K(H)eaps

K(H)eaps is a systematic study on Linux kernel exploit reliability problem. It identifies the unreliability factors during exploitation and investigates how existing stablization techniques mitigate the unreliability factors. Based on the knowledge, it proposes a new stablization technique and combined the new technique with existing techniques to outperform realworld exploits in reliability.

This repository contains the dataset we used in the paper and the platform we built for the reliability evaluation.

Paper

We describe our findings in the full paper

Playing for K(H)eaps: Understanding and Improving Linux Kernel Exploit Reliability

Kyle Zeng*, Yueqi Chen*, Haehyun Cho, Xinyu Xing, Adam Doupé, Yan Shoshitaishvili, and Tiffany Bao

* indicates equal contribution

In Proceedings of USENIX Security Symposium August 2022,

@inproceedings {280034,
title = {Playing for {K(H)eaps}: Understanding and Improving Linux Kernel Exploit Reliability},
booktitle = {31st USENIX Security Symposium (USENIX Security 22)},
year = {2022},
address = {Boston, MA},
url = {https://www.usenix.org/conference/usenixsecurity22/presentation/zeng},
publisher = {USENIX Association},
month = aug,
}

Setup

For the environment setup, we provide a dockerfile to ease the process.

To build the docker image, you need to run the following command in the source code root directory

cd scripts/create-image/ && ./create-image.sh && cd ../..
docker build -t kheap .

The command takes around 10 minutes to finish.

At this point, a docker image called kheap will be created.

Evaluation

The evaluation can be done completely inside the kheap docker image.

To launch a kheap docker container, you should use docker run --privileged -it kheap bash. Note that we need the privileged flag because the experiments are run in QEMU, which requires the access to kvm to speed up the virtual machines.

Then you can use the following command inside the docker container to evaluate all exploits for a specific CVE

cd /kheap/grader && python vuln_tester.py -c <cve_number> -n 5000 -r ./results -C 2 -m 2 -nl

The above command will run exploits for the specific CVE for 5000 times and save the result in results folder under 2CPU+2GB RAM setting

For example, cd /kheap/grader && python vuln_tester.py -c CVE-2010-2959 -n 5000 -r ./results -C 2 -m 2 -nl will start evaluation for CVE-2010-2959.

For detailed usage, please refer to the help message using python vuln_tester.py -h.

When each evaluation finishes, the result will be printed in stdout and also saved in the results folder in json format.

Note that a full-fledge 5000-run experiment for one CVE takes a 48-core machine 4-6 days to finish.