blob: 5624430332a37f972b89e316931b920cc88ddeef [file] [log] [blame]
Dan Gohman113902e2010-04-08 18:47:09 +00001//===-- Lint.cpp - Check for common errors in LLVM IR ---------------------===//
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// This pass statically checks for common and easily-identified constructs
11// which produce undefined or likely unintended behavior in LLVM IR.
12//
13// It is not a guarantee of correctness, in two ways. First, it isn't
14// comprehensive. There are checks which could be done statically which are
15// not yet implemented. Some of these are indicated by TODO comments, but
16// those aren't comprehensive either. Second, many conditions cannot be
17// checked statically. This pass does no dynamic instrumentation, so it
18// can't check for all possible problems.
19//
20// Another limitation is that it assumes all code will be executed. A store
21// through a null pointer in a basic block which is never reached is harmless,
Dan Gohmand3b6e412010-07-06 15:21:57 +000022// but this pass will warn about it anyway. This is the main reason why most
23// of these checks live here instead of in the Verifier pass.
Dan Gohman08833552010-04-22 01:30:05 +000024//
Dan Gohman113902e2010-04-08 18:47:09 +000025// Optimization passes may make conditions that this pass checks for more or
26// less obvious. If an optimization pass appears to be introducing a warning,
27// it may be that the optimization pass is merely exposing an existing
28// condition in the code.
29//
30// This code may be run before instcombine. In many cases, instcombine checks
31// for the same kinds of things and turns instructions with undefined behavior
32// into unreachable (or equivalent). Because of this, this pass makes some
33// effort to look through bitcasts and so on.
34//
35//===----------------------------------------------------------------------===//
36
37#include "llvm/Analysis/Passes.h"
38#include "llvm/Analysis/AliasAnalysis.h"
Dan Gohmanff26d4e2010-05-28 16:21:24 +000039#include "llvm/Analysis/InstructionSimplify.h"
40#include "llvm/Analysis/ConstantFolding.h"
41#include "llvm/Analysis/Dominators.h"
Dan Gohman113902e2010-04-08 18:47:09 +000042#include "llvm/Analysis/Lint.h"
Dan Gohmanff26d4e2010-05-28 16:21:24 +000043#include "llvm/Analysis/Loads.h"
Dan Gohman113902e2010-04-08 18:47:09 +000044#include "llvm/Analysis/ValueTracking.h"
45#include "llvm/Assembly/Writer.h"
46#include "llvm/Target/TargetData.h"
47#include "llvm/Pass.h"
48#include "llvm/PassManager.h"
49#include "llvm/IntrinsicInst.h"
50#include "llvm/Function.h"
51#include "llvm/Support/CallSite.h"
52#include "llvm/Support/Debug.h"
53#include "llvm/Support/InstVisitor.h"
54#include "llvm/Support/raw_ostream.h"
Dan Gohmanbe02b202010-04-09 01:39:53 +000055#include "llvm/ADT/STLExtras.h"
Dan Gohman113902e2010-04-08 18:47:09 +000056using namespace llvm;
57
58namespace {
Dan Gohman5b61b382010-04-30 19:05:00 +000059 namespace MemRef {
60 static unsigned Read = 1;
61 static unsigned Write = 2;
62 static unsigned Callee = 4;
63 static unsigned Branchee = 8;
64 }
65
Dan Gohman113902e2010-04-08 18:47:09 +000066 class Lint : public FunctionPass, public InstVisitor<Lint> {
67 friend class InstVisitor<Lint>;
68
Dan Gohmanbe02b202010-04-09 01:39:53 +000069 void visitFunction(Function &F);
70
Dan Gohman113902e2010-04-08 18:47:09 +000071 void visitCallSite(CallSite CS);
Dan Gohmanaec2a0d2010-05-28 21:43:57 +000072 void visitMemoryReference(Instruction &I, Value *Ptr,
73 unsigned Size, unsigned Align,
Dan Gohman5b61b382010-04-30 19:05:00 +000074 const Type *Ty, unsigned Flags);
Dan Gohman113902e2010-04-08 18:47:09 +000075
Dan Gohman113902e2010-04-08 18:47:09 +000076 void visitCallInst(CallInst &I);
77 void visitInvokeInst(InvokeInst &I);
78 void visitReturnInst(ReturnInst &I);
79 void visitLoadInst(LoadInst &I);
80 void visitStoreInst(StoreInst &I);
Dan Gohmanbe02b202010-04-09 01:39:53 +000081 void visitXor(BinaryOperator &I);
82 void visitSub(BinaryOperator &I);
Dan Gohmandd98c4d2010-04-08 23:05:57 +000083 void visitLShr(BinaryOperator &I);
84 void visitAShr(BinaryOperator &I);
85 void visitShl(BinaryOperator &I);
Dan Gohman113902e2010-04-08 18:47:09 +000086 void visitSDiv(BinaryOperator &I);
87 void visitUDiv(BinaryOperator &I);
88 void visitSRem(BinaryOperator &I);
89 void visitURem(BinaryOperator &I);
90 void visitAllocaInst(AllocaInst &I);
91 void visitVAArgInst(VAArgInst &I);
92 void visitIndirectBrInst(IndirectBrInst &I);
Dan Gohmandd98c4d2010-04-08 23:05:57 +000093 void visitExtractElementInst(ExtractElementInst &I);
94 void visitInsertElementInst(InsertElementInst &I);
Dan Gohmanbe02b202010-04-09 01:39:53 +000095 void visitUnreachableInst(UnreachableInst &I);
Dan Gohman113902e2010-04-08 18:47:09 +000096
Dan Gohmanff26d4e2010-05-28 16:21:24 +000097 Value *findValue(Value *V, bool OffsetOk) const;
Dan Gohman17d95962010-05-28 16:45:33 +000098 Value *findValueImpl(Value *V, bool OffsetOk,
99 SmallPtrSet<Value *, 4> &Visited) const;
Dan Gohmanff26d4e2010-05-28 16:21:24 +0000100
Dan Gohman113902e2010-04-08 18:47:09 +0000101 public:
102 Module *Mod;
103 AliasAnalysis *AA;
Dan Gohmanff26d4e2010-05-28 16:21:24 +0000104 DominatorTree *DT;
Dan Gohman113902e2010-04-08 18:47:09 +0000105 TargetData *TD;
106
107 std::string Messages;
108 raw_string_ostream MessagesStr;
109
110 static char ID; // Pass identification, replacement for typeid
111 Lint() : FunctionPass(&ID), MessagesStr(Messages) {}
112
113 virtual bool runOnFunction(Function &F);
114
115 virtual void getAnalysisUsage(AnalysisUsage &AU) const {
116 AU.setPreservesAll();
117 AU.addRequired<AliasAnalysis>();
Dan Gohmanff26d4e2010-05-28 16:21:24 +0000118 AU.addRequired<DominatorTree>();
Dan Gohman113902e2010-04-08 18:47:09 +0000119 }
120 virtual void print(raw_ostream &O, const Module *M) const {}
121
122 void WriteValue(const Value *V) {
123 if (!V) return;
124 if (isa<Instruction>(V)) {
125 MessagesStr << *V << '\n';
126 } else {
127 WriteAsOperand(MessagesStr, V, true, Mod);
128 MessagesStr << '\n';
129 }
130 }
131
132 void WriteType(const Type *T) {
133 if (!T) return;
134 MessagesStr << ' ';
135 WriteTypeSymbolic(MessagesStr, T, Mod);
136 }
137
138 // CheckFailed - A check failed, so print out the condition and the message
139 // that failed. This provides a nice place to put a breakpoint if you want
140 // to see why something is not correct.
141 void CheckFailed(const Twine &Message,
142 const Value *V1 = 0, const Value *V2 = 0,
143 const Value *V3 = 0, const Value *V4 = 0) {
144 MessagesStr << Message.str() << "\n";
145 WriteValue(V1);
146 WriteValue(V2);
147 WriteValue(V3);
148 WriteValue(V4);
149 }
150
151 void CheckFailed(const Twine &Message, const Value *V1,
152 const Type *T2, const Value *V3 = 0) {
153 MessagesStr << Message.str() << "\n";
154 WriteValue(V1);
155 WriteType(T2);
156 WriteValue(V3);
157 }
158
159 void CheckFailed(const Twine &Message, const Type *T1,
160 const Type *T2 = 0, const Type *T3 = 0) {
161 MessagesStr << Message.str() << "\n";
162 WriteType(T1);
163 WriteType(T2);
164 WriteType(T3);
165 }
166 };
167}
168
169char Lint::ID = 0;
Owen Andersond13db2c2010-07-21 22:09:45 +0000170INITIALIZE_PASS(Lint, "lint", "Statically lint-checks LLVM IR", false, true);
Dan Gohman113902e2010-04-08 18:47:09 +0000171
172// Assert - We know that cond should be true, if not print an error message.
173#define Assert(C, M) \
174 do { if (!(C)) { CheckFailed(M); return; } } while (0)
175#define Assert1(C, M, V1) \
176 do { if (!(C)) { CheckFailed(M, V1); return; } } while (0)
177#define Assert2(C, M, V1, V2) \
178 do { if (!(C)) { CheckFailed(M, V1, V2); return; } } while (0)
179#define Assert3(C, M, V1, V2, V3) \
180 do { if (!(C)) { CheckFailed(M, V1, V2, V3); return; } } while (0)
181#define Assert4(C, M, V1, V2, V3, V4) \
182 do { if (!(C)) { CheckFailed(M, V1, V2, V3, V4); return; } } while (0)
183
184// Lint::run - This is the main Analysis entry point for a
185// function.
186//
187bool Lint::runOnFunction(Function &F) {
188 Mod = F.getParent();
189 AA = &getAnalysis<AliasAnalysis>();
Dan Gohmanff26d4e2010-05-28 16:21:24 +0000190 DT = &getAnalysis<DominatorTree>();
Dan Gohman113902e2010-04-08 18:47:09 +0000191 TD = getAnalysisIfAvailable<TargetData>();
192 visit(F);
193 dbgs() << MessagesStr.str();
Dan Gohmana0f7ff32010-05-26 22:28:53 +0000194 Messages.clear();
Dan Gohman113902e2010-04-08 18:47:09 +0000195 return false;
196}
197
Dan Gohmanbe02b202010-04-09 01:39:53 +0000198void Lint::visitFunction(Function &F) {
199 // This isn't undefined behavior, it's just a little unusual, and it's a
200 // fairly common mistake to neglect to name a function.
201 Assert1(F.hasName() || F.hasLocalLinkage(),
202 "Unusual: Unnamed function with non-local linkage", &F);
Dan Gohman0ce24992010-07-06 15:23:00 +0000203
204 // TODO: Check for irreducible control flow.
Dan Gohman113902e2010-04-08 18:47:09 +0000205}
206
207void Lint::visitCallSite(CallSite CS) {
208 Instruction &I = *CS.getInstruction();
209 Value *Callee = CS.getCalledValue();
210
Dan Gohmanaec2a0d2010-05-28 21:43:57 +0000211 visitMemoryReference(I, Callee, ~0u, 0, 0, MemRef::Callee);
Dan Gohman113902e2010-04-08 18:47:09 +0000212
Dan Gohmanff26d4e2010-05-28 16:21:24 +0000213 if (Function *F = dyn_cast<Function>(findValue(Callee, /*OffsetOk=*/false))) {
Dan Gohman113902e2010-04-08 18:47:09 +0000214 Assert1(CS.getCallingConv() == F->getCallingConv(),
Dan Gohmanbe02b202010-04-09 01:39:53 +0000215 "Undefined behavior: Caller and callee calling convention differ",
216 &I);
Dan Gohman113902e2010-04-08 18:47:09 +0000217
218 const FunctionType *FT = F->getFunctionType();
219 unsigned NumActualArgs = unsigned(CS.arg_end()-CS.arg_begin());
220
221 Assert1(FT->isVarArg() ?
222 FT->getNumParams() <= NumActualArgs :
223 FT->getNumParams() == NumActualArgs,
Dan Gohmanbe02b202010-04-09 01:39:53 +0000224 "Undefined behavior: Call argument count mismatches callee "
225 "argument count", &I);
Dan Gohman113902e2010-04-08 18:47:09 +0000226
Dan Gohman545d0062010-07-12 18:02:04 +0000227 Assert1(FT->getReturnType() == I.getType(),
228 "Undefined behavior: Call return type mismatches "
229 "callee return type", &I);
230
Dan Gohmanaec2a0d2010-05-28 21:43:57 +0000231 // Check argument types (in case the callee was casted) and attributes.
Dan Gohman0ce24992010-07-06 15:23:00 +0000232 // TODO: Verify that caller and callee attributes are compatible.
Dan Gohmanaec2a0d2010-05-28 21:43:57 +0000233 Function::arg_iterator PI = F->arg_begin(), PE = F->arg_end();
234 CallSite::arg_iterator AI = CS.arg_begin(), AE = CS.arg_end();
235 for (; AI != AE; ++AI) {
236 Value *Actual = *AI;
237 if (PI != PE) {
238 Argument *Formal = PI++;
239 Assert1(Formal->getType() == Actual->getType(),
240 "Undefined behavior: Call argument type mismatches "
241 "callee parameter type", &I);
Dan Gohman10e77262010-06-01 20:51:40 +0000242
243 // Check that noalias arguments don't alias other arguments. The
244 // AliasAnalysis API isn't expressive enough for what we really want
245 // to do. Known partial overlap is not distinguished from the case
246 // where nothing is known.
Dan Gohmanaec2a0d2010-05-28 21:43:57 +0000247 if (Formal->hasNoAliasAttr() && Actual->getType()->isPointerTy())
Dan Gohman10e77262010-06-01 20:51:40 +0000248 for (CallSite::arg_iterator BI = CS.arg_begin(); BI != AE; ++BI) {
249 Assert1(AI == BI ||
250 AA->alias(*AI, ~0u, *BI, ~0u) != AliasAnalysis::MustAlias,
251 "Unusual: noalias argument aliases another argument", &I);
252 }
253
254 // Check that an sret argument points to valid memory.
Dan Gohmanaec2a0d2010-05-28 21:43:57 +0000255 if (Formal->hasStructRetAttr() && Actual->getType()->isPointerTy()) {
256 const Type *Ty =
257 cast<PointerType>(Formal->getType())->getElementType();
258 visitMemoryReference(I, Actual, AA->getTypeStoreSize(Ty),
259 TD ? TD->getABITypeAlignment(Ty) : 0,
260 Ty, MemRef::Read | MemRef::Write);
261 }
262 }
263 }
Dan Gohman113902e2010-04-08 18:47:09 +0000264 }
265
Dan Gohman113b3e22010-05-26 21:46:36 +0000266 if (CS.isCall() && cast<CallInst>(CS.getInstruction())->isTailCall())
267 for (CallSite::arg_iterator AI = CS.arg_begin(), AE = CS.arg_end();
268 AI != AE; ++AI) {
Dan Gohmanff26d4e2010-05-28 16:21:24 +0000269 Value *Obj = findValue(*AI, /*OffsetOk=*/true);
Dan Gohman078f8592010-05-28 16:34:49 +0000270 Assert1(!isa<AllocaInst>(Obj),
Dan Gohman113b3e22010-05-26 21:46:36 +0000271 "Undefined behavior: Call with \"tail\" keyword references "
Dan Gohman078f8592010-05-28 16:34:49 +0000272 "alloca", &I);
Dan Gohman113b3e22010-05-26 21:46:36 +0000273 }
274
Dan Gohman113902e2010-04-08 18:47:09 +0000275
276 if (IntrinsicInst *II = dyn_cast<IntrinsicInst>(&I))
277 switch (II->getIntrinsicID()) {
278 default: break;
279
280 // TODO: Check more intrinsics
281
282 case Intrinsic::memcpy: {
283 MemCpyInst *MCI = cast<MemCpyInst>(&I);
Dan Gohmanaec2a0d2010-05-28 21:43:57 +0000284 // TODO: If the size is known, use it.
285 visitMemoryReference(I, MCI->getDest(), ~0u, MCI->getAlignment(), 0,
Dan Gohman13ec30b2010-05-28 17:44:00 +0000286 MemRef::Write);
Dan Gohmanaec2a0d2010-05-28 21:43:57 +0000287 visitMemoryReference(I, MCI->getSource(), ~0u, MCI->getAlignment(), 0,
Dan Gohman5b61b382010-04-30 19:05:00 +0000288 MemRef::Read);
Dan Gohman113902e2010-04-08 18:47:09 +0000289
Dan Gohmanbe02b202010-04-09 01:39:53 +0000290 // Check that the memcpy arguments don't overlap. The AliasAnalysis API
291 // isn't expressive enough for what we really want to do. Known partial
292 // overlap is not distinguished from the case where nothing is known.
Dan Gohman113902e2010-04-08 18:47:09 +0000293 unsigned Size = 0;
294 if (const ConstantInt *Len =
Dan Gohmanff26d4e2010-05-28 16:21:24 +0000295 dyn_cast<ConstantInt>(findValue(MCI->getLength(),
296 /*OffsetOk=*/false)))
Dan Gohman113902e2010-04-08 18:47:09 +0000297 if (Len->getValue().isIntN(32))
298 Size = Len->getValue().getZExtValue();
299 Assert1(AA->alias(MCI->getSource(), Size, MCI->getDest(), Size) !=
300 AliasAnalysis::MustAlias,
Dan Gohmanbe02b202010-04-09 01:39:53 +0000301 "Undefined behavior: memcpy source and destination overlap", &I);
Dan Gohman113902e2010-04-08 18:47:09 +0000302 break;
303 }
304 case Intrinsic::memmove: {
305 MemMoveInst *MMI = cast<MemMoveInst>(&I);
Dan Gohmanaec2a0d2010-05-28 21:43:57 +0000306 // TODO: If the size is known, use it.
307 visitMemoryReference(I, MMI->getDest(), ~0u, MMI->getAlignment(), 0,
Dan Gohman13ec30b2010-05-28 17:44:00 +0000308 MemRef::Write);
Dan Gohmanaec2a0d2010-05-28 21:43:57 +0000309 visitMemoryReference(I, MMI->getSource(), ~0u, MMI->getAlignment(), 0,
Dan Gohman5b61b382010-04-30 19:05:00 +0000310 MemRef::Read);
Dan Gohman113902e2010-04-08 18:47:09 +0000311 break;
312 }
313 case Intrinsic::memset: {
314 MemSetInst *MSI = cast<MemSetInst>(&I);
Dan Gohmanaec2a0d2010-05-28 21:43:57 +0000315 // TODO: If the size is known, use it.
316 visitMemoryReference(I, MSI->getDest(), ~0u, MSI->getAlignment(), 0,
Dan Gohman5b61b382010-04-30 19:05:00 +0000317 MemRef::Write);
Dan Gohman113902e2010-04-08 18:47:09 +0000318 break;
319 }
320
321 case Intrinsic::vastart:
Dan Gohmanbe02b202010-04-09 01:39:53 +0000322 Assert1(I.getParent()->getParent()->isVarArg(),
323 "Undefined behavior: va_start called in a non-varargs function",
324 &I);
325
Dan Gohmanaec2a0d2010-05-28 21:43:57 +0000326 visitMemoryReference(I, CS.getArgument(0), ~0u, 0, 0,
Dan Gohman5b61b382010-04-30 19:05:00 +0000327 MemRef::Read | MemRef::Write);
Dan Gohman113902e2010-04-08 18:47:09 +0000328 break;
329 case Intrinsic::vacopy:
Dan Gohmanaec2a0d2010-05-28 21:43:57 +0000330 visitMemoryReference(I, CS.getArgument(0), ~0u, 0, 0, MemRef::Write);
331 visitMemoryReference(I, CS.getArgument(1), ~0u, 0, 0, MemRef::Read);
Dan Gohman113902e2010-04-08 18:47:09 +0000332 break;
333 case Intrinsic::vaend:
Dan Gohmanaec2a0d2010-05-28 21:43:57 +0000334 visitMemoryReference(I, CS.getArgument(0), ~0u, 0, 0,
Dan Gohman5b61b382010-04-30 19:05:00 +0000335 MemRef::Read | MemRef::Write);
Dan Gohman113902e2010-04-08 18:47:09 +0000336 break;
Dan Gohman882ddb42010-05-26 22:21:25 +0000337
338 case Intrinsic::stackrestore:
339 // Stackrestore doesn't read or write memory, but it sets the
340 // stack pointer, which the compiler may read from or write to
341 // at any time, so check it for both readability and writeability.
Dan Gohmanaec2a0d2010-05-28 21:43:57 +0000342 visitMemoryReference(I, CS.getArgument(0), ~0u, 0, 0,
Dan Gohman882ddb42010-05-26 22:21:25 +0000343 MemRef::Read | MemRef::Write);
344 break;
Dan Gohman113902e2010-04-08 18:47:09 +0000345 }
346}
347
348void Lint::visitCallInst(CallInst &I) {
349 return visitCallSite(&I);
350}
351
352void Lint::visitInvokeInst(InvokeInst &I) {
353 return visitCallSite(&I);
354}
355
356void Lint::visitReturnInst(ReturnInst &I) {
357 Function *F = I.getParent()->getParent();
358 Assert1(!F->doesNotReturn(),
Dan Gohmanbe02b202010-04-09 01:39:53 +0000359 "Unusual: Return statement in function with noreturn attribute",
360 &I);
Dan Gohman292fc872010-05-28 04:33:42 +0000361
362 if (Value *V = I.getReturnValue()) {
Dan Gohmanff26d4e2010-05-28 16:21:24 +0000363 Value *Obj = findValue(V, /*OffsetOk=*/true);
Dan Gohman078f8592010-05-28 16:34:49 +0000364 Assert1(!isa<AllocaInst>(Obj),
365 "Unusual: Returning alloca value", &I);
Dan Gohman292fc872010-05-28 04:33:42 +0000366 }
Dan Gohman113902e2010-04-08 18:47:09 +0000367}
368
Dan Gohmanaec2a0d2010-05-28 21:43:57 +0000369// TODO: Check that the reference is in bounds.
Dan Gohman0ce24992010-07-06 15:23:00 +0000370// TODO: Check readnone/readonly function attributes.
Dan Gohman113902e2010-04-08 18:47:09 +0000371void Lint::visitMemoryReference(Instruction &I,
Dan Gohmanaec2a0d2010-05-28 21:43:57 +0000372 Value *Ptr, unsigned Size, unsigned Align,
373 const Type *Ty, unsigned Flags) {
374 // If no memory is being referenced, it doesn't matter if the pointer
375 // is valid.
376 if (Size == 0)
377 return;
378
Dan Gohmanff26d4e2010-05-28 16:21:24 +0000379 Value *UnderlyingObject = findValue(Ptr, /*OffsetOk=*/true);
Dan Gohmanbe02b202010-04-09 01:39:53 +0000380 Assert1(!isa<ConstantPointerNull>(UnderlyingObject),
381 "Undefined behavior: Null pointer dereference", &I);
382 Assert1(!isa<UndefValue>(UnderlyingObject),
383 "Undefined behavior: Undef pointer dereference", &I);
Dan Gohmanaec2a0d2010-05-28 21:43:57 +0000384 Assert1(!isa<ConstantInt>(UnderlyingObject) ||
385 !cast<ConstantInt>(UnderlyingObject)->isAllOnesValue(),
386 "Unusual: All-ones pointer dereference", &I);
387 Assert1(!isa<ConstantInt>(UnderlyingObject) ||
388 !cast<ConstantInt>(UnderlyingObject)->isOne(),
389 "Unusual: Address one pointer dereference", &I);
Dan Gohman113902e2010-04-08 18:47:09 +0000390
Dan Gohman5b61b382010-04-30 19:05:00 +0000391 if (Flags & MemRef::Write) {
392 if (const GlobalVariable *GV = dyn_cast<GlobalVariable>(UnderlyingObject))
393 Assert1(!GV->isConstant(),
394 "Undefined behavior: Write to read-only memory", &I);
395 Assert1(!isa<Function>(UnderlyingObject) &&
396 !isa<BlockAddress>(UnderlyingObject),
397 "Undefined behavior: Write to text section", &I);
398 }
399 if (Flags & MemRef::Read) {
400 Assert1(!isa<Function>(UnderlyingObject),
401 "Unusual: Load from function body", &I);
402 Assert1(!isa<BlockAddress>(UnderlyingObject),
403 "Undefined behavior: Load from block address", &I);
404 }
405 if (Flags & MemRef::Callee) {
406 Assert1(!isa<BlockAddress>(UnderlyingObject),
407 "Undefined behavior: Call to block address", &I);
408 }
409 if (Flags & MemRef::Branchee) {
410 Assert1(!isa<Constant>(UnderlyingObject) ||
411 isa<BlockAddress>(UnderlyingObject),
412 "Undefined behavior: Branch to non-blockaddress", &I);
413 }
414
Dan Gohman113902e2010-04-08 18:47:09 +0000415 if (TD) {
416 if (Align == 0 && Ty) Align = TD->getABITypeAlignment(Ty);
417
418 if (Align != 0) {
419 unsigned BitWidth = TD->getTypeSizeInBits(Ptr->getType());
420 APInt Mask = APInt::getAllOnesValue(BitWidth),
421 KnownZero(BitWidth, 0), KnownOne(BitWidth, 0);
422 ComputeMaskedBits(Ptr, Mask, KnownZero, KnownOne, TD);
423 Assert1(!(KnownOne & APInt::getLowBitsSet(BitWidth, Log2_32(Align))),
Dan Gohmanbe02b202010-04-09 01:39:53 +0000424 "Undefined behavior: Memory reference address is misaligned", &I);
Dan Gohman113902e2010-04-08 18:47:09 +0000425 }
426 }
427}
428
429void Lint::visitLoadInst(LoadInst &I) {
Dan Gohmanaec2a0d2010-05-28 21:43:57 +0000430 visitMemoryReference(I, I.getPointerOperand(),
431 AA->getTypeStoreSize(I.getType()), I.getAlignment(),
432 I.getType(), MemRef::Read);
Dan Gohman113902e2010-04-08 18:47:09 +0000433}
434
435void Lint::visitStoreInst(StoreInst &I) {
Dan Gohmanaec2a0d2010-05-28 21:43:57 +0000436 visitMemoryReference(I, I.getPointerOperand(),
437 AA->getTypeStoreSize(I.getOperand(0)->getType()),
438 I.getAlignment(),
439 I.getOperand(0)->getType(), MemRef::Write);
Dan Gohman113902e2010-04-08 18:47:09 +0000440}
441
Dan Gohmanbe02b202010-04-09 01:39:53 +0000442void Lint::visitXor(BinaryOperator &I) {
443 Assert1(!isa<UndefValue>(I.getOperand(0)) ||
444 !isa<UndefValue>(I.getOperand(1)),
445 "Undefined result: xor(undef, undef)", &I);
446}
447
448void Lint::visitSub(BinaryOperator &I) {
449 Assert1(!isa<UndefValue>(I.getOperand(0)) ||
450 !isa<UndefValue>(I.getOperand(1)),
451 "Undefined result: sub(undef, undef)", &I);
452}
453
Dan Gohmandd98c4d2010-04-08 23:05:57 +0000454void Lint::visitLShr(BinaryOperator &I) {
455 if (ConstantInt *CI =
Dan Gohmanff26d4e2010-05-28 16:21:24 +0000456 dyn_cast<ConstantInt>(findValue(I.getOperand(1), /*OffsetOk=*/false)))
Dan Gohmandd98c4d2010-04-08 23:05:57 +0000457 Assert1(CI->getValue().ult(cast<IntegerType>(I.getType())->getBitWidth()),
Dan Gohmanbe02b202010-04-09 01:39:53 +0000458 "Undefined result: Shift count out of range", &I);
Dan Gohmandd98c4d2010-04-08 23:05:57 +0000459}
460
461void Lint::visitAShr(BinaryOperator &I) {
462 if (ConstantInt *CI =
Dan Gohmanff26d4e2010-05-28 16:21:24 +0000463 dyn_cast<ConstantInt>(findValue(I.getOperand(1), /*OffsetOk=*/false)))
Dan Gohmandd98c4d2010-04-08 23:05:57 +0000464 Assert1(CI->getValue().ult(cast<IntegerType>(I.getType())->getBitWidth()),
Dan Gohmanbe02b202010-04-09 01:39:53 +0000465 "Undefined result: Shift count out of range", &I);
Dan Gohmandd98c4d2010-04-08 23:05:57 +0000466}
467
468void Lint::visitShl(BinaryOperator &I) {
469 if (ConstantInt *CI =
Dan Gohmanff26d4e2010-05-28 16:21:24 +0000470 dyn_cast<ConstantInt>(findValue(I.getOperand(1), /*OffsetOk=*/false)))
Dan Gohmandd98c4d2010-04-08 23:05:57 +0000471 Assert1(CI->getValue().ult(cast<IntegerType>(I.getType())->getBitWidth()),
Dan Gohmanbe02b202010-04-09 01:39:53 +0000472 "Undefined result: Shift count out of range", &I);
Dan Gohmandd98c4d2010-04-08 23:05:57 +0000473}
474
Dan Gohman113902e2010-04-08 18:47:09 +0000475static bool isZero(Value *V, TargetData *TD) {
Dan Gohmanbe02b202010-04-09 01:39:53 +0000476 // Assume undef could be zero.
477 if (isa<UndefValue>(V)) return true;
478
Dan Gohman113902e2010-04-08 18:47:09 +0000479 unsigned BitWidth = cast<IntegerType>(V->getType())->getBitWidth();
480 APInt Mask = APInt::getAllOnesValue(BitWidth),
481 KnownZero(BitWidth, 0), KnownOne(BitWidth, 0);
482 ComputeMaskedBits(V, Mask, KnownZero, KnownOne, TD);
483 return KnownZero.isAllOnesValue();
484}
485
486void Lint::visitSDiv(BinaryOperator &I) {
Dan Gohmanbe02b202010-04-09 01:39:53 +0000487 Assert1(!isZero(I.getOperand(1), TD),
488 "Undefined behavior: Division by zero", &I);
Dan Gohman113902e2010-04-08 18:47:09 +0000489}
490
491void Lint::visitUDiv(BinaryOperator &I) {
Dan Gohmanbe02b202010-04-09 01:39:53 +0000492 Assert1(!isZero(I.getOperand(1), TD),
493 "Undefined behavior: Division by zero", &I);
Dan Gohman113902e2010-04-08 18:47:09 +0000494}
495
496void Lint::visitSRem(BinaryOperator &I) {
Dan Gohmanbe02b202010-04-09 01:39:53 +0000497 Assert1(!isZero(I.getOperand(1), TD),
498 "Undefined behavior: Division by zero", &I);
Dan Gohman113902e2010-04-08 18:47:09 +0000499}
500
501void Lint::visitURem(BinaryOperator &I) {
Dan Gohmanbe02b202010-04-09 01:39:53 +0000502 Assert1(!isZero(I.getOperand(1), TD),
503 "Undefined behavior: Division by zero", &I);
Dan Gohman113902e2010-04-08 18:47:09 +0000504}
505
506void Lint::visitAllocaInst(AllocaInst &I) {
507 if (isa<ConstantInt>(I.getArraySize()))
508 // This isn't undefined behavior, it's just an obvious pessimization.
509 Assert1(&I.getParent()->getParent()->getEntryBlock() == I.getParent(),
Dan Gohmanbe02b202010-04-09 01:39:53 +0000510 "Pessimization: Static alloca outside of entry block", &I);
Dan Gohman0ce24992010-07-06 15:23:00 +0000511
512 // TODO: Check for an unusual size (MSB set?)
Dan Gohman113902e2010-04-08 18:47:09 +0000513}
514
515void Lint::visitVAArgInst(VAArgInst &I) {
Dan Gohmanaec2a0d2010-05-28 21:43:57 +0000516 visitMemoryReference(I, I.getOperand(0), ~0u, 0, 0,
Dan Gohman5b61b382010-04-30 19:05:00 +0000517 MemRef::Read | MemRef::Write);
Dan Gohman113902e2010-04-08 18:47:09 +0000518}
519
520void Lint::visitIndirectBrInst(IndirectBrInst &I) {
Dan Gohmanaec2a0d2010-05-28 21:43:57 +0000521 visitMemoryReference(I, I.getAddress(), ~0u, 0, 0, MemRef::Branchee);
Dan Gohmana8afb2a2010-08-02 23:06:43 +0000522
523 Assert1(I.getNumDestinations() != 0,
524 "Undefined behavior: indirectbr with no destinations", &I);
Dan Gohman113902e2010-04-08 18:47:09 +0000525}
526
Dan Gohmandd98c4d2010-04-08 23:05:57 +0000527void Lint::visitExtractElementInst(ExtractElementInst &I) {
528 if (ConstantInt *CI =
Dan Gohmanff26d4e2010-05-28 16:21:24 +0000529 dyn_cast<ConstantInt>(findValue(I.getIndexOperand(),
530 /*OffsetOk=*/false)))
Dan Gohmandd98c4d2010-04-08 23:05:57 +0000531 Assert1(CI->getValue().ult(I.getVectorOperandType()->getNumElements()),
Dan Gohmanbe02b202010-04-09 01:39:53 +0000532 "Undefined result: extractelement index out of range", &I);
Dan Gohmandd98c4d2010-04-08 23:05:57 +0000533}
534
535void Lint::visitInsertElementInst(InsertElementInst &I) {
536 if (ConstantInt *CI =
Dan Gohmanff26d4e2010-05-28 16:21:24 +0000537 dyn_cast<ConstantInt>(findValue(I.getOperand(2),
538 /*OffsetOk=*/false)))
Dan Gohmandd98c4d2010-04-08 23:05:57 +0000539 Assert1(CI->getValue().ult(I.getType()->getNumElements()),
Dan Gohmanbe02b202010-04-09 01:39:53 +0000540 "Undefined result: insertelement index out of range", &I);
541}
542
543void Lint::visitUnreachableInst(UnreachableInst &I) {
544 // This isn't undefined behavior, it's merely suspicious.
545 Assert1(&I == I.getParent()->begin() ||
546 prior(BasicBlock::iterator(&I))->mayHaveSideEffects(),
547 "Unusual: unreachable immediately preceded by instruction without "
548 "side effects", &I);
Dan Gohmandd98c4d2010-04-08 23:05:57 +0000549}
550
Dan Gohmanff26d4e2010-05-28 16:21:24 +0000551/// findValue - Look through bitcasts and simple memory reference patterns
552/// to identify an equivalent, but more informative, value. If OffsetOk
553/// is true, look through getelementptrs with non-zero offsets too.
554///
555/// Most analysis passes don't require this logic, because instcombine
556/// will simplify most of these kinds of things away. But it's a goal of
557/// this Lint pass to be useful even on non-optimized IR.
558Value *Lint::findValue(Value *V, bool OffsetOk) const {
Dan Gohman17d95962010-05-28 16:45:33 +0000559 SmallPtrSet<Value *, 4> Visited;
560 return findValueImpl(V, OffsetOk, Visited);
561}
562
563/// findValueImpl - Implementation helper for findValue.
564Value *Lint::findValueImpl(Value *V, bool OffsetOk,
565 SmallPtrSet<Value *, 4> &Visited) const {
566 // Detect self-referential values.
567 if (!Visited.insert(V))
568 return UndefValue::get(V->getType());
569
Dan Gohmanff26d4e2010-05-28 16:21:24 +0000570 // TODO: Look through sext or zext cast, when the result is known to
571 // be interpreted as signed or unsigned, respectively.
Dan Gohmanaec2a0d2010-05-28 21:43:57 +0000572 // TODO: Look through eliminable cast pairs.
Dan Gohmanff26d4e2010-05-28 16:21:24 +0000573 // TODO: Look through calls with unique return values.
574 // TODO: Look through vector insert/extract/shuffle.
575 V = OffsetOk ? V->getUnderlyingObject() : V->stripPointerCasts();
576 if (LoadInst *L = dyn_cast<LoadInst>(V)) {
577 BasicBlock::iterator BBI = L;
578 BasicBlock *BB = L->getParent();
Dan Gohman13ec30b2010-05-28 17:44:00 +0000579 SmallPtrSet<BasicBlock *, 4> VisitedBlocks;
Dan Gohmanff26d4e2010-05-28 16:21:24 +0000580 for (;;) {
Dan Gohman13ec30b2010-05-28 17:44:00 +0000581 if (!VisitedBlocks.insert(BB)) break;
Dan Gohmanff26d4e2010-05-28 16:21:24 +0000582 if (Value *U = FindAvailableLoadedValue(L->getPointerOperand(),
583 BB, BBI, 6, AA))
Dan Gohman17d95962010-05-28 16:45:33 +0000584 return findValueImpl(U, OffsetOk, Visited);
Dan Gohman13ec30b2010-05-28 17:44:00 +0000585 if (BBI != BB->begin()) break;
586 BB = BB->getUniquePredecessor();
Dan Gohmanff26d4e2010-05-28 16:21:24 +0000587 if (!BB) break;
588 BBI = BB->end();
589 }
Dan Gohmanaec2a0d2010-05-28 21:43:57 +0000590 } else if (PHINode *PN = dyn_cast<PHINode>(V)) {
591 if (Value *W = PN->hasConstantValue(DT))
592 return findValueImpl(W, OffsetOk, Visited);
Dan Gohmanff26d4e2010-05-28 16:21:24 +0000593 } else if (CastInst *CI = dyn_cast<CastInst>(V)) {
594 if (CI->isNoopCast(TD ? TD->getIntPtrType(V->getContext()) :
595 Type::getInt64Ty(V->getContext())))
Dan Gohman17d95962010-05-28 16:45:33 +0000596 return findValueImpl(CI->getOperand(0), OffsetOk, Visited);
Dan Gohmanff26d4e2010-05-28 16:21:24 +0000597 } else if (ExtractValueInst *Ex = dyn_cast<ExtractValueInst>(V)) {
598 if (Value *W = FindInsertedValue(Ex->getAggregateOperand(),
599 Ex->idx_begin(),
600 Ex->idx_end()))
601 if (W != V)
Dan Gohman17d95962010-05-28 16:45:33 +0000602 return findValueImpl(W, OffsetOk, Visited);
Dan Gohmanaec2a0d2010-05-28 21:43:57 +0000603 } else if (ConstantExpr *CE = dyn_cast<ConstantExpr>(V)) {
604 // Same as above, but for ConstantExpr instead of Instruction.
605 if (Instruction::isCast(CE->getOpcode())) {
606 if (CastInst::isNoopCast(Instruction::CastOps(CE->getOpcode()),
607 CE->getOperand(0)->getType(),
608 CE->getType(),
609 TD ? TD->getIntPtrType(V->getContext()) :
610 Type::getInt64Ty(V->getContext())))
611 return findValueImpl(CE->getOperand(0), OffsetOk, Visited);
612 } else if (CE->getOpcode() == Instruction::ExtractValue) {
613 const SmallVector<unsigned, 4> &Indices = CE->getIndices();
614 if (Value *W = FindInsertedValue(CE->getOperand(0),
615 Indices.begin(),
616 Indices.end()))
617 if (W != V)
618 return findValueImpl(W, OffsetOk, Visited);
619 }
Dan Gohmanff26d4e2010-05-28 16:21:24 +0000620 }
621
622 // As a last resort, try SimplifyInstruction or constant folding.
623 if (Instruction *Inst = dyn_cast<Instruction>(V)) {
624 if (Value *W = SimplifyInstruction(Inst, TD))
625 if (W != Inst)
Dan Gohman17d95962010-05-28 16:45:33 +0000626 return findValueImpl(W, OffsetOk, Visited);
Dan Gohmanff26d4e2010-05-28 16:21:24 +0000627 } else if (ConstantExpr *CE = dyn_cast<ConstantExpr>(V)) {
628 if (Value *W = ConstantFoldConstantExpression(CE, TD))
629 if (W != V)
Dan Gohman17d95962010-05-28 16:45:33 +0000630 return findValueImpl(W, OffsetOk, Visited);
Dan Gohmanff26d4e2010-05-28 16:21:24 +0000631 }
632
633 return V;
634}
635
Dan Gohman113902e2010-04-08 18:47:09 +0000636//===----------------------------------------------------------------------===//
637// Implement the public interfaces to this file...
638//===----------------------------------------------------------------------===//
639
640FunctionPass *llvm::createLintPass() {
641 return new Lint();
642}
643
644/// lintFunction - Check a function for errors, printing messages on stderr.
645///
646void llvm::lintFunction(const Function &f) {
647 Function &F = const_cast<Function&>(f);
648 assert(!F.isDeclaration() && "Cannot lint external functions");
649
650 FunctionPassManager FPM(F.getParent());
651 Lint *V = new Lint();
652 FPM.add(V);
653 FPM.run(F);
654}
655
656/// lintModule - Check a module for errors, printing messages on stderr.
Dan Gohman113902e2010-04-08 18:47:09 +0000657///
Dan Gohmana0f7ff32010-05-26 22:28:53 +0000658void llvm::lintModule(const Module &M) {
Dan Gohman113902e2010-04-08 18:47:09 +0000659 PassManager PM;
660 Lint *V = new Lint();
661 PM.add(V);
662 PM.run(const_cast<Module&>(M));
Dan Gohman113902e2010-04-08 18:47:09 +0000663}