1//===-- llvm-bcanalyzer.cpp - Bitcode Analyzer --------------------------===//
2//
3//                     The LLVM Compiler Infrastructure
4//
5// This file is distributed under the University of Illinois Open Source
6// License. See LICENSE.TXT for details.
7//
8//===----------------------------------------------------------------------===//
9//
10// This tool may be invoked in the following manner:
11//  llvm-bcanalyzer [options]      - Read LLVM bitcode from stdin
12//  llvm-bcanalyzer [options] x.bc - Read LLVM bitcode from the x.bc file
13//
14//  Options:
15//      --help      - Output information about command line switches
16//      --dump      - Dump low-level bitcode structure in readable format
17//
18// This tool provides analytical information about a bitcode file. It is
19// intended as an aid to developers of bitcode reading and writing software. It
20// produces on std::out a summary of the bitcode file that shows various
21// statistics about the contents of the file. By default this information is
22// detailed and contains information about individual bitcode blocks and the
23// functions in the module.
24// The tool is also able to print a bitcode file in a straight forward text
25// format that shows the containment and relationships of the information in
26// the bitcode file (-dump option).
27//
28//===----------------------------------------------------------------------===//
29
30#include "llvm/Bitcode/BitstreamReader.h"
31#include "llvm/ADT/Optional.h"
32#include "llvm/Bitcode/LLVMBitCodes.h"
33#include "llvm/Bitcode/ReaderWriter.h"
34#include "llvm/IR/Verifier.h"
35#include "llvm/Support/CommandLine.h"
36#include "llvm/Support/Format.h"
37#include "llvm/Support/ManagedStatic.h"
38#include "llvm/Support/MemoryBuffer.h"
39#include "llvm/Support/PrettyStackTrace.h"
40#include "llvm/Support/Signals.h"
41#include "llvm/Support/raw_ostream.h"
42#include <algorithm>
43#include <cctype>
44#include <map>
45#include <system_error>
46using namespace llvm;
47
48static cl::opt<std::string>
49  InputFilename(cl::Positional, cl::desc("<input bitcode>"), cl::init("-"));
50
51static cl::opt<bool> Dump("dump", cl::desc("Dump low level bitcode trace"));
52
53//===----------------------------------------------------------------------===//
54// Bitcode specific analysis.
55//===----------------------------------------------------------------------===//
56
57static cl::opt<bool> NoHistogram("disable-histogram",
58                                 cl::desc("Do not print per-code histogram"));
59
60static cl::opt<bool>
61NonSymbolic("non-symbolic",
62            cl::desc("Emit numeric info in dump even if"
63                     " symbolic info is available"));
64
65static cl::opt<std::string>
66  BlockInfoFilename("block-info",
67                    cl::desc("Use the BLOCK_INFO from the given file"));
68
69static cl::opt<bool>
70  ShowBinaryBlobs("show-binary-blobs",
71                  cl::desc("Print binary blobs using hex escapes"));
72
73namespace {
74
75/// CurStreamTypeType - A type for CurStreamType
76enum CurStreamTypeType {
77  UnknownBitstream,
78  LLVMIRBitstream
79};
80
81}
82
83/// GetBlockName - Return a symbolic block name if known, otherwise return
84/// null.
85static const char *GetBlockName(unsigned BlockID,
86                                const BitstreamReader &StreamFile,
87                                CurStreamTypeType CurStreamType) {
88  // Standard blocks for all bitcode files.
89  if (BlockID < bitc::FIRST_APPLICATION_BLOCKID) {
90    if (BlockID == bitc::BLOCKINFO_BLOCK_ID)
91      return "BLOCKINFO_BLOCK";
92    return nullptr;
93  }
94
95  // Check to see if we have a blockinfo record for this block, with a name.
96  if (const BitstreamReader::BlockInfo *Info =
97        StreamFile.getBlockInfo(BlockID)) {
98    if (!Info->Name.empty())
99      return Info->Name.c_str();
100  }
101
102
103  if (CurStreamType != LLVMIRBitstream) return nullptr;
104
105  switch (BlockID) {
106  default:                             return nullptr;
107  case bitc::MODULE_BLOCK_ID:          return "MODULE_BLOCK";
108  case bitc::PARAMATTR_BLOCK_ID:       return "PARAMATTR_BLOCK";
109  case bitc::PARAMATTR_GROUP_BLOCK_ID: return "PARAMATTR_GROUP_BLOCK_ID";
110  case bitc::TYPE_BLOCK_ID_NEW:        return "TYPE_BLOCK_ID";
111  case bitc::CONSTANTS_BLOCK_ID:       return "CONSTANTS_BLOCK";
112  case bitc::FUNCTION_BLOCK_ID:        return "FUNCTION_BLOCK";
113  case bitc::IDENTIFICATION_BLOCK_ID:
114    return "IDENTIFICATION_BLOCK_ID";
115  case bitc::VALUE_SYMTAB_BLOCK_ID:    return "VALUE_SYMTAB";
116  case bitc::METADATA_BLOCK_ID:        return "METADATA_BLOCK";
117  case bitc::METADATA_KIND_BLOCK_ID:   return "METADATA_KIND_BLOCK";
118  case bitc::METADATA_ATTACHMENT_ID:   return "METADATA_ATTACHMENT_BLOCK";
119  case bitc::USELIST_BLOCK_ID:         return "USELIST_BLOCK_ID";
120  case bitc::FUNCTION_SUMMARY_BLOCK_ID:
121                                       return "FUNCTION_SUMMARY_BLOCK";
122  case bitc::MODULE_STRTAB_BLOCK_ID:   return "MODULE_STRTAB_BLOCK";
123  }
124}
125
126/// GetCodeName - Return a symbolic code name if known, otherwise return
127/// null.
