Revert "Merge main"

This reverts commit 89f95b15482030ed9d3aee8d3caab68c65f43055.
This commit is contained in:
Theodor S. Midtlien
2026-06-08 14:49:58 +02:00
parent 23c82b32ff
commit 0a61d7b20e
148 changed files with 1748 additions and 4314 deletions

View File

@@ -214,10 +214,7 @@ func runServer(cmd *cobra.Command, args []string) error {
return fmt.Errorf("invalid --trusted-proxies: %w", err)
}
ctx, stop := signal.NotifyContext(context.Background(), syscall.SIGTERM, syscall.SIGINT)
defer stop()
srv := proxy.New(ctx, proxy.Config{
srv := proxy.New(proxy.Config{
ListenAddr: addr,
Logger: logger,
Version: Version,
@@ -254,6 +251,9 @@ func runServer(cmd *cobra.Command, args []string) error {
CrowdSecAPIKey: crowdsecAPIKey,
})
ctx, stop := signal.NotifyContext(context.Background(), syscall.SIGTERM, syscall.SIGINT)
defer stop()
return srv.ListenAndServe(ctx, addr)
}

View File

@@ -5,7 +5,6 @@ import (
"crypto/tls"
"errors"
"fmt"
"io"
stdlog "log"
"net"
"net/http"
@@ -43,7 +42,7 @@ const privateInboundPortHTTPS = 443
const privateInboundPortHTTP = 80
// inboundManager wires per-account inbound listeners into the proxy
// pipeline when --private is enabled. When disabled the manager
// pipeline when --private-inbound is enabled. When disabled the manager
// is nil and every method on *Server that touches it short-circuits.
type inboundManager struct {
logger *log.Logger
@@ -56,18 +55,15 @@ type inboundManager struct {
}
// inboundEntry owns the listeners, router and HTTP servers for a single
// account's embedded netstack. errorLogWriters retain the logrus pipe
// writers backing each http.Server's ErrorLog so tearDown can close
// them — otherwise the pipe + its scanner goroutine leak per account.
// account's embedded netstack.
type inboundEntry struct {
router *nbtcp.Router
tlsListener net.Listener
plainListener net.Listener
httpsServer *http.Server
httpServer *http.Server
errorLogWriters []*io.PipeWriter
cancel context.CancelFunc
wg sync.WaitGroup
router *nbtcp.Router
tlsListener net.Listener
plainListener net.Listener
httpsServer *http.Server
httpServer *http.Server
cancel context.CancelFunc
wg sync.WaitGroup
}
// pendingInboundRoute holds a route that arrived before the account's
@@ -151,34 +147,30 @@ func (m *inboundManager) bringUp(ctx context.Context, accountID types.AccountID,
return types.WithOverlayOrigin(ctx)
}
httpsErrLog, httpsErrW := newInboundErrorLog(m.logger, "https", accountID)
httpErrLog, httpErrW := newInboundErrorLog(m.logger, "http", accountID)
httpsServer := &http.Server{
Handler: scopedHandler,
TLSConfig: m.tlsConfig,
ReadHeaderTimeout: httpInboundReadHeaderTimeout,
IdleTimeout: httpInboundIdleTimeout,
ErrorLog: httpsErrLog,
ErrorLog: newInboundErrorLog(m.logger, "https", accountID),
ConnContext: markOverlayOrigin,
}
httpServer := &http.Server{
Handler: scopedHandler,
ReadHeaderTimeout: httpInboundReadHeaderTimeout,
IdleTimeout: httpInboundIdleTimeout,
ErrorLog: httpErrLog,
ErrorLog: newInboundErrorLog(m.logger, "http", accountID),
ConnContext: markOverlayOrigin,
}
runCtx, cancel := context.WithCancel(ctx)
entry := &inboundEntry{
router: router,
tlsListener: tlsListener,
plainListener: plainListener,
httpsServer: httpsServer,
httpServer: httpServer,
errorLogWriters: []*io.PipeWriter{httpsErrW, httpErrW},
cancel: cancel,
router: router,
tlsListener: tlsListener,
plainListener: plainListener,
httpsServer: httpsServer,
httpServer: httpServer,
cancel: cancel,
}
entry.wg.Add(1)
@@ -245,14 +237,6 @@ func (m *inboundManager) tearDown(accountID types.AccountID, entry *inboundEntry
m.logger.Debugf("close per-account plain listener: %v", err)
}
entry.wg.Wait()
// Close the ErrorLog pipes only after the http.Servers have fully
// stopped so any straggling stdlib write doesn't race with the
// close. Each writer also tears down the logrus scanner goroutine.
for _, w := range entry.errorLogWriters {
if err := w.Close(); err != nil {
m.logger.Debugf("close per-account inbound error log writer: %v", err)
}
}
}
// AddRoute records an SNI/host route on the account's per-account router.
@@ -390,7 +374,7 @@ func (m *inboundManager) ListenerInfo(accountID types.AccountID) (InboundListene
}
// Snapshot returns the inbound listener state for every account that has
// a live listener at call time. Empty when --private is off or
// a live listener at call time. Empty when --private-inbound is off or
// no accounts have come up yet.
func (m *inboundManager) Snapshot() map[types.AccountID]InboundListenerInfo {
if m == nil {
@@ -513,7 +497,7 @@ func accountTunnelLookup(client *embed.Client) auth.TunnelLookupFunc {
// peerstore lookup to every request's context before delegating to next.
// Calling on the host-level listener is a no-op because that path never
// installs this wrapper, so the existing behaviour stays byte-for-byte
// identical when --private is off or the request didn't arrive
// identical when --private-inbound is off or the request didn't arrive
// on a per-account listener.
func withTunnelLookup(next http.Handler, lookup auth.TunnelLookupFunc) http.Handler {
if lookup == nil {
@@ -554,14 +538,10 @@ func (a inboundDebugAdapter) InboundListeners() map[types.AccountID]debug.Inboun
}
// newInboundErrorLog routes a per-account http.Server's stdlib error
// stream through logrus at warn level. The returned PipeWriter must be
// closed by the caller (tearDown) once the http.Server has shut down —
// otherwise the pipe and its scanner goroutine leak per account, see
// logrus.Entry.WriterLevel.
func newInboundErrorLog(logger *log.Logger, scheme string, accountID types.AccountID) (*stdlog.Logger, *io.PipeWriter) {
w := logger.WithFields(log.Fields{
// stream through logrus at warn level.
func newInboundErrorLog(logger *log.Logger, scheme string, accountID types.AccountID) *stdlog.Logger {
return stdlog.New(logger.WithFields(log.Fields{
"inbound-http": scheme,
"account_id": accountID,
}).WriterLevel(log.WarnLevel)
return stdlog.New(w, "", 0), w
}).WriterLevel(log.WarnLevel), "", 0)
}

