lib/ProfileData/SampleProfReader.cpp

   1 //===- SampleProfReader.cpp - Read 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 reads LLVM sample profiles. It
  10 // supports three file formats: text, binary and gcov.
  11 //
  12 // The textual representation is useful for debugging and testing purposes. The
  13 // binary representation is more compact, resulting in smaller file sizes.
  14 //
  15 // The gcov encoding is the one generated by GCC's AutoFDO profile creation
  16 // tool (https://github.com/google/autofdo)
  17 //
  18 // All three encodings can be used interchangeably as an input sample profile.
  19 //
  20 //===----------------------------------------------------------------------===//
  21
  22 #include "llvm/ProfileData/SampleProfReader.h"
  23 #include "llvm/ADT/DenseMap.h"
  24 #include "llvm/ADT/STLExtras.h"
  25 #include "llvm/ADT/StringRef.h"
  26 #include "llvm/IR/ProfileSummary.h"
  27 #include "llvm/ProfileData/ProfileCommon.h"
  28 #include "llvm/ProfileData/SampleProf.h"
  29 #include "llvm/Support/ErrorOr.h"
  30 #include "llvm/Support/LEB128.h"
  31 #include "llvm/Support/LineIterator.h"
  32 #include "llvm/Support/MD5.h"
  33 #include "llvm/Support/MemoryBuffer.h"
  34 #include "llvm/Support/raw_ostream.h"
  35 #include <algorithm>
  36 #include <cstddef>
  37 #include <cstdint>
  38 #include <limits>
  39 #include <memory>
  40 #include <system_error>
  41 #include <vector>
  42
  43 using namespace llvm;
  44 using namespace sampleprof;
  45
  46 /// Dump the function profile for \p FName.
  47 ///
  48 /// \param FName Name of the function to print.
  49 /// \param OS Stream to emit the output to.
  50 void SampleProfileReader::dumpFunctionProfile(StringRef FName,
  51                                               raw_ostream &OS) {
  52   OS << "Function: " << FName << ": " << Profiles[FName];
  53 }
  54
  55 /// Dump all the function profiles found on stream \p OS.
  56 void SampleProfileReader::dump(raw_ostream &OS) {
  57   for (const auto &I : Profiles)
  58     dumpFunctionProfile(I.getKey(), OS);
  59 }
  60
  61 /// Parse \p Input as function head.
  62 ///
  63 /// Parse one line of \p Input, and update function name in \p FName,
  64 /// function's total sample count in \p NumSamples, function's entry
  65 /// count in \p NumHeadSamples.
  66 ///
  67 /// \returns true if parsing is successful.
  68 static bool ParseHead(const StringRef &Input, StringRef &FName,
  69                       uint64_t &NumSamples, uint64_t &NumHeadSamples) {
  70   if (Input[0] == ' ')
  71     return false;
  72   size_t n2 = Input.rfind(':');
  73   size_t n1 = Input.rfind(':', n2 - 1);
  74   FName = Input.substr(0, n1);
  75   if (Input.substr(n1 + 1, n2 - n1 - 1).getAsInteger(10, NumSamples))
  76     return false;
  77   if (Input.substr(n2 + 1).getAsInteger(10, NumHeadSamples))
  78     return false;
  79   return true;
  80 }
  81
  82 /// Returns true if line offset \p L is legal (only has 16 bits).
  83 static bool isOffsetLegal(unsigned L) { return (L & 0xffff) == L; }
  84
  85 /// Parse \p Input as line sample.
  86 ///
  87 /// \param Input input line.
  88 /// \param IsCallsite true if the line represents an inlined callsite.
  89 /// \param Depth the depth of the inline stack.
  90 /// \param NumSamples total samples of the line/inlined callsite.
  91 /// \param LineOffset line offset to the start of the function.
  92 /// \param Discriminator discriminator of the line.
  93 /// \param TargetCountMap map from indirect call target to count.
  94 ///
  95 /// returns true if parsing is successful.
  96 static bool ParseLine(const StringRef &Input, bool &IsCallsite, uint32_t &Depth,
  97                       uint64_t &NumSamples, uint32_t &LineOffset,
  98                       uint32_t &Discriminator, StringRef &CalleeName,
  99                       DenseMap<StringRef, uint64_t> &TargetCountMap) {
 100   for (Depth = 0; Input[Depth] == ' '; Depth++)
 101     ;
 102   if (Depth == 0)
 103     return false;
 104
 105   size_t n1 = Input.find(':');
 106   StringRef Loc = Input.substr(Depth, n1 - Depth);
 107   size_t n2 = Loc.find('.');
 108   if (n2 == StringRef::npos) {
 109     if (Loc.getAsInteger(10, LineOffset) || !isOffsetLegal(LineOffset))
 110       return false;
 111     Discriminator = 0;
 112   } else {
 113     if (Loc.substr(0, n2).getAsInteger(10, LineOffset))
 114       return false;
 115     if (Loc.substr(n2 + 1).getAsInteger(10, Discriminator))
 116       return false;
 117   }
 118
 119   StringRef Rest = Input.substr(n1 + 2);
 120   if (Rest[0] >= '0' && Rest[0] <= '9') {
 121     IsCallsite = false;
 122     size_t n3 = Rest.find(' ');
 123     if (n3 == StringRef::npos) {
 124       if (Rest.getAsInteger(10, NumSamples))
 125         return false;
 126     } else {
 127       if (Rest.substr(0, n3).getAsInteger(10, NumSamples))
 128         return false;
 129     }
 130     // Find call targets and their sample counts.
 131     // Note: In some cases, there are symbols in the profile which are not
 132     // mangled. To accommodate such cases, use colon + integer pairs as the
 133     // anchor points.
