llvm.org GIT mirror llvm / 72f82e4 lib / Target / AMDGPU / SIRegisterInfo.cpp
72f82e4

Tree @72f82e4 (Download .tar.gz)

SIRegisterInfo.cpp @72f82e4raw · 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
//===-- SIRegisterInfo.cpp - SI Register Information ---------------------===//
//
//                     The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
/// \file
/// \brief SI implementation of the TargetRegisterInfo class.
//
//===----------------------------------------------------------------------===//

#include "SIRegisterInfo.h"
#include "SIInstrInfo.h"
#include "SIMachineFunctionInfo.h"
#include "AMDGPUSubtarget.h"
#include "llvm/CodeGen/MachineFrameInfo.h"
#include "llvm/CodeGen/MachineInstrBuilder.h"
#include "llvm/CodeGen/RegisterScavenging.h"
#include "llvm/IR/Function.h"
#include "llvm/IR/LLVMContext.h"

using namespace llvm;

static cl::opt<bool> EnableSpillSGPRToSMEM(
  "amdgpu-spill-sgpr-to-smem",
  cl::desc("Use scalar stores to spill SGPRs if supported by subtarget"),
  cl::init(false));


static bool hasPressureSet(const int *PSets, unsigned PSetID) {
  for (unsigned i = 0; PSets[i] != -1; ++i) {
    if (PSets[i] == (int)PSetID)
      return true;
  }
  return false;
}

void SIRegisterInfo::classifyPressureSet(unsigned PSetID, unsigned Reg,
                                         BitVector &PressureSets) const {
  for (MCRegUnitIterator U(Reg, this); U.isValid(); ++U) {
    const int *PSets = getRegUnitPressureSets(*U);
    if (hasPressureSet(PSets, PSetID)) {
      PressureSets.set(PSetID);
      break;
    }
  }
}

SIRegisterInfo::SIRegisterInfo() : AMDGPURegisterInfo(),
                                   SGPRPressureSets(getNumRegPressureSets()),
                                   VGPRPressureSets(getNumRegPressureSets()) {
  unsigned NumRegPressureSets = getNumRegPressureSets();

  SGPRSetID = NumRegPressureSets;
  VGPRSetID = NumRegPressureSets;

  for (unsigned i = 0; i < NumRegPressureSets; ++i) {
    classifyPressureSet(i, AMDGPU::SGPR0, SGPRPressureSets);
    classifyPressureSet(i, AMDGPU::VGPR0, VGPRPressureSets);
  }

  // Determine the number of reg units for each pressure set.
  std::vector<unsigned> PressureSetRegUnits(NumRegPressureSets, 0);
  for (unsigned i = 0, e = getNumRegUnits(); i != e; ++i) {
    const int *PSets = getRegUnitPressureSets(i);
    for (unsigned j = 0; PSets[j] != -1; ++j) {
      ++PressureSetRegUnits[PSets[j]];
    }
  }

  unsigned VGPRMax = 0, SGPRMax = 0;
  for (unsigned i = 0; i < NumRegPressureSets; ++i) {
    if (isVGPRPressureSet(i) && PressureSetRegUnits[i] > VGPRMax) {
      VGPRSetID = i;
      VGPRMax = PressureSetRegUnits[i];
      continue;
    }
    if (isSGPRPressureSet(i) && PressureSetRegUnits[i] > SGPRMax) {
      SGPRSetID = i;
      SGPRMax = PressureSetRegUnits[i];
    }
  }

  assert(SGPRSetID < NumRegPressureSets &&
         VGPRSetID < NumRegPressureSets);
}

void SIRegisterInfo::reserveRegisterTuples(BitVector &Reserved, unsigned Reg) const {
  MCRegAliasIterator R(Reg, this, true);

  for (; R.isValid(); ++R)
    Reserved.set(*R);
}

unsigned SIRegisterInfo::reservedPrivateSegmentBufferReg(
  const MachineFunction &MF) const {
  unsigned BaseIdx = alignDown(getMaxNumSGPRs(MF), 4) - 4;
  unsigned BaseReg(AMDGPU::SGPR_32RegClass.getRegister(BaseIdx));
  return getMatchingSuperReg(BaseReg, AMDGPU::sub0, &AMDGPU::SReg_128RegClass);
}

unsigned SIRegisterInfo::reservedPrivateSegmentWaveByteOffsetReg(
  const MachineFunction &MF) const {
  unsigned RegCount = getMaxNumSGPRs(MF);
  unsigned Reg;

  // Try to place it in a hole after PrivateSegmentbufferReg.
  if (RegCount & 3) {
    // We cannot put the segment buffer in (Idx - 4) ... (Idx - 1) due to
    // alignment constraints, so we have a hole where can put the wave offset.
    Reg = RegCount - 1;
  } else {
    // We can put the segment buffer in (Idx - 4) ... (Idx - 1) and put the
    // wave offset before it.
    Reg = RegCount - 5;
  }
  return AMDGPU::SGPR_32RegClass.getRegister(Reg);
}

BitVector SIRegisterInfo::getReservedRegs(const MachineFunction &MF) const {
  BitVector Reserved(getNumRegs());
  Reserved.set(AMDGPU::INDIRECT_BASE_ADDR);

  // EXEC_LO and EXEC_HI could be allocated and used as regular register, but
  // this seems likely to result in bugs, so I'm marking them as reserved.
  reserveRegisterTuples(Reserved, AMDGPU::EXEC);
  reserveRegisterTuples(Reserved, AMDGPU::FLAT_SCR);

  // Reserve Trap Handler registers - support is not implemented in Codegen.
  reserveRegisterTuples(Reserved, AMDGPU::TBA);
  reserveRegisterTuples(Reserved, AMDGPU::TMA);
  reserveRegisterTuples(Reserved, AMDGPU::TTMP0_TTMP1);
  reserveRegisterTuples(Reserved, AMDGPU::TTMP2_TTMP3);
  reserveRegisterTuples(Reserved, AMDGPU::TTMP4_TTMP5);
  reserveRegisterTuples(Reserved, AMDGPU::TTMP6_TTMP7);
  reserveRegisterTuples(Reserved, AMDGPU::TTMP8_TTMP9);
  reserveRegisterTuples(Reserved, AMDGPU::TTMP10_TTMP11);

  unsigned MaxNumSGPRs = getMaxNumSGPRs(MF);
  unsigned TotalNumSGPRs = AMDGPU::SGPR_32RegClass.getNumRegs();
  for (unsigned i = MaxNumSGPRs; i < TotalNumSGPRs; ++i) {
    unsigned Reg = AMDGPU::SGPR_32RegClass.getRegister(i);
    reserveRegisterTuples(Reserved, Reg);
  }

  unsigned MaxNumVGPRs = getMaxNumVGPRs(MF);
  unsigned TotalNumVGPRs = AMDGPU::VGPR_32RegClass.getNumRegs();
  for (unsigned i = MaxNumVGPRs; i < TotalNumVGPRs; ++i) {
    unsigned Reg = AMDGPU::VGPR_32RegClass.getRegister(i);
    reserveRegisterTuples(Reserved, Reg);
  }

  const SIMachineFunctionInfo *MFI = MF.getInfo<SIMachineFunctionInfo>();

  unsigned ScratchWaveOffsetReg = MFI->getScratchWaveOffsetReg();
  if (ScratchWaveOffsetReg != AMDGPU::NoRegister) {
    // Reserve 1 SGPR for scratch wave offset in case we need to spill.
    reserveRegisterTuples(Reserved, ScratchWaveOffsetReg);
  }

  unsigned ScratchRSrcReg = MFI->getScratchRSrcReg();
  if (ScratchRSrcReg != AMDGPU::NoRegister) {
    // Reserve 4 SGPRs for the scratch buffer resource descriptor in case we need
    // to spill.
    // TODO: May need to reserve a VGPR if doing LDS spilling.
    reserveRegisterTuples(Reserved, ScratchRSrcReg);
    assert(!isSubRegister(ScratchRSrcReg, ScratchWaveOffsetReg));
  }

