core, core/state: instrument BAL slow-block metrics (#34861)

A first attempt to get the entire slow_blocks to be
BAL&parallel-exec-aware.

- Sequential-path output is unchanged; the `bal` JSON field is
`omitempty`.
- All new types are pure-data snapshots (value receivers, no shared
mutation), so they're safe to copy across goroutine boundaries in the
parallel pipeline.

---------

Co-authored-by: Jared Wasinger <j-wasinger@hotmail.com>

core/blockchain.go

@@ -656,6 +656,20 @@ func (bc *BlockChain) processBlockWithAccessList(parentRoot common.Hash, block *
 	writeTime := time.Since(writeStart)
 	var stats ExecuteStats
 
+	wc := stateTransition.WrittenCounts()
+	d := stateTransition.Deletions()
+	codeLoaded, codeLoadBytes := prefetchReader.(state.CodeLoadTracker).CodeLoads()
+	stats.AccountLoaded = al.UniqueAccountCount()
+	stats.AccountUpdated = wc.Accounts - d.Accounts
+	stats.AccountDeleted = d.Accounts
+	stats.StorageLoaded = al.UniqueStorageSlotCount()
+	stats.StorageUpdated = wc.StorageSlots - d.Storage
+	stats.StorageDeleted = d.Storage
+	stats.CodeLoaded = codeLoaded
+	stats.CodeLoadBytes = codeLoadBytes
+	stats.CodeUpdated = wc.Codes
+	stats.CodeUpdateBytes = wc.CodeBytes
+
 	stats.ExecWall = res.ExecTime
 	stats.PostProcess = res.PostProcessTime
 
@@ -666,12 +680,9 @@ func (bc *BlockChain) processBlockWithAccessList(parentRoot common.Hash, block *
 		stats.DatabaseCommit = m.TrieDBCommits
 		stats.Prefetch = m.StatePrefetch
 	}
-
 	stats.Prefetch = prefetchReader.(state.PrefetcherMetricer).Metrics().Elapsed
 
-	if r, ok := prefetchReader.(state.ReaderStater); ok {
-		stats.StateReadCacheStats = r.GetStats()
-	}
+	stats.StateReadCacheStats = prefetchReader.(state.ReaderStater).GetStats()
 
 	elapsed := time.Since(startTime) + 1 // prevent zero division
 	stats.TotalTime = elapsed
@@ -2436,11 +2447,11 @@ func (bc *BlockChain) ProcessBlock(ctx context.Context, parentRoot common.Hash,
 		stats    = &ExecuteStats{}
 	)
 	// Update the metrics touched during block processing and validation
-	stats.AccountReads = statedb.AccountReads     // Account reads are complete(in processing)
-	stats.StorageReads = statedb.StorageReads     // Storage reads are complete(in processing)
-	stats.AccountUpdates = statedb.AccountUpdates // Account updates are complete(in validation)
-	stats.StorageUpdates = statedb.StorageUpdates // Storage updates are complete(in validation)
-	stats.AccountHashes = statedb.AccountHashes   // Account hashes are complete(in validation)
+	stats.AccountReads = statedb.AccountReads     // Account reads are complete (in processing)
+	stats.StorageReads = statedb.StorageReads     // Storage reads are complete (in processing)
+	stats.AccountUpdates = statedb.AccountUpdates // Account updates are complete (in validation)
+	stats.StorageUpdates = statedb.StorageUpdates // Storage updates are complete (in validation)
+	stats.AccountHashes = statedb.AccountHashes   // Account hashes are complete (in validation)
 	stats.CodeReads = statedb.CodeReads
 
 	stats.AccountLoaded = statedb.AccountLoaded
@@ -2455,9 +2466,9 @@ func (bc *BlockChain) ProcessBlock(ctx context.Context, parentRoot common.Hash,
 	stats.CodeUpdated = statedb.CodeUpdated
 	stats.CodeUpdateBytes = statedb.CodeUpdateBytes
 
-	stats.Execution = ptime - (statedb.AccountReads + statedb.StorageReads + statedb.CodeReads)          // The time spent on EVM processing
-	stats.Validation = vtime - (statedb.AccountHashes + statedb.AccountUpdates + statedb.StorageUpdates) // The time spent on block validation
-	stats.CrossValidation = xvtime                                                                       // The time spent on stateless cross validation
+	stats.Execution = ptime - (statedb.AccountReads + statedb.StorageReads + statedb.CodeReads)          // EVM processing time
+	stats.Validation = vtime - (statedb.AccountHashes + statedb.AccountUpdates + statedb.StorageUpdates) // Block validation time
+	stats.CrossValidation = xvtime
 
 	// Write the block to the chain and get the status.
 	var status WriteStatus

core/blockchain_stats.go

@@ -38,17 +38,16 @@ type ExecuteStats struct {
 	StorageCommits time.Duration // Time spent on the storage trie commit
 	CodeReads      time.Duration // Time spent on the contract code read
 
