1 //===-- llvm-bcanalyzer.cpp - Bitcode Analyzer --------------------------===//
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 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
15 // --help - Output information about command line switches
16 // --dump - Dump low-level bitcode structure in readable format
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).
28 //===----------------------------------------------------------------------===//
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"
45 #include <system_error>
48 static cl::opt<std::string>
49 InputFilename(cl::Positional, cl::desc("<input bitcode>"), cl::init("-"));
51 static cl::opt<bool> Dump("dump", cl::desc("Dump low level bitcode trace"));
53 //===----------------------------------------------------------------------===//
54 // Bitcode specific analysis.
55 //===----------------------------------------------------------------------===//
57 static cl::opt<bool> NoHistogram("disable-histogram",
58 cl::desc("Do not print per-code histogram"));
61 NonSymbolic("non-symbolic",
62 cl::desc("Emit numeric info in dump even if"
63 " symbolic info is available"));
65 static cl::opt<std::string>
66 BlockInfoFilename("block-info",
67 cl::desc("Use the BLOCK_INFO from the given file"));
70 ShowBinaryBlobs("show-binary-blobs",
71 cl::desc("Print binary blobs using hex escapes"));
75 /// CurStreamTypeType - A type for CurStreamType
76 enum CurStreamTypeType {
83 /// GetBlockName - Return a symbolic block name if known, otherwise return
85 static 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";
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();
103 if (CurStreamType != LLVMIRBitstream) return nullptr;
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::VALUE_SYMTAB_BLOCK_ID: return "VALUE_SYMTAB";
114 case bitc::METADATA_BLOCK_ID: return "METADATA_BLOCK";
115 case bitc::METADATA_ATTACHMENT_ID: return "METADATA_ATTACHMENT_BLOCK";
116 case bitc::USELIST_BLOCK_ID: return "USELIST_BLOCK_ID";
117 case bitc::FUNCTION_SUMMARY_BLOCK_ID:
118 return "FUNCTION_SUMMARY_BLOCK";
119 case bitc::MODULE_STRTAB_BLOCK_ID: return "MODULE_STRTAB_BLOCK";
123 /// GetCodeName - Return a symbolic code name if known, otherwise return
125 static const char *GetCodeName(unsigned CodeID, unsigned BlockID,
126 const BitstreamReader &StreamFile,
127 CurStreamTypeType CurStreamType) {
128 // Standard blocks for all bitcode files.
129 if (BlockID < bitc::FIRST_APPLICATION_BLOCKID) {
130 if (BlockID == bitc::BLOCKINFO_BLOCK_ID) {
132 default: return nullptr;
133 case bitc::BLOCKINFO_CODE_SETBID: return "SETBID";
134 case bitc::BLOCKINFO_CODE_BLOCKNAME: return "BLOCKNAME";
135 case bitc::BLOCKINFO_CODE_SETRECORDNAME: return "SETRECORDNAME";
141 // Check to see if we have a blockinfo record for this record, with a name.
142 if (const BitstreamReader::BlockInfo *Info =
143 StreamFile.getBlockInfo(BlockID)) {
144 for (unsigned i = 0, e = Info->RecordNames.size(); i != e; ++i)
145 if (Info->RecordNames[i].first == CodeID)
146 return Info->RecordNames[i].second.c_str();
150 if (CurStreamType != LLVMIRBitstream) return nullptr;
152 #define STRINGIFY_CODE(PREFIX, CODE) \
153 case bitc::PREFIX##_##CODE: \
156 default: return nullptr;
157 case bitc::MODULE_BLOCK_ID:
159 default: return nullptr;
160 STRINGIFY_CODE(MODULE_CODE, VERSION)
161 STRINGIFY_CODE(MODULE_CODE, TRIPLE)
162 STRINGIFY_CODE(MODULE_CODE, DATALAYOUT)
163 STRINGIFY_CODE(MODULE_CODE, ASM)
164 STRINGIFY_CODE(MODULE_CODE, SECTIONNAME)
165 STRINGIFY_CODE(MODULE_CODE, DEPLIB) // FIXME: Remove in 4.0
166 STRINGIFY_CODE(MODULE_CODE, GLOBALVAR)
167 STRINGIFY_CODE(MODULE_CODE, FUNCTION)
168 STRINGIFY_CODE(MODULE_CODE, ALIAS)
169 STRINGIFY_CODE(MODULE_CODE, PURGEVALS)
170 STRINGIFY_CODE(MODULE_CODE, GCNAME)
171 STRINGIFY_CODE(MODULE_CODE, VSTOFFSET)
173 case bitc::PARAMATTR_BLOCK_ID:
175 default: return nullptr;
176 // FIXME: Should these be different?
