lib/ProfileData/InstrProfWriter.cpp

   1 //===- InstrProfWriter.cpp - Instrumented profiling writer ----------------===//
   2 //
   3 //                     The LLVM Compiler Infrastructure
   4 //
   5 // This file is distributed under the University of Illinois Open Source
   6 // License. See LICENSE.TXT for details.
   7 //
   8 //===----------------------------------------------------------------------===//
   9 //
  10 // This file contains support for writing profiling data for clang's
  11 // instrumentation based PGO and coverage.
  12 //
  13 //===----------------------------------------------------------------------===//
  14
  15 #include "llvm/ProfileData/InstrProfWriter.h"
  16 #include "llvm/ADT/STLExtras.h"
  17 #include "llvm/ADT/StringRef.h"
  18 #include "llvm/IR/ProfileSummary.h"
  19 #include "llvm/ProfileData/InstrProf.h"
  20 #include "llvm/ProfileData/ProfileCommon.h"
  21 #include "llvm/Support/Endian.h"
  22 #include "llvm/Support/EndianStream.h"
  23 #include "llvm/Support/Error.h"
  24 #include "llvm/Support/MemoryBuffer.h"
  25 #include "llvm/Support/OnDiskHashTable.h"
  26 #include "llvm/Support/raw_ostream.h"
  27 #include <algorithm>
  28 #include <cstdint>
  29 #include <memory>
  30 #include <string>
  31 #include <tuple>
  32 #include <utility>
  33 #include <vector>
  34
  35 using namespace llvm;
  36
  37 // A struct to define how the data stream should be patched. For Indexed
  38 // profiling, only uint64_t data type is needed.
  39 struct PatchItem {
  40   uint64_t Pos; // Where to patch.
  41   uint64_t *D;  // Pointer to an array of source data.
  42   int N;        // Number of elements in \c D array.
  43 };
  44
  45 namespace llvm {
  46
  47 // A wrapper class to abstract writer stream with support of bytes
  48 // back patching.
  49 class ProfOStream {
  50 public:
  51   ProfOStream(raw_fd_ostream &FD)
  52       : IsFDOStream(true), OS(FD), LE(FD, support::little) {}
  53   ProfOStream(raw_string_ostream &STR)
  54       : IsFDOStream(false), OS(STR), LE(STR, support::little) {}
  55
  56   uint64_t tell() { return OS.tell(); }
  57   void write(uint64_t V) { LE.write<uint64_t>(V); }
  58
  59   // \c patch can only be called when all data is written and flushed.
  60   // For raw_string_ostream, the patch is done on the target string
  61   // directly and it won't be reflected in the stream's internal buffer.
  62   void patch(PatchItem *P, int NItems) {
  63     using namespace support;
  64
  65     if (IsFDOStream) {
  66       raw_fd_ostream &FDOStream = static_cast<raw_fd_ostream &>(OS);
  67       for (int K = 0; K < NItems; K++) {
  68         FDOStream.seek(P[K].Pos);
  69         for (int I = 0; I < P[K].N; I++)
  70           write(P[K].D[I]);
  71       }
  72     } else {
  73       raw_string_ostream &SOStream = static_cast<raw_string_ostream &>(OS);
  74       std::string &Data = SOStream.str(); // with flush
  75       for (int K = 0; K < NItems; K++) {
  76         for (int I = 0; I < P[K].N; I++) {
  77           uint64_t Bytes = endian::byte_swap<uint64_t, little>(P[K].D[I]);
  78           Data.replace(P[K].Pos + I * sizeof(uint64_t), sizeof(uint64_t),
  79                        (const char *)&Bytes, sizeof(uint64_t));
  80         }
  81       }
  82     }
  83   }
  84
  85   // If \c OS is an instance of \c raw_fd_ostream, this field will be
  86   // true. Otherwise, \c OS will be an raw_string_ostream.
