aboutsummaryrefslogtreecommitdiffstats
path: root/lib/Transforms/InstCombine/InstCombineMulDivRem.cpp
blob: 8c48dce81de71f446f3a62152bcf2f1dbf731a26 (plain)
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
//===- InstCombineMulDivRem.cpp -------------------------------------------===//
//
//                     The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This file implements the visit functions for mul, fmul, sdiv, udiv, fdiv,
// srem, urem, frem.
//
//===----------------------------------------------------------------------===//

#include "InstCombine.h"
#include "llvm/Analysis/InstructionSimplify.h"
#include "llvm/IR/IntrinsicInst.h"
#include "llvm/IR/PatternMatch.h"
using namespace llvm;
using namespace PatternMatch;

#define DEBUG_TYPE "instcombine"


/// simplifyValueKnownNonZero - The specific integer value is used in a context
/// where it is known to be non-zero.  If this allows us to simplify the
/// computation, do so and return the new operand, otherwise return null.
static Value *simplifyValueKnownNonZero(Value *V, InstCombiner &IC,
                                        Instruction *CxtI) {
  // If V has multiple uses, then we would have to do more analysis to determine
  // if this is safe.  For example, the use could be in dynamically unreached
  // code.
  if (!V->hasOneUse()) return nullptr;

  bool MadeChange = false;

  // ((1 << A) >>u B) --> (1 << (A-B))
  // Because V cannot be zero, we know that B is less than A.
  Value *A = nullptr, *B = nullptr, *One = nullptr;
  if (match(V, m_LShr(m_OneUse(m_Shl(m_Value(One), m_Value(A))), m_Value(B))) &&
      match(One, m_One())) {
    A = IC.Builder->CreateSub(A, B);
    return IC.Builder->CreateShl(One, A);
  }

  // (PowerOfTwo >>u B) --> isExact since shifting out the result would make it
  // inexact.  Similarly for <<.
  if (BinaryOperator *I = dyn_cast<BinaryOperator>(V))
    if (I->isLogicalShift() && isKnownToBeAPowerOfTwo(I->getOperand(0), false,
                                                      0, IC.getAssumptionTracker(),
                                                      CxtI,
                                                      IC.getDominatorTree())) {
      // We know that this is an exact/nuw shift and that the input is a
      // non-zero context as well.
      if (Value *V2 = simplifyValueKnownNonZero(I->getOperand(0), IC, CxtI)) {
        I->setOperand(0, V2);
        MadeChange = true;
      }

      if (I->getOpcode() == Instruction::LShr && !I->isExact()) {
        I->setIsExact();
        MadeChange = true;
      }

      if (I->getOpcode() == Instruction::Shl && !I->hasNoUnsignedWrap()) {
        I->setHasNoUnsignedWrap();
        MadeChange = true;
      }
    }

  // TODO: Lots more we could do here:
  //    If V is a phi node, we can call this on each of its operands.
  //    "select cond, X, 0" can simplify to "X".

  return MadeChange ? V : nullptr;
}


/// MultiplyOverflows - True if the multiply can not be expressed in an int
/// this size.
static bool MultiplyOverflows(const APInt &C1, const APInt &C2, APInt &Product,
                              bool IsSigned) {
  bool Overflow;
  if (IsSigned)
    Product = C1.smul_ov(C2, Overflow);
  else
    Product = C1.umul_ov(C2, Overflow);

  return Overflow;
}

/// \brief True if C2 is a multiple of C1. Quotient contains C2/C1.
static bool IsMultiple(const APInt &C1, const APInt &C2, APInt &Quotient,
                       bool IsSigned) {
  assert(C1.getBitWidth() == C2.getBitWidth() &&
         "Inconsistent width of constants!");

  APInt Remainder(C1.getBitWidth(), /*Val=*/0ULL, IsSigned);
  if (IsSigned)
    APInt::sdivrem(C1, C2, Quotient, Remainder);
  else
    APInt::udivrem(C1, C2, Quotient, Remainder);

  return Remainder.isMinValue();
}

/// \brief A helper routine of InstCombiner::visitMul().
///
/// If C is a vector of known powers of 2, then this function returns
/// a new vector obtained from C replacing each element with its logBase2.
/// Return a null pointer otherwise.
static Constant *getLogBase2Vector(ConstantDataVector *CV) {
  const APInt *IVal;
  SmallVector<Constant *, 4> Elts;

  for (unsigned I = 0, E = CV->getNumElements(); I != E; ++I) {
    Constant *Elt = CV->getElementAsConstant(I);
    if (!match(Elt, m_APInt(IVal)) || !IVal->isPowerOf2())
      return nullptr;
    Elts.push_back(ConstantInt::get(Elt->getType(), IVal->logBase2()));
  }

  return ConstantVector::get(Elts);
}

Instruction *InstCombiner::visitMul(BinaryOperator &I) {
  bool Changed = SimplifyAssociativeOrCommutative(I);
  Value *Op0 = I.getOperand(0), *Op1 = I.getOperand(1);

  if (Value *V = SimplifyVectorOp(I))
    return ReplaceInstUsesWith(I, V);

  if (Value *V = SimplifyMulInst(Op0, Op1, DL, TLI, DT, AT))
    return ReplaceInstUsesWith(I, V);

  if (Value *V = SimplifyUsingDistributiveLaws(I))
    return ReplaceInstUsesWith(I, V);

  if (match(Op1, m_AllOnes()))  // X * -1 == 0 - X
    return BinaryOperator::CreateNeg(Op0, I.getName());

  // Also allow combining multiply instructions on vectors.
  {
    Value *NewOp;
    Constant *C1, *C2;
    const APInt *IVal;
    if (match(&I, m_Mul(m_Shl(m_Value(NewOp), m_Constant(C2)),
                        m_Constant(C1))) &&
        match(C1, m_APInt(IVal)))
      // ((X << C1)*C2) == (X * (C2 << C1))
      return BinaryOperator::CreateMul(NewOp, ConstantExpr::getShl(C1, C2));

    if (match(&I, m_Mul(m_Value(NewOp), m_Constant(C1)))) {
      Constant *NewCst = nullptr;
      if (match(C1, m_APInt(IVal)) && IVal->isPowerOf2())
        // Replace X*(2^C) with X << C, where C is either a scalar or a splat.
        NewCst = ConstantInt::get(NewOp->getType(), IVal->logBase2());
      else if (ConstantDataVector *CV = dyn_cast<ConstantDataVector>(C1))
        // Replace X*(2^C) with X << C, where C is a vector of known
        // constant powers of 2.
        NewCst = getLogBase2Vector(CV);

      if (NewCst) {
        BinaryOperator *Shl = BinaryOperator::CreateShl(NewOp, NewCst);

        if (I.hasNoUnsignedWrap())
          Shl->setHasNoUnsignedWrap();

        return Shl;
      }
    }
  }

  if (ConstantInt *CI = dyn_cast<ConstantInt>(Op1)) {
    // (Y - X) * (-(2**n)) -> (X - Y) * (2**n), for positive nonzero n
    // (Y + const) * (-(2**n)) -> (-constY) * (2**n), for positive nonzero n
    // The "* (2**n)" thus becomes a potential shifting opportunity.
    {
      const APInt &   Val = CI->getValue();
      const APInt &PosVal = Val.abs();
      if (Val.isNegative() && PosVal.isPowerOf2()) {
        Value *X = nullptr, *Y = nullptr;
        if (Op0->hasOneUse()) {
          ConstantInt *C1;
          Value *Sub = nullptr;
          if (match(Op0, m_Sub(m_Value(Y), m_Value(X))))
            Sub = Builder->CreateSub(X, Y, "suba");
          else if (match(Op0, m_Add(m_Value(Y), m_ConstantInt(C1))))
            Sub = Builder->CreateSub(Builder->CreateNeg(C1), Y, "subc");
          if (Sub)
            return
              BinaryOperator::CreateMul(Sub,
                                        ConstantInt::get(Y->getType(), PosVal));
        }
      }
    }
  }