  return Reserved;
}

bool SIRegisterInfo::requiresRegisterScavenging(const MachineFunction &Fn) const {
  return Fn.getFrameInfo().hasStackObjects();
}

bool
SIRegisterInfo::requiresFrameIndexScavenging(const MachineFunction &MF) const {
  return MF.getFrameInfo().hasStackObjects();
}

bool SIRegisterInfo::requiresFrameIndexReplacementScavenging(
  const MachineFunction &MF) const {
  // m0 is needed for the scalar store offset. m0 is unallocatable, so we can't
  // create a virtual register for it during frame index elimination, so the
  // scavenger is directly needed.
  return MF.getFrameInfo().hasStackObjects() &&
         MF.getSubtarget<SISubtarget>().hasScalarStores() &&
         MF.getInfo<SIMachineFunctionInfo>()->hasSpilledSGPRs();
}

bool SIRegisterInfo::requiresVirtualBaseRegisters(
  const MachineFunction &) const {
  // There are no special dedicated stack or frame pointers.
  return true;
}

bool SIRegisterInfo::trackLivenessAfterRegAlloc(const MachineFunction &MF) const {
  // This helps catch bugs as verifier errors.
  return true;
}

int64_t SIRegisterInfo::getMUBUFInstrOffset(const MachineInstr *MI) const {
  assert(SIInstrInfo::isMUBUF(*MI));

  int OffIdx = AMDGPU::getNamedOperandIdx(MI->getOpcode(),
                                          AMDGPU::OpName::offset);
  return MI->getOperand(OffIdx).getImm();
}

int64_t SIRegisterInfo::getFrameIndexInstrOffset(const MachineInstr *MI,
                                                 int Idx) const {
  if (!SIInstrInfo::isMUBUF(*MI))
    return 0;

  assert(Idx == AMDGPU::getNamedOperandIdx(MI->getOpcode(),
                                           AMDGPU::OpName::vaddr) &&
         "Should never see frame index on non-address operand");

  return getMUBUFInstrOffset(MI);
}

bool SIRegisterInfo::needsFrameBaseReg(MachineInstr *MI, int64_t Offset) const {
  if (!MI->mayLoadOrStore())
    return false;

  int64_t FullOffset = Offset + getMUBUFInstrOffset(MI);

  return !isUInt<12>(FullOffset);
}

void SIRegisterInfo::materializeFrameBaseRegister(MachineBasicBlock *MBB,
                                                  unsigned BaseReg,
                                                  int FrameIdx,
                                                  int64_t Offset) const {
  MachineBasicBlock::iterator Ins = MBB->begin();
  DebugLoc DL; // Defaults to "unknown"

  if (Ins != MBB->end())
    DL = Ins->getDebugLoc();

  MachineFunction *MF = MBB->getParent();
  const SISubtarget &Subtarget = MF->getSubtarget<SISubtarget>();
  const SIInstrInfo *TII = Subtarget.getInstrInfo();

  if (Offset == 0) {
    BuildMI(*MBB, Ins, DL, TII->get(AMDGPU::V_MOV_B32_e32), BaseReg)
      .addFrameIndex(FrameIdx);
    return;
  }

  MachineRegisterInfo &MRI = MF->getRegInfo();
  unsigned UnusedCarry = MRI.createVirtualRegister(&AMDGPU::SReg_64RegClass);
  unsigned OffsetReg = MRI.createVirtualRegister(&AMDGPU::SReg_32_XM0RegClass);

  unsigned FIReg = MRI.createVirtualRegister(&AMDGPU::VGPR_32RegClass);

  BuildMI(*MBB, Ins, DL, TII->get(AMDGPU::S_MOV_B32), OffsetReg)
    .addImm(Offset);
  BuildMI(*MBB, Ins, DL, TII->get(AMDGPU::V_MOV_B32_e32), FIReg)
    .addFrameIndex(FrameIdx);

  BuildMI(*MBB, Ins, DL, TII->get(AMDGPU::V_ADD_I32_e64), BaseReg)
    .addReg(UnusedCarry, RegState::Define | RegState::Dead)
    .addReg(OffsetReg, RegState::Kill)
    .addReg(FIReg);
}

void SIRegisterInfo::resolveFrameIndex(MachineInstr &MI, unsigned BaseReg,
                                       int64_t Offset) const {

  MachineBasicBlock *MBB = MI.getParent();
  MachineFunction *MF = MBB->getParent();
  const SISubtarget &Subtarget = MF->getSubtarget<SISubtarget>();
  const SIInstrInfo *TII = Subtarget.getInstrInfo();

#ifndef NDEBUG
  // FIXME: Is it possible to be storing a frame index to itself?
  bool SeenFI = false;
  for (const MachineOperand &MO: MI.operands()) {
    if (MO.isFI()) {
      if (SeenFI)
        llvm_unreachable("should not see multiple frame indices");

      SeenFI = true;
    }
  }
#endif

  MachineOperand *FIOp = TII->getNamedOperand(MI, AMDGPU::OpName::vaddr);
  assert(FIOp && FIOp->isFI() && "frame index must be address operand");

  assert(TII->isMUBUF(MI));

  MachineOperand *OffsetOp = TII->getNamedOperand(MI, AMDGPU::OpName::offset);
  int64_t NewOffset = OffsetOp->getImm() + Offset;
  assert(isUInt<12>(NewOffset) && "offset should be legal");

  FIOp->ChangeToRegister(BaseReg, false);
  OffsetOp->setImm(NewOffset);
}

bool SIRegisterInfo::isFrameOffsetLegal(const MachineInstr *MI,
                                        unsigned BaseReg,
                                        int64_t Offset) const {
  if (!SIInstrInfo::isMUBUF(*MI))
    return false;

  int64_t NewOffset = Offset + getMUBUFInstrOffset(MI);

  return isUInt<12>(NewOffset);
}

const TargetRegisterClass *SIRegisterInfo::getPointerRegClass(
  const MachineFunction &MF, unsigned Kind) const {
  // This is inaccurate. It depends on the instruction and address space. The
  // only place where we should hit this is for dealing with frame indexes /
  // private accesses, so this is correct in that case.
  return &AMDGPU::VGPR_32RegClass;
}

static unsigned getNumSubRegsForSpillOp(unsigned Op) {

  switch (Op) {
  case AMDGPU::SI_SPILL_S512_SAVE:
  case AMDGPU::SI_SPILL_S512_RESTORE:
  case AMDGPU::SI_SPILL_V512_SAVE:
  case AMDGPU::SI_SPILL_V512_RESTORE:
    return 16;
  case AMDGPU::SI_SPILL_S256_SAVE:
  case AMDGPU::SI_SPILL_S256_RESTORE:
  case AMDGPU::SI_SPILL_V256_SAVE:
  case AMDGPU::SI_SPILL_V256_RESTORE:
    return 8;
  case AMDGPU::SI_SPILL_S128_SAVE:
  case AMDGPU::SI_SPILL_S128_RESTORE:
  case AMDGPU::SI_SPILL_V128_SAVE:
  case AMDGPU::SI_SPILL_V128_RESTORE:
    return 4;
  case AMDGPU::SI_SPILL_V96_SAVE:
  case AMDGPU::SI_SPILL_V96_RESTORE:
    return 3;
  case AMDGPU::SI_SPILL_S64_SAVE:
  case AMDGPU::SI_SPILL_S64_RESTORE:
  case AMDGPU::SI_SPILL_V64_SAVE:
  case AMDGPU::SI_SPILL_V64_RESTORE:
    return 2;
  case AMDGPU::SI_SPILL_S32_SAVE:
  case AMDGPU::SI_SPILL_S32_RESTORE:
  case AMDGPU::SI_SPILL_V32_SAVE:
  case AMDGPU::SI_SPILL_V32_RESTORE:
    return 1;
  default: llvm_unreachable("Invalid spill opcode");
  }
}

