llvm.org GIT mirror llvm / 15c435a lib / CodeGen / MachineLICM.cpp
15c435a

Tree @15c435a (Download .tar.gz)

MachineLICM.cpp @15c435araw · 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
//===-- MachineLICM.cpp - Machine Loop Invariant Code Motion Pass ---------===//
//
//                     The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This pass performs loop invariant code motion on machine instructions. We
// attempt to remove as much code from the body of a loop as possible.
//
// This pass does not attempt to throttle itself to limit register pressure.
// The register allocation phases are expected to perform rematerialization
// to recover when register pressure is high.
//
// This pass is not intended to be a replacement or a complete alternative
// for the LLVM-IR-level LICM pass. It is only designed to hoist simple
// constructs that are not exposed before lowering and instruction selection.
//
//===----------------------------------------------------------------------===//

#define DEBUG_TYPE "machine-licm"
#include "llvm/CodeGen/Passes.h"
#include "llvm/ADT/DenseMap.h"
#include "llvm/ADT/SmallSet.h"
#include "llvm/ADT/Statistic.h"
#include "llvm/Analysis/AliasAnalysis.h"
#include "llvm/CodeGen/MachineDominators.h"
#include "llvm/CodeGen/MachineFrameInfo.h"
#include "llvm/CodeGen/MachineLoopInfo.h"
#include "llvm/CodeGen/MachineMemOperand.h"
#include "llvm/CodeGen/MachineRegisterInfo.h"
#include "llvm/CodeGen/PseudoSourceValue.h"
#include "llvm/MC/MCInstrItineraries.h"
#include "llvm/Support/CommandLine.h"
#include "llvm/Support/Debug.h"
#include "llvm/Support/raw_ostream.h"
#include "llvm/Target/TargetInstrInfo.h"
#include "llvm/Target/TargetLowering.h"
#include "llvm/Target/TargetMachine.h"
#include "llvm/Target/TargetRegisterInfo.h"
using namespace llvm;

static cl::opt<bool>
AvoidSpeculation("avoid-speculation",
                 cl::desc("MachineLICM should avoid speculation"),
                 cl::init(true), cl::Hidden);

STATISTIC(NumHoisted,
          "Number of machine instructions hoisted out of loops");
STATISTIC(NumLowRP,
          "Number of instructions hoisted in low reg pressure situation");
STATISTIC(NumHighLatency,
          "Number of high latency instructions hoisted");
STATISTIC(NumCSEed,
          "Number of hoisted machine instructions CSEed");
STATISTIC(NumPostRAHoisted,
          "Number of machine instructions hoisted out of loops post regalloc");

namespace {
  class MachineLICM : public MachineFunctionPass {
    const TargetMachine   *TM;
    const TargetInstrInfo *TII;
    const TargetLoweringBase *TLI;
    const TargetRegisterInfo *TRI;
    const MachineFrameInfo *MFI;
    MachineRegisterInfo *MRI;
    const InstrItineraryData *InstrItins;
    bool PreRegAlloc;

    // Various analyses that we use...
    AliasAnalysis        *AA;      // Alias analysis info.
    MachineLoopInfo      *MLI;     // Current MachineLoopInfo
    MachineDominatorTree *DT;      // Machine dominator tree for the cur loop

    // State that is updated as we process loops
    bool         Changed;          // True if a loop is changed.
    bool         FirstInLoop;      // True if it's the first LICM in the loop.
    MachineLoop *CurLoop;          // The current loop we are working on.
    MachineBasicBlock *CurPreheader; // The preheader for CurLoop.

    // Exit blocks for CurLoop.
    SmallVector<MachineBasicBlock*, 8> ExitBlocks;

    bool isExitBlock(const MachineBasicBlock *MBB) const {
      return std::find(ExitBlocks.begin(), ExitBlocks.end(), MBB) !=
        ExitBlocks.end();
    }

    // Track 'estimated' register pressure.
    SmallSet<unsigned, 32> RegSeen;
    SmallVector<unsigned, 8> RegPressure;

    // Register pressure "limit" per register class. If the pressure
    // is higher than the limit, then it's considered high.
    SmallVector<unsigned, 8> RegLimit;

    // Register pressure on path leading from loop preheader to current BB.
    SmallVector<SmallVector<unsigned, 8>, 16> BackTrace;

    // For each opcode, keep a list of potential CSE instructions.
    DenseMap<unsigned, std::vector<const MachineInstr*> > CSEMap;

    enum {
      SpeculateFalse   = 0,
      SpeculateTrue    = 1,
      SpeculateUnknown = 2
    };

    // If a MBB does not dominate loop exiting blocks then it may not safe
    // to hoist loads from this block.
    // Tri-state: 0 - false, 1 - true, 2 - unknown
    unsigned SpeculationState;

  public:
    static char ID; // Pass identification, replacement for typeid
    MachineLICM() :
      MachineFunctionPass(ID), PreRegAlloc(true) {
        initializeMachineLICMPass(*PassRegistry::getPassRegistry());
      }

    explicit MachineLICM(bool PreRA) :
      MachineFunctionPass(ID), PreRegAlloc(PreRA) {
        initializeMachineLICMPass(*PassRegistry::getPassRegistry());
      }

    bool runOnMachineFunction(MachineFunction &MF) override;

    void getAnalysisUsage(AnalysisUsage &AU) const override {
      AU.addRequired<MachineLoopInfo>();
      AU.addRequired<MachineDominatorTree>();
      AU.addRequired<AliasAnalysis>();
      AU.addPreserved<MachineLoopInfo>();
      AU.addPreserved<MachineDominatorTree>();
      MachineFunctionPass::getAnalysisUsage(AU);
    }

    void releaseMemory() override {
      RegSeen.clear();
      RegPressure.clear();
      RegLimit.clear();
      BackTrace.clear();
      for (DenseMap<unsigned,std::vector<const MachineInstr*> >::iterator
             CI = CSEMap.begin(), CE = CSEMap.end(); CI != CE; ++CI)
        CI->second.clear();
      CSEMap.clear();
    }

  private:
    /// CandidateInfo - Keep track of information about hoisting candidates.
    struct CandidateInfo {
      MachineInstr *MI;
      unsigned      Def;
      int           FI;
      CandidateInfo(MachineInstr *mi, unsigned def, int fi)
        : MI(mi), Def(def), FI(fi) {}
    };

    /// HoistRegionPostRA - Walk the specified region of the CFG and hoist loop
    /// invariants out to the preheader.
    void HoistRegionPostRA();

    /// HoistPostRA - When an instruction is found to only use loop invariant
    /// operands that is safe to hoist, this instruction is called to do the
    /// dirty work.
    void HoistPostRA(MachineInstr *MI, unsigned Def);

    /// ProcessMI - Examine the instruction for potentai LICM candidate. Also
    /// gather register def and frame object update information.
    void ProcessMI(MachineInstr *MI,
                   BitVector &PhysRegDefs,
                   BitVector &PhysRegClobbers,
                   SmallSet<int, 32> &StoredFIs,
                   SmallVectorImpl<CandidateInfo> &Candidates);

    /// AddToLiveIns - Add register 'Reg' to the livein sets of BBs in the
    /// current loop.
    void AddToLiveIns(unsigned Reg);

    /// IsLICMCandidate - Returns true if the instruction may be a suitable
    /// candidate for LICM. e.g. If the instruction is a call, then it's
    /// obviously not safe to hoist it.
    bool IsLICMCandidate(MachineInstr &I);

