llvm.org GIT mirror llvm / master lib / Target / WebAssembly / WebAssemblyISelLowering.cpp
master

Tree @master (Download .tar.gz)

WebAssemblyISelLowering.cpp @masterraw · history · blame

   1
   2
   3
   4
   5
   6
   7
   8
   9
  10
  11
  12
  13
  14
  15
  16
  17
  18
  19
  20
  21
  22
  23
  24
  25
  26
  27
  28
  29
  30
  31
  32
  33
  34
  35
  36
  37
  38
  39
  40
  41
  42
  43
  44
  45
  46
  47
  48
  49
  50
  51
  52
  53
  54
  55
  56
  57
  58
  59
  60
  61
  62
  63
  64
  65
  66
  67
  68
  69
  70
  71
  72
  73
  74
  75
  76
  77
  78
  79
  80
  81
  82
  83
  84
  85
  86
  87
  88
  89
  90
  91
  92
  93
  94
  95
  96
  97
  98
  99
 100
 101
 102
 103
 104
 105
 106
 107
 108
 109
 110
 111
 112
 113
 114
 115
 116
 117
 118
 119
 120
 121
 122
 123
 124
 125
 126
 127
 128
 129
 130
 131
 132
 133
 134
 135
 136
 137
 138
 139
 140
 141
 142
 143
 144
 145
 146
 147
 148
 149
 150
 151
 152
 153
 154
 155
 156
 157
 158
 159
 160
 161
 162
 163
 164
 165
 166
 167
 168
 169
 170
 171
 172
 173
 174
 175
 176
 177
 178
 179
 180
 181
 182
 183
 184
 185
 186
 187
 188
 189
 190
 191
 192
 193
 194
 195
 196
 197
 198
 199
 200
 201
 202
 203
 204
 205
 206
 207
 208
 209
 210
 211
 212
 213
 214
 215
 216
 217
 218
 219
 220
 221
 222
 223
 224
 225
 226
 227
 228
 229
 230
 231
 232
 233
 234
 235
 236
 237
 238
 239
 240
 241
 242
 243
 244
 245
 246
 247
 248
 249
 250
 251
 252
 253
 254
 255
 256
 257
 258
 259
 260
 261
 262
 263
 264
 265
 266
 267
 268
 269
 270
 271
 272
 273
 274
 275
 276
 277
 278
 279
 280
 281
 282
 283
 284
 285
 286
 287
 288
 289
 290
 291
 292
 293
 294
 295
 296
 297
 298
 299
 300
 301
 302
 303
 304
 305
 306
 307
 308
 309
 310
 311
 312
 313
 314
 315
 316
 317
 318
 319
 320
 321
 322
 323
 324
 325
 326
 327
 328
 329
 330
 331
 332
 333
 334
 335
 336
 337
 338
 339
 340
 341
 342
 343
 344
 345
 346
 347
 348
 349
 350
 351
 352
 353
 354
 355
 356
 357
 358
 359
 360
 361
 362
 363
 364
 365
 366
 367
 368
 369
 370
 371
 372
 373
 374
 375
 376
 377
 378
 379
 380
 381
 382
 383
 384
 385
 386
 387
 388
 389
 390
 391
 392
 393
 394
 395
 396
 397
 398
 399
 400
 401
 402
 403
 404
 405
 406
 407
 408
 409
 410
 411
 412
 413
 414
 415
 416
 417
 418
 419
 420
 421
 422
 423
 424
 425
 426
 427
 428
 429
 430
 431
 432
 433
 434
 435
 436
 437
 438
 439
 440
 441
 442
 443
 444
 445
 446
 447
 448
 449
 450
 451
 452
 453
 454
 455
 456
 457
 458
 459
 460
 461
 462
 463
 464
 465
 466
 467
 468
 469
 470
 471
 472
 473
 474
 475
 476
 477
 478
 479
 480
 481
 482
 483
 484
 485
 486
 487
 488
 489
 490
 491
 492
 493
 494
 495
 496
 497
 498
 499
 500
 501
 502
 503
 504
 505
 506
 507
 508
 509
 510
 511
 512
 513
 514
 515
 516
 517
 518
 519
 520
 521
 522
 523
 524
 525
 526
 527
 528
 529
 530
 531
 532
 533
 534
 535
 536
 537
 538
 539
 540
 541
 542
 543
 544
 545
 546
 547
 548
 549
 550
 551
 552
 553
 554
 555
 556
 557
 558
 559
 560
 561
 562
 563
 564
 565
 566
 567
 568
 569
 570
 571
 572
 573
 574
 575
 576
 577
 578
 579
 580
 581
 582
 583
 584
 585
 586
 587
 588
 589
 590
 591
 592
 593
 594
 595
 596
 597
 598
 599
 600
 601
 602
 603
 604
 605
 606
 607
 608
 609
 610
 611
 612
 613
 614
 615
 616
 617
 618
 619
 620
 621
 622
 623
 624
 625
 626
 627
 628
 629
 630
 631
 632
 633
 634
 635
 636
 637
 638
 639
 640
 641
 642
 643
 644
 645
 646
 647
 648
 649
 650
 651
 652
 653
 654
 655
 656
 657
 658
 659
 660
 661
 662
 663
 664
 665
 666
 667
 668
 669
 670
 671
 672
 673
 674
 675
 676
 677
 678
 679
 680
 681
 682
 683
 684
 685
 686
 687
 688
 689
 690
 691
 692
 693
 694
 695
 696
 697
 698
 699
 700
 701
 702
 703
 704
 705
 706
 707
 708
 709
 710
 711
 712
 713
 714
 715
 716
 717
 718
 719
 720
 721
 722
 723
 724
 725
 726
 727
 728
 729
 730
 731
 732
 733
 734
 735
 736
 737
 738
 739
 740
 741
 742
 743
 744
 745
 746
 747
 748
 749
 750
 751
 752
 753
 754
 755
 756
 757
 758
 759
 760
 761
 762
 763
 764
 765
 766
 767
 768
 769
 770
 771
 772
 773
 774
 775
 776
 777
 778
 779
 780
 781
 782
 783
 784
 785
 786
 787
 788
 789
 790
 791
 792
 793
 794
 795
 796
 797
 798
 799
 800
 801
 802
 803
 804
 805
 806
 807
 808
 809
 810
 811
 812
 813
 814
 815
 816
 817
 818
 819
 820
 821
 822
 823
 824
 825
 826
 827
 828
 829
 830
 831
 832
 833
 834
 835
 836
 837
 838
 839
 840
 841
 842
 843
 844
 845
 846
 847
 848
 849
 850
 851
 852
 853
 854
 855
 856
 857
 858
 859
 860
 861
 862
 863
 864
 865
 866
 867
 868
 869
 870
 871
 872
 873
 874
 875
 876
 877
 878
 879
 880
 881
 882
 883
 884
 885
 886
 887
 888
 889
 890
 891
 892
 893
 894
 895
 896
 897
 898
 899
 900
 901
 902
 903
 904
 905
 906
 907
 908
 909
 910
 911
 912
 913
 914
 915
 916
 917
 918
 919
 920
 921
 922
 923
 924
 925
 926
 927
 928
 929
 930
 931
 932
 933
 934
 935
 936
 937
 938
 939
 940
 941
 942
 943
 944
 945
 946
 947
 948
 949
 950
 951
 952
 953
 954
 955
 956
 957
 958
 959
 960
 961
 962
 963
 964
 965
 966
 967
 968
 969
 970
 971
 972
 973
 974
 975
 976
 977
 978
 979
 980
 981
 982
 983
 984
 985
 986
 987
 988
 989
 990
 991
 992
 993
 994
 995
 996
 997
 998
 999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
//=- WebAssemblyISelLowering.cpp - WebAssembly DAG Lowering Implementation -==//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
///
/// \file
/// This file implements the WebAssemblyTargetLowering class.
///
//===----------------------------------------------------------------------===//

#include "WebAssemblyISelLowering.h"
#include "MCTargetDesc/WebAssemblyMCTargetDesc.h"
#include "WebAssemblyMachineFunctionInfo.h"
#include "WebAssemblySubtarget.h"
#include "WebAssemblyTargetMachine.h"
#include "llvm/CodeGen/Analysis.h"
#include "llvm/CodeGen/CallingConvLower.h"
#include "llvm/CodeGen/MachineInstrBuilder.h"
#include "llvm/CodeGen/MachineJumpTableInfo.h"
#include "llvm/CodeGen/MachineModuleInfo.h"
#include "llvm/CodeGen/MachineRegisterInfo.h"
#include "llvm/CodeGen/SelectionDAG.h"
#include "llvm/CodeGen/WasmEHFuncInfo.h"
#include "llvm/IR/DiagnosticInfo.h"
#include "llvm/IR/DiagnosticPrinter.h"
#include "llvm/IR/Function.h"
#include "llvm/IR/Intrinsics.h"
#include "llvm/Support/Debug.h"
#include "llvm/Support/ErrorHandling.h"
#include "llvm/Support/raw_ostream.h"
#include "llvm/Target/TargetOptions.h"
using namespace llvm;

#define DEBUG_TYPE "wasm-lower"

WebAssemblyTargetLowering::WebAssemblyTargetLowering(
    const TargetMachine &TM, const WebAssemblySubtarget &STI)
    : TargetLowering(TM), Subtarget(&STI) {
  auto MVTPtr = Subtarget->hasAddr64() ? MVT::i64 : MVT::i32;

