lld/COFF/InputFiles.h

   1 //===- InputFiles.h ---------------------------------------------*- C++ -*-===//
   2 //
   3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
   4 // See https://llvm.org/LICENSE.txt for license information.
   5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
   6 //
   7 //===----------------------------------------------------------------------===//
   8
   9 #ifndef LLD_COFF_INPUT_FILES_H
  10 #define LLD_COFF_INPUT_FILES_H
  11
  12 #include "Config.h"
  13 #include "lld/Common/LLVM.h"
  14 #include "llvm/ADT/ArrayRef.h"
  15 #include "llvm/ADT/DenseMap.h"
  16 #include "llvm/ADT/DenseSet.h"
  17 #include "llvm/ADT/StringSet.h"
  18 #include "llvm/BinaryFormat/Magic.h"
  19 #include "llvm/Object/Archive.h"
  20 #include "llvm/Object/COFF.h"
  21 #include "llvm/Support/StringSaver.h"
  22 #include <memory>
  23 #include <set>
  24 #include <vector>
  25
  26 namespace llvm {
  27 struct DILineInfo;
  28 namespace pdb {
  29 class DbiModuleDescriptorBuilder;
  30 class NativeSession;
  31 }
  32 namespace lto {
  33 class InputFile;
  34 }
  35 }
  36
  37 namespace lld {
  38 class DWARFCache;
  39
  40 namespace coff {
  41 class COFFLinkerContext;
  42
  43 std::vector<MemoryBufferRef> getArchiveMembers(llvm::object::Archive *file);
  44
  45 using llvm::COFF::IMAGE_FILE_MACHINE_UNKNOWN;
  46 using llvm::COFF::MachineTypes;
  47 using llvm::object::Archive;
  48 using llvm::object::COFFObjectFile;
  49 using llvm::object::COFFSymbolRef;
  50 using llvm::object::coff_import_header;
  51 using llvm::object::coff_section;
  52
  53 class Chunk;
  54 class Defined;
  55 class DefinedImportData;
  56 class DefinedImportThunk;
  57 class DefinedRegular;
  58 class ImportThunkChunk;
  59 class ImportThunkChunkARM64EC;
  60 class SectionChunk;
  61 class Symbol;
  62 class Undefined;
  63 class TpiSource;
  64
  65 // The root class of input files.
  66 class InputFile {
  67 public:
  68   enum Kind {
  69     ArchiveKind,
  70     ObjectKind,
  71     LazyObjectKind,
  72     PDBKind,
  73     ImportKind,
  74     BitcodeKind,
  75     DLLKind
  76   };
  77   Kind kind() const { return fileKind; }
  78   virtual ~InputFile() {}
  79
  80   // Returns the filename.
  81   StringRef getName() const { return mb.getBufferIdentifier(); }
  82
  83   // Reads a file (the constructor doesn't do that).
  84   virtual void parse() = 0;
  85
  86   // Returns the CPU type this file was compiled to.
  87   virtual MachineTypes getMachineType() const {
  88     return IMAGE_FILE_MACHINE_UNKNOWN;
  89   }
  90
  91   MemoryBufferRef mb;
  92
  93   // An archive file name if this file is created from an archive.
  94   StringRef parentName;
  95
  96   // Returns .drectve section contents if exist.
  97   StringRef getDirectives() { return directives; }
  98
  99   COFFLinkerContext &ctx;
 100
 101 protected:
 102   InputFile(COFFLinkerContext &c, Kind k, MemoryBufferRef m, bool lazy = false)
 103       : mb(m), ctx(c), fileKind(k), lazy(lazy) {}
 104
 105   StringRef directives;
 106
 107 private:
 108   const Kind fileKind;
 109
 110 public:
 111   // True if this is a lazy ObjFile or BitcodeFile.
 112   bool lazy = false;
 113 };
 114
 115 // .lib or .a file.
 116 class ArchiveFile : public InputFile {
 117 public:
 118   explicit ArchiveFile(COFFLinkerContext &ctx, MemoryBufferRef m);
 119   static bool classof(const InputFile *f) { return f->kind() == ArchiveKind; }
 120   void parse() override;
 121
 122   // Enqueues an archive member load for the given symbol. If we've already
 123   // enqueued a load for the same archive member, this function does nothing,
 124   // which ensures that we don't load the same member more than once.
 125   void addMember(const Archive::Symbol &sym);
 126
 127 private:
 128   std::unique_ptr<Archive> file;
 129   llvm::DenseSet<uint64_t> seen;
 130 };
 131
 132 // .obj or .o file. This may be a member of an archive file.