177 case bitc::PARAMATTR_CODE_ENTRY_OLD: return "ENTRY";
178 case bitc::PARAMATTR_CODE_ENTRY: return "ENTRY";
179 case bitc::PARAMATTR_GRP_CODE_ENTRY: return "ENTRY";
181 case bitc::TYPE_BLOCK_ID_NEW:
183 default: return nullptr;
184 STRINGIFY_CODE(TYPE_CODE, NUMENTRY)
185 STRINGIFY_CODE(TYPE_CODE, VOID)
186 STRINGIFY_CODE(TYPE_CODE, FLOAT)
187 STRINGIFY_CODE(TYPE_CODE, DOUBLE)
188 STRINGIFY_CODE(TYPE_CODE, LABEL)
189 STRINGIFY_CODE(TYPE_CODE, OPAQUE)
190 STRINGIFY_CODE(TYPE_CODE, INTEGER)
191 STRINGIFY_CODE(TYPE_CODE, POINTER)
192 STRINGIFY_CODE(TYPE_CODE, ARRAY)
193 STRINGIFY_CODE(TYPE_CODE, VECTOR)
194 STRINGIFY_CODE(TYPE_CODE, X86_FP80)
195 STRINGIFY_CODE(TYPE_CODE, FP128)
196 STRINGIFY_CODE(TYPE_CODE, PPC_FP128)
197 STRINGIFY_CODE(TYPE_CODE, METADATA)
198 STRINGIFY_CODE(TYPE_CODE, STRUCT_ANON)
199 STRINGIFY_CODE(TYPE_CODE, STRUCT_NAME)
200 STRINGIFY_CODE(TYPE_CODE, STRUCT_NAMED)
201 STRINGIFY_CODE(TYPE_CODE, FUNCTION)
204 case bitc::CONSTANTS_BLOCK_ID:
206 default: return nullptr;
207 STRINGIFY_CODE(CST_CODE, SETTYPE)
208 STRINGIFY_CODE(CST_CODE, NULL)
209 STRINGIFY_CODE(CST_CODE, UNDEF)
210 STRINGIFY_CODE(CST_CODE, INTEGER)
211 STRINGIFY_CODE(CST_CODE, WIDE_INTEGER)
212 STRINGIFY_CODE(CST_CODE, FLOAT)
213 STRINGIFY_CODE(CST_CODE, AGGREGATE)
214 STRINGIFY_CODE(CST_CODE, STRING)
215 STRINGIFY_CODE(CST_CODE, CSTRING)
216 STRINGIFY_CODE(CST_CODE, CE_BINOP)
217 STRINGIFY_CODE(CST_CODE, CE_CAST)
218 STRINGIFY_CODE(CST_CODE, CE_GEP)
219 STRINGIFY_CODE(CST_CODE, CE_INBOUNDS_GEP)
220 STRINGIFY_CODE(CST_CODE, CE_SELECT)
221 STRINGIFY_CODE(CST_CODE, CE_EXTRACTELT)
222 STRINGIFY_CODE(CST_CODE, CE_INSERTELT)
223 STRINGIFY_CODE(CST_CODE, CE_SHUFFLEVEC)
224 STRINGIFY_CODE(CST_CODE, CE_CMP)
225 STRINGIFY_CODE(CST_CODE, INLINEASM)
226 STRINGIFY_CODE(CST_CODE, CE_SHUFVEC_EX)
227 case bitc::CST_CODE_BLOCKADDRESS: return "CST_CODE_BLOCKADDRESS";
228 STRINGIFY_CODE(CST_CODE, DATA)
230 case bitc::FUNCTION_BLOCK_ID:
232 default: return nullptr;
233 STRINGIFY_CODE(FUNC_CODE, DECLAREBLOCKS)
234 STRINGIFY_CODE(FUNC_CODE, INST_BINOP)