  // Booleans always contain 0 or 1.
  setBooleanContents(ZeroOrOneBooleanContent);
  // Except in SIMD vectors
  setBooleanVectorContents(ZeroOrNegativeOneBooleanContent);
  // We don't know the microarchitecture here, so just reduce register pressure.
  setSchedulingPreference(Sched::RegPressure);
  // Tell ISel that we have a stack pointer.
  setStackPointerRegisterToSaveRestore(
      Subtarget->hasAddr64() ? WebAssembly::SP64 : WebAssembly::SP32);
  // Set up the register classes.
  addRegisterClass(MVT::i32, &WebAssembly::I32RegClass);
  addRegisterClass(MVT::i64, &WebAssembly::I64RegClass);
  addRegisterClass(MVT::f32, &WebAssembly::F32RegClass);
  addRegisterClass(MVT::f64, &WebAssembly::F64RegClass);
  if (Subtarget->hasSIMD128()) {
    addRegisterClass(MVT::v16i8, &WebAssembly::V128RegClass);
    addRegisterClass(MVT::v8i16, &WebAssembly::V128RegClass);
    addRegisterClass(MVT::v4i32, &WebAssembly::V128RegClass);
    addRegisterClass(MVT::v4f32, &WebAssembly::V128RegClass);
  }
  if (Subtarget->hasUnimplementedSIMD128()) {
    addRegisterClass(MVT::v2i64, &WebAssembly::V128RegClass);
    addRegisterClass(MVT::v2f64, &WebAssembly::V128RegClass);
  }
  // Compute derived properties from the register classes.
  computeRegisterProperties(Subtarget->getRegisterInfo());

  setOperationAction(ISD::GlobalAddress, MVTPtr, Custom);
  setOperationAction(ISD::ExternalSymbol, MVTPtr, Custom);
  setOperationAction(ISD::JumpTable, MVTPtr, Custom);
  setOperationAction(ISD::BlockAddress, MVTPtr, Custom);
  setOperationAction(ISD::BRIND, MVT::Other, Custom);

  // Take the default expansion for va_arg, va_copy, and va_end. There is no
  // default action for va_start, so we do that custom.
  setOperationAction(ISD::VASTART, MVT::Other, Custom);
  setOperationAction(ISD::VAARG, MVT::Other, Expand);
  setOperationAction(ISD::VACOPY, MVT::Other, Expand);
  setOperationAction(ISD::VAEND, MVT::Other, Expand);

  for (auto T : {MVT::f32, MVT::f64, MVT::v4f32, MVT::v2f64}) {
    // Don't expand the floating-point types to constant pools.
    setOperationAction(ISD::ConstantFP, T, Legal);
    // Expand floating-point comparisons.
    for (auto CC : {ISD::SETO, ISD::SETUO, ISD::SETUEQ, ISD::SETONE,
                    ISD::SETULT, ISD::SETULE, ISD::SETUGT, ISD::SETUGE})
      setCondCodeAction(CC, T, Expand);
    // Expand floating-point library function operators.
    for (auto Op :
         {ISD::FSIN, ISD::FCOS, ISD::FSINCOS, ISD::FPOW, ISD::FREM, ISD::FMA})
      setOperationAction(Op, T, Expand);
    // Note supported floating-point library function operators that otherwise
    // default to expand.
    for (auto Op :
         {ISD::FCEIL, ISD::FFLOOR, ISD::FTRUNC, ISD::FNEARBYINT, ISD::FRINT})
      setOperationAction(Op, T, Legal);
    // Support minimum and maximum, which otherwise default to expand.
    setOperationAction(ISD::FMINIMUM, T, Legal);
    setOperationAction(ISD::FMAXIMUM, T, Legal);
    // WebAssembly currently has no builtin f16 support.
    setOperationAction(ISD::FP16_TO_FP, T, Expand);
    setOperationAction(ISD::FP_TO_FP16, T, Expand);
    setLoadExtAction(ISD::EXTLOAD, T, MVT::f16, Expand);
    setTruncStoreAction(T, MVT::f16, Expand);
  }

  // Expand unavailable integer operations.
  for (auto Op :
       {ISD::BSWAP, ISD::SMUL_LOHI, ISD::UMUL_LOHI, ISD::MULHS, ISD::MULHU,
        ISD::SDIVREM, ISD::UDIVREM, ISD::SHL_PARTS, ISD::SRA_PARTS,
        ISD::SRL_PARTS, ISD::ADDC, ISD::ADDE, ISD::SUBC, ISD::SUBE}) {
    for (auto T : {MVT::i32, MVT::i64})
      setOperationAction(Op, T, Expand);
    if (Subtarget->hasSIMD128())
      for (auto T : {MVT::v16i8, MVT::v8i16, MVT::v4i32})
        setOperationAction(Op, T, Expand);
    if (Subtarget->hasUnimplementedSIMD128())
      setOperationAction(Op, MVT::v2i64, Expand);
  }

  // SIMD-specific configuration
  if (Subtarget->hasSIMD128()) {
    // Support saturating add for i8x16 and i16x8
    for (auto Op : {ISD::SADDSAT, ISD::UADDSAT})
      for (auto T : {MVT::v16i8, MVT::v8i16})
        setOperationAction(Op, T, Legal);

    // Custom lower BUILD_VECTORs to minimize number of replace_lanes
    for (auto T : {MVT::v16i8, MVT::v8i16, MVT::v4i32, MVT::v4f32})
      setOperationAction(ISD::BUILD_VECTOR, T, Custom);
    if (Subtarget->hasUnimplementedSIMD128())
      for (auto T : {MVT::v2i64, MVT::v2f64})
        setOperationAction(ISD::BUILD_VECTOR, T, Custom);

    // We have custom shuffle lowering to expose the shuffle mask
    for (auto T : {MVT::v16i8, MVT::v8i16, MVT::v4i32, MVT::v4f32})
      setOperationAction(ISD::VECTOR_SHUFFLE, T, Custom);
    if (Subtarget->hasUnimplementedSIMD128())
      for (auto T: {MVT::v2i64, MVT::v2f64})
        setOperationAction(ISD::VECTOR_SHUFFLE, T, Custom);

    // Custom lowering since wasm shifts must have a scalar shift amount
    for (auto Op : {ISD::SHL, ISD::SRA, ISD::SRL}) {
      for (auto T : {MVT::v16i8, MVT::v8i16, MVT::v4i32})
        setOperationAction(Op, T, Custom);
      if (Subtarget->hasUnimplementedSIMD128())
        setOperationAction(Op, MVT::v2i64, Custom);
    }

    // Custom lower lane accesses to expand out variable indices
    for (auto Op : {ISD::EXTRACT_VECTOR_ELT, ISD::INSERT_VECTOR_ELT}) {
      for (auto T : {MVT::v16i8, MVT::v8i16, MVT::v4i32, MVT::v4f32})
        setOperationAction(Op, T, Custom);
      if (Subtarget->hasUnimplementedSIMD128())
        for (auto T : {MVT::v2i64, MVT::v2f64})
          setOperationAction(Op, T, Custom);
    }

    // There is no i64x2.mul instruction
    setOperationAction(ISD::MUL, MVT::v2i64, Expand);

    // There are no vector select instructions
    for (auto Op : {ISD::VSELECT, ISD::SELECT_CC, ISD::SELECT}) {
      for (auto T : {MVT::v16i8, MVT::v8i16, MVT::v4i32, MVT::v4f32})
        setOperationAction(Op, T, Expand);
      if (Subtarget->hasUnimplementedSIMD128())
        for (auto T : {MVT::v2i64, MVT::v2f64})
          setOperationAction(Op, T, Expand);
    }

    // Expand integer operations supported for scalars but not SIMD
    for (auto Op : {ISD::CTLZ, ISD::CTTZ, ISD::CTPOP, ISD::SDIV, ISD::UDIV,
                    ISD::SREM, ISD::UREM, ISD::ROTL, ISD::ROTR}) {
      for (auto T : {MVT::v16i8, MVT::v8i16, MVT::v4i32})
        setOperationAction(Op, T, Expand);
      if (Subtarget->hasUnimplementedSIMD128())
        setOperationAction(Op, MVT::v2i64, Expand);
    }

    // Expand float operations supported for scalars but not SIMD
    for (auto Op : {ISD::FCEIL, ISD::FFLOOR, ISD::FTRUNC, ISD::FNEARBYINT,
                    ISD::FCOPYSIGN, ISD::FLOG, ISD::FLOG2, ISD::FLOG10,
                    ISD::FEXP, ISD::FEXP2, ISD::FRINT}) {
      setOperationAction(Op, MVT::v4f32, Expand);
      if (Subtarget->hasUnimplementedSIMD128())
        setOperationAction(Op, MVT::v2f64, Expand);
    }

    // Expand additional SIMD ops that V8 hasn't implemented yet
    if (!Subtarget->hasUnimplementedSIMD128()) {
      setOperationAction(ISD::FSQRT, MVT::v4f32, Expand);
      setOperationAction(ISD::FDIV, MVT::v4f32, Expand);
    }
  }

  // As a special case, these operators use the type to mean the type to
  // sign-extend from.
  setOperationAction(ISD::SIGN_EXTEND_INREG, MVT::i1, Expand);
  if (!Subtarget->hasSignExt()) {
    // Sign extends are legal only when extending a vector extract
    auto Action = Subtarget->hasSIMD128() ? Custom : Expand;
    for (auto T : {MVT::i8, MVT::i16, MVT::i32})
      setOperationAction(ISD::SIGN_EXTEND_INREG, T, Action);
  }
  for (auto T : MVT::integer_fixedlen_vector_valuetypes())
    setOperationAction(ISD::SIGN_EXTEND_INREG, T, Expand);

  // Dynamic stack allocation: use the default expansion.
  setOperationAction(ISD::STACKSAVE, MVT::Other, Expand);
  setOperationAction(ISD::STACKRESTORE, MVT::Other, Expand);
  setOperationAction(ISD::DYNAMIC_STACKALLOC, MVTPtr, Expand);

  setOperationAction(ISD::FrameIndex, MVT::i32, Custom);
  setOperationAction(ISD::CopyToReg, MVT::Other, Custom);

  // Expand these forms; we pattern-match the forms that we can handle in isel.
  for (auto T : {MVT::i32, MVT::i64, MVT::f32, MVT::f64})
    for (auto Op : {ISD::BR_CC, ISD::SELECT_CC})
      setOperationAction(Op, T, Expand);

  // We have custom switch handling.
  setOperationAction(ISD::BR_JT, MVT::Other, Custom);

