Johnathan Corgan
1077fd6a7d
wait_for_peers_exact read the connected-peer count through a pipeline ending in `|| echo 0`, so a container that never answered and a daemon that answered zero produced the same value. That is only harmless while every caller expects a non-zero count, which is true here today and is the reason this copy was left alone when the acl suite's copy was fixed. It leaves the trap armed for whoever adds the first caller expecting zero: the check would be satisfied on its first iteration without the property it exists to verify ever being observed. The read moves into its own function that returns the empty string when the container does not answer, and the caller treats empty as "no answer" rather than as a count. A run that never gets an answer now fails saying so, distinctly from one that answered the wrong count, and the diagnostic peer dump runs only on the latter, since dumping from a container that cannot answer prints a docker error rather than evidence. This is the shape the acl-allowlist suite already uses. The two copies had diverged on whether a silent container counts as an answer, which is the kind of disagreement that decides a security assertion in whichever file is read last. Checked by construction rather than by a green run: driving the old form against an absent container with an expected count of zero returns success on the first iteration, and the new form fails; with an expected count of one, the shape both real callers use, both forms still fail, and a genuine zero from a live daemon is still reported as zero.
Firewall Baseline Test
End-to-end exercise of the production fips0 nftables baseline at
packaging/common/fips.nft. Closes the v0.3.0 audit gap that the
default-deny + conntrack + drop-in semantics had no integration coverage.
What this exercises
The fips.nft baseline polices ONLY the fips0 mesh interface and
implements default-deny inbound. This suite asserts the four behaviors
documented in the file's header are actually true on a live mesh:
- (a) Unallowed inbound on fips0 is dropped
- (b) Outbound-initiated flows get their reply via the
ct state established,related acceptrule - (c) ICMPv6 echo-request is accepted (ping6 reachability)
- (d) A drop-in
.nftfile under/etc/fips/fips.d/adds an allowlisted port and that port is accepted
A drop-counter check after case (a) confirms the connection was actively DROP'd by the fips chain (not silently unrouted).
Topology
Two FIPS nodes peered over UDP on a Docker bridge network:
| Container | Hostname | Firewall |
|---|---|---|
fips-fw-container-a |
host-a |
none (probe) |
fips-fw-container-b |
host-b |
fips.nft + drop-in |
The bridge network requests no subnet, so docker assigns one from its own address pool and two concurrent runs never contend for a fixed range. No node's IPv4 address is therefore known before startup, and the generated peer stanzas address each other by docker hostname, resolved through the container's dnsmasq to docker's embedded DNS. The firewall assertions themselves are unaffected: they run over the fips0 overlay, whose addresses are derived from the node npubs.
node-b mounts the production packaging/common/fips.nft read-only at
/etc/fips/fips.nft, plus a drop-in at /etc/fips/fips.d/services.nft
containing tcp dport 22 accept. node-a is unfirewalled and serves
as the probe origin.
Both containers run the unified test image's default mode, which
starts dnsmasq + sshd (port 22) + iperf3 + python http.server on
port 8000 + the FIPS daemon.
fips-firewall.service activation
The production unit's ExecStart is:
ExecStart=/usr/sbin/nft -f /etc/fips/fips.nft
The unified test image does not run systemd, so test.sh invokes the
same nft -f command directly inside node-b after fips0 is up and
peering has converged. The deb-install harness covers the systemd
unit-enablement path under real systemd separately.
Run
Build the Linux binaries and test image:
./testing/scripts/build.sh --no-docker
Run the suite:
./testing/firewall/test.sh
test.sh regenerates fixtures automatically before starting Docker.
Use --skip-build to reuse the existing release binaries. Use
--keep-up to leave the containers running for inspection.
Expected output shape
=== Generating firewall fixtures
=== Starting firewall harness
=== Waiting for fips0 on both nodes
=== Waiting for peer convergence
=== Resolving fips0 addresses
node-a: fd97:...
node-b: fd97:...
=== Activating fips-firewall on fips-fw-container-b
PASS: fips-fw-container-b: fips.nft baseline + drop-in loaded
=== Case (c): ICMPv6 echo-request to firewalled node
PASS: (c) ICMPv6 ping node-a → node-b accepted
=== Case (a): unallowed inbound TCP/8000 from node-a → node-b
PASS: (a) inbound TCP/8000 dropped (curl rc=28, timed out as expected)
=== Case (b): node-b initiates outbound TCP, expects reply via conntrack
PASS: (b) outbound from node-b got HTTP 200 via conntrack reply path
=== Case (d): drop-in allowlisted TCP/22 from node-a → node-b
PASS: (d) drop-in allowlisted TCP/22 reachable
=== Drop counter incremented (case a should have ticked it)
PASS: drop counter = N (case a was actually dropped, not just unrouted)
=== Firewall integration test passed
Inspect the loaded ruleset
docker exec fips-fw-container-b nft list table inet fips
Stop and clean up
docker compose -f testing/firewall/docker-compose.yml down
Generated fixture location
testing/firewall/generated-configs/ (gitignored), or
generated-configs<suffix>/ when FIPS_CI_NAME_SUFFIX is set, which is how
concurrent runs keep their fixtures apart.