//
//===----------------------------------------------------------------------===//
-#define DEBUG_TYPE "asan"
-
-#include "FunctionBlackList.h"
-#include "llvm/Function.h"
-#include "llvm/IRBuilder.h"
-#include "llvm/InlineAsm.h"
-#include "llvm/IntrinsicInst.h"
-#include "llvm/LLVMContext.h"
-#include "llvm/Module.h"
-#include "llvm/Type.h"
+#include "llvm/Transforms/Instrumentation.h"
#include "llvm/ADT/ArrayRef.h"
-#include "llvm/ADT/OwningPtr.h"
+#include "llvm/ADT/DenseMap.h"
+#include "llvm/ADT/DenseSet.h"
+#include "llvm/ADT/DepthFirstIterator.h"
#include "llvm/ADT/SmallSet.h"
#include "llvm/ADT/SmallString.h"
#include "llvm/ADT/SmallVector.h"
+#include "llvm/ADT/Statistic.h"
#include "llvm/ADT/StringExtras.h"
#include "llvm/ADT/Triple.h"
+#include "llvm/IR/CallSite.h"
+#include "llvm/IR/DIBuilder.h"
+#include "llvm/IR/DataLayout.h"
+#include "llvm/IR/Function.h"
+#include "llvm/IR/IRBuilder.h"
+#include "llvm/IR/InlineAsm.h"
+#include "llvm/IR/InstVisitor.h"
+#include "llvm/IR/IntrinsicInst.h"
+#include "llvm/IR/LLVMContext.h"
+#include "llvm/IR/MDBuilder.h"
+#include "llvm/IR/Module.h"
+#include "llvm/IR/Type.h"
#include "llvm/Support/CommandLine.h"
#include "llvm/Support/DataTypes.h"
#include "llvm/Support/Debug.h"
-#include "llvm/Support/raw_ostream.h"
-#include "llvm/Support/system_error.h"
-#include "llvm/Target/TargetData.h"
-#include "llvm/Target/TargetMachine.h"
-#include "llvm/Transforms/Instrumentation.h"
+#include "llvm/Support/Endian.h"
+#include "llvm/Transforms/Scalar.h"
+#include "llvm/Transforms/Utils/ASanStackFrameLayout.h"
#include "llvm/Transforms/Utils/BasicBlockUtils.h"
+#include "llvm/Transforms/Utils/Cloning.h"
+#include "llvm/Transforms/Utils/Local.h"
#include "llvm/Transforms/Utils/ModuleUtils.h"
-
-#include <string>
#include <algorithm>
+#include <string>
+#include <system_error>
using namespace llvm;
+#define DEBUG_TYPE "asan"
+
static const uint64_t kDefaultShadowScale = 3;
static const uint64_t kDefaultShadowOffset32 = 1ULL << 29;
+static const uint64_t kIOSShadowOffset32 = 1ULL << 30;
static const uint64_t kDefaultShadowOffset64 = 1ULL << 44;
-static const uint64_t kDefaultShadowOffsetAndroid = 0;
+static const uint64_t kSmallX86_64ShadowOffset = 0x7FFF8000; // < 2G.
+static const uint64_t kPPC64_ShadowOffset64 = 1ULL << 41;
+static const uint64_t kMIPS32_ShadowOffset32 = 0x0aaa8000;
+static const uint64_t kFreeBSD_ShadowOffset32 = 1ULL << 30;
+static const uint64_t kFreeBSD_ShadowOffset64 = 1ULL << 46;
+static const size_t kMinStackMallocSize = 1 << 6; // 64B
static const size_t kMaxStackMallocSize = 1 << 16; // 64K
static const uintptr_t kCurrentStackFrameMagic = 0x41B58AB3;
static const uintptr_t kRetiredStackFrameMagic = 0x45E0360E;
-static const char *kAsanModuleCtorName = "asan.module_ctor";
-static const char *kAsanModuleDtorName = "asan.module_dtor";
-static const int kAsanCtorAndCtorPriority = 1;
-static const char *kAsanReportErrorTemplate = "__asan_report_";
-static const char *kAsanRegisterGlobalsName = "__asan_register_globals";
-static const char *kAsanUnregisterGlobalsName = "__asan_unregister_globals";
-static const char *kAsanInitName = "__asan_init";
-static const char *kAsanHandleNoReturnName = "__asan_handle_no_return";
-static const char *kAsanMappingOffsetName = "__asan_mapping_offset";
-static const char *kAsanMappingScaleName = "__asan_mapping_scale";
-static const char *kAsanStackMallocName = "__asan_stack_malloc";
-static const char *kAsanStackFreeName = "__asan_stack_free";
-
-static const int kAsanStackLeftRedzoneMagic = 0xf1;
-static const int kAsanStackMidRedzoneMagic = 0xf2;
-static const int kAsanStackRightRedzoneMagic = 0xf3;
-static const int kAsanStackPartialRedzoneMagic = 0xf4;
+static const char *const kAsanModuleCtorName = "asan.module_ctor";
+static const char *const kAsanModuleDtorName = "asan.module_dtor";
+static const uint64_t kAsanCtorAndDtorPriority = 1;
+static const char *const kAsanReportErrorTemplate = "__asan_report_";
+static const char *const kAsanReportLoadN = "__asan_report_load_n";
+static const char *const kAsanReportStoreN = "__asan_report_store_n";
+static const char *const kAsanRegisterGlobalsName = "__asan_register_globals";
+static const char *const kAsanUnregisterGlobalsName =
+ "__asan_unregister_globals";
+static const char *const kAsanPoisonGlobalsName = "__asan_before_dynamic_init";
+static const char *const kAsanUnpoisonGlobalsName = "__asan_after_dynamic_init";
+static const char *const kAsanInitName = "__asan_init_v4";
+static const char *const kAsanCovModuleInitName = "__sanitizer_cov_module_init";
+static const char *const kAsanCovName = "__sanitizer_cov";
+static const char *const kAsanCovIndirCallName = "__sanitizer_cov_indir_call16";
+static const char *const kAsanPtrCmp = "__sanitizer_ptr_cmp";
+static const char *const kAsanPtrSub = "__sanitizer_ptr_sub";
+static const char *const kAsanHandleNoReturnName = "__asan_handle_no_return";
+static const int kMaxAsanStackMallocSizeClass = 10;
+static const char *const kAsanStackMallocNameTemplate = "__asan_stack_malloc_";
+static const char *const kAsanStackFreeNameTemplate = "__asan_stack_free_";
+static const char *const kAsanGenPrefix = "__asan_gen_";
+static const char *const kAsanPoisonStackMemoryName =
+ "__asan_poison_stack_memory";
+static const char *const kAsanUnpoisonStackMemoryName =
+ "__asan_unpoison_stack_memory";
+
+static const char *const kAsanOptionDetectUAR =
+ "__asan_option_detect_stack_use_after_return";
+
+#ifndef NDEBUG
+static const int kAsanStackAfterReturnMagic = 0xf5;
+#endif
// Accesses sizes are powers of two: 1, 2, 4, 8, 16.
static const size_t kNumberOfAccessSizes = 5;
static cl::opt<bool> ClInstrumentAtomics("asan-instrument-atomics",
cl::desc("instrument atomic instructions (rmw, cmpxchg)"),
cl::Hidden, cl::init(true));
-static cl::opt<bool> ClMergeCallbacks("asan-merge-callbacks",
- cl::desc("merge __asan_report_ callbacks to create fewer BBs"),
+static cl::opt<bool> ClAlwaysSlowPath("asan-always-slow-path",
+ cl::desc("use instrumentation with slow path for all accesses"),
cl::Hidden, cl::init(false));
// This flag limits the number of instructions to be instrumented
// in any given BB. Normally, this should be set to unlimited (INT_MAX),
// This flag may need to be replaced with -f[no]asan-stack.
static cl::opt<bool> ClStack("asan-stack",
cl::desc("Handle stack memory"), cl::Hidden, cl::init(true));
-// This flag may need to be replaced with -f[no]asan-use-after-return.
static cl::opt<bool> ClUseAfterReturn("asan-use-after-return",
- cl::desc("Check return-after-free"), cl::Hidden, cl::init(false));
+ cl::desc("Check return-after-free"), cl::Hidden, cl::init(true));
// This flag may need to be replaced with -f[no]asan-globals.
static cl::opt<bool> ClGlobals("asan-globals",
cl::desc("Handle global objects"), cl::Hidden, cl::init(true));
-static cl::opt<bool> ClMemIntrin("asan-memintrin",
- cl::desc("Handle memset/memcpy/memmove"), cl::Hidden, cl::init(true));
-// This flag may need to be replaced with -fasan-blacklist.
-static cl::opt<std::string> ClBlackListFile("asan-blacklist",
- cl::desc("File containing the list of functions to ignore "
- "during instrumentation"), cl::Hidden);
+static cl::opt<int> ClCoverage("asan-coverage",
+ cl::desc("ASan coverage. 0: none, 1: entry block, 2: all blocks, "
+ "3: all blocks and critical edges, "
+ "4: above plus indirect calls"),
+ cl::Hidden, cl::init(false));
+static cl::opt<int> ClCoverageBlockThreshold("asan-coverage-block-threshold",
+ cl::desc("Add coverage instrumentation only to the entry block if there "
+ "are more than this number of blocks."),
+ cl::Hidden, cl::init(1500));
+static cl::opt<bool> ClInitializers("asan-initialization-order",
+ cl::desc("Handle C++ initializer order"), cl::Hidden, cl::init(true));
+static cl::opt<bool> ClInvalidPointerPairs("asan-detect-invalid-pointer-pair",
+ cl::desc("Instrument <, <=, >, >=, - with pointer operands"),
+ cl::Hidden, cl::init(false));
+static cl::opt<unsigned> ClRealignStack("asan-realign-stack",
+ cl::desc("Realign stack to the value of this flag (power of two)"),
+ cl::Hidden, cl::init(32));
+static cl::opt<int> ClInstrumentationWithCallsThreshold(
+ "asan-instrumentation-with-call-threshold",
+ cl::desc("If the function being instrumented contains more than "
+ "this number of memory accesses, use callbacks instead of "
+ "inline checks (-1 means never use callbacks)."),
+ cl::Hidden, cl::init(7000));
+static cl::opt<std::string> ClMemoryAccessCallbackPrefix(
+ "asan-memory-access-callback-prefix",
+ cl::desc("Prefix for memory access callbacks"), cl::Hidden,
+ cl::init("__asan_"));
+
+// This is an experimental feature that will allow to choose between
+// instrumented and non-instrumented code at link-time.
+// If this option is on, just before instrumenting a function we create its
+// clone; if the function is not changed by asan the clone is deleted.
+// If we end up with a clone, we put the instrumented function into a section
+// called "ASAN" and the uninstrumented function into a section called "NOASAN".
+//
+// This is still a prototype, we need to figure out a way to keep two copies of
+// a function so that the linker can easily choose one of them.
+static cl::opt<bool> ClKeepUninstrumented("asan-keep-uninstrumented-functions",
+ cl::desc("Keep uninstrumented copies of functions"),
+ cl::Hidden, cl::init(false));
// These flags allow to change the shadow mapping.
// The shadow mapping looks like
// Shadow = (Mem >> scale) + (1 << offset_log)
static cl::opt<int> ClMappingScale("asan-mapping-scale",
cl::desc("scale of asan shadow mapping"), cl::Hidden, cl::init(0));
-static cl::opt<int> ClMappingOffsetLog("asan-mapping-offset-log",
- cl::desc("offset of asan shadow mapping"), cl::Hidden, cl::init(-1));
// Optimization flags. Not user visible, used mostly for testing
// and benchmarking the tool.
static cl::opt<bool> ClOptGlobals("asan-opt-globals",
cl::desc("Don't instrument scalar globals"), cl::Hidden, cl::init(true));
+static cl::opt<bool> ClCheckLifetime("asan-check-lifetime",
+ cl::desc("Use llvm.lifetime intrinsics to insert extra checks"),
+ cl::Hidden, cl::init(false));
+
// Debug flags.
static cl::opt<int> ClDebug("asan-debug", cl::desc("debug"), cl::Hidden,
cl::init(0));
static cl::opt<int> ClDebugMax("asan-debug-max", cl::desc("Debug man inst"),
cl::Hidden, cl::init(-1));
-namespace {
+STATISTIC(NumInstrumentedReads, "Number of instrumented reads");
+STATISTIC(NumInstrumentedWrites, "Number of instrumented writes");
+STATISTIC(NumOptimizedAccessesToGlobalArray,
+ "Number of optimized accesses to global arrays");
+STATISTIC(NumOptimizedAccessesToGlobalVar,
+ "Number of optimized accesses to global vars");
-/// When the crash callbacks are merged, they receive some amount of arguments
-/// that are merged in a PHI node. This struct represents arguments from one
-/// call site.
-struct CrashArg {
- Value *Arg1;
- Value *Arg2;
+namespace {
+/// Frontend-provided metadata for source location.
+struct LocationMetadata {
+ StringRef Filename;
+ int LineNo;
+ int ColumnNo;
+
+ LocationMetadata() : Filename(), LineNo(0), ColumnNo(0) {}
+
+ bool empty() const { return Filename.empty(); }
+
+ void parse(MDNode *MDN) {
+ assert(MDN->getNumOperands() == 3);
+ MDString *MDFilename = cast<MDString>(MDN->getOperand(0));
+ Filename = MDFilename->getString();
+ LineNo = cast<ConstantInt>(MDN->getOperand(1))->getLimitedValue();
+ ColumnNo = cast<ConstantInt>(MDN->getOperand(2))->getLimitedValue();
+ }
};
-/// An object of this type is created while instrumenting every function.
-struct AsanFunctionContext {
- AsanFunctionContext(Function &Function) : F(Function), CrashBlock() { }
+/// Frontend-provided metadata for global variables.
+class GlobalsMetadata {
+ public:
+ struct Entry {
+ Entry()
+ : SourceLoc(), Name(), IsDynInit(false),
+ IsBlacklisted(false) {}
+ LocationMetadata SourceLoc;
+ StringRef Name;
+ bool IsDynInit;
+ bool IsBlacklisted;
+ };
+
+ GlobalsMetadata() : inited_(false) {}
+
+ void init(Module& M) {
+ assert(!inited_);
+ inited_ = true;
+ NamedMDNode *Globals = M.getNamedMetadata("llvm.asan.globals");
+ if (!Globals)
+ return;
+ for (auto MDN : Globals->operands()) {
+ // Metadata node contains the global and the fields of "Entry".
+ assert(MDN->getNumOperands() == 5);
+ Value *V = MDN->getOperand(0);
+ // The optimizer may optimize away a global entirely.
+ if (!V)
+ continue;
+ GlobalVariable *GV = cast<GlobalVariable>(V);
+ // We can already have an entry for GV if it was merged with another
+ // global.
