blob: 30f672a81d74dad2df390dc532942411b380f025 [file] [log] [blame]
Gabor Greif8ff70c22007-07-04 21:55:50 +00001//===-- llvm-bcanalyzer.cpp - Bitcode Analyzer --------------------------===//
Misha Brukman3da94ae2005-04-22 00:00:37 +00002//
Reid Spencerdac69c82004-06-07 17:53:43 +00003// The LLVM Compiler Infrastructure
4//
Chris Lattner21c62da2007-12-29 20:44:31 +00005// This file is distributed under the University of Illinois Open Source
6// License. See LICENSE.TXT for details.
Misha Brukman3da94ae2005-04-22 00:00:37 +00007//
Reid Spencerdac69c82004-06-07 17:53:43 +00008//===----------------------------------------------------------------------===//
9//
Reid Spencer96684ef2004-06-08 05:56:58 +000010// This tool may be invoked in the following manner:
Gabor Greif8ff70c22007-07-04 21:55:50 +000011// llvm-bcanalyzer [options] - Read LLVM bitcode from stdin
12// llvm-bcanalyzer [options] x.bc - Read LLVM bitcode from the x.bc file
Reid Spencerdac69c82004-06-07 17:53:43 +000013//
Reid Spencer96684ef2004-06-08 05:56:58 +000014// Options:
Reid Spencer23d46e72004-06-10 18:38:44 +000015// --help - Output information about command line switches
Gabor Greif8ff70c22007-07-04 21:55:50 +000016// --dump - Dump low-level bitcode structure in readable format
Reid Spencer96684ef2004-06-08 05:56:58 +000017//
Gabor Greif8ff70c22007-07-04 21:55:50 +000018// 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
Reid Spencer23d46e72004-06-10 18:38:44 +000021// statistics about the contents of the file. By default this information is
Gabor Greif8ff70c22007-07-04 21:55:50 +000022// detailed and contains information about individual bitcode blocks and the
Chris Lattner44dadff2007-05-06 09:29:57 +000023// functions in the module.
Gabor Greif8ff70c22007-07-04 21:55:50 +000024// The tool is also able to print a bitcode file in a straight forward text
Misha Brukman3da94ae2005-04-22 00:00:37 +000025// format that shows the containment and relationships of the information in
Gabor Greif8ff70c22007-07-04 21:55:50 +000026// the bitcode file (-dump option).
Chris Lattner63db4852007-04-29 05:51:00 +000027//
Reid Spencerdac69c82004-06-07 17:53:43 +000028//===----------------------------------------------------------------------===//
29
Reid Spencer86a9a7a2004-06-29 23:34:27 +000030#include "llvm/Analysis/Verifier.h"
Chris Lattner45e0f892007-04-29 08:12:22 +000031#include "llvm/Bitcode/BitstreamReader.h"
Chris Lattner4238d472007-04-29 20:00:02 +000032#include "llvm/Bitcode/LLVMBitCodes.h"
Chris Lattnere2a466b2009-04-06 20:54:32 +000033#include "llvm/Bitcode/ReaderWriter.h"
Reid Spencer551ccae2004-09-01 22:55:40 +000034#include "llvm/Support/CommandLine.h"
Chris Lattnerc30598b2006-12-06 01:18:01 +000035#include "llvm/Support/ManagedStatic.h"
Chris Lattner45e0f892007-04-29 08:12:22 +000036#include "llvm/Support/MemoryBuffer.h"
Chris Lattnercc14d252009-03-06 05:34:10 +000037#include "llvm/Support/PrettyStackTrace.h"
Reid Spencerdac69c82004-06-07 17:53:43 +000038#include "llvm/System/Signals.h"
Chris Lattner44dadff2007-05-06 09:29:57 +000039#include <map>
Reid Spencerdac69c82004-06-07 17:53:43 +000040#include <fstream>
41#include <iostream>
Chris Lattnerb1e85b52007-05-05 01:46:49 +000042#include <algorithm>
Reid Spencerdac69c82004-06-07 17:53:43 +000043using namespace llvm;
44
45static cl::opt<std::string>
Gabor Greif8ff70c22007-07-04 21:55:50 +000046 InputFilename(cl::Positional, cl::desc("<input bitcode>"), cl::init("-"));
Reid Spencerdac69c82004-06-07 17:53:43 +000047
Reid Spencer75937452004-08-21 21:00:24 +000048static cl::opt<std::string>
49 OutputFilename("-o", cl::init("-"), cl::desc("<output file>"));
50
Gabor Greif8ff70c22007-07-04 21:55:50 +000051static cl::opt<bool> Dump("dump", cl::desc("Dump low level bitcode trace"));
Chris Lattner32de6332007-04-29 08:31:14 +000052
53//===----------------------------------------------------------------------===//
54// Bitcode specific analysis.
