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::list<std::string>
50     DisassemblerOptions("M", cl::desc("Disassembler options"));
51 
52 static cl::opt<std::string> OutputFilename("o", cl::desc("Output filename"),
53                                            cl::value_desc("filename"),
54                                            cl::init("-"));
55 
56 static cl::opt<std::string> SplitDwarfFile("split-dwarf-file",
57                                            cl::desc("DWO output filename"),
58                                            cl::value_desc("filename"));
59 
60 static cl::opt<bool>
61 ShowEncoding("show-encoding", cl::desc("Show instruction encodings"));
62 
63 static cl::opt<bool> RelaxELFRel(
64     "relax-relocations", cl::init(true),
65     cl::desc("Emit R_X86_64_GOTPCRELX instead of R_X86_64_GOTPCREL"));
66 
67 static cl::opt<DebugCompressionType> CompressDebugSections(
68     "compress-debug-sections", cl::ValueOptional,
69     cl::init(DebugCompressionType::None),
70     cl::desc("Choose DWARF debug sections compression:"),
71     cl::values(clEnumValN(DebugCompressionType::None, "none", "No compression"),
72                clEnumValN(DebugCompressionType::Z, "zlib",
73                           "Use zlib compression"),
74                clEnumValN(DebugCompressionType::GNU, "zlib-gnu",
75                           "Use zlib-gnu compression (deprecated)")));
76 
77 static cl::opt<bool>
78 ShowInst("show-inst", cl::desc("Show internal instruction representation"));
79 
80 static cl::opt<bool>
81 ShowInstOperands("show-inst-operands",
82                  cl::desc("Show instructions operands as parsed"));
83 
84 static cl::opt<unsigned>
85 OutputAsmVariant("output-asm-variant",
86                  cl::desc("Syntax variant to use for output printing"));
87 
88 static cl::opt<bool>
89 PrintImmHex("print-imm-hex", cl::init(false),
90             cl::desc("Prefer hex format for immediate values"));
91 
92 static cl::list<std::string>
93 DefineSymbol("defsym", cl::desc("Defines a symbol to be an integer constant"));
94 
95 static cl::opt<bool>
96     PreserveComments("preserve-comments",
97                      cl::desc("Preserve Comments in outputted assembly"));
98 
99 enum OutputFileType {
100   OFT_Null,
101   OFT_AssemblyFile,
102   OFT_ObjectFile
103 };
104 static cl::opt<OutputFileType>
105 FileType("filetype", cl::init(OFT_AssemblyFile),
106   cl::desc("Choose an output file type:"),
107   cl::values(
108        clEnumValN(OFT_AssemblyFile, "asm",
109                   "Emit an assembly ('.s') file"),
110        clEnumValN(OFT_Null, "null",
111                   "Don't emit anything (for timing purposes)"),
112        clEnumValN(OFT_ObjectFile, "obj",
113                   "Emit a native object ('.o') file")));
114 
115 static cl::list<std::string>
116 IncludeDirs("I", cl::desc("Directory of include files"),
117             cl::value_desc("directory"), cl::Prefix);
118 
119 static cl::opt<std::string>
120 ArchName("arch", cl::desc("Target arch to assemble for, "
121                           "see -version for available targets"));
122 
123 static cl::opt<std::string>
124 TripleName("triple", cl::desc("Target triple to assemble for, "
125                               "see -version for available targets"));
126 
127 static cl::opt<std::string>
128 MCPU("mcpu",
129      cl::desc("Target a specific cpu type (-mcpu=help for details)"),
130      cl::value_desc("cpu-name"),
131      cl::init(""));
132 
133 static cl::list<std::string>
134 MAttrs("mattr",
135   cl::CommaSeparated,
136   cl::desc("Target specific attributes (-mattr=help for details)"),
137   cl::value_desc("a1,+a2,-a3,..."));
138 
139 static cl::opt<bool> PIC("position-independent",
140                          cl::desc("Position independent"), cl::init(false));
141 
142 static cl::opt<bool>
143     LargeCodeModel("large-code-model",
144                    cl::desc("Create cfi directives that assume the code might "
