llvm.org GIT mirror llvm / 15c435a lib / CodeGen / PeepholeOptimizer.cpp
15c435a

Tree @15c435a (Download .tar.gz)

PeepholeOptimizer.cpp @15c435araw · 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
//===-- PeepholeOptimizer.cpp - Peephole Optimizations --------------------===//
//
//                     The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// Perform peephole optimizations on the machine code:
//
// - Optimize Extensions
//
//     Optimization of sign / zero extension instructions. It may be extended to
//     handle other instructions with similar properties.
//
//     On some targets, some instructions, e.g. X86 sign / zero extension, may
//     leave the source value in the lower part of the result. This optimization
//     will replace some uses of the pre-extension value with uses of the
//     sub-register of the results.
//
// - Optimize Comparisons
//
//     Optimization of comparison instructions. For instance, in this code:
//
//       sub r1, 1
//       cmp r1, 0
//       bz  L1
//
//     If the "sub" instruction all ready sets (or could be modified to set) the
//     same flag that the "cmp" instruction sets and that "bz" uses, then we can
//     eliminate the "cmp" instruction.
//
//     Another instance, in this code:
//
//       sub r1, r3 | sub r1, imm
//       cmp r3, r1 or cmp r1, r3 | cmp r1, imm
//       bge L1
//
//     If the branch instruction can use flag from "sub", then we can replace
//     "sub" with "subs" and eliminate the "cmp" instruction.
//
// - Optimize Loads:
//
//     Loads that can be folded into a later instruction. A load is foldable
//     if it loads to virtual registers and the virtual register defined has 
//     a single use.
//
// - Optimize Copies and Bitcast:
//
//     Rewrite copies and bitcasts to avoid cross register bank copies
//     when possible.
//     E.g., Consider the following example, where capital and lower
//     letters denote different register file:
//     b = copy A <-- cross-bank copy
//     C = copy b <-- cross-bank copy
//   =>
//     b = copy A <-- cross-bank copy
//     C = copy A <-- same-bank copy
//
//     E.g., for bitcast:
//     b = bitcast A <-- cross-bank copy
//     C = bitcast b <-- cross-bank copy
//   =>
//     b = bitcast A <-- cross-bank copy
//     C = copy A    <-- same-bank copy
//===----------------------------------------------------------------------===//

#define DEBUG_TYPE "peephole-opt"
#include "llvm/CodeGen/Passes.h"
#include "llvm/ADT/DenseMap.h"
#include "llvm/ADT/SmallPtrSet.h"
#include "llvm/ADT/SmallSet.h"
#include "llvm/ADT/Statistic.h"
#include "llvm/CodeGen/MachineDominators.h"
#include "llvm/CodeGen/MachineInstrBuilder.h"
#include "llvm/CodeGen/MachineRegisterInfo.h"
#include "llvm/Support/CommandLine.h"
#include "llvm/Support/Debug.h"
#include "llvm/Target/TargetInstrInfo.h"
#include "llvm/Target/TargetRegisterInfo.h"
using namespace llvm;

// Optimize Extensions
static cl::opt<bool>
Aggressive("aggressive-ext-opt", cl::Hidden,
           cl::desc("Aggressive extension optimization"));

static cl::opt<bool>
DisablePeephole("disable-peephole", cl::Hidden, cl::init(false),
                cl::desc("Disable the peephole optimizer"));

STATISTIC(NumReuse,      "Number of extension results reused");
STATISTIC(NumCmps,       "Number of compares eliminated");
STATISTIC(NumImmFold,    "Number of move immediate folded");
STATISTIC(NumLoadFold,   "Number of loads folded");
STATISTIC(NumSelects,    "Number of selects optimized");
STATISTIC(NumCopiesBitcasts, "Number of copies/bitcasts optimized");

namespace {
  class PeepholeOptimizer : public MachineFunctionPass {
    const TargetMachine   *TM;
    const TargetInstrInfo *TII;
    MachineRegisterInfo   *MRI;
    MachineDominatorTree  *DT;  // Machine dominator tree

  public:
    static char ID; // Pass identification
    PeepholeOptimizer() : MachineFunctionPass(ID) {
      initializePeepholeOptimizerPass(*PassRegistry::getPassRegistry());
    }

    bool runOnMachineFunction(MachineFunction &MF) override;