  // WebAssembly doesn't have:
  //  - Floating-point extending loads.
  //  - Floating-point truncating stores.
  //  - i1 extending loads.
  //  - truncating SIMD stores and most extending loads
  setLoadExtAction(ISD::EXTLOAD, MVT::f64, MVT::f32, Expand);
  setTruncStoreAction(MVT::f64, MVT::f32, Expand);
  for (auto T : MVT::integer_valuetypes())
    for (auto Ext : {ISD::EXTLOAD, ISD::ZEXTLOAD, ISD::SEXTLOAD})
      setLoadExtAction(Ext, T, MVT::i1, Promote);
  if (Subtarget->hasSIMD128()) {
    for (auto T : {MVT::v16i8, MVT::v8i16, MVT::v4i32, MVT::v2i64, MVT::v4f32,
                   MVT::v2f64}) {
      for (auto MemT : MVT::fixedlen_vector_valuetypes()) {
        if (MVT(T) != MemT) {
          setTruncStoreAction(T, MemT, Expand);
          for (auto Ext : {ISD::EXTLOAD, ISD::ZEXTLOAD, ISD::SEXTLOAD})
            setLoadExtAction(Ext, T, MemT, Expand);
        }
      }
    }
    // But some vector extending loads are legal
    if (Subtarget->hasUnimplementedSIMD128()) {
      for (auto Ext : {ISD::EXTLOAD, ISD::SEXTLOAD, ISD::ZEXTLOAD}) {
        setLoadExtAction(Ext, MVT::v8i16, MVT::v8i8, Legal);
        setLoadExtAction(Ext, MVT::v4i32, MVT::v4i16, Legal);
        setLoadExtAction(Ext, MVT::v2i64, MVT::v2i32, Legal);
      }
    }
  }

  // Don't do anything clever with build_pairs
  setOperationAction(ISD::BUILD_PAIR, MVT::i64, Expand);

  // Trap lowers to wasm unreachable
  setOperationAction(ISD::TRAP, MVT::Other, Legal);

  // Exception handling intrinsics
  setOperationAction(ISD::INTRINSIC_WO_CHAIN, MVT::Other, Custom);
  setOperationAction(ISD::INTRINSIC_VOID, MVT::Other, Custom);

  setMaxAtomicSizeInBitsSupported(64);

  // Override the __gnu_f2h_ieee/__gnu_h2f_ieee names so that the f32 name is
  // consistent with the f64 and f128 names.
  setLibcallName(RTLIB::FPEXT_F16_F32, "__extendhfsf2");
  setLibcallName(RTLIB::FPROUND_F32_F16, "__truncsfhf2");

  // Define the emscripten name for return address helper.
  // TODO: when implementing other WASM backends, make this generic or only do
  // this on emscripten depending on what they end up doing.
  setLibcallName(RTLIB::RETURN_ADDRESS, "emscripten_return_address");

  // Always convert switches to br_tables unless there is only one case, which
  // is equivalent to a simple branch. This reduces code size for wasm, and we
  // defer possible jump table optimizations to the VM.
  setMinimumJumpTableEntries(2);
}

TargetLowering::AtomicExpansionKind
WebAssemblyTargetLowering::shouldExpandAtomicRMWInIR(AtomicRMWInst *AI) const {
  // We have wasm instructions for these
  switch (AI->getOperation()) {
  case AtomicRMWInst::Add:
  case AtomicRMWInst::Sub:
  case AtomicRMWInst::And:
  case AtomicRMWInst::Or:
  case AtomicRMWInst::Xor:
  case AtomicRMWInst::Xchg:
    return AtomicExpansionKind::None;
  default:
    break;
  }
  return AtomicExpansionKind::CmpXChg;
}

FastISel *WebAssemblyTargetLowering::createFastISel(
    FunctionLoweringInfo &FuncInfo, const TargetLibraryInfo *LibInfo) const {
  return WebAssembly::createFastISel(FuncInfo, LibInfo);
}

MVT WebAssemblyTargetLowering::getScalarShiftAmountTy(const DataLayout & /*DL*/,
                                                      EVT VT) const {
  unsigned BitWidth = NextPowerOf2(VT.getSizeInBits() - 1);
  if (BitWidth > 1 && BitWidth < 8)
    BitWidth = 8;

  if (BitWidth > 64) {
    // The shift will be lowered to a libcall, and compiler-rt libcalls expect
    // the count to be an i32.
    BitWidth = 32;
    assert(BitWidth >= Log2_32_Ceil(VT.getSizeInBits()) &&
           "32-bit shift counts ought to be enough for anyone");
  }

  MVT Result = MVT::getIntegerVT(BitWidth);
  assert(Result != MVT::INVALID_SIMPLE_VALUE_TYPE &&
         "Unable to represent scalar shift amount type");
  return Result;
}

// Lower an fp-to-int conversion operator from the LLVM opcode, which has an
// undefined result on invalid/overflow, to the WebAssembly opcode, which
// traps on invalid/overflow.
static MachineBasicBlock *LowerFPToInt(MachineInstr &MI, DebugLoc DL,
                                       MachineBasicBlock *BB,
                                       const TargetInstrInfo &TII,
                                       bool IsUnsigned, bool Int64,
                                       bool Float64, unsigned LoweredOpcode) {
  MachineRegisterInfo &MRI = BB->getParent()->getRegInfo();

  Register OutReg = MI.getOperand(0).getReg();
  Register InReg = MI.getOperand(1).getReg();

  unsigned Abs = Float64 ? WebAssembly::ABS_F64 : WebAssembly::ABS_F32;
  unsigned FConst = Float64 ? WebAssembly::CONST_F64 : WebAssembly::CONST_F32;
  unsigned LT = Float64 ? WebAssembly::LT_F64 : WebAssembly::LT_F32;
  unsigned GE = Float64 ? WebAssembly::GE_F64 : WebAssembly::GE_F32;
  unsigned IConst = Int64 ? WebAssembly::CONST_I64 : WebAssembly::CONST_I32;
  unsigned Eqz = WebAssembly::EQZ_I32;
  unsigned And = WebAssembly::AND_I32;
  int64_t Limit = Int64 ? INT64_MIN : INT32_MIN;
  int64_t Substitute = IsUnsigned ? 0 : Limit;
  double CmpVal = IsUnsigned ? -(double)Limit * 2.0 : -(double)Limit;
  auto &Context = BB->getParent()->getFunction().getContext();
  Type *Ty = Float64 ? Type::getDoubleTy(Context) : Type::getFloatTy(Context);

  const BasicBlock *LLVMBB = BB->getBasicBlock();
  MachineFunction *F = BB->getParent();
  MachineBasicBlock *TrueMBB = F->CreateMachineBasicBlock(LLVMBB);
  MachineBasicBlock *FalseMBB = F->CreateMachineBasicBlock(LLVMBB);
  MachineBasicBlock *DoneMBB = F->CreateMachineBasicBlock(LLVMBB);

  MachineFunction::iterator It = ++BB->getIterator();
  F->insert(It, FalseMBB);
  F->insert(It, TrueMBB);
  F->insert(It, DoneMBB);

  // Transfer the remainder of BB and its successor edges to DoneMBB.
  DoneMBB->splice(DoneMBB->begin(), BB, std::next(MI.getIterator()), BB->end());
  DoneMBB->transferSuccessorsAndUpdatePHIs(BB);

  BB->addSuccessor(TrueMBB);
  BB->addSuccessor(FalseMBB);
  TrueMBB->addSuccessor(DoneMBB);
  FalseMBB->addSuccessor(DoneMBB);

  unsigned Tmp0, Tmp1, CmpReg, EqzReg, FalseReg, TrueReg;
  Tmp0 = MRI.createVirtualRegister(MRI.getRegClass(InReg));
  Tmp1 = MRI.createVirtualRegister(MRI.getRegClass(InReg));
  CmpReg = MRI.createVirtualRegister(&WebAssembly::I32RegClass);
  EqzReg = MRI.createVirtualRegister(&WebAssembly::I32RegClass);
  FalseReg = MRI.createVirtualRegister(MRI.getRegClass(OutReg));
  TrueReg = MRI.createVirtualRegister(MRI.getRegClass(OutReg));

  MI.eraseFromParent();
  // For signed numbers, we can do a single comparison to determine whether
  // fabs(x) is within range.
  if (IsUnsigned) {
    Tmp0 = InReg;
  } else {
    BuildMI(BB, DL, TII.get(Abs), Tmp0).addReg(InReg);
  }
  BuildMI(BB, DL, TII.get(FConst), Tmp1)
      .addFPImm(cast<ConstantFP>(ConstantFP::get(Ty, CmpVal)));
  BuildMI(BB, DL, TII.get(LT), CmpReg).addReg(Tmp0).addReg(Tmp1);

  // For unsigned numbers, we have to do a separate comparison with zero.
  if (IsUnsigned) {
    Tmp1 = MRI.createVirtualRegister(MRI.getRegClass(InReg));
    Register SecondCmpReg =
        MRI.createVirtualRegister(&WebAssembly::I32RegClass);
    Register AndReg = MRI.createVirtualRegister(&WebAssembly::I32RegClass);
    BuildMI(BB, DL, TII.get(FConst), Tmp1)
        .addFPImm(cast<ConstantFP>(ConstantFP::get(Ty, 0.0)));
    BuildMI(BB, DL, TII.get(GE), SecondCmpReg).addReg(Tmp0).addReg(Tmp1);
    BuildMI(BB, DL, TII.get(And), AndReg).addReg(CmpReg).addReg(SecondCmpReg);
    CmpReg = AndReg;
  }

  BuildMI(BB, DL, TII.get(Eqz), EqzReg).addReg(CmpReg);

