mirror of https://github.com/jmcorgan/fips.git synced 2026-09-14 00:45:08 +00:00

Files

T

Johnathan Corgan 6a564e26ac Prepare the v0.5.0 release content

Everything the release needs except the version number, which stays at
0.5.0-dev until the tag.

The changelog entry covers only the work that is new on this line. The
point release's forty-six entries arrived under their own heading with the
forward merge and are left alone; the twenty that remained are regrouped by
topic and eight more added for changes no entry covered. Three of those
eight matter to someone upgrading. Five root modules and four re-exports
left the public library surface and Node::connections narrowed, none of it
recorded anywhere; the entry names what to use instead and distinguishes
the removed connection-phase enum from the Noise type of the same name,
which is a different type that still exists. Tracing targets moved, so an
existing RUST_LOG filter stops matching rather than erroring. And the
handshake resend interval key no longer governs the first resend, which is
now a constant, though it still governs later ones.

Seven more entries cover the work that landed after the first content pass
was written: the experimental native datagram API, the fipsctl probe
diagnostic, per-instance transport addressing, the app-owned UDP socket
seam, and the connect, disconnect and path-MTU fixes. The four bug fixes
among them all reach the deployed line, so the release notes no longer
claim this release carries exactly one fix for a shipped bug; it carries
four.

There is no security section, because after the split every security entry
belongs to the point release. The release notes say so plainly rather than
leaving a reader upgrading across both releases to conclude this one
carries no security work.

The notes are organized by audience, since the release spans OpenWrt
routers, embedders, FreeBSD, and the existing platforms, and a single list
serves none of them. The native datagram API is given a section of its own
rather than folded into the embedding seam: it is a client-facing API
rather than a way to host a node, and its one rule with no Berkeley-socket
counterpart, that the v1 wire carries no half-close, needs to be somewhere
a client author will read it. FreeBSD is advertised as supported on x86_64
only, stated wherever the platform appears. Android is advertised as an
embedding seam and not as a supported platform: a compile-gated library
surface with no artifact and no host application guide.

The configuration table rename is carried through every shipped file that
taught the old spelling: nine documentation files, the OpenWrt sample
config and a test generator, twenty-two sites in all. Guides written this
same cycle were among them, which is how the omission was found. The
documentation that arrived with the native API was checked for the same
omission and was already clean. The compatibility tests keep the old
spelling deliberately, since they exist to test the fold.

The changelog section is the fold of master's [Unreleased], not a snapshot
of it. An earlier version of this commit took a copy that then drifted, so
each section ended up holding a bullet the other did not and re-folding
them would have picked a winner silently. Both causes were fixed on master
instead — the NixOS module had never been recorded there, and the
pre-release batch of fixes was new — so [Unreleased] is a strict superset
and this is a copy rather than a merge. [0.5.0] carries all forty-six
bullets byte for byte, [Unreleased] is empty, and [0.4.2] is untouched,
checked by hashing it against master's copy.

The BLE work landed after the content pass and gets one summary entry in
the changelog and one section in the release notes rather than nine
bullets: the ble_available gate replacing target_os = "linux",
packet-boundary recovery for stream-oriented backends, peer recognition by
node identity instead of a rotating link address, the L2CAP PSM moving
into the backend seam and onto the advertisement, the embedder-supplied
Android radio, bounded probe retry, and inbound handshakes moved off the
accept loop.

The two release-notes copies no longer share their link paths. Relative
links resolve from one directory only, so the seven written for
docs/releases/ all 404ed from the root copy. The root copy now uses paths
from the repository root and the versioned copy keeps the ../ form; both
sets were resolved against the tree. The same two links are broken the
same way in the v0.4.0 through v0.4.2 notes, left as shipped history.

The contributor tallies are re-derived against maint..HEAD rather than
adjusted: twenty commits from outside the project and 171 from me, with
Arjen at fifteen and fr34aky at two. An earlier count of twelve and 138
was carried from a measurement taken three days before this content was
written, and the BLE branch widened the gap after it. Arjen's NixOS flake
module, the UDP sin6_scope_id fix and most of the BLE rework were
uncredited, as was fr34aky's L2CAP PSM seam. They want one last re-derive
at tag time if anything lands before the tag.

