cmd/tl-longchain: tool to re-sign nodes with long rotation signatures

In Tailnet Lock, there is an implicit limit on the number of rotation
signatures that can be chained before the signature becomes too long.

This program helps tailnet admins to identify nodes that have signatures
with long chains and prints commands to re-sign those node keys with a
fresh direct signature. It's a temporary mitigation measure, and we will
remove this tool as we design and implement a long-term approach for
rotation signatures.

Example output:

```
2024/08/20 18:25:03 Self: does not need re-signing
2024/08/20 18:25:03 Visible peers with valid signatures:
2024/08/20 18:25:03 Peer xxx2.yy.ts.net. (100.77.192.34) nodeid=nyDmhiZiGA11KTM59, current signature kind=direct: does not need re-signing
2024/08/20 18:25:03 Peer xxx3.yy.ts.net. (100.84.248.22) nodeid=ndQ64mDnaB11KTM59, current signature kind=direct: does not need re-signing
2024/08/20 18:25:03 Peer xxx4.yy.ts.net. (100.85.253.53) nodeid=nmZfVygzkB21KTM59, current signature kind=rotation: chain length 4, printing command to re-sign
tailscale lock sign nodekey:530bddbfbe69e91fe15758a1d6ead5337aa6307e55ac92dafad3794f8b3fc661 tlpub:4bf07597336703395f2149dce88e7c50dd8694ab5bbde3d7c2a1c7b3e231a3c2
```

To support this, the NetworkLockStatus localapi response now includes
information about signatures of all peers rather than just the invalid
ones. This is not displayed by default in `tailscale lock status`, but
will be surfaced in `tailscale lock status --json`.

Updates #13185

Signed-off-by: Anton Tolchanov <anton@tailscale.com>

cmd/tl-longchain/tl-longchain.go

@@ -0,0 +1,93 @@
+// Copyright (c) Tailscale Inc & AUTHORS
+// SPDX-License-Identifier: BSD-3-Clause
+
+// Program tl-longchain prints commands to re-sign Tailscale nodes that have
+// long rotation signature chains.
+//
+// There is an implicit limit on the number of rotation signatures that can
+// be chained before the signature becomes too long. This program helps
+// tailnet admins to identify nodes that have signatures with long chains and
+// prints commands to re-sign those node keys with a fresh direct signature.
+// Commands are printed to stdout, while log messages are printed to stderr.
+//
+// Note that the Tailscale client this command is executed on must have
+// ACL visibility to all other nodes to be able to see their signatures.
+// https://tailscale.com/kb/1087/device-visibility
+package main
+
+import (
+	"context"
+	"flag"
+	"fmt"
+	"log"
+	"time"
+
+	"tailscale.com/client/tailscale"
+	"tailscale.com/ipn/ipnstate"
+	"tailscale.com/tka"
+	"tailscale.com/types/key"
+)
+
+var (
+	flagSocket   = flag.String("socket", "", "custom path to tailscaled socket")
+	maxRotations = flag.Int("rotations", 10, "number of rotation signatures before re-signing (max 16)")
+	showFiltered = flag.Bool("show-filtered", false, "include nodes with invalid signatures")
+)
+
+func main() {
+	flag.Parse()
+
+	lc := tailscale.LocalClient{Socket: *flagSocket}
+	if lc.Socket != "" {
+		lc.UseSocketOnly = true
+	}
+
+	ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
+	defer cancel()
+
+	st, err := lc.NetworkLockStatus(ctx)
+	if err != nil {
+		log.Fatalf("could not get Tailnet Lock status: %v", err)
+	}
+	if !st.Enabled {
+		log.Print("Tailnet Lock is not enabled")
+		return
+	}
+	print("Self", *st.NodeKey, *st.NodeKeySignature)
+	if len(st.VisiblePeers) > 0 {
+		log.Print("Visible peers with valid signatures:")
+		for _, peer := range st.VisiblePeers {
+			print(peerInfo(peer), peer.NodeKey, peer.NodeKeySignature)
+		}
+	}
+	if *showFiltered && len(st.FilteredPeers) > 0 {
+		log.Print("Visible peers with invalid signatures:")
+		for _, peer := range st.FilteredPeers {
+			print(peerInfo(peer), peer.NodeKey, peer.NodeKeySignature)
+		}
+	}
+}
+
+// peerInfo returns a string with information about a peer.
+func peerInfo(peer *ipnstate.TKAPeer) string {
+	return fmt.Sprintf("Peer %s (%s) nodeid=%s, current signature kind=%v", peer.Name, peer.TailscaleIPs[0], peer.StableID, peer.NodeKeySignature.SigKind)
+}
+
+// print prints a message about a node key signature and a re-signing command if needed.
+func print(info string, nodeKey key.NodePublic, sig tka.NodeKeySignature) {
+	if l := chainLength(sig); l > *maxRotations {
+		log.Printf("%s: chain length %d, printing command to re-sign", info, l)
+		wrapping, _ := sig.UnverifiedWrappingPublic()
+		fmt.Printf("tailscale lock sign %s %s\n", nodeKey, key.NLPublicFromEd25519Unsafe(wrapping).CLIString())
+	} else {
+		log.Printf("%s: does not need re-signing", info)
+	}
+}
+
+// chainLength returns the length of the rotation signature chain.
+func chainLength(sig tka.NodeKeySignature) int {
+	if sig.SigKind != tka.SigRotation {
+		return 1
+	}
+	return 1 + chainLength(*sig.Nested)
+}

ipn/ipnlocal/network-lock.go

@@ -53,7 +53,7 @@ type tkaState struct {
 	profile   ipn.ProfileID
 	authority *tka.Authority
 	storage   *tka.FS
-	filtered  []ipnstate.TKAFilteredPeer
+	filtered  []ipnstate.TKAPeer
 }
 
