On 26/09/2017 15:27, Maxime Ripard wrote:
> On Tue, Sep 26, 2017 at 01:08:21PM +0000, Quentin Schulz wrote:
>> Hi Maxime,
>>
>> On 26/09/2017 15:00, Maxime Ripard wrote:
>>> On Tue, Sep 26, 2017 at 12:17:12PM +0000, Quentin Schulz wrote:
>>>> +static const struct axp20x_desc_pin axp209_pins[] = {
>>>> + AXP20X_PIN(AXP20X_PINCTRL_PIN(0, "GPIO0"),
>>>> + AXP20X_FUNCTION(0x0, "gpio_out"),
>>>> + AXP20X_FUNCTION(0x2, "gpio_in"),
>>>> + AXP20X_FUNCTION(0x3, "ldo"),
>>>> + AXP20X_FUNCTION(0x4, "adc")),
>>>> + AXP20X_PIN(AXP20X_PINCTRL_PIN(1, "GPIO1"),
>>>> + AXP20X_FUNCTION(0x0, "gpio_out"),
>>>> + AXP20X_FUNCTION(0x2, "gpio_in"),
>>>> + AXP20X_FUNCTION(0x3, "ldo"),
>>>> + AXP20X_FUNCTION(0x4, "adc")),
>>>> + AXP20X_PIN(AXP20X_PINCTRL_PIN(2, "GPIO2"),
>>>> + AXP20X_FUNCTION(0x0, "gpio_out"),
>>>> + AXP20X_FUNCTION(0x2, "gpio_in")),
>>>> +};
>>>
>>> If all the functions are the same, and at the same offset, can't we
>>> just hardcode it, instead of having (and duplicate) all the logic
>>> below?
>>>
>>
>> AXP20X_PIN(AXP20X_PINCTRL_PIN(0, "GPIO0"),
>> AXP20X_GPIO_OUT,
>> AXP20X_GPIO_IN,
>> AXP20X_LDO,
>> AXP20X_ADC))
>>
>> That's what you mean?
>
> What I mean is:
>
> static int axp20x_get_func(char *func)
> {
> if (!strcmp(func, "gpio_out"))
> return 0;
>
> if (!strcmp(func, "gpio_in"))
> return 2;
>
> if (!strcmp(func, "ldo"))
> return 3;
>
> if (!strcmp(func, "adc"))
> return 4;
>
> return -EINVAL;
> }
>
GPIO2 on AXP209 does not support ldo nor adc.
GPIO1 on AXP813 does not support adc.
I find it more complex to handle those two cases in a function than by
hardcoding it in structures like above.
Moreover, nothing tells us that it would be the same offset for other PMICs.
Quentin
--
Quentin Schulz, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com
Attachment:
signature.asc
Description: OpenPGP digital signature
