core/rawdb, trie: improve db APIs for accessing trie nodes (#29362)

* core/rawdb, trie: improve db APIs for accessing trie nodes

* triedb/pathdb: fix

Author: rjl493456442

cmd/devp2p/internal/ethtest/snap.go

@@ -32,7 +32,6 @@ import (
 	"github.com/ethereum/go-ethereum/internal/utesting"
 	"github.com/ethereum/go-ethereum/trie"
 	"github.com/ethereum/go-ethereum/trie/trienode"
-	"golang.org/x/crypto/sha3"
 )
 
 func (c *Conn) snapRequest(code uint64, msg any) (any, error) {
@@ -905,7 +904,7 @@ func (s *Suite) snapGetByteCodes(t *utesting.T, tc *byteCodesTest) error {
 	// that the serving node is missing
 	var (
 		bytecodes = res.Codes
-		hasher    = sha3.NewLegacyKeccak256().(crypto.KeccakState)
+		hasher    = crypto.NewKeccakState()
 		hash      = make([]byte, 32)
 		codes     = make([][]byte, len(req.Hashes))
 	)
@@ -964,7 +963,7 @@ func (s *Suite) snapGetTrieNodes(t *utesting.T, tc *trieNodesTest) error {
 
 	// Cross reference the requested trienodes with the response to find gaps
 	// that the serving node is missing
-	hasher := sha3.NewLegacyKeccak256().(crypto.KeccakState)
+	hasher := crypto.NewKeccakState()
 	hash := make([]byte, 32)
 	trienodes := res.Nodes
 	if got, want := len(trienodes), len(tc.expHashes); got != want {

cmd/geth/dbcmd.go

@@ -246,11 +246,17 @@ func removeDB(ctx *cli.Context) error {
 		ancientDir = config.Node.ResolvePath(ancientDir)
 	}
 	// Delete state data
-	statePaths := []string{rootDir, filepath.Join(ancientDir, rawdb.StateFreezerName)}
+	statePaths := []string{
+		rootDir,
+		filepath.Join(ancientDir, rawdb.StateFreezerName),
+	}
 	confirmAndRemoveDB(statePaths, "state data", ctx, removeStateDataFlag.Name)
 
 	// Delete ancient chain
-	chainPaths := []string{filepath.Join(ancientDir, rawdb.ChainFreezerName)}
+	chainPaths := []string{filepath.Join(
+		ancientDir,
+		rawdb.ChainFreezerName,
+	)}
 	confirmAndRemoveDB(chainPaths, "ancient chain", ctx, removeChainDataFlag.Name)
 	return nil
 }

core/genesis_test.go

@@ -322,7 +322,7 @@ func TestVerkleGenesisCommit(t *testing.T) {
 		t.Fatalf("expected trie to be verkle")
 	}
 
-	if !rawdb.ExistsAccountTrieNode(db, nil) {
+	if !rawdb.HasAccountTrieNode(db, nil) {
 		t.Fatal("could not find node")
 	}
 }

core/rawdb/accessors_trie.go

@@ -24,7 +24,6 @@ import (
 	"github.com/ethereum/go-ethereum/crypto"
 	"github.com/ethereum/go-ethereum/ethdb"
 	"github.com/ethereum/go-ethereum/log"
-	"golang.org/x/crypto/sha3"
 )
 
 // HashScheme is the legacy hash-based state scheme with which trie nodes are
@@ -50,7 +49,7 @@ const PathScheme = "path"
 type hasher struct{ sha crypto.KeccakState }
 
 var hasherPool = sync.Pool{
-	New: func() interface{} { return &hasher{sha: sha3.NewLegacyKeccak256().(crypto.KeccakState)} },
+	New: func() interface{} { return &hasher{sha: crypto.NewKeccakState()} },
 }
 
 func newHasher() *hasher {
@@ -65,33 +64,15 @@ func (h *hasher) release() {
 	hasherPool.Put(h)
 }
 
