llvm.org GIT mirror llvm / 8b170f7 lib / Transforms / Scalar / MemCpyOptimizer.cpp
8b170f7

Tree @8b170f7 (Download .tar.gz)

MemCpyOptimizer.cpp @8b170f7raw · 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
//===- MemCpyOptimizer.cpp - Optimize use of memcpy and friends -----------===//
//
//                     The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This pass performs various transformations related to eliminating memcpy
// calls, or transforming sets of stores into memset's.
//
//===----------------------------------------------------------------------===//

#include "llvm/Transforms/Scalar.h"
#include "llvm/ADT/SmallVector.h"
#include "llvm/ADT/Statistic.h"
#include "llvm/Analysis/AliasAnalysis.h"
#include "llvm/Analysis/AssumptionCache.h"
#include "llvm/Analysis/GlobalsModRef.h"
#include "llvm/Analysis/MemoryDependenceAnalysis.h"
#include "llvm/Analysis/TargetLibraryInfo.h"
#include "llvm/Analysis/ValueTracking.h"
#include "llvm/IR/DataLayout.h"
#include "llvm/IR/Dominators.h"
#include "llvm/IR/GetElementPtrTypeIterator.h"
#include "llvm/IR/GlobalVariable.h"
#include "llvm/IR/IRBuilder.h"
#include "llvm/IR/Instructions.h"
#include "llvm/IR/IntrinsicInst.h"
#include "llvm/Support/Debug.h"
#include "llvm/Support/raw_ostream.h"
#include "llvm/Transforms/Utils/Local.h"
#include <algorithm>
using namespace llvm;

#define DEBUG_TYPE "memcpyopt"

STATISTIC(NumMemCpyInstr, "Number of memcpy instructions deleted");
STATISTIC(NumMemSetInfer, "Number of memsets inferred");
STATISTIC(NumMoveToCpy,   "Number of memmoves converted to memcpy");
STATISTIC(NumCpyToSet,    "Number of memcpys converted to memset");

static int64_t GetOffsetFromIndex(const GEPOperator *GEP, unsigned Idx,
                                  bool &VariableIdxFound,
                                  const DataLayout &DL) {
  // Skip over the first indices.
  gep_type_iterator GTI = gep_type_begin(GEP);
  for (unsigned i = 1; i != Idx; ++i, ++GTI)
    /*skip along*/;

  // Compute the offset implied by the rest of the indices.
  int64_t Offset = 0;
  for (unsigned i = Idx, e = GEP->getNumOperands(); i != e; ++i, ++GTI) {
    ConstantInt *OpC = dyn_cast<ConstantInt>(GEP->getOperand(i));
    if (!OpC)
      return VariableIdxFound = true;
    if (OpC->isZero()) continue;  // No offset.

    // Handle struct indices, which add their field offset to the pointer.
    if (StructType *STy = dyn_cast<StructType>(*GTI)) {
      Offset += DL.getStructLayout(STy)->getElementOffset(OpC->getZExtValue());
      continue;
    }

    // Otherwise, we have a sequential type like an array or vector.  Multiply
    // the index by the ElementSize.
    uint64_t Size = DL.getTypeAllocSize(GTI.getIndexedType());
    Offset += Size*OpC->getSExtValue();
  }

  return Offset;
}

/// Return true if Ptr1 is provably equal to Ptr2 plus a constant offset, and
/// return that constant offset. For example, Ptr1 might be &A[42], and Ptr2
/// might be &A[40]. In this case offset would be -8.
static bool IsPointerOffset(Value *Ptr1, Value *Ptr2, int64_t &Offset,
                            const DataLayout &DL) {
  Ptr1 = Ptr1->stripPointerCasts();
  Ptr2 = Ptr2->stripPointerCasts();

  // Handle the trivial case first.
  if (Ptr1 == Ptr2) {
    Offset = 0;
    return true;
  }

  GEPOperator *GEP1 = dyn_cast<GEPOperator>(Ptr1);
  GEPOperator *GEP2 = dyn_cast<GEPOperator>(Ptr2);

  bool VariableIdxFound = false;

  // If one pointer is a GEP and the other isn't, then see if the GEP is a
  // constant offset from the base, as in "P" and "gep P, 1".
  if (GEP1 && !GEP2 && GEP1->getOperand(0)->stripPointerCasts() == Ptr2) {
    Offset = -GetOffsetFromIndex(GEP1, 1, VariableIdxFound, DL);
    return !VariableIdxFound;
  }

  if (GEP2 && !GEP1 && GEP2->getOperand(0)->stripPointerCasts() == Ptr1) {
    Offset = GetOffsetFromIndex(GEP2, 1, VariableIdxFound, DL);
    return !VariableIdxFound;
  }

  // Right now we handle the case when Ptr1/Ptr2 are both GEPs with an identical
  // base.  After that base, they may have some number of common (and
  // potentially variable) indices.  After that they handle some constant
  // offset, which determines their offset from each other.  At this point, we
  // handle no other case.
  if (!GEP1 || !GEP2 || GEP1->getOperand(0) != GEP2->getOperand(0))
    return false;

  // Skip any common indices and track the GEP types.
  unsigned Idx = 1;
  for (; Idx != GEP1->getNumOperands() && Idx != GEP2->getNumOperands(); ++Idx)
    if (GEP1->getOperand(Idx) != GEP2->getOperand(Idx))
      break;

  int64_t Offset1 = GetOffsetFromIndex(GEP1, Idx, VariableIdxFound, DL);
  int64_t Offset2 = GetOffsetFromIndex(GEP2, Idx, VariableIdxFound, DL);
  if (VariableIdxFound) return false;

  Offset = Offset2-Offset1;
  return true;
}


/// Represents a range of memset'd bytes with the ByteVal value.
/// This allows us to analyze stores like:
///   store 0 -> P+1
///   store 0 -> P+0
///   store 0 -> P+3
///   store 0 -> P+2
/// which sometimes happens with stores to arrays of structs etc.  When we see
/// the first store, we make a range [1, 2).  The second store extends the range
/// to [0, 2).  The third makes a new range [2, 3).  The fourth store joins the
/// two ranges into [0, 3) which is memset'able.
namespace {
struct MemsetRange {
  // Start/End - A semi range that describes the span that this range covers.
  // The range is closed at the start and open at the end: [Start, End).
  int64_t Start, End;