128static const char *GetCodeName(unsigned CodeID, unsigned BlockID,
129                               const BitstreamReader &StreamFile,
130                               CurStreamTypeType CurStreamType) {
131  // Standard blocks for all bitcode files.
132  if (BlockID < bitc::FIRST_APPLICATION_BLOCKID) {
133    if (BlockID == bitc::BLOCKINFO_BLOCK_ID) {
134      switch (CodeID) {
135      default: return nullptr;
136      case bitc::BLOCKINFO_CODE_SETBID:        return "SETBID";
137      case bitc::BLOCKINFO_CODE_BLOCKNAME:     return "BLOCKNAME";
138      case bitc::BLOCKINFO_CODE_SETRECORDNAME: return "SETRECORDNAME";
139      }
140    }
141    return nullptr;
142  }
143
144  // Check to see if we have a blockinfo record for this record, with a name.
145  if (const BitstreamReader::BlockInfo *Info =
146        StreamFile.getBlockInfo(BlockID)) {
147    for (unsigned i = 0, e = Info->RecordNames.size(); i != e; ++i)
148      if (Info->RecordNames[i].first == CodeID)
149        return Info->RecordNames[i].second.c_str();
150  }
151
152
153  if (CurStreamType != LLVMIRBitstream) return nullptr;
154
155#define STRINGIFY_CODE(PREFIX, CODE)                                           \
156  case bitc::PREFIX##_##CODE:                                                  \
157    return #CODE;
158  switch (BlockID) {
159  default: return nullptr;
160  case bitc::MODULE_BLOCK_ID:
161    switch (CodeID) {
162    default: return nullptr;
163      STRINGIFY_CODE(MODULE_CODE, VERSION)
164      STRINGIFY_CODE(MODULE_CODE, TRIPLE)
165      STRINGIFY_CODE(MODULE_CODE, DATALAYOUT)
166      STRINGIFY_CODE(MODULE_CODE, ASM)
167      STRINGIFY_CODE(MODULE_CODE, SECTIONNAME)
168      STRINGIFY_CODE(MODULE_CODE, DEPLIB) // FIXME: Remove in 4.0
169      STRINGIFY_CODE(MODULE_CODE, GLOBALVAR)
170      STRINGIFY_CODE(MODULE_CODE, FUNCTION)
171      STRINGIFY_CODE(MODULE_CODE, ALIAS)
172      STRINGIFY_CODE(MODULE_CODE, PURGEVALS)
173      STRINGIFY_CODE(MODULE_CODE, GCNAME)
174      STRINGIFY_CODE(MODULE_CODE, VSTOFFSET)
175      STRINGIFY_CODE(MODULE_CODE, METADATA_VALUES)
176    }
177  case bitc::IDENTIFICATION_BLOCK_ID:
178    switch (CodeID) {
179    default:
180      return nullptr;
181      STRINGIFY_CODE(IDENTIFICATION_CODE, STRING)
182      STRINGIFY_CODE(IDENTIFICATION_CODE, EPOCH)
183    }
184  case bitc::PARAMATTR_BLOCK_ID:
185    switch (CodeID) {
186    default: return nullptr;
187    // FIXME: Should these be different?