  // Create the CFG diamond to select between doing the conversion or using
  // the substitute value.
  BuildMI(BB, DL, TII.get(WebAssembly::BR_IF)).addMBB(TrueMBB).addReg(EqzReg);
  BuildMI(FalseMBB, DL, TII.get(LoweredOpcode), FalseReg).addReg(InReg);
  BuildMI(FalseMBB, DL, TII.get(WebAssembly::BR)).addMBB(DoneMBB);
  BuildMI(TrueMBB, DL, TII.get(IConst), TrueReg).addImm(Substitute);
  BuildMI(*DoneMBB, DoneMBB->begin(), DL, TII.get(TargetOpcode::PHI), OutReg)
      .addReg(FalseReg)
      .addMBB(FalseMBB)
      .addReg(TrueReg)
      .addMBB(TrueMBB);

  return DoneMBB;
}

MachineBasicBlock *WebAssemblyTargetLowering::EmitInstrWithCustomInserter(
    MachineInstr &MI, MachineBasicBlock *BB) const {
  const TargetInstrInfo &TII = *Subtarget->getInstrInfo();
  DebugLoc DL = MI.getDebugLoc();

  switch (MI.getOpcode()) {
  default:
    llvm_unreachable("Unexpected instr type to insert");
  case WebAssembly::FP_TO_SINT_I32_F32:
    return LowerFPToInt(MI, DL, BB, TII, false, false, false,
                        WebAssembly::I32_TRUNC_S_F32);
  case WebAssembly::FP_TO_UINT_I32_F32:
    return LowerFPToInt(MI, DL, BB, TII, true, false, false,
                        WebAssembly::I32_TRUNC_U_F32);
  case WebAssembly::FP_TO_SINT_I64_F32:
    return LowerFPToInt(MI, DL, BB, TII, false, true, false,
                        WebAssembly::I64_TRUNC_S_F32);
  case WebAssembly::FP_TO_UINT_I64_F32:
    return LowerFPToInt(MI, DL, BB, TII, true, true, false,
                        WebAssembly::I64_TRUNC_U_F32);
  case WebAssembly::FP_TO_SINT_I32_F64:
    return LowerFPToInt(MI, DL, BB, TII, false, false, true,
                        WebAssembly::I32_TRUNC_S_F64);
  case WebAssembly::FP_TO_UINT_I32_F64:
    return LowerFPToInt(MI, DL, BB, TII, true, false, true,
                        WebAssembly::I32_TRUNC_U_F64);
  case WebAssembly::FP_TO_SINT_I64_F64:
    return LowerFPToInt(MI, DL, BB, TII, false, true, true,
                        WebAssembly::I64_TRUNC_S_F64);
  case WebAssembly::FP_TO_UINT_I64_F64:
    return LowerFPToInt(MI, DL, BB, TII, true, true, true,
                        WebAssembly::I64_TRUNC_U_F64);
    llvm_unreachable("Unexpected instruction to emit with custom inserter");
  }
}

const char *
WebAssemblyTargetLowering::getTargetNodeName(unsigned Opcode) const {
  switch (static_cast<WebAssemblyISD::NodeType>(Opcode)) {
  case WebAssemblyISD::FIRST_NUMBER:
    break;
#define HANDLE_NODETYPE(NODE)                                                  \
  case WebAssemblyISD::NODE:                                                   \
    return "WebAssemblyISD::" #NODE;
#include "WebAssemblyISD.def"
#undef HANDLE_NODETYPE
  }
  return nullptr;
}

std::pair<unsigned, const TargetRegisterClass *>
WebAssemblyTargetLowering::getRegForInlineAsmConstraint(
    const TargetRegisterInfo *TRI, StringRef Constraint, MVT VT) const {
  // First, see if this is a constraint that directly corresponds to a
  // WebAssembly register class.
  if (Constraint.size() == 1) {
    switch (Constraint[0]) {
    case 'r':
      assert(VT != MVT::iPTR && "Pointer MVT not expected here");
      if (Subtarget->hasSIMD128() && VT.isVector()) {
        if (VT.getSizeInBits() == 128)
          return std::make_pair(0U, &WebAssembly::V128RegClass);
      }
      if (VT.isInteger() && !VT.isVector()) {
        if (VT.getSizeInBits() <= 32)
          return std::make_pair(0U, &WebAssembly::I32RegClass);
        if (VT.getSizeInBits() <= 64)
          return std::make_pair(0U, &WebAssembly::I64RegClass);
      }
      break;
    default:
      break;
    }
  }

  return TargetLowering::getRegForInlineAsmConstraint(TRI, Constraint, VT);
}

bool WebAssemblyTargetLowering::isCheapToSpeculateCttz() const {
  // Assume ctz is a relatively cheap operation.
  return true;
}

bool WebAssemblyTargetLowering::isCheapToSpeculateCtlz() const {
  // Assume clz is a relatively cheap operation.
  return true;
}

bool WebAssemblyTargetLowering::isLegalAddressingMode(const DataLayout &DL,
                                                      const AddrMode &AM,
                                                      Type *Ty, unsigned AS,
                                                      Instruction *I) const {
  // WebAssembly offsets are added as unsigned without wrapping. The
  // isLegalAddressingMode gives us no way to determine if wrapping could be
  // happening, so we approximate this by accepting only non-negative offsets.
  if (AM.BaseOffs < 0)
    return false;

  // WebAssembly has no scale register operands.
  if (AM.Scale != 0)
    return false;

  // Everything else is legal.
  return true;
}

bool WebAssemblyTargetLowering::allowsMisalignedMemoryAccesses(
    EVT /*VT*/, unsigned /*AddrSpace*/, unsigned /*Align*/,
    MachineMemOperand::Flags /*Flags*/, bool *Fast) const {
  // WebAssembly supports unaligned accesses, though it should be declared
  // with the p2align attribute on loads and stores which do so, and there
  // may be a performance impact. We tell LLVM they're "fast" because
  // for the kinds of things that LLVM uses this for (merging adjacent stores
  // of constants, etc.), WebAssembly implementations will either want the
  // unaligned access or they'll split anyway.
  if (Fast)
    *Fast = true;
  return true;
}

bool WebAssemblyTargetLowering::isIntDivCheap(EVT VT,
                                              AttributeList Attr) const {
  // The current thinking is that wasm engines will perform this optimization,
  // so we can save on code size.
  return true;
}

bool WebAssemblyTargetLowering::isVectorLoadExtDesirable(SDValue ExtVal) const {
  if (!Subtarget->hasUnimplementedSIMD128())
    return false;
  MVT ExtT = ExtVal.getSimpleValueType();
  MVT MemT = cast<LoadSDNode>(ExtVal->getOperand(0))->getSimpleValueType(0);
  return (ExtT == MVT::v8i16 && MemT == MVT::v8i8) ||
         (ExtT == MVT::v4i32 && MemT == MVT::v4i16) ||
         (ExtT == MVT::v2i64 && MemT == MVT::v2i32);
}

EVT WebAssemblyTargetLowering::getSetCCResultType(const DataLayout &DL,
                                                  LLVMContext &C,
                                                  EVT VT) const {
  if (VT.isVector())
    return VT.changeVectorElementTypeToInteger();

  return TargetLowering::getSetCCResultType(DL, C, VT);
}

bool WebAssemblyTargetLowering::getTgtMemIntrinsic(IntrinsicInfo &Info,
                                                   const CallInst &I,
                                                   MachineFunction &MF,
                                                   unsigned Intrinsic) const {
  switch (Intrinsic) {
  case Intrinsic::wasm_atomic_notify:
    Info.opc = ISD::INTRINSIC_W_CHAIN;
    Info.memVT = MVT::i32;
    Info.ptrVal = I.getArgOperand(0);
    Info.offset = 0;
    Info.align = Align(4);
    // atomic.notify instruction does not really load the memory specified with
    // this argument, but MachineMemOperand should either be load or store, so
    // we set this to a load.
    // FIXME Volatile isn't really correct, but currently all LLVM atomic
    // instructions are treated as volatiles in the backend, so we should be
    // consistent. The same applies for wasm_atomic_wait intrinsics too.
    Info.flags = MachineMemOperand::MOVolatile | MachineMemOperand::MOLoad;
    return true;
  case Intrinsic::wasm_atomic_wait_i32:
    Info.opc = ISD::INTRINSIC_W_CHAIN;
    Info.memVT = MVT::i32;
    Info.ptrVal = I.getArgOperand(0);
    Info.offset = 0;
    Info.align = Align(4);
    Info.flags = MachineMemOperand::MOVolatile | MachineMemOperand::MOLoad;
    return true;
  case Intrinsic::wasm_atomic_wait_i64:
    Info.opc = ISD::INTRINSIC_W_CHAIN;
    Info.memVT = MVT::i64;
    Info.ptrVal = I.getArgOperand(0);
    Info.offset = 0;
    Info.align = Align(8);
    Info.flags = MachineMemOperand::MOVolatile | MachineMemOperand::MOLoad;
    return true;
  default:
    return false;
  }
}

//===----------------------------------------------------------------------===//
// WebAssembly Lowering private implementation.
//===----------------------------------------------------------------------===//

//===----------------------------------------------------------------------===//
// Lowering Code
//===----------------------------------------------------------------------===//

static void fail(const SDLoc &DL, SelectionDAG &DAG, const char *Msg) {
  MachineFunction &MF = DAG.getMachineFunction();
  DAG.getContext()->diagnose(
      DiagnosticInfoUnsupported(MF.getFunction(), Msg, DL.getDebugLoc()));
}

// Test whether the given calling convention is supported.
static bool callingConvSupported(CallingConv::ID CallConv) {
  // We currently support the language-independent target-independent
  // conventions. We don't yet have a way to annotate calls with properties like
  // "cold", and we don't have any call-clobbered registers, so these are mostly
  // all handled the same.
  return CallConv == CallingConv::C || CallConv == CallingConv::Fast ||
         CallConv == CallingConv::Cold ||
         CallConv == CallingConv::PreserveMost ||
         CallConv == CallingConv::PreserveAll ||
         CallConv == CallingConv::CXX_FAST_TLS ||
         CallConv == CallingConv::WASM_EmscriptenInvoke;
}

SDValue
WebAssemblyTargetLowering::LowerCall(CallLoweringInfo &CLI,
                                     SmallVectorImpl<SDValue> &InVals) const {
  SelectionDAG &DAG = CLI.DAG;
  SDLoc DL = CLI.DL;
  SDValue Chain = CLI.Chain;
  SDValue Callee = CLI.Callee;
  MachineFunction &MF = DAG.getMachineFunction();
  auto Layout = MF.getDataLayout();

  CallingConv::ID CallConv = CLI.CallConv;
  if (!callingConvSupported(CallConv))
    fail(DL, DAG,
         "WebAssembly doesn't support language-specific or target-specific "
         "calling conventions yet");
  if (CLI.IsPatchPoint)
    fail(DL, DAG, "WebAssembly doesn't support patch point yet");

  if (CLI.IsTailCall) {
    bool MustTail = CLI.CS && CLI.CS.isMustTailCall();
    if (Subtarget->hasTailCall() && !CLI.IsVarArg) {
      // Do not tail call unless caller and callee return types match
      const Function &F = MF.getFunction();
      const TargetMachine &TM = getTargetMachine();
      Type *RetTy = F.getReturnType();
      SmallVector<MVT, 4> CallerRetTys;
      SmallVector<MVT, 4> CalleeRetTys;
      computeLegalValueVTs(F, TM, RetTy, CallerRetTys);
      computeLegalValueVTs(F, TM, CLI.RetTy, CalleeRetTys);
      bool TypesMatch = CallerRetTys.size() == CalleeRetTys.size() &&
                        std::equal(CallerRetTys.begin(), CallerRetTys.end(),
                                   CalleeRetTys.begin());
      if (!TypesMatch) {
        // musttail in this case would be an LLVM IR validation failure
        assert(!MustTail);
        CLI.IsTailCall = false;
      }
    } else {
      CLI.IsTailCall = false;
      if (MustTail) {
        if (CLI.IsVarArg) {
          // The return would pop the argument buffer
          fail(DL, DAG, "WebAssembly does not support varargs tail calls");
        } else {
          fail(DL, DAG, "WebAssembly 'tail-call' feature not enabled");
        }
      }
    }
  }

