path: root/core/consensus.go

// Copyright 2018 The dexon-consensus Authors
// This file is part of the dexon-consensus library.
//
// The dexon-consensus 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 dexon-consensus 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 dexon-consensus library. If not, see
// <http://www.gnu.org/licenses/>.

package core

import (
    "context"
    "encoding/hex"
    "fmt"
    "sync"
    "sync/atomic"
    "time"

    "github.com/dexon-foundation/dexon-consensus/common"
    "github.com/dexon-foundation/dexon-consensus/core/crypto"
    "github.com/dexon-foundation/dexon-consensus/core/db"
    "github.com/dexon-foundation/dexon-consensus/core/types"
    typesDKG "github.com/dexon-foundation/dexon-consensus/core/types/dkg"
    "github.com/dexon-foundation/dexon-consensus/core/utils"
)

// Errors for consensus core.
var (
    ErrProposerNotInNodeSet = fmt.Errorf(
        "proposer is not in node set")
    ErrIncorrectHash = fmt.Errorf(
        "hash of block is incorrect")
    ErrIncorrectSignature = fmt.Errorf(
        "signature of block is incorrect")
    ErrGenesisBlockNotEmpty = fmt.Errorf(
        "genesis block should be empty")
    ErrUnknownBlockProposed = fmt.Errorf(
        "unknown block is proposed")
    ErrIncorrectAgreementResultPosition = fmt.Errorf(
        "incorrect agreement result position")
    ErrNotEnoughVotes = fmt.Errorf(
        "not enought votes")
    ErrCRSNotReady = fmt.Errorf(
        "CRS not ready")
    ErrConfigurationNotReady = fmt.Errorf(
        "Configuration not ready")
)

// consensusBAReceiver implements agreementReceiver.
type consensusBAReceiver struct {
    // TODO(mission): consensus would be replaced by lattice and network.
    consensus        *Consensus
    agreementModule  *agreement
    chainID          uint32
    changeNotaryTime time.Time
    roundValue       *atomic.Value
    isNotary         bool
    restartNotary    chan types.Position
}

func (recv *consensusBAReceiver) round() uint64 {
    return recv.roundValue.Load().(uint64)
}

func (recv *consensusBAReceiver) ProposeVote(vote *types.Vote) {
    if !recv.isNotary {
        return
    }
    if err := recv.agreementModule.prepareVote(vote); err != nil {
        recv.consensus.logger.Error("Failed to prepare vote", "error", err)
        return
    }
    go func() {
        if err := recv.agreementModule.processVote(vote); err != nil {
            recv.consensus.logger.Error("Failed to process self vote",
                "error", err,
                "vote", vote)
            return
        }
        recv.consensus.logger.Debug("Calling Network.BroadcastVote",
            "vote", vote)
        recv.consensus.network.BroadcastVote(vote)
    }()
}

func (recv *consensusBAReceiver) ProposeBlock() common.Hash {
    if !recv.isNotary {
        return common.Hash{}
    }
    block := recv.consensus.proposeBlock(recv.chainID, recv.round())
    if block == nil {
        recv.consensus.logger.Error("unable to propose block")
        return nullBlockHash
    }
    go func() {
        if err := recv.consensus.preProcessBlock(block); err != nil {
            recv.consensus.logger.Error("Failed to pre-process block", "error", err)
            return
        }
        recv.consensus.logger.Debug("Calling Network.BroadcastBlock",
            "block", block)
        recv.consensus.network.BroadcastBlock(block)
    }()
    return block.Hash
}

