cmd/geth, eth, core: snapshot dump + unify with trie dump (ethereum#2…

…2795)

* cmd/geth, eth, core: snapshot dump + unify with trie dump

* cmd/evm: dump API fixes

* cmd/geth, core, eth: fix some remaining errors

* cmd/evm: dump - add limit, support address startkey, address review concerns

* cmd, core/state, eth: minor polishes, fix snap dump crash, unify format

Co-authored-by: Péter Szilágyi <peterke@gmail.com>

cmd/evm/internal/t8ntool/transition.go

@@ -212,16 +212,15 @@ func Main(ctx *cli.Context) error {
 	// Iterate over all the tests, run them and aggregate the results
 
 	// Run the test and aggregate the result
-	state, result, err := prestate.Apply(vmConfig, chainConfig, txs, ctx.Int64(RewardFlag.Name), getTracer)
+	s, result, err := prestate.Apply(vmConfig, chainConfig, txs, ctx.Int64(RewardFlag.Name), getTracer)
 	if err != nil {
 		return err
 	}
 	body, _ := rlp.EncodeToBytes(txs)
 	// Dump the excution result
 	collector := make(Alloc)
-	state.DumpToCollector(collector, false, false, false, nil, -1)
+	s.DumpToCollector(collector, nil)
 	return dispatchOutput(ctx, baseDir, result, collector, body)
-
 }
 
 // txWithKey is a helper-struct, to allow us to use the types.Transaction along with
@@ -303,7 +302,7 @@ func (g Alloc) OnAccount(addr common.Address, dumpAccount state.DumpAccount) {
 		}
 	}
 	genesisAccount := core.GenesisAccount{
-		Code:    common.FromHex(dumpAccount.Code),
+		Code:    dumpAccount.Code,
 		Storage: storage,
 		Balance: balance,
 		Nonce:   dumpAccount.Nonce,

cmd/evm/runner.go

@@ -270,7 +270,7 @@ func runCmd(ctx *cli.Context) error {
 	if ctx.GlobalBool(DumpFlag.Name) {
 		statedb.Commit(true)
 		statedb.IntermediateRoot(true)
-		fmt.Println(string(statedb.Dump(false, false, true)))
+		fmt.Println(string(statedb.Dump(nil)))
 	}
 
 	if memProfilePath := ctx.GlobalString(MemProfileFlag.Name); memProfilePath != "" {

cmd/evm/staterunner.go

@@ -98,16 +98,16 @@ func stateTestCmd(ctx *cli.Context) error {
 		for _, st := range test.Subtests() {
 			// Run the test and aggregate the result
 			result := &StatetestResult{Name: key, Fork: st.Fork, Pass: true}
-			_, state, err := test.Run(st, cfg, false)
+			_, s, err := test.Run(st, cfg, false)
 			// print state root for evmlab tracing
-			if ctx.GlobalBool(MachineFlag.Name) && state != nil {
-				fmt.Fprintf(os.Stderr, "{\"stateRoot\": \"%x\"}\n", state.IntermediateRoot(false))
+			if ctx.GlobalBool(MachineFlag.Name) && s != nil {
+				fmt.Fprintf(os.Stderr, "{\"stateRoot\": \"%x\"}\n", s.IntermediateRoot(false))
 			}
 			if err != nil {
 				// Test failed, mark as so and dump any state to aid debugging
 				result.Pass, result.Error = false, err.Error()
-				if ctx.GlobalBool(DumpFlag.Name) && state != nil {
-					dump := state.RawDump(false, false, true)
+				if ctx.GlobalBool(DumpFlag.Name) && s != nil {
+					dump := s.RawDump(nil)
 					result.State = &dump
 				}
 			}

cmd/geth/chaincmd.go

@@ -18,6 +18,7 @@ package main
 
 import (
 	"encoding/json"
+	"errors"
 	"fmt"
 	"os"
 	"runtime"
@@ -27,12 +28,16 @@ import (
 
