eth/filters: avoid block body retrieval when no matching logs (ethereum#25199)

Author: JukLee0ira

accounts/abi/bind/backends/simulated.go

@@ -555,6 +555,13 @@ func (fb *filterBackend) GetReceipts(ctx context.Context, hash common.Hash) (typ
 	return core.GetBlockReceipts(fb.db, hash, core.GetBlockNumber(fb.db, hash)), nil
 }
 
+func (fb *filterBackend) GetBody(ctx context.Context, hash common.Hash, number rpc.BlockNumber) (*types.Body, error) {
+	if body := fb.bc.GetBody(hash); body != nil {
+		return body, nil
+	}
+	return nil, errors.New("block body not found")
+}
+
 func (fb *filterBackend) PendingBlockAndReceipts() (*types.Block, types.Receipts) {
 	return fb.backend.pendingBlock, fb.backend.pendingReceipts
 }

core/rawdb/accessors_chain.go

@@ -29,6 +29,22 @@ import (
 	"github.com/XinFinOrg/XDPoSChain/rlp"
 )
 
+// WriteCanonicalHash stores the hash assigned to a canonical block number.
+func WriteCanonicalHash(db ethdb.KeyValueWriter, hash common.Hash, number uint64) {
+	if err := db.Put(headerHashKey(number), hash.Bytes()); err != nil {
+		log.Crit("Failed to store number to hash mapping", "err", err)
+	}
+}
+
+// WriteHeaderNumber stores the hash->number mapping.
+func WriteHeaderNumber(db ethdb.KeyValueWriter, hash common.Hash, number uint64) {
+	key := headerNumberKey(hash)
+	enc := encodeBlockNumber(number)
+	if err := db.Put(key, enc); err != nil {
+		log.Crit("Failed to store hash to number mapping", "err", err)
+	}
+}
+
 // ReadHeaderNumber returns the header number assigned to a hash.
 func ReadHeaderNumber(db ethdb.KeyValueReader, hash common.Hash) *uint64 {
 	data, _ := db.Get(headerNumberKey(hash))
@@ -39,6 +55,13 @@ func ReadHeaderNumber(db ethdb.KeyValueReader, hash common.Hash) *uint64 {
 	return &number
 }
 
+// WriteHeadBlockHash stores the head block's hash.
+func WriteHeadBlockHash(db ethdb.KeyValueWriter, hash common.Hash) {
+	if err := db.Put(headBlockKey, hash.Bytes()); err != nil {
+		log.Crit("Failed to store last block's hash", "err", err)
+	}
+}
+
 // ReadBodyRLP retrieves the block body (transactions and uncles) in RLP encoding.
 func ReadBodyRLP(db ethdb.Reader, hash common.Hash, number uint64) rlp.RawValue {
 	// First try to look up the data in ancient database. Extra hash
@@ -100,6 +123,27 @@ func WriteBody(db ethdb.KeyValueWriter, hash common.Hash, number uint64, body *t
 	WriteBodyRLP(db, hash, number, data)
 }
 
+// WriteHeader stores a block header into the database and also stores the hash-
+// to-number mapping.
+func WriteHeader(db ethdb.KeyValueWriter, header *types.Header) {
+	var (
+		hash   = header.Hash()
+		number = header.Number.Uint64()
+	)
+	// Write the hash -> number mapping
+	WriteHeaderNumber(db, hash, number)
+
+	// Write the encoded header
+	data, err := rlp.EncodeToBytes(header)
+	if err != nil {
+		log.Crit("Failed to RLP encode header", "err", err)
+	}
+	key := headerKey(number, hash)
+	if err := db.Put(key, data); err != nil {
+		log.Crit("Failed to store header", "err", err)
+	}
+}
+
 // ReadReceiptsRLP retrieves all the transaction receipts belonging to a block in RLP encoding.
 func ReadReceiptsRLP(db ethdb.Reader, hash common.Hash, number uint64) rlp.RawValue {
 	// First try to look up the data in ancient database. Extra hash
@@ -195,6 +239,12 @@ func WriteReceipts(db ethdb.KeyValueWriter, hash common.Hash, number uint64, rec
 	}
 }
 
+// WriteBlock serializes a block into the database, header and body separately.
+func WriteBlock(db ethdb.KeyValueWriter, block *types.Block) {
+	WriteBody(db, block.Hash(), block.NumberU64(), block.Body())
+	WriteHeader(db, block.Header())
+}
+
 // storedReceiptRLP is the storage encoding of a receipt.
 // Re-definition in core/types/receipt.go.
 type storedReceiptRLP struct {
@@ -248,9 +298,9 @@ func deriveLogFields(receipts []*receiptLogs, hash common.Hash, number uint64, t
 	return nil
 }
 
