blob: 7a8feca0e5c6b63bfbd6857420edcb82a2263930 [file] [log] [blame]
Dan Gohmanf17a25c2007-07-18 16:29:46 +00001//===-- llvm-bcanalyzer.cpp - Bitcode Analyzer --------------------------===//
2//
3// The LLVM Compiler Infrastructure
4//
Chris Lattner5f5a5732007-12-29 20:44:31 +00005// This file is distributed under the University of Illinois Open Source
6// License. See LICENSE.TXT for details.
Dan Gohmanf17a25c2007-07-18 16:29:46 +00007//
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/Analysis/Verifier.h"
31#include "llvm/Bitcode/BitstreamReader.h"
32#include "llvm/Bitcode/LLVMBitCodes.h"
Chris Lattnerabd0e442009-04-06 20:54:32 +000033#include "llvm/Bitcode/ReaderWriter.h"
Dan Gohmanf17a25c2007-07-18 16:29:46 +000034#include "llvm/Support/CommandLine.h"
35#include "llvm/Support/ManagedStatic.h"
36#include "llvm/Support/MemoryBuffer.h"
Chris Lattnere6012df2009-03-06 05:34:10 +000037#include "llvm/Support/PrettyStackTrace.h"
Chris Lattnerb1aa85b2009-08-23 22:45:37 +000038#include "llvm/Support/raw_ostream.h"
Dan Gohmanf17a25c2007-07-18 16:29:46 +000039#include "llvm/System/Signals.h"
40#include <map>
Dan Gohmanf17a25c2007-07-18 16:29:46 +000041#include <algorithm>
42using namespace llvm;
43
44static cl::opt<std::string>
45 InputFilename(cl::Positional, cl::desc("<input bitcode>"), cl::init("-"));
46
47static cl::opt<std::string>
48 OutputFilename("-o", cl::init("-"), cl::desc("<output file>"));
49
50static cl::opt<bool> Dump("dump", cl::desc("Dump low level bitcode trace"));
51
52//===----------------------------------------------------------------------===//
53// Bitcode specific analysis.
54//===----------------------------------------------------------------------===//
55
56static cl::opt<bool> NoHistogram("disable-histogram",
57 cl::desc("Do not print per-code histogram"));
58
59static cl::opt<bool>
60NonSymbolic("non-symbolic",
61 cl::desc("Emit numberic info in dump even if"
62 " symbolic info is available"));
63
64/// CurStreamType - If we can sniff the flavor of this stream, we can produce
65/// better dump info.
66static enum {
67 UnknownBitstream,
68 LLVMIRBitstream
69} CurStreamType;
70
71
72/// GetBlockName - Return a symbolic block name if known, otherwise return
73/// null.
Chris Lattner71101322009-04-26 22:21:57 +000074static const char *GetBlockName(unsigned BlockID,
75 const BitstreamReader &StreamFile) {
Dan Gohmanf17a25c2007-07-18 16:29:46 +000076 // Standard blocks for all bitcode files.
77 if (BlockID < bitc::FIRST_APPLICATION_BLOCKID) {
78 if (BlockID == bitc::BLOCKINFO_BLOCK_ID)
79 return "BLOCKINFO_BLOCK";
80 return 0;
81 }
82
Chris Lattner71101322009-04-26 22:21:57 +000083 // Check to see if we have a blockinfo record for this block, with a name.
84 if (const BitstreamReader::BlockInfo *Info =
85 StreamFile.getBlockInfo(BlockID)) {
86 if (!Info->Name.empty())
87 return Info->Name.c_str();
88 }
89
90
Dan Gohmanf17a25c2007-07-18 16:29:46 +000091 if (CurStreamType != LLVMIRBitstream) return 0;
92
93 switch (BlockID) {
94 default: return 0;
95 case bitc::MODULE_BLOCK_ID: return "MODULE_BLOCK";
96 case bitc::PARAMATTR_BLOCK_ID: return "PARAMATTR_BLOCK";
97 case bitc::TYPE_BLOCK_ID: return "TYPE_BLOCK";
98 case bitc::CONSTANTS_BLOCK_ID: return "CONSTANTS_BLOCK";
99 case bitc::FUNCTION_BLOCK_ID: return "FUNCTION_BLOCK";
100 case bitc::TYPE_SYMTAB_BLOCK_ID: return "TYPE_SYMTAB";
101 case bitc::VALUE_SYMTAB_BLOCK_ID: return "VALUE_SYMTAB";
Devang Patel0c0a6ca2009-07-22 17:43:22 +0000102 case bitc::METADATA_BLOCK_ID: return "METADATA_BLOCK";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000103 }
104}
105
106/// GetCodeName - Return a symbolic code name if known, otherwise return
107/// null.
