llvm/tools/llvm-ml/Disassembler.cpp

   1 //===- Disassembler.cpp - Disassembler for hex strings --------------------===//
   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 class implements the disassembler of strings of bytes written in
  10 // hexadecimal, from standard input or from a file.
  11 //
  12 //===----------------------------------------------------------------------===//
  13
  14 #include "Disassembler.h"
  15 #include "llvm/MC/MCAsmInfo.h"
  16 #include "llvm/MC/MCContext.h"
  17 #include "llvm/MC/MCDisassembler/MCDisassembler.h"
  18 #include "llvm/MC/MCInst.h"
  19 #include "llvm/MC/MCRegisterInfo.h"
  20 #include "llvm/MC/MCStreamer.h"
  21 #include "llvm/MC/MCSubtargetInfo.h"
  22 #include "llvm/MC/TargetRegistry.h"
  23 #include "llvm/Support/MemoryBuffer.h"
  24 #include "llvm/Support/SourceMgr.h"
  25 #include "llvm/Support/raw_ostream.h"
  26 #include "llvm/TargetParser/Triple.h"
  27
  28 using namespace llvm;
  29
  30 typedef std::pair<std::vector<unsigned char>, std::vector<const char *>>
  31     ByteArrayTy;
  32
  33 static bool PrintInsts(const MCDisassembler &DisAsm, const ByteArrayTy &Bytes,
  34                        SourceMgr &SM, raw_ostream &Out, MCStreamer &Streamer,
  35                        bool InAtomicBlock, const MCSubtargetInfo &STI) {
  36   ArrayRef<uint8_t> Data(Bytes.first.data(), Bytes.first.size());
  37
  38   // Disassemble it to strings.
  39   uint64_t Size;
  40   uint64_t Index;
  41
  42   for (Index = 0; Index < Bytes.first.size(); Index += Size) {
  43     MCInst Inst;
  44
  45     MCDisassembler::DecodeStatus S;
  46     S = DisAsm.getInstruction(Inst, Size, Data.slice(Index), Index, nulls());
  47     switch (S) {
  48     case MCDisassembler::Fail:
  49       SM.PrintMessage(SMLoc::getFromPointer(Bytes.second[Index]),
  50                       SourceMgr::DK_Warning, "invalid instruction encoding");
  51       // Don't try to resynchronise the stream in a block
  52       if (InAtomicBlock)
  53         return true;
  54
  55       if (Size == 0)
  56         Size = 1;  // skip illegible bytes
  57
  58       break;
  59
  60     case MCDisassembler::SoftFail:
  61       SM.PrintMessage(SMLoc::getFromPointer(Bytes.second[Index]),
  62                       SourceMgr::DK_Warning,
  63                       "potentially undefined instruction encoding");
  64       [[fallthrough]];
  65
  66     case MCDisassembler::Success:
  67       Streamer.emitInstruction(Inst, STI);
  68       break;
  69     }
  70   }
  71
  72   return false;
  73 }
  74
  75 static bool SkipToToken(StringRef &Str) {
  76   for (;;) {
  77     if (Str.empty())
  78       return false;
  79
  80     // Strip horizontal whitespace and commas.
  81     if (size_t Pos = Str.find_first_not_of(" \t\r\n,")) {
  82       Str = Str.substr(Pos);
  83       continue;
  84     }
  85
  86     // If this is the start of a comment, remove the rest of the line.
  87     if (Str[0] == '#') {
  88       Str = Str.substr(Str.find_first_of('\n'));
  89       continue;
  90     }
  91     return true;
  92   }
  93 }
  94
  95 static bool ByteArrayFromString(ByteArrayTy &ByteArray, StringRef &Str,
  96                                 SourceMgr &SM) {
  97   while (SkipToToken(Str)) {
  98     // Handled by higher level
  99     if (Str[0] == '[' || Str[0] == ']')
 100       return false;
 101
 102     // Get the current token.
 103     size_t Next = Str.find_first_of(" \t\n\r,#[]");
 104     StringRef Value = Str.substr(0, Next);
 105
 106     // Convert to a byte and add to the byte vector.