235 STRINGIFY_CODE(FUNC_CODE, INST_CAST)
236 STRINGIFY_CODE(FUNC_CODE, INST_GEP_OLD)
237 STRINGIFY_CODE(FUNC_CODE, INST_INBOUNDS_GEP_OLD)
238 STRINGIFY_CODE(FUNC_CODE, INST_SELECT)
239 STRINGIFY_CODE(FUNC_CODE, INST_EXTRACTELT)
240 STRINGIFY_CODE(FUNC_CODE, INST_INSERTELT)
241 STRINGIFY_CODE(FUNC_CODE, INST_SHUFFLEVEC)
242 STRINGIFY_CODE(FUNC_CODE, INST_CMP)
243 STRINGIFY_CODE(FUNC_CODE, INST_RET)
244 STRINGIFY_CODE(FUNC_CODE, INST_BR)
245 STRINGIFY_CODE(FUNC_CODE, INST_SWITCH)
246 STRINGIFY_CODE(FUNC_CODE, INST_INVOKE)
247 STRINGIFY_CODE(FUNC_CODE, INST_UNREACHABLE)
248 STRINGIFY_CODE(FUNC_CODE, INST_CLEANUPRET)
249 STRINGIFY_CODE(FUNC_CODE, INST_CATCHRET)
250 STRINGIFY_CODE(FUNC_CODE, INST_CATCHPAD)
251 STRINGIFY_CODE(FUNC_CODE, INST_CLEANUPENDPAD)
252 STRINGIFY_CODE(FUNC_CODE, INST_CATCHENDPAD)
253 STRINGIFY_CODE(FUNC_CODE, INST_TERMINATEPAD)
254 STRINGIFY_CODE(FUNC_CODE, INST_PHI)
255 STRINGIFY_CODE(FUNC_CODE, INST_ALLOCA)
256 STRINGIFY_CODE(FUNC_CODE, INST_LOAD)
257 STRINGIFY_CODE(FUNC_CODE, INST_VAARG)
258 STRINGIFY_CODE(FUNC_CODE, INST_STORE)
259 STRINGIFY_CODE(FUNC_CODE, INST_EXTRACTVAL)
260 STRINGIFY_CODE(FUNC_CODE, INST_INSERTVAL)
261 STRINGIFY_CODE(FUNC_CODE, INST_CMP2)
262 STRINGIFY_CODE(FUNC_CODE, INST_VSELECT)
263 STRINGIFY_CODE(FUNC_CODE, DEBUG_LOC_AGAIN)
264 STRINGIFY_CODE(FUNC_CODE, INST_CALL)
265 STRINGIFY_CODE(FUNC_CODE, DEBUG_LOC)
266 STRINGIFY_CODE(FUNC_CODE, INST_GEP)
268 case bitc::VALUE_SYMTAB_BLOCK_ID:
270 default: return nullptr;
271 STRINGIFY_CODE(VST_CODE, ENTRY)
272 STRINGIFY_CODE(VST_CODE, BBENTRY)
273 STRINGIFY_CODE(VST_CODE, FNENTRY)
274 STRINGIFY_CODE(VST_CODE, COMBINED_FNENTRY)
276 case bitc::MODULE_STRTAB_BLOCK_ID:
278 default: return nullptr;
279 STRINGIFY_CODE(MST_CODE, ENTRY)
281 case bitc::FUNCTION_SUMMARY_BLOCK_ID:
283 default: return nullptr;
284 STRINGIFY_CODE(FS_CODE, PERMODULE_ENTRY)
285 STRINGIFY_CODE(FS_CODE, COMBINED_ENTRY)
287 case bitc::METADATA_ATTACHMENT_ID:
289 default:return nullptr;
290 STRINGIFY_CODE(METADATA, ATTACHMENT)
292 case bitc::METADATA_BLOCK_ID:
294 default:return nullptr;