 134     // An example:
 135     // _M_construct<char *>:1000 string_view<std::allocator<char> >:437
 136     // ":1000" and ":437" are used as anchor points so the string above will
 137     // be interpreted as
 138     // target: _M_construct<char *>
 139     // count: 1000
 140     // target: string_view<std::allocator<char> >
 141     // count: 437
 142     while (n3 != StringRef::npos) {
 143       n3 += Rest.substr(n3).find_first_not_of(' ');
 144       Rest = Rest.substr(n3);
 145       n3 = Rest.find_first_of(':');
 146       if (n3 == StringRef::npos || n3 == 0)
 147         return false;
 148
 149       StringRef Target;
 150       uint64_t count, n4;
 151       while (true) {
 152         // Get the segment after the current colon.
 153         StringRef AfterColon = Rest.substr(n3 + 1);
 154         // Get the target symbol before the current colon.
 155         Target = Rest.substr(0, n3);
 156         // Check if the word after the current colon is an integer.
 157         n4 = AfterColon.find_first_of(' ');
 158         n4 = (n4 != StringRef::npos) ? n3 + n4 + 1 : Rest.size();
 159         StringRef WordAfterColon = Rest.substr(n3 + 1, n4 - n3 - 1);
 160         if (!WordAfterColon.getAsInteger(10, count))
 161           break;
 162
 163         // Try to find the next colon.
 164         uint64_t n5 = AfterColon.find_first_of(':');
 165         if (n5 == StringRef::npos)
 166           return false;
 167         n3 += n5 + 1;
 168       }
 169
 170       // An anchor point is found. Save the {target, count} pair
 171       TargetCountMap[Target] = count;
 172       if (n4 == Rest.size())
 173         break;
 174       // Change n3 to the next blank space after colon + integer pair.
 175       n3 = n4;
 176     }
 177   } else {
 178     IsCallsite = true;
 179     size_t n3 = Rest.find_last_of(':');
 180     CalleeName = Rest.substr(0, n3);
 181     if (Rest.substr(n3 + 1).getAsInteger(10, NumSamples))
 182       return false;
 183   }
 184   return true;
 185 }
 186
 187 /// Load samples from a text file.
 188 ///
 189 /// See the documentation at the top of the file for an explanation of
 190 /// the expected format.
 191 ///
 192 /// \returns true if the file was loaded successfully, false otherwise.
 193 std::error_code SampleProfileReaderText::read() {
 194   line_iterator LineIt(*Buffer, /*SkipBlanks=*/true, '#');
 195   sampleprof_error Result = sampleprof_error::success;
 196
 197   InlineCallStack InlineStack;
 198
 199   for (; !LineIt.is_at_eof(); ++LineIt) {
 200     if ((*LineIt)[(*LineIt).find_first_not_of(' ')] == '#')
 201       continue;
 202     // Read the header of each function.
 203     //
 204     // Note that for function identifiers we are actually expecting
 205     // mangled names, but we may not always get them. This happens when
 206     // the compiler decides not to emit the function (e.g., it was inlined
 207     // and removed). In this case, the binary will not have the linkage
 208     // name for the function, so the profiler will emit the function's
 209     // unmangled name, which may contain characters like ':' and '>' in its
 210     // name (member functions, templates, etc).
 211     //
 212     // The only requirement we place on the identifier, then, is that it
 213     // should not begin with a number.
 214     if ((*LineIt)[0] != ' ') {
 215       uint64_t NumSamples, NumHeadSamples;
 216       StringRef FName;
 217       if (!ParseHead(*LineIt, FName, NumSamples, NumHeadSamples)) {
 218         reportError(LineIt.line_number(),
 219                     "Expected 'mangled_name:NUM:NUM', found " + *LineIt);
 220         return sampleprof_error::malformed;
 221       }
 222       Profiles[FName] = FunctionSamples();
 223       FunctionSamples &FProfile = Profiles[FName];
 224       FProfile.setName(FName);
 225       MergeResult(Result, FProfile.addTotalSamples(NumSamples));
 226       MergeResult(Result, FProfile.addHeadSamples(NumHeadSamples));
 227       InlineStack.clear();
 228       InlineStack.push_back(&FProfile);
 229     } else {
 230       uint64_t NumSamples;
 231       StringRef FName;
 232       DenseMap<StringRef, uint64_t> TargetCountMap;
 233       bool IsCallsite;
 234       uint32_t Depth, LineOffset, Discriminator;
 235       if (!ParseLine(*LineIt, IsCallsite, Depth, NumSamples, LineOffset,
 236                      Discriminator, FName, TargetCountMap)) {
 237         reportError(LineIt.line_number(),
 238                     "Expected 'NUM[.NUM]: NUM[ mangled_name:NUM]*', found " +
 239                         *LineIt);
 240         return sampleprof_error::malformed;
 241       }
 242       if (IsCallsite) {
 243         while (InlineStack.size() > Depth) {
 244           InlineStack.pop_back();
 245         }
 246         FunctionSamples &FSamples = InlineStack.back()->functionSamplesAt(
 247             LineLocation(LineOffset, Discriminator))[FName];
 248         FSamples.setName(FName);
 249         MergeResult(Result, FSamples.addTotalSamples(NumSamples));
 250         InlineStack.push_back(&FSamples);
 251       } else {
 252         while (InlineStack.size() > Depth) {
 253           InlineStack.pop_back();
 254         }
 255         FunctionSamples &FProfile = *InlineStack.back();
 256         for (const auto &name_count : TargetCountMap) {
 257           MergeResult(Result, FProfile.addCalledTargetSamples(
 258                                   LineOffset, Discriminator, name_count.first,
 259                                   name_count.second));
 260         }
 261         MergeResult(Result, FProfile.addBodySamples(LineOffset, Discriminator,
 262                                                     NumSamples));
 263       }
 264     }
 265   }
 266   if (Result == sampleprof_error::success)
 267     computeSummary();
 268
 269   return Result;
 270 }
 271
 272 bool SampleProfileReaderText::hasFormat(const MemoryBuffer &Buffer) {
 273   bool result = false;
 274
 275   // Check that the first non-comment line is a valid function header.
