Type-safe Rust bindings for encoding/decoding Netlink messages generated from YAML specifications.
Netlink is a collection of APIs exposed by the kernel, unified by a similar style of encoding data. The general list of Netlink families can be found in the kernel documentation, or at least those families, that were possible to condense to a machine readable description. This list very likely includes all the families that you would want to interact with.
The goal of this project is to provide an easy-to-use type-safe interface, while also being reasonably fast and supporting all properties of all sensible Netlink families.
- Streamlined type-safe interface.
- All the properties described in yaml specifications are available in Rust.
- Decoded Netlink messages can be Debug-printed, with enum values and flags annotated.
- There is a tool to simplify working off of existing tools by decoding Netlink messages of other programs.
- Unified interface for both generic and classic flavors of netlink. Family resolution is automatic.
- Many quirks of netlink-legacy and netlink-raw protocols are supported, including: binary structures, nested types, multi-attribute arrays, indexed arrays, sub messages.
- Zero unsafe code in encoder/decoder logic.
[dependencies]
netlink-bindings = { version = "0.2", features = [ "wireguard" ] }
netlink-socket2 = { version = "0.2", features = [ ] }A typical Netlink family, say wireguard, support multiple operations, for example, "get-device" and "set-device". Each operation may also be of "do" or "dump" kind.
For example, to get info about a device you would use a "dump" kind of "get-device" request. That's usually what it means, although different subsystems may imply different things. A typical request looks like this:
use netlink_bindings::wireguard;
use netlink_socket2::NetlinkSocket;
let mut sock = NetlinkSocket::new();
let ifname = "wg0";
// All available requests are conveniently accessible using `family::Request`
let mut request = wireguard::Request::new()
.op_get_device_dump_request();
// Add contents to the request
request.encode()
.push_ifname_bytes(ifname.as_bytes());
let mut iter = sock.request(&request).unwrap();
while let Some(res) = iter.recv() {
// Each request may return an error (literal error code), in some cases
// with some additional info from the kernel, e.g. lacking a permission,
// if you missing CAP_NET_ADMIN capability for querying wireguard info.
let attrs = res.unwrap();
// A simple approach to get a specific property from an attribute set is
// following. Note that it's not guaranteed that the property was supplied,
// nor that it can be parsed correctly. If either occurs, the error will
// include error context, i.e. name of the attribute and it's parent set.
let listen_port = attrs.get_listen_port().unwrap();
println!("Interface {ifname:?} is listening on {listen_port}");
// Print out all the attributes using the debug formatter.
println!("{:#?}", attrs);
}Your LSP should be able to nicely suggest appropriate methods both for encoding and decoding as you type.
Let's say you have a network interface and want to assign an ip address to it. This is the domain of "rt-addr" family. It was one of the first ones created, inheriting some now-discouraged quirks, like a fixed-header - a struct that always present and may also carry some relevant data in some cases or may be just ignored (zeroed-out) for other requests.
The relevant operation is "newaddr" with "do" kind. You may also notice ".set_change()". This specifies an additional request flags. Similar to fixed-header, theses flags may invoke some additional behavior in certain operations, or do nothing in others.
use std::net::IpAddr;
use netlink_bindings::rt_addr;
use netlink_socket2::NetlinkSocket;
let mut sock = NetlinkSocket::new();
let ifindex: u32 = 1234; // Acquired via "get-addr" request
let addr: IpAddr = "10.0.0.1".parse().unwrap();
let prefix: u8 = 32; // stands for "/32" in "10.0.0.1/32"
// Create fixed-header for the request
let mut header = rt_addr::PushIfaddrmsg::new();
header.set_ifa_index(ifindex);
header.set_ifa_family(libc::AF_INET as u8); // aka ipv4
header.set_ifa_prefixlen(prefix);
let mut request = rt_addr::Request::new()
.set_change() // Don't fail if address already assigned
.op_newaddr_do_request(&header);
request.encode()
.push_local(addr);
sock.request(&request).unwrap()
.recv_ack().unwrap();See full code in the example.
