blob: 2f0b015dc9aba15ecc633ef2a34d806f1d4d1c8b [file] [log] [blame]
Kostya Serebryany800e03f2011-11-16 01:35:23 +00001//===-- AddressSanitizer.cpp - memory error detector ------------*- C++ -*-===//
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 AddressSanitizer, an address sanity checker.
11// Details of the algorithm:
12// http://code.google.com/p/address-sanitizer/wiki/AddressSanitizerAlgorithm
13//
14//===----------------------------------------------------------------------===//
15
16#define DEBUG_TYPE "asan"
17
Kostya Serebryanyb5b86d22012-08-24 16:44:47 +000018#include "BlackList.h"
Chandler Carruth06cb8ed2012-06-29 12:38:19 +000019#include "llvm/Function.h"
20#include "llvm/IRBuilder.h"
Kostya Serebryanyf7b08222012-07-20 09:54:50 +000021#include "llvm/InlineAsm.h"
Chandler Carruth06cb8ed2012-06-29 12:38:19 +000022#include "llvm/IntrinsicInst.h"
23#include "llvm/LLVMContext.h"
24#include "llvm/Module.h"
25#include "llvm/Type.h"
Kostya Serebryany800e03f2011-11-16 01:35:23 +000026#include "llvm/ADT/ArrayRef.h"
27#include "llvm/ADT/OwningPtr.h"
28#include "llvm/ADT/SmallSet.h"
29#include "llvm/ADT/SmallString.h"
30#include "llvm/ADT/SmallVector.h"
31#include "llvm/ADT/StringExtras.h"
Evgeniy Stepanov06fdbaa2012-05-23 11:52:12 +000032#include "llvm/ADT/Triple.h"
Kostya Serebryany800e03f2011-11-16 01:35:23 +000033#include "llvm/Support/CommandLine.h"
34#include "llvm/Support/DataTypes.h"
35#include "llvm/Support/Debug.h"
Kostya Serebryany800e03f2011-11-16 01:35:23 +000036#include "llvm/Support/raw_ostream.h"
37#include "llvm/Support/system_error.h"
38#include "llvm/Target/TargetData.h"
39#include "llvm/Target/TargetMachine.h"
40#include "llvm/Transforms/Instrumentation.h"
41#include "llvm/Transforms/Utils/BasicBlockUtils.h"
42#include "llvm/Transforms/Utils/ModuleUtils.h"
Kostya Serebryany800e03f2011-11-16 01:35:23 +000043
44#include <string>
45#include <algorithm>
46
47using namespace llvm;
48
49static const uint64_t kDefaultShadowScale = 3;
50static const uint64_t kDefaultShadowOffset32 = 1ULL << 29;
51static const uint64_t kDefaultShadowOffset64 = 1ULL << 44;
Evgeniy Stepanov06fdbaa2012-05-23 11:52:12 +000052static const uint64_t kDefaultShadowOffsetAndroid = 0;
Kostya Serebryany800e03f2011-11-16 01:35:23 +000053
54static const size_t kMaxStackMallocSize = 1 << 16; // 64K
55static const uintptr_t kCurrentStackFrameMagic = 0x41B58AB3;
56static const uintptr_t kRetiredStackFrameMagic = 0x45E0360E;
57
58static const char *kAsanModuleCtorName = "asan.module_ctor";
Kostya Serebryany7bcfc992011-12-15 21:59:03 +000059static const char *kAsanModuleDtorName = "asan.module_dtor";
60static const int kAsanCtorAndCtorPriority = 1;
Kostya Serebryany800e03f2011-11-16 01:35:23 +000061static const char *kAsanReportErrorTemplate = "__asan_report_";
62static const char *kAsanRegisterGlobalsName = "__asan_register_globals";
Kostya Serebryany7bcfc992011-12-15 21:59:03 +000063static const char *kAsanUnregisterGlobalsName = "__asan_unregister_globals";
Kostya Serebryany9b9f87a2012-08-21 08:24:25 +000064static const char *kAsanPoisonGlobalsName = "__asan_before_dynamic_init";
65static const char *kAsanUnpoisonGlobalsName = "__asan_after_dynamic_init";
Kostya Serebryany800e03f2011-11-16 01:35:23 +000066static const char *kAsanInitName = "__asan_init";
Kostya Serebryany95e3cf42012-02-08 21:36:17 +000067static const char *kAsanHandleNoReturnName = "__asan_handle_no_return";
Kostya Serebryany800e03f2011-11-16 01:35:23 +000068static const char *kAsanMappingOffsetName = "__asan_mapping_offset";
69static const char *kAsanMappingScaleName = "__asan_mapping_scale";
70static const char *kAsanStackMallocName = "__asan_stack_malloc";
71static const char *kAsanStackFreeName = "__asan_stack_free";
72
73static const int kAsanStackLeftRedzoneMagic = 0xf1;
74static const int kAsanStackMidRedzoneMagic = 0xf2;
75static const int kAsanStackRightRedzoneMagic = 0xf3;
76static const int kAsanStackPartialRedzoneMagic = 0xf4;
77
Kostya Serebryanyc0ed3e52012-07-16 16:15:40 +000078// Accesses sizes are powers of two: 1, 2, 4, 8, 16.
79static const size_t kNumberOfAccessSizes = 5;
80
Kostya Serebryany800e03f2011-11-16 01:35:23 +000081// Command-line flags.
82
83// This flag may need to be replaced with -f[no-]asan-reads.
84static cl::opt<bool> ClInstrumentReads("asan-instrument-reads",
85 cl::desc("instrument read instructions"), cl::Hidden, cl::init(true));
86static cl::opt<bool> ClInstrumentWrites("asan-instrument-writes",
87 cl::desc("instrument write instructions"), cl::Hidden, cl::init(true));
Kostya Serebryanye6cf2e02012-05-30 09:04:06 +000088static cl::opt<bool> ClInstrumentAtomics("asan-instrument-atomics",
89 cl::desc("instrument atomic instructions (rmw, cmpxchg)"),
90 cl::Hidden, cl::init(true));
Kostya Serebryany6e2d5062012-08-15 08:58:58 +000091static cl::opt<bool> ClAlwaysSlowPath("asan-always-slow-path",
92 cl::desc("use instrumentation with slow path for all accesses"),
93 cl::Hidden, cl::init(false));
Kostya Serebryanyc0ed3e52012-07-16 16:15:40 +000094// This flag limits the number of instructions to be instrumented
Kostya Serebryany324cbb82012-06-28 09:34:41 +000095// in any given BB. Normally, this should be set to unlimited (INT_MAX),
96// but due to http://llvm.org/bugs/show_bug.cgi?id=12652 we temporary
97// set it to 10000.
98static cl::opt<int> ClMaxInsnsToInstrumentPerBB("asan-max-ins-per-bb",
99 cl::init(10000),
100 cl::desc("maximal number of instructions to instrument in any given BB"),
101 cl::Hidden);
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000102// This flag may need to be replaced with -f[no]asan-stack.
103static cl::opt<bool> ClStack("asan-stack",
104 cl::desc("Handle stack memory"), cl::Hidden, cl::init(true));
105// This flag may need to be replaced with -f[no]asan-use-after-return.
106static cl::opt<bool> ClUseAfterReturn("asan-use-after-return",
107 cl::desc("Check return-after-free"), cl::Hidden, cl::init(false));
108// This flag may need to be replaced with -f[no]asan-globals.
109static cl::opt<bool> ClGlobals("asan-globals",
110 cl::desc("Handle global objects"), cl::Hidden, cl::init(true));
Kostya Serebryany9b9f87a2012-08-21 08:24:25 +0000111static cl::opt<bool> ClInitializers("asan-initialization-order",
112 cl::desc("Handle C++ initializer order"), cl::Hidden, cl::init(false));
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000113static cl::opt<bool> ClMemIntrin("asan-memintrin",
114 cl::desc("Handle memset/memcpy/memmove"), cl::Hidden, cl::init(true));
115// This flag may need to be replaced with -fasan-blacklist.
116static cl::opt<std::string> ClBlackListFile("asan-blacklist",
117 cl::desc("File containing the list of functions to ignore "
118 "during instrumentation"), cl::Hidden);
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000119
120// These flags allow to change the shadow mapping.
121// The shadow mapping looks like
122// Shadow = (Mem >> scale) + (1 << offset_log)
123static cl::opt<int> ClMappingScale("asan-mapping-scale",
124 cl::desc("scale of asan shadow mapping"), cl::Hidden, cl::init(0));
125static cl::opt<int> ClMappingOffsetLog("asan-mapping-offset-log",
126 cl::desc("offset of asan shadow mapping"), cl::Hidden, cl::init(-1));
127
128// Optimization flags. Not user visible, used mostly for testing
129// and benchmarking the tool.