    /// IsLoopInvariantInst - Returns true if the instruction is loop
    /// invariant. I.e., all virtual register operands are defined outside of
    /// the loop, physical registers aren't accessed (explicitly or implicitly),
    /// and the instruction is hoistable.
    ///
    bool IsLoopInvariantInst(MachineInstr &I);

    /// HasLoopPHIUse - Return true if the specified instruction is used by any
    /// phi node in the current loop.
    bool HasLoopPHIUse(const MachineInstr *MI) const;

    /// HasHighOperandLatency - Compute operand latency between a def of 'Reg'
    /// and an use in the current loop, return true if the target considered
    /// it 'high'.
    bool HasHighOperandLatency(MachineInstr &MI, unsigned DefIdx,
                               unsigned Reg) const;

    bool IsCheapInstruction(MachineInstr &MI) const;

    /// CanCauseHighRegPressure - Visit BBs from header to current BB,
    /// check if hoisting an instruction of the given cost matrix can cause high
    /// register pressure.
    bool CanCauseHighRegPressure(DenseMap<unsigned, int> &Cost, bool Cheap);

    /// UpdateBackTraceRegPressure - Traverse the back trace from header to
    /// the current block and update their register pressures to reflect the
    /// effect of hoisting MI from the current block to the preheader.
    void UpdateBackTraceRegPressure(const MachineInstr *MI);

    /// IsProfitableToHoist - Return true if it is potentially profitable to
    /// hoist the given loop invariant.
    bool IsProfitableToHoist(MachineInstr &MI);

    /// IsGuaranteedToExecute - Check if this mbb is guaranteed to execute.
    /// If not then a load from this mbb may not be safe to hoist.
    bool IsGuaranteedToExecute(MachineBasicBlock *BB);

    void EnterScope(MachineBasicBlock *MBB);

    void ExitScope(MachineBasicBlock *MBB);

    /// ExitScopeIfDone - Destroy scope for the MBB that corresponds to given
    /// dominator tree node if its a leaf or all of its children are done. Walk
    /// up the dominator tree to destroy ancestors which are now done.
    void ExitScopeIfDone(MachineDomTreeNode *Node,
                DenseMap<MachineDomTreeNode*, unsigned> &OpenChildren,
                DenseMap<MachineDomTreeNode*, MachineDomTreeNode*> &ParentMap);

    /// HoistOutOfLoop - Walk the specified loop in the CFG (defined by all
    /// blocks dominated by the specified header block, and that are in the
    /// current loop) in depth first order w.r.t the DominatorTree. This allows
    /// us to visit definitions before uses, allowing us to hoist a loop body in
    /// one pass without iteration.
    ///
    void HoistOutOfLoop(MachineDomTreeNode *LoopHeaderNode);
    void HoistRegion(MachineDomTreeNode *N, bool IsHeader);

    /// getRegisterClassIDAndCost - For a given MI, register, and the operand
    /// index, return the ID and cost of its representative register class by
    /// reference.
    void getRegisterClassIDAndCost(const MachineInstr *MI,
                                   unsigned Reg, unsigned OpIdx,
                                   unsigned &RCId, unsigned &RCCost) const;

    /// InitRegPressure - Find all virtual register references that are liveout
    /// of the preheader to initialize the starting "register pressure". Note
    /// this does not count live through (livein but not used) registers.
    void InitRegPressure(MachineBasicBlock *BB);

    /// UpdateRegPressure - Update estimate of register pressure after the
    /// specified instruction.
    void UpdateRegPressure(const MachineInstr *MI);

    /// ExtractHoistableLoad - Unfold a load from the given machineinstr if
    /// the load itself could be hoisted. Return the unfolded and hoistable
    /// load, or null if the load couldn't be unfolded or if it wouldn't
    /// be hoistable.
    MachineInstr *ExtractHoistableLoad(MachineInstr *MI);

    /// LookForDuplicate - Find an instruction amount PrevMIs that is a
    /// duplicate of MI. Return this instruction if it's found.
    const MachineInstr *LookForDuplicate(const MachineInstr *MI,
                                     std::vector<const MachineInstr*> &PrevMIs);

    /// EliminateCSE - Given a LICM'ed instruction, look for an instruction on
    /// the preheader that compute the same value. If it's found, do a RAU on
    /// with the definition of the existing instruction rather than hoisting
    /// the instruction to the preheader.
    bool EliminateCSE(MachineInstr *MI,
           DenseMap<unsigned, std::vector<const MachineInstr*> >::iterator &CI);

    /// MayCSE - Return true if the given instruction will be CSE'd if it's
    /// hoisted out of the loop.
    bool MayCSE(MachineInstr *MI);

    /// Hoist - When an instruction is found to only use loop invariant operands
    /// that is safe to hoist, this instruction is called to do the dirty work.
    /// It returns true if the instruction is hoisted.
    bool Hoist(MachineInstr *MI, MachineBasicBlock *Preheader);

    /// InitCSEMap - Initialize the CSE map with instructions that are in the
    /// current loop preheader that may become duplicates of instructions that
    /// are hoisted out of the loop.
    void InitCSEMap(MachineBasicBlock *BB);

    /// getCurPreheader - Get the preheader for the current loop, splitting
    /// a critical edge if needed.
    MachineBasicBlock *getCurPreheader();
  };
} // end anonymous namespace

char MachineLICM::ID = 0;
char &llvm::MachineLICMID = MachineLICM::ID;
INITIALIZE_PASS_BEGIN(MachineLICM, "machinelicm",
                "Machine Loop Invariant Code Motion", false, false)
INITIALIZE_PASS_DEPENDENCY(MachineLoopInfo)
INITIALIZE_PASS_DEPENDENCY(MachineDominatorTree)
INITIALIZE_AG_DEPENDENCY(AliasAnalysis)
INITIALIZE_PASS_END(MachineLICM, "machinelicm",
                "Machine Loop Invariant Code Motion", false, false)

/// LoopIsOuterMostWithPredecessor - Test if the given loop is the outer-most
/// loop that has a unique predecessor.
static bool LoopIsOuterMostWithPredecessor(MachineLoop *CurLoop) {
  // Check whether this loop even has a unique predecessor.
  if (!CurLoop->getLoopPredecessor())
    return false;
  // Ok, now check to see if any of its outer loops do.
  for (MachineLoop *L = CurLoop->getParentLoop(); L; L = L->getParentLoop())
    if (L->getLoopPredecessor())
      return false;
  // None of them did, so this is the outermost with a unique predecessor.
  return true;
}

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

  Changed = FirstInLoop = false;
  TM = &MF.getTarget();
  TII = TM->getInstrInfo();
  TLI = TM->getTargetLowering();
  TRI = TM->getRegisterInfo();
  MFI = MF.getFrameInfo();
  MRI = &MF.getRegInfo();
  InstrItins = TM->getInstrItineraryData();

  PreRegAlloc = MRI->isSSA();

  if (PreRegAlloc)
    DEBUG(dbgs() << "******** Pre-regalloc Machine LICM: ");
  else
    DEBUG(dbgs() << "******** Post-regalloc Machine LICM: ");
  DEBUG(dbgs() << MF.getName() << " ********\n");

  if (PreRegAlloc) {
    // Estimate register pressure during pre-regalloc pass.
    unsigned NumRC = TRI->getNumRegClasses();
    RegPressure.resize(NumRC);
    std::fill(RegPressure.begin(), RegPressure.end(), 0);
    RegLimit.resize(NumRC);
    for (TargetRegisterInfo::regclass_iterator I = TRI->regclass_begin(),
           E = TRI->regclass_end(); I != E; ++I)
      RegLimit[(*I)->getID()] = TRI->getRegPressureLimit(*I, MF);
  }