  SmallVectorImpl<ISD::InputArg> &Ins = CLI.Ins;
  if (Ins.size() > 1)
    fail(DL, DAG, "WebAssembly doesn't support more than 1 returned value yet");

  SmallVectorImpl<ISD::OutputArg> &Outs = CLI.Outs;
  SmallVectorImpl<SDValue> &OutVals = CLI.OutVals;

  // The generic code may have added an sret argument. If we're lowering an
  // invoke function, the ABI requires that the function pointer be the first
  // argument, so we may have to swap the arguments.
  if (CallConv == CallingConv::WASM_EmscriptenInvoke && Outs.size() >= 2 &&
      Outs[0].Flags.isSRet()) {
    std::swap(Outs[0], Outs[1]);
    std::swap(OutVals[0], OutVals[1]);
  }

  unsigned NumFixedArgs = 0;
  for (unsigned I = 0; I < Outs.size(); ++I) {
    const ISD::OutputArg &Out = Outs[I];
    SDValue &OutVal = OutVals[I];
    if (Out.Flags.isNest())
      fail(DL, DAG, "WebAssembly hasn't implemented nest arguments");
    if (Out.Flags.isInAlloca())
      fail(DL, DAG, "WebAssembly hasn't implemented inalloca arguments");
    if (Out.Flags.isInConsecutiveRegs())
      fail(DL, DAG, "WebAssembly hasn't implemented cons regs arguments");
    if (Out.Flags.isInConsecutiveRegsLast())
      fail(DL, DAG, "WebAssembly hasn't implemented cons regs last arguments");
    if (Out.Flags.isByVal() && Out.Flags.getByValSize() != 0) {
      auto &MFI = MF.getFrameInfo();
      int FI = MFI.CreateStackObject(Out.Flags.getByValSize(),
                                     Out.Flags.getByValAlign(),
                                     /*isSS=*/false);
      SDValue SizeNode =
          DAG.getConstant(Out.Flags.getByValSize(), DL, MVT::i32);
      SDValue FINode = DAG.getFrameIndex(FI, getPointerTy(Layout));
      Chain = DAG.getMemcpy(
          Chain, DL, FINode, OutVal, SizeNode, Out.Flags.getByValAlign(),
          /*isVolatile*/ false, /*AlwaysInline=*/false,
          /*isTailCall*/ false, MachinePointerInfo(), MachinePointerInfo());
      OutVal = FINode;
    }
    // Count the number of fixed args *after* legalization.
    NumFixedArgs += Out.IsFixed;
  }

  bool IsVarArg = CLI.IsVarArg;
  auto PtrVT = getPointerTy(Layout);

  // Analyze operands of the call, assigning locations to each operand.
  SmallVector<CCValAssign, 16> ArgLocs;
  CCState CCInfo(CallConv, IsVarArg, MF, ArgLocs, *DAG.getContext());

  if (IsVarArg) {
    // Outgoing non-fixed arguments are placed in a buffer. First
    // compute their offsets and the total amount of buffer space needed.
    for (unsigned I = NumFixedArgs; I < Outs.size(); ++I) {
      const ISD::OutputArg &Out = Outs[I];
      SDValue &Arg = OutVals[I];
      EVT VT = Arg.getValueType();
      assert(VT != MVT::iPTR && "Legalized args should be concrete");
      Type *Ty = VT.getTypeForEVT(*DAG.getContext());
      unsigned Align = std::max(Out.Flags.getOrigAlign(),
                                Layout.getABITypeAlignment(Ty));
      unsigned Offset = CCInfo.AllocateStack(Layout.getTypeAllocSize(Ty),
                                             Align);
      CCInfo.addLoc(CCValAssign::getMem(ArgLocs.size(), VT.getSimpleVT(),
                                        Offset, VT.getSimpleVT(),
                                        CCValAssign::Full));
    }
  }

  unsigned NumBytes = CCInfo.getAlignedCallFrameSize();

  SDValue FINode;
  if (IsVarArg && NumBytes) {
    // For non-fixed arguments, next emit stores to store the argument values
    // to the stack buffer at the offsets computed above.
    int FI = MF.getFrameInfo().CreateStackObject(NumBytes,
                                                 Layout.getStackAlignment(),
                                                 /*isSS=*/false);
    unsigned ValNo = 0;
    SmallVector<SDValue, 8> Chains;
    for (SDValue Arg :
         make_range(OutVals.begin() + NumFixedArgs, OutVals.end())) {
      assert(ArgLocs[ValNo].getValNo() == ValNo &&
             "ArgLocs should remain in order and only hold varargs args");
      unsigned Offset = ArgLocs[ValNo++].getLocMemOffset();
      FINode = DAG.getFrameIndex(FI, getPointerTy(Layout));
      SDValue Add = DAG.getNode(ISD::ADD, DL, PtrVT, FINode,
                                DAG.getConstant(Offset, DL, PtrVT));
      Chains.push_back(
          DAG.getStore(Chain, DL, Arg, Add,
                       MachinePointerInfo::getFixedStack(MF, FI, Offset), 0));
    }
    if (!Chains.empty())
      Chain = DAG.getNode(ISD::TokenFactor, DL, MVT::Other, Chains);
  } else if (IsVarArg) {
    FINode = DAG.getIntPtrConstant(0, DL);
  }

  if (Callee->getOpcode() == ISD::GlobalAddress) {
    // If the callee is a GlobalAddress node (quite common, every direct call
    // is) turn it into a TargetGlobalAddress node so that LowerGlobalAddress
    // doesn't at MO_GOT which is not needed for direct calls.
    GlobalAddressSDNode* GA = cast<GlobalAddressSDNode>(Callee);
    Callee = DAG.getTargetGlobalAddress(GA->getGlobal(), DL,
                                        getPointerTy(DAG.getDataLayout()),
                                        GA->getOffset());
    Callee = DAG.getNode(WebAssemblyISD::Wrapper, DL,
                         getPointerTy(DAG.getDataLayout()), Callee);
  }

  // Compute the operands for the CALLn node.
  SmallVector<SDValue, 16> Ops;
  Ops.push_back(Chain);
  Ops.push_back(Callee);

  // Add all fixed arguments. Note that for non-varargs calls, NumFixedArgs
  // isn't reliable.
  Ops.append(OutVals.begin(),
             IsVarArg ? OutVals.begin() + NumFixedArgs : OutVals.end());
  // Add a pointer to the vararg buffer.
  if (IsVarArg)
    Ops.push_back(FINode);

  SmallVector<EVT, 8> InTys;
  for (const auto &In : Ins) {
    assert(!In.Flags.isByVal() && "byval is not valid for return values");
    assert(!In.Flags.isNest() && "nest is not valid for return values");
    if (In.Flags.isInAlloca())
      fail(DL, DAG, "WebAssembly hasn't implemented inalloca return values");
    if (In.Flags.isInConsecutiveRegs())
      fail(DL, DAG, "WebAssembly hasn't implemented cons regs return values");
    if (In.Flags.isInConsecutiveRegsLast())
      fail(DL, DAG,
           "WebAssembly hasn't implemented cons regs last return values");
    // Ignore In.getOrigAlign() because all our arguments are passed in
    // registers.
    InTys.push_back(In.VT);
  }

  if (CLI.IsTailCall) {
    // ret_calls do not return values to the current frame
    SDVTList NodeTys = DAG.getVTList(MVT::Other, MVT::Glue);
    return DAG.getNode(WebAssemblyISD::RET_CALL, DL, NodeTys, Ops);
  }

  InTys.push_back(MVT::Other);
  SDVTList InTyList = DAG.getVTList(InTys);
  SDValue Res =
      DAG.getNode(Ins.empty() ? WebAssemblyISD::CALL0 : WebAssemblyISD::CALL1,
                  DL, InTyList, Ops);
  if (Ins.empty()) {
    Chain = Res;
  } else {
    InVals.push_back(Res);
    Chain = Res.getValue(1);
  }

  return Chain;
}

bool WebAssemblyTargetLowering::CanLowerReturn(
    CallingConv::ID /*CallConv*/, MachineFunction & /*MF*/, bool /*IsVarArg*/,
    const SmallVectorImpl<ISD::OutputArg> &Outs,
    LLVMContext & /*Context*/) const {
  // WebAssembly can only handle returning tuples with multivalue enabled
  return Subtarget->hasMultivalue() || Outs.size() <= 1;
}

SDValue WebAssemblyTargetLowering::LowerReturn(
    SDValue Chain, CallingConv::ID CallConv, bool /*IsVarArg*/,
    const SmallVectorImpl<ISD::OutputArg> &Outs,
    const SmallVectorImpl<SDValue> &OutVals, const SDLoc &DL,
    SelectionDAG &DAG) const {
  assert((Subtarget->hasMultivalue() || Outs.size() <= 1) &&
         "MVP WebAssembly can only return up to one value");
  if (!callingConvSupported(CallConv))
    fail(DL, DAG, "WebAssembly doesn't support non-C calling conventions");

  SmallVector<SDValue, 4> RetOps(1, Chain);
  RetOps.append(OutVals.begin(), OutVals.end());
  Chain = DAG.getNode(WebAssemblyISD::RETURN, DL, MVT::Other, RetOps);

