lib/ProfileData/SampleProfWriter.cpp

   1 //===- SampleProfWriter.cpp - Write LLVM sample profile data --------------===//
   2 //
   3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
   4 // See https://llvm.org/LICENSE.txt for license information.
   5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
   6 //
   7 //===----------------------------------------------------------------------===//
   8 //
   9 // This file implements the class that writes LLVM sample profiles. It
  10 // supports two file formats: text and binary. The textual representation
  11 // is useful for debugging and testing purposes. The binary representation
  12 // is more compact, resulting in smaller file sizes. However, they can
  13 // both be used interchangeably.
  14 //
  15 // See lib/ProfileData/SampleProfReader.cpp for documentation on each of the
  16 // supported formats.
  17 //
  18 //===----------------------------------------------------------------------===//
  19
  20 #include "llvm/ProfileData/SampleProfWriter.h"
  21 #include "llvm/ADT/StringRef.h"
  22 #include "llvm/ProfileData/ProfileCommon.h"
  23 #include "llvm/ProfileData/SampleProf.h"
  24 #include "llvm/Support/Endian.h"
  25 #include "llvm/Support/EndianStream.h"
  26 #include "llvm/Support/ErrorOr.h"
  27 #include "llvm/Support/FileSystem.h"
  28 #include "llvm/Support/LEB128.h"
  29 #include "llvm/Support/MD5.h"
  30 #include "llvm/Support/raw_ostream.h"
  31 #include <algorithm>
  32 #include <cstdint>
  33 #include <memory>
  34 #include <set>
  35 #include <system_error>
  36 #include <utility>
  37 #include <vector>
  38
  39 using namespace llvm;
  40 using namespace sampleprof;
  41
  42 std::error_code
  43 SampleProfileWriter::write(const StringMap<FunctionSamples> &ProfileMap) {
  44   if (std::error_code EC = writeHeader(ProfileMap))
  45     return EC;
  46
  47   // Sort the ProfileMap by total samples.
  48   typedef std::pair<StringRef, const FunctionSamples *> NameFunctionSamples;
  49   std::vector<NameFunctionSamples> V;
  50   for (const auto &I : ProfileMap)
  51     V.push_back(std::make_pair(I.getKey(), &I.second));
  52
  53   llvm::stable_sort(
  54       V, [](const NameFunctionSamples &A, const NameFunctionSamples &B) {
  55         if (A.second->getTotalSamples() == B.second->getTotalSamples())
  56           return A.first > B.first;
  57         return A.second->getTotalSamples() > B.second->getTotalSamples();
  58       });
  59
  60   for (const auto &I : V) {
  61     if (std::error_code EC = write(*I.second))
  62       return EC;
  63   }
  64   return sampleprof_error::success;
  65 }
  66
  67 std::error_code SampleProfileWriterCompactBinary::write(
  68     const StringMap<FunctionSamples> &ProfileMap) {
  69   if (std::error_code EC = SampleProfileWriter::write(ProfileMap))
  70     return EC;
  71   if (std::error_code EC = writeFuncOffsetTable())
  72     return EC;
  73   return sampleprof_error::success;
  74 }
  75
  76 /// Write samples to a text file.
  77 ///
  78 /// Note: it may be tempting to implement this in terms of
  79 /// FunctionSamples::print().  Please don't.  The dump functionality is intended
  80 /// for debugging and has no specified form.
  81 ///
  82 /// The format used here is more structured and deliberate because
  83 /// it needs to be parsed by the SampleProfileReaderText class.
