Schedulability and Priority Assignment for Multi-Segment Self-Suspending Real-Time Tasks under Fixed-Priority Scheduling

Self-suspension is becoming an increasingly prominent characteristic in real-time systems such as: (i) I/O-intensive systems (ii) multi-core processors, and (iii) computation offloading systems with coprocessors, like Graphics Processing Units (GPUs). In this paper, we study the schedulability of multisegment self-suspension tasks under fixed-priority scheduling, where the executions of a multi-segment self-suspension task alternate between per-defined computation segments and suspension intervals. In particular, we do not use any enforcement to control the releases of computation segments and suspension intervals. Such an enforcement can prevent jitter but may incur non-negligible overheads. This work presents a combined method using the proposed multi-segment workload function to compute the upper bound on the worst-case response time (WCRT) of multi-segment tasks. To the best of our knowledge, this is the first study that successfully provides a pseudo-polynomial-time test for multi-segment self-suspension, hard real-time systems under fixed-priority scheduling without any additional execution control. We also show that the proposed analysis is compatible with Audsley’s Priority Assignment. Our empirical investigations show that the proposed approach is highly effective in terms of the number of task sets deemed to be schedulable.