On Wed, Jun 8, 2022 at 10:17 PM Doug Anderson <dianders@xxxxxxxxxxxx> wrote:
>
> Hi,
>
> On Wed, Jun 8, 2022 at 2:48 AM Hsin-Yi Wang <hsinyi@xxxxxxxxxxxx> wrote:
> >
> > @@ -269,6 +280,31 @@ void drm_panel_bridge_remove(struct drm_bridge *bridge)
> > }
> > EXPORT_SYMBOL(drm_panel_bridge_remove);
> >
> > +/**
> > + * drm_panel_bridge_set_orientation - Set the connector's panel orientation
> > + * if the bridge is a panel bridge.
> > + *
> > + * @connector: The connector to be set panel orientation.
> > + * @bridge: The drm_bridge to be transformed to panel bridge.
>
> Ideally you should have a kernel doc to describe what you're returning.
>
>
> > + */
> > +int drm_panel_bridge_set_orientation(struct drm_connector *connector,
> > + struct drm_bridge *bridge)
> > +{
> > + struct panel_bridge *panel_bridge;
> > +
> > + if (!bridge)
> > + return 0;
> > +
> > + if (bridge->funcs != &panel_bridge_bridge_funcs)
> > + return 0;
>
> nit: Why do you need to handle NULL bridge and the case that someone
> calls you with something other than a panel-bridge? I'm not convinced
> that's useful. In general kernel style doesn't do massive validation
> of parameters unless there's a reason for it. ...if we do need to
> handle them then it feels like they should be returning -EINVAL or
> something, not 0.
The only caller had checked it. I can remove the check here.
>
>
> > @@ -917,10 +917,13 @@ void drm_bridge_hpd_notify(struct drm_bridge *bridge,
> > enum drm_connector_status status);
> >
> > #ifdef CONFIG_DRM_PANEL_BRIDGE
> > +bool drm_bridge_is_panel(const struct drm_bridge *bridge);
> > struct drm_bridge *drm_panel_bridge_add(struct drm_panel *panel);
> > struct drm_bridge *drm_panel_bridge_add_typed(struct drm_panel *panel,
> > u32 connector_type);
> > void drm_panel_bridge_remove(struct drm_bridge *bridge);
> > +int drm_panel_bridge_set_orientation(struct drm_connector *connector,
> > + struct drm_bridge *bridge);
>
> I suspect that you need some dummy versions of your new functions
> defined if "CONFIG_DRM_PANEL_BRIDGE" is not defined. Otherwise we're
> going to be yelled at by the kernel robot eventually.
>
> -Doug
