JIT.cpp revision 34c9433004cabd4760987dce4804a91c84908219 
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 bitcode 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
31#include "llvm/Config/config.h"
32
33using namespace llvm;
34
35#ifdef __APPLE__
36// Apple gcc defaults to -fuse-cxa-atexit (i.e. calls __cxa_atexit instead
37// of atexit). It passes the address of linker generated symbol __dso_handle
38// to the function.
39// This configuration change happened at version 5330.
40# include <AvailabilityMacros.h>
41# if defined(MAC_OS_X_VERSION_10_4) && \
42     ((MAC_OS_X_VERSION_MIN_REQUIRED > MAC_OS_X_VERSION_10_4) || \
43      (MAC_OS_X_VERSION_MIN_REQUIRED == MAC_OS_X_VERSION_10_4 && \
44       __APPLE_CC__ >= 5330))
45#  ifndef HAVE___DSO_HANDLE
46#   define HAVE___DSO_HANDLE 1
47#  endif
48# endif
49#endif
50
51#if HAVE___DSO_HANDLE
52extern void *__dso_handle __attribute__ ((__visibility__ ("hidden")));
53#endif
54
55static struct RegisterJIT {
56  RegisterJIT() { JIT::Register(); }
57} JITRegistrator;
58
59namespace llvm {
60  void LinkInJIT() {
61  }
62}
63
64/// createJIT - This is the factory method for creating a JIT for the current
65/// machine, it does not fall back to the interpreter.  This takes ownership
66/// of the module provider.
67ExecutionEngine *ExecutionEngine::createJIT(ModuleProvider *MP,
68                                            std::string *ErrorStr,
69                                            JITMemoryManager *JMM) {
70  ExecutionEngine *EE = JIT::createJIT(MP, ErrorStr, JMM);
71  if (!EE) return 0;
72
73
74  // Make sure we can resolve symbols in the program as well. The zero arg
75  // to the function tells DynamicLibrary to load the program, not a library.
76  sys::DynamicLibrary::LoadLibraryPermanently(0, ErrorStr);
77  return EE;
78}
79
80JIT::JIT(ModuleProvider *MP, TargetMachine &tm, TargetJITInfo &tji,
81         JITMemoryManager *JMM)
82  : ExecutionEngine(MP), TM(tm), TJI(tji), jitstate(MP) {
83  setTargetData(TM.getTargetData());
84
85  // Initialize MCE
86  MCE = createEmitter(*this, JMM);
87
88  // Add target data
89  MutexGuard locked(lock);
90  FunctionPassManager &PM = jitstate.getPM(locked);
91  PM.add(new TargetData(*TM.getTargetData()));
92
93  // Turn the machine code intermediate representation into bytes in memory that
94  // may be executed.
95  if (TM.addPassesToEmitMachineCode(PM, *MCE, false /*fast*/)) {
96    cerr << "Target does not support machine code emission!\n";
97    abort();
98  }
99
100  // Initialize passes.
101  PM.doInitialization();
102}
103
104JIT::~JIT() {
105  delete MCE;
106  delete &TM;
107}
108
109/// run - Start execution with the specified function and arguments.
110///
111GenericValue JIT::runFunction(Function *F,
112                              const std::vector<GenericValue> &ArgValues) {
113  assert(F && "Function *F was null at entry to run()");
114
115  void *FPtr = getPointerToFunction(F);
116  assert(FPtr && "Pointer to fn's code was null after getPointerToFunction");
117  const FunctionType *FTy = F->getFunctionType();
118  const Type *RetTy = FTy->getReturnType();
119
120  assert((FTy->getNumParams() <= ArgValues.size() || FTy->isVarArg()) &&
121         "Too many arguments passed into function!");
122  assert(FTy->getNumParams() == ArgValues.size() &&
123         "This doesn't support passing arguments through varargs (yet)!");
124
125  // Handle some common cases first.  These cases correspond to common `main'
126  // prototypes.
