On the Effectiveness of Kernel Debloating via Compile-time Configuration
Linux kernel contains a large number of features that not all systems need, while Linux distributors enable as many features as possible to make their distributions generic, leading to severe bloating problem. Intuitively, we can use the existing configuration system to remove unnecessary features. However, it is unclear whether this system is adequate for kernel debloating. In this study, we perform analysis to understand how much security benefit a user can obtain if she performs the kernel debloating through the compile-time configuration. Our study shows that existing configuration system provides a convenient and effective vector to reduce the attack surface while producing a functional kernel. With such result, we plan to spend more effort to build a secure kernel through the compile-time debloating.
Thu 19 Jul
|16:00 - 16:37|
|16:37 - 17:15|
Julian DolbyIBM Thomas J. Watson Research Center
|17:15 - 17:52|
Omer TrippGoogle Inc.
|17:52 - 18:30|