all: introduce codedb and simplify cachingDB

cmd/evm/internal/t8ntool/execution.go

@@ -30,6 +30,7 @@ import (
 	"github.com/ethereum/go-ethereum/core"
 	"github.com/ethereum/go-ethereum/core/rawdb"
 	"github.com/ethereum/go-ethereum/core/state"
+	"github.com/ethereum/go-ethereum/core/state/codedb"
 	"github.com/ethereum/go-ethereum/core/tracing"
 	"github.com/ethereum/go-ethereum/core/types"
 	"github.com/ethereum/go-ethereum/core/vm"
@@ -379,7 +380,7 @@ func (pre *Prestate) Apply(vmConfig vm.Config, chainConfig *params.ChainConfig,
 
 func MakePreState(db ethdb.Database, accounts types.GenesisAlloc, isBintrie bool) *state.StateDB {
 	tdb := triedb.NewDatabase(db, &triedb.Config{Preimages: true, IsVerkle: isBintrie})
-	sdb := state.NewDatabase(tdb, nil)
+	sdb := state.NewDatabase(tdb, codedb.New(db))
 
 	root := types.EmptyRootHash
 	if isBintrie {

cmd/evm/runner.go

@@ -35,6 +35,7 @@ import (
 	"github.com/ethereum/go-ethereum/core"
 	"github.com/ethereum/go-ethereum/core/rawdb"
 	"github.com/ethereum/go-ethereum/core/state"
+	"github.com/ethereum/go-ethereum/core/state/codedb"
 	"github.com/ethereum/go-ethereum/core/tracing"
 	"github.com/ethereum/go-ethereum/core/types"
 	"github.com/ethereum/go-ethereum/core/vm"
@@ -226,8 +227,9 @@ func runCmd(ctx *cli.Context) error {
 		HashDB:    hashdb.Defaults,
 	})
 	defer triedb.Close()
+
 	genesis := genesisConfig.MustCommit(db, triedb)
-	sdb := state.NewDatabase(triedb, nil)
+	sdb := state.NewDatabase(triedb, codedb.New(db))
 	prestate, _ = state.New(genesis.Root(), sdb)
 	chainConfig = genesisConfig.Config
 

cmd/geth/chaincmd.go

@@ -37,6 +37,7 @@ import (
 	"github.com/ethereum/go-ethereum/core/history"
 	"github.com/ethereum/go-ethereum/core/rawdb"
 	"github.com/ethereum/go-ethereum/core/state"
+	"github.com/ethereum/go-ethereum/core/state/codedb"
 	"github.com/ethereum/go-ethereum/core/types"
 	"github.com/ethereum/go-ethereum/crypto"
 	"github.com/ethereum/go-ethereum/ethdb"
@@ -650,7 +651,7 @@ func dump(ctx *cli.Context) error {
 	triedb := utils.MakeTrieDatabase(ctx, stack, db, true, true, false) // always enable preimage lookup
 	defer triedb.Close()
 
-	state, err := state.New(root, state.NewDatabase(triedb, nil))
+	state, err := state.New(root, state.NewDatabase(triedb, codedb.New(db)))
 	if err != nil {
 		return err
 	}

core/bintrie_witness_test.go

@@ -27,6 +27,7 @@ import (
 	"github.com/ethereum/go-ethereum/consensus/ethash"
 	"github.com/ethereum/go-ethereum/core/rawdb"
 	"github.com/ethereum/go-ethereum/core/state"
+	"github.com/ethereum/go-ethereum/core/state/codedb"
 	"github.com/ethereum/go-ethereum/core/tracing"
 	"github.com/ethereum/go-ethereum/core/types"
 	"github.com/ethereum/go-ethereum/core/vm"
@@ -220,7 +221,7 @@ func TestProcessParentBlockHash(t *testing.T) {
 		cacheConfig := DefaultConfig().WithStateScheme(rawdb.PathScheme)
 		cacheConfig.SnapshotLimit = 0
 		triedb := triedb.NewDatabase(db, cacheConfig.triedbConfig(true))
-		statedb, _ := state.New(types.EmptyVerkleHash, state.NewDatabase(triedb, nil))
+		statedb, _ := state.New(types.EmptyVerkleHash, state.NewDatabase(triedb, codedb.New(db)))
 		checkBlockHashes(statedb, true)
 	})
 }

