Re: [PATCH v2 3/4] hwmon: Add support for Amphenol ChipCap 2

[Guenter Roeck]

On 11/9/23 08:18, Krzysztof Kozlowski wrote:
> On 09/11/2023 15:55, Guenter Roeck wrote:
> > > > +	if (IS_ERR(data->hwmon)) {
> > > > +		ret = PTR_ERR(data->hwmon);
> > > > +		goto cleanup;
> > > > +	}
> > > > +
> > > > +	return 0;
> > > > +
> > > > +cleanup:
> > > > +	if (cc2_disable(data))
> > > > +		dev_dbg(dev, "Failed to disable device");
> > > > +
> > > > +	return dev_err_probe(dev, ret,
> > > > +			     "Unable to register hwmon device\n");
> > >
> > > Drop or move to each error path.
> > This actually follows Documentation/process/coding-style.rst, chapter 7
> > (Centralized exiting of functions).
>
> The point is that centralized message of error is useless. Probe failure
> is already handled by core, thus another message doing the same is
> redundant. What is needed to explain the true reason of failure, thus
> the error message should be next to each type of failure. Missing
> regulator? Say it. Missing clock? Say something else.

Agreed, you have a point.

Thanks,
Guenter