Files
EasyTier/easytier-core/src/peers/conn/peer_map.rs
T
6487a5e463 fix(peers): don't report unmeasured connections as 1 ms to peer center (#2768)
* fix(peers): don't report unmeasured connections as 1 ms to peer center

A freshly established connection has no latency sample yet, so its
WindowLatency reports 0 and direct_peer_info advertised the peer as
max(1, 0) = 1 ms. With latency_first enabled, every node could then
route through a slow or lossy link until the next report (up to ~60 s),
and a flapping link re-advertised 1 ms on every reconnect.

Skip connections that have no latency sample yet. A peer with no
measured connection is left out of the report; it is reported once its
first sample arrives, because the peer list changes then.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* ci(mobile): use the Android command-line tools build ID

setup-android expects a download build ID, not the displayed tools
version. The value 12.0 becomes 12 and produces a URL that returns 404
when the runner cannot reuse a preinstalled SDK.

Request build 11076708, which provides the same command-line tools
12.0. Keep the runner and other tool versions unchanged.

Validate the workflow YAML and confirm the archive URL returns HTTP 200.

---------

Co-authored-by: imoulei <188999141+imoulei@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: KKRainbow <5665404+KKRainbow@users.noreply.github.com>
2026-10-28 12:27:23 +08:00

508 lines
16 KiB
Rust

use std::{
collections::{BTreeSet, HashMap, HashSet},
net::{Ipv4Addr, Ipv6Addr},
sync::Arc,
};
use anyhow::Context;
use dashmap::{DashMap, DashSet};
use parking_lot::Mutex;
use tokio::sync::RwLock;
use crate::{
config::PeerId,
foundation::small_map_shard_amount,
packet::ZCPacket,
peers::{
context::{ArcPeerContext, NetworkIdentity, PeerEvent},
error::Error,
util::shrink_dashmap,
},
proto::{
core_peer::peer::{PeerConnInfo, Route as CoreRoute},
peer_rpc::{
DirectConnectedPeerInfo, PeerIdentityType, PeerInfoForGlobalMap, RoutePeerInfo,
},
},
tunnel::TunnelError,
};
use super::{
peer::Peer,
peer_conn::{PeerConn, PeerConnId},
};
use crate::peers::{
PacketRecvChan,
route::{ArcRoute, NextHopPolicy},
};
pub struct PeerMap {
context: ArcPeerContext,
my_peer_id: PeerId,
peer_map: DashMap<PeerId, Arc<Peer>>,
packet_send: PacketRecvChan,
routes: RwLock<Vec<ArcRoute>>,
alive_client_urls: Arc<Mutex<HashMap<url::Url, HashSet<PeerConnId>>>>,
}
impl PeerMap {
pub(crate) fn new(
packet_send: PacketRecvChan,
context: ArcPeerContext,
my_peer_id: PeerId,
) -> Self {
PeerMap {
context,
my_peer_id,
peer_map: DashMap::with_shard_amount(small_map_shard_amount()),
packet_send,
routes: RwLock::new(Vec::new()),
alive_client_urls: Arc::new(Mutex::new(HashMap::new())),
}
}
async fn add_new_peer(&self, peer: Peer) {
let peer_id = peer.peer_node_id;
self.peer_map.insert(peer_id, Arc::new(peer));
self.context.issue_event(PeerEvent::PeerAdded(peer_id));
}
pub async fn add_new_peer_conn(&self, peer_conn: PeerConn) -> Result<(), Error> {
let _ = self.maintain_alive_client_urls(&peer_conn);
let peer_id = peer_conn.get_peer_id();
let no_entry = self.peer_map.get(&peer_id).is_none();
if no_entry {
let new_peer = Peer::new(peer_id, self.packet_send.clone(), self.context.clone());
new_peer.add_peer_conn(peer_conn).await?;
self.add_new_peer(new_peer).await;
} else {
let peer = self.peer_map.get(&peer_id).unwrap().clone();
peer.add_peer_conn(peer_conn).await?;
