On Fri, 12 May 2023 at 21:03, Stephen Boyd <swboyd@xxxxxxxxxxxx> wrote: > > Quoting Dmitry Baryshkov (2023-05-11 17:54:19) > > On Fri, 12 May 2023 at 03:16, Kuogee Hsieh <quic_khsieh@xxxxxxxxxxx> wrote: > > > 1) DP with GPIO: No downstream drm_bridge are connected, is_edp = false > > > and internal HPD-logic is in used (internal_hpd = true). Power needs to > > > be on at all times etc. > > > > > > 2) DP without GPIO: Downstream drm_bridge connected, is_edp = false and > > > internal HPD-logic should not be used/enabled (internal_hpd = false). > > > Power doesn't need to be on unless hpd_notify is invoked to tell us that > > > there's something connected... > > > > > > - dp_bridge_hpd_notify(). What is the point of this check? <== i have > > > below two questions, > > > > > > 1) can you explain when/what this dp_bridge_hpd_notify() will be called? > > > > The call chain is drm_bridge_hpd_notify() -> > > drm_bridge_connector_hpd_notify() -> .hpd_notify() for all drm_bridge > > in chain > > > > One should add a call to drm_bridge_hpd_notify() when the hotplug > > event has been detected. > > > > Also please note the patch https://patchwork.freedesktop.org/patch/484432/ > > > > > > > > 2) is dp_bridge_hpd_notify() only will be called at above case #2? and > > > it will not be used by case #1? > > > > Once the driver calls drm_bridge_hpd_notify() in the hpd path, the > > hpd_notify callbacks will be called in case#1 too. > > > > BTW: I don't see drm_bridge_hpd_notify() or > > drm_kms_{,connector_}_hotplug_event() HPD notifications in the DP > > driver at all. This should be fixed. > > > > Is dp_bridge_hpd_notify() being called by > drm_helper_probe_single_connector_modes() when the connectors are > detected? > > I see drm_helper_probe_detect() calls connector->funcs->detect() which I > think calls > drm_bridge_connector_funcs::drm_bridge_connector_hpd_notify() but I > haven't confirmed yet. The 'detect' bridge is the DP bridge in msm > driver > > if (!dp_display->is_edp) { > bridge->ops = > DRM_BRIDGE_OP_DETECT | > > so if the bridge_connector is being used then I think when fill_modes() > is called we'll run the detect cycle and call the hpd_notify callbacks > on the bridge chain. Yes. This call chain is correct. drm_helper_probe_single_connector_modes() -> drm_bridge_connector_detect() -> drm_bridge_connector_hpd_notify(). However on HPD events the DP driver doesn't call into the drm core (which I believe should be fixed). -- With best wishes Dmitry
