forked from cerc-io/plugeth
a907d7e81a
This enables the following linters - typecheck - unused - staticcheck - bidichk - durationcheck - exportloopref - gosec WIth a few exceptions. - We use a deprecated protobuf in trezor. I didn't want to mess with that, since I cannot meaningfully test any changes there. - The deprecated TypeMux is used in a few places still, so the warning for it is silenced for now. - Using string type in context.WithValue is apparently wrong, one should use a custom type, to prevent collisions between different places in the hierarchy of callers. That should be fixed at some point, but may require some attention. - The warnings for using weak random generator are squashed, since we use a lot of random without need for cryptographic guarantees.
1718 lines
51 KiB
Go
1718 lines
51 KiB
Go
// Copyright 2021 The go-ethereum Authors
|
|
// This file is part of the go-ethereum library.
|
|
//
|
|
// The go-ethereum library is free software: you can redistribute it and/or modify
|
|
// it under the terms of the GNU Lesser General Public License as published by
|
|
// the Free Software Foundation, either version 3 of the License, or
|
|
// (at your option) any later version.
|
|
//
|
|
// The go-ethereum library is distributed in the hope that it will be useful,
|
|
// but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
// GNU Lesser General Public License for more details.
|
|
//
|
|
// You should have received a copy of the GNU Lesser General Public License
|
|
// along with the go-ethereum library. If not, see <http://www.gnu.org/licenses/>.
|
|
|
|
package snap
|
|
|
|
import (
|
|
"bytes"
|
|
"crypto/rand"
|
|
"encoding/binary"
|
|
"fmt"
|
|
"math/big"
|
|
"sort"
|
|
"sync"
|
|
"testing"
|
|
"time"
|
|
|
|
"github.com/ethereum/go-ethereum/common"
|
|
"github.com/ethereum/go-ethereum/core/rawdb"
|
|
"github.com/ethereum/go-ethereum/core/types"
|
|
"github.com/ethereum/go-ethereum/crypto"
|
|
"github.com/ethereum/go-ethereum/ethdb"
|
|
"github.com/ethereum/go-ethereum/light"
|
|
"github.com/ethereum/go-ethereum/log"
|
|
"github.com/ethereum/go-ethereum/rlp"
|
|
"github.com/ethereum/go-ethereum/trie"
|
|
"golang.org/x/crypto/sha3"
|
|
)
|
|
|
|
func TestHashing(t *testing.T) {
|
|
t.Parallel()
|
|
|
|
var bytecodes = make([][]byte, 10)
|
|
for i := 0; i < len(bytecodes); i++ {
|
|
buf := make([]byte, 100)
|
|
rand.Read(buf)
|
|
bytecodes[i] = buf
|
|
}
|
|
var want, got string
|
|
var old = func() {
|
|
hasher := sha3.NewLegacyKeccak256()
|
|
for i := 0; i < len(bytecodes); i++ {
|
|
hasher.Reset()
|
|
hasher.Write(bytecodes[i])
|
|
hash := hasher.Sum(nil)
|
|
got = fmt.Sprintf("%v\n%v", got, hash)
|
|
}
|
|
}
|
|
var new = func() {
|
|
hasher := sha3.NewLegacyKeccak256().(crypto.KeccakState)
|
|
var hash = make([]byte, 32)
|
|
for i := 0; i < len(bytecodes); i++ {
|
|
hasher.Reset()
|
|
hasher.Write(bytecodes[i])
|
|
hasher.Read(hash)
|
|
want = fmt.Sprintf("%v\n%v", want, hash)
|
|
}
|
|
}
|
|
old()
|
|
new()
|
|
if want != got {
|
|
t.Errorf("want\n%v\ngot\n%v\n", want, got)
|
|
}
|
|
}
|
|
|
|
func BenchmarkHashing(b *testing.B) {
|
|
var bytecodes = make([][]byte, 10000)
|
|
for i := 0; i < len(bytecodes); i++ {
|
|
buf := make([]byte, 100)
|
|
rand.Read(buf)
|
|
bytecodes[i] = buf
|
|
}
|
|
var old = func() {
|
|
hasher := sha3.NewLegacyKeccak256()
|
|
for i := 0; i < len(bytecodes); i++ {
|
|
hasher.Reset()
|
|
hasher.Write(bytecodes[i])
|
|
hasher.Sum(nil)
|
|
}
|
|
}
|
|
var new = func() {
|
|
hasher := sha3.NewLegacyKeccak256().(crypto.KeccakState)
|
|
var hash = make([]byte, 32)
|
|
for i := 0; i < len(bytecodes); i++ {
|
|
hasher.Reset()
|
|
hasher.Write(bytecodes[i])
|
|
hasher.Read(hash)
|
|
}
|
|
}
|
|
b.Run("old", func(b *testing.B) {
|
|
b.ReportAllocs()
|
|
for i := 0; i < b.N; i++ {
|
|
old()
|
|
}
|
|
})
|
|
b.Run("new", func(b *testing.B) {
|
|
b.ReportAllocs()
|
|
for i := 0; i < b.N; i++ {
|
|
new()
|
|
}
|
|
})
|
|
}
|
|
|
|
type (
|
|
accountHandlerFunc func(t *testPeer, requestId uint64, root common.Hash, origin common.Hash, limit common.Hash, cap uint64) error
|
|
storageHandlerFunc func(t *testPeer, requestId uint64, root common.Hash, accounts []common.Hash, origin, limit []byte, max uint64) error
|
|
trieHandlerFunc func(t *testPeer, requestId uint64, root common.Hash, paths []TrieNodePathSet, cap uint64) error
|
|
codeHandlerFunc func(t *testPeer, id uint64, hashes []common.Hash, max uint64) error
|
|
)
|
|
|
|
type testPeer struct {
|
|
id string
|
|
test *testing.T
|
|
remote *Syncer
|
|
logger log.Logger
|
|
accountTrie *trie.Trie
|
|
accountValues entrySlice
|
|
storageTries map[common.Hash]*trie.Trie
|
|
storageValues map[common.Hash]entrySlice
|
|
|
|
accountRequestHandler accountHandlerFunc
|
|
storageRequestHandler storageHandlerFunc
|
|
trieRequestHandler trieHandlerFunc
|
|
codeRequestHandler codeHandlerFunc
|
|
term func()
|
|
|
|
// counters
|
|
nAccountRequests int
|
|
nStorageRequests int
|
|
nBytecodeRequests int
|
|
nTrienodeRequests int
|
|
}
|
|
|
|
func newTestPeer(id string, t *testing.T, term func()) *testPeer {
|
|
peer := &testPeer{
|
|
id: id,
|
|
test: t,
|
|
logger: log.New("id", id),
|
|
accountRequestHandler: defaultAccountRequestHandler,
|
|
trieRequestHandler: defaultTrieRequestHandler,
|
|
storageRequestHandler: defaultStorageRequestHandler,
|
|
codeRequestHandler: defaultCodeRequestHandler,
|
|
term: term,
|
|
}
|
|
//stderrHandler := log.StreamHandler(os.Stderr, log.TerminalFormat(true))
|
|
//peer.logger.SetHandler(stderrHandler)
|
|
return peer
|
|
}
|
|
|
|
func (t *testPeer) ID() string { return t.id }
|
|
func (t *testPeer) Log() log.Logger { return t.logger }
|
|
|
|
func (t *testPeer) Stats() string {
|
|
return fmt.Sprintf(`Account requests: %d
|
|
Storage requests: %d
|
|
Bytecode requests: %d
|
|
Trienode requests: %d
|
|
`, t.nAccountRequests, t.nStorageRequests, t.nBytecodeRequests, t.nTrienodeRequests)
|
|
}
|
|
|
|
func (t *testPeer) RequestAccountRange(id uint64, root, origin, limit common.Hash, bytes uint64) error {
|
|
t.logger.Trace("Fetching range of accounts", "reqid", id, "root", root, "origin", origin, "limit", limit, "bytes", common.StorageSize(bytes))
|
|
t.nAccountRequests++
|
|
go t.accountRequestHandler(t, id, root, origin, limit, bytes)
|
|
return nil
|
|
}
|
|
|
|
func (t *testPeer) RequestTrieNodes(id uint64, root common.Hash, paths []TrieNodePathSet, bytes uint64) error {
|
|
t.logger.Trace("Fetching set of trie nodes", "reqid", id, "root", root, "pathsets", len(paths), "bytes", common.StorageSize(bytes))
|
|
t.nTrienodeRequests++
|
|
go t.trieRequestHandler(t, id, root, paths, bytes)
|
|
return nil
|
|
}
|
|
|
|
func (t *testPeer) RequestStorageRanges(id uint64, root common.Hash, accounts []common.Hash, origin, limit []byte, bytes uint64) error {
|
|
t.nStorageRequests++
|
|
if len(accounts) == 1 && origin != nil {
|
|
t.logger.Trace("Fetching range of large storage slots", "reqid", id, "root", root, "account", accounts[0], "origin", common.BytesToHash(origin), "limit", common.BytesToHash(limit), "bytes", common.StorageSize(bytes))
|
|
} else {
|
|
t.logger.Trace("Fetching ranges of small storage slots", "reqid", id, "root", root, "accounts", len(accounts), "first", accounts[0], "bytes", common.StorageSize(bytes))
|
|
}
|
|
go t.storageRequestHandler(t, id, root, accounts, origin, limit, bytes)
|
|
return nil
|
|
}
|
|
|
|
func (t *testPeer) RequestByteCodes(id uint64, hashes []common.Hash, bytes uint64) error {
|
|
t.nBytecodeRequests++
|
|
