1 //===-- Operator.cpp - Implement the LLVM operators -----------------------===//
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 file implements the non-inline methods for the LLVM Operator classes.
10 //
11 //===----------------------------------------------------------------------===//
12
13 #include "llvm/IR/Operator.h"
14 #include "llvm/IR/DataLayout.h"
15 #include "llvm/IR/GetElementPtrTypeIterator.h"
16 #include "llvm/IR/Instructions.h"
17
18 #include "ConstantsContext.h"
19
20 namespace llvm {
hasPoisonGeneratingFlags() const21 bool Operator::hasPoisonGeneratingFlags() const {
22 switch (getOpcode()) {
23 case Instruction::Add:
24 case Instruction::Sub:
25 case Instruction::Mul:
26 case Instruction::Shl: {
27 auto *OBO = cast<OverflowingBinaryOperator>(this);
28 return OBO->hasNoUnsignedWrap() || OBO->hasNoSignedWrap();
29 }
30 case Instruction::Trunc: {
31 if (auto *TI = dyn_cast<TruncInst>(this))
32 return TI->hasNoUnsignedWrap() || TI->hasNoSignedWrap();
33 return false;
34 }
35 case Instruction::UDiv:
36 case Instruction::SDiv:
37 case Instruction::AShr:
38 case Instruction::LShr:
39 return cast<PossiblyExactOperator>(this)->isExact();
40 case Instruction::Or:
41 return cast<PossiblyDisjointInst>(this)->isDisjoint();
42 case Instruction::GetElementPtr: {
43 auto *GEP = cast<GEPOperator>(this);
44 // Note: inrange exists on constexpr only
45 return GEP->getNoWrapFlags() != GEPNoWrapFlags::none() ||
46 GEP->getInRange() != std::nullopt;
47 }
48 case Instruction::UIToFP:
49 case Instruction::ZExt:
50 if (auto *NNI = dyn_cast<PossiblyNonNegInst>(this))
51 return NNI->hasNonNeg();
52 return false;
53 default:
54 if (const auto *FP = dyn_cast<FPMathOperator>(this))
55 return FP->hasNoNaNs() || FP->hasNoInfs();
56 return false;
57 }
58 }
59
hasPoisonGeneratingAnnotations() const60 bool Operator::hasPoisonGeneratingAnnotations() const {
61 if (hasPoisonGeneratingFlags())
62 return true;
63 auto *I = dyn_cast<Instruction>(this);
64 return I && (I->hasPoisonGeneratingReturnAttributes() ||
65 I->hasPoisonGeneratingMetadata());
66 }
67
getSourceElementType() const68 Type *GEPOperator::getSourceElementType() const {
69 if (auto *I = dyn_cast<GetElementPtrInst>(this))
70 return I->getSourceElementType();
71 return cast<GetElementPtrConstantExpr>(this)->getSourceElementType();
72 }
73
getResultElementType() const74 Type *GEPOperator::getResultElementType() const {
75 if (auto *I = dyn_cast<GetElementPtrInst>(this))
76 return I->getResultElementType();
77 return cast<GetElementPtrConstantExpr>(this)->getResultElementType();
78 }
79
getInRange() const80 std::optional<ConstantRange> GEPOperator::getInRange() const {
81 if (auto *CE = dyn_cast<GetElementPtrConstantExpr>(this))
82 return CE->getInRange();
83 return std::nullopt;
84 }
85
getMaxPreservedAlignment(const DataLayout & DL) const86 Align GEPOperator::getMaxPreservedAlignment(const DataLayout &DL) const {
87 /// compute the worse possible offset for every level of the GEP et accumulate
88 /// the minimum alignment into Result.
89
90 Align Result = Align(llvm::Value::MaximumAlignment);
91 for (gep_type_iterator GTI = gep_type_begin(this), GTE = gep_type_end(this);
92 GTI != GTE; ++GTI) {
93 uint64_t Offset;
94 ConstantInt *OpC = dyn_cast<ConstantInt>(GTI.getOperand());
95
96 if (StructType *STy = GTI.getStructTypeOrNull()) {
97 const StructLayout *SL = DL.getStructLayout(STy);
98 Offset = SL->getElementOffset(OpC->getZExtValue());
99 } else {
100 assert(GTI.isSequential() && "should be sequencial");
101 /// If the index isn't known, we take 1 because it is the index that will
102 /// give the worse alignment of the offset.
