• Felix Lange's avatar
    p2p: ensure Server.loop is ticking even if discovery hangs (#20573) · d5acc5ed
    Felix Lange authored
    This is a temporary fix for a problem which started happening when the
    dialer was changed to read nodes from an enode.Iterator. Before the
    iterator change, discovery queries would always return within a couple
    seconds even if there was no Internet access. Since the iterator won't
    return unless a node is actually found, discoverTask can take much
    longer. This means that the 'emergency connect' logic might not execute
    in time, leading to a stuck node.
    d5acc5ed
server.go 32.3 KB
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152
// Copyright 2014 The go-ethereum Authors
// This file is part of the go-ethereum library.
//
// The go-ethereum library is free software: you can redistribute it and/or modify
// it under the terms of the GNU Lesser General Public License as published by
// the Free Software Foundation, either version 3 of the License, or
// (at your option) any later version.
//
// The go-ethereum library is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
// GNU Lesser General Public License for more details.
//
// You should have received a copy of the GNU Lesser General Public License
// along with the go-ethereum library. If not, see <http://www.gnu.org/licenses/>.

// Package p2p implements the Ethereum p2p network protocols.
package p2p

import (
	"bytes"
	"crypto/ecdsa"
	"encoding/hex"
	"errors"
	"fmt"
	"net"
	"sort"
	"sync"
	"sync/atomic"
	"time"

	"github.com/ethereum/go-ethereum/common"
	"github.com/ethereum/go-ethereum/common/mclock"
	"github.com/ethereum/go-ethereum/crypto"
	"github.com/ethereum/go-ethereum/event"
	"github.com/ethereum/go-ethereum/log"
	"github.com/ethereum/go-ethereum/p2p/discover"
	"github.com/ethereum/go-ethereum/p2p/discv5"
	"github.com/ethereum/go-ethereum/p2p/enode"
	"github.com/ethereum/go-ethereum/p2p/enr"
	"github.com/ethereum/go-ethereum/p2p/nat"
	"github.com/ethereum/go-ethereum/p2p/netutil"
)

const (
	defaultDialTimeout = 15 * time.Second

	// This is the fairness knob for the discovery mixer. When looking for peers, we'll
	// wait this long for a single source of candidates before moving on and trying other
	// sources.
	discmixTimeout = 5 * time.Second

	// Connectivity defaults.
	maxActiveDialTasks     = 16
	defaultMaxPendingPeers = 50
	defaultDialRatio       = 3

	// This time limits inbound connection attempts per source IP.
	inboundThrottleTime = 30 * time.Second

	// Maximum time allowed for reading a complete message.
	// This is effectively the amount of time a connection can be idle.
	frameReadTimeout = 30 * time.Second

	// Maximum amount of time allowed for writing a complete message.
	frameWriteTimeout = 20 * time.Second
)

var errServerStopped = errors.New("server stopped")

// Config holds Server options.
type Config struct {
	// This field must be set to a valid secp256k1 private key.
	PrivateKey *ecdsa.PrivateKey `toml:"-"`

	// MaxPeers is the maximum number of peers that can be
	// connected. It must be greater than zero.
	MaxPeers int

	// MaxPendingPeers is the maximum number of peers that can be pending in the
	// handshake phase, counted separately for inbound and outbound connections.
	// Zero defaults to preset values.
	MaxPendingPeers int `toml:",omitempty"`

	// DialRatio controls the ratio of inbound to dialed connections.
	// Example: a DialRatio of 2 allows 1/2 of connections to be dialed.
	// Setting DialRatio to zero defaults it to 3.
	DialRatio int `toml:",omitempty"`

	// NoDiscovery can be used to disable the peer discovery mechanism.
	// Disabling is useful for protocol debugging (manual topology).
	NoDiscovery bool

	// DiscoveryV5 specifies whether the new topic-discovery based V5 discovery
	// protocol should be started or not.
	DiscoveryV5 bool `toml:",omitempty"`

	// Name sets the node name of this server.
	// Use common.MakeName to create a name that follows existing conventions.
	Name string `toml:"-"`

	// BootstrapNodes are used to establish connectivity
	// with the rest of the network.
	BootstrapNodes []*enode.Node

	// BootstrapNodesV5 are used to establish connectivity
	// with the rest of the network using the V5 discovery
	// protocol.
	BootstrapNodesV5 []*discv5.Node `toml:",omitempty"`

	// Static nodes are used as pre-configured connections which are always
	// maintained and re-connected on disconnects.
	StaticNodes []*enode.Node

	// Trusted nodes are used as pre-configured connections which are always
	// allowed to connect, even above the peer limit.
	TrustedNodes []*enode.Node

	// Connectivity can be restricted to certain IP networks.
	// If this option is set to a non-nil value, only hosts which match one of the
	// IP networks contained in the list are considered.
	NetRestrict *netutil.Netlist `toml:",omitempty"`