295 STRINGIFY_CODE(METADATA, STRING)
296 STRINGIFY_CODE(METADATA, NAME)
297 STRINGIFY_CODE(METADATA, KIND)
298 STRINGIFY_CODE(METADATA, NODE)
299 STRINGIFY_CODE(METADATA, VALUE)
300 STRINGIFY_CODE(METADATA, OLD_NODE)
301 STRINGIFY_CODE(METADATA, OLD_FN_NODE)
302 STRINGIFY_CODE(METADATA, NAMED_NODE)
303 STRINGIFY_CODE(METADATA, DISTINCT_NODE)
304 STRINGIFY_CODE(METADATA, LOCATION)
305 STRINGIFY_CODE(METADATA, GENERIC_DEBUG)
306 STRINGIFY_CODE(METADATA, SUBRANGE)
307 STRINGIFY_CODE(METADATA, ENUMERATOR)
308 STRINGIFY_CODE(METADATA, BASIC_TYPE)
309 STRINGIFY_CODE(METADATA, FILE)
310 STRINGIFY_CODE(METADATA, DERIVED_TYPE)
311 STRINGIFY_CODE(METADATA, COMPOSITE_TYPE)
312 STRINGIFY_CODE(METADATA, SUBROUTINE_TYPE)
313 STRINGIFY_CODE(METADATA, COMPILE_UNIT)
314 STRINGIFY_CODE(METADATA, SUBPROGRAM)
315 STRINGIFY_CODE(METADATA, LEXICAL_BLOCK)
316 STRINGIFY_CODE(METADATA, LEXICAL_BLOCK_FILE)
317 STRINGIFY_CODE(METADATA, NAMESPACE)
318 STRINGIFY_CODE(METADATA, TEMPLATE_TYPE)
319 STRINGIFY_CODE(METADATA, TEMPLATE_VALUE)
320 STRINGIFY_CODE(METADATA, GLOBAL_VAR)
321 STRINGIFY_CODE(METADATA, LOCAL_VAR)
322 STRINGIFY_CODE(METADATA, EXPRESSION)
323 STRINGIFY_CODE(METADATA, OBJC_PROPERTY)
324 STRINGIFY_CODE(METADATA, IMPORTED_ENTITY)
325 STRINGIFY_CODE(METADATA, MODULE)
327 case bitc::USELIST_BLOCK_ID:
329 default:return nullptr;
330 case bitc::USELIST_CODE_DEFAULT: return "USELIST_CODE_DEFAULT";
331 case bitc::USELIST_CODE_BB: return "USELIST_CODE_BB";
334 #undef STRINGIFY_CODE
337 struct PerRecordStats {
338 unsigned NumInstances;
342 PerRecordStats() : NumInstances(0), NumAbbrev(0), TotalBits(0) {}
345 struct PerBlockIDStats {
346 /// NumInstances - This the number of times this block ID has been seen.
347 unsigned NumInstances;
349 /// NumBits - The total size in bits of all of these blocks.
352 /// NumSubBlocks - The total number of blocks these blocks contain.
353 unsigned NumSubBlocks;
355 /// NumAbbrevs - The total number of abbreviations.
358 /// NumRecords - The total number of records these blocks contain, and the
359 /// number that are abbreviated.
