1 //===- MCAssembler.h - Object File Generation -------------------*- 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 #ifndef LLVM_MC_MCASSEMBLER_H 11 #define LLVM_MC_MCASSEMBLER_H 12 13 #include "llvm/ADT/ArrayRef.h" 14 #include "llvm/ADT/STLExtras.h" 15 #include "llvm/ADT/SmallPtrSet.h" 16 #include "llvm/ADT/StringRef.h" 17 #include "llvm/ADT/iterator.h" 18 #include "llvm/ADT/iterator_range.h" 19 #include "llvm/BinaryFormat/MachO.h" 20 #include "llvm/MC/MCDirectives.h" 21 #include "llvm/MC/MCDwarf.h" 22 #include "llvm/MC/MCFixup.h" 23 #include "llvm/MC/MCFragment.h" 24 #include "llvm/MC/MCLinkerOptimizationHint.h" 25 #include "llvm/MC/MCSymbol.h" 26 #include <cassert> 27 #include <cstddef> 28 #include <cstdint> 29 #include <string> 30 #include <utility> 31 #include <vector> 32 33 namespace llvm { 34 35 class MCAsmBackend; 36 class MCAsmLayout; 37 class MCContext; 38 class MCCodeEmitter; 39 class MCFragment; 40 class MCObjectWriter; 41 class MCSection; 42 class MCValue; 43 44 // FIXME: This really doesn't belong here. See comments below. 45 struct IndirectSymbolData { 46 MCSymbol *Symbol; 47 MCSection *Section; 48 }; 49 50 // FIXME: Ditto this. Purely so the Streamer and the ObjectWriter can talk 51 // to one another. 52 struct DataRegionData { 53 // This enum should be kept in sync w/ the mach-o definition in 54 // llvm/Object/MachOFormat.h. 55 enum KindTy { Data = 1, JumpTable8, JumpTable16, JumpTable32 } Kind; 56 MCSymbol *Start; 57 MCSymbol *End; 58 }; 59 60 class MCAssembler { 61 friend class MCAsmLayout; 62 63 public: 64 using SectionListType = std::vector<MCSection *>; 65 using SymbolDataListType = std::vector<const MCSymbol *>; 66 67 using const_iterator = pointee_iterator<SectionListType::const_iterator>; 68 using iterator = pointee_iterator<SectionListType::iterator>; 69 70 using const_symbol_iterator = 71 pointee_iterator<SymbolDataListType::const_iterator>; 72 using symbol_iterator = pointee_iterator<SymbolDataListType::iterator>; 73 74 using symbol_range = iterator_range<symbol_iterator>; 75 using const_symbol_range = iterator_range<const_symbol_iterator>; 76 77 using const_indirect_symbol_iterator = 78 std::vector<IndirectSymbolData>::const_iterator; 79 using indirect_symbol_iterator = std::vector<IndirectSymbolData>::iterator; 80 81 using const_data_region_iterator = 82 std::vector<DataRegionData>::const_iterator; 83 using data_region_iterator = std::vector<DataRegionData>::iterator; 84 85 /// MachO specific deployment target version info. 86 // A Major version of 0 indicates that no version information was supplied 87 // and so the corresponding load command should not be emitted. 88 using VersionInfoType = struct { 89 bool EmitBuildVersion; 90 union { 91 MCVersionMinType Type; ///< Used when EmitBuildVersion==false. 92 MachO::PlatformType Platform; ///< Used when EmitBuildVersion==true. 93 } TypeOrPlatform; 94 unsigned Major; 95 unsigned Minor; 96 unsigned Update; 97 }; 98 99 private: 100 MCContext &Context; 101 102 std::unique_ptr<MCAsmBackend> Backend; 103 104 std::unique_ptr<MCCodeEmitter> Emitter; 105 106 std::unique_ptr<MCObjectWriter> Writer; 107 108 SectionListType Sections; 109 110 SymbolDataListType Symbols; 111 112 std::vector<IndirectSymbolData> IndirectSymbols; 113 114 std::vector<DataRegionData> DataRegions; 115 116 /// The list of linker options to propagate into the object file. 117 std::vector<std::vector<std::string>> LinkerOptions; 118 119 /// List of declared file names 120 std::vector<std::string> FileNames; 121 122 MCDwarfLineTableParams LTParams; 123 124 /// The set of function symbols for which a .thumb_func directive has 125 /// been seen. 126 // 127 // FIXME: We really would like this in target specific code rather than 128 // here. Maybe when the relocation stuff moves to target specific, 129 // this can go with it? The streamer would need some target specific 130 // refactoring too. 131 mutable SmallPtrSet<const MCSymbol *, 32> ThumbFuncs; 132 133 /// The bundle alignment size currently set in the assembler. 