Re: [PATCH v2 1/5] iio: imu: inv_icm42600: Simplify pm_runtime setup

[Sean Nyekjaer]

On Sat, Aug 09, 2025 at 10:27:52PM +0100, Andy Shevchenko wrote:
> On Sat, Aug 9, 2025 at 8:06 PM Jonathan Cameron <jic23@xxxxxxxxxx> wrote:
> > On Fri, 8 Aug 2025 23:37:51 +0200
> > Andy Shevchenko <andy.shevchenko@xxxxxxxxx> wrote:
> > > On Fri, Aug 8, 2025 at 5:58 PM Sean Nyekjaer <sean@xxxxxxxxxx> wrote:
> 
> ...
> 
> > > > +       struct device *dev = regmap_get_device(st->map);
> > > >
> > > > +       if (!pm_runtime_status_suspended(dev))
> > > > +               regulator_disable(st->vddio_supply);
> > >
> > > I would rather use positive conditional as it seems to me more scalable
> >
> > To potentially save time when Sean looks at this.  I don't follow. Do you mean
> > something like
> >         if (pm_runtime_status_suspended(dev))
> >                 return;
> >
> >         regulator_disable(st->vddio_supply);
> >
> > ?
> 
> Yes.
> 
> > If so I'm not seeing why we'd want this to scale as it's a single use
> > devm_set_action_or_reset() callback doing just one thing.
> 
> While I agree in _this_ case, in general the check and return
> immediately is more scalable for reading purposes, e.g., indentation
> will be one level less. Also it won't require additional churn in
> adding {, i.e. changing conditional line just for that.
> 

I like the return early if pm_runtime_status_suspended() is true.

Andy, when doing reviews please keep the function name, as it's much
easier to add the changes.

Jonathan, do we think checking pm_runtime_status_suspended() is a viable
option? Should we ask on the linux-pm list?

If it's ok; I will patch:
drivers/iio/adc/ti-ads1100.c
drivers/iio/pressure/bmp280-core.c
drivers/iio/pressure/icp10100.c

/Sean