A developing field of interest for the distributed systems and applied cryptography communities is that of smart contracts: self-executing financial instruments that synchronize their state, often through a blockchain. One such smart contract system that has seen widespread practical adoption is Ethereum, which has grown to a market capacity of 100 billion USD and clears an excess of 500,000 daily transactions. Unfortunately, the rise of these technologies has been marred by a series of costly bugs and exploits. Increasingly, the Ethereum community has turned to formal methods and rigorous program analysis tools. This trend holds great promise due to the relative simplicity of smart contracts and bounded-time deterministic execution inherent to the Ethereum Virtual Machine (EVM). Here we present KEVM, an executable formal specification of the EVM's bytecode stack-based language built with the K Framework, designed to serve as a solid foundation for further formal analyses. We empirically evaluate the correctness and performance of KEVM using the official Ethereum test suite. To demonstrate the usability, several extensions of the semantics are presented. and two different-language implementations of the ERC20 Standard Token are verified against the ERC20 specification. These results are encouraging for the executable semantics approach to language prototyping and specification.

KEVM: A Complete Formal Semantics of the Ethereum Virtual Machine

This paper presents K-Java, a complete executable formal semantics of Java 1.4. K-Java was extensively tested with a test suite developed alongside the project, following the Test Driven Development methodology. In order to maintain clarity while handling the great size of Java, the semantics was split into two separate definitions -- a static semantics and a dynamic semantics. The output of the static semantics is a preprocessed Java program, which is passed as input to the dynamic semantics for execution. The preprocessed program is a valid Java program, which uses a subset of the features of Java. The semantics is applied to model-check multi-threaded programs. Both the test suite and the static semantics are generic and ready to be used in other Java-related projects.

K-Java: A Complete Semantics of Java

A developing field of interest for the distributed systems and applied cryptography community is that of smart contracts: self-executing financial instruments that synchronize their state, often through a blockchain. One such smart contract system that has seen widespread practical adoption is Ethereum, which has grown to secure approximately 30 billion USD of currency value and in excess of 300,000 daily transactions. Unfortunately, the rise of these technologies has been marred by a repeated series of security vulnerabilities and high profile contract failures. To address these failures, the Ethereum community has turned to formal verification and program analysis which show great promise due to the computational simplicity and boundedtime execution inherent to smart contracts. Despite this, no fully formal, rigorous, comprehensive, and executable semantics of the EVM (Ethereum Virtual Machine) currently exists, leaving a lack of rigor on which to base such tools. In this work, we present KEVM, the first fully executable formal semantics of the EVM, the bytecode language in which smart contracts are executed. We create this semantics in a framework for executable semantics, the K framework. We show that our semantics not only passes the official 40,683-test stress test suite for EVM implementations, but also reveals ambiguities and potential sources of error in the existing on-paper formalization of EVM semantics [45] on which our work is based. These properties make KEVM an ideal formal reference implementation against which other implementations can be evaluated. We proceed to argue for a semantics-first formal verification approach for EVM contracts, and demonstrate its practicality by using KEVM to verify practically important properties over the arithmetic operation of an example smart contract and the correct operation of a token transfer function in a second contract. We show that our approach is feasible and not computationally restrictive. We hope that our work serves as the base for the development of a wide range of useful formally derived tools for Ethereum, including model checkers, certified compilers, and program equivalence checkers.

KEVM: A Complete Semantics of the Ethereum Virtual Machine

JavaFAN is a Java program analysis framework, that can symbolically execute multithreaded programs, detect safety violations searching through an unbounded state space, and verify finite state programs by explicit state model checking. Both Java language and JVM bytecode analyses are possible. JavaFAN's implementation consists of only 3,000 lines of Maude code, specifying formally the semantics of Java and JVM in rewriting logic and then using the capabilities of Maude for efficient execution, search and LTL model checking of rewriting theories.

