1 //===-- llvm-mc.cpp - Machine Code Hacking Driver ---------------*- C++ -*-===//
2 //
3 //                     The LLVM Compiler Infrastructure
4 //
5 // This file is distributed under the University of Illinois Open Source
6 // License. See LICENSE.TXT for details.
7 //
8 //===----------------------------------------------------------------------===//
9 //
10 // This utility is a simple driver that allows command line hacking on machine
11 // code.
12 //
13 //===----------------------------------------------------------------------===//
14 
15 #include "Disassembler.h"
16 #include "llvm/MC/MCAsmBackend.h"
17 #include "llvm/MC/MCAsmInfo.h"
18 #include "llvm/MC/MCCodeEmitter.h"
19 #include "llvm/MC/MCContext.h"
20 #include "llvm/MC/MCInstPrinter.h"
21 #include "llvm/MC/MCInstrInfo.h"
22 #include "llvm/MC/MCObjectFileInfo.h"
23 #include "llvm/MC/MCObjectWriter.h"
24 #include "llvm/MC/MCParser/AsmLexer.h"
25 #include "llvm/MC/MCParser/MCTargetAsmParser.h"
26 #include "llvm/MC/MCRegisterInfo.h"
27 #include "llvm/MC/MCStreamer.h"
28 #include "llvm/MC/MCSubtargetInfo.h"
29 #include "llvm/MC/MCTargetOptionsCommandFlags.inc"
30 #include "llvm/Support/CommandLine.h"
31 #include "llvm/Support/Compression.h"
32 #include "llvm/Support/FileUtilities.h"
33 #include "llvm/Support/FormattedStream.h"
34 #include "llvm/Support/Host.h"
35 #include "llvm/Support/InitLLVM.h"
36 #include "llvm/Support/MemoryBuffer.h"
37 #include "llvm/Support/SourceMgr.h"
38 #include "llvm/Support/TargetRegistry.h"
39 #include "llvm/Support/TargetSelect.h"
40 #include "llvm/Support/ToolOutputFile.h"
41 #include "llvm/Support/WithColor.h"
42 
43 using namespace llvm;
44 
45 static cl::opt<std::string>
46 InputFilename(cl::Positional, cl::desc("<input file>"), cl::init("-"));
47 
48 static cl::opt<std::string> OutputFilename("o", cl::desc("Output filename"),
49                                            cl::value_desc("filename"),
50                                            cl::init("-"));
51 
52 static cl::opt<std::string> SplitDwarfFile("split-dwarf-file",
53                                            cl::desc("DWO output filename"),
54                                            cl::value_desc("filename"));
55 
56 static cl::opt<bool>
57 ShowEncoding("show-encoding", cl::desc("Show instruction encodings"));
58 
59 static cl::opt<bool> RelaxELFRel(
60     "relax-relocations", cl::init(true),
61     cl::desc("Emit R_X86_64_GOTPCRELX instead of R_X86_64_GOTPCREL"));
62 
63 static cl::opt<DebugCompressionType> CompressDebugSections(
64     "compress-debug-sections", cl::ValueOptional,
65     cl::init(DebugCompressionType::None),
66     cl::desc("Choose DWARF debug sections compression:"),
67     cl::values(clEnumValN(DebugCompressionType::None, "none", "No compression"),
68                clEnumValN(DebugCompressionType::Z, "zlib",
69                           "Use zlib compression"),
70                clEnumValN(DebugCompressionType::GNU, "zlib-gnu",
71                           "Use zlib-gnu compression (deprecated)")));
72 
73 static cl::opt<bool>
74 ShowInst("show-inst", cl::desc("Show internal instruction representation"));
75 
76 static cl::opt<bool>
77 ShowInstOperands("show-inst-operands",
78                  cl::desc("Show instructions operands as parsed"));
79 
80 static cl::opt<unsigned>
81 OutputAsmVariant("output-asm-variant",
82                  cl::desc("Syntax variant to use for output printing"));
83 
84 static cl::opt<bool>
85 PrintImmHex("print-imm-hex", cl::init(false),
86             cl::desc("Prefer hex format for immediate values"));
87 
88 static cl::list<std::string>
89 DefineSymbol("defsym", cl::desc("Defines a symbol to be an integer constant"));
90 
91 static cl::opt<bool>
92     PreserveComments("preserve-comments",
93                      cl::desc("Preserve Comments in outputted assembly"));
94 
95 enum OutputFileType {
96   OFT_Null,
97   OFT_AssemblyFile,
98   OFT_ObjectFile
99 };
100 static cl::opt<OutputFileType>
101 FileType("filetype", cl::init(OFT_AssemblyFile),
102   cl::desc("Choose an output file type:"),
103   cl::values(