	// NodeDatabase is the path to the database containing the previously seen
	// live nodes in the network.
	NodeDatabase string `toml:",omitempty"`

	// Protocols should contain the protocols supported
	// by the server. Matching protocols are launched for
	// each peer.
	Protocols []Protocol `toml:"-"`

	// If ListenAddr is set to a non-nil address, the server
	// will listen for incoming connections.
	//
	// If the port is zero, the operating system will pick a port. The
	// ListenAddr field will be updated with the actual address when
	// the server is started.
	ListenAddr string

	// If set to a non-nil value, the given NAT port mapper
	// is used to make the listening port available to the
	// Internet.
	NAT nat.Interface `toml:",omitempty"`

	// If Dialer is set to a non-nil value, the given Dialer
	// is used to dial outbound peer connections.
	Dialer NodeDialer `toml:"-"`

	// If NoDial is true, the server will not dial any peers.
	NoDial bool `toml:",omitempty"`

	// If EnableMsgEvents is set then the server will emit PeerEvents
	// whenever a message is sent to or received from a peer
	EnableMsgEvents bool

	// Logger is a custom logger to use with the p2p.Server.
	Logger log.Logger `toml:",omitempty"`
}

// Server manages all peer connections.
type Server struct {
	// Config fields may not be modified while the server is running.
	Config

	// Hooks for testing. These are useful because we can inhibit
	// the whole protocol stack.
	newTransport func(net.Conn) transport
	newPeerHook  func(*Peer)
	listenFunc   func(network, addr string) (net.Listener, error)

	lock    sync.Mutex // protects running
	running bool

	listener     net.Listener
	ourHandshake *protoHandshake
	loopWG       sync.WaitGroup // loop, listenLoop
	peerFeed     event.Feed
	log          log.Logger

	nodedb    *enode.DB
	localnode *enode.LocalNode
	ntab      *discover.UDPv4
	DiscV5    *discv5.Network
	discmix   *enode.FairMix

	staticNodeResolver nodeResolver

	// Channels into the run loop.
	quit                    chan struct{}
	addstatic               chan *enode.Node
	removestatic            chan *enode.Node
	addtrusted              chan *enode.Node
	removetrusted           chan *enode.Node
	peerOp                  chan peerOpFunc
	peerOpDone              chan struct{}
	delpeer                 chan peerDrop
	checkpointPostHandshake chan *conn
	checkpointAddPeer       chan *conn

	// State of run loop and listenLoop.
	inboundHistory expHeap
}

type peerOpFunc func(map[enode.ID]*Peer)

type peerDrop struct {
	*Peer
	err       error
	requested bool // true if signaled by the peer
}

type connFlag int32

const (
	dynDialedConn connFlag = 1 << iota
	staticDialedConn
	inboundConn
	trustedConn
)

// conn wraps a network connection with information gathered
// during the two handshakes.
type conn struct {
	fd net.Conn
	transport
	node  *enode.Node
	flags connFlag
	cont  chan error // The run loop uses cont to signal errors to SetupConn.
	caps  []Cap      // valid after the protocol handshake
	name  string     // valid after the protocol handshake
}

type transport interface {
	// The two handshakes.
	doEncHandshake(prv *ecdsa.PrivateKey, dialDest *ecdsa.PublicKey) (*ecdsa.PublicKey, error)
	doProtoHandshake(our *protoHandshake) (*protoHandshake, error)
	// The MsgReadWriter can only be used after the encryption
	// handshake has completed. The code uses conn.id to track this
	// by setting it to a non-nil value after the encryption handshake.
	MsgReadWriter
	// transports must provide Close because we use MsgPipe in some of
	// the tests. Closing the actual network connection doesn't do
	// anything in those tests because MsgPipe doesn't use it.
	close(err error)
}

func (c *conn) String() string {
	s := c.flags.String()
	if (c.node.ID() != enode.ID{}) {
		s += " " + c.node.ID().String()
	}
	s += " " + c.fd.RemoteAddr().String()
	return s
}

func (f connFlag) String() string {
	s := ""
	if f&trustedConn != 0 {
		s += "-trusted"
	}
	if f&dynDialedConn != 0 {
		s += "-dyndial"
	}
	if f&staticDialedConn != 0 {
		s += "-staticdial"
	}
	if f&inboundConn != 0 {
		s += "-inbound"
	}
	if s != "" {
		s = s[1:]
	}
	return s
}

func (c *conn) is(f connFlag) bool {
	flags := connFlag(atomic.LoadInt32((*int32)(&c.flags)))
	return flags&f != 0
}

func (c *conn) set(f connFlag, val bool) {
	for {
		oldFlags := connFlag(atomic.LoadInt32((*int32)(&c.flags)))
		flags := oldFlags
		if val {
			flags |= f
		} else {
			flags &= ^f
		}
		if atomic.CompareAndSwapInt32((*int32)(&c.flags), int32(oldFlags), int32(flags)) {
			return
		}
	}
}