A sweep of all 99 tracked markdown files against the tree corrected
fifty-three of them. Four told the reader to run a build.sh that does not
exist; the only harness builder is testing/scripts/build.sh. The BLE build
prerequisites were described as optional on the strength of a probe that
build.rs does not perform, and bluez was named a build prerequisite when
libdbus-sys asks only for libdbus-1-dev and pkg-config and bluez is the
runtime daemon. Link cost is the primary sort key in next-hop ranking, not
reserved for future use; Ethernet runs on macOS as well as Linux; the BLE
MTU is the L2CAP CoC MTU rather than a negotiated ATT_MTU; effective
Ethernet MTU is 1497; the LAN discovery subsystem is src/mdns and eight
citations still named a src/discovery that never existed here. The
connectivity states in three tutorials were invented, and their jq filters
matched nothing including healthy peers. One command filtered on a literal
fd97: address prefix, which only the first byte of fixes, so it returned
empty for all but one reader in 256 and every later step using the
variable failed silently. transports.tor.advertise_on_nostr was
undocumented despite being validated against node.rendezvous.nostr.enabled.

The transport design document gains the BLE section it never had, written
from the source: the backend cascade and its compile_error tripwire, the
platform gate, the PSM advertisement wire layout and the byte budget that
forces a 16-bit service-data key, and the probe and admission bounds.

Three source files carried the same class of staleness and are corrected
with the documentation: the OpenWrt ipk usage line and Makefile error text
both named a packaging/openwrt that does not exist, and chaos.sh parsed
--subnet without listing it.

Folded in with the content commit, having been prepared alongside it:

The three GitHub Action pins that had gone stale. Every third-party
action is pinned to a commit SHA, nothing reports that a pin has aged,
and re-resolving all ten against their tags found dorny/test-reporter@v2,
taiki-e/install-action@v2 and vmactions/freebsd-vm@v1 had moved. The
three install-action@nextest references stay unpinned, since that action
reads the tool to install from the ref name. check-action-pins.sh passes
at 75 references and all nine workflow files parse.

The lockfile refresh, which is the mutating half of the dependency sweep.
Thirty-six packages move to their latest semver-compatible versions and
every one is transitive; nothing declared in Cargo.toml changes version.
No advisory forces any of them. It was taken before the validation
battery, because a gate run against a lockfile that later moves proves
nothing about what ships.

The sha2 0.10 to 0.11, hkdf 0.12 to 0.13 and bech32 0.11 to 0.12 majors,
three of the four deferred at v0.4.0 for change surface rather than
security. All three land with no source change. sha2 and hkdf must move
together, since both depend on digest 0.11, and neither changes an
algorithm. That matters because the chaining-key KDF in the Noise
handshake is built on Hkdf::<Sha256>, where an output change would be a
wire break rather than a compile error; no known-answer vectors exist for
that path, so the wire-compatibility gate is what covers it. secp256k1
0.31 is deliberately absent, since nostr's own requirement would leave
two copies of the ECC library in the tree.

The README support matrix, rebuilt as one feature table broken out by
Linux variety. A single Linux column hid that Debian, Ubuntu, Arch and
NixOS are one glibc build differing in packaging, that OpenWrt is musl
and drops BLE, and that Android is not a daemon platform. Transport rows
sort by how many platforms carry them. A Native API row reads its
platform set from the cfg gates. The installer row becomes a package
format row naming the artifact, and only the .deb is exercised per
release.

Four changelog and release-note gaps the BLE re-walk found: a Bluetooth
LE bullet stranded inside the released 0.4.2 section, a missing Fixed
entry for the scan and probe loop counting a pool-refused connection as
an established link, the unnamed embedder call that installs an
application-owned radio, and the fact that stopping the transport now
stops scanning as well as advertising.