  // Simplify mul instructions with a constant RHS.
  if (isa<Constant>(Op1)) {
    // Try to fold constant mul into select arguments.
    if (SelectInst *SI = dyn_cast<SelectInst>(Op0))
      if (Instruction *R = FoldOpIntoSelect(I, SI))
        return R;

    if (isa<PHINode>(Op0))
      if (Instruction *NV = FoldOpIntoPhi(I))
        return NV;

    // Canonicalize (X+C1)*CI -> X*CI+C1*CI.
    {
      Value *X;
      Constant *C1;
      if (match(Op0, m_OneUse(m_Add(m_Value(X), m_Constant(C1))))) {
        Value *Mul = Builder->CreateMul(C1, Op1);
        // Only go forward with the transform if C1*CI simplifies to a tidier
        // constant.
        if (!match(Mul, m_Mul(m_Value(), m_Value())))
          return BinaryOperator::CreateAdd(Builder->CreateMul(X, Op1), Mul);
      }
    }
  }

  if (Value *Op0v = dyn_castNegVal(Op0))     // -X * -Y = X*Y
    if (Value *Op1v = dyn_castNegVal(Op1))
      return BinaryOperator::CreateMul(Op0v, Op1v);

  // (X / Y) *  Y = X - (X % Y)
  // (X / Y) * -Y = (X % Y) - X
  {
    Value *Op1C = Op1;
    BinaryOperator *BO = dyn_cast<BinaryOperator>(Op0);
    if (!BO ||
        (BO->getOpcode() != Instruction::UDiv &&
         BO->getOpcode() != Instruction::SDiv)) {
      Op1C = Op0;
      BO = dyn_cast<BinaryOperator>(Op1);
    }
    Value *Neg = dyn_castNegVal(Op1C);
    if (BO && BO->hasOneUse() &&
        (BO->getOperand(1) == Op1C || BO->getOperand(1) == Neg) &&
        (BO->getOpcode() == Instruction::UDiv ||
         BO->getOpcode() == Instruction::SDiv)) {
      Value *Op0BO = BO->getOperand(0), *Op1BO = BO->getOperand(1);

      // If the division is exact, X % Y is zero, so we end up with X or -X.
      if (PossiblyExactOperator *SDiv = dyn_cast<PossiblyExactOperator>(BO))
        if (SDiv->isExact()) {
          if (Op1BO == Op1C)
            return ReplaceInstUsesWith(I, Op0BO);
          return BinaryOperator::CreateNeg(Op0BO);
        }

      Value *Rem;
      if (BO->getOpcode() == Instruction::UDiv)
        Rem = Builder->CreateURem(Op0BO, Op1BO);
      else
        Rem = Builder->CreateSRem(Op0BO, Op1BO);
      Rem->takeName(BO);

      if (Op1BO == Op1C)
        return BinaryOperator::CreateSub(Op0BO, Rem);
      return BinaryOperator::CreateSub(Rem, Op0BO);
    }
  }

  /// i1 mul -> i1 and.
  if (I.getType()->getScalarType()->isIntegerTy(1))
    return BinaryOperator::CreateAnd(Op0, Op1);

  // X*(1 << Y) --> X << Y
  // (1 << Y)*X --> X << Y
  {
    Value *Y;
    if (match(Op0, m_Shl(m_One(), m_Value(Y))))
      return BinaryOperator::CreateShl(Op1, Y);
    if (match(Op1, m_Shl(m_One(), m_Value(Y))))
      return BinaryOperator::CreateShl(Op0, Y);
  }

  // If one of the operands of the multiply is a cast from a boolean value, then
  // we know the bool is either zero or one, so this is a 'masking' multiply.
  //   X * Y (where Y is 0 or 1) -> X & (0-Y)
  if (!I.getType()->isVectorTy()) {
    // -2 is "-1 << 1" so it is all bits set except the low one.
    APInt Negative2(I.getType()->getPrimitiveSizeInBits(), (uint64_t)-2, true);

    Value *BoolCast = nullptr, *OtherOp = nullptr;
    if (MaskedValueIsZero(Op0, Negative2, 0, &I))
      BoolCast = Op0, OtherOp = Op1;
    else if (MaskedValueIsZero(Op1, Negative2, 0, &I))
      BoolCast = Op1, OtherOp = Op0;

    if (BoolCast) {
      Value *V = Builder->CreateSub(Constant::getNullValue(I.getType()),
                                    BoolCast);
      return BinaryOperator::CreateAnd(V, OtherOp);
    }
  }

  return Changed ? &I : nullptr;
}

/// Detect pattern log2(Y * 0.5) with corresponding fast math flags.
static void detectLog2OfHalf(Value *&Op, Value *&Y, IntrinsicInst *&Log2) {
  if (!Op->hasOneUse())
    return;

  IntrinsicInst *II = dyn_cast<IntrinsicInst>(Op);
  if (!II)
    return;
  if (II->getIntrinsicID() != Intrinsic::log2 || !II->hasUnsafeAlgebra())
    return;
  Log2 = II;

  Value *OpLog2Of = II->getArgOperand(0);
  if (!OpLog2Of->hasOneUse())
    return;

  Instruction *I = dyn_cast<Instruction>(OpLog2Of);
  if (!I)
    return;
  if (I->getOpcode() != Instruction::FMul || !I->hasUnsafeAlgebra())
    return;

  if (match(I->getOperand(0), m_SpecificFP(0.5)))
    Y = I->getOperand(1);
  else if (match(I->getOperand(1), m_SpecificFP(0.5)))
    Y = I->getOperand(0);
}

static bool isFiniteNonZeroFp(Constant *C) {
  if (C->getType()->isVectorTy()) {
    for (unsigned I = 0, E = C->getType()->getVectorNumElements(); I != E;
         ++I) {
      ConstantFP *CFP = dyn_cast<ConstantFP>(C->getAggregateElement(I));
      if (!CFP || !CFP->getValueAPF().isFiniteNonZero())
        return false;
    }
    return true;
  }

  return isa<ConstantFP>(C) &&
         cast<ConstantFP>(C)->getValueAPF().isFiniteNonZero();
}

static bool isNormalFp(Constant *C) {
  if (C->getType()->isVectorTy()) {
    for (unsigned I = 0, E = C->getType()->getVectorNumElements(); I != E;
         ++I) {
      ConstantFP *CFP = dyn_cast<ConstantFP>(C->getAggregateElement(I));
      if (!CFP || !CFP->getValueAPF().isNormal())
        return false;
    }
    return true;
  }

  return isa<ConstantFP>(C) && cast<ConstantFP>(C)->getValueAPF().isNormal();
}

/// Helper function of InstCombiner::visitFMul(BinaryOperator(). It returns
/// true iff the given value is FMul or FDiv with one and only one operand
/// being a normal constant (i.e. not Zero/NaN/Infinity).
static bool isFMulOrFDivWithConstant(Value *V) {
  Instruction *I = dyn_cast<Instruction>(V);
  if (!I || (I->getOpcode() != Instruction::FMul &&
             I->getOpcode() != Instruction::FDiv))
    return false;

