blob: fcde8da60995d79d5e7d0dd92e98dc18b288d3a4 [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"
Dan Gohmanf17a25c2007-07-18 16:29:46 +000038#include "llvm/System/Signals.h"
39#include <map>
40#include <fstream>
41#include <iostream>
42#include <algorithm>
43using namespace llvm;
44
45static cl::opt<std::string>
46 InputFilename(cl::Positional, cl::desc("<input bitcode>"), cl::init("-"));
47
48static cl::opt<std::string>
49 OutputFilename("-o", cl::init("-"), cl::desc("<output file>"));
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 numberic info in dump even if"
63 " symbolic info is available"));
64
65/// CurStreamType - If we can sniff the flavor of this stream, we can produce
66/// better dump info.
67static enum {
68 UnknownBitstream,
69 LLVMIRBitstream
70} CurStreamType;
71
72
73/// GetBlockName - Return a symbolic block name if known, otherwise return
74/// null.
Chris Lattner71101322009-04-26 22:21:57 +000075static const char *GetBlockName(unsigned BlockID,
76 const BitstreamReader &StreamFile) {
Dan Gohmanf17a25c2007-07-18 16:29:46 +000077 // Standard blocks for all bitcode files.
78 if (BlockID < bitc::FIRST_APPLICATION_BLOCKID) {
79 if (BlockID == bitc::BLOCKINFO_BLOCK_ID)
80 return "BLOCKINFO_BLOCK";
81 return 0;
82 }
83
Chris Lattner71101322009-04-26 22:21:57 +000084 // Check to see if we have a blockinfo record for this block, with a name.
85 if (const BitstreamReader::BlockInfo *Info =
86 StreamFile.getBlockInfo(BlockID)) {
87 if (!Info->Name.empty())
88 return Info->Name.c_str();
89 }
90
91
Dan Gohmanf17a25c2007-07-18 16:29:46 +000092 if (CurStreamType != LLVMIRBitstream) return 0;
93
94 switch (BlockID) {
95 default: return 0;
96 case bitc::MODULE_BLOCK_ID: return "MODULE_BLOCK";
97 case bitc::PARAMATTR_BLOCK_ID: return "PARAMATTR_BLOCK";
98 case bitc::TYPE_BLOCK_ID: return "TYPE_BLOCK";
99 case bitc::CONSTANTS_BLOCK_ID: return "CONSTANTS_BLOCK";
100 case bitc::FUNCTION_BLOCK_ID: return "FUNCTION_BLOCK";
101 case bitc::TYPE_SYMTAB_BLOCK_ID: return "TYPE_SYMTAB";
102 case bitc::VALUE_SYMTAB_BLOCK_ID: return "VALUE_SYMTAB";
103 }
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";
199 case bitc::CST_CODE_MDSTRING: return "MDSTRING";
200 case bitc::CST_CODE_MDNODE: return "MDNODE";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000201 }
202 case bitc::FUNCTION_BLOCK_ID:
203 switch (CodeID) {
204 default: return 0;
205 case bitc::FUNC_CODE_DECLAREBLOCKS: return "DECLAREBLOCKS";
206
207 case bitc::FUNC_CODE_INST_BINOP: return "INST_BINOP";
208 case bitc::FUNC_CODE_INST_CAST: return "INST_CAST";
209 case bitc::FUNC_CODE_INST_GEP: return "INST_GEP";
210 case bitc::FUNC_CODE_INST_SELECT: return "INST_SELECT";
211 case bitc::FUNC_CODE_INST_EXTRACTELT: return "INST_EXTRACTELT";
212 case bitc::FUNC_CODE_INST_INSERTELT: return "INST_INSERTELT";
213 case bitc::FUNC_CODE_INST_SHUFFLEVEC: return "INST_SHUFFLEVEC";
214 case bitc::FUNC_CODE_INST_CMP: return "INST_CMP";
215
216 case bitc::FUNC_CODE_INST_RET: return "INST_RET";
217 case bitc::FUNC_CODE_INST_BR: return "INST_BR";
218 case bitc::FUNC_CODE_INST_SWITCH: return "INST_SWITCH";
219 case bitc::FUNC_CODE_INST_INVOKE: return "INST_INVOKE";
220 case bitc::FUNC_CODE_INST_UNWIND: return "INST_UNWIND";
