Fuzz your Rust code with Google-developped Honggfuzz !
Honggfuzz is a security oriented fuzzer with powerful analysis options. Supports evolutionary, feedback-driven fuzzing based on code coverage (software- and hardware-based).
- project homepage honggfuzz.com
- project repository github.com/google/honggfuzz
- this upstream project is maintained by Google, but ...
- this is NOT an official Google product
- Rust: stable, beta, nightly
- OS: GNU/Linux, macOS, FreeBSD, Android, WSL (Windows Subsystem for Linux)
- Arch: x86_64, x86, arm64-v8a, armeabi-v7a, armeabi
- Sanitizer: none, address, thread, leak
Install honggfuzz commands to build with instrumentation and fuzz
# installs hfuzz and honggfuzz subcommands in cargo
cargo install honggfuzz
Add to your dependencies
[dependencies]
honggfuzz = "0.5"
Create a target to fuzz
#[macro_use] extern crate honggfuzz;
fn main() {
// Here you can parse `std::env::args and
// setup / initialize your project
// You have full control over the loop but
// you're supposed to call `fuzz` ad vitam aeternam
loop {
// The fuzz macro gives an arbitrary object (see `arbitrary crate`)
// to a closure-like block of code.
// For performance reasons, it is recommended that you use the native type
// `&[u8]` when possible.
// Here, this slice will contain a "random" quantity of "random" data.
fuzz!(|data: &[u8]| {
if data.len() != 6 {return}
if data[0] != b'q' {return}
if data[1] != b'w' {return}
if data[2] != b'e' {return}
if data[3] != b'r' {return}
if data[4] != b't' {return}
if data[5] != b'y' {return}
panic!("BOOM")
});
}
}
Fuzz for fun and profit !
# builds with fuzzing instrumentation and then fuzz the "example" target
cargo hfuzz run example
Once you got a crash, replay it easily in a debug environment
# builds the target in debug mode and replays automatically the crash in rust-lldb
cargo hfuzz run-debug example fuzzing_workspace/*.fuzz
You can also build and run your project without compile-time software instrumentation (LLVM's SanCov passes)
This allows you for example to try hardware-only feedback driven fuzzing:
# builds without fuzzing instrumentation and then fuzz the "example" target using hardware-based feedback
HFUZZ_RUN_ARGS="--linux_perf_ipt_block --linux_perf_instr --linux_perf_branch" cargo hfuzz run-no-instr example
Clean
# a wrapper on "cargo clean" which cleans the fuzzing_target directory
cargo hfuzz clean
Version
cargo hfuzz version
You can use RUSTFLAGS
to send additional arguments to rustc
.
For instance, you can enable the use of LLVM's sanitizers.
This is a recommended option if you want to test your unsafe
rust code but it will have an impact on performance.
RUSTFLAGS="-Z sanitizer=address" cargo hfuzz run example
You can use HFUZZ_BUILD_ARGS
to send additional arguments to cargo build
.
You can use HFUZZ_RUN_ARGS
to send additional arguments to honggfuzz
.
See USAGE for the list of those.
For example:
# 1 second of timeout
# use 12 fuzzing thread
# be verbose
# stop after 1000000 fuzzing iteration
# exit upon crash
HFUZZ_RUN_ARGS="-t 1 -n 12 -v -N 1000000 --exit_upon_crash" cargo hfuzz run example
By default we use rust-lldb
but you can change it to rust-gdb
, gdb
, /usr/bin/lldb-7
...
Target compilation directory, defaults to hfuzz_target
to not clash with cargo build
's default target
directory.
Honggfuzz working directory, defaults to hfuzz_workspace
.
Honggfuzz input files (also called "corpus"), defaults to $HFUZZ_WORKSPACE/{TARGET}/input
.
Sometimes, it is necessary to make some specific adaptation to your code to yield a better fuzzing efficiency.
For instance:
- Make you software behavior as much as possible deterministic on the fuzzing input
- PRNGs must be seeded with a constant or the fuzzer input
- Behavior shouldn't change based on the computer's clock.
- Avoid potential undeterministic behavior from racing threads.
- ...
- Never ever call
std::process::exit()
. - Disable logging and other unnecessary functionnalities.
- Try to avoid modifying global state when possible.
When building with cargo hfuzz
, the argument --cfg fuzzing
is passed to rustc
to allow you to condition the compilation of thoses adaptations thanks to the cfg
macro like so:
#[cfg(fuzzing)]
let mut rng = rand::StdRng::from_seed(&[0]);
#[cfg(not(fuzzing))]
let mut rng = rand::thread_rng();
Also, when building in debug mode, the fuzzing_debug
argument is added in addition to fuzzing
.
For more information about conditional compilation, please see the reference.
There is other projects providing Rust fuzzing support at github.com/rust-fuzz.
You'll find support for AFL and LLVM's LibFuzzer and there is also a trophy case ;-) .
This crate was inspired by those projects!