 	"github.com/ethereum/go-ethereum/cmd/utils"
 	"github.com/ethereum/go-ethereum/common"
+	"github.com/ethereum/go-ethereum/common/hexutil"
 	"github.com/ethereum/go-ethereum/core"
 	"github.com/ethereum/go-ethereum/core/rawdb"
 	"github.com/ethereum/go-ethereum/core/state"
 	"github.com/ethereum/go-ethereum/core/types"
+	"github.com/ethereum/go-ethereum/crypto"
+	"github.com/ethereum/go-ethereum/ethdb"
 	"github.com/ethereum/go-ethereum/log"
 	"github.com/ethereum/go-ethereum/metrics"
+	"github.com/ethereum/go-ethereum/node"
 	"gopkg.in/urfave/cli.v1"
 )
 
@@ -152,20 +157,21 @@ The export-preimages command export hash preimages to an RLP encoded stream`,
 		Action:    utils.MigrateFlags(dump),
 		Name:      "dump",
 		Usage:     "Dump a specific block from storage",
-		ArgsUsage: "[<blockHash> | <blockNum>]...",
+		ArgsUsage: "[? <blockHash> | <blockNum>]",
 		Flags: []cli.Flag{
 			utils.DataDirFlag,
 			utils.CacheFlag,
-			utils.SyncModeFlag,
 			utils.IterativeOutputFlag,
 			utils.ExcludeCodeFlag,
 			utils.ExcludeStorageFlag,
 			utils.IncludeIncompletesFlag,
+			utils.StartKeyFlag,
+			utils.DumpLimitFlag,
 		},
 		Category: "BLOCKCHAIN COMMANDS",
 		Description: `
-The arguments are interpreted as block numbers or hashes.
-Use "ethereum dump 0" to dump the genesis block.`,
+This command dumps out the state for a given block (or latest, if none provided).
+`,
 	}
 )
 
@@ -373,47 +379,85 @@ func exportPreimages(ctx *cli.Context) error {
 	return nil
 }
 