  Constant *C0 = dyn_cast<Constant>(I->getOperand(0));
  Constant *C1 = dyn_cast<Constant>(I->getOperand(1));

  if (C0 && C1)
    return false;

  return (C0 && isFiniteNonZeroFp(C0)) || (C1 && isFiniteNonZeroFp(C1));
}

/// foldFMulConst() is a helper routine of InstCombiner::visitFMul().
/// The input \p FMulOrDiv is a FMul/FDiv with one and only one operand
/// being a constant (i.e. isFMulOrFDivWithConstant(FMulOrDiv) == true).
/// This function is to simplify "FMulOrDiv * C" and returns the
/// resulting expression. Note that this function could return NULL in
/// case the constants cannot be folded into a normal floating-point.
///
Value *InstCombiner::foldFMulConst(Instruction *FMulOrDiv, Constant *C,
                                   Instruction *InsertBefore) {
  assert(isFMulOrFDivWithConstant(FMulOrDiv) && "V is invalid");

  Value *Opnd0 = FMulOrDiv->getOperand(0);
  Value *Opnd1 = FMulOrDiv->getOperand(1);

  Constant *C0 = dyn_cast<Constant>(Opnd0);
  Constant *C1 = dyn_cast<Constant>(Opnd1);

  BinaryOperator *R = nullptr;

  // (X * C0) * C => X * (C0*C)
  if (FMulOrDiv->getOpcode() == Instruction::FMul) {
    Constant *F = ConstantExpr::getFMul(C1 ? C1 : C0, C);
    if (isNormalFp(F))
      R = BinaryOperator::CreateFMul(C1 ? Opnd0 : Opnd1, F);
  } else {
    if (C0) {
      // (C0 / X) * C => (C0 * C) / X
      if (FMulOrDiv->hasOneUse()) {
        // It would otherwise introduce another div.
        Constant *F = ConstantExpr::getFMul(C0, C);
        if (isNormalFp(F))
          R = BinaryOperator::CreateFDiv(F, Opnd1);
      }
    } else {
      // (X / C1) * C => X * (C/C1) if C/C1 is not a denormal
      Constant *F = ConstantExpr::getFDiv(C, C1);
      if (isNormalFp(F)) {
        R = BinaryOperator::CreateFMul(Opnd0, F);
      } else {
        // (X / C1) * C => X / (C1/C)
        Constant *F = ConstantExpr::getFDiv(C1, C);
        if (isNormalFp(F))
          R = BinaryOperator::CreateFDiv(Opnd0, F);
      }
    }
  }

  if (R) {
    R->setHasUnsafeAlgebra(true);
    InsertNewInstWith(R, *InsertBefore);
  }

  return R;
}

Instruction *InstCombiner::visitFMul(BinaryOperator &I) {
  bool Changed = SimplifyAssociativeOrCommutative(I);
  Value *Op0 = I.getOperand(0), *Op1 = I.getOperand(1);

  if (Value *V = SimplifyVectorOp(I))
    return ReplaceInstUsesWith(I, V);

  if (isa<Constant>(Op0))
    std::swap(Op0, Op1);

  if (Value *V = SimplifyFMulInst(Op0, Op1, I.getFastMathFlags(), DL, TLI,
                                  DT, AT))
    return ReplaceInstUsesWith(I, V);

  bool AllowReassociate = I.hasUnsafeAlgebra();

  // Simplify mul instructions with a constant RHS.
  if (isa<Constant>(Op1)) {
    // Try to fold constant mul into select arguments.
    if (SelectInst *SI = dyn_cast<SelectInst>(Op0))
      if (Instruction *R = FoldOpIntoSelect(I, SI))
        return R;

    if (isa<PHINode>(Op0))
      if (Instruction *NV = FoldOpIntoPhi(I))
        return NV;

    // (fmul X, -1.0) --> (fsub -0.0, X)
    if (match(Op1, m_SpecificFP(-1.0))) {
      Constant *NegZero = ConstantFP::getNegativeZero(Op1->getType());
      Instruction *RI = BinaryOperator::CreateFSub(NegZero, Op0);
      RI->copyFastMathFlags(&I);
      return RI;
    }

    Constant *C = cast<Constant>(Op1);
    if (AllowReassociate && isFiniteNonZeroFp(C)) {
      // Let MDC denote an expression in one of these forms:
      // X * C, C/X, X/C, where C is a constant.
      //
      // Try to simplify "MDC * Constant"
      if (isFMulOrFDivWithConstant(Op0))
        if (Value *V = foldFMulConst(cast<Instruction>(Op0), C, &I))
          return ReplaceInstUsesWith(I, V);

      // (MDC +/- C1) * C => (MDC * C) +/- (C1 * C)
      Instruction *FAddSub = dyn_cast<Instruction>(Op0);
      if (FAddSub &&
          (FAddSub->getOpcode() == Instruction::FAdd ||
           FAddSub->getOpcode() == Instruction::FSub)) {
        Value *Opnd0 = FAddSub->getOperand(0);
        Value *Opnd1 = FAddSub->getOperand(1);
        Constant *C0 = dyn_cast<Constant>(Opnd0);
        Constant *C1 = dyn_cast<Constant>(Opnd1);
        bool Swap = false;
        if (C0) {
          std::swap(C0, C1);
          std::swap(Opnd0, Opnd1);
          Swap = true;
        }

        if (C1 && isFiniteNonZeroFp(C1) && isFMulOrFDivWithConstant(Opnd0)) {
          Value *M1 = ConstantExpr::getFMul(C1, C);
          Value *M0 = isNormalFp(cast<Constant>(M1)) ?
                      foldFMulConst(cast<Instruction>(Opnd0), C, &I) :
                      nullptr;
          if (M0 && M1) {
            if (Swap && FAddSub->getOpcode() == Instruction::FSub)
              std::swap(M0, M1);

            Instruction *RI = (FAddSub->getOpcode() == Instruction::FAdd)
                                  ? BinaryOperator::CreateFAdd(M0, M1)
                                  : BinaryOperator::CreateFSub(M0, M1);
            RI->copyFastMathFlags(&I);
            return RI;
          }
        }
      }
    }
  }

  // sqrt(X) * sqrt(X) -> X
  if (AllowReassociate && (Op0 == Op1))
    if (IntrinsicInst *II = dyn_cast<IntrinsicInst>(Op0))
      if (II->getIntrinsicID() == Intrinsic::sqrt)
        return ReplaceInstUsesWith(I, II->getOperand(0));

