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/FileUtilities.h" 31 #include "llvm/Support/FormattedStream.h" 32 #include "llvm/Support/ManagedStatic.h" 33 #include "llvm/Support/MemoryBuffer.h" 34 #include "llvm/Support/PrettyStackTrace.h" 35 #include "llvm/Support/SourceMgr.h" 36 #include "llvm/Support/raw_ostream.h" 37 #include "llvm/System/Host.h" 38 #include "llvm/System/Signals.h" 39 #include "Disassembler.h" 40 using namespace llvm; 41 42 static cl::opt<std::string> 43 InputFilename(cl::Positional, cl::desc("<input file>"), cl::init("-")); 44 45 static cl::opt<std::string> 46 OutputFilename("o", cl::desc("Output filename"), 47 cl::value_desc("filename")); 48 49 static cl::opt<bool> 50 ShowEncoding("show-encoding", cl::desc("Show instruction encodings")); 51 52 static cl::opt<bool> 53 ShowInst("show-inst", cl::desc("Show internal instruction representation")); 54 55 static cl::opt<bool> 56 ShowInstOperands("show-inst-operands", 57 cl::desc("Show instructions operands as parsed")); 58 59 static cl::opt<unsigned> 60 OutputAsmVariant("output-asm-variant", 61 cl::desc("Syntax variant to use for output printing")); 62 63 static cl::opt<bool> 64 RelaxAll("mc-relax-all", cl::desc("Relax all fixups")); 65 66 static cl::opt<bool> 67 EnableLogging("enable-api-logging", cl::desc("Enable MC API logging")); 68 69 enum OutputFileType { 70 OFT_Null, 71 OFT_AssemblyFile, 72 OFT_ObjectFile 73 }; 74 static cl::opt<OutputFileType> 75 FileType("filetype", cl::init(OFT_AssemblyFile), 76 cl::desc("Choose an output file type:"), 77 cl::values( 78 clEnumValN(OFT_AssemblyFile, "asm", 79 "Emit an assembly ('.s') file"), 80 clEnumValN(OFT_Null, "null", 81 "Don't emit anything (for timing purposes)"), 82 clEnumValN(OFT_ObjectFile, "obj", 83 "Emit a native object ('.o') file"), 84 clEnumValEnd)); 85 86 static cl::list<std::string> 87 IncludeDirs("I", cl::desc("Directory of include files"), 88 cl::value_desc("directory"), cl::Prefix); 89 90 static cl::opt<std::string> 91 ArchName("arch", cl::desc("Target arch to assemble for, " 92 "see -version for available targets")); 93 94 static cl::opt<std::string> 95 TripleName("triple", cl::desc("Target triple to assemble for, " 96 "see -version for available targets")); 97 98 static cl::opt<bool> 99 NoInitialTextSection("n", cl::desc( 100 "Don't assume assembly file starts in the text section")); 101 102 enum ActionType { 103 AC_AsLex, 104 AC_Assemble, 105 AC_Disassemble, 106 AC_EDisassemble 107 }; 108 109 static cl::opt<ActionType> 110 Action(cl::desc("Action to perform:"), 111 cl::init(AC_Assemble), 112 cl::values(clEnumValN(AC_AsLex, "as-lex", 113 "Lex tokens from a .s file"), 114 clEnumValN(AC_Assemble, "assemble", 115 "Assemble a .s file (default)"), 116 clEnumValN(AC_Disassemble, "disassemble", 117 "Disassemble strings of hex bytes"), 118 clEnumValN(AC_EDisassemble, "edis", 119 "Enhanced disassembly of strings of hex bytes"), 120 clEnumValEnd)); 121 122 static const Target *GetTarget(const char *ProgName) { 123 // Figure out the target triple. 124 if (TripleName.empty()) 125 TripleName = sys::getHostTriple(); 126 if (!ArchName.empty()) { 127 llvm::Triple TT(TripleName); 128 TT.setArchName(ArchName); 129 TripleName = TT.str(); 130 } 131 132 // Get the target specific parser. 