Files
tsunagi/crates/tsunagi-wireguard/src/session.rs
T
Aleksandr Bogomiakov 2166ba4800 Separate internal WireGuard encryption from data carriers
AWG-only policy also selected a competing end-to-end plugin, so compatible
external links could still have no compatible encrypted data channel. Carrier
changes replaced tunnel state and status selected counters from different
engines. Use one internal WireGuard plugin and a separate DataCarrier contract
for external capabilities/lifecycle. Keep the logical link, keys, handshake,
interface and counter epoch across policy, fallback and direct/relay changes.
Control/discovery transport selection and membership remain independent.

Share the bounded per-neighbour UDP WireGuard engine between ordinary WG and
Amnezia masking. Encrypt the relay envelope inside the outer hop channel with
fresh pair keys/PSK. Remove duplicate WG key/device code and the TCP/TLS inner
cipher. TLS carries opaque internal WireGuard frames. Refuse missing native
transports instead of letting default iroh impersonate their protocol names.
Probe the actual authenticated carrier, independently of the shared IP tunnel;
failed local probe submissions count as loss too.

Apply the vendored iroh interface exclusion to actual direct candidates, not
just public discovery records. Exclude overlay subnets from introductions and
propagate the actual owned interface name, including kernel-assigned utun names,
to every carrier. Production logs showed paths through the overlay itself.
This removes a recursive transport path; it does not establish the historical
cause of the reported 100 GB absolute counter. Exit-node traffic is also real
peer IP traffic even when local applications are idle.

Hide inner cipher choices/handshake clutter in CLI and tray. Report external
routes, readiness and actionable policy/version failures. Count original IP
packets only when their ciphertext is accepted by the logical transport,
including queued packets' exact unpadded lengths. Exclude physical probes and
keepalives. Export counter epochs; sum per-peer rate deltas and discard stale
histories so joins/restarts cannot turn old totals into fresh traffic. Status
also prints exact IP byte totals for comparison with the tray.

Coordinated upgrade: data ALPN 6, internal wireguard-e2e capability v4,
wg/awg v2, QUIC carrier v1, TLS v2, local IPC v23. Control ALPN 8, device/network
identity derivation, signed records and saved network choices stay unchanged.
Keep the historical wg-quic key store; transactionally import missing keys from
wg then awg without replacing canonical keys. Forget removes legacy copies too,
on the blocking pool, to prevent resurrection after restart. Preserve historic
cipher options as shared settings, rejecting contradictory values.

Validation on Linux in the Nix development shell:
- cargo fmt --all -- --check
- cargo clippy --locked --workspace --all-targets -- -D warnings
- cargo test --locked --workspace --all-targets (562 passed, 3 ignored)
The inherited service socket/secret/group environment overrides were removed
for isolated CLI process tests. A new carrier test initially raced the first
peer announcement; its deadline polling now waits for a peer row explicitly.
Real loopback endpoints, membership, SQLite, WG/AWG/TLS encryption and memory
TUNs cover mixed-carrier relaying and one smoltcp TCP connection surviving
relay/direct/policy changes and a dropped transmission. Counter tests check
exact plaintext lengths, refused sends, epochs, joins and stale histories.
Public DHT, privileged kernel and manual benchmark tests remain ignored;
macOS/Windows native TUNs, desktop GUI and WAN/DPI behavior were not run.
2026-10-22 01:07:52 +01:00

