lib/ProfileData/InstrProfWriter.cpp

   1 //===- InstrProfWriter.cpp - Instrumented profiling writer ----------------===//
   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 contains support for writing profiling data for clang's
  10 // instrumentation based PGO and coverage.
  11 //
  12 //===----------------------------------------------------------------------===//
  13
  14 #include "llvm/ProfileData/InstrProfWriter.h"
  15 #include "llvm/ADT/STLExtras.h"
  16 #include "llvm/ADT/StringRef.h"
  17 #include "llvm/IR/ProfileSummary.h"
  18 #include "llvm/ProfileData/InstrProf.h"
  19 #include "llvm/ProfileData/ProfileCommon.h"
  20 #include "llvm/Support/Endian.h"
  21 #include "llvm/Support/EndianStream.h"
  22 #include "llvm/Support/Error.h"
  23 #include "llvm/Support/MemoryBuffer.h"
  24 #include "llvm/Support/OnDiskHashTable.h"
  25 #include "llvm/Support/raw_ostream.h"
  26 #include <algorithm>
  27 #include <cstdint>
  28 #include <memory>
  29 #include <string>
  30 #include <tuple>
  31 #include <utility>
  32 #include <vector>
  33
  34 using namespace llvm;
  35
  36 // A struct to define how the data stream should be patched. For Indexed
  37 // profiling, only uint64_t data type is needed.
  38 struct PatchItem {
  39   uint64_t Pos; // Where to patch.
  40   uint64_t *D;  // Pointer to an array of source data.
  41   int N;        // Number of elements in \c D array.
  42 };
  43
  44 namespace llvm {
  45
  46 // A wrapper class to abstract writer stream with support of bytes
  47 // back patching.
  48 class ProfOStream {
  49 public:
  50   ProfOStream(raw_fd_ostream &FD)
  51       : IsFDOStream(true), OS(FD), LE(FD, support::little) {}
  52   ProfOStream(raw_string_ostream &STR)
  53       : IsFDOStream(false), OS(STR), LE(STR, support::little) {}
  54
  55   uint64_t tell() { return OS.tell(); }
  56   void write(uint64_t V) { LE.write<uint64_t>(V); }
  57
  58   // \c patch can only be called when all data is written and flushed.
  59   // For raw_string_ostream, the patch is done on the target string
  60   // directly and it won't be reflected in the stream's internal buffer.
  61   void patch(PatchItem *P, int NItems) {
  62     using namespace support;
  63
  64     if (IsFDOStream) {
  65       raw_fd_ostream &FDOStream = static_cast<raw_fd_ostream &>(OS);
  66       for (int K = 0; K < NItems; K++) {
  67         FDOStream.seek(P[K].Pos);
  68         for (int I = 0; I < P[K].N; I++)
  69           write(P[K].D[I]);
  70       }
  71     } else {
  72       raw_string_ostream &SOStream = static_cast<raw_string_ostream &>(OS);
  73       std::string &Data = SOStream.str(); // with flush
  74       for (int K = 0; K < NItems; K++) {
  75         for (int I = 0; I < P[K].N; I++) {
  76           uint64_t Bytes = endian::byte_swap<uint64_t, little>(P[K].D[I]);
  77           Data.replace(P[K].Pos + I * sizeof(uint64_t), sizeof(uint64_t),
  78                        (const char *)&Bytes, sizeof(uint64_t));
  79         }
  80       }
  81     }
  82   }
  83
  84   // If \c OS is an instance of \c raw_fd_ostream, this field will be
  85   // true. Otherwise, \c OS will be an raw_string_ostream.
