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/None.h" 13 #include "llvm/ADT/STLExtras.h" 14 #include "llvm/ADT/SmallString.h" 15 #include "llvm/ADT/SmallVector.h" 16 #include "llvm/ADT/StringRef.h" 17 #include "llvm/ADT/Twine.h" 18 #include "llvm/MC/MCParser/MCAsmLexer.h" 19 #include "llvm/Support/SMLoc.h" 20 #include <cstdint> 21 #include <string> 22 #include <utility> 23 24 namespace llvm { 25 26 class MCAsmInfo; 27 class MCAsmParserExtension; 28 class MCContext; 29 class MCExpr; 30 class MCInstPrinter; 31 class MCInstrInfo; 32 class MCStreamer; 33 class MCTargetAsmParser; 34 class SourceMgr; 35 36 struct InlineAsmIdentifierInfo { 37 enum IdKind { 38 IK_Invalid, // Initial state. Unexpected after a successful parsing. 39 IK_Label, // Function/Label reference. 40 IK_EnumVal, // Value of enumeration type. 41 IK_Var // Variable. 42 }; 43 // Represents an Enum value 44 struct EnumIdentifier { 45 int64_t EnumVal; 46 }; 47 // Represents a label/function reference 48 struct LabelIdentifier { 49 void *Decl; 50 }; 51 // Represents a variable 52 struct VariableIdentifier { 53 void *Decl; 54 bool IsGlobalLV; 55 unsigned Length; 56 unsigned Size; 57 unsigned Type; 58 }; 59 // An InlineAsm identifier can only be one of those 60 union { 61 EnumIdentifier Enum; 62 LabelIdentifier Label; 63 VariableIdentifier Var; 64 }; isKindInlineAsmIdentifierInfo65 bool isKind(IdKind kind) const { return Kind == kind; } 66 // Initializers setEnumInlineAsmIdentifierInfo67 void setEnum(int64_t enumVal) { 68 assert(isKind(IK_Invalid) && "should be initialized only once"); 69 Kind = IK_EnumVal; 70 Enum.EnumVal = enumVal; 71 } setLabelInlineAsmIdentifierInfo72 void setLabel(void *decl) { 73 assert(isKind(IK_Invalid) && "should be initialized only once"); 74 Kind = IK_Label; 75 Label.Decl = decl; 76 } setVarInlineAsmIdentifierInfo77 void setVar(void *decl, bool isGlobalLV, unsigned size, unsigned type) { 78 assert(isKind(IK_Invalid) && "should be initialized only once"); 79 Kind = IK_Var; 80 Var.Decl = decl; 81 Var.IsGlobalLV = isGlobalLV; 82 Var.Size = size; 83 Var.Type = type; 84 Var.Length = size / type; 85 } InlineAsmIdentifierInfoInlineAsmIdentifierInfo86 InlineAsmIdentifierInfo() : Kind(IK_Invalid) {} 87 88 private: 89 // Discriminate using the current kind. 90 IdKind Kind; 91 }; 92 93 // Generic type information for an assembly object. 94 // All sizes measured in bytes. 95 struct AsmTypeInfo { 96 StringRef Name; 97 unsigned Size = 0; 98 unsigned ElementSize = 0; 99 unsigned Length = 0; 100 }; 101 102 struct AsmFieldInfo { 103 AsmTypeInfo Type; 104 unsigned Offset = 0; 105 }; 106 107 /// Generic Sema callback for assembly parser. 108 class MCAsmParserSemaCallback { 109 public: 110 virtual ~MCAsmParserSemaCallback(); 111 112 virtual void LookupInlineAsmIdentifier(StringRef &LineBuf, 113 InlineAsmIdentifierInfo &Info, 114 bool IsUnevaluatedContext) = 0; 115 virtual StringRef LookupInlineAsmLabel(StringRef Identifier, SourceMgr &SM, 116 SMLoc Location, bool Create) = 0; 117 virtual bool LookupInlineAsmField(StringRef Base, StringRef Member, 118 unsigned &Offset) = 0; 119 }; 120 121 /// Generic assembler parser interface, for use by target specific 122 /// assembly parsers. 123 class MCAsmParser { 124 public: 125 using DirectiveHandler = bool (*)(MCAsmParserExtension*, StringRef, SMLoc); 126 using ExtensionDirectiveHandler = 127 std::pair<MCAsmParserExtension*, DirectiveHandler>; 128 129 struct MCPendingError { 130 SMLoc Loc; 131 SmallString<64> Msg; 132 SMRange Range; 133 }; 134 135 private: 136 MCTargetAsmParser *TargetParser = nullptr; 137 138 protected: // Can only create subclasses. 