Merge branch 'master' of github.com:harmony-one/harmony into staking-service

.gitignore

@@ -51,3 +51,12 @@ db/
 
 # bootnode keystore
 .bnkey
+
+# harmony node keystore
+.hmykey
+
+# vendor directory
+vendor
+
+# node_modules
+node_modules/

.gitmodules

@@ -5,3 +5,6 @@
 [submodule "vendor/github.com/golang/protobuf"]
 	path = vendor/github.com/golang/protobuf
 	url = https://github.com/golang/protobuf
+[submodule "vendor/github.com/libp2p/go-libp2p-kad-dht"]
+	path = vendor/github.com/libp2p/go-libp2p-kad-dht
+	url = https://github.com/libp2p/go-libp2p-kad-dht

api/service/discovery/discovery_test.go

@@ -0,0 +1,31 @@
+package discovery
+
+import (
+	"testing"
+
+	"github.com/harmony-one/harmony/internal/utils"
+	"github.com/harmony-one/harmony/p2p"
+	"github.com/harmony-one/harmony/p2p/p2pimpl"
+)
+
+var (
+	ip      = "127.0.0.1"
+	port    = "7099"
+	service *Service
+)
+
+func TestDiscoveryService(t *testing.T) {
+	selfPeer := p2p.Peer{IP: ip, Port: port}
+	priKey, _, err := utils.GenKeyP2P(ip, port)
+
+	host, err := p2pimpl.NewHost(&selfPeer, priKey)
+	if err != nil {
+		t.Fatalf("unable to new host in harmony: %v", err)
+	}
+
+	service = New(host, "rendezvous")
+
+	if service == nil {
+		t.Fatalf("unable to create new discovery service")
+	}
+}

api/service/discovery/errors.go

@@ -0,0 +1,12 @@
+package discovery
+
+import "errors"
+
+// Errors of peer discovery
+var (
+	ErrGetPeers       = errors.New("[DISCOVERY]: get peer list failed")
+	ErrConnectionFull = errors.New("[DISCOVERY]: node's incoming connection full")
+	ErrPing           = errors.New("[DISCOVERY]: ping peer failed")
+	ErrPong           = errors.New("[DISCOVERY]: pong peer failed")
+	ErrDHTBootstrap   = errors.New("[DISCOVERY]: DHT bootstrap failed")
+)

