llvm.org GIT mirror llvm / master lib / Target / AMDGPU / SIFixSGPRCopies.cpp
master

Tree @master (Download .tar.gz)

SIFixSGPRCopies.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
//===- SIFixSGPRCopies.cpp - Remove potential VGPR => SGPR copies ---------===//
//
// 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
/// Copies from VGPR to SGPR registers are illegal and the register coalescer
/// will sometimes generate these illegal copies in situations like this:
///
///  Register Class <vsrc> is the union of <vgpr> and <sgpr>
///
/// BB0:
///   %0 <sgpr> = SCALAR_INST
///   %1 <vsrc> = COPY %0 <sgpr>
///    ...
///    BRANCH %cond BB1, BB2
///  BB1:
///    %2 <vgpr> = VECTOR_INST
///    %3 <vsrc> = COPY %2 <vgpr>
///  BB2:
///    %4 <vsrc> = PHI %1 <vsrc>, <%bb.0>, %3 <vrsc>, <%bb.1>
///    %5 <vgpr> = VECTOR_INST %4 <vsrc>
///
///
/// The coalescer will begin at BB0 and eliminate its copy, then the resulting
/// code will look like this:
///
/// BB0:
///   %0 <sgpr> = SCALAR_INST
///    ...
///    BRANCH %cond BB1, BB2
/// BB1:
///   %2 <vgpr> = VECTOR_INST
///   %3 <vsrc> = COPY %2 <vgpr>
/// BB2:
///   %4 <sgpr> = PHI %0 <sgpr>, <%bb.0>, %3 <vsrc>, <%bb.1>
///   %5 <vgpr> = VECTOR_INST %4 <sgpr>
///
/// Now that the result of the PHI instruction is an SGPR, the register
/// allocator is now forced to constrain the register class of %3 to
/// <sgpr> so we end up with final code like this:
///
/// BB0:
///   %0 <sgpr> = SCALAR_INST
///    ...
///    BRANCH %cond BB1, BB2
/// BB1:
///   %2 <vgpr> = VECTOR_INST
///   %3 <sgpr> = COPY %2 <vgpr>
/// BB2:
///   %4 <sgpr> = PHI %0 <sgpr>, <%bb.0>, %3 <sgpr>, <%bb.1>
///   %5 <vgpr> = VECTOR_INST %4 <sgpr>
///
/// Now this code contains an illegal copy from a VGPR to an SGPR.
///
/// In order to avoid this problem, this pass searches for PHI instructions
/// which define a <vsrc> register and constrains its definition class to
/// <vgpr> if the user of the PHI's definition register is a vector instruction.
/// If the PHI's definition class is constrained to <vgpr> then the coalescer
/// will be unable to perform the COPY removal from the above example  which
/// ultimately led to the creation of an illegal COPY.
//===----------------------------------------------------------------------===//

#include "AMDGPU.h"
#include "AMDGPUSubtarget.h"
#include "SIInstrInfo.h"
#include "SIRegisterInfo.h"
#include "MCTargetDesc/AMDGPUMCTargetDesc.h"
#include "llvm/ADT/DenseSet.h"
#include "llvm/ADT/STLExtras.h"
#include "llvm/ADT/SmallSet.h"
#include "llvm/ADT/SmallVector.h"
#include "llvm/CodeGen/MachineBasicBlock.h"
#include "llvm/CodeGen/MachineDominators.h"
#include "llvm/CodeGen/MachineFunction.h"
#include "llvm/CodeGen/MachineFunctionPass.h"
#include "llvm/CodeGen/MachineInstr.h"
#include "llvm/CodeGen/MachineInstrBuilder.h"
#include "llvm/CodeGen/MachineOperand.h"
#include "llvm/CodeGen/MachineRegisterInfo.h"
#include "llvm/CodeGen/TargetRegisterInfo.h"
#include "llvm/Pass.h"
#include "llvm/Support/CodeGen.h"
#include "llvm/Support/CommandLine.h"
#include "llvm/Support/Debug.h"
#include "llvm/Support/raw_ostream.h"
#include "llvm/Target/TargetMachine.h"
#include <cassert>
#include <cstdint>
#include <iterator>
#include <list>
#include <map>
#include <tuple>
#include <utility>

using namespace llvm;

