quorum/eth/downloader/queue.go

887 lines
31 KiB
Go
Raw Normal View History

2015-07-06 17:54:22 -07:00
// Copyright 2015 The go-ethereum Authors
// This file is part of the go-ethereum library.
2015-07-06 17:54:22 -07:00
//
// The go-ethereum library is free software: you can redistribute it and/or modify
2015-07-06 17:54:22 -07:00
// 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,
2015-07-06 17:54:22 -07:00
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
2015-07-06 17:54:22 -07:00
// 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/>.
2015-07-06 17:54:22 -07:00
// Contains the block download scheduler to collect download tasks and schedule
// them in an ordered, and throttled way.
2015-04-12 03:38:25 -07:00
package downloader
import (
"errors"
"fmt"
2015-04-12 03:38:25 -07:00
"sync"
"time"
"github.com/ethereum/go-ethereum/common"
2015-04-12 03:38:25 -07:00
"github.com/ethereum/go-ethereum/core/types"
"github.com/ethereum/go-ethereum/log"
"github.com/ethereum/go-ethereum/metrics"
"gopkg.in/karalabe/cookiejar.v2/collections/prque"
2015-04-12 03:38:25 -07:00
)
var (
blockCacheItems = 8192 // Maximum number of blocks to cache before throttling the download
blockCacheMemory = 64 * 1024 * 1024 // Maximum amount of memory to use for block caching
blockCacheSizeWeight = 0.1 // Multiplier to approximate the average block size based on past ones
)
var (
errNoFetchesPending = errors.New("no fetches pending")
errStaleDelivery = errors.New("stale delivery")
)
// fetchRequest is a currently running data retrieval operation.
type fetchRequest struct {
Peer *peerConnection // Peer to which the request was sent
From uint64 // [eth/62] Requested chain element index (used for skeleton fills only)
Headers []*types.Header // [eth/62] Requested headers, sorted by request order
Time time.Time // Time when the request was made
}
// fetchResult is a struct collecting partial results from data fetchers until
// all outstanding pieces complete and the result as a whole can be processed.
type fetchResult struct {
Pending int // Number of data fetches still pending
Hash common.Hash // Hash of the header to prevent recalculating
Header *types.Header
Uncles []*types.Header
Transactions types.Transactions
Receipts types.Receipts
}
2015-04-12 03:38:25 -07:00
// queue represents hashes that are either need fetching or are being fetched
type queue struct {
mode SyncMode // Synchronisation mode to decide on the block parts to schedule for fetching
// Headers are "special", they download in batches, supported by a skeleton chain
headerHead common.Hash // [eth/62] Hash of the last queued header to verify order
headerTaskPool map[uint64]*types.Header // [eth/62] Pending header retrieval tasks, mapping starting indexes to skeleton headers
headerTaskQueue *prque.Prque // [eth/62] Priority queue of the skeleton indexes to fetch the filling headers for
headerPeerMiss map[string]map[uint64]struct{} // [eth/62] Set of per-peer header batches known to be unavailable
headerPendPool map[string]*fetchRequest // [eth/62] Currently pending header retrieval operations
headerResults []*types.Header // [eth/62] Result cache accumulating the completed headers
headerProced int // [eth/62] Number of headers already processed from the results
headerOffset uint64 // [eth/62] Number of the first header in the result cache
headerContCh chan bool // [eth/62] Channel to notify when header download finishes
// All data retrievals below are based on an already assembles header chain
blockTaskPool map[common.Hash]*types.Header // [eth/62] Pending block (body) retrieval tasks, mapping hashes to headers
blockTaskQueue *prque.Prque // [eth/62] Priority queue of the headers to fetch the blocks (bodies) for
blockPendPool map[string]*fetchRequest // [eth/62] Currently pending block (body) retrieval operations
blockDonePool map[common.Hash]struct{} // [eth/62] Set of the completed block (body) fetches
2015-04-12 03:38:25 -07:00
receiptTaskPool map[common.Hash]*types.Header // [eth/63] Pending receipt retrieval tasks, mapping hashes to headers
receiptTaskQueue *prque.Prque // [eth/63] Priority queue of the headers to fetch the receipts for
receiptPendPool map[string]*fetchRequest // [eth/63] Currently pending receipt retrieval operations
receiptDonePool map[common.Hash]struct{} // [eth/63] Set of the completed receipt fetches
resultCache []*fetchResult // Downloaded but not yet delivered fetch results
resultOffset uint64 // Offset of the first cached fetch result in the block chain
resultSize common.StorageSize // Approximate size of a block (exponential moving average)
lock *sync.Mutex
active *sync.Cond
closed bool
2015-04-12 03:38:25 -07:00
}
// newQueue creates a new download queue for scheduling block retrieval.
