lib/Target/WebAssembly/AsmParser/WebAssemblyAsmParser.cpp

   1 //==- WebAssemblyAsmParser.cpp - Assembler for WebAssembly -*- 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 /// \file
  10 /// This file is part of the WebAssembly Assembler.
  11 ///
  12 /// It contains code to translate a parsed .s file into MCInsts.
  13 ///
  14 //===----------------------------------------------------------------------===//
  15
  16 #include "MCTargetDesc/WebAssemblyMCTargetDesc.h"
  17 #include "MCTargetDesc/WebAssemblyTargetStreamer.h"
  18 #include "TargetInfo/WebAssemblyTargetInfo.h"
  19 #include "WebAssembly.h"
  20 #include "llvm/MC/MCContext.h"
  21 #include "llvm/MC/MCExpr.h"
  22 #include "llvm/MC/MCInst.h"
  23 #include "llvm/MC/MCInstrInfo.h"
  24 #include "llvm/MC/MCParser/MCParsedAsmOperand.h"
  25 #include "llvm/MC/MCParser/MCTargetAsmParser.h"
  26 #include "llvm/MC/MCSectionWasm.h"
  27 #include "llvm/MC/MCStreamer.h"
  28 #include "llvm/MC/MCSubtargetInfo.h"
  29 #include "llvm/MC/MCSymbol.h"
  30 #include "llvm/MC/MCSymbolWasm.h"
  31 #include "llvm/Support/Endian.h"
  32 #include "llvm/Support/TargetRegistry.h"
  33
  34 using namespace llvm;
  35
  36 #define DEBUG_TYPE "wasm-asm-parser"
  37
  38 namespace {
  39
  40 /// WebAssemblyOperand - Instances of this class represent the operands in a
  41 /// parsed WASM machine instruction.
  42 struct WebAssemblyOperand : public MCParsedAsmOperand {
  43   enum KindTy { Token, Integer, Float, Symbol, BrList } Kind;
  44
  45   SMLoc StartLoc, EndLoc;
  46
  47   struct TokOp {
  48     StringRef Tok;
  49   };
  50
  51   struct IntOp {
  52     int64_t Val;
  53   };
  54
  55   struct FltOp {
  56     double Val;
  57   };
  58
  59   struct SymOp {
  60     const MCExpr *Exp;
  61   };
  62
  63   struct BrLOp {
  64     std::vector<unsigned> List;
  65   };
  66
  67   union {
  68     struct TokOp Tok;
  69     struct IntOp Int;
  70     struct FltOp Flt;
  71     struct SymOp Sym;
  72     struct BrLOp BrL;
  73   };
  74
  75   WebAssemblyOperand(KindTy K, SMLoc Start, SMLoc End, TokOp T)
  76       : Kind(K), StartLoc(Start), EndLoc(End), Tok(T) {}
  77   WebAssemblyOperand(KindTy K, SMLoc Start, SMLoc End, IntOp I)
  78       : Kind(K), StartLoc(Start), EndLoc(End), Int(I) {}
  79   WebAssemblyOperand(KindTy K, SMLoc Start, SMLoc End, FltOp F)
  80       : Kind(K), StartLoc(Start), EndLoc(End), Flt(F) {}
  81   WebAssemblyOperand(KindTy K, SMLoc Start, SMLoc End, SymOp S)
  82       : Kind(K), StartLoc(Start), EndLoc(End), Sym(S) {}
  83   WebAssemblyOperand(KindTy K, SMLoc Start, SMLoc End)
  84       : Kind(K), StartLoc(Start), EndLoc(End), BrL() {}
  85
  86   ~WebAssemblyOperand() {
  87     if (isBrList())
  88       BrL.~BrLOp();
  89   }
  90
  91   bool isToken() const override { return Kind == Token; }
  92   bool isImm() const override { return Kind == Integer || Kind == Symbol; }
  93   bool isFPImm() const { return Kind == Float; }
  94   bool isMem() const override { return false; }
  95   bool isReg() const override { return false; }
  96   bool isBrList() const { return Kind == BrList; }
  97
  98   unsigned getReg() const override {
  99     llvm_unreachable("Assembly inspects a register operand");
 100     return 0;
 101   }
 102
 103   StringRef getToken() const {
 104     assert(isToken());
 105     return Tok.Tok;
 106   }
 107
 108   SMLoc getStartLoc() const override { return StartLoc; }
 109   SMLoc getEndLoc() const override { return EndLoc; }
 110
 111   void addRegOperands(MCInst &, unsigned) const {
 112     // Required by the assembly matcher.
