plugeth/eth/protocols/eth/peer.go
Martin Holst Swende db03faa10d
core, eth: improve delivery speed on header requests (#23105)
This PR reduces the amount of work we do when answering header queries, e.g. when a peer
is syncing from us.

For some items, e.g block bodies, when we read the rlp-data from database, we plug it
directly into the response package. We didn't do that for headers, but instead read
headers-rlp, decode to types.Header, and re-encode to rlp. This PR changes that to keep it
in RLP-form as much as possible. When a node is syncing from us, it typically requests 192
contiguous headers. On master it has the following effect:

- For headers not in ancient: 2 db lookups. One for translating hash->number (even though
  the request is by number), and another for reading by hash (this latter one is sometimes
  cached).
  
- For headers in ancient: 1 file lookup/syscall for translating hash->number (even though
  the request is by number), and another for reading the header itself. After this, it
  also performes a hashing of the header, to ensure that the hash is what it expected. In
  this PR, I instead move the logic for "give me a sequence of blocks" into the lower
  layers, where the database can determine how and what to read from leveldb and/or
  ancients.

There are basically four types of requests; three of them are improved this way. The
fourth, by hash going backwards, is more tricky to optimize. However, since we know that
the gap is 0, we can look up by the parentHash, and stlil shave off all the number->hash
lookups.

The gapped collection can be optimized similarly, as a follow-up, at least in three out of
four cases.

Co-authored-by: Felix Lange <fjl@twurst.com>
2021-12-07 17:50:58 +01:00

523 lines
17 KiB
Go

// Copyright 2020 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 eth
import (
"math/big"
"math/rand"
"sync"
mapset "github.com/deckarep/golang-set"
"github.com/ethereum/go-ethereum/common"
"github.com/ethereum/go-ethereum/core/types"
"github.com/ethereum/go-ethereum/p2p"
"github.com/ethereum/go-ethereum/rlp"
)
const (
// maxKnownTxs is the maximum transactions hashes to keep in the known list
// before starting to randomly evict them.
maxKnownTxs = 32768
// maxKnownBlocks is the maximum block hashes to keep in the known list
// before starting to randomly evict them.
maxKnownBlocks = 1024
// maxQueuedTxs is the maximum number of transactions to queue up before dropping
// older broadcasts.
maxQueuedTxs = 4096
// maxQueuedTxAnns is the maximum number of transaction announcements to queue up
// before dropping older announcements.
maxQueuedTxAnns = 4096
// maxQueuedBlocks is the maximum number of block propagations to queue up before
// dropping broadcasts. There's not much point in queueing stale blocks, so a few
// that might cover uncles should be enough.
maxQueuedBlocks = 4
// maxQueuedBlockAnns is the maximum number of block announcements to queue up before
// dropping broadcasts. Similarly to block propagations, there's no point to queue
// above some healthy uncle limit, so use that.
maxQueuedBlockAnns = 4
)
// max is a helper function which returns the larger of the two given integers.
func max(a, b int) int {
if a > b {
return a
}
return b
}
// Peer is a collection of relevant information we have about a `eth` peer.
type Peer struct {
id string // Unique ID for the peer, cached
*p2p.Peer // The embedded P2P package peer
rw p2p.MsgReadWriter // Input/output streams for snap
version uint // Protocol version negotiated
head common.Hash // Latest advertised head block hash
td *big.Int // Latest advertised head block total difficulty
knownBlocks *knownCache // Set of block hashes known to be known by this peer
queuedBlocks chan *blockPropagation // Queue of blocks to broadcast to the peer
queuedBlockAnns chan *types.Block // Queue of blocks to announce to the peer
txpool TxPool // Transaction pool used by the broadcasters for liveness checks
knownTxs *knownCache // Set of transaction hashes known to be known by this peer
txBroadcast chan []common.Hash // Channel used to queue transaction propagation requests
txAnnounce chan []common.Hash // Channel used to queue transaction announcement requests
reqDispatch chan *request // Dispatch channel to send requests and track then until fulfilment
reqCancel chan *cancel // Dispatch channel to cancel pending requests and untrack them
resDispatch chan *response // Dispatch channel to fulfil pending requests and untrack them
term chan struct{} // Termination channel to stop the broadcasters
lock sync.RWMutex // Mutex protecting the internal fields
}
// NewPeer create a wrapper for a network connection and negotiated protocol
// version.