// LocalNode returns the local node record.
func (srv *Server) LocalNode() *enode.LocalNode {
	return srv.localnode
}

// Peers returns all connected peers.
func (srv *Server) Peers() []*Peer {
	var ps []*Peer
	select {
	// Note: We'd love to put this function into a variable but
	// that seems to cause a weird compiler error in some
	// environments.
	case srv.peerOp <- func(peers map[enode.ID]*Peer) {
		for _, p := range peers {
			ps = append(ps, p)
		}
	}:
		<-srv.peerOpDone
	case <-srv.quit:
	}
	return ps
}

// PeerCount returns the number of connected peers.
func (srv *Server) PeerCount() int {
	var count int
	select {
	case srv.peerOp <- func(ps map[enode.ID]*Peer) { count = len(ps) }:
		<-srv.peerOpDone
	case <-srv.quit:
	}
	return count
}

// AddPeer connects to the given node and maintains the connection until the
// server is shut down. If the connection fails for any reason, the server will
// attempt to reconnect the peer.
func (srv *Server) AddPeer(node *enode.Node) {
	select {
	case srv.addstatic <- node:
	case <-srv.quit:
	}
}

// RemovePeer disconnects from the given node
func (srv *Server) RemovePeer(node *enode.Node) {
	select {
	case srv.removestatic <- node:
	case <-srv.quit:
	}
}

// AddTrustedPeer adds the given node to a reserved whitelist which allows the
// node to always connect, even if the slot are full.
func (srv *Server) AddTrustedPeer(node *enode.Node) {
	select {
	case srv.addtrusted <- node:
	case <-srv.quit:
	}
}

// RemoveTrustedPeer removes the given node from the trusted peer set.
func (srv *Server) RemoveTrustedPeer(node *enode.Node) {
	select {
	case srv.removetrusted <- node:
	case <-srv.quit:
	}
}

// SubscribePeers subscribes the given channel to peer events
func (srv *Server) SubscribeEvents(ch chan *PeerEvent) event.Subscription {
	return srv.peerFeed.Subscribe(ch)
}

// Self returns the local node's endpoint information.
func (srv *Server) Self() *enode.Node {
	srv.lock.Lock()
	ln := srv.localnode
	srv.lock.Unlock()

	if ln == nil {
		return enode.NewV4(&srv.PrivateKey.PublicKey, net.ParseIP("0.0.0.0"), 0, 0)
	}
	return ln.Node()
}

// Stop terminates the server and all active peer connections.
// It blocks until all active connections have been closed.
func (srv *Server) Stop() {
	srv.lock.Lock()
	if !srv.running {
		srv.lock.Unlock()
		return
	}
	srv.running = false
	if srv.listener != nil {
		// this unblocks listener Accept
		srv.listener.Close()
	}
	close(srv.quit)
	srv.lock.Unlock()
	srv.loopWG.Wait()
}

// sharedUDPConn implements a shared connection. Write sends messages to the underlying connection while read returns
// messages that were found unprocessable and sent to the unhandled channel by the primary listener.
type sharedUDPConn struct {
	*net.UDPConn
	unhandled chan discover.ReadPacket
}

// ReadFromUDP implements discv5.conn
func (s *sharedUDPConn) ReadFromUDP(b []byte) (n int, addr *net.UDPAddr, err error) {
	packet, ok := <-s.unhandled
	if !ok {
		return 0, nil, errors.New("connection was closed")
	}
	l := len(packet.Data)
	if l > len(b) {
		l = len(b)
	}
	copy(b[:l], packet.Data[:l])
	return l, packet.Addr, nil
}

// Close implements discv5.conn
func (s *sharedUDPConn) Close() error {
	return nil
}

// Start starts running the server.
// Servers can not be re-used after stopping.
func (srv *Server) Start() (err error) {
	srv.lock.Lock()
	defer srv.lock.Unlock()
	if srv.running {
		return errors.New("server already running")
	}
	srv.running = true
	srv.log = srv.Config.Logger
	if srv.log == nil {
		srv.log = log.Root()
	}
	if srv.NoDial && srv.ListenAddr == "" {
		srv.log.Warn("P2P server will be useless, neither dialing nor listening")
	}