api/service/discovery/service.go

@@ -0,0 +1,131 @@
+package discovery
+
+import (
+	"context"
+	"sync"
+
+	"github.com/ethereum/go-ethereum/log"
+	proto_discovery "github.com/harmony-one/harmony/api/proto/discovery"
+	"github.com/harmony-one/harmony/internal/utils"
+	"github.com/harmony-one/harmony/p2p"
+
+	peerstore "github.com/libp2p/go-libp2p-peerstore"
+
+	libp2pdis "github.com/libp2p/go-libp2p-discovery"
+	libp2pdht "github.com/libp2p/go-libp2p-kad-dht"
+)
+
+// Constants for discovery service.
+const (
+	numIncoming = 128
+	numOutgoing = 16
+)
+
+// Service is the struct for discovery service.
+type Service struct {
+	Host       p2p.Host
+	DHT        *libp2pdht.IpfsDHT
+	Rendezvous string
+	ctx        context.Context
+	peerChan   <-chan peerstore.PeerInfo
+}
+
+// New returns discovery service.
+// h is the p2p host
+// r is the rendezvous string, we use shardID to start (TODO: leo, build two overlays of network)
+func New(h p2p.Host, r string) *Service {
+	ctx := context.Background()
+	dht, err := libp2pdht.New(ctx, h.GetP2PHost())
+	if err != nil {
+		panic(err)
+	}
+
+	return &Service{
+		Host:       h,
+		DHT:        dht,
+		Rendezvous: r,
+		ctx:        ctx,
+		peerChan:   make(<-chan peerstore.PeerInfo),
+	}
+}
+
+// StartService starts discovery service.
+func (s *Service) StartService() {
+	log.Info("Starting discovery service.")
+	err := s.Init()
+	if err != nil {
+		log.Error("StartService Aborted", "Error", err)
+		return
+	}
+
+	// We use a rendezvous point "shardID" to announce our location.
+	log.Info("Announcing ourselves...")
+	routingDiscovery := libp2pdis.NewRoutingDiscovery(s.DHT)
+	libp2pdis.Advertise(s.ctx, routingDiscovery, s.Rendezvous)
+	log.Debug("Successfully announced!")
+
+	log.Debug("Searching for other peers...")
+	s.peerChan, err = routingDiscovery.FindPeers(s.ctx, s.Rendezvous)
+	if err != nil {
+		log.Error("FindPeers", "error", err)
+	}
+}
+
+// StopService shutdowns discovery service.
+func (s *Service) StopService() {
+	log.Info("Shutting down discovery service.")
+}
+
+func (s *Service) foundPeers() {
+	select {
+	case peer := <-s.peerChan:
+		if peer.ID != s.Host.GetP2PHost().ID() {
+			log.Debug("Found Peer", "peer", peer.ID, "addr", peer.Addrs)
+			if len(peer.ID) > 0 {
+				p := p2p.Peer{PeerID: peer.ID, Addrs: peer.Addrs}
+				s.Host.AddPeer(&p)
+				// TODO: stop ping if pinged before
+				s.pingPeer(p)
+			}
+		}
+	}
+}
+
+// Init is to initialize for discoveryService.
+func (s *Service) Init() error {
+	log.Info("Init discovery service")
+
+	// Bootstrap the DHT. In the default configuration, this spawns a Background
+	// thread that will refresh the peer table every five minutes.
+	log.Debug("Bootstrapping the DHT")
+	if err := s.DHT.Bootstrap(s.ctx); err != nil {
+		return ErrDHTBootstrap
+	}
+
+	var wg sync.WaitGroup
+	for _, peerAddr := range utils.BootNodes {
+		peerinfo, _ := peerstore.InfoFromP2pAddr(peerAddr)
+		wg.Add(1)
+		go func() {
+			defer wg.Done()
+			if err := s.Host.GetP2PHost().Connect(s.ctx, *peerinfo); err != nil {
+				log.Warn("can't connect to bootnode", "error", err)
+			} else {
+				log.Info("connected to bootnode", "node", *peerinfo)
+			}
+		}()
+	}
+	wg.Wait()
+
+	go s.foundPeers()
+
+	return nil
+}
+
+func (s *Service) pingPeer(peer p2p.Peer) {
+	ping := proto_discovery.NewPingMessage(s.Host.GetSelfPeer())
+	buffer := ping.ConstructPingMessage()
+	log.Debug("Sending Ping Message to", "peer", peer)
+	s.Host.SendMessage(peer, buffer)
+	log.Debug("Sent Ping Message to", "peer", peer)
+}

api/service/manager.go

@@ -22,18 +22,19 @@ type Type byte
 // Constants for Type.
 const (
 	SupportSyncing Type = iota
-	SupportClient
+	ClientSupport
 	SupportExplorer
 	Consensus
 	BlockProposal
+	PeerDiscovery
 	Test
 	Done
 )
 
 func (t Type) String() string {
 	switch t {
-	case SupportClient:
-		return "SupportClient"
+	case ClientSupport:
+		return "ClientSupport"
 	case SupportSyncing:
 		return "SyncingSupport"
 	case SupportExplorer:
@@ -42,6 +43,8 @@ func (t Type) String() string {
 		return "Consensus"
 	case BlockProposal:
 		return "BlockProposal"
+	case PeerDiscovery:
+		return "PeerDiscovery"
 	case Test:
 		return "Test"
 	case Staking:

api/service/syncing/syncing.go

@@ -87,6 +87,8 @@ type StateSync struct {
 
 // AddLastMileBlock add the lastest a few block into queue for syncing
 func (ss *StateSync) AddLastMileBlock(block *types.Block) {
+	ss.syncMux.Lock()
+	defer ss.syncMux.Unlock()
 	ss.lastMileBlocks = append(ss.lastMileBlocks, block)
 }
 

cmd/harmony.go

@@ -216,9 +216,6 @@ func main() {
 	}
 
 	go currentNode.SupportSyncing()
-	if consensus.IsLeader {
-		go currentNode.SupportClient()
-	}
 	currentNode.ServiceManagerSetup()
 	currentNode.RunServices()
 	currentNode.StartServer()

consensus/consensus.go

@@ -92,7 +92,7 @@ type Consensus struct {
 	OnConsensusDone func(*types.Block)
 
 	// current consensus block to check if out of sync
-	ConsensusBlock chan *types.Block
+	ConsensusBlock chan *BFTBlockInfo
 	// verified block to state sync broadcast
 	VerifiedNewBlock chan *types.Block
 