55//===----------------------------------------------------------------------===//
56
Chris Lattnerb1e85b52007-05-05 01:46:49 +000057static cl::opt<bool> NoHistogram("disable-histogram",
58 cl::desc("Do not print per-code histogram"));
Chris Lattner45e0f892007-04-29 08:12:22 +000059
Chris Lattnerb30c9252007-04-29 21:48:19 +000060static cl::opt<bool>
61NonSymbolic("non-symbolic",
62 cl::desc("Emit numberic info in dump even if"
63 " symbolic info is available"));
64
Chris Lattner45e0f892007-04-29 08:12:22 +000065/// 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
Chris Lattner4238d472007-04-29 20:00:02 +000072
73/// GetBlockName - Return a symbolic block name if known, otherwise return
Chris Lattnerb30c9252007-04-29 21:48:19 +000074/// null.
Chris Lattnerf9a3ec82009-04-26 22:21:57 +000075static const char *GetBlockName(unsigned BlockID,
76 const BitstreamReader &StreamFile) {
Chris Lattner1772b122007-05-05 00:17:42 +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 Lattnerf9a3ec82009-04-26 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
Chris Lattnerb30c9252007-04-29 21:48:19 +000092 if (CurStreamType != LLVMIRBitstream) return 0;
Chris Lattner4238d472007-04-29 20:00:02 +000093
94 switch (BlockID) {
Chris Lattnerb30c9252007-04-29 21:48:19 +000095 default: return 0;
Chris Lattner4238d472007-04-29 20:00:02 +000096 case bitc::MODULE_BLOCK_ID: return "MODULE_BLOCK";
Chris Lattnercd5b7d72007-05-04 03:01:41 +000097 case bitc::PARAMATTR_BLOCK_ID: return "PARAMATTR_BLOCK";
Chris Lattner4238d472007-04-29 20:00:02 +000098 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
Chris Lattnerb30c9252007-04-29 21:48:19 +0000106/// GetCodeName - Return a symbolic code name if known, otherwise return
107/// null.
Chris Lattnerf9a3ec82009-04-26 22:21:57 +0000108static const char *GetCodeName(unsigned CodeID, unsigned BlockID,
109 const BitstreamReader &StreamFile) {
Chris Lattner1772b122007-05-05 00:17:42 +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 Lattnerf9a3ec82009-04-26 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";
Chris Lattner1772b122007-05-05 00:17:42 +0000118 }
119 }
120 return 0;
121 }
122
Chris Lattnerf9a3ec82009-04-26 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
Chris Lattnerb30c9252007-04-29 21:48:19 +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 Lewycky82b80d92008-11-07 14:52:51 +0000149 case bitc::MODULE_CODE_GCNAME: return "GCNAME";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000150 }
Chris Lattnercd5b7d72007-05-04 03:01:41 +0000151 case bitc::PARAMATTR_BLOCK_ID:
152 switch (CodeID) {
153 default: return 0;
154 case bitc::PARAMATTR_CODE_ENTRY: return "ENTRY";
155 }
Chris Lattnerb30c9252007-04-29 21:48:19 +0000156 case bitc::TYPE_BLOCK_ID:
157 switch (CodeID) {
158 default: return 0;
Nick Lewycky82b80d92008-11-07 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";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000174 }
175
176 case bitc::CONSTANTS_BLOCK_ID:
177 switch (CodeID) {
178 default: return 0;
179 case bitc::CST_CODE_SETTYPE: return "SETTYPE";
180 case bitc::CST_CODE_NULL: return "NULL";
181 case bitc::CST_CODE_UNDEF: return "UNDEF";
182 case bitc::CST_CODE_INTEGER: return "INTEGER";
183 case bitc::CST_CODE_WIDE_INTEGER: return "WIDE_INTEGER";
184 case bitc::CST_CODE_FLOAT: return "FLOAT";
185 case bitc::CST_CODE_AGGREGATE: return "AGGREGATE";
Chris Lattnercb3d91b2007-05-06 00:53:07 +0000186 case bitc::CST_CODE_STRING: return "STRING";