Three release-document gaps found walking the unsurveyed commits: the UDP
reuse-flag fix stated in the direction opposite to the one it was made,
with the silent second-daemon bind it prevents left unsaid; the corrected
native-API socket paragraph carried into both release-note copies, which
still named SOCK_SEQPACKET on FreeBSD and two kernels where three are
handled; and the coordinate-cache hardening, which shipped with no text
anywhere despite adding four operator-visible status fields. That last
entry states plainly that the checks are mitigations and not a closure,
since the coordinate is still not authenticated.

Also folded in, the documentation pass that followed the content commit:

A stage-pipeline diagram for the probe, embedded in the fipsctl
reference under the five-stage list. It draws the five stages left to
right with each stage's failure reasons below it, and the bypass that
skips both lookup stages when the coordinates are cached or the target
is a direct peer. Its branches come from the probe state machine rather
than from the report, so the path stage is drawn as the one failure that
does not stop the probe.

A rewrite of the README's "What FIPS does" section. It now opens with
what a machine running FIPS gets, rather than with the two deployment
modes, and gives the self-organizing and permissionless property its own
paragraph since it holds for both modes.

A regrouping of the README's feature list into the mesh, getting traffic
onto it, and running a node, with a bullet added for the native datagram
API, which had none despite sitting in the support matrix. The Quick
start now leads with the released packages rather than a source build.
It also fixes a real defect: the package enables fips.service and
fips-dns.service and starts neither on a fresh install, so .fips name
resolution was silently dead until the next reboot and neither page said
to start the service.

A rewrite of the release notes. They opened with seven subsections of
upgrade caveats and reached the first feature two hundred lines in; they
now open with a summary of the release and elaborate below it in the
same order. Android is stated as supported through an embedded crate
rather than as a standalone daemon, consistently across all three
documents. The OpenWrt pair is corrected: it is 802.11s between routers
with FIPS supplying encryption, authentication and routing, plus a
convention of an open !FIPS SSID a client joins over WiFi, not meshing
over a router's own radios. The probe's path output is described as the
least-common-ancestor walk, which is the worst-case fallback route
rather than the route a packet takes. Detail that did not change what a
reader does was cut from the notes and kept in the changelog.

2026-08-30 10:42:59 +00:00

16 KiB

Raw Blame History

Deploy `fips-gateway` (Manual Linux-Host Setup)

fips-gateway is a separate service that runs alongside the FIPS daemon and bridges a non-FIPS LAN to the FIPS mesh in two independent directions: outbound (LAN clients reach mesh services through DNS proxy + virtual-IP NAT) and inbound (mesh peers reach LAN services through 1:1 port forwards on fips0). This guide covers the manual Linux-host deployment path — wiring DNS forwarding, route distribution, and firewall integration on a server or non-OpenWrt router by hand.

Running OpenWrt? Use the tutorial instead. The OpenWrt ipk ships with the gateway: block pre-populated and the init script automates dnsmasq forwarding, RA route distribution, and the global IPv6 prefix on br-lan. The OpenWrt path is the canonical deployment of this feature; this how-to is the secondary path for operators with a different LAN-edge box (a Linux server already serving DHCP/DNS, a custom router distribution, etc.).

For the gateway design (NAT pipeline, virtual IP pool lifecycle, DNS resolution flow), see ../design/fips-gateway.md. For the full gateway.* configuration block, see the Gateway section of the configuration reference. For the fips-gateway binary's CLI flags, see ../reference/cli-fips-gateway.md.

The two halves

The gateway exposes two independent features that share a common control plane (the same binary, the same nftables table inet fips_gateway, the same control socket /run/fips/gateway.sock, the same gateway.* config block). You can configure either half on its own or both together.