  /// StartPtr - The getelementptr instruction that points to the start of the
  /// range.
  Value *StartPtr;

  /// Alignment - The known alignment of the first store.
  unsigned Alignment;

  /// TheStores - The actual stores that make up this range.
  SmallVector<Instruction*, 16> TheStores;

  bool isProfitableToUseMemset(const DataLayout &DL) const;
};
} // end anon namespace

bool MemsetRange::isProfitableToUseMemset(const DataLayout &DL) const {
  // If we found more than 4 stores to merge or 16 bytes, use memset.
  if (TheStores.size() >= 4 || End-Start >= 16) return true;

  // If there is nothing to merge, don't do anything.
  if (TheStores.size() < 2) return false;

  // If any of the stores are a memset, then it is always good to extend the
  // memset.
  for (unsigned i = 0, e = TheStores.size(); i != e; ++i)
    if (!isa<StoreInst>(TheStores[i]))
      return true;

  // Assume that the code generator is capable of merging pairs of stores
  // together if it wants to.
  if (TheStores.size() == 2) return false;

  // If we have fewer than 8 stores, it can still be worthwhile to do this.
  // For example, merging 4 i8 stores into an i32 store is useful almost always.
  // However, merging 2 32-bit stores isn't useful on a 32-bit architecture (the
  // memset will be split into 2 32-bit stores anyway) and doing so can
  // pessimize the llvm optimizer.
  //
  // Since we don't have perfect knowledge here, make some assumptions: assume
  // the maximum GPR width is the same size as the largest legal integer
  // size. If so, check to see whether we will end up actually reducing the
  // number of stores used.
  unsigned Bytes = unsigned(End-Start);
  unsigned MaxIntSize = DL.getLargestLegalIntTypeSize();
  if (MaxIntSize == 0)
    MaxIntSize = 1;
  unsigned NumPointerStores = Bytes / MaxIntSize;

  // Assume the remaining bytes if any are done a byte at a time.
  unsigned NumByteStores = Bytes - NumPointerStores * MaxIntSize;

  // If we will reduce the # stores (according to this heuristic), do the
  // transformation.  This encourages merging 4 x i8 -> i32 and 2 x i16 -> i32
  // etc.
  return TheStores.size() > NumPointerStores+NumByteStores;
}


namespace {
class MemsetRanges {
  /// A sorted list of the memset ranges.
  SmallVector<MemsetRange, 8> Ranges;
  typedef SmallVectorImpl<MemsetRange>::iterator range_iterator;
  const DataLayout &DL;
public:
  MemsetRanges(const DataLayout &DL) : DL(DL) {}

  typedef SmallVectorImpl<MemsetRange>::const_iterator const_iterator;
  const_iterator begin() const { return Ranges.begin(); }
  const_iterator end() const { return Ranges.end(); }
  bool empty() const { return Ranges.empty(); }

  void addInst(int64_t OffsetFromFirst, Instruction *Inst) {
    if (StoreInst *SI = dyn_cast<StoreInst>(Inst))
      addStore(OffsetFromFirst, SI);
    else
      addMemSet(OffsetFromFirst, cast<MemSetInst>(Inst));
  }

  void addStore(int64_t OffsetFromFirst, StoreInst *SI) {
    int64_t StoreSize = DL.getTypeStoreSize(SI->getOperand(0)->getType());

    addRange(OffsetFromFirst, StoreSize,
             SI->getPointerOperand(), SI->getAlignment(), SI);
  }

  void addMemSet(int64_t OffsetFromFirst, MemSetInst *MSI) {
    int64_t Size = cast<ConstantInt>(MSI->getLength())->getZExtValue();
    addRange(OffsetFromFirst, Size, MSI->getDest(), MSI->getDestAlignment(),
             MSI);
  }

  void addRange(int64_t Start, int64_t Size, Value *Ptr,
                unsigned Alignment, Instruction *Inst);

};

} // end anon namespace


/// Add a new store to the MemsetRanges data structure.  This adds a
/// new range for the specified store at the specified offset, merging into
/// existing ranges as appropriate.
void MemsetRanges::addRange(int64_t Start, int64_t Size, Value *Ptr,
                            unsigned Alignment, Instruction *Inst) {
  int64_t End = Start+Size;

  range_iterator I = std::lower_bound(Ranges.begin(), Ranges.end(), Start,
    [](const MemsetRange &LHS, int64_t RHS) { return LHS.End < RHS; });

  // We now know that I == E, in which case we didn't find anything to merge
  // with, or that Start <= I->End.  If End < I->Start or I == E, then we need
  // to insert a new range.  Handle this now.
  if (I == Ranges.end() || End < I->Start) {
    MemsetRange &R = *Ranges.insert(I, MemsetRange());
    R.Start        = Start;
    R.End          = End;
    R.StartPtr     = Ptr;
    R.Alignment    = Alignment;
    R.TheStores.push_back(Inst);
    return;
  }

  // This store overlaps with I, add it.
  I->TheStores.push_back(Inst);

  // At this point, we may have an interval that completely contains our store.
  // If so, just add it to the interval and return.
  if (I->Start <= Start && I->End >= End)
    return;

  // Now we know that Start <= I->End and End >= I->Start so the range overlaps
  // but is not entirely contained within the range.

  // See if the range extends the start of the range.  In this case, it couldn't
  // possibly cause it to join the prior range, because otherwise we would have
  // stopped on *it*.
  if (Start < I->Start) {
    I->Start = Start;
    I->StartPtr = Ptr;
    I->Alignment = Alignment;
  }

  // Now we know that Start <= I->End and Start >= I->Start (so the startpoint
  // is in or right at the end of I), and that End >= I->Start.  Extend I out to
  // End.
  if (End > I->End) {
    I->End = End;
    range_iterator NextI = I;
    while (++NextI != Ranges.end() && End >= NextI->Start) {
      // Merge the range in.
      I->TheStores.append(NextI->TheStores.begin(), NextI->TheStores.end());
      if (NextI->End > I->End)
        I->End = NextI->End;
      Ranges.erase(NextI);
      NextI = I;
    }
  }
}

