Subversion Repositories QNX 8.QNX8 LLVM/Clang compiler suite

Rev

Details | Last modification | View Log | RSS feed

Rev Author Line No. Line
14 pmbaty 1
//===- MemorySSA.h - Build Memory SSA ---------------------------*- C++ -*-===//
2
//
3
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4
// See https://llvm.org/LICENSE.txt for license information.
5
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6
//
7
//===----------------------------------------------------------------------===//
8
//
9
/// \file
10
/// This file exposes an interface to building/using memory SSA to
11
/// walk memory instructions using a use/def graph.
12
///
13
/// Memory SSA class builds an SSA form that links together memory access
14
/// instructions such as loads, stores, atomics, and calls. Additionally, it
15
/// does a trivial form of "heap versioning" Every time the memory state changes
16
/// in the program, we generate a new heap version. It generates
17
/// MemoryDef/Uses/Phis that are overlayed on top of the existing instructions.
18
///
19
/// As a trivial example,
20
/// define i32 @main() #0 {
21
/// entry:
22
///   %call = call noalias i8* @_Znwm(i64 4) #2
23
///   %0 = bitcast i8* %call to i32*
24
///   %call1 = call noalias i8* @_Znwm(i64 4) #2
25
///   %1 = bitcast i8* %call1 to i32*
26
///   store i32 5, i32* %0, align 4
27
///   store i32 7, i32* %1, align 4
28
///   %2 = load i32* %0, align 4
29
///   %3 = load i32* %1, align 4
30
///   %add = add nsw i32 %2, %3
31
///   ret i32 %add
32
/// }
33
///
34
/// Will become
35
/// define i32 @main() #0 {
36
/// entry:
37
///   ; 1 = MemoryDef(0)
38
///   %call = call noalias i8* @_Znwm(i64 4) #3
39
///   %2 = bitcast i8* %call to i32*
40
///   ; 2 = MemoryDef(1)
41
///   %call1 = call noalias i8* @_Znwm(i64 4) #3
42
///   %4 = bitcast i8* %call1 to i32*
43
///   ; 3 = MemoryDef(2)
44
///   store i32 5, i32* %2, align 4
45
///   ; 4 = MemoryDef(3)
46
///   store i32 7, i32* %4, align 4
47
///   ; MemoryUse(3)
48
///   %7 = load i32* %2, align 4
49
///   ; MemoryUse(4)
50
///   %8 = load i32* %4, align 4
51
///   %add = add nsw i32 %7, %8
52
///   ret i32 %add
53
/// }
54
///
55
/// Given this form, all the stores that could ever effect the load at %8 can be
56
/// gotten by using the MemoryUse associated with it, and walking from use to
57
/// def until you hit the top of the function.
58
///
59
/// Each def also has a list of users associated with it, so you can walk from
60
/// both def to users, and users to defs. Note that we disambiguate MemoryUses,
61
/// but not the RHS of MemoryDefs. You can see this above at %7, which would
62
/// otherwise be a MemoryUse(4). Being disambiguated means that for a given
63
/// store, all the MemoryUses on its use lists are may-aliases of that store
64
/// (but the MemoryDefs on its use list may not be).
65
///
66
/// MemoryDefs are not disambiguated because it would require multiple reaching
67
/// definitions, which would require multiple phis, and multiple memoryaccesses
68
/// per instruction.
69
///
70
/// In addition to the def/use graph described above, MemoryDefs also contain
71
/// an "optimized" definition use.  The "optimized" use points to some def
72
/// reachable through the memory def chain.  The optimized def *may* (but is
73
/// not required to) alias the original MemoryDef, but no def *closer* to the
74
/// source def may alias it.  As the name implies, the purpose of the optimized
75
/// use is to allow caching of clobber searches for memory defs.  The optimized
76
/// def may be nullptr, in which case clients must walk the defining access
77
/// chain.
78
///
79
/// When iterating the uses of a MemoryDef, both defining uses and optimized
80
/// uses will be encountered.  If only one type is needed, the client must
81
/// filter the use walk.
82
//
83
//===----------------------------------------------------------------------===//
84
 
85
#ifndef LLVM_ANALYSIS_MEMORYSSA_H
86
#define LLVM_ANALYSIS_MEMORYSSA_H
87
 
88
#include "llvm/ADT/DenseMap.h"
89
#include "llvm/ADT/SmallPtrSet.h"
90
#include "llvm/ADT/SmallVector.h"
91
#include "llvm/ADT/ilist_node.h"
92
#include "llvm/ADT/iterator_range.h"
93
#include "llvm/Analysis/AliasAnalysis.h"
94
#include "llvm/Analysis/MemoryLocation.h"
95
#include "llvm/Analysis/PHITransAddr.h"
96
#include "llvm/IR/DerivedUser.h"
97
#include "llvm/IR/Dominators.h"
98
#include "llvm/IR/Type.h"
99
#include "llvm/IR/User.h"
100
#include "llvm/Pass.h"
101
#include <algorithm>
102
#include <cassert>
103
#include <cstddef>
104
#include <iterator>
105
#include <memory>
106
#include <utility>
107
 
108
namespace llvm {
109
 
110
template <class GraphType> struct GraphTraits;
111
class BasicBlock;
112
class Function;
113
class Instruction;
114
class LLVMContext;
115
class MemoryAccess;
116
class MemorySSAWalker;
117
class Module;
118
class Use;
119
class Value;
120
class raw_ostream;
121
 
122
namespace MSSAHelpers {
123
 
124
struct AllAccessTag {};
125
struct DefsOnlyTag {};
126
 
127
} // end namespace MSSAHelpers
128
 
129
enum : unsigned {
130
  // Used to signify what the default invalid ID is for MemoryAccess's
131
  // getID()
132
  INVALID_MEMORYACCESS_ID = -1U
133
};
134
 
135
template <class T> class memoryaccess_def_iterator_base;
136
using memoryaccess_def_iterator = memoryaccess_def_iterator_base<MemoryAccess>;
137
using const_memoryaccess_def_iterator =
138
    memoryaccess_def_iterator_base<const MemoryAccess>;
139
 
140
// The base for all memory accesses. All memory accesses in a block are
141
// linked together using an intrusive list.
142
class MemoryAccess
143
    : public DerivedUser,
144
      public ilist_node<MemoryAccess, ilist_tag<MSSAHelpers::AllAccessTag>>,
145
      public ilist_node<MemoryAccess, ilist_tag<MSSAHelpers::DefsOnlyTag>> {
146
public:
147
  using AllAccessType =
148
      ilist_node<MemoryAccess, ilist_tag<MSSAHelpers::AllAccessTag>>;
149
  using DefsOnlyType =
150
      ilist_node<MemoryAccess, ilist_tag<MSSAHelpers::DefsOnlyTag>>;
151
 
152
  MemoryAccess(const MemoryAccess &) = delete;
153
  MemoryAccess &operator=(const MemoryAccess &) = delete;
154
 
155
  void *operator new(size_t) = delete;
156
 
157
  // Methods for support type inquiry through isa, cast, and
158
  // dyn_cast
159
  static bool classof(const Value *V) {
160
    unsigned ID = V->getValueID();
161
    return ID == MemoryUseVal || ID == MemoryPhiVal || ID == MemoryDefVal;
162
  }
163
 
164
  BasicBlock *getBlock() const { return Block; }
165
 
166
  void print(raw_ostream &OS) const;
167
  void dump() const;
168
 
169
  /// The user iterators for a memory access
170
  using iterator = user_iterator;
171
  using const_iterator = const_user_iterator;
172
 
173
  /// This iterator walks over all of the defs in a given
174
  /// MemoryAccess. For MemoryPhi nodes, this walks arguments. For
175
  /// MemoryUse/MemoryDef, this walks the defining access.
176
  memoryaccess_def_iterator defs_begin();
177
  const_memoryaccess_def_iterator defs_begin() const;
178
  memoryaccess_def_iterator defs_end();
179
  const_memoryaccess_def_iterator defs_end() const;
180
 
