core: store genesis allocation and recommit them if necessary (#24460)

* core: store genesis allocation and recommit them if necessary

* core: recover predefined genesis allocation if possible

core/blockchain.go

@@ -542,6 +542,19 @@ func (bc *BlockChain) setHeadBeyondRoot(head uint64, root common.Hash, repair bo
 						}
 					}
 					if beyondRoot || newHeadBlock.NumberU64() == 0 {
+						if newHeadBlock.NumberU64() == 0 {
+							// Recommit the genesis state into disk in case the rewinding destination
+							// is genesis block and the relevant state is gone. In the future this
+							// rewinding destination can be the earliest block stored in the chain
+							// if the historical chain pruning is enabled. In that case the logic
+							// needs to be improved here.
+							if !bc.HasState(bc.genesisBlock.Root()) {
+								if err := CommitGenesisState(bc.db, bc.genesisBlock.Hash()); err != nil {
+									log.Crit("Failed to commit genesis state", "err", err)
+								}
+								log.Debug("Recommitted genesis state to disk")
+							}
+						}
 						log.Debug("Rewound to block with state", "number", newHeadBlock.NumberU64(), "hash", newHeadBlock.Hash())
 						break
 					}

core/genesis.go

@@ -80,6 +80,81 @@ func (ga *GenesisAlloc) UnmarshalJSON(data []byte) error {
 	return nil
 }
 
+// flush adds allocated genesis accounts into a fresh new statedb and
+// commit the state changes into the given database handler.
+func (ga *GenesisAlloc) flush(db ethdb.Database) (common.Hash, error) {
+	statedb, err := state.New(common.Hash{}, state.NewDatabase(db), nil)
+	if err != nil {
+		return common.Hash{}, err
+	}
+	for addr, account := range *ga {
+		statedb.AddBalance(addr, account.Balance)
+		statedb.SetCode(addr, account.Code)
+		statedb.SetNonce(addr, account.Nonce)
+		for key, value := range account.Storage {
+			statedb.SetState(addr, key, value)
+		}
+	}
+	root, err := statedb.Commit(false)
+	if err != nil {
+		return common.Hash{}, err
+	}
+	err = statedb.Database().TrieDB().Commit(root, true, nil)
+	if err != nil {
+		return common.Hash{}, err
+	}
+	return root, nil
+}
+
+// write writes the json marshaled genesis state into database
+// with the given block hash as the unique identifier.
+func (ga *GenesisAlloc) write(db ethdb.KeyValueWriter, hash common.Hash) error {
+	blob, err := json.Marshal(ga)
+	if err != nil {
+		return err
+	}
+	rawdb.WriteGenesisState(db, hash, blob)
+	return nil
+}
+
+// CommitGenesisState loads the stored genesis state with the given block
+// hash and commits them into the given database handler.
+func CommitGenesisState(db ethdb.Database, hash common.Hash) error {
+	var alloc GenesisAlloc
+	blob := rawdb.ReadGenesisState(db, hash)
+	if len(blob) != 0 {
+		if err := alloc.UnmarshalJSON(blob); err != nil {
+			return err
+		}
+	} else {
+		// Genesis allocation is missing and there are several possibilities:
+		// the node is legacy which doesn't persist the genesis allocation or
+		// the persisted allocation is just lost.
+		// - supported networks(mainnet, testnets), recover with defined allocations
+		// - private network, can't recover
+		var genesis *Genesis
+		switch hash {
+		case params.MainnetGenesisHash:
+			genesis = DefaultGenesisBlock()
+		case params.RopstenGenesisHash:
+			genesis = DefaultRopstenGenesisBlock()
+		case params.RinkebyGenesisHash:
+			genesis = DefaultRinkebyGenesisBlock()
+		case params.GoerliGenesisHash:
+			genesis = DefaultGoerliGenesisBlock()
+		case params.SepoliaGenesisHash:
+			genesis = DefaultSepoliaGenesisBlock()
+		}
+		if genesis != nil {
+			alloc = genesis.Alloc
+		} else {
+			return errors.New("not found")
+		}
+	}
+	_, err := alloc.flush(db)
+	return err
+}
+
 // GenesisAccount is an account in the state of the genesis block.
 type GenesisAccount struct {
 	Code       []byte                      `json:"code,omitempty"`
@@ -267,19 +342,10 @@ func (g *Genesis) ToBlock(db ethdb.Database) *types.Block {
 	if db == nil {
 		db = rawdb.NewMemoryDatabase()
 	}
-	statedb, err := state.New(common.Hash{}, state.NewDatabase(db), nil)
+	root, err := g.Alloc.flush(db)
 	if err != nil {
 		panic(err)
 	}
-	for addr, account := range g.Alloc {
-		statedb.AddBalance(addr, account.Balance)
-		statedb.SetCode(addr, account.Code)
-		statedb.SetNonce(addr, account.Nonce)
-		for key, value := range account.Storage {
-			statedb.SetState(addr, key, value)
-		}
-	}
-	root := statedb.IntermediateRoot(false)
 	head := &types.Header{
 		Number:     new(big.Int).SetUint64(g.Number),
 		Nonce:      types.EncodeNonce(g.Nonce),
@@ -307,9 +373,6 @@ func (g *Genesis) ToBlock(db ethdb.Database) *types.Block {
 			head.BaseFee = new(big.Int).SetUint64(params.InitialBaseFee)
 		}
 	}
-	statedb.Commit(false)
-	statedb.Database().TrieDB().Commit(root, true, nil)
-
 	return types.NewBlock(head, nil, nil, nil, trie.NewStackTrie(nil))
 }
 
