blob: eeccc3e64495faf98c049a523be255de1fd13137 [file] [log] [blame]
Derek Schuffe4825972018-03-20 20:06:35 +00001//==- WebAssemblyAsmParser.cpp - Assembler for WebAssembly -*- C++ -*-==//
2//
Chandler Carruth2946cd72019-01-19 08:50:56 +00003// 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
Derek Schuffe4825972018-03-20 20:06:35 +00006//
7//===----------------------------------------------------------------------===//
8///
9/// \file
Adrian Prantl5f8f34e42018-05-01 15:54:18 +000010/// This file is part of the WebAssembly Assembler.
Derek Schuffe4825972018-03-20 20:06:35 +000011///
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 "WebAssembly.h"
19#include "llvm/MC/MCContext.h"
Wouter van Oortmerssen0b3cf242019-02-04 18:03:11 +000020#include "llvm/MC/MCExpr.h"
Derek Schuffe4825972018-03-20 20:06:35 +000021#include "llvm/MC/MCInst.h"
22#include "llvm/MC/MCInstrInfo.h"
Heejin Ahnf208f632018-09-05 01:27:38 +000023#include "llvm/MC/MCParser/MCParsedAsmOperand.h"
24#include "llvm/MC/MCParser/MCTargetAsmParser.h"
Wouter van Oortmerssen0b3cf242019-02-04 18:03:11 +000025#include "llvm/MC/MCSectionWasm.h"
Heejin Ahnf208f632018-09-05 01:27:38 +000026#include "llvm/MC/MCStreamer.h"
Derek Schuffe4825972018-03-20 20:06:35 +000027#include "llvm/MC/MCSubtargetInfo.h"
28#include "llvm/MC/MCSymbol.h"
Wouter van Oortmerssende28b5d2018-11-02 22:04:33 +000029#include "llvm/MC/MCSymbolWasm.h"
Derek Schuffe4825972018-03-20 20:06:35 +000030#include "llvm/Support/Endian.h"
31#include "llvm/Support/TargetRegistry.h"
32
33using namespace llvm;
34
35#define DEBUG_TYPE "wasm-asm-parser"
36
Derek Schuffe4825972018-03-20 20:06:35 +000037namespace {
38
Derek Schuffe4825972018-03-20 20:06:35 +000039/// WebAssemblyOperand - Instances of this class represent the operands in a
40/// parsed WASM machine instruction.
41struct WebAssemblyOperand : public MCParsedAsmOperand {
Wouter van Oortmerssend3c544a2018-12-17 22:04:44 +000042 enum KindTy { Token, Integer, Float, Symbol, BrList } Kind;
Derek Schuffe4825972018-03-20 20:06:35 +000043
44 SMLoc StartLoc, EndLoc;
45
46 struct TokOp {
47 StringRef Tok;
48 };
49
Derek Schuffe4825972018-03-20 20:06:35 +000050 struct IntOp {
51 int64_t Val;
52 };
53
54 struct FltOp {
55 double Val;
56 };
57
58 struct SymOp {
59 const MCExpr *Exp;
60 };
61
Wouter van Oortmerssend3c544a2018-12-17 22:04:44 +000062 struct BrLOp {
63 std::vector<unsigned> List;
64 };
65
Derek Schuffe4825972018-03-20 20:06:35 +000066 union {
67 struct TokOp Tok;
Derek Schuffe4825972018-03-20 20:06:35 +000068 struct IntOp Int;
69 struct FltOp Flt;
70 struct SymOp Sym;
Wouter van Oortmerssend3c544a2018-12-17 22:04:44 +000071 struct BrLOp BrL;
Derek Schuffe4825972018-03-20 20:06:35 +000072 };
73
74 WebAssemblyOperand(KindTy K, SMLoc Start, SMLoc End, TokOp T)
Heejin Ahnf208f632018-09-05 01:27:38 +000075 : Kind(K), StartLoc(Start), EndLoc(End), Tok(T) {}
Derek Schuffe4825972018-03-20 20:06:35 +000076 WebAssemblyOperand(KindTy K, SMLoc Start, SMLoc End, IntOp I)
Heejin Ahnf208f632018-09-05 01:27:38 +000077 : Kind(K), StartLoc(Start), EndLoc(End), Int(I) {}
Derek Schuffe4825972018-03-20 20:06:35 +000078 WebAssemblyOperand(KindTy K, SMLoc Start, SMLoc End, FltOp F)
Heejin Ahnf208f632018-09-05 01:27:38 +000079 : Kind(K), StartLoc(Start), EndLoc(End), Flt(F) {}
Derek Schuffe4825972018-03-20 20:06:35 +000080 WebAssemblyOperand(KindTy K, SMLoc Start, SMLoc End, SymOp S)
Heejin Ahnf208f632018-09-05 01:27:38 +000081 : Kind(K), StartLoc(Start), EndLoc(End), Sym(S) {}
Wouter van Oortmerssend3c544a2018-12-17 22:04:44 +000082 WebAssemblyOperand(KindTy K, SMLoc Start, SMLoc End)
83 : Kind(K), StartLoc(Start), EndLoc(End), BrL() {}
84
85 ~WebAssemblyOperand() {
86 if (isBrList())
87 BrL.~BrLOp();
88 }
Derek Schuffe4825972018-03-20 20:06:35 +000089
90 bool isToken() const override { return Kind == Token; }
Heejin Ahnf208f632018-09-05 01:27:38 +000091 bool isImm() const override {
92 return Kind == Integer || Kind == Float || Kind == Symbol;
93 }
Derek Schuffe4825972018-03-20 20:06:35 +000094 bool isMem() const override { return false; }
Wouter van Oortmerssen8a9cb242018-08-27 15:45:51 +000095 bool isReg() const override { return false; }
Wouter van Oortmerssend3c544a2018-12-17 22:04:44 +000096 bool isBrList() const { return Kind == BrList; }
Derek Schuffe4825972018-03-20 20:06:35 +000097
98 unsigned getReg() const override {
Wouter van Oortmerssen8a9cb242018-08-27 15:45:51 +000099 llvm_unreachable("Assembly inspects a register operand");
100 return 0;
Derek Schuffe4825972018-03-20 20:06:35 +0000101 }
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
Wouter van Oortmerssen8a9cb242018-08-27 15:45:51 +0000111 void addRegOperands(MCInst &, unsigned) const {
112 // Required by the assembly matcher.
