llvm.org GIT mirror llvm / e3e43d9 lib / CodeGen / BranchFolding.cpp
e3e43d9

Tree @e3e43d9 (Download .tar.gz)

BranchFolding.cpp @e3e43d9raw · history · blame

   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
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
//===-- BranchFolding.cpp - Fold machine code branch instructions ---------===//
//
//                     The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This pass forwards branches to unconditional branches to make them branch
// directly to the target block.  This pass often results in dead MBB's, which
// it then removes.
//
// Note that this pass must be run after register allocation, it cannot handle
// SSA form. It also must handle virtual registers for targets that emit virtual
// ISA (e.g. NVPTX).
//
//===----------------------------------------------------------------------===//

#include "BranchFolding.h"
#include "llvm/ADT/STLExtras.h"
#include "llvm/ADT/SmallSet.h"
#include "llvm/ADT/Statistic.h"
#include "llvm/CodeGen/Analysis.h"
#include "llvm/CodeGen/MachineBlockFrequencyInfo.h"
#include "llvm/CodeGen/MachineBranchProbabilityInfo.h"
#include "llvm/CodeGen/MachineFunctionPass.h"
#include "llvm/CodeGen/MachineJumpTableInfo.h"
#include "llvm/CodeGen/MachineLoopInfo.h"
#include "llvm/CodeGen/MachineMemOperand.h"
#include "llvm/CodeGen/MachineModuleInfo.h"
#include "llvm/CodeGen/MachineRegisterInfo.h"
#include "llvm/CodeGen/Passes.h"
#include "llvm/CodeGen/TargetPassConfig.h"
#include "llvm/IR/DebugInfoMetadata.h"
#include "llvm/IR/Function.h"
#include "llvm/Support/CommandLine.h"
#include "llvm/Support/Debug.h"
#include "llvm/Support/ErrorHandling.h"
#include "llvm/Support/raw_ostream.h"
#include "llvm/Target/TargetInstrInfo.h"
#include "llvm/Target/TargetRegisterInfo.h"
#include "llvm/Target/TargetSubtargetInfo.h"
#include <algorithm>
using namespace llvm;

#define DEBUG_TYPE "branch-folder"

STATISTIC(NumDeadBlocks, "Number of dead blocks removed");
STATISTIC(NumBranchOpts, "Number of branches optimized");
STATISTIC(NumTailMerge , "Number of block tails merged");
STATISTIC(NumHoist     , "Number of times common instructions are hoisted");
STATISTIC(NumTailCalls,  "Number of tail calls optimized");

static cl::opt<cl::boolOrDefault> FlagEnableTailMerge("enable-tail-merge",
                              cl::init(cl::BOU_UNSET), cl::Hidden);

// Throttle for huge numbers of predecessors (compile speed problems)
static cl::opt<unsigned>
TailMergeThreshold("tail-merge-threshold",
          cl::desc("Max number of predecessors to consider tail merging"),
          cl::init(150), cl::Hidden);

// Heuristic for tail merging (and, inversely, tail duplication).
// TODO: This should be replaced with a target query.
static cl::opt<unsigned>
TailMergeSize("tail-merge-size",
          cl::desc("Min number of instructions to consider tail merging"),
                              cl::init(3), cl::Hidden);

namespace {
  /// BranchFolderPass - Wrap branch folder in a machine function pass.
  class BranchFolderPass : public MachineFunctionPass {
  public:
    static char ID;
    explicit BranchFolderPass(): MachineFunctionPass(ID) {}

    bool runOnMachineFunction(MachineFunction &MF) override;

    void getAnalysisUsage(AnalysisUsage &AU) const override {
      AU.addRequired<MachineBlockFrequencyInfo>();
      AU.addRequired<MachineBranchProbabilityInfo>();
      AU.addRequired<TargetPassConfig>();
      MachineFunctionPass::getAnalysisUsage(AU);
    }
  };
}

char BranchFolderPass::ID = 0;
char &llvm::BranchFolderPassID = BranchFolderPass::ID;

INITIALIZE_PASS(BranchFolderPass, DEBUG_TYPE,
                "Control Flow Optimizer", false, false)

bool BranchFolderPass::runOnMachineFunction(MachineFunction &MF) {
  if (skipFunction(*MF.getFunction()))
    return false;

  TargetPassConfig *PassConfig = &getAnalysis<TargetPassConfig>();
  // TailMerge can create jump into if branches that make CFG irreducible for
  // HW that requires structurized CFG.
  bool EnableTailMerge = !MF.getTarget().requiresStructuredCFG() &&
                         PassConfig->getEnableTailMerge();
  BranchFolder::MBFIWrapper MBBFreqInfo(
      getAnalysis<MachineBlockFrequencyInfo>());
  BranchFolder Folder(EnableTailMerge, /*CommonHoist=*/true, MBBFreqInfo,
                      getAnalysis<MachineBranchProbabilityInfo>());
  return Folder.OptimizeFunction(MF, MF.getSubtarget().getInstrInfo(),
                                 MF.getSubtarget().getRegisterInfo(),
                                 getAnalysisIfAvailable<MachineModuleInfo>());
}

BranchFolder::BranchFolder(bool defaultEnableTailMerge, bool CommonHoist,
                           MBFIWrapper &FreqInfo,
                           const MachineBranchProbabilityInfo &ProbInfo,
                           unsigned MinTailLength)
    : EnableHoistCommonCode(CommonHoist), MinCommonTailLength(MinTailLength),
      MBBFreqInfo(FreqInfo), MBPI(ProbInfo) {
  if (MinCommonTailLength == 0)
    MinCommonTailLength = TailMergeSize;
  switch (FlagEnableTailMerge) {
  case cl::BOU_UNSET: EnableTailMerge = defaultEnableTailMerge; break;
  case cl::BOU_TRUE: EnableTailMerge = true; break;
  case cl::BOU_FALSE: EnableTailMerge = false; break;
  }
}

void BranchFolder::RemoveDeadBlock(MachineBasicBlock *MBB) {
  assert(MBB->pred_empty() && "MBB must be dead!");
  DEBUG(dbgs() << "\nRemoving MBB: " << *MBB);

  MachineFunction *MF = MBB->getParent();
  // drop all successors.
  while (!MBB->succ_empty())
    MBB->removeSuccessor(MBB->succ_end()-1);

  // Avoid matching if this pointer gets reused.
  TriedMerging.erase(MBB);

  // Remove the block.
  MF->erase(MBB);
  FuncletMembership.erase(MBB);
  if (MLI)
    MLI->removeBlock(MBB);
}

bool BranchFolder::OptimizeFunction(MachineFunction &MF,
                                    const TargetInstrInfo *tii,
                                    const TargetRegisterInfo *tri,
                                    MachineModuleInfo *mmi,
                                    MachineLoopInfo *mli, bool AfterPlacement) {
  if (!tii) return false;

  TriedMerging.clear();

  MachineRegisterInfo &MRI = MF.getRegInfo();
  AfterBlockPlacement = AfterPlacement;
  TII = tii;
  TRI = tri;
  MMI = mmi;
  MLI = mli;
  this->MRI = &MRI;

  UpdateLiveIns = MRI.tracksLiveness() && TRI->trackLivenessAfterRegAlloc(MF);
  if (!UpdateLiveIns)
    MRI.invalidateLiveness();

  // Fix CFG.  The later algorithms expect it to be right.
  bool MadeChange = false;
  for (MachineBasicBlock &MBB : MF) {
    MachineBasicBlock *TBB = nullptr, *FBB = nullptr;
    SmallVector<MachineOperand, 4> Cond;
    if (!TII->analyzeBranch(MBB, TBB, FBB, Cond, true))
      MadeChange |= MBB.CorrectExtraCFGEdges(TBB, FBB, !Cond.empty());
  }

  // Recalculate funclet membership.
  FuncletMembership = getFuncletMembership(MF);

  bool MadeChangeThisIteration = true;
  while (MadeChangeThisIteration) {
    MadeChangeThisIteration    = TailMergeBlocks(MF);
    // No need to clean up if tail merging does not change anything after the
    // block placement.
    if (!AfterBlockPlacement || MadeChangeThisIteration)
      MadeChangeThisIteration |= OptimizeBranches(MF);
    if (EnableHoistCommonCode)
      MadeChangeThisIteration |= HoistCommonCode(MF);
    MadeChange |= MadeChangeThisIteration;
  }

  // See if any jump tables have become dead as the code generator
  // did its thing.
  MachineJumpTableInfo *JTI = MF.getJumpTableInfo();
  if (!JTI)
    return MadeChange;

  // Walk the function to find jump tables that are live.
  BitVector JTIsLive(JTI->getJumpTables().size());
  for (const MachineBasicBlock &BB : MF) {
    for (const MachineInstr &I : BB)
      for (const MachineOperand &Op : I.operands()) {
        if (!Op.isJTI()) continue;

        // Remember that this JT is live.
        JTIsLive.set(Op.getIndex());
      }
  }

  // Finally, remove dead jump tables.  This happens when the
  // indirect jump was unreachable (and thus deleted).
  for (unsigned i = 0, e = JTIsLive.size(); i != e; ++i)
    if (!JTIsLive.test(i)) {
      JTI->RemoveJumpTable(i);
      MadeChange = true;
    }

  return MadeChange;
}

//===----------------------------------------------------------------------===//
//  Tail Merging of Blocks
//===----------------------------------------------------------------------===//

/// HashMachineInstr - Compute a hash value for MI and its operands.
static unsigned HashMachineInstr(const MachineInstr &MI) {
  unsigned Hash = MI.getOpcode();
  for (unsigned i = 0, e = MI.getNumOperands(); i != e; ++i) {
    const MachineOperand &Op = MI.getOperand(i);

