MIME-Version: 1.0
References: <20230301185432.3010939-1-robdclark@gmail.com> <6eb2cde8-f548-73ba-6091-131c1848690c@collabora.com>
 <20230302072531.375i6xetk72nis75@sirius.home.kraxel.org>
In-Reply-To: <20230302072531.375i6xetk72nis75@sirius.home.kraxel.org>
From:   Rob Clark <robdclark@gmail.com>
Date:   Thu, 2 Mar 2023 15:33:07 -0800
Message-ID: <CAF6AEGtzdty5m9ccRZqYNN=+OCa0KogNQxR41AD-LWC5fjbujw@mail.gmail.com>
Subject: Re: [PATCH v5] drm/virtio: Add option to disable KMS support
To:     Gerd Hoffmann <kraxel@redhat.com>
Cc:     Dmitry Osipenko <dmitry.osipenko@collabora.com>,
        dri-devel@lists.freedesktop.org, Chia-I Wu <olvaffe@gmail.com>,
        Ryan Neph <ryanneph@chromium.org>,
        Daniel Vetter <daniel@ffwll.ch>,
        Rob Clark <robdclark@chromium.org>,
        Javier Martinez Canillas <javierm@redhat.com>,
        David Airlie <airlied@redhat.com>,
        Gurchetan Singh <gurchetansingh@chromium.org>,
        "open list:VIRTIO GPU DRIVER" 
        <virtualization@lists.linux-foundation.org>,
        open list <linux-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Wed, Mar 1, 2023 at 11:25 PM Gerd Hoffmann <kraxel@redhat.com> wrote:
>
> On Thu, Mar 02, 2023 at 12:39:33AM +0300, Dmitry Osipenko wrote:
> > On 3/1/23 21:54, Rob Clark wrote:
> > >  /* virtgpu_display.c */
> > > +#if defined(CONFIG_DRM_VIRTIO_GPU_KMS)
> > >  int virtio_gpu_modeset_init(struct virtio_gpu_device *vgdev);
> > >  void virtio_gpu_modeset_fini(struct virtio_gpu_device *vgdev);
> > > +#else
> > > +static inline int virtio_gpu_modeset_init(struct virtio_gpu_device *vgdev)
> > > +{
> > > +   return 0;
> > > +}
> > > +static inline void virtio_gpu_modeset_fini(struct virtio_gpu_device *vgdev)
> > > +{
> > > +}
> > > +#endif
> >
> > In v4 Gerd wanted to keep building the virtgpu_display.o and instead add
> > the KSM config check to virtio_gpu_modeset_init/fini().
>
> The main point is that the code workflow should be the same in both
> cases.  The patch does that for virtio_gpu_modeset_init() but doesn't
> for virtio_gpu_modeset_fini().
>
> Return early in the functions (and drop the #ifdef here) is how I would
> implement this, but I wouldn't insist on that, there are other ways to
> solve this too ;)

Ahh, true, I guess omitting that one file doesn't save anything and
early return makes for a bit simpler/smaller patch

BR,
-R