188    case bitc::PARAMATTR_CODE_ENTRY_OLD: return "ENTRY";
189    case bitc::PARAMATTR_CODE_ENTRY:     return "ENTRY";
190    case bitc::PARAMATTR_GRP_CODE_ENTRY: return "ENTRY";
191    }
192  case bitc::TYPE_BLOCK_ID_NEW:
193    switch (CodeID) {
194    default: return nullptr;
195      STRINGIFY_CODE(TYPE_CODE, NUMENTRY)
196      STRINGIFY_CODE(TYPE_CODE, VOID)
197      STRINGIFY_CODE(TYPE_CODE, FLOAT)
198      STRINGIFY_CODE(TYPE_CODE, DOUBLE)
199      STRINGIFY_CODE(TYPE_CODE, LABEL)
200      STRINGIFY_CODE(TYPE_CODE, OPAQUE)
201      STRINGIFY_CODE(TYPE_CODE, INTEGER)
202      STRINGIFY_CODE(TYPE_CODE, POINTER)
203      STRINGIFY_CODE(TYPE_CODE, ARRAY)
204      STRINGIFY_CODE(TYPE_CODE, VECTOR)
205      STRINGIFY_CODE(TYPE_CODE, X86_FP80)
206      STRINGIFY_CODE(TYPE_CODE, FP128)
207      STRINGIFY_CODE(TYPE_CODE, PPC_FP128)
208      STRINGIFY_CODE(TYPE_CODE, METADATA)
209      STRINGIFY_CODE(TYPE_CODE, STRUCT_ANON)
210      STRINGIFY_CODE(TYPE_CODE, STRUCT_NAME)
211      STRINGIFY_CODE(TYPE_CODE, STRUCT_NAMED)
212      STRINGIFY_CODE(TYPE_CODE, FUNCTION)
213    }
214
215  case bitc::CONSTANTS_BLOCK_ID:
216    switch (CodeID) {
217    default: return nullptr;
218      STRINGIFY_CODE(CST_CODE, SETTYPE)
219      STRINGIFY_CODE(CST_CODE, NULL)
220      STRINGIFY_CODE(CST_CODE, UNDEF)
221      STRINGIFY_CODE(CST_CODE, INTEGER)
222      STRINGIFY_CODE(CST_CODE, WIDE_INTEGER)
223      STRINGIFY_CODE(CST_CODE, FLOAT)
224      STRINGIFY_CODE(CST_CODE, AGGREGATE)
225      STRINGIFY_CODE(CST_CODE, STRING)
226      STRINGIFY_CODE(CST_CODE, CSTRING)
227      STRINGIFY_CODE(CST_CODE, CE_BINOP)
228      STRINGIFY_CODE(CST_CODE, CE_CAST)
229      STRINGIFY_CODE(CST_CODE, CE_GEP)
230      STRINGIFY_CODE(CST_CODE, CE_INBOUNDS_GEP)
231      STRINGIFY_CODE(CST_CODE, CE_SELECT)
232      STRINGIFY_CODE(CST_CODE, CE_EXTRACTELT)
233      STRINGIFY_CODE(CST_CODE, CE_INSERTELT)
234      STRINGIFY_CODE(CST_CODE, CE_SHUFFLEVEC)
235      STRINGIFY_CODE(CST_CODE, CE_CMP)
236      STRINGIFY_CODE(CST_CODE, INLINEASM)
237      STRINGIFY_CODE(CST_CODE, CE_SHUFVEC_EX)
238    case bitc::CST_CODE_BLOCKADDRESS:    return "CST_CODE_BLOCKADDRESS";
239      STRINGIFY_CODE(CST_CODE, DATA)
240    }
241  case bitc::FUNCTION_BLOCK_ID:
242    switch (CodeID) {
243    default: return nullptr;
244      STRINGIFY_CODE(FUNC_CODE, DECLAREBLOCKS)
245      STRINGIFY_CODE(FUNC_CODE, INST_BINOP)
246      STRINGIFY_CODE(FUNC_CODE, INST_CAST)
247      STRINGIFY_CODE(FUNC_CODE, INST_GEP_OLD)
248      STRINGIFY_CODE(FUNC_CODE, INST_INBOUNDS_GEP_OLD)
249      STRINGIFY_CODE(FUNC_CODE, INST_SELECT)
250      STRINGIFY_CODE(FUNC_CODE, INST_EXTRACTELT)
251      STRINGIFY_CODE(FUNC_CODE, INST_INSERTELT)
252      STRINGIFY_CODE(FUNC_CODE, INST_SHUFFLEVEC)
253      STRINGIFY_CODE(FUNC_CODE, INST_CMP)
254      STRINGIFY_CODE(FUNC_CODE, INST_RET)
255      STRINGIFY_CODE(FUNC_CODE, INST_BR)
256      STRINGIFY_CODE(FUNC_CODE, INST_SWITCH)
257      STRINGIFY_CODE(FUNC_CODE, INST_INVOKE)
258      STRINGIFY_CODE(FUNC_CODE, INST_UNREACHABLE)
259      STRINGIFY_CODE(FUNC_CODE, INST_CLEANUPRET)
260      STRINGIFY_CODE(FUNC_CODE, INST_CATCHRET)
261      STRINGIFY_CODE(FUNC_CODE, INST_CATCHPAD)
262      STRINGIFY_CODE(FUNC_CODE, INST_PHI)
263      STRINGIFY_CODE(FUNC_CODE, INST_ALLOCA)
264      STRINGIFY_CODE(FUNC_CODE, INST_LOAD)
265      STRINGIFY_CODE(FUNC_CODE, INST_VAARG)
266      STRINGIFY_CODE(FUNC_CODE, INST_STORE)
267      STRINGIFY_CODE(FUNC_CODE, INST_EXTRACTVAL)
