blob: 4eaa60c7570468db4cfe5e785435e93d21606aa8 [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>
Chris Lattnerb1e85b52007-05-05 01:46:49 +000040#include <algorithm>
Reid Spencerdac69c82004-06-07 17:53:43 +000041using namespace llvm;
42
43static cl::opt<std::string>
Gabor Greif8ff70c22007-07-04 21:55:50 +000044 InputFilename(cl::Positional, cl::desc("<input bitcode>"), cl::init("-"));
Reid Spencerdac69c82004-06-07 17:53:43 +000045
Reid Spencer75937452004-08-21 21:00:24 +000046static cl::opt<std::string>
47 OutputFilename("-o", cl::init("-"), cl::desc("<output file>"));
48
Gabor Greif8ff70c22007-07-04 21:55:50 +000049static cl::opt<bool> Dump("dump", cl::desc("Dump low level bitcode trace"));
Chris Lattner32de6332007-04-29 08:31:14 +000050
51//===----------------------------------------------------------------------===//
52// Bitcode specific analysis.
53//===----------------------------------------------------------------------===//
54
Chris Lattnerb1e85b52007-05-05 01:46:49 +000055static cl::opt<bool> NoHistogram("disable-histogram",
56 cl::desc("Do not print per-code histogram"));
Chris Lattner45e0f892007-04-29 08:12:22 +000057
Chris Lattnerb30c9252007-04-29 21:48:19 +000058static cl::opt<bool>
59NonSymbolic("non-symbolic",
60 cl::desc("Emit numberic info in dump even if"
61 " symbolic info is available"));
62
Chris Lattner45e0f892007-04-29 08:12:22 +000063/// CurStreamType - If we can sniff the flavor of this stream, we can produce
64/// better dump info.
65static enum {
66 UnknownBitstream,
67 LLVMIRBitstream
68} CurStreamType;
69
Chris Lattner4238d472007-04-29 20:00:02 +000070
71/// GetBlockName - Return a symbolic block name if known, otherwise return
Chris Lattnerb30c9252007-04-29 21:48:19 +000072/// null.
Chris Lattnerf9a3ec82009-04-26 22:21:57 +000073static const char *GetBlockName(unsigned BlockID,
74 const BitstreamReader &StreamFile) {
Chris Lattner1772b122007-05-05 00:17:42 +000075 // Standard blocks for all bitcode files.
76 if (BlockID < bitc::FIRST_APPLICATION_BLOCKID) {
77 if (BlockID == bitc::BLOCKINFO_BLOCK_ID)
78 return "BLOCKINFO_BLOCK";
79 return 0;
80 }
81
Chris Lattnerf9a3ec82009-04-26 22:21:57 +000082 // Check to see if we have a blockinfo record for this block, with a name.
83 if (const BitstreamReader::BlockInfo *Info =
84 StreamFile.getBlockInfo(BlockID)) {
85 if (!Info->Name.empty())
86 return Info->Name.c_str();
87 }
88
89
Chris Lattnerb30c9252007-04-29 21:48:19 +000090 if (CurStreamType != LLVMIRBitstream) return 0;
Chris Lattner4238d472007-04-29 20:00:02 +000091
92 switch (BlockID) {
Chris Lattnerb30c9252007-04-29 21:48:19 +000093 default: return 0;
Chris Lattner4238d472007-04-29 20:00:02 +000094 case bitc::MODULE_BLOCK_ID: return "MODULE_BLOCK";
Chris Lattnercd5b7d72007-05-04 03:01:41 +000095 case bitc::PARAMATTR_BLOCK_ID: return "PARAMATTR_BLOCK";
Chris Lattner4238d472007-04-29 20:00:02 +000096 case bitc::TYPE_BLOCK_ID: return "TYPE_BLOCK";
97 case bitc::CONSTANTS_BLOCK_ID: return "CONSTANTS_BLOCK";
98 case bitc::FUNCTION_BLOCK_ID: return "FUNCTION_BLOCK";
99 case bitc::TYPE_SYMTAB_BLOCK_ID: return "TYPE_SYMTAB";
100 case bitc::VALUE_SYMTAB_BLOCK_ID: return "VALUE_SYMTAB";
Devang Patele54abc92009-07-22 17:43:22 +0000101 case bitc::METADATA_BLOCK_ID: return "METADATA_BLOCK";
Chris Lattner4238d472007-04-29 20:00:02 +0000102 }
103}
104
Chris Lattnerb30c9252007-04-29 21:48:19 +0000105/// GetCodeName - Return a symbolic code name if known, otherwise return
106/// null.
