triedb/pathdb: introduce lookup structure to optimize node query

core/blockchain.go

@@ -77,6 +77,7 @@ var (
 
 	snapshotCommitTimer = metrics.NewRegisteredResettingTimer("chain/snapshot/commits", nil)
 	triedbCommitTimer   = metrics.NewRegisteredResettingTimer("chain/triedb/commits", nil)
+	prefetchWaitTimer   = metrics.NewRegisteredResettingTimer("chain/prefetch/wait", nil)
 
 	blockInsertTimer          = metrics.NewRegisteredResettingTimer("chain/inserts", nil)
 	blockValidationTimer      = metrics.NewRegisteredResettingTimer("chain/validation", nil)
@@ -1952,12 +1953,13 @@ func (bc *BlockChain) processBlock(block *types.Block, statedb *state.StateDB, s
 	if statedb.StorageLoaded != 0 {
 		storageReadSingleTimer.Update(statedb.StorageReads / time.Duration(statedb.StorageLoaded))
 	}
-	accountUpdateTimer.Update(statedb.AccountUpdates)                                 // Account updates are complete(in validation)
+	accountUpdateTimer.Update(statedb.AccountUpdates - statedb.PrefetcherWait)        // Account updates are complete(in validation)
 	storageUpdateTimer.Update(statedb.StorageUpdates)                                 // Storage updates are complete(in validation)
 	accountHashTimer.Update(statedb.AccountHashes)                                    // Account hashes are complete(in validation)
 	triehash := statedb.AccountHashes                                                 // The time spent on tries hashing
 	trieUpdate := statedb.AccountUpdates + statedb.StorageUpdates                     // The time spent on tries update
 	blockExecutionTimer.Update(ptime - (statedb.AccountReads + statedb.StorageReads)) // The time spent on EVM processing
+	prefetchWaitTimer.Update(statedb.PrefetcherWait)                                  // The time spent on waiting prefetcher to finish preload tasks
 	blockValidationTimer.Update(vtime - (triehash + trieUpdate))                      // The time spent on block validation
 	blockCrossValidationTimer.Update(xvtime)                                          // The time spent on stateless cross validation
 

core/state/statedb.go

@@ -142,13 +142,15 @@ type StateDB struct {
 	witness *stateless.Witness
 
 	// Measurements gathered during execution for debugging purposes
-	AccountReads    time.Duration
-	AccountHashes   time.Duration
-	AccountUpdates  time.Duration
-	AccountCommits  time.Duration
-	StorageReads    time.Duration
-	StorageUpdates  time.Duration
-	StorageCommits  time.Duration
+	AccountReads   time.Duration
+	AccountHashes  time.Duration
+	AccountUpdates time.Duration
+	AccountCommits time.Duration
+	StorageReads   time.Duration
+	StorageUpdates time.Duration
+	StorageCommits time.Duration
+
+	PrefetcherWait  time.Duration
 	SnapshotCommits time.Duration
 	TrieDBCommits   time.Duration
 
@@ -866,6 +868,7 @@ func (s *StateDB) IntermediateRoot(deleteEmptyObjects bool) common.Hash {
 		} else {
 			s.trie = trie
 		}
+		s.PrefetcherWait = time.Since(start)
 	}
 	// Perform updates before deletions.  This prevents resolution of unnecessary trie nodes
 	// in circumstances similar to the following:

core/state/trie_prefetcher.go

@@ -19,6 +19,7 @@ package state
 import (
 	"errors"
 	"sync"
+	"time"
 
 	"github.com/ethereum/go-ethereum/common"
 	"github.com/ethereum/go-ethereum/log"
@@ -55,13 +56,15 @@ type triePrefetcher struct {
 	accountDupWriteMeter  metrics.Meter
 	accountDupCrossMeter  metrics.Meter
 	accountWasteMeter     metrics.Meter
+	accountLoadTimer      metrics.ResettingTimer
 
 	storageLoadReadMeter  metrics.Meter
 	storageLoadWriteMeter metrics.Meter
 	storageDupReadMeter   metrics.Meter
 	storageDupWriteMeter  metrics.Meter
 	storageDupCrossMeter  metrics.Meter
 	storageWasteMeter     metrics.Meter
+	storageLoadTimer      metrics.ResettingTimer
 }
 
