core, tests: track dirty objects in separate map

core/blockchain.go

@@ -197,7 +197,15 @@ func (self *BlockChain) loadLastState() error {
 			self.currentFastBlock = block
 		}
 	}
-	// Issue a status log and return
+	// Initialize a statedb cache to ensure singleton account bloom filter generation
+	statedb, err := state.New(self.currentBlock.Root(), self.chainDb)
+	if err != nil {
+		return err
+	}
+	self.stateCache = statedb
+	self.stateCache.GetAccount(common.Address{})
+
+	// Issue a status log for the user
 	headerTd := self.GetTd(currentHeader.Hash(), currentHeader.Number.Uint64())
 	blockTd := self.GetTd(self.currentBlock.Hash(), self.currentBlock.NumberU64())
 	fastTd := self.GetTd(self.currentFastBlock.Hash(), self.currentFastBlock.NumberU64())
@@ -894,8 +902,6 @@ func (self *BlockChain) InsertChain(chain types.Blocks) (int, error) {
 		// Create a new statedb using the parent block and report an
 		// error if it fails.
 		switch {
-		case self.stateCache == nil:
-			self.stateCache, err = state.New(self.GetBlock(block.ParentHash(), block.NumberU64()-1).Root(), self.chainDb)
 		case i == 0:
 			err = self.stateCache.Reset(self.GetBlock(block.ParentHash(), block.NumberU64()-1).Root())
 		default:

core/state/dump.go

@@ -52,7 +52,7 @@ func (self *StateDB) RawDump() Dump {
 			panic(err)
 		}
 
-		obj := NewObject(common.BytesToAddress(addr), data)
+		obj := NewObject(common.BytesToAddress(addr), data, nil)
 		account := DumpAccount{
 			Balance:  data.Balance.String(),
 			Nonce:    data.Nonce,

core/state/iterator_test.go

@@ -50,7 +50,7 @@ func TestNodeIteratorCoverage(t *testing.T) {
 		if bytes.HasPrefix(key, []byte("secure-key-")) {
 			continue
 		}
-		if bytes.HasPrefix(key, []byte("accounts-bloom")) {
+		if bytes.Equal(key, []byte("accounts-bloom")) {
 			continue
 		}
 		if _, ok := hashes[common.BytesToHash(key)]; !ok {

core/state/state_object.go

@@ -82,10 +82,10 @@ type StateObject struct {
 	// Cache flags.
 	// When an object is marked for deletion it will be delete from the trie
 	// during the "update" phase of the state transition
-	dirty     bool // true if anything has changed
 	dirtyCode bool // true if the code was updated
 	remove    bool
 	deleted   bool
+	onDirty   func(addr common.Address) // Callback method to mark a state object newly dirty
 }
 
 // Account is the Ethereum consensus representation of accounts.
@@ -100,14 +100,14 @@ type Account struct {
 }
 
 // NewObject creates a state object.
-func NewObject(address common.Address, data Account) *StateObject {
+func NewObject(address common.Address, data Account, onDirty func(addr common.Address)) *StateObject {
 	if data.Balance == nil {
 		data.Balance = new(big.Int)
 	}
 	if data.CodeHash == nil {
 		data.CodeHash = emptyCodeHash
 	}
-	return &StateObject{address: address, data: data, storage: make(Storage)}
+	return &StateObject{address: address, data: data, storage: make(Storage), onDirty: onDirty}
 }
 
 // EncodeRLP implements rlp.Encoder.
@@ -124,8 +124,10 @@ func (self *StateObject) setError(err error) {
 
 func (self *StateObject) MarkForDeletion() {
 	self.remove = true
-	self.dirty = true
+	if self.onDirty != nil {
-
+		self.onDirty(self.Address())
+		self.onDirty = nil
+	}
 	if glog.V(logger.Core) {
 		glog.Infof("%x: #%d %v X\n", self.Address(), self.Nonce(), self.Balance())
 	}
@@ -162,7 +164,10 @@ func (self *StateObject) GetState(db trie.Database, key common.Hash) common.Hash
 // SetState updates a value in account storage.
 func (self *StateObject) SetState(key, value common.Hash) {
 	self.storage[key] = value
-	self.dirty = true
+	if self.onDirty != nil {
+		self.onDirty(self.Address())
+		self.onDirty = nil
+	}
 }
 