127  if (RetTy == Type::Int32Ty || RetTy == Type::VoidTy) {
128    switch (ArgValues.size()) {
129    case 3:
130      if (FTy->getParamType(0) == Type::Int32Ty &&
131          isa<PointerType>(FTy->getParamType(1)) &&
132          isa<PointerType>(FTy->getParamType(2))) {
133        int (*PF)(int, char **, const char **) =
134          (int(*)(int, char **, const char **))(intptr_t)FPtr;
135
136        // Call the function.
137        GenericValue rv;
138        rv.IntVal = APInt(32, PF(ArgValues[0].IntVal.getZExtValue(),
139                                 (char **)GVTOP(ArgValues[1]),
140                                 (const char **)GVTOP(ArgValues[2])));
141        return rv;
142      }
143      break;
144    case 2:
145      if (FTy->getParamType(0) == Type::Int32Ty &&
146          isa<PointerType>(FTy->getParamType(1))) {
147        int (*PF)(int, char **) = (int(*)(int, char **))(intptr_t)FPtr;
148
149        // Call the function.
150        GenericValue rv;
151        rv.IntVal = APInt(32, PF(ArgValues[0].IntVal.getZExtValue(),
152                                 (char **)GVTOP(ArgValues[1])));
153        return rv;
154      }
155      break;
156    case 1:
157      if (FTy->getNumParams() == 1 &&
158          FTy->getParamType(0) == Type::Int32Ty) {
159        GenericValue rv;
160        int (*PF)(int) = (int(*)(int))(intptr_t)FPtr;
161        rv.IntVal = APInt(32, PF(ArgValues[0].IntVal.getZExtValue()));
162        return rv;
163      }
164      break;
165    }
166  }
167
168  // Handle cases where no arguments are passed first.
169  if (ArgValues.empty()) {
170    GenericValue rv;
171    switch (RetTy->getTypeID()) {
172    default: assert(0 && "Unknown return type for function call!");
173    case Type::IntegerTyID: {
174      unsigned BitWidth = cast<IntegerType>(RetTy)->getBitWidth();
175      if (BitWidth == 1)
176        rv.IntVal = APInt(BitWidth, ((bool(*)())(intptr_t)FPtr)());
177      else if (BitWidth <= 8)
178        rv.IntVal = APInt(BitWidth, ((char(*)())(intptr_t)FPtr)());
179      else if (BitWidth <= 16)
180        rv.IntVal = APInt(BitWidth, ((short(*)())(intptr_t)FPtr)());
181      else if (BitWidth <= 32)
182        rv.IntVal = APInt(BitWidth, ((int(*)())(intptr_t)FPtr)());
183      else if (BitWidth <= 64)
184        rv.IntVal = APInt(BitWidth, ((int64_t(*)())(intptr_t)FPtr)());
185      else
186        assert(0 && "Integer types > 64 bits not supported");
187      return rv;
188    }
189    case Type::VoidTyID:
190      rv.IntVal = APInt(32, ((int(*)())(intptr_t)FPtr)());
191      return rv;
192    case Type::FloatTyID:
193      rv.FloatVal = ((float(*)())(intptr_t)FPtr)();
194      return rv;
195    case Type::DoubleTyID:
196      rv.DoubleVal = ((double(*)())(intptr_t)FPtr)();
197      return rv;
198    case Type::X86_FP80TyID:
199    case Type::FP128TyID:
200    case Type::PPC_FP128TyID:
201      assert(0 && "long double not supported yet");
202      return rv;
203    case Type::PointerTyID:
204      return PTOGV(((void*(*)())(intptr_t)FPtr)());
205    }
206  }
207
208  // Okay, this is not one of our quick and easy cases.  Because we don't have a
209  // full FFI, we have to codegen a nullary stub function that just calls the
210  // function we are interested in, passing in constants for all of the
211  // arguments.  Make this function and return.
212
213  // First, create the function.