#define DEBUG_TYPE "si-fix-sgpr-copies"

static cl::opt<bool> EnableM0Merge(
  "amdgpu-enable-merge-m0",
  cl::desc("Merge and hoist M0 initializations"),
  cl::init(true));

namespace {

class SIFixSGPRCopies : public MachineFunctionPass {
  MachineDominatorTree *MDT;

public:
  static char ID;

  MachineRegisterInfo *MRI;
  const SIRegisterInfo *TRI;
  const SIInstrInfo *TII;

  SIFixSGPRCopies() : MachineFunctionPass(ID) {}

  bool runOnMachineFunction(MachineFunction &MF) override;

  void processPHINode(MachineInstr &MI);

  StringRef getPassName() const override { return "SI Fix SGPR copies"; }

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

} // end anonymous namespace

INITIALIZE_PASS_BEGIN(SIFixSGPRCopies, DEBUG_TYPE,
                     "SI Fix SGPR copies", false, false)
INITIALIZE_PASS_DEPENDENCY(MachineDominatorTree)
INITIALIZE_PASS_END(SIFixSGPRCopies, DEBUG_TYPE,
                     "SI Fix SGPR copies", false, false)

char SIFixSGPRCopies::ID = 0;

char &llvm::SIFixSGPRCopiesID = SIFixSGPRCopies::ID;

FunctionPass *llvm::createSIFixSGPRCopiesPass() {
  return new SIFixSGPRCopies();
}

static bool hasVectorOperands(const MachineInstr &MI,
                              const SIRegisterInfo *TRI) {
  const MachineRegisterInfo &MRI = MI.getParent()->getParent()->getRegInfo();
  for (unsigned i = 0, e = MI.getNumOperands(); i != e; ++i) {
    if (!MI.getOperand(i).isReg() ||
        !Register::isVirtualRegister(MI.getOperand(i).getReg()))
      continue;

    if (TRI->hasVectorRegisters(MRI.getRegClass(MI.getOperand(i).getReg())))
      return true;
  }
  return false;
}

static std::pair<const TargetRegisterClass *, const TargetRegisterClass *>
getCopyRegClasses(const MachineInstr &Copy,
                  const SIRegisterInfo &TRI,
                  const MachineRegisterInfo &MRI) {
  Register DstReg = Copy.getOperand(0).getReg();
  Register SrcReg = Copy.getOperand(1).getReg();

  const TargetRegisterClass *SrcRC = Register::isVirtualRegister(SrcReg)
                                         ? MRI.getRegClass(SrcReg)
                                         : TRI.getPhysRegClass(SrcReg);

  // We don't really care about the subregister here.
  // SrcRC = TRI.getSubRegClass(SrcRC, Copy.getOperand(1).getSubReg());

  const TargetRegisterClass *DstRC = Register::isVirtualRegister(DstReg)
                                         ? MRI.getRegClass(DstReg)
                                         : TRI.getPhysRegClass(DstReg);

  return std::make_pair(SrcRC, DstRC);
}

static bool isVGPRToSGPRCopy(const TargetRegisterClass *SrcRC,
                             const TargetRegisterClass *DstRC,
                             const SIRegisterInfo &TRI) {
  return SrcRC != &AMDGPU::VReg_1RegClass && TRI.isSGPRClass(DstRC) &&
         TRI.hasVectorRegisters(SrcRC);
}

static bool isSGPRToVGPRCopy(const TargetRegisterClass *SrcRC,
                             const TargetRegisterClass *DstRC,
                             const SIRegisterInfo &TRI) {
  return DstRC != &AMDGPU::VReg_1RegClass && TRI.isSGPRClass(SrcRC) &&
         TRI.hasVectorRegisters(DstRC);
}

static bool tryChangeVGPRtoSGPRinCopy(MachineInstr &MI,
                                      const SIRegisterInfo *TRI,
                                      const SIInstrInfo *TII) {
  MachineRegisterInfo &MRI = MI.getParent()->getParent()->getRegInfo();
  auto &Src = MI.getOperand(1);
  Register DstReg = MI.getOperand(0).getReg();
  Register SrcReg = Src.getReg();
  if (!Register::isVirtualRegister(SrcReg) ||
      !Register::isVirtualRegister(DstReg))
    return false;