core/blockchain.go

@@ -39,6 +39,7 @@ import (
 	"github.com/ethereum/go-ethereum/core/history"
 	"github.com/ethereum/go-ethereum/core/rawdb"
 	"github.com/ethereum/go-ethereum/core/state"
+	"github.com/ethereum/go-ethereum/core/state/codedb"
 	"github.com/ethereum/go-ethereum/core/state/snapshot"
 	"github.com/ethereum/go-ethereum/core/stateless"
 	"github.com/ethereum/go-ethereum/core/tracing"
@@ -72,9 +73,10 @@ var (
 	accountUpdateTimer = metrics.NewRegisteredResettingTimer("chain/account/updates", nil)
 	accountCommitTimer = metrics.NewRegisteredResettingTimer("chain/account/commits", nil)
 
-	storageReadTimer   = metrics.NewRegisteredResettingTimer("chain/storage/reads", nil)
-	storageUpdateTimer = metrics.NewRegisteredResettingTimer("chain/storage/updates", nil)
-	storageCommitTimer = metrics.NewRegisteredResettingTimer("chain/storage/commits", nil)
+	storageReadTimer    = metrics.NewRegisteredResettingTimer("chain/storage/reads", nil)
+	storageUpdateTimer  = metrics.NewRegisteredResettingTimer("chain/storage/updates", nil)
+	storageCommitTimer  = metrics.NewRegisteredResettingTimer("chain/storage/commits", nil)
+	databaseCommitTimer = metrics.NewRegisteredResettingTimer("chain/database/commits", nil)
 
 	accountCacheHitMeter  = metrics.NewRegisteredMeter("chain/account/reads/cache/process/hit", nil)
 	accountCacheMissMeter = metrics.NewRegisteredMeter("chain/account/reads/cache/process/miss", nil)
@@ -89,9 +91,6 @@ var (
 	accountReadSingleTimer = metrics.NewRegisteredResettingTimer("chain/account/single/reads", nil)
 	storageReadSingleTimer = metrics.NewRegisteredResettingTimer("chain/storage/single/reads", nil)
 
-	snapshotCommitTimer = metrics.NewRegisteredResettingTimer("chain/snapshot/commits", nil)
-	triedbCommitTimer   = metrics.NewRegisteredResettingTimer("chain/triedb/commits", nil)
-
 	blockInsertTimer          = metrics.NewRegisteredResettingTimer("chain/inserts", nil)
 	blockValidationTimer      = metrics.NewRegisteredResettingTimer("chain/validation", nil)
 	blockCrossValidationTimer = metrics.NewRegisteredResettingTimer("chain/crossvalidation", nil)
@@ -300,7 +299,7 @@ type BlockChain struct {
 	lastWrite     uint64                           // Last block when the state was flushed
 	flushInterval atomic.Int64                     // Time interval (processing time) after which to flush a state
 	triedb        *triedb.Database                 // The database handler for maintaining trie nodes.
-	statedb       *state.CachingDB                 // State database to reuse between imports (contains state cache)
+	codedb        *codedb.Database                 // The database handler for maintaining contract codes
 	txIndexer     *txIndexer                       // Transaction indexer, might be nil if not enabled
 