130static cl::opt<bool> ClOpt("asan-opt",
131 cl::desc("Optimize instrumentation"), cl::Hidden, cl::init(true));
132static cl::opt<bool> ClOptSameTemp("asan-opt-same-temp",
133 cl::desc("Instrument the same temp just once"), cl::Hidden,
134 cl::init(true));
135static cl::opt<bool> ClOptGlobals("asan-opt-globals",
136 cl::desc("Don't instrument scalar globals"), cl::Hidden, cl::init(true));
137
138// Debug flags.
139static cl::opt<int> ClDebug("asan-debug", cl::desc("debug"), cl::Hidden,
140 cl::init(0));
141static cl::opt<int> ClDebugStack("asan-debug-stack", cl::desc("debug stack"),
142 cl::Hidden, cl::init(0));
143static cl::opt<std::string> ClDebugFunc("asan-debug-func",
144 cl::Hidden, cl::desc("Debug func"));
145static cl::opt<int> ClDebugMin("asan-debug-min", cl::desc("Debug min inst"),
146 cl::Hidden, cl::init(-1));
147static cl::opt<int> ClDebugMax("asan-debug-max", cl::desc("Debug man inst"),
148 cl::Hidden, cl::init(-1));
149
150namespace {
151
Kostya Serebryanyc0ed3e52012-07-16 16:15:40 +0000152/// An object of this type is created while instrumenting every function.
153struct AsanFunctionContext {
Kostya Serebryany11c2a472012-08-13 14:08:46 +0000154 AsanFunctionContext(Function &Function) : F(Function) { }
Kostya Serebryanyc0ed3e52012-07-16 16:15:40 +0000155
Kostya Serebryany2735cf42012-07-16 17:12:07 +0000156 Function &F;
Kostya Serebryanyc0ed3e52012-07-16 16:15:40 +0000157};
158
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000159/// AddressSanitizer: instrument the code in module to find memory bugs.
160struct AddressSanitizer : public ModulePass {
161 AddressSanitizer();
Alexander Potapenko25878042012-01-23 11:22:43 +0000162 virtual const char *getPassName() const;
Kostya Serebryanyc0ed3e52012-07-16 16:15:40 +0000163 void instrumentMop(AsanFunctionContext &AFC, Instruction *I);
164 void instrumentAddress(AsanFunctionContext &AFC,
165 Instruction *OrigIns, IRBuilder<> &IRB,
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000166 Value *Addr, uint32_t TypeSize, bool IsWrite);
Kostya Serebryanyc0ed3e52012-07-16 16:15:40 +0000167 Value *createSlowPathCmp(IRBuilder<> &IRB, Value *AddrLong,
168 Value *ShadowValue, uint32_t TypeSize);
Kostya Serebryanyebd64542012-08-14 14:04:51 +0000169 Instruction *generateCrashCode(Instruction *InsertBefore, Value *Addr,
Kostya Serebryany2735cf42012-07-16 17:12:07 +0000170 bool IsWrite, size_t AccessSizeIndex);
Kostya Serebryanyc0ed3e52012-07-16 16:15:40 +0000171 bool instrumentMemIntrinsic(AsanFunctionContext &AFC, MemIntrinsic *MI);
172 void instrumentMemIntrinsicParam(AsanFunctionContext &AFC,
173 Instruction *OrigIns, Value *Addr,
174 Value *Size,
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000175 Instruction *InsertBefore, bool IsWrite);
176 Value *memToShadow(Value *Shadow, IRBuilder<> &IRB);
177 bool handleFunction(Module &M, Function &F);
Kostya Serebryany9b9f87a2012-08-21 08:24:25 +0000178 void createInitializerPoisonCalls(Module &M,
179 Value *FirstAddr, Value *LastAddr);
Kostya Serebryanya1a8a322012-01-30 23:50:10 +0000180 bool maybeInsertAsanInitAtFunctionEntry(Function &F);
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000181 bool poisonStackInFunction(Module &M, Function &F);
182 virtual bool runOnModule(Module &M);
183 bool insertGlobalRedzones(Module &M);
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000184 static char ID; // Pass identification, replacement for typeid
185
186 private:
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000187 uint64_t getAllocaSizeInBytes(AllocaInst *AI) {
188 Type *Ty = AI->getAllocatedType();
Evgeniy Stepanovd8313be2012-03-02 10:41:08 +0000189 uint64_t SizeInBytes = TD->getTypeAllocSize(Ty);
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000190 return SizeInBytes;
191 }
192 uint64_t getAlignedSize(uint64_t SizeInBytes) {
193 return ((SizeInBytes + RedzoneSize - 1)
194 / RedzoneSize) * RedzoneSize;
195 }
196 uint64_t getAlignedAllocaSize(AllocaInst *AI) {
197 uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
198 return getAlignedSize(SizeInBytes);
199 }
200
Alexander Potapenko55cabae2012-04-23 10:47:31 +0000201 Function *checkInterfaceFunction(Constant *FuncOrBitcast);
Kostya Serebryany9b9f87a2012-08-21 08:24:25 +0000202 bool ShouldInstrumentGlobal(GlobalVariable *G);
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000203 void PoisonStack(const ArrayRef<AllocaInst*> &AllocaVec, IRBuilder<> IRB,
204 Value *ShadowBase, bool DoPoison);
Kostya Serebryany5a3a9c92011-11-18 01:41:06 +0000205 bool LooksLikeCodeInBug11395(Instruction *I);
Kostya Serebryany9b9f87a2012-08-21 08:24:25 +0000206 void FindDynamicInitializers(Module &M);
207 bool HasDynamicInitializer(GlobalVariable *G);
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000208
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000209 LLVMContext *C;
210 TargetData *TD;
211 uint64_t MappingOffset;
212 int MappingScale;
213 size_t RedzoneSize;
214 int LongSize;
215 Type *IntptrTy;
216 Type *IntptrPtrTy;
217 Function *AsanCtorFunction;
218 Function *AsanInitFunction;
219 Instruction *CtorInsertBefore;
Kostya Serebryanyb5b86d22012-08-24 16:44:47 +0000220 OwningPtr<BlackList> BL;
Kostya Serebryany9db5b5f2012-07-16 14:09:42 +0000221 // This array is indexed by AccessIsWrite and log2(AccessSize).
222 Function *AsanErrorCallback[2][kNumberOfAccessSizes];
Kostya Serebryanyf7b08222012-07-20 09:54:50 +0000223 InlineAsm *EmptyAsm;
Kostya Serebryany9b9f87a2012-08-21 08:24:25 +0000224 SmallSet<GlobalValue*, 32> DynamicallyInitializedGlobals;
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000225};
Kostya Serebryanyc0ed3e52012-07-16 16:15:40 +0000226
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000227} // namespace
228
229char AddressSanitizer::ID = 0;
230INITIALIZE_PASS(AddressSanitizer, "asan",
231 "AddressSanitizer: detects use-after-free and out-of-bounds bugs.",
232 false, false)
233AddressSanitizer::AddressSanitizer() : ModulePass(ID) { }
234ModulePass *llvm::createAddressSanitizerPass() {
235 return new AddressSanitizer();
236}
237
Alexander Potapenko25878042012-01-23 11:22:43 +0000238const char *AddressSanitizer::getPassName() const {
239 return "AddressSanitizer";
240}
241
Kostya Serebryany2735cf42012-07-16 17:12:07 +0000242static size_t TypeSizeToSizeIndex(uint32_t TypeSize) {
243 size_t Res = CountTrailingZeros_32(TypeSize / 8);
244 assert(Res < kNumberOfAccessSizes);
245 return Res;
246}
247
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000248// Create a constant for Str so that we can pass it to the run-time lib.
249static GlobalVariable *createPrivateGlobalForString(Module &M, StringRef Str) {
Chris Lattner18c7f802012-02-05 02:29:43 +0000250 Constant *StrConst = ConstantDataArray::getString(M.getContext(), Str);
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000251 return new GlobalVariable(M, StrConst->getType(), true,
252 GlobalValue::PrivateLinkage, StrConst, "");
253}
254
255// Split the basic block and insert an if-then code.
256// Before:
257// Head
Kostya Serebryany56139bc2012-07-02 11:42:29 +0000258// Cmp
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000259// Tail
260// After:
261// Head
262// if (Cmp)
Kostya Serebryany56139bc2012-07-02 11:42:29 +0000263// ThenBlock
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000264// Tail
265//
Kostya Serebryanyebd64542012-08-14 14:04:51 +0000266// ThenBlock block is created and its terminator is returned.
