Re: [PATCH v2 1/2] iio: adc: ti-ads7924: add Texas Instruments ADS7924 driver

[Hugo Villeneuve]

On Thu, 12 Jan 2023 19:00:30 +0000
Jonathan Cameron <jic23@xxxxxxxxxx> wrote:

> On Tue, 10 Jan 2023 11:01:23 -0500
> Hugo Villeneuve <hugo@xxxxxxxxxxx> wrote:
> 
> > From: Hugo Villeneuve <hvilleneuve@xxxxxxxxxxxx>
> > 
> > The Texas Instruments ADS7924 is a 4 channels, 12-bit analog to
> > digital converter (ADC) with an I2C interface.
> > 
> > Datasheet: https://www.ti.com/lit/gpn/ads7924
> > 
> > Signed-off-by: Hugo Villeneuve <hvilleneuve@xxxxxxxxxxxx>
> 
> Hi Hugo,
> 
> A few minor additional comments from me as you are doing a v3
> (I might just have tidied them up whilst applying otherwise to save
> us all time!)

No problem, I will add them to my v3.

> Thanks,
> 
> Jonathan
> 
> > diff --git a/drivers/iio/adc/ti-ads7924.c b/drivers/iio/adc/ti-ads7924.c
> > new file mode 100644
> > index 000000000000..8384f3451236
> > --- /dev/null
> > +++ b/drivers/iio/adc/ti-ads7924.c
> > @@ -0,0 +1,492 @@
> > +// SPDX-License-Identifier: GPL-2.0
> > +/*
> > + * IIO driver for Texas Instruments ADS7924 ADC, 12-bit, 4-Channels, I2C
> > + *
> > + * iio/adc/ti-ads7924.c
> 
> Not a huge amount of point in having a file name in the file.  I don't mind
> if you really want it though.

Removed.

> 
> > + * Author: Hugo Villeneuve <hvilleneuve@xxxxxxxxxxxx>
> > + * Copyright 2022 DimOnOff
> > + *
> > + * based on iio/adc/ti-ads1015.c
> > + * Copyright (c) 2016, Intel Corporation.
> > + *
> > + * Datasheet: https://www.ti.com/lit/gpn/ads7924
> > + */
> 
> > +
> > +#define ADS7924_TOTAL_CONVTIME_US (ADS7924_PWRUPTIME_US + ADS7924_ACQTIME_US + \
> > +				   ADS7924_CONVTIME_US)
> > +
> > +#define ADS7924_SLEEP_DELAY_MS		2000
> 
> Unused, so drop for now.

Removed.

> > +static bool ads7924_is_writeable_reg(struct device *dev, unsigned int reg)
> > +{
> > +	switch (reg) {
> > +	case ADS7924_MODECNTRL_REG:
> > +	case ADS7924_INTCNTRL_REG:
> > +	case ADS7924_ULR0_REG:
> > +	case ADS7924_LLR0_REG:
> > +	case ADS7924_ULR1_REG:
> > +	case ADS7924_LLR1_REG:
> > +	case ADS7924_ULR2_REG:
> > +	case ADS7924_LLR2_REG:
> > +	case ADS7924_ULR3_REG:
> > +	case ADS7924_LLR3_REG:
> > +	case ADS7924_INTCONFIG_REG:
> > +	case ADS7924_SLPCONFIG_REG:
> > +	case ADS7924_ACQCONFIG_REG:
> > +	case ADS7924_PWRCONFIG_REG:
> > +	case ADS7924_RESET_REG:
> > +		return true;
> > +	default:
> > +		return false;
> > +	}
> > +}
> > +
> > +static const struct regmap_config ads7924_regmap_config = {
> > +	.reg_bits = 8,
> > +	.val_bits = 8,
> > +	.max_register = ADS7924_RESET_REG,
> > +	.writeable_reg = ads7924_is_writeable_reg,
> > +};
> > +
> > +static const struct iio_chan_spec ads7924_channels[] = {
> > +	ADS7924_V_CHAN(0, ADS7924_DATA0_U_REG),
> > +	ADS7924_V_CHAN(1, ADS7924_DATA1_U_REG),
> > +	ADS7924_V_CHAN(2, ADS7924_DATA2_U_REG),
> > +	ADS7924_V_CHAN(3, ADS7924_DATA3_U_REG),
> > +};
> > +
> > +static
> > +int ads7924_get_adc_result(struct ads7924_data *data,
> 
> Drop the unusual line break as not needed (probably refactoring left over!) 
> static int ...

