x..y is correctly NOT rejected by the slash/dot/empty/NUL guard; the recipe now documents this as an intentional regression test against over-rejection. Frames canonicalize-then-verify as Phase 17 follow-up hardening.
4.7 KiB
Testing pam_authforge_pending.so
This module gates first-login enrollment. Its full behavior is impossible to
exercise from a unit test because PAM modules run inside a real PAM stack that
needs root to install. The recipe below runs in a VM or container with
libpam0g-dev and pamtester installed.
Build + install
sudo apt-get install -y libpam0g-dev libargon2-dev pamtester
sudo make -C pam install
sudo install -D -m 0644 pam/test/authforge.pamd /etc/pam.d/authforge
libargon2-dev is required for the recovery-mode verification path (Phase 12).
make install drops the .so into
/usr/lib/$(dpkg-architecture -qDEB_HOST_MULTIARCH)/security/.
Smoke test 1: pending flag present → denied
sudo mkdir -p /var/lib/authforge/pending
sudo touch /var/lib/authforge/pending/$USER
pamtester -v authforge "$USER" authenticate
# Expect: "Account setup incomplete..." printed and a non-zero exit code.
Smoke test 2: pending flag cleared → falls through
sudo rm /var/lib/authforge/pending/$USER
pamtester -v authforge "$USER" authenticate
# Expect: success (the test PAM stack file falls through to pam_unix after
# our PAM_IGNORE; pamtester will prompt for the user's password).
Smoke test 3: path-traversal regression
The guard in safe_user rejects names that break the path-as-identifier
invariant: empty strings, ., .., anything containing / or NUL.
Anything else passes through — including unusual but legitimate names that
contain .. as a substring (LDAP / federated identities can produce
these). The loop below mixes both classes; the assertions live in the
comments:
for u in "../etc" "a/b" "" "." ".." "x..y"; do
pamtester -v authforge "$u" authenticate 2>&1 | head -3
echo "---"
done
# Expect rejected (logged "rejected unsafe username"): ../etc, a/b, "", ., ..
# Expect passed through (no rejection log; pam_unix runs and prompts): x..y
Keeping x..y in the loop is intentional — it documents the boundary and
acts as a regression test against accidental over-rejection if the guard
is ever tightened.
Future hardening (deferred — not blocking v1): the guard is a cheap
pre-filter, not the authoritative boundary. A belt-and-suspenders pattern
would be canonicalize-then-verify: build the per-user path, resolve it
via fs::canonicalize (Rust) / realpath (C), and confirm the resolved
path is still under the expected parent directory. That covers symlink
shenanigans and any future ..-substring attack surface uniformly.
Track as a Phase 17 item.
The module logs via pam_syslog without setting its own openlog ident,
so journal entries inherit the calling process's syslog identifier:
pamtester here, sudo / sshd / gdm / login in production. The
module name appears in the message body, so the always-works query is to
grep for the literal string:
journalctl --since "10 minutes ago" --no-pager | grep authforge_pending
For pamtester runs specifically, journalctl -t pamtester works too.
Older systems that route pam_syslog to /var/log/auth.log show the
same lines there.
Smoke test 4: recovery-code path (Phase 12)
The module's recovery branch activates when invoked with mode=recovery in
the PAM stack. The pam-auth-update profile rendered by Phase 4 puts this
line first in the auth stack — a successful recovery code short-circuits
the rest of the auth chain via [success=done default=ignore].
To exercise it manually with pamtester against a custom stack:
# 1. Issue a code via the daemon and stash the plaintext.
CODE=$(sudo authforgectl recovery generate "$USER" | tr -d '\n')
echo "issued $CODE"
# 2. Build a PAM stack file that runs ONLY the recovery-mode module.
sudo tee /etc/pam.d/authforge-recovery <<'EOF'
auth [success=done default=ignore] pam_authforge_pending.so mode=recovery
auth required pam_deny.so
EOF
# 3. Verify a matching code authenticates.
echo "$CODE" | pamtester -v authforge-recovery "$USER" authenticate
# Expect: PAM_SUCCESS, /var/lib/authforge/recovery/$USER deleted, and
# /var/lib/authforge/pending/$USER now contains re_enroll: true.
# 4. Verify a non-matching code falls through to pam_deny.
echo "$CODE" | pamtester -v authforge-recovery "$USER" authenticate
# Expect: failure (pam_deny) — the recovery file was already consumed.
# 5. Cleanup.
sudo rm /etc/pam.d/authforge-recovery /var/lib/authforge/pending/$USER
The recovery file is at /var/lib/authforge/recovery/<user> mode 0600.
The two-line format is <expires_unix>\n<argon2id-PHC>\n.
CI status
CI cannot run pamtester against a real PAM stack without root and a VM, so
this is the last manual gate before the Phase 14 PPA build. The CI pipeline
only verifies that make -C pam produces a valid ELF.