Details | Last modification | View Log | RSS feed
| Rev | Author | Line No. | Line | 
|---|---|---|---|
| 14 | pmbaty | 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/StringMapEntry.h" | ||
| 18 | #include "llvm/ADT/StringRef.h" | ||
| 19 | #include "llvm/MC/MCExpr.h" | ||
| 20 | #include "llvm/MC/MCFragment.h" | ||
| 21 | #include "llvm/Support/ErrorHandling.h" | ||
| 22 | #include "llvm/Support/MathExtras.h" | ||
| 23 | #include <cassert> | ||
| 24 | #include <cstddef> | ||
| 25 | #include <cstdint> | ||
| 26 | |||
| 27 | namespace llvm { | ||
| 28 | |||
| 29 | class MCAsmInfo; | ||
| 30 | class MCContext; | ||
| 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 | SymbolKindGOFF, | ||
| 50 | SymbolKindMachO, | ||
| 51 | SymbolKindWasm, | ||
| 52 | SymbolKindXCOFF, | ||
| 53 | }; | ||
| 54 | |||
| 55 |   /// A symbol can contain an Offset, or Value, or be Common, but never more | ||
| 56 |   /// than one of these. | ||
| 57 | enum Contents : uint8_t { | ||
| 58 | SymContentsUnset, | ||
| 59 | SymContentsOffset, | ||
| 60 | SymContentsVariable, | ||
| 61 | SymContentsCommon, | ||
| 62 |     SymContentsTargetCommon, // Index stores the section index | ||
| 63 | }; | ||
| 64 | |||
| 65 |   // Special sentinal value for the absolute pseudo fragment. | ||
| 66 | static MCFragment *AbsolutePseudoFragment; | ||
| 67 | |||
| 68 |   /// If a symbol has a Fragment, the section is implied, so we only need | ||
| 69 |   /// one pointer. | ||
| 70 |   /// The special AbsolutePseudoFragment value is for absolute symbols. | ||
| 71 |   /// If this is a variable symbol, this caches the variable value's fragment. | ||
| 72 |   /// FIXME: We might be able to simplify this by having the asm streamer create | ||
| 73 |   /// dummy fragments. | ||
| 74 |   /// If this is a section, then it gives the symbol is defined in. This is null | ||
| 75 |   /// for undefined symbols. | ||
| 76 |   /// | ||
| 77 |   /// If this is a fragment, then it gives the fragment this symbol's value is | ||
| 78 |   /// relative to, if any. | ||
| 79 |   /// | ||
| 80 |   /// For the 'HasName' integer, this is true if this symbol is named. | ||
| 81 |   /// A named symbol will have a pointer to the name allocated in the bytes | ||
| 82 |   /// immediately prior to the MCSymbol. | ||
| 83 | mutable PointerIntPair<MCFragment *, 1> FragmentAndHasName; | ||
| 84 | |||
| 85 |   /// IsTemporary - True if this is an assembler temporary label, which | ||
| 86 |   /// typically does not survive in the .o file's symbol table.  Usually | ||
| 87 |   /// "Lfoo" or ".foo". | ||
| 88 | unsigned IsTemporary : 1; | ||
| 89 | |||
| 90 |   /// True if this symbol can be redefined. | ||
| 91 | unsigned IsRedefinable : 1; | ||
| 92 | |||
| 93 |   /// IsUsed - True if this symbol has been used. | ||
| 94 | mutable unsigned IsUsed : 1; | ||
| 95 | |||
| 96 | mutable unsigned IsRegistered : 1; | ||
| 97 | |||
| 98 |   /// True if this symbol is visible outside this translation unit. Note: ELF | ||
| 99 |   /// uses binding instead of this bit. | ||
| 100 | mutable unsigned IsExternal : 1; | ||
| 101 | |||
| 102 |   /// This symbol is private extern. | ||
| 103 | mutable unsigned IsPrivateExtern : 1; | ||
| 104 | |||
| 105 |   /// LLVM RTTI discriminator. This is actually a SymbolKind enumerator, but is | ||
| 106 |   /// unsigned to avoid sign extension and achieve better bitpacking with MSVC. | ||
| 107 | unsigned Kind : 3; | ||
| 108 | |||
| 109 |   /// True if we have created a relocation that uses this symbol. | ||
