WebAssemblyDisassembler.cpp revision 360784 
1//==- WebAssemblyDisassembler.cpp - Disassembler 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 Disassembler.
11///
12/// It contains code to translate the data produced by the decoder into
13/// MCInsts.
14///
15//===----------------------------------------------------------------------===//
16
17#include "MCTargetDesc/WebAssemblyInstPrinter.h"
18#include "MCTargetDesc/WebAssemblyMCTargetDesc.h"
19#include "TargetInfo/WebAssemblyTargetInfo.h"
20#include "llvm/MC/MCContext.h"
21#include "llvm/MC/MCDisassembler/MCDisassembler.h"
22#include "llvm/MC/MCFixedLenDisassembler.h"
23#include "llvm/MC/MCInst.h"
24#include "llvm/MC/MCInstrInfo.h"
25#include "llvm/MC/MCSubtargetInfo.h"
26#include "llvm/MC/MCSymbol.h"
27#include "llvm/MC/MCSymbolWasm.h"
28#include "llvm/Support/Endian.h"
29#include "llvm/Support/LEB128.h"
30#include "llvm/Support/TargetRegistry.h"
31
32using namespace llvm;
33
34#define DEBUG_TYPE "wasm-disassembler"
35
36using DecodeStatus = MCDisassembler::DecodeStatus;
37
38#include "WebAssemblyGenDisassemblerTables.inc"
39
40namespace {
41static constexpr int WebAssemblyInstructionTableSize = 256;
42
43class WebAssemblyDisassembler final : public MCDisassembler {
44  std::unique_ptr<const MCInstrInfo> MCII;
45
46  DecodeStatus getInstruction(MCInst &Instr, uint64_t &Size,
47                              ArrayRef<uint8_t> Bytes, uint64_t Address,
48                              raw_ostream &CStream) const override;
49  DecodeStatus onSymbolStart(StringRef Name, uint64_t &Size,
50                             ArrayRef<uint8_t> Bytes, uint64_t Address,
51                             raw_ostream &CStream) const override;
52
53public:
54  WebAssemblyDisassembler(const MCSubtargetInfo &STI, MCContext &Ctx,
55                          std::unique_ptr<const MCInstrInfo> MCII)
56      : MCDisassembler(STI, Ctx), MCII(std::move(MCII)) {}
57};
58} // end anonymous namespace
59
60static MCDisassembler *createWebAssemblyDisassembler(const Target &T,
61                                                     const MCSubtargetInfo &STI,
62                                                     MCContext &Ctx) {
63  std::unique_ptr<const MCInstrInfo> MCII(T.createMCInstrInfo());
64  return new WebAssemblyDisassembler(STI, Ctx, std::move(MCII));
65}
66
67extern "C" LLVM_EXTERNAL_VISIBILITY void
68LLVMInitializeWebAssemblyDisassembler() {
69  // Register the disassembler for each target.
70  TargetRegistry::RegisterMCDisassembler(getTheWebAssemblyTarget32(),
71                                         createWebAssemblyDisassembler);
72  TargetRegistry::RegisterMCDisassembler(getTheWebAssemblyTarget64(),
73                                         createWebAssemblyDisassembler);
74}
75
76static int nextByte(ArrayRef<uint8_t> Bytes, uint64_t &Size) {
77  if (Size >= Bytes.size())
78    return -1;
79  auto V = Bytes[Size];
80  Size++;
81  return V;
82}
83
84static bool nextLEB(int64_t &Val, ArrayRef<uint8_t> Bytes, uint64_t &Size,
85                    bool Signed) {
86  unsigned N = 0;
87  const char *Error = nullptr;
88  Val = Signed ? decodeSLEB128(Bytes.data() + Size, &N,
89                               Bytes.data() + Bytes.size(), &Error)
90               : static_cast<int64_t>(decodeULEB128(Bytes.data() + Size, &N,
91                                                    Bytes.data() + Bytes.size(),
92                                                    &Error));
93  if (Error)
94    return false;
95  Size += N;
96  return true;
97}
98
99static bool parseLEBImmediate(MCInst &MI, uint64_t &Size,
100                              ArrayRef<uint8_t> Bytes, bool Signed) {
101  int64_t Val;
102  if (!nextLEB(Val, Bytes, Size, Signed))
103    return false;
104  MI.addOperand(MCOperand::createImm(Val));
105  return true;
106}
107
108template <typename T>
109bool parseImmediate(MCInst &MI, uint64_t &Size, ArrayRef<uint8_t> Bytes) {
110  if (Size + sizeof(T) > Bytes.size())
111    return false;
112  T Val = support::endian::read<T, support::endianness::little, 1>(
113      Bytes.data() + Size);
114  Size += sizeof(T);
115  if (std::is_floating_point<T>::value) {
116    MI.addOperand(MCOperand::createFPImm(static_cast<double>(Val)));
117  } else {
118    MI.addOperand(MCOperand::createImm(static_cast<int64_t>(Val)));
119  }
120  return true;
121}
122
123MCDisassembler::DecodeStatus WebAssemblyDisassembler::onSymbolStart(
124    StringRef Name, uint64_t &Size, ArrayRef<uint8_t> Bytes, uint64_t Address,
125    raw_ostream &CStream) const {
126  Size = 0;
127  if (Address == 0) {
128    // Start of a code section: we're parsing only the function count.