181
  /// Get the iterators for the all access list and the defs only list
182
  /// We default to the all access list.
183
  AllAccessType::self_iterator getIterator() {
184
    return this->AllAccessType::getIterator();
185
  }
186
  AllAccessType::const_self_iterator getIterator() const {
187
    return this->AllAccessType::getIterator();
188
  }
189
  AllAccessType::reverse_self_iterator getReverseIterator() {
190
    return this->AllAccessType::getReverseIterator();
191
  }
192
  AllAccessType::const_reverse_self_iterator getReverseIterator() const {
193
    return this->AllAccessType::getReverseIterator();
194
  }
195
  DefsOnlyType::self_iterator getDefsIterator() {
196
    return this->DefsOnlyType::getIterator();
197
  }
198
  DefsOnlyType::const_self_iterator getDefsIterator() const {
199
    return this->DefsOnlyType::getIterator();
200
  }
201
  DefsOnlyType::reverse_self_iterator getReverseDefsIterator() {
202
    return this->DefsOnlyType::getReverseIterator();
203
  }
204
  DefsOnlyType::const_reverse_self_iterator getReverseDefsIterator() const {
205
    return this->DefsOnlyType::getReverseIterator();
206
  }
207
 
208
protected:
209
  friend class MemoryDef;
210
  friend class MemoryPhi;
211
  friend class MemorySSA;
212
  friend class MemoryUse;
213
  friend class MemoryUseOrDef;
214
 
215
  /// Used by MemorySSA to change the block of a MemoryAccess when it is
216
  /// moved.
217
  void setBlock(BasicBlock *BB) { Block = BB; }
218
 
219
  /// Used for debugging and tracking things about MemoryAccesses.
220
  /// Guaranteed unique among MemoryAccesses, no guarantees otherwise.
221
  inline unsigned getID() const;
222
 
223
  MemoryAccess(LLVMContext &C, unsigned Vty, DeleteValueTy DeleteValue,
224
               BasicBlock *BB, unsigned NumOperands)
225
      : DerivedUser(Type::getVoidTy(C), Vty, nullptr, NumOperands, DeleteValue),
226
        Block(BB) {}
227
 
228
  // Use deleteValue() to delete a generic MemoryAccess.
229
  ~MemoryAccess() = default;
230
 
231
private:
232
  BasicBlock *Block;
233
};
234
 
235
template <>
236
struct ilist_alloc_traits<MemoryAccess> {
237
  static void deleteNode(MemoryAccess *MA) { MA->deleteValue(); }
238
};
239
 
240
inline raw_ostream &operator<<(raw_ostream &OS, const MemoryAccess &MA) {
241
  MA.print(OS);
242
  return OS;
243
}
244
 
245
/// Class that has the common methods + fields of memory uses/defs. It's
246
/// a little awkward to have, but there are many cases where we want either a
247
/// use or def, and there are many cases where uses are needed (defs aren't
248
/// acceptable), and vice-versa.
249
///
250
/// This class should never be instantiated directly; make a MemoryUse or
251
/// MemoryDef instead.
252
class MemoryUseOrDef : public MemoryAccess {
253
public:
254
  void *operator new(size_t) = delete;
255
 
256
  DECLARE_TRANSPARENT_OPERAND_ACCESSORS(MemoryAccess);
257
 
258
  /// Get the instruction that this MemoryUse represents.
259
  Instruction *getMemoryInst() const { return MemoryInstruction; }
260
 
261
  /// Get the access that produces the memory state used by this Use.
262
  MemoryAccess *getDefiningAccess() const { return getOperand(0); }
263
 
264
  static bool classof(const Value *MA) {
265
    return MA->getValueID() == MemoryUseVal || MA->getValueID() == MemoryDefVal;
266
  }
267
 
268
  /// Do we have an optimized use?
269
  inline bool isOptimized() const;
270
  /// Return the MemoryAccess associated with the optimized use, or nullptr.
271
  inline MemoryAccess *getOptimized() const;
272
  /// Sets the optimized use for a MemoryDef.
273
  inline void setOptimized(MemoryAccess *);
274
 
275
  /// Reset the ID of what this MemoryUse was optimized to, causing it to
276
  /// be rewalked by the walker if necessary.
277
  /// This really should only be called by tests.
278
  inline void resetOptimized();
279
 
280
protected:
281
  friend class MemorySSA;
282
  friend class MemorySSAUpdater;
283
 
284
  MemoryUseOrDef(LLVMContext &C, MemoryAccess *DMA, unsigned Vty,
285
                 DeleteValueTy DeleteValue, Instruction *MI, BasicBlock *BB,
286
                 unsigned NumOperands)
287
      : MemoryAccess(C, Vty, DeleteValue, BB, NumOperands),
288
        MemoryInstruction(MI) {
289
    setDefiningAccess(DMA);
290
  }
291
 
292
  // Use deleteValue() to delete a generic MemoryUseOrDef.
293
  ~MemoryUseOrDef() = default;
294
 
295
  void setDefiningAccess(MemoryAccess *DMA, bool Optimized = false) {
296
    if (!Optimized) {
297
      setOperand(0, DMA);
298
      return;
299
    }
300
    setOptimized(DMA);
301
  }
302
 
303
private:
304
  Instruction *MemoryInstruction;
305
};
306
 
307
/// Represents read-only accesses to memory
308
///
309
/// In particular, the set of Instructions that will be represented by
310
/// MemoryUse's is exactly the set of Instructions for which
311
/// AliasAnalysis::getModRefInfo returns "Ref".
312
class MemoryUse final : public MemoryUseOrDef {
313
public:
314
  DECLARE_TRANSPARENT_OPERAND_ACCESSORS(MemoryAccess);
315
 
316
  MemoryUse(LLVMContext &C, MemoryAccess *DMA, Instruction *MI, BasicBlock *BB)
317
      : MemoryUseOrDef(C, DMA, MemoryUseVal, deleteMe, MI, BB,
318
                       /*NumOperands=*/1) {}
319
 
320
  // allocate space for exactly one operand
321
  void *operator new(size_t S) { return User::operator new(S, 1); }
322
  void operator delete(void *Ptr) { User::operator delete(Ptr); }
323
 
324
  static bool classof(const Value *MA) {
325
    return MA->getValueID() == MemoryUseVal;
326
  }
327
 
328
  void print(raw_ostream &OS) const;
329
 
330
  void setOptimized(MemoryAccess *DMA) {
331
    OptimizedID = DMA->getID();
332
    setOperand(0, DMA);
333
  }
334
 
335
  /// Whether the MemoryUse is optimized. If ensureOptimizedUses() was called,
336
  /// uses will usually be optimized, but this is not guaranteed (e.g. due to
337
  /// invalidation and optimization limits.)
338
  bool isOptimized() const {
339
    return getDefiningAccess() && OptimizedID == getDefiningAccess()->getID();
340
  }
341
 
342
  MemoryAccess *getOptimized() const {
343
    return getDefiningAccess();
344
  }
345
 
346
  void resetOptimized() {
347
    OptimizedID = INVALID_MEMORYACCESS_ID;
348
  }
349
 
350
protected:
351
  friend class MemorySSA;
352
 
353
private:
354
  static void deleteMe(DerivedUser *Self);
355
 
356
  unsigned OptimizedID = INVALID_MEMORYACCESS_ID;
357
};
358
 
359
template <>
360
struct OperandTraits<MemoryUse> : public FixedNumOperandTraits<MemoryUse, 1> {};
361
DEFINE_TRANSPARENT_OPERAND_ACCESSORS(MemoryUse, MemoryAccess)
362
 
