core: speed up header import (#21967)

This PR implements the following modifications

- Don't shortcut check if block is present, thus avoid disk lookup
- Don't check hash ancestry in early-check (it's still done in parallel checker)
- Don't check time.Now for every single header

Charts and background info can be found here: https://github.com/holiman/headerimport/blob/main/README.md
With these changes, writing 1M headers goes down to from 80s to 62s.

consensus/ethash/consensus.go

@@ -39,11 +39,11 @@ import (
 
 // Ethash proof-of-work protocol constants.
 var (
-	FrontierBlockReward       = big.NewInt(5e+18) // Block reward in wei for successfully mining a block
+	FrontierBlockReward           = big.NewInt(5e+18) // Block reward in wei for successfully mining a block
-	ByzantiumBlockReward      = big.NewInt(3e+18) // Block reward in wei for successfully mining a block upward from Byzantium
+	ByzantiumBlockReward          = big.NewInt(3e+18) // Block reward in wei for successfully mining a block upward from Byzantium
-	ConstantinopleBlockReward = big.NewInt(2e+18) // Block reward in wei for successfully mining a block upward from Constantinople
+	ConstantinopleBlockReward     = big.NewInt(2e+18) // Block reward in wei for successfully mining a block upward from Constantinople
-	maxUncles                 = 2                 // Maximum number of uncles allowed in a single block
+	maxUncles                     = 2                 // Maximum number of uncles allowed in a single block
-	allowedFutureBlockTime    = 15 * time.Second  // Max time from current time allowed for blocks, before they're considered future blocks
+	allowedFutureBlockTimeSeconds = int64(15)         // Max seconds from current time allowed for blocks, before they're considered future blocks
 
 	// calcDifficultyEip2384 is the difficulty adjustment algorithm as specified by EIP 2384.
 	// It offsets the bomb 4M blocks from Constantinople, so in total 9M blocks.
@@ -102,7 +102,7 @@ func (ethash *Ethash) VerifyHeader(chain consensus.ChainHeaderReader, header *ty
 		return consensus.ErrUnknownAncestor
 	}
 	// Sanity checks passed, do a proper verification
-	return ethash.verifyHeader(chain, header, parent, false, seal)
+	return ethash.verifyHeader(chain, header, parent, false, seal, time.Now().Unix())
 }
 
 // VerifyHeaders is similar to VerifyHeader, but verifies a batch of headers
@@ -126,15 +126,16 @@ func (ethash *Ethash) VerifyHeaders(chain consensus.ChainHeaderReader, headers [
 
 	// Create a task channel and spawn the verifiers
 	var (
-		inputs = make(chan int)
+		inputs  = make(chan int)
-		done   = make(chan int, workers)
+		done    = make(chan int, workers)
-		errors = make([]error, len(headers))
+		errors  = make([]error, len(headers))
-		abort  = make(chan struct{})
+		abort   = make(chan struct{})
+		unixNow = time.Now().Unix()
 	)
 	for i := 0; i < workers; i++ {
 		go func() {
 			for index := range inputs {
-				errors[index] = ethash.verifyHeaderWorker(chain, headers, seals, index)
+				errors[index] = ethash.verifyHeaderWorker(chain, headers, seals, index, unixNow)
 				done <- index
 			}
 		}()
@@ -170,7 +171,7 @@ func (ethash *Ethash) VerifyHeaders(chain consensus.ChainHeaderReader, headers [
 	return abort, errorsOut
 }
 
-func (ethash *Ethash) verifyHeaderWorker(chain consensus.ChainHeaderReader, headers []*types.Header, seals []bool, index int) error {
+func (ethash *Ethash) verifyHeaderWorker(chain consensus.ChainHeaderReader, headers []*types.Header, seals []bool, index int, unixNow int64) error {
 	var parent *types.Header
 	if index == 0 {
 		parent = chain.GetHeader(headers[0].ParentHash, headers[0].Number.Uint64()-1)
@@ -180,10 +181,7 @@ func (ethash *Ethash) verifyHeaderWorker(chain consensus.ChainHeaderReader, head
 	if parent == nil {
 		return consensus.ErrUnknownAncestor
 	}
-	if chain.GetHeader(headers[index].Hash(), headers[index].Number.Uint64()) != nil {
+	return ethash.verifyHeader(chain, headers[index], parent, false, seals[index], unixNow)
-		return nil // known block
-	}
-	return ethash.verifyHeader(chain, headers[index], parent, false, seals[index])
 }
 