  for (const auto &MO : MRI.reg_nodbg_operands(DstReg)) {
    const auto *UseMI = MO.getParent();
    if (UseMI == &MI)
      continue;
    if (MO.isDef() || UseMI->getParent() != MI.getParent() ||
        UseMI->getOpcode() <= TargetOpcode::GENERIC_OP_END ||
        !TII->isOperandLegal(*UseMI, UseMI->getOperandNo(&MO), &Src))
      return false;
  }
  // Change VGPR to SGPR destination.
  MRI.setRegClass(DstReg, TRI->getEquivalentSGPRClass(MRI.getRegClass(DstReg)));
  return true;
}

// Distribute an SGPR->VGPR copy of a REG_SEQUENCE into a VGPR REG_SEQUENCE.
//
// SGPRx = ...
// SGPRy = REG_SEQUENCE SGPRx, sub0 ...
// VGPRz = COPY SGPRy
//
// ==>
//
// VGPRx = COPY SGPRx
// VGPRz = REG_SEQUENCE VGPRx, sub0
//
// This exposes immediate folding opportunities when materializing 64-bit
// immediates.
static bool foldVGPRCopyIntoRegSequence(MachineInstr &MI,
                                        const SIRegisterInfo *TRI,
                                        const SIInstrInfo *TII,
                                        MachineRegisterInfo &MRI) {
  assert(MI.isRegSequence());

  Register DstReg = MI.getOperand(0).getReg();
  if (!TRI->isSGPRClass(MRI.getRegClass(DstReg)))
    return false;

  if (!MRI.hasOneUse(DstReg))
    return false;

  MachineInstr &CopyUse = *MRI.use_instr_begin(DstReg);
  if (!CopyUse.isCopy())
    return false;

  // It is illegal to have vreg inputs to a physreg defining reg_sequence.
  if (Register::isPhysicalRegister(CopyUse.getOperand(0).getReg()))
    return false;

  const TargetRegisterClass *SrcRC, *DstRC;
  std::tie(SrcRC, DstRC) = getCopyRegClasses(CopyUse, *TRI, MRI);

  if (!isSGPRToVGPRCopy(SrcRC, DstRC, *TRI))
    return false;

  if (tryChangeVGPRtoSGPRinCopy(CopyUse, TRI, TII))
    return true;

  // TODO: Could have multiple extracts?
  unsigned SubReg = CopyUse.getOperand(1).getSubReg();
  if (SubReg != AMDGPU::NoSubRegister)
    return false;

  MRI.setRegClass(DstReg, DstRC);

  // SGPRx = ...
  // SGPRy = REG_SEQUENCE SGPRx, sub0 ...
  // VGPRz = COPY SGPRy

  // =>
  // VGPRx = COPY SGPRx
  // VGPRz = REG_SEQUENCE VGPRx, sub0

  MI.getOperand(0).setReg(CopyUse.getOperand(0).getReg());
  bool IsAGPR = TRI->hasAGPRs(DstRC);

  for (unsigned I = 1, N = MI.getNumOperands(); I != N; I += 2) {
    Register SrcReg = MI.getOperand(I).getReg();
    unsigned SrcSubReg = MI.getOperand(I).getSubReg();

    const TargetRegisterClass *SrcRC = MRI.getRegClass(SrcReg);
    assert(TRI->isSGPRClass(SrcRC) &&
           "Expected SGPR REG_SEQUENCE to only have SGPR inputs");

    SrcRC = TRI->getSubRegClass(SrcRC, SrcSubReg);
    const TargetRegisterClass *NewSrcRC = TRI->getEquivalentVGPRClass(SrcRC);

    Register TmpReg = MRI.createVirtualRegister(NewSrcRC);

    BuildMI(*MI.getParent(), &MI, MI.getDebugLoc(), TII->get(AMDGPU::COPY),
            TmpReg)
        .add(MI.getOperand(I));

    if (IsAGPR) {
      const TargetRegisterClass *NewSrcRC = TRI->getEquivalentAGPRClass(SrcRC);
      Register TmpAReg = MRI.createVirtualRegister(NewSrcRC);
      unsigned Opc = NewSrcRC == &AMDGPU::AGPR_32RegClass ?
        AMDGPU::V_ACCVGPR_WRITE_B32 : AMDGPU::COPY;
      BuildMI(*MI.getParent(), &MI, MI.getDebugLoc(), TII->get(Opc),
            TmpAReg)
        .addReg(TmpReg, RegState::Kill);
      TmpReg = TmpAReg;
    }

