lib/Transforms/Utils/ModuleUtils.cpp

   1 //===-- ModuleUtils.cpp - Functions to manipulate Modules -----------------===//
   2 //
   3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
   4 // See https://llvm.org/LICENSE.txt for license information.
   5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
   6 //
   7 //===----------------------------------------------------------------------===//
   8 //
   9 // This family of functions perform manipulations on Modules.
  10 //
  11 //===----------------------------------------------------------------------===//
  12
  13 #include "llvm/Transforms/Utils/ModuleUtils.h"
  14 #include "llvm/IR/DerivedTypes.h"
  15 #include "llvm/IR/Function.h"
  16 #include "llvm/IR/IRBuilder.h"
  17 #include "llvm/IR/Module.h"
  18 #include "llvm/Support/raw_ostream.h"
  19
  20 using namespace llvm;
  21
  22 static void appendToGlobalArray(const char *Array, Module &M, Function *F,
  23                                 int Priority, Constant *Data) {
  24   IRBuilder<> IRB(M.getContext());
  25   FunctionType *FnTy = FunctionType::get(IRB.getVoidTy(), false);
  26
  27   // Get the current set of static global constructors and add the new ctor
  28   // to the list.
  29   SmallVector<Constant *, 16> CurrentCtors;
  30   StructType *EltTy = StructType::get(
  31       IRB.getInt32Ty(), PointerType::getUnqual(FnTy), IRB.getInt8PtrTy());
  32   if (GlobalVariable *GVCtor = M.getNamedGlobal(Array)) {
  33     if (Constant *Init = GVCtor->getInitializer()) {
  34       unsigned n = Init->getNumOperands();
  35       CurrentCtors.reserve(n + 1);
  36       for (unsigned i = 0; i != n; ++i)
  37         CurrentCtors.push_back(cast<Constant>(Init->getOperand(i)));
  38     }
  39     GVCtor->eraseFromParent();
  40   }
  41
  42   // Build a 3 field global_ctor entry.  We don't take a comdat key.
  43   Constant *CSVals[3];
  44   CSVals[0] = IRB.getInt32(Priority);
  45   CSVals[1] = F;
  46   CSVals[2] = Data ? ConstantExpr::getPointerCast(Data, IRB.getInt8PtrTy())
  47                    : Constant::getNullValue(IRB.getInt8PtrTy());
  48   Constant *RuntimeCtorInit =
  49       ConstantStruct::get(EltTy, makeArrayRef(CSVals, EltTy->getNumElements()));
  50
  51   CurrentCtors.push_back(RuntimeCtorInit);
  52
  53   // Create a new initializer.
  54   ArrayType *AT = ArrayType::get(EltTy, CurrentCtors.size());
  55   Constant *NewInit = ConstantArray::get(AT, CurrentCtors);
  56
  57   // Create the new global variable and replace all uses of
  58   // the old global variable with the new one.
  59   (void)new GlobalVariable(M, NewInit->getType(), false,
  60                            GlobalValue::AppendingLinkage, NewInit, Array);
  61 }
  62
  63 void llvm::appendToGlobalCtors(Module &M, Function *F, int Priority, Constant *Data) {
  64   appendToGlobalArray("llvm.global_ctors", M, F, Priority, Data);
  65 }
  66
  67 void llvm::appendToGlobalDtors(Module &M, Function *F, int Priority, Constant *Data) {
  68   appendToGlobalArray("llvm.global_dtors", M, F, Priority, Data);
  69 }
  70
  71 static void appendToUsedList(Module &M, StringRef Name, ArrayRef<GlobalValue *> Values) {
  72   GlobalVariable *GV = M.getGlobalVariable(Name);
  73   SmallPtrSet<Constant *, 16> InitAsSet;
  74   SmallVector<Constant *, 16> Init;
  75   if (GV) {
  76     ConstantArray *CA = dyn_cast<ConstantArray>(GV->getInitializer());