 133 class ObjFile : public InputFile {
 134 public:
 135   explicit ObjFile(COFFLinkerContext &ctx, MemoryBufferRef m, bool lazy = false)
 136       : InputFile(ctx, ObjectKind, m, lazy) {}
 137   static bool classof(const InputFile *f) { return f->kind() == ObjectKind; }
 138   void parse() override;
 139   void parseLazy();
 140   MachineTypes getMachineType() const override;
 141   ArrayRef<Chunk *> getChunks() { return chunks; }
 142   ArrayRef<SectionChunk *> getDebugChunks() { return debugChunks; }
 143   ArrayRef<SectionChunk *> getSXDataChunks() { return sxDataChunks; }
 144   ArrayRef<SectionChunk *> getGuardFidChunks() { return guardFidChunks; }
 145   ArrayRef<SectionChunk *> getGuardIATChunks() { return guardIATChunks; }
 146   ArrayRef<SectionChunk *> getGuardLJmpChunks() { return guardLJmpChunks; }
 147   ArrayRef<SectionChunk *> getGuardEHContChunks() { return guardEHContChunks; }
 148   ArrayRef<Symbol *> getSymbols() { return symbols; }
 149
 150   MutableArrayRef<Symbol *> getMutableSymbols() { return symbols; }
 151
 152   ArrayRef<uint8_t> getDebugSection(StringRef secName);
 153
 154   // Returns a Symbol object for the symbolIndex'th symbol in the
 155   // underlying object file.
 156   Symbol *getSymbol(uint32_t symbolIndex) {
 157     return symbols[symbolIndex];
 158   }
 159
 160   // Returns the underlying COFF file.
 161   COFFObjectFile *getCOFFObj() { return coffObj.get(); }
 162
 163   // Add a symbol for a range extension thunk. Return the new symbol table
 164   // index. This index can be used to modify a relocation.
 165   uint32_t addRangeThunkSymbol(Symbol *thunk) {
 166     symbols.push_back(thunk);
 167     return symbols.size() - 1;
 168   }
 169
 170   void includeResourceChunks();
 171
 172   bool isResourceObjFile() const { return !resourceChunks.empty(); }
 173
 174   // Flags in the absolute @feat.00 symbol if it is present. These usually
 175   // indicate if an object was compiled with certain security features enabled
 176   // like stack guard, safeseh, /guard:cf, or other things.
 177   uint32_t feat00Flags = 0;
 178
 179   // True if this object file is compatible with SEH.  COFF-specific and
 180   // x86-only. COFF spec 5.10.1. The .sxdata section.
 181   bool hasSafeSEH() { return feat00Flags & 0x1; }
 182
 183   // True if this file was compiled with /guard:cf.
 184   bool hasGuardCF() { return feat00Flags & 0x800; }
 185
 186   // True if this file was compiled with /guard:ehcont.
 187   bool hasGuardEHCont() { return feat00Flags & 0x4000; }
 188
 189   // Pointer to the PDB module descriptor builder. Various debug info records
 190   // will reference object files by "module index", which is here. Things like
 191   // source files and section contributions are also recorded here. Will be null
 192   // if we are not producing a PDB.
 193   llvm::pdb::DbiModuleDescriptorBuilder *moduleDBI = nullptr;
 194
 195   const coff_section *addrsigSec = nullptr;
 196
 197   const coff_section *callgraphSec = nullptr;
 198
 199   // When using Microsoft precompiled headers, this is the PCH's key.
 200   // The same key is used by both the precompiled object, and objects using the
 201   // precompiled object. Any difference indicates out-of-date objects.
 202   std::optional<uint32_t> pchSignature;
 203
 204   // Whether this file was compiled with /hotpatch.
 205   bool hotPatchable = false;
 206
 207   // Whether the object was already merged into the final PDB.
 208   bool mergedIntoPDB = false;
 209
 210   // If the OBJ has a .debug$T stream, this tells how it will be handled.
 211   TpiSource *debugTypesObj = nullptr;
 212
 213   // The .debug$P or .debug$T section data if present. Empty otherwise.
