1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
package logs
import (
"errors"
"fmt"
"io"
"github.com/ethereum/go-ethereum/common"
"github.com/ethereum-optimism/optimism/op-service/eth"
"github.com/ethereum-optimism/optimism/op-supervisor/supervisor/backend/db/entrydb"
"github.com/ethereum-optimism/optimism/op-supervisor/supervisor/types"
)
// logContext is a buffer on top of the DB,
// where blocks and logs can be applied to.
//
// Rules:
//
// if entry_index % 256 == 0: must be type 0. For easy binary search.
// else if end_of_block: also type 0.
// else:
// after type 0: type 1
// after type 1: type 2 iff any event and space, otherwise type 0
// after type 2: type 3 iff executing, otherwise type 2 or 0
// after type 3: type 4
// after type 4: type 2 iff any event and space, otherwise type 0
// after type 5: any
//
// Type 0 can repeat: seal the block, then start a search checkpoint, then a single canonical hash.
// Type 0 may also be used as padding: type 2 only starts when it will not be interrupted by a search checkpoint.
//
// Types (<type> = 1 byte):
// type 0: "checkpoint" <type><uint64 block number: 8 bytes><uint32 logsSince count: 4 bytes><uint64 timestamp: 8 bytes> = 21 bytes
// type 1: "canonical hash" <type><parent blockhash: 32 bytes> = 33 bytes
// type 2: "initiating event" <type><event flags: 1 byte><event-hash: 32 bytes> = 34 bytes
// type 3: "executing link" <type><chain: 4 bytes><blocknum: 8 bytes><event index: 3 bytes><uint64 timestamp: 8 bytes> = 24 bytes
// type 4: "executing check" <type><event-hash: 32 bytes> = 33 bytes
// type 5: "padding" <type><padding: 33 bytes> = 34 bytes
// other types: future compat. E.g. for linking to L1, registering block-headers as a kind of initiating-event, tracking safe-head progression, etc.
//
// Right-pad each entry that is not 34 bytes.
//
// We insert a checkpoint for every search interval and block sealing event,
// and these may overlap as the same thing.
// Such seal has logsSince == 0, i.e. wrapping up the last block and starting a fresh list of logs.
//
// event-flags: each bit represents a boolean value, currently only two are defined
// * event-flags & 0x01 - true if the initiating event has an executing link that should follow. Allows detecting when the executing link failed to write.
// event-hash: H(origin, timestamp, payloadhash); enough to check identifier matches & payload matches.
type logContext struct {
// next entry index, including the contents of `out`
nextEntryIndex entrydb.EntryIdx
// blockHash of the last sealed block.
// A block is not considered sealed until we know its block hash.
// While we process logs we keep the parent-block of said logs around as sealed block.
blockHash common.Hash
// blockNum of the last sealed block
blockNum uint64
// timestamp of the last sealed block
timestamp uint64
// number of logs since the last sealed block
logsSince uint32
// payload-hash of the log-event that was last processed. (may not be fully processed, see doneLog)
logHash common.Hash
// executing message that might exist for the current log event.
// Might be incomplete; if !logDone while we already processed the initiating event,
// then we know an executing message is still coming.
execMsg *types.ExecutingMessage
need EntryTypeFlag
// buffer of entries not yet in the DB.
// This is generated as objects are applied.
// E.g. you can build multiple hypothetical blocks with log events on top of the state,
// before flushing the entries to a DB.
// However, no entries can be read from the DB while objects are being applied.
out []Entry
}
func (l *logContext) NextIndex() entrydb.EntryIdx {
return l.nextEntryIndex
}
// SealedBlock returns the block that we are building on top of, and if it is sealed.
func (l *logContext) SealedBlock() (hash common.Hash, num uint64, ok bool) {
if !l.hasCompleteBlock() {
return common.Hash{}, 0, false
}
return l.blockHash, l.blockNum, true
}
func (l *logContext) SealedTimestamp() (timestamp uint64, ok bool) {
if !l.hasCompleteBlock() {
return 0, false
}
return l.timestamp, true
}
func (l *logContext) hasCompleteBlock() bool {
return !l.need.Any(FlagCanonicalHash)
}
func (l *logContext) hasIncompleteLog() bool {
return l.need.Any(FlagInitiatingEvent | FlagExecutingLink | FlagExecutingCheck)
}
func (l *logContext) hasReadableLog() bool {
return l.logsSince > 0 && !l.hasIncompleteLog()
}
// InitMessage returns the current initiating message, if any is available.
func (l *logContext) InitMessage() (hash common.Hash, logIndex uint32, ok bool) {
if !l.hasReadableLog() {
return common.Hash{}, 0, false
}
return l.logHash, l.logsSince - 1, true
}
// ExecMessage returns the current executing message, if any is available.
func (l *logContext) ExecMessage() *types.ExecutingMessage {
if l.hasCompleteBlock() && l.hasReadableLog() && l.execMsg != nil {
return l.execMsg
}
return nil
}
// ApplyEntry applies an entry on top of the current state.
func (l *logContext) ApplyEntry(entry Entry) error {
// Wrap processEntry to add common useful error message info
err := l.processEntry(entry)
if err != nil {
return fmt.Errorf("failed to process type %s entry at idx %d (%x): %w", entry.Type().String(), l.nextEntryIndex, entry[:], err)
}
return nil
}
// processEntry decodes and applies an entry to the state.