Chris Lattner71101322009-04-26 22:21:57 +0000108static const char *GetCodeName(unsigned CodeID, unsigned BlockID,
109 const BitstreamReader &StreamFile) {
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000110 // Standard blocks for all bitcode files.
111 if (BlockID < bitc::FIRST_APPLICATION_BLOCKID) {
112 if (BlockID == bitc::BLOCKINFO_BLOCK_ID) {
113 switch (CodeID) {
114 default: return 0;
Chris Lattner71101322009-04-26 22:21:57 +0000115 case bitc::BLOCKINFO_CODE_SETBID: return "SETBID";
116 case bitc::BLOCKINFO_CODE_BLOCKNAME: return "BLOCKNAME";
117 case bitc::BLOCKINFO_CODE_SETRECORDNAME: return "SETRECORDNAME";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000118 }
119 }
120 return 0;
121 }
122
Chris Lattner71101322009-04-26 22:21:57 +0000123 // Check to see if we have a blockinfo record for this record, with a name.
124 if (const BitstreamReader::BlockInfo *Info =
125 StreamFile.getBlockInfo(BlockID)) {
126 for (unsigned i = 0, e = Info->RecordNames.size(); i != e; ++i)
127 if (Info->RecordNames[i].first == CodeID)
128 return Info->RecordNames[i].second.c_str();
129 }
130
131
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000132 if (CurStreamType != LLVMIRBitstream) return 0;
133
134 switch (BlockID) {
135 default: return 0;
136 case bitc::MODULE_BLOCK_ID:
137 switch (CodeID) {
138 default: return 0;
139 case bitc::MODULE_CODE_VERSION: return "VERSION";
140 case bitc::MODULE_CODE_TRIPLE: return "TRIPLE";
141 case bitc::MODULE_CODE_DATALAYOUT: return "DATALAYOUT";
142 case bitc::MODULE_CODE_ASM: return "ASM";
143 case bitc::MODULE_CODE_SECTIONNAME: return "SECTIONNAME";
144 case bitc::MODULE_CODE_DEPLIB: return "DEPLIB";
145 case bitc::MODULE_CODE_GLOBALVAR: return "GLOBALVAR";
146 case bitc::MODULE_CODE_FUNCTION: return "FUNCTION";
147 case bitc::MODULE_CODE_ALIAS: return "ALIAS";
148 case bitc::MODULE_CODE_PURGEVALS: return "PURGEVALS";
Nick Lewyckyb90b8412008-11-07 14:52:51 +0000149 case bitc::MODULE_CODE_GCNAME: return "GCNAME";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000150 }
151 case bitc::PARAMATTR_BLOCK_ID:
152 switch (CodeID) {
153 default: return 0;
154 case bitc::PARAMATTR_CODE_ENTRY: return "ENTRY";
155 }
156 case bitc::TYPE_BLOCK_ID:
157 switch (CodeID) {
158 default: return 0;
Nick Lewyckyb90b8412008-11-07 14:52:51 +0000159 case bitc::TYPE_CODE_NUMENTRY: return "NUMENTRY";
160 case bitc::TYPE_CODE_VOID: return "VOID";
161 case bitc::TYPE_CODE_FLOAT: return "FLOAT";
162 case bitc::TYPE_CODE_DOUBLE: return "DOUBLE";
163 case bitc::TYPE_CODE_LABEL: return "LABEL";
164 case bitc::TYPE_CODE_OPAQUE: return "OPAQUE";
165 case bitc::TYPE_CODE_INTEGER: return "INTEGER";
166 case bitc::TYPE_CODE_POINTER: return "POINTER";
167 case bitc::TYPE_CODE_FUNCTION: return "FUNCTION";
168 case bitc::TYPE_CODE_STRUCT: return "STRUCT";
169 case bitc::TYPE_CODE_ARRAY: return "ARRAY";
170 case bitc::TYPE_CODE_VECTOR: return "VECTOR";
171 case bitc::TYPE_CODE_X86_FP80: return "X86_FP80";