View File

@@ -4,7 +4,6 @@ import (
"bufio"
"context"
"crypto/tls"
"io"
"net"
"net/http"
"net/http/httptest"
@@ -111,7 +110,7 @@ func TestServer_PrivateInbound_Enabled_WiresLifecycle(t *testing.T) {
// Construct a NetBird transport. We can't actually start the embedded
// client here (that needs a real management server), but we can
// confirm that the lifecycle callbacks are registered.
s.netbird = roundtrip.NewNetBird(t.Context(), "test", "test", roundtrip.ClientConfig{
s.netbird = roundtrip.NewNetBird("test", "test", roundtrip.ClientConfig{
MgmtAddr: "http://invalid.test",
}, quietLogger(), nil, fakeMgmtClient{})
@@ -140,7 +139,7 @@ func TestInboundManager_AddRouteAfterReady_RegistersDirectly(t *testing.T) {
// TestPrivateCapability_DerivedFromPrivateOnly tests that the capability
// bit reported upstream tracks --private exclusively. The previous
// --private flag has been folded into --private.
// --private-inbound flag has been folded into --private.
func TestPrivateCapability_DerivedFromPrivateOnly(t *testing.T) {
tests := []struct {
name string
@@ -319,7 +318,7 @@ func TestInboundManager_ListenerInfo(t *testing.T) {
}
// TestInboundManager_NilManagerSafe ensures the observability accessors
// are safe to call when --private is off (nil manager).
// are safe to call when --private-inbound is off (nil manager).
func TestInboundManager_NilManagerSafe(t *testing.T) {
var mgr *inboundManager
_, ok := mgr.ListenerInfo("anything")
@@ -483,38 +482,6 @@ func selfSignedTLSConfig(t *testing.T) *tls.Config {
return &tls.Config{Certificates: []tls.Certificate{cert}, MinVersion: tls.VersionTLS12} //nolint:gosec
}
// TestNewInboundErrorLog_WriterIsCloseable guards the close path on the
// logrus PipeWriter that backs each per-account http.Server's ErrorLog.
// logrus.Entry.WriterLevel returns an *io.PipeWriter that owns a pipe +
// scanner goroutine; the caller must Close() it on teardown or the
// resources leak per account. The contract is verified two ways:
//
// - the constructor returns a non-nil writer the caller can keep,
// - writing to the writer after Close() fails with io.ErrClosedPipe,
// which is the only externally observable sign that Close was wired.
//
// A leaking refactor (forgetting to thread the writer to tearDown, or
// dropping the Close call) would still pass this test individually but
// fail an integration goleak check; this unit test is the cheap first
// line of defence.
func TestNewInboundErrorLog_WriterIsCloseable(t *testing.T) {
logger := quietLogger()
stdLog, writer := newInboundErrorLog(logger, "https", types.AccountID("acct-1"))
require.NotNil(t, stdLog, "newInboundErrorLog must return a non-nil *log.Logger")
require.NotNil(t, writer, "newInboundErrorLog must return the underlying PipeWriter so tearDown can Close it")
// First Close succeeds.
require.NoError(t, writer.Close(), "PipeWriter.Close should succeed the first time")
// After Close, the writer must refuse new writes — that's the only
// behavioural signal that the pipe (and its scanner goroutine) has
// shut down.
_, err := writer.Write([]byte("post-close write\n"))
require.ErrorIs(t, err, io.ErrClosedPipe,
"writes after Close must surface io.ErrClosedPipe so callers know the writer is gone")
}
// testCertPEM / testKeyPEM are a minimal RSA self-signed cert for
// 127.0.0.1 — only used by tests that need a working TLS handshake.
var testCertPEM = []byte(`-----BEGIN CERTIFICATE-----

View File

@@ -346,15 +346,13 @@ func (mw *Middleware) forwardWithSessionCookie(w http.ResponseWriter, r *http.Re
// management unreachable, peer unknown, user not in group) returns false so
// the caller falls back to the existing OIDC scheme dispatch.
//
// The fast-path is gated on TunnelLookupFromContext(r.Context()) being
// present — that context value is attached only by the per-account
// inbound (overlay) listener. The host listener never sets it, so a
// public client whose source IP happens to fall inside an RFC1918 / ULA
// / CGNAT range can't impersonate a mesh peer by colliding with a
// tunnel-IP. Once we know the request arrived over WireGuard the
// per-account peerstore lookup is consulted: a miss denies fast (no
// management round-trip), a hit gates the cached ValidateTunnelPeer RPC
// that mints the session JWT.
// Phase 3 adds a local-first short-circuit: when the request arrived on a
// per-account inbound listener the context carries a peerstore lookup
// (TunnelLookupFromContext). If the lookup says the IP isn't in the account's
// roster the proxy denies fast without calling management. If the lookup
// confirms a known peer the RPC still runs for the user-identity tail
// (UserID + group access), but its result is cached for tunnelCacheTTL so
// repeat requests skip management entirely.
func (mw *Middleware) forwardWithTunnelPeer(w http.ResponseWriter, r *http.Request, host string, config DomainConfig, next http.Handler) bool {
if mw.sessionValidator == nil {
return false
@@ -363,24 +361,18 @@ func (mw *Middleware) forwardWithTunnelPeer(w http.ResponseWriter, r *http.Reque
if !clientIP.IsValid() {
return false
}
// Anti-spoof: only honour the tunnel-peer fast-path on requests that
// were stamped by an overlay listener. Without that marker an
// attacker could send a request from a colliding RFC1918 / CGNAT
// source on the public listener and bypass operator auth.
lookup := TunnelLookupFromContext(r.Context())
if lookup == nil {
return false
}
if !isTunnelSourceIP(clientIP) {
return false
}
if _, ok := lookup(clientIP); !ok {
mw.logger.WithFields(log.Fields{
"host": host,
"remote": clientIP,
}).Debug("local peerstore: tunnel IP not in account roster; denying without RPC")
return false
if lookup := TunnelLookupFromContext(r.Context()); lookup != nil {
if _, ok := lookup(clientIP); !ok {
mw.logger.WithFields(log.Fields{
"host": host,
"remote": clientIP,
}).Debug("local peerstore: tunnel IP not in account roster; denying without RPC")
return false
}
}
resp, _, err := mw.tunnelCache.fetch(r.Context(), tunnelCacheKey{

