Re: [PATCH 5/7] iio: adc: rockchip_saradc: Use dev_err_probe

[Jonathan Cameron <jic23@kernel.org>]

On Wed, 17 May 2023 12:40:37 +0200
AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com> wrote:

> Il 17/05/23 01:00, Shreeya Patel ha scritto:
> > Use dev_err_probe instead of dev_err in probe function,
> > which simplifies code a little bit and prints the error
> > code.
> > 
> > Signed-off-by: Shreeya Patel <shreeya.patel@collabora.com>
> > ---
> >   drivers/iio/adc/rockchip_saradc.c | 45 ++++++++++++++-----------------
> >   1 file changed, 20 insertions(+), 25 deletions(-)
> > 
> > diff --git a/drivers/iio/adc/rockchip_saradc.c b/drivers/iio/adc/rockchip_saradc.c
> > index 5e1e8575bc76..a52021fd477d 100644
> > --- a/drivers/iio/adc/rockchip_saradc.c
> > +++ b/drivers/iio/adc/rockchip_saradc.c  
> 
> ..snip..
> 
> > @@ -494,23 +492,20 @@ static int rockchip_saradc_probe(struct platform_device *pdev)
> >   	 * This may become user-configurable in the future.
> >   	 */
> >   	ret = clk_set_rate(info->clk, info->data->clk_rate);
> > -	if (ret < 0) {
> > -		dev_err(&pdev->dev, "failed to set adc clk rate, %d\n", ret);
> > -		return ret;
> > -	}
> > +	if (ret < 0)
> > +		return dev_err_probe(&pdev->dev, ret,
> > +				     "failed to set adc clk rate\n");
> >   
> >   	ret = regulator_enable(info->vref);
> > -	if (ret < 0) {
> > -		dev_err(&pdev->dev, "failed to enable vref regulator\n");
> > -		return ret;
> > -	}
> > +	if (ret < 0)
> > +		return dev_err_probe(&pdev->dev, ret,
> > +				     "failed to enable vref regulator\n");
> > +
> >   	ret = devm_add_action_or_reset(&pdev->dev,
> >   				       rockchip_saradc_regulator_disable, info);
> > -	if (ret) {
> > -		dev_err(&pdev->dev, "failed to register devm action, %d\n",
> > -			ret);
> > -		return ret;
> > -	}
> > +	if (ret)
> > +		return dev_err_probe(&pdev->dev, ret,
> > +				     "failed to register devm action\n");  
> 
> It's not your fault - and it's about a pre-existing issue, but there's that:
> you're returning an error if devm_add_action_or_reset() fails (which is highly
> unlikely), but you're leaving the regulator enabled!
Which regulator?  The _or_reset() part means that the callback is called if
the devm_add_action_or_reset is going to return an error.  So in the
path you talk about
rockchip_saradc_regulator_disable() is called and disables vref

> 
> As for how to proceed here, I would suggest to fix this issue in a separated
> commit (before the dev_err_probe() conversion); it'd look like...
> 
> 	if (ret) {
> 		regulator_disable(info->vref);
> 		dev_err( .... blurb );
> 		return ret;
> 	}
> 
> and after the conversion it'd look like...
> 
> 	if (ret) {
> 		regulator_disable(info->vref);
> 		return dev_err_probe( ... blurb );
> 	}
> 
> Cheers,
> Angelo
> 
> >   
> >   	ret = regulator_get_voltage(info->vref);
> >   	if (ret < 0)  
>