172 case bitc::TYPE_CODE_FP128: return "FP128";
173 case bitc::TYPE_CODE_PPC_FP128: return "PPC_FP128";
Nick Lewycky15714342009-06-01 04:41:03 +0000174 case bitc::TYPE_CODE_METADATA: return "METADATA";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000175 }
176
177 case bitc::CONSTANTS_BLOCK_ID:
178 switch (CodeID) {
179 default: return 0;
180 case bitc::CST_CODE_SETTYPE: return "SETTYPE";
181 case bitc::CST_CODE_NULL: return "NULL";
182 case bitc::CST_CODE_UNDEF: return "UNDEF";
183 case bitc::CST_CODE_INTEGER: return "INTEGER";
184 case bitc::CST_CODE_WIDE_INTEGER: return "WIDE_INTEGER";
185 case bitc::CST_CODE_FLOAT: return "FLOAT";
186 case bitc::CST_CODE_AGGREGATE: return "AGGREGATE";
187 case bitc::CST_CODE_STRING: return "STRING";
188 case bitc::CST_CODE_CSTRING: return "CSTRING";
189 case bitc::CST_CODE_CE_BINOP: return "CE_BINOP";
190 case bitc::CST_CODE_CE_CAST: return "CE_CAST";
191 case bitc::CST_CODE_CE_GEP: return "CE_GEP";
192 case bitc::CST_CODE_CE_SELECT: return "CE_SELECT";
193 case bitc::CST_CODE_CE_EXTRACTELT: return "CE_EXTRACTELT";
194 case bitc::CST_CODE_CE_INSERTELT: return "CE_INSERTELT";
195 case bitc::CST_CODE_CE_SHUFFLEVEC: return "CE_SHUFFLEVEC";
196 case bitc::CST_CODE_CE_CMP: return "CE_CMP";
197 case bitc::CST_CODE_INLINEASM: return "INLINEASM";
Nick Lewycky15714342009-06-01 04:41:03 +0000198 case bitc::CST_CODE_CE_SHUFVEC_EX: return "CE_SHUFVEC_EX";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000199 }
200 case bitc::FUNCTION_BLOCK_ID:
201 switch (CodeID) {
202 default: return 0;
203 case bitc::FUNC_CODE_DECLAREBLOCKS: return "DECLAREBLOCKS";
204
205 case bitc::FUNC_CODE_INST_BINOP: return "INST_BINOP";
206 case bitc::FUNC_CODE_INST_CAST: return "INST_CAST";
207 case bitc::FUNC_CODE_INST_GEP: return "INST_GEP";
208 case bitc::FUNC_CODE_INST_SELECT: return "INST_SELECT";
209 case bitc::FUNC_CODE_INST_EXTRACTELT: return "INST_EXTRACTELT";
210 case bitc::FUNC_CODE_INST_INSERTELT: return "INST_INSERTELT";
211 case bitc::FUNC_CODE_INST_SHUFFLEVEC: return "INST_SHUFFLEVEC";
212 case bitc::FUNC_CODE_INST_CMP: return "INST_CMP";
213
214 case bitc::FUNC_CODE_INST_RET: return "INST_RET";
215 case bitc::FUNC_CODE_INST_BR: return "INST_BR";
216 case bitc::FUNC_CODE_INST_SWITCH: return "INST_SWITCH";
217 case bitc::FUNC_CODE_INST_INVOKE: return "INST_INVOKE";
218 case bitc::FUNC_CODE_INST_UNWIND: return "INST_UNWIND";
219 case bitc::FUNC_CODE_INST_UNREACHABLE: return "INST_UNREACHABLE";
220
221 case bitc::FUNC_CODE_INST_PHI: return "INST_PHI";
222 case bitc::FUNC_CODE_INST_MALLOC: return "INST_MALLOC";
223 case bitc::FUNC_CODE_INST_FREE: return "INST_FREE";
224 case bitc::FUNC_CODE_INST_ALLOCA: return "INST_ALLOCA";
225 case bitc::FUNC_CODE_INST_LOAD: return "INST_LOAD";
226 case bitc::FUNC_CODE_INST_STORE: return "INST_STORE";
227 case bitc::FUNC_CODE_INST_CALL: return "INST_CALL";
228 case bitc::FUNC_CODE_INST_VAARG: return "INST_VAARG";