134 /// 135 /// By default it's 0, which means bundling is disabled. 136 unsigned BundleAlignSize; 137 138 bool RelaxAll : 1; 139 bool SubsectionsViaSymbols : 1; 140 bool IncrementalLinkerCompatible : 1; 141 142 /// ELF specific e_header flags 143 // It would be good if there were an MCELFAssembler class to hold this. 144 // ELF header flags are used both by the integrated and standalone assemblers. 145 // Access to the flags is necessary in cases where assembler directives affect 146 // which flags to be set. 147 unsigned ELFHeaderEFlags; 148 149 /// Used to communicate Linker Optimization Hint information between 150 /// the Streamer and the .o writer 151 MCLOHContainer LOHContainer; 152 153 VersionInfoType VersionInfo; 154 155 /// Evaluate a fixup to a relocatable expression and the value which should be 156 /// placed into the fixup. 157 /// 158 /// \param Layout The layout to use for evaluation. 159 /// \param Fixup The fixup to evaluate. 160 /// \param DF The fragment the fixup is inside. 161 /// \param Target [out] On return, the relocatable expression the fixup 162 /// evaluates to. 163 /// \param Value [out] On return, the value of the fixup as currently laid 164 /// out. 165 /// \param WasForced [out] On return, the value in the fixup is set to the 166 /// correct value if WasForced is true, even if evaluateFixup returns false. 167 /// \return Whether the fixup value was fully resolved. This is true if the 168 /// \p Value result is fixed, otherwise the value may change due to 169 /// relocation. 170 bool evaluateFixup(const MCAsmLayout &Layout, const MCFixup &Fixup, 171 const MCFragment *DF, MCValue &Target, 172 uint64_t &Value, bool &WasForced) const; 173 174 /// Check whether a fixup can be satisfied, or whether it needs to be relaxed 175 /// (increased in size, in order to hold its value correctly). 176 bool fixupNeedsRelaxation(const MCFixup &Fixup, const MCRelaxableFragment *DF, 177 const MCAsmLayout &Layout) const; 178 179 /// Check whether the given fragment needs relaxation. 180 bool fragmentNeedsRelaxation(const MCRelaxableFragment *IF, 181 const MCAsmLayout &Layout) const; 182 183 /// Perform one layout iteration and return true if any offsets 184 /// were adjusted. 185 bool layoutOnce(MCAsmLayout &Layout); 186 187 /// Perform one layout iteration of the given section and return true 188 /// if any offsets were adjusted. 189 bool layoutSectionOnce(MCAsmLayout &Layout, MCSection &Sec); 190 191 bool relaxInstruction(MCAsmLayout &Layout, MCRelaxableFragment &IF); 192 193 bool relaxPaddingFragment(MCAsmLayout &Layout, MCPaddingFragment &PF); 194 195 bool relaxLEB(MCAsmLayout &Layout, MCLEBFragment &IF); 196 197 bool relaxDwarfLineAddr(MCAsmLayout &Layout, MCDwarfLineAddrFragment &DF); 198 bool relaxDwarfCallFrameFragment(MCAsmLayout &Layout, 199 MCDwarfCallFrameFragment &DF); 200 bool relaxCVInlineLineTable(MCAsmLayout &Layout, 201 MCCVInlineLineTableFragment &DF); 202 bool relaxCVDefRange(MCAsmLayout &Layout, MCCVDefRangeFragment &DF); 203 204 /// finishLayout - Finalize a layout, including fragment lowering. 205 void finishLayout(MCAsmLayout &Layout); 206 207 std::tuple<MCValue, uint64_t, bool> 208 handleFixup(const MCAsmLayout &Layout, MCFragment &F, const MCFixup &Fixup); 209 210 public: 211 std::vector<std::pair<StringRef, const MCSymbol *>> Symvers; 212 213 /// Construct a new assembler instance. 214 // 215 // FIXME: How are we going to parameterize this? Two obvious options are stay 216 // concrete and require clients to pass in a target like object. The other 217 // option is to make this abstract, and have targets provide concrete 218 // implementations as we do with AsmParser. 219 MCAssembler(MCContext &Context, std::unique_ptr<MCAsmBackend> Backend, 220 std::unique_ptr<MCCodeEmitter> Emitter, 221 std::unique_ptr<MCObjectWriter> Writer); 222 MCAssembler(const MCAssembler &) = delete; 223 MCAssembler &operator=(const MCAssembler &) = delete; 224 ~MCAssembler(); 225 226 /// Compute the effective fragment size assuming it is laid out at the given 227 /// \p SectionAddress and \p FragmentOffset. 