this is failing on linux hosts, force a triple.
[llvm/avr.git] / lib / Analysis / LoopVR.cpp
blob573bd3ea01620ca6df2425e9a26a956eea88ffce
1 //===- LoopVR.cpp - Value Range analysis driven by loop information -------===//
2 //
3 // The LLVM Compiler Infrastructure
4 //
5 // This file is distributed under the University of Illinois Open Source
6 // License. See LICENSE.TXT for details.
7 //
8 //===----------------------------------------------------------------------===//
9 //
10 // FIXME: What does this do?
12 //===----------------------------------------------------------------------===//
14 #define DEBUG_TYPE "loopvr"
15 #include "llvm/Analysis/LoopVR.h"
16 #include "llvm/Constants.h"
17 #include "llvm/Instructions.h"
18 #include "llvm/LLVMContext.h"
19 #include "llvm/Analysis/LoopInfo.h"
20 #include "llvm/Analysis/ScalarEvolutionExpressions.h"
21 #include "llvm/Assembly/Writer.h"
22 #include "llvm/Support/CFG.h"
23 #include "llvm/Support/Debug.h"
24 #include "llvm/Support/raw_ostream.h"
25 using namespace llvm;
27 char LoopVR::ID = 0;
28 static RegisterPass<LoopVR> X("loopvr", "Loop Value Ranges", false, true);
30 /// getRange - determine the range for a particular SCEV within a given Loop
31 ConstantRange LoopVR::getRange(const SCEV *S, Loop *L, ScalarEvolution &SE) {
32 const SCEV *T = SE.getBackedgeTakenCount(L);
33 if (isa<SCEVCouldNotCompute>(T))
34 return ConstantRange(cast<IntegerType>(S->getType())->getBitWidth(), true);
36 T = SE.getTruncateOrZeroExtend(T, S->getType());
37 return getRange(S, T, SE);
40 /// getRange - determine the range for a particular SCEV with a given trip count
41 ConstantRange LoopVR::getRange(const SCEV *S, const SCEV *T, ScalarEvolution &SE){
43 if (const SCEVConstant *C = dyn_cast<SCEVConstant>(S))
44 return ConstantRange(C->getValue()->getValue());
46 ConstantRange FullSet(cast<IntegerType>(S->getType())->getBitWidth(), true);
48 // {x,+,y,+,...z}. We detect overflow by checking the size of the set after
49 // summing the upper and lower.
50 if (const SCEVAddExpr *Add = dyn_cast<SCEVAddExpr>(S)) {
51 ConstantRange X = getRange(Add->getOperand(0), T, SE);
52 if (X.isFullSet()) return FullSet;
53 for (unsigned i = 1, e = Add->getNumOperands(); i != e; ++i) {
54 ConstantRange Y = getRange(Add->getOperand(i), T, SE);
55 if (Y.isFullSet()) return FullSet;
57 APInt Spread_X = X.getSetSize(), Spread_Y = Y.getSetSize();
58 APInt NewLower = X.getLower() + Y.getLower();
59 APInt NewUpper = X.getUpper() + Y.getUpper() - 1;
60 if (NewLower == NewUpper)
61 return FullSet;
63 X = ConstantRange(NewLower, NewUpper);
64 if (X.getSetSize().ult(Spread_X) || X.getSetSize().ult(Spread_Y))
65 return FullSet; // we've wrapped, therefore, full set.
67 return X;
70 // {x,*,y,*,...,z}. In order to detect overflow, we use k*bitwidth where
71 // k is the number of terms being multiplied.
72 if (const SCEVMulExpr *Mul = dyn_cast<SCEVMulExpr>(S)) {
73 ConstantRange X = getRange(Mul->getOperand(0), T, SE);
74 if (X.isFullSet()) return FullSet;
76 const IntegerType *Ty = IntegerType::get(SE.getContext(), X.getBitWidth());
77 const IntegerType *ExTy = IntegerType::get(SE.getContext(),
78 X.getBitWidth() * Mul->getNumOperands());
79 ConstantRange XExt = X.zeroExtend(ExTy->getBitWidth());
81 for (unsigned i = 1, e = Mul->getNumOperands(); i != e; ++i) {
82 ConstantRange Y = getRange(Mul->getOperand(i), T, SE);
83 if (Y.isFullSet()) return FullSet;
85 ConstantRange YExt = Y.zeroExtend(ExTy->getBitWidth());
86 XExt = ConstantRange(XExt.getLower() * YExt.getLower(),
87 ((XExt.getUpper()-1) * (YExt.getUpper()-1)) + 1);
89 return XExt.truncate(Ty->getBitWidth());
92 // X smax Y smax ... Z is: range(smax(X_smin, Y_smin, ..., Z_smin),
93 // smax(X_smax, Y_smax, ..., Z_smax))
94 // It doesn't matter if one of the SCEVs has FullSet because we're taking
95 // a maximum of the minimums across all of them.