t.logger.Trace("Fetching set of byte codes", "reqid", id, "hashes", len(hashes), "bytes", common.StorageSize(bytes))
|
|
go t.codeRequestHandler(t, id, hashes, bytes)
|
|
return nil
|
|
}
|
|
|
|
// defaultTrieRequestHandler is a well-behaving handler for trie healing requests
|
|
func defaultTrieRequestHandler(t *testPeer, requestId uint64, root common.Hash, paths []TrieNodePathSet, cap uint64) error {
|
|
// Pass the response
|
|
var nodes [][]byte
|
|
for _, pathset := range paths {
|
|
switch len(pathset) {
|
|
case 1:
|
|
blob, _, err := t.accountTrie.TryGetNode(pathset[0])
|
|
if err != nil {
|
|
t.logger.Info("Error handling req", "error", err)
|
|
break
|
|
}
|
|
nodes = append(nodes, blob)
|
|
default:
|
|
account := t.storageTries[(common.BytesToHash(pathset[0]))]
|
|
for _, path := range pathset[1:] {
|
|
blob, _, err := account.TryGetNode(path)
|
|
if err != nil {
|
|
t.logger.Info("Error handling req", "error", err)
|
|
break
|
|
}
|
|
nodes = append(nodes, blob)
|
|
}
|
|
}
|
|
}
|
|
t.remote.OnTrieNodes(t, requestId, nodes)
|
|
return nil
|
|
}
|
|
|
|
// defaultAccountRequestHandler is a well-behaving handler for AccountRangeRequests
|
|
func defaultAccountRequestHandler(t *testPeer, id uint64, root common.Hash, origin common.Hash, limit common.Hash, cap uint64) error {
|
|
keys, vals, proofs := createAccountRequestResponse(t, root, origin, limit, cap)
|
|
if err := t.remote.OnAccounts(t, id, keys, vals, proofs); err != nil {
|
|
t.test.Errorf("Remote side rejected our delivery: %v", err)
|
|
t.term()
|
|
return err
|
|
}
|
|
return nil
|
|
}
|
|
|
|
func createAccountRequestResponse(t *testPeer, root common.Hash, origin common.Hash, limit common.Hash, cap uint64) (keys []common.Hash, vals [][]byte, proofs [][]byte) {
|
|
var size uint64
|
|
if limit == (common.Hash{}) {
|
|
limit = common.HexToHash("0xffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff")
|
|
}
|
|
for _, entry := range t.accountValues {
|
|
if size > cap {
|
|
break
|
|
}
|
|
if bytes.Compare(origin[:], entry.k) <= 0 {
|
|
keys = append(keys, common.BytesToHash(entry.k))
|
|
vals = append(vals, entry.v)
|
|
size += uint64(32 + len(entry.v))
|
|
}
|
|
// If we've exceeded the request threshold, abort
|
|
if bytes.Compare(entry.k, limit[:]) >= 0 {
|
|
break
|
|
}
|
|
}
|
|
// Unless we send the entire trie, we need to supply proofs
|
|
// Actually, we need to supply proofs either way! This seems to be an implementation
|
|
// quirk in go-ethereum
|
|
proof := light.NewNodeSet()
|
|
if err := t.accountTrie.Prove(origin[:], 0, proof); err != nil {
|
|
t.logger.Error("Could not prove inexistence of origin", "origin", origin, "error", err)
|
|
}
|
|
if len(keys) > 0 {
|
|
lastK := (keys[len(keys)-1])[:]
|
|
if err := t.accountTrie.Prove(lastK, 0, proof); err != nil {
|
|
t.logger.Error("Could not prove last item", "error", err)
|
|
}
|
|
}
|
|
for _, blob := range proof.NodeList() {
|
|
proofs = append(proofs, blob)
|
|
}
|
|
return keys, vals, proofs
|
|
}
|
|
|
|
// defaultStorageRequestHandler is a well-behaving storage request handler
|
|
func defaultStorageRequestHandler(t *testPeer, requestId uint64, root common.Hash, accounts []common.Hash, bOrigin, bLimit []byte, max uint64) error {
|
|
hashes, slots, proofs := createStorageRequestResponse(t, root, accounts, bOrigin, bLimit, max)
|
|
if err := t.remote.OnStorage(t, requestId, hashes, slots, proofs); err != nil {
|
|
t.test.Errorf("Remote side rejected our delivery: %v", err)
|
|
t.term()
|
|
}
|
|
return nil
|
|
}
|
|
|
|
func defaultCodeRequestHandler(t *testPeer, id uint64, hashes []common.Hash, max uint64) error {
|
|
var bytecodes [][]byte
|
|
for _, h := range hashes {
|
|
bytecodes = append(bytecodes, getCodeByHash(h))
|
|
}
|
|
if err := t.remote.OnByteCodes(t, id, bytecodes); err != nil {
|
|
t.test.Errorf("Remote side rejected our delivery: %v", err)
|
|
t.term()
|
|
}
|
|
return nil
|
|
}
|
|
|
|
func createStorageRequestResponse(t *testPeer, root common.Hash, accounts []common.Hash, origin, limit []byte, max uint64) (hashes [][]common.Hash, slots [][][]byte, proofs [][]byte) {
|
|
var size uint64
|
|
for _, account := range accounts {
|
|
// The first account might start from a different origin and end sooner
|
|
var originHash common.Hash
|
|
if len(origin) > 0 {
|
|
originHash = common.BytesToHash(origin)
|
|
}
|
|
var limitHash = common.HexToHash("0xffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff")
|
|
if len(limit) > 0 {
|
|
limitHash = common.BytesToHash(limit)
|
|
}
|
|
var (
|
|
keys []common.Hash
|
|
vals [][]byte
|
|
abort bool
|
|
)
|
|
for _, entry := range t.storageValues[account] {
|
|
if size >= max {
|
|
abort = true
|
|
break
|
|
}
|
|
if bytes.Compare(entry.k, originHash[:]) < 0 {
|
|
continue
|
|
}
|
|
keys = append(keys, common.BytesToHash(entry.k))
|
|
vals = append(vals, entry.v)
|
|
size += uint64(32 + len(entry.v))
|
|
if bytes.Compare(entry.k, limitHash[:]) >= 0 {
|
|
break
|
|
}
|
|
}
|
|
if len(keys) > 0 {
|
|
hashes = append(hashes, keys)
|
|
slots = append(slots, vals)
|
|
}
|
|
// Generate the Merkle proofs for the first and last storage slot, but
|
|
// only if the response was capped. If the entire storage trie included
|
|
// in the response, no need for any proofs.
|
|
if originHash != (common.Hash{}) || (abort && len(keys) > 0) {
|
|
// If we're aborting, we need to prove the first and last item
|
|
// This terminates the response (and thus the loop)
|
|
proof := light.NewNodeSet()
|
|
stTrie := t.storageTries[account]
|
|
|
|
// Here's a potential gotcha: when constructing the proof, we cannot
|
|
// use the 'origin' slice directly, but must use the full 32-byte
|
|
// hash form.
|
|
if err := stTrie.Prove(originHash[:], 0, proof); err != nil {
|
|
t.logger.Error("Could not prove inexistence of origin", "origin", originHash, "error", err)
|
|
}
|
|
if len(keys) > 0 {
|
|
lastK := (keys[len(keys)-1])[:]
|
|
if err := stTrie.Prove(lastK, 0, proof); err != nil {
|
|
t.logger.Error("Could not prove last item", "error", err)
|
|
}
|
|
}
|
|
for _, blob := range proof.NodeList() {
|
|
proofs = append(proofs, blob)
|
|
}
|
|
break
|
|
}
|
|
}
|
|
return hashes, slots, proofs
|
|
}
|
|
|
|
// the createStorageRequestResponseAlwaysProve tests a cornercase, where it always
|
|
// supplies the proof for the last account, even if it is 'complete'.h
|
|
func createStorageRequestResponseAlwaysProve(t *testPeer, root common.Hash, accounts []common.Hash, bOrigin, bLimit []byte, max uint64) (hashes [][]common.Hash, slots [][][]byte, proofs [][]byte) {
|
|
var size uint64
|
|
max = max * 3 / 4
|
|
|
|
var origin common.Hash
|
|
if len(bOrigin) > 0 {
|
|
origin = common.BytesToHash(bOrigin)
|
|
}
|
|
var exit bool
|
|
for i, account := range accounts {
|
|
var keys []common.Hash
|
|
var vals [][]byte
|
|
for _, entry := range t.storageValues[account] {
|
|
if bytes.Compare(entry.k, origin[:]) < 0 {
|
|
exit = true
|
|
}
|
|
keys = append(keys, common.BytesToHash(entry.k))
|
|
vals = append(vals, entry.v)
|
|
size += uint64(32 + len(entry.v))
|
|
if size > max {
|
|
exit = true
|
|
}
|
|
}
|
|
if i == len(accounts)-1 {
|
|
exit = true
|
|
}
|
|
hashes = append(hashes, keys)
|
|
slots = append(slots, vals)
|
|
|
|
if exit {
|
|
// If we're aborting, we need to prove the first and last item
|
|
// This terminates the response (and thus the loop)
|
|
proof := light.NewNodeSet()
|
|
stTrie := t.storageTries[account]
|
|
|
|
// Here's a potential gotcha: when constructing the proof, we cannot
|
|
// use the 'origin' slice directly, but must use the full 32-byte
|
|
// hash form.
