[PATCH 2/3] ARM: AT91: IIO: Add AT91 ADC driver.

[maxime.ripard@free-electrons.com (Maxime Ripard)]

On 04/11/2011 11:27, Jonathan Cameron wrote:
> On 11/03/2011 10:11 AM, Maxime Ripard wrote:
>> diff --git a/drivers/staging/iio/adc/at91adc.c b/drivers/staging/iio/adc/at91adc.c
>> new file mode 100644
>> index 0000000..acf656d
>> --- /dev/null
>> +++ b/drivers/staging/iio/adc/at91adc.c
>> @@ -0,0 +1,310 @@
>> +/*
>> + * Driver for the TouchScreen ADC Controller present in the Atmel AT91
>> + * evaluation boards.
> I'd add something here to make it clear that it isn't being used for
> touchscreen input in this driver.

Indeed, I forgot to change it here as well.

>> + *
>> + * Copyright 2011 Free Electrons
>> + *
>> + * Licensed under the GPLv2 or later.
>> + */
>> +
>> +#include <linux/clk.h>
>> +#include <linux/err.h>
>> +#include <linux/io.h>
>> +#include <linux/interrupt.h>
>> +#include <linux/jiffies.h>
>> +#include <linux/kernel.h>
>> +#include <linux/module.h>
>> +#include <linux/platform_device.h>
>> +#include <linux/slab.h>
>> +#include <linux/wait.h>
>> +
>> +#include "../iio.h"
>> +
>> +#include <mach/at91_adc.h>
>> +#include <mach/board.h>
>> +
> I'd marginally prefer to see these as inlines rather than macros.  Far
> from critical though!

Ok.

>> +#define at91adc_reg_read(base, reg)        readl_relaxed((base) + (reg))
>> +#define at91adc_reg_write(base, reg, val)  writel_relaxed((val), (base) + (reg))
>> +
>> +struct at91adc_state {
>> +	struct clk *clk;
>> +	bool done;
>> +	struct mutex lock;
> why hold a local pointer?  Just put these
> directly into iio_dev->channels.
>> +	struct iio_chan_spec *channels;
> likewise, we already have iio_dev->num_channels.
>> +	int nb_chan;

Indeed

>> +	int irq;
>> +	wait_queue_head_t wq_data_avail;
> cryptic element.  Nice to have everything documented, but
> this one is most important as it isn't self explanatory.
>> +	u16 lcdr;
>> +	void __iomem *reg_base;
>> +	unsigned int vref_mv;
>> +};
>> +
>> +static irqreturn_t at91adc_eoc_trigger(int irq, void *private)
>> +{
>> +	int chan;
>> +	struct at91adc_state *st = private;
>> +	unsigned int status = at91adc_reg_read(st->reg_base, AT91_ADC_SR);
>> +
>> +	if (!(status & AT91_ADC_DRDY))
>> +		return IRQ_HANDLED;
>> +
> extra unneeded brackets for the for loop.

I'm confused here. While it makes sense, I find it more readable around
large blocks of code and checkpatch doesn't complain on that one. But
coding style is a huge troll, so will change.

>> +	for (chan = 0; chan < st->nb_chan; chan++) {
>> +		if (status & AT91_ADC_EOC(chan)) {
>> +			st->done = true;
>> +			st->lcdr = at91adc_reg_read(st->reg_base,
>> +						    AT91_ADC_CHR(chan));
>> +		}
>> +	}
>> +	wake_up_interruptible(&st->wq_data_avail);
>> +
>> +	return IRQ_HANDLED;
>> +}
>> +
>> +static int at91adc_channel_init(struct at91adc_state *st,
>> +				struct at91_adc_data *pdata)
>> +{
>> +	int i, idx = 0;
>> +	st->channels = kcalloc(pdata->num_channels_used,
>> +			       sizeof(struct iio_chan_spec), GFP_KERNEL);
>> +	if (st->channels == NULL)
>> +		return -ENOMEM;
>> +
> Again, extra brackets.  Also, some wrong spacing for the if.
> can you run checkpatch.pl over this again and make sure you
> have fixed everything.
>> +	for (i = 0; i < st->nb_chan; i++) {
> as stated for the previous patch, this should be a bitmap rather
> than array of u8's.   You can then use for_each_bit_set
> to clean this code up nicely ;)
>> +		if(pdata->channels_used[i]) {
>> +			struct iio_chan_spec *chan = st->channels + idx;
>> +			chan->type = IIO_VOLTAGE;
>> +			chan->indexed = 1;
>> +			chan->channel = i;
>> +			chan->scan_type.sign = 's';
>> +			chan->scan_type.realbits = 10;
>> +			chan->scan_type.storagebits = 32;
> That is pretty inefficient storage!  If we do implement buffering
> on this, given mass reads don't look to be quicker, we'll just
> munge it down to 16 bits.