 // updateTrie writes cached storage modifications into the object's storage trie.
@@ -222,22 +227,23 @@ func (c *StateObject) SubBalance(amount *big.Int) {
 	}
 }
 
-func (c *StateObject) SetBalance(amount *big.Int) {
+func (self *StateObject) SetBalance(amount *big.Int) {
-	c.data.Balance = amount
+	self.data.Balance = amount
-	c.dirty = true
+	if self.onDirty != nil {
+		self.onDirty(self.Address())
+		self.onDirty = nil
+	}
 }
 
 // Return the gas back to the origin. Used by the Virtual machine or Closures
 func (c *StateObject) ReturnGas(gas, price *big.Int) {}
 
-func (self *StateObject) Copy(db trie.Database) *StateObject {
+func (self *StateObject) Copy(db trie.Database, onDirty func(addr common.Address)) *StateObject {
-	stateObject := NewObject(self.address, self.data)
+	stateObject := NewObject(self.address, self.data, onDirty)
-	stateObject.data.Balance.Set(self.data.Balance)
 	stateObject.trie = self.trie
 	stateObject.code = self.code
 	stateObject.storage = self.storage.Copy()
 	stateObject.remove = self.remove
-	stateObject.dirty = self.dirty
 	stateObject.dirtyCode = self.dirtyCode
 	stateObject.deleted = self.deleted
 	return stateObject
@@ -282,12 +288,19 @@ func (self *StateObject) SetCode(code []byte) {
 	self.data.CodeHash = crypto.Keccak256(code)
 	self.data.codeSize = new(int)
 	*self.data.codeSize = len(code)
-	self.dirty, self.dirtyCode = true, true
+	self.dirtyCode = true
+	if self.onDirty != nil {
+		self.onDirty(self.Address())
+		self.onDirty = nil
+	}
 }
 
 func (self *StateObject) SetNonce(nonce uint64) {
 	self.data.Nonce = nonce
-	self.dirty = true
+	if self.onDirty != nil {
+		self.onDirty(self.Address())
+		self.onDirty = nil
+	}
 }
 
 func (self *StateObject) CodeHash() []byte {

core/state/state_test.go

@@ -151,9 +151,10 @@ func TestSnapshot2(t *testing.T) {
 	so0.SetCode([]byte{'c', 'a', 'f', 'e'})
 	so0.remove = false
 	so0.deleted = false
-	so0.dirty = true
 	state.SetStateObject(so0)
-	state.Commit()
+
+	root, _ := state.Commit()
+	state.Reset(root)
 
 	// and one with deleted == true
 	so1 := state.GetStateObject(stateobjaddr1)
@@ -162,7 +163,6 @@ func TestSnapshot2(t *testing.T) {
 	so1.SetCode([]byte{'c', 'a', 'f', 'e', '2'})
 	so1.remove = true
 	so1.deleted = true
-	so1.dirty = true
 	state.SetStateObject(so1)
 
 	so1 = state.GetStateObject(stateobjaddr1)
@@ -183,7 +183,7 @@ func TestSnapshot2(t *testing.T) {
 	// deleted should be nil, both before and after restore of state copy
 	so1Restored := state.GetStateObject(stateobjaddr1)
 	if so1Restored != nil {
-		t.Fatalf("deleted object not nil after restoring snapshot")
+		t.Fatalf("deleted object not nil after restoring snapshot: %+v", so1Restored)
 	}
 }
 
@@ -227,7 +227,4 @@ func compareStateObjects(so0, so1 *StateObject, t *testing.T) {
 	if so0.deleted != so1.deleted {
 		t.Fatalf("Deleted mismatch: have %v, want %v", so0.deleted, so1.deleted)
 	}
-	if so0.dirty != so1.dirty {
-		t.Fatalf("Dirty mismatch: have %v, want %v", so0.dirty, so1.dirty)
-	}
 }

core/state/statedb.go

@@ -54,9 +54,9 @@ type StateDB struct {
 	accountBloom      *boom.ScalableBloomFilter
 	accountBloomDirty bool
 
-	// This map holds 'live' objects, which will get modified
+	// This map holds 'live' objects, which will get modified while processing a state transition.
-	// while processing a state transition.
 	stateObjects      map[common.Address]*StateObject
+	stateObjectsDirty map[common.Address]struct{}
 
