Re: [PATCH v2 1/2] gpio: tps65219: add GPIO support for TPS65219 PMIC

[jerome Neanne]

On 22/05/2023 13:18, Andy Shevchenko wrote:
> On Mon, May 22, 2023 at 10:47 AM jerome Neanne <jneanne@xxxxxxxxxxxx> wrote:
> > On 20/05/2023 11:44, andy.shevchenko@xxxxxxxxx wrote:
> > > Mon, May 15, 2023 at 05:36:46PM +0200, Bartosz Golaszewski kirjoitti:
> > > > On Thu, May 11, 2023 at 4:09 PM Jerome Neanne <jneanne@xxxxxxxxxxxx> wrote:
>
> ...
>
> > > > > +       gpio->gpio_chip = tps65219_gpio_chip;
> > > >
> > > > Aren't you getting any warnings here about dropping the 'const' from
> > > > the global structure?
> > >
> > > But this is a copy of the contents and not the simple pointer.
>
> I commented on Bart's question.
>
> > In many other places where this is done, the struct is declared like:
> >
> > static const struct gpio_chip template_chip = {
> >
> > After internal review, I changed this to:
> >
> > static const struct gpio_chip tps65219_gpio_chip = {
> >
> > This is because I didn't want to have this "template" that sounds to me
> > like "dummy". Maybe I misunderstood and this "template" was used on
> > purpose because this const struct is just copied once to initialize
> > tps65219_gpio->gpio_chip during probe.
> >
> > Introducing tps65219_gpio_chip name is maybe confusing with
> > tps65219_gpio struct.
> >
> > I think the const should not be a problem here but the naming I used
> > might be misleading. If you have a suggestion of what is a good practice
> > to make this piece of code clearer. I'll follow your suggestion (use
> > template? more_explicit name like ???).
>
> It's up to Bart.
Bart, should I keep the code like this or do you suggest a name change 
so that's it's more appealing?