  84 std::error_code SampleProfileWriterText::write(const FunctionSamples &S) {
  85   auto &OS = *OutputStream;
  86   OS << S.getName() << ":" << S.getTotalSamples();
  87   if (Indent == 0)
  88     OS << ":" << S.getHeadSamples();
  89   OS << "\n";
  90
  91   SampleSorter<LineLocation, SampleRecord> SortedSamples(S.getBodySamples());
  92   for (const auto &I : SortedSamples.get()) {
  93     LineLocation Loc = I->first;
  94     const SampleRecord &Sample = I->second;
  95     OS.indent(Indent + 1);
  96     if (Loc.Discriminator == 0)
  97       OS << Loc.LineOffset << ": ";
  98     else
  99       OS << Loc.LineOffset << "." << Loc.Discriminator << ": ";
 100
 101     OS << Sample.getSamples();
 102
 103     for (const auto &J : Sample.getCallTargets())
 104       OS << " " << J.first() << ":" << J.second;
 105     OS << "\n";
 106   }
 107
 108   SampleSorter<LineLocation, FunctionSamplesMap> SortedCallsiteSamples(
 109       S.getCallsiteSamples());
 110   Indent += 1;
 111   for (const auto &I : SortedCallsiteSamples.get())
 112     for (const auto &FS : I->second) {
 113       LineLocation Loc = I->first;
 114       const FunctionSamples &CalleeSamples = FS.second;
 115       OS.indent(Indent);
 116       if (Loc.Discriminator == 0)
 117         OS << Loc.LineOffset << ": ";
 118       else
 119         OS << Loc.LineOffset << "." << Loc.Discriminator << ": ";
 120       if (std::error_code EC = write(CalleeSamples))
 121         return EC;
 122     }
 123   Indent -= 1;
 124
 125   return sampleprof_error::success;
 126 }
 127
 128 std::error_code SampleProfileWriterBinary::writeNameIdx(StringRef FName) {
 129   const auto &ret = NameTable.find(FName);
 130   if (ret == NameTable.end())
 131     return sampleprof_error::truncated_name_table;
 132   encodeULEB128(ret->second, *OutputStream);
 133   return sampleprof_error::success;
 134 }
 135
 136 void SampleProfileWriterBinary::addName(StringRef FName) {
 137   NameTable.insert(std::make_pair(FName, 0));
 138 }
 139
 140 void SampleProfileWriterBinary::addNames(const FunctionSamples &S) {
 141   // Add all the names in indirect call targets.
 142   for (const auto &I : S.getBodySamples()) {
 143     const SampleRecord &Sample = I.second;
 144     for (const auto &J : Sample.getCallTargets())
 145       addName(J.first());
 146   }
 147
 148   // Recursively add all the names for inlined callsites.
 149   for (const auto &J : S.getCallsiteSamples())
 150     for (const auto &FS : J.second) {
 151       const FunctionSamples &CalleeSamples = FS.second;
 152       addName(CalleeSamples.getName());
 153       addNames(CalleeSamples);
 154     }
 155 }
 156
 157 void SampleProfileWriterBinary::stablizeNameTable(std::set<StringRef> &V) {
 158   // Sort the names to make NameTable deterministic.
 159   for (const auto &I : NameTable)
 160     V.insert(I.first);
 161   int i = 0;
 162   for (const StringRef &N : V)
 163     NameTable[N] = i++;
 164 }
 165
 166 std::error_code SampleProfileWriterRawBinary::writeNameTable() {
 167   auto &OS = *OutputStream;
 168   std::set<StringRef> V;
 169   stablizeNameTable(V);
 170
 171   // Write out the name table.
 172   encodeULEB128(NameTable.size(), OS);
 173   for (auto N : V) {
 174     OS << N;
 175     encodeULEB128(0, OS);
 176   }
 177   return sampleprof_error::success;
 178 }
 179
 180 std::error_code SampleProfileWriterCompactBinary::writeFuncOffsetTable() {
 181   auto &OS = *OutputStream;
 182
 183   // Fill the slot remembered by TableOffset with the offset of FuncOffsetTable.
 184   auto &OFS = static_cast<raw_fd_ostream &>(OS);
 185   uint64_t FuncOffsetTableStart = OS.tell();
 186   if (OFS.seek(TableOffset) == (uint64_t)-1)
 187     return sampleprof_error::ostream_seek_unsupported;
 188   support::endian::Writer Writer(*OutputStream, support::little);
 189   Writer.write(FuncOffsetTableStart);
 190   if (OFS.seek(FuncOffsetTableStart) == (uint64_t)-1)
 191     return sampleprof_error::ostream_seek_unsupported;
 192
 193   // Write out the table size.
 194   encodeULEB128(FuncOffsetTable.size(), OS);
 195
 196   // Write out FuncOffsetTable.