Christopher Lamb44d62f62007-12-11 08:59:05 +0000229 case bitc::FUNC_CODE_INST_STORE2: return "INST_STORE2";
Nick Lewycky2c3eced2008-03-01 21:47:06 +0000230 case bitc::FUNC_CODE_INST_GETRESULT: return "INST_GETRESULT";
Nick Lewyckyb90b8412008-11-07 14:52:51 +0000231 case bitc::FUNC_CODE_INST_EXTRACTVAL: return "INST_EXTRACTVAL";
232 case bitc::FUNC_CODE_INST_INSERTVAL: return "INST_INSERTVAL";
233 case bitc::FUNC_CODE_INST_CMP2: return "INST_CMP2";
234 case bitc::FUNC_CODE_INST_VSELECT: return "INST_VSELECT";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000235 }
236 case bitc::TYPE_SYMTAB_BLOCK_ID:
237 switch (CodeID) {
238 default: return 0;
239 case bitc::TST_CODE_ENTRY: return "ENTRY";
240 }
241 case bitc::VALUE_SYMTAB_BLOCK_ID:
242 switch (CodeID) {
243 default: return 0;
244 case bitc::VST_CODE_ENTRY: return "ENTRY";
245 case bitc::VST_CODE_BBENTRY: return "BBENTRY";
246 }
Devang Patel0c0a6ca2009-07-22 17:43:22 +0000247 case bitc::METADATA_BLOCK_ID:
248 switch(CodeID) {
249 default:return 0;
250 case bitc::METADATA_STRING: return "MDSTRING";
Devang Patel54199ef2009-07-23 01:07:34 +0000251 case bitc::METADATA_NODE: return "MDNODE";
Devang Patel5288eae2009-07-30 23:03:19 +0000252 case bitc::METADATA_NAME: return "METADATA_NAME";
253 case bitc::METADATA_NAMED_NODE: return "NAMEDMDNODE";
Devang Patel0c0a6ca2009-07-22 17:43:22 +0000254 }
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000255 }
256}
257
Chris Lattnere4775482009-04-27 17:59:34 +0000258struct PerRecordStats {
259 unsigned NumInstances;
Chris Lattnerc40ab842009-04-27 18:15:27 +0000260 unsigned NumAbbrev;
261 uint64_t TotalBits;
262
263 PerRecordStats() : NumInstances(0), NumAbbrev(0), TotalBits(0) {}
Chris Lattnere4775482009-04-27 17:59:34 +0000264};
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000265
266struct PerBlockIDStats {
267 /// NumInstances - This the number of times this block ID has been seen.
268 unsigned NumInstances;
269
270 /// NumBits - The total size in bits of all of these blocks.
271 uint64_t NumBits;
272
273 /// NumSubBlocks - The total number of blocks these blocks contain.
274 unsigned NumSubBlocks;
275
276 /// NumAbbrevs - The total number of abbreviations.
277 unsigned NumAbbrevs;
278
279 /// NumRecords - The total number of records these blocks contain, and the
280 /// number that are abbreviated.
281 unsigned NumRecords, NumAbbreviatedRecords;
282
283 /// CodeFreq - Keep track of the number of times we see each code.
Chris Lattnere4775482009-04-27 17:59:34 +0000284 std::vector<PerRecordStats> CodeFreq;
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000285
286 PerBlockIDStats()
287 : NumInstances(0), NumBits(0),
288 NumSubBlocks(0), NumAbbrevs(0), NumRecords(0), NumAbbreviatedRecords(0) {}
289};
290
291static std::map<unsigned, PerBlockIDStats> BlockIDStats;
292
293
294
295/// Error - All bitcode analysis errors go through this function, making this a
296/// good place to breakpoint if debugging.