  // Record the number and types of the return values.
  for (const ISD::OutputArg &Out : Outs) {
    assert(!Out.Flags.isByVal() && "byval is not valid for return values");
    assert(!Out.Flags.isNest() && "nest is not valid for return values");
    assert(Out.IsFixed && "non-fixed return value is not valid");
    if (Out.Flags.isInAlloca())
      fail(DL, DAG, "WebAssembly hasn't implemented inalloca results");
    if (Out.Flags.isInConsecutiveRegs())
      fail(DL, DAG, "WebAssembly hasn't implemented cons regs results");
    if (Out.Flags.isInConsecutiveRegsLast())
      fail(DL, DAG, "WebAssembly hasn't implemented cons regs last results");
  }

  return Chain;
}

SDValue WebAssemblyTargetLowering::LowerFormalArguments(
    SDValue Chain, CallingConv::ID CallConv, bool IsVarArg,
    const SmallVectorImpl<ISD::InputArg> &Ins, const SDLoc &DL,
    SelectionDAG &DAG, SmallVectorImpl<SDValue> &InVals) const {
  if (!callingConvSupported(CallConv))
    fail(DL, DAG, "WebAssembly doesn't support non-C calling conventions");

  MachineFunction &MF = DAG.getMachineFunction();
  auto *MFI = MF.getInfo<WebAssemblyFunctionInfo>();

  // Set up the incoming ARGUMENTS value, which serves to represent the liveness
  // of the incoming values before they're represented by virtual registers.
  MF.getRegInfo().addLiveIn(WebAssembly::ARGUMENTS);

  for (const ISD::InputArg &In : Ins) {
    if (In.Flags.isInAlloca())
      fail(DL, DAG, "WebAssembly hasn't implemented inalloca arguments");
    if (In.Flags.isNest())
      fail(DL, DAG, "WebAssembly hasn't implemented nest arguments");
    if (In.Flags.isInConsecutiveRegs())
      fail(DL, DAG, "WebAssembly hasn't implemented cons regs arguments");
    if (In.Flags.isInConsecutiveRegsLast())
      fail(DL, DAG, "WebAssembly hasn't implemented cons regs last arguments");
    // Ignore In.getOrigAlign() because all our arguments are passed in
    // registers.
    InVals.push_back(In.Used ? DAG.getNode(WebAssemblyISD::ARGUMENT, DL, In.VT,
                                           DAG.getTargetConstant(InVals.size(),
                                                                 DL, MVT::i32))
                             : DAG.getUNDEF(In.VT));

    // Record the number and types of arguments.
    MFI->addParam(In.VT);
  }

  // Varargs are copied into a buffer allocated by the caller, and a pointer to
  // the buffer is passed as an argument.
  if (IsVarArg) {
    MVT PtrVT = getPointerTy(MF.getDataLayout());
    Register VarargVreg =
        MF.getRegInfo().createVirtualRegister(getRegClassFor(PtrVT));
    MFI->setVarargBufferVreg(VarargVreg);
    Chain = DAG.getCopyToReg(
        Chain, DL, VarargVreg,
        DAG.getNode(WebAssemblyISD::ARGUMENT, DL, PtrVT,
                    DAG.getTargetConstant(Ins.size(), DL, MVT::i32)));
    MFI->addParam(PtrVT);
  }

  // Record the number and types of arguments and results.
  SmallVector<MVT, 4> Params;
  SmallVector<MVT, 4> Results;
  computeSignatureVTs(MF.getFunction().getFunctionType(), MF.getFunction(),
                      DAG.getTarget(), Params, Results);
  for (MVT VT : Results)
    MFI->addResult(VT);
  // TODO: Use signatures in WebAssemblyMachineFunctionInfo too and unify
  // the param logic here with ComputeSignatureVTs
  assert(MFI->getParams().size() == Params.size() &&
         std::equal(MFI->getParams().begin(), MFI->getParams().end(),
                    Params.begin()));

  return Chain;
}

void WebAssemblyTargetLowering::ReplaceNodeResults(
    SDNode *N, SmallVectorImpl<SDValue> &Results, SelectionDAG &DAG) const {
  switch (N->getOpcode()) {
  case ISD::SIGN_EXTEND_INREG:
    // Do not add any results, signifying that N should not be custom lowered
    // after all. This happens because simd128 turns on custom lowering for
    // SIGN_EXTEND_INREG, but for non-vector sign extends the result might be an
    // illegal type.
    break;
  default:
    llvm_unreachable(
        "ReplaceNodeResults not implemented for this op for WebAssembly!");
  }
}

//===----------------------------------------------------------------------===//
//  Custom lowering hooks.
//===----------------------------------------------------------------------===//

SDValue WebAssemblyTargetLowering::LowerOperation(SDValue Op,
                                                  SelectionDAG &DAG) const {
  SDLoc DL(Op);
  switch (Op.getOpcode()) {
  default:
    llvm_unreachable("unimplemented operation lowering");
    return SDValue();
  case ISD::FrameIndex:
    return LowerFrameIndex(Op, DAG);
  case ISD::GlobalAddress:
    return LowerGlobalAddress(Op, DAG);
  case ISD::ExternalSymbol:
    return LowerExternalSymbol(Op, DAG);
  case ISD::JumpTable:
    return LowerJumpTable(Op, DAG);
  case ISD::BR_JT:
    return LowerBR_JT(Op, DAG);
  case ISD::VASTART:
    return LowerVASTART(Op, DAG);
  case ISD::BlockAddress:
  case ISD::BRIND:
    fail(DL, DAG, "WebAssembly hasn't implemented computed gotos");
    return SDValue();
  case ISD::RETURNADDR:
    return LowerRETURNADDR(Op, DAG);
  case ISD::FRAMEADDR:
    return LowerFRAMEADDR(Op, DAG);
  case ISD::CopyToReg:
    return LowerCopyToReg(Op, DAG);
  case ISD::EXTRACT_VECTOR_ELT:
  case ISD::INSERT_VECTOR_ELT:
    return LowerAccessVectorElement(Op, DAG);
  case ISD::INTRINSIC_VOID:
  case ISD::INTRINSIC_WO_CHAIN:
  case ISD::INTRINSIC_W_CHAIN:
    return LowerIntrinsic(Op, DAG);
  case ISD::SIGN_EXTEND_INREG:
    return LowerSIGN_EXTEND_INREG(Op, DAG);
  case ISD::BUILD_VECTOR:
    return LowerBUILD_VECTOR(Op, DAG);
  case ISD::VECTOR_SHUFFLE:
    return LowerVECTOR_SHUFFLE(Op, DAG);
  case ISD::SHL:
  case ISD::SRA:
  case ISD::SRL:
    return LowerShift(Op, DAG);
  }
}

SDValue WebAssemblyTargetLowering::LowerCopyToReg(SDValue Op,
                                                  SelectionDAG &DAG) const {
  SDValue Src = Op.getOperand(2);
  if (isa<FrameIndexSDNode>(Src.getNode())) {
    // CopyToReg nodes don't support FrameIndex operands. Other targets select
    // the FI to some LEA-like instruction, but since we don't have that, we
    // need to insert some kind of instruction that can take an FI operand and
    // produces a value usable by CopyToReg (i.e. in a vreg). So insert a dummy
    // local.copy between Op and its FI operand.
    SDValue Chain = Op.getOperand(0);
    SDLoc DL(Op);
    unsigned Reg = cast<RegisterSDNode>(Op.getOperand(1))->getReg();
    EVT VT = Src.getValueType();
    SDValue Copy(DAG.getMachineNode(VT == MVT::i32 ? WebAssembly::COPY_I32
                                                   : WebAssembly::COPY_I64,
                                    DL, VT, Src),
                 0);
    return Op.getNode()->getNumValues() == 1
               ? DAG.getCopyToReg(Chain, DL, Reg, Copy)
               : DAG.getCopyToReg(Chain, DL, Reg, Copy,
                                  Op.getNumOperands() == 4 ? Op.getOperand(3)
                                                           : SDValue());
  }
  return SDValue();
}

SDValue WebAssemblyTargetLowering::LowerFrameIndex(SDValue Op,
                                                   SelectionDAG &DAG) const {
  int FI = cast<FrameIndexSDNode>(Op)->getIndex();
  return DAG.getTargetFrameIndex(FI, Op.getValueType());
}

SDValue WebAssemblyTargetLowering::LowerRETURNADDR(SDValue Op,
                                                   SelectionDAG &DAG) const {
  SDLoc DL(Op);

  if (!Subtarget->getTargetTriple().isOSEmscripten()) {
    fail(DL, DAG,
         "Non-Emscripten WebAssembly hasn't implemented "
         "__builtin_return_address");
    return SDValue();
  }

  if (verifyReturnAddressArgumentIsConstant(Op, DAG))
    return SDValue();

  unsigned Depth = cast<ConstantSDNode>(Op.getOperand(0))->getZExtValue();
  MakeLibCallOptions CallOptions;
  return makeLibCall(DAG, RTLIB::RETURN_ADDRESS, Op.getValueType(),
                     {DAG.getConstant(Depth, DL, MVT::i32)}, CallOptions, DL)
      .first;
}

SDValue WebAssemblyTargetLowering::LowerFRAMEADDR(SDValue Op,
                                                  SelectionDAG &DAG) const {
  // Non-zero depths are not supported by WebAssembly currently. Use the
  // legalizer's default expansion, which is to return 0 (what this function is
  // documented to do).
  if (Op.getConstantOperandVal(0) > 0)
    return SDValue();

  DAG.getMachineFunction().getFrameInfo().setFrameAddressIsTaken(true);
  EVT VT = Op.getValueType();
  Register FP =
      Subtarget->getRegisterInfo()->getFrameRegister(DAG.getMachineFunction());
  return DAG.getCopyFromReg(DAG.getEntryNode(), SDLoc(Op), FP, VT);
}

