Re: [PATCHv2 2/2] iio: adc: add driver for the ti-adc084s021 chip

["Mårten Lindahl" <marten.lindahl@axis.com>]

On Sun, 2017-04-30 at 16:51 +0100, Jonathan Cameron wrote:
> On 30/04/17 13:53, Mårten Lindahl wrote:
> > From: Mårten Lindahl <martenli@axis.com>
> > 
> > This adds support for the Texas Instruments ADC084S021 ADC chip.
> > 
> > Signed-off-by: Mårten Lindahl <martenli@axis.com>
> A few more bits inline.  Mostly stuff that has come up
> in the V2 changes (and the inevitable bits I missed the
> first time!)
> 
> Jonathan

Hi Jonathan! Please see my comments. I will send v3 today.
Thanks,
Mårten

> > ---
> > Changes in v2:
> > - Removed most #defines in favor of inlines
> > - Corrected channel macro
> > - Removed configuration array with only one item
> > - Updated func adc084s021_adc_conversion to use be16_to_cpu
> > - Added IIO_CHAN_INFO_SCALE to func adc084s021_read_raw
> > - Use iio_device_claim_direct_mode in func adc084s021_read_raw
> > - Removed documentation for standard driver functions
> > - Changed retval to ret everywhere
> > - Removed dynamic alloc for data buffer in trigger handler
> > - Keeping mutex for all iterations in trigger handler
> > - Removed usage of events in this driver
> > - Removed info log in probe
> > - Use spi_message_init_with_transfers for spi message structs
> > - Use preenable and postdisable functions for regulator
> > - Inserted blank line before last return in all functions
> > 
> >  drivers/iio/adc/Kconfig         |  12 ++
> >  drivers/iio/adc/Makefile        |   1 +
> >  drivers/iio/adc/ti-adc084s021.c | 294 ++++++++++++++++++++++++++++++++++++++++
> >  3 files changed, 307 insertions(+)
> >  create mode 100644 drivers/iio/adc/ti-adc084s021.c
> > 
> > diff --git a/drivers/iio/adc/Kconfig b/drivers/iio/adc/Kconfig
> > index dedae7a..13141e5 100644
> > --- a/drivers/iio/adc/Kconfig
> > +++ b/drivers/iio/adc/Kconfig
> > @@ -560,6 +560,18 @@ config TI_ADC0832
> >  	  This driver can also be built as a module. If so, the module will be
> >  	  called ti-adc0832.
> >  
> > +config TI_ADC084S021
> > +	tristate "Texas Instruments ADC084S021"
> > +	depends on SPI
> > +	select IIO_BUFFER
> > +	select IIO_TRIGGERED_BUFFER
> > +	help
> > +	  If you say yes here you get support for Texas Instruments ADC084S021
> > +	  chips.
> > +
> > +	  This driver can also be built as a module. If so, the module will be
> > +	  called ti-adc084s021.
> > +
> >  config TI_ADC12138
> >  	tristate "Texas Instruments ADC12130/ADC12132/ADC12138"
> >  	depends on SPI
> > diff --git a/drivers/iio/adc/Makefile b/drivers/iio/adc/Makefile
> > index d001262..b1a6158 100644
> > --- a/drivers/iio/adc/Makefile
> > +++ b/drivers/iio/adc/Makefile
> > @@ -51,6 +51,7 @@ obj-$(CONFIG_STM32_ADC_CORE) += stm32-adc-core.o
> >  obj-$(CONFIG_STM32_ADC) += stm32-adc.o
> >  obj-$(CONFIG_TI_ADC081C) += ti-adc081c.o
> >  obj-$(CONFIG_TI_ADC0832) += ti-adc0832.o
> > +obj-$(CONFIG_TI_ADC084S021) += ti-adc084s021.o
> >  obj-$(CONFIG_TI_ADC12138) += ti-adc12138.o
> >  obj-$(CONFIG_TI_ADC128S052) += ti-adc128s052.o
> >  obj-$(CONFIG_TI_ADC161S626) += ti-adc161s626.o
> > diff --git a/drivers/iio/adc/ti-adc084s021.c b/drivers/iio/adc/ti-adc084s021.c
> > new file mode 100644
> > index 0000000..2dce257
> > --- /dev/null
> > +++ b/drivers/iio/adc/ti-adc084s021.c
> > @@ -0,0 +1,294 @@
> > +/**
> > + * Copyright (C) 2017 Axis Communications AB
> > + *
> > + * Driver for Texas Instruments' ADC084S021 ADC chip.