360 unsigned NumRecords, NumAbbreviatedRecords;
362 /// CodeFreq - Keep track of the number of times we see each code.
363 std::vector<PerRecordStats> CodeFreq;
366 : NumInstances(0), NumBits(0),
367 NumSubBlocks(0), NumAbbrevs(0), NumRecords(0), NumAbbreviatedRecords(0) {}
370 static std::map<unsigned, PerBlockIDStats> BlockIDStats;
374 /// Error - All bitcode analysis errors go through this function, making this a
375 /// good place to breakpoint if debugging.
376 static bool Error(const Twine &Err) {
377 errs() << Err << "\n";
381 /// ParseBlock - Read a block, updating statistics, etc.
382 static bool ParseBlock(BitstreamCursor &Stream, unsigned BlockID,
383 unsigned IndentLevel, CurStreamTypeType CurStreamType) {
384 std::string Indent(IndentLevel*2, ' ');
385 uint64_t BlockBitStart = Stream.GetCurrentBitNo();
387 // Get the statistics for this BlockID.
388 PerBlockIDStats &BlockStats = BlockIDStats[BlockID];
390 BlockStats.NumInstances++;
392 // BLOCKINFO is a special part of the stream.
393 if (BlockID == bitc::BLOCKINFO_BLOCK_ID) {
394 if (Dump) outs() << Indent << "<BLOCKINFO_BLOCK/>\n";
395 if (Stream.ReadBlockInfoBlock())
396 return Error("Malformed BlockInfoBlock");
397 uint64_t BlockBitEnd = Stream.GetCurrentBitNo();
398 BlockStats.NumBits += BlockBitEnd-BlockBitStart;
402 unsigned NumWords = 0;
403 if (Stream.EnterSubBlock(BlockID, &NumWords))
404 return Error("Malformed block record");
406 const char *BlockName = nullptr;
408 outs() << Indent << "<";
409 if ((BlockName = GetBlockName(BlockID, *Stream.getBitStreamReader(),
413 outs() << "UnknownBlock" << BlockID;
415 if (NonSymbolic && BlockName)
416 outs() << " BlockID=" << BlockID;
418 outs() << " NumWords=" << NumWords
419 << " BlockCodeSize=" << Stream.getAbbrevIDWidth() << ">\n";
422 SmallVector<uint64_t, 64> Record;
424 // Read all the records for this block.
426 if (Stream.AtEndOfStream())
427 return Error("Premature end of bitstream");
429 uint64_t RecordStartBit = Stream.GetCurrentBitNo();
431 BitstreamEntry Entry =
432 Stream.advance(BitstreamCursor::AF_DontAutoprocessAbbrevs);
434 switch (Entry.Kind) {
435 case BitstreamEntry::Error:
436 return Error("malformed bitcode file");
437 case BitstreamEntry::EndBlock: {
438 uint64_t BlockBitEnd = Stream.GetCurrentBitNo();
439 BlockStats.NumBits += BlockBitEnd-BlockBitStart;
441 outs() << Indent << "</";
443 outs() << BlockName << ">\n";
445 outs() << "UnknownBlock" << BlockID << ">\n";
450 case BitstreamEntry::SubBlock: {
451 uint64_t SubBlockBitStart = Stream.GetCurrentBitNo();
452 if (ParseBlock(Stream, Entry.ID, IndentLevel+1, CurStreamType))
454 ++BlockStats.NumSubBlocks;
455 uint64_t SubBlockBitEnd = Stream.GetCurrentBitNo();
457 // Don't include subblock sizes in the size of this block.
458 BlockBitStart += SubBlockBitEnd-SubBlockBitStart;
461 case BitstreamEntry::Record:
462 // The interesting case.
466 if (Entry.ID == bitc::DEFINE_ABBREV) {
467 Stream.ReadAbbrevRecord();
468 ++BlockStats.NumAbbrevs;
474 ++BlockStats.NumRecords;
477 unsigned Code = Stream.readRecord(Entry.ID, Record, &Blob);
479 // Increment the # occurrences of this code.