221 case bitc::FUNC_CODE_INST_UNREACHABLE: return "INST_UNREACHABLE";
222
223 case bitc::FUNC_CODE_INST_PHI: return "INST_PHI";
224 case bitc::FUNC_CODE_INST_MALLOC: return "INST_MALLOC";
225 case bitc::FUNC_CODE_INST_FREE: return "INST_FREE";
226 case bitc::FUNC_CODE_INST_ALLOCA: return "INST_ALLOCA";
227 case bitc::FUNC_CODE_INST_LOAD: return "INST_LOAD";
228 case bitc::FUNC_CODE_INST_STORE: return "INST_STORE";
229 case bitc::FUNC_CODE_INST_CALL: return "INST_CALL";
230 case bitc::FUNC_CODE_INST_VAARG: return "INST_VAARG";
Christopher Lamb44d62f62007-12-11 08:59:05 +0000231 case bitc::FUNC_CODE_INST_STORE2: return "INST_STORE2";
Nick Lewycky2c3eced2008-03-01 21:47:06 +0000232 case bitc::FUNC_CODE_INST_GETRESULT: return "INST_GETRESULT";
Nick Lewyckyb90b8412008-11-07 14:52:51 +0000233 case bitc::FUNC_CODE_INST_EXTRACTVAL: return "INST_EXTRACTVAL";
234 case bitc::FUNC_CODE_INST_INSERTVAL: return "INST_INSERTVAL";
235 case bitc::FUNC_CODE_INST_CMP2: return "INST_CMP2";
236 case bitc::FUNC_CODE_INST_VSELECT: return "INST_VSELECT";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000237 }
238 case bitc::TYPE_SYMTAB_BLOCK_ID:
239 switch (CodeID) {
240 default: return 0;
241 case bitc::TST_CODE_ENTRY: return "ENTRY";
242 }
243 case bitc::VALUE_SYMTAB_BLOCK_ID:
244 switch (CodeID) {
245 default: return 0;
246 case bitc::VST_CODE_ENTRY: return "ENTRY";
247 case bitc::VST_CODE_BBENTRY: return "BBENTRY";
248 }
249 }
250}
251
Chris Lattnere4775482009-04-27 17:59:34 +0000252struct PerRecordStats {
253 unsigned NumInstances;
Chris Lattnerc40ab842009-04-27 18:15:27 +0000254 unsigned NumAbbrev;
255 uint64_t TotalBits;
256
257 PerRecordStats() : NumInstances(0), NumAbbrev(0), TotalBits(0) {}
Chris Lattnere4775482009-04-27 17:59:34 +0000258};
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000259
260struct PerBlockIDStats {
261 /// NumInstances - This the number of times this block ID has been seen.
262 unsigned NumInstances;
263
264 /// NumBits - The total size in bits of all of these blocks.
265 uint64_t NumBits;
266
267 /// NumSubBlocks - The total number of blocks these blocks contain.
268 unsigned NumSubBlocks;
269
270 /// NumAbbrevs - The total number of abbreviations.
271 unsigned NumAbbrevs;
272
273 /// NumRecords - The total number of records these blocks contain, and the
274 /// number that are abbreviated.
275 unsigned NumRecords, NumAbbreviatedRecords;
276
277 /// CodeFreq - Keep track of the number of times we see each code.
Chris Lattnere4775482009-04-27 17:59:34 +0000278 std::vector<PerRecordStats> CodeFreq;
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000279
280 PerBlockIDStats()
281 : NumInstances(0), NumBits(0),
282 NumSubBlocks(0), NumAbbrevs(0), NumRecords(0), NumAbbreviatedRecords(0) {}
283};
284
285static std::map<unsigned, PerBlockIDStats> BlockIDStats;
286
287
288
289/// Error - All bitcode analysis errors go through this function, making this a
290/// good place to breakpoint if debugging.
291static bool Error(const std::string &Err) {
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000292 errs() << Err << "\n";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000293 return true;
294}
295
296/// ParseBlock - Read a block, updating statistics, etc.
