lib/Target/WebAssembly/WebAssemblyFixFunctionBitcasts.cpp

   1 //===-- WebAssemblyFixFunctionBitcasts.cpp - Fix function bitcasts --------===//
   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 /// \file
  10 /// Fix bitcasted functions.
  11 ///
  12 /// WebAssembly requires caller and callee signatures to match, however in LLVM,
  13 /// some amount of slop is vaguely permitted. Detect mismatch by looking for
  14 /// bitcasts of functions and rewrite them to use wrapper functions instead.
  15 ///
  16 /// This doesn't catch all cases, such as when a function's address is taken in
  17 /// one place and casted in another, but it works for many common cases.
  18 ///
  19 /// Note that LLVM already optimizes away function bitcasts in common cases by
  20 /// dropping arguments as needed, so this pass only ends up getting used in less
  21 /// common cases.
  22 ///
  23 //===----------------------------------------------------------------------===//
  24
  25 #include "WebAssembly.h"
  26 #include "llvm/IR/CallSite.h"
  27 #include "llvm/IR/Constants.h"
  28 #include "llvm/IR/Instructions.h"
  29 #include "llvm/IR/Module.h"
  30 #include "llvm/IR/Operator.h"
  31 #include "llvm/Pass.h"
  32 #include "llvm/Support/Debug.h"
  33 #include "llvm/Support/raw_ostream.h"
  34 using namespace llvm;
  35
  36 #define DEBUG_TYPE "wasm-fix-function-bitcasts"
  37
  38 namespace {
  39 class FixFunctionBitcasts final : public ModulePass {
  40   StringRef getPassName() const override {
  41     return "WebAssembly Fix Function Bitcasts";
  42   }
  43
  44   void getAnalysisUsage(AnalysisUsage &AU) const override {
  45     AU.setPreservesCFG();
  46     ModulePass::getAnalysisUsage(AU);
  47   }
  48
  49   bool runOnModule(Module &M) override;
  50
  51 public:
  52   static char ID;
  53   FixFunctionBitcasts() : ModulePass(ID) {}
  54 };
  55 } // End anonymous namespace
  56
  57 char FixFunctionBitcasts::ID = 0;
  58 INITIALIZE_PASS(FixFunctionBitcasts, DEBUG_TYPE,
  59                 "Fix mismatching bitcasts for WebAssembly", false, false)
  60
  61 ModulePass *llvm::createWebAssemblyFixFunctionBitcasts() {
  62   return new FixFunctionBitcasts();
  63 }
  64
  65 // Recursively descend the def-use lists from V to find non-bitcast users of
  66 // bitcasts of V.
  67 static void findUses(Value *V, Function &F,
  68                      SmallVectorImpl<std::pair<Use *, Function *>> &Uses,
  69                      SmallPtrSetImpl<Constant *> &ConstantBCs) {
  70   for (Use &U : V->uses()) {
  71     if (auto *BC = dyn_cast<BitCastOperator>(U.getUser()))
  72       findUses(BC, F, Uses, ConstantBCs);
  73     else if (auto *A = dyn_cast<GlobalAlias>(U.getUser()))
  74       findUses(A, F, Uses, ConstantBCs);
  75     else if (U.get()->getType() != F.getType()) {
  76       CallSite CS(U.getUser());
  77       if (!CS)
  78         // Skip uses that aren't immediately called
  79         continue;
  80       Value *Callee = CS.getCalledValue();
  81       if (Callee != V)
  82         // Skip calls where the function isn't the callee
  83         continue;
  84       if (isa<Constant>(U.get())) {
  85         // Only add constant bitcasts to the list once; they get RAUW'd
  86         auto C = ConstantBCs.insert(cast<Constant>(U.get()));
  87         if (!C.second)
  88           continue;
  89       }
  90       Uses.push_back(std::make_pair(&U, &F));
  91     }
  92   }
  93 }
  94
  95 // Create a wrapper function with type Ty that calls F (which may have a
  96 // different type). Attempt to support common bitcasted function idioms:
  97 //  - Call with more arguments than needed: arguments are dropped
  98 //  - Call with fewer arguments than needed: arguments are filled in with undef
  99 //  - Return value is not needed: drop it
 100 //  - Return value needed but not present: supply an undef
 101 //
 102 // If the all the argument types of trivially castable to one another (i.e.
 103 // I32 vs pointer type) then we don't create a wrapper at all (return nullptr
 104 // instead).
 105 //
 106 // If there is a type mismatch that we know would result in an invalid wasm
 107 // module then generate wrapper that contains unreachable (i.e. abort at
 108 // runtime).  Such programs are deep into undefined behaviour territory,
 109 // but we choose to fail at runtime rather than generate and invalid module
 110 // or fail at compiler time.  The reason we delay the error is that we want
 111 // to support the CMake which expects to be able to compile and link programs
 112 // that refer to functions with entirely incorrect signatures (this is how
 113 // CMake detects the existence of a function in a toolchain).
 114 //
 115 // For bitcasts that involve struct types we don't know at this stage if they
 116 // would be equivalent at the wasm level and so we can't know if we need to
 117 // generate a wrapper.
