Guarded Sections: Structuring Aid for Wait-Free Synchronisation

Wait-free synchronisation gives any process in the system strong progress guarantees, irrespective of number and behaviour of other processes simultaneously competing for shared resources (i.e., data structures and code sections). It ensures completion of any operation in a finite number of steps and, thus, provides the basis to derive bounded above or even constant executions times for each of these operations of the non-sequential programs. This characteristic is of special meaning for timedependent processes typical for real-time (embedded) systems. Unfortunately, wait-free synchronisation against the background of arbitrary data and code structures is no bed of roses. This paper is about a two-stage approach of organising nonsequential programs to the benefit of wait-free synchronisation. In a preceding constructional stage, conventional critical sections are designed as so called guarded sections. Other than critical sections, processes never block at entrance to a guarded section although only one process at a time is allowed to pass through. Competing processes are forced into bypass but, if necessary and by using futures, they can synchronise on concurrent state changes inside the respective section. In consequence of this measure, the execution model of guarded sections constrains the overlapping pattern of interacting (simultaneous) processes. Thereby, in the downstream transactional stage, efficient waitfree synchronisation of the “guarding operations” is gratifying by-product. First experimental results made with a 80-way multi-core system show that non-blocking wait-free synchronised guarded sections outperform MCS-locks.