Chris Lattnerf9a3ec82009-04-26 22:21:57 +0000107static const char *GetCodeName(unsigned CodeID, unsigned BlockID,
108 const BitstreamReader &StreamFile) {
Chris Lattner1772b122007-05-05 00:17:42 +0000109 // Standard blocks for all bitcode files.
110 if (BlockID < bitc::FIRST_APPLICATION_BLOCKID) {
111 if (BlockID == bitc::BLOCKINFO_BLOCK_ID) {
112 switch (CodeID) {
113 default: return 0;
Chris Lattnerf9a3ec82009-04-26 22:21:57 +0000114 case bitc::BLOCKINFO_CODE_SETBID: return "SETBID";
115 case bitc::BLOCKINFO_CODE_BLOCKNAME: return "BLOCKNAME";
116 case bitc::BLOCKINFO_CODE_SETRECORDNAME: return "SETRECORDNAME";
Chris Lattner1772b122007-05-05 00:17:42 +0000117 }
118 }
119 return 0;
120 }
121
Chris Lattnerf9a3ec82009-04-26 22:21:57 +0000122 // Check to see if we have a blockinfo record for this record, with a name.
123 if (const BitstreamReader::BlockInfo *Info =
124 StreamFile.getBlockInfo(BlockID)) {
125 for (unsigned i = 0, e = Info->RecordNames.size(); i != e; ++i)
126 if (Info->RecordNames[i].first == CodeID)
127 return Info->RecordNames[i].second.c_str();
128 }
129
130
Chris Lattnerb30c9252007-04-29 21:48:19 +0000131 if (CurStreamType != LLVMIRBitstream) return 0;
132
133 switch (BlockID) {
134 default: return 0;
135 case bitc::MODULE_BLOCK_ID:
136 switch (CodeID) {
137 default: return 0;
138 case bitc::MODULE_CODE_VERSION: return "VERSION";
139 case bitc::MODULE_CODE_TRIPLE: return "TRIPLE";
140 case bitc::MODULE_CODE_DATALAYOUT: return "DATALAYOUT";
141 case bitc::MODULE_CODE_ASM: return "ASM";
142 case bitc::MODULE_CODE_SECTIONNAME: return "SECTIONNAME";
143 case bitc::MODULE_CODE_DEPLIB: return "DEPLIB";
144 case bitc::MODULE_CODE_GLOBALVAR: return "GLOBALVAR";
145 case bitc::MODULE_CODE_FUNCTION: return "FUNCTION";
146 case bitc::MODULE_CODE_ALIAS: return "ALIAS";
147 case bitc::MODULE_CODE_PURGEVALS: return "PURGEVALS";
Nick Lewycky82b80d92008-11-07 14:52:51 +0000148 case bitc::MODULE_CODE_GCNAME: return "GCNAME";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000149 }
Chris Lattnercd5b7d72007-05-04 03:01:41 +0000150 case bitc::PARAMATTR_BLOCK_ID:
151 switch (CodeID) {
152 default: return 0;
153 case bitc::PARAMATTR_CODE_ENTRY: return "ENTRY";
154 }
Chris Lattnerb30c9252007-04-29 21:48:19 +0000155 case bitc::TYPE_BLOCK_ID:
156 switch (CodeID) {
157 default: return 0;
Nick Lewycky82b80d92008-11-07 14:52:51 +0000158 case bitc::TYPE_CODE_NUMENTRY: return "NUMENTRY";
159 case bitc::TYPE_CODE_VOID: return "VOID";
160 case bitc::TYPE_CODE_FLOAT: return "FLOAT";
161 case bitc::TYPE_CODE_DOUBLE: return "DOUBLE";
162 case bitc::TYPE_CODE_LABEL: return "LABEL";
163 case bitc::TYPE_CODE_OPAQUE: return "OPAQUE";
164 case bitc::TYPE_CODE_INTEGER: return "INTEGER";
165 case bitc::TYPE_CODE_POINTER: return "POINTER";
166 case bitc::TYPE_CODE_FUNCTION: return "FUNCTION";
167 case bitc::TYPE_CODE_STRUCT: return "STRUCT";
168 case bitc::TYPE_CODE_ARRAY: return "ARRAY";
169 case bitc::TYPE_CODE_VECTOR: return "VECTOR";
170 case bitc::TYPE_CODE_X86_FP80: return "X86_FP80";
171 case bitc::TYPE_CODE_FP128: return "FP128";
172 case bitc::TYPE_CODE_PPC_FP128: return "PPC_FP128";
Nick Lewycky079c0342009-06-01 04:41:03 +0000173 case bitc::TYPE_CODE_METADATA: return "METADATA";
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";
Nick Lewycky079c0342009-06-01 04:41:03 +0000197 case bitc::CST_CODE_CE_SHUFVEC_EX: return "CE_SHUFVEC_EX";
Nick Lewycky079c0342009-06-01 04:41:03 +0000198 case bitc::CST_CODE_MDNODE: return "MDNODE";