 113     llvm_unreachable("Assembly matcher creates register operands");
 114   }
 115
 116   void addImmOperands(MCInst &Inst, unsigned N) const {
 117     assert(N == 1 && "Invalid number of operands!");
 118     if (Kind == Integer)
 119       Inst.addOperand(MCOperand::createImm(Int.Val));
 120     else if (Kind == Symbol)
 121       Inst.addOperand(MCOperand::createExpr(Sym.Exp));
 122     else
 123       llvm_unreachable("Should be integer immediate or symbol!");
 124   }
 125
 126   void addFPImmOperands(MCInst &Inst, unsigned N) const {
 127     assert(N == 1 && "Invalid number of operands!");
 128     if (Kind == Float)
 129       Inst.addOperand(MCOperand::createFPImm(Flt.Val));
 130     else
 131       llvm_unreachable("Should be float immediate!");
 132   }
 133
 134   void addBrListOperands(MCInst &Inst, unsigned N) const {
 135     assert(N == 1 && isBrList() && "Invalid BrList!");
 136     for (auto Br : BrL.List)
 137       Inst.addOperand(MCOperand::createImm(Br));
 138   }
 139
 140   void print(raw_ostream &OS) const override {
 141     switch (Kind) {
 142     case Token:
 143       OS << "Tok:" << Tok.Tok;
 144       break;
 145     case Integer:
 146       OS << "Int:" << Int.Val;
 147       break;
 148     case Float:
 149       OS << "Flt:" << Flt.Val;
 150       break;
 151     case Symbol:
 152       OS << "Sym:" << Sym.Exp;
 153       break;
 154     case BrList:
 155       OS << "BrList:" << BrL.List.size();
 156       break;
 157     }
 158   }
 159 };
 160
 161 class WebAssemblyAsmParser final : public MCTargetAsmParser {
 162   MCAsmParser &Parser;
 163   MCAsmLexer &Lexer;
 164
 165   // Much like WebAssemblyAsmPrinter in the backend, we have to own these.
 166   std::vector<std::unique_ptr<wasm::WasmSignature>> Signatures;
 167
 168   // Order of labels, directives and instructions in a .s file have no
 169   // syntactical enforcement. This class is a callback from the actual parser,
 170   // and yet we have to be feeding data to the streamer in a very particular
 171   // order to ensure a correct binary encoding that matches the regular backend
 172   // (the streamer does not enforce this). This "state machine" enum helps
 173   // guarantee that correct order.
 174   enum ParserState {
 175     FileStart,
 176     Label,
 177     FunctionStart,
 178     FunctionLocals,
 179     Instructions,
 180     EndFunction,
 181     DataSection,
 182   } CurrentState = FileStart;
 183
 184   // For ensuring blocks are properly nested.
 185   enum NestingType {
 186     Function,
 187     Block,
 188     Loop,
 189     Try,
 190     If,
 191     Else,
 192     Undefined,
 193   };
 194   std::vector<NestingType> NestingStack;
 195
 196   // We track this to see if a .functype following a label is the same,
 197   // as this is how we recognize the start of a function.
 198   MCSymbol *LastLabel = nullptr;
 199   MCSymbol *LastFunctionLabel = nullptr;
 200
 201 public:
 202   WebAssemblyAsmParser(const MCSubtargetInfo &STI, MCAsmParser &Parser,
 203                        const MCInstrInfo &MII, const MCTargetOptions &Options)
 204       : MCTargetAsmParser(Options, STI, MII), Parser(Parser),
 205         Lexer(Parser.getLexer()) {
 206     setAvailableFeatures(ComputeAvailableFeatures(STI.getFeatureBits()));
 207   }
 208
 209 #define GET_ASSEMBLER_HEADER
 210 #include "WebAssemblyGenAsmMatcher.inc"
 211
 212   // TODO: This is required to be implemented, but appears unused.
