Merge main into ui-refactor; port MDM support to the Wails UI

Integrates main's MDM configuration-profile feature and adapts it to the
Wails UI (this branch had already replaced the Fyne UI).

Conflict resolution:
- go.mod/go.sum: take main's deps; howett.net/plist pinned to v1.0.2-... (tidy)
- client/proto/daemon.pb.go: regenerated from the merged daemon.proto
- client/internal/peer/status.go: union of ipToKey (main) + sessionExpiresAt (HEAD)
- client/server/server.go: main's intent/liveness model (connectionGoroutineRunning,
  clientRunning no longer cleared by the goroutine) + empty-PSK guard
- client/ui/client_ui.go, client/ui/profile.go: removed (dead Fyne UI)

MDM port (backend + tray):
- services/settings.go: expose MDMManagedFields plus a managedFields map keyed
  by Config field names so the settings form can gate a control without
  translating mdm.Key* names
- tray: gate Profiles / Exit Node menus on DisableProfiles / DisableNetworks via
  GetFeatures, refreshed on the config_changed system event (replaces the legacy
  2s poll); localized MDM policy-applied toast in all shipped locales
- client/proto/metadata.go: shared constants for the config_changed /
  policy_applied event markers

PreSharedKey: GetConfig now returns preSharedKeySet (bool) instead of the masked
value; the settings form provides its own placeholder and sends a new key only
when the user types one.
This commit is contained in:
Zoltan Papp
2026-06-12 15:27:23 +02:00
56 changed files with 3639 additions and 314 deletions

422
client/server/mdm.go Normal file
View File

@@ -0,0 +1,422 @@
package server
import (
"context"
"fmt"
"time"
log "github.com/sirupsen/logrus"
"google.golang.org/grpc/codes"
gstatus "google.golang.org/grpc/status"
"github.com/netbirdio/netbird/client/mdm"
"github.com/netbirdio/netbird/client/proto"
)
// preSharedKeyRedactedSentinel is the value GetConfig returns in place
// of an actual PSK, so a UI that round-trips the field back to the
// daemon (via SetConfig / Login) can be distinguished from a deliberate
// override. Any incoming PSK that equals this sentinel is treated as
// a no-op echo, never as a conflict with the policy.
const preSharedKeyRedactedSentinel = "**********"
// loadMDMPolicy is the indirection used by server handlers to read the
// active MDM policy. Tests override this to inject a fake policy.
var loadMDMPolicy = mdm.LoadPolicy
// conflictCheck is a value-aware comparison between a single field in
// the incoming request and the corresponding MDM-enforced value. It
// runs only when the field was actually set in the request (presence
// already filtered upstream); ok=true reports the policy value, ok=false
// means the policy is silent on the key — both are treated as conflicts
// to be safe (an MDM key declared as managed must hold a value).
type conflictCheck struct {
key string
check func(*mdm.Policy) (match bool)
}
// onMDMPolicyChange is invoked by the MDM reload ticker every time the
// OS-native managed-config store reports a diff vs the last observation.
//
// Restart sequence:
// 1. Cancel the active engine context (terminates connectWithRetryRuns).
// 2. Wait briefly for that goroutine to exit (giveUpChan is closed on exit).
// 3. Re-resolve Config from disk + MDM policy (Config.apply re-runs
// applyMDMPolicy with the freshly loaded Policy).
// 4. Spawn a fresh connectWithRetryRuns with the new context and config.
// 5. Broadcast a SystemEvent so any GUI / CLI subscriber (SubscribeEvents
// RPC) can refresh its cached config view without polling.
//
// The callback runs in the ticker's own goroutine. Ticker has already
// logged the per-key diff before invoking this hook.
func (s *Server) onMDMPolicyChange(_, _ *mdm.Policy) error {
log.Warn("MDM policy changed; restarting engine to apply new configuration")
// Hold s.mutex for the entire restart sequence (cancel + quiescence
// wait + re-spawn). Any concurrent Up/Down/Status arriving while
// MDM is restarting blocks on the Lock until we are done — they
// then observe the post-restart state coherently. This is safe
// because the connectWithRetryRuns goroutine no longer acquires
// s.mutex in its defer (intent vs. goroutine-alive concerns are
// fully separated; see the connectionGoroutineRunning helper).
s.mutex.Lock()
defer s.mutex.Unlock()
if !s.clientRunning {
// The client is not running, so there's no engine to restart.
return nil
}
if s.actCancel != nil {
s.actCancel()
}
// Wait for previous connectWithRetryRuns to exit so we don't end up
// with two goroutines fighting over the same status recorder + engine.
// The teardown engages a fan-out of engine goroutines (peer workers,
// signal handler, route manager, ...). close(clientGiveUpChan)
// happens in the function-scope defer of connectWithRetryRuns, on
// every exit path (ctx cancel, backoff exhausted, panic) — see the
// defer in server.go.
if s.clientGiveUpChan != nil {
select {
case <-s.clientGiveUpChan:
case <-time.After(10 * time.Second):
return fmt.Errorf("failed to restart the engine due to timeout")
}
}
if err := s.restartEngineForMDMLocked(); err != nil {
log.Errorf("MDM restart failed: %v", err)
return err
}
// publishConfigChangedEvent has already fired inside
// restartEngineForMDMLocked with source="mdm". Emit an MDM-specific
// user-visible toast so the operator knows their IT policy was
// applied (UserMessage != "" triggers the GUI notifier).
