drm: improve the documentation of connector hpd ops

While making the changes in [1], it was noted that the documentation
of the enable_hpd() and disable_hpd() does not make it clear that
these ops should not try to do hpd state maintenance and should only
enable/disable hpd related hardware for the connector.

The state management of these calls to make sure these calls are
balanced is handled by the DRM core and we should keep it that way
to minimize the overhead in the drivers which implement these ops.

[1]: https://patchwork.freedesktop.org/patch/558387/

Signed-off-by: Abhinav Kumar <[email protected]>
Suggested-by: Laurent Pinchart <[email protected]>
Reviewed-by: Laurent Pinchart <[email protected]>
Signed-off-by: Dmitry Baryshkov <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]

Author: Abhinav Kumar

include/drm/drm_modeset_helper_vtables.h

@@ -1154,6 +1154,11 @@ struct drm_connector_helper_funcs {
 	 * This operation is optional.
 	 *
 	 * This callback is used by the drm_kms_helper_poll_enable() helpers.
+	 *
+	 * This operation does not need to perform any hpd state tracking as
+	 * the DRM core handles that maintenance and ensures the calls to enable
+	 * and disable hpd are balanced.
+	 *
 	 */
 	void (*enable_hpd)(struct drm_connector *connector);
 
@@ -1165,6 +1170,11 @@ struct drm_connector_helper_funcs {
 	 * This operation is optional.
 	 *
 	 * This callback is used by the drm_kms_helper_poll_disable() helpers.
+	 *
+	 * This operation does not need to perform any hpd state tracking as
+	 * the DRM core handles that maintenance and ensures the calls to enable
+	 * and disable hpd are balanced.
+	 *
 	 */
 	void (*disable_hpd)(struct drm_connector *connector);
 };