//===----------------------------------------------------------------------===//
//                         MemCpyOpt Pass
//===----------------------------------------------------------------------===//

namespace {
  class MemCpyOpt : public FunctionPass {
    MemoryDependenceAnalysis *MD;
    TargetLibraryInfo *TLI;
  public:
    static char ID; // Pass identification, replacement for typeid
    MemCpyOpt() : FunctionPass(ID) {
      initializeMemCpyOptPass(*PassRegistry::getPassRegistry());
      MD = nullptr;
      TLI = nullptr;
    }

    bool runOnFunction(Function &F) override;

  private:
    // This transformation requires dominator postdominator info
    void getAnalysisUsage(AnalysisUsage &AU) const override {
      AU.setPreservesCFG();
      AU.addRequired<AssumptionCacheTracker>();
      AU.addRequired<DominatorTreeWrapperPass>();
      AU.addRequired<MemoryDependenceAnalysis>();
      AU.addRequired<AAResultsWrapperPass>();
      AU.addRequired<TargetLibraryInfoWrapperPass>();
      AU.addPreserved<GlobalsAAWrapperPass>();
      AU.addPreserved<MemoryDependenceAnalysis>();
    }

    // Helper functions
    bool processStore(StoreInst *SI, BasicBlock::iterator &BBI);
    bool processMemSet(MemSetInst *SI, BasicBlock::iterator &BBI);
    bool processMemCpy(MemCpyInst *M);
    bool processMemMove(MemMoveInst *M);
    bool performCallSlotOptzn(Instruction *cpy, Value *cpyDst, Value *cpySrc,
                              uint64_t cpyLen, unsigned cpyAlign, CallInst *C);
    bool processMemCpyMemCpyDependence(MemCpyInst *M, MemCpyInst *MDep);
    bool processMemSetMemCpyDependence(MemCpyInst *M, MemSetInst *MDep);
    bool performMemCpyToMemSetOptzn(MemCpyInst *M, MemSetInst *MDep);
    bool processByValArgument(CallSite CS, unsigned ArgNo);
    Instruction *tryMergingIntoMemset(Instruction *I, Value *StartPtr,
                                      Value *ByteVal);

    bool iterateOnFunction(Function &F);
  };

  char MemCpyOpt::ID = 0;
}

/// The public interface to this file...
FunctionPass *llvm::createMemCpyOptPass() { return new MemCpyOpt(); }

INITIALIZE_PASS_BEGIN(MemCpyOpt, "memcpyopt", "MemCpy Optimization",
                      false, false)
INITIALIZE_PASS_DEPENDENCY(AssumptionCacheTracker)
INITIALIZE_PASS_DEPENDENCY(DominatorTreeWrapperPass)
INITIALIZE_PASS_DEPENDENCY(MemoryDependenceAnalysis)
INITIALIZE_PASS_DEPENDENCY(TargetLibraryInfoWrapperPass)
INITIALIZE_PASS_DEPENDENCY(AAResultsWrapperPass)
INITIALIZE_PASS_DEPENDENCY(GlobalsAAWrapperPass)
INITIALIZE_PASS_END(MemCpyOpt, "memcpyopt", "MemCpy Optimization",
                    false, false)

/// When scanning forward over instructions, we look for some other patterns to
/// fold away. In particular, this looks for stores to neighboring locations of
/// memory. If it sees enough consecutive ones, it attempts to merge them
/// together into a memcpy/memset.
Instruction *MemCpyOpt::tryMergingIntoMemset(Instruction *StartInst,
                                             Value *StartPtr, Value *ByteVal) {
  const DataLayout &DL = StartInst->getModule()->getDataLayout();

  // Okay, so we now have a single store that can be splatable.  Scan to find
  // all subsequent stores of the same value to offset from the same pointer.
  // Join these together into ranges, so we can decide whether contiguous blocks
  // are stored.
  MemsetRanges Ranges(DL);

  BasicBlock::iterator BI(StartInst);
  for (++BI; !isa<TerminatorInst>(BI); ++BI) {
    if (!isa<StoreInst>(BI) && !isa<MemSetInst>(BI)) {
      // If the instruction is readnone, ignore it, otherwise bail out.  We
      // don't even allow readonly here because we don't want something like:
      // A[1] = 2; strlen(A); A[2] = 2; -> memcpy(A, ...); strlen(A).
      if (BI->mayWriteToMemory() || BI->mayReadFromMemory())
        break;
      continue;
    }

    if (StoreInst *NextStore = dyn_cast<StoreInst>(BI)) {
      // If this is a store, see if we can merge it in.
      if (!NextStore->isSimple()) break;

      // Check to see if this stored value is of the same byte-splattable value.
      if (ByteVal != isBytewiseValue(NextStore->getOperand(0)))
        break;

      // Check to see if this store is to a constant offset from the start ptr.
      int64_t Offset;
      if (!IsPointerOffset(StartPtr, NextStore->getPointerOperand(), Offset,
                           DL))
        break;

      Ranges.addStore(Offset, NextStore);
    } else {
      MemSetInst *MSI = cast<MemSetInst>(BI);

      if (MSI->isVolatile() || ByteVal != MSI->getValue() ||
          !isa<ConstantInt>(MSI->getLength()))
        break;

      // Check to see if this store is to a constant offset from the start ptr.
      int64_t Offset;
      if (!IsPointerOffset(StartPtr, MSI->getDest(), Offset, DL))
        break;

      Ranges.addMemSet(Offset, MSI);
    }
  }

  // If we have no ranges, then we just had a single store with nothing that
  // could be merged in.  This is a very common case of course.
  if (Ranges.empty())
    return nullptr;

  // If we had at least one store that could be merged in, add the starting
  // store as well.  We try to avoid this unless there is at least something
  // interesting as a small compile-time optimization.
  Ranges.addInst(0, StartInst);

  // If we create any memsets, we put it right before the first instruction that
  // isn't part of the memset block.  This ensure that the memset is dominated
  // by any addressing instruction needed by the start of the block.
  IRBuilder<> Builder(&*BI);

  // Now that we have full information about ranges, loop over the ranges and
  // emit memset's for anything big enough to be worthwhile.
  Instruction *AMemSet = nullptr;
  for (MemsetRanges::const_iterator I = Ranges.begin(), E = Ranges.end();
       I != E; ++I) {
    const MemsetRange &Range = *I;

    if (Range.TheStores.size() == 1) continue;

    // If it is profitable to lower this range to memset, do so now.
    if (!Range.isProfitableToUseMemset(DL))
      continue;

