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/SourceMgr.h"
26 #include "llvm/Support/TargetRegistry.h"
27 #include "llvm/Support/raw_ostream.h"
31 typedef std::pair<std::vector<unsigned char>, std::vector<const char *>>
34 static bool PrintInsts(const MCDisassembler &DisAsm,
35 const ByteArrayTy &Bytes,
36 SourceMgr &SM, raw_ostream &Out,
37 MCStreamer &Streamer, bool InAtomicBlock,
38 const MCSubtargetInfo &STI) {
39 // Wrap the vector in a MemoryObject.
40 ArrayRef<uint8_t> Data(Bytes.first.data(), Bytes.first.size());
42 // Disassemble it to strings.
46 for (Index = 0; Index < Bytes.first.size(); Index += Size) {
49 MCDisassembler::DecodeStatus S;
50 S = DisAsm.getInstruction(Inst, Size, Data.slice(Index), Index,
51 /*REMOVE*/ nulls(), nulls());
53 case MCDisassembler::Fail:
54 SM.PrintMessage(SMLoc::getFromPointer(Bytes.second[Index]),
55 SourceMgr::DK_Warning,
56 "invalid instruction encoding");
57 // Don't try to resynchronise the stream in a block
62 Size = 1; // skip illegible bytes
66 case MCDisassembler::SoftFail:
67 SM.PrintMessage(SMLoc::getFromPointer(Bytes.second[Index]),
68 SourceMgr::DK_Warning,
69 "potentially undefined instruction encoding");
72 case MCDisassembler::Success:
73 Streamer.EmitInstruction(Inst, STI);
81 static bool SkipToToken(StringRef &Str) {
86 // Strip horizontal whitespace and commas.
87 if (size_t Pos = Str.find_first_not_of(" \t\r\n,")) {
88 Str = Str.substr(Pos);
92 // If this is the start of a comment, remove the rest of the line.
94 Str = Str.substr(Str.find_first_of('\n'));
102 static bool ByteArrayFromString(ByteArrayTy &ByteArray,
105 while (SkipToToken(Str)) {
106 // Handled by higher level
107 if (Str[0] == '[' || Str[0] == ']')
110 // Get the current token.
111 size_t Next = Str.find_first_of(" \t\n\r,#[]");
112 StringRef Value = Str.substr(0, Next);
114 // Convert to a byte and add to the byte vector.
116 if (Value.getAsInteger(0, ByteVal) || ByteVal > 255) {
117 // If we have an error, print it and skip to the end of line.
118 SM.PrintMessage(SMLoc::getFromPointer(Value.data()), SourceMgr::DK_Error,
119 "invalid input token");
120 Str = Str.substr(Str.find('\n'));
121 ByteArray.first.clear();
122 ByteArray.second.clear();
126 ByteArray.first.push_back(ByteVal);
127 ByteArray.second.push_back(Value.data());
128 Str = Str.substr(Next);
134 int Disassembler::disassemble(const Target &T,
135 const std::string &Triple,
136 MCSubtargetInfo &STI,
137 MCStreamer &Streamer,
138 MemoryBuffer &Buffer,
142 std::unique_ptr<const MCRegisterInfo> MRI(T.createMCRegInfo(Triple));
144 errs() << "error: no register info for target " << Triple << "\n";
148 std::unique_ptr<const MCAsmInfo> MAI(T.createMCAsmInfo(*MRI, Triple));
150 errs() << "error: no assembly info for target " << Triple << "\n";
154 // Set up the MCContext for creating symbols and MCExpr's.
155 MCContext Ctx(MAI.get(), MRI.get(), nullptr);
157 std::unique_ptr<const MCDisassembler> DisAsm(
158 T.createMCDisassembler(STI, Ctx));
160 errs() << "error: no disassembler for target " << Triple << "\n";
164 // Set up initial section manually here
165 Streamer.InitSections(false);
167 bool ErrorOccurred = false;
169 // Convert the input to a vector for disassembly.
170 ByteArrayTy ByteArray;
171 StringRef Str = Buffer.getBuffer();
172 bool InAtomicBlock = false;
174 while (SkipToToken(Str)) {
175 ByteArray.first.clear();
176 ByteArray.second.clear();
180 SM.PrintMessage(SMLoc::getFromPointer(Str.data()), SourceMgr::DK_Error,
181 "nested atomic blocks make no sense");
182 ErrorOccurred = true;
184 InAtomicBlock = true;
185 Str = Str.drop_front();
187 } else if (Str[0] == ']') {
188 if (!InAtomicBlock) {
189 SM.PrintMessage(SMLoc::getFromPointer(Str.data()), SourceMgr::DK_Error,
190 "attempt to close atomic block without opening");
191 ErrorOccurred = true;
193 InAtomicBlock = false;
194 Str = Str.drop_front();
198 // It's a real token, get the bytes and emit them
199 ErrorOccurred |= ByteArrayFromString(ByteArray, Str, SM);
201 if (!ByteArray.first.empty())
202 ErrorOccurred |= PrintInsts(*DisAsm, ByteArray, SM, Out, Streamer,
207 SM.PrintMessage(SMLoc::getFromPointer(Str.data()), SourceMgr::DK_Error,
208 "unclosed atomic block");
209 ErrorOccurred = true;
212 return ErrorOccurred;