@@ -108,6 +108,13 @@ type Consensus struct {
 	OfflinePeerList []p2p.Peer
 }
 
+// BFTBlockInfo send the latest block that was in BFT consensus process as well as its consensusID to state syncing
+// consensusID is necessary to make sure the out of sync node can enter the correct view
+type BFTBlockInfo struct {
+	Block       *types.Block
+	ConsensusID uint32
+}
+
 // BlockConsensusStatus used to keep track of the consensus status of multiple blocks received so far
 // This is mainly used in the case that this node is lagging behind and needs to catch up.
 // For example, the consensus moved to round N and this node received message(N).
@@ -119,6 +126,16 @@ type BlockConsensusStatus struct {
 	state State  // the latest state of the consensus
 }
 
+// UpdateConsensusID is used to update latest consensusID for nodes that out of sync
+func (consensus *Consensus) UpdateConsensusID(consensusID uint32) {
+	consensus.mutex.Lock()
+	defer consensus.mutex.Unlock()
+	if consensus.consensusID < consensusID {
+		utils.GetLogInstance().Debug("update consensusID", "myConsensusID", consensus.consensusID, "newConsensusID", consensusID)
+		consensus.consensusID = consensusID
+	}
+}
+
 // New creates a new Consensus object
 func New(host p2p.Host, ShardID string, peers []p2p.Peer, leader p2p.Peer) *Consensus {
 	consensus := Consensus{}
@@ -197,7 +214,8 @@ func New(host p2p.Host, ShardID string, peers []p2p.Peer, leader p2p.Peer) *Cons
 }
 
 // Checks the basic meta of a consensus message.
-func (consensus *Consensus) checkConsensusMessage(message consensus_proto.Message, publicKey *bls.PublicKey) bool {
+//
+func (consensus *Consensus) checkConsensusMessage(message consensus_proto.Message, publicKey *bls.PublicKey) error {
 	consensusID := message.ConsensusId
 	blockHash := message.BlockHash
 
@@ -205,20 +223,20 @@ func (consensus *Consensus) checkConsensusMessage(message consensus_proto.Messag
 	err := verifyMessageSig(publicKey, message)
 	if err != nil {
 		utils.GetLogInstance().Warn("Failed to verify the message signature", "Error", err)
-		return false
+		return ErrInvalidConsensusMessage
 	}
 
 	// check consensus Id
 	if consensusID != consensus.consensusID {
 		utils.GetLogInstance().Warn("Wrong consensus Id", "myConsensusId", consensus.consensusID, "theirConsensusId", consensusID, "consensus", consensus)
-		return false
+		return ErrConsensusIDNotMatch
 	}
 
 	if !bytes.Equal(blockHash, consensus.blockHash[:]) {
 		utils.GetLogInstance().Warn("Wrong blockHash", "consensus", consensus)
-		return false
+		return ErrInvalidConsensusMessage
 	}
-	return true
+	return nil
 }
 
 // Gets the validator peer based on validator ID.

consensus/consensus_leader.go

@@ -123,7 +123,7 @@ func (consensus *Consensus) processPrepareMessage(message consensus_proto.Messag
 
 	validatorPeer := consensus.getValidatorPeerByID(validatorID)
 
-	if !consensus.checkConsensusMessage(message, validatorPeer.PubKey) {
+	if err := consensus.checkConsensusMessage(message, validatorPeer.PubKey); err != nil {
 		utils.GetLogInstance().Debug("Failed to check the validator message", "validatorID", validatorID)
 		return
 	}
@@ -185,7 +185,7 @@ func (consensus *Consensus) processCommitMessage(message consensus_proto.Message
 
 	validatorPeer := consensus.getValidatorPeerByID(validatorID)
 
-	if !consensus.checkConsensusMessage(message, validatorPeer.PubKey) {
+	if err := consensus.checkConsensusMessage(message, validatorPeer.PubKey); err != nil {
 		utils.GetLogInstance().Debug("Failed to check the validator message", "validatorID", validatorID)
 		return
 	}

consensus/consensus_validator.go

@@ -12,6 +12,30 @@ import (
 	"github.com/harmony-one/harmony/internal/utils"
 )
 