  // Under unsafe algebra do:
  // X * log2(0.5*Y) = X*log2(Y) - X
  if (AllowReassociate) {
    Value *OpX = nullptr;
    Value *OpY = nullptr;
    IntrinsicInst *Log2;
    detectLog2OfHalf(Op0, OpY, Log2);
    if (OpY) {
      OpX = Op1;
    } else {
      detectLog2OfHalf(Op1, OpY, Log2);
      if (OpY) {
        OpX = Op0;
      }
    }
    // if pattern detected emit alternate sequence
    if (OpX && OpY) {
      BuilderTy::FastMathFlagGuard Guard(*Builder);
      Builder->SetFastMathFlags(Log2->getFastMathFlags());
      Log2->setArgOperand(0, OpY);
      Value *FMulVal = Builder->CreateFMul(OpX, Log2);
      Value *FSub = Builder->CreateFSub(FMulVal, OpX);
      FSub->takeName(&I);
      return ReplaceInstUsesWith(I, FSub);
    }
  }

  // Handle symmetric situation in a 2-iteration loop
  Value *Opnd0 = Op0;
  Value *Opnd1 = Op1;
  for (int i = 0; i < 2; i++) {
    bool IgnoreZeroSign = I.hasNoSignedZeros();
    if (BinaryOperator::isFNeg(Opnd0, IgnoreZeroSign)) {
      BuilderTy::FastMathFlagGuard Guard(*Builder);
      Builder->SetFastMathFlags(I.getFastMathFlags());

      Value *N0 = dyn_castFNegVal(Opnd0, IgnoreZeroSign);
      Value *N1 = dyn_castFNegVal(Opnd1, IgnoreZeroSign);

      // -X * -Y => X*Y
      if (N1) {
        Value *FMul = Builder->CreateFMul(N0, N1);
        FMul->takeName(&I);
        return ReplaceInstUsesWith(I, FMul);
      }

      if (Opnd0->hasOneUse()) {
        // -X * Y => -(X*Y) (Promote negation as high as possible)
        Value *T = Builder->CreateFMul(N0, Opnd1);
        Value *Neg = Builder->CreateFNeg(T);
        Neg->takeName(&I);
        return ReplaceInstUsesWith(I, Neg);
      }
    }

    // (X*Y) * X => (X*X) * Y where Y != X
    //  The purpose is two-fold:
    //   1) to form a power expression (of X).
    //   2) potentially shorten the critical path: After transformation, the
    //  latency of the instruction Y is amortized by the expression of X*X,
    //  and therefore Y is in a "less critical" position compared to what it
    //  was before the transformation.
    //
    if (AllowReassociate) {
      Value *Opnd0_0, *Opnd0_1;
      if (Opnd0->hasOneUse() &&
          match(Opnd0, m_FMul(m_Value(Opnd0_0), m_Value(Opnd0_1)))) {
        Value *Y = nullptr;
        if (Opnd0_0 == Opnd1 && Opnd0_1 != Opnd1)
          Y = Opnd0_1;
        else if (Opnd0_1 == Opnd1 && Opnd0_0 != Opnd1)
          Y = Opnd0_0;

        if (Y) {
          BuilderTy::FastMathFlagGuard Guard(*Builder);
          Builder->SetFastMathFlags(I.getFastMathFlags());
          Value *T = Builder->CreateFMul(Opnd1, Opnd1);

          Value *R = Builder->CreateFMul(T, Y);
          R->takeName(&I);
          return ReplaceInstUsesWith(I, R);
        }
      }
    }

    if (!isa<Constant>(Op1))
      std::swap(Opnd0, Opnd1);
    else
      break;
  }

  return Changed ? &I : nullptr;
}

/// SimplifyDivRemOfSelect - Try to fold a divide or remainder of a select
/// instruction.
bool InstCombiner::SimplifyDivRemOfSelect(BinaryOperator &I) {
  SelectInst *SI = cast<SelectInst>(I.getOperand(1));

  // div/rem X, (Cond ? 0 : Y) -> div/rem X, Y
  int NonNullOperand = -1;
  if (Constant *ST = dyn_cast<Constant>(SI->getOperand(1)))
    if (ST->isNullValue())
      NonNullOperand = 2;
  // div/rem X, (Cond ? Y : 0) -> div/rem X, Y
  if (Constant *ST = dyn_cast<Constant>(SI->getOperand(2)))
    if (ST->isNullValue())
      NonNullOperand = 1;

  if (NonNullOperand == -1)
    return false;

  Value *SelectCond = SI->getOperand(0);

  // Change the div/rem to use 'Y' instead of the select.
  I.setOperand(1, SI->getOperand(NonNullOperand));

  // Okay, we know we replace the operand of the div/rem with 'Y' with no
  // problem.  However, the select, or the condition of the select may have
  // multiple uses.  Based on our knowledge that the operand must be non-zero,
  // propagate the known value for the select into other uses of it, and
  // propagate a known value of the condition into its other users.

  // If the select and condition only have a single use, don't bother with this,
  // early exit.
  if (SI->use_empty() && SelectCond->hasOneUse())
    return true;

  // Scan the current block backward, looking for other uses of SI.
  BasicBlock::iterator BBI = &I, BBFront = I.getParent()->begin();

  while (BBI != BBFront) {
    --BBI;
    // If we found a call to a function, we can't assume it will return, so
    // information from below it cannot be propagated above it.
    if (isa<CallInst>(BBI) && !isa<IntrinsicInst>(BBI))
      break;

    // Replace uses of the select or its condition with the known values.
    for (Instruction::op_iterator I = BBI->op_begin(), E = BBI->op_end();
         I != E; ++I) {
      if (*I == SI) {
        *I = SI->getOperand(NonNullOperand);
        Worklist.Add(BBI);
      } else if (*I == SelectCond) {
        *I = Builder->getInt1(NonNullOperand == 1);
        Worklist.Add(BBI);
      }
    }

    // If we past the instruction, quit looking for it.
    if (&*BBI == SI)
      SI = nullptr;
    if (&*BBI == SelectCond)
      SelectCond = nullptr;

    // If we ran out of things to eliminate, break out of the loop.
    if (!SelectCond && !SI)
      break;

  }
  return true;
}


/// This function implements the transforms common to both integer division
/// instructions (udiv and sdiv). It is called by the visitors to those integer
/// division instructions.
/// @brief Common integer divide transforms
Instruction *InstCombiner::commonIDivTransforms(BinaryOperator &I) {
  Value *Op0 = I.getOperand(0), *Op1 = I.getOperand(1);

  // The RHS is known non-zero.
  if (Value *V = simplifyValueKnownNonZero(I.getOperand(1), *this, &I)) {
    I.setOperand(1, V);
    return &I;
  }

  // Handle cases involving: [su]div X, (select Cond, Y, Z)
  // This does not apply for fdiv.
  if (isa<SelectInst>(Op1) && SimplifyDivRemOfSelect(I))
    return &I;

  if (Instruction *LHS = dyn_cast<Instruction>(Op0)) {
    const APInt *C2;
    if (match(Op1, m_APInt(C2))) {
      Value *X;
      const APInt *C1;
      bool IsSigned = I.getOpcode() == Instruction::SDiv;

      // (X / C1) / C2  -> X / (C1*C2)
      if ((IsSigned && match(LHS, m_SDiv(m_Value(X), m_APInt(C1)))) ||
          (!IsSigned && match(LHS, m_UDiv(m_Value(X), m_APInt(C1))))) {
        APInt Product(C1->getBitWidth(), /*Val=*/0ULL, IsSigned);
        if (!MultiplyOverflows(*C1, *C2, Product, IsSigned))
          return BinaryOperator::Create(I.getOpcode(), X,
                                        ConstantInt::get(I.getType(), Product));
      }

      if ((IsSigned && match(LHS, m_NSWMul(m_Value(X), m_APInt(C1)))) ||
          (!IsSigned && match(LHS, m_NUWMul(m_Value(X), m_APInt(C1))))) {
        APInt Quotient(C1->getBitWidth(), /*Val=*/0ULL, IsSigned);

