On Thu, May 14, 2026 at 01:24:19PM -0500, Aaron Kling wrote:
> On Thu, May 14, 2026 at 1:05 PM Conor Dooley <conor@kernel.org> wrote:
> >
> > On Thu, May 14, 2026 at 12:24:50PM -0500, Aaron Kling via B4 Relay wrote:
> >
> > > +static const struct of_device_id icna35xx_of_match[] = {
> > > +     { .compatible = "ayaneo,pocketds-panel-top", .data = &odin2portal_desc },
> > > +     { .compatible = "ayntec,odin2portal-panel", .data = &odin2portal_desc },
> > > +     { .compatible = "ayntec,odin3-panel", .data = &thor_top_desc },
> > > +     { .compatible = "ayntec,thor-panel-top", .data = &thor_top_desc },
> >
> > This looks like you could be using the fallback compatibles instead of
> > matching on the panel's. Why aren't you?
> 
> I'm uncertain if the panels are actually the same between these
> devices or if the init sequences are just similar enough to share. I
> am told that the thor top panel and the odin 3 panel are the same
> model. But I don't know about the portal and the pocket ds, just know
> that they both fire up and work as expected with the same init
> sequence. I'm trying to follow the bindings idea that these might be
> different physical panels that only share a ddic and thus driver
> handling.

That's a reasonable rationale. Could you include this in your commit
message please?

> 
> > > +     { /* sentinel */ }
> 
> Aaron