llvm/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/Analysis/TargetLibraryInfo.h"
  15 #include "llvm/Analysis/VectorUtils.h"
  16 #include "llvm/IR/DerivedTypes.h"
  17 #include "llvm/IR/Function.h"
  18 #include "llvm/IR/IRBuilder.h"
  19 #include "llvm/IR/Module.h"
  20 #include "llvm/Support/raw_ostream.h"
  21 using namespace llvm;
  22
  23 #define DEBUG_TYPE "moduleutils"
  24
  25 static void appendToGlobalArray(const char *Array, Module &M, Function *F,
  26                                 int Priority, Constant *Data) {
  27   IRBuilder<> IRB(M.getContext());
  28   FunctionType *FnTy = FunctionType::get(IRB.getVoidTy(), false);
  29
  30   // Get the current set of static global constructors and add the new ctor
  31   // to the list.
  32   SmallVector<Constant *, 16> CurrentCtors;
  33   StructType *EltTy = StructType::get(
  34       IRB.getInt32Ty(), PointerType::getUnqual(FnTy), IRB.getInt8PtrTy());
  35   if (GlobalVariable *GVCtor = M.getNamedGlobal(Array)) {
  36     if (Constant *Init = GVCtor->getInitializer()) {
  37       unsigned n = Init->getNumOperands();
  38       CurrentCtors.reserve(n + 1);
  39       for (unsigned i = 0; i != n; ++i)
  40         CurrentCtors.push_back(cast<Constant>(Init->getOperand(i)));
  41     }
  42     GVCtor->eraseFromParent();
  43   }
  44
  45   // Build a 3 field global_ctor entry.  We don't take a comdat key.
  46   Constant *CSVals[3];
  47   CSVals[0] = IRB.getInt32(Priority);
  48   CSVals[1] = F;
  49   CSVals[2] = Data ? ConstantExpr::getPointerCast(Data, IRB.getInt8PtrTy())
  50                    : Constant::getNullValue(IRB.getInt8PtrTy());
  51   Constant *RuntimeCtorInit =
  52       ConstantStruct::get(EltTy, makeArrayRef(CSVals, EltTy->getNumElements()));
  53
  54   CurrentCtors.push_back(RuntimeCtorInit);
  55
  56   // Create a new initializer.
  57   ArrayType *AT = ArrayType::get(EltTy, CurrentCtors.size());
  58   Constant *NewInit = ConstantArray::get(AT, CurrentCtors);
  59
  60   // Create the new global variable and replace all uses of
  61   // the old global variable with the new one.
  62   (void)new GlobalVariable(M, NewInit->getType(), false,
  63                            GlobalValue::AppendingLinkage, NewInit, Array);
  64 }
  65
  66 void llvm::appendToGlobalCtors(Module &M, Function *F, int Priority, Constant *Data) {
  67   appendToGlobalArray("llvm.global_ctors", M, F, Priority, Data);
  68 }
  69
  70 void llvm::appendToGlobalDtors(Module &M, Function *F, int Priority, Constant *Data) {
  71   appendToGlobalArray("llvm.global_dtors", M, F, Priority, Data);
  72 }
  73
  74 static void appendToUsedList(Module &M, StringRef Name, ArrayRef<GlobalValue *> Values) {
  75   GlobalVariable *GV = M.getGlobalVariable(Name);
  76   SmallPtrSet<Constant *, 16> InitAsSet;
  77   SmallVector<Constant *, 16> Init;
  78   if (GV) {
  79     if (GV->hasInitializer()) {
  80       auto *CA = cast<ConstantArray>(GV->getInitializer());
  81       for (auto &Op : CA->operands()) {
  82         Constant *C = cast_or_null<Constant>(Op);
  83         if (InitAsSet.insert(C).second)
  84           Init.push_back(C);
  85       }
  86     }
  87     GV->eraseFromParent();
  88   }
  89
  90   Type *Int8PtrTy = llvm::Type::getInt8PtrTy(M.getContext());
  91   for (auto *V : Values) {
  92     Constant *C = ConstantExpr::getPointerBitCastOrAddrSpaceCast(V, Int8PtrTy);
  93     if (InitAsSet.insert(C).second)
  94       Init.push_back(C);
  95   }
  96
  97   if (Init.empty())
  98     return;
  99
 100   ArrayType *ATy = ArrayType::get(Int8PtrTy, Init.size());
 101   GV = new llvm::GlobalVariable(M, ATy, false, GlobalValue::AppendingLinkage,
 102                                 ConstantArray::get(ATy, Init), Name);
 103   GV->setSection("llvm.metadata");
 104 }
 105
 106 void llvm::appendToUsed(Module &M, ArrayRef<GlobalValue *> Values) {