View File

@@ -1227,93 +1227,3 @@ func TestProtect_NonOIDCSchemes_PlainHTTP_NotBlocked(t *testing.T) {
assert.Equal(t, http.StatusUnauthorized, rec.Code, "PIN-only domain should serve the login page on plain HTTP")
}
// stubTunnelValidator records ValidateTunnelPeer calls so a test can
// assert whether the fast-path reached management.
type stubTunnelValidator struct {
called bool
resp *proto.ValidateTunnelPeerResponse
}
func (s *stubTunnelValidator) ValidateSession(context.Context, *proto.ValidateSessionRequest, ...grpc.CallOption) (*proto.ValidateSessionResponse, error) {
return nil, errors.New("not used in this test")
}
func (s *stubTunnelValidator) ValidateTunnelPeer(context.Context, *proto.ValidateTunnelPeerRequest, ...grpc.CallOption) (*proto.ValidateTunnelPeerResponse, error) {
s.called = true
return s.resp, nil
}
// TestProtect_TunnelPeerFastPath_RequiresInboundMarker guards the
// anti-spoof gate: a request with an RFC1918 source IP arriving on the
// public listener (no TunnelLookupFromContext attached) must not be
// allowed to take the tunnel-peer fast-path. Without this gate a public
// client whose source IP happens to fall inside an RFC1918 range could
// bypass the configured auth scheme by colliding with a known tunnel
// IP.
func TestProtect_TunnelPeerFastPath_RequiresInboundMarker(t *testing.T) {
validator := &stubTunnelValidator{
resp: &proto.ValidateTunnelPeerResponse{
Valid: true,
SessionToken: "should-not-be-used",
UserId: "user-1",
},
}
mw := NewMiddleware(log.StandardLogger(), validator, nil)
kp := generateTestKeyPair(t)
scheme := &stubScheme{method: auth.MethodPIN, promptID: "pin"}
require.NoError(t, mw.AddDomain("example.com", []Scheme{scheme}, kp.PublicKey, time.Hour, "", "", nil, false))
handler := mw.Protect(newPassthroughHandler())
// Request from an RFC1918 source IP on the public listener — no
// TunnelLookupFromContext attached. The fast-path must reject this
// and fall through to the PIN scheme (which renders 401 on plain
// HTTP for a non-authenticated request).
req := httptest.NewRequest(http.MethodGet, "http://example.com/", nil)
req.RemoteAddr = "100.64.0.5:5000"
rec := httptest.NewRecorder()
handler.ServeHTTP(rec, req)
assert.False(t, validator.called,
"ValidateTunnelPeer must not be invoked when the request lacks the inbound TunnelLookup marker")
assert.Equal(t, http.StatusUnauthorized, rec.Code,
"without the inbound marker the request must fall through to the operator auth scheme")
}
// TestProtect_TunnelPeerFastPath_TakesPathWithInboundMarker verifies
// the positive side: a request marked as overlay-origin (carrying the
// TunnelLookup context value) and matching a tunnel-IP range does take
// the fast-path and reach management.
func TestProtect_TunnelPeerFastPath_TakesPathWithInboundMarker(t *testing.T) {
validator := &stubTunnelValidator{
resp: &proto.ValidateTunnelPeerResponse{
Valid: true,
SessionToken: "tunnel-session-token",
UserId: "user-1",
},
}
mw := NewMiddleware(log.StandardLogger(), validator, nil)
kp := generateTestKeyPair(t)
scheme := &stubScheme{method: auth.MethodPIN, promptID: "pin"}
require.NoError(t, mw.AddDomain("example.com", []Scheme{scheme}, kp.PublicKey, time.Hour, "", "", nil, false))
handler := mw.Protect(newPassthroughHandler())
lookup := TunnelLookupFunc(func(_ netip.Addr) (PeerIdentity, bool) {
return PeerIdentity{}, true
})
req := httptest.NewRequest(http.MethodGet, "http://example.com/", nil)
req.RemoteAddr = "100.64.0.5:5000"
req = req.WithContext(WithTunnelLookup(req.Context(), lookup))
rec := httptest.NewRecorder()
handler.ServeHTTP(rec, req)
assert.True(t, validator.called,
"ValidateTunnelPeer must run when the request carries the inbound TunnelLookup marker")
assert.Equal(t, http.StatusOK, rec.Code,
"a successful tunnel-peer validation must forward to the next handler")
}