	// static fields
	if srv.PrivateKey == nil {
		return errors.New("Server.PrivateKey must be set to a non-nil key")
	}
	if srv.newTransport == nil {
		srv.newTransport = newRLPX
	}
	if srv.listenFunc == nil {
		srv.listenFunc = net.Listen
	}
	if srv.Dialer == nil {
		srv.Dialer = TCPDialer{&net.Dialer{Timeout: defaultDialTimeout}}
	}
	srv.quit = make(chan struct{})
	srv.delpeer = make(chan peerDrop)
	srv.checkpointPostHandshake = make(chan *conn)
	srv.checkpointAddPeer = make(chan *conn)
	srv.addstatic = make(chan *enode.Node)
	srv.removestatic = make(chan *enode.Node)
	srv.addtrusted = make(chan *enode.Node)
	srv.removetrusted = make(chan *enode.Node)
	srv.peerOp = make(chan peerOpFunc)
	srv.peerOpDone = make(chan struct{})

	if err := srv.setupLocalNode(); err != nil {
		return err
	}
	if srv.ListenAddr != "" {
		if err := srv.setupListening(); err != nil {
			return err
		}
	}
	if err := srv.setupDiscovery(); err != nil {
		return err
	}

	dynPeers := srv.maxDialedConns()
	dialer := newDialState(srv.localnode.ID(), dynPeers, &srv.Config)
	srv.loopWG.Add(1)
	go srv.run(dialer)
	return nil
}

func (srv *Server) setupLocalNode() error {
	// Create the devp2p handshake.
	pubkey := crypto.FromECDSAPub(&srv.PrivateKey.PublicKey)
	srv.ourHandshake = &protoHandshake{Version: baseProtocolVersion, Name: srv.Name, ID: pubkey[1:]}
	for _, p := range srv.Protocols {
		srv.ourHandshake.Caps = append(srv.ourHandshake.Caps, p.cap())
	}
	sort.Sort(capsByNameAndVersion(srv.ourHandshake.Caps))

	// Create the local node.
	db, err := enode.OpenDB(srv.Config.NodeDatabase)
	if err != nil {
		return err
	}
	srv.nodedb = db
	srv.localnode = enode.NewLocalNode(db, srv.PrivateKey)
	srv.localnode.SetFallbackIP(net.IP{127, 0, 0, 1})
	// TODO: check conflicts
	for _, p := range srv.Protocols {
		for _, e := range p.Attributes {
			srv.localnode.Set(e)
		}
	}
	switch srv.NAT.(type) {
	case nil:
		// No NAT interface, do nothing.
	case nat.ExtIP:
		// ExtIP doesn't block, set the IP right away.
		ip, _ := srv.NAT.ExternalIP()
		srv.localnode.SetStaticIP(ip)
	default:
		// Ask the router about the IP. This takes a while and blocks startup,
		// do it in the background.
		srv.loopWG.Add(1)
		go func() {
			defer srv.loopWG.Done()
			if ip, err := srv.NAT.ExternalIP(); err == nil {
				srv.localnode.SetStaticIP(ip)
			}
		}()
	}
	return nil
}

func (srv *Server) setupDiscovery() error {
	srv.discmix = enode.NewFairMix(discmixTimeout)

	// Add protocol-specific discovery sources.
	added := make(map[string]bool)
	for _, proto := range srv.Protocols {
		if proto.DialCandidates != nil && !added[proto.Name] {
			srv.discmix.AddSource(proto.DialCandidates)
			added[proto.Name] = true
		}
	}

	// Don't listen on UDP endpoint if DHT is disabled.
	if srv.NoDiscovery && !srv.DiscoveryV5 {
		return nil
	}

	addr, err := net.ResolveUDPAddr("udp", srv.ListenAddr)
	if err != nil {
		return err
	}
	conn, err := net.ListenUDP("udp", addr)
	if err != nil {
		return err
	}
	realaddr := conn.LocalAddr().(*net.UDPAddr)
	srv.log.Debug("UDP listener up", "addr", realaddr)
	if srv.NAT != nil {
		if !realaddr.IP.IsLoopback() {
			go nat.Map(srv.NAT, srv.quit, "udp", realaddr.Port, realaddr.Port, "ethereum discovery")
		}
	}
	srv.localnode.SetFallbackUDP(realaddr.Port)

	// Discovery V4
	var unhandled chan discover.ReadPacket
	var sconn *sharedUDPConn
	if !srv.NoDiscovery {
		if srv.DiscoveryV5 {
			unhandled = make(chan discover.ReadPacket, 100)
			sconn = &sharedUDPConn{conn, unhandled}
		}
		cfg := discover.Config{
			PrivateKey:  srv.PrivateKey,
			NetRestrict: srv.NetRestrict,
			Bootnodes:   srv.BootstrapNodes,
			Unhandled:   unhandled,
			Log:         srv.log,
		}
		ntab, err := discover.ListenUDP(conn, srv.localnode, cfg)
		if err != nil {
			return err
		}
		srv.ntab = ntab
		srv.discmix.AddSource(ntab.RandomNodes())
		srv.staticNodeResolver = ntab
	}

