Merge branch 'main' into embedded-vnc

client/internal/auth/pkce_flow.go

@@ -360,7 +360,13 @@ func isRedirectURLPortUsed(redirectURL string, excludedRanges []excludedPortRang
 		return true
 	}
 
-	addr := fmt.Sprintf(":%s", port)
+	// FreeBSD 15 disables connecting to INADDR_ANY (0.0.0.0) as a localhost
+	// alias by default, ensure explicit ip for localhost.
+	host := parsedURL.Hostname()
+	if host == "" {
+		host = "127.0.0.1"
+	}
+	addr := net.JoinHostPort(host, port)
 	conn, err := net.DialTimeout("tcp", addr, 3*time.Second)
 	if err != nil {
 		return false

client/internal/connect.go

@@ -6,6 +6,7 @@ import (
 	"fmt"
 	"net"
 	"net/netip"
+	"path/filepath"
 	"runtime"
 	"runtime/debug"
 	"strings"
@@ -346,6 +347,11 @@ func (c *ConnectClient) run(mobileDependency MobileDependency, runningChan chan
 			return wrapErr(err)
 		}
 		engineConfig.TempDir = mobileDependency.TempDir
+		// Leave StateDir empty when there is no state path so a disk-backed
+		// syncstore falls back to os.TempDir() instead of filepath.Dir("") == ".".
+		if path != "" {
+			engineConfig.StateDir = filepath.Dir(path)
+		}
 
 		relayManager := relayClient.NewManager(engineCtx, relayURLs, myPrivateKey.PublicKey().String(), engineConfig.MTU)
 		c.statusRecorder.SetRelayMgr(relayManager)

client/internal/debug/debug.go

@@ -254,6 +254,8 @@ type BundleGenerator struct {
 	capturePath    string
 	refreshStatus  func() // Optional callback to refresh status before bundle generation
 	clientMetrics  MetricsExporter
+	daemonVersion  string
+	cliVersion     string
 
 	anonymize         bool
 	includeSystemInfo bool
@@ -278,6 +280,8 @@ type GeneratorDependencies struct {
 	CapturePath    string
 	RefreshStatus  func()
 	ClientMetrics  MetricsExporter
+	DaemonVersion  string
+	CliVersion     string
 }
 
 func NewBundleGenerator(deps GeneratorDependencies, cfg BundleConfig) *BundleGenerator {
@@ -299,6 +303,8 @@ func NewBundleGenerator(deps GeneratorDependencies, cfg BundleConfig) *BundleGen
 		capturePath:    deps.CapturePath,
 		refreshStatus:  deps.RefreshStatus,
 		clientMetrics:  deps.ClientMetrics,
+		daemonVersion:  deps.DaemonVersion,
+		cliVersion:     deps.CliVersion,
 
 		anonymize:         cfg.Anonymize,
 		includeSystemInfo: cfg.IncludeSystemInfo,
@@ -459,9 +465,11 @@ func (g *BundleGenerator) addStatus() error {
 		protoFullStatus := nbstatus.ToProtoFullStatus(fullStatus)
 		protoFullStatus.Events = g.statusRecorder.GetEventHistory()
 		overview := nbstatus.ConvertToStatusOutputOverview(protoFullStatus, nbstatus.ConvertOptions{
-			Anonymize:   g.anonymize,
-			ProfileName: profName,
+			Anonymize:     g.anonymize,
+			ProfileName:   profName,
+			DaemonVersion: g.daemonVersion,
 		})
+		overview.CliVersion = g.cliVersion
 		statusOutput := overview.FullDetailSummary()
 
 		statusReader := strings.NewReader(statusOutput)
@@ -804,6 +812,8 @@ func (g *BundleGenerator) addSyncResponse() error {
 		AllowPartial:    true,
 	}
 
+	g.maskSecrets()
+
 	jsonBytes, err := options.Marshal(g.syncResponse)
 	if err != nil {
 		return fmt.Errorf("generate json: %w", err)
@@ -816,6 +826,27 @@ func (g *BundleGenerator) addSyncResponse() error {
 	return nil
 }
 
+func (g *BundleGenerator) maskSecrets() {
+	if g.syncResponse == nil || g.syncResponse.NetbirdConfig == nil {
+		return
+	}
+
+	if g.syncResponse.NetbirdConfig.Flow != nil {
+		g.syncResponse.NetbirdConfig.Flow.TokenPayload = maskedValue
+
+	}
+
+	if g.syncResponse.NetbirdConfig.Relay != nil {
+		g.syncResponse.NetbirdConfig.Relay.TokenPayload = maskedValue
+	}
+
+	for i := range g.syncResponse.NetbirdConfig.Turns {
+		if g.syncResponse.NetbirdConfig.Turns[i] != nil {
+			g.syncResponse.NetbirdConfig.Turns[i].Password = maskedValue
+		}
+	}
+}
+
 func (g *BundleGenerator) addStateFile() error {
 	sm := profilemanager.NewServiceManager("")
 	path := sm.GetStatePath()
@@ -1045,7 +1076,8 @@ func (g *BundleGenerator) addRotatedLogFiles(logDir string) {
 		return
 	}
 
-	pattern := filepath.Join(logDir, "client-*.log.gz")
+	// This regex will match both logs rotated by us and logrotate on linux
+	pattern := filepath.Join(logDir, "client*.log.*")
 	files, err := filepath.Glob(pattern)
 	if err != nil {
 		log.Warnf("failed to glob rotated logs: %v", err)
@@ -1078,7 +1110,12 @@ func (g *BundleGenerator) addRotatedLogFiles(logDir string) {
 
 	for i := 0; i < maxFiles; i++ {
 		name := filepath.Base(files[i])
-		if err := g.addSingleLogFileGz(files[i], name); err != nil {
+		if strings.HasSuffix(name, ".gz") {
+			err = g.addSingleLogFileGz(files[i], name)
+		} else {
+			err = g.addSingleLogfile(files[i], name)
+		}
+		if err != nil {
 			log.Warnf("failed to add rotated log %s: %v", name, err)
 		}
 	}

client/internal/debug/debug_logfiles_test.go

@@ -0,0 +1,103 @@
+package debug
+
+import (
+	"archive/zip"
+	"bytes"
+	"compress/gzip"
+	"io"
+	"os"
+	"path/filepath"
+	"testing"
+	"time"
+
+	"github.com/stretchr/testify/require"
+)
+
+// TestAddRotatedLogFiles_PicksUpAllVariants asserts that the rotated-log
+// glob picks up logs rotated by timberjack (gzipped) and by logrotate (plain
+// and gzipped), and skips unrelated files.
+func TestAddRotatedLogFiles_PicksUpAllVariants(t *testing.T) {
+	dir := t.TempDir()
+
+	writeFile(t, filepath.Join(dir, "client.log"), "active log\n")
+	writeFile(t, filepath.Join(dir, "other.log"), "unrelated\n")
+
+	timberjackRotated := "client-2026-05-21T10-30-45.000.log.gz"
+	writeGzFile(t, filepath.Join(dir, timberjackRotated), "timberjack rotated content\n")
+
+	logrotatePlain := "client.log.1"
+	writeFile(t, filepath.Join(dir, logrotatePlain), "logrotate plain content\n")
+
+	logrotateGz := "client.log.2.gz"
+	writeGzFile(t, filepath.Join(dir, logrotateGz), "logrotate gz content\n")
+
+	names := runAddRotatedLogFiles(t, dir, 10)
+
+	require.Contains(t, names, timberjackRotated, "timberjack rotated file should be in bundle")
+	require.Contains(t, names, logrotatePlain, "logrotate plain rotated file should be in bundle")
+	require.Contains(t, names, logrotateGz, "logrotate gzipped rotated file should be in bundle")
+	require.NotContains(t, names, "client.log", "active log should not be added by addRotatedLogFiles")
+	require.NotContains(t, names, "other.log", "unrelated files should not be in bundle")
+}
+
+// TestAddRotatedLogFiles_RespectsLogFileCount asserts that only the newest
+// logFileCount rotated files are bundled, ordered by mtime.
+func TestAddRotatedLogFiles_RespectsLogFileCount(t *testing.T) {
+	dir := t.TempDir()
+
+	oldest := filepath.Join(dir, "client.log.3")
+	middle := filepath.Join(dir, "client.log.2")
+	newest := filepath.Join(dir, "client.log.1")
+	writeFile(t, oldest, "old\n")
+	writeFile(t, middle, "mid\n")
+	writeFile(t, newest, "new\n")
+
+	now := time.Now()
+	require.NoError(t, os.Chtimes(oldest, now.Add(-2*time.Hour), now.Add(-2*time.Hour)))
+	require.NoError(t, os.Chtimes(middle, now.Add(-1*time.Hour), now.Add(-1*time.Hour)))
+	require.NoError(t, os.Chtimes(newest, now, now))
+
+	names := runAddRotatedLogFiles(t, dir, 2)
+
+	require.Contains(t, names, "client.log.1")
+	require.Contains(t, names, "client.log.2")
+	require.NotContains(t, names, "client.log.3", "oldest file should be dropped when logFileCount=2")
+}
+
+// runAddRotatedLogFiles calls addRotatedLogFiles against a fresh in-memory
+// zip writer and returns the set of entry names that ended up in the archive.
+func runAddRotatedLogFiles(t *testing.T, dir string, logFileCount uint32) map[string]struct{} {
+	t.Helper()
+
+	var buf bytes.Buffer
+	g := &BundleGenerator{
+		archive:      zip.NewWriter(&buf),
+		logFileCount: logFileCount,
+	}
+	g.addRotatedLogFiles(dir)
+	require.NoError(t, g.archive.Close())
+
+	zr, err := zip.NewReader(bytes.NewReader(buf.Bytes()), int64(buf.Len()))
+	require.NoError(t, err)
+
+	names := make(map[string]struct{}, len(zr.File))
+	for _, f := range zr.File {
+		names[f.Name] = struct{}{}
+	}
+	return names
+}
+
+func writeFile(t *testing.T, path, content string) {
+	t.Helper()
+	require.NoError(t, os.WriteFile(path, []byte(content), 0o644))
+}
+
+func writeGzFile(t *testing.T, path, content string) {
+	t.Helper()
+	var buf bytes.Buffer
+	gw := gzip.NewWriter(&buf)
+	_, err := io.WriteString(gw, content)
+	require.NoError(t, err)
+	require.NoError(t, gw.Close())
+	require.NoError(t, os.WriteFile(path, buf.Bytes(), 0o644))
+}

client/internal/dns/handler_chain.go

@@ -339,8 +339,7 @@ func (c *HandlerChain) isHandlerMatch(qname string, entry HandlerEntry) bool {
 	case entry.Pattern == ".":
 		return true
 	case entry.IsWildcard:
-		parts := strings.Split(strings.TrimSuffix(qname, entry.Pattern), ".")
-		return len(parts) >= 2 && strings.HasSuffix(qname, entry.Pattern)
+		return strings.HasSuffix(qname, "."+entry.Pattern)
 	default:
 		// For non-wildcard patterns:
 		// If handler wants subdomain matching, allow suffix match

client/internal/dns/handler_chain_test.go

@@ -164,6 +164,54 @@ func TestHandlerChain_ServeDNS_DomainMatching(t *testing.T) {
 			matchSubdomains: true,
 			shouldMatch:     true,
 		},
+		{
+			name:            "wildcard label-boundary mismatch (suffix overlap)",
+			handlerDomain:   "*.b.test.",
+			queryDomain:     "x.ab.test.",
+			isWildcard:      true,
+			matchSubdomains: false,
+			shouldMatch:     false,
+		},
+		{
+			name:            "wildcard label-boundary match",
+			handlerDomain:   "*.b.test.",
+			queryDomain:     "x.b.test.",
+			isWildcard:      true,
+			matchSubdomains: false,
+			shouldMatch:     true,
+		},
+		{
+			name:            "wildcard multi-label match",
+			handlerDomain:   "*.b.test.",
+			queryDomain:     "x.y.b.test.",
+			isWildcard:      true,
+			matchSubdomains: false,
+			shouldMatch:     true,
+		},
+		{
+			name:            "wildcard no match on multi-label apex",
+			handlerDomain:   "*.b.test.",
+			queryDomain:     "b.test.",
+			isWildcard:      true,
+			matchSubdomains: false,
+			shouldMatch:     false,
+		},
+		{
+			name:            "wildcard no match on unrelated suffix containment",
+			handlerDomain:   "*.example.com.",
+			queryDomain:     "notexample.com.",
+			isWildcard:      true,
+			matchSubdomains: false,
+			shouldMatch:     false,
+		},
+		{
+			name:            "wildcard accepts pattern registered without trailing dot",
+			handlerDomain:   "*.b.test",
+			queryDomain:     "x.b.test.",
+			isWildcard:      true,
+			matchSubdomains: false,
+			shouldMatch:     true,
+		},
 	}
 
 	for _, tt := range tests {
@@ -273,6 +321,19 @@ func TestHandlerChain_ServeDNS_OverlappingDomains(t *testing.T) {
 			expectedCalls:   1,
 			expectedHandler: 2, // highest priority matching handler should be called
 		},
+		{
+			name: "overlapping wildcard suffixes route to correct handler",
+			handlers: []struct {
+				pattern  string
+				priority int
+			}{
+				{pattern: "*.b.test.", priority: nbdns.PriorityDNSRoute},
+				{pattern: "*.ab.test.", priority: nbdns.PriorityDNSRoute},
+			},
+			queryDomain:     "app.ab.test.",
+			expectedCalls:   1,
+			expectedHandler: 1,
+		},
 		{
 			name: "root zone with specific domain",
 			handlers: []struct {

client/internal/dns/local/local.go

@@ -26,6 +26,19 @@ type resolver interface {
 	LookupNetIP(ctx context.Context, network, host string) ([]netip.Addr, error)
 }
 