297static bool Error(const std::string &Err) {
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000298 errs() << Err << "\n";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000299 return true;
300}
301
302/// ParseBlock - Read a block, updating statistics, etc.
Chris Lattner25a6abf2009-04-26 20:59:02 +0000303static bool ParseBlock(BitstreamCursor &Stream, unsigned IndentLevel) {
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000304 std::string Indent(IndentLevel*2, ' ');
305 uint64_t BlockBitStart = Stream.GetCurrentBitNo();
306 unsigned BlockID = Stream.ReadSubBlockID();
307
308 // Get the statistics for this BlockID.
309 PerBlockIDStats &BlockStats = BlockIDStats[BlockID];
310
311 BlockStats.NumInstances++;
312
313 // BLOCKINFO is a special part of the stream.
314 if (BlockID == bitc::BLOCKINFO_BLOCK_ID) {
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000315 if (Dump) errs() << Indent << "<BLOCKINFO_BLOCK/>\n";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000316 if (Stream.ReadBlockInfoBlock())
317 return Error("Malformed BlockInfoBlock");
318 uint64_t BlockBitEnd = Stream.GetCurrentBitNo();
319 BlockStats.NumBits += BlockBitEnd-BlockBitStart;
320 return false;
321 }
322
323 unsigned NumWords = 0;
324 if (Stream.EnterSubBlock(BlockID, &NumWords))
325 return Error("Malformed block record");
326
327 const char *BlockName = 0;
328 if (Dump) {
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000329 errs() << Indent << "<";
Chris Lattner71101322009-04-26 22:21:57 +0000330 if ((BlockName = GetBlockName(BlockID, *Stream.getBitStreamReader())))
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000331 errs() << BlockName;
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000332 else
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000333 errs() << "UnknownBlock" << BlockID;
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000334
335 if (NonSymbolic && BlockName)
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000336 errs() << " BlockID=" << BlockID;
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000337
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000338 errs() << " NumWords=" << NumWords
339 << " BlockCodeSize=" << Stream.GetAbbrevIDWidth() << ">\n";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000340 }
341
342 SmallVector<uint64_t, 64> Record;
343
344 // Read all the records for this block.
345 while (1) {
346 if (Stream.AtEndOfStream())
347 return Error("Premature end of bitstream");
348
Chris Lattnerc40ab842009-04-27 18:15:27 +0000349 uint64_t RecordStartBit = Stream.GetCurrentBitNo();
350
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000351 // Read the code for this record.
352 unsigned AbbrevID = Stream.ReadCode();
353 switch (AbbrevID) {
354 case bitc::END_BLOCK: {
355 if (Stream.ReadBlockEnd())
356 return Error("Error at end of block");
357 uint64_t BlockBitEnd = Stream.GetCurrentBitNo();
358 BlockStats.NumBits += BlockBitEnd-BlockBitStart;
359 if (Dump) {
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000360 errs() << Indent << "</";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000361 if (BlockName)
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000362 errs() << BlockName << ">\n";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000363 else
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000364 errs() << "UnknownBlock" << BlockID << ">\n";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000365 }
366 return false;
367 }
368 case bitc::ENTER_SUBBLOCK: {
369 uint64_t SubBlockBitStart = Stream.GetCurrentBitNo();
370 if (ParseBlock(Stream, IndentLevel+1))
371 return true;
372 ++BlockStats.NumSubBlocks;
373 uint64_t SubBlockBitEnd = Stream.GetCurrentBitNo();
374
375 // Don't include subblock sizes in the size of this block.