View File

@@ -101,10 +101,7 @@ func TestForwardWithTunnelPeer_GroupsPropagateToCapturedData(t *testing.T) {
w, r := newTunnelRequest("100.64.0.10:55555")
cd := proxy.NewCapturedData("")
lookup := TunnelLookupFunc(func(_ netip.Addr) (PeerIdentity, bool) {
return PeerIdentity{}, true
})
r = r.WithContext(proxy.WithCapturedData(WithTunnelLookup(r.Context(), lookup), cd))
r = r.WithContext(proxy.WithCapturedData(r.Context(), cd))
called := false
next := http.HandlerFunc(func(http.ResponseWriter, *http.Request) { called = true })
@@ -151,13 +148,9 @@ func TestForwardWithTunnelPeer_LocalLookupKnownPeerStillRPCs(t *testing.T) {
assert.Equal(t, int32(1), validator.tunnelCalls.Load(), "RPC must run for the user-identity tail when local lookup confirms the peer")
}
// TestForwardWithTunnelPeer_NoLookupRefusesFastPath guards the
// anti-spoof gate: requests that didn't arrive on the per-account
// inbound listener (no TunnelLookup attached) must never reach
// management's ValidateTunnelPeer, even when the source IP looks like
// a tunnel address. A colliding RFC1918 / CGNAT source on the public
// listener would otherwise impersonate a mesh peer.
func TestForwardWithTunnelPeer_NoLookupRefusesFastPath(t *testing.T) {
// TestForwardWithTunnelPeer_NoLookupKeepsLegacyPath ensures the existing
// behaviour stays intact on the host-level listener (no lookup attached).
func TestForwardWithTunnelPeer_NoLookupKeepsLegacyPath(t *testing.T) {
validator := &stubSessionValidator{
respFn: func(_ *proto.ValidateTunnelPeerRequest) *proto.ValidateTunnelPeerResponse {
return &proto.ValidateTunnelPeerResponse{Valid: true, SessionToken: "tok", UserId: "user-1"}
@@ -172,9 +165,9 @@ func TestForwardWithTunnelPeer_NoLookupRefusesFastPath(t *testing.T) {
config, _ := mw.getDomainConfig("svc.example")
handled := mw.forwardWithTunnelPeer(w, r, "svc.example", config, next)
assert.False(t, handled, "fast-path must refuse without the inbound marker")
assert.False(t, called, "next handler must not run")
assert.Equal(t, int32(0), validator.tunnelCalls.Load(), "ValidateTunnelPeer must not be invoked without the inbound marker")
assert.True(t, handled, "host-level path forwards on positive RPC result")
assert.True(t, called, "next handler runs on host-level success")
assert.Equal(t, int32(1), validator.tunnelCalls.Load(), "host-level path always RPCs (Phase 3 unchanged)")
}
// TestForwardWithTunnelPeer_RPCErrorFallsThrough validates that an RPC
@@ -208,13 +201,8 @@ func TestForwardWithTunnelPeer_CacheReusesPositiveResponse(t *testing.T) {
}
mw := newTunnelMiddleware(t, validator)
lookup := TunnelLookupFunc(func(_ netip.Addr) (PeerIdentity, bool) {
return PeerIdentity{}, true
})
for i := 0; i < 4; i++ {
w, r := newTunnelRequest("100.64.0.10:55555")
r = r.WithContext(WithTunnelLookup(r.Context(), lookup))
next := http.HandlerFunc(func(http.ResponseWriter, *http.Request) {})
config, _ := mw.getDomainConfig("svc.example")
handled := mw.forwardWithTunnelPeer(w, r, "svc.example", config, next)
@@ -238,21 +226,11 @@ func TestForwardWithTunnelPeer_RoutesAccountIDIntoCacheKey(t *testing.T) {
require.NoError(t, mw.AddDomain("svc-a.example", nil, "", 0, "acct-a", "svc-a", nil, false))
require.NoError(t, mw.AddDomain("svc-b.example", nil, "", 0, "acct-b", "svc-b", nil, false))
// The fast-path requires the inbound-listener marker on the context.
// The peerstore lookup itself is account-agnostic at this level
// (one TunnelLookupFunc per account is attached by inbound.go); a
// trivial "always hit" lookup is enough to exercise the cache-key
// branch this test covers.
lookup := TunnelLookupFunc(func(_ netip.Addr) (PeerIdentity, bool) {
return PeerIdentity{}, true
})
for _, host := range []string{"svc-a.example", "svc-b.example"} {
w := httptest.NewRecorder()
r := httptest.NewRequest(http.MethodGet, "https://"+host+"/", nil)
r.Host = host
r.RemoteAddr = "100.64.0.10:55555"
r = r.WithContext(WithTunnelLookup(r.Context(), lookup))
config, _ := mw.getDomainConfig(host)
handled := mw.forwardWithTunnelPeer(w, r, host, config, http.HandlerFunc(func(http.ResponseWriter, *http.Request) {}))
require.True(t, handled, "host %s should forward", host)
@@ -336,17 +314,9 @@ func TestPrivateService_ForwardsOnTunnelPeerSuccess(t *testing.T) {
w.WriteHeader(http.StatusOK)
}))
// Per-account inbound listener attaches WithTunnelLookup; without it
// forwardWithTunnelPeer refuses to take the fast-path. Mirror the
// real flow so this test exercises the post-gating success branch.
lookup := TunnelLookupFunc(func(_ netip.Addr) (PeerIdentity, bool) {
return PeerIdentity{}, true
})
req := httptest.NewRequest(http.MethodGet, "https://private.svc/", nil)
req.Host = "private.svc"
req.RemoteAddr = "100.64.0.10:55555"
req = req.WithContext(WithTunnelLookup(req.Context(), lookup))
w := httptest.NewRecorder()
handler.ServeHTTP(w, req)

View File

@@ -131,7 +131,7 @@ func (h *Handler) SetCertStatus(cs certStatus) {
// SetInboundProvider wires per-account inbound listener observability.
// Pass nil (or skip the call) to keep the inbound section out of debug
// responses on proxies that don't run --private.
// responses on proxies that don't run --private-inbound.
func (h *Handler) SetInboundProvider(p InboundProvider) {
h.inbound = p
}

