Re: [PATCH v4 2/8] drm/rockchip: vop2: Switch impossible pos conditional to WARN_ON

[Daniel Stone]

Hi,

On Thu, 11 Dec 2025 at 22:41, David Laight <david.laight.linux@xxxxxxxxx> wrote:
> On Thu, 11 Dec 2025 21:40:32 +0100
> Nicolas Frattaroli <nicolas.frattaroli@xxxxxxxxxxxxx> wrote:
> > -     if (dest->y1 + dsp_h > adjusted_mode->vdisplay) {
> > -             drm_dbg_kms(vop2->drm,
> > -                         "vp%d %s dest->y1[%d] + dsp_h[%d] exceed mode vdisplay[%d]\n",
> > -                         vp->id, win->data->name, dest->y1, dsp_h, adjusted_mode->vdisplay);
> > -             dsp_h = adjusted_mode->vdisplay - dest->y1;
> > -             if (dsp_h < 4)
> > -                     dsp_h = 4;
> > -             src_h = dsp_h * src_h / drm_rect_height(dest);
> > -     }
> > +     /* drm_atomic_helper_check_plane_state calls drm_rect_clip_scaled for
> > +      * us, which keeps our planes bounded within the CRTC active area
> > +      */
> > +     WARN_ON(dest->x1 + dsp_w > adjusted_mode->hdisplay);
> > +     WARN_ON(dest->y1 + dsp_h > adjusted_mode->vdisplay);
> > +     WARN_ON(dsp_w < 4);
> > +     WARN_ON(dsp_h < 4);
> > +     WARN_ON(src_w < 4);
> > +     WARN_ON(src_h < 4);
>
> You need to do something when the tests fail.
> Carrying on regardless is never right.

When we arrive at this point, because the load-bearing assumption in
the comment has not been met, our options would be:
* display random incorrect content, and hope that we aren't reading
out of bounds from a buffer that's too small
* turn the display off
* there is no third option

Which are you suggesting?

Cheers,
Daniel