 276   line_iterator LineIt(Buffer, /*SkipBlanks=*/true, '#');
 277   if (!LineIt.is_at_eof()) {
 278     if ((*LineIt)[0] != ' ') {
 279       uint64_t NumSamples, NumHeadSamples;
 280       StringRef FName;
 281       result = ParseHead(*LineIt, FName, NumSamples, NumHeadSamples);
 282     }
 283   }
 284
 285   return result;
 286 }
 287
 288 template <typename T> ErrorOr<T> SampleProfileReaderBinary::readNumber() {
 289   unsigned NumBytesRead = 0;
 290   std::error_code EC;
 291   uint64_t Val = decodeULEB128(Data, &NumBytesRead);
 292
 293   if (Val > std::numeric_limits<T>::max())
 294     EC = sampleprof_error::malformed;
 295   else if (Data + NumBytesRead > End)
 296     EC = sampleprof_error::truncated;
 297   else
 298     EC = sampleprof_error::success;
 299
 300   if (EC) {
 301     reportError(0, EC.message());
 302     return EC;
 303   }
 304
 305   Data += NumBytesRead;
 306   return static_cast<T>(Val);
 307 }
 308
 309 ErrorOr<StringRef> SampleProfileReaderBinary::readString() {
 310   std::error_code EC;
 311   StringRef Str(reinterpret_cast<const char *>(Data));
 312   if (Data + Str.size() + 1 > End) {
 313     EC = sampleprof_error::truncated;
 314     reportError(0, EC.message());
 315     return EC;
 316   }
 317
 318   Data += Str.size() + 1;
 319   return Str;
 320 }
 321
 322 template <typename T>
 323 ErrorOr<T> SampleProfileReaderBinary::readUnencodedNumber() {
 324   std::error_code EC;
 325
 326   if (Data + sizeof(T) > End) {
 327     EC = sampleprof_error::truncated;
 328     reportError(0, EC.message());
 329     return EC;
 330   }
 331
 332   using namespace support;
 333   T Val = endian::readNext<T, little, unaligned>(Data);
 334   return Val;
 335 }
 336
 337 template <typename T>
 338 inline ErrorOr<uint32_t> SampleProfileReaderBinary::readStringIndex(T &Table) {
 339   std::error_code EC;
 340   auto Idx = readNumber<uint32_t>();
 341   if (std::error_code EC = Idx.getError())
 342     return EC;
 343   if (*Idx >= Table.size())
 344     return sampleprof_error::truncated_name_table;
 345   return *Idx;
 346 }
 347
 348 ErrorOr<StringRef> SampleProfileReaderBinary::readStringFromTable() {
 349   auto Idx = readStringIndex(NameTable);
 350   if (std::error_code EC = Idx.getError())
 351     return EC;
 352
 353   return NameTable[*Idx];
 354 }
 355
 356 ErrorOr<StringRef> SampleProfileReaderCompactBinary::readStringFromTable() {
 357   auto Idx = readStringIndex(NameTable);
 358   if (std::error_code EC = Idx.getError())
 359     return EC;
 360
 361   return StringRef(NameTable[*Idx]);
 362 }
 363
 364 std::error_code
 365 SampleProfileReaderBinary::readProfile(FunctionSamples &FProfile) {
 366   auto NumSamples = readNumber<uint64_t>();
 367   if (std::error_code EC = NumSamples.getError())
 368     return EC;
 369   FProfile.addTotalSamples(*NumSamples);
 370
 371   // Read the samples in the body.
 372   auto NumRecords = readNumber<uint32_t>();
 373   if (std::error_code EC = NumRecords.getError())
 374     return EC;
 375
 376   for (uint32_t I = 0; I < *NumRecords; ++I) {
 377     auto LineOffset = readNumber<uint64_t>();
 378     if (std::error_code EC = LineOffset.getError())
 379       return EC;
 380
 381     if (!isOffsetLegal(*LineOffset)) {
 382       return std::error_code();
 383     }
 384
 385     auto Discriminator = readNumber<uint64_t>();
 386     if (std::error_code EC = Discriminator.getError())
 387       return EC;
 388
 389     auto NumSamples = readNumber<uint64_t>();
 390     if (std::error_code EC = NumSamples.getError())
 391       return EC;
 392
 393     auto NumCalls = readNumber<uint32_t>();
 394     if (std::error_code EC = NumCalls.getError())
 395       return EC;
 396
 397     for (uint32_t J = 0; J < *NumCalls; ++J) {
 398       auto CalledFunction(readStringFromTable());
 399       if (std::error_code EC = CalledFunction.getError())
 400         return EC;
 401
 402       auto CalledFunctionSamples = readNumber<uint64_t>();
 403       if (std::error_code EC = CalledFunctionSamples.getError())
 404         return EC;
 405
 406       FProfile.addCalledTargetSamples(*LineOffset, *Discriminator,
 407                                       *CalledFunction, *CalledFunctionSamples);
 408     }
 409
 410     FProfile.addBodySamples(*LineOffset, *Discriminator, *NumSamples);
 411   }
 412
 413   // Read all the samples for inlined function calls.
