utils/TableGen/WebAssemblyDisassemblerEmitter.cpp

   1 //===- WebAssemblyDisassemblerEmitter.cpp - Disassembler tables -*- 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 // This file is part of the WebAssembly Disassembler Emitter.
  10 // It contains the implementation of the disassembler tables.
  11 // Documentation for the disassembler emitter in general can be found in
  12 // WebAssemblyDisassemblerEmitter.h.
  13 //
  14 //===----------------------------------------------------------------------===//
  15
  16 #include "WebAssemblyDisassemblerEmitter.h"
  17 #include "llvm/TableGen/Record.h"
  18
  19 namespace llvm {
  20
  21 static constexpr int WebAssemblyInstructionTableSize = 256;
  22
  23 void emitWebAssemblyDisassemblerTables(
  24     raw_ostream &OS,
  25     const ArrayRef<const CodeGenInstruction *> &NumberedInstructions) {
  26   // First lets organize all opcodes by (prefix) byte. Prefix 0 is the
  27   // starting table.
  28   std::map<unsigned,
  29            std::map<unsigned, std::pair<unsigned, const CodeGenInstruction *>>>
  30       OpcodeTable;
  31   for (unsigned I = 0; I != NumberedInstructions.size(); ++I) {
  32     auto &CGI = *NumberedInstructions[I];
  33     auto &Def = *CGI.TheDef;
  34     if (!Def.getValue("Inst"))
  35       continue;
  36     auto &Inst = *Def.getValueAsBitsInit("Inst");
  37     auto Opc = static_cast<unsigned>(
  38         reinterpret_cast<IntInit *>(Inst.convertInitializerTo(IntRecTy::get()))
  39             ->getValue());
  40     if (Opc == 0xFFFFFFFF)
  41       continue; // No opcode defined.
  42     assert(Opc <= 0xFFFF);
  43     auto Prefix = Opc >> 8;
  44     Opc = Opc & 0xFF;
  45     auto &CGIP = OpcodeTable[Prefix][Opc];
  46     // All wasm instructions have a StackBased field of type string, we only
  47     // want the instructions for which this is "true".
  48     auto StackString =
  49         Def.getValue("StackBased")->getValue()->getCastTo(StringRecTy::get());
  50     auto IsStackBased =
  51         StackString &&
  52         reinterpret_cast<const StringInit *>(StackString)->getValue() == "true";
  53     if (!IsStackBased)
  54       continue;
  55     if (CGIP.second) {
  56       // We already have an instruction for this slot, so decide which one
  57       // should be the canonical one. This determines which variant gets
  58       // printed in a disassembly. We want e.g. "call" not "i32.call", and
  59       // "end" when we don't know if its "end_loop" or "end_block" etc.
  60       auto IsCanonicalExisting = CGIP.second->TheDef->getValue("IsCanonical")
  61                                      ->getValue()
  62                                      ->getAsString() == "1";
  63       // We already have one marked explicitly as canonical, so keep it.
  64       if (IsCanonicalExisting)
  65         continue;
  66       auto IsCanonicalNew =
  67           Def.getValue("IsCanonical")->getValue()->getAsString() == "1";
  68       // If the new one is explicitly marked as canonical, take it.
  69       if (!IsCanonicalNew) {
  70         // Neither the existing or new instruction is canonical.
  71         // Pick the one with the shortest name as heuristic.
  72         // Though ideally IsCanonical is always defined for at least one
  73         // variant so this never has to apply.
  74         if (CGIP.second->AsmString.size() <= CGI.AsmString.size())
  75           continue;
  76       }
  77     }
  78     // Set this instruction as the one to use.