 	hc               *HeaderChain
@@ -381,6 +380,7 @@ func NewBlockChain(db ethdb.Database, genesis *Genesis, engine consensus.Engine,
 		cfg:                cfg,
 		db:                 db,
 		triedb:             triedb,
+		codedb:             codedb.New(db),
 		triegc:             prque.New[int64, common.Hash](nil),
 		chainmu:            syncx.NewClosableMutex(),
 		bodyCache:          lru.NewCache[common.Hash, *types.Body](bodyCacheLimit),
@@ -397,7 +397,6 @@ func NewBlockChain(db ethdb.Database, genesis *Genesis, engine consensus.Engine,
 		return nil, err
 	}
 	bc.flushInterval.Store(int64(cfg.TrieTimeLimit))
-	bc.statedb = state.NewDatabase(bc.triedb, nil)
 	bc.validator = NewBlockValidator(chainConfig, bc)
 	bc.prefetcher = newStatePrefetcher(chainConfig, bc.hc)
 	bc.processor = NewStateProcessor(bc.hc)
@@ -574,9 +573,6 @@ func (bc *BlockChain) setupSnapshot() {
 			AsyncBuild: !bc.cfg.SnapshotWait,
 		}
 		bc.snaps, _ = snapshot.New(snapconfig, bc.db, bc.triedb, head.Root)
-
-		// Re-initialize the state database with snapshot
-		bc.statedb = state.NewDatabase(bc.triedb, bc.snaps)
 	}
 }
 
@@ -1610,14 +1606,20 @@ func (bc *BlockChain) writeBlockWithState(block *types.Block, receipts []*types.
 		log.Crit("Failed to write block into disk", "err", err)
 	}
 	// Commit all cached state changes into underlying memory database.
-	root, stateUpdate, err := statedb.CommitWithUpdate(block.NumberU64(), bc.chainConfig.IsEIP158(block.Number()), bc.chainConfig.IsCancun(block.Number(), block.Time()))
+	var (
+		err      error
+		root     common.Hash
+		isEIP158 = bc.chainConfig.IsEIP158(block.Number())
+		isCancun = bc.chainConfig.IsCancun(block.Number(), block.Time())
+	)
+	if bc.stateSizer == nil {
+		root, err = statedb.Commit(block.NumberU64(), isEIP158, isCancun)
+	} else {
+		root, err = statedb.CommitWithUpdate(block.NumberU64(), isEIP158, isCancun, bc.stateSizer)
+	}
 	if err != nil {
 		return err
 	}
-	// Emit the state update to the state sizestats if it's active
-	if bc.stateSizer != nil {
-		bc.stateSizer.Notify(stateUpdate)
-	}
 	// If node is running in path mode, skip explicit gc operation
 	// which is unnecessary in this mode.
 	if bc.triedb.Scheme() == rawdb.PathScheme {
@@ -2008,11 +2010,12 @@ func (bc *BlockChain) ProcessBlock(parentRoot common.Hash, block *types.Block, s
 		startTime = time.Now()
 		statedb   *state.StateDB
 		interrupt atomic.Bool
+		sdb       = state.NewDatabase(bc.triedb, bc.codedb).WithSnapshot(bc.snaps)
 	)
 	defer interrupt.Store(true) // terminate the prefetch at the end
 
 	if bc.cfg.NoPrefetch {
-		statedb, err = state.New(parentRoot, bc.statedb)
+		statedb, err = state.New(parentRoot, sdb)
 		if err != nil {
 			return nil, err
 		}
@@ -2022,15 +2025,15 @@ func (bc *BlockChain) ProcessBlock(parentRoot common.Hash, block *types.Block, s
 		//
 		// Note: the main processor and prefetcher share the same reader with a local
 		// cache for mitigating the overhead of state access.
-		prefetch, process, err := bc.statedb.ReadersWithCacheStats(parentRoot)
+		prefetch, process, err := sdb.ReadersWithCacheStats(parentRoot)
 		if err != nil {
 			return nil, err
 		}
-		throwaway, err := state.NewWithReader(parentRoot, bc.statedb, prefetch)
+		throwaway, err := state.NewWithReader(parentRoot, sdb, prefetch)
 		if err != nil {
 			return nil, err
 		}
-		statedb, err = state.NewWithReader(parentRoot, bc.statedb, process)
+		statedb, err = state.NewWithReader(parentRoot, sdb, process)
 		if err != nil {
 			return nil, err
 		}
@@ -2192,11 +2195,10 @@ func (bc *BlockChain) ProcessBlock(parentRoot common.Hash, block *types.Block, s
 	}
 
