Merge branch 'main' into feat/auto-upgrade

client/internal/acl/manager_test.go

@@ -9,6 +9,7 @@ import (
 	"github.com/stretchr/testify/require"
 
 	"github.com/netbirdio/netbird/client/firewall"
+	"github.com/netbirdio/netbird/client/iface"
 	"github.com/netbirdio/netbird/client/iface/wgaddr"
 	"github.com/netbirdio/netbird/client/internal/acl/mocks"
 	"github.com/netbirdio/netbird/client/internal/netflow"
@@ -52,7 +53,7 @@ func TestDefaultManager(t *testing.T) {
 	}).AnyTimes()
 	ifaceMock.EXPECT().GetWGDevice().Return(nil).AnyTimes()
 
-	fw, err := firewall.NewFirewall(ifaceMock, nil, flowLogger, false)
+	fw, err := firewall.NewFirewall(ifaceMock, nil, flowLogger, false, iface.DefaultMTU)
 	require.NoError(t, err)
 	defer func() {
 		err = fw.Close(nil)
@@ -170,7 +171,7 @@ func TestDefaultManagerStateless(t *testing.T) {
 	}).AnyTimes()
 	ifaceMock.EXPECT().GetWGDevice().Return(nil).AnyTimes()
 
-	fw, err := firewall.NewFirewall(ifaceMock, nil, flowLogger, false)
+	fw, err := firewall.NewFirewall(ifaceMock, nil, flowLogger, false, iface.DefaultMTU)
 	require.NoError(t, err)
 	defer func() {
 		err = fw.Close(nil)
@@ -321,7 +322,7 @@ func TestDefaultManagerEnableSSHRules(t *testing.T) {
 	}).AnyTimes()
 	ifaceMock.EXPECT().GetWGDevice().Return(nil).AnyTimes()
 
-	fw, err := firewall.NewFirewall(ifaceMock, nil, flowLogger, false)
+	fw, err := firewall.NewFirewall(ifaceMock, nil, flowLogger, false, iface.DefaultMTU)
 	require.NoError(t, err)
 	defer func() {
 		err = fw.Close(nil)

client/internal/auth/device_flow.go

@@ -128,9 +128,34 @@ func (d *DeviceAuthorizationFlow) RequestAuthInfo(ctx context.Context) (AuthFlow
 		deviceCode.VerificationURIComplete = deviceCode.VerificationURI
 	}
 
+	if d.providerConfig.LoginHint != "" {
+		deviceCode.VerificationURIComplete = appendLoginHint(deviceCode.VerificationURIComplete, d.providerConfig.LoginHint)
+		if deviceCode.VerificationURI != "" {
+			deviceCode.VerificationURI = appendLoginHint(deviceCode.VerificationURI, d.providerConfig.LoginHint)
+		}
+	}
+
 	return deviceCode, err
 }
 
+func appendLoginHint(uri, loginHint string) string {
+	if uri == "" || loginHint == "" {
+		return uri
+	}
+
+	parsedURL, err := url.Parse(uri)
+	if err != nil {
+		log.Debugf("failed to parse verification URI for login_hint: %v", err)
+		return uri
+	}
+
+	query := parsedURL.Query()
+	query.Set("login_hint", loginHint)
+	parsedURL.RawQuery = query.Encode()
+
+	return parsedURL.String()
+}
+
 func (d *DeviceAuthorizationFlow) requestToken(info AuthFlowInfo) (TokenRequestResponse, error) {
 	form := url.Values{}
 	form.Add("client_id", d.providerConfig.ClientID)

client/internal/auth/oauth.go

@@ -66,32 +66,34 @@ func (t TokenInfo) GetTokenToUse() string {
 // and if that also fails, the authentication process is deemed unsuccessful
 //
 // On Linux distros without desktop environment support, it only tries to initialize the Device Code Flow
-func NewOAuthFlow(ctx context.Context, config *profilemanager.Config, isUnixDesktopClient bool) (OAuthFlow, error) {
+func NewOAuthFlow(ctx context.Context, config *profilemanager.Config, isUnixDesktopClient bool, hint string) (OAuthFlow, error) {
 	if (runtime.GOOS == "linux" || runtime.GOOS == "freebsd") && !isUnixDesktopClient {
-		return authenticateWithDeviceCodeFlow(ctx, config)
+		return authenticateWithDeviceCodeFlow(ctx, config, hint)
 	}
 
-	pkceFlow, err := authenticateWithPKCEFlow(ctx, config)
+	pkceFlow, err := authenticateWithPKCEFlow(ctx, config, hint)
 	if err != nil {
-		// fallback to device code flow
 		log.Debugf("failed to initialize pkce authentication with error: %v\n", err)
 		log.Debug("falling back to device code flow")
-		return authenticateWithDeviceCodeFlow(ctx, config)
+		return authenticateWithDeviceCodeFlow(ctx, config, hint)
 	}
 	return pkceFlow, nil
 }
 
 // authenticateWithPKCEFlow initializes the Proof Key for Code Exchange flow auth flow
-func authenticateWithPKCEFlow(ctx context.Context, config *profilemanager.Config) (OAuthFlow, error) {
+func authenticateWithPKCEFlow(ctx context.Context, config *profilemanager.Config, hint string) (OAuthFlow, error) {
 	pkceFlowInfo, err := internal.GetPKCEAuthorizationFlowInfo(ctx, config.PrivateKey, config.ManagementURL, config.ClientCertKeyPair)
 	if err != nil {
 		return nil, fmt.Errorf("getting pkce authorization flow info failed with error: %v", err)
 	}
+
+	pkceFlowInfo.ProviderConfig.LoginHint = hint
+
 	return NewPKCEAuthorizationFlow(pkceFlowInfo.ProviderConfig)
 }
 
 // authenticateWithDeviceCodeFlow initializes the Device Code auth Flow
-func authenticateWithDeviceCodeFlow(ctx context.Context, config *profilemanager.Config) (OAuthFlow, error) {
+func authenticateWithDeviceCodeFlow(ctx context.Context, config *profilemanager.Config, hint string) (OAuthFlow, error) {
 	deviceFlowInfo, err := internal.GetDeviceAuthorizationFlowInfo(ctx, config.PrivateKey, config.ManagementURL)
 	if err != nil {
 		switch s, ok := gstatus.FromError(err); {
@@ -107,5 +109,7 @@ func authenticateWithDeviceCodeFlow(ctx context.Context, config *profilemanager.
 		}
 	}
 
+	deviceFlowInfo.ProviderConfig.LoginHint = hint
+
 	return NewDeviceAuthorizationFlow(deviceFlowInfo.ProviderConfig)
 }

client/internal/auth/pkce_flow.go

@@ -109,6 +109,9 @@ func (p *PKCEAuthorizationFlow) RequestAuthInfo(ctx context.Context) (AuthFlowIn
 			params = append(params, oauth2.SetAuthURLParam("max_age", "0"))
 		}
 	}
+	if p.providerConfig.LoginHint != "" {
+		params = append(params, oauth2.SetAuthURLParam("login_hint", p.providerConfig.LoginHint))
+	}
 
 	authURL := p.oAuthConfig.AuthCodeURL(state, params...)
 

client/internal/connect.go

@@ -293,15 +293,18 @@ func (c *ConnectClient) run(mobileDependency MobileDependency, runningChan chan
 		}
 
 		<-engineCtx.Done()
+
 		c.engineMutex.Lock()
-		if c.engine != nil && c.engine.wgInterface != nil {
-			log.Infof("ensuring %s is removed, Netbird engine context cancelled", c.engine.wgInterface.Name())
-			if err := c.engine.Stop(); err != nil {
+		engine := c.engine
+		c.engine = nil
+		c.engineMutex.Unlock()
+
+		if engine != nil && engine.wgInterface != nil {
+			log.Infof("ensuring %s is removed, Netbird engine context cancelled", engine.wgInterface.Name())
+			if err := engine.Stop(); err != nil {
 				log.Errorf("Failed to stop engine: %v", err)
 			}
-			c.engine = nil
 		}
-		c.engineMutex.Unlock()
 		c.statusRecorder.ClientTeardown()
 
 		backOff.Reset()
@@ -386,19 +389,12 @@ func (c *ConnectClient) Status() StatusType {
 }
 
 func (c *ConnectClient) Stop() error {
-	if c == nil {
-		return nil
+	engine := c.Engine()
+	if engine != nil {
+		if err := engine.Stop(); err != nil {
+			return fmt.Errorf("stop engine: %w", err)
+		}
 	}
-	c.engineMutex.Lock()
-	defer c.engineMutex.Unlock()
-
-	if c.engine == nil {
-		return nil
-	}
-	if err := c.engine.Stop(); err != nil {
-		return fmt.Errorf("stop engine: %w", err)
-	}
-
 	return nil
 }
 

client/internal/debug/debug.go

@@ -44,6 +44,8 @@ interfaces.txt: Anonymized network interface information, if --system-info flag
 ip_rules.txt: Detailed IP routing rules in tabular format including priority, source, destination, interfaces, table, and action information (Linux only), if --system-info flag was provided.
 iptables.txt: Anonymized iptables rules with packet counters, if --system-info flag was provided.
 nftables.txt: Anonymized nftables rules with packet counters, if --system-info flag was provided.
+resolv.conf: DNS resolver configuration from /etc/resolv.conf (Unix systems only), if --system-info flag was provided.
+scutil_dns.txt: DNS configuration from scutil --dns (macOS only), if --system-info flag was provided.
 resolved_domains.txt: Anonymized resolved domain IP addresses from the status recorder.
 config.txt: Anonymized configuration information of the NetBird client.
 network_map.json: Anonymized sync response containing peer configurations, routes, DNS settings, and firewall rules.
@@ -184,6 +186,20 @@ The ip_rules.txt file contains detailed IP routing rule information:
 The table format provides comprehensive visibility into the IP routing decision process, including how traffic is directed to different routing tables based on various criteria. This is valuable for troubleshooting advanced routing configurations and policy-based routing.
 
 For anonymized rules, IP addresses and prefixes are replaced as described above. Interface names are anonymized using string anonymization. Table names, actions, and other non-sensitive information remain unchanged.
+
+DNS Configuration
+The debug bundle includes platform-specific DNS configuration files:
+
+resolv.conf (Unix systems):
+- Contains DNS resolver configuration from /etc/resolv.conf
+- Includes nameserver entries, search domains, and resolver options
+- All IP addresses and domain names are anonymized following the same rules as other files
+
+scutil_dns.txt (macOS only):
+- Contains detailed DNS configuration from scutil --dns
+- Shows DNS configuration for all network interfaces
+- Includes search domains, nameservers, and DNS resolver settings
+- All IP addresses and domain names are anonymized
 `
 
 const (
@@ -357,6 +373,10 @@ func (g *BundleGenerator) addSystemInfo() {
 	if err := g.addFirewallRules(); err != nil {
 		log.Errorf("failed to add firewall rules to debug bundle: %v", err)
 	}
+
+	if err := g.addDNSInfo(); err != nil {
+		log.Errorf("failed to add DNS info to debug bundle: %v", err)
+	}
 }
 
 func (g *BundleGenerator) addReadme() error {

client/internal/debug/debug_darwin.go

@@ -0,0 +1,53 @@
+//go:build darwin && !ios
+
+package debug
+
+import (
+	"bytes"
+	"context"
+	"fmt"
+	"os/exec"
+	"strings"
+	"time"
+
+	log "github.com/sirupsen/logrus"
+)
+
+// addDNSInfo collects and adds DNS configuration information to the archive
+func (g *BundleGenerator) addDNSInfo() error {
+	if err := g.addResolvConf(); err != nil {
+		log.Errorf("failed to add resolv.conf: %v", err)
+	}
+
+	if err := g.addScutilDNS(); err != nil {
+		log.Errorf("failed to add scutil DNS output: %v", err)
+	}
+
+	return nil
+}
+
+func (g *BundleGenerator) addScutilDNS() error {
+	ctx, cancel := context.WithTimeout(context.Background(), 10*time.Second)
+	defer cancel()
+
+	cmd := exec.CommandContext(ctx, "scutil", "--dns")
+	output, err := cmd.CombinedOutput()
+	if err != nil {
+		return fmt.Errorf("execute scutil --dns: %w", err)
+	}
+
+	if len(bytes.TrimSpace(output)) == 0 {
+		return fmt.Errorf("no scutil DNS output")
+	}
+
+	content := string(output)
+	if g.anonymize {
+		content = g.anonymizer.AnonymizeString(content)
+	}
+
+	if err := g.addFileToZip(strings.NewReader(content), "scutil_dns.txt"); err != nil {
+		return fmt.Errorf("add scutil DNS output to zip: %w", err)
+	}
+
+	return nil
+}

client/internal/debug/debug_mobile.go

@@ -5,3 +5,7 @@ package debug
 func (g *BundleGenerator) addRoutes() error {
 	return nil
 }
+
+func (g *BundleGenerator) addDNSInfo() error {
+	return nil
+}

client/internal/debug/debug_nondarwin.go

@@ -0,0 +1,16 @@
+//go:build unix && !darwin && !android
+
+package debug
+
+import (
+	log "github.com/sirupsen/logrus"
+)
+
+// addDNSInfo collects and adds DNS configuration information to the archive
+func (g *BundleGenerator) addDNSInfo() error {
+	if err := g.addResolvConf(); err != nil {
+		log.Errorf("failed to add resolv.conf: %v", err)
+	}
+
+	return nil
+}

client/internal/debug/debug_nonunix.go

@@ -0,0 +1,7 @@
+//go:build !unix
+
+package debug
+
+func (g *BundleGenerator) addDNSInfo() error {
+	return nil
+}

client/internal/debug/debug_unix.go

@@ -0,0 +1,29 @@
+//go:build unix && !android
+
+package debug
+
+import (
+	"fmt"
+	"os"
+	"strings"
+)
+
+const resolvConfPath = "/etc/resolv.conf"
+
+func (g *BundleGenerator) addResolvConf() error {
+	data, err := os.ReadFile(resolvConfPath)
+	if err != nil {
+		return fmt.Errorf("read %s: %w", resolvConfPath, err)
+	}
+
+	content := string(data)
+	if g.anonymize {
+		content = g.anonymizer.AnonymizeString(content)
+	}
+
+	if err := g.addFileToZip(strings.NewReader(content), "resolv.conf"); err != nil {
+		return fmt.Errorf("add resolv.conf to zip: %w", err)
+	}
+
+	return nil
+}

client/internal/device_auth.go

@@ -38,6 +38,8 @@ type DeviceAuthProviderConfig struct {
 	Scope string
 	// UseIDToken indicates if the id token should be used for authentication
 	UseIDToken bool
+	// LoginHint is used to pre-fill the email/username field during authentication
+	LoginHint string
 }
 