Async functionality is available using the same interface, you just need to
enable it, and to add .await keyword in all places where async IO is expected.
[dependencies]
netlink-socket2 = { ... , features = [ "tokio" ] } # or "smol"An earlier example, but using async, would look like this:
use netlink_bindings::wireguard;
use netlink_socket2::NetlinkSocket;
let mut sock = NetlinkSocket::new();
let mut request = wireguard::Request::new()
.op_get_device_dump_request();
request.encode()
.push_ifname_bytes(b"wg0");
let mut iter = sock.request(&request).await.unwrap();
while let Some(res) = iter.recv().await {
println!("{:#?}", res);
}- wireguard-setup - Create and configure wireguard interface.
- conntrack - Dump tracked network
connections, similar to
conntrack -L. - extack - Showcase handing of extended ack attributes in error reporting.
- nftables - Create nftables rules.
- nftables-api - A high-level wrapper for creating nftables rules.
Under the hood, calling .encode() is just a convenience to pass the lead to
the correct encoding struct. The ecoder's job is to actually write the
attributes directly into provided buffer. All types relevant for encoding are
prefixed with Push. For example, directly encoding a "do" request of
"set-device" operation looks like the following:
use netlink_bindings::wireguard::{PushOpSetDeviceDoRequest, WgdeviceFlags};
let mut vec = Vec::new();
// Do set-device (request)
PushOpSetDeviceDoRequest::new(&mut vec)
.push_ifname(c"wg0") // &CStr
// .push_ifname_bytes("wg0".as_bytes()) // &[u8]
.push_flags(WgdeviceFlags::ReplacePeers as u32) // Remove existing peers
.array_peers()
.entry_nested()
.push_public_key(&[/* ... */]) // &[u8]
.push_endpoint("127.0.0.1:12345".parse().unwrap()) // SocketAddr
.array_allowedips()
.entry_nested()
.push_family(libc::AF_INET as u16) // aka ipv4
.push_ipaddr("0.0.0.0".parse().unwrap()) // IpAddr
.push_cidr_mask(0) // stands for "/0" in "0.0.0.0/0"
.end_nested()
// More allowed ips...
.end_array() // Explicitly closing isn't necessary
.end_nested()
// More peers...
.end_array();Additionally, check out all available methods, along with the in-line documentation.
Similarly, under the hood, receiving a reply yield an attribute decoder. The decoder itself is just a slice, therefore it can be cheaply cloned, copying it's frame. The low-level interface is based on iterators, with nice-to-use wrapper on top.
use netlink_bindings::wireguard::OpGetDeviceDumpReply;
let buf = vec![/* ... */];
// Dump get-device (reply)
let attrs = OpGetDeviceDumpReply::new(&buf);
println!("Ifname: {:?}", attrs.get_ifname().unwrap()); // &CStr
for peer in attrs.get_peers().unwrap() {
println!("Endpoint: {}", peer.get_endpoint().unwrap()); // SockAddr
for addr in peer.get_allowedips().unwrap() {
let ip = addr.get_ipaddr().unwrap(); // IpAddr
let mask = addr.get_cidr_mask().unwrap(); // u8
println!("Allowed ip: {ip}/{mask}");
}
}See full code in the example. And as previously, check out all available methods, along with the in-line documentation.
