1 //===-- llvm-mc.cpp - Machine Code Hacking Driver -------------------------===// 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 "llvm/MC/MCParser/AsmLexer.h" 16 #include "llvm/MC/MCParser/MCAsmLexer.h" 17 #include "llvm/MC/MCContext.h" 18 #include "llvm/MC/MCCodeEmitter.h" 19 #include "llvm/MC/MCInstPrinter.h" 20 #include "llvm/MC/MCSectionMachO.h" 21 #include "llvm/MC/MCStreamer.h" 22 #include "llvm/Target/TargetAsmBackend.h" 23 #include "llvm/Target/TargetAsmParser.h" 24 #include "llvm/Target/TargetData.h" 25 #include "llvm/Target/TargetRegistry.h" 26 #include "llvm/Target/TargetMachine.h" // FIXME. 27 #include "llvm/Target/TargetSelect.h" 28 #include "llvm/ADT/OwningPtr.h" 29 #include "llvm/Support/CommandLine.h" 30 #include "llvm/Support/FormattedStream.h" 31 #include "llvm/Support/ManagedStatic.h" 32 #include "llvm/Support/MemoryBuffer.h" 33 #include "llvm/Support/PrettyStackTrace.h" 34 #include "llvm/Support/SourceMgr.h" 35 #include "llvm/Support/raw_ostream.h" 36 #include "llvm/System/Host.h" 37 #include "llvm/System/Signals.h" 38 #include "Disassembler.h" 39 using namespace llvm; 40 41 static cl::opt<std::string> 42 InputFilename(cl::Positional, cl::desc("<input file>"), cl::init("-")); 43 44 static cl::opt<std::string> 45 OutputFilename("o", cl::desc("Output filename"), 46 cl::value_desc("filename")); 47 48 static cl::opt<bool> 49 ShowEncoding("show-encoding", cl::desc("Show instruction encodings")); 50 51 static cl::opt<bool> 52 ShowInst("show-inst", cl::desc("Show internal instruction representation")); 53 54 static cl::opt<bool> 55 ShowInstOperands("show-inst-operands", 56 cl::desc("Show instructions operands as parsed")); 57 58 static cl::opt<unsigned> 59 OutputAsmVariant("output-asm-variant", 60 cl::desc("Syntax variant to use for output printing")); 61 62 static cl::opt<bool> 63 RelaxAll("mc-relax-all", cl::desc("Relax all fixups")); 64 65 static cl::opt<bool> 66 EnableLogging("enable-api-logging", cl::desc("Enable MC API logging")); 67 68 enum OutputFileType { 69 OFT_Null, 70 OFT_AssemblyFile, 71 OFT_ObjectFile 72 }; 73 static cl::opt<OutputFileType> 74 FileType("filetype", cl::init(OFT_AssemblyFile), 75 cl::desc("Choose an output file type:"), 76 cl::values( 77 clEnumValN(OFT_AssemblyFile, "asm", 78 "Emit an assembly ('.s') file"), 79 clEnumValN(OFT_Null, "null", 80 "Don't emit anything (for timing purposes)"), 81 clEnumValN(OFT_ObjectFile, "obj", 82 "Emit a native object ('.o') file"), 83 clEnumValEnd)); 84 85 static cl::list<std::string> 86 IncludeDirs("I", cl::desc("Directory of include files"), 87 cl::value_desc("directory"), cl::Prefix); 88 89 static cl::opt<std::string> 90 ArchName("arch", cl::desc("Target arch to assemble for, " 91 "see -version for available targets")); 92 93 static cl::opt<std::string> 94 TripleName("triple", cl::desc("Target triple to assemble for, " 95 "see -version for available targets")); 96 97 static cl::opt<bool> 98 NoInitialTextSection("n", cl::desc( 99 "Don't assume assembly file starts in the text section")); 100 101 enum ActionType { 102 AC_AsLex, 103 AC_Assemble, 104 AC_Disassemble, 105 AC_EDisassemble 106 }; 