+// PeerConnectivity reports whether a tunnel IP belongs to a peer the
+// client knows about and whether that peer is currently connected. The
+// local resolver uses this to suppress A/AAAA answers whose RDATA points
+// at a disconnected peer (typical case: a synthesized private-service
+// record pointing at an embedded proxy peer that just went offline).
+//
+// known=false means the IP isn't in the local peerstore at all — the
+// record is left alone (it points at something outside our mesh, e.g.
+// a non-peer upstream).
+type PeerConnectivity interface {
+	IsConnectedByIP(ip string) (known, connected bool)
+}
+
 type Resolver struct {
 	mu      sync.RWMutex
 	records map[dns.Question][]dns.RR
@@ -33,6 +46,11 @@ type Resolver struct {
 	// zones maps zone domain -> NonAuthoritative (true = non-authoritative, user-created zone)
 	zones    map[domain.Domain]bool
 	resolver resolver
+	// peerConn, when non-nil, is consulted on every A/AAAA answer to
+	// drop records pointing at disconnected peers. nil disables the
+	// filter and preserves the legacy "return whatever is registered"
+	// behaviour for callers that never wire a status source.
+	peerConn PeerConnectivity
 
 	ctx    context.Context
 	cancel context.CancelFunc
@@ -49,6 +67,15 @@ func NewResolver() *Resolver {
 	}
 }
 
+// SetPeerConnectivity wires the per-IP connectivity check used to filter
+// out A/AAAA answers pointing at disconnected peers. Pass nil to disable.
+// Safe to call multiple times; the latest value wins.
+func (d *Resolver) SetPeerConnectivity(p PeerConnectivity) {
+	d.mu.Lock()
+	defer d.mu.Unlock()
+	d.peerConn = p
+}
+
 func (d *Resolver) MatchSubdomains() bool {
 	return true
 }
@@ -95,6 +122,7 @@ func (d *Resolver) ServeDNS(w dns.ResponseWriter, r *dns.Msg) {
 	replyMessage.RecursionAvailable = true
 
 	result := d.lookupRecords(logger, question)
+	result.records = d.filterDisconnectedPeerAnswers(logger, question, result.records)
 	replyMessage.Authoritative = !result.hasExternalData
 	replyMessage.Answer = result.records
 	replyMessage.Rcode = d.determineRcode(question, result)
@@ -436,6 +464,78 @@ func (d *Resolver) logDNSError(logger *log.Entry, hostname string, qtype uint16,
 	}
 }
 
+// filterDisconnectedPeerAnswers drops A/AAAA records whose RDATA matches
+// a known but disconnected peer. The synthesized private-service zones
+// emit one A record per connected proxy peer in a cluster; when a peer
+// goes offline, the server-side refresh removes the record from the
+// next netmap, but the client may still hold the previous netmap for a
+// short window. This filter is the local belt to that braces — even on
+// the stale netmap, the resolver hides the offline target.
+//
+// Records pointing at unknown IPs (outside the local peerstore, e.g.
+// non-mesh upstreams) are never dropped. Non-A/AAAA records pass
+// through untouched.
+//
+// Escape hatch: if filtering would leave the answer empty AND at least
+// one record was filtered, the original list is returned. Better to
+// hand the client a record that may not respond than NXDOMAIN it
+// completely when every proxy peer is offline (the upstream may still
+// be reachable some other way, or the peerstore may be stale).
+func (d *Resolver) filterDisconnectedPeerAnswers(logger *log.Entry, question dns.Question, records []dns.RR) []dns.RR {
+	if len(records) == 0 {
+		return records
+	}
+	d.mu.RLock()
+	checker := d.peerConn
+	d.mu.RUnlock()
+	if checker == nil {
+		return records
+	}
+
+	kept := make([]dns.RR, 0, len(records))
+	var dropped int
+	for _, rr := range records {
+		ip := extractRecordIP(rr)
+		if ip == "" {
+			kept = append(kept, rr)
+			continue
+		}
+		known, connected := checker.IsConnectedByIP(ip)
+		if known && !connected {
+			dropped++
+			continue
+		}
+		kept = append(kept, rr)
+	}
+	if dropped == 0 {
+		return records
+	}
+	if len(kept) == 0 {
+		logger.Debugf("all %d answers for %s point at disconnected peers; returning the original list", dropped, question.Name)
+		return records
+	}
+	logger.Tracef("dropped %d disconnected-peer answer(s) for %s, returning %d", dropped, question.Name, len(kept))
+	return kept
+}
+
+// extractRecordIP returns the dotted-decimal / colon-hex IP carried by
+// an A or AAAA record, or "" for any other record type.
+func extractRecordIP(rr dns.RR) string {
+	switch r := rr.(type) {
+	case *dns.A:
+		if r.A == nil {
+			return ""
+		}
+		return r.A.String()
+	case *dns.AAAA:
+		if r.AAAA == nil {
+			return ""
+		}
+		return r.AAAA.String()
+	}
+	return ""
+}
+
 // Update replaces all zones and their records
 func (d *Resolver) Update(customZones []nbdns.CustomZone) {
 	d.mu.Lock()

client/internal/dns/local/local_test.go

@@ -30,6 +30,21 @@ func (m *mockResolver) LookupNetIP(ctx context.Context, network, host string) ([
 	return nil, nil
 }
 
+// mockPeerConnectivity returns canned (known, connected) results per IP.
+// Used by the disconnected-peer filter tests below. IPs not in the map
+// are reported as unknown so the filter leaves them alone.
+type mockPeerConnectivity struct {
+	byIP map[string]struct{ known, connected bool }
+}
+
+func (m mockPeerConnectivity) IsConnectedByIP(ip string) (known, connected bool) {
+	v, ok := m.byIP[ip]
+	if !ok {
+		return false, false
+	}
+	return v.known, v.connected
+}
+
 func TestLocalResolver_ServeDNS(t *testing.T) {
 	recordA := nbdns.SimpleRecord{
 		Name:  "peera.netbird.cloud.",
@@ -2652,3 +2667,114 @@ func BenchmarkIsInManagedZone_ManyZones(b *testing.B) {
 		resolver.isInManagedZone(qname)
 	}
 }
+
+// TestLocalResolver_FilterDisconnectedPeerAnswers verifies the
+// connectivity-aware filtering layered on top of lookupRecords:
+// when an A record's IP belongs to a known peer that's disconnected,
+// the record is dropped from the answer. Records for unknown IPs pass
+// through. If filtering would empty the answer entirely and at least
+// one record was dropped, the original list is restored (escape hatch
+// for the "all proxies offline" case).
+func TestLocalResolver_FilterDisconnectedPeerAnswers(t *testing.T) {
+	zone := "svc.cluster.netbird."
+	connectedRec := nbdns.SimpleRecord{
+		Name:  zone,
+		Type:  int(dns.TypeA),
+		Class: nbdns.DefaultClass,
+		TTL:   5,
+		RData: "100.64.0.10",
+	}
+	disconnectedRec := nbdns.SimpleRecord{
+		Name:  zone,
+		Type:  int(dns.TypeA),
+		Class: nbdns.DefaultClass,
+		TTL:   5,
+		RData: "100.64.0.11",
+	}
+	unknownRec := nbdns.SimpleRecord{
+		Name:  zone,
+		Type:  int(dns.TypeA),
+		Class: nbdns.DefaultClass,
+		TTL:   5,
+		RData: "203.0.113.5",
+	}
+
+	type ipState struct{ known, connected bool }
+	tests := []struct {
+		name        string
+		records     []nbdns.SimpleRecord
+		connByIP    map[string]ipState
+		wantInOrder []string
+	}{
+		{
+			name:    "drops disconnected peer, keeps connected",
+			records: []nbdns.SimpleRecord{connectedRec, disconnectedRec},
+			connByIP: map[string]ipState{
+				"100.64.0.10": {known: true, connected: true},
+				"100.64.0.11": {known: true, connected: false},
+			},
+			wantInOrder: []string{"100.64.0.10"},
+		},
+		{
+			name:    "unknown IPs pass through untouched",
+			records: []nbdns.SimpleRecord{unknownRec, disconnectedRec},
+			connByIP: map[string]ipState{
+				"100.64.0.11": {known: true, connected: false},
+			},
+			wantInOrder: []string{"203.0.113.5"},
+		},
+		{
+			name:    "all disconnected falls back to original list",
+			records: []nbdns.SimpleRecord{disconnectedRec, connectedRec},
+			connByIP: map[string]ipState{
+				"100.64.0.10": {known: true, connected: false},
+				"100.64.0.11": {known: true, connected: false},
+			},
+			wantInOrder: []string{"100.64.0.11", "100.64.0.10"},
+		},
+		{
+			name:        "no checker wired returns all records",
+			records:     []nbdns.SimpleRecord{connectedRec, disconnectedRec},
+			connByIP:    nil,
+			wantInOrder: []string{"100.64.0.10", "100.64.0.11"},
+		},
+	}
+
+	for _, tc := range tests {
+		t.Run(tc.name, func(t *testing.T) {
+			resolver := NewResolver()
+			if tc.connByIP != nil {
+				cm := mockPeerConnectivity{byIP: make(map[string]struct{ known, connected bool }, len(tc.connByIP))}
+				for ip, st := range tc.connByIP {
+					cm.byIP[ip] = struct{ known, connected bool }{st.known, st.connected}
+				}
+				resolver.SetPeerConnectivity(cm)
+			}
+			resolver.Update([]nbdns.CustomZone{{
+				Domain:           strings.TrimSuffix(zone, "."),
+				Records:          tc.records,
+				NonAuthoritative: true,
+			}})
+
+			var got *dns.Msg
+			writer := &test.MockResponseWriter{
+				WriteMsgFunc: func(m *dns.Msg) error {
+					got = m
+					return nil
+				},
+			}
+			req := new(dns.Msg).SetQuestion(zone, dns.TypeA)
+			resolver.ServeDNS(writer, req)
+
+			require.NotNil(t, got, "resolver must produce a response")
+			require.Len(t, got.Answer, len(tc.wantInOrder),
+				"answer count must match expected: %v", tc.wantInOrder)
+			for i, want := range tc.wantInOrder {
+				a, ok := got.Answer[i].(*dns.A)
+				require.True(t, ok, "answer[%d] must be an A record", i)
+				assert.Equal(t, want, a.A.String(),
+					"answer[%d] expected %s got %s", i, want, a.A.String())
+			}
+		})
+	}
+}

client/internal/dns/server.go

@@ -301,6 +301,11 @@ func newDefaultServer(
 		warningDelayBase:  defaultWarningDelayBase,
 		healthRefresh:     make(chan struct{}, 1),
 	}
+	// Wire the local resolver against the peer status recorder so it can
+	// suppress A/AAAA answers that point at disconnected peers (typical
+	// case: synthesised private-service records pointing at an embedded
+	// proxy peer that just went offline).
+	defaultServer.localResolver.SetPeerConnectivity(localPeerConnectivity{statusRecorder})
 
