llvm/lib/ExecutionEngine/Interpreter/ExternalFunctions.cpp

   1 //===-- ExternalFunctions.cpp - Implement External Functions --------------===//
   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 //  This file contains both code to deal with invoking "external" functions, but
  10 //  also contains code that implements "exported" external functions.
  11 //
  12 //  There are currently two mechanisms for handling external functions in the
  13 //  Interpreter.  The first is to implement lle_* wrapper functions that are
  14 //  specific to well-known library functions which manually translate the
  15 //  arguments from GenericValues and make the call.  If such a wrapper does
  16 //  not exist, and libffi is available, then the Interpreter will attempt to
  17 //  invoke the function using libffi, after finding its address.
  18 //
  19 //===----------------------------------------------------------------------===//
  20
  21 #include "Interpreter.h"
  22 #include "llvm/ADT/APInt.h"
  23 #include "llvm/ADT/ArrayRef.h"
  24 #include "llvm/Config/config.h" // Detect libffi
  25 #include "llvm/ExecutionEngine/GenericValue.h"
  26 #include "llvm/IR/DataLayout.h"
  27 #include "llvm/IR/DerivedTypes.h"
  28 #include "llvm/IR/Function.h"
  29 #include "llvm/IR/Type.h"
  30 #include "llvm/Support/Casting.h"
  31 #include "llvm/Support/DynamicLibrary.h"
  32 #include "llvm/Support/ErrorHandling.h"
  33 #include "llvm/Support/ManagedStatic.h"
  34 #include "llvm/Support/Mutex.h"
  35 #include "llvm/Support/raw_ostream.h"
  36 #include <cassert>
  37 #include <cmath>
  38 #include <csignal>
  39 #include <cstdint>
  40 #include <cstdio>
  41 #include <cstring>
  42 #include <map>
  43 #include <mutex>
  44 #include <string>
  45 #include <utility>
  46 #include <vector>
  47
  48 #ifdef HAVE_FFI_CALL
  49 #ifdef HAVE_FFI_H
  50 #include <ffi.h>
  51 #define USE_LIBFFI
  52 #elif HAVE_FFI_FFI_H
  53 #include <ffi/ffi.h>
  54 #define USE_LIBFFI
  55 #endif
  56 #endif
  57
  58 using namespace llvm;
  59
  60 static ManagedStatic<sys::Mutex> FunctionsLock;
  61
  62 typedef GenericValue (*ExFunc)(FunctionType *, ArrayRef<GenericValue>);
  63 static ManagedStatic<std::map<const Function *, ExFunc> > ExportedFunctions;
  64 static ManagedStatic<std::map<std::string, ExFunc> > FuncNames;
  65
  66 #ifdef USE_LIBFFI
  67 typedef void (*RawFunc)();
  68 static ManagedStatic<std::map<const Function *, RawFunc> > RawFunctions;
  69 #endif
  70
  71 static Interpreter *TheInterpreter;
  72
  73 static char getTypeID(Type *Ty) {
  74   switch (Ty->getTypeID()) {
  75   case Type::VoidTyID:    return 'V';
  76   case Type::IntegerTyID:
  77     switch (cast<IntegerType>(Ty)->getBitWidth()) {
  78       case 1:  return 'o';
  79       case 8:  return 'B';
  80       case 16: return 'S';
  81       case 32: return 'I';
  82       case 64: return 'L';
  83       default: return 'N';
  84     }
  85   case Type::FloatTyID:   return 'F';
  86   case Type::DoubleTyID:  return 'D';
  87   case Type::PointerTyID: return 'P';
  88   case Type::FunctionTyID:return 'M';
  89   case Type::StructTyID:  return 'T';
  90   case Type::ArrayTyID:   return 'A';
  91   default: return 'U';
  92   }
  93 }
  94
  95 // Try to find address of external function given a Function object.
  96 // Please note, that interpreter doesn't know how to assemble a
  97 // real call in general case (this is JIT job), that's why it assumes,
  98 // that all external functions has the same (and pretty "general") signature.
