On Thu, Nov 22, 2018 at 5:26 PM Paul Cercueil <paul@xxxxxxxxxxxxxxx> wrote: > I think the driver should just call mmc_regulator_get_supply() in the > probe function, > and then call mmc_regulator_set_ocr() to power ON or OFF the card. The > board file would > then provide the power regulator that triggers the GPIO line. One does not exclude the other, feel free (or anyone else) to switch this to a fixed regulator. (Using descriptors!) I feel insecure about cold-coding that without being able to test it, I think I am on deep waters already. Someone with the proper hardware should do radical semantic patches like that. Introducing the regulator abstraction can lead to probe deferrals and whatnot. Yours, Linus Walleij
