core/state: semantic journalling (part 1) (#28880)

This is a follow-up to #29520, and a preparatory PR to a more thorough
change in the journalling system.

### API methods instead of `append` operations

This PR hides the journal-implementation details away, so that the
statedb invokes methods like `JournalCreate`, instead of explicitly
appending journal-events in a list. This means that it's up to the
journal whether to implement it as a sequence of events or
aggregate/merge events.

### Snapshot-management inside the journal 

This PR also makes it so that management of valid snapshots is moved
inside the journal, exposed via the methods `Snapshot() int` and
`RevertToSnapshot(revid int, s *StateDB)`.


### SetCode

JournalSetCode journals the setting of code: it is implicit that the
previous values were "no code" and emptyCodeHash. Therefore, we can
simplify the setCode journal.

### Selfdestruct

The self-destruct journalling is a bit strange: we allow the
selfdestruct operation to be journalled several times. This makes it so
that we also are forced to store whether the account was already
destructed.

What we can do instead, is to only journal the first destruction, and
after that only journal balance-changes, but not journal the
selfdestruct itself.

This simplifies the journalling, so that internals about state
management does not leak into the journal-API.

### Preimages

Preimages were, for some reason, integrated into the journal management,
despite not being a consensus-critical data structure. This PR undoes
that.

---------

Co-authored-by: Gary Rong <[email protected]>

Author: holiman

core/state/journal.go

@@ -17,12 +17,21 @@
 package state
 
 import (
+	"fmt"
 	"maps"
+	"slices"
+	"sort"
 
 	"github.com/ethereum/go-ethereum/common"
+	"github.com/ethereum/go-ethereum/core/types"
 	"github.com/holiman/uint256"
 )
 
+type revision struct {
+	id           int
+	journalIndex int
+}
+
 // journalEntry is a modification entry in the state change journal that can be
 // reverted on demand.
 type journalEntry interface {
@@ -42,6 +51,9 @@ type journalEntry interface {
 type journal struct {
 	entries []journalEntry         // Current changes tracked by the journal
 	dirties map[common.Address]int // Dirty accounts and the number of changes
+
+	validRevisions []revision
+	nextRevisionId int
 }
 
 // newJournal creates a new initialized journal.
@@ -51,6 +63,40 @@ func newJournal() *journal {
 	}
 }
 
+// reset clears the journal, after this operation the journal can be used anew.
+// It is semantically similar to calling 'newJournal', but the underlying slices
+// can be reused.
+func (j *journal) reset() {
+	j.entries = j.entries[:0]
+	j.validRevisions = j.validRevisions[:0]
+	clear(j.dirties)
+	j.nextRevisionId = 0
+}
+
+// snapshot returns an identifier for the current revision of the state.
+func (j *journal) snapshot() int {
+	id := j.nextRevisionId
+	j.nextRevisionId++
+	j.validRevisions = append(j.validRevisions, revision{id, j.length()})
+	return id
+}
+
+// revertToSnapshot reverts all state changes made since the given revision.
+func (j *journal) revertToSnapshot(revid int, s *StateDB) {
+	// Find the snapshot in the stack of valid snapshots.
+	idx := sort.Search(len(j.validRevisions), func(i int) bool {
+		return j.validRevisions[i].id >= revid
+	})
+	if idx == len(j.validRevisions) || j.validRevisions[idx].id != revid {
+		panic(fmt.Errorf("revision id %v cannot be reverted", revid))
+	}
+	snapshot := j.validRevisions[idx].journalIndex
+
+	// Replay the journal to undo changes and remove invalidated snapshots
+	j.revert(s, snapshot)
+	j.validRevisions = j.validRevisions[:idx]
+}
+
 // append inserts a new modification entry to the end of the change journal.
 func (j *journal) append(entry journalEntry) {
 	j.entries = append(j.entries, entry)
@@ -95,11 +141,90 @@ func (j *journal) copy() *journal {
 		entries = append(entries, j.entries[i].copy())
 	}
 	return &journal{
-		entries: entries,
-		dirties: maps.Clone(j.dirties),
+		entries:        entries,
+		dirties:        maps.Clone(j.dirties),
+		validRevisions: slices.Clone(j.validRevisions),
+		nextRevisionId: j.nextRevisionId,
+	}
+}
+
+func (j *journal) logChange(txHash common.Hash) {
+	j.append(addLogChange{txhash: txHash})
+}
+
+func (j *journal) createObject(addr common.Address) {
+	j.append(createObjectChange{account: &addr})
+}
+
+func (j *journal) createContract(addr common.Address) {
+	j.append(createContractChange{account: addr})
+}
+
+func (j *journal) destruct(addr common.Address) {
+	j.append(selfDestructChange{account: &addr})
+}
+
+func (j *journal) storageChange(addr common.Address, key, prev, origin common.Hash) {
+	j.append(storageChange{
+		account:   &addr,
+		key:       key,
+		prevvalue: prev,
+		origvalue: origin,
+	})
+}
+
+func (j *journal) transientStateChange(addr common.Address, key, prev common.Hash) {
+	j.append(transientStorageChange{
+		account:  &addr,
+		key:      key,
+		prevalue: prev,
+	})
+}
+
+func (j *journal) refundChange(previous uint64) {
+	j.append(refundChange{prev: previous})
+}
+
+func (j *journal) balanceChange(addr common.Address, previous *uint256.Int) {
+	j.append(balanceChange{
+		account: &addr,
+		prev:    previous.Clone(),
+	})
+}
+
+func (j *journal) setCode(address common.Address) {
+	j.append(codeChange{account: &address})
+}
+
+func (j *journal) nonceChange(address common.Address, prev uint64) {
+	j.append(nonceChange{
+		account: &address,
+		prev:    prev,
+	})
+}
+
+func (j *journal) touchChange(address common.Address) {
+	j.append(touchChange{
+		account: &address,
+	})
+	if address == ripemd {
+		// Explicitly put it in the dirty-cache, which is otherwise generated from
+		// flattened journals.
+		j.dirty(address)
 	}
 }
 
