Blocking analysis of real-time tasks with parallel critical sections under federated scheduling

Locks are among the most common mechanisms for coordinating concurrent access to exclusive shared resources. Although extensive research has examined locking protocols and schedulability analysis for sequential real-time tasks, comparatively little attention has been paid to parallel tasks. Existing approaches commonly assume that critical sections must execute sequentially—an assumption maybe violated by realistic parallel programs. In this paper, we focus on the analysis of parallel real-time tasks containing parallel critical sections. We enhanced the locking protocols with new queue structure, new progress mechanism and propose novel analytical methods that accommodate both light (with density $\le 1$) and heavy tasks (with density $>1$), as opposed to approaches that exclusively target heavy tasks. We evaluate our techniques through comprehensive experiments, demonstrating that they improve upon existing methods by providing more accurate and efficient schedulability assessments.