1 //===-- llvm-mc.cpp - Machine Code Hacking Driver ---------------*- C++ -*-===// 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 utility is a simple driver that allows command line hacking on machine 10 // code. 11 // 12 //===----------------------------------------------------------------------===// 13 14 #include "Disassembler.h" 15 #include "llvm/MC/MCAsmBackend.h" 16 #include "llvm/MC/MCAsmInfo.h" 17 #include "llvm/MC/MCCodeEmitter.h" 18 #include "llvm/MC/MCContext.h" 19 #include "llvm/MC/MCInstPrinter.h" 20 #include "llvm/MC/MCInstrInfo.h" 21 #include "llvm/MC/MCObjectFileInfo.h" 22 #include "llvm/MC/MCObjectWriter.h" 23 #include "llvm/MC/MCParser/AsmLexer.h" 24 #include "llvm/MC/MCParser/MCTargetAsmParser.h" 25 #include "llvm/MC/MCRegisterInfo.h" 26 #include "llvm/MC/MCStreamer.h" 27 #include "llvm/MC/MCSubtargetInfo.h" 28 #include "llvm/MC/MCTargetOptionsCommandFlags.inc" 29 #include "llvm/Support/CommandLine.h" 30 #include "llvm/Support/Compression.h" 31 #include "llvm/Support/FileUtilities.h" 32 #include "llvm/Support/FormattedStream.h" 33 #include "llvm/Support/Host.h" 34 #include "llvm/Support/InitLLVM.h" 35 #include "llvm/Support/MemoryBuffer.h" 36 #include "llvm/Support/SourceMgr.h" 37 #include "llvm/Support/TargetRegistry.h" 38 #include "llvm/Support/TargetSelect.h" 39 #include "llvm/Support/ToolOutputFile.h" 40 #include "llvm/Support/WithColor.h" 41 42 using namespace llvm; 43 44 static cl::opt<std::string> 45 InputFilename(cl::Positional, cl::desc("<input file>"), cl::init("-")); 46 47 static cl::opt<std::string> OutputFilename("o", cl::desc("Output filename"), 48 cl::value_desc("filename"), 49 cl::init("-")); 50 51 static cl::opt<std::string> SplitDwarfFile("split-dwarf-file", 52 cl::desc("DWO output filename"), 53 cl::value_desc("filename")); 54 55 static cl::opt<bool> 56 ShowEncoding("show-encoding", cl::desc("Show instruction encodings")); 57 58 static cl::opt<bool> RelaxELFRel( 59 "relax-relocations", cl::init(true), 60 cl::desc("Emit R_X86_64_GOTPCRELX instead of R_X86_64_GOTPCREL")); 61 62 static cl::opt<DebugCompressionType> CompressDebugSections( 63 "compress-debug-sections", cl::ValueOptional, 64 cl::init(DebugCompressionType::None), 65 cl::desc("Choose DWARF debug sections compression:"), 66 cl::values(clEnumValN(DebugCompressionType::None, "none", "No compression"), 67 clEnumValN(DebugCompressionType::Z, "zlib", 68 "Use zlib compression"), 69 clEnumValN(DebugCompressionType::GNU, "zlib-gnu", 70 "Use zlib-gnu compression (deprecated)"))); 71 72 static cl::opt<bool> 73 ShowInst("show-inst", cl::desc("Show internal instruction representation")); 74 75 static cl::opt<bool> 76 ShowInstOperands("show-inst-operands", 77 cl::desc("Show instructions operands as parsed")); 78 79 static cl::opt<unsigned> 80 OutputAsmVariant("output-asm-variant", 81 cl::desc("Syntax variant to use for output printing")); 82 83 static cl::opt<bool> 84 PrintImmHex("print-imm-hex", cl::init(false), 85 cl::desc("Prefer hex format for immediate values")); 86 87 static cl::list<std::string> 88 DefineSymbol("defsym", cl::desc("Defines a symbol to be an integer constant")); 89 90 static cl::opt<bool> 91 PreserveComments("preserve-comments", 92 cl::desc("Preserve Comments in outputted assembly")); 93 94 enum OutputFileType { 95 OFT_Null, 96 OFT_AssemblyFile, 97 OFT_ObjectFile 98 }; 99 static cl::opt<OutputFileType> 100 FileType("filetype", cl::init(OFT_AssemblyFile), 101 cl::desc("Choose an output file type:"), 102 cl::values( 103 