1 //===- llvm/MC/MCAsmParser.h - Abstract Asm Parser Interface ----*- 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 LLVM_MC_MCPARSER_MCASMPARSER_H 10 #define LLVM_MC_MCPARSER_MCASMPARSER_H 11 12 #include "llvm/ADT/STLFunctionalExtras.h" 13 #include "llvm/ADT/SmallString.h" 14 #include "llvm/ADT/SmallVector.h" 15 #include "llvm/ADT/StringRef.h" 16 #include "llvm/ADT/Twine.h" 17 #include "llvm/MC/MCAsmMacro.h" 18 #include "llvm/MC/MCContext.h" 19 #include "llvm/MC/MCParser/AsmLexer.h" 20 #include "llvm/Support/Compiler.h" 21 #include "llvm/Support/SMLoc.h" 22 #include <cstdint> 23 #include <string> 24 #include <utility> 25 26 namespace llvm { 27 28 class MCAsmInfo; 29 class MCAsmParserExtension; 30 class MCExpr; 31 class MCInstPrinter; 32 class MCInstrInfo; 33 class MCStreamer; 34 class MCTargetAsmParser; 35 class SourceMgr; 36 37 struct InlineAsmIdentifierInfo { 38 enum IdKind { 39 IK_Invalid, // Initial state. Unexpected after a successful parsing. 40 IK_Label, // Function/Label reference. 41 IK_EnumVal, // Value of enumeration type. 42 IK_Var // Variable. 43 }; 44 // Represents an Enum value 45 struct EnumIdentifier { 46 int64_t EnumVal; 47 }; 48 // Represents a label/function reference 49 struct LabelIdentifier { 50 void *Decl; 51 }; 52 // Represents a variable 53 struct VariableIdentifier { 54 void *Decl; 55 bool IsGlobalLV; 56 unsigned Length; 57 unsigned Size; 58 unsigned Type; 59 }; 60 // An InlineAsm identifier can only be one of those 61 union { 62 EnumIdentifier Enum; 63 LabelIdentifier Label; 64 VariableIdentifier Var; 65 }; isKindInlineAsmIdentifierInfo66 bool isKind(IdKind kind) const { return Kind == kind; } 67 // Initializers setEnumInlineAsmIdentifierInfo68 void setEnum(int64_t enumVal) { 69 assert(isKind(IK_Invalid) && "should be initialized only once"); 70 Kind = IK_EnumVal; 71 Enum.EnumVal = enumVal; 72 } setLabelInlineAsmIdentifierInfo73 void setLabel(void *decl) { 74 assert(isKind(IK_Invalid) && "should be initialized only once"); 75 Kind = IK_Label; 76 Label.Decl = decl; 77 } setVarInlineAsmIdentifierInfo78 void setVar(void *decl, bool isGlobalLV, unsigned size, unsigned type) { 79 assert(isKind(IK_Invalid) && "should be initialized only once"); 80 Kind = IK_Var; 81 Var.Decl = decl; 82 Var.IsGlobalLV = isGlobalLV; 83 Var.Size = size; 84 Var.Type = type; 85 Var.Length = size / type; 86 } 87 InlineAsmIdentifierInfo() = default; 88 89 private: 90 // Discriminate using the current kind. 91 IdKind Kind = IK_Invalid; 92 }; 93 94 // Generic type information for an assembly object. 95 // All sizes measured in bytes. 96 struct AsmTypeInfo { 97 StringRef Name; 98 unsigned Size = 0; 99 unsigned ElementSize = 0; 100 unsigned Length = 0; 101 }; 102 103 struct AsmFieldInfo { 104 AsmTypeInfo Type; 105 unsigned Offset = 0; 106 }; 107 108 /// Generic Sema callback for assembly parser. 109 class LLVM_ABI MCAsmParserSemaCallback { 110 public: 111 virtual ~MCAsmParserSemaCallback(); 112 113 virtual void LookupInlineAsmIdentifier(StringRef &LineBuf, 114 InlineAsmIdentifierInfo &Info, 115 bool IsUnevaluatedContext) = 0; 116 virtual StringRef LookupInlineAsmLabel(StringRef Identifier, SourceMgr &SM, 117 SMLoc Location, bool Create) = 0; 118 virtual bool LookupInlineAsmField(StringRef Base, StringRef Member, 119 unsigned &Offset) = 0; 120 }; 121 122 /// Generic assembler parser interface, for use by target specific 123 /// assembly parsers. 124 class LLVM_ABI MCAsmParser { 125 public: 126 using DirectiveHandler = bool (*)(MCAsmParserExtension*, StringRef, SMLoc); 127 using ExtensionDirectiveHandler = 128 std::pair<MCAsmParserExtension*, DirectiveHandler>; 129 130 struct MCPendingError { 131 SMLoc Loc; 132 SmallString<64> Msg; 133 SMRange Range; 134 }; 135 136 private: 137 MCTargetAsmParser *TargetParser = nullptr; 138 139 protected: // Can only create subclasses. 140 MCAsmParser(MCContext &, MCStreamer &, SourceMgr &, const MCAsmInfo &); 141 142 MCContext &Ctx; 143 MCStreamer &Out; 144 SourceMgr &SrcMgr; 145 const MCAsmInfo &MAI; 146 AsmLexer Lexer; 147 SmallVector<MCPendingError, 0> PendingErrors; 148 149 /// Flag tracking whether any errors have been encountered. 