On Thu, Mar 19, 2020 at 12:07:37PM +0200, Jani Nikula wrote:
On Tue, 17 Mar 2020, Manasi Navare manasi.d.navare@intel.com wrote:
This patch adds a hook in drm_connector_helper_funcs to get the support of the driver for adaptive sync functionality.
This can be called in the connector probe helper function after the connector detect() and get_modes() hooks to also query the adaptive sync support of the driver.
I can obviously see that from the patch. But this does not explain at all *why* we need another hook to begin with, and why it neeeds to be called from ->fill_modes that is set to drm_helper_probe_single_connector_modes().
This needs to be called after get_modes that ends up populating the monitor range after parsing the EDID.
I could have just modified the get_modes hook in the driver to set the vrr capabilities but that doesnt go with the definition of get_modes() hook which is purely to obtain modes from edid.
So i added a separate hook which will always return the vrr capabilities if this hook exists in the driver everytime a connector probe happens.
Would you suggest this elsewhere or some other design? I am open to better placement of this hook or reusing some other hook or other suggestions for getting the vrr capabilities in the driver
Manasi
Cc: Jani Nikula jani.nikula@linux.intel.com Cc: Ville Syrjälä ville.syrjala@linux.intel.com Cc: Harry Wentland harry.wentland@amd.com Cc: Nicholas Kazlauskas Nicholas.Kazlauskas@amd.com Signed-off-by: Manasi Navare manasi.d.navare@intel.com
drivers/gpu/drm/drm_probe_helper.c | 4 ++++ include/drm/drm_modeset_helper_vtables.h | 16 ++++++++++++++++ 2 files changed, 20 insertions(+)
diff --git a/drivers/gpu/drm/drm_probe_helper.c b/drivers/gpu/drm/drm_probe_helper.c index 576b4b7dcd89..4403817bfb02 100644 --- a/drivers/gpu/drm/drm_probe_helper.c +++ b/drivers/gpu/drm/drm_probe_helper.c @@ -482,6 +482,10 @@ int drm_helper_probe_single_connector_modes(struct drm_connector *connector,
count = (*connector_funcs->get_modes)(connector);
- /* Get the Adaptive Sync Support if helper exists */
- if (*connector_funcs->get_adaptive_sync_support)
(**connector_funcs->get_adaptive_sync_support)(connector);This is in the middle of a sequence figuring out the modes. First ->get_modes, then fallback to other mechanisms. Certainly we don't want to do something else in the middle.
/* * Fallback for when DDC probe failed in drm_get_edid() and thus skipped * override/firmware EDID. diff --git a/include/drm/drm_modeset_helper_vtables.h b/include/drm/drm_modeset_helper_vtables.h index 7c20b1c8b6a7..0b203fdd25df 100644 --- a/include/drm/drm_modeset_helper_vtables.h +++ b/include/drm/drm_modeset_helper_vtables.h @@ -1079,6 +1079,22 @@ struct drm_connector_helper_funcs { struct drm_writeback_job *job); void (*cleanup_writeback_job)(struct drm_writeback_connector *connector, struct drm_writeback_job *job);
- /**
* @get_adaptive_sync_support:** This hook is used by the probe helper to get the driver's support* for adaptive sync or variable refresh rate.* This is called from drm_helper_probe_single_connector_modes()* This is called after the @get_modes hook so that the connector modes* are already obtained and EDID is parsed to obtain the monitor* range descriptor information.** This hook is optional and defined only for the drivers and on* connectors that advertise adaptive sync support.**/- void (*get_adaptive_sync_support)(struct drm_connector *connector);
};
/**
-- Jani Nikula, Intel Open Source Graphics Center