104        clEnumValN(OFT_AssemblyFile, "asm",
105                   "Emit an assembly ('.s') file"),
106        clEnumValN(OFT_Null, "null",
107                   "Don't emit anything (for timing purposes)"),
108        clEnumValN(OFT_ObjectFile, "obj",
109                   "Emit a native object ('.o') file")));
110 
111 static cl::list<std::string>
112 IncludeDirs("I", cl::desc("Directory of include files"),
113             cl::value_desc("directory"), cl::Prefix);
114 
115 static cl::opt<std::string>
116 ArchName("arch", cl::desc("Target arch to assemble for, "
117                           "see -version for available targets"));
118 
119 static cl::opt<std::string>
120 TripleName("triple", cl::desc("Target triple to assemble for, "
121                               "see -version for available targets"));
122 
123 static cl::opt<std::string>
124 MCPU("mcpu",
125      cl::desc("Target a specific cpu type (-mcpu=help for details)"),
126      cl::value_desc("cpu-name"),
127      cl::init(""));
128 
129 static cl::list<std::string>
130 MAttrs("mattr",
131   cl::CommaSeparated,
132   cl::desc("Target specific attributes (-mattr=help for details)"),
133   cl::value_desc("a1,+a2,-a3,..."));
134 
135 static cl::opt<bool> PIC("position-independent",
136                          cl::desc("Position independent"), cl::init(false));
137 
138 static cl::opt<bool>
139     LargeCodeModel("large-code-model",
140                    cl::desc("Create cfi directives that assume the code might "
141                             "be more than 2gb away"));
142 
143 static cl::opt<bool>
144 NoInitialTextSection("n", cl::desc("Don't assume assembly file starts "
145                                    "in the text section"));
146 
147 static cl::opt<bool>
148 GenDwarfForAssembly("g", cl::desc("Generate dwarf debugging info for assembly "
149                                   "source files"));
150 
151 static cl::opt<std::string>
152 DebugCompilationDir("fdebug-compilation-dir",
153                     cl::desc("Specifies the debug info's compilation dir"));
154 
155 static cl::list<std::string>
156 DebugPrefixMap("fdebug-prefix-map",
157                cl::desc("Map file source paths in debug info"),
158                cl::value_desc("= separated key-value pairs"));
159 
160 static cl::opt<std::string>
161 MainFileName("main-file-name",
162              cl::desc("Specifies the name we should consider the input file"));
163 
164 static cl::opt<bool> SaveTempLabels("save-temp-labels",
165                                     cl::desc("Don't discard temporary labels"));
166 
167 static cl::opt<bool> NoExecStack("no-exec-stack",
168                                  cl::desc("File doesn't need an exec stack"));
169 
170 enum ActionType {
171   AC_AsLex,
172   AC_Assemble,
173   AC_Disassemble,
174   AC_MDisassemble,
175 };
176 
177 static cl::opt<ActionType>
178 Action(cl::desc("Action to perform:"),
179        cl::init(AC_Assemble),
180        cl::values(clEnumValN(AC_AsLex, "as-lex",
181                              "Lex tokens from a .s file"),
182                   clEnumValN(AC_Assemble, "assemble",
183                              "Assemble a .s file (default)"),
184                   clEnumValN(AC_Disassemble, "disassemble",
185                              "Disassemble strings of hex bytes"),
186                   clEnumValN(AC_MDisassemble, "mdis",
187                              "Marked up disassembly of strings of hex bytes")));
188 
189 static const Target *GetTarget(const char *ProgName) {
190   // Figure out the target triple.
191   if (TripleName.empty())
192     TripleName = sys::getDefaultTargetTriple();
193   Triple TheTriple(Triple::normalize(TripleName));
194 
195   // Get the target specific parser.
196   std::string Error;
197   const Target *TheTarget = TargetRegistry::lookupTarget(ArchName, TheTriple,
198                                                          Error);
199   if (!TheTarget) {
200     WithColor::error(errs(), ProgName) << Error;
201     return nullptr;
202   }
203 
204   // Update the triple name and return the found target.
205   TripleName = TheTriple.getTriple();
206   return TheTarget;
207 }
208 
209 static std::unique_ptr<ToolOutputFile> GetOutputStream(StringRef Path) {
210   std::error_code EC;
211   auto Out = llvm::make_unique<ToolOutputFile>(Path, EC, sys::fs::F_None);
212   if (EC) {
213     WithColor::error() << EC.message() << '\n';
214     return nullptr;
215   }
216 
217   return Out;
218 }
219 
220 static std::string DwarfDebugFlags;
221 static void setDwarfDebugFlags(int argc, char **argv) {
222   if (!getenv("RC_DEBUG_OPTIONS"))
223     return;
224   for (int i = 0; i < argc; i++) {
225     DwarfDebugFlags += argv[i];
226     if (i + 1 < argc)
227       DwarfDebugFlags += " ";
228   }
229 }
230 
231 static std::string DwarfDebugProducer;
232 static void setDwarfDebugProducer() {
233   if(!getenv("DEBUG_PRODUCER"))
234     return;
235   DwarfDebugProducer += getenv("DEBUG_PRODUCER");
236 }
237 
238 static int AsLexInput(SourceMgr &SrcMgr, MCAsmInfo &MAI,
239                       raw_ostream &OS) {
240 
241   AsmLexer Lexer(MAI);
242   Lexer.setBuffer(SrcMgr.getMemoryBuffer(SrcMgr.getMainFileID())->getBuffer());
243 
244   bool Error = false;
245   while (Lexer.Lex().isNot(AsmToken::Eof)) {
246     Lexer.getTok().dump(OS);
247     OS << "\n";
248     if (Lexer.getTok().getKind() == AsmToken::Error)
249       Error = true;
250   }
251 
252   return Error;
253 }
254 
255 static int fillCommandLineSymbols(MCAsmParser &Parser) {
256   for (auto &I: DefineSymbol) {
257     auto Pair = StringRef(I).split('=');
258     auto Sym = Pair.first;
259     auto Val = Pair.second;
260 
261     if (Sym.empty() || Val.empty()) {
262       WithColor::error() << "defsym must be of the form: sym=value: " << I
263                          << "\n";
264       return 1;
265     }
266     int64_t Value;
267     if (Val.getAsInteger(0, Value)) {
268       WithColor::error() << "value is not an integer: " << Val << "\n";
269       return 1;
270     }
271     Parser.getContext().setSymbolValue(Parser.getStreamer(), Sym, Value);
272   }
273   return 0;
274 }
275 
276 static int AssembleInput(const char *ProgName, const Target *TheTarget,
277                          SourceMgr &SrcMgr, MCContext &Ctx, MCStreamer &Str,
278                          MCAsmInfo &MAI, MCSubtargetInfo &STI,
279                          MCInstrInfo &MCII, MCTargetOptions &MCOptions) {
280   std::unique_ptr<MCAsmParser> Parser(
281       createMCAsmParser(SrcMgr, Ctx, Str, MAI));
282   std::unique_ptr<MCTargetAsmParser> TAP(
283       TheTarget->createMCAsmParser(STI, *Parser, MCII, MCOptions));
284 
285   if (!TAP) {
286     WithColor::error(errs(), ProgName)
287         << "this target does not support assembly parsing.\n";
288     return 1;
289   }
290 
291   int SymbolResult = fillCommandLineSymbols(*Parser);
292   if(SymbolResult)
293     return SymbolResult;
294   Parser->setShowParsedOperands(ShowInstOperands);
295   Parser->setTargetParser(*TAP);
296 
297   int Res = Parser->Run(NoInitialTextSection);
298 
299   return Res;
300 }
301 
302 int main(int argc, char **argv) {
303   InitLLVM X(argc, argv);
304 
305   // Initialize targets and assembly printers/parsers.
306   llvm::InitializeAllTargetInfos();
307   llvm::InitializeAllTargetMCs();
308   llvm::InitializeAllAsmParsers();
309   llvm::InitializeAllDisassemblers();
310 
311   // Register the target printer for --version.
312   cl::AddExtraVersionPrinter(TargetRegistry::printRegisteredTargetsForVersion);
313 
314   cl::ParseCommandLineOptions(argc, argv, "llvm machine code playground\n");
315   MCTargetOptions MCOptions = InitMCTargetOptionsFromFlags();
316   TripleName = Triple::normalize(TripleName);
317   setDwarfDebugFlags(argc, argv);
318 
319   setDwarfDebugProducer();
320 
321   const char *ProgName = argv[0];
322   const Target *TheTarget = GetTarget(ProgName);
323   if (!TheTarget)
324     return 1;
325   // Now that GetTarget() has (potentially) replaced TripleName, it's safe to
326   // construct the Triple object.
327   Triple TheTriple(TripleName);
328 
329   ErrorOr<std::unique_ptr<MemoryBuffer>> BufferPtr =
330       MemoryBuffer::getFileOrSTDIN(InputFilename);
331   if (std::error_code EC = BufferPtr.getError()) {
332     WithColor::error(errs(), ProgName)
333         << InputFilename << ": " << EC.message() << '\n';
334     return 1;
335   }
336   MemoryBuffer *Buffer = BufferPtr->get();
337 
338   SourceMgr SrcMgr;
339 
340   // Tell SrcMgr about this buffer, which is what the parser will pick up.