187 case bitc::CST_CODE_CSTRING: return "CSTRING";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000188 case bitc::CST_CODE_CE_BINOP: return "CE_BINOP";
189 case bitc::CST_CODE_CE_CAST: return "CE_CAST";
190 case bitc::CST_CODE_CE_GEP: return "CE_GEP";
191 case bitc::CST_CODE_CE_SELECT: return "CE_SELECT";
192 case bitc::CST_CODE_CE_EXTRACTELT: return "CE_EXTRACTELT";
193 case bitc::CST_CODE_CE_INSERTELT: return "CE_INSERTELT";
194 case bitc::CST_CODE_CE_SHUFFLEVEC: return "CE_SHUFFLEVEC";
195 case bitc::CST_CODE_CE_CMP: return "CE_CMP";
Chris Lattnerc5ff2cc2007-05-06 01:50:11 +0000196 case bitc::CST_CODE_INLINEASM: return "INLINEASM";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000197 }
198 case bitc::FUNCTION_BLOCK_ID:
199 switch (CodeID) {
200 default: return 0;
201 case bitc::FUNC_CODE_DECLAREBLOCKS: return "DECLAREBLOCKS";
202
203 case bitc::FUNC_CODE_INST_BINOP: return "INST_BINOP";
204 case bitc::FUNC_CODE_INST_CAST: return "INST_CAST";
205 case bitc::FUNC_CODE_INST_GEP: return "INST_GEP";
206 case bitc::FUNC_CODE_INST_SELECT: return "INST_SELECT";
207 case bitc::FUNC_CODE_INST_EXTRACTELT: return "INST_EXTRACTELT";
208 case bitc::FUNC_CODE_INST_INSERTELT: return "INST_INSERTELT";
209 case bitc::FUNC_CODE_INST_SHUFFLEVEC: return "INST_SHUFFLEVEC";
210 case bitc::FUNC_CODE_INST_CMP: return "INST_CMP";
211
212 case bitc::FUNC_CODE_INST_RET: return "INST_RET";
213 case bitc::FUNC_CODE_INST_BR: return "INST_BR";
214 case bitc::FUNC_CODE_INST_SWITCH: return "INST_SWITCH";
215 case bitc::FUNC_CODE_INST_INVOKE: return "INST_INVOKE";
216 case bitc::FUNC_CODE_INST_UNWIND: return "INST_UNWIND";
217 case bitc::FUNC_CODE_INST_UNREACHABLE: return "INST_UNREACHABLE";
218
Chris Lattner8f926682007-05-01 02:43:46 +0000219 case bitc::FUNC_CODE_INST_PHI: return "INST_PHI";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000220 case bitc::FUNC_CODE_INST_MALLOC: return "INST_MALLOC";
221 case bitc::FUNC_CODE_INST_FREE: return "INST_FREE";
222 case bitc::FUNC_CODE_INST_ALLOCA: return "INST_ALLOCA";
223 case bitc::FUNC_CODE_INST_LOAD: return "INST_LOAD";
224 case bitc::FUNC_CODE_INST_STORE: return "INST_STORE";
225 case bitc::FUNC_CODE_INST_CALL: return "INST_CALL";
226 case bitc::FUNC_CODE_INST_VAARG: return "INST_VAARG";
Christopher Lambfe63fb92007-12-11 08:59:05 +0000227 case bitc::FUNC_CODE_INST_STORE2: return "INST_STORE2";
Nick Lewycky33a834a2008-03-01 21:47:06 +0000228 case bitc::FUNC_CODE_INST_GETRESULT: return "INST_GETRESULT";
Nick Lewycky82b80d92008-11-07 14:52:51 +0000229 case bitc::FUNC_CODE_INST_EXTRACTVAL: return "INST_EXTRACTVAL";
230 case bitc::FUNC_CODE_INST_INSERTVAL: return "INST_INSERTVAL";
231 case bitc::FUNC_CODE_INST_CMP2: return "INST_CMP2";
232 case bitc::FUNC_CODE_INST_VSELECT: return "INST_VSELECT";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000233 }
234 case bitc::TYPE_SYMTAB_BLOCK_ID:
235 switch (CodeID) {
236 default: return 0;
237 case bitc::TST_CODE_ENTRY: return "ENTRY";
238 }
239 case bitc::VALUE_SYMTAB_BLOCK_ID:
240 switch (CodeID) {
241 default: return 0;
242 case bitc::VST_CODE_ENTRY: return "ENTRY";
Chris Lattnercd5b7d72007-05-04 03:01:41 +0000243 case bitc::VST_CODE_BBENTRY: return "BBENTRY";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000244 }
245 }
246}
247
Chris Lattner24437472009-04-27 17:59:34 +0000248struct PerRecordStats {
249 unsigned NumInstances;
250};
Chris Lattner4238d472007-04-29 20:00:02 +0000251
252struct PerBlockIDStats {
253 /// NumInstances - This the number of times this block ID has been seen.