 	// register with root zone, handler chain takes care of the routing
 	dnsService.RegisterMux(".", handlerChain)
@@ -772,13 +777,24 @@ func (s *DefaultServer) applyHostConfig() {
 // context is released rather than leaked until GC.
 func (s *DefaultServer) registerFallback() {
 	originalNameservers := s.hostManager.getOriginalNameservers()
-	if len(originalNameservers) == 0 {
+
+	serverIP := s.service.RuntimeIP()
+	var servers []netip.AddrPort
+	for _, ns := range originalNameservers {
+		if ns == serverIP {
+			log.Debugf("skipping original nameserver %s as it is the same as the server IP %s", ns, serverIP)
+			continue
+		}
+		servers = append(servers, netip.AddrPortFrom(ns, DefaultPort))
+	}
+
+	if len(servers) == 0 {
 		log.Debugf("no fallback upstreams to register; clearing PriorityFallback handler")
 		s.clearFallback()
 		return
 	}
 
-	log.Infof("registering original nameservers %v as upstream handlers with priority %d", originalNameservers, PriorityFallback)
+	log.Infof("registering original nameservers %v as upstream handlers with priority %d", servers, PriorityFallback)
 
 	handler, err := newUpstreamResolver(
 		s.ctx,
@@ -792,11 +808,6 @@ func (s *DefaultServer) registerFallback() {
 		return
 	}
 	handler.selectedRoutes = s.selectedRoutes
-
-	var servers []netip.AddrPort
-	for _, ns := range originalNameservers {
-		servers = append(servers, netip.AddrPortFrom(ns, DefaultPort))
-	}
 	handler.addRace(servers)
 
 	prev := s.fallbackHandler
@@ -1386,3 +1397,25 @@ func (s *DefaultServer) PopulateManagementDomain(mgmtURL *url.URL) error {
 	}
 	return nil
 }
+
+// localPeerConnectivity adapts *peer.Status to local.PeerConnectivity so
+// the local resolver can ask "is this IP a known peer and is it
+// connected?" without taking on the peer package as a dependency.
+// A nil status recorder always reports known=false so the resolver
+// short-circuits to the legacy "return everything" path.
+type localPeerConnectivity struct {
+	status *peer.Status
+}
+
+// IsConnectedByIP looks the IP up in the peerstore and surfaces both
+// the known and connected bits. Used by Resolver.filterDisconnectedPeerAnswers.
+func (l localPeerConnectivity) IsConnectedByIP(ip string) (known, connected bool) {
+	if l.status == nil {
+		return false, false
+	}
+	state, ok := l.status.PeerStateByIP(ip)
+	if !ok {
+		return false, false
+	}
+	return true, state.ConnStatus == peer.StatusConnected
+}

client/internal/engine.go

@@ -22,7 +22,6 @@ import (
 	log "github.com/sirupsen/logrus"
 	"golang.zx2c4.com/wireguard/tun/netstack"
 	"golang.zx2c4.com/wireguard/wgctrl/wgtypes"
-	"google.golang.org/protobuf/proto"
 
 	nberrors "github.com/netbirdio/netbird/client/errors"
 	"github.com/netbirdio/netbird/client/firewall"
@@ -57,6 +56,7 @@ import (
 	"github.com/netbirdio/netbird/client/internal/routemanager"
 	"github.com/netbirdio/netbird/client/internal/routemanager/systemops"
 	"github.com/netbirdio/netbird/client/internal/statemanager"
+	"github.com/netbirdio/netbird/client/internal/syncstore"
 	"github.com/netbirdio/netbird/client/internal/updater"
 	"github.com/netbirdio/netbird/client/jobexec"
 	cProto "github.com/netbirdio/netbird/client/proto"
@@ -73,6 +73,7 @@ import (
 	sProto "github.com/netbirdio/netbird/shared/signal/proto"
 	"github.com/netbirdio/netbird/util"
 	"github.com/netbirdio/netbird/util/capture"
+	"github.com/netbirdio/netbird/version"
 )
 
 // PeerConnectionTimeoutMax is a timeout of an initial connection attempt to a remote peer.
@@ -151,6 +152,10 @@ type EngineConfig struct {
 
 	LogPath string
 	TempDir string
+
+	// StateDir is the directory holding the state file. The sync response
+	// (network map) is serialized here on platforms that persist it to disk.
+	StateDir string
 }
 
 // EngineServices holds the external service dependencies required by the Engine.
@@ -231,10 +236,15 @@ type Engine struct {
 
 	afpacketCapture *capture.AFPacketCapture
 
-	// Sync response persistence (protected by syncRespMux)
-	syncRespMux         sync.RWMutex
-	persistSyncResponse bool
-	latestSyncResponse  *mgmProto.SyncResponse
+	// Sync response persistence (protected by syncRespMux).
+	// syncStore is nil unless persistence has been enabled; its presence is
+	// what marks persistence as active. The backend (disk or memory) is
+	// selected per-platform; see the syncstore package. syncStoreDir is where
+	// a disk-backed store serializes to.
+	syncRespMux  sync.RWMutex
+	syncStore    syncstore.Store
+	syncStoreDir string
+
 	flowManager         nftypes.FlowManager
 
 	// auto-update
@@ -298,6 +308,7 @@ func NewEngine(
 		jobExecutor:        jobexec.NewExecutor(),
 		clientMetrics:      services.ClientMetrics,
 		updateManager:      services.UpdateManager,
+		syncStoreDir:       config.StateDir,
 	}
 
 	log.Infof("I am: %s", config.WgPrivateKey.PublicKey().String())
@@ -923,19 +934,18 @@ func (e *Engine) handleSync(update *mgmProto.SyncResponse) error {
 	}
 
 	// Persist sync response under the dedicated lock (syncRespMux), not under syncMsgMux.
-	// Read the storage-enabled flag under the syncRespMux too.
+	// A non-nil syncStore is what marks persistence as enabled. Hold the lock for
+	// the whole Set so the store cannot be cleared (disabled / engine close)
+	// mid-call and have this write resurrect a file that was just removed.
 	e.syncRespMux.RLock()
-	enabled := e.persistSyncResponse
-	e.syncRespMux.RUnlock()
-
-	// Store sync response if persistence is enabled
-	if enabled {
-		e.syncRespMux.Lock()
-		e.latestSyncResponse = update
-		e.syncRespMux.Unlock()
-
-		log.Debugf("sync response persisted with serial %d", nm.GetSerial())
+	if e.syncStore != nil {
+		if err := e.syncStore.Set(update); err != nil {
+			log.Errorf("failed to persist sync response: %v", err)
+		} else {
+			log.Debugf("sync response persisted with serial %d", nm.GetSerial())
+		}
 	}
+	e.syncRespMux.RUnlock()
 
 	// only apply new changes and ignore old ones
 	if err := e.updateNetworkMap(nm); err != nil {
@@ -1078,6 +1088,7 @@ func (e *Engine) updateConfig(conf *mgmProto.PeerConfig) error {
 	state.PubKey = e.config.WgPrivateKey.PublicKey().String()
 	state.KernelInterface = !e.wgInterface.IsUserspaceBind()
 	state.FQDN = conf.GetFqdn()
+	state.WgPort = e.config.WgPort
 
 	e.statusRecorder.UpdateLocalPeerState(state)
 
@@ -1156,6 +1167,7 @@ func (e *Engine) handleBundle(params *mgmProto.BundleParameters) (*mgmProto.JobR
 		LogPath:        e.config.LogPath,
 		TempDir:        e.config.TempDir,
 		ClientMetrics:  e.clientMetrics,
+		DaemonVersion:  version.NetbirdVersion(),
 		RefreshStatus: func() {
 			e.RunHealthProbes(true)
 		},
@@ -1834,6 +1846,18 @@ func (e *Engine) close() {
 	if err := e.portForwardManager.GracefullyStop(ctx); err != nil {
 		log.Warnf("failed to gracefully stop port forwarding manager: %s", err)
 	}
+
+	// Drop any persisted sync response so its network map does not linger on
+	// disk after the engine stops (and cannot leak into a later run).
+	e.syncRespMux.Lock()
+	store := e.syncStore
+	e.syncStore = nil
+	e.syncRespMux.Unlock()
+	if store != nil {
+		if err := store.Clear(); err != nil {
+			log.Warnf("failed to clear persisted sync response on close: %v", err)
+		}
+	}
 }
 
 func (e *Engine) readInitialSettings() ([]*route.Route, *nbdns.Config, bool, error) {
@@ -1989,6 +2013,29 @@ func (e *Engine) GetClientMetrics() *metrics.ClientMetrics {
 	return e.clientMetrics
 }
 
+// Performance bundles runtime-adjustable tunnel pool knobs.
+// See Engine.SetPerformance. Nil fields are ignored.
+type Performance struct {
+	PreallocatedBuffersPerPool *uint32
+}
+
+// SetPerformance applies the given tuning to this engine's live Device.
+func (e *Engine) SetPerformance(t Performance) error {
+	e.syncMsgMux.Lock()
+	defer e.syncMsgMux.Unlock()
+	if e.wgInterface == nil {
+		return fmt.Errorf("wg interface not initialized")
+	}
+	dev := e.wgInterface.GetWGDevice()
+	if dev == nil {
+		return fmt.Errorf("wg device not initialized")
+	}
+	if t.PreallocatedBuffersPerPool != nil {
+		dev.SetPreallocatedBuffersPerPool(*t.PreallocatedBuffersPerPool)
+	}
+	return nil
+}
+
 func findIPFromInterfaceName(ifaceName string) (net.IP, error) {
 	iface, err := net.InterfaceByName(ifaceName)
 	if err != nil {
@@ -2141,45 +2188,42 @@ func (e *Engine) stopDNSServer() {
 	e.statusRecorder.UpdateDNSStates(nsGroupStates)
 }
 
-// SetSyncResponsePersistence enables or disables sync response persistence
+// SetSyncResponsePersistence enables or disables sync response persistence.
+// The store is only instantiated while persistence is enabled; construction
+// itself drops any stale data left over from an earlier run (see syncstore).
 func (e *Engine) SetSyncResponsePersistence(enabled bool) {
 	e.syncRespMux.Lock()
 	defer e.syncRespMux.Unlock()
 
-	if enabled == e.persistSyncResponse {
+	if enabled == (e.syncStore != nil) {
 		return
 	}
-	e.persistSyncResponse = enabled
 	log.Debugf("Sync response persistence is set to %t", enabled)
 
 	if !enabled {
-		e.latestSyncResponse = nil
+		if err := e.syncStore.Clear(); err != nil {
+			log.Warnf("failed to clear persisted sync response: %v", err)
+		}
+		e.syncStore = nil
+		return
 	}
+
+	e.syncStore = syncstore.New(e.syncStoreDir)
 }
 
 // GetLatestSyncResponse returns the stored sync response if persistence is enabled
 func (e *Engine) GetLatestSyncResponse() (*mgmProto.SyncResponse, error) {
+	// Hold the lock for the whole Get so the store cannot be cleared
+	// (disabled / engine close) mid-call.
 	e.syncRespMux.RLock()
-	enabled := e.persistSyncResponse
-	latest := e.latestSyncResponse
-	e.syncRespMux.RUnlock()
+	defer e.syncRespMux.RUnlock()
 
-	if !enabled {
+	if e.syncStore == nil {
 		return nil, errors.New("sync response persistence is disabled")
 	}
 