+func (j *journal) accessListAddAccount(addr common.Address) {
+	j.append(accessListAddAccountChange{&addr})
+}
+
+func (j *journal) accessListAddSlot(addr common.Address, slot common.Hash) {
+	j.append(accessListAddSlotChange{
+		address: &addr,
+		slot:    &slot,
+	})
+}
+
 type (
 	// Changes to the account trie.
 	createObjectChange struct {
@@ -114,9 +239,7 @@ type (
 	}
 
 	selfDestructChange struct {
-		account     *common.Address
-		prev        bool // whether account had already self-destructed
-		prevbalance *uint256.Int
+		account *common.Address
 	}
 
 	// Changes to individual accounts.
@@ -135,8 +258,7 @@ type (
 		origvalue common.Hash
 	}
 	codeChange struct {
-		account            *common.Address
-		prevcode, prevhash []byte
+		account *common.Address
 	}
 
 	// Changes to other state values.
@@ -146,9 +268,6 @@ type (
 	addLogChange struct {
 		txhash common.Hash
 	}
-	addPreimageChange struct {
-		hash common.Hash
-	}
 	touchChange struct {
 		account *common.Address
 	}
@@ -200,8 +319,7 @@ func (ch createContractChange) copy() journalEntry {
 func (ch selfDestructChange) revert(s *StateDB) {
 	obj := s.getStateObject(*ch.account)
 	if obj != nil {
-		obj.selfDestructed = ch.prev
-		obj.setBalance(ch.prevbalance)
+		obj.selfDestructed = false
 	}
 }
 
@@ -211,9 +329,7 @@ func (ch selfDestructChange) dirtied() *common.Address {
 
 func (ch selfDestructChange) copy() journalEntry {
 	return selfDestructChange{
-		account:     ch.account,
-		prev:        ch.prev,
-		prevbalance: new(uint256.Int).Set(ch.prevbalance),
+		account: ch.account,
 	}
 }
 
@@ -263,19 +379,15 @@ func (ch nonceChange) copy() journalEntry {
 }
 
 func (ch codeChange) revert(s *StateDB) {
-	s.getStateObject(*ch.account).setCode(common.BytesToHash(ch.prevhash), ch.prevcode)
+	s.getStateObject(*ch.account).setCode(types.EmptyCodeHash, nil)
 }
 
 func (ch codeChange) dirtied() *common.Address {
 	return ch.account
 }
 
 func (ch codeChange) copy() journalEntry {
-	return codeChange{
-		account:  ch.account,
-		prevhash: common.CopyBytes(ch.prevhash),
-		prevcode: common.CopyBytes(ch.prevcode),
-	}
+	return codeChange{account: ch.account}
 }
 
 func (ch storageChange) revert(s *StateDB) {
@@ -344,20 +456,6 @@ func (ch addLogChange) copy() journalEntry {
 	}
 }
 
