A note on efficient solutions for the linear bilevel programming problem