Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <CAK8P3a0i0WP24Z0TScmPqKxmM2ovtKnmm+qZq6+Tc1ju+hma0w@mail.gmail.com>
 <20210804141049.499767-1-kherbst@redhat.com> <CAK8P3a136c_L3yVn-841Sbfib9UMOf1M-pk+2SqWt0wD2zfRKQ@mail.gmail.com>
 <CACO55tsLpURTm=Jf=4gRVtYQbit5h2OBYw_MFb6Vf1PFvTV7dw@mail.gmail.com> <CACO55tuy5Am9zbcR490KWYYAg7MguBN5m82vbjzifGN5KpGbxw@mail.gmail.com>
In-Reply-To: <CACO55tuy5Am9zbcR490KWYYAg7MguBN5m82vbjzifGN5KpGbxw@mail.gmail.com>
From:   Arnd Bergmann <arnd@kernel.org>
Date:   Wed, 4 Aug 2021 23:09:51 +0200
Message-ID: <CAK8P3a3hZ7X5+kM5E+_Y+COUp49Kt6iDjiqMFtimiSbPk4byzQ@mail.gmail.com>
Subject: Re: [PATCH] depend on BACKLIGHT_CLASS_DEVICE for more devices
To:     Karol Herbst <kherbst@redhat.com>
Cc:     Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        Lyude Paul <lyude@redhat.com>, Ben Skeggs <bskeggs@redhat.com>,
        Randy Dunlap <rdunlap@infradead.org>,
        Daniel Vetter <daniel@ffwll.ch>,
        ML nouveau <nouveau@lists.freedesktop.org>,
        dri-devel <dri-devel@lists.freedesktop.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Wed, Aug 4, 2021 at 8:59 PM Karol Herbst <kherbst@redhat.com> wrote:
> On Wed, Aug 4, 2021 at 4:43 PM Karol Herbst <kherbst@redhat.com> wrote:
> > On Wed, Aug 4, 2021 at 4:19 PM Arnd Bergmann <arnd@kernel.org> wrote:
> > > On Wed, Aug 4, 2021 at 4:10 PM Karol Herbst <kherbst@redhat.com> wrote:
> > > >
> > > > playing around a little bit with this, I think the original "select
> > > > BACKLIGHT_CLASS_DEVICE" is fine. Atm we kind of have this weird mix of
> > > > drivers selecting and others depending on it. We could of course convert
> > > > everything over to depend, and break those cycling dependency issues with
> > > > this.
> > > >
> > > > Anyway this change on top of my initial patch is enough to make Kconfig
> > > > happy and has the advantage of not having to mess with the deps of nouveau
> > > > too much.
> > >
> > > Looks good to me. We'd probably want to make the BACKLIGHT_CLASS_DEVICE
> > > option itself 'default FB || DRM' though, to ensure that defconfigs
> > > keep working.
> > >
> >
> > okay cool. Will send out a proper updated patch series soonish.
> >
>
> mhh, actually that breaks drivers selecting FB_BACKLIGHT as now
> BACKLIGHT_CLASS_DEVICE might be disabled :(

Are you sure? It should already be the case that any driver that selects
FB_BACKLIGHT either 'depends on BACKLIGHT_CLASS_DEVICE'
or 'select BACKLIGHT_CLASS_DEVICE'.

If you change all the 'select BACKLIGHT_CLASS_DEVICE' to 'depends
on', I don't see a problem with doing 'select FB_BACKLIGHT' from
those.

I have applied your patch to my randconfig tree and built a few dozen
kernels, don't see any regressions so far, but will let it run over night.

      Arnd