@@ -330,6 +393,9 @@ func (g *Genesis) Commit(db ethdb.Database) (*types.Block, error) {
 	if config.Clique != nil && len(block.Extra()) == 0 {
 		return nil, errors.New("can't start clique chain without signers")
 	}
+	if err := g.Alloc.write(db, block.Hash()); err != nil {
+		return nil, err
+	}
 	rawdb.WriteTd(db, block.Hash(), block.NumberU64(), block.Difficulty())
 	rawdb.WriteBlock(db, block)
 	rawdb.WriteReceipts(db, block.Hash(), block.NumberU64(), nil)

core/genesis_test.go

@@ -213,3 +213,33 @@ func TestGenesis_Commit(t *testing.T) {
 		t.Errorf("inequal difficulty; stored: %v, genesisBlock: %v", stored, genesisBlock.Difficulty())
 	}
 }
+
+func TestReadWriteGenesisAlloc(t *testing.T) {
+	var (
+		db    = rawdb.NewMemoryDatabase()
+		alloc = &GenesisAlloc{
+			{1}: {Balance: big.NewInt(1), Storage: map[common.Hash]common.Hash{{1}: {1}}},
+			{2}: {Balance: big.NewInt(2), Storage: map[common.Hash]common.Hash{{2}: {2}}},
+		}
+		hash = common.HexToHash("0xdeadbeef")
+	)
+	alloc.write(db, hash)
+
+	var reload GenesisAlloc
+	err := reload.UnmarshalJSON(rawdb.ReadGenesisState(db, hash))
+	if err != nil {
+		t.Fatalf("Failed to load genesis state %v", err)
+	}
+	if len(reload) != len(*alloc) {
+		t.Fatal("Unexpected genesis allocation")
+	}
+	for addr, account := range reload {
+		want, ok := (*alloc)[addr]
+		if !ok {
+			t.Fatal("Account is not found")
+		}
+		if !reflect.DeepEqual(want, account) {
+			t.Fatal("Unexpected account")
+		}
+	}
+}

core/rawdb/accessors_metadata.go

@@ -81,6 +81,19 @@ func WriteChainConfig(db ethdb.KeyValueWriter, hash common.Hash, cfg *params.Cha
 	}
 }
 
+// ReadGenesisState retrieves the genesis state based on the given genesis hash.
+func ReadGenesisState(db ethdb.KeyValueReader, hash common.Hash) []byte {
+	data, _ := db.Get(genesisKey(hash))
+	return data
+}
+
+// WriteGenesisState writes the genesis state into the disk.
+func WriteGenesisState(db ethdb.KeyValueWriter, hash common.Hash, data []byte) {
+	if err := db.Put(genesisKey(hash), data); err != nil {
+		log.Crit("Failed to store genesis state", "err", err)
+	}
+}
+
 // crashList is a list of unclean-shutdown-markers, for rlp-encoding to the
 // database
 type crashList struct {

core/rawdb/database.go

@@ -386,6 +386,8 @@ func InspectDatabase(db ethdb.Database, keyPrefix, keyStart []byte) error {
 			preimages.Add(size)
 		case bytes.HasPrefix(key, configPrefix) && len(key) == (len(configPrefix)+common.HashLength):
 			metadata.Add(size)
+		case bytes.HasPrefix(key, genesisPrefix) && len(key) == (len(genesisPrefix)+common.HashLength):
+			metadata.Add(size)
 		case bytes.HasPrefix(key, bloomBitsPrefix) && len(key) == (len(bloomBitsPrefix)+10+common.HashLength):
 			bloomBits.Add(size)
 		case bytes.HasPrefix(key, BloomBitsIndexPrefix):

core/rawdb/schema.go

@@ -97,8 +97,9 @@ var (
 	CodePrefix            = []byte("c") // CodePrefix + code hash -> account code
 	skeletonHeaderPrefix  = []byte("S") // skeletonHeaderPrefix + num (uint64 big endian) -> header
 
-	PreimagePrefix = []byte("secure-key-")      // PreimagePrefix + hash -> preimage
-	configPrefix   = []byte("ethereum-config-") // config prefix for the db
+	PreimagePrefix = []byte("secure-key-")       // PreimagePrefix + hash -> preimage
+	configPrefix   = []byte("ethereum-config-")  // config prefix for the db
+	genesisPrefix  = []byte("ethereum-genesis-") // genesis state prefix for the db
 
 	// Chain index prefixes (use `i` + single byte to avoid mixing data types).
 	BloomBitsIndexPrefix = []byte("iB") // BloomBitsIndexPrefix is the data table of a chain indexer to track its progress
@@ -242,3 +243,8 @@ func IsCodeKey(key []byte) (bool, []byte) {
 func configKey(hash common.Hash) []byte {
 	return append(configPrefix, hash.Bytes()...)
 }
+
+// genesisKey = genesisPrefix + hash
+func genesisKey(hash common.Hash) []byte {
+	return append(genesisPrefix, hash.Bytes()...)
+}