SDValue WebAssemblyTargetLowering::LowerGlobalAddress(SDValue Op,
                                                      SelectionDAG &DAG) const {
  SDLoc DL(Op);
  const auto *GA = cast<GlobalAddressSDNode>(Op);
  EVT VT = Op.getValueType();
  assert(GA->getTargetFlags() == 0 &&
         "Unexpected target flags on generic GlobalAddressSDNode");
  if (GA->getAddressSpace() != 0)
    fail(DL, DAG, "WebAssembly only expects the 0 address space");

  unsigned OperandFlags = 0;
  if (isPositionIndependent()) {
    const GlobalValue *GV = GA->getGlobal();
    if (getTargetMachine().shouldAssumeDSOLocal(*GV->getParent(), GV)) {
      MachineFunction &MF = DAG.getMachineFunction();
      MVT PtrVT = getPointerTy(MF.getDataLayout());
      const char *BaseName;
      if (GV->getValueType()->isFunctionTy()) {
        BaseName = MF.createExternalSymbolName("__table_base");
        OperandFlags = WebAssemblyII::MO_TABLE_BASE_REL;
      }
      else {
        BaseName = MF.createExternalSymbolName("__memory_base");
        OperandFlags = WebAssemblyII::MO_MEMORY_BASE_REL;
      }
      SDValue BaseAddr =
          DAG.getNode(WebAssemblyISD::Wrapper, DL, PtrVT,
                      DAG.getTargetExternalSymbol(BaseName, PtrVT));

      SDValue SymAddr = DAG.getNode(
          WebAssemblyISD::WrapperPIC, DL, VT,
          DAG.getTargetGlobalAddress(GA->getGlobal(), DL, VT, GA->getOffset(),
                                     OperandFlags));

      return DAG.getNode(ISD::ADD, DL, VT, BaseAddr, SymAddr);
    } else {
      OperandFlags = WebAssemblyII::MO_GOT;
    }
  }

  return DAG.getNode(WebAssemblyISD::Wrapper, DL, VT,
                     DAG.getTargetGlobalAddress(GA->getGlobal(), DL, VT,
                                                GA->getOffset(), OperandFlags));
}

SDValue
WebAssemblyTargetLowering::LowerExternalSymbol(SDValue Op,
                                               SelectionDAG &DAG) const {
  SDLoc DL(Op);
  const auto *ES = cast<ExternalSymbolSDNode>(Op);
  EVT VT = Op.getValueType();
  assert(ES->getTargetFlags() == 0 &&
         "Unexpected target flags on generic ExternalSymbolSDNode");
  return DAG.getNode(WebAssemblyISD::Wrapper, DL, VT,
                     DAG.getTargetExternalSymbol(ES->getSymbol(), VT));
}

SDValue WebAssemblyTargetLowering::LowerJumpTable(SDValue Op,
                                                  SelectionDAG &DAG) const {
  // There's no need for a Wrapper node because we always incorporate a jump
  // table operand into a BR_TABLE instruction, rather than ever
  // materializing it in a register.
  const JumpTableSDNode *JT = cast<JumpTableSDNode>(Op);
  return DAG.getTargetJumpTable(JT->getIndex(), Op.getValueType(),
                                JT->getTargetFlags());
}

SDValue WebAssemblyTargetLowering::LowerBR_JT(SDValue Op,
                                              SelectionDAG &DAG) const {
  SDLoc DL(Op);
  SDValue Chain = Op.getOperand(0);
  const auto *JT = cast<JumpTableSDNode>(Op.getOperand(1));
  SDValue Index = Op.getOperand(2);
  assert(JT->getTargetFlags() == 0 && "WebAssembly doesn't set target flags");

  SmallVector<SDValue, 8> Ops;
  Ops.push_back(Chain);
  Ops.push_back(Index);

  MachineJumpTableInfo *MJTI = DAG.getMachineFunction().getJumpTableInfo();
  const auto &MBBs = MJTI->getJumpTables()[JT->getIndex()].MBBs;

  // Add an operand for each case.
  for (auto MBB : MBBs)
    Ops.push_back(DAG.getBasicBlock(MBB));

  // TODO: For now, we just pick something arbitrary for a default case for now.
  // We really want to sniff out the guard and put in the real default case (and
  // delete the guard).
  Ops.push_back(DAG.getBasicBlock(MBBs[0]));

  return DAG.getNode(WebAssemblyISD::BR_TABLE, DL, MVT::Other, Ops);
}

SDValue WebAssemblyTargetLowering::LowerVASTART(SDValue Op,
                                                SelectionDAG &DAG) const {
  SDLoc DL(Op);
  EVT PtrVT = getPointerTy(DAG.getMachineFunction().getDataLayout());

  auto *MFI = DAG.getMachineFunction().getInfo<WebAssemblyFunctionInfo>();
  const Value *SV = cast<SrcValueSDNode>(Op.getOperand(2))->getValue();

  SDValue ArgN = DAG.getCopyFromReg(DAG.getEntryNode(), DL,
                                    MFI->getVarargBufferVreg(), PtrVT);
  return DAG.getStore(Op.getOperand(0), DL, ArgN, Op.getOperand(1),
                      MachinePointerInfo(SV), 0);
}

SDValue WebAssemblyTargetLowering::LowerIntrinsic(SDValue Op,
                                                  SelectionDAG &DAG) const {
  MachineFunction &MF = DAG.getMachineFunction();
  unsigned IntNo;
  switch (Op.getOpcode()) {
  case ISD::INTRINSIC_VOID:
  case ISD::INTRINSIC_W_CHAIN:
    IntNo = cast<ConstantSDNode>(Op.getOperand(1))->getZExtValue();
    break;
  case ISD::INTRINSIC_WO_CHAIN:
    IntNo = cast<ConstantSDNode>(Op.getOperand(0))->getZExtValue();
    break;
  default:
    llvm_unreachable("Invalid intrinsic");
  }
  SDLoc DL(Op);

  switch (IntNo) {
  default:
    return SDValue(); // Don't custom lower most intrinsics.

  case Intrinsic::wasm_lsda: {
    EVT VT = Op.getValueType();
    const TargetLowering &TLI = DAG.getTargetLoweringInfo();
    MVT PtrVT = TLI.getPointerTy(DAG.getDataLayout());
    auto &Context = MF.getMMI().getContext();
    MCSymbol *S = Context.getOrCreateSymbol(Twine("GCC_except_table") +
                                            Twine(MF.getFunctionNumber()));
    return DAG.getNode(WebAssemblyISD::Wrapper, DL, VT,
                       DAG.getMCSymbol(S, PtrVT));
  }

  case Intrinsic::wasm_throw: {
    // We only support C++ exceptions for now
    int Tag = cast<ConstantSDNode>(Op.getOperand(2).getNode())->getZExtValue();
    if (Tag != CPP_EXCEPTION)
      llvm_unreachable("Invalid tag!");
    const TargetLowering &TLI = DAG.getTargetLoweringInfo();
    MVT PtrVT = TLI.getPointerTy(DAG.getDataLayout());
    const char *SymName = MF.createExternalSymbolName("__cpp_exception");
    SDValue SymNode = DAG.getNode(WebAssemblyISD::Wrapper, DL, PtrVT,
                                  DAG.getTargetExternalSymbol(SymName, PtrVT));
    return DAG.getNode(WebAssemblyISD::THROW, DL,
                       MVT::Other, // outchain type
                       {
                           Op.getOperand(0), // inchain
                           SymNode,          // exception symbol
                           Op.getOperand(3)  // thrown value
                       });
  }
  }
}

SDValue
WebAssemblyTargetLowering::LowerSIGN_EXTEND_INREG(SDValue Op,
                                                  SelectionDAG &DAG) const {
  SDLoc DL(Op);
  // If sign extension operations are disabled, allow sext_inreg only if operand
  // is a vector extract. SIMD does not depend on sign extension operations, but
  // allowing sext_inreg in this context lets us have simple patterns to select
  // extract_lane_s instructions. Expanding sext_inreg everywhere would be
  // simpler in this file, but would necessitate large and brittle patterns to
  // undo the expansion and select extract_lane_s instructions.
  assert(!Subtarget->hasSignExt() && Subtarget->hasSIMD128());
  if (Op.getOperand(0).getOpcode() == ISD::EXTRACT_VECTOR_ELT) {
    const SDValue &Extract = Op.getOperand(0);
    MVT VecT = Extract.getOperand(0).getSimpleValueType();
    MVT ExtractedLaneT = static_cast<VTSDNode *>(Op.getOperand(1).getNode())
                             ->getVT()
                             .getSimpleVT();
    MVT ExtractedVecT =
        MVT::getVectorVT(ExtractedLaneT, 128 / ExtractedLaneT.getSizeInBits());
    if (ExtractedVecT == VecT)
      return Op;
    // Bitcast vector to appropriate type to ensure ISel pattern coverage
    const SDValue &Index = Extract.getOperand(1);
    unsigned IndexVal =
        static_cast<ConstantSDNode *>(Index.getNode())->getZExtValue();
    unsigned Scale =
        ExtractedVecT.getVectorNumElements() / VecT.getVectorNumElements();
    assert(Scale > 1);
    SDValue NewIndex =
        DAG.getConstant(IndexVal * Scale, DL, Index.getValueType());
    SDValue NewExtract = DAG.getNode(
        ISD::EXTRACT_VECTOR_ELT, DL, Extract.getValueType(),
        DAG.getBitcast(ExtractedVecT, Extract.getOperand(0)), NewIndex);
    return DAG.getNode(ISD::SIGN_EXTEND_INREG, DL, Op.getValueType(),
                       NewExtract, Op.getOperand(1));
  }
  // Otherwise expand
  return SDValue();
}

SDValue WebAssemblyTargetLowering::LowerBUILD_VECTOR(SDValue Op,
                                                     SelectionDAG &DAG) const {
  SDLoc DL(Op);
  const EVT VecT = Op.getValueType();
  const EVT LaneT = Op.getOperand(0).getValueType();
  const size_t Lanes = Op.getNumOperands();
  bool CanSwizzle = Subtarget->hasUnimplementedSIMD128() && VecT == MVT::v16i8;

  // BUILD_VECTORs are lowered to the instruction that initializes the highest
  // possible number of lanes at once followed by a sequence of replace_lane
  // instructions to individually initialize any remaining lanes.

  // TODO: Tune this. For example, lanewise swizzling is very expensive, so
  // swizzled lanes should be given greater weight.

