Johnathan Corgan
1cc069f1a2
The TUN adapter, the .fips DNS responder, ICMPv6 generation and TCP MSS clamping are the host-side IPv6 plane, but the code was split between src/upper and inline bodies in the node's session handler and lifecycle. Gather it into one module, src/ipv6tun, and reduce its calls into the rest of the crate to a small explicit set. This prepares the TUN adapter to run later as a separate daemon over the native API, and gives embedders one place to find the host-side surface. Behaviour is unchanged apart from log target names. - Rename src/upper to src/ipv6tun. A `pub use ipv6tun as upper;` alias keeps every crate::upper:: and fips::upper:: path resolving, so no consumer has to change. - Move hosts.rs whole to src/hosts.rs, a top-level public module: the hosts file also serves peer display names, the peer ACL and fipsctl, so it is not host-side only. ipv6tun re-exports it for the old path. - Move the DNS socket helpers (dual-stack bind, IPV6_RECVPKTINFO, interface index lookup) from Node into ipv6tun::dns, unchanged. Node::mesh_ifindex, which reads the live TUN device name, becomes Handles::mesh_ifindex. - Move ICMPv6 Destination Unreachable and Packet Too Big sending into ipv6tun::icmp behind IcmpContext, which borrows the TUN channel, our address and the Packet Too Big rate limiter for one use. Packet Too Big is still rate limited and Destination Unreachable still is not. The discovery lookup timeout hands its queued packets over as one no-route report. - Split handle_tun_outbound. The host-side half, in ipv6tun::outbound, validates the packet, makes both Packet Too Big decisions and sends the ICMPv6 replies, reaching the mesh through a small Mesh trait Node implements. The mesh-side half, Node::send_outbound, stays with the pending queue. The checks run in the same order with the same thresholds, and Node::handle_tun_outbound remains the entry point. - Move the TUN and DNS child start and stop bodies into ipv6tun::lifecycle, and gather their nine supervisor fields and the node's TUN device name into one Handles struct the supervisor holds. The supervisor arms, their order and the child-exit reporting are unchanged. A TUN still counts as up when it has a device name, so an app-owned TUN produces no TUN teardown, and DNS counts as up while its task runs. The node passes a new peer-alias base to the running responder through Handles::publish_aliases. Node::tun_name, tun_tx, dns_local_addr and enable_app_owned_tun keep their behaviour; tests install a TUN sender through a test-only Node::install_tun. Tracing targets follow module paths, so lines from the moved code now log under fips::ipv6tun::* and fips::hosts instead of fips::upper::*, fips::node::lifecycle and fips::node::handlers::session. Update the RUST_LOG example in the MTU diagnosis guide and the test harness filters that relied on the old targets, and note the rename in the changelog.
Dynamic Interface Binding
Two FIPS daemons whose only transports are bound to network interfaces, exercised against a veth pair the harness creates, downs, deletes and recreates underneath them while they run.
node-a node-b
lab ve-lab0 required ── veth ── ve-lab0 required
dock fips-dock0 optional fips-dock0 optional
ve-lab0 does not exist when the daemons start. fips-dock0 never exists at
all, on any host, ever — it is the negative control for optional: true.
What it asserts
| Behavior | |
|---|---|
| (a) | A daemon whose only interface is missing starts, reports the transport absent, and reports Degraded — it does not exit on NoTransports, and it does not skip the transport for the life of the process |
| (b) | The interface appears; both daemons bind it with no restart, Degraded clears, and they discover and peer over it |
| (c) | The interface goes down and comes back; presence and health follow it in both directions, and the rebind is counted |
| (d) | The interface is deleted outright and recreated; both daemons rebind and re-peer — the case the old ENXIO beacon-socket reopen half-covered |
| (e) | An optional interface that never appears logs at info and never moves node health |
| Absence is logged once on the edge, not once per retry |
Health is asserted through fipsctl show status (state), presence through
fipsctl show transports (the per-transport interface block: presence,
policy, binds, since_secs).
Running
./test.sh # builds the image first
./test.sh --skip-build # reuse an existing image
./test.sh --keep-up # leave the containers running for inspection
Via the local CI runner:
./testing/ci-local.sh --only iface-binding
Notes
The containers run under FIPS_TEST_MODE=default, not chaos. The chaos
entrypoint waits up to 30 s for every configured Ethernet interface before
starting the daemon — which is exactly the workaround this mechanism retires.
The daemon has to do its own waiting here or the suite proves nothing.
Every ip link operation on the host network stack runs inside a short-lived
privileged container sharing the host network and PID namespaces, for the
reason chaos/sim/veth.py documents: on macOS the
containers live in the Docker VM, so ip(8) run on the macOS host could never
reach them, while on Linux the shared namespaces make it identical to running
ip(8) directly.