|
|
if err := stTrie.Prove(origin[:], 0, proof); err != nil {
|
|
t.logger.Error("Could not prove inexistence of origin", "origin", origin,
|
|
"error", err)
|
|
}
|
|
if len(keys) > 0 {
|
|
lastK := (keys[len(keys)-1])[:]
|
|
if err := stTrie.Prove(lastK, 0, proof); err != nil {
|
|
t.logger.Error("Could not prove last item", "error", err)
|
|
}
|
|
}
|
|
for _, blob := range proof.NodeList() {
|
|
proofs = append(proofs, blob)
|
|
}
|
|
break
|
|
}
|
|
}
|
|
return hashes, slots, proofs
|
|
}
|
|
|
|
// emptyRequestAccountRangeFn is a rejects AccountRangeRequests
|
|
func emptyRequestAccountRangeFn(t *testPeer, requestId uint64, root common.Hash, origin common.Hash, limit common.Hash, cap uint64) error {
|
|
t.remote.OnAccounts(t, requestId, nil, nil, nil)
|
|
return nil
|
|
}
|
|
|
|
func nonResponsiveRequestAccountRangeFn(t *testPeer, requestId uint64, root common.Hash, origin common.Hash, limit common.Hash, cap uint64) error {
|
|
return nil
|
|
}
|
|
|
|
func emptyTrieRequestHandler(t *testPeer, requestId uint64, root common.Hash, paths []TrieNodePathSet, cap uint64) error {
|
|
t.remote.OnTrieNodes(t, requestId, nil)
|
|
return nil
|
|
}
|
|
|
|
func nonResponsiveTrieRequestHandler(t *testPeer, requestId uint64, root common.Hash, paths []TrieNodePathSet, cap uint64) error {
|
|
return nil
|
|
}
|
|
|
|
func emptyStorageRequestHandler(t *testPeer, requestId uint64, root common.Hash, accounts []common.Hash, origin, limit []byte, max uint64) error {
|
|
t.remote.OnStorage(t, requestId, nil, nil, nil)
|
|
return nil
|
|
}
|
|
|
|
func nonResponsiveStorageRequestHandler(t *testPeer, requestId uint64, root common.Hash, accounts []common.Hash, origin, limit []byte, max uint64) error {
|
|
return nil
|
|
}
|
|
|
|
func proofHappyStorageRequestHandler(t *testPeer, requestId uint64, root common.Hash, accounts []common.Hash, origin, limit []byte, max uint64) error {
|
|
hashes, slots, proofs := createStorageRequestResponseAlwaysProve(t, root, accounts, origin, limit, max)
|
|
if err := t.remote.OnStorage(t, requestId, hashes, slots, proofs); err != nil {
|
|
t.test.Errorf("Remote side rejected our delivery: %v", err)
|
|
t.term()
|
|
}
|
|
return nil
|
|
}
|
|
|
|
//func emptyCodeRequestHandler(t *testPeer, id uint64, hashes []common.Hash, max uint64) error {
|
|
// var bytecodes [][]byte
|
|
// t.remote.OnByteCodes(t, id, bytecodes)
|
|
// return nil
|
|
//}
|
|
|
|
func corruptCodeRequestHandler(t *testPeer, id uint64, hashes []common.Hash, max uint64) error {
|
|
var bytecodes [][]byte
|
|
for _, h := range hashes {
|
|
// Send back the hashes
|
|
bytecodes = append(bytecodes, h[:])
|
|
}
|
|
if err := t.remote.OnByteCodes(t, id, bytecodes); err != nil {
|
|
t.logger.Info("remote error on delivery (as expected)", "error", err)
|
|
// Mimic the real-life handler, which drops a peer on errors
|
|
t.remote.Unregister(t.id)
|
|
}
|
|
return nil
|
|
}
|
|
|
|
func cappedCodeRequestHandler(t *testPeer, id uint64, hashes []common.Hash, max uint64) error {
|
|
var bytecodes [][]byte
|
|
for _, h := range hashes[:1] {
|
|
bytecodes = append(bytecodes, getCodeByHash(h))
|
|
}
|
|
// Missing bytecode can be retrieved again, no error expected
|
|
if err := t.remote.OnByteCodes(t, id, bytecodes); err != nil {
|
|
t.test.Errorf("Remote side rejected our delivery: %v", err)
|
|
t.term()
|
|
}
|
|
return nil
|
|
}
|
|
|
|
// starvingStorageRequestHandler is somewhat well-behaving storage handler, but it caps the returned results to be very small
|
|
func starvingStorageRequestHandler(t *testPeer, requestId uint64, root common.Hash, accounts []common.Hash, origin, limit []byte, max uint64) error {
|
|
return defaultStorageRequestHandler(t, requestId, root, accounts, origin, limit, 500)
|
|
}
|
|
|
|
func starvingAccountRequestHandler(t *testPeer, requestId uint64, root common.Hash, origin common.Hash, limit common.Hash, cap uint64) error {
|
|
return defaultAccountRequestHandler(t, requestId, root, origin, limit, 500)
|
|
}
|
|
|
|
//func misdeliveringAccountRequestHandler(t *testPeer, requestId uint64, root common.Hash, origin common.Hash, cap uint64) error {
|
|
// return defaultAccountRequestHandler(t, requestId-1, root, origin, 500)
|
|
//}
|
|
|
|
func corruptAccountRequestHandler(t *testPeer, requestId uint64, root common.Hash, origin common.Hash, limit common.Hash, cap uint64) error {
|
|
hashes, accounts, proofs := createAccountRequestResponse(t, root, origin, limit, cap)
|
|
if len(proofs) > 0 {
|
|
proofs = proofs[1:]
|
|
}
|
|
if err := t.remote.OnAccounts(t, requestId, hashes, accounts, proofs); err != nil {
|
|
t.logger.Info("remote error on delivery (as expected)", "error", err)
|
|
// Mimic the real-life handler, which drops a peer on errors
|
|
t.remote.Unregister(t.id)
|
|
}
|
|
return nil
|
|
}
|
|
|
|
// corruptStorageRequestHandler doesn't provide good proofs
|
|
func corruptStorageRequestHandler(t *testPeer, requestId uint64, root common.Hash, accounts []common.Hash, origin, limit []byte, max uint64) error {
|
|
hashes, slots, proofs := createStorageRequestResponse(t, root, accounts, origin, limit, max)
|
|
if len(proofs) > 0 {
|
|
proofs = proofs[1:]
|
|
}
|
|
if err := t.remote.OnStorage(t, requestId, hashes, slots, proofs); err != nil {
|
|
t.logger.Info("remote error on delivery (as expected)", "error", err)
|
|
// Mimic the real-life handler, which drops a peer on errors
|
|
t.remote.Unregister(t.id)
|
|
}
|
|
return nil
|
|
}
|
|
|
|
func noProofStorageRequestHandler(t *testPeer, requestId uint64, root common.Hash, accounts []common.Hash, origin, limit []byte, max uint64) error {
|
|
hashes, slots, _ := createStorageRequestResponse(t, root, accounts, origin, limit, max)
|
|
if err := t.remote.OnStorage(t, requestId, hashes, slots, nil); err != nil {
|
|
t.logger.Info("remote error on delivery (as expected)", "error", err)
|
|
// Mimic the real-life handler, which drops a peer on errors
|
|
t.remote.Unregister(t.id)
|
|
}
|
|
return nil
|
|
}
|
|
|
|
// TestSyncBloatedProof tests a scenario where we provide only _one_ value, but
|
|
// also ship the entire trie inside the proof. If the attack is successful,
|
|
// the remote side does not do any follow-up requests
|
|
func TestSyncBloatedProof(t *testing.T) {
|
|
t.Parallel()
|
|
|
|
var (
|
|
once sync.Once
|
|
cancel = make(chan struct{})
|
|
term = func() {
|
|
once.Do(func() {
|
|
close(cancel)
|
|
})
|
|
}
|
|
)
|
|
sourceAccountTrie, elems := makeAccountTrieNoStorage(100)
|
|
source := newTestPeer("source", t, term)
|
|
source.accountTrie = sourceAccountTrie
|
|
source.accountValues = elems
|
|
|
|
source.accountRequestHandler = func(t *testPeer, requestId uint64, root common.Hash, origin common.Hash, limit common.Hash, cap uint64) error {
|
|
var (
|
|
proofs [][]byte
|
|
keys []common.Hash
|
|
vals [][]byte
|
|
)
|
|
// The values
|
|
for _, entry := range t.accountValues {
|
|
if bytes.Compare(entry.k, origin[:]) < 0 {
|
|
continue
|
|
}
|
|
if bytes.Compare(entry.k, limit[:]) > 0 {
|
|
continue
|
|
}
|
|
keys = append(keys, common.BytesToHash(entry.k))
|
|
vals = append(vals, entry.v)
|
|
}
|
|
// The proofs
|
|
proof := light.NewNodeSet()
|
|
if err := t.accountTrie.Prove(origin[:], 0, proof); err != nil {
|
|
t.logger.Error("Could not prove origin", "origin", origin, "error", err)
|
|
}
|
|
// The bloat: add proof of every single element
|
|
for _, entry := range t.accountValues {
|
|
if err := t.accountTrie.Prove(entry.k, 0, proof); err != nil {
|
|
t.logger.Error("Could not prove item", "error", err)
|
|
}
|
|
}
|
|
// And remove one item from the elements
|
|
if len(keys) > 2 {
|
|
keys = append(keys[:1], keys[2:]...)