-	if latest == nil {
-		//nolint:nilnil
-		return nil, nil
-	}
-
-	log.Debugf("Retrieving latest sync response with size %d bytes", proto.Size(latest))
-	sr, ok := proto.Clone(latest).(*mgmProto.SyncResponse)
-	if !ok {
-		return nil, fmt.Errorf("failed to clone sync response")
-	}
-
-	return sr, nil
+	//nolint:nilnil
+	return e.syncStore.Get()
 }
 
 // GetWgAddr returns the wireguard address
@@ -2215,7 +2259,7 @@ func (e *Engine) updateDNSForwarder(
 	enabled bool,
 	fwdEntries []*dnsfwd.ForwarderEntry,
 ) {
-	if e.config.DisableServerRoutes {
+	if e.config.DisableServerRoutes || e.config.BlockInbound {
 		return
 	}
 

client/internal/engine_test.go

@@ -27,7 +27,7 @@ import (
 	"github.com/netbirdio/netbird/client/internal/stdnet"
 	"github.com/netbirdio/netbird/management/server/job"
 
-	"github.com/netbirdio/management-integrations/integrations"
+	"github.com/netbirdio/netbird/management/server/integrations/integrated_validator/validator"
 
 	"github.com/netbirdio/netbird/management/internals/controllers/network_map/controller"
 	"github.com/netbirdio/netbird/management/internals/controllers/network_map/update_channel"
@@ -66,8 +66,8 @@ import (
 	"github.com/netbirdio/netbird/route"
 	mgmt "github.com/netbirdio/netbird/shared/management/client"
 	mgmtProto "github.com/netbirdio/netbird/shared/management/proto"
-	relayClient "github.com/netbirdio/netbird/shared/relay/client"
 	"github.com/netbirdio/netbird/shared/netiputil"
+	relayClient "github.com/netbirdio/netbird/shared/relay/client"
 	signal "github.com/netbirdio/netbird/shared/signal/client"
 	"github.com/netbirdio/netbird/shared/signal/proto"
 	signalServer "github.com/netbirdio/netbird/signal/server"
@@ -1641,7 +1641,7 @@ func startManagement(t *testing.T, dataDir, testFile string) (*grpc.Server, stri
 		return nil, "", err
 	}
 
-	ia, _ := integrations.NewIntegratedValidator(context.Background(), peersManager, nil, eventStore, cacheStore)
+	ia, _ := validator.NewIntegratedValidator(context.Background(), peersManager, nil, eventStore, cacheStore)
 
 	metrics, err := telemetry.NewDefaultAppMetrics(context.Background())
 	require.NoError(t, err)

client/internal/lazyconn/support.go

@@ -4,6 +4,8 @@ import (
 	"strings"
 
 	"github.com/hashicorp/go-version"
+
+	nbversion "github.com/netbirdio/netbird/version"
 )
 
 var (
@@ -11,7 +13,7 @@ var (
 )
 
 func IsSupported(agentVersion string) bool {
-	if agentVersion == "development" {
+	if nbversion.IsDevelopmentVersion(agentVersion) {
 		return true
 	}
 

client/internal/networkmonitor/check_change_common.go

@@ -50,7 +50,7 @@ func routeCheck(ctx context.Context, fd int, nexthopv4, nexthopv6 systemops.Next
 		switch msg.Type {
 		// handle route changes
 		case unix.RTM_ADD, syscall.RTM_DELETE:
-			route, err := parseRouteMessage(buf[:n])
+			route, flags, err := parseRouteMessage(buf[:n])
 			if err != nil {
 				log.Debugf("Network monitor: error parsing routing message: %v", err)
 				continue
@@ -66,6 +66,10 @@ func routeCheck(ctx context.Context, fd int, nexthopv4, nexthopv6 systemops.Next
 			}
 			switch msg.Type {
 			case unix.RTM_ADD:
+				if systemops.IgnoreAddedDefaultRoute(flags) {
+					log.Debugf("Network monitor: ignoring added default route via %s, interface %s, flags %#x", route.Gw, intf, flags)
+					continue
+				}
 				log.Infof("Network monitor: default route changed: via %s, interface %s", route.Gw, intf)
 				return nil
 			case unix.RTM_DELETE:
@@ -78,22 +82,26 @@ func routeCheck(ctx context.Context, fd int, nexthopv4, nexthopv6 systemops.Next
 	}
 }
 
-func parseRouteMessage(buf []byte) (*systemops.Route, error) {
+func parseRouteMessage(buf []byte) (*systemops.Route, int, error) {
 	msgs, err := route.ParseRIB(route.RIBTypeRoute, buf)
 	if err != nil {
-		return nil, fmt.Errorf("parse RIB: %v", err)
+		return nil, 0, fmt.Errorf("parse RIB: %v", err)
 	}
 
 	if len(msgs) != 1 {
-		return nil, fmt.Errorf("unexpected RIB message msgs: %v", msgs)
+		return nil, 0, fmt.Errorf("unexpected RIB message msgs: %v", msgs)
 	}
 
 	msg, ok := msgs[0].(*route.RouteMessage)
 	if !ok {
-		return nil, fmt.Errorf("unexpected RIB message type: %T", msgs[0])
+		return nil, 0, fmt.Errorf("unexpected RIB message type: %T", msgs[0])
 	}
 
-	return systemops.MsgToRoute(msg)
+	r, err := systemops.MsgToRoute(msg)
+	if err != nil {
+		return nil, 0, err
+	}
+	return r, msg.Flags, nil
 }
 
 // waitReadable blocks until fd has data to read, or ctx is cancelled.

client/internal/peer/conn.go

@@ -23,6 +23,7 @@ import (
 	"github.com/netbirdio/netbird/client/internal/peer/id"
 	"github.com/netbirdio/netbird/client/internal/peer/worker"
 	"github.com/netbirdio/netbird/client/internal/portforward"
+	"github.com/netbirdio/netbird/client/internal/rosenpass"
 	"github.com/netbirdio/netbird/client/internal/stdnet"
 	"github.com/netbirdio/netbird/route"
 	relayClient "github.com/netbirdio/netbird/shared/relay/client"
@@ -899,7 +900,7 @@ func (conn *Conn) presharedKey(remoteRosenpassKey []byte) *wgtypes.Key {
 	}
 
 	// Fallback to deterministic key if no NetBird PSK is configured
-	determKey, err := conn.rosenpassDetermKey()
+	determKey, err := rosenpass.DeterministicSeedKey(conn.config.LocalKey, conn.config.Key)
 	if err != nil {
 		conn.Log.Errorf("failed to generate Rosenpass initial key: %v", err)
 		return nil
@@ -908,26 +909,6 @@ func (conn *Conn) presharedKey(remoteRosenpassKey []byte) *wgtypes.Key {
 	return determKey
 }
 
-// todo: move this logic into Rosenpass package
-func (conn *Conn) rosenpassDetermKey() (*wgtypes.Key, error) {
-	lk := []byte(conn.config.LocalKey)
-	rk := []byte(conn.config.Key) // remote key
-	var keyInput []byte
-	if string(lk) > string(rk) {
-		//nolint:gocritic
-		keyInput = append(lk[:16], rk[:16]...)
-	} else {
-		//nolint:gocritic
-		keyInput = append(rk[:16], lk[:16]...)
-	}
-
-	key, err := wgtypes.NewKey(keyInput)
-	if err != nil {
-		return nil, err
-	}
-	return &key, nil
-}
-
 func isController(config ConnConfig) bool {
 	return config.LocalKey > config.Key
 }

client/internal/peer/status.go

@@ -111,6 +111,7 @@ type LocalPeerState struct {
 	PubKey          string
 	KernelInterface bool
 	FQDN            string
+	WgPort          int
 	Routes          map[string]struct{}
 }
 
@@ -185,9 +186,12 @@ func (s *StatusChangeSubscription) Events() chan map[string]RouterState {
 	return s.eventsChan
 }
 
-// Status holds a state of peers, signal, management connections and relays
+// Status holds a state of peers, signal, management connections and relays.
+// mux is an RWMutex so hot read paths (notably PeerStateByIP, called for
+// every private-service request) don't contend against each other.
+// Pure read methods take RLock; anything that mutates state takes Lock.
 type Status struct {
-	mux                   sync.Mutex
+	mux                   sync.RWMutex
 	peers                 map[string]State
 	changeNotify          map[string]map[string]*StatusChangeSubscription // map[peerID]map[subscriptionID]*StatusChangeSubscription
 	signalState           bool
@@ -283,8 +287,8 @@ func (d *Status) AddPeer(peerPubKey string, fqdn string, ip string, ipv6 string)
 
 // GetPeer adds peer to Daemon status map
 func (d *Status) GetPeer(peerPubKey string) (State, error) {
-	d.mux.Lock()
-	defer d.mux.Unlock()
+	d.mux.RLock()
+	defer d.mux.RUnlock()
 
 	state, ok := d.peers[peerPubKey]
 	if !ok {
@@ -294,8 +298,8 @@ func (d *Status) GetPeer(peerPubKey string) (State, error) {
 }
 
 func (d *Status) PeerByIP(ip string) (string, bool) {
-	d.mux.Lock()
-	defer d.mux.Unlock()
+	d.mux.RLock()
+	defer d.mux.RUnlock()
 
 	for _, state := range d.peers {
 		if state.IP == ip {
@@ -305,6 +309,34 @@ func (d *Status) PeerByIP(ip string) (string, bool) {
 	return "", false
 }
 
+// PeerStateByIP returns the full peer State for the given tunnel IP.
+// Matches against either the IPv4 (State.IP) or IPv6 (State.IPv6) tunnel
+// address so dual-stack peers are reachable on either family. Searches
+// both d.peers and d.offlinePeers — peers that have been moved into
+// the offline slice by ReplaceOfflinePeers are still part of the
+// account's roster and callers (DNS filter, embed.Client.IdentityForIP)
+// need to recognise them rather than treating them as unknown. Returns
+// the zero State and false when no peer matches or the input is empty.
+func (d *Status) PeerStateByIP(ip string) (State, bool) {
+	if ip == "" {
+		return State{}, false
+	}
+	d.mux.RLock()
+	defer d.mux.RUnlock()
+
+	for _, state := range d.peers {
+		if (state.IP != "" && state.IP == ip) || (state.IPv6 != "" && state.IPv6 == ip) {
+			return state, true
+		}
+	}
+	for _, state := range d.offlinePeers {
+		if (state.IP != "" && state.IP == ip) || (state.IPv6 != "" && state.IPv6 == ip) {
+			return state, true
+		}
+	}
+	return State{}, false
+}
+
 // RemovePeer removes peer from Daemon status map
 func (d *Status) RemovePeer(peerPubKey string) error {
 	d.mux.Lock()
@@ -702,8 +734,8 @@ func (d *Status) UnsubscribePeerStateChanges(subscription *StatusChangeSubscript
 
 // GetLocalPeerState returns the local peer state
 func (d *Status) GetLocalPeerState() LocalPeerState {
-	d.mux.Lock()
-	defer d.mux.Unlock()
+	d.mux.RLock()
+	defer d.mux.RUnlock()
 	return d.localPeer.Clone()
 }
 
@@ -909,8 +941,8 @@ func (d *Status) DeleteResolvedDomainsStates(domain domain.Domain) {
 }
 
 func (d *Status) GetRosenpassState() RosenpassState {
-	d.mux.Lock()
-	defer d.mux.Unlock()
+	d.mux.RLock()
+	defer d.mux.RUnlock()
 	return RosenpassState{
 		d.rosenpassEnabled,
 		d.rosenpassPermissive,
@@ -918,14 +950,14 @@ func (d *Status) GetRosenpassState() RosenpassState {
 }
 
 func (d *Status) GetLazyConnection() bool {
-	d.mux.Lock()
-	defer d.mux.Unlock()
+	d.mux.RLock()
+	defer d.mux.RUnlock()
 	return d.lazyConnectionEnabled
 }
 
 func (d *Status) GetManagementState() ManagementState {
-	d.mux.Lock()
-	defer d.mux.Unlock()
+	d.mux.RLock()
+	defer d.mux.RUnlock()
 	return ManagementState{
 		d.mgmAddress,
 		d.managementState,
@@ -951,8 +983,8 @@ func (d *Status) UpdateLatency(pubKey string, latency time.Duration) error {
 
