lib/Target/WebAssembly/WebAssemblyFixFunctionBitcasts.cpp

   1 //===-- WebAssemblyFixFunctionBitcasts.cpp - Fix function bitcasts --------===//
   2 //
   3 //                     The LLVM Compiler Infrastructure
   4 //
   5 // This file is distributed under the University of Illinois Open Source
   6 // License. See LICENSE.TXT for details.
   7 //
   8 //===----------------------------------------------------------------------===//
   9 ///
  10 /// \file
  11 /// Fix bitcasted functions.
  12 ///
  13 /// WebAssembly requires caller and callee signatures to match, however in LLVM,
  14 /// some amount of slop is vaguely permitted. Detect mismatch by looking for
  15 /// bitcasts of functions and rewrite them to use wrapper functions instead.
  16 ///
  17 /// This doesn't catch all cases, such as when a function's address is taken in
  18 /// one place and casted in another, but it works for many common cases.
  19 ///
  20 /// Note that LLVM already optimizes away function bitcasts in common cases by
  21 /// dropping arguments as needed, so this pass only ends up getting used in less
  22 /// common cases.
  23 ///
  24 //===----------------------------------------------------------------------===//
  25
  26 #include "WebAssembly.h"
  27 #include "llvm/IR/CallSite.h"
  28 #include "llvm/IR/Constants.h"
  29 #include "llvm/IR/Instructions.h"
  30 #include "llvm/IR/Module.h"
  31 #include "llvm/IR/Operator.h"
  32 #include "llvm/Pass.h"
  33 #include "llvm/Support/Debug.h"
  34 #include "llvm/Support/raw_ostream.h"
  35 using namespace llvm;
  36
  37 #define DEBUG_TYPE "wasm-fix-function-bitcasts"
  38
  39 namespace {
  40 class FixFunctionBitcasts final : public ModulePass {
  41   StringRef getPassName() const override {
  42     return "WebAssembly Fix Function Bitcasts";
  43   }
  44
  45   void getAnalysisUsage(AnalysisUsage &AU) const override {
  46     AU.setPreservesCFG();
  47     ModulePass::getAnalysisUsage(AU);
  48   }
  49
  50   bool runOnModule(Module &M) override;
  51
  52 public:
  53   static char ID;
  54   FixFunctionBitcasts() : ModulePass(ID) {}
  55 };
  56 } // End anonymous namespace
  57
  58 char FixFunctionBitcasts::ID = 0;
  59 INITIALIZE_PASS(FixFunctionBitcasts, DEBUG_TYPE,
  60                 "Fix mismatching bitcasts for WebAssembly", false, false)
  61
  62 ModulePass *llvm::createWebAssemblyFixFunctionBitcasts() {
  63   return new FixFunctionBitcasts();
  64 }
  65
  66 // Recursively descend the def-use lists from V to find non-bitcast users of
  67 // bitcasts of V.
  68 static void FindUses(Value *V, Function &F,
  69                      SmallVectorImpl<std::pair<Use *, Function *>> &Uses,
  70                      SmallPtrSetImpl<Constant *> &ConstantBCs) {
  71   for (Use &U : V->uses()) {
  72     if (BitCastOperator *BC = dyn_cast<BitCastOperator>(U.getUser()))
  73       FindUses(BC, F, Uses, ConstantBCs);
  74     else if (U.get()->getType() != F.getType()) {
  75       CallSite CS(U.getUser());
  76       if (!CS)
  77         // Skip uses that aren't immediately called
  78         continue;
  79       Value *Callee = CS.getCalledValue();
  80       if (Callee != V)
  81         // Skip calls where the function isn't the callee
  82         continue;
  83       if (isa<Constant>(U.get())) {
  84         // Only add constant bitcasts to the list once; they get RAUW'd
  85         auto c = ConstantBCs.insert(cast<Constant>(U.get()));
  86         if (!c.second)
  87           continue;
  88       }
  89       Uses.push_back(std::make_pair(&U, &F));
  90     }
  91   }
  92 }
  93
  94 // Create a wrapper function with type Ty that calls F (which may have a
  95 // different type). Attempt to support common bitcasted function idioms:
  96 //  - Call with more arguments than needed: arguments are dropped
  97 //  - Call with fewer arguments than needed: arguments are filled in with undef
  98 //  - Return value is not needed: drop it
  99 //  - Return value needed but not present: supply an undef
 100 //
 101 // If the all the argument types of trivially castable to one another (i.e.
 102 // I32 vs pointer type) then we don't create a wrapper at all (return nullptr
 103 // instead).
 104 //
 105 // If there is a type mismatch that we know would result in an invalid wasm
 106 // module then generate wrapper that contains unreachable (i.e. abort at
 107 // runtime).  Such programs are deep into undefined behaviour territory,
 108 // but we choose to fail at runtime rather than generate and invalid module
 109 // or fail at compiler time.  The reason we delay the error is that we want
 110 // to support the CMake which expects to be able to compile and link programs
 111 // that refer to functions with entirely incorrect signatures (this is how
 112 // CMake detects the existence of a function in a toolchain).
