blob: b89691082e73e1f1dacde7614afd927facb8fe0a [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"
Chris Lattner74382b72009-08-23 22:45:37 +000038#include "llvm/Support/raw_ostream.h"
Reid Spencerdac69c82004-06-07 17:53:43 +000039#include "llvm/System/Signals.h"
Duncan Sands15af79c2009-08-24 10:59:01 +000040#include <cstdio>
Chris Lattner44dadff2007-05-06 09:29:57 +000041#include <map>
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) {
Devang Patele8e02132009-09-18 19:26:43 +000095 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 case bitc::METADATA_BLOCK_ID: return "METADATA_BLOCK";
104 case bitc::METADATA_ATTACHMENT_ID: return "METADATA_ATTACHMENT_BLOCK";
Chris Lattner4238d472007-04-29 20:00:02 +0000105 }
106}
107
Chris Lattnerb30c9252007-04-29 21:48:19 +0000108/// GetCodeName - Return a symbolic code name if known, otherwise return
109/// null.
Chris Lattnerf9a3ec82009-04-26 22:21:57 +0000110static const char *GetCodeName(unsigned CodeID, unsigned BlockID,
111 const BitstreamReader &StreamFile) {
Chris Lattner1772b122007-05-05 00:17:42 +0000112 // Standard blocks for all bitcode files.
113 if (BlockID < bitc::FIRST_APPLICATION_BLOCKID) {
114 if (BlockID == bitc::BLOCKINFO_BLOCK_ID) {
115 switch (CodeID) {
116 default: return 0;
Chris Lattnerf9a3ec82009-04-26 22:21:57 +0000117 case bitc::BLOCKINFO_CODE_SETBID: return "SETBID";
118 case bitc::BLOCKINFO_CODE_BLOCKNAME: return "BLOCKNAME";
119 case bitc::BLOCKINFO_CODE_SETRECORDNAME: return "SETRECORDNAME";
Chris Lattner1772b122007-05-05 00:17:42 +0000120 }
121 }
122 return 0;
123 }
124
Chris Lattnerf9a3ec82009-04-26 22:21:57 +0000125 // Check to see if we have a blockinfo record for this record, with a name.
126 if (const BitstreamReader::BlockInfo *Info =
127 StreamFile.getBlockInfo(BlockID)) {
128 for (unsigned i = 0, e = Info->RecordNames.size(); i != e; ++i)
129 if (Info->RecordNames[i].first == CodeID)
130 return Info->RecordNames[i].second.c_str();
131 }
132
133
Chris Lattnerb30c9252007-04-29 21:48:19 +0000134 if (CurStreamType != LLVMIRBitstream) return 0;
135
136 switch (BlockID) {
137 default: return 0;
138 case bitc::MODULE_BLOCK_ID:
139 switch (CodeID) {
140 default: return 0;
141 case bitc::MODULE_CODE_VERSION: return "VERSION";
142 case bitc::MODULE_CODE_TRIPLE: return "TRIPLE";
143 case bitc::MODULE_CODE_DATALAYOUT: return "DATALAYOUT";
144 case bitc::MODULE_CODE_ASM: return "ASM";
145 case bitc::MODULE_CODE_SECTIONNAME: return "SECTIONNAME";
146 case bitc::MODULE_CODE_DEPLIB: return "DEPLIB";
147 case bitc::MODULE_CODE_GLOBALVAR: return "GLOBALVAR";
148 case bitc::MODULE_CODE_FUNCTION: return "FUNCTION";
149 case bitc::MODULE_CODE_ALIAS: return "ALIAS";
150 case bitc::MODULE_CODE_PURGEVALS: return "PURGEVALS";
Nick Lewycky82b80d92008-11-07 14:52:51 +0000151 case bitc::MODULE_CODE_GCNAME: return "GCNAME";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000152 }
Chris Lattnercd5b7d72007-05-04 03:01:41 +0000153 case bitc::PARAMATTR_BLOCK_ID:
154 switch (CodeID) {
155 default: return 0;
156 case bitc::PARAMATTR_CODE_ENTRY: return "ENTRY";
157 }
Chris Lattnerb30c9252007-04-29 21:48:19 +0000158 case bitc::TYPE_BLOCK_ID:
159 switch (CodeID) {
160 default: return 0;
Nick Lewycky82b80d92008-11-07 14:52:51 +0000161 case bitc::TYPE_CODE_NUMENTRY: return "NUMENTRY";
