blob: ce9c29fdf922ca9006e18763fefe3989ec77fbd5 [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) {
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";
Devang Patele54abc92009-07-22 17:43:22 +0000103 case bitc::METADATA_BLOCK_ID: return "METADATA_BLOCK";
Chris Lattner4238d472007-04-29 20:00:02 +0000104 }
105}
106
Chris Lattnerb30c9252007-04-29 21:48:19 +0000107/// GetCodeName - Return a symbolic code name if known, otherwise return
108/// null.
Chris Lattnerf9a3ec82009-04-26 22:21:57 +0000109static const char *GetCodeName(unsigned CodeID, unsigned BlockID,
110 const BitstreamReader &StreamFile) {
Chris Lattner1772b122007-05-05 00:17:42 +0000111 // Standard blocks for all bitcode files.
112 if (BlockID < bitc::FIRST_APPLICATION_BLOCKID) {
113 if (BlockID == bitc::BLOCKINFO_BLOCK_ID) {
114 switch (CodeID) {
115 default: return 0;
Chris Lattnerf9a3ec82009-04-26 22:21:57 +0000116 case bitc::BLOCKINFO_CODE_SETBID: return "SETBID";
117 case bitc::BLOCKINFO_CODE_BLOCKNAME: return "BLOCKNAME";
118 case bitc::BLOCKINFO_CODE_SETRECORDNAME: return "SETRECORDNAME";
Chris Lattner1772b122007-05-05 00:17:42 +0000119 }
120 }
121 return 0;
122 }
123
Chris Lattnerf9a3ec82009-04-26 22:21:57 +0000124 // Check to see if we have a blockinfo record for this record, with a name.
125 if (const BitstreamReader::BlockInfo *Info =
126 StreamFile.getBlockInfo(BlockID)) {
127 for (unsigned i = 0, e = Info->RecordNames.size(); i != e; ++i)
128 if (Info->RecordNames[i].first == CodeID)
129 return Info->RecordNames[i].second.c_str();
130 }
131
132
Chris Lattnerb30c9252007-04-29 21:48:19 +0000133 if (CurStreamType != LLVMIRBitstream) return 0;
134
135 switch (BlockID) {
136 default: return 0;
137 case bitc::MODULE_BLOCK_ID:
138 switch (CodeID) {
139 default: return 0;
140 case bitc::MODULE_CODE_VERSION: return "VERSION";
141 case bitc::MODULE_CODE_TRIPLE: return "TRIPLE";
142 case bitc::MODULE_CODE_DATALAYOUT: return "DATALAYOUT";
143 case bitc::MODULE_CODE_ASM: return "ASM";
144 case bitc::MODULE_CODE_SECTIONNAME: return "SECTIONNAME";
145 case bitc::MODULE_CODE_DEPLIB: return "DEPLIB";
146 case bitc::MODULE_CODE_GLOBALVAR: return "GLOBALVAR";
147 case bitc::MODULE_CODE_FUNCTION: return "FUNCTION";
148 case bitc::MODULE_CODE_ALIAS: return "ALIAS";
149 case bitc::MODULE_CODE_PURGEVALS: return "PURGEVALS";
Nick Lewycky82b80d92008-11-07 14:52:51 +0000150 case bitc::MODULE_CODE_GCNAME: return "GCNAME";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000151 }
Chris Lattnercd5b7d72007-05-04 03:01:41 +0000152 case bitc::PARAMATTR_BLOCK_ID:
153 switch (CodeID) {
154 default: return 0;
155 case bitc::PARAMATTR_CODE_ENTRY: return "ENTRY";
156 }
Chris Lattnerb30c9252007-04-29 21:48:19 +0000157 case bitc::TYPE_BLOCK_ID:
158 switch (CodeID) {
159 default: return 0;
Nick Lewycky82b80d92008-11-07 14:52:51 +0000160 case bitc::TYPE_CODE_NUMENTRY: return "NUMENTRY";
161 case bitc::TYPE_CODE_VOID: return "VOID";
162 case bitc::TYPE_CODE_FLOAT: return "FLOAT";
163 case bitc::TYPE_CODE_DOUBLE: return "DOUBLE";
164 case bitc::TYPE_CODE_LABEL: return "LABEL";
165 case bitc::TYPE_CODE_OPAQUE: return "OPAQUE";
166 case bitc::TYPE_CODE_INTEGER: return "INTEGER";
167 case bitc::TYPE_CODE_POINTER: return "POINTER";
168 case bitc::TYPE_CODE_FUNCTION: return "FUNCTION";
169 case bitc::TYPE_CODE_STRUCT: return "STRUCT";
170 case bitc::TYPE_CODE_ARRAY: return "ARRAY";