103 const uint64_t ElemCount = OpC ? OpC->getZExtValue() : 1;
104 Offset = GTI.getSequentialElementStride(DL) * ElemCount;
105 }
106 Result = Align(MinAlign(Offset, Result.value()));
107 }
108 return Result;
109 }
110
accumulateConstantOffset(const DataLayout & DL,APInt & Offset,function_ref<bool (Value &,APInt &)> ExternalAnalysis) const111 bool GEPOperator::accumulateConstantOffset(
112 const DataLayout &DL, APInt &Offset,
113 function_ref<bool(Value &, APInt &)> ExternalAnalysis) const {
114 assert(Offset.getBitWidth() ==
115 DL.getIndexSizeInBits(getPointerAddressSpace()) &&
116 "The offset bit width does not match DL specification.");
117 SmallVector<const Value *> Index(llvm::drop_begin(operand_values()));
118 return GEPOperator::accumulateConstantOffset(getSourceElementType(), Index,
119 DL, Offset, ExternalAnalysis);
120 }
121
accumulateConstantOffset(Type * SourceType,ArrayRef<const Value * > Index,const DataLayout & DL,APInt & Offset,function_ref<bool (Value &,APInt &)> ExternalAnalysis)122 bool GEPOperator::accumulateConstantOffset(
123 Type *SourceType, ArrayRef<const Value *> Index, const DataLayout &DL,
124 APInt &Offset, function_ref<bool(Value &, APInt &)> ExternalAnalysis) {
125 // Fast path for canonical getelementptr i8 form.
126 if (SourceType->isIntegerTy(8) && !ExternalAnalysis) {
127 if (auto *CI = dyn_cast<ConstantInt>(Index.front())) {
128 Offset += CI->getValue().sextOrTrunc(Offset.getBitWidth());
129 return true;
130 }
131 return false;
132 }
133
134 bool UsedExternalAnalysis = false;
135 auto AccumulateOffset = [&](APInt Index, uint64_t Size) -> bool {
136 Index = Index.sextOrTrunc(Offset.getBitWidth());
137 APInt IndexedSize = APInt(Offset.getBitWidth(), Size);
138 // For array or vector indices, scale the index by the size of the type.
139 if (!UsedExternalAnalysis) {
140 Offset += Index * IndexedSize;
141 } else {
142 // External Analysis can return a result higher/lower than the value
143 // represents. We need to detect overflow/underflow.
144 bool Overflow = false;
145 APInt OffsetPlus = Index.smul_ov(IndexedSize, Overflow);
146 if (Overflow)
147 return false;
148 Offset = Offset.sadd_ov(OffsetPlus, Overflow);
149 if (Overflow)
150 return false;
151 }
152 return true;
153 };
154 auto begin = generic_gep_type_iterator<decltype(Index.begin())>::begin(
155 SourceType, Index.begin());
156 auto end = generic_gep_type_iterator<decltype(Index.end())>::end(Index.end());
157 for (auto GTI = begin, GTE = end; GTI != GTE; ++GTI) {
158 // Scalable vectors are multiplied by a runtime constant.
159 bool ScalableType = GTI.getIndexedType()->isScalableTy();
160
161 Value *V = GTI.getOperand();
162 StructType *STy = GTI.getStructTypeOrNull();
163 // Handle ConstantInt if possible.
164 if (auto ConstOffset = dyn_cast<ConstantInt>(V)) {
165 if (ConstOffset->isZero())
166 continue;
167 // if the type is scalable and the constant is not zero (vscale * n * 0 =
168 // 0) bailout.
169 if (ScalableType)
170 return false;
171 // Handle a struct index, which adds its field offset to the pointer.
172 if (STy) {
173 unsigned ElementIdx = ConstOffset->getZExtValue();
174 const StructLayout *SL = DL.getStructLayout(STy);
175 // Element offset is in bytes.