162 case bitc::TYPE_CODE_VOID: return "VOID";
163 case bitc::TYPE_CODE_FLOAT: return "FLOAT";
164 case bitc::TYPE_CODE_DOUBLE: return "DOUBLE";
165 case bitc::TYPE_CODE_LABEL: return "LABEL";
166 case bitc::TYPE_CODE_OPAQUE: return "OPAQUE";
167 case bitc::TYPE_CODE_INTEGER: return "INTEGER";
168 case bitc::TYPE_CODE_POINTER: return "POINTER";
169 case bitc::TYPE_CODE_FUNCTION: return "FUNCTION";
170 case bitc::TYPE_CODE_STRUCT: return "STRUCT";
171 case bitc::TYPE_CODE_ARRAY: return "ARRAY";
172 case bitc::TYPE_CODE_VECTOR: return "VECTOR";
173 case bitc::TYPE_CODE_X86_FP80: return "X86_FP80";
174 case bitc::TYPE_CODE_FP128: return "FP128";
175 case bitc::TYPE_CODE_PPC_FP128: return "PPC_FP128";
Nick Lewycky079c0342009-06-01 04:41:03 +0000176 case bitc::TYPE_CODE_METADATA: return "METADATA";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000177 }
178
179 case bitc::CONSTANTS_BLOCK_ID:
180 switch (CodeID) {
181 default: return 0;
182 case bitc::CST_CODE_SETTYPE: return "SETTYPE";
183 case bitc::CST_CODE_NULL: return "NULL";
184 case bitc::CST_CODE_UNDEF: return "UNDEF";
185 case bitc::CST_CODE_INTEGER: return "INTEGER";
186 case bitc::CST_CODE_WIDE_INTEGER: return "WIDE_INTEGER";
187 case bitc::CST_CODE_FLOAT: return "FLOAT";
188 case bitc::CST_CODE_AGGREGATE: return "AGGREGATE";
Chris Lattnercb3d91b2007-05-06 00:53:07 +0000189 case bitc::CST_CODE_STRING: return "STRING";
190 case bitc::CST_CODE_CSTRING: return "CSTRING";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000191 case bitc::CST_CODE_CE_BINOP: return "CE_BINOP";
192 case bitc::CST_CODE_CE_CAST: return "CE_CAST";
193 case bitc::CST_CODE_CE_GEP: return "CE_GEP";
194 case bitc::CST_CODE_CE_SELECT: return "CE_SELECT";
195 case bitc::CST_CODE_CE_EXTRACTELT: return "CE_EXTRACTELT";
196 case bitc::CST_CODE_CE_INSERTELT: return "CE_INSERTELT";
197 case bitc::CST_CODE_CE_SHUFFLEVEC: return "CE_SHUFFLEVEC";
198 case bitc::CST_CODE_CE_CMP: return "CE_CMP";
Chris Lattnerc5ff2cc2007-05-06 01:50:11 +0000199 case bitc::CST_CODE_INLINEASM: return "INLINEASM";
Nick Lewycky079c0342009-06-01 04:41:03 +0000200 case bitc::CST_CODE_CE_SHUFVEC_EX: return "CE_SHUFVEC_EX";
Chris Lattnerb30c9252007-04-29 21:48:19 +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
Chris Lattner8f926682007-05-01 02:43:46 +0000223 case bitc::FUNC_CODE_INST_PHI: return "INST_PHI";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000224 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 Lambfe63fb92007-12-11 08:59:05 +0000231 case bitc::FUNC_CODE_INST_STORE2: return "INST_STORE2";
Nick Lewycky33a834a2008-03-01 21:47:06 +0000232 case bitc::FUNC_CODE_INST_GETRESULT: return "INST_GETRESULT";
Nick Lewycky82b80d92008-11-07 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";
Chris Lattnerb30c9252007-04-29 21:48:19 +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";
Chris Lattnercd5b7d72007-05-04 03:01:41 +0000247 case bitc::VST_CODE_BBENTRY: return "BBENTRY";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000248 }
Devang Patele8e02132009-09-18 19:26:43 +0000249 case bitc::METADATA_ATTACHMENT_ID:
250 switch(CodeID) {
251 default:return 0;
252 case bitc::METADATA_ATTACHMENT: return "METADATA_ATTACHMENT";
253 }