 // GetDeviceAuthorizationFlowInfo initialize a DeviceAuthorizationFlow instance and return with it

client/internal/dns/host_darwin.go

@@ -13,6 +13,7 @@ import (
 	"strings"
 
 	log "github.com/sirupsen/logrus"
+	"golang.org/x/exp/maps"
 
 	"github.com/netbirdio/netbird/client/internal/statemanager"
 )
@@ -50,28 +51,21 @@ func (s *systemConfigurator) supportCustomPort() bool {
 }
 
 func (s *systemConfigurator) applyDNSConfig(config HostDNSConfig, stateManager *statemanager.Manager) error {
-	var err error
-
-	if err := stateManager.UpdateState(&ShutdownState{}); err != nil {
-		log.Errorf("failed to update shutdown state: %s", err)
-	}
-
 	var (
 		searchDomains []string
 		matchDomains  []string
 	)
 
-	err = s.recordSystemDNSSettings(true)
-	if err != nil {
+	if err := s.recordSystemDNSSettings(true); err != nil {
 		log.Errorf("unable to update record of System's DNS config: %s", err.Error())
 	}
 
 	if config.RouteAll {
 		searchDomains = append(searchDomains, "\"\"")
-		err = s.addLocalDNS()
-		if err != nil {
-			log.Infof("failed to enable split DNS")
+		if err := s.addLocalDNS(); err != nil {
+			log.Warnf("failed to add local DNS: %v", err)
 		}
+		s.updateState(stateManager)
 	}
 
 	for _, dConf := range config.Domains {
@@ -86,6 +80,7 @@ func (s *systemConfigurator) applyDNSConfig(config HostDNSConfig, stateManager *
 	}
 
 	matchKey := getKeyWithInput(netbirdDNSStateKeyFormat, matchSuffix)
+	var err error
 	if len(matchDomains) != 0 {
 		err = s.addMatchDomains(matchKey, strings.Join(matchDomains, " "), config.ServerIP, config.ServerPort)
 	} else {
@@ -95,6 +90,7 @@ func (s *systemConfigurator) applyDNSConfig(config HostDNSConfig, stateManager *
 	if err != nil {
 		return fmt.Errorf("add match domains: %w", err)
 	}
+	s.updateState(stateManager)
 
 	searchKey := getKeyWithInput(netbirdDNSStateKeyFormat, searchSuffix)
 	if len(searchDomains) != 0 {
@@ -106,6 +102,7 @@ func (s *systemConfigurator) applyDNSConfig(config HostDNSConfig, stateManager *
 	if err != nil {
 		return fmt.Errorf("add search domains: %w", err)
 	}
+	s.updateState(stateManager)
 
 	if err := s.flushDNSCache(); err != nil {
 		log.Errorf("failed to flush DNS cache: %v", err)
@@ -114,6 +111,12 @@ func (s *systemConfigurator) applyDNSConfig(config HostDNSConfig, stateManager *
 	return nil
 }
 
+func (s *systemConfigurator) updateState(stateManager *statemanager.Manager) {
+	if err := stateManager.UpdateState(&ShutdownState{CreatedKeys: maps.Keys(s.createdKeys)}); err != nil {
+		log.Errorf("failed to update shutdown state: %s", err)
+	}
+}
+
 func (s *systemConfigurator) string() string {
 	return "scutil"
 }
@@ -167,18 +170,20 @@ func (s *systemConfigurator) removeKeyFromSystemConfig(key string) error {
 func (s *systemConfigurator) addLocalDNS() error {
 	if !s.systemDNSSettings.ServerIP.IsValid() || len(s.systemDNSSettings.Domains) == 0 {
 		if err := s.recordSystemDNSSettings(true); err != nil {
-			log.Errorf("Unable to get system DNS configuration")
 			return fmt.Errorf("recordSystemDNSSettings(): %w", err)
 		}
 	}
 	localKey := getKeyWithInput(netbirdDNSStateKeyFormat, localSuffix)
-	if s.systemDNSSettings.ServerIP.IsValid() && len(s.systemDNSSettings.Domains) != 0 {
-		err := s.addSearchDomains(localKey, strings.Join(s.systemDNSSettings.Domains, " "), s.systemDNSSettings.ServerIP, s.systemDNSSettings.ServerPort)
-		if err != nil {
-			return fmt.Errorf("couldn't add local network DNS conf: %w", err)
-		}
-	} else {
+	if !s.systemDNSSettings.ServerIP.IsValid() || len(s.systemDNSSettings.Domains) == 0 {
 		log.Info("Not enabling local DNS server")
+		return nil
+	}
+
+	if err := s.addSearchDomains(
+		localKey,
+		strings.Join(s.systemDNSSettings.Domains, " "), s.systemDNSSettings.ServerIP, s.systemDNSSettings.ServerPort,
+	); err != nil {
+		return fmt.Errorf("add search domains: %w", err)
 	}
 
 	return nil

client/internal/dns/host_darwin_test.go

@@ -0,0 +1,111 @@
+//go:build !ios
+
+package dns
+
+import (
+	"context"
+	"net/netip"
+	"os/exec"
+	"path/filepath"
+	"strings"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+
+	"github.com/netbirdio/netbird/client/internal/statemanager"
+)
+
+func TestDarwinDNSUncleanShutdownCleanup(t *testing.T) {
+	if testing.Short() {
+		t.Skip("skipping scutil integration test in short mode")
+	}
+
+	tmpDir := t.TempDir()
+	stateFile := filepath.Join(tmpDir, "state.json")
+
+	sm := statemanager.New(stateFile)
+	sm.RegisterState(&ShutdownState{})
+	sm.Start()
+	defer func() {
+		require.NoError(t, sm.Stop(context.Background()))
+	}()
+
+	configurator := &systemConfigurator{
+		createdKeys: make(map[string]struct{}),
+	}
+
+	config := HostDNSConfig{
+		ServerIP:   netip.MustParseAddr("100.64.0.1"),
+		ServerPort: 53,
+		RouteAll:   true,
+		Domains: []DomainConfig{
+			{Domain: "example.com", MatchOnly: true},
+		},
+	}
+
+	err := configurator.applyDNSConfig(config, sm)
+	require.NoError(t, err)
+
+	require.NoError(t, sm.PersistState(context.Background()))
+
+	searchKey := getKeyWithInput(netbirdDNSStateKeyFormat, searchSuffix)
+	matchKey := getKeyWithInput(netbirdDNSStateKeyFormat, matchSuffix)
+	localKey := getKeyWithInput(netbirdDNSStateKeyFormat, localSuffix)
+
+	defer func() {
+		for _, key := range []string{searchKey, matchKey, localKey} {
+			_ = removeTestDNSKey(key)
+		}
+	}()
+
+	for _, key := range []string{searchKey, matchKey, localKey} {
+		exists, err := checkDNSKeyExists(key)
+		require.NoError(t, err)
+		if exists {
+			t.Logf("Key %s exists before cleanup", key)
+		}
+	}
+
+	sm2 := statemanager.New(stateFile)
+	sm2.RegisterState(&ShutdownState{})
+	err = sm2.LoadState(&ShutdownState{})
+	require.NoError(t, err)
+
+	state := sm2.GetState(&ShutdownState{})
+	if state == nil {
+		t.Skip("State not saved, skipping cleanup test")
+	}
+
+	shutdownState, ok := state.(*ShutdownState)
+	require.True(t, ok)
+
+	err = shutdownState.Cleanup()
+	require.NoError(t, err)
+
+	for _, key := range []string{searchKey, matchKey, localKey} {
+		exists, err := checkDNSKeyExists(key)
+		require.NoError(t, err)
+		assert.False(t, exists, "Key %s should NOT exist after cleanup", key)
+	}
+}
+
+func checkDNSKeyExists(key string) (bool, error) {
+	cmd := exec.Command(scutilPath)
+	cmd.Stdin = strings.NewReader("show " + key + "\nquit\n")
+	output, err := cmd.CombinedOutput()
+	if err != nil {
+		if strings.Contains(string(output), "No such key") {
+			return false, nil
+		}
+		return false, err
+	}
+	return !strings.Contains(string(output), "No such key"), nil
+}
+
+func removeTestDNSKey(key string) error {
+	cmd := exec.Command(scutilPath)
+	cmd.Stdin = strings.NewReader("remove " + key + "\nquit\n")
+	_, err := cmd.CombinedOutput()
+	return err
+}

client/internal/dns/host_windows.go

@@ -179,13 +179,7 @@ func (r *registryConfigurator) applyDNSConfig(config HostDNSConfig, stateManager
 		log.Infof("removed %s as main DNS forwarder for this peer", config.ServerIP)
 	}
 
-	if err := stateManager.UpdateState(&ShutdownState{
-		Guid:           r.guid,
-		GPO:            r.gpo,
-		NRPTEntryCount: r.nrptEntryCount,
-	}); err != nil {
-		log.Errorf("failed to update shutdown state: %s", err)
-	}
+	r.updateState(stateManager)
 
 	var searchDomains, matchDomains []string
 	for _, dConf := range config.Domains {
@@ -212,13 +206,7 @@ func (r *registryConfigurator) applyDNSConfig(config HostDNSConfig, stateManager
 		r.nrptEntryCount = 0
 	}
 
-	if err := stateManager.UpdateState(&ShutdownState{
-		Guid:           r.guid,
-		GPO:            r.gpo,
-		NRPTEntryCount: r.nrptEntryCount,
-	}); err != nil {
-		log.Errorf("failed to update shutdown state: %s", err)
-	}
+	r.updateState(stateManager)
 
 	if err := r.updateSearchDomains(searchDomains); err != nil {
 		return fmt.Errorf("update search domains: %w", err)
@@ -229,6 +217,16 @@ func (r *registryConfigurator) applyDNSConfig(config HostDNSConfig, stateManager
 	return nil
 }
 
+func (r *registryConfigurator) updateState(stateManager *statemanager.Manager) {
+	if err := stateManager.UpdateState(&ShutdownState{
+		Guid:           r.guid,
+		GPO:            r.gpo,
+		NRPTEntryCount: r.nrptEntryCount,
+	}); err != nil {
+		log.Errorf("failed to update shutdown state: %s", err)
+	}
+}
+
 func (r *registryConfigurator) addDNSSetupForAll(ip netip.Addr) error {
 	if err := r.setInterfaceRegistryKeyStringValue(interfaceConfigNameServerKey, ip.String()); err != nil {
 		return fmt.Errorf("adding dns setup for all failed: %w", err)

client/internal/dns/server.go

@@ -65,8 +65,9 @@ type hostManagerWithOriginalNS interface {
 
 // DefaultServer dns server object
 type DefaultServer struct {
-	ctx       context.Context
-	ctxCancel context.CancelFunc
+	ctx        context.Context
+	ctxCancel  context.CancelFunc
+	shutdownWg sync.WaitGroup
 	// disableSys disables system DNS management (e.g., /etc/resolv.conf updates) while keeping the DNS service running.
 	// This is different from ServiceEnable=false from management which completely disables the DNS service.
 	disableSys         bool
@@ -318,6 +319,7 @@ func (s *DefaultServer) DnsIP() netip.Addr {
 // Stop stops the server
 func (s *DefaultServer) Stop() {
 	s.ctxCancel()
+	s.shutdownWg.Wait()
 
 	s.mux.Lock()
 	defer s.mux.Unlock()
@@ -507,8 +509,9 @@ func (s *DefaultServer) applyConfiguration(update nbdns.Config) error {
 
 	s.applyHostConfig()
 
+	s.shutdownWg.Add(1)
 	go func() {
-		// persist dns state right away
+		defer s.shutdownWg.Done()
 		if err := s.stateManager.PersistState(s.ctx); err != nil {
 			log.Errorf("Failed to persist dns state: %v", err)
 		}

client/internal/dns/server_test.go

@@ -944,7 +944,7 @@ func createWgInterfaceWithBind(t *testing.T) (*iface.WGIface, error) {
 		return nil, err
 	}
 
-	pf, err := uspfilter.Create(wgIface, false, flowLogger)
+	pf, err := uspfilter.Create(wgIface, false, flowLogger, iface.DefaultMTU)
 	if err != nil {
 		t.Fatalf("failed to create uspfilter: %v", err)
 		return nil, err

client/internal/dns/unclean_shutdown_darwin.go

@@ -7,6 +7,7 @@ import (
 )
 
 type ShutdownState struct {
+	CreatedKeys []string
 }
 
 func (s *ShutdownState) Name() string {
@@ -19,6 +20,10 @@ func (s *ShutdownState) Cleanup() error {
 		return fmt.Errorf("create host manager: %w", err)
 	}
 
