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.h"
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
42using namespace llvm;
43
44static mc::RegisterMCTargetOptionsFlags MOF;
45
46static cl::opt<std::string>
47InputFilename(cl::Positional, cl::desc("<input file>"), cl::init("-"));
48
49static cl::opt<std::string> OutputFilename("o", cl::desc("Output filename"),
50                                           cl::value_desc("filename"),
51                                           cl::init("-"));
52
53static cl::opt<std::string> SplitDwarfFile("split-dwarf-file",
54                                           cl::desc("DWO output filename"),
55                                           cl::value_desc("filename"));
56
57static cl::opt<bool>
58ShowEncoding("show-encoding", cl::desc("Show instruction encodings"));
59
60static cl::opt<bool> RelaxELFRel(
61    "relax-relocations", cl::init(true),
62    cl::desc("Emit R_X86_64_GOTPCRELX instead of R_X86_64_GOTPCREL"));
63
64static cl::opt<DebugCompressionType> CompressDebugSections(
65    "compress-debug-sections", cl::ValueOptional,
66    cl::init(DebugCompressionType::None),
67    cl::desc("Choose DWARF debug sections compression:"),
68    cl::values(clEnumValN(DebugCompressionType::None, "none", "No compression"),
69               clEnumValN(DebugCompressionType::Z, "zlib",
70                          "Use zlib compression"),
71               clEnumValN(DebugCompressionType::GNU, "zlib-gnu",
72                          "Use zlib-gnu compression (deprecated)")));
73
74static cl::opt<bool>
75ShowInst("show-inst", cl::desc("Show internal instruction representation"));
76
77static cl::opt<bool>
78ShowInstOperands("show-inst-operands",
79                 cl::desc("Show instructions operands as parsed"));
80
81static cl::opt<unsigned>
82OutputAsmVariant("output-asm-variant",
83                 cl::desc("Syntax variant to use for output printing"));
84
85static cl::opt<bool>
86PrintImmHex("print-imm-hex", cl::init(false),
87            cl::desc("Prefer hex format for immediate values"));
88
89static cl::list<std::string>
90DefineSymbol("defsym", cl::desc("Defines a symbol to be an integer constant"));
91
92static cl::opt<bool>
93    PreserveComments("preserve-comments",
94                     cl::desc("Preserve Comments in outputted assembly"));
95
96enum OutputFileType {
97  OFT_Null,
98  OFT_AssemblyFile,
99  OFT_ObjectFile
100};
101static cl::opt<OutputFileType>
102FileType("filetype", cl::init(OFT_AssemblyFile),
103  cl::desc("Choose an output file type:"),
104  cl::values(
105       clEnumValN(OFT_AssemblyFile, "asm",
106                  "Emit an assembly ('.s') file"),
107       clEnumValN(OFT_Null, "null",
108                  "Don't emit anything (for timing purposes)"),
109       clEnumValN(OFT_ObjectFile, "obj",
110                  "Emit a native object ('.o') file")));
111
112static cl::list<std::string>
113IncludeDirs("I", cl::desc("Directory of include files"),
114            cl::value_desc("directory"), cl::Prefix);
115
116static cl::opt<std::string>
117ArchName("arch", cl::desc("Target arch to assemble for, "
118                          "see -version for available targets"));
119
120static cl::opt<std::string>
121TripleName("triple", cl::desc("Target triple to assemble for, "
122                              "see -version for available targets"));
123
124static cl::opt<std::string>
125MCPU("mcpu",
126     cl::desc("Target a specific cpu type (-mcpu=help for details)"),
127     cl::value_desc("cpu-name"),
128     cl::init(""));
129
130static cl::list<std::string>
131MAttrs("mattr",
132  cl::CommaSeparated,
133  cl::desc("Target specific attributes (-mattr=help for details)"),
134  cl::value_desc("a1,+a2,-a3,..."));
135
136static cl::opt<bool> PIC("position-independent",
137                         cl::desc("Position independent"), cl::init(false));
138
139static cl::opt<bool>
140    LargeCodeModel("large-code-model",
141                   cl::desc("Create cfi directives that assume the code might "
142                            "be more than 2gb away"));