214  FunctionType *STy=FunctionType::get(RetTy, std::vector<const Type*>(), false);
215  Function *Stub = new Function(STy, Function::InternalLinkage, "",
216                                F->getParent());
217
218  // Insert a basic block.
219  BasicBlock *StubBB = new BasicBlock("", Stub);
220
221  // Convert all of the GenericValue arguments over to constants.  Note that we
222  // currently don't support varargs.
223  SmallVector<Value*, 8> Args;
224  for (unsigned i = 0, e = ArgValues.size(); i != e; ++i) {
225    Constant *C = 0;
226    const Type *ArgTy = FTy->getParamType(i);
227    const GenericValue &AV = ArgValues[i];
228    switch (ArgTy->getTypeID()) {
229    default: assert(0 && "Unknown argument type for function call!");
230    case Type::IntegerTyID: C = ConstantInt::get(AV.IntVal); break;
231    case Type::FloatTyID:   C = ConstantFP ::get(ArgTy, APFloat(AV.FloatVal));
232                            break;
233    case Type::DoubleTyID:  C = ConstantFP ::get(ArgTy, APFloat(AV.DoubleVal));
234                            break;
235    case Type::PPC_FP128TyID:
236    case Type::X86_FP80TyID:
237    case Type::FP128TyID:   C = ConstantFP ::get(ArgTy, APFloat(AV.IntVal));
238                            break;
239    case Type::PointerTyID:
240      void *ArgPtr = GVTOP(AV);
241      if (sizeof(void*) == 4) {
242        C = ConstantInt::get(Type::Int32Ty, (int)(intptr_t)ArgPtr);
243      } else {
244        C = ConstantInt::get(Type::Int64Ty, (intptr_t)ArgPtr);
245      }
246      C = ConstantExpr::getIntToPtr(C, ArgTy);  // Cast the integer to pointer
247      break;
248    }
249    Args.push_back(C);
250  }
251
252  CallInst *TheCall = new CallInst(F, Args.begin(), Args.end(), "", StubBB);
253  TheCall->setTailCall();
254  if (TheCall->getType() != Type::VoidTy)
255    new ReturnInst(TheCall, StubBB);             // Return result of the call.
256  else
257    new ReturnInst(StubBB);                      // Just return void.
258
259  // Finally, return the value returned by our nullary stub function.
260  return runFunction(Stub, std::vector<GenericValue>());
261}
262
263/// runJITOnFunction - Run the FunctionPassManager full of
264/// just-in-time compilation passes on F, hopefully filling in
265/// GlobalAddress[F] with the address of F's machine code.
266///
267void JIT::runJITOnFunction(Function *F) {
268  static bool isAlreadyCodeGenerating = false;
269
270  MutexGuard locked(lock);
271  assert(!isAlreadyCodeGenerating && "Error: Recursive compilation detected!");
272
273  // JIT the function
274  isAlreadyCodeGenerating = true;
275  jitstate.getPM(locked).run(*F);
276  isAlreadyCodeGenerating = false;
277
278  // If the function referred to a global variable that had not yet been
279  // emitted, it allocates memory for the global, but doesn't emit it yet.  Emit
280  // all of these globals now.
281  while (!jitstate.getPendingGlobals(locked).empty()) {
282    const GlobalVariable *GV = jitstate.getPendingGlobals(locked).back();
283    jitstate.getPendingGlobals(locked).pop_back();
284    EmitGlobalVariable(GV);
285  }
286}
287
288/// getPointerToFunction - This method is used to get the address of the
289/// specified function, compiling it if neccesary.
290///
291void *JIT::getPointerToFunction(Function *F) {
292  MutexGuard locked(lock);
293
294  if (void *Addr = getPointerToGlobalIfAvailable(F))
295    return Addr;   // Check if function already code gen'd
296
297  // Make sure we read in the function if it exists in this Module.
298  if (F->hasNotBeenReadFromBitcode()) {
299    // Determine the module provider this function is provided by.
