1//===- Reassociate.cpp - Reassociate binary expressions -------------------===//
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// This pass reassociates commutative expressions in an order that is designed
10// to promote better constant propagation, GCSE, LICM, PRE, etc.
11//
12// For example: 4 + (x + 5) -> x + (4 + 5)
13//
14// In the implementation of this algorithm, constants are assigned rank = 0,
15// function arguments are rank = 1, and other values are assigned ranks
16// corresponding to the reverse post order traversal of current function
17// (starting at 2), which effectively gives values in deep loops higher rank
18// than values not in loops.
19//
20//===----------------------------------------------------------------------===//
21
22#include "llvm/Transforms/Scalar/Reassociate.h"
23#include "llvm/ADT/APFloat.h"
24#include "llvm/ADT/APInt.h"
25#include "llvm/ADT/DenseMap.h"
26#include "llvm/ADT/PostOrderIterator.h"
27#include "llvm/ADT/SmallPtrSet.h"
28#include "llvm/ADT/SmallSet.h"
29#include "llvm/ADT/SmallVector.h"
30#include "llvm/ADT/Statistic.h"
31#include "llvm/Analysis/BasicAliasAnalysis.h"
32#include "llvm/Analysis/ConstantFolding.h"
33#include "llvm/Analysis/GlobalsModRef.h"
34#include "llvm/Analysis/ValueTracking.h"
35#include "llvm/IR/Argument.h"
36#include "llvm/IR/BasicBlock.h"
37#include "llvm/IR/CFG.h"
38#include "llvm/IR/Constant.h"
39#include "llvm/IR/Constants.h"
40#include "llvm/IR/Function.h"
41#include "llvm/IR/IRBuilder.h"
42#include "llvm/IR/InstrTypes.h"
43#include "llvm/IR/Instruction.h"
44#include "llvm/IR/Instructions.h"
45#include "llvm/IR/Operator.h"
46#include "llvm/IR/PassManager.h"
47#include "llvm/IR/PatternMatch.h"
48#include "llvm/IR/Type.h"
49#include "llvm/IR/User.h"
50#include "llvm/IR/Value.h"
51#include "llvm/IR/ValueHandle.h"
52#include "llvm/InitializePasses.h"
53#include "llvm/Pass.h"
54#include "llvm/Support/Casting.h"
55#include "llvm/Support/CommandLine.h"
56#include "llvm/Support/Debug.h"
57#include "llvm/Support/raw_ostream.h"
58#include "llvm/Transforms/Scalar.h"
59#include "llvm/Transforms/Utils/Local.h"
60#include <algorithm>
61#include <cassert>
62#include <utility>
63
64using namespace llvm;
65using namespace reassociate;
66using namespace PatternMatch;
67
68#define DEBUG_TYPE "reassociate"
69
70STATISTIC(NumChanged, "Number of insts reassociated");
71STATISTIC(NumAnnihil, "Number of expr tree annihilated");
72STATISTIC(NumFactor , "Number of multiplies factored");
73
74static cl::opt<bool>
75 UseCSELocalOpt(DEBUG_TYPE "-use-cse-local",
76 cl::desc("Only reorder expressions within a basic block "
77 "when exposing CSE opportunities"),
78 cl::init(Val: true), cl::Hidden);
79
80#ifndef NDEBUG
81/// Print out the expression identified in the Ops list.
82static void PrintOps(Instruction *I, const SmallVectorImpl<ValueEntry> &Ops) {
83 Module *M = I->getModule();
84 dbgs() << Instruction::getOpcodeName(I->getOpcode()) << " "
85 << *Ops[0].Op->getType() << '\t';
86 for (const ValueEntry &Op : Ops) {
87 dbgs() << "[ ";
88 Op.Op->printAsOperand(dbgs(), false, M);
89 dbgs() << ", #" << Op.Rank << "] ";
90 }
91}
92#endif
93
94/// Utility class representing a non-constant Xor-operand. We classify
95/// non-constant Xor-Operands into two categories:
96/// C1) The operand is in the form "X & C", where C is a constant and C != ~0
97/// C2)
98/// C2.1) The operand is in the form of "X | C", where C is a non-zero
99/// constant.
100/// C2.2) Any operand E which doesn't fall into C1 and C2.1, we view this
101/// operand as "E | 0"
102class llvm::reassociate::XorOpnd {
103public:
104 XorOpnd(Value *V);
105
106 bool isInvalid() const { return SymbolicPart == nullptr; }
107 bool isOrExpr() const { return isOr; }
108 Value *getValue() const { return OrigVal; }
109 Value *getSymbolicPart() const { return SymbolicPart; }
110 unsigned getSymbolicRank() const { return SymbolicRank; }
111 const APInt &getConstPart() const { return ConstPart; }
112
113 void Invalidate() { SymbolicPart = OrigVal = nullptr; }
114 void setSymbolicRank(unsigned R) { SymbolicRank = R; }
115
116private:
117 Value *OrigVal;
118 Value *SymbolicPart;
119 APInt ConstPart;
120 unsigned SymbolicRank;
121 bool isOr;
122};
123
124XorOpnd::XorOpnd(Value *V) {
125 assert(!isa<ConstantInt>(V) && "No ConstantInt");
126 OrigVal = V;
127 Instruction *I = dyn_cast<Instruction>(Val: V);
128 SymbolicRank = 0;
129
130 if (I && (I->getOpcode() == Instruction::Or ||
131 I->getOpcode() == Instruction::And)) {
132 Value *V0 = I->getOperand(i: 0);
133 Value *V1 = I->getOperand(i: 1);
134 const APInt *C;
135 if (match(V: V0, P: m_APInt(Res&: C)))
136 std::swap(a&: V0, b&: V1);
137
138 if (match(V: V1, P: m_APInt(Res&: C))) {
139 ConstPart = *C;
140 SymbolicPart = V0;
141 isOr = (I->getOpcode() == Instruction::Or);
142 return;
143 }
144 }
145
146 // view the operand as "V | 0"
147 SymbolicPart = V;
148 ConstPart = APInt::getZero(numBits: V->getType()->getScalarSizeInBits());
149 isOr = true;
150}
151
152/// Return true if I is an instruction with the FastMathFlags that are needed
153/// for general reassociation set. This is not the same as testing
154/// Instruction::isAssociative() because it includes operations like fsub.
155/// (This routine is only intended to be called for floating-point operations.)
156static bool hasFPAssociativeFlags(Instruction *I) {
157 assert(I && isa<FPMathOperator>(I) && "Should only check FP ops");
158 return I->hasAllowReassoc() && I->hasNoSignedZeros();
159}
160
161/// Return true if V is an instruction of the specified opcode and if it
162/// only has one use.
163static BinaryOperator *isReassociableOp(Value *V, unsigned Opcode) {
164 auto *BO = dyn_cast<BinaryOperator>(Val: V);
165 if (BO && BO->hasOneUse() && BO->getOpcode() == Opcode)
166 if (!isa<FPMathOperator>(Val: BO) || hasFPAssociativeFlags(I: BO))
167 return BO;
168 return nullptr;
169}
170
171static BinaryOperator *isReassociableOp(Value *V, unsigned Opcode1,
172 unsigned Opcode2) {
173 auto *BO = dyn_cast<BinaryOperator>(Val: V);
174 if (BO && BO->hasOneUse() &&
175 (BO->getOpcode() == Opcode1 || BO->getOpcode() == Opcode2))
176 if (!isa<FPMathOperator>(Val: BO) || hasFPAssociativeFlags(I: BO))
177 return BO;
178 return nullptr;
179}
180
181void ReassociatePass::BuildRankMap(Function &F,
182 ReversePostOrderTraversal<Function*> &RPOT) {
183 unsigned Rank = 2;
184
185 // Assign distinct ranks to function arguments.
186 for (auto &Arg : F.args()) {
187 ValueRankMap[&Arg] = ++Rank;
188 LLVM_DEBUG(dbgs() << "Calculated Rank[" << Arg.getName() << "] = " << Rank
189 << "\n");
190 }
191
192 // Traverse basic blocks in ReversePostOrder.
193 for (BasicBlock *BB : RPOT) {
194 unsigned BBRank = RankMap[BB] = ++Rank << 16;
195
196 // Walk the basic block, adding precomputed ranks for any instructions that
197 // we cannot move. This ensures that the ranks for these instructions are
198 // all different in the block.
199 for (Instruction &I : *BB)
200 if (mayHaveNonDefUseDependency(I))
201 ValueRankMap[&I] = ++BBRank;
202 }
203}
204
205unsigned ReassociatePass::getRank(Value *V) {
206 Instruction *I = dyn_cast<Instruction>(Val: V);
207 if (!I) {
208 if (isa<Argument>(Val: V)) return ValueRankMap[V]; // Function argument.
209 return 0; // Otherwise it's a global or constant, rank 0.
210 }
211
212 if (unsigned Rank = ValueRankMap[I])
213 return Rank; // Rank already known?
214
215 // If this is an expression, return the 1+MAX(rank(LHS), rank(RHS)) so that
216 // we can reassociate expressions for code motion! Since we do not recurse
217 // for PHI nodes, we cannot have infinite recursion here, because there
218 // cannot be loops in the value graph that do not go through PHI nodes.
219 unsigned Rank = 0, MaxRank = RankMap[I->getParent()];
220 for (unsigned i = 0, e = I->getNumOperands(); i != e && Rank != MaxRank; ++i)
221 Rank = std::max(a: Rank, b: getRank(V: I->getOperand(i)));
222
223 // If this is a 'not' or 'neg' instruction, do not count it for rank. This
224 // assures us that X and ~X will have the same rank.
225 if (!match(V: I, P: m_Not(V: m_Value())) && !match(V: I, P: m_Neg(V: m_Value())) &&
226 !match(V: I, P: m_FNeg(X: m_Value())))
227 ++Rank;
228
229 LLVM_DEBUG(dbgs() << "Calculated Rank[" << V->getName() << "] = " << Rank
230 << "\n");
231
232 return ValueRankMap[I] = Rank;
233}
234
235// Canonicalize constants to RHS. Otherwise, sort the operands by rank.
236void ReassociatePass::canonicalizeOperands(Instruction *I) {
237 assert(isa<BinaryOperator>(I) && "Expected binary operator.");
238 assert(I->isCommutative() && "Expected commutative operator.");
239
240 Value *LHS = I->getOperand(i: 0);
241 Value *RHS = I->getOperand(i: 1);
242 if (LHS == RHS || isa<Constant>(Val: RHS))
243 return;
244 if (isa<Constant>(Val: LHS) || getRank(V: RHS) < getRank(V: LHS)) {
245 cast<BinaryOperator>(Val: I)->swapOperands();
246 MadeChange = true;
247 }
248}
249
250static BinaryOperator *CreateAdd(Value *S1, Value *S2, const Twine &Name,
251 BasicBlock::iterator InsertBefore,
252 Value *FlagsOp) {
253 if (S1->getType()->isIntOrIntVectorTy())
254 return BinaryOperator::CreateAdd(V1: S1, V2: S2, Name, InsertBefore);
255 else {
256 BinaryOperator *Res =
257 BinaryOperator::CreateFAdd(V1: S1, V2: S2, Name, InsertBefore);
258 Res->setFastMathFlags(cast<FPMathOperator>(Val: FlagsOp)->getFastMathFlags());
259 return Res;
260 }
261}
262
263static BinaryOperator *CreateMul(Value *S1, Value *S2, const Twine &Name,
264 BasicBlock::iterator InsertBefore,
265 Value *FlagsOp) {
266 if (S1->getType()->isIntOrIntVectorTy())
267 return BinaryOperator::CreateMul(V1: S1, V2: S2, Name, InsertBefore);
268 else {
269 BinaryOperator *Res =
270 BinaryOperator::CreateFMul(V1: S1, V2: S2, Name, InsertBefore);
271 Res->setFastMathFlags(cast<FPMathOperator>(Val: FlagsOp)->getFastMathFlags());
272 return Res;
273 }
274}
275
276static Instruction *CreateNeg(Value *S1, const Twine &Name,
277 BasicBlock::iterator InsertBefore,
278 Value *FlagsOp) {
279 if (S1->getType()->isIntOrIntVectorTy())
280 return BinaryOperator::CreateNeg(Op: S1, Name, InsertBefore);
281
282 if (auto *FMFSource = dyn_cast<Instruction>(Val: FlagsOp))
283 return UnaryOperator::CreateFNegFMF(Op: S1, FMFSource, Name, InsertBefore);
284
285 return UnaryOperator::CreateFNeg(V: S1, Name, InsertBefore);
286}
287
288/// Replace 0-X with X*-1.
289static BinaryOperator *LowerNegateToMultiply(Instruction *Neg) {
290 assert((isa<UnaryOperator>(Neg) || isa<BinaryOperator>(Neg)) &&
291 "Expected a Negate!");
292 // FIXME: It's not safe to lower a unary FNeg into a FMul by -1.0.
293 unsigned OpNo = isa<BinaryOperator>(Val: Neg) ? 1 : 0;
294 Type *Ty = Neg->getType();
295 Constant *NegOne = Ty->isIntOrIntVectorTy() ?
296 ConstantInt::getAllOnesValue(Ty) : ConstantFP::get(Ty, V: -1.0);
297
298 BinaryOperator *Res =
299 CreateMul(S1: Neg->getOperand(i: OpNo), S2: NegOne, Name: "", InsertBefore: Neg->getIterator(), FlagsOp: Neg);
300 Neg->setOperand(i: OpNo, Val: Constant::getNullValue(Ty)); // Drop use of op.
301 Res->takeName(V: Neg);
302 Neg->replaceAllUsesWith(V: Res);
303 Res->setDebugLoc(Neg->getDebugLoc());
304 return Res;
305}
306
307using RepeatedValue = std::pair<Value *, uint64_t>;
308
309/// Given an associative binary expression, return the leaf
310/// nodes in Ops along with their weights (how many times the leaf occurs). The
311/// original expression is the same as
312/// (Ops[0].first op Ops[0].first op ... Ops[0].first) <- Ops[0].second times
313/// op
314/// (Ops[1].first op Ops[1].first op ... Ops[1].first) <- Ops[1].second times
315/// op
316/// ...
317/// op
318/// (Ops[N].first op Ops[N].first op ... Ops[N].first) <- Ops[N].second times
319///
320/// Note that the values Ops[0].first, ..., Ops[N].first are all distinct.
321///
322/// This routine may modify the function, in which case it returns 'true'. The
323/// changes it makes may well be destructive, changing the value computed by 'I'
324/// to something completely different. Thus if the routine returns 'true' then
325/// you MUST either replace I with a new expression computed from the Ops array,
326/// or use RewriteExprTree to put the values back in.
327///
328/// A leaf node is either not a binary operation of the same kind as the root
329/// node 'I' (i.e. is not a binary operator at all, or is, but with a different
330/// opcode), or is the same kind of binary operator but has a use which either
331/// does not belong to the expression, or does belong to the expression but is
332/// a leaf node. Every leaf node has at least one use that is a non-leaf node
333/// of the expression, while for non-leaf nodes (except for the root 'I') every
334/// use is a non-leaf node of the expression.
335///
336/// For example:
337/// expression graph node names
338///
339/// + | I
340/// / \ |
341/// + + | A, B
342/// / \ / \ |
343/// * + * | C, D, E
344/// / \ / \ / \ |
345/// + * | F, G
346///
347/// The leaf nodes are C, E, F and G. The Ops array will contain (maybe not in
348/// that order) (C, 1), (E, 1), (F, 2), (G, 2).
349///
350/// The expression is maximal: if some instruction is a binary operator of the
351/// same kind as 'I', and all of its uses are non-leaf nodes of the expression,
352/// then the instruction also belongs to the expression, is not a leaf node of
353/// it, and its operands also belong to the expression (but may be leaf nodes).