  86   bool IsFDOStream;
  87   raw_ostream &OS;
  88   support::endian::Writer LE;
  89 };
  90
  91 class InstrProfRecordWriterTrait {
  92 public:
  93   using key_type = StringRef;
  94   using key_type_ref = StringRef;
  95
  96   using data_type = const InstrProfWriter::ProfilingData *const;
  97   using data_type_ref = const InstrProfWriter::ProfilingData *const;
  98
  99   using hash_value_type = uint64_t;
 100   using offset_type = uint64_t;
 101
 102   support::endianness ValueProfDataEndianness = support::little;
 103   InstrProfSummaryBuilder *SummaryBuilder;
 104   InstrProfSummaryBuilder *CSSummaryBuilder;
 105
 106   InstrProfRecordWriterTrait() = default;
 107
 108   static hash_value_type ComputeHash(key_type_ref K) {
 109     return IndexedInstrProf::ComputeHash(K);
 110   }
 111
 112   static std::pair<offset_type, offset_type>
 113   EmitKeyDataLength(raw_ostream &Out, key_type_ref K, data_type_ref V) {
 114     using namespace support;
 115
 116     endian::Writer LE(Out, little);
 117
 118     offset_type N = K.size();
 119     LE.write<offset_type>(N);
 120
 121     offset_type M = 0;
 122     for (const auto &ProfileData : *V) {
 123       const InstrProfRecord &ProfRecord = ProfileData.second;
 124       M += sizeof(uint64_t); // The function hash
 125       M += sizeof(uint64_t); // The size of the Counts vector
 126       M += ProfRecord.Counts.size() * sizeof(uint64_t);
 127
 128       // Value data
 129       M += ValueProfData::getSize(ProfileData.second);
 130     }
 131     LE.write<offset_type>(M);
 132
 133     return std::make_pair(N, M);
 134   }
 135
 136   void EmitKey(raw_ostream &Out, key_type_ref K, offset_type N) {
 137     Out.write(K.data(), N);
 138   }
 139
 140   void EmitData(raw_ostream &Out, key_type_ref, data_type_ref V, offset_type) {
 141     using namespace support;
 142
 143     endian::Writer LE(Out, little);
 144     for (const auto &ProfileData : *V) {
 145       const InstrProfRecord &ProfRecord = ProfileData.second;
 146       if (NamedInstrProfRecord::hasCSFlagInHash(ProfileData.first))
 147         CSSummaryBuilder->addRecord(ProfRecord);
 148       else
 149         SummaryBuilder->addRecord(ProfRecord);
 150
 151       LE.write<uint64_t>(ProfileData.first); // Function hash
 152       LE.write<uint64_t>(ProfRecord.Counts.size());
 153       for (uint64_t I : ProfRecord.Counts)
 154         LE.write<uint64_t>(I);
 155
 156       // Write value data
 157       std::unique_ptr<ValueProfData> VDataPtr =
 158           ValueProfData::serializeFrom(ProfileData.second);
 159       uint32_t S = VDataPtr->getSize();
 160       VDataPtr->swapBytesFromHost(ValueProfDataEndianness);
 161       Out.write((const char *)VDataPtr.get(), S);
 162     }
 163   }
 164 };
 165
 166 } // end namespace llvm
 167
 168 InstrProfWriter::InstrProfWriter(bool Sparse)
 169     : Sparse(Sparse), InfoObj(new InstrProfRecordWriterTrait()) {}
 170
 171 InstrProfWriter::~InstrProfWriter() { delete InfoObj; }
 172
 173 // Internal interface for testing purpose only.
 174 void InstrProfWriter::setValueProfDataEndianness(
 175     support::endianness Endianness) {
 176   InfoObj->ValueProfDataEndianness = Endianness;
 177 }
 178
 179 void InstrProfWriter::setOutputSparse(bool Sparse) {
 180   this->Sparse = Sparse;
 181 }
 182
 183 void InstrProfWriter::addRecord(NamedInstrProfRecord &&I, uint64_t Weight,
 184                                 function_ref<void(Error)> Warn) {
 185   auto Name = I.Name;
 186   auto Hash = I.Hash;
 187   addRecord(Name, Hash, std::move(I), Weight, Warn);
 188 }
 189
 190 void InstrProfWriter::overlapRecord(NamedInstrProfRecord &&Other,
 191                                     OverlapStats &Overlap,
 192                                     OverlapStats &FuncLevelOverlap,
 193                                     const OverlapFuncFilters &FuncFilter) {
 194   auto Name = Other.Name;
 195   auto Hash = Other.Hash;