|
|
vals = append(vals[:1], vals[2:]...)
|
|
}
|
|
for _, blob := range proof.NodeList() {
|
|
proofs = append(proofs, blob)
|
|
}
|
|
if err := t.remote.OnAccounts(t, requestId, keys, vals, proofs); err != nil {
|
|
t.logger.Info("remote error on delivery (as expected)", "error", err)
|
|
t.term()
|
|
// This is actually correct, signal to exit the test successfully
|
|
}
|
|
return nil
|
|
}
|
|
syncer := setupSyncer(source)
|
|
if err := syncer.Sync(sourceAccountTrie.Hash(), cancel); err == nil {
|
|
t.Fatal("No error returned from incomplete/cancelled sync")
|
|
}
|
|
}
|
|
|
|
func setupSyncer(peers ...*testPeer) *Syncer {
|
|
stateDb := rawdb.NewMemoryDatabase()
|
|
syncer := NewSyncer(stateDb)
|
|
for _, peer := range peers {
|
|
syncer.Register(peer)
|
|
peer.remote = syncer
|
|
}
|
|
return syncer
|
|
}
|
|
|
|
// TestSync tests a basic sync with one peer
|
|
func TestSync(t *testing.T) {
|
|
t.Parallel()
|
|
|
|
var (
|
|
once sync.Once
|
|
cancel = make(chan struct{})
|
|
term = func() {
|
|
once.Do(func() {
|
|
close(cancel)
|
|
})
|
|
}
|
|
)
|
|
sourceAccountTrie, elems := makeAccountTrieNoStorage(100)
|
|
|
|
mkSource := func(name string) *testPeer {
|
|
source := newTestPeer(name, t, term)
|
|
source.accountTrie = sourceAccountTrie
|
|
source.accountValues = elems
|
|
return source
|
|
}
|
|
syncer := setupSyncer(mkSource("source"))
|
|
if err := syncer.Sync(sourceAccountTrie.Hash(), cancel); err != nil {
|
|
t.Fatalf("sync failed: %v", err)
|
|
}
|
|
verifyTrie(syncer.db, sourceAccountTrie.Hash(), t)
|
|
}
|
|
|
|
// TestSyncTinyTriePanic tests a basic sync with one peer, and a tiny trie. This caused a
|
|
// panic within the prover
|
|
func TestSyncTinyTriePanic(t *testing.T) {
|
|
t.Parallel()
|
|
|
|
var (
|
|
once sync.Once
|
|
cancel = make(chan struct{})
|
|
term = func() {
|
|
once.Do(func() {
|
|
close(cancel)
|
|
})
|
|
}
|
|
)
|
|
sourceAccountTrie, elems := makeAccountTrieNoStorage(1)
|
|
|
|
mkSource := func(name string) *testPeer {
|
|
source := newTestPeer(name, t, term)
|
|
source.accountTrie = sourceAccountTrie
|
|
source.accountValues = elems
|
|
return source
|
|
}
|
|
syncer := setupSyncer(mkSource("source"))
|
|
done := checkStall(t, term)
|
|
if err := syncer.Sync(sourceAccountTrie.Hash(), cancel); err != nil {
|
|
t.Fatalf("sync failed: %v", err)
|
|
}
|
|
close(done)
|
|
verifyTrie(syncer.db, sourceAccountTrie.Hash(), t)
|
|
}
|
|
|
|
// TestMultiSync tests a basic sync with multiple peers
|
|
func TestMultiSync(t *testing.T) {
|
|
t.Parallel()
|
|
|
|
var (
|
|
once sync.Once
|
|
cancel = make(chan struct{})
|
|
term = func() {
|
|
once.Do(func() {
|
|
close(cancel)
|
|
})
|
|
}
|
|
)
|
|
sourceAccountTrie, elems := makeAccountTrieNoStorage(100)
|
|
|
|
mkSource := func(name string) *testPeer {
|
|
source := newTestPeer(name, t, term)
|
|
source.accountTrie = sourceAccountTrie
|
|
source.accountValues = elems
|
|
return source
|
|
}
|
|
syncer := setupSyncer(mkSource("sourceA"), mkSource("sourceB"))
|
|
done := checkStall(t, term)
|
|
if err := syncer.Sync(sourceAccountTrie.Hash(), cancel); err != nil {
|
|
t.Fatalf("sync failed: %v", err)
|
|
}
|
|
close(done)
|
|
verifyTrie(syncer.db, sourceAccountTrie.Hash(), t)
|
|
}
|
|
|
|
// TestSyncWithStorage tests basic sync using accounts + storage + code
|
|
func TestSyncWithStorage(t *testing.T) {
|
|
t.Parallel()
|
|
|
|
var (
|
|
once sync.Once
|
|
cancel = make(chan struct{})
|
|
term = func() {
|
|
once.Do(func() {
|
|
close(cancel)
|
|
})
|
|
}
|
|
)
|
|
sourceAccountTrie, elems, storageTries, storageElems := makeAccountTrieWithStorage(3, 3000, true, false)
|
|
|
|
mkSource := func(name string) *testPeer {
|
|
source := newTestPeer(name, t, term)
|
|
source.accountTrie = sourceAccountTrie
|
|
source.accountValues = elems
|
|
source.storageTries = storageTries
|
|
source.storageValues = storageElems
|
|
return source
|
|
}
|
|
syncer := setupSyncer(mkSource("sourceA"))
|
|
done := checkStall(t, term)
|
|
if err := syncer.Sync(sourceAccountTrie.Hash(), cancel); err != nil {
|
|
t.Fatalf("sync failed: %v", err)
|
|
}
|
|
close(done)
|
|
verifyTrie(syncer.db, sourceAccountTrie.Hash(), t)
|
|
}
|
|
|
|
// TestMultiSyncManyUseless contains one good peer, and many which doesn't return anything valuable at all
|
|
func TestMultiSyncManyUseless(t *testing.T) {
|
|
t.Parallel()
|
|
|
|
var (
|
|
once sync.Once
|
|
cancel = make(chan struct{})
|
|
term = func() {
|
|
once.Do(func() {
|
|
close(cancel)
|
|
})
|
|
}
|
|
)
|
|
sourceAccountTrie, elems, storageTries, storageElems := makeAccountTrieWithStorage(100, 3000, true, false)
|
|
|
|
mkSource := func(name string, noAccount, noStorage, noTrieNode bool) *testPeer {
|
|
source := newTestPeer(name, t, term)
|
|
source.accountTrie = sourceAccountTrie
|
|
source.accountValues = elems
|
|
source.storageTries = storageTries
|
|
source.storageValues = storageElems
|
|
|
|
if !noAccount {
|
|
source.accountRequestHandler = emptyRequestAccountRangeFn
|
|
}
|
|
if !noStorage {
|
|
source.storageRequestHandler = emptyStorageRequestHandler
|
|
}
|
|
if !noTrieNode {
|
|
source.trieRequestHandler = emptyTrieRequestHandler
|
|
}
|
|
return source
|
|
}
|
|
|
|
syncer := setupSyncer(
|
|
mkSource("full", true, true, true),
|
|
mkSource("noAccounts", false, true, true),
|
|
mkSource("noStorage", true, false, true),
|
|
mkSource("noTrie", true, true, false),
|
|
)
|
|
done := checkStall(t, term)
|
|
if err := syncer.Sync(sourceAccountTrie.Hash(), cancel); err != nil {
|
|
t.Fatalf("sync failed: %v", err)
|
|
}
|
|
close(done)
|
|
verifyTrie(syncer.db, sourceAccountTrie.Hash(), t)
|
|
}
|
|
|
|
// TestMultiSyncManyUseless contains one good peer, and many which doesn't return anything valuable at all
|
|
func TestMultiSyncManyUselessWithLowTimeout(t *testing.T) {
|
|
var (
|
|
once sync.Once
|
|
cancel = make(chan struct{})
|
|
term = func() {
|
|
once.Do(func() {
|
|
close(cancel)
|
|
})
|
|
}
|
|
)
|
|
sourceAccountTrie, elems, storageTries, storageElems := makeAccountTrieWithStorage(100, 3000, true, false)
|
|
|
|
mkSource := func(name string, noAccount, noStorage, noTrieNode bool) *testPeer {
|
|
source := newTestPeer(name, t, term)
|
|
source.accountTrie = sourceAccountTrie
|
|
source.accountValues = elems
|
|
source.storageTries = storageTries
|
|
source.storageValues = storageElems
|
|
|
|
if !noAccount {
|
|
source.accountRequestHandler = emptyRequestAccountRangeFn
|
|
}
|
|
if !noStorage {
|
|
source.storageRequestHandler = emptyStorageRequestHandler
|
|
}
|
|
if !noTrieNode {
|
|
source.trieRequestHandler = emptyTrieRequestHandler
|
|
}
|
|
return source
|
|
}
|
|
|
|
syncer := setupSyncer(
|
|
mkSource("full", true, true, true),
|
|
mkSource("noAccounts", false, true, true),
|
|
mkSource("noStorage", true, false, true),
|
|
mkSource("noTrie", true, true, false),
|
|
)
|
|
// We're setting the timeout to very low, to increase the chance of the timeout
|
|
// being triggered. This was previously a cause of panic, when a response
|
|
// arrived simultaneously as a timeout was triggered.
