533 lines · 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-exception6//7//===----------------------------------------------------------------------===//8//9// This file contains both code to deal with invoking "external" functions, but10// also contains code that implements "exported" external functions.11//12// There are currently two mechanisms for handling external functions in the13// Interpreter. The first is to implement lle_* wrapper functions that are14// specific to well-known library functions which manually translate the15// arguments from GenericValues and make the call. If such a wrapper does16// not exist, and libffi is available, then the Interpreter will attempt to17// 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 libffi25#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/Mutex.h"34#include "llvm/Support/raw_ostream.h"35#include <cassert>36#include <cmath>37#include <csignal>38#include <cstdint>39#include <cstdio>40#include <cstring>41#include <map>42#include <mutex>43#include <string>44#include <vector>45 46#ifdef HAVE_FFI_CALL47#ifdef HAVE_FFI_H48#include <ffi.h>49#define USE_LIBFFI50#elif HAVE_FFI_FFI_H51#include <ffi/ffi.h>52#define USE_LIBFFI53#endif54#endif55 56using namespace llvm;57 58namespace {59 60typedef GenericValue (*ExFunc)(FunctionType *, ArrayRef<GenericValue>);61typedef void (*RawFunc)();62 63struct Functions {64 sys::Mutex Lock;65 std::map<const Function *, ExFunc> ExportedFunctions;66 std::map<std::string, ExFunc> FuncNames;67#ifdef USE_LIBFFI68 std::map<const Function *, RawFunc> RawFunctions;69#endif70};71 72Functions &getFunctions() {73 static Functions F;74 return F;75}76 77} // anonymous namespace78 79static Interpreter *TheInterpreter;80 81static char getTypeID(Type *Ty) {82 switch (Ty->getTypeID()) {83 case Type::VoidTyID: return 'V';84 case Type::IntegerTyID:85 switch (cast<IntegerType>(Ty)->getBitWidth()) {86 case 1: return 'o';87 case 8: return 'B';88 case 16: return 'S';89 case 32: return 'I';90 case 64: return 'L';91 default: return 'N';92 }93 case Type::FloatTyID: return 'F';94 case Type::DoubleTyID: return 'D';95 case Type::PointerTyID: return 'P';96 case Type::FunctionTyID:return 'M';97 case Type::StructTyID: return 'T';98 case Type::ArrayTyID: return 'A';99 default: return 'U';100 }101}102 103// Try to find address of external function given a Function object.104// Please note, that interpreter doesn't know how to assemble a105// real call in general case (this is JIT job), that's why it assumes,106// that all external functions has the same (and pretty "general") signature.107// The typical example of such functions are "lle_X_" ones.108static ExFunc lookupFunction(const Function *F) {109 // Function not found, look it up... start by figuring out what the110 // composite function name should be.111 std::string ExtName = "lle_";112 FunctionType *FT = F->getFunctionType();113 ExtName += getTypeID(FT->getReturnType());114 for (Type *T : FT->params())115 ExtName += getTypeID(T);116 ExtName += ("_" + F->getName()).str();117 118 auto &Fns = getFunctions();119 sys::ScopedLock Writer(Fns.Lock);120 ExFunc FnPtr = Fns.FuncNames[ExtName];121 if (!FnPtr)122 FnPtr = Fns.FuncNames[("lle_X_" + F->getName()).str()];123 if (!FnPtr) // Try calling a generic function... if it exists...124 FnPtr = (ExFunc)(intptr_t)sys::DynamicLibrary::SearchForAddressOfSymbol(125 ("lle_X_" + F->getName()).str());126 if (FnPtr)127 Fns.ExportedFunctions.insert(std::make_pair(F, FnPtr)); // Cache for later128 return FnPtr;129}130 131#ifdef USE_LIBFFI132static ffi_type *ffiTypeFor(Type *Ty) {133 switch (Ty->getTypeID()) {134 case Type::VoidTyID: return &ffi_type_void;135 case Type::IntegerTyID:136 switch (cast<IntegerType>(Ty)->getBitWidth()) {137 case 8: return &ffi_type_sint8;138 case 16: return &ffi_type_sint16;139 case 32: return &ffi_type_sint32;140 case 64: return &ffi_type_sint64;141 }142 llvm_unreachable("Unhandled integer type bitwidth");143 case Type::FloatTyID: return &ffi_type_float;144 case Type::DoubleTyID: return &ffi_type_double;145 case Type::PointerTyID: return &ffi_type_pointer;146 default: break;147 }148 // TODO: Support other types such as StructTyID, ArrayTyID, OpaqueTyID, etc.149 report_fatal_error("Type could not be mapped for use with libffi.");150 return NULL;151}152 153static void *ffiValueFor(Type *Ty, const GenericValue &AV,154 void *ArgDataPtr) {155 switch (Ty->getTypeID()) {156 case Type::IntegerTyID:157 switch (cast<IntegerType>(Ty)->getBitWidth()) {158 case 8: {159 int8_t *I8Ptr = (int8_t *) ArgDataPtr;160 *I8Ptr = (int8_t) AV.IntVal.getZExtValue();161 return ArgDataPtr;162 }163 case 16: {164 int16_t *I16Ptr = (int16_t *) ArgDataPtr;165 *I16Ptr = (int16_t) AV.IntVal.getZExtValue();166 return ArgDataPtr;167 }168 case 32: {169 int32_t *I32Ptr = (int32_t *) ArgDataPtr;170 *I32Ptr = (int32_t) AV.IntVal.getZExtValue();171 return ArgDataPtr;172 }173 case 64: {174 int64_t *I64Ptr = (int64_t *) ArgDataPtr;175 *I64Ptr = (int64_t) AV.IntVal.getZExtValue();176 return ArgDataPtr;177 }178 }179 llvm_unreachable("Unhandled integer type bitwidth");180 case Type::FloatTyID: {181 float *FloatPtr = (float *) ArgDataPtr;182 *FloatPtr = AV.FloatVal;183 return ArgDataPtr;184 }185 case Type::DoubleTyID: {186 double *DoublePtr = (double *) ArgDataPtr;187 *DoublePtr = AV.DoubleVal;188 return ArgDataPtr;189 }190 case Type::PointerTyID: {191 void **PtrPtr = (void **) ArgDataPtr;192 *PtrPtr = GVTOP(AV);193 return ArgDataPtr;194 }195 default: break;196 }197 // TODO: Support other types such as StructTyID, ArrayTyID, OpaqueTyID, etc.198 report_fatal_error("Type value could not be mapped for use with libffi.");199 return NULL;200}201 202static bool ffiInvoke(RawFunc Fn, Function *F, ArrayRef<GenericValue> ArgVals,203 const DataLayout &TD, GenericValue &Result) {204 ffi_cif cif;205 FunctionType *FTy = F->getFunctionType();206 const unsigned NumArgs = F->arg_size();207 208 // TODO: We don't have type information about the remaining arguments, because209 // this information is never passed into ExecutionEngine::runFunction().210 if (ArgVals.size() > NumArgs && F->isVarArg()) {211 report_fatal_error("Calling external var arg function '" + F->getName()212 + "' is not supported by the Interpreter.");213 }214 215 unsigned ArgBytes = 0;216 217 std::vector<ffi_type*> args(NumArgs);218 for (Function::const_arg_iterator A = F->arg_begin(), E = F->arg_end();219 A != E; ++A) {220 const unsigned ArgNo = A->getArgNo();221 Type *ArgTy = FTy->getParamType(ArgNo);222 args[ArgNo] = ffiTypeFor(ArgTy);223 ArgBytes += TD.getTypeStoreSize(ArgTy);224 }225 226 SmallVector<uint8_t, 128> ArgData;227 ArgData.resize(ArgBytes);228 uint8_t *ArgDataPtr = ArgData.data();229 SmallVector<void*, 