113 llvm_unreachable("Assembly matcher creates register operands");
Derek Schuffe4825972018-03-20 20:06:35 +0000114 }
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 == Float)
121 Inst.addOperand(MCOperand::createFPImm(Flt.Val));
122 else if (Kind == Symbol)
123 Inst.addOperand(MCOperand::createExpr(Sym.Exp));
124 else
125 llvm_unreachable("Should be immediate or symbol!");
126 }
127
Wouter van Oortmerssend3c544a2018-12-17 22:04:44 +0000128 void addBrListOperands(MCInst &Inst, unsigned N) const {
129 assert(N == 1 && isBrList() && "Invalid BrList!");
130 for (auto Br : BrL.List)
131 Inst.addOperand(MCOperand::createImm(Br));
132 }
133
Derek Schuffe4825972018-03-20 20:06:35 +0000134 void print(raw_ostream &OS) const override {
135 switch (Kind) {
136 case Token:
137 OS << "Tok:" << Tok.Tok;
138 break;
Derek Schuffe4825972018-03-20 20:06:35 +0000139 case Integer:
140 OS << "Int:" << Int.Val;
141 break;
142 case Float:
143 OS << "Flt:" << Flt.Val;
144 break;
145 case Symbol:
146 OS << "Sym:" << Sym.Exp;
147 break;
Wouter van Oortmerssend3c544a2018-12-17 22:04:44 +0000148 case BrList:
149 OS << "BrList:" << BrL.List.size();
150 break;
Derek Schuffe4825972018-03-20 20:06:35 +0000151 }
152 }
153};
154
155class WebAssemblyAsmParser final : public MCTargetAsmParser {
156 MCAsmParser &Parser;
157 MCAsmLexer &Lexer;
Derek Schuffe4825972018-03-20 20:06:35 +0000158
Wouter van Oortmerssen49482f82018-11-19 17:10:36 +0000159 // Much like WebAssemblyAsmPrinter in the backend, we have to own these.
160 std::vector<std::unique_ptr<wasm::WasmSignature>> Signatures;
161
Wouter van Oortmerssenc7b89f02018-12-03 20:30:28 +0000162 // Order of labels, directives and instructions in a .s file have no
163 // syntactical enforcement. This class is a callback from the actual parser,
164 // and yet we have to be feeding data to the streamer in a very particular
165 // order to ensure a correct binary encoding that matches the regular backend
166 // (the streamer does not enforce this). This "state machine" enum helps
167 // guarantee that correct order.
168 enum ParserState {
169 FileStart,
170 Label,
171 FunctionStart,
172 FunctionLocals,
173 Instructions,
Wouter van Oortmerssen0b3cf242019-02-04 18:03:11 +0000174 EndFunction,
Wouter van Oortmerssenc7b89f02018-12-03 20:30:28 +0000175 } CurrentState = FileStart;
176
Wouter van Oortmerssen29c6ce52018-12-26 22:46:18 +0000177 // For ensuring blocks are properly nested.
178 enum NestingType {
179 Function,
180 Block,
181 Loop,
182 Try,
183 If,
184 Else,
185 Undefined,
186 };
187 std::vector<NestingType> NestingStack;
188
Wouter van Oortmerssenc7b89f02018-12-03 20:30:28 +0000189 // We track this to see if a .functype following a label is the same,
190 // as this is how we recognize the start of a function.
191 MCSymbol *LastLabel = nullptr;
Wouter van Oortmerssen0b3cf242019-02-04 18:03:11 +0000192 MCSymbol *LastFunctionLabel = nullptr;
Wouter van Oortmerssenc7b89f02018-12-03 20:30:28 +0000193
Derek Schuffe4825972018-03-20 20:06:35 +0000194public:
Wouter van Oortmerssende28b5d2018-11-02 22:04:33 +0000195 WebAssemblyAsmParser(const MCSubtargetInfo &STI, MCAsmParser &Parser,
196 const MCInstrInfo &MII, const MCTargetOptions &Options)
197 : MCTargetAsmParser(Options, STI, MII), Parser(Parser),
Wouter van Oortmerssencc75e772018-11-12 20:15:01 +0000198 Lexer(Parser.getLexer()) {
Wouter van Oortmerssende28b5d2018-11-02 22:04:33 +0000199 setAvailableFeatures(ComputeAvailableFeatures(STI.getFeatureBits()));
Derek Schuffe4825972018-03-20 20:06:35 +0000200 }
201
202#define GET_ASSEMBLER_HEADER
203#include "WebAssemblyGenAsmMatcher.inc"
204
205 // TODO: This is required to be implemented, but appears unused.
