A memory leak

[worker24h]

Hi, michalvasko:
I find a memory leak when i read code. The following code is:
nc_reply* ncds_apply_rpc(ncds_id id, const struct nc_session* session, const nc_rpc* rpc, struct nc_filter* shared_filter)
{
....
if (ds->transapis != NULL
&& (op == NC_OP_COMMIT || op == NC_OP_COPYCONFIG || (op == NC_OP_EDITCONFIG && (nc_rpc_get_testopt(rpc) != NC_EDIT_TESTOPT_TEST))) &&
(nc_rpc_get_target(rpc) == NC_DATASTORE_RUNNING)) {

	old_data = ds->func.getconfig(ds, session, NC_DATASTORE_RUNNING, &e);
	old = read_datastore_data(ds->id, old_data);
	if (old == NULL) {/* cannot get or parse data */
		pthread_mutex_unlock(&ds->lock);
		if (e == NULL) { /* error not set */
			e = nc_err_new(NC_ERR_OP_FAILED);
			nc_err_set(e, NC_ERR_PARAM_MSG, "TransAPI: Failed to get data from RUNNING datastore.");
		}

// not free data. this is bug??
return nc_reply_error(e);
}
free(old_data);
}
.....
}