clEnumValN(OFT_AssemblyFile, "asm", 104 "Emit an assembly ('.s') file"), 105 clEnumValN(OFT_Null, "null", 106 "Don't emit anything (for timing purposes)"), 107 clEnumValN(OFT_ObjectFile, "obj", 108 "Emit a native object ('.o') file"))); 109 110 static cl::list<std::string> 111 IncludeDirs("I", cl::desc("Directory of include files"), 112 cl::value_desc("directory"), cl::Prefix); 113 114 static cl::opt<std::string> 115 ArchName("arch", cl::desc("Target arch to assemble for, " 116 "see -version for available targets")); 117 118 static cl::opt<std::string> 119 TripleName("triple", cl::desc("Target triple to assemble for, " 120 "see -version for available targets")); 121 122 static cl::opt<std::string> 123 MCPU("mcpu", 124 cl::desc("Target a specific cpu type (-mcpu=help for details)"), 125 cl::value_desc("cpu-name"), 126 cl::init("")); 127 128 static cl::list<std::string> 129 MAttrs("mattr", 130 cl::CommaSeparated, 131 cl::desc("Target specific attributes (-mattr=help for details)"), 132 cl::value_desc("a1,+a2,-a3,...")); 133 134 static cl::opt<bool> PIC("position-independent", 135 cl::desc("Position independent"), cl::init(false)); 136 137 static cl::opt<bool> 138 LargeCodeModel("large-code-model", 139 cl::desc("Create cfi directives that assume the code might " 140 "be more than 2gb away")); 141 142 static cl::opt<bool> 143 NoInitialTextSection("n", cl::desc("Don't assume assembly file starts " 144 "in the text section")); 145 146 static cl::opt<bool> 147 GenDwarfForAssembly("g", cl::desc("Generate dwarf debugging info for assembly " 148 "source files")); 149 150 static cl::opt<std::string> 151 DebugCompilationDir("fdebug-compilation-dir", 152 cl::desc("Specifies the debug info's compilation dir")); 153 154 static cl::list<std::string> 155 DebugPrefixMap("fdebug-prefix-map", 156 cl::desc("Map file source paths in debug info"), 157 cl::value_desc("= separated key-value pairs")); 158 159 static cl::opt<std::string> 160 MainFileName("main-file-name", 161 cl::desc("Specifies the name we should consider the input file")); 162 163 static cl::opt<bool> SaveTempLabels("save-temp-labels", 164 cl::desc("Don't discard temporary labels")); 165 166 static cl::opt<bool> LexMasmIntegers( 167 "masm-integers", 168 cl::desc("Enable binary and hex masm integers (0b110 and 0ABCh)")); 169 170 static cl::opt<bool> NoExecStack("no-exec-stack", 171 cl::desc("File doesn't need an exec stack")); 172 173 enum ActionType { 174 AC_AsLex, 175 AC_Assemble, 176 AC_Disassemble, 177 AC_MDisassemble, 178 }; 179 180 static cl::opt<ActionType> 181 Action(cl::desc("Action to perform:"), 182 cl::init(AC_Assemble), 183 cl::values(clEnumValN(AC_AsLex, "as-lex", 184 "Lex tokens from a .s file"), 185 clEnumValN(AC_Assemble, "assemble", 186 "Assemble a .s file (default)"), 187 clEnumValN(AC_Disassemble, "disassemble", 188 "Disassemble strings of hex bytes"), 189 clEnumValN(AC_MDisassemble, "mdis", 190 "Marked up disassembly of strings of hex bytes"))); 191 192 static const Target *GetTarget(const char *ProgName) { 193 // Figure out the target triple. 194 if (TripleName.empty()) 195 TripleName = sys::getDefaultTargetTriple(); 196 Triple TheTriple(Triple::normalize(TripleName)); 197 198 // Get the target specific parser. 199 std::string Error; 200 const Target *TheTarget = TargetRegistry::lookupTarget(ArchName, TheTriple, 201 Error); 202 if (!TheTarget) { 203 WithColor::error(errs(), ProgName) << Error; 204 return nullptr; 205 } 206 207 // Update the triple name and return the found target. 