171 case bitc::TYPE_CODE_VECTOR: return "VECTOR";
172 case bitc::TYPE_CODE_X86_FP80: return "X86_FP80";
173 case bitc::TYPE_CODE_FP128: return "FP128";
174 case bitc::TYPE_CODE_PPC_FP128: return "PPC_FP128";
Nick Lewycky079c0342009-06-01 04:41:03 +0000175 case bitc::TYPE_CODE_METADATA: return "METADATA";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000176 }
177
178 case bitc::CONSTANTS_BLOCK_ID:
179 switch (CodeID) {
180 default: return 0;
181 case bitc::CST_CODE_SETTYPE: return "SETTYPE";
182 case bitc::CST_CODE_NULL: return "NULL";
183 case bitc::CST_CODE_UNDEF: return "UNDEF";
184 case bitc::CST_CODE_INTEGER: return "INTEGER";
185 case bitc::CST_CODE_WIDE_INTEGER: return "WIDE_INTEGER";
186 case bitc::CST_CODE_FLOAT: return "FLOAT";
187 case bitc::CST_CODE_AGGREGATE: return "AGGREGATE";
Chris Lattnercb3d91b2007-05-06 00:53:07 +0000188 case bitc::CST_CODE_STRING: return "STRING";
189 case bitc::CST_CODE_CSTRING: return "CSTRING";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000190 case bitc::CST_CODE_CE_BINOP: return "CE_BINOP";
191 case bitc::CST_CODE_CE_CAST: return "CE_CAST";
192 case bitc::CST_CODE_CE_GEP: return "CE_GEP";
193 case bitc::CST_CODE_CE_SELECT: return "CE_SELECT";
194 case bitc::CST_CODE_CE_EXTRACTELT: return "CE_EXTRACTELT";
195 case bitc::CST_CODE_CE_INSERTELT: return "CE_INSERTELT";
196 case bitc::CST_CODE_CE_SHUFFLEVEC: return "CE_SHUFFLEVEC";
197 case bitc::CST_CODE_CE_CMP: return "CE_CMP";
Chris Lattnerc5ff2cc2007-05-06 01:50:11 +0000198 case bitc::CST_CODE_INLINEASM: return "INLINEASM";
Nick Lewycky079c0342009-06-01 04:41:03 +0000199 case bitc::CST_CODE_CE_SHUFVEC_EX: return "CE_SHUFVEC_EX";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000200 }
201 case bitc::FUNCTION_BLOCK_ID:
202 switch (CodeID) {
203 default: return 0;
204 case bitc::FUNC_CODE_DECLAREBLOCKS: return "DECLAREBLOCKS";
205
206 case bitc::FUNC_CODE_INST_BINOP: return "INST_BINOP";
207 case bitc::FUNC_CODE_INST_CAST: return "INST_CAST";
208 case bitc::FUNC_CODE_INST_GEP: return "INST_GEP";
209 case bitc::FUNC_CODE_INST_SELECT: return "INST_SELECT";
210 case bitc::FUNC_CODE_INST_EXTRACTELT: return "INST_EXTRACTELT";
211 case bitc::FUNC_CODE_INST_INSERTELT: return "INST_INSERTELT";
212 case bitc::FUNC_CODE_INST_SHUFFLEVEC: return "INST_SHUFFLEVEC";
213 case bitc::FUNC_CODE_INST_CMP: return "INST_CMP";
214
215 case bitc::FUNC_CODE_INST_RET: return "INST_RET";
216 case bitc::FUNC_CODE_INST_BR: return "INST_BR";
217 case bitc::FUNC_CODE_INST_SWITCH: return "INST_SWITCH";
218 case bitc::FUNC_CODE_INST_INVOKE: return "INST_INVOKE";
219 case bitc::FUNC_CODE_INST_UNWIND: return "INST_UNWIND";
220 case bitc::FUNC_CODE_INST_UNREACHABLE: return "INST_UNREACHABLE";
221
Chris Lattner8f926682007-05-01 02:43:46 +0000222 case bitc::FUNC_CODE_INST_PHI: return "INST_PHI";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000223 case bitc::FUNC_CODE_INST_MALLOC: return "INST_MALLOC";
224 case bitc::FUNC_CODE_INST_FREE: return "INST_FREE";
225 case bitc::FUNC_CODE_INST_ALLOCA: return "INST_ALLOCA";
226 case bitc::FUNC_CODE_INST_LOAD: return "INST_LOAD";
227 case bitc::FUNC_CODE_INST_STORE: return "INST_STORE";
228 case bitc::FUNC_CODE_INST_CALL: return "INST_CALL";
229 case bitc::FUNC_CODE_INST_VAARG: return "INST_VAARG";
Christopher Lambfe63fb92007-12-11 08:59:05 +0000230 case bitc::FUNC_CODE_INST_STORE2: return "INST_STORE2";