static int getOffsetMUBUFStore(unsigned Opc) {
  switch (Opc) {
  case AMDGPU::BUFFER_STORE_DWORD_OFFEN:
    return AMDGPU::BUFFER_STORE_DWORD_OFFSET;
  case AMDGPU::BUFFER_STORE_BYTE_OFFEN:
    return AMDGPU::BUFFER_STORE_BYTE_OFFSET;
  case AMDGPU::BUFFER_STORE_SHORT_OFFEN:
    return AMDGPU::BUFFER_STORE_SHORT_OFFSET;
  case AMDGPU::BUFFER_STORE_DWORDX2_OFFEN:
    return AMDGPU::BUFFER_STORE_DWORDX2_OFFSET;
  case AMDGPU::BUFFER_STORE_DWORDX4_OFFEN:
    return AMDGPU::BUFFER_STORE_DWORDX4_OFFSET;
  default:
    return -1;
  }
}

static int getOffsetMUBUFLoad(unsigned Opc) {
  switch (Opc) {
  case AMDGPU::BUFFER_LOAD_DWORD_OFFEN:
    return AMDGPU::BUFFER_LOAD_DWORD_OFFSET;
  case AMDGPU::BUFFER_LOAD_UBYTE_OFFEN:
    return AMDGPU::BUFFER_LOAD_UBYTE_OFFSET;
  case AMDGPU::BUFFER_LOAD_SBYTE_OFFEN:
    return AMDGPU::BUFFER_LOAD_SBYTE_OFFSET;
  case AMDGPU::BUFFER_LOAD_USHORT_OFFEN:
    return AMDGPU::BUFFER_LOAD_USHORT_OFFSET;
  case AMDGPU::BUFFER_LOAD_SSHORT_OFFEN:
    return AMDGPU::BUFFER_LOAD_SSHORT_OFFSET;
  case AMDGPU::BUFFER_LOAD_DWORDX2_OFFEN:
    return AMDGPU::BUFFER_LOAD_DWORDX2_OFFSET;
  case AMDGPU::BUFFER_LOAD_DWORDX4_OFFEN:
    return AMDGPU::BUFFER_LOAD_DWORDX4_OFFSET;
  default:
    return -1;
  }
}

// This differs from buildSpillLoadStore by only scavenging a VGPR. It does not
// need to handle the case where an SGPR may need to be spilled while spilling.
static bool buildMUBUFOffsetLoadStore(const SIInstrInfo *TII,
                                      MachineFrameInfo &MFI,
                                      MachineBasicBlock::iterator MI,
                                      int Index,
                                      int64_t Offset) {
  MachineBasicBlock *MBB = MI->getParent();
  const DebugLoc &DL = MI->getDebugLoc();
  bool IsStore = MI->mayStore();

  unsigned Opc = MI->getOpcode();
  int LoadStoreOp = IsStore ?
    getOffsetMUBUFStore(Opc) : getOffsetMUBUFLoad(Opc);
  if (LoadStoreOp == -1)
    return false;

  unsigned Reg = TII->getNamedOperand(*MI, AMDGPU::OpName::vdata)->getReg();

  BuildMI(*MBB, MI, DL, TII->get(LoadStoreOp))
    .addReg(Reg, getDefRegState(!IsStore))
    .addOperand(*TII->getNamedOperand(*MI, AMDGPU::OpName::srsrc))
    .addOperand(*TII->getNamedOperand(*MI, AMDGPU::OpName::soffset))
    .addImm(Offset)
    .addImm(0) // glc
    .addImm(0) // slc
    .addImm(0) // tfe
    .setMemRefs(MI->memoperands_begin(), MI->memoperands_end());
  return true;
}

void SIRegisterInfo::buildSpillLoadStore(MachineBasicBlock::iterator MI,
                                         unsigned LoadStoreOp,
                                         int Index,
                                         unsigned ValueReg,
                                         bool IsKill,
                                         unsigned ScratchRsrcReg,
                                         unsigned ScratchOffsetReg,
                                         int64_t InstOffset,
                                         MachineMemOperand *MMO,
                                         RegScavenger *RS) const {
  MachineBasicBlock *MBB = MI->getParent();
  MachineFunction *MF = MI->getParent()->getParent();
  const SISubtarget &ST =  MF->getSubtarget<SISubtarget>();
  const SIInstrInfo *TII = ST.getInstrInfo();
  const MachineFrameInfo &MFI = MF->getFrameInfo();

  const MCInstrDesc &Desc = TII->get(LoadStoreOp);
  const DebugLoc &DL = MI->getDebugLoc();
  bool IsStore = Desc.mayStore();

  bool RanOutOfSGPRs = false;
  bool Scavenged = false;
  unsigned SOffset = ScratchOffsetReg;

  const TargetRegisterClass *RC = getRegClassForReg(MF->getRegInfo(), ValueReg);
  unsigned NumSubRegs = AMDGPU::getRegBitWidth(RC->getID()) / 32;
  unsigned Size = NumSubRegs * 4;
  int64_t Offset = InstOffset + MFI.getObjectOffset(Index);
  const int64_t OriginalImmOffset = Offset;

  unsigned Align = MFI.getObjectAlignment(Index);
  const MachinePointerInfo &BasePtrInfo = MMO->getPointerInfo();

  if (!isUInt<12>(Offset + Size)) {
    SOffset = AMDGPU::NoRegister;

    // We don't have access to the register scavenger if this function is called
    // during  PEI::scavengeFrameVirtualRegs().
    if (RS)
      SOffset = RS->FindUnusedReg(&AMDGPU::SGPR_32RegClass);

    if (SOffset == AMDGPU::NoRegister) {
      // There are no free SGPRs, and since we are in the process of spilling
      // VGPRs too.  Since we need a VGPR in order to spill SGPRs (this is true
      // on SI/CI and on VI it is true until we implement spilling using scalar
      // stores), we have no way to free up an SGPR.  Our solution here is to
      // add the offset directly to the ScratchOffset register, and then
      // subtract the offset after the spill to return ScratchOffset to it's
      // original value.
      RanOutOfSGPRs = true;
      SOffset = ScratchOffsetReg;
    } else {
      Scavenged = true;
    }

    BuildMI(*MBB, MI, DL, TII->get(AMDGPU::S_ADD_U32), SOffset)
      .addReg(ScratchOffsetReg)
      .addImm(Offset);

    Offset = 0;
  }

  const unsigned EltSize = 4;

  for (unsigned i = 0, e = NumSubRegs; i != e; ++i, Offset += EltSize) {
    unsigned SubReg = NumSubRegs == 1 ?
      ValueReg : getSubReg(ValueReg, getSubRegFromChannel(i));

    unsigned SOffsetRegState = 0;
    unsigned SrcDstRegState = getDefRegState(!IsStore);
    if (i + 1 == e) {
      SOffsetRegState |= getKillRegState(Scavenged);
      // The last implicit use carries the "Kill" flag.
      SrcDstRegState |= getKillRegState(IsKill);
    }

    MachinePointerInfo PInfo = BasePtrInfo.getWithOffset(EltSize * i);
    MachineMemOperand *NewMMO
      = MF->getMachineMemOperand(PInfo, MMO->getFlags(),
                                 EltSize, MinAlign(Align, EltSize * i));

    auto MIB = BuildMI(*MBB, MI, DL, Desc)
      .addReg(SubReg, getDefRegState(!IsStore) | getKillRegState(IsKill))
      .addReg(ScratchRsrcReg)
      .addReg(SOffset, SOffsetRegState)
      .addImm(Offset)
      .addImm(0) // glc
      .addImm(0) // slc
      .addImm(0) // tfe
      .addMemOperand(NewMMO);

    if (NumSubRegs > 1)
      MIB.addReg(ValueReg, RegState::Implicit | SrcDstRegState);
  }

  if (RanOutOfSGPRs) {
    // Subtract the offset we added to the ScratchOffset register.
    BuildMI(*MBB, MI, DL, TII->get(AMDGPU::S_SUB_U32), ScratchOffsetReg)
      .addReg(ScratchOffsetReg)
      .addImm(OriginalImmOffset);
  }
}

static std::pair<unsigned, unsigned> getSpillEltSize(unsigned SuperRegSize,
                                                     bool Store) {
  if (SuperRegSize % 16 == 0) {
    return { 16, Store ? AMDGPU::S_BUFFER_STORE_DWORDX4_SGPR :
                         AMDGPU::S_BUFFER_LOAD_DWORDX4_SGPR };
  }