+ Entry &E = Entries[GV];
+ if (Value *Loc = MDN->getOperand(1))
+ E.SourceLoc.parse(cast<MDNode>(Loc));
+ if (Value *Name = MDN->getOperand(2)) {
+ MDString *MDName = cast<MDString>(Name);
+ E.Name = MDName->getString();
+ }
+ ConstantInt *IsDynInit = cast<ConstantInt>(MDN->getOperand(3));
+ E.IsDynInit |= IsDynInit->isOne();
+ ConstantInt *IsBlacklisted = cast<ConstantInt>(MDN->getOperand(4));
+ E.IsBlacklisted |= IsBlacklisted->isOne();
+ }
+ }
- Function &F;
- // These are initially zero. If we require at least one call to
- // __asan_report_{read,write}{1,2,4,8,16}, an appropriate BB is created.
- BasicBlock *CrashBlock[2][kNumberOfAccessSizes];
- typedef SmallVector<CrashArg, 8> CrashArgsVec;
- CrashArgsVec CrashArgs[2][kNumberOfAccessSizes];
+ /// Returns metadata entry for a given global.
+ Entry get(GlobalVariable *G) const {
+ auto Pos = Entries.find(G);
+ return (Pos != Entries.end()) ? Pos->second : Entry();
+ }
+
+ private:
+ bool inited_;
+ DenseMap<GlobalVariable*, Entry> Entries;
+};
+
+/// This struct defines the shadow mapping using the rule:
+/// shadow = (mem >> Scale) ADD-or-OR Offset.
+struct ShadowMapping {
+ int Scale;
+ uint64_t Offset;
+ bool OrShadowOffset;
};
+static ShadowMapping getShadowMapping(const Module &M, int LongSize) {
+ llvm::Triple TargetTriple(M.getTargetTriple());
+ bool IsAndroid = TargetTriple.getEnvironment() == llvm::Triple::Android;
+ bool IsIOS = TargetTriple.isiOS();
+ bool IsFreeBSD = TargetTriple.getOS() == llvm::Triple::FreeBSD;
+ bool IsLinux = TargetTriple.getOS() == llvm::Triple::Linux;
+ bool IsPPC64 = TargetTriple.getArch() == llvm::Triple::ppc64 ||
+ TargetTriple.getArch() == llvm::Triple::ppc64le;
+ bool IsX86_64 = TargetTriple.getArch() == llvm::Triple::x86_64;
+ bool IsMIPS32 = TargetTriple.getArch() == llvm::Triple::mips ||
+ TargetTriple.getArch() == llvm::Triple::mipsel;
+
+ ShadowMapping Mapping;
+
+ if (LongSize == 32) {
+ if (IsAndroid)
+ Mapping.Offset = 0;
+ else if (IsMIPS32)
+ Mapping.Offset = kMIPS32_ShadowOffset32;
+ else if (IsFreeBSD)
+ Mapping.Offset = kFreeBSD_ShadowOffset32;
+ else if (IsIOS)
+ Mapping.Offset = kIOSShadowOffset32;
+ else
+ Mapping.Offset = kDefaultShadowOffset32;
+ } else { // LongSize == 64
+ if (IsPPC64)
+ Mapping.Offset = kPPC64_ShadowOffset64;
+ else if (IsFreeBSD)
+ Mapping.Offset = kFreeBSD_ShadowOffset64;
+ else if (IsLinux && IsX86_64)
+ Mapping.Offset = kSmallX86_64ShadowOffset;
+ else
+ Mapping.Offset = kDefaultShadowOffset64;
+ }
+
+ Mapping.Scale = kDefaultShadowScale;
+ if (ClMappingScale) {
+ Mapping.Scale = ClMappingScale;
+ }
+
+ // OR-ing shadow offset if more efficient (at least on x86) if the offset
+ // is a power of two, but on ppc64 we have to use add since the shadow
+ // offset is not necessary 1/8-th of the address space.
+ Mapping.OrShadowOffset = !IsPPC64 && !(Mapping.Offset & (Mapping.Offset - 1));
+
+ return Mapping;
+}
+
+static size_t RedzoneSizeForScale(int MappingScale) {
+ // Redzone used for stack and globals is at least 32 bytes.
+ // For scales 6 and 7, the redzone has to be 64 and 128 bytes respectively.
+ return std::max(32U, 1U << MappingScale);
+}
+
/// AddressSanitizer: instrument the code in module to find memory bugs.
-struct AddressSanitizer : public ModulePass {
- AddressSanitizer();
- virtual const char *getPassName() const;
- void instrumentMop(AsanFunctionContext &AFC, Instruction *I);
- void instrumentAddress(AsanFunctionContext &AFC,
- Instruction *OrigIns, IRBuilder<> &IRB,
- Value *Addr, uint32_t TypeSize, bool IsWrite);
+struct AddressSanitizer : public FunctionPass {
+ AddressSanitizer() : FunctionPass(ID) {
+ initializeBreakCriticalEdgesPass(*PassRegistry::getPassRegistry());
+ }
+ const char *getPassName() const override {
+ return "AddressSanitizerFunctionPass";
+ }
+ void instrumentMop(Instruction *I, bool UseCalls);
+ void instrumentPointerComparisonOrSubtraction(Instruction *I);
+ void instrumentAddress(Instruction *OrigIns, Instruction *InsertBefore,
+ Value *Addr, uint32_t TypeSize, bool IsWrite,
+ Value *SizeArgument, bool UseCalls);
Value *createSlowPathCmp(IRBuilder<> &IRB, Value *AddrLong,
Value *ShadowValue, uint32_t TypeSize);
- Instruction *generateCrashCode(BasicBlock *BB, Value *Addr, Value *PC,
- bool IsWrite, size_t AccessSizeIndex);
- bool instrumentMemIntrinsic(AsanFunctionContext &AFC, MemIntrinsic *MI);
- void instrumentMemIntrinsicParam(AsanFunctionContext &AFC,
- Instruction *OrigIns, Value *Addr,
- Value *Size,
- Instruction *InsertBefore, bool IsWrite);
+ Instruction *generateCrashCode(Instruction *InsertBefore, Value *Addr,
+ bool IsWrite, size_t AccessSizeIndex,
+ Value *SizeArgument);
+ void instrumentMemIntrinsic(MemIntrinsic *MI);
Value *memToShadow(Value *Shadow, IRBuilder<> &IRB);
- bool handleFunction(Module &M, Function &F);
+ bool runOnFunction(Function &F) override;
bool maybeInsertAsanInitAtFunctionEntry(Function &F);
- bool poisonStackInFunction(Module &M, Function &F);
- virtual bool runOnModule(Module &M);
- bool insertGlobalRedzones(Module &M);
+ bool doInitialization(Module &M) override;
static char ID; // Pass identification, replacement for typeid
- private:
-
- uint64_t getAllocaSizeInBytes(AllocaInst *AI) {
- Type *Ty = AI->getAllocatedType();
- uint64_t SizeInBytes = TD->getTypeAllocSize(Ty);
- return SizeInBytes;
- }
- uint64_t getAlignedSize(uint64_t SizeInBytes) {
- return ((SizeInBytes + RedzoneSize - 1)
- / RedzoneSize) * RedzoneSize;
- }
- uint64_t getAlignedAllocaSize(AllocaInst *AI) {
- uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
- return getAlignedSize(SizeInBytes);
+ void getAnalysisUsage(AnalysisUsage &AU) const override {
+ if (ClCoverage >= 3)
+ AU.addRequiredID(BreakCriticalEdgesID);
}
- Function *checkInterfaceFunction(Constant *FuncOrBitcast);
- void PoisonStack(const ArrayRef<AllocaInst*> &AllocaVec, IRBuilder<> IRB,
- Value *ShadowBase, bool DoPoison);
+ private:
+ void initializeCallbacks(Module &M);
+
bool LooksLikeCodeInBug11395(Instruction *I);
+ bool GlobalIsLinkerInitialized(GlobalVariable *G);
+ void InjectCoverageForIndirectCalls(Function &F,
+ ArrayRef<Instruction *> IndirCalls);
+ bool InjectCoverage(Function &F, ArrayRef<BasicBlock *> AllBlocks,
+ ArrayRef<Instruction *> IndirCalls);
+ void InjectCoverageAtBlock(Function &F, BasicBlock &BB);
LLVMContext *C;
- TargetData *TD;
- uint64_t MappingOffset;
- int MappingScale;
- size_t RedzoneSize;
+ const DataLayout *DL;
int LongSize;
Type *IntptrTy;
- Type *IntptrPtrTy;
+ ShadowMapping Mapping;
Function *AsanCtorFunction;
Function *AsanInitFunction;
- Instruction *CtorInsertBefore;
- OwningPtr<FunctionBlackList> BL;
+ Function *AsanHandleNoReturnFunc;
+ Function *AsanCovFunction;
+ Function *AsanCovIndirCallFunction;
+ Function *AsanPtrCmpFunction, *AsanPtrSubFunction;
// This array is indexed by AccessIsWrite and log2(AccessSize).
Function *AsanErrorCallback[2][kNumberOfAccessSizes];
+ Function *AsanMemoryAccessCallback[2][kNumberOfAccessSizes];
+ // This array is indexed by AccessIsWrite.
+ Function *AsanErrorCallbackSized[2],
+ *AsanMemoryAccessCallbackSized[2];
+ Function *AsanMemmove, *AsanMemcpy, *AsanMemset;
InlineAsm *EmptyAsm;
+ GlobalsMetadata GlobalsMD;
+
+ friend struct FunctionStackPoisoner;
+};
+
+class AddressSanitizerModule : public ModulePass {
+ public:
+ AddressSanitizerModule() : ModulePass(ID) {}
+ bool runOnModule(Module &M) override;
+ static char ID; // Pass identification, replacement for typeid
+ const char *getPassName() const override {
+ return "AddressSanitizerModule";
+ }
+
+ private:
+ void initializeCallbacks(Module &M);
+
+ bool InstrumentGlobals(IRBuilder<> &IRB, Module &M);
+ bool ShouldInstrumentGlobal(GlobalVariable *G);
+ void poisonOneInitializer(Function &GlobalInit, GlobalValue *ModuleName);
+ void createInitializerPoisonCalls(Module &M, GlobalValue *ModuleName);
+ size_t MinRedzoneSizeForGlobal() const {
+ return RedzoneSizeForScale(Mapping.Scale);
+ }
+
+ GlobalsMetadata GlobalsMD;
+ Type *IntptrTy;
+ LLVMContext *C;
+ const DataLayout *DL;
+ ShadowMapping Mapping;
+ Function *AsanPoisonGlobals;
+ Function *AsanUnpoisonGlobals;
+ Function *AsanRegisterGlobals;
+ Function *AsanUnregisterGlobals;
+ Function *AsanCovModuleInit;
+};
+
+// Stack poisoning does not play well with exception handling.
+// When an exception is thrown, we essentially bypass the code
+// that unpoisones the stack. This is why the run-time library has
+// to intercept __cxa_throw (as well as longjmp, etc) and unpoison the entire
+// stack in the interceptor. This however does not work inside the
+// actual function which catches the exception. Most likely because the
+// compiler hoists the load of the shadow value somewhere too high.
+// This causes asan to report a non-existing bug on 453.povray.
+// It sounds like an LLVM bug.
+struct FunctionStackPoisoner : public InstVisitor<FunctionStackPoisoner> {
+ Function &F;
+ AddressSanitizer &ASan;
+ DIBuilder DIB;
+ LLVMContext *C;
+ Type *IntptrTy;
+ Type *IntptrPtrTy;
+ ShadowMapping Mapping;
+
+ SmallVector<AllocaInst*, 16> AllocaVec;
+ SmallVector<Instruction*, 8> RetVec;
+ unsigned StackAlignment;
+
+ Function *AsanStackMallocFunc[kMaxAsanStackMallocSizeClass + 1],
+ *AsanStackFreeFunc[kMaxAsanStackMallocSizeClass + 1];
+ Function *AsanPoisonStackMemoryFunc, *AsanUnpoisonStackMemoryFunc;
+
+ // Stores a place and arguments of poisoning/unpoisoning call for alloca.
+ struct AllocaPoisonCall {
+ IntrinsicInst *InsBefore;
+ AllocaInst *AI;
+ uint64_t Size;
+ bool DoPoison;
+ };
+ SmallVector<AllocaPoisonCall, 8> AllocaPoisonCallVec;
+
+ // Maps Value to an AllocaInst from which the Value is originated.
+ typedef DenseMap<Value*, AllocaInst*> AllocaForValueMapTy;
+ AllocaForValueMapTy AllocaForValue;
+
+ FunctionStackPoisoner(Function &F, AddressSanitizer &ASan)
+ : F(F), ASan(ASan), DIB(*F.getParent()), C(ASan.C),
+ IntptrTy(ASan.IntptrTy), IntptrPtrTy(PointerType::get(IntptrTy, 0)),
+ Mapping(ASan.Mapping),
+ StackAlignment(1 << Mapping.Scale) {}
+
+ bool runOnFunction() {
+ if (!ClStack) return false;
+ // Collect alloca, ret, lifetime instructions etc.
+ for (BasicBlock *BB : depth_first(&F.getEntryBlock()))
+ visit(*BB);
+
+ if (AllocaVec.empty()) return false;
+
+ initializeCallbacks(*F.getParent());
+
+ poisonStack();
+
+ if (ClDebugStack) {
+ DEBUG(dbgs() << F);
+ }
+ return true;
+ }
+
+ // Finds all static Alloca instructions and puts
+ // poisoned red zones around all of them.
+ // Then unpoison everything back before the function returns.
+ void poisonStack();
+
+ // ----------------------- Visitors.
+ /// \brief Collect all Ret instructions.
+ void visitReturnInst(ReturnInst &RI) {
+ RetVec.push_back(&RI);
+ }
+
+ /// \brief Collect Alloca instructions we want (and can) handle.
+ void visitAllocaInst(AllocaInst &AI) {
+ if (!isInterestingAlloca(AI)) return;
+
+ StackAlignment = std::max(StackAlignment, AI.getAlignment());
+ AllocaVec.push_back(&AI);
+ }
+
+ /// \brief Collect lifetime intrinsic calls to check for use-after-scope
+ /// errors.
+ void visitIntrinsicInst(IntrinsicInst &II) {
+ if (!ClCheckLifetime) return;
+ Intrinsic::ID ID = II.getIntrinsicID();
+ if (ID != Intrinsic::lifetime_start &&
+ ID != Intrinsic::lifetime_end)
+ return;
+ // Found lifetime intrinsic, add ASan instrumentation if necessary.
+ ConstantInt *Size = dyn_cast<ConstantInt>(II.getArgOperand(0));
+ // If size argument is undefined, don't do anything.
+ if (Size->isMinusOne()) return;
+ // Check that size doesn't saturate uint64_t and can
+ // be stored in IntptrTy.
+ const uint64_t SizeValue = Size->getValue().getLimitedValue();
+ if (SizeValue == ~0ULL ||
+ !ConstantInt::isValueValidForType(IntptrTy, SizeValue))
+ return;
+ // Find alloca instruction that corresponds to llvm.lifetime argument.
+ AllocaInst *AI = findAllocaForValue(II.getArgOperand(1));
+ if (!AI) return;
+ bool DoPoison = (ID == Intrinsic::lifetime_end);
+ AllocaPoisonCall APC = {&II, AI, SizeValue, DoPoison};
+ AllocaPoisonCallVec.push_back(APC);
+ }
+
+ // ---------------------- Helpers.
+ void initializeCallbacks(Module &M);
+
+ // Check if we want (and can) handle this alloca.