View File

@@ -66,22 +66,6 @@ func (p *ReverseProxy) ServeHTTP(w http.ResponseWriter, r *http.Request) {
return
}
// Loop guard for private services: a peer that hosts the target
// dialing its own service URL would round-trip its own traffic
// through the proxy and back over WG to itself. Refuse the request
// with 421 (Misdirected Request) so the caller sees an explicit
// error instead of silently doubling tunnel traffic.
if p.isSelfTargetLoop(r, result.target.URL) {
if cd := CapturedDataFromContext(r.Context()); cd != nil {
cd.SetOrigin(OriginNoRoute)
}
requestID := getRequestID(r)
web.ServeErrorPage(w, r, http.StatusMisdirectedRequest, "Loop Detected",
"This peer is the target of the requested service. Reach the backend directly instead of dialing the public service URL from the same machine.",
requestID, web.ErrorStatus{Proxy: true, Destination: false})
return
}
ctx := r.Context()
// Set the account ID in the context for the roundtripper to use.
ctx = roundtrip.WithAccountID(ctx, result.accountID)
@@ -123,32 +107,6 @@ func (p *ReverseProxy) ServeHTTP(w http.ResponseWriter, r *http.Request) {
rp.ServeHTTP(w, r.WithContext(ctx))
}
// isSelfTargetLoop reports whether an overlay-origin request is about to
// be forwarded back to the very peer that initiated it. The detection
// is intentionally narrow: it only fires when the request arrived on
// the per-account inbound (overlay) listener (so we're confident the
// source address is the caller's tunnel IP), and only when the resolved
// target host matches that tunnel IP. Catching this here returns 421 to
// the caller instead of letting the proxy round-trip its own traffic
// over WG twice.
func (p *ReverseProxy) isSelfTargetLoop(r *http.Request, target *url.URL) bool {
if target == nil {
return false
}
if !types.IsOverlayOrigin(r.Context()) {
return false
}
srcIP := extractHostIP(r.RemoteAddr)
if !srcIP.IsValid() {
return false
}
targetIP, err := netip.ParseAddr(target.Hostname())
if err != nil {
return false
}
return srcIP.Unmap() == targetIP.Unmap()
}
// rewriteFunc returns a Rewrite function for httputil.ReverseProxy that rewrites
// inbound requests to target the backend service while setting security-relevant
// forwarding headers and stripping proxy authentication credentials.

View File

@@ -20,7 +20,6 @@ import (
"github.com/netbirdio/netbird/proxy/auth"
"github.com/netbirdio/netbird/proxy/internal/roundtrip"
"github.com/netbirdio/netbird/proxy/internal/types"
"github.com/netbirdio/netbird/proxy/web"
)
@@ -1286,103 +1285,6 @@ func TestStampNetBirdIdentity_OmitsGroupsHeaderWhenAllInvalid(t *testing.T) {
"X-NetBird-Groups must not be set when every group label is rejected")
}
// nopOKTransport returns 200 for every request without dialing — used
// by the self-target-loop tests so the non-loop cases don't pay a real
// TCP-dial timeout.
type nopOKTransport struct{}
func (nopOKTransport) RoundTrip(*http.Request) (*http.Response, error) {
return &http.Response{StatusCode: http.StatusOK, Body: http.NoBody, Header: http.Header{}}, nil
}
// TestServeHTTP_SelfTargetLoopReturns421 covers the loop guard for
// private services: when a peer dials a service whose only target is
// the peer itself, the proxy must refuse with 421 (Misdirected
// Request) rather than round-tripping the request back over WG to
// the same peer.
func TestServeHTTP_SelfTargetLoopReturns421(t *testing.T) {
rp := NewReverseProxy(nopOKTransport{}, "auto", nil, nil)
rp.AddMapping(Mapping{
ID: "svc-1",
AccountID: "acct-1",
Host: "private.svc",
Paths: map[string]*PathTarget{
"/": {
URL: &url.URL{Scheme: "http", Host: "100.64.0.5:8080"},
},
},
})
req := httptest.NewRequest(http.MethodGet, "http://private.svc/", nil)
req.Host = "private.svc"
req.RemoteAddr = "100.64.0.5:55555"
req = req.WithContext(types.WithOverlayOrigin(req.Context()))
rec := httptest.NewRecorder()
rp.ServeHTTP(rec, req)
assert.Equal(t, http.StatusMisdirectedRequest, rec.Code,
"a peer dialing a service whose target is itself must get 421")
}
// TestServeHTTP_SelfTargetLoop_NonOverlayRequestPassesThrough verifies
// the guard is scoped to overlay-origin requests. A public-listener
// request that happens to share a source IP with the target host must
// not be misinterpreted as a loop — the gating relies on the inbound
// marker being attached only by the per-account overlay listener.
func TestServeHTTP_SelfTargetLoop_NonOverlayRequestPassesThrough(t *testing.T) {
rp := NewReverseProxy(nopOKTransport{}, "auto", nil, nil)
rp.AddMapping(Mapping{
ID: "svc-1",
AccountID: "acct-1",
Host: "public.svc",
Paths: map[string]*PathTarget{
"/": {
URL: &url.URL{Scheme: "http", Host: "100.64.0.5:8080"},
},
},
})
req := httptest.NewRequest(http.MethodGet, "http://public.svc/", nil)
req.Host = "public.svc"
req.RemoteAddr = "100.64.0.5:55555"
// No WithOverlayOrigin → the guard must not fire.
rec := httptest.NewRecorder()
rp.ServeHTTP(rec, req)
assert.NotEqual(t, http.StatusMisdirectedRequest, rec.Code,
"a non-overlay request with a colliding source IP must not be flagged as a loop")
}
// TestServeHTTP_SelfTargetLoop_OverlayDifferentIPPassesThrough confirms
// that overlay-origin requests with a source IP that does *not* match
// the target host are forwarded normally.
func TestServeHTTP_SelfTargetLoop_OverlayDifferentIPPassesThrough(t *testing.T) {
rp := NewReverseProxy(nopOKTransport{}, "auto", nil, nil)
rp.AddMapping(Mapping{
ID: "svc-1",
AccountID: "acct-1",
Host: "private.svc",
Paths: map[string]*PathTarget{
"/": {
URL: &url.URL{Scheme: "http", Host: "100.64.0.5:8080"},
},
},
})
req := httptest.NewRequest(http.MethodGet, "http://private.svc/", nil)
req.Host = "private.svc"
req.RemoteAddr = "100.64.0.99:55555" // different from the target
req = req.WithContext(types.WithOverlayOrigin(req.Context()))
rec := httptest.NewRecorder()
rp.ServeHTTP(rec, req)
assert.NotEqual(t, http.StatusMisdirectedRequest, rec.Code,
"overlay request with a non-matching source IP must not be flagged as a loop")
}
// TestStampNetBirdIdentity_CapturedDataPresentButEmpty covers requests
// that carry CapturedData with no identity fields populated (e.g. the
// auth middleware ran but the request didn't authenticate). Both