16> values(NumArgs);230 for (Function::const_arg_iterator A = F->arg_begin(), E = F->arg_end();231 A != E; ++A) {232 const unsigned ArgNo = A->getArgNo();233 Type *ArgTy = FTy->getParamType(ArgNo);234 values[ArgNo] = ffiValueFor(ArgTy, ArgVals[ArgNo], ArgDataPtr);235 ArgDataPtr += TD.getTypeStoreSize(ArgTy);236 }237 238 Type *RetTy = FTy->getReturnType();239 ffi_type *rtype = ffiTypeFor(RetTy);240 241 if (ffi_prep_cif(&cif, FFI_DEFAULT_ABI, NumArgs, rtype, args.data()) ==242 FFI_OK) {243 SmallVector<uint8_t, 128> ret;244 if (RetTy->getTypeID() != Type::VoidTyID)245 ret.resize(TD.getTypeStoreSize(RetTy));246 ffi_call(&cif, Fn, ret.data(), values.data());247 switch (RetTy->getTypeID()) {248 case Type::IntegerTyID:249 switch (cast<IntegerType>(RetTy)->getBitWidth()) {250 case 8: Result.IntVal = APInt(8 , *(int8_t *) ret.data()); break;251 case 16: Result.IntVal = APInt(16, *(int16_t*) ret.data()); break;252 case 32: Result.IntVal = APInt(32, *(int32_t*) ret.data()); break;253 case 64: Result.IntVal = APInt(64, *(int64_t*) ret.data()); break;254 }255 break;256 case Type::FloatTyID: Result.FloatVal = *(float *) ret.data(); break;257 case Type::DoubleTyID: Result.DoubleVal = *(double*) ret.data(); break;258 case Type::PointerTyID: Result.PointerVal = *(void **) ret.data(); break;259 default: break;260 }261 return true;262 }263 264 return false;265}266#endif // USE_LIBFFI267 268GenericValue Interpreter::callExternalFunction(Function *F,269 ArrayRef<GenericValue> ArgVals) {270 TheInterpreter = this;271 272 auto &Fns = getFunctions();273 std::unique_lock<sys::Mutex> Guard(Fns.Lock);274 275 // Do a lookup to see if the function is in our cache... this should just be a276 // deferred annotation!277 std::map<const Function *, ExFunc>::iterator FI =278 Fns.ExportedFunctions.find(F);279 if (ExFunc Fn = (FI == Fns.ExportedFunctions.end()) ? lookupFunction(F)280 : FI->second) {281 Guard.unlock();282 return Fn(F->getFunctionType(), ArgVals);283 }284 285#ifdef USE_LIBFFI286 std::map<const Function *, RawFunc>::iterator RF = Fns.RawFunctions.find(F);287 RawFunc RawFn;288 if (RF == Fns.RawFunctions.end()) {289 RawFn = (RawFunc)(intptr_t)290 sys::DynamicLibrary::SearchForAddressOfSymbol(std::string(F->getName()));291 if (!RawFn)292 RawFn = (RawFunc)(intptr_t)getPointerToGlobalIfAvailable(F);293 if (RawFn != 0)294 Fns.RawFunctions.insert(std::make_pair(F, RawFn)); // Cache for later295 } else {296 RawFn = RF->second;297 }298 299 Guard.unlock();300 301 GenericValue Result;302 if (RawFn != 0 && ffiInvoke(RawFn, F, ArgVals, getDataLayout(), Result))303 return Result;304#endif // USE_LIBFFI305 306 if (F->getName() == "__main")307 errs() << "Tried to execute an unknown external function: "308 << *F->getType() << " __main\n";309 else310 report_fatal_error("Tried to execute an unknown external function: " +311 F->getName());312#ifndef USE_LIBFFI313 errs() << "Recompiling LLVM with --enable-libffi might help.