> > + * Datasheets can be found here:
> > + * http://www.ti.com/lit/ds/symlink/adc084s021.pdf
> > + *
> > + * This program is free software; you can redistribute it and/or modify
> > + * it under the terms of the GNU General Public License version 2 as
> > + * published by the Free Software Foundation.
> > + */
> > +
> > +#include <linux/err.h>
> > +#include <linux/spi/spi.h>
> > +#include <linux/module.h>
> > +#include <linux/interrupt.h>
> > +#include <linux/iio/iio.h>
> > +#include <linux/iio/buffer.h>
> > +#include <linux/iio/triggered_buffer.h>
> > +#include <linux/iio/trigger_consumer.h>
> > +#include <linux/regulator/consumer.h>
> > +
> > +#define ADC084S021_DRIVER_NAME "adc084s021"
> > +
> > +struct adc084s021 {
> > +	struct spi_device *spi;
> > +	struct spi_message message;
> > +	struct spi_transfer spi_trans[2];
> > +	struct regulator *reg;
> > +	struct mutex lock;
> > +	/*
> > +	 * DMA (thus cache coherency maintenance) requires the
> > +	 * transfer buffers to live in their own cache lines.
> > +	 */
> > +	union {
> > +		u16 tx_buf;
> > +		__be16 rx_buf;
> > +	} ____cacheline_aligned;
> > +};
> > +
> > +/**
> > + * Channel specification
> > + */
> Comment doesn't add much so I'd drop it.
Fixed in v3.
> > +#define ADC084S021_VOLTAGE_CHANNEL(num)                  \
> > +	{                                                      \
> > +		.type = IIO_VOLTAGE,                                 \
> > +		.channel = (num),                                    \
> > +		.address = (num) << 3,                               \
> This does feel a little pointless as you can just do the shift inline and
> use the channel value instead.  Doesn't really matter though.
Ok, fixed in v3.
> > +		.indexed = 1,                                        \
> > +		.scan_index = (num),                                 \
> > +		.scan_type = {                                       \
> > +			.sign = 'u',                                       \
> > +			.realbits = 8,                                     \
> > +			.storagebits = 16,                                 \
> > +			.shift = 4,                                        \
> > +			.endianness = IIO_BE,                              \
> > +		},                                                   \
> > +		.info_mask_separate = BIT(IIO_CHAN_INFO_RAW),        \
> > +		.info_mask_shared_by_type = BIT(IIO_CHAN_INFO_SCALE),\
> > +	}
> > +
> > +static const struct iio_chan_spec adc084s021_channels[] = {
> > +	ADC084S021_VOLTAGE_CHANNEL(0),
> > +	ADC084S021_VOLTAGE_CHANNEL(1),
> > +	ADC084S021_VOLTAGE_CHANNEL(2),
> > +	ADC084S021_VOLTAGE_CHANNEL(3),
> > +	IIO_CHAN_SOFT_TIMESTAMP(4),
> > +};
> > +
> > +/**
> > + * Read an ADC channel and return its value.
> > + *
> > + * @adc: The ADC SPI data.
> > + * @channel: The IIO channel data structure.