    // Merge in bits from the operand if easy. We can't use MachineOperand's
    // hash_code here because it's not deterministic and we sort by hash value
    // later.
    unsigned OperandHash = 0;
    switch (Op.getType()) {
    case MachineOperand::MO_Register:
      OperandHash = Op.getReg();
      break;
    case MachineOperand::MO_Immediate:
      OperandHash = Op.getImm();
      break;
    case MachineOperand::MO_MachineBasicBlock:
      OperandHash = Op.getMBB()->getNumber();
      break;
    case MachineOperand::MO_FrameIndex:
    case MachineOperand::MO_ConstantPoolIndex:
    case MachineOperand::MO_JumpTableIndex:
      OperandHash = Op.getIndex();
      break;
    case MachineOperand::MO_GlobalAddress:
    case MachineOperand::MO_ExternalSymbol:
      // Global address / external symbol are too hard, don't bother, but do
      // pull in the offset.
      OperandHash = Op.getOffset();
      break;
    default:
      break;
    }

    Hash += ((OperandHash << 3) | Op.getType()) << (i & 31);
  }
  return Hash;
}

/// HashEndOfMBB - Hash the last instruction in the MBB.
static unsigned HashEndOfMBB(const MachineBasicBlock &MBB) {
  MachineBasicBlock::const_iterator I = MBB.getLastNonDebugInstr();
  if (I == MBB.end())
    return 0;

  return HashMachineInstr(*I);
}

/// ComputeCommonTailLength - Given two machine basic blocks, compute the number
/// of instructions they actually have in common together at their end.  Return
/// iterators for the first shared instruction in each block.
static unsigned ComputeCommonTailLength(MachineBasicBlock *MBB1,
                                        MachineBasicBlock *MBB2,
                                        MachineBasicBlock::iterator &I1,
                                        MachineBasicBlock::iterator &I2) {
  I1 = MBB1->end();
  I2 = MBB2->end();

  unsigned TailLen = 0;
  while (I1 != MBB1->begin() && I2 != MBB2->begin()) {
    --I1; --I2;
    // Skip debugging pseudos; necessary to avoid changing the code.
    while (I1->isDebugValue()) {
      if (I1==MBB1->begin()) {
        while (I2->isDebugValue()) {
          if (I2==MBB2->begin())
            // I1==DBG at begin; I2==DBG at begin
            return TailLen;
          --I2;
        }
        ++I2;
        // I1==DBG at begin; I2==non-DBG, or first of DBGs not at begin
        return TailLen;
      }
      --I1;
    }
    // I1==first (untested) non-DBG preceding known match
    while (I2->isDebugValue()) {
      if (I2==MBB2->begin()) {
        ++I1;
        // I1==non-DBG, or first of DBGs not at begin; I2==DBG at begin
        return TailLen;
      }
      --I2;
    }
    // I1, I2==first (untested) non-DBGs preceding known match
    if (!I1->isIdenticalTo(*I2) ||
        // FIXME: This check is dubious. It's used to get around a problem where
        // people incorrectly expect inline asm directives to remain in the same
        // relative order. This is untenable because normal compiler
        // optimizations (like this one) may reorder and/or merge these
        // directives.
        I1->isInlineAsm()) {
      ++I1; ++I2;
      break;
    }
    ++TailLen;
  }
  // Back past possible debugging pseudos at beginning of block.  This matters
  // when one block differs from the other only by whether debugging pseudos
  // are present at the beginning. (This way, the various checks later for
  // I1==MBB1->begin() work as expected.)
  if (I1 == MBB1->begin() && I2 != MBB2->begin()) {
    --I2;
    while (I2->isDebugValue()) {
      if (I2 == MBB2->begin())
        return TailLen;
      --I2;
    }
    ++I2;
  }
  if (I2 == MBB2->begin() && I1 != MBB1->begin()) {
    --I1;
    while (I1->isDebugValue()) {
      if (I1 == MBB1->begin())
        return TailLen;
      --I1;
    }
    ++I1;
  }
  return TailLen;
}

void BranchFolder::ReplaceTailWithBranchTo(MachineBasicBlock::iterator OldInst,
                                           MachineBasicBlock *NewDest) {
  TII->ReplaceTailWithBranchTo(OldInst, NewDest);

  if (UpdateLiveIns) {
    NewDest->clearLiveIns();
    computeLiveIns(LiveRegs, *MRI, *NewDest);
  }

  ++NumTailMerge;
}

MachineBasicBlock *BranchFolder::SplitMBBAt(MachineBasicBlock &CurMBB,
                                            MachineBasicBlock::iterator BBI1,
                                            const BasicBlock *BB) {
  if (!TII->isLegalToSplitMBBAt(CurMBB, BBI1))
    return nullptr;

  MachineFunction &MF = *CurMBB.getParent();

  // Create the fall-through block.
  MachineFunction::iterator MBBI = CurMBB.getIterator();
  MachineBasicBlock *NewMBB =MF.CreateMachineBasicBlock(BB);
  CurMBB.getParent()->insert(++MBBI, NewMBB);

  // Move all the successors of this block to the specified block.
  NewMBB->transferSuccessors(&CurMBB);

  // Add an edge from CurMBB to NewMBB for the fall-through.
  CurMBB.addSuccessor(NewMBB);

  // Splice the code over.
  NewMBB->splice(NewMBB->end(), &CurMBB, BBI1, CurMBB.end());

  // NewMBB belongs to the same loop as CurMBB.
  if (MLI)
    if (MachineLoop *ML = MLI->getLoopFor(&CurMBB))
      ML->addBasicBlockToLoop(NewMBB, MLI->getBase());

  // NewMBB inherits CurMBB's block frequency.
  MBBFreqInfo.setBlockFreq(NewMBB, MBBFreqInfo.getBlockFreq(&CurMBB));

  if (UpdateLiveIns)
    computeLiveIns(LiveRegs, *MRI, *NewMBB);

  // Add the new block to the funclet.
  const auto &FuncletI = FuncletMembership.find(&CurMBB);
  if (FuncletI != FuncletMembership.end()) {
    auto n = FuncletI->second;
    FuncletMembership[NewMBB] = n;
  }

  return NewMBB;
}

/// EstimateRuntime - Make a rough estimate for how long it will take to run
/// the specified code.
static unsigned EstimateRuntime(MachineBasicBlock::iterator I,
                                MachineBasicBlock::iterator E) {
  unsigned Time = 0;
  for (; I != E; ++I) {
    if (I->isDebugValue())
      continue;
    if (I->isCall())
      Time += 10;
    else if (I->mayLoad() || I->mayStore())
      Time += 2;
    else
      ++Time;
  }
  return Time;
}

// CurMBB needs to add an unconditional branch to SuccMBB (we removed these
// branches temporarily for tail merging).  In the case where CurMBB ends
// with a conditional branch to the next block, optimize by reversing the
// test and conditionally branching to SuccMBB instead.
static void FixTail(MachineBasicBlock *CurMBB, MachineBasicBlock *SuccBB,
                    const TargetInstrInfo *TII) {
  MachineFunction *MF = CurMBB->getParent();
  MachineFunction::iterator I = std::next(MachineFunction::iterator(CurMBB));
  MachineBasicBlock *TBB = nullptr, *FBB = nullptr;
  SmallVector<MachineOperand, 4> Cond;
  DebugLoc dl = CurMBB->findBranchDebugLoc();
  if (I != MF->end() && !TII->analyzeBranch(*CurMBB, TBB, FBB, Cond, true)) {
    MachineBasicBlock *NextBB = &*I;
    if (TBB == NextBB && !Cond.empty() && !FBB) {
      if (!TII->reverseBranchCondition(Cond)) {
        TII->removeBranch(*CurMBB);
        TII->insertBranch(*CurMBB, SuccBB, nullptr, Cond, dl);
        return;
      }
    }
  }
  TII->insertBranch(*CurMBB, SuccBB, nullptr,
                    SmallVector<MachineOperand, 0>(), dl);
}

bool
BranchFolder::MergePotentialsElt::operator<(const MergePotentialsElt &o) const {
  if (getHash() < o.getHash())
    return true;
  if (getHash() > o.getHash())
    return false;
  if (getBlock()->getNumber() < o.getBlock()->getNumber())
    return true;
  if (getBlock()->getNumber() > o.getBlock()->getNumber())
    return false;
  // _GLIBCXX_DEBUG checks strict weak ordering, which involves comparing
  // an object with itself.
#ifndef _GLIBCXX_DEBUG
  llvm_unreachable("Predecessor appears twice");
#else
  return false;
#endif
}

BlockFrequency
BranchFolder::MBFIWrapper::getBlockFreq(const MachineBasicBlock *MBB) const {
  auto I = MergedBBFreq.find(MBB);

  if (I != MergedBBFreq.end())
    return I->second;

  return MBFI.getBlockFreq(MBB);
}

void BranchFolder::MBFIWrapper::setBlockFreq(const MachineBasicBlock *MBB,
                                             BlockFrequency F) {
  MergedBBFreq[MBB] = F;
}

raw_ostream &
BranchFolder::MBFIWrapper::printBlockFreq(raw_ostream &OS,
                                          const MachineBasicBlock *MBB) const {
  return MBFI.printBlockFreq(OS, getBlockFreq(MBB));
}

raw_ostream &
BranchFolder::MBFIWrapper::printBlockFreq(raw_ostream &OS,
                                          const BlockFrequency Freq) const {
  return MBFI.printBlockFreq(OS, Freq);
}

void BranchFolder::MBFIWrapper::view(const Twine &Name, bool isSimple) {
  MBFI.view(Name, isSimple);
}

uint64_t
BranchFolder::MBFIWrapper::getEntryFreq() const {
  return MBFI.getEntryFreq();
}

/// CountTerminators - Count the number of terminators in the given
/// block and set I to the position of the first non-terminator, if there
/// is one, or MBB->end() otherwise.
static unsigned CountTerminators(MachineBasicBlock *MBB,
                                 MachineBasicBlock::iterator &I) {
  I = MBB->end();
  unsigned NumTerms = 0;
  for (;;) {
    if (I == MBB->begin()) {
      I = MBB->end();
      break;
    }
    --I;
    if (!I->isTerminator()) break;
    ++NumTerms;
  }
  return NumTerms;
}

/// A no successor, non-return block probably ends in unreachable and is cold.
/// Also consider a block that ends in an indirect branch to be a return block,
/// since many targets use plain indirect branches to return.
static bool blockEndsInUnreachable(const MachineBasicBlock *MBB) {
  if (!MBB->succ_empty())
    return false;
  if (MBB->empty())
    return true;
  return !(MBB->back().isReturn() || MBB->back().isIndirectBranch());
}

