[PATCH v2 1/4] iio: adc: rockchip_saradc: reset saradc controller before programming it

[Caesar Wang]

On 2016年07月27日 10:00, Guenter Roeck wrote:
> On 07/26/2016 05:42 PM, Caesar Wang wrote:
>>
>> On 2016年07月27日 01:00, Guenter Roeck wrote:
>>> On Tue, Jul 26, 2016 at 10:47:16PM +0800, Caesar Wang wrote:
>>>> On 2016年07月26日 21:39, Guenter Roeck wrote:
>>>>>> static int rockchip_saradc_probe(struct platform_device *pdev)
>>>>>> {
>>>>>> struct rockchip_saradc *info = NULL;
>>>>>> @@ -218,6 +231,21 @@ static int rockchip_saradc_probe(struct
>>>>>> platform_device *pdev)
>>>>>> if (IS_ERR(info->regs))
>>>>>> return PTR_ERR(info->regs);
>>>>>>
>>>>>> + /*
>>>>>> + * The reset should be an optional property, as it should work
>>>>>> + * with old devicetrees as well
>>>>>> + */
>>>>>> + info->reset = devm_reset_control_get_optional(&pdev->dev,
>>>>>> + "saradc-apb");
>>>>> Does anyone know what the _optional API is for ? It seems to be 
>>>>> exactly
>>>>> the same
>>>>> as devm_reset_control_get().
>>>> “
>>>> As far as I see, the difference is WARN_ON(1)
>>>> when  is not defined.
>>>>
>>>>
>>>> The _optional functions were introduced by the following commit:
>>>>
>>>> ----------------->8-----------------
>>>> commit b424080a9e086e683ad5fdc624a7cf3c024e0c0f
>>>> Author: Philipp Zabel <p.zabel at xxxxxxxxxxxxxx>
>>>> Date: Fri Mar 7 15:18:47 2014 +0100
>>>>
>>>> reset: Add optional resets and stubs
>>>>
>>>> This patch adds device_reset_optional and
>>>> variants that drivers can use to indicate they can function without 
>>>> control
>>>> over the reset line. For those functions, stubs are added so the 
>>>> drivers can
>>>> be compiled with CONFIG_RESET_CONTROLLER disabled.
>>>> Also, device_reset is annotated with __must_check. Drivers
>>>> ignoring the return
>>>> value should use device_reset_optional instead.
>>>>
>>> Is that really what we are looking for here ? CONFIG_RESET_CONTROLLER
>>> is required for other functions of rk3399, isn't it ?
>>
>> Right, as the DRM and thermal are depend on the 
>> CONFIG_RESET_CONTROLLER .....
>>
>
> Since "optional" doesn't really mean "the reset property is optional"
> but "CONFIG_RESET_CONTROLLER is optional", I would suggest to use
> devm_reset_control_get().

Agree, free riding.
Maybe the API (_optional) will be changed in later.
As someone make a point present an idea on 
http://www.spinics.net/lists/kernel/msg2306677.html

Okay,  devm_reset_control_get() will be better, anyway the driver has 
been depend on the CONFIG_RESET_CONTROLLER.
--

I will send a new patch for upstream if nobody object it on today.


Sorry for noisy!


>
> Guenter
>
>
> _______________________________________________
> Linux-rockchip mailing list
> Linux-rockchip at lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-rockchip


-- 
caesar wang | software engineer | wxt at rock-chip.com