 214   ArrayRef<uint8_t> debugTypes;
 215
 216   std::optional<std::pair<StringRef, uint32_t>>
 217   getVariableLocation(StringRef var);
 218
 219   std::optional<llvm::DILineInfo> getDILineInfo(uint32_t offset,
 220                                                 uint32_t sectionIndex);
 221
 222 private:
 223   const coff_section* getSection(uint32_t i);
 224   const coff_section *getSection(COFFSymbolRef sym) {
 225     return getSection(sym.getSectionNumber());
 226   }
 227
 228   void enqueuePdbFile(StringRef path, ObjFile *fromFile);
 229
 230   void initializeChunks();
 231   void initializeSymbols();
 232   void initializeFlags();
 233   void initializeDependencies();
 234   void initializeECThunks();
 235
 236   SectionChunk *
 237   readSection(uint32_t sectionNumber,
 238               const llvm::object::coff_aux_section_definition *def,
 239               StringRef leaderName);
 240
 241   void readAssociativeDefinition(
 242       COFFSymbolRef coffSym,
 243       const llvm::object::coff_aux_section_definition *def);
 244
 245   void readAssociativeDefinition(
 246       COFFSymbolRef coffSym,
 247       const llvm::object::coff_aux_section_definition *def,
 248       uint32_t parentSection);
 249
 250   void recordPrevailingSymbolForMingw(
 251       COFFSymbolRef coffSym,
 252       llvm::DenseMap<StringRef, uint32_t> &prevailingSectionMap);
 253
 254   void maybeAssociateSEHForMingw(
 255       COFFSymbolRef sym, const llvm::object::coff_aux_section_definition *def,
 256       const llvm::DenseMap<StringRef, uint32_t> &prevailingSectionMap);
 257
 258   // Given a new symbol Sym with comdat selection Selection, if the new
 259   // symbol is not (yet) Prevailing and the existing comdat leader set to
 260   // Leader, emits a diagnostic if the new symbol and its selection doesn't
 261   // match the existing symbol and its selection. If either old or new
 262   // symbol have selection IMAGE_COMDAT_SELECT_LARGEST, Sym might replace
 263   // the existing leader. In that case, Prevailing is set to true.
 264   void
 265   handleComdatSelection(COFFSymbolRef sym, llvm::COFF::COMDATType &selection,
 266                         bool &prevailing, DefinedRegular *leader,
 267                         const llvm::object::coff_aux_section_definition *def);
 268
 269   std::optional<Symbol *>
 270   createDefined(COFFSymbolRef sym,
 271                 std::vector<const llvm::object::coff_aux_section_definition *>
 272                     &comdatDefs,
 273                 bool &prevailingComdat);
 274   Symbol *createRegular(COFFSymbolRef sym);
 275   Symbol *createUndefined(COFFSymbolRef sym, bool overrideLazy);
 276
 277   std::unique_ptr<COFFObjectFile> coffObj;
 278
 279   // List of all chunks defined by this file. This includes both section
 280   // chunks and non-section chunks for common symbols.
 281   std::vector<Chunk *> chunks;
 282
 283   std::vector<SectionChunk *> resourceChunks;
 284
 285   // CodeView debug info sections.
 286   std::vector<SectionChunk *> debugChunks;
 287
 288   // Chunks containing symbol table indices of exception handlers. Only used for
 289   // 32-bit x86.
 290   std::vector<SectionChunk *> sxDataChunks;
 291
 292   // Chunks containing symbol table indices of address taken symbols, address
 293   // taken IAT entries, longjmp and ehcont targets. These are not linked into
 294   // the final binary when /guard:cf is set.
 295   std::vector<SectionChunk *> guardFidChunks;
 296   std::vector<SectionChunk *> guardIATChunks;
 297   std::vector<SectionChunk *> guardLJmpChunks;
 298   std::vector<SectionChunk *> guardEHContChunks;
 299
 300   std::vector<SectionChunk *> hybmpChunks;
 301
 302   // This vector contains a list of all symbols defined or referenced by this
 303   // file. They are indexed such that you can get a Symbol by symbol
 304   // index. Nonexistent indices (which are occupied by auxiliary
 305   // symbols in the real symbol table) are filled with null pointers.
 306   std::vector<Symbol *> symbols;
 307
 308   // This vector contains the same chunks as Chunks, but they are
 309   // indexed such that you can get a SectionChunk by section index.
 310   // Nonexistent section indices are filled with null pointers.
 311   // (Because section number is 1-based, the first slot is always a
 312   // null pointer.) This vector is only valid during initialization.
 313   std::vector<SectionChunk *> sparseChunks;
 314
 315   DWARFCache *dwarf = nullptr;
 316 };
 317
 318 // This is a PDB type server dependency, that is not a input file per se, but
 319 // needs to be treated like one. Such files are discovered from the debug type
 320 // stream.
