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 
42 using namespace llvm;
43 
44 static mc::RegisterMCTargetOptionsFlags MOF;
45 
46 static cl::opt<std::string>
47 InputFilename(cl::Positional, cl::desc("<input file>"), cl::init("-"));
48 
49 static cl::opt<std::string> OutputFilename("o", cl::desc("Output filename"),
50                                            cl::value_desc("filename"),
51                                            cl::init("-"));
52 
53 static cl::opt<std::string> SplitDwarfFile("split-dwarf-file",
54                                            cl::desc("DWO output filename"),
55                                            cl::value_desc("filename"));
56 
57 static cl::opt<bool>
58 ShowEncoding("show-encoding", cl::desc("Show instruction encodings"));
59 
60 static 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 
64 static 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 
74 static cl::opt<bool>
75 ShowInst("show-inst", cl::desc("Show internal instruction representation"));
76 
77 static cl::opt<bool>
78 ShowInstOperands("show-inst-operands",
79                  cl::desc("Show instructions operands as parsed"));
80 
81 static cl::opt<unsigned>
82 OutputAsmVariant("output-asm-variant",
83                  cl::desc("Syntax variant to use for output printing"));
84 
85 static cl::opt<bool>
86 PrintImmHex("print-imm-hex", cl::init(false),
87             cl::desc("Prefer hex format for immediate values"));
88 
89 static cl::list<std::string>
90 DefineSymbol("defsym", cl::desc("Defines a symbol to be an integer constant"));
91 
92 static cl::opt<bool>
93     PreserveComments("preserve-comments",
94                      cl::desc("Preserve Comments in outputted assembly"));
95 
96 enum OutputFileType {
97   OFT_Null,
98   OFT_AssemblyFile,
99   OFT_ObjectFile
100 };
101 static cl::opt<OutputFileType>
102 FileType("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 
112 static cl::list<std::string>
113 IncludeDirs("I", cl::desc("Directory of include files"),
114             cl::value_desc("directory"), cl::Prefix);
115 
116 static cl::opt<std::string>
117 ArchName("arch", cl::desc("Target arch to assemble for, "
118                           "see -version for available targets"));
119 
120 static cl::opt<std::string>
121 TripleName("triple", cl::desc("Target triple to assemble for, "
122                               "see -version for available targets"));
123 
124 static cl::opt<std::string>
125 MCPU("mcpu",
126      cl::desc("Target a specific cpu type (-mcpu=help for details)"),
127      cl::value_desc("cpu-name"),
128      cl::init(""));
129 
130 static cl::list<std::string>
131 MAttrs("mattr",
132   cl::CommaSeparated,
133   cl::desc("Target specific attributes (-mattr=help for details)"),
134   cl::value_desc("a1,+a2,-a3,..."));
135 
136 static cl::opt<bool> PIC("position-independent",
137                          cl::desc("Position independent"), cl::init(false));
138 
139 static 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 
144 static cl::opt<bool>
145 NoInitialTextSection("n", cl::desc("Don't assume assembly file starts "
146                                    "in the text section"));
147 
148 static cl::opt<bool>
149 GenDwarfForAssembly("g", cl::desc("Generate dwarf debugging info for assembly "
150                                   "source files"));
151 
152 static cl::opt<std::string>
153 DebugCompilationDir("fdebug-compilation-dir",
154                     cl::desc("Specifies the debug info's compilation dir"));
155 
156 static cl::list<std::string>
157 DebugPrefixMap("fdebug-prefix-map",
158                cl::desc("Map file source paths in debug info"),
159                cl::value_desc("= separated key-value pairs"));
160 
161 static cl::opt<std::string>
162 MainFileName("main-file-name",
163              cl::desc("Specifies the name we should consider the input file"));
164 
165 static cl::opt<bool> SaveTempLabels("save-temp-labels",
166                                     cl::desc("Don't discard temporary labels"));
167 
168 static cl::opt<bool> LexMasmIntegers(
169     "masm-integers",
170     cl::desc("Enable binary and hex masm integers (0b110 and 0ABCh)"));
171 
172 static cl::opt<bool> LexMasmHexFloats(
173     "masm-hexfloats",
174     cl::desc("Enable MASM-style hex float initializers (3F800000r)"));
175 
176 static cl::opt<bool> LexMotorolaIntegers(
177     "motorola-integers",
178     cl::desc("Enable binary and hex Motorola integers (%110 and $ABC)"));
179 
180 static cl::opt<bool> NoExecStack("no-exec-stack",
181                                  cl::desc("File doesn't need an exec stack"));
182 
183 enum ActionType {
184   AC_AsLex,
185   AC_Assemble,
186   AC_Disassemble,
187   AC_MDisassemble,
188 };
189 
190 static cl::opt<ActionType>
191 Action(cl::desc("Action to perform:"),
192        cl::init(AC_Assemble),
193        cl::values(clEnumValN(AC_AsLex, "as-lex",
194                              "Lex tokens from a .s file"),
195                   clEnumValN(AC_Assemble, "assemble",
196                              "Assemble a .s file (default)"),
197                   clEnumValN(AC_Disassemble, "disassemble",
198                              "Disassemble strings of hex bytes"),
199                   clEnumValN(AC_MDisassemble, "mdis",
200                              "Marked up disassembly of strings of hex bytes")));
201 
202 static const Target *GetTarget(const char *ProgName) {
203   // Figure out the target triple.
