28 Sep
2023
28 Sep
'23
2:47 p.m.
On Thu, Sep 28, 2023 at 01:34:56PM +0100, Robin Murphy wrote:
On 27/09/2023 11:47 pm, Linus Walleij wrote:
- As the probe() code is very insistent on getting valid GPIOs out of the device before it will continue, there is no point to carry all the code handling the GPIOs as optional, drop all these checks.
Isn't it allowing them to be optional as long as of_get_named_gpio() returns -ENODEV (which I guess may come out of the chip->of_xlate callback)? Or is it implied that that's never actually been able to happen?
Right, I *think* it's just trying to open code optional GPIO support (possibly it predates the core helpers?).