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

  private:
    bool optimizeCmpInstr(MachineInstr *MI, MachineBasicBlock *MBB);
    bool optimizeExtInstr(MachineInstr *MI, MachineBasicBlock *MBB,
                          SmallPtrSet<MachineInstr*, 8> &LocalMIs);
    bool optimizeSelect(MachineInstr *MI);
    bool optimizeCopyOrBitcast(MachineInstr *MI);
    bool isMoveImmediate(MachineInstr *MI,
                         SmallSet<unsigned, 4> &ImmDefRegs,
                         DenseMap<unsigned, MachineInstr*> &ImmDefMIs);
    bool foldImmediate(MachineInstr *MI, MachineBasicBlock *MBB,
                       SmallSet<unsigned, 4> &ImmDefRegs,
                       DenseMap<unsigned, MachineInstr*> &ImmDefMIs);
    bool isLoadFoldable(MachineInstr *MI,
                        SmallSet<unsigned, 16> &FoldAsLoadDefCandidates);
  };
}

char PeepholeOptimizer::ID = 0;
char &llvm::PeepholeOptimizerID = PeepholeOptimizer::ID;
INITIALIZE_PASS_BEGIN(PeepholeOptimizer, "peephole-opts",
                "Peephole Optimizations", false, false)
INITIALIZE_PASS_DEPENDENCY(MachineDominatorTree)
INITIALIZE_PASS_END(PeepholeOptimizer, "peephole-opts",
                "Peephole Optimizations", false, false)

/// optimizeExtInstr - If instruction is a copy-like instruction, i.e. it reads
/// a single register and writes a single register and it does not modify the
/// source, and if the source value is preserved as a sub-register of the
/// result, then replace all reachable uses of the source with the subreg of the
/// result.
///
/// Do not generate an EXTRACT that is used only in a debug use, as this changes
/// the code. Since this code does not currently share EXTRACTs, just ignore all
/// debug uses.
bool PeepholeOptimizer::
optimizeExtInstr(MachineInstr *MI, MachineBasicBlock *MBB,
                 SmallPtrSet<MachineInstr*, 8> &LocalMIs) {
  unsigned SrcReg, DstReg, SubIdx;
  if (!TII->isCoalescableExtInstr(*MI, SrcReg, DstReg, SubIdx))
    return false;

  if (TargetRegisterInfo::isPhysicalRegister(DstReg) ||
      TargetRegisterInfo::isPhysicalRegister(SrcReg))
    return false;

  if (MRI->hasOneNonDBGUse(SrcReg))
    // No other uses.
    return false;

  // Ensure DstReg can get a register class that actually supports
  // sub-registers. Don't change the class until we commit.
  const TargetRegisterClass *DstRC = MRI->getRegClass(DstReg);
  DstRC = TM->getRegisterInfo()->getSubClassWithSubReg(DstRC, SubIdx);
  if (!DstRC)
    return false;

  // The ext instr may be operating on a sub-register of SrcReg as well.
  // PPC::EXTSW is a 32 -> 64-bit sign extension, but it reads a 64-bit
  // register.
  // If UseSrcSubIdx is Set, SubIdx also applies to SrcReg, and only uses of
  // SrcReg:SubIdx should be replaced.
  bool UseSrcSubIdx = TM->getRegisterInfo()->
    getSubClassWithSubReg(MRI->getRegClass(SrcReg), SubIdx) != 0;

  // The source has other uses. See if we can replace the other uses with use of
  // the result of the extension.
  SmallPtrSet<MachineBasicBlock*, 4> ReachedBBs;
  for (MachineInstr &UI : MRI->use_nodbg_instructions(DstReg))
    ReachedBBs.insert(UI.getParent());