Chris Lattner25a6abf2009-04-26 20:59:02 +0000297static bool ParseBlock(BitstreamCursor &Stream, unsigned IndentLevel) {
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000298 std::string Indent(IndentLevel*2, ' ');
299 uint64_t BlockBitStart = Stream.GetCurrentBitNo();
300 unsigned BlockID = Stream.ReadSubBlockID();
301
302 // Get the statistics for this BlockID.
303 PerBlockIDStats &BlockStats = BlockIDStats[BlockID];
304
305 BlockStats.NumInstances++;
306
307 // BLOCKINFO is a special part of the stream.
308 if (BlockID == bitc::BLOCKINFO_BLOCK_ID) {
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000309 if (Dump) errs() << Indent << "<BLOCKINFO_BLOCK/>\n";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000310 if (Stream.ReadBlockInfoBlock())
311 return Error("Malformed BlockInfoBlock");
312 uint64_t BlockBitEnd = Stream.GetCurrentBitNo();
313 BlockStats.NumBits += BlockBitEnd-BlockBitStart;
314 return false;
315 }
316
317 unsigned NumWords = 0;
318 if (Stream.EnterSubBlock(BlockID, &NumWords))
319 return Error("Malformed block record");
320
321 const char *BlockName = 0;
322 if (Dump) {
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000323 errs() << Indent << "<";
Chris Lattner71101322009-04-26 22:21:57 +0000324 if ((BlockName = GetBlockName(BlockID, *Stream.getBitStreamReader())))
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000325 errs() << BlockName;
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000326 else
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000327 errs() << "UnknownBlock" << BlockID;
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000328
329 if (NonSymbolic && BlockName)
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000330 errs() << " BlockID=" << BlockID;
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000331
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000332 errs() << " NumWords=" << NumWords
333 << " BlockCodeSize=" << Stream.GetAbbrevIDWidth() << ">\n";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000334 }
335
336 SmallVector<uint64_t, 64> Record;
337
338 // Read all the records for this block.
339 while (1) {
340 if (Stream.AtEndOfStream())
341 return Error("Premature end of bitstream");
342
Chris Lattnerc40ab842009-04-27 18:15:27 +0000343 uint64_t RecordStartBit = Stream.GetCurrentBitNo();
344
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000345 // Read the code for this record.
346 unsigned AbbrevID = Stream.ReadCode();
347 switch (AbbrevID) {
348 case bitc::END_BLOCK: {
349 if (Stream.ReadBlockEnd())
350 return Error("Error at end of block");
351 uint64_t BlockBitEnd = Stream.GetCurrentBitNo();
352 BlockStats.NumBits += BlockBitEnd-BlockBitStart;
353 if (Dump) {
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000354 errs() << Indent << "</";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000355 if (BlockName)
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000356 errs() << BlockName << ">\n";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000357 else
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000358 errs() << "UnknownBlock" << BlockID << ">\n";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000359 }
360 return false;
361 }
362 case bitc::ENTER_SUBBLOCK: {
363 uint64_t SubBlockBitStart = Stream.GetCurrentBitNo();
364 if (ParseBlock(Stream, IndentLevel+1))
365 return true;
366 ++BlockStats.NumSubBlocks;
367 uint64_t SubBlockBitEnd = Stream.GetCurrentBitNo();
368
369 // Don't include subblock sizes in the size of this block.
370 BlockBitStart += SubBlockBitEnd-SubBlockBitStart;
371 break;
372 }
373 case bitc::DEFINE_ABBREV:
374 Stream.ReadAbbrevRecord();
375 ++BlockStats.NumAbbrevs;
376 break;
377 default:
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000378 Record.clear();
Chris Lattner7cbe0072009-04-06 22:44:40 +0000379
380 ++BlockStats.NumRecords;
Chris Lattnerc1894162009-04-07 02:56:46 +0000381 if (AbbrevID != bitc::UNABBREV_RECORD)
Chris Lattner7cbe0072009-04-06 22:44:40 +0000382 ++BlockStats.NumAbbreviatedRecords;
Chris Lattner7cbe0072009-04-06 22:44:40 +0000383
Chris Lattner7cbe0072009-04-06 22:44:40 +0000384 const char *BlobStart = 0;
385 unsigned BlobLen = 0;
Chris Lattnerc1894162009-04-07 02:56:46 +0000386 unsigned Code = Stream.ReadRecord(AbbrevID, Record, BlobStart, BlobLen);
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000387
Chris Lattnerc40ab842009-04-27 18:15:27 +0000388
389
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000390 // Increment the # occurrences of this code.