+ bool isInterestingAlloca(AllocaInst &AI) const {
+ return (!AI.isArrayAllocation() && AI.isStaticAlloca() &&
+ AI.getAllocatedType()->isSized() &&
+ // alloca() may be called with 0 size, ignore it.
+ getAllocaSizeInBytes(&AI) > 0);
+ }
+
+ uint64_t getAllocaSizeInBytes(AllocaInst *AI) const {
+ Type *Ty = AI->getAllocatedType();
+ uint64_t SizeInBytes = ASan.DL->getTypeAllocSize(Ty);
+ return SizeInBytes;
+ }
+ /// Finds alloca where the value comes from.
+ AllocaInst *findAllocaForValue(Value *V);
+ void poisonRedZones(ArrayRef<uint8_t> ShadowBytes, IRBuilder<> &IRB,
+ Value *ShadowBase, bool DoPoison);
+ void poisonAlloca(Value *V, uint64_t Size, IRBuilder<> &IRB, bool DoPoison);
+
+ void SetShadowToStackAfterReturnInlined(IRBuilder<> &IRB, Value *ShadowBase,
+ int Size);
};
} // namespace
INITIALIZE_PASS(AddressSanitizer, "asan",
"AddressSanitizer: detects use-after-free and out-of-bounds bugs.",
false, false)
-AddressSanitizer::AddressSanitizer() : ModulePass(ID) { }
-ModulePass *llvm::createAddressSanitizerPass() {
+FunctionPass *llvm::createAddressSanitizerFunctionPass() {
return new AddressSanitizer();
}
-const char *AddressSanitizer::getPassName() const {
- return "AddressSanitizer";
+char AddressSanitizerModule::ID = 0;
+INITIALIZE_PASS(AddressSanitizerModule, "asan-module",
+ "AddressSanitizer: detects use-after-free and out-of-bounds bugs."
+ "ModulePass", false, false)
+ModulePass *llvm::createAddressSanitizerModulePass() {
+ return new AddressSanitizerModule();
}
static size_t TypeSizeToSizeIndex(uint32_t TypeSize) {
- size_t Res = CountTrailingZeros_32(TypeSize / 8);
+ size_t Res = countTrailingZeros(TypeSize / 8);
assert(Res < kNumberOfAccessSizes);
return Res;
}
-// Create a constant for Str so that we can pass it to the run-time lib.
-static GlobalVariable *createPrivateGlobalForString(Module &M, StringRef Str) {
+// \brief Create a constant for Str so that we can pass it to the run-time lib.
+static GlobalVariable *createPrivateGlobalForString(
+ Module &M, StringRef Str, bool AllowMerging) {
Constant *StrConst = ConstantDataArray::getString(M.getContext(), Str);
- return new GlobalVariable(M, StrConst->getType(), true,
- GlobalValue::PrivateLinkage, StrConst, "");
+ // We use private linkage for module-local strings. If they can be merged
+ // with another one, we set the unnamed_addr attribute.
+ GlobalVariable *GV =
+ new GlobalVariable(M, StrConst->getType(), true,
+ GlobalValue::PrivateLinkage, StrConst, kAsanGenPrefix);
+ if (AllowMerging)
+ GV->setUnnamedAddr(true);
+ GV->setAlignment(1); // Strings may not be merged w/o setting align 1.
+ return GV;
}
-// Split the basic block and insert an if-then code.
-// Before:
-// Head
-// Cmp
-// Tail
-// After:
-// Head
-// if (Cmp)
-// ThenBlock
-// Tail
-//
-// If ThenBlock is zero, a new block is created and its terminator is returned.
-// Otherwize 0 is returned.
-static BranchInst *splitBlockAndInsertIfThen(Value *Cmp,
- BasicBlock *ThenBlock = 0) {
- Instruction *SplitBefore = cast<Instruction>(Cmp)->getNextNode();
- BasicBlock *Head = SplitBefore->getParent();
- BasicBlock *Tail = Head->splitBasicBlock(SplitBefore);
- TerminatorInst *HeadOldTerm = Head->getTerminator();
- BranchInst *CheckTerm = 0;
- if (!ThenBlock) {
- LLVMContext &C = Head->getParent()->getParent()->getContext();
- ThenBlock = BasicBlock::Create(C, "", Head->getParent(), Tail);
- CheckTerm = BranchInst::Create(Tail, ThenBlock);
- }
- BranchInst *HeadNewTerm =
- BranchInst::Create(/*ifTrue*/ThenBlock, /*ifFalse*/Tail, Cmp);
- ReplaceInstWithInst(HeadOldTerm, HeadNewTerm);
+/// \brief Create a global describing a source location.
+static GlobalVariable *createPrivateGlobalForSourceLoc(Module &M,
+ LocationMetadata MD) {
+ Constant *LocData[] = {
+ createPrivateGlobalForString(M, MD.Filename, true),
+ ConstantInt::get(Type::getInt32Ty(M.getContext()), MD.LineNo),
+ ConstantInt::get(Type::getInt32Ty(M.getContext()), MD.ColumnNo),
+ };
+ auto LocStruct = ConstantStruct::getAnon(LocData);
+ auto GV = new GlobalVariable(M, LocStruct->getType(), true,
+ GlobalValue::PrivateLinkage, LocStruct,
+ kAsanGenPrefix);
+ GV->setUnnamedAddr(true);
+ return GV;
+}
- return CheckTerm;
+static bool GlobalWasGeneratedByAsan(GlobalVariable *G) {
+ return G->getName().find(kAsanGenPrefix) == 0;
}
Value *AddressSanitizer::memToShadow(Value *Shadow, IRBuilder<> &IRB) {
// Shadow >> scale
- Shadow = IRB.CreateLShr(Shadow, MappingScale);
- if (MappingOffset == 0)
+ Shadow = IRB.CreateLShr(Shadow, Mapping.Scale);
+ if (Mapping.Offset == 0)
return Shadow;
// (Shadow >> scale) | offset
- return IRB.CreateOr(Shadow, ConstantInt::get(IntptrTy,
- MappingOffset));
-}
-
-void AddressSanitizer::instrumentMemIntrinsicParam(
- AsanFunctionContext &AFC, Instruction *OrigIns,
- Value *Addr, Value *Size, Instruction *InsertBefore, bool IsWrite) {
- // Check the first byte.
- {
- IRBuilder<> IRB(InsertBefore);
- instrumentAddress(AFC, OrigIns, IRB, Addr, 8, IsWrite);
- }
- // Check the last byte.
- {
- IRBuilder<> IRB(InsertBefore);
- Value *SizeMinusOne = IRB.CreateSub(
- Size, ConstantInt::get(Size->getType(), 1));
- SizeMinusOne = IRB.CreateIntCast(SizeMinusOne, IntptrTy, false);
- Value *AddrLong = IRB.CreatePointerCast(Addr, IntptrTy);
- Value *AddrPlusSizeMinisOne = IRB.CreateAdd(AddrLong, SizeMinusOne);
- instrumentAddress(AFC, OrigIns, IRB, AddrPlusSizeMinisOne, 8, IsWrite);
- }
+ if (Mapping.OrShadowOffset)
+ return IRB.CreateOr(Shadow, ConstantInt::get(IntptrTy, Mapping.Offset));
+ else
+ return IRB.CreateAdd(Shadow, ConstantInt::get(IntptrTy, Mapping.Offset));
}
// Instrument memset/memmove/memcpy
-bool AddressSanitizer::instrumentMemIntrinsic(AsanFunctionContext &AFC,
- MemIntrinsic *MI) {
- Value *Dst = MI->getDest();
- MemTransferInst *MemTran = dyn_cast<MemTransferInst>(MI);
- Value *Src = MemTran ? MemTran->getSource() : 0;
- Value *Length = MI->getLength();
-
- Constant *ConstLength = dyn_cast<Constant>(Length);
- Instruction *InsertBefore = MI;
- if (ConstLength) {
- if (ConstLength->isNullValue()) return false;
- } else {
- // The size is not a constant so it could be zero -- check at run-time.
- IRBuilder<> IRB(InsertBefore);
-
- Value *Cmp = IRB.CreateICmpNE(Length,
- Constant::getNullValue(Length->getType()));
- InsertBefore = splitBlockAndInsertIfThen(Cmp);
+void AddressSanitizer::instrumentMemIntrinsic(MemIntrinsic *MI) {
+ IRBuilder<> IRB(MI);
+ if (isa<MemTransferInst>(MI)) {
+ IRB.CreateCall3(
+ isa<MemMoveInst>(MI) ? AsanMemmove : AsanMemcpy,
+ IRB.CreatePointerCast(MI->getOperand(0), IRB.getInt8PtrTy()),
+ IRB.CreatePointerCast(MI->getOperand(1), IRB.getInt8PtrTy()),
+ IRB.CreateIntCast(MI->getOperand(2), IntptrTy, false));
+ } else if (isa<MemSetInst>(MI)) {
+ IRB.CreateCall3(
+ AsanMemset,
+ IRB.CreatePointerCast(MI->getOperand(0), IRB.getInt8PtrTy()),
+ IRB.CreateIntCast(MI->getOperand(1), IRB.getInt32Ty(), false),
+ IRB.CreateIntCast(MI->getOperand(2), IntptrTy, false));
}
-
- instrumentMemIntrinsicParam(AFC, MI, Dst, Length, InsertBefore, true);
- if (Src)
- instrumentMemIntrinsicParam(AFC, MI, Src, Length, InsertBefore, false);
- return true;
+ MI->eraseFromParent();
}
// If I is an interesting memory access, return the PointerOperand
-// and set IsWrite. Otherwise return NULL.
-static Value *isInterestingMemoryAccess(Instruction *I, bool *IsWrite) {
+// and set IsWrite/Alignment. Otherwise return NULL.
+static Value *isInterestingMemoryAccess(Instruction *I, bool *IsWrite,
+ unsigned *Alignment) {
+ // Skip memory accesses inserted by another instrumentation.
+ if (I->getMetadata("nosanitize"))
+ return nullptr;
if (LoadInst *LI = dyn_cast<LoadInst>(I)) {
- if (!ClInstrumentReads) return NULL;
+ if (!ClInstrumentReads) return nullptr;
*IsWrite = false;
+ *Alignment = LI->getAlignment();
return LI->getPointerOperand();
}
if (StoreInst *SI = dyn_cast<StoreInst>(I)) {
- if (!ClInstrumentWrites) return NULL;
+ if (!ClInstrumentWrites) return nullptr;
*IsWrite = true;
+ *Alignment = SI->getAlignment();
return SI->getPointerOperand();
}
if (AtomicRMWInst *RMW = dyn_cast<AtomicRMWInst>(I)) {
- if (!ClInstrumentAtomics) return NULL;
+ if (!ClInstrumentAtomics) return nullptr;
*IsWrite = true;
+ *Alignment = 0;
return RMW->getPointerOperand();
}
if (AtomicCmpXchgInst *XCHG = dyn_cast<AtomicCmpXchgInst>(I)) {
- if (!ClInstrumentAtomics) return NULL;
+ if (!ClInstrumentAtomics) return nullptr;
*IsWrite = true;
+ *Alignment = 0;
return XCHG->getPointerOperand();
}
- return NULL;
+ return nullptr;
}
-void AddressSanitizer::instrumentMop(AsanFunctionContext &AFC, Instruction *I) {
- bool IsWrite;
- Value *Addr = isInterestingMemoryAccess(I, &IsWrite);
+static bool isPointerOperand(Value *V) {
+ return V->getType()->isPointerTy() || isa<PtrToIntInst>(V);
+}
+
+// This is a rough heuristic; it may cause both false positives and
+// false negatives. The proper implementation requires cooperation with
+// the frontend.
+static bool isInterestingPointerComparisonOrSubtraction(Instruction *I) {
+ if (ICmpInst *Cmp = dyn_cast<ICmpInst>(I)) {
+ if (!Cmp->isRelational())
+ return false;
+ } else if (BinaryOperator *BO = dyn_cast<BinaryOperator>(I)) {
+ if (BO->getOpcode() != Instruction::Sub)
+ return false;
+ } else {
+ return false;
+ }
+ if (!isPointerOperand(I->getOperand(0)) ||
+ !isPointerOperand(I->getOperand(1)))
+ return false;
+ return true;
+}
+
+bool AddressSanitizer::GlobalIsLinkerInitialized(GlobalVariable *G) {
+ // If a global variable does not have dynamic initialization we don't
+ // have to instrument it. However, if a global does not have initializer
+ // at all, we assume it has dynamic initializer (in other TU).
+ return G->hasInitializer() && !GlobalsMD.get(G).IsDynInit;
+}
+
+void
+AddressSanitizer::instrumentPointerComparisonOrSubtraction(Instruction *I) {
+ IRBuilder<> IRB(I);
+ Function *F = isa<ICmpInst>(I) ? AsanPtrCmpFunction : AsanPtrSubFunction;
+ Value *Param[2] = {I->getOperand(0), I->getOperand(1)};
+ for (int i = 0; i < 2; i++) {
+ if (Param[i]->getType()->isPointerTy())
+ Param[i] = IRB.CreatePointerCast(Param[i], IntptrTy);
+ }
+ IRB.CreateCall2(F, Param[0], Param[1]);
+}
+
+void AddressSanitizer::instrumentMop(Instruction *I, bool UseCalls) {
+ bool IsWrite = false;
+ unsigned Alignment = 0;
+ Value *Addr = isInterestingMemoryAccess(I, &IsWrite, &Alignment);
assert(Addr);
- if (ClOpt && ClOptGlobals && isa<GlobalVariable>(Addr)) {
- // We are accessing a global scalar variable. Nothing to catch here.
- return;
+ if (ClOpt && ClOptGlobals) {
+ if (GlobalVariable *G = dyn_cast<GlobalVariable>(Addr)) {
+ // If initialization order checking is disabled, a simple access to a
+ // dynamically initialized global is always valid.
+ if (!ClInitializers || GlobalIsLinkerInitialized(G)) {
+ NumOptimizedAccessesToGlobalVar++;
+ return;
+ }
+ }
+ ConstantExpr *CE = dyn_cast<ConstantExpr>(Addr);
+ if (CE && CE->isGEPWithNoNotionalOverIndexing()) {
+ if (GlobalVariable *G = dyn_cast<GlobalVariable>(CE->getOperand(0))) {
+ if (CE->getOperand(1)->isNullValue() && GlobalIsLinkerInitialized(G)) {
+ NumOptimizedAccessesToGlobalArray++;
+ return;
+ }
+ }
+ }
}
+
Type *OrigPtrTy = Addr->getType();
Type *OrigTy = cast<PointerType>(OrigPtrTy)->getElementType();
assert(OrigTy->isSized());
- uint32_t TypeSize = TD->getTypeStoreSizeInBits(OrigTy);
+ uint32_t TypeSize = DL->getTypeStoreSizeInBits(OrigTy);
- if (TypeSize != 8 && TypeSize != 16 &&
- TypeSize != 32 && TypeSize != 64 && TypeSize != 128) {
- // Ignore all unusual sizes.
- return;
- }
+ assert((TypeSize % 8) == 0);
+ if (IsWrite)
+ NumInstrumentedWrites++;
+ else
+ NumInstrumentedReads++;
+
+ unsigned Granularity = 1 << Mapping.Scale;
+ // Instrument a 1-, 2-, 4-, 8-, or 16- byte access with one check
+ // if the data is properly aligned.
