Re: [tip:irq/core] gpio: mb86s7x: Enable ACPI support

[Ard Biesheuvel]

On Fri, 5 Jul 2019 at 10:58, Linus Walleij <linus.walleij@xxxxxxxxxx> wrote:
>
> On Thu, Jul 4, 2019 at 8:18 PM Ard Biesheuvel <ard.biesheuvel@xxxxxxxxxx> wrote:
> > On Thu, 4 Jul 2019 at 09:52, Linus Walleij <linus.walleij@xxxxxxxxxx> wrote:
> > >
> > > On Wed, Jul 3, 2019 at 3:50 PM Marc Zyngier <marc.zyngier@xxxxxxx> wrote:
> > > > On 03/07/2019 13:26, Linus Walleij wrote:
> > > > > On Wed, Jul 3, 2019 at 11:24 AM tip-bot for Ard Biesheuvel
> > > > > <tipbot@xxxxxxxxx> wrote:
> > > > >
> > > > >> Committer:  Marc Zyngier <marc.zyngier@xxxxxxx>
> > > > >> CommitDate: Wed, 29 May 2019 10:42:19 +0100
> > > > >>
> > > > >> gpio: mb86s7x: Enable ACPI support
> > > > >>
> > > > >> Make the mb86s7x GPIO block discoverable via ACPI. In addition, add
> > > > >> support for ACPI GPIO interrupts routed via platform interrupts, by
> > > > >> wiring the two together via the to_irq() gpiochip callback.
> > > > >>
> > > > >> Reviewed-by: Linus Walleij <linus.walleij@xxxxxxxxxx>
> > > > >> Reviewed-by: Mika Westerberg <mika.westerberg@xxxxxxxxxxxxxxx>
> > > > >> Signed-off-by: Ard Biesheuvel <ard.biesheuvel@xxxxxxxxxx>
> > > > >> Signed-off-by: Marc Zyngier <marc.zyngier@xxxxxxx>
> > > > >
> > > > > OK!
> > > > >
> > > > >> +#include "gpiolib.h"
> > > > >> +
> > > > >
> > > > > But this isn't needed anymore, is it?
> > > >
> > > > You tell me! ;-)
> > > >
> > > > > I can try to remember to remove it later though.
> > > >
> > > > Yeah, please send a separate patch. tip is stable, and we can't roll
> > > > this back.
> > >
> > > I'll just fix it in the GPIO tree after -rc1.
> > > Made a personal TODO note!
> >
> > Why wouldn't it be needed anymore?
>
> I looked over the code like 5 times and I can't see it touching any
> gpiolib internals anymore, but maybe I'm still wrong?
>
> Normally GPIO drivers should get all symbols it needs from
> <linux/gpio/driver.h> and "gpiolib.h" is not something drivers
> should include.
>
> Anyways I will get to know from compile tests if I'm wrong.
>

IIRC it had something to do with the ACPI helpers that this driver
calls directly.