s.statusRecorder.PublishEvent(
proto.SystemEvent_INFO,
proto.SystemEvent_SYSTEM,
"MDM policy applied",
"NetBird configuration was updated by your IT policy.",
map[string]string{
proto.MetadataSourceKey: proto.MetadataSourceMDM,
proto.MetadataTypeKey: proto.MetadataTypePolicyApplied,
},
)
return nil
}
// publishConfigChangedEvent broadcasts a SystemEvent informing any active
// SubscribeEvents subscriber (typically the GUI tray) that the daemon's
// effective Config has been replaced and any cached client-side view
// should be refreshed. Callers pass a stable `source` label so the GUI
// can distinguish a startup spawn from a user-triggered Up or an
// MDM-driven restart. Reusing the SYSTEM category keeps the proto enum
// stable; metadata.type="config_changed" routes to the GUI's refresh
// handler. UserMessage is left empty so the system tray does not toast
// for every internal restart; the MDM path emits a separate
// "policy_applied" event (with UserMessage) for that purpose.
func (s *Server) publishConfigChangedEvent(source string) {
if s.statusRecorder == nil {
return
}
s.statusRecorder.PublishEvent(
proto.SystemEvent_INFO,
proto.SystemEvent_SYSTEM,
fmt.Sprintf("daemon config changed (source=%s)", source),
"",
map[string]string{
proto.MetadataSourceKey: source,
proto.MetadataTypeKey: proto.MetadataTypeConfigChanged,
},
)
}
// restartEngineForMDMLocked re-resolves the active profile config
// (re-running applyMDMPolicy via Config.apply) and re-spawns
// connectWithRetryRuns. Mirrors the tail of Server.Start so a runtime
// MDM change behaves identically to a fresh boot under the new policy.
//
// MUST be called with s.mutex held — onMDMPolicyChange holds the lock
// for the entire restart sequence (cancel + quiescence wait + re-spawn)
// so concurrent Up/Down/Status RPCs observe a coherent post-restart
// state.
func (s *Server) restartEngineForMDMLocked() error {
activeProf, err := s.profileManager.GetActiveProfileState()
if err != nil {
return fmt.Errorf("get active profile state: %w", err)
}
config, _, err := s.getConfig(activeProf)
if err != nil {
return fmt.Errorf("get active profile config: %w", err)
}
s.config = config
s.statusRecorder.UpdateManagementAddress(config.ManagementURL.String())
s.statusRecorder.UpdateRosenpass(config.RosenpassEnabled, config.RosenpassPermissive)
s.statusRecorder.UpdateLazyConnection(config.LazyConnectionEnabled)
ctx, cancel := context.WithCancel(s.rootCtx)
s.actCancel = cancel
s.clientRunning = true
s.clientRunningChan = make(chan struct{})
s.clientGiveUpChan = make(chan struct{})
log.Info("MDM restart: spawning connectWithRetryRuns with re-resolved config")
go s.connectWithRetryRuns(ctx, config, s.statusRecorder, s.clientRunningChan, s.clientGiveUpChan)
s.publishConfigChangedEvent(proto.MetadataSourceMDM)
return nil
}
// conflictBool builds a conflictCheck for a boolean MDM key. If p is nil
// the field is treated as matching (no override requested); otherwise the
// check returns true only when the policy contains the key and its
// boolean value equals *p.
func conflictBool(key string, p *bool) conflictCheck {
return conflictCheck{
key: key,
check: func(pol *mdm.Policy) bool {
if p == nil {
return true // absent → match by definition
}
want, ok := pol.GetBool(key)
return ok && want == *p
},
}
}
// conflictString builds a conflictCheck for a string MDM key. An empty
// `got` is treated as "field not set" (no override requested); otherwise
// the check returns true only when the policy contains the key and its
// value equals got.
func conflictString(key, got string) conflictCheck {
return conflictCheck{
key: key,
check: func(pol *mdm.Policy) bool {
if got == "" {
return true
}
want, ok := pol.GetString(key)
return ok && want == got
},
}
}
// conflictInt64 builds a conflictCheck for an integer MDM key. If p is
// nil the field is treated as matching; otherwise the check returns
// true only when the policy contains the key and its int value equals *p.
func conflictInt64(key string, p *int64) conflictCheck {
return conflictCheck{
key: key,
check: func(pol *mdm.Policy) bool {
if p == nil {
return true
}
want, ok := pol.GetInt(key)
return ok && want == *p
},
}
}
// resolveConflicts walks the per-field checks against the active MDM
// policy and returns the names of keys whose requested value diverges
// from the policy-enforced value. Keys not present in the policy are
// skipped silently (the gate fires only for keys the admin has
// actually pushed). Returns nil for an empty policy.
func resolveConflicts(policy *mdm.Policy, checks []conflictCheck) []string {
if policy.IsEmpty() {
return nil
}
var conflicts []string
for _, c := range checks {
if !policy.HasKey(c.key) {
continue
}
if !c.check(policy) {
conflicts = append(conflicts, c.key)
}
}
return conflicts
}
// mdmManagedFieldConflicts returns the names of MDM-managed keys whose
// requested value in the SetConfigRequest differs from the MDM-enforced
// value. A field set to the same value the policy already enforces is
// treated as a no-op echo (the GUI tray sends a full Config snapshot on
// every toggle, so most fields in a typical request match the policy
// exactly and must NOT be flagged as conflicts). The redacted PSK
// sentinel ("**********") returned by GetConfig is recognised and
// treated as no-op so the UI can safely round-trip it.
func mdmManagedFieldConflicts(msg *proto.SetConfigRequest, policy *mdm.Policy) []string {
if msg == nil {
return nil
}
// PSK round-trip echo: collapse the sentinel to empty so the
// shared check treats it as "field not set".