|
|
syncer.rates.OverrideTTLLimit = time.Millisecond
|
|
|
|
done := checkStall(t, term)
|
|
if err := syncer.Sync(sourceAccountTrie.Hash(), cancel); err != nil {
|
|
t.Fatalf("sync failed: %v", err)
|
|
}
|
|
close(done)
|
|
verifyTrie(syncer.db, sourceAccountTrie.Hash(), t)
|
|
}
|
|
|
|
// TestMultiSyncManyUnresponsive contains one good peer, and many which doesn't respond at all
|
|
func TestMultiSyncManyUnresponsive(t *testing.T) {
|
|
var (
|
|
once sync.Once
|
|
cancel = make(chan struct{})
|
|
term = func() {
|
|
once.Do(func() {
|
|
close(cancel)
|
|
})
|
|
}
|
|
)
|
|
sourceAccountTrie, elems, storageTries, storageElems := makeAccountTrieWithStorage(100, 3000, true, false)
|
|
|
|
mkSource := func(name string, noAccount, noStorage, noTrieNode bool) *testPeer {
|
|
source := newTestPeer(name, t, term)
|
|
source.accountTrie = sourceAccountTrie
|
|
source.accountValues = elems
|
|
source.storageTries = storageTries
|
|
source.storageValues = storageElems
|
|
|
|
if !noAccount {
|
|
source.accountRequestHandler = nonResponsiveRequestAccountRangeFn
|
|
}
|
|
if !noStorage {
|
|
source.storageRequestHandler = nonResponsiveStorageRequestHandler
|
|
}
|
|
if !noTrieNode {
|
|
source.trieRequestHandler = nonResponsiveTrieRequestHandler
|
|
}
|
|
return source
|
|
}
|
|
|
|
syncer := setupSyncer(
|
|
mkSource("full", true, true, true),
|
|
mkSource("noAccounts", false, true, true),
|
|
mkSource("noStorage", true, false, true),
|
|
mkSource("noTrie", true, true, false),
|
|
)
|
|
// We're setting the timeout to very low, to make the test run a bit faster
|
|
syncer.rates.OverrideTTLLimit = time.Millisecond
|
|
|
|
done := checkStall(t, term)
|
|
if err := syncer.Sync(sourceAccountTrie.Hash(), cancel); err != nil {
|
|
t.Fatalf("sync failed: %v", err)
|
|
}
|
|
close(done)
|
|
verifyTrie(syncer.db, sourceAccountTrie.Hash(), t)
|
|
}
|
|
|
|
func checkStall(t *testing.T, term func()) chan struct{} {
|
|
testDone := make(chan struct{})
|
|
go func() {
|
|
select {
|
|
case <-time.After(time.Minute): // TODO(karalabe): Make tests smaller, this is too much
|
|
t.Log("Sync stalled")
|
|
term()
|
|
case <-testDone:
|
|
return
|
|
}
|
|
}()
|
|
return testDone
|
|
}
|
|
|
|
// TestSyncBoundaryAccountTrie tests sync against a few normal peers, but the
|
|
// account trie has a few boundary elements.
|
|
func TestSyncBoundaryAccountTrie(t *testing.T) {
|
|
t.Parallel()
|
|
|
|
var (
|
|
once sync.Once
|
|
cancel = make(chan struct{})
|
|
term = func() {
|
|
once.Do(func() {
|
|
close(cancel)
|
|
})
|
|
}
|
|
)
|
|
sourceAccountTrie, elems := makeBoundaryAccountTrie(3000)
|
|
|
|
mkSource := func(name string) *testPeer {
|
|
source := newTestPeer(name, t, term)
|
|
source.accountTrie = sourceAccountTrie
|
|
source.accountValues = elems
|
|
return source
|
|
}
|
|
syncer := setupSyncer(
|
|
mkSource("peer-a"),
|
|
mkSource("peer-b"),
|
|
)
|
|
done := checkStall(t, term)
|
|
if err := syncer.Sync(sourceAccountTrie.Hash(), cancel); err != nil {
|
|
t.Fatalf("sync failed: %v", err)
|
|
}
|
|
close(done)
|
|
verifyTrie(syncer.db, sourceAccountTrie.Hash(), t)
|
|
}
|
|
|
|
// TestSyncNoStorageAndOneCappedPeer tests sync using accounts and no storage, where one peer is
|
|
// consistently returning very small results
|
|
func TestSyncNoStorageAndOneCappedPeer(t *testing.T) {
|
|
t.Parallel()
|
|
|
|
var (
|
|
once sync.Once
|
|
cancel = make(chan struct{})
|
|
term = func() {
|
|
once.Do(func() {
|
|
close(cancel)
|
|
})
|
|
}
|
|
)
|
|
sourceAccountTrie, elems := makeAccountTrieNoStorage(3000)
|
|
|
|
mkSource := func(name string, slow bool) *testPeer {
|
|
source := newTestPeer(name, t, term)
|
|
source.accountTrie = sourceAccountTrie
|
|
source.accountValues = elems
|
|
|
|
if slow {
|
|
source.accountRequestHandler = starvingAccountRequestHandler
|
|
}
|
|
return source
|
|
}
|
|
|
|
syncer := setupSyncer(
|
|
mkSource("nice-a", false),
|
|
mkSource("nice-b", false),
|
|
mkSource("nice-c", false),
|
|
mkSource("capped", true),
|
|
)
|
|
done := checkStall(t, term)
|
|
if err := syncer.Sync(sourceAccountTrie.Hash(), cancel); err != nil {
|
|
t.Fatalf("sync failed: %v", err)
|
|
}
|
|
close(done)
|
|
verifyTrie(syncer.db, sourceAccountTrie.Hash(), t)
|
|
}
|
|
|
|
// TestSyncNoStorageAndOneCodeCorruptPeer has one peer which doesn't deliver
|
|
// code requests properly.
|
|
func TestSyncNoStorageAndOneCodeCorruptPeer(t *testing.T) {
|
|
t.Parallel()
|
|
|
|
var (
|
|
once sync.Once
|
|
cancel = make(chan struct{})
|
|
term = func() {
|
|
once.Do(func() {
|
|
close(cancel)
|
|
})
|
|
}
|
|
)
|
|
sourceAccountTrie, elems := makeAccountTrieNoStorage(3000)
|
|
|
|
mkSource := func(name string, codeFn codeHandlerFunc) *testPeer {
|
|
source := newTestPeer(name, t, term)
|
|
source.accountTrie = sourceAccountTrie
|
|
source.accountValues = elems
|
|
source.codeRequestHandler = codeFn
|
|
return source
|
|
}
|
|
// One is capped, one is corrupt. If we don't use a capped one, there's a 50%
|
|
// chance that the full set of codes requested are sent only to the
|
|
// non-corrupt peer, which delivers everything in one go, and makes the
|
|
// test moot
|
|
syncer := setupSyncer(
|
|
mkSource("capped", cappedCodeRequestHandler),
|
|
mkSource("corrupt", corruptCodeRequestHandler),
|
|
)
|
|
done := checkStall(t, term)
|
|
if err := syncer.Sync(sourceAccountTrie.Hash(), cancel); err != nil {
|
|
t.Fatalf("sync failed: %v", err)
|
|
}
|
|
close(done)
|
|
verifyTrie(syncer.db, sourceAccountTrie.Hash(), t)
|
|
}
|
|
|
|
func TestSyncNoStorageAndOneAccountCorruptPeer(t *testing.T) {
|
|
t.Parallel()
|
|
|
|
var (
|
|
once sync.Once
|
|
cancel = make(chan struct{})
|
|
term = func() {
|
|
once.Do(func() {
|
|
close(cancel)
|
|
})
|
|
}
|
|
)
|
|
sourceAccountTrie, elems := makeAccountTrieNoStorage(3000)
|
|
|
|
mkSource := func(name string, accFn accountHandlerFunc) *testPeer {
|
|
source := newTestPeer(name, t, term)
|
|
source.accountTrie = sourceAccountTrie
|
|
source.accountValues = elems
|
|
source.accountRequestHandler = accFn
|
|
return source
|
|
}
|
|
// One is capped, one is corrupt. If we don't use a capped one, there's a 50%
|
|
// chance that the full set of codes requested are sent only to the
|
|
// non-corrupt peer, which delivers everything in one go, and makes the
|
|
// test moot
|
|
syncer := setupSyncer(
|
|
mkSource("capped", defaultAccountRequestHandler),
|
|
mkSource("corrupt", corruptAccountRequestHandler),
|
|
)
|
|
done := checkStall(t, term)
|
|
if err := syncer.Sync(sourceAccountTrie.Hash(), cancel); err != nil {
|
|
t.Fatalf("sync failed: %v", err)
|
|
}
|
|
close(done)
|
|
verifyTrie(syncer.db, sourceAccountTrie.Hash(), t)
|
|
}
|
|
|
|
// TestSyncNoStorageAndOneCodeCappedPeer has one peer which delivers code hashes
|
|
// one by one
|
|
func TestSyncNoStorageAndOneCodeCappedPeer(t *testing.T) {
|
|
t.Parallel()
|
|
|
|
var (
|
|
once sync.Once
|
|
cancel = make(chan struct{})
|
|
term = func() {
|
|
once.Do(func() {
|
|
close(cancel)
|
|
})
|
|
}
|
|
)
|
|
sourceAccountTrie, elems := makeAccountTrieNoStorage(3000)
|
|
|
|
mkSource := func(name string, codeFn codeHandlerFunc) *testPeer {
|
|
source := newTestPeer(name, t, term)
|
|
source.accountTrie = sourceAccountTrie
|
|
source.accountValues = elems
|
|
source.codeRequestHandler = codeFn
|
|
return source
|
|
}
|
|
// Count how many times it's invoked. Remember, there are only 8 unique hashes,
|
|
// so it shouldn't be more than that
|
|
var counter int
|
|
syncer := setupSyncer(
|
|
mkSource("capped", func(t *testPeer, id uint64, hashes []common.Hash, max uint64) error {
|
|
counter++
|
|
return cappedCodeRequestHandler(t, id, hashes, max)
|
|
}),
|
|
)
|
|
done := checkStall(t, term)
|
|
if err := syncer.Sync(sourceAccountTrie.Hash(), cancel); err != nil {
|
|
t.Fatalf("sync failed: %v", err)
|
|
}
|
|
close(done)
|
|
|
|
// There are only 8 unique hashes, and 3K accounts. However, the code
|
|
// deduplication is per request batch. If it were a perfect global dedup,
|
|
// we would expect only 8 requests. If there were no dedup, there would be
|
|
// 3k requests.