+	for _, key := range s.CreatedKeys {
+		manager.createdKeys[key] = struct{}{}
+	}
+
 	if err := manager.restoreUncleanShutdownDNS(); err != nil {
 		return fmt.Errorf("restore unclean shutdown dns: %w", err)
 	}

client/internal/dnsfwd/cache_test.go

@@ -83,4 +83,3 @@ func TestCacheMiss(t *testing.T) {
 		t.Fatalf("expected cache miss, got=%v ok=%v", got, ok)
 	}
 }
-

client/internal/dnsfwd/forwarder.go

@@ -14,6 +14,7 @@ import (
 	"github.com/hashicorp/go-multierror"
 	"github.com/miekg/dns"
 	log "github.com/sirupsen/logrus"
+	"golang.zx2c4.com/wireguard/tun/netstack"
 
 	nberrors "github.com/netbirdio/netbird/client/errors"
 	firewall "github.com/netbirdio/netbird/client/firewall/manager"
@@ -33,7 +34,7 @@ type firewaller interface {
 }
 
 type DNSForwarder struct {
-	listenAddress  string
+	listenAddress  netip.AddrPort
 	ttl            uint32
 	statusRecorder *peer.Status
 
@@ -47,9 +48,11 @@ type DNSForwarder struct {
 	firewall   firewaller
 	resolver   resolver
 	cache      *cache
+
+	wgIface wgIface
 }
 
-func NewDNSForwarder(listenAddress string, ttl uint32, firewall firewaller, statusRecorder *peer.Status) *DNSForwarder {
+func NewDNSForwarder(listenAddress netip.AddrPort, ttl uint32, firewall firewaller, statusRecorder *peer.Status, wgIface wgIface) *DNSForwarder {
 	log.Debugf("creating DNS forwarder with listen_address=%s ttl=%d", listenAddress, ttl)
 	return &DNSForwarder{
 		listenAddress:  listenAddress,
@@ -58,30 +61,46 @@ func NewDNSForwarder(listenAddress string, ttl uint32, firewall firewaller, stat
 		statusRecorder: statusRecorder,
 		resolver:       net.DefaultResolver,
 		cache:          newCache(),
+		wgIface:        wgIface,
 	}
 }
 
 func (f *DNSForwarder) Listen(entries []*ForwarderEntry) error {
-	log.Infof("starting DNS forwarder on address=%s", f.listenAddress)
+	var netstackNet *netstack.Net
+	if f.wgIface != nil {
+		netstackNet = f.wgIface.GetNet()
+	}
+
+	addrDesc := f.listenAddress.String()
+	if netstackNet != nil {
+		addrDesc = fmt.Sprintf("netstack %s", f.listenAddress)
+	}
+	log.Infof("starting DNS forwarder on address=%s", addrDesc)
+
+	udpLn, err := f.createUDPListener(netstackNet)
+	if err != nil {
+		return fmt.Errorf("create UDP listener: %w", err)
+	}
+
+	tcpLn, err := f.createTCPListener(netstackNet)
+	if err != nil {
+		return fmt.Errorf("create TCP listener: %w", err)
+	}
 
-	// UDP server
 	mux := dns.NewServeMux()
 	f.mux = mux
 	mux.HandleFunc(".", f.handleDNSQueryUDP)
 	f.dnsServer = &dns.Server{
-		Addr:    f.listenAddress,
-		Net:     "udp",
-		Handler: mux,
+		PacketConn: udpLn,
+		Handler:    mux,
 	}
 
-	// TCP server
 	tcpMux := dns.NewServeMux()
 	f.tcpMux = tcpMux
 	tcpMux.HandleFunc(".", f.handleDNSQueryTCP)
 	f.tcpServer = &dns.Server{
-		Addr:    f.listenAddress,
-		Net:     "tcp",
-		Handler: tcpMux,
+		Listener: tcpLn,
+		Handler:  tcpMux,
 	}
 
 	f.UpdateDomains(entries)
@@ -89,18 +108,33 @@ func (f *DNSForwarder) Listen(entries []*ForwarderEntry) error {
 	errCh := make(chan error, 2)
 
 	go func() {
-		log.Infof("DNS UDP listener running on %s", f.listenAddress)
-		errCh <- f.dnsServer.ListenAndServe()
+		log.Infof("DNS UDP listener running on %s", addrDesc)
+		errCh <- f.dnsServer.ActivateAndServe()
 	}()
 	go func() {
-		log.Infof("DNS TCP listener running on %s", f.listenAddress)
-		errCh <- f.tcpServer.ListenAndServe()
+		log.Infof("DNS TCP listener running on %s", addrDesc)
+		errCh <- f.tcpServer.ActivateAndServe()
 	}()
 
-	// return the first error we get (e.g. bind failure or shutdown)
 	return <-errCh
 }
 
+func (f *DNSForwarder) createUDPListener(netstackNet *netstack.Net) (net.PacketConn, error) {
+	if netstackNet != nil {
+		return netstackNet.ListenUDPAddrPort(f.listenAddress)
+	}
+
+	return net.ListenUDP("udp", net.UDPAddrFromAddrPort(f.listenAddress))
+}
+
+func (f *DNSForwarder) createTCPListener(netstackNet *netstack.Net) (net.Listener, error) {
+	if netstackNet != nil {
+		return netstackNet.ListenTCPAddrPort(f.listenAddress)
+	}
+
+	return net.ListenTCP("tcp", net.TCPAddrFromAddrPort(f.listenAddress))
+}
+
 func (f *DNSForwarder) UpdateDomains(entries []*ForwarderEntry) {
 	f.mutex.Lock()
 	defer f.mutex.Unlock()

client/internal/dnsfwd/forwarder_test.go

@@ -297,7 +297,7 @@ func TestDNSForwarder_UnauthorizedDomainAccess(t *testing.T) {
 				mockResolver.On("LookupNetIP", mock.Anything, "ip4", dns.Fqdn(tt.queryDomain)).Return([]netip.Addr{fakeIP}, nil)
 			}
 
-			forwarder := NewDNSForwarder("127.0.0.1:0", 300, mockFirewall, &peer.Status{})
+			forwarder := NewDNSForwarder(netip.MustParseAddrPort("127.0.0.1:0"), 300, mockFirewall, &peer.Status{}, nil)
 			forwarder.resolver = mockResolver
 
 			d, err := domain.FromString(tt.configuredDomain)
@@ -402,7 +402,7 @@ func TestDNSForwarder_FirewallSetUpdates(t *testing.T) {
 			mockResolver := &MockResolver{}
 
 			// Set up forwarder
-			forwarder := NewDNSForwarder("127.0.0.1:0", 300, mockFirewall, &peer.Status{})
+			forwarder := NewDNSForwarder(netip.MustParseAddrPort("127.0.0.1:0"), 300, mockFirewall, &peer.Status{}, nil)
 			forwarder.resolver = mockResolver
 
 			// Create entries and track sets
@@ -489,7 +489,7 @@ func TestDNSForwarder_MultipleIPsInSingleUpdate(t *testing.T) {
 	mockFirewall := &MockFirewall{}
 	mockResolver := &MockResolver{}
 
-	forwarder := NewDNSForwarder("127.0.0.1:0", 300, mockFirewall, &peer.Status{})
+	forwarder := NewDNSForwarder(netip.MustParseAddrPort("127.0.0.1:0"), 300, mockFirewall, &peer.Status{}, nil)
 	forwarder.resolver = mockResolver
 
 	// Configure a single domain
@@ -584,7 +584,7 @@ func TestDNSForwarder_ResponseCodes(t *testing.T) {
 
 	for _, tt := range tests {
 		t.Run(tt.name, func(t *testing.T) {
-			forwarder := NewDNSForwarder("127.0.0.1:0", 300, nil, &peer.Status{})
+			forwarder := NewDNSForwarder(netip.MustParseAddrPort("127.0.0.1:0"), 300, nil, &peer.Status{}, nil)
 
 			d, err := domain.FromString(tt.configured)
 			require.NoError(t, err)
@@ -616,7 +616,7 @@ func TestDNSForwarder_ResponseCodes(t *testing.T) {
 func TestDNSForwarder_TCPTruncation(t *testing.T) {
 	// Test that large UDP responses are truncated with TC bit set
 	mockResolver := &MockResolver{}
-	forwarder := NewDNSForwarder("127.0.0.1:0", 300, nil, &peer.Status{})
+	forwarder := NewDNSForwarder(netip.MustParseAddrPort("127.0.0.1:0"), 300, nil, &peer.Status{}, nil)
 	forwarder.resolver = mockResolver
 
 	d, _ := domain.FromString("example.com")
@@ -652,7 +652,7 @@ func TestDNSForwarder_TCPTruncation(t *testing.T) {
 // a subsequent upstream failure still returns a successful response from cache.
 func TestDNSForwarder_ServeFromCacheOnUpstreamFailure(t *testing.T) {
 	mockResolver := &MockResolver{}
-	forwarder := NewDNSForwarder("127.0.0.1:0", 300, nil, &peer.Status{})
+	forwarder := NewDNSForwarder(netip.MustParseAddrPort("127.0.0.1:0"), 300, nil, &peer.Status{}, nil)
 	forwarder.resolver = mockResolver
 
 	d, err := domain.FromString("example.com")
@@ -696,7 +696,7 @@ func TestDNSForwarder_ServeFromCacheOnUpstreamFailure(t *testing.T) {
 // Verifies that cache normalization works across casing and trailing dot variations.
 func TestDNSForwarder_CacheNormalizationCasingAndDot(t *testing.T) {
 	mockResolver := &MockResolver{}
-	forwarder := NewDNSForwarder("127.0.0.1:0", 300, nil, &peer.Status{})
+	forwarder := NewDNSForwarder(netip.MustParseAddrPort("127.0.0.1:0"), 300, nil, &peer.Status{}, nil)
 	forwarder.resolver = mockResolver
 
 	d, err := domain.FromString("ExAmPlE.CoM")
@@ -742,7 +742,7 @@ func TestDNSForwarder_MultipleOverlappingPatterns(t *testing.T) {
 	mockFirewall := &MockFirewall{}
 	mockResolver := &MockResolver{}
 
-	forwarder := NewDNSForwarder("127.0.0.1:0", 300, mockFirewall, &peer.Status{})
+	forwarder := NewDNSForwarder(netip.MustParseAddrPort("127.0.0.1:0"), 300, mockFirewall, &peer.Status{}, nil)
 	forwarder.resolver = mockResolver
 
 	// Set up complex overlapping patterns
@@ -804,7 +804,7 @@ func TestDNSForwarder_NodataVsNxdomain(t *testing.T) {
 	mockFirewall := &MockFirewall{}
 	mockResolver := &MockResolver{}
 
-	forwarder := NewDNSForwarder("127.0.0.1:0", 300, mockFirewall, &peer.Status{})
+	forwarder := NewDNSForwarder(netip.MustParseAddrPort("127.0.0.1:0"), 300, mockFirewall, &peer.Status{}, nil)
 	forwarder.resolver = mockResolver
 
 	d, err := domain.FromString("example.com")
@@ -925,7 +925,7 @@ func TestDNSForwarder_NodataVsNxdomain(t *testing.T) {
 
 func TestDNSForwarder_EmptyQuery(t *testing.T) {
 	// Test handling of malformed query with no questions
-	forwarder := NewDNSForwarder("127.0.0.1:0", 300, nil, &peer.Status{})
+	forwarder := NewDNSForwarder(netip.MustParseAddrPort("127.0.0.1:0"), 300, nil, &peer.Status{}, nil)
 
 	query := &dns.Msg{}
 	// Don't set any question

client/internal/dnsfwd/manager.go

@@ -4,27 +4,34 @@ import (
 	"context"
 	"fmt"
 	"net"
-	"sync"
+	"net/netip"
+	"os"
+	"strconv"
 
 	"github.com/hashicorp/go-multierror"
 	log "github.com/sirupsen/logrus"
+	"golang.zx2c4.com/wireguard/tun/netstack"
 
 	nberrors "github.com/netbirdio/netbird/client/errors"
 	firewall "github.com/netbirdio/netbird/client/firewall/manager"
+	"github.com/netbirdio/netbird/client/iface/wgaddr"
+	nftypes "github.com/netbirdio/netbird/client/internal/netflow/types"
 	"github.com/netbirdio/netbird/client/internal/peer"
+	nbdns "github.com/netbirdio/netbird/dns"
 	"github.com/netbirdio/netbird/route"
 	"github.com/netbirdio/netbird/shared/management/domain"
 )
 
-var (
-	// ListenPort is the port that the DNS forwarder listens on. It has been used by the client peers also
-	listenPort   uint16 = 5353
-	listenPortMu sync.RWMutex
+const (
+	dnsTTL        = 60
+	envServerPort = "NB_DNS_FORWARDER_PORT"
 )
 
-const (
-	dnsTTL = 60 //seconds
-)
+// wgIface defines the interface for WireGuard interface operations needed by the DNS forwarder.
+type wgIface interface {
+	GetNet() *netstack.Net
+	Address() wgaddr.Address
+}
 
 // ForwarderEntry is a mapping from a domain to a resource ID and a hash of the parent domain list.
 type ForwarderEntry struct {
@@ -36,28 +43,30 @@ type ForwarderEntry struct {
 type Manager struct {
 	firewall       firewall.Manager
 	statusRecorder *peer.Status
+	wgIface        wgIface
+	serverPort     uint16
 
 	fwRules      []firewall.Rule
 	tcpRules     []firewall.Rule
 	dnsForwarder *DNSForwarder
 }
 
