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 "Disassembler.h"
16 #include "llvm/MC/MCAsmBackend.h"
17 #include "llvm/MC/MCAsmInfo.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/MCParser/AsmLexer.h"
23 #include "llvm/MC/MCRegisterInfo.h"
24 #include "llvm/MC/MCSectionMachO.h"
25 #include "llvm/MC/MCStreamer.h"
26 #include "llvm/MC/MCSubtargetInfo.h"
27 #include "llvm/MC/MCTargetAsmParser.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/ManagedStatic.h"
35 #include "llvm/Support/MemoryBuffer.h"
36 #include "llvm/Support/PrettyStackTrace.h"
37 #include "llvm/Support/Signals.h"
38 #include "llvm/Support/SourceMgr.h"
39 #include "llvm/Support/TargetRegistry.h"
40 #include "llvm/Support/TargetSelect.h"
41 #include "llvm/Support/ToolOutputFile.h"
42 using namespace llvm;
43
44 static cl::opt<std::string>
45 InputFilename(cl::Positional, cl::desc("<input file>"), cl::init("-"));
46
47 static cl::opt<std::string>
48 OutputFilename("o", cl::desc("Output filename"),
49 cl::value_desc("filename"));
50
51 static cl::opt<bool>
52 ShowEncoding("show-encoding", cl::desc("Show instruction encodings"));
53
54 static cl::opt<bool>
55 CompressDebugSections("compress-debug-sections",
56 cl::desc("Compress DWARF debug sections"));
57
58 static cl::opt<bool>
59 ShowInst("show-inst", cl::desc("Show internal instruction representation"));
60
61 static cl::opt<bool>
62 ShowInstOperands("show-inst-operands",
63 cl::desc("Show instructions operands as parsed"));
64
65 static cl::opt<unsigned>
66 OutputAsmVariant("output-asm-variant",
67 cl::desc("Syntax variant to use for output printing"));
68
69 static cl::opt<bool>
70 PrintImmHex("print-imm-hex", cl::init(false),
71 cl::desc("Prefer hex format for immediate values"));
72
73 enum OutputFileType {
74 OFT_Null,
75 OFT_AssemblyFile,
76 OFT_ObjectFile
77 };
78 static cl::opt<OutputFileType>
79 FileType("filetype", cl::init(OFT_AssemblyFile),
80 cl::desc("Choose an output file type:"),
81 cl::values(
82 clEnumValN(OFT_AssemblyFile, "asm",
83 "Emit an assembly ('.s') file"),
84 clEnumValN(OFT_Null, "null",
85 "Don't emit anything (for timing purposes)"),
86 clEnumValN(OFT_ObjectFile, "obj",
87 "Emit a native object ('.o') file"),
88 clEnumValEnd));
89
90 static cl::list<std::string>
91 IncludeDirs("I", cl::desc("Directory of include files"),
92 cl::value_desc("directory"), cl::Prefix);
93
94 static cl::opt<std::string>
95 ArchName("arch", cl::desc("Target arch to assemble for, "
96 "see -version for available targets"));
97
98 static cl::opt<std::string>
99 TripleName("triple", cl::desc("Target triple to assemble for, "
100 "see -version for available targets"));
101
102 static cl::opt<std::string>
103 MCPU("mcpu",
104 cl::desc("Target a specific cpu type (-mcpu=help for details)"),
105 cl::value_desc("cpu-name"),
106 cl::init(""));
107
108 static cl::list<std::string>
109 MAttrs("mattr",
110 cl::CommaSeparated,
111 cl::desc("Target specific attributes (-mattr=help for details)"),
112 cl::value_desc("a1,+a2,-a3,..."));
113
114 static cl::opt<Reloc::Model>
115 RelocModel("relocation-model",
116 cl::desc("Choose relocation model"),
117 cl::init(Reloc::Default),
118 cl::values(
119 clEnumValN(Reloc::Default, "default",
120 "Target default relocation model"),
121 clEnumValN(Reloc::Static, "static",
122 "Non-relocatable code"),
123 clEnumValN(Reloc::PIC_, "pic",
124 "Fully relocatable, position independent code"),
125 clEnumValN(Reloc::DynamicNoPIC, "dynamic-no-pic",
126 "Relocatable external references, non-relocatable code"),
127 clEnumValEnd));
128
129 static cl::opt<llvm::CodeModel::Model>
130 CMModel("code-model",
131 cl::desc("Choose code model"),
132 cl::init(CodeModel::Default),
133 cl::values(clEnumValN(CodeModel::Default, "default",
134 "Target default code model"),
