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