View File

@@ -152,7 +152,6 @@ type managementClient interface {
// backed by underlying NetBird connections.
// Clients are keyed by AccountID, allowing multiple services to share the same connection.
type NetBird struct {
ctx context.Context
proxyID string
proxyAddr string
clientCfg ClientConfig
@@ -214,11 +213,7 @@ func (n *NetBird) AddPeer(ctx context.Context, accountID types.AccountID, key Se
}).Debug("registered service with existing client")
if started && n.statusNotifier != nil {
// Use a background context, not the caller's: the management
// connection notification must land even if the request /
// stream that triggered this registration is cancelled.
// Mirrors the async runClientStartup path.
if err := n.statusNotifier.NotifyStatus(context.Background(), accountID, serviceID, true); err != nil {
if err := n.statusNotifier.NotifyStatus(ctx, accountID, serviceID, true); err != nil {
n.logger.WithFields(log.Fields{
"account_id": accountID,
"service_key": key,
@@ -247,10 +242,8 @@ func (n *NetBird) AddPeer(ctx context.Context, accountID types.AccountID, key Se
}).Info("created new client for account")
// Attempt to start the client in the background; if this fails we will
// retry on the first request via RoundTrip. runClientStartup uses its
// own background context so the caller's request-scoped ctx can't
// cancel the inbound bring-up.
go n.runClientStartup(accountID, entry.client)
// retry on the first request via RoundTrip.
go n.runClientStartup(ctx, accountID, entry.client)
return nil
}
@@ -314,7 +307,7 @@ func (n *NetBird) createClientEntry(ctx context.Context, accountID types.Account
ManagementURL: n.clientCfg.MgmtAddr,
PrivateKey: privateKey.String(),
LogLevel: log.WarnLevel.String(),
BlockInbound: n.clientCfg.BlockInbound,
BlockInbound: n.clientCfg.BlockInbound,
// The embedded proxy peer must never be a stepping stone into
// the proxy host's LAN: it only exists to reach NetBird mesh
// targets or, when direct_upstream is set, the host network
@@ -362,14 +355,8 @@ func (n *NetBird) createClientEntry(ctx context.Context, accountID types.Account
}, nil
}
// runClientStartup starts the client and notifies registered services on
// success. This function runs in a goroutine launched from AddPeer, so it
// must never inherit the caller's request-scoped context — a canceled
// request must not abort the inbound listener bring-up or the management
// status notification. The embedded client.Start gets its own bounded
// startCtx; once Start succeeds, notifyClientReady takes over with a
// fresh context.Background() (see that function for the contract).
func (n *NetBird) runClientStartup(accountID types.AccountID, client *embed.Client) {
// runClientStartup starts the client and notifies registered services on success.
func (n *NetBird) runClientStartup(ctx context.Context, accountID types.AccountID, client *embed.Client) {
startCtx, cancel := context.WithTimeout(context.Background(), 30*time.Second)
defer cancel()
@@ -382,17 +369,7 @@ func (n *NetBird) runClientStartup(accountID types.AccountID, client *embed.Clie
return
}
n.notifyClientReady(accountID, client)
}
// notifyClientReady marks the account's client as started, fires the
// readyHandler hook, and notifies management of the new tunnel
// connection for every registered service. It is split out of
// runClientStartup so a regression test can drive the post-Start tail
// without needing a live embedded client. The contract that the
// hooks/notifier see context.Background() — never the AddPeer caller's
// ctx — lives here.
func (n *NetBird) notifyClientReady(accountID types.AccountID, client *embed.Client) {
// Mark client as started and collect services to notify outside the lock.
n.clientsMux.Lock()
entry, exists := n.clients[accountID]
if exists {
@@ -408,7 +385,7 @@ func (n *NetBird) notifyClientReady(accountID types.AccountID, client *embed.Cli
n.clientsMux.Unlock()
if readyHandler != nil {
state := readyHandler(n.ctx, accountID, client)
state := readyHandler(ctx, accountID, client)
n.clientsMux.Lock()
if e, ok := n.clients[accountID]; ok {
e.inbound = state
@@ -427,7 +404,7 @@ func (n *NetBird) notifyClientReady(accountID types.AccountID, client *embed.Cli
return
}
for _, sn := range toNotify {
if err := n.statusNotifier.NotifyStatus(n.ctx, accountID, sn.serviceID, true); err != nil {
if err := n.statusNotifier.NotifyStatus(ctx, accountID, sn.serviceID, true); err != nil {
n.logger.WithFields(log.Fields{
"account_id": accountID,
"service_key": sn.key,
@@ -689,12 +666,11 @@ func (n *NetBird) ListClientsForStartup() map[types.AccountID]*embed.Client {
// NewNetBird creates a new NetBird transport. Set clientCfg.WGPort to 0 for a random
// OS-assigned port. A fixed port only works with single-account deployments;
// multiple accounts will fail to bind the same port.
func NewNetBird(ctx context.Context, proxyID, proxyAddr string, clientCfg ClientConfig, logger *log.Logger, notifier statusNotifier, mgmtClient managementClient) *NetBird {
func NewNetBird(proxyID, proxyAddr string, clientCfg ClientConfig, logger *log.Logger, notifier statusNotifier, mgmtClient managementClient) *NetBird {
if logger == nil {
logger = log.StandardLogger()
}
return &NetBird{
ctx: ctx,
proxyID: proxyID,
proxyAddr: proxyAddr,
clientCfg: clientCfg,