Outbound gateway (LAN → mesh). Non-FIPS LAN workstations resolve <npub>.fips names against the gateway's DNS listener and receive AAAA answers from the gateway's virtual-IP pool. Outbound traffic to those addresses is DNAT'd to the real mesh address and SNAT'd (masqueraded) onto fips0 under the gateway's mesh identity. The audience is unmodified LAN clients.
Inbound gateway (mesh → LAN). A static (listen_port, proto) → [target_addr]:target_port table — configured in gateway.port_forwards[] — exposes selected LAN services to the mesh as <gateway-npub>.fips:<listen_port>. Mesh peers connect to the gateway's mesh address; the gateway DNATs to the LAN target and masquerades on the LAN side so return traffic flows through conntrack. The audience is mesh peers reaching a service that happens to live on this LAN.

The two halves are independent. Configure the outbound half if you want LAN clients to reach the mesh; configure the inbound half if you want mesh peers to reach into the LAN; configure both if you want both.

Common gateway-host setup

Both halves require the same host preparation. Work through this section first, then jump to whichever half (or both) you need.

FIPS daemon prerequisites

The gateway runs alongside a fips daemon on the same host:

The daemon must be running with the TUN adapter enabled (the fips0 interface must exist).
The daemon's DNS resolver must be enabled (dns.enabled: true, default) and reachable from fips-gateway. By default that means [::1]:5354 (IPv6 loopback). The gateway's default dns.upstream matches this; a v4 upstream like 127.0.0.1:5354 cannot reach a daemon bound on [::1]:5354 because Linux IPv6 sockets bound to explicit ::1 do not accept v4-mapped traffic.

If the daemon is not yet running with these features, set up the daemon first — see persistent-identity.md and ../reference/configuration.md.

Kernel sysctls

sudo sysctl -w net.ipv6.conf.all.forwarding=1
sudo sysctl -w net.ipv6.conf.all.proxy_ndp=1

forwarding lets the host route IPv6 packets between the LAN interface and fips0. proxy_ndp lets the gateway answer Neighbor Solicitation requests for virtual-pool addresses so LAN clients can resolve their link-layer addresses (only relevant for the outbound half, but harmless if you only run the inbound half).

Persist via a drop-in:

sudo tee /etc/sysctl.d/60-fips-gateway.conf <<'EOF'
net.ipv6.conf.all.forwarding = 1
net.ipv6.conf.all.proxy_ndp = 1
EOF
sudo sysctl --system

Capability

fips-gateway requires CAP_NET_ADMIN to manage its nftables table (inet fips_gateway) and proxy-NDP entries. The packaged systemd unit (fips-gateway.service) runs as root, which satisfies this. For non-package installs, set the file capability:

sudo setcap cap_net_admin+ep /usr/bin/fips-gateway

Pool route

At startup fips-gateway adds local <pool-cidr> dev lo to the local routing table. This tells the kernel to accept packets destined for pool addresses as locally-owned, enabling the NAT processing path. The route is cleaned up on shutdown. You do not need to install it manually; if you see "destination unreachable" errors for pool addresses on the gateway host, verify the route is present:

ip -6 route show table local | grep <pool-cidr>

Minimum configuration

In /etc/fips/fips.yaml, populate the gateway block with at minimum enabled: true, pool, and lan_interface:

gateway:
  enabled: true
  pool: "fd01::/112"
  lan_interface: "enp3s0"

Pick a pool CIDR that does not overlap with any address space in use on the LAN or in the mesh (the FIPS mesh occupies fd00::/8; pick a different fdXX::/N). The /112 size yields 65 535 usable virtual IPs, which is the gateway's hard cap regardless of CIDR width.

This minimum config is enough to start the gateway. The dns.* block is optional and defaults to listen: "[::1]:5353" and upstream: "[::1]:5354". The full block — including dns.*, pool_grace_period, conntrack.*, and port_forwards[] — is documented in ../reference/configuration.md#gateway-gateway.

Start the service

sudo systemctl enable --now fips-gateway

Verify the unit came up:

sudo systemctl status fips-gateway
sudo journalctl -u fips-gateway -e

The startup log will report Gateway config loaded, DNS upstream is reachable, Created nftables table 'fips_gateway', and finally fips-gateway running. The unit's ExecStartPre waits up to 30 s for fips0 to appear, which covers the cold-boot race where the daemon is still bringing up its TUN.

