Assert the login status after the SSO flow starts

Return the login failure instead of demanding SSO when management is unreachable
[client] Support Andorid session expiry handling (#6945 )
2026-07-31 04:41:28 +02:00 · 2026-07-30 19:49:53 +02:00 · 2026-07-30 14:34:34 +02:00 · 2026-07-30 10:27:52 +02:00
4 changed files with 455 additions and 5 deletions
--- a/client/android/client.go
+++ b/client/android/client.go
@@ -76,6 +76,24 @@ type Client struct {
 	connectClient *internal.ConnectClient
 	config        *profilemanager.Config
 	cacheDir      string
+
+	stateChangeMu    sync.Mutex
+	stateChangeSubID string
+	eventSub         *peer.EventSubscription
+	// Closed to stop the watch goroutines from delivering buffered items to a
+	// listener that has been removed or replaced. See stopStateChangeWatchLocked.
+	stateChangeDone chan struct{}
+
+	// Latched "the server wants an interactive login": survives the engine
+	// restarts that replace the run loop's context state. See Client.Status.
+	// Guarded by loginRequiredMu together with loginCleared, which counts
+	// clears so a stale observation cannot re-latch over one.
+	loginRequiredMu sync.Mutex
+	loginRequired   bool
+	loginCleared    uint64
+
+	extendMu     sync.Mutex
+	extendCancel context.CancelFunc
 }

 func (c *Client) setState(cfg *profilemanager.Config, cacheDir string, cc *internal.ConnectClient) {
@@ -149,11 +167,16 @@ func (c *Client) Run(platformFiles PlatformFiles, urlOpener URLOpener, isAndroid
 	if err != nil {
 		return err
 	}
-
 	// todo do not throw error in case of cancelled context
 	ctx = internal.CtxInitState(ctx)
 	connectClient := internal.NewConnectClient(ctx, cfg, c.recorder)
 	c.setState(cfg, cacheDir, connectClient)
+	// This path runs the interactive SSO flow, so reaching here means the peer
+	// is authenticated again — release the latch Status() reports from. Clear
+	// only once the fresh connect client is installed: until then Status()
+	// still reads the previous run's context state, which holds the NeedsLogin
+	// that prompted this login, and would re-latch what was just cleared.
+	c.clearLoginRequired()
 	return connectClient.RunOnAndroid(c.tunAdapter, c.iFaceDiscover, c.networkChangeListener, slices.Clone(dns.items), dnsReadyListener, stateFile, cacheDir)
 }