135 clEnumValN(CodeModel::Small, "small",
136 "Small code model"),
137 clEnumValN(CodeModel::Kernel, "kernel",
138 "Kernel code model"),
139 clEnumValN(CodeModel::Medium, "medium",
140 "Medium code model"),
141 clEnumValN(CodeModel::Large, "large",
142 "Large code model"),
143 clEnumValEnd));
144
145 static cl::opt<bool>
146 NoInitialTextSection("n", cl::desc("Don't assume assembly file starts "
147 "in the text section"));
148
149 static cl::opt<bool>
150 GenDwarfForAssembly("g", cl::desc("Generate dwarf debugging info for assembly "
151 "source files"));
152
153 static cl::opt<std::string>
154 DebugCompilationDir("fdebug-compilation-dir",
155 cl::desc("Specifies the debug info's compilation dir"));
156
157 static cl::opt<std::string>
158 MainFileName("main-file-name",
159 cl::desc("Specifies the name we should consider the input file"));
160
161 static cl::opt<bool> SaveTempLabels("save-temp-labels",
162 cl::desc("Don't discard temporary labels"));
163
164 static cl::opt<bool> NoExecStack("no-exec-stack",
165 cl::desc("File doesn't need an exec stack"));
166
167 enum ActionType {
168 AC_AsLex,
169 AC_Assemble,
170 AC_Disassemble,
171 AC_MDisassemble,
172 };
173
174 static cl::opt<ActionType>
175 Action(cl::desc("Action to perform:"),
176 cl::init(AC_Assemble),
177 cl::values(clEnumValN(AC_AsLex, "as-lex",
178 "Lex tokens from a .s file"),
179 clEnumValN(AC_Assemble, "assemble",
180 "Assemble a .s file (default)"),
181 clEnumValN(AC_Disassemble, "disassemble",
182 "Disassemble strings of hex bytes"),
183 clEnumValN(AC_MDisassemble, "mdis",
184 "Marked up disassembly of strings of hex bytes"),
185 clEnumValEnd));
186
GetTarget(const char * ProgName)187 static const Target *GetTarget(const char *ProgName) {
188 // Figure out the target triple.
189 if (TripleName.empty())
190 TripleName = sys::getDefaultTargetTriple();
191 Triple TheTriple(Triple::normalize(TripleName));
192
193 // Get the target specific parser.
194 std::string Error;
195 const Target *TheTarget = TargetRegistry::lookupTarget(ArchName, TheTriple,
196 Error);
197 if (!TheTarget) {
198 errs() << ProgName << ": " << Error;
199 return nullptr;
200 }
201
202 // Update the triple name and return the found target.
203 TripleName = TheTriple.getTriple();
204 return TheTarget;
205 }
206
GetOutputStream()207 static tool_output_file *GetOutputStream() {
208 if (OutputFilename == "")
209 OutputFilename = "-";
210
211 std::string Err;
212 tool_output_file *Out =
213 new tool_output_file(OutputFilename.c_str(), Err, sys::fs::F_None);
214 if (!Err.empty()) {
215 errs() << Err << '\n';
216 delete Out;
217 return nullptr;
218 }
219
220 return Out;
221 }
222
223 static std::string DwarfDebugFlags;
setDwarfDebugFlags(int argc,char ** argv)224 static void setDwarfDebugFlags(int argc, char **argv) {
225 if (!getenv("RC_DEBUG_OPTIONS"))
226 return;
227 for (int i = 0; i < argc; i++) {
228 DwarfDebugFlags += argv[i];
229 if (i + 1 < argc)
230 DwarfDebugFlags += " ";
231 }
232 }
233
234 static std::string DwarfDebugProducer;
setDwarfDebugProducer(void)235 static void setDwarfDebugProducer(void) {
236 if(!getenv("DEBUG_PRODUCER"))
237 return;
238 DwarfDebugProducer += getenv("DEBUG_PRODUCER");
239 }
240
AsLexInput(SourceMgr & SrcMgr,MCAsmInfo & MAI,tool_output_file * Out)241 static int AsLexInput(SourceMgr &SrcMgr, MCAsmInfo &MAI,
242 tool_output_file *Out) {
243
244 AsmLexer Lexer(MAI);
245 Lexer.setBuffer(SrcMgr.getMemoryBuffer(SrcMgr.getMainFileID())->getBuffer());
246
247 bool Error = false;
248 while (Lexer.Lex().isNot(AsmToken::Eof)) {
249 AsmToken Tok = Lexer.getTok();
250
251 switch (Tok.getKind()) {
252 default:
253 SrcMgr.PrintMessage(Lexer.getLoc(), SourceMgr::DK_Warning,
254 "unknown token");
255 Error = true;
256 break;
257 case AsmToken::Error:
258 Error = true; // error already printed.
259 break;
260 case AsmToken::Identifier:
261 Out->os() << "identifier: " << Lexer.getTok().getString();
262 break;
263 case AsmToken::Integer:
264 Out->os() << "int: " << Lexer.getTok().getString();
265 break;
266 case AsmToken::Real:
267 Out->os() << "real: " << Lexer.getTok().getString();
268 break;
269 case AsmToken::String:
270 Out->os() << "string: " << Lexer.getTok().getString();
271 break;
272
273 case AsmToken::Amp: Out->os() << "Amp"; break;
274 case AsmToken::AmpAmp: Out->os() << "AmpAmp"; break;
275 case AsmToken::At: Out->os() << "At"; break;
276 case AsmToken::Caret: Out->os() << "Caret"; break;
277 case AsmToken::Colon: Out->os() << "Colon"; break;
278 case AsmToken::Comma: Out->os() << "Comma"; break;
279 case AsmToken::Dollar: Out->os() << "Dollar"; break;
280 case AsmToken::Dot: Out->os() << "Dot"; break;
281 case AsmToken::EndOfStatement: Out->os() << "EndOfStatement"; break;
282 case AsmToken::Eof: Out->os() << "Eof"; break;
283 case AsmToken::Equal: Out->os() << "Equal"; break;
284 case AsmToken::EqualEqual: Out->os() << "EqualEqual"; break;
285 case AsmToken::Exclaim: Out->os() << "Exclaim"; break;
286 case AsmToken::ExclaimEqual: Out->os() << "ExclaimEqual"; break;
287 case AsmToken::Greater: Out->os() << "Greater"; break;
288 case AsmToken::GreaterEqual: Out->os() << "GreaterEqual"; break;
289 case AsmToken::GreaterGreater: Out->os() << "GreaterGreater"; break;
290 case AsmToken::Hash: Out->os() << "Hash"; break;
291 case AsmToken::LBrac: Out->os() << "LBrac"; break;
292 case AsmToken::LCurly: Out->os() << "LCurly"; break;
293 case AsmToken::LParen: Out->os() << "LParen"; break;
294 case AsmToken::Less: Out->os() << "Less"; break;
295 case AsmToken::LessEqual: Out->os() << "LessEqual"; break;
296 case AsmToken::LessGreater: Out->os() << "LessGreater"; break;
297 case AsmToken::LessLess: Out->os() << "LessLess"; break;
298 case AsmToken::Minus: Out->os() << "Minus"; break;
299 case AsmToken::Percent: Out->os() << "Percent"; break;
300 case AsmToken::Pipe: Out->os() << "Pipe"; break;
301 case AsmToken::PipePipe: Out->os() << "PipePipe"; break;
302 case AsmToken::Plus: Out->os() << "Plus"; break;
303 case AsmToken::RBrac: Out->os() << "RBrac"; break;
304 case AsmToken::RCurly: Out->os() << "RCurly"; break;
305 case AsmToken::RParen: Out->os() << "RParen"; break;
306 case AsmToken::Slash: Out->os() << "Slash"; break;
307 case AsmToken::Star: Out->os() << "Star"; break;
308 case AsmToken::Tilde: Out->os() << "Tilde"; break;
309 }
310
311 // Print the token string.
