense to rename this older 
> peripheral, to make it more obvious to everyone that this version lies 
> between ADC5 and ADC5 Gen3.
> 
> The patches were organized like  this:
> 
> Patch 1 - Update documentation to add gen2 compatible and update 
> examples(without removing older compatible). Add new binding files 
> equivalent to existing ADC7 files, just with macros and file names 
> updated to use "adc5_gen2" instead of "adc7"
> 
> Patch 2 - Update driver files to replace usage of "adc7" with "adc5 
> gen2", adding new compatible for adc5 gen2 without removing exsiting one 
> for adc7.
> 
> Patch 3 - Update compatible, macros and binding files included in all 
> devicetree files, based on the earlier two changes.
> 
> Patch 4 - Delete all instances of adc7 compatible from documentation 
> files. Delete all older binding files
> 
> Patch 5 - Delete the adc7 compatible from the driver
> 
> 
> Based on the comments I got, I understand I cannot proceed as such with 
> patches 4 and 5, I can amend/drop them. But to get back to your above 
> point about my overall approach, how exactly would you like me to 
> structure my patch series?
> 
> Should I make one big patch for documentation, bindings, driver and 
> devicetree changes where I update the naming and deprecate adc7 usage? 
> This may be straightforward but also hard to review.
> 
> 
> Or a patch series like this:
> 
> One patch to update documentation
> 
> One patch to update the bindings (headers) (Or one patch per header file?)
> 
> One patch to update driver file (adding new compatible and comment to 
> deprecate old one)
> 
> One patch to update all devicetree files (or separate patches?)

It must remain buildable at all times.  That can either be done by
duplicating everything, or by pushing through a patch that performs
all renames (maybe excluding bindings as we care less about that).
The all renames in single patch is a lot easier to review as can
see both sides of the change in a single patch.

Breaking that up into sets of renames will keep it manageable.

Jonathan

> 
> Please let me know what you think.
> 
> > A few other comments inline,
> >
> > Jonathan
> >
> >  
> >>   
> >>   properties:
> >> @@ -27,6 +27,7 @@ properties:
> >>             - qcom,spmi-adc5
> >>             - qcom,spmi-adc-rev2
> >>             - qcom,spmi-adc7
> >> +          - qcom,spmi-adc5-gen2  
> > Alphabetical order (roughly given currently list). So I'd stick
> > this after qcom,spmi-adc5  
> 
> 
> Will reorder them in the next patchset.
> 
> 
> >>   
> >>     reg:
> >>       description: VADC base address in the SPMI PMIC register map
> >> @@ -71,7 +72,7 @@ patternProperties:
> >>           description: |
> >>             ADC channel number.
> >>             See include/dt-bindings/iio/qcom,spmi-vadc.h
> >> -          For PMIC7 ADC, the channel numbers are specified separately per PMIC
> >> +          For PMIC5 Gen2 ADC, the channel numbers are specified separately per PMIC
> >>             in the PMIC-specific files in include/dt-bindings/iio/.
> >>   
> >>         label:
> >> @@ -114,7 +115,7 @@ patternProperties:
> >>                 channel calibration. If property is not found, channel will be
> >>                 calibrated with 0.625V and 1.25V reference channels, also
> >>                 known as absolute calibration.
> >> -            - For compatible property "qcom,spmi-adc5", "qcom,spmi-adc7" and
> >> +            - For compatible property "qcom,spmi-adc5", "qcom,spmi-adc5-gen2" and
> >>                 "qcom,spmi-adc-rev2", if this property is specified VADC will use
> >>                 the VDD reference (1.875V) and GND for channel calibration. If
> >>                 property is not found, channel will be calibrated with 0V and 1.25V
> >> @@ -213,7 +214,9 @@ allOf:
> >>         properties:
> >>           compatible:
> >>             contains:
> >> -            const: qcom,spmi-adc7
> >> +            enum :
> >> +                - qcom,spmi-adc7  
> > There is a deprecated marking for dt-bindings. Might be good to use it here.  
> 
> 
> Thanks for your suggestion, I'll do this in the next patchset.
> 
> 
> >  
> >> +                - qcom,spmi-adc5-gen2
> >>   
> >>       then:  
> 
> Thanks,
> 
> Jishnu
> 
> >>