354///
355/// NOTE: This routine will set operands of non-leaf non-root nodes to undef in
356/// order to ensure that every non-root node in the expression has *exactly one*
357/// use by a non-leaf node of the expression. This destruction means that the
358/// caller MUST either replace 'I' with a new expression or use something like
359/// RewriteExprTree to put the values back in if the routine indicates that it
360/// made a change by returning 'true'.
361///
362/// In the above example either the right operand of A or the left operand of B
363/// will be replaced by undef. If it is B's operand then this gives:
364///
365/// + | I
366/// / \ |
367/// + + | A, B - operand of B replaced with undef
368/// / \ \ |
369/// * + * | C, D, E
370/// / \ / \ / \ |
371/// + * | F, G
372///
373/// Note that such undef operands can only be reached by passing through 'I'.
374/// For example, if you visit operands recursively starting from a leaf node
375/// then you will never see such an undef operand unless you get back to 'I',
376/// which requires passing through a phi node.
377///
378/// Note that this routine may also mutate binary operators of the wrong type
379/// that have all uses inside the expression (i.e. only used by non-leaf nodes
380/// of the expression) if it can turn them into binary operators of the right
381/// type and thus make the expression bigger.
382static bool LinearizeExprTree(Instruction *I,
383 SmallVectorImpl<RepeatedValue> &Ops,
384 ReassociatePass::OrderedSet &ToRedo,
385 OverflowTracking &Flags) {
386 assert((isa<UnaryOperator>(I) || isa<BinaryOperator>(I)) &&
387 "Expected a UnaryOperator or BinaryOperator!");
388 LLVM_DEBUG(dbgs() << "LINEARIZE: " << *I << '\n');
389 unsigned Opcode = I->getOpcode();
390 assert(I->isAssociative() && I->isCommutative() &&
391 "Expected an associative and commutative operation!");
392
393 // Visit all operands of the expression, keeping track of their weight (the
394 // number of paths from the expression root to the operand, or if you like
395 // the number of times that operand occurs in the linearized expression).
396 // For example, if I = X + A, where X = A + B, then I, X and B have weight 1
397 // while A has weight two.
398
399 // Worklist of non-leaf nodes (their operands are in the expression too) along
400 // with their weights, representing a certain number of paths to the operator.
401 // If an operator occurs in the worklist multiple times then we found multiple
402 // ways to get to it.
403 SmallVector<std::pair<Instruction *, uint64_t>, 8> Worklist; // (Op, Weight)
404 Worklist.push_back(Elt: std::make_pair(x&: I, y: 1));
405 bool Changed = false;
406
407 // Leaves of the expression are values that either aren't the right kind of
408 // operation (eg: a constant, or a multiply in an add tree), or are, but have
409 // some uses that are not inside the expression. For example, in I = X + X,
410 // X = A + B, the value X has two uses (by I) that are in the expression. If
411 // X has any other uses, for example in a return instruction, then we consider
412 // X to be a leaf, and won't analyze it further. When we first visit a value,
413 // if it has more than one use then at first we conservatively consider it to
414 // be a leaf. Later, as the expression is explored, we may discover some more
415 // uses of the value from inside the expression. If all uses turn out to be
416 // from within the expression (and the value is a binary operator of the right
417 // kind) then the value is no longer considered to be a leaf, and its operands
418 // are explored.
419
420 // Leaves - Keeps track of the set of putative leaves as well as the number of
421 // paths to each leaf seen so far.
422 using LeafMap = DenseMap<Value *, uint64_t>;
423 LeafMap Leaves; // Leaf -> Total weight so far.
424 SmallVector<Value *, 8> LeafOrder; // Ensure deterministic leaf output order.
425 const DataLayout &DL = I->getDataLayout();
426
427#ifndef NDEBUG
428 SmallPtrSet<Value *, 8> Visited; // For checking the iteration scheme.
429#endif
430 while (!Worklist.empty()) {
431 // We examine the operands of this binary operator.
432 auto [I, Weight] = Worklist.pop_back_val();
433
434 Flags.mergeFlags(I&: *I);
435
436 for (unsigned OpIdx = 0; OpIdx < I->getNumOperands(); ++OpIdx) { // Visit operands.
437 Value *Op = I->getOperand(i: OpIdx);
438 LLVM_DEBUG(dbgs() << "OPERAND: " << *Op << " (" << Weight << ")\n");
439 assert((!Op->hasUseList() || !Op->use_empty()) &&
440 "No uses, so how did we get to it?!");
441
442 // If this is a binary operation of the right kind with only one use then
443 // add its operands to the expression.
444 if (BinaryOperator *BO = isReassociableOp(V: Op, Opcode)) {
445 assert(Visited.insert(Op).second && "Not first visit!");
446 LLVM_DEBUG(dbgs() << "DIRECT ADD: " << *Op << " (" << Weight << ")\n");
447 Worklist.push_back(Elt: std::make_pair(x&: BO, y&: Weight));
448 continue;
449 }
450
451 // Appears to be a leaf. Is the operand already in the set of leaves?
452 LeafMap::iterator It = Leaves.find(Val: Op);
453 if (It == Leaves.end()) {
454 // Not in the leaf map. Must be the first time we saw this operand.
455 assert(Visited.insert(Op).second && "Not first visit!");
456 if (!Op->hasOneUse()) {
457 // This value has uses not accounted for by the expression, so it is
458 // not safe to modify. Mark it as being a leaf.
459 LLVM_DEBUG(dbgs()
460 << "ADD USES LEAF: " << *Op << " (" << Weight << ")\n");
461 LeafOrder.push_back(Elt: Op);
462 Leaves[Op] = Weight;
463 continue;
464 }
465 // No uses outside the expression, try morphing it.
466 } else {
467 // Already in the leaf map.
468 assert(It != Leaves.end() && Visited.count(Op) &&
469 "In leaf map but not visited!");
470
471 // Update the number of paths to the leaf.
472 It->second += Weight;
473 assert(It->second >= Weight && "Weight overflows");
474
475 // If we still have uses that are not accounted for by the expression
476 // then it is not safe to modify the value.
477 if (!Op->hasOneUse())
478 continue;
479
480 // No uses outside the expression, try morphing it.
481 Weight = It->second;
482 Leaves.erase(I: It); // Since the value may be morphed below.
483 }
484
485 // At this point we have a value which, first of all, is not a binary
486 // expression of the right kind, and secondly, is only used inside the
487 // expression. This means that it can safely be modified. See if we
488 // can usefully morph it into an expression of the right kind.
489 assert((!isa<Instruction>(Op) ||
490 cast<Instruction>(Op)->getOpcode() != Opcode
491 || (isa<FPMathOperator>(Op) &&
492 !hasFPAssociativeFlags(cast<Instruction>(Op)))) &&
493 "Should have been handled above!");
494 assert(Op->hasOneUse() && "Has uses outside the expression tree!");
495
496 // If this is a multiply expression, turn any internal negations into
497 // multiplies by -1 so they can be reassociated. Add any users of the
498 // newly created multiplication by -1 to the redo list, so any
499 // reassociation opportunities that are exposed will be reassociated
500 // further.
501 Instruction *Neg;
502 if (((Opcode == Instruction::Mul && match(V: Op, P: m_Neg(V: m_Value()))) ||
503 (Opcode == Instruction::FMul && match(V: Op, P: m_FNeg(X: m_Value())))) &&
504 match(V: Op, P: m_Instruction(I&: Neg))) {
505 LLVM_DEBUG(dbgs()
506 << "MORPH LEAF: " << *Op << " (" << Weight << ") TO ");
507 Instruction *Mul = LowerNegateToMultiply(Neg);
508 LLVM_DEBUG(dbgs() << *Mul << '\n');
509 Worklist.push_back(Elt: std::make_pair(x&: Mul, y&: Weight));
510 for (User *U : Mul->users()) {
511 if (BinaryOperator *UserBO = dyn_cast<BinaryOperator>(Val: U))
512 ToRedo.insert(X: UserBO);
513 }
514 ToRedo.insert(X: Neg);
515 Changed = true;
516 continue;
517 }
518
519 // Failed to morph into an expression of the right type. This really is
520 // a leaf.
521 LLVM_DEBUG(dbgs() << "ADD LEAF: " << *Op << " (" << Weight << ")\n");
522 assert(!isReassociableOp(Op, Opcode) && "Value was morphed?");
523 LeafOrder.push_back(Elt: Op);
524 Leaves[Op] = Weight;
525 }
526 }
527
528 // The leaves, repeated according to their weights, represent the linearized
529 // form of the expression.
530 for (Value *V : LeafOrder) {
531 LeafMap::iterator It = Leaves.find(Val: V);
532 if (It == Leaves.end())
533 // Node initially thought to be a leaf wasn't.
534 continue;
535 assert(!isReassociableOp(V, Opcode) && "Shouldn't be a leaf!");
536 uint64_t Weight = It->second;
537 // Ensure the leaf is only output once.
538 It->second = 0;
539 Ops.push_back(Elt: std::make_pair(x&: V, y&: Weight));
540 if (Opcode == Instruction::Add && Flags.AllKnownNonNegative && Flags.HasNSW)
541 Flags.AllKnownNonNegative &= isKnownNonNegative(V, SQ: SimplifyQuery(DL));
542 else if (Opcode == Instruction::Mul) {
543 // To preserve NUW we need all inputs non-zero.
544 // To preserve NSW we need all inputs strictly positive.
545 if (Flags.AllKnownNonZero &&
546 (Flags.HasNUW || (Flags.HasNSW && Flags.AllKnownNonNegative))) {
547 Flags.AllKnownNonZero &= isKnownNonZero(V, Q: SimplifyQuery(DL));
548 if (Flags.HasNSW && Flags.AllKnownNonNegative)
549 Flags.AllKnownNonNegative &= isKnownNonNegative(V, SQ: SimplifyQuery(DL));
550 }
551 }
552 }
553
554 // For nilpotent operations or addition there may be no operands, for example
555 // because the expression was "X xor X" or consisted of 2^Bitwidth additions:
556 // in both cases the weight reduces to 0 causing the value to be skipped.
557 if (Ops.empty()) {
558 Constant *Identity = ConstantExpr::getBinOpIdentity(Opcode, Ty: I->getType());
559 assert(Identity && "Associative operation without identity!");
560 Ops.emplace_back(Args&: Identity, Args: 1);
561 }
562
563 return Changed;
564}
565
566/// Now that the operands for this expression tree are
567/// linearized and optimized, emit them in-order.
568void ReassociatePass::RewriteExprTree(BinaryOperator *I,
569 SmallVectorImpl<ValueEntry> &Ops,
570 OverflowTracking Flags) {
571 assert(Ops.size() > 1 && "Single values should be used directly!");
572
573 // Since our optimizations should never increase the number of operations, the
574 // new expression can usually be written reusing the existing binary operators
575 // from the original expression tree, without creating any new instructions,
576 // though the rewritten expression may have a completely different topology.
577 // We take care to not change anything if the new expression will be the same
578 // as the original. If more than trivial changes (like commuting operands)
579 // were made then we are obliged to clear out any optional subclass data like
580 // nsw flags.
581
582 /// NodesToRewrite - Nodes from the original expression available for writing
583 /// the new expression into.
584 SmallVector<BinaryOperator*, 8> NodesToRewrite;
585 unsigned Opcode = I->getOpcode();
586 BinaryOperator *Op = I;
587
588 /// NotRewritable - The operands being written will be the leaves of the new
589 /// expression and must not be used as inner nodes (via NodesToRewrite) by
590 /// mistake. Inner nodes are always reassociable, and usually leaves are not
591 /// (if they were they would have been incorporated into the expression and so
592 /// would not be leaves), so most of the time there is no danger of this. But
593 /// in rare cases a leaf may become reassociable if an optimization kills uses
594 /// of it, or it may momentarily become reassociable during rewriting (below)
595 /// due it being removed as an operand of one of its uses. Ensure that misuse
596 /// of leaf nodes as inner nodes cannot occur by remembering all of the future
597 /// leaves and refusing to reuse any of them as inner nodes.
598 SmallPtrSet<Value*, 8> NotRewritable;
599 for (const ValueEntry &Op : Ops)
600 NotRewritable.insert(Ptr: Op.Op);
601
602 // ExpressionChangedStart - Non-null if the rewritten expression differs from
603 // the original in some non-trivial way, requiring the clearing of optional
604 // flags. Flags are cleared from the operator in ExpressionChangedStart up to
605 // ExpressionChangedEnd inclusive.
606 BinaryOperator *ExpressionChangedStart = nullptr,
607 *ExpressionChangedEnd = nullptr;
608 for (unsigned i = 0; ; ++i) {
609 // The last operation (which comes earliest in the IR) is special as both
610 // operands will come from Ops, rather than just one with the other being
611 // a subexpression.
612 if (i+2 == Ops.size()) {
613 Value *NewLHS = Ops[i].Op;
614 Value *NewRHS = Ops[i+1].Op;
615 Value *OldLHS = Op->getOperand(i_nocapture: 0);
616 Value *OldRHS = Op->getOperand(i_nocapture: 1);
617
618 if (NewLHS == OldLHS && NewRHS == OldRHS)
619 // Nothing changed, leave it alone.
620 break;
621
622 if (NewLHS == OldRHS && NewRHS == OldLHS) {
623 // The order of the operands was reversed. Swap them.
624 LLVM_DEBUG(dbgs() << "RA: " << *Op << '\n');
625 Op->swapOperands();
626 LLVM_DEBUG(dbgs() << "TO: " << *Op << '\n');
627 MadeChange = true;
628 ++NumChanged;
629 break;
630 }
631
632 // The new operation differs non-trivially from the original. Overwrite
633 // the old operands with the new ones.
634 LLVM_DEBUG(dbgs() << "RA: " << *Op << '\n');
635 if (NewLHS != OldLHS) {
636 BinaryOperator *BO = isReassociableOp(V: OldLHS, Opcode);
637 if (BO && !NotRewritable.count(Ptr: BO))
638 NodesToRewrite.push_back(Elt: BO);
639 Op->setOperand(i_nocapture: 0, Val_nocapture: NewLHS);
640 }
641 if (NewRHS != OldRHS) {
642 BinaryOperator *BO = isReassociableOp(V: OldRHS, Opcode);
643 if (BO && !NotRewritable.count(Ptr: BO))
644 NodesToRewrite.push_back(Elt: BO);
645 Op->setOperand(i_nocapture: 1, Val_nocapture: NewRHS);
646 }
647 LLVM_DEBUG(dbgs() << "TO: " << *Op << '\n');
648
649 ExpressionChangedStart = Op;
650 if (!ExpressionChangedEnd)
651 ExpressionChangedEnd = Op;
652 MadeChange = true;
653 ++NumChanged;
654
655 break;
656 }
657
658 // Not the last operation. The left-hand side will be a sub-expression
659 // while the right-hand side will be the current element of Ops.
660 Value *NewRHS = Ops[i].Op;
661 if (NewRHS != Op->getOperand(i_nocapture: 1)) {
662 LLVM_DEBUG(dbgs() << "RA: " << *Op << '\n');
663 if (NewRHS == Op->getOperand(i_nocapture: 0)) {
664 // The new right-hand side was already present as the left operand. If
665 // we are lucky then swapping the operands will sort out both of them.
666 Op->swapOperands();
667 } else {
668 // Overwrite with the new right-hand side.