eth/downloader: separate state sync from queue (#14460) * eth/downloader: separate state sync from queue Scheduling of state node downloads hogged the downloader queue lock when new requests were scheduled. This caused timeouts for other requests. With this change, state sync is fully independent of all other downloads and doesn't involve the queue at all. State sync is started and checked on in processContent. This is slightly awkward because processContent doesn't have a select loop. Instead, the queue is closed by an auxiliary goroutine when state sync fails. We tried several alternatives to this but settled on the current approach because it's the least amount of change overall. Handling of the pivot block has changed slightly: the queue previously prevented import of pivot block receipts before the state of the pivot block was available. In this commit, the receipt will be imported before the state. This causes an annoyance where the pivot block is committed as fast block head even when state downloads fail. Stay tuned for more updates in this area ;) * eth/downloader: remove cancelTimeout channel * eth/downloader: retry state requests on timeout * eth/downloader: improve comment * eth/downloader: mark peers idle when state sync is done * eth/downloader: move pivot block splitting to processContent This change also ensures that pivot block receipts aren't imported before the pivot block itself. * eth/downloader: limit state node retries * eth/downloader: improve state node error handling and retry check * eth/downloader: remove maxStateNodeRetries It fails the sync too much. * eth/downloader: remove last use of cancelCh in statesync.go Fixes TestDeliverHeadersHang*Fast and (hopefully) the weird cancellation behaviour at the end of fast sync. * eth/downloader: fix leak in runStateSync * eth/downloader: don't run processFullSyncContent in LightSync mode * eth/downloader: improve comments * eth/downloader: fix vet, megacheck * eth/downloader: remove unrequested tasks anyway * eth/downloader, trie: various polishes around duplicate items This commit explicitly tracks duplicate and unexpected state delieveries done against a trie Sync structure, also adding there to import info logs. The commit moves the db batch used to commit trie changes one level deeper so its flushed after every node insertion. This is needed to avoid a lot of duplicate retrievals caused by inconsistencies between Sync internals and database. A better approach is to track not-yet-written states in trie.Sync and flush on commit, but I'm focuing on correctness first now. The commit fixes a regression around pivot block fail count. The counter previously was reset to 1 if and only if a sync cycle progressed (inserted at least 1 entry to the database). The current code reset it already if a node was delivered, which is not stong enough, because unless it ends up written to disk, an attacker can just loop and attack ad infinitum. The commit also fixes a regression around state deliveries and timeouts. The old downloader tracked if a delivery is stale (none of the deliveries were requestedt), in which case it didn't mark the node idle and did not send further requests, since it signals a past timeout. The current code did mark it idle even on stale deliveries, which eventually caused two requests to be in flight at the same time, making the deliveries always stale and mass duplicating retrievals between multiple peers. * eth/downloader: fix state request leak This commit fixes the hang seen sometimes while doing the state sync. The cause of the hang was a rare combination of events: request state data from peer, peer drops and reconnects almost immediately. This caused a new download task to be assigned to the peer, overwriting the old one still waiting for a timeout, which in turned leaked the requests out, never to be retried. The fix is to ensure that a task assignment moves any pending one back into the retry queue. The commit also fixes a regression with peer dropping due to stalls. The current code considered a peer stalling if they timed out delivering 1 item. However, the downloader never requests only one, the minimum is 2 (attempt to fine tune estimated latency/bandwidth). The fix is simply to drop if a timeout is detected at 2 items. Apart from the above bugfixes, the commit contains some code polishes I made while debugging the hang. * core, eth, trie: support batched trie sync db writes * trie: rename SyncMemCache to syncMemBatch
2017-06-22 05:26:03 -07:00
func newQueue() *queue {
lock := new(sync.Mutex)
2015-04-12 03:38:25 -07:00
return &queue{
headerPendPool: make(map[string]*fetchRequest),
headerContCh: make(chan bool),
blockTaskPool: make(map[common.Hash]*types.Header),
blockTaskQueue: prque.New(),
blockPendPool: make(map[string]*fetchRequest),
blockDonePool: make(map[common.Hash]struct{}),
receiptTaskPool: make(map[common.Hash]*types.Header),
receiptTaskQueue: prque.New(),
receiptPendPool: make(map[string]*fetchRequest),
receiptDonePool: make(map[common.Hash]struct{}),
resultCache: make([]*fetchResult, blockCacheItems),
active: sync.NewCond(lock),
lock: lock,
2015-04-12 03:38:25 -07:00
}
}
// Reset clears out the queue contents.
func (q *queue) Reset() {
q.lock.Lock()
defer q.lock.Unlock()
q.closed = false
q.mode = FullSync
q.headerHead = common.Hash{}
q.headerPendPool = make(map[string]*fetchRequest)
q.blockTaskPool = make(map[common.Hash]*types.Header)
q.blockTaskQueue.Reset()
q.blockPendPool = make(map[string]*fetchRequest)
q.blockDonePool = make(map[common.Hash]struct{})
q.receiptTaskPool = make(map[common.Hash]*types.Header)
q.receiptTaskQueue.Reset()
q.receiptPendPool = make(map[string]*fetchRequest)
q.receiptDonePool = make(map[common.Hash]struct{})
q.resultCache = make([]*fetchResult, blockCacheItems)
q.resultOffset = 0
}
// Close marks the end of the sync, unblocking WaitResults.