145                             "be more than 2gb away"));
146 
147 static cl::opt<bool>
148 NoInitialTextSection("n", cl::desc("Don't assume assembly file starts "
149                                    "in the text section"));
150 
151 static cl::opt<bool>
152 GenDwarfForAssembly("g", cl::desc("Generate dwarf debugging info for assembly "
153                                   "source files"));
154 
155 static cl::opt<std::string>
156 DebugCompilationDir("fdebug-compilation-dir",
157                     cl::desc("Specifies the debug info's compilation dir"));
158 
159 static cl::list<std::string>
160 DebugPrefixMap("fdebug-prefix-map",
161                cl::desc("Map file source paths in debug info"),
162                cl::value_desc("= separated key-value pairs"));
163 
164 static cl::opt<std::string>
165 MainFileName("main-file-name",
166              cl::desc("Specifies the name we should consider the input file"));
167 
168 static cl::opt<bool> SaveTempLabels("save-temp-labels",
169                                     cl::desc("Don't discard temporary labels"));
170 
171 static cl::opt<bool> LexMasmIntegers(
172     "masm-integers",
173     cl::desc("Enable binary and hex masm integers (0b110 and 0ABCh)"));
174 
175 static cl::opt<bool> LexMasmHexFloats(
176     "masm-hexfloats",
177     cl::desc("Enable MASM-style hex float initializers (3F800000r)"));
178 
179 static cl::opt<bool> LexMotorolaIntegers(
180     "motorola-integers",
181     cl::desc("Enable binary and hex Motorola integers (%110 and $ABC)"));
182 
183 static cl::opt<bool> NoExecStack("no-exec-stack",
184                                  cl::desc("File doesn't need an exec stack"));
185 
186 enum ActionType {
187   AC_AsLex,
188   AC_Assemble,
189   AC_Disassemble,
190   AC_MDisassemble,
191 };
192 
193 static cl::opt<ActionType>
194 Action(cl::desc("Action to perform:"),
195        cl::init(AC_Assemble),
196        cl::values(clEnumValN(AC_AsLex, "as-lex",
197                              "Lex tokens from a .s file"),
198                   clEnumValN(AC_Assemble, "assemble",
199                              "Assemble a .s file (default)"),
200                   clEnumValN(AC_Disassemble, "disassemble",
201                              "Disassemble strings of hex bytes"),
202                   clEnumValN(AC_MDisassemble, "mdis",
203                              "Marked up disassembly of strings of hex bytes")));
204 
205 static const Target *GetTarget(const char *ProgName) {
206   // Figure out the target triple.
207   if (TripleName.empty())
208     TripleName = sys::getDefaultTargetTriple();
209   Triple TheTriple(Triple::normalize(TripleName));
210 
211   // Get the target specific parser.
212   std::string Error;
213   const Target *TheTarget = TargetRegistry::lookupTarget(ArchName, TheTriple,
214                                                          Error);
215   if (!TheTarget) {
216     WithColor::error(errs(), ProgName) << Error;
217     return nullptr;
218   }
219 
220   // Update the triple name and return the found target.
221   TripleName = TheTriple.getTriple();
222   return TheTarget;
223 }
224 
225 static std::unique_ptr<ToolOutputFile> GetOutputStream(StringRef Path,
226     sys::fs::OpenFlags Flags) {
227   std::error_code EC;
228   auto Out = std::make_unique<ToolOutputFile>(Path, EC, Flags);
229   if (EC) {
230     WithColor::error() << EC.message() << '\n';
231     return nullptr;
232   }
233 
234   return Out;
235 }
236 
237 static std::string DwarfDebugFlags;
238 static void setDwarfDebugFlags(int argc, char **argv) {
239   if (!getenv("RC_DEBUG_OPTIONS"))
240     return;
241   for (int i = 0; i < argc; i++) {
242     DwarfDebugFlags += argv[i];
243     if (i + 1 < argc)
244       DwarfDebugFlags += " ";
245   }
246 }
247 
248 static std::string DwarfDebugProducer;
249 static void setDwarfDebugProducer() {
250   if(!getenv("DEBUG_PRODUCER"))
251     return;
252   DwarfDebugProducer += getenv("DEBUG_PRODUCER");
253 }
254 
255 static int AsLexInput(SourceMgr &SrcMgr, MCAsmInfo &MAI,
256                       raw_ostream &OS) {
257 
258   AsmLexer Lexer(MAI);
259   Lexer.setBuffer(SrcMgr.getMemoryBuffer(SrcMgr.getMainFileID())->getBuffer());
260 
261   bool Error = false;
262   while (Lexer.Lex().isNot(AsmToken::Eof)) {
263     Lexer.getTok().dump(OS);
264     OS << "\n";
265     if (Lexer.getTok().getKind() == AsmToken::Error)
266       Error = true;
267   }
268 
269   return Error;
270 }
271 
272 static int fillCommandLineSymbols(MCAsmParser &Parser) {
273   for (auto &I: DefineSymbol) {
274     auto Pair = StringRef(I).split('=');
275     auto Sym = Pair.first;
276     auto Val = Pair.second;
277 
278     if (Sym.empty() || Val.empty()) {
279       WithColor::error() << "defsym must be of the form: sym=value: " << I
280                          << "\n";
281       return 1;
282     }
283     int64_t Value;
284     if (Val.getAsInteger(0, Value)) {
285       WithColor::error() << "value is not an integer: " << Val << "\n";
286       return 1;
287     }
288     Parser.getContext().setSymbolValue(Parser.getStreamer(), Sym, Value);
289   }
290   return 0;
291 }
292 
293 static int AssembleInput(const char *ProgName, const Target *TheTarget,
294                          SourceMgr &SrcMgr, MCContext &Ctx, MCStreamer &Str,
295                          MCAsmInfo &MAI, MCSubtargetInfo &STI,
296                          MCInstrInfo &MCII, MCTargetOptions const &MCOptions) {
297   std::unique_ptr<MCAsmParser> Parser(
298       createMCAsmParser(SrcMgr, Ctx, Str, MAI));
299   std::unique_ptr<MCTargetAsmParser> TAP(
300       TheTarget->createMCAsmParser(STI, *Parser, MCII, MCOptions));
301 
302   if (!TAP) {
303     WithColor::error(errs(), ProgName)
304         << "this target does not support assembly parsing.\n";
305     return 1;
306   }
307 
308   int SymbolResult = fillCommandLineSymbols(*Parser);
309   if(SymbolResult)
310     return SymbolResult;
311   Parser->setShowParsedOperands(ShowInstOperands);
312   Parser->setTargetParser(*TAP);
313   Parser->getLexer().setLexMasmIntegers(LexMasmIntegers);
314   Parser->getLexer().setLexMasmHexFloats(LexMasmHexFloats);
315   Parser->getLexer().setLexMotorolaIntegers(LexMotorolaIntegers);
316 
317   int Res = Parser->Run(NoInitialTextSection);
318 
319   return Res;
320 }
321 
322 int main(int argc, char **argv) {
323   InitLLVM X(argc, argv);
324 
325   // Initialize targets and assembly printers/parsers.
326   llvm::InitializeAllTargetInfos();
327   llvm::InitializeAllTargetMCs();
328   llvm::InitializeAllAsmParsers();
329   llvm::InitializeAllDisassemblers();
330 
331   // Register the target printer for --version.
332   cl::AddExtraVersionPrinter(TargetRegistry::printRegisteredTargetsForVersion);
333 
334   cl::ParseCommandLineOptions(argc, argv, "llvm machine code playground\n");
335   const MCTargetOptions MCOptions = mc::InitMCTargetOptionsFromFlags();
336   setDwarfDebugFlags(argc, argv);
337 
338   setDwarfDebugProducer();
339 
340   const char *ProgName = argv[0];
341   const Target *TheTarget = GetTarget(ProgName);
342   if (!TheTarget)
343     return 1;
344   // Now that GetTarget() has (potentially) replaced TripleName, it's safe to
345   // construct the Triple object.
346   Triple TheTriple(TripleName);
347 
348   ErrorOr<std::unique_ptr<MemoryBuffer>> BufferPtr =
349       MemoryBuffer::getFileOrSTDIN(InputFilename, /*IsText=*/true);
350   if (std::error_code EC = BufferPtr.getError()) {
351     WithColor::error(errs(), ProgName)
352         << InputFilename << ": " << EC.message() << '\n';
353     return 1;
354   }
355   MemoryBuffer *Buffer = BufferPtr->get();
356 
357   SourceMgr SrcMgr;
358 
359   // Tell SrcMgr about this buffer, which is what the parser will pick up.
360   SrcMgr.AddNewSourceBuffer(std::move(*BufferPtr), SMLoc());
361 
362   // Record the location of the include directories so that the lexer can find
363   // it later.