669 BinaryOperator *BO = isReassociableOp(V: Op->getOperand(i_nocapture: 1), Opcode);
670 if (BO && !NotRewritable.count(Ptr: BO))
671 NodesToRewrite.push_back(Elt: BO);
672 Op->setOperand(i_nocapture: 1, Val_nocapture: NewRHS);
673 ExpressionChangedStart = Op;
674 if (!ExpressionChangedEnd)
675 ExpressionChangedEnd = Op;
676 }
677 LLVM_DEBUG(dbgs() << "TO: " << *Op << '\n');
678 MadeChange = true;
679 ++NumChanged;
680 }
681
682 // Now deal with the left-hand side. If this is already an operation node
683 // from the original expression then just rewrite the rest of the expression
684 // into it.
685 BinaryOperator *BO = isReassociableOp(V: Op->getOperand(i_nocapture: 0), Opcode);
686 if (BO && !NotRewritable.count(Ptr: BO)) {
687 Op = BO;
688 continue;
689 }
690
691 // Otherwise, grab a spare node from the original expression and use that as
692 // the left-hand side. If there are no nodes left then the optimizers made
693 // an expression with more nodes than the original! This usually means that
694 // they did something stupid but it might mean that the problem was just too
695 // hard (finding the mimimal number of multiplications needed to realize a
696 // multiplication expression is NP-complete). Whatever the reason, smart or
697 // stupid, create a new node if there are none left.
698 BinaryOperator *NewOp;
699 if (NodesToRewrite.empty()) {
700 Constant *Poison = PoisonValue::get(T: I->getType());
701 NewOp = BinaryOperator::Create(Op: Instruction::BinaryOps(Opcode), S1: Poison,
702 S2: Poison, Name: "", InsertBefore: I->getIterator());
703 if (isa<FPMathOperator>(Val: NewOp))
704 NewOp->setFastMathFlags(I->getFastMathFlags());
705 } else {
706 NewOp = NodesToRewrite.pop_back_val();
707 }
708
709 LLVM_DEBUG(dbgs() << "RA: " << *Op << '\n');
710 Op->setOperand(i_nocapture: 0, Val_nocapture: NewOp);
711 LLVM_DEBUG(dbgs() << "TO: " << *Op << '\n');
712 ExpressionChangedStart = Op;
713 if (!ExpressionChangedEnd)
714 ExpressionChangedEnd = Op;
715 MadeChange = true;
716 ++NumChanged;
717 Op = NewOp;
718 }
719
720 // If the expression changed non-trivially then clear out all subclass data
721 // starting from the operator specified in ExpressionChanged, and compactify
722 // the operators to just before the expression root to guarantee that the
723 // expression tree is dominated by all of Ops.
724 if (ExpressionChangedStart) {
725 bool ClearFlags = true;
726 do {
727 // Preserve flags.
728 if (ClearFlags) {
729 if (isa<FPMathOperator>(Val: I)) {
730 FastMathFlags Flags = I->getFastMathFlags();
731 ExpressionChangedStart->clearSubclassOptionalData();
732 ExpressionChangedStart->setFastMathFlags(Flags);
733 } else {
734 Flags.applyFlags(I&: *ExpressionChangedStart);
735 }
736 }
737
738 if (ExpressionChangedStart == ExpressionChangedEnd)
739 ClearFlags = false;
740 if (ExpressionChangedStart == I)
741 break;
742
743 // Discard any debug info related to the expressions that has changed (we
744 // can leave debug info related to the root and any operation that didn't
745 // change, since the result of the expression tree should be the same
746 // even after reassociation).
747 if (ClearFlags)
748 replaceDbgUsesWithUndef(I: ExpressionChangedStart);
749
750 ExpressionChangedStart->moveBefore(InsertPos: I->getIterator());
751 ExpressionChangedStart =
752 cast<BinaryOperator>(Val: *ExpressionChangedStart->user_begin());
753 } while (true);
754 }
755
756 // Throw away any left over nodes from the original expression.
757 RedoInsts.insert_range(R&: NodesToRewrite);
758}
759
760/// Insert instructions before the instruction pointed to by BI,
761/// that computes the negative version of the value specified. The negative
762/// version of the value is returned, and BI is left pointing at the instruction
763/// that should be processed next by the reassociation pass.
764/// Also add intermediate instructions to the redo list that are modified while
765/// pushing the negates through adds. These will be revisited to see if
766/// additional opportunities have been exposed.
767static Value *NegateValue(Value *V, Instruction *BI,
768 ReassociatePass::OrderedSet &ToRedo) {
769 if (auto *C = dyn_cast<Constant>(Val: V)) {
770 const DataLayout &DL = BI->getDataLayout();
771 Constant *Res = C->getType()->isFPOrFPVectorTy()
772 ? ConstantFoldUnaryOpOperand(Opcode: Instruction::FNeg, Op: C, DL)
773 : ConstantExpr::getNeg(C);
774 if (Res)
775 return Res;
776 }
777
778 // We are trying to expose opportunity for reassociation. One of the things
779 // that we want to do to achieve this is to push a negation as deep into an
780 // expression chain as possible, to expose the add instructions. In practice,
781 // this means that we turn this:
782 // X = -(A+12+C+D) into X = -A + -12 + -C + -D = -12 + -A + -C + -D
783 // so that later, a: Y = 12+X could get reassociated with the -12 to eliminate
784 // the constants. We assume that instcombine will clean up the mess later if
785 // we introduce tons of unnecessary negation instructions.
786 //
787 if (BinaryOperator *I =
788 isReassociableOp(V, Opcode1: Instruction::Add, Opcode2: Instruction::FAdd)) {
789 // Push the negates through the add.
790 I->setOperand(i_nocapture: 0, Val_nocapture: NegateValue(V: I->getOperand(i_nocapture: 0), BI, ToRedo));
791 I->setOperand(i_nocapture: 1, Val_nocapture: NegateValue(V: I->getOperand(i_nocapture: 1), BI, ToRedo));
792 if (I->getOpcode() == Instruction::Add) {
793 I->setHasNoUnsignedWrap(false);
794 I->setHasNoSignedWrap(false);
795 }
796
797 // We must move the add instruction here, because the neg instructions do
798 // not dominate the old add instruction in general. By moving it, we are
799 // assured that the neg instructions we just inserted dominate the
800 // instruction we are about to insert after them.
801 //
802 I->moveBefore(InsertPos: BI->getIterator());
803 I->setName(I->getName()+".neg");
804
805 // Add the intermediate negates to the redo list as processing them later
806 // could expose more reassociating opportunities.
807 ToRedo.insert(X: I);
808 return I;
809 }
810
811 // Okay, we need to materialize a negated version of V with an instruction.
812 // Scan the use lists of V to see if we have one already.
813 for (User *U : V->users()) {
814 if (!match(V: U, P: m_Neg(V: m_Value())) && !match(V: U, P: m_FNeg(X: m_Value())))
815 continue;
816
817 // We found one! Now we have to make sure that the definition dominates
818 // this use. We do this by moving it to the entry block (if it is a
819 // non-instruction value) or right after the definition. These negates will
820 // be zapped by reassociate later, so we don't need much finesse here.
821 Instruction *TheNeg = dyn_cast<Instruction>(Val: U);
822
823 // We can't safely propagate a vector zero constant with poison/undef lanes.
824 Constant *C;
825 if (match(V: TheNeg, P: m_BinOp(L: m_Constant(C), R: m_Value())) &&
826 C->containsUndefOrPoisonElement())
827 continue;
828
829 // Verify that the negate is in this function, V might be a constant expr.
830 if (!TheNeg ||
831 TheNeg->getParent()->getParent() != BI->getParent()->getParent())
832 continue;
833
834 BasicBlock::iterator InsertPt;
835 if (Instruction *InstInput = dyn_cast<Instruction>(Val: V)) {
836 auto InsertPtOpt = InstInput->getInsertionPointAfterDef();
837 if (!InsertPtOpt)
838 continue;
839 InsertPt = *InsertPtOpt;
840 } else {
841 InsertPt = TheNeg->getFunction()
842 ->getEntryBlock()
843 .getFirstNonPHIOrDbg()
844 ->getIterator();
845 }
846
847 // Check that if TheNeg is moved out of its parent block, we drop its
848 // debug location to avoid extra coverage.
849 // See test dropping_debugloc_the_neg.ll for a detailed example.
850 if (TheNeg->getParent() != InsertPt->getParent())
851 TheNeg->dropLocation();
852 TheNeg->moveBefore(BB&: *InsertPt->getParent(), I: InsertPt);
853
854 if (TheNeg->getOpcode() == Instruction::Sub) {
855 TheNeg->setHasNoUnsignedWrap(false);
856 TheNeg->setHasNoSignedWrap(false);
857 } else {
858 TheNeg->andIRFlags(V: BI);
859 }
860 ToRedo.insert(X: TheNeg);
861 return TheNeg;
862 }
863
864 // Insert a 'neg' instruction that subtracts the value from zero to get the
865 // negation.
866 Instruction *NewNeg =
867 CreateNeg(S1: V, Name: V->getName() + ".neg", InsertBefore: BI->getIterator(), FlagsOp: BI);
868 // NewNeg is generated to potentially replace BI, so use its DebugLoc.
869 NewNeg->setDebugLoc(BI->getDebugLoc());
870 ToRedo.insert(X: NewNeg);
871 return NewNeg;
872}
873
874// See if this `or` looks like an load widening reduction, i.e. that it
875// consists of an `or`/`shl`/`zext`/`load` nodes only. Note that we don't
876// ensure that the pattern is *really* a load widening reduction,
877// we do not ensure that it can really be replaced with a widened load,
878// only that it mostly looks like one.
879static bool isLoadCombineCandidate(Instruction *Or) {
880 SmallVector<Instruction *, 8> Worklist;
881 SmallSet<Instruction *, 8> Visited;
882
883 auto Enqueue = [&](Value *V) {
884 auto *I = dyn_cast<Instruction>(Val: V);
885 // Each node of an `or` reduction must be an instruction,
886 if (!I)
887 return false; // Node is certainly not part of an `or` load reduction.
888 // Only process instructions we have never processed before.
889 if (Visited.insert(Ptr: I).second)
890 Worklist.emplace_back(Args&: I);
891 return true; // Will need to look at parent nodes.
892 };
893
894 if (!Enqueue(Or))
895 return false; // Not an `or` reduction pattern.
896
897 while (!Worklist.empty()) {
898 auto *I = Worklist.pop_back_val();
899
900 // Okay, which instruction is this node?
901 switch (I->getOpcode()) {
902 case Instruction::Or:
903 // Got an `or` node. That's fine, just recurse into it's operands.
904 for (Value *Op : I->operands())
905 if (!Enqueue(Op))
906 return false; // Not an `or` reduction pattern.
907 continue;
908
909 case Instruction::Shl:
910 case Instruction::ZExt:
911 // `shl`/`zext` nodes are fine, just recurse into their base operand.
912 if (!Enqueue(I->getOperand(i: 0)))
913 return false; // Not an `or` reduction pattern.
914 continue;
915
916 case Instruction::Load:
917 // Perfect, `load` node means we've reached an edge of the graph.
918 continue;
919
920 default: // Unknown node.
921 return false; // Not an `or` reduction pattern.
922 }
923 }
924
925 return true;
926}
927
928/// Return true if it may be profitable to convert this (X|Y) into (X+Y).
929static bool shouldConvertOrWithNoCommonBitsToAdd(Instruction *Or) {
930 // Don't bother to convert this up unless either the LHS is an associable add
931 // or subtract or mul or if this is only used by one of the above.
932 // This is only a compile-time improvement, it is not needed for correctness!
933 auto isInteresting = [](Value *V) {
934 for (auto Op : {Instruction::Add, Instruction::Sub, Instruction::Mul,
935 Instruction::Shl})
936 if (isReassociableOp(V, Opcode: Op))
937 return true;
938 return false;
939 };
940
941 if (any_of(Range: Or->operands(), P: isInteresting))
942 return true;
943
944 Value *VB = Or->user_back();
945 if (Or->hasOneUse() && isInteresting(VB))
946 return true;
947
948 return false;
949}
950
951/// If we have (X|Y), and iff X and Y have no common bits set,
952/// transform this into (X+Y) to allow arithmetics reassociation.
953static BinaryOperator *convertOrWithNoCommonBitsToAdd(Instruction *Or) {
954 // Convert an or into an add.
955 BinaryOperator *New = CreateAdd(S1: Or->getOperand(i: 0), S2: Or->getOperand(i: 1), Name: "",
956 InsertBefore: Or->getIterator(), FlagsOp: Or);
957 New->setHasNoSignedWrap();
958 New->setHasNoUnsignedWrap();
959 New->takeName(V: Or);
960
961 // Everyone now refers to the add instruction.
962 Or->replaceAllUsesWith(V: New);
963 New->setDebugLoc(Or->getDebugLoc());
964
965 LLVM_DEBUG(dbgs() << "Converted or into an add: " << *New << '\n');
966 return New;
967}
968
969/// Return true if we should break up this subtract of X-Y into (X + -Y).
970static bool ShouldBreakUpSubtract(Instruction *Sub) {
971 // If this is a negation, we can't split it up!
972 if (match(V: Sub, P: m_Neg(V: m_Value())) || match(V: Sub, P: m_FNeg(X: m_Value())))
973 return false;
974
975 // Don't breakup X - undef.
976 if (isa<UndefValue>(Val: Sub->getOperand(i: 1)))
977 return false;
978
979 // Don't bother to break this up unless either the LHS is an associable add or
980 // subtract or if this is only used by one.
981 Value *V0 = Sub->getOperand(i: 0);
982 if (isReassociableOp(V: V0, Opcode1: Instruction::Add, Opcode2: Instruction::FAdd) ||
983 isReassociableOp(V: V0, Opcode1: Instruction::Sub, Opcode2: Instruction::FSub))
984 return true;
985 Value *V1 = Sub->getOperand(i: 1);
986 if (isReassociableOp(V: V1, Opcode1: Instruction::Add, Opcode2: Instruction::FAdd) ||
987 isReassociableOp(V: V1, Opcode1: Instruction::Sub, Opcode2: Instruction::FSub))
988 return true;
989 Value *VB = Sub->user_back();
990 if (Sub->hasOneUse() &&
991 (isReassociableOp(V: VB, Opcode1: Instruction::Add, Opcode2: Instruction::FAdd) ||
992 isReassociableOp(V: VB, Opcode1: Instruction::Sub, Opcode2: Instruction::FSub)))
993 return true;
994
995 return false;
996}
997
998/// If we have (X-Y), and if either X is an add, or if this is only used by an
999/// add, transform this into (X+(0-Y)) to promote better reassociation.
1000static BinaryOperator *BreakUpSubtract(Instruction *Sub,
1001 ReassociatePass::OrderedSet &ToRedo) {
1002 // Convert a subtract into an add and a neg instruction. This allows sub
1003 // instructions to be commuted with other add instructions.
1004 //
1005 // Calculate the negative value of Operand 1 of the sub instruction,
1006 // and set it as the RHS of the add instruction we just made.
1007 Value *NegVal = NegateValue(V: Sub->getOperand(i: 1), BI: Sub, ToRedo);
1008 BinaryOperator *New =
1009 CreateAdd(S1: Sub->getOperand(i: 0), S2: NegVal, Name: "", InsertBefore: Sub->getIterator(), FlagsOp: Sub);
1010 Sub->setOperand(i: 0, Val: Constant::getNullValue(Ty: Sub->getType())); // Drop use of op.
1011 Sub->setOperand(i: 1, Val: Constant::getNullValue(Ty: Sub->getType())); // Drop use of op.
1012 New->takeName(V: Sub);
1013
1014 // Everyone now refers to the add instruction.