pskGot := ""
if msg.OptionalPreSharedKey != nil && *msg.OptionalPreSharedKey != preSharedKeyRedactedSentinel {
pskGot = *msg.OptionalPreSharedKey
}
return resolveConflicts(policy, []conflictCheck{
conflictString(mdm.KeyManagementURL, msg.ManagementUrl),
conflictString(mdm.KeyPreSharedKey, pskGot),
conflictBool(mdm.KeyRosenpassEnabled, msg.RosenpassEnabled),
conflictBool(mdm.KeyRosenpassPermissive, msg.RosenpassPermissive),
conflictBool(mdm.KeyDisableAutoConnect, msg.DisableAutoConnect),
conflictBool(mdm.KeyAllowServerSSH, msg.ServerSSHAllowed),
conflictBool(mdm.KeyDisableClientRoutes, msg.DisableClientRoutes),
conflictBool(mdm.KeyDisableServerRoutes, msg.DisableServerRoutes),
conflictBool(mdm.KeyBlockInbound, msg.BlockInbound),
conflictInt64(mdm.KeyWireguardPort, msg.WireguardPort),
})
}
// setConfigRequestHasConfigOverrides reports whether the SetConfigRequest
// carries ANY field that would actually mutate the persisted config.
// The CLI builds a SetConfigRequest unconditionally on every
// `netbird up` (see setupSetConfigReq in cmd/up.go) — a plain
// `netbird up` produces a request with every field at its zero value;
// the gate must skip such no-op invocations or it would always fire
// even when the user did not pass any --flag. Returns false on a nil
// msg; true when any management/admin URL, PSK, DNS/NAT list+clean
// flag, interface/port/MTU, or any optional bool/duration field is set.
func setConfigRequestHasConfigOverrides(msg *proto.SetConfigRequest) bool {
if msg == nil {
return false
}
return msg.ManagementUrl != "" ||
msg.AdminURL != "" ||
msg.OptionalPreSharedKey != nil ||
len(msg.CustomDNSAddress) > 0 ||
len(msg.NatExternalIPs) > 0 || msg.CleanNATExternalIPs ||
len(msg.ExtraIFaceBlacklist) > 0 ||
len(msg.DnsLabels) > 0 || msg.CleanDNSLabels ||
msg.DnsRouteInterval != nil ||
msg.RosenpassEnabled != nil ||
msg.RosenpassPermissive != nil ||
msg.InterfaceName != nil ||
msg.WireguardPort != nil ||
msg.Mtu != nil ||
msg.DisableAutoConnect != nil ||
msg.ServerSSHAllowed != nil ||
msg.NetworkMonitor != nil ||
msg.DisableClientRoutes != nil ||
msg.DisableServerRoutes != nil ||
msg.DisableDns != nil ||
msg.DisableFirewall != nil ||
msg.BlockLanAccess != nil ||
msg.DisableNotifications != nil ||
msg.LazyConnectionEnabled != nil ||
msg.BlockInbound != nil ||
msg.DisableIpv6 != nil ||
msg.EnableSSHRoot != nil ||
msg.EnableSSHSFTP != nil ||
msg.EnableSSHLocalPortForwarding != nil ||
msg.EnableSSHRemotePortForwarding != nil ||
msg.DisableSSHAuth != nil ||
msg.SshJWTCacheTTL != nil
}
// loginRequestHasConfigOverrides reports whether the LoginRequest
// carries ANY field that would mutate persisted daemon configuration
// (as opposed to pure-auth fields like setupKey, hostname, hint,
// profileName, username). Used by the Login handler to decide whether
// the `--disable-update-settings` / MDM gates must run: a re-auth that
// changes nothing about the configuration is always allowed.
func loginRequestHasConfigOverrides(msg *proto.LoginRequest) bool {
if msg == nil {
return false
}
return msg.ManagementUrl != "" ||
msg.AdminURL != "" ||
msg.PreSharedKey != "" || //nolint:staticcheck // SA1019: legacy proto field still accepted by Login
msg.OptionalPreSharedKey != nil ||
len(msg.CustomDNSAddress) > 0 ||
len(msg.NatExternalIPs) > 0 || msg.CleanNATExternalIPs ||
msg.RosenpassEnabled != nil ||
msg.InterfaceName != nil ||
msg.WireguardPort != nil ||
msg.DisableAutoConnect != nil ||
msg.ServerSSHAllowed != nil ||
msg.RosenpassPermissive != nil ||
len(msg.ExtraIFaceBlacklist) > 0 ||
msg.NetworkMonitor != nil ||
msg.DnsRouteInterval != nil ||
msg.DisableClientRoutes != nil ||
msg.DisableServerRoutes != nil ||
msg.DisableDns != nil ||
msg.DisableFirewall != nil ||
msg.BlockLanAccess != nil ||
msg.DisableNotifications != nil ||
len(msg.DnsLabels) > 0 || msg.CleanDNSLabels ||
msg.LazyConnectionEnabled != nil ||
msg.BlockInbound != nil
}
// loginRequestMDMConflicts mirrors mdmManagedFieldConflicts but for the
// LoginRequest surface. Same value-aware semantics: a field set to the
// MDM-enforced value is a no-op echo, not a conflict; only a divergent
// value is flagged. PSK has two proto fields — PreSharedKey (deprecated)
// and OptionalPreSharedKey (current); either route trips the gate if it
// diverges from the MDM-enforced PSK. OptionalPreSharedKey wins when
// both are set; the redaction sentinel ("**********") is accepted as
// a no-op echo.
func loginRequestMDMConflicts(msg *proto.LoginRequest, policy *mdm.Policy) []string {
if msg == nil {
return nil
}
// Collapse the two PSK fields + the redaction sentinel down to a
// single "got" string the shared check can compare against the
// policy: OptionalPreSharedKey wins if set; PreSharedKey (deprecated)
// is the fallback; sentinel echo is treated as "field not set".
