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
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
|
// Copyright 2015 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 downloader contains the manual full chain synchronisation.
package downloader
import (
"errors"
"fmt"
"math/big"
"sync"
"sync/atomic"
"time"
ethereum "github.com/dexon-foundation/dexon"
"github.com/dexon-foundation/dexon/common"
"github.com/dexon-foundation/dexon/core/rawdb"
"github.com/dexon-foundation/dexon/core/types"
"github.com/dexon-foundation/dexon/ethdb"
"github.com/dexon-foundation/dexon/event"
"github.com/dexon-foundation/dexon/log"
"github.com/dexon-foundation/dexon/metrics"
"github.com/dexon-foundation/dexon/params"
)
var (
MaxHashFetch = 512 // Amount of hashes to be fetched per retrieval request
MaxBlockFetch = 128 // Amount of blocks to be fetched per retrieval request
MaxHeaderFetch = 192 // Amount of block headers to be fetched per retrieval request
MaxSkeletonSize = 128 // Number of header fetches to need for a skeleton assembly
MaxBodyFetch = 128 // Amount of block bodies to be fetched per retrieval request
MaxReceiptFetch = 256 // Amount of transaction receipts to allow fetching per request
MaxStateFetch = 384 // Amount of node state values to allow fetching per request
MaxForkAncestry = 3 * params.EpochDuration // Maximum chain reorganisation
rttMinEstimate = 2 * time.Second // Minimum round-trip time to target for download requests
rttMaxEstimate = 20 * time.Second // Maximum round-trip time to target for download requests
rttMinConfidence = 0.1 // Worse confidence factor in our estimated RTT value
ttlScaling = 3 // Constant scaling factor for RTT -> TTL conversion
ttlLimit = time.Minute // Maximum TTL allowance to prevent reaching crazy timeouts
qosTuningPeers = 5 // Number of peers to tune based on (best peers)
qosConfidenceCap = 10 // Number of peers above which not to modify RTT confidence
qosTuningImpact = 0.25 // Impact that a new tuning target has on the previous value
maxQueuedHeaders = 32 * 1024 // [eth/62] Maximum number of headers to queue for import (DOS protection)
maxHeadersProcess = 2048 // Number of header download results to import at once into the chain
maxResultsProcess = 2048 // Number of content download results to import at once into the chain
reorgProtThreshold = 48 // Threshold number of recent blocks to disable mini reorg protection
reorgProtHeaderDelay = 2 // Number of headers to delay delivering to cover mini reorgs
fsHeaderCheckFrequency = 100 // Verification frequency of the downloaded headers during fast sync
fsHeaderSafetyNet = 2048 // Number of headers to discard in case a chain violation is detected
fsHeaderForceVerify = 24 // Number of headers to verify before and after the pivot to accept it
fsHeaderContCheck = 3 * time.Second // Time interval to check for header continuations during state download
fsMinFullBlocks = 64 // Number of blocks to retrieve fully even in fast sync
)
var (
errBusy = errors.New("busy")
errUnknownPeer = errors.New("peer is unknown or unhealthy")
errBadPeer = errors.New("action from bad peer ignored")
errStallingPeer = errors.New("peer is stalling")
errNoPeers = errors.New("no peers to keep download active")
errTimeout = errors.New("timeout")
errEmptyHeaderSet = errors.New("empty header set by peer")
errPeersUnavailable = errors.New("no peers available or all tried for download")
errInvalidAncestor = errors.New("retrieved ancestor is invalid")
errInvalidChain = errors.New("retrieved hash chain is invalid")
errInvalidBlock = errors.New("retrieved block is invalid")
errInvalidBody = errors.New("retrieved block body is invalid")
errInvalidReceipt = errors.New("retrieved receipt is invalid")
errCancelBlockFetch = errors.New("block download canceled (requested)")
errCancelHeaderFetch = errors.New("block header download canceled (requested)")
errCancelBodyFetch = errors.New("block body download canceled (requested)")
errCancelReceiptFetch = errors.New("receipt download canceled (requested)")
errCancelStateFetch = errors.New("state data download canceled (requested)")
errCancelHeaderProcessing = errors.New("header processing canceled (requested)")
errCancelContentProcessing = errors.New("content processing canceled (requested)")
errNoSyncActive = errors.New("no sync active")
errTooOld = errors.New("peer doesn't speak recent enough protocol version (need version >= 62)")
)
type Downloader struct {
mode SyncMode // Synchronisation mode defining the strategy used (per sync cycle)
mux *event.TypeMux // Event multiplexer to announce sync operation events
queue *queue // Scheduler for selecting the hashes to download
peers *peerSet // Set of active peers from which download can proceed
stateDB ethdb.Database
rttEstimate uint64 // Round trip time to target for download requests
rttConfidence uint64 // Confidence in the estimated RTT (unit: millionths to allow atomic ops)
// Statistics
syncStatsChainOrigin uint64 // Origin block number where syncing started at
syncStatsChainHeight uint64 // Highest block number known when syncing started
syncStatsState stateSyncStats
syncStatsLock sync.RWMutex // Lock protecting the sync stats fields
lightchain LightChain
blockchain BlockChain
// Callbacks
dropPeer peerDropFn // Drops a peer for misbehaving
// Status
synchroniseMock func(id string, hash common.Hash) error // Replacement for synchronise during testing
synchronising int32
notified int32
committed int32
// Channels
headerCh chan dataPack // [eth/62] Channel receiving inbound block headers
bodyCh chan dataPack // [eth/62] Channel receiving inbound block bodies
receiptCh chan dataPack // [eth/63] Channel receiving inbound receipts
bodyWakeCh chan bool // [eth/62] Channel to signal the block body fetcher of new tasks
receiptWakeCh chan bool // [eth/63] Channel to signal the receipt fetcher of new tasks
headerProcCh chan []*types.Header // [eth/62] Channel to feed the header processor new tasks
// for stateFetcher
stateSyncStart chan *stateSync
trackStateReq chan *stateReq
stateCh chan dataPack // [eth/63] Channel receiving inbound node state data
// Cancellation and termination
cancelPeer string // Identifier of the peer currently being used as the master (cancel on drop)
cancelCh chan struct{} // Channel to cancel mid-flight syncs
cancelLock sync.RWMutex // Lock to protect the cancel channel and peer in delivers
cancelWg sync.WaitGroup // Make sure all fetcher goroutines have exited.
quitCh chan struct{} // Quit channel to signal termination
quitLock sync.RWMutex // Lock to prevent double closes
// Testing hooks
syncInitHook func(uint64, uint64) // Method to call upon initiating a new sync run
bodyFetchHook func([]*types.Header) // Method to call upon starting a block body fetch
receiptFetchHook func([]*types.Header) // Method to call upon starting a receipt fetch
chainInsertHook func([]*fetchResult) // Method to call upon inserting a chain of blocks (possibly in multiple invocations)
}
// LightChain encapsulates functions required to synchronise a light chain.
type LightChain interface {
// HasHeader verifies a header's presence in the local chain.
HasHeader(common.Hash, uint64) bool
// GetHeaderByHash retrieves a header from the local chain.
GetHeaderByHash(common.Hash) *types.Header
// CurrentHeader retrieves the head header from the local chain.
CurrentHeader() *types.Header
// GetTd returns the total difficulty of a local block.
GetTd(common.Hash, uint64) *big.Int
// InsertHeaderChain inserts a batch of headers into the local chain.
InsertHeaderChain([]*types.Header, int) (int, error)
// Rollback removes a few recently added elements from the local chain.