129    int64_t FunctionCount;
130    if (!nextLEB(FunctionCount, Bytes, Size, false))
131      return MCDisassembler::Fail;
132    outs() << "        # " << FunctionCount << " functions in section.";
133  } else {
134    // Parse the start of a single function.
135    int64_t BodySize, LocalEntryCount;
136    if (!nextLEB(BodySize, Bytes, Size, false) ||
137        !nextLEB(LocalEntryCount, Bytes, Size, false))
138      return MCDisassembler::Fail;
139    if (LocalEntryCount) {
140      outs() << "        .local ";
141      for (int64_t I = 0; I < LocalEntryCount; I++) {
142        int64_t Count, Type;
143        if (!nextLEB(Count, Bytes, Size, false) ||
144            !nextLEB(Type, Bytes, Size, false))
145          return MCDisassembler::Fail;
146        for (int64_t J = 0; J < Count; J++) {
147          if (I || J)
148            outs() << ", ";
149          outs() << WebAssembly::anyTypeToString(Type);
150        }
151      }
152    }
153  }
154  outs() << "\n";
155  return MCDisassembler::Success;
156}
157
158MCDisassembler::DecodeStatus WebAssemblyDisassembler::getInstruction(
159    MCInst &MI, uint64_t &Size, ArrayRef<uint8_t> Bytes, uint64_t /*Address*/,
160    raw_ostream &CS) const {
161  CommentStream = &CS;
162  Size = 0;
163  int Opc = nextByte(Bytes, Size);
164  if (Opc < 0)
165    return MCDisassembler::Fail;
166  const auto *WasmInst = &InstructionTable0[Opc];
167  // If this is a prefix byte, indirect to another table.
168  if (WasmInst->ET == ET_Prefix) {
169    WasmInst = nullptr;
170    // Linear search, so far only 2 entries.
171    for (auto PT = PrefixTable; PT->Table; PT++) {
172      if (PT->Prefix == Opc) {
173        WasmInst = PT->Table;
174        break;
175      }
176    }
177    if (!WasmInst)
178      return MCDisassembler::Fail;
179    int64_t PrefixedOpc;
180    if (!nextLEB(PrefixedOpc, Bytes, Size, false))
181      return MCDisassembler::Fail;
182    if (PrefixedOpc < 0 || PrefixedOpc >= WebAssemblyInstructionTableSize)
183      return MCDisassembler::Fail;
184    WasmInst += PrefixedOpc;
185  }
186  if (WasmInst->ET == ET_Unused)
187    return MCDisassembler::Fail;
188  // At this point we must have a valid instruction to decode.
189  assert(WasmInst->ET == ET_Instruction);
190  MI.setOpcode(WasmInst->Opcode);
191  // Parse any operands.
