> One minor addition, please add a sanity check to > nvmet_fc_register_targetport() making sure if one registers a > queue_create() a matching queue_delete() has to be provided as well, > otherwise someone could theoretically build a memory leak. ok -- james