 213   bool ParseRegister(unsigned & /*RegNo*/, SMLoc & /*StartLoc*/,
 214                      SMLoc & /*EndLoc*/) override {
 215     llvm_unreachable("ParseRegister is not implemented.");
 216   }
 217
 218   bool error(const Twine &Msg, const AsmToken &Tok) {
 219     return Parser.Error(Tok.getLoc(), Msg + Tok.getString());
 220   }
 221
 222   bool error(const Twine &Msg) {
 223     return Parser.Error(Lexer.getTok().getLoc(), Msg);
 224   }
 225
 226   void addSignature(std::unique_ptr<wasm::WasmSignature> &&Sig) {
 227     Signatures.push_back(std::move(Sig));
 228   }
 229
 230   std::pair<StringRef, StringRef> nestingString(NestingType NT) {
 231     switch (NT) {
 232     case Function:
 233       return {"function", "end_function"};
 234     case Block:
 235       return {"block", "end_block"};
 236     case Loop:
 237       return {"loop", "end_loop"};
 238     case Try:
 239       return {"try", "end_try"};
 240     case If:
 241       return {"if", "end_if"};
 242     case Else:
 243       return {"else", "end_if"};
 244     default:
 245       llvm_unreachable("unknown NestingType");
 246     }
 247   }
 248
 249   void push(NestingType NT) { NestingStack.push_back(NT); }
 250
 251   bool pop(StringRef Ins, NestingType NT1, NestingType NT2 = Undefined) {
 252     if (NestingStack.empty())
 253       return error(Twine("End of block construct with no start: ") + Ins);
 254     auto Top = NestingStack.back();
 255     if (Top != NT1 && Top != NT2)
 256       return error(Twine("Block construct type mismatch, expected: ") +
 257                    nestingString(Top).second + ", instead got: " + Ins);
 258     NestingStack.pop_back();
 259     return false;
 260   }
 261
 262   bool ensureEmptyNestingStack() {
 263     auto Err = !NestingStack.empty();
 264     while (!NestingStack.empty()) {
 265       error(Twine("Unmatched block construct(s) at function end: ") +
 266             nestingString(NestingStack.back()).first);
 267       NestingStack.pop_back();
 268     }
 269     return Err;
 270   }
 271
 272   bool isNext(AsmToken::TokenKind Kind) {
 273     auto Ok = Lexer.is(Kind);
 274     if (Ok)
 275       Parser.Lex();
 276     return Ok;
 277   }
 278
 279   bool expect(AsmToken::TokenKind Kind, const char *KindName) {
 280     if (!isNext(Kind))
 281       return error(std::string("Expected ") + KindName + ", instead got: ",
 282                    Lexer.getTok());
 283     return false;
 284   }
 285
 286   StringRef expectIdent() {
 287     if (!Lexer.is(AsmToken::Identifier)) {
 288       error("Expected identifier, got: ", Lexer.getTok());
 289       return StringRef();
 290     }
 291     auto Name = Lexer.getTok().getString();
 292     Parser.Lex();
 293     return Name;
 294   }
 295
 296   Optional<wasm::ValType> parseType(const StringRef &Type) {
 297     // FIXME: can't use StringSwitch because wasm::ValType doesn't have a
 298     // "invalid" value.
 299     if (Type == "i32")
 300       return wasm::ValType::I32;
 301     if (Type == "i64")
 302       return wasm::ValType::I64;
 303     if (Type == "f32")
 304       return wasm::ValType::F32;
 305     if (Type == "f64")
 306       return wasm::ValType::F64;
 307     if (Type == "v128" || Type == "i8x16" || Type == "i16x8" ||
 308         Type == "i32x4" || Type == "i64x2" || Type == "f32x4" ||
 309         Type == "f64x2")
 310       return wasm::ValType::V128;
 311     if (Type == "exnref")
 312       return wasm::ValType::EXNREF;
 313     return Optional<wasm::ValType>();
 314   }
 315
 316   WebAssembly::ExprType parseBlockType(StringRef ID) {
 317     return StringSwitch<WebAssembly::ExprType>(ID)
 318         .Case("i32", WebAssembly::ExprType::I32)
 319         .Case("i64", WebAssembly::ExprType::I64)
 320         .Case("f32", WebAssembly::ExprType::F32)
 321         .Case("f64", WebAssembly::ExprType::F64)
 322         .Case("v128", WebAssembly::ExprType::V128)
 323         .Case("exnref", WebAssembly::ExprType::Exnref)
 324         .Case("void", WebAssembly::ExprType::Void)
 325         .Default(WebAssembly::ExprType::Invalid);
 326   }
 327
 328   bool parseRegTypeList(SmallVectorImpl<wasm::ValType> &Types) {
 329     while (Lexer.is(AsmToken::Identifier)) {
 330       auto Type = parseType(Lexer.getTok().getString());
 331       if (!Type)
 332         return error("unknown type: ", Lexer.getTok());
 333       Types.push_back(Type.getValue());
 334       Parser.Lex();
 335       if (!isNext(AsmToken::Comma))
 336         break;
 337     }
 338     return false;
 339   }
 340
 341   void parseSingleInteger(bool IsNegative, OperandVector &Operands) {
 342     auto &Int = Lexer.getTok();
 343     int64_t Val = Int.getIntVal();
 344     if (IsNegative)