 	// The refund counter, also used by state transitioning.
 	refund *big.Int
@@ -78,6 +78,7 @@ func New(root common.Hash, db ethdb.Database) (*StateDB, error) {
 		trie:              tr,
 		all:               make(map[common.Address]Account),
 		stateObjects:      make(map[common.Address]*StateObject),
+		stateObjectsDirty: make(map[common.Address]struct{}),
 		refund:            new(big.Int),
 		logs:              make(map[common.Hash]vm.Logs),
 	}, nil
@@ -104,6 +105,7 @@ func (self *StateDB) Reset(root common.Hash) error {
 		all:               all,
 		accountBloom:      self.accountBloom,
 		stateObjects:      make(map[common.Address]*StateObject),
+		stateObjectsDirty: make(map[common.Address]struct{}),
 		refund:            new(big.Int),
 		logs:              make(map[common.Hash]vm.Logs),
 	}
@@ -267,7 +269,6 @@ func (self *StateDB) DeleteStateObject(stateObject *StateObject) {
 
 	addr := stateObject.Address()
 	self.trie.Delete(addr[:])
-	//delete(self.stateObjects, addr)
 }
 
 // Retrieve a state object given my the address. Returns nil if not found.
@@ -281,7 +282,7 @@ func (self *StateDB) GetStateObject(addr common.Address) (stateObject *StateObje
 	}
 	// Use cached account data from the canon state if possible.
 	if data, ok := self.all[addr]; ok {
-		obj := NewObject(addr, data)
+		obj := NewObject(addr, data, self.MarkStateObjectDirty)
 		self.SetStateObject(obj)
 		return obj
 	}
@@ -304,7 +305,7 @@ func (self *StateDB) GetStateObject(addr common.Address) (stateObject *StateObje
 	// The object we just loaded has no storage trie and code yet.
 	self.all[addr] = data
 	// Insert into the live set.
-	obj := NewObject(addr, data)
+	obj := NewObject(addr, data, self.MarkStateObjectDirty)
 	self.SetStateObject(obj)
 	return obj
 }
@@ -328,9 +329,8 @@ func (self *StateDB) newStateObject(addr common.Address) *StateObject {
 	if glog.V(logger.Core) {
 		glog.Infof("(+) %x\n", addr)
 	}
-	obj := NewObject(addr, Account{})
+	obj := NewObject(addr, Account{}, self.MarkStateObjectDirty)
-	obj.dirty = true
+	obj.SetNonce(StartingNonce) // sets the object to dirty
-	obj.SetNonce(StartingNonce)
 	self.stateObjects[addr] = obj
 	if !self.getAccountBloom().TestAndAdd(self.trie.HashKey(addr[:])) {
 		self.accountBloomDirty = true
@@ -338,6 +338,12 @@ func (self *StateDB) newStateObject(addr common.Address) *StateObject {
 	return obj
 }
 
+// MarkStateObjectDirty adds the specified object to the dirty map to avoid costly
+// state object cache iteration to find a handful of modified ones.
+func (self *StateDB) MarkStateObjectDirty(addr common.Address) {
+	self.stateObjectsDirty[addr] = struct{}{}
+}
+
 // Creates creates a new state object and takes ownership.
 func (self *StateDB) CreateStateObject(addr common.Address) *StateObject {
 	// Get previous (if any)
@@ -361,32 +367,35 @@ func (self *StateDB) CreateAccount(addr common.Address) vm.Account {
 //
 
 func (self *StateDB) Copy() *StateDB {
-	// ignore error - we assume state-to-be-copied always exists
+	// Copy all the basic fields, initialize the memory ones
-	state, _ := New(common.Hash{}, self.db)
+	state := &StateDB{
-	state.trie = self.trie
+		db:                self.db,
-	state.all = self.all
+		trie:              self.trie,
-	state.accountBloom = self.accountBloom
+		all:               self.all,
-	state.accountBloomDirty = self.accountBloomDirty
+		accountBloom:      self.accountBloom,
-	for addr, stateObject := range self.stateObjects {
+		accountBloomDirty: self.accountBloomDirty,
-		if stateObject.dirty {
+		stateObjects:      make(map[common.Address]*StateObject, len(self.stateObjectsDirty)),
-			state.stateObjects[addr] = stateObject.Copy(self.db)
+		stateObjectsDirty: make(map[common.Address]struct{}, len(self.stateObjectsDirty)),
+		refund:            new(big.Int).Set(self.refund),
+		logs:              make(map[common.Hash]vm.Logs, len(self.logs)),
+		logSize:           self.logSize,
 	}
+	// Copy the dirty states and logs
+	for addr, _ := range self.stateObjectsDirty {
+		state.stateObjects[addr] = self.stateObjects[addr].Copy(self.db, state.MarkStateObjectDirty)
+		state.stateObjectsDirty[addr] = struct{}{}
 	}
-
-	state.refund.Set(self.refund)
-
 	for hash, logs := range self.logs {
 		state.logs[hash] = make(vm.Logs, len(logs))
 		copy(state.logs[hash], logs)
 	}
-	state.logSize = self.logSize
-
 	return state
 }
 