341   SrcMgr.AddNewSourceBuffer(std::move(*BufferPtr), SMLoc());
342 
343   // Record the location of the include directories so that the lexer can find
344   // it later.
345   SrcMgr.setIncludeDirs(IncludeDirs);
346 
347   std::unique_ptr<MCRegisterInfo> MRI(TheTarget->createMCRegInfo(TripleName));
348   assert(MRI && "Unable to create target register info!");
349 
350   std::unique_ptr<MCAsmInfo> MAI(TheTarget->createMCAsmInfo(*MRI, TripleName));
351   assert(MAI && "Unable to create target asm info!");
352 
353   MAI->setRelaxELFRelocations(RelaxELFRel);
354 
355   if (CompressDebugSections != DebugCompressionType::None) {
356     if (!zlib::isAvailable()) {
357       WithColor::error(errs(), ProgName)
358           << "build tools with zlib to enable -compress-debug-sections";
359       return 1;
360     }
361     MAI->setCompressDebugSections(CompressDebugSections);
362   }
363   MAI->setPreserveAsmComments(PreserveComments);
364 
365   // FIXME: This is not pretty. MCContext has a ptr to MCObjectFileInfo and
366   // MCObjectFileInfo needs a MCContext reference in order to initialize itself.
367   MCObjectFileInfo MOFI;
368   MCContext Ctx(MAI.get(), MRI.get(), &MOFI, &SrcMgr);
369   MOFI.InitMCObjectFileInfo(TheTriple, PIC, Ctx, LargeCodeModel);
370 
371   if (SaveTempLabels)
372     Ctx.setAllowTemporaryLabels(false);
373 
374   Ctx.setGenDwarfForAssembly(GenDwarfForAssembly);
375   // Default to 4 for dwarf version.
376   unsigned DwarfVersion = MCOptions.DwarfVersion ? MCOptions.DwarfVersion : 4;
377   if (DwarfVersion < 2 || DwarfVersion > 5) {
378     errs() << ProgName << ": Dwarf version " << DwarfVersion
379            << " is not supported." << '\n';
380     return 1;
381   }
382   Ctx.setDwarfVersion(DwarfVersion);
383   if (!DwarfDebugFlags.empty())
384     Ctx.setDwarfDebugFlags(StringRef(DwarfDebugFlags));
385   if (!DwarfDebugProducer.empty())
386     Ctx.setDwarfDebugProducer(StringRef(DwarfDebugProducer));
387   if (!DebugCompilationDir.empty())
388     Ctx.setCompilationDir(DebugCompilationDir);
389   else {
390     // If no compilation dir is set, try to use the current directory.
391     SmallString<128> CWD;
392     if (!sys::fs::current_path(CWD))
393       Ctx.setCompilationDir(CWD);
394   }
395   for (const auto &Arg : DebugPrefixMap) {
396     const auto &KV = StringRef(Arg).split('=');
397     Ctx.addDebugPrefixMapEntry(KV.first, KV.second);
398   }
399   if (!MainFileName.empty())
400     Ctx.setMainFileName(MainFileName);
401   if (GenDwarfForAssembly && DwarfVersion >= 5) {
402     // DWARF v5 needs the root file as well as the compilation directory.
403     // If we find a '.file 0' directive that will supersede these values.