func (recv *consensusBAReceiver) ConfirmBlock(
    hash common.Hash, votes map[types.NodeID]*types.Vote) {
    var block *types.Block
    isEmptyBlockConfirmed := hash == common.Hash{}
    if isEmptyBlockConfirmed {
        aID := recv.agreementModule.agreementID()
        recv.consensus.logger.Info("Empty block is confirmed",
            "position", &aID)
        var err error
        block, err = recv.consensus.proposeEmptyBlock(recv.round(), recv.chainID)
        if err != nil {
            recv.consensus.logger.Error("Propose empty block failed", "error", err)
            return
        }
    } else {
        var exist bool
        block, exist = recv.agreementModule.findBlockNoLock(hash)
        if !exist {
            recv.consensus.logger.Error("Unknown block confirmed",
                "hash", hash.String()[:6],
                "chainID", recv.chainID)
            ch := make(chan *types.Block)
            func() {
                recv.consensus.lock.Lock()
                defer recv.consensus.lock.Unlock()
                recv.consensus.baConfirmedBlock[hash] = ch
            }()
            go func() {
                hashes := common.Hashes{hash}
            PullBlockLoop:
                for {
                    recv.consensus.logger.Debug("Calling Network.PullBlock for BA block",
                        "hash", hash)
                    recv.consensus.network.PullBlocks(hashes)
                    select {
                    case block = <-ch:
                        break PullBlockLoop
                    case <-time.After(1 * time.Second):
                    }
                }
                recv.consensus.logger.Info("Receive unknown block",
                    "hash", hash.String()[:6],
                    "position", &block.Position,
                    "chainID", recv.chainID)
                recv.agreementModule.addCandidateBlock(block)
                recv.agreementModule.lock.Lock()
                defer recv.agreementModule.lock.Unlock()
                recv.ConfirmBlock(block.Hash, votes)
            }()
            return
        }
    }
    recv.consensus.ccModule.registerBlock(block)
    if block.Position.Height != 0 &&
        !recv.consensus.lattice.Exist(block.ParentHash) {
        go func(hash common.Hash) {
            parentHash := hash
            for {
                recv.consensus.logger.Warn("Parent block not confirmed",
                    "parent-hash", parentHash.String()[:6],
                    "cur-position", &block.Position)
                ch := make(chan *types.Block)
                if !func() bool {
                    recv.consensus.lock.Lock()
                    defer recv.consensus.lock.Unlock()
                    if _, exist := recv.consensus.baConfirmedBlock[parentHash]; exist {
                        return false
                    }
                    recv.consensus.baConfirmedBlock[parentHash] = ch
                    return true
                }() {
                    return
                }
                var block *types.Block
            PullBlockLoop:
                for {
                    recv.consensus.logger.Debug("Calling Network.PullBlock for parent",
                        "hash", parentHash)
                    recv.consensus.network.PullBlocks(common.Hashes{parentHash})
                    select {
                    case block = <-ch:
                        break PullBlockLoop
                    case <-time.After(1 * time.Second):
                    }
                }
                recv.consensus.logger.Info("Receive parent block",
                    "parent-hash", block.ParentHash.String()[:6],
                    "cur-position", &block.Position,
                    "chainID", recv.chainID)
                recv.consensus.ccModule.registerBlock(block)
                if err := recv.consensus.processBlock(block); err != nil {
                    recv.consensus.logger.Error("Failed to process block",
                        "block", block,
                        "error", err)
                    return
                }
                parentHash = block.ParentHash
                if block.Position.Height == 0 ||
                    recv.consensus.lattice.Exist(parentHash) {
                    return
                }
            }
        }(block.ParentHash)
    }
    if recv.isNotary {
        voteList := make([]types.Vote, 0, len(votes))
        for _, vote := range votes {
            if vote.BlockHash != hash {
                continue
            }
            voteList = append(voteList, *vote)
        }
        result := &types.AgreementResult{
            BlockHash:    block.Hash,
            Position:     block.Position,
            Votes:        voteList,
            IsEmptyBlock: isEmptyBlockConfirmed,
        }
        recv.consensus.logger.Debug("Propose AgreementResult",
            "result", result)
        recv.consensus.network.BroadcastAgreementResult(result)
    }
    if err := recv.consensus.processBlock(block); err != nil {
        recv.consensus.logger.Error("Failed to process block",
            "block", block,
            "error", err)
        return
    }
    // Clean the restartNotary channel so BA will not stuck by deadlock.
CleanChannelLoop:
    for {
        select {
        case <-recv.restartNotary:
        default:
            break CleanChannelLoop
        }
    }
    newPos := block.Position
    if block.Timestamp.After(recv.changeNotaryTime) {
        recv.roundValue.Store(recv.round() + 1)
        newPos.Round++
    }
    recv.restartNotary <- newPos
}

func (recv *consensusBAReceiver) PullBlocks(hashes common.Hashes) {
    if !recv.isNotary {
        return
    }
    recv.consensus.logger.Debug("Calling Network.PullBlocks", "hashes", hashes)
    recv.consensus.network.PullBlocks(hashes)
}

func (recv *consensusBAReceiver) ReportForkVote(v1, v2 *types.Vote) {
    recv.consensus.gov.ReportForkVote(v1, v2)
}

func (recv *consensusBAReceiver) ReportForkBlock(b1, b2 *types.Block) {
    recv.consensus.gov.ReportForkBlock(b1, b2)
}

// consensusDKGReceiver implements dkgReceiver.
type consensusDKGReceiver struct {
    ID           types.NodeID
    gov          Governance
    signer       *utils.Signer
    nodeSetCache *utils.NodeSetCache
    cfgModule    *configurationChain
    network      Network
    logger       common.Logger
}

// ProposeDKGComplaint proposes a DKGComplaint.
func (recv *consensusDKGReceiver) ProposeDKGComplaint(
    complaint *typesDKG.Complaint) {
    if err := recv.signer.SignDKGComplaint(complaint); err != nil {
        recv.logger.Error("Failed to sign DKG complaint", "error", err)
        return
    }
    recv.logger.Debug("Calling Governace.AddDKGComplaint",
        "complaint", complaint)
    recv.gov.AddDKGComplaint(complaint.Round, complaint)
}

// ProposeDKGMasterPublicKey propose a DKGMasterPublicKey.
func (recv *consensusDKGReceiver) ProposeDKGMasterPublicKey(
    mpk *typesDKG.MasterPublicKey) {
    if err := recv.signer.SignDKGMasterPublicKey(mpk); err != nil {
        recv.logger.Error("Failed to sign DKG master public key", "error", err)
        return
    }
    recv.logger.Debug("Calling Governance.AddDKGMasterPublicKey", "key", mpk)
    recv.gov.AddDKGMasterPublicKey(mpk.Round, mpk)
}

// ProposeDKGPrivateShare propose a DKGPrivateShare.
func (recv *consensusDKGReceiver) ProposeDKGPrivateShare(
    prv *typesDKG.PrivateShare) {
    if err := recv.signer.SignDKGPrivateShare(prv); err != nil {
        recv.logger.Error("Failed to sign DKG private share", "error", err)
        return
    }
    receiverPubKey, exists := recv.nodeSetCache.GetPublicKey(prv.ReceiverID)
    if !exists {
        recv.logger.Error("Public key for receiver not found",
            "receiver", prv.ReceiverID.String()[:6])
        return
    }
    if prv.ReceiverID == recv.ID {
        go func() {
            if err := recv.cfgModule.processPrivateShare(prv); err != nil {
                recv.logger.Error("Failed to process self private share", "prvShare", prv)
            }
        }()
    } else {
        recv.logger.Debug("Calling Network.SendDKGPrivateShare",
            "receiver", hex.EncodeToString(receiverPubKey.Bytes()))
        recv.network.SendDKGPrivateShare(receiverPubKey, prv)
    }
}

