eth/downloader: refactor downloader + queue ethereum#21263 (XinFinOrg#1041)

* eth/downloader: refactor downloader + queue

downloader, fetcher: throttle-metrics, fetcher filter improvements, standalone resultcache

downloader: more accurate deliverytime calculation, less mem overhead in state requests

downloader/queue: increase underlying buffer of results, new throttle mechanism

eth/downloader: updates to tests

eth/downloader: fix up some review concerns

eth/downloader/queue: minor fixes

eth/downloader: minor fixes after review call

eth/downloader: testcases for queue.go

eth/downloader: minor change, don't set progress unless progress...

eth/downloader: fix flaw which prevented useless peers from being dropped

eth/downloader: try to fix tests

eth/downloader: verify non-deliveries against advertised remote head

eth/downloader: fix flaw with checking closed-status causing hang

eth/downloader: hashing avoidance

eth/downloader: review concerns + simplify resultcache and queue

eth/downloader: add back some locks, address review concerns

downloader/queue: fix remaining lock flaw

* eth/downloader: nitpick fixes

* eth/downloader: remove the *2*3/4 throttling threshold dance

* eth/downloader: print correct throttle threshold in stats

Co-authored-by: Martin Holst Swende <[email protected]>
Co-authored-by: Péter Szilágyi <[email protected]>

Author: 3 people

core/types/block.go

@@ -153,6 +153,17 @@ func (h *Header) Size() common.StorageSize {
 	return common.StorageSize(unsafe.Sizeof(*h)) + common.StorageSize(len(h.Extra)+(h.Difficulty.BitLen()+h.Number.BitLen()+h.Time.BitLen())/8)
 }
 
+// EmptyBody returns true if there is no additional 'body' to complete the header
+// that is: no transactions and no uncles.
+func (h *Header) EmptyBody() bool {
+	return h.TxHash == EmptyRootHash && h.UncleHash == EmptyUncleHash
+}
+
+// EmptyReceipts returns true if there are no receipts for this header/block.
+func (h *Header) EmptyReceipts() bool {
+	return h.ReceiptHash == EmptyRootHash
+}
+
 // Body is a simple (mutable, non-safe) data container for storing and moving
 // a block's data contents (transactions and uncles) together.
 type Body struct {

eth/downloader/downloader.go

@@ -213,7 +213,7 @@ func New(stateDb ethdb.Database, mux *event.TypeMux, chain BlockChain, lightchai
 	dl := &Downloader{
 		stateDB:             stateDb,
 		mux:                 mux,
-		queue:               newQueue(),
+		queue:               newQueue(blockCacheItems),
 		peers:               newPeerSet(),
 		rttEstimate:         uint64(rttMaxEstimate),
 		rttConfidence:       uint64(1000000),
@@ -359,7 +359,7 @@ func (d *Downloader) synchronise(id string, hash common.Hash, td *big.Int, mode
 		log.Info("Block synchronisation started")
 	}
 	// Reset the queue, peer set and wake channels to clean any internal leftover state
-	d.queue.Reset()
+	d.queue.Reset(blockCacheItems)
 	d.peers.Reset()
 