-	// TODO: code bytes loaded
-	AccountLoaded   int // Number of accounts loaded
-	AccountUpdated  int // Number of accounts updated
-	AccountDeleted  int // Number of accounts deleted
-	StorageLoaded   int // Number of storage slots loaded
-	StorageUpdated  int // Number of storage slots updated
-	StorageDeleted  int // Number of storage slots deleted
-	CodeLoaded      int // Number of contract code loaded
-	CodeLoadBytes   int // Number of bytes read from contract code
-	CodeUpdated     int // Number of contract code written (CREATE/CREATE2 + EIP-7702)
-	CodeUpdateBytes int // Total bytes of code written
+	AccountLoaded   int
+	AccountUpdated  int
+	AccountDeleted  int
+	StorageLoaded   int
+	StorageUpdated  int
+	StorageDeleted  int
+	CodeLoaded      int
+	CodeLoadBytes   int
+	CodeUpdated     int
+	CodeUpdateBytes int
 
 	Execution       time.Duration // Time spent on the EVM execution
 	Validation      time.Duration // Time spent on the block validation
@@ -223,7 +222,7 @@ func buildSlowBlockLog(s *ExecuteStats, block *types.Block) slowBlockLog {
 		},
 		Timing: slowBlockTime{
 			ExecutionMs: durationToMs(s.Execution),
-			StateReadMs: durationToMs(s.AccountReads + s.StorageReads + s.CodeReads),
+			StateReadMs: durationToMs(s.AccountReads + s.StorageReads + s.CodeReads + s.Prefetch),
 			StateHashMs: durationToMs(s.AccountHashes + s.AccountUpdates + s.StorageUpdates),
 			CommitMs:    durationToMs(max(s.AccountCommits, s.StorageCommits) + s.DatabaseCommit + s.BlockWrite),
 			TotalMs:     durationToMs(s.TotalTime),
@@ -304,8 +303,8 @@ func (s *ExecuteStats) logSlow(block *types.Block, slowBlockThreshold time.Durat
 }
 
 func (s *ExecuteStats) reportBALMetrics() {
-	accountCommitTimer.Update(s.AccountCommits) // Account commits are complete, we can mark them
-	storageCommitTimer.Update(s.StorageCommits) // Storage commits are complete, we can mark them
+	accountCommitTimer.Update(s.AccountCommits)
+	storageCommitTimer.Update(s.StorageCommits)
 
 	if m := s.balTransitionStats; m != nil {
 		stateTriePrefetchTimer.Update(m.StatePrefetch)

core/parallel_state_processor.go

@@ -14,16 +14,13 @@ import (
 	"golang.org/x/sync/errgroup"
 )
 
-// ProcessResultWithMetrics wraps ProcessResult with some metrics that are
-// emitted when executing blocks containing access lists.
+// ProcessResultWithMetrics wraps ProcessResult with timing breakdown for BAL block processing.
 type ProcessResultWithMetrics struct {
 	ProcessResult          *ProcessResult
 	PreProcessTime         time.Duration
 	StateTransitionMetrics *state.BALStateTransitionMetrics
-	// the time it took to execute all txs in the block
-	ExecTime        time.Duration
-	PostProcessTime time.Duration
-	// TODO: have the prefetch metric in here as well?
+	ExecTime               time.Duration
+	PostProcessTime        time.Duration
 }
 
 // ParallelStateProcessor is used to execute and verify blocks containing

core/state/bal_state_transition.go

@@ -3,6 +3,7 @@ package state
 import (
 	"maps"
 	"sync"
+	"sync/atomic"
 	"time"
 
 	"github.com/ethereum/go-ethereum/common"
@@ -20,6 +21,7 @@ import (
 // this object is only used for a single block.
 type BALStateTransition struct {
 	accessList bal.AccessListReader
+	written    bal.WrittenCounts
 	db         Database
 	reader     Reader
 	stateTrie  Trie
@@ -40,6 +42,11 @@ type BALStateTransition struct {
 	tries     sync.Map //map[common.Address]Trie
 	deletions map[common.Address]struct{}
 
+	// Deletion counters; not derivable from the BAL alone (selfdestruct vs
+	// balance/nonce reset is indistinguishable without prestate).
+	accountDeleted int
+	storageDeleted atomic.Int64
+
 	stateUpdate *stateUpdate
 
 	metrics   BALStateTransitionMetrics
@@ -52,6 +59,19 @@ func (s *BALStateTransition) Metrics() *BALStateTransitionMetrics {
 	return &s.metrics
 }
 
