WebAssemblyFixFunctionBitcasts.cpp revision 341825 
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"
35using namespace llvm;
36
37#define DEBUG_TYPE "wasm-fix-function-bitcasts"
38
39static cl::opt<bool> TemporaryWorkarounds(
40  "wasm-temporary-workarounds",
41  cl::desc("Apply certain temporary workarounds"),
42  cl::init(true), cl::Hidden);
43
44namespace {
45class FixFunctionBitcasts final : public ModulePass {
46  StringRef getPassName() const override {
47    return "WebAssembly Fix Function Bitcasts";
48  }
49
50  void getAnalysisUsage(AnalysisUsage &AU) const override {
51    AU.setPreservesCFG();
52    ModulePass::getAnalysisUsage(AU);
53  }
54
55  bool runOnModule(Module &M) override;
56
57public:
58  static char ID;
59  FixFunctionBitcasts() : ModulePass(ID) {}
60};
61} // End anonymous namespace
62
63char FixFunctionBitcasts::ID = 0;
64INITIALIZE_PASS(FixFunctionBitcasts, DEBUG_TYPE,
65                "Fix mismatching bitcasts for WebAssembly", false, false)
66
67ModulePass *llvm::createWebAssemblyFixFunctionBitcasts() {
68  return new FixFunctionBitcasts();
69}
70
71// Recursively descend the def-use lists from V to find non-bitcast users of
72// bitcasts of V.
73static void FindUses(Value *V, Function &F,
74                     SmallVectorImpl<std::pair<Use *, Function *>> &Uses,
75                     SmallPtrSetImpl<Constant *> &ConstantBCs) {
76  for (Use &U : V->uses()) {
77    if (BitCastOperator *BC = dyn_cast<BitCastOperator>(U.getUser()))
78      FindUses(BC, F, Uses, ConstantBCs);
79    else if (U.get()->getType() != F.getType()) {
80      CallSite CS(U.getUser());
81      if (!CS)
82        // Skip uses that aren't immediately called
83        continue;
84      Value *Callee = CS.getCalledValue();
85      if (Callee != V)
86        // Skip calls where the function isn't the callee
87        continue;
88      if (isa<Constant>(U.get())) {
89        // Only add constant bitcasts to the list once; they get RAUW'd
90        auto c = ConstantBCs.insert(cast<Constant>(U.get()));
91        if (!c.second)
92          continue;
93      }
94      Uses.push_back(std::make_pair(&U, &F));
95    }
96  }
97}
98
99// Create a wrapper function with type Ty that calls F (which may have a
100// different type). Attempt to support common bitcasted function idioms:
101//  - Call with more arguments than needed: arguments are dropped
102//  - Call with fewer arguments than needed: arguments are filled in with undef
103//  - Return value is not needed: drop it
104//  - Return value needed but not present: supply an undef
105//
106// For now, return nullptr without creating a wrapper if the wrapper cannot
107// be generated due to incompatible types.
108static Function *CreateWrapper(Function *F, FunctionType *Ty) {
109  Module *M = F->getParent();
110
111  Function *Wrapper =
112      Function::Create(Ty, Function::PrivateLinkage, "bitcast", M);
113  BasicBlock *BB = BasicBlock::Create(M->getContext(), "body", Wrapper);
114
115  // Determine what arguments to pass.
116  SmallVector<Value *, 4> Args;
117  Function::arg_iterator AI = Wrapper->arg_begin();
118  Function::arg_iterator AE = Wrapper->arg_end();
119  FunctionType::param_iterator PI = F->getFunctionType()->param_begin();
120  FunctionType::param_iterator PE = F->getFunctionType()->param_end();
121  for (; AI != AE && PI != PE; ++AI, ++PI) {
122    if (AI->getType() != *PI) {
123      Wrapper->eraseFromParent();
124      return nullptr;
125    }
126    Args.push_back(&*AI);
127  }
128  for (; PI != PE; ++PI)
129    Args.push_back(UndefValue::get(*PI));
130  if (F->isVarArg())
131    for (; AI != AE; ++AI)
132      Args.push_back(&*AI);
133
134  CallInst *Call = CallInst::Create(F, Args, "", BB);
135
136  // Determine what value to return.
