MundoFuzz: Hypervisor Fuzzing with Statistical Coverage Testing and Grammar Inference
Cheolwoo Myung, Gwangmu Lee, Byoungyoung Lee
摘要
A hypervisor is system software, managing and running virtual machines. Since the hypervisor is placed at the lowestlevel in the typical systems software stack, it has critical security implications. Once compromised, the entire software components running on top of the hypervisor (including all guest virtual machines and applications running within each guest virtual machine) are compromised as well, as the hypervisor has all the privileges to access those. This paper proposes MUNDOFUZZ, a hypervisor fuzzer to enable both coverage-guided and grammar-aware fuzzing. We find that the coverage measurement in hypervisors suffers from noises due to the hypervisor's asynchronous system event handling. In order to filter out such noises, MUNDOFUZZ develops a statistical differential coverage measurement methods, allowing MUNDOFUZZ to capture the clean coverage information for hypervisor inputs. Moreover, we observe that hypervisor inputs have complex input grammars because it supports many different devices and each device has its own input format. Thus, MUNDOFUZZ learns the input grammar through inspecting the coverage characteristics of the given hypervisor input, which is based on the idea that the hypervisor behaves in a different way if the grammatically correct (or incorrect) input is given. We evaluated MUNDOFUZZ with popular hypervisors, QEMU and Bhyve, and MUNDOFUZZ outperformed other state-of-the-art hypervisor fuzzers ranging from 4.91% to 6.60% in terms of coverage. More importantly, MUNDOFUZZ identified 40 previously unknown bugs (including 9 CVEs), demonstrating its strong practical effectiveness in finding real-world hypervisor vulnerabilities. * Corresponding author which enables a single host machine to run multiple guest VMs. This advantage becomes particularly important as the computing trends are rapidly shifting towards cloud computing. Using the hypervisor, cloud computing platforms can facilitate flexible resource managements on virtualized platforms, which can efficiently service for billions of users [18] . From the security point of view, hypervisors introduce crucial attack surface to be protected. Importantly, a hypervisor is running at the lowest layer in the typical systems software stack, so it is given the highest security privilege of the host machine. Consequently, any successful compromise exploiting hypervisor vulnerabilities would allow attackers to gain privilege of the entire host machine, including the hypervisor itself as well as all guest VMs running on top of them.
问问这篇 Paper
智能体会读完全文。
Lune 把这篇 Paper 索引到了最后一个公式,引用它的顶会 Paper 也一样。你提问,回答直接引用原文。
引用它的顶会 Paper9
- SoK: Prudent Evaluation Practices for FuzzingMoritz Schloegel, Nils Bars, Nico Schiller, Lukas Bernhard 等S&P 2024 · 被引用 69 次
- HYPERPILL: Fuzzing for Hypervisor-bugs by leveraging the Hardware Virtualization InterfaceAlexander Bulekov, Qiang Liu, Manuel Egele, Mathias PayerUSENIX Security 2024 · 被引用 15 次
- ORANalyst: Systematic Testing Framework for Open RAN ImplementationsTianchang Yang, Syed Md. Mukit Rashid, Ali Ranjbar, Gang Tan 等USENIX Security 2024 · 被引用 10 次
- Breaking Isolation: A New Perspective on Hypervisor Exploitation via Cross-Domain AttacksGaoning Pan, Yiming Tao, Qinying Wang, Chunming Wu 等NDSS 2026
- AidFuzzer: Adaptive Interrupt-Driven Firmware Fuzzing via Run-Time State RecognitionJianqiang Wang, Qinying Wang, Tobias Scharnowski, Li Shi 等USENIX Security 2025
它引用的顶会 Paper8
- Coverage-based Greybox Fuzzing as Markov ChainMarcel Böhme, Van-Thuan Pham, Abhik RoychoudhuryCCS 2016 · 被引用 1,026 次
- Evaluating Fuzz TestingGeorge Klees, Andrew Ruef, Benji Cooper, Shiyi Wei 等CCS 2018 · 被引用 753 次
- MoonShine: Optimizing OS Fuzzer Seed Selection with Trace DistillationShankara Pailoor, Andrew Aday, Suman JanaUSENIX Security 2018 · 被引用 180 次
- IMF: Inferred Model-based FuzzerHyungSeok Han, Sang Kil ChaCCS 2017 · 被引用 139 次
- DifuzzRTL: Differential Fuzz Testing to Find CPU BugsJaewon Hur, Suhwan Song, Dongup Kwon, Eunjin Baek 等S&P 2021 · 被引用 126 次
相关 Paper
- HYPER-CUBE: High-Dimensional Hypervisor FuzzingSergej Schumilo, Cornelius Aschermann, Ali Abbasi, Simon Wörner 等NDSS 2020
- Morphuzz: Bending (Input) Space to Fuzz Virtual DevicesAlexander Bulekov, Bandan Das, Stefan Hajnoczi, Manuel EgeleUSENIX Security 2022
- Nyx: Greybox Hypervisor Fuzzing using Fast Snapshots and Affine TypesSergej Schumilo, Cornelius Aschermann, Ali Abbasi, Simon Wörner 等USENIX Security 2021 · 被引用 102 次
- HyperMirage: Direct State Manipulation in Hybrid Virtual CPU FuzzingManuel Andreas, Fabian Specht, Marius MomeuNDSS 2026
- HyperFuzzer: An Efficient Hybrid Fuzzer for Virtual CPUsXinyang Ge, Ben Niu, Robert Brotzman, Yaohui Chen 等CCS 2021 · 被引用 9 次