 345       Val = -Val;
 346     Operands.push_back(std::make_unique<WebAssemblyOperand>(
 347         WebAssemblyOperand::Integer, Int.getLoc(), Int.getEndLoc(),
 348         WebAssemblyOperand::IntOp{Val}));
 349     Parser.Lex();
 350   }
 351
 352   bool parseSingleFloat(bool IsNegative, OperandVector &Operands) {
 353     auto &Flt = Lexer.getTok();
 354     double Val;
 355     if (Flt.getString().getAsDouble(Val, false))
 356       return error("Cannot parse real: ", Flt);
 357     if (IsNegative)
 358       Val = -Val;
 359     Operands.push_back(std::make_unique<WebAssemblyOperand>(
 360         WebAssemblyOperand::Float, Flt.getLoc(), Flt.getEndLoc(),
 361         WebAssemblyOperand::FltOp{Val}));
 362     Parser.Lex();
 363     return false;
 364   }
 365
 366   bool parseSpecialFloatMaybe(bool IsNegative, OperandVector &Operands) {
 367     if (Lexer.isNot(AsmToken::Identifier))
 368       return true;
 369     auto &Flt = Lexer.getTok();
 370     auto S = Flt.getString();
 371     double Val;
 372     if (S.compare_lower("infinity") == 0) {
 373       Val = std::numeric_limits<double>::infinity();
 374     } else if (S.compare_lower("nan") == 0) {
 375       Val = std::numeric_limits<double>::quiet_NaN();
 376     } else {
 377       return true;
 378     }
 379     if (IsNegative)
 380       Val = -Val;
 381     Operands.push_back(std::make_unique<WebAssemblyOperand>(
 382         WebAssemblyOperand::Float, Flt.getLoc(), Flt.getEndLoc(),
 383         WebAssemblyOperand::FltOp{Val}));
 384     Parser.Lex();
 385     return false;
 386   }
 387
 388   bool checkForP2AlignIfLoadStore(OperandVector &Operands, StringRef InstName) {
 389     // FIXME: there is probably a cleaner way to do this.
 390     auto IsLoadStore = InstName.find(".load") != StringRef::npos ||
 391                        InstName.find(".store") != StringRef::npos;
 392     auto IsAtomic = InstName.find("atomic.") != StringRef::npos;
 393     if (IsLoadStore || IsAtomic) {
 394       // Parse load/store operands of the form: offset:p2align=align
 395       if (IsLoadStore && isNext(AsmToken::Colon)) {
 396         auto Id = expectIdent();
 397         if (Id != "p2align")
 398           return error("Expected p2align, instead got: " + Id);
 399         if (expect(AsmToken::Equal, "="))
 400           return true;
 401         if (!Lexer.is(AsmToken::Integer))
 402           return error("Expected integer constant");
 403         parseSingleInteger(false, Operands);
 404       } else {
 405         // Alignment not specified (or atomics, must use default alignment).
 406         // We can't just call WebAssembly::GetDefaultP2Align since we don't have
 407         // an opcode until after the assembly matcher, so set a default to fix
 408         // up later.
 409         auto Tok = Lexer.getTok();
 410         Operands.push_back(std::make_unique<WebAssemblyOperand>(
 411             WebAssemblyOperand::Integer, Tok.getLoc(), Tok.getEndLoc(),
 412             WebAssemblyOperand::IntOp{-1}));
 413       }
 414     }
 415     return false;
 416   }
 417
 418   void addBlockTypeOperand(OperandVector &Operands, SMLoc NameLoc,
 419                            WebAssembly::ExprType BT) {
 420     Operands.push_back(std::make_unique<WebAssemblyOperand>(
 421         WebAssemblyOperand::Integer, NameLoc, NameLoc,
 422         WebAssemblyOperand::IntOp{static_cast<int64_t>(BT)}));
 423   }
 424
 425   bool ParseInstruction(ParseInstructionInfo & /*Info*/, StringRef Name,
 426                         SMLoc NameLoc, OperandVector &Operands) override {
 427     // Note: Name does NOT point into the sourcecode, but to a local, so
 428     // use NameLoc instead.
 429     Name = StringRef(NameLoc.getPointer(), Name.size());
 430
 431     // WebAssembly has instructions with / in them, which AsmLexer parses
 432     // as seperate tokens, so if we find such tokens immediately adjacent (no
 433     // whitespace), expand the name to include them:
 434     for (;;) {
 435       auto &Sep = Lexer.getTok();
 436       if (Sep.getLoc().getPointer() != Name.end() ||
 437           Sep.getKind() != AsmToken::Slash)
 438         break;
 439       // Extend name with /
 440       Name = StringRef(Name.begin(), Name.size() + Sep.getString().size());
 441       Parser.Lex();
 442       // We must now find another identifier, or error.
 443       auto &Id = Lexer.getTok();
 444       if (Id.getKind() != AsmToken::Identifier ||
 445           Id.getLoc().getPointer() != Name.end())
 446         return error("Incomplete instruction name: ", Id);
 447       Name = StringRef(Name.begin(), Name.size() + Id.getString().size());
 448       Parser.Lex();
 449     }
 450
 451     // Now construct the name as first operand.
