Deadlock

Summary

When each process in a group need two or more shared resource instances of a resource type, and every process holds a resource and waits to acquire a resource instance that another process holds, there is indefinite waiting for all processes in the group. This state is called deadlock because each process is waiting for an event that will never occur. The five philosophers problem is a classical synchronization problem that is used here to illustrate deadlock. The three general methods that operating systems use to handle deadlock are prevention, avoidance and detection.

Four different simulation models, which implement the five philosophers problem, were presented in this chapter to show in detail how and why deadlock occurs, and three possible solutions were discussed. Different values of performance measures were shown from the simulation runs.