133 std::string Error; 134 const Target *TheTarget = TargetRegistry::lookupTarget(TripleName, Error); 135 if (TheTarget) 136 return TheTarget; 137 138 errs() << ProgName << ": error: unable to get target for '" << TripleName 139 << "', see --version and --triple.\n"; 140 return 0; 141 } 142 143 static tool_output_file *GetOutputStream() { 144 if (OutputFilename == "") 145 OutputFilename = "-"; 146 147 std::string Err; 148 tool_output_file *Out = new tool_output_file(OutputFilename.c_str(), Err, 149 raw_fd_ostream::F_Binary); 150 if (!Err.empty()) { 151 errs() << Err << '\n'; 152 delete Out; 153 return 0; 154 } 155 156 return Out; 157 } 158 159 static int AsLexInput(const char *ProgName) { 160 std::string ErrorMessage; 161 MemoryBuffer *Buffer = MemoryBuffer::getFileOrSTDIN(InputFilename, 162 &ErrorMessage); 163 if (Buffer == 0) { 164 errs() << ProgName << ": "; 165 if (ErrorMessage.size()) 166 errs() << ErrorMessage << "\n"; 167 else 168 errs() << "input file didn't read correctly.\n"; 169 return 1; 170 } 171 172 SourceMgr SrcMgr; 173 174 // Tell SrcMgr about this buffer, which is what TGParser will pick up. 175 SrcMgr.AddNewSourceBuffer(Buffer, SMLoc()); 176 177 // Record the location of the include directories so that the lexer can find 178 // it later. 179 SrcMgr.setIncludeDirs(IncludeDirs); 180 181 const Target *TheTarget = GetTarget(ProgName); 182 if (!TheTarget) 183 return 1; 184 185 llvm::OwningPtr<MCAsmInfo> MAI(TheTarget->createAsmInfo(TripleName)); 186 assert(MAI && "Unable to create target asm info!"); 187 188 AsmLexer Lexer(*MAI); 189 Lexer.setBuffer(SrcMgr.getMemoryBuffer(0)); 190 191 OwningPtr<tool_output_file> Out(GetOutputStream()); 192 if (!Out) 193 return 1; 194 195 bool Error = false; 196 while (Lexer.Lex().isNot(AsmToken::Eof)) { 197 switch (Lexer.getKind()) { 198 default: 199 SrcMgr.PrintMessage(Lexer.getLoc(), "unknown token", "warning"); 200 Error = true; 201 break; 202 case AsmToken::Error: 203 Error = true; // error already printed. 204 break; 205 case AsmToken::Identifier: 206 Out->os() << "identifier: " << Lexer.getTok().getString() << '\n'; 207 break; 208 case AsmToken::Integer: 209 Out->os() << "int: " << Lexer.getTok().getString() << '\n'; 210 break; 211 case AsmToken::Real: 212 Out->os() << "real: " << Lexer.getTok().getString() << '\n'; 213 break; 214 case AsmToken::Register: 215 Out->os() << "register: " << Lexer.getTok().getRegVal() << '\n'; 216 break; 217 case AsmToken::String: 218 Out->os() << "string: " << Lexer.getTok().getString() << '\n'; 219 break; 220 221 case AsmToken::Amp: Out->os() << "Amp\n"; break; 222 case AsmToken::AmpAmp: Out->os() << "AmpAmp\n"; break; 223 case AsmToken::At: Out->os() << "At\n"; break; 224 case AsmToken::Caret: Out->os() << "Caret\n"; break; 225 case AsmToken::Colon: Out->os() << "Colon\n"; break; 226 case AsmToken::Comma: Out->os() << "Comma\n"; break; 227 case AsmToken::Dollar: Out->os() << "Dollar\n"; break; 228 case AsmToken::Dot: Out->os() << "Dot\n"; break; 229 case AsmToken::EndOfStatement: Out->os() << "EndOfStatement\n"; break; 230 case AsmToken::Eof: Out->os() << "Eof\n"; break; 231 case AsmToken::Equal: Out->os() << "Equal\n"; break; 232 case AsmToken::EqualEqual: Out->os() << "EqualEqual\n"; break; 233 case AsmToken::Exclaim: Out->os() << "Exclaim\n"; break; 234 case AsmToken::ExclaimEqual: Out->os() << "ExclaimEqual\n"; break; 235 case AsmToken::Greater: Out->os() << "Greater\n"; break; 236 case AsmToken::GreaterEqual: Out->os() << "GreaterEqual\n"; break; 237 case AsmToken::GreaterGreater: Out->os() << "GreaterGreater\n"; break; 238 case AsmToken::Hash: Out->os() << "Hash\n"; break; 239 case AsmToken::LBrac: Out->os() << "LBrac\n"; break; 240 case AsmToken::LCurly: Out->os() << "LCurly\n"; break; 241 case AsmToken::LParen: Out->os() << "LParen\n"; break; 242 case AsmToken::Less: Out->os() << "Less\n"; break; 243 case AsmToken::LessEqual: Out->os() << "LessEqual\n"; break; 244 case AsmToken::LessGreater: Out->os() << "LessGreater\n"; break; 245 case AsmToken::LessLess: Out->os() << "LessLess\n"; break; 246 case AsmToken::Minus: Out->os() << "Minus\n"; break; 247 case AsmToken::Percent: Out->os() << "Percent\n"; break; 248 case AsmToken::Pipe: Out->os() << "Pipe\n"; break; 249 case AsmToken::PipePipe: Out->os() << "PipePipe\n"; break; 250 case AsmToken::Plus: Out->os() << "Plus\n"; break; 251 case AsmToken::RBrac: Out->os() << "RBrac\n"; break; 252 case AsmToken::RCurly: Out->os() << "RCurly\n"; break; 253 case AsmToken::RParen: Out->os() << "RParen\n"; break; 254 case AsmToken::Slash: Out->os() << "Slash\n"; break; 255 case AsmToken::Star: Out->os() << "Star\n"; break; 256 case AsmToken::Tilde: Out->os() << "Tilde\n"; break; 257 } 258 } 259 260 // Keep output if no errors. 261 if (Error == 0) Out->keep(); 262 263 return Error; 264 } 265 266 static int AssembleInput(const char *ProgName) { 267 const Target *TheTarget = GetTarget(ProgName); 268 if (!TheTarget) 269 return 1; 270 271 std::string Error; 272 MemoryBuffer *Buffer = MemoryBuffer::getFileOrSTDIN(InputFilename, &Error); 273 if (Buffer == 0) { 274 errs() << ProgName << ": "; 275 if (Error.size()) 276 errs() << Error << "\n"; 277 else 278 errs() << "input file didn't read correctly.\n"; 279 return 1; 280 } 281 282 SourceMgr SrcMgr; 283 284 // Tell SrcMgr about this buffer, which is what the parser will pick up. 285 SrcMgr.AddNewSourceBuffer(Buffer, SMLoc()); 286 287 // Record the location of the include directories so that the lexer can find 288 // it later. 289 SrcMgr.setIncludeDirs(IncludeDirs); 290 291 292 llvm::OwningPtr<MCAsmInfo> MAI(TheTarget->createAsmInfo(TripleName)); 293 assert(MAI && "Unable to create target asm info!"); 294 295 MCContext Ctx(*MAI); 296 297 // FIXME: We shouldn't need to do this (and link in codegen). 298 OwningPtr<TargetMachine> TM(TheTarget->createTargetMachine(TripleName, "")); 299 300 if (!TM) { 301 errs() << ProgName << ": error: could not create target for triple '" 302 << TripleName << "'.