tools/llvm-mc-assemble-fuzzer/llvm-mc-assemble-fuzzer.cpp

   1 //===-- llvm-mc-assemble-fuzzer.cpp - Fuzzer for the MC layer -------------===//
   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 //===----------------------------------------------------------------------===//
  10
  11 #include "llvm-c/Target.h"
  12 #include "llvm/MC/SubtargetFeature.h"
  13 #include "llvm/MC/MCAsmBackend.h"
  14 #include "llvm/MC/MCAsmInfo.h"
  15 #include "llvm/MC/MCCodeEmitter.h"
  16 #include "llvm/MC/MCContext.h"
  17 #include "llvm/MC/MCInstPrinter.h"
  18 #include "llvm/MC/MCInstrInfo.h"
  19 #include "llvm/MC/MCObjectFileInfo.h"
  20 #include "llvm/MC/MCObjectWriter.h"
  21 #include "llvm/MC/MCParser/AsmLexer.h"
  22 #include "llvm/MC/MCParser/MCTargetAsmParser.h"
  23 #include "llvm/MC/MCRegisterInfo.h"
  24 #include "llvm/MC/MCSectionMachO.h"
  25 #include "llvm/MC/MCStreamer.h"
  26 #include "llvm/MC/MCSubtargetInfo.h"
  27 #include "llvm/MC/MCTargetOptionsCommandFlags.inc"
  28 #include "llvm/Support/MemoryBuffer.h"
  29 #include "llvm/Support/CommandLine.h"
  30 #include "llvm/Support/FileUtilities.h"
  31 #include "llvm/Support/raw_ostream.h"
  32 #include "llvm/Support/SourceMgr.h"
  33 #include "llvm/Support/TargetSelect.h"
  34 #include "llvm/Support/TargetRegistry.h"
  35 #include "llvm/Support/ToolOutputFile.h"
  36
  37 using namespace llvm;
  38
  39 static cl::opt<std::string>
  40     TripleName("triple", cl::desc("Target triple to assemble for, "
  41                                   "see -version for available targets"));
  42
  43 static cl::opt<std::string>
  44     MCPU("mcpu",
  45          cl::desc("Target a specific cpu type (-mcpu=help for details)"),
  46          cl::value_desc("cpu-name"), cl::init(""));
  47
  48 // This is useful for variable-length instruction sets.
  49 static cl::opt<unsigned> InsnLimit(
  50     "insn-limit",
  51     cl::desc("Limit the number of instructions to process (0 for no limit)"),
  52     cl::value_desc("count"), cl::init(0));
  53
  54 static cl::list<std::string>
  55     MAttrs("mattr", cl::CommaSeparated,
  56            cl::desc("Target specific attributes (-mattr=help for details)"),
  57            cl::value_desc("a1,+a2,-a3,..."));
  58 // The feature string derived from -mattr's values.
  59 std::string FeaturesStr;
  60
  61 static cl::list<std::string>
  62     FuzzerArgs("fuzzer-args", cl::Positional,
  63                cl::desc("Options to pass to the fuzzer"), cl::ZeroOrMore,
  64                cl::PositionalEatsArgs);
  65 static std::vector<char *> ModifiedArgv;
  66
  67 enum OutputFileType {
  68   OFT_Null,
  69   OFT_AssemblyFile,
  70   OFT_ObjectFile
  71 };
  72 static cl::opt<OutputFileType>
  73 FileType("filetype", cl::init(OFT_AssemblyFile),
  74   cl::desc("Choose an output file type:"),
  75   cl::values(
  76        clEnumValN(OFT_AssemblyFile, "asm",
  77                   "Emit an assembly ('.s') file"),
  78        clEnumValN(OFT_Null, "null",
  79                   "Don't emit anything (for timing purposes)"),
  80        clEnumValN(OFT_ObjectFile, "obj",
  81                   "Emit a native object ('.o') file")));
  82
  83
  84 class LLVMFuzzerInputBuffer : public MemoryBuffer
  85 {
  86   public:
  87     LLVMFuzzerInputBuffer(const uint8_t *data_, size_t size_)
  88       : Data(reinterpret_cast<const char *>(data_)),
  89         Size(size_) {
  90         init(Data, Data+Size, false);
  91       }
  92
  93
  94     virtual BufferKind getBufferKind() const {
  95       return MemoryBuffer_Malloc; // it's not disk-backed so I think that's
  96                                   // the intent ... though AFAIK it
  97                                   // probably came from an mmap or sbrk
  98     }
  99
 100   private:
 101     const char *Data;
 102     size_t Size;
 103 };
 104
 105 static int AssembleInput(const char *ProgName, const Target *TheTarget,
 106                          SourceMgr &SrcMgr, MCContext &Ctx, MCStreamer &Str,
 107                          MCAsmInfo &MAI, MCSubtargetInfo &STI,
 108                          MCInstrInfo &MCII, MCTargetOptions &MCOptions) {
 109   static const bool NoInitialTextSection = false;
 110
 111   std::unique_ptr<MCAsmParser> Parser(
 112     createMCAsmParser(SrcMgr, Ctx, Str, MAI));
 113
 114   std::unique_ptr<MCTargetAsmParser> TAP(
 115     TheTarget->createMCAsmParser(STI, *Parser, MCII, MCOptions));
 116
 117   if (!TAP) {
 118     errs() << ProgName
 119            << ": error: this target '" << TripleName
 120            << "', does not support assembly parsing.\n";
 121     abort();
 122   }
 123
 124   Parser->setTargetParser(*TAP);
 125
 126   return Parser->Run(NoInitialTextSection);
 127 }
 128
 129
 130 int AssembleOneInput(const uint8_t *Data, size_t Size) {
 131   const bool ShowInst = false;
 132   const bool AsmVerbose = false;
 133   const bool UseDwarfDirectory = true;
 134
 135   Triple TheTriple(Triple::normalize(TripleName));
 136
 137   SourceMgr SrcMgr;
 138
 139   std::unique_ptr<MemoryBuffer> BufferPtr(new LLVMFuzzerInputBuffer(Data, Size));
 140
 141   // Tell SrcMgr about this buffer, which is what the parser will pick up.