208 TripleName = TheTriple.getTriple(); 209 return TheTarget; 210 } 211 212 static std::unique_ptr<ToolOutputFile> GetOutputStream(StringRef Path, 213 sys::fs::OpenFlags Flags) { 214 std::error_code EC; 215 auto Out = std::make_unique<ToolOutputFile>(Path, EC, Flags); 216 if (EC) { 217 WithColor::error() << EC.message() << '\n'; 218 return nullptr; 219 } 220 221 return Out; 222 } 223 224 static std::string DwarfDebugFlags; 225 static void setDwarfDebugFlags(int argc, char **argv) { 226 if (!getenv("RC_DEBUG_OPTIONS")) 227 return; 228 for (int i = 0; i < argc; i++) { 229 DwarfDebugFlags += argv[i]; 230 if (i + 1 < argc) 231 DwarfDebugFlags += " "; 232 } 233 } 234 235 static std::string DwarfDebugProducer; 236 static void setDwarfDebugProducer() { 237 if(!getenv("DEBUG_PRODUCER")) 238 return; 239 DwarfDebugProducer += getenv("DEBUG_PRODUCER"); 240 } 241 242 static int AsLexInput(SourceMgr &SrcMgr, MCAsmInfo &MAI, 243 raw_ostream &OS) { 244 245 AsmLexer Lexer(MAI); 246 Lexer.setBuffer(SrcMgr.getMemoryBuffer(SrcMgr.getMainFileID())->getBuffer()); 247 248 bool Error = false; 249 while (Lexer.Lex().isNot(AsmToken::Eof)) { 250 Lexer.getTok().dump(OS); 251 OS << "\n"; 252 if (Lexer.getTok().getKind() == AsmToken::Error) 253 Error = true; 254 } 255 256 return Error; 257 } 258 259 static int fillCommandLineSymbols(MCAsmParser &Parser) { 260 for (auto &I: DefineSymbol) { 261 auto Pair = StringRef(I).split('='); 262 auto Sym = Pair.first; 263 auto Val = Pair.second; 264 265 if (Sym.empty() || Val.empty()) { 266 WithColor::error() << "defsym must be of the form: sym=value: " << I 267 << "\n"; 268 return 1; 269 } 270 int64_t Value; 271 if (Val.getAsInteger(0, Value)) { 272 WithColor::error() << "value is not an integer: " << Val << "\n"; 273 return 1; 274 } 275 Parser.getContext().setSymbolValue(Parser.getStreamer(), Sym, Value); 276 } 277 return 0; 278 } 279 280 static int AssembleInput(const char *ProgName, const Target *TheTarget, 281 SourceMgr &SrcMgr, MCContext &Ctx, MCStreamer &Str, 282 MCAsmInfo &MAI, MCSubtargetInfo &STI, 283 MCInstrInfo &MCII, MCTargetOptions const &MCOptions) { 284 std::unique_ptr<MCAsmParser> Parser( 285 createMCAsmParser(SrcMgr, Ctx, Str, MAI)); 286 std::unique_ptr<MCTargetAsmParser> TAP( 287 TheTarget->createMCAsmParser(STI, *Parser, MCII, MCOptions)); 288 289 if (!TAP) { 290 WithColor::error(errs(), ProgName) 291 << "this target does not support assembly parsing.\n"; 292 return 1; 293 } 294 295 int SymbolResult = fillCommandLineSymbols(*Parser); 296 if(SymbolResult) 297 return SymbolResult; 298 Parser->setShowParsedOperands(ShowInstOperands); 299 Parser->setTargetParser(*TAP); 300 Parser->getLexer().setLexMasmIntegers(LexMasmIntegers); 301 302 int Res = Parser->Run(NoInitialTextSection); 303 304 return Res; 305 } 306 307 int main(int argc, char **argv) { 308 InitLLVM X(argc, argv); 309 310 // Initialize targets and assembly printers/parsers. 311 llvm::InitializeAllTargetInfos(); 312 llvm::InitializeAllTargetMCs(); 313 llvm::InitializeAllAsmParsers(); 314 llvm::InitializeAllDisassemblers(); 315 316 // Register the target printer for --version. 317 cl::AddExtraVersionPrinter(TargetRegistry::printRegisteredTargetsForVersion); 318 319 cl::ParseCommandLineOptions(argc, argv, "llvm machine code playground\n"); 320 const MCTargetOptions MCOptions = InitMCTargetOptionsFromFlags(); 321 setDwarfDebugFlags(argc, argv); 322 323 setDwarfDebugProducer(); 324 325 const char *ProgName = argv[0]; 326 const Target *TheTarget = GetTarget(ProgName); 327 if (!TheTarget) 328 return 1; 329 // Now that GetTarget() has (potentially) replaced TripleName, it's safe to 330 // construct the Triple object. 