Struct netlink_sys::Socket

pub struct Socket(_);

A netlink socket.

Example

In this example we:

1. open a new socket
2. send a message to the kernel
3. read the reponse

use netlink_sys::{protocols::NETLINK_ROUTE, Socket, SocketAddr};
use std::process;

// open a new socket for the NETLINK_ROUTE subsystem (see "man 7 rtnetlink")
let mut socket = Socket::new(NETLINK_ROUTE).unwrap();
// address of the remote peer we'll send a message to. This particular address is for the kernel
let kernel_addr = SocketAddr::new(0, 0);
// this is a valid message for listing the network links on the system
let pkt = vec![
    0x14, 0x00, 0x00, 0x00, 0x12, 0x00, 0x01, 0x03, 0xfd, 0xfe, 0x38, 0x5c, 0x00, 0x00, 0x00,
    0x00, 0x00, 0x00, 0x00, 0x00,
];
// send the message to the kernel
let n_sent = socket.send_to(&pkt[..], &kernel_addr, 0).unwrap();
assert_eq!(n_sent, pkt.len());
// buffer for receiving the response
let mut buf = vec![0; 4096];
loop {
    // receive a datagram
    let (n_received, sender_addr) = socket.recv_from(&mut &mut buf[..], 0).unwrap();
    assert_eq!(sender_addr, kernel_addr);
    println!("received datagram {:?}", &buf[..n_received]);
    if buf[4] == 2 && buf[5] == 0 {
        println!("the kernel responded with an error");
        return;
    }
    if buf[4] == 3 && buf[5] == 0 {
        println!("end of dump");
        return;
    }
}

Implementations

impl Socket

pub fn new(protocol: isize) -> Result<Self>

Open a new socket for the given netlink subsystem. protocol must be one of the netlink_sys::protocols constants.

pub fn bind(&mut self, addr: &SocketAddr) -> Result<()>

Bind the socket to the given address

pub fn bind_auto(&mut self) -> Result<SocketAddr>

Bind the socket to an address assigned by the kernel, and return that address.

pub fn get_address(&self, addr: &mut SocketAddr) -> Result<()>

Get the socket address

pub fn set_non_blocking(&self, non_blocking: bool) -> Result<()>

Make this socket non-blocking

pub fn connect(&self, remote_addr: &SocketAddr) -> Result<()>

Connect the socket to the given address. Netlink is a connection-less protocol, so a socket can communicate with multiple peers with the Socket::send_to and Socket::recv_from methods. However, if the socket only needs to communicate with one peer, it is convenient not to have to bother with the peer address. This is what connect is for. After calling connect, Socket::send and Socket::recv respectively send and receive datagrams to and from remote_addr.

Examples

In this example we:

1. open a socket
2. connect it to the kernel with Socket::connect
3. send a request to the kernel with Socket::send
4. read the response (which can span over several messages) Socket::recv

use netlink_sys::{protocols::NETLINK_ROUTE, Socket, SocketAddr};
use std::process;

let mut socket = Socket::new(NETLINK_ROUTE).unwrap();
let _ = socket.bind_auto().unwrap();
let kernel_addr = SocketAddr::new(0, 0);
socket.connect(&kernel_addr).unwrap();
// This is a valid message for listing the network links on the system
let msg = vec![
    0x14, 0x00, 0x00, 0x00, 0x12, 0x00, 0x01, 0x03, 0xfd, 0xfe, 0x38, 0x5c, 0x00, 0x00, 0x00,
    0x00, 0x00, 0x00, 0x00, 0x00,
];
let n_sent = socket.send(&msg[..], 0).unwrap();
assert_eq!(n_sent, msg.len());
// buffer for receiving the response
let mut buf = vec![0; 4096];
loop {
    let mut n_received = socket.recv(&mut &mut buf[..], 0).unwrap();
    println!("received {:?}", &buf[..n_received]);
    if buf[4] == 2 && buf[5] == 0 {
        println!("the kernel responded with an error");
        return;
    }
    if buf[4] == 3 && buf[5] == 0 {
        println!("end of dump");
        return;
    }
}

pub fn recv_from<B>(
    &self,
    buf: &mut B,
    flags: c_int
) -> Result<(usize, SocketAddr)>where
    B: BufMut,