204   if (TripleName.empty())
205     TripleName = sys::getDefaultTargetTriple();
206   Triple TheTriple(Triple::normalize(TripleName));
207 
208   // Get the target specific parser.
209   std::string Error;
210   const Target *TheTarget = TargetRegistry::lookupTarget(ArchName, TheTriple,
211                                                          Error);
212   if (!TheTarget) {
213     WithColor::error(errs(), ProgName) << Error;
214     return nullptr;
215   }
216 
217   // Update the triple name and return the found target.
218   TripleName = TheTriple.getTriple();
219   return TheTarget;
220 }
221 
222 static std::unique_ptr<ToolOutputFile> GetOutputStream(StringRef Path,
223     sys::fs::OpenFlags Flags) {
224   std::error_code EC;
225   auto Out = std::make_unique<ToolOutputFile>(Path, EC, Flags);
226   if (EC) {
227     WithColor::error() << EC.message() << '\n';
228     return nullptr;
229   }
230 
231   return Out;
232 }
233 
234 static std::string DwarfDebugFlags;
235 static void setDwarfDebugFlags(int argc, char **argv) {
236   if (!getenv("RC_DEBUG_OPTIONS"))
237     return;
238   for (int i = 0; i < argc; i++) {
239     DwarfDebugFlags += argv[i];
240     if (i + 1 < argc)
241       DwarfDebugFlags += " ";
242   }
243 }
244 
245 static std::string DwarfDebugProducer;
246 static void setDwarfDebugProducer() {
247   if(!getenv("DEBUG_PRODUCER"))
248     return;
249   DwarfDebugProducer += getenv("DEBUG_PRODUCER");
250 }
251 
252 static int AsLexInput(SourceMgr &SrcMgr, MCAsmInfo &MAI,
253                       raw_ostream &OS) {
254 
255   AsmLexer Lexer(MAI);
256   Lexer.setBuffer(SrcMgr.getMemoryBuffer(SrcMgr.getMainFileID())->getBuffer());
257 
258   bool Error = false;
259   while (Lexer.Lex().isNot(AsmToken::Eof)) {
260     Lexer.getTok().dump(OS);
261     OS << "\n";
262     if (Lexer.getTok().getKind() == AsmToken::Error)
263       Error = true;
264   }
265 
266   return Error;
267 }
268 
269 static int fillCommandLineSymbols(MCAsmParser &Parser) {
270   for (auto &I: DefineSymbol) {
271     auto Pair = StringRef(I).split('=');
272     auto Sym = Pair.first;
273     auto Val = Pair.second;
274 
275     if (Sym.empty() || Val.empty()) {
276       WithColor::error() << "defsym must be of the form: sym=value: " << I
277                          << "\n";
278       return 1;
279     }
280     int64_t Value;
281     if (Val.getAsInteger(0, Value)) {
282       WithColor::error() << "value is not an integer: " << Val << "\n";
283       return 1;
284     }
285     Parser.getContext().setSymbolValue(Parser.getStreamer(), Sym, Value);
286   }
287   return 0;
288 }
289 
290 static int AssembleInput(const char *ProgName, const Target *TheTarget,
291                          SourceMgr &SrcMgr, MCContext &Ctx, MCStreamer &Str,
292                          MCAsmInfo &MAI, MCSubtargetInfo &STI,
293                          MCInstrInfo &MCII, MCTargetOptions const &MCOptions) {
294   std::unique_ptr<MCAsmParser> Parser(
295       createMCAsmParser(SrcMgr, Ctx, Str, MAI));
296   std::unique_ptr<MCTargetAsmParser> TAP(
297       TheTarget->createMCAsmParser(STI, *Parser, MCII, MCOptions));
298 
299   if (!TAP) {