 // VerifyUncles verifies that the given block's uncles conform to the consensus
@@ -234,7 +232,7 @@ func (ethash *Ethash) VerifyUncles(chain consensus.ChainReader, block *types.Blo
 		if ancestors[uncle.ParentHash] == nil || uncle.ParentHash == block.ParentHash() {
 			return errDanglingUncle
 		}
-		if err := ethash.verifyHeader(chain, uncle, ancestors[uncle.ParentHash], true, true); err != nil {
+		if err := ethash.verifyHeader(chain, uncle, ancestors[uncle.ParentHash], true, true, time.Now().Unix()); err != nil {
 			return err
 		}
 	}
@@ -244,14 +242,14 @@ func (ethash *Ethash) VerifyUncles(chain consensus.ChainReader, block *types.Blo
 // verifyHeader checks whether a header conforms to the consensus rules of the
 // stock Ethereum ethash engine.
 // See YP section 4.3.4. "Block Header Validity"
-func (ethash *Ethash) verifyHeader(chain consensus.ChainHeaderReader, header, parent *types.Header, uncle bool, seal bool) error {
+func (ethash *Ethash) verifyHeader(chain consensus.ChainHeaderReader, header, parent *types.Header, uncle bool, seal bool, unixNow int64) error {
 	// Ensure that the header's extra-data section is of a reasonable size
 	if uint64(len(header.Extra)) > params.MaximumExtraDataSize {
 		return fmt.Errorf("extra-data too long: %d > %d", len(header.Extra), params.MaximumExtraDataSize)
 	}
 	// Verify the header's timestamp
 	if !uncle {
-		if header.Time > uint64(time.Now().Add(allowedFutureBlockTime).Unix()) {
+		if header.Time > uint64(unixNow+allowedFutureBlockTimeSeconds) {
 			return consensus.ErrFutureBlock
 		}
 	}

core/headerchain.go

@@ -299,17 +299,18 @@ func (hc *HeaderChain) writeHeaders(headers []*types.Header) (result *headerWrit
 func (hc *HeaderChain) ValidateHeaderChain(chain []*types.Header, checkFreq int) (int, error) {
 	// Do a sanity check that the provided chain is actually ordered and linked
 	for i := 1; i < len(chain); i++ {
-		parentHash := chain[i-1].Hash()
+		if chain[i].Number.Uint64() != chain[i-1].Number.Uint64()+1 {
-		if chain[i].Number.Uint64() != chain[i-1].Number.Uint64()+1 || chain[i].ParentHash != parentHash {
+			hash := chain[i].Hash()
+			parentHash := chain[i-1].Hash()
 			// Chain broke ancestry, log a message (programming error) and skip insertion
-			log.Error("Non contiguous header insert", "number", chain[i].Number, "hash", chain[i].Hash(),
+			log.Error("Non contiguous header insert", "number", chain[i].Number, "hash", hash,
 				"parent", chain[i].ParentHash, "prevnumber", chain[i-1].Number, "prevhash", parentHash)
 
 			return 0, fmt.Errorf("non contiguous insert: item %d is #%d [%x…], item %d is #%d [%x…] (parent [%x…])", i-1, chain[i-1].Number,
-				parentHash.Bytes()[:4], i, chain[i].Number, chain[i].Hash().Bytes()[:4], chain[i].ParentHash[:4])
+				parentHash.Bytes()[:4], i, chain[i].Number, hash.Bytes()[:4], chain[i].ParentHash[:4])
 		}
 		// If the header is a banned one, straight out abort
-		if BadHashes[parentHash] {
+		if BadHashes[chain[i].ParentHash] {
 			return i - 1, ErrBlacklistedHash
 		}
 		// If it's the last header in the cunk, we need to check it too

core/rawdb/freezer_table.go

@@ -103,6 +103,11 @@ type freezerTable struct {
 	lock   sync.RWMutex // Mutex protecting the data file descriptors
 }
 
+// NewFreezerTable opens the given path as a freezer table.
+func NewFreezerTable(path, name string, disableSnappy bool) (*freezerTable, error) {
+	return newTable(path, name, metrics.NilMeter{}, metrics.NilMeter{}, metrics.NilGauge{}, disableSnappy)
+}
+
 // newTable opens a freezer table with default settings - 2G files
 func newTable(path string, name string, readMeter metrics.Meter, writeMeter metrics.Meter, sizeGauge metrics.Gauge, disableSnappy bool) (*freezerTable, error) {
 	return newCustomTable(path, name, readMeter, writeMeter, sizeGauge, 2*1000*1000*1000, disableSnappy)