480 if (BlockStats.CodeFreq.size() <= Code)
481 BlockStats.CodeFreq.resize(Code+1);
482 BlockStats.CodeFreq[Code].NumInstances++;
483 BlockStats.CodeFreq[Code].TotalBits +=
484 Stream.GetCurrentBitNo()-RecordStartBit;
485 if (Entry.ID != bitc::UNABBREV_RECORD) {
486 BlockStats.CodeFreq[Code].NumAbbrev++;
487 ++BlockStats.NumAbbreviatedRecords;
491 outs() << Indent << " <";
492 if (const char *CodeName =
493 GetCodeName(Code, BlockID, *Stream.getBitStreamReader(),
497 outs() << "UnknownCode" << Code;
499 GetCodeName(Code, BlockID, *Stream.getBitStreamReader(),
501 outs() << " codeid=" << Code;
502 const BitCodeAbbrev *Abbv = nullptr;
503 if (Entry.ID != bitc::UNABBREV_RECORD) {
504 Abbv = Stream.getAbbrev(Entry.ID);
505 outs() << " abbrevid=" << Entry.ID;
508 for (unsigned i = 0, e = Record.size(); i != e; ++i)
509 outs() << " op" << i << "=" << (int64_t)Record[i];
514 for (unsigned i = 1, e = Abbv->getNumOperandInfos(); i != e; ++i) {
515 const BitCodeAbbrevOp &Op = Abbv->getOperandInfo(i);
516 if (!Op.isEncoding() || Op.getEncoding() != BitCodeAbbrevOp::Array)
518 assert(i + 2 == e && "Array op not second to last");
520 bool ArrayIsPrintable = true;
521 for (unsigned j = i - 1, je = Record.size(); j != je; ++j) {
522 if (!isprint(static_cast<unsigned char>(Record[j]))) {
523 ArrayIsPrintable = false;
526 Str += (char)Record[j];
528 if (ArrayIsPrintable) outs() << " record string = '" << Str << "'";
534 outs() << " blob data = ";
535 if (ShowBinaryBlobs) {
537 outs().write_escaped(Blob, /*hex=*/true) << "'";
539 bool BlobIsPrintable = true;
540 for (unsigned i = 0, e = Blob.size(); i != e; ++i)
541 if (!isprint(static_cast<unsigned char>(Blob[i]))) {
542 BlobIsPrintable = false;
547 outs() << "'" << Blob << "'";
549 outs() << "unprintable, " << Blob.size() << " bytes.";
558 static void PrintSize(double Bits) {
559 outs() << format("%.2f/%.2fB/%luW", Bits, Bits/8,(unsigned long)(Bits/32));
561 static void PrintSize(uint64_t Bits) {
562 outs() << format("%lub/%.2fB/%luW", (unsigned long)Bits,
563 (double)Bits/8, (unsigned long)(Bits/32));
566 static bool openBitcodeFile(StringRef Path,
567 std::unique_ptr<MemoryBuffer> &MemBuf,
568 BitstreamReader &StreamFile,
569 BitstreamCursor &Stream,
570 CurStreamTypeType &CurStreamType) {
571 // Read the input file.
572 ErrorOr<std::unique_ptr<MemoryBuffer>> MemBufOrErr =
573 MemoryBuffer::getFileOrSTDIN(Path);
574 if (std::error_code EC = MemBufOrErr.getError())
575 return Error(Twine("Error reading '") + Path + "': " + EC.message());
576 MemBuf = std::move(MemBufOrErr.get());
578 if (MemBuf->getBufferSize() & 3)
579 return Error("Bitcode stream should be a multiple of 4 bytes in length");
581 const unsigned char *BufPtr = (const unsigned char *)MemBuf->getBufferStart();
582 const unsigned char *EndBufPtr = BufPtr + MemBuf->getBufferSize();
584 // If we have a wrapper header, parse it and ignore the non-bc file contents.