Configure the outbound half

The outbound half lets LAN clients resolve .fips names and reach mesh destinations. Three operator decisions are involved: pool CIDR, DNS listen address, and how LAN clients learn the route to the pool and the resolver address.

Choose the pool CIDR

gateway:
  pool: "fd01::/112"

Constraints:

Must not overlap with fd00::/8 (the FIPS mesh address space).
Must not overlap with any LAN-side IPv6 prefix already in use.
/112 is the practical width — wider just wastes address space because the pool is hard-capped at 65 535 usable entries. Narrower is fine if you want a smaller pool, but you'll reject DNS lookups faster under churn.

Choose the DNS listen address

gateway:
  dns:
    listen: "[::1]:5353"
    upstream: "[::1]:5354"
    ttl: 60

Common cases:

Another resolver on the host (the canonical case): the default listen: "[::1]:5353" is loopback-only on an unprivileged port, so it never conflicts with dnsmasq, systemd-resolved, or BIND holding 53. Configure the existing resolver to forward .fips queries to [::1]:5353 and you are done — this is what the OpenWrt ipk does automatically.
No other resolver on the host: set listen: "[::]:53" explicitly and LAN clients can query the gateway directly.
systemd-resolved is on port 53: the default already side-steps this — leave the listen address at [::1]:5353 and configure the stub or a small forwarder to delegate .fips to the gateway. If you would rather have the gateway on 53 directly, disable the systemd stub listener (DNSStubListener=no in /etc/systemd/resolved.conf) and switch listen to "[::]:53". See troubleshoot-gateway.md.
Bind on the LAN address only: listen: "192.168.1.1:53" exposes the resolver only to LAN clients, not loopback.

The gateway returns REFUSED for any non-.fips query — clients that point at it directly need a fallback resolver, or you should front it with a stub forwarder.

Distribute the route to LAN clients

Each LAN client must route the gateway's pool CIDR to the gateway's LAN-side IPv6 address. Three options, in order of preference for production:

RA Route Information Option (RFC 4191). If the LAN's RA daemon (radvd, dnsmasq --enable-ra, OpenWrt's odhcpd) supports publishing route options, configure it to advertise the pool CIDR with the gateway as next-hop. Clients pick this up automatically.
Static route on the LAN router. If clients route through a central LAN router, add a static route entry there — the router then handles forwarding to the gateway. The exact syntax depends on the router OS.

Per-host static route (testing or single-client deployments):

sudo ip -6 route add fd01::/112 via fe80::<gateway-link-local>%<iface>
# or, if the gateway has a stable global LAN address:
sudo ip -6 route add fd01::/112 via <gateway-lan-addr>

Distribute the resolver to LAN clients

LAN clients also need to send .fips queries to the gateway. Two patterns:

Forward .fips from the LAN's main resolver. If the LAN runs Pi-hole, Unbound, dnsmasq, or systemd-resolved as the central resolver, configure a conditional forward for fips.. Unbound example:
```
forward-zone:
    name: "fips."
    forward-addr: <gateway-lan-addr>@53
```
dnsmasq example:
```
server=/fips/<gateway-lan-addr>
```
Clients keep their existing DNS settings; only .fips queries are diverted.
Point clients directly at the gateway. Simpler for testing, but the gateway returns REFUSED for non-.fips queries, so each client must also have a fallback resolver configured.

Verify the outbound path

From a LAN client:

dig @<gateway-lan-addr> hostname.fips AAAA
# Expect an AAAA from the pool CIDR

ping6 hostname.fips
# Should succeed via the gateway

If either step fails, see troubleshoot-gateway.md.