View File

@@ -11,7 +11,6 @@ import (
"github.com/stretchr/testify/require"
"google.golang.org/grpc"
"github.com/netbirdio/netbird/client/embed"
"github.com/netbirdio/netbird/proxy/internal/types"
"github.com/netbirdio/netbird/shared/management/proto"
)
@@ -31,15 +30,12 @@ type statusCall struct {
accountID types.AccountID
serviceID types.ServiceID
connected bool
// ctx is captured so tests can assert the notifier received a
// fresh background context rather than an inherited request ctx.
ctx context.Context
}
func (m *mockStatusNotifier) NotifyStatus(ctx context.Context, accountID types.AccountID, serviceID types.ServiceID, connected bool) error {
func (m *mockStatusNotifier) NotifyStatus(_ context.Context, accountID types.AccountID, serviceID types.ServiceID, connected bool) error {
m.mu.Lock()
defer m.mu.Unlock()
m.statuses = append(m.statuses, statusCall{accountID, serviceID, connected, ctx})
m.statuses = append(m.statuses, statusCall{accountID, serviceID, connected})
return nil
}
@@ -52,7 +48,7 @@ func (m *mockStatusNotifier) calls() []statusCall {
// mockNetBird creates a NetBird instance for testing without actually connecting.
// It uses an invalid management URL to prevent real connections.
func mockNetBird() *NetBird {
return NewNetBird(context.Background(), "test-proxy", "invalid.test", ClientConfig{
return NewNetBird("test-proxy", "invalid.test", ClientConfig{
MgmtAddr: "http://invalid.test:9999",
WGPort: 0,
PreSharedKey: "",
@@ -283,7 +279,7 @@ func TestNetBird_RoundTrip_RequiresExistingClient(t *testing.T) {
func TestNetBird_AddPeer_ExistingStartedClient_NotifiesStatus(t *testing.T) {
notifier := &mockStatusNotifier{}
nb := NewNetBird(context.Background(), "test-proxy", "invalid.test", ClientConfig{
nb := NewNetBird("test-proxy", "invalid.test", ClientConfig{
MgmtAddr: "http://invalid.test:9999",
WGPort: 0,
PreSharedKey: "",
@@ -299,12 +295,8 @@ func TestNetBird_AddPeer_ExistingStartedClient_NotifiesStatus(t *testing.T) {
nb.clients[accountID].started = true
nb.clientsMux.Unlock()
// Add second service with an already-cancelled caller context —
// should notify immediately (client is started) AND the notification
// must not inherit the cancelled ctx.
cancelledCtx, cancel := context.WithCancel(context.Background())
cancel()
err = nb.AddPeer(cancelledCtx, accountID, "domain2.test", "key-1", types.ServiceID("svc-2"))
// Add second service — should notify immediately since client is already started.
err = nb.AddPeer(context.Background(), accountID, "domain2.test", "key-1", types.ServiceID("svc-2"))
require.NoError(t, err)
calls := notifier.calls()
@@ -312,9 +304,6 @@ func TestNetBird_AddPeer_ExistingStartedClient_NotifiesStatus(t *testing.T) {
assert.Equal(t, accountID, calls[0].accountID)
assert.Equal(t, types.ServiceID("svc-2"), calls[0].serviceID)
assert.True(t, calls[0].connected)
require.NotNil(t, calls[0].ctx, "NotifyStatus must receive a context")
require.NoError(t, calls[0].ctx.Err(),
"already-started NotifyStatus must use a background ctx, not the cancelled caller ctx")
}
// TestNetBird_IdentityForIP_UnknownAccountReturnsFalse confirms that the
@@ -349,7 +338,7 @@ func TestClientEntry_IdentityForIP_InvalidIPReturnsFalse(t *testing.T) {
func TestNetBird_RemovePeer_NotifiesDisconnection(t *testing.T) {
notifier := &mockStatusNotifier{}
nb := NewNetBird(context.Background(), "test-proxy", "invalid.test", ClientConfig{
nb := NewNetBird("test-proxy", "invalid.test", ClientConfig{
MgmtAddr: "http://invalid.test:9999",
WGPort: 0,
PreSharedKey: "",
@@ -371,53 +360,3 @@ func TestNetBird_RemovePeer_NotifiesDisconnection(t *testing.T) {
assert.Equal(t, types.ServiceID("svc-1"), calls[0].serviceID)
assert.False(t, calls[0].connected)
}
// TestNotifyClientReady_UsesBackgroundCtx pins the contract that the
// post-Start hooks (readyHandler + statusNotifier.NotifyStatus) run on
// a fresh context.Background() rather than inheriting the AddPeer
// caller's request- or stream-scoped ctx. Without this, a cancelled
// caller ctx could abort the inbound listener bring-up or cause the
// management status notification to fail spuriously and leave the
// account in a half-connected state.
func TestNotifyClientReady_UsesBackgroundCtx(t *testing.T) {
notifier := &mockStatusNotifier{}
nb := NewNetBird(context.Background(), "test-proxy", "invalid.test", ClientConfig{
MgmtAddr: "http://invalid.test:9999",
}, nil, notifier, &mockMgmtClient{})
accountID := types.AccountID("acct-async")
// Pre-populate a client entry so notifyClientReady has something
// to mark started + something to enumerate for NotifyStatus.
nb.clientsMux.Lock()
nb.clients[accountID] = &clientEntry{
services: map[ServiceKey]serviceInfo{
DomainServiceKey("svc.example"): {serviceID: types.ServiceID("svc-1")},
},
}
nb.clientsMux.Unlock()
var capturedReadyCtx context.Context
nb.SetClientLifecycle(
func(ctx context.Context, _ types.AccountID, _ *embed.Client) any {
capturedReadyCtx = ctx
return nil
},
nil,
)
// Drive the post-Start path directly; a real client.Start would
// need a working management URL.
nb.notifyClientReady(accountID, nil)
require.NotNil(t, capturedReadyCtx, "readyHandler must have been invoked")
require.NoError(t, capturedReadyCtx.Err(),
"readyHandler must receive a background context, not an inherited cancelled one")
deadline, ok := capturedReadyCtx.Deadline()
assert.False(t, ok, "readyHandler ctx must have no deadline (background); got %v", deadline)
calls := notifier.calls()
require.Len(t, calls, 1, "NotifyStatus must be invoked once per registered service")
require.NotNil(t, calls[0].ctx, "NotifyStatus must receive a context")
require.NoError(t, calls[0].ctx.Err(),
"NotifyStatus must receive a background context, not an inherited cancelled one")
}