// It may be called even if the queue is already closed.
func (q *queue) Close() {
q.lock.Lock()
q.closed = true
q.lock.Unlock()
q.active.Broadcast()
}
// PendingHeaders retrieves the number of header requests pending for retrieval.
func (q *queue) PendingHeaders() int {
q.lock.Lock()
defer q.lock.Unlock()
return q.headerTaskQueue.Size()
}
// PendingBlocks retrieves the number of block (body) requests pending for retrieval.
func (q *queue) PendingBlocks() int {
q.lock.Lock()
defer q.lock.Unlock()
return q.blockTaskQueue.Size()
}
// PendingReceipts retrieves the number of block receipts pending for retrieval.
func (q *queue) PendingReceipts() int {
q.lock.Lock()
defer q.lock.Unlock()
2015-04-12 03:38:25 -07:00
return q.receiptTaskQueue.Size()
}
2015-04-12 03:38:25 -07:00
// InFlightHeaders retrieves whether there are header fetch requests currently
// in flight.
func (q *queue) InFlightHeaders() bool {
q.lock.Lock()
defer q.lock.Unlock()
return len(q.headerPendPool) > 0
}
// InFlightBlocks retrieves whether there are block fetch requests currently in
// flight.
func (q *queue) InFlightBlocks() bool {
q.lock.Lock()
defer q.lock.Unlock()
return len(q.blockPendPool) > 0
}
// InFlightReceipts retrieves whether there are receipt fetch requests currently
// in flight.
func (q *queue) InFlightReceipts() bool {
q.lock.Lock()
defer q.lock.Unlock()
2015-04-12 03:38:25 -07:00
return len(q.receiptPendPool) > 0
}
eth/downloader: separate state sync from queue (#14460) * eth/downloader: separate state sync from queue Scheduling of state node downloads hogged the downloader queue lock when new requests were scheduled. This caused timeouts for other requests. With this change, state sync is fully independent of all other downloads and doesn't involve the queue at all. State sync is started and checked on in processContent. This is slightly awkward because processContent doesn't have a select loop. Instead, the queue is closed by an auxiliary goroutine when state sync fails. We tried several alternatives to this but settled on the current approach because it's the least amount of change overall. Handling of the pivot block has changed slightly: the queue previously prevented import of pivot block receipts before the state of the pivot block was available. In this commit, the receipt will be imported before the state. This causes an annoyance where the pivot block is committed as fast block head even when state downloads fail. Stay tuned for more updates in this area ;) * eth/downloader: remove cancelTimeout channel * eth/downloader: retry state requests on timeout * eth/downloader: improve comment * eth/downloader: mark peers idle when state sync is done * eth/downloader: move pivot block splitting to processContent This change also ensures that pivot block receipts aren't imported before the pivot block itself. * eth/downloader: limit state node retries * eth/downloader: improve state node error handling and retry check * eth/downloader: remove maxStateNodeRetries It fails the sync too much. * eth/downloader: remove last use of cancelCh in statesync.go Fixes TestDeliverHeadersHang*Fast and (hopefully) the weird cancellation behaviour at the end of fast sync. * eth/downloader: fix leak in runStateSync * eth/downloader: don't run processFullSyncContent in LightSync mode * eth/downloader: improve comments * eth/downloader: fix vet, megacheck * eth/downloader: remove unrequested tasks anyway * eth/downloader, trie: various polishes around duplicate items This commit explicitly tracks duplicate and unexpected state delieveries done against a trie Sync structure, also adding there to import info logs. The commit moves the db batch used to commit trie changes one level deeper so its flushed after every node insertion. This is needed to avoid a lot of duplicate retrievals caused by inconsistencies between Sync internals and database. A better approach is to track not-yet-written states in trie.Sync and flush on commit, but I'm focuing on correctness first now. The commit fixes a regression around pivot block fail count. The counter previously was reset to 1 if and only if a sync cycle progressed (inserted at least 1 entry to the database). The current code reset it already if a node was delivered, which is not stong enough, because unless it ends up written to disk, an attacker can just loop and attack ad infinitum. The commit also fixes a regression around state deliveries and timeouts. The old downloader tracked if a delivery is stale (none of the deliveries were requestedt), in which case it didn't mark the node idle and did not send further requests, since it signals a past timeout. The current code did mark it idle even on stale deliveries, which eventually caused two requests to be in flight at the same time, making the deliveries always stale and mass duplicating retrievals between multiple peers. * eth/downloader: fix state request leak This commit fixes the hang seen sometimes while doing the state sync. The cause of the hang was a rare combination of events: request state data from peer, peer drops and reconnects almost immediately. This caused a new download task to be assigned to the peer, overwriting the old one still waiting for a timeout, which in turned leaked the requests out, never to be retried. The fix is to ensure that a task assignment moves any pending one back into the retry queue. The commit also fixes a regression with peer dropping due to stalls. The current code considered a peer stalling if they timed out delivering 1 item. However, the downloader never requests only one, the minimum is 2 (attempt to fine tune estimated latency/bandwidth). The fix is simply to drop if a timeout is detected at 2 items. Apart from the above bugfixes, the commit contains some code polishes I made while debugging the hang. * core, eth, trie: support batched trie sync db writes * trie: rename SyncMemCache to syncMemBatch
2017-06-22 05:26:03 -07:00
// Idle returns if the queue is fully idle or has some data still inside.
