> >Seems reasonable to me at the first glance. > >We need to call IGB_CORE_LOCK_DESTROY(adapter) before returning though. > Not necessary. igb_detach does that.