1015 Sub->replaceAllUsesWith(V: New);
1016 New->setDebugLoc(Sub->getDebugLoc());
1017
1018 LLVM_DEBUG(dbgs() << "Negated: " << *New << '\n');
1019 return New;
1020}
1021
1022/// If this is a shift of a reassociable multiply or is used by one, change
1023/// this into a multiply by a constant to assist with further reassociation.
1024static BinaryOperator *ConvertShiftToMul(Instruction *Shl) {
1025 Constant *MulCst = ConstantInt::get(Ty: Shl->getType(), V: 1);
1026 auto *SA = cast<ConstantInt>(Val: Shl->getOperand(i: 1));
1027 MulCst = ConstantFoldBinaryInstruction(Opcode: Instruction::Shl, V1: MulCst, V2: SA);
1028 assert(MulCst && "Constant folding of immediate constants failed");
1029
1030 BinaryOperator *Mul = BinaryOperator::CreateMul(V1: Shl->getOperand(i: 0), V2: MulCst,
1031 Name: "", InsertBefore: Shl->getIterator());
1032 Shl->setOperand(i: 0, Val: PoisonValue::get(T: Shl->getType())); // Drop use of op.
1033 Mul->takeName(V: Shl);
1034
1035 // Everyone now refers to the mul instruction.
1036 Shl->replaceAllUsesWith(V: Mul);
1037 Mul->setDebugLoc(Shl->getDebugLoc());
1038
1039 // We can safely preserve the nuw flag in all cases. It's also safe to turn a
1040 // nuw nsw shl into a nuw nsw mul. However, nsw in isolation requires special
1041 // handling. It can be preserved as long as we're not left shifting by
1042 // bitwidth - 1.
1043 bool NSW = cast<BinaryOperator>(Val: Shl)->hasNoSignedWrap();
1044 bool NUW = cast<BinaryOperator>(Val: Shl)->hasNoUnsignedWrap();
1045 unsigned BitWidth = Shl->getType()->getScalarSizeInBits();
1046 if (NSW && (NUW || SA->getValue().ult(RHS: BitWidth - 1)))
1047 Mul->setHasNoSignedWrap(true);
1048 Mul->setHasNoUnsignedWrap(NUW);
1049 return Mul;
1050}
1051
1052/// Scan backwards and forwards among values with the same rank as element i
1053/// to see if X exists. If X does not exist, return i. This is useful when
1054/// scanning for 'x' when we see '-x' because they both get the same rank.
1055static unsigned FindInOperandList(const SmallVectorImpl<ValueEntry> &Ops,
1056 unsigned i, Value *X) {
1057 unsigned XRank = Ops[i].Rank;
1058 unsigned e = Ops.size();
1059 for (unsigned j = i+1; j != e && Ops[j].Rank == XRank; ++j) {
1060 if (Ops[j].Op == X)
1061 return j;
1062 if (Instruction *I1 = dyn_cast<Instruction>(Val: Ops[j].Op))
1063 if (Instruction *I2 = dyn_cast<Instruction>(Val: X))
1064 if (I1->isIdenticalTo(I: I2))
1065 return j;
1066 }
1067 // Scan backwards.
1068 for (unsigned j = i-1; j != ~0U && Ops[j].Rank == XRank; --j) {
1069 if (Ops[j].Op == X)
1070 return j;
1071 if (Instruction *I1 = dyn_cast<Instruction>(Val: Ops[j].Op))
1072 if (Instruction *I2 = dyn_cast<Instruction>(Val: X))
1073 if (I1->isIdenticalTo(I: I2))
1074 return j;
1075 }
1076 return i;
1077}
1078
1079/// Emit a tree of add instructions, summing Ops together
1080/// and returning the result. Insert the tree before I.
1081static Value *EmitAddTreeOfValues(Instruction *I,
1082 SmallVectorImpl<WeakTrackingVH> &Ops) {
1083 if (Ops.size() == 1) return Ops.back();
1084
1085 Value *V1 = Ops.pop_back_val();
1086 Value *V2 = EmitAddTreeOfValues(I, Ops);
1087 auto *NewAdd = CreateAdd(S1: V2, S2: V1, Name: "reass.add", InsertBefore: I->getIterator(), FlagsOp: I);
1088 NewAdd->setDebugLoc(I->getDebugLoc());
1089 return NewAdd;
1090}
1091
1092/// If V is an expression tree that is a multiplication sequence,
1093/// and if this sequence contains a multiply by Factor,
1094/// remove Factor from the tree and return the new tree.
1095/// If new instructions are inserted to generate this tree, DL should be used
1096/// as the DebugLoc for these instructions.
1097Value *ReassociatePass::RemoveFactorFromExpression(Value *V, Value *Factor,
1098 DebugLoc DL) {
1099 BinaryOperator *BO = isReassociableOp(V, Opcode1: Instruction::Mul, Opcode2: Instruction::FMul);
1100 if (!BO)
1101 return nullptr;
1102
1103 SmallVector<RepeatedValue, 8> Tree;
1104 OverflowTracking Flags;
1105 MadeChange |= LinearizeExprTree(I: BO, Ops&: Tree, ToRedo&: RedoInsts, Flags);
1106 SmallVector<ValueEntry, 8> Factors;
1107 Factors.reserve(N: Tree.size());
1108 for (const RepeatedValue &E : Tree)
1109 Factors.append(NumInputs: E.second, Elt: ValueEntry(getRank(V: E.first), E.first));
1110
1111 bool FoundFactor = false;
1112 bool NeedsNegate = false;
1113 for (unsigned i = 0, e = Factors.size(); i != e; ++i) {
1114 if (Factors[i].Op == Factor) {
1115 FoundFactor = true;
1116 Factors.erase(CI: Factors.begin()+i);
1117 break;
1118 }
1119
1120 // If this is a negative version of this factor, remove it.
1121 if (ConstantInt *FC1 = dyn_cast<ConstantInt>(Val: Factor)) {
1122 if (ConstantInt *FC2 = dyn_cast<ConstantInt>(Val: Factors[i].Op))
1123 if (FC1->getValue() == -FC2->getValue()) {
1124 FoundFactor = NeedsNegate = true;
1125 Factors.erase(CI: Factors.begin()+i);
1126 break;
1127 }
1128 } else if (ConstantFP *FC1 = dyn_cast<ConstantFP>(Val: Factor)) {
1129 if (ConstantFP *FC2 = dyn_cast<ConstantFP>(Val: Factors[i].Op)) {
1130 const APFloat &F1 = FC1->getValueAPF();
1131 APFloat F2(FC2->getValueAPF());
1132 F2.changeSign();
1133 if (F1 == F2) {
1134 FoundFactor = NeedsNegate = true;
1135 Factors.erase(CI: Factors.begin() + i);
1136 break;
1137 }
1138 }
1139 }
1140 }
1141
1142 if (!FoundFactor) {
1143 // Make sure to restore the operands to the expression tree.
1144 RewriteExprTree(I: BO, Ops&: Factors, Flags);
1145 return nullptr;
1146 }
1147
1148 BasicBlock::iterator InsertPt = ++BO->getIterator();
1149
1150 // If this was just a single multiply, remove the multiply and return the only
1151 // remaining operand.
1152 if (Factors.size() == 1) {
1153 RedoInsts.insert(X: BO);
1154 V = Factors[0].Op;
1155 } else {
1156 RewriteExprTree(I: BO, Ops&: Factors, Flags);
1157 V = BO;
1158 }
1159
1160 if (NeedsNegate) {
1161 V = CreateNeg(S1: V, Name: "neg", InsertBefore: InsertPt, FlagsOp: BO);
1162 cast<Instruction>(Val: V)->setDebugLoc(DL);
1163 }
1164
1165 return V;
1166}
1167
1168/// If V is a single-use multiply, recursively add its operands as factors,
1169/// otherwise add V to the list of factors.
1170///
1171/// Ops is the top-level list of add operands we're trying to factor.
1172static void FindSingleUseMultiplyFactors(Value *V,
1173 SmallVectorImpl<Value*> &Factors) {
1174 BinaryOperator *BO = isReassociableOp(V, Opcode1: Instruction::Mul, Opcode2: Instruction::FMul);
1175 if (!BO) {
1176 Factors.push_back(Elt: V);
1177 return;
1178 }
1179
1180 // Otherwise, add the LHS and RHS to the list of factors.
1181 FindSingleUseMultiplyFactors(V: BO->getOperand(i_nocapture: 1), Factors);
1182 FindSingleUseMultiplyFactors(V: BO->getOperand(i_nocapture: 0), Factors);
1183}
1184
1185/// Optimize a series of operands to an 'and', 'or', or 'xor' instruction.
1186/// This optimizes based on identities. If it can be reduced to a single Value,
1187/// it is returned, otherwise the Ops list is mutated as necessary.
1188static Value *OptimizeAndOrXor(unsigned Opcode,
1189 SmallVectorImpl<ValueEntry> &Ops) {
1190 // Scan the operand lists looking for X and ~X pairs, along with X,X pairs.
1191 // If we find any, we can simplify the expression. X&~X == 0, X|~X == -1.
1192 for (unsigned i = 0, e = Ops.size(); i != e; ++i) {
1193 // First, check for X and ~X in the operand list.
1194 assert(i < Ops.size());
1195 Value *X;
1196 if (match(V: Ops[i].Op, P: m_Not(V: m_Value(V&: X)))) { // Cannot occur for ^.
1197 unsigned FoundX = FindInOperandList(Ops, i, X);
1198 if (FoundX != i) {
1199 if (Opcode == Instruction::And) // ...&X&~X = 0
1200 return Constant::getNullValue(Ty: X->getType());
1201
1202 if (Opcode == Instruction::Or) // ...|X|~X = -1
1203 return Constant::getAllOnesValue(Ty: X->getType());
1204 }
1205 }
1206
1207 // Next, check for duplicate pairs of values, which we assume are next to
1208 // each other, due to our sorting criteria.
1209 assert(i < Ops.size());
1210 if (i+1 != Ops.size() && Ops[i+1].Op == Ops[i].Op) {
1211 if (Opcode == Instruction::And || Opcode == Instruction::Or) {
1212 // Drop duplicate values for And and Or.
1213 Ops.erase(CI: Ops.begin()+i);
1214 --i; --e;
1215 ++NumAnnihil;
1216 continue;
1217 }
1218
1219 // Drop pairs of values for Xor.
1220 assert(Opcode == Instruction::Xor);
1221 if (e == 2)
1222 return Constant::getNullValue(Ty: Ops[0].Op->getType());
1223
1224 // Y ^ X^X -> Y
1225 Ops.erase(CS: Ops.begin()+i, CE: Ops.begin()+i+2);
1226 i -= 1; e -= 2;
1227 ++NumAnnihil;
1228 }
1229 }
1230 return nullptr;
1231}
1232
1233/// Helper function of CombineXorOpnd(). It creates a bitwise-and
1234/// instruction with the given two operands, and return the resulting
1235/// instruction. There are two special cases: 1) if the constant operand is 0,
1236/// it will return NULL. 2) if the constant is ~0, the symbolic operand will
1237/// be returned.
1238static Value *createAndInstr(BasicBlock::iterator InsertBefore, Value *Opnd,
1239 const APInt &ConstOpnd) {
1240 if (ConstOpnd.isZero())
1241 return nullptr;
1242
1243 if (ConstOpnd.isAllOnes())
1244 return Opnd;
1245
1246 Instruction *I = BinaryOperator::CreateAnd(
1247 V1: Opnd, V2: ConstantInt::get(Ty: Opnd->getType(), V: ConstOpnd), Name: "and.ra",
1248 InsertBefore);
1249 I->setDebugLoc(InsertBefore->getDebugLoc());
1250 return I;
1251}
1252
1253// Helper function of OptimizeXor(). It tries to simplify "Opnd1 ^ ConstOpnd"
1254// into "R ^ C", where C would be 0, and R is a symbolic value.
1255//
1256// If it was successful, true is returned, and the "R" and "C" is returned
1257// via "Res" and "ConstOpnd", respectively; otherwise, false is returned,
1258// and both "Res" and "ConstOpnd" remain unchanged.
1259bool ReassociatePass::CombineXorOpnd(BasicBlock::iterator It, XorOpnd *Opnd1,
1260 APInt &ConstOpnd, Value *&Res) {
1261 // Xor-Rule 1: (x | c1) ^ c2 = (x | c1) ^ (c1 ^ c1) ^ c2
1262 // = ((x | c1) ^ c1) ^ (c1 ^ c2)
1263 // = (x & ~c1) ^ (c1 ^ c2)
1264 // It is useful only when c1 == c2.
1265 if (!Opnd1->isOrExpr() || Opnd1->getConstPart().isZero())
1266 return false;
1267
1268 if (!Opnd1->getValue()->hasOneUse())
1269 return false;
1270
1271 const APInt &C1 = Opnd1->getConstPart();
1272 if (C1 != ConstOpnd)
1273 return false;
1274
1275 Value *X = Opnd1->getSymbolicPart();
1276 Res = createAndInstr(InsertBefore: It, Opnd: X, ConstOpnd: ~C1);
1277 // ConstOpnd was C2, now C1 ^ C2.
1278 ConstOpnd ^= C1;
1279
1280 if (Instruction *T = dyn_cast<Instruction>(Val: Opnd1->getValue()))
1281 RedoInsts.insert(X: T);
1282 return true;
1283}
1284
1285// Helper function of OptimizeXor(). It tries to simplify
1286// "Opnd1 ^ Opnd2 ^ ConstOpnd" into "R ^ C", where C would be 0, and R is a
1287// symbolic value.
1288//
1289// If it was successful, true is returned, and the "R" and "C" is returned
1290// via "Res" and "ConstOpnd", respectively (If the entire expression is
1291// evaluated to a constant, the Res is set to NULL); otherwise, false is
1292// returned, and both "Res" and "ConstOpnd" remain unchanged.
1293bool ReassociatePass::CombineXorOpnd(BasicBlock::iterator It, XorOpnd *Opnd1,
1294 XorOpnd *Opnd2, APInt &ConstOpnd,
1295 Value *&Res) {
1296 Value *X = Opnd1->getSymbolicPart();
1297 if (X != Opnd2->getSymbolicPart())
1298 return false;
1299
1300 // This many instruction become dead.(At least "Opnd1 ^ Opnd2" will die.)
1301 int DeadInstNum = 1;
1302 if (Opnd1->getValue()->hasOneUse())
1303 DeadInstNum++;
1304 if (Opnd2->getValue()->hasOneUse())
1305 DeadInstNum++;
1306
1307 // Xor-Rule 2:
1308 // (x | c1) ^ (x & c2)
1309 // = (x|c1) ^ (x&c2) ^ (c1 ^ c1) = ((x|c1) ^ c1) ^ (x & c2) ^ c1
1310 // = (x & ~c1) ^ (x & c2) ^ c1 // Xor-Rule 1
1311 // = (x & c3) ^ c1, where c3 = ~c1 ^ c2 // Xor-rule 3
1312 //
1313 if (Opnd1->isOrExpr() != Opnd2->isOrExpr()) {
1314 if (Opnd2->isOrExpr())
1315 std::swap(a&: Opnd1, b&: Opnd2);
1316
1317 const APInt &C1 = Opnd1->getConstPart();
1318 const APInt &C2 = Opnd2->getConstPart();
1319 APInt C3((~C1) ^ C2);
1320
1321 // Do not increase code size!