A low-level decoding interface is exposed as an iterator, that yields enum variants, containing either a target type, e.g. SockAddr, or another iterator, in case of a nested attribute set. But as you can see, using it directly quickly turns very ugly.
use netlink_bindings::wireguard::{OpGetDeviceDumpReply, Wgpeer};
let buf = vec![/* ... */];
for attr in OpGetDeviceDumpReply::new(&buf) {
match attr.unwrap() {
OpGetDeviceDumpReply::Ifname(n) => println!("Ifname: {n:?}"),
OpGetDeviceDumpReply::Peers(iter) => {
for entry in iter {
for attr in entry.unwrap() {
match attr.unwrap() {
Wgpeer::Endpoint(e) => println!("Endpoint: {e:?}"),
_ => {}
}
}
}
}
_ => {}
}
}Both don't use codegen to generate bindings, hence many Netlink families are not supported. Another difference is that they represent netlink messages as lists of rust enums, while this project works with the binary representation directly, with a separate interfaces for encoding and decoding: a builder pattern-like interface for encoding, and an iterator interface for decoding (internally).
- ✅ - supported, has tests.
- ✔️ - compiles, testing needed.
- • - compiles.
- ? - not attempted.
- ❌ - doesn't compile (needs adaptations in codegen).
| subsystem | ? | comment |
|---|---|---|
| nlctrl | ✅ | |
| rt-addr | ✅ | |
| rt-link | ✅ | |
| wireguard | ✅ | |
| nftables | ✅ | |
| conntrack | ✔️ | |
| rt-neigh | ✔️ | |
| rt-route | ✔️ | |
| rt-rule | ✔️ | |
| tc | ✔️ | |
| devlink | • | |
| ethtool | • | |
| dpll | ? | |
| fou | ? | |
| handshake | ? | |
| lockd | ? | |
| mptcp_pm | ? | |
| net-shaper | ? | |
| netdev | ? | |
| nfsd | ? | |
| nl80211 | ? | |
| ovpn | ? | |
| ovs_datapath | ? | |
| ovs_flow | ? | |
| ovs_vport | ? | |
| tcp_metrics | ? | |
| team | ? |
The following netlink features are not implemented (yet):
- Attributes denoting a C array (i.e. attributes with type binary and sub-type u32/u64).
- Events/notifications/multicast messages.
- Sub messages with a selector attribute outside of the parent attribute set.
If there's an existing tool using Netlink, you can use reverse-lookup binary
from this project to decipher it's Netlink communications, and work off of
that. Let's say you want to inspect what wg command does.
$ strace -o ./output_file --decode-fd=socket -e %network --{write,read}=$(seq -s, 0 100) -- wg
$ cargo run --bin reverse-lookup --features=wireguard,nlctrl,rt-link -- ./output_file
Decoding request in family ROUTE flags=[REQUEST,ACK,DUMP,REPLACE,EXCL] Raw { protonum: 0, request_type: 0 }
...Update protocol bindings using the yaml description from the protocol directory.
$ cargo run --bin codegen -- -d ./netlink-bindings/src/wireguard/
Writing "netlink-bindings/src/wireguard/mod.rs"
Dumping "netlink-bindings/src/wireguard/wireguard.md"
Dumping all "netlink-bindings/src/wireguard/wireguard-all.md"Generate protocol bindings for a new family, copying yaml specification from somewhere else.
$ cargo run --bin codegen -- -d ./netlink-bindings/src/ linux/Documentation/netlink/specs/wireguard.yaml
Writing "netlink-bindings/src/wireguard/mod.rs"
Dumping "netlink-bindings/src/wireguard/wireguard.md"
Dumping all "netlink-bindings/src/wireguard/wireguard-all.md"- Testing (for each sensible netlink family and for parsing primitives).
- Simplify codegen logic.
- Clean up unintentional panics in encoding/decoding.
- Optimize generated code size, e.g. leave out code encoding kernel replies by default.
- Improve user interface (better error reporting, tooling, etc).
- Benchmarks, fuzzing.
If your want to contribute, you can, for example:
- Add testing: collect netlink messages and check that they are parsed correctly. See wireguard tests as an example. Additional examples are also very welcome.
- Implement yet unsupported netlink functionality.
- Improve compilation time, reduce the size of generated bindings.
- Experiment with a better Rust interface (for encoding/decoding and the sockets).
- Sponsor the project (contact the author for options).