254 unsigned NumInstances;
255
256 /// NumBits - The total size in bits of all of these blocks.
257 uint64_t NumBits;
258
259 /// NumSubBlocks - The total number of blocks these blocks contain.
260 unsigned NumSubBlocks;
261
262 /// NumAbbrevs - The total number of abbreviations.
263 unsigned NumAbbrevs;
264
265 /// NumRecords - The total number of records these blocks contain, and the
266 /// number that are abbreviated.
267 unsigned NumRecords, NumAbbreviatedRecords;
268
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000269 /// CodeFreq - Keep track of the number of times we see each code.
Chris Lattner24437472009-04-27 17:59:34 +0000270 std::vector<PerRecordStats> CodeFreq;
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000271
Chris Lattner4238d472007-04-29 20:00:02 +0000272 PerBlockIDStats()
273 : NumInstances(0), NumBits(0),
274 NumSubBlocks(0), NumAbbrevs(0), NumRecords(0), NumAbbreviatedRecords(0) {}
275};
276
277static std::map<unsigned, PerBlockIDStats> BlockIDStats;
278
279
280
Chris Lattner32de6332007-04-29 08:31:14 +0000281/// Error - All bitcode analysis errors go through this function, making this a
282/// good place to breakpoint if debugging.
283static bool Error(const std::string &Err) {
284 std::cerr << Err << "\n";
285 return true;
286}
287
288/// ParseBlock - Read a block, updating statistics, etc.
Chris Lattner962dde32009-04-26 20:59:02 +0000289static bool ParseBlock(BitstreamCursor &Stream, unsigned IndentLevel) {
Chris Lattner1772b122007-05-05 00:17:42 +0000290 std::string Indent(IndentLevel*2, ' ');
Chris Lattner4238d472007-04-29 20:00:02 +0000291 uint64_t BlockBitStart = Stream.GetCurrentBitNo();
Chris Lattner32de6332007-04-29 08:31:14 +0000292 unsigned BlockID = Stream.ReadSubBlockID();
Chris Lattnerb30c9252007-04-29 21:48:19 +0000293
Chris Lattner4238d472007-04-29 20:00:02 +0000294 // Get the statistics for this BlockID.
295 PerBlockIDStats &BlockStats = BlockIDStats[BlockID];
296
297 BlockStats.NumInstances++;
Chris Lattner32de6332007-04-29 08:31:14 +0000298
Chris Lattner1772b122007-05-05 00:17:42 +0000299 // BLOCKINFO is a special part of the stream.
300 if (BlockID == bitc::BLOCKINFO_BLOCK_ID) {
301 if (Dump) std::cerr << Indent << "<BLOCKINFO_BLOCK/>\n";
302 if (Stream.ReadBlockInfoBlock())
303 return Error("Malformed BlockInfoBlock");
304 uint64_t BlockBitEnd = Stream.GetCurrentBitNo();
305 BlockStats.NumBits += BlockBitEnd-BlockBitStart;
306 return false;
307 }
308
Chris Lattnerb30c9252007-04-29 21:48:19 +0000309 unsigned NumWords = 0;
Chris Lattner1772b122007-05-05 00:17:42 +0000310 if (Stream.EnterSubBlock(BlockID, &NumWords))
Chris Lattner32de6332007-04-29 08:31:14 +0000311 return Error("Malformed block record");
312
Chris Lattnerb30c9252007-04-29 21:48:19 +0000313 const char *BlockName = 0;
314 if (Dump) {
315 std::cerr << Indent << "<";
Chris Lattnerf9a3ec82009-04-26 22:21:57 +0000316 if ((BlockName = GetBlockName(BlockID, *Stream.getBitStreamReader())))
Chris Lattnerb30c9252007-04-29 21:48:19 +0000317 std::cerr << BlockName;
318 else
319 std::cerr << "UnknownBlock" << BlockID;
320
321 if (NonSymbolic && BlockName)
322 std::cerr << " BlockID=" << BlockID;
323
324 std::cerr << " NumWords=" << NumWords
325 << " BlockCodeSize=" << Stream.GetAbbrevIDWidth() << ">\n";
326 }
327
Chris Lattner32de6332007-04-29 08:31:14 +0000328 SmallVector<uint64_t, 64> Record;
329
330 // Read all the records for this block.
