include/llvm/Object/IRSymtab.h

   1 //===- IRSymtab.h - data definitions for IR symbol tables -------*- 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 // This file contains data definitions and a reader and builder for a symbol
  10 // table for LLVM IR. Its purpose is to allow linkers and other consumers of
  11 // bitcode files to efficiently read the symbol table for symbol resolution
  12 // purposes without needing to construct a module in memory.
  13 //
  14 // As with most object files the symbol table has two parts: the symbol table
  15 // itself and a string table which is referenced by the symbol table.
  16 //
  17 // A symbol table corresponds to a single bitcode file, which may consist of
  18 // multiple modules, so symbol tables may likewise contain symbols for multiple
  19 // modules.
  20 //
  21 //===----------------------------------------------------------------------===//
  22
  23 #ifndef LLVM_OBJECT_IRSYMTAB_H
  24 #define LLVM_OBJECT_IRSYMTAB_H
  25
  26 #include "llvm/ADT/ArrayRef.h"
  27 #include "llvm/ADT/StringRef.h"
  28 #include "llvm/ADT/iterator_range.h"
  29 #include "llvm/IR/GlobalValue.h"
  30 #include "llvm/Object/SymbolicFile.h"
  31 #include "llvm/Support/Endian.h"
  32 #include "llvm/Support/Error.h"
  33 #include <cassert>
  34 #include <cstdint>
  35 #include <vector>
  36
  37 namespace llvm {
  38
  39 struct BitcodeFileContents;
  40 class StringTableBuilder;
  41
  42 namespace irsymtab {
  43
  44 namespace storage {
  45
  46 // The data structures in this namespace define the low-level serialization
  47 // format. Clients that just want to read a symbol table should use the
  48 // irsymtab::Reader class.
  49
  50 using Word = support::ulittle32_t;
  51
  52 /// A reference to a string in the string table.
  53 struct Str {
  54   Word Offset, Size;
  55
  56   StringRef get(StringRef Strtab) const {
  57     return {Strtab.data() + Offset, Size};
  58   }
  59 };
  60
  61 /// A reference to a range of objects in the symbol table.
  62 template <typename T> struct Range {
  63   Word Offset, Size;
  64
  65   ArrayRef<T> get(StringRef Symtab) const {
  66     return {reinterpret_cast<const T *>(Symtab.data() + Offset), Size};
  67   }
  68 };
  69
  70 /// Describes the range of a particular module's symbols within the symbol
  71 /// table.
  72 struct Module {
  73   Word Begin, End;
  74
  75   /// The index of the first Uncommon for this Module.
  76   Word UncBegin;
  77 };
  78
  79 /// This is equivalent to an IR comdat.
  80 struct Comdat {
  81   Str Name;
  82 };
  83
  84 /// Contains the information needed by linkers for symbol resolution, as well as
  85 /// by the LTO implementation itself.
  86 struct Symbol {
  87   /// The mangled symbol name.
  88   Str Name;
  89
  90   /// The unmangled symbol name, or the empty string if this is not an IR
  91   /// symbol.
  92   Str IRName;
  93
  94   /// The index into Header::Comdats, or -1 if not a comdat member.
  95   Word ComdatIndex;
  96
  97   Word Flags;
  98   enum FlagBits {
  99     FB_visibility, // 2 bits
 100     FB_has_uncommon = FB_visibility + 2,
 101     FB_undefined,
 102     FB_weak,
 103     FB_common,
 104     FB_indirect,
 105     FB_used,
 106     FB_tls,
 107     FB_may_omit,
 108     FB_global,
 109     FB_format_specific,
 110     FB_unnamed_addr,
 111     FB_executable,
 112   };
 113 };
 114
 115 /// This data structure contains rarely used symbol fields and is optionally
 116 /// referenced by a Symbol.
 117 struct Uncommon {
 118   Word CommonSize, CommonAlign;
 119
 120   /// COFF-specific: the name of the symbol that a weak external resolves to
 121   /// if not defined.
 122   Str COFFWeakExternFallbackName;
 123
 124   /// Specified section name, if any.
 125   Str SectionName;
 126 };
 127
 128 struct Header {
 129   /// Version number of the symtab format. This number should be incremented
 130   /// when the format changes, but it does not need to be incremented if a
 131   /// change to LLVM would cause it to create a different symbol table.
 132   Word Version;
 133   enum { kCurrentVersion = 1 };
 134
 135   /// The producer's version string (LLVM_VERSION_STRING " " LLVM_REVISION).
 136   /// Consumers should rebuild the symbol table from IR if the producer's
 137   /// version does not match the consumer's version due to potential differences
 138   /// in symbol table format, symbol enumeration order and so on.