268      STRINGIFY_CODE(FUNC_CODE, INST_INSERTVAL)
269      STRINGIFY_CODE(FUNC_CODE, INST_CMP2)
270      STRINGIFY_CODE(FUNC_CODE, INST_VSELECT)
271      STRINGIFY_CODE(FUNC_CODE, DEBUG_LOC_AGAIN)
272      STRINGIFY_CODE(FUNC_CODE, INST_CALL)
273      STRINGIFY_CODE(FUNC_CODE, DEBUG_LOC)
274      STRINGIFY_CODE(FUNC_CODE, INST_GEP)
275    }
276  case bitc::VALUE_SYMTAB_BLOCK_ID:
277    switch (CodeID) {
278    default: return nullptr;
279    STRINGIFY_CODE(VST_CODE, ENTRY)
280    STRINGIFY_CODE(VST_CODE, BBENTRY)
281    STRINGIFY_CODE(VST_CODE, FNENTRY)
282    STRINGIFY_CODE(VST_CODE, COMBINED_FNENTRY)
283    }
284  case bitc::MODULE_STRTAB_BLOCK_ID:
285    switch (CodeID) {
286    default:
287      return nullptr;
288      STRINGIFY_CODE(MST_CODE, ENTRY)
289    }
290  case bitc::FUNCTION_SUMMARY_BLOCK_ID:
291    switch (CodeID) {
292    default:
293      return nullptr;
294      STRINGIFY_CODE(FS_CODE, PERMODULE_ENTRY)
295      STRINGIFY_CODE(FS_CODE, COMBINED_ENTRY)
296    }
297  case bitc::METADATA_ATTACHMENT_ID:
298    switch(CodeID) {
299    default:return nullptr;
300      STRINGIFY_CODE(METADATA, ATTACHMENT)
301    }
302  case bitc::METADATA_BLOCK_ID:
303    switch(CodeID) {
304    default:return nullptr;
305      STRINGIFY_CODE(METADATA, STRING)
306      STRINGIFY_CODE(METADATA, NAME)
307      STRINGIFY_CODE(METADATA, KIND) // Older bitcode has it in a MODULE_BLOCK
308      STRINGIFY_CODE(METADATA, NODE)
309      STRINGIFY_CODE(METADATA, VALUE)
310      STRINGIFY_CODE(METADATA, OLD_NODE)
311      STRINGIFY_CODE(METADATA, OLD_FN_NODE)
312      STRINGIFY_CODE(METADATA, NAMED_NODE)
313      STRINGIFY_CODE(METADATA, DISTINCT_NODE)
314      STRINGIFY_CODE(METADATA, LOCATION)
315      STRINGIFY_CODE(METADATA, GENERIC_DEBUG)
316      STRINGIFY_CODE(METADATA, SUBRANGE)
317      STRINGIFY_CODE(METADATA, ENUMERATOR)
318      STRINGIFY_CODE(METADATA, BASIC_TYPE)
319      STRINGIFY_CODE(METADATA, FILE)
320      STRINGIFY_CODE(METADATA, DERIVED_TYPE)
321      STRINGIFY_CODE(METADATA, COMPOSITE_TYPE)
322      STRINGIFY_CODE(METADATA, SUBROUTINE_TYPE)
323      STRINGIFY_CODE(METADATA, COMPILE_UNIT)
324      STRINGIFY_CODE(METADATA, SUBPROGRAM)
325      STRINGIFY_CODE(METADATA, LEXICAL_BLOCK)
326      STRINGIFY_CODE(METADATA, LEXICAL_BLOCK_FILE)
327      STRINGIFY_CODE(METADATA, NAMESPACE)
328      STRINGIFY_CODE(METADATA, TEMPLATE_TYPE)
329      STRINGIFY_CODE(METADATA, TEMPLATE_VALUE)
330      STRINGIFY_CODE(METADATA, GLOBAL_VAR)
331      STRINGIFY_CODE(METADATA, LOCAL_VAR)
332      STRINGIFY_CODE(METADATA, EXPRESSION)
333      STRINGIFY_CODE(METADATA, OBJC_PROPERTY)
334      STRINGIFY_CODE(METADATA, IMPORTED_ENTITY)
335      STRINGIFY_CODE(METADATA, MODULE)
336    }
337  case bitc::METADATA_KIND_BLOCK_ID:
338    switch (CodeID) {
339    default:
340      return nullptr;
341      STRINGIFY_CODE(METADATA, KIND)
342    }
343  case bitc::USELIST_BLOCK_ID:
344    switch(CodeID) {
345    default:return nullptr;
346    case bitc::USELIST_CODE_DEFAULT: return "USELIST_CODE_DEFAULT";
347    case bitc::USELIST_CODE_BB:      return "USELIST_CODE_BB";
348    }
349  }
350#undef STRINGIFY_CODE
351}
352
353struct PerRecordStats {
354  unsigned NumInstances;
355  unsigned NumAbbrev;
356  uint64_t TotalBits;
357
358  PerRecordStats() : NumInstances(0), NumAbbrev(0), TotalBits(0) {}
359};
360
361struct PerBlockIDStats {
362  /// NumInstances - This the number of times this block ID has been seen.