Nick Lewycky33a834a2008-03-01 21:47:06 +0000231 case bitc::FUNC_CODE_INST_GETRESULT: return "INST_GETRESULT";
Nick Lewycky82b80d92008-11-07 14:52:51 +0000232 case bitc::FUNC_CODE_INST_EXTRACTVAL: return "INST_EXTRACTVAL";
233 case bitc::FUNC_CODE_INST_INSERTVAL: return "INST_INSERTVAL";
234 case bitc::FUNC_CODE_INST_CMP2: return "INST_CMP2";
235 case bitc::FUNC_CODE_INST_VSELECT: return "INST_VSELECT";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000236 }
237 case bitc::TYPE_SYMTAB_BLOCK_ID:
238 switch (CodeID) {
239 default: return 0;
240 case bitc::TST_CODE_ENTRY: return "ENTRY";
241 }
242 case bitc::VALUE_SYMTAB_BLOCK_ID:
243 switch (CodeID) {
244 default: return 0;
245 case bitc::VST_CODE_ENTRY: return "ENTRY";
Chris Lattnercd5b7d72007-05-04 03:01:41 +0000246 case bitc::VST_CODE_BBENTRY: return "BBENTRY";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000247 }
Devang Patele54abc92009-07-22 17:43:22 +0000248 case bitc::METADATA_BLOCK_ID:
249 switch(CodeID) {
250 default:return 0;
251 case bitc::METADATA_STRING: return "MDSTRING";
Devang Patel104cf9e2009-07-23 01:07:34 +0000252 case bitc::METADATA_NODE: return "MDNODE";
Devang Patel124e6eb2009-07-30 23:03:19 +0000253 case bitc::METADATA_NAME: return "METADATA_NAME";
254 case bitc::METADATA_NAMED_NODE: return "NAMEDMDNODE";
Devang Patele54abc92009-07-22 17:43:22 +0000255 }
Chris Lattnerb30c9252007-04-29 21:48:19 +0000256 }
257}
258
Chris Lattner24437472009-04-27 17:59:34 +0000259struct PerRecordStats {
260 unsigned NumInstances;
Chris Lattnerc167cac2009-04-27 18:15:27 +0000261 unsigned NumAbbrev;
262 uint64_t TotalBits;
263
264 PerRecordStats() : NumInstances(0), NumAbbrev(0), TotalBits(0) {}
Chris Lattner24437472009-04-27 17:59:34 +0000265};
Chris Lattner4238d472007-04-29 20:00:02 +0000266
267struct PerBlockIDStats {
268 /// NumInstances - This the number of times this block ID has been seen.
269 unsigned NumInstances;
270
271 /// NumBits - The total size in bits of all of these blocks.
272 uint64_t NumBits;
273
274 /// NumSubBlocks - The total number of blocks these blocks contain.
275 unsigned NumSubBlocks;
276
277 /// NumAbbrevs - The total number of abbreviations.
278 unsigned NumAbbrevs;
279
280 /// NumRecords - The total number of records these blocks contain, and the
281 /// number that are abbreviated.
282 unsigned NumRecords, NumAbbreviatedRecords;
283
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000284 /// CodeFreq - Keep track of the number of times we see each code.
Chris Lattner24437472009-04-27 17:59:34 +0000285 std::vector<PerRecordStats> CodeFreq;
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000286
Chris Lattner4238d472007-04-29 20:00:02 +0000287 PerBlockIDStats()
288 : NumInstances(0), NumBits(0),
289 NumSubBlocks(0), NumAbbrevs(0), NumRecords(0), NumAbbreviatedRecords(0) {}
290};
291
292static std::map<unsigned, PerBlockIDStats> BlockIDStats;
293
294
295
Chris Lattner32de6332007-04-29 08:31:14 +0000296/// Error - All bitcode analysis errors go through this function, making this a
297/// good place to breakpoint if debugging.
298static bool Error(const std::string &Err) {
Dan Gohman65f57c22009-07-15 16:35:29 +0000299 errs() << Err << "\n";
Chris Lattner32de6332007-04-29 08:31:14 +0000300 return true;
301}
302
303/// ParseBlock - Read a block, updating statistics, etc.