1322 if (!C3.isZero() && !C3.isAllOnes()) {
1323 int NewInstNum = ConstOpnd.getBoolValue() ? 1 : 2;
1324 if (NewInstNum > DeadInstNum)
1325 return false;
1326 }
1327
1328 Res = createAndInstr(InsertBefore: It, Opnd: X, ConstOpnd: C3);
1329 ConstOpnd ^= C1;
1330 } else if (Opnd1->isOrExpr()) {
1331 // Xor-Rule 3: (x | c1) ^ (x | c2) = (x & c3) ^ c3 where c3 = c1 ^ c2
1332 //
1333 const APInt &C1 = Opnd1->getConstPart();
1334 const APInt &C2 = Opnd2->getConstPart();
1335 APInt C3 = C1 ^ C2;
1336
1337 // Do not increase code size
1338 if (!C3.isZero() && !C3.isAllOnes()) {
1339 int NewInstNum = ConstOpnd.getBoolValue() ? 1 : 2;
1340 if (NewInstNum > DeadInstNum)
1341 return false;
1342 }
1343
1344 Res = createAndInstr(InsertBefore: It, Opnd: X, ConstOpnd: C3);
1345 ConstOpnd ^= C3;
1346 } else {
1347 // Xor-Rule 4: (x & c1) ^ (x & c2) = (x & (c1^c2))
1348 //
1349 const APInt &C1 = Opnd1->getConstPart();
1350 const APInt &C2 = Opnd2->getConstPart();
1351 APInt C3 = C1 ^ C2;
1352 Res = createAndInstr(InsertBefore: It, Opnd: X, ConstOpnd: C3);
1353 }
1354
1355 // Put the original operands in the Redo list; hope they will be deleted
1356 // as dead code.
1357 if (Instruction *T = dyn_cast<Instruction>(Val: Opnd1->getValue()))
1358 RedoInsts.insert(X: T);
1359 if (Instruction *T = dyn_cast<Instruction>(Val: Opnd2->getValue()))
1360 RedoInsts.insert(X: T);
1361
1362 return true;
1363}
1364
1365/// Optimize a series of operands to an 'xor' instruction. If it can be reduced
1366/// to a single Value, it is returned, otherwise the Ops list is mutated as
1367/// necessary.
1368Value *ReassociatePass::OptimizeXor(Instruction *I,
1369 SmallVectorImpl<ValueEntry> &Ops) {
1370 if (Value *V = OptimizeAndOrXor(Opcode: Instruction::Xor, Ops))
1371 return V;
1372
1373 if (Ops.size() == 1)
1374 return nullptr;
1375
1376 SmallVector<XorOpnd, 8> Opnds;
1377 SmallVector<XorOpnd*, 8> OpndPtrs;
1378 Type *Ty = Ops[0].Op->getType();
1379 APInt ConstOpnd(Ty->getScalarSizeInBits(), 0);
1380
1381 // Step 1: Convert ValueEntry to XorOpnd
1382 for (const ValueEntry &Op : Ops) {
1383 Value *V = Op.Op;
1384 const APInt *C;
1385 // TODO: Support non-splat vectors.
1386 if (match(V, P: m_APInt(Res&: C))) {
1387 ConstOpnd ^= *C;
1388 } else {
1389 XorOpnd O(V);
1390 O.setSymbolicRank(getRank(V: O.getSymbolicPart()));
1391 Opnds.push_back(Elt: O);
1392 }
1393 }
1394
1395 // NOTE: From this point on, do *NOT* add/delete element to/from "Opnds".
1396 // It would otherwise invalidate the "Opnds"'s iterator, and hence invalidate
1397 // the "OpndPtrs" as well. For the similar reason, do not fuse this loop
1398 // with the previous loop --- the iterator of the "Opnds" may be invalidated
1399 // when new elements are added to the vector.
1400 for (XorOpnd &Op : Opnds)
1401 OpndPtrs.push_back(Elt: &Op);
1402
1403 // Step 2: Sort the Xor-Operands in a way such that the operands containing
1404 // the same symbolic value cluster together. For instance, the input operand
1405 // sequence ("x | 123", "y & 456", "x & 789") will be sorted into:
1406 // ("x | 123", "x & 789", "y & 456").
1407 //
1408 // The purpose is twofold:
1409 // 1) Cluster together the operands sharing the same symbolic-value.
1410 // 2) Operand having smaller symbolic-value-rank is permuted earlier, which
1411 // could potentially shorten crital path, and expose more loop-invariants.
1412 // Note that values' rank are basically defined in RPO order (FIXME).
1413 // So, if Rank(X) < Rank(Y) < Rank(Z), it means X is defined earlier
1414 // than Y which is defined earlier than Z. Permute "x | 1", "Y & 2",
1415 // "z" in the order of X-Y-Z is better than any other orders.
1416 llvm::stable_sort(Range&: OpndPtrs, C: [](XorOpnd *LHS, XorOpnd *RHS) {
1417 return LHS->getSymbolicRank() < RHS->getSymbolicRank();
1418 });
1419
1420 // Step 3: Combine adjacent operands
1421 XorOpnd *PrevOpnd = nullptr;
1422 bool Changed = false;
1423 for (unsigned i = 0, e = Opnds.size(); i < e; i++) {
1424 XorOpnd *CurrOpnd = OpndPtrs[i];
1425 // The combined value
1426 Value *CV;
1427
1428 // Step 3.1: Try simplifying "CurrOpnd ^ ConstOpnd"
1429 if (!ConstOpnd.isZero() &&
1430 CombineXorOpnd(It: I->getIterator(), Opnd1: CurrOpnd, ConstOpnd, Res&: CV)) {
1431 Changed = true;
1432 if (CV)
1433 *CurrOpnd = XorOpnd(CV);
1434 else {
1435 CurrOpnd->Invalidate();
1436 continue;
1437 }
1438 }
1439
1440 if (!PrevOpnd || CurrOpnd->getSymbolicPart() != PrevOpnd->getSymbolicPart()) {
1441 PrevOpnd = CurrOpnd;
1442 continue;
1443 }
1444
1445 // step 3.2: When previous and current operands share the same symbolic
1446 // value, try to simplify "PrevOpnd ^ CurrOpnd ^ ConstOpnd"
1447 if (CombineXorOpnd(It: I->getIterator(), Opnd1: CurrOpnd, Opnd2: PrevOpnd, ConstOpnd, Res&: CV)) {
1448 // Remove previous operand
1449 PrevOpnd->Invalidate();
1450 if (CV) {
1451 *CurrOpnd = XorOpnd(CV);
1452 PrevOpnd = CurrOpnd;
1453 } else {
1454 CurrOpnd->Invalidate();
1455 PrevOpnd = nullptr;
1456 }
1457 Changed = true;
1458 }
1459 }
1460
1461 // Step 4: Reassemble the Ops
1462 if (Changed) {
1463 Ops.clear();
1464 for (const XorOpnd &O : Opnds) {
1465 if (O.isInvalid())
1466 continue;
1467 ValueEntry VE(getRank(V: O.getValue()), O.getValue());
1468 Ops.push_back(Elt: VE);
1469 }
1470 if (!ConstOpnd.isZero()) {
1471 Value *C = ConstantInt::get(Ty, V: ConstOpnd);
1472 ValueEntry VE(getRank(V: C), C);
1473 Ops.push_back(Elt: VE);
1474 }
1475 unsigned Sz = Ops.size();
1476 if (Sz == 1)
1477 return Ops.back().Op;
1478 if (Sz == 0) {
1479 assert(ConstOpnd.isZero());
1480 return ConstantInt::get(Ty, V: ConstOpnd);
1481 }
1482 }
1483
1484 return nullptr;
1485}
1486
1487/// Optimize a series of operands to an 'add' instruction. This
1488/// optimizes based on identities. If it can be reduced to a single Value, it
1489/// is returned, otherwise the Ops list is mutated as necessary.
1490Value *ReassociatePass::OptimizeAdd(Instruction *I,
1491 SmallVectorImpl<ValueEntry> &Ops) {
1492 // Scan the operand lists looking for X and -X pairs. If we find any, we
1493 // can simplify expressions like X+-X == 0 and X+~X ==-1. While we're at it,
1494 // scan for any
1495 // duplicates. We want to canonicalize Y+Y+Y+Z -> 3*Y+Z.
1496
1497 for (unsigned i = 0, e = Ops.size(); i != e; ++i) {
1498 Value *TheOp = Ops[i].Op;
1499 // Check to see if we've seen this operand before. If so, we factor all
1500 // instances of the operand together. Due to our sorting criteria, we know
1501 // that these need to be next to each other in the vector.
1502 if (i+1 != Ops.size() && Ops[i+1].Op == TheOp) {
1503 // Rescan the list, remove all instances of this operand from the expr.
1504 unsigned NumFound = 0;
1505 do {
1506 Ops.erase(CI: Ops.begin()+i);
1507 ++NumFound;
1508 } while (i != Ops.size() && Ops[i].Op == TheOp);
1509
1510 LLVM_DEBUG(dbgs() << "\nFACTORING [" << NumFound << "]: " << *TheOp
1511 << '\n');
1512 ++NumFactor;
1513
1514 // Insert a new multiply.
1515 Type *Ty = TheOp->getType();
1516 Constant *C = Ty->isIntOrIntVectorTy() ?
1517 ConstantInt::get(Ty, V: NumFound) : ConstantFP::get(Ty, V: NumFound);
1518 Instruction *Mul = CreateMul(S1: TheOp, S2: C, Name: "factor", InsertBefore: I->getIterator(), FlagsOp: I);
1519 Mul->setDebugLoc(I->getDebugLoc());
1520
1521 // Now that we have inserted a multiply, optimize it. This allows us to
1522 // handle cases that require multiple factoring steps, such as this:
1523 // (X*2) + (X*2) + (X*2) -> (X*2)*3 -> X*6
1524 RedoInsts.insert(X: Mul);
1525
1526 // If every add operand was a duplicate, return the multiply.
1527 if (Ops.empty())
1528 return Mul;
1529
1530 // Otherwise, we had some input that didn't have the dupe, such as
1531 // "A + A + B" -> "A*2 + B". Add the new multiply to the list of
1532 // things being added by this operation.
1533 Ops.insert(I: Ops.begin(), Elt: ValueEntry(getRank(V: Mul), Mul));
1534
1535 --i;
1536 e = Ops.size();
1537 continue;
1538 }
1539
1540 // Check for X and -X or X and ~X in the operand list.
1541 Value *X;
1542 if (!match(V: TheOp, P: m_Neg(V: m_Value(V&: X))) && !match(V: TheOp, P: m_Not(V: m_Value(V&: X))) &&
1543 !match(V: TheOp, P: m_FNeg(X: m_Value(V&: X))))
1544 continue;
1545
1546 unsigned FoundX = FindInOperandList(Ops, i, X);
1547 if (FoundX == i)
1548 continue;
1549
1550 // Remove X and -X from the operand list.
1551 if (Ops.size() == 2 &&
1552 (match(V: TheOp, P: m_Neg(V: m_Value())) || match(V: TheOp, P: m_FNeg(X: m_Value()))))
1553 return Constant::getNullValue(Ty: X->getType());
1554
1555 // Remove X and ~X from the operand list.
1556 if (Ops.size() == 2 && match(V: TheOp, P: m_Not(V: m_Value())))
1557 return Constant::getAllOnesValue(Ty: X->getType());
1558
1559 Ops.erase(CI: Ops.begin()+i);
1560 if (i < FoundX)
1561 --FoundX;
1562 else
1563 --i; // Need to back up an extra one.
1564 Ops.erase(CI: Ops.begin()+FoundX);
1565 ++NumAnnihil;
1566 --i; // Revisit element.
1567 e -= 2; // Removed two elements.
1568
1569 // if X and ~X we append -1 to the operand list.
1570 if (match(V: TheOp, P: m_Not(V: m_Value()))) {
1571 Value *V = Constant::getAllOnesValue(Ty: X->getType());
1572 Ops.insert(I: Ops.end(), Elt: ValueEntry(getRank(V), V));
1573 e += 1;
1574 }
1575 }
1576
1577 // Scan the operand list, checking to see if there are any common factors
1578 // between operands. Consider something like A*A+A*B*C+D. We would like to
1579 // reassociate this to A*(A+B*C)+D, which reduces the number of multiplies.
1580 // To efficiently find this, we count the number of times a factor occurs
1581 // for any ADD operands that are MULs.
1582 DenseMap<Value*, unsigned> FactorOccurrences;
1583
1584 // Keep track of each multiply we see, to avoid triggering on (X*4)+(X*4)
1585 // where they are actually the same multiply.
1586 unsigned MaxOcc = 0;
1587 Value *MaxOccVal = nullptr;
1588 for (const ValueEntry &Op : Ops) {
1589 BinaryOperator *BOp =
1590 isReassociableOp(V: Op.Op, Opcode1: Instruction::Mul, Opcode2: Instruction::FMul);
1591 if (!BOp)
1592 continue;
1593
1594 // Compute all of the factors of this added value.
1595 SmallVector<Value*, 8> Factors;
1596 FindSingleUseMultiplyFactors(V: BOp, Factors);
1597 assert(Factors.size() > 1 && "Bad linearize!");
1598
1599 // Add one to FactorOccurrences for each unique factor in this op.
1600 SmallPtrSet<Value*, 8> Duplicates;
1601 for (Value *Factor : Factors) {
1602 if (!Duplicates.insert(Ptr: Factor).second)
1603 continue;
1604
1605 unsigned Occ = ++FactorOccurrences[Factor];
1606 if (Occ > MaxOcc) {
1607 MaxOcc = Occ;
1608 MaxOccVal = Factor;
1609 }
1610
1611 // If Factor is a negative constant, add the negated value as a factor
1612 // because we can percolate the negate out. Watch for minint, which
1613 // cannot be positivified.
1614 if (ConstantInt *CI = dyn_cast<ConstantInt>(Val: Factor)) {
1615 if (CI->isNegative() && !CI->isMinValue(IsSigned: true)) {
1616 Factor = ConstantInt::get(Context&: CI->getContext(), V: -CI->getValue());
1617 if (!Duplicates.insert(Ptr: Factor).second)
1618 continue;
1619 unsigned Occ = ++FactorOccurrences[Factor];
1620 if (Occ > MaxOcc) {
1621 MaxOcc = Occ;
1622 MaxOccVal = Factor;
1623 }
1624 }
1625 } else if (ConstantFP *CF = dyn_cast<ConstantFP>(Val: Factor)) {
1626 if (CF->isNegative()) {
1627 APFloat F(CF->getValueAPF());
1628 F.changeSign();
1629 Factor = ConstantFP::get(Context&: CF->getContext(), V: F);
1630 if (!Duplicates.insert(Ptr: Factor).second)
1631 continue;
1632 unsigned Occ = ++FactorOccurrences[Factor];
1633 if (Occ > MaxOcc) {
1634 MaxOcc = Occ;
1635 MaxOccVal = Factor;
1636 }
1637 }
1638 }
1639 }
1640 }
1641
1642 // If any factor occurred more than one time, we can pull it out.
1643 if (MaxOcc > 1) {
1644 LLVM_DEBUG(dbgs() << "\nFACTORING [" << MaxOcc << "]: " << *MaxOccVal
1645 << '\n');
1646 ++NumFactor;
1647
1648 // Create a new instruction that uses the MaxOccVal twice. If we don't do
1649 // this, we could otherwise run into situations where removing a factor
1650 // from an expression will drop a use of maxocc, and this can cause
1651 // RemoveFactorFromExpression on successive values to behave differently.
1652 Instruction *DummyInst =
1653 I->getType()->isIntOrIntVectorTy()
1654 ? BinaryOperator::CreateAdd(V1: MaxOccVal, V2: MaxOccVal)
1655 : BinaryOperator::CreateFAdd(V1: MaxOccVal, V2: MaxOccVal);
1656
1657 SmallVector<WeakTrackingVH, 4> NewMulOps;
1658 for (unsigned i = 0; i != Ops.size(); ++i) {
1659 // Only try to remove factors from expressions we're allowed to.