-func ListenPort() uint16 {
-	listenPortMu.RLock()
-	defer listenPortMu.RUnlock()
-	return listenPort
-}
+func NewManager(fw firewall.Manager, statusRecorder *peer.Status, wgIface wgIface) *Manager {
+	serverPort := nbdns.ForwarderServerPort
+	if envPort := os.Getenv(envServerPort); envPort != "" {
+		if port, err := strconv.ParseUint(envPort, 10, 16); err == nil && port > 0 {
+			serverPort = uint16(port)
+			log.Infof("using custom DNS forwarder port from %s: %d", envServerPort, serverPort)
+		} else {
+			log.Warnf("invalid %s value %q, using default %d", envServerPort, envPort, nbdns.ForwarderServerPort)
+		}
+	}
 
-func SetListenPort(port uint16) {
-	listenPortMu.Lock()
-	listenPort = port
-	listenPortMu.Unlock()
-}
-
-func NewManager(fw firewall.Manager, statusRecorder *peer.Status) *Manager {
 	return &Manager{
 		firewall:       fw,
 		statusRecorder: statusRecorder,
+		wgIface:        wgIface,
+		serverPort:     serverPort,
 	}
 }
 
@@ -71,7 +80,25 @@ func (m *Manager) Start(fwdEntries []*ForwarderEntry) error {
 		return err
 	}
 
-	m.dnsForwarder = NewDNSForwarder(fmt.Sprintf(":%d", ListenPort()), dnsTTL, m.firewall, m.statusRecorder)
+	localAddr := m.wgIface.Address().IP
+
+	if localAddr.IsValid() && m.firewall != nil {
+		if err := m.firewall.AddInboundDNAT(localAddr, firewall.ProtocolUDP, nbdns.ForwarderClientPort, m.serverPort); err != nil {
+			log.Warnf("failed to add DNS UDP DNAT rule: %v", err)
+		} else {
+			log.Infof("added DNS UDP DNAT rule: %s:%d -> %s:%d", localAddr, nbdns.ForwarderClientPort, localAddr, m.serverPort)
+		}
+
+		if err := m.firewall.AddInboundDNAT(localAddr, firewall.ProtocolTCP, nbdns.ForwarderClientPort, m.serverPort); err != nil {
+			log.Warnf("failed to add DNS TCP DNAT rule: %v", err)
+		} else {
+			log.Infof("added DNS TCP DNAT rule: %s:%d -> %s:%d", localAddr, nbdns.ForwarderClientPort, localAddr, m.serverPort)
+		}
+	}
+
+	listenAddress := netip.AddrPortFrom(localAddr, m.serverPort)
+	m.dnsForwarder = NewDNSForwarder(listenAddress, dnsTTL, m.firewall, m.statusRecorder, m.wgIface)
+
 	go func() {
 		if err := m.dnsForwarder.Listen(fwdEntries); err != nil {
 			// todo handle close error if it is exists
@@ -96,6 +123,20 @@ func (m *Manager) Stop(ctx context.Context) error {
 	}
 
 	var mErr *multierror.Error
+
+	localAddr := m.wgIface.Address().IP
+	if localAddr.IsValid() && m.firewall != nil {
+		if err := m.firewall.RemoveInboundDNAT(localAddr, firewall.ProtocolUDP, nbdns.ForwarderClientPort, m.serverPort); err != nil {
+			mErr = multierror.Append(mErr, fmt.Errorf("remove DNS UDP DNAT rule: %w", err))
+		}
+
+		if err := m.firewall.RemoveInboundDNAT(localAddr, firewall.ProtocolTCP, nbdns.ForwarderClientPort, m.serverPort); err != nil {
+			mErr = multierror.Append(mErr, fmt.Errorf("remove DNS TCP DNAT rule: %w", err))
+		}
+	}
+
+	m.unregisterNetstackServices()
+
 	if err := m.dropDNSFirewall(); err != nil {
 		mErr = multierror.Append(mErr, err)
 	}
@@ -111,7 +152,7 @@ func (m *Manager) Stop(ctx context.Context) error {
 func (m *Manager) allowDNSFirewall() error {
 	dport := &firewall.Port{
 		IsRange: false,
-		Values:  []uint16{ListenPort()},
+		Values:  []uint16{m.serverPort},
 	}
 
 	if m.firewall == nil {
@@ -120,21 +161,50 @@ func (m *Manager) allowDNSFirewall() error {
 
 	dnsRules, err := m.firewall.AddPeerFiltering(nil, net.IP{0, 0, 0, 0}, firewall.ProtocolUDP, nil, dport, firewall.ActionAccept, "")
 	if err != nil {
-		log.Errorf("failed to add allow DNS router rules, err: %v", err)
-		return err
+		return fmt.Errorf("add udp firewall rule: %w", err)
 	}
-	m.fwRules = dnsRules
 
 	tcpRules, err := m.firewall.AddPeerFiltering(nil, net.IP{0, 0, 0, 0}, firewall.ProtocolTCP, nil, dport, firewall.ActionAccept, "")
 	if err != nil {
-		log.Errorf("failed to add allow DNS router rules, err: %v", err)
-		return err
+		return fmt.Errorf("add tcp firewall rule: %w", err)
 	}
+
+	if err := m.firewall.Flush(); err != nil {
+		return fmt.Errorf("flush: %w", err)
+	}
+
+	m.fwRules = dnsRules
 	m.tcpRules = tcpRules
 
+	m.registerNetstackServices()
+
 	return nil
 }
 
+func (m *Manager) registerNetstackServices() {
+	if netstackNet := m.wgIface.GetNet(); netstackNet != nil {
+		if registrar, ok := m.firewall.(interface {
+			RegisterNetstackService(protocol nftypes.Protocol, port uint16)
+		}); ok {
+			registrar.RegisterNetstackService(nftypes.TCP, m.serverPort)
+			registrar.RegisterNetstackService(nftypes.UDP, m.serverPort)
+			log.Debugf("registered DNS forwarder service with netstack for UDP/TCP:%d", m.serverPort)
+		}
+	}
+}
+
+func (m *Manager) unregisterNetstackServices() {
+	if netstackNet := m.wgIface.GetNet(); netstackNet != nil {
+		if registrar, ok := m.firewall.(interface {
+			UnregisterNetstackService(protocol nftypes.Protocol, port uint16)
+		}); ok {
+			registrar.UnregisterNetstackService(nftypes.TCP, m.serverPort)
+			registrar.UnregisterNetstackService(nftypes.UDP, m.serverPort)
+			log.Debugf("unregistered DNS forwarder service with netstack for UDP/TCP:%d", m.serverPort)
+		}
+	}
+}
+
 func (m *Manager) dropDNSFirewall() error {
 	var mErr *multierror.Error
 	for _, rule := range m.fwRules {

client/internal/engine.go

@@ -150,6 +150,8 @@ type Engine struct {
 
 	// syncMsgMux is used to guarantee sequential Management Service message processing
 	syncMsgMux *sync.Mutex
+	// sshMux protects sshServer field access
+	sshMux sync.Mutex
 
 	config    *EngineConfig
 	mobileDep MobileDependency
@@ -205,11 +207,12 @@ type Engine struct {
 	updateManager *updatemanager.UpdateManager
 
 	// WireGuard interface monitor
-	wgIfaceMonitor   *WGIfaceMonitor
-	wgIfaceMonitorWg sync.WaitGroup
+	wgIfaceMonitor *WGIfaceMonitor
 
-	// dns forwarder port
-	dnsFwdPort uint16
+	// shutdownWg tracks all long-running goroutines to ensure clean shutdown
+	shutdownWg sync.WaitGroup
+
+	probeStunTurn *relay.StunTurnProbe
 }
 
 // Peer is an instance of the Connection Peer
@@ -252,7 +255,7 @@ func NewEngine(
 		statusRecorder: statusRecorder,
 		checks:         checks,
 		connSemaphore:  semaphoregroup.NewSemaphoreGroup(connInitLimit),
-		dnsFwdPort:     dnsfwd.ListenPort(),
+		probeStunTurn:  relay.NewStunTurnProbe(relay.DefaultCacheTTL),
 	}
 
 	sm := profilemanager.NewServiceManager("")
@@ -304,17 +307,12 @@ func (e *Engine) Stop() error {
 		e.ingressGatewayMgr = nil
 	}
 
+	e.stopDNSForwarder()
+
 	if e.routeManager != nil {
 		e.routeManager.Stop(e.stateManager)
 	}
 
-	if e.dnsForwardMgr != nil {
-		if err := e.dnsForwardMgr.Stop(context.Background()); err != nil {
-			log.Errorf("failed to stop DNS forward: %v", err)
-		}
-		e.dnsForwardMgr = nil
-	}
-
 	if e.srWatcher != nil {
 		e.srWatcher.Close()
 	}
@@ -335,10 +333,6 @@ func (e *Engine) Stop() error {
 		e.cancel()
 	}
 
-	// very ugly but we want to remove peers from the WireGuard interface first before removing interface.
-	// Removing peers happens in the conn.Close() asynchronously
-	time.Sleep(500 * time.Millisecond)
-
 	e.close()
 
 	// stop flow manager after wg interface is gone
@@ -346,8 +340,6 @@ func (e *Engine) Stop() error {
 		e.flowManager.Close()
 	}
 
-	log.Infof("stopped Netbird Engine")
-
 	ctx, cancel := context.WithTimeout(context.Background(), 3*time.Second)
 	defer cancel()
 
@@ -358,12 +350,52 @@ func (e *Engine) Stop() error {
 		log.Errorf("failed to persist state: %v", err)
 	}
 
-	// Stop WireGuard interface monitor and wait for it to exit
-	e.wgIfaceMonitorWg.Wait()
+	timeout := e.calculateShutdownTimeout()
+	log.Debugf("waiting for goroutines to finish with timeout: %v", timeout)
+	shutdownCtx, cancel := context.WithTimeout(context.Background(), timeout)
+	defer cancel()
+
+	if err := waitWithContext(shutdownCtx, &e.shutdownWg); err != nil {
+		log.Warnf("shutdown timeout exceeded after %v, some goroutines may still be running", timeout)
+	}
+
+	log.Infof("stopped Netbird Engine")
 
 	return nil
 }
 
+// calculateShutdownTimeout returns shutdown timeout: 10s base + 100ms per peer, capped at 30s.
+func (e *Engine) calculateShutdownTimeout() time.Duration {
+	peerCount := len(e.peerStore.PeersPubKey())
+
+	baseTimeout := 10 * time.Second
+	perPeerTimeout := time.Duration(peerCount) * 100 * time.Millisecond
+	timeout := baseTimeout + perPeerTimeout
+
+	maxTimeout := 30 * time.Second
+	if timeout > maxTimeout {
+		timeout = maxTimeout
+	}
+
+	return timeout
+}
+
+// waitWithContext waits for WaitGroup with timeout, returns ctx.Err() on timeout.
+func waitWithContext(ctx context.Context, wg *sync.WaitGroup) error {
+	done := make(chan struct{})
+	go func() {
+		wg.Wait()
+		close(done)
+	}()
+
+	select {
+	case <-done:
+		return nil
+	case <-ctx.Done():
+		return ctx.Err()
+	}
+}
+
 // Start creates a new WireGuard tunnel interface and listens to events from Signal and Management services
 // Connections to remote peers are not established here.
 // However, they will be established once an event with a list of peers to connect to will be received from Management Service
@@ -493,14 +525,14 @@ func (e *Engine) Start(netbirdConfig *mgmProto.NetbirdConfig, mgmtURL *url.URL)
 
 	// monitor WireGuard interface lifecycle and restart engine on changes
 	e.wgIfaceMonitor = NewWGIfaceMonitor()
-	e.wgIfaceMonitorWg.Add(1)
+	e.shutdownWg.Add(1)
 
 	go func() {
-		defer e.wgIfaceMonitorWg.Done()
+		defer e.shutdownWg.Done()
 
 		if shouldRestart, err := e.wgIfaceMonitor.Start(e.ctx, e.wgInterface.Name()); shouldRestart {
 			log.Infof("WireGuard interface monitor: %s, restarting engine", err)
-			e.restartEngine()
+			e.triggerClientRestart()
 		} else if err != nil {
 			log.Warnf("WireGuard interface monitor: %s", err)
 		}
@@ -529,7 +561,7 @@ func (e *Engine) createFirewall() error {
 	}
 
 	var err error
-	e.firewall, err = firewall.NewFirewall(e.wgInterface, e.stateManager, e.flowManager.GetLogger(), e.config.DisableServerRoutes)
+	e.firewall, err = firewall.NewFirewall(e.wgInterface, e.stateManager, e.flowManager.GetLogger(), e.config.DisableServerRoutes, e.config.MTU)
 	if err != nil || e.firewall == nil {
 		log.Errorf("failed creating firewall manager: %s", err)
 		return nil
@@ -697,9 +729,11 @@ func (e *Engine) removeAllPeers() error {
 func (e *Engine) removePeer(peerKey string) error {
 	log.Debugf("removing peer from engine %s", peerKey)
 
+	e.sshMux.Lock()
 	if !isNil(e.sshServer) {
 		e.sshServer.RemoveAuthorizedKey(peerKey)
 	}
+	e.sshMux.Unlock()
 
 	e.connMgr.RemovePeerConn(peerKey)
 