// ProposeDKGAntiNackComplaint propose a DKGPrivateShare as an anti complaint.
func (recv *consensusDKGReceiver) ProposeDKGAntiNackComplaint(
    prv *typesDKG.PrivateShare) {
    if prv.ProposerID == recv.ID {
        if err := recv.signer.SignDKGPrivateShare(prv); err != nil {
            recv.logger.Error("Failed sign DKG private share", "error", err)
            return
        }
    }
    recv.logger.Debug("Calling Network.BroadcastDKGPrivateShare", "share", prv)
    recv.network.BroadcastDKGPrivateShare(prv)
}

// ProposeDKGMPKReady propose a DKGMPKReady message.
func (recv *consensusDKGReceiver) ProposeDKGMPKReady(ready *typesDKG.MPKReady) {
    if err := recv.signer.SignDKGMPKReady(ready); err != nil {
        recv.logger.Error("Failed to sign DKG ready", "error", err)
        return
    }
    recv.logger.Debug("Calling Governance.AddDKGMPKReady", "ready", ready)
    recv.gov.AddDKGMPKReady(ready.Round, ready)
}

// ProposeDKGFinalize propose a DKGFinalize message.
func (recv *consensusDKGReceiver) ProposeDKGFinalize(final *typesDKG.Finalize) {
    if err := recv.signer.SignDKGFinalize(final); err != nil {
        recv.logger.Error("Failed to sign DKG finalize", "error", err)
        return
    }
    recv.logger.Debug("Calling Governance.AddDKGFinalize", "final", final)
    recv.gov.AddDKGFinalize(final.Round, final)
}

// Consensus implements DEXON Consensus algorithm.
type Consensus struct {
    // Node Info.
    ID     types.NodeID
    signer *utils.Signer

    // BA.
    baMgr            *agreementMgr
    baConfirmedBlock map[common.Hash]chan<- *types.Block

    // DKG.
    dkgRunning int32
    dkgReady   *sync.Cond
    cfgModule  *configurationChain

    // Dexon consensus v1's modules.
    lattice  *Lattice
    ccModule *compactionChain
    toSyncer *totalOrderingSyncer

    // Interfaces.
    db       db.Database
    app      Application
    debugApp Debug
    gov      Governance
    network  Network

    // Misc.
    dMoment                    time.Time
    nodeSetCache               *utils.NodeSetCache
    round                      uint64
    roundToNotify              uint64
    lock                       sync.RWMutex
    ctx                        context.Context
    ctxCancel                  context.CancelFunc
    event                      *common.Event
    logger                     common.Logger
    nonFinalizedBlockDelivered bool
    resetRandomnessTicker      chan struct{}
    resetDeliveryGuardTicker   chan struct{}
    msgChan                    chan interface{}
    waitGroup                  sync.WaitGroup

    // Context of Dummy receiver during switching from syncer.
    dummyCancel    context.CancelFunc
    dummyFinished  <-chan struct{}
    dummyMsgBuffer []interface{}
}

// NewConsensus construct an Consensus instance.
func NewConsensus(
    dMoment time.Time,
    app Application,
    gov Governance,
    db db.Database,
    network Network,
    prv crypto.PrivateKey,
    logger common.Logger) *Consensus {
    return newConsensusForRound(
        &types.Block{}, dMoment, app, gov, db, network, prv, logger, nil, true)
}

// NewConsensusForSimulation creates an instance of Consensus for simulation,
// the only difference with NewConsensus is nonblocking of app.
func NewConsensusForSimulation(
    dMoment time.Time,
    app Application,
    gov Governance,
    db db.Database,
    network Network,
    prv crypto.PrivateKey,
    logger common.Logger) *Consensus {
    return newConsensusForRound(
        &types.Block{}, dMoment, app, gov, db, network, prv, logger, nil, false)
}

// NewConsensusFromSyncer constructs an Consensus instance from information
// provided from syncer.
//
// You need to provide the initial block for this newly created Consensus
// instance to bootstrap with. A proper choice is the last finalized block you
// delivered to syncer.
//
// NOTE: those confirmed blocks should be organized by chainID and sorted by
//       their positions, in ascending order.
func NewConsensusFromSyncer(
    initBlock *types.Block,
    initRoundBeginTime time.Time,
    app Application,
    gov Governance,
    db db.Database,
    networkModule Network,
    prv crypto.PrivateKey,
    latticeModule *Lattice,
    confirmedBlocks [][]*types.Block,
    randomnessResults []*types.BlockRandomnessResult,
    cachedMessages []interface{},
    logger common.Logger) (*Consensus, error) {
    // Setup Consensus instance.
    con := newConsensusForRound(initBlock, initRoundBeginTime, app, gov, db,
        networkModule, prv, logger, latticeModule, true)
    // Launch a dummy receiver before we start receiving from network module.
    con.dummyMsgBuffer = cachedMessages
    con.dummyCancel, con.dummyFinished = utils.LaunchDummyReceiver(
        con.ctx, networkModule.ReceiveChan(), func(msg interface{}) {
            con.dummyMsgBuffer = append(con.dummyMsgBuffer, msg)
        })
    // Dump all BA-confirmed blocks to the consensus instance, make sure these
    // added blocks forming a DAG.
    for {
        updated := false
        for idx, bs := range confirmedBlocks {
            for bIdx, b := range bs {
                // Only when its parent block is already added to lattice, we can
                // then add this block. If not, our pulling mechanism would stop at
                // the block we added, and lost its parent block forever.
                if !latticeModule.Exist(b.ParentHash) {
                    logger.Debug("Skip discontinuous confirmed block",
                        "from", b,
                        "until", bs[len(bs)-1])
                    confirmedBlocks[idx] = bs[bIdx:]
                    break
                }
                con.ccModule.registerBlock(b)
                if err := con.processBlock(b); err != nil {
                    return nil, err
                }
            }
        }
        if !updated {
            break
        }
    }
    // Dump all randomness result to the consensus instance.
    for _, r := range randomnessResults {
        if err := con.ProcessBlockRandomnessResult(r, false); err != nil {
            con.logger.Error("failed to process randomness result when syncing",
                "result", r)
            continue
        }
    }
    return con, nil
}