  99 // The typical example of such functions are "lle_X_" ones.
 100 static ExFunc lookupFunction(const Function *F) {
 101   // Function not found, look it up... start by figuring out what the
 102   // composite function name should be.
 103   std::string ExtName = "lle_";
 104   FunctionType *FT = F->getFunctionType();
 105   ExtName += getTypeID(FT->getReturnType());
 106   for (Type *T : FT->params())
 107     ExtName += getTypeID(T);
 108   ExtName += ("_" + F->getName()).str();
 109
 110   sys::ScopedLock Writer(*FunctionsLock);
 111   ExFunc FnPtr = (*FuncNames)[ExtName];
 112   if (!FnPtr)
 113     FnPtr = (*FuncNames)[("lle_X_" + F->getName()).str()];
 114   if (!FnPtr)  // Try calling a generic function... if it exists...
 115     FnPtr = (ExFunc)(intptr_t)sys::DynamicLibrary::SearchForAddressOfSymbol(
 116         ("lle_X_" + F->getName()).str());
 117   if (FnPtr)
 118     ExportedFunctions->insert(std::make_pair(F, FnPtr));  // Cache for later
 119   return FnPtr;
 120 }
 121
 122 #ifdef USE_LIBFFI
 123 static ffi_type *ffiTypeFor(Type *Ty) {
 124   switch (Ty->getTypeID()) {
 125     case Type::VoidTyID: return &ffi_type_void;
 126     case Type::IntegerTyID:
 127       switch (cast<IntegerType>(Ty)->getBitWidth()) {
 128         case 8:  return &ffi_type_sint8;
 129         case 16: return &ffi_type_sint16;
 130         case 32: return &ffi_type_sint32;
 131         case 64: return &ffi_type_sint64;
 132       }
 133       llvm_unreachable("Unhandled integer type bitwidth");
 134     case Type::FloatTyID:   return &ffi_type_float;
 135     case Type::DoubleTyID:  return &ffi_type_double;
 136     case Type::PointerTyID: return &ffi_type_pointer;
 137     default: break;
 138   }
 139   // TODO: Support other types such as StructTyID, ArrayTyID, OpaqueTyID, etc.
 140   report_fatal_error("Type could not be mapped for use with libffi.");
 141   return NULL;
 142 }
 143
 144 static void *ffiValueFor(Type *Ty, const GenericValue &AV,
 145                          void *ArgDataPtr) {
 146   switch (Ty->getTypeID()) {
 147     case Type::IntegerTyID:
 148       switch (cast<IntegerType>(Ty)->getBitWidth()) {
 149         case 8: {
 150           int8_t *I8Ptr = (int8_t *) ArgDataPtr;
 151           *I8Ptr = (int8_t) AV.IntVal.getZExtValue();
 152           return ArgDataPtr;
 153         }
 154         case 16: {
 155           int16_t *I16Ptr = (int16_t *) ArgDataPtr;
 156           *I16Ptr = (int16_t) AV.IntVal.getZExtValue();
 157           return ArgDataPtr;
 158         }
 159         case 32: {
 160           int32_t *I32Ptr = (int32_t *) ArgDataPtr;
 161           *I32Ptr = (int32_t) AV.IntVal.getZExtValue();
 162           return ArgDataPtr;
 163         }
 164         case 64: {
 165           int64_t *I64Ptr = (int64_t *) ArgDataPtr;
 166           *I64Ptr = (int64_t) AV.IntVal.getZExtValue();
 167           return ArgDataPtr;
 168         }
 169       }
 170       llvm_unreachable("Unhandled integer type bitwidth");
 171     case Type::FloatTyID: {
 172       float *FloatPtr = (float *) ArgDataPtr;
 173       *FloatPtr = AV.FloatVal;
 174       return ArgDataPtr;
 175     }
 176     case Type::DoubleTyID: {
 177       double *DoublePtr = (double *) ArgDataPtr;
 178       *DoublePtr = AV.DoubleVal;
 179       return ArgDataPtr;
 180     }
 181     case Type::PointerTyID: {
 182       void **PtrPtr = (void **) ArgDataPtr;
 183       *PtrPtr = GVTOP(AV);
 184       return ArgDataPtr;
 185     }
 186     default: break;
 187   }
 188   // TODO: Support other types such as StructTyID, ArrayTyID, OpaqueTyID, etc.