391 if (BlockStats.CodeFreq.size() <= Code)
392 BlockStats.CodeFreq.resize(Code+1);
Chris Lattnere4775482009-04-27 17:59:34 +0000393 BlockStats.CodeFreq[Code].NumInstances++;
Chris Lattnerc40ab842009-04-27 18:15:27 +0000394 BlockStats.CodeFreq[Code].TotalBits +=
395 Stream.GetCurrentBitNo()-RecordStartBit;
396 if (AbbrevID != bitc::UNABBREV_RECORD)
397 BlockStats.CodeFreq[Code].NumAbbrev++;
398
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000399 if (Dump) {
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000400 errs() << Indent << " <";
Chris Lattner71101322009-04-26 22:21:57 +0000401 if (const char *CodeName =
402 GetCodeName(Code, BlockID, *Stream.getBitStreamReader()))
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000403 errs() << CodeName;
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000404 else
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000405 errs() << "UnknownCode" << Code;
Chris Lattner71101322009-04-26 22:21:57 +0000406 if (NonSymbolic &&
407 GetCodeName(Code, BlockID, *Stream.getBitStreamReader()))
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000408 errs() << " codeid=" << Code;
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000409 if (AbbrevID != bitc::UNABBREV_RECORD)
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000410 errs() << " abbrevid=" << AbbrevID;
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000411
412 for (unsigned i = 0, e = Record.size(); i != e; ++i)
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000413 errs() << " op" << i << "=" << (int64_t)Record[i];
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000414
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000415 errs() << "/>";
Chris Lattnerc1894162009-04-07 02:56:46 +0000416
417 if (BlobStart) {
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000418 errs() << " blob data = ";
Chris Lattnerc1894162009-04-07 02:56:46 +0000419 bool BlobIsPrintable = true;
420 for (unsigned i = 0; i != BlobLen; ++i)
421 if (!isprint(BlobStart[i])) {
422 BlobIsPrintable = false;
423 break;
424 }
425
426 if (BlobIsPrintable)
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000427 errs() << "'" << std::string(BlobStart, BlobStart+BlobLen) <<"'";
Chris Lattnerc1894162009-04-07 02:56:46 +0000428 else
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000429 errs() << "unprintable, " << BlobLen << " bytes.";
Chris Lattnerc1894162009-04-07 02:56:46 +0000430 }
431
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000432 errs() << "\n";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000433 }
434
435 break;
436 }
437 }
438}
439
440static void PrintSize(double Bits) {
Chris Lattnere4775482009-04-27 17:59:34 +0000441 fprintf(stderr, "%.2f/%.2fB/%lluW", Bits, Bits/8,(unsigned long long)Bits/32);
442}
443static void PrintSize(uint64_t Bits) {
444 fprintf(stderr, "%llub/%.2fB/%lluW", (unsigned long long)Bits,
445 (double)Bits/8, (unsigned long long)Bits/32);
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000446}
447
448
449/// AnalyzeBitcode - Analyze the bitcode file specified by InputFilename.
450static int AnalyzeBitcode() {
451 // Read the input file.
Chris Lattnerabd0e442009-04-06 20:54:32 +0000452 MemoryBuffer *MemBuf = MemoryBuffer::getFileOrSTDIN(InputFilename.c_str());
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000453
Chris Lattnerabd0e442009-04-06 20:54:32 +0000454 if (MemBuf == 0)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000455 return Error("Error reading '" + InputFilename + "'.");
456
Chris Lattnerabd0e442009-04-06 20:54:32 +0000457 if (MemBuf->getBufferSize() & 3)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000458 return Error("Bitcode stream should be a multiple of 4 bytes in length");
459
Chris Lattnerabd0e442009-04-06 20:54:32 +0000460 unsigned char *BufPtr = (unsigned char *)MemBuf->getBufferStart();
461 unsigned char *EndBufPtr = BufPtr+MemBuf->getBufferSize();
462
463 // If we have a wrapper header, parse it and ignore the non-bc file contents.