@@ -935,6 +969,7 @@ func (e *Engine) updateSSH(sshConf *mgmProto.SSHConfig) error {
 			log.Warnf("running SSH server on %s is not supported", runtime.GOOS)
 			return nil
 		}
+		e.sshMux.Lock()
 		// start SSH server if it wasn't running
 		if isNil(e.sshServer) {
 			listenAddr := fmt.Sprintf("%s:%d", e.wgInterface.Address().IP.String(), nbssh.DefaultSSHPort)
@@ -942,34 +977,42 @@ func (e *Engine) updateSSH(sshConf *mgmProto.SSHConfig) error {
 				listenAddr = fmt.Sprintf("127.0.0.1:%d", nbssh.DefaultSSHPort)
 			}
 			// nil sshServer means it has not yet been started
-			var err error
-			e.sshServer, err = e.sshServerFunc(e.config.SSHKey, listenAddr)
-
+			server, err := e.sshServerFunc(e.config.SSHKey, listenAddr)
 			if err != nil {
+				e.sshMux.Unlock()
 				return fmt.Errorf("create ssh server: %w", err)
 			}
+
+			e.sshServer = server
+			e.sshMux.Unlock()
+
 			go func() {
 				// blocking
-				err = e.sshServer.Start()
+				err = server.Start()
 				if err != nil {
 					// will throw error when we stop it even if it is a graceful stop
 					log.Debugf("stopped SSH server with error %v", err)
 				}
-				e.syncMsgMux.Lock()
-				defer e.syncMsgMux.Unlock()
+				e.sshMux.Lock()
 				e.sshServer = nil
+				e.sshMux.Unlock()
 				log.Infof("stopped SSH server")
 			}()
 		} else {
+			e.sshMux.Unlock()
 			log.Debugf("SSH server is already running")
 		}
-	} else if !isNil(e.sshServer) {
-		// Disable SSH server request, so stop it if it was running
-		err := e.sshServer.Stop()
-		if err != nil {
-			log.Warnf("failed to stop SSH server %v", err)
+	} else {
+		e.sshMux.Lock()
+		if !isNil(e.sshServer) {
+			// Disable SSH server request, so stop it if it was running
+			err := e.sshServer.Stop()
+			if err != nil {
+				log.Warnf("failed to stop SSH server %v", err)
+			}
+			e.sshServer = nil
 		}
-		e.sshServer = nil
+		e.sshMux.Unlock()
 	}
 	return nil
 }
@@ -1006,7 +1049,9 @@ func (e *Engine) updateConfig(conf *mgmProto.PeerConfig) error {
 // receiveManagementEvents connects to the Management Service event stream to receive updates from the management service
 // E.g. when a new peer has been registered and we are allowed to connect to it.
 func (e *Engine) receiveManagementEvents() {
+	e.shutdownWg.Add(1)
 	go func() {
+		defer e.shutdownWg.Done()
 		info, err := system.GetInfoWithChecks(e.ctx, e.checks)
 		if err != nil {
 			log.Warnf("failed to get system info with checks: %v", err)
@@ -1116,10 +1161,14 @@ func (e *Engine) updateNetworkMap(networkMap *mgmProto.NetworkMap) error {
 		protoDNSConfig = &mgmProto.DNSConfig{}
 	}
 
-	if err := e.dnsServer.UpdateDNSServer(serial, toDNSConfig(protoDNSConfig, e.wgInterface.Address().Network)); err != nil {
+	dnsConfig := toDNSConfig(protoDNSConfig, e.wgInterface.Address().Network)
+
+	if err := e.dnsServer.UpdateDNSServer(serial, dnsConfig); err != nil {
 		log.Errorf("failed to update dns server, err: %v", err)
 	}
 
+	e.routeManager.SetDNSForwarderPort(dnsConfig.ForwarderPort)
+
 	// apply routes first, route related actions might depend on routing being enabled
 	routes := toRoutes(networkMap.GetRoutes())
 	serverRoutes, clientRoutes := e.routeManager.ClassifyRoutes(routes)
@@ -1140,7 +1189,7 @@ func (e *Engine) updateNetworkMap(networkMap *mgmProto.NetworkMap) error {
 	}
 
 	fwdEntries := toRouteDomains(e.config.WgPrivateKey.PublicKey().String(), routes)
-	e.updateDNSForwarder(dnsRouteFeatureFlag, fwdEntries, uint16(protoDNSConfig.ForwarderPort))
+	e.updateDNSForwarder(dnsRouteFeatureFlag, fwdEntries)
 
 	// Ingress forward rules
 	forwardingRules, err := e.updateForwardRules(networkMap.GetForwardingRules())
@@ -1178,6 +1227,7 @@ func (e *Engine) updateNetworkMap(networkMap *mgmProto.NetworkMap) error {
 		e.statusRecorder.FinishPeerListModifications()
 
 		// update SSHServer by adding remote peer SSH keys
+		e.sshMux.Lock()
 		if !isNil(e.sshServer) {
 			for _, config := range networkMap.GetRemotePeers() {
 				if config.GetSshConfig() != nil && config.GetSshConfig().GetSshPubKey() != nil {
@@ -1188,6 +1238,7 @@ func (e *Engine) updateNetworkMap(networkMap *mgmProto.NetworkMap) error {
 				}
 			}
 		}
+		e.sshMux.Unlock()
 	}
 
 	// must set the exclude list after the peers are added. Without it the manager can not figure out the peers parameters from the store
@@ -1264,10 +1315,16 @@ func toRouteDomains(myPubKey string, routes []*route.Route) []*dnsfwd.ForwarderE
 }
 
 func toDNSConfig(protoDNSConfig *mgmProto.DNSConfig, network netip.Prefix) nbdns.Config {
+	forwarderPort := uint16(protoDNSConfig.GetForwarderPort())
+	if forwarderPort == 0 {
+		forwarderPort = nbdns.ForwarderClientPort
+	}
+
 	dnsUpdate := nbdns.Config{
 		ServiceEnable:    protoDNSConfig.GetServiceEnable(),
 		CustomZones:      make([]nbdns.CustomZone, 0),
 		NameServerGroups: make([]*nbdns.NameServerGroup, 0),
+		ForwarderPort:    forwarderPort,
 	}
 
 	for _, zone := range protoDNSConfig.GetCustomZones() {
@@ -1424,7 +1481,9 @@ func (e *Engine) createPeerConn(pubKey string, allowedIPs []netip.Prefix, agentV
 
 // receiveSignalEvents connects to the Signal Service event stream to negotiate connection with remote peers
 func (e *Engine) receiveSignalEvents() {
+	e.shutdownWg.Add(1)
 	go func() {
+		defer e.shutdownWg.Done()
 		// connect to a stream of messages coming from the signal server
 		err := e.signal.Receive(e.ctx, func(msg *sProto.Message) error {
 			e.syncMsgMux.Lock()
@@ -1541,12 +1600,14 @@ func (e *Engine) close() {
 		e.statusRecorder.SetWgIface(nil)
 	}
 
+	e.sshMux.Lock()
 	if !isNil(e.sshServer) {
 		err := e.sshServer.Stop()
 		if err != nil {
 			log.Warnf("failed stopping the SSH server: %v", err)
 		}
 	}
+	e.sshMux.Unlock()
 
 	if e.firewall != nil {
 		err := e.firewall.Close(e.stateManager)
@@ -1723,7 +1784,7 @@ func (e *Engine) getRosenpassAddr() string {
 
 // RunHealthProbes executes health checks for Signal, Management, Relay and WireGuard services
 // and updates the status recorder with the latest states.
-func (e *Engine) RunHealthProbes() bool {
+func (e *Engine) RunHealthProbes(waitForResult bool) bool {
 	e.syncMsgMux.Lock()
 
 	signalHealthy := e.signal.IsHealthy()
@@ -1755,8 +1816,12 @@ func (e *Engine) RunHealthProbes() bool {
 	}
 
 	e.syncMsgMux.Unlock()
-
-	results := e.probeICE(stuns, turns)
+	var results []relay.ProbeResult
+	if waitForResult {
+		results = e.probeStunTurn.ProbeAllWaitResult(e.ctx, stuns, turns)
+	} else {
+		results = e.probeStunTurn.ProbeAll(e.ctx, stuns, turns)
+	}
 	e.statusRecorder.UpdateRelayStates(results)
 
 	relayHealthy := true
@@ -1773,15 +1838,10 @@ func (e *Engine) RunHealthProbes() bool {
 	return allHealthy
 }
 
-func (e *Engine) probeICE(stuns, turns []*stun.URI) []relay.ProbeResult {
-	return append(
-		relay.ProbeAll(e.ctx, relay.ProbeSTUN, stuns),
-		relay.ProbeAll(e.ctx, relay.ProbeTURN, turns)...,
-	)
-}
-
-// restartEngine restarts the engine by cancelling the client context
-func (e *Engine) restartEngine() {
+// triggerClientRestart triggers a full client restart by cancelling the client context.
+// Note: This does NOT just restart the engine - it cancels the entire client context,
+// which causes the connect client's retry loop to create a completely new engine.
+func (e *Engine) triggerClientRestart() {
 	e.syncMsgMux.Lock()
 	defer e.syncMsgMux.Unlock()
 
@@ -1803,7 +1863,9 @@ func (e *Engine) startNetworkMonitor() {
 	}
 
 	e.networkMonitor = networkmonitor.New()
+	e.shutdownWg.Add(1)
 	go func() {
+		defer e.shutdownWg.Done()
 		if err := e.networkMonitor.Listen(e.ctx); err != nil {
 			if errors.Is(err, context.Canceled) {
 				log.Infof("network monitor stopped")
@@ -1813,8 +1875,8 @@ func (e *Engine) startNetworkMonitor() {
 			return
 		}
 
-		log.Infof("Network monitor: detected network change, restarting engine")
-		e.restartEngine()
+		log.Infof("Network monitor: detected network change, triggering client restart")
+		e.triggerClientRestart()
 	}()
 }
 
@@ -1899,64 +1961,50 @@ func (e *Engine) GetWgAddr() netip.Addr {
 func (e *Engine) updateDNSForwarder(
 	enabled bool,
 	fwdEntries []*dnsfwd.ForwarderEntry,
-	forwarderPort uint16,
 ) {
 	if e.config.DisableServerRoutes {
 		return
 	}
 
-	if forwarderPort > 0 {
-		dnsfwd.SetListenPort(forwarderPort)
-	}
-
 	if !enabled {
-		if e.dnsForwardMgr == nil {
-			return
-		}
-		if err := e.dnsForwardMgr.Stop(context.Background()); err != nil {
-			log.Errorf("failed to stop DNS forward: %v", err)
-		}
+		e.stopDNSForwarder()
 		return
 	}
 
 	if len(fwdEntries) > 0 {
-		switch {
-		case e.dnsForwardMgr == nil:
-			e.dnsForwardMgr = dnsfwd.NewManager(e.firewall, e.statusRecorder)
-			if err := e.dnsForwardMgr.Start(fwdEntries); err != nil {
-				log.Errorf("failed to start DNS forward: %v", err)
-				e.dnsForwardMgr = nil
-			}
-			log.Infof("started domain router service with %d entries", len(fwdEntries))
-		case e.dnsFwdPort != forwarderPort:
-			log.Infof("updating domain router service port from %d to %d", e.dnsFwdPort, forwarderPort)
-			e.restartDnsFwd(fwdEntries, forwarderPort)
-			e.dnsFwdPort = forwarderPort
-
-		default:
+		if e.dnsForwardMgr == nil {
+			e.startDNSForwarder(fwdEntries)
+		} else {
 			e.dnsForwardMgr.UpdateDomains(fwdEntries)
 		}
 	} else if e.dnsForwardMgr != nil {
 		log.Infof("disable domain router service")
-		if err := e.dnsForwardMgr.Stop(context.Background()); err != nil {
-			log.Errorf("failed to stop DNS forward: %v", err)
-		}
-		e.dnsForwardMgr = nil
+		e.stopDNSForwarder()
 	}
-
 }
 
-func (e *Engine) restartDnsFwd(fwdEntries []*dnsfwd.ForwarderEntry, forwarderPort uint16) {
-	log.Infof("updating domain router service port from %d to %d", e.dnsFwdPort, forwarderPort)
-	// stop and start the forwarder to apply the new port
-	if err := e.dnsForwardMgr.Stop(context.Background()); err != nil {
-		log.Errorf("failed to stop DNS forward: %v", err)
-	}
-	e.dnsForwardMgr = dnsfwd.NewManager(e.firewall, e.statusRecorder)
+func (e *Engine) startDNSForwarder(fwdEntries []*dnsfwd.ForwarderEntry) {
+	e.dnsForwardMgr = dnsfwd.NewManager(e.firewall, e.statusRecorder, e.wgInterface)
+
 	if err := e.dnsForwardMgr.Start(fwdEntries); err != nil {
 		log.Errorf("failed to start DNS forward: %v", err)
 		e.dnsForwardMgr = nil
+		return
 	}
+
+	log.Infof("started domain router service with %d entries", len(fwdEntries))
+}
+
+func (e *Engine) stopDNSForwarder() {
+	if e.dnsForwardMgr == nil {
+		return
+	}
+
+	if err := e.dnsForwardMgr.Stop(context.Background()); err != nil {
+		log.Errorf("failed to stop DNS forward: %v", err)
+	}
+
+	e.dnsForwardMgr = nil
 }
 
 func (e *Engine) GetNet() (*netstack.Net, error) {

client/internal/netflow/logger/logger.go

@@ -10,10 +10,10 @@ import (
 	"github.com/google/uuid"
 	log "github.com/sirupsen/logrus"
 