+ if ((TypeSize == 8 || TypeSize == 16 || TypeSize == 32 || TypeSize == 64 ||
+ TypeSize == 128) &&
+ (Alignment >= Granularity || Alignment == 0 || Alignment >= TypeSize / 8))
+ return instrumentAddress(I, I, Addr, TypeSize, IsWrite, nullptr, UseCalls);
+ // Instrument unusual size or unusual alignment.
+ // We can not do it with a single check, so we do 1-byte check for the first
+ // and the last bytes. We call __asan_report_*_n(addr, real_size) to be able
+ // to report the actual access size.
IRBuilder<> IRB(I);
- instrumentAddress(AFC, I, IRB, Addr, TypeSize, IsWrite);
+ Value *Size = ConstantInt::get(IntptrTy, TypeSize / 8);
+ Value *AddrLong = IRB.CreatePointerCast(Addr, IntptrTy);
+ if (UseCalls) {
+ IRB.CreateCall2(AsanMemoryAccessCallbackSized[IsWrite], AddrLong, Size);
+ } else {
+ Value *LastByte = IRB.CreateIntToPtr(
+ IRB.CreateAdd(AddrLong, ConstantInt::get(IntptrTy, TypeSize / 8 - 1)),
+ OrigPtrTy);
+ instrumentAddress(I, I, Addr, 8, IsWrite, Size, false);
+ instrumentAddress(I, I, LastByte, 8, IsWrite, Size, false);
+ }
}
// Validate the result of Module::getOrInsertFunction called for an interface
// function of AddressSanitizer. If the instrumented module defines a function
// with the same name, their prototypes must match, otherwise
// getOrInsertFunction returns a bitcast.
-Function *AddressSanitizer::checkInterfaceFunction(Constant *FuncOrBitcast) {
+static Function *checkInterfaceFunction(Constant *FuncOrBitcast) {
if (isa<Function>(FuncOrBitcast)) return cast<Function>(FuncOrBitcast);
FuncOrBitcast->dump();
report_fatal_error("trying to redefine an AddressSanitizer "
}
Instruction *AddressSanitizer::generateCrashCode(
- BasicBlock *BB, Value *Addr, Value *PC,
- bool IsWrite, size_t AccessSizeIndex) {
- IRBuilder<> IRB(BB->getFirstNonPHI());
- CallInst *Call;
- if (PC)
- Call = IRB.CreateCall2(AsanErrorCallback[IsWrite][AccessSizeIndex],
- Addr, PC);
- else
- Call = IRB.CreateCall(AsanErrorCallback[IsWrite][AccessSizeIndex], Addr);
+ Instruction *InsertBefore, Value *Addr,
+ bool IsWrite, size_t AccessSizeIndex, Value *SizeArgument) {
+ IRBuilder<> IRB(InsertBefore);
+ CallInst *Call = SizeArgument
+ ? IRB.CreateCall2(AsanErrorCallbackSized[IsWrite], Addr, SizeArgument)
+ : IRB.CreateCall(AsanErrorCallback[IsWrite][AccessSizeIndex], Addr);
+
// We don't do Call->setDoesNotReturn() because the BB already has
// UnreachableInst at the end.
// This EmptyAsm is required to avoid callback merge.
Value *AddressSanitizer::createSlowPathCmp(IRBuilder<> &IRB, Value *AddrLong,
Value *ShadowValue,
uint32_t TypeSize) {
- size_t Granularity = 1 << MappingScale;
+ size_t Granularity = 1 << Mapping.Scale;
// Addr & (Granularity - 1)
Value *LastAccessedByte = IRB.CreateAnd(
AddrLong, ConstantInt::get(IntptrTy, Granularity - 1));
LastAccessedByte, ConstantInt::get(IntptrTy, TypeSize / 8 - 1));
// (uint8_t) ((Addr & (Granularity-1)) + size - 1)
LastAccessedByte = IRB.CreateIntCast(
- LastAccessedByte, IRB.getInt8Ty(), false);
+ LastAccessedByte, ShadowValue->getType(), false);
// ((uint8_t) ((Addr & (Granularity-1)) + size - 1)) >= ShadowValue
return IRB.CreateICmpSGE(LastAccessedByte, ShadowValue);
}
-void AddressSanitizer::instrumentAddress(AsanFunctionContext &AFC,
- Instruction *OrigIns,
- IRBuilder<> &IRB, Value *Addr,
- uint32_t TypeSize, bool IsWrite) {
+void AddressSanitizer::instrumentAddress(Instruction *OrigIns,
+ Instruction *InsertBefore, Value *Addr,
+ uint32_t TypeSize, bool IsWrite,
+ Value *SizeArgument, bool UseCalls) {
+ IRBuilder<> IRB(InsertBefore);
Value *AddrLong = IRB.CreatePointerCast(Addr, IntptrTy);
+ size_t AccessSizeIndex = TypeSizeToSizeIndex(TypeSize);
+
+ if (UseCalls) {
+ IRB.CreateCall(AsanMemoryAccessCallback[IsWrite][AccessSizeIndex],
+ AddrLong);
+ return;
+ }
Type *ShadowTy = IntegerType::get(
- *C, std::max(8U, TypeSize >> MappingScale));
+ *C, std::max(8U, TypeSize >> Mapping.Scale));
Type *ShadowPtrTy = PointerType::get(ShadowTy, 0);
Value *ShadowPtr = memToShadow(AddrLong, IRB);
Value *CmpVal = Constant::getNullValue(ShadowTy);
IRB.CreateIntToPtr(ShadowPtr, ShadowPtrTy));
Value *Cmp = IRB.CreateICmpNE(ShadowValue, CmpVal);
-
- BasicBlock *CrashBlock = 0;
- if (ClMergeCallbacks) {
- size_t AccessSizeIndex = TypeSizeToSizeIndex(TypeSize);
- BasicBlock **Cached = &AFC.CrashBlock[IsWrite][AccessSizeIndex];
- if (!*Cached) {
- std::string BBName("crash_bb-");
- BBName += (IsWrite ? "w-" : "r-") + itostr(1 << AccessSizeIndex);
- BasicBlock *BB = BasicBlock::Create(*C, BBName, &AFC.F);
- new UnreachableInst(*C, BB);
- *Cached = BB;
- }
- CrashBlock = *Cached;
- // We need to pass the PC as the second parameter to __asan_report_*.
- // There are few problems:
- // - Some architectures (e.g. x86_32) don't have a cheap way to get the PC.
- // - LLVM doesn't have the appropriate intrinsic.
- // For now, put a random number into the PC, just to allow experiments.
- Value *PC = ConstantInt::get(IntptrTy, rand());
- CrashArg Arg = {AddrLong, PC};
- AFC.CrashArgs[IsWrite][AccessSizeIndex].push_back(Arg);
- } else {
- CrashBlock = BasicBlock::Create(*C, "crash_bb", &AFC.F);
- new UnreachableInst(*C, CrashBlock);
- size_t AccessSizeIndex = TypeSizeToSizeIndex(TypeSize);
- Instruction *Crash =
- generateCrashCode(CrashBlock, AddrLong, 0, IsWrite, AccessSizeIndex);
- Crash->setDebugLoc(OrigIns->getDebugLoc());
- }
-
- size_t Granularity = 1 << MappingScale;
- if (TypeSize < 8 * Granularity) {
- BranchInst *CheckTerm = splitBlockAndInsertIfThen(Cmp);
- assert(CheckTerm->isUnconditional());
+ size_t Granularity = 1 << Mapping.Scale;
+ TerminatorInst *CrashTerm = nullptr;
+
+ if (ClAlwaysSlowPath || (TypeSize < 8 * Granularity)) {
+ // We use branch weights for the slow path check, to indicate that the slow
+ // path is rarely taken. This seems to be the case for SPEC benchmarks.
+ TerminatorInst *CheckTerm =
+ SplitBlockAndInsertIfThen(Cmp, InsertBefore, false,
+ MDBuilder(*C).createBranchWeights(1, 100000));
+ assert(dyn_cast<BranchInst>(CheckTerm)->isUnconditional());
BasicBlock *NextBB = CheckTerm->getSuccessor(0);
IRB.SetInsertPoint(CheckTerm);
Value *Cmp2 = createSlowPathCmp(IRB, AddrLong, ShadowValue, TypeSize);
+ BasicBlock *CrashBlock =
+ BasicBlock::Create(*C, "", NextBB->getParent(), NextBB);
+ CrashTerm = new UnreachableInst(*C, CrashBlock);
BranchInst *NewTerm = BranchInst::Create(CrashBlock, NextBB, Cmp2);
ReplaceInstWithInst(CheckTerm, NewTerm);
} else {
- splitBlockAndInsertIfThen(Cmp, CrashBlock);
+ CrashTerm = SplitBlockAndInsertIfThen(Cmp, InsertBefore, true);
}
+
+ Instruction *Crash = generateCrashCode(
+ CrashTerm, AddrLong, IsWrite, AccessSizeIndex, SizeArgument);
+ Crash->setDebugLoc(OrigIns->getDebugLoc());
}
-// This function replaces all global variables with new variables that have
-// trailing redzones. It also creates a function that poisons
-// redzones and inserts this function into llvm.global_ctors.
-bool AddressSanitizer::insertGlobalRedzones(Module &M) {
- SmallVector<GlobalVariable *, 16> GlobalsToChange;
+void AddressSanitizerModule::poisonOneInitializer(Function &GlobalInit,
+ GlobalValue *ModuleName) {
+ // Set up the arguments to our poison/unpoison functions.
+ IRBuilder<> IRB(GlobalInit.begin()->getFirstInsertionPt());
- for (Module::GlobalListType::iterator G = M.getGlobalList().begin(),
- E = M.getGlobalList().end(); G != E; ++G) {
- Type *Ty = cast<PointerType>(G->getType())->getElementType();
- DEBUG(dbgs() << "GLOBAL: " << *G);
-
- if (!Ty->isSized()) continue;
- if (!G->hasInitializer()) continue;
- // Touch only those globals that will not be defined in other modules.
- // Don't handle ODR type linkages since other modules may be built w/o asan.
- if (G->getLinkage() != GlobalVariable::ExternalLinkage &&
- G->getLinkage() != GlobalVariable::PrivateLinkage &&
- G->getLinkage() != GlobalVariable::InternalLinkage)
- continue;
- // Two problems with thread-locals:
- // - The address of the main thread's copy can't be computed at link-time.
- // - Need to poison all copies, not just the main thread's one.
- if (G->isThreadLocal())
- continue;
- // For now, just ignore this Alloca if the alignment is large.
- if (G->getAlignment() > RedzoneSize) continue;
-
- // Ignore all the globals with the names starting with "\01L_OBJC_".
- // Many of those are put into the .cstring section. The linker compresses
- // that section by removing the spare \0s after the string terminator, so
- // our redzones get broken.
- if ((G->getName().find("\01L_OBJC_") == 0) ||
- (G->getName().find("\01l_OBJC_") == 0)) {
- DEBUG(dbgs() << "Ignoring \\01L_OBJC_* global: " << *G);
+ // Add a call to poison all external globals before the given function starts.
+ Value *ModuleNameAddr = ConstantExpr::getPointerCast(ModuleName, IntptrTy);
+ IRB.CreateCall(AsanPoisonGlobals, ModuleNameAddr);
+
+ // Add calls to unpoison all globals before each return instruction.
+ for (auto &BB : GlobalInit.getBasicBlockList())
+ if (ReturnInst *RI = dyn_cast<ReturnInst>(BB.getTerminator()))
+ CallInst::Create(AsanUnpoisonGlobals, "", RI);
+}
+
+void AddressSanitizerModule::createInitializerPoisonCalls(
+ Module &M, GlobalValue *ModuleName) {
+ GlobalVariable *GV = M.getGlobalVariable("llvm.global_ctors");
+
+ ConstantArray *CA = cast<ConstantArray>(GV->getInitializer());
+ for (Use &OP : CA->operands()) {
+ if (isa<ConstantAggregateZero>(OP))
continue;
+ ConstantStruct *CS = cast<ConstantStruct>(OP);
+
+ // Must have a function or null ptr.
+ if (Function* F = dyn_cast<Function>(CS->getOperand(1))) {
+ if (F->getName() == kAsanModuleCtorName) continue;
+ ConstantInt *Priority = dyn_cast<ConstantInt>(CS->getOperand(0));
+ // Don't instrument CTORs that will run before asan.module_ctor.
+ if (Priority->getLimitedValue() <= kAsanCtorAndDtorPriority) continue;
+ poisonOneInitializer(*F, ModuleName);
}
+ }
+}
- if (G->hasSection()) {
- StringRef Section(G->getSection());
- // Ignore the globals from the __OBJC section. The ObjC runtime assumes
- // those conform to /usr/lib/objc/runtime.h, so we can't add redzones to
- // them.
- if ((Section.find("__OBJC,") == 0) ||
- (Section.find("__DATA, __objc_") == 0)) {
- DEBUG(dbgs() << "Ignoring ObjC runtime global: " << *G);
- continue;
- }
- // See http://code.google.com/p/address-sanitizer/issues/detail?id=32
- // Constant CFString instances are compiled in the following way:
- // -- the string buffer is emitted into
- // __TEXT,__cstring,cstring_literals
- // -- the constant NSConstantString structure referencing that buffer
- // is placed into __DATA,__cfstring
- // Therefore there's no point in placing redzones into __DATA,__cfstring.
- // Moreover, it causes the linker to crash on OS X 10.7
- if (Section.find("__DATA,__cfstring") == 0) {
- DEBUG(dbgs() << "Ignoring CFString: " << *G);
- continue;
- }
+bool AddressSanitizerModule::ShouldInstrumentGlobal(GlobalVariable *G) {
+ Type *Ty = cast<PointerType>(G->getType())->getElementType();
+ DEBUG(dbgs() << "GLOBAL: " << *G << "\n");
+
+ if (GlobalsMD.get(G).IsBlacklisted) return false;
+ if (!Ty->isSized()) return false;
+ if (!G->hasInitializer()) return false;
+ if (GlobalWasGeneratedByAsan(G)) return false; // Our own global.
+ // Touch only those globals that will not be defined in other modules.
+ // Don't handle ODR linkage types and COMDATs since other modules may be built
+ // without ASan.
+ if (G->getLinkage() != GlobalVariable::ExternalLinkage &&
+ G->getLinkage() != GlobalVariable::PrivateLinkage &&
+ G->getLinkage() != GlobalVariable::InternalLinkage)
+ return false;
+ if (G->hasComdat())
+ return false;
+ // Two problems with thread-locals:
+ // - The address of the main thread's copy can't be computed at link-time.
+ // - Need to poison all copies, not just the main thread's one.
+ if (G->isThreadLocal())
+ return false;
+ // For now, just ignore this Global if the alignment is large.
+ if (G->getAlignment() > MinRedzoneSizeForGlobal()) return false;
+
+ // Ignore all the globals with the names starting with "\01L_OBJC_".
+ // Many of those are put into the .cstring section. The linker compresses
+ // that section by removing the spare \0s after the string terminator, so
+ // our redzones get broken.