Well, I thought that storage bits was here to represent how the hardware
stored the values. Since these values are stored on 10 bits alone in a
32 bits register, I thought that these were the right values, but we can
definitely lower the storagebits to 16 here.

> 
>> +			chan->scan_type.shift = 0;
> don't bother initialising this value as it is zero (and that's an
> obvious default).
>> +			chan->info_mask = IIO_CHAN_INFO_SCALE_SHARED_BIT;
>> +			++idx;
>> +		}
>> +	}
>> +
>> +	return st->nb_chan;
>> +}
>> +
>> +static int at91adc_read_raw(struct iio_dev *idev,
>> +			    struct iio_chan_spec const *chan,
>> +			    int *val, int *val2, long mask)
>> +{
>> +	struct at91adc_state *st = iio_priv(idev);
>> +	unsigned int scale_uv;
>> +
>> +	switch (mask) {
>> +	case 0:
>> +		mutex_lock(&st->lock);
>> +
>> +		at91adc_reg_write(st->reg_base, AT91_ADC_CHER,
>> +				  AT91_ADC_CH(chan->channel));
>> +		at91adc_reg_write(st->reg_base, AT91_ADC_IER,
>> +				  AT91_ADC_EOC(chan->channel));
>> +		at91adc_reg_write(st->reg_base, AT91_ADC_CR, AT91_ADC_START);
>> +
>> +		wait_event_interruptible_timeout(st->wq_data_avail, st->done,
>> +						 msecs_to_jiffies(10 * 1000));
>> +		*val = st->lcdr;
>> +
>> +		at91adc_reg_write(st->reg_base, AT91_ADC_CHDR,
>> +				  AT91_ADC_CH(chan->channel));
>> +		at91adc_reg_write(st->reg_base, AT91_ADC_IDR,
>> +				  AT91_ADC_EOC(chan->channel));
>> +
>> +		st->lcdr = 0;
>> +		st->done = false;
>> +		mutex_unlock(&st->lock);
>> +		return IIO_VAL_INT;
>> +	case IIO_CHAN_INFO_SCALE:
>> +		scale_uv = (st->vref_mv * 1000) >> chan->scan_type.realbits;
>> +		*val = scale_uv / 1000;
>> +		*val2 = (scale_uv % 1000) * 1000;
>> +		return IIO_VAL_INT_PLUS_MICRO;
>> +	default:
>> +		break;
>> +	}
>> +	return -EINVAL;
>> +}
>> +
>> +static const struct iio_info at91adc_info = {
>> +	.driver_module = THIS_MODULE,
>> +	.read_raw = &at91adc_read_raw,
>> +};
>> +
>> +static int __devinit at91adc_probe(struct platform_device *pdev)
>> +{
>> +	unsigned int prsc, mstrclk, ticks;
>> +	int ret;
>> +	struct iio_dev *idev;
>> +	struct at91adc_state *st;
>> +	struct resource *res;
>> +	struct at91_adc_data *pdata = pdev->dev.platform_data;
>> +
>> +	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
>> +	if (!res) {
>> +		dev_err(&pdev->dev, "No resource defined\n");
>> +		ret = -ENXIO;
>> +		goto error_ret;
>> +	}
>> +
>> +	idev = iio_allocate_device(sizeof(struct at91adc_state));
> sizeof(*st) is a little neater?

Well, I find sizeof(struct at91adc_state) clearer, but ok.

>> +	if (idev == NULL) {
> I'm not sure this message adds anything or is even necessarily
> correct as that function doesn't just allocate memory!