 452     Operands.push_back(std::make_unique<WebAssemblyOperand>(
 453         WebAssemblyOperand::Token, NameLoc, SMLoc::getFromPointer(Name.end()),
 454         WebAssemblyOperand::TokOp{Name}));
 455
 456     // If this instruction is part of a control flow structure, ensure
 457     // proper nesting.
 458     bool ExpectBlockType = false;
 459     if (Name == "block") {
 460       push(Block);
 461       ExpectBlockType = true;
 462     } else if (Name == "loop") {
 463       push(Loop);
 464       ExpectBlockType = true;
 465     } else if (Name == "try") {
 466       push(Try);
 467       ExpectBlockType = true;
 468     } else if (Name == "if") {
 469       push(If);
 470       ExpectBlockType = true;
 471     } else if (Name == "else") {
 472       if (pop(Name, If))
 473         return true;
 474       push(Else);
 475     } else if (Name == "catch") {
 476       if (pop(Name, Try))
 477         return true;
 478       push(Try);
 479     } else if (Name == "end_if") {
 480       if (pop(Name, If, Else))
 481         return true;
 482     } else if (Name == "end_try") {
 483       if (pop(Name, Try))
 484         return true;
 485     } else if (Name == "end_loop") {
 486       if (pop(Name, Loop))
 487         return true;
 488     } else if (Name == "end_block") {
 489       if (pop(Name, Block))
 490         return true;
 491     } else if (Name == "end_function") {
 492       ensureLocals(getStreamer());
 493       CurrentState = EndFunction;
 494       if (pop(Name, Function) || ensureEmptyNestingStack())
 495         return true;
 496     } else if (Name == "call_indirect" || Name == "return_call_indirect") {
 497       // This has a special TYPEINDEX operand which in text we
 498       // represent as a signature, such that we can re-build this signature,
 499       // attach it to an anonymous symbol, which is what WasmObjectWriter
 500       // expects to be able to recreate the actual unique-ified type indices.
 501       auto Loc = Parser.getTok();
 502       auto Signature = std::make_unique<wasm::WasmSignature>();
 503       if (parseSignature(Signature.get()))
 504         return true;
 505       auto &Ctx = getStreamer().getContext();
 506       // The "true" here will cause this to be a nameless symbol.
 507       MCSymbol *Sym = Ctx.createTempSymbol("typeindex", true);
 508       auto *WasmSym = cast<MCSymbolWasm>(Sym);
 509       WasmSym->setSignature(Signature.get());
 510       addSignature(std::move(Signature));
 511       WasmSym->setType(wasm::WASM_SYMBOL_TYPE_FUNCTION);
 512       const MCExpr *Expr = MCSymbolRefExpr::create(
 513           WasmSym, MCSymbolRefExpr::VK_WASM_TYPEINDEX, Ctx);
 514       Operands.push_back(std::make_unique<WebAssemblyOperand>(
 515           WebAssemblyOperand::Symbol, Loc.getLoc(), Loc.getEndLoc(),
 516           WebAssemblyOperand::SymOp{Expr}));
 517     }
 518
 519     while (Lexer.isNot(AsmToken::EndOfStatement)) {
 520       auto &Tok = Lexer.getTok();
 521       switch (Tok.getKind()) {
 522       case AsmToken::Identifier: {
 523         if (!parseSpecialFloatMaybe(false, Operands))
 524           break;
 525         auto &Id = Lexer.getTok();
 526         if (ExpectBlockType) {
 527           // Assume this identifier is a block_type.
 528           auto BT = parseBlockType(Id.getString());
 529           if (BT == WebAssembly::ExprType::Invalid)
 530             return error("Unknown block type: ", Id);
 531           addBlockTypeOperand(Operands, NameLoc, BT);
 532           Parser.Lex();
 533         } else {
 534           // Assume this identifier is a label.