267// If Unreachable, ThenBlock is terminated with UnreachableInst, otherwise
268// it is terminated with BranchInst to Tail.
269static TerminatorInst *splitBlockAndInsertIfThen(Value *Cmp, bool Unreachable) {
Kostya Serebryany56139bc2012-07-02 11:42:29 +0000270 Instruction *SplitBefore = cast<Instruction>(Cmp)->getNextNode();
Chandler Carruthc3c8db92012-07-16 08:58:53 +0000271 BasicBlock *Head = SplitBefore->getParent();
Chandler Carruth349f14c2012-07-16 10:01:02 +0000272 BasicBlock *Tail = Head->splitBasicBlock(SplitBefore);
Chandler Carruthc3c8db92012-07-16 08:58:53 +0000273 TerminatorInst *HeadOldTerm = Head->getTerminator();
Kostya Serebryanyebd64542012-08-14 14:04:51 +0000274 LLVMContext &C = Head->getParent()->getParent()->getContext();
275 BasicBlock *ThenBlock = BasicBlock::Create(C, "", Head->getParent(), Tail);
276 TerminatorInst *CheckTerm;
277 if (Unreachable)
278 CheckTerm = new UnreachableInst(C, ThenBlock);
279 else
Kostya Serebryanyc0ed3e52012-07-16 16:15:40 +0000280 CheckTerm = BranchInst::Create(Tail, ThenBlock);
Chandler Carruth349f14c2012-07-16 10:01:02 +0000281 BranchInst *HeadNewTerm =
282 BranchInst::Create(/*ifTrue*/ThenBlock, /*ifFalse*/Tail, Cmp);
283 ReplaceInstWithInst(HeadOldTerm, HeadNewTerm);
Chandler Carruth349f14c2012-07-16 10:01:02 +0000284 return CheckTerm;
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000285}
286
287Value *AddressSanitizer::memToShadow(Value *Shadow, IRBuilder<> &IRB) {
288 // Shadow >> scale
289 Shadow = IRB.CreateLShr(Shadow, MappingScale);
290 if (MappingOffset == 0)
291 return Shadow;
292 // (Shadow >> scale) | offset
293 return IRB.CreateOr(Shadow, ConstantInt::get(IntptrTy,
294 MappingOffset));
295}
296
Kostya Serebryanyc0ed3e52012-07-16 16:15:40 +0000297void AddressSanitizer::instrumentMemIntrinsicParam(
298 AsanFunctionContext &AFC, Instruction *OrigIns,
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000299 Value *Addr, Value *Size, Instruction *InsertBefore, bool IsWrite) {
300 // Check the first byte.
301 {
302 IRBuilder<> IRB(InsertBefore);
Kostya Serebryanyc0ed3e52012-07-16 16:15:40 +0000303 instrumentAddress(AFC, OrigIns, IRB, Addr, 8, IsWrite);
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000304 }
305 // Check the last byte.
306 {
307 IRBuilder<> IRB(InsertBefore);
308 Value *SizeMinusOne = IRB.CreateSub(
309 Size, ConstantInt::get(Size->getType(), 1));
310 SizeMinusOne = IRB.CreateIntCast(SizeMinusOne, IntptrTy, false);
311 Value *AddrLong = IRB.CreatePointerCast(Addr, IntptrTy);
312 Value *AddrPlusSizeMinisOne = IRB.CreateAdd(AddrLong, SizeMinusOne);
Kostya Serebryanyc0ed3e52012-07-16 16:15:40 +0000313 instrumentAddress(AFC, OrigIns, IRB, AddrPlusSizeMinisOne, 8, IsWrite);
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000314 }
315}
316
317// Instrument memset/memmove/memcpy
Kostya Serebryanyc0ed3e52012-07-16 16:15:40 +0000318bool AddressSanitizer::instrumentMemIntrinsic(AsanFunctionContext &AFC,
319 MemIntrinsic *MI) {
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000320 Value *Dst = MI->getDest();
321 MemTransferInst *MemTran = dyn_cast<MemTransferInst>(MI);
Kostya Serebryany2735cf42012-07-16 17:12:07 +0000322 Value *Src = MemTran ? MemTran->getSource() : 0;
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000323 Value *Length = MI->getLength();
324
325 Constant *ConstLength = dyn_cast<Constant>(Length);
326 Instruction *InsertBefore = MI;
327 if (ConstLength) {
328 if (ConstLength->isNullValue()) return false;
329 } else {
330 // The size is not a constant so it could be zero -- check at run-time.
331 IRBuilder<> IRB(InsertBefore);
332
333 Value *Cmp = IRB.CreateICmpNE(Length,
Kostya Serebryany56139bc2012-07-02 11:42:29 +0000334 Constant::getNullValue(Length->getType()));
Kostya Serebryanyebd64542012-08-14 14:04:51 +0000335 InsertBefore = splitBlockAndInsertIfThen(Cmp, false);
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000336 }
337
Kostya Serebryanyc0ed3e52012-07-16 16:15:40 +0000338 instrumentMemIntrinsicParam(AFC, MI, Dst, Length, InsertBefore, true);
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000339 if (Src)
Kostya Serebryanyc0ed3e52012-07-16 16:15:40 +0000340 instrumentMemIntrinsicParam(AFC, MI, Src, Length, InsertBefore, false);
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000341 return true;
342}
343
Kostya Serebryanye6cf2e02012-05-30 09:04:06 +0000344// If I is an interesting memory access, return the PointerOperand
345// and set IsWrite. Otherwise return NULL.
346static Value *isInterestingMemoryAccess(Instruction *I, bool *IsWrite) {
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000347 if (LoadInst *LI = dyn_cast<LoadInst>(I)) {
Kostya Serebryanye6cf2e02012-05-30 09:04:06 +0000348 if (!ClInstrumentReads) return NULL;
349 *IsWrite = false;
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000350 return LI->getPointerOperand();
351 }
Kostya Serebryanye6cf2e02012-05-30 09:04:06 +0000352 if (StoreInst *SI = dyn_cast<StoreInst>(I)) {
353 if (!ClInstrumentWrites) return NULL;
354 *IsWrite = true;
355 return SI->getPointerOperand();
356 }
357 if (AtomicRMWInst *RMW = dyn_cast<AtomicRMWInst>(I)) {
358 if (!ClInstrumentAtomics) return NULL;
359 *IsWrite = true;
360 return RMW->getPointerOperand();
361 }
362 if (AtomicCmpXchgInst *XCHG = dyn_cast<AtomicCmpXchgInst>(I)) {
363 if (!ClInstrumentAtomics) return NULL;
364 *IsWrite = true;
365 return XCHG->getPointerOperand();
366 }
367 return NULL;
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000368}
369
Kostya Serebryany9b9f87a2012-08-21 08:24:25 +0000370void AddressSanitizer::FindDynamicInitializers(Module& M) {
371 // Clang generates metadata identifying all dynamically initialized globals.
372 NamedMDNode *DynamicGlobals =
373 M.getNamedMetadata("llvm.asan.dynamically_initialized_globals");
374 if (!DynamicGlobals)
375 return;
376 for (int i = 0, n = DynamicGlobals->getNumOperands(); i < n; ++i) {
377 MDNode *MDN = DynamicGlobals->getOperand(i);
378 assert(MDN->getNumOperands() == 1);
379 Value *VG = MDN->getOperand(0);
380 // The optimizer may optimize away a global entirely, in which case we
381 // cannot instrument access to it.
382 if (!VG)
383 continue;
384
385 GlobalVariable *G = cast<GlobalVariable>(VG);
386 DynamicallyInitializedGlobals.insert(G);
387 }
388}
389// Returns true if a global variable is initialized dynamically in this TU.
390bool AddressSanitizer::HasDynamicInitializer(GlobalVariable *G) {
391 return DynamicallyInitializedGlobals.count(G);
392}
393
Kostya Serebryanyc0ed3e52012-07-16 16:15:40 +0000394void AddressSanitizer::instrumentMop(AsanFunctionContext &AFC, Instruction *I) {
Kostya Serebryanye6cf2e02012-05-30 09:04:06 +0000395 bool IsWrite;
396 Value *Addr = isInterestingMemoryAccess(I, &IsWrite);
397 assert(Addr);
Kostya Serebryany9b9f87a2012-08-21 08:24:25 +0000398 if (ClOpt && ClOptGlobals) {
399 if (GlobalVariable *G = dyn_cast<GlobalVariable>(Addr)) {
400 // If initialization order checking is disabled, a simple access to a
401 // dynamically initialized global is always valid.
402 if (!ClInitializers)
403 return;
404 // If a global variable does not have dynamic initialization we don't
405 // have to instrument it. However, if a global has external linkage, we
406 // assume it has dynamic initialization, as it may have an initializer
407 // in a different TU.
