On Sun, Feb 17, 2019 at 10:57:26PM +0100, Marek Vasut wrote: > On 2/17/19 9:00 PM, Harald Geyer wrote:
> > If this is the case, why would we need to set an initial state instead > > of just waiting what the first consumer requests. I still don't see > > a use case (aside from don't accidentally shut down something important > > during probing). > Possibly to prevent a state which might be harmful until the first > consumer comes up. However this is rather theoretical. This is the GPIO API restrictions again - there was a big push at one point to try to ensure that we set the output state when requesting output GPIOs, and a lot of the time requesting will set a state anyway when it puts things into output mode so it may as wel be one we know about.
Description: PGP signature