Struct NlSocketHandle

Source

pub struct NlSocketHandle { /* private fields */ }

Expand description

Tokio-enabled Netlink socket struct

Implementations§

Source §

impl NlSocketHandle

Source

pub fn connect( proto: NlFamily, pid: Option<u32>, groups: Groups, ) -> Result<Self, SocketError>

Set up asynchronous socket handle.

Source

pub fn add_mcast_membership(&self, groups: Groups) -> Result<(), SocketError>

Join multicast groups for a socket.

Source

pub fn drop_mcast_membership(&self, groups: Groups) -> Result<(), SocketError>

Leave multicast groups for a socket.

Source

pub fn list_mcast_membership(&self) -> Result<NetlinkBitArray, SocketError>

List joined groups for a socket.

Source

pub fn pid(&self) -> u32

Get the PID for the current socket.

Source

pub async fn send<T, P>(&self, msg: &Nlmsghdr<T, P>) -> Result<(), SocketError>
where T: NlType, P: Size + ToBytes,

Send a message on the socket asynchronously.

Source

pub async fn recv<T, P>( &self, ) -> Result<(NlBufferIter<T, P, BufferPoolGuard<'_>>, Groups), SocketError>
where T: NlType, P: Size + FromBytesWithInput<Input = usize>,

Receive a message from the socket asynchronously.

Source

pub async fn recv_all<T, P>( &self, ) -> Result<(NlBuffer<T, P>, Groups), SocketError>
where T: NlType + Debug, P: Size + FromBytesWithInput<Input = usize> + Debug,

Parse all Nlmsghdr structs sent in one network packet and return them all in a list.

Failure to parse any packet will cause the entire operation to fail. If an error is detected at the application level, this method will discard any non-error Nlmsghdr structs and only return the error. For a more granular approach, use either NlSocketHandle::recv.

Source

pub fn set_recv_buffer_size(&self, size: usize) -> Result<(), SocketError>

Set the size of the receive buffer for the socket.

This can be useful when communicating with a service that sends a high volume of messages (especially multicast), and your application cannot process them fast enough, leading to the kernel dropping messages. A larger buffer may help mitigate this.

The value passed is a hint to the kernel to set the size of the receive buffer. The kernel will double the value provided to account for bookkeeping overhead. The doubled value is capped by the value in /proc/sys/net/core/rmem_max.

The default value is /proc/sys/net/core/rmem_default

See socket(7) documentation for SO_RCVBUF for more information.

Source

pub fn enable_ext_ack(&self, enable: bool) -> Result<(), SocketError>

If true is passed in, enable extended ACKs for this socket. If false is passed in, disable extended ACKs for this socket.

Source

pub fn get_ext_ack_enabled(&self) -> Result<bool, SocketError>

Return true if an extended ACK is enabled for this socket.

Source

pub fn enable_strict_checking(&self, enable: bool) -> Result<(), SocketError>

If true is passed in, enable strict checking for this socket. If false is passed in, disable strict checking for for this socket. Only supported by NlFamily::Route sockets. Requires Linux >= 4.20.

Source