Towards efficient and effective kernel driver fuzzing

————————————————————————–

▣ 연사(Speaker) : Prof. Dokyung Song (Yonsei University)

▣ 초청(Host) : Prof. Gwangsun Kim

▣ 일시(Date &Time) : 2022.5.27(Fri), 4pm ~

▣ 언어(Language) : 한국어(Korean)

▣ Zoom URL : https://postech-ac-kr.zoom.us/j/97754060329?pwd=TUlkdjlqVGV6U1BiZzl4aEZVL2Y3Zz09

Zoom ID : 977 5406 0329 Passcode : 251664
—————————————————————————————————————————

Fuzzing has been successfully applied to a variety of software systems, revealing countless vulnerabilities in them. OS kernels are not an exception. Numerous vulnerabilities in OS kernels have been discovered by kernel fuzzers such as Google’s Syzkaller. Despite a huge success, multiple problems remain to be solved, especially for OS kernels, due to unique challenges posed by the kernels. The challenges can broadly be classified into two: the efficiency and effectiveness challenge. In this talk, I will introduce my past and current research on addressing these two challenges in OS kernel fuzzing. First, I will talk about our research that tackles the efficiency challenge; in particular, I will introduce Agamotto, a system that can transparently accelerate kernel fuzzers via lightweight virtual machine checkpointing. Then, I will briefly introduce our efforts currently ongoging in my research group to tackle the effectiveness challenge (or, more precisely, the dependency challenge) of kernel driver fuzzing. I will conclude the talk by unifying these two lines of efforts into a high-level research theme “systems approaches to security analysis”, which hopefully inspires further research and collaboration in the area.

Dokyung Song is an assistant professor in the Department Computer Science at Yonsei University and the director of Yonsei University’s Cyber Security Lab. He received his B.S. degree in Electrical and Computer Engineering from Seoul National University in 2014, and his M.S. and Ph.D. degree in Computer Science from UC Irvine in 2019 and 2020, respectively. During his Ph.D. studies, he worked as an intern in the C++ dynamic analysis team and the Fuchsia OS security team at Google, and also in the product security team at Qualcomm. He also worked in the server technologies group at Oracle as a senior member of technical staff. His research interest lies in the broad area of systems security, and his current focus is on developing new systems, compiler and machine learning techniques that can better analyze the security of OS kernels as well as binary-only software.