Chris Lattnerb30c9252007-04-29 21:48:19 +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
Chris Lattner8f926682007-05-01 02:43:46 +0000221 case bitc::FUNC_CODE_INST_PHI: return "INST_PHI";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000222 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 Lambfe63fb92007-12-11 08:59:05 +0000229 case bitc::FUNC_CODE_INST_STORE2: return "INST_STORE2";
Nick Lewycky33a834a2008-03-01 21:47:06 +0000230 case bitc::FUNC_CODE_INST_GETRESULT: return "INST_GETRESULT";
Nick Lewycky82b80d92008-11-07 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";
Chris Lattnerb30c9252007-04-29 21:48:19 +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";
Chris Lattnercd5b7d72007-05-04 03:01:41 +0000245 case bitc::VST_CODE_BBENTRY: return "BBENTRY";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000246 }
Devang Patele54abc92009-07-22 17:43:22 +0000247 case bitc::METADATA_BLOCK_ID:
248 switch(CodeID) {
249 default:return 0;
250 case bitc::METADATA_STRING: return "MDSTRING";
251 }
Chris Lattnerb30c9252007-04-29 21:48:19 +0000252 }
253}
254
Chris Lattner24437472009-04-27 17:59:34 +0000255struct PerRecordStats {
256 unsigned NumInstances;
Chris Lattnerc167cac2009-04-27 18:15:27 +0000257 unsigned NumAbbrev;
258 uint64_t TotalBits;
259
260 PerRecordStats() : NumInstances(0), NumAbbrev(0), TotalBits(0) {}
Chris Lattner24437472009-04-27 17:59:34 +0000261};
Chris Lattner4238d472007-04-29 20:00:02 +0000262
263struct PerBlockIDStats {
264 /// NumInstances - This the number of times this block ID has been seen.
265 unsigned NumInstances;
266
267 /// NumBits - The total size in bits of all of these blocks.
268 uint64_t NumBits;
269
270 /// NumSubBlocks - The total number of blocks these blocks contain.
271 unsigned NumSubBlocks;
272
273 /// NumAbbrevs - The total number of abbreviations.
274 unsigned NumAbbrevs;
275
276 /// NumRecords - The total number of records these blocks contain, and the
277 /// number that are abbreviated.
278 unsigned NumRecords, NumAbbreviatedRecords;
279
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000280 /// CodeFreq - Keep track of the number of times we see each code.
Chris Lattner24437472009-04-27 17:59:34 +0000281 std::vector<PerRecordStats> CodeFreq;
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000282
Chris Lattner4238d472007-04-29 20:00:02 +0000283 PerBlockIDStats()
284 : NumInstances(0), NumBits(0),
285 NumSubBlocks(0), NumAbbrevs(0), NumRecords(0), NumAbbreviatedRecords(0) {}
286};
287
288static std::map<unsigned, PerBlockIDStats> BlockIDStats;
289
290
291
Chris Lattner32de6332007-04-29 08:31:14 +0000292/// Error - All bitcode analysis errors go through this function, making this a
293/// good place to breakpoint if debugging.
294static bool Error(const std::string &Err) {
Dan Gohman65f57c22009-07-15 16:35:29 +0000295 errs() << Err << "\n";
Chris Lattner32de6332007-04-29 08:31:14 +0000296 return true;
297}
298
299/// ParseBlock - Read a block, updating statistics, etc.
Chris Lattner962dde32009-04-26 20:59:02 +0000300static bool ParseBlock(BitstreamCursor &Stream, unsigned IndentLevel) {
Chris Lattner1772b122007-05-05 00:17:42 +0000301 std::string Indent(IndentLevel*2, ' ');
Chris Lattner4238d472007-04-29 20:00:02 +0000302 uint64_t BlockBitStart = Stream.GetCurrentBitNo();
Chris Lattner32de6332007-04-29 08:31:14 +0000303 unsigned BlockID = Stream.ReadSubBlockID();
Chris Lattnerb30c9252007-04-29 21:48:19 +0000304
Chris Lattner4238d472007-04-29 20:00:02 +0000305 // Get the statistics for this BlockID.