 // tkaFilterNetmapLocked checks the signatures on each node key, dropping
@@ -99,7 +99,7 @@ func (b *LocalBackend) tkaFilterNetmapLocked(nm *netmap.NetworkMap) {
 	// nm.Peers is ordered, so deletion must be order-preserving.
 	if len(toDelete) > 0 || len(obsoleteByRotation) > 0 {
 		peers := make([]tailcfg.NodeView, 0, len(nm.Peers))
-		filtered := make([]ipnstate.TKAFilteredPeer, 0, len(toDelete)+len(obsoleteByRotation))
+		filtered := make([]ipnstate.TKAPeer, 0, len(toDelete)+len(obsoleteByRotation))
 		for i, p := range nm.Peers {
 			if !toDelete[i] && !obsoleteByRotation.Contains(p.Key()) {
 				peers = append(peers, p)
@@ -108,20 +108,7 @@ func (b *LocalBackend) tkaFilterNetmapLocked(nm *netmap.NetworkMap) {
 					b.logf("Network lock is dropping peer %v(%v) due to key rotation", p.ID(), p.StableID())
 				}
 				// Record information about the node we filtered out.
-				fp := ipnstate.TKAFilteredPeer{
+				filtered = append(filtered, tkaStateFromPeer(p))
-					Name:         p.Name(),
-					ID:           p.ID(),
-					StableID:     p.StableID(),
-					TailscaleIPs: make([]netip.Addr, p.Addresses().Len()),
-					NodeKey:      p.Key(),
-				}
-				for i := range p.Addresses().Len() {
-					addr := p.Addresses().At(i)
-					if addr.IsSingleIP() && tsaddr.IsTailscaleIP(addr.Addr()) {
-						fp.TailscaleIPs[i] = addr.Addr()
-					}
-				}
-				filtered = append(filtered, fp)
 			}
 		}
 		nm.Peers = peers
@@ -546,11 +533,17 @@ func (b *LocalBackend) NetworkLockStatus() *ipnstate.NetworkLockStatus {
 		}
 	}
 
-	filtered := make([]*ipnstate.TKAFilteredPeer, len(b.tka.filtered))
+	filtered := make([]*ipnstate.TKAPeer, len(b.tka.filtered))
 	for i := range len(filtered) {
 		filtered[i] = b.tka.filtered[i].Clone()
 	}
 
+	visible := make([]*ipnstate.TKAPeer, len(b.netMap.Peers))
+	for i, p := range b.netMap.Peers {
+		s := tkaStateFromPeer(p)
+		visible[i] = &s
+	}
+
 	stateID1, _ := b.tka.authority.StateIDs()
 
 	return &ipnstate.NetworkLockStatus{
@@ -562,10 +555,32 @@ func (b *LocalBackend) NetworkLockStatus() *ipnstate.NetworkLockStatus {
 		NodeKeySignature: nodeKeySignature,
 		TrustedKeys:      outKeys,
 		FilteredPeers:    filtered,
+		VisiblePeers:     visible,
 		StateID:          stateID1,
 	}
 }
 
+func tkaStateFromPeer(p tailcfg.NodeView) ipnstate.TKAPeer {
+	fp := ipnstate.TKAPeer{
+		Name:         p.Name(),
+		ID:           p.ID(),
+		StableID:     p.StableID(),
+		TailscaleIPs: make([]netip.Addr, 0, p.Addresses().Len()),
+		NodeKey:      p.Key(),
+	}
+	for i := range p.Addresses().Len() {
+		addr := p.Addresses().At(i)
+		if addr.IsSingleIP() && tsaddr.IsTailscaleIP(addr.Addr()) {
+			fp.TailscaleIPs = append(fp.TailscaleIPs, addr.Addr())
+		}
+	}
+	var decoded tka.NodeKeySignature
+	if err := decoded.Unserialize(p.KeySignature().AsSlice()); err == nil {
+		fp.NodeKeySignature = decoded
+	}
+	return fp
+}
+
 // NetworkLockInit enables network-lock for the tailnet, with the tailnets'
 // key authority initialized to trust the provided keys.
 //

ipn/ipnstate/ipnstate.go

@@ -26,7 +26,7 @@ import (
 	"tailscale.com/version"
 )
 
-//go:generate go run tailscale.com/cmd/cloner  -clonefunc=false -type=TKAFilteredPeer
+//go:generate go run tailscale.com/cmd/cloner  -clonefunc=false -type=TKAPeer
 