 113 //
 114 // For bitcasts that involve struct types we don't know at this stage if they
 115 // would be equivalent at the wasm level and so we can't know if we need to
 116 // generate a wrapper.
 117 static Function *CreateWrapper(Function *F, FunctionType *Ty) {
 118   Module *M = F->getParent();
 119
 120   Function *Wrapper = Function::Create(Ty, Function::PrivateLinkage,
 121                                        F->getName() + "_bitcast", M);
 122   BasicBlock *BB = BasicBlock::Create(M->getContext(), "body", Wrapper);
 123   const DataLayout &DL = BB->getModule()->getDataLayout();
 124
 125   // Determine what arguments to pass.
 126   SmallVector<Value *, 4> Args;
 127   Function::arg_iterator AI = Wrapper->arg_begin();
 128   Function::arg_iterator AE = Wrapper->arg_end();
 129   FunctionType::param_iterator PI = F->getFunctionType()->param_begin();
 130   FunctionType::param_iterator PE = F->getFunctionType()->param_end();
 131   bool TypeMismatch = false;
 132   bool WrapperNeeded = false;
 133
 134   Type *ExpectedRtnType = F->getFunctionType()->getReturnType();
 135   Type *RtnType = Ty->getReturnType();
 136
 137   if ((F->getFunctionType()->getNumParams() != Ty->getNumParams()) ||
 138       (F->getFunctionType()->isVarArg() != Ty->isVarArg()) ||
 139       (ExpectedRtnType != RtnType))
 140     WrapperNeeded = true;
 141
 142   for (; AI != AE && PI != PE; ++AI, ++PI) {
 143     Type *ArgType = AI->getType();
 144     Type *ParamType = *PI;
 145
 146     if (ArgType == ParamType) {
 147       Args.push_back(&*AI);
 148     } else {
 149       if (CastInst::isBitOrNoopPointerCastable(ArgType, ParamType, DL)) {
 150         Instruction *PtrCast =
 151             CastInst::CreateBitOrPointerCast(AI, ParamType, "cast");
 152         BB->getInstList().push_back(PtrCast);
 153         Args.push_back(PtrCast);
 154       } else if (ArgType->isStructTy() || ParamType->isStructTy()) {
 155         LLVM_DEBUG(dbgs() << "CreateWrapper: struct param type in bitcast: "
 156                           << F->getName() << "\n");
 157         WrapperNeeded = false;
 158       } else {
 159         LLVM_DEBUG(dbgs() << "CreateWrapper: arg type mismatch calling: "
 160                           << F->getName() << "\n");
 161         LLVM_DEBUG(dbgs() << "Arg[" << Args.size() << "] Expected: "
 162                           << *ParamType << " Got: " << *ArgType << "\n");
 163         TypeMismatch = true;
 164         break;
 165       }
 166     }
 167   }
 168
 169   if (WrapperNeeded && !TypeMismatch) {
 170     for (; PI != PE; ++PI)
 171       Args.push_back(UndefValue::get(*PI));
 172     if (F->isVarArg())
 173       for (; AI != AE; ++AI)
 174         Args.push_back(&*AI);
 175
 176     CallInst *Call = CallInst::Create(F, Args, "", BB);
 177
 178     Type *ExpectedRtnType = F->getFunctionType()->getReturnType();
 179     Type *RtnType = Ty->getReturnType();
 180     // Determine what value to return.
 181     if (RtnType->isVoidTy()) {
 182       ReturnInst::Create(M->getContext(), BB);
 183     } else if (ExpectedRtnType->isVoidTy()) {
 184       LLVM_DEBUG(dbgs() << "Creating dummy return: " << *RtnType << "\n");
 185       ReturnInst::Create(M->getContext(), UndefValue::get(RtnType), BB);
 186     } else if (RtnType == ExpectedRtnType) {
 187       ReturnInst::Create(M->getContext(), Call, BB);
 188     } else if (CastInst::isBitOrNoopPointerCastable(ExpectedRtnType, RtnType,
 189                                                     DL)) {
 190       Instruction *Cast =
 191           CastInst::CreateBitOrPointerCast(Call, RtnType, "cast");
 192       BB->getInstList().push_back(Cast);
 193       ReturnInst::Create(M->getContext(), Cast, BB);
 194     } else if (RtnType->isStructTy() || ExpectedRtnType->isStructTy()) {
 195       LLVM_DEBUG(dbgs() << "CreateWrapper: struct return type in bitcast: "
 196                         << F->getName() << "\n");
 197       WrapperNeeded = false;
 198     } else {
 199       LLVM_DEBUG(dbgs() << "CreateWrapper: return type mismatch calling: "
 200                         << F->getName() << "\n");
 201       LLVM_DEBUG(dbgs() << "Expected: " << *ExpectedRtnType
 202                         << " Got: " << *RtnType << "\n");
 203       TypeMismatch = true;
 204     }
 205   }
 206
 207   if (TypeMismatch) {
 208     // Create a new wrapper that simply contains `unreachable`.