176 if (!AccumulateOffset(
177 APInt(Offset.getBitWidth(), SL->getElementOffset(ElementIdx)),
178 1))
179 return false;
180 continue;
181 }
182 if (!AccumulateOffset(ConstOffset->getValue(),
183 GTI.getSequentialElementStride(DL)))
184 return false;
185 continue;
186 }
187
188 // The operand is not constant, check if an external analysis was provided.
189 // External analsis is not applicable to a struct type.
190 if (!ExternalAnalysis || STy || ScalableType)
191 return false;
192 APInt AnalysisIndex;
193 if (!ExternalAnalysis(*V, AnalysisIndex))
194 return false;
195 UsedExternalAnalysis = true;
196 if (!AccumulateOffset(AnalysisIndex, GTI.getSequentialElementStride(DL)))
197 return false;
198 }
199 return true;
200 }
201
collectOffset(const DataLayout & DL,unsigned BitWidth,MapVector<Value *,APInt> & VariableOffsets,APInt & ConstantOffset) const202 bool GEPOperator::collectOffset(
203 const DataLayout &DL, unsigned BitWidth,
204 MapVector<Value *, APInt> &VariableOffsets,
205 APInt &ConstantOffset) const {
206 assert(BitWidth == DL.getIndexSizeInBits(getPointerAddressSpace()) &&
207 "The offset bit width does not match DL specification.");
208
209 auto CollectConstantOffset = [&](APInt Index, uint64_t Size) {
210 Index = Index.sextOrTrunc(BitWidth);
211 APInt IndexedSize = APInt(BitWidth, Size);
212 ConstantOffset += Index * IndexedSize;
213 };
214
215 for (gep_type_iterator GTI = gep_type_begin(this), GTE = gep_type_end(this);
216 GTI != GTE; ++GTI) {
217 // Scalable vectors are multiplied by a runtime constant.
218 bool ScalableType = GTI.getIndexedType()->isScalableTy();
219
220 Value *V = GTI.getOperand();
221 StructType *STy = GTI.getStructTypeOrNull();
222 // Handle ConstantInt if possible.
223 if (auto ConstOffset = dyn_cast<ConstantInt>(V)) {
224 if (ConstOffset->isZero())
225 continue;
226 // If the type is scalable and the constant is not zero (vscale * n * 0 =
227 // 0) bailout.
228 // TODO: If the runtime value is accessible at any point before DWARF
229 // emission, then we could potentially keep a forward reference to it
230 // in the debug value to be filled in later.
231 if (ScalableType)
232 return false;
233 // Handle a struct index, which adds its field offset to the pointer.
234 if (STy) {
235 unsigned ElementIdx = ConstOffset->getZExtValue();
236 const StructLayout *SL = DL.getStructLayout(STy);
237 // Element offset is in bytes.
238 CollectConstantOffset(APInt(BitWidth, SL->getElementOffset(ElementIdx)),
239 1);
240 continue;
241 }
242 CollectConstantOffset(ConstOffset->getValue(),
243 GTI.getSequentialElementStride(DL));
244 continue;
245 }
246
247 if (STy || ScalableType)
248 return false;
249 APInt IndexedSize = APInt(BitWidth, GTI.getSequentialElementStride(DL));
250 // Insert an initial offset of 0 for V iff none exists already, then
251 // increment the offset by IndexedSize.
252 if (!IndexedSize.isZero()) {
253 auto *It = VariableOffsets.insert({V, APInt(BitWidth, 0)}).first;
254 It->second += IndexedSize;
255 }
256 }
257 return true;
258 }
259
print(raw_ostream & O) const260 void FastMathFlags::print(raw_ostream &O) const {
261 if (all())
262 O << " fast";
263 else {
264 if (allowReassoc())
265 O << " reassoc";
266 if (noNaNs())
267 O << " nnan";
268 if (noInfs())
269 O << " ninf";
270 if (noSignedZeros())
271 O << " nsz";
272 if (allowReciprocal())
273 O << " arcp";
274 if (allowContract())
275 O << " contract";
276 if (approxFunc())
277 O << " afn";
278 }
279 }
280 } // namespace llvm
281