pskGot := ""
if msg.OptionalPreSharedKey != nil {
pskGot = *msg.OptionalPreSharedKey
} else if msg.PreSharedKey != "" { //nolint:staticcheck // SA1019: legacy proto field still accepted by Login
pskGot = msg.PreSharedKey //nolint:staticcheck // SA1019
}
if pskGot == preSharedKeyRedactedSentinel {
pskGot = ""
}
return resolveConflicts(policy, []conflictCheck{
conflictString(mdm.KeyManagementURL, msg.ManagementUrl),
conflictString(mdm.KeyPreSharedKey, pskGot),
conflictBool(mdm.KeyRosenpassEnabled, msg.RosenpassEnabled),
conflictBool(mdm.KeyRosenpassPermissive, msg.RosenpassPermissive),
conflictBool(mdm.KeyDisableAutoConnect, msg.DisableAutoConnect),
conflictBool(mdm.KeyAllowServerSSH, msg.ServerSSHAllowed),
conflictBool(mdm.KeyDisableClientRoutes, msg.DisableClientRoutes),
conflictBool(mdm.KeyDisableServerRoutes, msg.DisableServerRoutes),
conflictBool(mdm.KeyBlockInbound, msg.BlockInbound),
conflictInt64(mdm.KeyWireguardPort, msg.WireguardPort),
})
}
// rejectMDMManagedFieldConflicts returns a FailedPrecondition gRPC error
// with an MDMManagedFieldsViolation detail when any of the requested
// fields tries to change an MDM-enforced value to something else, and
// nil otherwise. The whole request is rejected on any conflict; non-
// conflicting fields in the same request are not applied either (no
// partial apply).
func rejectMDMManagedFieldConflicts(conflicts []string) error {
if len(conflicts) == 0 {
return nil
}
log.Warnf("MDM rejected request: tried to modify %d managed key(s): %v",
len(conflicts), conflicts)
st := gstatus.New(
codes.FailedPrecondition,
fmt.Sprintf("fields managed by MDM cannot be modified: %v", conflicts),
)
detailed, err := st.WithDetails(&proto.MDMManagedFieldsViolation{Fields: conflicts})
if err != nil {
// Detail attachment is best-effort; fall back to the plain status
// so the caller still gets a usable FailedPrecondition.
return st.Err()
}
return detailed.Err()
}

View File

@@ -30,7 +30,7 @@ func (s *Server) ListNetworks(context.Context, *proto.ListNetworksRequest) (*pro
s.mutex.Lock()
defer s.mutex.Unlock()
if s.networksDisabled {
if s.checkNetworksDisabled() {
return nil, gstatus.Errorf(codes.Unavailable, errNetworksDisabled)
}
@@ -143,7 +143,7 @@ func (s *Server) SelectNetworks(_ context.Context, req *proto.SelectNetworksRequ
s.mutex.Lock()
defer s.mutex.Unlock()
if s.networksDisabled {
if s.checkNetworksDisabled() {
return nil, gstatus.Errorf(codes.Unavailable, errNetworksDisabled)
}
@@ -206,7 +206,7 @@ func (s *Server) DeselectNetworks(_ context.Context, req *proto.SelectNetworksRe
s.mutex.Lock()
defer s.mutex.Unlock()
if s.networksDisabled {
if s.checkNetworksDisabled() {
return nil, gstatus.Errorf(codes.Unavailable, errNetworksDisabled)
}

View File

@@ -25,6 +25,7 @@ import (
"github.com/netbirdio/netbird/client/internal/expose"
"github.com/netbirdio/netbird/client/internal/profilemanager"
sleephandler "github.com/netbirdio/netbird/client/internal/sleep/handler"
"github.com/netbirdio/netbird/client/mdm"
"github.com/netbirdio/netbird/client/system"
mgm "github.com/netbirdio/netbird/shared/management/client"
"github.com/netbirdio/netbird/shared/management/domain"
@@ -84,7 +85,13 @@ type Server struct {
mutex sync.Mutex
config *profilemanager.Config
proto.UnimplementedDaemonServiceServer
clientRunning bool // protected by mutex
// clientRunning tracks "the daemon wants to be connected" — set true by
// Start / Up, cleared by Down / Logout. Persists across retry
// loops, signal disconnects, and ErrResetConnection cycles. NOT
// changed by connectWithRetryRuns goroutine exit — for that
// (goroutine-still-alive) check, see connectionGoroutineRunning() which
// derives from clientGiveUpChan close state. Protected by s.mutex.
clientRunning bool
clientRunningChan chan struct{}
clientGiveUpChan chan struct{} // closed when connectWithRetryRuns goroutine exits
@@ -111,6 +118,11 @@ type Server struct {
sleepHandler *sleephandler.SleepHandler
// mdmTicker periodically re-reads the OS-native MDM policy and triggers
// an engine restart when the policy changes. Launched once by Start;
// stopped by the rootCtx cancellation.
mdmTicker *mdm.Ticker
updateManager *updater.Manager
jwtCache *jwtCache
@@ -179,6 +191,17 @@ func (s *Server) Start() error {
s.updateManager.CheckUpdateSuccess(s.rootCtx)
}
// MDM policy reload ticker: every minute the desktop daemon re-reads
// the OS-native managed-config store and, on diff vs the previous
// observation, cancels the active engine context so connectWithRetry-
// Runs re-resolves Config (re-running profilemanager.Config.apply which
// applies the freshly-read MDM policy as the last layer) and brings
// the engine back with the new values.
if s.mdmTicker == nil {
s.mdmTicker = mdm.NewTicker(mdm.DefaultReloadInterval)
go s.mdmTicker.Run(s.rootCtx, s.onMDMPolicyChange)
}
// if current state contains any error, return it
// in all other cases we can continue execution only if status is idle and up command was
// not in the progress or already successfully established connection.
