On Tue, Apr 15, 2025 at 12:10:06PM +0300, Dmitry Baryshkov wrote:
> On 14/04/2025 17:52, Maxime Ripard wrote:
> > Hi,
> >
> > On Mon, Apr 07, 2025 at 06:11:00PM +0300, Dmitry Baryshkov wrote:
> > > +/**
> > > + * struct drm_connector_cec - DRM Connector CEC-related structure
> > > + */
> > > +struct drm_connector_cec {
> > > + /**
> > > + * @mutex: protects all fields in this structure.
> > > + */
> > > + struct mutex mutex;
> > > +
> > > + /**
> > > + * @funcs: CEC Control Functions
> > > + */
> > > + const struct drm_connector_cec_funcs *funcs;
> > > +
> > > + /**
> > > + * @data: CEC implementation-specific data
> > > + */
> > > + void *data;
> >
> > Is there a reason we don't just skip that data? The only user I'm seeing
> > so far are the helpers, and they only put the cec_adapter pointer in
> > there.
> >
> > Can't we pass the connector to CEC and make the adapter part of
> > drm_connector_cec?
>
> It will be either cec_notifier or cec_adapter +
> drm_connector_hdmi_cec_funcs. Initially I sketched a union here, but then I
> thought that a void pointer makes more sense. It allows us to make CEC data
> helper-specific. For example, cec-pin might store platform callbacks here.
> DP CEC might need to store AUX pointer, etc.Ah I see, that makes sense. Thanks! Maxime
signature.asc
Description: PGP signature