  87   bool IsFDOStream;
  88   raw_ostream &OS;
  89   support::endian::Writer LE;
  90 };
  91
  92 class InstrProfRecordWriterTrait {
  93 public:
  94   using key_type = StringRef;
  95   using key_type_ref = StringRef;
  96
  97   using data_type = const InstrProfWriter::ProfilingData *const;
  98   using data_type_ref = const InstrProfWriter::ProfilingData *const;
  99
 100   using hash_value_type = uint64_t;
 101   using offset_type = uint64_t;
 102
 103   support::endianness ValueProfDataEndianness = support::little;
 104   InstrProfSummaryBuilder *SummaryBuilder;
 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       SummaryBuilder->addRecord(ProfRecord);
 147
 148       LE.write<uint64_t>(ProfileData.first); // Function hash
 149       LE.write<uint64_t>(ProfRecord.Counts.size());
 150       for (uint64_t I : ProfRecord.Counts)
 151         LE.write<uint64_t>(I);
 152
 153       // Write value data
 154       std::unique_ptr<ValueProfData> VDataPtr =
 155           ValueProfData::serializeFrom(ProfileData.second);
 156       uint32_t S = VDataPtr->getSize();
 157       VDataPtr->swapBytesFromHost(ValueProfDataEndianness);
 158       Out.write((const char *)VDataPtr.get(), S);
 159     }
 160   }
 161 };
 162
 163 } // end namespace llvm
 164
 165 InstrProfWriter::InstrProfWriter(bool Sparse)
 166     : Sparse(Sparse), InfoObj(new InstrProfRecordWriterTrait()) {}
 167
 168 InstrProfWriter::~InstrProfWriter() { delete InfoObj; }
 169
 170 // Internal interface for testing purpose only.
 171 void InstrProfWriter::setValueProfDataEndianness(
 172     support::endianness Endianness) {
 173   InfoObj->ValueProfDataEndianness = Endianness;
 174 }
 175
 176 void InstrProfWriter::setOutputSparse(bool Sparse) {
 177   this->Sparse = Sparse;
 178 }
 179
 180 void InstrProfWriter::addRecord(NamedInstrProfRecord &&I, uint64_t Weight,
 181                                 function_ref<void(Error)> Warn) {
 182   auto Name = I.Name;
 183   auto Hash = I.Hash;
 184   addRecord(Name, Hash, std::move(I), Weight, Warn);
 185 }
 186
 187 void InstrProfWriter::addRecord(StringRef Name, uint64_t Hash,
 188                                 InstrProfRecord &&I, uint64_t Weight,
 189                                 function_ref<void(Error)> Warn) {
 190   auto &ProfileDataMap = FunctionData[Name];
 191
 192   bool NewFunc;
 193   ProfilingData::iterator Where;
 194   std::tie(Where, NewFunc) =
 195       ProfileDataMap.insert(std::make_pair(Hash, InstrProfRecord()));
 196   InstrProfRecord &Dest = Where->second;
 197
 198   auto MapWarn = [&](instrprof_error E) {
 199     Warn(make_error<InstrProfError>(E));
 200   };
 201
 202   if (NewFunc) {
 203     // We've never seen a function with this name and hash, add it.
 204     Dest = std::move(I);
 205     if (Weight > 1)
 206       Dest.scale(Weight, MapWarn);
 207   } else {
 208     // We're updating a function we've seen before.