Devang Patele54abc92009-07-22 17:43:22 +0000254 case bitc::METADATA_BLOCK_ID:
255 switch(CodeID) {
256 default:return 0;
257 case bitc::METADATA_STRING: return "MDSTRING";
Devang Patel104cf9e2009-07-23 01:07:34 +0000258 case bitc::METADATA_NODE: return "MDNODE";
Devang Patel124e6eb2009-07-30 23:03:19 +0000259 case bitc::METADATA_NAME: return "METADATA_NAME";
260 case bitc::METADATA_NAMED_NODE: return "NAMEDMDNODE";
Devang Patele8e02132009-09-18 19:26:43 +0000261 case bitc::METADATA_KIND: return "METADATA_KIND";
Devang Patele54abc92009-07-22 17:43:22 +0000262 }
Chris Lattnerb30c9252007-04-29 21:48:19 +0000263 }
264}
265
Chris Lattner24437472009-04-27 17:59:34 +0000266struct PerRecordStats {
267 unsigned NumInstances;
Chris Lattnerc167cac2009-04-27 18:15:27 +0000268 unsigned NumAbbrev;
269 uint64_t TotalBits;
270
271 PerRecordStats() : NumInstances(0), NumAbbrev(0), TotalBits(0) {}
Chris Lattner24437472009-04-27 17:59:34 +0000272};
Chris Lattner4238d472007-04-29 20:00:02 +0000273
274struct PerBlockIDStats {
275 /// NumInstances - This the number of times this block ID has been seen.
276 unsigned NumInstances;
277
278 /// NumBits - The total size in bits of all of these blocks.
279 uint64_t NumBits;
280
281 /// NumSubBlocks - The total number of blocks these blocks contain.
282 unsigned NumSubBlocks;
283
284 /// NumAbbrevs - The total number of abbreviations.
285 unsigned NumAbbrevs;
286
287 /// NumRecords - The total number of records these blocks contain, and the
288 /// number that are abbreviated.
289 unsigned NumRecords, NumAbbreviatedRecords;
290
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000291 /// CodeFreq - Keep track of the number of times we see each code.
Chris Lattner24437472009-04-27 17:59:34 +0000292 std::vector<PerRecordStats> CodeFreq;
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000293
Chris Lattner4238d472007-04-29 20:00:02 +0000294 PerBlockIDStats()
295 : NumInstances(0), NumBits(0),
296 NumSubBlocks(0), NumAbbrevs(0), NumRecords(0), NumAbbreviatedRecords(0) {}
297};
298
299static std::map<unsigned, PerBlockIDStats> BlockIDStats;
300
301
302
Chris Lattner32de6332007-04-29 08:31:14 +0000303/// Error - All bitcode analysis errors go through this function, making this a
304/// good place to breakpoint if debugging.
305static bool Error(const std::string &Err) {
Dan Gohman65f57c22009-07-15 16:35:29 +0000306 errs() << Err << "\n";
Chris Lattner32de6332007-04-29 08:31:14 +0000307 return true;
308}
309
310/// ParseBlock - Read a block, updating statistics, etc.
Chris Lattner962dde32009-04-26 20:59:02 +0000311static bool ParseBlock(BitstreamCursor &Stream, unsigned IndentLevel) {
Chris Lattner1772b122007-05-05 00:17:42 +0000312 std::string Indent(IndentLevel*2, ' ');
Chris Lattner4238d472007-04-29 20:00:02 +0000313 uint64_t BlockBitStart = Stream.GetCurrentBitNo();
Chris Lattner32de6332007-04-29 08:31:14 +0000314 unsigned BlockID = Stream.ReadSubBlockID();
Chris Lattnerb30c9252007-04-29 21:48:19 +0000315
Chris Lattner4238d472007-04-29 20:00:02 +0000316 // Get the statistics for this BlockID.
317 PerBlockIDStats &BlockStats = BlockIDStats[BlockID];
318
319 BlockStats.NumInstances++;
Chris Lattner32de6332007-04-29 08:31:14 +0000320
Chris Lattner1772b122007-05-05 00:17:42 +0000321 // BLOCKINFO is a special part of the stream.