107 108 static cl::opt<ActionType> 109 Action(cl::desc("Action to perform:"), 110 cl::init(AC_Assemble), 111 cl::values(clEnumValN(AC_AsLex, "as-lex", 112 "Lex tokens from a .s file"), 113 clEnumValN(AC_Assemble, "assemble", 114 "Assemble a .s file (default)"), 115 clEnumValN(AC_Disassemble, "disassemble", 116 "Disassemble strings of hex bytes"), 117 clEnumValN(AC_EDisassemble, "edis", 118 "Enhanced disassembly of strings of hex bytes"), 119 clEnumValEnd)); 120 121 static const Target *GetTarget(const char *ProgName) { 122 // Figure out the target triple. 123 if (TripleName.empty()) 124 TripleName = sys::getHostTriple(); 125 if (!ArchName.empty()) { 126 llvm::Triple TT(TripleName); 127 TT.setArchName(ArchName); 128 TripleName = TT.str(); 129 } 130 131 // Get the target specific parser. 132 std::string Error; 133 const Target *TheTarget = TargetRegistry::lookupTarget(TripleName, Error); 134 if (TheTarget) 135 return TheTarget; 136 137 errs() << ProgName << ": error: unable to get target for '" << TripleName 138 << "', see --version and --triple.\n"; 139 return 0; 140 } 141 142 static int AsLexInput(const char *ProgName) { 143 std::string ErrorMessage; 144 MemoryBuffer *Buffer = MemoryBuffer::getFileOrSTDIN(InputFilename, 145 &ErrorMessage); 146 if (Buffer == 0) { 147 errs() << ProgName << ": "; 148 if (ErrorMessage.size()) 149 errs() << ErrorMessage << "\n"; 150 else 151 errs() << "input file didn't read correctly.\n"; 152 return 1; 153 } 154 155 SourceMgr SrcMgr; 156 157 // Tell SrcMgr about this buffer, which is what TGParser will pick up. 158 SrcMgr.AddNewSourceBuffer(Buffer, SMLoc()); 159 160 // Record the location of the include directories so that the lexer can find 161 // it later. 162 SrcMgr.setIncludeDirs(IncludeDirs); 163 164 const Target *TheTarget = GetTarget(ProgName); 165 if (!TheTarget) 166 return 1; 167 168 llvm::OwningPtr<MCAsmInfo> MAI(TheTarget->createAsmInfo(TripleName)); 169 assert(MAI && "Unable to create target asm info!"); 170 171 AsmLexer Lexer(*MAI); 172 Lexer.setBuffer(SrcMgr.getMemoryBuffer(0)); 173 174 bool Error = false; 175 while (Lexer.Lex().isNot(AsmToken::Eof)) { 176 switch (Lexer.getKind()) { 177 default: 178 SrcMgr.PrintMessage(Lexer.getLoc(), "unknown token", "warning"); 179 Error = true; 180 break; 181 case AsmToken::Error: 182 Error = true; // error already printed. 183 break; 184 case AsmToken::Identifier: 185 outs() << "identifier: " << Lexer.getTok().getString() << '\n'; 186 break; 187 case AsmToken::String: 188 outs() << "string: " << Lexer.getTok().getString() << '\n'; 189 break; 190 case AsmToken::Integer: 191 outs() << "int: " << Lexer.getTok().getString() << '\n'; 192 break; 193 194 case AsmToken::Amp: outs() << "Amp\n"; break; 195 case AsmToken::AmpAmp: outs() << "AmpAmp\n"; break; 196 case AsmToken::Caret: outs() << "Caret\n"; break; 197 case AsmToken::Colon: outs() << "Colon\n"; break; 198 case AsmToken::Comma: outs() << "Comma\n"; break; 199 case AsmToken::Dollar: outs() << "Dollar\n"; break; 200 case AsmToken::EndOfStatement: outs() << "EndOfStatement\n"; break; 201 case AsmToken::Eof: outs() << "Eof\n"; break; 202 case