// Entries may not be applied if we are in the process of generating entries from objects.
// These outputs need to be flushed before inputs can be accepted.
func (l *logContext) processEntry(entry Entry) error {
if len(l.out) != 0 {
panic("can only apply without appending if the state is still empty")
}
switch entry.Type() {
case TypeSearchCheckpoint:
current, err := newSearchCheckpointFromEntry(entry)
if err != nil {
return err
}
l.blockNum = current.blockNum
l.blockHash = common.Hash{}
l.logsSince = current.logsSince
l.timestamp = current.timestamp
l.need.Add(FlagCanonicalHash)
// Log data after the block we are sealing remains to be seen
if l.logsSince == 0 {
l.logHash = common.Hash{}
l.execMsg = nil
}
case TypeCanonicalHash:
if !l.need.Any(FlagCanonicalHash) {
return errors.New("not ready for canonical hash entry, already sealed the last block")
}
canonHash, err := newCanonicalHashFromEntry(entry)
if err != nil {
return err
}
l.blockHash = canonHash.hash
l.need.Remove(FlagCanonicalHash)
case TypeInitiatingEvent:
if !l.hasCompleteBlock() {
return errors.New("did not complete block seal, cannot add log")
}
if l.hasIncompleteLog() {
return errors.New("cannot process log before last log completes")
}
evt, err := newInitiatingEventFromEntry(entry)
if err != nil {
return err
}
l.execMsg = nil // clear the old state
l.logHash = evt.logHash
if evt.hasExecMsg {
l.need.Add(FlagExecutingLink | FlagExecutingCheck)
} else {
l.logsSince += 1
}
l.need.Remove(FlagInitiatingEvent)
case TypeExecutingLink:
if !l.need.Any(FlagExecutingLink) {
return errors.New("unexpected executing-link")
}
link, err := newExecutingLinkFromEntry(entry)
if err != nil {
return err
}
l.execMsg = &types.ExecutingMessage{
Chain: types.ChainIndex(link.chain), // TODO(#11105): translate chain ID to chain index
BlockNum: link.blockNum,
LogIdx: link.logIdx,
Timestamp: link.timestamp,
Hash: common.Hash{}, // not known yet
}
l.need.Remove(FlagExecutingLink)
l.need.Add(FlagExecutingCheck)
case TypeExecutingCheck:
if l.need.Any(FlagExecutingLink) {
return errors.New("need executing link to be applied before the check part")
}
if !l.need.Any(FlagExecutingCheck) {
return errors.New("unexpected executing check")
}
link, err := newExecutingCheckFromEntry(entry)
if err != nil {
return err
}
l.execMsg.Hash = link.hash
l.need.Remove(FlagExecutingCheck)
l.logsSince += 1
case TypePadding:
if l.need.Any(FlagPadding) {
l.need.Remove(FlagPadding)
} else {
l.need.Remove(FlagPadding2)
}
default:
return fmt.Errorf("unknown entry type: %s", entry.Type())
}
l.nextEntryIndex += 1
return nil
}
// appendEntry add the entry to the output-buffer,
// and registers it as last processed entry type, and increments the next entry-index.
func (l *logContext) appendEntry(obj EntryObj) {
entry := obj.encode()
l.out = append(l.out, entry)
l.nextEntryIndex += 1
}
// infer advances the logContext in cases where complex entries contain multiple implied entries
// eg. a SearchCheckpoint implies a CannonicalHash will follow
// this also handles inserting the searchCheckpoint at the set frequency, and padding entries
func (l *logContext) infer() error {
// We force-insert a checkpoint whenever we hit the known fixed interval.
if l.nextEntryIndex%searchCheckpointFrequency == 0 {
l.need.Add(FlagSearchCheckpoint)
}
if l.need.Any(FlagSearchCheckpoint) {
l.appendEntry(newSearchCheckpoint(l.blockNum, l.logsSince, l.timestamp))
l.need.Add(FlagCanonicalHash) // always follow with a canonical hash
l.need.Remove(FlagSearchCheckpoint)
return nil
}
if l.need.Any(FlagCanonicalHash) {
l.appendEntry(newCanonicalHash(l.blockHash))
l.need.Remove(FlagCanonicalHash)
return nil
}
if l.need.Any(FlagPadding) {
l.appendEntry(paddingEntry{})
l.need.Remove(FlagPadding)
return nil
}
if l.need.Any(FlagPadding2) {
l.appendEntry(paddingEntry{})
l.need.Remove(FlagPadding2)
return nil
}
if l.need.Any(FlagInitiatingEvent) {
// If we are running out of space for log-event data,
// write some checkpoints as padding, to pass the checkpoint.