+// sendBFTBlockToStateSyncing will send the latest BFT consensus block to state syncing checkingjjkkkkkkkkkkkkkkkjnjk
+func (consensus *Consensus) sendBFTBlockToStateSyncing(consensusID uint32) {
+	// validator send consensus block to state syncing
+	if val, ok := consensus.blocksReceived[consensusID]; ok {
+		consensus.mutex.Lock()
+		delete(consensus.blocksReceived, consensusID)
+		consensus.mutex.Unlock()
+
+		var blockObj types.Block
+		err := rlp.DecodeBytes(val.block, &blockObj)
+		if err != nil {
+			utils.GetLogInstance().Debug("failed to construct the cached block")
+			return
+		}
+		blockInfo := &BFTBlockInfo{Block: &blockObj, ConsensusID: consensusID}
+		select {
+		case consensus.ConsensusBlock <- blockInfo:
+		default:
+			utils.GetLogInstance().Warn("consensus block unable to sent to state sync", "height", blockObj.NumberU64(), "blockHash", blockObj.Hash().Hex())
+		}
+	}
+	return
+}
+
 // ProcessMessageValidator dispatches validator's consensus message.
 func (consensus *Consensus) ProcessMessageValidator(payload []byte) {
 	message := consensus_proto.Message{}
@@ -40,11 +64,20 @@ func (consensus *Consensus) processAnnounceMessage(message consensus_proto.Messa
 	blockHash := message.BlockHash
 	block := message.Payload
 
+	// Add block to received block cache
+	consensus.mutex.Lock()
+	consensus.blocksReceived[consensusID] = &BlockConsensusStatus{block, consensus.state}
+	consensus.mutex.Unlock()
+
 	copy(consensus.blockHash[:], blockHash[:])
 	consensus.block = block
 
-	if !consensus.checkConsensusMessage(message, consensus.leader.PubKey) {
+	if err := consensus.checkConsensusMessage(message, consensus.leader.PubKey); err != nil {
 		utils.GetLogInstance().Debug("Failed to check the leader message")
+		if err == ErrConsensusIDNotMatch {
+			utils.GetLogInstance().Debug("sending bft block to state syncing")
+			consensus.sendBFTBlockToStateSyncing(consensusID)
+		}
 		return
 	}
 
@@ -56,11 +89,6 @@ func (consensus *Consensus) processAnnounceMessage(message consensus_proto.Messa
 		return
 	}
 
-	// Add block to received block cache
-	consensus.mutex.Lock()
-	consensus.blocksReceived[consensusID] = &BlockConsensusStatus{block, consensus.state}
-	consensus.mutex.Unlock()
-
 	// Add attack model of IncorrectResponse
 	if attack.GetInstance().IncorrectResponse() {
 		utils.GetLogInstance().Warn("IncorrectResponse attacked")
@@ -102,8 +130,8 @@ func (consensus *Consensus) processPreparedMessage(message consensus_proto.Messa
 	// Update readyByConsensus for attack.
 	attack.GetInstance().UpdateConsensusReady(consensusID)
 
-	if !consensus.checkConsensusMessage(message, consensus.leader.PubKey) {
-		utils.GetLogInstance().Debug("Failed to check the leader message")
+	if err := consensus.checkConsensusMessage(message, consensus.leader.PubKey); err != nil {
+		utils.GetLogInstance().Debug("processPreparedMessage error", "error", err)
 		return
 	}
 
@@ -161,8 +189,8 @@ func (consensus *Consensus) processCommittedMessage(message consensus_proto.Mess
 	// Update readyByConsensus for attack.
 	attack.GetInstance().UpdateConsensusReady(consensusID)
 
-	if !consensus.checkConsensusMessage(message, consensus.leader.PubKey) {
-		utils.GetLogInstance().Debug("Failed to check the leader message")
+	if err := consensus.checkConsensusMessage(message, consensus.leader.PubKey); err != nil {
+		utils.GetLogInstance().Debug("processCommittedMessage error", "error", err)
 		return
 	}
 
@@ -195,6 +223,7 @@ func (consensus *Consensus) processCommittedMessage(message consensus_proto.Mess
 	consensus.state = CommittedDone
 	// TODO: the block catch up logic is a temporary workaround for full failure node catchup. Implement the full node catchup logic
 	// The logic is to roll up to the latest blocks one by one to try catching up with the leader.
+	// but because of checkConsensusMessage, the catchup logic will never be used here
 	for {
 		val, ok := consensus.blocksReceived[consensus.consensusID]
 		if ok {
@@ -205,10 +234,6 @@ func (consensus *Consensus) processCommittedMessage(message consensus_proto.Mess
 
 			var blockObj types.Block
 			err := rlp.DecodeBytes(val.block, &blockObj)
-			if err != nil {
-				utils.GetLogInstance().Warn("Unparseable block header data", "error", err)
-				return
-			}
 			if err != nil {
 				utils.GetLogInstance().Debug("failed to construct the new block after consensus")
 			}