331 while (1) {
332 if (Stream.AtEndOfStream())
333 return Error("Premature end of bitstream");
334
335 // Read the code for this record.
336 unsigned AbbrevID = Stream.ReadCode();
337 switch (AbbrevID) {
Chris Lattner4238d472007-04-29 20:00:02 +0000338 case bitc::END_BLOCK: {
Chris Lattner32de6332007-04-29 08:31:14 +0000339 if (Stream.ReadBlockEnd())
340 return Error("Error at end of block");
Chris Lattner4238d472007-04-29 20:00:02 +0000341 uint64_t BlockBitEnd = Stream.GetCurrentBitNo();
342 BlockStats.NumBits += BlockBitEnd-BlockBitStart;
Chris Lattnerb30c9252007-04-29 21:48:19 +0000343 if (Dump) {
344 std::cerr << Indent << "</";
345 if (BlockName)
346 std::cerr << BlockName << ">\n";
347 else
348 std::cerr << "UnknownBlock" << BlockID << ">\n";
349 }
Chris Lattner32de6332007-04-29 08:31:14 +0000350 return false;
Chris Lattner4238d472007-04-29 20:00:02 +0000351 }
Chris Lattner44b0f102007-05-05 01:29:31 +0000352 case bitc::ENTER_SUBBLOCK: {
353 uint64_t SubBlockBitStart = Stream.GetCurrentBitNo();
Chris Lattnerb30c9252007-04-29 21:48:19 +0000354 if (ParseBlock(Stream, IndentLevel+1))
Chris Lattner32de6332007-04-29 08:31:14 +0000355 return true;
Chris Lattner4238d472007-04-29 20:00:02 +0000356 ++BlockStats.NumSubBlocks;
Chris Lattner44b0f102007-05-05 01:29:31 +0000357 uint64_t SubBlockBitEnd = Stream.GetCurrentBitNo();
358
359 // Don't include subblock sizes in the size of this block.
360 BlockBitStart += SubBlockBitEnd-SubBlockBitStart;
Chris Lattner32de6332007-04-29 08:31:14 +0000361 break;
Chris Lattner44b0f102007-05-05 01:29:31 +0000362 }
Chris Lattner32de6332007-04-29 08:31:14 +0000363 case bitc::DEFINE_ABBREV:
364 Stream.ReadAbbrevRecord();
Chris Lattner4238d472007-04-29 20:00:02 +0000365 ++BlockStats.NumAbbrevs;
Chris Lattner32de6332007-04-29 08:31:14 +0000366 break;
367 default:
368 Record.clear();
Chris Lattner3f75d312009-04-06 22:44:40 +0000369
370 ++BlockStats.NumRecords;
Chris Lattnerae7dd802009-04-07 02:56:46 +0000371 if (AbbrevID != bitc::UNABBREV_RECORD)
Chris Lattner3f75d312009-04-06 22:44:40 +0000372 ++BlockStats.NumAbbreviatedRecords;
Chris Lattner3f75d312009-04-06 22:44:40 +0000373
Chris Lattner3f75d312009-04-06 22:44:40 +0000374 const char *BlobStart = 0;
375 unsigned BlobLen = 0;
Chris Lattnerae7dd802009-04-07 02:56:46 +0000376 unsigned Code = Stream.ReadRecord(AbbrevID, Record, BlobStart, BlobLen);
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000377
378 // Increment the # occurrences of this code.
