Re: [PATCH v1 3/7] power: reset: qcom-pon: Migrate to devm_spmi_subdevice_alloc_and_add()

[AngeloGioacchino Del Regno]

Il 21/07/25 19:14, Andy Shevchenko ha scritto:
> On Mon, Jul 21, 2025 at 03:05:46PM +0200, AngeloGioacchino Del Regno wrote:
> > Il 21/07/25 13:36, Andy Shevchenko ha scritto:
> > > On Mon, Jul 21, 2025 at 09:55:21AM +0200, AngeloGioacchino Del Regno wrote:
>
> ...
>
> > > > +	if (!pdev->dev.parent)
> > > > +		return -ENODEV;
> > >
> > > You can start using
> > >
> > > 	struct device *dev = &pdev->dev;
> > >
> > > here and perhaps one may convert the rest to it...
> > >
> > > ...
> > >
> > > >    	error = of_property_read_u32(pdev->dev.of_node, "reg",
> > >
> > > ...including, but not limited to, use of device_property_read_u32(dev, ...) here.
> >
> > I didn't do that for one single reason: I did not want to add noise to the commits
> > and wanted those to exclusively migrate the drivers to the new API, literally
> > without doing *anything* else unnecessary, even if I have located some almost
> > effortless improvements that I could've done to those drivers.
> >
> > Please - I prefer to keep it this way: these are the first commits that add the
> > usage of the new functions and of the concept of SPMI subdevices, and I really
> > want those to contain just that and nothing else - because I suspect that these
> > will be taken as example and will be read by the next person that is implementing
> > a new SPMI (sub)driver or converting any remaining ones to subdevice.
>
> You can introduce a temporary variable in this change and use it only in the
> lines you have added/touched. We have similar approach in several drivers.
> Then somebody (not specifically should be you) can move it forward.

Makes sense, okay!

Cheers,
Angelo