Re: [PATCH v2 12/13] iio: light: al3010: Implement regmap support

[Jonathan Cameron <jic23@kernel.org>]

On Sun, 30 Mar 2025 18:45:38 +0100
Jonathan Cameron <jic23@kernel.org> wrote:

> On Wed, 19 Mar 2025 21:59:51 +0100
> David Heidelberg via B4 Relay <devnull+david.ixit.cz@kernel.org> wrote:
> 
> > From: David Heidelberg <david@ixit.cz>
> > 
> > Modernize and make driver a bit cleaner.
> > 
> > After the regmap implementation, the compiler is able to produce
> > much smaller module.
> > 
> > Size before: 72 kB
> > Size after:  58 kB
> > 
> > Signed-off-by: David Heidelberg <david@ixit.cz>  
> Just one real comment on your code. 
> Also a passing musing that you can feel free to ignore!
> 
> Jonathan
> 
> 
> >  static void al3010_set_pwr_off(void *_data)
> >  {
> >  	struct al3010_data *data = _data;
> > +	struct device *dev = regmap_get_device(data->regmap);
> > +	int ret;
> >  
> > -	i2c_smbus_write_byte_data(data->client, AL3010_REG_SYSTEM,
> > -				  AL3010_CONFIG_DISABLE);
> > +	ret = regmap_write(data->regmap, AL3010_REG_SYSTEM, AL3010_CONFIG_DISABLE);
> > +	if (ret)
> > +		dev_err(dev, "failed to write system register\n");  
> 
> Unrelated to your patch, but I wonder if there is appetite for
> 		regmap_err(data->regmap, "failed to write system register\n");
> 
> A lot of drivers carry a local dev pointer because the regmap_get_device() stuff
> is a bit clunky. They only use it in many paths for prints like this.
> 
> >  }  
> 
> 
> > @@ -204,7 +210,9 @@ static int al3010_suspend(struct device *dev)
> >  
> >  static int al3010_resume(struct device *dev)
> >  {
> > -	return al3010_set_pwr_on(to_i2c_client(dev));
> > +	struct al3010_data *data = iio_priv(dev_get_drvdata(dev));  
> 
> Why this change?
Ah. Obvious.  Ignore me.
> 
> > +
> > +	return al3010_set_pwr_on(data);
> >  }
> >  
> >  static DEFINE_SIMPLE_DEV_PM_OPS(al3010_pm_ops, al3010_suspend, al3010_resume);
> >   
>