// newConsensus creates a Consensus instance.
func newConsensusForRound(
    initBlock *types.Block,
    initRoundBeginTime time.Time,
    app Application,
    gov Governance,
    db db.Database,
    network Network,
    prv crypto.PrivateKey,
    logger common.Logger,
    latticeModule *Lattice,
    usingNonBlocking bool) *Consensus {

    // TODO(w): load latest blockHeight from DB, and use config at that height.
    nodeSetCache := utils.NewNodeSetCache(gov)
    // Setup signer module.
    signer := utils.NewSigner(prv)
    // Check if the application implement Debug interface.
    var debugApp Debug
    if a, ok := app.(Debug); ok {
        debugApp = a
    }
    // Get configuration for bootstrap round.
    initRound := initBlock.Position.Round
    initConfig := utils.GetConfigWithPanic(gov, initRound, logger)
    // Init lattice.
    if latticeModule == nil {
        latticeModule = NewLattice(initRoundBeginTime, initRound, initConfig,
            signer, app, debugApp, db, logger)
    }
    // Init configuration chain.
    ID := types.NewNodeID(prv.PublicKey())
    recv := &consensusDKGReceiver{
        ID:           ID,
        gov:          gov,
        signer:       signer,
        nodeSetCache: nodeSetCache,
        network:      network,
        logger:       logger,
    }
    cfgModule := newConfigurationChain(ID, recv, gov, nodeSetCache, db, logger)
    recv.cfgModule = cfgModule
    appModule := app
    if usingNonBlocking {
        appModule = newNonBlocking(app, debugApp)
    }
    // Construct Consensus instance.
    con := &Consensus{
        ID:                       ID,
        ccModule:                 newCompactionChain(gov),
        lattice:                  latticeModule,
        app:                      appModule,
        debugApp:                 debugApp,
        gov:                      gov,
        db:                       db,
        network:                  network,
        baConfirmedBlock:         make(map[common.Hash]chan<- *types.Block),
        dkgReady:                 sync.NewCond(&sync.Mutex{}),
        cfgModule:                cfgModule,
        dMoment:                  initRoundBeginTime,
        nodeSetCache:             nodeSetCache,
        signer:                   signer,
        event:                    common.NewEvent(),
        logger:                   logger,
        resetRandomnessTicker:    make(chan struct{}),
        resetDeliveryGuardTicker: make(chan struct{}),
        msgChan:                  make(chan interface{}, 1024),
    }
    con.ctx, con.ctxCancel = context.WithCancel(context.Background())
    con.baMgr = newAgreementMgr(con, initRound, initRoundBeginTime)
    if err := con.prepare(initBlock); err != nil {
        panic(err)
    }
    return con
}

// prepare the Consensus instance to be ready for blocks after 'initBlock'.
// 'initBlock' could be either:
//  - an empty block
//  - the last finalized block
func (con *Consensus) prepare(initBlock *types.Block) error {
    // The block past from full node should be delivered already or known by
    // full node. We don't have to notify it.
    con.roundToNotify = initBlock.Position.Round + 1
    initRound := initBlock.Position.Round
    initConfig := utils.GetConfigWithPanic(con.gov, initRound, con.logger)
    // Setup context.
    con.ccModule.init(initBlock)
    con.logger.Debug("Calling Governance.CRS", "round", initRound)
    initCRS := con.gov.CRS(initRound)
    if (initCRS == common.Hash{}) {
        return ErrCRSNotReady
    }
    if err := con.baMgr.appendConfig(initRound, initConfig, initCRS); err != nil {
        return err
    }
    // Setup lattice module.
    initPlusOneCfg := utils.GetConfigWithPanic(con.gov, initRound+1, con.logger)
    if err := con.lattice.AppendConfig(initRound+1, initPlusOneCfg); err != nil {
        if err == ErrRoundNotIncreasing {
            err = nil
        } else {
            return err
        }
    }
    // Register events.
    dkgSet, err := con.nodeSetCache.GetDKGSet(initRound)
    if err != nil {
        return err
    }
    if _, exist := dkgSet[con.ID]; exist {
        con.logger.Info("Selected as DKG set", "round", initRound)
        go func() {
            // Sleep until dMoment come.
            time.Sleep(con.dMoment.Sub(time.Now().UTC()))
            con.cfgModule.registerDKG(initRound, getDKGThreshold(initConfig))
            con.event.RegisterTime(con.dMoment.Add(initConfig.RoundInterval/4),
                func(time.Time) {
                    con.runDKG(initRound, initConfig)
                })
        }()
    }
    con.initialRound(con.dMoment, initRound, initConfig)
    return nil
}