Chris Lattner962dde32009-04-26 20:59:02 +0000304static bool ParseBlock(BitstreamCursor &Stream, unsigned IndentLevel) {
Chris Lattner1772b122007-05-05 00:17:42 +0000305 std::string Indent(IndentLevel*2, ' ');
Chris Lattner4238d472007-04-29 20:00:02 +0000306 uint64_t BlockBitStart = Stream.GetCurrentBitNo();
Chris Lattner32de6332007-04-29 08:31:14 +0000307 unsigned BlockID = Stream.ReadSubBlockID();
Chris Lattnerb30c9252007-04-29 21:48:19 +0000308
Chris Lattner4238d472007-04-29 20:00:02 +0000309 // Get the statistics for this BlockID.
310 PerBlockIDStats &BlockStats = BlockIDStats[BlockID];
311
312 BlockStats.NumInstances++;
Chris Lattner32de6332007-04-29 08:31:14 +0000313
Chris Lattner1772b122007-05-05 00:17:42 +0000314 // BLOCKINFO is a special part of the stream.
315 if (BlockID == bitc::BLOCKINFO_BLOCK_ID) {
Dan Gohman65f57c22009-07-15 16:35:29 +0000316 if (Dump) errs() << Indent << "<BLOCKINFO_BLOCK/>\n";
Chris Lattner1772b122007-05-05 00:17:42 +0000317 if (Stream.ReadBlockInfoBlock())
318 return Error("Malformed BlockInfoBlock");
319 uint64_t BlockBitEnd = Stream.GetCurrentBitNo();
320 BlockStats.NumBits += BlockBitEnd-BlockBitStart;
321 return false;
322 }
323
Chris Lattnerb30c9252007-04-29 21:48:19 +0000324 unsigned NumWords = 0;
Chris Lattner1772b122007-05-05 00:17:42 +0000325 if (Stream.EnterSubBlock(BlockID, &NumWords))
Chris Lattner32de6332007-04-29 08:31:14 +0000326 return Error("Malformed block record");
327
Chris Lattnerb30c9252007-04-29 21:48:19 +0000328 const char *BlockName = 0;
329 if (Dump) {
Dan Gohman65f57c22009-07-15 16:35:29 +0000330 errs() << Indent << "<";
Chris Lattnerf9a3ec82009-04-26 22:21:57 +0000331 if ((BlockName = GetBlockName(BlockID, *Stream.getBitStreamReader())))
Dan Gohman65f57c22009-07-15 16:35:29 +0000332 errs() << BlockName;
Chris Lattnerb30c9252007-04-29 21:48:19 +0000333 else
Dan Gohman65f57c22009-07-15 16:35:29 +0000334 errs() << "UnknownBlock" << BlockID;
Chris Lattnerb30c9252007-04-29 21:48:19 +0000335
336 if (NonSymbolic && BlockName)
Dan Gohman65f57c22009-07-15 16:35:29 +0000337 errs() << " BlockID=" << BlockID;
Chris Lattnerb30c9252007-04-29 21:48:19 +0000338
Dan Gohman65f57c22009-07-15 16:35:29 +0000339 errs() << " NumWords=" << NumWords
340 << " BlockCodeSize=" << Stream.GetAbbrevIDWidth() << ">\n";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000341 }
342
Chris Lattner32de6332007-04-29 08:31:14 +0000343 SmallVector<uint64_t, 64> Record;
344
345 // Read all the records for this block.
346 while (1) {
347 if (Stream.AtEndOfStream())
348 return Error("Premature end of bitstream");
349
Chris Lattnerc167cac2009-04-27 18:15:27 +0000350 uint64_t RecordStartBit = Stream.GetCurrentBitNo();
351
Chris Lattner32de6332007-04-29 08:31:14 +0000352 // Read the code for this record.
353 unsigned AbbrevID = Stream.ReadCode();
354 switch (AbbrevID) {
Chris Lattner4238d472007-04-29 20:00:02 +0000355 case bitc::END_BLOCK: {
Chris Lattner32de6332007-04-29 08:31:14 +0000356 if (Stream.ReadBlockEnd())
357 return Error("Error at end of block");
Chris Lattner4238d472007-04-29 20:00:02 +0000358 uint64_t BlockBitEnd = Stream.GetCurrentBitNo();
359 BlockStats.NumBits += BlockBitEnd-BlockBitStart;
Chris Lattnerb30c9252007-04-29 21:48:19 +0000360 if (Dump) {
Dan Gohman65f57c22009-07-15 16:35:29 +0000361 errs() << Indent << "</";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000362 if (BlockName)
Dan Gohman65f57c22009-07-15 16:35:29 +0000363 errs() << BlockName << ">\n";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000364 else
Dan Gohman65f57c22009-07-15 16:35:29 +0000365 errs() << "UnknownBlock" << BlockID << ">\n";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000366 }
Chris Lattner32de6332007-04-29 08:31:14 +0000367 return false;
Chris Lattner4238d472007-04-29 20:00:02 +0000368 }
Chris Lattner44b0f102007-05-05 01:29:31 +0000369 case bitc::ENTER_SUBBLOCK: {
370 uint64_t SubBlockBitStart = Stream.GetCurrentBitNo();
Chris Lattnerb30c9252007-04-29 21:48:19 +0000371 if (ParseBlock(Stream, IndentLevel+1))
Chris Lattner32de6332007-04-29 08:31:14 +0000372 return true;
Chris Lattner4238d472007-04-29 20:00:02 +0000373 ++BlockStats.NumSubBlocks;
Chris Lattner44b0f102007-05-05 01:29:31 +0000374 uint64_t SubBlockBitEnd = Stream.GetCurrentBitNo();
375
376 // Don't include subblock sizes in the size of this block.