 535           const MCExpr *Val;
 536           SMLoc End;
 537           if (Parser.parseExpression(Val, End))
 538             return error("Cannot parse symbol: ", Lexer.getTok());
 539           Operands.push_back(std::make_unique<WebAssemblyOperand>(
 540               WebAssemblyOperand::Symbol, Id.getLoc(), Id.getEndLoc(),
 541               WebAssemblyOperand::SymOp{Val}));
 542           if (checkForP2AlignIfLoadStore(Operands, Name))
 543             return true;
 544         }
 545         break;
 546       }
 547       case AsmToken::Minus:
 548         Parser.Lex();
 549         if (Lexer.is(AsmToken::Integer)) {
 550           parseSingleInteger(true, Operands);
 551           if (checkForP2AlignIfLoadStore(Operands, Name))
 552             return true;
 553         } else if(Lexer.is(AsmToken::Real)) {
 554           if (parseSingleFloat(true, Operands))
 555             return true;
 556         } else if (!parseSpecialFloatMaybe(true, Operands)) {
 557         } else {
 558           return error("Expected numeric constant instead got: ",
 559                        Lexer.getTok());
 560         }
 561         break;
 562       case AsmToken::Integer:
 563         parseSingleInteger(false, Operands);
 564         if (checkForP2AlignIfLoadStore(Operands, Name))
 565           return true;
 566         break;
 567       case AsmToken::Real: {
 568         if (parseSingleFloat(false, Operands))
 569           return true;
 570         break;
 571       }
 572       case AsmToken::LCurly: {
 573         Parser.Lex();
 574         auto Op = std::make_unique<WebAssemblyOperand>(
 575             WebAssemblyOperand::BrList, Tok.getLoc(), Tok.getEndLoc());
 576         if (!Lexer.is(AsmToken::RCurly))
 577           for (;;) {
 578             Op->BrL.List.push_back(Lexer.getTok().getIntVal());
 579             expect(AsmToken::Integer, "integer");
 580             if (!isNext(AsmToken::Comma))
 581               break;
 582           }
 583         expect(AsmToken::RCurly, "}");
 584         Operands.push_back(std::move(Op));
 585         break;
 586       }
 587       default:
 588         return error("Unexpected token in operand: ", Tok);
 589       }
 590       if (Lexer.isNot(AsmToken::EndOfStatement)) {
 591         if (expect(AsmToken::Comma, ","))
 592           return true;
 593       }
 594     }
 595     if (ExpectBlockType && Operands.size() == 1) {
 596       // Support blocks with no operands as default to void.
 597       addBlockTypeOperand(Operands, NameLoc, WebAssembly::ExprType::Void);
 598     }
 599     Parser.Lex();
 600     return false;
 601   }
 602
 603   void onLabelParsed(MCSymbol *Symbol) override {
 604     LastLabel = Symbol;
 605     CurrentState = Label;
 606   }
 607
 608   bool parseSignature(wasm::WasmSignature *Signature) {
 609     if (expect(AsmToken::LParen, "("))
 610       return true;
 611     if (parseRegTypeList(Signature->Params))
 612       return true;
 613     if (expect(AsmToken::RParen, ")"))
 614       return true;
 615     if (expect(AsmToken::MinusGreater, "->"))
 616       return true;
 617     if (expect(AsmToken::LParen, "("))
 618       return true;
 619     if (parseRegTypeList(Signature->Returns))
 620       return true;
 621     if (expect(AsmToken::RParen, ")"))
 622       return true;
 623     return false;
 624   }
 625
 626   bool CheckDataSection() {
 627     if (CurrentState != DataSection) {
 628       auto WS = cast<MCSectionWasm>(getStreamer().getCurrentSection().first);
 629       if (WS && WS->getKind().isText())
 630         return error("data directive must occur in a data segment: ",
 631                      Lexer.getTok());
 632     }
 633     CurrentState = DataSection;
 634     return false;
 635   }
 636
 637   // This function processes wasm-specific directives streamed to
 638   // WebAssemblyTargetStreamer, all others go to the generic parser
 639   // (see WasmAsmParser).
 640   bool ParseDirective(AsmToken DirectiveID) override {
 641     // This function has a really weird return value behavior that is different
 642     // from all the other parsing functions:
 643     // - return true && no tokens consumed -> don't know this directive / let
 644     //   the generic parser handle it.
 645     // - return true && tokens consumed -> a parsing error occurred.
 646     // - return false -> processed this directive successfully.
 647     assert(DirectiveID.getKind() == AsmToken::Identifier);
 648     auto &Out = getStreamer();
 649     auto &TOut =
 650         reinterpret_cast<WebAssemblyTargetStreamer &>(*Out.getTargetStreamer());
 651     auto &Ctx = Out.getContext();
 652
 653     // TODO: any time we return an error, at least one token must have been
 654     // consumed, otherwise this will not signal an error to the caller.
 655     if (DirectiveID.getString() == ".globaltype") {
 656       auto SymName = expectIdent();
 657       if (SymName.empty())
 658         return true;
 659       if (expect(AsmToken::Comma, ","))
 660         return true;
 661       auto TypeTok = Lexer.getTok();
 662       auto TypeName = expectIdent();
 663       if (TypeName.empty())
 664         return true;
 665       auto Type = parseType(TypeName);
 666       if (!Type)
 667         return error("Unknown type in .globaltype directive: ", TypeTok);
 668       // Now set this symbol with the correct type.