| 110 | mutable unsigned IsUsedInReloc : 1; | ||
| 111 | |||
| 112 |   /// This is actually a Contents enumerator, but is unsigned to avoid sign | ||
| 113 |   /// extension and achieve better bitpacking with MSVC. | ||
| 114 | unsigned SymbolContents : 3; | ||
| 115 | |||
| 116 |   /// The alignment of the symbol if it is 'common'. | ||
| 117 |   /// | ||
| 118 |   /// Internally, this is stored as log2(align) + 1. | ||
| 119 |   /// We reserve 5 bits to encode this value which allows the following values | ||
| 120 |   /// 0b00000 -> unset | ||
| 121 |   /// 0b00001 -> 1ULL <<  0 = 1 | ||
| 122 |   /// 0b00010 -> 1ULL <<  1 = 2 | ||
| 123 |   /// 0b00011 -> 1ULL <<  2 = 4 | ||
| 124 |   /// ... | ||
| 125 |   /// 0b11111 -> 1ULL << 30 = 1 GiB | ||
| 126 | enum : unsigned { NumCommonAlignmentBits = 5 }; | ||
| 127 | unsigned CommonAlignLog2 : NumCommonAlignmentBits; | ||
| 128 | |||
| 129 |   /// The Flags field is used by object file implementations to store | ||
| 130 |   /// additional per symbol information which is not easily classified. | ||
| 131 | enum : unsigned { NumFlagsBits = 16 }; | ||
| 132 | mutable uint32_t Flags : NumFlagsBits; | ||
| 133 | |||
| 134 |   /// Index field, for use by the object file implementation. | ||
| 135 | mutable uint32_t Index = 0; | ||
| 136 | |||
| 137 | union { | ||
| 138 |     /// The offset to apply to the fragment address to form this symbol's value. | ||
| 139 | uint64_t Offset; | ||
| 140 | |||
| 141 |     /// The size of the symbol, if it is 'common'. | ||
| 142 | uint64_t CommonSize; | ||
| 143 | |||
| 144 |     /// If non-null, the value for a variable symbol. | ||
| 145 | const MCExpr *Value; | ||
| 146 | }; | ||
| 147 | |||
| 148 |   // MCContext creates and uniques these. | ||
| 149 | friend class MCExpr; | ||
| 150 | friend class MCContext; | ||
| 151 | |||
| 152 |   /// The name for a symbol. | ||
| 153 |   /// MCSymbol contains a uint64_t so is probably aligned to 8.  On a 32-bit | ||
| 154 |   /// system, the name is a pointer so isn't going to satisfy the 8 byte | ||
| 155 |   /// alignment of uint64_t.  Account for that here. | ||
| 156 | using NameEntryStorageTy = union { | ||
| 157 | const StringMapEntry<bool> *NameEntry; | ||
| 158 | uint64_t AlignmentPadding; | ||
| 159 | }; | ||
| 160 | |||
| 161 | MCSymbol(SymbolKind Kind, const StringMapEntry<bool> *Name, bool isTemporary) | ||
| 162 | : IsTemporary(isTemporary), IsRedefinable(false), IsUsed(false), | ||
| 163 | IsRegistered(false), IsExternal(false), IsPrivateExtern(false), | ||
| 164 | Kind(Kind), IsUsedInReloc(false), SymbolContents(SymContentsUnset), | ||
| 165 | CommonAlignLog2(0), Flags(0) { | ||
| 166 | Offset = 0; | ||
| 167 | FragmentAndHasName.setInt(!!Name); | ||
| 168 | if (Name) | ||
| 169 | getNameEntryPtr() = Name; | ||
| 170 |   } | ||
| 171 | |||
| 172 |   // Provide custom new/delete as we will only allocate space for a name | ||
| 173 |   // if we need one. | ||
| 174 | void *operator new(size_t s, const StringMapEntry<bool> *Name, | ||
| 175 | MCContext &Ctx); | ||
| 176 | |||
| 177 | private: | ||
| 178 | void operator delete(void *); | ||
| 179 |   /// Placement delete - required by std, but never called. | ||
| 180 | void operator delete(void*, unsigned) { | ||
| 181 | llvm_unreachable("Constructor throws?"); | ||
| 182 |   } | ||
| 183 |   /// Placement delete - required by std, but never called. | ||
| 184 | void operator delete(void*, unsigned, bool) { | ||
| 185 | llvm_unreachable("Constructor throws?"); | ||