464 // The magic number is 0x0B17C0DE stored in little endian.
465 if (isBitcodeWrapper(BufPtr, EndBufPtr))
466 if (SkipBitcodeWrapperHeader(BufPtr, EndBufPtr))
467 return Error("Invalid bitcode wrapper header");
468
Chris Lattner25a6abf2009-04-26 20:59:02 +0000469 BitstreamReader StreamFile(BufPtr, EndBufPtr);
470 BitstreamCursor Stream(StreamFile);
Chris Lattnerfe9f9c12009-04-27 20:04:08 +0000471 StreamFile.CollectBlockInfoNames();
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000472
473 // Read the stream signature.
474 char Signature[6];
475 Signature[0] = Stream.Read(8);
476 Signature[1] = Stream.Read(8);
477 Signature[2] = Stream.Read(4);
478 Signature[3] = Stream.Read(4);
479 Signature[4] = Stream.Read(4);
480 Signature[5] = Stream.Read(4);
481
482 // Autodetect the file contents, if it is one we know.
483 CurStreamType = UnknownBitstream;
484 if (Signature[0] == 'B' && Signature[1] == 'C' &&
485 Signature[2] == 0x0 && Signature[3] == 0xC &&
486 Signature[4] == 0xE && Signature[5] == 0xD)
487 CurStreamType = LLVMIRBitstream;
488
489 unsigned NumTopBlocks = 0;
490
491 // Parse the top-level structure. We only allow blocks at the top-level.
492 while (!Stream.AtEndOfStream()) {
493 unsigned Code = Stream.ReadCode();
494 if (Code != bitc::ENTER_SUBBLOCK)
495 return Error("Invalid record at top-level");
496
497 if (ParseBlock(Stream, 0))
498 return true;
499 ++NumTopBlocks;
500 }
501
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000502 if (Dump) errs() << "\n\n";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000503
Chris Lattnerabd0e442009-04-06 20:54:32 +0000504 uint64_t BufferSizeBits = (EndBufPtr-BufPtr)*CHAR_BIT;
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000505 // Print a summary of the read file.
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000506 errs() << "Summary of " << InputFilename << ":\n";
507 errs() << " Total size: ";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000508 PrintSize(BufferSizeBits);
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000509 errs() << "\n";
510 errs() << " Stream type: ";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000511 switch (CurStreamType) {
512 default: assert(0 && "Unknown bitstream type");
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000513 case UnknownBitstream: errs() << "unknown\n"; break;
514 case LLVMIRBitstream: errs() << "LLVM IR\n"; break;
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000515 }
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000516 errs() << " # Toplevel Blocks: " << NumTopBlocks << "\n";
517 errs() << "\n";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000518
519 // Emit per-block stats.