 196   Other.accumulateCounts(FuncLevelOverlap.Test);
 197   if (FunctionData.find(Name) == FunctionData.end()) {
 198     Overlap.addOneUnique(FuncLevelOverlap.Test);
 199     return;
 200   }
 201   if (FuncLevelOverlap.Test.CountSum < 1.0f) {
 202     Overlap.Overlap.NumEntries += 1;
 203     return;
 204   }
 205   auto &ProfileDataMap = FunctionData[Name];
 206   bool NewFunc;
 207   ProfilingData::iterator Where;
 208   std::tie(Where, NewFunc) =
 209       ProfileDataMap.insert(std::make_pair(Hash, InstrProfRecord()));
 210   if (NewFunc) {
 211     Overlap.addOneMismatch(FuncLevelOverlap.Test);
 212     return;
 213   }
 214   InstrProfRecord &Dest = Where->second;
 215
 216   uint64_t ValueCutoff = FuncFilter.ValueCutoff;
 217   if (!FuncFilter.NameFilter.empty() &&
 218       Name.find(FuncFilter.NameFilter) != Name.npos)
 219     ValueCutoff = 0;
 220
 221   Dest.overlap(Other, Overlap, FuncLevelOverlap, ValueCutoff);
 222 }
 223
 224 void InstrProfWriter::addRecord(StringRef Name, uint64_t Hash,
 225                                 InstrProfRecord &&I, uint64_t Weight,
 226                                 function_ref<void(Error)> Warn) {
 227   auto &ProfileDataMap = FunctionData[Name];
 228
 229   bool NewFunc;
 230   ProfilingData::iterator Where;
 231   std::tie(Where, NewFunc) =
 232       ProfileDataMap.insert(std::make_pair(Hash, InstrProfRecord()));
 233   InstrProfRecord &Dest = Where->second;
 234
 235   auto MapWarn = [&](instrprof_error E) {
 236     Warn(make_error<InstrProfError>(E));
 237   };
 238
 239   if (NewFunc) {
 240     // We've never seen a function with this name and hash, add it.
 241     Dest = std::move(I);
 242     if (Weight > 1)
 243       Dest.scale(Weight, MapWarn);
 244   } else {
 245     // We're updating a function we've seen before.
 246     Dest.merge(I, Weight, MapWarn);
 247   }
 248
 249   Dest.sortValueData();
 250 }
 251
 252 void InstrProfWriter::mergeRecordsFromWriter(InstrProfWriter &&IPW,
 253                                              function_ref<void(Error)> Warn) {
 254   for (auto &I : IPW.FunctionData)
 255     for (auto &Func : I.getValue())
 256       addRecord(I.getKey(), Func.first, std::move(Func.second), 1, Warn);
 257 }
 258
 259 bool InstrProfWriter::shouldEncodeData(const ProfilingData &PD) {
 260   if (!Sparse)
 261     return true;
 262   for (const auto &Func : PD) {
 263     const InstrProfRecord &IPR = Func.second;
 264     if (llvm::any_of(IPR.Counts, [](uint64_t Count) { return Count > 0; }))
 265       return true;
 266   }
 267   return false;
 268 }
 269
 270 static void setSummary(IndexedInstrProf::Summary *TheSummary,
 271                        ProfileSummary &PS) {
 272   using namespace IndexedInstrProf;
 273
 274   std::vector<ProfileSummaryEntry> &Res = PS.getDetailedSummary();
 275   TheSummary->NumSummaryFields = Summary::NumKinds;
 276   TheSummary->NumCutoffEntries = Res.size();
 277   TheSummary->set(Summary::MaxFunctionCount, PS.getMaxFunctionCount());
 278   TheSummary->set(Summary::MaxBlockCount, PS.getMaxCount());
 279   TheSummary->set(Summary::MaxInternalBlockCount, PS.getMaxInternalCount());
 280   TheSummary->set(Summary::TotalBlockCount, PS.getTotalCount());
 281   TheSummary->set(Summary::TotalNumBlocks, PS.getNumCounts());
 282   TheSummary->set(Summary::TotalNumFunctions, PS.getNumFunctions());
 283   for (unsigned I = 0; I < Res.size(); I++)
 284     TheSummary->setEntry(I, Res[I]);
 285 }
 286
 287 void InstrProfWriter::writeImpl(ProfOStream &OS) {
 288   using namespace IndexedInstrProf;
 289
 290   OnDiskChainedHashTableGenerator<InstrProfRecordWriterTrait> Generator;
 291
 292   InstrProfSummaryBuilder ISB(ProfileSummaryBuilder::DefaultCutoffs);
 293   InfoObj->SummaryBuilder = &ISB;
 294   InstrProfSummaryBuilder CSISB(ProfileSummaryBuilder::DefaultCutoffs);
 295   InfoObj->CSSummaryBuilder = &CSISB;
 296
 297   // Populate the hash table generator.