322 if (BlockID == bitc::BLOCKINFO_BLOCK_ID) {
Dan Gohman65f57c22009-07-15 16:35:29 +0000323 if (Dump) errs() << Indent << "<BLOCKINFO_BLOCK/>\n";
Chris Lattner1772b122007-05-05 00:17:42 +0000324 if (Stream.ReadBlockInfoBlock())
325 return Error("Malformed BlockInfoBlock");
326 uint64_t BlockBitEnd = Stream.GetCurrentBitNo();
327 BlockStats.NumBits += BlockBitEnd-BlockBitStart;
328 return false;
329 }
330
Chris Lattnerb30c9252007-04-29 21:48:19 +0000331 unsigned NumWords = 0;
Chris Lattner1772b122007-05-05 00:17:42 +0000332 if (Stream.EnterSubBlock(BlockID, &NumWords))
Chris Lattner32de6332007-04-29 08:31:14 +0000333 return Error("Malformed block record");
334
Chris Lattnerb30c9252007-04-29 21:48:19 +0000335 const char *BlockName = 0;
336 if (Dump) {
Dan Gohman65f57c22009-07-15 16:35:29 +0000337 errs() << Indent << "<";
Chris Lattnerf9a3ec82009-04-26 22:21:57 +0000338 if ((BlockName = GetBlockName(BlockID, *Stream.getBitStreamReader())))
Dan Gohman65f57c22009-07-15 16:35:29 +0000339 errs() << BlockName;
Chris Lattnerb30c9252007-04-29 21:48:19 +0000340 else
Dan Gohman65f57c22009-07-15 16:35:29 +0000341 errs() << "UnknownBlock" << BlockID;
Chris Lattnerb30c9252007-04-29 21:48:19 +0000342
343 if (NonSymbolic && BlockName)
Dan Gohman65f57c22009-07-15 16:35:29 +0000344 errs() << " BlockID=" << BlockID;
Chris Lattnerb30c9252007-04-29 21:48:19 +0000345
Dan Gohman65f57c22009-07-15 16:35:29 +0000346 errs() << " NumWords=" << NumWords
347 << " BlockCodeSize=" << Stream.GetAbbrevIDWidth() << ">\n";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000348 }
349
Chris Lattner32de6332007-04-29 08:31:14 +0000350 SmallVector<uint64_t, 64> Record;
351
352 // Read all the records for this block.
353 while (1) {
354 if (Stream.AtEndOfStream())
355 return Error("Premature end of bitstream");
356
Chris Lattnerc167cac2009-04-27 18:15:27 +0000357 uint64_t RecordStartBit = Stream.GetCurrentBitNo();
358
Chris Lattner32de6332007-04-29 08:31:14 +0000359 // Read the code for this record.
360 unsigned AbbrevID = Stream.ReadCode();
361 switch (AbbrevID) {
Chris Lattner4238d472007-04-29 20:00:02 +0000362 case bitc::END_BLOCK: {
Chris Lattner32de6332007-04-29 08:31:14 +0000363 if (Stream.ReadBlockEnd())
364 return Error("Error at end of block");
Chris Lattner4238d472007-04-29 20:00:02 +0000365 uint64_t BlockBitEnd = Stream.GetCurrentBitNo();
366 BlockStats.NumBits += BlockBitEnd-BlockBitStart;
Chris Lattnerb30c9252007-04-29 21:48:19 +0000367 if (Dump) {
Dan Gohman65f57c22009-07-15 16:35:29 +0000368 errs() << Indent << "</";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000369 if (BlockName)
Dan Gohman65f57c22009-07-15 16:35:29 +0000370 errs() << BlockName << ">\n";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000371 else
Dan Gohman65f57c22009-07-15 16:35:29 +0000372 errs() << "UnknownBlock" << BlockID << ">\n";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000373 }
Chris Lattner32de6332007-04-29 08:31:14 +0000374 return false;
Chris Lattner4238d472007-04-29 20:00:02 +0000375 }
Chris Lattner44b0f102007-05-05 01:29:31 +0000376 case bitc::ENTER_SUBBLOCK: {
377 uint64_t SubBlockBitStart = Stream.GetCurrentBitNo();
Chris Lattnerb30c9252007-04-29 21:48:19 +0000378 if (ParseBlock(Stream, IndentLevel+1))
Chris Lattner32de6332007-04-29 08:31:14 +0000379 return true;
Chris Lattner4238d472007-04-29 20:00:02 +0000380 ++BlockStats.NumSubBlocks;
Chris Lattner44b0f102007-05-05 01:29:31 +0000381 uint64_t SubBlockBitEnd = Stream.GetCurrentBitNo();
382
383 // Don't include subblock sizes in the size of this block.