408 if (G->getLinkage() != GlobalVariable::ExternalLinkage &&
409 !HasDynamicInitializer(G))
410 return;
411 }
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000412 }
Kostya Serebryany9b9f87a2012-08-21 08:24:25 +0000413
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000414 Type *OrigPtrTy = Addr->getType();
415 Type *OrigTy = cast<PointerType>(OrigPtrTy)->getElementType();
416
417 assert(OrigTy->isSized());
418 uint32_t TypeSize = TD->getTypeStoreSizeInBits(OrigTy);
419
420 if (TypeSize != 8 && TypeSize != 16 &&
421 TypeSize != 32 && TypeSize != 64 && TypeSize != 128) {
422 // Ignore all unusual sizes.
423 return;
424 }
425
426 IRBuilder<> IRB(I);
Kostya Serebryanyc0ed3e52012-07-16 16:15:40 +0000427 instrumentAddress(AFC, I, IRB, Addr, TypeSize, IsWrite);
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000428}
429
Alexander Potapenko55cabae2012-04-23 10:47:31 +0000430// Validate the result of Module::getOrInsertFunction called for an interface
431// function of AddressSanitizer. If the instrumented module defines a function
432// with the same name, their prototypes must match, otherwise
433// getOrInsertFunction returns a bitcast.
434Function *AddressSanitizer::checkInterfaceFunction(Constant *FuncOrBitcast) {
435 if (isa<Function>(FuncOrBitcast)) return cast<Function>(FuncOrBitcast);
436 FuncOrBitcast->dump();
437 report_fatal_error("trying to redefine an AddressSanitizer "
438 "interface function");
439}
440
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000441Instruction *AddressSanitizer::generateCrashCode(
Kostya Serebryanyebd64542012-08-14 14:04:51 +0000442 Instruction *InsertBefore, Value *Addr,
Kostya Serebryany4f0c6962012-07-17 11:04:12 +0000443 bool IsWrite, size_t AccessSizeIndex) {
Kostya Serebryanyebd64542012-08-14 14:04:51 +0000444 IRBuilder<> IRB(InsertBefore);
445 CallInst *Call = IRB.CreateCall(AsanErrorCallback[IsWrite][AccessSizeIndex],
446 Addr);
Kostya Serebryanyf7b08222012-07-20 09:54:50 +0000447 // We don't do Call->setDoesNotReturn() because the BB already has
448 // UnreachableInst at the end.
449 // This EmptyAsm is required to avoid callback merge.
450 IRB.CreateCall(EmptyAsm);
Kostya Serebryany3c7faae2012-01-06 18:09:21 +0000451 return Call;
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000452}
453
Kostya Serebryany2735cf42012-07-16 17:12:07 +0000454Value *AddressSanitizer::createSlowPathCmp(IRBuilder<> &IRB, Value *AddrLong,
Kostya Serebryanyc0ed3e52012-07-16 16:15:40 +0000455 Value *ShadowValue,
456 uint32_t TypeSize) {
457 size_t Granularity = 1 << MappingScale;
458 // Addr & (Granularity - 1)
459 Value *LastAccessedByte = IRB.CreateAnd(
460 AddrLong, ConstantInt::get(IntptrTy, Granularity - 1));
461 // (Addr & (Granularity - 1)) + size - 1
462 if (TypeSize / 8 > 1)
463 LastAccessedByte = IRB.CreateAdd(
464 LastAccessedByte, ConstantInt::get(IntptrTy, TypeSize / 8 - 1));
465 // (uint8_t) ((Addr & (Granularity-1)) + size - 1)
466 LastAccessedByte = IRB.CreateIntCast(
Kostya Serebryany6e2d5062012-08-15 08:58:58 +0000467 LastAccessedByte, ShadowValue->getType(), false);
Kostya Serebryanyc0ed3e52012-07-16 16:15:40 +0000468 // ((uint8_t) ((Addr & (Granularity-1)) + size - 1)) >= ShadowValue
469 return IRB.CreateICmpSGE(LastAccessedByte, ShadowValue);
470}
471
472void AddressSanitizer::instrumentAddress(AsanFunctionContext &AFC,
473 Instruction *OrigIns,
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000474 IRBuilder<> &IRB, Value *Addr,
475 uint32_t TypeSize, bool IsWrite) {
476 Value *AddrLong = IRB.CreatePointerCast(Addr, IntptrTy);
477
478 Type *ShadowTy = IntegerType::get(
479 *C, std::max(8U, TypeSize >> MappingScale));
480 Type *ShadowPtrTy = PointerType::get(ShadowTy, 0);
481 Value *ShadowPtr = memToShadow(AddrLong, IRB);
482 Value *CmpVal = Constant::getNullValue(ShadowTy);
483 Value *ShadowValue = IRB.CreateLoad(
484 IRB.CreateIntToPtr(ShadowPtr, ShadowPtrTy));
485
486 Value *Cmp = IRB.CreateICmpNE(ShadowValue, CmpVal);
Kostya Serebryany11c2a472012-08-13 14:08:46 +0000487 size_t AccessSizeIndex = TypeSizeToSizeIndex(TypeSize);
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000488 size_t Granularity = 1 << MappingScale;
Kostya Serebryanyebd64542012-08-14 14:04:51 +0000489 TerminatorInst *CrashTerm = 0;
490
Kostya Serebryany6e2d5062012-08-15 08:58:58 +0000491 if (ClAlwaysSlowPath || (TypeSize < 8 * Granularity)) {
Kostya Serebryanyebd64542012-08-14 14:04:51 +0000492 TerminatorInst *CheckTerm = splitBlockAndInsertIfThen(Cmp, false);
493 assert(dyn_cast<BranchInst>(CheckTerm)->isUnconditional());
Kostya Serebryanyf7b08222012-07-20 09:54:50 +0000494 BasicBlock *NextBB = CheckTerm->getSuccessor(0);
Kostya Serebryanyc0ed3e52012-07-16 16:15:40 +0000495 IRB.SetInsertPoint(CheckTerm);
496 Value *Cmp2 = createSlowPathCmp(IRB, AddrLong, ShadowValue, TypeSize);
Kostya Serebryanyebd64542012-08-14 14:04:51 +0000497 BasicBlock *CrashBlock = BasicBlock::Create(*C, "", &AFC.F, NextBB);
498 CrashTerm = new UnreachableInst(*C, CrashBlock);
Kostya Serebryanyf7b08222012-07-20 09:54:50 +0000499 BranchInst *NewTerm = BranchInst::Create(CrashBlock, NextBB, Cmp2);
500 ReplaceInstWithInst(CheckTerm, NewTerm);
Kostya Serebryanyc0ed3e52012-07-16 16:15:40 +0000501 } else {
Kostya Serebryanyebd64542012-08-14 14:04:51 +0000502 CrashTerm = splitBlockAndInsertIfThen(Cmp, true);
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000503 }
Kostya Serebryanyebd64542012-08-14 14:04:51 +0000504
505 Instruction *Crash =
506 generateCrashCode(CrashTerm, AddrLong, IsWrite, AccessSizeIndex);
507 Crash->setDebugLoc(OrigIns->getDebugLoc());
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000508}
509
Kostya Serebryany9b9f87a2012-08-21 08:24:25 +0000510void AddressSanitizer::createInitializerPoisonCalls(Module &M,
511 Value *FirstAddr,
512 Value *LastAddr) {
513 // We do all of our poisoning and unpoisoning within _GLOBAL__I_a.
514 Function *GlobalInit = M.getFunction("_GLOBAL__I_a");
515 // If that function is not present, this TU contains no globals, or they have
516 // all been optimized away
517 if (!GlobalInit)
518 return;
519
520 // Set up the arguments to our poison/unpoison functions.
521 IRBuilder<> IRB(GlobalInit->begin()->getFirstInsertionPt());
522
523 // Declare our poisoning and unpoisoning functions.
524 Function *AsanPoisonGlobals = checkInterfaceFunction(M.getOrInsertFunction(
525 kAsanPoisonGlobalsName, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
526 AsanPoisonGlobals->setLinkage(Function::ExternalLinkage);
527 Function *AsanUnpoisonGlobals = checkInterfaceFunction(M.getOrInsertFunction(
528 kAsanUnpoisonGlobalsName, IRB.getVoidTy(), NULL));
529 AsanUnpoisonGlobals->setLinkage(Function::ExternalLinkage);
530
531 // Add a call to poison all external globals before the given function starts.
532 IRB.CreateCall2(AsanPoisonGlobals, FirstAddr, LastAddr);
533
534 // Add calls to unpoison all globals before each return instruction.