 	// Update the metrics touched during block commit
-	stats.AccountCommits = statedb.AccountCommits  // Account commits are complete, we can mark them
-	stats.StorageCommits = statedb.StorageCommits  // Storage commits are complete, we can mark them
-	stats.SnapshotCommit = statedb.SnapshotCommits // Snapshot commits are complete, we can mark them
-	stats.TrieDBCommit = statedb.TrieDBCommits     // Trie database commits are complete, we can mark them
-	stats.BlockWrite = time.Since(wstart) - max(statedb.AccountCommits, statedb.StorageCommits) /* concurrent */ - statedb.SnapshotCommits - statedb.TrieDBCommits
+	stats.AccountCommits = statedb.AccountCommits // Account commits are complete, we can mark them
+	stats.StorageCommits = statedb.StorageCommits // Storage commits are complete, we can mark them
+	stats.DBCommit = statedb.DatabaseCommits      // Trie database commits are complete, we can mark them
+	stats.BlockWrite = time.Since(wstart) - max(statedb.AccountCommits, statedb.StorageCommits) /* concurrent */ - statedb.DatabaseCommits
 
 	elapsed := time.Since(startTime) + 1 // prevent zero division
 	stats.TotalTime = elapsed

core/blockchain_reader.go

@@ -26,6 +26,7 @@ import (
 	"github.com/ethereum/go-ethereum/consensus/misc/eip4844"
 	"github.com/ethereum/go-ethereum/core/rawdb"
 	"github.com/ethereum/go-ethereum/core/state"
+	"github.com/ethereum/go-ethereum/core/state/codedb"
 	"github.com/ethereum/go-ethereum/core/state/snapshot"
 	"github.com/ethereum/go-ethereum/core/types"
 	"github.com/ethereum/go-ethereum/core/vm"
@@ -371,7 +372,7 @@ func (bc *BlockChain) TxIndexDone() bool {
 
 // HasState checks if state trie is fully present in the database or not.
 func (bc *BlockChain) HasState(hash common.Hash) bool {
-	_, err := bc.statedb.OpenTrie(hash)
+	_, err := bc.triedb.NodeReader(hash)
 	return err == nil
 }
 
@@ -403,7 +404,7 @@ func (bc *BlockChain) stateRecoverable(root common.Hash) bool {
 func (bc *BlockChain) ContractCodeWithPrefix(hash common.Hash) []byte {
 	// TODO(rjl493456442) The associated account address is also required
 	// in Verkle scheme. Fix it once snap-sync is supported for Verkle.
-	return bc.statedb.ContractCodeWithPrefix(common.Address{}, hash)
+	return bc.codedb.Reader().CodeWithPrefix(common.Address{}, hash)
 }
 
 // State returns a new mutable state based on the current HEAD block.
@@ -413,14 +414,14 @@ func (bc *BlockChain) State() (*state.StateDB, error) {
 
 // StateAt returns a new mutable state based on a particular point in time.
 func (bc *BlockChain) StateAt(root common.Hash) (*state.StateDB, error) {
-	return state.New(root, bc.statedb)
+	return state.New(root, state.NewDatabase(bc.triedb, bc.codedb).WithSnapshot(bc.snaps))
 }
 
 // HistoricState returns a historic state specified by the given root.
 // Live states are not available and won't be served, please use `State`
 // or `StateAt` instead.
 func (bc *BlockChain) HistoricState(root common.Hash) (*state.StateDB, error) {
-	return state.New(root, state.NewHistoricDatabase(bc.db, bc.triedb))
+	return state.New(root, state.NewHistoricDatabase(bc.triedb, bc.codedb))
 }
 
 // Config retrieves the chain's fork configuration.
@@ -444,11 +445,6 @@ func (bc *BlockChain) Processor() Processor {
 	return bc.processor
 }
 