AsmToken::Equal: outs() << "Equal\n"; break; 203 case AsmToken::EqualEqual: outs() << "EqualEqual\n"; break; 204 case AsmToken::Exclaim: outs() << "Exclaim\n"; break; 205 case AsmToken::ExclaimEqual: outs() << "ExclaimEqual\n"; break; 206 case AsmToken::Greater: outs() << "Greater\n"; break; 207 case AsmToken::GreaterEqual: outs() << "GreaterEqual\n"; break; 208 case AsmToken::GreaterGreater: outs() << "GreaterGreater\n"; break; 209 case AsmToken::LParen: outs() << "LParen\n"; break; 210 case AsmToken::Less: outs() << "Less\n"; break; 211 case AsmToken::LessEqual: outs() << "LessEqual\n"; break; 212 case AsmToken::LessGreater: outs() << "LessGreater\n"; break; 213 case AsmToken::LessLess: outs() << "LessLess\n"; break; 214 case AsmToken::Minus: outs() << "Minus\n"; break; 215 case AsmToken::Percent: outs() << "Percent\n"; break; 216 case AsmToken::Pipe: outs() << "Pipe\n"; break; 217 case AsmToken::PipePipe: outs() << "PipePipe\n"; break; 218 case AsmToken::Plus: outs() << "Plus\n"; break; 219 case AsmToken::RParen: outs() << "RParen\n"; break; 220 case AsmToken::Slash: outs() << "Slash\n"; break; 221 case AsmToken::Star: outs() << "Star\n"; break; 222 case AsmToken::Tilde: outs() << "Tilde\n"; break; 223 } 224 } 225 226 return Error; 227 } 228 229 static formatted_raw_ostream *GetOutputStream() { 230 if (OutputFilename == "") 231 OutputFilename = "-"; 232 233 // Make sure that the Out file gets unlinked from the disk if we get a 234 // SIGINT. 235 if (OutputFilename != "-") 236 sys::RemoveFileOnSignal(sys::Path(OutputFilename)); 237 238 std::string Err; 239 raw_fd_ostream *Out = new raw_fd_ostream(OutputFilename.c_str(), Err, 240 raw_fd_ostream::F_Binary); 241 if (!Err.empty()) { 242 errs() << Err << '\n'; 243 delete Out; 244 return 0; 245 } 246 247 return new formatted_raw_ostream(*Out, formatted_raw_ostream::DELETE_STREAM); 248 } 249 250 static int AssembleInput(const char *ProgName) { 251 const Target *TheTarget = GetTarget(ProgName); 252 if (!TheTarget) 253 return 1; 254 255 std::string Error; 256 MemoryBuffer *Buffer = MemoryBuffer::getFileOrSTDIN(InputFilename, &Error); 257 if (Buffer == 0) { 258 errs() << ProgName << ": "; 259 if (Error.size()) 260 errs() << Error << "\n"; 261 else 262 errs() << "input file didn't read correctly.\n"; 263 return 1; 264 } 265 266 SourceMgr SrcMgr; 267 268 // Tell SrcMgr about this buffer, which is what the parser will pick up. 269 SrcMgr.AddNewSourceBuffer(Buffer, SMLoc()); 270 271 // Record the location of the include directories so that the lexer can find 272 // it later. 273 SrcMgr.setIncludeDirs(IncludeDirs); 274 275 276 llvm::OwningPtr<MCAsmInfo> MAI(TheTarget->createAsmInfo(TripleName)); 277 assert(MAI && "Unable to create target asm info!"); 278 279 MCContext Ctx(*MAI); 280 formatted_raw_ostream *Out = GetOutputStream(); 281 if (!Out) 282 return 1; 283 284 285 // FIXME: We shouldn't need to do this (and link in codegen). 286 OwningPtr<TargetMachine> TM(TheTarget->createTargetMachine(TripleName, "")); 287 288 if (!TM) { 289 errs() << ProgName << ": error: could not create target for triple '" 290 << TripleName << "'.