Rollback([]common.Hash)
}
// BlockChain encapsulates functions required to sync a (full or fast) blockchain.
type BlockChain interface {
LightChain
// HasBlock verifies a block's presence in the local chain.
HasBlock(common.Hash, uint64) bool
// GetBlockByHash retrieves a block from the local chain.
GetBlockByHash(common.Hash) *types.Block
// CurrentBlock retrieves the head block from the local chain.
CurrentBlock() *types.Block
// CurrentFastBlock retrieves the head fast block from the local chain.
CurrentFastBlock() *types.Block
// FastSyncCommitHead directly commits the head block to a certain entity.
FastSyncCommitHead(common.Hash) error
// InsertChain inserts a batch of blocks into the local chain.
InsertChain(types.Blocks) (int, error)
// InsertReceiptChain inserts a batch of receipts into the local chain.
InsertReceiptChain(types.Blocks, []types.Receipts) (int, error)
}
// New creates a new downloader to fetch hashes and blocks from remote peers.
func New(mode SyncMode, stateDb ethdb.Database, mux *event.TypeMux, chain BlockChain, lightchain LightChain, dropPeer peerDropFn) *Downloader {
if lightchain == nil {
lightchain = chain
}
dl := &Downloader{
mode: mode,
stateDB: stateDb,
mux: mux,
queue: newQueue(),
peers: newPeerSet(),
rttEstimate: uint64(rttMaxEstimate),
rttConfidence: uint64(1000000),
blockchain: chain,
lightchain: lightchain,
dropPeer: dropPeer,
headerCh: make(chan dataPack, 1),
bodyCh: make(chan dataPack, 1),
receiptCh: make(chan dataPack, 1),
bodyWakeCh: make(chan bool, 1),
receiptWakeCh: make(chan bool, 1),
headerProcCh: make(chan []*types.Header, 1),
quitCh: make(chan struct{}),
stateCh: make(chan dataPack),
stateSyncStart: make(chan *stateSync),
syncStatsState: stateSyncStats{
processed: rawdb.ReadFastTrieProgress(stateDb),
},
trackStateReq: make(chan *stateReq),
}
go dl.qosTuner()
go dl.stateFetcher()
return dl
}
// Progress retrieves the synchronisation boundaries, specifically the origin
// block where synchronisation started at (may have failed/suspended); the block
// or header sync is currently at; and the latest known block which the sync targets.
//
// In addition, during the state download phase of fast synchronisation the number
// of processed and the total number of known states are also returned. Otherwise
// these are zero.
func (d *Downloader) Progress() ethereum.SyncProgress {
// Lock the current stats and return the progress
d.syncStatsLock.RLock()
defer d.syncStatsLock.RUnlock()
current := uint64(0)
switch d.mode {
case FullSync:
current = d.blockchain.CurrentBlock().NumberU64()
case FastSync:
current = d.blockchain.CurrentFastBlock().NumberU64()
case LightSync:
current = d.lightchain.CurrentHeader().Number.Uint64()
}
return ethereum.SyncProgress{
StartingBlock: d.syncStatsChainOrigin,
CurrentBlock: current,
HighestBlock: d.syncStatsChainHeight,
PulledStates: d.syncStatsState.processed,
KnownStates: d.syncStatsState.processed + d.syncStatsState.pending,
}
}
// Synchronising returns whether the downloader is currently retrieving blocks.
func (d *Downloader) Synchronising() bool {
return atomic.LoadInt32(&d.synchronising) > 0
}
// RegisterPeer injects a new download peer into the set of block source to be
// used for fetching hashes and blocks from.
func (d *Downloader) RegisterPeer(id string, version int, peer Peer) error {
logger := log.New("peer", id)
logger.Trace("Registering sync peer")
if err := d.peers.Register(newPeerConnection(id, version, peer, logger)); err != nil {
logger.Error("Failed to register sync peer", "err", err)
return err
}
d.qosReduceConfidence()
return nil
}
// RegisterLightPeer injects a light client peer, wrapping it so it appears as a regular peer.
func (d *Downloader) RegisterLightPeer(id string, version int, peer LightPeer) error {
return d.RegisterPeer(id, version, &lightPeerWrapper{peer})
}
// UnregisterPeer remove a peer from the known list, preventing any action from
// the specified peer. An effort is also made to return any pending fetches into
// the queue.
func (d *Downloader) UnregisterPeer(id string) error {
// Unregister the peer from the active peer set and revoke any fetch tasks
logger := log.New("peer", id)
logger.Trace("Unregistering sync peer")
if err := d.peers.Unregister(id); err != nil {
logger.Error("Failed to unregister sync peer", "err", err)
return err
}
d.queue.Revoke(id)
// If this peer was the master peer, abort sync immediately
d.cancelLock.RLock()
master := id == d.cancelPeer
d.cancelLock.RUnlock()
if master {
d.cancel()
}
return nil
}
// Synchronise tries to sync up our local block chain with a remote peer, both
// adding various sanity checks as well as wrapping it with various log entries.
func (d *Downloader) Synchronise(id string, head common.Hash, td *big.Int, mode SyncMode) error {
err := d.synchronise(id, head, td, mode)
switch err {
case nil:
case errBusy:
case errTimeout, errBadPeer, errStallingPeer,
errEmptyHeaderSet, errPeersUnavailable, errTooOld,
errInvalidAncestor, errInvalidChain:
log.Warn("Synchronisation failed, dropping peer", "peer", id, "err", err)
if d.dropPeer == nil {
// The dropPeer method is nil when `--copydb` is used for a local copy.
// Timeouts can occur if e.g. compaction hits at the wrong time, and can be ignored
log.Warn("Downloader wants to drop peer, but peerdrop-function is not set", "peer", id)
} else {
d.dropPeer(id)
}
default:
log.Warn("Synchronisation failed, retrying", "err", err)
}
return err
}
// synchronise will select the peer and use it for synchronising. If an empty string is given
// it will use the best peer possible and synchronize if its TD is higher than our own. If any of the
// checks fail an error will be returned. This method is synchronous
func (d *Downloader) synchronise(id string, hash common.Hash, td *big.Int, mode SyncMode) error {
// Mock out the synchronisation if testing
if d.synchroniseMock != nil {
return d.synchroniseMock(id, hash)
}
// Make sure only one goroutine is ever allowed past this point at once
if !atomic.CompareAndSwapInt32(&d.synchronising, 0, 1) {
return errBusy
}
defer atomic.StoreInt32(&d.synchronising, 0)
// Post a user notification of the sync (only once per session)
if atomic.CompareAndSwapInt32(&d.notified, 0, 1) {
log.Info("Block synchronisation started")
}
// Reset the queue, peer set and wake channels to clean any internal leftover state
d.queue.Reset()
d.peers.Reset()
for _, ch := range []chan bool{d.bodyWakeCh, d.receiptWakeCh} {
select {
case <-ch:
default:
}
}
for _, ch := range []chan dataPack{d.headerCh, d.bodyCh, d.receiptCh} {
for empty := false; !empty; {
select {
case <-ch:
default:
empty = true
}
}
}
for empty := false; !empty; {
select {
case <-d.headerProcCh:
default:
empty = true
}
}
// Create cancel channel for aborting mid-flight and mark the master peer
d.cancelLock.Lock()
d.cancelCh = make(chan struct{})
d.cancelPeer = id
d.cancelLock.Unlock()
defer d.Cancel() // No matter what, we can't leave the cancel channel open
// Set the requested sync mode, unless it's forbidden
d.mode = mode
// Retrieve the origin peer and initiate the downloading process
p := d.peers.Peer(id)
if p == nil {
return errUnknownPeer
}
return d.syncWithPeer(p, hash, td)
}
// syncWithPeer starts a block synchronization based on the hash chain from the
// specified peer and head hash.