139 MCAsmParser(); 140 141 SmallVector<MCPendingError, 0> PendingErrors; 142 143 /// Flag tracking whether any errors have been encountered. 144 bool HadError = false; 145 146 bool ShowParsedOperands = false; 147 148 public: 149 MCAsmParser(const MCAsmParser &) = delete; 150 MCAsmParser &operator=(const MCAsmParser &) = delete; 151 virtual ~MCAsmParser(); 152 153 virtual void addDirectiveHandler(StringRef Directive, 154 ExtensionDirectiveHandler Handler) = 0; 155 156 virtual void addAliasForDirective(StringRef Directive, StringRef Alias) = 0; 157 158 virtual SourceMgr &getSourceManager() = 0; 159 160 virtual MCAsmLexer &getLexer() = 0; getLexer()161 const MCAsmLexer &getLexer() const { 162 return const_cast<MCAsmParser*>(this)->getLexer(); 163 } 164 165 virtual MCContext &getContext() = 0; 166 167 /// Return the output streamer for the assembler. 168 virtual MCStreamer &getStreamer() = 0; 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 void *AsmLoc, std::string &AsmString, unsigned &NumOutputs, 206 unsigned &NumInputs, SmallVectorImpl<std::pair<void *, bool>> &OpDecls, 207 SmallVectorImpl<std::string> &Constraints, 208 SmallVectorImpl<std::string> &Clobbers, const MCInstrInfo *MII, 209 const 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, SMRange Range = None) = 0; 213 214 /// Emit a warning at the location \p L, with the message \p Msg. 215 /// 216 /// \return The return value is true, if warnings are fatal. 217 virtual bool Warning(SMLoc L, const Twine &Msg, SMRange Range = None) = 0; 218 219 /// Return an error at the location \p L, with the message \p Msg. This 220 /// may be modified before being emitted. 221 /// 222 /// \return The return value is always true, as an idiomatic convenience to 223 /// clients. 224 bool Error(SMLoc L, const Twine &Msg, SMRange Range = None); 225 226 /// Emit an error at the location \p L, with the message \p Msg. 227 /// 228 /// \return The return value is always true, as an idiomatic convenience to 229 /// clients. 230 virtual bool printError(SMLoc L, const Twine &Msg, SMRange Range = None) = 0; 231 hasPendingError()232 bool hasPendingError() { return !PendingErrors.empty(); } 233 printPendingErrors()234 bool printPendingErrors() { 235 bool rv = !PendingErrors.empty(); 236 for (auto Err : PendingErrors) { 237 printError(Err.Loc, Twine(Err.Msg), Err.Range); 238 } 239 PendingErrors.clear(); 240 return rv; 241 } 242 clearPendingErrors()243 void clearPendingErrors() { PendingErrors.clear(); } 244 245 bool addErrorSuffix(const Twine &Suffix); 246 247 /// Get the next AsmToken in the stream, possibly handling file 248 /// inclusion first. 249 virtual const AsmToken &Lex() = 0; 250 251 /// Get the current AsmToken from the stream. 252 const AsmToken &getTok() const; 253 254 /// Report an error at the current lexer location. 255 bool TokError(const Twine &Msg, SMRange Range = None); 256 257 bool parseTokenLoc(SMLoc &Loc); 258 bool parseToken(AsmToken::TokenKind T, const Twine &Msg = "unexpected token"); 259 /// Attempt to parse and consume token, returning true on 260 /// success. 261 bool parseOptionalToken(AsmToken::TokenKind T); 262 parseComma()263 bool parseComma() { return parseToken(AsmToken::Comma, "expected comma"); } 264 bool parseEOL(); 265 bool parseEOL(const Twine &ErrMsg); 266 267 bool parseMany(function_ref<bool()> parseOne, bool hasComma = true); 268 269 bool parseIntToken(int64_t &V, const Twine &ErrMsg); 270 271 bool check(bool P, const Twine &Msg); 272 bool check(bool P, SMLoc Loc, const Twine &Msg); 273 274 /// Parse an identifier or string (as a quoted identifier) and set \p 275 /// Res to the identifier contents. 