Read a datagram from the socket and return the number of bytes that have been read and the address of the sender. The data being read is copied into buf. If buf is too small, the datagram is truncated. The supported flags are the MSG_* described in man 2 recvmsg

Warning

In datagram oriented protocols, recv and recvfrom receive normally only ONE datagram, but this seems not to be always true for netlink sockets: with some protocols like NETLINK_AUDIT, multiple netlink packets can be read with a single call.

pub fn recv<B>(&self, buf: &mut B, flags: c_int) -> Result<usize>where
    B: BufMut,

For a connected socket, recv reads a datagram from the socket. The sender is the remote peer the socket is connected to (see Socket::connect). See also Socket::recv_from

pub fn recv_from_full(&self) -> Result<(Vec<u8>, SocketAddr)>

Receive a full message. Unlike Socket::recv_from, which truncates messages that exceed the length of the buffer passed as argument, this method always reads a whole message, no matter its size.

pub fn send_to(&self, buf: &[u8], addr: &SocketAddr, flags: c_int) -> Result<usize>

Send the given buffer buf to the remote peer with address addr. The supported flags are the MSG_* values documented in man 2 send.

pub fn send(&self, buf: &[u8], flags: c_int) -> Result<usize>

For a connected socket, send sends the given buffer buf to the remote peer the socket is connected to. See also Socket::connect and Socket::send_to.

pub fn set_pktinfo(&mut self, value: bool) -> Result<()>

pub fn get_pktinfo(&self) -> Result<bool>

pub fn add_membership(&mut self, group: u32) -> Result<()>

pub fn drop_membership(&mut self, group: u32) -> Result<()>

pub fn set_broadcast_error(&mut self, value: bool) -> Result<()>

NETLINK_BROADCAST_ERROR (since Linux 2.6.30). When not set, netlink_broadcast() only reports ESRCH errors and silently ignore NOBUFS errors.

pub fn get_broadcast_error(&self) -> Result<bool>

pub fn set_no_enobufs(&mut self, value: bool) -> Result<()>

NETLINK_NO_ENOBUFS (since Linux 2.6.30). This flag can be used by unicast and broadcast listeners to avoid receiving ENOBUFS errors.

pub fn get_no_enobufs(&self) -> Result<bool>

pub fn set_listen_all_namespaces(&mut self, value: bool) -> Result<()>

NETLINK_LISTEN_ALL_NSID (since Linux 4.2). When set, this socket will receive netlink notifications from all network namespaces that have an nsid assigned into the network namespace where the socket has been opened. The nsid is sent to user space via an ancillary data.

pub fn get_listen_all_namespaces(&self) -> Result<bool>

pub fn set_cap_ack(&mut self, value: bool) -> Result<()>

NETLINK_CAP_ACK (since Linux 4.2). The kernel may fail to allocate the necessary room for the acknowledgment message back to user space. This option trims off the payload of the original netlink message. The netlink message header is still included, so the user can guess from the sequence number which message triggered the acknowledgment.

pub fn get_cap_ack(&self) -> Result<bool>

Trait Implementations

impl AsRawFd for Socket

fn as_raw_fd(&self) -> RawFd

Extracts the raw file descriptor.

impl Clone for Socket

fn clone(&self) -> Socket

Returns a copy of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for Socket

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Drop for Socket

fn drop(&mut self)

Executes the destructor for this type.

impl FromRawFd for Socket

unsafe fn from_raw_fd(fd: RawFd) -> Self

Constructs a new instance of Self from the given raw file descriptor.