Oriel — AI & technology worth noticing, every day.

Tips for more effective fuzz testing with AFL++

Web page · https://nullprogram.com/blog/2025/02/05

First seen by Oriel on .

AI summary

This checklist presents practical AFL++ techniques for finding more defects in source-available C and C++ programs on glibc-based Linux systems.

  • The advice is based on fuzz testing approximately 1,000 projects over six years and complements AFL++ documentation.
  • Use ASan and UBSan with abort_on_error=1, halt_on_error=1, and symbolize=0; enable detect_stack_use_after_return and larger allocation filling for greater sensitivity.
  • Prefer persistent mode, typically an order of magnitude faster, and include selected source files directly in one translation unit to reach static functions and simplify builds.
  • Isolate persistent tests by resetting global state and initializing memory, while copying AFL++ input into a correctly sized buffer because direct buffer use can hide read overflows.
  • Use memfd_create with /proc/self/fd paths for file-based targets, and reduce oversized fixed buffers so fuzzing reaches boundary conditions sooner.

#afl++ #fuzz testing #c++ #address sanitizer

Generated by gpt-5.6-sol. Check the original for details.

Metric trends

Last 30 days · Up to 120 recorded samples per source and metric, with reposts kept separate · First and latest retained

Points

Score reported by the source.

  • Hacker News: 153, latest observation ; +149 over observed period since

Comments

Comment count reported by the source.

  • Hacker News: 28, latest observation ; +28 over observed period since

Aliases

SourceExternal IDURL
Hacker News50271106https://nullprogram.com/blog/2025/02/05/ Read discussion