  // Get our Loop information...
  MLI = &getAnalysis<MachineLoopInfo>();
  DT  = &getAnalysis<MachineDominatorTree>();
  AA  = &getAnalysis<AliasAnalysis>();

  SmallVector<MachineLoop *, 8> Worklist(MLI->begin(), MLI->end());
  while (!Worklist.empty()) {
    CurLoop = Worklist.pop_back_val();
    CurPreheader = 0;
    ExitBlocks.clear();

    // If this is done before regalloc, only visit outer-most preheader-sporting
    // loops.
    if (PreRegAlloc && !LoopIsOuterMostWithPredecessor(CurLoop)) {
      Worklist.append(CurLoop->begin(), CurLoop->end());
      continue;
    }

    CurLoop->getExitBlocks(ExitBlocks);

    if (!PreRegAlloc)
      HoistRegionPostRA();
    else {
      // CSEMap is initialized for loop header when the first instruction is
      // being hoisted.
      MachineDomTreeNode *N = DT->getNode(CurLoop->getHeader());
      FirstInLoop = true;
      HoistOutOfLoop(N);
      CSEMap.clear();
    }
  }

  return Changed;
}

/// InstructionStoresToFI - Return true if instruction stores to the
/// specified frame.
static bool InstructionStoresToFI(const MachineInstr *MI, int FI) {
  for (MachineInstr::mmo_iterator o = MI->memoperands_begin(),
         oe = MI->memoperands_end(); o != oe; ++o) {
    if (!(*o)->isStore() || !(*o)->getValue())
      continue;
    if (const FixedStackPseudoSourceValue *Value =
        dyn_cast<const FixedStackPseudoSourceValue>((*o)->getValue())) {
      if (Value->getFrameIndex() == FI)
        return true;
    }
  }
  return false;
}

/// ProcessMI - Examine the instruction for potentai LICM candidate. Also
/// gather register def and frame object update information.
void MachineLICM::ProcessMI(MachineInstr *MI,
                            BitVector &PhysRegDefs,
                            BitVector &PhysRegClobbers,
                            SmallSet<int, 32> &StoredFIs,
                            SmallVectorImpl<CandidateInfo> &Candidates) {
  bool RuledOut = false;
  bool HasNonInvariantUse = false;
  unsigned Def = 0;
  for (unsigned i = 0, e = MI->getNumOperands(); i != e; ++i) {
    const MachineOperand &MO = MI->getOperand(i);
    if (MO.isFI()) {
      // Remember if the instruction stores to the frame index.
      int FI = MO.getIndex();
      if (!StoredFIs.count(FI) &&
          MFI->isSpillSlotObjectIndex(FI) &&
          InstructionStoresToFI(MI, FI))
        StoredFIs.insert(FI);
      HasNonInvariantUse = true;
      continue;
    }

    // We can't hoist an instruction defining a physreg that is clobbered in
    // the loop.
    if (MO.isRegMask()) {
      PhysRegClobbers.setBitsNotInMask(MO.getRegMask());
      continue;
    }

    if (!MO.isReg())
      continue;
    unsigned Reg = MO.getReg();
    if (!Reg)
      continue;
    assert(TargetRegisterInfo::isPhysicalRegister(Reg) &&
           "Not expecting virtual register!");

    if (!MO.isDef()) {
      if (Reg && (PhysRegDefs.test(Reg) || PhysRegClobbers.test(Reg)))
        // If it's using a non-loop-invariant register, then it's obviously not
        // safe to hoist.
        HasNonInvariantUse = true;
      continue;
    }

    if (MO.isImplicit()) {
      for (MCRegAliasIterator AI(Reg, TRI, true); AI.isValid(); ++AI)
        PhysRegClobbers.set(*AI);
      if (!MO.isDead())
        // Non-dead implicit def? This cannot be hoisted.
        RuledOut = true;
      // No need to check if a dead implicit def is also defined by
      // another instruction.
      continue;
    }

    // FIXME: For now, avoid instructions with multiple defs, unless
    // it's a dead implicit def.
    if (Def)
      RuledOut = true;
    else
      Def = Reg;

    // If we have already seen another instruction that defines the same
    // register, then this is not safe.  Two defs is indicated by setting a
    // PhysRegClobbers bit.
    for (MCRegAliasIterator AS(Reg, TRI, true); AS.isValid(); ++AS) {
      if (PhysRegDefs.test(*AS))
        PhysRegClobbers.set(*AS);
      PhysRegDefs.set(*AS);
    }
    if (PhysRegClobbers.test(Reg))
      // MI defined register is seen defined by another instruction in
      // the loop, it cannot be a LICM candidate.
      RuledOut = true;
  }

  // Only consider reloads for now and remats which do not have register
  // operands. FIXME: Consider unfold load folding instructions.
  if (Def && !RuledOut) {
    int FI = INT_MIN;
    if ((!HasNonInvariantUse && IsLICMCandidate(*MI)) ||
        (TII->isLoadFromStackSlot(MI, FI) && MFI->isSpillSlotObjectIndex(FI)))
      Candidates.push_back(CandidateInfo(MI, Def, FI));
  }
}

/// HoistRegionPostRA - Walk the specified region of the CFG and hoist loop
/// invariants out to the preheader.
void MachineLICM::HoistRegionPostRA() {
  MachineBasicBlock *Preheader = getCurPreheader();
  if (!Preheader)
    return;

  unsigned NumRegs = TRI->getNumRegs();
  BitVector PhysRegDefs(NumRegs); // Regs defined once in the loop.
  BitVector PhysRegClobbers(NumRegs); // Regs defined more than once.

  SmallVector<CandidateInfo, 32> Candidates;
  SmallSet<int, 32> StoredFIs;

  // Walk the entire region, count number of defs for each register, and
  // collect potential LICM candidates.
  const std::vector<MachineBasicBlock *> &Blocks = CurLoop->getBlocks();
  for (unsigned i = 0, e = Blocks.size(); i != e; ++i) {
    MachineBasicBlock *BB = Blocks[i];

    // If the header of the loop containing this basic block is a landing pad,
    // then don't try to hoist instructions out of this loop.
    const MachineLoop *ML = MLI->getLoopFor(BB);
    if (ML && ML->getHeader()->isLandingPad()) continue;

    // Conservatively treat live-in's as an external def.
    // FIXME: That means a reload that're reused in successor block(s) will not
    // be LICM'ed.
    for (MachineBasicBlock::livein_iterator I = BB->livein_begin(),
           E = BB->livein_end(); I != E; ++I) {
      unsigned Reg = *I;
      for (MCRegAliasIterator AI(Reg, TRI, true); AI.isValid(); ++AI)
        PhysRegDefs.set(*AI);
    }

    SpeculationState = SpeculateUnknown;
    for (MachineBasicBlock::iterator
           MII = BB->begin(), E = BB->end(); MII != E; ++MII) {
      MachineInstr *MI = &*MII;
      ProcessMI(MI, PhysRegDefs, PhysRegClobbers, StoredFIs, Candidates);
    }
  }

