Control applications require defining several parallel activities to model the environment. Periodic tasks model the activities to be executed at periodic instants of time. While the process of control design is focused on obtaining the regulator, later on translated into an algorithm, the software design is focused on producing pieces of software that will be executed concurrently under a scheduler. Nowadays, more and more applications require complex computation and the use of complex algorithms that can compromise the response time of the system. The activities involving a control loop task can be structured in some parts: data acquisition, computation of the control action, optional activities and output of the control action. This decomposition is useful to improve the control performance and reduce delays due to the scheduler. This paper shows how to implement complex real-time control applications by means of periodic tasks in RT-Linux, using a task decomposition