    MI.getOperand(I).setReg(TmpReg);
  }

  CopyUse.eraseFromParent();
  return true;
}

static bool isSafeToFoldImmIntoCopy(const MachineInstr *Copy,
                                    const MachineInstr *MoveImm,
                                    const SIInstrInfo *TII,
                                    unsigned &SMovOp,
                                    int64_t &Imm) {
  if (Copy->getOpcode() != AMDGPU::COPY)
    return false;

  if (!MoveImm->isMoveImmediate())
    return false;

  const MachineOperand *ImmOp =
      TII->getNamedOperand(*MoveImm, AMDGPU::OpName::src0);
  if (!ImmOp->isImm())
    return false;

  // FIXME: Handle copies with sub-regs.
  if (Copy->getOperand(0).getSubReg())
    return false;

  switch (MoveImm->getOpcode()) {
  default:
    return false;
  case AMDGPU::V_MOV_B32_e32:
    SMovOp = AMDGPU::S_MOV_B32;
    break;
  case AMDGPU::V_MOV_B64_PSEUDO:
    SMovOp = AMDGPU::S_MOV_B64;
    break;
  }
  Imm = ImmOp->getImm();
  return true;
}

template <class UnaryPredicate>
bool searchPredecessors(const MachineBasicBlock *MBB,
                        const MachineBasicBlock *CutOff,
                        UnaryPredicate Predicate) {
  if (MBB == CutOff)
    return false;

  DenseSet<const MachineBasicBlock *> Visited;
  SmallVector<MachineBasicBlock *, 4> Worklist(MBB->pred_begin(),
                                               MBB->pred_end());

  while (!Worklist.empty()) {
    MachineBasicBlock *MBB = Worklist.pop_back_val();

    if (!Visited.insert(MBB).second)
      continue;
    if (MBB == CutOff)
      continue;
    if (Predicate(MBB))
      return true;

    Worklist.append(MBB->pred_begin(), MBB->pred_end());
  }

  return false;
}

// Checks if there is potential path From instruction To instruction.
// If CutOff is specified and it sits in between of that path we ignore
// a higher portion of the path and report it is not reachable.
static bool isReachable(const MachineInstr *From,
                        const MachineInstr *To,
                        const MachineBasicBlock *CutOff,
                        MachineDominatorTree &MDT) {
  // If either From block dominates To block or instructions are in the same
  // block and From is higher.
  if (MDT.dominates(From, To))
    return true;

  const MachineBasicBlock *MBBFrom = From->getParent();
  const MachineBasicBlock *MBBTo = To->getParent();
  if (MBBFrom == MBBTo)
    return false;

  // Instructions are in different blocks, do predecessor search.
  // We should almost never get here since we do not usually produce M0 stores
  // other than -1.
  return searchPredecessors(MBBTo, CutOff, [MBBFrom]
           (const MachineBasicBlock *MBB) { return MBB == MBBFrom; });
}

// Return the first non-prologue instruction in the block.
static MachineBasicBlock::iterator
getFirstNonPrologue(MachineBasicBlock *MBB, const TargetInstrInfo *TII) {
  MachineBasicBlock::iterator I = MBB->getFirstNonPHI();
  while (I != MBB->end() && TII->isBasicBlockPrologue(*I))
    ++I;

  return I;
}

// Hoist and merge identical SGPR initializations into a common predecessor.
// This is intended to combine M0 initializations, but can work with any
// SGPR. A VGPR cannot be processed since we cannot guarantee vector
// executioon.
static bool hoistAndMergeSGPRInits(unsigned Reg,
                                   const MachineRegisterInfo &MRI,
                                   const TargetRegisterInfo *TRI,
                                   MachineDominatorTree &MDT,
                                   const TargetInstrInfo *TII) {
  // List of inits by immediate value.
  using InitListMap = std::map<unsigned, std::list<MachineInstr *>>;
  InitListMap Inits;
  // List of clobbering instructions.
  SmallVector<MachineInstr*, 8> Clobbers;
  // List of instructions marked for deletion.
  SmallSet<MachineInstr*, 8> MergedInstrs;