 139   Str Producer;
 140
 141   Range<Module> Modules;
 142   Range<Comdat> Comdats;
 143   Range<Symbol> Symbols;
 144   Range<Uncommon> Uncommons;
 145
 146   Str TargetTriple, SourceFileName;
 147
 148   /// COFF-specific: linker directives.
 149   Str COFFLinkerOpts;
 150 };
 151
 152 } // end namespace storage
 153
 154 /// Fills in Symtab and StrtabBuilder with a valid symbol and string table for
 155 /// Mods.
 156 Error build(ArrayRef<Module *> Mods, SmallVector<char, 0> &Symtab,
 157             StringTableBuilder &StrtabBuilder, BumpPtrAllocator &Alloc);
 158
 159 /// This represents a symbol that has been read from a storage::Symbol and
 160 /// possibly a storage::Uncommon.
 161 struct Symbol {
 162   // Copied from storage::Symbol.
 163   StringRef Name, IRName;
 164   int ComdatIndex;
 165   uint32_t Flags;
 166
 167   // Copied from storage::Uncommon.
 168   uint32_t CommonSize, CommonAlign;
 169   StringRef COFFWeakExternFallbackName;
 170   StringRef SectionName;
 171
 172   /// Returns the mangled symbol name.
 173   StringRef getName() const { return Name; }
 174
 175   /// Returns the unmangled symbol name, or the empty string if this is not an
 176   /// IR symbol.
 177   StringRef getIRName() const { return IRName; }
 178
 179   /// Returns the index into the comdat table (see Reader::getComdatTable()), or
 180   /// -1 if not a comdat member.
 181   int getComdatIndex() const { return ComdatIndex; }
 182
 183   using S = storage::Symbol;
 184
 185   GlobalValue::VisibilityTypes getVisibility() const {
 186     return GlobalValue::VisibilityTypes((Flags >> S::FB_visibility) & 3);
 187   }
 188
 189   bool isUndefined() const { return (Flags >> S::FB_undefined) & 1; }
 190   bool isWeak() const { return (Flags >> S::FB_weak) & 1; }
 191   bool isCommon() const { return (Flags >> S::FB_common) & 1; }
 192   bool isIndirect() const { return (Flags >> S::FB_indirect) & 1; }
 193   bool isUsed() const { return (Flags >> S::FB_used) & 1; }
 194   bool isTLS() const { return (Flags >> S::FB_tls) & 1; }
 195
 196   bool canBeOmittedFromSymbolTable() const {
 197     return (Flags >> S::FB_may_omit) & 1;
 198   }
 199
 200   bool isGlobal() const { return (Flags >> S::FB_global) & 1; }
 201   bool isFormatSpecific() const { return (Flags >> S::FB_format_specific) & 1; }
 202   bool isUnnamedAddr() const { return (Flags >> S::FB_unnamed_addr) & 1; }
 203   bool isExecutable() const { return (Flags >> S::FB_executable) & 1; }
 204
 205   uint64_t getCommonSize() const {
 206     assert(isCommon());
 207     return CommonSize;
 208   }
 209
 210   uint32_t getCommonAlignment() const {
 211     assert(isCommon());
 212     return CommonAlign;
 213   }
 214
 215   /// COFF-specific: for weak externals, returns the name of the symbol that is
 216   /// used as a fallback if the weak external remains undefined.
 217   StringRef getCOFFWeakExternalFallback() const {
 218     assert(isWeak() && isIndirect());
 219     return COFFWeakExternFallbackName;
 220   }
 221
 222   StringRef getSectionName() const { return SectionName; }
 223 };
 224
 225 /// This class can be used to read a Symtab and Strtab produced by
 226 /// irsymtab::build.
 227 class Reader {
 228   StringRef Symtab, Strtab;
 229
 230   ArrayRef<storage::Module> Modules;
 231   ArrayRef<storage::Comdat> Comdats;
 232   ArrayRef<storage::Symbol> Symbols;
 233   ArrayRef<storage::Uncommon> Uncommons;
 234
 235   StringRef str(storage::Str S) const { return S.get(Strtab); }
 236
 237   template <typename T> ArrayRef<T> range(storage::Range<T> R) const {
 238     return R.get(Symtab);
 239   }
 240
 241   const storage::Header &header() const {
 242     return *reinterpret_cast<const storage::Header *>(Symtab.data());
 243   }
 244
 245 public:
 246   class SymbolRef;
 247
 248   Reader() = default;
 249   Reader(StringRef Symtab, StringRef Strtab) : Symtab(Symtab), Strtab(Strtab) {
 250     Modules = range(header().Modules);
 251     Comdats = range(header().Comdats);
 252     Symbols = range(header().Symbols);
 253     Uncommons = range(header().Uncommons);
 254   }
 255
 256   using symbol_range = iterator_range<object::content_iterator<SymbolRef>>;
 257
 258   /// Returns the symbol table for the entire bitcode file.