 func newTriePrefetcher(db Database, root common.Hash, namespace string, noreads bool) *triePrefetcher {
@@ -78,13 +81,15 @@ func newTriePrefetcher(db Database, root common.Hash, namespace string, noreads
 
 		accountLoadReadMeter:  metrics.GetOrRegisterMeter(prefix+"/account/load/read", nil),
 		accountLoadWriteMeter: metrics.GetOrRegisterMeter(prefix+"/account/load/write", nil),
+		accountLoadTimer:      metrics.GetOrRegisterResettingTimer(prefix+"/account/load/time", nil),
 		accountDupReadMeter:   metrics.GetOrRegisterMeter(prefix+"/account/dup/read", nil),
 		accountDupWriteMeter:  metrics.GetOrRegisterMeter(prefix+"/account/dup/write", nil),
 		accountDupCrossMeter:  metrics.GetOrRegisterMeter(prefix+"/account/dup/cross", nil),
 		accountWasteMeter:     metrics.GetOrRegisterMeter(prefix+"/account/waste", nil),
 
 		storageLoadReadMeter:  metrics.GetOrRegisterMeter(prefix+"/storage/load/read", nil),
 		storageLoadWriteMeter: metrics.GetOrRegisterMeter(prefix+"/storage/load/write", nil),
+		storageLoadTimer:      metrics.GetOrRegisterResettingTimer(prefix+"/storage/load/time", nil),
 		storageDupReadMeter:   metrics.GetOrRegisterMeter(prefix+"/storage/dup/read", nil),
 		storageDupWriteMeter:  metrics.GetOrRegisterMeter(prefix+"/storage/dup/write", nil),
 		storageDupCrossMeter:  metrics.GetOrRegisterMeter(prefix+"/storage/dup/cross", nil),
@@ -120,7 +125,10 @@ func (p *triePrefetcher) report() {
 		if fetcher.root == p.root {
 			p.accountLoadReadMeter.Mark(int64(len(fetcher.seenRead)))
 			p.accountLoadWriteMeter.Mark(int64(len(fetcher.seenWrite)))
-
+			total := len(fetcher.seenRead) + len(fetcher.seenWrite)
+			if total > 0 {
+				p.accountLoadTimer.Update(fetcher.readTime / time.Duration(total))
+			}
 			p.accountDupReadMeter.Mark(int64(fetcher.dupsRead))
 			p.accountDupWriteMeter.Mark(int64(fetcher.dupsWrite))
 			p.accountDupCrossMeter.Mark(int64(fetcher.dupsCross))
@@ -133,7 +141,10 @@ func (p *triePrefetcher) report() {
 		} else {
 			p.storageLoadReadMeter.Mark(int64(len(fetcher.seenRead)))
 			p.storageLoadWriteMeter.Mark(int64(len(fetcher.seenWrite)))
-
+			total := len(fetcher.seenRead) + len(fetcher.seenWrite)
+			if total > 0 {
+				p.storageLoadTimer.Update(fetcher.readTime / time.Duration(total))
+			}
 			p.storageDupReadMeter.Mark(int64(fetcher.dupsRead))
 			p.storageDupWriteMeter.Mark(int64(fetcher.dupsWrite))
 			p.storageDupCrossMeter.Mark(int64(fetcher.dupsCross))
@@ -237,6 +248,7 @@ type subfetcher struct {
 
 	seenRead  map[string]struct{} // Tracks the entries already loaded via read operations
 	seenWrite map[string]struct{} // Tracks the entries already loaded via write operations
+	readTime  time.Duration       // Total time spent on resolving states
 
 	dupsRead  int // Number of duplicate preload tasks via reads only
 	dupsWrite int // Number of duplicate preload tasks via writes only
@@ -378,6 +390,7 @@ func (sf *subfetcher) loop() {
 			sf.tasks = nil
 			sf.lock.Unlock()
 