379 if (BlockStats.CodeFreq.size() <= Code)
380 BlockStats.CodeFreq.resize(Code+1);
Chris Lattner24437472009-04-27 17:59:34 +0000381 BlockStats.CodeFreq[Code].NumInstances++;
Chris Lattnerb30c9252007-04-29 21:48:19 +0000382
383 if (Dump) {
384 std::cerr << Indent << " <";
Chris Lattnerf9a3ec82009-04-26 22:21:57 +0000385 if (const char *CodeName =
386 GetCodeName(Code, BlockID, *Stream.getBitStreamReader()))
Chris Lattnerb30c9252007-04-29 21:48:19 +0000387 std::cerr << CodeName;
388 else
389 std::cerr << "UnknownCode" << Code;
Chris Lattnerf9a3ec82009-04-26 22:21:57 +0000390 if (NonSymbolic &&
391 GetCodeName(Code, BlockID, *Stream.getBitStreamReader()))
Chris Lattnerb30c9252007-04-29 21:48:19 +0000392 std::cerr << " codeid=" << Code;
393 if (AbbrevID != bitc::UNABBREV_RECORD)
394 std::cerr << " abbrevid=" << AbbrevID;
395
396 for (unsigned i = 0, e = Record.size(); i != e; ++i)
397 std::cerr << " op" << i << "=" << (int64_t)Record[i];
398
Chris Lattnerae7dd802009-04-07 02:56:46 +0000399 std::cerr << "/>";
400
401 if (BlobStart) {
402 std::cerr << " blob data = ";
403 bool BlobIsPrintable = true;
404 for (unsigned i = 0; i != BlobLen; ++i)
405 if (!isprint(BlobStart[i])) {
406 BlobIsPrintable = false;
407 break;
408 }
409
410 if (BlobIsPrintable)
411 std::cerr << "'" << std::string(BlobStart, BlobStart+BlobLen) <<"'";
412 else
413 std::cerr << "unprintable, " << BlobLen << " bytes.";
414 }
415
416 std::cerr << "\n";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000417 }
418
Chris Lattner32de6332007-04-29 08:31:14 +0000419 break;
420 }
421 }
422}
423
Chris Lattner4238d472007-04-29 20:00:02 +0000424static void PrintSize(double Bits) {
Chris Lattner24437472009-04-27 17:59:34 +0000425 fprintf(stderr, "%.2f/%.2fB/%lluW", Bits, Bits/8,(unsigned long long)Bits/32);
426}
427static void PrintSize(uint64_t Bits) {
428 fprintf(stderr, "%llub/%.2fB/%lluW", (unsigned long long)Bits,
429 (double)Bits/8, (unsigned long long)Bits/32);
Chris Lattner4238d472007-04-29 20:00:02 +0000430}
431
432
Chris Lattner45e0f892007-04-29 08:12:22 +0000433/// AnalyzeBitcode - Analyze the bitcode file specified by InputFilename.
434static int AnalyzeBitcode() {
435 // Read the input file.
Chris Lattnere2a466b2009-04-06 20:54:32 +0000436 MemoryBuffer *MemBuf = MemoryBuffer::getFileOrSTDIN(InputFilename.c_str());
Chris Lattner45e0f892007-04-29 08:12:22 +0000437
Chris Lattnere2a466b2009-04-06 20:54:32 +0000438 if (MemBuf == 0)
Chris Lattner32de6332007-04-29 08:31:14 +0000439 return Error("Error reading '" + InputFilename + "'.");
Chris Lattner45e0f892007-04-29 08:12:22 +0000440
Chris Lattnere2a466b2009-04-06 20:54:32 +0000441 if (MemBuf->getBufferSize() & 3)
Chris Lattner32de6332007-04-29 08:31:14 +0000442 return Error("Bitcode stream should be a multiple of 4 bytes in length");
Chris Lattner45e0f892007-04-29 08:12:22 +0000443
Chris Lattnere2a466b2009-04-06 20:54:32 +0000444 unsigned char *BufPtr = (unsigned char *)MemBuf->getBufferStart();
445 unsigned char *EndBufPtr = BufPtr+MemBuf->getBufferSize();
446
447 // If we have a wrapper header, parse it and ignore the non-bc file contents.
448 // The magic number is 0x0B17C0DE stored in little endian.
449 if (isBitcodeWrapper(BufPtr, EndBufPtr))
450 if (SkipBitcodeWrapperHeader(BufPtr, EndBufPtr))
451 return Error("Invalid bitcode wrapper header");
452
Chris Lattner962dde32009-04-26 20:59:02 +0000453 BitstreamReader StreamFile(BufPtr, EndBufPtr);
454 BitstreamCursor Stream(StreamFile);
Chris Lattner45e0f892007-04-29 08:12:22 +0000455
456 // Read the stream signature.