  // Uses that are in the same BB of uses of the result of the instruction.
  SmallVector<MachineOperand*, 8> Uses;

  // Uses that the result of the instruction can reach.
  SmallVector<MachineOperand*, 8> ExtendedUses;

  bool ExtendLife = true;
  for (MachineOperand &UseMO : MRI->use_nodbg_operands(SrcReg)) {
    MachineInstr *UseMI = UseMO.getParent();
    if (UseMI == MI)
      continue;

    if (UseMI->isPHI()) {
      ExtendLife = false;
      continue;
    }

    // Only accept uses of SrcReg:SubIdx.
    if (UseSrcSubIdx && UseMO.getSubReg() != SubIdx)
      continue;

    // It's an error to translate this:
    //
    //    %reg1025 = <sext> %reg1024
    //     ...
    //    %reg1026 = SUBREG_TO_REG 0, %reg1024, 4
    //
    // into this:
    //
    //    %reg1025 = <sext> %reg1024
    //     ...
    //    %reg1027 = COPY %reg1025:4
    //    %reg1026 = SUBREG_TO_REG 0, %reg1027, 4
    //
    // The problem here is that SUBREG_TO_REG is there to assert that an
    // implicit zext occurs. It doesn't insert a zext instruction. If we allow
    // the COPY here, it will give us the value after the <sext>, not the
    // original value of %reg1024 before <sext>.
    if (UseMI->getOpcode() == TargetOpcode::SUBREG_TO_REG)
      continue;

    MachineBasicBlock *UseMBB = UseMI->getParent();
    if (UseMBB == MBB) {
      // Local uses that come after the extension.
      if (!LocalMIs.count(UseMI))
        Uses.push_back(&UseMO);
    } else if (ReachedBBs.count(UseMBB)) {
      // Non-local uses where the result of the extension is used. Always
      // replace these unless it's a PHI.
      Uses.push_back(&UseMO);
    } else if (Aggressive && DT->dominates(MBB, UseMBB)) {
      // We may want to extend the live range of the extension result in order
      // to replace these uses.
      ExtendedUses.push_back(&UseMO);
    } else {
      // Both will be live out of the def MBB anyway. Don't extend live range of
      // the extension result.
      ExtendLife = false;
      break;
    }
  }

  if (ExtendLife && !ExtendedUses.empty())
    // Extend the liveness of the extension result.
    std::copy(ExtendedUses.begin(), ExtendedUses.end(),
              std::back_inserter(Uses));

  // Now replace all uses.
  bool Changed = false;
  if (!Uses.empty()) {
    SmallPtrSet<MachineBasicBlock*, 4> PHIBBs;

    // Look for PHI uses of the extended result, we don't want to extend the
    // liveness of a PHI input. It breaks all kinds of assumptions down
    // stream. A PHI use is expected to be the kill of its source values.
    for (MachineInstr &UI : MRI->use_nodbg_instructions(DstReg))
      if (UI.isPHI())
        PHIBBs.insert(UI.getParent());

    const TargetRegisterClass *RC = MRI->getRegClass(SrcReg);
    for (unsigned i = 0, e = Uses.size(); i != e; ++i) {
      MachineOperand *UseMO = Uses[i];
      MachineInstr *UseMI = UseMO->getParent();
      MachineBasicBlock *UseMBB = UseMI->getParent();
      if (PHIBBs.count(UseMBB))
        continue;

