Re: [PATCH] hte: tegra194: add GPIOLIB dependency

Bartosz Golaszewski <[email protected]> Mon, 23 Oct 2023 14:02:06 +0200
Newsgroups dev.linux.lists.timestamp,org.kernel.vger.linux-kernel
Message-ID <CAMRc=Mc7PCEimVEJ_JYg=4+X5PAJ1Lh2SxXsNEFpuKctbEPiYg@mail.gmail.com>
On Mon, Oct 23, 2023 at 2:00=E2=80=AFPM Bartosz Golaszewski <[email protected]>=
 wrote:
>
> On Mon, Oct 23, 2023 at 1:57=E2=80=AFPM Arnd Bergmann <[email protected]> w=
rote:
> >
> > From: Arnd Bergmann <[email protected]>
> >
> > The driver started calling into a few interfaces that are part of GPIOL=
IB and
> > don't have stub implementations otherwise:
> >
> > drivers/hte/hte-tegra194.c: In function 'tegra_hte_line_xlate':
> > drivers/hte/hte-tegra194.c:424:48: error: implicit declaration of funct=
ion 'gpio_device_get_base'; did you mean 'gpio_device_get_desc'? [-Werror=
=3Dimplicit-function-declaration]
> >   424 |                 line_id =3D desc->attr.line_id - gpio_device_ge=
t_base(gs->gdev);
> >       |                                                ^~~~~~~~~~~~~~~~=
~~~~
> >       |                                                gpio_device_get_=
desc
> >
> > Add a Kconfig dependency to only allow building when this is defined.
> >
> > Fixes: dc850faa28ee0 ("hte: tegra194: don't access struct gpio_chip")
> > Signed-off-by: Arnd Bergmann <[email protected]>
> > ---
> >  drivers/hte/Kconfig | 1 +
> >  1 file changed, 1 insertion(+)
> >
> > diff --git a/drivers/hte/Kconfig b/drivers/hte/Kconfig
> > index 083e67492bf2e..641af722b555d 100644
> > --- a/drivers/hte/Kconfig
> > +++ b/drivers/hte/Kconfig
> > @@ -17,6 +17,7 @@ if HTE
> >  config HTE_TEGRA194
> >         tristate "NVIDIA Tegra194 HTE Support"
> >         depends on (ARCH_TEGRA_194_SOC || COMPILE_TEST)
> > +       depends on GPIOLIB
> >         help
> >           Enable this option for integrated hardware timestamping engin=
e also
> >           known as generic timestamping engine (GTE) support on NVIDIA =
Tegra194
> > --
> > 2.39.2
> >
>
> Thanks. None of the new interfaces are "consumer" APIs and are not
> meant to have stubs so depending on GPIOLIB is correct here.
>
> Reviewed-by: Bartosz Golaszewski <[email protected]>

Ah, this goes on top of the patch that is in my tree so I'll wait for
Dipen's Ack and apply it.

Bart