Defining Deadlock with Fungible Resources

摘要

In a previous paper we discussed resource deadlock within a classification of all dead states. We differentiated between the different types of anomalies that are called deadlock in the computer science literature. Our primary aim was to correct inconsistencies that are currently being taught in operating systems' textbooks. We introduced a model that allowed us to precisely define deadlock, but we only provided a partial definition of resource deadlock, which, as we noted, was limited to deadlocks with uniquely identifiable resource units. In response to readers of OSR, we now distinguish between different types of resource deadlock and provide a comprehensive definition. We also suggest that it may be desirable to maintain separate definitions.