/// ProfitableToMerge - Check if two machine basic blocks have a common tail
/// and decide if it would be profitable to merge those tails.  Return the
/// length of the common tail and iterators to the first common instruction
/// in each block.
/// MBB1, MBB2      The blocks to check
/// MinCommonTailLength  Minimum size of tail block to be merged.
/// CommonTailLen   Out parameter to record the size of the shared tail between
///                 MBB1 and MBB2
/// I1, I2          Iterator references that will be changed to point to the first
///                 instruction in the common tail shared by MBB1,MBB2
/// SuccBB          A common successor of MBB1, MBB2 which are in a canonical form
///                 relative to SuccBB
/// PredBB          The layout predecessor of SuccBB, if any.
/// FuncletMembership  map from block to funclet #.
/// AfterPlacement  True if we are merging blocks after layout. Stricter
///                 thresholds apply to prevent undoing tail-duplication.
static bool
ProfitableToMerge(MachineBasicBlock *MBB1, MachineBasicBlock *MBB2,
                  unsigned MinCommonTailLength, unsigned &CommonTailLen,
                  MachineBasicBlock::iterator &I1,
                  MachineBasicBlock::iterator &I2, MachineBasicBlock *SuccBB,
                  MachineBasicBlock *PredBB,
                  DenseMap<const MachineBasicBlock *, int> &FuncletMembership,
                  bool AfterPlacement) {
  // It is never profitable to tail-merge blocks from two different funclets.
  if (!FuncletMembership.empty()) {
    auto Funclet1 = FuncletMembership.find(MBB1);
    assert(Funclet1 != FuncletMembership.end());
    auto Funclet2 = FuncletMembership.find(MBB2);
    assert(Funclet2 != FuncletMembership.end());
    if (Funclet1->second != Funclet2->second)
      return false;
  }

  CommonTailLen = ComputeCommonTailLength(MBB1, MBB2, I1, I2);
  if (CommonTailLen == 0)
    return false;
  DEBUG(dbgs() << "Common tail length of BB#" << MBB1->getNumber()
               << " and BB#" << MBB2->getNumber() << " is " << CommonTailLen
               << '\n');

  // It's almost always profitable to merge any number of non-terminator
  // instructions with the block that falls through into the common successor.
  // This is true only for a single successor. For multiple successors, we are
  // trading a conditional branch for an unconditional one.
  // TODO: Re-visit successor size for non-layout tail merging.
  if ((MBB1 == PredBB || MBB2 == PredBB) &&
      (!AfterPlacement || MBB1->succ_size() == 1)) {
    MachineBasicBlock::iterator I;
    unsigned NumTerms = CountTerminators(MBB1 == PredBB ? MBB2 : MBB1, I);
    if (CommonTailLen > NumTerms)
      return true;
  }

  // If these are identical non-return blocks with no successors, merge them.
  // Such blocks are typically cold calls to noreturn functions like abort, and
  // are unlikely to become a fallthrough target after machine block placement.
  // Tail merging these blocks is unlikely to create additional unconditional
  // branches, and will reduce the size of this cold code.
  if (I1 == MBB1->begin() && I2 == MBB2->begin() &&
      blockEndsInUnreachable(MBB1) && blockEndsInUnreachable(MBB2))
    return true;

  // If one of the blocks can be completely merged and happens to be in
  // a position where the other could fall through into it, merge any number
  // of instructions, because it can be done without a branch.
  // TODO: If the blocks are not adjacent, move one of them so that they are?
  if (MBB1->isLayoutSuccessor(MBB2) && I2 == MBB2->begin())
    return true;
  if (MBB2->isLayoutSuccessor(MBB1) && I1 == MBB1->begin())
    return true;

  // If both blocks are identical and end in a branch, merge them unless they
  // both have a fallthrough predecessor and successor.
  // We can only do this after block placement because it depends on whether
  // there are fallthroughs, and we don't know until after layout.
  if (AfterPlacement && I1 == MBB1->begin() && I2 == MBB2->begin()) {
    auto BothFallThrough = [](MachineBasicBlock *MBB) {
      if (MBB->succ_size() != 0 && !MBB->canFallThrough())
        return false;
      MachineFunction::iterator I(MBB);
      MachineFunction *MF = MBB->getParent();
      return (MBB != &*MF->begin()) && std::prev(I)->canFallThrough();
    };
    if (!BothFallThrough(MBB1) || !BothFallThrough(MBB2))
      return true;
  }

  // If both blocks have an unconditional branch temporarily stripped out,
  // count that as an additional common instruction for the following
  // heuristics. This heuristic is only accurate for single-succ blocks, so to
  // make sure that during layout merging and duplicating don't crash, we check
  // for that when merging during layout.
  unsigned EffectiveTailLen = CommonTailLen;
  if (SuccBB && MBB1 != PredBB && MBB2 != PredBB &&
      (MBB1->succ_size() == 1 || !AfterPlacement) &&
      !MBB1->back().isBarrier() &&
      !MBB2->back().isBarrier())
    ++EffectiveTailLen;

  // Check if the common tail is long enough to be worthwhile.
  if (EffectiveTailLen >= MinCommonTailLength)
    return true;

  // If we are optimizing for code size, 2 instructions in common is enough if
  // we don't have to split a block.  At worst we will be introducing 1 new
  // branch instruction, which is likely to be smaller than the 2
  // instructions that would be deleted in the merge.
  MachineFunction *MF = MBB1->getParent();
  return EffectiveTailLen >= 2 && MF->getFunction()->optForSize() &&
         (I1 == MBB1->begin() || I2 == MBB2->begin());
}

unsigned BranchFolder::ComputeSameTails(unsigned CurHash,
                                        unsigned MinCommonTailLength,
                                        MachineBasicBlock *SuccBB,
                                        MachineBasicBlock *PredBB) {
  unsigned maxCommonTailLength = 0U;
  SameTails.clear();
  MachineBasicBlock::iterator TrialBBI1, TrialBBI2;
  MPIterator HighestMPIter = std::prev(MergePotentials.end());
  for (MPIterator CurMPIter = std::prev(MergePotentials.end()),
                  B = MergePotentials.begin();
       CurMPIter != B && CurMPIter->getHash() == CurHash; --CurMPIter) {
    for (MPIterator I = std::prev(CurMPIter); I->getHash() == CurHash; --I) {
      unsigned CommonTailLen;
      if (ProfitableToMerge(CurMPIter->getBlock(), I->getBlock(),
                            MinCommonTailLength,
                            CommonTailLen, TrialBBI1, TrialBBI2,
                            SuccBB, PredBB,
                            FuncletMembership,
                            AfterBlockPlacement)) {
        if (CommonTailLen > maxCommonTailLength) {
          SameTails.clear();
          maxCommonTailLength = CommonTailLen;
          HighestMPIter = CurMPIter;
          SameTails.push_back(SameTailElt(CurMPIter, TrialBBI1));
        }
        if (HighestMPIter == CurMPIter &&
            CommonTailLen == maxCommonTailLength)
          SameTails.push_back(SameTailElt(I, TrialBBI2));
      }
      if (I == B)
        break;
    }
  }
  return maxCommonTailLength;
}

void BranchFolder::RemoveBlocksWithHash(unsigned CurHash,
                                        MachineBasicBlock *SuccBB,
                                        MachineBasicBlock *PredBB) {
  MPIterator CurMPIter, B;
  for (CurMPIter = std::prev(MergePotentials.end()),
      B = MergePotentials.begin();
       CurMPIter->getHash() == CurHash; --CurMPIter) {
    // Put the unconditional branch back, if we need one.
    MachineBasicBlock *CurMBB = CurMPIter->getBlock();
    if (SuccBB && CurMBB != PredBB)
      FixTail(CurMBB, SuccBB, TII);
    if (CurMPIter == B)
      break;
  }
  if (CurMPIter->getHash() != CurHash)
    CurMPIter++;
  MergePotentials.erase(CurMPIter, MergePotentials.end());
}

bool BranchFolder::CreateCommonTailOnlyBlock(MachineBasicBlock *&PredBB,
                                             MachineBasicBlock *SuccBB,
                                             unsigned maxCommonTailLength,
                                             unsigned &commonTailIndex) {
  commonTailIndex = 0;
  unsigned TimeEstimate = ~0U;
  for (unsigned i = 0, e = SameTails.size(); i != e; ++i) {
    // Use PredBB if possible; that doesn't require a new branch.
    if (SameTails[i].getBlock() == PredBB) {
      commonTailIndex = i;
      break;
    }
    // Otherwise, make a (fairly bogus) choice based on estimate of
    // how long it will take the various blocks to execute.
    unsigned t = EstimateRuntime(SameTails[i].getBlock()->begin(),
                                 SameTails[i].getTailStartPos());
    if (t <= TimeEstimate) {
      TimeEstimate = t;
      commonTailIndex = i;
    }
  }

  MachineBasicBlock::iterator BBI =
    SameTails[commonTailIndex].getTailStartPos();
  MachineBasicBlock *MBB = SameTails[commonTailIndex].getBlock();

  DEBUG(dbgs() << "\nSplitting BB#" << MBB->getNumber() << ", size "
               << maxCommonTailLength);

  // If the split block unconditionally falls-thru to SuccBB, it will be
  // merged. In control flow terms it should then take SuccBB's name. e.g. If
  // SuccBB is an inner loop, the common tail is still part of the inner loop.
  const BasicBlock *BB = (SuccBB && MBB->succ_size() == 1) ?
    SuccBB->getBasicBlock() : MBB->getBasicBlock();
  MachineBasicBlock *newMBB = SplitMBBAt(*MBB, BBI, BB);
  if (!newMBB) {
    DEBUG(dbgs() << "... failed!");
    return false;
  }

  SameTails[commonTailIndex].setBlock(newMBB);
  SameTails[commonTailIndex].setTailStartPos(newMBB->begin());