228 uint64_t computeFragmentSize(const MCAsmLayout &Layout, 229 const MCFragment &F) const; 230 231 /// Find the symbol which defines the atom containing the given symbol, or 232 /// null if there is no such symbol. 233 const MCSymbol *getAtom(const MCSymbol &S) const; 234 235 /// Check whether a particular symbol is visible to the linker and is required 236 /// in the symbol table, or whether it can be discarded by the assembler. This 237 /// also effects whether the assembler treats the label as potentially 238 /// defining a separate atom. 239 bool isSymbolLinkerVisible(const MCSymbol &SD) const; 240 241 /// Emit the section contents to \p OS. 242 void writeSectionData(raw_ostream &OS, const MCSection *Section, 243 const MCAsmLayout &Layout) const; 244 245 /// Check whether a given symbol has been flagged with .thumb_func. 246 bool isThumbFunc(const MCSymbol *Func) const; 247 248 /// Flag a function symbol as the target of a .thumb_func directive. 249 void setIsThumbFunc(const MCSymbol *Func) { ThumbFuncs.insert(Func); } 250 251 /// ELF e_header flags 252 unsigned getELFHeaderEFlags() const { return ELFHeaderEFlags; } 253 void setELFHeaderEFlags(unsigned Flags) { ELFHeaderEFlags = Flags; } 254 255 /// MachO deployment target version information. 256 const VersionInfoType &getVersionInfo() const { return VersionInfo; } 257 void setVersionMin(MCVersionMinType Type, unsigned Major, unsigned Minor, 258 unsigned Update) { 259 VersionInfo.EmitBuildVersion = false; 260 VersionInfo.TypeOrPlatform.Type = Type; 261 VersionInfo.Major = Major; 262 VersionInfo.Minor = Minor; 263 VersionInfo.Update = Update; 264 } 265 void setBuildVersion(MachO::PlatformType Platform, unsigned Major, 266 unsigned Minor, unsigned Update) { 267 VersionInfo.EmitBuildVersion = true; 268 VersionInfo.TypeOrPlatform.Platform = Platform; 269 VersionInfo.Major = Major; 270 VersionInfo.Minor = Minor; 271 VersionInfo.Update = Update; 272 } 273 274 /// Reuse an assembler instance 275 /// 276 void reset(); 277 278 MCContext &getContext() const { return Context; } 279 280 MCAsmBackend *getBackendPtr() const { return Backend.get(); } 281 282 MCCodeEmitter *getEmitterPtr() const { return Emitter.get(); } 283 284 MCObjectWriter *getWriterPtr() const { return Writer.get(); } 285 286 MCAsmBackend &getBackend() const { return *Backend; } 287 288 MCCodeEmitter &getEmitter() const { return *Emitter; } 289 290 MCObjectWriter &getWriter() const { return *Writer; } 291 292 MCDwarfLineTableParams getDWARFLinetableParams() const { return LTParams; } 293 void setDWARFLinetableParams(MCDwarfLineTableParams P) { LTParams = P; } 294 295 /// Finish - Do final processing and write the object to the output stream. 296 /// \p Writer is used for custom object writer (as the MCJIT does), 297 /// if not specified it is automatically created from backend. 298 void Finish(); 299 300 // Layout all section and prepare them for emission. 301 void layout(MCAsmLayout &Layout); 302 303 // FIXME: This does not belong here. 304 bool getSubsectionsViaSymbols() const { return SubsectionsViaSymbols; } 305 void setSubsectionsViaSymbols(bool Value) { SubsectionsViaSymbols = Value; } 306 307 bool isIncrementalLinkerCompatible() const { 308 return IncrementalLinkerCompatible; 309 } 310 void setIncrementalLinkerCompatible(bool Value) { 311 IncrementalLinkerCompatible = Value; 312 } 313 314 bool getRelaxAll() const { return RelaxAll; } 315 void setRelaxAll(bool Value) { RelaxAll = Value; } 316 317 bool isBundlingEnabled() const { return BundleAlignSize != 0; } 318 319 unsigned getBundleAlignSize() const { return BundleAlignSize; } 320 321 void setBundleAlignSize(unsigned Size) { 322 assert((Size == 0 || !