	// Discovery V5
	if srv.DiscoveryV5 {
		var ntab *discv5.Network
		var err error
		if sconn != nil {
			ntab, err = discv5.ListenUDP(srv.PrivateKey, sconn, "", srv.NetRestrict)
		} else {
			ntab, err = discv5.ListenUDP(srv.PrivateKey, conn, "", srv.NetRestrict)
		}
		if err != nil {
			return err
		}
		if err := ntab.SetFallbackNodes(srv.BootstrapNodesV5); err != nil {
			return err
		}
		srv.DiscV5 = ntab
	}
	return nil
}

func (srv *Server) setupListening() error {
	// Launch the listener.
	listener, err := srv.listenFunc("tcp", srv.ListenAddr)
	if err != nil {
		return err
	}
	srv.listener = listener
	srv.ListenAddr = listener.Addr().String()

	// Update the local node record and map the TCP listening port if NAT is configured.
	if tcp, ok := listener.Addr().(*net.TCPAddr); ok {
		srv.localnode.Set(enr.TCP(tcp.Port))
		if !tcp.IP.IsLoopback() && srv.NAT != nil {
			srv.loopWG.Add(1)
			go func() {
				nat.Map(srv.NAT, srv.quit, "tcp", tcp.Port, tcp.Port, "ethereum p2p")
				srv.loopWG.Done()
			}()
		}
	}

	srv.loopWG.Add(1)
	go srv.listenLoop()
	return nil
}

type dialer interface {
	newTasks(running int, peers map[enode.ID]*Peer, now time.Time) []task
	taskDone(task, time.Time)
	addStatic(*enode.Node)
	removeStatic(*enode.Node)
}

func (srv *Server) run(dialstate dialer) {
	srv.log.Info("Started P2P networking", "self", srv.localnode.Node().URLv4())
	defer srv.loopWG.Done()
	defer srv.nodedb.Close()
	defer srv.discmix.Close()

	var (
		peers        = make(map[enode.ID]*Peer)
		inboundCount = 0
		trusted      = make(map[enode.ID]bool, len(srv.TrustedNodes))
		taskdone     = make(chan task, maxActiveDialTasks)
		tick         = time.NewTicker(30 * time.Second)
		runningTasks []task
		queuedTasks  []task // tasks that can't run yet
	)
	defer tick.Stop()

	// Put trusted nodes into a map to speed up checks.
	// Trusted peers are loaded on startup or added via AddTrustedPeer RPC.
	for _, n := range srv.TrustedNodes {
		trusted[n.ID()] = true
	}

	// removes t from runningTasks
	delTask := func(t task) {
		for i := range runningTasks {
			if runningTasks[i] == t {
				runningTasks = append(runningTasks[:i], runningTasks[i+1:]...)
				break
			}
		}
	}
	// starts until max number of active tasks is satisfied
	startTasks := func(ts []task) (rest []task) {
		i := 0
		for ; len(runningTasks) < maxActiveDialTasks && i < len(ts); i++ {
			t := ts[i]
			srv.log.Trace("New dial task", "task", t)
			go func() { t.Do(srv); taskdone <- t }()
			runningTasks = append(runningTasks, t)
		}
		return ts[i:]
	}
	scheduleTasks := func() {
		// Start from queue first.
		queuedTasks = append(queuedTasks[:0], startTasks(queuedTasks)...)
		// Query dialer for new tasks and start as many as possible now.
		if len(runningTasks) < maxActiveDialTasks {
			nt := dialstate.newTasks(len(runningTasks)+len(queuedTasks), peers, time.Now())
			queuedTasks = append(queuedTasks, startTasks(nt)...)
		}
	}

running:
	for {
		scheduleTasks()

		select {
		case <-tick.C:
			// This is just here to ensure the dial scheduler runs occasionally.

		case <-srv.quit:
			// The server was stopped. Run the cleanup logic.
			break running

		case n := <-srv.addstatic:
			// This channel is used by AddPeer to add to the
			// ephemeral static peer list. Add it to the dialer,
			// it will keep the node connected.
			srv.log.Trace("Adding static node", "node", n)
			dialstate.addStatic(n)

		case n := <-srv.removestatic:
			// This channel is used by RemovePeer to send a
			// disconnect request to a peer and begin the
			// stop keeping the node connected.
			srv.log.Trace("Removing static node", "node", n)
			dialstate.removeStatic(n)
			if p, ok := peers[n.ID()]; ok {
				p.Disconnect(DiscRequested)
			}

		case n := <-srv.addtrusted:
			// This channel is used by AddTrustedPeer to add an enode
			// to the trusted node set.
			srv.log.Trace("Adding trusted node", "node", n)
			trusted[n.ID()] = true
			// Mark any already-connected peer as trusted
			if p, ok := peers[n.ID()]; ok {
				p.rw.set(trustedConn, true)
			}

		case n := <-srv.removetrusted:
			// This channel is used by RemoveTrustedPeer to remove an enode
			// from the trusted node set.
			srv.log.Trace("Removing trusted node", "node", n)
			delete(trusted, n.ID())

