Re: [PATCH] iio/adc/nau7802: Improve unlocking of a mutex in nau7802_read_raw()

[Jonathan Cameron]

On Tue, 13 Mar 2018 20:56:25 +0100
SF Markus Elfring <elfring@xxxxxxxxxxxxxxxxxxxxx> wrote:

> From: Markus Elfring <elfring@xxxxxxxxxxxxxxxxxxxxx>
> Date: Tue, 13 Mar 2018 20:52:26 +0100
> 
> * Add a jump target so that a call of the function "mutex_unlock" is stored
>   only once in this function implementation.
> 
> * Replace two calls by goto statements.
> 
> This issue was detected by using the Coccinelle software.
> 
> Signed-off-by: Markus Elfring <elfring@xxxxxxxxxxxxxxxxxxxxx>
Similar to other cases, this non obvious error handling flow rather implies
that this block of code might work better as a utility function called from
the case statement.

> ---
>  drivers/iio/adc/nau7802.c | 15 +++++----------
>  1 file changed, 5 insertions(+), 10 deletions(-)
> 
> diff --git a/drivers/iio/adc/nau7802.c b/drivers/iio/adc/nau7802.c
> index 8997e74a8847..68d06a492760 100644
> --- a/drivers/iio/adc/nau7802.c
> +++ b/drivers/iio/adc/nau7802.c
> @@ -303,10 +303,8 @@ static int nau7802_read_raw(struct iio_dev *indio_dev,
>  		 *   - Channel 2 is value 1 in the CHS register
>  		 */
>  		ret = i2c_smbus_read_byte_data(st->client, NAU7802_REG_CTRL2);
> -		if (ret < 0) {
> -			mutex_unlock(&st->lock);
> -			return ret;
> -		}
> +		if (ret < 0)
> +			goto unlock;
>  
>  		if (((ret & NAU7802_CTRL2_CHS_BIT) && !chan->channel) ||
>  				(!(ret & NAU7802_CTRL2_CHS_BIT) &&
> @@ -316,18 +314,15 @@ static int nau7802_read_raw(struct iio_dev *indio_dev,
>  					NAU7802_REG_CTRL2,
>  					NAU7802_CTRL2_CHS(chan->channel) |
>  					NAU7802_CTRL2_CRS(st->sample_rate));
> -
> -			if (ret < 0) {
> -				mutex_unlock(&st->lock);
> -				return ret;
> -			}
> +			if (ret < 0)
> +				goto unlock;
>  		}
>  
>  		if (st->client->irq)
>  			ret = nau7802_read_irq(indio_dev, chan, val);
>  		else
>  			ret = nau7802_read_poll(indio_dev, chan, val);
> -
> +unlock:
>  		mutex_unlock(&st->lock);
blank line before simple returns.  It isn't a major enough part to be
worth the noise of adding them to existing code, but there is no reason
to take them out!

>  		return ret;
>