-// StateCache returns the caching database underpinning the blockchain instance.
-func (bc *BlockChain) StateCache() state.Database {
-	return bc.statedb
-}
-
 // GasLimit returns the gas limit of the current HEAD block.
 func (bc *BlockChain) GasLimit() uint64 {
 	return bc.CurrentBlock().GasLimit
@@ -492,6 +488,11 @@ func (bc *BlockChain) TrieDB() *triedb.Database {
 	return bc.triedb
 }
 
+// CodeDB retrieves the low level contract code database used for data storage.
+func (bc *BlockChain) CodeDB() *codedb.Database {
+	return bc.codedb
+}
+
 // HeaderChain returns the underlying header chain.
 func (bc *BlockChain) HeaderChain() *HeaderChain {
 	return bc.hc

core/blockchain_sethead_test.go

@@ -30,7 +30,6 @@ import (
 	"github.com/ethereum/go-ethereum/common"
 	"github.com/ethereum/go-ethereum/consensus/ethash"
 	"github.com/ethereum/go-ethereum/core/rawdb"
-	"github.com/ethereum/go-ethereum/core/state"
 	"github.com/ethereum/go-ethereum/core/types"
 	"github.com/ethereum/go-ethereum/ethdb/pebble"
 	"github.com/ethereum/go-ethereum/params"
@@ -2041,7 +2040,6 @@ func testSetHeadWithScheme(t *testing.T, tt *rewindTest, snapshots bool, scheme
 		dbconfig.HashDB = hashdb.Defaults
 	}
 	chain.triedb = triedb.NewDatabase(chain.db, dbconfig)
-	chain.statedb = state.NewDatabase(chain.triedb, chain.snaps)
 
 	// Force run a freeze cycle
 	type freezer interface {

core/blockchain_stats.go

@@ -48,8 +48,7 @@ type ExecuteStats struct {
 	Execution       time.Duration // Time spent on the EVM execution
 	Validation      time.Duration // Time spent on the block validation
 	CrossValidation time.Duration // Optional, time spent on the block cross validation
-	SnapshotCommit  time.Duration // Time spent on snapshot commit
-	TrieDBCommit    time.Duration // Time spent on database commit
+	DBCommit        time.Duration // Time spent on database commit
 	BlockWrite      time.Duration // Time spent on block write
 	TotalTime       time.Duration // The total time spent on block execution
 	MgasPerSecond   float64       // The million gas processed per second
@@ -80,8 +79,7 @@ func (s *ExecuteStats) reportMetrics() {
 	blockExecutionTimer.Update(s.Execution)                 // The time spent on EVM processing
 	blockValidationTimer.Update(s.Validation)               // The time spent on block validation
 	blockCrossValidationTimer.Update(s.CrossValidation)     // The time spent on stateless cross validation
-	snapshotCommitTimer.Update(s.SnapshotCommit)            // Snapshot commits are complete, we can mark them
-	triedbCommitTimer.Update(s.TrieDBCommit)                // Trie database commits are complete, we can mark them
+	databaseCommitTimer.Update(s.DBCommit)                  // Trie database commits are complete, we can mark them
 	blockWriteTimer.Update(s.BlockWrite)                    // The time spent on block write
 	blockInsertTimer.Update(s.TotalTime)                    // The total time spent on block execution
 	chainMgaspsMeter.Update(time.Duration(s.MgasPerSecond)) // TODO(rjl493456442) generalize the ResettingTimer
@@ -127,7 +125,7 @@ Block write: %v
 		common.PrettyDuration(s.StorageReads), s.StorageLoaded,
 		common.PrettyDuration(s.AccountHashes+s.AccountCommits+s.AccountUpdates),
 		common.PrettyDuration(s.StorageCommits+s.StorageUpdates),
-		common.PrettyDuration(s.TrieDBCommit+s.SnapshotCommit), common.PrettyDuration(s.BlockWrite),
+		common.PrettyDuration(s.DBCommit), common.PrettyDuration(s.BlockWrite),
 		s.StateReadCacheStats)
 	for _, line := range strings.Split(msg, "\n") {
 		if line == "" {