// Run starts running DEXON Consensus.
func (con *Consensus) Run() {
    // Launch BA routines.
    con.baMgr.run()
    // Launch network handler.
    con.logger.Debug("Calling Network.ReceiveChan")
    con.waitGroup.Add(1)
    go con.deliverNetworkMsg()
    con.waitGroup.Add(1)
    go con.processMsg()
    // Sleep until dMoment come.
    time.Sleep(con.dMoment.Sub(time.Now().UTC()))
    // Take some time to bootstrap.
    time.Sleep(3 * time.Second)
    con.waitGroup.Add(1)
    go con.pullRandomness()
    // Stop dummy receiver if launched.
    if con.dummyCancel != nil {
        con.logger.Trace("Stop dummy receiver")
        con.dummyCancel()
        <-con.dummyFinished
        // Replay those cached messages.
        con.logger.Trace("Dummy receiver stoped, start dumping cached messages",
            "count", len(con.dummyMsgBuffer))
        for _, msg := range con.dummyMsgBuffer {
        loop:
            for {
                select {
                case con.msgChan <- msg:
                    break loop
                case <-time.After(50 * time.Millisecond):
                    con.logger.Debug(
                        "internal message channel is full when syncing")
                }
            }
        }
        con.logger.Trace("Finish dumping cached messages")
    }
    con.waitGroup.Add(1)
    go con.deliveryGuard()
    // Block until done.
    select {
    case <-con.ctx.Done():
    }
}

// runDKG starts running DKG protocol.
func (con *Consensus) runDKG(round uint64, config *types.Config) {
    con.dkgReady.L.Lock()
    defer con.dkgReady.L.Unlock()
    if con.dkgRunning != 0 {
        return
    }
    con.dkgRunning = 1
    go func() {
        startTime := time.Now().UTC()
        defer func() {
            con.dkgReady.L.Lock()
            defer con.dkgReady.L.Unlock()
            con.dkgReady.Broadcast()
            con.dkgRunning = 2
            DKGTime := time.Now().Sub(startTime)
            if DKGTime.Nanoseconds() >=
                config.RoundInterval.Nanoseconds()/2 {
                con.logger.Warn("Your computer cannot finish DKG on time!",
                    "nodeID", con.ID.String())
            }
        }()
        if err := con.cfgModule.runDKG(round); err != nil {
            con.logger.Error("Failed to runDKG", "error", err)
        }
    }()
}

func (con *Consensus) runCRS(round uint64) {
    for {
        con.logger.Debug("Calling Governance.CRS to check if already proposed",
            "round", round+1)
        if (con.gov.CRS(round+1) != common.Hash{}) {
            con.logger.Debug("CRS already proposed", "round", round+1)
            return
        }
        con.logger.Debug("Calling Governance.IsDKGFinal to check if ready to run CRS",
            "round", round)
        if con.cfgModule.isDKGFinal(round) {
            break
        }
        con.logger.Debug("DKG is not ready for running CRS. Retry later...",
            "round", round)
        time.Sleep(500 * time.Millisecond)
    }
    // Start running next round CRS.
    con.logger.Debug("Calling Governance.CRS", "round", round)
    psig, err := con.cfgModule.preparePartialSignature(
        round, utils.GetCRSWithPanic(con.gov, round, con.logger))
    if err != nil {
        con.logger.Error("Failed to prepare partial signature", "error", err)
    } else if err = con.signer.SignDKGPartialSignature(psig); err != nil {
        con.logger.Error("Failed to sign DKG partial signature", "error", err)
    } else if err = con.cfgModule.processPartialSignature(psig); err != nil {
        con.logger.Error("Failed to process partial signature", "error", err)
    } else {
        con.logger.Debug("Calling Network.BroadcastDKGPartialSignature",
            "proposer", psig.ProposerID,
            "round", psig.Round,
            "hash", psig.Hash)
        con.network.BroadcastDKGPartialSignature(psig)
        con.logger.Debug("Calling Governance.CRS", "round", round)
        crs, err := con.cfgModule.runCRSTSig(
            round, utils.GetCRSWithPanic(con.gov, round, con.logger))
        if err != nil {
            con.logger.Error("Failed to run CRS Tsig", "error", err)
        } else {
            con.logger.Debug("Calling Governance.ProposeCRS",
                "round", round+1,
                "crs", hex.EncodeToString(crs))
            con.gov.ProposeCRS(round+1, crs)
        }
    }
}