 // IsLoginRequired determines if a peer's login has expired.
 func (d *Status) IsLoginRequired() bool {
-	d.mux.Lock()
-	defer d.mux.Unlock()
+	d.mux.RLock()
+	defer d.mux.RUnlock()
 
 	// if peer is connected to the management then login is not expired
 	if d.managementState {
@@ -967,8 +999,8 @@ func (d *Status) IsLoginRequired() bool {
 }
 
 func (d *Status) GetSignalState() SignalState {
-	d.mux.Lock()
-	defer d.mux.Unlock()
+	d.mux.RLock()
+	defer d.mux.RUnlock()
 	return SignalState{
 		d.signalAddress,
 		d.signalState,
@@ -978,8 +1010,8 @@ func (d *Status) GetSignalState() SignalState {
 
 // GetRelayStates returns the stun/turn/permanent relay states
 func (d *Status) GetRelayStates() []relay.ProbeResult {
-	d.mux.Lock()
-	defer d.mux.Unlock()
+	d.mux.RLock()
+	defer d.mux.RUnlock()
 	if d.relayMgr == nil {
 		return d.relayStates
 	}
@@ -1008,8 +1040,8 @@ func (d *Status) GetRelayStates() []relay.ProbeResult {
 }
 
 func (d *Status) ForwardingRules() []firewall.ForwardRule {
-	d.mux.Lock()
-	defer d.mux.Unlock()
+	d.mux.RLock()
+	defer d.mux.RUnlock()
 	if d.ingressGwMgr == nil {
 		return nil
 	}
@@ -1018,16 +1050,16 @@ func (d *Status) ForwardingRules() []firewall.ForwardRule {
 }
 
 func (d *Status) GetDNSStates() []NSGroupState {
-	d.mux.Lock()
-	defer d.mux.Unlock()
+	d.mux.RLock()
+	defer d.mux.RUnlock()
 
 	// shallow copy is good enough, as slices fields are currently not updated
 	return slices.Clone(d.nsGroupStates)
 }
 
 func (d *Status) GetResolvedDomainsStates() map[domain.Domain]ResolvedDomainInfo {
-	d.mux.Lock()
-	defer d.mux.Unlock()
+	d.mux.RLock()
+	defer d.mux.RUnlock()
 	return maps.Clone(d.resolvedDomainsStates)
 }
 
@@ -1043,8 +1075,8 @@ func (d *Status) GetFullStatus() FullStatus {
 		LazyConnectionEnabled: d.GetLazyConnection(),
 	}
 
-	d.mux.Lock()
-	defer d.mux.Unlock()
+	d.mux.RLock()
+	defer d.mux.RUnlock()
 
 	fullStatus.LocalPeerState = d.localPeer
 
@@ -1228,8 +1260,8 @@ func (d *Status) SetWgIface(wgInterface WGIfaceStatus) {
 }
 
 func (d *Status) PeersStatus() (*configurer.Stats, error) {
-	d.mux.Lock()
-	defer d.mux.Unlock()
+	d.mux.RLock()
+	defer d.mux.RUnlock()
 	if d.wgIface == nil {
 		return nil, fmt.Errorf("wgInterface is nil, cannot retrieve peers status")
 	}
@@ -1335,6 +1367,7 @@ func (fs FullStatus) ToProto() *proto.FullStatus {
 	pbFullStatus.LocalPeerState.PubKey = fs.LocalPeerState.PubKey
 	pbFullStatus.LocalPeerState.KernelInterface = fs.LocalPeerState.KernelInterface
 	pbFullStatus.LocalPeerState.Fqdn = fs.LocalPeerState.FQDN
+	pbFullStatus.LocalPeerState.WgPort = int32(fs.LocalPeerState.WgPort)
 	pbFullStatus.LocalPeerState.RosenpassPermissive = fs.RosenpassState.Permissive
 	pbFullStatus.LocalPeerState.RosenpassEnabled = fs.RosenpassState.Enabled
 	pbFullStatus.NumberOfForwardingRules = int32(fs.NumOfForwardingRules)

client/internal/peer/status_test.go

@@ -63,6 +63,55 @@ func TestUpdatePeerState(t *testing.T) {
 	assert.Equal(t, ip, state.IP, "ip should be equal")
 }
 
+func TestStatus_PeerStateByIP(t *testing.T) {
+	status := NewRecorder("https://mgm")
+	req := require.New(t)
+
+	req.NoError(status.AddPeer("pk-1", "peer-1.netbird", "100.64.0.10", ""))
+	req.NoError(status.AddPeer("pk-2", "peer-2.netbird", "100.64.0.11", ""))
+
+	state, ok := status.PeerStateByIP("100.64.0.10")
+	req.True(ok, "known tunnel IP should resolve to a peer state")
+	req.Equal("pk-1", state.PubKey, "matching state must carry the right pub key")
+	req.Equal("peer-1.netbird", state.FQDN, "matching state must carry the right FQDN")
+
+	_, ok = status.PeerStateByIP("100.64.0.99")
+	req.False(ok, "unknown IP must report ok=false")
+}
+
+func TestStatus_PeerStateByIP_MatchesIPv6(t *testing.T) {
+	status := NewRecorder("https://mgm")
+	req := require.New(t)
+
+	req.NoError(status.AddPeer("pk-1", "peer-1.netbird", "100.64.0.10", "fd00::1"))
+
+	state, ok := status.PeerStateByIP("fd00::1")
+	req.True(ok, "IPv6-only match must resolve to the peer state")
+	req.Equal("pk-1", state.PubKey, "matching state must carry the right pub key")
+}
+
+// TestStatus_PeerStateByIP_MatchesOfflinePeers covers peers that have
+// been moved into the offline slice via ReplaceOfflinePeers. Callers
+// (DNS filter, embed.Client.IdentityForIP) need to treat them as known
+// rather than unknown — otherwise authentication / DNS filtering treats
+// known-but-offline peers as foreign IPs.
+func TestStatus_PeerStateByIP_MatchesOfflinePeers(t *testing.T) {
+	status := NewRecorder("https://mgm")
+	req := require.New(t)
+
+	status.ReplaceOfflinePeers([]State{
+		{PubKey: "pk-offline", FQDN: "offline.netbird", IP: "100.64.0.20", IPv6: "fd00::20"},
+	})
+
+	state, ok := status.PeerStateByIP("100.64.0.20")
+	req.True(ok, "offline peer must resolve by IPv4 tunnel address")
+	req.Equal("pk-offline", state.PubKey, "matching state must carry the offline peer's pub key")
+
+	state, ok = status.PeerStateByIP("fd00::20")
+	req.True(ok, "offline peer must resolve by IPv6 tunnel address")
+	req.Equal("pk-offline", state.PubKey, "IPv6 match must carry the offline peer's pub key")
+}
+
 func TestStatus_UpdatePeerFQDN(t *testing.T) {
 	key := "abc"
 	fqdn := "peer-a.netbird.local"

client/internal/portforward/pcp/nat.go

@@ -179,8 +179,10 @@ func getDefaultGateway() (gateway net.IP, localIP net.IP, err error) {
 	}
 
 	dst := net.IPv4zero
-	if runtime.GOOS == "linux" {
-		// go-netroute v0.4.0 rejects unspecified destinations client-side on Linux.
+	if runtime.GOOS == "linux" || runtime.GOOS == "android" {
+		// go-netroute v0.4.0 rejects unspecified destinations client-side on Linux/Android.
+		// TODO: on android/ios, use platform APIs (ConnectivityManager.getLinkProperties /
+		// NWPathMonitor) when netlink-based lookup is restricted or unavailable.
 		dst = net.IPv4(0, 0, 0, 1)
 	}
 	_, gateway, localIP, err = router.Route(dst)
@@ -203,7 +205,7 @@ func getDefaultGateway6() (gateway net.IP, localIP net.IP, err error) {
 	}
 
 	dst := net.IPv6zero
-	if runtime.GOOS == "linux" {
+	if runtime.GOOS == "linux" || runtime.GOOS == "android" {
 		// ::2
 		dst = net.IP{0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2}
 	}

client/internal/rosenpass/manager.go

@@ -28,6 +28,15 @@ func hashRosenpassKey(key []byte) string {
 	return hex.EncodeToString(hasher.Sum(nil))
 }
 
+// rpServer is the subset of rp.Server used by Manager. Defined as an interface
+// so tests can substitute a mock without spinning up a real UDP server.
+type rpServer interface {
+	AddPeer(rp.PeerConfig) (rp.PeerID, error)
+	RemovePeer(rp.PeerID) error
+	Run() error
+	Close() error
+}
+
 type Manager struct {
 	ifaceName    string
 	spk          []byte
@@ -36,7 +45,7 @@ type Manager struct {
 	preSharedKey *[32]byte
 	rpPeerIDs    map[string]*rp.PeerID
 	rpWgHandler  *NetbirdHandler
-	server       *rp.Server
+	server       rpServer
 	lock         sync.Mutex
 	port         int
 	wgIface      PresharedKeySetter
@@ -51,7 +60,22 @@ func NewManager(preSharedKey *wgtypes.Key, wgIfaceName string) (*Manager, error)
 
 	rpKeyHash := hashRosenpassKey(public)
 	log.Tracef("generated new rosenpass key pair with public key %s", rpKeyHash)
-	return &Manager{ifaceName: wgIfaceName, rpKeyHash: rpKeyHash, spk: public, ssk: secret, preSharedKey: (*[32]byte)(preSharedKey), rpPeerIDs: make(map[string]*rp.PeerID), lock: sync.Mutex{}}, nil
+	return &Manager{
+		ifaceName:    wgIfaceName,
+		rpKeyHash:    rpKeyHash,
+		spk:          public,
+		ssk:          secret,
+		preSharedKey: (*[32]byte)(preSharedKey),
+		rpPeerIDs:    make(map[string]*rp.PeerID),
+		// rpWgHandler is created here (instead of only in generateConfig) so it
+		// is never nil between NewManager and Run(). Otherwise an early
+		// OnConnected call (race observed on Android, issue #4341) panics on
+		// nil receiver in addPeer -> m.rpWgHandler.AddPeer. generateConfig will
+		// replace it with a fresh handler on each Run() to clear stale peer
+		// state from previous engine sessions.
+		rpWgHandler: NewNetbirdHandler(),
+		lock:        sync.Mutex{},
+	}, nil
 }
 
 func (m *Manager) GetPubKey() []byte {
@@ -65,6 +89,16 @@ func (m *Manager) GetAddress() *net.UDPAddr {
 
 // addPeer adds a new peer to the Rosenpass server
 func (m *Manager) addPeer(rosenpassPubKey []byte, rosenpassAddr string, wireGuardIP string, wireGuardPubKey string) error {
+	// Defense in depth against issue #4341 (Android crash): if Run() has not
+	// completed yet, m.server / m.rpWgHandler may be nil. Return an explicit
+	// error instead of panicking on nil-receiver dereference.
+	if m.server == nil {
+		return fmt.Errorf("rosenpass server not initialized")
+	}
+	if m.rpWgHandler == nil {
+		return fmt.Errorf("rosenpass wg handler not initialized")
+	}
+
 	var err error
 	pcfg := rp.PeerConfig{PublicKey: rosenpassPubKey}
 	if m.preSharedKey != nil {
@@ -79,6 +113,16 @@ func (m *Manager) addPeer(rosenpassPubKey []byte, rosenpassAddr string, wireGuar
 		if pcfg.Endpoint, err = net.ResolveUDPAddr("udp", peerAddr); err != nil {
 			return fmt.Errorf("failed to resolve peer endpoint address: %w", err)
 		}
+		// Our local Rosenpass UDP server binds on the IPv6 wildcard ([::]) — see
+		// GetAddress(). The remote peer's endpoint (pcfg.Endpoint) is the destination
+		// our server will sendto when initiating handshakes. ResolveUDPAddr returns a
+		// 4-byte IPv4 for IPv4 hosts, which the kernel rejects (EDESTADDRREQ) when
+		// sent from an AF_INET6 socket. Normalize the remote endpoint to IPv4-mapped
+		// IPv6 so its address family matches our listening socket.
+		// TODO: maybe bind the Rosenpass UDP server to the peer wg IP addr
+		if v4 := pcfg.Endpoint.IP.To4(); v4 != nil {
+			pcfg.Endpoint.IP = v4.To16()
+		}
 	}
 	peerID, err := m.server.AddPeer(pcfg)
 	if err != nil {
@@ -182,24 +226,31 @@ func (m *Manager) Run() error {
 		return err
 	}
 