+ if ((G->getName().find("\01L_OBJC_") == 0) ||
+ (G->getName().find("\01l_OBJC_") == 0)) {
+ DEBUG(dbgs() << "Ignoring \\01L_OBJC_* global: " << *G << "\n");
+ return false;
+ }
+
+ if (G->hasSection()) {
+ StringRef Section(G->getSection());
+ // Ignore the globals from the __OBJC section. The ObjC runtime assumes
+ // those conform to /usr/lib/objc/runtime.h, so we can't add redzones to
+ // them.
+ if (Section.startswith("__OBJC,") ||
+ Section.startswith("__DATA, __objc_")) {
+ DEBUG(dbgs() << "Ignoring ObjC runtime global: " << *G << "\n");
+ return false;
+ }
+ // See http://code.google.com/p/address-sanitizer/issues/detail?id=32
+ // Constant CFString instances are compiled in the following way:
+ // -- the string buffer is emitted into
+ // __TEXT,__cstring,cstring_literals
+ // -- the constant NSConstantString structure referencing that buffer
+ // is placed into __DATA,__cfstring
+ // Therefore there's no point in placing redzones into __DATA,__cfstring.
+ // Moreover, it causes the linker to crash on OS X 10.7
+ if (Section.startswith("__DATA,__cfstring")) {
+ DEBUG(dbgs() << "Ignoring CFString: " << *G << "\n");
+ return false;
+ }
+ // The linker merges the contents of cstring_literals and removes the
+ // trailing zeroes.
+ if (Section.startswith("__TEXT,__cstring,cstring_literals")) {
+ DEBUG(dbgs() << "Ignoring a cstring literal: " << *G << "\n");
+ return false;
+ }
+
+ // Callbacks put into the CRT initializer/terminator sections
+ // should not be instrumented.
+ // See https://code.google.com/p/address-sanitizer/issues/detail?id=305
+ // and http://msdn.microsoft.com/en-US/en-en/library/bb918180(v=vs.120).aspx
+ if (Section.startswith(".CRT")) {
+ DEBUG(dbgs() << "Ignoring a global initializer callback: " << *G << "\n");
+ return false;
}
- GlobalsToChange.push_back(G);
+ // Globals from llvm.metadata aren't emitted, do not instrument them.
+ if (Section == "llvm.metadata") return false;
+ }
+
+ return true;
+}
+
+void AddressSanitizerModule::initializeCallbacks(Module &M) {
+ IRBuilder<> IRB(*C);
+ // Declare our poisoning and unpoisoning functions.
+ AsanPoisonGlobals = checkInterfaceFunction(M.getOrInsertFunction(
+ kAsanPoisonGlobalsName, IRB.getVoidTy(), IntptrTy, NULL));
+ AsanPoisonGlobals->setLinkage(Function::ExternalLinkage);
+ AsanUnpoisonGlobals = checkInterfaceFunction(M.getOrInsertFunction(
+ kAsanUnpoisonGlobalsName, IRB.getVoidTy(), NULL));
+ AsanUnpoisonGlobals->setLinkage(Function::ExternalLinkage);
+ // Declare functions that register/unregister globals.
+ AsanRegisterGlobals = checkInterfaceFunction(M.getOrInsertFunction(
+ kAsanRegisterGlobalsName, IRB.getVoidTy(),
+ IntptrTy, IntptrTy, NULL));
+ AsanRegisterGlobals->setLinkage(Function::ExternalLinkage);
+ AsanUnregisterGlobals = checkInterfaceFunction(M.getOrInsertFunction(
+ kAsanUnregisterGlobalsName,
+ IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
+ AsanUnregisterGlobals->setLinkage(Function::ExternalLinkage);
+ AsanCovModuleInit = checkInterfaceFunction(M.getOrInsertFunction(
+ kAsanCovModuleInitName,
+ IRB.getVoidTy(), IntptrTy, NULL));
+ AsanCovModuleInit->setLinkage(Function::ExternalLinkage);
+}
+
+// This function replaces all global variables with new variables that have
+// trailing redzones. It also creates a function that poisons
+// redzones and inserts this function into llvm.global_ctors.
+bool AddressSanitizerModule::InstrumentGlobals(IRBuilder<> &IRB, Module &M) {
+ GlobalsMD.init(M);
+
+ SmallVector<GlobalVariable *, 16> GlobalsToChange;
+
+ for (auto &G : M.globals()) {
+ if (ShouldInstrumentGlobal(&G))
+ GlobalsToChange.push_back(&G);
}
size_t n = GlobalsToChange.size();
// size_t size;
// size_t size_with_redzone;
// const char *name;
+ // const char *module_name;
+ // size_t has_dynamic_init;
+ // void *source_location;
// We initialize an array of such structures and pass it to a run-time call.
- StructType *GlobalStructTy = StructType::get(IntptrTy, IntptrTy,
- IntptrTy, IntptrTy, NULL);
+ StructType *GlobalStructTy =
+ StructType::get(IntptrTy, IntptrTy, IntptrTy, IntptrTy, IntptrTy,
+ IntptrTy, IntptrTy, NULL);
SmallVector<Constant *, 16> Initializers(n);
- IRBuilder<> IRB(CtorInsertBefore);
+ bool HasDynamicallyInitializedGlobals = false;
+
+ // We shouldn't merge same module names, as this string serves as unique
+ // module ID in runtime.
+ GlobalVariable *ModuleName = createPrivateGlobalForString(
+ M, M.getModuleIdentifier(), /*AllowMerging*/false);
for (size_t i = 0; i < n; i++) {
+ static const uint64_t kMaxGlobalRedzone = 1 << 18;
GlobalVariable *G = GlobalsToChange[i];
+
+ auto MD = GlobalsMD.get(G);
+ // Create string holding the global name (use global name from metadata
+ // if it's available, otherwise just write the name of global variable).
+ GlobalVariable *Name = createPrivateGlobalForString(
+ M, MD.Name.empty() ? G->getName() : MD.Name,
+ /*AllowMerging*/ true);
+
PointerType *PtrTy = cast<PointerType>(G->getType());
Type *Ty = PtrTy->getElementType();
- uint64_t SizeInBytes = TD->getTypeAllocSize(Ty);
- uint64_t RightRedzoneSize = RedzoneSize +
- (RedzoneSize - (SizeInBytes % RedzoneSize));
+ uint64_t SizeInBytes = DL->getTypeAllocSize(Ty);
+ uint64_t MinRZ = MinRedzoneSizeForGlobal();
+ // MinRZ <= RZ <= kMaxGlobalRedzone
+ // and trying to make RZ to be ~ 1/4 of SizeInBytes.
+ uint64_t RZ = std::max(MinRZ,
+ std::min(kMaxGlobalRedzone,
+ (SizeInBytes / MinRZ / 4) * MinRZ));
+ uint64_t RightRedzoneSize = RZ;
+ // Round up to MinRZ
+ if (SizeInBytes % MinRZ)
+ RightRedzoneSize += MinRZ - (SizeInBytes % MinRZ);
+ assert(((RightRedzoneSize + SizeInBytes) % MinRZ) == 0);
Type *RightRedZoneTy = ArrayType::get(IRB.getInt8Ty(), RightRedzoneSize);
StructType *NewTy = StructType::get(Ty, RightRedZoneTy, NULL);
NewTy, G->getInitializer(),
Constant::getNullValue(RightRedZoneTy), NULL);
- SmallString<2048> DescriptionOfGlobal = G->getName();
- DescriptionOfGlobal += " (";
- DescriptionOfGlobal += M.getModuleIdentifier();
- DescriptionOfGlobal += ")";
- GlobalVariable *Name = createPrivateGlobalForString(M, DescriptionOfGlobal);
-
// Create a new global variable with enough space for a redzone.
+ GlobalValue::LinkageTypes Linkage = G->getLinkage();
+ if (G->isConstant() && Linkage == GlobalValue::PrivateLinkage)
+ Linkage = GlobalValue::InternalLinkage;
GlobalVariable *NewGlobal = new GlobalVariable(
- M, NewTy, G->isConstant(), G->getLinkage(),
+ M, NewTy, G->isConstant(), Linkage,
NewInitializer, "", G, G->getThreadLocalMode());
NewGlobal->copyAttributesFrom(G);
- NewGlobal->setAlignment(RedzoneSize);
+ NewGlobal->setAlignment(MinRZ);
Value *Indices2[2];
Indices2[0] = IRB.getInt32(0);
NewGlobal->takeName(G);
G->eraseFromParent();
+ Constant *SourceLoc;
+ if (!MD.SourceLoc.empty()) {
+ auto SourceLocGlobal = createPrivateGlobalForSourceLoc(M, MD.SourceLoc);
+ SourceLoc = ConstantExpr::getPointerCast(SourceLocGlobal, IntptrTy);
+ } else {
+ SourceLoc = ConstantInt::get(IntptrTy, 0);
+ }
+
Initializers[i] = ConstantStruct::get(
- GlobalStructTy,
- ConstantExpr::getPointerCast(NewGlobal, IntptrTy),
+ GlobalStructTy, ConstantExpr::getPointerCast(NewGlobal, IntptrTy),
ConstantInt::get(IntptrTy, SizeInBytes),
ConstantInt::get(IntptrTy, SizeInBytes + RightRedzoneSize),
ConstantExpr::getPointerCast(Name, IntptrTy),
- NULL);
- DEBUG(dbgs() << "NEW GLOBAL:\n" << *NewGlobal);
+ ConstantExpr::getPointerCast(ModuleName, IntptrTy),
+ ConstantInt::get(IntptrTy, MD.IsDynInit), SourceLoc, NULL);
+
+ if (ClInitializers && MD.IsDynInit)
+ HasDynamicallyInitializedGlobals = true;
+
+ DEBUG(dbgs() << "NEW GLOBAL: " << *NewGlobal << "\n");
}
ArrayType *ArrayOfGlobalStructTy = ArrayType::get(GlobalStructTy, n);
GlobalVariable *AllGlobals = new GlobalVariable(
- M, ArrayOfGlobalStructTy, false, GlobalVariable::PrivateLinkage,
+ M, ArrayOfGlobalStructTy, false, GlobalVariable::InternalLinkage,
ConstantArray::get(ArrayOfGlobalStructTy, Initializers), "");
- Function *AsanRegisterGlobals = checkInterfaceFunction(M.getOrInsertFunction(
- kAsanRegisterGlobalsName, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
- AsanRegisterGlobals->setLinkage(Function::ExternalLinkage);
-
+ // Create calls for poisoning before initializers run and unpoisoning after.
+ if (HasDynamicallyInitializedGlobals)
+ createInitializerPoisonCalls(M, ModuleName);
IRB.CreateCall2(AsanRegisterGlobals,
IRB.CreatePointerCast(AllGlobals, IntptrTy),
ConstantInt::get(IntptrTy, n));
GlobalValue::InternalLinkage, kAsanModuleDtorName, &M);
BasicBlock *AsanDtorBB = BasicBlock::Create(*C, "", AsanDtorFunction);
IRBuilder<> IRB_Dtor(ReturnInst::Create(*C, AsanDtorBB));
- Function *AsanUnregisterGlobals =
- checkInterfaceFunction(M.getOrInsertFunction(
- kAsanUnregisterGlobalsName,
- IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
- AsanUnregisterGlobals->setLinkage(Function::ExternalLinkage);
-
IRB_Dtor.CreateCall2(AsanUnregisterGlobals,
IRB.CreatePointerCast(AllGlobals, IntptrTy),
ConstantInt::get(IntptrTy, n));
- appendToGlobalDtors(M, AsanDtorFunction, kAsanCtorAndCtorPriority);
+ appendToGlobalDtors(M, AsanDtorFunction, kAsanCtorAndDtorPriority);
DEBUG(dbgs() << M);
return true;
}
-// virtual
-bool AddressSanitizer::runOnModule(Module &M) {
- // Initialize the private fields. No one has accessed them before.
- TD = getAnalysisIfAvailable<TargetData>();
- if (!TD)
+bool AddressSanitizerModule::runOnModule(Module &M) {
+ DataLayoutPass *DLP = getAnalysisIfAvailable<DataLayoutPass>();
+ if (!DLP)
return false;
- BL.reset(new FunctionBlackList(ClBlackListFile));
-
+ DL = &DLP->getDataLayout();
C = &(M.getContext());
- LongSize = TD->getPointerSizeInBits();
+ int LongSize = DL->getPointerSizeInBits();
IntptrTy = Type::getIntNTy(*C, LongSize);
- IntptrPtrTy = PointerType::get(IntptrTy, 0);
+ Mapping = getShadowMapping(M, LongSize);
+ initializeCallbacks(M);
- AsanCtorFunction = Function::Create(
- FunctionType::get(Type::getVoidTy(*C), false),
- GlobalValue::InternalLinkage, kAsanModuleCtorName, &M);
- BasicBlock *AsanCtorBB = BasicBlock::Create(*C, "", AsanCtorFunction);
- CtorInsertBefore = ReturnInst::Create(*C, AsanCtorBB);
+ bool Changed = false;
- // call __asan_init in the module ctor.
- IRBuilder<> IRB(CtorInsertBefore);
- AsanInitFunction = checkInterfaceFunction(
- M.getOrInsertFunction(kAsanInitName, IRB.getVoidTy(), NULL));
- AsanInitFunction->setLinkage(Function::ExternalLinkage);
- IRB.CreateCall(AsanInitFunction);
+ Function *CtorFunc = M.getFunction(kAsanModuleCtorName);
+ assert(CtorFunc);
+ IRBuilder<> IRB(CtorFunc->getEntryBlock().getTerminator());
+ if (ClCoverage > 0) {
+ Function *CovFunc = M.getFunction(kAsanCovName);
+ int nCov = CovFunc ? CovFunc->getNumUses() : 0;
+ IRB.CreateCall(AsanCovModuleInit, ConstantInt::get(IntptrTy, nCov));
+ Changed = true;
+ }
+
+ if (ClGlobals)
+ Changed |= InstrumentGlobals(IRB, M);
+
+ return Changed;
+}
+
+void AddressSanitizer::initializeCallbacks(Module &M) {
+ IRBuilder<> IRB(*C);
// Create __asan_report* callbacks.
for (size_t AccessIsWrite = 0; AccessIsWrite <= 1; AccessIsWrite++) {
for (size_t AccessSizeIndex = 0; AccessSizeIndex < kNumberOfAccessSizes;
AccessSizeIndex++) {
// IsWrite and TypeSize are encoded in the function name.
- std::string FunctionName = std::string(kAsanReportErrorTemplate) +
+ std::string Suffix =
(AccessIsWrite ? "store" : "load") + itostr(1 << AccessSizeIndex);
- // If we are merging crash callbacks, they have two parameters.