363
/// Represents a read-write access to memory, whether it is a must-alias,
364
/// or a may-alias.
365
///
366
/// In particular, the set of Instructions that will be represented by
367
/// MemoryDef's is exactly the set of Instructions for which
368
/// AliasAnalysis::getModRefInfo returns "Mod" or "ModRef".
369
/// Note that, in order to provide def-def chains, all defs also have a use
370
/// associated with them. This use points to the nearest reaching
371
/// MemoryDef/MemoryPhi.
372
class MemoryDef final : public MemoryUseOrDef {
373
public:
374
  friend class MemorySSA;
375
 
376
  DECLARE_TRANSPARENT_OPERAND_ACCESSORS(MemoryAccess);
377
 
378
  MemoryDef(LLVMContext &C, MemoryAccess *DMA, Instruction *MI, BasicBlock *BB,
379
            unsigned Ver)
380
      : MemoryUseOrDef(C, DMA, MemoryDefVal, deleteMe, MI, BB,
381
                       /*NumOperands=*/2),
382
        ID(Ver) {}
383
 
384
  // allocate space for exactly two operands
385
  void *operator new(size_t S) { return User::operator new(S, 2); }
386
  void operator delete(void *Ptr) { User::operator delete(Ptr); }
387
 
388
  static bool classof(const Value *MA) {
389
    return MA->getValueID() == MemoryDefVal;
390
  }
391
 
392
  void setOptimized(MemoryAccess *MA) {
393
    setOperand(1, MA);
394
    OptimizedID = MA->getID();
395
  }
396
 
397
  MemoryAccess *getOptimized() const {
398
    return cast_or_null<MemoryAccess>(getOperand(1));
399
  }
400
 
401
  bool isOptimized() const {
402
    return getOptimized() && OptimizedID == getOptimized()->getID();
403
  }
404
 
405
  void resetOptimized() {
406
    OptimizedID = INVALID_MEMORYACCESS_ID;
407
    setOperand(1, nullptr);
408
  }
409
 
410
  void print(raw_ostream &OS) const;
411
 
412
  unsigned getID() const { return ID; }
413
 
414
private:
415
  static void deleteMe(DerivedUser *Self);
416
 
417
  const unsigned ID;
418
  unsigned OptimizedID = INVALID_MEMORYACCESS_ID;
419
};
420
 
421
template <>
422
struct OperandTraits<MemoryDef> : public FixedNumOperandTraits<MemoryDef, 2> {};
423
DEFINE_TRANSPARENT_OPERAND_ACCESSORS(MemoryDef, MemoryAccess)
424
 
425
template <>
426
struct OperandTraits<MemoryUseOrDef> {
427
  static Use *op_begin(MemoryUseOrDef *MUD) {
428
    if (auto *MU = dyn_cast<MemoryUse>(MUD))
429
      return OperandTraits<MemoryUse>::op_begin(MU);
430
    return OperandTraits<MemoryDef>::op_begin(cast<MemoryDef>(MUD));
431
  }
432
 
433
  static Use *op_end(MemoryUseOrDef *MUD) {
434
    if (auto *MU = dyn_cast<MemoryUse>(MUD))
435
      return OperandTraits<MemoryUse>::op_end(MU);
436
    return OperandTraits<MemoryDef>::op_end(cast<MemoryDef>(MUD));
437
  }
438
 
439
  static unsigned operands(const MemoryUseOrDef *MUD) {
440
    if (const auto *MU = dyn_cast<MemoryUse>(MUD))
441
      return OperandTraits<MemoryUse>::operands(MU);
442
    return OperandTraits<MemoryDef>::operands(cast<MemoryDef>(MUD));
443
  }
444
};
445
DEFINE_TRANSPARENT_OPERAND_ACCESSORS(MemoryUseOrDef, MemoryAccess)
446
 
447
/// Represents phi nodes for memory accesses.
448
///
449
/// These have the same semantic as regular phi nodes, with the exception that
450
/// only one phi will ever exist in a given basic block.
451
/// Guaranteeing one phi per block means guaranteeing there is only ever one
452
/// valid reaching MemoryDef/MemoryPHI along each path to the phi node.
453
/// This is ensured by not allowing disambiguation of the RHS of a MemoryDef or
454
/// a MemoryPhi's operands.
455
/// That is, given
456
/// if (a) {
457
///   store %a
458
///   store %b
459
/// }
460
/// it *must* be transformed into
461
/// if (a) {
462
///    1 = MemoryDef(liveOnEntry)
463
///    store %a
464
///    2 = MemoryDef(1)
465
///    store %b
466
/// }
467
/// and *not*
468
/// if (a) {
469
///    1 = MemoryDef(liveOnEntry)
470
///    store %a
471
///    2 = MemoryDef(liveOnEntry)
472
///    store %b
473
/// }
474
/// even if the two stores do not conflict. Otherwise, both 1 and 2 reach the
475
/// end of the branch, and if there are not two phi nodes, one will be
476
/// disconnected completely from the SSA graph below that point.
477
/// Because MemoryUse's do not generate new definitions, they do not have this
478
/// issue.
479
class MemoryPhi final : public MemoryAccess {
480
  // allocate space for exactly zero operands
481
  void *operator new(size_t S) { return User::operator new(S); }
482
 
483
public:
484
  void operator delete(void *Ptr) { User::operator delete(Ptr); }
485
 
486
  /// Provide fast operand accessors
487
  DECLARE_TRANSPARENT_OPERAND_ACCESSORS(MemoryAccess);
488
 
489
  MemoryPhi(LLVMContext &C, BasicBlock *BB, unsigned Ver, unsigned NumPreds = 0)
490
      : MemoryAccess(C, MemoryPhiVal, deleteMe, BB, 0), ID(Ver),
491
        ReservedSpace(NumPreds) {
492
    allocHungoffUses(ReservedSpace);
493
  }
494
 
495
  // Block iterator interface. This provides access to the list of incoming
496
  // basic blocks, which parallels the list of incoming values.
497
  using block_iterator = BasicBlock **;
498
  using const_block_iterator = BasicBlock *const *;
499
 
500
  block_iterator block_begin() {
501
    return reinterpret_cast<block_iterator>(op_begin() + ReservedSpace);
502
  }
503
 
504
  const_block_iterator block_begin() const {
505
    return reinterpret_cast<const_block_iterator>(op_begin() + ReservedSpace);
506
  }
507
 
508
  block_iterator block_end() { return block_begin() + getNumOperands(); }
509
 
510
  const_block_iterator block_end() const {
511
    return block_begin() + getNumOperands();
512
  }
513
 
514
  iterator_range<block_iterator> blocks() {
515
    return make_range(block_begin(), block_end());
516
  }
517
 
518
  iterator_range<const_block_iterator> blocks() const {
519
    return make_range(block_begin(), block_end());
520
  }
521
 
522
  op_range incoming_values() { return operands(); }
523
 
524
  const_op_range incoming_values() const { return operands(); }
525
 
526
  /// Return the number of incoming edges
527
  unsigned getNumIncomingValues() const { return getNumOperands(); }
528
 
529
  /// Return incoming value number x
530
  MemoryAccess *getIncomingValue(unsigned I) const { return getOperand(I); }
531
  void setIncomingValue(unsigned I, MemoryAccess *V) {
532
    assert(V && "PHI node got a null value!");
533
    setOperand(I, V);
534
  }
535
 
536
  static unsigned getOperandNumForIncomingValue(unsigned I) { return I; }
537
  static unsigned getIncomingValueNumForOperand(unsigned I) { return I; }
538
 
539
  /// Return incoming basic block number @p i.
540
  BasicBlock *getIncomingBlock(unsigned I) const { return block_begin()[I]; }
541
 
542
  /// Return incoming basic block corresponding
543
  /// to an operand of the PHI.
544
  BasicBlock *getIncomingBlock(const Use &U) const {
545
    assert(this == U.getUser() && "Iterator doesn't point to PHI's Uses?");
546
    return getIncomingBlock(unsigned(&U - op_begin()));
547
  }
548
 