 669       auto WasmSym = cast<MCSymbolWasm>(Ctx.getOrCreateSymbol(SymName));
 670       WasmSym->setType(wasm::WASM_SYMBOL_TYPE_GLOBAL);
 671       WasmSym->setGlobalType(
 672           wasm::WasmGlobalType{uint8_t(Type.getValue()), true});
 673       // And emit the directive again.
 674       TOut.emitGlobalType(WasmSym);
 675       return expect(AsmToken::EndOfStatement, "EOL");
 676     }
 677
 678     if (DirectiveID.getString() == ".functype") {
 679       // This code has to send things to the streamer similar to
 680       // WebAssemblyAsmPrinter::EmitFunctionBodyStart.
 681       // TODO: would be good to factor this into a common function, but the
 682       // assembler and backend really don't share any common code, and this code
 683       // parses the locals seperately.
 684       auto SymName = expectIdent();
 685       if (SymName.empty())
 686         return true;
 687       auto WasmSym = cast<MCSymbolWasm>(Ctx.getOrCreateSymbol(SymName));
 688       if (CurrentState == Label && WasmSym == LastLabel) {
 689         // This .functype indicates a start of a function.
 690         if (ensureEmptyNestingStack())
 691           return true;
 692         CurrentState = FunctionStart;
 693         LastFunctionLabel = LastLabel;
 694         push(Function);
 695       }
 696       auto Signature = std::make_unique<wasm::WasmSignature>();
 697       if (parseSignature(Signature.get()))
 698         return true;
 699       WasmSym->setSignature(Signature.get());
 700       addSignature(std::move(Signature));
 701       WasmSym->setType(wasm::WASM_SYMBOL_TYPE_FUNCTION);
 702       TOut.emitFunctionType(WasmSym);
 703       // TODO: backend also calls TOut.emitIndIdx, but that is not implemented.
 704       return expect(AsmToken::EndOfStatement, "EOL");
 705     }
 706
 707     if (DirectiveID.getString() == ".eventtype") {
 708       auto SymName = expectIdent();
 709       if (SymName.empty())
 710         return true;
 711       auto WasmSym = cast<MCSymbolWasm>(Ctx.getOrCreateSymbol(SymName));
 712       auto Signature = std::make_unique<wasm::WasmSignature>();
 713       if (parseRegTypeList(Signature->Params))
 714         return true;
 715       WasmSym->setSignature(Signature.get());
 716       addSignature(std::move(Signature));
 717       WasmSym->setType(wasm::WASM_SYMBOL_TYPE_EVENT);
 718       TOut.emitEventType(WasmSym);
 719       // TODO: backend also calls TOut.emitIndIdx, but that is not implemented.
 720       return expect(AsmToken::EndOfStatement, "EOL");
 721     }
 722
 723     if (DirectiveID.getString() == ".local") {
 724       if (CurrentState != FunctionStart)
 725         return error(".local directive should follow the start of a function",
 726                      Lexer.getTok());
 727       SmallVector<wasm::ValType, 4> Locals;
 728       if (parseRegTypeList(Locals))
 729         return true;
 730       TOut.emitLocal(Locals);
 731       CurrentState = FunctionLocals;
 732       return expect(AsmToken::EndOfStatement, "EOL");
 733     }
 734
 735     if (DirectiveID.getString() == ".int8" ||
 736         DirectiveID.getString() == ".int16" ||
 737         DirectiveID.getString() == ".int32" ||
 738         DirectiveID.getString() == ".int64") {
 739       if (CheckDataSection()) return true;
 740       const MCExpr *Val;
 741       SMLoc End;
 742       if (Parser.parseExpression(Val, End))
 743         return error("Cannot parse .int expression: ", Lexer.getTok());
 744       size_t NumBits = 0;
 745       DirectiveID.getString().drop_front(4).getAsInteger(10, NumBits);
 746       Out.EmitValue(Val, NumBits / 8, End);
 747       return expect(AsmToken::EndOfStatement, "EOL");
 748     }
 749
 750     if (DirectiveID.getString() == ".asciz") {
 751       if (CheckDataSection()) return true;
 752       std::string S;
 753       if (Parser.parseEscapedString(S))
 754         return error("Cannot parse string constant: ", Lexer.getTok());
 755       Out.EmitBytes(StringRef(S.c_str(), S.length() + 1));
 756       return expect(AsmToken::EndOfStatement, "EOL");
 757     }
 758
 759     return true; // We didn't process this directive.
 760   }
 761
 762   // Called either when the first instruction is parsed of the function ends.
 763   void ensureLocals(MCStreamer &Out) {
 764     if (CurrentState == FunctionStart) {
 765       // We haven't seen a .local directive yet. The streamer requires locals to
 766       // be encoded as a prelude to the instructions, so emit an empty list of
 767       // locals here.