 142   SrcMgr.AddNewSourceBuffer(std::move(BufferPtr), SMLoc());
 143
 144   static const std::vector<std::string> NoIncludeDirs;
 145   SrcMgr.setIncludeDirs(NoIncludeDirs);
 146
 147   static std::string ArchName;
 148   std::string Error;
 149   const Target *TheTarget = TargetRegistry::lookupTarget(ArchName, TheTriple,
 150       Error);
 151   if (!TheTarget) {
 152     errs() << "error: this target '" << TheTriple.normalize()
 153       << "/" << ArchName << "', was not found: '" << Error << "'\n";
 154
 155     abort();
 156   }
 157
 158   std::unique_ptr<MCRegisterInfo> MRI(TheTarget->createMCRegInfo(TripleName));
 159   if (!MRI) {
 160     errs() << "Unable to create target register info!";
 161     abort();
 162   }
 163
 164   std::unique_ptr<MCAsmInfo> MAI(TheTarget->createMCAsmInfo(*MRI, TripleName));
 165   if (!MAI) {
 166     errs() << "Unable to create target asm info!";
 167     abort();
 168   }
 169
 170
 171   MCObjectFileInfo MOFI;
 172   MCContext Ctx(MAI.get(), MRI.get(), &MOFI, &SrcMgr);
 173
 174   static const bool UsePIC = false;
 175   MOFI.InitMCObjectFileInfo(TheTriple, UsePIC, Ctx);
 176
 177   const unsigned OutputAsmVariant = 0;
 178   std::unique_ptr<MCInstrInfo> MCII(TheTarget->createMCInstrInfo());
 179   MCInstPrinter *IP = TheTarget->createMCInstPrinter(Triple(TripleName), OutputAsmVariant,
 180       *MAI, *MCII, *MRI);
 181   if (!IP) {
 182     errs()
 183       << "error: unable to create instruction printer for target triple '"
 184       << TheTriple.normalize() << "' with assembly variant "
 185       << OutputAsmVariant << ".\n";
 186
 187     abort();
 188   }
 189
 190   const char *ProgName = "llvm-mc-fuzzer";
 191   std::unique_ptr<MCSubtargetInfo> STI(
 192       TheTarget->createMCSubtargetInfo(TripleName, MCPU, FeaturesStr));
 193   std::unique_ptr<MCCodeEmitter> CE = nullptr;
 194   std::unique_ptr<MCAsmBackend> MAB = nullptr;
 195
 196   MCTargetOptions MCOptions = InitMCTargetOptionsFromFlags();
 197
 198   std::string OutputString;
 199   raw_string_ostream Out(OutputString);
 200   auto FOut = std::make_unique<formatted_raw_ostream>(Out);
 201
 202   std::unique_ptr<MCStreamer> Str;
 203
 204   if (FileType == OFT_AssemblyFile) {
 205     Str.reset(TheTarget->createAsmStreamer(Ctx, std::move(FOut), AsmVerbose,
 206                                            UseDwarfDirectory, IP, std::move(CE),
 207                                            std::move(MAB), ShowInst));
 208   } else {
 209     assert(FileType == OFT_ObjectFile && "Invalid file type!");
 210
 211     std::error_code EC;
 212     const std::string OutputFilename = "-";
 213     auto Out =
 214         std::make_unique<ToolOutputFile>(OutputFilename, EC, sys::fs::OF_None);
 215     if (EC) {
 216       errs() << EC.message() << '\n';
 217       abort();
 218     }
 219
 220     // Don't waste memory on names of temp labels.