1660 BinaryOperator *BOp =
1661 isReassociableOp(V: Ops[i].Op, Opcode1: Instruction::Mul, Opcode2: Instruction::FMul);
1662 if (!BOp)
1663 continue;
1664
1665 if (Value *V = RemoveFactorFromExpression(V: Ops[i].Op, Factor: MaxOccVal,
1666 DL: I->getDebugLoc())) {
1667 // The factorized operand may occur several times. Convert them all in
1668 // one fell swoop.
1669 for (unsigned j = Ops.size(); j != i;) {
1670 --j;
1671 if (Ops[j].Op == Ops[i].Op) {
1672 NewMulOps.push_back(Elt: V);
1673 Ops.erase(CI: Ops.begin()+j);
1674 }
1675 }
1676 --i;
1677 }
1678 }
1679
1680 // No need for extra uses anymore.
1681 DummyInst->deleteValue();
1682
1683 unsigned NumAddedValues = NewMulOps.size();
1684 Value *V = EmitAddTreeOfValues(I, Ops&: NewMulOps);
1685
1686 // Now that we have inserted the add tree, optimize it. This allows us to
1687 // handle cases that require multiple factoring steps, such as this:
1688 // A*A*B + A*A*C --> A*(A*B+A*C) --> A*(A*(B+C))
1689 assert(NumAddedValues > 1 && "Each occurrence should contribute a value");
1690 (void)NumAddedValues;
1691 if (Instruction *VI = dyn_cast<Instruction>(Val: V))
1692 RedoInsts.insert(X: VI);
1693
1694 // Create the multiply.
1695 Instruction *V2 = CreateMul(S1: V, S2: MaxOccVal, Name: "reass.mul", InsertBefore: I->getIterator(), FlagsOp: I);
1696 V2->setDebugLoc(I->getDebugLoc());
1697
1698 // Rerun associate on the multiply in case the inner expression turned into
1699 // a multiply. We want to make sure that we keep things in canonical form.
1700 RedoInsts.insert(X: V2);
1701
1702 // If every add operand included the factor (e.g. "A*B + A*C"), then the
1703 // entire result expression is just the multiply "A*(B+C)".
1704 if (Ops.empty())
1705 return V2;
1706
1707 // Otherwise, we had some input that didn't have the factor, such as
1708 // "A*B + A*C + D" -> "A*(B+C) + D". Add the new multiply to the list of
1709 // things being added by this operation.
1710 Ops.insert(I: Ops.begin(), Elt: ValueEntry(getRank(V: V2), V2));
1711 }
1712
1713 return nullptr;
1714}
1715
1716/// Build up a vector of value/power pairs factoring a product.
1717///
1718/// Given a series of multiplication operands, build a vector of factors and
1719/// the powers each is raised to when forming the final product. Sort them in
1720/// the order of descending power.
1721///
1722/// (x*x) -> [(x, 2)]
1723/// ((x*x)*x) -> [(x, 3)]
1724/// ((((x*y)*x)*y)*x) -> [(x, 3), (y, 2)]
1725///
1726/// \returns Whether any factors have a power greater than one.
1727static bool collectMultiplyFactors(SmallVectorImpl<ValueEntry> &Ops,
1728 SmallVectorImpl<Factor> &Factors) {
1729 // FIXME: Have Ops be (ValueEntry, Multiplicity) pairs, simplifying this.
1730 // Compute the sum of powers of simplifiable factors.
1731 unsigned FactorPowerSum = 0;
1732 for (unsigned Idx = 1, Size = Ops.size(); Idx < Size; ++Idx) {
1733 Value *Op = Ops[Idx-1].Op;
1734
1735 // Count the number of occurrences of this value.
1736 unsigned Count = 1;
1737 for (; Idx < Size && Ops[Idx].Op == Op; ++Idx)
1738 ++Count;
1739 // Track for simplification all factors which occur 2 or more times.
1740 if (Count > 1)
1741 FactorPowerSum += Count;
1742 }
1743
1744 // We can only simplify factors if the sum of the powers of our simplifiable
1745 // factors is 4 or higher. When that is the case, we will *always* have
1746 // a simplification. This is an important invariant to prevent cyclicly
1747 // trying to simplify already minimal formations.
1748 if (FactorPowerSum < 4)
1749 return false;
1750
1751 // Now gather the simplifiable factors, removing them from Ops.
1752 FactorPowerSum = 0;
1753 for (unsigned Idx = 1; Idx < Ops.size(); ++Idx) {
1754 Value *Op = Ops[Idx-1].Op;
1755
1756 // Count the number of occurrences of this value.
1757 unsigned Count = 1;
1758 for (; Idx < Ops.size() && Ops[Idx].Op == Op; ++Idx)
1759 ++Count;
1760 if (Count == 1)
1761 continue;
1762 // Move an even number of occurrences to Factors.
1763 Count &= ~1U;
1764 Idx -= Count;
1765 FactorPowerSum += Count;
1766 Factors.push_back(Elt: Factor(Op, Count));
1767 Ops.erase(CS: Ops.begin()+Idx, CE: Ops.begin()+Idx+Count);
1768 }
1769
1770 // None of the adjustments above should have reduced the sum of factor powers
1771 // below our mininum of '4'.
1772 assert(FactorPowerSum >= 4);
1773
1774 llvm::stable_sort(Range&: Factors, C: [](const Factor &LHS, const Factor &RHS) {
1775 return LHS.Power > RHS.Power;
1776 });
1777 return true;
1778}
1779
1780/// Build a tree of multiplies, computing the product of Ops.
1781static Value *buildMultiplyTree(IRBuilderBase &Builder,
1782 SmallVectorImpl<Value*> &Ops) {
1783 if (Ops.size() == 1)
1784 return Ops.back();
1785
1786 Value *LHS = Ops.pop_back_val();
1787 do {
1788 if (LHS->getType()->isIntOrIntVectorTy())
1789 LHS = Builder.CreateMul(LHS, RHS: Ops.pop_back_val());
1790 else
1791 LHS = Builder.CreateFMul(L: LHS, R: Ops.pop_back_val());
1792 } while (!Ops.empty());
1793
1794 return LHS;
1795}
1796
1797/// Build a minimal multiplication DAG for (a^x)*(b^y)*(c^z)*...
1798///
1799/// Given a vector of values raised to various powers, where no two values are
1800/// equal and the powers are sorted in decreasing order, compute the minimal
1801/// DAG of multiplies to compute the final product, and return that product
1802/// value.
1803Value *
1804ReassociatePass::buildMinimalMultiplyDAG(IRBuilderBase &Builder,
1805 SmallVectorImpl<Factor> &Factors) {
1806 assert(Factors[0].Power);
1807 SmallVector<Value *, 4> OuterProduct;
1808 for (unsigned LastIdx = 0, Idx = 1, Size = Factors.size();
1809 Idx < Size && Factors[Idx].Power > 0; ++Idx) {
1810 if (Factors[Idx].Power != Factors[LastIdx].Power) {
1811 LastIdx = Idx;
1812 continue;
1813 }
1814
1815 // We want to multiply across all the factors with the same power so that
1816 // we can raise them to that power as a single entity. Build a mini tree
1817 // for that.
1818 SmallVector<Value *, 4> InnerProduct;
1819 InnerProduct.push_back(Elt: Factors[LastIdx].Base);
1820 do {
1821 InnerProduct.push_back(Elt: Factors[Idx].Base);
1822 ++Idx;
1823 } while (Idx < Size && Factors[Idx].Power == Factors[LastIdx].Power);
1824
1825 // Reset the base value of the first factor to the new expression tree.
1826 // We'll remove all the factors with the same power in a second pass.
1827 Value *M = Factors[LastIdx].Base = buildMultiplyTree(Builder, Ops&: InnerProduct);
1828 if (Instruction *MI = dyn_cast<Instruction>(Val: M))
1829 RedoInsts.insert(X: MI);
1830
1831 LastIdx = Idx;
1832 }
1833 // Unique factors with equal powers -- we've folded them into the first one's
1834 // base.
1835 Factors.erase(CS: llvm::unique(R&: Factors,
1836 P: [](const Factor &LHS, const Factor &RHS) {
1837 return LHS.Power == RHS.Power;
1838 }),
1839 CE: Factors.end());
1840
1841 // Iteratively collect the base of each factor with an add power into the
1842 // outer product, and halve each power in preparation for squaring the
1843 // expression.
1844 for (Factor &F : Factors) {
1845 if (F.Power & 1)
1846 OuterProduct.push_back(Elt: F.Base);
1847 F.Power >>= 1;
1848 }
1849 if (Factors[0].Power) {
1850 Value *SquareRoot = buildMinimalMultiplyDAG(Builder, Factors);
1851 OuterProduct.push_back(Elt: SquareRoot);
1852 OuterProduct.push_back(Elt: SquareRoot);
1853 }
1854 if (OuterProduct.size() == 1)
1855 return OuterProduct.front();
1856
1857 Value *V = buildMultiplyTree(Builder, Ops&: OuterProduct);
1858 return V;
1859}
1860
1861Value *ReassociatePass::OptimizeMul(BinaryOperator *I,
1862 SmallVectorImpl<ValueEntry> &Ops) {
1863 // We can only optimize the multiplies when there is a chain of more than
1864 // three, such that a balanced tree might require fewer total multiplies.
1865 if (Ops.size() < 4)
1866 return nullptr;
1867
1868 // Try to turn linear trees of multiplies without other uses of the
1869 // intermediate stages into minimal multiply DAGs with perfect sub-expression
1870 // re-use.
1871 SmallVector<Factor, 4> Factors;
1872 if (!collectMultiplyFactors(Ops, Factors))
1873 return nullptr; // All distinct factors, so nothing left for us to do.
1874
1875 IRBuilder<> Builder(I);
1876 // The reassociate transformation for FP operations is performed only
1877 // if unsafe algebra is permitted by FastMathFlags. Propagate those flags
1878 // to the newly generated operations.
1879 if (auto FPI = dyn_cast<FPMathOperator>(Val: I))
1880 Builder.setFastMathFlags(FPI->getFastMathFlags());
1881
1882 Value *V = buildMinimalMultiplyDAG(Builder, Factors);
1883 if (Ops.empty())
1884 return V;
1885
1886 ValueEntry NewEntry = ValueEntry(getRank(V), V);
1887 Ops.insert(I: llvm::lower_bound(Range&: Ops, Value&: NewEntry), Elt: NewEntry);
1888 return nullptr;
1889}
1890
1891Value *ReassociatePass::OptimizeExpression(BinaryOperator *I,
1892 SmallVectorImpl<ValueEntry> &Ops) {
1893 // Now that we have the linearized expression tree, try to optimize it.
1894 // Start by folding any constants that we found.
1895 const DataLayout &DL = I->getDataLayout();
1896 Constant *Cst = nullptr;
1897 unsigned Opcode = I->getOpcode();
1898 while (!Ops.empty()) {
1899 if (auto *C = dyn_cast<Constant>(Val: Ops.back().Op)) {
1900 if (!Cst) {
1901 Ops.pop_back();
1902 Cst = C;
1903 continue;
1904 }
1905 if (Constant *Res = ConstantFoldBinaryOpOperands(Opcode, LHS: C, RHS: Cst, DL)) {
1906 Ops.pop_back();
1907 Cst = Res;
1908 continue;
1909 }
1910 }
1911 break;
1912 }
1913 // If there was nothing but constants then we are done.
1914 if (Ops.empty())
1915 return Cst;
1916
1917 // Put the combined constant back at the end of the operand list, except if
1918 // there is no point. For example, an add of 0 gets dropped here, while a
1919 // multiplication by zero turns the whole expression into zero.
1920 if (Cst && Cst != ConstantExpr::getBinOpIdentity(Opcode, Ty: I->getType())) {
1921 if (Cst == ConstantExpr::getBinOpAbsorber(Opcode, Ty: I->getType()))
1922 return Cst;
1923 Ops.push_back(Elt: ValueEntry(0, Cst));
1924 }
1925
1926 if (Ops.size() == 1) return Ops[0].Op;
1927
1928 // Handle destructive annihilation due to identities between elements in the
1929 // argument list here.
1930 unsigned NumOps = Ops.size();
1931 switch (Opcode) {
1932 default: break;
1933 case Instruction::And:
1934 case Instruction::Or:
1935 if (Value *Result = OptimizeAndOrXor(Opcode, Ops))
1936 return Result;
1937 break;
1938
1939 case Instruction::Xor:
1940 if (Value *Result = OptimizeXor(I, Ops))
1941 return Result;
1942 break;
1943
1944 case Instruction::Add:
1945 case Instruction::FAdd:
1946 if (Value *Result = OptimizeAdd(I, Ops))
1947 return Result;
1948 break;
1949
1950 case Instruction::Mul:
1951 case Instruction::FMul:
1952 if (Value *Result = OptimizeMul(I, Ops))
1953 return Result;
1954 break;
1955 }
1956
1957 if (Ops.size() != NumOps)
1958 return OptimizeExpression(I, Ops);
1959 return nullptr;
1960}
1961
1962// Remove dead instructions and if any operands are trivially dead add them to
1963// Insts so they will be removed as well.
1964void ReassociatePass::RecursivelyEraseDeadInsts(Instruction *I,
1965 OrderedSet &Insts) {
1966 assert(isInstructionTriviallyDead(I) && "Trivially dead instructions only!");
1967 SmallVector<Value *, 4> Ops(I->operands());
1968 ValueRankMap.erase(Val: I);
1969 Insts.remove(X: I);
1970 RedoInsts.remove(X: I);
1971 llvm::salvageDebugInfo(I&: *I);
1972 I->eraseFromParent();
1973 for (auto *Op : Ops)
1974 if (Instruction *OpInst = dyn_cast<Instruction>(Val: Op))
1975 if (OpInst->use_empty())
1976 Insts.insert(X: OpInst);
1977}
1978
1979/// Zap the given instruction, adding interesting operands to the work list.
1980void ReassociatePass::EraseInst(Instruction *I) {
1981 assert(isInstructionTriviallyDead(I) && "Trivially dead instructions only!");
1982 LLVM_DEBUG(dbgs() << "Erasing dead inst: "; I->dump());
1983
1984 SmallVector<Value *, 8> Ops(I->operands());
1985 // Erase the dead instruction.
1986 ValueRankMap.erase(Val: I);
1987 RedoInsts.remove(X: I);
1988 llvm::salvageDebugInfo(I&: *I);
1989 I->eraseFromParent();
1990 // Optimize its operands.
1991 SmallPtrSet<Instruction *, 8> Visited; // Detect self-referential nodes.
1992 for (Value *V : Ops)
1993 if (Instruction *Op = dyn_cast<Instruction>(Val: V)) {
1994 // If this is a node in an expression tree, climb to the expression root
1995 // and add that since that's where optimization actually happens.
1996 unsigned Opcode = Op->getOpcode();
1997 while (Op->hasOneUse() && Op->user_back()->getOpcode() == Opcode &&
1998 Visited.insert(Ptr: Op).second)
1999 Op = Op->user_back();
2000
2001 // The instruction we're going to push may be coming from a
2002 // dead block, and Reassociate skips the processing of unreachable
2003 // blocks because it's a waste of time and also because it can
2004 // lead to infinite loop due to LLVM's non-standard definition
2005 // of dominance.
2006 if (ValueRankMap.contains(Val: Op))
2007 RedoInsts.insert(X: Op);
2008 }
2009
2010 MadeChange = true;
2011}
2012
2013/// Recursively analyze an expression to build a list of instructions that have
2014/// negative floating-point constant operands. The caller can then transform
2015/// the list to create positive constants for better reassociation and CSE.
2016static void getNegatibleInsts(Value *V,
2017 SmallVectorImpl<Instruction *> &Candidates) {
2018 // Handle only one-use instructions. Combining negations does not justify
2019 // replicating instructions.
