include/llvm/MC/MCSymbol.h

   1 //===- MCSymbol.h - Machine Code Symbols ------------------------*- 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 the declaration of the MCSymbol class.
  10 //
  11 //===----------------------------------------------------------------------===//
  12
  13 #ifndef LLVM_MC_MCSYMBOL_H
  14 #define LLVM_MC_MCSYMBOL_H
  15
  16 #include "llvm/ADT/PointerIntPair.h"
  17 #include "llvm/ADT/StringMap.h"
  18 #include "llvm/ADT/StringRef.h"
  19 #include "llvm/MC/MCFragment.h"
  20 #include "llvm/Support/ErrorHandling.h"
  21 #include "llvm/Support/MathExtras.h"
  22 #include <cassert>
  23 #include <cstddef>
  24 #include <cstdint>
  25
  26 namespace llvm {
  27
  28 class MCAsmInfo;
  29 class MCContext;
  30 class MCExpr;
  31 class MCSection;
  32 class raw_ostream;
  33
  34 /// MCSymbol - Instances of this class represent a symbol name in the MC file,
  35 /// and MCSymbols are created and uniqued by the MCContext class.  MCSymbols
  36 /// should only be constructed with valid names for the object file.
  37 ///
  38 /// If the symbol is defined/emitted into the current translation unit, the
  39 /// Section member is set to indicate what section it lives in.  Otherwise, if
  40 /// it is a reference to an external entity, it has a null section.
  41 class MCSymbol {
  42 protected:
  43   /// The kind of the symbol.  If it is any value other than unset then this
  44   /// class is actually one of the appropriate subclasses of MCSymbol.
  45   enum SymbolKind {
  46     SymbolKindUnset,
  47     SymbolKindCOFF,
  48     SymbolKindELF,
  49     SymbolKindMachO,
  50     SymbolKindWasm,
  51   };
  52
  53   /// A symbol can contain an Offset, or Value, or be Common, but never more
  54   /// than one of these.
  55   enum Contents : uint8_t {
  56     SymContentsUnset,
  57     SymContentsOffset,
  58     SymContentsVariable,
  59     SymContentsCommon,
  60   };
  61
  62   // Special sentinal value for the absolute pseudo fragment.
  63   static MCFragment *AbsolutePseudoFragment;
  64
  65   /// If a symbol has a Fragment, the section is implied, so we only need
  66   /// one pointer.
  67   /// The special AbsolutePseudoFragment value is for absolute symbols.
  68   /// If this is a variable symbol, this caches the variable value's fragment.
  69   /// FIXME: We might be able to simplify this by having the asm streamer create
  70   /// dummy fragments.
  71   /// If this is a section, then it gives the symbol is defined in. This is null
  72   /// for undefined symbols.
  73   ///
  74   /// If this is a fragment, then it gives the fragment this symbol's value is
  75   /// relative to, if any.
  76   ///
  77   /// For the 'HasName' integer, this is true if this symbol is named.
  78   /// A named symbol will have a pointer to the name allocated in the bytes
  79   /// immediately prior to the MCSymbol.
  80   mutable PointerIntPair<MCFragment *, 1> FragmentAndHasName;
  81
  82   /// IsTemporary - True if this is an assembler temporary label, which
  83   /// typically does not survive in the .o file's symbol table.  Usually
  84   /// "Lfoo" or ".foo".
  85   unsigned IsTemporary : 1;
  86
  87   /// True if this symbol can be redefined.
  88   unsigned IsRedefinable : 1;
  89
  90   /// IsUsed - True if this symbol has been used.
  91   mutable unsigned IsUsed : 1;
  92
  93   mutable unsigned IsRegistered : 1;
  94
  95   /// This symbol is visible outside this translation unit.
  96   mutable unsigned IsExternal : 1;
  97
  98   /// This symbol is private extern.
  99   mutable unsigned IsPrivateExtern : 1;
 100
 101   /// LLVM RTTI discriminator. This is actually a SymbolKind enumerator, but is
 102   /// unsigned to avoid sign extension and achieve better bitpacking with MSVC.
 103   unsigned Kind : 3;
 104
 105   /// True if we have created a relocation that uses this symbol.