 414   auto NumCallsites = readNumber<uint32_t>();
 415   if (std::error_code EC = NumCallsites.getError())
 416     return EC;
 417
 418   for (uint32_t J = 0; J < *NumCallsites; ++J) {
 419     auto LineOffset = readNumber<uint64_t>();
 420     if (std::error_code EC = LineOffset.getError())
 421       return EC;
 422
 423     auto Discriminator = readNumber<uint64_t>();
 424     if (std::error_code EC = Discriminator.getError())
 425       return EC;
 426
 427     auto FName(readStringFromTable());
 428     if (std::error_code EC = FName.getError())
 429       return EC;
 430
 431     FunctionSamples &CalleeProfile = FProfile.functionSamplesAt(
 432         LineLocation(*LineOffset, *Discriminator))[*FName];
 433     CalleeProfile.setName(*FName);
 434     if (std::error_code EC = readProfile(CalleeProfile))
 435       return EC;
 436   }
 437
 438   return sampleprof_error::success;
 439 }
 440
 441 std::error_code SampleProfileReaderBinary::readFuncProfile() {
 442   auto NumHeadSamples = readNumber<uint64_t>();
 443   if (std::error_code EC = NumHeadSamples.getError())
 444     return EC;
 445
 446   auto FName(readStringFromTable());
 447   if (std::error_code EC = FName.getError())
 448     return EC;
 449
 450   Profiles[*FName] = FunctionSamples();
 451   FunctionSamples &FProfile = Profiles[*FName];
 452   FProfile.setName(*FName);
 453
 454   FProfile.addHeadSamples(*NumHeadSamples);
 455
 456   if (std::error_code EC = readProfile(FProfile))
 457     return EC;
 458   return sampleprof_error::success;
 459 }
 460
 461 std::error_code SampleProfileReaderBinary::read() {
 462   while (!at_eof()) {
 463     if (std::error_code EC = readFuncProfile())
 464       return EC;
 465   }
 466
 467   return sampleprof_error::success;
 468 }
 469
 470 std::error_code
 471 SampleProfileReaderExtBinary::readOneSection(const uint8_t *Start,
 472                                              uint64_t Size, SecType Type) {
 473   Data = Start;
 474   switch (Type) {
 475   case SecProfSummary:
 476     if (std::error_code EC = readSummary())
 477       return EC;
 478     break;
 479   case SecNameTable:
 480     if (std::error_code EC = readNameTable())
 481       return EC;
 482     break;
 483   case SecLBRProfile:
 484     while (Data < Start + Size) {
 485       if (std::error_code EC = readFuncProfile())
 486         return EC;
 487     }
 488     break;
 489   case SecProfileSymbolList:
 490     if (std::error_code EC = readProfileSymbolList())
 491       return EC;
 492     break;
 493   default:
 494     break;
 495   }
 496   return sampleprof_error::success;
 497 }
 498
 499 std::error_code SampleProfileReaderExtBinary::readProfileSymbolList() {
 500   auto UncompressSize = readNumber<uint64_t>();
 501   if (std::error_code EC = UncompressSize.getError())
 502     return EC;
 503
 504   auto CompressSize = readNumber<uint64_t>();
 505   if (std::error_code EC = CompressSize.getError())
 506     return EC;
 507
 508   if (!ProfSymList)
 509     ProfSymList = std::make_unique<ProfileSymbolList>();
 510
 511   if (std::error_code EC =
 512           ProfSymList->read(*CompressSize, *UncompressSize, Data))
 513     return EC;
 514
 515   // CompressSize is zero only when ProfileSymbolList is not compressed.
 516   if (*CompressSize == 0)
 517     Data = Data + *UncompressSize;
 518   else
 519     Data = Data + *CompressSize;
 520   return sampleprof_error::success;
 521 }
 522
 523 std::error_code SampleProfileReaderExtBinaryBase::read() {
 524   const uint8_t *BufStart =
 525       reinterpret_cast<const uint8_t *>(Buffer->getBufferStart());
 526
 527   for (auto &Entry : SecHdrTable) {
 528     // Skip empty section.
 529     if (!Entry.Size)
 530       continue;
 531     const uint8_t *SecStart = BufStart + Entry.Offset;
 532     if (std::error_code EC = readOneSection(SecStart, Entry.Size, Entry.Type))
 533       return EC;
 534     if (Data != SecStart + Entry.Size)
 535       return sampleprof_error::malformed;
 536   }
 537
 538   return sampleprof_error::success;
 539 }
 540
 541 std::error_code SampleProfileReaderCompactBinary::read() {
 542   std::vector<uint64_t> OffsetsToUse;
 543   if (UseAllFuncs) {
 544     for (auto FuncEntry : FuncOffsetTable) {
 545       OffsetsToUse.push_back(FuncEntry.second);
 546     }
 547   }
 548   else {
 549     for (auto Name : FuncsToUse) {
 550       auto GUID = std::to_string(MD5Hash(Name));
 551       auto iter = FuncOffsetTable.find(StringRef(GUID));
 552       if (iter == FuncOffsetTable.end())
 553         continue;
 554       OffsetsToUse.push_back(iter->second);
 555     }
 556   }
 557
 558   for (auto Offset : OffsetsToUse) {
 559     const uint8_t *SavedData = Data;
 560     Data = reinterpret_cast<const uint8_t *>(Buffer->getBufferStart()) +
 561            Offset;
 562     if (std::error_code EC = readFuncProfile())
 563       return EC;
 564     Data = SavedData;
 565   }
 566   return sampleprof_error::success;
 567 }
 568
 569 std::error_code SampleProfileReaderRawBinary::verifySPMagic(uint64_t Magic) {
 570   if (Magic == SPMagic())
 571     return sampleprof_error::success;
 572   return sampleprof_error::bad_magic;
 573 }
 574
 575 std::error_code SampleProfileReaderExtBinary::verifySPMagic(uint64_t Magic) {
 576   if (Magic == SPMagic(SPF_Ext_Binary))
 577     return sampleprof_error::success;
 578   return sampleprof_error::bad_magic;
 579 }
 580
 581 std::error_code
 582 SampleProfileReaderCompactBinary::verifySPMagic(uint64_t Magic) {
 583   if (Magic == SPMagic(SPF_Compact_Binary))
 584     return sampleprof_error::success;
 585   return sampleprof_error::bad_magic;
 586 }
 587
 588 std::error_code SampleProfileReaderBinary::readNameTable() {
 589   auto Size = readNumber<uint32_t>();
 590   if (std::error_code EC = Size.getError())
 591     return EC;
 592   NameTable.reserve(*Size);
 593   for (uint32_t I = 0; I < *Size; ++I) {
 594     auto Name(readString());
 595     if (std::error_code EC = Name.getError())
 596       return EC;
 597     NameTable.push_back(*Name);
 598   }
 599