  // If we split PredBB, newMBB is the new predecessor.
  if (PredBB == MBB)
    PredBB = newMBB;

  return true;
}

void BranchFolder::MergeCommonTailDebugLocs(unsigned commonTailIndex) {
  MachineBasicBlock *MBB = SameTails[commonTailIndex].getBlock();

  std::vector<MachineBasicBlock::iterator> NextCommonInsts(SameTails.size());
  for (unsigned int i = 0 ; i != SameTails.size() ; ++i) {
    if (i != commonTailIndex)
      NextCommonInsts[i] = SameTails[i].getTailStartPos();
    else {
      assert(SameTails[i].getTailStartPos() == MBB->begin() &&
          "MBB is not a common tail only block");
    }
  }

  for (auto &MI : *MBB) {
    if (MI.isDebugValue())
      continue;
    DebugLoc DL = MI.getDebugLoc();
    for (unsigned int i = 0 ; i < NextCommonInsts.size() ; i++) {
      if (i == commonTailIndex)
        continue;

      auto &Pos = NextCommonInsts[i];
      assert(Pos != SameTails[i].getBlock()->end() &&
          "Reached BB end within common tail");
      while (Pos->isDebugValue()) {
        ++Pos;
        assert(Pos != SameTails[i].getBlock()->end() &&
            "Reached BB end within common tail");
      }
      assert(MI.isIdenticalTo(*Pos) && "Expected matching MIIs!");
      DL = DILocation::getMergedLocation(DL, Pos->getDebugLoc());
      NextCommonInsts[i] = ++Pos;
    }
    MI.setDebugLoc(DL);
  }
}

static void
mergeOperations(MachineBasicBlock::iterator MBBIStartPos,
                MachineBasicBlock &MBBCommon) {
  MachineBasicBlock *MBB = MBBIStartPos->getParent();
  // Note CommonTailLen does not necessarily matches the size of
  // the common BB nor all its instructions because of debug
  // instructions differences.
  unsigned CommonTailLen = 0;
  for (auto E = MBB->end(); MBBIStartPos != E; ++MBBIStartPos)
    ++CommonTailLen;

  MachineBasicBlock::reverse_iterator MBBI = MBB->rbegin();
  MachineBasicBlock::reverse_iterator MBBIE = MBB->rend();
  MachineBasicBlock::reverse_iterator MBBICommon = MBBCommon.rbegin();
  MachineBasicBlock::reverse_iterator MBBIECommon = MBBCommon.rend();

  while (CommonTailLen--) {
    assert(MBBI != MBBIE && "Reached BB end within common tail length!");
    (void)MBBIE;

    if (MBBI->isDebugValue()) {
      ++MBBI;
      continue;
    }

    while ((MBBICommon != MBBIECommon) && MBBICommon->isDebugValue())
      ++MBBICommon;

    assert(MBBICommon != MBBIECommon &&
           "Reached BB end within common tail length!");
    assert(MBBICommon->isIdenticalTo(*MBBI) && "Expected matching MIIs!");

    // Merge MMOs from memory operations in the common block.
    if (MBBICommon->mayLoad() || MBBICommon->mayStore())
      MBBICommon->setMemRefs(MBBICommon->mergeMemRefsWith(*MBBI));
    // Drop undef flags if they aren't present in all merged instructions.
    for (unsigned I = 0, E = MBBICommon->getNumOperands(); I != E; ++I) {
      MachineOperand &MO = MBBICommon->getOperand(I);
      if (MO.isReg() && MO.isUndef()) {
        const MachineOperand &OtherMO = MBBI->getOperand(I);
        if (!OtherMO.isUndef())
          MO.setIsUndef(false);
      }
    }

    ++MBBI;
    ++MBBICommon;
  }
}

// See if any of the blocks in MergePotentials (which all have SuccBB as a
// successor, or all have no successor if it is null) can be tail-merged.
// If there is a successor, any blocks in MergePotentials that are not
// tail-merged and are not immediately before Succ must have an unconditional
// branch to Succ added (but the predecessor/successor lists need no
// adjustment). The lone predecessor of Succ that falls through into Succ,
// if any, is given in PredBB.
// MinCommonTailLength - Except for the special cases below, tail-merge if
// there are at least this many instructions in common.
bool BranchFolder::TryTailMergeBlocks(MachineBasicBlock *SuccBB,
                                      MachineBasicBlock *PredBB,
                                      unsigned MinCommonTailLength) {
  bool MadeChange = false;

  DEBUG(dbgs() << "\nTryTailMergeBlocks: ";
        for (unsigned i = 0, e = MergePotentials.size(); i != e; ++i)
          dbgs() << "BB#" << MergePotentials[i].getBlock()->getNumber()
                 << (i == e-1 ? "" : ", ");
        dbgs() << "\n";
        if (SuccBB) {
          dbgs() << "  with successor BB#" << SuccBB->getNumber() << '\n';
          if (PredBB)
            dbgs() << "  which has fall-through from BB#"
                   << PredBB->getNumber() << "\n";
        }
        dbgs() << "Looking for common tails of at least "
               << MinCommonTailLength << " instruction"
               << (MinCommonTailLength == 1 ? "" : "s") << '\n';
       );

  // Sort by hash value so that blocks with identical end sequences sort
  // together.
  array_pod_sort(MergePotentials.begin(), MergePotentials.end());

  // Walk through equivalence sets looking for actual exact matches.
  while (MergePotentials.size() > 1) {
    unsigned CurHash = MergePotentials.back().getHash();

    // Build SameTails, identifying the set of blocks with this hash code
    // and with the maximum number of instructions in common.
    unsigned maxCommonTailLength = ComputeSameTails(CurHash,
                                                    MinCommonTailLength,
                                                    SuccBB, PredBB);

    // If we didn't find any pair that has at least MinCommonTailLength
    // instructions in common, remove all blocks with this hash code and retry.
    if (SameTails.empty()) {
      RemoveBlocksWithHash(CurHash, SuccBB, PredBB);
      continue;
    }

    // If one of the blocks is the entire common tail (and not the entry
    // block, which we can't jump to), we can treat all blocks with this same
    // tail at once.  Use PredBB if that is one of the possibilities, as that
    // will not introduce any extra branches.
    MachineBasicBlock *EntryBB =
        &MergePotentials.front().getBlock()->getParent()->front();
    unsigned commonTailIndex = SameTails.size();
    // If there are two blocks, check to see if one can be made to fall through
    // into the other.
    if (SameTails.size() == 2 &&
        SameTails[0].getBlock()->isLayoutSuccessor(SameTails[1].getBlock()) &&
        SameTails[1].tailIsWholeBlock())
      commonTailIndex = 1;
    else if (SameTails.size() == 2 &&
             SameTails[1].getBlock()->isLayoutSuccessor(
                                                     SameTails[0].getBlock()) &&
             SameTails[0].tailIsWholeBlock())
      commonTailIndex = 0;
    else {
      // Otherwise just pick one, favoring the fall-through predecessor if
      // there is one.
      for (unsigned i = 0, e = SameTails.size(); i != e; ++i) {
        MachineBasicBlock *MBB = SameTails[i].getBlock();
        if (MBB == EntryBB && SameTails[i].tailIsWholeBlock())
          continue;
        if (MBB == PredBB) {
          commonTailIndex = i;
          break;
        }
        if (SameTails[i].tailIsWholeBlock())
          commonTailIndex = i;
      }
    }

    if (commonTailIndex == SameTails.size() ||
        (SameTails[commonTailIndex].getBlock() == PredBB &&
         !SameTails[commonTailIndex].tailIsWholeBlock())) {
      // None of the blocks consist entirely of the common tail.
      // Split a block so that one does.
      if (!CreateCommonTailOnlyBlock(PredBB, SuccBB,
                                     maxCommonTailLength, commonTailIndex)) {
        RemoveBlocksWithHash(CurHash, SuccBB, PredBB);
        continue;
      }
    }

    MachineBasicBlock *MBB = SameTails[commonTailIndex].getBlock();

    // Recompute common tail MBB's edge weights and block frequency.
    setCommonTailEdgeWeights(*MBB);

    // Merge debug locations across identical instructions for common tail.
    MergeCommonTailDebugLocs(commonTailIndex);

    // MBB is common tail.  Adjust all other BB's to jump to this one.
    // Traversal must be forwards so erases work.
    DEBUG(dbgs() << "\nUsing common tail in BB#" << MBB->getNumber()
                 << " for ");
    for (unsigned int i=0, e = SameTails.size(); i != e; ++i) {
      if (commonTailIndex == i)
        continue;
      DEBUG(dbgs() << "BB#" << SameTails[i].getBlock()->getNumber()
                   << (i == e-1 ? "" : ", "));
      // Merge operations (MMOs, undef flags)
      mergeOperations(SameTails[i].getTailStartPos(), *MBB);
      // Hack the end off BB i, making it jump to BB commonTailIndex instead.
      ReplaceTailWithBranchTo(SameTails[i].getTailStartPos(), MBB);
      // BB i is no longer a predecessor of SuccBB; remove it from the worklist.
      MergePotentials.erase(SameTails[i].getMPIter());
    }
    DEBUG(dbgs() << "\n");
    // We leave commonTailIndex in the worklist in case there are other blocks
    // that match it with a smaller number of instructions.
    MadeChange = true;
  }
  return MadeChange;
}

bool BranchFolder::TailMergeBlocks(MachineFunction &MF) {
  bool MadeChange = false;
  if (!EnableTailMerge) return MadeChange;

  // First find blocks with no successors.
  // Block placement does not create new tail merging opportunities for these
  // blocks.
  if (!AfterBlockPlacement) {
    MergePotentials.clear();
    for (MachineBasicBlock &MBB : MF) {
      if (MergePotentials.size() == TailMergeThreshold)
        break;
      if (!TriedMerging.count(&MBB) && MBB.succ_empty())
        MergePotentials.push_back(MergePotentialsElt(HashEndOfMBB(MBB), &MBB));
    }

    // If this is a large problem, avoid visiting the same basic blocks
    // multiple times.
    if (MergePotentials.size() == TailMergeThreshold)
      for (unsigned i = 0, e = MergePotentials.size(); i != e; ++i)
        TriedMerging.insert(MergePotentials[i].getBlock());

    // See if we can do any tail merging on those.
    if (MergePotentials.size() >= 2)
      MadeChange |= TryTailMergeBlocks(nullptr, nullptr, MinCommonTailLength);
  }