-func dump(ctx *cli.Context) error {
-	stack, _ := makeConfigNode(ctx)
-	defer stack.Close()
-
+func parseDumpConfig(ctx *cli.Context, stack *node.Node) (*state.DumpConfig, ethdb.Database, common.Hash, error) {
 	db := utils.MakeChainDatabase(ctx, stack, true)
-	for _, arg := range ctx.Args() {
-		var header *types.Header
+	var header *types.Header
+	if ctx.NArg() > 1 {
+		return nil, nil, common.Hash{}, fmt.Errorf("expected 1 argument (number or hash), got %d", ctx.NArg())
+	}
+	if ctx.NArg() == 1 {
+		arg := ctx.Args().First()
 		if hashish(arg) {
 			hash := common.HexToHash(arg)
-			number := rawdb.ReadHeaderNumber(db, hash)
-			if number != nil {
+			if number := rawdb.ReadHeaderNumber(db, hash); number != nil {
 				header = rawdb.ReadHeader(db, hash, *number)
+			} else {
+				return nil, nil, common.Hash{}, fmt.Errorf("block %x not found", hash)
 			}
 		} else {
-			number, _ := strconv.Atoi(arg)
-			hash := rawdb.ReadCanonicalHash(db, uint64(number))
-			if hash != (common.Hash{}) {
-				header = rawdb.ReadHeader(db, hash, uint64(number))
-			}
-		}
-		if header == nil {
-			fmt.Println("{}")
-			utils.Fatalf("block not found")
-		} else {
-			state, err := state.New(header.Root, state.NewDatabase(db), nil)
+			number, err := strconv.Atoi(arg)
 			if err != nil {
-				utils.Fatalf("could not create new state: %v", err)
+				return nil, nil, common.Hash{}, err
 			}
-			excludeCode := ctx.Bool(utils.ExcludeCodeFlag.Name)
-			excludeStorage := ctx.Bool(utils.ExcludeStorageFlag.Name)
-			includeMissing := ctx.Bool(utils.IncludeIncompletesFlag.Name)
-			if ctx.Bool(utils.IterativeOutputFlag.Name) {
-				state.IterativeDump(excludeCode, excludeStorage, !includeMissing, json.NewEncoder(os.Stdout))
+			if hash := rawdb.ReadCanonicalHash(db, uint64(number)); hash != (common.Hash{}) {
+				header = rawdb.ReadHeader(db, hash, uint64(number))
 			} else {
-				if includeMissing {
-					fmt.Printf("If you want to include accounts with missing preimages, you need iterative output, since" +
-						" otherwise the accounts will overwrite each other in the resulting mapping.")
-				}
-				fmt.Printf("%v %s\n", includeMissing, state.Dump(excludeCode, excludeStorage, false))
+				return nil, nil, common.Hash{}, fmt.Errorf("header for block %d not found", number)
 			}
 		}
+	} else {
+		// Use latest
+		header = rawdb.ReadHeadHeader(db)
+	}
+	if header == nil {
+		return nil, nil, common.Hash{}, errors.New("no head block found")
+	}
+	startArg := common.FromHex(ctx.String(utils.StartKeyFlag.Name))
+	var start common.Hash
+	switch len(startArg) {
+	case 0: // common.Hash
+	case 32:
+		start = common.BytesToHash(startArg)
+	case 20:
+		start = crypto.Keccak256Hash(startArg)
+		log.Info("Converting start-address to hash", "address", common.BytesToAddress(startArg), "hash", start.Hex())
+	default:
+		return nil, nil, common.Hash{}, fmt.Errorf("invalid start argument: %x. 20 or 32 hex-encoded bytes required", startArg)
+	}
+	var conf = &state.DumpConfig{
+		SkipCode:          ctx.Bool(utils.ExcludeCodeFlag.Name),
+		SkipStorage:       ctx.Bool(utils.ExcludeStorageFlag.Name),
+		OnlyWithAddresses: !ctx.Bool(utils.IncludeIncompletesFlag.Name),
+		Start:             start.Bytes(),
+		Max:               ctx.Uint64(utils.DumpLimitFlag.Name),
+	}
+	log.Info("State dump configured", "block", header.Number, "hash", header.Hash().Hex(),
+		"skipcode", conf.SkipCode, "skipstorage", conf.SkipStorage,
+		"start", hexutil.Encode(conf.Start), "limit", conf.Max)
+	return conf, db, header.Root, nil
+}
+
+func dump(ctx *cli.Context) error {
+	stack, _ := makeConfigNode(ctx)
+	defer stack.Close()
+
+	conf, db, root, err := parseDumpConfig(ctx, stack)
+	if err != nil {
+		return err
+	}
+	state, err := state.New(root, state.NewDatabase(db), nil)
+	if err != nil {
+		return err
+	}
+	if ctx.Bool(utils.IterativeOutputFlag.Name) {
+		state.IterativeDump(conf, json.NewEncoder(os.Stdout))
+	} else {
+		if conf.OnlyWithAddresses {
+			fmt.Fprintf(os.Stderr, "If you want to include accounts with missing preimages, you need iterative output, since"+
+				" otherwise the accounts will overwrite each other in the resulting mapping.")
+			return fmt.Errorf("incompatible options")
+		}
+		fmt.Println(string(state.Dump(conf)))
 	}
 	return nil
 }

cmd/geth/snapshot.go

@@ -18,7 +18,9 @@ package main
 
 import (
 	"bytes"
+	"encoding/json"
 	"errors"
+	"os"
 	"time"
 
 	"github.com/ethereum/go-ethereum/cmd/utils"
@@ -142,6 +144,31 @@ verification. The default checking target is the HEAD state. It's basically iden
 to traverse-state, but the check granularity is smaller. 
 