585 // The magic number is 0x0B17C0DE stored in little endian.
586 if (isBitcodeWrapper(BufPtr, EndBufPtr))
587 if (SkipBitcodeWrapperHeader(BufPtr, EndBufPtr, true))
588 return Error("Invalid bitcode wrapper header");
590 StreamFile = BitstreamReader(BufPtr, EndBufPtr);
591 Stream = BitstreamCursor(StreamFile);
592 StreamFile.CollectBlockInfoNames();
594 // Read the stream signature.
596 Signature[0] = Stream.Read(8);
597 Signature[1] = Stream.Read(8);
598 Signature[2] = Stream.Read(4);
599 Signature[3] = Stream.Read(4);
600 Signature[4] = Stream.Read(4);
601 Signature[5] = Stream.Read(4);
603 // Autodetect the file contents, if it is one we know.
604 CurStreamType = UnknownBitstream;
605 if (Signature[0] == 'B' && Signature[1] == 'C' &&
606 Signature[2] == 0x0 && Signature[3] == 0xC &&
607 Signature[4] == 0xE && Signature[5] == 0xD)
608 CurStreamType = LLVMIRBitstream;
613 /// AnalyzeBitcode - Analyze the bitcode file specified by InputFilename.
614 static int AnalyzeBitcode() {
615 std::unique_ptr<MemoryBuffer> StreamBuffer;
616 BitstreamReader StreamFile;
617 BitstreamCursor Stream;
618 CurStreamTypeType CurStreamType;
619 if (openBitcodeFile(InputFilename, StreamBuffer, StreamFile, Stream,
623 // Read block info from BlockInfoFilename, if specified.
624 // The block info must be a top-level block.
625 if (!BlockInfoFilename.empty()) {
626 std::unique_ptr<MemoryBuffer> BlockInfoBuffer;
627 BitstreamReader BlockInfoFile;
628 BitstreamCursor BlockInfoCursor;
629 CurStreamTypeType BlockInfoStreamType;
630 if (openBitcodeFile(BlockInfoFilename, BlockInfoBuffer, BlockInfoFile,
631 BlockInfoCursor, BlockInfoStreamType))
634 while (!BlockInfoCursor.AtEndOfStream()) {
635 unsigned Code = BlockInfoCursor.ReadCode();
636 if (Code != bitc::ENTER_SUBBLOCK)
637 return Error("Invalid record at top-level in block info file");
639 unsigned BlockID = BlockInfoCursor.ReadSubBlockID();
640 if (BlockID == bitc::BLOCKINFO_BLOCK_ID) {
641 if (BlockInfoCursor.ReadBlockInfoBlock())
642 return Error("Malformed BlockInfoBlock in block info file");
646 BlockInfoCursor.SkipBlock();
649 StreamFile.takeBlockInfo(std::move(BlockInfoFile));
652 unsigned NumTopBlocks = 0;
654 // Parse the top-level structure. We only allow blocks at the top-level.
655 while (!Stream.AtEndOfStream()) {
656 unsigned Code = Stream.ReadCode();
657 if (Code != bitc::ENTER_SUBBLOCK)
658 return Error("Invalid record at top-level");
660 unsigned BlockID = Stream.ReadSubBlockID();
662 if (ParseBlock(Stream, BlockID, 0, CurStreamType))
667 if (Dump) outs() << "\n\n";
669 uint64_t BufferSizeBits = StreamFile.getBitcodeBytes().getExtent() * CHAR_BIT;
670 // Print a summary of the read file.
671 outs() << "Summary of " << InputFilename << ":\n";
672 outs() << " Total size: ";
673 PrintSize(BufferSizeBits);
675 outs() << " Stream type: ";
676 switch (CurStreamType) {
677 case UnknownBitstream: outs() << "unknown\n"; break;
678 case LLVMIRBitstream: outs() << "LLVM IR\n"; break;
680 outs() << " # Toplevel Blocks: " << NumTopBlocks << "\n";
683 // Emit per-block stats.