143
144static cl::opt<bool>
145NoInitialTextSection("n", cl::desc("Don't assume assembly file starts "
146                                   "in the text section"));
147
148static cl::opt<bool>
149GenDwarfForAssembly("g", cl::desc("Generate dwarf debugging info for assembly "
150                                  "source files"));
151
152static cl::opt<std::string>
153DebugCompilationDir("fdebug-compilation-dir",
154                    cl::desc("Specifies the debug info's compilation dir"));
155
156static cl::list<std::string>
157DebugPrefixMap("fdebug-prefix-map",
158               cl::desc("Map file source paths in debug info"),
159               cl::value_desc("= separated key-value pairs"));
160
161static cl::opt<std::string>
162MainFileName("main-file-name",
163             cl::desc("Specifies the name we should consider the input file"));
164
165static cl::opt<bool> SaveTempLabels("save-temp-labels",
166                                    cl::desc("Don't discard temporary labels"));
167
168static cl::opt<bool> LexMasmIntegers(
169    "masm-integers",
170    cl::desc("Enable binary and hex masm integers (0b110 and 0ABCh)"));
171
172static cl::opt<bool> NoExecStack("no-exec-stack",
173                                 cl::desc("File doesn't need an exec stack"));
174
175enum ActionType {
176  AC_AsLex,
177  AC_Assemble,
178  AC_Disassemble,
179  AC_MDisassemble,
180};
181
182static cl::opt<ActionType>
183Action(cl::desc("Action to perform:"),
184       cl::init(AC_Assemble),
185       cl::values(clEnumValN(AC_AsLex, "as-lex",
186                             "Lex tokens from a .s file"),
187                  clEnumValN(AC_Assemble, "assemble",
188                             "Assemble a .s file (default)"),
189                  clEnumValN(AC_Disassemble, "disassemble",
190                             "Disassemble strings of hex bytes"),
191                  clEnumValN(AC_MDisassemble, "mdis",
192                             "Marked up disassembly of strings of hex bytes")));
193
194static const Target *GetTarget(const char *ProgName) {
195  // Figure out the target triple.
196  if (TripleName.empty())
197    TripleName = sys::getDefaultTargetTriple();
198  Triple TheTriple(Triple::normalize(TripleName));
199
200  // Get the target specific parser.
201  std::string Error;
202  const Target *TheTarget = TargetRegistry::lookupTarget(ArchName, TheTriple,
203                                                         Error);
204  if (!TheTarget) {
205    WithColor::error(errs(), ProgName) << Error;
206    return nullptr;
207  }
208
209  // Update the triple name and return the found target.
210  TripleName = TheTriple.getTriple();
211  return TheTarget;
212}
213
214static std::unique_ptr<ToolOutputFile> GetOutputStream(StringRef Path,
215    sys::fs::OpenFlags Flags) {
216  std::error_code EC;
217  auto Out = std::make_unique<ToolOutputFile>(Path, EC, Flags);
218  if (EC) {
219    WithColor::error() << EC.message() << '\n';
220    return nullptr;
221  }
222
223  return Out;
224}
225
226static std::string DwarfDebugFlags;
227static void setDwarfDebugFlags(int argc, char **argv) {
228  if (!getenv("RC_DEBUG_OPTIONS"))
229    return;
230  for (int i = 0; i < argc; i++) {
231    DwarfDebugFlags += argv[i];
232    if (i + 1 < argc)
233      DwarfDebugFlags += " ";
234  }
235}
236
237static std::string DwarfDebugProducer;
238static void setDwarfDebugProducer() {
239  if(!getenv("DEBUG_PRODUCER"))
240    return;
241  DwarfDebugProducer += getenv("DEBUG_PRODUCER");
242}
243
244static int AsLexInput(SourceMgr &SrcMgr, MCAsmInfo &MAI,
245                      raw_ostream &OS) {
246
247  AsmLexer Lexer(MAI);
248  Lexer.setBuffer(SrcMgr.getMemoryBuffer(SrcMgr.getMainFileID())->getBuffer());
249
250  bool Error = false;
251  while (Lexer.Lex().isNot(AsmToken::Eof)) {
252    Lexer.getTok().dump(OS);
253    OS << "\n";
254    if (Lexer.getTok().getKind() == AsmToken::Error)
255      Error = true;
256  }
257
258  return Error;
259}
260
261static int fillCommandLineSymbols(MCAsmParser &Parser) {
262  for (auto &I: DefineSymbol) {