 118 static Function *createWrapper(Function *F, FunctionType *Ty) {
 119   Module *M = F->getParent();
 120
 121   Function *Wrapper = Function::Create(Ty, Function::PrivateLinkage,
 122                                        F->getName() + "_bitcast", M);
 123   BasicBlock *BB = BasicBlock::Create(M->getContext(), "body", Wrapper);
 124   const DataLayout &DL = BB->getModule()->getDataLayout();
 125
 126   // Determine what arguments to pass.
 127   SmallVector<Value *, 4> Args;
 128   Function::arg_iterator AI = Wrapper->arg_begin();
 129   Function::arg_iterator AE = Wrapper->arg_end();
 130   FunctionType::param_iterator PI = F->getFunctionType()->param_begin();
 131   FunctionType::param_iterator PE = F->getFunctionType()->param_end();
 132   bool TypeMismatch = false;
 133   bool WrapperNeeded = false;
 134
 135   Type *ExpectedRtnType = F->getFunctionType()->getReturnType();
 136   Type *RtnType = Ty->getReturnType();
 137
 138   if ((F->getFunctionType()->getNumParams() != Ty->getNumParams()) ||
 139       (F->getFunctionType()->isVarArg() != Ty->isVarArg()) ||
 140       (ExpectedRtnType != RtnType))
 141     WrapperNeeded = true;
 142
 143   for (; AI != AE && PI != PE; ++AI, ++PI) {
 144     Type *ArgType = AI->getType();
 145     Type *ParamType = *PI;
 146
 147     if (ArgType == ParamType) {
 148       Args.push_back(&*AI);
 149     } else {
 150       if (CastInst::isBitOrNoopPointerCastable(ArgType, ParamType, DL)) {
 151         Instruction *PtrCast =
 152             CastInst::CreateBitOrPointerCast(AI, ParamType, "cast");
 153         BB->getInstList().push_back(PtrCast);
 154         Args.push_back(PtrCast);
 155       } else if (ArgType->isStructTy() || ParamType->isStructTy()) {
 156         LLVM_DEBUG(dbgs() << "createWrapper: struct param type in bitcast: "
 157                           << F->getName() << "\n");
 158         WrapperNeeded = false;
 159       } else {
 160         LLVM_DEBUG(dbgs() << "createWrapper: arg type mismatch calling: "
 161                           << F->getName() << "\n");
 162         LLVM_DEBUG(dbgs() << "Arg[" << Args.size() << "] Expected: "
 163                           << *ParamType << " Got: " << *ArgType << "\n");
 164         TypeMismatch = true;
 165         break;
 166       }
 167     }
 168   }
 169
 170   if (WrapperNeeded && !TypeMismatch) {
 171     for (; PI != PE; ++PI)
 172       Args.push_back(UndefValue::get(*PI));
 173     if (F->isVarArg())
 174       for (; AI != AE; ++AI)
 175         Args.push_back(&*AI);
 176
 177     CallInst *Call = CallInst::Create(F, Args, "", BB);
 178
 179     Type *ExpectedRtnType = F->getFunctionType()->getReturnType();
 180     Type *RtnType = Ty->getReturnType();
 181     // Determine what value to return.
 182     if (RtnType->isVoidTy()) {
 183       ReturnInst::Create(M->getContext(), BB);
 184     } else if (ExpectedRtnType->isVoidTy()) {
 185       LLVM_DEBUG(dbgs() << "Creating dummy return: " << *RtnType << "\n");
 186       ReturnInst::Create(M->getContext(), UndefValue::get(RtnType), BB);
 187     } else if (RtnType == ExpectedRtnType) {
 188       ReturnInst::Create(M->getContext(), Call, BB);
 189     } else if (CastInst::isBitOrNoopPointerCastable(ExpectedRtnType, RtnType,
 190                                                     DL)) {
 191       Instruction *Cast =
 192           CastInst::CreateBitOrPointerCast(Call, RtnType, "cast");
 193       BB->getInstList().push_back(Cast);
 194       ReturnInst::Create(M->getContext(), Cast, BB);
 195     } else if (RtnType->isStructTy() || ExpectedRtnType->isStructTy()) {
 196       LLVM_DEBUG(dbgs() << "createWrapper: struct return type in bitcast: "
 197                         << F->getName() << "\n");
 198       WrapperNeeded = false;
 199     } else {
 200       LLVM_DEBUG(dbgs() << "createWrapper: return type mismatch calling: "
 201                         << F->getName() << "\n");
 202       LLVM_DEBUG(dbgs() << "Expected: " << *ExpectedRtnType
 203                         << " Got: " << *RtnType << "\n");
 204       TypeMismatch = true;
 205     }
 206   }
 207
 208   if (TypeMismatch) {
 209     // Create a new wrapper that simply contains `unreachable`.