535 for (Function::iterator I = GlobalInit->begin(), E = GlobalInit->end();
536 I != E; ++I) {
537 if (ReturnInst *RI = dyn_cast<ReturnInst>(I->getTerminator())) {
538 CallInst::Create(AsanUnpoisonGlobals, "", RI);
539 }
540 }
541}
542
543bool AddressSanitizer::ShouldInstrumentGlobal(GlobalVariable *G) {
544 Type *Ty = cast<PointerType>(G->getType())->getElementType();
545 DEBUG(dbgs() << "GLOBAL: " << *G);
546
547 if (!Ty->isSized()) return false;
548 if (!G->hasInitializer()) return false;
549 // Touch only those globals that will not be defined in other modules.
550 // Don't handle ODR type linkages since other modules may be built w/o asan.
551 if (G->getLinkage() != GlobalVariable::ExternalLinkage &&
552 G->getLinkage() != GlobalVariable::PrivateLinkage &&
553 G->getLinkage() != GlobalVariable::InternalLinkage)
554 return false;
555 // Two problems with thread-locals:
556 // - The address of the main thread's copy can't be computed at link-time.
557 // - Need to poison all copies, not just the main thread's one.
558 if (G->isThreadLocal())
559 return false;
560 // For now, just ignore this Alloca if the alignment is large.
561 if (G->getAlignment() > RedzoneSize) return false;
562
563 // Ignore all the globals with the names starting with "\01L_OBJC_".
564 // Many of those are put into the .cstring section. The linker compresses
565 // that section by removing the spare \0s after the string terminator, so
566 // our redzones get broken.
567 if ((G->getName().find("\01L_OBJC_") == 0) ||
568 (G->getName().find("\01l_OBJC_") == 0)) {
569 DEBUG(dbgs() << "Ignoring \\01L_OBJC_* global: " << *G);
570 return false;
571 }
572
573 if (G->hasSection()) {
574 StringRef Section(G->getSection());
575 // Ignore the globals from the __OBJC section. The ObjC runtime assumes
576 // those conform to /usr/lib/objc/runtime.h, so we can't add redzones to
577 // them.
578 if ((Section.find("__OBJC,") == 0) ||
579 (Section.find("__DATA, __objc_") == 0)) {
580 DEBUG(dbgs() << "Ignoring ObjC runtime global: " << *G);
581 return false;
582 }
583 // See http://code.google.com/p/address-sanitizer/issues/detail?id=32
584 // Constant CFString instances are compiled in the following way:
585 // -- the string buffer is emitted into
586 // __TEXT,__cstring,cstring_literals
587 // -- the constant NSConstantString structure referencing that buffer
588 // is placed into __DATA,__cfstring
589 // Therefore there's no point in placing redzones into __DATA,__cfstring.
590 // Moreover, it causes the linker to crash on OS X 10.7
591 if (Section.find("__DATA,__cfstring") == 0) {
592 DEBUG(dbgs() << "Ignoring CFString: " << *G);
593 return false;
594 }
595 }
596
597 return true;
598}
599
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000600// This function replaces all global variables with new variables that have
601// trailing redzones. It also creates a function that poisons
602// redzones and inserts this function into llvm.global_ctors.
603bool AddressSanitizer::insertGlobalRedzones(Module &M) {
604 SmallVector<GlobalVariable *, 16> GlobalsToChange;
605
Kostya Serebryany9b9f87a2012-08-21 08:24:25 +0000606 for (Module::GlobalListType::iterator G = M.global_begin(),
607 E = M.global_end(); G != E; ++G) {
608 if (ShouldInstrumentGlobal(G))
609 GlobalsToChange.push_back(G);
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000610 }
611
612 size_t n = GlobalsToChange.size();
613 if (n == 0) return false;
614
615 // A global is described by a structure
616 // size_t beg;
617 // size_t size;
618 // size_t size_with_redzone;
619 // const char *name;
Kostya Serebryany9b9f87a2012-08-21 08:24:25 +0000620 // size_t has_dynamic_init;
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000621 // We initialize an array of such structures and pass it to a run-time call.
622 StructType *GlobalStructTy = StructType::get(IntptrTy, IntptrTy,
Kostya Serebryany9b9f87a2012-08-21 08:24:25 +0000623 IntptrTy, IntptrTy,
624 IntptrTy, NULL);
625 SmallVector<Constant *, 16> Initializers(n), DynamicInit;
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000626
627 IRBuilder<> IRB(CtorInsertBefore);
628
Kostya Serebryany9b9f87a2012-08-21 08:24:25 +0000629 if (ClInitializers)
630 FindDynamicInitializers(M);
631
632 // The addresses of the first and last dynamically initialized globals in
633 // this TU. Used in initialization order checking.
634 Value *FirstDynamic = 0, *LastDynamic = 0;
635
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000636 for (size_t i = 0; i < n; i++) {
637 GlobalVariable *G = GlobalsToChange[i];
638 PointerType *PtrTy = cast<PointerType>(G->getType());
639 Type *Ty = PtrTy->getElementType();
Kostya Serebryany208a4ff2012-03-21 15:28:50 +0000640 uint64_t SizeInBytes = TD->getTypeAllocSize(Ty);
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000641 uint64_t RightRedzoneSize = RedzoneSize +
642 (RedzoneSize - (SizeInBytes % RedzoneSize));
643 Type *RightRedZoneTy = ArrayType::get(IRB.getInt8Ty(), RightRedzoneSize);
Kostya Serebryany9b9f87a2012-08-21 08:24:25 +0000644 // Determine whether this global should be poisoned in initialization.
645 bool GlobalHasDynamicInitializer = HasDynamicInitializer(G);
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000646
647 StructType *NewTy = StructType::get(Ty, RightRedZoneTy, NULL);
648 Constant *NewInitializer = ConstantStruct::get(
649 NewTy, G->getInitializer(),
650 Constant::getNullValue(RightRedZoneTy), NULL);
651
Kostya Serebryanya4b2b1d2011-12-15 22:55:55 +0000652 SmallString<2048> DescriptionOfGlobal = G->getName();
653 DescriptionOfGlobal += " (";
654 DescriptionOfGlobal += M.getModuleIdentifier();
655 DescriptionOfGlobal += ")";
656 GlobalVariable *Name = createPrivateGlobalForString(M, DescriptionOfGlobal);
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000657
658 // Create a new global variable with enough space for a redzone.
659 GlobalVariable *NewGlobal = new GlobalVariable(
660 M, NewTy, G->isConstant(), G->getLinkage(),
Hans Wennborgce718ff2012-06-23 11:37:03 +0000661 NewInitializer, "", G, G->getThreadLocalMode());
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000662 NewGlobal->copyAttributesFrom(G);
663 NewGlobal->setAlignment(RedzoneSize);
664
665 Value *Indices2[2];
666 Indices2[0] = IRB.getInt32(0);
667 Indices2[1] = IRB.getInt32(0);
668
669 G->replaceAllUsesWith(
Kostya Serebryanyf1639ab2012-01-28 04:27:16 +0000670 ConstantExpr::getGetElementPtr(NewGlobal, Indices2, true));
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000671 NewGlobal->takeName(G);
672 G->eraseFromParent();
673
674 Initializers[i] = ConstantStruct::get(
675 GlobalStructTy,
676 ConstantExpr::getPointerCast(NewGlobal, IntptrTy),
677 ConstantInt::get(IntptrTy, SizeInBytes),
678 ConstantInt::get(IntptrTy, SizeInBytes + RightRedzoneSize),
679 ConstantExpr::getPointerCast(Name, IntptrTy),
Kostya Serebryany9b9f87a2012-08-21 08:24:25 +0000680 ConstantInt::get(IntptrTy, GlobalHasDynamicInitializer),
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000681 NULL);
Kostya Serebryany9b9f87a2012-08-21 08:24:25 +0000682
683 // Populate the first and last globals declared in this TU.
684 if (ClInitializers && GlobalHasDynamicInitializer) {
685 LastDynamic = ConstantExpr::getPointerCast(NewGlobal, IntptrTy);
686 if (FirstDynamic == 0)
687 FirstDynamic = LastDynamic;
688 }
689
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000690 DEBUG(dbgs() << "NEW GLOBAL:\n" << *NewGlobal);
691 }
692
693 ArrayType *ArrayOfGlobalStructTy = ArrayType::get(GlobalStructTy, n);
694 GlobalVariable *AllGlobals = new GlobalVariable(
695 M, ArrayOfGlobalStructTy, false, GlobalVariable::PrivateLinkage,
696 ConstantArray::get(ArrayOfGlobalStructTy, Initializers), "");
697
Kostya Serebryany9b9f87a2012-08-21 08:24:25 +0000698 // Create calls for poisoning before initializers run and unpoisoning after.
