On Fri, Jul 31, 2026 at 07:19:28PM +0300, Cristian Ciocaltea wrote: > In preparation for supporting SCDC status synchronization on hotplug > events, extend drm_atomic_helper_connector_hdmi_hotplug() and > drm_atomic_helper_connector_hdmi_update() to take a struct > drm_modeset_acquire_ctx argument. This will be used when a CRTC reset > is required. > > Additionally, change their return type from void to int to allow > propagation of errors such as -EDEADLK caused by lock contention. > > Update existing callers accordingly. Note that in the VC4 case, the > return code is ignored, as it cannot fail: being in the .detect_ctx() > call path, the connection mutex is already held through ctx, and it > bails out early because the driver relies on a local SCDC > state-management implementation, and therefore does not yet expose the > scrambler capability through the common helpers. > > Similarly, introduce drm_atomic_helper_connector_hdmi_force_ctx() as the > HDMI connector implementation of the .force_ctx() callback. Since > migrating from the .force() callback is somewhat more involved, its > users will be migrated incrementally, after which > drm_atomic_helper_connector_hdmi_force() will be removed.
The force_ctx introduction should be a separate patch. Once split, you have (for both patches) Reviewed-by: Maxime Ripard <[email protected]> Maxime
signature.asc
Description: PGP signature
