Re: [PATCH v2] pinctrl: dra: dt-bindings: Fix output pull up/down

[Nishanth Menon <nm@ti.com>]

On 11/03/2014 08:44 AM, Roger Quadros wrote:
> On 11/03/2014 04:30 PM, Nishanth Menon wrote:
>> On 12:09-20141103, Roger Quadros wrote:
>>> For PIN_OUTPUT_PULLUP and PIN_OUTPUT_PULLDOWN we must not set the
>>> PULL_DIS bit which disables the PULLs.
>>>
>>> PULL_ENA is a 0 and using it in an OR operation is a NOP, so don't
>>> use it in the PIN_OUTPUT_PULLUP/DOWN macros.
>>>
>>> Fixes: 23d9cec07c58 ("pinctrl: dra: dt-bindings: Fix pull enable/disable")
>>>
>>> Signed-off-by: Roger Quadros <rogerq@ti.com>
>>> ---
>>>  include/dt-bindings/pinctrl/dra.h | 4 ++--
>>>  1 file changed, 2 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/include/dt-bindings/pinctrl/dra.h b/include/dt-bindings/pinctrl/dra.h
>>> index 3d33794..7448edf 100644
>>> --- a/include/dt-bindings/pinctrl/dra.h
>>> +++ b/include/dt-bindings/pinctrl/dra.h
>>> @@ -40,8 +40,8 @@
>>>  
>>>  /* Active pin states */
>>>  #define PIN_OUTPUT		(0 | PULL_DIS)
>>> -#define PIN_OUTPUT_PULLUP	(PIN_OUTPUT | PULL_ENA | PULL_UP)
>>> -#define PIN_OUTPUT_PULLDOWN	(PIN_OUTPUT | PULL_ENA)
>>> +#define PIN_OUTPUT_PULLUP	(PULL_UP)
>>> +#define PIN_OUTPUT_PULLDOWN	(0)
>>>  #define PIN_INPUT		(INPUT_EN | PULL_DIS)
>>>  #define PIN_INPUT_SLEW		(INPUT_EN | SLEWCONTROL)
>>>  #define PIN_INPUT_PULLUP	(PULL_ENA | INPUT_EN | PULL_UP)
>>
>> You are right, we do have an issue with using PIN_OUTPUT along with
>> remaining setting.
>>
>> For a moment, I wondered why input was not impacted - then I realized
>> that INPUT_EN was being used instead of PIN_INPUT - following that
>> convention. With the intent being explicitly using macros that
>> clearly indicate what each setting combination is is, how about the
>> following?
>>
>>
>> diff --git a/include/dt-bindings/pinctrl/dra.h b/include/dt-bindings/pinctrl/dra.h
>> index 3d33794..d4037e7 100644
>> --- a/include/dt-bindings/pinctrl/dra.h
>> +++ b/include/dt-bindings/pinctrl/dra.h
>> @@ -34,14 +34,15 @@
>>  #define PULL_DIS		(1 << 16)
>>  #define PULL_UP			(1 << 17)
>>  #define INPUT_EN		(1 << 18)
>> +#define OUTPUT_EN		(0 << 18)
>>  #define SLEWCONTROL		(1 << 19)
>>  #define WAKEUP_EN		(1 << 24)
>>  #define WAKEUP_EVENT		(1 << 25)
>>  
>>  /* Active pin states */
>> -#define PIN_OUTPUT		(0 | PULL_DIS)
>> -#define PIN_OUTPUT_PULLUP	(PIN_OUTPUT | PULL_ENA | PULL_UP)
>> -#define PIN_OUTPUT_PULLDOWN	(PIN_OUTPUT | PULL_ENA)
>> +#define PIN_OUTPUT		(OUTPUT_EN | PULL_DIS)
>> +#define PIN_OUTPUT_PULLUP	(OUTPUT_EN | PULL_ENA | PULL_UP)
>> +#define PIN_OUTPUT_PULLDOWN	(OUTPUT_EN | PULL_ENA)
> 
> To me it adds more confusion and this change is a NOP as we're ORing 0 here
> with OUTPUT_EN.

look at this this way:

PIN_OUTPUT_PULLDOWN	(OUTPUT_EN | PULL_ENA)

should probably trigger in mind (what about PULLDOWN?)

PIN_OUTPUT_PULLDOWN	(OUTPUT_EN | PULL_ENA | PULL_DOWN)

then verify values of each OUTPUT_EN  -> 0 in bit 18, ok, etc.


if we ensure that PIN_XX macros use just the basic primitives, it is
easier to prevent the mistake like the one I made. the other option of
not defining macros whose values are 0 implies that the reviewer has
to recheck against trm to ensure all the right "1" bits are set.


just my view here.

> 
>>  #define PIN_INPUT		(INPUT_EN | PULL_DIS)
>>  #define PIN_INPUT_SLEW		(INPUT_EN | SLEWCONTROL)
>>  #define PIN_INPUT_PULLUP	(PULL_ENA | INPUT_EN | PULL_UP)
>>
> 
> cheers,
> -roger
> 


-- 
Regards,
Nishanth Menon