300     WithColor::error(errs(), ProgName)
301         << "this target does not support assembly parsing.\n";
302     return 1;
303   }
304 
305   int SymbolResult = fillCommandLineSymbols(*Parser);
306   if(SymbolResult)
307     return SymbolResult;
308   Parser->setShowParsedOperands(ShowInstOperands);
309   Parser->setTargetParser(*TAP);
310   Parser->getLexer().setLexMasmIntegers(LexMasmIntegers);
311   Parser->getLexer().setLexMasmHexFloats(LexMasmHexFloats);
312   Parser->getLexer().setLexMotorolaIntegers(LexMotorolaIntegers);
313 
314   int Res = Parser->Run(NoInitialTextSection);
315 
316   return Res;
317 }
318 
319 int main(int argc, char **argv) {
320   InitLLVM X(argc, argv);
321 
322   // Initialize targets and assembly printers/parsers.
323   llvm::InitializeAllTargetInfos();
324   llvm::InitializeAllTargetMCs();
325   llvm::InitializeAllAsmParsers();
326   llvm::InitializeAllDisassemblers();
327 
328   // Register the target printer for --version.
329   cl::AddExtraVersionPrinter(TargetRegistry::printRegisteredTargetsForVersion);
330 
331   cl::ParseCommandLineOptions(argc, argv, "llvm machine code playground\n");
332   const MCTargetOptions MCOptions = mc::InitMCTargetOptionsFromFlags();
333   setDwarfDebugFlags(argc, argv);
334 
335   setDwarfDebugProducer();
336 
337   const char *ProgName = argv[0];
338   const Target *TheTarget = GetTarget(ProgName);
339   if (!TheTarget)
340     return 1;
341   // Now that GetTarget() has (potentially) replaced TripleName, it's safe to
342   // construct the Triple object.
343   Triple TheTriple(TripleName);
344 
345   ErrorOr<std::unique_ptr<MemoryBuffer>> BufferPtr =
346       MemoryBuffer::getFileOrSTDIN(InputFilename, /*IsText=*/true);
347   if (std::error_code EC = BufferPtr.getError()) {
348     WithColor::error(errs(), ProgName)
349         << InputFilename << ": " << EC.message() << '\n';
350     return 1;
351   }
352   MemoryBuffer *Buffer = BufferPtr->get();
353 
354   SourceMgr SrcMgr;
355 
356   // Tell SrcMgr about this buffer, which is what the parser will pick up.
357   SrcMgr.AddNewSourceBuffer(std::move(*BufferPtr), SMLoc());
358 
359   // Record the location of the include directories so that the lexer can find
360   // it later.
361   SrcMgr.setIncludeDirs(IncludeDirs);
362 
363   std::unique_ptr<MCRegisterInfo> MRI(TheTarget->createMCRegInfo(TripleName));
364   assert(MRI && "Unable to create target register info!");
365 
366   std::unique_ptr<MCAsmInfo> MAI(
367       TheTarget->createMCAsmInfo(*MRI, TripleName, MCOptions));
368   assert(MAI && "Unable to create target asm info!");
369 
370   MAI->setRelaxELFRelocations(RelaxELFRel);
371 
372   if (CompressDebugSections != DebugCompressionType::None) {
373     if (!zlib::isAvailable()) {
374       WithColor::error(errs(), ProgName)
375           << "build tools with zlib to enable -compress-debug-sections";
376       return 1;
377     }
378     MAI->setCompressDebugSections(CompressDebugSections);
379   }
380   MAI->setPreserveAsmComments(PreserveComments);
381 
382   // FIXME: This is not pretty. MCContext has a ptr to MCObjectFileInfo and
383   // MCObjectFileInfo needs a MCContext reference in order to initialize itself.