96 if (const SCEVSMaxExpr *SMax = dyn_cast<SCEVSMaxExpr>(S)) {
97 ConstantRange X = getRange(SMax->getOperand(0), T, SE);
98 if (X.isFullSet()) return FullSet;
100 APInt smin = X.getSignedMin(), smax = X.getSignedMax();
101 for (unsigned i = 1, e = SMax->getNumOperands(); i != e; ++i) {
102 ConstantRange Y = getRange(SMax->getOperand(i), T, SE);
103 smin = APIntOps::smax(smin, Y.getSignedMin());
104 smax = APIntOps::smax(smax, Y.getSignedMax());
106 if (smax + 1 == smin) return FullSet;
107 return ConstantRange(smin, smax + 1);
110 // X umax Y umax ... Z is: range(umax(X_umin, Y_umin, ..., Z_umin),
111 // umax(X_umax, Y_umax, ..., Z_umax))
112 // It doesn't matter if one of the SCEVs has FullSet because we're taking
113 // a maximum of the minimums across all of them.
114 if (const SCEVUMaxExpr *UMax = dyn_cast<SCEVUMaxExpr>(S)) {
115 ConstantRange X = getRange(UMax->getOperand(0), T, SE);
116 if (X.isFullSet()) return FullSet;
118 APInt umin = X.getUnsignedMin(), umax = X.getUnsignedMax();
119 for (unsigned i = 1, e = UMax->getNumOperands(); i != e; ++i) {
120 ConstantRange Y = getRange(UMax->getOperand(i), T, SE);
121 umin = APIntOps::umax(umin, Y.getUnsignedMin());
122 umax = APIntOps::umax(umax, Y.getUnsignedMax());
124 if (umax + 1 == umin) return FullSet;
125 return ConstantRange(umin, umax + 1);
128 // L udiv R. Luckily, there's only ever 2 sides to a udiv.
129 if (const SCEVUDivExpr *UDiv = dyn_cast<SCEVUDivExpr>(S)) {
130 ConstantRange L = getRange(UDiv->getLHS(), T, SE);
131 ConstantRange R = getRange(UDiv->getRHS(), T, SE);
132 if (L.isFullSet() && R.isFullSet()) return FullSet;
134 if (R.getUnsignedMax() == 0) {
135 // RHS must be single-element zero. Return an empty set.
136 return ConstantRange(R.getBitWidth(), false);
139 APInt Lower = L.getUnsignedMin().udiv(R.getUnsignedMax());
141 APInt Upper;
143 if (R.getUnsignedMin() == 0) {
144 // Just because it contains zero, doesn't mean it will also contain one.
145 ConstantRange NotZero(APInt(L.getBitWidth(), 1),
146 APInt::getNullValue(L.getBitWidth()));
147 R = R.intersectWith(NotZero);
150 // But, the intersection might still include zero. If it does, then we know
151 // it also included one.
152 if (R.contains(APInt::getNullValue(L.getBitWidth())))
153 Upper = L.getUnsignedMax();
154 else
155 Upper = L.getUnsignedMax().udiv(R.getUnsignedMin());
157 return ConstantRange(Lower, Upper);
160 // ConstantRange already implements the cast operators.
