"Arnd Bergmann" <arnd@xxxxxxxx> writes:
> On Thu, Feb 1, 2024, at 14:17, Andy Shevchenko wrote:
>> On Thu, Feb 01, 2024 at 01:20:16PM +0100, Arnd Bergmann wrote:
>>> On Wed, Jan 31, 2024, at 23:37, Linus Walleij wrote:
>>
>>> + } else if (ah->led_pin < 0) {
>>
>> ...
>>
>>> + if (sc->sc_ah->led_gpio)
>>
>> Dup check
>
> I don't know what you mean here. To explain what I'm
> trying to do: The idea is that the LED is always backed
> by either gpiolib or the internal gpio controller on
> the PCI device. This means every access to an LED must
> be guarded with
>
> if (gpiodesc)
> gpio_*(gpiodesc);
> else
> internal(ah);
>
> We could probably go a little further in the cleanup and
> throw out the gpiolib path entirely, instead relying
> on the existing leds-gpio driver. Since there are currently
> no upstream users of the gpiolib path, that would likely
> lead to cleaner code but require more changes to any
> out-of-tree users that rely on the platform_data to
> pass the GPIOs today.
There being exactly one such out of tree user (per your up-thread
email) in OpenWrt? Or are you aware of others?
-Toke