384   MCObjectFileInfo MOFI;
385   MCContext Ctx(MAI.get(), MRI.get(), &MOFI, &SrcMgr, &MCOptions);
386   MOFI.InitMCObjectFileInfo(TheTriple, PIC, Ctx, LargeCodeModel);
387 
388   if (SaveTempLabels)
389     Ctx.setAllowTemporaryLabels(false);
390 
391   Ctx.setGenDwarfForAssembly(GenDwarfForAssembly);
392   // Default to 4 for dwarf version.
393   unsigned DwarfVersion = MCOptions.DwarfVersion ? MCOptions.DwarfVersion : 4;
394   if (DwarfVersion < 2 || DwarfVersion > 5) {
395     errs() << ProgName << ": Dwarf version " << DwarfVersion
396            << " is not supported." << '\n';
397     return 1;
398   }
399   Ctx.setDwarfVersion(DwarfVersion);
400   if (MCOptions.Dwarf64) {
401     // The 64-bit DWARF format was introduced in DWARFv3.
402     if (DwarfVersion < 3) {
403       errs() << ProgName
404              << ": the 64-bit DWARF format is not supported for DWARF versions "
405                 "prior to 3\n";
406       return 1;
407     }
408     // 32-bit targets don't support DWARF64, which requires 64-bit relocations.
409     if (MAI->getCodePointerSize() < 8) {
410       errs() << ProgName
411              << ": the 64-bit DWARF format is only supported for 64-bit "
412                 "targets\n";
413       return 1;
414     }
415     // If needsDwarfSectionOffsetDirective is true, we would eventually call
416     // MCStreamer::emitSymbolValue() with IsSectionRelative = true, but that
417     // is supported only for 4-byte long references.
418     if (MAI->needsDwarfSectionOffsetDirective()) {
419       errs() << ProgName << ": the 64-bit DWARF format is not supported for "
420              << TheTriple.normalize() << "\n";
421       return 1;
422     }
423     Ctx.setDwarfFormat(dwarf::DWARF64);
424   }
425   if (!DwarfDebugFlags.empty())
426     Ctx.setDwarfDebugFlags(StringRef(DwarfDebugFlags));
427   if (!DwarfDebugProducer.empty())
428     Ctx.setDwarfDebugProducer(StringRef(DwarfDebugProducer));
429   if (!DebugCompilationDir.empty())
430     Ctx.setCompilationDir(DebugCompilationDir);
431   else {
432     // If no compilation dir is set, try to use the current directory.
433     SmallString<128> CWD;
434     if (!sys::fs::current_path(CWD))
435       Ctx.setCompilationDir(CWD);
436   }
437   for (const auto &Arg : DebugPrefixMap) {
438     const auto &KV = StringRef(Arg).split('=');
439     Ctx.addDebugPrefixMapEntry(std::string(KV.first), std::string(KV.second));
440   }
441   if (!MainFileName.empty())
442     Ctx.setMainFileName(MainFileName);
443   if (GenDwarfForAssembly)
444     Ctx.setGenDwarfRootFile(InputFilename, Buffer->getBuffer());
445 
446   // Package up features to be passed to target/subtarget
447   std::string FeaturesStr;
448   if (MAttrs.size()) {
449     SubtargetFeatures Features;
450     for (unsigned i = 0; i != MAttrs.size(); ++i)
451       Features.AddFeature(MAttrs[i]);
452     FeaturesStr = Features.getString();
453   }
454 
455   sys::fs::OpenFlags Flags = (FileType == OFT_AssemblyFile)
456                                  ? sys::fs::OF_TextWithCRLF
457                                  : sys::fs::OF_None;
458   std::unique_ptr<ToolOutputFile> Out = GetOutputStream(OutputFilename, Flags);
459   if (!Out)
460     return 1;
461 
462   std::unique_ptr<ToolOutputFile> DwoOut;
463   if (!SplitDwarfFile.empty()) {
464     if (FileType != OFT_ObjectFile) {
465       WithColor::error() << "dwo output only supported with object files\n";
466       return 1;
467     }
468     DwoOut = GetOutputStream(SplitDwarfFile, sys::fs::OF_None);
469     if (!DwoOut)
470       return 1;
471   }
472 
473   std::unique_ptr<buffer_ostream> BOS;
474   raw_pwrite_stream *OS = &Out->os();
475   std::unique_ptr<MCStreamer> Str;
476 
477   std::unique_ptr<MCInstrInfo> MCII(TheTarget->createMCInstrInfo());
478   assert(MCII && "Unable to create instruction info!");
479 
480   std::unique_ptr<MCSubtargetInfo> STI(
481       TheTarget->createMCSubtargetInfo(TripleName, MCPU, FeaturesStr));
482   assert(STI && "Unable to create subtarget info!");
483 
484   MCInstPrinter *IP = nullptr;
485   if (FileType == OFT_AssemblyFile) {
486     IP = TheTarget->createMCInstPrinter(Triple(TripleName), OutputAsmVariant,
487                                         *MAI, *MCII, *MRI);
488 
489     if (!IP) {
490       WithColor::error()
491           << "unable to create instruction printer for target triple '"
492           << TheTriple.normalize() << "' with assembly variant "
493           << OutputAsmVariant << ".\n";
494       return 1;
495     }
496 
497     // Set the display preference for hex vs. decimal immediates.