384 BlockBitStart += SubBlockBitEnd-SubBlockBitStart;
Chris Lattner32de6332007-04-29 08:31:14 +0000385 break;
Chris Lattner44b0f102007-05-05 01:29:31 +0000386 }
Chris Lattner32de6332007-04-29 08:31:14 +0000387 case bitc::DEFINE_ABBREV:
388 Stream.ReadAbbrevRecord();
Chris Lattner4238d472007-04-29 20:00:02 +0000389 ++BlockStats.NumAbbrevs;
Chris Lattner32de6332007-04-29 08:31:14 +0000390 break;
391 default:
392 Record.clear();
Chris Lattner3f75d312009-04-06 22:44:40 +0000393
394 ++BlockStats.NumRecords;
Chris Lattnerae7dd802009-04-07 02:56:46 +0000395 if (AbbrevID != bitc::UNABBREV_RECORD)
Chris Lattner3f75d312009-04-06 22:44:40 +0000396 ++BlockStats.NumAbbreviatedRecords;
Chris Lattner3f75d312009-04-06 22:44:40 +0000397
Chris Lattner3f75d312009-04-06 22:44:40 +0000398 const char *BlobStart = 0;
399 unsigned BlobLen = 0;
Chris Lattnerae7dd802009-04-07 02:56:46 +0000400 unsigned Code = Stream.ReadRecord(AbbrevID, Record, BlobStart, BlobLen);
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000401
Chris Lattnerc167cac2009-04-27 18:15:27 +0000402
403
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000404 // Increment the # occurrences of this code.
405 if (BlockStats.CodeFreq.size() <= Code)
406 BlockStats.CodeFreq.resize(Code+1);
Chris Lattner24437472009-04-27 17:59:34 +0000407 BlockStats.CodeFreq[Code].NumInstances++;
Chris Lattnerc167cac2009-04-27 18:15:27 +0000408 BlockStats.CodeFreq[Code].TotalBits +=
409 Stream.GetCurrentBitNo()-RecordStartBit;
410 if (AbbrevID != bitc::UNABBREV_RECORD)
411 BlockStats.CodeFreq[Code].NumAbbrev++;
412
Chris Lattnerb30c9252007-04-29 21:48:19 +0000413 if (Dump) {
Dan Gohman65f57c22009-07-15 16:35:29 +0000414 errs() << Indent << " <";
Chris Lattnerf9a3ec82009-04-26 22:21:57 +0000415 if (const char *CodeName =
416 GetCodeName(Code, BlockID, *Stream.getBitStreamReader()))
Dan Gohman65f57c22009-07-15 16:35:29 +0000417 errs() << CodeName;
Chris Lattnerb30c9252007-04-29 21:48:19 +0000418 else
Dan Gohman65f57c22009-07-15 16:35:29 +0000419 errs() << "UnknownCode" << Code;
Chris Lattnerf9a3ec82009-04-26 22:21:57 +0000420 if (NonSymbolic &&
421 GetCodeName(Code, BlockID, *Stream.getBitStreamReader()))
Dan Gohman65f57c22009-07-15 16:35:29 +0000422 errs() << " codeid=" << Code;
Chris Lattnerb30c9252007-04-29 21:48:19 +0000423 if (AbbrevID != bitc::UNABBREV_RECORD)
Dan Gohman65f57c22009-07-15 16:35:29 +0000424 errs() << " abbrevid=" << AbbrevID;
Chris Lattnerb30c9252007-04-29 21:48:19 +0000425
426 for (unsigned i = 0, e = Record.size(); i != e; ++i)
Dan Gohman65f57c22009-07-15 16:35:29 +0000427 errs() << " op" << i << "=" << (int64_t)Record[i];
Chris Lattnerb30c9252007-04-29 21:48:19 +0000428
Dan Gohman65f57c22009-07-15 16:35:29 +0000429 errs() << "/>";
Chris Lattnerae7dd802009-04-07 02:56:46 +0000430
431 if (BlobStart) {
Dan Gohman65f57c22009-07-15 16:35:29 +0000432 errs() << " blob data = ";
Chris Lattnerae7dd802009-04-07 02:56:46 +0000433 bool BlobIsPrintable = true;
434 for (unsigned i = 0; i != BlobLen; ++i)
435 if (!isprint(BlobStart[i])) {
436 BlobIsPrintable = false;
437 break;
438 }
439
440 if (BlobIsPrintable)
Dan Gohman65f57c22009-07-15 16:35:29 +0000441 errs() << "'" << std::string(BlobStart, BlobStart+BlobLen) <<"'";
Chris Lattnerae7dd802009-04-07 02:56:46 +0000442 else