-	"github.com/netbirdio/netbird/client/internal/dnsfwd"
 	"github.com/netbirdio/netbird/client/internal/netflow/store"
 	"github.com/netbirdio/netbird/client/internal/netflow/types"
 	"github.com/netbirdio/netbird/client/internal/peer"
+	"github.com/netbirdio/netbird/dns"
 )
 
 type rcvChan chan *types.EventFields
@@ -138,7 +138,8 @@ func (l *Logger) UpdateConfig(dnsCollection, exitNodeCollection bool) {
 
 func (l *Logger) shouldStore(event *types.EventFields, isExitNode bool) bool {
 	// check dns collection
-	if !l.dnsCollection.Load() && event.Protocol == types.UDP && (event.DestPort == 53 || event.DestPort == uint16(dnsfwd.ListenPort())) {
+	if !l.dnsCollection.Load() && event.Protocol == types.UDP &&
+		(event.DestPort == 53 || event.DestPort == dns.ForwarderClientPort || event.DestPort == dns.ForwarderServerPort) {
 		return false
 	}
 

client/internal/netflow/manager.go

@@ -24,6 +24,7 @@ import (
 // Manager handles netflow tracking and logging
 type Manager struct {
 	mux            sync.Mutex
+	shutdownWg     sync.WaitGroup
 	logger         nftypes.FlowLogger
 	flowConfig     *nftypes.FlowConfig
 	conntrack      nftypes.ConnTracker
@@ -105,8 +106,15 @@ func (m *Manager) resetClient() error {
 	ctx, cancel := context.WithCancel(context.Background())
 	m.cancel = cancel
 
-	go m.receiveACKs(ctx, flowClient)
-	go m.startSender(ctx)
+	m.shutdownWg.Add(2)
+	go func() {
+		defer m.shutdownWg.Done()
+		m.receiveACKs(ctx, flowClient)
+	}()
+	go func() {
+		defer m.shutdownWg.Done()
+		m.startSender(ctx)
+	}()
 
 	return nil
 }
@@ -176,11 +184,12 @@ func (m *Manager) Update(update *nftypes.FlowConfig) error {
 // Close cleans up all resources
 func (m *Manager) Close() {
 	m.mux.Lock()
-	defer m.mux.Unlock()
-
 	if err := m.disableFlow(); err != nil {
 		log.Warnf("failed to disable flow manager: %v", err)
 	}
+	m.mux.Unlock()
+
+	m.shutdownWg.Wait()
 }
 
 // GetLogger returns the flow logger

client/internal/networkmonitor/check_change_bsd.go

@@ -1,4 +1,4 @@
-//go:build (darwin && !ios) || dragonfly || freebsd || netbsd || openbsd
+//go:build dragonfly || freebsd || netbsd || openbsd
 
 package networkmonitor
 
@@ -6,21 +6,19 @@ import (
 	"context"
 	"errors"
 	"fmt"
-	"syscall"
-	"unsafe"
 
 	log "github.com/sirupsen/logrus"
-	"golang.org/x/net/route"
 	"golang.org/x/sys/unix"
 
 	"github.com/netbirdio/netbird/client/internal/routemanager/systemops"
 )
 
 func checkChange(ctx context.Context, nexthopv4, nexthopv6 systemops.Nexthop) error {
-	fd, err := unix.Socket(syscall.AF_ROUTE, syscall.SOCK_RAW, syscall.AF_UNSPEC)
+	fd, err := prepareFd()
 	if err != nil {
 		return fmt.Errorf("open routing socket: %v", err)
 	}
+
 	defer func() {
 		err := unix.Close(fd)
 		if err != nil && !errors.Is(err, unix.EBADF) {
@@ -28,72 +26,5 @@ func checkChange(ctx context.Context, nexthopv4, nexthopv6 systemops.Nexthop) er
 		}
 	}()
 
-	for {
-		select {
-		case <-ctx.Done():
-			return ctx.Err()
-		default:
-			buf := make([]byte, 2048)
-			n, err := unix.Read(fd, buf)
-			if err != nil {
-				if !errors.Is(err, unix.EBADF) && !errors.Is(err, unix.EINVAL) {
-					log.Warnf("Network monitor: failed to read from routing socket: %v", err)
-				}
-				continue
-			}
-			if n < unix.SizeofRtMsghdr {
-				log.Debugf("Network monitor: read from routing socket returned less than expected: %d bytes", n)
-				continue
-			}
-
-			msg := (*unix.RtMsghdr)(unsafe.Pointer(&buf[0]))
-
-			switch msg.Type {
-			// handle route changes
-			case unix.RTM_ADD, syscall.RTM_DELETE:
-				route, err := parseRouteMessage(buf[:n])
-				if err != nil {
-					log.Debugf("Network monitor: error parsing routing message: %v", err)
-					continue
-				}
-
-				if route.Dst.Bits() != 0 {
-					continue
-				}
-
-				intf := "<nil>"
-				if route.Interface != nil {
-					intf = route.Interface.Name
-				}
-				switch msg.Type {
-				case unix.RTM_ADD:
-					log.Infof("Network monitor: default route changed: via %s, interface %s", route.Gw, intf)
-					return nil
-				case unix.RTM_DELETE:
-					if nexthopv4.Intf != nil && route.Gw.Compare(nexthopv4.IP) == 0 || nexthopv6.Intf != nil && route.Gw.Compare(nexthopv6.IP) == 0 {
-						log.Infof("Network monitor: default route removed: via %s, interface %s", route.Gw, intf)
-						return nil
-					}
-				}
-			}
-		}
-	}
-}
-
-func parseRouteMessage(buf []byte) (*systemops.Route, error) {
-	msgs, err := route.ParseRIB(route.RIBTypeRoute, buf)
-	if err != nil {
-		return nil, fmt.Errorf("parse RIB: %v", err)
-	}
-
-	if len(msgs) != 1 {
-		return nil, 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 systemops.MsgToRoute(msg)
+	return routeCheck(ctx, fd, nexthopv4, nexthopv6)
 }

client/internal/networkmonitor/check_change_common.go

@@ -0,0 +1,92 @@
+//go:build dragonfly || freebsd || netbsd || openbsd || darwin
+
+package networkmonitor
+
+import (
+	"context"
+	"errors"
+	"fmt"
+	"syscall"
+	"unsafe"
+
+	log "github.com/sirupsen/logrus"
+	"golang.org/x/net/route"
+	"golang.org/x/sys/unix"
+
+	"github.com/netbirdio/netbird/client/internal/routemanager/systemops"
+)
+
+func prepareFd() (int, error) {
+	return unix.Socket(syscall.AF_ROUTE, syscall.SOCK_RAW, syscall.AF_UNSPEC)
+}
+
+func routeCheck(ctx context.Context, fd int, nexthopv4, nexthopv6 systemops.Nexthop) error {
+	for {
+		select {
+		case <-ctx.Done():
+			return ctx.Err()
+		default:
+			buf := make([]byte, 2048)
+			n, err := unix.Read(fd, buf)
+			if err != nil {
+				if !errors.Is(err, unix.EBADF) && !errors.Is(err, unix.EINVAL) {
+					log.Warnf("Network monitor: failed to read from routing socket: %v", err)
+				}
+				continue
+			}
+			if n < unix.SizeofRtMsghdr {
+				log.Debugf("Network monitor: read from routing socket returned less than expected: %d bytes", n)
+				continue
+			}
+
+			msg := (*unix.RtMsghdr)(unsafe.Pointer(&buf[0]))
+
+			switch msg.Type {
+			// handle route changes
+			case unix.RTM_ADD, syscall.RTM_DELETE:
+				route, err := parseRouteMessage(buf[:n])
+				if err != nil {
+					log.Debugf("Network monitor: error parsing routing message: %v", err)
+					continue
+				}
+
+				if route.Dst.Bits() != 0 {
+					continue
+				}
+
+				intf := "<nil>"
+				if route.Interface != nil {
+					intf = route.Interface.Name
+				}
+				switch msg.Type {
+				case unix.RTM_ADD:
+					log.Infof("Network monitor: default route changed: via %s, interface %s", route.Gw, intf)
+					return nil
+				case unix.RTM_DELETE:
+					if nexthopv4.Intf != nil && route.Gw.Compare(nexthopv4.IP) == 0 || nexthopv6.Intf != nil && route.Gw.Compare(nexthopv6.IP) == 0 {
+						log.Infof("Network monitor: default route removed: via %s, interface %s", route.Gw, intf)
+						return nil
+					}
+				}
+			}
+		}
+	}
+}
+
+func parseRouteMessage(buf []byte) (*systemops.Route, error) {
+	msgs, err := route.ParseRIB(route.RIBTypeRoute, buf)
+	if err != nil {
+		return nil, fmt.Errorf("parse RIB: %v", err)
+	}
+
+	if len(msgs) != 1 {
+		return nil, 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 systemops.MsgToRoute(msg)
+}

client/internal/networkmonitor/check_change_darwin.go

@@ -0,0 +1,149 @@
+//go:build darwin && !ios
+
+package networkmonitor
+
+import (
+	"context"
+	"errors"
+	"fmt"
+	"hash/fnv"
+	"os/exec"
+	"time"
+
+	log "github.com/sirupsen/logrus"
+	"golang.org/x/sys/unix"
+
+	"github.com/netbirdio/netbird/client/internal/routemanager/systemops"
+)
+
+// todo: refactor to not use static functions
+
+func checkChange(ctx context.Context, nexthopv4, nexthopv6 systemops.Nexthop) error {
+	fd, err := prepareFd()
+	if err != nil {
+		return fmt.Errorf("open routing socket: %v", err)
+	}
+
+	defer func() {
+		if err := unix.Close(fd); err != nil {
+			if !errors.Is(err, unix.EBADF) {
+				log.Warnf("Network monitor: failed to close routing socket: %v", err)
+			}
+		}
+	}()
+
+	routeChanged := make(chan struct{})
+	go func() {
+		_ = routeCheck(ctx, fd, nexthopv4, nexthopv6)
+		close(routeChanged)
+	}()
+
+	wakeUp := make(chan struct{})
+	go func() {
+		wakeUpListen(ctx)
+		close(wakeUp)
+	}()
+
+	select {
+	case <-ctx.Done():
+		return ctx.Err()
+	case <-routeChanged:
+		if ctx.Err() != nil {
+			return ctx.Err()
+		}
+		log.Infof("route change detected")
+		return nil
+	case <-wakeUp:
+		if ctx.Err() != nil {
+			return ctx.Err()
+		}
+		log.Infof("wakeup detected")
+		return nil
+	}
+}
+
+func wakeUpListen(ctx context.Context) {
+	log.Infof("start to watch for system wakeups")
+	var (
+		initialHash uint32
+		err         error
+	)
+
+	// Keep retrying until initial sysctl succeeds or context is canceled
+	for {
+		select {
+		case <-ctx.Done():
+			log.Info("exit from wakeUpListen initial hash detection due to context cancellation")
+			return
+		default:
+			initialHash, err = readSleepTimeHash()
+			if err != nil {
+				log.Errorf("failed to detect initial sleep time: %v", err)
+				select {
+				case <-ctx.Done():
+					log.Info("exit from wakeUpListen initial hash detection due to context cancellation")
+					return
+				case <-time.After(3 * time.Second):
+					continue
+				}
+			}
+			log.Debugf("initial wakeup hash: %d", initialHash)
+			break
+		}
+		break
+	}
+
+	ticker := time.NewTicker(5 * time.Second)
+	defer ticker.Stop()
+
+	for {
+		select {
+		case <-ctx.Done():
+			log.Info("context canceled, stopping wakeUpListen")
+			return
+
+		case <-ticker.C:
+			newHash, err := readSleepTimeHash()
+			if err != nil {
+				log.Errorf("failed to read sleep time hash: %v", err)
+				continue
+			}
+
+			if newHash == initialHash {
+				log.Tracef("no wakeup detected")
+				continue
+			}
+
+			upOut, err := exec.Command("uptime").Output()
+			if err != nil {
+				log.Errorf("failed to run uptime command: %v", err)
+				upOut = []byte("unknown")
+			}
+			log.Infof("Wakeup detected: %d -> %d, uptime: %s", initialHash, newHash, upOut)
+			return
+		}
+	}
+}
+
+func readSleepTimeHash() (uint32, error) {
+	cmd := exec.Command("sysctl", "kern.sleeptime")
+	out, err := cmd.Output()
+	if err != nil {
+		return 0, fmt.Errorf("failed to run sysctl: %w", err)
+	}
+
+	h, err := hash(out)
+	if err != nil {
+		return 0, fmt.Errorf("failed to compute hash: %w", err)
+	}
+
+	return h, nil
+}
+
+func hash(data []byte) (uint32, error) {
+	hasher := fnv.New32a() // Create a new 32-bit FNV-1a hasher
+	if _, err := hasher.Write(data); err != nil {
+		return 0, err
+	}
+	return hasher.Sum32(), nil
+}

client/internal/networkmonitor/monitor.go

@@ -88,6 +88,7 @@ func (nw *NetworkMonitor) Listen(ctx context.Context) (err error) {
 	event := make(chan struct{}, 1)
 	go nw.checkChanges(ctx, event, nexthop4, nexthop6)
 
+	log.Infof("start watching for network changes")
 	// debounce changes
 	timer := time.NewTimer(0)
 	timer.Stop()

client/internal/peer/guard/sr_watcher.go

@@ -19,11 +19,10 @@ type SRWatcher struct {
 	signalClient chNotifier
 	relayManager chNotifier
 
