Re: [PATCH] input: zforce_ts: add DT support for reset GPIO polarity

From: Dirk Behme <dirk.behme@de.bosch.com>
To: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Cc: linux-input@vger.kernel.org,
	Knut Wohlrab <Knut.Wohlrab@de.bosch.com>,
	Oleksij Rempel <external.Oleksij.Rempel@de.bosch.com>
Subject: Re: [PATCH] input: zforce_ts: add DT support for reset GPIO polarity
Date: Tue, 30 Jun 2015 07:40:04 +0200	[thread overview]
Message-ID: <55922BB4.40809@de.bosch.com> (raw)
In-Reply-To: <20150630000858.GA35823@dtor-ws>

Hi Dmitry,

On 30.06.2015 02:08, Dmitry Torokhov wrote:
> On Wed, Jun 24, 2015 at 07:18:38AM +0200, Dirk Behme wrote:
>> On 17.06.2015 01:57, Dmitry Torokhov wrote:
>>> On Mon, Jun 15, 2015 at 07:54:25AM +0200, Dirk Behme wrote:
>>>> From: Knut Wohlrab <Knut.Wohlrab@de.bosch.com>
>>>>
>>>> According to
>>>>
>>>> Documentation/devicetree/bindings/input/touchscreen/zforce_ts.txt
>>>>
>>>> the RST GPIO is supposed to provide a polarity flag parameter
>>>>
>>>> gpios = <&GPIO_BANK GPIO_NUMBER GPIO_POLARITY>
>>>>
>>>> with GPIO_POLARITY
>>>>
>>>> reset active low  = 1 (GPIO_ACTIVE_LOW)
>>>> reset active high = 0 (GPIO_ACTIVE_HIGH)
>>>>
>>>> Example for GPIO_ACTIVE_LOW (1) reset GPIO:
>>>>
>>>> 	zforce_ts@50 { /* Neonode zForce I2C */
>>>> 		compatible = "neonode,zforce-ts";
>>>> 		...
>>>> 		gpios = <&gpio6 14 GPIO_ACTIVE_HIGH>, /* INT */
>>>> 			<&gpio1 29 GPIO_ACTIVE_LOW>; /* RST */
>>>> 		...
>>>> 	};
>>>>
>>>> Add the missing polarity flag evaluation to the driver.
>>>>
>>>> Signed-off-by: Knut Wohlrab <Knut.Wohlrab@de.bosch.com>
>>>> Signed-off-by: Oleksij Rempel <external.Oleksij.Rempel@de.bosch.com>
>>>> Signed-off-by: Dirk Behme <dirk.behme@de.bosch.com>
>>>> ---
>>>>   drivers/input/touchscreen/zforce_ts.c   | 27 +++++++++++++++++++++++----
>>>>   include/linux/platform_data/zforce_ts.h |  3 +++
>>>>   2 files changed, 26 insertions(+), 4 deletions(-)
>>>>
>>>> diff --git a/drivers/input/touchscreen/zforce_ts.c b/drivers/input/touchscreen/zforce_ts.c
>>>> index 19880c7..125311d 100644
>>>> --- a/drivers/input/touchscreen/zforce_ts.c
>>>> +++ b/drivers/input/touchscreen/zforce_ts.c
>>>> @@ -162,6 +162,20 @@ static int zforce_command(struct zforce_ts *ts, u8 cmd)
>>>>   	return 0;
>>>>   }
>>>>
>>>> +static void zforce_reset_assert(struct zforce_ts *ts)
>>>> +{
>>>> +	const struct zforce_ts_platdata *pdata = ts->pdata;
>>>> +
>>>> +	gpio_set_value(pdata->gpio_rst, pdata->reset_active_low ? 0 : 1);
>>>
>>> Instead of doing this I'd rather we converted the driver to use gpiod
>>> that handles polarity automatically.
>>
>>
>> Thanks, we'll look into that.
>>
>> Just to understand correctly: Converting this driver to gpiod will
>> be an additional patch on top of this patch as it doesn't change any
>> functionality provided by this patch, but does some clean up.
>> Correct?
>
> No, gpiod understands annotations on gpio descriptions and converts the
> logical "active"/"inactive" value into appropriate output depending on
> the polarity specified for gpio. So zforce_reset_assert() woudl only
> need to do
>
> 	gpiod_set_value[_cansleep](ts->gpio_rst, 1);

Yes, this is understood. But my question has been a different one ;) Let 
me rephrase it:

Based on the existing driver this patch adds support for the (missing) 
GPIO polarity on the reset pin. Based on the existing GPIO framework 
used by the driver. Not talking about gpiod for the moment, I hope this 
is fine (?).

Then, you've asked to convert the driver to gpiod. That's fine, too. And 
will make the driver cleaner. And yes, it will handle the polarity, too.

With this, my proposal is to have two patches:

1) the existing one adding the polarity based on the existing GPIO 
framework (handling the polarity 'manually')

2) converting the driver to gpiod, handling the polarity implicitly, then.

I'd propose to have 2 patches as described above as it would separate 
the additional functionality (patch #1) and the gpiod framework change 
(patch #2).

So my question is about the structuring of the patches, not about 
details of gpiod_xxx().

Best regards

Dirk