  if (SuperRegSize % 8 == 0) {
    return { 8, Store ? AMDGPU::S_BUFFER_STORE_DWORDX2_SGPR :
                        AMDGPU::S_BUFFER_LOAD_DWORDX2_SGPR };
  }

  return { 4, Store ? AMDGPU::S_BUFFER_STORE_DWORD_SGPR :
                      AMDGPU::S_BUFFER_LOAD_DWORD_SGPR};
}

void SIRegisterInfo::spillSGPR(MachineBasicBlock::iterator MI,
                               int Index,
                               RegScavenger *RS) const {
  MachineBasicBlock *MBB = MI->getParent();
  MachineFunction *MF = MBB->getParent();
  MachineRegisterInfo &MRI = MF->getRegInfo();
  const SISubtarget &ST =  MF->getSubtarget<SISubtarget>();
  const SIInstrInfo *TII = ST.getInstrInfo();

  unsigned SuperReg = MI->getOperand(0).getReg();
  bool IsKill = MI->getOperand(0).isKill();
  const DebugLoc &DL = MI->getDebugLoc();

  SIMachineFunctionInfo *MFI = MF->getInfo<SIMachineFunctionInfo>();
  MachineFrameInfo &FrameInfo = MF->getFrameInfo();

  bool SpillToSMEM = ST.hasScalarStores() && EnableSpillSGPRToSMEM;

  assert(SuperReg != AMDGPU::M0 && "m0 should never spill");

  unsigned OffsetReg = AMDGPU::M0;
  unsigned M0CopyReg = AMDGPU::NoRegister;

  if (SpillToSMEM) {
    if (RS->isRegUsed(AMDGPU::M0)) {
      M0CopyReg = MRI.createVirtualRegister(&AMDGPU::SReg_32_XM0RegClass);
      BuildMI(*MBB, MI, DL, TII->get(AMDGPU::COPY), M0CopyReg)
        .addReg(AMDGPU::M0);
    }
  }

  unsigned ScalarStoreOp;
  unsigned EltSize = 4;
  const TargetRegisterClass *RC = getPhysRegClass(SuperReg);
  if (SpillToSMEM && isSGPRClass(RC)) {
    // XXX - if private_element_size is larger than 4 it might be useful to be
    // able to spill wider vmem spills.
    std::tie(EltSize, ScalarStoreOp) = getSpillEltSize(RC->getSize(), true);
  }

  ArrayRef<int16_t> SplitParts = getRegSplitParts(RC, EltSize);
  unsigned NumSubRegs = SplitParts.empty() ? 1 : SplitParts.size();

  // SubReg carries the "Kill" flag when SubReg == SuperReg.
  unsigned SubKillState = getKillRegState((NumSubRegs == 1) && IsKill);
  for (unsigned i = 0, e = NumSubRegs; i < e; ++i) {
    unsigned SubReg = NumSubRegs == 1 ?
      SuperReg : getSubReg(SuperReg, SplitParts[i]);

    if (SpillToSMEM) {
      int64_t FrOffset = FrameInfo.getObjectOffset(Index);

      // The allocated memory size is really the wavefront size * the frame
      // index size. The widest register class is 64 bytes, so a 4-byte scratch
      // allocation is enough to spill this in a single stack object.
      //
      // FIXME: Frame size/offsets are computed earlier than this, so the extra
      // space is still unnecessarily allocated.

      unsigned Align = FrameInfo.getObjectAlignment(Index);
      MachinePointerInfo PtrInfo
        = MachinePointerInfo::getFixedStack(*MF, Index, EltSize * i);
      MachineMemOperand *MMO
        = MF->getMachineMemOperand(PtrInfo, MachineMemOperand::MOStore,
                                   EltSize, MinAlign(Align, EltSize * i));

      // SMEM instructions only support a single offset, so increment the wave
      // offset.

      int64_t Offset = (ST.getWavefrontSize() * FrOffset) + (EltSize * i);
      if (Offset != 0) {
        BuildMI(*MBB, MI, DL, TII->get(AMDGPU::S_ADD_U32), OffsetReg)
          .addReg(MFI->getScratchWaveOffsetReg())
          .addImm(Offset);
      } else {
        BuildMI(*MBB, MI, DL, TII->get(AMDGPU::S_MOV_B32), OffsetReg)
          .addReg(MFI->getScratchWaveOffsetReg());
      }

      BuildMI(*MBB, MI, DL, TII->get(ScalarStoreOp))
        .addReg(SubReg, getKillRegState(IsKill)) // sdata
        .addReg(MFI->getScratchRSrcReg())        // sbase
        .addReg(OffsetReg, RegState::Kill)       // soff
        .addImm(0)                               // glc
        .addMemOperand(MMO);

      continue;
    }

    struct SIMachineFunctionInfo::SpilledReg Spill =
      MFI->getSpilledReg(MF, Index, i);
    if (Spill.hasReg()) {
      BuildMI(*MBB, MI, DL,
              TII->getMCOpcodeFromPseudo(AMDGPU::V_WRITELANE_B32),
              Spill.VGPR)
        .addReg(SubReg, getKillRegState(IsKill))
        .addImm(Spill.Lane);

      // FIXME: Since this spills to another register instead of an actual
      // frame index, we should delete the frame index when all references to
      // it are fixed.
    } else {
      // Spill SGPR to a frame index.
      // TODO: Should VI try to spill to VGPR and then spill to SMEM?
      unsigned TmpReg = MRI.createVirtualRegister(&AMDGPU::VGPR_32RegClass);
      // TODO: Should VI try to spill to VGPR and then spill to SMEM?

      MachineInstrBuilder Mov
        = BuildMI(*MBB, MI, DL, TII->get(AMDGPU::V_MOV_B32_e32), TmpReg)
        .addReg(SubReg, SubKillState);


      // There could be undef components of a spilled super register.
      // TODO: Can we detect this and skip the spill?
      if (NumSubRegs > 1) {
        // The last implicit use of the SuperReg carries the "Kill" flag.
        unsigned SuperKillState = 0;
        if (i + 1 == e)
          SuperKillState |= getKillRegState(IsKill);
        Mov.addReg(SuperReg, RegState::Implicit | SuperKillState);
      }

      unsigned Align = FrameInfo.getObjectAlignment(Index);
      MachinePointerInfo PtrInfo
        = MachinePointerInfo::getFixedStack(*MF, Index, EltSize * i);
      MachineMemOperand *MMO
        = MF->getMachineMemOperand(PtrInfo, MachineMemOperand::MOStore,
                                   EltSize, MinAlign(Align, EltSize * i));
      BuildMI(*MBB, MI, DL, TII->get(AMDGPU::SI_SPILL_V32_SAVE))
        .addReg(TmpReg, RegState::Kill)         // src
        .addFrameIndex(Index)                   // vaddr
        .addReg(MFI->getScratchRSrcReg())       // srrsrc
        .addReg(MFI->getScratchWaveOffsetReg()) // soffset
        .addImm(i * 4)                          // offset
        .addMemOperand(MMO);
    }
  }

  if (M0CopyReg != AMDGPU::NoRegister) {
    BuildMI(*MBB, MI, DL, TII->get(AMDGPU::COPY), AMDGPU::M0)
      .addReg(M0CopyReg, RegState::Kill);
  }

