JIT.cpp revision 5c72a3ae106d8d1dd6aa7e573948dac4102aca7b 
1//===-- JIT.cpp - LLVM Just in Time Compiler ------------------------------===//
2//
3//                     The LLVM Compiler Infrastructure
4//
5// This file was developed by the LLVM research group and is distributed under
6// the University of Illinois Open Source License. See LICENSE.TXT for details.
7//
8//===----------------------------------------------------------------------===//
9//
10// This tool implements a just-in-time compiler for LLVM, allowing direct
11// execution of LLVM bytecode in an efficient manner.
12//
13//===----------------------------------------------------------------------===//
14
15#include "JIT.h"
16#include "llvm/Constants.h"
17#include "llvm/DerivedTypes.h"
18#include "llvm/Function.h"
19#include "llvm/GlobalVariable.h"
20#include "llvm/Instructions.h"
21#include "llvm/ModuleProvider.h"
22#include "llvm/CodeGen/MachineCodeEmitter.h"
23#include "llvm/CodeGen/MachineFunction.h"
24#include "llvm/ExecutionEngine/GenericValue.h"
25#include "llvm/Support/MutexGuard.h"
26#include "llvm/System/DynamicLibrary.h"
27#include "llvm/Target/TargetData.h"
28#include "llvm/Target/TargetMachine.h"
29#include "llvm/Target/TargetJITInfo.h"
30#include <iostream>
31using namespace llvm;
32
33static struct RegisterJIT {
34  RegisterJIT() { JIT::Register(); }
35} JITRegistrator;
36
37namespace llvm {
38  void LinkInJIT() {
39  }
40}
41
42JIT::JIT(ModuleProvider *MP, TargetMachine &tm, TargetJITInfo &tji)
43  : ExecutionEngine(MP), TM(tm), TJI(tji), state(MP) {
44  setTargetData(TM.getTargetData());
45
46  // Initialize MCE
47  MCE = createEmitter(*this);
48
49  // Add target data
50  MutexGuard locked(lock);
51  FunctionPassManager& PM = state.getPM(locked);
52  PM.add(new TargetData(*TM.getTargetData()));
53
54  // Compile LLVM Code down to machine code in the intermediate representation
55  TJI.addPassesToJITCompile(PM);
56
57  // Turn the machine code intermediate representation into bytes in memory that
58  // may be executed.
59  if (TM.addPassesToEmitMachineCode(PM, *MCE)) {
60    std::cerr << "Target '" << TM.getName()
61              << "' doesn't support machine code emission!\n";
62    abort();
63  }
64}
65
66JIT::~JIT() {
67  delete MCE;
68  delete &TM;
69}
70
71/// run - Start execution with the specified function and arguments.
72///
73GenericValue JIT::runFunction(Function *F,
74                              const std::vector<GenericValue> &ArgValues) {
75  assert(F && "Function *F was null at entry to run()");
76
77  void *FPtr = getPointerToFunction(F);
78  assert(FPtr && "Pointer to fn's code was null after getPointerToFunction");
79  const FunctionType *FTy = F->getFunctionType();
80  const Type *RetTy = FTy->getReturnType();
81
82  assert((FTy->getNumParams() <= ArgValues.size() || FTy->isVarArg()) &&
83         "Too many arguments passed into function!");
84  assert(FTy->getNumParams() == ArgValues.size() &&
85         "This doesn't support passing arguments through varargs (yet)!");
86
87  // Handle some common cases first.  These cases correspond to common `main'
88  // prototypes.
89  if (RetTy == Type::IntTy || RetTy == Type::UIntTy || RetTy == Type::VoidTy) {
90    switch (ArgValues.size()) {
91    case 3:
92      if ((FTy->getParamType(0) == Type::IntTy ||
93           FTy->getParamType(0) == Type::UIntTy) &&
94          isa<PointerType>(FTy->getParamType(1)) &&
95          isa<PointerType>(FTy->getParamType(2))) {
96        int (*PF)(int, char **, const char **) =
97          (int(*)(int, char **, const char **))(intptr_t)FPtr;
98
99        // Call the function.
100        GenericValue rv;
101        rv.IntVal = PF(ArgValues[0].IntVal, (char **)GVTOP(ArgValues[1]),
102                       (const char **)GVTOP(ArgValues[2]));
103        return rv;
104      }
105      break;
106    case 2:
107      if ((FTy->getParamType(0) == Type::IntTy ||
108           FTy->getParamType(0) == Type::UIntTy) &&
109          isa<PointerType>(FTy->getParamType(1))) {
110        int (*PF)(int, char **) = (int(*)(int, char **))(intptr_t)FPtr;
111
112        // Call the function.