func (q *queue) Idle() bool {
q.lock.Lock()
defer q.lock.Unlock()
eth/downloader: separate state sync from queue (#14460) * eth/downloader: separate state sync from queue Scheduling of state node downloads hogged the downloader queue lock when new requests were scheduled. This caused timeouts for other requests. With this change, state sync is fully independent of all other downloads and doesn't involve the queue at all. State sync is started and checked on in processContent. This is slightly awkward because processContent doesn't have a select loop. Instead, the queue is closed by an auxiliary goroutine when state sync fails. We tried several alternatives to this but settled on the current approach because it's the least amount of change overall. Handling of the pivot block has changed slightly: the queue previously prevented import of pivot block receipts before the state of the pivot block was available. In this commit, the receipt will be imported before the state. This causes an annoyance where the pivot block is committed as fast block head even when state downloads fail. Stay tuned for more updates in this area ;) * eth/downloader: remove cancelTimeout channel * eth/downloader: retry state requests on timeout * eth/downloader: improve comment * eth/downloader: mark peers idle when state sync is done * eth/downloader: move pivot block splitting to processContent This change also ensures that pivot block receipts aren't imported before the pivot block itself. * eth/downloader: limit state node retries * eth/downloader: improve state node error handling and retry check * eth/downloader: remove maxStateNodeRetries It fails the sync too much. * eth/downloader: remove last use of cancelCh in statesync.go Fixes TestDeliverHeadersHang*Fast and (hopefully) the weird cancellation behaviour at the end of fast sync. * eth/downloader: fix leak in runStateSync * eth/downloader: don't run processFullSyncContent in LightSync mode * eth/downloader: improve comments * eth/downloader: fix vet, megacheck * eth/downloader: remove unrequested tasks anyway * eth/downloader, trie: various polishes around duplicate items This commit explicitly tracks duplicate and unexpected state delieveries done against a trie Sync structure, also adding there to import info logs. The commit moves the db batch used to commit trie changes one level deeper so its flushed after every node insertion. This is needed to avoid a lot of duplicate retrievals caused by inconsistencies between Sync internals and database. A better approach is to track not-yet-written states in trie.Sync and flush on commit, but I'm focuing on correctness first now. The commit fixes a regression around pivot block fail count. The counter previously was reset to 1 if and only if a sync cycle progressed (inserted at least 1 entry to the database). The current code reset it already if a node was delivered, which is not stong enough, because unless it ends up written to disk, an attacker can just loop and attack ad infinitum. The commit also fixes a regression around state deliveries and timeouts. The old downloader tracked if a delivery is stale (none of the deliveries were requestedt), in which case it didn't mark the node idle and did not send further requests, since it signals a past timeout. The current code did mark it idle even on stale deliveries, which eventually caused two requests to be in flight at the same time, making the deliveries always stale and mass duplicating retrievals between multiple peers. * eth/downloader: fix state request leak This commit fixes the hang seen sometimes while doing the state sync. The cause of the hang was a rare combination of events: request state data from peer, peer drops and reconnects almost immediately. This caused a new download task to be assigned to the peer, overwriting the old one still waiting for a timeout, which in turned leaked the requests out, never to be retried. The fix is to ensure that a task assignment moves any pending one back into the retry queue. The commit also fixes a regression with peer dropping due to stalls. The current code considered a peer stalling if they timed out delivering 1 item. However, the downloader never requests only one, the minimum is 2 (attempt to fine tune estimated latency/bandwidth). The fix is simply to drop if a timeout is detected at 2 items. Apart from the above bugfixes, the commit contains some code polishes I made while debugging the hang. * core, eth, trie: support batched trie sync db writes * trie: rename SyncMemCache to syncMemBatch
2017-06-22 05:26:03 -07:00
queued := q.blockTaskQueue.Size() + q.receiptTaskQueue.Size()
pending := len(q.blockPendPool) + len(q.receiptPendPool)
cached := len(q.blockDonePool) + len(q.receiptDonePool)
return (queued + pending + cached) == 0
}
// ShouldThrottleBlocks checks if the download should be throttled (active block (body)
// fetches exceed block cache).
func (q *queue) ShouldThrottleBlocks() bool {
q.lock.Lock()
defer q.lock.Unlock()
return q.resultSlots(q.blockPendPool, q.blockDonePool) <= 0
}
// ShouldThrottleReceipts checks if the download should be throttled (active receipt
// fetches exceed block cache).
func (q *queue) ShouldThrottleReceipts() bool {
q.lock.Lock()
defer q.lock.Unlock()
return q.resultSlots(q.receiptPendPool, q.receiptDonePool) <= 0
}
// resultSlots calculates the number of results slots available for requests
// whilst adhering to both the item and the memory limit too of the results
// cache.
func (q *queue) resultSlots(pendPool map[string]*fetchRequest, donePool map[common.Hash]struct{}) int {
// Calculate the maximum length capped by the memory limit
limit := len(q.resultCache)
if common.StorageSize(len(q.resultCache))*q.resultSize > common.StorageSize(blockCacheMemory) {
limit = int((common.StorageSize(blockCacheMemory) + q.resultSize - 1) / q.resultSize)
}
// Calculate the number of slots already finished
finished := 0
for _, result := range q.resultCache[:limit] {
if result == nil {
break
}
if _, ok := donePool[result.Hash]; ok {
finished++
}
}
// Calculate the number of slots currently downloading
pending := 0
for _, request := range pendPool {
for _, header := range request.Headers {
if header.Number.Uint64() < q.resultOffset+uint64(limit) {
pending++
}
}
}
// Return the free slots to distribute
return limit - finished - pending
}
// ScheduleSkeleton adds a batch of header retrieval tasks to the queue to fill
// up an already retrieved header skeleton.