func (con *Consensus) initialRound(
    startTime time.Time, round uint64, config *types.Config) {
    select {
    case <-con.ctx.Done():
        return
    default:
    }
    curDkgSet, err := con.nodeSetCache.GetDKGSet(round)
    if err != nil {
        con.logger.Error("Error getting DKG set", "round", round, "error", err)
        curDkgSet = make(map[types.NodeID]struct{})
    }
    // Initiate CRS routine.
    if _, exist := curDkgSet[con.ID]; exist {
        con.event.RegisterTime(startTime.Add(config.RoundInterval/2),
            func(time.Time) {
                go func() {
                    con.runCRS(round)
                }()
            })
    }
    // checkCRS is a generator of checker to check if CRS for that round is
    // ready or not.
    checkCRS := func(round uint64) func() bool {
        return func() bool {
            nextCRS := con.gov.CRS(round)
            if (nextCRS != common.Hash{}) {
                return true
            }
            con.logger.Debug("CRS is not ready yet. Try again later...",
                "nodeID", con.ID,
                "round", round)
            return false
        }
    }
    // Initiate BA modules.
    con.event.RegisterTime(
        startTime.Add(config.RoundInterval/2+config.LambdaDKG),
        func(time.Time) {
            go func(nextRound uint64) {
                if !checkWithCancel(
                    con.ctx, 500*time.Millisecond, checkCRS(nextRound)) {
                    con.logger.Debug("unable to prepare CRS for baMgr",
                        "round", nextRound)
                    return
                }
                // Notify BA for new round.
                nextConfig := utils.GetConfigWithPanic(
                    con.gov, nextRound, con.logger)
                nextCRS := utils.GetCRSWithPanic(
                    con.gov, nextRound, con.logger)
                con.logger.Info("appendConfig for baMgr", "round", nextRound)
                if err := con.baMgr.appendConfig(
                    nextRound, nextConfig, nextCRS); err != nil {
                    panic(err)
                }
            }(round + 1)
        })
    // Initiate DKG for this round.
    con.event.RegisterTime(startTime.Add(config.RoundInterval/2+config.LambdaDKG),
        func(time.Time) {
            go func(nextRound uint64) {
                // Normally, gov.CRS would return non-nil. Use this for in case of
                // unexpected network fluctuation and ensure the robustness.
                if !checkWithCancel(
                    con.ctx, 500*time.Millisecond, checkCRS(nextRound)) {
                    con.logger.Debug("unable to prepare CRS for DKG set",
                        "round", nextRound)
                    return
                }
                nextDkgSet, err := con.nodeSetCache.GetDKGSet(nextRound)
                if err != nil {
                    con.logger.Error("Error getting DKG set",
                        "round", nextRound,
                        "error", err)
                    return
                }
                if _, exist := nextDkgSet[con.ID]; !exist {
                    return
                }
                con.logger.Info("Selected as DKG set", "round", nextRound)
                con.cfgModule.registerDKG(nextRound, getDKGThreshold(config))
                con.event.RegisterTime(
                    startTime.Add(config.RoundInterval*2/3),
                    func(time.Time) {
                        func() {
                            con.dkgReady.L.Lock()
                            defer con.dkgReady.L.Unlock()
                            con.dkgRunning = 0
                        }()
                        nextConfig := utils.GetConfigWithPanic(
                            con.gov, nextRound, con.logger)
                        con.runDKG(nextRound, nextConfig)
                    })
            }(round + 1)
        })
    // Prepare lattice module for next round and next "initialRound" routine.
    con.event.RegisterTime(startTime.Add(config.RoundInterval),
        func(time.Time) {
            // Change round.
            // Get configuration for next round.
            nextRound := round + 1
            nextConfig := utils.GetConfigWithPanic(con.gov, nextRound, con.logger)
            con.initialRound(
                startTime.Add(config.RoundInterval), nextRound, nextConfig)
        })
}

// Stop the Consensus core.
func (con *Consensus) Stop() {
    con.ctxCancel()
    con.baMgr.stop()
    con.event.Reset()
    con.waitGroup.Wait()
}

func (con *Consensus) deliverNetworkMsg() {
    defer con.waitGroup.Done()
    recv := con.network.ReceiveChan()
    for {
        select {
        case <-con.ctx.Done():
            return
        default:
        }
        select {
        case msg := <-recv:
        innerLoop:
            for {
                select {
                case con.msgChan <- msg:
                    break innerLoop
                case <-time.After(500 * time.Millisecond):
                    con.logger.Debug("internal message channel is full",
                        "pending", msg)
                }
            }
        case <-con.ctx.Done():
            return
        }
    }
}

func (con *Consensus) processMsg() {
    defer con.waitGroup.Done()
MessageLoop:
    for {
        select {
        case <-con.ctx.Done():
            return
        default:
        }
        var msg interface{}
        select {
        case msg = <-con.msgChan:
        case <-con.ctx.Done():
            return
        }
        switch val := msg.(type) {
        case *types.Block:
            if ch, exist := func() (chan<- *types.Block, bool) {
                con.lock.RLock()
                defer con.lock.RUnlock()
                ch, e := con.baConfirmedBlock[val.Hash]
                return ch, e
            }(); exist {
                if err := con.lattice.SanityCheck(val, false); err != nil {
                    if err == ErrRetrySanityCheckLater {
                        err = nil
                    } else {
                        con.logger.Error("SanityCheck failed", "error", err)
                        continue MessageLoop
                    }
                }
                func() {
                    con.lock.Lock()
                    defer con.lock.Unlock()
                    // In case of multiple delivered block.
                    if _, exist := con.baConfirmedBlock[val.Hash]; !exist {
                        return
                    }
                    delete(con.baConfirmedBlock, val.Hash)
                    ch <- val
                }()
            } else if val.IsFinalized() {
                // For sync mode.
                if err := con.processFinalizedBlock(val); err != nil {
                    con.logger.Error("Failed to process finalized block",
                        "block", val,
                        "error", err)
                }
            } else {
                if err := con.preProcessBlock(val); err != nil {
                    con.logger.Error("Failed to pre process block",
                        "block", val,
                        "error", err)
                }
            }
        case *types.Vote:
            if err := con.ProcessVote(val); err != nil {
                con.logger.Error("Failed to process vote",
                    "vote", val,
                    "error", err)
            }
        case *types.AgreementResult:
            if err := con.ProcessAgreementResult(val); err != nil {
                con.logger.Error("Failed to process agreement result",
                    "result", val,
                    "error", err)
            }
        case *types.BlockRandomnessResult:
            if err := con.ProcessBlockRandomnessResult(val, true); err != nil {
                con.logger.Error("Failed to process block randomness result",
                    "hash", val.BlockHash.String()[:6],
                    "position", &val.Position,
                    "error", err)
            }
        case *typesDKG.PrivateShare:
            if err := con.cfgModule.processPrivateShare(val); err != nil {
                con.logger.Error("Failed to process private share",
                    "error", err)
            }