-	listeners     map[chan struct{}]struct{}
-	mu            sync.Mutex
-	iFaceDiscover stdnet.ExternalIFaceDiscover
-	iceConfig     ice.Config
-
+	listeners        map[chan struct{}]struct{}
+	mu               sync.Mutex
+	iFaceDiscover    stdnet.ExternalIFaceDiscover
+	iceConfig        ice.Config
 	cancelIceMonitor context.CancelFunc
 }
 

client/internal/peer/worker_ice.go

@@ -411,7 +411,7 @@ func (w *WorkerICE) onConnectionStateChange(agent *icemaker.ThreadSafeAgent, dia
 
 func (w *WorkerICE) turnAgentDial(ctx context.Context, agent *icemaker.ThreadSafeAgent, remoteOfferAnswer *OfferAnswer) (*ice.Conn, error) {
 	if isController(w.config) {
-		return w.agent.Dial(ctx, remoteOfferAnswer.IceCredentials.UFrag, remoteOfferAnswer.IceCredentials.Pwd)
+		return agent.Dial(ctx, remoteOfferAnswer.IceCredentials.UFrag, remoteOfferAnswer.IceCredentials.Pwd)
 	} else {
 		return agent.Accept(ctx, remoteOfferAnswer.IceCredentials.UFrag, remoteOfferAnswer.IceCredentials.Pwd)
 	}

client/internal/pkce_auth.go

@@ -44,6 +44,8 @@ type PKCEAuthProviderConfig struct {
 	DisablePromptLogin bool
 	// LoginFlag is used to configure the PKCE flow login behavior
 	LoginFlag common.LoginFlag
+	// LoginHint is used to pre-fill the email/username field during authentication
+	LoginHint string
 }
 
 // GetPKCEAuthorizationFlowInfo initialize a PKCEAuthorizationFlow instance and return with it

client/internal/relay/relay.go

@@ -2,6 +2,8 @@ package relay
 
 import (
 	"context"
+	"crypto/sha256"
+	"errors"
 	"fmt"
 	"net"
 	"sync"
@@ -15,6 +17,15 @@ import (
 	nbnet "github.com/netbirdio/netbird/client/net"
 )
 
+const (
+	DefaultCacheTTL = 20 * time.Second
+	probeTimeout    = 6 * time.Second
+)
+
+var (
+	ErrCheckInProgress = errors.New("probe check is already in progress")
+)
+
 // ProbeResult holds the info about the result of a relay probe request
 type ProbeResult struct {
 	URI  string
@@ -22,8 +33,164 @@ type ProbeResult struct {
 	Addr string
 }
 
+type StunTurnProbe struct {
+	cacheResults    []ProbeResult
+	cacheTimestamp  time.Time
+	cacheKey        string
+	cacheTTL        time.Duration
+	probeInProgress bool
+	probeDone       chan struct{}
+	mu              sync.Mutex
+}
+
+func NewStunTurnProbe(cacheTTL time.Duration) *StunTurnProbe {
+	return &StunTurnProbe{
+		cacheTTL: cacheTTL,
+	}
+}
+
+func (p *StunTurnProbe) ProbeAllWaitResult(ctx context.Context, stuns []*stun.URI, turns []*stun.URI) []ProbeResult {
+	cacheKey := generateCacheKey(stuns, turns)
+
+	p.mu.Lock()
+	if p.probeInProgress {
+		doneChan := p.probeDone
+		p.mu.Unlock()
+
+		select {
+		case <-ctx.Done():
+			log.Debugf("Context cancelled while waiting for probe results")
+			return createErrorResults(stuns, turns)
+		case <-doneChan:
+			return p.getCachedResults(cacheKey, stuns, turns)
+		}
+	}
+
+	p.probeInProgress = true
+	probeDone := make(chan struct{})
+	p.probeDone = probeDone
+	p.mu.Unlock()
+
+	p.doProbe(ctx, stuns, turns, cacheKey)
+	close(probeDone)
+
+	return p.getCachedResults(cacheKey, stuns, turns)
+}
+
+// ProbeAll probes all given servers asynchronously and returns the results
+func (p *StunTurnProbe) ProbeAll(ctx context.Context, stuns []*stun.URI, turns []*stun.URI) []ProbeResult {
+	cacheKey := generateCacheKey(stuns, turns)
+
+	p.mu.Lock()
+
+	if results := p.checkCache(cacheKey); results != nil {
+		p.mu.Unlock()
+		return results
+	}
+
+	if p.probeInProgress {
+		p.mu.Unlock()
+		return createErrorResults(stuns, turns)
+	}
+
+	p.probeInProgress = true
+	probeDone := make(chan struct{})
+	p.probeDone = probeDone
+	log.Infof("started new probe for STUN, TURN servers")
+	go func() {
+		p.doProbe(ctx, stuns, turns, cacheKey)
+		close(probeDone)
+	}()
+
+	p.mu.Unlock()
+
+	timer := time.NewTimer(1300 * time.Millisecond)
+	defer timer.Stop()
+
+	select {
+	case <-ctx.Done():
+		log.Debugf("Context cancelled while waiting for probe results")
+		return createErrorResults(stuns, turns)
+	case <-probeDone:
+		// when the probe is return fast, return the results right away
+		return p.getCachedResults(cacheKey, stuns, turns)
+	case <-timer.C:
+		// if the probe takes longer than 1.3s, return error results to avoid blocking
+		return createErrorResults(stuns, turns)
+	}
+}
+
+func (p *StunTurnProbe) checkCache(cacheKey string) []ProbeResult {
+	if p.cacheKey == cacheKey && len(p.cacheResults) > 0 {
+		age := time.Since(p.cacheTimestamp)
+		if age < p.cacheTTL {
+			results := append([]ProbeResult(nil), p.cacheResults...)
+			log.Debugf("returning cached probe results (age: %v)", age)
+			return results
+		}
+	}
+	return nil
+}
+
+func (p *StunTurnProbe) getCachedResults(cacheKey string, stuns []*stun.URI, turns []*stun.URI) []ProbeResult {
+	p.mu.Lock()
+	defer p.mu.Unlock()
+
+	if p.cacheKey == cacheKey && len(p.cacheResults) > 0 {
+		return append([]ProbeResult(nil), p.cacheResults...)
+	}
+	return createErrorResults(stuns, turns)
+}
+
+func (p *StunTurnProbe) doProbe(ctx context.Context, stuns []*stun.URI, turns []*stun.URI, cacheKey string) {
+	defer func() {
+		p.mu.Lock()
+		p.probeInProgress = false
+		p.mu.Unlock()
+	}()
+	results := make([]ProbeResult, len(stuns)+len(turns))
+
+	var wg sync.WaitGroup
+	for i, uri := range stuns {
+		wg.Add(1)
+		go func(idx int, stunURI *stun.URI) {
+			defer wg.Done()
+
+			probeCtx, cancel := context.WithTimeout(ctx, probeTimeout)
+			defer cancel()
+
+			results[idx].URI = stunURI.String()
+			results[idx].Addr, results[idx].Err = p.probeSTUN(probeCtx, stunURI)
+		}(i, uri)
+	}
+
+	stunOffset := len(stuns)
+	for i, uri := range turns {
+		wg.Add(1)
+		go func(idx int, turnURI *stun.URI) {
+			defer wg.Done()
+
+			probeCtx, cancel := context.WithTimeout(ctx, probeTimeout)
+			defer cancel()
+
+			results[idx].URI = turnURI.String()
+			results[idx].Addr, results[idx].Err = p.probeTURN(probeCtx, turnURI)
+		}(stunOffset+i, uri)
+	}
+
+	wg.Wait()
+
+	p.mu.Lock()
+	p.cacheResults = results
+	p.cacheTimestamp = time.Now()
+	p.cacheKey = cacheKey
+	p.mu.Unlock()
+
+	log.Debug("Stored new probe results in cache")
+}
+
 // ProbeSTUN tries binding to the given STUN uri and acquiring an address
-func ProbeSTUN(ctx context.Context, uri *stun.URI) (addr string, probeErr error) {
+func (p *StunTurnProbe) probeSTUN(ctx context.Context, uri *stun.URI) (addr string, probeErr error) {
 	defer func() {
 		if probeErr != nil {
 			log.Debugf("stun probe error from %s: %s", uri, probeErr)
@@ -83,7 +250,7 @@ func ProbeSTUN(ctx context.Context, uri *stun.URI) (addr string, probeErr error)
 }
 
 // ProbeTURN tries allocating a session from the given TURN URI
-func ProbeTURN(ctx context.Context, uri *stun.URI) (addr string, probeErr error) {
+func (p *StunTurnProbe) probeTURN(ctx context.Context, uri *stun.URI) (addr string, probeErr error) {
 	defer func() {
 		if probeErr != nil {
 			log.Debugf("turn probe error from %s: %s", uri, probeErr)
@@ -160,28 +327,28 @@ func ProbeTURN(ctx context.Context, uri *stun.URI) (addr string, probeErr error)
 	return relayConn.LocalAddr().String(), nil
 }
 
-// ProbeAll probes all given servers asynchronously and returns the results
-func ProbeAll(
-	ctx context.Context,
-	fn func(ctx context.Context, uri *stun.URI) (addr string, probeErr error),
-	relays []*stun.URI,
-) []ProbeResult {
-	results := make([]ProbeResult, len(relays))
+func createErrorResults(stuns []*stun.URI, turns []*stun.URI) []ProbeResult {
+	total := len(stuns) + len(turns)
+	results := make([]ProbeResult, total)
 
-	var wg sync.WaitGroup
-	for i, uri := range relays {
-		ctx, cancel := context.WithTimeout(ctx, 6*time.Second)
-		defer cancel()
-
-		wg.Add(1)
-		go func(res *ProbeResult, stunURI *stun.URI) {
-			defer wg.Done()
-			res.URI = stunURI.String()
-			res.Addr, res.Err = fn(ctx, stunURI)
-		}(&results[i], uri)
+	allURIs := append(append([]*stun.URI{}, stuns...), turns...)
+	for i, uri := range allURIs {
+		results[i] = ProbeResult{
+			URI: uri.String(),
+			Err: ErrCheckInProgress,
+		}
 	}
 
-	wg.Wait()
-
 	return results
 }
+
+func generateCacheKey(stuns []*stun.URI, turns []*stun.URI) string {
+	h := sha256.New()
+	for _, uri := range stuns {
+		h.Write([]byte(uri.String()))
+	}
+	for _, uri := range turns {
+		h.Write([]byte(uri.String()))
+	}
+	return fmt.Sprintf("%x", h.Sum(nil))
+}

client/internal/routemanager/common/params.go

@@ -1,6 +1,7 @@
 package common
 
 import (
+	"sync/atomic"
 	"time"
 
 	"github.com/netbirdio/netbird/client/firewall/manager"
@@ -25,4 +26,5 @@ type HandlerParams struct {
 	UseNewDNSRoute       bool
 	Firewall             manager.Manager
 	FakeIPManager        *fakeip.Manager
+	ForwarderPort        *atomic.Uint32
 }

client/internal/routemanager/dnsinterceptor/handler.go

@@ -8,6 +8,7 @@ import (
 	"runtime"
 	"strings"
 	"sync"
+	"sync/atomic"
 	"time"
 
 	"github.com/hashicorp/go-multierror"
@@ -18,7 +19,6 @@ import (
 	firewall "github.com/netbirdio/netbird/client/firewall/manager"
 	"github.com/netbirdio/netbird/client/iface/wgaddr"
 	nbdns "github.com/netbirdio/netbird/client/internal/dns"
-	"github.com/netbirdio/netbird/client/internal/dnsfwd"
 	"github.com/netbirdio/netbird/client/internal/peer"
 	"github.com/netbirdio/netbird/client/internal/peerstore"
 	"github.com/netbirdio/netbird/client/internal/routemanager/common"
@@ -55,6 +55,7 @@ type DnsInterceptor struct {
 	peerStore            *peerstore.Store
 	firewall             firewall.Manager
 	fakeIPManager        *fakeip.Manager
+	forwarderPort        *atomic.Uint32
 }
 
 func New(params common.HandlerParams) *DnsInterceptor {
@@ -69,6 +70,7 @@ func New(params common.HandlerParams) *DnsInterceptor {
 		firewall:             params.Firewall,
 		fakeIPManager:        params.FakeIPManager,
 		interceptedDomains:   make(domainMap),
+		forwarderPort:        params.ForwarderPort,
 	}
 }
 
@@ -257,7 +259,7 @@ func (d *DnsInterceptor) ServeDNS(w dns.ResponseWriter, r *dns.Msg) {
 		r.MsgHdr.AuthenticatedData = true
 	}
 
-	upstream := fmt.Sprintf("%s:%d", upstreamIP.String(), dnsfwd.ListenPort())
+	upstream := fmt.Sprintf("%s:%d", upstreamIP.String(), uint16(d.forwarderPort.Load()))
 	ctx, cancel := context.WithTimeout(context.Background(), dnsTimeout)
 	defer cancel()
 

client/internal/routemanager/manager.go

@@ -10,6 +10,7 @@ import (
 	"runtime"
 	"slices"
 	"sync"
+	"sync/atomic"
 	"time"
 
 	"github.com/google/uuid"
@@ -23,6 +24,7 @@ import (
 	"github.com/netbirdio/netbird/client/iface/netstack"
 	"github.com/netbirdio/netbird/client/internal/dns"
 	"github.com/netbirdio/netbird/client/internal/listener"
+	nbdns "github.com/netbirdio/netbird/dns"
 	"github.com/netbirdio/netbird/client/internal/peer"
 	"github.com/netbirdio/netbird/client/internal/peerstore"
 	"github.com/netbirdio/netbird/client/internal/routemanager/client"
@@ -54,6 +56,7 @@ type Manager interface {
 	SetRouteChangeListener(listener listener.NetworkChangeListener)
 	InitialRouteRange() []string
 	SetFirewall(firewall.Manager) error
+	SetDNSForwarderPort(port uint16)
 	Stop(stateManager *statemanager.Manager)
 }
 