\n";314#endif315 return GenericValue();316}317 318//===----------------------------------------------------------------------===//319// Functions "exported" to the running application...320//321 322// void atexit(Function*)323static GenericValue lle_X_atexit(FunctionType *FT,324 ArrayRef<GenericValue> Args) {325 assert(Args.size() == 1);326 TheInterpreter->addAtExitHandler((Function*)GVTOP(Args[0]));327 GenericValue GV;328 GV.IntVal = 0;329 return GV;330}331 332// void exit(int)333static GenericValue lle_X_exit(FunctionType *FT, ArrayRef<GenericValue> Args) {334 TheInterpreter->exitCalled(Args[0]);335 return GenericValue();336}337 338// void abort(void)339static GenericValue lle_X_abort(FunctionType *FT, ArrayRef<GenericValue> Args) {340 //FIXME: should we report or raise here?341 //report_fatal_error("Interpreted program raised SIGABRT");342 raise (SIGABRT);343 return GenericValue();344}345 346// Silence warnings about sprintf. (See also347// https://github.com/llvm/llvm-project/issues/58086)348#if defined(__clang__)349#pragma clang diagnostic push350#pragma clang diagnostic ignored "-Wdeprecated-declarations"351#endif352// int sprintf(char *, const char *, ...) - a very rough implementation to make353// output useful.354static GenericValue lle_X_sprintf(FunctionType *FT,355 ArrayRef<GenericValue> Args) {356 char *OutputBuffer = (char *)GVTOP(Args[0]);357 const char *FmtStr = (const char *)GVTOP(Args[1]);358 unsigned ArgNo = 2;359 360 // printf should return # chars printed. This is completely incorrect, but361 // close enough for now.362 GenericValue GV;363 GV.IntVal = APInt(32, strlen(FmtStr));364 while (true) {365 switch (*FmtStr) {366 case 0: return GV; // Null terminator...367 default: // Normal nonspecial character368 sprintf(OutputBuffer++, "%c", *FmtStr++);369 break;370 case '\\': { // Handle escape codes371 sprintf(OutputBuffer, "%c%c", *FmtStr, *(FmtStr+1));372 FmtStr += 2; OutputBuffer += 2;373 break;374 }375 case '%': { // Handle format specifiers376 char FmtBuf[100] = "", Buffer[1000] = "";377 char *FB = FmtBuf;378 *FB++ = *FmtStr++;379 char Last = *FB++ = *FmtStr++;380 unsigned HowLong = 0;381 while (Last != 'c' && Last != 'd' && Last != 'i' && Last != 'u' &&382 Last != 'o' && Last != 'x' && Last != 'X' && Last != 'e' &&383 Last != 'E' && Last != 'g' && Last != 'G' && Last != 'f' &&384 Last != 'p' && Last != 's' && Last != '%') {385 if (Last == 'l' || Last == 'L') HowLong++; // Keep track of l's386 Last = *FB++ = *FmtStr++;387 }388 *FB = 0;389 390 switch (Last) {391 case '%':392 memcpy(Buffer, "%", 2); break;393 case 'c':394 sprintf(Buffer, FmtBuf, uint32_t(Args[ArgNo++].IntVal.getZExtValue()));395 break;396 case 'd': case 'i':397 case 'u': case 'o':398 case 'x': case 'X':399 if (HowLong >= 1) {400 if (HowLong == 1 &&401 TheInterpreter->getDataLayout().getPointerSizeInBits() == 64 &&402 sizeof(long) < sizeof(int64_t)) {403 // Make sure we use %lld with a 64 bit argument because we might be404 // compiling LLI on a 32 bit compiler.405 unsigned Size = strlen(FmtBuf);406 FmtBuf[Size] = FmtBuf[Size-1];407 FmtBuf[Size+1] = 0;408 FmtBuf[Size-1] = 'l';409 }410 sprintf(Buffer, FmtBuf, Args[ArgNo++].IntVal.getZExtValue());411 } else412 sprintf(Buffer, FmtBuf,uint32_t(Args[ArgNo++].IntVal.getZExtValue()));413 break;414 case 'e': case 'E': case 'g': case 'G': case 'f':415 sprintf(Buffer, FmtBuf, Args[ArgNo++].DoubleVal); break;416 case 'p':417 sprintf(Buffer, FmtBuf, (void*)GVTOP(Args[ArgNo++])); break;418 case 's':419 sprintf(Buffer, FmtBuf, (char*)GVTOP(Args[ArgNo++])); break;420 default:421 errs() << "<unknown printf code '" << *FmtStr << "'!