| 186 |   } | ||
| 187 | |||
| 188 |   /// Get a reference to the name field.  Requires that we have a name | ||
| 189 | const StringMapEntry<bool> *&getNameEntryPtr() { | ||
| 190 | assert(FragmentAndHasName.getInt() && "Name is required"); | ||
| 191 | NameEntryStorageTy *Name = reinterpret_cast<NameEntryStorageTy *>(this); | ||
| 192 | return (*(Name - 1)).NameEntry; | ||
| 193 |   } | ||
| 194 | const StringMapEntry<bool> *&getNameEntryPtr() const { | ||
| 195 | return const_cast<MCSymbol*>(this)->getNameEntryPtr(); | ||
| 196 |   } | ||
| 197 | |||
| 198 | public: | ||
| 199 | MCSymbol(const MCSymbol &) = delete; | ||
| 200 | MCSymbol &operator=(const MCSymbol &) = delete; | ||
| 201 | |||
| 202 |   /// getName - Get the symbol name. | ||
| 203 | StringRef getName() const { | ||
| 204 | if (!FragmentAndHasName.getInt()) | ||
| 205 | return StringRef(); | ||
| 206 | |||
| 207 | return getNameEntryPtr()->first(); | ||
| 208 |   } | ||
| 209 | |||
| 210 | bool isRegistered() const { return IsRegistered; } | ||
| 211 | void setIsRegistered(bool Value) const { IsRegistered = Value; } | ||
| 212 | |||
| 213 | void setUsedInReloc() const { IsUsedInReloc = true; } | ||
| 214 | bool isUsedInReloc() const { return IsUsedInReloc; } | ||
| 215 | |||
| 216 |   /// \name Accessors | ||
| 217 |   /// @{ | ||
| 218 | |||
| 219 |   /// isTemporary - Check if this is an assembler temporary symbol. | ||
| 220 | bool isTemporary() const { return IsTemporary; } | ||
| 221 | |||
| 222 |   /// isUsed - Check if this is used. | ||
| 223 | bool isUsed() const { return IsUsed; } | ||
| 224 | |||
| 225 |   /// Check if this symbol is redefinable. | ||
| 226 | bool isRedefinable() const { return IsRedefinable; } | ||
| 227 |   /// Mark this symbol as redefinable. | ||
| 228 | void setRedefinable(bool Value) { IsRedefinable = Value; } | ||
| 229 |   /// Prepare this symbol to be redefined. | ||
| 230 | void redefineIfPossible() { | ||
| 231 | if (IsRedefinable) { | ||
| 232 | if (SymbolContents == SymContentsVariable) { | ||
| 233 | Value = nullptr; | ||
| 234 | SymbolContents = SymContentsUnset; | ||
| 235 |       } | ||
| 236 | setUndefined(); | ||
| 237 | IsRedefinable = false; | ||
| 238 |     } | ||
| 239 |   } | ||
| 240 | |||
| 241 |   /// @} | ||
| 242 |   /// \name Associated Sections | ||
| 243 |   /// @{ | ||
| 244 | |||
| 245 |   /// isDefined - Check if this symbol is defined (i.e., it has an address). | ||
| 246 |   /// | ||
| 247 |   /// Defined symbols are either absolute or in some section. | ||
| 248 | bool isDefined() const { return !isUndefined(); } | ||
| 249 | |||
| 250 |   /// isInSection - Check if this symbol is defined in some section (i.e., it | ||
| 251 |   /// is defined but not absolute). | ||
| 252 | bool isInSection() const { | ||
| 253 | return isDefined() && !isAbsolute(); | ||
| 254 |   } | ||
| 255 | |||
| 256 |   /// isUndefined - Check if this symbol undefined (i.e., implicitly defined). | ||
| 257 | bool isUndefined(bool SetUsed = true) const { | ||
| 258 | return getFragment(SetUsed) == nullptr; | ||
| 259 |   } | ||
| 260 | |||
| 261 |   /// isAbsolute - Check if this is an absolute symbol. | ||
| 262 | bool isAbsolute() const { | ||
| 263 | return getFragment() == AbsolutePseudoFragment; | ||
| 264 |   } | ||
| 265 | |||
| 266 |   /// Get the section associated with a defined, non-absolute symbol. | ||
| 267 | MCSection &getSection() const { | ||
| 268 | assert(isInSection() && "Invalid accessor!"); | ||