684 outs() << "Per-block Summary:\n";
685 for (std::map<unsigned, PerBlockIDStats>::iterator I = BlockIDStats.begin(),
686 E = BlockIDStats.end(); I != E; ++I) {
687 outs() << " Block ID #" << I->first;
688 if (const char *BlockName = GetBlockName(I->first, StreamFile,
690 outs() << " (" << BlockName << ")";
693 const PerBlockIDStats &Stats = I->second;
694 outs() << " Num Instances: " << Stats.NumInstances << "\n";
695 outs() << " Total Size: ";
696 PrintSize(Stats.NumBits);
698 double pct = (Stats.NumBits * 100.0) / BufferSizeBits;
699 outs() << " Percent of file: " << format("%2.4f%%", pct) << "\n";
700 if (Stats.NumInstances > 1) {
701 outs() << " Average Size: ";
702 PrintSize(Stats.NumBits/(double)Stats.NumInstances);
704 outs() << " Tot/Avg SubBlocks: " << Stats.NumSubBlocks << "/"
705 << Stats.NumSubBlocks/(double)Stats.NumInstances << "\n";
706 outs() << " Tot/Avg Abbrevs: " << Stats.NumAbbrevs << "/"
707 << Stats.NumAbbrevs/(double)Stats.NumInstances << "\n";
708 outs() << " Tot/Avg Records: " << Stats.NumRecords << "/"
709 << Stats.NumRecords/(double)Stats.NumInstances << "\n";
711 outs() << " Num SubBlocks: " << Stats.NumSubBlocks << "\n";
712 outs() << " Num Abbrevs: " << Stats.NumAbbrevs << "\n";
713 outs() << " Num Records: " << Stats.NumRecords << "\n";
715 if (Stats.NumRecords) {
716 double pct = (Stats.NumAbbreviatedRecords * 100.0) / Stats.NumRecords;
717 outs() << " Percent Abbrevs: " << format("%2.4f%%", pct) << "\n";
721 // Print a histogram of the codes we see.
722 if (!NoHistogram && !Stats.CodeFreq.empty()) {
723 std::vector<std::pair<unsigned, unsigned> > FreqPairs; // <freq,code>
724 for (unsigned i = 0, e = Stats.CodeFreq.size(); i != e; ++i)
725 if (unsigned Freq = Stats.CodeFreq[i].NumInstances)
726 FreqPairs.push_back(std::make_pair(Freq, i));
727 std::stable_sort(FreqPairs.begin(), FreqPairs.end());
728 std::reverse(FreqPairs.begin(), FreqPairs.end());
730 outs() << "\tRecord Histogram:\n";
731 outs() << "\t\t Count # Bits %% Abv Record Kind\n";
732 for (unsigned i = 0, e = FreqPairs.size(); i != e; ++i) {
733 const PerRecordStats &RecStats = Stats.CodeFreq[FreqPairs[i].second];
735 outs() << format("\t\t%7d %9lu",
736 RecStats.NumInstances,
737 (unsigned long)RecStats.TotalBits);
739 if (RecStats.NumAbbrev)
742 (double)RecStats.NumAbbrev/RecStats.NumInstances*100);
746 if (const char *CodeName =
747 GetCodeName(FreqPairs[i].second, I->first, StreamFile,
749 outs() << CodeName << "\n";
751 outs() << "UnknownCode" << FreqPairs[i].second << "\n";
761 int main(int argc, char **argv) {
762 // Print a stack trace if we signal out.
763 sys::PrintStackTraceOnErrorSignal();
764 PrettyStackTraceProgram X(argc, argv);
765 llvm_shutdown_obj Y; // Call llvm_shutdown() on exit.
766 cl::ParseCommandLineOptions(argc, argv, "llvm-bcanalyzer file analyzer\n");
768 return AnalyzeBitcode();