699 if (ClInitializers && FirstDynamic && LastDynamic)
700 createInitializerPoisonCalls(M, FirstDynamic, LastDynamic);
701
Alexander Potapenko55cabae2012-04-23 10:47:31 +0000702 Function *AsanRegisterGlobals = checkInterfaceFunction(M.getOrInsertFunction(
Kostya Serebryany9b9f87a2012-08-21 08:24:25 +0000703 kAsanRegisterGlobalsName, IRB.getVoidTy(),
704 IntptrTy, IntptrTy, NULL));
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000705 AsanRegisterGlobals->setLinkage(Function::ExternalLinkage);
706
707 IRB.CreateCall2(AsanRegisterGlobals,
708 IRB.CreatePointerCast(AllGlobals, IntptrTy),
709 ConstantInt::get(IntptrTy, n));
710
Kostya Serebryany7bcfc992011-12-15 21:59:03 +0000711 // We also need to unregister globals at the end, e.g. when a shared library
712 // gets closed.
713 Function *AsanDtorFunction = Function::Create(
714 FunctionType::get(Type::getVoidTy(*C), false),
715 GlobalValue::InternalLinkage, kAsanModuleDtorName, &M);
716 BasicBlock *AsanDtorBB = BasicBlock::Create(*C, "", AsanDtorFunction);
717 IRBuilder<> IRB_Dtor(ReturnInst::Create(*C, AsanDtorBB));
Alexander Potapenko55cabae2012-04-23 10:47:31 +0000718 Function *AsanUnregisterGlobals =
719 checkInterfaceFunction(M.getOrInsertFunction(
720 kAsanUnregisterGlobalsName,
721 IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
Kostya Serebryany7bcfc992011-12-15 21:59:03 +0000722 AsanUnregisterGlobals->setLinkage(Function::ExternalLinkage);
723
724 IRB_Dtor.CreateCall2(AsanUnregisterGlobals,
725 IRB.CreatePointerCast(AllGlobals, IntptrTy),
726 ConstantInt::get(IntptrTy, n));
727 appendToGlobalDtors(M, AsanDtorFunction, kAsanCtorAndCtorPriority);
728
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000729 DEBUG(dbgs() << M);
730 return true;
731}
732
733// virtual
734bool AddressSanitizer::runOnModule(Module &M) {
735 // Initialize the private fields. No one has accessed them before.
736 TD = getAnalysisIfAvailable<TargetData>();
737 if (!TD)
738 return false;
Kostya Serebryanyb5b86d22012-08-24 16:44:47 +0000739 BL.reset(new BlackList(ClBlackListFile));
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000740
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000741 C = &(M.getContext());
742 LongSize = TD->getPointerSizeInBits();
743 IntptrTy = Type::getIntNTy(*C, LongSize);
744 IntptrPtrTy = PointerType::get(IntptrTy, 0);
745
746 AsanCtorFunction = Function::Create(
747 FunctionType::get(Type::getVoidTy(*C), false),
748 GlobalValue::InternalLinkage, kAsanModuleCtorName, &M);
749 BasicBlock *AsanCtorBB = BasicBlock::Create(*C, "", AsanCtorFunction);
750 CtorInsertBefore = ReturnInst::Create(*C, AsanCtorBB);
751
752 // call __asan_init in the module ctor.
753 IRBuilder<> IRB(CtorInsertBefore);
Alexander Potapenko55cabae2012-04-23 10:47:31 +0000754 AsanInitFunction = checkInterfaceFunction(
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000755 M.getOrInsertFunction(kAsanInitName, IRB.getVoidTy(), NULL));
756 AsanInitFunction->setLinkage(Function::ExternalLinkage);
757 IRB.CreateCall(AsanInitFunction);
758
Kostya Serebryany9db5b5f2012-07-16 14:09:42 +0000759 // Create __asan_report* callbacks.
760 for (size_t AccessIsWrite = 0; AccessIsWrite <= 1; AccessIsWrite++) {
761 for (size_t AccessSizeIndex = 0; AccessSizeIndex < kNumberOfAccessSizes;
762 AccessSizeIndex++) {
763 // IsWrite and TypeSize are encoded in the function name.
764 std::string FunctionName = std::string(kAsanReportErrorTemplate) +
765 (AccessIsWrite ? "store" : "load") + itostr(1 << AccessSizeIndex);
Kostya Serebryany4f0c6962012-07-17 11:04:12 +0000766 // If we are merging crash callbacks, they have two parameters.
Kostya Serebryany11c2a472012-08-13 14:08:46 +0000767 AsanErrorCallback[AccessIsWrite][AccessSizeIndex] = cast<Function>(
Kostya Serebryany4f0c6962012-07-17 11:04:12 +0000768 M.getOrInsertFunction(FunctionName, IRB.getVoidTy(), IntptrTy, NULL));
Kostya Serebryany9db5b5f2012-07-16 14:09:42 +0000769 }
770 }
Kostya Serebryanyf7b08222012-07-20 09:54:50 +0000771 // We insert an empty inline asm after __asan_report* to avoid callback merge.
772 EmptyAsm = InlineAsm::get(FunctionType::get(IRB.getVoidTy(), false),
773 StringRef(""), StringRef(""),
774 /*hasSideEffects=*/true);
Kostya Serebryany9db5b5f2012-07-16 14:09:42 +0000775
Evgeniy Stepanov06fdbaa2012-05-23 11:52:12 +0000776 llvm::Triple targetTriple(M.getTargetTriple());
777 bool isAndroid = targetTriple.getEnvironment() == llvm::Triple::ANDROIDEABI;
778
779 MappingOffset = isAndroid ? kDefaultShadowOffsetAndroid :
780 (LongSize == 32 ? kDefaultShadowOffset32 : kDefaultShadowOffset64);
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000781 if (ClMappingOffsetLog >= 0) {
782 if (ClMappingOffsetLog == 0) {
783 // special case
784 MappingOffset = 0;
785 } else {
786 MappingOffset = 1ULL << ClMappingOffsetLog;
787 }
788 }
789 MappingScale = kDefaultShadowScale;
790 if (ClMappingScale) {
791 MappingScale = ClMappingScale;
792 }
793 // Redzone used for stack and globals is at least 32 bytes.
794 // For scales 6 and 7, the redzone has to be 64 and 128 bytes respectively.
795 RedzoneSize = std::max(32, (int)(1 << MappingScale));
796
797 bool Res = false;
798
799 if (ClGlobals)
800 Res |= insertGlobalRedzones(M);
801
Kostya Serebryany8c0134a2012-03-19 16:40:35 +0000802 if (ClMappingOffsetLog >= 0) {
803 // Tell the run-time the current values of mapping offset and scale.
804 GlobalValue *asan_mapping_offset =
805 new GlobalVariable(M, IntptrTy, true, GlobalValue::LinkOnceODRLinkage,
806 ConstantInt::get(IntptrTy, MappingOffset),
807 kAsanMappingOffsetName);
808 // Read the global, otherwise it may be optimized away.
809 IRB.CreateLoad(asan_mapping_offset, true);
810 }
811 if (ClMappingScale) {
812 GlobalValue *asan_mapping_scale =
813 new GlobalVariable(M, IntptrTy, true, GlobalValue::LinkOnceODRLinkage,
814 ConstantInt::get(IntptrTy, MappingScale),
815 kAsanMappingScaleName);
816 // Read the global, otherwise it may be optimized away.
817 IRB.CreateLoad(asan_mapping_scale, true);
818 }
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000819
820
821 for (Module::iterator F = M.begin(), E = M.end(); F != E; ++F) {
822 if (F->isDeclaration()) continue;
823 Res |= handleFunction(M, *F);
824 }
825
Kostya Serebryany7bcfc992011-12-15 21:59:03 +0000826 appendToGlobalCtors(M, AsanCtorFunction, kAsanCtorAndCtorPriority);
Kostya Serebryany9b027412011-12-12 18:01:46 +0000827
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000828 return Res;
829}
830
Kostya Serebryanya1a8a322012-01-30 23:50:10 +0000831bool AddressSanitizer::maybeInsertAsanInitAtFunctionEntry(Function &F) {
832 // For each NSObject descendant having a +load method, this method is invoked
833 // by the ObjC runtime before any of the static constructors is called.
834 // Therefore we need to instrument such methods with a call to __asan_init
835 // at the beginning in order to initialize our runtime before any access to
836 // the shadow memory.
837 // We cannot just ignore these methods, because they may call other
838 // instrumented functions.