300    Module *M = F->getParent();
301    ModuleProvider *MP = 0;
302    for (unsigned i = 0, e = Modules.size(); i != e; ++i) {
303      if (Modules[i]->getModule() == M) {
304        MP = Modules[i];
305        break;
306      }
307    }
308    assert(MP && "Function isn't in a module we know about!");
309
310    std::string ErrorMsg;
311    if (MP->materializeFunction(F, &ErrorMsg)) {
312      cerr << "Error reading function '" << F->getName()
313           << "' from bitcode file: " << ErrorMsg << "\n";
314      abort();
315    }
316  }
317
318  if (F->isDeclaration()) {
319    void *Addr = getPointerToNamedFunction(F->getName());
320    addGlobalMapping(F, Addr);
321    return Addr;
322  }
323
324  runJITOnFunction(F);
325
326  void *Addr = getPointerToGlobalIfAvailable(F);
327  assert(Addr && "Code generation didn't add function to GlobalAddress table!");
328  return Addr;
329}
330
331/// getOrEmitGlobalVariable - Return the address of the specified global
332/// variable, possibly emitting it to memory if needed.  This is used by the
333/// Emitter.
334void *JIT::getOrEmitGlobalVariable(const GlobalVariable *GV) {
335  MutexGuard locked(lock);
336
337  void *Ptr = getPointerToGlobalIfAvailable(GV);
338  if (Ptr) return Ptr;
339
340  // If the global is external, just remember the address.
341  if (GV->isDeclaration()) {
342#if HAVE___DSO_HANDLE
343    if (GV->getName() == "__dso_handle")
344      return (void*)&__dso_handle;
345#endif
346    Ptr = sys::DynamicLibrary::SearchForAddressOfSymbol(GV->getName().c_str());
347    if (Ptr == 0) {
348      cerr << "Could not resolve external global address: "
349           << GV->getName() << "\n";
350      abort();
351    }
352  } else {
353    // If the global hasn't been emitted to memory yet, allocate space.  We will
354    // actually initialize the global after current function has finished
355    // compilation.
356    const Type *GlobalType = GV->getType()->getElementType();
357    size_t S = getTargetData()->getABITypeSize(GlobalType);
358    size_t A = getTargetData()->getPrefTypeAlignment(GlobalType);
359    if (A <= 8) {
360      Ptr = malloc(S);
361    } else {
362      // Allocate S+A bytes of memory, then use an aligned pointer within that
363      // space.
364      Ptr = malloc(S+A);
365      unsigned MisAligned = ((intptr_t)Ptr & (A-1));
366      Ptr = (char*)Ptr + (MisAligned ? (A-MisAligned) : 0);
367    }
368    jitstate.getPendingGlobals(locked).push_back(GV);
369  }
370  addGlobalMapping(GV, Ptr);
371  return Ptr;
372}
373
374
375/// recompileAndRelinkFunction - This method is used to force a function
376/// which has already been compiled, to be compiled again, possibly
377/// after it has been modified. Then the entry to the old copy is overwritten
378/// with a branch to the new copy. If there was no old copy, this acts
379/// just like JIT::getPointerToFunction().
380///
381void *JIT::recompileAndRelinkFunction(Function *F) {
382  void *OldAddr = getPointerToGlobalIfAvailable(F);
383
384  // If it's not already compiled there is no reason to patch it up.
385  if (OldAddr == 0) { return getPointerToFunction(F); }
386
387  // Delete the old function mapping.
388  addGlobalMapping(F, 0);
389
390  // Recodegen the function
391  runJITOnFunction(F);
392
393  // Update state, forward the old function to the new function.
394  void *Addr = getPointerToGlobalIfAvailable(F);
395  assert(Addr && "Code generation didn't add function to GlobalAddress table!");
396  TJI.replaceMachineCodeForFunction(OldAddr, Addr);
397  return Addr;
398}
399
400