--- a/client/android/session.go
+++ b/client/android/session.go
@@ -0,0 +1,309 @@
+//go:build android
+
+package android
+
+import (
+	"context"
+	"fmt"
+
+	log "github.com/sirupsen/logrus"
+
+	"github.com/netbirdio/netbird/client/internal"
+	"github.com/netbirdio/netbird/client/internal/auth"
+	"github.com/netbirdio/netbird/client/internal/auth/sessionwatch"
+	"github.com/netbirdio/netbird/client/internal/peer"
+	cProto "github.com/netbirdio/netbird/client/proto"
+)
+
+// StateChangeListener receives client state notifications.
+//
+// OnStateChanged is a payload-free wake-up whenever the state snapshot
+// changed: connection state, the run-loop status label (e.g. NeedsLogin) or
+// the session deadline. It mirrors the daemon's SubscribeStatus stream
+// trigger — on each signal the consumer pulls the fresh values via
+// Status() / SessionExpiresAtUnix().
+//
+// OnSessionExpiring forwards the engine's session-expiry warnings, fired at
+// sessionwatch.WarningLead before the deadline and again at FinalWarningLead
+// (finalWarning true). The second one is suppressed when the user dismissed
+// the first via DismissSessionWarning. The daemon turns the same events into
+// its tray notification.
+type StateChangeListener interface {
+	OnStateChanged()
+	OnSessionExpiring(expiresAtUnix int64, leadMinutes int64, finalWarning bool)
+}
+
+// Status returns the connect run-loop's status label — the same value the
+// desktop daemon serves in StatusResponse.Status. "NeedsLogin" means the
+// management server rejected the peer and an interactive login is required.
+//
+// The label is latched: the run loop keeps its status in a per-run context
+// state, which a restart replaces with a fresh Idle one, so an engine restart
+// (network change, always-on) would otherwise erase the fact that the peer
+// still needs to log in. Only a successful interactive login or extend clears
+// it — see clearLoginRequired.
+func (c *Client) Status() string {
+	latched, generation := c.loginRequiredState()
+	if latched {
+		return string(internal.StatusNeedsLogin)
+	}
+	cc := c.getConnectClient()
+	if cc == nil {
+		return string(internal.StatusIdle)
+	}
+	status := cc.Status()
+	if status == internal.StatusNeedsLogin {
+		c.latchLoginRequired(generation)
+	}
+	return string(status)
+}
+
+func (c *Client) loginRequiredState() (bool, uint64) {
+	c.loginRequiredMu.Lock()
+	defer c.loginRequiredMu.Unlock()
+	return c.loginRequired, c.loginCleared
+}
+
+// latchLoginRequired records a NeedsLogin observation, unless a clear landed
+// while the caller was reading the run loop's status: cc.Status() is read
+// outside the lock, so a login or extend completing in that window would
+// otherwise be undone by this stale observation, stranding the UI on
+// "login required" over a healthy session.
+func (c *Client) latchLoginRequired(observedGeneration uint64) {
+	c.loginRequiredMu.Lock()
+	defer c.loginRequiredMu.Unlock()
+	if c.loginCleared != observedGeneration {
+		return
+	}
+	c.loginRequired = true
+}
+
+// clearLoginRequired releases the latch after a successful interactive login
+// or session extend, and invalidates any observation already in flight.
+func (c *Client) clearLoginRequired() {
+	c.loginRequiredMu.Lock()
+	defer c.loginRequiredMu.Unlock()
+	c.loginRequired = false
+	c.loginCleared++
+}
+
+// SessionExpiresAtUnix returns the SSO session deadline as unix seconds, or 0
+// when no deadline is known (not SSO-registered, expiry disabled, or the
+// engine has not received one yet). A past value means the session expired.
+// Mirror of StatusResponse.sessionExpiresAt on the desktop daemon.
+func (c *Client) SessionExpiresAtUnix() int64 {
+	deadline := c.recorder.GetSessionExpiresAt()
+	if deadline.IsZero() {
+		return 0
+	}
+	return deadline.Unix()
+}
+
+// SetStateChangeListener registers the state notification listener.
+// Replaces any previously registered listener; remove it with
+// RemoveStateChangeListener.
+func (c *Client) SetStateChangeListener(listener StateChangeListener) {
+	c.stateChangeMu.Lock()
+	defer c.stateChangeMu.Unlock()
+	c.stopStateChangeWatchLocked()
+	if listener == nil {
+		return
+	}
+
+	// Both subscriptions are buffered (one pending tick, ten pending events),
+	// so unsubscribing is not enough to stop callbacks: the loops would drain
+	// what is already queued and deliver it to a listener the caller has
+	// already removed or replaced. Gate every callback on this registration's
+	// own signal, which is closed before unsubscribing.
+	done := make(chan struct{})
+	c.stateChangeDone = done
+
+	id, ch := c.recorder.SubscribeToStateChanges()
+	c.stateChangeSubID = id
+	// The channel is closed by UnsubscribeFromStateChanges, which ends the
+	// goroutine. Ticks are coalesced (buffer of one), so a burst of changes
+	// wakes the listener once.
+	go func() {