  bool Changed = false;

  for (auto &MI : MRI.def_instructions(Reg)) {
    MachineOperand *Imm = nullptr;
    for (auto &MO : MI.operands()) {
      if ((MO.isReg() && ((MO.isDef() && MO.getReg() != Reg) || !MO.isDef())) ||
          (!MO.isImm() && !MO.isReg()) || (MO.isImm() && Imm)) {
        Imm = nullptr;
        break;
      } else if (MO.isImm())
        Imm = &MO;
    }
    if (Imm)
      Inits[Imm->getImm()].push_front(&MI);
    else
      Clobbers.push_back(&MI);
  }

  for (auto &Init : Inits) {
    auto &Defs = Init.second;

    for (auto I1 = Defs.begin(), E = Defs.end(); I1 != E; ) {
      MachineInstr *MI1 = *I1;

      for (auto I2 = std::next(I1); I2 != E; ) {
        MachineInstr *MI2 = *I2;

        // Check any possible interference
        auto interferes = [&](MachineBasicBlock::iterator From,
                              MachineBasicBlock::iterator To) -> bool {

          assert(MDT.dominates(&*To, &*From));

          auto interferes = [&MDT, From, To](MachineInstr* &Clobber) -> bool {
            const MachineBasicBlock *MBBFrom = From->getParent();
            const MachineBasicBlock *MBBTo = To->getParent();
            bool MayClobberFrom = isReachable(Clobber, &*From, MBBTo, MDT);
            bool MayClobberTo = isReachable(Clobber, &*To, MBBTo, MDT);
            if (!MayClobberFrom && !MayClobberTo)
              return false;
            if ((MayClobberFrom && !MayClobberTo) ||
                (!MayClobberFrom && MayClobberTo))
              return true;
            // Both can clobber, this is not an interference only if both are
            // dominated by Clobber and belong to the same block or if Clobber
            // properly dominates To, given that To >> From, so it dominates
            // both and located in a common dominator.
            return !((MBBFrom == MBBTo &&
                      MDT.dominates(Clobber, &*From) &&
                      MDT.dominates(Clobber, &*To)) ||
                     MDT.properlyDominates(Clobber->getParent(), MBBTo));
          };

          return (llvm::any_of(Clobbers, interferes)) ||
                 (llvm::any_of(Inits, [&](InitListMap::value_type &C) {
                    return C.first != Init.first &&
                           llvm::any_of(C.second, interferes);
                  }));
        };

        if (MDT.dominates(MI1, MI2)) {
          if (!interferes(MI2, MI1)) {
            LLVM_DEBUG(dbgs()
                       << "Erasing from "
                       << printMBBReference(*MI2->getParent()) << " " << *MI2);
            MergedInstrs.insert(MI2);
            Changed = true;
            ++I2;
            continue;
          }
        } else if (MDT.dominates(MI2, MI1)) {
          if (!interferes(MI1, MI2)) {
            LLVM_DEBUG(dbgs()
                       << "Erasing from "
                       << printMBBReference(*MI1->getParent()) << " " << *MI1);
            MergedInstrs.insert(MI1);
            Changed = true;
            ++I1;
            break;
          }
        } else {
          auto *MBB = MDT.findNearestCommonDominator(MI1->getParent(),
                                                     MI2->getParent());
          if (!MBB) {
            ++I2;
            continue;
          }

          MachineBasicBlock::iterator I = getFirstNonPrologue(MBB, TII);
          if (!interferes(MI1, I) && !interferes(MI2, I)) {
            LLVM_DEBUG(dbgs()
                       << "Erasing from "
                       << printMBBReference(*MI1->getParent()) << " " << *MI1
                       << "and moving from "
                       << printMBBReference(*MI2->getParent()) << " to "
                       << printMBBReference(*I->getParent()) << " " << *MI2);
            I->getParent()->splice(I, MI2->getParent(), MI2);
            MergedInstrs.insert(MI1);
            Changed = true;
            ++I1;
            break;
          }
        }
        ++I2;
      }
      ++I1;
    }
  }