404     MD5 Hash;
405     MD5::MD5Result *Cksum =
406         (MD5::MD5Result *)Ctx.allocate(sizeof(MD5::MD5Result), 1);
407     Hash.update(Buffer->getBuffer());
408     Hash.final(*Cksum);
409     Ctx.setMCLineTableRootFile(
410         /*CUID=*/0, Ctx.getCompilationDir(),
411         !MainFileName.empty() ? MainFileName : InputFilename, Cksum, None);
412   }
413 
414   // Package up features to be passed to target/subtarget
415   std::string FeaturesStr;
416   if (MAttrs.size()) {
417     SubtargetFeatures Features;
418     for (unsigned i = 0; i != MAttrs.size(); ++i)
419       Features.AddFeature(MAttrs[i]);
420     FeaturesStr = Features.getString();
421   }
422 
423   std::unique_ptr<ToolOutputFile> Out = GetOutputStream(OutputFilename);
424   if (!Out)
425     return 1;
426 
427   std::unique_ptr<ToolOutputFile> DwoOut;
428   if (!SplitDwarfFile.empty()) {
429     if (FileType != OFT_ObjectFile) {
430       WithColor::error() << "dwo output only supported with object files\n";
431       return 1;
432     }
433     DwoOut = GetOutputStream(SplitDwarfFile);
434     if (!DwoOut)
435       return 1;
436   }
437 
438   std::unique_ptr<buffer_ostream> BOS;
439   raw_pwrite_stream *OS = &Out->os();
440   std::unique_ptr<MCStreamer> Str;
441 
442   std::unique_ptr<MCInstrInfo> MCII(TheTarget->createMCInstrInfo());
443   std::unique_ptr<MCSubtargetInfo> STI(
444       TheTarget->createMCSubtargetInfo(TripleName, MCPU, FeaturesStr));
445 
446   MCInstPrinter *IP = nullptr;
447   if (FileType == OFT_AssemblyFile) {
448     IP = TheTarget->createMCInstPrinter(Triple(TripleName), OutputAsmVariant,
449                                         *MAI, *MCII, *MRI);
450 
451     if (!IP) {
452       WithColor::error()
453           << "unable to create instruction printer for target triple '"
454           << TheTriple.normalize() << "' with assembly variant "
455           << OutputAsmVariant << ".\n";
456       return 1;
457     }
458 
459     // Set the display preference for hex vs. decimal immediates.
460     IP->setPrintImmHex(PrintImmHex);
461 
462     // Set up the AsmStreamer.
463     std::unique_ptr<MCCodeEmitter> CE;
464     if (ShowEncoding)
465       CE.reset(TheTarget->createMCCodeEmitter(*MCII, *MRI, Ctx));
466 
467     std::unique_ptr<MCAsmBackend> MAB(
468         TheTarget->createMCAsmBackend(*STI, *MRI, MCOptions));
469     auto FOut = llvm::make_unique<formatted_raw_ostream>(*OS);
470     Str.reset(
471         TheTarget->createAsmStreamer(Ctx, std::move(FOut), /*asmverbose*/ true,
472                                      /*useDwarfDirectory*/ true, IP,
473                                      std::move(CE), std::move(MAB), ShowInst));
474 
475   } else if (FileType == OFT_Null) {
476     Str.reset(TheTarget->createNullStreamer(Ctx));
477   } else {
478     assert(FileType == OFT_ObjectFile && "Invalid file type!");
479 
480     // Don't waste memory on names of temp labels.
481     Ctx.setUseNamesOnTempLabels(false);
482 
483     if (!Out->os().supportsSeeking()) {
484       BOS = make_unique<buffer_ostream>(Out->os());
485       OS = BOS.get();
486     }
487 
488     MCCodeEmitter *CE = TheTarget->createMCCodeEmitter(*MCII, *MRI, Ctx);
489     MCAsmBackend *MAB = TheTarget->createMCAsmBackend(*STI, *MRI, MCOptions);
490     Str.reset(TheTarget->createMCObjectStreamer(
491         TheTriple, Ctx, std::unique_ptr<MCAsmBackend>(MAB),
492         DwoOut ? MAB->createDwoObjectWriter(*OS, DwoOut->os())
493                : MAB->createObjectWriter(*OS),
494         std::unique_ptr<MCCodeEmitter>(CE), *STI, MCOptions.MCRelaxAll,
495         MCOptions.MCIncrementalLinkerCompatible,
496         /*DWARFMustBeAtTheEnd*/ false));
497     if (NoExecStack)
498       Str->InitSections(true);
499   }
500 
501   // Use Assembler information for parsing.
502   Str->setUseAssemblerInfoForParsing(true);
503 
504   int Res = 1;
505   bool disassemble = false;
506   switch (Action) {
507   case AC_AsLex:
508     Res = AsLexInput(SrcMgr, *MAI, Out->os());
509     break;
510   case AC_Assemble:
511     Res = AssembleInput(ProgName, TheTarget, SrcMgr, Ctx, *Str, *MAI, *STI,
512                         *MCII, MCOptions);
513     break;
514   case AC_MDisassemble:
515     assert(IP && "Expected assembly output");
516     IP->setUseMarkup(1);
517     disassemble = true;
518     break;
519   case AC_Disassemble:
520     disassemble = true;
521     break;
522   }
523   if (disassemble)
524     Res = Disassembler::disassemble(*TheTarget, TripleName, *STI, *Str,
525                                     *Buffer, SrcMgr, Out->os());
526 
527   // Keep output if no errors.
528   if (Res == 0) {
529     Out->keep();
530     if (DwoOut)
531       DwoOut->keep();
532   }
533   return Res;
534 }
535