    // Otherwise, we do want to transform this!  Create a new memset.
    // Get the starting pointer of the block.
    StartPtr = Range.StartPtr;

    // Determine alignment
    unsigned Alignment = Range.Alignment;
    if (Alignment == 0) {
      Type *EltType =
        cast<PointerType>(StartPtr->getType())->getElementType();
      Alignment = DL.getABITypeAlignment(EltType);
    }

    AMemSet =
      Builder.CreateMemSet(StartPtr, ByteVal, Range.End-Range.Start, Alignment);

    DEBUG(dbgs() << "Replace stores:\n";
          for (unsigned i = 0, e = Range.TheStores.size(); i != e; ++i)
            dbgs() << *Range.TheStores[i] << '\n';
          dbgs() << "With: " << *AMemSet << '\n');

    if (!Range.TheStores.empty())
      AMemSet->setDebugLoc(Range.TheStores[0]->getDebugLoc());

    // Zap all the stores.
    for (SmallVectorImpl<Instruction *>::const_iterator
         SI = Range.TheStores.begin(),
         SE = Range.TheStores.end(); SI != SE; ++SI) {
      MD->removeInstruction(*SI);
      (*SI)->eraseFromParent();
    }
    ++NumMemSetInfer;
  }

  return AMemSet;
}


bool MemCpyOpt::processStore(StoreInst *SI, BasicBlock::iterator &BBI) {
  if (!SI->isSimple()) return false;

  // Avoid merging nontemporal stores since the resulting
  // memcpy/memset would not be able to preserve the nontemporal hint.
  // In theory we could teach how to propagate the !nontemporal metadata to
  // memset calls. However, that change would force the backend to
  // conservatively expand !nontemporal memset calls back to sequences of
  // store instructions (effectively undoing the merging).
  if (SI->getMetadata(LLVMContext::MD_nontemporal))
    return false;

  const DataLayout &DL = SI->getModule()->getDataLayout();

  // Detect cases where we're performing call slot forwarding, but
  // happen to be using a load-store pair to implement it, rather than
  // a memcpy.
  if (LoadInst *LI = dyn_cast<LoadInst>(SI->getOperand(0))) {
    if (LI->isSimple() && LI->hasOneUse() &&
        LI->getParent() == SI->getParent()) {
      MemDepResult ldep = MD->getDependency(LI);
      CallInst *C = nullptr;
      if (ldep.isClobber() && !isa<MemCpyInst>(ldep.getInst()))
        C = dyn_cast<CallInst>(ldep.getInst());

      if (C) {
        // Check that nothing touches the dest of the "copy" between
        // the call and the store.
        AliasAnalysis &AA = getAnalysis<AAResultsWrapperPass>().getAAResults();
        MemoryLocation StoreLoc = MemoryLocation::get(SI);
        for (BasicBlock::iterator I = --SI->getIterator(), E = C->getIterator();
             I != E; --I) {
          if (AA.getModRefInfo(&*I, StoreLoc) != MRI_NoModRef) {
            C = nullptr;
            break;
          }
        }
      }

      if (C) {
        unsigned storeAlign = SI->getAlignment();
        if (!storeAlign)
          storeAlign = DL.getABITypeAlignment(SI->getOperand(0)->getType());
        unsigned loadAlign = LI->getAlignment();
        if (!loadAlign)
          loadAlign = DL.getABITypeAlignment(LI->getType());

        bool changed = performCallSlotOptzn(
            LI, SI->getPointerOperand()->stripPointerCasts(),
            LI->getPointerOperand()->stripPointerCasts(),
            DL.getTypeStoreSize(SI->getOperand(0)->getType()),
            std::min(storeAlign, loadAlign), C);
        if (changed) {
          MD->removeInstruction(SI);
          SI->eraseFromParent();
          MD->removeInstruction(LI);
          LI->eraseFromParent();
          ++NumMemCpyInstr;
          return true;
        }
      }
    }
  }

  // There are two cases that are interesting for this code to handle: memcpy
  // and memset.  Right now we only handle memset.

  // Ensure that the value being stored is something that can be memset'able a
  // byte at a time like "0" or "-1" or any width, as well as things like
  // 0xA0A0A0A0 and 0.0.
  if (Value *ByteVal = isBytewiseValue(SI->getOperand(0)))
    if (Instruction *I = tryMergingIntoMemset(SI, SI->getPointerOperand(),
                                              ByteVal)) {
      BBI = I->getIterator(); // Don't invalidate iterator.
      return true;
    }

  return false;
}

bool MemCpyOpt::processMemSet(MemSetInst *MSI, BasicBlock::iterator &BBI) {
  // See if there is another memset or store neighboring this memset which
  // allows us to widen out the memset to do a single larger store.
  if (isa<ConstantInt>(MSI->getLength()) && !MSI->isVolatile())
    if (Instruction *I = tryMergingIntoMemset(MSI, MSI->getDest(),
                                              MSI->getValue())) {
      BBI = I->getIterator(); // Don't invalidate iterator.
      return true;
    }
  return false;
}


/// Takes a memcpy and a call that it depends on,
/// and checks for the possibility of a call slot optimization by having
/// the call write its result directly into the destination of the memcpy.
bool MemCpyOpt::performCallSlotOptzn(Instruction *cpy,
                                     Value *cpyDest, Value *cpySrc,
                                     uint64_t cpyLen, unsigned cpyAlign,
                                     CallInst *C) {
  // The general transformation to keep in mind is
  //
  //   call @func(..., src, ...)
  //   memcpy(dest, src, ...)
  //
  // ->
  //
  //   memcpy(dest, src, ...)
  //   call @func(..., dest, ...)
  //
  // Since moving the memcpy is technically awkward, we additionally check that
  // src only holds uninitialized values at the moment of the call, meaning that
  // the memcpy can be discarded rather than moved.