  77     for (auto &Op : CA->operands()) {
  78       Constant *C = cast_or_null<Constant>(Op);
  79       if (InitAsSet.insert(C).second)
  80         Init.push_back(C);
  81     }
  82     GV->eraseFromParent();
  83   }
  84
  85   Type *Int8PtrTy = llvm::Type::getInt8PtrTy(M.getContext());
  86   for (auto *V : Values) {
  87     Constant *C = ConstantExpr::getBitCast(V, Int8PtrTy);
  88     if (InitAsSet.insert(C).second)
  89       Init.push_back(C);
  90   }
  91
  92   if (Init.empty())
  93     return;
  94
  95   ArrayType *ATy = ArrayType::get(Int8PtrTy, Init.size());
  96   GV = new llvm::GlobalVariable(M, ATy, false, GlobalValue::AppendingLinkage,
  97                                 ConstantArray::get(ATy, Init), Name);
  98   GV->setSection("llvm.metadata");
  99 }
 100
 101 void llvm::appendToUsed(Module &M, ArrayRef<GlobalValue *> Values) {
 102   appendToUsedList(M, "llvm.used", Values);
 103 }
 104
 105 void llvm::appendToCompilerUsed(Module &M, ArrayRef<GlobalValue *> Values) {
 106   appendToUsedList(M, "llvm.compiler.used", Values);
 107 }
 108
 109 FunctionCallee
 110 llvm::declareSanitizerInitFunction(Module &M, StringRef InitName,
 111                                    ArrayRef<Type *> InitArgTypes) {
 112   assert(!InitName.empty() && "Expected init function name");
 113   return M.getOrInsertFunction(
 114       InitName,
 115       FunctionType::get(Type::getVoidTy(M.getContext()), InitArgTypes, false),
 116       AttributeList());
 117 }
 118
 119 std::pair<Function *, FunctionCallee> llvm::createSanitizerCtorAndInitFunctions(
 120     Module &M, StringRef CtorName, StringRef InitName,
 121     ArrayRef<Type *> InitArgTypes, ArrayRef<Value *> InitArgs,
 122     StringRef VersionCheckName) {
 123   assert(!InitName.empty() && "Expected init function name");
 124   assert(InitArgs.size() == InitArgTypes.size() &&
 125          "Sanitizer's init function expects different number of arguments");
 126   FunctionCallee InitFunction =
 127       declareSanitizerInitFunction(M, InitName, InitArgTypes);
 128   Function *Ctor = Function::Create(
 129       FunctionType::get(Type::getVoidTy(M.getContext()), false),
 130       GlobalValue::InternalLinkage, CtorName, &M);
 131   BasicBlock *CtorBB = BasicBlock::Create(M.getContext(), "", Ctor);
 132   IRBuilder<> IRB(ReturnInst::Create(M.getContext(), CtorBB));
 133   IRB.CreateCall(InitFunction, InitArgs);
 134   if (!VersionCheckName.empty()) {
 135     FunctionCallee VersionCheckFunction = M.getOrInsertFunction(
 136         VersionCheckName, FunctionType::get(IRB.getVoidTy(), {}, false),
 137         AttributeList());
 138     IRB.CreateCall(VersionCheckFunction, {});
 139   }
 140   return std::make_pair(Ctor, InitFunction);
 141 }
 142
 143 std::pair<Function *, FunctionCallee>
 144 llvm::getOrCreateSanitizerCtorAndInitFunctions(
 145     Module &M, StringRef CtorName, StringRef InitName,
 146     ArrayRef<Type *> InitArgTypes, ArrayRef<Value *> InitArgs,
 147     function_ref<void(Function *, FunctionCallee)> FunctionsCreatedCallback,
 148     StringRef VersionCheckName) {
 149   assert(!CtorName.empty() && "Expected ctor function name");
 150
 151   if (Function *Ctor = M.getFunction(CtorName))
 152     // FIXME: Sink this logic into the module, similar to the handling of
 153     // globals. This will make moving to a concurrent model much easier.