  // Remove initializations that were merged into another.
  for (auto &Init : Inits) {
    auto &Defs = Init.second;
    auto I = Defs.begin();
    while (I != Defs.end()) {
      if (MergedInstrs.count(*I)) {
        (*I)->eraseFromParent();
        I = Defs.erase(I);
      } else
        ++I;
    }
  }

  // Try to schedule SGPR initializations as early as possible in the MBB.
  for (auto &Init : Inits) {
    auto &Defs = Init.second;
    for (auto MI : Defs) {
      auto MBB = MI->getParent();
      MachineInstr &BoundaryMI = *getFirstNonPrologue(MBB, TII);
      MachineBasicBlock::reverse_iterator B(BoundaryMI);
      // Check if B should actually be a boundary. If not set the previous
      // instruction as the boundary instead.
      if (!TII->isBasicBlockPrologue(*B))
        B++;

      auto R = std::next(MI->getReverseIterator());
      const unsigned Threshold = 50;
      // Search until B or Threshold for a place to insert the initialization.
      for (unsigned I = 0; R != B && I < Threshold; ++R, ++I)
        if (R->readsRegister(Reg, TRI) || R->definesRegister(Reg, TRI) ||
            TII->isSchedulingBoundary(*R, MBB, *MBB->getParent()))
          break;

      // Move to directly after R.
      if (&*--R != MI)
        MBB->splice(*R, MBB, MI);
    }
  }

  if (Changed)
    MRI.clearKillFlags(Reg);

  return Changed;
}

bool SIFixSGPRCopies::runOnMachineFunction(MachineFunction &MF) {
  const GCNSubtarget &ST = MF.getSubtarget<GCNSubtarget>();
  MRI = &MF.getRegInfo();
  TRI = ST.getRegisterInfo();
  TII = ST.getInstrInfo();
  MDT = &getAnalysis<MachineDominatorTree>();

  SmallVector<MachineInstr *, 16> Worklist;

  for (MachineFunction::iterator BI = MF.begin(), BE = MF.end();
                                                  BI != BE; ++BI) {
    MachineBasicBlock &MBB = *BI;
    for (MachineBasicBlock::iterator I = MBB.begin(), E = MBB.end();
         I != E; ++I) {
      MachineInstr &MI = *I;

      switch (MI.getOpcode()) {
      default:
        continue;
      case AMDGPU::COPY:
      case AMDGPU::WQM:
      case AMDGPU::SOFT_WQM:
      case AMDGPU::WWM: {
        Register DstReg = MI.getOperand(0).getReg();

        const TargetRegisterClass *SrcRC, *DstRC;
        std::tie(SrcRC, DstRC) = getCopyRegClasses(MI, *TRI, *MRI);

        if (!Register::isVirtualRegister(DstReg)) {
          // If the destination register is a physical register there isn't
          // really much we can do to fix this.
          // Some special instructions use M0 as an input. Some even only use
          // the first lane. Insert a readfirstlane and hope for the best.
          if (DstReg == AMDGPU::M0 && TRI->hasVectorRegisters(SrcRC)) {
            Register TmpReg
              = MRI->createVirtualRegister(&AMDGPU::SReg_32_XM0RegClass);

            BuildMI(MBB, MI, MI.getDebugLoc(),
                    TII->get(AMDGPU::V_READFIRSTLANE_B32), TmpReg)
              .add(MI.getOperand(1));
            MI.getOperand(1).setReg(TmpReg);
          }

          continue;
        }

        if (isVGPRToSGPRCopy(SrcRC, DstRC, *TRI)) {
          Register SrcReg = MI.getOperand(1).getReg();
          if (!Register::isVirtualRegister(SrcReg)) {
            TII->moveToVALU(MI, MDT);
            break;
          }