> > + */
> > +static int adc084s021_adc_conversion(struct adc084s021 *adc,
> > +			   struct iio_chan_spec const *channel)
> > +{
> > +	u8 value;
> > +	int ret;
> > +
> > +	adc->tx_buf = channel->address;
> > +
> > +	/* Do the transfer */
> > +	ret = spi_sync(adc->spi, &adc->message);
> > +	if (ret < 0)
> > +		return ret;
> > +
> > +	value = (be16_to_cpu(adc->rx_buf) >> channel->scan_type.shift) & 0xff;You want to do this for the read_raw sysfs calls, but not the buffered ones
> (where this shift and mask is made userspace's problem).
Fixed in v3.
> > +
> > +	dev_dbg(&adc->spi->dev, "value 0x%02X on channel %d\n",
> > +			     value, channel->channel);
> > +
> > +	return value;
> > +}
> > +
> > +static int adc084s021_read_raw(struct iio_dev *indio_dev,
> > +			   struct iio_chan_spec const *channel, int *val,
> > +			   int *val2, long mask)
> > +{
> > +	struct adc084s021 *adc = iio_priv(indio_dev);
> > +	int ret;
> > +
> > +	switch (mask) {
> > +	case IIO_CHAN_INFO_RAW:
> > +		ret = iio_device_claim_direct_mode(indio_dev);
> > +		if (ret < 0)
> > +			return ret;
> > +
> Unless I'm dozing off this morning, you don't have any power..
> So you'll need to enable the regulator first in this path.
Fixed in v3.
> 
> Note that the runtime power management autosuspend stuff can
> be good for this as it stops the power cycling if a short burst
> of readings is taken.
I don't have much experience of the pm_runtime system, so I looked
around to the other drivers but I see very few users of it. I can
absolutely look into it if you think it should be applied to this
driver.
> > +		ret = adc084s021_adc_conversion(adc, channel);
> > +		iio_device_release_direct_mode(indio_dev);
> > +		if (ret < 0)
> > +			return ret;
> > +
> > +		*val = ret;
> > +
> > +		return IIO_VAL_INT;
> > +	case IIO_CHAN_INFO_SCALE:
> > +		ret = regulator_get_voltage(adc->reg);
> Given the regulator is currently off as far as this device is concerned
> it's possible the answer will be 0...
Fixed in v3.
> > +		if (ret < 0)
> > +			return ret;
> > +
> > +		*val = ret / 1000;
> > +
> > +		return IIO_VAL_INT;
> > +	default:
> > +		return -EINVAL;
> > +	}
> > +}
> > +
> > +/**
> > + * Read enabled ADC channels and push data to the buffer.
> > + *
> > + * @irq: The interrupt number (not used).
> > + * @pollfunc: Pointer to the poll func.
> > + */
> > +static irqreturn_t adc084s021_buffer_trigger_handler(int irq, void *pollfunc)
> > +{
> > +	struct iio_poll_func *pf = pollfunc;
> > +	struct iio_dev *indio_dev = pf->indio_dev;
> > +	struct adc084s021 *adc = iio_priv(indio_dev);
> > +	__be16 data[8] = {0}; /* 4 * 16-bit words of data + 8 bytes timestamp */
> > +	int scan_index;
> > +	int i = 0;
> > +
> > +	mutex_lock(&adc->lock);
> > +
> > +	for_each_set_bit(scan_index, indio_dev->active_scan_mask,
> > +			 indio_dev->masklength) {
> > +		const struct iio_chan_spec *channel =
> > +			&indio_dev->channels[scan_index];
> > +		data[i++] = adc084s021_adc_conversion(adc, channel);
> This is clearly a rather simplistic way of handling the read outs.
> Perfectly good for an initial version (which may never get improved
> on!) but you could do the spi transfers for a set of channels much
> more efficiently.
> 
> Figure 1 on the datasheet shows how the control register for the next
> read can be written in parallel with the previous read. That would
> mean each scan took N + 1 2 byte transfers rather than 2N as currently.
