Over the last decade, step-indices have been widely used for the construction of operationally-based logical relations in the presence of various kinds of recursion. We first give an argument that stepindices, or something like them, seem to be required for defining realizability relations between high-level source languages and lowlevel targets, in the case that the low-level allows egregiously intensional operations such as reflection or comparison of code pointers. We then show how, much to our annoyance, step-indices also seem to prevent us from exploiting such operations as aggressively as we would like in proving program transformations.

This paper presents Vellvm (verified LLVM), a framework for reasoning about programs expressed in LLVM’s intermediate representation and transformations that operate on it. Vellvm provides a mechanized formal semantics of LLVM’s intermediate representation, its type system, and properties of its SSA form. The framework is built using the Coq interactive theorem prover. It includes multiple operational semantics and proves relations among them to facilitate different reasoning styles and proof techniques. To validate Vellvm’s design, we extract an interpreter from the Coq formal semantics that can execute programs from LLVM test suite and thus be compared against LLVM reference implementations. To demonstrate Vellvm’s practicality, we formalize and verify a previously proposed transformation that hardens C programs against spatial memory safety violations. Vellvm’s tools allow us to

Separation logic is a powerful tool for reasoning about structured, imperative programs that manipulate pointers. However, its application to unstructured, lower-level languages such as assembly language or machine code remains challenging. In this paper we describe a separation logic tailored for this purpose that we have applied to x86 machine-code programs. The logic is built from an assertion logic on machine states over which we construct a specification logic that encapsulates uses of frames and step indexing. The traditional notion of Hoare triple is not applicable directly to unstructured machine code, where code and data are mixed together and programs do not in general run to completion, so instead we adopt a continuation-passing style of specification with preconditions alone. Nevertheless, the range of primitives provided by the specification logic, which include a higher-order frame connective, a novel read-only frame connective, and a ‘later ’ modality, support the definition of derived forms to support structured-programming-style reasoning for common cases, in which standard rules for Hoare triples are derived as lemmas. Furthermore, our encoding of scoped assembly-language labels lets us give definitions and proof rules for powerful assemblylanguage ‘macros ’ such as while loops, conditionals and procedures. We have applied the framework to a model of sequential x86 machine code built entirely within the Coq proof assistant, including tactic support based on computational reflection. Categories and Subject Descriptors F.3.1 [Logics and meanings

Reasoning about program equivalence is one of the oldest problems in semantics. In recent years, useful techniques have been developed, based on bisimulations and logical relations, for reasoning about equivalence in the setting of increasingly realistic languages—languages nearly as complex as ML or Haskell. Much of the recent work in this direction has considered the interesting representation independence principles enabled by the use of local state, but it is also important to understand the principles that powerful features like higher-order state and control effects disable. This latter topic has been broached extensively within the framework of game semantics, resulting in what Abramsky dubbed the “semantic cube”: fully abstract game-semantic characterizations of various axes in the design space of ML-like languages. But when it comes to reasoning about many actual examples, game semantics does not yet supply a useful technique for proving equivalences. In this paper, we marry the aspirations of the semantic cube to the powerful proof method of stepindexed Kripke logical relations. Building on recent work of Ahmed, Dreyer, and Rossberg, we define the first fully abstract logical relation for an ML-like language with recursive types, abstract types, general references and call/cc. We then show how, under orthogonal restrictions to the expressive power of our language—namely, the restriction to first-order state and/or the removal of call/cc—we can enhance the proving power of our possible-worlds model in correspondingly orthogonal ways, and we demonstrate this proving power on a range of interesting examples. Central to our story is the use of state transition systems to model the way in which properties of local state evolve over time.