|
|
// We expect somewhere below 100 requests for these 8 unique hashes. But
|
|
// the number can be flaky, so don't limit it so strictly.
|
|
if threshold := 100; counter > threshold {
|
|
t.Logf("Error, expected < %d invocations, got %d", threshold, counter)
|
|
}
|
|
verifyTrie(syncer.db, sourceAccountTrie.Hash(), t)
|
|
}
|
|
|
|
// TestSyncBoundaryStorageTrie tests sync against a few normal peers, but the
|
|
// storage trie has a few boundary elements.
|
|
func TestSyncBoundaryStorageTrie(t *testing.T) {
|
|
t.Parallel()
|
|
|
|
var (
|
|
once sync.Once
|
|
cancel = make(chan struct{})
|
|
term = func() {
|
|
once.Do(func() {
|
|
close(cancel)
|
|
})
|
|
}
|
|
)
|
|
sourceAccountTrie, elems, storageTries, storageElems := makeAccountTrieWithStorage(10, 1000, false, true)
|
|
|
|
mkSource := func(name string) *testPeer {
|
|
source := newTestPeer(name, t, term)
|
|
source.accountTrie = sourceAccountTrie
|
|
source.accountValues = elems
|
|
source.storageTries = storageTries
|
|
source.storageValues = storageElems
|
|
return source
|
|
}
|
|
syncer := setupSyncer(
|
|
mkSource("peer-a"),
|
|
mkSource("peer-b"),
|
|
)
|
|
done := checkStall(t, term)
|
|
if err := syncer.Sync(sourceAccountTrie.Hash(), cancel); err != nil {
|
|
t.Fatalf("sync failed: %v", err)
|
|
}
|
|
close(done)
|
|
verifyTrie(syncer.db, sourceAccountTrie.Hash(), t)
|
|
}
|
|
|
|
// TestSyncWithStorageAndOneCappedPeer tests sync using accounts + storage, where one peer is
|
|
// consistently returning very small results
|
|
func TestSyncWithStorageAndOneCappedPeer(t *testing.T) {
|
|
t.Parallel()
|
|
|
|
var (
|
|
once sync.Once
|
|
cancel = make(chan struct{})
|
|
term = func() {
|
|
once.Do(func() {
|
|
close(cancel)
|
|
})
|
|
}
|
|
)
|
|
sourceAccountTrie, elems, storageTries, storageElems := makeAccountTrieWithStorage(300, 1000, false, false)
|
|
|
|
mkSource := func(name string, slow bool) *testPeer {
|
|
source := newTestPeer(name, t, term)
|
|
source.accountTrie = sourceAccountTrie
|
|
source.accountValues = elems
|
|
source.storageTries = storageTries
|
|
source.storageValues = storageElems
|
|
|
|
if slow {
|
|
source.storageRequestHandler = starvingStorageRequestHandler
|
|
}
|
|
return source
|
|
}
|
|
|
|
syncer := setupSyncer(
|
|
mkSource("nice-a", false),
|
|
mkSource("slow", true),
|
|
)
|
|
done := checkStall(t, term)
|
|
if err := syncer.Sync(sourceAccountTrie.Hash(), cancel); err != nil {
|
|
t.Fatalf("sync failed: %v", err)
|
|
}
|
|
close(done)
|
|
verifyTrie(syncer.db, sourceAccountTrie.Hash(), t)
|
|
}
|
|
|
|
// TestSyncWithStorageAndCorruptPeer tests sync using accounts + storage, where one peer is
|
|
// sometimes sending bad proofs
|
|
func TestSyncWithStorageAndCorruptPeer(t *testing.T) {
|
|
t.Parallel()
|
|
|
|
var (
|
|
once sync.Once
|
|
cancel = make(chan struct{})
|
|
term = func() {
|
|
once.Do(func() {
|
|
close(cancel)
|
|
})
|
|
}
|
|
)
|
|
sourceAccountTrie, elems, storageTries, storageElems := makeAccountTrieWithStorage(100, 3000, true, false)
|
|
|
|
mkSource := func(name string, handler storageHandlerFunc) *testPeer {
|
|
source := newTestPeer(name, t, term)
|
|
source.accountTrie = sourceAccountTrie
|
|
source.accountValues = elems
|
|
source.storageTries = storageTries
|
|
source.storageValues = storageElems
|
|
source.storageRequestHandler = handler
|
|
return source
|
|
}
|
|
|
|
syncer := setupSyncer(
|
|
mkSource("nice-a", defaultStorageRequestHandler),
|
|
mkSource("nice-b", defaultStorageRequestHandler),
|
|
mkSource("nice-c", defaultStorageRequestHandler),
|
|
mkSource("corrupt", corruptStorageRequestHandler),
|
|
)
|
|
done := checkStall(t, term)
|
|
if err := syncer.Sync(sourceAccountTrie.Hash(), cancel); err != nil {
|
|
t.Fatalf("sync failed: %v", err)
|
|
}
|
|
close(done)
|
|
verifyTrie(syncer.db, sourceAccountTrie.Hash(), t)
|
|
}
|
|
|
|
func TestSyncWithStorageAndNonProvingPeer(t *testing.T) {
|
|
t.Parallel()
|
|
|
|
var (
|
|
once sync.Once
|
|
cancel = make(chan struct{})
|
|
term = func() {
|
|
once.Do(func() {
|
|
close(cancel)
|
|
})
|
|
}
|
|
)
|
|
sourceAccountTrie, elems, storageTries, storageElems := makeAccountTrieWithStorage(100, 3000, true, false)
|
|
|
|
mkSource := func(name string, handler storageHandlerFunc) *testPeer {
|
|
source := newTestPeer(name, t, term)
|
|
source.accountTrie = sourceAccountTrie
|
|
source.accountValues = elems
|
|
source.storageTries = storageTries
|
|
source.storageValues = storageElems
|
|
source.storageRequestHandler = handler
|
|
return source
|
|
}
|
|
syncer := setupSyncer(
|
|
mkSource("nice-a", defaultStorageRequestHandler),
|
|
mkSource("nice-b", defaultStorageRequestHandler),
|
|
mkSource("nice-c", defaultStorageRequestHandler),
|
|
mkSource("corrupt", noProofStorageRequestHandler),
|
|
)
|
|
done := checkStall(t, term)
|
|
if err := syncer.Sync(sourceAccountTrie.Hash(), cancel); err != nil {
|
|
t.Fatalf("sync failed: %v", err)
|
|
}
|
|
close(done)
|
|
verifyTrie(syncer.db, sourceAccountTrie.Hash(), t)
|
|
}
|
|
|
|
// TestSyncWithStorage tests basic sync using accounts + storage + code, against
|
|
// a peer who insists on delivering full storage sets _and_ proofs. This triggered
|
|
// an error, where the recipient erroneously clipped the boundary nodes, but
|
|
// did not mark the account for healing.