839 if (F.getName().find(" load]") != std::string::npos) {
840 IRBuilder<> IRB(F.begin()->begin());
841 IRB.CreateCall(AsanInitFunction);
842 return true;
843 }
844 return false;
845}
846
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000847bool AddressSanitizer::handleFunction(Module &M, Function &F) {
848 if (BL->isIn(F)) return false;
849 if (&F == AsanCtorFunction) return false;
Kostya Serebryanya1a8a322012-01-30 23:50:10 +0000850
851 // If needed, insert __asan_init before checking for AddressSafety attr.
852 maybeInsertAsanInitAtFunctionEntry(F);
853
Kostya Serebryany0307b9a2012-01-24 19:34:43 +0000854 if (!F.hasFnAttr(Attribute::AddressSafety)) return false;
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000855
856 if (!ClDebugFunc.empty() && ClDebugFunc != F.getName())
857 return false;
858 // We want to instrument every address only once per basic block
859 // (unless there are calls between uses).
860 SmallSet<Value*, 16> TempsToInstrument;
861 SmallVector<Instruction*, 16> ToInstrument;
Kostya Serebryany95e3cf42012-02-08 21:36:17 +0000862 SmallVector<Instruction*, 8> NoReturnCalls;
Kostya Serebryanye6cf2e02012-05-30 09:04:06 +0000863 bool IsWrite;
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000864
865 // Fill the set of memory operations to instrument.
866 for (Function::iterator FI = F.begin(), FE = F.end();
867 FI != FE; ++FI) {
868 TempsToInstrument.clear();
Kostya Serebryany324cbb82012-06-28 09:34:41 +0000869 int NumInsnsPerBB = 0;
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000870 for (BasicBlock::iterator BI = FI->begin(), BE = FI->end();
871 BI != BE; ++BI) {
Kostya Serebryanybcb55ce2012-01-11 18:15:23 +0000872 if (LooksLikeCodeInBug11395(BI)) return false;
Kostya Serebryanye6cf2e02012-05-30 09:04:06 +0000873 if (Value *Addr = isInterestingMemoryAccess(BI, &IsWrite)) {
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000874 if (ClOpt && ClOptSameTemp) {
875 if (!TempsToInstrument.insert(Addr))
876 continue; // We've seen this temp in the current BB.
877 }
878 } else if (isa<MemIntrinsic>(BI) && ClMemIntrin) {
879 // ok, take it.
880 } else {
Kostya Serebryany95e3cf42012-02-08 21:36:17 +0000881 if (CallInst *CI = dyn_cast<CallInst>(BI)) {
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000882 // A call inside BB.
883 TempsToInstrument.clear();
Kostya Serebryany95e3cf42012-02-08 21:36:17 +0000884 if (CI->doesNotReturn()) {
885 NoReturnCalls.push_back(CI);
886 }
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000887 }
888 continue;
889 }
890 ToInstrument.push_back(BI);
Kostya Serebryany324cbb82012-06-28 09:34:41 +0000891 NumInsnsPerBB++;
892 if (NumInsnsPerBB >= ClMaxInsnsToInstrumentPerBB)
893 break;
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000894 }
895 }
896
Kostya Serebryany2735cf42012-07-16 17:12:07 +0000897 AsanFunctionContext AFC(F);
Kostya Serebryanyc0ed3e52012-07-16 16:15:40 +0000898
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000899 // Instrument.
900 int NumInstrumented = 0;
901 for (size_t i = 0, n = ToInstrument.size(); i != n; i++) {
902 Instruction *Inst = ToInstrument[i];
903 if (ClDebugMin < 0 || ClDebugMax < 0 ||
904 (NumInstrumented >= ClDebugMin && NumInstrumented <= ClDebugMax)) {
Kostya Serebryanye6cf2e02012-05-30 09:04:06 +0000905 if (isInterestingMemoryAccess(Inst, &IsWrite))
Kostya Serebryanyc0ed3e52012-07-16 16:15:40 +0000906 instrumentMop(AFC, Inst);
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000907 else
Kostya Serebryanyc0ed3e52012-07-16 16:15:40 +0000908 instrumentMemIntrinsic(AFC, cast<MemIntrinsic>(Inst));
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000909 }
910 NumInstrumented++;
911 }
912
913 DEBUG(dbgs() << F);
914
915 bool ChangedStack = poisonStackInFunction(M, F);
Kostya Serebryany95e3cf42012-02-08 21:36:17 +0000916
917 // We must unpoison the stack before every NoReturn call (throw, _exit, etc).
918 // See e.g. http://code.google.com/p/address-sanitizer/issues/detail?id=37
919 for (size_t i = 0, n = NoReturnCalls.size(); i != n; i++) {
920 Instruction *CI = NoReturnCalls[i];
921 IRBuilder<> IRB(CI);
922 IRB.CreateCall(M.getOrInsertFunction(kAsanHandleNoReturnName,
923 IRB.getVoidTy(), NULL));
924 }
925
926 return NumInstrumented > 0 || ChangedStack || !NoReturnCalls.empty();
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000927}
928
929static uint64_t ValueForPoison(uint64_t PoisonByte, size_t ShadowRedzoneSize) {
930 if (ShadowRedzoneSize == 1) return PoisonByte;
931 if (ShadowRedzoneSize == 2) return (PoisonByte << 8) + PoisonByte;
932 if (ShadowRedzoneSize == 4)
933 return (PoisonByte << 24) + (PoisonByte << 16) +
934 (PoisonByte << 8) + (PoisonByte);
Craig Topper85814382012-02-07 05:05:23 +0000935 llvm_unreachable("ShadowRedzoneSize is either 1, 2 or 4");
Kostya Serebryany800e03f2011-11-16 01:35:23 +0000936}
937
938static void PoisonShadowPartialRightRedzone(uint8_t *Shadow,
939 size_t Size,
940 size_t RedzoneSize,
941 size_t ShadowGranularity,
942 uint8_t Magic) {
943 for (size_t i = 0; i < RedzoneSize;
944 i+= ShadowGranularity, Shadow++) {
945 if (i + ShadowGranularity <= Size) {
946 *Shadow = 0; // fully addressable
947 } else if (i >= Size) {
948 *Shadow = Magic; // unaddressable
949 } else {
950 *Shadow = Size - i; // first Size-i bytes are addressable
951 }
952 }
953}
954
955void AddressSanitizer::PoisonStack(const ArrayRef<AllocaInst*> &AllocaVec,
956 IRBuilder<> IRB,
957 Value *ShadowBase, bool DoPoison) {
958 size_t ShadowRZSize = RedzoneSize >> MappingScale;
959 assert(ShadowRZSize >= 1 && ShadowRZSize <= 4);
960 Type *RZTy = Type::getIntNTy(*C, ShadowRZSize * 8);
961 Type *RZPtrTy = PointerType::get(RZTy, 0);
962
963 Value *PoisonLeft = ConstantInt::get(RZTy,
964 ValueForPoison(DoPoison ? kAsanStackLeftRedzoneMagic : 0LL, ShadowRZSize));
965 Value *PoisonMid = ConstantInt::get(RZTy,
966 ValueForPoison(DoPoison ? kAsanStackMidRedzoneMagic : 0LL, ShadowRZSize));
967 Value *PoisonRight = ConstantInt::get(RZTy,
968 ValueForPoison(DoPoison ? kAsanStackRightRedzoneMagic : 0LL, ShadowRZSize));
969
970 // poison the first red zone.
971 IRB.CreateStore(PoisonLeft, IRB.CreateIntToPtr(ShadowBase, RZPtrTy));
972
973 // poison all other red zones.
974 uint64_t Pos = RedzoneSize;
975 for (size_t i = 0, n = AllocaVec.size(); i < n; i++) {
976 AllocaInst *AI = AllocaVec[i];
977 uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
978 uint64_t AlignedSize = getAlignedAllocaSize(AI);
979 assert(AlignedSize - SizeInBytes < RedzoneSize);
980 Value *Ptr = NULL;
981
982 Pos += AlignedSize;
983
984 assert(ShadowBase->getType() == IntptrTy);
985 if (SizeInBytes < AlignedSize) {
986 // Poison the partial redzone at right
987 Ptr = IRB.CreateAdd(
988 ShadowBase, ConstantInt::get(IntptrTy,
989 (Pos >> MappingScale) - ShadowRZSize));
990 size_t AddressableBytes = RedzoneSize - (AlignedSize - SizeInBytes);
991 uint32_t Poison = 0;
992 if (DoPoison) {
993 PoisonShadowPartialRightRedzone((uint8_t*)&Poison, AddressableBytes,
994 RedzoneSize,
995 1ULL << MappingScale,
996 kAsanStackPartialRedzoneMagic);
997 }
998 Value *PartialPoison = ConstantInt::get(RZTy, Poison);
999 IRB.CreateStore(PartialPoison, IRB.CreateIntToPtr(Ptr, RZPtrTy));
1000 }
1001
1002 // Poison the full redzone at right.