 197   for (auto entry : FuncOffsetTable) {
 198     writeNameIdx(entry.first);
 199     encodeULEB128(entry.second, OS);
 200   }
 201   return sampleprof_error::success;
 202 }
 203
 204 std::error_code SampleProfileWriterCompactBinary::writeNameTable() {
 205   auto &OS = *OutputStream;
 206   std::set<StringRef> V;
 207   stablizeNameTable(V);
 208
 209   // Write out the name table.
 210   encodeULEB128(NameTable.size(), OS);
 211   for (auto N : V) {
 212     encodeULEB128(MD5Hash(N), OS);
 213   }
 214   return sampleprof_error::success;
 215 }
 216
 217 std::error_code SampleProfileWriterRawBinary::writeMagicIdent() {
 218   auto &OS = *OutputStream;
 219   // Write file magic identifier.
 220   encodeULEB128(SPMagic(), OS);
 221   encodeULEB128(SPVersion(), OS);
 222   return sampleprof_error::success;
 223 }
 224
 225 std::error_code SampleProfileWriterCompactBinary::writeMagicIdent() {
 226   auto &OS = *OutputStream;
 227   // Write file magic identifier.
 228   encodeULEB128(SPMagic(SPF_Compact_Binary), OS);
 229   encodeULEB128(SPVersion(), OS);
 230   return sampleprof_error::success;
 231 }
 232
 233 std::error_code SampleProfileWriterBinary::writeHeader(
 234     const StringMap<FunctionSamples> &ProfileMap) {
 235   writeMagicIdent();
 236
 237   computeSummary(ProfileMap);
 238   if (auto EC = writeSummary())
 239     return EC;
 240
 241   // Generate the name table for all the functions referenced in the profile.
 242   for (const auto &I : ProfileMap) {
 243     addName(I.first());
 244     addNames(I.second);
 245   }
 246
 247   writeNameTable();
 248   return sampleprof_error::success;
 249 }
 250
 251 std::error_code SampleProfileWriterCompactBinary::writeHeader(
 252     const StringMap<FunctionSamples> &ProfileMap) {
 253   support::endian::Writer Writer(*OutputStream, support::little);
 254   if (auto EC = SampleProfileWriterBinary::writeHeader(ProfileMap))
 255     return EC;
 256
 257   // Reserve a slot for the offset of function offset table. The slot will
 258   // be populated with the offset of FuncOffsetTable later.
 259   TableOffset = OutputStream->tell();
 260   Writer.write(static_cast<uint64_t>(-2));
 261   return sampleprof_error::success;
 262 }
 263
 264 std::error_code SampleProfileWriterBinary::writeSummary() {
 265   auto &OS = *OutputStream;
 266   encodeULEB128(Summary->getTotalCount(), OS);
 267   encodeULEB128(Summary->getMaxCount(), OS);
 268   encodeULEB128(Summary->getMaxFunctionCount(), OS);
 269   encodeULEB128(Summary->getNumCounts(), OS);
 270   encodeULEB128(Summary->getNumFunctions(), OS);
 271   std::vector<ProfileSummaryEntry> &Entries = Summary->getDetailedSummary();
 272   encodeULEB128(Entries.size(), OS);
 273   for (auto Entry : Entries) {
 274     encodeULEB128(Entry.Cutoff, OS);
 275     encodeULEB128(Entry.MinCount, OS);
 276     encodeULEB128(Entry.NumCounts, OS);
 277   }
 278   return sampleprof_error::success;
 279 }
 280 std::error_code SampleProfileWriterBinary::writeBody(const FunctionSamples &S) {
 281   auto &OS = *OutputStream;
 282
 283   if (std::error_code EC = writeNameIdx(S.getName()))
 284     return EC;
 285
 286   encodeULEB128(S.getTotalSamples(), OS);
 287
 288   // Emit all the body samples.
 289   encodeULEB128(S.getBodySamples().size(), OS);
 290   for (const auto &I : S.getBodySamples()) {
 291     LineLocation Loc = I.first;
 292     const SampleRecord &Sample = I.second;
 293     encodeULEB128(Loc.LineOffset, OS);
 294     encodeULEB128(Loc.Discriminator, OS);
 295     encodeULEB128(Sample.getSamples(), OS);
 296     encodeULEB128(Sample.getCallTargets().size(), OS);
 297     for (const auto &J : Sample.getCallTargets()) {
 298       StringRef Callee = J.first();
 299       uint64_t CalleeSamples = J.second;
 300       if (std::error_code EC = writeNameIdx(Callee))
 301         return EC;
 302       encodeULEB128(CalleeSamples, OS);
 303     }
 304   }
 305
 306   // Recursively emit all the callsite samples.