 It's also usable without snapshot enabled.
+`,
+			},
+			{
+				Name:      "dump",
+				Usage:     "Dump a specific block from storage (same as 'geth dump' but using snapshots)",
+				ArgsUsage: "[? <blockHash> | <blockNum>]",
+				Action:    utils.MigrateFlags(dumpState),
+				Category:  "MISCELLANEOUS COMMANDS",
+				Flags: []cli.Flag{
+					utils.DataDirFlag,
+					utils.AncientFlag,
+					utils.RopstenFlag,
+					utils.RinkebyFlag,
+					utils.GoerliFlag,
+					utils.ExcludeCodeFlag,
+					utils.ExcludeStorageFlag,
+					utils.StartKeyFlag,
+					utils.DumpLimitFlag,
+				},
+				Description: `
+This command is semantically equivalent to 'geth dump', but uses the snapshots
+as the backend data source, making this command a lot faster. 
+
+The argument is interpreted as block number or hash. If none is provided, the latest
+block is used.
 `,
 			},
 		},
@@ -430,3 +457,73 @@ func parseRoot(input string) (common.Hash, error) {
 	}
 	return h, nil
 }
+
+func dumpState(ctx *cli.Context) error {
+	stack, _ := makeConfigNode(ctx)
+	defer stack.Close()
+
+	conf, db, root, err := parseDumpConfig(ctx, stack)
+	if err != nil {
+		return err
+	}
+	snaptree, err := snapshot.New(db, trie.NewDatabase(db), 256, root, false, false, false)
+	if err != nil {
+		return err
+	}
+	accIt, err := snaptree.AccountIterator(root, common.BytesToHash(conf.Start))
+	if err != nil {
+		return err
+	}
+	defer accIt.Release()
+
+	log.Info("Snapshot dumping started", "root", root)
+	var (
+		start    = time.Now()
+		logged   = time.Now()
+		accounts uint64
+	)
+	enc := json.NewEncoder(os.Stdout)
+	enc.Encode(struct {
+		Root common.Hash `json:"root"`
+	}{root})
+	for accIt.Next() {
+		account, err := snapshot.FullAccount(accIt.Account())
+		if err != nil {
+			return err
+		}
+		da := &state.DumpAccount{
+			Balance:   account.Balance.String(),
+			Nonce:     account.Nonce,
+			Root:      account.Root,
+			CodeHash:  account.CodeHash,
+			SecureKey: accIt.Hash().Bytes(),
+		}
+		if !conf.SkipCode && !bytes.Equal(account.CodeHash, emptyCode) {
+			da.Code = rawdb.ReadCode(db, common.BytesToHash(account.CodeHash))
+		}
+		if !conf.SkipStorage {
+			da.Storage = make(map[common.Hash]string)
+
+			stIt, err := snaptree.StorageIterator(root, accIt.Hash(), common.Hash{})
+			if err != nil {
+				return err
+			}
+			for stIt.Next() {
+				da.Storage[stIt.Hash()] = common.Bytes2Hex(stIt.Slot())
+			}
+		}
+		enc.Encode(da)
+		accounts++
+		if time.Since(logged) > 8*time.Second {
+			log.Info("Snapshot dumping in progress", "at", accIt.Hash(), "accounts", accounts,
+				"elapsed", common.PrettyDuration(time.Since(start)))
+			logged = time.Now()
+		}
+		if conf.Max > 0 && accounts >= conf.Max {
+			break
+		}
+	}
+	log.Info("Snapshot dumping complete", "accounts", accounts,
+		"elapsed", common.PrettyDuration(time.Since(start)))
+	return nil
+}

cmd/utils/flags.go

@@ -184,7 +184,7 @@ var (
 		Name:  "exitwhensynced",
 		Usage: "Exits after block synchronisation completes",
 	}
-	IterativeOutputFlag = cli.BoolFlag{
+	IterativeOutputFlag = cli.BoolTFlag{
 		Name:  "iterative",
 		Usage: "Print streaming JSON iteratively, delimited by newlines",
 	}
@@ -200,6 +200,16 @@ var (
 		Name:  "nocode",
 		Usage: "Exclude contract code (save db lookups)",
 	}
+	StartKeyFlag = cli.StringFlag{
+		Name:  "start",
+		Usage: "Start position. Either a hash or address",
+		Value: "0x0000000000000000000000000000000000000000000000000000000000000000",
+	}
+	DumpLimitFlag = cli.Uint64Flag{
+		Name:  "limit",
+		Usage: "Max number of elements (0 = no limit)",
+		Value: 0,
+	}
 	defaultSyncMode = ethconfig.Defaults.SyncMode
 	SyncModeFlag    = TextMarshalerFlag{
 		Name:  "syncmode",