func (q *queue) ScheduleSkeleton(from uint64, skeleton []*types.Header) {
q.lock.Lock()
defer q.lock.Unlock()
// No skeleton retrieval can be in progress, fail hard if so (huge implementation bug)
if q.headerResults != nil {
panic("skeleton assembly already in progress")
}
// Shedule all the header retrieval tasks for the skeleton assembly
q.headerTaskPool = make(map[uint64]*types.Header)
q.headerTaskQueue = prque.New()
q.headerPeerMiss = make(map[string]map[uint64]struct{}) // Reset availability to correct invalid chains
q.headerResults = make([]*types.Header, len(skeleton)*MaxHeaderFetch)
q.headerProced = 0
q.headerOffset = from
q.headerContCh = make(chan bool, 1)
for i, header := range skeleton {
index := from + uint64(i*MaxHeaderFetch)
q.headerTaskPool[index] = header
q.headerTaskQueue.Push(index, -float32(index))
}
}
// RetrieveHeaders retrieves the header chain assemble based on the scheduled
// skeleton.
func (q *queue) RetrieveHeaders() ([]*types.Header, int) {
q.lock.Lock()
defer q.lock.Unlock()
headers, proced := q.headerResults, q.headerProced
q.headerResults, q.headerProced = nil, 0
return headers, proced
}
// Schedule adds a set of headers for the download queue for scheduling, returning
// the new headers encountered.
func (q *queue) Schedule(headers []*types.Header, from uint64) []*types.Header {
q.lock.Lock()
defer q.lock.Unlock()
// Insert all the headers prioritised by the contained block number
inserts := make([]*types.Header, 0, len(headers))
for _, header := range headers {
// Make sure chain order is honoured and preserved throughout
hash := header.Hash()
if header.Number == nil || header.Number.Uint64() != from {
log.Warn("Header broke chain ordering", "number", header.Number, "hash", hash, "expected", from)
break
}
if q.headerHead != (common.Hash{}) && q.headerHead != header.ParentHash {
log.Warn("Header broke chain ancestry", "number", header.Number, "hash", hash)
break
}
// Make sure no duplicate requests are executed
if _, ok := q.blockTaskPool[hash]; ok {
log.Warn("Header already scheduled for block fetch", "number", header.Number, "hash", hash)
continue
}
if _, ok := q.receiptTaskPool[hash]; ok {
log.Warn("Header already scheduled for receipt fetch", "number", header.Number, "hash", hash)
continue
}
// Queue the header for content retrieval
q.blockTaskPool[hash] = header
q.blockTaskQueue.Push(header, -float32(header.Number.Uint64()))
if q.mode == FastSync {
q.receiptTaskPool[hash] = header
q.receiptTaskQueue.Push(header, -float32(header.Number.Uint64()))
}
inserts = append(inserts, header)
q.headerHead = hash
from++
}
return inserts
}
// Results retrieves and permanently removes a batch of fetch results from
// the cache. the result slice will be empty if the queue has been closed.
func (q *queue) Results(block bool) []*fetchResult {
q.lock.Lock()
defer q.lock.Unlock()
// Count the number of items available for processing
nproc := q.countProcessableItems()
for nproc == 0 && !q.closed {
if !block {
return nil
}
q.active.Wait()
nproc = q.countProcessableItems()
}
// Since we have a batch limit, don't pull more into "dangling" memory
if nproc > maxResultsProcess {
nproc = maxResultsProcess
}
results := make([]*fetchResult, nproc)
copy(results, q.resultCache[:nproc])
if len(results) > 0 {
// Mark results as done before dropping them from the cache.
for _, result := range results {
hash := result.Header.Hash()
delete(q.blockDonePool, hash)
delete(q.receiptDonePool, hash)
}
// Delete the results from the cache and clear the tail.
copy(q.resultCache, q.resultCache[nproc:])
for i := len(q.resultCache) - nproc; i < len(q.resultCache); i++ {
q.resultCache[i] = nil
}
// Advance the expected block number of the first cache entry.
q.resultOffset += uint64(nproc)
// Recalculate the result item weights to prevent memory exhaustion
for _, result := range results {
size := result.Header.Size()
for _, uncle := range result.Uncles {
size += uncle.Size()
}
for _, receipt := range result.Receipts {
size += receipt.Size()
}
for _, tx := range result.Transactions {
size += tx.Size()
}
q.resultSize = common.StorageSize(blockCacheSizeWeight)*size + (1-common.StorageSize(blockCacheSizeWeight))*q.resultSize
}
}
return results
}
// countProcessableItems counts the processable items.
func (q *queue) countProcessableItems() int {
for i, result := range q.resultCache {
if result == nil || result.Pending > 0 {
return i
}
}
return len(q.resultCache)
}
// ReserveHeaders reserves a set of headers for the given peer, skipping any
// previously failed batches.
func (q *queue) ReserveHeaders(p *peerConnection, count int) *fetchRequest {
q.lock.Lock()
defer q.lock.Unlock()
// Short circuit if the peer's already downloading something (sanity check to
// not corrupt state)
if _, ok := q.headerPendPool[p.id]; ok {
return nil
}
// Retrieve a batch of hashes, skipping previously failed ones
send, skip := uint64(0), []uint64{}
for send == 0 && !q.headerTaskQueue.Empty() {
from, _ := q.headerTaskQueue.Pop()
if q.headerPeerMiss[p.id] != nil {
if _, ok := q.headerPeerMiss[p.id][from.(uint64)]; ok {
skip = append(skip, from.(uint64))
continue
}
}
send = from.(uint64)
}
// Merge all the skipped batches back
for _, from := range skip {
q.headerTaskQueue.Push(from, -float32(from))
}
// Assemble and return the block download request
if send == 0 {
return nil
}
request := &fetchRequest{
Peer: p,
From: send,
Time: time.Now(),
}
q.headerPendPool[p.id] = request
return request
}
// ReserveBodies reserves a set of body fetches for the given peer, skipping any
// previously failed downloads. Beside the next batch of needed fetches, it also
// returns a flag whether empty blocks were queued requiring processing.