363  unsigned NumInstances;
364
365  /// NumBits - The total size in bits of all of these blocks.
366  uint64_t NumBits;
367
368  /// NumSubBlocks - The total number of blocks these blocks contain.
369  unsigned NumSubBlocks;
370
371  /// NumAbbrevs - The total number of abbreviations.
372  unsigned NumAbbrevs;
373
374  /// NumRecords - The total number of records these blocks contain, and the
375  /// number that are abbreviated.
376  unsigned NumRecords, NumAbbreviatedRecords;
377
378  /// CodeFreq - Keep track of the number of times we see each code.
379  std::vector<PerRecordStats> CodeFreq;
380
381  PerBlockIDStats()
382    : NumInstances(0), NumBits(0),
383      NumSubBlocks(0), NumAbbrevs(0), NumRecords(0), NumAbbreviatedRecords(0) {}
384};
385
386static std::map<unsigned, PerBlockIDStats> BlockIDStats;
387
388
389
390/// Error - All bitcode analysis errors go through this function, making this a
391/// good place to breakpoint if debugging.
392static bool Error(const Twine &Err) {
393  errs() << Err << "\n";
394  return true;
395}
396
397/// ParseBlock - Read a block, updating statistics, etc.
398static bool ParseBlock(BitstreamCursor &Stream, unsigned BlockID,
399                       unsigned IndentLevel, CurStreamTypeType CurStreamType) {
400  std::string Indent(IndentLevel*2, ' ');
401  uint64_t BlockBitStart = Stream.GetCurrentBitNo();
402
403  // Get the statistics for this BlockID.
404  PerBlockIDStats &BlockStats = BlockIDStats[BlockID];
405
406  BlockStats.NumInstances++;
407
408  // BLOCKINFO is a special part of the stream.
409  if (BlockID == bitc::BLOCKINFO_BLOCK_ID) {
410    if (Dump) outs() << Indent << "<BLOCKINFO_BLOCK/>\n";
411    if (Stream.ReadBlockInfoBlock())
412      return Error("Malformed BlockInfoBlock");
413    uint64_t BlockBitEnd = Stream.GetCurrentBitNo();
414    BlockStats.NumBits += BlockBitEnd-BlockBitStart;
415    return false;
416  }
417
418  unsigned NumWords = 0;
419  if (Stream.EnterSubBlock(BlockID, &NumWords))
420    return Error("Malformed block record");
421
422  const char *BlockName = nullptr;
423  if (Dump) {
424    outs() << Indent << "<";
425    if ((BlockName = GetBlockName(BlockID, *Stream.getBitStreamReader(),
426                                  CurStreamType)))
427      outs() << BlockName;
428    else
429      outs() << "UnknownBlock" << BlockID;
430
431    if (NonSymbolic && BlockName)
432      outs() << " BlockID=" << BlockID;
433
434    outs() << " NumWords=" << NumWords
435           << " BlockCodeSize=" << Stream.getAbbrevIDWidth() << ">\n";
436  }
437
438  SmallVector<uint64_t, 64> Record;
439
440  // Read all the records for this block.
441  while (1) {
442    if (Stream.AtEndOfStream())
443      return Error("Premature end of bitstream");
444
445    uint64_t RecordStartBit = Stream.GetCurrentBitNo();
446
447    BitstreamEntry Entry =
448      Stream.advance(BitstreamCursor::AF_DontAutoprocessAbbrevs);
449
450    switch (Entry.Kind) {
451    case BitstreamEntry::Error:
452      return Error("malformed bitcode file");
453    case BitstreamEntry::EndBlock: {
454      uint64_t BlockBitEnd = Stream.GetCurrentBitNo();
455      BlockStats.NumBits += BlockBitEnd-BlockBitStart;
456      if (Dump) {
457        outs() << Indent << "</";
458        if (BlockName)
459          outs() << BlockName << ">\n";
460        else
461          outs() << "UnknownBlock" << BlockID << ">\n";
462      }
463      return false;
464    }
465
466    case BitstreamEntry::SubBlock: {
467      uint64_t SubBlockBitStart = Stream.GetCurrentBitNo();
468      if (ParseBlock(Stream, Entry.ID, IndentLevel+1, CurStreamType))
469        return true;
470      ++BlockStats.NumSubBlocks;
471      uint64_t SubBlockBitEnd = Stream.GetCurrentBitNo();
472
473      // Don't include subblock sizes in the size of this block.