\n"; 303 return 1; 304 } 305 306 OwningPtr<tool_output_file> Out(GetOutputStream()); 307 if (!Out) 308 return 1; 309 310 formatted_raw_ostream FOS(Out->os()); 311 OwningPtr<MCStreamer> Str; 312 313 if (FileType == OFT_AssemblyFile) { 314 MCInstPrinter *IP = 315 TheTarget->createMCInstPrinter(OutputAsmVariant, *MAI); 316 MCCodeEmitter *CE = 0; 317 if (ShowEncoding) 318 CE = TheTarget->createCodeEmitter(*TM, Ctx); 319 Str.reset(createAsmStreamer(Ctx, FOS, 320 TM->getTargetData()->isLittleEndian(), 321 /*asmverbose*/true, IP, CE, ShowInst)); 322 } else if (FileType == OFT_Null) { 323 Str.reset(createNullStreamer(Ctx)); 324 } else { 325 assert(FileType == OFT_ObjectFile && "Invalid file type!"); 326 MCCodeEmitter *CE = TheTarget->createCodeEmitter(*TM, Ctx); 327 TargetAsmBackend *TAB = TheTarget->createAsmBackend(TripleName); 328 Str.reset(TheTarget->createObjectStreamer(TripleName, Ctx, *TAB, 329 FOS, CE, RelaxAll)); 330 } 331 332 if (EnableLogging) { 333 Str.reset(createLoggingStreamer(Str.take(), errs())); 334 } 335 336 OwningPtr<MCAsmParser> Parser(createMCAsmParser(*TheTarget, SrcMgr, Ctx, 337 *Str.get(), *MAI)); 338 OwningPtr<TargetAsmParser> TAP(TheTarget->createAsmParser(*Parser, *TM)); 339 if (!TAP) { 340 errs() << ProgName 341 << ": error: this target does not support assembly parsing.\n"; 342 return 1; 343 } 344 345 Parser->setShowParsedOperands(ShowInstOperands); 346 Parser->setTargetParser(*TAP.get()); 347 348 int Res = Parser->Run(NoInitialTextSection); 349 350 // Keep output if no errors. 351 if (Res == 0) Out->keep(); 352 353 return Res; 354 } 355 356 static int DisassembleInput(const char *ProgName, bool Enhanced) { 357 const Target *TheTarget = GetTarget(ProgName); 358 if (!TheTarget) 359 return 0; 360 361 std::string ErrorMessage; 362 363 MemoryBuffer *Buffer = MemoryBuffer::getFileOrSTDIN(InputFilename, 364 &ErrorMessage); 365 366 if (Buffer == 0) { 367 errs() << ProgName << ": "; 368 if (ErrorMessage.size()) 369 errs() << ErrorMessage << "\n"; 370 else 371 errs() << "input file didn't read correctly.\n"; 372 return 1; 373 } 374 375 OwningPtr<tool_output_file> Out(GetOutputStream()); 376 if (!Out) 377 return 1; 378 379 int Res; 380 if (Enhanced) 381 Res = Disassembler::disassembleEnhanced(TripleName, *Buffer, Out->os()); 382 else 383 Res = Disassembler::disassemble(*TheTarget, TripleName, *Buffer, Out->os()); 384 385 // Keep output if no errors. 386 if (Res == 0) Out->keep(); 387 388 return Res; 389 } 390 391 392 int main(int argc, char **argv) { 393 // Print a stack trace if we signal out. 394 sys::PrintStackTraceOnErrorSignal(); 395 PrettyStackTraceProgram X(argc, argv); 396 llvm_shutdown_obj Y; // Call llvm_shutdown() on exit. 397 398 // Initialize targets and assembly printers/parsers. 399 llvm::InitializeAllTargetInfos(); 400 // FIXME: We shouldn't need to initialize the Target(Machine)s. 401 llvm::InitializeAllTargets(); 402 llvm::InitializeAllAsmPrinters(); 403 llvm::InitializeAllAsmParsers(); 404 llvm::InitializeAllDisassemblers(); 405 406 cl::ParseCommandLineOptions(argc, argv, "llvm machine code playground\n"); 407 TripleName = Triple::normalize(TripleName); 408 409 switch (Action) { 410 default: 411 case AC_AsLex: 412 return AsLexInput(argv[0]); 413 case AC_Assemble: 414 return AssembleInput(argv[0]); 415 case AC_Disassemble: 416 return DisassembleInput(argv[0], false); 417 case AC_EDisassemble: 418 return DisassembleInput(argv[0], true); 419 } 420 421 return 0; 422 } 423 424