Formal analysis of Java programs in JavaFAN

This paper presents KJS, the most complete and throughly tested formal semantics of JavaScript to date. Being executable, KJS has been tested against the ECMAScript 5.1 conformance test suite, and passes all 2,782 core language tests. Among the existing implementations of JavaScript, only Chrome V8's passes all the tests, and no other semantics passes more than 90%. In addition to a reference implementation for JavaScript, KJS also yields a simple coverage metric for a test suite: the set of semantic rules it exercises. Our semantics revealed that the ECMAScript 5.1 conformance test suite fails to cover several semantic rules. Guided by the semantics, we wrote tests to exercise those rules. The new tests revealed bugs both in production JavaScript engines (Chrome V8, Safari WebKit, Firefox SpiderMonkey) and in other semantics. KJS is symbolically executable, thus it can be used for formal analysis and verification of JavaScript programs. We verified non-trivial programs and found a known security vulnerability.

KJS: a complete formal semantics of JavaScript

Android apps require permissions when accessing resources related to privacy or system integrity. Starting from Android 6, these permissions have to be asked at runtime. However, migrating to the new permission model poses multiple challenges for developers. First, developers have to discover where the app uses permissions, which requires a permission specification. To date several such specifications have been built, yet these are either imprecise, incomplete or don't support all types of protected resources. 
We first present DPSpec, a novel permission specification built from several documentation formats supplied with the Android SDK. Compared with the state the art specification, it contains 2.5x as many entries for protected methods and detects dangerous permission usages in more than twice as many apps. 
A second challenge for developers is where to insert permission requests, with possible locations restricted by the request mechanism. 
We also present DPerm, a static analysis for Android apps that recommends locations for permission requests in code. It achieves high precision through context sensitivity and improves recall through a general call graph augmentation algorithm for incomplete code. Our empirical evaluation on 32 apps shows a precision of 96% and recall of 89%.

DPerm: Assisting the Migration of Android Apps to Runtime Permissions.

A programming language definition consists of two parts: syntax and semantics. K Framework is an actively developed formalism aimed to address the semantics part. While K has its own capabilities to define syntax and to perform parsing, those are often not expressive enough to handle real world programming languages. Therefore K Framework was designed to be connected to an external parser. In this case the semantics is defined over the representation of programs as AST (Abstract Syntax Tree), and is called label-based. This technique was successfully used to define C language. In this paper we present a parser generator which accepts as input a grammar in SDF format, and produces a parser ready to be integrated with K. We illustrate the methodology of defining a language using the generated parser based on an example. The example definition emphasizes distinctive features of the integration of SDF syntax and K semantics. It also illustrates syntax constructs which are difficult to handle with K Framework alone, but are straightforward to express in SDF + K combination.

Label-Based Programming Language Semantics in K Framework with SDF

In this section we present the fragment of con guration used by runtime method invocation. The gure below contains the cells and their sorts. μ ≠≠ 〈K〉 k 〈 List [ (K,Bag) ] 〉 stack 〈〈Map [Id 7→ Int]〉 env 〈ClassType〉 crntClass 〈Int〉 location〉methodContext∑ thread∗∑ threads μμ 〈ClassType〉 classType 〈ClassMetaType〉 classMetaType 〈Map [Signature 7→ ClassType]〉methods ¥¥ 〈Signature〉methodSignature 〈 List [ MethodParam ] 〉methodParams 〈K〉methodBody 〈AccessMode〉methodAccessMode 〈ContextType〉methodContextType æ

K-Java: Runtime Semantics for Method Invocation and Object Instantiation

To protect user’s privacy and system’s integrity, mobile platforms use permission models to control accesses to protected resources such as GPS location, Contacts, etc. The previous major version of Android used a static permission model, which compromised the security and privacy of apps. Android 6 overhauled its permission model to ask permissions at runtime which reduces the risk of permission abuse. However, migrating to the runtime permission model requires significant effort from the app developers. In this paper we first present a large-scale formative study to understand how app developers use and migrate to the new permission model. Inspired by these findings, we designed, implemented, and evaluated a tool suite that (i) recommends locations where to insert permission requests and (ii) automatically inserts all the permission-related code. Our empirical evaluations on a diverse corpus of real-world apps show that our tools are highly applicable and accurate.

/pdf/analysis-and-transformations-in-support-of-android-privacy-3fomnf8owd.pdf

Denis Bogdanas

Papers

K-Java: A Complete Semantics of Java

DPerm: Assisting the Migration of Android Apps to Runtime Permissions.

Label-Based Programming Language Semantics in K Framework with SDF

K-Java: Runtime Semantics for Method Invocation and Object Instantiation

Analysis and Transformations in Support of Android Privacy