474      BlockBitStart += SubBlockBitEnd-SubBlockBitStart;
475      continue;
476    }
477    case BitstreamEntry::Record:
478      // The interesting case.
479      break;
480    }
481
482    if (Entry.ID == bitc::DEFINE_ABBREV) {
483      Stream.ReadAbbrevRecord();
484      ++BlockStats.NumAbbrevs;
485      continue;
486    }
487
488    Record.clear();
489
490    ++BlockStats.NumRecords;
491
492    StringRef Blob;
493    unsigned Code = Stream.readRecord(Entry.ID, Record, &Blob);
494
495    // Increment the # occurrences of this code.
496    if (BlockStats.CodeFreq.size() <= Code)
497      BlockStats.CodeFreq.resize(Code+1);
498    BlockStats.CodeFreq[Code].NumInstances++;
499    BlockStats.CodeFreq[Code].TotalBits +=
500      Stream.GetCurrentBitNo()-RecordStartBit;
501    if (Entry.ID != bitc::UNABBREV_RECORD) {
502      BlockStats.CodeFreq[Code].NumAbbrev++;
503      ++BlockStats.NumAbbreviatedRecords;
504    }
505
506    if (Dump) {
507      outs() << Indent << "  <";
508      if (const char *CodeName =
509            GetCodeName(Code, BlockID, *Stream.getBitStreamReader(),
510                        CurStreamType))
511        outs() << CodeName;
512      else
513        outs() << "UnknownCode" << Code;
514      if (NonSymbolic &&
515          GetCodeName(Code, BlockID, *Stream.getBitStreamReader(),
516                      CurStreamType))
517        outs() << " codeid=" << Code;
518      const BitCodeAbbrev *Abbv = nullptr;
519      if (Entry.ID != bitc::UNABBREV_RECORD) {
520        Abbv = Stream.getAbbrev(Entry.ID);
521        outs() << " abbrevid=" << Entry.ID;
522      }
523
524      for (unsigned i = 0, e = Record.size(); i != e; ++i)
525        outs() << " op" << i << "=" << (int64_t)Record[i];
526
527      outs() << "/>";
528
529      if (Abbv) {
530        for (unsigned i = 1, e = Abbv->getNumOperandInfos(); i != e; ++i) {
531          const BitCodeAbbrevOp &Op = Abbv->getOperandInfo(i);
532          if (!Op.isEncoding() || Op.getEncoding() != BitCodeAbbrevOp::Array)
533            continue;
534          assert(i + 2 == e && "Array op not second to last");
535          std::string Str;
536          bool ArrayIsPrintable = true;
537          for (unsigned j = i - 1, je = Record.size(); j != je; ++j) {
538            if (!isprint(static_cast<unsigned char>(Record[j]))) {
539              ArrayIsPrintable = false;
540              break;
541            }
542            Str += (char)Record[j];
543          }
544          if (ArrayIsPrintable)
545            outs() << " record string = '" << Str << "'";
546          break;
547        }
548      }
549
550      if (Blob.data()) {
551        outs() << " blob data = ";
552        if (ShowBinaryBlobs) {
553          outs() << "'";
554          outs().write_escaped(Blob, /*hex=*/true) << "'";
555        } else {
556          bool BlobIsPrintable = true;
557          for (unsigned i = 0, e = Blob.size(); i != e; ++i)
558            if (!isprint(static_cast<unsigned char>(Blob[i]))) {
559              BlobIsPrintable = false;
560              break;
561            }
562
563          if (BlobIsPrintable)
564            outs() << "'" << Blob << "'";
565          else
566            outs() << "unprintable, " << Blob.size() << " bytes.";
567        }
568      }
569
570      outs() << "\n";
571    }
572  }
573}
574
575static void PrintSize(double Bits) {
576  outs() << format("%.2f/%.2fB/%luW", Bits, Bits/8,(unsigned long)(Bits/32));
577}
578static void PrintSize(uint64_t Bits) {
579  outs() << format("%lub/%.2fB/%luW", (unsigned long)Bits,
580                   (double)Bits/8, (unsigned long)(Bits/32));
581}
582
583static bool openBitcodeFile(StringRef Path,
584                            std::unique_ptr<MemoryBuffer> &MemBuf,
585                            BitstreamReader &StreamFile,
586                            BitstreamCursor &Stream,
587                            CurStreamTypeType &CurStreamType) {
588  // Read the input file.
