On Thu, 2017-03-30 at 12:36 +0200, Maarten Lankhorst wrote:
Op 30-03-17 om 10:42 schreef Dhinakaran Pandiyan:
From: "Pandiyan, Dhinakaran" dhinakaran.pandiyan@intel.com
Having an ->atomic_release callback is useful to release shared resources that get allocated in compute_config(). This function is expected to be called in the atomic_check() phase before new resources are acquired.
v4: Document that the function is conditionally called and before other atomic_checks() (Daniel) v3: Use the new 'for_each_oldnew_connector_in_state()' macro. v2: Moved the caller hunk to this patch (Daniel)
Cc: Daniel Vetter daniel.vetter@ffwll.ch Cc: Maarten Lankhorst maarten.lankhorst@linux.intel.com Cc: Archit Taneja architt@codeaurora.org Cc: Chris Wilson chris@chris-wilson.co.uk Cc: Harry Wentland Harry.wentland@amd.com Reviewed-by: Maarten Lankhorst maarten.lankhorst@linux.intel.com Reviewed-by: Daniel Vetter daniel.vetter@ffwll.ch Suggested-by: Daniel Vetter daniel.vetter@ffwll.ch Signed-off-by: Dhinakaran Pandiyan dhinakaran.pandiyan@intel.com
drivers/gpu/drm/drm_atomic_helper.c | 19 +++++++++++++++++++ include/drm/drm_modeset_helper_vtables.h | 16 ++++++++++++++++ 2 files changed, 35 insertions(+)
diff --git a/drivers/gpu/drm/drm_atomic_helper.c b/drivers/gpu/drm/drm_atomic_helper.c index d14094d..9d07669 100644 --- a/drivers/gpu/drm/drm_atomic_helper.c +++ b/drivers/gpu/drm/drm_atomic_helper.c @@ -586,6 +586,25 @@ drm_atomic_helper_check_modeset(struct drm_device *dev, } }
- for_each_oldnew_connector_in_state(state, connector, old_connector_state, new_connector_state, i) {
const struct drm_connector_helper_funcs *conn_funcs;struct drm_crtc_state *crtc_state;conn_funcs = connector->helper_private;if (!conn_funcs->atomic_release)continue;if (!old_connector_state->crtc)continue;crtc_state = drm_atomic_get_existing_crtc_state(state, old_connector_state->crtc);if (crtc_state->connectors_changed ||crtc_state->mode_changed ||(crtc_state->active_changed && !crtc_state->active))conn_funcs->atomic_release(connector, new_connector_state);- }
- return mode_fixup(state);
}
Oops, I'm just looking at patch 5 again..
atomic_release should return an int to allow error propogation. There's no good reason why it shouldn't. This would allow handling errors in patch 5 more gracefully.
Makes sense, will change that. This made me think about how intel_dp_mst_atomic_release() handles an invalid mst_port reference i.e., in case the port is gone. I'll fix both and send a new version.
-DK
EXPORT_SYMBOL(drm_atomic_helper_check_modeset); diff --git a/include/drm/drm_modeset_helper_vtables.h b/include/drm/drm_modeset_helper_vtables.h index 091c422..84e80aa 100644 --- a/include/drm/drm_modeset_helper_vtables.h +++ b/include/drm/drm_modeset_helper_vtables.h @@ -836,6 +836,22 @@ struct drm_connector_helper_funcs { */ struct drm_encoder *(*atomic_best_encoder)(struct drm_connector *connector, struct drm_connector_state *connector_state);
- /**
* @atomic_release:** This function is conditionally called to release shared resources* when the attached CRTC's mode changes or it's connectors change or* becomes inactive. It is called before the corresponding* &drm_crtc_helper_funcs.atomic_check or* &drm_crtc_helper_funcs.mode_fixup hooks are called.** NOTE:** This function is called in the check phase of an atomic update.*/- void (*atomic_release)(struct drm_connector *connector,
struct drm_connector_state *connector_state);};
/**
Intel-gfx mailing list Intel-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/intel-gfx