func (q *queue) ReserveBodies(p *peerConnection, count int) (*fetchRequest, bool, error) {
isNoop := func(header *types.Header) bool {
return header.TxHash == types.EmptyRootHash && header.UncleHash == types.EmptyUncleHash
}
q.lock.Lock()
defer q.lock.Unlock()
return q.reserveHeaders(p, count, q.blockTaskPool, q.blockTaskQueue, q.blockPendPool, q.blockDonePool, isNoop)
}
// ReserveReceipts reserves a set of receipt fetches for the given peer, skipping
// any previously failed downloads. Beside the next batch of needed fetches, it
// also returns a flag whether empty receipts were queued requiring importing.
func (q *queue) ReserveReceipts(p *peerConnection, count int) (*fetchRequest, bool, error) {
isNoop := func(header *types.Header) bool {
return header.ReceiptHash == types.EmptyRootHash
}
q.lock.Lock()
defer q.lock.Unlock()
return q.reserveHeaders(p, count, q.receiptTaskPool, q.receiptTaskQueue, q.receiptPendPool, q.receiptDonePool, isNoop)
}
// reserveHeaders reserves a set of data download operations for a given peer,
// skipping any previously failed ones. This method is a generic version used
// by the individual special reservation functions.
//
// Note, this method expects the queue lock to be already held for writing. The
// reason the lock is not obtained in here is because the parameters already need
// to access the queue, so they already need a lock anyway.
func (q *queue) reserveHeaders(p *peerConnection, count int, taskPool map[common.Hash]*types.Header, taskQueue *prque.Prque,
pendPool map[string]*fetchRequest, donePool map[common.Hash]struct{}, isNoop func(*types.Header) bool) (*fetchRequest, bool, error) {
// Short circuit if the pool has been depleted, or if the peer's already
// downloading something (sanity check not to corrupt state)
if taskQueue.Empty() {
return nil, false, nil
}
if _, ok := pendPool[p.id]; ok {
return nil, false, nil
}
// Calculate an upper limit on the items we might fetch (i.e. throttling)
space := q.resultSlots(pendPool, donePool)
// Retrieve a batch of tasks, skipping previously failed ones
send := make([]*types.Header, 0, count)
skip := make([]*types.Header, 0)
progress := false
for proc := 0; proc < space && len(send) < count && !taskQueue.Empty(); proc++ {
header := taskQueue.PopItem().(*types.Header)
hash := header.Hash()
// If we're the first to request this task, initialise the result container
index := int(header.Number.Int64() - int64(q.resultOffset))
if index >= len(q.resultCache) || index < 0 {
common.Report("index allocation went beyond available resultCache space")
return nil, false, errInvalidChain
}
if q.resultCache[index] == nil {
components := 1
if q.mode == FastSync {
components = 2
}
q.resultCache[index] = &fetchResult{
Pending: components,
Hash: hash,
Header: header,
}
}
// If this fetch task is a noop, skip this fetch operation
if isNoop(header) {
donePool[hash] = struct{}{}
delete(taskPool, hash)
space, proc = space-1, proc-1
q.resultCache[index].Pending--
progress = true
continue
}
// Otherwise unless the peer is known not to have the data, add to the retrieve list
if p.Lacks(hash) {
skip = append(skip, header)
} else {
send = append(send, header)
}
}
// Merge all the skipped headers back
for _, header := range skip {
taskQueue.Push(header, -float32(header.Number.Uint64()))
}
if progress {
// Wake WaitResults, resultCache was modified
q.active.Signal()
}
// Assemble and return the block download request
if len(send) == 0 {
return nil, progress, nil
}
request := &fetchRequest{
Peer: p,
Headers: send,
Time: time.Now(),
}
pendPool[p.id] = request
return request, progress, nil
}
// CancelHeaders aborts a fetch request, returning all pending skeleton indexes to the queue.
func (q *queue) CancelHeaders(request *fetchRequest) {
q.cancel(request, q.headerTaskQueue, q.headerPendPool)
}
// CancelBodies aborts a body fetch request, returning all pending headers to the
// task queue.
func (q *queue) CancelBodies(request *fetchRequest) {
q.cancel(request, q.blockTaskQueue, q.blockPendPool)
}
// CancelReceipts aborts a body fetch request, returning all pending headers to
// the task queue.
func (q *queue) CancelReceipts(request *fetchRequest) {
q.cancel(request, q.receiptTaskQueue, q.receiptPendPool)
}
// Cancel aborts a fetch request, returning all pending hashes to the task queue.