150 bool HadError = false; 151 152 bool ShowParsedOperands = false; 153 154 public: 155 MCAsmParser(const MCAsmParser &) = delete; 156 MCAsmParser &operator=(const MCAsmParser &) = delete; 157 virtual ~MCAsmParser(); 158 159 virtual void addDirectiveHandler(StringRef Directive, 160 ExtensionDirectiveHandler Handler) = 0; 161 162 virtual void addAliasForDirective(StringRef Directive, StringRef Alias) = 0; 163 getContext()164 MCContext &getContext() { return Ctx; } getStreamer()165 MCStreamer &getStreamer() { return Out; } getSourceManager()166 SourceMgr &getSourceManager() { return SrcMgr; } getLexer()167 AsmLexer &getLexer() { return Lexer; } getLexer()168 const AsmLexer &getLexer() const { return Lexer; } 169 getTargetParser()170 MCTargetAsmParser &getTargetParser() const { return *TargetParser; } 171 void setTargetParser(MCTargetAsmParser &P); 172 getAssemblerDialect()173 virtual unsigned getAssemblerDialect() { return 0;} setAssemblerDialect(unsigned i)174 virtual void setAssemblerDialect(unsigned i) { } 175 getShowParsedOperands()176 bool getShowParsedOperands() const { return ShowParsedOperands; } setShowParsedOperands(bool Value)177 void setShowParsedOperands(bool Value) { ShowParsedOperands = Value; } 178 179 /// Run the parser on the input source buffer. 180 virtual bool Run(bool NoInitialTextSection, bool NoFinalize = false) = 0; 181 182 virtual void setParsingMSInlineAsm(bool V) = 0; 183 virtual bool isParsingMSInlineAsm() = 0; 184 discardLTOSymbol(StringRef)185 virtual bool discardLTOSymbol(StringRef) const { return false; } 186 isParsingMasm()187 virtual bool isParsingMasm() const { return false; } 188 defineMacro(StringRef Name,StringRef Value)189 virtual bool defineMacro(StringRef Name, StringRef Value) { return true; } 190 lookUpField(StringRef Name,AsmFieldInfo & Info)191 virtual bool lookUpField(StringRef Name, AsmFieldInfo &Info) const { 192 return true; 193 } lookUpField(StringRef Base,StringRef Member,AsmFieldInfo & Info)194 virtual bool lookUpField(StringRef Base, StringRef Member, 195 AsmFieldInfo &Info) const { 196 return true; 197 } 198 lookUpType(StringRef Name,AsmTypeInfo & Info)199 virtual bool lookUpType(StringRef Name, AsmTypeInfo &Info) const { 200 return true; 201 } 202 203 /// Parse MS-style inline assembly. 204 virtual bool parseMSInlineAsm( 205 std::string &AsmString, unsigned &NumOutputs, unsigned &NumInputs, 206 SmallVectorImpl<std::pair<void *, bool>> &OpDecls, 207 SmallVectorImpl<std::string> &Constraints, 208 SmallVectorImpl<std::string> &Clobbers, const MCInstrInfo *MII, 209 MCInstPrinter *IP, MCAsmParserSemaCallback &SI) = 0; 210 211 /// Emit a note at the location \p L, with the message \p Msg. 212 virtual void Note(SMLoc L, const Twine &Msg, 213 SMRange Range = std::nullopt) = 0; 214 215 /// Emit a warning at the location \p L, with the message \p Msg. 216 /// 217 /// \return The return value is true, if warnings are fatal. 218 virtual bool Warning(SMLoc L, const Twine &Msg, 219 SMRange Range = std::nullopt) = 0; 220 221 /// Return an error at the location \p L, with the message \p Msg. This 222 /// may be modified before being emitted. 223 /// 224 /// \return The return value is always true, as an idiomatic convenience to 225 /// clients. 226 bool Error(SMLoc L, const Twine &Msg, SMRange Range = std::nullopt); 227 228 /// Emit an error at the location \p L, with the message \p Msg. 229 /// 230 /// \return The return value is always true, as an idiomatic convenience to 231 /// clients. 232 virtual bool printError(SMLoc L, const Twine &Msg, 233 SMRange Range = std::nullopt) = 0; 234 hasPendingError()235 bool hasPendingError() { return !PendingErrors.empty(); } 236 printPendingErrors()237 bool printPendingErrors() { 238 bool rv = !PendingErrors.empty(); 239 for (auto &Err : PendingErrors) { 240 printError(Err.Loc, Twine(Err.Msg), Err.Range); 241 } 242 PendingErrors.clear(); 243 return rv; 244 } 245 clearPendingErrors()246 void clearPendingErrors() { PendingErrors.clear(); } 247 248 bool addErrorSuffix(const Twine &Suffix); 249 250 /// Get the next AsmToken in the stream, possibly handling file 251 /// inclusion first. 252 virtual const AsmToken &Lex() = 0; 253 254 /// Get the current AsmToken from the stream. 