On Fri, Jun 26, 2026 at 3:00 PM Maxime Ripard <[email protected]> wrote: > > On Fri, Jun 26, 2026 at 02:03:27PM +0200, Albert Esteve wrote: > > of_drm_find_panel() and drm_of_find_panel_or_bridge() now return a > > counted reference. In drivers that immediately wrap the panel in a > > bridge via devm_drm_panel_bridge_add() or equivalent, the bridge > > acquires its own reference, so the caller's lookup reference must be > > released right afterwards. > > > > Also handle the cases where a panel is found but cannot be used, > > dropping the reference immediately in those paths. > > > > Assisted-by: Claude:claude-opus-4-6 > > Signed-off-by: Albert Esteve <[email protected]> > > drm_of_find_panel_or_bridge() does indeed return a refcounted pointer > now, but afaik the doc wasn't updated to reflect that.
True, I'll fix that in the next version. > > More importantly, I feel like with both of_drm_find_panel and > drm_of_find_panel_or_bridge we update a path that is considered legacy > anyway now, and we should rather focus on providing a safe alternative. Oh, I missed that this code path is considered legacy. > > But none of the functions you updated are unsafe, so it won't be more > unsafe, or provide any illusion of safety to the caller. Idk. > > Either way, this should all be on its way out if Luca creates a bridge > for every panel, and we'll consolidate on bridges only, so maybe it's > not such a big deal to merge this patch. I see. Given what you wrote, I think it'd make sense to correct them while this code isn't completely dead. BR, Albert. > > Maxime