 189   report_fatal_error("Type value could not be mapped for use with libffi.");
 190   return NULL;
 191 }
 192
 193 static bool ffiInvoke(RawFunc Fn, Function *F, ArrayRef<GenericValue> ArgVals,
 194                       const DataLayout &TD, GenericValue &Result) {
 195   ffi_cif cif;
 196   FunctionType *FTy = F->getFunctionType();
 197   const unsigned NumArgs = F->arg_size();
 198
 199   // TODO: We don't have type information about the remaining arguments, because
 200   // this information is never passed into ExecutionEngine::runFunction().
 201   if (ArgVals.size() > NumArgs && F->isVarArg()) {
 202     report_fatal_error("Calling external var arg function '" + F->getName()
 203                       + "' is not supported by the Interpreter.");
 204   }
 205
 206   unsigned ArgBytes = 0;
 207
 208   std::vector<ffi_type*> args(NumArgs);
 209   for (Function::const_arg_iterator A = F->arg_begin(), E = F->arg_end();
 210        A != E; ++A) {
 211     const unsigned ArgNo = A->getArgNo();
 212     Type *ArgTy = FTy->getParamType(ArgNo);
 213     args[ArgNo] = ffiTypeFor(ArgTy);
 214     ArgBytes += TD.getTypeStoreSize(ArgTy);
 215   }
 216
 217   SmallVector<uint8_t, 128> ArgData;
 218   ArgData.resize(ArgBytes);
 219   uint8_t *ArgDataPtr = ArgData.data();
 220   SmallVector<void*, 16> values(NumArgs);
 221   for (Function::const_arg_iterator A = F->arg_begin(), E = F->arg_end();
 222        A != E; ++A) {
 223     const unsigned ArgNo = A->getArgNo();
 224     Type *ArgTy = FTy->getParamType(ArgNo);
 225     values[ArgNo] = ffiValueFor(ArgTy, ArgVals[ArgNo], ArgDataPtr);
 226     ArgDataPtr += TD.getTypeStoreSize(ArgTy);
 227   }
 228
 229   Type *RetTy = FTy->getReturnType();
 230   ffi_type *rtype = ffiTypeFor(RetTy);
 231
 232   if (ffi_prep_cif(&cif, FFI_DEFAULT_ABI, NumArgs, rtype, args.data()) ==
 233       FFI_OK) {
 234     SmallVector<uint8_t, 128> ret;
 235     if (RetTy->getTypeID() != Type::VoidTyID)
 236       ret.resize(TD.getTypeStoreSize(RetTy));
 237     ffi_call(&cif, Fn, ret.data(), values.data());
 238     switch (RetTy->getTypeID()) {
 239       case Type::IntegerTyID:
 240         switch (cast<IntegerType>(RetTy)->getBitWidth()) {
 241           case 8:  Result.IntVal = APInt(8 , *(int8_t *) ret.data()); break;
 242           case 16: Result.IntVal = APInt(16, *(int16_t*) ret.data()); break;
 243           case 32: Result.IntVal = APInt(32, *(int32_t*) ret.data()); break;
 244           case 64: Result.IntVal = APInt(64, *(int64_t*) ret.data()); break;
 245         }
 246         break;
 247       case Type::FloatTyID:   Result.FloatVal   = *(float *) ret.data(); break;
 248       case Type::DoubleTyID:  Result.DoubleVal  = *(double*) ret.data(); break;
 249       case Type::PointerTyID: Result.PointerVal = *(void **) ret.data(); break;
 250       default: break;
 251     }
 252     return true;
 253   }
 254
 255   return false;
 256 }
 257 #endif // USE_LIBFFI
 258
 259 GenericValue Interpreter::callExternalFunction(Function *F,
 260                                                ArrayRef<GenericValue> ArgVals) {
 261   TheInterpreter = this;
 262
 263   std::unique_lock<sys::Mutex> Guard(*FunctionsLock);
 264
 265   // Do a lookup to see if the function is in our cache... this should just be a
 266   // deferred annotation!