  // Look at blocks (IBB) with multiple predecessors (PBB).
  // We change each predecessor to a canonical form, by
  // (1) temporarily removing any unconditional branch from the predecessor
  // to IBB, and
  // (2) alter conditional branches so they branch to the other block
  // not IBB; this may require adding back an unconditional branch to IBB
  // later, where there wasn't one coming in.  E.g.
  //   Bcc IBB
  //   fallthrough to QBB
  // here becomes
  //   Bncc QBB
  // with a conceptual B to IBB after that, which never actually exists.
  // With those changes, we see whether the predecessors' tails match,
  // and merge them if so.  We change things out of canonical form and
  // back to the way they were later in the process.  (OptimizeBranches
  // would undo some of this, but we can't use it, because we'd get into
  // a compile-time infinite loop repeatedly doing and undoing the same
  // transformations.)

  for (MachineFunction::iterator I = std::next(MF.begin()), E = MF.end();
       I != E; ++I) {
    if (I->pred_size() < 2) continue;
    SmallPtrSet<MachineBasicBlock *, 8> UniquePreds;
    MachineBasicBlock *IBB = &*I;
    MachineBasicBlock *PredBB = &*std::prev(I);
    MergePotentials.clear();
    MachineLoop *ML;

    // Bail if merging after placement and IBB is the loop header because
    // -- If merging predecessors that belong to the same loop as IBB, the
    // common tail of merged predecessors may become the loop top if block
    // placement is called again and the predecessors may branch to this common
    // tail and require more branches. This can be relaxed if
    // MachineBlockPlacement::findBestLoopTop is more flexible.
    // --If merging predecessors that do not belong to the same loop as IBB, the
    // loop info of IBB's loop and the other loops may be affected. Calling the
    // block placement again may make big change to the layout and eliminate the
    // reason to do tail merging here.
    if (AfterBlockPlacement && MLI) {
      ML = MLI->getLoopFor(IBB);
      if (ML && IBB == ML->getHeader())
        continue;
    }

    for (MachineBasicBlock *PBB : I->predecessors()) {
      if (MergePotentials.size() == TailMergeThreshold)
        break;

      if (TriedMerging.count(PBB))
        continue;

      // Skip blocks that loop to themselves, can't tail merge these.
      if (PBB == IBB)
        continue;

      // Visit each predecessor only once.
      if (!UniquePreds.insert(PBB).second)
        continue;

      // Skip blocks which may jump to a landing pad. Can't tail merge these.
      if (PBB->hasEHPadSuccessor())
        continue;

      // After block placement, only consider predecessors that belong to the
      // same loop as IBB.  The reason is the same as above when skipping loop
      // header.
      if (AfterBlockPlacement && MLI)
        if (ML != MLI->getLoopFor(PBB))
          continue;

      MachineBasicBlock *TBB = nullptr, *FBB = nullptr;
      SmallVector<MachineOperand, 4> Cond;
      if (!TII->analyzeBranch(*PBB, TBB, FBB, Cond, true)) {
        // Failing case: IBB is the target of a cbr, and we cannot reverse the
        // branch.
        SmallVector<MachineOperand, 4> NewCond(Cond);
        if (!Cond.empty() && TBB == IBB) {
          if (TII->reverseBranchCondition(NewCond))
            continue;
          // This is the QBB case described above
          if (!FBB) {
            auto Next = ++PBB->getIterator();
            if (Next != MF.end())
              FBB = &*Next;
          }
        }

        // Failing case: the only way IBB can be reached from PBB is via
        // exception handling.  Happens for landing pads.  Would be nice to have
        // a bit in the edge so we didn't have to do all this.
        if (IBB->isEHPad()) {
          MachineFunction::iterator IP = ++PBB->getIterator();
          MachineBasicBlock *PredNextBB = nullptr;
          if (IP != MF.end())
            PredNextBB = &*IP;
          if (!TBB) {
            if (IBB != PredNextBB)      // fallthrough
              continue;
          } else if (FBB) {
            if (TBB != IBB && FBB != IBB)   // cbr then ubr
              continue;
          } else if (Cond.empty()) {
            if (TBB != IBB)               // ubr
              continue;
          } else {
            if (TBB != IBB && IBB != PredNextBB)  // cbr
              continue;
          }
        }

        // Remove the unconditional branch at the end, if any.
        if (TBB && (Cond.empty() || FBB)) {
          DebugLoc dl = PBB->findBranchDebugLoc();
          TII->removeBranch(*PBB);
          if (!Cond.empty())
            // reinsert conditional branch only, for now
            TII->insertBranch(*PBB, (TBB == IBB) ? FBB : TBB, nullptr,
                              NewCond, dl);
        }

        MergePotentials.push_back(MergePotentialsElt(HashEndOfMBB(*PBB), PBB));
      }
    }

    // If this is a large problem, avoid visiting the same basic blocks multiple
    // times.
    if (MergePotentials.size() == TailMergeThreshold)
      for (unsigned i = 0, e = MergePotentials.size(); i != e; ++i)
        TriedMerging.insert(MergePotentials[i].getBlock());

    if (MergePotentials.size() >= 2)
      MadeChange |= TryTailMergeBlocks(IBB, PredBB, MinCommonTailLength);

    // Reinsert an unconditional branch if needed. The 1 below can occur as a
    // result of removing blocks in TryTailMergeBlocks.
    PredBB = &*std::prev(I); // this may have been changed in TryTailMergeBlocks
    if (MergePotentials.size() == 1 &&
        MergePotentials.begin()->getBlock() != PredBB)
      FixTail(MergePotentials.begin()->getBlock(), IBB, TII);
  }

  return MadeChange;
}

void BranchFolder::setCommonTailEdgeWeights(MachineBasicBlock &TailMBB) {
  SmallVector<BlockFrequency, 2> EdgeFreqLs(TailMBB.succ_size());
  BlockFrequency AccumulatedMBBFreq;

  // Aggregate edge frequency of successor edge j:
  //  edgeFreq(j) = sum (freq(bb) * edgeProb(bb, j)),
  //  where bb is a basic block that is in SameTails.
  for (const auto &Src : SameTails) {
    const MachineBasicBlock *SrcMBB = Src.getBlock();
    BlockFrequency BlockFreq = MBBFreqInfo.getBlockFreq(SrcMBB);
    AccumulatedMBBFreq += BlockFreq;

    // It is not necessary to recompute edge weights if TailBB has less than two
    // successors.
    if (TailMBB.succ_size() <= 1)
      continue;

    auto EdgeFreq = EdgeFreqLs.begin();

    for (auto SuccI = TailMBB.succ_begin(), SuccE = TailMBB.succ_end();
         SuccI != SuccE; ++SuccI, ++EdgeFreq)
      *EdgeFreq += BlockFreq * MBPI.getEdgeProbability(SrcMBB, *SuccI);
  }

  MBBFreqInfo.setBlockFreq(&TailMBB, AccumulatedMBBFreq);

  if (TailMBB.succ_size() <= 1)
    return;

  auto SumEdgeFreq =
      std::accumulate(EdgeFreqLs.begin(), EdgeFreqLs.end(), BlockFrequency(0))
          .getFrequency();
  auto EdgeFreq = EdgeFreqLs.begin();

  if (SumEdgeFreq > 0) {
    for (auto SuccI = TailMBB.succ_begin(), SuccE = TailMBB.succ_end();
         SuccI != SuccE; ++SuccI, ++EdgeFreq) {
      auto Prob = BranchProbability::getBranchProbability(
          EdgeFreq->getFrequency(), SumEdgeFreq);
      TailMBB.setSuccProbability(SuccI, Prob);
    }
  }
}

//===----------------------------------------------------------------------===//
//  Branch Optimization
//===----------------------------------------------------------------------===//

bool BranchFolder::OptimizeBranches(MachineFunction &MF) {
  bool MadeChange = false;

  // Make sure blocks are numbered in order
  MF.RenumberBlocks();
  // Renumbering blocks alters funclet membership, recalculate it.
  FuncletMembership = getFuncletMembership(MF);

  for (MachineFunction::iterator I = std::next(MF.begin()), E = MF.end();
       I != E; ) {
    MachineBasicBlock *MBB = &*I++;
    MadeChange |= OptimizeBlock(MBB);

    // If it is dead, remove it.
    if (MBB->pred_empty()) {
      RemoveDeadBlock(MBB);
      MadeChange = true;
      ++NumDeadBlocks;
    }
  }

  return MadeChange;
}

// Blocks should be considered empty if they contain only debug info;
// else the debug info would affect codegen.
static bool IsEmptyBlock(MachineBasicBlock *MBB) {
  return MBB->getFirstNonDebugInstr() == MBB->end();
}

// Blocks with only debug info and branches should be considered the same
// as blocks with only branches.
static bool IsBranchOnlyBlock(MachineBasicBlock *MBB) {
  MachineBasicBlock::iterator I = MBB->getFirstNonDebugInstr();
  assert(I != MBB->end() && "empty block!");
  return I->isBranch();
}

/// IsBetterFallthrough - Return true if it would be clearly better to
/// fall-through to MBB1 than to fall through into MBB2.  This has to return
/// a strict ordering, returning true for both (MBB1,MBB2) and (MBB2,MBB1) will
/// result in infinite loops.
static bool IsBetterFallthrough(MachineBasicBlock *MBB1,
                                MachineBasicBlock *MBB2) {
  // Right now, we use a simple heuristic.  If MBB2 ends with a call, and
  // MBB1 doesn't, we prefer to fall through into MBB1.  This allows us to
  // optimize branches that branch to either a return block or an assert block
  // into a fallthrough to the return.
  MachineBasicBlock::iterator MBB1I = MBB1->getLastNonDebugInstr();
  MachineBasicBlock::iterator MBB2I = MBB2->getLastNonDebugInstr();
  if (MBB1I == MBB1->end() || MBB2I == MBB2->end())
    return false;

