On Wed, Jul 30, 2014 at 7:42 PM, Benjamin Gaignard <benjamin.gaignard@xxxxxxxxxx> wrote: > @@ -87,11 +90,50 @@ static int sti_compositor_bind(struct device *dev, struct device *master, > struct sti_compositor *compo = dev_get_drvdata(dev); > struct drm_device *drm_dev = data; > unsigned int i, crtc = 0, plane = 0; > + struct sti_drm_private *dev_priv = drm_dev->dev_private; > + struct drm_plane *cursor = NULL; > + struct drm_plane *primary = NULL; > + > + dev_priv->compo = compo; > > drm_vblank_init(drm_dev, crtc); This looks strange - you should pass this the total number of crtcs (the same that eventually ends up in dev->mode_config.num_crtc), not 0. And the assignement of cursors to crtcs looks a bit strange on first read-through, but I have no clue about the sti hw. And in any case those pointers really only matter for backwards compat with existing pageflip and cursor ioctls, so doesn't really matter too much. Anyway didn't spot anything else which would need to be upgrade to never kms interfaces, so ack from my side for that. Only looked at that since right now I'm refreshing drm docs in those areas ;-) -Daniel -- Daniel Vetter Software Engineer, Intel Corporation +41 (0) 79 365 57 48 - http://blog.ffwll.ch _______________________________________________ dri-devel mailing list dri-devel@xxxxxxxxxxxxxxxxxxxxx http://lists.freedesktop.org/mailman/listinfo/dri-devel