+// DeletionCounts holds per-block deletion counters from the parallel root-pass.
+type DeletionCounts struct {
+	Accounts int
+	Storage  int
+}
+
+func (s *BALStateTransition) Deletions() DeletionCounts {
+	return DeletionCounts{
+		Accounts: s.accountDeleted,
+		Storage:  int(s.storageDeleted.Load()),
+	}
+}
+
 type BALStateTransitionMetrics struct {
 	// trie hashing metrics
 	AccountUpdate time.Duration
@@ -75,6 +95,7 @@ func NewBALStateTransition(block *types.Block, prefetchReader Reader, db Databas
 
 	return &BALStateTransition{
 		accessList:  bal.NewAccessListReader(*block.AccessList()),
+		written:     block.AccessList().WrittenCounts(),
 		db:          db,
 		reader:      prefetchReader,
 		stateTrie:   stateTrie,
@@ -90,6 +111,11 @@ func NewBALStateTransition(block *types.Block, prefetchReader Reader, db Databas
 	}, nil
 }
 
+// WrittenCounts returns the cached BAL write counts (computed once per block).
+func (s *BALStateTransition) WrittenCounts() bal.WrittenCounts {
+	return s.written
+}
+
 func (s *BALStateTransition) Error() error {
 	return s.err
 }
@@ -334,15 +360,11 @@ func (s *BALStateTransition) CommitWithUpdate(block uint64, deleteEmptyObjects b
 		return common.Hash{}, nil, err
 	}
 
-	/*
-			TODO: derive these from the BAL
-			^ I think even then, there is a semantic difference with how these metrics were calculated previously
-		    I don't know if it makes sense to recompute those, or just derive new ones from the BAL
-			accountUpdatedMeter.Mark(int64(s.accountUpdated))
-			storageUpdatedMeter.Mark(s.storageUpdated.Load())
-			accountDeletedMeter.Mark(int64(s.accountDeleted))
-			storageDeletedMeter.Mark(s.storageDeleted.Load())
-	*/
+	storageDeleted := s.storageDeleted.Load()
+	accountUpdatedMeter.Mark(int64(s.written.Accounts - s.accountDeleted))
+	storageUpdatedMeter.Mark(int64(s.written.StorageSlots) - storageDeleted)
+	accountDeletedMeter.Mark(int64(s.accountDeleted))
+	storageDeletedMeter.Mark(storageDeleted)
 	accountTrieUpdatedMeter.Mark(int64(accountTrieNodesUpdated))
 	accountTrieDeletedMeter.Mark(int64(accountTrieNodesDeleted))
 	storageTriesUpdatedMeter.Mark(int64(storageTrieNodesUpdated))
@@ -477,6 +499,7 @@ func (s *BALStateTransition) IntermediateRoot(_ bool) common.Hash {
 				return common.Hash{}
 			}
 			s.deletions[mutatedAddr] = struct{}{}
+			s.accountDeleted++
 		} else {
 			acct, code := s.updateAccount(mutatedAddr)
 

core/state/reader.go

@@ -18,6 +18,9 @@ package state
 
 import (
 	"errors"
+	"sync"
+	"sync/atomic"
+
 	"github.com/ethereum/go-ethereum/common"
 	"github.com/ethereum/go-ethereum/core/overlay"
 	"github.com/ethereum/go-ethereum/core/types"
@@ -28,8 +31,6 @@ import (
 	"github.com/ethereum/go-ethereum/trie/transitiontrie"
 	"github.com/ethereum/go-ethereum/triedb"
 	"github.com/ethereum/go-ethereum/triedb/database"
-	"sync"
-	"sync/atomic"
 )
 
 // ContractCodeReader defines the interface for accessing contract code.
@@ -530,6 +531,8 @@ type reader struct {
 	ContractCodeReader
 	StateReader
 	PrefetcherMetricer
+
+	codeLoaded sync.Map // common.Address → int (first-seen len(code))
 }
 
 // newReader constructs a reader with the supplied code reader and state reader.
@@ -548,6 +551,33 @@ func newReaderWithPrefetch(codeReader ContractCodeReader, stateReader StateReade
 	}
 }
 
+func (r *reader) Code(addr common.Address, codeHash common.Hash) []byte {
+	code := r.ContractCodeReader.Code(addr, codeHash)
+	if len(code) > 0 {
+		r.codeLoaded.LoadOrStore(addr, len(code))
+	}
+	return code
+}
+
+func (r *reader) CodeSize(addr common.Address, codeHash common.Hash) (int, error) {
+	size, err := r.ContractCodeReader.CodeSize(addr, codeHash)
+	if err == nil && size > 0 {
+		r.codeLoaded.LoadOrStore(addr, size)
+	}
+	return size, err
+}
+
+// CodeLoads returns the count of unique contracts whose code was fetched and
+// the sum of their first-seen byte lengths. Call after Reader use has quiesced.
+func (r *reader) CodeLoads() (count, bytes int) {
+	r.codeLoaded.Range(func(_, v any) bool {
+		count++
+		bytes += v.(int)
+		return true
+	})
+	return
+}
+
 // GetCodeStats returns the statistics of code access.
 func (r *reader) GetCodeStats() ContractCodeReaderStats {
 	if stater, ok := r.ContractCodeReader.(ContractCodeReaderStater); ok {