+		for range ch {
+			select {
+			case <-done:
+				return
+			default:
+			}
+			listener.OnStateChanged()
+		}
+	}()
+
+	c.eventSub = c.recorder.SubscribeToEvents()
+	go watchSessionWarnings(c.eventSub, listener, done)
+}
+
+// RemoveStateChangeListener unregisters the state notification listener.
+func (c *Client) RemoveStateChangeListener() {
+	c.stateChangeMu.Lock()
+	defer c.stateChangeMu.Unlock()
+	c.stopStateChangeWatchLocked()
+}
+
+// DismissSessionWarning records the user's "Dismiss" on the first expiry
+// warning and suppresses the final one for the current deadline. A refreshed
+// deadline re-arms both. No-op while the engine is not running.
+func (c *Client) DismissSessionWarning() {
+	cc := c.getConnectClient()
+	if cc == nil {
+		return
+	}
+	engine := cc.Engine()
+	if engine == nil {
+		return
+	}
+	engine.DismissSessionWarning()
+}
+
+// ExtendAuthSession runs the interactive SSO flow to obtain a fresh JWT and
+// asks the management server to extend the session deadline. The tunnel is
+// untouched: no resync, no reconnect. Async; the result arrives on the
+// listener. Mirror of the daemon's RequestExtendAuthSession /
+// WaitExtendAuthSession RPC pair, with URLOpener playing the "UI opens the
+// browser" role.
+//
+// Only one flow may be in flight: the PKCE step binds a fixed loopback port,
+// so a second concurrent flow would fail on that bind. Call
+// CancelExtendAuthSession when the user abandons the browser.
+func (c *Client) ExtendAuthSession(urlOpener URLOpener, isAndroidTV bool, resultListener ErrListener) {
+	ctx, err := c.beginExtend()
+	if err != nil {
+		resultListener.OnError(err)
+		return
+	}
+
+	go func() {
+		defer c.endExtend()
+		if err := c.extendAuthSession(ctx, urlOpener, isAndroidTV); err != nil {
+			resultListener.OnError(err)
+			return
+		}
+		resultListener.OnSuccess()
+	}()
+}
+
+// CancelExtendAuthSession aborts an in-flight ExtendAuthSession. The tunnel is
+// left alone — unlike the login flow, which cancels the whole client context
+// by stopping the engine. Without this the abandoned PKCE wait keeps its
+// loopback port for the full flow timeout and blocks every later attempt.
+// No-op when no flow is running.
+func (c *Client) CancelExtendAuthSession() {
+	c.extendMu.Lock()
+	defer c.extendMu.Unlock()
+	if c.extendCancel != nil {
+		c.extendCancel()
+	}
+}
+
+func (c *Client) stopStateChangeWatchLocked() {
+	// Signal first, unsubscribe second: closing the channels only stops new
+	// items, and the loops would still hand whatever is buffered to a listener
+	// that is no longer registered.
+	if c.stateChangeDone != nil {
+		close(c.stateChangeDone)
+		c.stateChangeDone = nil
+	}
+	if c.stateChangeSubID != "" {
+		c.recorder.UnsubscribeFromStateChanges(c.stateChangeSubID)
+		c.stateChangeSubID = ""
+	}
+	if c.eventSub != nil {
+		// Closes the channel, which ends watchSessionWarnings.
+		c.recorder.UnsubscribeFromEvents(c.eventSub)
+		c.eventSub = nil
+	}
+}
+
+// watchSessionWarnings forwards the engine's session-expiry warnings to the
+// listener. The event stream also carries unrelated traffic — network-map
+// updates on every sync, DNS and route errors — so everything but an
+// AUTHENTICATION event carrying the session-warning marker is dropped. Exits
+// when the subscription is closed by UnsubscribeFromEvents, or earlier when
+// done is closed — the stream buffers up to ten events, and a deregistered
+// listener must not receive the ones already queued.
+func watchSessionWarnings(sub *peer.EventSubscription, listener StateChangeListener, done <-chan struct{}) {
+	for ev := range sub.Events() {
+		select {
+		case <-done:
+			return
+		default:
+		}
+		if ev.GetCategory() != cProto.SystemEvent_AUTHENTICATION {
+			continue
+		}
+		meta := ev.GetMetadata()
+		if meta[sessionwatch.MetaSessionWarning] != "true" {
+			// Other AUTHENTICATION events exist (e.g. a deadline rejected as
+			// out of range); they carry no warning marker.
+			continue
+		}
+		deadline, err := sessionwatch.ParseExpiresAt(meta[sessionwatch.MetaSessionExpiresAt])
+		if err != nil {
+			log.Warnf("session warning event with unparsable deadline: %v", err)
+			continue
+		}
+		lead, err := sessionwatch.ParseLeadMinutes(meta[sessionwatch.MetaSessionLeadMinutes])
+		if err != nil {
+			// Informational only — the deadline above is what drives the UI.
+			lead = 0
+		}
+		listener.OnSessionExpiring(deadline.Unix(), int64(lead),
+			meta[sessionwatch.MetaSessionFinal] == "true")
+	}
+}
+
+func (c *Client) beginExtend() (context.Context, error) {
+	c.extendMu.Lock()
+	defer c.extendMu.Unlock()
+	if c.extendCancel != nil {
+		return nil, fmt.Errorf("session extend already in progress")
+	}
+	ctx, cancel := context.WithCancel(context.Background())
+	c.extendCancel = cancel
+	return ctx, nil
+}
+
+func (c *Client) endExtend() {
+	c.extendMu.Lock()
