In this paper we investigate some optimal convex control problems, with mixed constraints on the state and the control. We give a general condition which slows us to set optimality conditions for nonqualified problems (in the Slater sense). Then we give some applications and examples involving generalized bang-bang results.