On 23/10/16 10:57, H. Nikolaus Schaller wrote:
> Hi,
>
>> Am 23.10.2016 um 11:24 schrieb Jonathan Cameron <jic23@xxxxxxxxxx>:
>>
>> On 22/10/16 21:46, H. Nikolaus Schaller wrote:
>>> Hi Jonathan,
>>>
>>>> Am 22.10.2016 um 20:33 schrieb Jonathan Cameron <jic23@xxxxxxxxxx>:
>>>>
>>>> On 17/10/16 14:57, H. Nikolaus Schaller wrote:
>>>>> The tsc2007 chip not only has a resistive touch screen controller but
>>>>> also an external AUX adc imput which can be used for an ambient
>>>>> light sensor, battery voltage monitoring or any general purpose.
>>>>>
>>>>> Additionally it can measure the chip temperature.
>>>>>
>>>>> This extension provides an iio interface for these adc channels.
>>>>>
>>>>> Since it is not wasting much resources and is very straightforward,
>>>>> we simply provide all other adc channels as optional iio interfaces
>>>>> as weel. This can be used for debugging or special applications.
>>>> well
>>>>>
>>>>> Signed-off-by: H. Nikolaus Schaller <hns@xxxxxxxxxxxxx>
>>>> This could be cleaner done perhaps by factoring out the IIO stuff into a separate
>>>> file and using a header with stubs to deal with the no available case.
>>>>
>>>> There will only be a handful of stubs and it'll give you a lot cleaner code
>>>> in here.
>>>>
>>>> If def fun in .c files is always harder to deal with than in a header
>>>> where stubs are really obvious.
>>>
>>> Yes, it became a lot of #ifdefs spread over the source file.
>>>
>>> The easiest thing would be to require IIO to be enabled :)
>>>
>>> With your proposal to consider refactoring, I think the crucial part
>>> is the conditional allocation either through devm_iio_device_alloc()
>>> or devm_kzalloc(). This can be refactored into some conditional
>>> tsc2007_alloc().
>>>
>>> I have tried some draft (not tested and not tidied up) to check if the
>>> direction is good.
>>>
>>> This reduces the number of #ifdef CONFIG_IIO from 7 to 2 without introducing
>>> new files or includes. There are also 2 other #ifdef CONFIG_OF so it doesn't
>>> seem to be very complex now in comparison. And the patch itself has only a
>>> handful of hunks (8).
>>>
>>> Moving tsc2007_alloc into a separate file tsc2007_iio.c would only move around
>>> one #ifdef CONFIG_OF from tsc2007.c but IMHO makes it more difficult to understand
>>> because it is not really iio specific and one has to switch between two source
>>> files. And I would have to touch the Makefile as well.
>>>
>>> What do you think?
>> I'd still split it. The only bit of the IIO block that isn't specific is
>> a tiny chunk of the allocation code (as you've highlighted).
>>
>> Even that can be avoided by adding a tiny bit more indirection than would
>> otherwise be needed (it's not pretty but it would give a clean separation).
>
> I hope I understand what you mean (which is an indication that the result
> may be much easier to read for you but not me...).
>
>> It's pretty much the way this sort of optional functionality should always
>> be done - means that if you don't care (i.e. it's not enabled) you don't
>> even have to see the code.
>
>>
>> Jonathan
>>>
>>> If generally ok, I can include that in [PATCH v5].
>>>
>>> BR and thanks,
>>> Nikolaus
>>>
>>>
>>> diff --git a/drivers/input/touchscreen/tsc2007.c b/drivers/input/touchscreen/tsc2007.c
>>> index 5e3c4bf..691e79f 100644
>>> --- a/drivers/input/touchscreen/tsc2007.c
>>> +++ b/drivers/input/touchscreen/tsc2007.c
>>> @@ -30,6 +30,7 @@
>>> #include <linux/of.h>
>>> #include <linux/of_gpio.h>
>>> #include <linux/input/touchscreen.h>
>>> +#include <linux/iio/iio.h>
>>>
>>> #define TSC2007_MEASURE_TEMP0 (0x0 << 4)
>>> #define TSC2007_MEASURE_AUX (0x2 << 4)
>>> @@ -69,9 +70,13 @@ struct ts_event {
>>>
>>> struct tsc2007 {
>>> struct input_dev *input;
>>> +#ifdef CONFIG_IIO
>>> + struct iio_dev *indio;
>>> +#endif
>> I wouldn't bother with this one. Just have
>> struct iio_dev; before this and it'll waste a whole
>> one pointer (+ you shouldn't need to have iio.h included
>> in here once you have spit the files).