  // Deliberately get the source and destination with bitcasts stripped away,
  // because we'll need to do type comparisons based on the underlying type.
  CallSite CS(C);

  // Require that src be an alloca.  This simplifies the reasoning considerably.
  AllocaInst *srcAlloca = dyn_cast<AllocaInst>(cpySrc);
  if (!srcAlloca)
    return false;

  ConstantInt *srcArraySize = dyn_cast<ConstantInt>(srcAlloca->getArraySize());
  if (!srcArraySize)
    return false;

  const DataLayout &DL = cpy->getModule()->getDataLayout();
  uint64_t srcSize = DL.getTypeAllocSize(srcAlloca->getAllocatedType()) *
                     srcArraySize->getZExtValue();

  if (cpyLen < srcSize)
    return false;

  // Check that accessing the first srcSize bytes of dest will not cause a
  // trap.  Otherwise the transform is invalid since it might cause a trap
  // to occur earlier than it otherwise would.
  if (AllocaInst *A = dyn_cast<AllocaInst>(cpyDest)) {
    // The destination is an alloca.  Check it is larger than srcSize.
    ConstantInt *destArraySize = dyn_cast<ConstantInt>(A->getArraySize());
    if (!destArraySize)
      return false;

    uint64_t destSize = DL.getTypeAllocSize(A->getAllocatedType()) *
                        destArraySize->getZExtValue();

    if (destSize < srcSize)
      return false;
  } else if (Argument *A = dyn_cast<Argument>(cpyDest)) {
    if (A->getDereferenceableBytes() < srcSize) {
      // If the destination is an sret parameter then only accesses that are
      // outside of the returned struct type can trap.
      if (!A->hasStructRetAttr())
        return false;

      Type *StructTy = cast<PointerType>(A->getType())->getElementType();
      if (!StructTy->isSized()) {
        // The call may never return and hence the copy-instruction may never
        // be executed, and therefore it's not safe to say "the destination
        // has at least <cpyLen> bytes, as implied by the copy-instruction",
        return false;
      }

      uint64_t destSize = DL.getTypeAllocSize(StructTy);
      if (destSize < srcSize)
        return false;
    }
  } else {
    return false;
  }

  // Check that dest points to memory that is at least as aligned as src.
  unsigned srcAlign = srcAlloca->getAlignment();
  if (!srcAlign)
    srcAlign = DL.getABITypeAlignment(srcAlloca->getAllocatedType());
  bool isDestSufficientlyAligned = srcAlign <= cpyAlign;
  // If dest is not aligned enough and we can't increase its alignment then
  // bail out.
  if (!isDestSufficientlyAligned && !isa<AllocaInst>(cpyDest))
    return false;

  // Check that src is not accessed except via the call and the memcpy.  This
  // guarantees that it holds only undefined values when passed in (so the final
  // memcpy can be dropped), that it is not read or written between the call and
  // the memcpy, and that writing beyond the end of it is undefined.
  SmallVector<User*, 8> srcUseList(srcAlloca->user_begin(),
                                   srcAlloca->user_end());
  while (!srcUseList.empty()) {
    User *U = srcUseList.pop_back_val();

    if (isa<BitCastInst>(U) || isa<AddrSpaceCastInst>(U)) {
      for (User *UU : U->users())
        srcUseList.push_back(UU);
      continue;
    }
    if (GetElementPtrInst *G = dyn_cast<GetElementPtrInst>(U)) {
      if (!G->hasAllZeroIndices())
        return false;

      for (User *UU : U->users())
        srcUseList.push_back(UU);
      continue;
    }
    if (const IntrinsicInst *IT = dyn_cast<IntrinsicInst>(U))
      if (IT->getIntrinsicID() == Intrinsic::lifetime_start ||
          IT->getIntrinsicID() == Intrinsic::lifetime_end)
        continue;

    if (U != C && U != cpy)
      return false;
  }

  // Check that src isn't captured by the called function since the
  // transformation can cause aliasing issues in that case.
  for (unsigned i = 0, e = CS.arg_size(); i != e; ++i)
    if (CS.getArgument(i) == cpySrc && !CS.doesNotCapture(i))
      return false;

  // Since we're changing the parameter to the callsite, we need to make sure
  // that what would be the new parameter dominates the callsite.
  DominatorTree &DT = getAnalysis<DominatorTreeWrapperPass>().getDomTree();
  if (Instruction *cpyDestInst = dyn_cast<Instruction>(cpyDest))
    if (!DT.dominates(cpyDestInst, C))
      return false;

  // In addition to knowing that the call does not access src in some
  // unexpected manner, for example via a global, which we deduce from
  // the use analysis, we also need to know that it does not sneakily
  // access dest.  We rely on AA to figure this out for us.
  AliasAnalysis &AA = getAnalysis<AAResultsWrapperPass>().getAAResults();
  ModRefInfo MR = AA.getModRefInfo(C, cpyDest, srcSize);
  // If necessary, perform additional analysis.
  if (MR != MRI_NoModRef)
    MR = AA.callCapturesBefore(C, cpyDest, srcSize, &DT);
  if (MR != MRI_NoModRef)
    return false;

  // All the checks have passed, so do the transformation.
  bool changedArgument = false;
  for (unsigned i = 0; i < CS.arg_size(); ++i)
    if (CS.getArgument(i)->stripPointerCasts() == cpySrc) {
      Value *Dest = cpySrc->getType() == cpyDest->getType() ?  cpyDest
        : CastInst::CreatePointerCast(cpyDest, cpySrc->getType(),
                                      cpyDest->getName(), C);
      changedArgument = true;
      if (CS.getArgument(i)->getType() == Dest->getType())
        CS.setArgument(i, Dest);
      else
        CS.setArgument(i, CastInst::CreatePointerCast(Dest,
                          CS.getArgument(i)->getType(), Dest->getName(), C));
    }

  if (!changedArgument)
    return false;

  // If the destination wasn't sufficiently aligned then increase its alignment.
  if (!isDestSufficientlyAligned) {
    assert(isa<AllocaInst>(cpyDest) && "Can only increase alloca alignment!");
    cast<AllocaInst>(cpyDest)->setAlignment(srcAlign);
  }

  // Drop any cached information about the call, because we may have changed
  // its dependence information by changing its parameter.
  MD->removeInstruction(C);

  // Update AA metadata
  // FIXME: MD_tbaa_struct and MD_mem_parallel_loop_access should also be
  // handled here, but combineMetadata doesn't support them yet
  unsigned KnownIDs[] = {LLVMContext::MD_tbaa, LLVMContext::MD_alias_scope,
                         LLVMContext::MD_noalias,
                         LLVMContext::MD_invariant_group};
  combineMetadata(C, cpy, KnownIDs);