549
  /// Return incoming basic block corresponding
550
  /// to value use iterator.
551
  BasicBlock *getIncomingBlock(MemoryAccess::const_user_iterator I) const {
552
    return getIncomingBlock(I.getUse());
553
  }
554
 
555
  void setIncomingBlock(unsigned I, BasicBlock *BB) {
556
    assert(BB && "PHI node got a null basic block!");
557
    block_begin()[I] = BB;
558
  }
559
 
560
  /// Add an incoming value to the end of the PHI list
561
  void addIncoming(MemoryAccess *V, BasicBlock *BB) {
562
    if (getNumOperands() == ReservedSpace)
563
      growOperands(); // Get more space!
564
    // Initialize some new operands.
565
    setNumHungOffUseOperands(getNumOperands() + 1);
566
    setIncomingValue(getNumOperands() - 1, V);
567
    setIncomingBlock(getNumOperands() - 1, BB);
568
  }
569
 
570
  /// Return the first index of the specified basic
571
  /// block in the value list for this PHI.  Returns -1 if no instance.
572
  int getBasicBlockIndex(const BasicBlock *BB) const {
573
    for (unsigned I = 0, E = getNumOperands(); I != E; ++I)
574
      if (block_begin()[I] == BB)
575
        return I;
576
    return -1;
577
  }
578
 
579
  MemoryAccess *getIncomingValueForBlock(const BasicBlock *BB) const {
580
    int Idx = getBasicBlockIndex(BB);
581
    assert(Idx >= 0 && "Invalid basic block argument!");
582
    return getIncomingValue(Idx);
583
  }
584
 
585
  // After deleting incoming position I, the order of incoming may be changed.
586
  void unorderedDeleteIncoming(unsigned I) {
587
    unsigned E = getNumOperands();
588
    assert(I < E && "Cannot remove out of bounds Phi entry.");
589
    // MemoryPhi must have at least two incoming values, otherwise the MemoryPhi
590
    // itself should be deleted.
591
    assert(E >= 2 && "Cannot only remove incoming values in MemoryPhis with "
592
                     "at least 2 values.");
593
    setIncomingValue(I, getIncomingValue(E - 1));
594
    setIncomingBlock(I, block_begin()[E - 1]);
595
    setOperand(E - 1, nullptr);
596
    block_begin()[E - 1] = nullptr;
597
    setNumHungOffUseOperands(getNumOperands() - 1);
598
  }
599
 
600
  // After deleting entries that satisfy Pred, remaining entries may have
601
  // changed order.
602
  template <typename Fn> void unorderedDeleteIncomingIf(Fn &&Pred) {
603
    for (unsigned I = 0, E = getNumOperands(); I != E; ++I)
604
      if (Pred(getIncomingValue(I), getIncomingBlock(I))) {
605
        unorderedDeleteIncoming(I);
606
        E = getNumOperands();
607
        --I;
608
      }
609
    assert(getNumOperands() >= 1 &&
610
           "Cannot remove all incoming blocks in a MemoryPhi.");
611
  }
612
 
613
  // After deleting incoming block BB, the incoming blocks order may be changed.
614
  void unorderedDeleteIncomingBlock(const BasicBlock *BB) {
615
    unorderedDeleteIncomingIf(
616
        [&](const MemoryAccess *, const BasicBlock *B) { return BB == B; });
617
  }
618
 
619
  // After deleting incoming memory access MA, the incoming accesses order may
620
  // be changed.
621
  void unorderedDeleteIncomingValue(const MemoryAccess *MA) {
622
    unorderedDeleteIncomingIf(
623
        [&](const MemoryAccess *M, const BasicBlock *) { return MA == M; });
624
  }
625
 
626
  static bool classof(const Value *V) {
627
    return V->getValueID() == MemoryPhiVal;
628
  }
629
 
630
  void print(raw_ostream &OS) const;
631
 
632
  unsigned getID() const { return ID; }
633
 
634
protected:
635
  friend class MemorySSA;
636
 
637
  /// this is more complicated than the generic
638
  /// User::allocHungoffUses, because we have to allocate Uses for the incoming
639
  /// values and pointers to the incoming blocks, all in one allocation.
640
  void allocHungoffUses(unsigned N) {
641
    User::allocHungoffUses(N, /* IsPhi */ true);
642
  }
643
 
644
private:
645
  // For debugging only
646
  const unsigned ID;
647
  unsigned ReservedSpace;
648
 
649
  /// This grows the operand list in response to a push_back style of
650
  /// operation.  This grows the number of ops by 1.5 times.
651
  void growOperands() {
652
    unsigned E = getNumOperands();
653
    // 2 op PHI nodes are VERY common, so reserve at least enough for that.
654
    ReservedSpace = std::max(E + E / 2, 2u);
655
    growHungoffUses(ReservedSpace, /* IsPhi */ true);
656
  }
657
 
658
  static void deleteMe(DerivedUser *Self);
659
};
660
 
661
inline unsigned MemoryAccess::getID() const {
662
  assert((isa<MemoryDef>(this) || isa<MemoryPhi>(this)) &&
663
         "only memory defs and phis have ids");
664
  if (const auto *MD = dyn_cast<MemoryDef>(this))
665
    return MD->getID();
666
  return cast<MemoryPhi>(this)->getID();
667
}
668
 
669
inline bool MemoryUseOrDef::isOptimized() const {
670
  if (const auto *MD = dyn_cast<MemoryDef>(this))
671
    return MD->isOptimized();
672
  return cast<MemoryUse>(this)->isOptimized();
673
}
674
 
675
inline MemoryAccess *MemoryUseOrDef::getOptimized() const {
676
  if (const auto *MD = dyn_cast<MemoryDef>(this))
677
    return MD->getOptimized();
678
  return cast<MemoryUse>(this)->getOptimized();
679
}
680
 
681
inline void MemoryUseOrDef::setOptimized(MemoryAccess *MA) {
682
  if (auto *MD = dyn_cast<MemoryDef>(this))
683
    MD->setOptimized(MA);
684
  else
685
    cast<MemoryUse>(this)->setOptimized(MA);
686
}
687
 
688
inline void MemoryUseOrDef::resetOptimized() {
689
  if (auto *MD = dyn_cast<MemoryDef>(this))
690
    MD->resetOptimized();
691
  else
692
    cast<MemoryUse>(this)->resetOptimized();
693
}
694
 
695
template <> struct OperandTraits<MemoryPhi> : public HungoffOperandTraits<2> {};
696
DEFINE_TRANSPARENT_OPERAND_ACCESSORS(MemoryPhi, MemoryAccess)
697
 
698
/// Encapsulates MemorySSA, including all data associated with memory
699
/// accesses.
700
class MemorySSA {
701
public:
702
  MemorySSA(Function &, AliasAnalysis *, DominatorTree *);
703
 
704
  // MemorySSA must remain where it's constructed; Walkers it creates store
705
  // pointers to it.
706
  MemorySSA(MemorySSA &&) = delete;
707
 
708
  ~MemorySSA();
709
 
710
  MemorySSAWalker *getWalker();
711
  MemorySSAWalker *getSkipSelfWalker();
712
 
713
  /// Given a memory Mod/Ref'ing instruction, get the MemorySSA
714
  /// access associated with it. If passed a basic block gets the memory phi
715
  /// node that exists for that block, if there is one. Otherwise, this will get
716
  /// a MemoryUseOrDef.
717
  MemoryUseOrDef *getMemoryAccess(const Instruction *I) const {
718
    return cast_or_null<MemoryUseOrDef>(ValueToMemoryAccess.lookup(I));
719
  }
720
 
721
  MemoryPhi *getMemoryAccess(const BasicBlock *BB) const {
722
    return cast_or_null<MemoryPhi>(ValueToMemoryAccess.lookup(cast<Value>(BB)));
723
  }
724
 