192  for (uint8_t OPI = 0; OPI < WasmInst->NumOperands; OPI++) {
193    auto OT = OperandTable[WasmInst->OperandStart + OPI];
194    switch (OT) {
195    // ULEB operands:
196    case WebAssembly::OPERAND_BASIC_BLOCK:
197    case WebAssembly::OPERAND_LOCAL:
198    case WebAssembly::OPERAND_GLOBAL:
199    case WebAssembly::OPERAND_FUNCTION32:
200    case WebAssembly::OPERAND_OFFSET32:
201    case WebAssembly::OPERAND_P2ALIGN:
202    case WebAssembly::OPERAND_TYPEINDEX:
203    case WebAssembly::OPERAND_EVENT:
204    case MCOI::OPERAND_IMMEDIATE: {
205      if (!parseLEBImmediate(MI, Size, Bytes, false))
206        return MCDisassembler::Fail;
207      break;
208    }
209    // SLEB operands:
210    case WebAssembly::OPERAND_I32IMM:
211    case WebAssembly::OPERAND_I64IMM: {
212      if (!parseLEBImmediate(MI, Size, Bytes, true))
213        return MCDisassembler::Fail;
214      break;
215    }
216    // block_type operands:
217    case WebAssembly::OPERAND_SIGNATURE: {
218      int64_t Val;
219      uint64_t PrevSize = Size;
220      if (!nextLEB(Val, Bytes, Size, true))
221        return MCDisassembler::Fail;
222      if (Val < 0) {
223        // Negative values are single septet value types or empty types
224        if (Size != PrevSize + 1) {
225          MI.addOperand(
226              MCOperand::createImm(int64_t(WebAssembly::BlockType::Invalid)));
227        } else {
228          MI.addOperand(MCOperand::createImm(Val & 0x7f));
229        }
230      } else {
231        // We don't have access to the signature, so create a symbol without one
232        MCSymbol *Sym = getContext().createTempSymbol("typeindex", true);
233        auto *WasmSym = cast<MCSymbolWasm>(Sym);
234        WasmSym->setType(wasm::WASM_SYMBOL_TYPE_FUNCTION);
235        const MCExpr *Expr = MCSymbolRefExpr::create(
236            WasmSym, MCSymbolRefExpr::VK_WASM_TYPEINDEX, getContext());
237        MI.addOperand(MCOperand::createExpr(Expr));
238      }
239      break;
240    }
241    // FP operands.
242    case WebAssembly::OPERAND_F32IMM: {
243      if (!parseImmediate<float>(MI, Size, Bytes))
244        return MCDisassembler::Fail;
245      break;
246    }
247    case WebAssembly::OPERAND_F64IMM: {
248      if (!parseImmediate<double>(MI, Size, Bytes))
249        return MCDisassembler::Fail;
250      break;
251    }
252    // Vector lane operands (not LEB encoded).
253    case WebAssembly::OPERAND_VEC_I8IMM: {
254      if (!parseImmediate<uint8_t>(MI, Size, Bytes))
255        return MCDisassembler::Fail;
256      break;
257    }
258    case WebAssembly::OPERAND_VEC_I16IMM: {
259      if (!parseImmediate<uint16_t>(MI, Size, Bytes))
260        return MCDisassembler::Fail;
261      break;
262    }
263    case WebAssembly::OPERAND_VEC_I32IMM: {
264      if (!parseImmediate<uint32_t>(MI, Size, Bytes))
265        return MCDisassembler::Fail;
266      break;
267    }
268    case WebAssembly::OPERAND_VEC_I64IMM: {
269      if (!parseImmediate<uint64_t>(MI, Size, Bytes))
270        return MCDisassembler::Fail;
271      break;
272    }
273    case WebAssembly::OPERAND_BRLIST: {
274      int64_t TargetTableLen;
275      if (!nextLEB(TargetTableLen, Bytes, Size, false))
276        return MCDisassembler::Fail;
277      for (int64_t I = 0; I < TargetTableLen; I++) {
278        if (!parseLEBImmediate(MI, Size, Bytes, false))
279          return MCDisassembler::Fail;
280      }
281      // Default case.
282      if (!parseLEBImmediate(MI, Size, Bytes, false))
283        return MCDisassembler::Fail;
284      break;
285    }
286    case MCOI::OPERAND_REGISTER:
287      // The tablegen header currently does not have any register operands since
288      // we use only the stack (_S) instructions.
289      // If you hit this that probably means a bad instruction definition in
290      // tablegen.
291      llvm_unreachable("Register operand in WebAssemblyDisassembler");
292    default:
293      llvm_unreachable("Unknown operand type in WebAssemblyDisassembler");
294    }
295  }
296  return MCDisassembler::Success;
297}
298