tools/llvm-mc/llvm-mc.cpp

   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   std::error_code EC;
 214   auto Out = llvm::make_unique<ToolOutputFile>(Path, EC, sys::fs::F_None);
 215   if (EC) {
 216     WithColor::error() << EC.message() << '\n';
 217     return nullptr;
 218   }
 219
 220   return Out;
 221 }
 222
 223 static std::string DwarfDebugFlags;
 224 static void setDwarfDebugFlags(int argc, char **argv) {
 225   if (!getenv("RC_DEBUG_OPTIONS"))
 226     return;
 227   for (int i = 0; i < argc; i++) {
 228     DwarfDebugFlags += argv[i];
 229     if (i + 1 < argc)
 230       DwarfDebugFlags += " ";
 231   }
 232 }
 233
 234 static std::string DwarfDebugProducer;
 235 static void setDwarfDebugProducer() {
 236   if(!getenv("DEBUG_PRODUCER"))
 237     return;
 238   DwarfDebugProducer += getenv("DEBUG_PRODUCER");
 239 }
 240
 241 static int AsLexInput(SourceMgr &SrcMgr, MCAsmInfo &MAI,
 242                       raw_ostream &OS) {
 243
 244   AsmLexer Lexer(MAI);
 245   Lexer.setBuffer(SrcMgr.getMemoryBuffer(SrcMgr.getMainFileID())->getBuffer());
 246
 247   bool Error = false;
 248   while (Lexer.Lex().isNot(AsmToken::Eof)) {
 249     Lexer.getTok().dump(OS);
 250     OS << "\n";
 251     if (Lexer.getTok().getKind() == AsmToken::Error)
 252       Error = true;
 253   }
 254
 255   return Error;
 256 }
 257
 258 static int fillCommandLineSymbols(MCAsmParser &Parser) {
 259   for (auto &I: DefineSymbol) {
 260     auto Pair = StringRef(I).split('=');
 261     auto Sym = Pair.first;
 262     auto Val = Pair.second;
 263
 264     if (Sym.empty() || Val.empty()) {
 265       WithColor::error() << "defsym must be of the form: sym=value: " << I
 266                          << "\n";
 267       return 1;
 268     }
 269     int64_t Value;
 270     if (Val.getAsInteger(0, Value)) {
 271       WithColor::error() << "value is not an integer: " << Val << "\n";
 272       return 1;
 273     }
 274     Parser.getContext().setSymbolValue(Parser.getStreamer(), Sym, Value);
 275   }
 276   return 0;
 277 }
 278
 279 static int AssembleInput(const char *ProgName, const Target *TheTarget,
 280                          SourceMgr &SrcMgr, MCContext &Ctx, MCStreamer &Str,
 281                          MCAsmInfo &MAI, MCSubtargetInfo &STI,
 282                          MCInstrInfo &MCII, MCTargetOptions &MCOptions) {
 283   std::unique_ptr<MCAsmParser> Parser(
 284       createMCAsmParser(SrcMgr, Ctx, Str, MAI));
 285   std::unique_ptr<MCTargetAsmParser> TAP(
 286       TheTarget->createMCAsmParser(STI, *Parser, MCII, MCOptions));
 287
 288   if (!TAP) {
 289     WithColor::error(errs(), ProgName)
 290         << "this target does not support assembly parsing.\n";
 291     return 1;
 292   }
 293
 294   int SymbolResult = fillCommandLineSymbols(*Parser);
 295   if(SymbolResult)
 296     return SymbolResult;
 297   Parser->setShowParsedOperands(ShowInstOperands);
 298   Parser->setTargetParser(*TAP);
 299   Parser->getLexer().setLexMasmIntegers(LexMasmIntegers);
 300
 301   int Res = Parser->Run(NoInitialTextSection);
 302
 303   return Res;
 304 }
 305
 306 int main(int argc, char **argv) {
 307   InitLLVM X(argc, argv);
 308
 309   // Initialize targets and assembly printers/parsers.
 310   llvm::InitializeAllTargetInfos();
 311   llvm::InitializeAllTargetMCs();
 312   llvm::InitializeAllAsmParsers();
 313   llvm::InitializeAllDisassemblers();
 314
 315   // Register the target printer for --version.
 316   cl::AddExtraVersionPrinter(TargetRegistry::printRegisteredTargetsForVersion);
 317
 318   cl::ParseCommandLineOptions(argc, argv, "llvm machine code playground\n");
 319   MCTargetOptions MCOptions = InitMCTargetOptionsFromFlags();
 320   setDwarfDebugFlags(argc, argv);
 321
 322   setDwarfDebugProducer();
 323
 324   const char *ProgName = argv[0];
 325   const Target *TheTarget = GetTarget(ProgName);
 326   if (!TheTarget)
 327     return 1;
 328   // Now that GetTarget() has (potentially) replaced TripleName, it's safe to
 329   // construct the Triple object.