-	m.server, err = rp.NewUDPServer(conf)
+	server, err := rp.NewUDPServer(conf)
 	if err != nil {
 		return err
 	}
 
+	m.lock.Lock()
+	m.server = server
+	m.lock.Unlock()
+
 	log.Infof("starting rosenpass server on port %d", m.port)
 
-	return m.server.Run()
+	return server.Run()
 }
 
 // Close closes the Rosenpass server
 func (m *Manager) Close() error {
-	if m.server != nil {
-		err := m.server.Close()
-		if err != nil {
-			log.Errorf("failed closing local rosenpass server")
-		}
-		m.server = nil
+	m.lock.Lock()
+	server := m.server
+	m.server = nil
+	m.lock.Unlock()
+	if server == nil {
+		return nil
+	}
+	if err := server.Close(); err != nil {
+		log.Errorf("failed closing local rosenpass server: %v", err)
 	}
 	return nil
 }

client/internal/rosenpass/manager_test.go

@@ -1,14 +1,412 @@
 package rosenpass
 
 import (
+	"errors"
+	"os"
+	"sync"
 	"testing"
 
+	rp "cunicu.li/go-rosenpass"
 	"github.com/stretchr/testify/require"
+	"golang.zx2c4.com/wireguard/wgctrl/wgtypes"
 )
 
+// --- test doubles -----------------------------------------------------------
+
+type addPeerCall struct {
+	cfg rp.PeerConfig
+}
+
+type removePeerCall struct {
+	id rp.PeerID
+}
+
+type mockServer struct {
+	mu         sync.Mutex
+	addCalls   []addPeerCall
+	removed    []removePeerCall
+	nextID     rp.PeerID
+	addErr     error
+	removeErr  error
+	closed     bool
+	ran        bool
+}
+
+func (m *mockServer) AddPeer(cfg rp.PeerConfig) (rp.PeerID, error) {
+	m.mu.Lock()
+	defer m.mu.Unlock()
+	m.addCalls = append(m.addCalls, addPeerCall{cfg: cfg})
+	if m.addErr != nil {
+		return rp.PeerID{}, m.addErr
+	}
+	// Increment a byte in nextID so distinct peers get distinct IDs.
+	m.nextID[0]++
+	return m.nextID, nil
+}
+
+func (m *mockServer) RemovePeer(id rp.PeerID) error {
+	m.mu.Lock()
+	defer m.mu.Unlock()
+	m.removed = append(m.removed, removePeerCall{id: id})
+	return m.removeErr
+}
+
+func (m *mockServer) Run() error  { m.ran = true; return nil }
+func (m *mockServer) Close() error { m.closed = true; return nil }
+
+type setPSKCall struct {
+	peerKey    string
+	psk        wgtypes.Key
+	updateOnly bool
+}
+
+type mockIface struct {
+	mu    sync.Mutex
+	calls []setPSKCall
+	err   error
+}
+
+func (m *mockIface) SetPresharedKey(peerKey string, psk wgtypes.Key, updateOnly bool) error {
+	m.mu.Lock()
+	defer m.mu.Unlock()
+	m.calls = append(m.calls, setPSKCall{peerKey: peerKey, psk: psk, updateOnly: updateOnly})
+	return m.err
+}
+
+// newTestManager builds a Manager with deterministic spk so tie-break
+// against a peer pubkey is controllable from tests. The provided spk byte
+// becomes the first byte; remaining bytes are zero.
+func newTestManager(spkFirstByte byte, mock *mockServer) *Manager {
+	spk := make([]byte, 32)
+	spk[0] = spkFirstByte
+	return &Manager{
+		ifaceName:   "wt0",
+		spk:         spk,
+		ssk:         make([]byte, 32),
+		rpKeyHash:   "test-hash",
+		rpPeerIDs:   make(map[string]*rp.PeerID),
+		rpWgHandler: NewNetbirdHandler(),
+		server:      mock,
+	}
+}
+
+// validWGKey returns a deterministic 32-byte wireguard public key (base64).
+func validWGKey(t *testing.T, lastByte byte) string {
+	t.Helper()
+	var k wgtypes.Key
+	k[31] = lastByte
+	return k.String()
+}
+
+// --- pure helpers ----------------------------------------------------------
+
+func TestHashRosenpassKey_Deterministic(t *testing.T) {
+	key := []byte("hello-rosenpass")
+	require.Equal(t, hashRosenpassKey(key), hashRosenpassKey(key))
+	require.Len(t, hashRosenpassKey(key), 64) // sha256 hex
+}
+
+func TestHashRosenpassKey_DifferentInputsDifferOutputs(t *testing.T) {
+	require.NotEqual(t, hashRosenpassKey([]byte("a")), hashRosenpassKey([]byte("b")))
+}
+
+func TestGetLogLevel_DefaultWhenUnset(t *testing.T) {
+	// Snapshot + unset to exercise the LookupEnv ok=false branch. t.Setenv
+	// can only set, not delete, so do it manually with restore via t.Cleanup.
+	prev, hadPrev := os.LookupEnv(defaultLogLevelVar)
+	require.NoError(t, os.Unsetenv(defaultLogLevelVar))
+	t.Cleanup(func() {
+		if hadPrev {
+			_ = os.Setenv(defaultLogLevelVar, prev)
+		} else {
+			_ = os.Unsetenv(defaultLogLevelVar)
+		}
+	})
+	require.Equal(t, defaultLog.String(), getLogLevel().String())
+}
+
+func TestGetLogLevel_Cases(t *testing.T) {
+	cases := map[string]string{
+		"debug":   "DEBUG",
+		"info":    "INFO",
+		"warn":    "WARN",
+		"error":   "ERROR",
+		"unknown": "INFO", // default fallback
+	}
+	for input, wantStr := range cases {
+		input, wantStr := input, wantStr
+		t.Run(input, func(t *testing.T) {
+			t.Setenv(defaultLogLevelVar, input)
+			require.Equal(t, wantStr, getLogLevel().String())
+		})
+	}
+}
+
 func TestFindRandomAvailableUDPPort(t *testing.T) {
 	port, err := findRandomAvailableUDPPort()
 	require.NoError(t, err)
 	require.Greater(t, port, 0)
 	require.LessOrEqual(t, port, 65535)
 }
+
+// --- addPeer ---------------------------------------------------------------
+
+func TestAddPeer_HigherLocalPubkey_SetsEndpoint(t *testing.T) {
+	srv := &mockServer{}
+	m := newTestManager(0xFF, srv) // local spk lexicographically larger
+
+	remotePubKey := make([]byte, 32) // remote spk = all zeros (smaller)
+	err := m.addPeer(remotePubKey, "rosenpass-host:7000", "100.1.1.1", validWGKey(t, 1))
+	require.NoError(t, err)
+	require.Len(t, srv.addCalls, 1)
+
+	ep := srv.addCalls[0].cfg.Endpoint
+	require.NotNil(t, ep, "initiator side must set Endpoint")
+	require.Equal(t, 7000, ep.Port)
+	require.Equal(t, "100.1.1.1", ep.IP.String())
+}
+
+func TestAddPeer_HigherLocalPubkey_EndpointIPIsIPv4Mapped(t *testing.T) {
+	// Regression guard for the EDESTADDRREQ fix: Endpoint.IP must be 16-byte
+	// (IPv4-mapped IPv6) so it matches the AF_INET6 listening socket family.
+	srv := &mockServer{}
+	m := newTestManager(0xFF, srv)
+
+	err := m.addPeer(make([]byte, 32), "rp:5000", "100.1.1.1", validWGKey(t, 1))
+	require.NoError(t, err)
+
+	ep := srv.addCalls[0].cfg.Endpoint
+	require.NotNil(t, ep)
+	require.Len(t, ep.IP, 16, "IPv4 endpoint must be normalized to 16-byte v4-mapped form")
+	require.True(t, ep.IP.To4() != nil, "Endpoint must still be detected as IPv4")
+}
+
+func TestAddPeer_LowerLocalPubkey_LeavesEndpointNil(t *testing.T) {
+	srv := &mockServer{}
+	m := newTestManager(0x00, srv) // local spk smaller
+
+	remotePubKey := make([]byte, 32)
+	remotePubKey[0] = 0xFF
+	err := m.addPeer(remotePubKey, "rp:5000", "100.1.1.1", validWGKey(t, 2))
+	require.NoError(t, err)
+
+	require.Nil(t, srv.addCalls[0].cfg.Endpoint, "responder side must NOT set Endpoint")
+}
+
+func TestAddPeer_PresharedKeyPropagated(t *testing.T) {
+	srv := &mockServer{}
+	psk := &wgtypes.Key{0x42}
+	m := newTestManager(0xFF, srv)
+	m.preSharedKey = (*[32]byte)(psk)
+
+	err := m.addPeer(make([]byte, 32), "rp:5000", "100.1.1.1", validWGKey(t, 3))
+	require.NoError(t, err)
+	require.Equal(t, [32]byte(*psk), [32]byte(srv.addCalls[0].cfg.PresharedKey))
+}
+
+func TestAddPeer_InvalidRosenpassAddr_ReturnsError(t *testing.T) {
+	srv := &mockServer{}
+	m := newTestManager(0xFF, srv) // initiator path → parses rosenpassAddr
+
+	err := m.addPeer(make([]byte, 32), "not-a-host-port", "100.1.1.1", validWGKey(t, 1))
+	require.Error(t, err)
+	require.Empty(t, srv.addCalls, "server.AddPeer must not run when address parse fails")
+}
+
+func TestAddPeer_InvalidWireGuardPubKey_ReturnsError(t *testing.T) {
+	srv := &mockServer{}
+	m := newTestManager(0xFF, srv)
+
+	err := m.addPeer(make([]byte, 32), "rp:5000", "100.1.1.1", "not-a-valid-key")
+	require.Error(t, err)
+}
+
+func TestAddPeer_ServerError_Propagates(t *testing.T) {
+	srv := &mockServer{addErr: errors.New("boom")}
+	m := newTestManager(0xFF, srv)
+
+	err := m.addPeer(make([]byte, 32), "rp:5000", "100.1.1.1", validWGKey(t, 1))
+	require.Error(t, err)
+}
+
+// Regression guard for issue #4341 (Android crash). If Run() has not completed
+// before OnConnected fires, m.rpWgHandler or m.server may be nil. Without the
+// nil guards, m.rpWgHandler.AddPeer panics on nil receiver.
+func TestAddPeer_NilHandler_ReturnsErrorNoCrash(t *testing.T) {
+	srv := &mockServer{}
+	m := newTestManager(0xFF, srv)
+	m.rpWgHandler = nil // simulate Run() not yet completed
+
+	err := m.addPeer(make([]byte, 32), "rp:5000", "100.1.1.1", validWGKey(t, 1))
+	require.Error(t, err)
+	require.Contains(t, err.Error(), "wg handler not initialized")
+}
+
+func TestAddPeer_NilServer_ReturnsErrorNoCrash(t *testing.T) {
+	m := newTestManager(0xFF, nil)
+	m.server = nil // simulate Run() not yet completed
+
+	err := m.addPeer(make([]byte, 32), "rp:5000", "100.1.1.1", validWGKey(t, 1))
+	require.Error(t, err)
+	require.Contains(t, err.Error(), "server not initialized")
+}
+
+// NewManager must pre-initialize rpWgHandler so the nil-receiver crash from
+// issue #4341 cannot occur in the window between NewManager and Run().
+func TestNewManager_PreInitializesHandler(t *testing.T) {
+	psk := wgtypes.Key{}
+	m, err := NewManager(&psk, "wt0")
+	require.NoError(t, err)
+	require.NotNil(t, m.rpWgHandler, "rpWgHandler must be initialized in NewManager")
+}
+
+func TestAddPeer_RecordsPeerID(t *testing.T) {
+	srv := &mockServer{}
+	m := newTestManager(0xFF, srv)
+
+	wgKey := validWGKey(t, 5)
+	err := m.addPeer(make([]byte, 32), "rp:5000", "100.1.1.1", wgKey)
+	require.NoError(t, err)
+	require.Contains(t, m.rpPeerIDs, wgKey)
+}
+
+// --- OnConnected / OnDisconnected ------------------------------------------
+
+func TestOnConnected_NilRemotePubKey_NoAddPeer(t *testing.T) {
+	srv := &mockServer{}
+	m := newTestManager(0xFF, srv)
+
+	m.OnConnected(validWGKey(t, 1), nil, "100.1.1.1", "rp:5000")
+	require.Empty(t, srv.addCalls, "nil remote rosenpass pubkey must skip AddPeer")
+	require.Empty(t, m.rpPeerIDs)
+}
+
+func TestOnConnected_ValidPubKey_CallsAddPeer(t *testing.T) {
+	srv := &mockServer{}
+	m := newTestManager(0xFF, srv)
+
+	wgKey := validWGKey(t, 1)
+	m.OnConnected(wgKey, make([]byte, 32), "100.1.1.1", "rp:5000")
+	require.Len(t, srv.addCalls, 1)
+	require.Contains(t, m.rpPeerIDs, wgKey)
+}
+
+func TestOnDisconnected_UnknownPeer_NoOp(t *testing.T) {
+	srv := &mockServer{}
+	m := newTestManager(0xFF, srv)
+
+	m.OnDisconnected(validWGKey(t, 99))
+	require.Empty(t, srv.removed, "unknown peer key must not call RemovePeer")
+}
+
+func TestOnDisconnected_KnownPeer_CallsRemoveAndForgets(t *testing.T) {
+	srv := &mockServer{}
+	m := newTestManager(0xFF, srv)
+
+	wgKey := validWGKey(t, 1)
+	require.NoError(t, m.addPeer(make([]byte, 32), "rp:5000", "100.1.1.1", wgKey))
+	require.Contains(t, m.rpPeerIDs, wgKey)
+
+	m.OnDisconnected(wgKey)
+	require.Len(t, srv.removed, 1)
+	require.NotContains(t, m.rpPeerIDs, wgKey, "peer must be forgotten after disconnect")
+}
+
+// --- IsPresharedKeyInitialized ---------------------------------------------
+
+func TestIsPresharedKeyInitialized_UnknownPeer_ReturnsFalse(t *testing.T) {
+	srv := &mockServer{}
+	m := newTestManager(0xFF, srv)
+	require.False(t, m.IsPresharedKeyInitialized(validWGKey(t, 1)))
+}
+
+func TestIsPresharedKeyInitialized_AddedButNotHandshaken_ReturnsFalse(t *testing.T) {
+	srv := &mockServer{}
+	m := newTestManager(0xFF, srv)
+
+	wgKey := validWGKey(t, 2)
+	require.NoError(t, m.addPeer(make([]byte, 32), "rp:5000", "100.1.1.1", wgKey))
+	require.False(t, m.IsPresharedKeyInitialized(wgKey))
+}
+
+// --- NetbirdHandler.outputKey ----------------------------------------------
+
+func TestHandler_OutputKey_FirstCallUsesUpdateOnlyFalse(t *testing.T) {
+	h := NewNetbirdHandler()
+	iface := &mockIface{}
+	h.SetInterface(iface)
+
+	pid := rp.PeerID{0x01}
+	wgKey := wgtypes.Key{0xAA}
+	h.AddPeer(pid, "wt0", rp.Key(wgKey))
+
+	psk := rp.Key{0xBB}
+	h.HandshakeCompleted(pid, psk)
+
+	require.Len(t, iface.calls, 1)
+	require.False(t, iface.calls[0].updateOnly, "first PSK rotation must use updateOnly=false")
+	require.Equal(t, wgKey.String(), iface.calls[0].peerKey)
+}
+
+func TestHandler_OutputKey_SubsequentCallsUseUpdateOnlyTrue(t *testing.T) {
+	h := NewNetbirdHandler()
+	iface := &mockIface{}
+	h.SetInterface(iface)
+
+	pid := rp.PeerID{0x02}
+	h.AddPeer(pid, "wt0", rp.Key(wgtypes.Key{0xCC}))
+
+	h.HandshakeCompleted(pid, rp.Key{0x01}) // first
+	h.HandshakeCompleted(pid, rp.Key{0x02}) // second
+
+	require.Len(t, iface.calls, 2)
+	require.False(t, iface.calls[0].updateOnly)
+	require.True(t, iface.calls[1].updateOnly, "subsequent rotations must use updateOnly=true")
+}
+
+func TestHandler_OutputKey_NilInterface_NoCrashNoCall(t *testing.T) {
+	h := NewNetbirdHandler()
+	// no SetInterface — iface remains nil
+	pid := rp.PeerID{0x03}
+	h.AddPeer(pid, "wt0", rp.Key(wgtypes.Key{}))
+
+	// Must not panic.
+	h.HandshakeCompleted(pid, rp.Key{})
+}
+
+func TestHandler_OutputKey_UnknownPeer_NoCall(t *testing.T) {
+	h := NewNetbirdHandler()
+	iface := &mockIface{}
+	h.SetInterface(iface)
+
+	h.HandshakeCompleted(rp.PeerID{0xFF}, rp.Key{})
+	require.Empty(t, iface.calls, "unknown peer id must not trigger SetPresharedKey")
+}
+
+func TestHandler_RemovePeer_ClearsInitializedState(t *testing.T) {
+	h := NewNetbirdHandler()
+	iface := &mockIface{}
+	h.SetInterface(iface)
+
+	pid := rp.PeerID{0x04}
+	h.AddPeer(pid, "wt0", rp.Key(wgtypes.Key{0xDD}))
+	h.HandshakeCompleted(pid, rp.Key{0x01})
+	require.True(t, h.IsPeerInitialized(pid))
+
+	h.RemovePeer(pid)
+	require.False(t, h.IsPeerInitialized(pid), "RemovePeer must clear initialized flag")
+}
+
+func TestHandler_SetInterfaceAfterAddPeer_StillReceivesKey(t *testing.T) {
+	h := NewNetbirdHandler()
+	pid := rp.PeerID{0x05}
+	wgKey := wgtypes.Key{0xEE}
+	h.AddPeer(pid, "wt0", rp.Key(wgKey))
+
+	iface := &mockIface{}
+	h.SetInterface(iface) // set after AddPeer
+
+	h.HandshakeCompleted(pid, rp.Key{0x42})
+	require.Len(t, iface.calls, 1)
+	require.Equal(t, wgKey.String(), iface.calls[0].peerKey)
+}