 // Status represents the entire state of the IPN network.
 type Status struct {
@@ -94,15 +94,14 @@ type TKAKey struct {
 	Votes    uint
 }
 
-// TKAFilteredPeer describes a peer which was removed from the netmap
+// TKAPeer describes a peer and its network lock details.
-// (i.e. no connectivity) because it failed tailnet lock
+type TKAPeer struct {
-// checks.
+	Name             string // DNS
-type TKAFilteredPeer struct {
+	ID               tailcfg.NodeID
-	Name         string // DNS
+	StableID         tailcfg.StableNodeID
-	ID           tailcfg.NodeID
+	TailscaleIPs     []netip.Addr // Tailscale IP(s) assigned to this node
-	StableID     tailcfg.StableNodeID
+	NodeKey          key.NodePublic
-	TailscaleIPs []netip.Addr // Tailscale IP(s) assigned to this node
+	NodeKeySignature tka.NodeKeySignature
-	NodeKey      key.NodePublic
 }
 
 // NetworkLockStatus represents whether network-lock is enabled,
@@ -134,10 +133,14 @@ type NetworkLockStatus struct {
 	// to network-lock.
 	TrustedKeys []TKAKey
 
+	// VisiblePeers describes peers which are visible in the netmap that
+	// have valid Tailnet Lock signatures signatures.
+	VisiblePeers []*TKAPeer
+
 	// FilteredPeers describes peers which were removed from the netmap
 	// (i.e. no connectivity) because they failed tailnet lock
 	// checks.
-	FilteredPeers []*TKAFilteredPeer
+	FilteredPeers []*TKAPeer
 
 	// StateID is a nonce associated with the network lock authority,
 	// generated upon enablement. This field is not populated if the

ipn/ipnstate/ipnstate_clone.go

@@ -9,26 +9,29 @@ import (
 	"net/netip"
 
 	"tailscale.com/tailcfg"
+	"tailscale.com/tka"
 	"tailscale.com/types/key"
 )
 
-// Clone makes a deep copy of TKAFilteredPeer.
+// Clone makes a deep copy of TKAPeer.
 // The result aliases no memory with the original.
-func (src *TKAFilteredPeer) Clone() *TKAFilteredPeer {
+func (src *TKAPeer) Clone() *TKAPeer {
 	if src == nil {
 		return nil
 	}
-	dst := new(TKAFilteredPeer)
+	dst := new(TKAPeer)
 	*dst = *src
 	dst.TailscaleIPs = append(src.TailscaleIPs[:0:0], src.TailscaleIPs...)
+	dst.NodeKeySignature = *src.NodeKeySignature.Clone()
 	return dst
 }
 
 // A compilation failure here means this code must be regenerated, with the command at the top of this file.
-var _TKAFilteredPeerCloneNeedsRegeneration = TKAFilteredPeer(struct {
+var _TKAPeerCloneNeedsRegeneration = TKAPeer(struct {
-	Name         string
+	Name             string
-	ID           tailcfg.NodeID
+	ID               tailcfg.NodeID
-	StableID     tailcfg.StableNodeID
+	StableID         tailcfg.StableNodeID
-	TailscaleIPs []netip.Addr
+	TailscaleIPs     []netip.Addr
-	NodeKey      key.NodePublic
+	NodeKey          key.NodePublic
+	NodeKeySignature tka.NodeKeySignature
 }{})

tka/sig.go

@@ -19,6 +19,8 @@ import (
 	"tailscale.com/types/tkatype"
 )
 
+//go:generate go run tailscale.com/cmd/cloner  -clonefunc=false -type=NodeKeySignature
+
 // SigKind describes valid NodeKeySignature types.
 type SigKind uint8
 

tka/tka_clone.go

@@ -0,0 +1,32 @@
+// Copyright (c) Tailscale Inc & AUTHORS
+// SPDX-License-Identifier: BSD-3-Clause
+
+// Code generated by tailscale.com/cmd/cloner; DO NOT EDIT.
+
+package tka
+
+// Clone makes a deep copy of NodeKeySignature.
+// The result aliases no memory with the original.
+func (src *NodeKeySignature) Clone() *NodeKeySignature {
+	if src == nil {
+		return nil
+	}
+	dst := new(NodeKeySignature)
+	*dst = *src
+	dst.Pubkey = append(src.Pubkey[:0:0], src.Pubkey...)
+	dst.KeyID = append(src.KeyID[:0:0], src.KeyID...)
+	dst.Signature = append(src.Signature[:0:0], src.Signature...)
+	dst.Nested = src.Nested.Clone()
+	dst.WrappingPubkey = append(src.WrappingPubkey[:0:0], src.WrappingPubkey...)
+	return dst
+}
+
+// A compilation failure here means this code must be regenerated, with the command at the top of this file.
+var _NodeKeySignatureCloneNeedsRegeneration = NodeKeySignature(struct {
+	SigKind        SigKind
+	Pubkey         []byte
+	KeyID          []byte
+	Signature      []byte
+	Nested         *NodeKeySignature
+	WrappingPubkey []byte
+}{})