Bor-Yuh Evan Chang

Boogie: a modular reusable verifier for object-oriented programs

A program verifier is a complex system that uses compiler technology, program semantics, property inference, verification-condition generation, automatic decision procedures, and a user interface. This paper describes the architecture of a state-of-the-art program verifier for object-oriented programs.

Relational inductive shape analysis

Shape analyses are concerned with precise abstractions of the heap to capture detailed structural properties. To do so, they need to build and decompose summaries of disjoint memory regions. Unfortunately, many data structure invariants require relations be tracked across disjoint regions, such as intricate numerical data invariants or structural invariants concerning back and cross pointers. In this paper, we identify issues inherent to analyzing relational structures and design an abstract domain that is parameterized both by an abstract domain for pure data properties and by user-supplied specifications of the data structure invariants to check. Particularly, it supports hybrid invariants about shape and data and features a generic mechanism for materializing summaries at the beginning, middle, or end of inductive structures. Around this domain, we build a shape analysis whose interesting components include a pre-analysis on the user-supplied specifications that guides the abstract interpretation and a widening operator over the combined shape and data domain. We then demonstrate our techniques on the proof of preservation of the red-black tree invariants during insertion.

In defense of soundiness: a manifesto

Soundy is the new sound.

Shape analysis with structural invariant checkers

Developer-supplied data structure specifications are important to shape analyses, as they tell the analysis what information should be tracked in order to obtain the desired shape invariants. We observe that data structure checking code (e.g., used in testing or dynamic analysis) provides shape information that can also be used in static analysis. In this paper, we propose a lightweight, automatic shape analysis based on these developer-supplied structural invariant checkers. In particular, we set up a parametric abstract domain, which is instantiated with such checker specifications to summarize memory regions using both notions of complete and partial checker evaluations. The analysis then automatically derives a strategy for canonicalizing or weakening shape invariants

The open verifier framework for foundational verifiers

We present the Open Verifier approach for verifying untrusted code using customized verifiers. This approach can be viewed as an instance of foundational proof-carrying code where an untrusted program can be checked using the verifier most natural for it instead of using a single generic type system. In this paper we focus on a specialized architecture designed to reduce the burden of expressing both type-based and Hoare-style verifiers.A new verifier is created by providing an untrusted executable extension module, which can incorporate directly pre-existing non-foundational verifiers based on dataflow analysis or type checking. The extensions control virtually all aspects of the verification by carrying on a dialogue with the Open Verifier using a language designed both to correspond closely to common verification actions and to carry simple adeQuacy proofs for those actions.We describe the design of the trusted core of the Open Verifier, along with our experience implementing proof-carrying code, typed assembly language, and dataflow or abstract interpretation based verifiers in this unified setting.

Trustless Grid Computing in ConCert

We believe that fundamental to the establishment of a grid computing framework where all (not just large organizations) are able to effectively tap into the resources available on the global network is the establishment of trust between grid application developers and resource donors. Resource donors must be able to trust that their security, safety, and privacy policies will be respected by programs that use their systems.In this paper, we present a novel solution based on the notion of certified code that upholds safety, security, and privacy policies by examining intrinsic properties of code. Certified code complements authentication and provides a foundation for a safe, secure, and efficient framework that executes native code. We describe the implementation of such a framework known as the ConCert software.

Thresher: precise refutations for heap reachability

We present a precise, path-sensitive static analysis for reasoning about heap reachability, that is, whether an object can be reached from another variable or object via pointer dereferences. Precise reachability information is useful for a number of clients, including static detection of a class of Android memory leaks. For this client, we found the heap reachability information computed by a state-of-the-art points-to analysis was too imprecise, leading to numerous false-positive leak reports. Our analysis combines a symbolic execution capable of path-sensitivity and strong updates with abstract heap information computed by an initial flow-insensitive points-to analysis. This novel mixed representation allows us to achieve both precision and scalability by leveraging the pre-computed points-to facts to guide execution and prune infeasible paths. We have evaluated our techniques in the Thresher tool, which we used to find several developer-confirmed leaks in Android applications.

Analysis of low-level code using cooperating decompilers

Analysis or verification of low-level code is useful for minimizing the disconnect between what is verified and what is actually executed and is necessary when source code is unavailable or is, say, intermingled with inline assembly. We present a modular framework for building pipelines of cooperating decompilers that gradually lift the level of the language to something appropriate for source-level tools. Each decompilation stage contains an abstract interpreter that encapsulates its findings about the program by translating the program into a higher-level intermediate language. We provide evidence for the modularity of this framework through the implementation of multiple decompilation pipelines for both x86 and MIPS assembly produced by gcc, gcj, and coolc (a compiler for a pedagogical Java-like language) that share several low-level components. Finally, we discuss our experimental results that apply the BLAST model checker for C and the Cqual analyzer to decompiled assembly.

Calling context abstraction with shapes

Interprocedural program analysis is often performed by computing procedure summaries. While possible, computing adequate summaries is difficult, particularly in the presence of recursive procedures. In this paper, we propose a complementary framework for interprocedural analysis based on a direct abstraction of the calling context. Specifically, our approach exploits the inductive structure of a calling context by treating it directly as a stack of activation records. We then build an abstraction based on separation logic with inductive definitions. A key element of this abstract domain is the use of parameters to refine the meaning of such call stack summaries and thus express relations across activation records and with the heap. In essence, we define an abstract interpretation-based analysis framework for recursive programs that permits a fluid per call site abstraction of the call stack--much like how shape analyzers enable a fluid per program point abstraction of the heap.

Mixing type checking and symbolic execution

Static analysis designers must carefully balance precision and efficiency. In our experience, many static analysis tools are built around an elegant, core algorithm, but that algorithm is then extensively tweaked to add just enough precision for the coding idioms seen in practice, without sacrificing too much efficiency. There are several downsides to adding precision in this way: the tools implementation becomes much more complicated; it can be hard for an end-user to interpret the tools results; and as software systems vary tremendously in their coding styles, it may require significant algorithmic engineering to enhance a tool to perform well in a particular software domain. In this paper, we present Mix, a novel system that mixes type checking and symbolic execution. The key aspect of our approach is that these analyses are applied independently on disjoint parts of the program, in an off-the-shelf manner. At the boundaries between nested type checked and symbolically executed code regions, we use special mix rules to communicate information between the off-the-shelf systems. The resulting mixture is a provably sound analysis that is more precise than type checking alone and more efficient than exclusive symbolic execution. In addition, we also describe a prototype implementation, Mixy, for C. Mixy checks for potential null dereferences by mixing a null/non-null type qualifier inference system with a symbolic executor.