  // Remove the memcpy.
  MD->removeInstruction(cpy);
  ++NumMemCpyInstr;

  return true;
}

/// We've found that the (upward scanning) memory dependence of memcpy 'M' is
/// the memcpy 'MDep'. Try to simplify M to copy from MDep's input if we can.
bool MemCpyOpt::processMemCpyMemCpyDependence(MemCpyInst *M, MemCpyInst *MDep) {
  // We can only transforms memcpy's where the dest of one is the source of the
  // other.
  if (M->getSource() != MDep->getDest() || MDep->isVolatile())
    return false;

  // If dep instruction is reading from our current input, then it is a noop
  // transfer and substituting the input won't change this instruction.  Just
  // ignore the input and let someone else zap MDep.  This handles cases like:
  //    memcpy(a <- a)
  //    memcpy(b <- a)
  if (M->getSource() == MDep->getSource())
    return false;

  // Second, the length of the memcpy's must be the same, or the preceding one
  // must be larger than the following one.
  ConstantInt *MDepLen = dyn_cast<ConstantInt>(MDep->getLength());
  ConstantInt *MLen = dyn_cast<ConstantInt>(M->getLength());
  if (!MDepLen || !MLen || MDepLen->getZExtValue() < MLen->getZExtValue())
    return false;

  AliasAnalysis &AA = getAnalysis<AAResultsWrapperPass>().getAAResults();

  // Verify that the copied-from memory doesn't change in between the two
  // transfers.  For example, in:
  //    memcpy(a <- b)
  //    *b = 42;
  //    memcpy(c <- a)
  // It would be invalid to transform the second memcpy into memcpy(c <- b).
  //
  // TODO: If the code between M and MDep is transparent to the destination "c",
  // then we could still perform the xform by moving M up to the first memcpy.
  //
  // NOTE: This is conservative, it will stop on any read from the source loc,
  // not just the defining memcpy.
  MemDepResult SourceDep =
      MD->getPointerDependencyFrom(MemoryLocation::getForSource(MDep), false,
                                   M->getIterator(), M->getParent());
  if (!SourceDep.isClobber() || SourceDep.getInst() != MDep)
    return false;

  // If the dest of the second might alias the source of the first, then the
  // source and dest might overlap.  We still want to eliminate the intermediate
  // value, but we have to generate a memmove instead of memcpy.
  bool UseMemMove = false;
  if (!AA.isNoAlias(MemoryLocation::getForDest(M),
                    MemoryLocation::getForSource(MDep)))
    UseMemMove = true;

  // If all checks passed, then we can transform M.

  // TODO: Is this worth it if we're creating a less aligned memcpy? For
  // example we could be moving from movaps -> movq on x86.
  IRBuilder<> Builder(M);
  if (UseMemMove)
    Builder.CreateMemMove(M->getRawDest(), MDep->getRawSource(), M->getLength(),
                          M->getDestAlignment(), MDep->getSrcAlignment(),
                          M->isVolatile());
  else
    Builder.CreateMemCpy(M->getRawDest(), MDep->getRawSource(), M->getLength(),
                         M->getDestAlignment(), MDep->getSrcAlignment(),
                         M->isVolatile());

  // Remove the instruction we're replacing.
  MD->removeInstruction(M);
  M->eraseFromParent();
  ++NumMemCpyInstr;
  return true;
}

/// We've found that the (upward scanning) memory dependence of \p MemCpy is
/// \p MemSet.  Try to simplify \p MemSet to only set the trailing bytes that
/// weren't copied over by \p MemCpy.
///
/// In other words, transform:
/// \code
///   memset(dst, c, dst_size);
///   memcpy(dst, src, src_size);
/// \endcode
/// into:
/// \code
///   memcpy(dst, src, src_size);
///   memset(dst + src_size, c, dst_size <= src_size ? 0 : dst_size - src_size);
/// \endcode
bool MemCpyOpt::processMemSetMemCpyDependence(MemCpyInst *MemCpy,
                                              MemSetInst *MemSet) {
  // We can only transform memset/memcpy with the same destination.
  if (MemSet->getDest() != MemCpy->getDest())
    return false;

  // Check that there are no other dependencies on the memset destination.
  MemDepResult DstDepInfo =
      MD->getPointerDependencyFrom(MemoryLocation::getForDest(MemSet), false,
                                   MemCpy->getIterator(), MemCpy->getParent());
  if (DstDepInfo.getInst() != MemSet)
    return false;

  // Use the same i8* dest as the memcpy, killing the memset dest if different.
  Value *Dest = MemCpy->getRawDest();
  Value *DestSize = MemSet->getLength();
  Value *SrcSize = MemCpy->getLength();

  // By default, create an unaligned memset.
  unsigned Align = 1;
  // If Dest is aligned, and SrcSize is constant, use the minimum alignment
  // of the sum.
  const unsigned DestAlign =
      std::max(MemSet->getDestAlignment(), MemCpy->getDestAlignment());
  if (DestAlign > 1)
    if (ConstantInt *SrcSizeC = dyn_cast<ConstantInt>(SrcSize))
      Align = MinAlign(SrcSizeC->getZExtValue(), DestAlign);

  IRBuilder<> Builder(MemCpy);

  // If the sizes have different types, zext the smaller one.
  if (DestSize->getType() != SrcSize->getType()) {
    if (DestSize->getType()->getIntegerBitWidth() >
        SrcSize->getType()->getIntegerBitWidth())
      SrcSize = Builder.CreateZExt(SrcSize, DestSize->getType());
    else
      DestSize = Builder.CreateZExt(DestSize, SrcSize->getType());
  }

  Value *MemsetLen =
      Builder.CreateSelect(Builder.CreateICmpULE(DestSize, SrcSize),
                           ConstantInt::getNullValue(DestSize->getType()),
                           Builder.CreateSub(DestSize, SrcSize));
  Builder.CreateMemSet(Builder.CreateGEP(Dest, SrcSize), MemSet->getOperand(1),
                       MemsetLen, Align);