  // If there is a clear successor ordering we make sure that one block
  // will fall through to the next
  if (MBB1->isSuccessor(MBB2)) return true;
  if (MBB2->isSuccessor(MBB1)) return false;

  return MBB2I->isCall() && !MBB1I->isCall();
}

/// getBranchDebugLoc - Find and return, if any, the DebugLoc of the branch
/// instructions on the block.
static DebugLoc getBranchDebugLoc(MachineBasicBlock &MBB) {
  MachineBasicBlock::iterator I = MBB.getLastNonDebugInstr();
  if (I != MBB.end() && I->isBranch())
    return I->getDebugLoc();
  return DebugLoc();
}

bool BranchFolder::OptimizeBlock(MachineBasicBlock *MBB) {
  bool MadeChange = false;
  MachineFunction &MF = *MBB->getParent();
ReoptimizeBlock:

  MachineFunction::iterator FallThrough = MBB->getIterator();
  ++FallThrough;

  // Make sure MBB and FallThrough belong to the same funclet.
  bool SameFunclet = true;
  if (!FuncletMembership.empty() && FallThrough != MF.end()) {
    auto MBBFunclet = FuncletMembership.find(MBB);
    assert(MBBFunclet != FuncletMembership.end());
    auto FallThroughFunclet = FuncletMembership.find(&*FallThrough);
    assert(FallThroughFunclet != FuncletMembership.end());
    SameFunclet = MBBFunclet->second == FallThroughFunclet->second;
  }

  // If this block is empty, make everyone use its fall-through, not the block
  // explicitly.  Landing pads should not do this since the landing-pad table
  // points to this block.  Blocks with their addresses taken shouldn't be
  // optimized away.
  if (IsEmptyBlock(MBB) && !MBB->isEHPad() && !MBB->hasAddressTaken() &&
      SameFunclet) {
    // Dead block?  Leave for cleanup later.
    if (MBB->pred_empty()) return MadeChange;

    if (FallThrough == MF.end()) {
      // TODO: Simplify preds to not branch here if possible!
    } else if (FallThrough->isEHPad()) {
      // Don't rewrite to a landing pad fallthough.  That could lead to the case
      // where a BB jumps to more than one landing pad.
      // TODO: Is it ever worth rewriting predecessors which don't already
      // jump to a landing pad, and so can safely jump to the fallthrough?
    } else if (MBB->isSuccessor(&*FallThrough)) {
      // Rewrite all predecessors of the old block to go to the fallthrough
      // instead.
      while (!MBB->pred_empty()) {
        MachineBasicBlock *Pred = *(MBB->pred_end()-1);
        Pred->ReplaceUsesOfBlockWith(MBB, &*FallThrough);
      }
      // If MBB was the target of a jump table, update jump tables to go to the
      // fallthrough instead.
      if (MachineJumpTableInfo *MJTI = MF.getJumpTableInfo())
        MJTI->ReplaceMBBInJumpTables(MBB, &*FallThrough);
      MadeChange = true;
    }
    return MadeChange;
  }

  // Check to see if we can simplify the terminator of the block before this
  // one.
  MachineBasicBlock &PrevBB = *std::prev(MachineFunction::iterator(MBB));

  MachineBasicBlock *PriorTBB = nullptr, *PriorFBB = nullptr;
  SmallVector<MachineOperand, 4> PriorCond;
  bool PriorUnAnalyzable =
      TII->analyzeBranch(PrevBB, PriorTBB, PriorFBB, PriorCond, true);
  if (!PriorUnAnalyzable) {
    // If the CFG for the prior block has extra edges, remove them.
    MadeChange |= PrevBB.CorrectExtraCFGEdges(PriorTBB, PriorFBB,
                                              !PriorCond.empty());

    // If the previous branch is conditional and both conditions go to the same
    // destination, remove the branch, replacing it with an unconditional one or
    // a fall-through.
    if (PriorTBB && PriorTBB == PriorFBB) {
      DebugLoc dl = getBranchDebugLoc(PrevBB);
      TII->removeBranch(PrevBB);
      PriorCond.clear();
      if (PriorTBB != MBB)
        TII->insertBranch(PrevBB, PriorTBB, nullptr, PriorCond, dl);
      MadeChange = true;
      ++NumBranchOpts;
      goto ReoptimizeBlock;
    }

    // If the previous block unconditionally falls through to this block and
    // this block has no other predecessors, move the contents of this block
    // into the prior block. This doesn't usually happen when SimplifyCFG
    // has been used, but it can happen if tail merging splits a fall-through
    // predecessor of a block.
    // This has to check PrevBB->succ_size() because EH edges are ignored by
    // AnalyzeBranch.
    if (PriorCond.empty() && !PriorTBB && MBB->pred_size() == 1 &&
        PrevBB.succ_size() == 1 &&
        !MBB->hasAddressTaken() && !MBB->isEHPad()) {
      DEBUG(dbgs() << "\nMerging into block: " << PrevBB
                   << "From MBB: " << *MBB);
      // Remove redundant DBG_VALUEs first.
      if (PrevBB.begin() != PrevBB.end()) {
        MachineBasicBlock::iterator PrevBBIter = PrevBB.end();
        --PrevBBIter;
        MachineBasicBlock::iterator MBBIter = MBB->begin();
        // Check if DBG_VALUE at the end of PrevBB is identical to the
        // DBG_VALUE at the beginning of MBB.
        while (PrevBBIter != PrevBB.begin() && MBBIter != MBB->end()
               && PrevBBIter->isDebugValue() && MBBIter->isDebugValue()) {
          if (!MBBIter->isIdenticalTo(*PrevBBIter))
            break;
          MachineInstr &DuplicateDbg = *MBBIter;
          ++MBBIter; -- PrevBBIter;
          DuplicateDbg.eraseFromParent();
        }
      }
      PrevBB.splice(PrevBB.end(), MBB, MBB->begin(), MBB->end());
      PrevBB.removeSuccessor(PrevBB.succ_begin());
      assert(PrevBB.succ_empty());
      PrevBB.transferSuccessors(MBB);
      MadeChange = true;
      return MadeChange;
    }

    // If the previous branch *only* branches to *this* block (conditional or
    // not) remove the branch.
    if (PriorTBB == MBB && !PriorFBB) {
      TII->removeBranch(PrevBB);
      MadeChange = true;
      ++NumBranchOpts;
      goto ReoptimizeBlock;
    }

    // If the prior block branches somewhere else on the condition and here if
    // the condition is false, remove the uncond second branch.
    if (PriorFBB == MBB) {
      DebugLoc dl = getBranchDebugLoc(PrevBB);
      TII->removeBranch(PrevBB);
      TII->insertBranch(PrevBB, PriorTBB, nullptr, PriorCond, dl);
      MadeChange = true;
      ++NumBranchOpts;
      goto ReoptimizeBlock;
    }

    // If the prior block branches here on true and somewhere else on false, and
    // if the branch condition is reversible, reverse the branch to create a
    // fall-through.
    if (PriorTBB == MBB) {
      SmallVector<MachineOperand, 4> NewPriorCond(PriorCond);
      if (!TII->reverseBranchCondition(NewPriorCond)) {
        DebugLoc dl = getBranchDebugLoc(PrevBB);
        TII->removeBranch(PrevBB);
        TII->insertBranch(PrevBB, PriorFBB, nullptr, NewPriorCond, dl);
        MadeChange = true;
        ++NumBranchOpts;
        goto ReoptimizeBlock;
      }
    }

    // If this block has no successors (e.g. it is a return block or ends with
    // a call to a no-return function like abort or __cxa_throw) and if the pred
    // falls through into this block, and if it would otherwise fall through
    // into the block after this, move this block to the end of the function.
    //
    // We consider it more likely that execution will stay in the function (e.g.
    // due to loops) than it is to exit it.  This asserts in loops etc, moving
    // the assert condition out of the loop body.
    if (MBB->succ_empty() && !PriorCond.empty() && !PriorFBB &&
        MachineFunction::iterator(PriorTBB) == FallThrough &&
        !MBB->canFallThrough()) {
      bool DoTransform = true;

      // We have to be careful that the succs of PredBB aren't both no-successor
      // blocks.  If neither have successors and if PredBB is the second from
      // last block in the function, we'd just keep swapping the two blocks for
      // last.  Only do the swap if one is clearly better to fall through than
      // the other.
      if (FallThrough == --MF.end() &&
          !IsBetterFallthrough(PriorTBB, MBB))
        DoTransform = false;

      if (DoTransform) {
        // Reverse the branch so we will fall through on the previous true cond.
        SmallVector<MachineOperand, 4> NewPriorCond(PriorCond);
        if (!TII->reverseBranchCondition(NewPriorCond)) {
          DEBUG(dbgs() << "\nMoving MBB: " << *MBB
                       << "To make fallthrough to: " << *PriorTBB << "\n");

          DebugLoc dl = getBranchDebugLoc(PrevBB);
          TII->removeBranch(PrevBB);
          TII->insertBranch(PrevBB, MBB, nullptr, NewPriorCond, dl);

          // Move this block to the end of the function.
          MBB->moveAfter(&MF.back());
          MadeChange = true;
          ++NumBranchOpts;
          return MadeChange;
        }
      }
    }
  }

  if (!IsEmptyBlock(MBB) && MBB->pred_size() == 1 &&
      MF.getFunction()->optForSize()) {
    // Changing "Jcc foo; foo: jmp bar;" into "Jcc bar;" might change the branch
    // direction, thereby defeating careful block placement and regressing
    // performance. Therefore, only consider this for optsize functions.
    MachineInstr &TailCall = *MBB->getFirstNonDebugInstr();
    if (TII->isUnconditionalTailCall(TailCall)) {
      MachineBasicBlock *Pred = *MBB->pred_begin();
      MachineBasicBlock *PredTBB = nullptr, *PredFBB = nullptr;
      SmallVector<MachineOperand, 4> PredCond;
      bool PredAnalyzable =
          !TII->analyzeBranch(*Pred, PredTBB, PredFBB, PredCond, true);

      if (PredAnalyzable && !PredCond.empty() && PredTBB == MBB) {
        // The predecessor has a conditional branch to this block which consists
        // of only a tail call. Try to fold the tail call into the conditional
        // branch.
        if (TII->canMakeTailCallConditional(PredCond, TailCall)) {
          // TODO: It would be nice if analyzeBranch() could provide a pointer
          // to the branch insturction so replaceBranchWithTailCall() doesn't
          // have to search for it.
          TII->replaceBranchWithTailCall(*Pred, PredCond, TailCall);
          ++NumTailCalls;
          Pred->removeSuccessor(MBB);
          MadeChange = true;
          return MadeChange;
        }
      }
      // If the predecessor is falling through to this block, we could reverse
      // the branch condition and fold the tail call into that. However, after
      // that we might have to re-arrange the CFG to fall through to the other
      // block and there is a high risk of regressing code size rather than
      // improving it.
    }
  }

