RVDebloater: Mode-based Adaptive Firmware Debloating for Robotic Vehicles
summary
RVDebloater claims mode-based, reversible firmware debloating that identifies per-mode needed functions with static and dynamic analysis, instruments indirect control flows with LLVM-based instrumentation and enforces allowed-call tables at runtime without special hardware. Implemented and tested on six robotic vehicles (four simulated, two real) it reduces reachable functions by 40% (static) to 85% (dynamic), prunes call graphs ~45%, adds ~3.9% runtime and ~4% memory overhead (~0.25 MB), causes no mission failures and prevents three real attacks.
AI-written summary · about
chapters (6)
- 0:00Robotic vehicle firmware risks
- 3:41Threat model limits
- 5:06RVD loader approach
- 7:55Runtime monitoring
- 9:20Evaluation results
- 12:10Conclusion
related talks
55mStrong software for a safer future
35mDevOps Days MTL 2019- Yannick Brosseau
29mBackdooring of Real Time Automotive OS Devices
13mInsight of Linux Distro Optimization for RISC-V: Using VisionFive 2 Debian as Exam...
26mDebloating Software through Piece-Wise Compilation and Loading
01h 08mAutomated Dynamic Fireware Analysis At Scale: A Case Study on Embedded Web Interfaces
34mRISC-V-Based Sandboxing for Secure and Efficient Software Execution
39mBackdooring Hardware Devices by Injecting Malicious Payloads on Microcontrollers