func (q *queue) cancel(request *fetchRequest, taskQueue *prque.Prque, pendPool map[string]*fetchRequest) {
q.lock.Lock()
defer q.lock.Unlock()
2015-04-12 03:38:25 -07:00
if request.From > 0 {
taskQueue.Push(request.From, -float32(request.From))
}
for _, header := range request.Headers {
taskQueue.Push(header, -float32(header.Number.Uint64()))
}
delete(pendPool, request.Peer.id)
2015-04-12 03:38:25 -07:00
}
// Revoke cancels all pending requests belonging to a given peer. This method is
// meant to be called during a peer drop to quickly reassign owned data fetches
// to remaining nodes.
func (q *queue) Revoke(peerId string) {
q.lock.Lock()
defer q.lock.Unlock()
if request, ok := q.blockPendPool[peerId]; ok {
for _, header := range request.Headers {
q.blockTaskQueue.Push(header, -float32(header.Number.Uint64()))
}
delete(q.blockPendPool, peerId)
}
if request, ok := q.receiptPendPool[peerId]; ok {
for _, header := range request.Headers {
q.receiptTaskQueue.Push(header, -float32(header.Number.Uint64()))
}
delete(q.receiptPendPool, peerId)
}
}
// ExpireHeaders checks for in flight requests that exceeded a timeout allowance,
// canceling them and returning the responsible peers for penalisation.
func (q *queue) ExpireHeaders(timeout time.Duration) map[string]int {
q.lock.Lock()
defer q.lock.Unlock()
return q.expire(timeout, q.headerPendPool, q.headerTaskQueue, headerTimeoutMeter)
}
// ExpireBodies checks for in flight block body requests that exceeded a timeout
// allowance, canceling them and returning the responsible peers for penalisation.
func (q *queue) ExpireBodies(timeout time.Duration) map[string]int {
q.lock.Lock()
defer q.lock.Unlock()
return q.expire(timeout, q.blockPendPool, q.blockTaskQueue, bodyTimeoutMeter)
}
// ExpireReceipts checks for in flight receipt requests that exceeded a timeout
// allowance, canceling them and returning the responsible peers for penalisation.
func (q *queue) ExpireReceipts(timeout time.Duration) map[string]int {
q.lock.Lock()
defer q.lock.Unlock()
return q.expire(timeout, q.receiptPendPool, q.receiptTaskQueue, receiptTimeoutMeter)
}
// expire is the generic check that move expired tasks from a pending pool back
// into a task pool, returning all entities caught with expired tasks.
//
// Note, this method expects the queue lock to be already held. The
// reason the lock is not obtained in here is because the parameters already need
// to access the queue, so they already need a lock anyway.
func (q *queue) expire(timeout time.Duration, pendPool map[string]*fetchRequest, taskQueue *prque.Prque, timeoutMeter metrics.Meter) map[string]int {
// Iterate over the expired requests and return each to the queue
expiries := make(map[string]int)
for id, request := range pendPool {
if time.Since(request.Time) > timeout {
// Update the metrics with the timeout
timeoutMeter.Mark(1)
// Return any non satisfied requests to the pool
if request.From > 0 {
taskQueue.Push(request.From, -float32(request.From))
}
for _, header := range request.Headers {
taskQueue.Push(header, -float32(header.Number.Uint64()))
}
// Add the peer to the expiry report along the the number of failed requests
expiries[id] = len(request.Headers)
}
}
// Remove the expired requests from the pending pool
2017-01-06 06:52:03 -08:00
for id := range expiries {
delete(pendPool, id)
}
return expiries
2015-04-12 03:38:25 -07:00
}
// DeliverHeaders injects a header retrieval response into the header results
// cache. This method either accepts all headers it received, or none of them
// if they do not map correctly to the skeleton.
//
// If the headers are accepted, the method makes an attempt to deliver the set
// of ready headers to the processor to keep the pipeline full. However it will
// not block to prevent stalling other pending deliveries.
func (q *queue) DeliverHeaders(id string, headers []*types.Header, headerProcCh chan []*types.Header) (int, error) {
q.lock.Lock()
defer q.lock.Unlock()
// Short circuit if the data was never requested
request := q.headerPendPool[id]
if request == nil {
return 0, errNoFetchesPending
}
headerReqTimer.UpdateSince(request.Time)
delete(q.headerPendPool, id)
// Ensure headers can be mapped onto the skeleton chain
target := q.headerTaskPool[request.From].Hash()
accepted := len(headers) == MaxHeaderFetch
if accepted {
if headers[0].Number.Uint64() != request.From {
log.Trace("First header broke chain ordering", "peer", id, "number", headers[0].Number, "hash", headers[0].Hash(), request.From)
accepted = false
} else if headers[len(headers)-1].Hash() != target {
log.Trace("Last header broke skeleton structure ", "peer", id, "number", headers[len(headers)-1].Number, "hash", headers[len(headers)-1].Hash(), "expected", target)
accepted = false
}
}
if accepted {
for i, header := range headers[1:] {
hash := header.Hash()
if want := request.From + 1 + uint64(i); header.Number.Uint64() != want {
log.Warn("Header broke chain ordering", "peer", id, "number", header.Number, "hash", hash, "expected", want)
accepted = false
break
}
if headers[i].Hash() != header.ParentHash {
log.Warn("Header broke chain ancestry", "peer", id, "number", header.Number, "hash", hash)
accepted = false
break
}
}
}
// If the batch of headers wasn't accepted, mark as unavailable
if !accepted {
log.Trace("Skeleton filling not accepted", "peer", id, "from", request.From)
miss := q.headerPeerMiss[id]
if miss == nil {
q.headerPeerMiss[id] = make(map[uint64]struct{})
miss = q.headerPeerMiss[id]
}
miss[request.From] = struct{}{}
q.headerTaskQueue.Push(request.From, -float32(request.From))
return 0, errors.New("delivery not accepted")
}
// Clean up a successful fetch and try to deliver any sub-results
copy(q.headerResults[request.From-q.headerOffset:], headers)
delete(q.headerTaskPool, request.From)
ready := 0
for q.headerProced+ready < len(q.headerResults) && q.headerResults[q.headerProced+ready] != nil {
ready += MaxHeaderFetch
}
if ready > 0 {
// Headers are ready for delivery, gather them and push forward (non blocking)
process := make([]*types.Header, ready)
copy(process, q.headerResults[q.headerProced:q.headerProced+ready])
select {
case headerProcCh <- process:
log.Trace("Pre-scheduled new headers", "peer", id, "count", len(process), "from", process[0].Number)
q.headerProced += len(process)
default:
}
}
// Check for termination and return
if len(q.headerTaskPool) == 0 {
q.headerContCh <- false
}
return len(headers), nil
}
// DeliverBodies injects a block body retrieval response into the results queue.