>";422 ArgNo++; break;423 }424 size_t Len = strlen(Buffer);425 memcpy(OutputBuffer, Buffer, Len + 1);426 OutputBuffer += Len;427 }428 break;429 }430 }431 return GV;432}433#if defined(__clang__)434#pragma clang diagnostic pop435#endif436 437// int printf(const char *, ...) - a very rough implementation to make output438// useful.439static GenericValue lle_X_printf(FunctionType *FT,440 ArrayRef<GenericValue> Args) {441 char Buffer[10000];442 std::vector<GenericValue> NewArgs;443 NewArgs.push_back(PTOGV((void*)&Buffer[0]));444 llvm::append_range(NewArgs, Args);445 GenericValue GV = lle_X_sprintf(FT, NewArgs);446 outs() << Buffer;447 return GV;448}449 450// int sscanf(const char *format, ...);451static GenericValue lle_X_sscanf(FunctionType *FT,452 ArrayRef<GenericValue> args) {453 assert(args.size() < 10 && "Only handle up to 10 args to sscanf right now!");454 455 char *Args[10];456 for (unsigned i = 0; i < args.size(); ++i)457 Args[i] = (char*)GVTOP(args[i]);458 459 GenericValue GV;460 GV.IntVal = APInt(32, sscanf(Args[0], Args[1], Args[2], Args[3], Args[4],461 Args[5], Args[6], Args[7], Args[8], Args[9]));462 return GV;463}464 465// int scanf(const char *format, ...);466static GenericValue lle_X_scanf(FunctionType *FT, ArrayRef<GenericValue> args) {467 assert(args.size() < 10 && "Only handle up to 10 args to scanf right now!");468 469 char *Args[10];470 for (unsigned i = 0; i < args.size(); ++i)471 Args[i] = (char*)GVTOP(args[i]);472 473 GenericValue GV;474 GV.IntVal = APInt(32, scanf( Args[0], Args[1], Args[2], Args[3], Args[4],475 Args[5], Args[6], Args[7], Args[8], Args[9]));476 return GV;477}478 479// int fprintf(FILE *, const char *, ...) - a very rough implementation to make480// output useful.481static GenericValue lle_X_fprintf(FunctionType *FT,482 ArrayRef<GenericValue> Args) {483 assert(Args.size() >= 2);484 char Buffer[10000];485 std::vector<GenericValue> NewArgs;486 NewArgs.push_back(PTOGV(Buffer));487 llvm::append_range(NewArgs, llvm::drop_begin(Args));488 GenericValue GV = lle_X_sprintf(FT, NewArgs);489 490 fputs(Buffer, (FILE *) GVTOP(Args[0]));491 return GV;492}493 494static GenericValue lle_X_memset(FunctionType *FT,495 ArrayRef<GenericValue> Args) {496 int val = (int)Args[1].IntVal.getSExtValue();497 size_t len = (size_t)Args[2].IntVal.getZExtValue();498 memset((void *)GVTOP(Args[0]), val, len);499 // llvm.memset.* returns void, lle_X_* returns GenericValue,500 // so here we return GenericValue with IntVal set to zero501 GenericValue GV;502 GV.IntVal = 0;503 return GV;504}505 506static GenericValue lle_X_memcpy(FunctionType *FT,507 ArrayRef<GenericValue> Args) {508 memcpy(GVTOP(Args[0]), GVTOP(Args[1]),509 (size_t)(Args[2].IntVal.getLimitedValue()));510 511 // llvm.memcpy* returns void, lle_X_* returns GenericValue,512 // so here we return GenericValue with IntVal set to zero513 GenericValue GV;514 GV.IntVal = 0;515 return GV;516}517 518void Interpreter::initializeExternalFunctions() {519 auto &Fns = getFunctions();520 sys::ScopedLock Writer(Fns.Lock);521 Fns.FuncNames["lle_X_atexit"] = lle_X_atexit;522 Fns.FuncNames["lle_X_exit"] = lle_X_exit;523 Fns.FuncNames["lle_X_abort"] = lle_X_abort;524 525 Fns.FuncNames["lle_X_printf"] = lle_X_printf;526 Fns.FuncNames["lle_X_sprintf"] = lle_X_sprintf;527 Fns.FuncNames["lle_X_sscanf"] = lle_X_sscanf;528 Fns.FuncNames["lle_X_scanf"] = lle_X_scanf;529 Fns.FuncNames["lle_X_fprintf"] = lle_X_fprintf;530 Fns.FuncNames["lle_X_memset"] = lle_X_memset;531 Fns.FuncNames["lle_X_memcpy"] = lle_X_memcpy;532}533