  // Gather the registers read / clobbered by the terminator.
  BitVector TermRegs(NumRegs);
  MachineBasicBlock::iterator TI = Preheader->getFirstTerminator();
  if (TI != Preheader->end()) {
    for (unsigned i = 0, e = TI->getNumOperands(); i != e; ++i) {
      const MachineOperand &MO = TI->getOperand(i);
      if (!MO.isReg())
        continue;
      unsigned Reg = MO.getReg();
      if (!Reg)
        continue;
      for (MCRegAliasIterator AI(Reg, TRI, true); AI.isValid(); ++AI)
        TermRegs.set(*AI);
    }
  }

  // Now evaluate whether the potential candidates qualify.
  // 1. Check if the candidate defined register is defined by another
  //    instruction in the loop.
  // 2. If the candidate is a load from stack slot (always true for now),
  //    check if the slot is stored anywhere in the loop.
  // 3. Make sure candidate def should not clobber
  //    registers read by the terminator. Similarly its def should not be
  //    clobbered by the terminator.
  for (unsigned i = 0, e = Candidates.size(); i != e; ++i) {
    if (Candidates[i].FI != INT_MIN &&
        StoredFIs.count(Candidates[i].FI))
      continue;

    unsigned Def = Candidates[i].Def;
    if (!PhysRegClobbers.test(Def) && !TermRegs.test(Def)) {
      bool Safe = true;
      MachineInstr *MI = Candidates[i].MI;
      for (unsigned j = 0, ee = MI->getNumOperands(); j != ee; ++j) {
        const MachineOperand &MO = MI->getOperand(j);
        if (!MO.isReg() || MO.isDef() || !MO.getReg())
          continue;
        unsigned Reg = MO.getReg();
        if (PhysRegDefs.test(Reg) ||
            PhysRegClobbers.test(Reg)) {
          // If it's using a non-loop-invariant register, then it's obviously
          // not safe to hoist.
          Safe = false;
          break;
        }
      }
      if (Safe)
        HoistPostRA(MI, Candidates[i].Def);
    }
  }
}

/// AddToLiveIns - Add register 'Reg' to the livein sets of BBs in the current
/// loop, and make sure it is not killed by any instructions in the loop.
void MachineLICM::AddToLiveIns(unsigned Reg) {
  const std::vector<MachineBasicBlock *> &Blocks = CurLoop->getBlocks();
  for (unsigned i = 0, e = Blocks.size(); i != e; ++i) {
    MachineBasicBlock *BB = Blocks[i];
    if (!BB->isLiveIn(Reg))
      BB->addLiveIn(Reg);
    for (MachineBasicBlock::iterator
           MII = BB->begin(), E = BB->end(); MII != E; ++MII) {
      MachineInstr *MI = &*MII;
      for (unsigned i = 0, e = MI->getNumOperands(); i != e; ++i) {
        MachineOperand &MO = MI->getOperand(i);
        if (!MO.isReg() || !MO.getReg() || MO.isDef()) continue;
        if (MO.getReg() == Reg || TRI->isSuperRegister(Reg, MO.getReg()))
          MO.setIsKill(false);
      }
    }
  }
}

/// HoistPostRA - When an instruction is found to only use loop invariant
/// operands that is safe to hoist, this instruction is called to do the
/// dirty work.
void MachineLICM::HoistPostRA(MachineInstr *MI, unsigned Def) {
  MachineBasicBlock *Preheader = getCurPreheader();

  // Now move the instructions to the predecessor, inserting it before any
  // terminator instructions.
  DEBUG(dbgs() << "Hoisting to BB#" << Preheader->getNumber() << " from BB#"
               << MI->getParent()->getNumber() << ": " << *MI);

  // Splice the instruction to the preheader.
  MachineBasicBlock *MBB = MI->getParent();
  Preheader->splice(Preheader->getFirstTerminator(), MBB, MI);

  // Add register to livein list to all the BBs in the current loop since a
  // loop invariant must be kept live throughout the whole loop. This is
  // important to ensure later passes do not scavenge the def register.
  AddToLiveIns(Def);

  ++NumPostRAHoisted;
  Changed = true;
}

// IsGuaranteedToExecute - Check if this mbb is guaranteed to execute.
// If not then a load from this mbb may not be safe to hoist.
bool MachineLICM::IsGuaranteedToExecute(MachineBasicBlock *BB) {
  if (SpeculationState != SpeculateUnknown)
    return SpeculationState == SpeculateFalse;

  if (BB != CurLoop->getHeader()) {
    // Check loop exiting blocks.
    SmallVector<MachineBasicBlock*, 8> CurrentLoopExitingBlocks;
    CurLoop->getExitingBlocks(CurrentLoopExitingBlocks);
    for (unsigned i = 0, e = CurrentLoopExitingBlocks.size(); i != e; ++i)
      if (!DT->dominates(BB, CurrentLoopExitingBlocks[i])) {
        SpeculationState = SpeculateTrue;
        return false;
      }
  }

  SpeculationState = SpeculateFalse;
  return true;
}

void MachineLICM::EnterScope(MachineBasicBlock *MBB) {
  DEBUG(dbgs() << "Entering: " << MBB->getName() << '\n');

  // Remember livein register pressure.
  BackTrace.push_back(RegPressure);
}

void MachineLICM::ExitScope(MachineBasicBlock *MBB) {
  DEBUG(dbgs() << "Exiting: " << MBB->getName() << '\n');
  BackTrace.pop_back();
}

/// ExitScopeIfDone - Destroy scope for the MBB that corresponds to the given
/// dominator tree node if its a leaf or all of its children are done. Walk
/// up the dominator tree to destroy ancestors which are now done.
void MachineLICM::ExitScopeIfDone(MachineDomTreeNode *Node,
                DenseMap<MachineDomTreeNode*, unsigned> &OpenChildren,
                DenseMap<MachineDomTreeNode*, MachineDomTreeNode*> &ParentMap) {
  if (OpenChildren[Node])
    return;

  // Pop scope.
  ExitScope(Node->getBlock());

  // Now traverse upwards to pop ancestors whose offsprings are all done.
  while (MachineDomTreeNode *Parent = ParentMap[Node]) {
    unsigned Left = --OpenChildren[Parent];
    if (Left != 0)
      break;
    ExitScope(Parent->getBlock());
    Node = Parent;
  }
}

/// HoistOutOfLoop - Walk the specified loop in the CFG (defined by all
/// blocks dominated by the specified header block, and that are in the
/// current loop) in depth first order w.r.t the DominatorTree. This allows
/// us to visit definitions before uses, allowing us to hoist a loop body in
/// one pass without iteration.
///
void MachineLICM::HoistOutOfLoop(MachineDomTreeNode *HeaderN) {
  SmallVector<MachineDomTreeNode*, 32> Scopes;
  SmallVector<MachineDomTreeNode*, 8> WorkList;
  DenseMap<MachineDomTreeNode*, MachineDomTreeNode*> ParentMap;
  DenseMap<MachineDomTreeNode*, unsigned> OpenChildren;

  // Perform a DFS walk to determine the order of visit.
  WorkList.push_back(HeaderN);
  do {
    MachineDomTreeNode *Node = WorkList.pop_back_val();
    assert(Node != 0 && "Null dominator tree node?");
    MachineBasicBlock *BB = Node->getBlock();

    // If the header of the loop containing this basic block is a landing pad,
    // then don't try to hoist instructions out of this loop.
    const MachineLoop *ML = MLI->getLoopFor(BB);
    if (ML && ML->getHeader()->isLandingPad())
      continue;

    // If this subregion is not in the top level loop at all, exit.
    if (!CurLoop->contains(BB))
      continue;