Dan Gohman65f57c22009-07-15 16:35:29 +0000443 errs() << "unprintable, " << BlobLen << " bytes.";
Chris Lattnerae7dd802009-04-07 02:56:46 +0000444 }
445
Dan Gohman65f57c22009-07-15 16:35:29 +0000446 errs() << "\n";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000447 }
448
Chris Lattner32de6332007-04-29 08:31:14 +0000449 break;
450 }
451 }
452}
453
Chris Lattner4238d472007-04-29 20:00:02 +0000454static void PrintSize(double Bits) {
Chris Lattner24437472009-04-27 17:59:34 +0000455 fprintf(stderr, "%.2f/%.2fB/%lluW", Bits, Bits/8,(unsigned long long)Bits/32);
456}
457static void PrintSize(uint64_t Bits) {
458 fprintf(stderr, "%llub/%.2fB/%lluW", (unsigned long long)Bits,
459 (double)Bits/8, (unsigned long long)Bits/32);
Chris Lattner4238d472007-04-29 20:00:02 +0000460}
461
462
Chris Lattner45e0f892007-04-29 08:12:22 +0000463/// AnalyzeBitcode - Analyze the bitcode file specified by InputFilename.
464static int AnalyzeBitcode() {
465 // Read the input file.
Chris Lattnere2a466b2009-04-06 20:54:32 +0000466 MemoryBuffer *MemBuf = MemoryBuffer::getFileOrSTDIN(InputFilename.c_str());
Chris Lattner45e0f892007-04-29 08:12:22 +0000467
Chris Lattnere2a466b2009-04-06 20:54:32 +0000468 if (MemBuf == 0)
Chris Lattner32de6332007-04-29 08:31:14 +0000469 return Error("Error reading '" + InputFilename + "'.");
Chris Lattner45e0f892007-04-29 08:12:22 +0000470
Chris Lattnere2a466b2009-04-06 20:54:32 +0000471 if (MemBuf->getBufferSize() & 3)
Chris Lattner32de6332007-04-29 08:31:14 +0000472 return Error("Bitcode stream should be a multiple of 4 bytes in length");
Chris Lattner45e0f892007-04-29 08:12:22 +0000473
Chris Lattnere2a466b2009-04-06 20:54:32 +0000474 unsigned char *BufPtr = (unsigned char *)MemBuf->getBufferStart();
475 unsigned char *EndBufPtr = BufPtr+MemBuf->getBufferSize();
476
477 // If we have a wrapper header, parse it and ignore the non-bc file contents.
478 // The magic number is 0x0B17C0DE stored in little endian.
479 if (isBitcodeWrapper(BufPtr, EndBufPtr))
480 if (SkipBitcodeWrapperHeader(BufPtr, EndBufPtr))
481 return Error("Invalid bitcode wrapper header");
482
Chris Lattner962dde32009-04-26 20:59:02 +0000483 BitstreamReader StreamFile(BufPtr, EndBufPtr);
484 BitstreamCursor Stream(StreamFile);
Chris Lattner0370cc62009-04-27 20:04:08 +0000485 StreamFile.CollectBlockInfoNames();
Chris Lattner45e0f892007-04-29 08:12:22 +0000486
487 // Read the stream signature.
488 char Signature[6];
489 Signature[0] = Stream.Read(8);
490 Signature[1] = Stream.Read(8);
491 Signature[2] = Stream.Read(4);
492 Signature[3] = Stream.Read(4);
493 Signature[4] = Stream.Read(4);
494 Signature[5] = Stream.Read(4);
495
Chris Lattner32de6332007-04-29 08:31:14 +0000496 // Autodetect the file contents, if it is one we know.
Chris Lattner45e0f892007-04-29 08:12:22 +0000497 CurStreamType = UnknownBitstream;
498 if (Signature[0] == 'B' && Signature[1] == 'C' &&
499 Signature[2] == 0x0 && Signature[3] == 0xC &&
500 Signature[4] == 0xE && Signature[5] == 0xD)
501 CurStreamType = LLVMIRBitstream;
502
Chris Lattner4238d472007-04-29 20:00:02 +0000503 unsigned NumTopBlocks = 0;
504
Chris Lattner32de6332007-04-29 08:31:14 +0000505 // Parse the top-level structure. We only allow blocks at the top-level.
