Re: [PATCH v5 3/3] dt-bindings: leds: Convert pwm to yaml

[Alexander Dahl]

Hello Rob,

Am Dienstag, 22. September 2020, 17:42:58 CEST schrieb Rob Herring:
> On Sat, 19 Sep 2020 07:31:45 +0200, Alexander Dahl wrote:
> > The example was adapted slightly to make use of the 'function' and
> > 'color' properties.  License discussed with the original author.
> > 
> > Suggested-by: Jacek Anaszewski <jacek.anaszewski@xxxxxxxxx>
> > Signed-off-by: Alexander Dahl <post@xxxxxxxxxxx>
> > Cc: Peter Ujfalusi <peter.ujfalusi@xxxxxx>
> > ---
> > 
> > Notes:
> >     v4 -> v5:
> >       * updated based on feedback by Rob Herring
> >       * removed Acked-by
> >     
> >     v3 -> v4:
> >       * added Cc to original author of the binding
> >     
> >     v2 -> v3:
> >       * changed license identifier to recommended one
> >       * added Acked-by
> >     
> >     v2:
> >       * added this patch to series (Suggested-by: Jacek Anaszewski)
> >  
> >  .../devicetree/bindings/leds/leds-pwm.txt     | 50 -----------
> >  .../devicetree/bindings/leds/leds-pwm.yaml    | 82 +++++++++++++++++++
> >  2 files changed, 82 insertions(+), 50 deletions(-)
> >  delete mode 100644 Documentation/devicetree/bindings/leds/leds-pwm.txt
> >  create mode 100644 Documentation/devicetree/bindings/leds/leds-pwm.yaml
> 
> My bot found errors running 'make dt_binding_check' on your patch:
> 
> /builds/robherring/linux-dt-review/Documentation/devicetree/bindings/mfd/iqs
> 62x.example.dt.yaml: pwmleds: 'panel' does not match any of the regexes:
> '^led(-[0-9a-f]+)?$', 'pinctrl-[0-9]+' From schema:
> /builds/robherring/linux-dt-review/Documentation/devicetree/bindings/leds/l
> eds-pwm.yaml

I somehow expected errors on those checks, because I got actually two 
different recommendations from you:

In feedback on v4 of this patch (series) you recommended '^led(-[0-9a-f]+)?$' 
for the (pwm) led node name, which I used in v5.  Or just allow any node name 
with ".*" like in gpio-keys.yaml …

I just checked all in-tree dts files using "pwm-leds" and each also defines 
the "label" property, so renaming those nodes should not alter the paths in 
sysfs, if I understood everything correctly.  So I see two options now:

1) Go with the stricter check and fix all failing dts files and examples.

2) Just use the very loose check.

If 1), which patch would go first, renaming nodes in dts and examples or 
converting bindings to yaml enabling the stricter check?

> Please check and re-submit.

Will do, maybe I split the patch series and send both remaining patches 
separately?

Alex