1003 Ptr = IRB.CreateAdd(ShadowBase,
1004 ConstantInt::get(IntptrTy, Pos >> MappingScale));
1005 Value *Poison = i == AllocaVec.size() - 1 ? PoisonRight : PoisonMid;
1006 IRB.CreateStore(Poison, IRB.CreateIntToPtr(Ptr, RZPtrTy));
1007
1008 Pos += RedzoneSize;
1009 }
1010}
1011
Kostya Serebryany5a3a9c92011-11-18 01:41:06 +00001012// Workaround for bug 11395: we don't want to instrument stack in functions
1013// with large assembly blobs (32-bit only), otherwise reg alloc may crash.
Kostya Serebryanyd2703de2011-11-23 02:10:54 +00001014// FIXME: remove once the bug 11395 is fixed.
Kostya Serebryany5a3a9c92011-11-18 01:41:06 +00001015bool AddressSanitizer::LooksLikeCodeInBug11395(Instruction *I) {
1016 if (LongSize != 32) return false;
1017 CallInst *CI = dyn_cast<CallInst>(I);
1018 if (!CI || !CI->isInlineAsm()) return false;
1019 if (CI->getNumArgOperands() <= 5) return false;
1020 // We have inline assembly with quite a few arguments.
1021 return true;
1022}
1023
Kostya Serebryany800e03f2011-11-16 01:35:23 +00001024// Find all static Alloca instructions and put
1025// poisoned red zones around all of them.
1026// Then unpoison everything back before the function returns.
1027//
1028// Stack poisoning does not play well with exception handling.
1029// When an exception is thrown, we essentially bypass the code
1030// that unpoisones the stack. This is why the run-time library has
1031// to intercept __cxa_throw (as well as longjmp, etc) and unpoison the entire
1032// stack in the interceptor. This however does not work inside the
1033// actual function which catches the exception. Most likely because the
1034// compiler hoists the load of the shadow value somewhere too high.
1035// This causes asan to report a non-existing bug on 453.povray.
1036// It sounds like an LLVM bug.
1037bool AddressSanitizer::poisonStackInFunction(Module &M, Function &F) {
1038 if (!ClStack) return false;
1039 SmallVector<AllocaInst*, 16> AllocaVec;
1040 SmallVector<Instruction*, 8> RetVec;
1041 uint64_t TotalSize = 0;
1042
1043 // Filter out Alloca instructions we want (and can) handle.
1044 // Collect Ret instructions.
1045 for (Function::iterator FI = F.begin(), FE = F.end();
1046 FI != FE; ++FI) {
1047 BasicBlock &BB = *FI;
1048 for (BasicBlock::iterator BI = BB.begin(), BE = BB.end();
1049 BI != BE; ++BI) {
Kostya Serebryany800e03f2011-11-16 01:35:23 +00001050 if (isa<ReturnInst>(BI)) {
1051 RetVec.push_back(BI);
1052 continue;
1053 }
1054
1055 AllocaInst *AI = dyn_cast<AllocaInst>(BI);
1056 if (!AI) continue;
1057 if (AI->isArrayAllocation()) continue;
1058 if (!AI->isStaticAlloca()) continue;
1059 if (!AI->getAllocatedType()->isSized()) continue;
1060 if (AI->getAlignment() > RedzoneSize) continue;
1061 AllocaVec.push_back(AI);
1062 uint64_t AlignedSize = getAlignedAllocaSize(AI);
1063 TotalSize += AlignedSize;
1064 }
1065 }
1066
1067 if (AllocaVec.empty()) return false;
1068
1069 uint64_t LocalStackSize = TotalSize + (AllocaVec.size() + 1) * RedzoneSize;
1070
1071 bool DoStackMalloc = ClUseAfterReturn
1072 && LocalStackSize <= kMaxStackMallocSize;
1073
1074 Instruction *InsBefore = AllocaVec[0];
1075 IRBuilder<> IRB(InsBefore);
1076
1077
1078 Type *ByteArrayTy = ArrayType::get(IRB.getInt8Ty(), LocalStackSize);
1079 AllocaInst *MyAlloca =
1080 new AllocaInst(ByteArrayTy, "MyAlloca", InsBefore);
1081 MyAlloca->setAlignment(RedzoneSize);
1082 assert(MyAlloca->isStaticAlloca());
1083 Value *OrigStackBase = IRB.CreatePointerCast(MyAlloca, IntptrTy);
1084 Value *LocalStackBase = OrigStackBase;
1085
1086 if (DoStackMalloc) {
1087 Value *AsanStackMallocFunc = M.getOrInsertFunction(
1088 kAsanStackMallocName, IntptrTy, IntptrTy, IntptrTy, NULL);
1089 LocalStackBase = IRB.CreateCall2(AsanStackMallocFunc,
1090 ConstantInt::get(IntptrTy, LocalStackSize), OrigStackBase);
1091 }
1092
1093 // This string will be parsed by the run-time (DescribeStackAddress).
1094 SmallString<2048> StackDescriptionStorage;
1095 raw_svector_ostream StackDescription(StackDescriptionStorage);
1096 StackDescription << F.getName() << " " << AllocaVec.size() << " ";
1097
1098 uint64_t Pos = RedzoneSize;
1099 // Replace Alloca instructions with base+offset.
1100 for (size_t i = 0, n = AllocaVec.size(); i < n; i++) {
1101 AllocaInst *AI = AllocaVec[i];
1102 uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
1103 StringRef Name = AI->getName();
1104 StackDescription << Pos << " " << SizeInBytes << " "
1105 << Name.size() << " " << Name << " ";
1106 uint64_t AlignedSize = getAlignedAllocaSize(AI);
1107 assert((AlignedSize % RedzoneSize) == 0);
1108 AI->replaceAllUsesWith(
1109 IRB.CreateIntToPtr(
1110 IRB.CreateAdd(LocalStackBase, ConstantInt::get(IntptrTy, Pos)),
1111 AI->getType()));
1112 Pos += AlignedSize + RedzoneSize;
1113 }
1114 assert(Pos == LocalStackSize);
1115
1116 // Write the Magic value and the frame description constant to the redzone.
1117 Value *BasePlus0 = IRB.CreateIntToPtr(LocalStackBase, IntptrPtrTy);
1118 IRB.CreateStore(ConstantInt::get(IntptrTy, kCurrentStackFrameMagic),
1119 BasePlus0);
1120 Value *BasePlus1 = IRB.CreateAdd(LocalStackBase,
1121 ConstantInt::get(IntptrTy, LongSize/8));
1122 BasePlus1 = IRB.CreateIntToPtr(BasePlus1, IntptrPtrTy);
1123 Value *Description = IRB.CreatePointerCast(
1124 createPrivateGlobalForString(M, StackDescription.str()),
1125 IntptrTy);
1126 IRB.CreateStore(Description, BasePlus1);
1127
1128 // Poison the stack redzones at the entry.
1129 Value *ShadowBase = memToShadow(LocalStackBase, IRB);
1130 PoisonStack(ArrayRef<AllocaInst*>(AllocaVec), IRB, ShadowBase, true);
1131
1132 Value *AsanStackFreeFunc = NULL;
1133 if (DoStackMalloc) {
1134 AsanStackFreeFunc = M.getOrInsertFunction(
1135 kAsanStackFreeName, IRB.getVoidTy(),
1136 IntptrTy, IntptrTy, IntptrTy, NULL);
1137 }
1138
1139 // Unpoison the stack before all ret instructions.
1140 for (size_t i = 0, n = RetVec.size(); i < n; i++) {
1141 Instruction *Ret = RetVec[i];
1142 IRBuilder<> IRBRet(Ret);
1143
1144 // Mark the current frame as retired.
1145 IRBRet.CreateStore(ConstantInt::get(IntptrTy, kRetiredStackFrameMagic),
1146 BasePlus0);
1147 // Unpoison the stack.
1148 PoisonStack(ArrayRef<AllocaInst*>(AllocaVec), IRBRet, ShadowBase, false);
1149
1150 if (DoStackMalloc) {
1151 IRBRet.CreateCall3(AsanStackFreeFunc, LocalStackBase,
1152 ConstantInt::get(IntptrTy, LocalStackSize),
1153 OrigStackBase);
1154 }
1155 }
1156
1157 if (ClDebugStack) {
1158 DEBUG(dbgs() << F);
1159 }
1160
1161 return true;
1162}