Configure the inbound half

The inbound half exposes a LAN-side service to mesh peers. Configured under gateway.port_forwards[]:

gateway:
  port_forwards:
    - listen_port: 8080
      proto: tcp
      target: "[fd12:3456::10]:80"
    - listen_port: 2222
      proto: tcp
      target: "[fd12:3456::20]:22"
    - listen_port: 5353
      proto: udp
      target: "[fd12:3456::10]:53"

Field reference:

listen_port — port on the gateway's fips0 mesh-side address that mesh peers connect to. Must be non-zero. Each (listen_port, proto) pair must be unique across the list (the same port on TCP and UDP is allowed; the same port twice on the same proto is rejected at config-load time).
proto — tcp or udp.
target — IPv6 LAN destination as [addr]:port. IPv4 targets are rejected at parse time by the YAML deserializer (the field is typed SocketAddrV6). If the LAN host is reachable only by IPv4, put a small IPv6-aware reverse proxy in front of it on the gateway itself.

Worked example: HTTP and DNS

Suppose the gateway runs on a LAN with an HTTP server at [fd12:3456::10]:80 and a recursive resolver at [fd12:3456::10]:53, and you want mesh peers to reach them as <gateway-npub>.fips:8080 (HTTP) and <gateway-npub>.fips:5353 (DNS). Add to the gateway's fips.yaml:

gateway:
  port_forwards:
    - listen_port: 8080
      proto: tcp
      target: "[fd12:3456::10]:80"
    - listen_port: 5353
      proto: udp
      target: "[fd12:3456::10]:53"

Reload:

sudo systemctl restart fips-gateway

From any mesh peer (the host name gateway is whatever the gateway's npub maps to in the local hosts file or via Nostr advert):

curl http://gateway.fips:8080/
dig @gateway.fips -p 5353 example.com A

Each mesh-side request enters fips0 on the listen port, gets DNAT'd to the LAN target, and the LAN-side masquerade rule rewrites the source to the gateway's LAN address so return traffic flows back through conntrack.

Compose with the mesh firewall

gateway.port_forwards[] opens mesh-side listeners on fips0. If the host's mesh firewall is enabled (see enable-mesh-firewall.md), inbound TCP/UDP on fips0 for these ports must be permitted in the baseline or via a drop-in. The default baseline allows established/related and ICMPv6 only, so without an explicit allow rule, mesh peers will see TCP RSTs or silent drops on the listen port.

A typical drop-in for the worked example:

# /etc/fips/fips.d/gateway-inbound.nft
tcp dport 8080 accept
udp dport 5353 accept

Reload the firewall:

sudo systemctl reload-or-restart fips-firewall.service

If the inbound half doesn't need access control beyond the listen port itself, no source filter is needed. To restrict to specific mesh peers, follow the ip6 saddr <addr> tcp dport <port> accept pattern from the firewall guide.

Verify the inbound path

From a mesh peer (any FIPS node):

curl -v http://<gateway-npub>.fips:8080/

A successful response confirms the full path: mesh ingress on fips0, DNAT to the LAN target, LAN-side masquerade, and conntrack- tracked return. If it fails, see troubleshoot-gateway.md.

Operate and verify

fips-gateway exposes its own control socket at /run/fips/gateway.sock, separate from the daemon's /run/fips/control.sock. There is no fipsctl gateway subcommand — talk to it directly:

echo '{"command":"show_gateway"}' | sudo nc -U /run/fips/gateway.sock
echo '{"command":"show_mappings"}' | sudo nc -U /run/fips/gateway.sock

show_gateway returns pool counters (pool_total, pool_allocated, pool_active, pool_draining, pool_free), nat_mappings, dns_listen, uptime_secs, and the active config snapshot. show_mappings returns the per-allocation list with virtual IP, mesh address, npub-derived node_addr, dns name, state (Allocated, Active, Draining), session count, and ages. For the full schema see ../reference/control-socket.md#gateway-command-catalog.

The journal is the other primary signal:

sudo systemctl status fips-gateway
sudo journalctl -u fips-gateway -e

Expect MappingCreated/MappingRemoved debug lines as DNS-driven allocations come and go (run with --log-level debug to see them), and Final pool status on shutdown. Errors in adding NAT rules or proxy-NDP entries surface here.

16 KiB Raw Blame History

Deploy fips-gateway (Manual Linux-Host Setup)