      // About to add uses of DstReg, clear DstReg's kill flags.
      if (!Changed) {
        MRI->clearKillFlags(DstReg);
        MRI->constrainRegClass(DstReg, DstRC);
      }

      unsigned NewVR = MRI->createVirtualRegister(RC);
      MachineInstr *Copy = BuildMI(*UseMBB, UseMI, UseMI->getDebugLoc(),
                                   TII->get(TargetOpcode::COPY), NewVR)
        .addReg(DstReg, 0, SubIdx);
      // SubIdx applies to both SrcReg and DstReg when UseSrcSubIdx is set.
      if (UseSrcSubIdx) {
        Copy->getOperand(0).setSubReg(SubIdx);
        Copy->getOperand(0).setIsUndef();
      }
      UseMO->setReg(NewVR);
      ++NumReuse;
      Changed = true;
    }
  }

  return Changed;
}

/// optimizeCmpInstr - If the instruction is a compare and the previous
/// instruction it's comparing against all ready sets (or could be modified to
/// set) the same flag as the compare, then we can remove the comparison and use
/// the flag from the previous instruction.
bool PeepholeOptimizer::optimizeCmpInstr(MachineInstr *MI,
                                         MachineBasicBlock *MBB) {
  // If this instruction is a comparison against zero and isn't comparing a
  // physical register, we can try to optimize it.
  unsigned SrcReg, SrcReg2;
  int CmpMask, CmpValue;
  if (!TII->analyzeCompare(MI, SrcReg, SrcReg2, CmpMask, CmpValue) ||
      TargetRegisterInfo::isPhysicalRegister(SrcReg) ||
      (SrcReg2 != 0 && TargetRegisterInfo::isPhysicalRegister(SrcReg2)))
    return false;

  // Attempt to optimize the comparison instruction.
  if (TII->optimizeCompareInstr(MI, SrcReg, SrcReg2, CmpMask, CmpValue, MRI)) {
    ++NumCmps;
    return true;
  }

  return false;
}

/// Optimize a select instruction.
bool PeepholeOptimizer::optimizeSelect(MachineInstr *MI) {
  unsigned TrueOp = 0;
  unsigned FalseOp = 0;
  bool Optimizable = false;
  SmallVector<MachineOperand, 4> Cond;
  if (TII->analyzeSelect(MI, Cond, TrueOp, FalseOp, Optimizable))
    return false;
  if (!Optimizable)
    return false;
  if (!TII->optimizeSelect(MI))
    return false;
  MI->eraseFromParent();
  ++NumSelects;
  return true;
}

/// \brief Check if the registers defined by the pair (RegisterClass, SubReg)
/// share the same register file.
static bool shareSameRegisterFile(const TargetRegisterInfo &TRI,
                                  const TargetRegisterClass *DefRC,
                                  unsigned DefSubReg,
                                  const TargetRegisterClass *SrcRC,
                                  unsigned SrcSubReg) {
  // Same register class.
  if (DefRC == SrcRC)
    return true;

  // Both operands are sub registers. Check if they share a register class.
  unsigned SrcIdx, DefIdx;
  if (SrcSubReg && DefSubReg)
    return TRI.getCommonSuperRegClass(SrcRC, SrcSubReg, DefRC, DefSubReg,
                                      SrcIdx, DefIdx) != NULL;
  // At most one of the register is a sub register, make it Src to avoid
  // duplicating the test.
  if (!SrcSubReg) {
    std::swap(DefSubReg, SrcSubReg);
    std::swap(DefRC, SrcRC);
  }

  // One of the register is a sub register, check if we can get a superclass.
  if (SrcSubReg)
    return TRI.getMatchingSuperRegClass(SrcRC, DefRC, SrcSubReg) != NULL;
  // Plain copy.
  return TRI.getCommonSubClass(DefRC, SrcRC) != NULL;
}