Done.

> > +			   struct iio_chan_spec const *chan, int *val)
> > +{
> > +	int ret;
> > +	__be16 be_val;
> > +
> > +	if (chan->channel < 0 || chan->channel >= ADS7924_CHANNELS)
> > +		return -EINVAL;
> > +
> > +	if (data->conv_invalid) {
> > +		int conv_time;
> > +
> > +		conv_time = ADS7924_TOTAL_CONVTIME_US;
> > +		/* Allow 10% for internal clock inaccuracy. */
> > +		conv_time += conv_time / 10;
> > +		usleep_range(conv_time, conv_time + 1);
> > +		data->conv_invalid = false;
> > +	}
> > +
> > +	ret = regmap_raw_read(data->regmap, ADS7924_AUTO_INCREMENT_BIT |
> > +			      chan->address, &be_val, sizeof(be_val));
> > +	if (ret)
> > +		return ret;
> > +
> > +	*val = be16_to_cpu(be_val);
> > +	*val = *val >> ADS7924_DATA_SHIFT;
> 	*val = be16_to_cpu(be_val) >> ADS7924_DATA_SHIFT;
> 
> 	return 0;

Yes, simpler and clearer. Done

> 
> > +
> > +	return ret;
> > +}
> > +
> > +static int ads7924_read_raw(struct iio_dev *indio_dev,
> > +			    struct iio_chan_spec const *chan, int *val,
> > +			    int *val2, long mask)
> > +{
> > +	int ret, vref_uv;
> > +	struct ads7924_data *data = iio_priv(indio_dev);
> > +
> > +	switch (mask) {
> > +	case IIO_CHAN_INFO_RAW: {
> > +		mutex_lock(&data->lock);
> > +		ret = ads7924_get_adc_result(data, chan, val);
> > +		mutex_unlock(&data->lock);
> > +		if (ret < 0)
> > +			dev_dbg(data->dev, "%s error1\n", __func__);
> > +		else
> > +			ret = IIO_VAL_INT;
> 		if (ret < 0) {
> 			... 
> 			return ret;
> 		}
> 
> 		return IIO_VAL_INT;
> 		

Done. I also have removed the not-very-informative "error1" message...

> > +		break;
> > +	}
> > +	case IIO_CHAN_INFO_SCALE:
> > +		vref_uv = regulator_get_voltage(data->vref_reg);
> > +		if (vref_uv < 0) {
> > +			/* dummy regulator "get_voltage" returns -EINVAL */
> > +			ret = -EINVAL;
> 			return -EINVAL;
> > +		} else {
> > +			*val =  vref_uv / 1000; /* Convert reg voltage to mV */
> > +			*val2 = ADS7924_BITS;
> > +			ret = IIO_VAL_FRACTIONAL_LOG2;
> 			return IIO_VAL_FR...
> 
> > +		}
> > +		break;
> > +	default:
> > +		ret = -EINVAL;
> 		return -EINVAL;
> > +		break;
> > +	}
> > +
> > +	return ret;
> > +}

Done. With these changes, I propose to also remove last "return ret" (like in rcar-gyroadc.c). Then, maybe also remove break statements?