364   SrcMgr.setIncludeDirs(IncludeDirs);
365 
366   std::unique_ptr<MCRegisterInfo> MRI(TheTarget->createMCRegInfo(TripleName));
367   assert(MRI && "Unable to create target register info!");
368 
369   std::unique_ptr<MCAsmInfo> MAI(
370       TheTarget->createMCAsmInfo(*MRI, TripleName, MCOptions));
371   assert(MAI && "Unable to create target asm info!");
372 
373   MAI->setRelaxELFRelocations(RelaxELFRel);
374 
375   if (CompressDebugSections != DebugCompressionType::None) {
376     if (!zlib::isAvailable()) {
377       WithColor::error(errs(), ProgName)
378           << "build tools with zlib to enable -compress-debug-sections";
379       return 1;
380     }
381     MAI->setCompressDebugSections(CompressDebugSections);
382   }
383   MAI->setPreserveAsmComments(PreserveComments);
384 
385   // Package up features to be passed to target/subtarget
386   std::string FeaturesStr;
387   if (MAttrs.size()) {
388     SubtargetFeatures Features;
389     for (unsigned i = 0; i != MAttrs.size(); ++i)
390       Features.AddFeature(MAttrs[i]);
391     FeaturesStr = Features.getString();
392   }
393 
394   std::unique_ptr<MCSubtargetInfo> STI(
395       TheTarget->createMCSubtargetInfo(TripleName, MCPU, FeaturesStr));
396   assert(STI && "Unable to create subtarget info!");
397 
398   // FIXME: This is not pretty. MCContext has a ptr to MCObjectFileInfo and
399   // MCObjectFileInfo needs a MCContext reference in order to initialize itself.
400   MCContext Ctx(TheTriple, MAI.get(), MRI.get(), STI.get(), &SrcMgr,
401                 &MCOptions);
402   std::unique_ptr<MCObjectFileInfo> MOFI(
403       TheTarget->createMCObjectFileInfo(Ctx, PIC, LargeCodeModel));
404   Ctx.setObjectFileInfo(MOFI.get());
405 
406   if (SaveTempLabels)
407     Ctx.setAllowTemporaryLabels(false);
408 
409   Ctx.setGenDwarfForAssembly(GenDwarfForAssembly);
410   // Default to 4 for dwarf version.
411   unsigned DwarfVersion = MCOptions.DwarfVersion ? MCOptions.DwarfVersion : 4;
412   if (DwarfVersion < 2 || DwarfVersion > 5) {
413     errs() << ProgName << ": Dwarf version " << DwarfVersion
414            << " is not supported." << '\n';
415     return 1;
416   }
417   Ctx.setDwarfVersion(DwarfVersion);
418   if (MCOptions.Dwarf64) {
419     // The 64-bit DWARF format was introduced in DWARFv3.
420     if (DwarfVersion < 3) {
421       errs() << ProgName
422              << ": the 64-bit DWARF format is not supported for DWARF versions "
423                 "prior to 3\n";
424       return 1;
425     }
426     // 32-bit targets don't support DWARF64, which requires 64-bit relocations.
427     if (MAI->getCodePointerSize() < 8) {
428       errs() << ProgName
429              << ": the 64-bit DWARF format is only supported for 64-bit "
430                 "targets\n";
431       return 1;
432     }
433     // If needsDwarfSectionOffsetDirective is true, we would eventually call
434     // MCStreamer::emitSymbolValue() with IsSectionRelative = true, but that
435     // is supported only for 4-byte long references.
436     if (MAI->needsDwarfSectionOffsetDirective()) {
437       errs() << ProgName << ": the 64-bit DWARF format is not supported for "
438              << TheTriple.normalize() << "\n";
439       return 1;
440     }
441     Ctx.setDwarfFormat(dwarf::DWARF64);
442   }
443   if (!DwarfDebugFlags.empty())
444     Ctx.setDwarfDebugFlags(StringRef(DwarfDebugFlags));
445   if (!DwarfDebugProducer.empty())
446     Ctx.setDwarfDebugProducer(StringRef(DwarfDebugProducer));
447   if (!DebugCompilationDir.empty())
448     Ctx.setCompilationDir(DebugCompilationDir);
449   else {
450     // If no compilation dir is set, try to use the current directory.