          MachineInstr *DefMI = MRI->getVRegDef(SrcReg);
          unsigned SMovOp;
          int64_t Imm;
          // If we are just copying an immediate, we can replace the copy with
          // s_mov_b32.
          if (isSafeToFoldImmIntoCopy(&MI, DefMI, TII, SMovOp, Imm)) {
            MI.getOperand(1).ChangeToImmediate(Imm);
            MI.addImplicitDefUseOperands(MF);
            MI.setDesc(TII->get(SMovOp));
            break;
          }
          TII->moveToVALU(MI, MDT);
        } else if (isSGPRToVGPRCopy(SrcRC, DstRC, *TRI)) {
          tryChangeVGPRtoSGPRinCopy(MI, TRI, TII);
        }

        break;
      }
      case AMDGPU::PHI: {
        processPHINode(MI);
        break;
      }
      case AMDGPU::REG_SEQUENCE:
        if (TRI->hasVectorRegisters(TII->getOpRegClass(MI, 0)) ||
            !hasVectorOperands(MI, TRI)) {
          foldVGPRCopyIntoRegSequence(MI, TRI, TII, *MRI);
          continue;
        }

        LLVM_DEBUG(dbgs() << "Fixing REG_SEQUENCE: " << MI);

        TII->moveToVALU(MI, MDT);
        break;
      case AMDGPU::INSERT_SUBREG: {
        const TargetRegisterClass *DstRC, *Src0RC, *Src1RC;
        DstRC = MRI->getRegClass(MI.getOperand(0).getReg());
        Src0RC = MRI->getRegClass(MI.getOperand(1).getReg());
        Src1RC = MRI->getRegClass(MI.getOperand(2).getReg());
        if (TRI->isSGPRClass(DstRC) &&
            (TRI->hasVectorRegisters(Src0RC) ||
             TRI->hasVectorRegisters(Src1RC))) {
          LLVM_DEBUG(dbgs() << " Fixing INSERT_SUBREG: " << MI);
          TII->moveToVALU(MI, MDT);
        }
        break;
      }
      case AMDGPU::V_WRITELANE_B32: {
        // Some architectures allow more than one constant bus access without
        // SGPR restriction
        if (ST.getConstantBusLimit(MI.getOpcode()) != 1)
          break;

        // Writelane is special in that it can use SGPR and M0 (which would
        // normally count as using the constant bus twice - but in this case it
        // is allowed since the lane selector doesn't count as a use of the
        // constant bus). However, it is still required to abide by the 1 SGPR
        // rule. Apply a fix here as we might have multiple SGPRs after
        // legalizing VGPRs to SGPRs
        int Src0Idx =
            AMDGPU::getNamedOperandIdx(MI.getOpcode(), AMDGPU::OpName::src0);
        int Src1Idx =
            AMDGPU::getNamedOperandIdx(MI.getOpcode(), AMDGPU::OpName::src1);
        MachineOperand &Src0 = MI.getOperand(Src0Idx);
        MachineOperand &Src1 = MI.getOperand(Src1Idx);

        // Check to see if the instruction violates the 1 SGPR rule
        if ((Src0.isReg() && TRI->isSGPRReg(*MRI, Src0.getReg()) &&
             Src0.getReg() != AMDGPU::M0) &&
            (Src1.isReg() && TRI->isSGPRReg(*MRI, Src1.getReg()) &&
             Src1.getReg() != AMDGPU::M0)) {

          // Check for trivially easy constant prop into one of the operands
          // If this is the case then perform the operation now to resolve SGPR
          // issue. If we don't do that here we will always insert a mov to m0
          // that can't be resolved in later operand folding pass
          bool Resolved = false;
          for (MachineOperand *MO : {&Src0, &Src1}) {
            if (Register::isVirtualRegister(MO->getReg())) {
              MachineInstr *DefMI = MRI->getVRegDef(MO->getReg());
              if (DefMI && TII->isFoldableCopy(*DefMI)) {
                const MachineOperand &Def = DefMI->getOperand(0);
                if (Def.isReg() &&
                    MO->getReg() == Def.getReg() &&
                    MO->getSubReg() == Def.getSubReg()) {
                  const MachineOperand &Copied = DefMI->getOperand(1);
                  if (Copied.isImm() &&
                      TII->isInlineConstant(APInt(64, Copied.getImm(), true))) {
                    MO->ChangeToImmediate(Copied.getImm());
                    Resolved = true;
                    break;
                  }
                }
              }
            }
          }