725
  DominatorTree &getDomTree() const { return *DT; }
726
 
727
  void dump() const;
728
  void print(raw_ostream &) const;
729
 
730
  /// Return true if \p MA represents the live on entry value
731
  ///
732
  /// Loads and stores from pointer arguments and other global values may be
733
  /// defined by memory operations that do not occur in the current function, so
734
  /// they may be live on entry to the function. MemorySSA represents such
735
  /// memory state by the live on entry definition, which is guaranteed to occur
736
  /// before any other memory access in the function.
737
  inline bool isLiveOnEntryDef(const MemoryAccess *MA) const {
738
    return MA == LiveOnEntryDef.get();
739
  }
740
 
741
  inline MemoryAccess *getLiveOnEntryDef() const {
742
    return LiveOnEntryDef.get();
743
  }
744
 
745
  // Sadly, iplists, by default, owns and deletes pointers added to the
746
  // list. It's not currently possible to have two iplists for the same type,
747
  // where one owns the pointers, and one does not. This is because the traits
748
  // are per-type, not per-tag.  If this ever changes, we should make the
749
  // DefList an iplist.
750
  using AccessList = iplist<MemoryAccess, ilist_tag<MSSAHelpers::AllAccessTag>>;
751
  using DefsList =
752
      simple_ilist<MemoryAccess, ilist_tag<MSSAHelpers::DefsOnlyTag>>;
753
 
754
  /// Return the list of MemoryAccess's for a given basic block.
755
  ///
756
  /// This list is not modifiable by the user.
757
  const AccessList *getBlockAccesses(const BasicBlock *BB) const {
758
    return getWritableBlockAccesses(BB);
759
  }
760
 
761
  /// Return the list of MemoryDef's and MemoryPhi's for a given basic
762
  /// block.
763
  ///
764
  /// This list is not modifiable by the user.
765
  const DefsList *getBlockDefs(const BasicBlock *BB) const {
766
    return getWritableBlockDefs(BB);
767
  }
768
 
769
  /// Given two memory accesses in the same basic block, determine
770
  /// whether MemoryAccess \p A dominates MemoryAccess \p B.
771
  bool locallyDominates(const MemoryAccess *A, const MemoryAccess *B) const;
772
 
773
  /// Given two memory accesses in potentially different blocks,
774
  /// determine whether MemoryAccess \p A dominates MemoryAccess \p B.
775
  bool dominates(const MemoryAccess *A, const MemoryAccess *B) const;
776
 
777
  /// Given a MemoryAccess and a Use, determine whether MemoryAccess \p A
778
  /// dominates Use \p B.
779
  bool dominates(const MemoryAccess *A, const Use &B) const;
780
 
781
  enum class VerificationLevel { Fast, Full };
782
  /// Verify that MemorySSA is self consistent (IE definitions dominate
783
  /// all uses, uses appear in the right places).  This is used by unit tests.
784
  void verifyMemorySSA(VerificationLevel = VerificationLevel::Fast) const;
785
 
786
  /// Used in various insertion functions to specify whether we are talking
787
  /// about the beginning or end of a block.
788
  enum InsertionPlace { Beginning, End, BeforeTerminator };
789
 
790
  /// By default, uses are *not* optimized during MemorySSA construction.
791
  /// Calling this method will attempt to optimize all MemoryUses, if this has
792
  /// not happened yet for this MemorySSA instance. This should be done if you
793
  /// plan to query the clobbering access for most uses, or if you walk the
794
  /// def-use chain of uses.
795
  void ensureOptimizedUses();
796
 
797
  AliasAnalysis &getAA() { return *AA; }
798
 
799
protected:
800
  // Used by Memory SSA dumpers and wrapper pass
801
  friend class MemorySSAPrinterLegacyPass;
802
  friend class MemorySSAUpdater;
803
 
804
  void verifyOrderingDominationAndDefUses(
805
      Function &F, VerificationLevel = VerificationLevel::Fast) const;
806
  void verifyDominationNumbers(const Function &F) const;
807
  void verifyPrevDefInPhis(Function &F) const;
808
 
809
  // This is used by the use optimizer and updater.
810
  AccessList *getWritableBlockAccesses(const BasicBlock *BB) const {
811
    auto It = PerBlockAccesses.find(BB);
812
    return It == PerBlockAccesses.end() ? nullptr : It->second.get();
813
  }
814
 
815
  // This is used by the use optimizer and updater.
816
  DefsList *getWritableBlockDefs(const BasicBlock *BB) const {
817
    auto It = PerBlockDefs.find(BB);
818
    return It == PerBlockDefs.end() ? nullptr : It->second.get();
819
  }
820
 
821
  // These is used by the updater to perform various internal MemorySSA
822
  // machinsations.  They do not always leave the IR in a correct state, and
823
  // relies on the updater to fixup what it breaks, so it is not public.
824
 
825
  void moveTo(MemoryUseOrDef *What, BasicBlock *BB, AccessList::iterator Where);
826
  void moveTo(MemoryAccess *What, BasicBlock *BB, InsertionPlace Point);
827
 
828
  // Rename the dominator tree branch rooted at BB.
829
  void renamePass(BasicBlock *BB, MemoryAccess *IncomingVal,
830
                  SmallPtrSetImpl<BasicBlock *> &Visited) {
831
    renamePass(DT->getNode(BB), IncomingVal, Visited, true, true);
832
  }
833
 
834
  void removeFromLookups(MemoryAccess *);
835
  void removeFromLists(MemoryAccess *, bool ShouldDelete = true);
836
  void insertIntoListsForBlock(MemoryAccess *, const BasicBlock *,
837
                               InsertionPlace);
838
  void insertIntoListsBefore(MemoryAccess *, const BasicBlock *,
839
                             AccessList::iterator);
840
  MemoryUseOrDef *createDefinedAccess(Instruction *, MemoryAccess *,
841
                                      const MemoryUseOrDef *Template = nullptr,
842
                                      bool CreationMustSucceed = true);
843
 
844
private:
845
  class ClobberWalkerBase;
846
  class CachingWalker;
847
  class SkipSelfWalker;
848
  class OptimizeUses;
849
 
850
  CachingWalker *getWalkerImpl();
851
  void buildMemorySSA(BatchAAResults &BAA);
852
 
853
  void prepareForMoveTo(MemoryAccess *, BasicBlock *);
854
  void verifyUseInDefs(MemoryAccess *, MemoryAccess *) const;
855
 
856
  using AccessMap = DenseMap<const BasicBlock *, std::unique_ptr<AccessList>>;
857
  using DefsMap = DenseMap<const BasicBlock *, std::unique_ptr<DefsList>>;
858
 
859
  void markUnreachableAsLiveOnEntry(BasicBlock *BB);
860
  MemoryPhi *createMemoryPhi(BasicBlock *BB);
861
  template <typename AliasAnalysisType>
862
  MemoryUseOrDef *createNewAccess(Instruction *, AliasAnalysisType *,
863
                                  const MemoryUseOrDef *Template = nullptr);
864
  void placePHINodes(const SmallPtrSetImpl<BasicBlock *> &);
865
  MemoryAccess *renameBlock(BasicBlock *, MemoryAccess *, bool);
866
  void renameSuccessorPhis(BasicBlock *, MemoryAccess *, bool);
867
  void renamePass(DomTreeNode *, MemoryAccess *IncomingVal,
868
                  SmallPtrSetImpl<BasicBlock *> &Visited,
869
                  bool SkipVisited = false, bool RenameAllUses = false);
870
  AccessList *getOrCreateAccessList(const BasicBlock *);
871
  DefsList *getOrCreateDefsList(const BasicBlock *);
872
  void renumberBlock(const BasicBlock *) const;
873
  AliasAnalysis *AA = nullptr;
874
  DominatorTree *DT;
875
  Function &F;
876
 