Right

>> +		dev_err(&pdev->dev, "Failed to allocate memory.\n");
>> +		ret = -ENOMEM;
>> +		goto error_ret;
>> +	}
>> +	platform_set_drvdata(pdev, idev);
>> +
>> +	idev->dev.parent = &pdev->dev;
>> +	idev->name = dev_name(&pdev->dev);
>> +	idev->modes = INDIO_DIRECT_MODE;
>> +	idev->info = &at91adc_info;
>> +
>> +	st = iio_priv(idev);
>> +	st->irq = platform_get_irq(pdev, 0);
>> +	if (st->irq < 0) {
>> +		dev_err(&pdev->dev, "No IRQ ID is designated\n");
>> +		ret = -ENODEV;
>> +		goto error_free_device;
>> +	}
>> +
>> +	if (!request_mem_region(res->start, resource_size(res),
>> +				"AT91 adc registers")) {
>> +		dev_err(&pdev->dev, "Resources are unavailable.\n");
>> +		ret = -EBUSY;
>> +		goto error_free_device;
>> +	}
>> +
>> +	st->reg_base = ioremap(res->start, resource_size(res));
>> +	if (!st->reg_base) {
>> +		dev_err(&pdev->dev, "Failed to map registers.\n");
>> +		ret = -ENOMEM;
>> +		goto error_release_mem;
>> +	}
>> +
>> +	/*
>> +	 * Disable all IRQs before setting up the handler
>> +	 */
>> +	at91adc_reg_write(st->reg_base, AT91_ADC_CR, AT91_ADC_SWRST);
>> +	at91adc_reg_write(st->reg_base, AT91_ADC_IDR, 0xFFFFFFFF);
>> +	ret = request_irq(st->irq,
>> +			  at91adc_eoc_trigger, 0, pdev->dev.driver->name, st);
>> +	if (ret) {
>> +		dev_err(&pdev->dev, "Failed to allocate IRQ.\n");
>> +		goto error_unmap_reg;
>> +	}
>> +
>> +	st->clk = clk_get(&pdev->dev, "adc_clk");
>> +	if (IS_ERR(st->clk)) {
>> +		dev_err(&pdev->dev, "Failed to get the clock.\n");
>> +		ret = PTR_ERR(st->clk);
>> +		goto error_free_irq;
>> +	}
>> +
>> +	clk_enable(st->clk);
>> +	mstrclk = clk_get_rate(st->clk);
>> +
>> +	if (!pdata) {
>> +		dev_err(&pdev->dev, "No platform data available.\n");
>> +		ret = -EINVAL;
>> +		goto error_free_clk;
>> +	}
>> +
>> +	if (!pdata->adc_clock) {
>> +		dev_err(&pdev->dev, "No ADCClock available.\n");
>> +		ret = -EINVAL;
>> +		goto error_free_clk;
>> +	}
>> +
>> +	prsc = (mstrclk / (2 * pdata->adc_clock)) - 1;
>> +
>> +	if (!pdata->startup_time) {
>> +		dev_err(&pdev->dev, "No startup time available.\n");
>> +		ret = -EINVAL;
>> +		goto error_free_clk;
>> +	}
>> +
>> +	ticks = round_up((pdata->startup_time * pdata->adc_clock /
>> +			  1000000) - 1, 8) / 8;
>> +	at91adc_reg_write(st->reg_base, AT91_ADC_MR,
>> +			  (AT91_ADC_PRESCAL_(prsc) & AT91_ADC_PRESCAL) |
>> +			  (AT91_ADC_STARTUP_(ticks) & AT91_ADC_STARTUP));
>> +
>> +	/* Setup the ADC channels available on the board */
>> +	st->nb_chan = pdata->num_channels;
>> +	ret = at91adc_channel_init(st, pdata);
>> +	if (ret < 0) {
> no brackets please. Again, checkpatch would probably have shouted
> about this.
>> +		goto error_free_clk;
>> +	}
>> +
> This is what I meant when I asked why you have two copies of the
> pointer and count?
>> +	idev->channels = st->channels;
>> +	idev->num_channels = pdata->num_channels_used;
>> +
>> +	init_waitqueue_head(&st->wq_data_avail);
>> +	mutex_init(&st->lock);
>> +
>> +	st->vref_mv = pdata->vref;
>> +
>> +	ret = iio_device_register(idev);
> again incorrect brackets for the code style.
>> +	if (ret < 0) {
> unwind the channel init?
Oooh, right...

>> +		goto error_free_clk;
>> +	}
>> +
>> +	return 0;
>> +
>> +error_free_clk:
>> +	clk_disable(st->clk);
>> +	clk_put(st->clk);
>> +error_free_irq:
>> +	free_irq(st->irq, st);
>> +error_unmap_reg:
>> +	iounmap(st->reg_base);
>> +error_release_mem:
>> +	release_mem_region(res->start, resource_size(res));
>> +error_free_device:
>> +	iio_free_device(idev);
>> +error_ret:
>> +	return ret;
>> +}
>> +
>> +static int __devexit at91adc_remove(struct platform_device *pdev)
>> +{
>> +	struct iio_dev *idev = platform_get_drvdata(pdev);
>> +	struct resource *res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
>> +	struct at91adc_state *st = iio_priv(idev);
>> +
>> +	iio_device_unregister(idev);
> Free channels array?
>> +	free_irq(st->irq, st);
>> +	iounmap(st->reg_base);
>> +	release_mem_region(res->start, resource_size(res));
>> +	iio_free_device(idev);
>> +
>> +	return 0;
>> +}
>> +
>> +static struct platform_driver at91adc_driver = {
>> +	.probe = at91adc_probe,
>> +	.remove = __devexit_p(at91adc_remove),
>> +	.driver = {
>> +		   .name = "at91adc",
>> +		   },
>> +};
>> +
>> +static int __init at91adc_init(void)
>> +{
>> +	return platform_driver_register(&at91adc_driver);
>> +}
>> +
>> +static void __exit at91adc_exit(void)
>> +{
>> +	platform_driver_unregister(&at91adc_driver);
>> +}
>> +
> 
> Some neat boiler plate removal stuff was just merged for
> platform drivers.  See module_platform_driver in include/linux/platform.h
> 
> Definitely want to use that!  I'm hoping the spi and i2c versions
> turn up soon as it will save us a lot of silly cut and paste code.
Aah, nice :)

Thanks,

-- 
Maxime Ripard, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com