377 BlockBitStart += SubBlockBitEnd-SubBlockBitStart;
Chris Lattner32de6332007-04-29 08:31:14 +0000378 break;
Chris Lattner44b0f102007-05-05 01:29:31 +0000379 }
Chris Lattner32de6332007-04-29 08:31:14 +0000380 case bitc::DEFINE_ABBREV:
381 Stream.ReadAbbrevRecord();
Chris Lattner4238d472007-04-29 20:00:02 +0000382 ++BlockStats.NumAbbrevs;
Chris Lattner32de6332007-04-29 08:31:14 +0000383 break;
384 default:
385 Record.clear();
Chris Lattner3f75d312009-04-06 22:44:40 +0000386
387 ++BlockStats.NumRecords;
Chris Lattnerae7dd802009-04-07 02:56:46 +0000388 if (AbbrevID != bitc::UNABBREV_RECORD)
Chris Lattner3f75d312009-04-06 22:44:40 +0000389 ++BlockStats.NumAbbreviatedRecords;
Chris Lattner3f75d312009-04-06 22:44:40 +0000390
Chris Lattner3f75d312009-04-06 22:44:40 +0000391 const char *BlobStart = 0;
392 unsigned BlobLen = 0;
Chris Lattnerae7dd802009-04-07 02:56:46 +0000393 unsigned Code = Stream.ReadRecord(AbbrevID, Record, BlobStart, BlobLen);
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000394
Chris Lattnerc167cac2009-04-27 18:15:27 +0000395
396
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000397 // Increment the # occurrences of this code.
398 if (BlockStats.CodeFreq.size() <= Code)
399 BlockStats.CodeFreq.resize(Code+1);
Chris Lattner24437472009-04-27 17:59:34 +0000400 BlockStats.CodeFreq[Code].NumInstances++;
Chris Lattnerc167cac2009-04-27 18:15:27 +0000401 BlockStats.CodeFreq[Code].TotalBits +=
402 Stream.GetCurrentBitNo()-RecordStartBit;
403 if (AbbrevID != bitc::UNABBREV_RECORD)
404 BlockStats.CodeFreq[Code].NumAbbrev++;
405
Chris Lattnerb30c9252007-04-29 21:48:19 +0000406 if (Dump) {
Dan Gohman65f57c22009-07-15 16:35:29 +0000407 errs() << Indent << " <";
Chris Lattnerf9a3ec82009-04-26 22:21:57 +0000408 if (const char *CodeName =
409 GetCodeName(Code, BlockID, *Stream.getBitStreamReader()))
Dan Gohman65f57c22009-07-15 16:35:29 +0000410 errs() << CodeName;
Chris Lattnerb30c9252007-04-29 21:48:19 +0000411 else
Dan Gohman65f57c22009-07-15 16:35:29 +0000412 errs() << "UnknownCode" << Code;
Chris Lattnerf9a3ec82009-04-26 22:21:57 +0000413 if (NonSymbolic &&
414 GetCodeName(Code, BlockID, *Stream.getBitStreamReader()))
Dan Gohman65f57c22009-07-15 16:35:29 +0000415 errs() << " codeid=" << Code;
Chris Lattnerb30c9252007-04-29 21:48:19 +0000416 if (AbbrevID != bitc::UNABBREV_RECORD)
Dan Gohman65f57c22009-07-15 16:35:29 +0000417 errs() << " abbrevid=" << AbbrevID;
Chris Lattnerb30c9252007-04-29 21:48:19 +0000418
419 for (unsigned i = 0, e = Record.size(); i != e; ++i)
Dan Gohman65f57c22009-07-15 16:35:29 +0000420 errs() << " op" << i << "=" << (int64_t)Record[i];
Chris Lattnerb30c9252007-04-29 21:48:19 +0000421
Dan Gohman65f57c22009-07-15 16:35:29 +0000422 errs() << "/>";
Chris Lattnerae7dd802009-04-07 02:56:46 +0000423
424 if (BlobStart) {
Dan Gohman65f57c22009-07-15 16:35:29 +0000425 errs() << " blob data = ";
Chris Lattnerae7dd802009-04-07 02:56:46 +0000426 bool BlobIsPrintable = true;
427 for (unsigned i = 0; i != BlobLen; ++i)
