Weak updates and separation logic

Abstract

Separation Logic (SL) provides a simple but powerful technique for reasoning about imperative programs that use shared data structures. Unfortunately, SL supports only updates, in which mutation to a location is safe only if a unique reference is owned. This limits the applicability of SL when reasoning about the interaction between many high-level languages (e.g., ML, Java, C#) and low-level ones since these high-level languages do not support strong updates. Instead, they adopt the discipline of updates, in which there is a global heap type to enforce the invariant of type-preserving updates. We present SLw, a logic that extends SL with reference types and elegantly reasons about the interaction between strong and weak updates. We also describe a semantic framework for reference types; this framework is used to prove the soundness of SLw.