 267   std::map<const Function *, ExFunc>::iterator FI = ExportedFunctions->find(F);
 268   if (ExFunc Fn = (FI == ExportedFunctions->end()) ? lookupFunction(F)
 269                                                    : FI->second) {
 270     Guard.unlock();
 271     return Fn(F->getFunctionType(), ArgVals);
 272   }
 273
 274 #ifdef USE_LIBFFI
 275   std::map<const Function *, RawFunc>::iterator RF = RawFunctions->find(F);
 276   RawFunc RawFn;
 277   if (RF == RawFunctions->end()) {
 278     RawFn = (RawFunc)(intptr_t)
 279       sys::DynamicLibrary::SearchForAddressOfSymbol(std::string(F->getName()));
 280     if (!RawFn)
 281       RawFn = (RawFunc)(intptr_t)getPointerToGlobalIfAvailable(F);
 282     if (RawFn != 0)
 283       RawFunctions->insert(std::make_pair(F, RawFn));  // Cache for later
 284   } else {
 285     RawFn = RF->second;
 286   }
 287
 288   Guard.unlock();
 289
 290   GenericValue Result;
 291   if (RawFn != 0 && ffiInvoke(RawFn, F, ArgVals, getDataLayout(), Result))
 292     return Result;
 293 #endif // USE_LIBFFI
 294
 295   if (F->getName() == "__main")
 296     errs() << "Tried to execute an unknown external function: "
 297       << *F->getType() << " __main\n";
 298   else
 299     report_fatal_error("Tried to execute an unknown external function: " +
 300                        F->getName());
 301 #ifndef USE_LIBFFI
 302   errs() << "Recompiling LLVM with --enable-libffi might help.\n";
 303 #endif
 304   return GenericValue();
 305 }
 306
 307 //===----------------------------------------------------------------------===//
 308 //  Functions "exported" to the running application...
 309 //
 310
 311 // void atexit(Function*)
 312 static GenericValue lle_X_atexit(FunctionType *FT,
 313                                  ArrayRef<GenericValue> Args) {
 314   assert(Args.size() == 1);
 315   TheInterpreter->addAtExitHandler((Function*)GVTOP(Args[0]));
 316   GenericValue GV;
 317   GV.IntVal = 0;
 318   return GV;
 319 }
 320
 321 // void exit(int)
 322 static GenericValue lle_X_exit(FunctionType *FT, ArrayRef<GenericValue> Args) {
 323   TheInterpreter->exitCalled(Args[0]);
 324   return GenericValue();
 325 }
 326
 327 // void abort(void)
 328 static GenericValue lle_X_abort(FunctionType *FT, ArrayRef<GenericValue> Args) {
 329   //FIXME: should we report or raise here?
 330   //report_fatal_error("Interpreted program raised SIGABRT");
 331   raise (SIGABRT);
 332   return GenericValue();
 333 }
 334
 335 // int sprintf(char *, const char *, ...) - a very rough implementation to make
 336 // output useful.
 337 static GenericValue lle_X_sprintf(FunctionType *FT,
 338                                   ArrayRef<GenericValue> Args) {
 339   char *OutputBuffer = (char *)GVTOP(Args[0]);
 340   const char *FmtStr = (const char *)GVTOP(Args[1]);
 341   unsigned ArgNo = 2;
 342
 343   // printf should return # chars printed.  This is completely incorrect, but
 344   // close enough for now.
 345   GenericValue GV;
 346   GV.IntVal = APInt(32, strlen(FmtStr));
 347   while (true) {
 348     switch (*FmtStr) {
 349     case 0: return GV;             // Null terminator...