func (d *Downloader) syncWithPeer(p *peerConnection, hash common.Hash, td *big.Int) (err error) {
d.mux.Post(StartEvent{})
defer func() {
// reset on error
if err != nil {
d.mux.Post(FailedEvent{err})
} else {
d.mux.Post(DoneEvent{})
}
}()
if p.version < 62 {
return errTooOld
}
log.Debug("Synchronising with the network", "peer", p.id, "eth", p.version, "head", hash, "td", td, "mode", d.mode)
defer func(start time.Time) {
log.Debug("Synchronisation terminated", "elapsed", time.Since(start))
}(time.Now())
// Look up the sync boundaries: the common ancestor and the target block
latest, err := d.fetchHeight(p)
if err != nil {
return err
}
height := latest.Number.Uint64()
origin, err := d.findAncestor(p, height)
if err != nil {
return err
}
d.syncStatsLock.Lock()
if d.syncStatsChainHeight <= origin || d.syncStatsChainOrigin > origin {
d.syncStatsChainOrigin = origin
}
d.syncStatsChainHeight = height
d.syncStatsLock.Unlock()
// Ensure our origin point is below any fast sync pivot point
pivot := uint64(0)
if d.mode == FastSync {
if height <= uint64(fsMinFullBlocks) {
origin = 0
} else {
pivot = height - uint64(fsMinFullBlocks)
if pivot <= origin {
origin = pivot - 1
}
}
}
d.committed = 1
if d.mode == FastSync && pivot != 0 {
d.committed = 0
}
// Initiate the sync using a concurrent header and content retrieval algorithm
d.queue.Prepare(origin+1, d.mode)
if d.syncInitHook != nil {
d.syncInitHook(origin, height)
}
fetchers := []func() error{
func() error { return d.fetchHeaders(p, origin+1, pivot) }, // Headers are always retrieved
func() error { return d.fetchBodies(origin + 1) }, // Bodies are retrieved during normal and fast sync
func() error { return d.fetchReceipts(origin + 1) }, // Receipts are retrieved during fast sync
func() error { return d.processHeaders(origin+1, pivot, td) },
}
if d.mode == FastSync {
fetchers = append(fetchers, func() error { return d.processFastSyncContent(latest) })
} else if d.mode == FullSync {
fetchers = append(fetchers, d.processFullSyncContent)
}
return d.spawnSync(fetchers)
}
// spawnSync runs d.process and all given fetcher functions to completion in
// separate goroutines, returning the first error that appears.
func (d *Downloader) spawnSync(fetchers []func() error) error {
errc := make(chan error, len(fetchers))
d.cancelWg.Add(len(fetchers))
for _, fn := range fetchers {
fn := fn
go func() { defer d.cancelWg.Done(); errc <- fn() }()
}
// Wait for the first error, then terminate the others.
var err error
for i := 0; i < len(fetchers); i++ {
if i == len(fetchers)-1 {
// Close the queue when all fetchers have exited.
// This will cause the block processor to end when
// it has processed the queue.
d.queue.Close()
}
if err = <-errc; err != nil {
break
}
}
d.queue.Close()
d.Cancel()
return err
}
// cancel aborts all of the operations and resets the queue. However, cancel does
// not wait for the running download goroutines to finish. This method should be
// used when cancelling the downloads from inside the downloader.
func (d *Downloader) cancel() {
// Close the current cancel channel
d.cancelLock.Lock()
if d.cancelCh != nil {
select {
case <-d.cancelCh:
// Channel was already closed
default:
close(d.cancelCh)
}
}
d.cancelLock.Unlock()
}
// Cancel aborts all of the operations and waits for all download goroutines to
// finish before returning.
func (d *Downloader) Cancel() {
d.cancel()
d.cancelWg.Wait()
}
// Terminate interrupts the downloader, canceling all pending operations.
// The downloader cannot be reused after calling Terminate.
func (d *Downloader) Terminate() {
// Close the termination channel (make sure double close is allowed)
d.quitLock.Lock()
select {
case <-d.quitCh:
default:
close(d.quitCh)
}
d.quitLock.Unlock()
// Cancel any pending download requests
d.Cancel()
}
// fetchHeight retrieves the head header of the remote peer to aid in estimating
// the total time a pending synchronisation would take.
func (d *Downloader) fetchHeight(p *peerConnection) (*types.Header, error) {
p.log.Debug("Retrieving remote chain height")
// Request the advertised remote head block and wait for the response
head, _ := p.peer.Head()
go p.peer.RequestHeadersByHash(head, 1, 0, false)
ttl := d.requestTTL()
timeout := time.After(ttl)
for {
select {
case <-d.cancelCh:
return nil, errCancelBlockFetch
case packet := <-d.headerCh:
// Discard anything not from the origin peer
if packet.PeerId() != p.id {
log.Debug("Received headers from incorrect peer", "peer", packet.PeerId())
break
}
// Make sure the peer actually gave something valid
headers := packet.(*headerPack).headers
if len(headers) != 1 {
p.log.Debug("Multiple headers for single request", "headers", len(headers))
return nil, errBadPeer
}
head := headers[0]
p.log.Debug("Remote head header identified", "number", head.Number, "hash", head.Hash())
return head, nil
case <-timeout:
p.log.Debug("Waiting for head header timed out", "elapsed", ttl)
return nil, errTimeout
case <-d.bodyCh:
case <-d.receiptCh:
// Out of bounds delivery, ignore
}
}
}
// findAncestor tries to locate the common ancestor link of the local chain and
// a remote peers blockchain. In the general case when our node was in sync and
// on the correct chain, checking the top N links should already get us a match.
// In the rare scenario when we ended up on a long reorganisation (i.e. none of
// the head links match), we do a binary search to find the common ancestor.