}
Ok(())
}
fn maintain_alive_client_urls(&self, peer_conn: &PeerConn) -> Option<()> {
let conn_info = peer_conn.get_conn_info();
if !conn_info.is_client {
return None;
}
let close_notifier = peer_conn.get_close_notifier();
let alive_conns_weak = Arc::downgrade(&self.alive_client_urls);
let conn_id = close_notifier.get_conn_id();
let alive_client_url: url::Url = conn_info.tunnel?.remote_addr?.into();
self.alive_client_urls
.lock()
.entry(alive_client_url.clone())
.or_default()
.insert(conn_id);
tokio::spawn(async move {
if let Some(mut waiter) = close_notifier.get_waiter().await {
let _ = waiter.recv().await;
}
let Some(alive_conns) = alive_conns_weak.upgrade() else {
return;
};
let mut guard = alive_conns.lock();
if let Some(conn_ids) = guard.get_mut(&alive_client_url) {
conn_ids.retain(|id| id != &conn_id);
if conn_ids.is_empty() {
guard.remove(&alive_client_url);
}
}
let alive_conn_count = guard.len();
drop(guard);
tracing::debug!(
?conn_id,
"peer conn is closed, current alive conns: {}",
alive_conn_count
);
});
Some(())
}
pub fn is_client_url_alive(&self, url: &url::Url) -> bool {
self.alive_client_urls.lock().contains_key(url)
}
pub fn get_peer_by_id(&self, peer_id: PeerId) -> Option<Arc<Peer>> {
self.peer_map.get(&peer_id).map(|v| v.clone())
}
pub fn get_directly_connections_by_peer_id(&self, peer_id: PeerId) -> DashSet<uuid::Uuid> {
if let Some(peer) = self.get_peer_by_id(peer_id) {
return peer.get_directly_connections();
}
DashSet::new()
}
pub fn has_peer(&self, peer_id: PeerId) -> bool {
peer_id == self.my_peer_id || self.peer_map.contains_key(&peer_id)
}
pub(crate) fn has_direct_attached_peer(&self, peer_id: PeerId) -> bool {
self.get_peer_by_id(peer_id)
.is_some_and(|peer| peer.has_direct_attached_conn())
}
pub(crate) fn is_self(&self, peer_id: PeerId) -> bool {
peer_id == self.my_peer_id
}
pub async fn send_msg_directly(&self, msg: ZCPacket, dst_peer_id: PeerId) -> Result<(), Error> {
if dst_peer_id == self.my_peer_id {
let packet_send = self.packet_send.clone();
tokio::spawn(async move {
let ret = packet_send
.send(msg)
.await
.with_context(|| "send msg to self failed");
if ret.is_err() {
tracing::error!("send msg to self failed: {:?}", ret);
}
});
return Ok(());
}
match self.get_peer_by_id(dst_peer_id) {
Some(peer) => {
peer.send_msg(msg).await?;
}
None => {
tracing::error!("no peer for dst_peer_id: {}", dst_peer_id);
return Err(Error::RouteError(Some(format!(
"peer map sengmsg directly no connected dst_peer_id: {}",
dst_peer_id
))));
}
}
Ok(())
}
pub async fn get_gateway_peer_id(
&self,
dst_peer_id: PeerId,
policy: NextHopPolicy,
) -> Option<PeerId> {
if dst_peer_id == self.my_peer_id {
return Some(dst_peer_id);
}
if self.has_peer(dst_peer_id) && matches!(policy, NextHopPolicy::LeastHop) {
return Some(dst_peer_id);
}
// get route info
for route in self.routes.read().await.iter() {
if let Some(gateway_peer_id) = route
.get_next_hop_with_policy(dst_peer_id, policy.clone())
.await
{
// NOTIC: for foreign network, gateway_peer_id may not connect to me
return Some(gateway_peer_id);
}
}
None
}
pub async fn list_peers_own_foreign_network(
&self,
network_identity: &NetworkIdentity,
) -> Vec<PeerId> {
let mut ret = Vec::new();
for route in self.routes.read().await.iter() {
let peers = route.list_peers_own_foreign_network(network_identity).await;
ret.extend(peers);
}
ret
}
pub async fn send_msg(
&self,
msg: ZCPacket,
dst_peer_id: PeerId,