-// ReadLogs retrieves the logs for all transactions in a block. The log fields
-// are populated with metadata. In case the receipts or the block body
-// are not found, a nil is returned.
+// ReadLogs retrieves the logs for all transactions in a block. In case
+// receipts is not found, a nil is returned.
+// Note: ReadLogs does not derive unstored log fields.
 func ReadLogs(db ethdb.Reader, hash common.Hash, number uint64) [][]*types.Log {
 	// Retrieve the flattened receipt slice
 	data := ReadReceiptsRLP(db, hash, number)
@@ -263,15 +313,6 @@ func ReadLogs(db ethdb.Reader, hash common.Hash, number uint64) [][]*types.Log {
 		return nil
 	}
 
-	body := ReadBody(db, hash, number)
-	if body == nil {
-		log.Error("Missing body but have receipt", "hash", hash, "number", number)
-		return nil
-	}
-	if err := deriveLogFields(receipts, hash, number, body.Transactions); err != nil {
-		log.Error("Failed to derive block receipts fields", "hash", hash, "number", number, "err", err)
-		return nil
-	}
 	logs := make([][]*types.Log, len(receipts))
 	for i, receipt := range receipts {
 		logs[i] = receipt.Logs

core/rawdb/accessors_chain_test.go

@@ -116,10 +116,6 @@ func TestReadLogs(t *testing.T) {
 		t.Fatalf("unexpected number of logs[1] returned, have %d want %d", have, want)
 	}
 
-	// Fill in log fields so we can compare their rlp encoding
-	if err := types.Receipts(receipts).DeriveFields(params.TestChainConfig, hash, 0, body.Transactions); err != nil {
-		t.Fatal(err)
-	}
 	for i, pr := range receipts {
 		for j, pl := range pr.Logs {
 			rlpHave, err := rlp.EncodeToBytes(newFullLogRLP(logs[i][j]))

core/rawdb/schema.go

@@ -25,10 +25,13 @@ import (
 
 // The fields below define the low level database schema prefixing.
 var (
+	headBlockKey = []byte("LastBlock")
 	// Data item prefixes (use single byte to avoid mixing data types, avoid `i`, used for indexes).
+	headerPrefix       = []byte("h") // headerPrefix + num (uint64 big endian) + hash -> header
+	headerHashSuffix   = []byte("n") // headerPrefix + num (uint64 big endian) + headerHashSuffix -> hash
 	headerNumberPrefix = []byte("H") // headerNumberPrefix + hash -> num (uint64 big endian)
-	blockBodyPrefix    = []byte("b") // blockBodyPrefix + num (uint64 big endian) + hash -> block body
 
+	blockBodyPrefix     = []byte("b") // blockBodyPrefix + num (uint64 big endian) + hash -> block body
 	blockReceiptsPrefix = []byte("r") // blockReceiptsPrefix + num (uint64 big endian) + hash -> block receipts
 )
 
@@ -50,6 +53,16 @@ func encodeBlockNumber(number uint64) []byte {
 	return enc
 }
 
+// headerKey = headerPrefix + num (uint64 big endian) + hash
+func headerKey(number uint64, hash common.Hash) []byte {
+	return append(append(headerPrefix, encodeBlockNumber(number)...), hash.Bytes()...)
+}
+
+// headerHashKey = headerPrefix + num (uint64 big endian) + headerHashSuffix
+func headerHashKey(number uint64) []byte {
+	return append(append(headerPrefix, encodeBlockNumber(number)...), headerHashSuffix...)
+}
+
 // headerNumberKey = headerNumberPrefix + hash
 func headerNumberKey(hash common.Hash) []byte {
 	return append(headerNumberPrefix, hash.Bytes()...)

eth/api_backend.go

@@ -159,6 +159,17 @@ func (b *EthApiBackend) BlockByHash(ctx context.Context, hash common.Hash) (*typ
 	return b.eth.blockchain.GetBlockByHash(hash), nil
 }
 
+// GetBody returns body of a block. It does not resolve special block numbers.
+func (b *EthApiBackend) GetBody(ctx context.Context, hash common.Hash, number rpc.BlockNumber) (*types.Body, error) {
+	if number < 0 || hash == (common.Hash{}) {
+		return nil, errors.New("invalid arguments; expect hash and no special block numbers")
+	}
+	if body := b.eth.blockchain.GetBody(hash); body != nil {
+		return body, nil
+	}
+	return nil, errors.New("block body not found")
+}
+
 func (b *EthApiBackend) BlockByNumberOrHash(ctx context.Context, blockNrOrHash rpc.BlockNumberOrHash) (*types.Block, error) {
 	if blockNr, ok := blockNrOrHash.Number(); ok {
 		return b.BlockByNumber(ctx, blockNr)