/// \brief Get the index of the definition and source for \p Copy
/// instruction.
/// \pre Copy.isCopy() or Copy.isBitcast().
/// \return True if the Copy instruction has only one register source
/// and one register definition. Otherwise, \p DefIdx and \p SrcIdx
/// are invalid.
static bool getCopyOrBitcastDefUseIdx(const MachineInstr &Copy,
                                      unsigned &DefIdx, unsigned &SrcIdx) {
  assert((Copy.isCopy() || Copy.isBitcast()) && "Wrong operation type.");
  if (Copy.isCopy()) {
    // Copy instruction are supposed to be: Def = Src.
     if (Copy.getDesc().getNumOperands() != 2)
       return false;
     DefIdx = 0;
     SrcIdx = 1;
     assert(Copy.getOperand(DefIdx).isDef() && "Use comes before def!");
     return true;
  }
  // Bitcast case.
  // Bitcasts with more than one def are not supported.
  if (Copy.getDesc().getNumDefs() != 1)
    return false;
  // Initialize SrcIdx to an undefined operand.
  SrcIdx = Copy.getDesc().getNumOperands();
  for (unsigned OpIdx = 0, EndOpIdx = SrcIdx; OpIdx != EndOpIdx; ++OpIdx) {
    const MachineOperand &MO = Copy.getOperand(OpIdx);
    if (!MO.isReg() || !MO.getReg())
      continue;
    if (MO.isDef())
      DefIdx = OpIdx;
    else if (SrcIdx != EndOpIdx)
      // Multiple sources?
      return false;
    SrcIdx = OpIdx;
  }
  return true;
}

/// \brief Optimize a copy or bitcast instruction to avoid cross
/// register bank copy. The optimization looks through a chain of
/// copies and try to find a source that has a compatible register
/// class.
/// Two register classes are considered to be compatible if they share
/// the same register bank.
/// New copies issued by this optimization are register allocator
/// friendly. This optimization does not remove any copy as it may
/// overconstraint the register allocator, but replaces some when
/// possible.
/// \pre \p MI is a Copy (MI->isCopy() is true)
/// \return True, when \p MI has been optimized. In that case, \p MI has
/// been removed from its parent.
bool PeepholeOptimizer::optimizeCopyOrBitcast(MachineInstr *MI) {
  unsigned DefIdx, SrcIdx;
  if (!MI || !getCopyOrBitcastDefUseIdx(*MI, DefIdx, SrcIdx))
    return false;

  const MachineOperand &MODef = MI->getOperand(DefIdx);
  assert(MODef.isReg() && "Copies must be between registers.");
  unsigned Def = MODef.getReg();

  if (TargetRegisterInfo::isPhysicalRegister(Def))
    return false;

  const TargetRegisterClass *DefRC = MRI->getRegClass(Def);
  unsigned DefSubReg = MODef.getSubReg();

  unsigned Src;
  unsigned SrcSubReg;
  bool ShouldRewrite = false;
  MachineInstr *Copy = MI;
  const TargetRegisterInfo &TRI = *TM->getRegisterInfo();

  // Follow the chain of copies until we reach the top or find a
  // more suitable source.
  do {
    unsigned CopyDefIdx, CopySrcIdx;
    if (!getCopyOrBitcastDefUseIdx(*Copy, CopyDefIdx, CopySrcIdx))
      break;
    const MachineOperand &MO = Copy->getOperand(CopySrcIdx);
    assert(MO.isReg() && "Copies must be between registers.");
    Src = MO.getReg();

    if (TargetRegisterInfo::isPhysicalRegister(Src))
      break;

    const TargetRegisterClass *SrcRC = MRI->getRegClass(Src);
    SrcSubReg = MO.getSubReg();

    // If this source does not incur a cross register bank copy, use it.
    ShouldRewrite = shareSameRegisterFile(TRI, DefRC, DefSubReg, SrcRC,
                                          SrcSubReg);
    // Follow the chain of copies: get the definition of Src.
    Copy = MRI->getVRegDef(Src);
  } while (!ShouldRewrite && Copy && (Copy->isCopy() || Copy->isBitcast()));

  // If we did not find a more suitable source, there is nothing to optimize.
  if (!ShouldRewrite || Src == MI->getOperand(SrcIdx).getReg())
    return false;

  // Rewrite the copy to avoid a cross register bank penalty. 
  unsigned NewVR = TargetRegisterInfo::isPhysicalRegister(Def) ? Def :
    MRI->createVirtualRegister(DefRC);
  MachineInstr *NewCopy = BuildMI(*MI->getParent(), MI, MI->getDebugLoc(),
                                  TII->get(TargetOpcode::COPY), NewVR)
    .addReg(Src, 0, SrcSubReg);
  NewCopy->getOperand(0).setSubReg(DefSubReg);

