Re: [PATCH v2] iio: afe: iio-rescale: Support processed channels

[Peter Rosin <peda@axentia.se>]

Hi!

Thanks for the update!

On 2021-05-10 00:59, Linus Walleij wrote:
> It happens that an ADC will only provide raw or processed
> voltage conversion channels. (adc/ab8500-gpadc.c).
> On the Samsung GT-I9070 this is used for a light sensor
> and current sense amplifier so we need to think of something.
> 
> The idea is to allow processed channels and scale them
> with 1/1 and then the rescaler can modify the result
> on top.
> 
> Link: https://lore.kernel.org/linux-iio/20201101232211.1194304-1-linus.walleij@linaro.org/
> Cc: Peter Rosin <peda@axentia.se>
> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
> ---
> ChangeLog v1->v2:
> - Deny calls to .read_avail() for processed channels, and
>   insert a comment.
> - Move an assignment of IIO_VAL_FRACTIONAL down to the end
>   of the block for better readability.
> ---
>  drivers/iio/afe/iio-rescale.c | 39 +++++++++++++++++++++++++++++++----
>  1 file changed, 35 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/iio/afe/iio-rescale.c b/drivers/iio/afe/iio-rescale.c
> index e42ea2b1707d..ddbb760ae8df 100644
> --- a/drivers/iio/afe/iio-rescale.c
> +++ b/drivers/iio/afe/iio-rescale.c
> @@ -29,6 +29,7 @@ struct rescale {
>  	struct iio_channel *source;
>  	struct iio_chan_spec chan;
>  	struct iio_chan_spec_ext_info *ext_info;
> +	bool chan_processed;
>  	s32 numerator;
>  	s32 denominator;
>  };
> @@ -43,10 +44,27 @@ static int rescale_read_raw(struct iio_dev *indio_dev,
>  
>  	switch (mask) {
>  	case IIO_CHAN_INFO_RAW:
> -		return iio_read_channel_raw(rescale->source, val);
> +		if (rescale->chan_processed)
> +			/*
> +			 * When only processed channels are supported, we
> +			 * read the processed data and scale it by 1/1
> +			 * augmented with whatever the rescaler has calculated.
> +			 */
> +			return iio_read_channel_processed(rescale->source, val);
> +		else
> +			return iio_read_channel_raw(rescale->source, val);
>  
>  	case IIO_CHAN_INFO_SCALE:
> -		ret = iio_read_channel_scale(rescale->source, val, val2);
> +		if (rescale->chan_processed) {
> +			/*
> +			 * Processed channels are scaled 1-to-1
> +			 */
> +			*val = 1;
> +			*val2 = 1;
> +			ret = IIO_VAL_FRACTIONAL;
> +		} else {
> +			ret = iio_read_channel_scale(rescale->source, val, val2);
> +		}
>  		switch (ret) {
>  		case IIO_VAL_FRACTIONAL:
>  			*val *= rescale->numerator;
> @@ -82,6 +100,14 @@ static int rescale_read_avail(struct iio_dev *indio_dev,
>  
>  	switch (mask) {
>  	case IIO_CHAN_INFO_RAW:
> +		/*
> +		 * Using .read_avail() is fringe to begin with and makes
> +		 * no sense whatsoever for processed channels, so if called
> +		 * on a processed channel, we just error out here.
> +		 */
> +		if (rescale->chan_processed)
> +			return -EINVAL;
> +

I believe this is not the correct place to block requests for .read_avail.
It's too late; the driver should not first promise that it is possible to
call .read_avail just to later return -EINVAL.

It's better to just avoid adding the IIO_CHAN_INFO_RAW bit to
chan->info_mask_separate_available in the rescale_configure_channel
function. See some suggested -++ below at the end...

>  		*type = IIO_VAL_INT;
>  		return iio_read_avail_channel_raw(rescale->source,
>  						  vals, length);
> @@ -132,8 +158,13 @@ static int rescale_configure_channel(struct device *dev,
>  
>  	if (!iio_channel_has_info(schan, IIO_CHAN_INFO_RAW) ||
>  	    !iio_channel_has_info(schan, IIO_CHAN_INFO_SCALE)) {
> -		dev_err(dev, "source channel does not support raw/scale\n");
> -		return -EINVAL;
> +		if (iio_channel_has_info(schan, IIO_CHAN_INFO_PROCESSED)) {
> +			dev_info(dev, "using processed channel\n");
> +			rescale->chan_processed = true;
> +		} else {
> +			dev_err(dev, "source channel does not support raw+scale or processed data\n");
> +			return -EINVAL;
> +		}
>  	}
I'd also rather have this logic reversed/flattened:

	if (iio_channel_has_info(schan, IIO_CHAN_INFO_RAW) &&
	    iio_channel_has_info(schan, IIO_CHAN_INFO_SCALE)) {
		dev_info(dev, "using raw+scale source channel\n");
	} else if (iio_channel_has_info(schan, IIO_CHAN_INFO_PROCESSED)) {
		dev_info(dev, "using processed source channel\n");
		rescale->chan_processed = true;
	} else {
		dev_err(dev, "source channel is not supported\n");
		return -EINVAL;
	}

>  
>  	chan->info_mask_separate = BIT(IIO_CHAN_INFO_RAW) |
		BIT(IIO_CHAN_INFO_SCALE);
 
-	if (iio_channel_has_available(schan, IIO_CHAN_INFO_RAW))
+	if (iio_channel_has_available(schan, IIO_CHAN_INFO_RAW) &&
+	    !rescale->chan_processed)
 		chan->info_mask_separate_available |= BIT(IIO_CHAN_INFO_RAW);

(disclaimer, all new code untested and written directly in the mail client)

CHeers,
Peter