263    auto Pair = StringRef(I).split('=');
264    auto Sym = Pair.first;
265    auto Val = Pair.second;
266
267    if (Sym.empty() || Val.empty()) {
268      WithColor::error() << "defsym must be of the form: sym=value: " << I
269                         << "\n";
270      return 1;
271    }
272    int64_t Value;
273    if (Val.getAsInteger(0, Value)) {
274      WithColor::error() << "value is not an integer: " << Val << "\n";
275      return 1;
276    }
277    Parser.getContext().setSymbolValue(Parser.getStreamer(), Sym, Value);
278  }
279  return 0;
280}
281
282static int AssembleInput(const char *ProgName, const Target *TheTarget,
283                         SourceMgr &SrcMgr, MCContext &Ctx, MCStreamer &Str,
284                         MCAsmInfo &MAI, MCSubtargetInfo &STI,
285                         MCInstrInfo &MCII, MCTargetOptions const &MCOptions) {
286  std::unique_ptr<MCAsmParser> Parser(
287      createMCAsmParser(SrcMgr, Ctx, Str, MAI));
288  std::unique_ptr<MCTargetAsmParser> TAP(
289      TheTarget->createMCAsmParser(STI, *Parser, MCII, MCOptions));
290
291  if (!TAP) {
292    WithColor::error(errs(), ProgName)
293        << "this target does not support assembly parsing.\n";
294    return 1;
295  }
296
297  int SymbolResult = fillCommandLineSymbols(*Parser);
298  if(SymbolResult)
299    return SymbolResult;
300  Parser->setShowParsedOperands(ShowInstOperands);
301  Parser->setTargetParser(*TAP);
302  Parser->getLexer().setLexMasmIntegers(LexMasmIntegers);
303
304  int Res = Parser->Run(NoInitialTextSection);
305
306  return Res;
307}
308
309int main(int argc, char **argv) {
310  InitLLVM X(argc, argv);
311
312  // Initialize targets and assembly printers/parsers.
313  llvm::InitializeAllTargetInfos();
314  llvm::InitializeAllTargetMCs();
315  llvm::InitializeAllAsmParsers();
316  llvm::InitializeAllDisassemblers();
317
318  // Register the target printer for --version.
319  cl::AddExtraVersionPrinter(TargetRegistry::printRegisteredTargetsForVersion);
320
321  cl::ParseCommandLineOptions(argc, argv, "llvm machine code playground\n");
322  const MCTargetOptions MCOptions = mc::InitMCTargetOptionsFromFlags();
323  setDwarfDebugFlags(argc, argv);
324
325  setDwarfDebugProducer();
326
327  const char *ProgName = argv[0];
328  const Target *TheTarget = GetTarget(ProgName);
329  if (!TheTarget)
330    return 1;
331  // Now that GetTarget() has (potentially) replaced TripleName, it's safe to
332  // construct the Triple object.
333  Triple TheTriple(TripleName);
334
335  ErrorOr<std::unique_ptr<MemoryBuffer>> BufferPtr =
336      MemoryBuffer::getFileOrSTDIN(InputFilename);
337  if (std::error_code EC = BufferPtr.getError()) {
338    WithColor::error(errs(), ProgName)
339        << InputFilename << ": " << EC.message() << '\n';
340    return 1;
341  }
342  MemoryBuffer *Buffer = BufferPtr->get();
343
344  SourceMgr SrcMgr;
345
346  // Tell SrcMgr about this buffer, which is what the parser will pick up.
347  SrcMgr.AddNewSourceBuffer(std::move(*BufferPtr), SMLoc());
348
349  // Record the location of the include directories so that the lexer can find
350  // it later.
351  SrcMgr.setIncludeDirs(IncludeDirs);
352
353  std::unique_ptr<MCRegisterInfo> MRI(TheTarget->createMCRegInfo(TripleName));
354  assert(MRI && "Unable to create target register info!");
355
356  std::unique_ptr<MCAsmInfo> MAI(
357      TheTarget->createMCAsmInfo(*MRI, TripleName, MCOptions));
358  assert(MAI && "Unable to create target asm info!");
359
360  MAI->setRelaxELFRelocations(RelaxELFRel);
361
362  if (CompressDebugSections != DebugCompressionType::None) {
363    if (!zlib::isAvailable()) {
364      WithColor::error(errs(), ProgName)
365          << "build tools with zlib to enable -compress-debug-sections";
366      return 1;
367    }
368    MAI->setCompressDebugSections(CompressDebugSections);
369  }
370  MAI->setPreserveAsmComments(PreserveComments);
371
372  // FIXME: This is not pretty. MCContext has a ptr to MCObjectFileInfo and
373  // MCObjectFileInfo needs a MCContext reference in order to initialize itself.