 298   for (const auto &I : FunctionData)
 299     if (shouldEncodeData(I.getValue()))
 300       Generator.insert(I.getKey(), &I.getValue());
 301   // Write the header.
 302   IndexedInstrProf::Header Header;
 303   Header.Magic = IndexedInstrProf::Magic;
 304   Header.Version = IndexedInstrProf::ProfVersion::CurrentVersion;
 305   if (ProfileKind == PF_IRLevel)
 306     Header.Version |= VARIANT_MASK_IR_PROF;
 307   if (ProfileKind == PF_IRLevelWithCS) {
 308     Header.Version |= VARIANT_MASK_IR_PROF;
 309     Header.Version |= VARIANT_MASK_CSIR_PROF;
 310   }
 311   Header.Unused = 0;
 312   Header.HashType = static_cast<uint64_t>(IndexedInstrProf::HashType);
 313   Header.HashOffset = 0;
 314   int N = sizeof(IndexedInstrProf::Header) / sizeof(uint64_t);
 315
 316   // Only write out all the fields except 'HashOffset'. We need
 317   // to remember the offset of that field to allow back patching
 318   // later.
 319   for (int I = 0; I < N - 1; I++)
 320     OS.write(reinterpret_cast<uint64_t *>(&Header)[I]);
 321
 322   // Save the location of Header.HashOffset field in \c OS.
 323   uint64_t HashTableStartFieldOffset = OS.tell();
 324   // Reserve the space for HashOffset field.
 325   OS.write(0);
 326
 327   // Reserve space to write profile summary data.
 328   uint32_t NumEntries = ProfileSummaryBuilder::DefaultCutoffs.size();
 329   uint32_t SummarySize = Summary::getSize(Summary::NumKinds, NumEntries);
 330   // Remember the summary offset.
 331   uint64_t SummaryOffset = OS.tell();
 332   for (unsigned I = 0; I < SummarySize / sizeof(uint64_t); I++)
 333     OS.write(0);
 334   uint64_t CSSummaryOffset = 0;
 335   uint64_t CSSummarySize = 0;
 336   if (ProfileKind == PF_IRLevelWithCS) {
 337     CSSummaryOffset = OS.tell();
 338     CSSummarySize = SummarySize / sizeof(uint64_t);
 339     for (unsigned I = 0; I < CSSummarySize; I++)
 340       OS.write(0);
 341   }
 342
 343   // Write the hash table.
 344   uint64_t HashTableStart = Generator.Emit(OS.OS, *InfoObj);
 345
 346   // Allocate space for data to be serialized out.
 347   std::unique_ptr<IndexedInstrProf::Summary> TheSummary =
 348       IndexedInstrProf::allocSummary(SummarySize);
 349   // Compute the Summary and copy the data to the data
 350   // structure to be serialized out (to disk or buffer).
 351   std::unique_ptr<ProfileSummary> PS = ISB.getSummary();
 352   setSummary(TheSummary.get(), *PS);
 353   InfoObj->SummaryBuilder = nullptr;
 354
 355   // For Context Sensitive summary.
 356   std::unique_ptr<IndexedInstrProf::Summary> TheCSSummary = nullptr;
 357   if (ProfileKind == PF_IRLevelWithCS) {
 358     TheCSSummary = IndexedInstrProf::allocSummary(SummarySize);
 359     std::unique_ptr<ProfileSummary> CSPS = CSISB.getSummary();
 360     setSummary(TheCSSummary.get(), *CSPS);
 361   }
 362   InfoObj->CSSummaryBuilder = nullptr;
 363
 364   // Now do the final patch:
 365   PatchItem PatchItems[] = {
 366       // Patch the Header.HashOffset field.
 367       {HashTableStartFieldOffset, &HashTableStart, 1},
 368       // Patch the summary data.
 369       {SummaryOffset, reinterpret_cast<uint64_t *>(TheSummary.get()),
 370        (int)(SummarySize / sizeof(uint64_t))},
 371       {CSSummaryOffset, reinterpret_cast<uint64_t *>(TheCSSummary.get()),
 372        (int)CSSummarySize}};
 373
 374   OS.patch(PatchItems, sizeof(PatchItems) / sizeof(*PatchItems));
 375 }
 376
 377 void InstrProfWriter::write(raw_fd_ostream &OS) {
 378   // Write the hash table.