113        GenericValue rv;
114        rv.IntVal = PF(ArgValues[0].IntVal, (char **)GVTOP(ArgValues[1]));
115        return rv;
116      }
117      break;
118    case 1:
119      if (FTy->getNumParams() == 1 &&
120          (FTy->getParamType(0) == Type::IntTy ||
121           FTy->getParamType(0) == Type::UIntTy)) {
122        GenericValue rv;
123        int (*PF)(int) = (int(*)(int))(intptr_t)FPtr;
124        rv.IntVal = PF(ArgValues[0].IntVal);
125        return rv;
126      }
127      break;
128    }
129  }
130
131  // Handle cases where no arguments are passed first.
132  if (ArgValues.empty()) {
133    GenericValue rv;
134    switch (RetTy->getTypeID()) {
135    default: assert(0 && "Unknown return type for function call!");
136    case Type::BoolTyID:
137      rv.BoolVal = ((bool(*)())(intptr_t)FPtr)();
138      return rv;
139    case Type::SByteTyID:
140    case Type::UByteTyID:
141      rv.SByteVal = ((char(*)())(intptr_t)FPtr)();
142      return rv;
143    case Type::ShortTyID:
144    case Type::UShortTyID:
145      rv.ShortVal = ((short(*)())(intptr_t)FPtr)();
146      return rv;
147    case Type::VoidTyID:
148    case Type::IntTyID:
149    case Type::UIntTyID:
150      rv.IntVal = ((int(*)())(intptr_t)FPtr)();
151      return rv;
152    case Type::LongTyID:
153    case Type::ULongTyID:
154      rv.LongVal = ((int64_t(*)())(intptr_t)FPtr)();
155      return rv;
156    case Type::FloatTyID:
157      rv.FloatVal = ((float(*)())(intptr_t)FPtr)();
158      return rv;
159    case Type::DoubleTyID:
160      rv.DoubleVal = ((double(*)())(intptr_t)FPtr)();
161      return rv;
162    case Type::PointerTyID:
163      return PTOGV(((void*(*)())(intptr_t)FPtr)());
164    }
165  }
166
167  // Okay, this is not one of our quick and easy cases.  Because we don't have a
168  // full FFI, we have to codegen a nullary stub function that just calls the
169  // function we are interested in, passing in constants for all of the
170  // arguments.  Make this function and return.
171
172  // First, create the function.
173  FunctionType *STy=FunctionType::get(RetTy, std::vector<const Type*>(), false);
174  Function *Stub = new Function(STy, Function::InternalLinkage, "",
175                                F->getParent());
176
177  // Insert a basic block.
178  BasicBlock *StubBB = new BasicBlock("", Stub);
179
180  // Convert all of the GenericValue arguments over to constants.  Note that we
181  // currently don't support varargs.
182  std::vector<Value*> Args;
183  for (unsigned i = 0, e = ArgValues.size(); i != e; ++i) {
184    Constant *C = 0;
185    const Type *ArgTy = FTy->getParamType(i);
186    const GenericValue &AV = ArgValues[i];
187    switch (ArgTy->getTypeID()) {
188    default: assert(0 && "Unknown argument type for function call!");
189    case Type::BoolTyID:   C = ConstantBool::get(AV.BoolVal); break;
190    case Type::SByteTyID:  C = ConstantSInt::get(ArgTy, AV.SByteVal);  break;
191    case Type::UByteTyID:  C = ConstantUInt::get(ArgTy, AV.UByteVal);  break;
192    case Type::ShortTyID:  C = ConstantSInt::get(ArgTy, AV.ShortVal);  break;
193    case Type::UShortTyID: C = ConstantUInt::get(ArgTy, AV.UShortVal); break;
194    case Type::IntTyID:    C = ConstantSInt::get(ArgTy, AV.IntVal);    break;
195    case Type::UIntTyID:   C = ConstantUInt::get(ArgTy, AV.UIntVal);   break;
196    case Type::LongTyID:   C = ConstantSInt::get(ArgTy, AV.LongVal);   break;
197    case Type::ULongTyID:  C = ConstantUInt::get(ArgTy, AV.ULongVal);  break;
198    case Type::FloatTyID:  C = ConstantFP  ::get(ArgTy, AV.FloatVal);  break;
199    case Type::DoubleTyID: C = ConstantFP  ::get(ArgTy, AV.DoubleVal); break;
200    case Type::PointerTyID:
201      void *ArgPtr = GVTOP(AV);
202      if (sizeof(void*) == 4) {
203        C = ConstantSInt::get(Type::IntTy, (int)(intptr_t)ArgPtr);
204      } else {
205        C = ConstantSInt::get(Type::LongTy, (intptr_t)ArgPtr);
206      }
207      C = ConstantExpr::getCast(C, ArgTy);  // Cast the integer to pointer
208      break;
209    }
210    Args.push_back(C);
211  }
212
213  CallInst *TheCall = new CallInst(F, Args, "", StubBB);
214  TheCall->setTailCall();
215  if (TheCall->getType() != Type::VoidTy)
216    new ReturnInst(TheCall, StubBB);             // Return result of the call.
217  else
218    new ReturnInst(StubBB);                      // Just return void.
219
220  // Finally, return the value returned by our nullary stub function.