func (d *Downloader) findAncestor(p *peerConnection, height uint64) (uint64, error) {
// Figure out the valid ancestor range to prevent rewrite attacks
floor, ceil := int64(-1), d.lightchain.CurrentHeader().Number.Uint64()
if d.mode == FullSync {
ceil = d.blockchain.CurrentBlock().NumberU64()
} else if d.mode == FastSync {
ceil = d.blockchain.CurrentFastBlock().NumberU64()
}
if ceil >= MaxForkAncestry {
floor = int64(ceil - MaxForkAncestry)
}
p.log.Debug("Looking for common ancestor", "local", ceil, "remote", height)
// Request the topmost blocks to short circuit binary ancestor lookup
head := ceil
if head > height {
head = height
}
from := int64(head) - int64(MaxHeaderFetch)
if from < 0 {
from = 0
}
// Span out with 15 block gaps into the future to catch bad head reports
limit := 2 * MaxHeaderFetch / 16
count := 1 + int((int64(ceil)-from)/16)
if count > limit {
count = limit
}
go p.peer.RequestHeadersByNumber(uint64(from), count, 15, false)
// Wait for the remote response to the head fetch
number, hash := uint64(0), common.Hash{}
ttl := d.requestTTL()
timeout := time.After(ttl)
for finished := false; !finished; {
select {
case <-d.cancelCh:
return 0, errCancelHeaderFetch
case packet := <-d.headerCh:
// Discard anything not from the origin peer
if packet.PeerId() != p.id {
log.Debug("Received headers from incorrect peer", "peer", packet.PeerId())
break
}
// Make sure the peer actually gave something valid
headers := packet.(*headerPack).headers
if len(headers) == 0 {
p.log.Warn("Empty head header set")
return 0, errEmptyHeaderSet
}
// Make sure the peer's reply conforms to the request
for i := 0; i < len(headers); i++ {
if number := headers[i].Number.Int64(); number != from+int64(i)*16 {
p.log.Warn("Head headers broke chain ordering", "index", i, "requested", from+int64(i)*16, "received", number)
return 0, errInvalidChain
}
}
// Check if a common ancestor was found
finished = true
for i := len(headers) - 1; i >= 0; i-- {
// Skip any headers that underflow/overflow our requested set
if headers[i].Number.Int64() < from || headers[i].Number.Uint64() > ceil {
continue
}
// Otherwise check if we already know the header or not
h := headers[i].Hash()
n := headers[i].Number.Uint64()
if (d.mode == FullSync && d.blockchain.HasBlock(h, n)) || (d.mode != FullSync && d.lightchain.HasHeader(h, n)) {
number, hash = n, h
// If every header is known, even future ones, the peer straight out lied about its head
if number > height && i == limit-1 {
p.log.Warn("Lied about chain head", "reported", height, "found", number)
return 0, errStallingPeer
}
break
}
}
case <-timeout:
p.log.Debug("Waiting for head header timed out", "elapsed", ttl)
return 0, errTimeout
case <-d.bodyCh:
case <-d.receiptCh:
// Out of bounds delivery, ignore
}
}
// If the head fetch already found an ancestor, return
if hash != (common.Hash{}) {
if int64(number) <= floor {
p.log.Warn("Ancestor below allowance", "number", number, "hash", hash, "allowance", floor)
return 0, errInvalidAncestor
}
p.log.Debug("Found common ancestor", "number", number, "hash", hash)
return number, nil
}
// Ancestor not found, we need to binary search over our chain
start, end := uint64(0), head
if floor > 0 {
start = uint64(floor)
}
for start+1 < end {
// Split our chain interval in two, and request the hash to cross check
check := (start + end) / 2
ttl := d.requestTTL()
timeout := time.After(ttl)
go p.peer.RequestHeadersByNumber(check, 1, 0, false)
// Wait until a reply arrives to this request
for arrived := false; !arrived; {
select {
case <-d.cancelCh:
return 0, errCancelHeaderFetch
case packer := <-d.headerCh:
// Discard anything not from the origin peer
if packer.PeerId() != p.id {
log.Debug("Received headers from incorrect peer", "peer", packer.PeerId())
break
}
// Make sure the peer actually gave something valid
headers := packer.(*headerPack).headers
if len(headers) != 1 {
p.log.Debug("Multiple headers for single request", "headers", len(headers))
return 0, errBadPeer
}
arrived = true
// Modify the search interval based on the response
h := headers[0].Hash()
n := headers[0].Number.Uint64()
if (d.mode == FullSync && !d.blockchain.HasBlock(h, n)) || (d.mode != FullSync && !d.lightchain.HasHeader(h, n)) {
end = check
break
}
header := d.lightchain.GetHeaderByHash(h) // Independent of sync mode, header surely exists
if header.Number.Uint64() != check {
p.log.Debug("Received non requested header", "number", header.Number, "hash", header.Hash(), "request", check)
return 0, errBadPeer
}
start = check
hash = h
case <-timeout:
p.log.Debug("Waiting for search header timed out", "elapsed", ttl)
return 0, errTimeout
case <-d.bodyCh:
case <-d.receiptCh:
// Out of bounds delivery, ignore
}
}
}
// Ensure valid ancestry and return
if int64(start) <= floor {
p.log.Warn("Ancestor below allowance", "number", start, "hash", hash, "allowance", floor)
return 0, errInvalidAncestor
}
p.log.Debug("Found common ancestor", "number", start, "hash", hash)
return start, nil
}
// fetchHeaders keeps retrieving headers concurrently from the number
// requested, until no more are returned, potentially throttling on the way. To
// facilitate concurrency but still protect against malicious nodes sending bad
// headers, we construct a header chain skeleton using the "origin" peer we are
// syncing with, and fill in the missing headers using anyone else. Headers from
// other peers are only accepted if they map cleanly to the skeleton. If no one
// can fill in the skeleton - not even the origin peer - it's assumed invalid and
// the origin is dropped.
func (d *Downloader) fetchHeaders(p *peerConnection, from uint64, pivot uint64) error {
p.log.Debug("Directing header downloads", "origin", from)
defer p.log.Debug("Header download terminated")
// Create a timeout timer, and the associated header fetcher
skeleton := true // Skeleton assembly phase or finishing up
request := time.Now() // time of the last skeleton fetch request
timeout := time.NewTimer(0) // timer to dump a non-responsive active peer
<-timeout.C // timeout channel should be initially empty
defer timeout.Stop()
var ttl time.Duration
getHeaders := func(from uint64) {
request = time.Now()
ttl = d.requestTTL()
timeout.Reset(ttl)
if skeleton {
p.log.Trace("Fetching skeleton headers", "count", MaxHeaderFetch, "from", from)
go p.peer.RequestHeadersByNumber(from+uint64(MaxHeaderFetch)-1, MaxSkeletonSize, MaxHeaderFetch-1, false)
} else {
p.log.Trace("Fetching full headers", "count", MaxHeaderFetch, "from", from)
go p.peer.RequestHeadersByNumber(from, MaxHeaderFetch, 0, false)
}
}
// Start pulling the header chain skeleton until all is done
getHeaders(from)
for {
select {
case <-d.cancelCh:
return errCancelHeaderFetch
case packet := <-d.headerCh:
// Make sure the active peer is giving us the skeleton headers
if packet.PeerId() != p.id {
log.Debug("Received skeleton from incorrect peer", "peer", packet.PeerId())
break
}
headerReqTimer.UpdateSince(request)
timeout.Stop()
// If the skeleton's finished, pull any remaining head headers directly from the origin
if packet.Items() == 0 && skeleton {
skeleton = false
getHeaders(from)
continue
}
// If no more headers are inbound, notify the content fetchers and return
if packet.Items() == 0 {
// Don't abort header fetches while the pivot is downloading
if atomic.LoadInt32(&d.committed) == 0 && pivot <= from {
p.log.Debug("No headers, waiting for pivot commit")
select {
case <-time.After(fsHeaderContCheck):
getHeaders(from)
continue
case <-d.cancelCh:
return errCancelHeaderFetch
}
}
// Pivot done (or not in fast sync) and no more headers, terminate the process
p.log.Debug("No more headers available")
select {
case d.headerProcCh <- nil:
return nil
case <-d.cancelCh:
return errCancelHeaderFetch
}
}
headers := packet.(*headerPack).headers
// If we received a skeleton batch, resolve internals concurrently
if skeleton {
filled, proced, err := d.fillHeaderSkeleton(from, headers)
if err != nil {
p.log.Debug("Skeleton chain invalid", "err", err)
return errInvalidChain
}
headers = filled[proced:]
from += uint64(proced)
} else {
// If we're closing in on the chain head, but haven't yet reached it, delay
// the last few headers so mini reorgs on the head don't cause invalid hash
// chain errors.
if n := len(headers); n > 0 {
// Retrieve the current head we're at
head := uint64(0)
if d.mode == LightSync {
head = d.lightchain.CurrentHeader().Number.Uint64()
} else {
head = d.blockchain.CurrentFastBlock().NumberU64()
if full := d.blockchain.CurrentBlock().NumberU64(); head < full {
head = full
}
}
// If the head is way older than this batch, delay the last few headers
if head+uint64(reorgProtThreshold) < headers[n-1].Number.Uint64() {
delay := reorgProtHeaderDelay
if delay > n {
delay = n
}
headers = headers[:n-delay]
}
}
}
// Insert all the new headers and fetch the next batch
if len(headers) > 0 {
p.log.Trace("Scheduling new headers", "count", len(headers), "from", from)
select {
case d.headerProcCh <- headers:
case <-d.cancelCh:
return errCancelHeaderFetch
}
from += uint64(len(headers))
getHeaders(from)
} else {
// No headers delivered, or all of them being delayed, sleep a bit and retry
p.log.Trace("All headers delayed, waiting")
select {
case <-time.After(fsHeaderContCheck):
getHeaders(from)
continue
case <-d.cancelCh:
return errCancelHeaderFetch
}
}
case <-timeout.C:
if d.dropPeer == nil {
// The dropPeer method is nil when `--copydb` is used for a local copy.