Heejin Ahnf208f632018-09-05 01:27:38 +0000206 bool ParseRegister(unsigned & /*RegNo*/, SMLoc & /*StartLoc*/,
207 SMLoc & /*EndLoc*/) override {
Derek Schuffe4825972018-03-20 20:06:35 +0000208 llvm_unreachable("ParseRegister is not implemented.");
209 }
210
Wouter van Oortmerssen29c6ce52018-12-26 22:46:18 +0000211 bool error(const Twine &Msg, const AsmToken &Tok) {
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000212 return Parser.Error(Tok.getLoc(), Msg + Tok.getString());
Derek Schuffe4825972018-03-20 20:06:35 +0000213 }
214
Wouter van Oortmerssen29c6ce52018-12-26 22:46:18 +0000215 bool error(const Twine &Msg) {
216 return Parser.Error(Lexer.getTok().getLoc(), Msg);
217 }
218
219 void addSignature(std::unique_ptr<wasm::WasmSignature> &&Sig) {
220 Signatures.push_back(std::move(Sig));
221 }
222
223 std::pair<StringRef, StringRef> nestingString(NestingType NT) {
224 switch (NT) {
225 case Function:
226 return {"function", "end_function"};
227 case Block:
228 return {"block", "end_block"};
229 case Loop:
230 return {"loop", "end_loop"};
231 case Try:
232 return {"try", "end_try"};
233 case If:
234 return {"if", "end_if"};
235 case Else:
236 return {"else", "end_if"};
237 default:
238 llvm_unreachable("unknown NestingType");
239 }
240 }
241
242 void push(NestingType NT) { NestingStack.push_back(NT); }
243
244 bool pop(StringRef Ins, NestingType NT1, NestingType NT2 = Undefined) {
245 if (NestingStack.empty())
246 return error(Twine("End of block construct with no start: ") + Ins);
247 auto Top = NestingStack.back();
248 if (Top != NT1 && Top != NT2)
249 return error(Twine("Block construct type mismatch, expected: ") +
250 nestingString(Top).second + ", instead got: " + Ins);
251 NestingStack.pop_back();
252 return false;
253 }
254
255 bool ensureEmptyNestingStack() {
Heejin Ahn18c56a02019-02-04 19:13:39 +0000256 auto Err = !NestingStack.empty();
Wouter van Oortmerssen29c6ce52018-12-26 22:46:18 +0000257 while (!NestingStack.empty()) {
258 error(Twine("Unmatched block construct(s) at function end: ") +
259 nestingString(NestingStack.back()).first);
260 NestingStack.pop_back();
261 }
Heejin Ahn18c56a02019-02-04 19:13:39 +0000262 return Err;
Wouter van Oortmerssen29c6ce52018-12-26 22:46:18 +0000263 }
264
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000265 bool isNext(AsmToken::TokenKind Kind) {
266 auto Ok = Lexer.is(Kind);
267 if (Ok)
Heejin Ahnf208f632018-09-05 01:27:38 +0000268 Parser.Lex();
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000269 return Ok;
Derek Schuffe4825972018-03-20 20:06:35 +0000270 }
271
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000272 bool expect(AsmToken::TokenKind Kind, const char *KindName) {
273 if (!isNext(Kind))
274 return error(std::string("Expected ") + KindName + ", instead got: ",
Derek Schuffe4825972018-03-20 20:06:35 +0000275 Lexer.getTok());
276 return false;
277 }
278
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000279 StringRef expectIdent() {
Wouter van Oortmerssen49482f82018-11-19 17:10:36 +0000280 if (!Lexer.is(AsmToken::Identifier)) {
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000281 error("Expected identifier, got: ", Lexer.getTok());
Wouter van Oortmerssen49482f82018-11-19 17:10:36 +0000282 return StringRef();
283 }
284 auto Name = Lexer.getTok().getString();
285 Parser.Lex();
286 return Name;
Derek Schuffe4825972018-03-20 20:06:35 +0000287 }
288
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000289 Optional<wasm::ValType> parseType(const StringRef &Type) {
Wouter van Oortmerssen49482f82018-11-19 17:10:36 +0000290 // FIXME: can't use StringSwitch because wasm::ValType doesn't have a
291 // "invalid" value.