 	for _, ch := range []chan bool{d.bodyWakeCh, d.receiptWakeCh} {
@@ -575,7 +575,7 @@ func (d *Downloader) fetchHeight(p *peerConnection, hash common.Hash) (*types.He
 			// Make sure the peer actually gave something valid
 			headers := packet.(*headerPack).headers
 			if len(headers) != 1 {
-				p.log.Debug("Multiple headers for single request", "headers", len(headers))
+				p.log.Warn("Multiple headers for single request", "headers", len(headers))
 				return nil, fmt.Errorf("%w: multiple headers (%d) for single request", errBadPeer, len(headers))
 			}
 			head := headers[0]
@@ -784,7 +784,7 @@ func (d *Downloader) findAncestor(p *peerConnection, remoteHeader *types.Header)
 				// Make sure the peer actually gave something valid
 				headers := packer.(*headerPack).headers
 				if len(headers) != 1 {
-					p.log.Debug("Multiple headers for single request", "headers", len(headers))
+					p.log.Warn("Multiple headers for single request", "headers", len(headers))
 					return 0, fmt.Errorf("%w: multiple headers (%d) for single request", errBadPeer, len(headers))
 				}
 				arrived = true
@@ -808,7 +808,7 @@ func (d *Downloader) findAncestor(p *peerConnection, remoteHeader *types.Header)
 				}
 				header := d.lightchain.GetHeaderByHash(h) // Independent of sync mode, header surely exists
 				if header.Number.Uint64() != check {
-					p.log.Debug("Received non requested header", "number", header.Number, "hash", header.Hash(), "request", check)
+					p.log.Warn("Received non requested header", "number", header.Number, "hash", header.Hash(), "request", check)
 					return 0, fmt.Errorf("%w: non-requested header (%d)", errBadPeer, header.Number)
 				}
 				start = check
@@ -1017,17 +1017,18 @@ func (d *Downloader) fillHeaderSkeleton(from uint64, skeleton []*types.Header) (
 			pack := packet.(*headerPack)
 			return d.queue.DeliverHeaders(pack.peerId, pack.headers, d.headerProcCh)
 		}
-		expire   = func() map[string]int { return d.queue.ExpireHeaders(d.requestTTL()) }
-		throttle = func() bool { return false }
-		reserve  = func(p *peerConnection, count int) (*fetchRequest, bool, error) {
-			return d.queue.ReserveHeaders(p, count), false, nil
+		expire  = func() map[string]int { return d.queue.ExpireHeaders(d.requestTTL()) }
+		reserve = func(p *peerConnection, count int) (*fetchRequest, bool, bool) {
+			return d.queue.ReserveHeaders(p, count), false, false
 		}
 		fetch    = func(p *peerConnection, req *fetchRequest) error { return p.FetchHeaders(req.From, MaxHeaderFetch) }
 		capacity = func(p *peerConnection) int { return p.HeaderCapacity(d.requestRTT()) }
-		setIdle  = func(p *peerConnection, accepted int) { p.SetHeadersIdle(accepted) }
+		setIdle  = func(p *peerConnection, accepted int, deliveryTime time.Time) {
+			p.SetHeadersIdle(accepted, deliveryTime)
+		}
 	)
 	err := d.fetchParts(d.headerCh, deliver, d.queue.headerContCh, expire,
-		d.queue.PendingHeaders, d.queue.InFlightHeaders, throttle, reserve,
+		d.queue.PendingHeaders, d.queue.InFlightHeaders, reserve,
 		nil, fetch, d.queue.CancelHeaders, capacity, d.peers.HeaderIdlePeers, setIdle, "headers")
 
 	log.Debug("Skeleton fill terminated", "err", err)
@@ -1050,10 +1051,10 @@ func (d *Downloader) fetchBodies(from uint64) error {
 		expire   = func() map[string]int { return d.queue.ExpireBodies(d.requestTTL()) }
 		fetch    = func(p *peerConnection, req *fetchRequest) error { return p.FetchBodies(req) }
 		capacity = func(p *peerConnection) int { return p.BlockCapacity(d.requestRTT()) }
-		setIdle  = func(p *peerConnection, accepted int) { p.SetBodiesIdle(accepted) }
+		setIdle  = func(p *peerConnection, accepted int, deliveryTime time.Time) { p.SetBodiesIdle(accepted, deliveryTime) }
 	)
 	err := d.fetchParts(d.bodyCh, deliver, d.bodyWakeCh, expire,
-		d.queue.PendingBlocks, d.queue.InFlightBlocks, d.queue.ShouldThrottleBlocks, d.queue.ReserveBodies,
+		d.queue.PendingBlocks, d.queue.InFlightBlocks, d.queue.ReserveBodies,
 		d.bodyFetchHook, fetch, d.queue.CancelBodies, capacity, d.peers.BodyIdlePeers, setIdle, "bodies")
 
 	log.Debug("Block body download terminated", "err", err)
@@ -1074,10 +1075,12 @@ func (d *Downloader) fetchReceipts(from uint64) error {
 		expire   = func() map[string]int { return d.queue.ExpireReceipts(d.requestTTL()) }
 		fetch    = func(p *peerConnection, req *fetchRequest) error { return p.FetchReceipts(req) }
 		capacity = func(p *peerConnection) int { return p.ReceiptCapacity(d.requestRTT()) }
-		setIdle  = func(p *peerConnection, accepted int) { p.SetReceiptsIdle(accepted) }
+		setIdle  = func(p *peerConnection, accepted int, deliveryTime time.Time) {
+			p.SetReceiptsIdle(accepted, deliveryTime)
+		}
 	)
 	err := d.fetchParts(d.receiptCh, deliver, d.receiptWakeCh, expire,
-		d.queue.PendingReceipts, d.queue.InFlightReceipts, d.queue.ShouldThrottleReceipts, d.queue.ReserveReceipts,
+		d.queue.PendingReceipts, d.queue.InFlightReceipts, d.queue.ReserveReceipts,
 		d.receiptFetchHook, fetch, d.queue.CancelReceipts, capacity, d.peers.ReceiptIdlePeers, setIdle, "receipts")
 