if l.execMsg != nil { // takes 3 total. Need to avoid the checkpoint.
switch l.nextEntryIndex % searchCheckpointFrequency {
case searchCheckpointFrequency - 1:
l.need.Add(FlagPadding)
return nil
case searchCheckpointFrequency - 2:
l.need.Add(FlagPadding | FlagPadding2)
return nil
}
}
evt := newInitiatingEvent(l.logHash, l.execMsg != nil)
l.appendEntry(evt)
l.need.Remove(FlagInitiatingEvent)
if l.execMsg == nil {
l.logsSince += 1
}
return nil
}
if l.need.Any(FlagExecutingLink) {
link, err := newExecutingLink(*l.execMsg)
if err != nil {
return fmt.Errorf("failed to create executing link: %w", err)
}
l.appendEntry(link)
l.need.Remove(FlagExecutingLink)
return nil
}
if l.need.Any(FlagExecutingCheck) {
l.appendEntry(newExecutingCheck(l.execMsg.Hash))
l.need.Remove(FlagExecutingCheck)
l.logsSince += 1
return nil
}
return io.EOF
}
// inferFull advances the logContext until it cannot infer any more entries.
func (l *logContext) inferFull() error {
for i := 0; i < 10; i++ {
err := l.infer()
if err == nil {
continue
}
if err == io.EOF { // wrapped io.EOF does not count.
return nil
} else {
return err
}
}
panic("hit sanity limit")
}
// forceBlock force-overwrites the state, to match the given sealed block as starting point (excl)
func (l *logContext) forceBlock(upd eth.BlockID, timestamp uint64) error {
if l.nextEntryIndex != 0 {
return errors.New("can only bootstrap on top of an empty state")
}
l.blockHash = upd.Hash
l.blockNum = upd.Number
l.timestamp = timestamp
l.logsSince = 0
l.execMsg = nil
l.logHash = common.Hash{}
l.need = 0
l.out = nil
return l.inferFull() // apply to the state as much as possible
}
// SealBlock applies a block header on top of the current state.
// This seals the state; no further logs of this block may be added with ApplyLog.
func (l *logContext) SealBlock(parent common.Hash, upd eth.BlockID, timestamp uint64) error {
// If we don't have any entries yet, allow any block to start things off
if l.nextEntryIndex != 0 {
if err := l.inferFull(); err != nil { // ensure we can start applying
return err
}
if l.blockHash != parent {
return fmt.Errorf("%w: cannot apply block %s (parent %s) on top of %s", types.ErrConflict, upd, parent, l.blockHash)
}
if l.blockHash != (common.Hash{}) && l.blockNum+1 != upd.Number {
return fmt.Errorf("%w: cannot apply block %d on top of %d", types.ErrConflict, upd.Number, l.blockNum)
}
if l.timestamp > timestamp {
return fmt.Errorf("%w: block timestamp %d must be equal or larger than current timestamp %d", types.ErrConflict, timestamp, l.timestamp)
}
}
l.blockHash = upd.Hash
l.blockNum = upd.Number
l.timestamp = timestamp
l.logsSince = 0
l.execMsg = nil
l.logHash = common.Hash{}
l.need.Add(FlagSearchCheckpoint)
return l.inferFull() // apply to the state as much as possible
}
// ApplyLog applies a log on top of the current state.
// The parent-block that the log comes after must be applied with ApplyBlock first.
func (l *logContext) ApplyLog(parentBlock eth.BlockID, logIdx uint32, logHash common.Hash, execMsg *types.ExecutingMessage) error {
if parentBlock == (eth.BlockID{}) {
return fmt.Errorf("genesis does not have logs: %w", types.ErrOutOfOrder)
}
if err := l.inferFull(); err != nil { // ensure we can start applying
return err
}
if !l.hasCompleteBlock() {
if l.blockNum == 0 {
return fmt.Errorf("%w: should not have logs in block 0", types.ErrOutOfOrder)
} else {
return errors.New("cannot append log before last known block is sealed")
}
}
// check parent block
if l.blockHash != parentBlock.Hash {
return fmt.Errorf("%w: log builds on top of block %s, but have block %s", types.ErrOutOfOrder, parentBlock, l.blockHash)
}
if l.blockNum != parentBlock.Number {
return fmt.Errorf("%w: log builds on top of block %d, but have block %d", types.ErrOutOfOrder, parentBlock.Number, l.blockNum)
}
// check if log fits on top. The length so far == the index of the next log.
if logIdx != l.logsSince {
return fmt.Errorf("%w: expected event index %d, cannot append %d", types.ErrOutOfOrder, l.logsSince, logIdx)
}
l.logHash = logHash
l.execMsg = execMsg
l.need.Add(FlagInitiatingEvent)
if execMsg != nil {
l.need.Add(FlagExecutingLink | FlagExecutingCheck)
}
return l.inferFull() // apply to the state as much as possible
}