589  ErrorOr<std::unique_ptr<MemoryBuffer>> MemBufOrErr =
590      MemoryBuffer::getFileOrSTDIN(Path);
591  if (std::error_code EC = MemBufOrErr.getError())
592    return Error(Twine("Error reading '") + Path + "': " + EC.message());
593  MemBuf = std::move(MemBufOrErr.get());
594
595  if (MemBuf->getBufferSize() & 3)
596    return Error("Bitcode stream should be a multiple of 4 bytes in length");
597
598  const unsigned char *BufPtr = (const unsigned char *)MemBuf->getBufferStart();
599  const unsigned char *EndBufPtr = BufPtr + MemBuf->getBufferSize();
600
601  // If we have a wrapper header, parse it and ignore the non-bc file contents.
602  // The magic number is 0x0B17C0DE stored in little endian.
603  if (isBitcodeWrapper(BufPtr, EndBufPtr))
604    if (SkipBitcodeWrapperHeader(BufPtr, EndBufPtr, true))
605      return Error("Invalid bitcode wrapper header");
606
607  StreamFile = BitstreamReader(BufPtr, EndBufPtr);
608  Stream = BitstreamCursor(StreamFile);
609  StreamFile.CollectBlockInfoNames();
610
611  // Read the stream signature.
612  char Signature[6];
613  Signature[0] = Stream.Read(8);
614  Signature[1] = Stream.Read(8);
615  Signature[2] = Stream.Read(4);
616  Signature[3] = Stream.Read(4);
617  Signature[4] = Stream.Read(4);
618  Signature[5] = Stream.Read(4);
619
620  // Autodetect the file contents, if it is one we know.
621  CurStreamType = UnknownBitstream;
622  if (Signature[0] == 'B' && Signature[1] == 'C' &&
623      Signature[2] == 0x0 && Signature[3] == 0xC &&
624      Signature[4] == 0xE && Signature[5] == 0xD)
625    CurStreamType = LLVMIRBitstream;
626
627  return false;
628}
629
630/// AnalyzeBitcode - Analyze the bitcode file specified by InputFilename.
631static int AnalyzeBitcode() {
632  std::unique_ptr<MemoryBuffer> StreamBuffer;
633  BitstreamReader StreamFile;
634  BitstreamCursor Stream;
635  CurStreamTypeType CurStreamType;
636  if (openBitcodeFile(InputFilename, StreamBuffer, StreamFile, Stream,
637                      CurStreamType))
638    return true;
639
640  // Read block info from BlockInfoFilename, if specified.
641  // The block info must be a top-level block.
642  if (!BlockInfoFilename.empty()) {
643    std::unique_ptr<MemoryBuffer> BlockInfoBuffer;
644    BitstreamReader BlockInfoFile;
645    BitstreamCursor BlockInfoCursor;
646    CurStreamTypeType BlockInfoStreamType;
647    if (openBitcodeFile(BlockInfoFilename, BlockInfoBuffer, BlockInfoFile,
648                        BlockInfoCursor, BlockInfoStreamType))
649      return true;
650
651    while (!BlockInfoCursor.AtEndOfStream()) {
652      unsigned Code = BlockInfoCursor.ReadCode();
653      if (Code != bitc::ENTER_SUBBLOCK)
654        return Error("Invalid record at top-level in block info file");
655
656      unsigned BlockID = BlockInfoCursor.ReadSubBlockID();
657      if (BlockID == bitc::BLOCKINFO_BLOCK_ID) {
658        if (BlockInfoCursor.ReadBlockInfoBlock())
659          return Error("Malformed BlockInfoBlock in block info file");
660        break;
661      }
662
663      BlockInfoCursor.SkipBlock();
664    }
665
666    StreamFile.takeBlockInfo(std::move(BlockInfoFile));
667  }
668
669  unsigned NumTopBlocks = 0;
670
671  // Parse the top-level structure.  We only allow blocks at the top-level.
672  while (!Stream.AtEndOfStream()) {
673    unsigned Code = Stream.ReadCode();
674    if (Code != bitc::ENTER_SUBBLOCK)
675      return Error("Invalid record at top-level");
676
677    unsigned BlockID = Stream.ReadSubBlockID();
678
679    if (ParseBlock(Stream, BlockID, 0, CurStreamType))
680      return true;
681    ++NumTopBlocks;
682  }
683
684  if (Dump) outs() << "\n\n";
685
686  uint64_t BufferSizeBits = StreamFile.getBitcodeBytes().getExtent() * CHAR_BIT;
687  // Print a summary of the read file.