|
|
func TestSyncWithStorageMisbehavingProve(t *testing.T) {
|
|
t.Parallel()
|
|
var (
|
|
once sync.Once
|
|
cancel = make(chan struct{})
|
|
term = func() {
|
|
once.Do(func() {
|
|
close(cancel)
|
|
})
|
|
}
|
|
)
|
|
sourceAccountTrie, elems, storageTries, storageElems := makeAccountTrieWithStorageWithUniqueStorage(10, 30, false)
|
|
|
|
mkSource := func(name string) *testPeer {
|
|
source := newTestPeer(name, t, term)
|
|
source.accountTrie = sourceAccountTrie
|
|
source.accountValues = elems
|
|
source.storageTries = storageTries
|
|
source.storageValues = storageElems
|
|
source.storageRequestHandler = proofHappyStorageRequestHandler
|
|
return source
|
|
}
|
|
syncer := setupSyncer(mkSource("sourceA"))
|
|
if err := syncer.Sync(sourceAccountTrie.Hash(), cancel); err != nil {
|
|
t.Fatalf("sync failed: %v", err)
|
|
}
|
|
verifyTrie(syncer.db, sourceAccountTrie.Hash(), t)
|
|
}
|
|
|
|
type kv struct {
|
|
k, v []byte
|
|
}
|
|
|
|
// Some helpers for sorting
|
|
type entrySlice []*kv
|
|
|
|
func (p entrySlice) Len() int { return len(p) }
|
|
func (p entrySlice) Less(i, j int) bool { return bytes.Compare(p[i].k, p[j].k) < 0 }
|
|
func (p entrySlice) Swap(i, j int) { p[i], p[j] = p[j], p[i] }
|
|
|
|
func key32(i uint64) []byte {
|
|
key := make([]byte, 32)
|
|
binary.LittleEndian.PutUint64(key, i)
|
|
return key
|
|
}
|
|
|
|
var (
|
|
codehashes = []common.Hash{
|
|
crypto.Keccak256Hash([]byte{0}),
|
|
crypto.Keccak256Hash([]byte{1}),
|
|
crypto.Keccak256Hash([]byte{2}),
|
|
crypto.Keccak256Hash([]byte{3}),
|
|
crypto.Keccak256Hash([]byte{4}),
|
|
crypto.Keccak256Hash([]byte{5}),
|
|
crypto.Keccak256Hash([]byte{6}),
|
|
crypto.Keccak256Hash([]byte{7}),
|
|
}
|
|
)
|
|
|
|
// getCodeHash returns a pseudo-random code hash
|
|
func getCodeHash(i uint64) []byte {
|
|
h := codehashes[int(i)%len(codehashes)]
|
|
return common.CopyBytes(h[:])
|
|
}
|
|
|
|
// getCodeByHash convenience function to lookup the code from the code hash
|
|
func getCodeByHash(hash common.Hash) []byte {
|
|
if hash == emptyCode {
|
|
return nil
|
|
}
|
|
for i, h := range codehashes {
|
|
if h == hash {
|
|
return []byte{byte(i)}
|
|
}
|
|
}
|
|
return nil
|
|
}
|
|
|
|
// makeAccountTrieNoStorage spits out a trie, along with the leafs
|
|
func makeAccountTrieNoStorage(n int) (*trie.Trie, entrySlice) {
|
|
db := trie.NewDatabase(rawdb.NewMemoryDatabase())
|
|
accTrie := trie.NewEmpty(db)
|
|
var entries entrySlice
|
|
for i := uint64(1); i <= uint64(n); i++ {
|
|
value, _ := rlp.EncodeToBytes(&types.StateAccount{
|
|
Nonce: i,
|
|
Balance: big.NewInt(int64(i)),
|
|
Root: emptyRoot,
|
|
CodeHash: getCodeHash(i),
|
|
})
|
|
key := key32(i)
|
|
elem := &kv{key, value}
|
|
accTrie.Update(elem.k, elem.v)
|
|
entries = append(entries, elem)
|
|
}
|
|
sort.Sort(entries)
|
|
accTrie.Commit(nil)
|
|
return accTrie, entries
|
|
}
|
|
|
|
// makeBoundaryAccountTrie constructs an account trie. Instead of filling
|
|
// accounts normally, this function will fill a few accounts which have
|
|
// boundary hash.
|
|
func makeBoundaryAccountTrie(n int) (*trie.Trie, entrySlice) {
|
|
var (
|
|
entries entrySlice
|
|
boundaries []common.Hash
|
|
|
|
db = trie.NewDatabase(rawdb.NewMemoryDatabase())
|
|
trie = trie.NewEmpty(db)
|
|
)
|
|
// Initialize boundaries
|
|
var next common.Hash
|
|
step := new(big.Int).Sub(
|
|
new(big.Int).Div(
|
|
new(big.Int).Exp(common.Big2, common.Big256, nil),
|
|
big.NewInt(int64(accountConcurrency)),
|
|
), common.Big1,
|
|
)
|
|
for i := 0; i < accountConcurrency; i++ {
|
|
last := common.BigToHash(new(big.Int).Add(next.Big(), step))
|
|
if i == accountConcurrency-1 {
|
|
last = common.HexToHash("0xffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff")
|
|
}
|
|
boundaries = append(boundaries, last)
|
|
next = common.BigToHash(new(big.Int).Add(last.Big(), common.Big1))
|
|
}
|
|
// Fill boundary accounts
|
|
for i := 0; i < len(boundaries); i++ {
|
|
value, _ := rlp.EncodeToBytes(&types.StateAccount{
|
|
Nonce: uint64(0),
|
|
Balance: big.NewInt(int64(i)),
|
|
Root: emptyRoot,
|
|
CodeHash: getCodeHash(uint64(i)),
|
|
})
|
|
elem := &kv{boundaries[i].Bytes(), value}
|
|
trie.Update(elem.k, elem.v)
|
|
entries = append(entries, elem)
|
|
}
|
|
// Fill other accounts if required
|
|
for i := uint64(1); i <= uint64(n); i++ {
|
|
value, _ := rlp.EncodeToBytes(&types.StateAccount{
|
|
Nonce: i,
|
|
Balance: big.NewInt(int64(i)),
|
|
Root: emptyRoot,
|
|
CodeHash: getCodeHash(i),
|
|
})
|
|
elem := &kv{key32(i), value}
|
|
trie.Update(elem.k, elem.v)
|
|
entries = append(entries, elem)
|
|
}
|
|
sort.Sort(entries)
|
|
trie.Commit(nil)
|
|
return trie, entries
|
|
}
|
|
|
|
// makeAccountTrieWithStorageWithUniqueStorage creates an account trie where each accounts
|
|
// has a unique storage set.
|
|
func makeAccountTrieWithStorageWithUniqueStorage(accounts, slots int, code bool) (*trie.Trie, entrySlice, map[common.Hash]*trie.Trie, map[common.Hash]entrySlice) {
|
|
var (
|
|
db = trie.NewDatabase(rawdb.NewMemoryDatabase())
|
|
accTrie = trie.NewEmpty(db)
|
|
entries entrySlice
|
|
storageTries = make(map[common.Hash]*trie.Trie)
|
|
storageEntries = make(map[common.Hash]entrySlice)
|
|
)
|
|
// Create n accounts in the trie
|
|
for i := uint64(1); i <= uint64(accounts); i++ {
|
|
key := key32(i)
|
|
codehash := emptyCode[:]
|
|
if code {
|
|
codehash = getCodeHash(i)
|
|
}
|
|
// Create a storage trie
|
|
stTrie, stEntries := makeStorageTrieWithSeed(common.BytesToHash(key), uint64(slots), i, db)
|
|
stRoot := stTrie.Hash()
|
|
stTrie.Commit(nil)
|
|
value, _ := rlp.EncodeToBytes(&types.StateAccount{
|
|
Nonce: i,
|
|
Balance: big.NewInt(int64(i)),
|
|
Root: stRoot,
|
|
CodeHash: codehash,
|
|
})
|
|
elem := &kv{key, value}
|
|
accTrie.Update(elem.k, elem.v)
|
|
entries = append(entries, elem)
|
|
|
|
storageTries[common.BytesToHash(key)] = stTrie
|
|
storageEntries[common.BytesToHash(key)] = stEntries
|
|
}
|
|
sort.Sort(entries)
|
|
|
|
accTrie.Commit(nil)
|
|
return accTrie, entries, storageTries, storageEntries
|
|
}
|
|
|
|
// makeAccountTrieWithStorage spits out a trie, along with the leafs
|
|
func makeAccountTrieWithStorage(accounts, slots int, code, boundary bool) (*trie.Trie, entrySlice, map[common.Hash]*trie.Trie, map[common.Hash]entrySlice) {
|
|
var (
|
|
db = trie.NewDatabase(rawdb.NewMemoryDatabase())
|
|
accTrie = trie.NewEmpty(db)
|
|
entries entrySlice
|
|
storageTries = make(map[common.Hash]*trie.Trie)
|
|
storageEntries = make(map[common.Hash]entrySlice)
|
|
)
|
|
// Create n accounts in the trie
|
|
for i := uint64(1); i <= uint64(accounts); i++ {
|
|
key := key32(i)
|
|
codehash := emptyCode[:]
|
|
if code {
|
|
codehash = getCodeHash(i)
|
|
}
|
|
// Make a storage trie
|
|
var (
|
|
stTrie *trie.Trie
|
|
stEntries entrySlice
|
|
)
|
|
if boundary {
|
|
stTrie, stEntries = makeBoundaryStorageTrie(common.BytesToHash(key), slots, db)
|
|
} else {
|
|
stTrie, stEntries = makeStorageTrieWithSeed(common.BytesToHash(key), uint64(slots), 0, db)
|
|
}
|
|
stRoot := stTrie.Hash()
|
|
stTrie.Commit(nil)
|
|
|
|
value, _ := rlp.EncodeToBytes(&types.StateAccount{
|
|
Nonce: i,
|
|
Balance: big.NewInt(int64(i)),
|
|
Root: stRoot,
|
|
CodeHash: codehash,
|
|
})
|
|
elem := &kv{key, value}
|
|
accTrie.Update(elem.k, elem.v)
|
|
entries = append(entries, elem)
|
|
// we reuse the same one for all accounts
|
|
storageTries[common.BytesToHash(key)] = stTrie
|
|
storageEntries[common.BytesToHash(key)] = stEntries
|
|
}
|
|
sort.Sort(entries)
|
|
accTrie.Commit(nil)
|
|
return accTrie, entries, storageTries, storageEntries
|
|
}
|
|
|
|
// makeStorageTrieWithSeed fills a storage trie with n items, returning the
|
|
// not-yet-committed trie and the sorted entries. The seeds can be used to ensure
|
|
// that tries are unique.