  MI->eraseFromParent();
  MFI->addToSpilledSGPRs(NumSubRegs);
}

void SIRegisterInfo::restoreSGPR(MachineBasicBlock::iterator MI,
                                 int Index,
                                 RegScavenger *RS) const {
  MachineFunction *MF = MI->getParent()->getParent();
  MachineRegisterInfo &MRI = MF->getRegInfo();
  MachineBasicBlock *MBB = MI->getParent();
  SIMachineFunctionInfo *MFI = MF->getInfo<SIMachineFunctionInfo>();
  MachineFrameInfo &FrameInfo = MF->getFrameInfo();
  const SISubtarget &ST =  MF->getSubtarget<SISubtarget>();
  const SIInstrInfo *TII = ST.getInstrInfo();
  const DebugLoc &DL = MI->getDebugLoc();

  unsigned SuperReg = MI->getOperand(0).getReg();
  bool SpillToSMEM = ST.hasScalarStores() && EnableSpillSGPRToSMEM;

  assert(SuperReg != AMDGPU::M0 && "m0 should never spill");

  unsigned OffsetReg = AMDGPU::M0;
  unsigned M0CopyReg = AMDGPU::NoRegister;

  if (SpillToSMEM) {
    if (RS->isRegUsed(AMDGPU::M0)) {
      M0CopyReg = MRI.createVirtualRegister(&AMDGPU::SReg_32_XM0RegClass);
      BuildMI(*MBB, MI, DL, TII->get(AMDGPU::COPY), M0CopyReg)
        .addReg(AMDGPU::M0);
    }
  }

  unsigned EltSize = 4;
  unsigned ScalarLoadOp;

  const TargetRegisterClass *RC = getPhysRegClass(SuperReg);
  if (SpillToSMEM && isSGPRClass(RC)) {
    // XXX - if private_element_size is larger than 4 it might be useful to be
    // able to spill wider vmem spills.
    std::tie(EltSize, ScalarLoadOp) = getSpillEltSize(RC->getSize(), false);
  }

  ArrayRef<int16_t> SplitParts = getRegSplitParts(RC, EltSize);
  unsigned NumSubRegs = SplitParts.empty() ? 1 : SplitParts.size();

  // SubReg carries the "Kill" flag when SubReg == SuperReg.
  int64_t FrOffset = FrameInfo.getObjectOffset(Index);

  for (unsigned i = 0, e = NumSubRegs; i < e; ++i) {
    unsigned SubReg = NumSubRegs == 1 ?
      SuperReg : getSubReg(SuperReg, SplitParts[i]);

    if (SpillToSMEM) {
      // FIXME: Size may be > 4 but extra bytes wasted.
      unsigned Align = FrameInfo.getObjectAlignment(Index);
      MachinePointerInfo PtrInfo
        = MachinePointerInfo::getFixedStack(*MF, Index, EltSize * i);
      MachineMemOperand *MMO
        = MF->getMachineMemOperand(PtrInfo, MachineMemOperand::MOLoad,
                                   EltSize, MinAlign(Align, EltSize * i));

      // Add i * 4 offset
      int64_t Offset = (ST.getWavefrontSize() * FrOffset) + (EltSize * i);
      if (Offset != 0) {
        BuildMI(*MBB, MI, DL, TII->get(AMDGPU::S_ADD_U32), OffsetReg)
          .addReg(MFI->getScratchWaveOffsetReg())
          .addImm(Offset);
      } else {
        BuildMI(*MBB, MI, DL, TII->get(AMDGPU::S_MOV_B32), OffsetReg)
          .addReg(MFI->getScratchWaveOffsetReg());
      }

      auto MIB =
        BuildMI(*MBB, MI, DL, TII->get(ScalarLoadOp), SubReg)
        .addReg(MFI->getScratchRSrcReg()) // sbase
        .addReg(OffsetReg, RegState::Kill)                // soff
        .addImm(0)                        // glc
        .addMemOperand(MMO);

      if (NumSubRegs > 1)
        MIB.addReg(SuperReg, RegState::ImplicitDefine);

      continue;
    }

    SIMachineFunctionInfo::SpilledReg Spill
      = MFI->getSpilledReg(MF, Index, i);

    if (Spill.hasReg()) {
      auto MIB =
        BuildMI(*MBB, MI, DL, TII->getMCOpcodeFromPseudo(AMDGPU::V_READLANE_B32),
                SubReg)
        .addReg(Spill.VGPR)
        .addImm(Spill.Lane);

      if (NumSubRegs > 1)
        MIB.addReg(SuperReg, RegState::ImplicitDefine);
    } else {
      // Restore SGPR from a stack slot.
      // FIXME: We should use S_LOAD_DWORD here for VI.
      unsigned TmpReg = MRI.createVirtualRegister(&AMDGPU::VGPR_32RegClass);
      unsigned Align = FrameInfo.getObjectAlignment(Index);

      MachinePointerInfo PtrInfo
        = MachinePointerInfo::getFixedStack(*MF, Index, EltSize * i);

      MachineMemOperand *MMO = MF->getMachineMemOperand(PtrInfo,
        MachineMemOperand::MOLoad, EltSize,
        MinAlign(Align, EltSize * i));

      BuildMI(*MBB, MI, DL, TII->get(AMDGPU::SI_SPILL_V32_RESTORE), TmpReg)
        .addFrameIndex(Index)                   // vaddr
        .addReg(MFI->getScratchRSrcReg())       // srsrc
        .addReg(MFI->getScratchWaveOffsetReg()) // soffset
        .addImm(i * 4)                          // offset
        .addMemOperand(MMO);

      auto MIB =
        BuildMI(*MBB, MI, DL, TII->get(AMDGPU::V_READFIRSTLANE_B32), SubReg)
        .addReg(TmpReg, RegState::Kill);

      if (NumSubRegs > 1)
        MIB.addReg(MI->getOperand(0).getReg(), RegState::ImplicitDefine);
    }
  }

  if (M0CopyReg != AMDGPU::NoRegister) {
    BuildMI(*MBB, MI, DL, TII->get(AMDGPU::COPY), AMDGPU::M0)
      .addReg(M0CopyReg, RegState::Kill);
  }

  MI->eraseFromParent();
}

void SIRegisterInfo::eliminateFrameIndex(MachineBasicBlock::iterator MI,
                                        int SPAdj, unsigned FIOperandNum,
                                        RegScavenger *RS) const {
  MachineFunction *MF = MI->getParent()->getParent();
  MachineRegisterInfo &MRI = MF->getRegInfo();
  MachineBasicBlock *MBB = MI->getParent();
  SIMachineFunctionInfo *MFI = MF->getInfo<SIMachineFunctionInfo>();
  MachineFrameInfo &FrameInfo = MF->getFrameInfo();
  const SISubtarget &ST =  MF->getSubtarget<SISubtarget>();
  const SIInstrInfo *TII = ST.getInstrInfo();
  DebugLoc DL = MI->getDebugLoc();

  MachineOperand &FIOp = MI->getOperand(FIOperandNum);
  int Index = MI->getOperand(FIOperandNum).getIndex();

  switch (MI->getOpcode()) {
    // SGPR register spill
    case AMDGPU::SI_SPILL_S512_SAVE:
    case AMDGPU::SI_SPILL_S256_SAVE:
    case AMDGPU::SI_SPILL_S128_SAVE:
    case AMDGPU::SI_SPILL_S64_SAVE:
    case AMDGPU::SI_SPILL_S32_SAVE: {
      spillSGPR(MI, Index, RS);
      break;
    }

    // SGPR register restore
    case AMDGPU::SI_SPILL_S512_RESTORE:
    case AMDGPU::SI_SPILL_S256_RESTORE:
    case AMDGPU::SI_SPILL_S128_RESTORE:
    case AMDGPU::SI_SPILL_S64_RESTORE:
    case AMDGPU::SI_SPILL_S32_RESTORE: {
      restoreSGPR(MI, Index, RS);
      break;
    }