 350     default:                       // Normal nonspecial character
 351       sprintf(OutputBuffer++, "%c", *FmtStr++);
 352       break;
 353     case '\\': {                   // Handle escape codes
 354       sprintf(OutputBuffer, "%c%c", *FmtStr, *(FmtStr+1));
 355       FmtStr += 2; OutputBuffer += 2;
 356       break;
 357     }
 358     case '%': {                    // Handle format specifiers
 359       char FmtBuf[100] = "", Buffer[1000] = "";
 360       char *FB = FmtBuf;
 361       *FB++ = *FmtStr++;
 362       char Last = *FB++ = *FmtStr++;
 363       unsigned HowLong = 0;
 364       while (Last != 'c' && Last != 'd' && Last != 'i' && Last != 'u' &&
 365              Last != 'o' && Last != 'x' && Last != 'X' && Last != 'e' &&
 366              Last != 'E' && Last != 'g' && Last != 'G' && Last != 'f' &&
 367              Last != 'p' && Last != 's' && Last != '%') {
 368         if (Last == 'l' || Last == 'L') HowLong++;  // Keep track of l's
 369         Last = *FB++ = *FmtStr++;
 370       }
 371       *FB = 0;
 372
 373       switch (Last) {
 374       case '%':
 375         memcpy(Buffer, "%", 2); break;
 376       case 'c':
 377         sprintf(Buffer, FmtBuf, uint32_t(Args[ArgNo++].IntVal.getZExtValue()));
 378         break;
 379       case 'd': case 'i':
 380       case 'u': case 'o':
 381       case 'x': case 'X':
 382         if (HowLong >= 1) {
 383           if (HowLong == 1 &&
 384               TheInterpreter->getDataLayout().getPointerSizeInBits() == 64 &&
 385               sizeof(long) < sizeof(int64_t)) {
 386             // Make sure we use %lld with a 64 bit argument because we might be
 387             // compiling LLI on a 32 bit compiler.
 388             unsigned Size = strlen(FmtBuf);
 389             FmtBuf[Size] = FmtBuf[Size-1];
 390             FmtBuf[Size+1] = 0;
 391             FmtBuf[Size-1] = 'l';
 392           }
 393           sprintf(Buffer, FmtBuf, Args[ArgNo++].IntVal.getZExtValue());
 394         } else
 395           sprintf(Buffer, FmtBuf,uint32_t(Args[ArgNo++].IntVal.getZExtValue()));
 396         break;
 397       case 'e': case 'E': case 'g': case 'G': case 'f':
 398         sprintf(Buffer, FmtBuf, Args[ArgNo++].DoubleVal); break;
 399       case 'p':
 400         sprintf(Buffer, FmtBuf, (void*)GVTOP(Args[ArgNo++])); break;
 401       case 's':
 402         sprintf(Buffer, FmtBuf, (char*)GVTOP(Args[ArgNo++])); break;
 403       default:
 404         errs() << "<unknown printf code '" << *FmtStr << "'!>";
 405         ArgNo++; break;
 406       }
 407       size_t Len = strlen(Buffer);
 408       memcpy(OutputBuffer, Buffer, Len + 1);
 409       OutputBuffer += Len;
 410       }
 411       break;
 412     }
 413   }
 414   return GV;
 415 }
 416
 417 // int printf(const char *, ...) - a very rough implementation to make output
 418 // useful.