415 lines
15 KiB
Rust

//! One cancellable actor and UDP socket per pair. All payload and health frames
//! are inside the Amnezia-masked authenticated WireGuard tunnel.
use crate::profile::PairProfile;
use boringtun::noise::{Tunn, TunnResult};
use bytes::Bytes;
use iroh::EndpointId;
use std::collections::{HashMap, VecDeque};
use std::net::{Ipv4Addr, SocketAddr};
use std::sync::atomic::{AtomicBool, AtomicU64, Ordering};
use std::sync::{Arc, Mutex};
use std::time::{Duration, Instant};
use tokio::net::UdpSocket;
use tokio::sync::{mpsc, watch};
use tsunagi::dataplane::transport::{PacketLink, TransportError};
use tsunagi::{BoxFuture, NetworkId};
pub(crate) const MAX_PAYLOAD: usize = tsunagi::config::MAX_DATA_DATAGRAM;
const CHUNK: usize = 700;
const QUEUE: usize = 64;
const WINDOW: Duration = Duration::from_secs(20);
pub(crate) struct PairLink {
network: NetworkId,
protocol: &'static str,
peer: EndpointId,
pub(crate) attempt: [u8; 16],
tx: mpsc::Sender<Bytes>,
rx: tokio::sync::Mutex<mpsc::Receiver<Bytes>>,
stop: watch::Sender<bool>,
closed: AtomicBool,
remote: Mutex<Option<SocketAddr>>,
reason: Mutex<Option<String>>,
checks: Arc<AtomicU64>,
}
impl std::fmt::Debug for PairLink {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
f.debug_struct("AwgPair")
.field("network", &self.network)
.field("peer", &self.peer)
.finish_non_exhaustive()
}
}
impl PairLink {
pub(crate) fn delivery_checks(&self) -> u64 {
self.checks.load(Ordering::Relaxed)
}
pub(crate) fn has_authenticated_path(&self) -> bool {
match self.remote.lock() {
Ok(remote) => remote.is_some(),
Err(_) => false,
}
}
}
impl PacketLink for PairLink {
fn network(&self) -> NetworkId {
self.network
}
fn peer(&self) -> EndpointId {
self.peer
}
fn max_datagram_size(&self) -> usize {
MAX_PAYLOAD
}
fn send(&self, payload: Bytes) -> Result<(), TransportError> {
if self.is_closed() {
return Err(TransportError::Closed);
}
if payload.len() > MAX_PAYLOAD {
return Err(TransportError::TooLarge {
size: payload.len(),
limit: MAX_PAYLOAD,
});
}
self.tx
.try_send(payload)
.map_err(|_| TransportError::Other("AWG send queue full or closed".into()))
}
fn recv(&self) -> BoxFuture<'_, Option<Bytes>> {
Box::pin(async move {
let mut stop = self.stop.subscribe();
if *stop.borrow() {
return None;
}
let mut rx = self.rx.lock().await;
tokio::select! {biased; _=stop.changed()=>None,packet=rx.recv()=>packet}
})
}
fn closed(&self) -> BoxFuture<'_, ()> {
Box::pin(async move {
let mut stop = self.stop.subscribe();
if !*stop.borrow() {
let _ = stop.changed().await;
}
})
}
fn is_closed(&self) -> bool {
self.closed.load(Ordering::Acquire) || *self.stop.borrow()
}
fn close(&self) {
self.stop.send_replace(true);
}
fn failure_reason(&self) -> Option<String> {
match self.reason.lock() {
Ok(reason) => reason.clone(),
Err(_) => None,
}
}
fn path_description(&self) -> String {
match self.remote.lock() {
Ok(addr) => match *addr {
Some(addr) => format!("{} UDP via {addr}", self.protocol),
None => format!("{} UDP, establishing", self.protocol),
},
Err(_) => "awg UDP".into(),
}
}
}
impl Drop for PairLink {
fn drop(&mut self) {
self.stop.send_replace(true);
}
}
pub(crate) fn start(
pair: (NetworkId, EndpointId),
socket: UdpSocket,
tunn: Tunn,
profile: PairProfile,
candidates: Vec<SocketAddr>,
protocol: &'static str,
parent: watch::Receiver<bool>,
) -> Arc<PairLink> {
let (network, peer) = pair;
let (tx, rx) = mpsc::channel(QUEUE);
let (in_tx, in_rx) = mpsc::channel(QUEUE);
let (stop, _) = watch::channel(false);
let link = Arc::new(PairLink {
network,
protocol,
peer,
attempt: profile.id(),
tx,
rx: tokio::sync::Mutex::new(in_rx),
stop: stop.clone(),
closed: AtomicBool::new(false),
remote: Mutex::new(None),
reason: Mutex::new(None),
checks: Arc::new(AtomicU64::new(0)),
});
let weak = Arc::downgrade(&link);
let checks = link.checks.clone();
tokio::spawn(async move {