-// ReadAccountTrieNode retrieves the account trie node and the associated node
-// hash with the specified node path.
-func ReadAccountTrieNode(db ethdb.KeyValueReader, path []byte) ([]byte, common.Hash) {
-	data, err := db.Get(accountTrieNodeKey(path))
-	if err != nil {
-		return nil, common.Hash{}
-	}
-	h := newHasher()
-	defer h.release()
-	return data, h.hash(data)
-}
-
-// HasAccountTrieNode checks the account trie node presence with the specified
-// node path and the associated node hash.
-func HasAccountTrieNode(db ethdb.KeyValueReader, path []byte, hash common.Hash) bool {
-	data, err := db.Get(accountTrieNodeKey(path))
-	if err != nil {
-		return false
-	}
-	h := newHasher()
-	defer h.release()
-	return h.hash(data) == hash
+// ReadAccountTrieNode retrieves the account trie node with the specified node path.
+func ReadAccountTrieNode(db ethdb.KeyValueReader, path []byte) []byte {
+	data, _ := db.Get(accountTrieNodeKey(path))
+	return data
 }
 
-// ExistsAccountTrieNode checks the presence of the account trie node with the
+// HasAccountTrieNode checks the presence of the account trie node with the
 // specified node path, regardless of the node hash.
-func ExistsAccountTrieNode(db ethdb.KeyValueReader, path []byte) bool {
+func HasAccountTrieNode(db ethdb.KeyValueReader, path []byte) bool {
 	has, err := db.Has(accountTrieNodeKey(path))
 	if err != nil {
 		return false
@@ -113,33 +94,15 @@ func DeleteAccountTrieNode(db ethdb.KeyValueWriter, path []byte) {
 	}
 }
 
-// ReadStorageTrieNode retrieves the storage trie node and the associated node
-// hash with the specified node path.
-func ReadStorageTrieNode(db ethdb.KeyValueReader, accountHash common.Hash, path []byte) ([]byte, common.Hash) {
-	data, err := db.Get(storageTrieNodeKey(accountHash, path))
-	if err != nil {
-		return nil, common.Hash{}
-	}
-	h := newHasher()
-	defer h.release()
-	return data, h.hash(data)
-}
-
-// HasStorageTrieNode checks the storage trie node presence with the provided
-// node path and the associated node hash.
-func HasStorageTrieNode(db ethdb.KeyValueReader, accountHash common.Hash, path []byte, hash common.Hash) bool {
-	data, err := db.Get(storageTrieNodeKey(accountHash, path))
-	if err != nil {
-		return false
-	}
-	h := newHasher()
-	defer h.release()
-	return h.hash(data) == hash
+// ReadStorageTrieNode retrieves the storage trie node with the specified node path.
+func ReadStorageTrieNode(db ethdb.KeyValueReader, accountHash common.Hash, path []byte) []byte {
+	data, _ := db.Get(storageTrieNodeKey(accountHash, path))
+	return data
 }
 
-// ExistsStorageTrieNode checks the presence of the storage trie node with the
+// HasStorageTrieNode checks the presence of the storage trie node with the
 // specified account hash and node path, regardless of the node hash.
-func ExistsStorageTrieNode(db ethdb.KeyValueReader, accountHash common.Hash, path []byte) bool {
+func HasStorageTrieNode(db ethdb.KeyValueReader, accountHash common.Hash, path []byte) bool {
 	has, err := db.Has(storageTrieNodeKey(accountHash, path))
 	if err != nil {
 		return false
@@ -198,54 +161,54 @@ func HasTrieNode(db ethdb.KeyValueReader, owner common.Hash, path []byte, hash c
 	case HashScheme:
 		return HasLegacyTrieNode(db, hash)
 	case PathScheme:
+		var blob []byte
 		if owner == (common.Hash{}) {
-			return HasAccountTrieNode(db, path, hash)
+			blob = ReadAccountTrieNode(db, path)
+		} else {
+			blob = ReadStorageTrieNode(db, owner, path)
 		}
-		return HasStorageTrieNode(db, owner, path, hash)
+		if len(blob) == 0 {
+			return false
+		}
+		h := newHasher()
+		defer h.release()
+		return h.hash(blob) == hash // exists but not match
 	default:
 		panic(fmt.Sprintf("Unknown scheme %v", scheme))
 	}
 }
 
