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 "Common/CodeGenInstruction.h"
18#include "llvm/ADT/STLExtras.h"
19#include "llvm/Support/raw_ostream.h"
20#include "llvm/TableGen/Record.h"
21
22namespace llvm {
23
24static constexpr int WebAssemblyInstructionTableSize = 256;
25
26void emitWebAssemblyDisassemblerTables(
27 raw_ostream &OS,
28 const ArrayRef<const CodeGenInstruction *> &NumberedInstructions) {
29 // First lets organize all opcodes by (prefix) byte. Prefix 0 is the
30 // starting table.
31 std::map<unsigned,
32 std::map<unsigned, std::pair<unsigned, const CodeGenInstruction *>>>
33 OpcodeTable;
34 for (unsigned I = 0; I != NumberedInstructions.size(); ++I) {
35 auto &CGI = *NumberedInstructions[I];
36 auto &Def = *CGI.TheDef;
37 if (!Def.getValue(Name: "Inst"))
38 continue;
39 auto &Inst = *Def.getValueAsBitsInit(FieldName: "Inst");
40 RecordKeeper &RK = Inst.getRecordKeeper();
41 unsigned Opc = static_cast<unsigned>(
42 cast<IntInit>(Val: Inst.convertInitializerTo(Ty: IntRecTy::get(RK)))
43 ->getValue());
44 if (Opc == 0xFFFFFFFF)
45 continue; // No opcode defined.
46 assert(Opc <= 0xFFFFFF);
47 unsigned Prefix;
48 if (Opc <= 0xFFFF) {
49 Prefix = Opc >> 8;
50 Opc = Opc & 0xFF;
51 } else {
52 Prefix = Opc >> 16;
53 Opc = Opc & 0xFFFF;
54 }
55 auto &CGIP = OpcodeTable[Prefix][Opc];
56 // All wasm instructions have a StackBased field of type string, we only
57 // want the instructions for which this is "true".
58 bool IsStackBased = Def.getValueAsBit(FieldName: "StackBased");
59 if (!IsStackBased)
60 continue;
61 if (CGIP.second) {
62 // We already have an instruction for this slot, so decide which one
63 // should be the canonical one. This determines which variant gets
64 // printed in a disassembly. We want e.g. "call" not "i32.call", and
65 // "end" when we don't know if its "end_loop" or "end_block" etc.
66 bool IsCanonicalExisting =
67 CGIP.second->TheDef->getValueAsBit(FieldName: "IsCanonical");
68 // We already have one marked explicitly as canonical, so keep it.
69 if (IsCanonicalExisting)
70 continue;
71 bool IsCanonicalNew = Def.getValueAsBit(FieldName: "IsCanonical");
72 // If the new one is explicitly marked as canonical, take it.
73 if (!IsCanonicalNew) {
74 // Neither the existing or new instruction is canonical.
75 // Pick the one with the shortest name as heuristic.
76 // Though ideally IsCanonical is always defined for at least one
77 // variant so this never has to apply.
78 if (CGIP.second->AsmString.size() <= CGI.AsmString.size())
79 continue;
80 }
81 }
82 // Set this instruction as the one to use.
83 CGIP = std::pair(I, &CGI);
84 }
85 OS << "#include \"MCTargetDesc/WebAssemblyMCTargetDesc.h\"\n";
86 OS << "\n";
87 OS << "namespace llvm {\n\n";
88 OS << "static constexpr int WebAssemblyInstructionTableSize = ";
89 OS << WebAssemblyInstructionTableSize << ";\n\n";
90 OS << "enum EntryType : uint8_t { ";
91 OS << "ET_Unused, ET_Prefix, ET_Instruction };\n\n";
92 OS << "struct WebAssemblyInstruction {\n";
93 OS << " uint16_t Opcode;\n";
94 OS << " EntryType ET;\n";
95 OS << " uint8_t NumOperands;\n";
96 OS << " uint16_t OperandStart;\n";
97 OS << "};\n\n";
98 std::vector<std::string> OperandTable, CurOperandList;
99 // Output one table per prefix.
100 for (auto &PrefixPair : OpcodeTable) {
101 if (PrefixPair.second.empty())
102 continue;
103 OS << "WebAssemblyInstruction InstructionTable" << PrefixPair.first;
104 OS << "[] = {\n";
105 for (unsigned I = 0; I < WebAssemblyInstructionTableSize; I++) {
106 auto InstIt = PrefixPair.second.find(x: I);
107 if (InstIt != PrefixPair.second.end()) {
108 // Regular instruction.
109 assert(InstIt->second.second);
110 auto &CGI = *InstIt->second.second;
111 OS << " // 0x";
112 OS.write_hex(N: static_cast<unsigned long long>(I));
113 OS << ": " << CGI.AsmString << "\n";
114 OS << " { " << InstIt->second.first << ", ET_Instruction, ";
115 OS << CGI.Operands.OperandList.size() << ", ";
116 // Collect operand types for storage in a shared list.
117 CurOperandList.clear();
118 for (auto &Op : CGI.Operands.OperandList) {
119 assert(Op.OperandType != "MCOI::OPERAND_UNKNOWN");
120 CurOperandList.push_back(x: Op.OperandType);
121 }
122 // See if we already have stored this sequence before. This is not
123 // strictly necessary but makes the table really small.
124 size_t OperandStart = OperandTable.size();
125 if (CurOperandList.size() <= OperandTable.size()) {
126 for (size_t J = 0; J <= OperandTable.size() - CurOperandList.size();
127 ++J) {
128 size_t K = 0;
129 for (; K < CurOperandList.size(); ++K) {
130 if (OperandTable[J + K] != CurOperandList[K])
131 break;
132 }
133 if (K == CurOperandList.size()) {
134 OperandStart = J;
135 break;
136 }
137 }
138 }
139 // Store operands if no prior occurrence.
140 if (OperandStart == OperandTable.size()) {
141 llvm::append_range(C&: OperandTable, R&: CurOperandList);
142 }
143 OS << OperandStart;
144 } else {
145 auto PrefixIt = OpcodeTable.find(x: I);
146 // If we have a non-empty table for it that's not 0, this is a prefix.
147 if (PrefixIt != OpcodeTable.end() && I && !PrefixPair.first) {
148 OS << " { 0, ET_Prefix, 0, 0";
149 } else {
150 OS << " { 0, ET_Unused, 0, 0";
151 }
152 }
153 OS << " },\n";
154 }
155 OS << "};\n\n";
156 }
157 // Create a table of all operands:
158 OS << "const uint8_t OperandTable[] = {\n";
159 for (auto &Op : OperandTable) {
160 OS << " " << Op << ",\n";
161 }
162 OS << "};\n\n";
163 // Create a table of all extension tables:
164 OS << "struct { uint8_t Prefix; const WebAssemblyInstruction *Table; }\n";
165 OS << "PrefixTable[] = {\n";
166 for (auto &PrefixPair : OpcodeTable) {
167 if (PrefixPair.second.empty() || !PrefixPair.first)
168 continue;
169 OS << " { " << PrefixPair.first << ", InstructionTable"
170 << PrefixPair.first;
171 OS << " },\n";
172 }
173 OS << " { 0, nullptr }\n};\n\n";
174 OS << "} // end namespace llvm\n";
175}
176
177} // namespace llvm
178