func NewPeer(version uint, p *p2p.Peer, rw p2p.MsgReadWriter, txpool TxPool) *Peer {
peer := &Peer{
id: p.ID().String(),
Peer: p,
rw: rw,
version: version,
knownTxs: newKnownCache(maxKnownTxs),
knownBlocks: newKnownCache(maxKnownBlocks),
queuedBlocks: make(chan *blockPropagation, maxQueuedBlocks),
queuedBlockAnns: make(chan *types.Block, maxQueuedBlockAnns),
txBroadcast: make(chan []common.Hash),
txAnnounce: make(chan []common.Hash),
reqDispatch: make(chan *request),
reqCancel: make(chan *cancel),
resDispatch: make(chan *response),
txpool: txpool,
term: make(chan struct{}),
}
// Start up all the broadcasters
go peer.broadcastBlocks()
go peer.broadcastTransactions()
go peer.announceTransactions()
go peer.dispatcher()
return peer
}
// Close signals the broadcast goroutine to terminate. Only ever call this if
// you created the peer yourself via NewPeer. Otherwise let whoever created it
// clean it up!
func (p *Peer) Close() {
close(p.term)
}
// ID retrieves the peer's unique identifier.
func (p *Peer) ID() string {
return p.id
}
// Version retrieves the peer's negoatiated `eth` protocol version.
func (p *Peer) Version() uint {
return p.version
}
// Head retrieves the current head hash and total difficulty of the peer.
func (p *Peer) Head() (hash common.Hash, td *big.Int) {
p.lock.RLock()
defer p.lock.RUnlock()
copy(hash[:], p.head[:])
return hash, new(big.Int).Set(p.td)
}
// SetHead updates the head hash and total difficulty of the peer.
func (p *Peer) SetHead(hash common.Hash, td *big.Int) {
p.lock.Lock()
defer p.lock.Unlock()
copy(p.head[:], hash[:])
p.td.Set(td)
}
// KnownBlock returns whether peer is known to already have a block.
func (p *Peer) KnownBlock(hash common.Hash) bool {
return p.knownBlocks.Contains(hash)
}
// KnownTransaction returns whether peer is known to already have a transaction.
func (p *Peer) KnownTransaction(hash common.Hash) bool {
return p.knownTxs.Contains(hash)
}
// markBlock marks a block as known for the peer, ensuring that the block will
// never be propagated to this particular peer.
func (p *Peer) markBlock(hash common.Hash) {
// If we reached the memory allowance, drop a previously known block hash
p.knownBlocks.Add(hash)
}
// markTransaction marks a transaction as known for the peer, ensuring that it
// will never be propagated to this particular peer.
func (p *Peer) markTransaction(hash common.Hash) {
// If we reached the memory allowance, drop a previously known transaction hash
p.knownTxs.Add(hash)
}
// SendTransactions sends transactions to the peer and includes the hashes
// in its transaction hash set for future reference.
//
// This method is a helper used by the async transaction sender. Don't call it
// directly as the queueing (memory) and transmission (bandwidth) costs should
// not be managed directly.