306 PerBlockIDStats &BlockStats = BlockIDStats[BlockID];
307
308 BlockStats.NumInstances++;
Chris Lattner32de6332007-04-29 08:31:14 +0000309
Chris Lattner1772b122007-05-05 00:17:42 +0000310 // BLOCKINFO is a special part of the stream.
311 if (BlockID == bitc::BLOCKINFO_BLOCK_ID) {
Dan Gohman65f57c22009-07-15 16:35:29 +0000312 if (Dump) errs() << Indent << "<BLOCKINFO_BLOCK/>\n";
Chris Lattner1772b122007-05-05 00:17:42 +0000313 if (Stream.ReadBlockInfoBlock())
314 return Error("Malformed BlockInfoBlock");
315 uint64_t BlockBitEnd = Stream.GetCurrentBitNo();
316 BlockStats.NumBits += BlockBitEnd-BlockBitStart;
317 return false;
318 }
319
Chris Lattnerb30c9252007-04-29 21:48:19 +0000320 unsigned NumWords = 0;
Chris Lattner1772b122007-05-05 00:17:42 +0000321 if (Stream.EnterSubBlock(BlockID, &NumWords))
Chris Lattner32de6332007-04-29 08:31:14 +0000322 return Error("Malformed block record");
323
Chris Lattnerb30c9252007-04-29 21:48:19 +0000324 const char *BlockName = 0;
325 if (Dump) {
Dan Gohman65f57c22009-07-15 16:35:29 +0000326 errs() << Indent << "<";
Chris Lattnerf9a3ec82009-04-26 22:21:57 +0000327 if ((BlockName = GetBlockName(BlockID, *Stream.getBitStreamReader())))
Dan Gohman65f57c22009-07-15 16:35:29 +0000328 errs() << BlockName;
Chris Lattnerb30c9252007-04-29 21:48:19 +0000329 else
Dan Gohman65f57c22009-07-15 16:35:29 +0000330 errs() << "UnknownBlock" << BlockID;
Chris Lattnerb30c9252007-04-29 21:48:19 +0000331
332 if (NonSymbolic && BlockName)
Dan Gohman65f57c22009-07-15 16:35:29 +0000333 errs() << " BlockID=" << BlockID;
Chris Lattnerb30c9252007-04-29 21:48:19 +0000334
Dan Gohman65f57c22009-07-15 16:35:29 +0000335 errs() << " NumWords=" << NumWords
336 << " BlockCodeSize=" << Stream.GetAbbrevIDWidth() << ">\n";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000337 }
338
Chris Lattner32de6332007-04-29 08:31:14 +0000339 SmallVector<uint64_t, 64> Record;
340
341 // Read all the records for this block.
342 while (1) {
343 if (Stream.AtEndOfStream())
344 return Error("Premature end of bitstream");
345
Chris Lattnerc167cac2009-04-27 18:15:27 +0000346 uint64_t RecordStartBit = Stream.GetCurrentBitNo();
347
Chris Lattner32de6332007-04-29 08:31:14 +0000348 // Read the code for this record.
349 unsigned AbbrevID = Stream.ReadCode();
350 switch (AbbrevID) {
Chris Lattner4238d472007-04-29 20:00:02 +0000351 case bitc::END_BLOCK: {
Chris Lattner32de6332007-04-29 08:31:14 +0000352 if (Stream.ReadBlockEnd())
353 return Error("Error at end of block");
Chris Lattner4238d472007-04-29 20:00:02 +0000354 uint64_t BlockBitEnd = Stream.GetCurrentBitNo();
355 BlockStats.NumBits += BlockBitEnd-BlockBitStart;
Chris Lattnerb30c9252007-04-29 21:48:19 +0000356 if (Dump) {
Dan Gohman65f57c22009-07-15 16:35:29 +0000357 errs() << Indent << "</";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000358 if (BlockName)
Dan Gohman65f57c22009-07-15 16:35:29 +0000359 errs() << BlockName << ">\n";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000360 else
Dan Gohman65f57c22009-07-15 16:35:29 +0000361 errs() << "UnknownBlock" << BlockID << ">\n";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000362 }
Chris Lattner32de6332007-04-29 08:31:14 +0000363 return false;
Chris Lattner4238d472007-04-29 20:00:02 +0000364 }
Chris Lattner44b0f102007-05-05 01:29:31 +0000365 case bitc::ENTER_SUBBLOCK: {
366 uint64_t SubBlockBitStart = Stream.GetCurrentBitNo();
Chris Lattnerb30c9252007-04-29 21:48:19 +0000367 if (ParseBlock(Stream, IndentLevel+1))
Chris Lattner32de6332007-04-29 08:31:14 +0000368 return true;
Chris Lattner4238d472007-04-29 20:00:02 +0000369 ++BlockStats.NumSubBlocks;
Chris Lattner44b0f102007-05-05 01:29:31 +0000370 uint64_t SubBlockBitEnd = Stream.GetCurrentBitNo();
371
372 // Don't include subblock sizes in the size of this block.