428 if (!isprint(BlobStart[i])) {
429 BlobIsPrintable = false;
430 break;
431 }
432
433 if (BlobIsPrintable)
Dan Gohman65f57c22009-07-15 16:35:29 +0000434 errs() << "'" << std::string(BlobStart, BlobStart+BlobLen) <<"'";
Chris Lattnerae7dd802009-04-07 02:56:46 +0000435 else
Dan Gohman65f57c22009-07-15 16:35:29 +0000436 errs() << "unprintable, " << BlobLen << " bytes.";
Chris Lattnerae7dd802009-04-07 02:56:46 +0000437 }
438
Dan Gohman65f57c22009-07-15 16:35:29 +0000439 errs() << "\n";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000440 }
441
Chris Lattner32de6332007-04-29 08:31:14 +0000442 break;
443 }
444 }
445}
446
Chris Lattner4238d472007-04-29 20:00:02 +0000447static void PrintSize(double Bits) {
Chris Lattner24437472009-04-27 17:59:34 +0000448 fprintf(stderr, "%.2f/%.2fB/%lluW", Bits, Bits/8,(unsigned long long)Bits/32);
449}
450static void PrintSize(uint64_t Bits) {
451 fprintf(stderr, "%llub/%.2fB/%lluW", (unsigned long long)Bits,
452 (double)Bits/8, (unsigned long long)Bits/32);
Chris Lattner4238d472007-04-29 20:00:02 +0000453}
454
455
Chris Lattner45e0f892007-04-29 08:12:22 +0000456/// AnalyzeBitcode - Analyze the bitcode file specified by InputFilename.
457static int AnalyzeBitcode() {
458 // Read the input file.
Chris Lattnere2a466b2009-04-06 20:54:32 +0000459 MemoryBuffer *MemBuf = MemoryBuffer::getFileOrSTDIN(InputFilename.c_str());
Chris Lattner45e0f892007-04-29 08:12:22 +0000460
Chris Lattnere2a466b2009-04-06 20:54:32 +0000461 if (MemBuf == 0)
Chris Lattner32de6332007-04-29 08:31:14 +0000462 return Error("Error reading '" + InputFilename + "'.");
Chris Lattner45e0f892007-04-29 08:12:22 +0000463
Chris Lattnere2a466b2009-04-06 20:54:32 +0000464 if (MemBuf->getBufferSize() & 3)
Chris Lattner32de6332007-04-29 08:31:14 +0000465 return Error("Bitcode stream should be a multiple of 4 bytes in length");
Chris Lattner45e0f892007-04-29 08:12:22 +0000466
Chris Lattnere2a466b2009-04-06 20:54:32 +0000467 unsigned char *BufPtr = (unsigned char *)MemBuf->getBufferStart();
468 unsigned char *EndBufPtr = BufPtr+MemBuf->getBufferSize();
469
470 // If we have a wrapper header, parse it and ignore the non-bc file contents.
471 // The magic number is 0x0B17C0DE stored in little endian.
472 if (isBitcodeWrapper(BufPtr, EndBufPtr))
473 if (SkipBitcodeWrapperHeader(BufPtr, EndBufPtr))
474 return Error("Invalid bitcode wrapper header");
475
Chris Lattner962dde32009-04-26 20:59:02 +0000476 BitstreamReader StreamFile(BufPtr, EndBufPtr);
477 BitstreamCursor Stream(StreamFile);
Chris Lattner0370cc62009-04-27 20:04:08 +0000478 StreamFile.CollectBlockInfoNames();
Chris Lattner45e0f892007-04-29 08:12:22 +0000479
480 // Read the stream signature.
481 char Signature[6];
482 Signature[0] = Stream.Read(8);
483 Signature[1] = Stream.Read(8);
484 Signature[2] = Stream.Read(4);
485 Signature[3] = Stream.Read(4);
486 Signature[4] = Stream.Read(4);
487 Signature[5] = Stream.Read(4);
488
Chris Lattner32de6332007-04-29 08:31:14 +0000489 // Autodetect the file contents, if it is one we know.