    Scopes.push_back(Node);
    const std::vector<MachineDomTreeNode*> &Children = Node->getChildren();
    unsigned NumChildren = Children.size();

    // Don't hoist things out of a large switch statement.  This often causes
    // code to be hoisted that wasn't going to be executed, and increases
    // register pressure in a situation where it's likely to matter.
    if (BB->succ_size() >= 25)
      NumChildren = 0;

    OpenChildren[Node] = NumChildren;
    // Add children in reverse order as then the next popped worklist node is
    // the first child of this node.  This means we ultimately traverse the
    // DOM tree in exactly the same order as if we'd recursed.
    for (int i = (int)NumChildren-1; i >= 0; --i) {
      MachineDomTreeNode *Child = Children[i];
      ParentMap[Child] = Node;
      WorkList.push_back(Child);
    }
  } while (!WorkList.empty());

  if (Scopes.size() != 0) {
    MachineBasicBlock *Preheader = getCurPreheader();
    if (!Preheader)
      return;

    // Compute registers which are livein into the loop headers.
    RegSeen.clear();
    BackTrace.clear();
    InitRegPressure(Preheader);
  }

  // Now perform LICM.
  for (unsigned i = 0, e = Scopes.size(); i != e; ++i) {
    MachineDomTreeNode *Node = Scopes[i];
    MachineBasicBlock *MBB = Node->getBlock();

    MachineBasicBlock *Preheader = getCurPreheader();
    if (!Preheader)
      continue;

    EnterScope(MBB);

    // Process the block
    SpeculationState = SpeculateUnknown;
    for (MachineBasicBlock::iterator
         MII = MBB->begin(), E = MBB->end(); MII != E; ) {
      MachineBasicBlock::iterator NextMII = MII; ++NextMII;
      MachineInstr *MI = &*MII;
      if (!Hoist(MI, Preheader))
        UpdateRegPressure(MI);
      MII = NextMII;
    }

    // If it's a leaf node, it's done. Traverse upwards to pop ancestors.
    ExitScopeIfDone(Node, OpenChildren, ParentMap);
  }
}

static bool isOperandKill(const MachineOperand &MO, MachineRegisterInfo *MRI) {
  return MO.isKill() || MRI->hasOneNonDBGUse(MO.getReg());
}

/// getRegisterClassIDAndCost - For a given MI, register, and the operand
/// index, return the ID and cost of its representative register class.
void
MachineLICM::getRegisterClassIDAndCost(const MachineInstr *MI,
                                       unsigned Reg, unsigned OpIdx,
                                       unsigned &RCId, unsigned &RCCost) const {
  const TargetRegisterClass *RC = MRI->getRegClass(Reg);
  MVT VT = *RC->vt_begin();
  if (VT == MVT::Untyped) {
    RCId = RC->getID();
    RCCost = 1;
  } else {
    RCId = TLI->getRepRegClassFor(VT)->getID();
    RCCost = TLI->getRepRegClassCostFor(VT);
  }
}

/// InitRegPressure - Find all virtual register references that are liveout of
/// the preheader to initialize the starting "register pressure". Note this
/// does not count live through (livein but not used) registers.
void MachineLICM::InitRegPressure(MachineBasicBlock *BB) {
  std::fill(RegPressure.begin(), RegPressure.end(), 0);

  // If the preheader has only a single predecessor and it ends with a
  // fallthrough or an unconditional branch, then scan its predecessor for live
  // defs as well. This happens whenever the preheader is created by splitting
  // the critical edge from the loop predecessor to the loop header.
  if (BB->pred_size() == 1) {
    MachineBasicBlock *TBB = 0, *FBB = 0;
    SmallVector<MachineOperand, 4> Cond;
    if (!TII->AnalyzeBranch(*BB, TBB, FBB, Cond, false) && Cond.empty())
      InitRegPressure(*BB->pred_begin());
  }

  for (MachineBasicBlock::iterator MII = BB->begin(), E = BB->end();
       MII != E; ++MII) {
    MachineInstr *MI = &*MII;
    for (unsigned i = 0, e = MI->getDesc().getNumOperands(); i != e; ++i) {
      const MachineOperand &MO = MI->getOperand(i);
      if (!MO.isReg() || MO.isImplicit())
        continue;
      unsigned Reg = MO.getReg();
      if (!TargetRegisterInfo::isVirtualRegister(Reg))
        continue;

      bool isNew = RegSeen.insert(Reg);
      unsigned RCId, RCCost;
      getRegisterClassIDAndCost(MI, Reg, i, RCId, RCCost);
      if (MO.isDef())
        RegPressure[RCId] += RCCost;
      else {
        bool isKill = isOperandKill(MO, MRI);
        if (isNew && !isKill)
          // Haven't seen this, it must be a livein.
          RegPressure[RCId] += RCCost;
        else if (!isNew && isKill)
          RegPressure[RCId] -= RCCost;
      }
    }
  }
}

/// UpdateRegPressure - Update estimate of register pressure after the
/// specified instruction.
void MachineLICM::UpdateRegPressure(const MachineInstr *MI) {
  if (MI->isImplicitDef())
    return;

  SmallVector<unsigned, 4> Defs;
  for (unsigned i = 0, e = MI->getDesc().getNumOperands(); i != e; ++i) {
    const MachineOperand &MO = MI->getOperand(i);
    if (!MO.isReg() || MO.isImplicit())
      continue;
    unsigned Reg = MO.getReg();
    if (!TargetRegisterInfo::isVirtualRegister(Reg))
      continue;

    bool isNew = RegSeen.insert(Reg);
    if (MO.isDef())
      Defs.push_back(Reg);
    else if (!isNew && isOperandKill(MO, MRI)) {
      unsigned RCId, RCCost;
      getRegisterClassIDAndCost(MI, Reg, i, RCId, RCCost);
      if (RCCost > RegPressure[RCId])
        RegPressure[RCId] = 0;
      else
        RegPressure[RCId] -= RCCost;
    }
  }

  unsigned Idx = 0;
  while (!Defs.empty()) {
    unsigned Reg = Defs.pop_back_val();
    unsigned RCId, RCCost;
    getRegisterClassIDAndCost(MI, Reg, Idx, RCId, RCCost);
    RegPressure[RCId] += RCCost;
    ++Idx;
  }
}

/// isLoadFromGOTOrConstantPool - Return true if this machine instruction
/// loads from global offset table or constant pool.
static bool isLoadFromGOTOrConstantPool(MachineInstr &MI) {
  assert (MI.mayLoad() && "Expected MI that loads!");
  for (MachineInstr::mmo_iterator I = MI.memoperands_begin(),
         E = MI.memoperands_end(); I != E; ++I) {
    if (const Value *V = (*I)->getValue()) {
      if (const PseudoSourceValue *PSV = dyn_cast<PseudoSourceValue>(V))
        if (PSV == PSV->getGOT() || PSV == PSV->getConstantPool())
          return true;
    }
  }
  return false;
}

/// IsLICMCandidate - Returns true if the instruction may be a suitable
/// candidate for LICM. e.g. If the instruction is a call, then it's obviously
/// not safe to hoist it.
bool MachineLICM::IsLICMCandidate(MachineInstr &I) {
  // Check if it's safe to move the instruction.
  bool DontMoveAcrossStore = true;
  if (!I.isSafeToMove(TII, AA, DontMoveAcrossStore))
    return false;