 379   ProfOStream POS(OS);
 380   writeImpl(POS);
 381 }
 382
 383 std::unique_ptr<MemoryBuffer> InstrProfWriter::writeBuffer() {
 384   std::string Data;
 385   raw_string_ostream OS(Data);
 386   ProfOStream POS(OS);
 387   // Write the hash table.
 388   writeImpl(POS);
 389   // Return this in an aligned memory buffer.
 390   return MemoryBuffer::getMemBufferCopy(Data);
 391 }
 392
 393 static const char *ValueProfKindStr[] = {
 394 #define VALUE_PROF_KIND(Enumerator, Value, Descr) #Enumerator,
 395 #include "llvm/ProfileData/InstrProfData.inc"
 396 };
 397
 398 void InstrProfWriter::writeRecordInText(StringRef Name, uint64_t Hash,
 399                                         const InstrProfRecord &Func,
 400                                         InstrProfSymtab &Symtab,
 401                                         raw_fd_ostream &OS) {
 402   OS << Name << "\n";
 403   OS << "# Func Hash:\n" << Hash << "\n";
 404   OS << "# Num Counters:\n" << Func.Counts.size() << "\n";
 405   OS << "# Counter Values:\n";
 406   for (uint64_t Count : Func.Counts)
 407     OS << Count << "\n";
 408
 409   uint32_t NumValueKinds = Func.getNumValueKinds();
 410   if (!NumValueKinds) {
 411     OS << "\n";
 412     return;
 413   }
 414
 415   OS << "# Num Value Kinds:\n" << Func.getNumValueKinds() << "\n";
 416   for (uint32_t VK = 0; VK < IPVK_Last + 1; VK++) {
 417     uint32_t NS = Func.getNumValueSites(VK);
 418     if (!NS)
 419       continue;
 420     OS << "# ValueKind = " << ValueProfKindStr[VK] << ":\n" << VK << "\n";
 421     OS << "# NumValueSites:\n" << NS << "\n";
 422     for (uint32_t S = 0; S < NS; S++) {
 423       uint32_t ND = Func.getNumValueDataForSite(VK, S);
 424       OS << ND << "\n";
 425       std::unique_ptr<InstrProfValueData[]> VD = Func.getValueForSite(VK, S);
 426       for (uint32_t I = 0; I < ND; I++) {
 427         if (VK == IPVK_IndirectCallTarget)
 428           OS << Symtab.getFuncNameOrExternalSymbol(VD[I].Value) << ":"
 429              << VD[I].Count << "\n";
 430         else
 431           OS << VD[I].Value << ":" << VD[I].Count << "\n";
 432       }
 433     }
 434   }
 435
 436   OS << "\n";
 437 }
 438
 439 Error InstrProfWriter::writeText(raw_fd_ostream &OS) {
 440   if (ProfileKind == PF_IRLevel)
 441     OS << "# IR level Instrumentation Flag\n:ir\n";
 442   else if (ProfileKind == PF_IRLevelWithCS)
 443     OS << "# CSIR level Instrumentation Flag\n:csir\n";
 444   InstrProfSymtab Symtab;
 445
 446   using FuncPair = detail::DenseMapPair<uint64_t, InstrProfRecord>;
 447   using RecordType = std::pair<StringRef, FuncPair>;
 448   SmallVector<RecordType, 4> OrderedFuncData;
 449
 450   for (const auto &I : FunctionData) {
 451     if (shouldEncodeData(I.getValue())) {
 452       if (Error E = Symtab.addFuncName(I.getKey()))
 453         return E;
 454       for (const auto &Func : I.getValue())
 455         OrderedFuncData.push_back(std::make_pair(I.getKey(), Func));
 456     }
 457   }
 458
 459   llvm::sort(OrderedFuncData, [](const RecordType &A, const RecordType &B) {
 460     return std::tie(A.first, A.second.first) <
 461            std::tie(B.first, B.second.first);
 462   });
 463
 464   for (const auto &record : OrderedFuncData) {
 465     const StringRef &Name = record.first;
 466     const FuncPair &Func = record.second;
 467     writeRecordInText(Name, Func.first, Func.second, Symtab, OS);
 468   }
 469
 470   return Error::success();
 471 }