 600   return sampleprof_error::success;
 601 }
 602
 603 std::error_code SampleProfileReaderCompactBinary::readNameTable() {
 604   auto Size = readNumber<uint64_t>();
 605   if (std::error_code EC = Size.getError())
 606     return EC;
 607   NameTable.reserve(*Size);
 608   for (uint32_t I = 0; I < *Size; ++I) {
 609     auto FID = readNumber<uint64_t>();
 610     if (std::error_code EC = FID.getError())
 611       return EC;
 612     NameTable.push_back(std::to_string(*FID));
 613   }
 614   return sampleprof_error::success;
 615 }
 616
 617 std::error_code SampleProfileReaderExtBinaryBase::readSecHdrTableEntry() {
 618   SecHdrTableEntry Entry;
 619   auto Type = readUnencodedNumber<uint64_t>();
 620   if (std::error_code EC = Type.getError())
 621     return EC;
 622   Entry.Type = static_cast<SecType>(*Type);
 623
 624   auto Flag = readUnencodedNumber<uint64_t>();
 625   if (std::error_code EC = Flag.getError())
 626     return EC;
 627   Entry.Flag = *Flag;
 628
 629   auto Offset = readUnencodedNumber<uint64_t>();
 630   if (std::error_code EC = Offset.getError())
 631     return EC;
 632   Entry.Offset = *Offset;
 633
 634   auto Size = readUnencodedNumber<uint64_t>();
 635   if (std::error_code EC = Size.getError())
 636     return EC;
 637   Entry.Size = *Size;
 638
 639   SecHdrTable.push_back(std::move(Entry));
 640   return sampleprof_error::success;
 641 }
 642
 643 std::error_code SampleProfileReaderExtBinaryBase::readSecHdrTable() {
 644   auto EntryNum = readUnencodedNumber<uint64_t>();
 645   if (std::error_code EC = EntryNum.getError())
 646     return EC;
 647
 648   for (uint32_t i = 0; i < (*EntryNum); i++)
 649     if (std::error_code EC = readSecHdrTableEntry())
 650       return EC;
 651
 652   return sampleprof_error::success;
 653 }
 654
 655 std::error_code SampleProfileReaderExtBinaryBase::readHeader() {
 656   const uint8_t *BufStart =
 657       reinterpret_cast<const uint8_t *>(Buffer->getBufferStart());
 658   Data = BufStart;
 659   End = BufStart + Buffer->getBufferSize();
 660
 661   if (std::error_code EC = readMagicIdent())
 662     return EC;
 663
 664   if (std::error_code EC = readSecHdrTable())
 665     return EC;
 666
 667   return sampleprof_error::success;
 668 }
 669
 670 std::error_code SampleProfileReaderBinary::readMagicIdent() {
 671   // Read and check the magic identifier.
 672   auto Magic = readNumber<uint64_t>();
 673   if (std::error_code EC = Magic.getError())
 674     return EC;
 675   else if (std::error_code EC = verifySPMagic(*Magic))
 676     return EC;
 677
 678   // Read the version number.
 679   auto Version = readNumber<uint64_t>();
 680   if (std::error_code EC = Version.getError())
 681     return EC;
 682   else if (*Version != SPVersion())
 683     return sampleprof_error::unsupported_version;
 684
 685   return sampleprof_error::success;
 686 }
 687
 688 std::error_code SampleProfileReaderBinary::readHeader() {
 689   Data = reinterpret_cast<const uint8_t *>(Buffer->getBufferStart());
 690   End = Data + Buffer->getBufferSize();
 691
 692   if (std::error_code EC = readMagicIdent())
 693     return EC;
 694
 695   if (std::error_code EC = readSummary())
 696     return EC;
 697
 698   if (std::error_code EC = readNameTable())
 699     return EC;
 700   return sampleprof_error::success;
 701 }
 702
 703 std::error_code SampleProfileReaderCompactBinary::readHeader() {
 704   SampleProfileReaderBinary::readHeader();
 705   if (std::error_code EC = readFuncOffsetTable())
 706     return EC;
 707   return sampleprof_error::success;
 708 }
 709
 710 std::error_code SampleProfileReaderCompactBinary::readFuncOffsetTable() {
 711   auto TableOffset = readUnencodedNumber<uint64_t>();
 712   if (std::error_code EC = TableOffset.getError())
 713     return EC;
 714
 715   const uint8_t *SavedData = Data;
 716   const uint8_t *TableStart =
 717       reinterpret_cast<const uint8_t *>(Buffer->getBufferStart()) +
 718       *TableOffset;
 719   Data = TableStart;
 720
 721   auto Size = readNumber<uint64_t>();
 722   if (std::error_code EC = Size.getError())
 723     return EC;
 724
 725   FuncOffsetTable.reserve(*Size);
 726   for (uint32_t I = 0; I < *Size; ++I) {
 727     auto FName(readStringFromTable());
 728     if (std::error_code EC = FName.getError())
 729       return EC;
 730
 731     auto Offset = readNumber<uint64_t>();
 732     if (std::error_code EC = Offset.getError())
 733       return EC;
 734
 735     FuncOffsetTable[*FName] = *Offset;
 736   }
 737   End = TableStart;
 738   Data = SavedData;
 739   return sampleprof_error::success;
 740 }
 741
 742 void SampleProfileReaderCompactBinary::collectFuncsToUse(const Module &M) {
 743   UseAllFuncs = false;
 744   FuncsToUse.clear();
 745   for (auto &F : M) {
 746     StringRef CanonName = FunctionSamples::getCanonicalFnName(F);
 747     FuncsToUse.insert(CanonName);
 748   }
 749 }
 750
 751 std::error_code SampleProfileReaderBinary::readSummaryEntry(
 752     std::vector<ProfileSummaryEntry> &Entries) {
 753   auto Cutoff = readNumber<uint64_t>();
 754   if (std::error_code EC = Cutoff.getError())
 755     return EC;
 756
 757   auto MinBlockCount = readNumber<uint64_t>();
 758   if (std::error_code EC = MinBlockCount.getError())
 759     return EC;
 760
 761   auto NumBlocks = readNumber<uint64_t>();
 762   if (std::error_code EC = NumBlocks.getError())
 763     return EC;
 764
 765   Entries.emplace_back(*Cutoff, *MinBlockCount, *NumBlocks);
 766   return sampleprof_error::success;
 767 }
 768
 769 std::error_code SampleProfileReaderBinary::readSummary() {
 770   auto TotalCount = readNumber<uint64_t>();
 771   if (std::error_code EC = TotalCount.getError())
 772     return EC;
 773
 774   auto MaxBlockCount = readNumber<uint64_t>();
 775   if (std::error_code EC = MaxBlockCount.getError())
 776     return EC;
 777
 778   auto MaxFunctionCount = readNumber<uint64_t>();
 779   if (std::error_code EC = MaxFunctionCount.getError())
 780     return EC;
 781
 782   auto NumBlocks = readNumber<uint64_t>();
 783   if (std::error_code EC = NumBlocks.getError())
