eBPF is a pure Go library that provides utilities for loading, compiling, and debugging eBPF programs. It has minimal external dependencies and is intended to be used in long running processes.

The library is maintained by Cloudflare and Cilium.

See ebpf.io for other projects from the eBPF ecosystem.

A small collection of Go and eBPF programs that serve as examples for building your own tools can be found under examples/.

Contributions are highly encouraged, as they highlight certain use cases of eBPF and the library, and help shape the future of the project.

Please join the #ebpf-go channel on Slack if you have questions regarding the library.

This library includes the following packages:

• asm contains a basic assembler, allowing you to write eBPF assembly instructions directly within your Go code. (You don't need to use this if you prefer to write your eBPF program in C.)
• cmd/bpf2go allows compiling and embedding eBPF programs written in C within Go code. As well as compiling the C code, it auto-generates Go code for loading and manipulating the eBPF program and map objects.
• link allows attaching eBPF to various hooks
• perf allows reading from a PERF_EVENT_ARRAY
• ringbuf allows reading from a BPF_MAP_TYPE_RINGBUF map
• features implements the equivalent of bpftool feature probe for discovering BPF-related kernel features using native Go.
• rlimit provides a convenient API to lift the RLIMIT_MEMLOCK constraint on kernels before 5.11.

• A version of Go that is supported by upstream
• Linux >= 4.4. CI is run against LTS releases.

Run make in the root of this repository to rebuild testdata in all subpackages. This requires Docker, as it relies on a standardized build environment to keep the build output stable.

It is possible to regenerate data using Podman by overriding the CONTAINER_* variables: CONTAINER_ENGINE=podman CONTAINER_RUN_ARGS= make.

The toolchain image build files are kept in testdata/docker/.

MIT

The eBPF honeygopher is based on the Go gopher designed by Renee French.