373 BlockBitStart += SubBlockBitEnd-SubBlockBitStart;
Chris Lattner32de6332007-04-29 08:31:14 +0000374 break;
Chris Lattner44b0f102007-05-05 01:29:31 +0000375 }
Chris Lattner32de6332007-04-29 08:31:14 +0000376 case bitc::DEFINE_ABBREV:
377 Stream.ReadAbbrevRecord();
Chris Lattner4238d472007-04-29 20:00:02 +0000378 ++BlockStats.NumAbbrevs;
Chris Lattner32de6332007-04-29 08:31:14 +0000379 break;
380 default:
381 Record.clear();
Chris Lattner3f75d312009-04-06 22:44:40 +0000382
383 ++BlockStats.NumRecords;
Chris Lattnerae7dd802009-04-07 02:56:46 +0000384 if (AbbrevID != bitc::UNABBREV_RECORD)
Chris Lattner3f75d312009-04-06 22:44:40 +0000385 ++BlockStats.NumAbbreviatedRecords;
Chris Lattner3f75d312009-04-06 22:44:40 +0000386
Chris Lattner3f75d312009-04-06 22:44:40 +0000387 const char *BlobStart = 0;
388 unsigned BlobLen = 0;
Chris Lattnerae7dd802009-04-07 02:56:46 +0000389 unsigned Code = Stream.ReadRecord(AbbrevID, Record, BlobStart, BlobLen);
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000390
Chris Lattnerc167cac2009-04-27 18:15:27 +0000391
392
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000393 // Increment the # occurrences of this code.
394 if (BlockStats.CodeFreq.size() <= Code)
395 BlockStats.CodeFreq.resize(Code+1);
Chris Lattner24437472009-04-27 17:59:34 +0000396 BlockStats.CodeFreq[Code].NumInstances++;
Chris Lattnerc167cac2009-04-27 18:15:27 +0000397 BlockStats.CodeFreq[Code].TotalBits +=
398 Stream.GetCurrentBitNo()-RecordStartBit;
399 if (AbbrevID != bitc::UNABBREV_RECORD)
400 BlockStats.CodeFreq[Code].NumAbbrev++;
401
Chris Lattnerb30c9252007-04-29 21:48:19 +0000402 if (Dump) {
Dan Gohman65f57c22009-07-15 16:35:29 +0000403 errs() << Indent << " <";
Chris Lattnerf9a3ec82009-04-26 22:21:57 +0000404 if (const char *CodeName =
405 GetCodeName(Code, BlockID, *Stream.getBitStreamReader()))
Dan Gohman65f57c22009-07-15 16:35:29 +0000406 errs() << CodeName;
Chris Lattnerb30c9252007-04-29 21:48:19 +0000407 else
Dan Gohman65f57c22009-07-15 16:35:29 +0000408 errs() << "UnknownCode" << Code;
Chris Lattnerf9a3ec82009-04-26 22:21:57 +0000409 if (NonSymbolic &&
410 GetCodeName(Code, BlockID, *Stream.getBitStreamReader()))
Dan Gohman65f57c22009-07-15 16:35:29 +0000411 errs() << " codeid=" << Code;
Chris Lattnerb30c9252007-04-29 21:48:19 +0000412 if (AbbrevID != bitc::UNABBREV_RECORD)
Dan Gohman65f57c22009-07-15 16:35:29 +0000413 errs() << " abbrevid=" << AbbrevID;
Chris Lattnerb30c9252007-04-29 21:48:19 +0000414
415 for (unsigned i = 0, e = Record.size(); i != e; ++i)
Dan Gohman65f57c22009-07-15 16:35:29 +0000416 errs() << " op" << i << "=" << (int64_t)Record[i];
Chris Lattnerb30c9252007-04-29 21:48:19 +0000417
Dan Gohman65f57c22009-07-15 16:35:29 +0000418 errs() << "/>";
Chris Lattnerae7dd802009-04-07 02:56:46 +0000419
420 if (BlobStart) {
Dan Gohman65f57c22009-07-15 16:35:29 +0000421 errs() << " blob data = ";
Chris Lattnerae7dd802009-04-07 02:56:46 +0000422 bool BlobIsPrintable = true;
423 for (unsigned i = 0; i != BlobLen; ++i)