 330   Triple TheTriple(TripleName);
 331
 332   ErrorOr<std::unique_ptr<MemoryBuffer>> BufferPtr =
 333       MemoryBuffer::getFileOrSTDIN(InputFilename);
 334   if (std::error_code EC = BufferPtr.getError()) {
 335     WithColor::error(errs(), ProgName)
 336         << InputFilename << ": " << EC.message() << '\n';
 337     return 1;
 338   }
 339   MemoryBuffer *Buffer = BufferPtr->get();
 340
 341   SourceMgr SrcMgr;
 342
 343   // Tell SrcMgr about this buffer, which is what the parser will pick up.
 344   SrcMgr.AddNewSourceBuffer(std::move(*BufferPtr), SMLoc());
 345
 346   // Record the location of the include directories so that the lexer can find
 347   // it later.
 348   SrcMgr.setIncludeDirs(IncludeDirs);
 349
 350   std::unique_ptr<MCRegisterInfo> MRI(TheTarget->createMCRegInfo(TripleName));
 351   assert(MRI && "Unable to create target register info!");
 352
 353   std::unique_ptr<MCAsmInfo> MAI(TheTarget->createMCAsmInfo(*MRI, TripleName));
 354   assert(MAI && "Unable to create target asm info!");
 355
 356   MAI->setRelaxELFRelocations(RelaxELFRel);
 357
 358   if (CompressDebugSections != DebugCompressionType::None) {
 359     if (!zlib::isAvailable()) {
 360       WithColor::error(errs(), ProgName)
 361           << "build tools with zlib to enable -compress-debug-sections";
 362       return 1;
 363     }
 364     MAI->setCompressDebugSections(CompressDebugSections);
 365   }
 366   MAI->setPreserveAsmComments(PreserveComments);
 367
 368   // FIXME: This is not pretty. MCContext has a ptr to MCObjectFileInfo and
 369   // MCObjectFileInfo needs a MCContext reference in order to initialize itself.
 370   MCObjectFileInfo MOFI;
 371   MCContext Ctx(MAI.get(), MRI.get(), &MOFI, &SrcMgr);
 372   MOFI.InitMCObjectFileInfo(TheTriple, PIC, Ctx, LargeCodeModel);
 373
 374   if (SaveTempLabels)
 375     Ctx.setAllowTemporaryLabels(false);
 376
 377   Ctx.setGenDwarfForAssembly(GenDwarfForAssembly);
 378   // Default to 4 for dwarf version.
 379   unsigned DwarfVersion = MCOptions.DwarfVersion ? MCOptions.DwarfVersion : 4;
 380   if (DwarfVersion < 2 || DwarfVersion > 5) {
 381     errs() << ProgName << ": Dwarf version " << DwarfVersion
 382            << " is not supported." << '\n';
 383     return 1;
 384   }
 385   Ctx.setDwarfVersion(DwarfVersion);
 386   if (!DwarfDebugFlags.empty())
 387     Ctx.setDwarfDebugFlags(StringRef(DwarfDebugFlags));
 388   if (!DwarfDebugProducer.empty())
 389     Ctx.setDwarfDebugProducer(StringRef(DwarfDebugProducer));
 390   if (!DebugCompilationDir.empty())
 391     Ctx.setCompilationDir(DebugCompilationDir);
 392   else {
 393     // If no compilation dir is set, try to use the current directory.
 394     SmallString<128> CWD;
 395     if (!sys::fs::current_path(CWD))
 396       Ctx.setCompilationDir(CWD);
 397   }
 398   for (const auto &Arg : DebugPrefixMap) {
 399     const auto &KV = StringRef(Arg).split('=');
 400     Ctx.addDebugPrefixMapEntry(KV.first, KV.second);
 401   }
 402   if (!MainFileName.empty())
 403     Ctx.setMainFileName(MainFileName);
 404   if (GenDwarfForAssembly && DwarfVersion >= 5) {
 405     // DWARF v5 needs the root file as well as the compilation directory.
 406     // If we find a '.file 0' directive that will supersede these values.