 209     Dest.merge(I, Weight, MapWarn);
 210   }
 211
 212   Dest.sortValueData();
 213 }
 214
 215 void InstrProfWriter::mergeRecordsFromWriter(InstrProfWriter &&IPW,
 216                                              function_ref<void(Error)> Warn) {
 217   for (auto &I : IPW.FunctionData)
 218     for (auto &Func : I.getValue())
 219       addRecord(I.getKey(), Func.first, std::move(Func.second), 1, Warn);
 220 }
 221
 222 bool InstrProfWriter::shouldEncodeData(const ProfilingData &PD) {
 223   if (!Sparse)
 224     return true;
 225   for (const auto &Func : PD) {
 226     const InstrProfRecord &IPR = Func.second;
 227     if (llvm::any_of(IPR.Counts, [](uint64_t Count) { return Count > 0; }))
 228       return true;
 229   }
 230   return false;
 231 }
 232
 233 static void setSummary(IndexedInstrProf::Summary *TheSummary,
 234                        ProfileSummary &PS) {
 235   using namespace IndexedInstrProf;
 236
 237   std::vector<ProfileSummaryEntry> &Res = PS.getDetailedSummary();
 238   TheSummary->NumSummaryFields = Summary::NumKinds;
 239   TheSummary->NumCutoffEntries = Res.size();
 240   TheSummary->set(Summary::MaxFunctionCount, PS.getMaxFunctionCount());
 241   TheSummary->set(Summary::MaxBlockCount, PS.getMaxCount());
 242   TheSummary->set(Summary::MaxInternalBlockCount, PS.getMaxInternalCount());
 243   TheSummary->set(Summary::TotalBlockCount, PS.getTotalCount());
 244   TheSummary->set(Summary::TotalNumBlocks, PS.getNumCounts());
 245   TheSummary->set(Summary::TotalNumFunctions, PS.getNumFunctions());
 246   for (unsigned I = 0; I < Res.size(); I++)
 247     TheSummary->setEntry(I, Res[I]);
 248 }
 249
 250 void InstrProfWriter::writeImpl(ProfOStream &OS) {
 251   using namespace IndexedInstrProf;
 252
 253   OnDiskChainedHashTableGenerator<InstrProfRecordWriterTrait> Generator;
 254
 255   InstrProfSummaryBuilder ISB(ProfileSummaryBuilder::DefaultCutoffs);
 256   InfoObj->SummaryBuilder = &ISB;
 257
 258   // Populate the hash table generator.
 259   for (const auto &I : FunctionData)
 260     if (shouldEncodeData(I.getValue()))
 261       Generator.insert(I.getKey(), &I.getValue());
 262   // Write the header.
 263   IndexedInstrProf::Header Header;
 264   Header.Magic = IndexedInstrProf::Magic;
 265   Header.Version = IndexedInstrProf::ProfVersion::CurrentVersion;
 266   if (ProfileKind == PF_IRLevel)
 267     Header.Version |= VARIANT_MASK_IR_PROF;
 268   Header.Unused = 0;
 269   Header.HashType = static_cast<uint64_t>(IndexedInstrProf::HashType);
 270   Header.HashOffset = 0;
 271   int N = sizeof(IndexedInstrProf::Header) / sizeof(uint64_t);
 272
 273   // Only write out all the fields except 'HashOffset'. We need
 274   // to remember the offset of that field to allow back patching
 275   // later.
 276   for (int I = 0; I < N - 1; I++)
 277     OS.write(reinterpret_cast<uint64_t *>(&Header)[I]);
 278
 279   // Save the location of Header.HashOffset field in \c OS.
 280   uint64_t HashTableStartFieldOffset = OS.tell();
 281   // Reserve the space for HashOffset field.
 282   OS.write(0);
 283
 284   // Reserve space to write profile summary data.
 285   uint32_t NumEntries = ProfileSummaryBuilder::DefaultCutoffs.size();
 286   uint32_t SummarySize = Summary::getSize(Summary::NumKinds, NumEntries);
 287   // Remember the summary offset.
 288   uint64_t SummaryOffset = OS.tell();
 289   for (unsigned I = 0; I < SummarySize / sizeof(uint64_t); I++)
 290     OS.write(0);
 291
 292   // Write the hash table.
 293   uint64_t HashTableStart = Generator.Emit(OS.OS, *InfoObj);
 294
 295   // Allocate space for data to be serialized out.
 296   std::unique_ptr<IndexedInstrProf::Summary> TheSummary =
 297       IndexedInstrProf::allocSummary(SummarySize);
 298   // Compute the Summary and copy the data to the data
 299   // structure to be serialized out (to disk or buffer).
 300   std::unique_ptr<ProfileSummary> PS = ISB.getSummary();
 301   setSummary(TheSummary.get(), *PS);
 302   InfoObj->SummaryBuilder = nullptr;
 303
 304   // Now do the final patch:
 305   PatchItem PatchItems[] = {
 306       // Patch the Header.HashOffset field.
