Re: [PATCH v7 5/9] serial: sc16is7xx: fix regression with GPIO configuration

[Hugo Villeneuve]

On Sun, 4 Jun 2023 22:31:04 +0300
Andy Shevchenko <andy.shevchenko@xxxxxxxxx> wrote:

> On Sun, Jun 4, 2023 at 8:45 PM Hugo Villeneuve <hugo@xxxxxxxxxxx> wrote:
> >
> > On Sun, 4 Jun 2023 14:57:31 +0300
> > Andy Shevchenko <andy.shevchenko@xxxxxxxxx> wrote:
> >
> > > On Sun, Jun 4, 2023 at 10:47 AM Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx> wrote:
> > > > On Fri, Jun 02, 2023 at 11:26:21AM -0400, Hugo Villeneuve wrote:
> > >
> > > ...
> > >
> > > > > +static u8 sc16is7xx_setup_mctrl_ports(struct device *dev)
> > > >
> > > > This returns what, mctrl?  If so, please document that, it doesn't look
> > > > obvious.
> > >
> > > Good suggestion. Because I also stumbled over the returned type.
> > >
> > > >  And as the kernel test robot reported, you do nothing with the
> > > > return value so why compute it?
> > >
> > > It seems that the entire function and respective call has to be moved
> > > under #ifdef CONFIG_GPIOLIB.
> >
> > Hi,
> > it cannot. See my explanations in response to Greg's comments.
> 
> Then as Greg suggested, store in the structure and make this function
> to return an error code (with int), with this amendment you don't need
> to add a comment about the returned variable anymore.

Hi,
Yes, that is what I have done for V8. Simplifies/clean things a lot.

Hugo.