 106   mutable unsigned IsUsedInReloc : 1;
 107
 108   /// This is actually a Contents enumerator, but is unsigned to avoid sign
 109   /// extension and achieve better bitpacking with MSVC.
 110   unsigned SymbolContents : 2;
 111
 112   /// The alignment of the symbol, if it is 'common', or -1.
 113   ///
 114   /// The alignment is stored as log2(align) + 1.  This allows all values from
 115   /// 0 to 2^31 to be stored which is every power of 2 representable by an
 116   /// unsigned.
 117   enum : unsigned { NumCommonAlignmentBits = 5 };
 118   unsigned CommonAlignLog2 : NumCommonAlignmentBits;
 119
 120   /// The Flags field is used by object file implementations to store
 121   /// additional per symbol information which is not easily classified.
 122   enum : unsigned { NumFlagsBits = 16 };
 123   mutable uint32_t Flags : NumFlagsBits;
 124
 125   /// Index field, for use by the object file implementation.
 126   mutable uint32_t Index = 0;
 127
 128   union {
 129     /// The offset to apply to the fragment address to form this symbol's value.
 130     uint64_t Offset;
 131
 132     /// The size of the symbol, if it is 'common'.
 133     uint64_t CommonSize;
 134
 135     /// If non-null, the value for a variable symbol.
 136     const MCExpr *Value;
 137   };
 138
 139   // MCContext creates and uniques these.
 140   friend class MCExpr;
 141   friend class MCContext;
 142
 143   /// The name for a symbol.
 144   /// MCSymbol contains a uint64_t so is probably aligned to 8.  On a 32-bit
 145   /// system, the name is a pointer so isn't going to satisfy the 8 byte
 146   /// alignment of uint64_t.  Account for that here.
 147   using NameEntryStorageTy = union {
 148     const StringMapEntry<bool> *NameEntry;
 149     uint64_t AlignmentPadding;
 150   };
 151
 152   MCSymbol(SymbolKind Kind, const StringMapEntry<bool> *Name, bool isTemporary)
 153       : IsTemporary(isTemporary), IsRedefinable(false), IsUsed(false),
 154         IsRegistered(false), IsExternal(false), IsPrivateExtern(false),
 155         Kind(Kind), IsUsedInReloc(false), SymbolContents(SymContentsUnset),
 156         CommonAlignLog2(0), Flags(0) {
 157     Offset = 0;
 158     FragmentAndHasName.setInt(!!Name);
 159     if (Name)
 160       getNameEntryPtr() = Name;
 161   }
 162
 163   // Provide custom new/delete as we will only allocate space for a name
 164   // if we need one.
 165   void *operator new(size_t s, const StringMapEntry<bool> *Name,
 166                      MCContext &Ctx);
 167
 168 private:
 169   void operator delete(void *);
 170   /// Placement delete - required by std, but never called.
 171   void operator delete(void*, unsigned) {
 172     llvm_unreachable("Constructor throws?");
 173   }
 174   /// Placement delete - required by std, but never called.
 175   void operator delete(void*, unsigned, bool) {
 176     llvm_unreachable("Constructor throws?");
 177   }
 178
 179   MCSection *getSectionPtr() const {
 180     if (MCFragment *F = getFragment()) {
 181       assert(F != AbsolutePseudoFragment);
 182       return F->getParent();
 183     }
 184     return nullptr;
 185   }
 186
 187   /// Get a reference to the name field.  Requires that we have a name
 188   const StringMapEntry<bool> *&getNameEntryPtr() {
 189     assert(FragmentAndHasName.getInt() && "Name is required");
 190     NameEntryStorageTy *Name = reinterpret_cast<NameEntryStorageTy *>(this);
 191     return (*(Name - 1)).NameEntry;
 192   }
 193   const StringMapEntry<bool> *&getNameEntryPtr() const {
 194     return const_cast<MCSymbol*>(this)->getNameEntryPtr();
 195   }
 196
 197 public:
 198   MCSymbol(const MCSymbol &) = delete;
 199   MCSymbol &operator=(const MCSymbol &) = delete;
 200
 201   /// getName - Get the symbol name.