 307       {HashTableStartFieldOffset, &HashTableStart, 1},
 308       // Patch the summary data.
 309       {SummaryOffset, reinterpret_cast<uint64_t *>(TheSummary.get()),
 310        (int)(SummarySize / sizeof(uint64_t))}};
 311   OS.patch(PatchItems, sizeof(PatchItems) / sizeof(*PatchItems));
 312 }
 313
 314 void InstrProfWriter::write(raw_fd_ostream &OS) {
 315   // Write the hash table.
 316   ProfOStream POS(OS);
 317   writeImpl(POS);
 318 }
 319
 320 std::unique_ptr<MemoryBuffer> InstrProfWriter::writeBuffer() {
 321   std::string Data;
 322   raw_string_ostream OS(Data);
 323   ProfOStream POS(OS);
 324   // Write the hash table.
 325   writeImpl(POS);
 326   // Return this in an aligned memory buffer.
 327   return MemoryBuffer::getMemBufferCopy(Data);
 328 }
 329
 330 static const char *ValueProfKindStr[] = {
 331 #define VALUE_PROF_KIND(Enumerator, Value) #Enumerator,
 332 #include "llvm/ProfileData/InstrProfData.inc"
 333 };
 334
 335 void InstrProfWriter::writeRecordInText(StringRef Name, uint64_t Hash,
 336                                         const InstrProfRecord &Func,
 337                                         InstrProfSymtab &Symtab,
 338                                         raw_fd_ostream &OS) {
 339   OS << Name << "\n";
 340   OS << "# Func Hash:\n" << Hash << "\n";
 341   OS << "# Num Counters:\n" << Func.Counts.size() << "\n";
 342   OS << "# Counter Values:\n";
 343   for (uint64_t Count : Func.Counts)
 344     OS << Count << "\n";
 345
 346   uint32_t NumValueKinds = Func.getNumValueKinds();
 347   if (!NumValueKinds) {
 348     OS << "\n";
 349     return;
 350   }
 351
 352   OS << "# Num Value Kinds:\n" << Func.getNumValueKinds() << "\n";
 353   for (uint32_t VK = 0; VK < IPVK_Last + 1; VK++) {
 354     uint32_t NS = Func.getNumValueSites(VK);
 355     if (!NS)
 356       continue;
 357     OS << "# ValueKind = " << ValueProfKindStr[VK] << ":\n" << VK << "\n";
 358     OS << "# NumValueSites:\n" << NS << "\n";
 359     for (uint32_t S = 0; S < NS; S++) {
 360       uint32_t ND = Func.getNumValueDataForSite(VK, S);
 361       OS << ND << "\n";
 362       std::unique_ptr<InstrProfValueData[]> VD = Func.getValueForSite(VK, S);
 363       for (uint32_t I = 0; I < ND; I++) {
 364         if (VK == IPVK_IndirectCallTarget)
 365           OS << Symtab.getFuncNameOrExternalSymbol(VD[I].Value) << ":"
 366              << VD[I].Count << "\n";
 367         else
 368           OS << VD[I].Value << ":" << VD[I].Count << "\n";
 369       }
 370     }
 371   }
 372
 373   OS << "\n";
 374 }
 375
 376 Error InstrProfWriter::writeText(raw_fd_ostream &OS) {
 377   if (ProfileKind == PF_IRLevel)
 378     OS << "# IR level Instrumentation Flag\n:ir\n";
 379   InstrProfSymtab Symtab;
 380   for (const auto &I : FunctionData)
 381     if (shouldEncodeData(I.getValue()))
 382       if (Error E = Symtab.addFuncName(I.getKey()))
 383         return E;
 384
 385   for (const auto &I : FunctionData)
 386     if (shouldEncodeData(I.getValue()))
 387       for (const auto &Func : I.getValue())
 388         writeRecordInText(I.getKey(), Func.first, Func.second, Symtab, OS);
 389   return Error::success();
 390 }