2020 Instruction *I;
2021 if (!match(V, P: m_OneUse(SubPattern: m_Instruction(I))))
2022 return;
2023
2024 // Handle expressions of multiplications and divisions.
2025 // TODO: This could look through floating-point casts.
2026 const APFloat *C;
2027 switch (I->getOpcode()) {
2028 case Instruction::FMul:
2029 // Not expecting non-canonical code here. Bail out and wait.
2030 if (match(V: I->getOperand(i: 0), P: m_Constant()))
2031 break;
2032
2033 if (match(V: I->getOperand(i: 1), P: m_APFloat(Res&: C)) && C->isNegative()) {
2034 Candidates.push_back(Elt: I);
2035 LLVM_DEBUG(dbgs() << "FMul with negative constant: " << *I << '\n');
2036 }
2037 getNegatibleInsts(V: I->getOperand(i: 0), Candidates);
2038 getNegatibleInsts(V: I->getOperand(i: 1), Candidates);
2039 break;
2040 case Instruction::FDiv:
2041 // Not expecting non-canonical code here. Bail out and wait.
2042 if (match(V: I->getOperand(i: 0), P: m_Constant()) &&
2043 match(V: I->getOperand(i: 1), P: m_Constant()))
2044 break;
2045
2046 if ((match(V: I->getOperand(i: 0), P: m_APFloat(Res&: C)) && C->isNegative()) ||
2047 (match(V: I->getOperand(i: 1), P: m_APFloat(Res&: C)) && C->isNegative())) {
2048 Candidates.push_back(Elt: I);
2049 LLVM_DEBUG(dbgs() << "FDiv with negative constant: " << *I << '\n');
2050 }
2051 getNegatibleInsts(V: I->getOperand(i: 0), Candidates);
2052 getNegatibleInsts(V: I->getOperand(i: 1), Candidates);
2053 break;
2054 default:
2055 break;
2056 }
2057}
2058
2059/// Given an fadd/fsub with an operand that is a one-use instruction
2060/// (the fadd/fsub), try to change negative floating-point constants into
2061/// positive constants to increase potential for reassociation and CSE.
2062Instruction *ReassociatePass::canonicalizeNegFPConstantsForOp(Instruction *I,
2063 Instruction *Op,
2064 Value *OtherOp) {
2065 assert((I->getOpcode() == Instruction::FAdd ||
2066 I->getOpcode() == Instruction::FSub) && "Expected fadd/fsub");
2067
2068 // Collect instructions with negative FP constants from the subtree that ends
2069 // in Op.
2070 SmallVector<Instruction *, 4> Candidates;
2071 getNegatibleInsts(V: Op, Candidates);
2072 if (Candidates.empty())
2073 return nullptr;
2074
2075 // Don't canonicalize x + (-Constant * y) -> x - (Constant * y), if the
2076 // resulting subtract will be broken up later. This can get us into an
2077 // infinite loop during reassociation.
2078 bool IsFSub = I->getOpcode() == Instruction::FSub;
2079 bool NeedsSubtract = !IsFSub && Candidates.size() % 2 == 1;
2080 if (NeedsSubtract && ShouldBreakUpSubtract(Sub: I))
2081 return nullptr;
2082
2083 for (Instruction *Negatible : Candidates) {
2084 const APFloat *C;
2085 if (match(V: Negatible->getOperand(i: 0), P: m_APFloat(Res&: C))) {
2086 assert(!match(Negatible->getOperand(1), m_Constant()) &&
2087 "Expecting only 1 constant operand");
2088 assert(C->isNegative() && "Expected negative FP constant");
2089 Negatible->setOperand(i: 0, Val: ConstantFP::get(Ty: Negatible->getType(), V: abs(X: *C)));
2090 MadeChange = true;
2091 }
2092 if (match(V: Negatible->getOperand(i: 1), P: m_APFloat(Res&: C))) {
2093 assert(!match(Negatible->getOperand(0), m_Constant()) &&
2094 "Expecting only 1 constant operand");
2095 assert(C->isNegative() && "Expected negative FP constant");
2096 Negatible->setOperand(i: 1, Val: ConstantFP::get(Ty: Negatible->getType(), V: abs(X: *C)));
2097 MadeChange = true;
2098 }
2099 }
2100 assert(MadeChange == true && "Negative constant candidate was not changed");
2101
2102 // Negations cancelled out.
2103 if (Candidates.size() % 2 == 0)
2104 return I;
2105
2106 // Negate the final operand in the expression by flipping the opcode of this
2107 // fadd/fsub.
2108 assert(Candidates.size() % 2 == 1 && "Expected odd number");
2109 IRBuilder<> Builder(I);
2110 Value *NewInst = IsFSub ? Builder.CreateFAddFMF(L: OtherOp, R: Op, FMFSource: I)
2111 : Builder.CreateFSubFMF(L: OtherOp, R: Op, FMFSource: I);
2112 I->replaceAllUsesWith(V: NewInst);
2113 RedoInsts.insert(X: I);
2114 return dyn_cast<Instruction>(Val: NewInst);
2115}
2116
2117/// Canonicalize expressions that contain a negative floating-point constant
2118/// of the following form:
2119/// OtherOp + (subtree) -> OtherOp {+/-} (canonical subtree)
2120/// (subtree) + OtherOp -> OtherOp {+/-} (canonical subtree)
2121/// OtherOp - (subtree) -> OtherOp {+/-} (canonical subtree)
2122///
2123/// The fadd/fsub opcode may be switched to allow folding a negation into the
2124/// input instruction.
2125Instruction *ReassociatePass::canonicalizeNegFPConstants(Instruction *I) {
2126 LLVM_DEBUG(dbgs() << "Combine negations for: " << *I << '\n');
2127 Value *X;
2128 Instruction *Op;
2129 if (match(V: I, P: m_FAdd(L: m_Value(V&: X), R: m_OneUse(SubPattern: m_Instruction(I&: Op)))))
2130 if (Instruction *R = canonicalizeNegFPConstantsForOp(I, Op, OtherOp: X))
2131 I = R;
2132 if (match(V: I, P: m_FAdd(L: m_OneUse(SubPattern: m_Instruction(I&: Op)), R: m_Value(V&: X))))
2133 if (Instruction *R = canonicalizeNegFPConstantsForOp(I, Op, OtherOp: X))
2134 I = R;
2135 if (match(V: I, P: m_FSub(L: m_Value(V&: X), R: m_OneUse(SubPattern: m_Instruction(I&: Op)))))
2136 if (Instruction *R = canonicalizeNegFPConstantsForOp(I, Op, OtherOp: X))
2137 I = R;
2138 return I;
2139}
2140
2141/// Inspect and optimize the given instruction. Note that erasing
2142/// instructions is not allowed.
2143void ReassociatePass::OptimizeInst(Instruction *I) {
2144 // Only consider operations that we understand.
2145 if (!isa<UnaryOperator>(Val: I) && !isa<BinaryOperator>(Val: I))
2146 return;
2147
2148 if (I->getOpcode() == Instruction::Shl && isa<ConstantInt>(Val: I->getOperand(i: 1)))
2149 // If an operand of this shift is a reassociable multiply, or if the shift
2150 // is used by a reassociable multiply or add, turn into a multiply.
2151 if (isReassociableOp(V: I->getOperand(i: 0), Opcode: Instruction::Mul) ||
2152 (I->hasOneUse() &&
2153 (isReassociableOp(V: I->user_back(), Opcode: Instruction::Mul) ||
2154 isReassociableOp(V: I->user_back(), Opcode: Instruction::Add)))) {
2155 Instruction *NI = ConvertShiftToMul(Shl: I);
2156 RedoInsts.insert(X: I);
2157 MadeChange = true;
2158 I = NI;
2159 }
2160
2161 // Commute binary operators, to canonicalize the order of their operands.
2162 // This can potentially expose more CSE opportunities, and makes writing other
2163 // transformations simpler.
2164 if (I->isCommutative())
2165 canonicalizeOperands(I);
2166
2167 // Canonicalize negative constants out of expressions.
2168 if (Instruction *Res = canonicalizeNegFPConstants(I))
2169 I = Res;
2170
2171 // Don't optimize floating-point instructions unless they have the
2172 // appropriate FastMathFlags for reassociation enabled.
2173 if (isa<FPMathOperator>(Val: I) && !hasFPAssociativeFlags(I))
2174 return;
2175
2176 // Do not reassociate boolean (i1/vXi1) expressions. We want to preserve the
2177 // original order of evaluation for short-circuited comparisons that
2178 // SimplifyCFG has folded to AND/OR expressions. If the expression
2179 // is not further optimized, it is likely to be transformed back to a
2180 // short-circuited form for code gen, and the source order may have been
2181 // optimized for the most likely conditions. For vector boolean expressions,
2182 // we should be optimizing for ILP and not serializing the logical operations.
2183 if (I->getType()->isIntOrIntVectorTy(BitWidth: 1))
2184 return;
2185
2186 // If this is a bitwise or instruction of operands
2187 // with no common bits set, convert it to X+Y.
2188 if (I->getOpcode() == Instruction::Or &&
2189 shouldConvertOrWithNoCommonBitsToAdd(Or: I) && !isLoadCombineCandidate(Or: I) &&
2190 (cast<PossiblyDisjointInst>(Val: I)->isDisjoint() ||
2191 haveNoCommonBitsSet(LHSCache: I->getOperand(i: 0), RHSCache: I->getOperand(i: 1),
2192 SQ: SimplifyQuery(I->getDataLayout(),
2193 /*DT=*/nullptr, /*AC=*/nullptr, I)))) {
2194 Instruction *NI = convertOrWithNoCommonBitsToAdd(Or: I);
2195 RedoInsts.insert(X: I);
2196 MadeChange = true;
2197 I = NI;
2198 }
2199
2200 // If this is a subtract instruction which is not already in negate form,
2201 // see if we can convert it to X+-Y.
2202 if (I->getOpcode() == Instruction::Sub) {
2203 if (ShouldBreakUpSubtract(Sub: I)) {
2204 Instruction *NI = BreakUpSubtract(Sub: I, ToRedo&: RedoInsts);
2205 RedoInsts.insert(X: I);
2206 MadeChange = true;
2207 I = NI;
2208 } else if (match(V: I, P: m_Neg(V: m_Value()))) {
2209 // Otherwise, this is a negation. See if the operand is a multiply tree
2210 // and if this is not an inner node of a multiply tree.
2211 if (isReassociableOp(V: I->getOperand(i: 1), Opcode: Instruction::Mul) &&
2212 (!I->hasOneUse() ||
2213 !isReassociableOp(V: I->user_back(), Opcode: Instruction::Mul))) {
2214 Instruction *NI = LowerNegateToMultiply(Neg: I);
2215 // If the negate was simplified, revisit the users to see if we can
2216 // reassociate further.
2217 for (User *U : NI->users()) {
2218 if (BinaryOperator *Tmp = dyn_cast<BinaryOperator>(Val: U))
2219 RedoInsts.insert(X: Tmp);
2220 }
2221 RedoInsts.insert(X: I);
2222 MadeChange = true;
2223 I = NI;
2224 }
2225 }
2226 } else if (I->getOpcode() == Instruction::FNeg ||
2227 I->getOpcode() == Instruction::FSub) {
2228 if (ShouldBreakUpSubtract(Sub: I)) {
2229 Instruction *NI = BreakUpSubtract(Sub: I, ToRedo&: RedoInsts);
2230 RedoInsts.insert(X: I);
2231 MadeChange = true;
2232 I = NI;
2233 } else if (match(V: I, P: m_FNeg(X: m_Value()))) {
2234 // Otherwise, this is a negation. See if the operand is a multiply tree
2235 // and if this is not an inner node of a multiply tree.
2236 Value *Op = isa<BinaryOperator>(Val: I) ? I->getOperand(i: 1) :
2237 I->getOperand(i: 0);
2238 if (isReassociableOp(V: Op, Opcode: Instruction::FMul) &&
2239 (!I->hasOneUse() ||
2240 !isReassociableOp(V: I->user_back(), Opcode: Instruction::FMul))) {
2241 // If the negate was simplified, revisit the users to see if we can
2242 // reassociate further.
2243 Instruction *NI = LowerNegateToMultiply(Neg: I);
2244 for (User *U : NI->users()) {
2245 if (BinaryOperator *Tmp = dyn_cast<BinaryOperator>(Val: U))
2246 RedoInsts.insert(X: Tmp);
2247 }
2248 RedoInsts.insert(X: I);
2249 MadeChange = true;
2250 I = NI;
2251 }
2252 }
2253 }
2254
2255 // If this instruction is an associative binary operator, process it.
2256 if (!I->isAssociative()) return;
2257 BinaryOperator *BO = cast<BinaryOperator>(Val: I);
2258
2259 // If this is an interior node of a reassociable tree, ignore it until we
2260 // get to the root of the tree, to avoid N^2 analysis.
2261 unsigned Opcode = BO->getOpcode();
2262 if (BO->hasOneUse() && BO->user_back()->getOpcode() == Opcode) {
2263 // During the initial run we will get to the root of the tree.
2264 // But if we get here while we are redoing instructions, there is no
2265 // guarantee that the root will be visited. So Redo later
2266 if (BO->user_back() != BO &&
2267 BO->getParent() == BO->user_back()->getParent())
2268 RedoInsts.insert(X: BO->user_back());
2269 return;
2270 }
2271
2272 // If this is an add tree that is used by a sub instruction, ignore it
2273 // until we process the subtract.
2274 if (BO->hasOneUse() && BO->getOpcode() == Instruction::Add &&
2275 cast<Instruction>(Val: BO->user_back())->getOpcode() == Instruction::Sub)
2276 return;
2277 if (BO->hasOneUse() && BO->getOpcode() == Instruction::FAdd &&
2278 cast<Instruction>(Val: BO->user_back())->getOpcode() == Instruction::FSub)
2279 return;
2280
2281 ReassociateExpression(I: BO);
2282}
2283
2284void ReassociatePass::ReassociateExpression(BinaryOperator *I) {
2285 // First, walk the expression tree, linearizing the tree, collecting the
2286 // operand information.
2287 SmallVector<RepeatedValue, 8> Tree;
2288 OverflowTracking Flags;
2289 MadeChange |= LinearizeExprTree(I, Ops&: Tree, ToRedo&: RedoInsts, Flags);
2290 SmallVector<ValueEntry, 8> Ops;
2291 Ops.reserve(N: Tree.size());
2292 for (const RepeatedValue &E : Tree)
2293 Ops.append(NumInputs: E.second, Elt: ValueEntry(getRank(V: E.first), E.first));
2294
2295 LLVM_DEBUG(dbgs() << "RAIn:\t"; PrintOps(I, Ops); dbgs() << '\n');
2296
2297 // Now that we have linearized the tree to a list and have gathered all of
2298 // the operands and their ranks, sort the operands by their rank. Use a
2299 // stable_sort so that values with equal ranks will have their relative
2300 // positions maintained (and so the compiler is deterministic). Note that
2301 // this sorts so that the highest ranking values end up at the beginning of
2302 // the vector.
2303 llvm::stable_sort(Range&: Ops);
2304
2305 // Now that we have the expression tree in a convenient
2306 // sorted form, optimize it globally if possible.
2307 if (Value *V = OptimizeExpression(I, Ops)) {
2308 if (V == I)
2309 // Self-referential expression in unreachable code.
2310 return;
2311 // This expression tree simplified to something that isn't a tree,
2312 // eliminate it.
