Automotive firmware extraction and analysis techniques

Van den Herrewegen, Jan (2021). Automotive firmware extraction and analysis techniques. University of Birmingham. Ph.D.

Preview

VandenHerrewegen2021PhD.pdf
Text - Accepted Version
Available under License All rights reserved.
Download (1MB) | Preview

Abstract

An intricate network of embedded devices, called Electronic Control Units (ECUs), is responsible for the functionality of a modern vehicle. Every module processes a myriad of information and forwards it on to other nodes on the network, typically an automotive bus such as the Controller Area Network (CAN). Analysing embedded device software, and automotive in particular, brings many challenges.

The analyst must, especially in the notoriously secretive automotive industry, first lift the ECU firmware from the hardware, which typically prevents unauthorised access. In this thesis, we address this problem in two ways:

- We detail and bypass the access control mechanism used in diagnostic protocols in ECU firmware. Using existing diagnostic functionality, we present a generic technique to download code to RAM and execute it, without requiring physical access to the ECU. We propose a generic firmware readout framework on top of this, which only requires access to the CAN bus.
- We analyse various embedded bootloaders and combine dynamic analysis with low-level hardware fault attacks, resulting in several fault-injection attacks which bypass on-chip readout protection.

We then apply these firmware extraction techniques to acquire immobiliser firmware by two different manufacturers, from which we reverse engineer the DST80 cipher and present it in full detail here. Furthermore, we point out flaws in the key generation procedure, also recovered from the ECU firmware, leading to a full key recovery based on publicly readable transponder pages.

Type of Work:

Thesis (Doctorates > Ph.D.)

Award Type:

Doctorates > Ph.D.

Supervisor(s):