@@ -78,6 +81,7 @@ type DefaultManager struct {
 	ctx                  context.Context
 	stop                 context.CancelFunc
 	mux                  sync.Mutex
+	shutdownWg           sync.WaitGroup
 	clientNetworks       map[route.HAUniqueID]*client.Watcher
 	routeSelector        *routeselector.RouteSelector
 	serverRouter         *server.Router
@@ -101,12 +105,13 @@ type DefaultManager struct {
 	disableServerRoutes bool
 	activeRoutes        map[route.HAUniqueID]client.RouteHandler
 	fakeIPManager       *fakeip.Manager
+	dnsForwarderPort    atomic.Uint32
 }
 
 func NewManager(config ManagerConfig) *DefaultManager {
 	mCTX, cancel := context.WithCancel(config.Context)
 	notifier := notifier.NewNotifier()
-	sysOps := systemops.NewSysOps(config.WGInterface, notifier)
+	sysOps := systemops.New(config.WGInterface, notifier)
 
 	if runtime.GOOS == "windows" && config.WGInterface != nil {
 		nbnet.SetVPNInterfaceName(config.WGInterface.Name())
@@ -130,6 +135,7 @@ func NewManager(config ManagerConfig) *DefaultManager {
 		disableServerRoutes: config.DisableServerRoutes,
 		activeRoutes:        make(map[route.HAUniqueID]client.RouteHandler),
 	}
+	dm.dnsForwarderPort.Store(uint32(nbdns.ForwarderClientPort))
 
 	useNoop := netstack.IsEnabled() || config.DisableClientRoutes
 	dm.setupRefCounters(useNoop)
@@ -270,9 +276,15 @@ func (m *DefaultManager) SetFirewall(firewall firewall.Manager) error {
 	return nil
 }
 
+// SetDNSForwarderPort sets the DNS forwarder port for route handlers
+func (m *DefaultManager) SetDNSForwarderPort(port uint16) {
+	m.dnsForwarderPort.Store(uint32(port))
+}
+
 // Stop stops the manager watchers and clean firewall rules
 func (m *DefaultManager) Stop(stateManager *statemanager.Manager) {
 	m.stop()
+	m.shutdownWg.Wait()
 	if m.serverRouter != nil {
 		m.serverRouter.CleanUp()
 	}
@@ -345,6 +357,7 @@ func (m *DefaultManager) updateSystemRoutes(newRoutes route.HAMap) error {
 			UseNewDNSRoute:       m.useNewDNSRoute,
 			Firewall:             m.firewall,
 			FakeIPManager:        m.fakeIPManager,
+			ForwarderPort:        &m.dnsForwarderPort,
 		}
 		handler := client.HandlerFromRoute(params)
 		if err := handler.AddRoute(m.ctx); err != nil {
@@ -474,7 +487,11 @@ func (m *DefaultManager) TriggerSelection(networks route.HAMap) {
 		}
 		clientNetworkWatcher := client.NewWatcher(config)
 		m.clientNetworks[id] = clientNetworkWatcher
-		go clientNetworkWatcher.Start()
+		m.shutdownWg.Add(1)
+		go func() {
+			defer m.shutdownWg.Done()
+			clientNetworkWatcher.Start()
+		}()
 		clientNetworkWatcher.SendUpdate(client.RoutesUpdate{Routes: routes})
 	}
 
@@ -516,7 +533,11 @@ func (m *DefaultManager) updateClientNetworks(updateSerial uint64, networks rout
 			}
 			clientNetworkWatcher = client.NewWatcher(config)
 			m.clientNetworks[id] = clientNetworkWatcher
-			go clientNetworkWatcher.Start()
+			m.shutdownWg.Add(1)
+			go func() {
+				defer m.shutdownWg.Done()
+				clientNetworkWatcher.Start()
+			}()
 		}
 		update := client.RoutesUpdate{
 			UpdateSerial: updateSerial,

client/internal/routemanager/mock.go

@@ -90,6 +90,10 @@ func (m *MockManager) SetFirewall(firewall.Manager) error {
 	panic("implement me")
 }
 
+// SetDNSForwarderPort mock implementation of SetDNSForwarderPort from Manager interface
+func (m *MockManager) SetDNSForwarderPort(port uint16) {
+}
+
 // Stop mock implementation of Stop from Manager interface
 func (m *MockManager) Stop(stateManager *statemanager.Manager) {
 	if m.StopFunc != nil {

client/internal/routemanager/systemops/flush_nonbsd.go

@@ -0,0 +1,8 @@
+//go:build !((darwin && !ios) || dragonfly || freebsd || netbsd || openbsd)
+
+package systemops
+
+// FlushMarkedRoutes is a no-op on non-BSD platforms.
+func (r *SysOps) FlushMarkedRoutes() error {
+	return nil
+}

client/internal/routemanager/systemops/state.go

@@ -13,11 +13,11 @@ func (s *ShutdownState) Name() string {
 }
 
 func (s *ShutdownState) Cleanup() error {
-	sysops := NewSysOps(nil, nil)
-	sysops.refCounter = refcounter.New[netip.Prefix, struct{}, Nexthop](nil, sysops.removeFromRouteTable)
-	sysops.refCounter.LoadData((*ExclusionCounter)(s))
+	sysOps := New(nil, nil)
+	sysOps.refCounter = refcounter.New[netip.Prefix, struct{}, Nexthop](nil, sysOps.removeFromRouteTable)
+	sysOps.refCounter.LoadData((*ExclusionCounter)(s))
 
-	return sysops.refCounter.Flush()
+	return sysOps.refCounter.Flush()
 }
 
 func (s *ShutdownState) MarshalJSON() ([]byte, error) {

client/internal/routemanager/systemops/systemops.go

@@ -83,7 +83,7 @@ type SysOps struct {
 	localSubnetsCacheTime time.Time
 }
 
-func NewSysOps(wgInterface wgIface, notifier *notifier.Notifier) *SysOps {
+func New(wgInterface wgIface, notifier *notifier.Notifier) *SysOps {
 	return &SysOps{
 		wgInterface: wgInterface,
 		notifier:    notifier,

client/internal/routemanager/systemops/systemops_bsd_test.go

@@ -42,7 +42,7 @@ func TestConcurrentRoutes(t *testing.T) {
 	_, intf = setupDummyInterface(t)
 	nexthop = Nexthop{netip.Addr{}, intf}
 
-	r := NewSysOps(nil, nil)
+	r := New(nil, nil)
 
 	var wg sync.WaitGroup
 	for i := 0; i < 1024; i++ {
@@ -146,7 +146,7 @@ func createAndSetupDummyInterface(t *testing.T, intf string, ipAddressCIDR strin
 
 	nexthop := Nexthop{netip.Addr{}, netIntf}
 
-	r := NewSysOps(nil, nil)
+	r := New(nil, nil)
 	err = r.addToRouteTable(prefix, nexthop)
 	require.NoError(t, err, "Failed to add route to table")
 

client/internal/routemanager/systemops/systemops_generic_test.go

@@ -143,7 +143,7 @@ func TestAddVPNRoute(t *testing.T) {
 
 			wgInterface := createWGInterface(t, fmt.Sprintf("utun53%d", n), "100.65.75.2/24", 33100+n)
 
-			r := NewSysOps(wgInterface, nil)
+			r := New(wgInterface, nil)
 			advancedRouting := nbnet.AdvancedRouting()
 			err := r.SetupRouting(nil, nil, advancedRouting)
 			require.NoError(t, err)
@@ -342,7 +342,7 @@ func TestAddRouteToNonVPNIntf(t *testing.T) {
 
 			wgInterface := createWGInterface(t, fmt.Sprintf("utun54%d", n), "100.65.75.2/24", 33200+n)
 
-			r := NewSysOps(wgInterface, nil)
+			r := New(wgInterface, nil)
 			advancedRouting := nbnet.AdvancedRouting()
 			err := r.SetupRouting(nil, nil, advancedRouting)
 			require.NoError(t, err)
@@ -486,7 +486,7 @@ func setupTestEnv(t *testing.T) {
 		assert.NoError(t, wgInterface.Close())
 	})
 
-	r := NewSysOps(wgInterface, nil)
+	r := New(wgInterface, nil)
 	advancedRouting := nbnet.AdvancedRouting()
 	err := r.SetupRouting(nil, nil, advancedRouting)
 	require.NoError(t, err, "setupRouting should not return err")

client/internal/routemanager/systemops/systemops_unix.go

@@ -7,19 +7,39 @@ import (
 	"fmt"
 	"net"
 	"net/netip"
+	"os"
 	"strconv"
 	"syscall"
 	"time"
 	"unsafe"
 
 	"github.com/cenkalti/backoff/v4"
+	"github.com/hashicorp/go-multierror"
 	log "github.com/sirupsen/logrus"
 	"golang.org/x/net/route"
 	"golang.org/x/sys/unix"
 
+	nberrors "github.com/netbirdio/netbird/client/errors"
 	"github.com/netbirdio/netbird/client/internal/statemanager"
 )
 
+const (
+	envRouteProtoFlag = "NB_ROUTE_PROTO_FLAG"
+)
+
+var routeProtoFlag int
+
+func init() {
+	switch os.Getenv(envRouteProtoFlag) {
+	case "2":
+		routeProtoFlag = unix.RTF_PROTO2
+	case "3":
+		routeProtoFlag = unix.RTF_PROTO3
+	default:
+		routeProtoFlag = unix.RTF_PROTO1
+	}
+}
+
 func (r *SysOps) SetupRouting(initAddresses []net.IP, stateManager *statemanager.Manager, advancedRouting bool) error {
 	return r.setupRefCounter(initAddresses, stateManager)
 }
@@ -28,6 +48,62 @@ func (r *SysOps) CleanupRouting(stateManager *statemanager.Manager, advancedRout
 	return r.cleanupRefCounter(stateManager)
 }
 
+// FlushMarkedRoutes removes single IP exclusion routes marked with the configured RTF_PROTO flag.
+func (r *SysOps) FlushMarkedRoutes() error {
+	rib, err := retryFetchRIB()
+	if err != nil {
+		return fmt.Errorf("fetch routing table: %w", err)
+	}
+
+	msgs, err := route.ParseRIB(route.RIBTypeRoute, rib)
+	if err != nil {
+		return fmt.Errorf("parse routing table: %w", err)
+	}
+
+	var merr *multierror.Error
+	flushedCount := 0
+
+	for _, msg := range msgs {
+		rtMsg, ok := msg.(*route.RouteMessage)
+		if !ok {
+			continue
+		}
+
+		if rtMsg.Flags&routeProtoFlag == 0 {
+			continue
+		}
+
+		routeInfo, err := MsgToRoute(rtMsg)
+		if err != nil {
+			log.Debugf("Skipping route flush: %v", err)
+			continue
+		}
+
+		if !routeInfo.Dst.IsValid() || !routeInfo.Dst.IsSingleIP() {
+			continue
+		}
+
+		nexthop := Nexthop{
+			IP:   routeInfo.Gw,
+			Intf: routeInfo.Interface,
+		}
+
+		if err := r.removeFromRouteTable(routeInfo.Dst, nexthop); err != nil {
+			merr = multierror.Append(merr, fmt.Errorf("remove route %s: %w", routeInfo.Dst, err))
+			continue
+		}
+
+		flushedCount++
+		log.Debugf("Flushed marked route: %s", routeInfo.Dst)
+	}
+
+	if flushedCount > 0 {
+		log.Infof("Flushed %d residual NetBird routes from previous session", flushedCount)
+	}
+
+	return nberrors.FormatErrorOrNil(merr)
+}
+
 func (r *SysOps) addToRouteTable(prefix netip.Prefix, nexthop Nexthop) error {
 	return r.routeSocket(unix.RTM_ADD, prefix, nexthop)
 }
@@ -105,7 +181,7 @@ func (r *SysOps) routeOp(action int, prefix netip.Prefix, nexthop Nexthop) func(
 func (r *SysOps) buildRouteMessage(action int, prefix netip.Prefix, nexthop Nexthop) (msg *route.RouteMessage, err error) {
 	msg = &route.RouteMessage{
 		Type:    action,
-		Flags:   unix.RTF_UP,
+		Flags:   unix.RTF_UP | routeProtoFlag,
 		Version: unix.RTM_VERSION,
 		Seq:     r.getSeq(),
 	}

client/internal/routeselector/routeselector.go

@@ -9,8 +9,6 @@ import (
 	"github.com/hashicorp/go-multierror"
 	"golang.org/x/exp/maps"
 
-	log "github.com/sirupsen/logrus"
-
 	"github.com/netbirdio/netbird/client/errors"
 	"github.com/netbirdio/netbird/route"
 )
@@ -128,13 +126,11 @@ func (rs *RouteSelector) IsSelected(routeID route.NetID) bool {
 	defer rs.mu.RUnlock()
 
 	if rs.deselectAll {
-		log.Debugf("Route %s not selected (deselect all)", routeID)
 		return false
 	}
 
 	_, deselected := rs.deselectedRoutes[routeID]
 	isSelected := !deselected
-	log.Debugf("Route %s selection status: %v (deselected: %v)", routeID, isSelected, deselected)
 	return isSelected
 }
 

client/internal/statemanager/manager.go

@@ -295,7 +295,7 @@ func (m *Manager) loadStateFile(deleteCorrupt bool) (map[string]json.RawMessage,
 	data, err := os.ReadFile(m.filePath)
 	if err != nil {
 		if errors.Is(err, fs.ErrNotExist) {
-			log.Debug("state file does not exist")
+			log.Debugf("state file %s does not exist", m.filePath)
 			return nil, nil // nolint:nilnil
 		}
 		return nil, fmt.Errorf("read state file: %w", err)