//
// The reasons this is public is to allow packages using this protocol to write
// tests that directly send messages without having to do the asyn queueing.
func (p *Peer) SendTransactions(txs types.Transactions) error {
// Mark all the transactions as known, but ensure we don't overflow our limits
for _, tx := range txs {
p.knownTxs.Add(tx.Hash())
}
return p2p.Send(p.rw, TransactionsMsg, txs)
}
// AsyncSendTransactions queues a list of transactions (by hash) to eventually
// propagate to a remote peer. The number of pending sends are capped (new ones
// will force old sends to be dropped)
func (p *Peer) AsyncSendTransactions(hashes []common.Hash) {
select {
case p.txBroadcast <- hashes:
// Mark all the transactions as known, but ensure we don't overflow our limits
p.knownTxs.Add(hashes...)
case <-p.term:
p.Log().Debug("Dropping transaction propagation", "count", len(hashes))
}
}
// sendPooledTransactionHashes sends transaction hashes to the peer and includes
// them in its transaction hash set for future reference.
//
// This method is a helper used by the async transaction announcer. Don't call it
// directly as the queueing (memory) and transmission (bandwidth) costs should
// not be managed directly.
func (p *Peer) sendPooledTransactionHashes(hashes []common.Hash) error {
// Mark all the transactions as known, but ensure we don't overflow our limits
p.knownTxs.Add(hashes...)
return p2p.Send(p.rw, NewPooledTransactionHashesMsg, NewPooledTransactionHashesPacket(hashes))
}
// AsyncSendPooledTransactionHashes queues a list of transactions hashes to eventually
// announce to a remote peer. The number of pending sends are capped (new ones
// will force old sends to be dropped)
func (p *Peer) AsyncSendPooledTransactionHashes(hashes []common.Hash) {
select {
case p.txAnnounce <- hashes:
// Mark all the transactions as known, but ensure we don't overflow our limits
p.knownTxs.Add(hashes...)
case <-p.term:
p.Log().Debug("Dropping transaction announcement", "count", len(hashes))
}
}
// ReplyPooledTransactionsRLP is the eth/66 version of SendPooledTransactionsRLP.
func (p *Peer) ReplyPooledTransactionsRLP(id uint64, hashes []common.Hash, txs []rlp.RawValue) error {
// Mark all the transactions as known, but ensure we don't overflow our limits
p.knownTxs.Add(hashes...)
// Not packed into PooledTransactionsPacket to avoid RLP decoding
return p2p.Send(p.rw, PooledTransactionsMsg, PooledTransactionsRLPPacket66{
RequestId: id,
PooledTransactionsRLPPacket: txs,
})
}
// SendNewBlockHashes announces the availability of a number of blocks through
// a hash notification.
func (p *Peer) SendNewBlockHashes(hashes []common.Hash, numbers []uint64) error {
// Mark all the block hashes as known, but ensure we don't overflow our limits
p.knownBlocks.Add(hashes...)
request := make(NewBlockHashesPacket, len(hashes))
for i := 0; i < len(hashes); i++ {
request[i].Hash = hashes[i]
request[i].Number = numbers[i]
}
return p2p.Send(p.rw, NewBlockHashesMsg, request)
}
// AsyncSendNewBlockHash queues the availability of a block for propagation to a
// remote peer. If the peer's broadcast queue is full, the event is silently
// dropped.
func (p *Peer) AsyncSendNewBlockHash(block *types.Block) {
select {
case p.queuedBlockAnns <- block:
// Mark all the block hash as known, but ensure we don't overflow our limits
p.knownBlocks.Add(block.Hash())
default:
p.Log().Debug("Dropping block announcement", "number", block.NumberU64(), "hash", block.Hash())
}
}
// SendNewBlock propagates an entire block to a remote peer.
func (p *Peer) SendNewBlock(block *types.Block, td *big.Int) error {
// Mark all the block hash as known, but ensure we don't overflow our limits
p.knownBlocks.Add(block.Hash())
return p2p.Send(p.rw, NewBlockMsg, &NewBlockPacket{
Block: block,
TD: td,
})
}
// AsyncSendNewBlock queues an entire block for propagation to a remote peer. If
// the peer's broadcast queue is full, the event is silently dropped.