 107   appendToUsedList(M, "llvm.used", Values);
 108 }
 109
 110 void llvm::appendToCompilerUsed(Module &M, ArrayRef<GlobalValue *> Values) {
 111   appendToUsedList(M, "llvm.compiler.used", Values);
 112 }
 113
 114 FunctionCallee
 115 llvm::declareSanitizerInitFunction(Module &M, StringRef InitName,
 116                                    ArrayRef<Type *> InitArgTypes) {
 117   assert(!InitName.empty() && "Expected init function name");
 118   return M.getOrInsertFunction(
 119       InitName,
 120       FunctionType::get(Type::getVoidTy(M.getContext()), InitArgTypes, false),
 121       AttributeList());
 122 }
 123
 124 Function *llvm::createSanitizerCtor(Module &M, StringRef CtorName) {
 125   Function *Ctor = Function::createWithDefaultAttr(
 126       FunctionType::get(Type::getVoidTy(M.getContext()), false),
 127       GlobalValue::InternalLinkage, 0, CtorName, &M);
 128   Ctor->addFnAttr(Attribute::NoUnwind);
 129   BasicBlock *CtorBB = BasicBlock::Create(M.getContext(), "", Ctor);
 130   ReturnInst::Create(M.getContext(), CtorBB);
 131   // Ensure Ctor cannot be discarded, even if in a comdat.
 132   appendToUsed(M, {Ctor});
 133   return Ctor;
 134 }
 135
 136 std::pair<Function *, FunctionCallee> llvm::createSanitizerCtorAndInitFunctions(
 137     Module &M, StringRef CtorName, StringRef InitName,
 138     ArrayRef<Type *> InitArgTypes, ArrayRef<Value *> InitArgs,
 139     StringRef VersionCheckName) {
 140   assert(!InitName.empty() && "Expected init function name");
 141   assert(InitArgs.size() == InitArgTypes.size() &&
 142          "Sanitizer's init function expects different number of arguments");
 143   FunctionCallee InitFunction =
 144       declareSanitizerInitFunction(M, InitName, InitArgTypes);
 145   Function *Ctor = createSanitizerCtor(M, CtorName);
 146   IRBuilder<> IRB(Ctor->getEntryBlock().getTerminator());
 147   IRB.CreateCall(InitFunction, InitArgs);
 148   if (!VersionCheckName.empty()) {
 149     FunctionCallee VersionCheckFunction = M.getOrInsertFunction(
 150         VersionCheckName, FunctionType::get(IRB.getVoidTy(), {}, false),
 151         AttributeList());
 152     IRB.CreateCall(VersionCheckFunction, {});
 153   }
 154   return std::make_pair(Ctor, InitFunction);
 155 }
 156
 157 std::pair<Function *, FunctionCallee>
 158 llvm::getOrCreateSanitizerCtorAndInitFunctions(
 159     Module &M, StringRef CtorName, StringRef InitName,
 160     ArrayRef<Type *> InitArgTypes, ArrayRef<Value *> InitArgs,
 161     function_ref<void(Function *, FunctionCallee)> FunctionsCreatedCallback,
 162     StringRef VersionCheckName) {
 163   assert(!CtorName.empty() && "Expected ctor function name");
 164
 165   if (Function *Ctor = M.getFunction(CtorName))
 166     // FIXME: Sink this logic into the module, similar to the handling of
 167     // globals. This will make moving to a concurrent model much easier.
 168     if (Ctor->arg_size() == 0 ||
 169         Ctor->getReturnType() == Type::getVoidTy(M.getContext()))
 170       return {Ctor, declareSanitizerInitFunction(M, InitName, InitArgTypes)};
 171
 172   Function *Ctor;
 173   FunctionCallee InitFunction;
 174   std::tie(Ctor, InitFunction) = llvm::createSanitizerCtorAndInitFunctions(
 175       M, CtorName, InitName, InitArgTypes, InitArgs, VersionCheckName);
 176   FunctionsCreatedCallback(Ctor, InitFunction);
 177   return std::make_pair(Ctor, InitFunction);
 178 }
 179
 180 void llvm::filterDeadComdatFunctions(
 181     Module &M, SmallVectorImpl<Function *> &DeadComdatFunctions) {
 182   // Build a map from the comdat to the number of entries in that comdat we
 183   // think are dead. If this fully covers the comdat group, then the entire
 184   // group is dead. If we find another entry in the comdat group though, we'll
 185   // have to preserve the whole group.