137  if (Ty->getReturnType()->isVoidTy())
138    ReturnInst::Create(M->getContext(), BB);
139  else if (F->getFunctionType()->getReturnType()->isVoidTy())
140    ReturnInst::Create(M->getContext(), UndefValue::get(Ty->getReturnType()),
141                       BB);
142  else if (F->getFunctionType()->getReturnType() == Ty->getReturnType())
143    ReturnInst::Create(M->getContext(), Call, BB);
144  else {
145    Wrapper->eraseFromParent();
146    return nullptr;
147  }
148
149  return Wrapper;
150}
151
152bool FixFunctionBitcasts::runOnModule(Module &M) {
153  Function *Main = nullptr;
154  CallInst *CallMain = nullptr;
155  SmallVector<std::pair<Use *, Function *>, 0> Uses;
156  SmallPtrSet<Constant *, 2> ConstantBCs;
157
158  // Collect all the places that need wrappers.
159  for (Function &F : M) {
160    FindUses(&F, F, Uses, ConstantBCs);
161
162    // If we have a "main" function, and its type isn't
163    // "int main(int argc, char *argv[])", create an artificial call with it
164    // bitcasted to that type so that we generate a wrapper for it, so that
165    // the C runtime can call it.
166    if (!TemporaryWorkarounds && !F.isDeclaration() && F.getName() == "main") {
167      Main = &F;
168      LLVMContext &C = M.getContext();
169      Type *MainArgTys[] = {
170        PointerType::get(Type::getInt8PtrTy(C), 0),
171        Type::getInt32Ty(C)
172      };
173      FunctionType *MainTy = FunctionType::get(Type::getInt32Ty(C), MainArgTys,
174                                               /*isVarArg=*/false);
175      if (F.getFunctionType() != MainTy) {
176        Value *Args[] = {
177          UndefValue::get(MainArgTys[0]),
178          UndefValue::get(MainArgTys[1])
179        };
180        Value *Casted = ConstantExpr::getBitCast(Main,
181                                                 PointerType::get(MainTy, 0));
182        CallMain = CallInst::Create(Casted, Args, "call_main");
183        Use *UseMain = &CallMain->getOperandUse(2);
184        Uses.push_back(std::make_pair(UseMain, &F));
185      }
186    }
187  }
188
189  DenseMap<std::pair<Function *, FunctionType *>, Function *> Wrappers;
190
191  for (auto &UseFunc : Uses) {
192    Use *U = UseFunc.first;
193    Function *F = UseFunc.second;
194    PointerType *PTy = cast<PointerType>(U->get()->getType());
195    FunctionType *Ty = dyn_cast<FunctionType>(PTy->getElementType());
196
197    // If the function is casted to something like i8* as a "generic pointer"
198    // to be later casted to something else, we can't generate a wrapper for it.
199    // Just ignore such casts for now.
200    if (!Ty)
201      continue;
202
203    // Bitcasted vararg functions occur in Emscripten's implementation of
204    // EM_ASM, so suppress wrappers for them for now.
205    if (TemporaryWorkarounds && (Ty->isVarArg() || F->isVarArg()))
206      continue;
207
208    auto Pair = Wrappers.insert(std::make_pair(std::make_pair(F, Ty), nullptr));
209    if (Pair.second)
210      Pair.first->second = CreateWrapper(F, Ty);
211
212    Function *Wrapper = Pair.first->second;
213    if (!Wrapper)
214      continue;
215
216    if (isa<Constant>(U->get()))
217      U->get()->replaceAllUsesWith(Wrapper);
218    else
219      U->set(Wrapper);
220  }
221
222  // If we created a wrapper for main, rename the wrapper so that it's the
223  // one that gets called from startup.
224  if (CallMain) {
225    Main->setName("__original_main");
226    Function *MainWrapper =
227        cast<Function>(CallMain->getCalledValue()->stripPointerCasts());
228    MainWrapper->setName("main");
229    MainWrapper->setLinkage(Main->getLinkage());
230    MainWrapper->setVisibility(Main->getVisibility());
231    Main->setLinkage(Function::PrivateLinkage);
232    Main->setVisibility(Function::DefaultVisibility);
233    delete CallMain;
234  }
235
236  return true;
237}
238