policy: NextHopPolicy,
) -> Result<(), Error> {
let Some(gateway_peer_id) = self.get_gateway_peer_id(dst_peer_id, policy).await else {
return Err(Error::RouteError(Some(format!(
"peer map sengmsg no gateway for dst_peer_id: {}",
dst_peer_id
))));
};
self.send_msg_directly(msg, gateway_peer_id).await?;
Ok(())
}
pub async fn get_peer_id_by_ipv4(&self, ipv4: &Ipv4Addr) -> Option<PeerId> {
for route in self.routes.read().await.iter() {
let peer_id = route.get_peer_id_by_ipv4(ipv4).await;
if peer_id.is_some() {
return peer_id;
}
}
None
}
pub async fn get_peer_id_by_ipv6(&self, ipv6: &Ipv6Addr) -> Option<PeerId> {
for route in self.routes.read().await.iter() {
let peer_id = route.get_peer_id_by_ipv6(ipv6).await;
if peer_id.is_some() {
return peer_id;
}
}
None
}
pub async fn get_route_peer_info(&self, peer_id: PeerId) -> Option<RoutePeerInfo> {
for route in self.routes.read().await.iter() {
if let Some(info) = route.get_peer_info(peer_id).await {
return Some(info);
}
}
None
}
pub async fn get_origin_my_peer_id(
&self,
network_name: &str,
foreign_my_peer_id: PeerId,
) -> Option<PeerId> {
for route in self.routes.read().await.iter() {
let origin_peer_id = route
.get_origin_my_peer_id(network_name, foreign_my_peer_id)
.await;
if origin_peer_id.is_some() {
return origin_peer_id;
}
}
None
}
pub fn is_empty(&self) -> bool {
self.peer_map.is_empty()
}
pub fn list_peers(&self) -> Vec<PeerId> {
let mut ret = Vec::new();
for item in self.peer_map.iter() {
let peer_id = item.key();
ret.push(*peer_id);
}
ret
}
pub async fn list_peers_with_conn(&self) -> Vec<PeerId> {
let mut ret = Vec::new();
for item in self.peer_map.iter() {
if item.value().has_live_conns() {
ret.push(*item.key());
}
}
ret
}
pub async fn list_peer_conns(&self, peer_id: PeerId) -> Option<Vec<PeerConnInfo>> {
if let Some(p) = self.get_peer_by_id(peer_id) {
Some(p.list_peer_conns().await)
} else {
None
}
}
pub async fn get_peer_default_conn_id(&self, peer_id: PeerId) -> Option<PeerConnId> {
self.get_peer_by_id(peer_id)
.map(|p| p.get_default_conn_id())
}
pub fn get_peer_identity_type(&self, peer_id: PeerId) -> Option<PeerIdentityType> {
self.get_peer_by_id(peer_id)
.and_then(|p| p.get_peer_identity_type())
}
pub fn get_peer_public_key(&self, peer_id: PeerId) -> Option<Vec<u8>> {
self.get_peer_by_id(peer_id)
.and_then(|p| p.get_peer_public_key())
}
pub async fn close_peer_conn(
&self,
peer_id: PeerId,
conn_id: &PeerConnId,
) -> Result<(), Error> {
if let Some(p) = self.get_peer_by_id(peer_id) {
p.close_peer_conn(conn_id).await
} else {
Err(Error::NotFound)
}
}
pub async fn close_peer(&self, peer_id: PeerId) -> Result<(), TunnelError> {
let remove_ret = self.peer_map.remove(&peer_id);
shrink_dashmap(&self.peer_map, None);
self.context.issue_event(PeerEvent::PeerRemoved(peer_id));
tracing::info!(
?peer_id,
has_old_value = ?remove_ret.is_some(),
peer_ref_counter = ?remove_ret.map(|v| Arc::strong_count(&v.1)),
"peer is closed"
);
Ok(())
}
pub async fn add_route(&self, route: ArcRoute) {
let mut routes = self.routes.write().await;
routes.insert(0, route);
}
pub(crate) async fn clear_resources(&self) {
for peer_id in self.list_peers() {
let _ = self.close_peer(peer_id).await;
}
let routes = {
let mut routes = self.routes.write().await;
std::mem::take(&mut *routes)
};
for route in routes {
route.close().await;
}
self.alive_client_urls.lock().clear();
}
pub async fn clean_peer_without_conn(&self) {