> 
> I'm not particularly suggesting you do this, but thought I'd just
> comment on the possibility as it is a common situation with spi
> ADCs (sometimes you get a greater lag between setup and read out
> making this even fiddlier!)
> 
> If you do want to do this, the trick is to do your transfer setup
> and creation stuff in preenable so that it can be customised for
> whatever channels are enabled.
Ok, Fixed in v3.
I have looked into this and I think I improved the design according to
your suggestion. Please note: According to the datasheet (section 8.3)
after each power up the ADC will read out last scanned/first channel,
which will result in a very first read of 16 bits that is ignored. Also,
the readings are setup by only one transfer segment now.
> > +	}
> > +
> > +	iio_push_to_buffers_with_timestamp(indio_dev, data,
> > +					   iio_get_time_ns(indio_dev));
> > +	mutex_unlock(&adc->lock);
> > +	iio_trigger_notify_done(indio_dev->trig);
> > +
> > +	return IRQ_HANDLED;
> > +}
> > +
> > +static int adc084s021_buffer_preenable(struct iio_dev *indio_dev)
> > +{
> > +	struct adc084s021 *adc = iio_priv(indio_dev);
> > +
> > +	return regulator_enable(adc->reg);
> > +}
> > +
> > +static int adc084s021_buffer_postdisable(struct iio_dev *indio_dev)
> > +{
> > +	struct adc084s021 *adc = iio_priv(indio_dev);
> > +
> > +	return regulator_disable(adc->reg);
> > +}
> > +
> > +static const struct iio_info adc084s021_info = {
> > +	.read_raw = adc084s021_read_raw,
> > +	.driver_module = THIS_MODULE,
> > +};
> > +
> > +static const struct iio_buffer_setup_ops adc084s021_buffer_setup_ops = {
> > +	.preenable = adc084s021_buffer_preenable,
> > +	.postenable = iio_triggered_buffer_postenable,
> > +	.predisable = iio_triggered_buffer_predisable,
> > +	.postdisable = adc084s021_buffer_postdisable,
> > +};
> > +
> > +static int adc084s021_probe(struct spi_device *spi)
> > +{
> > +	struct iio_dev *indio_dev;
> > +	struct adc084s021 *adc;
> > +	int ret;
> > +
> > +	indio_dev = devm_iio_device_alloc(&spi->dev, sizeof(*adc));
> > +	if (!indio_dev) {
> > +		dev_err(&spi->dev, "Failed to allocate IIO device\n");
> > +		return -ENOMEM;
> > +	}
> > +
> > +	adc = iio_priv(indio_dev);
> > +	adc->spi = spi;
> > +	spi->bits_per_word = 8;
> > +
> > +	/* Update the SPI device with config and connect the iio dev */
> > +	ret = spi_setup(spi);
> > +	if (ret) {
> > +		dev_err(&spi->dev, "Failed to update SPI device\n");
> > +		return ret;
> > +	}
> > +	spi_set_drvdata(spi, indio_dev);
> > +
> > +	/* Initiate the Industrial I/O device */
> > +	indio_dev->dev.parent = &spi->dev;
> > +	indio_dev->dev.of_node = spi->dev.of_node;
> > +	indio_dev->name = spi_get_device_id(spi)->name;
> > +	indio_dev->modes = INDIO_DIRECT_MODE;
> > +	indio_dev->info = &adc084s021_info;
> > +	indio_dev->channels = adc084s021_channels;
> > +	indio_dev->num_channels = ARRAY_SIZE(adc084s021_channels);
> > +
> > +	/* Create SPI transfer for channel reads */
> > +	adc->spi_trans[0].tx_buf = &adc->tx_buf;
> > +	adc->spi_trans[0].len = 2;
> > +	adc->spi_trans[0].speed_hz = spi->max_speed_hz;
> You don't need to set these for individual transfers unless they
> are different from how the spi bus has been set up...
