1 //===-- ThreadSanitizer.cpp - race detector -------------------------------===//
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 file is a part of ThreadSanitizer, a race detector.
11 //
12 // The tool is under development, for the details about previous versions see
13 // http://code.google.com/p/data-race-test
14 //
15 // The instrumentation phase is quite simple:
16 //   - Insert calls to run-time library before every memory access.
17 //      - Optimizations may apply to avoid instrumenting some of the accesses.
18 //   - Insert calls at function entry/exit.
19 // The rest is handled by the run-time library.
20 //===----------------------------------------------------------------------===//
21 
22 #define DEBUG_TYPE "tsan"
23 
24 #include "llvm/ADT/SmallString.h"
25 #include "llvm/ADT/SmallVector.h"
26 #include "llvm/ADT/StringExtras.h"
27 #include "llvm/Intrinsics.h"
28 #include "llvm/Function.h"
29 #include "llvm/Module.h"
30 #include "llvm/Support/Debug.h"
31 #include "llvm/Support/IRBuilder.h"
32 #include "llvm/Support/MathExtras.h"
33 #include "llvm/Target/TargetData.h"
34 #include "llvm/Transforms/Instrumentation.h"
35 #include "llvm/Transforms/Utils/ModuleUtils.h"
36 #include "llvm/Type.h"
37 
38 using namespace llvm;
39 
40 namespace {
41 /// ThreadSanitizer: instrument the code in module to find races.
42 struct ThreadSanitizer : public FunctionPass {
43   ThreadSanitizer();
44   bool runOnFunction(Function &F);
45   bool doInitialization(Module &M);
46   bool instrumentLoadOrStore(Instruction *I);
47   static char ID;  // Pass identification, replacement for typeid.
48 
49  private:
50   TargetData *TD;
51   // Callbacks to run-time library are computed in doInitialization.
52   Value *TsanFuncEntry;
53   Value *TsanFuncExit;
54   // Accesses sizes are powers of two: 1, 2, 4, 8, 16.
55   static const size_t kNumberOfAccessSizes = 5;
56   Value *TsanRead[kNumberOfAccessSizes];
57   Value *TsanWrite[kNumberOfAccessSizes];
58 };
59 }  // namespace
60 
61 char ThreadSanitizer::ID = 0;
62 INITIALIZE_PASS(ThreadSanitizer, "tsan",
63     "ThreadSanitizer: detects data races.",
64     false, false)
65 
66 ThreadSanitizer::ThreadSanitizer()
67   : FunctionPass(ID),
68   TD(NULL) {
69 }
70 
71 FunctionPass *llvm::createThreadSanitizerPass() {
72   return new ThreadSanitizer();
73 }
74 
75 bool ThreadSanitizer::doInitialization(Module &M) {
76   TD = getAnalysisIfAvailable<TargetData>();
77   if (!TD)
78     return false;
79   // Always insert a call to __tsan_init into the module's CTORs.
80   IRBuilder<> IRB(M.getContext());
81   Value *TsanInit = M.getOrInsertFunction("__tsan_init",
82                                           IRB.getVoidTy(), NULL);
83   appendToGlobalCtors(M, cast<Function>(TsanInit), 0);
84 
85   // Initialize the callbacks.
86   TsanFuncEntry = M.getOrInsertFunction("__tsan_func_entry", IRB.getVoidTy(),
87                                         IRB.getInt8PtrTy(), NULL);
88   TsanFuncExit = M.getOrInsertFunction("__tsan_func_exit", IRB.getVoidTy(),
89                                        NULL);
90   for (size_t i = 0; i < kNumberOfAccessSizes; ++i) {
91     SmallString<32> ReadName("__tsan_read");
92     ReadName += itostr(1 << i);
93     TsanRead[i] = M.getOrInsertFunction(ReadName, IRB.getVoidTy(),
94                                         IRB.getInt8PtrTy(), NULL);
95     SmallString<32> WriteName("__tsan_write");
96     WriteName += itostr(1 << i);
97     TsanWrite[i] = M.getOrInsertFunction(WriteName, IRB.getVoidTy(),
98                                          IRB.getInt8PtrTy(), NULL);
99   }
100   return true;
101 }
102 
103 bool ThreadSanitizer::runOnFunction(Function &F) {
104   if (!TD) return false;
105   SmallVector<Instruction*, 8> RetVec;
106   SmallVector<Instruction*, 8> LoadsAndStores;
107   bool Res = false;
108   bool HasCalls = false;
109 
110   // Traverse all instructions, collect loads/stores/returns, check for calls.
111   for (Function::iterator FI = F.begin(), FE = F.end();
112        FI != FE; ++FI) {
113     BasicBlock &BB = *FI;
114     for (BasicBlock::iterator BI = BB.begin(), BE = BB.end();
115          BI != BE; ++BI) {
116       if (isa<LoadInst>(BI) || isa<StoreInst>(BI))
117         LoadsAndStores.push_back(BI);
118       else if (isa<ReturnInst>(BI))
119         RetVec.push_back(BI);
120       else if (isa<CallInst>(BI) || isa<InvokeInst>(BI))
121         HasCalls = true;
122     }
123   }
124 
125   // We have collected all loads and stores.
126   // FIXME: many of these accesses do not need to be checked for races
127   // (e.g. variables that do not escape, etc).
128 
129   // Instrument memory accesses.
130   for (size_t i = 0, n = LoadsAndStores.size(); i < n; ++i) {
131     Res |= instrumentLoadOrStore(LoadsAndStores[i]);
132   }
133 
134   // Instrument function entry/exit points if there were instrumented accesses.
135   if (Res || HasCalls) {
136     IRBuilder<> IRB(F.getEntryBlock().getFirstNonPHI());
137     Value *ReturnAddress = IRB.CreateCall(
138         Intrinsic::getDeclaration(F.getParent(), Intrinsic::returnaddress),
139         IRB.getInt32(0));
140     IRB.CreateCall(TsanFuncEntry, ReturnAddress);
141     for (size_t i = 0, n = RetVec.size(); i < n; ++i) {
142       IRBuilder<> IRBRet(RetVec[i]);
143       IRBRet.CreateCall(TsanFuncExit);
144     }
145   }
146   return Res;
147 }
148 
149 bool ThreadSanitizer::instrumentLoadOrStore(Instruction *I) {
150   IRBuilder<> IRB(I);
151   bool IsWrite = isa<StoreInst>(*I);
152   Value *Addr = IsWrite
153       ? cast<StoreInst>(I)->getPointerOperand()
154       : cast<LoadInst>(I)->getPointerOperand();
155   Type *OrigPtrTy = Addr->getType();
156   Type *OrigTy = cast<PointerType>(OrigPtrTy)->getElementType();
157   assert(OrigTy->isSized());
158   uint32_t TypeSize = TD->getTypeStoreSizeInBits(OrigTy);
159   if (TypeSize != 8  && TypeSize != 16 &&
160       TypeSize != 32 && TypeSize != 64 && TypeSize != 128) {
161     // Ignore all unusual sizes.
162     return false;
163   }
164   size_t Idx = CountTrailingZeros_32(TypeSize / 8);
165   assert(Idx < kNumberOfAccessSizes);
166   Value *OnAccessFunc = IsWrite ? TsanWrite[Idx] : TsanRead[Idx];
167   IRB.CreateCall(OnAccessFunc, IRB.CreatePointerCast(Addr, IRB.getInt8PtrTy()));
168   return true;
169 }
170