consensus/errors.go

@@ -34,4 +34,10 @@ var (
 	// ErrInvalidNumber is returned if a block's number doesn't equal it's parent's
 	// plus one.
 	ErrInvalidNumber = errors.New("invalid block number")
+
+	// ErrConsensusIDNotMatch is returned if the current consensusID is not equal message's consensusID
+	ErrConsensusIDNotMatch = errors.New("consensusID not match")
+
+	// ErrInvalidConsensusMessage is returned is the consensus message received is invalid
+	ErrInvalidConsensusMessage = errors.New("invalid consensus message")
 )

internal/utils/flags.go

@@ -46,7 +46,10 @@ func StringsToAddrs(addrStrings []string) (maddrs []ma.Multiaddr, err error) {
 	return
 }
 
-// DefaultBootNodeAddrStrings is a list of Harmony bootnodes. Used to find other peers in the network.
+// DefaultBootNodeAddrStrings is a list of Harmony bootnodes address. Used to find other peers in the network.
 var DefaultBootNodeAddrStrings = []string{
 	"/ip4/127.0.0.1/tcp/9876/p2p/QmayB8NwxmfGE4Usb4H61M8uwbfc7LRbmXb3ChseJgbVuf",
 }
+
+// BootNodes is a list of boot nodes. It is populated either from default or from user CLI input.
+var BootNodes AddrList

node/node.go

@@ -26,6 +26,7 @@ import (
 	proto_node "github.com/harmony-one/harmony/api/proto/node"
 	service_manager "github.com/harmony-one/harmony/api/service"
 	blockproposal "github.com/harmony-one/harmony/api/service/blockproposal"
+	"github.com/harmony-one/harmony/api/service/clientsupport"
 	consensus_service "github.com/harmony-one/harmony/api/service/consensus"
 	"github.com/harmony-one/harmony/api/service/explorer"
 	"github.com/harmony-one/harmony/api/service/syncing"
@@ -91,7 +92,7 @@ func (state State) String() string {
 // Constants related to doing syncing.
 const (
 	lastMileThreshold = 4
-	inSyncThreshold   = 2
+	inSyncThreshold   = 1
 )
 
 const (
@@ -280,7 +281,7 @@ func New(host p2p.Host, consensus *bft.Consensus, db ethdb.Database) *Node {
 		node.Worker = worker.New(params.TestChainConfig, chain, node.Consensus, pki.GetAddressFromPublicKey(node.SelfPeer.PubKey), node.Consensus.ShardID)
 
 		node.AddSmartContractsToPendingTransactions()
-		node.Consensus.ConsensusBlock = make(chan *types.Block)
+		node.Consensus.ConsensusBlock = make(chan *bft.BFTBlockInfo)
 		node.Consensus.VerifiedNewBlock = make(chan *types.Block)
 	}
 
@@ -301,11 +302,16 @@ func New(host p2p.Host, consensus *bft.Consensus, db ethdb.Database) *Node {
 }
 