        // (X * C1) / C2 -> X / (C2 / C1) if C2 is a multiple of C1.
        if (IsMultiple(*C2, *C1, Quotient, IsSigned)) {
          BinaryOperator *BO = BinaryOperator::Create(
              I.getOpcode(), X, ConstantInt::get(X->getType(), Quotient));
          BO->setIsExact(I.isExact());
          return BO;
        }

        // (X * C1) / C2 -> X * (C1 / C2) if C1 is a multiple of C2.
        if (IsMultiple(*C1, *C2, Quotient, IsSigned)) {
          BinaryOperator *BO = BinaryOperator::Create(
              Instruction::Mul, X, ConstantInt::get(X->getType(), Quotient));
          BO->setHasNoUnsignedWrap(
              !IsSigned &&
              cast<OverflowingBinaryOperator>(LHS)->hasNoUnsignedWrap());
          BO->setHasNoSignedWrap(
              cast<OverflowingBinaryOperator>(LHS)->hasNoSignedWrap());
          return BO;
        }
      }

      if ((IsSigned && match(LHS, m_NSWShl(m_Value(X), m_APInt(C1))) &&
           *C1 != C1->getBitWidth() - 1) ||
          (!IsSigned && match(LHS, m_NUWShl(m_Value(X), m_APInt(C1))))) {
        APInt Quotient(C1->getBitWidth(), /*Val=*/0ULL, IsSigned);
        APInt C1Shifted = APInt::getOneBitSet(
            C1->getBitWidth(), static_cast<unsigned>(C1->getLimitedValue()));

        // (X << C1) / C2 -> X / (C2 >> C1) if C2 is a multiple of C1.
        if (IsMultiple(*C2, C1Shifted, Quotient, IsSigned)) {
          BinaryOperator *BO = BinaryOperator::Create(
              I.getOpcode(), X, ConstantInt::get(X->getType(), Quotient));
          BO->setIsExact(I.isExact());
          return BO;
        }

        // (X << C1) / C2 -> X * (C2 >> C1) if C1 is a multiple of C2.
        if (IsMultiple(C1Shifted, *C2, Quotient, IsSigned)) {
          BinaryOperator *BO = BinaryOperator::Create(
              Instruction::Mul, X, ConstantInt::get(X->getType(), Quotient));
          BO->setHasNoUnsignedWrap(
              !IsSigned &&
              cast<OverflowingBinaryOperator>(LHS)->hasNoUnsignedWrap());
          BO->setHasNoSignedWrap(
              cast<OverflowingBinaryOperator>(LHS)->hasNoSignedWrap());
          return BO;
        }
      }

      if (*C2 != 0) { // avoid X udiv 0
        if (SelectInst *SI = dyn_cast<SelectInst>(Op0))
          if (Instruction *R = FoldOpIntoSelect(I, SI))
            return R;
        if (isa<PHINode>(Op0))
          if (Instruction *NV = FoldOpIntoPhi(I))
            return NV;
      }
    }
  }

  if (ConstantInt *One = dyn_cast<ConstantInt>(Op0)) {
    if (One->isOne() && !I.getType()->isIntegerTy(1)) {
      bool isSigned = I.getOpcode() == Instruction::SDiv;
      if (isSigned) {
        // If Op1 is 0 then it's undefined behaviour, if Op1 is 1 then the
        // result is one, if Op1 is -1 then the result is minus one, otherwise
        // it's zero.
        Value *Inc = Builder->CreateAdd(Op1, One);
        Value *Cmp = Builder->CreateICmpULT(
                         Inc, ConstantInt::get(I.getType(), 3));
        return SelectInst::Create(Cmp, Op1, ConstantInt::get(I.getType(), 0));
      } else {
        // If Op1 is 0 then it's undefined behaviour. If Op1 is 1 then the
        // result is one, otherwise it's zero.
        return new ZExtInst(Builder->CreateICmpEQ(Op1, One), I.getType());
      }
    }
  }

  // See if we can fold away this div instruction.
  if (SimplifyDemandedInstructionBits(I))
    return &I;

  // (X - (X rem Y)) / Y -> X / Y; usually originates as ((X / Y) * Y) / Y
  Value *X = nullptr, *Z = nullptr;
  if (match(Op0, m_Sub(m_Value(X), m_Value(Z)))) { // (X - Z) / Y; Y = Op1
    bool isSigned = I.getOpcode() == Instruction::SDiv;
    if ((isSigned && match(Z, m_SRem(m_Specific(X), m_Specific(Op1)))) ||
        (!isSigned && match(Z, m_URem(m_Specific(X), m_Specific(Op1)))))
      return BinaryOperator::Create(I.getOpcode(), X, Op1);
  }

  return nullptr;
}

/// dyn_castZExtVal - Checks if V is a zext or constant that can
/// be truncated to Ty without losing bits.
static Value *dyn_castZExtVal(Value *V, Type *Ty) {
  if (ZExtInst *Z = dyn_cast<ZExtInst>(V)) {
    if (Z->getSrcTy() == Ty)
      return Z->getOperand(0);
  } else if (ConstantInt *C = dyn_cast<ConstantInt>(V)) {
    if (C->getValue().getActiveBits() <= cast<IntegerType>(Ty)->getBitWidth())
      return ConstantExpr::getTrunc(C, Ty);
  }
  return nullptr;
}

namespace {
const unsigned MaxDepth = 6;
typedef Instruction *(*FoldUDivOperandCb)(Value *Op0, Value *Op1,
                                          const BinaryOperator &I,
                                          InstCombiner &IC);

/// \brief Used to maintain state for visitUDivOperand().
struct UDivFoldAction {
  FoldUDivOperandCb FoldAction; ///< Informs visitUDiv() how to fold this
                                ///< operand.  This can be zero if this action
                                ///< joins two actions together.

  Value *OperandToFold;         ///< Which operand to fold.
  union {
    Instruction *FoldResult;    ///< The instruction returned when FoldAction is
                                ///< invoked.

    size_t SelectLHSIdx;        ///< Stores the LHS action index if this action
                                ///< joins two actions together.
  };

  UDivFoldAction(FoldUDivOperandCb FA, Value *InputOperand)
      : FoldAction(FA), OperandToFold(InputOperand), FoldResult(nullptr) {}
  UDivFoldAction(FoldUDivOperandCb FA, Value *InputOperand, size_t SLHS)
      : FoldAction(FA), OperandToFold(InputOperand), SelectLHSIdx(SLHS) {}
};
}

// X udiv 2^C -> X >> C
static Instruction *foldUDivPow2Cst(Value *Op0, Value *Op1,
                                    const BinaryOperator &I, InstCombiner &IC) {
  const APInt &C = cast<Constant>(Op1)->getUniqueInteger();
  BinaryOperator *LShr = BinaryOperator::CreateLShr(
      Op0, ConstantInt::get(Op0->getType(), C.logBase2()));
  if (I.isExact())
    LShr->setIsExact();
  return LShr;
}

