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
 105   InstrProfRecordWriterTrait() = default;
 106
 107   static hash_value_type ComputeHash(key_type_ref K) {
 108     return IndexedInstrProf::ComputeHash(K);
 109   }
 110
 111   static std::pair<offset_type, offset_type>
 112   EmitKeyDataLength(raw_ostream &Out, key_type_ref K, data_type_ref V) {
 113     using namespace support;
 114
 115     endian::Writer LE(Out, little);
 116
 117     offset_type N = K.size();
 118     LE.write<offset_type>(N);
 119
 120     offset_type M = 0;
 121     for (const auto &ProfileData : *V) {
 122       const InstrProfRecord &ProfRecord = ProfileData.second;
 123       M += sizeof(uint64_t); // The function hash
 124       M += sizeof(uint64_t); // The size of the Counts vector
 125       M += ProfRecord.Counts.size() * sizeof(uint64_t);
 126
 127       // Value data
 128       M += ValueProfData::getSize(ProfileData.second);
 129     }
 130     LE.write<offset_type>(M);
 131
 132     return std::make_pair(N, M);
 133   }
 134
 135   void EmitKey(raw_ostream &Out, key_type_ref K, offset_type N) {
 136     Out.write(K.data(), N);
 137   }
 138
 139   void EmitData(raw_ostream &Out, key_type_ref, data_type_ref V, offset_type) {
 140     using namespace support;
 141
 142     endian::Writer LE(Out, little);
 143     for (const auto &ProfileData : *V) {
 144       const InstrProfRecord &ProfRecord = ProfileData.second;
 145       SummaryBuilder->addRecord(ProfRecord);
 146
 147       LE.write<uint64_t>(ProfileData.first); // Function hash
 148       LE.write<uint64_t>(ProfRecord.Counts.size());
 149       for (uint64_t I : ProfRecord.Counts)
 150         LE.write<uint64_t>(I);
 151
 152       // Write value data
 153       std::unique_ptr<ValueProfData> VDataPtr =
 154           ValueProfData::serializeFrom(ProfileData.second);
 155       uint32_t S = VDataPtr->getSize();
 156       VDataPtr->swapBytesFromHost(ValueProfDataEndianness);
 157       Out.write((const char *)VDataPtr.get(), S);
 158     }
 159   }
 160 };
 161
 162 } // end namespace llvm
 163
 164 InstrProfWriter::InstrProfWriter(bool Sparse)
 165     : Sparse(Sparse), InfoObj(new InstrProfRecordWriterTrait()) {}
 166
 167 InstrProfWriter::~InstrProfWriter() { delete InfoObj; }
 168
 169 // Internal interface for testing purpose only.
 170 void InstrProfWriter::setValueProfDataEndianness(
 171     support::endianness Endianness) {
 172   InfoObj->ValueProfDataEndianness = Endianness;
 173 }
 174
 175 void InstrProfWriter::setOutputSparse(bool Sparse) {
 176   this->Sparse = Sparse;
 177 }
 178
 179 void InstrProfWriter::addRecord(NamedInstrProfRecord &&I, uint64_t Weight,
 180                                 function_ref<void(Error)> Warn) {
 181   auto Name = I.Name;
 182   auto Hash = I.Hash;
 183   addRecord(Name, Hash, std::move(I), Weight, Warn);
 184 }
 185
 186 void InstrProfWriter::addRecord(StringRef Name, uint64_t Hash,
 187                                 InstrProfRecord &&I, uint64_t Weight,
 188                                 function_ref<void(Error)> Warn) {
 189   auto &ProfileDataMap = FunctionData[Name];
 190
 191   bool NewFunc;
 192   ProfilingData::iterator Where;
 193   std::tie(Where, NewFunc) =
 194       ProfileDataMap.insert(std::make_pair(Hash, InstrProfRecord()));
 195   InstrProfRecord &Dest = Where->second;
 196
 197   auto MapWarn = [&](instrprof_error E) {
 198     Warn(make_error<InstrProfError>(E));
 199   };
 200
 201   if (NewFunc) {
 202     // We've never seen a function with this name and hash, add it.
 203     Dest = std::move(I);
 204     if (Weight > 1)
 205       Dest.scale(Weight, MapWarn);
 206   } else {
 207     // We're updating a function we've seen before.