  79     CGIP = std::make_pair(I, &CGI);
  80   }
  81   OS << "#include \"MCTargetDesc/WebAssemblyMCTargetDesc.h\"\n";
  82   OS << "\n";
  83   OS << "namespace llvm {\n\n";
  84   OS << "static constexpr int WebAssemblyInstructionTableSize = ";
  85   OS << WebAssemblyInstructionTableSize << ";\n\n";
  86   OS << "enum EntryType : uint8_t { ";
  87   OS << "ET_Unused, ET_Prefix, ET_Instruction };\n\n";
  88   OS << "struct WebAssemblyInstruction {\n";
  89   OS << "  uint16_t Opcode;\n";
  90   OS << "  EntryType ET;\n";
  91   OS << "  uint8_t NumOperands;\n";
  92   OS << "  uint16_t OperandStart;\n";
  93   OS << "};\n\n";
  94   std::vector<std::string> OperandTable, CurOperandList;
  95   // Output one table per prefix.
  96   for (auto &PrefixPair : OpcodeTable) {
  97     if (PrefixPair.second.empty())
  98       continue;
  99     OS << "WebAssemblyInstruction InstructionTable" << PrefixPair.first;
 100     OS << "[] = {\n";
 101     for (unsigned I = 0; I < WebAssemblyInstructionTableSize; I++) {
 102       auto InstIt = PrefixPair.second.find(I);
 103       if (InstIt != PrefixPair.second.end()) {
 104         // Regular instruction.
 105         assert(InstIt->second.second);
 106         auto &CGI = *InstIt->second.second;
 107         OS << "  // 0x";
 108         OS.write_hex(static_cast<unsigned long long>(I));
 109         OS << ": " << CGI.AsmString << "\n";
 110         OS << "  { " << InstIt->second.first << ", ET_Instruction, ";
 111         OS << CGI.Operands.OperandList.size() << ", ";
 112         // Collect operand types for storage in a shared list.
 113         CurOperandList.clear();
 114         for (auto &Op : CGI.Operands.OperandList) {
 115           assert(Op.OperandType != "MCOI::OPERAND_UNKNOWN");
 116           CurOperandList.push_back(Op.OperandType);
 117         }
 118         // See if we already have stored this sequence before. This is not
 119         // strictly necessary but makes the table really small.
 120         size_t OperandStart = OperandTable.size();
 121         if (CurOperandList.size() <= OperandTable.size()) {
 122           for (size_t J = 0; J <= OperandTable.size() - CurOperandList.size();
 123                ++J) {
 124             size_t K = 0;
 125             for (; K < CurOperandList.size(); ++K) {
 126               if (OperandTable[J + K] != CurOperandList[K]) break;
 127             }
 128             if (K == CurOperandList.size()) {
 129               OperandStart = J;
 130               break;
 131             }
 132           }
 133         }
 134         // Store operands if no prior occurrence.
 135         if (OperandStart == OperandTable.size()) {
 136           OperandTable.insert(OperandTable.end(), CurOperandList.begin(),
 137                               CurOperandList.end());
 138         }
 139         OS << OperandStart;
 140       } else {
 141         auto PrefixIt = OpcodeTable.find(I);
 142         // If we have a non-empty table for it that's not 0, this is a prefix.
 143         if (PrefixIt != OpcodeTable.end() && I && !PrefixPair.first) {
 144           OS << "  { 0, ET_Prefix, 0, 0";
 145         } else {
 146           OS << "  { 0, ET_Unused, 0, 0";
 147         }
 148       }
 149       OS << "  },\n";
 150     }
 151     OS << "};\n\n";
 152   }
 153   // Create a table of all operands:
 154   OS << "const uint8_t OperandTable[] = {\n";
 155   for (auto &Op : OperandTable) {
 156     OS << "  " << Op << ",\n";
 157   }
 158   OS << "};\n\n";
 159   // Create a table of all extension tables:
 160   OS << "struct { uint8_t Prefix; const WebAssemblyInstruction *Table; }\n";
 161   OS << "PrefixTable[] = {\n";
 162   for (auto &PrefixPair : OpcodeTable) {
 163     if (PrefixPair.second.empty() || !PrefixPair.first)
 164       continue;
 165     OS << "  { " << PrefixPair.first << ", InstructionTable"
 166        << PrefixPair.first;
 167     OS << " },\n";
 168   }
 169   OS << "  { 0, nullptr }\n};\n\n";
 170   OS << "} // End llvm namespace\n";
 171 }
 172
 173 } // namespace llvm