    // VGPR register spill
    case AMDGPU::SI_SPILL_V512_SAVE:
    case AMDGPU::SI_SPILL_V256_SAVE:
    case AMDGPU::SI_SPILL_V128_SAVE:
    case AMDGPU::SI_SPILL_V96_SAVE:
    case AMDGPU::SI_SPILL_V64_SAVE:
    case AMDGPU::SI_SPILL_V32_SAVE: {
      const MachineOperand *VData = TII->getNamedOperand(*MI,
                                                         AMDGPU::OpName::vdata);
      buildSpillLoadStore(MI, AMDGPU::BUFFER_STORE_DWORD_OFFSET,
            Index,
            VData->getReg(), VData->isKill(),
            TII->getNamedOperand(*MI, AMDGPU::OpName::srsrc)->getReg(),
            TII->getNamedOperand(*MI, AMDGPU::OpName::soffset)->getReg(),
            TII->getNamedOperand(*MI, AMDGPU::OpName::offset)->getImm(),
            *MI->memoperands_begin(),
            RS);
      MFI->addToSpilledVGPRs(getNumSubRegsForSpillOp(MI->getOpcode()));
      MI->eraseFromParent();
      break;
    }
    case AMDGPU::SI_SPILL_V32_RESTORE:
    case AMDGPU::SI_SPILL_V64_RESTORE:
    case AMDGPU::SI_SPILL_V96_RESTORE:
    case AMDGPU::SI_SPILL_V128_RESTORE:
    case AMDGPU::SI_SPILL_V256_RESTORE:
    case AMDGPU::SI_SPILL_V512_RESTORE: {
      const MachineOperand *VData = TII->getNamedOperand(*MI,
                                                         AMDGPU::OpName::vdata);

      buildSpillLoadStore(MI, AMDGPU::BUFFER_LOAD_DWORD_OFFSET,
            Index,
            VData->getReg(), VData->isKill(),
            TII->getNamedOperand(*MI, AMDGPU::OpName::srsrc)->getReg(),
            TII->getNamedOperand(*MI, AMDGPU::OpName::soffset)->getReg(),
            TII->getNamedOperand(*MI, AMDGPU::OpName::offset)->getImm(),
            *MI->memoperands_begin(),
            RS);
      MI->eraseFromParent();
      break;
    }

    default: {
      if (TII->isMUBUF(*MI)) {
        // Disable offen so we don't need a 0 vgpr base.
        assert(static_cast<int>(FIOperandNum) ==
               AMDGPU::getNamedOperandIdx(MI->getOpcode(),
                                          AMDGPU::OpName::vaddr));

        int64_t Offset = FrameInfo.getObjectOffset(Index);
        int64_t OldImm
          = TII->getNamedOperand(*MI, AMDGPU::OpName::offset)->getImm();
        int64_t NewOffset = OldImm + Offset;

        if (isUInt<12>(NewOffset) &&
            buildMUBUFOffsetLoadStore(TII, FrameInfo, MI, Index, NewOffset)) {
          MI->eraseFromParent();
          break;
        }
      }

      int64_t Offset = FrameInfo.getObjectOffset(Index);
      FIOp.ChangeToImmediate(Offset);
      if (!TII->isImmOperandLegal(*MI, FIOperandNum, FIOp)) {
        unsigned TmpReg = MRI.createVirtualRegister(&AMDGPU::VGPR_32RegClass);
        BuildMI(*MBB, MI, MI->getDebugLoc(),
                TII->get(AMDGPU::V_MOV_B32_e32), TmpReg)
                .addImm(Offset);
        FIOp.ChangeToRegister(TmpReg, false, false, true);
      }
    }
  }
}

// FIXME: This is very slow. It might be worth creating a map from physreg to
// register class.
const TargetRegisterClass *SIRegisterInfo::getPhysRegClass(unsigned Reg) const {
  assert(!TargetRegisterInfo::isVirtualRegister(Reg));

  static const TargetRegisterClass *const BaseClasses[] = {
    &AMDGPU::VGPR_32RegClass,
    &AMDGPU::SReg_32RegClass,
    &AMDGPU::VReg_64RegClass,
    &AMDGPU::SReg_64RegClass,
    &AMDGPU::VReg_96RegClass,
    &AMDGPU::VReg_128RegClass,
    &AMDGPU::SReg_128RegClass,
    &AMDGPU::VReg_256RegClass,
    &AMDGPU::SReg_256RegClass,
    &AMDGPU::VReg_512RegClass,
    &AMDGPU::SReg_512RegClass,
    &AMDGPU::SCC_CLASSRegClass,
  };

  for (const TargetRegisterClass *BaseClass : BaseClasses) {
    if (BaseClass->contains(Reg)) {
      return BaseClass;
    }
  }
  return nullptr;
}

// TODO: It might be helpful to have some target specific flags in
// TargetRegisterClass to mark which classes are VGPRs to make this trivial.
bool SIRegisterInfo::hasVGPRs(const TargetRegisterClass *RC) const {
  switch (RC->getSize()) {
  case 0: return false;
  case 1: return false;
  case 4:
    return getCommonSubClass(&AMDGPU::VGPR_32RegClass, RC) != nullptr;
  case 8:
    return getCommonSubClass(&AMDGPU::VReg_64RegClass, RC) != nullptr;
  case 12:
    return getCommonSubClass(&AMDGPU::VReg_96RegClass, RC) != nullptr;
  case 16:
    return getCommonSubClass(&AMDGPU::VReg_128RegClass, RC) != nullptr;
  case 32:
    return getCommonSubClass(&AMDGPU::VReg_256RegClass, RC) != nullptr;
  case 64:
    return getCommonSubClass(&AMDGPU::VReg_512RegClass, RC) != nullptr;
  default:
    llvm_unreachable("Invalid register class size");
  }
}

const TargetRegisterClass *SIRegisterInfo::getEquivalentVGPRClass(
                                         const TargetRegisterClass *SRC) const {
  switch (SRC->getSize()) {
  case 4:
    return &AMDGPU::VGPR_32RegClass;
  case 8:
    return &AMDGPU::VReg_64RegClass;
  case 12:
    return &AMDGPU::VReg_96RegClass;
  case 16:
    return &AMDGPU::VReg_128RegClass;
  case 32:
    return &AMDGPU::VReg_256RegClass;
  case 64:
    return &AMDGPU::VReg_512RegClass;
  default:
    llvm_unreachable("Invalid register class size");
  }
}

const TargetRegisterClass *SIRegisterInfo::getEquivalentSGPRClass(
                                         const TargetRegisterClass *VRC) const {
  switch (VRC->getSize()) {
  case 4:
    return &AMDGPU::SGPR_32RegClass;
  case 8:
    return &AMDGPU::SReg_64RegClass;
  case 16:
    return &AMDGPU::SReg_128RegClass;
  case 32:
    return &AMDGPU::SReg_256RegClass;
  case 64:
    return &AMDGPU::SReg_512RegClass;
  default:
    llvm_unreachable("Invalid register class size");
  }
}

const TargetRegisterClass *SIRegisterInfo::getSubRegClass(
                         const TargetRegisterClass *RC, unsigned SubIdx) const {
  if (SubIdx == AMDGPU::NoSubRegister)
    return RC;

  // We can assume that each lane corresponds to one 32-bit register.
  LaneBitmask::Type Mask = getSubRegIndexLaneMask(SubIdx).getAsInteger();
  unsigned Count = countPopulation(Mask);
  if (isSGPRClass(RC)) {
    switch (Count) {
    case 1:
      return &AMDGPU::SGPR_32RegClass;
    case 2:
      return &AMDGPU::SReg_64RegClass;
    case 4:
      return &AMDGPU::SReg_128RegClass;
    case 8:
      return &AMDGPU::SReg_256RegClass;
    case 16: /* fall-through */
    default:
      llvm_unreachable("Invalid sub-register class size");
    }
  } else {
    switch (Count) {
    case 1:
      return &AMDGPU::VGPR_32RegClass;
    case 2:
      return &AMDGPU::VReg_64RegClass;
    case 3:
      return &AMDGPU::VReg_96RegClass;
    case 4:
      return &AMDGPU::VReg_128RegClass;
    case 8:
      return &AMDGPU::VReg_256RegClass;
    case 16: /* fall-through */
    default:
      llvm_unreachable("Invalid sub-register class size");
    }
  }
}

bool SIRegisterInfo::shouldRewriteCopySrc(
  const TargetRegisterClass *DefRC,
  unsigned DefSubReg,
  const TargetRegisterClass *SrcRC,
  unsigned SrcSubReg) const {
  // We want to prefer the smallest register class possible, so we don't want to
  // stop and rewrite on anything that looks like a subregister
  // extract. Operations mostly don't care about the super register class, so we
  // only want to stop on the most basic of copies between the same register
  // class.
  //
  // e.g. if we have something like
  // vreg0 = ...
  // vreg1 = ...
  // vreg2 = REG_SEQUENCE vreg0, sub0, vreg1, sub1, vreg2, sub2
  // vreg3 = COPY vreg2, sub0
  //
  // We want to look through the COPY to find:
  //  => vreg3 = COPY vreg0