 419 static GenericValue lle_X_printf(FunctionType *FT,
 420                                  ArrayRef<GenericValue> Args) {
 421   char Buffer[10000];
 422   std::vector<GenericValue> NewArgs;
 423   NewArgs.push_back(PTOGV((void*)&Buffer[0]));
 424   llvm::append_range(NewArgs, Args);
 425   GenericValue GV = lle_X_sprintf(FT, NewArgs);
 426   outs() << Buffer;
 427   return GV;
 428 }
 429
 430 // int sscanf(const char *format, ...);
 431 static GenericValue lle_X_sscanf(FunctionType *FT,
 432                                  ArrayRef<GenericValue> args) {
 433   assert(args.size() < 10 && "Only handle up to 10 args to sscanf right now!");
 434
 435   char *Args[10];
 436   for (unsigned i = 0; i < args.size(); ++i)
 437     Args[i] = (char*)GVTOP(args[i]);
 438
 439   GenericValue GV;
 440   GV.IntVal = APInt(32, sscanf(Args[0], Args[1], Args[2], Args[3], Args[4],
 441                     Args[5], Args[6], Args[7], Args[8], Args[9]));
 442   return GV;
 443 }
 444
 445 // int scanf(const char *format, ...);
 446 static GenericValue lle_X_scanf(FunctionType *FT, ArrayRef<GenericValue> args) {
 447   assert(args.size() < 10 && "Only handle up to 10 args to scanf right now!");
 448
 449   char *Args[10];
 450   for (unsigned i = 0; i < args.size(); ++i)
 451     Args[i] = (char*)GVTOP(args[i]);
 452
 453   GenericValue GV;
 454   GV.IntVal = APInt(32, scanf( Args[0], Args[1], Args[2], Args[3], Args[4],
 455                     Args[5], Args[6], Args[7], Args[8], Args[9]));
 456   return GV;
 457 }
 458
 459 // int fprintf(FILE *, const char *, ...) - a very rough implementation to make
 460 // output useful.
 461 static GenericValue lle_X_fprintf(FunctionType *FT,
 462                                   ArrayRef<GenericValue> Args) {
 463   assert(Args.size() >= 2);
 464   char Buffer[10000];
 465   std::vector<GenericValue> NewArgs;
 466   NewArgs.push_back(PTOGV(Buffer));
 467   NewArgs.insert(NewArgs.end(), Args.begin()+1, Args.end());
 468   GenericValue GV = lle_X_sprintf(FT, NewArgs);
 469
 470   fputs(Buffer, (FILE *) GVTOP(Args[0]));
 471   return GV;
 472 }
 473
 474 static GenericValue lle_X_memset(FunctionType *FT,
 475                                  ArrayRef<GenericValue> Args) {
 476   int val = (int)Args[1].IntVal.getSExtValue();
 477   size_t len = (size_t)Args[2].IntVal.getZExtValue();
 478   memset((void *)GVTOP(Args[0]), val, len);
 479   // llvm.memset.* returns void, lle_X_* returns GenericValue,
 480   // so here we return GenericValue with IntVal set to zero
 481   GenericValue GV;
 482   GV.IntVal = 0;
 483   return GV;
 484 }
 485
 486 static GenericValue lle_X_memcpy(FunctionType *FT,
 487                                  ArrayRef<GenericValue> Args) {
 488   memcpy(GVTOP(Args[0]), GVTOP(Args[1]),
 489          (size_t)(Args[2].IntVal.getLimitedValue()));
 490
 491   // llvm.memcpy* returns void, lle_X_* returns GenericValue,
 492   // so here we return GenericValue with IntVal set to zero
 493   GenericValue GV;
 494   GV.IntVal = 0;
 495   return GV;
 496 }
 497
 498 void Interpreter::initializeExternalFunctions() {
 499   sys::ScopedLock Writer(*FunctionsLock);
 500   (*FuncNames)["lle_X_atexit"]       = lle_X_atexit;
 501   (*FuncNames)["lle_X_exit"]         = lle_X_exit;
 502   (*FuncNames)["lle_X_abort"]        = lle_X_abort;
 503
 504   (*FuncNames)["lle_X_printf"]       = lle_X_printf;
 505   (*FuncNames)["lle_X_sprintf"]      = lle_X_sprintf;
 506   (*FuncNames)["lle_X_sscanf"]       = lle_X_sscanf;
 507   (*FuncNames)["lle_X_scanf"]        = lle_X_scanf;
 508   (*FuncNames)["lle_X_fprintf"]      = lle_X_fprintf;
 509   (*FuncNames)["lle_X_memset"]       = lle_X_memset;
 510   (*FuncNames)["lle_X_memcpy"]       = lle_X_memcpy;
 511 }