        case *typesDKG.PartialSignature:
            if err := con.cfgModule.processPartialSignature(val); err != nil {
                con.logger.Error("Failed to process partial signature",
                    "error", err)
            }
        }
    }
}

func (con *Consensus) proposeBlock(chainID uint32, round uint64) *types.Block {
    block := &types.Block{
        Position: types.Position{
            ChainID: chainID,
            Round:   round,
        },
    }
    if err := con.prepareBlock(block, time.Now().UTC()); err != nil {
        con.logger.Error("Failed to prepare block", "error", err)
        return nil
    }
    return block
}

func (con *Consensus) proposeEmptyBlock(
    round uint64, chainID uint32) (*types.Block, error) {
    block := &types.Block{
        Position: types.Position{
            Round:   round,
            ChainID: chainID,
        },
    }
    if err := con.lattice.PrepareEmptyBlock(block); err != nil {
        return nil, err
    }
    return block, nil
}

// ProcessVote is the entry point to submit ont vote to a Consensus instance.
func (con *Consensus) ProcessVote(vote *types.Vote) (err error) {
    v := vote.Clone()
    err = con.baMgr.processVote(v)
    return
}

// ProcessAgreementResult processes the randomness request.
func (con *Consensus) ProcessAgreementResult(
    rand *types.AgreementResult) error {
    // Sanity Check.
    if err := VerifyAgreementResult(rand, con.nodeSetCache); err != nil {
        return err
    }
    con.lattice.AddShallowBlock(rand.BlockHash, rand.Position)
    // Syncing BA Module.
    if err := con.baMgr.processAgreementResult(rand); err != nil {
        return err
    }
    // Calculating randomness.
    if rand.Position.Round == 0 {
        return nil
    }
    // TODO(mission): find a way to avoid spamming by older agreement results.
    // Sanity check done.
    if !con.cfgModule.touchTSigHash(rand.BlockHash) {
        return nil
    }
    con.logger.Debug("Rebroadcast AgreementResult",
        "result", rand)
    con.network.BroadcastAgreementResult(rand)
    dkgSet, err := con.nodeSetCache.GetDKGSet(rand.Position.Round)
    if err != nil {
        return err
    }
    if _, exist := dkgSet[con.ID]; !exist {
        return nil
    }
    psig, err := con.cfgModule.preparePartialSignature(rand.Position.Round, rand.BlockHash)
    if err != nil {
        return err
    }
    if err = con.signer.SignDKGPartialSignature(psig); err != nil {
        return err
    }
    if err = con.cfgModule.processPartialSignature(psig); err != nil {
        return err
    }
    con.logger.Debug("Calling Network.BroadcastDKGPartialSignature",
        "proposer", psig.ProposerID,
        "round", psig.Round,
        "hash", psig.Hash.String()[:6])
    con.network.BroadcastDKGPartialSignature(psig)
    go func() {
        tsig, err := con.cfgModule.runTSig(rand.Position.Round, rand.BlockHash)
        if err != nil {
            if err != ErrTSigAlreadyRunning {
                con.logger.Error("Failed to run TSIG",
                    "position", &rand.Position,
                    "hash", rand.BlockHash.String()[:6],
                    "error", err)
            }
            return
        }
        result := &types.BlockRandomnessResult{
            BlockHash:  rand.BlockHash,
            Position:   rand.Position,
            Randomness: tsig.Signature,
        }
        if err := con.ProcessBlockRandomnessResult(result, true); err != nil {
            con.logger.Error("Failed to process randomness result",
                "error", err)
            return
        }
    }()
    return nil
}

// ProcessBlockRandomnessResult processes the randomness result.
func (con *Consensus) ProcessBlockRandomnessResult(
    rand *types.BlockRandomnessResult, needBroadcast bool) error {
    if rand.Position.Round == 0 {
        return nil
    }
    if err := con.ccModule.processBlockRandomnessResult(rand); err != nil {
        if err == ErrBlockNotRegistered {
            err = nil
        } else {
            return err
        }
    }
    if needBroadcast {
        con.logger.Debug("Calling Network.BroadcastRandomnessResult",
            "randomness", rand)
        con.network.BroadcastRandomnessResult(rand)
    }
    return con.deliverFinalizedBlocks()
}

// preProcessBlock performs Byzantine Agreement on the block.
func (con *Consensus) preProcessBlock(b *types.Block) (err error) {
    err = con.baMgr.processBlock(b)
    if err == nil && con.debugApp != nil {
        con.debugApp.BlockReceived(b.Hash)
    }
    return
}

func (con *Consensus) pullRandomness() {
    defer con.waitGroup.Done()
    for {
        select {
        case <-con.ctx.Done():
            return
        default:
        }
        select {
        case <-con.ctx.Done():
            return
        case <-con.resetRandomnessTicker:
        case <-time.After(1500 * time.Millisecond):
            // TODO(jimmy): pulling period should be related to lambdaBA.
            hashes := con.ccModule.pendingBlocksWithoutRandomness()
            if len(hashes) > 0 {
                con.logger.Debug(
                    "Calling Network.PullRandomness", "blocks", hashes)
                con.network.PullRandomness(hashes)
            }
        }
    }
}

func (con *Consensus) deliveryGuard() {
    defer con.waitGroup.Done()
    time.Sleep(con.dMoment.Sub(time.Now()))
    // Node takes time to start.
    time.Sleep(60 * time.Second)
    for {
        select {
        case <-con.ctx.Done():
            return
        default:
        }
        select {
        case <-con.ctx.Done():
            return
        case <-con.resetDeliveryGuardTicker:
        case <-time.After(60 * time.Second):
            con.logger.Error("no blocks delivered for too long", "ID", con.ID)
            panic(fmt.Errorf("no blocks delivered for too long"))
        }
    }
}