506 while (!Stream.AtEndOfStream()) {
507 unsigned Code = Stream.ReadCode();
508 if (Code != bitc::ENTER_SUBBLOCK)
509 return Error("Invalid record at top-level");
510
Chris Lattnerb30c9252007-04-29 21:48:19 +0000511 if (ParseBlock(Stream, 0))
Chris Lattner32de6332007-04-29 08:31:14 +0000512 return true;
Chris Lattner4238d472007-04-29 20:00:02 +0000513 ++NumTopBlocks;
Chris Lattner32de6332007-04-29 08:31:14 +0000514 }
515
Dan Gohman65f57c22009-07-15 16:35:29 +0000516 if (Dump) errs() << "\n\n";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000517
Chris Lattnere2a466b2009-04-06 20:54:32 +0000518 uint64_t BufferSizeBits = (EndBufPtr-BufPtr)*CHAR_BIT;
Chris Lattner32de6332007-04-29 08:31:14 +0000519 // Print a summary of the read file.
Dan Gohman65f57c22009-07-15 16:35:29 +0000520 errs() << "Summary of " << InputFilename << ":\n";
521 errs() << " Total size: ";
Chris Lattner8f926682007-05-01 02:43:46 +0000522 PrintSize(BufferSizeBits);
Dan Gohman65f57c22009-07-15 16:35:29 +0000523 errs() << "\n";
524 errs() << " Stream type: ";
Chris Lattner45e0f892007-04-29 08:12:22 +0000525 switch (CurStreamType) {
526 default: assert(0 && "Unknown bitstream type");
Dan Gohman65f57c22009-07-15 16:35:29 +0000527 case UnknownBitstream: errs() << "unknown\n"; break;
528 case LLVMIRBitstream: errs() << "LLVM IR\n"; break;
Chris Lattner45e0f892007-04-29 08:12:22 +0000529 }
Dan Gohman65f57c22009-07-15 16:35:29 +0000530 errs() << " # Toplevel Blocks: " << NumTopBlocks << "\n";
531 errs() << "\n";
Chris Lattner4238d472007-04-29 20:00:02 +0000532
533 // Emit per-block stats.
Dan Gohman65f57c22009-07-15 16:35:29 +0000534 errs() << "Per-block Summary:\n";
Chris Lattner4238d472007-04-29 20:00:02 +0000535 for (std::map<unsigned, PerBlockIDStats>::iterator I = BlockIDStats.begin(),
536 E = BlockIDStats.end(); I != E; ++I) {
Dan Gohman65f57c22009-07-15 16:35:29 +0000537 errs() << " Block ID #" << I->first;
Chris Lattnerf9a3ec82009-04-26 22:21:57 +0000538 if (const char *BlockName = GetBlockName(I->first, StreamFile))
Dan Gohman65f57c22009-07-15 16:35:29 +0000539 errs() << " (" << BlockName << ")";
540 errs() << ":\n";
Chris Lattner4238d472007-04-29 20:00:02 +0000541
542 const PerBlockIDStats &Stats = I->second;
Dan Gohman65f57c22009-07-15 16:35:29 +0000543 errs() << " Num Instances: " << Stats.NumInstances << "\n";
544 errs() << " Total Size: ";
Chris Lattner4238d472007-04-29 20:00:02 +0000545 PrintSize(Stats.NumBits);
Dan Gohman65f57c22009-07-15 16:35:29 +0000546 errs() << "\n";
547 errs() << " % of file: "
548 << Stats.NumBits/(double)BufferSizeBits*100 << "\n";
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000549 if (Stats.NumInstances > 1) {
Dan Gohman65f57c22009-07-15 16:35:29 +0000550 errs() << " Average Size: ";
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000551 PrintSize(Stats.NumBits/(double)Stats.NumInstances);
Dan Gohman65f57c22009-07-15 16:35:29 +0000552 errs() << "\n";
553 errs() << " Tot/Avg SubBlocks: " << Stats.NumSubBlocks << "/"
554 << Stats.NumSubBlocks/(double)Stats.NumInstances << "\n";
555 errs() << " Tot/Avg Abbrevs: " << Stats.NumAbbrevs << "/"
556 << Stats.NumAbbrevs/(double)Stats.NumInstances << "\n";
557 errs() << " Tot/Avg Records: " << Stats.NumRecords << "/"
558 << Stats.NumRecords/(double)Stats.NumInstances << "\n";
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000559 } else {
Dan Gohman65f57c22009-07-15 16:35:29 +0000560 errs() << " Num SubBlocks: " << Stats.NumSubBlocks << "\n";
561 errs() << " Num Abbrevs: " << Stats.NumAbbrevs << "\n";
562 errs() << " Num Records: " << Stats.NumRecords << "\n";
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000563 }
Chris Lattner1772b122007-05-05 00:17:42 +0000564 if (Stats.NumRecords)
Dan Gohman65f57c22009-07-15 16:35:29 +0000565 errs() << " % Abbrev Recs: " << (Stats.NumAbbreviatedRecords/
566 (double)Stats.NumRecords)*100 << "\n";
567 errs() << "\n";
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000568
569 // Print a histogram of the codes we see.