>
> Looks as if I have to make a knot in my brain before I start to understand...
>
> How can I use struct iio_dev here w/o including iio.h?
you aren't using it. You have a pointer to it.
So it (before this definition) you have a line that says
struct iio_dev; you let the compiler know such a structure exists.
At that point you don't actually have to provide a definition of
what is in it as long as all you use is a pointer (they are always
the same size).
>
>>> char phys[32];
>>>
>>> struct i2c_client *client;
>>> + struct mutex mlock;
>>>
>>> u16 model;
>>> u16 x_plate_ohms;
>>> @@ -192,7 +197,10 @@ static irqreturn_t tsc2007_soft_irq(int irq, void *handle)
>>> while (!ts->stopped && tsc2007_is_pen_down(ts)) {
>>>
>>> /* pen is down, continue with the measurement */
>>> +
>>> + mutex_lock(&ts->mlock);
>>> tsc2007_read_values(ts, &tc);
>>> + mutex_unlock(&ts->mlock);
>>>
>>> rt = tsc2007_calculate_resistance(ts, &tc);
>>>
>>> @@ -319,6 +327,162 @@ static void tsc2007_close(struct input_dev *input_dev)
>>> tsc2007_stop(ts);
>>> }
>>>
>>> +#ifdef CONFIG_IIO
>>> +
>>> +#define TSC2007_CHAN_IIO(_chan, _name, _type, _chan_info) \
>>> +{ \
>>> + .datasheet_name = _name, \
>>> + .type = _type, \
>>> + .info_mask_separate = BIT(IIO_CHAN_INFO_RAW) | \
>>> + BIT(_chan_info), \
>>> + .indexed = 1, \
>>> + .channel = _chan, \
>>> +}
>>> +
>>> +static const struct iio_chan_spec tsc2007_iio_channel[] = {
>>> + TSC2007_CHAN_IIO(0, "x", IIO_VOLTAGE, IIO_CHAN_INFO_RAW),
>>> + TSC2007_CHAN_IIO(1, "y", IIO_VOLTAGE, IIO_CHAN_INFO_RAW),
>>> + TSC2007_CHAN_IIO(2, "z1", IIO_VOLTAGE, IIO_CHAN_INFO_RAW),
>>> + TSC2007_CHAN_IIO(3, "z2", IIO_VOLTAGE, IIO_CHAN_INFO_RAW),
>>> + TSC2007_CHAN_IIO(4, "adc", IIO_VOLTAGE, IIO_CHAN_INFO_RAW),
>>> + TSC2007_CHAN_IIO(5, "rt", IIO_VOLTAGE, IIO_CHAN_INFO_RAW), /* Ohms? */
>>> + TSC2007_CHAN_IIO(6, "pen", IIO_PRESSURE, IIO_CHAN_INFO_RAW),
>>> + TSC2007_CHAN_IIO(7, "temp0", IIO_TEMP, IIO_CHAN_INFO_RAW),
>>> + TSC2007_CHAN_IIO(8, "temp1", IIO_TEMP, IIO_CHAN_INFO_RAW),
>>> +};
>>> +
>>> +static int tsc2007_read_raw(struct iio_dev *indio_dev,
>>> + struct iio_chan_spec const *chan, int *val, int *val2, long mask)
>>> +{
>>> + struct tsc2007 *tsc = iio_priv(indio_dev);
>>> + int adc_chan = chan->channel;
>>> + int ret = 0;
>>> +
>>> + if (adc_chan >= ARRAY_SIZE(tsc2007_iio_channel))
>>> + return -EINVAL;
>>> +
>>> + if (mask != IIO_CHAN_INFO_RAW)
>>> + return -EINVAL;
>>> +
>>> + mutex_lock(&tsc->mlock);
>>> +
>>> + switch (chan->channel) {
>>> + case 0:
>>> + *val = tsc2007_xfer(tsc, READ_X);
>>> + break;
>>> + case 1:
>>> + *val = tsc2007_xfer(tsc, READ_Y);
>>> + break;
>>> + case 2:
>>> + *val = tsc2007_xfer(tsc, READ_Z1);
>>> + break;
>>> + case 3:
>>> + *val = tsc2007_xfer(tsc, READ_Z2);
>>> + break;
>>> + case 4:
>>> + *val = tsc2007_xfer(tsc, (ADC_ON_12BIT | TSC2007_MEASURE_AUX));
>>> + break;
>>> + case 5: {
>>> + struct ts_event tc;
>>> +
>>> + tc.x = tsc2007_xfer(tsc, READ_X);
>>> + tc.z1 = tsc2007_xfer(tsc, READ_Z1);
>>> + tc.z2 = tsc2007_xfer(tsc, READ_Z2);
>>> + *val = tsc2007_calculate_resistance(tsc, &tc);