View File

@@ -1781,14 +1781,11 @@ func TestRouter_PlainHTTP_RoutesToPlainChannel(t *testing.T) {
}
}()
tlsListener, ok := router.HTTPListener().(*chanListener)
require.True(t, ok, "router.HTTPListener() must be the test's chanListener; the test relies on observing its channel directly")
select {
case conn := <-acceptDone:
require.NotNil(t, conn)
_ = conn.Close()
case <-tlsListener.ch:
case <-router.HTTPListener().(*chanListener).ch:
t.Fatal("plain HTTP request leaked into TLS channel")
case <-time.After(3 * time.Second):
t.Fatal("plain HTTP connection never reached plain channel")

View File

@@ -1,7 +1,6 @@
package proxy
import (
"context"
"net/netip"
"time"
@@ -21,17 +20,14 @@ import (
type Config struct {
// ListenAddr is the TCP address the main listener binds. Required.
ListenAddr string
// ID identifies this proxy instance to management. Empty values are
// replaced with a timestamped default at Server.Start time (see
// initDefaults), not in New.
// ID identifies this proxy instance to management. Empty value lets
// New generate a timestamped default.
ID string
// Logger is the logrus logger used everywhere. Empty values fall
// back to log.StandardLogger() at Server.Start time (see
// initDefaults), not in New.
// Logger is the logrus logger used everywhere. Empty value falls back
// to log.StandardLogger().
Logger *log.Logger
// Version is the build version string reported to management. Empty
// values are replaced with "dev" at Server.Start time (see
// initDefaults), not in New.
// becomes "dev".
Version string
// ProxyURL is the public address operators use to reach this proxy.
ProxyURL string
@@ -129,9 +125,8 @@ type Config struct {
// bound — call Start to bring the proxy up. Returning a fully-formed
// Server keeps the standalone code path (which still constructs Server
// directly) byte-for-byte equivalent.
func New(ctx context.Context, cfg Config) *Server {
func New(cfg Config) *Server {
return &Server{
ctx: ctx,
ListenAddr: cfg.ListenAddr,
ID: cfg.ID,
Logger: cfg.Logger,

View File

@@ -73,7 +73,7 @@ func benchServerWithLatency(b *testing.B, createPeerDelay, statusDelay time.Dura
statusUpdateDelay: statusDelay,
}
nb := roundtrip.NewNetBird(b.Context(), "bench-proxy", "bench.test",
nb := roundtrip.NewNetBird("bench-proxy", "bench.test",
roundtrip.ClientConfig{MgmtAddr: "http://bench.test:9999"},
logger, nil, mgmtClient)

View File

@@ -75,7 +75,6 @@ type portRouter struct {
}
type Server struct {
ctx context.Context
mgmtClient proto.ProxyServiceClient
proxy *proxy.ReverseProxy
netbird *roundtrip.NetBird
@@ -282,7 +281,7 @@ func (s *Server) NotifyCertificateIssued(ctx context.Context, accountID types.Ac
}
// inboundListenerProto resolves the per-account inbound listener state for
// the SendStatusUpdate payload. Returns nil when --private is off
// the SendStatusUpdate payload. Returns nil when --private-inbound is off
// or the account has no live listener so management treats the field as
// absent.
func (s *Server) inboundListenerProto(accountID types.AccountID) *proto.ProxyInboundListener {
@@ -529,10 +528,10 @@ func (s *Server) initManagementClient() error {
}
// initNetBirdClient builds the multi-tenant embedded NetBird client used
// for outbound RoundTripping and (when --private is on) per-account
// for outbound RoundTripping and (when --private-inbound is on) per-account
// inbound listeners.
func (s *Server) initNetBirdClient() {
s.netbird = roundtrip.NewNetBird(s.ctx, s.ID, s.ProxyURL, roundtrip.ClientConfig{
s.netbird = roundtrip.NewNetBird(s.ID, s.ProxyURL, roundtrip.ClientConfig{
MgmtAddr: s.ManagementAddress,
WGPort: s.WireguardPort,
PreSharedKey: s.PreSharedKey,

View File

@@ -64,7 +64,7 @@ func quietLifecycleLogger() *log.Logger {
}
func TestStopBeforeStartIsNoOp(t *testing.T) {
srv := New(t.Context(), Config{Logger: quietLifecycleLogger()})
srv := New(Config{Logger: quietLifecycleLogger()})
ctx, cancel := context.WithTimeout(context.Background(), time.Second)
defer cancel()
@@ -77,7 +77,7 @@ func TestStopBeforeStartIsNoOp(t *testing.T) {
}
func TestStartFailsWithoutManagement(t *testing.T) {
srv := New(t.Context(), Config{
srv := New(Config{
Logger: quietLifecycleLogger(),
ListenAddr: "127.0.0.1:0",
ManagementAddress: "://broken-url",
@@ -137,7 +137,7 @@ func TestRecordRunErrPreservesFirstFailure(t *testing.T) {
}
func TestStopSkipsShutdownWhenNeverStarted(t *testing.T) {
srv := New(t.Context(), Config{Logger: quietLifecycleLogger()})
srv := New(Config{Logger: quietLifecycleLogger()})
ctx, cancel := context.WithCancel(context.Background())
cancel()