877
  // Memory SSA mappings
878
  DenseMap<const Value *, MemoryAccess *> ValueToMemoryAccess;
879
 
880
  // These two mappings contain the main block to access/def mappings for
881
  // MemorySSA. The list contained in PerBlockAccesses really owns all the
882
  // MemoryAccesses.
883
  // Both maps maintain the invariant that if a block is found in them, the
884
  // corresponding list is not empty, and if a block is not found in them, the
885
  // corresponding list is empty.
886
  AccessMap PerBlockAccesses;
887
  DefsMap PerBlockDefs;
888
  std::unique_ptr<MemoryAccess, ValueDeleter> LiveOnEntryDef;
889
 
890
  // Domination mappings
891
  // Note that the numbering is local to a block, even though the map is
892
  // global.
893
  mutable SmallPtrSet<const BasicBlock *, 16> BlockNumberingValid;
894
  mutable DenseMap<const MemoryAccess *, unsigned long> BlockNumbering;
895
 
896
  // Memory SSA building info
897
  std::unique_ptr<ClobberWalkerBase> WalkerBase;
898
  std::unique_ptr<CachingWalker> Walker;
899
  std::unique_ptr<SkipSelfWalker> SkipWalker;
900
  unsigned NextID = 0;
901
  bool IsOptimized = false;
902
};
903
 
904
/// Enables verification of MemorySSA.
905
///
906
/// The checks which this flag enables is exensive and disabled by default
907
/// unless `EXPENSIVE_CHECKS` is defined.  The flag `-verify-memoryssa` can be
908
/// used to selectively enable the verification without re-compilation.
909
extern bool VerifyMemorySSA;
910
 
911
// Internal MemorySSA utils, for use by MemorySSA classes and walkers
912
class MemorySSAUtil {
913
protected:
914
  friend class GVNHoist;
915
  friend class MemorySSAWalker;
916
 
917
  // This function should not be used by new passes.
918
  static bool defClobbersUseOrDef(MemoryDef *MD, const MemoryUseOrDef *MU,
919
                                  AliasAnalysis &AA);
920
};
921
 
922
// This pass does eager building and then printing of MemorySSA. It is used by
923
// the tests to be able to build, dump, and verify Memory SSA.
924
class MemorySSAPrinterLegacyPass : public FunctionPass {
925
public:
926
  MemorySSAPrinterLegacyPass();
927
 
928
  bool runOnFunction(Function &) override;
929
  void getAnalysisUsage(AnalysisUsage &AU) const override;
930
 
931
  static char ID;
932
};
933
 
934
/// An analysis that produces \c MemorySSA for a function.
935
///
936
class MemorySSAAnalysis : public AnalysisInfoMixin<MemorySSAAnalysis> {
937
  friend AnalysisInfoMixin<MemorySSAAnalysis>;
938
 
939
  static AnalysisKey Key;
940
 
941
public:
942
  // Wrap MemorySSA result to ensure address stability of internal MemorySSA
943
  // pointers after construction.  Use a wrapper class instead of plain
944
  // unique_ptr<MemorySSA> to avoid build breakage on MSVC.
945
  struct Result {
946
    Result(std::unique_ptr<MemorySSA> &&MSSA) : MSSA(std::move(MSSA)) {}
947
 
948
    MemorySSA &getMSSA() { return *MSSA.get(); }
949
 
950
    std::unique_ptr<MemorySSA> MSSA;
951
 
952
    bool invalidate(Function &F, const PreservedAnalyses &PA,
953
                    FunctionAnalysisManager::Invalidator &Inv);
954
  };
955
 
956
  Result run(Function &F, FunctionAnalysisManager &AM);
957
};
958
 
959
/// Printer pass for \c MemorySSA.
960
class MemorySSAPrinterPass : public PassInfoMixin<MemorySSAPrinterPass> {
961
  raw_ostream &OS;
962
 
963
public:
964
  explicit MemorySSAPrinterPass(raw_ostream &OS) : OS(OS) {}
965
 
966
  PreservedAnalyses run(Function &F, FunctionAnalysisManager &AM);
967
};
968
 
969
/// Printer pass for \c MemorySSA via the walker.
970
class MemorySSAWalkerPrinterPass
971
    : public PassInfoMixin<MemorySSAWalkerPrinterPass> {
972
  raw_ostream &OS;
973
 
974
public:
975
  explicit MemorySSAWalkerPrinterPass(raw_ostream &OS) : OS(OS) {}
976
 
977
  PreservedAnalyses run(Function &F, FunctionAnalysisManager &AM);
978
};
979
 
980
/// Verifier pass for \c MemorySSA.
981
struct MemorySSAVerifierPass : PassInfoMixin<MemorySSAVerifierPass> {
982
  PreservedAnalyses run(Function &F, FunctionAnalysisManager &AM);
983
};
984
 
985
/// Legacy analysis pass which computes \c MemorySSA.
986
class MemorySSAWrapperPass : public FunctionPass {
987
public:
988
  MemorySSAWrapperPass();
989
 
990
  static char ID;
991
 
992
  bool runOnFunction(Function &) override;
993
  void releaseMemory() override;
994
  MemorySSA &getMSSA() { return *MSSA; }
995
  const MemorySSA &getMSSA() const { return *MSSA; }
996
 
997
  void getAnalysisUsage(AnalysisUsage &AU) const override;
998
 
999
  void verifyAnalysis() const override;
1000
  void print(raw_ostream &OS, const Module *M = nullptr) const override;
1001
 
1002
private:
1003
  std::unique_ptr<MemorySSA> MSSA;
1004
};
1005
 
1006
/// This is the generic walker interface for walkers of MemorySSA.
1007
/// Walkers are used to be able to further disambiguate the def-use chains
1008
/// MemorySSA gives you, or otherwise produce better info than MemorySSA gives
1009
/// you.
1010
/// In particular, while the def-use chains provide basic information, and are
1011
/// guaranteed to give, for example, the nearest may-aliasing MemoryDef for a
1012
/// MemoryUse as AliasAnalysis considers it, a user mant want better or other
1013
/// information. In particular, they may want to use SCEV info to further
1014
/// disambiguate memory accesses, or they may want the nearest dominating
1015
/// may-aliasing MemoryDef for a call or a store. This API enables a
1016
/// standardized interface to getting and using that info.
1017
class MemorySSAWalker {
1018
public:
1019
  MemorySSAWalker(MemorySSA *);
1020
  virtual ~MemorySSAWalker() = default;
1021
 
1022
  using MemoryAccessSet = SmallVector<MemoryAccess *, 8>;
1023
 
1024
  /// Given a memory Mod/Ref/ModRef'ing instruction, calling this
1025
  /// will give you the nearest dominating MemoryAccess that Mod's the location
1026
  /// the instruction accesses (by skipping any def which AA can prove does not
1027
  /// alias the location(s) accessed by the instruction given).
1028
  ///
1029
  /// Note that this will return a single access, and it must dominate the
1030
  /// Instruction, so if an operand of a MemoryPhi node Mod's the instruction,
1031
  /// this will return the MemoryPhi, not the operand. This means that
1032
  /// given:
1033
  /// if (a) {
1034
  ///   1 = MemoryDef(liveOnEntry)
1035
  ///   store %a
1036
  /// } else {
1037
  ///   2 = MemoryDef(liveOnEntry)
1038
  ///   store %b
1039
  /// }
1040
  /// 3 = MemoryPhi(2, 1)
1041
  /// MemoryUse(3)
1042
  /// load %a
1043
  ///
1044
  /// calling this API on load(%a) will return the MemoryPhi, not the MemoryDef
1045
  /// in the if (a) branch.
1046
  MemoryAccess *getClobberingMemoryAccess(const Instruction *I,
1047
                                          BatchAAResults &AA) {
1048
    MemoryAccess *MA = MSSA->getMemoryAccess(I);
1049
    assert(MA && "Handed an instruction that MemorySSA doesn't recognize?");
1050
    return getClobberingMemoryAccess(MA, AA);
1051
  }
1052
 