@@ -237,24 +260,28 @@ func (s *Server) Start() error {
s.clientRunningChan = make(chan struct{})
s.clientGiveUpChan = make(chan struct{})
go s.connectWithRetryRuns(ctx, config, s.statusRecorder, s.clientRunningChan, s.clientGiveUpChan)
s.publishConfigChangedEvent(proto.MetadataSourceStartup)
return nil
}
// connectWithRetryRuns runs the client connection with a backoff strategy where we retry the operation as additional
// mechanism to keep the client connected even when the connection is lost.
// we cancel retry if the client receive a stop or down command, or if disable auto connect is configured.
//
// The goroutine's exit is signalled to the daemon via close(giveUpChan)
// — placed in the function-scope defer so every return path (panic,
// DisableAutoConnect early-exit, backoff exhausted, ctx cancel) closes
// it. Callers that need to observe "is the goroutine still alive?" use
// Server.connectionGoroutineRunning() which non-blockingly checks the close state
// of clientGiveUpChan. The defer does NOT touch s.mutex; the daemon's
// "intent" (clientRunning) is maintained by the RPC handlers, not by this
// goroutine.
func (s *Server) connectWithRetryRuns(ctx context.Context, profileConfig *profilemanager.Config, statusRecorder *peer.Status, runningChan chan struct{}, giveUpChan chan struct{}) {
// close(giveUpChan) MUST run on every exit path (DisableAutoConnect
// return, backoff.Retry return, panic) — Down() blocks for up to 5s
// waiting on this signal before flipping the state to Idle, and a
// missed close leaves Down() always hitting the timeout. The signal
// fires AFTER clientRunning=false is committed under the mutex so a
// Down/Up racing with the goroutine exit never observes a half-state
// (chan closed but clientRunning still true).
// missed close leaves Down() always hitting the timeout.
defer func() {
s.mutex.Lock()
s.clientRunning = false
s.mutex.Unlock()
if giveUpChan != nil {
close(giveUpChan)
}
@@ -312,6 +339,27 @@ func (s *Server) connectWithRetryRuns(ctx context.Context, profileConfig *profil
if err := backoff.Retry(runOperation, backOff); err != nil {
log.Errorf("operation failed: %v", err)
}
// giveUpChan is closed by the function-scope defer.
}
// connectionGoroutineRunning reports whether the connectWithRetryRuns goroutine is
// still running. Returns false when no goroutine has ever been started
// AND when the most recent one has already closed clientGiveUpChan on
// exit (whether due to ctx cancel, DisableAutoConnect single-shot
// completion, or backoff retry exhaustion).
//
// MUST be called with s.mutex held — accesses s.clientGiveUpChan which
// is written by Start/Up under the same lock.
func (s *Server) connectionGoroutineRunning() bool {
if s.clientGiveUpChan == nil {
return false
}
select {
case <-s.clientGiveUpChan:
return false
default:
return true
}
}
// loginAttempt attempts to login using the provided information. it returns a status in case something fails
@@ -343,52 +391,85 @@ func (s *Server) SetConfig(callerCtx context.Context, msg *proto.SetConfigReques
s.mutex.Lock()
defer s.mutex.Unlock()
if s.checkUpdateSettingsDisabled() {
return nil, gstatus.Errorf(codes.Unavailable, errUpdateSettingsDisabled)
// Skip the update-settings gate when the request carries no actual
// overrides: the CLI builds a SetConfigRequest unconditionally on
// every `netbird up` (setupSetConfigReq in cmd/up.go), so a plain
// `netbird up` would otherwise always trip the gate and surface a
// misleading "setConfig method is not available" warning, even when
// the user did not pass any config flag.
if setConfigRequestHasConfigOverrides(msg) {
if s.checkUpdateSettingsDisabled() {
return nil, gstatus.Errorf(codes.Unavailable, errUpdateSettingsDisabled)
}
}
// MDM gate: refuse the whole request if any of its fields is enforced
// by the active MDM policy. The error carries an MDMManagedFields-
// Violation detail listing the offending key names. Non-conflicting
// fields in the same request are not applied either.
policy := loadMDMPolicy()
if err := rejectMDMManagedFieldConflicts(mdmManagedFieldConflicts(msg, policy)); err != nil {
return nil, err
}
config, err := setConfigInputFromRequest(msg)
if err != nil {
return nil, err
}
if _, err := profilemanager.UpdateConfig(config); err != nil {
log.Errorf("failed to update profile config: %v", err)
return nil, fmt.Errorf("failed to update profile config: %w", err)
}
return &proto.SetConfigResponse{}, nil
}
// setConfigInputFromRequest translates a SetConfigRequest into the
// profilemanager.ConfigInput that profilemanager.UpdateConfig consumes.
// Pure mapping with no business logic beyond presence-aware copying of
// optional fields and the "empty / clean" semantics for the two slice
// fields (DNS labels, NAT external IPs). Extracted from SetConfig to
// keep the handler's cognitive complexity below the SonarCube
// threshold; the body is intentionally linear because each proto
// field is its own optional case. Returns the resolved ConfigInput
// and a non-nil error only when the active profile file path cannot
// be determined.