331 Triple TheTriple(TripleName); 332 333 ErrorOr<std::unique_ptr<MemoryBuffer>> BufferPtr = 334 MemoryBuffer::getFileOrSTDIN(InputFilename); 335 if (std::error_code EC = BufferPtr.getError()) { 336 WithColor::error(errs(), ProgName) 337 << InputFilename << ": " << EC.message() << '\n'; 338 return 1; 339 } 340 MemoryBuffer *Buffer = BufferPtr->get(); 341 342 SourceMgr SrcMgr; 343 344 // Tell SrcMgr about this buffer, which is what the parser will pick up. 345 SrcMgr.AddNewSourceBuffer(std::move(*BufferPtr), SMLoc()); 346 347 // Record the location of the include directories so that the lexer can find 348 // it later. 349 SrcMgr.setIncludeDirs(IncludeDirs); 350 351 std::unique_ptr<MCRegisterInfo> MRI(TheTarget->createMCRegInfo(TripleName)); 352 assert(MRI && "Unable to create target register info!"); 353 354 std::unique_ptr<MCAsmInfo> MAI(TheTarget->createMCAsmInfo(*MRI, TripleName)); 355 assert(MAI && "Unable to create target asm info!"); 356 357 MAI->setRelaxELFRelocations(RelaxELFRel); 358 359 if (CompressDebugSections != DebugCompressionType::None) { 360 if (!zlib::isAvailable()) { 361 WithColor::error(errs(), ProgName) 362 << "build tools with zlib to enable -compress-debug-sections"; 363 return 1; 364 } 365 MAI->setCompressDebugSections(CompressDebugSections); 366 } 367 MAI->setPreserveAsmComments(PreserveComments); 368 369 // FIXME: This is not pretty. MCContext has a ptr to MCObjectFileInfo and 370 // MCObjectFileInfo needs a MCContext reference in order to initialize itself. 371 MCObjectFileInfo MOFI; 372 MCContext Ctx(MAI.get(), MRI.get(), &MOFI, &SrcMgr, &MCOptions); 373 MOFI.InitMCObjectFileInfo(TheTriple, PIC, Ctx, LargeCodeModel); 374 375 if (SaveTempLabels) 376 Ctx.setAllowTemporaryLabels(false); 377 378 Ctx.setGenDwarfForAssembly(GenDwarfForAssembly); 379 // Default to 4 for dwarf version. 380 unsigned DwarfVersion = MCOptions.DwarfVersion ? MCOptions.DwarfVersion : 4; 381 if (DwarfVersion < 2 || DwarfVersion > 5) { 382 errs() << ProgName << ": Dwarf version " << DwarfVersion 383 << " is not supported." << '\n'; 384 return 1; 385 } 386 Ctx.setDwarfVersion(DwarfVersion); 387 if (!DwarfDebugFlags.empty()) 388 Ctx.setDwarfDebugFlags(StringRef(DwarfDebugFlags)); 389 if (!DwarfDebugProducer.empty()) 390 Ctx.setDwarfDebugProducer(StringRef(DwarfDebugProducer)); 391 if (!DebugCompilationDir.empty()) 392 Ctx.setCompilationDir(DebugCompilationDir); 393 else { 394 // If no compilation dir is set, try to use the current directory. 395 SmallString<128> CWD; 396 if (!sys::fs::current_path(CWD)) 397 Ctx.setCompilationDir(CWD); 398 } 399 for (const auto &Arg : DebugPrefixMap) { 400 const auto &KV = StringRef(Arg).split('='); 401 Ctx.addDebugPrefixMapEntry(KV.first, KV.second); 402 } 403 if (!MainFileName.empty()) 404 Ctx.setMainFileName(MainFileName); 405 if (GenDwarfForAssembly) 406 Ctx.setGenDwarfRootFile(InputFilename, Buffer->getBuffer()); 407 408 // Package up features to be passed to target/subtarget 409 std::string FeaturesStr; 410 if (MAttrs.size()) { 411 SubtargetFeatures Features; 412 for (unsigned i = 0; i != MAttrs.size(); ++i) 413 Features.AddFeature(MAttrs[i]); 414 FeaturesStr = Features.getString(); 415 } 416 417 sys::fs::OpenFlags Flags = (FileType == OFT_AssemblyFile) ? sys::fs::OF_Text 418 : sys::fs::OF_None; 419 std::unique_ptr<ToolOutputFile> Out = GetOutputStream(OutputFilename, Flags); 420 if (!Out) 421 return 1; 422 423 std::unique_ptr<ToolOutputFile> DwoOut; 424 if (!SplitDwarfFile.empty()) { 425 if (FileType != OFT_ObjectFile) { 426 WithColor::error() << "dwo output only supported with object files\n"; 427 return 1; 428 } 429 DwoOut = GetOutputStream(SplitDwarfFile, sys::fs::OF_None); 430 if (!DwoOut) 431 return 1; 432 } 433 434 std::unique_ptr<buffer_ostream> BOS; 435 raw_pwrite_stream *OS = &Out->os(); 436 std::unique_ptr<MCStreamer> Str; 437 438 std::unique_ptr<MCInstrInfo> MCII(TheTarget->createMCInstrInfo()); 439 std::unique_ptr<MCSubtargetInfo> STI( 440 TheTarget->createMCSubtargetInfo(TripleName, MCPU, FeaturesStr)); 441 442 MCInstPrinter *IP = nullptr; 443 if (FileType == OFT_AssemblyFile) { 444 IP = TheTarget->createMCInstPrinter(Triple(TripleName), OutputAsmVariant, 445 *MAI, *MCII, *MRI); 446 447 if (!IP) { 448 WithColor::error() 449 << "unable to create instruction printer for target triple '" 450 << TheTriple.normalize() << "' with assembly variant " 451 << OutputAsmVariant << ".\n"; 452 return 1; 453 } 454 455 // Set the display preference for hex vs. decimal immediates. 456 IP->setPrintImmHex(PrintImmHex); 457 458 // Set up the AsmStreamer. 459 std::unique_ptr<MCCodeEmitter> CE; 460 if (ShowEncoding) 461 CE.reset(TheTarget->createMCCodeEmitter(*MCII, *MRI, Ctx)); 462 463 std::unique_ptr<MCAsmBackend> MAB( 464 TheTarget->createMCAsmBackend(*STI, *MRI, MCOptions)); 465 auto FOut = std::make_unique<formatted_raw_ostream>(*OS); 466 Str.reset( 467 TheTarget->createAsmStreamer(Ctx, std::move(FOut), /*asmverbose*/ true, 468 /*useDwarfDirectory*/ true, IP, 469 std::move(CE), std::move(MAB), ShowInst)); 470 471 } else if (FileType == OFT_Null) { 472 Str.reset(TheTarget->createNullStreamer(Ctx)); 473 } else { 474 assert(FileType == OFT_ObjectFile && "Invalid file type!"); 475 476 // Don't waste memory on names of temp labels. 477 Ctx.setUseNamesOnTempLabels(false); 478 479 if (!Out->os().supportsSeeking()) { 480 BOS = std::make_unique<buffer_ostream>(Out->os()); 481 OS = BOS.get(); 482 } 483 484 MCCodeEmitter *CE = TheTarget->createMCCodeEmitter(*MCII, *MRI, Ctx); 485 MCAsmBackend *MAB = TheTarget->createMCAsmBackend(*STI, *MRI, MCOptions); 486 Str.reset(TheTarget->createMCObjectStreamer( 487 TheTriple, Ctx, std::unique_ptr<MCAsmBackend>(MAB), 488 DwoOut ? MAB->createDwoObjectWriter(*OS, DwoOut->os()) 489 : MAB->createObjectWriter(*OS), 490 std::unique_ptr<MCCodeEmitter>(CE), *STI, MCOptions.MCRelaxAll, 491 MCOptions.MCIncrementalLinkerCompatible, 492 /*DWARFMustBeAtTheEnd*/ false)); 493 if (NoExecStack) 494 Str->InitSections(true); 495 } 496 497 // Use Assembler information for parsing. 498 Str->setUseAssemblerInfoForParsing(true); 499 500 int Res = 1; 501 bool disassemble = false; 502 switch (Action) { 503 case AC_AsLex: 504 Res = AsLexInput(SrcMgr, *MAI, Out->os()); 505 break; 506 case AC_Assemble: 507 Res = AssembleInput(ProgName, TheTarget, SrcMgr, Ctx, *Str, *MAI, *STI, 508 *MCII, MCOptions); 509 break; 510 case AC_MDisassemble: 511 assert(IP && "Expected assembly output"); 512 IP->setUseMarkup(true); 513 disassemble = true; 514 break; 515 case AC_Disassemble: 516 disassemble = true; 517 break; 518 } 519 if (disassemble) 520 Res = Disassembler::disassemble(*TheTarget, TripleName, *STI, *Str, *Buffer, 521 SrcMgr, Ctx, Out->os()); 522 523 // Keep output if no errors. 524 if (Res == 0) { 525 Out->keep(); 526 if (DwoOut) 527 DwoOut->keep(); 528 } 529 return Res; 530 } 531