 // IsOutOfSync checks whether the node is out of sync by comparing latest block with consensus block
-func (node *Node) IsOutOfSync(consensusBlock *types.Block) bool {
+func (node *Node) IsOutOfSync(consensusBlockInfo *bft.BFTBlockInfo) bool {
+	consensusBlock := consensusBlockInfo.Block
+	consensusID := consensusBlockInfo.ConsensusID
+
 	myHeight := node.blockchain.CurrentBlock().NumberU64()
 	newHeight := consensusBlock.NumberU64()
 	utils.GetLogInstance().Debug("[SYNC]", "myHeight", myHeight, "newHeight", newHeight)
 	if newHeight-myHeight <= inSyncThreshold {
+		node.stateSync.AddLastMileBlock(consensusBlock)
+		node.Consensus.UpdateConsensusID(consensusID + 1)
 		return false
 	}
 	// cache latest blocks for last mile catch up
@@ -321,21 +327,24 @@ func (node *Node) DoSyncing() {
 		select {
 		// in current implementation logic, timeout means in sync
 		case <-time.After(5 * time.Second):
+			//myHeight := node.blockchain.CurrentBlock().NumberU64()
+			//utils.GetLogInstance().Debug("[SYNC]", "currentHeight", myHeight)
 			node.stateMutex.Lock()
 			node.State = NodeReadyForConsensus
 			node.stateMutex.Unlock()
 			continue
-		case consensusBlock := <-node.Consensus.ConsensusBlock:
-			// never reached from chao
-			if !node.IsOutOfSync(consensusBlock) {
+		case consensusBlockInfo := <-node.Consensus.ConsensusBlock:
+			if !node.IsOutOfSync(consensusBlockInfo) {
 				if node.State == NodeNotInSync {
 					utils.GetLogInstance().Info("[SYNC] Node is now IN SYNC!")
-					node.stateSync.CloseConnections()
-					node.stateSync = nil
 				}
 				node.stateMutex.Lock()
 				node.State = NodeReadyForConsensus
 				node.stateMutex.Unlock()
+				// wait for last mile block finish; think a better way
+				time.Sleep(200 * time.Millisecond)
+				node.stateSync.CloseConnections()
+				node.stateSync = nil
 				continue
 			} else {
 				utils.GetLogInstance().Debug("[SYNC] node is out of sync")
@@ -642,6 +651,8 @@ func (node *Node) setupForShardLeader() {
 	node.serviceManager.RegisterService(service_manager.Consensus, consensus_service.NewService(node.BlockChannel, node.Consensus))
 	// Register new block service.
 	node.serviceManager.RegisterService(service_manager.BlockProposal, blockproposal.NewService(node.Consensus.ReadySignal, node.WaitForConsensusReady))
+	// Register client support service.
+	node.serviceManager.RegisterService(service_manager.ClientSupport, clientsupport.NewService(node.blockchain.State, node.CallFaucetContract, node.SelfPeer.IP, node.SelfPeer.Port))
 }
 
 func (node *Node) setupForShardValidator() {

node/node_handler.go

@@ -200,17 +200,8 @@ func (node *Node) VerifyNewBlock(newBlock *types.Block) bool {
 	err := node.blockchain.ValidateNewBlock(newBlock, pki.GetAddressFromPublicKey(node.SelfPeer.PubKey))
 	if err != nil {
 		utils.GetLogInstance().Debug("Failed verifying new block", "Error", err, "tx", newBlock.Transactions()[0])
-
-		// send consensus block to state syncing
-		select {
-		case node.Consensus.ConsensusBlock <- newBlock:
-		default:
-			utils.GetLogInstance().Warn("consensus block unable to sent to state sync", "height", newBlock.NumberU64(), "blockHash", newBlock.Hash().Hex())
-		}
-
 		return false
 	}
-
 	return true
 }
 

p2p/host/hostv2/hostv2.go

@@ -13,6 +13,7 @@ import (
 	net "github.com/libp2p/go-libp2p-net"
 	peer "github.com/libp2p/go-libp2p-peer"
 	peerstore "github.com/libp2p/go-libp2p-peerstore"
+	protocol "github.com/libp2p/go-libp2p-protocol"
 	p2p_config "github.com/libp2p/go-libp2p/config"
 	ma "github.com/multiformats/go-multiaddr"
 )
@@ -103,7 +104,7 @@ func (host *HostV2) GetSelfPeer() p2p.Peer {
 
 // BindHandlerAndServe bind a streamHandler to the harmony protocol.
 func (host *HostV2) BindHandlerAndServe(handler p2p.StreamHandler) {
-	host.h.SetStreamHandler(ProtocolID, func(s net.Stream) {
+	host.h.SetStreamHandler(protocol.ID(ProtocolID), func(s net.Stream) {
 		handler(s)
 	})
 	// Hang forever
@@ -113,7 +114,8 @@ func (host *HostV2) BindHandlerAndServe(handler p2p.StreamHandler) {
 // SendMessage a p2p message sending function with signature compatible to p2pv1.
 func (host *HostV2) SendMessage(p p2p.Peer, message []byte) error {
 	logger := log.New("from", host.self, "to", p, "PeerID", p.PeerID)
-	s, err := host.h.NewStream(context.Background(), p.PeerID, ProtocolID)
+	host.Peerstore().AddProtocols(p.PeerID, ProtocolID)
+	s, err := host.h.NewStream(context.Background(), p.PeerID, protocol.ID(ProtocolID))
 	if err != nil {
 		logger.Error("NewStream() failed", "peerID", p.PeerID,
 			"protocolID", ProtocolID, "error", err)

vendor/github.com/libp2p/go-libp2p-kad-dht

@@ -0,0 +1 @@
+Subproject commit 838d43da02fc33899794e1c63fe4bd4d0bfd749a