1 //===- Disassembler.cpp - Disassembler for hex strings --------------------===//
3 // The LLVM Compiler Infrastructure
5 // This file is distributed under the University of Illinois Open Source
6 // License. See LICENSE.TXT for details.
8 //===----------------------------------------------------------------------===//
10 // This class implements the disassembler of strings of bytes written in
11 // hexadecimal, from standard input or from a file.
13 //===----------------------------------------------------------------------===//
15 #include "Disassembler.h"
16 #include "llvm/ADT/Triple.h"
17 #include "llvm/MC/MCAsmInfo.h"
18 #include "llvm/MC/MCContext.h"
19 #include "llvm/MC/MCDisassembler.h"
20 #include "llvm/MC/MCInst.h"
21 #include "llvm/MC/MCRegisterInfo.h"
22 #include "llvm/MC/MCStreamer.h"
23 #include "llvm/MC/MCSubtargetInfo.h"
24 #include "llvm/Support/MemoryBuffer.h"
25 #include "llvm/Support/StringRefMemoryObject.h"
26 #include "llvm/Support/SourceMgr.h"
27 #include "llvm/Support/TargetRegistry.h"
28 #include "llvm/Support/raw_ostream.h"
32 typedef std::pair<std::vector<unsigned char>, std::vector<const char *>>
35 static bool PrintInsts(const MCDisassembler &DisAsm,
36 const ByteArrayTy &Bytes,
37 SourceMgr &SM, raw_ostream &Out,
38 MCStreamer &Streamer, bool InAtomicBlock,
39 const MCSubtargetInfo &STI) {
40 // Wrap the vector in a MemoryObject.
41 ArrayRef<uint8_t> Data(Bytes.first.data(), Bytes.first.size());
43 // Disassemble it to strings.
47 for (Index = 0; Index < Bytes.first.size(); Index += Size) {
50 MCDisassembler::DecodeStatus S;
51 S = DisAsm.getInstruction(Inst, Size, Data.slice(Index), Index,
52 /*REMOVE*/ nulls(), nulls());
54 case MCDisassembler::Fail:
55 SM.PrintMessage(SMLoc::getFromPointer(Bytes.second[Index]),
56 SourceMgr::DK_Warning,
57 "invalid instruction encoding");
58 // Don't try to resynchronise the stream in a block
63 Size = 1; // skip illegible bytes
67 case MCDisassembler::SoftFail:
68 SM.PrintMessage(SMLoc::getFromPointer(Bytes.second[Index]),
69 SourceMgr::DK_Warning,
70 "potentially undefined instruction encoding");
73 case MCDisassembler::Success:
74 Streamer.EmitInstruction(Inst, STI);
82 static bool SkipToToken(StringRef &Str) {
87 // Strip horizontal whitespace and commas.
88 if (size_t Pos = Str.find_first_not_of(" \t\r\n,")) {
89 Str = Str.substr(Pos);
93 // If this is the start of a comment, remove the rest of the line.
95 Str = Str.substr(Str.find_first_of('\n'));
103 static bool ByteArrayFromString(ByteArrayTy &ByteArray,
106 while (SkipToToken(Str)) {
107 // Handled by higher level
108 if (Str[0] == '[' || Str[0] == ']')
111 // Get the current token.
112 size_t Next = Str.find_first_of(" \t\n\r,#[]");
113 StringRef Value = Str.substr(0, Next);
115 // Convert to a byte and add to the byte vector.
117 if (Value.getAsInteger(0, ByteVal) || ByteVal > 255) {
118 // If we have an error, print it and skip to the end of line.
119 SM.PrintMessage(SMLoc::getFromPointer(Value.data()), SourceMgr::DK_Error,
120 "invalid input token");
121 Str = Str.substr(Str.find('\n'));
122 ByteArray.first.clear();
123 ByteArray.second.clear();
127 ByteArray.first.push_back(ByteVal);
128 ByteArray.second.push_back(Value.data());
129 Str = Str.substr(Next);
135 int Disassembler::disassemble(const Target &T,
136 const std::string &Triple,
137 MCSubtargetInfo &STI,
138 MCStreamer &Streamer,
139 MemoryBuffer &Buffer,
143 std::unique_ptr<const MCRegisterInfo> MRI(T.createMCRegInfo(Triple));
145 errs() << "error: no register info for target " << Triple << "\n";
149 std::unique_ptr<const MCAsmInfo> MAI(T.createMCAsmInfo(*MRI, Triple));
151 errs() << "error: no assembly info for target " << Triple << "\n";
155 // Set up the MCContext for creating symbols and MCExpr's.
156 MCContext Ctx(MAI.get(), MRI.get(), nullptr);
158 std::unique_ptr<const MCDisassembler> DisAsm(
159 T.createMCDisassembler(STI, Ctx));
161 errs() << "error: no disassembler for target " << Triple << "\n";
165 // Set up initial section manually here
166 Streamer.InitSections(false);
168 bool ErrorOccurred = false;
170 // Convert the input to a vector for disassembly.
171 ByteArrayTy ByteArray;
172 StringRef Str = Buffer.getBuffer();
173 bool InAtomicBlock = false;
175 while (SkipToToken(Str)) {
176 ByteArray.first.clear();
177 ByteArray.second.clear();
181 SM.PrintMessage(SMLoc::getFromPointer(Str.data()), SourceMgr::DK_Error,
182 "nested atomic blocks make no sense");
183 ErrorOccurred = true;
185 InAtomicBlock = true;
186 Str = Str.drop_front();
188 } else if (Str[0] == ']') {
189 if (!InAtomicBlock) {
190 SM.PrintMessage(SMLoc::getFromPointer(Str.data()), SourceMgr::DK_Error,
191 "attempt to close atomic block without opening");
192 ErrorOccurred = true;
194 InAtomicBlock = false;
195 Str = Str.drop_front();
199 // It's a real token, get the bytes and emit them
200 ErrorOccurred |= ByteArrayFromString(ByteArray, Str, SM);
202 if (!ByteArray.first.empty())
203 ErrorOccurred |= PrintInsts(*DisAsm, ByteArray, SM, Out, Streamer,
208 SM.PrintMessage(SMLoc::getFromPointer(Str.data()), SourceMgr::DK_Error,
209 "unclosed atomic block");
210 ErrorOccurred = true;
213 return ErrorOccurred;