+	defer c.extendMu.Unlock()
+	if c.extendCancel != nil {
+		c.extendCancel()
+		c.extendCancel = nil
+	}
+}
+
+func (c *Client) extendAuthSession(ctx context.Context, urlOpener URLOpener, isAndroidTV bool) error {
+	cfg, _, cc := c.stateSnapshot()
+	if cfg == nil || cc == nil {
+		return fmt.Errorf("engine is not running")
+	}
+	engine := cc.Engine()
+	if engine == nil {
+		return fmt.Errorf("engine is not initialized")
+	}
+
+	authClient, err := auth.NewAuth(ctx, cfg.PrivateKey, cfg.ManagementURL, cfg)
+	if err != nil {
+		return fmt.Errorf("failed to create auth client: %v", err)
+	}
+	defer authClient.Close()
+
+	a := &Auth{ctx: ctx, config: cfg}
+	tokenInfo, err := a.foregroundGetTokenInfo(authClient, urlOpener, isAndroidTV)
+	if err != nil {
+		return fmt.Errorf("interactive sso login failed: %v", err)
+	}
+
+	if _, err := engine.ExtendAuthSession(ctx, tokenInfo.GetTokenToUse()); err != nil {
+		return err
+	}
+	c.clearLoginRequired()
+
+	go urlOpener.OnLoginSuccess()
+	return nil
+}
--- a/client/server/login_outcome_test.go
+++ b/client/server/login_outcome_test.go
@@ -0,0 +1,89 @@
+package server
+
+import (
+	"context"
+	"encoding/json"
+	"errors"
+	"os"
+	"testing"
+
+	"github.com/stretchr/testify/require"
+	"google.golang.org/grpc/codes"
+	gstatus "google.golang.org/grpc/status"
+
+	"github.com/netbirdio/netbird/client/internal"
+	"github.com/netbirdio/netbird/client/proto"
+)
+
+// A login that never reached Management is not a decision about the peer's
+// credentials, so it must come back as a retryable error rather than an SSO
+// prompt: the user cannot finish a browser login while Management is down, and
+// the CLI's own backoff resolves the outage on its own once the daemon reports
+// the failure. Reproduces `netbird down; netbird up` printing a device-code URL
+// because Management happened to be restarting when the daemon dialed it.
+func TestLogin_ManagementUnreachableIsReturnedInsteadOfDemandingSSO(t *testing.T) {
+	s, _, _, username, _ := setupServerWithProfile(t)
+	s.rootCtx = internal.CtxInitState(context.Background())
+
+	unreachable := errors.New("create connection: dial context: context deadline exceeded")
+	attempts := 0
+	s.loginAttemptFn = func(context.Context, string, string) (internal.StatusType, error) {
+		attempts++
+		return internal.StatusLoginFailed, unreachable
+	}
+
+	resp, err := s.Login(userCtx(), &proto.LoginRequest{Username: &username})
+	require.Error(t, err)
+	require.ErrorIs(t, err, unreachable, "the transport failure was replaced by something else")
+	require.Nil(t, resp, "a failed login must not answer with a login response")
+	require.Equal(t, 1, attempts)
+	require.Nil(t, s.oauthAuthFlow.flow, "the daemon started an SSO flow for a peer whose login was never decided")
+
+	status, err := internal.CtxGetState(s.rootCtx).Status()
+	require.NoError(t, err)
+	require.Equal(t, internal.StatusLoginFailed, status,
+		"a peer that could not reach Management is not waiting on a login")
+}
+
+// The counterpart: Management refusing the peer's credentials is a decision, and
+// the SSO flow still has to start for it. The profile carries an unusable
+// private key so the flow setup fails immediately instead of dialing, which is
+// enough to show the branch was entered — the refusal itself is never what comes
+// back out.
+func TestLogin_AuthRefusalStartsSSOFlow(t *testing.T) {
+	s, _, _, username, cfgPath := setupServerWithProfile(t)
+	s.rootCtx = internal.CtxInitState(context.Background())
+	breakProfilePrivateKey(t, cfgPath)
+
+	refused := gstatus.Error(codes.PermissionDenied, "peer is not registered")
+	s.loginAttemptFn = func(context.Context, string, string) (internal.StatusType, error) {
+		return internal.StatusNeedsLogin, refused
+	}
+
+	_, err := s.Login(userCtx(), &proto.LoginRequest{Username: &username})
+	require.Error(t, err)
+	require.NotErrorIs(t, err, refused,
+		"the refusal was handed back to the caller instead of starting the SSO flow")
+
+	status, stateErr := internal.CtxGetState(s.rootCtx).Status()
+	require.NoError(t, stateErr)
+	require.Equal(t, internal.StatusLoginFailed, status,
+		"the SSO flow setup was never reached with the broken key")
+}
+
+// breakProfilePrivateKey replaces the profile's private key with an unparseable
+// one, which makes any attempt to build a Management client fail on the spot.
+func breakProfilePrivateKey(t *testing.T, cfgPath string) {
+	t.Helper()
+
+	raw, err := os.ReadFile(cfgPath)
+	require.NoError(t, err)
+
+	var cfg map[string]any
+	require.NoError(t, json.Unmarshal(raw, &cfg))
+	cfg["PrivateKey"] = "not-a-key"
+
+	patched, err := json.Marshal(cfg)
+	require.NoError(t, err)
+	require.NoError(t, os.WriteFile(cfgPath, patched, 0o600))
+}
--- a/client/server/server.go
+++ b/client/server/server.go
@@ -132,6 +132,11 @@ type Server struct {
 	updateManager *updater.Manager

