A Systematic Study of Elastic Objects in Kernel Exploitation

Yueqi Chen, Zhenpeng Lin, Xinyu Xing

Abstract: Recent research has proposed various methods to perform kernel exploitation and bypass kernel protection. For example, security researchers have demonstrated an exploitation method that utilizes the characteristic of elastic kernel objects to bypass KASLR, disclose stack/heap cookies, and even perform arbitrary read in the kernel. While this exploitation method is considered a commonly adopted approach to disclosing critical kernel information, there is no evidence indicating a strong need for developing a new defense mechanism to limit this exploitation method. It is because the effectiveness of this exploitation method is demonstrated only on anecdotal kernel vulnerabilities. It is unclear whether such a method is useful for a majority of kernel vulnerabilities. To answer this question, we propose a systematic approach. It utilizes static/dynamic analysis methods to pinpoint elastic kernel objects and then employs constraint solving to pair them to corresponding kernel vulnerabilities. In this work, we implement our proposed method as a tool - ELOISE. Using this tool on three popular OSes (Linux, FreeBSD, and XNU), we discover that elastic objects are pervasive in general caches. Evaluating the effectiveness of these elastic objects on 40 kernel vulnerabilities across three OSes, we observe that they can enable most of the vulnerabilities to bypass KASLR and heap cookie protector. Besides, we also observe that these elastic objects can even escalate the exploitability of some vulnerabilities allowing them to perform arbitrary read in the kernel. Motivated by these observations, we further introduce a new defense mechanism to mitigate the threat of elastic kernel objects. We prototype our defense mechanism on Linux, showing this mechanism introduces negligible overhead.