  MRI->replaceRegWith(Def, NewVR);
  MRI->clearKillFlags(NewVR);
  MI->eraseFromParent();
  ++NumCopiesBitcasts;
  return true;
}

/// isLoadFoldable - Check whether MI is a candidate for folding into a later
/// instruction. We only fold loads to virtual registers and the virtual
/// register defined has a single use.
bool PeepholeOptimizer::isLoadFoldable(
                              MachineInstr *MI,
                              SmallSet<unsigned, 16> &FoldAsLoadDefCandidates) {
  if (!MI->canFoldAsLoad() || !MI->mayLoad())
    return false;
  const MCInstrDesc &MCID = MI->getDesc();
  if (MCID.getNumDefs() != 1)
    return false;

  unsigned Reg = MI->getOperand(0).getReg();
  // To reduce compilation time, we check MRI->hasOneNonDBGUse when inserting
  // loads. It should be checked when processing uses of the load, since
  // uses can be removed during peephole.
  if (!MI->getOperand(0).getSubReg() &&
      TargetRegisterInfo::isVirtualRegister(Reg) &&
      MRI->hasOneNonDBGUse(Reg)) {
    FoldAsLoadDefCandidates.insert(Reg);
    return true;
  }
  return false;
}

bool PeepholeOptimizer::isMoveImmediate(MachineInstr *MI,
                                        SmallSet<unsigned, 4> &ImmDefRegs,
                                 DenseMap<unsigned, MachineInstr*> &ImmDefMIs) {
  const MCInstrDesc &MCID = MI->getDesc();
  if (!MI->isMoveImmediate())
    return false;
  if (MCID.getNumDefs() != 1)
    return false;
  unsigned Reg = MI->getOperand(0).getReg();
  if (TargetRegisterInfo::isVirtualRegister(Reg)) {
    ImmDefMIs.insert(std::make_pair(Reg, MI));
    ImmDefRegs.insert(Reg);
    return true;
  }

  return false;
}

/// foldImmediate - Try folding register operands that are defined by move
/// immediate instructions, i.e. a trivial constant folding optimization, if
/// and only if the def and use are in the same BB.
bool PeepholeOptimizer::foldImmediate(MachineInstr *MI, MachineBasicBlock *MBB,
                                      SmallSet<unsigned, 4> &ImmDefRegs,
                                 DenseMap<unsigned, MachineInstr*> &ImmDefMIs) {
  for (unsigned i = 0, e = MI->getDesc().getNumOperands(); i != e; ++i) {
    MachineOperand &MO = MI->getOperand(i);
    if (!MO.isReg() || MO.isDef())
      continue;
    unsigned Reg = MO.getReg();
    if (!TargetRegisterInfo::isVirtualRegister(Reg))
      continue;
    if (ImmDefRegs.count(Reg) == 0)
      continue;
    DenseMap<unsigned, MachineInstr*>::iterator II = ImmDefMIs.find(Reg);
    assert(II != ImmDefMIs.end());
    if (TII->FoldImmediate(MI, II->second, Reg, MRI)) {
      ++NumImmFold;
      return true;
    }
  }
  return false;
}

bool PeepholeOptimizer::runOnMachineFunction(MachineFunction &MF) {
  if (skipOptnoneFunction(*MF.getFunction()))
    return false;

  DEBUG(dbgs() << "********** PEEPHOLE OPTIMIZER **********\n");
  DEBUG(dbgs() << "********** Function: " << MF.getName() << '\n');

  if (DisablePeephole)
    return false;

  TM  = &MF.getTarget();
  TII = TM->getInstrInfo();
  MRI = &MF.getRegInfo();
  DT  = Aggressive ? &getAnalysis<MachineDominatorTree>() : 0;

  bool Changed = false;