 func (self *StateDB) Set(state *StateDB) {
 	self.trie = state.trie
 	self.stateObjects = state.stateObjects
+	self.stateObjectsDirty = state.stateObjectsDirty
 	self.all = state.all
 	self.accountBloom = state.accountBloom
 	self.accountBloomDirty = state.accountBloomDirty
@@ -405,8 +414,8 @@ func (self *StateDB) GetRefund() *big.Int {
 // goes into transaction receipts.
 func (s *StateDB) IntermediateRoot() common.Hash {
 	s.refund = new(big.Int)
-	for _, stateObject := range s.stateObjects {
+	for addr, _ := range s.stateObjectsDirty {
-		if stateObject.dirty {
+		stateObject := s.stateObjects[addr]
 		if stateObject.remove {
 			s.DeleteStateObject(stateObject)
 		} else {
@@ -414,7 +423,6 @@ func (s *StateDB) IntermediateRoot() common.Hash {
 			s.UpdateStateObject(stateObject)
 		}
 	}
-	}
 	return s.trie.Hash()
 }
 
@@ -427,15 +435,15 @@ func (s *StateDB) DeleteSuicides() {
 	// Reset refund so that any used-gas calculations can use
 	// this method.
 	s.refund = new(big.Int)
-	for _, stateObject := range s.stateObjects {
+	for addr, _ := range s.stateObjectsDirty {
-		if stateObject.dirty {
+		stateObject := s.stateObjects[addr]
+
 		// If the object has been removed by a suicide
 		// flag the object as deleted.
 		if stateObject.remove {
 			stateObject.deleted = true
 		}
-			stateObject.dirty = false
+		delete(s.stateObjectsDirty, addr)
-		}
 	}
 }
 
@@ -470,7 +478,7 @@ func (s *StateDB) commit(dbw trie.DatabaseWriter) (root common.Hash, err error)
 			// and just mark it for deletion in the trie.
 			s.DeleteStateObject(stateObject)
 			delete(s.all, addr)
-		} else if stateObject.dirty {
+		} else if _, ok := s.stateObjectsDirty[addr]; ok {
 			// Write any contract code associated with the state object
 			if stateObject.code != nil && stateObject.dirtyCode {
 				if err := dbw.Put(stateObject.CodeHash(), stateObject.code); err != nil {
@@ -486,7 +494,7 @@ func (s *StateDB) commit(dbw trie.DatabaseWriter) (root common.Hash, err error)
 			s.UpdateStateObject(stateObject)
 			s.all[addr] = stateObject.data
 		}
-		stateObject.dirty = false
+		delete(s.stateObjectsDirty, addr)
 	}
 	// Write account bloom and trie changes
 	if s.accountBloomDirty {
@@ -522,7 +530,7 @@ func (self *StateDB) getAccountBloom() *boom.ScalableBloomFilter {
 		return self.accountBloom
 	}
 	// If the database does not yet contain a bloom filter, generate a new one
-	glog.V(logger.Info).Infof("generating accounts bloom, please wait...")
+	glog.V(logger.Info).Infof("generating account bloom, please wait...")
 
 	filter := boom.NewDefaultScalableBloomFilter(0.01)
 	start, pref := time.Now(), byte(0x00)
@@ -531,13 +539,17 @@ func (self *StateDB) getAccountBloom() *boom.ScalableBloomFilter {
 	for it.Next() {
 		if pref != it.Key[0] {
 			pref = it.Key[0]
-			glog.V(logger.Info).Infof("generating accounts bloom, at %.2f%%, please wait...", float64(pref)/2.56)
+			glog.V(logger.Info).Infof("generating account bloom, at %.2f%%, please wait...", float64(pref)/2.56)
 		}
 		filter.Add(it.Key)
 	}
-	glog.V(logger.Info).Infof("accounts bloom generated in %v", time.Since(start))
+	glog.V(logger.Info).Infof("account bloom generated in %v", time.Since(start))
 