  MD->removeInstruction(MemSet);
  MemSet->eraseFromParent();
  return true;
}

/// Transform memcpy to memset when its source was just memset.
/// In other words, turn:
/// \code
///   memset(dst1, c, dst1_size);
///   memcpy(dst2, dst1, dst2_size);
/// \endcode
/// into:
/// \code
///   memset(dst1, c, dst1_size);
///   memset(dst2, c, dst2_size);
/// \endcode
/// When dst2_size <= dst1_size.
///
/// The \p MemCpy must have a Constant length.
bool MemCpyOpt::performMemCpyToMemSetOptzn(MemCpyInst *MemCpy,
                                           MemSetInst *MemSet) {
  // This only makes sense on memcpy(..., memset(...), ...).
  if (MemSet->getRawDest() != MemCpy->getRawSource())
    return false;

  ConstantInt *CopySize = cast<ConstantInt>(MemCpy->getLength());
  ConstantInt *MemSetSize = dyn_cast<ConstantInt>(MemSet->getLength());
  // Make sure the memcpy doesn't read any more than what the memset wrote.
  // Don't worry about sizes larger than i64.
  if (!MemSetSize || CopySize->getZExtValue() > MemSetSize->getZExtValue())
    return false;

  IRBuilder<> Builder(MemCpy);
  Builder.CreateMemSet(MemCpy->getRawDest(), MemSet->getOperand(1),
                       CopySize, MemCpy->getDestAlignment());
  return true;
}

/// Perform simplification of memcpy's.  If we have memcpy A
/// which copies X to Y, and memcpy B which copies Y to Z, then we can rewrite
/// B to be a memcpy from X to Z (or potentially a memmove, depending on
/// circumstances). This allows later passes to remove the first memcpy
/// altogether.
bool MemCpyOpt::processMemCpy(MemCpyInst *M) {
  // We can only optimize non-volatile memcpy's.
  if (M->isVolatile()) return false;

  // If the source and destination of the memcpy are the same, then zap it.
  if (M->getSource() == M->getDest()) {
    MD->removeInstruction(M);
    M->eraseFromParent();
    return false;
  }

  // If copying from a constant, try to turn the memcpy into a memset.
  if (GlobalVariable *GV = dyn_cast<GlobalVariable>(M->getSource()))
    if (GV->isConstant() && GV->hasDefinitiveInitializer())
      if (Value *ByteVal = isBytewiseValue(GV->getInitializer())) {
        IRBuilder<> Builder(M);
        Builder.CreateMemSet(M->getRawDest(), ByteVal, M->getLength(),
                             M->getDestAlignment(), false);
        MD->removeInstruction(M);
        M->eraseFromParent();
        ++NumCpyToSet;
        return true;
      }

  MemDepResult DepInfo = MD->getDependency(M);

  // Try to turn a partially redundant memset + memcpy into
  // memcpy + smaller memset.  We don't need the memcpy size for this.
  if (DepInfo.isClobber())
    if (MemSetInst *MDep = dyn_cast<MemSetInst>(DepInfo.getInst()))
      if (processMemSetMemCpyDependence(M, MDep))
        return true;

  // The optimizations after this point require the memcpy size.
  ConstantInt *CopySize = dyn_cast<ConstantInt>(M->getLength());
  if (!CopySize) return false;

  // There are four possible optimizations we can do for memcpy:
  //   a) memcpy-memcpy xform which exposes redundance for DSE.
  //   b) call-memcpy xform for return slot optimization.
  //   c) memcpy from freshly alloca'd space or space that has just started its
  //      lifetime copies undefined data, and we can therefore eliminate the
  //      memcpy in favor of the data that was already at the destination.
  //   d) memcpy from a just-memset'd source can be turned into memset.
  if (DepInfo.isClobber()) {
    if (CallInst *C = dyn_cast<CallInst>(DepInfo.getInst())) {
      // FIXME: Can we pass in either of dest/src alignment here instead of
      // convervatively taking the minimum?
      unsigned Align = std::min(M->getDestAlignment(), M->getSrcAlignment());
      if (performCallSlotOptzn(M, M->getDest(), M->getSource(),
                               CopySize->getZExtValue(), Align,
                               C)) {
        MD->removeInstruction(M);
        M->eraseFromParent();
        return true;
      }
    }
  }

  MemoryLocation SrcLoc = MemoryLocation::getForSource(M);
  MemDepResult SrcDepInfo = MD->getPointerDependencyFrom(
      SrcLoc, true, M->getIterator(), M->getParent());

  if (SrcDepInfo.isClobber()) {
    if (MemCpyInst *MDep = dyn_cast<MemCpyInst>(SrcDepInfo.getInst()))
      return processMemCpyMemCpyDependence(M, MDep);
  } else if (SrcDepInfo.isDef()) {
    Instruction *I = SrcDepInfo.getInst();
    bool hasUndefContents = false;

    if (isa<AllocaInst>(I)) {
      hasUndefContents = true;
    } else if (IntrinsicInst *II = dyn_cast<IntrinsicInst>(I)) {
      if (II->getIntrinsicID() == Intrinsic::lifetime_start)
        if (ConstantInt *LTSize = dyn_cast<ConstantInt>(II->getArgOperand(0)))
          if (LTSize->getZExtValue() >= CopySize->getZExtValue())
            hasUndefContents = true;
    }

    if (hasUndefContents) {
      MD->removeInstruction(M);
      M->eraseFromParent();
      ++NumMemCpyInstr;
      return true;
    }
  }

  if (SrcDepInfo.isClobber())
    if (MemSetInst *MDep = dyn_cast<MemSetInst>(SrcDepInfo.getInst()))
      if (performMemCpyToMemSetOptzn(M, MDep)) {
        MD->removeInstruction(M);
        M->eraseFromParent();
        ++NumCpyToSet;
        return true;
      }

  return false;
}

/// Transforms memmove calls to memcpy calls when the src/dst are guaranteed
/// not to alias.
bool MemCpyOpt::processMemMove(MemMoveInst *M) {
  AliasAnalysis &AA = getAnalysis<AAResultsWrapperPass>().getAAResults();

  if (!TLI->has(LibFunc::memmove))
    return false;

  // See if the pointers alias.
  if (!AA.isNoAlias(MemoryLocation::getForDest(M),
                    MemoryLocation::getForSource(M)))
    return false;

  DEBUG(dbgs() << "MemCpyOpt: Optimizing memmove -> memcpy: " << *M << "\n");