255 const AsmToken &getTok() const; 256 257 /// Report an error at the current lexer location. 258 bool TokError(const Twine &Msg, SMRange Range = std::nullopt); 259 260 bool parseTokenLoc(SMLoc &Loc); 261 bool parseToken(AsmToken::TokenKind T, const Twine &Msg = "unexpected token"); 262 /// Attempt to parse and consume token, returning true on 263 /// success. 264 bool parseOptionalToken(AsmToken::TokenKind T); 265 parseComma()266 bool parseComma() { return parseToken(AsmToken::Comma, "expected comma"); } parseRParen()267 bool parseRParen() { return parseToken(AsmToken::RParen, "expected ')'"); } 268 bool parseEOL(); 269 bool parseEOL(const Twine &ErrMsg); 270 271 bool parseMany(function_ref<bool()> parseOne, bool hasComma = true); 272 273 bool parseIntToken(int64_t &V, const Twine &ErrMsg = "expected integer"); 274 275 bool check(bool P, const Twine &Msg); 276 bool check(bool P, SMLoc Loc, const Twine &Msg); 277 278 /// Parse an identifier or string (as a quoted identifier) and set \p 279 /// Res to the identifier contents. 280 virtual bool parseIdentifier(StringRef &Res) = 0; 281 282 /// Parse identifier and get or create symbol for it. 283 bool parseSymbol(MCSymbol *&Res); 284 285 /// Parse up to the end of statement and return the contents from the 286 /// current token until the end of the statement; the current token on exit 287 /// will be either the EndOfStatement or EOF. 288 virtual StringRef parseStringToEndOfStatement() = 0; 289 290 /// Parse the current token as a string which may include escaped 291 /// characters and return the string contents. 292 virtual bool parseEscapedString(std::string &Data) = 0; 293 294 /// Parse an angle-bracket delimited string at the current position if one is 295 /// present, returning the string contents. 296 virtual bool parseAngleBracketString(std::string &Data) = 0; 297 298 /// Skip to the end of the current statement, for error recovery. 299 virtual void eatToEndOfStatement() = 0; 300 301 /// Parse an arbitrary expression. 302 /// 303 /// \param Res - The value of the expression. The result is undefined 304 /// on error. 305 /// \return - False on success. 306 virtual bool parseExpression(const MCExpr *&Res, SMLoc &EndLoc) = 0; 307 bool parseExpression(const MCExpr *&Res); 308 309 /// Parse a primary expression. 310 /// 311 /// \param Res - The value of the expression. The result is undefined 312 /// on error. 313 /// \return - False on success. 314 virtual bool parsePrimaryExpr(const MCExpr *&Res, SMLoc &EndLoc, 315 AsmTypeInfo *TypeInfo = nullptr) = 0; 316 317 /// Parse an arbitrary expression, assuming that an initial '(' has 318 /// already been consumed. 319 /// 320 /// \param Res - The value of the expression. The result is undefined 321 /// on error. 322 /// \return - False on success. 323 virtual bool parseParenExpression(const MCExpr *&Res, SMLoc &EndLoc) = 0; 324 325 /// Parse an expression which must evaluate to an absolute value. 326 /// 327 /// \param Res - The value of the absolute expression. The result is undefined 328 /// on error. 329 /// \return - False on success. 330 virtual bool parseAbsoluteExpression(int64_t &Res) = 0; 331 332 /// Ensure that we have a valid section set in the streamer. Otherwise, 333 /// report an error and switch to .text. 334 /// \return - False on success. 335 virtual bool checkForValidSection() = 0; 336 337 /// Parse a .gnu_attribute. 338 bool parseGNUAttribute(SMLoc L, int64_t &Tag, int64_t &IntegerValue); 339 340 bool parseAtSpecifier(const MCExpr *&Res, SMLoc &EndLoc); 341 const MCExpr *applySpecifier(const MCExpr *E, uint32_t Variant); 342 }; 343 344 /// Create an MCAsmParser instance for parsing assembly similar to gas syntax 345 LLVM_ABI MCAsmParser *createMCAsmParser(SourceMgr &, MCContext &, MCStreamer &, 346 const MCAsmInfo &, unsigned CB = 0); 347 348 /// Create an MCAsmParser instance for parsing Microsoft MASM-style assembly 349 LLVM_ABI MCAsmParser *createMCMasmParser(SourceMgr &, MCContext &, MCStreamer &, 350 const MCAsmInfo &, struct tm, 351 unsigned CB = 0); 352 353 } // end namespace llvm 354 355 #endif // LLVM_MC_MCPARSER_MCASMPARSER_H 356