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000292 if (Type == "i32")
293 return wasm::ValType::I32;
294 if (Type == "i64")
295 return wasm::ValType::I64;
296 if (Type == "f32")
297 return wasm::ValType::F32;
298 if (Type == "f64")
299 return wasm::ValType::F64;
Wouter van Oortmerssen49482f82018-11-19 17:10:36 +0000300 if (Type == "v128" || Type == "i8x16" || Type == "i16x8" ||
301 Type == "i32x4" || Type == "i64x2" || Type == "f32x4" ||
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000302 Type == "f64x2")
303 return wasm::ValType::V128;
Wouter van Oortmerssenf3b762a2019-01-17 18:14:09 +0000304 if (Type == "except_ref")
305 return wasm::ValType::EXCEPT_REF;
Wouter van Oortmerssen49482f82018-11-19 17:10:36 +0000306 return Optional<wasm::ValType>();
307 }
308
Wouter van Oortmerssenad72f682019-01-02 23:23:51 +0000309 WebAssembly::ExprType parseBlockType(StringRef ID) {
310 return StringSwitch<WebAssembly::ExprType>(ID)
311 .Case("i32", WebAssembly::ExprType::I32)
312 .Case("i64", WebAssembly::ExprType::I64)
313 .Case("f32", WebAssembly::ExprType::F32)
314 .Case("f64", WebAssembly::ExprType::F64)
315 .Case("v128", WebAssembly::ExprType::V128)
316 .Case("except_ref", WebAssembly::ExprType::ExceptRef)
317 .Case("void", WebAssembly::ExprType::Void)
318 .Default(WebAssembly::ExprType::Invalid);
319 }
320
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000321 bool parseRegTypeList(SmallVectorImpl<wasm::ValType> &Types) {
Wouter van Oortmerssencc75e772018-11-12 20:15:01 +0000322 while (Lexer.is(AsmToken::Identifier)) {
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000323 auto Type = parseType(Lexer.getTok().getString());
Wouter van Oortmerssen49482f82018-11-19 17:10:36 +0000324 if (!Type)
Wouter van Oortmerssenf3b762a2019-01-17 18:14:09 +0000325 return error("unknown type: ", Lexer.getTok());
Wouter van Oortmerssen49482f82018-11-19 17:10:36 +0000326 Types.push_back(Type.getValue());
Wouter van Oortmerssencc75e772018-11-12 20:15:01 +0000327 Parser.Lex();
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000328 if (!isNext(AsmToken::Comma))
Wouter van Oortmerssencc75e772018-11-12 20:15:01 +0000329 break;
330 }
Wouter van Oortmerssen49482f82018-11-19 17:10:36 +0000331 return false;
Wouter van Oortmerssencc75e772018-11-12 20:15:01 +0000332 }
333
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000334 void parseSingleInteger(bool IsNegative, OperandVector &Operands) {
Derek Schuffe4825972018-03-20 20:06:35 +0000335 auto &Int = Lexer.getTok();
336 int64_t Val = Int.getIntVal();
Heejin Ahnf208f632018-09-05 01:27:38 +0000337 if (IsNegative)
338 Val = -Val;
Derek Schuffe4825972018-03-20 20:06:35 +0000339 Operands.push_back(make_unique<WebAssemblyOperand>(
Heejin Ahnf208f632018-09-05 01:27:38 +0000340 WebAssemblyOperand::Integer, Int.getLoc(), Int.getEndLoc(),
341 WebAssemblyOperand::IntOp{Val}));
Derek Schuffe4825972018-03-20 20:06:35 +0000342 Parser.Lex();
343 }
344
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000345 bool parseOperandStartingWithInteger(bool IsNegative, OperandVector &Operands,
Wouter van Oortmerssen8a9cb242018-08-27 15:45:51 +0000346 StringRef InstName) {
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000347 parseSingleInteger(IsNegative, Operands);
Wouter van Oortmerssen8a9cb242018-08-27 15:45:51 +0000348 // FIXME: there is probably a cleaner way to do this.
349 auto IsLoadStore = InstName.startswith("load") ||
350 InstName.startswith("store") ||
Heejin Ahn3477bd12019-02-20 01:14:36 +0000351 InstName.startswith("atomic");
Wouter van Oortmerssen8a9cb242018-08-27 15:45:51 +0000352 if (IsLoadStore) {
353 // Parse load/store operands of the form: offset align
354 auto &Offset = Lexer.getTok();
355 if (Offset.is(AsmToken::Integer)) {
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000356 parseSingleInteger(false, Operands);
Derek Schuffe4825972018-03-20 20:06:35 +0000357 } else {
358 // Alignment not specified.
359 // FIXME: correctly derive a default from the instruction.
Wouter van Oortmerssen8a9cb242018-08-27 15:45:51 +0000360 // We can't just call WebAssembly::GetDefaultP2Align since we don't have
361 // an opcode until after the assembly matcher.
Derek Schuffe4825972018-03-20 20:06:35 +0000362 Operands.push_back(make_unique<WebAssemblyOperand>(
Heejin Ahnf208f632018-09-05 01:27:38 +0000363 WebAssemblyOperand::Integer, Offset.getLoc(), Offset.getEndLoc(),
364 WebAssemblyOperand::IntOp{0}));
Derek Schuffe4825972018-03-20 20:06:35 +0000365 }
366 }
367 return false;
368 }
369
Wouter van Oortmerssenad72f682019-01-02 23:23:51 +0000370 void addBlockTypeOperand(OperandVector &Operands, SMLoc NameLoc,
371 WebAssembly::ExprType BT) {
372 Operands.push_back(make_unique<WebAssemblyOperand>(
373 WebAssemblyOperand::Integer, NameLoc, NameLoc,
374 WebAssemblyOperand::IntOp{static_cast<int64_t>(BT)}));
375 }
376
Heejin Ahnf208f632018-09-05 01:27:38 +0000377 bool ParseInstruction(ParseInstructionInfo & /*Info*/, StringRef Name,
Derek Schuffe4825972018-03-20 20:06:35 +0000378 SMLoc NameLoc, OperandVector &Operands) override {
Wouter van Oortmerssen0c83c3f2018-10-01 17:20:31 +0000379 // Note: Name does NOT point into the sourcecode, but to a local, so
380 // use NameLoc instead.