276 virtual bool parseIdentifier(StringRef &Res) = 0; 277 278 /// Parse up to the end of statement and return the contents from the 279 /// current token until the end of the statement; the current token on exit 280 /// will be either the EndOfStatement or EOF. 281 virtual StringRef parseStringToEndOfStatement() = 0; 282 283 /// Parse the current token as a string which may include escaped 284 /// characters and return the string contents. 285 virtual bool parseEscapedString(std::string &Data) = 0; 286 287 /// Parse an angle-bracket delimited string at the current position if one is 288 /// present, returning the string contents. 289 virtual bool parseAngleBracketString(std::string &Data) = 0; 290 291 /// Skip to the end of the current statement, for error recovery. 292 virtual void eatToEndOfStatement() = 0; 293 294 /// Parse an arbitrary expression. 295 /// 296 /// \param Res - The value of the expression. The result is undefined 297 /// on error. 298 /// \return - False on success. 299 virtual bool parseExpression(const MCExpr *&Res, SMLoc &EndLoc) = 0; 300 bool parseExpression(const MCExpr *&Res); 301 302 /// Parse a primary expression. 303 /// 304 /// \param Res - The value of the expression. The result is undefined 305 /// on error. 306 /// \return - False on success. 307 virtual bool parsePrimaryExpr(const MCExpr *&Res, SMLoc &EndLoc, 308 AsmTypeInfo *TypeInfo) = 0; 309 310 /// Parse an arbitrary expression, assuming that an initial '(' has 311 /// already been consumed. 312 /// 313 /// \param Res - The value of the expression. The result is undefined 314 /// on error. 315 /// \return - False on success. 316 virtual bool parseParenExpression(const MCExpr *&Res, SMLoc &EndLoc) = 0; 317 318 /// Parse an expression which must evaluate to an absolute value. 319 /// 320 /// \param Res - The value of the absolute expression. The result is undefined 321 /// on error. 322 /// \return - False on success. 323 virtual bool parseAbsoluteExpression(int64_t &Res) = 0; 324 325 /// Ensure that we have a valid section set in the streamer. Otherwise, 326 /// report an error and switch to .text. 327 /// \return - False on success. 328 virtual bool checkForValidSection() = 0; 329 330 /// Parse an arbitrary expression of a specified parenthesis depth, 331 /// assuming that the initial '(' characters have already been consumed. 332 /// 333 /// \param ParenDepth - Specifies how many trailing expressions outside the 334 /// current parentheses we have to parse. 335 /// \param Res - The value of the expression. The result is undefined 336 /// on error. 337 /// \return - False on success. 338 virtual bool parseParenExprOfDepth(unsigned ParenDepth, const MCExpr *&Res, 339 SMLoc &EndLoc) = 0; 340 }; 341 342 /// Create an MCAsmParser instance for parsing assembly similar to gas syntax 343 MCAsmParser *createMCAsmParser(SourceMgr &, MCContext &, MCStreamer &, 344 const MCAsmInfo &, unsigned CB = 0); 345 346 /// Create an MCAsmParser instance for parsing Microsoft MASM-style assembly 347 MCAsmParser *createMCMasmParser(SourceMgr &, MCContext &, MCStreamer &, 348 const MCAsmInfo &, unsigned CB = 0); 349 350 } // end namespace llvm 351 352 #endif // LLVM_MC_MCPARSER_MCASMPARSER_H 353