- if (ClMergeCallbacks)
- AsanErrorCallback[AccessIsWrite][AccessSizeIndex] = cast<Function>(
- M.getOrInsertFunction(FunctionName, IRB.getVoidTy(), IntptrTy,
- IntptrTy, NULL));
- else
- AsanErrorCallback[AccessIsWrite][AccessSizeIndex] = cast<Function>(
- M.getOrInsertFunction(FunctionName, IRB.getVoidTy(), IntptrTy, NULL));
+ AsanErrorCallback[AccessIsWrite][AccessSizeIndex] =
+ checkInterfaceFunction(
+ M.getOrInsertFunction(kAsanReportErrorTemplate + Suffix,
+ IRB.getVoidTy(), IntptrTy, NULL));
+ AsanMemoryAccessCallback[AccessIsWrite][AccessSizeIndex] =
+ checkInterfaceFunction(
+ M.getOrInsertFunction(ClMemoryAccessCallbackPrefix + Suffix,
+ IRB.getVoidTy(), IntptrTy, NULL));
}
}
+ AsanErrorCallbackSized[0] = checkInterfaceFunction(M.getOrInsertFunction(
+ kAsanReportLoadN, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
+ AsanErrorCallbackSized[1] = checkInterfaceFunction(M.getOrInsertFunction(
+ kAsanReportStoreN, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
+
+ AsanMemoryAccessCallbackSized[0] = checkInterfaceFunction(
+ M.getOrInsertFunction(ClMemoryAccessCallbackPrefix + "loadN",
+ IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
+ AsanMemoryAccessCallbackSized[1] = checkInterfaceFunction(
+ M.getOrInsertFunction(ClMemoryAccessCallbackPrefix + "storeN",
+ IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
+
+ AsanMemmove = checkInterfaceFunction(M.getOrInsertFunction(
+ ClMemoryAccessCallbackPrefix + "memmove", IRB.getInt8PtrTy(),
+ IRB.getInt8PtrTy(), IRB.getInt8PtrTy(), IntptrTy, NULL));
+ AsanMemcpy = checkInterfaceFunction(M.getOrInsertFunction(
+ ClMemoryAccessCallbackPrefix + "memcpy", IRB.getInt8PtrTy(),
+ IRB.getInt8PtrTy(), IRB.getInt8PtrTy(), IntptrTy, NULL));
+ AsanMemset = checkInterfaceFunction(M.getOrInsertFunction(
+ ClMemoryAccessCallbackPrefix + "memset", IRB.getInt8PtrTy(),
+ IRB.getInt8PtrTy(), IRB.getInt32Ty(), IntptrTy, NULL));
+
+ AsanHandleNoReturnFunc = checkInterfaceFunction(
+ M.getOrInsertFunction(kAsanHandleNoReturnName, IRB.getVoidTy(), NULL));
+ AsanCovFunction = checkInterfaceFunction(M.getOrInsertFunction(
+ kAsanCovName, IRB.getVoidTy(), NULL));
+ AsanCovIndirCallFunction = checkInterfaceFunction(M.getOrInsertFunction(
+ kAsanCovIndirCallName, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
+
+ AsanPtrCmpFunction = checkInterfaceFunction(M.getOrInsertFunction(
+ kAsanPtrCmp, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
+ AsanPtrSubFunction = checkInterfaceFunction(M.getOrInsertFunction(
+ kAsanPtrSub, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
// We insert an empty inline asm after __asan_report* to avoid callback merge.
EmptyAsm = InlineAsm::get(FunctionType::get(IRB.getVoidTy(), false),
StringRef(""), StringRef(""),
/*hasSideEffects=*/true);
+}
- llvm::Triple targetTriple(M.getTargetTriple());
- bool isAndroid = targetTriple.getEnvironment() == llvm::Triple::ANDROIDEABI;
-
- MappingOffset = isAndroid ? kDefaultShadowOffsetAndroid :
- (LongSize == 32 ? kDefaultShadowOffset32 : kDefaultShadowOffset64);
- if (ClMappingOffsetLog >= 0) {
- if (ClMappingOffsetLog == 0) {
- // special case
- MappingOffset = 0;
- } else {
- MappingOffset = 1ULL << ClMappingOffsetLog;
- }
- }
- MappingScale = kDefaultShadowScale;
- if (ClMappingScale) {
- MappingScale = ClMappingScale;
- }
- // Redzone used for stack and globals is at least 32 bytes.
- // For scales 6 and 7, the redzone has to be 64 and 128 bytes respectively.
- RedzoneSize = std::max(32, (int)(1 << MappingScale));
-
- bool Res = false;
+// virtual
+bool AddressSanitizer::doInitialization(Module &M) {
+ // Initialize the private fields. No one has accessed them before.
+ DataLayoutPass *DLP = getAnalysisIfAvailable<DataLayoutPass>();
+ if (!DLP)
+ report_fatal_error("data layout missing");
+ DL = &DLP->getDataLayout();
- if (ClGlobals)
- Res |= insertGlobalRedzones(M);
-
- if (ClMappingOffsetLog >= 0) {
- // Tell the run-time the current values of mapping offset and scale.
- GlobalValue *asan_mapping_offset =
- new GlobalVariable(M, IntptrTy, true, GlobalValue::LinkOnceODRLinkage,
- ConstantInt::get(IntptrTy, MappingOffset),
- kAsanMappingOffsetName);
- // Read the global, otherwise it may be optimized away.
- IRB.CreateLoad(asan_mapping_offset, true);
- }
- if (ClMappingScale) {
- GlobalValue *asan_mapping_scale =
- new GlobalVariable(M, IntptrTy, true, GlobalValue::LinkOnceODRLinkage,
- ConstantInt::get(IntptrTy, MappingScale),
- kAsanMappingScaleName);
- // Read the global, otherwise it may be optimized away.
- IRB.CreateLoad(asan_mapping_scale, true);
- }
+ GlobalsMD.init(M);
+ C = &(M.getContext());
+ LongSize = DL->getPointerSizeInBits();
+ IntptrTy = Type::getIntNTy(*C, LongSize);
- for (Module::iterator F = M.begin(), E = M.end(); F != E; ++F) {
- if (F->isDeclaration()) continue;
- Res |= handleFunction(M, *F);
- }
+ AsanCtorFunction = Function::Create(
+ FunctionType::get(Type::getVoidTy(*C), false),
+ GlobalValue::InternalLinkage, kAsanModuleCtorName, &M);
+ BasicBlock *AsanCtorBB = BasicBlock::Create(*C, "", AsanCtorFunction);
+ // call __asan_init in the module ctor.
+ IRBuilder<> IRB(ReturnInst::Create(*C, AsanCtorBB));
+ AsanInitFunction = checkInterfaceFunction(
+ M.getOrInsertFunction(kAsanInitName, IRB.getVoidTy(), NULL));
+ AsanInitFunction->setLinkage(Function::ExternalLinkage);
+ IRB.CreateCall(AsanInitFunction);
- appendToGlobalCtors(M, AsanCtorFunction, kAsanCtorAndCtorPriority);
+ Mapping = getShadowMapping(M, LongSize);
- return Res;
+ appendToGlobalCtors(M, AsanCtorFunction, kAsanCtorAndDtorPriority);
+ return true;
}
bool AddressSanitizer::maybeInsertAsanInitAtFunctionEntry(Function &F) {
return false;
}
-bool AddressSanitizer::handleFunction(Module &M, Function &F) {
- if (BL->isIn(F)) return false;
+void AddressSanitizer::InjectCoverageAtBlock(Function &F, BasicBlock &BB) {
+ BasicBlock::iterator IP = BB.getFirstInsertionPt(), BE = BB.end();
+ // Skip static allocas at the top of the entry block so they don't become
+ // dynamic when we split the block. If we used our optimized stack layout,
+ // then there will only be one alloca and it will come first.
+ for (; IP != BE; ++IP) {
+ AllocaInst *AI = dyn_cast<AllocaInst>(IP);
+ if (!AI || !AI->isStaticAlloca())
+ break;
+ }
+
+ DebugLoc EntryLoc = &BB == &F.getEntryBlock()
+ ? IP->getDebugLoc().getFnDebugLoc(*C)
+ : IP->getDebugLoc();
+ IRBuilder<> IRB(IP);
+ IRB.SetCurrentDebugLocation(EntryLoc);
+ Type *Int8Ty = IRB.getInt8Ty();
+ GlobalVariable *Guard = new GlobalVariable(
+ *F.getParent(), Int8Ty, false, GlobalValue::PrivateLinkage,
+ Constant::getNullValue(Int8Ty), "__asan_gen_cov_" + F.getName());
+ LoadInst *Load = IRB.CreateLoad(Guard);
+ Load->setAtomic(Monotonic);
+ Load->setAlignment(1);
+ Value *Cmp = IRB.CreateICmpEQ(Constant::getNullValue(Int8Ty), Load);
+ Instruction *Ins = SplitBlockAndInsertIfThen(
+ Cmp, IP, false, MDBuilder(*C).createBranchWeights(1, 100000));
+ IRB.SetInsertPoint(Ins);
+ IRB.SetCurrentDebugLocation(EntryLoc);
+ // __sanitizer_cov gets the PC of the instruction using GET_CALLER_PC.
+ IRB.CreateCall(AsanCovFunction);
+ StoreInst *Store = IRB.CreateStore(ConstantInt::get(Int8Ty, 1), Guard);
+ Store->setAtomic(Monotonic);
+ Store->setAlignment(1);
+}
+
+// Poor man's coverage that works with ASan.
+// We create a Guard boolean variable with the same linkage
+// as the function and inject this code into the entry block (-asan-coverage=1)
+// or all blocks (-asan-coverage=2):
+// if (*Guard) {
+// __sanitizer_cov();
+// *Guard = 1;
+// }
+// The accesses to Guard are atomic. The rest of the logic is
+// in __sanitizer_cov (it's fine to call it more than once).
+//
+// This coverage implementation provides very limited data:
+// it only tells if a given function (block) was ever executed.
+// No counters, no per-edge data.
+// But for many use cases this is what we need and the added slowdown
+// is negligible. This simple implementation will probably be obsoleted
+// by the upcoming Clang-based coverage implementation.
+// By having it here and now we hope to
+// a) get the functionality to users earlier and
+// b) collect usage statistics to help improve Clang coverage design.
+bool AddressSanitizer::InjectCoverage(Function &F,
+ ArrayRef<BasicBlock *> AllBlocks,
+ ArrayRef<Instruction*> IndirCalls) {
+ if (!ClCoverage) return false;
+
+ if (ClCoverage == 1 ||
+ (unsigned)ClCoverageBlockThreshold < AllBlocks.size()) {
+ InjectCoverageAtBlock(F, F.getEntryBlock());
+ } else {
+ for (auto BB : AllBlocks)
+ InjectCoverageAtBlock(F, *BB);
+ }
+ InjectCoverageForIndirectCalls(F, IndirCalls);
+ return true;
+}
+
+// On every indirect call we call a run-time function
+// __sanitizer_cov_indir_call* with two parameters:
+// - callee address,
+// - global cache array that contains kCacheSize pointers (zero-initialed).
+// The cache is used to speed up recording the caller-callee pairs.
+// The address of the caller is passed implicitly via caller PC.
+// kCacheSize is encoded in the name of the run-time function.
+void AddressSanitizer::InjectCoverageForIndirectCalls(
+ Function &F, ArrayRef<Instruction *> IndirCalls) {
+ if (ClCoverage < 4 || IndirCalls.empty()) return;
+ const int kCacheSize = 16;
+ const int kCacheAlignment = 64; // Align for better performance.
+ Type *Ty = ArrayType::get(IntptrTy, kCacheSize);
+ GlobalVariable *CalleeCache =
+ new GlobalVariable(*F.getParent(), Ty, false, GlobalValue::PrivateLinkage,
+ Constant::getNullValue(Ty), "__asan_gen_callee_cache");
+ CalleeCache->setAlignment(kCacheAlignment);
+ for (auto I : IndirCalls) {
+ IRBuilder<> IRB(I);
+ CallSite CS(I);
+ IRB.CreateCall2(AsanCovIndirCallFunction,
+ IRB.CreatePointerCast(CS.getCalledValue(), IntptrTy),
+ IRB.CreatePointerCast(CalleeCache, IntptrTy));
+ }
+}
+
+bool AddressSanitizer::runOnFunction(Function &F) {
if (&F == AsanCtorFunction) return false;
+ if (F.getLinkage() == GlobalValue::AvailableExternallyLinkage) return false;
+ DEBUG(dbgs() << "ASAN instrumenting:\n" << F << "\n");
+ initializeCallbacks(*F.getParent());
- // If needed, insert __asan_init before checking for AddressSafety attr.
+ // If needed, insert __asan_init before checking for SanitizeAddress attr.
maybeInsertAsanInitAtFunctionEntry(F);
- if (!F.hasFnAttr(Attribute::AddressSafety)) return false;
+ if (!F.hasFnAttribute(Attribute::SanitizeAddress))
+ return false;
if (!ClDebugFunc.empty() && ClDebugFunc != F.getName())
return false;
- // We want to instrument every address only once per basic block
- // (unless there are calls between uses).
+
+ // We want to instrument every address only once per basic block (unless there
+ // are calls between uses).
SmallSet<Value*, 16> TempsToInstrument;
SmallVector<Instruction*, 16> ToInstrument;
SmallVector<Instruction*, 8> NoReturnCalls;
+ SmallVector<BasicBlock*, 16> AllBlocks;
+ SmallVector<Instruction*, 16> PointerComparisonsOrSubtracts;
+ SmallVector<Instruction*, 8> IndirCalls;
+ int NumAllocas = 0;
bool IsWrite;
+ unsigned Alignment;
// Fill the set of memory operations to instrument.
- for (Function::iterator FI = F.begin(), FE = F.end();
- FI != FE; ++FI) {
+ for (auto &BB : F) {
+ AllBlocks.push_back(&BB);
TempsToInstrument.clear();
int NumInsnsPerBB = 0;
- for (BasicBlock::iterator BI = FI->begin(), BE = FI->end();
- BI != BE; ++BI) {
- if (LooksLikeCodeInBug11395(BI)) return false;
- if (Value *Addr = isInterestingMemoryAccess(BI, &IsWrite)) {
+ for (auto &Inst : BB) {
+ if (LooksLikeCodeInBug11395(&Inst)) return false;
+ if (Value *Addr =
+ isInterestingMemoryAccess(&Inst, &IsWrite, &Alignment)) {
if (ClOpt && ClOptSameTemp) {
if (!TempsToInstrument.insert(Addr))
continue; // We've seen this temp in the current BB.
}
- } else if (isa<MemIntrinsic>(BI) && ClMemIntrin) {
+ } else if (ClInvalidPointerPairs &&
+ isInterestingPointerComparisonOrSubtraction(&Inst)) {
+ PointerComparisonsOrSubtracts.push_back(&Inst);
+ continue;
+ } else if (isa<MemIntrinsic>(Inst)) {
// ok, take it.
} else {
- if (CallInst *CI = dyn_cast<CallInst>(BI)) {
+ if (isa<AllocaInst>(Inst))
+ NumAllocas++;
+ CallSite CS(&Inst);
+ if (CS) {
// A call inside BB.
