Fuzzing JavaScript Engines by Fusing JavaScript and WebAssembly
Jiayi Lin, Changhua Luo, Mingxue Zhang, Lanteng Lin, Penghui Li, Chenxiong Qian
摘要
JavaScript engines are a fundamental part of modern browsers, and many efforts have been invested in testing them to enhance their security. However, the incorporation of WebAssembly into JavaScript engines introduces new attack surfaces that have not received sufficient attention. Existing fuzzers for JavaScript engines primarily focus on JavaScript, neglecting WebAssembly code and its interactions with JavaScript. We introduce Mad-Eye, the first fuzzer that can test the JavaScript-WebAssembly interaction using a novel cross-language code fusion technique. Evaluations of Mad-Eye on V8, SpiderMonkey, and JavaScriptCore detected 21 previously unknown vulnerabilities, with 20 confirmed and 18 fixed and merged into mainstream browsers by the developers, who acknowledged our reports with vulnerability bounties.
问问这篇 Paper
问问你的智能体。
Lune 读过与它相关的顶会 Paper,每个回答都会注明依据哪几篇。
相关 Paper
- FuzzJIT: Oracle-Enhanced Fuzzing for JavaScript Engine JIT CompilerJunjie Wang, Zhiyi Zhang, Shuang Liu, Xiaoning Du 等USENIX Security 2023
- RGFuzz: Rule-Guided Fuzzer for WebAssembly RuntimesJunyoung Park, Yunho Kim, Insu YunS&P 2025
- FUZZILLI: Fuzzing for JavaScript JIT Compiler VulnerabilitiesSamuel Groß, Simon Koch, Lukas Bernhard, Thorsten Holz 等NDSS 2023
- CodeAlchemist: Semantics-Aware Code Generation to Find Vulnerabilities in JavaScript EnginesHyungSeok Han, DongHyeon Oh, Sang Kil ChaNDSS 2019 · 被引用 178 次
- Montage: A Neural Network Language Model-Guided JavaScript Engine FuzzerSuyoung Lee, HyungSeok Han, Sang Kil Cha, Sooel SonUSENIX Security 2020