451     SmallString<128> CWD;
452     if (!sys::fs::current_path(CWD))
453       Ctx.setCompilationDir(CWD);
454   }
455   for (const auto &Arg : DebugPrefixMap) {
456     const auto &KV = StringRef(Arg).split('=');
457     Ctx.addDebugPrefixMapEntry(std::string(KV.first), std::string(KV.second));
458   }
459   if (!MainFileName.empty())
460     Ctx.setMainFileName(MainFileName);
461   if (GenDwarfForAssembly)
462     Ctx.setGenDwarfRootFile(InputFilename, Buffer->getBuffer());
463 
464   sys::fs::OpenFlags Flags = (FileType == OFT_AssemblyFile)
465                                  ? sys::fs::OF_TextWithCRLF
466                                  : sys::fs::OF_None;
467   std::unique_ptr<ToolOutputFile> Out = GetOutputStream(OutputFilename, Flags);
468   if (!Out)
469     return 1;
470 
471   std::unique_ptr<ToolOutputFile> DwoOut;
472   if (!SplitDwarfFile.empty()) {
473     if (FileType != OFT_ObjectFile) {
474       WithColor::error() << "dwo output only supported with object files\n";
475       return 1;
476     }
477     DwoOut = GetOutputStream(SplitDwarfFile, sys::fs::OF_None);
478     if (!DwoOut)
479       return 1;
480   }
481 
482   std::unique_ptr<buffer_ostream> BOS;
483   raw_pwrite_stream *OS = &Out->os();
484   std::unique_ptr<MCStreamer> Str;
485 
486   std::unique_ptr<MCInstrInfo> MCII(TheTarget->createMCInstrInfo());
487   assert(MCII && "Unable to create instruction info!");
488 
489   MCInstPrinter *IP = nullptr;
490   if (FileType == OFT_AssemblyFile) {
491     IP = TheTarget->createMCInstPrinter(Triple(TripleName), OutputAsmVariant,
492                                         *MAI, *MCII, *MRI);
493 
494     if (!IP) {
495       WithColor::error()
496           << "unable to create instruction printer for target triple '"
497           << TheTriple.normalize() << "' with assembly variant "
498           << OutputAsmVariant << ".\n";
499       return 1;
500     }
501 
502     for (StringRef Opt : DisassemblerOptions)
503       if (!IP->applyTargetSpecificCLOption(Opt)) {
504         WithColor::error() << "invalid disassembler option '" << Opt << "'\n";
505         return 1;
506       }
507 
508     // Set the display preference for hex vs. decimal immediates.
509     IP->setPrintImmHex(PrintImmHex);
510 
511     // Set up the AsmStreamer.
512     std::unique_ptr<MCCodeEmitter> CE;
513     if (ShowEncoding)
514       CE.reset(TheTarget->createMCCodeEmitter(*MCII, *MRI, Ctx));
515 
516     std::unique_ptr<MCAsmBackend> MAB(
517         TheTarget->createMCAsmBackend(*STI, *MRI, MCOptions));
518     auto FOut = std::make_unique<formatted_raw_ostream>(*OS);
519     Str.reset(
520         TheTarget->createAsmStreamer(Ctx, std::move(FOut), /*asmverbose*/ true,
521                                      /*useDwarfDirectory*/ true, IP,
522                                      std::move(CE), std::move(MAB), ShowInst));
523 
524   } else if (FileType == OFT_Null) {
525     Str.reset(TheTarget->createNullStreamer(Ctx));
526   } else {
527     assert(FileType == OFT_ObjectFile && "Invalid file type!");
528 
529     if (!Out->os().supportsSeeking()) {
530       BOS = std::make_unique<buffer_ostream>(Out->os());
531       OS = BOS.get();
532     }
533 
534     MCCodeEmitter *CE = TheTarget->createMCCodeEmitter(*MCII, *MRI, Ctx);
535     MCAsmBackend *MAB = TheTarget->createMCAsmBackend(*STI, *MRI, MCOptions);
536     Str.reset(TheTarget->createMCObjectStreamer(
537         TheTriple, Ctx, std::unique_ptr<MCAsmBackend>(MAB),
538         DwoOut ? MAB->createDwoObjectWriter(*OS, DwoOut->os())
539                : MAB->createObjectWriter(*OS),
540         std::unique_ptr<MCCodeEmitter>(CE), *STI, MCOptions.MCRelaxAll,
541         MCOptions.MCIncrementalLinkerCompatible,
542         /*DWARFMustBeAtTheEnd*/ false));
543     if (NoExecStack)
544       Str->InitSections(true);
545   }
546 
547   // Use Assembler information for parsing.
548   Str->setUseAssemblerInfoForParsing(true);
549 
550   int Res = 1;
551   bool disassemble = false;
552   switch (Action) {
553   case AC_AsLex:
554     Res = AsLexInput(SrcMgr, *MAI, Out->os());
555     break;
556   case AC_Assemble:
557     Res = AssembleInput(ProgName, TheTarget, SrcMgr, Ctx, *Str, *MAI, *STI,
558                         *MCII, MCOptions);
559     break;
560   case AC_MDisassemble:
561     assert(IP && "Expected assembly output");
562     IP->setUseMarkup(true);
563     disassemble = true;
564     break;
565   case AC_Disassemble:
566     disassemble = true;
567     break;
568   }
569   if (disassemble)
570     Res = Disassembler::disassemble(*TheTarget, TripleName, *STI, *Str, *Buffer,
571                                     SrcMgr, Ctx, Out->os(), MCOptions);
572 
573   // Keep output if no errors.
574   if (Res == 0) {
575     Out->keep();
576     if (DwoOut)
577       DwoOut->keep();
578   }
579   return Res;
580 }
581