 208     Dest.merge(I, Weight, MapWarn);
 209   }
 210
 211   Dest.sortValueData();
 212 }
 213
 214 void InstrProfWriter::mergeRecordsFromWriter(InstrProfWriter &&IPW,
 215                                              function_ref<void(Error)> Warn) {
 216   for (auto &I : IPW.FunctionData)
 217     for (auto &Func : I.getValue())
 218       addRecord(I.getKey(), Func.first, std::move(Func.second), 1, Warn);
 219 }
 220
 221 bool InstrProfWriter::shouldEncodeData(const ProfilingData &PD) {
 222   if (!Sparse)
 223     return true;
 224   for (const auto &Func : PD) {
 225     const InstrProfRecord &IPR = Func.second;
 226     if (llvm::any_of(IPR.Counts, [](uint64_t Count) { return Count > 0; }))
 227       return true;
 228   }
 229   return false;
 230 }
 231
 232 static void setSummary(IndexedInstrProf::Summary *TheSummary,
 233                        ProfileSummary &PS) {
 234   using namespace IndexedInstrProf;
 235
 236   std::vector<ProfileSummaryEntry> &Res = PS.getDetailedSummary();
 237   TheSummary->NumSummaryFields = Summary::NumKinds;
 238   TheSummary->NumCutoffEntries = Res.size();
 239   TheSummary->set(Summary::MaxFunctionCount, PS.getMaxFunctionCount());
 240   TheSummary->set(Summary::MaxBlockCount, PS.getMaxCount());
 241   TheSummary->set(Summary::MaxInternalBlockCount, PS.getMaxInternalCount());
 242   TheSummary->set(Summary::TotalBlockCount, PS.getTotalCount());
 243   TheSummary->set(Summary::TotalNumBlocks, PS.getNumCounts());
 244   TheSummary->set(Summary::TotalNumFunctions, PS.getNumFunctions());
 245   for (unsigned I = 0; I < Res.size(); I++)
 246     TheSummary->setEntry(I, Res[I]);
 247 }
 248
 249 void InstrProfWriter::writeImpl(ProfOStream &OS) {
 250   using namespace IndexedInstrProf;
 251
 252   OnDiskChainedHashTableGenerator<InstrProfRecordWriterTrait> Generator;
 253
 254   InstrProfSummaryBuilder ISB(ProfileSummaryBuilder::DefaultCutoffs);
 255   InfoObj->SummaryBuilder = &ISB;
 256
 257   // Populate the hash table generator.
 258   for (const auto &I : FunctionData)
 259     if (shouldEncodeData(I.getValue()))
 260       Generator.insert(I.getKey(), &I.getValue());
 261   // Write the header.
 262   IndexedInstrProf::Header Header;
 263   Header.Magic = IndexedInstrProf::Magic;
 264   Header.Version = IndexedInstrProf::ProfVersion::CurrentVersion;
 265   if (ProfileKind == PF_IRLevel)
 266     Header.Version |= VARIANT_MASK_IR_PROF;
 267   Header.Unused = 0;
 268   Header.HashType = static_cast<uint64_t>(IndexedInstrProf::HashType);
 269   Header.HashOffset = 0;
 270   int N = sizeof(IndexedInstrProf::Header) / sizeof(uint64_t);
 271
 272   // Only write out all the fields except 'HashOffset'. We need
 273   // to remember the offset of that field to allow back patching
 274   // later.
 275   for (int I = 0; I < N - 1; I++)
 276     OS.write(reinterpret_cast<uint64_t *>(&Header)[I]);
 277
 278   // Save the location of Header.HashOffset field in \c OS.
 279   uint64_t HashTableStartFieldOffset = OS.tell();
 280   // Reserve the space for HashOffset field.
 281   OS.write(0);
 282
 283   // Reserve space to write profile summary data.
 284   uint32_t NumEntries = ProfileSummaryBuilder::DefaultCutoffs.size();
 285   uint32_t SummarySize = Summary::getSize(Summary::NumKinds, NumEntries);
 286   // Remember the summary offset.
 287   uint64_t SummaryOffset = OS.tell();
 288   for (unsigned I = 0; I < SummarySize / sizeof(uint64_t); I++)
 289     OS.write(0);
 290
 291   // Write the hash table.
 292   uint64_t HashTableStart = Generator.Emit(OS.OS, *InfoObj);
 293
 294   // Allocate space for data to be serialized out.