  // If it is load then check if it is guaranteed to execute by making sure that
  // it dominates all exiting blocks. If it doesn't, then there is a path out of
  // the loop which does not execute this load, so we can't hoist it. Loads
  // from constant memory are not safe to speculate all the time, for example
  // indexed load from a jump table.
  // Stores and side effects are already checked by isSafeToMove.
  if (I.mayLoad() && !isLoadFromGOTOrConstantPool(I) &&
      !IsGuaranteedToExecute(I.getParent()))
    return false;

  return true;
}

/// IsLoopInvariantInst - Returns true if the instruction is loop
/// invariant. I.e., all virtual register operands are defined outside of the
/// loop, physical registers aren't accessed explicitly, and there are no side
/// effects that aren't captured by the operands or other flags.
///
bool MachineLICM::IsLoopInvariantInst(MachineInstr &I) {
  if (!IsLICMCandidate(I))
    return false;

  // The instruction is loop invariant if all of its operands are.
  for (unsigned i = 0, e = I.getNumOperands(); i != e; ++i) {
    const MachineOperand &MO = I.getOperand(i);

    if (!MO.isReg())
      continue;

    unsigned Reg = MO.getReg();
    if (Reg == 0) continue;

    // Don't hoist an instruction that uses or defines a physical register.
    if (TargetRegisterInfo::isPhysicalRegister(Reg)) {
      if (MO.isUse()) {
        // If the physreg has no defs anywhere, it's just an ambient register
        // and we can freely move its uses. Alternatively, if it's allocatable,
        // it could get allocated to something with a def during allocation.
        if (!MRI->isConstantPhysReg(Reg, *I.getParent()->getParent()))
          return false;
        // Otherwise it's safe to move.
        continue;
      } else if (!MO.isDead()) {
        // A def that isn't dead. We can't move it.
        return false;
      } else if (CurLoop->getHeader()->isLiveIn(Reg)) {
        // If the reg is live into the loop, we can't hoist an instruction
        // which would clobber it.
        return false;
      }
    }

    if (!MO.isUse())
      continue;

    assert(MRI->getVRegDef(Reg) &&
           "Machine instr not mapped for this vreg?!");

    // If the loop contains the definition of an operand, then the instruction
    // isn't loop invariant.
    if (CurLoop->contains(MRI->getVRegDef(Reg)))
      return false;
  }

  // If we got this far, the instruction is loop invariant!
  return true;
}


/// HasLoopPHIUse - Return true if the specified instruction is used by a
/// phi node and hoisting it could cause a copy to be inserted.
bool MachineLICM::HasLoopPHIUse(const MachineInstr *MI) const {
  SmallVector<const MachineInstr*, 8> Work(1, MI);
  do {
    MI = Work.pop_back_val();
    for (ConstMIOperands MO(MI); MO.isValid(); ++MO) {
      if (!MO->isReg() || !MO->isDef())
        continue;
      unsigned Reg = MO->getReg();
      if (!TargetRegisterInfo::isVirtualRegister(Reg))
        continue;
      for (MachineInstr &UseMI : MRI->use_instructions(Reg)) {
        // A PHI may cause a copy to be inserted.
        if (UseMI.isPHI()) {
          // A PHI inside the loop causes a copy because the live range of Reg is
          // extended across the PHI.
          if (CurLoop->contains(&UseMI))
            return true;
          // A PHI in an exit block can cause a copy to be inserted if the PHI
          // has multiple predecessors in the loop with different values.
          // For now, approximate by rejecting all exit blocks.
          if (isExitBlock(UseMI.getParent()))
            return true;
          continue;
        }
        // Look past copies as well.
        if (UseMI.isCopy() && CurLoop->contains(&UseMI))
          Work.push_back(&UseMI);
      }
    }
  } while (!Work.empty());
  return false;
}

/// HasHighOperandLatency - Compute operand latency between a def of 'Reg'
/// and an use in the current loop, return true if the target considered
/// it 'high'.
bool MachineLICM::HasHighOperandLatency(MachineInstr &MI,
                                        unsigned DefIdx, unsigned Reg) const {
  if (!InstrItins || InstrItins->isEmpty() || MRI->use_nodbg_empty(Reg))
    return false;

  for (MachineInstr &UseMI : MRI->use_nodbg_instructions(Reg)) {
    if (UseMI.isCopyLike())
      continue;
    if (!CurLoop->contains(UseMI.getParent()))
      continue;
    for (unsigned i = 0, e = UseMI.getNumOperands(); i != e; ++i) {
      const MachineOperand &MO = UseMI.getOperand(i);
      if (!MO.isReg() || !MO.isUse())
        continue;
      unsigned MOReg = MO.getReg();
      if (MOReg != Reg)
        continue;

      if (TII->hasHighOperandLatency(InstrItins, MRI, &MI, DefIdx, &UseMI, i))
        return true;
    }

    // Only look at the first in loop use.
    break;
  }

  return false;
}

/// IsCheapInstruction - Return true if the instruction is marked "cheap" or
/// the operand latency between its def and a use is one or less.
bool MachineLICM::IsCheapInstruction(MachineInstr &MI) const {
  if (MI.isAsCheapAsAMove() || MI.isCopyLike())
    return true;
  if (!InstrItins || InstrItins->isEmpty())
    return false;

  bool isCheap = false;
  unsigned NumDefs = MI.getDesc().getNumDefs();
  for (unsigned i = 0, e = MI.getNumOperands(); NumDefs && i != e; ++i) {
    MachineOperand &DefMO = MI.getOperand(i);
    if (!DefMO.isReg() || !DefMO.isDef())
      continue;
    --NumDefs;
    unsigned Reg = DefMO.getReg();
    if (TargetRegisterInfo::isPhysicalRegister(Reg))
      continue;

    if (!TII->hasLowDefLatency(InstrItins, &MI, i))
      return false;
    isCheap = true;
  }

  return isCheap;
}

/// CanCauseHighRegPressure - Visit BBs from header to current BB, check
/// if hoisting an instruction of the given cost matrix can cause high
/// register pressure.
bool MachineLICM::CanCauseHighRegPressure(DenseMap<unsigned, int> &Cost,
                                          bool CheapInstr) {
  for (DenseMap<unsigned, int>::iterator CI = Cost.begin(), CE = Cost.end();
       CI != CE; ++CI) {
    if (CI->second <= 0)
      continue;

    unsigned RCId = CI->first;
    unsigned Limit = RegLimit[RCId];
    int Cost = CI->second;

    // Don't hoist cheap instructions if they would increase register pressure,
    // even if we're under the limit.
    if (CheapInstr)
      return true;

    for (unsigned i = BackTrace.size(); i != 0; --i) {
      SmallVectorImpl<unsigned> &RP = BackTrace[i-1];
      if (RP[RCId] + Cost >= Limit)
        return true;
    }
  }

  return false;
}

/// UpdateBackTraceRegPressure - Traverse the back trace from header to the
/// current block and update their register pressures to reflect the effect
/// of hoisting MI from the current block to the preheader.
void MachineLICM::UpdateBackTraceRegPressure(const MachineInstr *MI) {
  if (MI->isImplicitDef())
    return;