570 if (!NoHistogram && !Stats.CodeFreq.empty()) {
571 std::vector<std::pair<unsigned, unsigned> > FreqPairs; // <freq,code>
572 for (unsigned i = 0, e = Stats.CodeFreq.size(); i != e; ++i)
Chris Lattner24437472009-04-27 17:59:34 +0000573 if (unsigned Freq = Stats.CodeFreq[i].NumInstances)
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000574 FreqPairs.push_back(std::make_pair(Freq, i));
575 std::stable_sort(FreqPairs.begin(), FreqPairs.end());
576 std::reverse(FreqPairs.begin(), FreqPairs.end());
577
Dan Gohman65f57c22009-07-15 16:35:29 +0000578 errs() << "\tRecord Histogram:\n";
Chris Lattnerc167cac2009-04-27 18:15:27 +0000579 fprintf(stderr, "\t\t Count # Bits %% Abv Record Kind\n");
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000580 for (unsigned i = 0, e = FreqPairs.size(); i != e; ++i) {
Chris Lattnerc167cac2009-04-27 18:15:27 +0000581 const PerRecordStats &RecStats = Stats.CodeFreq[FreqPairs[i].second];
582
583 fprintf(stderr, "\t\t%7d %9llu ", RecStats.NumInstances,
Dan Gohmanfbccdef2009-05-01 16:33:33 +0000584 (unsigned long long)RecStats.TotalBits);
Chris Lattnerc167cac2009-04-27 18:15:27 +0000585
586 if (RecStats.NumAbbrev)
587 fprintf(stderr, "%7.2f ",
588 (double)RecStats.NumAbbrev/RecStats.NumInstances*100);
589 else
590 fprintf(stderr, " ");
Chris Lattner24437472009-04-27 17:59:34 +0000591
Chris Lattnerf9a3ec82009-04-26 22:21:57 +0000592 if (const char *CodeName =
593 GetCodeName(FreqPairs[i].second, I->first, StreamFile))
Chris Lattner24437472009-04-27 17:59:34 +0000594 fprintf(stderr, "%s\n", CodeName);
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000595 else
Chris Lattner24437472009-04-27 17:59:34 +0000596 fprintf(stderr, "UnknownCode%d\n", FreqPairs[i].second);
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000597 }
Dan Gohman65f57c22009-07-15 16:35:29 +0000598 errs() << "\n";
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000599
600 }
Chris Lattner4238d472007-04-29 20:00:02 +0000601 }
Chris Lattner45e0f892007-04-29 08:12:22 +0000602 return 0;
603}
Reid Spencerdac69c82004-06-07 17:53:43 +0000604
Chris Lattner32de6332007-04-29 08:31:14 +0000605
Chris Lattnerc30598b2006-12-06 01:18:01 +0000606int main(int argc, char **argv) {
Chris Lattnercc14d252009-03-06 05:34:10 +0000607 // Print a stack trace if we signal out.
Chris Lattner45e0f892007-04-29 08:12:22 +0000608 sys::PrintStackTraceOnErrorSignal();
Chris Lattnercc14d252009-03-06 05:34:10 +0000609 PrettyStackTraceProgram X(argc, argv);
610 llvm_shutdown_obj Y; // Call llvm_shutdown() on exit.
611 cl::ParseCommandLineOptions(argc, argv, "llvm-bcanalyzer file analyzer\n");
Chris Lattner45e0f892007-04-29 08:12:22 +0000612
Chris Lattner44dadff2007-05-06 09:29:57 +0000613 return AnalyzeBitcode();
Reid Spencerdac69c82004-06-07 17:53:43 +0000614}