let mut to_remove = vec![];
for peer_id in self.list_peers() {
let conns = self.list_peer_conns(peer_id).await;
if conns.is_none() || conns.as_ref().unwrap().is_empty() {
to_remove.push(peer_id);
}
}
for peer_id in to_remove {
self.close_peer(peer_id).await.unwrap();
}
}
pub async fn list_routes(&self) -> DashMap<PeerId, PeerId> {
let route_map = DashMap::new();
for route in self.routes.read().await.iter() {
for item in route.list_routes().await.iter() {
route_map.insert(item.peer_id, item.next_hop_peer_id);
}
}
route_map
}
pub async fn list_route_infos(&self) -> Vec<CoreRoute> {
if let Some(route) = self.routes.read().await.iter().next() {
return route.list_routes().await;
}
vec![]
}
pub async fn need_relay_by_foreign_network(&self, dst_peer_id: PeerId) -> Result<bool, Error> {
// if gateway_peer_id is not connected to me, means need relay by foreign network
let gateway_id = self
.get_gateway_peer_id(dst_peer_id, NextHopPolicy::LeastHop)
.await
.ok_or(Error::RouteError(Some(format!(
"peer map need_relay_by_foreign_network no gateway for dst_peer_id: {}",
dst_peer_id
))))?;
Ok(!self.has_peer(gateway_id))
}
pub fn my_peer_id(&self) -> PeerId {
self.my_peer_id
}
}
impl Drop for PeerMap {
fn drop(&mut self) {
tracing::debug!(
self.my_peer_id,
network = ?self.context.network_identity(),
"PeerMap is dropped"
);
}
}
/// Aggregates the directly-connected peers of several peer maps into the
/// peer-center reporting format, keeping the lowest observed latency per peer.
pub(crate) async fn direct_peer_info(peer_maps: &[Arc<PeerMap>]) -> PeerInfoForGlobalMap {
let mut peers = BTreeSet::new();
for peer_map in peer_maps {
peers.extend(peer_map.list_peers());
}
let mut ret = PeerInfoForGlobalMap::default();
for peer in peers {
let mut conns = None;
for peer_map in peer_maps {
if let Some(found) = peer_map.list_peer_conns(peer).await {
conns = Some(found);
break;
}
}
let Some(min_lat) = min_measured_latency_us(
conns
.into_iter()
.flatten()
.map(|conn| conn.stats.as_ref().unwrap().latency_us),
) else {
continue;
};
ret.direct_peers.insert(
peer,
DirectConnectedPeerInfo {
latency_ms: std::cmp::max(1, (min_lat as u32 / 1000) as i32),
},
);
}
ret
}
/// Lowest latency among the connections that already have a latency sample.
///
/// A freshly established connection reports `latency_us == 0` until its first
/// ping completes. Treating that as a real value would advertise the peer as
/// 1 ms away to the peer center, so latency-first routing could detour through
/// a slow or lossy link until the next report. Unmeasured connections are
/// skipped instead; a peer with none measured is left out of the report and is
/// reported once its first sample arrives, because the peer list changes then.
fn min_measured_latency_us(latencies_us: impl IntoIterator<Item = u64>) -> Option<u64> {
latencies_us.into_iter().filter(|lat| *lat > 0).min()
}
#[cfg(test)]
mod tests {
use super::min_measured_latency_us;
#[test]
fn unmeasured_connections_are_not_reported_as_fastest() {
assert_eq!(min_measured_latency_us([0, 306_000]), Some(306_000));
assert_eq!(min_measured_latency_us([306_000, 0, 93_000]), Some(93_000));
}
#[test]
fn peer_without_measured_connection_is_skipped() {
assert_eq!(min_measured_latency_us([0, 0]), None);
assert_eq!(min_measured_latency_us([]), None);
}
#[test]
fn sub_millisecond_lan_latency_is_still_reported() {
assert_eq!(min_measured_latency_us([400]), Some(400));
}
}