457 char Signature[6];
458 Signature[0] = Stream.Read(8);
459 Signature[1] = Stream.Read(8);
460 Signature[2] = Stream.Read(4);
461 Signature[3] = Stream.Read(4);
462 Signature[4] = Stream.Read(4);
463 Signature[5] = Stream.Read(4);
464
Chris Lattner32de6332007-04-29 08:31:14 +0000465 // Autodetect the file contents, if it is one we know.
Chris Lattner45e0f892007-04-29 08:12:22 +0000466 CurStreamType = UnknownBitstream;
467 if (Signature[0] == 'B' && Signature[1] == 'C' &&
468 Signature[2] == 0x0 && Signature[3] == 0xC &&
469 Signature[4] == 0xE && Signature[5] == 0xD)
470 CurStreamType = LLVMIRBitstream;
471
Chris Lattner4238d472007-04-29 20:00:02 +0000472 unsigned NumTopBlocks = 0;
473
Chris Lattner32de6332007-04-29 08:31:14 +0000474 // Parse the top-level structure. We only allow blocks at the top-level.
475 while (!Stream.AtEndOfStream()) {
476 unsigned Code = Stream.ReadCode();
477 if (Code != bitc::ENTER_SUBBLOCK)
478 return Error("Invalid record at top-level");
479
Chris Lattnerb30c9252007-04-29 21:48:19 +0000480 if (ParseBlock(Stream, 0))
Chris Lattner32de6332007-04-29 08:31:14 +0000481 return true;
Chris Lattner4238d472007-04-29 20:00:02 +0000482 ++NumTopBlocks;
Chris Lattner32de6332007-04-29 08:31:14 +0000483 }
484
Chris Lattnerb30c9252007-04-29 21:48:19 +0000485 if (Dump) std::cerr << "\n\n";
486
Chris Lattnere2a466b2009-04-06 20:54:32 +0000487 uint64_t BufferSizeBits = (EndBufPtr-BufPtr)*CHAR_BIT;
Chris Lattner32de6332007-04-29 08:31:14 +0000488 // Print a summary of the read file.
Chris Lattner45e0f892007-04-29 08:12:22 +0000489 std::cerr << "Summary of " << InputFilename << ":\n";
Chris Lattner4238d472007-04-29 20:00:02 +0000490 std::cerr << " Total size: ";
Chris Lattner8f926682007-05-01 02:43:46 +0000491 PrintSize(BufferSizeBits);
Chris Lattner4238d472007-04-29 20:00:02 +0000492 std::cerr << "\n";
493 std::cerr << " Stream type: ";
Chris Lattner45e0f892007-04-29 08:12:22 +0000494 switch (CurStreamType) {
495 default: assert(0 && "Unknown bitstream type");
496 case UnknownBitstream: std::cerr << "unknown\n"; break;
497 case LLVMIRBitstream: std::cerr << "LLVM IR\n"; break;
498 }
Chris Lattner4238d472007-04-29 20:00:02 +0000499 std::cerr << " # Toplevel Blocks: " << NumTopBlocks << "\n";
500 std::cerr << "\n";
501
502 // Emit per-block stats.