 784     return EC;
 785
 786   auto NumFunctions = readNumber<uint64_t>();
 787   if (std::error_code EC = NumFunctions.getError())
 788     return EC;
 789
 790   auto NumSummaryEntries = readNumber<uint64_t>();
 791   if (std::error_code EC = NumSummaryEntries.getError())
 792     return EC;
 793
 794   std::vector<ProfileSummaryEntry> Entries;
 795   for (unsigned i = 0; i < *NumSummaryEntries; i++) {
 796     std::error_code EC = readSummaryEntry(Entries);
 797     if (EC != sampleprof_error::success)
 798       return EC;
 799   }
 800   Summary = std::make_unique<ProfileSummary>(
 801       ProfileSummary::PSK_Sample, Entries, *TotalCount, *MaxBlockCount, 0,
 802       *MaxFunctionCount, *NumBlocks, *NumFunctions);
 803
 804   return sampleprof_error::success;
 805 }
 806
 807 bool SampleProfileReaderRawBinary::hasFormat(const MemoryBuffer &Buffer) {
 808   const uint8_t *Data =
 809       reinterpret_cast<const uint8_t *>(Buffer.getBufferStart());
 810   uint64_t Magic = decodeULEB128(Data);
 811   return Magic == SPMagic();
 812 }
 813
 814 bool SampleProfileReaderExtBinary::hasFormat(const MemoryBuffer &Buffer) {
 815   const uint8_t *Data =
 816       reinterpret_cast<const uint8_t *>(Buffer.getBufferStart());
 817   uint64_t Magic = decodeULEB128(Data);
 818   return Magic == SPMagic(SPF_Ext_Binary);
 819 }
 820
 821 bool SampleProfileReaderCompactBinary::hasFormat(const MemoryBuffer &Buffer) {
 822   const uint8_t *Data =
 823       reinterpret_cast<const uint8_t *>(Buffer.getBufferStart());
 824   uint64_t Magic = decodeULEB128(Data);
 825   return Magic == SPMagic(SPF_Compact_Binary);
 826 }
 827
 828 std::error_code SampleProfileReaderGCC::skipNextWord() {
 829   uint32_t dummy;
 830   if (!GcovBuffer.readInt(dummy))
 831     return sampleprof_error::truncated;
 832   return sampleprof_error::success;
 833 }
 834
 835 template <typename T> ErrorOr<T> SampleProfileReaderGCC::readNumber() {
 836   if (sizeof(T) <= sizeof(uint32_t)) {
 837     uint32_t Val;
 838     if (GcovBuffer.readInt(Val) && Val <= std::numeric_limits<T>::max())
 839       return static_cast<T>(Val);
 840   } else if (sizeof(T) <= sizeof(uint64_t)) {
 841     uint64_t Val;
 842     if (GcovBuffer.readInt64(Val) && Val <= std::numeric_limits<T>::max())
 843       return static_cast<T>(Val);
 844   }
 845
 846   std::error_code EC = sampleprof_error::malformed;
 847   reportError(0, EC.message());
 848   return EC;
 849 }
 850
 851 ErrorOr<StringRef> SampleProfileReaderGCC::readString() {
 852   StringRef Str;
 853   if (!GcovBuffer.readString(Str))
 854     return sampleprof_error::truncated;
 855   return Str;
 856 }
 857
 858 std::error_code SampleProfileReaderGCC::readHeader() {
 859   // Read the magic identifier.
 860   if (!GcovBuffer.readGCDAFormat())
 861     return sampleprof_error::unrecognized_format;
 862
 863   // Read the version number. Note - the GCC reader does not validate this
 864   // version, but the profile creator generates v704.
 865   GCOV::GCOVVersion version;
 866   if (!GcovBuffer.readGCOVVersion(version))
 867     return sampleprof_error::unrecognized_format;
 868
 869   if (version != GCOV::V704)
 870     return sampleprof_error::unsupported_version;
 871
 872   // Skip the empty integer.
 873   if (std::error_code EC = skipNextWord())
 874     return EC;
 875
 876   return sampleprof_error::success;
 877 }
 878
 879 std::error_code SampleProfileReaderGCC::readSectionTag(uint32_t Expected) {
 880   uint32_t Tag;
 881   if (!GcovBuffer.readInt(Tag))
 882     return sampleprof_error::truncated;
 883
 884   if (Tag != Expected)
 885     return sampleprof_error::malformed;
 886
 887   if (std::error_code EC = skipNextWord())
 888     return EC;
 889
 890   return sampleprof_error::success;
 891 }
 892
 893 std::error_code SampleProfileReaderGCC::readNameTable() {
 894   if (std::error_code EC = readSectionTag(GCOVTagAFDOFileNames))
 895     return EC;
 896
 897   uint32_t Size;
 898   if (!GcovBuffer.readInt(Size))
 899     return sampleprof_error::truncated;
 900
 901   for (uint32_t I = 0; I < Size; ++I) {
 902     StringRef Str;
 903     if (!GcovBuffer.readString(Str))
 904       return sampleprof_error::truncated;
 905     Names.push_back(Str);
 906   }
 907
 908   return sampleprof_error::success;
 909 }
 910
 911 std::error_code SampleProfileReaderGCC::readFunctionProfiles() {
 912   if (std::error_code EC = readSectionTag(GCOVTagAFDOFunction))
 913     return EC;
 914
 915   uint32_t NumFunctions;
 916   if (!GcovBuffer.readInt(NumFunctions))
 917     return sampleprof_error::truncated;
 918
 919   InlineCallStack Stack;
 920   for (uint32_t I = 0; I < NumFunctions; ++I)
 921     if (std::error_code EC = readOneFunctionProfile(Stack, true, 0))
 922       return EC;
 923
 924   computeSummary();
 925   return sampleprof_error::success;
 926 }
 927
 928 std::error_code SampleProfileReaderGCC::readOneFunctionProfile(
 929     const InlineCallStack &InlineStack, bool Update, uint32_t Offset) {
 930   uint64_t HeadCount = 0;
 931   if (InlineStack.size() == 0)
 932     if (!GcovBuffer.readInt64(HeadCount))
 933       return sampleprof_error::truncated;
 934
 935   uint32_t NameIdx;
 936   if (!GcovBuffer.readInt(NameIdx))
 937     return sampleprof_error::truncated;
 938
 939   StringRef Name(Names[NameIdx]);
 940
 941   uint32_t NumPosCounts;
 942   if (!GcovBuffer.readInt(NumPosCounts))
 943     return sampleprof_error::truncated;
 944
 945   uint32_t NumCallsites;
 946   if (!GcovBuffer.readInt(NumCallsites))
 947     return sampleprof_error::truncated;
 948
 949   FunctionSamples *FProfile = nullptr;
 950   if (InlineStack.size() == 0) {
 951     // If this is a top function that we have already processed, do not
 952     // update its profile again.  This happens in the presence of
 953     // function aliases.  Since these aliases share the same function
 954     // body, there will be identical replicated profiles for the
 955     // original function.  In this case, we simply not bother updating
 956     // the profile of the original function.