381 Name = StringRef(NameLoc.getPointer(), Name.size());
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000382
Wouter van Oortmerssen0c83c3f2018-10-01 17:20:31 +0000383 // WebAssembly has instructions with / in them, which AsmLexer parses
384 // as seperate tokens, so if we find such tokens immediately adjacent (no
385 // whitespace), expand the name to include them:
386 for (;;) {
387 auto &Sep = Lexer.getTok();
388 if (Sep.getLoc().getPointer() != Name.end() ||
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000389 Sep.getKind() != AsmToken::Slash)
390 break;
Wouter van Oortmerssen0c83c3f2018-10-01 17:20:31 +0000391 // Extend name with /
392 Name = StringRef(Name.begin(), Name.size() + Sep.getString().size());
393 Parser.Lex();
394 // We must now find another identifier, or error.
395 auto &Id = Lexer.getTok();
396 if (Id.getKind() != AsmToken::Identifier ||
397 Id.getLoc().getPointer() != Name.end())
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000398 return error("Incomplete instruction name: ", Id);
Wouter van Oortmerssen0c83c3f2018-10-01 17:20:31 +0000399 Name = StringRef(Name.begin(), Name.size() + Id.getString().size());
400 Parser.Lex();
401 }
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000402
Wouter van Oortmerssen0c83c3f2018-10-01 17:20:31 +0000403 // Now construct the name as first operand.
Heejin Ahnf208f632018-09-05 01:27:38 +0000404 Operands.push_back(make_unique<WebAssemblyOperand>(
Wouter van Oortmerssen0c83c3f2018-10-01 17:20:31 +0000405 WebAssemblyOperand::Token, NameLoc, SMLoc::getFromPointer(Name.end()),
406 WebAssemblyOperand::TokOp{Name}));
Derek Schuffe4825972018-03-20 20:06:35 +0000407 auto NamePair = Name.split('.');
408 // If no '.', there is no type prefix.
Wouter van Oortmerssen0c83c3f2018-10-01 17:20:31 +0000409 auto BaseName = NamePair.second.empty() ? NamePair.first : NamePair.second;
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000410
Wouter van Oortmerssen29c6ce52018-12-26 22:46:18 +0000411 // If this instruction is part of a control flow structure, ensure
412 // proper nesting.
Wouter van Oortmerssenad72f682019-01-02 23:23:51 +0000413 bool ExpectBlockType = false;
Wouter van Oortmerssen29c6ce52018-12-26 22:46:18 +0000414 if (BaseName == "block") {
415 push(Block);
Wouter van Oortmerssenad72f682019-01-02 23:23:51 +0000416 ExpectBlockType = true;
Wouter van Oortmerssen29c6ce52018-12-26 22:46:18 +0000417 } else if (BaseName == "loop") {
418 push(Loop);
Wouter van Oortmerssenad72f682019-01-02 23:23:51 +0000419 ExpectBlockType = true;
Wouter van Oortmerssen29c6ce52018-12-26 22:46:18 +0000420 } else if (BaseName == "try") {
421 push(Try);
Wouter van Oortmerssenad72f682019-01-02 23:23:51 +0000422 ExpectBlockType = true;
Wouter van Oortmerssen29c6ce52018-12-26 22:46:18 +0000423 } else if (BaseName == "if") {
424 push(If);
Wouter van Oortmerssenad72f682019-01-02 23:23:51 +0000425 ExpectBlockType = true;
Wouter van Oortmerssen29c6ce52018-12-26 22:46:18 +0000426 } else if (BaseName == "else") {
427 if (pop(BaseName, If))
428 return true;
429 push(Else);
430 } else if (BaseName == "catch") {
431 if (pop(BaseName, Try))
432 return true;
433 push(Try);
434 } else if (BaseName == "catch_all") {
435 if (pop(BaseName, Try))
436 return true;
437 push(Try);
438 } else if (BaseName == "end_if") {
439 if (pop(BaseName, If, Else))
440 return true;
441 } else if (BaseName == "end_try") {
442 if (pop(BaseName, Try))
443 return true;
444 } else if (BaseName == "end_loop") {
445 if (pop(BaseName, Loop))
446 return true;
447 } else if (BaseName == "end_block") {
448 if (pop(BaseName, Block))
449 return true;
450 } else if (BaseName == "end_function") {
Wouter van Oortmerssen0b3cf242019-02-04 18:03:11 +0000451 CurrentState = EndFunction;
Wouter van Oortmerssen29c6ce52018-12-26 22:46:18 +0000452 if (pop(BaseName, Function) || ensureEmptyNestingStack())
453 return true;
454 }
455
Derek Schuffe4825972018-03-20 20:06:35 +0000456 while (Lexer.isNot(AsmToken::EndOfStatement)) {
457 auto &Tok = Lexer.getTok();
458 switch (Tok.getKind()) {
Derek Schuffe4825972018-03-20 20:06:35 +0000459 case AsmToken::Identifier: {
460 auto &Id = Lexer.getTok();
Wouter van Oortmerssenad72f682019-01-02 23:23:51 +0000461 if (ExpectBlockType) {
462 // Assume this identifier is a block_type.
463 auto BT = parseBlockType(Id.getString());
464 if (BT == WebAssembly::ExprType::Invalid)
465 return error("Unknown block type: ", Id);
466 addBlockTypeOperand(Operands, NameLoc, BT);
467 Parser.Lex();
468 } else {
469 // Assume this identifier is a label.
