Hello, I'm currently trying to get the TC9595 (tc358767.c) running on my imx93 platform. For probing the EDID/DPCD information I need to put the DSI lanes into LP-11 (stop) state. For this I was trying to call drm_atomic_bridge_chain_pre_enable() inside tc_connector_get_modes(), if bridge is not enabled not enabled yet. Please refer to [1] why pre_enable has to be called. But at that point connector->state->state is NULL, while connector->state has some state. Thus drm_atomic_bridge_call_pre_enable(), called from drm_atomic_bridge_chain_pre_enable(), does nothing. But this seems bogus to me, how can the backpointer to drm_atomic_state be NULL if the drm_connector_state from drm_connector is still valid? Maybe I do not fully grasp the data organization and lifetime to it's fullest yet. Maybe someone has a suggestion how to pre_enable the bridge chain while get_modes is called. Thanks and best regards, Alexander [1] https://lore.kernel.org/all/20221205173328.1395350-7-dave.stevenson@xxxxxxxxxxxxxxx/ -- TQ-Systems GmbH | Mühlstraße 2, Gut Delling | 82229 Seefeld, Germany Amtsgericht München, HRB 105018 Geschäftsführer: Detlef Schneider, Rüdiger Stahl, Stefan Schneider http://www.tq-group.com/
