Merge pull request #619 from fitzgen/introduce-wasmtime-fuzzing-crate

Introduce the `wasmtime-fuzzing` crate

crates/fuzzing/Cargo.toml

@@ -0,0 +1,17 @@
+[package]
+authors = ["The Wasmtime Project Developers"]
+description = "Fuzzing infrastructure for Wasmtime"
+edition = "2018"
+name = "wasmtime-fuzzing"
+publish = false
+version = "0.1.0"
+
+# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
+
+[dependencies]
+arbitrary = "0.2.0"
+binaryen = "0.8.2"
+cranelift-codegen = "0.50.0"
+cranelift-native = "0.50.0"
+wasmparser = "0.42.1"
+wasmtime-jit = { path = "../jit" }

crates/fuzzing/README.md

@@ -0,0 +1,14 @@
+# Fuzzing Infrastructure for Wasmtime
+
+This crate provides test case generators and oracles for use with fuzzing.
+
+These generators and oracles are generally independent of the fuzzing engine
+that might be using them and driving the whole fuzzing process (e.g. libFuzzer
+or AFL). As such, this crate does *not* contain any actual fuzz targets
+itself. Those are generally just a couple lines of glue code that plug raw input
+from (for example) `libFuzzer` into a generator, and then run one or more
+oracles on the generated test case.
+
+If you're looking for the actual fuzz target definitions we currently have, they
+live in `wasmtime/fuzz/fuzz_targets/*` and are driven by `cargo fuzz` and
+`libFuzzer`.

crates/fuzzing/src/generators.rs

@@ -0,0 +1,29 @@
+//! Test case generators.
+//!
+//! Test case generators take raw, unstructured input from a fuzzer
+//! (e.g. libFuzzer) and translate that into a structured test case (e.g. a
+//! valid Wasm binary).
+//!
+//! These are generally implementations of the `Arbitrary` trait, or some
+//! wrapper over an external tool, such that the wrapper implements the
+//! `Arbitrary` trait for the wrapped external tool.
+
+use arbitrary::{Arbitrary, Unstructured};
+
+/// A Wasm test case generator that is powered by Binaryen's `wasm-opt -ttf`.
+pub struct WasmOptTtf {
+    /// The raw, encoded Wasm bytes.
+    pub wasm: Vec<u8>,
+}
+
+impl Arbitrary for WasmOptTtf {
+    fn arbitrary<U>(input: &mut U) -> Result<Self, U::Error>
+    where
+        U: Unstructured + ?Sized,
+    {
+        let seed: Vec<u8> = Arbitrary::arbitrary(input)?;
+        let module = binaryen::tools::translate_to_fuzz_mvp(&seed);
+        let wasm = module.write();
+        Ok(WasmOptTtf { wasm })
+    }
+}

crates/fuzzing/src/lib.rs

@@ -0,0 +1,2 @@
+pub mod generators;
+pub mod oracles;

crates/fuzzing/src/oracles.rs

@@ -0,0 +1,66 @@
+//! Oracles.
+//!
+//! Oracles take a test case and determine whether we have a bug. For example,
+//! one of the simplest oracles is to take a Wasm binary as our input test case,
+//! validate and instantiate it, and (implicitly) check that no assertions
+//! failed or segfaults happened. A more complicated oracle might compare the
+//! result of executing a Wasm file with and without optimizations enabled, and
+//! make sure that the two executions are observably identical.
+//!
+//! When an oracle finds a bug, it should report it to the fuzzing engine by
+//! panicking.
+
+use cranelift_codegen::settings;
+use std::cell::RefCell;
+use std::collections::HashMap;
+use std::rc::Rc;
+use wasmtime_jit::{CompilationStrategy, CompiledModule, Compiler, NullResolver};
+
+fn host_isa() -> Box<dyn cranelift_codegen::isa::TargetIsa> {
+    let flag_builder = settings::builder();
+    let isa_builder = cranelift_native::builder().expect("host machine is not a supported target");
+    isa_builder.finish(settings::Flags::new(flag_builder))
+}
+
+/// Instantiate the Wasm buffer, and implicitly fail if we have an unexpected
+/// panic or segfault or anything else that can be detected "passively".
+///
+/// Performs initial validation, and returns early if the Wasm is invalid.
+///
+/// You can control which compiler is used via passing a `CompilationStrategy`.
+pub fn instantiate(wasm: &[u8], compilation_strategy: CompilationStrategy) {
+    if wasmparser::validate(wasm, None).is_err() {
+        return;
+    }
+
+    let isa = host_isa();
+    let mut compiler = Compiler::new(isa, compilation_strategy);
+    let mut imports_resolver = NullResolver {};
+
+    wasmtime_jit::instantiate(
+        &mut compiler,
+        wasm,
+        &mut imports_resolver,
+        Default::default(),
+        true,
+    )
+    .expect("failed to instantiate valid Wasm!");
+}
+
+/// Compile the Wasm buffer, and implicitly fail if we have an unexpected
+/// panic or segfault or anything else that can be detected "passively".
+///
+/// Performs initial validation, and returns early if the Wasm is invalid.
+///
+/// You can control which compiler is used via passing a `CompilationStrategy`.
+pub fn compile(wasm: &[u8], compilation_strategy: CompilationStrategy) {
+    if wasmparser::validate(wasm, None).is_err() {
+        return;
+    }
+
+    let isa = host_isa();
+    let mut compiler = Compiler::new(isa, compilation_strategy);
+    let mut resolver = NullResolver {};
+    let global_exports = Rc::new(RefCell::new(HashMap::new()));
+    let _ = CompiledModule::new(&mut compiler, wasm, &mut resolver, global_exports, false);
+}