 // ReadTrieNode retrieves the trie node from database with the provided node info
 // and associated node hash.
-// hashScheme-based lookup requires the following:
-//   - hash
-//
-// pathScheme-based lookup requires the following:
-//   - owner
-//   - path
 func ReadTrieNode(db ethdb.KeyValueReader, owner common.Hash, path []byte, hash common.Hash, scheme string) []byte {
 	switch scheme {
 	case HashScheme:
 		return ReadLegacyTrieNode(db, hash)
 	case PathScheme:
-		var (
-			blob  []byte
-			nHash common.Hash
-		)
+		var blob []byte
 		if owner == (common.Hash{}) {
-			blob, nHash = ReadAccountTrieNode(db, path)
+			blob = ReadAccountTrieNode(db, path)
 		} else {
-			blob, nHash = ReadStorageTrieNode(db, owner, path)
+			blob = ReadStorageTrieNode(db, owner, path)
 		}
-		if nHash != hash {
+		if len(blob) == 0 {
 			return nil
 		}
+		h := newHasher()
+		defer h.release()
+		if h.hash(blob) != hash {
+			return nil // exists but not match
+		}
 		return blob
 	default:
 		panic(fmt.Sprintf("Unknown scheme %v", scheme))
 	}
 }
 
-// WriteTrieNode writes the trie node into database with the provided node info
-// and associated node hash.
-// hashScheme-based lookup requires the following:
-//   - hash
+// WriteTrieNode writes the trie node into database with the provided node info.
 //
-// pathScheme-based lookup requires the following:
-//   - owner
-//   - path
+// hash-scheme requires the node hash as the identifier.
+// path-scheme requires the node owner and path as the identifier.
 func WriteTrieNode(db ethdb.KeyValueWriter, owner common.Hash, path []byte, hash common.Hash, node []byte, scheme string) {
 	switch scheme {
 	case HashScheme:
@@ -261,14 +224,10 @@ func WriteTrieNode(db ethdb.KeyValueWriter, owner common.Hash, path []byte, hash
 	}
 }
 
-// DeleteTrieNode deletes the trie node from database with the provided node info
-// and associated node hash.
-// hashScheme-based lookup requires the following:
-//   - hash
+// DeleteTrieNode deletes the trie node from database with the provided node info.
 //
-// pathScheme-based lookup requires the following:
-//   - owner
-//   - path
+// hash-scheme requires the node hash as the identifier.
+// path-scheme requires the node owner and path as the identifier.
 func DeleteTrieNode(db ethdb.KeyValueWriter, owner common.Hash, path []byte, hash common.Hash, scheme string) {
 	switch scheme {
 	case HashScheme:
@@ -287,9 +246,8 @@ func DeleteTrieNode(db ethdb.KeyValueWriter, owner common.Hash, path []byte, has
 // ReadStateScheme reads the state scheme of persistent state, or none
 // if the state is not present in database.
 func ReadStateScheme(db ethdb.Reader) string {
-	// Check if state in path-based scheme is present
-	blob, _ := ReadAccountTrieNode(db, nil)
-	if len(blob) != 0 {
+	// Check if state in path-based scheme is present.
+	if HasAccountTrieNode(db, nil) {
 		return PathScheme
 	}
 	// The root node might be deleted during the initial snap sync, check
@@ -304,8 +262,7 @@ func ReadStateScheme(db ethdb.Reader) string {
 	if header == nil {
 		return "" // empty datadir
 	}
-	blob = ReadLegacyTrieNode(db, header.Root)
-	if len(blob) == 0 {
+	if !HasLegacyTrieNode(db, header.Root) {
 		return "" // no state in disk
 	}
 	return HashScheme

core/rawdb/ancient_utils.go

@@ -89,20 +89,17 @@ func inspectFreezers(db ethdb.Database) ([]freezerInfo, error) {
 			infos = append(infos, info)
 
 		case StateFreezerName:
-			if ReadStateScheme(db) != PathScheme {
-				continue
-			}
 			datadir, err := db.AncientDatadir()
 			if err != nil {
 				return nil, err
 			}
 			f, err := NewStateFreezer(datadir, true)
 			if err != nil {
-				return nil, err
+				continue // might be possible the state freezer is not existent
 			}
 			defer f.Close()
 
-			info, err := inspect(StateFreezerName, stateFreezerNoSnappy, f)
+			info, err := inspect(freezer, stateFreezerNoSnappy, f)
 			if err != nil {
 				return nil, err
 			}

eth/handler.go

@@ -42,7 +42,6 @@ import (
 	"github.com/ethereum/go-ethereum/p2p"
 	"github.com/ethereum/go-ethereum/p2p/enode"
 	"github.com/ethereum/go-ethereum/triedb/pathdb"
-	"golang.org/x/crypto/sha3"
 )
 
 const (
@@ -480,7 +479,7 @@ func (h *handler) BroadcastTransactions(txs types.Transactions) {
 
