Challenges in Collecting Code Coverage from Blackbox Binaries During Fuzz Testing

This article discusses the evolution of fuzz testing, specifically focusing on the challenges and techniques for collecting code coverage from blackbox binaries. It contrasts older, feedback-free fuzzing approaches with modern coverage-guided methods, and explores the "wrong way" to approach binary coverage — likely involving inefficient or incorrect instrumentation techniques. The piece provides technical depth on how coverage data can be gathered from compiled binaries without source code access, and the pitfalls that come with naive implementations.