| 269 | return *getFragment()->getParent(); | ||
| 270 |   } | ||
| 271 | |||
| 272 |   /// Mark the symbol as defined in the fragment \p F. | ||
| 273 | void setFragment(MCFragment *F) const { | ||
| 274 | assert(!isVariable() && "Cannot set fragment of variable"); | ||
| 275 | FragmentAndHasName.setPointer(F); | ||
| 276 |   } | ||
| 277 | |||
| 278 |   /// Mark the symbol as undefined. | ||
| 279 | void setUndefined() { FragmentAndHasName.setPointer(nullptr); } | ||
| 280 | |||
| 281 | bool isELF() const { return Kind == SymbolKindELF; } | ||
| 282 | |||
| 283 | bool isCOFF() const { return Kind == SymbolKindCOFF; } | ||
| 284 | |||
| 285 | bool isGOFF() const { return Kind == SymbolKindGOFF; } | ||
| 286 | |||
| 287 | bool isMachO() const { return Kind == SymbolKindMachO; } | ||
| 288 | |||
| 289 | bool isWasm() const { return Kind == SymbolKindWasm; } | ||
| 290 | |||
| 291 | bool isXCOFF() const { return Kind == SymbolKindXCOFF; } | ||
| 292 | |||
| 293 |   /// @} | ||
| 294 |   /// \name Variable Symbols | ||
| 295 |   /// @{ | ||
| 296 | |||
| 297 |   /// isVariable - Check if this is a variable symbol. | ||
| 298 | bool isVariable() const { | ||
| 299 | return SymbolContents == SymContentsVariable; | ||
| 300 |   } | ||
| 301 | |||
| 302 |   /// getVariableValue - Get the value for variable symbols. | ||
| 303 | const MCExpr *getVariableValue(bool SetUsed = true) const { | ||
| 304 | assert(isVariable() && "Invalid accessor!"); | ||
| 305 | IsUsed |= SetUsed; | ||
| 306 | return Value; | ||
| 307 |   } | ||
| 308 | |||
| 309 | void setVariableValue(const MCExpr *Value); | ||
| 310 | |||
| 311 |   /// @} | ||
| 312 | |||
| 313 |   /// Get the (implementation defined) index. | ||
| 314 | uint32_t getIndex() const { | ||
| 315 | return Index; | ||
| 316 |   } | ||
| 317 | |||
| 318 |   /// Set the (implementation defined) index. | ||
| 319 | void setIndex(uint32_t Value) const { | ||
| 320 | Index = Value; | ||
| 321 |   } | ||
| 322 | |||
| 323 | bool isUnset() const { return SymbolContents == SymContentsUnset; } | ||
| 324 | |||
| 325 | uint64_t getOffset() const { | ||
| 326 | assert((SymbolContents == SymContentsUnset || | ||
| 327 | SymbolContents == SymContentsOffset) && | ||
| 328 | "Cannot get offset for a common/variable symbol"); | ||
| 329 | return Offset; | ||
| 330 |   } | ||
| 331 | void setOffset(uint64_t Value) { | ||
| 332 | assert((SymbolContents == SymContentsUnset || | ||
| 333 | SymbolContents == SymContentsOffset) && | ||
| 334 | "Cannot set offset for a common/variable symbol"); | ||
| 335 | Offset = Value; | ||
| 336 | SymbolContents = SymContentsOffset; | ||
| 337 |   } | ||
| 338 | |||
| 339 |   /// Return the size of a 'common' symbol. | ||
| 340 | uint64_t getCommonSize() const { | ||
| 341 | assert(isCommon() && "Not a 'common' symbol!"); | ||
| 342 | return CommonSize; | ||
| 343 |   } | ||
| 344 | |||
| 345 |   /// Mark this symbol as being 'common'. | ||
| 346 |   /// | ||
| 347 |   /// \param Size - The size of the symbol. | ||
| 348 |   /// \param Alignment - The alignment of the symbol. | ||
| 349 |   /// \param Target - Is the symbol a target-specific common-like symbol. | ||
| 350 | void setCommon(uint64_t Size, Align Alignment, bool Target = false) { | ||
| 351 | assert(getOffset() == 0); | ||
| 352 | CommonSize = Size; | ||
| 353 | SymbolContents = Target ? SymContentsTargetCommon : SymContentsCommon; | ||
| 354 | |||
| 355 | unsigned Log2Align = encode(Alignment); | ||
| 356 | assert(Log2Align < (1U << NumCommonAlignmentBits) && | ||
