1# Fuzzing
2
3## Test Case Generators and Oracles
4
5Test case generators and oracles live in the `wasmtime-fuzzing` crate, located
6in the `crates/fuzzing` directory.
7
8A *test case generator* takes raw, unstructured input from a fuzzer and
9translates that into a test case. This might involve interpreting the raw input
10as "DNA" or pre-determined choices through a decision tree and using it to
11generate an in-memory data structure, or it might be a no-op where we interpret
12the raw bytes as if they were Wasm.
13
14An *oracle* takes a test case and determines whether we have a bug. For example,
15one of the simplest oracles is to take a Wasm binary as an input test case,
16validate and instantiate it, and (implicitly) check that no assertions failed or
17segfaults happened. A more complicated oracle might compare the result of
18executing a Wasm file with and without optimizations enabled, and make sure that
19the two executions are observably identical.
20
21Our test case generators and oracles strive to be fuzzer-agnostic: they can be
22reused with libFuzzer or AFL or any other fuzzing engine or driver.
23
24## libFuzzer and `cargo fuzz` Fuzz Targets
25
26We combine a test case generator and one more oracles into a *fuzz
27target*. Because the target needs to pipe the raw input from a fuzzer into the
28test case generator, it is specific to a particular fuzzer. This is generally
29fine, since they're only a couple of lines of glue code.
30
31Currently, all of our fuzz targets are written for
32[libFuzzer](https://www.llvm.org/docs/LibFuzzer.html) and [`cargo
33fuzz`](https://rust-fuzz.github.io/book/cargo-fuzz.html). They are defined in
34the `fuzz` subdirectory.
35
36See
37[`fuzz/README.md`](https://github.com/bytecodealliance/wasmtime/blob/main/fuzz/README.md)
38for details on how to run these fuzz targets and set up a corpus of seed inputs.
39