Chris Lattner45e0f892007-04-29 08:12:22 +0000490 CurStreamType = UnknownBitstream;
491 if (Signature[0] == 'B' && Signature[1] == 'C' &&
492 Signature[2] == 0x0 && Signature[3] == 0xC &&
493 Signature[4] == 0xE && Signature[5] == 0xD)
494 CurStreamType = LLVMIRBitstream;
495
Chris Lattner4238d472007-04-29 20:00:02 +0000496 unsigned NumTopBlocks = 0;
497
Chris Lattner32de6332007-04-29 08:31:14 +0000498 // Parse the top-level structure. We only allow blocks at the top-level.
499 while (!Stream.AtEndOfStream()) {
500 unsigned Code = Stream.ReadCode();
501 if (Code != bitc::ENTER_SUBBLOCK)
502 return Error("Invalid record at top-level");
503
Chris Lattnerb30c9252007-04-29 21:48:19 +0000504 if (ParseBlock(Stream, 0))
Chris Lattner32de6332007-04-29 08:31:14 +0000505 return true;
Chris Lattner4238d472007-04-29 20:00:02 +0000506 ++NumTopBlocks;
Chris Lattner32de6332007-04-29 08:31:14 +0000507 }
508
Dan Gohman65f57c22009-07-15 16:35:29 +0000509 if (Dump) errs() << "\n\n";
Chris Lattnerb30c9252007-04-29 21:48:19 +0000510
Chris Lattnere2a466b2009-04-06 20:54:32 +0000511 uint64_t BufferSizeBits = (EndBufPtr-BufPtr)*CHAR_BIT;
Chris Lattner32de6332007-04-29 08:31:14 +0000512 // Print a summary of the read file.
Dan Gohman65f57c22009-07-15 16:35:29 +0000513 errs() << "Summary of " << InputFilename << ":\n";
514 errs() << " Total size: ";
Chris Lattner8f926682007-05-01 02:43:46 +0000515 PrintSize(BufferSizeBits);
Dan Gohman65f57c22009-07-15 16:35:29 +0000516 errs() << "\n";
517 errs() << " Stream type: ";
Chris Lattner45e0f892007-04-29 08:12:22 +0000518 switch (CurStreamType) {
519 default: assert(0 && "Unknown bitstream type");
Dan Gohman65f57c22009-07-15 16:35:29 +0000520 case UnknownBitstream: errs() << "unknown\n"; break;
521 case LLVMIRBitstream: errs() << "LLVM IR\n"; break;
Chris Lattner45e0f892007-04-29 08:12:22 +0000522 }
Dan Gohman65f57c22009-07-15 16:35:29 +0000523 errs() << " # Toplevel Blocks: " << NumTopBlocks << "\n";
524 errs() << "\n";
Chris Lattner4238d472007-04-29 20:00:02 +0000525
526 // Emit per-block stats.
Dan Gohman65f57c22009-07-15 16:35:29 +0000527 errs() << "Per-block Summary:\n";
Chris Lattner4238d472007-04-29 20:00:02 +0000528 for (std::map<unsigned, PerBlockIDStats>::iterator I = BlockIDStats.begin(),
529 E = BlockIDStats.end(); I != E; ++I) {
Dan Gohman65f57c22009-07-15 16:35:29 +0000530 errs() << " Block ID #" << I->first;
Chris Lattnerf9a3ec82009-04-26 22:21:57 +0000531 if (const char *BlockName = GetBlockName(I->first, StreamFile))
Dan Gohman65f57c22009-07-15 16:35:29 +0000532 errs() << " (" << BlockName << ")";
533 errs() << ":\n";
Chris Lattner4238d472007-04-29 20:00:02 +0000534
535 const PerBlockIDStats &Stats = I->second;
Dan Gohman65f57c22009-07-15 16:35:29 +0000536 errs() << " Num Instances: " << Stats.NumInstances << "\n";
537 errs() << " Total Size: ";
Chris Lattner4238d472007-04-29 20:00:02 +0000538 PrintSize(Stats.NumBits);
Dan Gohman65f57c22009-07-15 16:35:29 +0000539 errs() << "\n";
540 errs() << " % of file: "
541 << Stats.NumBits/(double)BufferSizeBits*100 << "\n";
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000542 if (Stats.NumInstances > 1) {
Dan Gohman65f57c22009-07-15 16:35:29 +0000543 errs() << " Average Size: ";
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000544 PrintSize(Stats.NumBits/(double)Stats.NumInstances);
Dan Gohman65f57c22009-07-15 16:35:29 +0000545 errs() << "\n";
546 errs() << " Tot/Avg SubBlocks: " << Stats.NumSubBlocks << "/"
547 << Stats.NumSubBlocks/(double)Stats.NumInstances << "\n";
548 errs() << " Tot/Avg Abbrevs: " << Stats.NumAbbrevs << "/"
549 << Stats.NumAbbrevs/(double)Stats.NumInstances << "\n";
550 errs() << " Tot/Avg Records: " << Stats.NumRecords << "/"
551 << Stats.NumRecords/(double)Stats.NumInstances << "\n";
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000552 } else {
Dan Gohman65f57c22009-07-15 16:35:29 +0000553 errs() << " Num SubBlocks: " << Stats.NumSubBlocks << "\n";
554 errs() << " Num Abbrevs: " << Stats.NumAbbrevs << "\n";
555 errs() << " Num Records: " << Stats.NumRecords << "\n";
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000556 }
Chris Lattner1772b122007-05-05 00:17:42 +0000557 if (Stats.NumRecords)
Dan Gohman65f57c22009-07-15 16:35:29 +0000558 errs() << " % Abbrev Recs: " << (Stats.NumAbbreviatedRecords/
559 (double)Stats.NumRecords)*100 << "\n";
560 errs() << "\n";
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000561
562 // Print a histogram of the codes we see.