312 Out->os() << " (\"";
313 Out->os().write_escaped(Tok.getString());
314 Out->os() << "\")\n";
315 }
316
317 return Error;
318 }
319
AssembleInput(const char * ProgName,const Target * TheTarget,SourceMgr & SrcMgr,MCContext & Ctx,MCStreamer & Str,MCAsmInfo & MAI,MCSubtargetInfo & STI,MCInstrInfo & MCII,MCTargetOptions & MCOptions)320 static int AssembleInput(const char *ProgName, const Target *TheTarget,
321 SourceMgr &SrcMgr, MCContext &Ctx, MCStreamer &Str,
322 MCAsmInfo &MAI, MCSubtargetInfo &STI,
323 MCInstrInfo &MCII, MCTargetOptions &MCOptions) {
324 std::unique_ptr<MCAsmParser> Parser(
325 createMCAsmParser(SrcMgr, Ctx, Str, MAI));
326 std::unique_ptr<MCTargetAsmParser> TAP(
327 TheTarget->createMCAsmParser(STI, *Parser, MCII, MCOptions));
328
329 if (!TAP) {
330 errs() << ProgName
331 << ": error: this target does not support assembly parsing.\n";
332 return 1;
333 }
334
335 Parser->setShowParsedOperands(ShowInstOperands);
336 Parser->setTargetParser(*TAP.get());
337
338 int Res = Parser->Run(NoInitialTextSection);
339
340 return Res;
341 }
342
main(int argc,char ** argv)343 int main(int argc, char **argv) {
344 // Print a stack trace if we signal out.
345 sys::PrintStackTraceOnErrorSignal();
346 PrettyStackTraceProgram X(argc, argv);
347 llvm_shutdown_obj Y; // Call llvm_shutdown() on exit.
348
349 // Initialize targets and assembly printers/parsers.
350 llvm::InitializeAllTargetInfos();
351 llvm::InitializeAllTargetMCs();
352 llvm::InitializeAllAsmParsers();
353 llvm::InitializeAllDisassemblers();
354
355 // Register the target printer for --version.
356 cl::AddExtraVersionPrinter(TargetRegistry::printRegisteredTargetsForVersion);
357
358 cl::ParseCommandLineOptions(argc, argv, "llvm machine code playground\n");
359 MCTargetOptions MCOptions = InitMCTargetOptionsFromFlags();
360 TripleName = Triple::normalize(TripleName);
361 setDwarfDebugFlags(argc, argv);
362
363 setDwarfDebugProducer();
364
365 const char *ProgName = argv[0];
366 const Target *TheTarget = GetTarget(ProgName);
367 if (!TheTarget)
368 return 1;
369
370 ErrorOr<std::unique_ptr<MemoryBuffer>> BufferPtr =
371 MemoryBuffer::getFileOrSTDIN(InputFilename);
372 if (std::error_code EC = BufferPtr.getError()) {
373 errs() << ProgName << ": " << EC.message() << '\n';
374 return 1;
375 }
376 MemoryBuffer *Buffer = BufferPtr->release();
377
378 SourceMgr SrcMgr;
379
380 // Tell SrcMgr about this buffer, which is what the parser will pick up.
381 SrcMgr.AddNewSourceBuffer(Buffer, SMLoc());
382
383 // Record the location of the include directories so that the lexer can find
384 // it later.
385 SrcMgr.setIncludeDirs(IncludeDirs);
386
387 std::unique_ptr<MCRegisterInfo> MRI(TheTarget->createMCRegInfo(TripleName));
388 assert(MRI && "Unable to create target register info!");
389
390 std::unique_ptr<MCAsmInfo> MAI(TheTarget->createMCAsmInfo(*MRI, TripleName));
391 assert(MAI && "Unable to create target asm info!");
392
393 if (CompressDebugSections) {
394 if (!zlib::isAvailable()) {
395 errs() << ProgName
396 << ": build tools with zlib to enable -compress-debug-sections";
397 return 1;
398 }
399 MAI->setCompressDebugSections(true);
400 }
401
402 // FIXME: This is not pretty. MCContext has a ptr to MCObjectFileInfo and
403 // MCObjectFileInfo needs a MCContext reference in order to initialize itself.
404 MCObjectFileInfo MOFI;
405 MCContext Ctx(MAI.get(), MRI.get(), &MOFI, &SrcMgr);
406 MOFI.InitMCObjectFileInfo(TripleName, RelocModel, CMModel, Ctx);
407
408 if (SaveTempLabels)
409 Ctx.setAllowTemporaryLabels(false);
410
411 Ctx.setGenDwarfForAssembly(GenDwarfForAssembly);
412 // Default to 4 for dwarf version.