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

    bool SeenMoveImm = false;
    SmallPtrSet<MachineInstr*, 8> LocalMIs;
    SmallSet<unsigned, 4> ImmDefRegs;
    DenseMap<unsigned, MachineInstr*> ImmDefMIs;
    SmallSet<unsigned, 16> FoldAsLoadDefCandidates;

    for (MachineBasicBlock::iterator
           MII = I->begin(), MIE = I->end(); MII != MIE; ) {
      MachineInstr *MI = &*MII;
      // We may be erasing MI below, increment MII now.
      ++MII;
      LocalMIs.insert(MI);

      // Skip debug values. They should not affect this peephole optimization.
      if (MI->isDebugValue())
          continue;

      // If there exists an instruction which belongs to the following
      // categories, we will discard the load candidates.
      if (MI->isPosition() || MI->isPHI() || MI->isImplicitDef() ||
          MI->isKill() || MI->isInlineAsm() ||
          MI->hasUnmodeledSideEffects()) {
        FoldAsLoadDefCandidates.clear();
        continue;
      }
      if (MI->mayStore() || MI->isCall())
        FoldAsLoadDefCandidates.clear();

      if (((MI->isBitcast() || MI->isCopy()) && optimizeCopyOrBitcast(MI)) ||
          (MI->isCompare() && optimizeCmpInstr(MI, MBB)) ||
          (MI->isSelect() && optimizeSelect(MI))) {
        // MI is deleted.
        LocalMIs.erase(MI);
        Changed = true;
        continue;
      }

      if (isMoveImmediate(MI, ImmDefRegs, ImmDefMIs)) {
        SeenMoveImm = true;
      } else {
        Changed |= optimizeExtInstr(MI, MBB, LocalMIs);
        // optimizeExtInstr might have created new instructions after MI
        // and before the already incremented MII. Adjust MII so that the
        // next iteration sees the new instructions.
        MII = MI;
        ++MII;
        if (SeenMoveImm)
          Changed |= foldImmediate(MI, MBB, ImmDefRegs, ImmDefMIs);
      }

      // Check whether MI is a load candidate for folding into a later
      // instruction. If MI is not a candidate, check whether we can fold an
      // earlier load into MI.
      if (!isLoadFoldable(MI, FoldAsLoadDefCandidates) &&
          !FoldAsLoadDefCandidates.empty()) {
        const MCInstrDesc &MIDesc = MI->getDesc();
        for (unsigned i = MIDesc.getNumDefs(); i != MIDesc.getNumOperands();
             ++i) {
          const MachineOperand &MOp = MI->getOperand(i);
          if (!MOp.isReg())
            continue;
          unsigned FoldAsLoadDefReg = MOp.getReg();
          if (FoldAsLoadDefCandidates.count(FoldAsLoadDefReg)) {
            // We need to fold load after optimizeCmpInstr, since
            // optimizeCmpInstr can enable folding by converting SUB to CMP.
            // Save FoldAsLoadDefReg because optimizeLoadInstr() resets it and
            // we need it for markUsesInDebugValueAsUndef().
            unsigned FoldedReg = FoldAsLoadDefReg;
            MachineInstr *DefMI = 0;
            MachineInstr *FoldMI = TII->optimizeLoadInstr(MI, MRI,
                                                          FoldAsLoadDefReg,
                                                          DefMI);
            if (FoldMI) {
              // Update LocalMIs since we replaced MI with FoldMI and deleted
              // DefMI.
              DEBUG(dbgs() << "Replacing: " << *MI);
              DEBUG(dbgs() << "     With: " << *FoldMI);
              LocalMIs.erase(MI);
              LocalMIs.erase(DefMI);
              LocalMIs.insert(FoldMI);
              MI->eraseFromParent();
              DefMI->eraseFromParent();
              MRI->markUsesInDebugValueAsUndef(FoldedReg);
              FoldAsLoadDefCandidates.erase(FoldedReg);
              ++NumLoadFold;
              // MI is replaced with FoldMI.
              Changed = true;
              break;
            }
          }
        }
      }
    }
  }

  return Changed;
}