// deliverBlock deliver a block to application layer.
func (con *Consensus) deliverBlock(b *types.Block) {
    select {
    case con.resetRandomnessTicker <- struct{}{}:
    default:
    }
    select {
    case con.resetDeliveryGuardTicker <- struct{}{}:
    default:
    }
    if err := con.db.UpdateBlock(*b); err != nil {
        panic(err)
    }
    if err := con.db.PutCompactionChainTipInfo(
        b.Hash, b.Finalization.Height); err != nil {
        panic(err)
    }
    con.cfgModule.untouchTSigHash(b.Hash)
    con.logger.Debug("Calling Application.BlockDelivered", "block", b)
    con.app.BlockDelivered(b.Hash, b.Position, b.Finalization.Clone())
    if b.Position.Round == con.roundToNotify {
        // Get configuration for the round next to next round. Configuration
        // for that round should be ready at this moment and is required for
        // lattice module. This logic is related to:
        //  - roundShift
        //  - notifyGenesisRound
        futureRound := con.roundToNotify + 1
        futureConfig := utils.GetConfigWithPanic(con.gov, futureRound, con.logger)
        con.logger.Debug("Append Config", "round", futureRound)
        if err := con.lattice.AppendConfig(
            futureRound, futureConfig); err != nil {
            con.logger.Debug("Unable to append config",
                "round", futureRound,
                "error", err)
            panic(err)
        }
        // Only the first block delivered of that round would
        // trigger this noitification.
        con.logger.Debug("Calling Governance.NotifyRoundHeight",
            "round", con.roundToNotify,
            "height", b.Finalization.Height)
        con.gov.NotifyRoundHeight(
            con.roundToNotify, b.Finalization.Height)
        con.roundToNotify++
    }
    if con.debugApp != nil {
        con.debugApp.BlockReady(b.Hash)
    }
}

// deliverFinalizedBlocks extracts and delivers finalized blocks to application
// layer.
func (con *Consensus) deliverFinalizedBlocks() error {
    con.lock.Lock()
    defer con.lock.Unlock()
    return con.deliverFinalizedBlocksWithoutLock()
}

func (con *Consensus) deliverFinalizedBlocksWithoutLock() (err error) {
    deliveredBlocks := con.ccModule.extractBlocks()
    con.logger.Debug("Last blocks in compaction chain",
        "delivered", con.ccModule.lastDeliveredBlock(),
        "pending", con.ccModule.lastPendingBlock())
    for _, b := range deliveredBlocks {
        con.deliverBlock(b)
    }
    err = con.lattice.PurgeBlocks(deliveredBlocks)
    return
}

// processBlock is the entry point to submit one block to a Consensus instance.
func (con *Consensus) processBlock(block *types.Block) (err error) {
    if err = con.db.PutBlock(*block); err != nil && err != db.ErrBlockExists {
        return
    }
    con.lock.Lock()
    defer con.lock.Unlock()
    // Block processed by lattice can be out-of-order. But the output of lattice
    // (deliveredBlocks) cannot.
    deliveredBlocks, err := con.lattice.ProcessBlock(block)
    if err != nil {
        return
    }
    // Pass delivered blocks to compaction chain.
    for _, b := range deliveredBlocks {
        if b.IsFinalized() {
            if con.nonFinalizedBlockDelivered {
                panic(fmt.Errorf("attempting to skip finalized block: %s", b))
            }
            con.logger.Debug("skip delivery of finalized block",
                "block", b,
                "finalization-height", b.Finalization.Height)
            continue
        } else {
            // Mark that some non-finalized block delivered. After this flag
            // turned on, it's not allowed to deliver finalized blocks anymore.
            con.nonFinalizedBlockDelivered = true
        }
        if err = con.ccModule.processBlock(b); err != nil {
            return
        }
        go con.event.NotifyTime(b.Finalization.Timestamp)
    }
    if err = con.deliverFinalizedBlocksWithoutLock(); err != nil {
        return
    }
    return
}

// processFinalizedBlock is the entry point for handling finalized blocks.
func (con *Consensus) processFinalizedBlock(block *types.Block) error {
    return con.ccModule.processFinalizedBlock(block)
}

// PrepareBlock would setup header fields of block based on its ProposerID.
func (con *Consensus) prepareBlock(b *types.Block,
    proposeTime time.Time) (err error) {
    if err = con.lattice.PrepareBlock(b, proposeTime); err != nil {
        return
    }
    con.logger.Debug("Calling Governance.CRS", "round", b.Position.Round)
    crs := con.gov.CRS(b.Position.Round)
    if crs.Equal(common.Hash{}) {
        con.logger.Error("CRS for round is not ready, unable to prepare block",
            "position", &b.Position)
        err = ErrCRSNotReady
        return
    }
    err = con.signer.SignCRS(b, crs)
    return
}

// PrepareGenesisBlock would setup header fields for genesis block.
func (con *Consensus) PrepareGenesisBlock(b *types.Block,
    proposeTime time.Time) (err error) {
    if err = con.prepareBlock(b, proposeTime); err != nil {
        return
    }
    if len(b.Payload) != 0 {
        err = ErrGenesisBlockNotEmpty
        return
    }
    return
}