424 if (!isprint(BlobStart[i])) {
425 BlobIsPrintable = false;
426 break;
427 }
428
429 if (BlobIsPrintable)
Dan Gohman65f57c22009-07-15 16:35:29 +0000430 errs() << "'" << std::string(BlobStart, BlobStart+BlobLen) <<"'";
Chris Lattnerae7dd802009-04-07 02:56:46 +0000431 else
Dan Gohman65f57c22009-07-15 16:35:29 +0000432 errs() << "unprintable, " << BlobLen << " bytes.";
Chris Lattnerae7dd802009-04-07 02:56:46 +0000433 }
434
Dan Gohman65f57c22009-07-15 16:35:29 +0000435 errs() << "\n";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000436 }
437
Chris Lattner32de6332007-04-29 08:31:14 +0000438 break;
439 }
440 }
441}
442
Chris Lattner4238d472007-04-29 20:00:02 +0000443static void PrintSize(double Bits) {
Chris Lattner24437472009-04-27 17:59:34 +0000444 fprintf(stderr, "%.2f/%.2fB/%lluW", Bits, Bits/8,(unsigned long long)Bits/32);
445}
446static void PrintSize(uint64_t Bits) {
447 fprintf(stderr, "%llub/%.2fB/%lluW", (unsigned long long)Bits,
448 (double)Bits/8, (unsigned long long)Bits/32);
Chris Lattner4238d472007-04-29 20:00:02 +0000449}
450
451
Chris Lattner45e0f892007-04-29 08:12:22 +0000452/// AnalyzeBitcode - Analyze the bitcode file specified by InputFilename.
453static int AnalyzeBitcode() {
454 // Read the input file.
Chris Lattnere2a466b2009-04-06 20:54:32 +0000455 MemoryBuffer *MemBuf = MemoryBuffer::getFileOrSTDIN(InputFilename.c_str());
Chris Lattner45e0f892007-04-29 08:12:22 +0000456
Chris Lattnere2a466b2009-04-06 20:54:32 +0000457 if (MemBuf == 0)
Chris Lattner32de6332007-04-29 08:31:14 +0000458 return Error("Error reading '" + InputFilename + "'.");
Chris Lattner45e0f892007-04-29 08:12:22 +0000459
Chris Lattnere2a466b2009-04-06 20:54:32 +0000460 if (MemBuf->getBufferSize() & 3)
Chris Lattner32de6332007-04-29 08:31:14 +0000461 return Error("Bitcode stream should be a multiple of 4 bytes in length");
Chris Lattner45e0f892007-04-29 08:12:22 +0000462
Chris Lattnere2a466b2009-04-06 20:54:32 +0000463 unsigned char *BufPtr = (unsigned char *)MemBuf->getBufferStart();
464 unsigned char *EndBufPtr = BufPtr+MemBuf->getBufferSize();
465
466 // If we have a wrapper header, parse it and ignore the non-bc file contents.
467 // The magic number is 0x0B17C0DE stored in little endian.
468 if (isBitcodeWrapper(BufPtr, EndBufPtr))
469 if (SkipBitcodeWrapperHeader(BufPtr, EndBufPtr))
470 return Error("Invalid bitcode wrapper header");
471
Chris Lattner962dde32009-04-26 20:59:02 +0000472 BitstreamReader StreamFile(BufPtr, EndBufPtr);
473 BitstreamCursor Stream(StreamFile);
Chris Lattner0370cc62009-04-27 20:04:08 +0000474 StreamFile.CollectBlockInfoNames();
Chris Lattner45e0f892007-04-29 08:12:22 +0000475
476 // Read the stream signature.
477 char Signature[6];
478 Signature[0] = Stream.Read(8);
479 Signature[1] = Stream.Read(8);
480 Signature[2] = Stream.Read(4);
481 Signature[3] = Stream.Read(4);
482 Signature[4] = Stream.Read(4);
483 Signature[5] = Stream.Read(4);
484
Chris Lattner32de6332007-04-29 08:31:14 +0000485 // Autodetect the file contents, if it is one we know.