374  MCObjectFileInfo MOFI;
375  MCContext Ctx(MAI.get(), MRI.get(), &MOFI, &SrcMgr, &MCOptions);
376  MOFI.InitMCObjectFileInfo(TheTriple, PIC, Ctx, LargeCodeModel);
377
378  if (SaveTempLabels)
379    Ctx.setAllowTemporaryLabels(false);
380
381  Ctx.setGenDwarfForAssembly(GenDwarfForAssembly);
382  // Default to 4 for dwarf version.
383  unsigned DwarfVersion = MCOptions.DwarfVersion ? MCOptions.DwarfVersion : 4;
384  if (DwarfVersion < 2 || DwarfVersion > 5) {
385    errs() << ProgName << ": Dwarf version " << DwarfVersion
386           << " is not supported." << '\n';
387    return 1;
388  }
389  Ctx.setDwarfVersion(DwarfVersion);
390  if (MCOptions.Dwarf64) {
391    // The 64-bit DWARF format was introduced in DWARFv3.
392    if (DwarfVersion < 3) {
393      errs() << ProgName
394             << ": the 64-bit DWARF format is not supported for DWARF versions "
395                "prior to 3\n";
396      return 1;
397    }
398    // 32-bit targets don't support DWARF64, which requires 64-bit relocations.
399    if (MAI->getCodePointerSize() < 8) {
400      errs() << ProgName
401             << ": the 64-bit DWARF format is only supported for 64-bit "
402                "targets\n";
403      return 1;
404    }
405    // If needsDwarfSectionOffsetDirective is true, we would eventually call
406    // MCStreamer::emitSymbolValue() with IsSectionRelative = true, but that
407    // is supported only for 4-byte long references.
408    if (MAI->needsDwarfSectionOffsetDirective()) {
409      errs() << ProgName << ": the 64-bit DWARF format is not supported for "
410             << TheTriple.normalize() << "\n";
411      return 1;
412    }
413    Ctx.setDwarfFormat(dwarf::DWARF64);
414  }
415  if (!DwarfDebugFlags.empty())
416    Ctx.setDwarfDebugFlags(StringRef(DwarfDebugFlags));
417  if (!DwarfDebugProducer.empty())
418    Ctx.setDwarfDebugProducer(StringRef(DwarfDebugProducer));
419  if (!DebugCompilationDir.empty())
420    Ctx.setCompilationDir(DebugCompilationDir);
421  else {
422    // If no compilation dir is set, try to use the current directory.
423    SmallString<128> CWD;
424    if (!sys::fs::current_path(CWD))
425      Ctx.setCompilationDir(CWD);
426  }
427  for (const auto &Arg : DebugPrefixMap) {
428    const auto &KV = StringRef(Arg).split('=');
429    Ctx.addDebugPrefixMapEntry(std::string(KV.first), std::string(KV.second));
430  }
431  if (!MainFileName.empty())
432    Ctx.setMainFileName(MainFileName);
433  if (GenDwarfForAssembly)
434    Ctx.setGenDwarfRootFile(InputFilename, Buffer->getBuffer());
435
436  // Package up features to be passed to target/subtarget
437  std::string FeaturesStr;
438  if (MAttrs.size()) {
439    SubtargetFeatures Features;
440    for (unsigned i = 0; i != MAttrs.size(); ++i)
441      Features.AddFeature(MAttrs[i]);
442    FeaturesStr = Features.getString();
443  }
444
445  sys::fs::OpenFlags Flags = (FileType == OFT_AssemblyFile) ? sys::fs::OF_Text
446                                                            : sys::fs::OF_None;
447  std::unique_ptr<ToolOutputFile> Out = GetOutputStream(OutputFilename, Flags);
448  if (!Out)
449    return 1;
450
451  std::unique_ptr<ToolOutputFile> DwoOut;
452  if (!SplitDwarfFile.empty()) {
453    if (FileType != OFT_ObjectFile) {
454      WithColor::error() << "dwo output only supported with object files\n";
455      return 1;
456    }
457    DwoOut = GetOutputStream(SplitDwarfFile, sys::fs::OF_None);
458    if (!DwoOut)
459      return 1;
460  }
461
462  std::unique_ptr<buffer_ostream> BOS;
463  raw_pwrite_stream *OS = &Out->os();
464  std::unique_ptr<MCStreamer> Str;
465
466  std::unique_ptr<MCInstrInfo> MCII(TheTarget->createMCInstrInfo());
467  std::unique_ptr<MCSubtargetInfo> STI(
468      TheTarget->createMCSubtargetInfo(TripleName, MCPU, FeaturesStr));
469
470  MCInstPrinter *IP = nullptr;
471  if (FileType == OFT_AssemblyFile) {
472    IP = TheTarget->createMCInstPrinter(Triple(TripleName), OutputAsmVariant,
473                                        *MAI, *MCII, *MRI);
474
475    if (!IP) {
476      WithColor::error()
477          << "unable to create instruction printer for target triple '"
478          << TheTriple.normalize() << "' with assembly variant "
479          << OutputAsmVariant << ".\n";
480      return 1;
481    }
482
483    // Set the display preference for hex vs. decimal immediates.