func setConfigInputFromRequest(msg *proto.SetConfigRequest) (profilemanager.ConfigInput, error) {
var config profilemanager.ConfigInput
profState := profilemanager.ActiveProfileState{
Name: msg.ProfileName,
Username: msg.Username,
}
profPath, err := profState.FilePath()
if err != nil {
log.Errorf("failed to get active profile file path: %v", err)
return nil, fmt.Errorf("failed to get active profile file path: %w", err)
return config, fmt.Errorf("failed to get active profile file path: %w", err)
}
var config profilemanager.ConfigInput
config.ConfigPath = profPath
if msg.ManagementUrl != "" {
config.ManagementURL = msg.ManagementUrl
}
if msg.AdminURL != "" {
config.AdminURL = msg.AdminURL
}
if msg.InterfaceName != nil {
config.InterfaceName = msg.InterfaceName
}
if msg.WireguardPort != nil {
wgPort := int(*msg.WireguardPort)
config.WireguardPort = &wgPort
}
if msg.OptionalPreSharedKey != nil {
if msg.OptionalPreSharedKey != nil && *msg.OptionalPreSharedKey != "" {
config.PreSharedKey = msg.OptionalPreSharedKey
}
if msg.CleanDNSLabels {
config.DNSLabels = domain.List{}
} else if msg.DnsLabels != nil {
dnsLabels := domain.FromPunycodeList(msg.DnsLabels)
config.DNSLabels = dnsLabels
config.DNSLabels = domain.FromPunycodeList(msg.DnsLabels)
}
if msg.CleanNATExternalIPs {
@@ -401,7 +482,6 @@ func (s *Server) SetConfig(callerCtx context.Context, msg *proto.SetConfigReques
if string(msg.CustomDNSAddress) == "empty" {
config.CustomDNSAddress = []byte{}
}
config.ExtraIFaceBlackList = msg.ExtraIFaceBlacklist
if msg.DnsRouteInterval != nil {
@@ -434,22 +514,31 @@ func (s *Server) SetConfig(callerCtx context.Context, msg *proto.SetConfigReques
ttl := int(*msg.SshJWTCacheTTL)
config.SSHJWTCacheTTL = &ttl
}
if msg.Mtu != nil {
mtu := uint16(*msg.Mtu)
config.MTU = &mtu
}
if _, err := profilemanager.UpdateConfig(config); err != nil {
log.Errorf("failed to update profile config: %v", err)
return nil, fmt.Errorf("failed to update profile config: %w", err)
}
return &proto.SetConfigResponse{}, nil
return config, nil
}
// Login uses setup key to prepare configuration for the daemon.
func (s *Server) Login(callerCtx context.Context, msg *proto.LoginRequest) (*proto.LoginResponse, error) {
// Config-override gates. LoginRequest carries the same surface as
// SetConfigRequest (managementUrl, PSK, ssh/rosenpass/port toggles,
// ...), so the same protections must apply. Without these the CLI
// command `netbird up --management-url=X` (which falls through to
// Login when SetConfig is rejected — see cmd/up.go) would silently
// bypass `--disable-update-settings` and any MDM policy.
if loginRequestHasConfigOverrides(msg) {
if s.checkUpdateSettingsDisabled() {
return nil, gstatus.Errorf(codes.Unavailable, errUpdateSettingsDisabled)
}
policy := loadMDMPolicy()
if err := rejectMDMManagedFieldConflicts(loginRequestMDMConflicts(msg, policy)); err != nil {
return nil, err
}
}
s.mutex.Lock()
if s.actCancel != nil {
s.actCancel()
@@ -770,7 +859,13 @@ func (s *Server) WaitSSOLogin(callerCtx context.Context, msg *proto.WaitSSOLogin
// Up starts engine work in the daemon.
func (s *Server) Up(callerCtx context.Context, msg *proto.UpRequest) (*proto.UpResponse, error) {
s.mutex.Lock()
if s.clientRunning {
// clientRunning is the daemon-intent flag (set by previous Up/Start, cleared
// by Down). connectionGoroutineRunning() reports whether the previous retry-loop
// goroutine is still trying. When intent is up AND goroutine is alive,
// the existing engine is on the job — just wait for it. When intent
// is up but the goroutine has given up (backoff exhausted) OR when
// intent is down, fall through to spawn a fresh retry loop.
if s.clientRunning && s.connectionGoroutineRunning() {
state := internal.CtxGetState(s.rootCtx)
status, err := state.Status()
if err != nil {
@@ -877,6 +972,7 @@ func (s *Server) Up(callerCtx context.Context, msg *proto.UpRequest) (*proto.UpR
s.clientGiveUpChan = make(chan struct{})
go s.connectWithRetryRuns(ctx, s.config, s.statusRecorder, s.clientRunningChan, s.clientGiveUpChan)
s.publishConfigChangedEvent(proto.MetadataSourceUpRPC)
s.mutex.Unlock()
if msg.GetAsync() {
@@ -1026,6 +1122,12 @@ func (s *Server) cleanupConnection() error {
return ErrServiceNotUp
}
// Daemon intent flips to "down" — all callers (Down RPC,
// Logout RPC handlers) tear down the connection because the user
// explicitly asked for it. MDM restart does NOT go through this
// path, so its clientRunning stays true.
s.clientRunning = false
// Capture the engine reference before cancelling the context.
// After actCancel(), the connectWithRetryRuns goroutine wakes up
// and sets connectClient.engine = nil, causing connectClient.Stop()
@@ -1241,10 +1343,14 @@ func (s *Server) Status(
msg *proto.StatusRequest,
) (*proto.StatusResponse, error) {
s.mutex.Lock()
clientRunning := s.clientRunning
// Only wait if the retry-loop goroutine is alive and making
// progress. clientRunning=true with connectionGoroutineRunning=false means the
// backoff has given up — there is nothing to wait for; let the
// caller observe the failed status directly.