 	jwtCache *jwtCache
+
+	// loginAttemptFn stands in for the Management login round trip. Tests set
+	// it to drive the login outcomes that need a server on the other end;
+	// production leaves it nil, and every login goes through loginAttempt.
+	loginAttemptFn func(ctx context.Context, setupKey, jwtToken string) (internal.StatusType, error)
 }

 type oauthAuthFlow struct {
@@ -367,7 +372,19 @@ func (s *Server) connectionGoroutineRunning() bool {
 	}
 }

-// loginAttempt attempts to login using the provided information. it returns a status in case something fails
+// attemptLogin runs a login round trip against Management, or the stand-in a
+// test installed in place of it.
+func (s *Server) attemptLogin(ctx context.Context, setupKey, jwtToken string) (internal.StatusType, error) {
+	if s.loginAttemptFn != nil {
+		return s.loginAttemptFn(ctx, setupKey, jwtToken)
+	}
+	return s.loginAttempt(ctx, setupKey, jwtToken)
+}
+
+// loginAttempt attempts to login using the provided information. It returns
+// StatusNeedsLogin when Management refused the peer's credentials and
+// StatusLoginFailed for every other failure, so callers can tell an
+// authentication decision apart from a login that never got made.
 func (s *Server) loginAttempt(ctx context.Context, setupKey, jwtToken string) (internal.StatusType, error) {
 	authClient, err := auth.NewAuth(ctx, s.config.PrivateKey, s.config.ManagementURL, s.config)
 	if err != nil {
@@ -616,11 +633,23 @@ func (s *Server) Login(callerCtx context.Context, msg *proto.LoginRequest) (*pro
 	s.config = config
 	s.mutex.Unlock()

-	if _, err := s.loginAttempt(ctx, "", ""); err == nil {
+	loginStatus, err := s.attemptLogin(ctx, "", "")
+	if err == nil {
 		state.Set(internal.StatusIdle)
 		return &proto.LoginResponse{}, nil
 	}

+	// Only an authentication refusal means the peer has to (re-)authenticate.
+	// Any other failure leaves the login undecided: Management unreachable, a
+	// restart mid-request, an internal error. Those are returned for the caller
+	// to retry, because turning them into an SSO prompt asks the user to solve
+	// something that is not theirs to solve, and a browser login cannot succeed
+	// while Management is unreachable anyway.
+	if loginStatus != internal.StatusNeedsLogin {
+		state.Set(loginStatus)
+		return nil, err
+	}
+
 	if msg.SetupKey == "" {
 		hint := ""
 		if msg.Hint != nil {
@@ -677,7 +706,7 @@ func (s *Server) Login(callerCtx context.Context, msg *proto.LoginRequest) (*pro
 	// which returns NeedsLogin and parks on the browser leg.
 	state.Set(internal.StatusConnecting)

-	if loginStatus, err := s.loginAttempt(ctx, msg.SetupKey, ""); err != nil {
+	if loginStatus, err := s.attemptLogin(ctx, msg.SetupKey, ""); err != nil {
 		state.Set(loginStatus)
 		return nil, err
 	}
@@ -832,7 +861,7 @@ func (s *Server) WaitSSOLogin(callerCtx context.Context, msg *proto.WaitSSOLogin
 	s.oauthAuthFlow.expiresAt = time.Now()
 	s.mutex.Unlock()

-	if loginStatus, err := s.loginAttempt(ctx, "", tokenInfo.GetTokenToUse()); err != nil {
+	if loginStatus, err := s.attemptLogin(ctx, "", tokenInfo.GetTokenToUse()); err != nil {
 		state.Set(loginStatus)
 		return nil, err
 	}