core/blockchain_test.go

@@ -156,7 +156,8 @@ func testBlockChainImport(chain types.Blocks, blockchain *BlockChain) error {
 			}
 			return err
 		}
-		statedb, err := state.New(blockchain.GetBlockByHash(block.ParentHash()).Root(), blockchain.statedb)
+		sdb := state.NewDatabase(blockchain.triedb, blockchain.codedb)
+		statedb, err := state.New(blockchain.GetBlockByHash(block.ParentHash()).Root(), sdb)
 		if err != nil {
 			return err
 		}

core/chain_makers.go

@@ -27,6 +27,7 @@ import (
 	"github.com/ethereum/go-ethereum/consensus/misc/eip4844"
 	"github.com/ethereum/go-ethereum/core/rawdb"
 	"github.com/ethereum/go-ethereum/core/state"
+	"github.com/ethereum/go-ethereum/core/state/codedb"
 	"github.com/ethereum/go-ethereum/core/types"
 	"github.com/ethereum/go-ethereum/core/vm"
 	"github.com/ethereum/go-ethereum/ethdb"
@@ -432,9 +433,10 @@ func GenerateChain(config *params.ChainConfig, parent *types.Block, engine conse
 	}
 	triedb := triedb.NewDatabase(db, triedbConfig)
 	defer triedb.Close()
+	codedb := codedb.New(db)
 
 	for i := 0; i < n; i++ {
-		statedb, err := state.New(parent.Root(), state.NewDatabase(triedb, nil))
+		statedb, err := state.New(parent.Root(), state.NewDatabase(triedb, codedb))
 		if err != nil {
 			panic(err)
 		}

core/genesis.go

@@ -29,6 +29,7 @@ import (
 	"github.com/ethereum/go-ethereum/common/math"
 	"github.com/ethereum/go-ethereum/core/rawdb"
 	"github.com/ethereum/go-ethereum/core/state"
+	"github.com/ethereum/go-ethereum/core/state/codedb"
 	"github.com/ethereum/go-ethereum/core/tracing"
 	"github.com/ethereum/go-ethereum/core/types"
 	"github.com/ethereum/go-ethereum/crypto"
@@ -145,7 +146,7 @@ func hashAlloc(ga *types.GenesisAlloc, isVerkle bool) (common.Hash, error) {
 		emptyRoot = types.EmptyVerkleHash
 	}
 	db := rawdb.NewMemoryDatabase()
-	statedb, err := state.New(emptyRoot, state.NewDatabase(triedb.NewDatabase(db, config), nil))
+	statedb, err := state.New(emptyRoot, state.NewDatabase(triedb.NewDatabase(db, config), codedb.New(db)))
 	if err != nil {
 		return common.Hash{}, err
 	}
@@ -164,12 +165,12 @@ func hashAlloc(ga *types.GenesisAlloc, isVerkle bool) (common.Hash, error) {
 
 // flushAlloc is very similar with hash, but the main difference is all the
 // generated states will be persisted into the given database.
-func flushAlloc(ga *types.GenesisAlloc, triedb *triedb.Database) (common.Hash, error) {
+func flushAlloc(ga *types.GenesisAlloc, db ethdb.KeyValueStore, triedb *triedb.Database) (common.Hash, error) {
 	emptyRoot := types.EmptyRootHash
 	if triedb.IsVerkle() {
 		emptyRoot = types.EmptyVerkleHash
 	}
-	statedb, err := state.New(emptyRoot, state.NewDatabase(triedb, nil))
+	statedb, err := state.New(emptyRoot, state.NewDatabase(triedb, codedb.New(db)))
 	if err != nil {
 		return common.Hash{}, err
 	}
@@ -552,7 +553,7 @@ func (g *Genesis) Commit(db ethdb.Database, triedb *triedb.Database) (*types.Blo
 		return nil, errors.New("can't start clique chain without signers")
 	}
 	// flush the data to disk and compute the state root
-	root, err := flushAlloc(&g.Alloc, triedb)
+	root, err := flushAlloc(&g.Alloc, db, triedb)
 	if err != nil {
 		return nil, err
 	}