 186   SmallDenseMap<Comdat *, int, 16> ComdatEntriesCovered;
 187   for (Function *F : DeadComdatFunctions) {
 188     Comdat *C = F->getComdat();
 189     assert(C && "Expected all input GVs to be in a comdat!");
 190     ComdatEntriesCovered[C] += 1;
 191   }
 192
 193   auto CheckComdat = [&](Comdat &C) {
 194     auto CI = ComdatEntriesCovered.find(&C);
 195     if (CI == ComdatEntriesCovered.end())
 196       return;
 197
 198     // If this could have been covered by a dead entry, just subtract one to
 199     // account for it.
 200     if (CI->second > 0) {
 201       CI->second -= 1;
 202       return;
 203     }
 204
 205     // If we've already accounted for all the entries that were dead, the
 206     // entire comdat is alive so remove it from the map.
 207     ComdatEntriesCovered.erase(CI);
 208   };
 209
 210   auto CheckAllComdats = [&] {
 211     for (Function &F : M.functions())
 212       if (Comdat *C = F.getComdat()) {
 213         CheckComdat(*C);
 214         if (ComdatEntriesCovered.empty())
 215           return;
 216       }
 217     for (GlobalVariable &GV : M.globals())
 218       if (Comdat *C = GV.getComdat()) {
 219         CheckComdat(*C);
 220         if (ComdatEntriesCovered.empty())
 221           return;
 222       }
 223     for (GlobalAlias &GA : M.aliases())
 224       if (Comdat *C = GA.getComdat()) {
 225         CheckComdat(*C);
 226         if (ComdatEntriesCovered.empty())
 227           return;
 228       }
 229   };
 230   CheckAllComdats();
 231
 232   if (ComdatEntriesCovered.empty()) {
 233     DeadComdatFunctions.clear();
 234     return;
 235   }
 236
 237   // Remove the entries that were not covering.
 238   erase_if(DeadComdatFunctions, [&](GlobalValue *GV) {
 239     return ComdatEntriesCovered.find(GV->getComdat()) ==
 240            ComdatEntriesCovered.end();
 241   });
 242 }
 243
 244 std::string llvm::getUniqueModuleId(Module *M) {
 245   MD5 Md5;
 246   bool ExportsSymbols = false;
 247   auto AddGlobal = [&](GlobalValue &GV) {
 248     if (GV.isDeclaration() || GV.getName().startswith("llvm.") ||
 249         !GV.hasExternalLinkage() || GV.hasComdat())
 250       return;
 251     ExportsSymbols = true;
 252     Md5.update(GV.getName());
 253     Md5.update(ArrayRef<uint8_t>{0});
 254   };
 255
 256   for (auto &F : *M)
 257     AddGlobal(F);
 258   for (auto &GV : M->globals())
 259     AddGlobal(GV);
 260   for (auto &GA : M->aliases())
 261     AddGlobal(GA);
 262   for (auto &IF : M->ifuncs())
 263     AddGlobal(IF);
 264
 265   if (!ExportsSymbols)
 266     return "";
 267
 268   MD5::MD5Result R;
 269   Md5.final(R);
 270
 271   SmallString<32> Str;
 272   MD5::stringifyResult(R, Str);
 273   return ("." + Str).str();
 274 }
 275
 276 void VFABI::setVectorVariantNames(
 277     CallInst *CI, const SmallVector<std::string, 8> &VariantMappings) {
 278   if (VariantMappings.empty())
 279     return;
 280
 281   SmallString<256> Buffer;
 282   llvm::raw_svector_ostream Out(Buffer);
 283   for (const std::string &VariantMapping : VariantMappings)
 284     Out << VariantMapping << ",";
 285   // Get rid of the trailing ','.
 286   assert(!Buffer.str().empty() && "Must have at least one char.");
 287   Buffer.pop_back();
 288
 289   Module *M = CI->getModule();
 290 #ifndef NDEBUG
 291   for (const std::string &VariantMapping : VariantMappings) {
 292     LLVM_DEBUG(dbgs() << "VFABI: adding mapping '" << VariantMapping << "'\n");
 293     Optional<VFInfo> VI = VFABI::tryDemangleForVFABI(VariantMapping, *M);
 294     assert(VI.hasValue() && "Cannot add an invalid VFABI name.");
 295     assert(M->getNamedValue(VI.getValue().VectorName) &&
 296            "Cannot add variant to attribute: "
 297            "vector function declaration is missing.");
 298   }
 299 #endif
 300   CI->addFnAttr(
 301       Attribute::get(M->getContext(), MappingsAttrName, Buffer.str()));
 302 }