 407     MD5 Hash;
 408     MD5::MD5Result *Cksum =
 409         (MD5::MD5Result *)Ctx.allocate(sizeof(MD5::MD5Result), 1);
 410     Hash.update(Buffer->getBuffer());
 411     Hash.final(*Cksum);
 412     Ctx.setMCLineTableRootFile(
 413         /*CUID=*/0, Ctx.getCompilationDir(),
 414         !MainFileName.empty() ? MainFileName : InputFilename, Cksum, None);
 415   }
 416
 417   // Package up features to be passed to target/subtarget
 418   std::string FeaturesStr;
 419   if (MAttrs.size()) {
 420     SubtargetFeatures Features;
 421     for (unsigned i = 0; i != MAttrs.size(); ++i)
 422       Features.AddFeature(MAttrs[i]);
 423     FeaturesStr = Features.getString();
 424   }
 425
 426   std::unique_ptr<ToolOutputFile> Out = GetOutputStream(OutputFilename);
 427   if (!Out)
 428     return 1;
 429
 430   std::unique_ptr<ToolOutputFile> DwoOut;
 431   if (!SplitDwarfFile.empty()) {
 432     if (FileType != OFT_ObjectFile) {
 433       WithColor::error() << "dwo output only supported with object files\n";
 434       return 1;
 435     }
 436     DwoOut = GetOutputStream(SplitDwarfFile);
 437     if (!DwoOut)
 438       return 1;
 439   }
 440
 441   std::unique_ptr<buffer_ostream> BOS;
 442   raw_pwrite_stream *OS = &Out->os();
 443   std::unique_ptr<MCStreamer> Str;
 444
 445   std::unique_ptr<MCInstrInfo> MCII(TheTarget->createMCInstrInfo());
 446   std::unique_ptr<MCSubtargetInfo> STI(
 447       TheTarget->createMCSubtargetInfo(TripleName, MCPU, FeaturesStr));
 448
 449   MCInstPrinter *IP = nullptr;
 450   if (FileType == OFT_AssemblyFile) {
 451     IP = TheTarget->createMCInstPrinter(Triple(TripleName), OutputAsmVariant,
 452                                         *MAI, *MCII, *MRI);
 453
 454     if (!IP) {
 455       WithColor::error()
 456           << "unable to create instruction printer for target triple '"
 457           << TheTriple.normalize() << "' with assembly variant "
 458           << OutputAsmVariant << ".\n";
 459       return 1;
 460     }
 461
 462     // Set the display preference for hex vs. decimal immediates.
 463     IP->setPrintImmHex(PrintImmHex);
 464
 465     // Set up the AsmStreamer.
 466     std::unique_ptr<MCCodeEmitter> CE;
 467     if (ShowEncoding)
 468       CE.reset(TheTarget->createMCCodeEmitter(*MCII, *MRI, Ctx));
 469
 470     std::unique_ptr<MCAsmBackend> MAB(
 471         TheTarget->createMCAsmBackend(*STI, *MRI, MCOptions));
 472     auto FOut = llvm::make_unique<formatted_raw_ostream>(*OS);
 473     Str.reset(
 474         TheTarget->createAsmStreamer(Ctx, std::move(FOut), /*asmverbose*/ true,
 475                                      /*useDwarfDirectory*/ true, IP,
 476                                      std::move(CE), std::move(MAB), ShowInst));
 477
 478   } else if (FileType == OFT_Null) {
 479     Str.reset(TheTarget->createNullStreamer(Ctx));
 480   } else {
 481     assert(FileType == OFT_ObjectFile && "Invalid file type!");
 482
 483     // Don't waste memory on names of temp labels.
 484     Ctx.setUseNamesOnTempLabels(false);
 485
 486     if (!Out->os().supportsSeeking()) {
 487       BOS = make_unique<buffer_ostream>(Out->os());
 488       OS = BOS.get();
 489     }
 490
 491     MCCodeEmitter *CE = TheTarget->createMCCodeEmitter(*MCII, *MRI, Ctx);
 492     MCAsmBackend *MAB = TheTarget->createMCAsmBackend(*STI, *MRI, MCOptions);
 493     Str.reset(TheTarget->createMCObjectStreamer(
 494         TheTriple, Ctx, std::unique_ptr<MCAsmBackend>(MAB),
 495         DwoOut ? MAB->createDwoObjectWriter(*OS, DwoOut->os())
 496                : MAB->createObjectWriter(*OS),
 497         std::unique_ptr<MCCodeEmitter>(CE), *STI, MCOptions.MCRelaxAll,
 498         MCOptions.MCIncrementalLinkerCompatible,
 499         /*DWARFMustBeAtTheEnd*/ false));
 500     if (NoExecStack)
 501       Str->InitSections(true);
 502   }
 503
 504   // Use Assembler information for parsing.
 505   Str->setUseAssemblerInfoForParsing(true);
 506
 507   int Res = 1;
 508   bool disassemble = false;
 509   switch (Action) {
 510   case AC_AsLex:
 511     Res = AsLexInput(SrcMgr, *MAI, Out->os());
 512     break;
 513   case AC_Assemble:
 514     Res = AssembleInput(ProgName, TheTarget, SrcMgr, Ctx, *Str, *MAI, *STI,
 515                         *MCII, MCOptions);
 516     break;
 517   case AC_MDisassemble:
 518     assert(IP && "Expected assembly output");
 519     IP->setUseMarkup(1);
 520     disassemble = true;
 521     break;
 522   case AC_Disassemble:
 523     disassemble = true;
 524     break;
 525   }
 526   if (disassemble)
 527     Res = Disassembler::disassemble(*TheTarget, TripleName, *STI, *Str,
 528                                     *Buffer, SrcMgr, Out->os());
 529
 530   // Keep output if no errors.
 531   if (Res == 0) {
 532     Out->keep();
 533     if (DwoOut)
 534       DwoOut->keep();
 535   }
 536   return Res;
 537 }