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000520 errs() << "Per-block Summary:\n";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000521 for (std::map<unsigned, PerBlockIDStats>::iterator I = BlockIDStats.begin(),
522 E = BlockIDStats.end(); I != E; ++I) {
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000523 errs() << " Block ID #" << I->first;
Chris Lattner71101322009-04-26 22:21:57 +0000524 if (const char *BlockName = GetBlockName(I->first, StreamFile))
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000525 errs() << " (" << BlockName << ")";
526 errs() << ":\n";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000527
528 const PerBlockIDStats &Stats = I->second;
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000529 errs() << " Num Instances: " << Stats.NumInstances << "\n";
530 errs() << " Total Size: ";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000531 PrintSize(Stats.NumBits);
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000532 errs() << "\n";
533 errs() << " % of file: "
534 << Stats.NumBits/(double)BufferSizeBits*100 << "\n";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000535 if (Stats.NumInstances > 1) {
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000536 errs() << " Average Size: ";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000537 PrintSize(Stats.NumBits/(double)Stats.NumInstances);
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000538 errs() << "\n";
539 errs() << " Tot/Avg SubBlocks: " << Stats.NumSubBlocks << "/"
540 << Stats.NumSubBlocks/(double)Stats.NumInstances << "\n";
541 errs() << " Tot/Avg Abbrevs: " << Stats.NumAbbrevs << "/"
542 << Stats.NumAbbrevs/(double)Stats.NumInstances << "\n";
543 errs() << " Tot/Avg Records: " << Stats.NumRecords << "/"
544 << Stats.NumRecords/(double)Stats.NumInstances << "\n";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000545 } else {
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000546 errs() << " Num SubBlocks: " << Stats.NumSubBlocks << "\n";
547 errs() << " Num Abbrevs: " << Stats.NumAbbrevs << "\n";
548 errs() << " Num Records: " << Stats.NumRecords << "\n";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000549 }
550 if (Stats.NumRecords)
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000551 errs() << " % Abbrev Recs: " << (Stats.NumAbbreviatedRecords/
552 (double)Stats.NumRecords)*100 << "\n";
553 errs() << "\n";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000554
555 // Print a histogram of the codes we see.
556 if (!NoHistogram && !Stats.CodeFreq.empty()) {
557 std::vector<std::pair<unsigned, unsigned> > FreqPairs; // <freq,code>
558 for (unsigned i = 0, e = Stats.CodeFreq.size(); i != e; ++i)
Chris Lattnere4775482009-04-27 17:59:34 +0000559 if (unsigned Freq = Stats.CodeFreq[i].NumInstances)
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000560 FreqPairs.push_back(std::make_pair(Freq, i));
561 std::stable_sort(FreqPairs.begin(), FreqPairs.end());
562 std::reverse(FreqPairs.begin(), FreqPairs.end());
563
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000564 errs() << "\tRecord Histogram:\n";
Chris Lattnerc40ab842009-04-27 18:15:27 +0000565 fprintf(stderr, "\t\t Count # Bits %% Abv Record Kind\n");
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000566 for (unsigned i = 0, e = FreqPairs.size(); i != e; ++i) {
Chris Lattnerc40ab842009-04-27 18:15:27 +0000567 const PerRecordStats &RecStats = Stats.CodeFreq[FreqPairs[i].second];
568
569 fprintf(stderr, "\t\t%7d %9llu ", RecStats.NumInstances,
Dan Gohmand31c4192009-05-01 16:33:33 +0000570 (unsigned long long)RecStats.TotalBits);
Chris Lattnerc40ab842009-04-27 18:15:27 +0000571
572 if (RecStats.NumAbbrev)
573 fprintf(stderr, "%7.2f ",
574 (double)RecStats.NumAbbrev/RecStats.NumInstances*100);
575 else
576 fprintf(stderr, " ");
Chris Lattnere4775482009-04-27 17:59:34 +0000577
Chris Lattner71101322009-04-26 22:21:57 +0000578 if (const char *CodeName =
579 GetCodeName(FreqPairs[i].second, I->first, StreamFile))
Chris Lattnere4775482009-04-27 17:59:34 +0000580 fprintf(stderr, "%s\n", CodeName);
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000581 else
Chris Lattnere4775482009-04-27 17:59:34 +0000582 fprintf(stderr, "UnknownCode%d\n", FreqPairs[i].second);
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000583 }
Dan Gohmanf8b81bf2009-07-15 16:35:29 +0000584 errs() << "\n";
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000585
586 }
587 }
588 return 0;
589}
590
591
592int main(int argc, char **argv) {
Chris Lattnere6012df2009-03-06 05:34:10 +0000593 // Print a stack trace if we signal out.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000594 sys::PrintStackTraceOnErrorSignal();
Chris Lattnere6012df2009-03-06 05:34:10 +0000595 PrettyStackTraceProgram X(argc, argv);
596 llvm_shutdown_obj Y; // Call llvm_shutdown() on exit.
597 cl::ParseCommandLineOptions(argc, argv, "llvm-bcanalyzer file analyzer\n");
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000598
599 return AnalyzeBitcode();
600}