 	var (
 		signer = types.LatestSignerForChainID(h.chain.Config().ChainID) // Don't care about chain status, we just need *a* sender
-		hasher = sha3.NewLegacyKeccak256().(crypto.KeccakState)
+		hasher = crypto.NewKeccakState()
 		hash   = make([]byte, 32)
 	)
 	for _, tx := range txs {

eth/protocols/snap/gentrie.go

@@ -164,7 +164,7 @@ func (t *pathTrie) deleteAccountNode(path []byte, inner bool) {
 	} else {
 		accountOuterLookupGauge.Inc(1)
 	}
-	if !rawdb.ExistsAccountTrieNode(t.db, path) {
+	if !rawdb.HasAccountTrieNode(t.db, path) {
 		return
 	}
 	if inner {
@@ -181,7 +181,7 @@ func (t *pathTrie) deleteStorageNode(path []byte, inner bool) {
 	} else {
 		storageOuterLookupGauge.Inc(1)
 	}
-	if !rawdb.ExistsStorageTrieNode(t.db, t.owner, path) {
+	if !rawdb.HasStorageTrieNode(t.db, t.owner, path) {
 		return
 	}
 	if inner {

eth/protocols/snap/sync.go

@@ -42,7 +42,6 @@ import (
 	"github.com/ethereum/go-ethereum/rlp"
 	"github.com/ethereum/go-ethereum/trie"
 	"github.com/ethereum/go-ethereum/trie/trienode"
-	"golang.org/x/crypto/sha3"
 )
 
 const (
@@ -2653,7 +2652,7 @@ func (s *Syncer) onByteCodes(peer SyncPeer, id uint64, bytecodes [][]byte) error
 
 	// Cross reference the requested bytecodes with the response to find gaps
 	// that the serving node is missing
-	hasher := sha3.NewLegacyKeccak256().(crypto.KeccakState)
+	hasher := crypto.NewKeccakState()
 	hash := make([]byte, 32)
 
 	codes := make([][]byte, len(req.hashes))
@@ -2901,7 +2900,7 @@ func (s *Syncer) OnTrieNodes(peer SyncPeer, id uint64, trienodes [][]byte) error
 	// Cross reference the requested trienodes with the response to find gaps
 	// that the serving node is missing
 	var (
-		hasher = sha3.NewLegacyKeccak256().(crypto.KeccakState)
+		hasher = crypto.NewKeccakState()
 		hash   = make([]byte, 32)
 		nodes  = make([][]byte, len(req.hashes))
 		fills  uint64
@@ -3007,7 +3006,7 @@ func (s *Syncer) onHealByteCodes(peer SyncPeer, id uint64, bytecodes [][]byte) e
 
 	// Cross reference the requested bytecodes with the response to find gaps
 	// that the serving node is missing
-	hasher := sha3.NewLegacyKeccak256().(crypto.KeccakState)
+	hasher := crypto.NewKeccakState()
 	hash := make([]byte, 32)
 
 	codes := make([][]byte, len(req.hashes))

eth/protocols/snap/sync_test.go

@@ -64,7 +64,7 @@ func TestHashing(t *testing.T) {
 		}
 	}
 	var new = func() {
-		hasher := sha3.NewLegacyKeccak256().(crypto.KeccakState)
+		hasher := crypto.NewKeccakState()
 		var hash = make([]byte, 32)
 		for i := 0; i < len(bytecodes); i++ {
 			hasher.Reset()
@@ -96,7 +96,7 @@ func BenchmarkHashing(b *testing.B) {
 		}
 	}
 	var new = func() {
-		hasher := sha3.NewLegacyKeccak256().(crypto.KeccakState)
+		hasher := crypto.NewKeccakState()
 		var hash = make([]byte, 32)
 		for i := 0; i < len(bytecodes); i++ {
 			hasher.Reset()

trie/hasher.go

@@ -21,7 +21,6 @@ import (
 
 	"github.com/ethereum/go-ethereum/crypto"
 	"github.com/ethereum/go-ethereum/rlp"
-	"golang.org/x/crypto/sha3"
 )
 
 // hasher is a type used for the trie Hash operation. A hasher has some
@@ -38,7 +37,7 @@ var hasherPool = sync.Pool{
 	New: func() interface{} {
 		return &hasher{
 			tmp:    make([]byte, 0, 550), // cap is as large as a full fullNode.
-			sha:    sha3.NewLegacyKeccak256().(crypto.KeccakState),
+			sha:    crypto.NewKeccakState(),
 			encbuf: rlp.NewEncoderBuffer(nil),
 		}
 	},