 957     FProfile = &Profiles[Name];
 958     FProfile->addHeadSamples(HeadCount);
 959     if (FProfile->getTotalSamples() > 0)
 960       Update = false;
 961   } else {
 962     // Otherwise, we are reading an inlined instance. The top of the
 963     // inline stack contains the profile of the caller. Insert this
 964     // callee in the caller's CallsiteMap.
 965     FunctionSamples *CallerProfile = InlineStack.front();
 966     uint32_t LineOffset = Offset >> 16;
 967     uint32_t Discriminator = Offset & 0xffff;
 968     FProfile = &CallerProfile->functionSamplesAt(
 969         LineLocation(LineOffset, Discriminator))[Name];
 970   }
 971   FProfile->setName(Name);
 972
 973   for (uint32_t I = 0; I < NumPosCounts; ++I) {
 974     uint32_t Offset;
 975     if (!GcovBuffer.readInt(Offset))
 976       return sampleprof_error::truncated;
 977
 978     uint32_t NumTargets;
 979     if (!GcovBuffer.readInt(NumTargets))
 980       return sampleprof_error::truncated;
 981
 982     uint64_t Count;
 983     if (!GcovBuffer.readInt64(Count))
 984       return sampleprof_error::truncated;
 985
 986     // The line location is encoded in the offset as:
 987     //   high 16 bits: line offset to the start of the function.
 988     //   low 16 bits: discriminator.
 989     uint32_t LineOffset = Offset >> 16;
 990     uint32_t Discriminator = Offset & 0xffff;
 991
 992     InlineCallStack NewStack;
 993     NewStack.push_back(FProfile);
 994     NewStack.insert(NewStack.end(), InlineStack.begin(), InlineStack.end());
 995     if (Update) {
 996       // Walk up the inline stack, adding the samples on this line to
 997       // the total sample count of the callers in the chain.
 998       for (auto CallerProfile : NewStack)
 999         CallerProfile->addTotalSamples(Count);
1000
1001       // Update the body samples for the current profile.
1002       FProfile->addBodySamples(LineOffset, Discriminator, Count);
1003     }
1004
1005     // Process the list of functions called at an indirect call site.
1006     // These are all the targets that a function pointer (or virtual
1007     // function) resolved at runtime.
1008     for (uint32_t J = 0; J < NumTargets; J++) {
1009       uint32_t HistVal;
1010       if (!GcovBuffer.readInt(HistVal))
1011         return sampleprof_error::truncated;
1012
1013       if (HistVal != HIST_TYPE_INDIR_CALL_TOPN)
1014         return sampleprof_error::malformed;
1015
1016       uint64_t TargetIdx;
1017       if (!GcovBuffer.readInt64(TargetIdx))
1018         return sampleprof_error::truncated;
1019       StringRef TargetName(Names[TargetIdx]);
1020
1021       uint64_t TargetCount;
1022       if (!GcovBuffer.readInt64(TargetCount))
1023         return sampleprof_error::truncated;
1024
1025       if (Update)
1026         FProfile->addCalledTargetSamples(LineOffset, Discriminator,
1027                                          TargetName, TargetCount);
1028     }
1029   }
1030
1031   // Process all the inlined callers into the current function. These
1032   // are all the callsites that were inlined into this function.
1033   for (uint32_t I = 0; I < NumCallsites; I++) {
1034     // The offset is encoded as:
1035     //   high 16 bits: line offset to the start of the function.
1036     //   low 16 bits: discriminator.
1037     uint32_t Offset;
1038     if (!GcovBuffer.readInt(Offset))
1039       return sampleprof_error::truncated;
1040     InlineCallStack NewStack;
1041     NewStack.push_back(FProfile);
1042     NewStack.insert(NewStack.end(), InlineStack.begin(), InlineStack.end());
1043     if (std::error_code EC = readOneFunctionProfile(NewStack, Update, Offset))
1044       return EC;
1045   }
1046
1047   return sampleprof_error::success;
1048 }
1049
1050 /// Read a GCC AutoFDO profile.
1051 ///
1052 /// This format is generated by the Linux Perf conversion tool at
1053 /// https://github.com/google/autofdo.
1054 std::error_code SampleProfileReaderGCC::read() {
1055   // Read the string table.
1056   if (std::error_code EC = readNameTable())
1057     return EC;
1058
1059   // Read the source profile.