162 if (const SCEVZeroExtendExpr *ZExt = dyn_cast<SCEVZeroExtendExpr>(S)) {
163 T = SE.getTruncateOrZeroExtend(T, ZExt->getOperand()->getType());
164 ConstantRange X = getRange(ZExt->getOperand(), T, SE);
165 return X.zeroExtend(cast<IntegerType>(ZExt->getType())->getBitWidth());
168 if (const SCEVSignExtendExpr *SExt = dyn_cast<SCEVSignExtendExpr>(S)) {
169 T = SE.getTruncateOrZeroExtend(T, SExt->getOperand()->getType());
170 ConstantRange X = getRange(SExt->getOperand(), T, SE);
171 return X.signExtend(cast<IntegerType>(SExt->getType())->getBitWidth());
174 if (const SCEVTruncateExpr *Trunc = dyn_cast<SCEVTruncateExpr>(S)) {
175 T = SE.getTruncateOrZeroExtend(T, Trunc->getOperand()->getType());
176 ConstantRange X = getRange(Trunc->getOperand(), T, SE);
177 if (X.isFullSet()) return FullSet;
178 return X.truncate(cast<IntegerType>(Trunc->getType())->getBitWidth());
181 if (const SCEVAddRecExpr *AddRec = dyn_cast<SCEVAddRecExpr>(S)) {
182 const SCEVConstant *Trip = dyn_cast<SCEVConstant>(T);
183 if (!Trip) return FullSet;
185 if (AddRec->isAffine()) {
186 const SCEV *StartHandle = AddRec->getStart();
187 const SCEV *StepHandle = AddRec->getOperand(1);
189 const SCEVConstant *Step = dyn_cast<SCEVConstant>(StepHandle);
190 if (!Step) return FullSet;
192 uint32_t ExWidth = 2 * Trip->getValue()->getBitWidth();
193 APInt TripExt = Trip->getValue()->getValue(); TripExt.zext(ExWidth);
194 APInt StepExt = Step->getValue()->getValue(); StepExt.zext(ExWidth);
195 if ((TripExt * StepExt).ugt(APInt::getLowBitsSet(ExWidth, ExWidth >> 1)))
196 return FullSet;
198 const SCEV *EndHandle = SE.getAddExpr(StartHandle,
199 SE.getMulExpr(T, StepHandle));
200 const SCEVConstant *Start = dyn_cast<SCEVConstant>(StartHandle);
201 const SCEVConstant *End = dyn_cast<SCEVConstant>(EndHandle);
202 if (!Start || !End) return FullSet;
204 const APInt &StartInt = Start->getValue()->getValue();
205 const APInt &EndInt = End->getValue()->getValue();
206 const APInt &StepInt = Step->getValue()->getValue();
208 if (StepInt.isNegative()) {
209 if (EndInt == StartInt + 1) return FullSet;
210 return ConstantRange(EndInt, StartInt + 1);
211 } else {
212 if (StartInt == EndInt + 1) return FullSet;
213 return ConstantRange(StartInt, EndInt + 1);
218 // TODO: non-affine addrec, udiv, SCEVUnknown (narrowed from elsewhere)?
220 return FullSet;
223 void LoopVR::getAnalysisUsage(AnalysisUsage &AU) const {
224 AU.addRequiredTransitive<LoopInfo>();
225 AU.addRequiredTransitive<ScalarEvolution>();
226 AU.setPreservesAll();
229 bool LoopVR::runOnFunction(Function &F) { Map.clear(); return false; }
231 void LoopVR::print(raw_ostream &OS, const Module *) const {
232 for (std::map<Value *, ConstantRange *>::const_iterator I = Map.begin(),
233 E = Map.end(); I != E; ++I) {
234 OS << *I->first << ": " << *I->second << '\n';
238 void LoopVR::releaseMemory() {
239 for (std::map<Value *, ConstantRange *>::iterator I = Map.begin(),
240 E = Map.end(); I != E; ++I) {
241 delete I->second;
244 Map.clear();
247 ConstantRange LoopVR::compute(Value *V) {
248 if (ConstantInt *CI = dyn_cast<ConstantInt>(V))
249 return ConstantRange(CI->getValue());
251 Instruction *I = dyn_cast<Instruction>(V);
252 if (!I)
253 return ConstantRange(cast<IntegerType>(V->getType())->getBitWidth(), false);
255 LoopInfo &LI = getAnalysis<LoopInfo>();
257 Loop *L = LI.getLoopFor(I->getParent());
258 if (!L || L->isLoopInvariant(I))
259 return ConstantRange(cast<IntegerType>(V->getType())->getBitWidth(), false);
261 ScalarEvolution &SE = getAnalysis<ScalarEvolution>();
263 const SCEV *S = SE.getSCEV(I);
264 if (isa<SCEVUnknown>(S) || isa<SCEVCouldNotCompute>(S))
265 return ConstantRange(cast<IntegerType>(V->getType())->getBitWidth(), false);
267 return ConstantRange(getRange(S, L, SE));
270 ConstantRange LoopVR::get(Value *V) {
271 std::map<Value *, ConstantRange *>::iterator I = Map.find(V);
272 if (I == Map.end()) {
273 ConstantRange *CR = new ConstantRange(compute(V));
274 Map[V] = CR;
275 return *CR;
278 return *I->second;
281 void LoopVR::remove(Value *V) {
282 std::map<Value *, ConstantRange *>::iterator I = Map.find(V);
283 if (I != Map.end()) {
284 delete I->second;
285 Map.erase(I);
289 void LoopVR::narrow(Value *V, const ConstantRange &CR) {
290 if (CR.isFullSet()) return;
292 std::map<Value *, ConstantRange *>::iterator I = Map.find(V);
293 if (I == Map.end())
294 Map[V] = new ConstantRange(CR);
295 else
296 Map[V] = new ConstantRange(Map[V]->intersectWith(CR));