 307   uint64_t NumCallsites = 0;
 308   for (const auto &J : S.getCallsiteSamples())
 309     NumCallsites += J.second.size();
 310   encodeULEB128(NumCallsites, OS);
 311   for (const auto &J : S.getCallsiteSamples())
 312     for (const auto &FS : J.second) {
 313       LineLocation Loc = J.first;
 314       const FunctionSamples &CalleeSamples = FS.second;
 315       encodeULEB128(Loc.LineOffset, OS);
 316       encodeULEB128(Loc.Discriminator, OS);
 317       if (std::error_code EC = writeBody(CalleeSamples))
 318         return EC;
 319     }
 320
 321   return sampleprof_error::success;
 322 }
 323
 324 /// Write samples of a top-level function to a binary file.
 325 ///
 326 /// \returns true if the samples were written successfully, false otherwise.
 327 std::error_code SampleProfileWriterBinary::write(const FunctionSamples &S) {
 328   encodeULEB128(S.getHeadSamples(), *OutputStream);
 329   return writeBody(S);
 330 }
 331
 332 std::error_code
 333 SampleProfileWriterCompactBinary::write(const FunctionSamples &S) {
 334   uint64_t Offset = OutputStream->tell();
 335   StringRef Name = S.getName();
 336   FuncOffsetTable[Name] = Offset;
 337   encodeULEB128(S.getHeadSamples(), *OutputStream);
 338   return writeBody(S);
 339 }
 340
 341 /// Create a sample profile file writer based on the specified format.
 342 ///
 343 /// \param Filename The file to create.
 344 ///
 345 /// \param Format Encoding format for the profile file.
 346 ///
 347 /// \returns an error code indicating the status of the created writer.
 348 ErrorOr<std::unique_ptr<SampleProfileWriter>>
 349 SampleProfileWriter::create(StringRef Filename, SampleProfileFormat Format) {
 350   std::error_code EC;
 351   std::unique_ptr<raw_ostream> OS;
 352   if (Format == SPF_Binary || Format == SPF_Compact_Binary)
 353     OS.reset(new raw_fd_ostream(Filename, EC, sys::fs::F_None));
 354   else
 355     OS.reset(new raw_fd_ostream(Filename, EC, sys::fs::F_Text));
 356   if (EC)
 357     return EC;
 358
 359   return create(OS, Format);
 360 }
 361
 362 /// Create a sample profile stream writer based on the specified format.
 363 ///
 364 /// \param OS The output stream to store the profile data to.
 365 ///
 366 /// \param Format Encoding format for the profile file.
 367 ///
 368 /// \returns an error code indicating the status of the created writer.
 369 ErrorOr<std::unique_ptr<SampleProfileWriter>>
 370 SampleProfileWriter::create(std::unique_ptr<raw_ostream> &OS,
 371                             SampleProfileFormat Format) {
 372   std::error_code EC;
 373   std::unique_ptr<SampleProfileWriter> Writer;
 374
 375   if (Format == SPF_Binary)
 376     Writer.reset(new SampleProfileWriterRawBinary(OS));
 377   else if (Format == SPF_Compact_Binary)
 378     Writer.reset(new SampleProfileWriterCompactBinary(OS));
 379   else if (Format == SPF_Text)
 380     Writer.reset(new SampleProfileWriterText(OS));
 381   else if (Format == SPF_GCC)
 382     EC = sampleprof_error::unsupported_writing_format;
 383   else
 384     EC = sampleprof_error::unrecognized_format;
 385
 386   if (EC)
 387     return EC;
 388
 389   return std::move(Writer);
 390 }
 391
 392 void SampleProfileWriter::computeSummary(
 393     const StringMap<FunctionSamples> &ProfileMap) {
 394   SampleProfileSummaryBuilder Builder(ProfileSummaryBuilder::DefaultCutoffs);
 395   for (const auto &I : ProfileMap) {
 396     const FunctionSamples &Profile = I.second;
 397     Builder.addRecord(Profile);
 398   }
 399   Summary = Builder.getSummary();
 400 }