470 const MCExpr *Val;
471 SMLoc End;
472 if (Parser.parsePrimaryExpr(Val, End))
473 return error("Cannot parse symbol: ", Lexer.getTok());
474 Operands.push_back(make_unique<WebAssemblyOperand>(
475 WebAssemblyOperand::Symbol, Id.getLoc(), Id.getEndLoc(),
476 WebAssemblyOperand::SymOp{Val}));
477 }
Derek Schuffe4825972018-03-20 20:06:35 +0000478 break;
479 }
480 case AsmToken::Minus:
481 Parser.Lex();
482 if (Lexer.isNot(AsmToken::Integer))
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000483 return error("Expected integer instead got: ", Lexer.getTok());
484 if (parseOperandStartingWithInteger(true, Operands, BaseName))
Derek Schuffe4825972018-03-20 20:06:35 +0000485 return true;
486 break;
487 case AsmToken::Integer:
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000488 if (parseOperandStartingWithInteger(false, Operands, BaseName))
Derek Schuffe4825972018-03-20 20:06:35 +0000489 return true;
490 break;
491 case AsmToken::Real: {
492 double Val;
493 if (Tok.getString().getAsDouble(Val, false))
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000494 return error("Cannot parse real: ", Tok);
Derek Schuffe4825972018-03-20 20:06:35 +0000495 Operands.push_back(make_unique<WebAssemblyOperand>(
Heejin Ahnf208f632018-09-05 01:27:38 +0000496 WebAssemblyOperand::Float, Tok.getLoc(), Tok.getEndLoc(),
497 WebAssemblyOperand::FltOp{Val}));
Derek Schuffe4825972018-03-20 20:06:35 +0000498 Parser.Lex();
499 break;
500 }
Wouter van Oortmerssend3c544a2018-12-17 22:04:44 +0000501 case AsmToken::LCurly: {
502 Parser.Lex();
503 auto Op = make_unique<WebAssemblyOperand>(
504 WebAssemblyOperand::BrList, Tok.getLoc(), Tok.getEndLoc());
505 if (!Lexer.is(AsmToken::RCurly))
506 for (;;) {
507 Op->BrL.List.push_back(Lexer.getTok().getIntVal());
508 expect(AsmToken::Integer, "integer");
509 if (!isNext(AsmToken::Comma))
510 break;
511 }
512 expect(AsmToken::RCurly, "}");
513 Operands.push_back(std::move(Op));
514 break;
515 }
Derek Schuffe4825972018-03-20 20:06:35 +0000516 default:
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000517 return error("Unexpected token in operand: ", Tok);
Derek Schuffe4825972018-03-20 20:06:35 +0000518 }
519 if (Lexer.isNot(AsmToken::EndOfStatement)) {
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000520 if (expect(AsmToken::Comma, ","))
Heejin Ahnf208f632018-09-05 01:27:38 +0000521 return true;
Derek Schuffe4825972018-03-20 20:06:35 +0000522 }
523 }
Wouter van Oortmerssenad72f682019-01-02 23:23:51 +0000524 if (ExpectBlockType && Operands.size() == 1) {
525 // Support blocks with no operands as default to void.
526 addBlockTypeOperand(Operands, NameLoc, WebAssembly::ExprType::Void);
Derek Schuffe4825972018-03-20 20:06:35 +0000527 }
Wouter van Oortmerssenad72f682019-01-02 23:23:51 +0000528 Parser.Lex();
Derek Schuffe4825972018-03-20 20:06:35 +0000529 return false;
530 }
531
Wouter van Oortmerssenc7b89f02018-12-03 20:30:28 +0000532 void onLabelParsed(MCSymbol *Symbol) override {
533 LastLabel = Symbol;
534 CurrentState = Label;
535 }
536
Heejin Ahnbe5e5872018-12-11 01:11:04 +0000537 bool parseSignature(wasm::WasmSignature *Signature) {
538 if (expect(AsmToken::LParen, "("))
539 return true;
540 if (parseRegTypeList(Signature->Params))
541 return true;
542 if (expect(AsmToken::RParen, ")"))
543 return true;
544 if (expect(AsmToken::MinusGreater, "->"))
545 return true;
546 if (expect(AsmToken::LParen, "("))
547 return true;
548 if (parseRegTypeList(Signature->Returns))
549 return true;
550 if (expect(AsmToken::RParen, ")"))
551 return true;
552 return false;
553 }
554
Wouter van Oortmerssencc75e772018-11-12 20:15:01 +0000555 // This function processes wasm-specific directives streamed to
556 // WebAssemblyTargetStreamer, all others go to the generic parser
557 // (see WasmAsmParser).
Derek Schuffe4825972018-03-20 20:06:35 +0000558 bool ParseDirective(AsmToken DirectiveID) override {
Wouter van Oortmerssende28b5d2018-11-02 22:04:33 +0000559 // This function has a really weird return value behavior that is different
560 // from all the other parsing functions:
561 // - return true && no tokens consumed -> don't know this directive / let
562 // the generic parser handle it.
563 // - return true && tokens consumed -> a parsing error occurred.
564 // - return false -> processed this directive successfully.
Derek Schuffe4825972018-03-20 20:06:35 +0000565 assert(DirectiveID.getKind() == AsmToken::Identifier);
566 auto &Out = getStreamer();
Heejin Ahnf208f632018-09-05 01:27:38 +0000567 auto &TOut =
568 reinterpret_cast<WebAssemblyTargetStreamer &>(*Out.getTargetStreamer());
Wouter van Oortmerssenf3b762a2019-01-17 18:14:09 +0000569 auto &Ctx = Out.getContext();
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000570
Wouter van Oortmerssende28b5d2018-11-02 22:04:33 +0000571 // TODO: any time we return an error, at least one token must have been
572 // consumed, otherwise this will not signal an error to the caller.
