Magma

A ground-truth binary fuzzing benchmark suite based on real programs with real bugs

Install Docker Engine by following steps from this link

Then, we need to install some other dependencies and clone the repository

sudo apt install -y util-linux inotify-tools git
git clone --branch v1.0.4 https://github.com/HexHive/magma.git

Now, you just need to configure the parameters of captainrc file located in the magma/tools/captain directory for selecting different fuzzers and their targets with the time, no. of cores of CPU and no. of cycles alloted to them. Then, execute ./run.sh

For fuzzer based on AFL, you need to edit magma/tools/benchd/fuzzers/afl.py file and uncomment the part below.

os.system("sudo bash -c 'echo core >/proc/sys/kernel/core_pattern'")
os.system("sudo bash -c 'cd /sys/devices/system/cpu; echo performance | tee cpu*/cpufreq/scaling_governor'")

After executing the run.sh script, it will generate the output and other files in workdir directory in the same directory. For information related to what each files in it meant and to configure the run.sh, follow this link

This fuzzing benchmark is still in its infancy and requires a lot of work to be done like documentation on how to add other fuzzers as well, toolset required to visualize the results. But, still it is well documented and easy to setup as each fuzzer with its target runs in a separate docker container