// X udiv C, where C >= signbit
static Instruction *foldUDivNegCst(Value *Op0, Value *Op1,
                                   const BinaryOperator &I, InstCombiner &IC) {
  Value *ICI = IC.Builder->CreateICmpULT(Op0, cast<ConstantInt>(Op1));

  return SelectInst::Create(ICI, Constant::getNullValue(I.getType()),
                            ConstantInt::get(I.getType(), 1));
}

// X udiv (C1 << N), where C1 is "1<<C2"  -->  X >> (N+C2)
static Instruction *foldUDivShl(Value *Op0, Value *Op1, const BinaryOperator &I,
                                InstCombiner &IC) {
  Instruction *ShiftLeft = cast<Instruction>(Op1);
  if (isa<ZExtInst>(ShiftLeft))
    ShiftLeft = cast<Instruction>(ShiftLeft->getOperand(0));

  const APInt &CI =
      cast<Constant>(ShiftLeft->getOperand(0))->getUniqueInteger();
  Value *N = ShiftLeft->getOperand(1);
  if (CI != 1)
    N = IC.Builder->CreateAdd(N, ConstantInt::get(N->getType(), CI.logBase2()));
  if (ZExtInst *Z = dyn_cast<ZExtInst>(Op1))
    N = IC.Builder->CreateZExt(N, Z->getDestTy());
  BinaryOperator *LShr = BinaryOperator::CreateLShr(Op0, N);
  if (I.isExact())
    LShr->setIsExact();
  return LShr;
}

// \brief Recursively visits the possible right hand operands of a udiv
// instruction, seeing through select instructions, to determine if we can
// replace the udiv with something simpler.  If we find that an operand is not
// able to simplify the udiv, we abort the entire transformation.
static size_t visitUDivOperand(Value *Op0, Value *Op1, const BinaryOperator &I,
                               SmallVectorImpl<UDivFoldAction> &Actions,
                               unsigned Depth = 0) {
  // Check to see if this is an unsigned division with an exact power of 2,
  // if so, convert to a right shift.
  if (match(Op1, m_Power2())) {
    Actions.push_back(UDivFoldAction(foldUDivPow2Cst, Op1));
    return Actions.size();
  }

  if (ConstantInt *C = dyn_cast<ConstantInt>(Op1))
    // X udiv C, where C >= signbit
    if (C->getValue().isNegative()) {
      Actions.push_back(UDivFoldAction(foldUDivNegCst, C));
      return Actions.size();
    }

  // X udiv (C1 << N), where C1 is "1<<C2"  -->  X >> (N+C2)
  if (match(Op1, m_Shl(m_Power2(), m_Value())) ||
      match(Op1, m_ZExt(m_Shl(m_Power2(), m_Value())))) {
    Actions.push_back(UDivFoldAction(foldUDivShl, Op1));
    return Actions.size();
  }

  // The remaining tests are all recursive, so bail out if we hit the limit.
  if (Depth++ == MaxDepth)
    return 0;

  if (SelectInst *SI = dyn_cast<SelectInst>(Op1))
    if (size_t LHSIdx =
            visitUDivOperand(Op0, SI->getOperand(1), I, Actions, Depth))
      if (visitUDivOperand(Op0, SI->getOperand(2), I, Actions, Depth)) {
        Actions.push_back(UDivFoldAction(nullptr, Op1, LHSIdx - 1));
        return Actions.size();
      }

  return 0;
}

Instruction *InstCombiner::visitUDiv(BinaryOperator &I) {
  Value *Op0 = I.getOperand(0), *Op1 = I.getOperand(1);

  if (Value *V = SimplifyVectorOp(I))
    return ReplaceInstUsesWith(I, V);

  if (Value *V = SimplifyUDivInst(Op0, Op1, DL, TLI, DT, AT))
    return ReplaceInstUsesWith(I, V);

  // Handle the integer div common cases
  if (Instruction *Common = commonIDivTransforms(I))
    return Common;

  // (x lshr C1) udiv C2 --> x udiv (C2 << C1)
  {
    Value *X;
    const APInt *C1, *C2;
    if (match(Op0, m_LShr(m_Value(X), m_APInt(C1))) &&
        match(Op1, m_APInt(C2))) {
      bool Overflow;
      APInt C2ShlC1 = C2->ushl_ov(*C1, Overflow);
      if (!Overflow)
        return BinaryOperator::CreateUDiv(
            X, ConstantInt::get(X->getType(), C2ShlC1));
    }
  }

  // (zext A) udiv (zext B) --> zext (A udiv B)
  if (ZExtInst *ZOp0 = dyn_cast<ZExtInst>(Op0))
    if (Value *ZOp1 = dyn_castZExtVal(Op1, ZOp0->getSrcTy()))
      return new ZExtInst(
          Builder->CreateUDiv(ZOp0->getOperand(0), ZOp1, "div", I.isExact()),
          I.getType());

  // (LHS udiv (select (select (...)))) -> (LHS >> (select (select (...))))
  SmallVector<UDivFoldAction, 6> UDivActions;
  if (visitUDivOperand(Op0, Op1, I, UDivActions))
    for (unsigned i = 0, e = UDivActions.size(); i != e; ++i) {
      FoldUDivOperandCb Action = UDivActions[i].FoldAction;
      Value *ActionOp1 = UDivActions[i].OperandToFold;
      Instruction *Inst;
      if (Action)
        Inst = Action(Op0, ActionOp1, I, *this);
      else {
        // This action joins two actions together.  The RHS of this action is
        // simply the last action we processed, we saved the LHS action index in
        // the joining action.
        size_t SelectRHSIdx = i - 1;
        Value *SelectRHS = UDivActions[SelectRHSIdx].FoldResult;
        size_t SelectLHSIdx = UDivActions[i].SelectLHSIdx;
        Value *SelectLHS = UDivActions[SelectLHSIdx].FoldResult;
        Inst = SelectInst::Create(cast<SelectInst>(ActionOp1)->getCondition(),
                                  SelectLHS, SelectRHS);
      }

      // If this is the last action to process, return it to the InstCombiner.
      // Otherwise, we insert it before the UDiv and record it so that we may
      // use it as part of a joining action (i.e., a SelectInst).
      if (e - i != 1) {
        Inst->insertBefore(&I);
        UDivActions[i].FoldResult = Inst;
      } else
        return Inst;
    }

  return nullptr;
}

Instruction *InstCombiner::visitSDiv(BinaryOperator &I) {
  Value *Op0 = I.getOperand(0), *Op1 = I.getOperand(1);

  if (Value *V = SimplifyVectorOp(I))
    return ReplaceInstUsesWith(I, V);

  if (Value *V = SimplifySDivInst(Op0, Op1, DL, TLI, DT, AT))
    return ReplaceInstUsesWith(I, V);

  // Handle the integer div common cases
  if (Instruction *Common = commonIDivTransforms(I))
    return Common;

  // sdiv X, -1 == -X
  if (match(Op1, m_AllOnes()))
    return BinaryOperator::CreateNeg(Op0);

  if (ConstantInt *RHS = dyn_cast<ConstantInt>(Op1)) {
    // sdiv X, C  -->  ashr exact X, log2(C)
    if (I.isExact() && RHS->getValue().isNonNegative() &&
        RHS->getValue().isPowerOf2()) {
      Value *ShAmt = llvm::ConstantInt::get(RHS->getType(),
                                            RHS->getValue().exactLogBase2());
      return BinaryOperator::CreateExactAShr(Op0, ShAmt, I.getName());
    }
  }

  if (Constant *RHS = dyn_cast<Constant>(Op1)) {
    // X/INT_MIN -> X == INT_MIN
    if (RHS->isMinSignedValue())
      return new ZExtInst(Builder->CreateICmpEQ(Op0, Op1), I.getType());