413 unsigned DwarfVersion = MCOptions.DwarfVersion ? MCOptions.DwarfVersion : 4;
414 if (DwarfVersion < 2 || DwarfVersion > 4) {
415 errs() << ProgName << ": Dwarf version " << DwarfVersion
416 << " is not supported." << '\n';
417 return 1;
418 }
419 Ctx.setDwarfVersion(DwarfVersion);
420 if (!DwarfDebugFlags.empty())
421 Ctx.setDwarfDebugFlags(StringRef(DwarfDebugFlags));
422 if (!DwarfDebugProducer.empty())
423 Ctx.setDwarfDebugProducer(StringRef(DwarfDebugProducer));
424 if (!DebugCompilationDir.empty())
425 Ctx.setCompilationDir(DebugCompilationDir);
426 if (!MainFileName.empty())
427 Ctx.setMainFileName(MainFileName);
428
429 // Package up features to be passed to target/subtarget
430 std::string FeaturesStr;
431 if (MAttrs.size()) {
432 SubtargetFeatures Features;
433 for (unsigned i = 0; i != MAttrs.size(); ++i)
434 Features.AddFeature(MAttrs[i]);
435 FeaturesStr = Features.getString();
436 }
437
438 std::unique_ptr<tool_output_file> Out(GetOutputStream());
439 if (!Out)
440 return 1;
441
442 formatted_raw_ostream FOS(Out->os());
443 std::unique_ptr<MCStreamer> Str;
444
445 std::unique_ptr<MCInstrInfo> MCII(TheTarget->createMCInstrInfo());
446 std::unique_ptr<MCSubtargetInfo> STI(
447 TheTarget->createMCSubtargetInfo(TripleName, MCPU, FeaturesStr));
448
449 MCInstPrinter *IP = nullptr;
450 if (FileType == OFT_AssemblyFile) {
451 IP =
452 TheTarget->createMCInstPrinter(OutputAsmVariant, *MAI, *MCII, *MRI, *STI);
453
454 // Set the display preference for hex vs. decimal immediates.
455 IP->setPrintImmHex(PrintImmHex);
456
457 // Set up the AsmStreamer.
458 MCCodeEmitter *CE = nullptr;
459 MCAsmBackend *MAB = nullptr;
460 if (ShowEncoding) {
461 CE = TheTarget->createMCCodeEmitter(*MCII, *MRI, *STI, Ctx);
462 MAB = TheTarget->createMCAsmBackend(*MRI, TripleName, MCPU);
463 }
464 Str.reset(TheTarget->createAsmStreamer(Ctx, FOS, /*asmverbose*/ true,
465 /*useDwarfDirectory*/ true, IP, CE,
466 MAB, ShowInst));
467
468 } else if (FileType == OFT_Null) {
469 Str.reset(createNullStreamer(Ctx));
470 } else {
471 assert(FileType == OFT_ObjectFile && "Invalid file type!");
472 MCCodeEmitter *CE = TheTarget->createMCCodeEmitter(*MCII, *MRI, *STI, Ctx);
473 MCAsmBackend *MAB = TheTarget->createMCAsmBackend(*MRI, TripleName, MCPU);
474 Str.reset(TheTarget->createMCObjectStreamer(TripleName, Ctx, *MAB,
475 FOS, CE, *STI, RelaxAll,
476 NoExecStack));
477 }
478
479 int Res = 1;
480 bool disassemble = false;
481 switch (Action) {
482 case AC_AsLex:
483 Res = AsLexInput(SrcMgr, *MAI, Out.get());
484 break;
485 case AC_Assemble:
486 Res = AssembleInput(ProgName, TheTarget, SrcMgr, Ctx, *Str, *MAI, *STI,
487 *MCII, MCOptions);
488 break;
489 case AC_MDisassemble:
490 assert(IP && "Expected assembly output");
491 IP->setUseMarkup(1);
492 disassemble = true;
493 break;
494 case AC_Disassemble:
495 disassemble = true;
496 break;
497 }
498 if (disassemble)
499 Res = Disassembler::disassemble(*TheTarget, TripleName, *STI, *Str,
500 *Buffer, SrcMgr, Out->os());
501
502 // Keep output if no errors.
503 if (Res == 0) Out->keep();
504 return Res;
505 }
506