alive := s.connectionGoroutineRunning()
s.mutex.Unlock()
if msg.WaitForReady != nil && *msg.WaitForReady && clientRunning {
if msg.WaitForReady != nil && *msg.WaitForReady && alive {
state := internal.CtxGetState(s.rootCtx)
status, err := state.Status()
if err != nil {
@@ -1865,6 +1971,7 @@ func (s *Server) GetConfig(ctx context.Context, req *proto.GetConfigRequest) (*p
EnableSSHRemotePortForwarding: enableSSHRemotePortForwarding,
DisableSSHAuth: disableSSHAuth,
SshJWTCacheTTL: sshJWTCacheTTL,
MDMManagedFields: cfg.Policy().ManagedKeys(),
}, nil
}
@@ -2007,7 +2114,7 @@ func (s *Server) GetFeatures(ctx context.Context, msg *proto.GetFeaturesRequest)
features := &proto.GetFeaturesResponse{
DisableProfiles: s.checkProfilesDisabled(),
DisableUpdateSettings: s.checkUpdateSettingsDisabled(),
DisableNetworks: s.networksDisabled,
DisableNetworks: s.checkNetworksDisabled(),
}
return features, nil
@@ -2029,22 +2136,46 @@ func (s *Server) connect(ctx context.Context, config *profilemanager.Config, sta
return nil
}
// MDM authority: when the platform-native MDM source sets a kill switch
// key (regardless of true/false value), that value wins. The CLI flag
// supplied at service install time is the fallback used only when the
// MDM source is silent on the key. This honors the "MDM decides
// everything" semantic agreed for NET-1214 — an admin pushing
// disableX=false via MDM explicitly re-enables the feature even on a
// box installed with --disable-X.
func (s *Server) checkProfilesDisabled() bool {
// Check if the environment variable is set to disable profiles
if s.profilesDisabled {
return true
if s.config != nil {
if v, ok := s.config.Policy().GetBool(mdm.KeyDisableProfiles); ok {
return v
}
}
return s.profilesDisabled
}
return false
// checkNetworksDisabled reports whether the networks/exit-node feature
// is disabled on this daemon instance. Resolved MDM-first: when the
// active policy declares mdm.KeyDisableNetworks the policy value wins
// (regardless of true/false), so an admin can re-enable the feature
// via MDM even on a host that was installed with --disable-networks.
// Falls back to the s.networksDisabled CLI flag when the policy is
// silent on the key. Mirrors checkProfilesDisabled and
// checkUpdateSettingsDisabled.
func (s *Server) checkNetworksDisabled() bool {
if s.config != nil {
if v, ok := s.config.Policy().GetBool(mdm.KeyDisableNetworks); ok {
return v
}
}
return s.networksDisabled
}
func (s *Server) checkUpdateSettingsDisabled() bool {
// Check if the environment variable is set to disable profiles
if s.updateSettingsDisabled {
return true
if s.config != nil {
if v, ok := s.config.Policy().GetBool(mdm.KeyDisableUpdateSettings); ok {
return v
}
}
return false
return s.updateSettingsDisabled
}
func (s *Server) startUpdateManagerForGUI() {

View File

@@ -101,6 +101,7 @@ func TestCleanupConnection_ClearsConnectClient(t *testing.T) {
require.NoError(t, err)
assert.Nil(t, s.connectClient, "connectClient should be nil after cleanup")
assert.False(t, s.clientRunning, "clientRunning should be cleared after cleanup (intent = down)")
}
// TestCleanState_NilConnectClient validates that CleanState doesn't panic
@@ -144,17 +145,20 @@ func TestDownThenUp_StaleRunningChan(t *testing.T) {
_, cancel := context.WithCancel(context.Background())
s.actCancel = cancel
// Simulate Down(): cleanupConnection sets connectClient = nil
// Simulate Down(): cleanupConnection sets connectClient = nil and
// flips clientRunning to false (intent = down). The connectionGoroutineRunning state
// remains independent of intent — derived from clientGiveUpChan.
s.mutex.Lock()
err := s.cleanupConnection()
s.mutex.Unlock()
require.NoError(t, err)
// After cleanup: connectClient is nil, clientRunning still true
// (goroutine hasn't exited yet)
// After cleanup: connectClient is nil, clientRunning is false (intent
// cleared by cleanupConnection), connectionGoroutineRunning may still be true
// (goroutine teardown is independent of the intent flag).
s.mutex.Lock()
assert.Nil(t, s.connectClient, "connectClient should be nil after cleanup")
assert.True(t, s.clientRunning, "clientRunning still true until goroutine exits")
assert.False(t, s.clientRunning, "clientRunning should be cleared by cleanupConnection (intent = down)")
s.mutex.Unlock()
// waitForUp() returns immediately due to stale closed clientRunningChan

View File

@@ -0,0 +1,198 @@
package server
import (
"context"
"os/user"
"path/filepath"
"testing"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
"google.golang.org/grpc/codes"
gstatus "google.golang.org/grpc/status"
"github.com/netbirdio/netbird/client/internal/profilemanager"
"github.com/netbirdio/netbird/client/mdm"
"github.com/netbirdio/netbird/client/proto"
)
// withMDMPolicy temporarily overrides the server-package loadMDMPolicy hook
// so SetConfig observes the supplied Policy. Restores the original loader
// at test cleanup.
func withMDMPolicy(t *testing.T, policy *mdm.Policy) {
t.Helper()
prev := loadMDMPolicy
loadMDMPolicy = func() *mdm.Policy { return policy }
t.Cleanup(func() { loadMDMPolicy = prev })
}
// setupServerWithProfile mirrors the boilerplate of TestSetConfig_AllFieldsSaved:
// overrides profilemanager paths to a temp dir, seeds a profile, sets it
// active, and constructs a Server instance. Returns the constructed server
// plus context + profile name + username + cfgPath for the seeded profile.