1053
  /// Does the same thing as getClobberingMemoryAccess(const Instruction *I),
1054
  /// but takes a MemoryAccess instead of an Instruction.
1055
  virtual MemoryAccess *getClobberingMemoryAccess(MemoryAccess *,
1056
                                                  BatchAAResults &AA) = 0;
1057
 
1058
  /// Given a potentially clobbering memory access and a new location,
1059
  /// calling this will give you the nearest dominating clobbering MemoryAccess
1060
  /// (by skipping non-aliasing def links).
1061
  ///
1062
  /// This version of the function is mainly used to disambiguate phi translated
1063
  /// pointers, where the value of a pointer may have changed from the initial
1064
  /// memory access. Note that this expects to be handed either a MemoryUse,
1065
  /// or an already potentially clobbering access. Unlike the above API, if
1066
  /// given a MemoryDef that clobbers the pointer as the starting access, it
1067
  /// will return that MemoryDef, whereas the above would return the clobber
1068
  /// starting from the use side of  the memory def.
1069
  virtual MemoryAccess *getClobberingMemoryAccess(MemoryAccess *,
1070
                                                  const MemoryLocation &,
1071
                                                  BatchAAResults &AA) = 0;
1072
 
1073
  MemoryAccess *getClobberingMemoryAccess(const Instruction *I) {
1074
    BatchAAResults BAA(MSSA->getAA());
1075
    return getClobberingMemoryAccess(I, BAA);
1076
  }
1077
 
1078
  MemoryAccess *getClobberingMemoryAccess(MemoryAccess *MA) {
1079
    BatchAAResults BAA(MSSA->getAA());
1080
    return getClobberingMemoryAccess(MA, BAA);
1081
  }
1082
 
1083
  MemoryAccess *getClobberingMemoryAccess(MemoryAccess *MA,
1084
                                          const MemoryLocation &Loc) {
1085
    BatchAAResults BAA(MSSA->getAA());
1086
    return getClobberingMemoryAccess(MA, Loc, BAA);
1087
  }
1088
 
1089
  /// Given a memory access, invalidate anything this walker knows about
1090
  /// that access.
1091
  /// This API is used by walkers that store information to perform basic cache
1092
  /// invalidation.  This will be called by MemorySSA at appropriate times for
1093
  /// the walker it uses or returns.
1094
  virtual void invalidateInfo(MemoryAccess *) {}
1095
 
1096
protected:
1097
  friend class MemorySSA; // For updating MSSA pointer in MemorySSA move
1098
                          // constructor.
1099
  MemorySSA *MSSA;
1100
};
1101
 
1102
/// A MemorySSAWalker that does no alias queries, or anything else. It
1103
/// simply returns the links as they were constructed by the builder.
1104
class DoNothingMemorySSAWalker final : public MemorySSAWalker {
1105
public:
1106
  // Keep the overrides below from hiding the Instruction overload of
1107
  // getClobberingMemoryAccess.
1108
  using MemorySSAWalker::getClobberingMemoryAccess;
1109
 
1110
  MemoryAccess *getClobberingMemoryAccess(MemoryAccess *,
1111
                                          BatchAAResults &) override;
1112
  MemoryAccess *getClobberingMemoryAccess(MemoryAccess *,
1113
                                          const MemoryLocation &,
1114
                                          BatchAAResults &) override;
1115
};
1116
 
1117
using MemoryAccessPair = std::pair<MemoryAccess *, MemoryLocation>;
1118
using ConstMemoryAccessPair = std::pair<const MemoryAccess *, MemoryLocation>;
1119
 
1120
/// Iterator base class used to implement const and non-const iterators
1121
/// over the defining accesses of a MemoryAccess.
1122
template <class T>
1123
class memoryaccess_def_iterator_base
1124
    : public iterator_facade_base<memoryaccess_def_iterator_base<T>,
1125
                                  std::forward_iterator_tag, T, ptrdiff_t, T *,
1126
                                  T *> {
1127
  using BaseT = typename memoryaccess_def_iterator_base::iterator_facade_base;
1128
 
1129
public:
1130
  memoryaccess_def_iterator_base(T *Start) : Access(Start) {}
1131
  memoryaccess_def_iterator_base() = default;
1132
 
1133
  bool operator==(const memoryaccess_def_iterator_base &Other) const {
1134
    return Access == Other.Access && (!Access || ArgNo == Other.ArgNo);
1135
  }
1136
 
1137
  // This is a bit ugly, but for MemoryPHI's, unlike PHINodes, you can't get the
1138
  // block from the operand in constant time (In a PHINode, the uselist has
1139
  // both, so it's just subtraction). We provide it as part of the
1140
  // iterator to avoid callers having to linear walk to get the block.
1141
  // If the operation becomes constant time on MemoryPHI's, this bit of
1142
  // abstraction breaking should be removed.
1143
  BasicBlock *getPhiArgBlock() const {
1144
    MemoryPhi *MP = dyn_cast<MemoryPhi>(Access);
1145
    assert(MP && "Tried to get phi arg block when not iterating over a PHI");
1146
    return MP->getIncomingBlock(ArgNo);
1147
  }
1148
 
1149
  typename std::iterator_traits<BaseT>::pointer operator*() const {
1150
    assert(Access && "Tried to access past the end of our iterator");
1151
    // Go to the first argument for phis, and the defining access for everything
1152
    // else.
1153
    if (const MemoryPhi *MP = dyn_cast<MemoryPhi>(Access))
1154
      return MP->getIncomingValue(ArgNo);
1155
    return cast<MemoryUseOrDef>(Access)->getDefiningAccess();
1156
  }
1157
 
1158
  using BaseT::operator++;
1159
  memoryaccess_def_iterator_base &operator++() {
1160
    assert(Access && "Hit end of iterator");
1161
    if (const MemoryPhi *MP = dyn_cast<MemoryPhi>(Access)) {
1162
      if (++ArgNo >= MP->getNumIncomingValues()) {
1163
        ArgNo = 0;
1164
        Access = nullptr;
1165
      }
1166
    } else {
1167
      Access = nullptr;
1168
    }
1169
    return *this;
1170
  }
1171
 
1172
private:
1173
  T *Access = nullptr;
1174
  unsigned ArgNo = 0;
1175
};
1176
 
1177
inline memoryaccess_def_iterator MemoryAccess::defs_begin() {
1178
  return memoryaccess_def_iterator(this);
1179
}
1180
 
1181
inline const_memoryaccess_def_iterator MemoryAccess::defs_begin() const {
1182
  return const_memoryaccess_def_iterator(this);
1183
}
1184
 
1185
inline memoryaccess_def_iterator MemoryAccess::defs_end() {
1186
  return memoryaccess_def_iterator();
1187
}
1188
 
1189
inline const_memoryaccess_def_iterator MemoryAccess::defs_end() const {
1190
  return const_memoryaccess_def_iterator();
1191
}
1192
 
1193
/// GraphTraits for a MemoryAccess, which walks defs in the normal case,
1194
/// and uses in the inverse case.
1195
template <> struct GraphTraits<MemoryAccess *> {
1196
  using NodeRef = MemoryAccess *;
1197
  using ChildIteratorType = memoryaccess_def_iterator;
1198
 
1199
  static NodeRef getEntryNode(NodeRef N) { return N; }
1200
  static ChildIteratorType child_begin(NodeRef N) { return N->defs_begin(); }
1201
  static ChildIteratorType child_end(NodeRef N) { return N->defs_end(); }
1202
};
1203
 
1204
template <> struct GraphTraits<Inverse<MemoryAccess *>> {
1205
  using NodeRef = MemoryAccess *;
1206
  using ChildIteratorType = MemoryAccess::iterator;
1207
 