 209     Wrapper->eraseFromParent();
 210     Wrapper = Function::Create(Ty, Function::PrivateLinkage,
 211                                F->getName() + "_bitcast_invalid", M);
 212     BasicBlock *BB = BasicBlock::Create(M->getContext(), "body", Wrapper);
 213     new UnreachableInst(M->getContext(), BB);
 214     Wrapper->setName(F->getName() + "_bitcast_invalid");
 215   } else if (!WrapperNeeded) {
 216     LLVM_DEBUG(dbgs() << "CreateWrapper: no wrapper needed: " << F->getName()
 217                       << "\n");
 218     Wrapper->eraseFromParent();
 219     return nullptr;
 220   }
 221   LLVM_DEBUG(dbgs() << "CreateWrapper: " << F->getName() << "\n");
 222   return Wrapper;
 223 }
 224
 225 bool FixFunctionBitcasts::runOnModule(Module &M) {
 226   Function *Main = nullptr;
 227   CallInst *CallMain = nullptr;
 228   SmallVector<std::pair<Use *, Function *>, 0> Uses;
 229   SmallPtrSet<Constant *, 2> ConstantBCs;
 230
 231   // Collect all the places that need wrappers.
 232   for (Function &F : M) {
 233     FindUses(&F, F, Uses, ConstantBCs);
 234
 235     // If we have a "main" function, and its type isn't
 236     // "int main(int argc, char *argv[])", create an artificial call with it
 237     // bitcasted to that type so that we generate a wrapper for it, so that
 238     // the C runtime can call it.
 239     if (!F.isDeclaration() && F.getName() == "main") {
 240       Main = &F;
 241       LLVMContext &C = M.getContext();
 242       Type *MainArgTys[] = {Type::getInt32Ty(C),
 243                             PointerType::get(Type::getInt8PtrTy(C), 0)};
 244       FunctionType *MainTy = FunctionType::get(Type::getInt32Ty(C), MainArgTys,
 245                                                /*isVarArg=*/false);
 246       if (F.getFunctionType() != MainTy) {
 247         LLVM_DEBUG(dbgs() << "Found `main` function with incorrect type: "
 248                           << *F.getFunctionType() << "\n");
 249         Value *Args[] = {UndefValue::get(MainArgTys[0]),
 250                          UndefValue::get(MainArgTys[1])};
 251         Value *Casted =
 252             ConstantExpr::getBitCast(Main, PointerType::get(MainTy, 0));
 253         CallMain = CallInst::Create(Casted, Args, "call_main");
 254         Use *UseMain = &CallMain->getOperandUse(2);
 255         Uses.push_back(std::make_pair(UseMain, &F));
 256       }
 257     }
 258   }
 259
 260   DenseMap<std::pair<Function *, FunctionType *>, Function *> Wrappers;
 261
 262   for (auto &UseFunc : Uses) {
 263     Use *U = UseFunc.first;
 264     Function *F = UseFunc.second;
 265     PointerType *PTy = cast<PointerType>(U->get()->getType());
 266     FunctionType *Ty = dyn_cast<FunctionType>(PTy->getElementType());
 267
 268     // If the function is casted to something like i8* as a "generic pointer"
 269     // to be later casted to something else, we can't generate a wrapper for it.
 270     // Just ignore such casts for now.
 271     if (!Ty)
 272       continue;
 273
 274     auto Pair = Wrappers.insert(std::make_pair(std::make_pair(F, Ty), nullptr));
 275     if (Pair.second)
 276       Pair.first->second = CreateWrapper(F, Ty);
 277
 278     Function *Wrapper = Pair.first->second;
 279     if (!Wrapper)
 280       continue;
 281
 282     if (isa<Constant>(U->get()))
 283       U->get()->replaceAllUsesWith(Wrapper);
 284     else
 285       U->set(Wrapper);
 286   }
 287
 288   // If we created a wrapper for main, rename the wrapper so that it's the
 289   // one that gets called from startup.
 290   if (CallMain) {
 291     Main->setName("__original_main");
 292     Function *MainWrapper =
 293         cast<Function>(CallMain->getCalledValue()->stripPointerCasts());
 294     MainWrapper->setName("main");
 295     MainWrapper->setLinkage(Main->getLinkage());
 296     MainWrapper->setVisibility(Main->getVisibility());
 297     Main->setLinkage(Function::PrivateLinkage);
 298     Main->setVisibility(Function::DefaultVisibility);
 299     delete CallMain;
 300   }
 301
 302   return true;
 303 }