 210     Wrapper->eraseFromParent();
 211     Wrapper = Function::Create(Ty, Function::PrivateLinkage,
 212                                F->getName() + "_bitcast_invalid", M);
 213     BasicBlock *BB = BasicBlock::Create(M->getContext(), "body", Wrapper);
 214     new UnreachableInst(M->getContext(), BB);
 215     Wrapper->setName(F->getName() + "_bitcast_invalid");
 216   } else if (!WrapperNeeded) {
 217     LLVM_DEBUG(dbgs() << "createWrapper: no wrapper needed: " << F->getName()
 218                       << "\n");
 219     Wrapper->eraseFromParent();
 220     return nullptr;
 221   }
 222   LLVM_DEBUG(dbgs() << "createWrapper: " << F->getName() << "\n");
 223   return Wrapper;
 224 }
 225
 226 // Test whether a main function with type FuncTy should be rewritten to have
 227 // type MainTy.
 228 static bool shouldFixMainFunction(FunctionType *FuncTy, FunctionType *MainTy) {
 229   // Only fix the main function if it's the standard zero-arg form. That way,
 230   // the standard cases will work as expected, and users will see signature
 231   // mismatches from the linker for non-standard cases.
 232   return FuncTy->getReturnType() == MainTy->getReturnType() &&
 233          FuncTy->getNumParams() == 0 &&
 234          !FuncTy->isVarArg();
 235 }
 236
 237 bool FixFunctionBitcasts::runOnModule(Module &M) {
 238   LLVM_DEBUG(dbgs() << "********** Fix Function Bitcasts **********\n");
 239
 240   Function *Main = nullptr;
 241   CallInst *CallMain = nullptr;
 242   SmallVector<std::pair<Use *, Function *>, 0> Uses;
 243   SmallPtrSet<Constant *, 2> ConstantBCs;
 244
 245   // Collect all the places that need wrappers.
 246   for (Function &F : M) {
 247     findUses(&F, F, Uses, ConstantBCs);
 248
 249     // If we have a "main" function, and its type isn't
 250     // "int main(int argc, char *argv[])", create an artificial call with it
 251     // bitcasted to that type so that we generate a wrapper for it, so that
 252     // the C runtime can call it.
 253     if (F.getName() == "main") {
 254       Main = &F;
 255       LLVMContext &C = M.getContext();
 256       Type *MainArgTys[] = {Type::getInt32Ty(C),
 257                             PointerType::get(Type::getInt8PtrTy(C), 0)};
 258       FunctionType *MainTy = FunctionType::get(Type::getInt32Ty(C), MainArgTys,
 259                                                /*isVarArg=*/false);
 260       if (shouldFixMainFunction(F.getFunctionType(), MainTy)) {
 261         LLVM_DEBUG(dbgs() << "Found `main` function with incorrect type: "
 262                           << *F.getFunctionType() << "\n");
 263         Value *Args[] = {UndefValue::get(MainArgTys[0]),
 264                          UndefValue::get(MainArgTys[1])};
 265         Value *Casted =
 266             ConstantExpr::getBitCast(Main, PointerType::get(MainTy, 0));
 267         CallMain = CallInst::Create(MainTy, Casted, Args, "call_main");
 268         Use *UseMain = &CallMain->getOperandUse(2);
 269         Uses.push_back(std::make_pair(UseMain, &F));
 270       }
 271     }
 272   }
 273
 274   DenseMap<std::pair<Function *, FunctionType *>, Function *> Wrappers;
 275
 276   for (auto &UseFunc : Uses) {
 277     Use *U = UseFunc.first;
 278     Function *F = UseFunc.second;
 279     auto *PTy = cast<PointerType>(U->get()->getType());
 280     auto *Ty = dyn_cast<FunctionType>(PTy->getElementType());
 281
 282     // If the function is casted to something like i8* as a "generic pointer"
 283     // to be later casted to something else, we can't generate a wrapper for it.
 284     // Just ignore such casts for now.
 285     if (!Ty)
 286       continue;
 287
 288     auto Pair = Wrappers.insert(std::make_pair(std::make_pair(F, Ty), nullptr));
 289     if (Pair.second)
 290       Pair.first->second = createWrapper(F, Ty);
 291
 292     Function *Wrapper = Pair.first->second;
 293     if (!Wrapper)
 294       continue;
 295
 296     if (isa<Constant>(U->get()))
 297       U->get()->replaceAllUsesWith(Wrapper);
 298     else
 299       U->set(Wrapper);
 300   }
 301
 302   // If we created a wrapper for main, rename the wrapper so that it's the
 303   // one that gets called from startup.
 304   if (CallMain) {
 305     Main->setName("__original_main");
 306     auto *MainWrapper =
 307         cast<Function>(CallMain->getCalledValue()->stripPointerCasts());
 308     delete CallMain;
 309     if (Main->isDeclaration()) {
 310       // The wrapper is not needed in this case as we don't need to export
 311       // it to anyone else.
 312       MainWrapper->eraseFromParent();
 313     } else {
 314       // Otherwise give the wrapper the same linkage as the original main
 315       // function, so that it can be called from the same places.
 316       MainWrapper->setName("main");
 317       MainWrapper->setLinkage(Main->getLinkage());
 318       MainWrapper->setVisibility(Main->getVisibility());
 319     }
 320   }
 321
 322   return true;
 323 }