376 BlockBitStart += SubBlockBitEnd-SubBlockBitStart;
377 break;
378 }
379 case bitc::DEFINE_ABBREV:
380 Stream.ReadAbbrevRecord();
381 ++BlockStats.NumAbbrevs;
382 break;
383 default:
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000384 Record.clear();
Chris Lattner7cbe0072009-04-06 22:44:40 +0000385
386 ++BlockStats.NumRecords;
Chris Lattnerc1894162009-04-07 02:56:46 +0000387 if (AbbrevID != bitc::UNABBREV_RECORD)
Chris Lattner7cbe0072009-04-06 22:44:40 +0000388 ++BlockStats.NumAbbreviatedRecords;
Chris Lattner7cbe0072009-04-06 22:44:40 +0000389
Chris Lattner7cbe0072009-04-06 22:44:40 +0000390 const char *BlobStart = 0;
391 unsigned BlobLen = 0;
Chris Lattnerc1894162009-04-07 02:56:46 +0000392 unsigned Code = Stream.ReadRecord(AbbrevID, Record, BlobStart, BlobLen);
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000393
Chris Lattnerc40ab842009-04-27 18:15:27 +0000394
395
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000396 // Increment the # occurrences of this code.
397 if (BlockStats.CodeFreq.size() <= Code)
398 BlockStats.CodeFreq.resize(Code+1);
Chris Lattnere4775482009-04-27 17:59:34 +0000399 BlockStats.CodeFreq[Code].NumInstances++;
Chris Lattnerc40ab842009-04-27 18:15:27 +0000400 BlockStats.CodeFreq[Code].TotalBits +=
401 Stream.GetCurrentBitNo()-RecordStartBit;
402 if (AbbrevID != bitc::UNABBREV_RECORD)
403 BlockStats.CodeFreq[Code].NumAbbrev++;
404
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000405 if (Dump) {
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000406 errs() << Indent << " <";
Chris Lattner71101322009-04-26 22:21:57 +0000407 if (const char *CodeName =
408 GetCodeName(Code, BlockID, *Stream.getBitStreamReader()))
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000409 errs() << CodeName;
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000410 else
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000411 errs() << "UnknownCode" << Code;
Chris Lattner71101322009-04-26 22:21:57 +0000412 if (NonSymbolic &&
413 GetCodeName(Code, BlockID, *Stream.getBitStreamReader()))
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000414 errs() << " codeid=" << Code;
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000415 if (AbbrevID != bitc::UNABBREV_RECORD)
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000416 errs() << " abbrevid=" << AbbrevID;
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000417
418 for (unsigned i = 0, e = Record.size(); i != e; ++i)
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000419 errs() << " op" << i << "=" << (int64_t)Record[i];
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000420
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000421 errs() << "/>";
Chris Lattnerc1894162009-04-07 02:56:46 +0000422
423 if (BlobStart) {
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000424 errs() << " blob data = ";
Chris Lattnerc1894162009-04-07 02:56:46 +0000425 bool BlobIsPrintable = true;
426 for (unsigned i = 0; i != BlobLen; ++i)
427 if (!isprint(BlobStart[i])) {
428 BlobIsPrintable = false;
429 break;
430 }
431
432 if (BlobIsPrintable)
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000433 errs() << "'" << std::string(BlobStart, BlobStart+BlobLen) <<"'";
Chris Lattnerc1894162009-04-07 02:56:46 +0000434 else
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000435 errs() << "unprintable, " << BlobLen << " bytes.";
Chris Lattnerc1894162009-04-07 02:56:46 +0000436 }
437
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000438 errs() << "\n";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000439 }
440
441 break;
442 }
443 }
444}
445
446static void PrintSize(double Bits) {
Chris Lattnere4775482009-04-27 17:59:34 +0000447 fprintf(stderr, "%.2f/%.2fB/%lluW", Bits, Bits/8,(unsigned long long)Bits/32);
448}
449static void PrintSize(uint64_t Bits) {
450 fprintf(stderr, "%llub/%.2fB/%lluW", (unsigned long long)Bits,
451 (double)Bits/8, (unsigned long long)Bits/32);
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000452}
453
454
455/// AnalyzeBitcode - Analyze the bitcode file specified by InputFilename.
456static int AnalyzeBitcode() {
457 // Read the input file.