 259   /// The symbols enumerated by this method are ephemeral, but they can be
 260   /// copied into an irsymtab::Symbol object.
 261   symbol_range symbols() const;
 262
 263   size_t getNumModules() const { return Modules.size(); }
 264
 265   /// Returns a slice of the symbol table for the I'th module in the file.
 266   /// The symbols enumerated by this method are ephemeral, but they can be
 267   /// copied into an irsymtab::Symbol object.
 268   symbol_range module_symbols(unsigned I) const;
 269
 270   StringRef getTargetTriple() const { return str(header().TargetTriple); }
 271
 272   /// Returns the source file path specified at compile time.
 273   StringRef getSourceFileName() const { return str(header().SourceFileName); }
 274
 275   /// Returns a table with all the comdats used by this file.
 276   std::vector<StringRef> getComdatTable() const {
 277     std::vector<StringRef> ComdatTable;
 278     ComdatTable.reserve(Comdats.size());
 279     for (auto C : Comdats)
 280       ComdatTable.push_back(str(C.Name));
 281     return ComdatTable;
 282   }
 283
 284   /// COFF-specific: returns linker options specified in the input file.
 285   StringRef getCOFFLinkerOpts() const { return str(header().COFFLinkerOpts); }
 286 };
 287
 288 /// Ephemeral symbols produced by Reader::symbols() and
 289 /// Reader::module_symbols().
 290 class Reader::SymbolRef : public Symbol {
 291   const storage::Symbol *SymI, *SymE;
 292   const storage::Uncommon *UncI;
 293   const Reader *R;
 294
 295   void read() {
 296     if (SymI == SymE)
 297       return;
 298
 299     Name = R->str(SymI->Name);
 300     IRName = R->str(SymI->IRName);
 301     ComdatIndex = SymI->ComdatIndex;
 302     Flags = SymI->Flags;
 303
 304     if (Flags & (1 << storage::Symbol::FB_has_uncommon)) {
 305       CommonSize = UncI->CommonSize;
 306       CommonAlign = UncI->CommonAlign;
 307       COFFWeakExternFallbackName = R->str(UncI->COFFWeakExternFallbackName);
 308       SectionName = R->str(UncI->SectionName);
 309     } else
 310       // Reset this field so it can be queried unconditionally for all symbols.
 311       SectionName = "";
 312   }
 313
 314 public:
 315   SymbolRef(const storage::Symbol *SymI, const storage::Symbol *SymE,
 316             const storage::Uncommon *UncI, const Reader *R)
 317       : SymI(SymI), SymE(SymE), UncI(UncI), R(R) {
 318     read();
 319   }
 320
 321   void moveNext() {
 322     ++SymI;
 323     if (Flags & (1 << storage::Symbol::FB_has_uncommon))
 324       ++UncI;
 325     read();
 326   }
 327
 328   bool operator==(const SymbolRef &Other) const { return SymI == Other.SymI; }
 329 };
 330
 331 inline Reader::symbol_range Reader::symbols() const {
 332   return {SymbolRef(Symbols.begin(), Symbols.end(), Uncommons.begin(), this),
 333           SymbolRef(Symbols.end(), Symbols.end(), nullptr, this)};
 334 }
 335
 336 inline Reader::symbol_range Reader::module_symbols(unsigned I) const {
 337   const storage::Module &M = Modules[I];
 338   const storage::Symbol *MBegin = Symbols.begin() + M.Begin,
 339                         *MEnd = Symbols.begin() + M.End;
 340   return {SymbolRef(MBegin, MEnd, Uncommons.begin() + M.UncBegin, this),
 341           SymbolRef(MEnd, MEnd, nullptr, this)};
 342 }
 343
 344 /// The contents of the irsymtab in a bitcode file. Any underlying data for the
 345 /// irsymtab are owned by Symtab and Strtab.
 346 struct FileContents {
 347   SmallVector<char, 0> Symtab, Strtab;
 348   Reader TheReader;
 349 };
 350
 351 /// Reads the contents of a bitcode file, creating its irsymtab if necessary.
 352 Expected<FileContents> readBitcode(const BitcodeFileContents &BFC);
 353
 354 } // end namespace irsymtab
 355 } // end namespace llvm
 356
 357 #endif // LLVM_OBJECT_IRSYMTAB_H