  // Analyze the branch in the current block.
  MachineBasicBlock *CurTBB = nullptr, *CurFBB = nullptr;
  SmallVector<MachineOperand, 4> CurCond;
  bool CurUnAnalyzable =
      TII->analyzeBranch(*MBB, CurTBB, CurFBB, CurCond, true);
  if (!CurUnAnalyzable) {
    // If the CFG for the prior block has extra edges, remove them.
    MadeChange |= MBB->CorrectExtraCFGEdges(CurTBB, CurFBB, !CurCond.empty());

    // If this is a two-way branch, and the FBB branches to this block, reverse
    // the condition so the single-basic-block loop is faster.  Instead of:
    //    Loop: xxx; jcc Out; jmp Loop
    // we want:
    //    Loop: xxx; jncc Loop; jmp Out
    if (CurTBB && CurFBB && CurFBB == MBB && CurTBB != MBB) {
      SmallVector<MachineOperand, 4> NewCond(CurCond);
      if (!TII->reverseBranchCondition(NewCond)) {
        DebugLoc dl = getBranchDebugLoc(*MBB);
        TII->removeBranch(*MBB);
        TII->insertBranch(*MBB, CurFBB, CurTBB, NewCond, dl);
        MadeChange = true;
        ++NumBranchOpts;
        goto ReoptimizeBlock;
      }
    }

    // If this branch is the only thing in its block, see if we can forward
    // other blocks across it.
    if (CurTBB && CurCond.empty() && !CurFBB &&
        IsBranchOnlyBlock(MBB) && CurTBB != MBB &&
        !MBB->hasAddressTaken() && !MBB->isEHPad()) {
      DebugLoc dl = getBranchDebugLoc(*MBB);
      // This block may contain just an unconditional branch.  Because there can
      // be 'non-branch terminators' in the block, try removing the branch and
      // then seeing if the block is empty.
      TII->removeBranch(*MBB);
      // If the only things remaining in the block are debug info, remove these
      // as well, so this will behave the same as an empty block in non-debug
      // mode.
      if (IsEmptyBlock(MBB)) {
        // Make the block empty, losing the debug info (we could probably
        // improve this in some cases.)
        MBB->erase(MBB->begin(), MBB->end());
      }
      // If this block is just an unconditional branch to CurTBB, we can
      // usually completely eliminate the block.  The only case we cannot
      // completely eliminate the block is when the block before this one
      // falls through into MBB and we can't understand the prior block's branch
      // condition.
      if (MBB->empty()) {
        bool PredHasNoFallThrough = !PrevBB.canFallThrough();
        if (PredHasNoFallThrough || !PriorUnAnalyzable ||
            !PrevBB.isSuccessor(MBB)) {
          // If the prior block falls through into us, turn it into an
          // explicit branch to us to make updates simpler.
          if (!PredHasNoFallThrough && PrevBB.isSuccessor(MBB) &&
              PriorTBB != MBB && PriorFBB != MBB) {
            if (!PriorTBB) {
              assert(PriorCond.empty() && !PriorFBB &&
                     "Bad branch analysis");
              PriorTBB = MBB;
            } else {
              assert(!PriorFBB && "Machine CFG out of date!");
              PriorFBB = MBB;
            }
            DebugLoc pdl = getBranchDebugLoc(PrevBB);
            TII->removeBranch(PrevBB);
            TII->insertBranch(PrevBB, PriorTBB, PriorFBB, PriorCond, pdl);
          }

          // Iterate through all the predecessors, revectoring each in-turn.
          size_t PI = 0;
          bool DidChange = false;
          bool HasBranchToSelf = false;
          while(PI != MBB->pred_size()) {
            MachineBasicBlock *PMBB = *(MBB->pred_begin() + PI);
            if (PMBB == MBB) {
              // If this block has an uncond branch to itself, leave it.
              ++PI;
              HasBranchToSelf = true;
            } else {
              DidChange = true;
              PMBB->ReplaceUsesOfBlockWith(MBB, CurTBB);
              // If this change resulted in PMBB ending in a conditional
              // branch where both conditions go to the same destination,
              // change this to an unconditional branch (and fix the CFG).
              MachineBasicBlock *NewCurTBB = nullptr, *NewCurFBB = nullptr;
              SmallVector<MachineOperand, 4> NewCurCond;
              bool NewCurUnAnalyzable = TII->analyzeBranch(
                  *PMBB, NewCurTBB, NewCurFBB, NewCurCond, true);
              if (!NewCurUnAnalyzable && NewCurTBB && NewCurTBB == NewCurFBB) {
                DebugLoc pdl = getBranchDebugLoc(*PMBB);
                TII->removeBranch(*PMBB);
                NewCurCond.clear();
                TII->insertBranch(*PMBB, NewCurTBB, nullptr, NewCurCond, pdl);
                MadeChange = true;
                ++NumBranchOpts;
                PMBB->CorrectExtraCFGEdges(NewCurTBB, nullptr, false);
              }
            }
          }

          // Change any jumptables to go to the new MBB.
          if (MachineJumpTableInfo *MJTI = MF.getJumpTableInfo())
            MJTI->ReplaceMBBInJumpTables(MBB, CurTBB);
          if (DidChange) {
            ++NumBranchOpts;
            MadeChange = true;
            if (!HasBranchToSelf) return MadeChange;
          }
        }
      }

      // Add the branch back if the block is more than just an uncond branch.
      TII->insertBranch(*MBB, CurTBB, nullptr, CurCond, dl);
    }
  }

  // If the prior block doesn't fall through into this block, and if this
  // block doesn't fall through into some other block, see if we can find a
  // place to move this block where a fall-through will happen.
  if (!PrevBB.canFallThrough()) {

    // Now we know that there was no fall-through into this block, check to
    // see if it has a fall-through into its successor.
    bool CurFallsThru = MBB->canFallThrough();

    if (!MBB->isEHPad()) {
      // Check all the predecessors of this block.  If one of them has no fall
      // throughs, move this block right after it.
      for (MachineBasicBlock *PredBB : MBB->predecessors()) {
        // Analyze the branch at the end of the pred.
        MachineBasicBlock *PredTBB = nullptr, *PredFBB = nullptr;
        SmallVector<MachineOperand, 4> PredCond;
        if (PredBB != MBB && !PredBB->canFallThrough() &&
            !TII->analyzeBranch(*PredBB, PredTBB, PredFBB, PredCond, true) &&
            (!CurFallsThru || !CurTBB || !CurFBB) &&
            (!CurFallsThru || MBB->getNumber() >= PredBB->getNumber())) {
          // If the current block doesn't fall through, just move it.
          // If the current block can fall through and does not end with a
          // conditional branch, we need to append an unconditional jump to
          // the (current) next block.  To avoid a possible compile-time
          // infinite loop, move blocks only backward in this case.
          // Also, if there are already 2 branches here, we cannot add a third;
          // this means we have the case
          // Bcc next
          // B elsewhere
          // next:
          if (CurFallsThru) {
            MachineBasicBlock *NextBB = &*std::next(MBB->getIterator());
            CurCond.clear();
            TII->insertBranch(*MBB, NextBB, nullptr, CurCond, DebugLoc());
          }
          MBB->moveAfter(PredBB);
          MadeChange = true;
          goto ReoptimizeBlock;
        }
      }
    }

    if (!CurFallsThru) {
      // Check all successors to see if we can move this block before it.
      for (MachineBasicBlock *SuccBB : MBB->successors()) {
        // Analyze the branch at the end of the block before the succ.
        MachineFunction::iterator SuccPrev = --SuccBB->getIterator();

        // If this block doesn't already fall-through to that successor, and if
        // the succ doesn't already have a block that can fall through into it,
        // and if the successor isn't an EH destination, we can arrange for the
        // fallthrough to happen.
        if (SuccBB != MBB && &*SuccPrev != MBB &&
            !SuccPrev->canFallThrough() && !CurUnAnalyzable &&
            !SuccBB->isEHPad()) {
          MBB->moveBefore(SuccBB);
          MadeChange = true;
          goto ReoptimizeBlock;
        }
      }

      // Okay, there is no really great place to put this block.  If, however,
      // the block before this one would be a fall-through if this block were
      // removed, move this block to the end of the function. There is no real
      // advantage in "falling through" to an EH block, so we don't want to
      // perform this transformation for that case.
      //
      // Also, Windows EH introduced the possibility of an arbitrary number of
      // successors to a given block.  The analyzeBranch call does not consider
      // exception handling and so we can get in a state where a block
      // containing a call is followed by multiple EH blocks that would be
      // rotated infinitely at the end of the function if the transformation
      // below were performed for EH "FallThrough" blocks.  Therefore, even if
      // that appears not to be happening anymore, we should assume that it is
      // possible and not remove the "!FallThrough()->isEHPad" condition below.
      MachineBasicBlock *PrevTBB = nullptr, *PrevFBB = nullptr;
      SmallVector<MachineOperand, 4> PrevCond;
      if (FallThrough != MF.end() &&
          !FallThrough->isEHPad() &&
          !TII->analyzeBranch(PrevBB, PrevTBB, PrevFBB, PrevCond, true) &&
          PrevBB.isSuccessor(&*FallThrough)) {
        MBB->moveAfter(&MF.back());
        MadeChange = true;
        return MadeChange;
      }
    }
  }

  return MadeChange;
}

//===----------------------------------------------------------------------===//
//  Hoist Common Code
//===----------------------------------------------------------------------===//

bool BranchFolder::HoistCommonCode(MachineFunction &MF) {
  bool MadeChange = false;
  for (MachineFunction::iterator I = MF.begin(), E = MF.end(); I != E; ) {
    MachineBasicBlock *MBB = &*I++;
    MadeChange |= HoistCommonCodeInSuccs(MBB);
  }

  return MadeChange;
}

/// findFalseBlock - BB has a fallthrough. Find its 'false' successor given
/// its 'true' successor.
static MachineBasicBlock *findFalseBlock(MachineBasicBlock *BB,
                                         MachineBasicBlock *TrueBB) {
  for (MachineBasicBlock *SuccBB : BB->successors())
    if (SuccBB != TrueBB)
      return SuccBB;
  return nullptr;
}

template <class Container>
static void addRegAndItsAliases(unsigned Reg, const TargetRegisterInfo *TRI,
                                Container &Set) {
  if (TargetRegisterInfo::isPhysicalRegister(Reg)) {
    for (MCRegAliasIterator AI(Reg, TRI, true); AI.isValid(); ++AI)
      Set.insert(*AI);
  } else {
    Set.insert(Reg);
  }
}

