fr34akyandJohnathan Corgan
429d77731b
pfSense is FreeBSD underneath, but the FreeBSD package does not work there, failing in three silent ways. pfSense runs only /usr/local/etc/rc.d/*.sh at boot and re-runs them when WAN gets a new address, so a suffixless rc script never starts; unbound.conf is generated from config.xml with no conf.d, so a drop-in is never read; and on a firewall where the default-on "Allow IPv6" has been turned off, unbound is then generated with do-ip6: no and a responder on ::1 is unreachable. So this ships fips.sh, wires the fips. zone into the DNS Resolver through config.xml, and binds the responder on 127.0.0.1 for robustness against that last case. The rc script is plain sh: what pfSense imposes is the .sh name and that a re-run leave a running daemon alone and exit 0. It identifies the daemon by process name and recovers an orphaned daemon(8) supervisor found via fstat, since a locked empty pidfile makes daemon(8) report pid -1. The DNS setup is a manual step, never run from post-install, and validates the merged options with unbound-checkconf (pfSense's test_unbound_config) before touching config.xml, so a bad merge cannot take DNS from every client behind the firewall. The daemon runs under daemon(8) -H so newsyslog can rotate its log by signalling a reopen. Packages link statically by default: pfSense runs a FreeBSD base that cannot be obtained to link against. A firmware upgrade keeps the package (pfSense-upgrade removes only pfSense-pkg-*; confirmed on a live Plus 26.03.1 -> 26.07 upgrade, aarch64 — the package survived and the daemon restarted at boot. That is a minor, FreeBSD 16 -> 16 change; the cross-major compat case is still only source-reasoned). aarch64 is refused, where a static binary faults at posix_spawn. The mechanics the two builders share — version derivation, the stage layout, the manifest fields, the @sample scripts and pkg create — live in packaging/common/pkg-lib.sh, which both source; the FreeBSD package is byte-identical before and after that extraction. One ABI can serve more than one product: CE 2.9 and Plus 26.x on Intel are both FreeBSD:16:amd64 with a byte-identical artifact, named ...-ce2.9-plus26-amd64.pkg. The pfSense package is built and checked in its own CI job — separate from the FreeBSD package, and not a dependency of the release job, so a pfSense-only failure reds that job alone and is never a release asset. It is kept as a workflow artifact until it has been installed on a real pfSense box. CI produces the CE 2.8.1 (FreeBSD:15:amd64) package; CE 2.9, Plus 26.x Intel and ARM need a FreeBSD 16 build host the CI does not have, and ARM stays build-it-yourself because rustup ships no toolchain for it. testing/check-pfsense-pkg.sh validates a built package on any FreeBSD host and runs in that CI job: contents, modes, a positive boot-script lifecycle against a stub daemon, php -l and a fips_strip_block unit test of the config.xml helper. Installing on a real pfSense box, and the firmware-upgrade behaviour, are covered only by an aarch64 hardware run and pfSense-upgrade's source; the README records what is and is not tested. Co-authored-by: Johnathan Corgan <johnathan@corganlabs.com>
FIPS FreeBSD packaging
Builds a native FreeBSD .pkg shipping fips, fipsctl, fipstop,
rc.d services, and .fips DNS integration. fips-gateway is excluded
(its NAT backend is nftables, Linux-only).
On pfSense? Use packaging/pfsense/
instead. pfSense is FreeBSD underneath, but it boots packages, wires up
DNS and handles upgrades differently enough that this package does not
work there; the pfSense one addresses each difference.
Platform notes: the Ethernet and BLE transports are not available on
FreeBSD (UDP, TCP, Tor, and Nym are). The UDP datapath deliberately
uses the portable single-packet receive loop — FreeBSD's recvmmsg(2)
is a libc loop over recvmsg, not a kernel batch, so the Linux/macOS
batched arm would gain nothing — and the connected-UDP fast path is
not compiled pending FreeBSD-specific SO_REUSEPORT validation.
Build
./packaging/freebsd/build-pkg.sh # cargo build --release + pkg create
./packaging/freebsd/build-pkg.sh --no-build # package existing release binaries
Output: deploy/fips-<version>-freebsd-<arch>.pkg. pkg versions cannot
contain - or +, so both are mapped to .: a Cargo version of
<x.y.z>-dev becomes fips-<x.y.z>.dev-freebsd-amd64.pkg.
Install
pkg add ./deploy/fips-<version>-freebsd-<arch>.pkg
# post-install seeds this from the sample if absent, at mode 0600
vi /usr/local/etc/fips/fips.yaml
sysrc fips_enable=YES fips_dns_enable=YES
service fips start
service fips_dns start
fipsctl show status
Config installs sample-style (@sample semantics via manifest
scripts), so an edited fips.yaml survives upgrade/removal.
fips.yaml is installed 0600 — it may hold the node's private key
(nsec:). The daemon runs under daemon(8) with pidfile
/var/run/fips/fips.pid and logs to /var/log/fips.log (rc.conf
knobs: fips_config, fips_flags, fips_logfile).
The package creates a fips group; members can run fipsctl and
fipstop without root (pw groupmod fips -m <user>, then re-login).
On pkg upgrade the services are stopped before the binaries are
replaced and started again afterwards if enabled; on pkg delete they
are stopped and the .fips resolver drop-in is removed.
.fips DNS integration
The daemon answers .fips queries on [::1]:5354. fips_dns points
the system resolver's fips. zone there; backends tried in order:
base local_unbound (drop-in /var/unbound/conf.d/fips.conf), pkg
unbound, pkg dnsmasq. The unbound drop-in must (and does) set:
do-not-query-localhost: no— unbound's default silently refuses loopback forwarders, SERVFAILing every.fipsquery.do-ip6: yes— the daemon binds::1only.domain-insecure: "fips."— the zone is unsigned.
One-time host resolver setup (NOT automated)
The package configures the fips. zone only; making the local
resolver the system resolver is an operator decision. On a typical
box:
sysrc local_unbound_enable=YES
local-unbound-setup 1.1.1.1 9.9.9.9 # explicit upstreams — see below
service local_unbound restart
Field-tested caveats (fips-dns-setup detects and warns about each):
/etc/resolv.confmust list a loopback nameserver (ideally only127.0.0.1), or nothing ever queries unbound and.fipscannot resolve.- Do not use a home-router DNS proxy as unbound's upstream. Many CPE forwarders are EDNS-broken; unbound always sends EDNS, so every public query SERVFAILs. Forward to real resolvers (ISP or public).
- Remove
options edns0from/etc/resolv.confif present, and setresolv_conf_options=""in/etc/resolvconf.confso resolvconf(8) does not re-add it — against an EDNS-broken router it breaks libc resolution outright. - Never run
local-unbound-setupwith no arguments while resolv.conf already points at 127.0.0.1 — it snapshots that as upstream and forwards unbound to itself.
Debugging
drill -p 5354 <npub>.fips @::1 AAAA # daemon directly (bypasses unbound)
drill <npub>.fips AAAA # full chain; SERVER: must be 127.0.0.1
cat /var/run/fips/dns-backend # which backend fips_dns configured
The TUN interface gets a kernel-assigned name (tun0, tun1, ...),
like utun on macOS, and is destroyed automatically when the daemon
exits. ifconfig <name> prints Opened by PID <n> for the process
holding a tun device.