Chris Lattnerabd0e442009-04-06 20:54:32 +0000458 MemoryBuffer *MemBuf = MemoryBuffer::getFileOrSTDIN(InputFilename.c_str());
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000459
Chris Lattnerabd0e442009-04-06 20:54:32 +0000460 if (MemBuf == 0)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000461 return Error("Error reading '" + InputFilename + "'.");
462
Chris Lattnerabd0e442009-04-06 20:54:32 +0000463 if (MemBuf->getBufferSize() & 3)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000464 return Error("Bitcode stream should be a multiple of 4 bytes in length");
465
Chris Lattnerabd0e442009-04-06 20:54:32 +0000466 unsigned char *BufPtr = (unsigned char *)MemBuf->getBufferStart();
467 unsigned char *EndBufPtr = BufPtr+MemBuf->getBufferSize();
468
469 // If we have a wrapper header, parse it and ignore the non-bc file contents.
470 // The magic number is 0x0B17C0DE stored in little endian.
471 if (isBitcodeWrapper(BufPtr, EndBufPtr))
472 if (SkipBitcodeWrapperHeader(BufPtr, EndBufPtr))
473 return Error("Invalid bitcode wrapper header");
474
Chris Lattner25a6abf2009-04-26 20:59:02 +0000475 BitstreamReader StreamFile(BufPtr, EndBufPtr);
476 BitstreamCursor Stream(StreamFile);
Chris Lattnerfe9f9c12009-04-27 20:04:08 +0000477 StreamFile.CollectBlockInfoNames();
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000478
479 // Read the stream signature.
480 char Signature[6];
481 Signature[0] = Stream.Read(8);
482 Signature[1] = Stream.Read(8);
483 Signature[2] = Stream.Read(4);
484 Signature[3] = Stream.Read(4);
485 Signature[4] = Stream.Read(4);
486 Signature[5] = Stream.Read(4);
487
488 // Autodetect the file contents, if it is one we know.
489 CurStreamType = UnknownBitstream;
490 if (Signature[0] == 'B' && Signature[1] == 'C' &&
491 Signature[2] == 0x0 && Signature[3] == 0xC &&
492 Signature[4] == 0xE && Signature[5] == 0xD)
493 CurStreamType = LLVMIRBitstream;
494
495 unsigned NumTopBlocks = 0;
496
497 // Parse the top-level structure. We only allow blocks at the top-level.
498 while (!Stream.AtEndOfStream()) {
499 unsigned Code = Stream.ReadCode();
500 if (Code != bitc::ENTER_SUBBLOCK)
501 return Error("Invalid record at top-level");
502
503 if (ParseBlock(Stream, 0))
504 return true;
505 ++NumTopBlocks;
506 }
507
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000508 if (Dump) errs() << "\n\n";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000509
Chris Lattnerabd0e442009-04-06 20:54:32 +0000510 uint64_t BufferSizeBits = (EndBufPtr-BufPtr)*CHAR_BIT;
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000511 // Print a summary of the read file.
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000512 errs() << "Summary of " << InputFilename << ":\n";
513 errs() << " Total size: ";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000514 PrintSize(BufferSizeBits);
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000515 errs() << "\n";
516 errs() << " Stream type: ";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000517 switch (CurStreamType) {
518 default: assert(0 && "Unknown bitstream type");
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000519 case UnknownBitstream: errs() << "unknown\n"; break;
520 case LLVMIRBitstream: errs() << "LLVM IR\n"; break;
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000521 }
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000522 errs() << " # Toplevel Blocks: " << NumTopBlocks << "\n";
523 errs() << "\n";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000524
525 // Emit per-block stats.
