The Youla parametrization of two-degree-of-freedom controllers reveals that the design of the prefilter and the design of the regulator can be done independently. In this paper, a unified framework is proposed to address the design of the feedforward part of a two-degree-of-freedom controller. The approach deals both with tracking of a causal reference signal and with rejection of a causal measurable disturbance. It is shown that the determination of a feedforward filter, with a constraint on the nominal control effort, reduces to the solution of a particular full-information problem. The approach is attractive because it leads to a significant reduction in the number of decision variables in the design process. The method also extends naturally to plants in a polytope. Its effectiveness is demonstrated on two numerical examples. In the first example, the design of a linear time-invariant feedforward controller is carried out to reject a measurable disturbance and to follow a reference signal; while in the second example, we address the design of a robust gain scheduled two-degree-of-freedom controller.