			// Unmark any already-connected peer as trusted
			if p, ok := peers[n.ID()]; ok {
				p.rw.set(trustedConn, false)
			}

		case op := <-srv.peerOp:
			// This channel is used by Peers and PeerCount.
			op(peers)
			srv.peerOpDone <- struct{}{}

		case t := <-taskdone:
			// A task got done. Tell dialstate about it so it
			// can update its state and remove it from the active
			// tasks list.
			srv.log.Trace("Dial task done", "task", t)
			dialstate.taskDone(t, time.Now())
			delTask(t)

		case c := <-srv.checkpointPostHandshake:
			// A connection has passed the encryption handshake so
			// the remote identity is known (but hasn't been verified yet).
			if trusted[c.node.ID()] {
				// Ensure that the trusted flag is set before checking against MaxPeers.
				c.flags |= trustedConn
			}
			// TODO: track in-progress inbound node IDs (pre-Peer) to avoid dialing them.
			c.cont <- srv.postHandshakeChecks(peers, inboundCount, c)

		case c := <-srv.checkpointAddPeer:
			// At this point the connection is past the protocol handshake.
			// Its capabilities are known and the remote identity is verified.
			err := srv.addPeerChecks(peers, inboundCount, c)
			if err == nil {
				// The handshakes are done and it passed all checks.
				p := newPeer(srv.log, c, srv.Protocols)
				// If message events are enabled, pass the peerFeed
				// to the peer
				if srv.EnableMsgEvents {
					p.events = &srv.peerFeed
				}
				name := truncateName(c.name)
				p.log.Debug("Adding p2p peer", "addr", p.RemoteAddr(), "peers", len(peers)+1, "name", name)
				go srv.runPeer(p)
				peers[c.node.ID()] = p
				if p.Inbound() {
					inboundCount++
				}
				if conn, ok := c.fd.(*meteredConn); ok {
					conn.handshakeDone(p)
				}
			}
			// The dialer logic relies on the assumption that
			// dial tasks complete after the peer has been added or
			// discarded. Unblock the task last.
			c.cont <- err

		case pd := <-srv.delpeer:
			// A peer disconnected.
			d := common.PrettyDuration(mclock.Now() - pd.created)
			pd.log.Debug("Removing p2p peer", "addr", pd.RemoteAddr(), "peers", len(peers)-1, "duration", d, "req", pd.requested, "err", pd.err)
			delete(peers, pd.ID())
			if pd.Inbound() {
				inboundCount--
			}
		}
	}

	srv.log.Trace("P2P networking is spinning down")

	// Terminate discovery. If there is a running lookup it will terminate soon.
	if srv.ntab != nil {
		srv.ntab.Close()
	}
	if srv.DiscV5 != nil {
		srv.DiscV5.Close()
	}
	// Disconnect all peers.
	for _, p := range peers {
		p.Disconnect(DiscQuitting)
	}
	// Wait for peers to shut down. Pending connections and tasks are
	// not handled here and will terminate soon-ish because srv.quit
	// is closed.
	for len(peers) > 0 {
		p := <-srv.delpeer
		p.log.Trace("<-delpeer (spindown)", "remainingTasks", len(runningTasks))
		delete(peers, p.ID())
	}
}

func (srv *Server) postHandshakeChecks(peers map[enode.ID]*Peer, inboundCount int, c *conn) error {
	switch {
	case !c.is(trustedConn|staticDialedConn) && len(peers) >= srv.MaxPeers:
		return DiscTooManyPeers
	case !c.is(trustedConn) && c.is(inboundConn) && inboundCount >= srv.maxInboundConns():
		return DiscTooManyPeers
	case peers[c.node.ID()] != nil:
		return DiscAlreadyConnected
	case c.node.ID() == srv.localnode.ID():
		return DiscSelf
	default:
		return nil
	}
}

func (srv *Server) addPeerChecks(peers map[enode.ID]*Peer, inboundCount int, c *conn) error {
	// Drop connections with no matching protocols.
	if len(srv.Protocols) > 0 && countMatchingProtocols(srv.Protocols, c.caps) == 0 {
		return DiscUselessPeer
	}
	// Repeat the post-handshake checks because the
	// peer set might have changed since those checks were performed.
	return srv.postHandshakeChecks(peers, inboundCount, c)
}

func (srv *Server) maxInboundConns() int {
	return srv.MaxPeers - srv.maxDialedConns()
}

func (srv *Server) maxDialedConns() int {
	if srv.NoDiscovery || srv.NoDial {
		return 0
	}
	r := srv.DialRatio
	if r == 0 {
		r = defaultDialRatio
	}
	return srv.MaxPeers / r
}

// listenLoop runs in its own goroutine and accepts
// inbound connections.
func (srv *Server) listenLoop() {
	srv.log.Debug("TCP listener up", "addr", srv.listener.Addr())

