Re: [PATCH 5/5] iio:adc:lpc32xx Add scale feature

[Gregory CLEMENT <gregory.clement@bootlin.com>]

Hi Jonathan,
 
 On sam., févr. 09 2019, Jonathan Cameron <jic23@kernel.org> wrote:

> On Fri,  8 Feb 2019 17:09:44 +0100
> Gregory CLEMENT <gregory.clement@bootlin.com> wrote:
>
>> Until now this driver only exposed the raw value of the channels. With
>> this patch, the scale value is also exposed.
>> 
>> It depends of a regulator supply, and unlike most of the other driver, do
>> not having this regulator won't prevent to use the driver. The reason for
>> it is to allow to continue to use this driver with an old device tree. If
>> there is no regulator supply then the scale won't be exposed.
>> 
>> Signed-off-by: Gregory CLEMENT <gregory.clement@bootlin.com>
> Hi Gregory,
>
> A few minor comments inline.
>
> I'll admit to being surprised to see any patches at all for this driver given
> how long it has been since we had any known users.  We nearly dropped it as
> unused years ago IIRC!  Good thing we didn't it seems.
>
> Thanks,
>
> Jonathan
>
>> ---
>>  drivers/iio/adc/lpc32xx_adc.c | 27 +++++++++++++++++++++++----
>>  1 file changed, 23 insertions(+), 4 deletions(-)
>> 
>> diff --git a/drivers/iio/adc/lpc32xx_adc.c b/drivers/iio/adc/lpc32xx_adc.c
>> index f391c1e10136..e36ca307f065 100644
>> --- a/drivers/iio/adc/lpc32xx_adc.c
>> +++ b/drivers/iio/adc/lpc32xx_adc.c
>> @@ -14,6 +14,7 @@
>>  #include <linux/io.h>
>>  #include <linux/module.h>
>>  #include <linux/platform_device.h>
>> +#include <linux/regulator/consumer.h>
>>  
>>  /*
>>   * LPC32XX registers definitions
>> @@ -45,6 +46,7 @@ struct lpc32xx_adc_state {
>>  	void __iomem *adc_base;
>>  	struct clk *clk;
>>  	struct completion completion;
>> +	struct regulator *vref;
>>  
>>  	u32 value;
>>  };
>> @@ -57,7 +59,9 @@ static int lpc32xx_read_raw(struct iio_dev *indio_dev,
>>  {
>>  	struct lpc32xx_adc_state *st = iio_priv(indio_dev);
>>  	int ret;
>> -	if (mask == IIO_CHAN_INFO_RAW) {
>> +
>> +	switch (mask) {
>> +	case IIO_CHAN_INFO_RAW:
>>  		mutex_lock(&indio_dev->mlock);
>>  		ret = clk_prepare_enable(st->clk);
>>  		if (ret) {
>> @@ -77,6 +81,12 @@ static int lpc32xx_read_raw(struct iio_dev *indio_dev,
>>  		mutex_unlock(&indio_dev->mlock);
>>  
>>  		return IIO_VAL_INT;
>> +
>> +	case IIO_CHAN_INFO_SCALE:
>> +		*val = regulator_get_voltage(st->vref) / 1000;
>> +		*val2 =  chan->scan_type.realbits;
>> +
>> +		return IIO_VAL_FRACTIONAL_LOG2;
>
> Please add a default, otherwise we are going to get some compiler
> warnings that are irritating as it will think we 'forgot' to handle
> the other cases.

Sure, I will do it.

>
>>  	}
>>  
>>  	return -EINVAL;
>> @@ -93,9 +103,10 @@ static const struct iio_info lpc32xx_adc_iio_info = {
>>  	.info_mask_separate = BIT(IIO_CHAN_INFO_RAW),	\
>>  	.address = LPC32XXAD_IN * _index,		\
>>  	.scan_index = _index,				\
>> +	.scan_type.realbits = 10			\
>
> Given scan_type is only used in the core in buffered modes
> that this driver doesn't support this is a little 'unusual'.

I found it in other drivers and thought it was expected to store the bit
resolution here.

>
> Also, only one value, so why bother rather than just putting it
> in the one place it is used?

OK I will just use it in the lpc32xx_read_raw function then.

>
>>  }
>>  
>> -static const struct iio_chan_spec lpc32xx_adc_iio_channels[] = {
>> +static struct iio_chan_spec lpc32xx_adc_iio_channels[] = {
> Please provide two const versions and pick between them rather than
> modifying the one.
>
> Clearly we might 'know' there is only ever one such ADC on these devices
> but it's not obvious to a reviewer who isn't familiar with the hardware,
> making this look like a bug.

OK

>
>>  	LPC32XX_ADC_CHANNEL(0),
>>  	LPC32XX_ADC_CHANNEL(1),
>>  	LPC32XX_ADC_CHANNEL(2),
>> @@ -119,7 +130,7 @@ static int lpc32xx_adc_probe(struct platform_device *pdev)
>>  	struct resource *res;
>>  	int retval = -ENODEV;
>>  	struct iio_dev *iodev = NULL;
>> -	int irq;
>> +	int irq, i;
>>  
>>  	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
>>  	if (!res) {
>> @@ -159,6 +170,15 @@ static int lpc32xx_adc_probe(struct platform_device *pdev)
>>  		return retval;
>>  	}
>>  
>> +	st->vref = devm_regulator_get(&pdev->dev, "vref");
>> +	if (IS_ERR(st->vref))
>> +		dev_err(&pdev->dev,
>> +			"Missing vref regulator: No scaling available\n");
>> +	else
>> +		for (i = 0; i <  ARRAY_SIZE(lpc32xx_adc_iio_channels); i++)
>> +			lpc32xx_adc_iio_channels[i].info_mask_shared_by_type =
>> +				BIT(IIO_CHAN_INFO_SCALE);
>> +
>>  	platform_set_drvdata(pdev, iodev);
>>  
>>  	init_completion(&st->completion);
>> @@ -175,7 +195,6 @@ static int lpc32xx_adc_probe(struct platform_device *pdev)
>>  		return retval;
>>  
>>  	dev_info(&pdev->dev, "LPC32XX ADC driver loaded, IRQ %d\n", irq);
>> -
> Unrelated (and in my view) bad whitespace change.

Indeed, it was a mistake.

Thanks,

Gregory

>>  	return 0;
>>  }
>>  
>
>
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

-- 
Gregory Clement, Bootlin
Embedded Linux and Kernel engineering
http://bootlin.com

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel