Lint.cpp revision 749be11f4daf859c53638112b8f0620a503ed0cc 
1//===-- 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,
22// but this pass will warn about it anyway.
23//
24// Optimization passes may make conditions that this pass checks for more or
25// less obvious. If an optimization pass appears to be introducing a warning,
26// it may be that the optimization pass is merely exposing an existing
27// condition in the code.
28//
29// This code may be run before instcombine. In many cases, instcombine checks
30// for the same kinds of things and turns instructions with undefined behavior
31// into unreachable (or equivalent). Because of this, this pass makes some
32// effort to look through bitcasts and so on.
33//
34//===----------------------------------------------------------------------===//
35
36#include "llvm/Analysis/Passes.h"
37#include "llvm/Analysis/AliasAnalysis.h"
38#include "llvm/Analysis/Lint.h"
39#include "llvm/Analysis/ValueTracking.h"
40#include "llvm/Assembly/Writer.h"
41#include "llvm/Target/TargetData.h"
42#include "llvm/Pass.h"
43#include "llvm/PassManager.h"
44#include "llvm/IntrinsicInst.h"
45#include "llvm/Function.h"
46#include "llvm/Support/CallSite.h"
47#include "llvm/Support/Debug.h"
48#include "llvm/Support/InstVisitor.h"
49#include "llvm/Support/raw_ostream.h"
50#include "llvm/ADT/STLExtras.h"
51using namespace llvm;
52
53namespace {
54  namespace MemRef {
55    static unsigned Read     = 1;
56    static unsigned Write    = 2;
57    static unsigned Callee   = 4;
58    static unsigned Branchee = 8;
59  }
60
61  class Lint : public FunctionPass, public InstVisitor<Lint> {
62    friend class InstVisitor<Lint>;
63
64    void visitFunction(Function &F);
65
66    void visitCallSite(CallSite CS);
67    void visitMemoryReference(Instruction &I, Value *Ptr, unsigned Align,
68                              const Type *Ty, unsigned Flags);
69
70    void visitCallInst(CallInst &I);
71    void visitInvokeInst(InvokeInst &I);
72    void visitReturnInst(ReturnInst &I);
73    void visitLoadInst(LoadInst &I);
74    void visitStoreInst(StoreInst &I);
75    void visitXor(BinaryOperator &I);
76    void visitSub(BinaryOperator &I);
77    void visitLShr(BinaryOperator &I);
78    void visitAShr(BinaryOperator &I);
79    void visitShl(BinaryOperator &I);
80    void visitSDiv(BinaryOperator &I);
81    void visitUDiv(BinaryOperator &I);
82    void visitSRem(BinaryOperator &I);
83    void visitURem(BinaryOperator &I);
84    void visitAllocaInst(AllocaInst &I);
85    void visitVAArgInst(VAArgInst &I);
86    void visitIndirectBrInst(IndirectBrInst &I);
87    void visitExtractElementInst(ExtractElementInst &I);
88    void visitInsertElementInst(InsertElementInst &I);
89    void visitUnreachableInst(UnreachableInst &I);
90
91  public:
92    Module *Mod;
93    AliasAnalysis *AA;
94    TargetData *TD;
95
96    std::string Messages;
97    raw_string_ostream MessagesStr;
98
99    static char ID; // Pass identification, replacement for typeid
100    Lint() : FunctionPass(&ID), MessagesStr(Messages) {}
101
102    virtual bool runOnFunction(Function &F);
103
104    virtual void getAnalysisUsage(AnalysisUsage &AU) const {
105      AU.setPreservesAll();
106      AU.addRequired<AliasAnalysis>();
107    }
108    virtual void print(raw_ostream &O, const Module *M) const {}
109
110    void WriteValue(const Value *V) {
111      if (!V) return;
112      if (isa<Instruction>(V)) {
113        MessagesStr << *V << '\n';
114      } else {
115        WriteAsOperand(MessagesStr, V, true, Mod);
116        MessagesStr << '\n';
117      }
118    }
119
120    void WriteType(const Type *T) {
121      if (!T) return;
122      MessagesStr << ' ';
123      WriteTypeSymbolic(MessagesStr, T, Mod);
124    }
125
126    // CheckFailed - A check failed, so print out the condition and the message
127    // that failed.  This provides a nice place to put a breakpoint if you want
128    // to see why something is not correct.
129    void CheckFailed(const Twine &Message,
130                     const Value *V1 = 0, const Value *V2 = 0,
131                     const Value *V3 = 0, const Value *V4 = 0) {
132      MessagesStr << Message.str() << "\n";
133      WriteValue(V1);
134      WriteValue(V2);
135      WriteValue(V3);
136      WriteValue(V4);
137    }
138
139    void CheckFailed(const Twine &Message, const Value *V1,
140                     const Type *T2, const Value *V3 = 0) {
141      MessagesStr << Message.str() << "\n";
142      WriteValue(V1);
143      WriteType(T2);
144      WriteValue(V3);
145    }
146
147    void CheckFailed(const Twine &Message, const Type *T1,
148                     const Type *T2 = 0, const Type *T3 = 0) {
149      MessagesStr << Message.str() << "\n";
150      WriteType(T1);
151      WriteType(T2);
152      WriteType(T3);
153    }
154  };
155}
156
157char Lint::ID = 0;
158static RegisterPass<Lint>
159X("lint", "Statically lint-checks LLVM IR", false, true);
160
161// Assert - We know that cond should be true, if not print an error message.
162#define Assert(C, M) \
163    do { if (!(C)) { CheckFailed(M); return; } } while (0)
164#define Assert1(C, M, V1) \
165    do { if (!(C)) { CheckFailed(M, V1); return; } } while (0)
166#define Assert2(C, M, V1, V2) \
167    do { if (!(C)) { CheckFailed(M, V1, V2); return; } } while (0)
168#define Assert3(C, M, V1, V2, V3) \
169    do { if (!(C)) { CheckFailed(M, V1, V2, V3); return; } } while (0)
170#define Assert4(C, M, V1, V2, V3, V4) \
171    do { if (!(C)) { CheckFailed(M, V1, V2, V3, V4); return; } } while (0)
172
173// Lint::run - This is the main Analysis entry point for a
174// function.
175//
176bool Lint::runOnFunction(Function &F) {
177  Mod = F.getParent();
178  AA = &getAnalysis<AliasAnalysis>();
179  TD = getAnalysisIfAvailable<TargetData>();
180  visit(F);
181  dbgs() << MessagesStr.str();
182  return false;
183}
184
185void Lint::visitFunction(Function &F) {
186  // This isn't undefined behavior, it's just a little unusual, and it's a
187  // fairly common mistake to neglect to name a function.
188  Assert1(F.hasName() || F.hasLocalLinkage(),
189          "Unusual: Unnamed function with non-local linkage", &F);
190}
191
192void Lint::visitCallSite(CallSite CS) {
193  Instruction &I = *CS.getInstruction();
194  Value *Callee = CS.getCalledValue();
195
196  visitMemoryReference(I, Callee, 0, 0, MemRef::Callee);
197
198  if (Function *F = dyn_cast<Function>(Callee->stripPointerCasts())) {
199    Assert1(CS.getCallingConv() == F->getCallingConv(),
200            "Undefined behavior: Caller and callee calling convention differ",
201            &I);
202
203    const FunctionType *FT = F->getFunctionType();
204    unsigned NumActualArgs = unsigned(CS.arg_end()-CS.arg_begin());
205
206    Assert1(FT->isVarArg() ?
207              FT->getNumParams() <= NumActualArgs :
208              FT->getNumParams() == NumActualArgs,
209            "Undefined behavior: Call argument count mismatches callee "
210            "argument count", &I);
211
212    // TODO: Check argument types (in case the callee was casted)
213
214    // TODO: Check ABI-significant attributes.
215
216    // TODO: Check noalias attribute.
217
218    // TODO: Check sret attribute.
219  }
220
221  if (CS.isCall() && cast<CallInst>(CS.getInstruction())->isTailCall())
222    for (CallSite::arg_iterator AI = CS.arg_begin(), AE = CS.arg_end();
223         AI != AE; ++AI) {
224      Value *Obj = (*AI)->getUnderlyingObject();
225      Assert1(!isa<AllocaInst>(Obj) && !isa<VAArgInst>(Obj),
226              "Undefined behavior: Call with \"tail\" keyword references "
227              "alloca or va_arg", &I);
228    }
229
230
231  if (IntrinsicInst *II = dyn_cast<IntrinsicInst>(&I))
232    switch (II->getIntrinsicID()) {
233    default: break;
234
235    // TODO: Check more intrinsics
236
237    case Intrinsic::memcpy: {
238      MemCpyInst *MCI = cast<MemCpyInst>(&I);
239      visitMemoryReference(I, MCI->getSource(), MCI->getAlignment(), 0,
240                           MemRef::Write);
241      visitMemoryReference(I, MCI->getDest(), MCI->getAlignment(), 0,
242                           MemRef::Read);
243
244      // Check that the memcpy arguments don't overlap. The AliasAnalysis API
245      // isn't expressive enough for what we really want to do. Known partial
246      // overlap is not distinguished from the case where nothing is known.
247      unsigned Size = 0;
248      if (const ConstantInt *Len =
249            dyn_cast<ConstantInt>(MCI->getLength()->stripPointerCasts()))
250        if (Len->getValue().isIntN(32))
251          Size = Len->getValue().getZExtValue();
252      Assert1(AA->alias(MCI->getSource(), Size, MCI->getDest(), Size) !=
253              AliasAnalysis::MustAlias,
254              "Undefined behavior: memcpy source and destination overlap", &I);
255      break;
256    }
257    case Intrinsic::memmove: {
258      MemMoveInst *MMI = cast<MemMoveInst>(&I);
259      visitMemoryReference(I, MMI->getSource(), MMI->getAlignment(), 0,
260                           MemRef::Write);
261      visitMemoryReference(I, MMI->getDest(), MMI->getAlignment(), 0,
262                           MemRef::Read);
263      break;
264    }
265    case Intrinsic::memset: {
266      MemSetInst *MSI = cast<MemSetInst>(&I);
267      visitMemoryReference(I, MSI->getDest(), MSI->getAlignment(), 0,
268                           MemRef::Write);
269      break;
270    }
271
272    case Intrinsic::vastart:
273      Assert1(I.getParent()->getParent()->isVarArg(),
274              "Undefined behavior: va_start called in a non-varargs function",
275              &I);
276
277      visitMemoryReference(I, CS.getArgument(0), 0, 0,
278                           MemRef::Read | MemRef::Write);
279      break;
280    case Intrinsic::vacopy:
281      visitMemoryReference(I, CS.getArgument(0), 0, 0, MemRef::Write);
282      visitMemoryReference(I, CS.getArgument(1), 0, 0, MemRef::Read);
283      break;
284    case Intrinsic::vaend:
285      visitMemoryReference(I, CS.getArgument(0), 0, 0,
286                           MemRef::Read | MemRef::Write);
287      break;
288    }
289}
290
291void Lint::visitCallInst(CallInst &I) {
292  return visitCallSite(&I);
293}
294
295void Lint::visitInvokeInst(InvokeInst &I) {
296  return visitCallSite(&I);
297}
298
299void Lint::visitReturnInst(ReturnInst &I) {
300  Function *F = I.getParent()->getParent();
301  Assert1(!F->doesNotReturn(),
302          "Unusual: Return statement in function with noreturn attribute",
303          &I);
304}
305
306// TODO: Add a length argument and check that the reference is in bounds
307void Lint::visitMemoryReference(Instruction &I,
308                                Value *Ptr, unsigned Align, const Type *Ty,
309                                unsigned Flags) {
310  Value *UnderlyingObject = Ptr->getUnderlyingObject();
311  Assert1(!isa<ConstantPointerNull>(UnderlyingObject),
312          "Undefined behavior: Null pointer dereference", &I);
313  Assert1(!isa<UndefValue>(UnderlyingObject),
314          "Undefined behavior: Undef pointer dereference", &I);
315
316  if (Flags & MemRef::Write) {
317    if (const GlobalVariable *GV = dyn_cast<GlobalVariable>(UnderlyingObject))
318      Assert1(!GV->isConstant(),
319              "Undefined behavior: Write to read-only memory", &I);
320    Assert1(!isa<Function>(UnderlyingObject) &&
321            !isa<BlockAddress>(UnderlyingObject),
322            "Undefined behavior: Write to text section", &I);
323  }
324  if (Flags & MemRef::Read) {
325    Assert1(!isa<Function>(UnderlyingObject),
326            "Unusual: Load from function body", &I);
327    Assert1(!isa<BlockAddress>(UnderlyingObject),
328            "Undefined behavior: Load from block address", &I);
329  }
330  if (Flags & MemRef::Callee) {
331    Assert1(!isa<BlockAddress>(UnderlyingObject),
332            "Undefined behavior: Call to block address", &I);
333  }
334  if (Flags & MemRef::Branchee) {
335    Assert1(!isa<Constant>(UnderlyingObject) ||
336            isa<BlockAddress>(UnderlyingObject),
337            "Undefined behavior: Branch to non-blockaddress", &I);
338  }
339
340  if (TD) {
341    if (Align == 0 && Ty) Align = TD->getABITypeAlignment(Ty);
342
343    if (Align != 0) {
344      unsigned BitWidth = TD->getTypeSizeInBits(Ptr->getType());
345      APInt Mask = APInt::getAllOnesValue(BitWidth),
346                   KnownZero(BitWidth, 0), KnownOne(BitWidth, 0);
347      ComputeMaskedBits(Ptr, Mask, KnownZero, KnownOne, TD);
348      Assert1(!(KnownOne & APInt::getLowBitsSet(BitWidth, Log2_32(Align))),
349              "Undefined behavior: Memory reference address is misaligned", &I);
350    }
351  }
352}
353
354void Lint::visitLoadInst(LoadInst &I) {
355  visitMemoryReference(I, I.getPointerOperand(), I.getAlignment(), I.getType(),
356                       MemRef::Read);
357}
358
359void Lint::visitStoreInst(StoreInst &I) {
360  visitMemoryReference(I, I.getPointerOperand(), I.getAlignment(),
361                  I.getOperand(0)->getType(), MemRef::Write);
362}
363
364void Lint::visitXor(BinaryOperator &I) {
365  Assert1(!isa<UndefValue>(I.getOperand(0)) ||
366          !isa<UndefValue>(I.getOperand(1)),
367          "Undefined result: xor(undef, undef)", &I);
368}
369
370void Lint::visitSub(BinaryOperator &I) {
371  Assert1(!isa<UndefValue>(I.getOperand(0)) ||
372          !isa<UndefValue>(I.getOperand(1)),
373          "Undefined result: sub(undef, undef)", &I);
374}
375
376void Lint::visitLShr(BinaryOperator &I) {
377  if (ConstantInt *CI =
378        dyn_cast<ConstantInt>(I.getOperand(1)->stripPointerCasts()))
379    Assert1(CI->getValue().ult(cast<IntegerType>(I.getType())->getBitWidth()),
380            "Undefined result: Shift count out of range", &I);
381}
382
383void Lint::visitAShr(BinaryOperator &I) {
384  if (ConstantInt *CI =
385        dyn_cast<ConstantInt>(I.getOperand(1)->stripPointerCasts()))
386    Assert1(CI->getValue().ult(cast<IntegerType>(I.getType())->getBitWidth()),
387            "Undefined result: Shift count out of range", &I);
388}
389
390void Lint::visitShl(BinaryOperator &I) {
391  if (ConstantInt *CI =
392        dyn_cast<ConstantInt>(I.getOperand(1)->stripPointerCasts()))
393    Assert1(CI->getValue().ult(cast<IntegerType>(I.getType())->getBitWidth()),
394            "Undefined result: Shift count out of range", &I);
395}
396
397static bool isZero(Value *V, TargetData *TD) {
398  // Assume undef could be zero.
399  if (isa<UndefValue>(V)) return true;
400
401  unsigned BitWidth = cast<IntegerType>(V->getType())->getBitWidth();
402  APInt Mask = APInt::getAllOnesValue(BitWidth),
403               KnownZero(BitWidth, 0), KnownOne(BitWidth, 0);
404  ComputeMaskedBits(V, Mask, KnownZero, KnownOne, TD);
405  return KnownZero.isAllOnesValue();
406}
407
408void Lint::visitSDiv(BinaryOperator &I) {
409  Assert1(!isZero(I.getOperand(1), TD),
410          "Undefined behavior: Division by zero", &I);
411}
412
413void Lint::visitUDiv(BinaryOperator &I) {
414  Assert1(!isZero(I.getOperand(1), TD),
415          "Undefined behavior: Division by zero", &I);
416}
417
418void Lint::visitSRem(BinaryOperator &I) {
419  Assert1(!isZero(I.getOperand(1), TD),
420          "Undefined behavior: Division by zero", &I);
421}
422
423void Lint::visitURem(BinaryOperator &I) {
424  Assert1(!isZero(I.getOperand(1), TD),
425          "Undefined behavior: Division by zero", &I);
426}
427
428void Lint::visitAllocaInst(AllocaInst &I) {
429  if (isa<ConstantInt>(I.getArraySize()))
430    // This isn't undefined behavior, it's just an obvious pessimization.
431    Assert1(&I.getParent()->getParent()->getEntryBlock() == I.getParent(),
432            "Pessimization: Static alloca outside of entry block", &I);
433}
434
435void Lint::visitVAArgInst(VAArgInst &I) {
436  visitMemoryReference(I, I.getOperand(0), 0, 0,
437                       MemRef::Read | MemRef::Write);
438}
439
440void Lint::visitIndirectBrInst(IndirectBrInst &I) {
441  visitMemoryReference(I, I.getAddress(), 0, 0, MemRef::Branchee);
442}
443
444void Lint::visitExtractElementInst(ExtractElementInst &I) {
445  if (ConstantInt *CI =
446        dyn_cast<ConstantInt>(I.getIndexOperand()->stripPointerCasts()))
447    Assert1(CI->getValue().ult(I.getVectorOperandType()->getNumElements()),
448            "Undefined result: extractelement index out of range", &I);
449}
450
451void Lint::visitInsertElementInst(InsertElementInst &I) {
452  if (ConstantInt *CI =
453        dyn_cast<ConstantInt>(I.getOperand(2)->stripPointerCasts()))
454    Assert1(CI->getValue().ult(I.getType()->getNumElements()),
455            "Undefined result: insertelement index out of range", &I);
456}
457
458void Lint::visitUnreachableInst(UnreachableInst &I) {
459  // This isn't undefined behavior, it's merely suspicious.
460  Assert1(&I == I.getParent()->begin() ||
461          prior(BasicBlock::iterator(&I))->mayHaveSideEffects(),
462          "Unusual: unreachable immediately preceded by instruction without "
463          "side effects", &I);
464}
465
466//===----------------------------------------------------------------------===//
467//  Implement the public interfaces to this file...
468//===----------------------------------------------------------------------===//
469
470FunctionPass *llvm::createLintPass() {
471  return new Lint();
472}
473
474/// lintFunction - Check a function for errors, printing messages on stderr.
475///
476void llvm::lintFunction(const Function &f) {
477  Function &F = const_cast<Function&>(f);
478  assert(!F.isDeclaration() && "Cannot lint external functions");
479
480  FunctionPassManager FPM(F.getParent());
481  Lint *V = new Lint();
482  FPM.add(V);
483  FPM.run(F);
484}
485
486/// lintModule - Check a module for errors, printing messages on stderr.
487/// Return true if the module is corrupt.
488///
489void llvm::lintModule(const Module &M, std::string *ErrorInfo) {
490  PassManager PM;
491  Lint *V = new Lint();
492  PM.add(V);
493  PM.run(const_cast<Module&>(M));
494
495  if (ErrorInfo)
496    *ErrorInfo = V->MessagesStr.str();
497}
498