 107     unsigned ByteVal;
 108     if (Value.getAsInteger(0, ByteVal) || ByteVal > 255) {
 109       // If we have an error, print it and skip to the end of line.
 110       SM.PrintMessage(SMLoc::getFromPointer(Value.data()), SourceMgr::DK_Error,
 111                       "invalid input token");
 112       Str = Str.substr(Str.find('\n'));
 113       ByteArray.first.clear();
 114       ByteArray.second.clear();
 115       continue;
 116     }
 117
 118     ByteArray.first.push_back(ByteVal);
 119     ByteArray.second.push_back(Value.data());
 120     Str = Str.substr(Next);
 121   }
 122
 123   return false;
 124 }
 125
 126 int Disassembler::disassemble(const Target &T, const std::string &TripleName,
 127                               MCSubtargetInfo &STI, MCStreamer &Streamer,
 128                               MemoryBuffer &Buffer, SourceMgr &SM,
 129                               raw_ostream &Out) {
 130   std::unique_ptr<const MCRegisterInfo> MRI(T.createMCRegInfo(TripleName));
 131   if (!MRI) {
 132     errs() << "error: no register info for target " << TripleName << "\n";
 133     return -1;
 134   }
 135
 136   MCTargetOptions MCOptions;
 137   std::unique_ptr<const MCAsmInfo> MAI(
 138       T.createMCAsmInfo(*MRI, TripleName, MCOptions));
 139   if (!MAI) {
 140     errs() << "error: no assembly info for target " << TripleName << "\n";
 141     return -1;
 142   }
 143
 144   // Set up the MCContext for creating symbols and MCExpr's.
 145   MCContext Ctx(Triple(TripleName), MAI.get(), MRI.get(), &STI);
 146
 147   std::unique_ptr<const MCDisassembler> DisAsm(
 148       T.createMCDisassembler(STI, Ctx));
 149   if (!DisAsm) {
 150     errs() << "error: no disassembler for target " << TripleName << "\n";
 151     return -1;
 152   }
 153
 154   // Set up initial section manually here
 155   Streamer.initSections(false, STI);
 156
 157   bool ErrorOccurred = false;
 158
 159   // Convert the input to a vector for disassembly.
 160   ByteArrayTy ByteArray;
 161   StringRef Str = Buffer.getBuffer();
 162   bool InAtomicBlock = false;
 163
 164   while (SkipToToken(Str)) {
 165     ByteArray.first.clear();
 166     ByteArray.second.clear();
 167
 168     if (Str[0] == '[') {
 169       if (InAtomicBlock) {
 170         SM.PrintMessage(SMLoc::getFromPointer(Str.data()), SourceMgr::DK_Error,
 171                         "nested atomic blocks make no sense");
 172         ErrorOccurred = true;
 173       }
 174       InAtomicBlock = true;
 175       Str = Str.drop_front();
 176       continue;
 177     } else if (Str[0] == ']') {
 178       if (!InAtomicBlock) {
 179         SM.PrintMessage(SMLoc::getFromPointer(Str.data()), SourceMgr::DK_Error,
 180                         "attempt to close atomic block without opening");
 181         ErrorOccurred = true;
 182       }
 183       InAtomicBlock = false;
 184       Str = Str.drop_front();
 185       continue;
 186     }
 187
 188     // It's a real token, get the bytes and emit them
 189     ErrorOccurred |= ByteArrayFromString(ByteArray, Str, SM);
 190
 191     if (!ByteArray.first.empty())
 192       ErrorOccurred |=
 193           PrintInsts(*DisAsm, ByteArray, SM, Out, Streamer, InAtomicBlock, STI);
 194   }
 195
 196   if (InAtomicBlock) {
 197     SM.PrintMessage(SMLoc::getFromPointer(Str.data()), SourceMgr::DK_Error,
 198                     "unclosed atomic block");
 199     ErrorOccurred = true;
 200   }
 201
 202   return ErrorOccurred;
 203 }