+			start := time.Now()
 			for _, task := range tasks {
 				key := string(task.key)
 				if task.read {
@@ -410,6 +423,10 @@ func (sf *subfetcher) loop() {
 					sf.seenWrite[key] = struct{}{}
 				}
 			}
+			// Count the time being spent on state resolving. While it's not very
+			// accurate due to some additional operations (e.g., filter out duplicated
+			// task), but it's already good enough for monitoring.
+			sf.readTime += time.Since(start)
 
 		case <-sf.stop:
 			// Termination is requested, abort if no more tasks are pending. If

triedb/pathdb/database.go

@@ -86,6 +86,9 @@ type layer interface {
 	// This is meant to be used during shutdown to persist the layer without
 	// flattening everything down (bad for reorgs).
 	journal(w io.Writer) error
+
+	// isStale returns whether this layer has become stale or if it's still live.
+	isStale() bool
 }
 
 // Config contains the settings for database.

triedb/pathdb/difflayer.go

@@ -41,7 +41,8 @@ type diffLayer struct {
 	memory uint64                                    // Approximate guess as to how much memory we use
 
 	parent layer        // Parent layer modified by this one, never nil, **can be changed**
-	lock   sync.RWMutex // Lock used to protect parent
+	stale  bool         // Signals that the layer became stale (referenced disk layer became stale)
+	lock   sync.RWMutex // Lock used to protect parent and stale fields
 }
 
 // newDiffLayer creates a new diff layer on top of an existing layer.
@@ -95,6 +96,25 @@ func (dl *diffLayer) parentLayer() layer {
 	return dl.parent
 }
 
+// isStale returns whether this layer has become stale or if it's still live.
+func (dl *diffLayer) isStale() bool {
+	dl.lock.RLock()
+	defer dl.lock.RUnlock()
+
+	return dl.stale
+}
+
+// markStale sets the stale flag as true.
+func (dl *diffLayer) markStale() {
+	dl.lock.Lock()
+	defer dl.lock.Unlock()
+
+	if dl.stale {
+		panic("triedb diff layer is stale")
+	}
+	dl.stale = true
+}
+
 // node implements the layer interface, retrieving the trie node blob with the
 // provided node information. No error will be returned if the node is not found.
 func (dl *diffLayer) node(owner common.Hash, path []byte, depth int) ([]byte, common.Hash, *nodeLoc, error) {
@@ -103,6 +123,9 @@ func (dl *diffLayer) node(owner common.Hash, path []byte, depth int) ([]byte, co
 	dl.lock.RLock()
 	defer dl.lock.RUnlock()
 
+	if dl.stale {
+		return nil, common.Hash{}, nil, errSnapshotStale
+	}
 	// If the trie node is known locally, return it
 	subset, ok := dl.nodes[owner]
 	if ok {
@@ -125,7 +148,7 @@ func (dl *diffLayer) update(root common.Hash, id uint64, block uint64, nodes map
 }
 
 // persist flushes the diff layer and all its parent layers to disk layer.
-func (dl *diffLayer) persist(force bool) (layer, error) {
+func (dl *diffLayer) persist(force bool) (*diskLayer, error) {
 	if parent, ok := dl.parentLayer().(*diffLayer); ok {
 		// Hold the lock to prevent any read operation until the new
 		// parent is linked correctly.
@@ -147,7 +170,7 @@ func (dl *diffLayer) persist(force bool) (layer, error) {
 
 // diffToDisk merges a bottom-most diff into the persistent disk layer underneath
 // it. The method will panic if called onto a non-bottom-most diff layer.
-func diffToDisk(layer *diffLayer, force bool) (layer, error) {
+func diffToDisk(layer *diffLayer, force bool) (*diskLayer, error) {
 	disk, ok := layer.parentLayer().(*diskLayer)
 	if !ok {
 		panic(fmt.Sprintf("unknown layer type: %T", layer.parentLayer()))

triedb/pathdb/disklayer.go

@@ -73,7 +73,7 @@ func (dl *diskLayer) parentLayer() layer {
 	return nil
 }
 
-// isStale return whether this layer has become stale (was flattened across) or if
+// isStale returns whether this layer has become stale (was flattened across) or if
 // it's still live.
 func (dl *diskLayer) isStale() bool {
 	dl.lock.RLock()