 221     Ctx.setUseNamesOnTempLabels(false);
 222
 223     std::unique_ptr<buffer_ostream> BOS;
 224     raw_pwrite_stream *OS = &Out->os();
 225     if (!Out->os().supportsSeeking()) {
 226       BOS = std::make_unique<buffer_ostream>(Out->os());
 227       OS = BOS.get();
 228     }
 229
 230     MCCodeEmitter *CE = TheTarget->createMCCodeEmitter(*MCII, *MRI, Ctx);
 231     MCAsmBackend *MAB = TheTarget->createMCAsmBackend(*STI, *MRI, MCOptions);
 232     Str.reset(TheTarget->createMCObjectStreamer(
 233         TheTriple, Ctx, std::unique_ptr<MCAsmBackend>(MAB),
 234         MAB->createObjectWriter(*OS), std::unique_ptr<MCCodeEmitter>(CE), *STI,
 235         MCOptions.MCRelaxAll, MCOptions.MCIncrementalLinkerCompatible,
 236         /*DWARFMustBeAtTheEnd*/ false));
 237   }
 238   const int Res = AssembleInput(ProgName, TheTarget, SrcMgr, Ctx, *Str, *MAI, *STI,
 239       *MCII, MCOptions);
 240
 241   (void) Res;
 242
 243   return 0;
 244 }
 245
 246 extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) {
 247   return AssembleOneInput(Data, Size);
 248 }
 249
 250 extern "C" LLVM_ATTRIBUTE_USED int LLVMFuzzerInitialize(int *argc,
 251                                                         char ***argv) {
 252   // The command line is unusual compared to other fuzzers due to the need to
 253   // specify the target. Options like -triple, -mcpu, and -mattr work like
 254   // their counterparts in llvm-mc, while -fuzzer-args collects options for the
 255   // fuzzer itself.
 256   //
 257   // Examples:
 258   //
 259   // Fuzz the big-endian MIPS32R6 disassembler using 100,000 inputs of up to
 260   // 4-bytes each and use the contents of ./corpus as the test corpus:
 261   //   llvm-mc-fuzzer -triple mips-linux-gnu -mcpu=mips32r6 -disassemble \
 262   //       -fuzzer-args -max_len=4 -runs=100000 ./corpus
 263   //
 264   // Infinitely fuzz the little-endian MIPS64R2 disassembler with the MSA
 265   // feature enabled using up to 64-byte inputs:
 266   //   llvm-mc-fuzzer -triple mipsel-linux-gnu -mcpu=mips64r2 -mattr=msa \
 267   //       -disassemble -fuzzer-args ./corpus
 268   //
 269   // If your aim is to find instructions that are not tested, then it is
 270   // advisable to constrain the maximum input size to a single instruction
 271   // using -max_len as in the first example. This results in a test corpus of
 272   // individual instructions that test unique paths. Without this constraint,
 273   // there will be considerable redundancy in the corpus.
 274
 275   char **OriginalArgv = *argv;
 276
 277   LLVMInitializeAllTargetInfos();
 278   LLVMInitializeAllTargetMCs();
 279   LLVMInitializeAllAsmParsers();
 280
 281   cl::ParseCommandLineOptions(*argc, OriginalArgv);
 282
 283   // Rebuild the argv without the arguments llvm-mc-fuzzer consumed so that
 284   // the driver can parse its arguments.
 285   //
 286   // FuzzerArgs cannot provide the non-const pointer that OriginalArgv needs.
 287   // Re-use the strings from OriginalArgv instead of copying FuzzerArg to a
 288   // non-const buffer to avoid the need to clean up when the fuzzer terminates.
 289   ModifiedArgv.push_back(OriginalArgv[0]);
 290   for (const auto &FuzzerArg : FuzzerArgs) {
 291     for (int i = 1; i < *argc; ++i) {
 292       if (FuzzerArg == OriginalArgv[i])
 293         ModifiedArgv.push_back(OriginalArgv[i]);
 294     }
 295   }
 296   *argc = ModifiedArgv.size();
 297   *argv = ModifiedArgv.data();
 298
 299   // Package up features to be passed to target/subtarget
 300   // We have to pass it via a global since the callback doesn't
 301   // permit any user data.
 302   if (MAttrs.size()) {
 303     SubtargetFeatures Features;
 304     for (unsigned i = 0; i != MAttrs.size(); ++i)
 305       Features.AddFeature(MAttrs[i]);
 306     FeaturesStr = Features.getString();
 307   }
 308
 309   if (TripleName.empty())
 310     TripleName = sys::getDefaultTargetTriple();
 311
 312   return 0;
 313 }