Wouter van Oortmerssencc75e772018-11-12 20:15:01 +0000573 if (DirectiveID.getString() == ".globaltype") {
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000574 auto SymName = expectIdent();
575 if (SymName.empty())
576 return true;
577 if (expect(AsmToken::Comma, ","))
578 return true;
Wouter van Oortmerssen49482f82018-11-19 17:10:36 +0000579 auto TypeTok = Lexer.getTok();
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000580 auto TypeName = expectIdent();
581 if (TypeName.empty())
582 return true;
583 auto Type = parseType(TypeName);
Wouter van Oortmerssen49482f82018-11-19 17:10:36 +0000584 if (!Type)
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000585 return error("Unknown type in .globaltype directive: ", TypeTok);
Wouter van Oortmerssende28b5d2018-11-02 22:04:33 +0000586 // Now set this symbol with the correct type.
Wouter van Oortmerssenf3b762a2019-01-17 18:14:09 +0000587 auto WasmSym = cast<MCSymbolWasm>(Ctx.getOrCreateSymbol(SymName));
Wouter van Oortmerssende28b5d2018-11-02 22:04:33 +0000588 WasmSym->setType(wasm::WASM_SYMBOL_TYPE_GLOBAL);
Wouter van Oortmerssen49482f82018-11-19 17:10:36 +0000589 WasmSym->setGlobalType(
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000590 wasm::WasmGlobalType{uint8_t(Type.getValue()), true});
Wouter van Oortmerssende28b5d2018-11-02 22:04:33 +0000591 // And emit the directive again.
592 TOut.emitGlobalType(WasmSym);
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000593 return expect(AsmToken::EndOfStatement, "EOL");
594 }
595
596 if (DirectiveID.getString() == ".functype") {
Wouter van Oortmerssenc7b89f02018-12-03 20:30:28 +0000597 // This code has to send things to the streamer similar to
598 // WebAssemblyAsmPrinter::EmitFunctionBodyStart.
599 // TODO: would be good to factor this into a common function, but the
600 // assembler and backend really don't share any common code, and this code
601 // parses the locals seperately.
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000602 auto SymName = expectIdent();
603 if (SymName.empty())
604 return true;
Wouter van Oortmerssenf3b762a2019-01-17 18:14:09 +0000605 auto WasmSym = cast<MCSymbolWasm>(Ctx.getOrCreateSymbol(SymName));
Wouter van Oortmerssenc7b89f02018-12-03 20:30:28 +0000606 if (CurrentState == Label && WasmSym == LastLabel) {
607 // This .functype indicates a start of a function.
Wouter van Oortmerssen29c6ce52018-12-26 22:46:18 +0000608 if (ensureEmptyNestingStack())
609 return true;
Wouter van Oortmerssenc7b89f02018-12-03 20:30:28 +0000610 CurrentState = FunctionStart;
Wouter van Oortmerssen0b3cf242019-02-04 18:03:11 +0000611 LastFunctionLabel = LastLabel;
Wouter van Oortmerssen29c6ce52018-12-26 22:46:18 +0000612 push(Function);
Wouter van Oortmerssenc7b89f02018-12-03 20:30:28 +0000613 }
Wouter van Oortmerssen49482f82018-11-19 17:10:36 +0000614 auto Signature = make_unique<wasm::WasmSignature>();
Heejin Ahnbe5e5872018-12-11 01:11:04 +0000615 if (parseSignature(Signature.get()))
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000616 return true;
Wouter van Oortmerssen49482f82018-11-19 17:10:36 +0000617 WasmSym->setSignature(Signature.get());
618 addSignature(std::move(Signature));
619 WasmSym->setType(wasm::WASM_SYMBOL_TYPE_FUNCTION);
620 TOut.emitFunctionType(WasmSym);
Wouter van Oortmerssenc7b89f02018-12-03 20:30:28 +0000621 // TODO: backend also calls TOut.emitIndIdx, but that is not implemented.
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000622 return expect(AsmToken::EndOfStatement, "EOL");
623 }
624
Heejin Ahnbe5e5872018-12-11 01:11:04 +0000625 if (DirectiveID.getString() == ".eventtype") {
626 auto SymName = expectIdent();
627 if (SymName.empty())
628 return true;
Wouter van Oortmerssenf3b762a2019-01-17 18:14:09 +0000629 auto WasmSym = cast<MCSymbolWasm>(Ctx.getOrCreateSymbol(SymName));
Heejin Ahnbe5e5872018-12-11 01:11:04 +0000630 auto Signature = make_unique<wasm::WasmSignature>();
631 if (parseRegTypeList(Signature->Params))
632 return true;
633 WasmSym->setSignature(Signature.get());
634 addSignature(std::move(Signature));
635 WasmSym->setType(wasm::WASM_SYMBOL_TYPE_EVENT);
636 TOut.emitEventType(WasmSym);
637 // TODO: backend also calls TOut.emitIndIdx, but that is not implemented.