Chris Lattner45e0f892007-04-29 08:12:22 +0000486 CurStreamType = UnknownBitstream;
487 if (Signature[0] == 'B' && Signature[1] == 'C' &&
488 Signature[2] == 0x0 && Signature[3] == 0xC &&
489 Signature[4] == 0xE && Signature[5] == 0xD)
490 CurStreamType = LLVMIRBitstream;
491
Chris Lattner4238d472007-04-29 20:00:02 +0000492 unsigned NumTopBlocks = 0;
493
Chris Lattner32de6332007-04-29 08:31:14 +0000494 // Parse the top-level structure. We only allow blocks at the top-level.
495 while (!Stream.AtEndOfStream()) {
496 unsigned Code = Stream.ReadCode();
497 if (Code != bitc::ENTER_SUBBLOCK)
498 return Error("Invalid record at top-level");
499
Chris Lattnerb30c9252007-04-29 21:48:19 +0000500 if (ParseBlock(Stream, 0))
Chris Lattner32de6332007-04-29 08:31:14 +0000501 return true;
Chris Lattner4238d472007-04-29 20:00:02 +0000502 ++NumTopBlocks;
Chris Lattner32de6332007-04-29 08:31:14 +0000503 }
504
Dan Gohman65f57c22009-07-15 16:35:29 +0000505 if (Dump) errs() << "\n\n";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000506
Chris Lattnere2a466b2009-04-06 20:54:32 +0000507 uint64_t BufferSizeBits = (EndBufPtr-BufPtr)*CHAR_BIT;
Chris Lattner32de6332007-04-29 08:31:14 +0000508 // Print a summary of the read file.
Dan Gohman65f57c22009-07-15 16:35:29 +0000509 errs() << "Summary of " << InputFilename << ":\n";
510 errs() << " Total size: ";
Chris Lattner8f926682007-05-01 02:43:46 +0000511 PrintSize(BufferSizeBits);
Dan Gohman65f57c22009-07-15 16:35:29 +0000512 errs() << "\n";
513 errs() << " Stream type: ";
Chris Lattner45e0f892007-04-29 08:12:22 +0000514 switch (CurStreamType) {
515 default: assert(0 && "Unknown bitstream type");
Dan Gohman65f57c22009-07-15 16:35:29 +0000516 case UnknownBitstream: errs() << "unknown\n"; break;
517 case LLVMIRBitstream: errs() << "LLVM IR\n"; break;
Chris Lattner45e0f892007-04-29 08:12:22 +0000518 }
Dan Gohman65f57c22009-07-15 16:35:29 +0000519 errs() << " # Toplevel Blocks: " << NumTopBlocks << "\n";
520 errs() << "\n";
Chris Lattner4238d472007-04-29 20:00:02 +0000521
522 // Emit per-block stats.
Dan Gohman65f57c22009-07-15 16:35:29 +0000523 errs() << "Per-block Summary:\n";
Chris Lattner4238d472007-04-29 20:00:02 +0000524 for (std::map<unsigned, PerBlockIDStats>::iterator I = BlockIDStats.begin(),
525 E = BlockIDStats.end(); I != E; ++I) {
Dan Gohman65f57c22009-07-15 16:35:29 +0000526 errs() << " Block ID #" << I->first;
Chris Lattnerf9a3ec82009-04-26 22:21:57 +0000527 if (const char *BlockName = GetBlockName(I->first, StreamFile))
Dan Gohman65f57c22009-07-15 16:35:29 +0000528 errs() << " (" << BlockName << ")";
529 errs() << ":\n";
Chris Lattner4238d472007-04-29 20:00:02 +0000530
531 const PerBlockIDStats &Stats = I->second;
Dan Gohman65f57c22009-07-15 16:35:29 +0000532 errs() << " Num Instances: " << Stats.NumInstances << "\n";
533 errs() << " Total Size: ";
Chris Lattner4238d472007-04-29 20:00:02 +0000534 PrintSize(Stats.NumBits);
Dan Gohman65f57c22009-07-15 16:35:29 +0000535 errs() << "\n";
536 errs() << " % of file: "
537 << Stats.NumBits/(double)BufferSizeBits*100 << "\n";
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000538 if (Stats.NumInstances > 1) {
Dan Gohman65f57c22009-07-15 16:35:29 +0000539 errs() << " Average Size: ";
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000540 PrintSize(Stats.NumBits/(double)Stats.NumInstances);
Dan Gohman65f57c22009-07-15 16:35:29 +0000541 errs() << "\n";
542 errs() << " Tot/Avg SubBlocks: " << Stats.NumSubBlocks << "/"
543 << Stats.NumSubBlocks/(double)Stats.NumInstances << "\n";
544 errs() << " Tot/Avg Abbrevs: " << Stats.NumAbbrevs << "/"
545 << Stats.NumAbbrevs/(double)Stats.NumInstances << "\n";
546 errs() << " Tot/Avg Records: " << Stats.NumRecords << "/"
547 << Stats.NumRecords/(double)Stats.NumInstances << "\n";
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000548 } else {
Dan Gohman65f57c22009-07-15 16:35:29 +0000549 errs() << " Num SubBlocks: " << Stats.NumSubBlocks << "\n";
550 errs() << " Num Abbrevs: " << Stats.NumAbbrevs << "\n";
551 errs() << " Num Records: " << Stats.NumRecords << "\n";
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000552 }
Chris Lattner1772b122007-05-05 00:17:42 +0000553 if (Stats.NumRecords)
Dan Gohman65f57c22009-07-15 16:35:29 +0000554 errs() << " % Abbrev Recs: " << (Stats.NumAbbreviatedRecords/
555 (double)Stats.NumRecords)*100 << "\n";
556 errs() << "\n";
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000557
558 // Print a histogram of the codes we see.