688  outs() << "Summary of " << InputFilename << ":\n";
689  outs() << "         Total size: ";
690  PrintSize(BufferSizeBits);
691  outs() << "\n";
692  outs() << "        Stream type: ";
693  switch (CurStreamType) {
694  case UnknownBitstream: outs() << "unknown\n"; break;
695  case LLVMIRBitstream:  outs() << "LLVM IR\n"; break;
696  }
697  outs() << "  # Toplevel Blocks: " << NumTopBlocks << "\n";
698  outs() << "\n";
699
700  // Emit per-block stats.
701  outs() << "Per-block Summary:\n";
702  for (std::map<unsigned, PerBlockIDStats>::iterator I = BlockIDStats.begin(),
703       E = BlockIDStats.end(); I != E; ++I) {
704    outs() << "  Block ID #" << I->first;
705    if (const char *BlockName = GetBlockName(I->first, StreamFile,
706                                             CurStreamType))
707      outs() << " (" << BlockName << ")";
708    outs() << ":\n";
709
710    const PerBlockIDStats &Stats = I->second;
711    outs() << "      Num Instances: " << Stats.NumInstances << "\n";
712    outs() << "         Total Size: ";
713    PrintSize(Stats.NumBits);
714    outs() << "\n";
715    double pct = (Stats.NumBits * 100.0) / BufferSizeBits;
716    outs() << "    Percent of file: " << format("%2.4f%%", pct) << "\n";
717    if (Stats.NumInstances > 1) {
718      outs() << "       Average Size: ";
719      PrintSize(Stats.NumBits/(double)Stats.NumInstances);
720      outs() << "\n";
721      outs() << "  Tot/Avg SubBlocks: " << Stats.NumSubBlocks << "/"
722             << Stats.NumSubBlocks/(double)Stats.NumInstances << "\n";
723      outs() << "    Tot/Avg Abbrevs: " << Stats.NumAbbrevs << "/"
724             << Stats.NumAbbrevs/(double)Stats.NumInstances << "\n";
725      outs() << "    Tot/Avg Records: " << Stats.NumRecords << "/"
726             << Stats.NumRecords/(double)Stats.NumInstances << "\n";
727    } else {
728      outs() << "      Num SubBlocks: " << Stats.NumSubBlocks << "\n";
729      outs() << "        Num Abbrevs: " << Stats.NumAbbrevs << "\n";
730      outs() << "        Num Records: " << Stats.NumRecords << "\n";
731    }
732    if (Stats.NumRecords) {
733      double pct = (Stats.NumAbbreviatedRecords * 100.0) / Stats.NumRecords;
734      outs() << "    Percent Abbrevs: " << format("%2.4f%%", pct) << "\n";
735    }
736    outs() << "\n";
737
738    // Print a histogram of the codes we see.
739    if (!NoHistogram && !Stats.CodeFreq.empty()) {
740      std::vector<std::pair<unsigned, unsigned> > FreqPairs;  // <freq,code>
741      for (unsigned i = 0, e = Stats.CodeFreq.size(); i != e; ++i)
742        if (unsigned Freq = Stats.CodeFreq[i].NumInstances)
743          FreqPairs.push_back(std::make_pair(Freq, i));
744      std::stable_sort(FreqPairs.begin(), FreqPairs.end());
745      std::reverse(FreqPairs.begin(), FreqPairs.end());
746
747      outs() << "\tRecord Histogram:\n";
748      outs() << "\t\t  Count    # Bits   %% Abv  Record Kind\n";
749      for (unsigned i = 0, e = FreqPairs.size(); i != e; ++i) {
750        const PerRecordStats &RecStats = Stats.CodeFreq[FreqPairs[i].second];
751
752        outs() << format("\t\t%7d %9lu",
753                         RecStats.NumInstances,
754                         (unsigned long)RecStats.TotalBits);
755
756        if (RecStats.NumAbbrev)
757          outs() <<
758              format("%7.2f  ",
759                     (double)RecStats.NumAbbrev/RecStats.NumInstances*100);
760        else
761          outs() << "         ";
762
763        if (const char *CodeName =
764              GetCodeName(FreqPairs[i].second, I->first, StreamFile,
765                          CurStreamType))
766          outs() << CodeName << "\n";
767        else
768          outs() << "UnknownCode" << FreqPairs[i].second << "\n";
769      }
770      outs() << "\n";
771
772    }
773  }
774  return 0;
775}
776
777
778int main(int argc, char **argv) {
779  // Print a stack trace if we signal out.
780  sys::PrintStackTraceOnErrorSignal();
781  PrettyStackTraceProgram X(argc, argv);
782  llvm_shutdown_obj Y;  // Call llvm_shutdown() on exit.
783  cl::ParseCommandLineOptions(argc, argv, "llvm-bcanalyzer file analyzer\n");
784
785  return AnalyzeBitcode();
786}
787