638 return expect(AsmToken::EndOfStatement, "EOL");
639 }
640
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000641 if (DirectiveID.getString() == ".local") {
Wouter van Oortmerssenc7b89f02018-12-03 20:30:28 +0000642 if (CurrentState != FunctionStart)
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000643 return error(".local directive should follow the start of a function",
Wouter van Oortmerssenc7b89f02018-12-03 20:30:28 +0000644 Lexer.getTok());
Wouter van Oortmerssen49482f82018-11-19 17:10:36 +0000645 SmallVector<wasm::ValType, 4> Locals;
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000646 if (parseRegTypeList(Locals))
647 return true;
Derek Schuffe4825972018-03-20 20:06:35 +0000648 TOut.emitLocal(Locals);
Wouter van Oortmerssenc7b89f02018-12-03 20:30:28 +0000649 CurrentState = FunctionLocals;
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000650 return expect(AsmToken::EndOfStatement, "EOL");
Derek Schuffe4825972018-03-20 20:06:35 +0000651 }
Heejin Ahn7ce5edf2018-12-07 21:35:37 +0000652
653 return true; // We didn't process this directive.
Derek Schuffe4825972018-03-20 20:06:35 +0000654 }
655
Heejin Ahnf208f632018-09-05 01:27:38 +0000656 bool MatchAndEmitInstruction(SMLoc IDLoc, unsigned & /*Opcode*/,
657 OperandVector &Operands, MCStreamer &Out,
658 uint64_t &ErrorInfo,
Derek Schuffe4825972018-03-20 20:06:35 +0000659 bool MatchingInlineAsm) override {
660 MCInst Inst;
661 unsigned MatchResult =
662 MatchInstructionImpl(Operands, Inst, ErrorInfo, MatchingInlineAsm);
663 switch (MatchResult) {
664 case Match_Success: {
Wouter van Oortmerssenc7b89f02018-12-03 20:30:28 +0000665 if (CurrentState == FunctionStart) {
666 // This is the first instruction in a function, but we haven't seen
667 // a .local directive yet. The streamer requires locals to be encoded
668 // as a prelude to the instructions, so emit an empty list of locals
669 // here.
670 auto &TOut = reinterpret_cast<WebAssemblyTargetStreamer &>(
671 *Out.getTargetStreamer());
672 TOut.emitLocal(SmallVector<wasm::ValType, 0>());
673 }
Derek Schuffe4825972018-03-20 20:06:35 +0000674 Out.EmitInstruction(Inst, getSTI());
Wouter van Oortmerssen0b3cf242019-02-04 18:03:11 +0000675 if (CurrentState == EndFunction) {
676 onEndOfFunction();
677 } else {
678 CurrentState = Instructions;
679 }
Derek Schuffe4825972018-03-20 20:06:35 +0000680 return false;
681 }
682 case Match_MissingFeature:
Heejin Ahnf208f632018-09-05 01:27:38 +0000683 return Parser.Error(
684 IDLoc, "instruction requires a WASM feature not currently enabled");
Derek Schuffe4825972018-03-20 20:06:35 +0000685 case Match_MnemonicFail:
686 return Parser.Error(IDLoc, "invalid instruction");
687 case Match_NearMisses:
688 return Parser.Error(IDLoc, "ambiguous instruction");
689 case Match_InvalidTiedOperand:
690 case Match_InvalidOperand: {
691 SMLoc ErrorLoc = IDLoc;
692 if (ErrorInfo != ~0ULL) {
693 if (ErrorInfo >= Operands.size())
694 return Parser.Error(IDLoc, "too few operands for instruction");
695 ErrorLoc = Operands[ErrorInfo]->getStartLoc();
696 if (ErrorLoc == SMLoc())
697 ErrorLoc = IDLoc;
698 }
699 return Parser.Error(ErrorLoc, "invalid operand for instruction");
700 }
701 }
702 llvm_unreachable("Implement any new match types added!");
703 }
Wouter van Oortmerssen29c6ce52018-12-26 22:46:18 +0000704
Wouter van Oortmerssen0b3cf242019-02-04 18:03:11 +0000705 void doBeforeLabelEmit(MCSymbol *Symbol) override {
706 // Start a new section for the next function automatically, since our
707 // object writer expects each function to have its own section. This way
708 // The user can't forget this "convention".
709 auto SymName = Symbol->getName();
710 if (SymName.startswith(".L"))
711 return; // Local Symbol.
712 auto SecName = ".text." + SymName;
713 auto WS = getContext().getWasmSection(SecName, SectionKind::getText());
714 getStreamer().SwitchSection(WS);
715 }
716
717 void onEndOfFunction() {
718 // Automatically output a .size directive, so it becomes optional for the
719 // user.
720 auto TempSym = getContext().createLinkerPrivateTempSymbol();
721 getStreamer().EmitLabel(TempSym);
722 auto Start = MCSymbolRefExpr::create(LastLabel, getContext());
723 auto End = MCSymbolRefExpr::create(TempSym, getContext());
724 auto Expr =
725 MCBinaryExpr::create(MCBinaryExpr::Sub, End, Start, getContext());
726 getStreamer().emitELFSize(LastFunctionLabel, Expr);
727 }
728
Wouter van Oortmerssen29c6ce52018-12-26 22:46:18 +0000729 void onEndOfFile() override { ensureEmptyNestingStack(); }
Derek Schuffe4825972018-03-20 20:06:35 +0000730};
731} // end anonymous namespace
732
733// Force static initialization.
734extern "C" void LLVMInitializeWebAssemblyAsmParser() {
735 RegisterMCAsmParser<WebAssemblyAsmParser> X(getTheWebAssemblyTarget32());
736 RegisterMCAsmParser<WebAssemblyAsmParser> Y(getTheWebAssemblyTarget64());
737}
738
739#define GET_REGISTER_MATCHER
740#define GET_MATCHER_IMPLEMENTATION
741#include "WebAssemblyGenAsmMatcher.inc"