484    IP->setPrintImmHex(PrintImmHex);
485
486    // Set up the AsmStreamer.
487    std::unique_ptr<MCCodeEmitter> CE;
488    if (ShowEncoding)
489      CE.reset(TheTarget->createMCCodeEmitter(*MCII, *MRI, Ctx));
490
491    std::unique_ptr<MCAsmBackend> MAB(
492        TheTarget->createMCAsmBackend(*STI, *MRI, MCOptions));
493    auto FOut = std::make_unique<formatted_raw_ostream>(*OS);
494    Str.reset(
495        TheTarget->createAsmStreamer(Ctx, std::move(FOut), /*asmverbose*/ true,
496                                     /*useDwarfDirectory*/ true, IP,
497                                     std::move(CE), std::move(MAB), ShowInst));
498
499  } else if (FileType == OFT_Null) {
500    Str.reset(TheTarget->createNullStreamer(Ctx));
501  } else {
502    assert(FileType == OFT_ObjectFile && "Invalid file type!");
503
504    if (!Out->os().supportsSeeking()) {
505      BOS = std::make_unique<buffer_ostream>(Out->os());
506      OS = BOS.get();
507    }
508
509    MCCodeEmitter *CE = TheTarget->createMCCodeEmitter(*MCII, *MRI, Ctx);
510    MCAsmBackend *MAB = TheTarget->createMCAsmBackend(*STI, *MRI, MCOptions);
511    Str.reset(TheTarget->createMCObjectStreamer(
512        TheTriple, Ctx, std::unique_ptr<MCAsmBackend>(MAB),
513        DwoOut ? MAB->createDwoObjectWriter(*OS, DwoOut->os())
514               : MAB->createObjectWriter(*OS),
515        std::unique_ptr<MCCodeEmitter>(CE), *STI, MCOptions.MCRelaxAll,
516        MCOptions.MCIncrementalLinkerCompatible,
517        /*DWARFMustBeAtTheEnd*/ false));
518    if (NoExecStack)
519      Str->InitSections(true);
520  }
521
522  // Use Assembler information for parsing.
523  Str->setUseAssemblerInfoForParsing(true);
524
525  int Res = 1;
526  bool disassemble = false;
527  switch (Action) {
528  case AC_AsLex:
529    Res = AsLexInput(SrcMgr, *MAI, Out->os());
530    break;
531  case AC_Assemble:
532    Res = AssembleInput(ProgName, TheTarget, SrcMgr, Ctx, *Str, *MAI, *STI,
533                        *MCII, MCOptions);
534    break;
535  case AC_MDisassemble:
536    assert(IP && "Expected assembly output");
537    IP->setUseMarkup(true);
538    disassemble = true;
539    break;
540  case AC_Disassemble:
541    disassemble = true;
542    break;
543  }
544  if (disassemble)
545    Res = Disassembler::disassemble(*TheTarget, TripleName, *STI, *Str, *Buffer,
546                                    SrcMgr, Ctx, Out->os(), MCOptions);
547
548  // Keep output if no errors.
549  if (Res == 0) {
550    Out->keep();
551    if (DwoOut)
552      DwoOut->keep();
553  }
554  return Res;
555}
556