	// The slots channel limits accepts of new connections.
	tokens := defaultMaxPendingPeers
	if srv.MaxPendingPeers > 0 {
		tokens = srv.MaxPendingPeers
	}
	slots := make(chan struct{}, tokens)
	for i := 0; i < tokens; i++ {
		slots <- struct{}{}
	}

	// Wait for slots to be returned on exit. This ensures all connection goroutines
	// are down before listenLoop returns.
	defer srv.loopWG.Done()
	defer func() {
		for i := 0; i < cap(slots); i++ {
			<-slots
		}
	}()

	for {
		// Wait for a free slot before accepting.
		<-slots

		var (
			fd  net.Conn
			err error
		)
		for {
			fd, err = srv.listener.Accept()
			if netutil.IsTemporaryError(err) {
				srv.log.Debug("Temporary read error", "err", err)
				continue
			} else if err != nil {
				srv.log.Debug("Read error", "err", err)
				slots <- struct{}{}
				return
			}
			break
		}

		remoteIP := netutil.AddrIP(fd.RemoteAddr())
		if err := srv.checkInboundConn(fd, remoteIP); err != nil {
			srv.log.Debug("Rejected inbound connnection", "addr", fd.RemoteAddr(), "err", err)
			fd.Close()
			slots <- struct{}{}
			continue
		}
		if remoteIP != nil {
			var addr *net.TCPAddr
			if tcp, ok := fd.RemoteAddr().(*net.TCPAddr); ok {
				addr = tcp
			}
			fd = newMeteredConn(fd, true, addr)
			srv.log.Trace("Accepted connection", "addr", fd.RemoteAddr())
		}
		go func() {
			srv.SetupConn(fd, inboundConn, nil)
			slots <- struct{}{}
		}()
	}
}

func (srv *Server) checkInboundConn(fd net.Conn, remoteIP net.IP) error {
	if remoteIP != nil {
		// Reject connections that do not match NetRestrict.
		if srv.NetRestrict != nil && !srv.NetRestrict.Contains(remoteIP) {
			return fmt.Errorf("not whitelisted in NetRestrict")
		}
		// Reject Internet peers that try too often.
		srv.inboundHistory.expire(time.Now())
		if !netutil.IsLAN(remoteIP) && srv.inboundHistory.contains(remoteIP.String()) {
			return fmt.Errorf("too many attempts")
		}
		srv.inboundHistory.add(remoteIP.String(), time.Now().Add(inboundThrottleTime))
	}
	return nil
}

// SetupConn runs the handshakes and attempts to add the connection
// as a peer. It returns when the connection has been added as a peer
// or the handshakes have failed.
func (srv *Server) SetupConn(fd net.Conn, flags connFlag, dialDest *enode.Node) error {
	c := &conn{fd: fd, transport: srv.newTransport(fd), flags: flags, cont: make(chan error)}
	err := srv.setupConn(c, flags, dialDest)
	if err != nil {
		c.close(err)
		srv.log.Trace("Setting up connection failed", "addr", fd.RemoteAddr(), "err", err)
	}
	return err
}

func (srv *Server) setupConn(c *conn, flags connFlag, dialDest *enode.Node) error {
	// Prevent leftover pending conns from entering the handshake.
	srv.lock.Lock()
	running := srv.running
	srv.lock.Unlock()
	if !running {
		return errServerStopped
	}

	// If dialing, figure out the remote public key.
	var dialPubkey *ecdsa.PublicKey
	if dialDest != nil {
		dialPubkey = new(ecdsa.PublicKey)
		if err := dialDest.Load((*enode.Secp256k1)(dialPubkey)); err != nil {
			return errors.New("dial destination doesn't have a secp256k1 public key")
		}
	}

	// Run the RLPx handshake.
	remotePubkey, err := c.doEncHandshake(srv.PrivateKey, dialPubkey)
	if err != nil {
		srv.log.Trace("Failed RLPx handshake", "addr", c.fd.RemoteAddr(), "conn", c.flags, "err", err)
		return err
	}
	if dialDest != nil {
		// For dialed connections, check that the remote public key matches.
		if dialPubkey.X.Cmp(remotePubkey.X) != 0 || dialPubkey.Y.Cmp(remotePubkey.Y) != 0 {
			return DiscUnexpectedIdentity
		}
		c.node = dialDest
	} else {
		c.node = nodeFromConn(remotePubkey, c.fd)
	}
	clog := srv.log.New("id", c.node.ID(), "addr", c.fd.RemoteAddr(), "conn", c.flags)
	err = srv.checkpoint(c, srv.checkpointPostHandshake)
	if err != nil {
		clog.Trace("Rejected peer", "err", err)
		return err
	}