// Timeouts can occur if e.g. compaction hits at the wrong time, and can be ignored
p.log.Warn("Downloader wants to drop peer, but peerdrop-function is not set", "peer", p.id)
break
}
// Header retrieval timed out, consider the peer bad and drop
p.log.Debug("Header request timed out", "elapsed", ttl)
headerTimeoutMeter.Mark(1)
d.dropPeer(p.id)
// Finish the sync gracefully instead of dumping the gathered data though
for _, ch := range []chan bool{d.bodyWakeCh, d.receiptWakeCh} {
select {
case ch <- false:
case <-d.cancelCh:
}
}
select {
case d.headerProcCh <- nil:
case <-d.cancelCh:
}
return errBadPeer
}
}
}
// fillHeaderSkeleton concurrently retrieves headers from all our available peers
// and maps them to the provided skeleton header chain.
//
// Any partial results from the beginning of the skeleton is (if possible) forwarded
// immediately to the header processor to keep the rest of the pipeline full even
// in the case of header stalls.
//
// The method returns the entire filled skeleton and also the number of headers
// already forwarded for processing.
func (d *Downloader) fillHeaderSkeleton(from uint64, skeleton []*types.Header) ([]*types.Header, int, error) {
log.Debug("Filling up skeleton", "from", from)
d.queue.ScheduleSkeleton(from, skeleton)
var (
deliver = func(packet dataPack) (int, error) {
pack := packet.(*headerPack)
return d.queue.DeliverHeaders(pack.peerID, pack.headers, d.headerProcCh)
}
expire = func() map[string]int { return d.queue.ExpireHeaders(d.requestTTL()) }
throttle = func() bool { return false }
reserve = func(p *peerConnection, count int) (*fetchRequest, bool, error) {
return d.queue.ReserveHeaders(p, count), false, nil
}
fetch = func(p *peerConnection, req *fetchRequest) error { return p.FetchHeaders(req.From, MaxHeaderFetch) }
capacity = func(p *peerConnection) int { return p.HeaderCapacity(d.requestRTT()) }
setIdle = func(p *peerConnection, accepted int) { p.SetHeadersIdle(accepted) }
)
err := d.fetchParts(errCancelHeaderFetch, d.headerCh, deliver, d.queue.headerContCh, expire,
d.queue.PendingHeaders, d.queue.InFlightHeaders, throttle, reserve,
nil, fetch, d.queue.CancelHeaders, capacity, d.peers.HeaderIdlePeers, setIdle, "headers")
log.Debug("Skeleton fill terminated", "err", err)
filled, proced := d.queue.RetrieveHeaders()
return filled, proced, err
}
// fetchBodies iteratively downloads the scheduled block bodies, taking any
// available peers, reserving a chunk of blocks for each, waiting for delivery
// and also periodically checking for timeouts.
func (d *Downloader) fetchBodies(from uint64) error {
log.Debug("Downloading block bodies", "origin", from)
var (
deliver = func(packet dataPack) (int, error) {
pack := packet.(*bodyPack)
return d.queue.DeliverBodies(pack.peerID, pack.transactions, pack.uncles)
}
expire = func() map[string]int { return d.queue.ExpireBodies(d.requestTTL()) }
fetch = func(p *peerConnection, req *fetchRequest) error { return p.FetchBodies(req) }
capacity = func(p *peerConnection) int { return p.BlockCapacity(d.requestRTT()) }
setIdle = func(p *peerConnection, accepted int) { p.SetBodiesIdle(accepted) }
)
err := d.fetchParts(errCancelBodyFetch, d.bodyCh, deliver, d.bodyWakeCh, expire,
d.queue.PendingBlocks, d.queue.InFlightBlocks, d.queue.ShouldThrottleBlocks, d.queue.ReserveBodies,
d.bodyFetchHook, fetch, d.queue.CancelBodies, capacity, d.peers.BodyIdlePeers, setIdle, "bodies")
log.Debug("Block body download terminated", "err", err)
return err
}
// fetchReceipts iteratively downloads the scheduled block receipts, taking any
// available peers, reserving a chunk of receipts for each, waiting for delivery
// and also periodically checking for timeouts.
func (d *Downloader) fetchReceipts(from uint64) error {
log.Debug("Downloading transaction receipts", "origin", from)
var (
deliver = func(packet dataPack) (int, error) {
pack := packet.(*receiptPack)
return d.queue.DeliverReceipts(pack.peerID, pack.receipts)
}
expire = func() map[string]int { return d.queue.ExpireReceipts(d.requestTTL()) }
fetch = func(p *peerConnection, req *fetchRequest) error { return p.FetchReceipts(req) }
capacity = func(p *peerConnection) int { return p.ReceiptCapacity(d.requestRTT()) }
setIdle = func(p *peerConnection, accepted int) { p.SetReceiptsIdle(accepted) }
)
err := d.fetchParts(errCancelReceiptFetch, d.receiptCh, deliver, d.receiptWakeCh, expire,
d.queue.PendingReceipts, d.queue.InFlightReceipts, d.queue.ShouldThrottleReceipts, d.queue.ReserveReceipts,
d.receiptFetchHook, fetch, d.queue.CancelReceipts, capacity, d.peers.ReceiptIdlePeers, setIdle, "receipts")
log.Debug("Transaction receipt download terminated", "err", err)
return err
}
// fetchParts iteratively downloads scheduled block parts, taking any available
// peers, reserving a chunk of fetch requests for each, waiting for delivery and
// also periodically checking for timeouts.
//
// As the scheduling/timeout logic mostly is the same for all downloaded data
// types, this method is used by each for data gathering and is instrumented with
// various callbacks to handle the slight differences between processing them.