1060   if (std::error_code EC = readFunctionProfiles())
1061     return EC;
1062
1063   return sampleprof_error::success;
1064 }
1065
1066 bool SampleProfileReaderGCC::hasFormat(const MemoryBuffer &Buffer) {
1067   StringRef Magic(reinterpret_cast<const char *>(Buffer.getBufferStart()));
1068   return Magic == "adcg*704";
1069 }
1070
1071 std::error_code SampleProfileReaderItaniumRemapper::read() {
1072   // If the underlying data is in compact format, we can't remap it because
1073   // we don't know what the original function names were.
1074   if (getFormat() == SPF_Compact_Binary) {
1075     Ctx.diagnose(DiagnosticInfoSampleProfile(
1076         Buffer->getBufferIdentifier(),
1077         "Profile data remapping cannot be applied to profile data "
1078         "in compact format (original mangled names are not available).",
1079         DS_Warning));
1080     return sampleprof_error::success;
1081   }
1082
1083   if (Error E = Remappings.read(*Buffer)) {
1084     handleAllErrors(
1085         std::move(E), [&](const SymbolRemappingParseError &ParseError) {
1086           reportError(ParseError.getLineNum(), ParseError.getMessage());
1087         });
1088     return sampleprof_error::malformed;
1089   }
1090
1091   for (auto &Sample : getProfiles())
1092     if (auto Key = Remappings.insert(Sample.first()))
1093       SampleMap.insert({Key, &Sample.second});
1094
1095   return sampleprof_error::success;
1096 }
1097
1098 FunctionSamples *
1099 SampleProfileReaderItaniumRemapper::getSamplesFor(StringRef Fname) {
1100   if (auto Key = Remappings.lookup(Fname))
1101     return SampleMap.lookup(Key);
1102   return SampleProfileReader::getSamplesFor(Fname);
1103 }
1104
1105 /// Prepare a memory buffer for the contents of \p Filename.
1106 ///
1107 /// \returns an error code indicating the status of the buffer.
1108 static ErrorOr<std::unique_ptr<MemoryBuffer>>
1109 setupMemoryBuffer(const Twine &Filename) {
1110   auto BufferOrErr = MemoryBuffer::getFileOrSTDIN(Filename);
1111   if (std::error_code EC = BufferOrErr.getError())
1112     return EC;
1113   auto Buffer = std::move(BufferOrErr.get());
1114
1115   // Sanity check the file.
1116   if (uint64_t(Buffer->getBufferSize()) > std::numeric_limits<uint32_t>::max())
1117     return sampleprof_error::too_large;
1118
1119   return std::move(Buffer);
1120 }
1121
1122 /// Create a sample profile reader based on the format of the input file.
1123 ///
1124 /// \param Filename The file to open.
1125 ///
1126 /// \param C The LLVM context to use to emit diagnostics.
1127 ///
1128 /// \returns an error code indicating the status of the created reader.
1129 ErrorOr<std::unique_ptr<SampleProfileReader>>
1130 SampleProfileReader::create(const Twine &Filename, LLVMContext &C) {
1131   auto BufferOrError = setupMemoryBuffer(Filename);
1132   if (std::error_code EC = BufferOrError.getError())
1133     return EC;
1134   return create(BufferOrError.get(), C);
1135 }
1136
1137 /// Create a sample profile remapper from the given input, to remap the
1138 /// function names in the given profile data.
1139 ///
1140 /// \param Filename The file to open.
1141 ///
1142 /// \param C The LLVM context to use to emit diagnostics.
1143 ///
1144 /// \param Underlying The underlying profile data reader to remap.
1145 ///
1146 /// \returns an error code indicating the status of the created reader.
1147 ErrorOr<std::unique_ptr<SampleProfileReader>>
1148 SampleProfileReaderItaniumRemapper::create(
1149     const Twine &Filename, LLVMContext &C,
1150     std::unique_ptr<SampleProfileReader> Underlying) {
1151   auto BufferOrError = setupMemoryBuffer(Filename);
1152   if (std::error_code EC = BufferOrError.getError())
1153     return EC;
1154   return std::make_unique<SampleProfileReaderItaniumRemapper>(
1155       std::move(BufferOrError.get()), C, std::move(Underlying));
1156 }
1157
1158 /// Create a sample profile reader based on the format of the input data.
1159 ///
1160 /// \param B The memory buffer to create the reader from (assumes ownership).
1161 ///
1162 /// \param C The LLVM context to use to emit diagnostics.
1163 ///
1164 /// \returns an error code indicating the status of the created reader.
1165 ErrorOr<std::unique_ptr<SampleProfileReader>>
1166 SampleProfileReader::create(std::unique_ptr<MemoryBuffer> &B, LLVMContext &C) {
1167   std::unique_ptr<SampleProfileReader> Reader;
1168   if (SampleProfileReaderRawBinary::hasFormat(*B))
1169     Reader.reset(new SampleProfileReaderRawBinary(std::move(B), C));
1170   else if (SampleProfileReaderExtBinary::hasFormat(*B))
1171     Reader.reset(new SampleProfileReaderExtBinary(std::move(B), C));
1172   else if (SampleProfileReaderCompactBinary::hasFormat(*B))
1173     Reader.reset(new SampleProfileReaderCompactBinary(std::move(B), C));
1174   else if (SampleProfileReaderGCC::hasFormat(*B))
1175     Reader.reset(new SampleProfileReaderGCC(std::move(B), C));
1176   else if (SampleProfileReaderText::hasFormat(*B))
1177     Reader.reset(new SampleProfileReaderText(std::move(B), C));
1178   else
1179     return sampleprof_error::unrecognized_format;
1180
1181   FunctionSamples::Format = Reader->getFormat();
1182   if (std::error_code EC = Reader->readHeader()) {
1183     return EC;
1184   }
1185
1186   return std::move(Reader);
1187 }
1188
1189 // For text and GCC file formats, we compute the summary after reading the
1190 // profile. Binary format has the profile summary in its header.
1191 void SampleProfileReader::computeSummary() {
1192   SampleProfileSummaryBuilder Builder(ProfileSummaryBuilder::DefaultCutoffs);
1193   for (const auto &I : Profiles) {
1194     const FunctionSamples &Profile = I.second;
1195     Builder.addRecord(Profile);
1196   }
1197   Summary = Builder.getSummary();
1198 }