563 if (!NoHistogram && !Stats.CodeFreq.empty()) {
564 std::vector<std::pair<unsigned, unsigned> > FreqPairs; // <freq,code>
565 for (unsigned i = 0, e = Stats.CodeFreq.size(); i != e; ++i)
Chris Lattner24437472009-04-27 17:59:34 +0000566 if (unsigned Freq = Stats.CodeFreq[i].NumInstances)
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000567 FreqPairs.push_back(std::make_pair(Freq, i));
568 std::stable_sort(FreqPairs.begin(), FreqPairs.end());
569 std::reverse(FreqPairs.begin(), FreqPairs.end());
570
Dan Gohman65f57c22009-07-15 16:35:29 +0000571 errs() << "\tRecord Histogram:\n";
Chris Lattnerc167cac2009-04-27 18:15:27 +0000572 fprintf(stderr, "\t\t Count # Bits %% Abv Record Kind\n");
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000573 for (unsigned i = 0, e = FreqPairs.size(); i != e; ++i) {
Chris Lattnerc167cac2009-04-27 18:15:27 +0000574 const PerRecordStats &RecStats = Stats.CodeFreq[FreqPairs[i].second];
575
576 fprintf(stderr, "\t\t%7d %9llu ", RecStats.NumInstances,
Dan Gohmanfbccdef2009-05-01 16:33:33 +0000577 (unsigned long long)RecStats.TotalBits);
Chris Lattnerc167cac2009-04-27 18:15:27 +0000578
579 if (RecStats.NumAbbrev)
580 fprintf(stderr, "%7.2f ",
581 (double)RecStats.NumAbbrev/RecStats.NumInstances*100);
582 else
583 fprintf(stderr, " ");
Chris Lattner24437472009-04-27 17:59:34 +0000584
Chris Lattnerf9a3ec82009-04-26 22:21:57 +0000585 if (const char *CodeName =
586 GetCodeName(FreqPairs[i].second, I->first, StreamFile))
Chris Lattner24437472009-04-27 17:59:34 +0000587 fprintf(stderr, "%s\n", CodeName);
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000588 else
Chris Lattner24437472009-04-27 17:59:34 +0000589 fprintf(stderr, "UnknownCode%d\n", FreqPairs[i].second);
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000590 }
Dan Gohman65f57c22009-07-15 16:35:29 +0000591 errs() << "\n";
Chris Lattnerb1e85b52007-05-05 01:46:49 +0000592
593 }
Chris Lattner4238d472007-04-29 20:00:02 +0000594 }
Chris Lattner45e0f892007-04-29 08:12:22 +0000595 return 0;
596}
Reid Spencerdac69c82004-06-07 17:53:43 +0000597
Chris Lattner32de6332007-04-29 08:31:14 +0000598
Chris Lattnerc30598b2006-12-06 01:18:01 +0000599int main(int argc, char **argv) {
Chris Lattnercc14d252009-03-06 05:34:10 +0000600 // Print a stack trace if we signal out.
Chris Lattner45e0f892007-04-29 08:12:22 +0000601 sys::PrintStackTraceOnErrorSignal();
Chris Lattnercc14d252009-03-06 05:34:10 +0000602 PrettyStackTraceProgram X(argc, argv);
603 llvm_shutdown_obj Y; // Call llvm_shutdown() on exit.
604 cl::ParseCommandLineOptions(argc, argv, "llvm-bcanalyzer file analyzer\n");
Chris Lattner45e0f892007-04-29 08:12:22 +0000605
Chris Lattner44dadff2007-05-06 09:29:57 +0000606 return AnalyzeBitcode();
Reid Spencerdac69c82004-06-07 17:53:43 +0000607}