/// findHoistingInsertPosAndDeps - Find the location to move common instructions
/// in successors to. The location is usually just before the terminator,
/// however if the terminator is a conditional branch and its previous
/// instruction is the flag setting instruction, the previous instruction is
/// the preferred location. This function also gathers uses and defs of the
/// instructions from the insertion point to the end of the block. The data is
/// used by HoistCommonCodeInSuccs to ensure safety.
static
MachineBasicBlock::iterator findHoistingInsertPosAndDeps(MachineBasicBlock *MBB,
                                                  const TargetInstrInfo *TII,
                                                  const TargetRegisterInfo *TRI,
                                                  SmallSet<unsigned,4> &Uses,
                                                  SmallSet<unsigned,4> &Defs) {
  MachineBasicBlock::iterator Loc = MBB->getFirstTerminator();
  if (!TII->isUnpredicatedTerminator(*Loc))
    return MBB->end();

  for (const MachineOperand &MO : Loc->operands()) {
    if (!MO.isReg())
      continue;
    unsigned Reg = MO.getReg();
    if (!Reg)
      continue;
    if (MO.isUse()) {
      addRegAndItsAliases(Reg, TRI, Uses);
    } else {
      if (!MO.isDead())
        // Don't try to hoist code in the rare case the terminator defines a
        // register that is later used.
        return MBB->end();

      // If the terminator defines a register, make sure we don't hoist
      // the instruction whose def might be clobbered by the terminator.
      addRegAndItsAliases(Reg, TRI, Defs);
    }
  }

  if (Uses.empty())
    return Loc;
  if (Loc == MBB->begin())
    return MBB->end();

  // The terminator is probably a conditional branch, try not to separate the
  // branch from condition setting instruction.
  MachineBasicBlock::iterator PI =
    skipDebugInstructionsBackward(std::prev(Loc), MBB->begin());

  bool IsDef = false;
  for (const MachineOperand &MO : PI->operands()) {
    // If PI has a regmask operand, it is probably a call. Separate away.
    if (MO.isRegMask())
      return Loc;
    if (!MO.isReg() || MO.isUse())
      continue;
    unsigned Reg = MO.getReg();
    if (!Reg)
      continue;
    if (Uses.count(Reg)) {
      IsDef = true;
      break;
    }
  }
  if (!IsDef)
    // The condition setting instruction is not just before the conditional
    // branch.
    return Loc;

  // Be conservative, don't insert instruction above something that may have
  // side-effects. And since it's potentially bad to separate flag setting
  // instruction from the conditional branch, just abort the optimization
  // completely.
  // Also avoid moving code above predicated instruction since it's hard to
  // reason about register liveness with predicated instruction.
  bool DontMoveAcrossStore = true;
  if (!PI->isSafeToMove(nullptr, DontMoveAcrossStore) || TII->isPredicated(*PI))
    return MBB->end();


  // Find out what registers are live. Note this routine is ignoring other live
  // registers which are only used by instructions in successor blocks.
  for (const MachineOperand &MO : PI->operands()) {
    if (!MO.isReg())
      continue;
    unsigned Reg = MO.getReg();
    if (!Reg)
      continue;
    if (MO.isUse()) {
      addRegAndItsAliases(Reg, TRI, Uses);
    } else {
      if (Uses.erase(Reg)) {
        if (TargetRegisterInfo::isPhysicalRegister(Reg)) {
          for (MCSubRegIterator SubRegs(Reg, TRI); SubRegs.isValid(); ++SubRegs)
            Uses.erase(*SubRegs); // Use sub-registers to be conservative
        }
      }
      addRegAndItsAliases(Reg, TRI, Defs);
    }
  }

  return PI;
}

bool BranchFolder::HoistCommonCodeInSuccs(MachineBasicBlock *MBB) {
  MachineBasicBlock *TBB = nullptr, *FBB = nullptr;
  SmallVector<MachineOperand, 4> Cond;
  if (TII->analyzeBranch(*MBB, TBB, FBB, Cond, true) || !TBB || Cond.empty())
    return false;

  if (!FBB) FBB = findFalseBlock(MBB, TBB);
  if (!FBB)
    // Malformed bcc? True and false blocks are the same?
    return false;

  // Restrict the optimization to cases where MBB is the only predecessor,
  // it is an obvious win.
  if (TBB->pred_size() > 1 || FBB->pred_size() > 1)
    return false;

  // Find a suitable position to hoist the common instructions to. Also figure
  // out which registers are used or defined by instructions from the insertion
  // point to the end of the block.
  SmallSet<unsigned, 4> Uses, Defs;
  MachineBasicBlock::iterator Loc =
    findHoistingInsertPosAndDeps(MBB, TII, TRI, Uses, Defs);
  if (Loc == MBB->end())
    return false;

  bool HasDups = false;
  SmallVector<unsigned, 4> LocalDefs, LocalKills;
  SmallSet<unsigned, 4> ActiveDefsSet, AllDefsSet;
  MachineBasicBlock::iterator TIB = TBB->begin();
  MachineBasicBlock::iterator FIB = FBB->begin();
  MachineBasicBlock::iterator TIE = TBB->end();
  MachineBasicBlock::iterator FIE = FBB->end();
  while (TIB != TIE && FIB != FIE) {
    // Skip dbg_value instructions. These do not count.
    TIB = skipDebugInstructionsForward(TIB, TIE);
    FIB = skipDebugInstructionsForward(FIB, FIE);
    if (TIB == TIE || FIB == FIE)
      break;

    if (!TIB->isIdenticalTo(*FIB, MachineInstr::CheckKillDead))
      break;

    if (TII->isPredicated(*TIB))
      // Hard to reason about register liveness with predicated instruction.
      break;

    bool IsSafe = true;
    for (MachineOperand &MO : TIB->operands()) {
      // Don't attempt to hoist instructions with register masks.
      if (MO.isRegMask()) {
        IsSafe = false;
        break;
      }
      if (!MO.isReg())
        continue;
      unsigned Reg = MO.getReg();
      if (!Reg)
        continue;
      if (MO.isDef()) {
        if (Uses.count(Reg)) {
          // Avoid clobbering a register that's used by the instruction at
          // the point of insertion.
          IsSafe = false;
          break;
        }

        if (Defs.count(Reg) && !MO.isDead()) {
          // Don't hoist the instruction if the def would be clobber by the
          // instruction at the point insertion. FIXME: This is overly
          // conservative. It should be possible to hoist the instructions
          // in BB2 in the following example:
          // BB1:
          // r1, eflag = op1 r2, r3
          // brcc eflag
          //
          // BB2:
          // r1 = op2, ...
          //    = op3, r1<kill>
          IsSafe = false;
          break;
        }
      } else if (!ActiveDefsSet.count(Reg)) {
        if (Defs.count(Reg)) {
          // Use is defined by the instruction at the point of insertion.
          IsSafe = false;
          break;
        }

        if (MO.isKill() && Uses.count(Reg))
          // Kills a register that's read by the instruction at the point of
          // insertion. Remove the kill marker.
          MO.setIsKill(false);
      }
    }
    if (!IsSafe)
      break;

    bool DontMoveAcrossStore = true;
    if (!TIB->isSafeToMove(nullptr, DontMoveAcrossStore))
      break;

    // Remove kills from ActiveDefsSet, these registers had short live ranges.
    for (const MachineOperand &MO : TIB->operands()) {
      if (!MO.isReg() || !MO.isUse() || !MO.isKill())
        continue;
      unsigned Reg = MO.getReg();
      if (!Reg)
        continue;
      if (!AllDefsSet.count(Reg)) {
        LocalKills.push_back(Reg);
        continue;
      }
      if (TargetRegisterInfo::isPhysicalRegister(Reg)) {
        for (MCRegAliasIterator AI(Reg, TRI, true); AI.isValid(); ++AI)
          ActiveDefsSet.erase(*AI);
      } else {
        ActiveDefsSet.erase(Reg);
      }
    }

    // Track local defs so we can update liveins.
    for (const MachineOperand &MO : TIB->operands()) {
      if (!MO.isReg() || !MO.isDef() || MO.isDead())
        continue;
      unsigned Reg = MO.getReg();
      if (!Reg || TargetRegisterInfo::isVirtualRegister(Reg))
        continue;
      LocalDefs.push_back(Reg);
      addRegAndItsAliases(Reg, TRI, ActiveDefsSet);
      addRegAndItsAliases(Reg, TRI, AllDefsSet);
    }

    HasDups = true;
    ++TIB;
    ++FIB;
  }

  if (!HasDups)
    return false;

  MBB->splice(Loc, TBB, TBB->begin(), TIB);
  FBB->erase(FBB->begin(), FIB);

  // Update livein's.
  bool ChangedLiveIns = false;
  for (unsigned i = 0, e = LocalDefs.size(); i != e; ++i) {
    unsigned Def = LocalDefs[i];
    if (ActiveDefsSet.count(Def)) {
      TBB->addLiveIn(Def);
      FBB->addLiveIn(Def);
      ChangedLiveIns = true;
    }
  }
  for (unsigned K : LocalKills) {
    TBB->removeLiveIn(K);
    FBB->removeLiveIn(K);
    ChangedLiveIns = true;
  }

  if (ChangedLiveIns) {
    TBB->sortUniqueLiveIns();
    FBB->sortUniqueLiveIns();
  }

  ++NumHoist;
  return true;
}