//
// The instrumentation parameters:
// - errCancel: error type to return if the fetch operation is cancelled (mostly makes logging nicer)
// - deliveryCh: channel from which to retrieve downloaded data packets (merged from all concurrent peers)
// - deliver: processing callback to deliver data packets into type specific download queues (usually within `queue`)
// - wakeCh: notification channel for waking the fetcher when new tasks are available (or sync completed)
// - expire: task callback method to abort requests that took too long and return the faulty peers (traffic shaping)
// - pending: task callback for the number of requests still needing download (detect completion/non-completability)
// - inFlight: task callback for the number of in-progress requests (wait for all active downloads to finish)
// - throttle: task callback to check if the processing queue is full and activate throttling (bound memory use)
// - reserve: task callback to reserve new download tasks to a particular peer (also signals partial completions)
// - fetchHook: tester callback to notify of new tasks being initiated (allows testing the scheduling logic)
// - fetch: network callback to actually send a particular download request to a physical remote peer
// - cancel: task callback to abort an in-flight download request and allow rescheduling it (in case of lost peer)
// - capacity: network callback to retrieve the estimated type-specific bandwidth capacity of a peer (traffic shaping)
// - idle: network callback to retrieve the currently (type specific) idle peers that can be assigned tasks
// - setIdle: network callback to set a peer back to idle and update its estimated capacity (traffic shaping)
// - kind: textual label of the type being downloaded to display in log mesages
func (d *Downloader) fetchParts(errCancel error, deliveryCh chan dataPack, deliver func(dataPack) (int, error), wakeCh chan bool,
expire func() map[string]int, pending func() int, inFlight func() bool, throttle func() bool, reserve func(*peerConnection, int) (*fetchRequest, bool, error),
fetchHook func([]*types.Header), fetch func(*peerConnection, *fetchRequest) error, cancel func(*fetchRequest), capacity func(*peerConnection) int,
idle func() ([]*peerConnection, int), setIdle func(*peerConnection, int), kind string) error {
// Create a ticker to detect expired retrieval tasks
ticker := time.NewTicker(100 * time.Millisecond)
defer ticker.Stop()
update := make(chan struct{}, 1)
// Prepare the queue and fetch block parts until the block header fetcher's done
finished := false
for {
select {
case <-d.cancelCh:
return errCancel
case packet := <-deliveryCh:
// If the peer was previously banned and failed to deliver its pack
// in a reasonable time frame, ignore its message.
if peer := d.peers.Peer(packet.PeerId()); peer != nil {
// Deliver the received chunk of data and check chain validity
accepted, err := deliver(packet)
if err == errInvalidChain {
return err
}
// Unless a peer delivered something completely else than requested (usually
// caused by a timed out request which came through in the end), set it to
// idle. If the delivery's stale, the peer should have already been idled.
if err != errStaleDelivery {
setIdle(peer, accepted)
}
// Issue a log to the user to see what's going on
switch {
case err == nil && packet.Items() == 0:
peer.log.Trace("Requested data not delivered", "type", kind)
case err == nil:
peer.log.Trace("Delivered new batch of data", "type", kind, "count", packet.Stats())
default:
peer.log.Trace("Failed to deliver retrieved data", "type", kind, "err", err)
}
}
// Blocks assembled, try to update the progress
select {
case update <- struct{}{}:
default:
}
case cont := <-wakeCh:
// The header fetcher sent a continuation flag, check if it's done
if !cont {
finished = true
}
// Headers arrive, try to update the progress
select {
case update <- struct{}{}:
default:
}
case <-ticker.C:
// Sanity check update the progress
select {
case update <- struct{}{}:
default:
}
case <-update:
// Short circuit if we lost all our peers
if d.peers.Len() == 0 {
return errNoPeers
}
// Check for fetch request timeouts and demote the responsible peers
for pid, fails := range expire() {
if peer := d.peers.Peer(pid); peer != nil {
// If a lot of retrieval elements expired, we might have overestimated the remote peer or perhaps
// ourselves. Only reset to minimal throughput but don't drop just yet. If even the minimal times
// out that sync wise we need to get rid of the peer.
//
// The reason the minimum threshold is 2 is because the downloader tries to estimate the bandwidth
// and latency of a peer separately, which requires pushing the measures capacity a bit and seeing
// how response times reacts, to it always requests one more than the minimum (i.e. min 2).
if fails > 2 {
peer.log.Trace("Data delivery timed out", "type", kind)
setIdle(peer, 0)
} else {
peer.log.Debug("Stalling delivery, dropping", "type", kind)
if d.dropPeer == nil {
// The dropPeer method is nil when `--copydb` is used for a local copy.
// Timeouts can occur if e.g. compaction hits at the wrong time, and can be ignored
peer.log.Warn("Downloader wants to drop peer, but peerdrop-function is not set", "peer", pid)
} else {
d.dropPeer(pid)
}
}
}
}
// If there's nothing more to fetch, wait or terminate
if pending() == 0 {
if !inFlight() && finished {
log.Debug("Data fetching completed", "type", kind)
return nil
}
break
}
// Send a download request to all idle peers, until throttled
progressed, throttled, running := false, false, inFlight()
idles, total := idle()
for _, peer := range idles {
// Short circuit if throttling activated
if throttle() {
throttled = true
break
}
// Short circuit if there is no more available task.
if pending() == 0 {
break
}
// Reserve a chunk of fetches for a peer. A nil can mean either that
// no more headers are available, or that the peer is known not to
// have them.
request, progress, err := reserve(peer, capacity(peer))
if err != nil {
return err
}
if progress {
progressed = true
}
if request == nil {
continue
}
if request.From > 0 {
peer.log.Trace("Requesting new batch of data", "type", kind, "from", request.From)
} else {
peer.log.Trace("Requesting new batch of data", "type", kind, "count", len(request.Headers), "from", request.Headers[0].Number)
}
// Fetch the chunk and make sure any errors return the hashes to the queue
if fetchHook != nil {
fetchHook(request.Headers)
}
if err := fetch(peer, request); err != nil {
// Although we could try and make an attempt to fix this, this error really
// means that we've double allocated a fetch task to a peer. If that is the
// case, the internal state of the downloader and the queue is very wrong so
// better hard crash and note the error instead of silently accumulating into
// a much bigger issue.
panic(fmt.Sprintf("%v: %s fetch assignment failed", peer, kind))
}
running = true
}
// Make sure that we have peers available for fetching. If all peers have been tried
// and all failed throw an error
if !progressed && !throttled && !running && len(idles) == total && pending() > 0 {
return errPeersUnavailable
}
}
}
}
// processHeaders takes batches of retrieved headers from an input channel and
// keeps processing and scheduling them into the header chain and downloader's
// queue until the stream ends or a failure occurs.
func (d *Downloader) processHeaders(origin uint64, pivot uint64, td *big.Int) error {
// Keep a count of uncertain headers to roll back
rollback := []*types.Header{}
defer func() {
if len(rollback) > 0 {
// Flatten the headers and roll them back
hashes := make([]common.Hash, len(rollback))
for i, header := range rollback {
hashes[i] = header.Hash()
}
lastHeader, lastFastBlock, lastBlock := d.lightchain.CurrentHeader().Number, common.Big0, common.Big0
if d.mode != LightSync {
lastFastBlock = d.blockchain.CurrentFastBlock().Number()
lastBlock = d.blockchain.CurrentBlock().Number()
}
d.lightchain.Rollback(hashes)
curFastBlock, curBlock := common.Big0, common.Big0
if d.mode != LightSync {
curFastBlock = d.blockchain.CurrentFastBlock().Number()
curBlock = d.blockchain.CurrentBlock().Number()
}
log.Warn("Rolled back headers", "count", len(hashes),
"header", fmt.Sprintf("%d->%d", lastHeader, d.lightchain.CurrentHeader().Number),
"fast", fmt.Sprintf("%d->%d", lastFastBlock, curFastBlock),
"block", fmt.Sprintf("%d->%d", lastBlock, curBlock))
}
}()
// Wait for batches of headers to process
gotHeaders := false
for {
select {
case <-d.cancelCh:
return errCancelHeaderProcessing
case headers := <-d.headerProcCh:
// Terminate header processing if we synced up
if len(headers) == 0 {
// Notify everyone that headers are fully processed
for _, ch := range []chan bool{d.bodyWakeCh, d.receiptWakeCh} {
select {
case ch <- false:
case <-d.cancelCh:
}
}
// If no headers were retrieved at all, the peer violated its TD promise that it had a
// better chain compared to ours. The only exception is if its promised blocks were
// already imported by other means (e.g. fecher):
//
// R <remote peer>, L <local node>: Both at block 10
// R: Mine block 11, and propagate it to L
// L: Queue block 11 for import
// L: Notice that R's head and TD increased compared to ours, start sync
// L: Import of block 11 finishes
// L: Sync begins, and finds common ancestor at 11
// L: Request new headers up from 11 (R's TD was higher, it must have something)
// R: Nothing to give
if d.mode != LightSync {
head := d.blockchain.CurrentBlock()
if !gotHeaders && td.Cmp(d.blockchain.GetTd(head.Hash(), head.NumberU64())) > 0 {
return errStallingPeer
}
}
// If fast or light syncing, ensure promised headers are indeed delivered. This is
// needed to detect scenarios where an attacker feeds a bad pivot and then bails out
// of delivering the post-pivot blocks that would flag the invalid content.