  // First compute the 'cost' of the instruction, i.e. its contribution
  // to register pressure.
  DenseMap<unsigned, int> Cost;
  for (unsigned i = 0, e = MI->getDesc().getNumOperands(); i != e; ++i) {
    const MachineOperand &MO = MI->getOperand(i);
    if (!MO.isReg() || MO.isImplicit())
      continue;
    unsigned Reg = MO.getReg();
    if (!TargetRegisterInfo::isVirtualRegister(Reg))
      continue;

    unsigned RCId, RCCost;
    getRegisterClassIDAndCost(MI, Reg, i, RCId, RCCost);
    if (MO.isDef()) {
      DenseMap<unsigned, int>::iterator CI = Cost.find(RCId);
      if (CI != Cost.end())
        CI->second += RCCost;
      else
        Cost.insert(std::make_pair(RCId, RCCost));
    } else if (isOperandKill(MO, MRI)) {
      DenseMap<unsigned, int>::iterator CI = Cost.find(RCId);
      if (CI != Cost.end())
        CI->second -= RCCost;
      else
        Cost.insert(std::make_pair(RCId, -RCCost));
    }
  }

  // Update register pressure of blocks from loop header to current block.
  for (unsigned i = 0, e = BackTrace.size(); i != e; ++i) {
    SmallVectorImpl<unsigned> &RP = BackTrace[i];
    for (DenseMap<unsigned, int>::iterator CI = Cost.begin(), CE = Cost.end();
         CI != CE; ++CI) {
      unsigned RCId = CI->first;
      RP[RCId] += CI->second;
    }
  }
}

/// IsProfitableToHoist - Return true if it is potentially profitable to hoist
/// the given loop invariant.
bool MachineLICM::IsProfitableToHoist(MachineInstr &MI) {
  if (MI.isImplicitDef())
    return true;

  // Besides removing computation from the loop, hoisting an instruction has
  // these effects:
  //
  // - The value defined by the instruction becomes live across the entire
  //   loop. This increases register pressure in the loop.
  //
  // - If the value is used by a PHI in the loop, a copy will be required for
  //   lowering the PHI after extending the live range.
  //
  // - When hoisting the last use of a value in the loop, that value no longer
  //   needs to be live in the loop. This lowers register pressure in the loop.

  bool CheapInstr = IsCheapInstruction(MI);
  bool CreatesCopy = HasLoopPHIUse(&MI);

  // Don't hoist a cheap instruction if it would create a copy in the loop.
  if (CheapInstr && CreatesCopy) {
    DEBUG(dbgs() << "Won't hoist cheap instr with loop PHI use: " << MI);
    return false;
  }

  // Rematerializable instructions should always be hoisted since the register
  // allocator can just pull them down again when needed.
  if (TII->isTriviallyReMaterializable(&MI, AA))
    return true;

  // Estimate register pressure to determine whether to LICM the instruction.
  // In low register pressure situation, we can be more aggressive about
  // hoisting. Also, favors hoisting long latency instructions even in
  // moderately high pressure situation.
  // Cheap instructions will only be hoisted if they don't increase register
  // pressure at all.
  // FIXME: If there are long latency loop-invariant instructions inside the
  // loop at this point, why didn't the optimizer's LICM hoist them?
  DenseMap<unsigned, int> Cost;
  for (unsigned i = 0, e = MI.getDesc().getNumOperands(); i != e; ++i) {
    const MachineOperand &MO = MI.getOperand(i);
    if (!MO.isReg() || MO.isImplicit())
      continue;
    unsigned Reg = MO.getReg();
    if (!TargetRegisterInfo::isVirtualRegister(Reg))
      continue;

    unsigned RCId, RCCost;
    getRegisterClassIDAndCost(&MI, Reg, i, RCId, RCCost);
    if (MO.isDef()) {
      if (HasHighOperandLatency(MI, i, Reg)) {
        DEBUG(dbgs() << "Hoist High Latency: " << MI);
        ++NumHighLatency;
        return true;
      }
      Cost[RCId] += RCCost;
    } else if (isOperandKill(MO, MRI)) {
      // Is a virtual register use is a kill, hoisting it out of the loop
      // may actually reduce register pressure or be register pressure
      // neutral.
      Cost[RCId] -= RCCost;
    }
  }

  // Visit BBs from header to current BB, if hoisting this doesn't cause
  // high register pressure, then it's safe to proceed.
  if (!CanCauseHighRegPressure(Cost, CheapInstr)) {
    DEBUG(dbgs() << "Hoist non-reg-pressure: " << MI);
    ++NumLowRP;
    return true;
  }

  // Don't risk increasing register pressure if it would create copies.
  if (CreatesCopy) {
    DEBUG(dbgs() << "Won't hoist instr with loop PHI use: " << MI);
    return false;
  }

  // Do not "speculate" in high register pressure situation. If an
  // instruction is not guaranteed to be executed in the loop, it's best to be
  // conservative.
  if (AvoidSpeculation &&
      (!IsGuaranteedToExecute(MI.getParent()) && !MayCSE(&MI))) {
    DEBUG(dbgs() << "Won't speculate: " << MI);
    return false;
  }

  // High register pressure situation, only hoist if the instruction is going
  // to be remat'ed.
  if (!TII->isTriviallyReMaterializable(&MI, AA) &&
      !MI.isInvariantLoad(AA)) {
    DEBUG(dbgs() << "Can't remat / high reg-pressure: " << MI);
    return false;
  }

  return true;
}

MachineInstr *MachineLICM::ExtractHoistableLoad(MachineInstr *MI) {
  // Don't unfold simple loads.
  if (MI->canFoldAsLoad())
    return 0;

  // If not, we may be able to unfold a load and hoist that.
  // First test whether the instruction is loading from an amenable
  // memory location.
  if (!MI->isInvariantLoad(AA))
    return 0;

  // Next determine the register class for a temporary register.
  unsigned LoadRegIndex;
  unsigned NewOpc =
    TII->getOpcodeAfterMemoryUnfold(MI->getOpcode(),
                                    /*UnfoldLoad=*/true,
                                    /*UnfoldStore=*/false,
                                    &LoadRegIndex);
  if (NewOpc == 0) return 0;
  const MCInstrDesc &MID = TII->get(NewOpc);
  if (MID.getNumDefs() != 1) return 0;
  MachineFunction &MF = *MI->getParent()->getParent();
  const TargetRegisterClass *RC = TII->getRegClass(MID, LoadRegIndex, TRI, MF);
  // Ok, we're unfolding. Create a temporary register and do the unfold.
  unsigned Reg = MRI->createVirtualRegister(RC);

  SmallVector<MachineInstr *, 2> NewMIs;
  bool Success =
    TII->unfoldMemoryOperand(MF, MI, Reg,
                             /*UnfoldLoad=*/true, /*UnfoldStore=*/false,
                             NewMIs);
  (void)Success;
  assert(Success &&
         "unfoldMemoryOperand failed when getOpcodeAfterMemoryUnfold "
         "succeeded!");
  assert(NewMIs.size() == 2 &&
         "Unfolded a load into multiple instructions!");
  MachineBasicBlock *MBB = MI->getParent();
  MachineBasicBlock::iterator Pos = MI;
  MBB->insert(Pos, NewMIs[0]);
  MBB->insert(Pos, NewMIs[1]);
  // If unfolding produced a load that wasn't loop-invariant or profitable to
  // hoist, discard the new instructions and bail.
  if (!IsLoopInvariantInst(*NewMIs[0]) || !IsProfitableToHoist(*NewMIs[0])) {
    NewMIs[0]->eraseFromParent();
    NewMIs[1]->eraseFromParent();
    return 0;
  }

  // Update register pressure for the unfolded instruction.
  UpdateRegPressure(NewMIs[1]);

