Identifying Safety Weaknesses and Fault Propagation in the Linux Kernel
summary
A reproducible fault-injection tool targets kernel control structures on ARM64 by letting a monitor spawn a victim process, locating its instance via a structure DB, and applying deterministic corruptions (write zeros, 0xFF, random pointers) from an interference generator. The workflow measures whether safety monitors report the fault reliably and within the car-relevant timing budget (tens of milliseconds) and computes hit rates over many runs; hypervisor trapping costs (~5,000 cycles) and which fields actually break safety are reported, and the tool can run against post-release kernels.
AI-written summary · about
chapters (6)
- 0:00Introduction and context
- 1:02The problem of safety
- 6:58Theory and practices
- 14:23Reusing and recycling methods
- 24:00Attempted safety remedies
- 32:12Fault injection methodology
related talks
26mIt's Coming From Inside the House: Kernelspace Fault Injection with KRF
44mDeep Analysis of Exploitable Linux Kernel Vulnerabilities 2017-2019
28mkAFL: Hardware-Assisted Feedback Fuzzing for OS Kernels
28mComplexity Killed Security
32mHow to Backdoor the Linux Kernel (and fail?)
41mUsing Microarchitectural Design to Break KASLR and More
48mCapturing 0Day Exploits With Perfectly Placed Hardware Traps
43mUsing Undocumented CPU Behavior to See Into Kernel Mode and Break KASLR in the Process
