Philipp Zieris

Konrad Hohentanner, Philipp Zieris, Julian Horsch

Memory safety bugs remain in the top ranks of security vulnerabilities, even after decades of research on their detection and prevention. Various mitigations have been proposed for C/C++, ranging from language dialects to instrumentation. Among these, compiler-based instrumentation is particularly promising, not requiring manual code modifications and being able to achieve precise memory safety. Unfortunately, existing compiler-based solutions compromise in many areas, including performance but also usability and memory safety guarantees. New developments in hardware can help improve performance and security of compiler-based memory safety. ARM Pointer Authentication, added in the ARMv8.3 architecture, is intended to enable hardware-assisted Control Flow Integrity (CFI). But since its operations are generic, it also enables other, more comprehensive hardware-supported runtime integrity approaches. As such, we propose CryptSan, a memory safety approach based on ARM Pointer Authentication. CryptSan uses pointer signatures to retrofit memory safety to C/C++ programs, protecting heap, stack, and globals against temporal and spatial vulnerabilities. We present a full LLVM-based prototype implementation, running on an M1 MacBook Pro, i.e., on actual ARMv8.3 hardware. Our prototype evaluation shows that the system outperforms similar approaches under real-world conditions. This, together with its interoperability with uninstrumented libraries and cryptographic protection against attacks on metadata, makes CryptSan a viable solution for retrofitting memory safety to C/C++ programs.

Mathias Morbitzer, Benedikt Kopf, Philipp Zieris

The majority of cloud providers offers users the possibility to deploy Trusted Execution Environments (TEEs) to protect their data and processes from high privileged adversaries. This offer is intended to address concerns of users when moving critical tasks into the cloud. However, TEEs only allow to attest the integrity of the environment at launch-time. To also enable the attestation of a TEE's integrity at run-time, we present GuaranTEE. GuaranTEE uses control-flow attestation to ensure the integrity of a service running within a TEE. By additionally placing all components of GuaranTEE in TEEs, we are able to not only detect a compromised target, but are also able to protect ourselves from malicious administrators. We show the practicability of GuaranTEE by providing a detailed performance and security evaluation of our prototype based on Intel SGX in Microsoft Azure. Our evaluation shows that the need to transfer information between TEEs and the additional verification process add considerable overhead under high CPU load. Yet, we are able to reduce this overhead by securely caching collected information and by performing the analysis in parallel to executing the application. In summary, our results show that GuaranTEE provides a practical solution for cloud users focused on protecting the integrity of their data and processes at run-time.