  // Otherwise we successfully unfolded a load that we can hoist.
  MI->eraseFromParent();
  return NewMIs[0];
}

void MachineLICM::InitCSEMap(MachineBasicBlock *BB) {
  for (MachineBasicBlock::iterator I = BB->begin(),E = BB->end(); I != E; ++I) {
    const MachineInstr *MI = &*I;
    unsigned Opcode = MI->getOpcode();
    DenseMap<unsigned, std::vector<const MachineInstr*> >::iterator
      CI = CSEMap.find(Opcode);
    if (CI != CSEMap.end())
      CI->second.push_back(MI);
    else {
      std::vector<const MachineInstr*> CSEMIs;
      CSEMIs.push_back(MI);
      CSEMap.insert(std::make_pair(Opcode, CSEMIs));
    }
  }
}

const MachineInstr*
MachineLICM::LookForDuplicate(const MachineInstr *MI,
                              std::vector<const MachineInstr*> &PrevMIs) {
  for (unsigned i = 0, e = PrevMIs.size(); i != e; ++i) {
    const MachineInstr *PrevMI = PrevMIs[i];
    if (TII->produceSameValue(MI, PrevMI, (PreRegAlloc ? MRI : 0)))
      return PrevMI;
  }
  return 0;
}

bool MachineLICM::EliminateCSE(MachineInstr *MI,
          DenseMap<unsigned, std::vector<const MachineInstr*> >::iterator &CI) {
  // Do not CSE implicit_def so ProcessImplicitDefs can properly propagate
  // the undef property onto uses.
  if (CI == CSEMap.end() || MI->isImplicitDef())
    return false;

  if (const MachineInstr *Dup = LookForDuplicate(MI, CI->second)) {
    DEBUG(dbgs() << "CSEing " << *MI << " with " << *Dup);

    // Replace virtual registers defined by MI by their counterparts defined
    // by Dup.
    SmallVector<unsigned, 2> Defs;
    for (unsigned i = 0, e = MI->getNumOperands(); i != e; ++i) {
      const MachineOperand &MO = MI->getOperand(i);

      // Physical registers may not differ here.
      assert((!MO.isReg() || MO.getReg() == 0 ||
              !TargetRegisterInfo::isPhysicalRegister(MO.getReg()) ||
              MO.getReg() == Dup->getOperand(i).getReg()) &&
             "Instructions with different phys regs are not identical!");

      if (MO.isReg() && MO.isDef() &&
          !TargetRegisterInfo::isPhysicalRegister(MO.getReg()))
        Defs.push_back(i);
    }

    SmallVector<const TargetRegisterClass*, 2> OrigRCs;
    for (unsigned i = 0, e = Defs.size(); i != e; ++i) {
      unsigned Idx = Defs[i];
      unsigned Reg = MI->getOperand(Idx).getReg();
      unsigned DupReg = Dup->getOperand(Idx).getReg();
      OrigRCs.push_back(MRI->getRegClass(DupReg));

      if (!MRI->constrainRegClass(DupReg, MRI->getRegClass(Reg))) {
        // Restore old RCs if more than one defs.
        for (unsigned j = 0; j != i; ++j)
          MRI->setRegClass(Dup->getOperand(Defs[j]).getReg(), OrigRCs[j]);
        return false;
      }
    }

    for (unsigned i = 0, e = Defs.size(); i != e; ++i) {
      unsigned Idx = Defs[i];
      unsigned Reg = MI->getOperand(Idx).getReg();
      unsigned DupReg = Dup->getOperand(Idx).getReg();
      MRI->replaceRegWith(Reg, DupReg);
      MRI->clearKillFlags(DupReg);
    }

    MI->eraseFromParent();
    ++NumCSEed;
    return true;
  }
  return false;
}

/// MayCSE - Return true if the given instruction will be CSE'd if it's
/// hoisted out of the loop.
bool MachineLICM::MayCSE(MachineInstr *MI) {
  unsigned Opcode = MI->getOpcode();
  DenseMap<unsigned, std::vector<const MachineInstr*> >::iterator
    CI = CSEMap.find(Opcode);
  // Do not CSE implicit_def so ProcessImplicitDefs can properly propagate
  // the undef property onto uses.
  if (CI == CSEMap.end() || MI->isImplicitDef())
    return false;

  return LookForDuplicate(MI, CI->second) != 0;
}

/// Hoist - When an instruction is found to use only loop invariant operands
/// that are safe to hoist, this instruction is called to do the dirty work.
///
bool MachineLICM::Hoist(MachineInstr *MI, MachineBasicBlock *Preheader) {
  // First check whether we should hoist this instruction.
  if (!IsLoopInvariantInst(*MI) || !IsProfitableToHoist(*MI)) {
    // If not, try unfolding a hoistable load.
    MI = ExtractHoistableLoad(MI);
    if (!MI) return false;
  }

  // Now move the instructions to the predecessor, inserting it before any
  // terminator instructions.
  DEBUG({
      dbgs() << "Hoisting " << *MI;
      if (Preheader->getBasicBlock())
        dbgs() << " to MachineBasicBlock "
               << Preheader->getName();
      if (MI->getParent()->getBasicBlock())
        dbgs() << " from MachineBasicBlock "
               << MI->getParent()->getName();
      dbgs() << "\n";
    });

  // If this is the first instruction being hoisted to the preheader,
  // initialize the CSE map with potential common expressions.
  if (FirstInLoop) {
    InitCSEMap(Preheader);
    FirstInLoop = false;
  }

  // Look for opportunity to CSE the hoisted instruction.
  unsigned Opcode = MI->getOpcode();
  DenseMap<unsigned, std::vector<const MachineInstr*> >::iterator
    CI = CSEMap.find(Opcode);
  if (!EliminateCSE(MI, CI)) {
    // Otherwise, splice the instruction to the preheader.
    Preheader->splice(Preheader->getFirstTerminator(),MI->getParent(),MI);

    // Update register pressure for BBs from header to this block.
    UpdateBackTraceRegPressure(MI);

    // Clear the kill flags of any register this instruction defines,
    // since they may need to be live throughout the entire loop
    // rather than just live for part of it.
    for (unsigned i = 0, e = MI->getNumOperands(); i != e; ++i) {
      MachineOperand &MO = MI->getOperand(i);
      if (MO.isReg() && MO.isDef() && !MO.isDead())
        MRI->clearKillFlags(MO.getReg());
    }

    // Add to the CSE map.
    if (CI != CSEMap.end())
      CI->second.push_back(MI);
    else {
      std::vector<const MachineInstr*> CSEMIs;
      CSEMIs.push_back(MI);
      CSEMap.insert(std::make_pair(Opcode, CSEMIs));
    }
  }

  ++NumHoisted;
  Changed = true;

  return true;
}

MachineBasicBlock *MachineLICM::getCurPreheader() {
  // Determine the block to which to hoist instructions. If we can't find a
  // suitable loop predecessor, we can't do any hoisting.

  // If we've tried to get a preheader and failed, don't try again.
  if (CurPreheader == reinterpret_cast<MachineBasicBlock *>(-1))
    return 0;

  if (!CurPreheader) {
    CurPreheader = CurLoop->getLoopPreheader();
    if (!CurPreheader) {
      MachineBasicBlock *Pred = CurLoop->getLoopPredecessor();
      if (!Pred) {
        CurPreheader = reinterpret_cast<MachineBasicBlock *>(-1);
        return 0;
      }

      CurPreheader = Pred->SplitCriticalEdge(CurLoop->getHeader(), this);
      if (!CurPreheader) {
        CurPreheader = reinterpret_cast<MachineBasicBlock *>(-1);
        return 0;
      }
    }
  }
  return CurPreheader;
}