-func (ch addPreimageChange) revert(s *StateDB) {
-	delete(s.preimages, ch.hash)
-}
-
-func (ch addPreimageChange) dirtied() *common.Address {
-	return nil
-}
-
-func (ch addPreimageChange) copy() journalEntry {
-	return addPreimageChange{
-		hash: ch.hash,
-	}
-}
-
 func (ch accessListAddAccountChange) revert(s *StateDB) {
 	/*
 		One important invariant here, is that whenever a (addr, slot) is added, if the

core/state/state_object.go

@@ -114,14 +114,7 @@ func (s *stateObject) markSelfdestructed() {
 }
 
 func (s *stateObject) touch() {
-	s.db.journal.append(touchChange{
-		account: &s.address,
-	})
-	if s.address == ripemd {
-		// Explicitly put it in the dirty-cache, which is otherwise generated from
-		// flattened journals.
-		s.db.journal.dirty(s.address)
-	}
+	s.db.journal.touchChange(s.address)
 }
 
 // getTrie returns the associated storage trie. The trie will be opened if it's
@@ -252,16 +245,11 @@ func (s *stateObject) SetState(key, value common.Hash) {
 		return
 	}
 	// New value is different, update and journal the change
-	s.db.journal.append(storageChange{
-		account:   &s.address,
-		key:       key,
-		prevvalue: prev,
-		origvalue: origin,
-	})
+	s.db.journal.storageChange(s.address, key, prev, origin)
+	s.setState(key, value, origin)
 	if s.db.logger != nil && s.db.logger.OnStorageChange != nil {
 		s.db.logger.OnStorageChange(s.address, key, prev, value)
 	}
-	s.setState(key, value, origin)
 }
 
 // setState updates a value in account dirty storage. The dirtiness will be
@@ -511,10 +499,7 @@ func (s *stateObject) SubBalance(amount *uint256.Int, reason tracing.BalanceChan
 }
 
 func (s *stateObject) SetBalance(amount *uint256.Int, reason tracing.BalanceChangeReason) {
-	s.db.journal.append(balanceChange{
-		account: &s.address,
-		prev:    new(uint256.Int).Set(s.data.Balance),
-	})
+	s.db.journal.balanceChange(s.address, s.data.Balance)
 	if s.db.logger != nil && s.db.logger.OnBalanceChange != nil {
 		s.db.logger.OnBalanceChange(s.address, s.Balance().ToBig(), amount.ToBig(), reason)
 	}
@@ -590,14 +575,10 @@ func (s *stateObject) CodeSize() int {
 }
 
 func (s *stateObject) SetCode(codeHash common.Hash, code []byte) {
-	prevcode := s.Code()
-	s.db.journal.append(codeChange{
-		account:  &s.address,
-		prevhash: s.CodeHash(),
-		prevcode: prevcode,
-	})
+	s.db.journal.setCode(s.address)
 	if s.db.logger != nil && s.db.logger.OnCodeChange != nil {
-		s.db.logger.OnCodeChange(s.address, common.BytesToHash(s.CodeHash()), prevcode, codeHash, code)
+		// TODO remove prevcode from this callback
+		s.db.logger.OnCodeChange(s.address, common.BytesToHash(s.CodeHash()), nil, codeHash, code)
 	}
 	s.setCode(codeHash, code)
 }
@@ -609,10 +590,7 @@ func (s *stateObject) setCode(codeHash common.Hash, code []byte) {
 }
 
 func (s *stateObject) SetNonce(nonce uint64) {
-	s.db.journal.append(nonceChange{
-		account: &s.address,
-		prev:    s.data.Nonce,
-	})
+	s.db.journal.nonceChange(s.address, s.data.Nonce)
 	if s.db.logger != nil && s.db.logger.OnNonceChange != nil {
 		s.db.logger.OnNonceChange(s.address, s.data.Nonce, nonce)
 	}