    // -X/C  -->  X/-C  provided the negation doesn't overflow.
    if (SubOperator *Sub = dyn_cast<SubOperator>(Op0))
      if (match(Sub->getOperand(0), m_Zero()) && Sub->hasNoSignedWrap())
        return BinaryOperator::CreateSDiv(Sub->getOperand(1),
                                          ConstantExpr::getNeg(RHS));
  }

  // If the sign bits of both operands are zero (i.e. we can prove they are
  // unsigned inputs), turn this into a udiv.
  if (I.getType()->isIntegerTy()) {
    APInt Mask(APInt::getSignBit(I.getType()->getPrimitiveSizeInBits()));
    if (MaskedValueIsZero(Op0, Mask, 0, &I)) {
      if (MaskedValueIsZero(Op1, Mask, 0, &I)) {
        // X sdiv Y -> X udiv Y, iff X and Y don't have sign bit set
        return BinaryOperator::CreateUDiv(Op0, Op1, I.getName());
      }

      if (match(Op1, m_Shl(m_Power2(), m_Value()))) {
        // X sdiv (1 << Y) -> X udiv (1 << Y) ( -> X u>> Y)
        // Safe because the only negative value (1 << Y) can take on is
        // INT_MIN, and X sdiv INT_MIN == X udiv INT_MIN == 0 if X doesn't have
        // the sign bit set.
        return BinaryOperator::CreateUDiv(Op0, Op1, I.getName());
      }
    }
  }

  return nullptr;
}

/// CvtFDivConstToReciprocal tries to convert X/C into X*1/C if C not a special
/// FP value and:
///    1) 1/C is exact, or
///    2) reciprocal is allowed.
/// If the conversion was successful, the simplified expression "X * 1/C" is
/// returned; otherwise, NULL is returned.
///
static Instruction *CvtFDivConstToReciprocal(Value *Dividend, Constant *Divisor,
                                             bool AllowReciprocal) {
  if (!isa<ConstantFP>(Divisor)) // TODO: handle vectors.
    return nullptr;

  const APFloat &FpVal = cast<ConstantFP>(Divisor)->getValueAPF();
  APFloat Reciprocal(FpVal.getSemantics());
  bool Cvt = FpVal.getExactInverse(&Reciprocal);

  if (!Cvt && AllowReciprocal && FpVal.isFiniteNonZero()) {
    Reciprocal = APFloat(FpVal.getSemantics(), 1.0f);
    (void)Reciprocal.divide(FpVal, APFloat::rmNearestTiesToEven);
    Cvt = !Reciprocal.isDenormal();
  }

  if (!Cvt)
    return nullptr;

  ConstantFP *R;
  R = ConstantFP::get(Dividend->getType()->getContext(), Reciprocal);
  return BinaryOperator::CreateFMul(Dividend, R);
}

Instruction *InstCombiner::visitFDiv(BinaryOperator &I) {
  Value *Op0 = I.getOperand(0), *Op1 = I.getOperand(1);

  if (Value *V = SimplifyVectorOp(I))
    return ReplaceInstUsesWith(I, V);

  if (Value *V = SimplifyFDivInst(Op0, Op1, DL, TLI, DT, AT))
    return ReplaceInstUsesWith(I, V);

  if (isa<Constant>(Op0))
    if (SelectInst *SI = dyn_cast<SelectInst>(Op1))
      if (Instruction *R = FoldOpIntoSelect(I, SI))
        return R;

  bool AllowReassociate = I.hasUnsafeAlgebra();
  bool AllowReciprocal = I.hasAllowReciprocal();

  if (Constant *Op1C = dyn_cast<Constant>(Op1)) {
    if (SelectInst *SI = dyn_cast<SelectInst>(Op0))
      if (Instruction *R = FoldOpIntoSelect(I, SI))
        return R;

    if (AllowReassociate) {
      Constant *C1 = nullptr;
      Constant *C2 = Op1C;
      Value *X;
      Instruction *Res = nullptr;

      if (match(Op0, m_FMul(m_Value(X), m_Constant(C1)))) {
        // (X*C1)/C2 => X * (C1/C2)
        //
        Constant *C = ConstantExpr::getFDiv(C1, C2);
        if (isNormalFp(C))
          Res = BinaryOperator::CreateFMul(X, C);
      } else if (match(Op0, m_FDiv(m_Value(X), m_Constant(C1)))) {
        // (X/C1)/C2 => X /(C2*C1) [=> X * 1/(C2*C1) if reciprocal is allowed]
        //
        Constant *C = ConstantExpr::getFMul(C1, C2);
        if (isNormalFp(C)) {
          Res = CvtFDivConstToReciprocal(X, C, AllowReciprocal);
          if (!Res)
            Res = BinaryOperator::CreateFDiv(X, C);
        }
      }

      if (Res) {
        Res->setFastMathFlags(I.getFastMathFlags());
        return Res;
      }
    }

    // X / C => X * 1/C
    if (Instruction *T = CvtFDivConstToReciprocal(Op0, Op1C, AllowReciprocal)) {
      T->copyFastMathFlags(&I);
      return T;
    }

    return nullptr;
  }

  if (AllowReassociate && isa<Constant>(Op0)) {
    Constant *C1 = cast<Constant>(Op0), *C2;
    Constant *Fold = nullptr;
    Value *X;
    bool CreateDiv = true;

    // C1 / (X*C2) => (C1/C2) / X
    if (match(Op1, m_FMul(m_Value(X), m_Constant(C2))))
      Fold = ConstantExpr::getFDiv(C1, C2);
    else if (match(Op1, m_FDiv(m_Value(X), m_Constant(C2)))) {
      // C1 / (X/C2) => (C1*C2) / X
      Fold = ConstantExpr::getFMul(C1, C2);
    } else if (match(Op1, m_FDiv(m_Constant(C2), m_Value(X)))) {
      // C1 / (C2/X) => (C1/C2) * X
      Fold = ConstantExpr::getFDiv(C1, C2);
      CreateDiv = false;
    }

    if (Fold && isNormalFp(Fold)) {
      Instruction *R = CreateDiv ? BinaryOperator::CreateFDiv(Fold, X)
                                 : BinaryOperator::CreateFMul(X, Fold);
      R->setFastMathFlags(I.getFastMathFlags());
      return R;
    }
    return nullptr;
  }

  if (AllowReassociate) {
    Value *X, *Y;
    Value *NewInst = nullptr;
    Instruction *SimpR = nullptr;