 	log.Debug("Transaction receipt download terminated", "err", err)
@@ -1110,9 +1113,9 @@ func (d *Downloader) fetchReceipts(from uint64) error {
 //   - setIdle:     network callback to set a peer back to idle and update its estimated capacity (traffic shaping)
 //   - kind:        textual label of the type being downloaded to display in log mesages
 func (d *Downloader) fetchParts(deliveryCh chan dataPack, deliver func(dataPack) (int, error), wakeCh chan bool,
-	expire func() map[string]int, pending func() int, inFlight func() bool, throttle func() bool, reserve func(*peerConnection, int) (*fetchRequest, bool, error),
+	expire func() map[string]int, pending func() int, inFlight func() bool, reserve func(*peerConnection, int) (*fetchRequest, bool, bool),
 	fetchHook func([]*types.Header), fetch func(*peerConnection, *fetchRequest) error, cancel func(*fetchRequest), capacity func(*peerConnection) int,
-	idle func() ([]*peerConnection, int), setIdle func(*peerConnection, int), kind string) error {
+	idle func() ([]*peerConnection, int), setIdle func(*peerConnection, int, time.Time), kind string) error {
 
 	// Create a ticker to detect expired retrieval tasks
 	ticker := time.NewTicker(100 * time.Millisecond)
@@ -1128,6 +1131,7 @@ func (d *Downloader) fetchParts(deliveryCh chan dataPack, deliver func(dataPack)
 			return errCanceled
 
 		case packet := <-deliveryCh:
+			deliveryTime := time.Now()
 			// If the peer was previously banned and failed to deliver its pack
 			// in a reasonable time frame, ignore its message.
 			if peer := d.peers.Peer(packet.PeerId()); peer != nil {
@@ -1140,7 +1144,7 @@ func (d *Downloader) fetchParts(deliveryCh chan dataPack, deliver func(dataPack)
 				// caused by a timed out request which came through in the end), set it to
 				// idle. If the delivery's stale, the peer should have already been idled.
 				if !errors.Is(err, errStaleDelivery) {
-					setIdle(peer, accepted)
+					setIdle(peer, accepted, deliveryTime)
 				}
 				// Issue a log to the user to see what's going on
 				switch {
@@ -1193,7 +1197,7 @@ func (d *Downloader) fetchParts(deliveryCh chan dataPack, deliver func(dataPack)
 					// how response times reacts, to it always requests one more than the minimum (i.e. min 2).
 					if fails > 2 {
 						peer.log.Trace("Data delivery timed out", "type", kind)
-						setIdle(peer, 0)
+						setIdle(peer, 0, time.Now())
 					} else {
 						peer.log.Debug("Stalling delivery, dropping", "type", kind)
 