  // Plain copy.
  return getCommonSubClass(DefRC, SrcRC) != nullptr;
}

// FIXME: Most of these are flexible with HSA and we don't need to reserve them
// as input registers if unused. Whether the dispatch ptr is necessary should be
// easy to detect from used intrinsics. Scratch setup is harder to know.
unsigned SIRegisterInfo::getPreloadedValue(const MachineFunction &MF,
                                           enum PreloadedValue Value) const {

  const SIMachineFunctionInfo *MFI = MF.getInfo<SIMachineFunctionInfo>();
  const SISubtarget &ST = MF.getSubtarget<SISubtarget>();
  (void)ST;
  switch (Value) {
  case SIRegisterInfo::WORKGROUP_ID_X:
    assert(MFI->hasWorkGroupIDX());
    return MFI->WorkGroupIDXSystemSGPR;
  case SIRegisterInfo::WORKGROUP_ID_Y:
    assert(MFI->hasWorkGroupIDY());
    return MFI->WorkGroupIDYSystemSGPR;
  case SIRegisterInfo::WORKGROUP_ID_Z:
    assert(MFI->hasWorkGroupIDZ());
    return MFI->WorkGroupIDZSystemSGPR;
  case SIRegisterInfo::PRIVATE_SEGMENT_WAVE_BYTE_OFFSET:
    return MFI->PrivateSegmentWaveByteOffsetSystemSGPR;
  case SIRegisterInfo::PRIVATE_SEGMENT_BUFFER:
    if (ST.isAmdCodeObjectV2(MF)) {
      assert(MFI->hasPrivateSegmentBuffer());
      return MFI->PrivateSegmentBufferUserSGPR;
    }
    assert(MFI->hasPrivateMemoryInputPtr());
    return MFI->PrivateMemoryPtrUserSGPR;
  case SIRegisterInfo::KERNARG_SEGMENT_PTR:
    assert(MFI->hasKernargSegmentPtr());
    return MFI->KernargSegmentPtrUserSGPR;
  case SIRegisterInfo::DISPATCH_ID:
    assert(MFI->hasDispatchID());
    return MFI->DispatchIDUserSGPR;
  case SIRegisterInfo::FLAT_SCRATCH_INIT:
    assert(MFI->hasFlatScratchInit());
    return MFI->FlatScratchInitUserSGPR;
  case SIRegisterInfo::DISPATCH_PTR:
    assert(MFI->hasDispatchPtr());
    return MFI->DispatchPtrUserSGPR;
  case SIRegisterInfo::QUEUE_PTR:
    assert(MFI->hasQueuePtr());
    return MFI->QueuePtrUserSGPR;
  case SIRegisterInfo::WORKITEM_ID_X:
    assert(MFI->hasWorkItemIDX());
    return AMDGPU::VGPR0;
  case SIRegisterInfo::WORKITEM_ID_Y:
    assert(MFI->hasWorkItemIDY());
    return AMDGPU::VGPR1;
  case SIRegisterInfo::WORKITEM_ID_Z:
    assert(MFI->hasWorkItemIDZ());
    return AMDGPU::VGPR2;
  }
  llvm_unreachable("unexpected preloaded value type");
}

/// \brief Returns a register that is not used at any point in the function.
///        If all registers are used, then this function will return
//         AMDGPU::NoRegister.
unsigned
SIRegisterInfo::findUnusedRegister(const MachineRegisterInfo &MRI,
                                   const TargetRegisterClass *RC,
                                   const MachineFunction &MF) const {

  for (unsigned Reg : *RC)
    if (MRI.isAllocatable(Reg) && !MRI.isPhysRegUsed(Reg))
      return Reg;
  return AMDGPU::NoRegister;
}

unsigned SIRegisterInfo::getTotalNumSGPRs(const SISubtarget &ST) const {
  if (ST.getGeneration() >= AMDGPUSubtarget::VOLCANIC_ISLANDS)
    return 800;
  return 512;
}

unsigned SIRegisterInfo::getNumAddressableSGPRs(const SISubtarget &ST) const {
  if (ST.getGeneration() >= AMDGPUSubtarget::VOLCANIC_ISLANDS)
    return 102;
  return 104;
}

unsigned SIRegisterInfo::getNumReservedSGPRs(const SISubtarget &ST,
                                             const SIMachineFunctionInfo &MFI) const {
  if (MFI.hasFlatScratchInit()) {
    if (ST.getGeneration() >= AMDGPUSubtarget::VOLCANIC_ISLANDS)
      return 6; // FLAT_SCRATCH, XNACK, VCC (in that order)

    if (ST.getGeneration() == AMDGPUSubtarget::SEA_ISLANDS)
      return 4; // FLAT_SCRATCH, VCC (in that order)
  }

  if (ST.isXNACKEnabled())
    return 4; // XNACK, VCC (in that order)

  return 2; // VCC.
}

unsigned SIRegisterInfo::getMinNumSGPRs(const SISubtarget &ST,
                                        unsigned WavesPerEU) const {
  if (ST.getGeneration() >= AMDGPUSubtarget::VOLCANIC_ISLANDS) {
    switch (WavesPerEU) {
      case 0:  return 0;
      case 10: return 0;
      case 9:  return 0;
      case 8:  return 81;
      default: return 97;
    }
  } else {
    switch (WavesPerEU) {
      case 0:  return 0;
      case 10: return 0;
      case 9:  return 49;
      case 8:  return 57;
      case 7:  return 65;
      case 6:  return 73;
      case 5:  return 81;
      default: return 97;
    }
  }
}

unsigned SIRegisterInfo::getMaxNumSGPRs(const SISubtarget &ST,
                                        unsigned WavesPerEU,
                                        bool Addressable) const {
  if (ST.getGeneration() >= AMDGPUSubtarget::VOLCANIC_ISLANDS) {
    switch (WavesPerEU) {
      case 0:  return 80;
      case 10: return 80;
      case 9:  return 80;
      case 8:  return 96;
      default: return Addressable ? getNumAddressableSGPRs(ST) : 112;
    }
  } else {
    switch (WavesPerEU) {
      case 0:  return 48;
      case 10: return 48;
      case 9:  return 56;
      case 8:  return 64;
      case 7:  return 72;
      case 6:  return 80;
      case 5:  return 96;
      default: return getNumAddressableSGPRs(ST);
    }
  }
}

unsigned SIRegisterInfo::getMaxNumSGPRs(const MachineFunction &MF) const {
  const Function &F = *MF.getFunction();

  const SISubtarget &ST = MF.getSubtarget<SISubtarget>();
  const SIMachineFunctionInfo &MFI = *MF.getInfo<SIMachineFunctionInfo>();

  // Compute maximum number of SGPRs function can use using default/requested
  // minimum number of waves per execution unit.
  std::pair<unsigned, unsigned> WavesPerEU = MFI.getWavesPerEU();
  unsigned MaxNumSGPRs = getMaxNumSGPRs(ST, WavesPerEU.first, false);
  unsigned MaxNumAddressableSGPRs = getMaxNumSGPRs(ST, WavesPerEU.first, true);

  // Check if maximum number of SGPRs was explicitly requested using
  // "amdgpu-num-sgpr" attribute.
  if (F.hasFnAttribute("amdgpu-num-sgpr")) {
    unsigned Requested = AMDGPU::getIntegerAttribute(
      F, "amdgpu-num-sgpr", MaxNumSGPRs);

    // Make sure requested value does not violate subtarget's specifications.
    if (Requested && (Requested <= getNumReservedSGPRs(ST, MFI)))
      Requested = 0;