  // TODO: Investigate building vectors by shuffling together vectors built by
  // separately specialized means.

  auto IsConstant = [](const SDValue &V) {
    return V.getOpcode() == ISD::Constant || V.getOpcode() == ISD::ConstantFP;
  };

  // Returns the source vector and index vector pair if they exist. Checks for:
  //   (extract_vector_elt
  //     $src,
  //     (sign_extend_inreg (extract_vector_elt $indices, $i))
  //   )
  auto GetSwizzleSrcs = [](size_t I, const SDValue &Lane) {
    auto Bail = std::make_pair(SDValue(), SDValue());
    if (Lane->getOpcode() != ISD::EXTRACT_VECTOR_ELT)
      return Bail;
    const SDValue &SwizzleSrc = Lane->getOperand(0);
    const SDValue &IndexExt = Lane->getOperand(1);
    if (IndexExt->getOpcode() != ISD::SIGN_EXTEND_INREG)
      return Bail;
    const SDValue &Index = IndexExt->getOperand(0);
    if (Index->getOpcode() != ISD::EXTRACT_VECTOR_ELT)
      return Bail;
    const SDValue &SwizzleIndices = Index->getOperand(0);
    if (SwizzleSrc.getValueType() != MVT::v16i8 ||
        SwizzleIndices.getValueType() != MVT::v16i8 ||
        Index->getOperand(1)->getOpcode() != ISD::Constant ||
        Index->getConstantOperandVal(1) != I)
      return Bail;
    return std::make_pair(SwizzleSrc, SwizzleIndices);
  };

  using ValueEntry = std::pair<SDValue, size_t>;
  SmallVector<ValueEntry, 16> SplatValueCounts;

  using SwizzleEntry = std::pair<std::pair<SDValue, SDValue>, size_t>;
  SmallVector<SwizzleEntry, 16> SwizzleCounts;

  auto AddCount = [](auto &Counts, const auto &Val) {
    auto CountIt = std::find_if(Counts.begin(), Counts.end(),
                                [&Val](auto E) { return E.first == Val; });
    if (CountIt == Counts.end()) {
      Counts.emplace_back(Val, 1);
    } else {
      CountIt->second++;
    }
  };

  auto GetMostCommon = [](auto &Counts) {
    auto CommonIt =
        std::max_element(Counts.begin(), Counts.end(),
                         [](auto A, auto B) { return A.second < B.second; });
    assert(CommonIt != Counts.end() && "Unexpected all-undef build_vector");
    return *CommonIt;
  };

  size_t NumConstantLanes = 0;

  // Count eligible lanes for each type of vector creation op
  for (size_t I = 0; I < Lanes; ++I) {
    const SDValue &Lane = Op->getOperand(I);
    if (Lane.isUndef())
      continue;

    AddCount(SplatValueCounts, Lane);

    if (IsConstant(Lane)) {
      NumConstantLanes++;
    } else if (CanSwizzle) {
      auto SwizzleSrcs = GetSwizzleSrcs(I, Lane);
      if (SwizzleSrcs.first)
        AddCount(SwizzleCounts, SwizzleSrcs);
    }
  }

  SDValue SplatValue;
  size_t NumSplatLanes;
  std::tie(SplatValue, NumSplatLanes) = GetMostCommon(SplatValueCounts);

  SDValue SwizzleSrc;
  SDValue SwizzleIndices;
  size_t NumSwizzleLanes = 0;
  if (SwizzleCounts.size())
    std::forward_as_tuple(std::tie(SwizzleSrc, SwizzleIndices),
                          NumSwizzleLanes) = GetMostCommon(SwizzleCounts);

  // Predicate returning true if the lane is properly initialized by the
  // original instruction
  std::function<bool(size_t, const SDValue &)> IsLaneConstructed;
  SDValue Result;
  if (Subtarget->hasUnimplementedSIMD128()) {
    // Prefer swizzles over vector consts over splats
    if (NumSwizzleLanes >= NumSplatLanes &&
        NumSwizzleLanes >= NumConstantLanes) {
      Result = DAG.getNode(WebAssemblyISD::SWIZZLE, DL, VecT, SwizzleSrc,
                           SwizzleIndices);
      auto Swizzled = std::make_pair(SwizzleSrc, SwizzleIndices);
      IsLaneConstructed = [&, Swizzled](size_t I, const SDValue &Lane) {
        return Swizzled == GetSwizzleSrcs(I, Lane);
      };
    } else if (NumConstantLanes >= NumSplatLanes) {
      SmallVector<SDValue, 16> ConstLanes;
      for (const SDValue &Lane : Op->op_values()) {
        if (IsConstant(Lane)) {
          ConstLanes.push_back(Lane);
        } else if (LaneT.isFloatingPoint()) {
          ConstLanes.push_back(DAG.getConstantFP(0, DL, LaneT));
        } else {
          ConstLanes.push_back(DAG.getConstant(0, DL, LaneT));
        }
      }
      Result = DAG.getBuildVector(VecT, DL, ConstLanes);
      IsLaneConstructed = [&](size_t _, const SDValue &Lane) {
        return IsConstant(Lane);
      };
    }
  }
  if (!Result) {
    // Use a splat, but possibly a load_splat
    LoadSDNode *SplattedLoad;
    if (Subtarget->hasUnimplementedSIMD128() &&
        (SplattedLoad = dyn_cast<LoadSDNode>(SplatValue)) &&
        SplattedLoad->getMemoryVT() == VecT.getVectorElementType()) {
      Result = DAG.getNode(WebAssemblyISD::LOAD_SPLAT, DL, VecT, SplatValue);
    } else {
      Result = DAG.getSplatBuildVector(VecT, DL, SplatValue);
    }
    IsLaneConstructed = [&](size_t _, const SDValue &Lane) {
      return Lane == SplatValue;
    };
  }

  // Add replace_lane instructions for any unhandled values
  for (size_t I = 0; I < Lanes; ++I) {
    const SDValue &Lane = Op->getOperand(I);
    if (!Lane.isUndef() && !IsLaneConstructed(I, Lane))
      Result = DAG.getNode(ISD::INSERT_VECTOR_ELT, DL, VecT, Result, Lane,
                           DAG.getConstant(I, DL, MVT::i32));
  }

  return Result;
}

SDValue
WebAssemblyTargetLowering::LowerVECTOR_SHUFFLE(SDValue Op,
                                               SelectionDAG &DAG) const {
  SDLoc DL(Op);
  ArrayRef<int> Mask = cast<ShuffleVectorSDNode>(Op.getNode())->getMask();
  MVT VecType = Op.getOperand(0).getSimpleValueType();
  assert(VecType.is128BitVector() && "Unexpected shuffle vector type");
  size_t LaneBytes = VecType.getVectorElementType().getSizeInBits() / 8;

  // Space for two vector args and sixteen mask indices
  SDValue Ops[18];
  size_t OpIdx = 0;
  Ops[OpIdx++] = Op.getOperand(0);
  Ops[OpIdx++] = Op.getOperand(1);

  // Expand mask indices to byte indices and materialize them as operands
  for (int M : Mask) {
    for (size_t J = 0; J < LaneBytes; ++J) {
      // Lower undefs (represented by -1 in mask) to zero
      uint64_t ByteIndex = M == -1 ? 0 : (uint64_t)M * LaneBytes + J;
      Ops[OpIdx++] = DAG.getConstant(ByteIndex, DL, MVT::i32);
    }
  }

  return DAG.getNode(WebAssemblyISD::SHUFFLE, DL, Op.getValueType(), Ops);
}

SDValue
WebAssemblyTargetLowering::LowerAccessVectorElement(SDValue Op,
                                                    SelectionDAG &DAG) const {
  // Allow constant lane indices, expand variable lane indices
  SDNode *IdxNode = Op.getOperand(Op.getNumOperands() - 1).getNode();
  if (isa<ConstantSDNode>(IdxNode) || IdxNode->isUndef())
    return Op;
  else
    // Perform default expansion
    return SDValue();
}

static SDValue unrollVectorShift(SDValue Op, SelectionDAG &DAG) {
  EVT LaneT = Op.getSimpleValueType().getVectorElementType();
  // 32-bit and 64-bit unrolled shifts will have proper semantics
  if (LaneT.bitsGE(MVT::i32))
    return DAG.UnrollVectorOp(Op.getNode());
  // Otherwise mask the shift value to get proper semantics from 32-bit shift
  SDLoc DL(Op);
  SDValue ShiftVal = Op.getOperand(1);
  uint64_t MaskVal = LaneT.getSizeInBits() - 1;
  SDValue MaskedShiftVal = DAG.getNode(
      ISD::AND,                    // mask opcode
      DL, ShiftVal.getValueType(), // masked value type
      ShiftVal,                    // original shift value operand
      DAG.getConstant(MaskVal, DL, ShiftVal.getValueType()) // mask operand
  );

  return DAG.UnrollVectorOp(
      DAG.getNode(Op.getOpcode(),        // original shift opcode
                  DL, Op.getValueType(), // original return type
                  Op.getOperand(0),      // original vector operand,
                  MaskedShiftVal         // new masked shift value operand
                  )
          .getNode());
}

SDValue WebAssemblyTargetLowering::LowerShift(SDValue Op,
                                              SelectionDAG &DAG) const {
  SDLoc DL(Op);

  // Only manually lower vector shifts
  assert(Op.getSimpleValueType().isVector());

  // Unroll non-splat vector shifts
  BuildVectorSDNode *ShiftVec;
  SDValue SplatVal;
  if (!(ShiftVec = dyn_cast<BuildVectorSDNode>(Op.getOperand(1).getNode())) ||
      !(SplatVal = ShiftVec->getSplatValue()))
    return unrollVectorShift(Op, DAG);

  // All splats except i64x2 const splats are handled by patterns
  auto *SplatConst = dyn_cast<ConstantSDNode>(SplatVal);
  if (!SplatConst || Op.getSimpleValueType() != MVT::v2i64)
    return Op;

  // i64x2 const splats are custom lowered to avoid unnecessary wraps
  unsigned Opcode;
  switch (Op.getOpcode()) {
  case ISD::SHL:
    Opcode = WebAssemblyISD::VEC_SHL;
    break;
  case ISD::SRA:
    Opcode = WebAssemblyISD::VEC_SHR_S;
    break;
  case ISD::SRL:
    Opcode = WebAssemblyISD::VEC_SHR_U;
    break;
  default:
    llvm_unreachable("unexpected opcode");
  }
  APInt Shift = SplatConst->getAPIntValue().zextOrTrunc(32);
  return DAG.getNode(Opcode, DL, Op.getValueType(), Op.getOperand(0),
                     DAG.getConstant(Shift, DL, MVT::i32));
}

//===----------------------------------------------------------------------===//
//                          WebAssembly Optimization Hooks
//===----------------------------------------------------------------------===//