mirror of
https://github.com/jmcorgan/fips.git
synced 2026-10-06 11:38:24 +00:00
Johnathan Corgan
5a79d3fc5e
Brings up the two release lines' work: the maint harness and guard fixes, the documentation corrections, master's probe, onion and epoch fixes, and both rebuilt changelog blocks. Three conflicts. The readme conflicted on the badge pair. Resolved by taking the Rust badge that no longer asserts a version, since rust-toolchain.toml is the only place that states one, and keeping this line's own v0.6.0-dev status badge. The peer machine conflicted, and the resolution is an adaptation rather than a pick. Master deleted PeerMachine.remote_epoch on the grounds that nothing read it, and that reasoning had to be re-derived here because this line's machine is the XX rewrite and shares almost no text with it. It holds. The shadow's only production write is inbound_msg3, which is where XX crystallizes identity, so it is inbound-only exactly as the msg1 write was on the other lines; conn carries the same value written from both legs, complete_handshake on the outbound one and complete_handshake_msg3 on the inbound; the only read is the cutover action payload, whose executor arm binds nothing; and the live consumer reads conn_remote_epoch. So an initiator cutover, which runs on an outbound machine, carried a zeroed epoch here too. One thing differs and needed handling. This line has an `established` constructor the others do not, and it writes the shadow and conn from the same argument, which would have made the field direction-correct. Its only caller is in the test module and its own doc comment calls the machine inert, so it is a seam that is not wired yet rather than a production path, and it does not rescue the field. Its assignment goes with the rest; the parameter stays, because conn still needs the value. The adaptation is folded into this merge rather than left to a follow-on, because master's half of the same change reached peer_actions.rs through a clean auto-merge. Keeping this line's field while accepting that auto-merge would have left the machine emitting a payload field the executor no longer has, which is a break that only the test build shows. The changelog conflicted because both lines had rebuilt their unreleased block. The Breaking section stays at the top untouched; Unreleased now holds the ten entries that are this line's own; and the other two lines' work sits below under 0.5.0 and 0.4.2 headings, neither dated, matching how master already carries 0.4.2. Four entries existed on both sides in branch-adapted form and were merged rather than picked, so each keeps the rework's wording and this line's accuracy: the OpenWrt entry drops its IK reference, the msg1 classifier keeps the promotion-state paragraph, the SessionAck entry keeps the two XX-only exits, and the msg3 epoch entry counts six sites here against master's five.
NAT Lab Harness
Real Docker-based NAT traversal integration tests for the mainline FIPS Nostr/STUN bootstrap path.
This harness spins up:
- two FIPS nodes
- a local Nostr relay
- a local STUN server
- one or two Linux router containers performing NAT with
iptables
For the NAT scenarios, the node LAN interfaces are not attached to
Docker bridge networks. The harness creates explicit veth pairs and
moves them into the node and router namespaces after docker compose up
so every packet must traverse the router namespace.
It covers three scenarios:
cone: both peers behind explicit namespace/veth full-cone emulation, UDP traversal succeedssymmetric: both peers behind symmetric-style NAT, UDP traversal fails, TCP fallback succeedslan: both peers share a LAN subnet, LAN targets are preferred over reflexive addresses
NAT model notes
The harness does not rely on plain Docker MASQUERADE for the cone case.
cone- uses explicit full-cone emulation in the router namespace
- outbound UDP is
SNATed to the router WAN address while preserving the source port - inbound UDP to the router WAN address is
DNATed back to the single LAN host regardless of remote source
symmetric- uses UDP
MASQUERADE --random-fully - outbound mappings may be port-randomized and are only reopened by matching conntrack state
- uses UDP
This distinction matters because plain MASQUERADE is convenient source NAT, but it does not by itself model the "accept from any remote once mapped" behavior expected from a full-cone NAT.
Prerequisites
- Docker with Compose support
- locally built
fips-test:latest
Build the test image with:
./testing/scripts/build.sh
Run
Run all scenarios:
./testing/nat/scripts/nat-test.sh
Run one scenario:
./testing/nat/scripts/nat-test.sh cone
./testing/nat/scripts/nat-test.sh symmetric
./testing/nat/scripts/nat-test.sh lan
Layout
docker-compose.yml- relay/STUN/WAN topology plus container definitions
node/- node bootstrap wrapper that waits for the injected veth interface
router/- NAT router image and
iptablessetup
- NAT router image and
stun/- minimal STUN binding responder
relay/- local
strfryconfig
- local
scripts/generate-configs.sh- derives ephemeral identities and writes per-scenario FIPS configs
scripts/setup-topology.sh- injects and configures the NAT LAN
vethpairs in the container namespaces
- injects and configures the NAT LAN
scripts/nat-test.sh- boots the lab, waits for convergence, and asserts the resulting path
Assertions
-
cone- both nodes connect
- connected transport is UDP
- active link remote addresses are on the WAN NAT subnet
-
symmetric- NAT bootstrap does not establish a UDP link
- fallback converges
- connected transport is TCP via router-published WAN addresses
-
lan- both nodes connect
- connected transport is UDP
- active link remote addresses stay on the shared LAN subnet