>>> + break;
>>> + }
>>> + case 6:
>>> + *val = tsc2007_is_pen_down(tsc);
>>> + break;
>>> + case 7:
>>> + *val = tsc2007_xfer(tsc,
>>> + (ADC_ON_12BIT | TSC2007_MEASURE_TEMP0));
>>> + break;
>>> + case 8:
>>> + *val = tsc2007_xfer(tsc,
>>> + (ADC_ON_12BIT | TSC2007_MEASURE_TEMP1));
>>> + break;
>>> + }
>>> +
>>> + /* Prepare for next touch reading - power down ADC, enable PENIRQ */
>>> + tsc2007_xfer(tsc, PWRDOWN);
>>> +
>>> + mutex_unlock(&tsc->mlock);
>>> +
>>> + ret = IIO_VAL_INT;
>>> +
>>> + return ret;
>>> +}
>>> +
>>> +static const struct iio_info tsc2007_iio_info = {
>>> + .read_raw = tsc2007_read_raw,
>>> + .driver_module = THIS_MODULE,
>>> +};
>>> +
>>> +static int tsc2007_alloc(struct i2c_client *client, struct tsc2007 **ts,
>>> + struct input_dev **input_dev)
>>> +{
>>> + int err;
>>> + struct iio_dev *indio_dev;
>>> +
>>> + indio_dev = devm_iio_device_alloc(&client->dev, sizeof(*ts));
>> Instead of doing this to reduce the delta between versions make
>> iio_priv a struct tsc2007 **
>>
>> That is have a single pointer in there and do your allocation of struct
>> tsc2007 separately.
>
> Sorry, but I think I do not completely understand what you mean here.
>
> The problem is that we need to allocate some struct tsc2007 in both cases.
> But in one case managed directly by &client->dev and in the other managed
> indirectly. This is why I use the private area of struct iio_dev to store
> the full struct tsc2007 and not just a pointer.
>
No you don't need to do what you are currently doing.
You need to have some means to navigate from struct iio_dev to the
struct tsc2007 - that doesn't have to be because it actually is
in iio_priv.
You can instead put a point to it in iio_priv (and only that) and allocate
it the same way in both paths (stashing a copy of the address in the
pointer in iio_priv).
> What I mean is:
>
>>> + indio_dev = devm_iio_device_alloc(&client->dev, sizeof(*ts));
>>> *ts = iio_priv(indio_dev);
>
> vs.
>
>>> *ts = devm_kzalloc(&client->dev, sizeof(struct tsc2007), GFP_KERNEL);
>
>
> So how can the IIO case just extend/wrap devm_kzalloc(&client->dev...) and still
> be managed as well?
Not relevant if you just allocate it the same way both times.
>
>>
>> Having doing that, you can have this CONFIG_IIO block as just
>> doing the iio stuff with the input elements pulled back into the main
>> probe function.
>>
>> Then define something like
>>
>> iio_configure (stubbed to nothing if no IIO)
>> and
>> iio_unconfigure (also stubbed to nothing if no IIO).
>>
>> A couple of additions in the header to make it all work
>> (the struct tsc2007 and tsc2007_xfer() + a few of the
>> register defines..
>>
>> Nothing big and gets all the CONFIG_IIO into some really
>> obvious stubbing out in the header.
>
>
> Is there some example driver which is doing it that way to be optionally IIO
> compatible? That might be easier to understand and copy than a description.
This particular combination is unusual - but it similar to how we
do optional buffer or trigger support in various iio drivers.
