Abstract:Separation logic is often praised for its ability to closely mimic the locality of state updates when reasoning about them at the level of assertions. The prover only needs to concern themselves with the footprint of the computation at hand, i.e., the part of the state that is actually being accessed and manipulated. Modern concurrent separation logics lift this local reasoning principle from the physical state to abstract ghost state. For instance, these logics allow one to abstract the state of a fine-grained concurrent data structure by a predicate that provides a client the illusion of atomic access to the underlying state. However, these abstractions inadvertently increase the footprint of a computation: when reasoning about a local low-level state update, one needs to account for its effect on the abstraction, which encompasses a possibly unbounded portion of the low-level state. Often this gives the reasoning a global character. We present context-aware (concurrent) separation logic (Co(Co)SL) to provide new opportunities for local reasoning in the presence of rich ghost state abstractions. Co(Co)SL introduces the notion of a context of a computation, the part of the concrete state that is only affected on the abstract level. Contexts give rise to a new proof rule that allows one to reduce the footprint by the context, provided the computation preserves the context as an invariant. The context rule complements the frame rule of separation logic by enabling more local reasoning in cases where the predicate to be framed is known in advance. We instantiate our developed theory for the flow framework, enabling contextual reasoning about programs manipulating general heap graphs, and describe two other applications of the logic. We have implemented the flow instantiation of the logic in a proof outline checker and used it to verify two highly-concurrent binary search trees.

Satisfiability of Compositional Separation Logic with Tree Predicates and Data Constraints.

A Complete Decision Procedure for Linearly Compositional Separation Logic with Data Constraints.

Tractability of Separation Logic with Inductive Definitions: Beyond Lists.

A separation logic for sequences in pointer programs and its decidability

A Decision Procedure for Guarded Separation Logic Complete Entailment Checking for Separation Logic with Inductive Definitions

Deciding Boolean Separation Logic via Small Models (Technical Report)

Constraint Automata on Infinite Data Trees: From CTL(Z)/CTL*(Z) To Decision Procedures

Model Checking Constraint LTL over Trees

SMT Solving for the Theory of Ordering Constraints.

Effective MSO-Definability for Tree-width Bounded Models of an Inductive Separation Logic of Relations

Model Checking Computation Tree Logic over Finite Lattices

Predicate Formal System Based on 0-Level Universal and Operator and Its Completeness

Inferring Properties in Computation Tree Logic

Deciding Separation Logic with Pointer Arithmetic and Inductive Definitions

Context-Aware Separation Logic

Decidable Logics Combining Word Equations, Regular Expressions and Length Constraints

Specification and Verification for Semi-Structured Data

HyperLTL Satisfiability Is Highly Undecidable, HyperCTL* is Even Harder

Expressiveness Results for an Inductive Logic of Separated Relations

Model Checking Fuzzy Computation Tree Logic.

An Extended Formalism To Constraint Logic Programming For Decision Analysis