    // If more SGPRs are required to support the input user/system SGPRs,
    // increase to accommodate them.
    //
    // FIXME: This really ends up using the requested number of SGPRs + number
    // of reserved special registers in total. Theoretically you could re-use
    // the last input registers for these special registers, but this would
    // require a lot of complexity to deal with the weird aliasing.
    unsigned NumInputSGPRs = MFI.getNumPreloadedSGPRs();
    if (Requested && Requested < NumInputSGPRs)
      Requested = NumInputSGPRs;

    // Make sure requested value is compatible with values implied by
    // default/requested minimum/maximum number of waves per execution unit.
    if (Requested && Requested > getMaxNumSGPRs(ST, WavesPerEU.first, false))
      Requested = 0;
    if (WavesPerEU.second &&
        Requested && Requested < getMinNumSGPRs(ST, WavesPerEU.second))
      Requested = 0;

    if (Requested)
      MaxNumSGPRs = Requested;
  }

  if (ST.hasSGPRInitBug())
    MaxNumSGPRs = SISubtarget::FIXED_SGPR_COUNT_FOR_INIT_BUG;

  return std::min(MaxNumSGPRs - getNumReservedSGPRs(ST, MFI),
                  MaxNumAddressableSGPRs);
}

unsigned SIRegisterInfo::getNumDebuggerReservedVGPRs(
  const SISubtarget &ST) const {
  if (ST.debuggerReserveRegs())
    return 4;
  return 0;
}

unsigned SIRegisterInfo::getMinNumVGPRs(unsigned WavesPerEU) const {
  switch (WavesPerEU) {
    case 0:  return 0;
    case 10: return 0;
    case 9:  return 25;
    case 8:  return 29;
    case 7:  return 33;
    case 6:  return 37;
    case 5:  return 41;
    case 4:  return 49;
    case 3:  return 65;
    case 2:  return 85;
    default: return 129;
  }
}

unsigned SIRegisterInfo::getMaxNumVGPRs(unsigned WavesPerEU) const {
  switch (WavesPerEU) {
    case 0:  return 24;
    case 10: return 24;
    case 9:  return 28;
    case 8:  return 32;
    case 7:  return 36;
    case 6:  return 40;
    case 5:  return 48;
    case 4:  return 64;
    case 3:  return 84;
    case 2:  return 128;
    default: return getTotalNumVGPRs();
  }
}

unsigned SIRegisterInfo::getMaxNumVGPRs(const MachineFunction &MF) const {
  const Function &F = *MF.getFunction();

  const SISubtarget &ST = MF.getSubtarget<SISubtarget>();
  const SIMachineFunctionInfo &MFI = *MF.getInfo<SIMachineFunctionInfo>();

  // Compute maximum number of VGPRs function can use using default/requested
  // minimum number of waves per execution unit.
  std::pair<unsigned, unsigned> WavesPerEU = MFI.getWavesPerEU();
  unsigned MaxNumVGPRs = getMaxNumVGPRs(WavesPerEU.first);

  // Check if maximum number of VGPRs was explicitly requested using
  // "amdgpu-num-vgpr" attribute.
  if (F.hasFnAttribute("amdgpu-num-vgpr")) {
    unsigned Requested = AMDGPU::getIntegerAttribute(
      F, "amdgpu-num-vgpr", MaxNumVGPRs);

    // Make sure requested value does not violate subtarget's specifications.
    if (Requested && Requested <= getNumDebuggerReservedVGPRs(ST))
      Requested = 0;

    // Make sure requested value is compatible with values implied by
    // default/requested minimum/maximum number of waves per execution unit.
    if (Requested && Requested > getMaxNumVGPRs(WavesPerEU.first))
      Requested = 0;
    if (WavesPerEU.second &&
        Requested && Requested < getMinNumVGPRs(WavesPerEU.second))
      Requested = 0;

    if (Requested)
      MaxNumVGPRs = Requested;
  }

  return MaxNumVGPRs - getNumDebuggerReservedVGPRs(ST);
}

ArrayRef<int16_t> SIRegisterInfo::getRegSplitParts(const TargetRegisterClass *RC,
                                                   unsigned EltSize) const {
  if (EltSize == 4) {
    static const int16_t Sub0_15[] = {
      AMDGPU::sub0, AMDGPU::sub1, AMDGPU::sub2, AMDGPU::sub3,
      AMDGPU::sub4, AMDGPU::sub5, AMDGPU::sub6, AMDGPU::sub7,
      AMDGPU::sub8, AMDGPU::sub9, AMDGPU::sub10, AMDGPU::sub11,
      AMDGPU::sub12, AMDGPU::sub13, AMDGPU::sub14, AMDGPU::sub15,
    };

    static const int16_t Sub0_7[] = {
      AMDGPU::sub0, AMDGPU::sub1, AMDGPU::sub2, AMDGPU::sub3,
      AMDGPU::sub4, AMDGPU::sub5, AMDGPU::sub6, AMDGPU::sub7,
    };

    static const int16_t Sub0_3[] = {
      AMDGPU::sub0, AMDGPU::sub1, AMDGPU::sub2, AMDGPU::sub3,
    };

    static const int16_t Sub0_2[] = {
      AMDGPU::sub0, AMDGPU::sub1, AMDGPU::sub2,
    };

    static const int16_t Sub0_1[] = {
      AMDGPU::sub0, AMDGPU::sub1,
    };

    switch (AMDGPU::getRegBitWidth(*RC->MC)) {
    case 32:
      return {};
    case 64:
      return makeArrayRef(Sub0_1);
    case 96:
      return makeArrayRef(Sub0_2);
    case 128:
      return makeArrayRef(Sub0_3);
    case 256:
      return makeArrayRef(Sub0_7);
    case 512:
      return makeArrayRef(Sub0_15);
    default:
      llvm_unreachable("unhandled register size");
    }
  }

  if (EltSize == 8) {
    static const int16_t Sub0_15_64[] = {
      AMDGPU::sub0_sub1, AMDGPU::sub2_sub3,
      AMDGPU::sub4_sub5, AMDGPU::sub6_sub7,
      AMDGPU::sub8_sub9, AMDGPU::sub10_sub11,
      AMDGPU::sub12_sub13, AMDGPU::sub14_sub15
    };

    static const int16_t Sub0_7_64[] = {
      AMDGPU::sub0_sub1, AMDGPU::sub2_sub3,
      AMDGPU::sub4_sub5, AMDGPU::sub6_sub7
    };


    static const int16_t Sub0_3_64[] = {
      AMDGPU::sub0_sub1, AMDGPU::sub2_sub3
    };

    switch (AMDGPU::getRegBitWidth(*RC->MC)) {
    case 64:
      return {};
    case 128:
      return makeArrayRef(Sub0_3_64);
    case 256:
      return makeArrayRef(Sub0_7_64);
    case 512:
      return makeArrayRef(Sub0_15_64);
    default:
      llvm_unreachable("unhandled register size");
    }
  }

  assert(EltSize == 16 && "unhandled register spill split size");

  static const int16_t Sub0_15_128[] = {
    AMDGPU::sub0_sub1_sub2_sub3,
    AMDGPU::sub4_sub5_sub6_sub7,
    AMDGPU::sub8_sub9_sub10_sub11,
    AMDGPU::sub12_sub13_sub14_sub15
  };

  static const int16_t Sub0_7_128[] = {
    AMDGPU::sub0_sub1_sub2_sub3,
    AMDGPU::sub4_sub5_sub6_sub7
  };

  switch (AMDGPU::getRegBitWidth(*RC->MC)) {
  case 128:
    return {};
  case 256:
    return makeArrayRef(Sub0_7_128);
  case 512:
    return makeArrayRef(Sub0_15_128);
  default:
    llvm_unreachable("unhandled register size");
  }
}

const TargetRegisterClass*
SIRegisterInfo::getRegClassForReg(const MachineRegisterInfo &MRI,
                                  unsigned Reg) const {
  if (TargetRegisterInfo::isVirtualRegister(Reg))
    return  MRI.getRegClass(Reg);

  return getPhysRegClass(Reg);
}

bool SIRegisterInfo::isVGPR(const MachineRegisterInfo &MRI,
                            unsigned Reg) const {
  return hasVGPRs(getRegClassForReg(MRI, Reg));
}