503 std::cerr << "Per-block Summary:\n";
504 for (std::map<unsigned, PerBlockIDStats>::iterator I = BlockIDStats.begin(),
505 E = BlockIDStats.end(); I != E; ++I) {
506 std::cerr << " Block ID #" << I->first;
Chris Lattnerf9a3ec82009-04-26 22:21:57 +0000507 if (const char *BlockName = GetBlockName(I->first, StreamFile))
Chris Lattner4238d472007-04-29 20:00:02 +0000508 std::cerr << " (" << BlockName << ")";
509 std::cerr << ":\n";
510
511 const PerBlockIDStats &Stats = I->second;
512 std::cerr << " Num Instances: " << Stats.NumInstances << "\n";
513 std::cerr << " Total Size: ";
514 PrintSize(Stats.NumBits);
515 std::cerr << "\n";
Chris Lattner8f926682007-05-01 02:43:46 +0000516 std::cerr << " % of file: "
517 << Stats.NumBits/(double)BufferSizeBits*100 << "\n";
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000518 if (Stats.NumInstances > 1) {
519 std::cerr << " Average Size: ";
520 PrintSize(Stats.NumBits/(double)Stats.NumInstances);
521 std::cerr << "\n";
522 std::cerr << " Tot/Avg SubBlocks: " << Stats.NumSubBlocks << "/"
523 << Stats.NumSubBlocks/(double)Stats.NumInstances << "\n";
524 std::cerr << " Tot/Avg Abbrevs: " << Stats.NumAbbrevs << "/"
525 << Stats.NumAbbrevs/(double)Stats.NumInstances << "\n";
526 std::cerr << " Tot/Avg Records: " << Stats.NumRecords << "/"
527 << Stats.NumRecords/(double)Stats.NumInstances << "\n";
528 } else {
529 std::cerr << " Num SubBlocks: " << Stats.NumSubBlocks << "\n";
530 std::cerr << " Num Abbrevs: " << Stats.NumAbbrevs << "\n";
531 std::cerr << " Num Records: " << Stats.NumRecords << "\n";
532 }
Chris Lattner1772b122007-05-05 00:17:42 +0000533 if (Stats.NumRecords)
534 std::cerr << " % Abbrev Recs: " << (Stats.NumAbbreviatedRecords/
535 (double)Stats.NumRecords)*100 << "\n";
Chris Lattner4238d472007-04-29 20:00:02 +0000536 std::cerr << "\n";
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000537
538 // Print a histogram of the codes we see.
539 if (!NoHistogram && !Stats.CodeFreq.empty()) {
540 std::vector<std::pair<unsigned, unsigned> > FreqPairs; // <freq,code>
541 for (unsigned i = 0, e = Stats.CodeFreq.size(); i != e; ++i)
Chris Lattner24437472009-04-27 17:59:34 +0000542 if (unsigned Freq = Stats.CodeFreq[i].NumInstances)
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000543 FreqPairs.push_back(std::make_pair(Freq, i));
544 std::stable_sort(FreqPairs.begin(), FreqPairs.end());
545 std::reverse(FreqPairs.begin(), FreqPairs.end());
546
Chris Lattner24437472009-04-27 17:59:34 +0000547 std::cerr << "\tRecord Histogram:\n";
548 fprintf(stderr, "\t\t Count Record Kind\n");
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000549 for (unsigned i = 0, e = FreqPairs.size(); i != e; ++i) {
Chris Lattner24437472009-04-27 17:59:34 +0000550 fprintf(stderr, "\t\t%7d ", FreqPairs[i].first);
551
Chris Lattnerf9a3ec82009-04-26 22:21:57 +0000552 if (const char *CodeName =
553 GetCodeName(FreqPairs[i].second, I->first, StreamFile))
Chris Lattner24437472009-04-27 17:59:34 +0000554 fprintf(stderr, "%s\n", CodeName);
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000555 else
Chris Lattner24437472009-04-27 17:59:34 +0000556 fprintf(stderr, "UnknownCode%d\n", FreqPairs[i].second);
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000557 }
558 std::cerr << "\n";
559
560 }
Chris Lattner4238d472007-04-29 20:00:02 +0000561 }
Chris Lattner45e0f892007-04-29 08:12:22 +0000562 return 0;
563}
Reid Spencerdac69c82004-06-07 17:53:43 +0000564
Chris Lattner32de6332007-04-29 08:31:14 +0000565
Chris Lattnerc30598b2006-12-06 01:18:01 +0000566int main(int argc, char **argv) {
Chris Lattnercc14d252009-03-06 05:34:10 +0000567 // Print a stack trace if we signal out.
Chris Lattner45e0f892007-04-29 08:12:22 +0000568 sys::PrintStackTraceOnErrorSignal();
Chris Lattnercc14d252009-03-06 05:34:10 +0000569 PrettyStackTraceProgram X(argc, argv);
570 llvm_shutdown_obj Y; // Call llvm_shutdown() on exit.
571 cl::ParseCommandLineOptions(argc, argv, "llvm-bcanalyzer file analyzer\n");
Chris Lattner45e0f892007-04-29 08:12:22 +0000572
Chris Lattner44dadff2007-05-06 09:29:57 +0000573 return AnalyzeBitcode();
Reid Spencerdac69c82004-06-07 17:53:43 +0000574}