-	self.accountBloom, self.accountBloomDirty = filter, true
+	self.accountBloom = filter
+	if err := self.writeAccountsBloom(self.db); err != nil {
+		glog.V(logger.Error).Infof("failed to write initial account bloom filter: %v", err)
+		return nil
+	}
 	return self.accountBloom
 }
 

core/state/statedb_test.go

@@ -17,6 +17,7 @@
 package state
 
 import (
+	"bytes"
 	"math/big"
 	"testing"
 
@@ -47,6 +48,9 @@ func TestUpdateLeaks(t *testing.T) {
 	// Ensure that no data was leaked into the database
 	for _, key := range db.Keys() {
 		value, _ := db.Get(key)
+		if bytes.Equal(key, []byte("accounts-bloom")) {
+			continue // This is written on first read
+		}
 		t.Errorf("State leaked into database: %x -> %x", key, value)
 	}
 }

tests/state_test_util.go

@@ -97,7 +97,7 @@ func benchStateTest(ruleSet RuleSet, test VmTest, env map[string]string, b *test
 	db, _ := ethdb.NewMemDatabase()
 	statedb, _ := state.New(common.Hash{}, db)
 	for addr, account := range test.Pre {
-		obj := StateObjectFromAccount(db, addr, account)
+		obj := StateObjectFromAccount(db, addr, account, statedb.MarkStateObjectDirty)
 		statedb.SetStateObject(obj)
 		for a, v := range account.Storage {
 			obj.SetState(common.HexToHash(a), common.HexToHash(v))
@@ -136,7 +136,7 @@ func runStateTest(ruleSet RuleSet, test VmTest) error {
 	db, _ := ethdb.NewMemDatabase()
 	statedb, _ := state.New(common.Hash{}, db)
 	for addr, account := range test.Pre {
-		obj := StateObjectFromAccount(db, addr, account)
+		obj := StateObjectFromAccount(db, addr, account, statedb.MarkStateObjectDirty)
 		statedb.SetStateObject(obj)
 		for a, v := range account.Storage {
 			obj.SetState(common.HexToHash(a), common.HexToHash(v))

tests/util.go

@@ -103,7 +103,7 @@ func (self Log) Topics() [][]byte {
 	return t
 }
 
-func StateObjectFromAccount(db ethdb.Database, addr string, account Account) *state.StateObject {
+func StateObjectFromAccount(db ethdb.Database, addr string, account Account, onDirty func(common.Address)) *state.StateObject {
 	if common.IsHex(account.Code) {
 		account.Code = account.Code[2:]
 	}
@@ -112,7 +112,7 @@ func StateObjectFromAccount(db ethdb.Database, addr string, account Account) *st
 		Balance:  common.Big(account.Balance),
 		CodeHash: crypto.Keccak256(code),
 		Nonce:    common.Big(account.Nonce).Uint64(),
-	})
+	}, onDirty)
 	obj.SetCode(code)
 	return obj
 }

tests/vm_test_util.go

@@ -103,7 +103,7 @@ func benchVmTest(test VmTest, env map[string]string, b *testing.B) {
 	db, _ := ethdb.NewMemDatabase()
 	statedb, _ := state.New(common.Hash{}, db)
 	for addr, account := range test.Pre {
-		obj := StateObjectFromAccount(db, addr, account)
+		obj := StateObjectFromAccount(db, addr, account, statedb.MarkStateObjectDirty)
 		statedb.SetStateObject(obj)
 		for a, v := range account.Storage {
 			obj.SetState(common.HexToHash(a), common.HexToHash(v))
@@ -154,7 +154,7 @@ func runVmTest(test VmTest) error {
 	db, _ := ethdb.NewMemDatabase()
 	statedb, _ := state.New(common.Hash{}, db)
 	for addr, account := range test.Pre {
-		obj := StateObjectFromAccount(db, addr, account)
+		obj := StateObjectFromAccount(db, addr, account, statedb.MarkStateObjectDirty)
 		statedb.SetStateObject(obj)
 		for a, v := range account.Storage {
 			obj.SetState(common.HexToHash(a), common.HexToHash(v))