let mut engine = Engine {
stop: stop.subscribe(),
socket,
tunn,
profile,
candidates,
remote: None,
inbox: in_tx,
rx,
counter: rand::random(),
sequence: 0,
pending: VecDeque::new(),
samples: VecDeque::new(),
assemblies: HashMap::new(),
started: Instant::now(),
last_authenticated: Instant::now(),
checks,
};
engine.run(parent, &weak).await;
if let Some(link) = weak.upgrade() {
link.closed.store(true, Ordering::Release);
link.stop.send_replace(true);
}
});
link
}
struct Assembly {
created: Instant,
count: u16,
chunks: Vec<Option<Bytes>>,
}
struct Engine {
stop: watch::Receiver<bool>,
socket: UdpSocket,
tunn: Tunn,
profile: PairProfile,
candidates: Vec<SocketAddr>,
remote: Option<SocketAddr>,
inbox: mpsc::Sender<Bytes>,
rx: mpsc::Receiver<Bytes>,
counter: u64,
sequence: u64,
pending: VecDeque<(u64, Instant)>,
samples: VecDeque<bool>,
assemblies: HashMap<u64, Assembly>,
started: Instant,
last_authenticated: Instant,
checks: Arc<AtomicU64>,
}
enum ResultOwned {
Wire(Vec<u8>),
Plain(Vec<u8>),
Done,
}
fn own(result: TunnResult<'_>) -> ResultOwned {
match result {
TunnResult::WriteToNetwork(p) => ResultOwned::Wire(p.to_vec()),
TunnResult::WriteToTunnelV4(p, _) => ResultOwned::Plain(p.to_vec()),
_ => ResultOwned::Done,
}
}
fn ip(payload: &[u8]) -> Bytes {
let mut packet = vec![0; 20];
packet[0] = 0x45;
packet[8] = 64;
packet[9] = 253;
packet[12..16].copy_from_slice(&Ipv4Addr::new(192, 0, 2, 1).octets());
packet[16..20].copy_from_slice(&Ipv4Addr::new(192, 0, 2, 2).octets());
packet[2..4].copy_from_slice(&((20 + payload.len()) as u16).to_be_bytes());
let sum = packet
.as_chunks::<2>()
.0
.iter()
.map(|b| u32::from(u16::from_be_bytes([b[0], b[1]])))
.sum::<u32>();
let sum = (sum & 0xffff) + (sum >> 16);
packet[10..12].copy_from_slice(&(!(sum as u16)).to_be_bytes());
packet.extend_from_slice(payload);
Bytes::from(packet)
}
impl Engine {
async fn run(&mut self, mut stop: watch::Receiver<bool>, link: &std::sync::Weak<PairLink>) {
if *self.stop.borrow() || *stop.borrow() {
return;
}
let mut buffer = [0; 1201];
let mut scratch = [0; 4096];
let mut timer = tokio::time::interval(Duration::from_millis(250));
timer.set_missed_tick_behavior(tokio::time::MissedTickBehavior::Skip);
let mut last_probe = Instant::now() - Duration::from_secs(2);
loop {
tokio::select! {
biased;
_=stop.changed()=>break,
_=self.stop.changed()=>break,
_=timer.tick()=>{
let result=own(self.tunn.update_timers(&mut scratch));self.process(result,None).await;
if last_probe.elapsed()>=Duration::from_secs(1) {
last_probe=Instant::now();self.sequence=self.sequence.wrapping_add(1);
self.pending.push_back((self.sequence,Instant::now()));
let mut probe=vec![1];probe.extend_from_slice(&self.sequence.to_be_bytes());
probe.resize(713,0);rand::fill(&mut probe[9..]);self.encrypt(&probe).await;
}
while self.pending.front().is_some_and(|(_,at)|at.elapsed()>=Duration::from_secs(3)) {
self.pending.pop_front();self.samples.push_back(false);
}
while self.samples.len()>20 {self.samples.pop_front();}
self.assemblies.retain(|_,a|a.created.elapsed()<Duration::from_secs(5));
if self.started.elapsed()>WINDOW && (self.last_authenticated.elapsed()>Duration::from_secs(10)
|| (self.samples.len()>=15&&self.samples.iter().filter(|ok|!**ok).count()*100>=self.samples.len()*60)) {
if let Some(link)=link.upgrade() && let Ok(mut reason)=link.reason.lock() {*reason=Some("AWG data delivery checks: persistent loss or timeout".into());}
break;
}
}
packet=self.socket.recv_from(&mut buffer)=>{
if let Ok((len,source))=packet && let Some(message)=self.profile.decode(&buffer[..len]) {
let result=own(self.tunn.decapsulate(None,&message,&mut scratch));
if matches!(result,ResultOwned::Plain(_)) {
self.remote=Some(source);self.last_authenticated=Instant::now();
if let Some(link)=link.upgrade() && let Ok(mut remote)=link.remote.lock(){*remote=Some(source);}
}
