On Thu, Sep 14, 2023 at 10:30 AM Paul Cercueil <paul@xxxxxxxxxxxxxxx> wrote: > > Hi, > > Le jeudi 14 septembre 2023 à 10:02 +0300, Andy Shevchenko a écrit : > > On Wed, Sep 13, 2023 at 11:23 PM Miquel Raynal > > <miquel.raynal@xxxxxxxxxxx> wrote: > > > linus.walleij@xxxxxxxxxx wrote on Wed, 13 Sep 2023 22:12:40 +0200: > > > > On Wed, Sep 13, 2023 at 10:05 PM Andy Shevchenko > > > > <andy.shevchenko@xxxxxxxxx> wrote: > > > > > On Wed, Sep 13, 2023 at 2:50 PM Bartosz Golaszewski > > > > > <brgl@xxxxxxxx> wrote: > > > > ... > > > > > > > Why not moving this quirk to gpiolib-of.c? > > > > > > > > That's a better idea here I think, it's clearly a quirk for a > > > > buggy device tree. > > > > > > Agreed, it's just for backward compatibility purposes in a single > > > driver. I believe it should stay here. > > > > I believe Linus was for moving. > > Which Linus? Because the one who's also the gpio maintainer just wrote > above that it was better to keep it in the driver. > I'm also under the impression that Linus meant moving it to gpiolib-of.c. Let's Linus: Could you clarify? Bart > Cheers, > -Paul > > > > > gpiolib-of.c contains a lot of quirks, including this one. Calling > > these new (or old) APIs for overriding polarity in many cases > > shouldn't be needed if were no issues with DT or something like that. > > >
