net-lattice-backend-linux

Linux backend for Net Lattice using Netlink and /etc/resolv.conf mechanisms. It implements native inspection, mutation, and monitoring behind the generic net-lattice-platform contracts.

What it provides

• interface, address, route, neighbor, and resolver inspection;
• route, address, resolver, and interface MTU/administrative-state mutation;
• Netlink change subscriptions and optional native async delivery;
• translation of native errors and state into portable Net Lattice types.

Applications should normally use the net-lattice facade, which selects this backend automatically on Linux. Direct use is intended for backend integration and diagnostics.

Direct usage

use net_lattice_platform::InterfaceProvider;

fn main() -> net_lattice_core::Result<()> {
    let backend = net_lattice_backend_linux::LinuxBackend::new()?;
    for interface in backend.interfaces()? {
        println!("{interface:?}");
    }
    Ok(())
}

Privileges and safety

Read-only operations generally require no elevated privilege. Interface, address, and route changes require CAP_NET_ADMIN; resolver replacement also depends on filesystem permissions and the host resolver manager. An interface configuration request may carry MTU and administrative-state changes together, which Linux can reject after applying one field, so callers must re-read state after an error and use explicit transaction compensation if restoration is required. Privileged tests are ignored in ordinary test runs and restore the observed interface configuration on every exit path.