Correctness of Mutual Exclusion Algorithms

One way of characterizing non-token-based mutual exclusion (m.e.) algorithms is in terms of the underlying information structure. The information structure for a given m.e. algorithm specifies which particular processes interact with which other processes before entering their critical sections, and which processes they interact with when leaving their critical sections. By focusing on the information structure of an m.e. algorithm, we can compare different algorithms by comparing the respective information structures. Further, we can characterize the correctness of an m.e. algorithm in terms of conditions that its information structure must satisfy. In this paper, we propose a necessary and sufficient condition for ensuring the correctness of the m.e. algorithm and show that this condition is superior to the currently accepted condition.