> 
> It's handled in __spi_validate in drivers/spi/spi.c
> 
> 	list_for_each_entry(xfer, &message->transfers, transfer_list) {
> 		message->frame_length += xfer->len;
> 		if (!xfer->bits_per_word)
> 			xfer->bits_per_word = spi->bits_per_word;
> 
> 		if (!xfer->speed_hz)
> 			xfer->speed_hz = spi->max_speed_hz;
> ...
> 
> So it will set them spi->... in the core if you haven't overridden.
Fixed in v3.
> 
> > +	adc->spi_trans[0].bits_per_word = spi->bits_per_word;
> > +	adc->spi_trans[1].rx_buf = &adc->rx_buf;
> > +	adc->spi_trans[1].len = 2;
> > +	adc->spi_trans[1].speed_hz = spi->max_speed_hz;
> > +	adc->spi_trans[1].bits_per_word = spi->bits_per_word;
> > +
> > +	spi_message_init_with_transfers(&adc->message, adc->spi_trans, 2);
> > +
> > +	adc->reg = devm_regulator_get(&spi->dev, "vref");
> > +	if (IS_ERR(adc->reg))
> > +		return PTR_ERR(adc->reg);
> > +
> > +	mutex_init(&adc->lock);
> > +
> > +	/* Setup triggered buffer with pollfunction */
> > +	ret = iio_triggered_buffer_setup(indio_dev, NULL,
> > +					    adc084s021_buffer_trigger_handler,
> > +					    &adc084s021_buffer_setup_ops);
> > +	if (ret) {
> > +		dev_err(&spi->dev, "Failed to setup triggered buffer\n");
> > +		return ret;
> > +	}
> > +
> > +	ret = iio_device_register(indio_dev);
> > +	if (ret) {
> > +		dev_err(&spi->dev, "Failed to register IIO device\n");
> > +		iio_triggered_buffer_cleanup(indio_dev);
> With devm usage this won't be needed any more.
Fixed in v3.
> 
> Hmm. We should improve the error message from the core as it'll allow us
> to remove a lot of boiler plate in drivers. Ah well, a project for
> another day.
I removed this message in v3.
> > +	}
> > +
> > +	return ret;
> > +}
> > +
> > +static int adc084s021_remove(struct spi_device *spi)
> > +{
> > +	struct iio_dev *indio_dev = spi_get_drvdata(spi);
> > +
> > +	iio_device_unregister(indio_dev);
> > +	iio_triggered_buffer_cleanup(indio_dev);
> Now you have moved to dynamically enabling the regulator, it makes
> sense to use the devm_ versions of iio_device_register and 
> iio_triggered_buffer setup.
> 
> With those, you'll be able to get rid of the remove callback entirely as
> there will be nothing to do.
Fixed in v3.
> > +
> > +	return 0;
> > +}
> > +
> > +static const struct of_device_id adc084s021_of_match[] = {
> > +	{ .compatible = "ti,adc084s021", },
> > +	{},
> > +};
> > +MODULE_DEVICE_TABLE(of, adc084s021_of_match);
> > +
> > +static const struct spi_device_id adc084s021_id[] = {
> > +	{ ADC084S021_DRIVER_NAME, 0},
> > +	{}
> > +};
> > +MODULE_DEVICE_TABLE(spi, adc084s021_id);
> > +
> > +static struct spi_driver adc084s021_driver = {
> > +	.driver = {
> > +		.name = ADC084S021_DRIVER_NAME,
> > +		.of_match_table = of_match_ptr(adc084s021_of_match),
> > +	},
> > +	.probe = adc084s021_probe,
> > +	.remove = adc084s021_remove,
> > +	.id_table = adc084s021_id,
> > +};
> > +module_spi_driver(adc084s021_driver);
> > +
> > +MODULE_AUTHOR("Mårten Lindahl <martenli@axis.com>");
> > +MODULE_DESCRIPTION("Texas Instruments ADC084S021");
> > +MODULE_LICENSE("GPL v2");
> > +MODULE_VERSION("1.0");
> > 
>