client/internal/rosenpass/seed.go

@@ -0,0 +1,42 @@
+package rosenpass
+
+import (
+	"fmt"
+
+	"golang.zx2c4.com/wireguard/wgctrl/wgtypes"
+)
+
+// DeterministicSeedKey derives a 32-byte WireGuard preshared key from a pair
+// of peer public keys. Both peers, given the same key pair, produce the same
+// output regardless of which side runs the function: the inputs are ordered
+// lexicographically before concatenation.
+//
+// NetBird uses this value as the initial Rosenpass-side preshared key when no
+// explicit account-level PSK is configured, so both peers converge on the same
+// PSK before the first post-quantum handshake completes.
+//
+// The resulting key MUST NOT be treated as quantum-safe: it is deterministic
+// from public keys and exists only to seed WireGuard until Rosenpass rotates
+// in a real post-quantum PSK.
+func DeterministicSeedKey(localKey, remoteKey string) (*wgtypes.Key, error) {
+	lk := []byte(localKey)
+	rk := []byte(remoteKey)
+	if len(lk) < 16 || len(rk) < 16 {
+		return nil, fmt.Errorf("rosenpass: peer keys must be at least 16 bytes (got local=%d, remote=%d)", len(lk), len(rk))
+	}
+
+	var keyInput []byte
+	if localKey > remoteKey {
+		keyInput = append(keyInput, lk[:16]...)
+		keyInput = append(keyInput, rk[:16]...)
+	} else {
+		keyInput = append(keyInput, rk[:16]...)
+		keyInput = append(keyInput, lk[:16]...)
+	}
+
+	key, err := wgtypes.NewKey(keyInput)
+	if err != nil {
+		return nil, fmt.Errorf("rosenpass: deterministic seed key: %w", err)
+	}
+	return &key, nil
+}

client/internal/rosenpass/seed_test.go

@@ -0,0 +1,44 @@
+package rosenpass
+
+import (
+	"strings"
+	"testing"
+
+	"github.com/stretchr/testify/require"
+)
+
+func TestDeterministicSeedKey_SameForBothSides(t *testing.T) {
+	// Peer A and peer B must derive the same PSK regardless of which side
+	// computes it: the function orders inputs internally.
+	a := strings.Repeat("a", 32)
+	b := strings.Repeat("b", 32)
+
+	keyAB, err := DeterministicSeedKey(a, b)
+	require.NoError(t, err)
+	keyBA, err := DeterministicSeedKey(b, a)
+	require.NoError(t, err)
+	require.Equal(t, keyAB.String(), keyBA.String(), "swapping arguments must yield identical key")
+}
+
+func TestDeterministicSeedKey_ChangesWithKeys(t *testing.T) {
+	a := strings.Repeat("a", 32)
+	b := strings.Repeat("b", 32)
+	c := strings.Repeat("c", 32)
+
+	keyAB, err := DeterministicSeedKey(a, b)
+	require.NoError(t, err)
+	keyAC, err := DeterministicSeedKey(a, c)
+	require.NoError(t, err)
+	require.NotEqual(t, keyAB.String(), keyAC.String(), "different peer pair must yield different key")
+}
+
+func TestDeterministicSeedKey_TooShortKey_ReturnsError(t *testing.T) {
+	short := "short" // < 16 bytes
+	long := strings.Repeat("x", 32)
+
+	_, err := DeterministicSeedKey(short, long)
+	require.Error(t, err)
+	_, err = DeterministicSeedKey(long, short)
+	require.Error(t, err)
+}
+

client/internal/routemanager/manager.go

@@ -700,6 +700,13 @@ func resolveURLsToIPs(urls []string) []net.IP {
 
 // updateRouteSelectorFromManagement updates the route selector based on the isSelected status from the management server
 func (m *DefaultManager) updateRouteSelectorFromManagement(clientRoutes route.HAMap) {
+	// An explicit user "deselect all" must not be overridden by management auto-apply.
+	// Auto-applying an exit node here would call SelectRoutes, which clears the
+	// deselect-all flag and re-enables every route the user turned off.
+	if m.routeSelector.IsDeselectAll() {
+		return
+	}
+
 	exitNodeInfo := m.collectExitNodeInfo(clientRoutes)
 	if len(exitNodeInfo.allIDs) == 0 {
 		return