    if (Op0->hasOneUse() && match(Op0, m_FDiv(m_Value(X), m_Value(Y)))) {
      // (X/Y) / Z => X / (Y*Z)
      //
      if (!isa<Constant>(Y) || !isa<Constant>(Op1)) {
        NewInst = Builder->CreateFMul(Y, Op1);
        if (Instruction *RI = dyn_cast<Instruction>(NewInst)) {
          FastMathFlags Flags = I.getFastMathFlags();
          Flags &= cast<Instruction>(Op0)->getFastMathFlags();
          RI->setFastMathFlags(Flags);
        }
        SimpR = BinaryOperator::CreateFDiv(X, NewInst);
      }
    } else if (Op1->hasOneUse() && match(Op1, m_FDiv(m_Value(X), m_Value(Y)))) {
      // Z / (X/Y) => Z*Y / X
      //
      if (!isa<Constant>(Y) || !isa<Constant>(Op0)) {
        NewInst = Builder->CreateFMul(Op0, Y);
        if (Instruction *RI = dyn_cast<Instruction>(NewInst)) {
          FastMathFlags Flags = I.getFastMathFlags();
          Flags &= cast<Instruction>(Op1)->getFastMathFlags();
          RI->setFastMathFlags(Flags);
        }
        SimpR = BinaryOperator::CreateFDiv(NewInst, X);
      }
    }

    if (NewInst) {
      if (Instruction *T = dyn_cast<Instruction>(NewInst))
        T->setDebugLoc(I.getDebugLoc());
      SimpR->setFastMathFlags(I.getFastMathFlags());
      return SimpR;
    }
  }

  return nullptr;
}

/// This function implements the transforms common to both integer remainder
/// instructions (urem and srem). It is called by the visitors to those integer
/// remainder instructions.
/// @brief Common integer remainder transforms
Instruction *InstCombiner::commonIRemTransforms(BinaryOperator &I) {
  Value *Op0 = I.getOperand(0), *Op1 = I.getOperand(1);

  // The RHS is known non-zero.
  if (Value *V = simplifyValueKnownNonZero(I.getOperand(1), *this, &I)) {
    I.setOperand(1, V);
    return &I;
  }

  // Handle cases involving: rem X, (select Cond, Y, Z)
  if (isa<SelectInst>(Op1) && SimplifyDivRemOfSelect(I))
    return &I;

  if (isa<Constant>(Op1)) {
    if (Instruction *Op0I = dyn_cast<Instruction>(Op0)) {
      if (SelectInst *SI = dyn_cast<SelectInst>(Op0I)) {
        if (Instruction *R = FoldOpIntoSelect(I, SI))
          return R;
      } else if (isa<PHINode>(Op0I)) {
        if (Instruction *NV = FoldOpIntoPhi(I))
          return NV;
      }

      // See if we can fold away this rem instruction.
      if (SimplifyDemandedInstructionBits(I))
        return &I;
    }
  }

  return nullptr;
}

Instruction *InstCombiner::visitURem(BinaryOperator &I) {
  Value *Op0 = I.getOperand(0), *Op1 = I.getOperand(1);

  if (Value *V = SimplifyVectorOp(I))
    return ReplaceInstUsesWith(I, V);

  if (Value *V = SimplifyURemInst(Op0, Op1, DL, TLI, DT, AT))
    return ReplaceInstUsesWith(I, V);

  if (Instruction *common = commonIRemTransforms(I))
    return common;

  // (zext A) urem (zext B) --> zext (A urem B)
  if (ZExtInst *ZOp0 = dyn_cast<ZExtInst>(Op0))
    if (Value *ZOp1 = dyn_castZExtVal(Op1, ZOp0->getSrcTy()))
      return new ZExtInst(Builder->CreateURem(ZOp0->getOperand(0), ZOp1),
                          I.getType());

  // X urem Y -> X and Y-1, where Y is a power of 2,
  if (isKnownToBeAPowerOfTwo(Op1, /*OrZero*/true, 0, AT, &I, DT)) {
    Constant *N1 = Constant::getAllOnesValue(I.getType());
    Value *Add = Builder->CreateAdd(Op1, N1);
    return BinaryOperator::CreateAnd(Op0, Add);
  }

  // 1 urem X -> zext(X != 1)
  if (match(Op0, m_One())) {
    Value *Cmp = Builder->CreateICmpNE(Op1, Op0);
    Value *Ext = Builder->CreateZExt(Cmp, I.getType());
    return ReplaceInstUsesWith(I, Ext);
  }

  return nullptr;
}

Instruction *InstCombiner::visitSRem(BinaryOperator &I) {
  Value *Op0 = I.getOperand(0), *Op1 = I.getOperand(1);

  if (Value *V = SimplifyVectorOp(I))
    return ReplaceInstUsesWith(I, V);

  if (Value *V = SimplifySRemInst(Op0, Op1, DL, TLI, DT, AT))
    return ReplaceInstUsesWith(I, V);

  // Handle the integer rem common cases
  if (Instruction *Common = commonIRemTransforms(I))
    return Common;

  {
    const APInt *Y;
    // X % -Y -> X % Y
    if (match(Op1, m_APInt(Y)) && Y->isNegative() && !Y->isMinSignedValue()) {
      Worklist.AddValue(I.getOperand(1));
      I.setOperand(1, ConstantInt::get(I.getType(), -*Y));
      return &I;
    }
  }

  // If the sign bits of both operands are zero (i.e. we can prove they are
  // unsigned inputs), turn this into a urem.
  if (I.getType()->isIntegerTy()) {
    APInt Mask(APInt::getSignBit(I.getType()->getPrimitiveSizeInBits()));
    if (MaskedValueIsZero(Op1, Mask, 0, &I) &&
        MaskedValueIsZero(Op0, Mask, 0, &I)) {
      // X srem Y -> X urem Y, iff X and Y don't have sign bit set
      return BinaryOperator::CreateURem(Op0, Op1, I.getName());
    }
  }

  // If it's a constant vector, flip any negative values positive.
  if (isa<ConstantVector>(Op1) || isa<ConstantDataVector>(Op1)) {
    Constant *C = cast<Constant>(Op1);
    unsigned VWidth = C->getType()->getVectorNumElements();

    bool hasNegative = false;
    bool hasMissing = false;
    for (unsigned i = 0; i != VWidth; ++i) {
      Constant *Elt = C->getAggregateElement(i);
      if (!Elt) {
        hasMissing = true;
        break;
      }

      if (ConstantInt *RHS = dyn_cast<ConstantInt>(Elt))
        if (RHS->isNegative())
          hasNegative = true;
    }

    if (hasNegative && !hasMissing) {
      SmallVector<Constant *, 16> Elts(VWidth);
      for (unsigned i = 0; i != VWidth; ++i) {
        Elts[i] = C->getAggregateElement(i);  // Handle undef, etc.
        if (ConstantInt *RHS = dyn_cast<ConstantInt>(Elts[i])) {
          if (RHS->isNegative())
            Elts[i] = cast<ConstantInt>(ConstantExpr::getNeg(RHS));
        }
      }

      Constant *NewRHSV = ConstantVector::get(Elts);
      if (NewRHSV != C) {  // Don't loop on -MININT
        Worklist.AddValue(I.getOperand(1));
        I.setOperand(1, NewRHSV);
        return &I;
      }
    }
  }

  return nullptr;
}

Instruction *InstCombiner::visitFRem(BinaryOperator &I) {
  Value *Op0 = I.getOperand(0), *Op1 = I.getOperand(1);

  if (Value *V = SimplifyVectorOp(I))
    return ReplaceInstUsesWith(I, V);

  if (Value *V = SimplifyFRemInst(Op0, Op1, DL, TLI, DT, AT))
    return ReplaceInstUsesWith(I, V);

  // Handle cases involving: rem X, (select Cond, Y, Z)
  if (isa<SelectInst>(Op1) && SimplifyDivRemOfSelect(I))
    return &I;

  return nullptr;
}