| 357 | "Out of range alignment"); | ||
| 358 | CommonAlignLog2 = Log2Align; | ||
| 359 |   } | ||
| 360 | |||
| 361 |   ///  Return the alignment of a 'common' symbol. | ||
| 362 | MaybeAlign getCommonAlignment() const { | ||
| 363 | assert(isCommon() && "Not a 'common' symbol!"); | ||
| 364 | return decodeMaybeAlign(CommonAlignLog2); | ||
| 365 |   } | ||
| 366 | |||
| 367 |   /// Declare this symbol as being 'common'. | ||
| 368 |   /// | ||
| 369 |   /// \param Size - The size of the symbol. | ||
| 370 |   /// \param Alignment - The alignment of the symbol. | ||
| 371 |   /// \param Target - Is the symbol a target-specific common-like symbol. | ||
| 372 |   /// \return True if symbol was already declared as a different type | ||
| 373 | bool declareCommon(uint64_t Size, Align Alignment, bool Target = false) { | ||
| 374 | assert(isCommon() || getOffset() == 0); | ||
| 375 | if(isCommon()) { | ||
| 376 | if (CommonSize != Size || getCommonAlignment() != Alignment || | ||
| 377 | isTargetCommon() != Target) | ||
| 378 | return true; | ||
| 379 | } else | ||
| 380 | setCommon(Size, Alignment, Target); | ||
| 381 | return false; | ||
| 382 |   } | ||
| 383 | |||
| 384 |   /// Is this a 'common' symbol. | ||
| 385 | bool isCommon() const { | ||
| 386 | return SymbolContents == SymContentsCommon || | ||
| 387 | SymbolContents == SymContentsTargetCommon; | ||
| 388 |   } | ||
| 389 | |||
| 390 |   /// Is this a target-specific common-like symbol. | ||
| 391 | bool isTargetCommon() const { | ||
| 392 | return SymbolContents == SymContentsTargetCommon; | ||
| 393 |   } | ||
| 394 | |||
| 395 | MCFragment *getFragment(bool SetUsed = true) const { | ||
| 396 | MCFragment *Fragment = FragmentAndHasName.getPointer(); | ||
| 397 | if (Fragment || !isVariable()) | ||
| 398 | return Fragment; | ||
| 399 | Fragment = getVariableValue(SetUsed)->findAssociatedFragment(); | ||
| 400 | FragmentAndHasName.setPointer(Fragment); | ||
| 401 | return Fragment; | ||
| 402 |   } | ||
| 403 | |||
| 404 | bool isExternal() const { return IsExternal; } | ||
| 405 | void setExternal(bool Value) const { IsExternal = Value; } | ||
| 406 | |||
| 407 | bool isPrivateExtern() const { return IsPrivateExtern; } | ||
| 408 | void setPrivateExtern(bool Value) { IsPrivateExtern = Value; } | ||
| 409 | |||
| 410 |   /// print - Print the value to the stream \p OS. | ||
| 411 | void print(raw_ostream &OS, const MCAsmInfo *MAI) const; | ||
| 412 | |||
| 413 |   /// dump - Print the value to stderr. | ||
| 414 | void dump() const; | ||
| 415 | |||
| 416 | protected: | ||
| 417 |   /// Get the (implementation defined) symbol flags. | ||
| 418 | uint32_t getFlags() const { return Flags; } | ||
| 419 | |||
| 420 |   /// Set the (implementation defined) symbol flags. | ||
| 421 | void setFlags(uint32_t Value) const { | ||
| 422 | assert(Value < (1U << NumFlagsBits) && "Out of range flags"); | ||
| 423 | Flags = Value; | ||
| 424 |   } | ||
| 425 | |||
| 426 |   /// Modify the flags via a mask | ||
| 427 | void modifyFlags(uint32_t Value, uint32_t Mask) const { | ||
| 428 | assert(Value < (1U << NumFlagsBits) && "Out of range flags"); | ||
| 429 | Flags = (Flags & ~Mask) | Value; | ||
| 430 |   } | ||
| 431 | }; | ||
| 432 | |||
| 433 | inline raw_ostream &operator<<(raw_ostream &OS, const MCSymbol &Sym) { | ||
| 434 | Sym.print(OS, nullptr); | ||
| 435 | return OS; | ||
| 436 | } | ||
| 437 | |||
| 438 | } // end namespace llvm | ||
| 439 | |||
| 440 | #endif // LLVM_MC_MCSYMBOL_H |