 202   StringRef getName() const {
 203     if (!FragmentAndHasName.getInt())
 204       return StringRef();
 205
 206     return getNameEntryPtr()->first();
 207   }
 208
 209   bool isRegistered() const { return IsRegistered; }
 210   void setIsRegistered(bool Value) const { IsRegistered = Value; }
 211
 212   void setUsedInReloc() const { IsUsedInReloc = true; }
 213   bool isUsedInReloc() const { return IsUsedInReloc; }
 214
 215   /// \name Accessors
 216   /// @{
 217
 218   /// isTemporary - Check if this is an assembler temporary symbol.
 219   bool isTemporary() const { return IsTemporary; }
 220
 221   /// isUsed - Check if this is used.
 222   bool isUsed() const { return IsUsed; }
 223
 224   /// Check if this symbol is redefinable.
 225   bool isRedefinable() const { return IsRedefinable; }
 226   /// Mark this symbol as redefinable.
 227   void setRedefinable(bool Value) { IsRedefinable = Value; }
 228   /// Prepare this symbol to be redefined.
 229   void redefineIfPossible() {
 230     if (IsRedefinable) {
 231       if (SymbolContents == SymContentsVariable) {
 232         Value = nullptr;
 233         SymbolContents = SymContentsUnset;
 234       }
 235       setUndefined();
 236       IsRedefinable = false;
 237     }
 238   }
 239
 240   /// @}
 241   /// \name Associated Sections
 242   /// @{
 243
 244   /// isDefined - Check if this symbol is defined (i.e., it has an address).
 245   ///
 246   /// Defined symbols are either absolute or in some section.
 247   bool isDefined() const { return !isUndefined(); }
 248
 249   /// isInSection - Check if this symbol is defined in some section (i.e., it
 250   /// is defined but not absolute).
 251   bool isInSection() const {
 252     return isDefined() && !isAbsolute();
 253   }
 254
 255   /// isUndefined - Check if this symbol undefined (i.e., implicitly defined).
 256   bool isUndefined(bool SetUsed = true) const {
 257     return getFragment(SetUsed) == nullptr;
 258   }
 259
 260   /// isAbsolute - Check if this is an absolute symbol.
 261   bool isAbsolute() const {
 262     return getFragment() == AbsolutePseudoFragment;
 263   }
 264
 265   /// Get the section associated with a defined, non-absolute symbol.
 266   MCSection &getSection() const {
 267     assert(isInSection() && "Invalid accessor!");
 268     return *getSectionPtr();
 269   }
 270
 271   /// Mark the symbol as defined in the fragment \p F.
 272   void setFragment(MCFragment *F) const {
 273     assert(!isVariable() && "Cannot set fragment of variable");
 274     FragmentAndHasName.setPointer(F);
 275   }
 276
 277   /// Mark the symbol as undefined.
 278   void setUndefined() { FragmentAndHasName.setPointer(nullptr); }
 279
 280   bool isELF() const { return Kind == SymbolKindELF; }
 281
 282   bool isCOFF() const { return Kind == SymbolKindCOFF; }
 283
 284   bool isMachO() const { return Kind == SymbolKindMachO; }
 285
 286   bool isWasm() const { return Kind == SymbolKindWasm; }
 287
 288   /// @}
 289   /// \name Variable Symbols
 290   /// @{
 291
 292   /// isVariable - Check if this is a variable symbol.
 293   bool isVariable() const {
 294     return SymbolContents == SymContentsVariable;
 295   }
 296
 297   /// getVariableValue - Get the value for variable symbols.
 298   const MCExpr *getVariableValue(bool SetUsed = true) const {
 299     assert(isVariable() && "Invalid accessor!");
 300     IsUsed |= SetUsed;
 301     return Value;
 302   }
 303
 304   void setVariableValue(const MCExpr *Value);
 305
 306   /// @}
 307
 308   /// Get the (implementation defined) index.
 309   uint32_t getIndex() const {
 310     return Index;
 311   }
 312
 313   /// Set the (implementation defined) index.