core/state/reader_eip_7928.go

@@ -382,3 +382,16 @@ func (r *readerTracker) TouchStorage(addr common.Address, slot common.Hash) {
 	}
 	list[slot] = struct{}{}
 }
+
+func (r *readerTracker) CodeLoads() (count, bytes int) {
+	return r.Reader.(CodeLoadTracker).CodeLoads()
+}
+
+// GetStateStats forwards stats from the wrapped reader; without this, BAL
+// blocks would emit zero cache hit/miss counts.
+func (r *prefetchStateReader) GetStateStats() StateReaderStats {
+	if stater, ok := r.StateReader.(StateReaderStater); ok {
+		return stater.GetStateStats()
+	}
+	return StateReaderStats{}
+}

core/state/reader_eip_7928_test.go

@@ -263,3 +263,31 @@ func TestTrackerSurvivesStateDBCache(t *testing.T) {
 		t.Fatal("slot must be tracked on cache hit (storage)")
 	}
 }
+
+// TestPrefetchStateReaderForwardsStats locks down that prefetchStateReader
+// exposes the underlying stateReaderWithStats counters via GetStateStats.
+func TestPrefetchStateReaderForwardsStats(t *testing.T) {
+	stub := newRefStateReader()
+	addr := testrand.Address()
+
+	cached := newStateReaderWithCache(stub)
+	withStats := newStateReaderWithStats(cached)
+	prefetch := newPrefetchStateReaderInternal(withStats, nil, 1)
+
+	if _, err := prefetch.Account(addr); err != nil {
+		t.Fatalf("Account: %v", err)
+	}
+	if _, err := prefetch.Account(addr); err != nil {
+		t.Fatalf("Account (second): %v", err)
+	}
+
+	stats := withStats.GetStateStats()
+	if stats.AccountCacheHit == 0 || stats.AccountCacheMiss == 0 {
+		t.Fatalf("inner stats not populated: %+v", stats)
+	}
+	gotStats := prefetch.GetStateStats()
+	if gotStats != stats {
+		t.Fatalf("forward mismatch: got %+v, want %+v", gotStats, stats)
+	}
+}
+

core/state/reader_stater.go

@@ -16,6 +16,11 @@
 
 package state
 
+// CodeLoadTracker exposes a Reader's deduplicated code-load count and bytes.
+type CodeLoadTracker interface {
+	CodeLoads() (count, bytes int)
+}
+
 // ContractCodeReaderStats aggregates statistics for the contract code reader.
 type ContractCodeReaderStats struct {
 	CacheHit       int64 // Number of cache hits

core/types/bal/bal_encoding.go

@@ -44,6 +44,49 @@ import (
 // BlockAccessList is the encoding format of AccessListBuilder.
 type BlockAccessList []AccountAccess
 
+// UniqueAccountCount returns the number of distinct account addresses in
+// the block access list.
+func (e BlockAccessList) UniqueAccountCount() int {
+	return len(e)
+}
+
+// UniqueStorageSlotCount returns the total number of distinct (address, slot)
+// pairs accessed across all accounts. Reads and writes are disjoint per
+// account by spec validation, so we can sum them directly.
+func (e BlockAccessList) UniqueStorageSlotCount() int {
+	var n int
+	for i := range e {
+		n += len(e[i].StorageReads) + len(e[i].StorageChanges)
+	}
+	return n
+}
+
+// WrittenCounts groups per-block aggregate write counts derived from the BAL.
+type WrittenCounts struct {
+	Accounts     int
+	StorageSlots int
+	Codes        int
+	CodeBytes    int
+}
+
+// WrittenCounts walks the BAL once and returns the aggregate write counts.
+func (e BlockAccessList) WrittenCounts() WrittenCounts {
+	var w WrittenCounts
+	for i := range e {
+		a := &e[i]
+		if len(a.StorageChanges) > 0 || len(a.BalanceChanges) > 0 ||
+			len(a.NonceChanges) > 0 || len(a.CodeChanges) > 0 {
+			w.Accounts++
+		}
+		w.StorageSlots += len(a.StorageChanges)
+		if n := len(a.CodeChanges); n > 0 {
+			w.Codes++
+			w.CodeBytes += len(a.CodeChanges[n-1].Code)
+		}
+	}
+	return w
+}
+
 func (e BlockAccessList) EncodeRLP(_w io.Writer) error {
 	w := rlp.NewEncoderBuffer(_w)
 	l := w.List()