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000526 errs() << "Per-block Summary:\n";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000527 for (std::map<unsigned, PerBlockIDStats>::iterator I = BlockIDStats.begin(),
528 E = BlockIDStats.end(); I != E; ++I) {
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000529 errs() << " Block ID #" << I->first;
Chris Lattner71101322009-04-26 22:21:57 +0000530 if (const char *BlockName = GetBlockName(I->first, StreamFile))
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000531 errs() << " (" << BlockName << ")";
532 errs() << ":\n";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000533
534 const PerBlockIDStats &Stats = I->second;
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000535 errs() << " Num Instances: " << Stats.NumInstances << "\n";
536 errs() << " Total Size: ";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000537 PrintSize(Stats.NumBits);
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000538 errs() << "\n";
539 errs() << " % of file: "
540 << Stats.NumBits/(double)BufferSizeBits*100 << "\n";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000541 if (Stats.NumInstances > 1) {
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000542 errs() << " Average Size: ";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000543 PrintSize(Stats.NumBits/(double)Stats.NumInstances);
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000544 errs() << "\n";
545 errs() << " Tot/Avg SubBlocks: " << Stats.NumSubBlocks << "/"
546 << Stats.NumSubBlocks/(double)Stats.NumInstances << "\n";
547 errs() << " Tot/Avg Abbrevs: " << Stats.NumAbbrevs << "/"
548 << Stats.NumAbbrevs/(double)Stats.NumInstances << "\n";
549 errs() << " Tot/Avg Records: " << Stats.NumRecords << "/"
550 << Stats.NumRecords/(double)Stats.NumInstances << "\n";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000551 } else {
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000552 errs() << " Num SubBlocks: " << Stats.NumSubBlocks << "\n";
553 errs() << " Num Abbrevs: " << Stats.NumAbbrevs << "\n";
554 errs() << " Num Records: " << Stats.NumRecords << "\n";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000555 }
556 if (Stats.NumRecords)
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000557 errs() << " % Abbrev Recs: " << (Stats.NumAbbreviatedRecords/
558 (double)Stats.NumRecords)*100 << "\n";
559 errs() << "\n";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000560
561 // Print a histogram of the codes we see.
562 if (!NoHistogram && !Stats.CodeFreq.empty()) {
563 std::vector<std::pair<unsigned, unsigned> > FreqPairs; // <freq,code>
564 for (unsigned i = 0, e = Stats.CodeFreq.size(); i != e; ++i)
Chris Lattnere4775482009-04-27 17:59:34 +0000565 if (unsigned Freq = Stats.CodeFreq[i].NumInstances)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000566 FreqPairs.push_back(std::make_pair(Freq, i));
567 std::stable_sort(FreqPairs.begin(), FreqPairs.end());
568 std::reverse(FreqPairs.begin(), FreqPairs.end());
569
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000570 errs() << "\tRecord Histogram:\n";
Chris Lattnerc40ab842009-04-27 18:15:27 +0000571 fprintf(stderr, "\t\t Count # Bits %% Abv Record Kind\n");
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000572 for (unsigned i = 0, e = FreqPairs.size(); i != e; ++i) {
Chris Lattnerc40ab842009-04-27 18:15:27 +0000573 const PerRecordStats &RecStats = Stats.CodeFreq[FreqPairs[i].second];
574
575 fprintf(stderr, "\t\t%7d %9llu ", RecStats.NumInstances,
Dan Gohmand31c4192009-05-01 16:33:33 +0000576 (unsigned long long)RecStats.TotalBits);
Chris Lattnerc40ab842009-04-27 18:15:27 +0000577
578 if (RecStats.NumAbbrev)
579 fprintf(stderr, "%7.2f ",
580 (double)RecStats.NumAbbrev/RecStats.NumInstances*100);
581 else
582 fprintf(stderr, " ");
Chris Lattnere4775482009-04-27 17:59:34 +0000583
Chris Lattner71101322009-04-26 22:21:57 +0000584 if (const char *CodeName =
585 GetCodeName(FreqPairs[i].second, I->first, StreamFile))
Chris Lattnere4775482009-04-27 17:59:34 +0000586 fprintf(stderr, "%s\n", CodeName);
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000587 else
Chris Lattnere4775482009-04-27 17:59:34 +0000588 fprintf(stderr, "UnknownCode%d\n", FreqPairs[i].second);
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000589 }
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000590 errs() << "\n";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000591
592 }
593 }
594 return 0;
595}
596
597
598int main(int argc, char **argv) {
Chris Lattnere6012df2009-03-06 05:34:10 +0000599 // Print a stack trace if we signal out.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000600 sys::PrintStackTraceOnErrorSignal();
Chris Lattnere6012df2009-03-06 05:34:10 +0000601 PrettyStackTraceProgram X(argc, argv);
602 llvm_shutdown_obj Y; // Call llvm_shutdown() on exit.
603 cl::ParseCommandLineOptions(argc, argv, "llvm-bcanalyzer file analyzer\n");
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000604
605 return AnalyzeBitcode();
606}