2313 LLVM_DEBUG(dbgs() << "Reassoc to scalar: " << *V << '\n');
2314 I->replaceAllUsesWith(V);
2315 if (Instruction *VI = dyn_cast<Instruction>(Val: V))
2316 if (I->getDebugLoc())
2317 VI->setDebugLoc(I->getDebugLoc());
2318 RedoInsts.insert(X: I);
2319 ++NumAnnihil;
2320 return;
2321 }
2322
2323 // We want to sink immediates as deeply as possible except in the case where
2324 // this is a multiply tree used only by an add, and the immediate is a -1.
2325 // In this case we reassociate to put the negation on the outside so that we
2326 // can fold the negation into the add: (-X)*Y + Z -> Z-X*Y
2327 if (I->hasOneUse()) {
2328 if (I->getOpcode() == Instruction::Mul &&
2329 cast<Instruction>(Val: I->user_back())->getOpcode() == Instruction::Add &&
2330 isa<ConstantInt>(Val: Ops.back().Op) &&
2331 cast<ConstantInt>(Val: Ops.back().Op)->isMinusOne()) {
2332 ValueEntry Tmp = Ops.pop_back_val();
2333 Ops.insert(I: Ops.begin(), Elt: Tmp);
2334 } else if (I->getOpcode() == Instruction::FMul &&
2335 cast<Instruction>(Val: I->user_back())->getOpcode() ==
2336 Instruction::FAdd &&
2337 isa<ConstantFP>(Val: Ops.back().Op) &&
2338 cast<ConstantFP>(Val: Ops.back().Op)->isExactlyValue(V: -1.0)) {
2339 ValueEntry Tmp = Ops.pop_back_val();
2340 Ops.insert(I: Ops.begin(), Elt: Tmp);
2341 }
2342 }
2343
2344 LLVM_DEBUG(dbgs() << "RAOut:\t"; PrintOps(I, Ops); dbgs() << '\n');
2345
2346 if (Ops.size() == 1) {
2347 if (Ops[0].Op == I)
2348 // Self-referential expression in unreachable code.
2349 return;
2350
2351 // This expression tree simplified to something that isn't a tree,
2352 // eliminate it.
2353 I->replaceAllUsesWith(V: Ops[0].Op);
2354 if (Instruction *OI = dyn_cast<Instruction>(Val: Ops[0].Op))
2355 OI->setDebugLoc(I->getDebugLoc());
2356 RedoInsts.insert(X: I);
2357 return;
2358 }
2359
2360 if (Ops.size() > 2 && Ops.size() <= GlobalReassociateLimit) {
2361 // Find the pair with the highest count in the pairmap and move it to the
2362 // back of the list so that it can later be CSE'd.
2363 // example:
2364 // a*b*c*d*e
2365 // if c*e is the most "popular" pair, we can express this as
2366 // (((c*e)*d)*b)*a
2367 unsigned Max = 1;
2368 unsigned BestRank = 0;
2369 std::pair<unsigned, unsigned> BestPair;
2370 unsigned Idx = I->getOpcode() - Instruction::BinaryOpsBegin;
2371 unsigned LimitIdx = 0;
2372 // With the CSE-driven heuristic, we are about to slap two values at the
2373 // beginning of the expression whereas they could live very late in the CFG.
2374 // When using the CSE-local heuristic we avoid creating dependences from
2375 // completely unrelated part of the CFG by limiting the expression
2376 // reordering on the values that live in the first seen basic block.
2377 // The main idea is that we want to avoid forming expressions that would
2378 // become loop dependent.
2379 if (UseCSELocalOpt) {
2380 const BasicBlock *FirstSeenBB = nullptr;
2381 int StartIdx = Ops.size() - 1;
2382 // Skip the first value of the expression since we need at least two
2383 // values to materialize an expression. I.e., even if this value is
2384 // anchored in a different basic block, the actual first sub expression
2385 // will be anchored on the second value.
2386 for (int i = StartIdx - 1; i != -1; --i) {
2387 const Value *Val = Ops[i].Op;
2388 const auto *CurrLeafInstr = dyn_cast<Instruction>(Val);
2389 const BasicBlock *SeenBB = nullptr;
2390 if (!CurrLeafInstr) {
2391 // The value is free of any CFG dependencies.
2392 // Do as if it lives in the entry block.
2393 //
2394 // We do this to make sure all the values falling on this path are
2395 // seen through the same anchor point. The rationale is these values
2396 // can be combined together to from a sub expression free of any CFG
2397 // dependencies so we want them to stay together.
2398 // We could be cleverer and postpone the anchor down to the first
2399 // anchored value, but that's likely complicated to get right.
2400 // E.g., we wouldn't want to do that if that means being stuck in a
2401 // loop.
2402 //
2403 // For instance, we wouldn't want to change:
2404 // res = arg1 op arg2 op arg3 op ... op loop_val1 op loop_val2 ...
2405 // into
2406 // res = loop_val1 op arg1 op arg2 op arg3 op ... op loop_val2 ...
2407 // Because all the sub expressions with arg2..N would be stuck between
2408 // two loop dependent values.
2409 SeenBB = &I->getParent()->getParent()->getEntryBlock();
2410 } else {
2411 SeenBB = CurrLeafInstr->getParent();
2412 }
2413
2414 if (!FirstSeenBB) {
2415 FirstSeenBB = SeenBB;
2416 continue;
2417 }
2418 if (FirstSeenBB != SeenBB) {
2419 // ith value is in a different basic block.
2420 // Rewind the index once to point to the last value on the same basic
2421 // block.
2422 LimitIdx = i + 1;
2423 LLVM_DEBUG(dbgs() << "CSE reordering: Consider values between ["
2424 << LimitIdx << ", " << StartIdx << "]\n");
2425 break;
2426 }
2427 }
2428 }
2429 for (unsigned i = Ops.size() - 1; i > LimitIdx; --i) {
2430 // We must use int type to go below zero when LimitIdx is 0.
2431 for (int j = i - 1; j >= (int)LimitIdx; --j) {
2432 unsigned Score = 0;
2433 Value *Op0 = Ops[i].Op;
2434 Value *Op1 = Ops[j].Op;
2435 if (std::less<Value *>()(Op1, Op0))
2436 std::swap(a&: Op0, b&: Op1);
2437 auto it = PairMap[Idx].find(Val: {Op0, Op1});
2438 if (it != PairMap[Idx].end()) {
2439 // Functions like BreakUpSubtract() can erase the Values we're using
2440 // as keys and create new Values after we built the PairMap. There's a
2441 // small chance that the new nodes can have the same address as
2442 // something already in the table. We shouldn't accumulate the stored
2443 // score in that case as it refers to the wrong Value.
2444 if (it->second.isValid())
2445 Score += it->second.Score;
2446 }
2447
2448 unsigned MaxRank = std::max(a: Ops[i].Rank, b: Ops[j].Rank);
2449
2450 // By construction, the operands are sorted in reverse order of their
2451 // topological order.
2452 // So we tend to form (sub) expressions with values that are close to
2453 // each other.
2454 //
2455 // Now to expose more CSE opportunities we want to expose the pair of
2456 // operands that occur the most (as statically computed in
2457 // BuildPairMap.) as the first sub-expression.
2458 //
2459 // If two pairs occur as many times, we pick the one with the
2460 // lowest rank, meaning the one with both operands appearing first in
2461 // the topological order.
2462 if (Score > Max || (Score == Max && MaxRank < BestRank)) {
2463 BestPair = {j, i};
2464 Max = Score;
2465 BestRank = MaxRank;
2466 }
2467 }
2468 }
2469 if (Max > 1) {
2470 auto Op0 = Ops[BestPair.first];
2471 auto Op1 = Ops[BestPair.second];
2472 Ops.erase(CI: &Ops[BestPair.second]);
2473 Ops.erase(CI: &Ops[BestPair.first]);
2474 Ops.push_back(Elt: Op0);
2475 Ops.push_back(Elt: Op1);
2476 }
2477 }
2478 LLVM_DEBUG(dbgs() << "RAOut after CSE reorder:\t"; PrintOps(I, Ops);
2479 dbgs() << '\n');
2480 // Now that we ordered and optimized the expressions, splat them back into
2481 // the expression tree, removing any unneeded nodes.
2482 RewriteExprTree(I, Ops, Flags);
2483}
2484
2485void
2486ReassociatePass::BuildPairMap(ReversePostOrderTraversal<Function *> &RPOT) {
2487 // Make a "pairmap" of how often each operand pair occurs.
2488 for (BasicBlock *BI : RPOT) {
2489 for (Instruction &I : *BI) {
2490 if (!I.isAssociative() || !I.isBinaryOp())
2491 continue;
2492
2493 // Ignore nodes that aren't at the root of trees.
2494 if (I.hasOneUse() && I.user_back()->getOpcode() == I.getOpcode())
2495 continue;
2496
2497 // Collect all operands in a single reassociable expression.
2498 // Since Reassociate has already been run once, we can assume things
2499 // are already canonical according to Reassociation's regime.
2500 SmallVector<Value *, 8> Worklist = { I.getOperand(i: 0), I.getOperand(i: 1) };
2501 SmallVector<Value *, 8> Ops;
2502 while (!Worklist.empty() && Ops.size() <= GlobalReassociateLimit) {
2503 Value *Op = Worklist.pop_back_val();
2504 Instruction *OpI = dyn_cast<Instruction>(Val: Op);
2505 if (!OpI || OpI->getOpcode() != I.getOpcode() || !OpI->hasOneUse()) {
2506 Ops.push_back(Elt: Op);
2507 continue;
2508 }
2509 // Be paranoid about self-referencing expressions in unreachable code.
2510 if (OpI->getOperand(i: 0) != OpI)
2511 Worklist.push_back(Elt: OpI->getOperand(i: 0));
2512 if (OpI->getOperand(i: 1) != OpI)
2513 Worklist.push_back(Elt: OpI->getOperand(i: 1));
2514 }
2515 // Skip extremely long expressions.
2516 if (Ops.size() > GlobalReassociateLimit)
2517 continue;
2518
2519 // Add all pairwise combinations of operands to the pair map.
2520 unsigned BinaryIdx = I.getOpcode() - Instruction::BinaryOpsBegin;
2521 SmallSet<std::pair<Value *, Value*>, 32> Visited;
2522 for (unsigned i = 0; i < Ops.size() - 1; ++i) {
2523 for (unsigned j = i + 1; j < Ops.size(); ++j) {
2524 // Canonicalize operand orderings.
2525 Value *Op0 = Ops[i];
2526 Value *Op1 = Ops[j];
2527 if (std::less<Value *>()(Op1, Op0))
2528 std::swap(a&: Op0, b&: Op1);
2529 if (!Visited.insert(V: {Op0, Op1}).second)
2530 continue;
2531 auto res = PairMap[BinaryIdx].insert(KV: {{Op0, Op1}, {.Value1: Op0, .Value2: Op1, .Score: 1}});
2532 if (!res.second) {
2533 // If either key value has been erased then we've got the same
2534 // address by coincidence. That can't happen here because nothing is
2535 // erasing values but it can happen by the time we're querying the
2536 // map.
2537 assert(res.first->second.isValid() && "WeakVH invalidated");
2538 ++res.first->second.Score;
2539 }
2540 }
2541 }
2542 }
2543 }
2544}
2545
2546PreservedAnalyses ReassociatePass::run(Function &F, FunctionAnalysisManager &) {
2547 // Get the functions basic blocks in Reverse Post Order. This order is used by
2548 // BuildRankMap to pre calculate ranks correctly. It also excludes dead basic
2549 // blocks (it has been seen that the analysis in this pass could hang when
2550 // analysing dead basic blocks).
2551 ReversePostOrderTraversal<Function *> RPOT(&F);
2552
2553 // Calculate the rank map for F.
2554 BuildRankMap(F, RPOT);
2555
2556 // Build the pair map before running reassociate.
2557 // Technically this would be more accurate if we did it after one round
2558 // of reassociation, but in practice it doesn't seem to help much on
2559 // real-world code, so don't waste the compile time running reassociate
2560 // twice.
2561 // If a user wants, they could expicitly run reassociate twice in their
2562 // pass pipeline for further potential gains.
2563 // It might also be possible to update the pair map during runtime, but the
2564 // overhead of that may be large if there's many reassociable chains.
2565 BuildPairMap(RPOT);
2566
2567 MadeChange = false;
2568
2569 // Traverse the same blocks that were analysed by BuildRankMap.
2570 for (BasicBlock *BI : RPOT) {
2571 assert(RankMap.count(&*BI) && "BB should be ranked.");
2572 // Optimize every instruction in the basic block.
2573 for (BasicBlock::iterator II = BI->begin(), IE = BI->end(); II != IE;)
2574 if (isInstructionTriviallyDead(I: &*II)) {
2575 EraseInst(I: &*II++);
2576 } else {
2577 OptimizeInst(I: &*II);
2578 assert(II->getParent() == &*BI && "Moved to a different block!");
2579 ++II;
2580 }
2581
2582 // Make a copy of all the instructions to be redone so we can remove dead
2583 // instructions.
2584 OrderedSet ToRedo(RedoInsts);
2585 // Iterate over all instructions to be reevaluated and remove trivially dead
2586 // instructions. If any operand of the trivially dead instruction becomes
2587 // dead mark it for deletion as well. Continue this process until all
2588 // trivially dead instructions have been removed.
2589 while (!ToRedo.empty()) {
2590 Instruction *I = ToRedo.pop_back_val();
2591 if (isInstructionTriviallyDead(I)) {
2592 RecursivelyEraseDeadInsts(I, Insts&: ToRedo);
2593 MadeChange = true;
2594 }
2595 }
2596
2597 // Now that we have removed dead instructions, we can reoptimize the
2598 // remaining instructions.
2599 while (!RedoInsts.empty()) {
2600 Instruction *I = RedoInsts.front();
2601 RedoInsts.erase(I: RedoInsts.begin());
2602 if (isInstructionTriviallyDead(I))
2603 EraseInst(I);
2604 else
2605 OptimizeInst(I);
2606 }
2607 }
2608
2609 // We are done with the rank map and pair map.
2610 RankMap.clear();
2611 ValueRankMap.clear();
2612 for (auto &Entry : PairMap)
2613 Entry.clear();
2614
2615 if (MadeChange) {
2616 PreservedAnalyses PA;
2617 PA.preserveSet<CFGAnalyses>();
2618 return PA;
2619 }
2620
2621 return PreservedAnalyses::all();
2622}
2623
2624namespace {
2625
2626 class ReassociateLegacyPass : public FunctionPass {
2627 ReassociatePass Impl;
2628
2629 public:
2630 static char ID; // Pass identification, replacement for typeid
2631
2632 ReassociateLegacyPass() : FunctionPass(ID) {
2633 initializeReassociateLegacyPassPass(*PassRegistry::getPassRegistry());
2634 }
2635
2636 bool runOnFunction(Function &F) override {
2637 if (skipFunction(F))
2638 return false;
2639
2640 FunctionAnalysisManager DummyFAM;
2641 auto PA = Impl.run(F, DummyFAM);
2642 return !PA.areAllPreserved();
2643 }
2644
2645 void getAnalysisUsage(AnalysisUsage &AU) const override {
2646 AU.setPreservesCFG();
2647 AU.addPreserved<AAResultsWrapperPass>();
2648 AU.addPreserved<BasicAAWrapperPass>();
2649 AU.addPreserved<GlobalsAAWrapperPass>();
2650 }
2651 };
2652
2653} // end anonymous namespace
2654
2655char ReassociateLegacyPass::ID = 0;
2656
2657INITIALIZE_PASS(ReassociateLegacyPass, "reassociate",
2658 "Reassociate expressions", false, false)
2659
2660// Public interface to the Reassociate pass
2661FunctionPass *llvm::createReassociatePass() {
2662 return new ReassociateLegacyPass();
2663}
2664