559 if (!NoHistogram && !Stats.CodeFreq.empty()) {
560 std::vector<std::pair<unsigned, unsigned> > FreqPairs; // <freq,code>
561 for (unsigned i = 0, e = Stats.CodeFreq.size(); i != e; ++i)
Chris Lattner24437472009-04-27 17:59:34 +0000562 if (unsigned Freq = Stats.CodeFreq[i].NumInstances)
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000563 FreqPairs.push_back(std::make_pair(Freq, i));
564 std::stable_sort(FreqPairs.begin(), FreqPairs.end());
565 std::reverse(FreqPairs.begin(), FreqPairs.end());
566
Dan Gohman65f57c22009-07-15 16:35:29 +0000567 errs() << "\tRecord Histogram:\n";
Chris Lattnerc167cac2009-04-27 18:15:27 +0000568 fprintf(stderr, "\t\t Count # Bits %% Abv Record Kind\n");
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000569 for (unsigned i = 0, e = FreqPairs.size(); i != e; ++i) {
Chris Lattnerc167cac2009-04-27 18:15:27 +0000570 const PerRecordStats &RecStats = Stats.CodeFreq[FreqPairs[i].second];
571
572 fprintf(stderr, "\t\t%7d %9llu ", RecStats.NumInstances,
Dan Gohmanfbccdef2009-05-01 16:33:33 +0000573 (unsigned long long)RecStats.TotalBits);
Chris Lattnerc167cac2009-04-27 18:15:27 +0000574
575 if (RecStats.NumAbbrev)
576 fprintf(stderr, "%7.2f ",
577 (double)RecStats.NumAbbrev/RecStats.NumInstances*100);
578 else
579 fprintf(stderr, " ");
Chris Lattner24437472009-04-27 17:59:34 +0000580
Chris Lattnerf9a3ec82009-04-26 22:21:57 +0000581 if (const char *CodeName =
582 GetCodeName(FreqPairs[i].second, I->first, StreamFile))
Chris Lattner24437472009-04-27 17:59:34 +0000583 fprintf(stderr, "%s\n", CodeName);
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000584 else
Chris Lattner24437472009-04-27 17:59:34 +0000585 fprintf(stderr, "UnknownCode%d\n", FreqPairs[i].second);
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000586 }
Dan Gohman65f57c22009-07-15 16:35:29 +0000587 errs() << "\n";
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000588
589 }
Chris Lattner4238d472007-04-29 20:00:02 +0000590 }
Chris Lattner45e0f892007-04-29 08:12:22 +0000591 return 0;
592}
Reid Spencerdac69c82004-06-07 17:53:43 +0000593
Chris Lattner32de6332007-04-29 08:31:14 +0000594
Chris Lattnerc30598b2006-12-06 01:18:01 +0000595int main(int argc, char **argv) {
Chris Lattnercc14d252009-03-06 05:34:10 +0000596 // Print a stack trace if we signal out.
Chris Lattner45e0f892007-04-29 08:12:22 +0000597 sys::PrintStackTraceOnErrorSignal();
Chris Lattnercc14d252009-03-06 05:34:10 +0000598 PrettyStackTraceProgram X(argc, argv);
599 llvm_shutdown_obj Y; // Call llvm_shutdown() on exit.
600 cl::ParseCommandLineOptions(argc, argv, "llvm-bcanalyzer file analyzer\n");
Chris Lattner45e0f892007-04-29 08:12:22 +0000601
Chris Lattner44dadff2007-05-06 09:29:57 +0000602 return AnalyzeBitcode();
Reid Spencerdac69c82004-06-07 17:53:43 +0000603}