@@ -1228,27 +1232,27 @@ func (d *Downloader) fetchParts(deliveryCh chan dataPack, deliver func(dataPack)
 			// Send a download request to all idle peers, until throttled
 			progressed, throttled, running := false, false, inFlight()
 			idles, total := idle()
-
+			pendCount := pending()
 			for _, peer := range idles {
 				// Short circuit if throttling activated
-				if throttle() {
-					throttled = true
+				if throttled {
 					break
 				}
 				// Short circuit if there is no more available task.
-				if pending() == 0 {
+				if pendCount = pending(); pendCount == 0 {
 					break
 				}
 				// Reserve a chunk of fetches for a peer. A nil can mean either that
 				// no more headers are available, or that the peer is known not to
 				// have them.
-				request, progress, err := reserve(peer, capacity(peer))
-				if err != nil {
-					return err
-				}
+				request, progress, throttle := reserve(peer, capacity(peer))
 				if progress {
 					progressed = true
 				}
+				if throttle {
+					throttled = true
+					throttleCounter.Inc(1)
+				}
 				if request == nil {
 					continue
 				}
@@ -1273,7 +1277,7 @@ func (d *Downloader) fetchParts(deliveryCh chan dataPack, deliver func(dataPack)
 			}
 			// Make sure that we have peers available for fetching. If all peers have been tried
 			// and all failed throw an error
-			if !progressed && !throttled && !running && len(idles) == total && pending() > 0 {
+			if !progressed && !throttled && !running && len(idles) == total && pendCount > 0 {
 				return errPeersUnavailable
 			}
 		}
@@ -1285,8 +1289,11 @@ func (d *Downloader) fetchParts(deliveryCh chan dataPack, deliver func(dataPack)
 // queue until the stream ends or a failure occurs.
 func (d *Downloader) processHeaders(origin uint64, pivot uint64, td *big.Int) error {
 	// Keep a count of uncertain headers to roll back
-	rollback := []*types.Header{}
-	mode := d.getMode()
+	var (
+		rollback    []*types.Header
+		rollbackErr error
+		mode        = d.getMode()
+	)
 	defer func() {
 		if len(rollback) > 0 {
 			// Flatten the headers and roll them back
@@ -1308,7 +1315,7 @@ func (d *Downloader) processHeaders(origin uint64, pivot uint64, td *big.Int) er
 			log.Warn("Rolled back headers", "count", len(hashes),
 				"header", fmt.Sprintf("%d->%d", lastHeader, d.lightchain.CurrentHeader().Number),
 				"fast", fmt.Sprintf("%d->%d", lastFastBlock, curFastBlock),
-				"block", fmt.Sprintf("%d->%d", lastBlock, curBlock))
+				"block", fmt.Sprintf("%d->%d", lastBlock, curBlock), "reason", rollbackErr)
 		}
 	}()
 
@@ -1318,6 +1325,7 @@ func (d *Downloader) processHeaders(origin uint64, pivot uint64, td *big.Int) er
 	for {
 		select {
 		case <-d.cancelCh:
+			rollbackErr = errCanceled
 			return errCanceled
 
 		case headers := <-d.headerProcCh:
@@ -1371,6 +1379,7 @@ func (d *Downloader) processHeaders(origin uint64, pivot uint64, td *big.Int) er
 				// Terminate if something failed in between processing chunks
 				select {
 				case <-d.cancelCh:
+					rollbackErr = errCanceled
 					return errCanceled
 				default:
 				}
@@ -1395,11 +1404,12 @@ func (d *Downloader) processHeaders(origin uint64, pivot uint64, td *big.Int) er
 						frequency = 1
 					}
 					if n, err := d.lightchain.InsertHeaderChain(chunk, frequency); err != nil {
+						rollbackErr = err
 						// If some headers were inserted, add them too to the rollback list
 						if n > 0 {
 							rollback = append(rollback, chunk[:n]...)
 						}
-						log.Debug("Invalid header encountered", "number", chunk[n].Number, "hash", chunk[n].Hash(), "err", err)
+						log.Debug("Invalid header encountered", "number", chunk[n].Number, "hash", chunk[n].Hash(), "parent", chunk[n].ParentHash, "err", err)
 						return fmt.Errorf("%w: %v", errInvalidChain, err)
 					}
 					// All verifications passed, store newly found uncertain headers
@@ -1414,14 +1424,15 @@ func (d *Downloader) processHeaders(origin uint64, pivot uint64, td *big.Int) er
 					for d.queue.PendingBlocks() >= maxQueuedHeaders || d.queue.PendingReceipts() >= maxQueuedHeaders {
 						select {
 						case <-d.cancelCh:
+							rollbackErr = errCanceled
 							return errCanceled
 						case <-time.After(time.Second):
 						}
 					}
 					// Otherwise insert the headers for content retrieval
 					inserts := d.queue.Schedule(chunk, origin)
 					if len(inserts) != len(chunk) {
-						log.Debug("Stale headers")
+						rollbackErr = fmt.Errorf("stale headers: len inserts %v len(chunk) %v", len(inserts), len(chunk))
 						return fmt.Errorf("%w: stale headers", errBadPeer)
 					}
 				}
@@ -1638,6 +1649,14 @@ func (d *Downloader) processFastSyncContent(latest *types.Header) error {
 }
 
 func splitAroundPivot(pivot uint64, results []*fetchResult) (p *fetchResult, before, after []*fetchResult) {
+	if len(results) == 0 {
+		return nil, nil, nil
+	}
+	if lastNum := results[len(results)-1].Header.Number.Uint64(); lastNum < pivot {
+		// the pivot is somewhere in the future
+		return nil, results, nil
+	}
+	// This can also be optimized, but only happens very seldom
 	for _, result := range results {
 		num := result.Header.Number.Uint64()
 		switch {