TempsToInstrument.clear();
- if (CI->doesNotReturn()) {
- NoReturnCalls.push_back(CI);
- }
+ if (CS.doesNotReturn())
+ NoReturnCalls.push_back(CS.getInstruction());
+ if (ClCoverage >= 4 && !CS.getCalledFunction())
+ IndirCalls.push_back(&Inst);
}
continue;
}
- ToInstrument.push_back(BI);
+ ToInstrument.push_back(&Inst);
NumInsnsPerBB++;
if (NumInsnsPerBB >= ClMaxInsnsToInstrumentPerBB)
break;
}
}
- AsanFunctionContext AFC(F);
+ Function *UninstrumentedDuplicate = nullptr;
+ bool LikelyToInstrument =
+ !NoReturnCalls.empty() || !ToInstrument.empty() || (NumAllocas > 0);
+ if (ClKeepUninstrumented && LikelyToInstrument) {
+ ValueToValueMapTy VMap;
+ UninstrumentedDuplicate = CloneFunction(&F, VMap, false);
+ UninstrumentedDuplicate->removeFnAttr(Attribute::SanitizeAddress);
+ UninstrumentedDuplicate->setName("NOASAN_" + F.getName());
+ F.getParent()->getFunctionList().push_back(UninstrumentedDuplicate);
+ }
+
+ bool UseCalls = false;
+ if (ClInstrumentationWithCallsThreshold >= 0 &&
+ ToInstrument.size() > (unsigned)ClInstrumentationWithCallsThreshold)
+ UseCalls = true;
// Instrument.
int NumInstrumented = 0;
- for (size_t i = 0, n = ToInstrument.size(); i != n; i++) {
- Instruction *Inst = ToInstrument[i];
+ for (auto Inst : ToInstrument) {
if (ClDebugMin < 0 || ClDebugMax < 0 ||
(NumInstrumented >= ClDebugMin && NumInstrumented <= ClDebugMax)) {
- if (isInterestingMemoryAccess(Inst, &IsWrite))
- instrumentMop(AFC, Inst);
+ if (isInterestingMemoryAccess(Inst, &IsWrite, &Alignment))
+ instrumentMop(Inst, UseCalls);
else
- instrumentMemIntrinsic(AFC, cast<MemIntrinsic>(Inst));
+ instrumentMemIntrinsic(cast<MemIntrinsic>(Inst));
}
NumInstrumented++;
}
- // Create PHI nodes and crash callbacks if we are merging crash callbacks.
- if (NumInstrumented) {
- for (size_t IsWrite = 0; IsWrite <= 1; IsWrite++) {
- for (size_t AccessSizeIndex = 0; AccessSizeIndex < kNumberOfAccessSizes;
- AccessSizeIndex++) {
- BasicBlock *BB = AFC.CrashBlock[IsWrite][AccessSizeIndex];
- if (!BB) continue;
- assert(ClMergeCallbacks);
- AsanFunctionContext::CrashArgsVec &Args =
- AFC.CrashArgs[IsWrite][AccessSizeIndex];
- IRBuilder<> IRB(BB->getFirstNonPHI());
- size_t n = Args.size();
- PHINode *PN1 = IRB.CreatePHI(IntptrTy, n);
- PHINode *PN2 = IRB.CreatePHI(IntptrTy, n);
- // We need to match crash parameters and the predecessors.
- for (pred_iterator PI = pred_begin(BB), PE = pred_end(BB);
- PI != PE; ++PI) {
- n--;
- PN1->addIncoming(Args[n].Arg1, *PI);
- PN2->addIncoming(Args[n].Arg2, *PI);
- }
- assert(n == 0);
- generateCrashCode(BB, PN1, PN2, IsWrite, AccessSizeIndex);
- }
- }
- }
-
- DEBUG(dbgs() << F);
-
- bool ChangedStack = poisonStackInFunction(M, F);
+ FunctionStackPoisoner FSP(F, *this);
+ bool ChangedStack = FSP.runOnFunction();
// We must unpoison the stack before every NoReturn call (throw, _exit, etc).
// See e.g. http://code.google.com/p/address-sanitizer/issues/detail?id=37
- for (size_t i = 0, n = NoReturnCalls.size(); i != n; i++) {
- Instruction *CI = NoReturnCalls[i];
+ for (auto CI : NoReturnCalls) {
IRBuilder<> IRB(CI);
- IRB.CreateCall(M.getOrInsertFunction(kAsanHandleNoReturnName,
- IRB.getVoidTy(), NULL));
+ IRB.CreateCall(AsanHandleNoReturnFunc);
}
- return NumInstrumented > 0 || ChangedStack || !NoReturnCalls.empty();
-}
+ for (auto Inst : PointerComparisonsOrSubtracts) {
+ instrumentPointerComparisonOrSubtraction(Inst);
+ NumInstrumented++;
+ }
-static uint64_t ValueForPoison(uint64_t PoisonByte, size_t ShadowRedzoneSize) {
- if (ShadowRedzoneSize == 1) return PoisonByte;
- if (ShadowRedzoneSize == 2) return (PoisonByte << 8) + PoisonByte;
- if (ShadowRedzoneSize == 4)
- return (PoisonByte << 24) + (PoisonByte << 16) +
- (PoisonByte << 8) + (PoisonByte);
- llvm_unreachable("ShadowRedzoneSize is either 1, 2 or 4");
-}
+ bool res = NumInstrumented > 0 || ChangedStack || !NoReturnCalls.empty();
+
+ if (InjectCoverage(F, AllBlocks, IndirCalls))
+ res = true;
-static void PoisonShadowPartialRightRedzone(uint8_t *Shadow,
- size_t Size,
- size_t RedzoneSize,
- size_t ShadowGranularity,
- uint8_t Magic) {
- for (size_t i = 0; i < RedzoneSize;
- i+= ShadowGranularity, Shadow++) {
- if (i + ShadowGranularity <= Size) {
- *Shadow = 0; // fully addressable
- } else if (i >= Size) {
- *Shadow = Magic; // unaddressable
+ DEBUG(dbgs() << "ASAN done instrumenting: " << res << " " << F << "\n");
+
+ if (ClKeepUninstrumented) {
+ if (!res) {
+ // No instrumentation is done, no need for the duplicate.
+ if (UninstrumentedDuplicate)
+ UninstrumentedDuplicate->eraseFromParent();
} else {
- *Shadow = Size - i; // first Size-i bytes are addressable
+ // The function was instrumented. We must have the duplicate.
+ assert(UninstrumentedDuplicate);
+ UninstrumentedDuplicate->setSection("NOASAN");
+ assert(!F.hasSection());
+ F.setSection("ASAN");
}
}
-}
-void AddressSanitizer::PoisonStack(const ArrayRef<AllocaInst*> &AllocaVec,
- IRBuilder<> IRB,
- Value *ShadowBase, bool DoPoison) {
- size_t ShadowRZSize = RedzoneSize >> MappingScale;
- assert(ShadowRZSize >= 1 && ShadowRZSize <= 4);
- Type *RZTy = Type::getIntNTy(*C, ShadowRZSize * 8);
- Type *RZPtrTy = PointerType::get(RZTy, 0);
-
- Value *PoisonLeft = ConstantInt::get(RZTy,
- ValueForPoison(DoPoison ? kAsanStackLeftRedzoneMagic : 0LL, ShadowRZSize));
- Value *PoisonMid = ConstantInt::get(RZTy,
- ValueForPoison(DoPoison ? kAsanStackMidRedzoneMagic : 0LL, ShadowRZSize));
- Value *PoisonRight = ConstantInt::get(RZTy,
- ValueForPoison(DoPoison ? kAsanStackRightRedzoneMagic : 0LL, ShadowRZSize));
-
- // poison the first red zone.
- IRB.CreateStore(PoisonLeft, IRB.CreateIntToPtr(ShadowBase, RZPtrTy));
-
- // poison all other red zones.
- uint64_t Pos = RedzoneSize;
- for (size_t i = 0, n = AllocaVec.size(); i < n; i++) {
- AllocaInst *AI = AllocaVec[i];
- uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
- uint64_t AlignedSize = getAlignedAllocaSize(AI);
- assert(AlignedSize - SizeInBytes < RedzoneSize);
- Value *Ptr = NULL;
-
- Pos += AlignedSize;
-
- assert(ShadowBase->getType() == IntptrTy);
- if (SizeInBytes < AlignedSize) {
- // Poison the partial redzone at right
- Ptr = IRB.CreateAdd(
- ShadowBase, ConstantInt::get(IntptrTy,
- (Pos >> MappingScale) - ShadowRZSize));
- size_t AddressableBytes = RedzoneSize - (AlignedSize - SizeInBytes);
- uint32_t Poison = 0;
- if (DoPoison) {
- PoisonShadowPartialRightRedzone((uint8_t*)&Poison, AddressableBytes,
- RedzoneSize,
- 1ULL << MappingScale,
- kAsanStackPartialRedzoneMagic);
- }
- Value *PartialPoison = ConstantInt::get(RZTy, Poison);
- IRB.CreateStore(PartialPoison, IRB.CreateIntToPtr(Ptr, RZPtrTy));
- }
-
- // Poison the full redzone at right.
- Ptr = IRB.CreateAdd(ShadowBase,
- ConstantInt::get(IntptrTy, Pos >> MappingScale));
- Value *Poison = i == AllocaVec.size() - 1 ? PoisonRight : PoisonMid;
- IRB.CreateStore(Poison, IRB.CreateIntToPtr(Ptr, RZPtrTy));
-
- Pos += RedzoneSize;
- }
+ return res;
}
// Workaround for bug 11395: we don't want to instrument stack in functions
return true;
}
-// Find all static Alloca instructions and put
-// poisoned red zones around all of them.
-// Then unpoison everything back before the function returns.
-//
-// Stack poisoning does not play well with exception handling.
-// When an exception is thrown, we essentially bypass the code
-// that unpoisones the stack. This is why the run-time library has
-// to intercept __cxa_throw (as well as longjmp, etc) and unpoison the entire
-// stack in the interceptor. This however does not work inside the
-// actual function which catches the exception. Most likely because the
-// compiler hoists the load of the shadow value somewhere too high.
-// This causes asan to report a non-existing bug on 453.povray.
-// It sounds like an LLVM bug.
-bool AddressSanitizer::poisonStackInFunction(Module &M, Function &F) {
- if (!ClStack) return false;
- SmallVector<AllocaInst*, 16> AllocaVec;
- SmallVector<Instruction*, 8> RetVec;
- uint64_t TotalSize = 0;
-
- // Filter out Alloca instructions we want (and can) handle.
- // Collect Ret instructions.
- for (Function::iterator FI = F.begin(), FE = F.end();
- FI != FE; ++FI) {
- BasicBlock &BB = *FI;
- for (BasicBlock::iterator BI = BB.begin(), BE = BB.end();
- BI != BE; ++BI) {
- if (isa<ReturnInst>(BI)) {
- RetVec.push_back(BI);
- continue;
- }
+void FunctionStackPoisoner::initializeCallbacks(Module &M) {
+ IRBuilder<> IRB(*C);
+ for (int i = 0; i <= kMaxAsanStackMallocSizeClass; i++) {
+ std::string Suffix = itostr(i);
+ AsanStackMallocFunc[i] = checkInterfaceFunction(
+ M.getOrInsertFunction(kAsanStackMallocNameTemplate + Suffix, IntptrTy,
+ IntptrTy, IntptrTy, NULL));
+ AsanStackFreeFunc[i] = checkInterfaceFunction(M.getOrInsertFunction(
+ kAsanStackFreeNameTemplate + Suffix, IRB.getVoidTy(), IntptrTy,
+ IntptrTy, IntptrTy, NULL));
+ }
+ AsanPoisonStackMemoryFunc = checkInterfaceFunction(M.getOrInsertFunction(
+ kAsanPoisonStackMemoryName, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
+ AsanUnpoisonStackMemoryFunc = checkInterfaceFunction(M.getOrInsertFunction(
+ kAsanUnpoisonStackMemoryName, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
+}
- AllocaInst *AI = dyn_cast<AllocaInst>(BI);
- if (!AI) continue;
- if (AI->isArrayAllocation()) continue;
- if (!AI->isStaticAlloca()) continue;
- if (!AI->getAllocatedType()->isSized()) continue;
- if (AI->getAlignment() > RedzoneSize) continue;
- AllocaVec.push_back(AI);
- uint64_t AlignedSize = getAlignedAllocaSize(AI);
- TotalSize += AlignedSize;
+void
+FunctionStackPoisoner::poisonRedZones(ArrayRef<uint8_t> ShadowBytes,
+ IRBuilder<> &IRB, Value *ShadowBase,
+ bool DoPoison) {
+ size_t n = ShadowBytes.size();
+ size_t i = 0;
+ // We need to (un)poison n bytes of stack shadow. Poison as many as we can
+ // using 64-bit stores (if we are on 64-bit arch), then poison the rest
+ // with 32-bit stores, then with 16-byte stores, then with 8-byte stores.
+ for (size_t LargeStoreSizeInBytes = ASan.LongSize / 8;
+ LargeStoreSizeInBytes != 0; LargeStoreSizeInBytes /= 2) {
+ for (; i + LargeStoreSizeInBytes - 1 < n; i += LargeStoreSizeInBytes) {
+ uint64_t Val = 0;
+ for (size_t j = 0; j < LargeStoreSizeInBytes; j++) {
+ if (ASan.DL->isLittleEndian())
+ Val |= (uint64_t)ShadowBytes[i + j] << (8 * j);
+ else
+ Val = (Val << 8) | ShadowBytes[i + j];
+ }
+ if (!Val) continue;
+ Value *Ptr = IRB.CreateAdd(ShadowBase, ConstantInt::get(IntptrTy, i));
+ Type *StoreTy = Type::getIntNTy(*C, LargeStoreSizeInBytes * 8);
+ Value *Poison = ConstantInt::get(StoreTy, DoPoison ? Val : 0);
+ IRB.CreateStore(Poison, IRB.CreateIntToPtr(Ptr, StoreTy->getPointerTo()));
}
}
+}
+
+// Fake stack allocator (asan_fake_stack.h) has 11 size classes
+// for every power of 2 from kMinStackMallocSize to kMaxAsanStackMallocSizeClass
+static int StackMallocSizeClass(uint64_t LocalStackSize) {
+ assert(LocalStackSize <= kMaxStackMallocSize);
+ uint64_t MaxSize = kMinStackMallocSize;
+ for (int i = 0; ; i++, MaxSize *= 2)
+ if (LocalStackSize <= MaxSize)
+ return i;
+ llvm_unreachable("impossible LocalStackSize");
+}
- if (AllocaVec.empty()) return false;
+// Set Size bytes starting from ShadowBase to kAsanStackAfterReturnMagic.
+// We can not use MemSet intrinsic because it may end up calling the actual
+// memset. Size is a multiple of 8.
+// Currently this generates 8-byte stores on x86_64; it may be better to
+// generate wider stores.