  // If not, then we know we can transform this.
  Module *Mod = M->getParent()->getParent()->getParent();
  Type *ArgTys[3] = { M->getRawDest()->getType(),
                      M->getRawSource()->getType(),
                      M->getLength()->getType() };
  M->setCalledFunction(Intrinsic::getDeclaration(Mod, Intrinsic::memcpy,
                                                 ArgTys));

  // MemDep may have over conservative information about this instruction, just
  // conservatively flush it from the cache.
  MD->removeInstruction(M);

  ++NumMoveToCpy;
  return true;
}

/// This is called on every byval argument in call sites.
bool MemCpyOpt::processByValArgument(CallSite CS, unsigned ArgNo) {
  const DataLayout &DL = CS.getCaller()->getParent()->getDataLayout();
  // Find out what feeds this byval argument.
  Value *ByValArg = CS.getArgument(ArgNo);
  Type *ByValTy = cast<PointerType>(ByValArg->getType())->getElementType();
  uint64_t ByValSize = DL.getTypeAllocSize(ByValTy);
  MemDepResult DepInfo = MD->getPointerDependencyFrom(
      MemoryLocation(ByValArg, ByValSize), true,
      CS.getInstruction()->getIterator(), CS.getInstruction()->getParent());
  if (!DepInfo.isClobber())
    return false;

  // If the byval argument isn't fed by a memcpy, ignore it.  If it is fed by
  // a memcpy, see if we can byval from the source of the memcpy instead of the
  // result.
  MemCpyInst *MDep = dyn_cast<MemCpyInst>(DepInfo.getInst());
  if (!MDep || MDep->isVolatile() ||
      ByValArg->stripPointerCasts() != MDep->getDest())
    return false;

  // The length of the memcpy must be larger or equal to the size of the byval.
  ConstantInt *C1 = dyn_cast<ConstantInt>(MDep->getLength());
  if (!C1 || C1->getValue().getZExtValue() < ByValSize)
    return false;

  // Get the alignment of the byval.  If the call doesn't specify the alignment,
  // then it is some target specific value that we can't know.
  unsigned ByValAlign = CS.getParamAlignment(ArgNo+1);
  if (ByValAlign == 0) return false;

  // If it is greater than the memcpy, then we check to see if we can force the
  // source of the memcpy to the alignment we need.  If we fail, we bail out.
  AssumptionCache &AC =
      getAnalysis<AssumptionCacheTracker>().getAssumptionCache(
          *CS->getParent()->getParent());
  DominatorTree &DT = getAnalysis<DominatorTreeWrapperPass>().getDomTree();
  // FIXME: Can we use either of dest/src alignment here instead of
  // convervatively taking the minimum?
  unsigned MinAlign = std::min(MDep->getDestAlignment(),
                               MDep->getSrcAlignment());
  if (MinAlign < ByValAlign &&
      getOrEnforceKnownAlignment(MDep->getSource(), ByValAlign, DL,
                                 CS.getInstruction(), &AC, &DT) < ByValAlign)
    return false;

  // Verify that the copied-from memory doesn't change in between the memcpy and
  // the byval call.
  //    memcpy(a <- b)
  //    *b = 42;
  //    foo(*a)
  // It would be invalid to transform the second memcpy into foo(*b).
  //
  // NOTE: This is conservative, it will stop on any read from the source loc,
  // not just the defining memcpy.
  MemDepResult SourceDep = MD->getPointerDependencyFrom(
      MemoryLocation::getForSource(MDep), false,
      CS.getInstruction()->getIterator(), MDep->getParent());
  if (!SourceDep.isClobber() || SourceDep.getInst() != MDep)
    return false;

  Value *TmpCast = MDep->getSource();
  if (MDep->getSource()->getType() != ByValArg->getType())
    TmpCast = new BitCastInst(MDep->getSource(), ByValArg->getType(),
                              "tmpcast", CS.getInstruction());

  DEBUG(dbgs() << "MemCpyOpt: Forwarding memcpy to byval:\n"
               << "  " << *MDep << "\n"
               << "  " << *CS.getInstruction() << "\n");

  // Otherwise we're good!  Update the byval argument.
  CS.setArgument(ArgNo, TmpCast);
  ++NumMemCpyInstr;
  return true;
}

/// Executes one iteration of MemCpyOpt.
bool MemCpyOpt::iterateOnFunction(Function &F) {
  bool MadeChange = false;

  // Walk all instruction in the function.
  for (Function::iterator BB = F.begin(), BBE = F.end(); BB != BBE; ++BB) {
    for (BasicBlock::iterator BI = BB->begin(), BE = BB->end(); BI != BE;) {
      // Avoid invalidating the iterator.
      Instruction *I = &*BI++;

      bool RepeatInstruction = false;

      if (StoreInst *SI = dyn_cast<StoreInst>(I))
        MadeChange |= processStore(SI, BI);
      else if (MemSetInst *M = dyn_cast<MemSetInst>(I))
        RepeatInstruction = processMemSet(M, BI);
      else if (MemCpyInst *M = dyn_cast<MemCpyInst>(I))
        RepeatInstruction = processMemCpy(M);
      else if (MemMoveInst *M = dyn_cast<MemMoveInst>(I))
        RepeatInstruction = processMemMove(M);
      else if (auto CS = CallSite(I)) {
        for (unsigned i = 0, e = CS.arg_size(); i != e; ++i)
          if (CS.isByValArgument(i))
            MadeChange |= processByValArgument(CS, i);
      }

      // Reprocess the instruction if desired.
      if (RepeatInstruction) {
        if (BI != BB->begin()) --BI;
        MadeChange = true;
      }
    }
  }

  return MadeChange;
}

/// This is the main transformation entry point for a function.
bool MemCpyOpt::runOnFunction(Function &F) {
  if (skipOptnoneFunction(F))
    return false;

  bool MadeChange = false;
  MD = &getAnalysis<MemoryDependenceAnalysis>();
  TLI = &getAnalysis<TargetLibraryInfoWrapperPass>().getTLI();

  // If we don't have at least memset and memcpy, there is little point of doing
  // anything here.  These are required by a freestanding implementation, so if
  // even they are disabled, there is no point in trying hard.
  if (!TLI->has(LibFunc::memset) || !TLI->has(LibFunc::memcpy))
    return false;

  while (1) {
    if (!iterateOnFunction(F))
      break;
    MadeChange = true;
  }

  MD = nullptr;
  return MadeChange;
}