221  return runFunction(Stub, std::vector<GenericValue>());
222}
223
224/// runJITOnFunction - Run the FunctionPassManager full of
225/// just-in-time compilation passes on F, hopefully filling in
226/// GlobalAddress[F] with the address of F's machine code.
227///
228void JIT::runJITOnFunction(Function *F) {
229  static bool isAlreadyCodeGenerating = false;
230  assert(!isAlreadyCodeGenerating && "Error: Recursive compilation detected!");
231
232  MutexGuard locked(lock);
233
234  // JIT the function
235  isAlreadyCodeGenerating = true;
236  state.getPM(locked).run(*F);
237  isAlreadyCodeGenerating = false;
238
239  // If the function referred to a global variable that had not yet been
240  // emitted, it allocates memory for the global, but doesn't emit it yet.  Emit
241  // all of these globals now.
242  while (!state.getPendingGlobals(locked).empty()) {
243    const GlobalVariable *GV = state.getPendingGlobals(locked).back();
244    state.getPendingGlobals(locked).pop_back();
245    EmitGlobalVariable(GV);
246  }
247}
248
249/// getPointerToFunction - This method is used to get the address of the
250/// specified function, compiling it if neccesary.
251///
252void *JIT::getPointerToFunction(Function *F) {
253  MutexGuard locked(lock);
254
255  if (void *Addr = getPointerToGlobalIfAvailable(F))
256    return Addr;   // Check if function already code gen'd
257
258  // Make sure we read in the function if it exists in this Module
259  if (F->hasNotBeenReadFromBytecode()) {
260    std::string ErrorMsg;
261    if (MP->materializeFunction(F, &ErrorMsg)) {
262      std::cerr << "Error reading function '" << F->getName()
263                << "' from bytecode file: " << ErrorMsg << "\n";
264      abort();
265    }
266  }
267
268  if (F->isExternal()) {
269    void *Addr = getPointerToNamedFunction(F->getName());
270    addGlobalMapping(F, Addr);
271    return Addr;
272  }
273
274  runJITOnFunction(F);
275
276  void *Addr = getPointerToGlobalIfAvailable(F);
277  assert(Addr && "Code generation didn't add function to GlobalAddress table!");
278  return Addr;
279}
280
281/// getOrEmitGlobalVariable - Return the address of the specified global
282/// variable, possibly emitting it to memory if needed.  This is used by the
283/// Emitter.
284void *JIT::getOrEmitGlobalVariable(const GlobalVariable *GV) {
285  MutexGuard locked(lock);
286
287  void *Ptr = getPointerToGlobalIfAvailable(GV);
288  if (Ptr) return Ptr;
289
290  // If the global is external, just remember the address.
291  if (GV->isExternal()) {
292    Ptr = sys::DynamicLibrary::SearchForAddressOfSymbol(GV->getName().c_str());
293    if (Ptr == 0) {
294      std::cerr << "Could not resolve external global address: "
295                << GV->getName() << "\n";
296      abort();
297    }
298  } else {
299    // If the global hasn't been emitted to memory yet, allocate space.  We will
300    // actually initialize the global after current function has finished
301    // compilation.
302    const Type *GlobalType = GV->getType()->getElementType();
303    size_t S = getTargetData()->getTypeSize(GlobalType);
304    size_t A = getTargetData()->getTypeAlignment(GlobalType);
305    if (A <= 8) {
306      Ptr = malloc(S);
307    } else {
308      // Allocate S+A bytes of memory, then use an aligned pointer within that
309      // space.
310      Ptr = malloc(S+A);
311      unsigned MisAligned = ((intptr_t)Ptr & (A-1));
312      unsigned Offset = MisAligned ? (A-MisAligned) : 0;
313
314      // Trim the tail off the memory block.
315      realloc(Ptr, S+Offset);
316      Ptr = (char*)Ptr + Offset;
317    }
318    state.getPendingGlobals(locked).push_back(GV);
319  }
320  addGlobalMapping(GV, Ptr);
321  return Ptr;
322}
323
324
325/// recompileAndRelinkFunction - This method is used to force a function
326/// which has already been compiled, to be compiled again, possibly
327/// after it has been modified. Then the entry to the old copy is overwritten
328/// with a branch to the new copy. If there was no old copy, this acts
329/// just like JIT::getPointerToFunction().
330///
331void *JIT::recompileAndRelinkFunction(Function *F) {
332  void *OldAddr = getPointerToGlobalIfAvailable(F);
333
334  // If it's not already compiled there is no reason to patch it up.
335  if (OldAddr == 0) { return getPointerToFunction(F); }
336
337  // Delete the old function mapping.
338  addGlobalMapping(F, 0);
339
340  // Recodegen the function
341  runJITOnFunction(F);
342
343  // Update state, forward the old function to the new function.
344  void *Addr = getPointerToGlobalIfAvailable(F);
345  assert(Addr && "Code generation didn't add function to GlobalAddress table!");
346  TJI.replaceMachineCodeForFunction(OldAddr, Addr);
347  return Addr;
348}
349
350