On Sat, 3 Mar 2018 17:44:44 +0200
Andy Shevchenko <andy.shevche...@gmail.com> wrote:
> On Sat, Mar 3, 2018 at 5:37 PM, Jonathan Cameron <ji...@kernel.org> wrote:
> > On Wed, 28 Feb 2018 17:06:09 +0200
> >> On Wed, Feb 28, 2018 at 2:15 AM, Pierre Bourdon <delr...@google.com>
> >> wrote:
>
> Better to address even minors before submission.
Absolutely.
>
> >> > + if (itime <= 0 || itime > 255)
> >>
> >> Just side note: Suprisingly how many in_range() implementations we
> >> have in kernel...
> > I guess one of those things that is so simple it's not worth having
> > one true in_range to rule them all ;)
>
> We have already several implementations of the macro.
>
> >> > +static int bh1730_adjust_gain(struct bh1730_data *bh1730)
> >> > +{
> >> > + int visible, ir, highest, gain, ret, i;
> >>
> >> int visible, ir, highest, gain;
> >> unsigned int i;
> >
> > Is there a strong reason for this one that I'm missing?
> > (beyond personal taste!)
>
> First of all, I'm far from being fan of mixing int ret into other
> variable definitions.
>
> unsigned int i OTOH shows explicitly that we have counter which is not
> supposed to be negative.
Given it's specifically indexing over an enum (which can have any definition
it likes) I wouldn't normally care, but fair enough.
>
> int i in most of the cases will work, so, it's a minor. I'm not
> insisting, though having counter variable on separate line is also a
> good thing.
>
> In general, having different things in one line is a bad idea for my opinion.
Agreed.
>
> >> int ret;
>
