[U-Boot] [PATCH V2 2/2] sunxi: gpio: convert bind() to use driver data

[Simon Glass]

On 12 May 2016 at 20:42, Simon Glass <sjg at chromium.org> wrote:
> Hi Hans,
>
> On 12 May 2016 at 11:51, Hans de Goede <hdegoede at redhat.com> wrote:
>> Hi,
>>
>>
>> On 12-05-16 19:50, Stephen Warren wrote:
>>>
>>> On 05/12/2016 11:43 AM, Simon Glass wrote:
>>>>
>>>> Hi Stephen,
>>>>
>>>> On 11 May 2016 at 15:26, Stephen Warren <swarren at wwwdotorg.org> wrote:
>>>>>
>>>>> From: Stephen Warren <swarren at nvidia.com>
>>>>>
>>>>> Now that the DM core sets driver_data before calling bind(), this driver
>>>>> can make use of driver_data to determine the set of child devices to
>>>>> create, rather than manually re-implementing the matching logic in code.
>>>
>>>
>>>>> diff --git a/drivers/gpio/sunxi_gpio.c b/drivers/gpio/sunxi_gpio.c
>>>
>>>
>>>>> +#define ID(_compat_, _soc_data_) \
>>>>> +       { .compatible = _compat_, .data = (ulong)&soc_data_##_soc_data_
>>>>> }
>>>>> +
>>>>>   static const struct udevice_id sunxi_gpio_ids[] = {
>>>>> -       { .compatible = "allwinner,sun4i-a10-pinctrl" },
>>>>> -       { .compatible = "allwinner,sun5i-a10s-pinctrl" },
>>>>> -       { .compatible = "allwinner,sun5i-a13-pinctrl" },
>>>
>>> ...
>>>>>
>>>>> +       ID("allwinner,sun4i-a10-pinctrl",       a_all),
>>>>> +       ID("allwinner,sun5i-a10s-pinctrl",      a_all),
>>>>
>>>>
>>>> I don't think the #define adds a lot of value - consider removing it
>>>> an writing things out in full?
>>>
>>>
>>> I originally did that, but you either end up with lines over 80 columns
>>> which checkpatch complains about, or multiple lines per entry which makes it
>>> harder to read for such a large table. Still, I can convert it if you want.
>>
>>
>> I'm fine with keeping the ID define:
>>
>> Reviewed-by: Hans de Goede <hdegoede at redhat.com>
>>
>> Simon, I assume that you will upstream this one through your
>> tree ?
>
> Yes, thanks.
>
> Regards,
> Simon

Applied to u-boot-dm, thanks!