// The method returns the number of blocks bodies accepted from the delivery and
// also wakes any threads waiting for data delivery.
func (q *queue) DeliverBodies(id string, txLists [][]*types.Transaction, uncleLists [][]*types.Header) (int, error) {
q.lock.Lock()
defer q.lock.Unlock()
reconstruct := func(header *types.Header, index int, result *fetchResult) error {
if types.DeriveSha(types.Transactions(txLists[index])) != header.TxHash || types.CalcUncleHash(uncleLists[index]) != header.UncleHash {
return errInvalidBody
}
result.Transactions = txLists[index]
result.Uncles = uncleLists[index]
return nil
}
return q.deliver(id, q.blockTaskPool, q.blockTaskQueue, q.blockPendPool, q.blockDonePool, bodyReqTimer, len(txLists), reconstruct)
}
// DeliverReceipts injects a receipt retrieval response into the results queue.
// The method returns the number of transaction receipts accepted from the delivery
// and also wakes any threads waiting for data delivery.
func (q *queue) DeliverReceipts(id string, receiptList [][]*types.Receipt) (int, error) {
q.lock.Lock()
defer q.lock.Unlock()
reconstruct := func(header *types.Header, index int, result *fetchResult) error {
if types.DeriveSha(types.Receipts(receiptList[index])) != header.ReceiptHash {
return errInvalidReceipt
}
result.Receipts = receiptList[index]
return nil
}
return q.deliver(id, q.receiptTaskPool, q.receiptTaskQueue, q.receiptPendPool, q.receiptDonePool, receiptReqTimer, len(receiptList), reconstruct)
}
// deliver injects a data retrieval response into the results queue.
//
// Note, this method expects the queue lock to be already held for writing. The
// reason the lock is not obtained in here is because the parameters already need
// to access the queue, so they already need a lock anyway.
func (q *queue) deliver(id string, taskPool map[common.Hash]*types.Header, taskQueue *prque.Prque,
pendPool map[string]*fetchRequest, donePool map[common.Hash]struct{}, reqTimer metrics.Timer,
results int, reconstruct func(header *types.Header, index int, result *fetchResult) error) (int, error) {
// Short circuit if the data was never requested
request := pendPool[id]
if request == nil {
return 0, errNoFetchesPending
}
reqTimer.UpdateSince(request.Time)
delete(pendPool, id)
// If no data items were retrieved, mark them as unavailable for the origin peer
if results == 0 {
for _, header := range request.Headers {
request.Peer.MarkLacking(header.Hash())
}
}
// Assemble each of the results with their headers and retrieved data parts
var (
accepted int
failure error
useful bool
)
for i, header := range request.Headers {
// Short circuit assembly if no more fetch results are found
if i >= results {
break
}
// Reconstruct the next result if contents match up
index := int(header.Number.Int64() - int64(q.resultOffset))
if index >= len(q.resultCache) || index < 0 || q.resultCache[index] == nil {
failure = errInvalidChain
break
}
if err := reconstruct(header, i, q.resultCache[index]); err != nil {
failure = err
break
}
hash := header.Hash()
donePool[hash] = struct{}{}
q.resultCache[index].Pending--
useful = true
accepted++
// Clean up a successful fetch
request.Headers[i] = nil
delete(taskPool, hash)
}
// Return all failed or missing fetches to the queue
for _, header := range request.Headers {
if header != nil {
taskQueue.Push(header, -float32(header.Number.Uint64()))
}
}
// Wake up WaitResults
if accepted > 0 {
q.active.Signal()
}
// If none of the data was good, it's a stale delivery
switch {
case failure == nil || failure == errInvalidChain:
return accepted, failure
case useful:
return accepted, fmt.Errorf("partial failure: %v", failure)
default:
return accepted, errStaleDelivery
}
}
// Prepare configures the result cache to allow accepting and caching inbound
// fetch results.
func (q *queue) Prepare(offset uint64, mode SyncMode) {
q.lock.Lock()
defer q.lock.Unlock()
// Prepare the queue for sync results
if q.resultOffset < offset {
q.resultOffset = offset
}
q.mode = mode
2015-04-12 03:38:25 -07:00
}