 321 class PDBInputFile : public InputFile {
 322 public:
 323   explicit PDBInputFile(COFFLinkerContext &ctx, MemoryBufferRef m);
 324   ~PDBInputFile();
 325   static bool classof(const InputFile *f) { return f->kind() == PDBKind; }
 326   void parse() override;
 327
 328   static PDBInputFile *findFromRecordPath(const COFFLinkerContext &ctx,
 329                                           StringRef path, ObjFile *fromFile);
 330
 331   // Record possible errors while opening the PDB file
 332   std::optional<std::string> loadErrorStr;
 333
 334   // This is the actual interface to the PDB (if it was opened successfully)
 335   std::unique_ptr<llvm::pdb::NativeSession> session;
 336
 337   // If the PDB has a .debug$T stream, this tells how it will be handled.
 338   TpiSource *debugTypesObj = nullptr;
 339 };
 340
 341 // This type represents import library members that contain DLL names
 342 // and symbols exported from the DLLs. See Microsoft PE/COFF spec. 7
 343 // for details about the format.
 344 class ImportFile : public InputFile {
 345 public:
 346   explicit ImportFile(COFFLinkerContext &ctx, MemoryBufferRef m);
 347
 348   static bool classof(const InputFile *f) { return f->kind() == ImportKind; }
 349   MachineTypes getMachineType() const override;
 350
 351   DefinedImportData *impSym = nullptr;
 352   Defined *thunkSym = nullptr;
 353   ImportThunkChunkARM64EC *impchkThunk = nullptr;
 354   std::string dllName;
 355
 356 private:
 357   void parse() override;
 358   ImportThunkChunk *makeImportThunk();
 359
 360 public:
 361   StringRef externalName;
 362   const coff_import_header *hdr;
 363   Chunk *location = nullptr;
 364
 365   // Auxiliary IAT symbols and chunks on ARM64EC.
 366   DefinedImportData *impECSym = nullptr;
 367   Chunk *auxLocation = nullptr;
 368   Defined *auxThunkSym = nullptr;
 369   DefinedImportData *auxImpCopySym = nullptr;
 370   Chunk *auxCopyLocation = nullptr;
 371
 372   // We want to eliminate dllimported symbols if no one actually refers to them.
 373   // These "Live" bits are used to keep track of which import library members
 374   // are actually in use.
 375   //
 376   // If the Live bit is turned off by MarkLive, Writer will ignore dllimported
 377   // symbols provided by this import library member.
 378   bool live;
 379 };
 380
 381 // Used for LTO.
 382 class BitcodeFile : public InputFile {
 383 public:
 384   explicit BitcodeFile(COFFLinkerContext &ctx, MemoryBufferRef mb,
 385                        StringRef archiveName, uint64_t offsetInArchive,
 386                        bool lazy);
 387   ~BitcodeFile();
 388   static bool classof(const InputFile *f) { return f->kind() == BitcodeKind; }
 389   ArrayRef<Symbol *> getSymbols() { return symbols; }
 390   MachineTypes getMachineType() const override;
 391   void parseLazy();
 392   std::unique_ptr<llvm::lto::InputFile> obj;
 393
 394 private:
 395   void parse() override;
 396
 397   std::vector<Symbol *> symbols;
 398 };
 399
 400 // .dll file. MinGW only.
 401 class DLLFile : public InputFile {
 402 public:
 403   explicit DLLFile(COFFLinkerContext &ctx, MemoryBufferRef m)
 404       : InputFile(ctx, DLLKind, m) {}
 405   static bool classof(const InputFile *f) { return f->kind() == DLLKind; }
 406   void parse() override;
 407   MachineTypes getMachineType() const override;
 408
 409   struct Symbol {
 410     StringRef dllName;
 411     StringRef symbolName;
 412     llvm::COFF::ImportNameType nameType;
 413     llvm::COFF::ImportType importType;
 414   };
 415
 416   void makeImport(Symbol *s);
 417
 418 private:
 419   std::unique_ptr<COFFObjectFile> coffObj;
 420   llvm::StringSet<> seen;
 421 };
 422
 423 inline bool isBitcode(MemoryBufferRef mb) {
 424   return identify_magic(mb.getBuffer()) == llvm::file_magic::bitcode;
 425 }
 426
 427 std::string replaceThinLTOSuffix(StringRef path, StringRef suffix,
 428                                  StringRef repl);
 429 } // namespace coff
 430
 431 std::string toString(const coff::InputFile *file);
 432 } // namespace lld
 433
 434 #endif