          if (!Resolved) {
            // Haven't managed to resolve by replacing an SGPR with an immediate
            // Move src1 to be in M0
            BuildMI(*MI.getParent(), MI, MI.getDebugLoc(),
                    TII->get(AMDGPU::COPY), AMDGPU::M0)
                .add(Src1);
            Src1.ChangeToRegister(AMDGPU::M0, false);
          }
        }
        break;
      }
      }
    }
  }

  if (MF.getTarget().getOptLevel() > CodeGenOpt::None && EnableM0Merge)
    hoistAndMergeSGPRInits(AMDGPU::M0, *MRI, TRI, *MDT, TII);

  return true;
}

void SIFixSGPRCopies::processPHINode(MachineInstr &MI) {
  unsigned numVGPRUses = 0;
  bool AllAGPRUses = true;
  SetVector<const MachineInstr *> worklist;
  SmallSet<const MachineInstr *, 4> Visited;
  worklist.insert(&MI);
  Visited.insert(&MI);
  while (!worklist.empty()) {
    const MachineInstr *Instr = worklist.pop_back_val();
    unsigned Reg = Instr->getOperand(0).getReg();
    for (const auto &Use : MRI->use_operands(Reg)) {
      const MachineInstr *UseMI = Use.getParent();
      AllAGPRUses &= (UseMI->isCopy() &&
                      TRI->isAGPR(*MRI, UseMI->getOperand(0).getReg())) ||
                     TRI->isAGPR(*MRI, Use.getReg());
      if (UseMI->isCopy() || UseMI->isRegSequence()) {
        if (UseMI->isCopy() &&
          UseMI->getOperand(0).getReg().isPhysical() &&
          !TRI->isSGPRReg(*MRI, UseMI->getOperand(0).getReg())) {
          numVGPRUses++;
        }
        if (Visited.insert(UseMI).second)
          worklist.insert(UseMI);

        continue;
      }

      if (UseMI->isPHI()) {
        const TargetRegisterClass *UseRC = MRI->getRegClass(Use.getReg());
        if (!TRI->isSGPRReg(*MRI, Use.getReg()) &&
          UseRC != &AMDGPU::VReg_1RegClass)
          numVGPRUses++;
        continue;
      }

      const TargetRegisterClass *OpRC =
        TII->getOpRegClass(*UseMI, UseMI->getOperandNo(&Use));
      if (!TRI->isSGPRClass(OpRC) && OpRC != &AMDGPU::VS_32RegClass &&
        OpRC != &AMDGPU::VS_64RegClass) {
        numVGPRUses++;
      }
    }
  }

  Register PHIRes = MI.getOperand(0).getReg();
  const TargetRegisterClass *RC0 = MRI->getRegClass(PHIRes);
  if (AllAGPRUses && numVGPRUses && !TRI->hasAGPRs(RC0)) {
    LLVM_DEBUG(dbgs() << "Moving PHI to AGPR: " << MI);
    MRI->setRegClass(PHIRes, TRI->getEquivalentAGPRClass(RC0));
  }

  bool hasVGPRInput = false;
  for (unsigned i = 1; i < MI.getNumOperands(); i += 2) {
    unsigned InputReg = MI.getOperand(i).getReg();
    MachineInstr *Def = MRI->getVRegDef(InputReg);
    if (TRI->isVectorRegister(*MRI, InputReg)) {
      if (Def->isCopy()) {
        unsigned SrcReg = Def->getOperand(1).getReg();
        const TargetRegisterClass *RC =
          TRI->getRegClassForReg(*MRI, SrcReg);
        if (TRI->isSGPRClass(RC))
          continue;
      }
      hasVGPRInput = true;
      break;
    }
    else if (Def->isCopy() &&
      TRI->isVectorRegister(*MRI, Def->getOperand(1).getReg())) {
      hasVGPRInput = true;
      break;
    }
  }

  if ((!TRI->isVectorRegister(*MRI, PHIRes) &&
       RC0 != &AMDGPU::VReg_1RegClass) &&
    (hasVGPRInput || numVGPRUses > 1)) {
    LLVM_DEBUG(dbgs() << "Fixing PHI: " << MI);
    TII->moveToVALU(MI);
  }
  else {
    LLVM_DEBUG(dbgs() << "Legalizing PHI: " << MI);
    TII->legalizeOperands(MI, MDT);
  }

}