 314   void setIndex(uint32_t Value) const {
 315     Index = Value;
 316   }
 317
 318   bool isUnset() const { return SymbolContents == SymContentsUnset; }
 319
 320   uint64_t getOffset() const {
 321     assert((SymbolContents == SymContentsUnset ||
 322             SymbolContents == SymContentsOffset) &&
 323            "Cannot get offset for a common/variable symbol");
 324     return Offset;
 325   }
 326   void setOffset(uint64_t Value) {
 327     assert((SymbolContents == SymContentsUnset ||
 328             SymbolContents == SymContentsOffset) &&
 329            "Cannot set offset for a common/variable symbol");
 330     Offset = Value;
 331     SymbolContents = SymContentsOffset;
 332   }
 333
 334   /// Return the size of a 'common' symbol.
 335   uint64_t getCommonSize() const {
 336     assert(isCommon() && "Not a 'common' symbol!");
 337     return CommonSize;
 338   }
 339
 340   /// Mark this symbol as being 'common'.
 341   ///
 342   /// \param Size - The size of the symbol.
 343   /// \param Align - The alignment of the symbol.
 344   void setCommon(uint64_t Size, unsigned Align) {
 345     assert(getOffset() == 0);
 346     CommonSize = Size;
 347     SymbolContents = SymContentsCommon;
 348
 349     assert((!Align || isPowerOf2_32(Align)) &&
 350            "Alignment must be a power of 2");
 351     unsigned Log2Align = Log2_32(Align) + 1;
 352     assert(Log2Align < (1U << NumCommonAlignmentBits) &&
 353            "Out of range alignment");
 354     CommonAlignLog2 = Log2Align;
 355   }
 356
 357   ///  Return the alignment of a 'common' symbol.
 358   unsigned getCommonAlignment() const {
 359     assert(isCommon() && "Not a 'common' symbol!");
 360     return CommonAlignLog2 ? (1U << (CommonAlignLog2 - 1)) : 0;
 361   }
 362
 363   /// Declare this symbol as being 'common'.
 364   ///
 365   /// \param Size - The size of the symbol.
 366   /// \param Align - The alignment of the symbol.
 367   /// \return True if symbol was already declared as a different type
 368   bool declareCommon(uint64_t Size, unsigned Align) {
 369     assert(isCommon() || getOffset() == 0);
 370     if(isCommon()) {
 371       if(CommonSize != Size || getCommonAlignment() != Align)
 372        return true;
 373     } else
 374       setCommon(Size, Align);
 375     return false;
 376   }
 377
 378   /// Is this a 'common' symbol.
 379   bool isCommon() const {
 380     return SymbolContents == SymContentsCommon;
 381   }
 382
 383   MCFragment *getFragment(bool SetUsed = true) const {
 384     MCFragment *Fragment = FragmentAndHasName.getPointer();
 385     if (Fragment || !isVariable())
 386       return Fragment;
 387     Fragment = getVariableValue(SetUsed)->findAssociatedFragment();
 388     FragmentAndHasName.setPointer(Fragment);
 389     return Fragment;
 390   }
 391
 392   bool isExternal() const { return IsExternal; }
 393   void setExternal(bool Value) const { IsExternal = Value; }
 394
 395   bool isPrivateExtern() const { return IsPrivateExtern; }
 396   void setPrivateExtern(bool Value) { IsPrivateExtern = Value; }
 397
 398   /// print - Print the value to the stream \p OS.
 399   void print(raw_ostream &OS, const MCAsmInfo *MAI) const;
 400
 401   /// dump - Print the value to stderr.
 402   void dump() const;
 403
 404 protected:
 405   /// Get the (implementation defined) symbol flags.
 406   uint32_t getFlags() const { return Flags; }
 407
 408   /// Set the (implementation defined) symbol flags.
 409   void setFlags(uint32_t Value) const {
 410     assert(Value < (1U << NumFlagsBits) && "Out of range flags");
 411     Flags = Value;
 412   }
 413
 414   /// Modify the flags via a mask
 415   void modifyFlags(uint32_t Value, uint32_t Mask) const {
 416     assert(Value < (1U << NumFlagsBits) && "Out of range flags");
 417     Flags = (Flags & ~Mask) | Value;
 418   }
 419 };
 420
 421 inline raw_ostream &operator<<(raw_ostream &OS, const MCSymbol &Sym) {
 422   Sym.print(OS, nullptr);
 423   return OS;
 424 }
 425
 426 } // end namespace llvm
 427
 428 #endif // LLVM_MC_MCSYMBOL_H