The process of integrating a project written in Go with OSS-Fuzz is very similar to the general Setting up a new project process. The key specifics of integrating a Go project are outlined below.
OSS-Fuzz supports go-fuzz in the libFuzzer compatible mode only. In that mode, fuzz targets for Go use the libFuzzer engine with native Go coverage instrumentation. Binaries compiled in this mode provide the same libFuzzer command line interface as non-Go fuzz targets.
First, you need to write a Go fuzz target that accepts a stream of bytes and calls the program API with that. This fuzz target should reside in your project repository (example).
The structure of the project directory in OSS-Fuzz repository doesn’t differ for projects written in Go. The project files have the following Go specific aspects.
language attribute must be specified.
The only supported fuzzing engine and sanitizer are
address, respectively. Example:
fuzzing_engines: - libfuzzer sanitizers: - address
The OSS-Fuzz builder image has the latest stable release of Golang installed. In order to install dependencies of your project, add
RUN go get ... command to your Dockerfile. Example:
# Dependency for one of the fuzz targets. RUN go get github.com/ianlancetaylor/demangle
In the case you are using modules, the best practice is to
git clone the repository into the expected
A usage example from go-coredns project is
RUN git clone --depth 1 https://github.com/coredns/coredns $GOPATH/src/github.com/coredns/coredns
In order to build a Go fuzz target, you need to call
go-fuzz command first, and then link the resulting
.a file against
$LIB_FUZZING_ENGINE using the
$CXX $CXXFLAGS ... command.
The best way to do this is by using a
compile_go_fuzzer script, as it also supports coverage builds.
A usage example from go-dns project is
compile_go_fuzzer github.com/miekg/dns FuzzNewRR fuzz_newrr fuzz
Arguments are :
- path of the package with the fuzz target
- name of the fuzz function
- name of the fuzzer to be built
- optional tag to be used by
go buildand such