498     IP->setPrintImmHex(PrintImmHex);
499 
500     // Set up the AsmStreamer.
501     std::unique_ptr<MCCodeEmitter> CE;
502     if (ShowEncoding)
503       CE.reset(TheTarget->createMCCodeEmitter(*MCII, *MRI, Ctx));
504 
505     std::unique_ptr<MCAsmBackend> MAB(
506         TheTarget->createMCAsmBackend(*STI, *MRI, MCOptions));
507     auto FOut = std::make_unique<formatted_raw_ostream>(*OS);
508     Str.reset(
509         TheTarget->createAsmStreamer(Ctx, std::move(FOut), /*asmverbose*/ true,
510                                      /*useDwarfDirectory*/ true, IP,
511                                      std::move(CE), std::move(MAB), ShowInst));
512 
513   } else if (FileType == OFT_Null) {
514     Str.reset(TheTarget->createNullStreamer(Ctx));
515   } else {
516     assert(FileType == OFT_ObjectFile && "Invalid file type!");
517 
518     if (!Out->os().supportsSeeking()) {
519       BOS = std::make_unique<buffer_ostream>(Out->os());
520       OS = BOS.get();
521     }
522 
523     MCCodeEmitter *CE = TheTarget->createMCCodeEmitter(*MCII, *MRI, Ctx);
524     MCAsmBackend *MAB = TheTarget->createMCAsmBackend(*STI, *MRI, MCOptions);
525     Str.reset(TheTarget->createMCObjectStreamer(
526         TheTriple, Ctx, std::unique_ptr<MCAsmBackend>(MAB),
527         DwoOut ? MAB->createDwoObjectWriter(*OS, DwoOut->os())
528                : MAB->createObjectWriter(*OS),
529         std::unique_ptr<MCCodeEmitter>(CE), *STI, MCOptions.MCRelaxAll,
530         MCOptions.MCIncrementalLinkerCompatible,
531         /*DWARFMustBeAtTheEnd*/ false));
532     if (NoExecStack)
533       Str->InitSections(true);
534   }
535 
536   // Use Assembler information for parsing.
537   Str->setUseAssemblerInfoForParsing(true);
538 
539   int Res = 1;
540   bool disassemble = false;
541   switch (Action) {
542   case AC_AsLex:
543     Res = AsLexInput(SrcMgr, *MAI, Out->os());
544     break;
545   case AC_Assemble:
546     Res = AssembleInput(ProgName, TheTarget, SrcMgr, Ctx, *Str, *MAI, *STI,
547                         *MCII, MCOptions);
548     break;
549   case AC_MDisassemble:
550     assert(IP && "Expected assembly output");
551     IP->setUseMarkup(true);
552     disassemble = true;
553     break;
554   case AC_Disassemble:
555     disassemble = true;
556     break;
557   }
558   if (disassemble)
559     Res = Disassembler::disassemble(*TheTarget, TripleName, *STI, *Str, *Buffer,
560                                     SrcMgr, Ctx, Out->os(), MCOptions);
561 
562   // Keep output if no errors.
563   if (Res == 0) {
564     Out->keep();
565     if (DwoOut)
566       DwoOut->keep();
567   }
568   return Res;
569 }
570