guard/src/app/shutdown.rs
loki5512344 c1a076a4a3
fix(core): close auth bypass, hangs and memory leaks found by review
Security:
- An empty API_PASSWORD passed validation (only the literal "changeme" was
  rejected) and verify_password("", "") returned true, so a login with an
  empty password minted an admin token. Require >= 16 characters and >= 32
  non-whitespace bytes for JWT_SECRET; verify rejects an empty stored secret.
- syn_challenge_enabled required a real challenge_secret_hex: the public
  placeholder from config.h is now rejected in every spelling.
- docker-compose: Redis/ClickHouse/Grafana are loopback-only and every
  password uses ${VAR:?} instead of a default. An exposed Redis was a way to
  lift bans and publish unbans to the whole fleet with one redis-cli.
- Empty CORS_ORIGIN no longer means "allow any origin".
- ban_cidr / userspace blacklist validate and mask prefixes; CIDR bans are
  stored as ranges, so a /24 or /64 ban actually covers the subnet (before,
  only the network address was blocked).

Availability:
- ClickHouse client has connect/request timeouts and no mutex held across
  await: a hung analytics backend can no longer freeze the ban path. Events
  go through a bounded buffer with a drop counter.
- Redis uses a ConnectionManager with per-command timeouts; pubsub and
  reconcile reset the backoff after success; publish cannot block shutdown.
- Ban durations from pubsub are clamped and use checked_add; the sync task
  survives panics instead of silently dying.
- Blacklist TOCTOU (remove_if) and replace_redis no longer empties the table
  during a Redis reconnect.
- RateLimiter and reputation have hard caps with incremental eviction instead
  of an unbounded DashMap and a stop-the-world retain.
- The attack detector scores before updating its baseline and needs several
  quiet windows to leave UnderAttack, so a sustained flood keeps banning new
  offenders instead of ending after one tick.
- accept() backs off on EMFILE/ENFILE, shutdown Err is treated as shutdown,
  a global semaphore bounds in-flight connections, rate limiting runs before
  PoW, whitelisted IPs are never banned, and connect/handshake/relay have
  timeouts with an idle budget.
- PoW uses the configured difficulty as a floor, the clamp is enforced, and
  nonce framing survives split segments and coalesced handshake bytes.
- Subnet tracker memory is capped (prefixes and per-prefix sources) and the
  monitor bans with the configured TTL.

CI and docs:
- CI runs the test suite with --no-default-features --features protocol-http;
  make ebpf fails on errors; the secrets gate catches weak compose defaults.
- doctor skips protected endpoints without a token and exits non-zero on
  failure; a drain timeout returns exit code 1.
- TODO §2b records the review findings, what is fixed and what is deferred;
  configuration/testing docs describe the new XDP and secret semantics.
2026-11-16 12:06:23 +01:00

58 lines
1.9 KiB
Rust

//! Edge daemon shutdown drain: signal handling, drain timer and exit code.
use std::time::Duration;
use tokio::sync::oneshot;
use tokio::sync::watch;
/// Daemon exit outcome. The exit code is computed here, but `process::exit` is
/// only called from the binary entry point (`src/bin/rampart.rs`), which keeps
/// the library free of the `exit` lint.
#[derive(Debug, Clone, Copy, PartialEq, Eq)]
pub enum RunExit {
/// The listener finished cleanly: the drain fit into the timeout.
Ok,
/// The drain timeout expired: the binary must terminate the process immediately
/// (replaces the former `process::exit(0)` inside the shutdown timer).
DrainTimeout,
}
impl RunExit {
/// Process exit code for this outcome.
///
/// A forced drain timeout loses connections that did not finish in time, so
/// it must not look like a clean shutdown to supervisors and scripts.
#[must_use]
pub const fn code(self) -> i32 {
match self {
Self::Ok => 0,
Self::DrainTimeout => 1,
}
}
}
/// Spawns the background shutdown task.
///
/// The timer must end the process, but `process::exit` now lives in the binary,
/// so this only reports the drain timeout to `run` over a oneshot channel. The
/// process is given five seconds to drain after the signal.
pub(crate) fn spawn_shutdown_drain(shutdown_tx: watch::Sender<bool>, drain_tx: oneshot::Sender<()>) {
tokio::spawn(async move {
wait_for_signal().await;
tracing::info!("shutdown signal received, draining connections...");
let _ = shutdown_tx.send(true);
tokio::time::sleep(Duration::from_secs(5)).await;
tracing::info!("shutdown timeout reached, exiting");
let _ = drain_tx.send(());
});
}
async fn wait_for_signal() {
let ctrl_c = tokio::signal::ctrl_c();
let mut term = tokio::signal::unix::signal(tokio::signal::unix::SignalKind::terminate())
.expect("failed to install SIGTERM handler");
tokio::select! {
_ = ctrl_c => {}
_ = term.recv() => {}
}
}