KFuzzTest: Targeted Fuzzing of Internal Kernel Functions
summary
KFuzzTest exposes stateless or low‑state internal kernel functions to userspace fuzzers by embedding metadata into vmlinux (DWARF/ELF sections) and invoking targets over a debugfs interface. It provides a macro-based API for constrained inputs and annotations, integrates with syzkaller, and demonstrates it finds overflow in ~30s on a laptop for an injected PKCS7 off‑by‑one. A custom serialization format handles pointer relocation and sandboxing; a simple-blob macro and an in-tree bridge tool reduce that burden. Work is under upstream review (third patch iteration).
AI-written summary · about
chapters (6)
- 0:00Introduction to KFuzzTest
- 1:23How KFuzzTest works
- 4:55Usage example with PKCS7
- 6:19Serialization complexity
- 7:25Current state and future
- 9:30Q&A session
related talks
46mBypass Kernel Barriers: Fuzzing Linux Kernel in Userspace With LKL
28mkAFL: Hardware-Assisted Feedback Fuzzing for OS Kernels
31m“Fuzz or lose..."
21mModern Testing and Fuzzing in Bitcoin Core
26mFuzzing Host-to Guest Attack Surface in Android Protected KVM
50mWhat the Fuzz
29mVirtual Device Fuzzing Support in QEMU
01h 15mMake your programs more reliable with Fuzzing