+void FunctionStackPoisoner::SetShadowToStackAfterReturnInlined(
+ IRBuilder<> &IRB, Value *ShadowBase, int Size) {
+ assert(!(Size % 8));
+ assert(kAsanStackAfterReturnMagic == 0xf5);
+ for (int i = 0; i < Size; i += 8) {
+ Value *p = IRB.CreateAdd(ShadowBase, ConstantInt::get(IntptrTy, i));
+ IRB.CreateStore(ConstantInt::get(IRB.getInt64Ty(), 0xf5f5f5f5f5f5f5f5ULL),
+ IRB.CreateIntToPtr(p, IRB.getInt64Ty()->getPointerTo()));
+ }
+}
- uint64_t LocalStackSize = TotalSize + (AllocaVec.size() + 1) * RedzoneSize;
+static DebugLoc getFunctionEntryDebugLocation(Function &F) {
+ for (const auto &Inst : F.getEntryBlock())
+ if (!isa<AllocaInst>(Inst))
+ return Inst.getDebugLoc();
+ return DebugLoc();
+}
- bool DoStackMalloc = ClUseAfterReturn
- && LocalStackSize <= kMaxStackMallocSize;
+void FunctionStackPoisoner::poisonStack() {
+ int StackMallocIdx = -1;
+ DebugLoc EntryDebugLocation = getFunctionEntryDebugLocation(F);
+ assert(AllocaVec.size() > 0);
Instruction *InsBefore = AllocaVec[0];
IRBuilder<> IRB(InsBefore);
-
+ IRB.SetCurrentDebugLocation(EntryDebugLocation);
+
+ SmallVector<ASanStackVariableDescription, 16> SVD;
+ SVD.reserve(AllocaVec.size());
+ for (AllocaInst *AI : AllocaVec) {
+ ASanStackVariableDescription D = { AI->getName().data(),
+ getAllocaSizeInBytes(AI),
+ AI->getAlignment(), AI, 0};
+ SVD.push_back(D);
+ }
+ // Minimal header size (left redzone) is 4 pointers,
+ // i.e. 32 bytes on 64-bit platforms and 16 bytes in 32-bit platforms.
+ size_t MinHeaderSize = ASan.LongSize / 2;
+ ASanStackFrameLayout L;
+ ComputeASanStackFrameLayout(SVD, 1UL << Mapping.Scale, MinHeaderSize, &L);
+ DEBUG(dbgs() << L.DescriptionString << " --- " << L.FrameSize << "\n");
+ uint64_t LocalStackSize = L.FrameSize;
+ bool DoStackMalloc =
+ ClUseAfterReturn && LocalStackSize <= kMaxStackMallocSize;
Type *ByteArrayTy = ArrayType::get(IRB.getInt8Ty(), LocalStackSize);
AllocaInst *MyAlloca =
new AllocaInst(ByteArrayTy, "MyAlloca", InsBefore);
- MyAlloca->setAlignment(RedzoneSize);
+ MyAlloca->setDebugLoc(EntryDebugLocation);
+ assert((ClRealignStack & (ClRealignStack - 1)) == 0);
+ size_t FrameAlignment = std::max(L.FrameAlignment, (size_t)ClRealignStack);
+ MyAlloca->setAlignment(FrameAlignment);
assert(MyAlloca->isStaticAlloca());
Value *OrigStackBase = IRB.CreatePointerCast(MyAlloca, IntptrTy);
Value *LocalStackBase = OrigStackBase;
if (DoStackMalloc) {
- Value *AsanStackMallocFunc = M.getOrInsertFunction(
- kAsanStackMallocName, IntptrTy, IntptrTy, IntptrTy, NULL);
- LocalStackBase = IRB.CreateCall2(AsanStackMallocFunc,
+ // LocalStackBase = OrigStackBase
+ // if (__asan_option_detect_stack_use_after_return)
+ // LocalStackBase = __asan_stack_malloc_N(LocalStackBase, OrigStackBase);
+ StackMallocIdx = StackMallocSizeClass(LocalStackSize);
+ assert(StackMallocIdx <= kMaxAsanStackMallocSizeClass);
+ Constant *OptionDetectUAR = F.getParent()->getOrInsertGlobal(
+ kAsanOptionDetectUAR, IRB.getInt32Ty());
+ Value *Cmp = IRB.CreateICmpNE(IRB.CreateLoad(OptionDetectUAR),
+ Constant::getNullValue(IRB.getInt32Ty()));
+ Instruction *Term = SplitBlockAndInsertIfThen(Cmp, InsBefore, false);
+ BasicBlock *CmpBlock = cast<Instruction>(Cmp)->getParent();
+ IRBuilder<> IRBIf(Term);
+ IRBIf.SetCurrentDebugLocation(EntryDebugLocation);
+ LocalStackBase = IRBIf.CreateCall2(
+ AsanStackMallocFunc[StackMallocIdx],
ConstantInt::get(IntptrTy, LocalStackSize), OrigStackBase);
+ BasicBlock *SetBlock = cast<Instruction>(LocalStackBase)->getParent();
+ IRB.SetInsertPoint(InsBefore);
+ IRB.SetCurrentDebugLocation(EntryDebugLocation);
+ PHINode *Phi = IRB.CreatePHI(IntptrTy, 2);
+ Phi->addIncoming(OrigStackBase, CmpBlock);
+ Phi->addIncoming(LocalStackBase, SetBlock);
+ LocalStackBase = Phi;
}
- // This string will be parsed by the run-time (DescribeStackAddress).
- SmallString<2048> StackDescriptionStorage;
- raw_svector_ostream StackDescription(StackDescriptionStorage);
- StackDescription << F.getName() << " " << AllocaVec.size() << " ";
+ // Insert poison calls for lifetime intrinsics for alloca.
+ bool HavePoisonedAllocas = false;
+ for (const auto &APC : AllocaPoisonCallVec) {
+ assert(APC.InsBefore);
+ assert(APC.AI);
+ IRBuilder<> IRB(APC.InsBefore);
+ poisonAlloca(APC.AI, APC.Size, IRB, APC.DoPoison);
+ HavePoisonedAllocas |= APC.DoPoison;
+ }
- uint64_t Pos = RedzoneSize;
// Replace Alloca instructions with base+offset.
- for (size_t i = 0, n = AllocaVec.size(); i < n; i++) {
- AllocaInst *AI = AllocaVec[i];
- uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
- StringRef Name = AI->getName();
- StackDescription << Pos << " " << SizeInBytes << " "
- << Name.size() << " " << Name << " ";
- uint64_t AlignedSize = getAlignedAllocaSize(AI);
- assert((AlignedSize % RedzoneSize) == 0);
- AI->replaceAllUsesWith(
- IRB.CreateIntToPtr(
- IRB.CreateAdd(LocalStackBase, ConstantInt::get(IntptrTy, Pos)),
- AI->getType()));
- Pos += AlignedSize + RedzoneSize;
+ for (const auto &Desc : SVD) {
+ AllocaInst *AI = Desc.AI;
+ Value *NewAllocaPtr = IRB.CreateIntToPtr(
+ IRB.CreateAdd(LocalStackBase, ConstantInt::get(IntptrTy, Desc.Offset)),
+ AI->getType());
+ replaceDbgDeclareForAlloca(AI, NewAllocaPtr, DIB);
+ AI->replaceAllUsesWith(NewAllocaPtr);
}
- assert(Pos == LocalStackSize);
- // Write the Magic value and the frame description constant to the redzone.
+ // The left-most redzone has enough space for at least 4 pointers.
+ // Write the Magic value to redzone[0].
Value *BasePlus0 = IRB.CreateIntToPtr(LocalStackBase, IntptrPtrTy);
IRB.CreateStore(ConstantInt::get(IntptrTy, kCurrentStackFrameMagic),
BasePlus0);
- Value *BasePlus1 = IRB.CreateAdd(LocalStackBase,
- ConstantInt::get(IntptrTy, LongSize/8));
- BasePlus1 = IRB.CreateIntToPtr(BasePlus1, IntptrPtrTy);
- Value *Description = IRB.CreatePointerCast(
- createPrivateGlobalForString(M, StackDescription.str()),
- IntptrTy);
+ // Write the frame description constant to redzone[1].
+ Value *BasePlus1 = IRB.CreateIntToPtr(
+ IRB.CreateAdd(LocalStackBase, ConstantInt::get(IntptrTy, ASan.LongSize/8)),
+ IntptrPtrTy);
+ GlobalVariable *StackDescriptionGlobal =
+ createPrivateGlobalForString(*F.getParent(), L.DescriptionString,
+ /*AllowMerging*/true);
+ Value *Description = IRB.CreatePointerCast(StackDescriptionGlobal,
+ IntptrTy);
IRB.CreateStore(Description, BasePlus1);
+ // Write the PC to redzone[2].
+ Value *BasePlus2 = IRB.CreateIntToPtr(
+ IRB.CreateAdd(LocalStackBase, ConstantInt::get(IntptrTy,
+ 2 * ASan.LongSize/8)),
+ IntptrPtrTy);
+ IRB.CreateStore(IRB.CreatePointerCast(&F, IntptrTy), BasePlus2);
// Poison the stack redzones at the entry.
- Value *ShadowBase = memToShadow(LocalStackBase, IRB);
- PoisonStack(ArrayRef<AllocaInst*>(AllocaVec), IRB, ShadowBase, true);
+ Value *ShadowBase = ASan.memToShadow(LocalStackBase, IRB);
+ poisonRedZones(L.ShadowBytes, IRB, ShadowBase, true);
- Value *AsanStackFreeFunc = NULL;
- if (DoStackMalloc) {
- AsanStackFreeFunc = M.getOrInsertFunction(
- kAsanStackFreeName, IRB.getVoidTy(),
- IntptrTy, IntptrTy, IntptrTy, NULL);
- }
-
- // Unpoison the stack before all ret instructions.
- for (size_t i = 0, n = RetVec.size(); i < n; i++) {
- Instruction *Ret = RetVec[i];
+ // (Un)poison the stack before all ret instructions.
+ for (auto Ret : RetVec) {
IRBuilder<> IRBRet(Ret);
-
// Mark the current frame as retired.
IRBRet.CreateStore(ConstantInt::get(IntptrTy, kRetiredStackFrameMagic),
BasePlus0);
- // Unpoison the stack.
- PoisonStack(ArrayRef<AllocaInst*>(AllocaVec), IRBRet, ShadowBase, false);
-
if (DoStackMalloc) {
- IRBRet.CreateCall3(AsanStackFreeFunc, LocalStackBase,
- ConstantInt::get(IntptrTy, LocalStackSize),
- OrigStackBase);
+ assert(StackMallocIdx >= 0);
+ // if LocalStackBase != OrigStackBase:
+ // // In use-after-return mode, poison the whole stack frame.
+ // if StackMallocIdx <= 4
+ // // For small sizes inline the whole thing:
+ // memset(ShadowBase, kAsanStackAfterReturnMagic, ShadowSize);
+ // **SavedFlagPtr(LocalStackBase) = 0
+ // else
+ // __asan_stack_free_N(LocalStackBase, OrigStackBase)
+ // else
+ // <This is not a fake stack; unpoison the redzones>
+ Value *Cmp = IRBRet.CreateICmpNE(LocalStackBase, OrigStackBase);
+ TerminatorInst *ThenTerm, *ElseTerm;
+ SplitBlockAndInsertIfThenElse(Cmp, Ret, &ThenTerm, &ElseTerm);
+
+ IRBuilder<> IRBPoison(ThenTerm);
+ if (StackMallocIdx <= 4) {
+ int ClassSize = kMinStackMallocSize << StackMallocIdx;
+ SetShadowToStackAfterReturnInlined(IRBPoison, ShadowBase,
+ ClassSize >> Mapping.Scale);
+ Value *SavedFlagPtrPtr = IRBPoison.CreateAdd(
+ LocalStackBase,
+ ConstantInt::get(IntptrTy, ClassSize - ASan.LongSize / 8));
+ Value *SavedFlagPtr = IRBPoison.CreateLoad(
+ IRBPoison.CreateIntToPtr(SavedFlagPtrPtr, IntptrPtrTy));
+ IRBPoison.CreateStore(
+ Constant::getNullValue(IRBPoison.getInt8Ty()),
+ IRBPoison.CreateIntToPtr(SavedFlagPtr, IRBPoison.getInt8PtrTy()));
+ } else {
+ // For larger frames call __asan_stack_free_*.
+ IRBPoison.CreateCall3(AsanStackFreeFunc[StackMallocIdx], LocalStackBase,
+ ConstantInt::get(IntptrTy, LocalStackSize),
+ OrigStackBase);
+ }
+
+ IRBuilder<> IRBElse(ElseTerm);
+ poisonRedZones(L.ShadowBytes, IRBElse, ShadowBase, false);
+ } else if (HavePoisonedAllocas) {
+ // If we poisoned some allocas in llvm.lifetime analysis,
+ // unpoison whole stack frame now.
+ assert(LocalStackBase == OrigStackBase);
+ poisonAlloca(LocalStackBase, LocalStackSize, IRBRet, false);
+ } else {
+ poisonRedZones(L.ShadowBytes, IRBRet, ShadowBase, false);
}
}
- if (ClDebugStack) {
- DEBUG(dbgs() << F);
- }
+ // We are done. Remove the old unused alloca instructions.
+ for (auto AI : AllocaVec)
+ AI->eraseFromParent();
+}
- return true;
+void FunctionStackPoisoner::poisonAlloca(Value *V, uint64_t Size,
+ IRBuilder<> &IRB, bool DoPoison) {
+ // For now just insert the call to ASan runtime.
+ Value *AddrArg = IRB.CreatePointerCast(V, IntptrTy);
+ Value *SizeArg = ConstantInt::get(IntptrTy, Size);
+ IRB.CreateCall2(DoPoison ? AsanPoisonStackMemoryFunc
+ : AsanUnpoisonStackMemoryFunc,
+ AddrArg, SizeArg);
+}
+
+// Handling llvm.lifetime intrinsics for a given %alloca:
+// (1) collect all llvm.lifetime.xxx(%size, %value) describing the alloca.
+// (2) if %size is constant, poison memory for llvm.lifetime.end (to detect
+// invalid accesses) and unpoison it for llvm.lifetime.start (the memory
+// could be poisoned by previous llvm.lifetime.end instruction, as the
+// variable may go in and out of scope several times, e.g. in loops).
+// (3) if we poisoned at least one %alloca in a function,
+// unpoison the whole stack frame at function exit.
+
+AllocaInst *FunctionStackPoisoner::findAllocaForValue(Value *V) {
+ if (AllocaInst *AI = dyn_cast<AllocaInst>(V))
+ // We're intested only in allocas we can handle.
+ return isInterestingAlloca(*AI) ? AI : nullptr;
+ // See if we've already calculated (or started to calculate) alloca for a
+ // given value.
+ AllocaForValueMapTy::iterator I = AllocaForValue.find(V);
+ if (I != AllocaForValue.end())
+ return I->second;
+ // Store 0 while we're calculating alloca for value V to avoid
+ // infinite recursion if the value references itself.
+ AllocaForValue[V] = nullptr;
+ AllocaInst *Res = nullptr;
+ if (CastInst *CI = dyn_cast<CastInst>(V))
+ Res = findAllocaForValue(CI->getOperand(0));
+ else if (PHINode *PN = dyn_cast<PHINode>(V)) {
+ for (unsigned i = 0, e = PN->getNumIncomingValues(); i != e; ++i) {
+ Value *IncValue = PN->getIncomingValue(i);
+ // Allow self-referencing phi-nodes.
+ if (IncValue == PN) continue;
+ AllocaInst *IncValueAI = findAllocaForValue(IncValue);
+ // AI for incoming values should exist and should all be equal.
+ if (IncValueAI == nullptr || (Res != nullptr && IncValueAI != Res))
+ return nullptr;
+ Res = IncValueAI;
+ }
+ }
+ if (Res)
+ AllocaForValue[V] = Res;
+ return Res;
}