\n"; 291 return 1; 292 } 293 294 OwningPtr<MCStreamer> Str; 295 296 if (FileType == OFT_AssemblyFile) { 297 MCInstPrinter *IP = 298 TheTarget->createMCInstPrinter(OutputAsmVariant, *MAI); 299 MCCodeEmitter *CE = 0; 300 if (ShowEncoding) 301 CE = TheTarget->createCodeEmitter(*TM, Ctx); 302 Str.reset(createAsmStreamer(Ctx, *Out,TM->getTargetData()->isLittleEndian(), 303 /*asmverbose*/true, IP, CE, ShowInst)); 304 } else if (FileType == OFT_Null) { 305 Str.reset(createNullStreamer(Ctx)); 306 } else { 307 assert(FileType == OFT_ObjectFile && "Invalid file type!"); 308 MCCodeEmitter *CE = TheTarget->createCodeEmitter(*TM, Ctx); 309 TargetAsmBackend *TAB = TheTarget->createAsmBackend(TripleName); 310 Str.reset(TheTarget->createObjectStreamer(TripleName, Ctx, *TAB, 311 *Out, CE, RelaxAll)); 312 } 313 314 if (EnableLogging) { 315 Str.reset(createLoggingStreamer(Str.take(), errs())); 316 } 317 318 OwningPtr<MCAsmParser> Parser(createMCAsmParser(*TheTarget, SrcMgr, Ctx, 319 *Str.get(), *MAI)); 320 OwningPtr<TargetAsmParser> TAP(TheTarget->createAsmParser(*Parser, *TM)); 321 if (!TAP) { 322 errs() << ProgName 323 << ": error: this target does not support assembly parsing.\n"; 324 return 1; 325 } 326 327 Parser->setShowParsedOperands(ShowInstOperands); 328 Parser->setTargetParser(*TAP.get()); 329 330 int Res = Parser->Run(NoInitialTextSection); 331 delete Out; 332 333 // Delete output on errors. 334 if (Res && OutputFilename != "-") 335 sys::Path(OutputFilename).eraseFromDisk(); 336 337 return Res; 338 } 339 340 static int DisassembleInput(const char *ProgName, bool Enhanced) { 341 const Target *TheTarget = GetTarget(ProgName); 342 if (!TheTarget) 343 return 0; 344 345 std::string ErrorMessage; 346 347 MemoryBuffer *Buffer = MemoryBuffer::getFileOrSTDIN(InputFilename, 348 &ErrorMessage); 349 350 if (Buffer == 0) { 351 errs() << ProgName << ": "; 352 if (ErrorMessage.size()) 353 errs() << ErrorMessage << "\n"; 354 else 355 errs() << "input file didn't read correctly.\n"; 356 return 1; 357 } 358 359 if (Enhanced) 360 return Disassembler::disassembleEnhanced(TripleName, *Buffer); 361 else 362 return Disassembler::disassemble(*TheTarget, TripleName, *Buffer); 363 } 364 365 366 int main(int argc, char **argv) { 367 // Print a stack trace if we signal out. 368 sys::PrintStackTraceOnErrorSignal(); 369 PrettyStackTraceProgram X(argc, argv); 370 llvm_shutdown_obj Y; // Call llvm_shutdown() on exit. 371 372 // Initialize targets and assembly printers/parsers. 373 llvm::InitializeAllTargetInfos(); 374 // FIXME: We shouldn't need to initialize the Target(Machine)s. 375 llvm::InitializeAllTargets(); 376 llvm::InitializeAllAsmPrinters(); 377 llvm::InitializeAllAsmParsers(); 378 llvm::InitializeAllDisassemblers(); 379 380 cl::ParseCommandLineOptions(argc, argv, "llvm machine code playground\n"); 381 382 switch (Action) { 383 default: 384 case AC_AsLex: 385 return AsLexInput(argv[0]); 386 case AC_Assemble: 387 return AssembleInput(argv[0]); 388 case AC_Disassemble: 389 return DisassembleInput(argv[0], false); 390 case AC_EDisassemble: 391 return DisassembleInput(argv[0], true); 392 } 393 394 return 0; 395 } 396 397