self.process(result,Some(source)).await;
for _ in 0..16 {
let result=own(self.tunn.decapsulate(None,&[],&mut scratch));
if matches!(result,ResultOwned::Done){break;}
self.process(result,Some(source)).await;
}
}
}
packet=self.rx.recv()=>{
let Some(packet)=packet else{break;};
let count=packet.len().max(1).div_ceil(CHUNK) as u16;
self.counter=self.counter.wrapping_add(1);
for (index,chunk) in packet.chunks(CHUNK).enumerate() {
let mut payload=vec![3];payload.extend_from_slice(&self.counter.to_be_bytes());
payload.extend_from_slice(&(index as u16).to_be_bytes());payload.extend_from_slice(&count.to_be_bytes());
payload.extend_from_slice(chunk);self.encrypt(&payload).await;
// Large logical datagrams must let local readers and other
// pairs drain their bounded socket/command queues.
if (index + 1).is_multiple_of(16) {tokio::task::yield_now().await;}
}
}
}
}
}
async fn encrypt(&mut self, payload: &[u8]) {
let packet = self.profile.pad(ip(payload));
let mut scratch = [0; 4096];
let result = own(self.tunn.encapsulate(&packet, &mut scratch));
self.process(result, None).await;
}
async fn wire(&self, raw: &[u8], target: Option<SocketAddr>) {
let Some(wire) = self.profile.encode(raw) else {
return;
};
let targets: Vec<_> = target
.or(self.remote)
.map(|a| vec![a])
.unwrap_or_else(|| self.candidates.clone());
for addr in targets {
if raw.first() == Some(&1) {
for signature in self.profile.signature_packets() {
let _ = self.send_wire(&signature, addr).await;
}
for junk in self.profile.junk_packets() {
let _ = self.send_wire(&junk, addr).await;
}
}
let _ = self.send_wire(&wire, addr).await;
}
}
async fn send_wire(&self, bytes: &[u8], address: SocketAddr) {
let mut stop = self.stop.clone();
if *stop.borrow() {
return;
}
tokio::select! {biased; _=stop.changed()=>{}, _=tokio::time::timeout(Duration::from_millis(100),self.socket.send_to(bytes,address))=>{}}
}
async fn process(&mut self, result: ResultOwned, target: Option<SocketAddr>) {
match result {
ResultOwned::Wire(raw) => self.wire(&raw, target).await,
ResultOwned::Plain(packet) => {
let Some(payload) = packet.get(20..) else {
return;
};
let Some(seq) = payload
.get(1..9)
.and_then(|b| b.try_into().ok())
.map(u64::from_be_bytes)
else {
return;
};
match payload.first() {
Some(1) => {
let mut pong = vec![2];
pong.extend_from_slice(&seq.to_be_bytes());
pong.resize(713, 0);
let packet = self.profile.pad(ip(&pong));
let mut scratch = [0; 4096];
if let ResultOwned::Wire(raw) =
own(self.tunn.encapsulate(&packet, &mut scratch))
{
self.wire(&raw, target).await;
}
}
Some(2) => {
if let Some(index) =
self.pending.iter().position(|(pending, _)| *pending == seq)
{
self.pending.remove(index);
self.samples.push_back(true);
self.checks.fetch_add(1, Ordering::Relaxed);
}
}
Some(3) => self.fragment(seq, payload),
_ => {}
}
}
ResultOwned::Done => {}
}
}
fn fragment(&mut self, id: u64, payload: &[u8]) {
let Some(header) = payload.get(9..13) else {
return;
};
let index = u16::from_be_bytes([header[0], header[1]]);
let count = u16::from_be_bytes([header[2], header[3]]);
let chunk = &payload[13..];
if count == 0
|| usize::from(count) > MAX_PAYLOAD.div_ceil(CHUNK)
|| index >= count
|| chunk.len() > CHUNK
{
return;
}
if !self.assemblies.contains_key(&id) && self.assemblies.len() >= 8 {
return;
}
let assembly = self.assemblies.entry(id).or_insert_with(|| Assembly {
created: Instant::now(),
count,
chunks: vec![None; count.into()],
});
if assembly.count != count {
return;
}
assembly.chunks[usize::from(index)] = Some(Bytes::copy_from_slice(chunk));
if assembly.chunks.iter().all(Option::is_some)
&& let Some(assembly) = self.assemblies.remove(&id)
{
let mut packet = Vec::new();
for chunk in assembly.chunks.into_iter().flatten() {
packet.extend_from_slice(&chunk);
}
if packet.len() <= MAX_PAYLOAD {
let _ = self.inbox.try_send(Bytes::from(packet));
}
}
}
}