func setupServerWithProfile(t *testing.T) (s *Server, ctx context.Context, profName, username, cfgPath string) {
t.Helper()
tempDir := t.TempDir()
origDefaultProfileDir := profilemanager.DefaultConfigPathDir
origDefaultConfigPath := profilemanager.DefaultConfigPath
origActiveProfileStatePath := profilemanager.ActiveProfileStatePath
profilemanager.ConfigDirOverride = tempDir
profilemanager.DefaultConfigPathDir = tempDir
profilemanager.ActiveProfileStatePath = tempDir + "/active_profile.json"
profilemanager.DefaultConfigPath = filepath.Join(tempDir, "default.json")
t.Cleanup(func() {
profilemanager.DefaultConfigPathDir = origDefaultProfileDir
profilemanager.ActiveProfileStatePath = origActiveProfileStatePath
profilemanager.DefaultConfigPath = origDefaultConfigPath
profilemanager.ConfigDirOverride = ""
})
currUser, err := user.Current()
require.NoError(t, err)
profName = "test-profile-mdm"
cfgPath = filepath.Join(tempDir, profName+".json")
_, err = profilemanager.UpdateOrCreateConfig(profilemanager.ConfigInput{
ConfigPath: cfgPath,
ManagementURL: "https://api.netbird.io:443",
})
require.NoError(t, err)
pm := profilemanager.ServiceManager{}
require.NoError(t, pm.SetActiveProfileState(&profilemanager.ActiveProfileState{
Name: profName,
Username: currUser.Username,
}))
ctx = context.Background()
s = New(ctx, "console", "", false, false, false, false)
return s, ctx, profName, currUser.Username, cfgPath
}
// extractViolation pulls the MDMManagedFieldsViolation detail from a
// FailedPrecondition error. Fails the test if absent or malformed.
func extractViolation(t *testing.T, err error) *proto.MDMManagedFieldsViolation {
t.Helper()
require.Error(t, err)
st, ok := gstatus.FromError(err)
require.True(t, ok, "error must be a gRPC status: %v", err)
require.Equal(t, codes.FailedPrecondition, st.Code(), "expected FailedPrecondition, got %s", st.Code())
for _, d := range st.Details() {
if v, ok := d.(*proto.MDMManagedFieldsViolation); ok {
return v
}
}
t.Fatalf("MDMManagedFieldsViolation detail not found on status; details: %v", st.Details())
return nil
}
func TestSetConfig_MDMReject_SingleField(t *testing.T) {
withMDMPolicy(t, mdm.NewPolicy(map[string]any{
mdm.KeyManagementURL: "https://mdm.example.com:443",
}))
s, ctx, profName, username, _ := setupServerWithProfile(t)
_, err := s.SetConfig(ctx, &proto.SetConfigRequest{
ProfileName: profName,
Username: username,
ManagementUrl: "https://user.tried.this.com:443",
})
v := extractViolation(t, err)
assert.Equal(t, []string{mdm.KeyManagementURL}, v.GetFields())
}
func TestSetConfig_MDMReject_MultipleFields(t *testing.T) {
withMDMPolicy(t, mdm.NewPolicy(map[string]any{
mdm.KeyManagementURL: "https://mdm.example.com:443",
mdm.KeyBlockInbound: true,
mdm.KeyRosenpassEnabled: true,
}))
s, ctx, profName, username, _ := setupServerWithProfile(t)
blockInbound := false
rosenpassEnabled := false
_, err := s.SetConfig(ctx, &proto.SetConfigRequest{
ProfileName: profName,
Username: username,
ManagementUrl: "https://user.tried.this.com:443",
BlockInbound: &blockInbound,
RosenpassEnabled: &rosenpassEnabled,
})
v := extractViolation(t, err)
assert.ElementsMatch(t, []string{
mdm.KeyManagementURL,
mdm.KeyBlockInbound,
mdm.KeyRosenpassEnabled,
}, v.GetFields())
}
func TestSetConfig_MDMReject_AllOrNothing(t *testing.T) {
// MDM enforces ManagementURL only; user request touches both the
// enforced field AND a non-enforced field (RosenpassEnabled).
// The whole request must be rejected — non-conflicting fields are not
// applied either.
withMDMPolicy(t, mdm.NewPolicy(map[string]any{
mdm.KeyManagementURL: "https://mdm.example.com:443",
}))
s, ctx, profName, username, cfgPath := setupServerWithProfile(t)
rosenpassEnabled := true
_, err := s.SetConfig(ctx, &proto.SetConfigRequest{
ProfileName: profName,
Username: username,
ManagementUrl: "https://user.tried.this.com:443",
RosenpassEnabled: &rosenpassEnabled,
})
v := extractViolation(t, err)
assert.Equal(t, []string{mdm.KeyManagementURL}, v.GetFields())
// Confirm RosenpassEnabled was NOT applied even though it was not
// in the conflict list: the request was rejected as a whole.
reloaded, err := profilemanager.GetConfig(cfgPath)
require.NoError(t, err)
assert.False(t, reloaded.RosenpassEnabled, "non-conflicting field must not be applied when request is rejected")
}
func TestSetConfig_MDMAllow_NonManagedFields(t *testing.T) {
// MDM enforces ManagementURL but the user only writes RosenpassEnabled.
// Request must succeed.
withMDMPolicy(t, mdm.NewPolicy(map[string]any{
mdm.KeyManagementURL: "https://mdm.example.com:443",
}))
s, ctx, profName, username, _ := setupServerWithProfile(t)
rosenpassEnabled := true
resp, err := s.SetConfig(ctx, &proto.SetConfigRequest{
ProfileName: profName,
Username: username,
RosenpassEnabled: &rosenpassEnabled,
})
require.NoError(t, err)
require.NotNil(t, resp)
}
func TestSetConfig_MDMEmpty_NoEnforcement(t *testing.T) {
// No MDM policy active: any field can be written.
withMDMPolicy(t, mdm.NewPolicy(nil))
s, ctx, profName, username, _ := setupServerWithProfile(t)
resp, err := s.SetConfig(ctx, &proto.SetConfigRequest{
ProfileName: profName,
Username: username,
ManagementUrl: "https://user.changed.url.com:443",
})
require.NoError(t, err)
require.NotNil(t, resp)
}