1208
  static NodeRef getEntryNode(NodeRef N) { return N; }
1209
  static ChildIteratorType child_begin(NodeRef N) { return N->user_begin(); }
1210
  static ChildIteratorType child_end(NodeRef N) { return N->user_end(); }
1211
};
1212
 
1213
/// Provide an iterator that walks defs, giving both the memory access,
1214
/// and the current pointer location, updating the pointer location as it
1215
/// changes due to phi node translation.
1216
///
1217
/// This iterator, while somewhat specialized, is what most clients actually
1218
/// want when walking upwards through MemorySSA def chains. It takes a pair of
1219
/// <MemoryAccess,MemoryLocation>, and walks defs, properly translating the
1220
/// memory location through phi nodes for the user.
1221
class upward_defs_iterator
1222
    : public iterator_facade_base<upward_defs_iterator,
1223
                                  std::forward_iterator_tag,
1224
                                  const MemoryAccessPair> {
1225
  using BaseT = upward_defs_iterator::iterator_facade_base;
1226
 
1227
public:
1228
  upward_defs_iterator(const MemoryAccessPair &Info, DominatorTree *DT)
1229
      : DefIterator(Info.first), Location(Info.second),
1230
        OriginalAccess(Info.first), DT(DT) {
1231
    CurrentPair.first = nullptr;
1232
 
1233
    WalkingPhi = Info.first && isa<MemoryPhi>(Info.first);
1234
    fillInCurrentPair();
1235
  }
1236
 
1237
  upward_defs_iterator() { CurrentPair.first = nullptr; }
1238
 
1239
  bool operator==(const upward_defs_iterator &Other) const {
1240
    return DefIterator == Other.DefIterator;
1241
  }
1242
 
1243
  typename std::iterator_traits<BaseT>::reference operator*() const {
1244
    assert(DefIterator != OriginalAccess->defs_end() &&
1245
           "Tried to access past the end of our iterator");
1246
    return CurrentPair;
1247
  }
1248
 
1249
  using BaseT::operator++;
1250
  upward_defs_iterator &operator++() {
1251
    assert(DefIterator != OriginalAccess->defs_end() &&
1252
           "Tried to access past the end of the iterator");
1253
    ++DefIterator;
1254
    if (DefIterator != OriginalAccess->defs_end())
1255
      fillInCurrentPair();
1256
    return *this;
1257
  }
1258
 
1259
  BasicBlock *getPhiArgBlock() const { return DefIterator.getPhiArgBlock(); }
1260
 
1261
private:
1262
  /// Returns true if \p Ptr is guaranteed to be loop invariant for any possible
1263
  /// loop. In particular, this guarantees that it only references a single
1264
  /// MemoryLocation during execution of the containing function.
1265
  bool IsGuaranteedLoopInvariant(const Value *Ptr) const;
1266
 
1267
  void fillInCurrentPair() {
1268
    CurrentPair.first = *DefIterator;
1269
    CurrentPair.second = Location;
1270
    if (WalkingPhi && Location.Ptr) {
1271
      PHITransAddr Translator(
1272
          const_cast<Value *>(Location.Ptr),
1273
          OriginalAccess->getBlock()->getModule()->getDataLayout(), nullptr);
1274
 
1275
      if (!Translator.PHITranslateValue(OriginalAccess->getBlock(),
1276
                                        DefIterator.getPhiArgBlock(), DT, true))
1277
        if (Translator.getAddr() != CurrentPair.second.Ptr)
1278
          CurrentPair.second =
1279
              CurrentPair.second.getWithNewPtr(Translator.getAddr());
1280
 
1281
      // Mark size as unknown, if the location is not guaranteed to be
1282
      // loop-invariant for any possible loop in the function. Setting the size
1283
      // to unknown guarantees that any memory accesses that access locations
1284
      // after the pointer are considered as clobbers, which is important to
1285
      // catch loop carried dependences.
1286
      if (!IsGuaranteedLoopInvariant(CurrentPair.second.Ptr))
1287
        CurrentPair.second = CurrentPair.second.getWithNewSize(
1288
            LocationSize::beforeOrAfterPointer());
1289
    }
1290
  }
1291
 
1292
  MemoryAccessPair CurrentPair;
1293
  memoryaccess_def_iterator DefIterator;
1294
  MemoryLocation Location;
1295
  MemoryAccess *OriginalAccess = nullptr;
1296
  DominatorTree *DT = nullptr;
1297
  bool WalkingPhi = false;
1298
};
1299
 
1300
inline upward_defs_iterator
1301
upward_defs_begin(const MemoryAccessPair &Pair, DominatorTree &DT) {
1302
  return upward_defs_iterator(Pair, &DT);
1303
}
1304
 
1305
inline upward_defs_iterator upward_defs_end() { return upward_defs_iterator(); }
1306
 
1307
inline iterator_range<upward_defs_iterator>
1308
upward_defs(const MemoryAccessPair &Pair, DominatorTree &DT) {
1309
  return make_range(upward_defs_begin(Pair, DT), upward_defs_end());
1310
}
1311
 
1312
/// Walks the defining accesses of MemoryDefs. Stops after we hit something that
1313
/// has no defining use (e.g. a MemoryPhi or liveOnEntry). Note that, when
1314
/// comparing against a null def_chain_iterator, this will compare equal only
1315
/// after walking said Phi/liveOnEntry.
1316
///
1317
/// The UseOptimizedChain flag specifies whether to walk the clobbering
1318
/// access chain, or all the accesses.
1319
///
1320
/// Normally, MemoryDef are all just def/use linked together, so a def_chain on
1321
/// a MemoryDef will walk all MemoryDefs above it in the program until it hits
1322
/// a phi node.  The optimized chain walks the clobbering access of a store.
1323
/// So if you are just trying to find, given a store, what the next
1324
/// thing that would clobber the same memory is, you want the optimized chain.
1325
template <class T, bool UseOptimizedChain = false>
1326
struct def_chain_iterator
1327
    : public iterator_facade_base<def_chain_iterator<T, UseOptimizedChain>,
1328
                                  std::forward_iterator_tag, MemoryAccess *> {
1329
  def_chain_iterator() : MA(nullptr) {}
1330
  def_chain_iterator(T MA) : MA(MA) {}
1331
 
1332
  T operator*() const { return MA; }
1333
 
1334
  def_chain_iterator &operator++() {
1335
    // N.B. liveOnEntry has a null defining access.
1336
    if (auto *MUD = dyn_cast<MemoryUseOrDef>(MA)) {
1337
      if (UseOptimizedChain && MUD->isOptimized())
1338
        MA = MUD->getOptimized();
1339
      else
1340
        MA = MUD->getDefiningAccess();
1341
    } else {
1342
      MA = nullptr;
1343
    }
1344
 
1345
    return *this;
1346
  }
1347
 
1348
  bool operator==(const def_chain_iterator &O) const { return MA == O.MA; }
1349
 
1350
private:
1351
  T MA;
1352
};
1353
 
1354
template <class T>
1355
inline iterator_range<def_chain_iterator<T>>
1356
def_chain(T MA, MemoryAccess *UpTo = nullptr) {
1357
#ifdef EXPENSIVE_CHECKS
1358
  assert((!UpTo || find(def_chain(MA), UpTo) != def_chain_iterator<T>()) &&
1359
         "UpTo isn't in the def chain!");
1360
#endif
1361
  return make_range(def_chain_iterator<T>(MA), def_chain_iterator<T>(UpTo));
1362
}
1363
 
1364
template <class T>
1365
inline iterator_range<def_chain_iterator<T, true>> optimized_def_chain(T MA) {
1366
  return make_range(def_chain_iterator<T, true>(MA),
1367
                    def_chain_iterator<T, true>(nullptr));
1368
}
1369
 
1370
} // end namespace llvm
1371
 
1372
#endif // LLVM_ANALYSIS_MEMORYSSA_H