(Size & (Size - 1))) && 323 "Expect a power-of-two bundle align size"); 324 BundleAlignSize = Size; 325 } 326 327 /// \name Section List Access 328 /// @{ 329 330 iterator begin() { return Sections.begin(); } 331 const_iterator begin() const { return Sections.begin(); } 332 333 iterator end() { return Sections.end(); } 334 const_iterator end() const { return Sections.end(); } 335 336 size_t size() const { return Sections.size(); } 337 338 /// @} 339 /// \name Symbol List Access 340 /// @{ 341 symbol_iterator symbol_begin() { return Symbols.begin(); } 342 const_symbol_iterator symbol_begin() const { return Symbols.begin(); } 343 344 symbol_iterator symbol_end() { return Symbols.end(); } 345 const_symbol_iterator symbol_end() const { return Symbols.end(); } 346 347 symbol_range symbols() { return make_range(symbol_begin(), symbol_end()); } 348 const_symbol_range symbols() const { 349 return make_range(symbol_begin(), symbol_end()); 350 } 351 352 size_t symbol_size() const { return Symbols.size(); } 353 354 /// @} 355 /// \name Indirect Symbol List Access 356 /// @{ 357 358 // FIXME: This is a total hack, this should not be here. Once things are 359 // factored so that the streamer has direct access to the .o writer, it can 360 // disappear. 361 std::vector<IndirectSymbolData> &getIndirectSymbols() { 362 return IndirectSymbols; 363 } 364 365 indirect_symbol_iterator indirect_symbol_begin() { 366 return IndirectSymbols.begin(); 367 } 368 const_indirect_symbol_iterator indirect_symbol_begin() const { 369 return IndirectSymbols.begin(); 370 } 371 372 indirect_symbol_iterator indirect_symbol_end() { 373 return IndirectSymbols.end(); 374 } 375 const_indirect_symbol_iterator indirect_symbol_end() const { 376 return IndirectSymbols.end(); 377 } 378 379 size_t indirect_symbol_size() const { return IndirectSymbols.size(); } 380 381 /// @} 382 /// \name Linker Option List Access 383 /// @{ 384 385 std::vector<std::vector<std::string>> &getLinkerOptions() { 386 return LinkerOptions; 387 } 388 389 /// @} 390 /// \name Data Region List Access 391 /// @{ 392 393 // FIXME: This is a total hack, this should not be here. Once things are 394 // factored so that the streamer has direct access to the .o writer, it can 395 // disappear. 396 std::vector<DataRegionData> &getDataRegions() { return DataRegions; } 397 398 data_region_iterator data_region_begin() { return DataRegions.begin(); } 399 const_data_region_iterator data_region_begin() const { 400 return DataRegions.begin(); 401 } 402 403 data_region_iterator data_region_end() { return DataRegions.end(); } 404 const_data_region_iterator data_region_end() const { 405 return DataRegions.end(); 406 } 407 408 size_t data_region_size() const { return DataRegions.size(); } 409 410 /// @} 411 /// \name Data Region List Access 412 /// @{ 413 414 // FIXME: This is a total hack, this should not be here. Once things are 415 // factored so that the streamer has direct access to the .o writer, it can 416 // disappear. 417 MCLOHContainer &getLOHContainer() { return LOHContainer; } 418 const MCLOHContainer &getLOHContainer() const { 419 return const_cast<MCAssembler *>(this)->getLOHContainer(); 420 } 421 422 struct CGProfileEntry { 423 const MCSymbolRefExpr *From; 424 const MCSymbolRefExpr *To; 425 uint64_t Count; 426 }; 427 std::vector<CGProfileEntry> CGProfile; 428 /// @} 429 /// \name Backend Data Access 430 /// @{ 431 432 bool registerSection(MCSection &Section); 433 434 void registerSymbol(const MCSymbol &Symbol, bool *Created = nullptr); 435 436 ArrayRef<std::string> getFileNames() { return FileNames; } 437 438 void addFileName(StringRef FileName) { 439 if (!is_contained(FileNames, FileName)) 440 FileNames.push_back(FileName); 441 } 442 443 /// Write the necessary bundle padding to \p OS. 444 /// Expects a fragment \p F containing instructions and its size \p FSize. 445 void writeFragmentPadding(raw_ostream &OS, const MCEncodedFragment &F, 446 uint64_t FSize) const; 447 448 /// @} 449 450 void dump() const; 451 }; 452 453 /// Compute the amount of padding required before the fragment \p F to 454 /// obey bundling restrictions, where \p FOffset is the fragment's offset in 455 /// its section and \p FSize is the fragment's size. 456 uint64_t computeBundlePadding(const MCAssembler &Assembler, 457 const MCEncodedFragment *F, uint64_t FOffset, 458 uint64_t FSize); 459 460 } // end namespace llvm 461 462 #endif // LLVM_MC_MCASSEMBLER_H 463