//
// This check cannot be executed "as is" for full imports, since blocks may still be
// queued for processing when the header download completes. However, as long as the
// peer gave us something useful, we're already happy/progressed (above check).
if d.mode == FastSync || d.mode == LightSync {
head := d.lightchain.CurrentHeader()
if td.Cmp(d.lightchain.GetTd(head.Hash(), head.Number.Uint64())) > 0 {
return errStallingPeer
}
}
// Disable any rollback and return
rollback = nil
return nil
}
// Otherwise split the chunk of headers into batches and process them
gotHeaders = true
for len(headers) > 0 {
// Terminate if something failed in between processing chunks
select {
case <-d.cancelCh:
return errCancelHeaderProcessing
default:
}
// Select the next chunk of headers to import
limit := maxHeadersProcess
if limit > len(headers) {
limit = len(headers)
}
chunk := headers[:limit]
// In case of header only syncing, validate the chunk immediately
if d.mode == FastSync || d.mode == LightSync {
// Collect the yet unknown headers to mark them as uncertain
unknown := make([]*types.Header, 0, len(headers))
for _, header := range chunk {
if !d.lightchain.HasHeader(header.Hash(), header.Number.Uint64()) {
unknown = append(unknown, header)
}
}
// If we're importing pure headers, verify based on their recentness
frequency := fsHeaderCheckFrequency
if chunk[len(chunk)-1].Number.Uint64()+uint64(fsHeaderForceVerify) > pivot {
frequency = 1
}
if n, err := d.lightchain.InsertHeaderChain(chunk, frequency); err != nil {
// If some headers were inserted, add them too to the rollback list
if n > 0 {
rollback = append(rollback, chunk[:n]...)
}
log.Debug("Invalid header encountered", "number", chunk[n].Number, "hash", chunk[n].Hash(), "err", err)
return errInvalidChain
}
// All verifications passed, store newly found uncertain headers
rollback = append(rollback, unknown...)
if len(rollback) > fsHeaderSafetyNet {
rollback = append(rollback[:0], rollback[len(rollback)-fsHeaderSafetyNet:]...)
}
}
// Unless we're doing light chains, schedule the headers for associated content retrieval
if d.mode == FullSync || d.mode == FastSync {
// If we've reached the allowed number of pending headers, stall a bit
for d.queue.PendingBlocks() >= maxQueuedHeaders || d.queue.PendingReceipts() >= maxQueuedHeaders {
select {
case <-d.cancelCh:
return errCancelHeaderProcessing
case <-time.After(time.Second):
}
}
// Otherwise insert the headers for content retrieval
inserts := d.queue.Schedule(chunk, origin)
if len(inserts) != len(chunk) {
log.Debug("Stale headers")
return errBadPeer
}
}
headers = headers[limit:]
origin += uint64(limit)
}
// Update the highest block number we know if a higher one is found.
d.syncStatsLock.Lock()
if d.syncStatsChainHeight < origin {
d.syncStatsChainHeight = origin - 1
}
d.syncStatsLock.Unlock()
// Signal the content downloaders of the availablility of new tasks
for _, ch := range []chan bool{d.bodyWakeCh, d.receiptWakeCh} {
select {
case ch <- true:
default:
}
}
}
}
}
// processFullSyncContent takes fetch results from the queue and imports them into the chain.
func (d *Downloader) processFullSyncContent() error {
for {
results := d.queue.Results(true)
if len(results) == 0 {
return nil
}
if d.chainInsertHook != nil {
d.chainInsertHook(results)
}
if err := d.importBlockResults(results); err != nil {
return err
}
}
}
func (d *Downloader) importBlockResults(results []*fetchResult) error {
// Check for any early termination requests
if len(results) == 0 {
return nil
}
select {
case <-d.quitCh:
return errCancelContentProcessing
default:
}
// Retrieve the a batch of results to import
first, last := results[0].Header, results[len(results)-1].Header
log.Debug("Inserting downloaded chain", "items", len(results),
"firstnum", first.Number, "firsthash", first.Hash(),
"lastnum", last.Number, "lasthash", last.Hash(),
)
blocks := make([]*types.Block, len(results))
for i, result := range results {
blocks[i] = types.NewBlockWithHeader(result.Header).WithBody(result.Transactions, result.Uncles)
}
if index, err := d.blockchain.InsertChain(blocks); err != nil {
log.Debug("Downloaded item processing failed", "number", results[index].Header.Number, "hash", results[index].Header.Hash(), "err", err)
return errInvalidChain
}
return nil
}
// processFastSyncContent takes fetch results from the queue and writes them to the
// database. It also controls the synchronisation of state nodes of the pivot block.
func (d *Downloader) processFastSyncContent(latest *types.Header) error {
// Start syncing state of the reported head block. This should get us most of
// the state of the pivot block.
stateSync := d.syncState(latest.Root)
defer stateSync.Cancel()
go func() {
if err := stateSync.Wait(); err != nil && err != errCancelStateFetch {
d.queue.Close() // wake up WaitResults
}
}()
// Figure out the ideal pivot block. Note, that this goalpost may move if the
// sync takes long enough for the chain head to move significantly.
pivot := uint64(0)
if height := latest.Number.Uint64(); height > uint64(fsMinFullBlocks) {
pivot = height - uint64(fsMinFullBlocks)
}
// To cater for moving pivot points, track the pivot block and subsequently
// accumulated download results separately.
var (
oldPivot *fetchResult // Locked in pivot block, might change eventually
oldTail []*fetchResult // Downloaded content after the pivot
)
for {
// Wait for the next batch of downloaded data to be available, and if the pivot
// block became stale, move the goalpost
results := d.queue.Results(oldPivot == nil) // Block if we're not monitoring pivot staleness
if len(results) == 0 {
// If pivot sync is done, stop
if oldPivot == nil {
return stateSync.Cancel()
}
// If sync failed, stop
select {
case <-d.cancelCh:
return stateSync.Cancel()
default:
}
}
if d.chainInsertHook != nil {
d.chainInsertHook(results)
}
if oldPivot != nil {
results = append(append([]*fetchResult{oldPivot}, oldTail...), results...)