func (p *Peer) AsyncSendNewBlock(block *types.Block, td *big.Int) {
select {
case p.queuedBlocks <- &blockPropagation{block: block, td: td}:
// Mark all the block hash as known, but ensure we don't overflow our limits
p.knownBlocks.Add(block.Hash())
default:
p.Log().Debug("Dropping block propagation", "number", block.NumberU64(), "hash", block.Hash())
}
}
// ReplyBlockHeaders is the eth/66 version of SendBlockHeaders.
func (p *Peer) ReplyBlockHeadersRLP(id uint64, headers []rlp.RawValue) error {
return p2p.Send(p.rw, BlockHeadersMsg, BlockHeadersRLPPacket66{
RequestId: id,
BlockHeadersRLPPacket: headers,
})
}
// ReplyBlockBodiesRLP is the eth/66 version of SendBlockBodiesRLP.
func (p *Peer) ReplyBlockBodiesRLP(id uint64, bodies []rlp.RawValue) error {
// Not packed into BlockBodiesPacket to avoid RLP decoding
return p2p.Send(p.rw, BlockBodiesMsg, BlockBodiesRLPPacket66{
RequestId: id,
BlockBodiesRLPPacket: bodies,
})
}
// ReplyNodeData is the eth/66 response to GetNodeData.
func (p *Peer) ReplyNodeData(id uint64, data [][]byte) error {
return p2p.Send(p.rw, NodeDataMsg, NodeDataPacket66{
RequestId: id,
NodeDataPacket: data,
})
}
// ReplyReceiptsRLP is the eth/66 response to GetReceipts.
func (p *Peer) ReplyReceiptsRLP(id uint64, receipts []rlp.RawValue) error {
return p2p.Send(p.rw, ReceiptsMsg, ReceiptsRLPPacket66{
RequestId: id,
ReceiptsRLPPacket: receipts,
})
}
// RequestOneHeader is a wrapper around the header query functions to fetch a
// single header. It is used solely by the fetcher.
func (p *Peer) RequestOneHeader(hash common.Hash, sink chan *Response) (*Request, error) {
p.Log().Debug("Fetching single header", "hash", hash)
id := rand.Uint64()
req := &Request{
id: id,
sink: sink,
code: GetBlockHeadersMsg,
want: BlockHeadersMsg,
data: &GetBlockHeadersPacket66{
RequestId: id,
GetBlockHeadersPacket: &GetBlockHeadersPacket{
Origin: HashOrNumber{Hash: hash},
Amount: uint64(1),
Skip: uint64(0),
Reverse: false,
},
},
}
if err := p.dispatchRequest(req); err != nil {
return nil, err
}
return req, nil
}
// RequestHeadersByHash fetches a batch of blocks' headers corresponding to the
// specified header query, based on the hash of an origin block.
func (p *Peer) RequestHeadersByHash(origin common.Hash, amount int, skip int, reverse bool, sink chan *Response) (*Request, error) {
p.Log().Debug("Fetching batch of headers", "count", amount, "fromhash", origin, "skip", skip, "reverse", reverse)
id := rand.Uint64()
req := &Request{
id: id,
sink: sink,
code: GetBlockHeadersMsg,
want: BlockHeadersMsg,
data: &GetBlockHeadersPacket66{
RequestId: id,
GetBlockHeadersPacket: &GetBlockHeadersPacket{
Origin: HashOrNumber{Hash: origin},
Amount: uint64(amount),
Skip: uint64(skip),
Reverse: reverse,
},
},
}
if err := p.dispatchRequest(req); err != nil {
return nil, err
}
return req, nil
}
// RequestHeadersByNumber fetches a batch of blocks' headers corresponding to the
// specified header query, based on the number of an origin block.
