Hyorm/RA_fuzzing

python code of the fuzzing book to the C language

RA_fuzzing

fuzzing algorithm, C with fuzzing book

• 20-1 Research Internship
• Personal Project with Professor

Summary

I performed the task of changing the python code of the fuzzing book to the C language. This process allows me to learn the fuzzing algorithm. The code of the fuzzing book is implemented as an object-oriented algorithm, so I overcomed the limitations of the c language using the structure of the c language. After completing this process, I conducted a fuzzing test using Google/fuzzer-test-suite with Anora, AFL, and LibFuzzer and performed a comparative analysis of the differences and performance of each fuzzer. The comparison data will be uploaded later.

Comparison(draft)🔧

Target Program

• guetzli
• libarchive
• lcms
• libjpeg-turbo
• Harfbuzz

Reason

• Google/fuzzer-test-suite : 24
• Condition
  • Seeds - 12
  • Not too big to angora dry run - 11
  • the libfuzzing engine not too tightly attached to the target program – 8
  • Gcov must be attachable -5

Graph (draft, some results per module)

This project has not been terminated. Consequently, results and information cannot be fully disclosed.

• Orenge: Angora
• Blue: AFL

Figure 1. Guetzli jpeg_data_reader

Figure 2. harfbuzz hp-op-shape

Figure 3. Lcms cmsio0

Figure 4. Libarchive archive_read

Figure 5. Libjpeg jdmarker