 154     if (Ctor->arg_size() == 0 ||
 155         Ctor->getReturnType() == Type::getVoidTy(M.getContext()))
 156       return {Ctor, declareSanitizerInitFunction(M, InitName, InitArgTypes)};
 157
 158   Function *Ctor;
 159   FunctionCallee InitFunction;
 160   std::tie(Ctor, InitFunction) = llvm::createSanitizerCtorAndInitFunctions(
 161       M, CtorName, InitName, InitArgTypes, InitArgs, VersionCheckName);
 162   FunctionsCreatedCallback(Ctor, InitFunction);
 163   return std::make_pair(Ctor, InitFunction);
 164 }
 165
 166 Function *llvm::getOrCreateInitFunction(Module &M, StringRef Name) {
 167   assert(!Name.empty() && "Expected init function name");
 168   if (Function *F = M.getFunction(Name)) {
 169     if (F->arg_size() != 0 ||
 170         F->getReturnType() != Type::getVoidTy(M.getContext())) {
 171       std::string Err;
 172       raw_string_ostream Stream(Err);
 173       Stream << "Sanitizer interface function defined with wrong type: " << *F;
 174       report_fatal_error(Err);
 175     }
 176     return F;
 177   }
 178   Function *F =
 179       cast<Function>(M.getOrInsertFunction(Name, AttributeList(),
 180                                            Type::getVoidTy(M.getContext()))
 181                          .getCallee());
 182
 183   appendToGlobalCtors(M, F, 0);
 184
 185   return F;
 186 }
 187
 188 void llvm::filterDeadComdatFunctions(
 189     Module &M, SmallVectorImpl<Function *> &DeadComdatFunctions) {
 190   // Build a map from the comdat to the number of entries in that comdat we
 191   // think are dead. If this fully covers the comdat group, then the entire
 192   // group is dead. If we find another entry in the comdat group though, we'll
 193   // have to preserve the whole group.
 194   SmallDenseMap<Comdat *, int, 16> ComdatEntriesCovered;
 195   for (Function *F : DeadComdatFunctions) {
 196     Comdat *C = F->getComdat();
 197     assert(C && "Expected all input GVs to be in a comdat!");
 198     ComdatEntriesCovered[C] += 1;
 199   }
 200
 201   auto CheckComdat = [&](Comdat &C) {
 202     auto CI = ComdatEntriesCovered.find(&C);
 203     if (CI == ComdatEntriesCovered.end())
 204       return;
 205
 206     // If this could have been covered by a dead entry, just subtract one to
 207     // account for it.
 208     if (CI->second > 0) {
 209       CI->second -= 1;
 210       return;
 211     }
 212
 213     // If we've already accounted for all the entries that were dead, the
 214     // entire comdat is alive so remove it from the map.
 215     ComdatEntriesCovered.erase(CI);
 216   };
 217
 218   auto CheckAllComdats = [&] {
 219     for (Function &F : M.functions())
 220       if (Comdat *C = F.getComdat()) {
 221         CheckComdat(*C);
 222         if (ComdatEntriesCovered.empty())
 223           return;
 224       }
 225     for (GlobalVariable &GV : M.globals())
 226       if (Comdat *C = GV.getComdat()) {
 227         CheckComdat(*C);
 228         if (ComdatEntriesCovered.empty())
 229           return;
 230       }
 231     for (GlobalAlias &GA : M.aliases())
 232       if (Comdat *C = GA.getComdat()) {
 233         CheckComdat(*C);
 234         if (ComdatEntriesCovered.empty())
 235           return;
 236       }
 237   };
 238   CheckAllComdats();
 239
 240   if (ComdatEntriesCovered.empty()) {
 241     DeadComdatFunctions.clear();
 242     return;
 243   }
 244
 245   // Remove the entries that were not covering.
 246   erase_if(DeadComdatFunctions, [&](GlobalValue *GV) {
 247     return ComdatEntriesCovered.find(GV->getComdat()) ==
 248            ComdatEntriesCovered.end();
 249   });
 250 }
 251
 252 std::string llvm::getUniqueModuleId(Module *M) {
 253   MD5 Md5;
 254   bool ExportsSymbols = false;
 255   auto AddGlobal = [&](GlobalValue &GV) {
 256     if (GV.isDeclaration() || GV.getName().startswith("llvm.") ||
 257         !GV.hasExternalLinkage() || GV.hasComdat())
 258       return;
 259     ExportsSymbols = true;
 260     Md5.update(GV.getName());
 261     Md5.update(ArrayRef<uint8_t>{0});
 262   };
 263
 264   for (auto &F : *M)
 265     AddGlobal(F);
 266   for (auto &GV : M->globals())
 267     AddGlobal(GV);
 268   for (auto &GA : M->aliases())
 269     AddGlobal(GA);
 270   for (auto &IF : M->ifuncs())
 271     AddGlobal(IF);
 272
 273   if (!ExportsSymbols)
 274     return "";
 275
 276   MD5::MD5Result R;
 277   Md5.final(R);
 278
 279   SmallString<32> Str;
 280   MD5::stringifyResult(R, Str);
 281   return ("$" + Str).str();
 282 }