Perhaps see include/linux/iio/common/st_sensors.h and look for CONFIG_IIO_TRIGGER
>
>>
>>> + if (!indio_dev) {
>>> + dev_err(&client->dev, "iio_device_alloc failed\n");
>>> + return -ENOMEM;
>>> + }
>>> +
>>> + *ts = iio_priv(indio_dev);
>>> +
>>> + *input_dev = devm_input_allocate_device(&client->dev);
>>> + if (!*input_dev)
>>> + return -ENOMEM;
>>> +
>>> + i2c_set_clientdata(client, *ts);
>>> + (*ts)->indio = indio_dev;
>>> +
>>> + indio_dev->name = "tsc2007";
>>> + indio_dev->dev.parent = &client->dev;
>>> + indio_dev->info = &tsc2007_iio_info;
>>> + indio_dev->modes = INDIO_DIRECT_MODE;
>>> + indio_dev->channels = tsc2007_iio_channel;
>>> + indio_dev->num_channels = ARRAY_SIZE(tsc2007_iio_channel);
>>> +
>>> + err = iio_device_register(indio_dev);
>>> + if (err < 0) {
>>> + dev_err(&client->dev, "iio_device_register() failed: %d\n",
>>> + err);
>>> + return err;
>>> + }
>>> +
>>> + return 0;
>>> +}
>>> +
>>> +#define tsc2007_iio_device_unregister(ts) iio_device_unregister(ts->indio)
>>> +
>>> +#else /* CONFIG_IIO */
>>> +
>>> +static int tsc2007_alloc(struct i2c_client *client, struct tsc2007 **ts,
>>> + struct input_dev **input_dev)
>>> +{
>>> + int err;
>>> +
>>> + *ts = devm_kzalloc(&client->dev, sizeof(struct tsc2007), GFP_KERNEL);
>>> + if (!*ts)
>>> + return -ENOMEM;
>>> +
>>> + *input_dev = devm_input_allocate_device(&client->dev);
>>> + if (!*input_dev)
>>> + return -ENOMEM;
>>> +
>>> + i2c_set_clientdata(client, *ts);
>>> +
>>> + return 0;
>>> +}
>>> +
>>> +#define tsc2007_iio_device_unregister(ts) /* not needed */
>> That's rather ugly and fragile. I'd stub it out as an actual function
>> with no content and let the compiler drop it.
>
> Well, it is a quick and dirty draft.
> Should indeed better be a static (inline) function with empty body.
>
>>> +
>>> +#endif /* CONFIG_IIO */
>>> +
>>> #ifdef CONFIG_OF
>>> static int tsc2007_get_pendown_state_gpio(struct device *dev)
>>> {
>>> @@ -459,20 +623,15 @@ static int tsc2007_probe(struct i2c_client *client,
>>> I2C_FUNC_SMBUS_READ_WORD_DATA))
>>> return -EIO;
>>>
>>> - ts = devm_kzalloc(&client->dev, sizeof(struct tsc2007), GFP_KERNEL);
>>> - if (!ts)
>>> - return -ENOMEM;
>>> -
>>> - input_dev = devm_input_allocate_device(&client->dev);
>>> - if (!input_dev)
>>> - return -ENOMEM;
>>> -
>>> - i2c_set_clientdata(client, ts);
>>> + err = tsc2007_alloc(client, &ts, &input_dev);
>>> + if (err < 0)
>>> + return err;
>>>
>>> ts->client = client;
>>> ts->irq = client->irq;
>>> ts->input = input_dev;
>>> init_waitqueue_head(&ts->wait);
>>> + mutex_init(&ts->mlock);
>>>
>>> snprintf(ts->phys, sizeof(ts->phys),
>>> "%s/input0", dev_name(&client->dev));
>>> @@ -543,6 +702,7 @@ static int tsc2007_probe(struct i2c_client *client,
>>> if (err < 0) {
>>> dev_err(&client->dev,
>>> "Failed to setup chip: %d\n", err);
>>> + tsc2007_iio_device_unregister(ts);
>>> return err; /* usually, chip does not respond */
>>> }
>>>
>>> @@ -556,6 +716,14 @@ static int tsc2007_probe(struct i2c_client *client,
>>> return 0;
>>> }
>>>
>>> +static int tsc2007_remove(struct i2c_client *client)
>>> +{
>>> + struct tsc2007 *ts = i2c_get_clientdata(client);
>>> + input_unregister_device(ts->input);
>>> + tsc2007_iio_device_unregister(ts);
>>> + return 0;
>>> +}
>>> +
>>> static const struct i2c_device_id tsc2007_idtable[] = {
>>> { "tsc2007", 0 },
>>> { }
>>> @@ -578,6 +746,7 @@ static struct i2c_driver tsc2007_driver = {
>>> },
>>> .id_table = tsc2007_idtable,
>>> .probe = tsc2007_probe,
>>> + .remove = tsc2007_remove,
>>> };
>>>
>>> module_i2c_driver(tsc2007_driver);
>>>
>>
>
> BR and thanks,
> Nikolaus
>
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