client/internal/routemanager/selector_management_test.go

@@ -0,0 +1,71 @@
+package routemanager
+
+import (
+	"net/netip"
+	"testing"
+
+	"github.com/stretchr/testify/require"
+
+	"github.com/netbirdio/netbird/client/internal/routeselector"
+	"github.com/netbirdio/netbird/route"
+)
+
+func exitNodeRoutes(netID route.NetID, skipAutoApply bool) route.HAMap {
+	haID := route.HAUniqueID(string(netID) + "|0.0.0.0/0")
+	return route.HAMap{
+		haID: []*route.Route{
+			{
+				ID:            "r-" + route.ID(netID),
+				NetID:         netID,
+				Network:       netip.MustParsePrefix("0.0.0.0/0"),
+				NetworkType:   route.IPv4Network,
+				Enabled:       true,
+				SkipAutoApply: skipAutoApply,
+			},
+		},
+	}
+}
+
+func TestUpdateRouteSelectorFromManagement(t *testing.T) {
+	t.Run("management auto-apply selects exit node without user selection", func(t *testing.T) {
+		m := &DefaultManager{routeSelector: routeselector.NewRouteSelector()}
+		routes := exitNodeRoutes("exit1", false)
+
+		m.updateRouteSelectorFromManagement(routes)
+
+		require.True(t, m.routeSelector.IsSelected("exit1"), "auto-apply exit node should be selected")
+		require.Len(t, m.routeSelector.FilterSelectedExitNodes(routes), 1, "selected exit node should pass the filter")
+	})
+
+	t.Run("management SkipAutoApply leaves exit node deselected", func(t *testing.T) {
+		m := &DefaultManager{routeSelector: routeselector.NewRouteSelector()}
+		routes := exitNodeRoutes("exit1", true)
+
+		m.updateRouteSelectorFromManagement(routes)
+
+		require.False(t, m.routeSelector.IsSelected("exit1"), "SkipAutoApply exit node should not be selected")
+		require.Empty(t, m.routeSelector.FilterSelectedExitNodes(routes), "deselected exit node should be filtered out")
+	})
+
+	t.Run("user selection is not overridden by management", func(t *testing.T) {
+		m := &DefaultManager{routeSelector: routeselector.NewRouteSelector()}
+		require.NoError(t, m.routeSelector.SelectRoutes([]route.NetID{"exit1"}, true, []route.NetID{"exit1"}))
+		routes := exitNodeRoutes("exit1", true)
+
+		m.updateRouteSelectorFromManagement(routes)
+
+		require.True(t, m.routeSelector.IsSelected("exit1"), "explicit user selection must survive a management sync that wants to skip auto-apply")
+		require.Len(t, m.routeSelector.FilterSelectedExitNodes(routes), 1, "user-selected exit node should pass the filter")
+	})
+
+	t.Run("deselect-all is preserved across a management sync", func(t *testing.T) {
+		m := &DefaultManager{routeSelector: routeselector.NewRouteSelector()}
+		m.routeSelector.DeselectAllRoutes()
+		routes := exitNodeRoutes("exit1", false)
+
+		m.updateRouteSelectorFromManagement(routes)
+
+		require.True(t, m.routeSelector.IsDeselectAll(), "an explicit deselect-all must not be cleared by management auto-apply")
+		require.Empty(t, m.routeSelector.FilterSelectedExitNodes(routes), "no routes should be selected while deselect-all is set")
+	})
+}

client/internal/routemanager/systemops/routeflags_addfilter_bsd.go

@@ -0,0 +1,9 @@
+//go:build dragonfly || freebsd || netbsd || openbsd
+
+package systemops
+
+// IgnoreAddedDefaultRoute reports whether an RTM_ADD default route with the
+// given flags should be ignored by the network monitor.
+func IgnoreAddedDefaultRoute(flags int) bool {
+	return filterRoutesByFlags(flags)
+}

client/internal/routemanager/systemops/routeflags_addfilter_darwin.go

@@ -0,0 +1,21 @@
+//go:build darwin
+
+package systemops
+
+import "golang.org/x/sys/unix"
+
+// IgnoreAddedDefaultRoute reports whether an RTM_ADD default route with the
+// given flags should be ignored by the network monitor. Scoped routes
+// (RTF_IFSCOPE) are tied to a specific interface index and cannot replace the
+// unscoped default the kernel uses for general egress, so flapping ones (e.g.
+// Wi-Fi calling IMS tunnels on ipsec0, Docker bridges, scoped utun defaults)
+// must not trigger an engine restart.
+func IgnoreAddedDefaultRoute(flags int) bool {
+	if filterRoutesByFlags(flags) {
+		return true
+	}
+	if flags&unix.RTF_IFSCOPE != 0 {
+		return true
+	}
+	return false
+}

client/internal/routeselector/routeselector.go

@@ -116,6 +116,14 @@ func (rs *RouteSelector) DeselectAllRoutes() {
 	clear(rs.selectedRoutes)
 }
 
+// IsDeselectAll reports whether the user has explicitly deselected all routes.
+func (rs *RouteSelector) IsDeselectAll() bool {
+	rs.mu.RLock()
+	defer rs.mu.RUnlock()
+
+	return rs.deselectAll
+}
+
 // IsSelected checks if a specific route is selected.
 func (rs *RouteSelector) IsSelected(routeID route.NetID) bool {
 	rs.mu.RLock()

client/internal/statemanager/manager.go

@@ -104,17 +104,19 @@ func (m *Manager) Stop(ctx context.Context) error {
 	}
 
 	m.mu.Lock()
-	defer m.mu.Unlock()
+	cancel := m.cancel
+	done := m.done
+	m.mu.Unlock()
 
-	if m.cancel == nil {
+	if cancel == nil {
 		return nil
 	}
-	m.cancel()
+	cancel()
 
 	select {
 	case <-ctx.Done():
 		return ctx.Err()
-	case <-m.done:
+	case <-done:
 	}
 
 	return nil

client/internal/syncstore/disk.go

@@ -0,0 +1,99 @@
+package syncstore
+
+import (
+	"context"
+	"errors"
+	"fmt"
+	"os"
+	"path/filepath"
+	"sync"
+
+	log "github.com/sirupsen/logrus"
+	"google.golang.org/protobuf/proto"
+
+	mgmProto "github.com/netbirdio/netbird/shared/management/proto"
+	"github.com/netbirdio/netbird/util"
+)
+
+// syncResponseFileName is the name of the file the sync response is serialized
+// to, placed inside the configured directory (the state directory).
+const syncResponseFileName = "networkmap.pb"
+
+// diskStore serializes the latest sync response to a file on disk instead of
+// keeping it in memory. This trades disk I/O for a much smaller memory
+// footprint, which matters on memory-constrained platforms (iOS).
+type diskStore struct {
+	mu   sync.Mutex
+	path string
+}
+
+// NewDiskStore returns a Store that serializes the sync response to a file in
+// the given directory. If dir is empty it falls back to the OS temp directory.
+//
+// Any file left over from a previous run is removed on construction so a fresh
+// store never reads stale data (e.g. another profile's network map).
+func NewDiskStore(dir string) Store {
+	if dir == "" {
+		dir = os.TempDir()
+	}
+	s := &diskStore{
+		path: filepath.Join(dir, syncResponseFileName),
+	}
+	if err := s.Clear(); err != nil {
+		log.Warnf("failed to clear stale sync response file: %v", err)
+	}
+	return s
+}
+
+func (s *diskStore) Set(resp *mgmProto.SyncResponse) error {
+	if resp == nil {
+		return s.Clear()
+	}
+
+	bs, err := proto.Marshal(resp)
+	if err != nil {
+		return fmt.Errorf("marshal sync response: %w", err)
+	}
+
+	s.mu.Lock()
+	defer s.mu.Unlock()
+
+	if err := util.WriteBytesWithRestrictedPermission(context.Background(), s.path, bs); err != nil {
+		return fmt.Errorf("write sync response to %s: %w", s.path, err)
+	}
+
+	log.Debugf("sync response persisted to %s (%d bytes)", s.path, len(bs))
+	return nil
+}
+
+func (s *diskStore) Get() (*mgmProto.SyncResponse, error) {
+	s.mu.Lock()
+	defer s.mu.Unlock()
+
+	bs, err := os.ReadFile(s.path)
+	if err != nil {
+		if errors.Is(err, os.ErrNotExist) {
+			//nolint:nilnil // nil,nil means "nothing stored", per the Store contract; preserve the original behaviour
+			return nil, nil
+		}
+		return nil, fmt.Errorf("read sync response from %s: %w", s.path, err)
+	}
+
+	resp := &mgmProto.SyncResponse{}
+	if err := proto.Unmarshal(bs, resp); err != nil {
+		return nil, fmt.Errorf("unmarshal sync response: %w", err)
+	}
+
+	log.Debugf("retrieving latest sync response from %s (%d bytes)", s.path, len(bs))
+	return resp, nil
+}
+
+func (s *diskStore) Clear() error {
+	s.mu.Lock()
+	defer s.mu.Unlock()
+
+	if err := os.Remove(s.path); err != nil && !errors.Is(err, os.ErrNotExist) {
+		return fmt.Errorf("remove sync response file %s: %w", s.path, err)
+	}
+	return nil
+}

client/internal/syncstore/factory_ios.go

@@ -0,0 +1,9 @@
+//go:build ios
+
+package syncstore
+
+// New returns the platform default store. On iOS the sync response is
+// serialized to disk (in dir) to keep it out of the constrained process memory.
+func New(dir string) Store {
+	return NewDiskStore(dir)
+}

client/internal/syncstore/factory_other.go

@@ -0,0 +1,9 @@
+//go:build !ios
+
+package syncstore
+
+// New returns the platform default store. On all non-iOS platforms the sync
+// response is kept in memory; dir is unused.
+func New(_ string) Store {
+	return NewMemoryStore()
+}

client/internal/syncstore/memory.go

@@ -0,0 +1,56 @@
+package syncstore
+
+import (
+	"fmt"
+	"sync"
+
+	log "github.com/sirupsen/logrus"
+	"google.golang.org/protobuf/proto"
+
+	mgmProto "github.com/netbirdio/netbird/shared/management/proto"
+)
+
+// memoryStore keeps the latest sync response in memory.
+type memoryStore struct {
+	mu     sync.RWMutex
+	latest *mgmProto.SyncResponse
+}
+
+// NewMemoryStore returns a Store that keeps the sync response in memory.
+func NewMemoryStore() Store {
+	return &memoryStore{}
+}
+
+func (s *memoryStore) Set(resp *mgmProto.SyncResponse) error {
+	s.mu.Lock()
+	defer s.mu.Unlock()
+
+	s.latest = resp
+	return nil
+}
+
+func (s *memoryStore) Get() (*mgmProto.SyncResponse, error) {
+	s.mu.RLock()
+	latest := s.latest
+	s.mu.RUnlock()
+
+	if latest == nil {
+		//nolint:nilnil // nil,nil means "nothing stored", per the Store contract; preserve the original behaviour
+		return nil, nil
+	}
+
+	log.Debugf("retrieving latest sync response with size %d bytes", proto.Size(latest))
+	sr, ok := proto.Clone(latest).(*mgmProto.SyncResponse)
+	if !ok {
+		return nil, fmt.Errorf("clone sync response")
+	}
+	return sr, nil
+}
+
+func (s *memoryStore) Clear() error {
+	s.mu.Lock()
+	defer s.mu.Unlock()
+
+	s.latest = nil
+	return nil
+}

client/internal/syncstore/syncstore.go

@@ -0,0 +1,29 @@
+// Package syncstore stores the latest Management sync response (which carries
+// the network map) for debug bundle generation.
+//
+// The storage backend is selected at build time per operating system: on iOS
+// the response is serialized to disk to keep it out of the (tightly
+// constrained) process memory, while on all other platforms it is kept in
+// memory. The backend is chosen by the New constructor; see factory_ios.go and
+// factory_other.go.
+package syncstore
+
+import (
+	mgmProto "github.com/netbirdio/netbird/shared/management/proto"
+)
+
+// Store persists the latest sync response and returns it on demand.
+//
+// Implementations must be safe for concurrent use.
+type Store interface {
+	// Set stores the given sync response, replacing any previously stored one.
+	Set(resp *mgmProto.SyncResponse) error
+
+	// Get returns the stored sync response, or nil if none is stored.
+	// The returned value is an independent copy that the caller may retain.
+	Get() (*mgmProto.SyncResponse, error)
+
+	// Clear removes any stored sync response. It is safe to call when nothing
+	// is stored.
+	Clear() error
+}

client/internal/updater/manager.go

@@ -19,8 +19,6 @@ import (
 
 const (
 	latestVersion = "latest"
-	// this version will be ignored
-	developmentVersion = "development"
 )
 
 var errNoUpdateState = errors.New("no update state found")
@@ -483,7 +481,7 @@ func (m *Manager) loadAndDeleteUpdateState(ctx context.Context) (*UpdateState, e
 }
 
 func (m *Manager) shouldUpdate(updateVersion *v.Version, forceUpdate bool) bool {
-	if m.currentVersion == developmentVersion {
+	if version.IsDevelopmentVersion(m.currentVersion) {
 		log.Debugf("skipping auto-update, running development version")
 		return false
 	}