eth/filters/filter.go

@@ -104,7 +104,7 @@ func (f *Filter) Logs(ctx context.Context) ([]*types.Log, error) {
 		if header == nil {
 			return nil, errors.New("unknown block")
 		}
-		return f.blockLogs(ctx, header, false)
+		return f.blockLogs(ctx, header)
 	}
 	// Short-cut if all we care about is pending logs
 	if f.begin == rpc.PendingBlockNumber.Int64() {
@@ -192,7 +192,7 @@ func (f *Filter) indexedLogs(ctx context.Context, end uint64) ([]*types.Log, err
 			if header == nil || err != nil {
 				return logs, err
 			}
-			found, err := f.blockLogs(ctx, header, true)
+			found, err := f.blockLogs(ctx, header)
 			if err != nil {
 				return logs, err
 			}
@@ -214,7 +214,7 @@ func (f *Filter) unindexedLogs(ctx context.Context, end uint64) ([]*types.Log, e
 		if header == nil || err != nil {
 			return logs, err
 		}
-		found, err := f.blockLogs(ctx, header, false)
+		found, err := f.blockLogs(ctx, header)
 		if err != nil {
 			return logs, err
 		}
@@ -224,46 +224,46 @@ func (f *Filter) unindexedLogs(ctx context.Context, end uint64) ([]*types.Log, e
 }
 
 // blockLogs returns the logs matching the filter criteria within a single block.
-func (f *Filter) blockLogs(ctx context.Context, header *types.Header, skipBloom bool) ([]*types.Log, error) {
-	// Fast track: no filtering criteria
-	if len(f.addresses) == 0 && len(f.topics) == 0 {
-		list, err := f.sys.cachedGetLogs(ctx, header.Hash(), header.Number.Uint64())
-		if err != nil {
-			return nil, err
-		}
-		return flatten(list), nil
-	} else if skipBloom || bloomFilter(header.Bloom, f.addresses, f.topics) {
+func (f *Filter) blockLogs(ctx context.Context, header *types.Header) ([]*types.Log, error) {
+	if bloomFilter(header.Bloom, f.addresses, f.topics) {
 		return f.checkMatches(ctx, header)
 	}
 	return nil, nil
 }
 
 // checkMatches checks if the receipts belonging to the given header contain any log events that
 // match the filter criteria. This function is called when the bloom filter signals a potential match.
+// skipFilter signals all logs of the given block are requested.
 func (f *Filter) checkMatches(ctx context.Context, header *types.Header) ([]*types.Log, error) {
-	logsList, err := f.sys.cachedGetLogs(ctx, header.Hash(), header.Number.Uint64())
+	hash := header.Hash()
+	// Logs in cache are partially filled with context data
+	// such as tx index, block hash, etc.
+	// Notably tx hash is NOT filled in because it needs
+	// access to block body data.
+	cached, err := f.sys.cachedLogElem(ctx, hash, header.Number.Uint64())
 	if err != nil {
 		return nil, err
 	}
-
-	unfiltered := flatten(logsList)
-	logs := filterLogs(unfiltered, nil, nil, f.addresses, f.topics)
-	if len(logs) > 0 {
-		// We have matching logs, check if we need to resolve full logs via the light client
-		if logs[0].TxHash == (common.Hash{}) {
-			receipts, err := f.sys.backend.GetReceipts(ctx, header.Hash())
-			if err != nil {
-				return nil, err
-			}
-			unfiltered = unfiltered[:0]
-			for _, receipt := range receipts {
-				unfiltered = append(unfiltered, receipt.Logs...)
-			}
-			logs = filterLogs(unfiltered, nil, nil, f.addresses, f.topics)
-		}
+	logs := filterLogs(cached.logs, nil, nil, f.addresses, f.topics)
+	if len(logs) == 0 {
+		return nil, nil
+	}
+	// Most backends will deliver un-derived logs, but check nevertheless.
+	if len(logs) > 0 && logs[0].TxHash != (common.Hash{}) {
 		return logs, nil
 	}
-	return nil, nil
+
+	body, err := f.sys.cachedGetBody(ctx, cached, hash, header.Number.Uint64())
+	if err != nil {
+		return nil, err
+	}
+	for i, log := range logs {
+		// Copy log not to modify cache elements
+		logcopy := *log
+		logcopy.TxHash = body.Transactions[logcopy.TxIndex].Hash()
+		logs[i] = &logcopy
+	}
+	return logs, nil
 }
 
 // pendingLogs returns the logs matching the filter criteria within the pending block.
@@ -353,11 +353,3 @@ func bloomFilter(bloom types.Bloom, addresses []common.Address, topics [][]commo
 	}
 	return true
 }
-
-func flatten(list [][]*types.Log) []*types.Log {
-	var flat []*types.Log
-	for _, logs := range list {
-		flat = append(flat, logs...)
-	}
-	return flat
-}