|
|
func makeStorageTrieWithSeed(owner common.Hash, n, seed uint64, db *trie.Database) (*trie.Trie, entrySlice) {
|
|
trie, _ := trie.New(owner, common.Hash{}, db)
|
|
var entries entrySlice
|
|
for i := uint64(1); i <= n; i++ {
|
|
// store 'x' at slot 'x'
|
|
slotValue := key32(i + seed)
|
|
rlpSlotValue, _ := rlp.EncodeToBytes(common.TrimLeftZeroes(slotValue[:]))
|
|
|
|
slotKey := key32(i)
|
|
key := crypto.Keccak256Hash(slotKey[:])
|
|
|
|
elem := &kv{key[:], rlpSlotValue}
|
|
trie.Update(elem.k, elem.v)
|
|
entries = append(entries, elem)
|
|
}
|
|
sort.Sort(entries)
|
|
trie.Commit(nil)
|
|
return trie, entries
|
|
}
|
|
|
|
// makeBoundaryStorageTrie constructs a storage trie. Instead of filling
|
|
// storage slots normally, this function will fill a few slots which have
|
|
// boundary hash.
|
|
func makeBoundaryStorageTrie(owner common.Hash, n int, db *trie.Database) (*trie.Trie, entrySlice) {
|
|
var (
|
|
entries entrySlice
|
|
boundaries []common.Hash
|
|
trie, _ = trie.New(owner, common.Hash{}, db)
|
|
)
|
|
// Initialize boundaries
|
|
var next common.Hash
|
|
step := new(big.Int).Sub(
|
|
new(big.Int).Div(
|
|
new(big.Int).Exp(common.Big2, common.Big256, nil),
|
|
big.NewInt(int64(accountConcurrency)),
|
|
), common.Big1,
|
|
)
|
|
for i := 0; i < accountConcurrency; i++ {
|
|
last := common.BigToHash(new(big.Int).Add(next.Big(), step))
|
|
if i == accountConcurrency-1 {
|
|
last = common.HexToHash("0xffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff")
|
|
}
|
|
boundaries = append(boundaries, last)
|
|
next = common.BigToHash(new(big.Int).Add(last.Big(), common.Big1))
|
|
}
|
|
// Fill boundary slots
|
|
for i := 0; i < len(boundaries); i++ {
|
|
key := boundaries[i]
|
|
val := []byte{0xde, 0xad, 0xbe, 0xef}
|
|
|
|
elem := &kv{key[:], val}
|
|
trie.Update(elem.k, elem.v)
|
|
entries = append(entries, elem)
|
|
}
|
|
// Fill other slots if required
|
|
for i := uint64(1); i <= uint64(n); i++ {
|
|
slotKey := key32(i)
|
|
key := crypto.Keccak256Hash(slotKey[:])
|
|
|
|
slotValue := key32(i)
|
|
rlpSlotValue, _ := rlp.EncodeToBytes(common.TrimLeftZeroes(slotValue[:]))
|
|
|
|
elem := &kv{key[:], rlpSlotValue}
|
|
trie.Update(elem.k, elem.v)
|
|
entries = append(entries, elem)
|
|
}
|
|
sort.Sort(entries)
|
|
trie.Commit(nil)
|
|
return trie, entries
|
|
}
|
|
|
|
func verifyTrie(db ethdb.KeyValueStore, root common.Hash, t *testing.T) {
|
|
t.Helper()
|
|
triedb := trie.NewDatabase(db)
|
|
accTrie, err := trie.New(common.Hash{}, root, triedb)
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
accounts, slots := 0, 0
|
|
accIt := trie.NewIterator(accTrie.NodeIterator(nil))
|
|
for accIt.Next() {
|
|
var acc struct {
|
|
Nonce uint64
|
|
Balance *big.Int
|
|
Root common.Hash
|
|
CodeHash []byte
|
|
}
|
|
if err := rlp.DecodeBytes(accIt.Value, &acc); err != nil {
|
|
log.Crit("Invalid account encountered during snapshot creation", "err", err)
|
|
}
|
|
accounts++
|
|
if acc.Root != emptyRoot {
|
|
storeTrie, err := trie.NewSecure(common.BytesToHash(accIt.Key), acc.Root, triedb)
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
storeIt := trie.NewIterator(storeTrie.NodeIterator(nil))
|
|
for storeIt.Next() {
|
|
slots++
|
|
}
|
|
if err := storeIt.Err; err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
}
|
|
}
|
|
if err := accIt.Err; err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
t.Logf("accounts: %d, slots: %d", accounts, slots)
|
|
}
|
|
|
|
// TestSyncAccountPerformance tests how efficient the snap algo is at minimizing
|
|
// state healing
|
|
func TestSyncAccountPerformance(t *testing.T) {
|
|
// Set the account concurrency to 1. This _should_ result in the
|
|
// range root to become correct, and there should be no healing needed
|
|
defer func(old int) { accountConcurrency = old }(accountConcurrency)
|
|
accountConcurrency = 1
|
|
|
|
var (
|
|
once sync.Once
|
|
cancel = make(chan struct{})
|
|
term = func() {
|
|
once.Do(func() {
|
|
close(cancel)
|
|
})
|
|
}
|
|
)
|
|
sourceAccountTrie, elems := makeAccountTrieNoStorage(100)
|
|
|
|
mkSource := func(name string) *testPeer {
|
|
source := newTestPeer(name, t, term)
|
|
source.accountTrie = sourceAccountTrie
|
|
source.accountValues = elems
|
|
return source
|
|
}
|
|
src := mkSource("source")
|
|
syncer := setupSyncer(src)
|
|
if err := syncer.Sync(sourceAccountTrie.Hash(), cancel); err != nil {
|
|
t.Fatalf("sync failed: %v", err)
|
|
}
|
|
verifyTrie(syncer.db, sourceAccountTrie.Hash(), t)
|
|
// The trie root will always be requested, since it is added when the snap
|
|
// sync cycle starts. When popping the queue, we do not look it up again.
|
|
// Doing so would bring this number down to zero in this artificial testcase,
|
|
// but only add extra IO for no reason in practice.
|
|
if have, want := src.nTrienodeRequests, 1; have != want {
|
|
fmt.Print(src.Stats())
|
|
t.Errorf("trie node heal requests wrong, want %d, have %d", want, have)
|
|
}
|
|
}
|
|
|
|
func TestSlotEstimation(t *testing.T) {
|
|
for i, tc := range []struct {
|
|
last common.Hash
|
|
count int
|
|
want uint64
|
|
}{
|
|
{
|
|
// Half the space
|
|
common.HexToHash("0x7fffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff"),
|
|
100,
|
|
100,
|
|
},
|
|
{
|
|
// 1 / 16th
|
|
common.HexToHash("0x0fffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff"),
|
|
100,
|
|
1500,
|
|
},
|
|
{
|
|
// Bit more than 1 / 16th
|
|
common.HexToHash("0x1000000000000000000000000000000000000000000000000000000000000000"),
|
|
100,
|
|
1499,
|
|
},
|
|
{
|
|
// Almost everything
|
|
common.HexToHash("0xF000000000000000000000000000000000000000000000000000000000000000"),
|
|
100,
|
|
6,
|
|
},
|
|
{
|
|
// Almost nothing -- should lead to error
|
|
common.HexToHash("0x0000000000000000000000000000000000000000000000000000000000000001"),
|
|
1,
|
|
0,
|
|
},
|
|
{
|
|
// Nothing -- should lead to error
|
|
common.Hash{},
|
|
100,
|
|
0,
|
|
},
|
|
} {
|
|
have, _ := estimateRemainingSlots(tc.count, tc.last)
|
|
if want := tc.want; have != want {
|
|
t.Errorf("test %d: have %d want %d", i, have, want)
|
|
}
|
|
}
|
|
}
|