}
// Split around the pivot block and process the two sides via fast/full sync
if atomic.LoadInt32(&d.committed) == 0 {
latest = results[len(results)-1].Header
if height := latest.Number.Uint64(); height > pivot+2*uint64(fsMinFullBlocks) {
log.Warn("Pivot became stale, moving", "old", pivot, "new", height-uint64(fsMinFullBlocks))
pivot = height - uint64(fsMinFullBlocks)
}
}
P, beforeP, afterP := splitAroundPivot(pivot, results)
if err := d.commitFastSyncData(beforeP, stateSync); err != nil {
return err
}
if P != nil {
// If new pivot block found, cancel old state retrieval and restart
if oldPivot != P {
stateSync.Cancel()
stateSync = d.syncState(P.Header.Root)
defer stateSync.Cancel()
go func() {
if err := stateSync.Wait(); err != nil && err != errCancelStateFetch {
d.queue.Close() // wake up WaitResults
}
}()
oldPivot = P
}
// Wait for completion, occasionally checking for pivot staleness
select {
case <-stateSync.done:
if stateSync.err != nil {
return stateSync.err
}
if err := d.commitPivotBlock(P); err != nil {
return err
}
oldPivot = nil
case <-time.After(time.Second):
oldTail = afterP
continue
}
}
// Fast sync done, pivot commit done, full import
if err := d.importBlockResults(afterP); err != nil {
return err
}
}
}
func splitAroundPivot(pivot uint64, results []*fetchResult) (p *fetchResult, before, after []*fetchResult) {
for _, result := range results {
num := result.Header.Number.Uint64()
switch {
case num < pivot:
before = append(before, result)
case num == pivot:
p = result
default:
after = append(after, result)
}
}
return p, before, after
}
func (d *Downloader) commitFastSyncData(results []*fetchResult, stateSync *stateSync) error {
// Check for any early termination requests
if len(results) == 0 {
return nil
}
select {
case <-d.quitCh:
return errCancelContentProcessing
case <-stateSync.done:
if err := stateSync.Wait(); err != nil {
return err
}
default:
}
// Retrieve the a batch of results to import
first, last := results[0].Header, results[len(results)-1].Header
log.Debug("Inserting fast-sync blocks", "items", len(results),
"firstnum", first.Number, "firsthash", first.Hash(),
"lastnumn", last.Number, "lasthash", last.Hash(),
)
blocks := make([]*types.Block, len(results))
receipts := make([]types.Receipts, len(results))
for i, result := range results {
blocks[i] = types.NewBlockWithHeader(result.Header).WithBody(result.Transactions, result.Uncles)
receipts[i] = result.Receipts
}
if index, err := d.blockchain.InsertReceiptChain(blocks, receipts); err != nil {
log.Debug("Downloaded item processing failed", "number", results[index].Header.Number, "hash", results[index].Header.Hash(), "err", err)
return errInvalidChain
}
return nil
}
func (d *Downloader) commitPivotBlock(result *fetchResult) error {
block := types.NewBlockWithHeader(result.Header).WithBody(result.Transactions, result.Uncles)
log.Debug("Committing fast sync pivot as new head", "number", block.Number(), "hash", block.Hash())
if _, err := d.blockchain.InsertReceiptChain([]*types.Block{block}, []types.Receipts{result.Receipts}); err != nil {
return err
}
if err := d.blockchain.FastSyncCommitHead(block.Hash()); err != nil {
return err
}
atomic.StoreInt32(&d.committed, 1)
return nil
}
// DeliverHeaders injects a new batch of block headers received from a remote
// node into the download schedule.
func (d *Downloader) DeliverHeaders(id string, headers []*types.Header) (err error) {
return d.deliver(id, d.headerCh, &headerPack{id, headers}, headerInMeter, headerDropMeter)
}
// DeliverBodies injects a new batch of block bodies received from a remote node.
func (d *Downloader) DeliverBodies(id string, transactions [][]*types.Transaction, uncles [][]*types.Header) (err error) {
return d.deliver(id, d.bodyCh, &bodyPack{id, transactions, uncles}, bodyInMeter, bodyDropMeter)
}
// DeliverReceipts injects a new batch of receipts received from a remote node.
func (d *Downloader) DeliverReceipts(id string, receipts [][]*types.Receipt) (err error) {
return d.deliver(id, d.receiptCh, &receiptPack{id, receipts}, receiptInMeter, receiptDropMeter)
}
// DeliverNodeData injects a new batch of node state data received from a remote node.
func (d *Downloader) DeliverNodeData(id string, data [][]byte) (err error) {
return d.deliver(id, d.stateCh, &statePack{id, data}, stateInMeter, stateDropMeter)
}
// deliver injects a new batch of data received from a remote node.
func (d *Downloader) deliver(id string, destCh chan dataPack, packet dataPack, inMeter, dropMeter metrics.Meter) (err error) {
// Update the delivery metrics for both good and failed deliveries
inMeter.Mark(int64(packet.Items()))
defer func() {
if err != nil {
dropMeter.Mark(int64(packet.Items()))
}
}()
// Deliver or abort if the sync is canceled while queuing
d.cancelLock.RLock()
cancel := d.cancelCh
d.cancelLock.RUnlock()
if cancel == nil {
return errNoSyncActive
}
select {
case destCh <- packet:
return nil
case <-cancel:
return errNoSyncActive
}
}
// qosTuner is the quality of service tuning loop that occasionally gathers the
// peer latency statistics and updates the estimated request round trip time.
func (d *Downloader) qosTuner() {
for {
// Retrieve the current median RTT and integrate into the previoust target RTT
rtt := time.Duration((1-qosTuningImpact)*float64(atomic.LoadUint64(&d.rttEstimate)) + qosTuningImpact*float64(d.peers.medianRTT()))
atomic.StoreUint64(&d.rttEstimate, uint64(rtt))
// A new RTT cycle passed, increase our confidence in the estimated RTT
conf := atomic.LoadUint64(&d.rttConfidence)
conf = conf + (1000000-conf)/2
atomic.StoreUint64(&d.rttConfidence, conf)
// Log the new QoS values and sleep until the next RTT
log.Debug("Recalculated downloader QoS values", "rtt", rtt, "confidence", float64(conf)/1000000.0, "ttl", d.requestTTL())
select {
case <-d.quitCh:
return
case <-time.After(rtt):
}
}
}
// qosReduceConfidence is meant to be called when a new peer joins the downloader's
// peer set, needing to reduce the confidence we have in out QoS estimates.
func (d *Downloader) qosReduceConfidence() {
// If we have a single peer, confidence is always 1
peers := uint64(d.peers.Len())
if peers == 0 {
// Ensure peer connectivity races don't catch us off guard
return
}
if peers == 1 {
atomic.StoreUint64(&d.rttConfidence, 1000000)
return
}
// If we have a ton of peers, don't drop confidence)
if peers >= uint64(qosConfidenceCap) {
return
}
// Otherwise drop the confidence factor
conf := atomic.LoadUint64(&d.rttConfidence) * (peers - 1) / peers
if float64(conf)/1000000 < rttMinConfidence {
conf = uint64(rttMinConfidence * 1000000)
}
atomic.StoreUint64(&d.rttConfidence, conf)
rtt := time.Duration(atomic.LoadUint64(&d.rttEstimate))
log.Debug("Relaxed downloader QoS values", "rtt", rtt, "confidence", float64(conf)/1000000.0, "ttl", d.requestTTL())
}
// requestRTT returns the current target round trip time for a download request
// to complete in.
//
// Note, the returned RTT is .9 of the actually estimated RTT. The reason is that
// the downloader tries to adapt queries to the RTT, so multiple RTT values can
// be adapted to, but smaller ones are preferred (stabler download stream).
func (d *Downloader) requestRTT() time.Duration {
return time.Duration(atomic.LoadUint64(&d.rttEstimate)) * 9 / 10
}
// requestTTL returns the current timeout allowance for a single download request
// to finish under.
func (d *Downloader) requestTTL() time.Duration {
var (
rtt = time.Duration(atomic.LoadUint64(&d.rttEstimate))
conf = float64(atomic.LoadUint64(&d.rttConfidence)) / 1000000.0
)
ttl := time.Duration(ttlScaling) * time.Duration(float64(rtt)/conf)
if ttl > ttlLimit {
ttl = ttlLimit
}
return ttl
}
|