 768       auto &TOut = reinterpret_cast<WebAssemblyTargetStreamer &>(
 769           *Out.getTargetStreamer());
 770       TOut.emitLocal(SmallVector<wasm::ValType, 0>());
 771       CurrentState = FunctionLocals;
 772     }
 773   }
 774
 775   bool MatchAndEmitInstruction(SMLoc IDLoc, unsigned & /*Opcode*/,
 776                                OperandVector &Operands, MCStreamer &Out,
 777                                uint64_t &ErrorInfo,
 778                                bool MatchingInlineAsm) override {
 779     MCInst Inst;
 780     Inst.setLoc(IDLoc);
 781     unsigned MatchResult =
 782         MatchInstructionImpl(Operands, Inst, ErrorInfo, MatchingInlineAsm);
 783     switch (MatchResult) {
 784     case Match_Success: {
 785       ensureLocals(Out);
 786       // Fix unknown p2align operands.
 787       auto Align = WebAssembly::GetDefaultP2AlignAny(Inst.getOpcode());
 788       if (Align != -1U) {
 789         auto &Op0 = Inst.getOperand(0);
 790         if (Op0.getImm() == -1)
 791           Op0.setImm(Align);
 792       }
 793       Out.EmitInstruction(Inst, getSTI());
 794       if (CurrentState == EndFunction) {
 795         onEndOfFunction();
 796       } else {
 797         CurrentState = Instructions;
 798       }
 799       return false;
 800     }
 801     case Match_MissingFeature:
 802       return Parser.Error(
 803           IDLoc, "instruction requires a WASM feature not currently enabled");
 804     case Match_MnemonicFail:
 805       return Parser.Error(IDLoc, "invalid instruction");
 806     case Match_NearMisses:
 807       return Parser.Error(IDLoc, "ambiguous instruction");
 808     case Match_InvalidTiedOperand:
 809     case Match_InvalidOperand: {
 810       SMLoc ErrorLoc = IDLoc;
 811       if (ErrorInfo != ~0ULL) {
 812         if (ErrorInfo >= Operands.size())
 813           return Parser.Error(IDLoc, "too few operands for instruction");
 814         ErrorLoc = Operands[ErrorInfo]->getStartLoc();
 815         if (ErrorLoc == SMLoc())
 816           ErrorLoc = IDLoc;
 817       }
 818       return Parser.Error(ErrorLoc, "invalid operand for instruction");
 819     }
 820     }
 821     llvm_unreachable("Implement any new match types added!");
 822   }
 823
 824   void doBeforeLabelEmit(MCSymbol *Symbol) override {
 825     // Start a new section for the next function automatically, since our
 826     // object writer expects each function to have its own section. This way
 827     // The user can't forget this "convention".
 828     auto SymName = Symbol->getName();
 829     if (SymName.startswith(".L"))
 830       return; // Local Symbol.
 831     // Only create a new text section if we're already in one.
 832     auto CWS = cast<MCSectionWasm>(getStreamer().getCurrentSection().first);
 833     if (!CWS || !CWS->getKind().isText())
 834       return;
 835     auto SecName = ".text." + SymName;
 836     auto WS = getContext().getWasmSection(SecName, SectionKind::getText());
 837     getStreamer().SwitchSection(WS);
 838   }
 839
 840   void onEndOfFunction() {
 841     // Automatically output a .size directive, so it becomes optional for the
 842     // user.
 843     if (!LastFunctionLabel) return;
 844     auto TempSym = getContext().createLinkerPrivateTempSymbol();
 845     getStreamer().EmitLabel(TempSym);
 846     auto Start = MCSymbolRefExpr::create(LastFunctionLabel, getContext());
 847     auto End = MCSymbolRefExpr::create(TempSym, getContext());
 848     auto Expr =
 849         MCBinaryExpr::create(MCBinaryExpr::Sub, End, Start, getContext());
 850     getStreamer().emitELFSize(LastFunctionLabel, Expr);
 851   }
 852
 853   void onEndOfFile() override { ensureEmptyNestingStack(); }
 854 };
 855 } // end anonymous namespace
 856
 857 // Force static initialization.
 858 extern "C" void LLVMInitializeWebAssemblyAsmParser() {
 859   RegisterMCAsmParser<WebAssemblyAsmParser> X(getTheWebAssemblyTarget32());
 860   RegisterMCAsmParser<WebAssemblyAsmParser> Y(getTheWebAssemblyTarget64());
 861 }
 862
 863 #define GET_REGISTER_MATCHER
 864 #define GET_MATCHER_IMPLEMENTATION
 865 #include "WebAssemblyGenAsmMatcher.inc"