 295   std::unique_ptr<IndexedInstrProf::Summary> TheSummary =
 296       IndexedInstrProf::allocSummary(SummarySize);
 297   // Compute the Summary and copy the data to the data
 298   // structure to be serialized out (to disk or buffer).
 299   std::unique_ptr<ProfileSummary> PS = ISB.getSummary();
 300   setSummary(TheSummary.get(), *PS);
 301   InfoObj->SummaryBuilder = nullptr;
 302
 303   // Now do the final patch:
 304   PatchItem PatchItems[] = {
 305       // Patch the Header.HashOffset field.
 306       {HashTableStartFieldOffset, &HashTableStart, 1},
 307       // Patch the summary data.
 308       {SummaryOffset, reinterpret_cast<uint64_t *>(TheSummary.get()),
 309        (int)(SummarySize / sizeof(uint64_t))}};
 310   OS.patch(PatchItems, sizeof(PatchItems) / sizeof(*PatchItems));
 311 }
 312
 313 void InstrProfWriter::write(raw_fd_ostream &OS) {
 314   // Write the hash table.
 315   ProfOStream POS(OS);
 316   writeImpl(POS);
 317 }
 318
 319 std::unique_ptr<MemoryBuffer> InstrProfWriter::writeBuffer() {
 320   std::string Data;
 321   raw_string_ostream OS(Data);
 322   ProfOStream POS(OS);
 323   // Write the hash table.
 324   writeImpl(POS);
 325   // Return this in an aligned memory buffer.
 326   return MemoryBuffer::getMemBufferCopy(Data);
 327 }
 328
 329 static const char *ValueProfKindStr[] = {
 330 #define VALUE_PROF_KIND(Enumerator, Value) #Enumerator,
 331 #include "llvm/ProfileData/InstrProfData.inc"
 332 };
 333
 334 void InstrProfWriter::writeRecordInText(StringRef Name, uint64_t Hash,
 335                                         const InstrProfRecord &Func,
 336                                         InstrProfSymtab &Symtab,
 337                                         raw_fd_ostream &OS) {
 338   OS << Name << "\n";
 339   OS << "# Func Hash:\n" << Hash << "\n";
 340   OS << "# Num Counters:\n" << Func.Counts.size() << "\n";
 341   OS << "# Counter Values:\n";
 342   for (uint64_t Count : Func.Counts)
 343     OS << Count << "\n";
 344
 345   uint32_t NumValueKinds = Func.getNumValueKinds();
 346   if (!NumValueKinds) {
 347     OS << "\n";
 348     return;
 349   }
 350
 351   OS << "# Num Value Kinds:\n" << Func.getNumValueKinds() << "\n";
 352   for (uint32_t VK = 0; VK < IPVK_Last + 1; VK++) {
 353     uint32_t NS = Func.getNumValueSites(VK);
 354     if (!NS)
 355       continue;
 356     OS << "# ValueKind = " << ValueProfKindStr[VK] << ":\n" << VK << "\n";
 357     OS << "# NumValueSites:\n" << NS << "\n";
 358     for (uint32_t S = 0; S < NS; S++) {
 359       uint32_t ND = Func.getNumValueDataForSite(VK, S);
 360       OS << ND << "\n";
 361       std::unique_ptr<InstrProfValueData[]> VD = Func.getValueForSite(VK, S);
 362       for (uint32_t I = 0; I < ND; I++) {
 363         if (VK == IPVK_IndirectCallTarget)
 364           OS << Symtab.getFuncNameOrExternalSymbol(VD[I].Value) << ":"
 365              << VD[I].Count << "\n";
 366         else
 367           OS << VD[I].Value << ":" << VD[I].Count << "\n";
 368       }
 369     }
 370   }
 371
 372   OS << "\n";
 373 }
 374
 375 Error InstrProfWriter::writeText(raw_fd_ostream &OS) {
 376   if (ProfileKind == PF_IRLevel)
 377     OS << "# IR level Instrumentation Flag\n:ir\n";
 378   InstrProfSymtab Symtab;
 379   for (const auto &I : FunctionData)
 380     if (shouldEncodeData(I.getValue()))
 381       if (Error E = Symtab.addFuncName(I.getKey()))
 382         return E;
 383
 384   for (const auto &I : FunctionData)
 385     if (shouldEncodeData(I.getValue()))
 386       for (const auto &Func : I.getValue())
 387         writeRecordInText(I.getKey(), Func.first, Func.second, Symtab, OS);
 388   return Error::success();
 389 }