> > +
> > +static const struct iio_info ads7924_info = {
> > +	.read_raw	= ads7924_read_raw,
> The tabbed indent doesn't add anything so 
> 	.read_raw = ads...
> 
> Aligning things is often a bad idea anyway as it just adds noise when inevitably
> the alignment needs increasing...  Definitely not worth doing for one item.

Ok, done.

> > +};
> > +
> > +static int ads7924_get_channels_config(struct i2c_client *client,
> > +				       struct iio_dev *indio_dev)
> > +{
> > +	struct ads7924_data *priv = iio_priv(indio_dev);
> > +	struct device *dev = priv->dev;
> > +	struct fwnode_handle *node;
> > +	int num_channels = 0;
> > +
> > +	device_for_each_child_node(dev, node) {
> > +		u32 pval;
> > +		unsigned int channel;
> > +
> > +		if (fwnode_property_read_u32(node, "reg", &pval)) {
> > +			dev_err(dev, "invalid reg on %pfw\n", node);
> > +			continue;
> > +		}
> > +
> > +		channel = pval;
> > +		if (channel >= ADS7924_CHANNELS) {
> > +			dev_err(dev, "invalid channel index %d on %pfw\n",
> > +				channel, node);
> > +			continue;
> > +		}
> > +
> > +		num_channels++;
> > +	}
> > +
> > +	if (num_channels > 0) {
> > +		dev_dbg(dev, "found %d ADC channels\n", num_channels);
> > +		return 0;
> > +	} else {
> 
> As per other review.  Give us what we expect which is error paths
> as out of line.

Already done as suggested by Christophe:

    if (num_channels <= 0)
        return -EINVAL;

    dev_dbg(dev, "found %d ADC channels\n", num_channels);
    return 0;

Although I do not fully understand what you mean by "...error paths as out of line"? Do you mean to drop the debug message?

> 
> > +		return -EINVAL;
> > +	}
> > +}
> 
> > +static int ads7924_reset(struct iio_dev *indio_dev)
> > +{
> > +	struct ads7924_data *data = iio_priv(indio_dev);
> > +
> > +	if (data->reset_gpio) {
> > +		gpiod_set_value(data->reset_gpio, 1); /* Assert. */
> > +		/* Educated guess: assert time not specified in datasheet... */
> > +		mdelay(100);
> > +		gpiod_set_value(data->reset_gpio, 0); /* Deassert. */
> 
> No real point in shared exit path. Early returns mean less code to look at
> when following particular flows which is always nice.

Agreed, done.


> 		return 0;
> 
> > +	} else {
> 
> if returning above, no need to have the else {

Done

> 
> > +		int ret;
> > +
> > +		/*
> > +		 * A write of 10101010 to this register will generate a
> > +		 * software reset of the ADS7924.
> > +		 */
> > +		ret = regmap_write(data->regmap, ADS7924_RESET_REG, 0b10101010);
> > +		if (ret)
> > +			return ret;
> 		return regmap_write(....)
> 
> > +	}
> > +
> > +	return 0;
> and drop this.

Done. also remove now unused variable "ret".

> > +};
> 
> > +static int ads7924_probe(struct i2c_client *client)
> > +{
> 
> ...
> 
> > +	ret = devm_add_action_or_reset(dev, ads7924_set_idle_mode, data);
> 
> The _or_reset() part of this is there so that if this function fails
> it will call the callback before returning an error code.  As a result
> you will turn the device, before you've ever turned it on.
> 
> Register the callback after turning it on, so ...

Makes sense, done.

> > +	if (ret)
> > +		return dev_err_probe(dev, ret,
> > +				     "failed to add idle mode action\n");
> > +
> > +	ret = ads7924_set_conv_mode(data, ADS7924_MODECNTRL_AUTO_SCAN);
> > +	if (ret)
> > +		return dev_err_probe(dev, ret,
> > +				     "failed to set conversion mode\n");
> 
> ... here.  

Thank you a lot for all your comments and meaningfull explanations!

-- 
Hugo Villeneuve <hugo@xxxxxxxxxxx>