	// Run the capability negotiation handshake.
	phs, err := c.doProtoHandshake(srv.ourHandshake)
	if err != nil {
		clog.Trace("Failed proto handshake", "err", err)
		return err
	}
	if id := c.node.ID(); !bytes.Equal(crypto.Keccak256(phs.ID), id[:]) {
		clog.Trace("Wrong devp2p handshake identity", "phsid", hex.EncodeToString(phs.ID))
		return DiscUnexpectedIdentity
	}
	c.caps, c.name = phs.Caps, phs.Name
	err = srv.checkpoint(c, srv.checkpointAddPeer)
	if err != nil {
		clog.Trace("Rejected peer", "err", err)
		return err
	}

	// If the checks completed successfully, the connection has been added as a peer and
	// runPeer has been launched.
	clog.Trace("Connection set up", "inbound", dialDest == nil)
	return nil
}

func nodeFromConn(pubkey *ecdsa.PublicKey, conn net.Conn) *enode.Node {
	var ip net.IP
	var port int
	if tcp, ok := conn.RemoteAddr().(*net.TCPAddr); ok {
		ip = tcp.IP
		port = tcp.Port
	}
	return enode.NewV4(pubkey, ip, port, port)
}

func truncateName(s string) string {
	if len(s) > 20 {
		return s[:20] + "..."
	}
	return s
}

// checkpoint sends the conn to run, which performs the
// post-handshake checks for the stage (posthandshake, addpeer).
func (srv *Server) checkpoint(c *conn, stage chan<- *conn) error {
	select {
	case stage <- c:
	case <-srv.quit:
		return errServerStopped
	}
	return <-c.cont
}

// runPeer runs in its own goroutine for each peer.
// it waits until the Peer logic returns and removes
// the peer.
func (srv *Server) runPeer(p *Peer) {
	if srv.newPeerHook != nil {
		srv.newPeerHook(p)
	}

	// broadcast peer add
	srv.peerFeed.Send(&PeerEvent{
		Type:          PeerEventTypeAdd,
		Peer:          p.ID(),
		RemoteAddress: p.RemoteAddr().String(),
		LocalAddress:  p.LocalAddr().String(),
	})

	// run the protocol
	remoteRequested, err := p.run()

	// broadcast peer drop
	srv.peerFeed.Send(&PeerEvent{
		Type:          PeerEventTypeDrop,
		Peer:          p.ID(),
		Error:         err.Error(),
		RemoteAddress: p.RemoteAddr().String(),
		LocalAddress:  p.LocalAddr().String(),
	})

	// Note: run waits for existing peers to be sent on srv.delpeer
	// before returning, so this send should not select on srv.quit.
	srv.delpeer <- peerDrop{p, err, remoteRequested}
}

// NodeInfo represents a short summary of the information known about the host.
type NodeInfo struct {
	ID    string `json:"id"`    // Unique node identifier (also the encryption key)
	Name  string `json:"name"`  // Name of the node, including client type, version, OS, custom data
	Enode string `json:"enode"` // Enode URL for adding this peer from remote peers
	ENR   string `json:"enr"`   // Ethereum Node Record
	IP    string `json:"ip"`    // IP address of the node
	Ports struct {
		Discovery int `json:"discovery"` // UDP listening port for discovery protocol
		Listener  int `json:"listener"`  // TCP listening port for RLPx
	} `json:"ports"`
	ListenAddr string                 `json:"listenAddr"`
	Protocols  map[string]interface{} `json:"protocols"`
}

// NodeInfo gathers and returns a collection of metadata known about the host.
func (srv *Server) NodeInfo() *NodeInfo {
	// Gather and assemble the generic node infos
	node := srv.Self()
	info := &NodeInfo{
		Name:       srv.Name,
		Enode:      node.URLv4(),
		ID:         node.ID().String(),
		IP:         node.IP().String(),
		ListenAddr: srv.ListenAddr,
		Protocols:  make(map[string]interface{}),
	}
	info.Ports.Discovery = node.UDP()
	info.Ports.Listener = node.TCP()
	info.ENR = node.String()

	// Gather all the running protocol infos (only once per protocol type)
	for _, proto := range srv.Protocols {
		if _, ok := info.Protocols[proto.Name]; !ok {
			nodeInfo := interface{}("unknown")
			if query := proto.NodeInfo; query != nil {
				nodeInfo = proto.NodeInfo()
			}
			info.Protocols[proto.Name] = nodeInfo
		}
	}
	return info
}

// PeersInfo returns an array of metadata objects describing connected peers.
func (srv *Server) PeersInfo() []*PeerInfo {
	// Gather all the generic and sub-protocol specific infos
	infos := make([]*PeerInfo, 0, srv.PeerCount())
	for _, peer := range srv.Peers() {
		if peer != nil {
			infos = append(infos, peer.Info())
		}
	}
	// Sort the result array alphabetically by node identifier
	for i := 0; i < len(infos); i++ {
		for j := i + 1; j < len(infos); j++ {
			if infos[i].ID > infos[j].ID {
				infos[i], infos[j] = infos[j], infos[i]
			}
		}
	}
	return infos
}