func (p *Peer) RequestHeadersByNumber(origin uint64, amount int, skip int, reverse bool, sink chan *Response) (*Request, error) {
p.Log().Debug("Fetching batch of headers", "count", amount, "fromnum", origin, "skip", skip, "reverse", reverse)
id := rand.Uint64()
req := &Request{
id: id,
sink: sink,
code: GetBlockHeadersMsg,
want: BlockHeadersMsg,
data: &GetBlockHeadersPacket66{
RequestId: id,
GetBlockHeadersPacket: &GetBlockHeadersPacket{
Origin: HashOrNumber{Number: origin},
Amount: uint64(amount),
Skip: uint64(skip),
Reverse: reverse,
},
},
}
if err := p.dispatchRequest(req); err != nil {
return nil, err
}
return req, nil
}
// RequestBodies fetches a batch of blocks' bodies corresponding to the hashes
// specified.
func (p *Peer) RequestBodies(hashes []common.Hash, sink chan *Response) (*Request, error) {
p.Log().Debug("Fetching batch of block bodies", "count", len(hashes))
id := rand.Uint64()
req := &Request{
id: id,
sink: sink,
code: GetBlockBodiesMsg,
want: BlockBodiesMsg,
data: &GetBlockBodiesPacket66{
RequestId: id,
GetBlockBodiesPacket: hashes,
},
}
if err := p.dispatchRequest(req); err != nil {
return nil, err
}
return req, nil
}
// RequestNodeData fetches a batch of arbitrary data from a node's known state
// data, corresponding to the specified hashes.
func (p *Peer) RequestNodeData(hashes []common.Hash, sink chan *Response) (*Request, error) {
p.Log().Debug("Fetching batch of state data", "count", len(hashes))
id := rand.Uint64()
req := &Request{
id: id,
sink: sink,
code: GetNodeDataMsg,
want: NodeDataMsg,
data: &GetNodeDataPacket66{
RequestId: id,
GetNodeDataPacket: hashes,
},
}
if err := p.dispatchRequest(req); err != nil {
return nil, err
}
return req, nil
}
// RequestReceipts fetches a batch of transaction receipts from a remote node.
func (p *Peer) RequestReceipts(hashes []common.Hash, sink chan *Response) (*Request, error) {
p.Log().Debug("Fetching batch of receipts", "count", len(hashes))
id := rand.Uint64()
req := &Request{
id: id,
sink: sink,
code: GetReceiptsMsg,
want: ReceiptsMsg,
data: &GetReceiptsPacket66{
RequestId: id,
GetReceiptsPacket: hashes,
},
}
if err := p.dispatchRequest(req); err != nil {
return nil, err
}
return req, nil
}
// RequestTxs fetches a batch of transactions from a remote node.
func (p *Peer) RequestTxs(hashes []common.Hash) error {
p.Log().Debug("Fetching batch of transactions", "count", len(hashes))
id := rand.Uint64()
requestTracker.Track(p.id, p.version, GetPooledTransactionsMsg, PooledTransactionsMsg, id)
return p2p.Send(p.rw, GetPooledTransactionsMsg, &GetPooledTransactionsPacket66{
RequestId: id,
GetPooledTransactionsPacket: hashes,
})
}
// knownCache is a cache for known hashes.
type knownCache struct {
hashes mapset.Set
max int
}
// newKnownCache creates a new knownCache with a max capacity.
func newKnownCache(max int) *knownCache {
return &knownCache{
max: max,
hashes: mapset.NewSet(),
}
}
// Add adds a list of elements to the set.
func (k *knownCache) Add(hashes ...common.Hash) {
for k.hashes.Cardinality() > max(0, k.max-len(hashes)) {
k.hashes.Pop()
}
for _, hash := range hashes {
k.hashes.Add(hash)
}
}
// Contains returns whether the given item is in the set.
func (k *knownCache) Contains(hash common.Hash) bool {
return k.hashes.Contains(hash)
}
// Cardinality returns the number of elements in the set.
func (k *knownCache) Cardinality() int {
return k.hashes.Cardinality()
}