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> SampleProfileReaderRawBinary::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 SampleProfileReaderCompactBinary::read() {
 471   for (auto Name : FuncsToUse) {
 472     auto GUID = std::to_string(MD5Hash(Name));
 473     auto iter = FuncOffsetTable.find(StringRef(GUID));
 474     if (iter == FuncOffsetTable.end())
 475       continue;
 476     const uint8_t *SavedData = Data;
 477     Data = reinterpret_cast<const uint8_t *>(Buffer->getBufferStart()) +
 478            iter->second;
 479     if (std::error_code EC = readFuncProfile())
 480       return EC;
 481     Data = SavedData;
 482   }
 483   return sampleprof_error::success;
 484 }
 485
 486 std::error_code SampleProfileReaderRawBinary::verifySPMagic(uint64_t Magic) {
 487   if (Magic == SPMagic())
 488     return sampleprof_error::success;
 489   return sampleprof_error::bad_magic;
 490 }
 491
 492 std::error_code
 493 SampleProfileReaderCompactBinary::verifySPMagic(uint64_t Magic) {
 494   if (Magic == SPMagic(SPF_Compact_Binary))
 495     return sampleprof_error::success;
 496   return sampleprof_error::bad_magic;
 497 }
 498
 499 std::error_code SampleProfileReaderRawBinary::readNameTable() {
 500   auto Size = readNumber<uint32_t>();
 501   if (std::error_code EC = Size.getError())
 502     return EC;
 503   NameTable.reserve(*Size);
 504   for (uint32_t I = 0; I < *Size; ++I) {
 505     auto Name(readString());
 506     if (std::error_code EC = Name.getError())
 507       return EC;
 508     NameTable.push_back(*Name);
 509   }
 510
 511   return sampleprof_error::success;
 512 }
 513
 514 std::error_code SampleProfileReaderCompactBinary::readNameTable() {
 515   auto Size = readNumber<uint64_t>();
 516   if (std::error_code EC = Size.getError())
 517     return EC;
 518   NameTable.reserve(*Size);
 519   for (uint32_t I = 0; I < *Size; ++I) {
 520     auto FID = readNumber<uint64_t>();
 521     if (std::error_code EC = FID.getError())
 522       return EC;
 523     NameTable.push_back(std::to_string(*FID));
 524   }
 525   return sampleprof_error::success;
 526 }
 527
 528 std::error_code SampleProfileReaderBinary::readHeader() {
 529   Data = reinterpret_cast<const uint8_t *>(Buffer->getBufferStart());
 530   End = Data + Buffer->getBufferSize();
 531
 532   // Read and check the magic identifier.
 533   auto Magic = readNumber<uint64_t>();
 534   if (std::error_code EC = Magic.getError())
 535     return EC;
 536   else if (std::error_code EC = verifySPMagic(*Magic))
 537     return EC;
 538
 539   // Read the version number.
 540   auto Version = readNumber<uint64_t>();
 541   if (std::error_code EC = Version.getError())
 542     return EC;
 543   else if (*Version != SPVersion())
 544     return sampleprof_error::unsupported_version;
 545
 546   if (std::error_code EC = readSummary())
 547     return EC;
 548
 549   if (std::error_code EC = readNameTable())
 550     return EC;
 551   return sampleprof_error::success;
 552 }
 553
 554 std::error_code SampleProfileReaderCompactBinary::readHeader() {
 555   SampleProfileReaderBinary::readHeader();
 556   if (std::error_code EC = readFuncOffsetTable())
 557     return EC;
 558   return sampleprof_error::success;
 559 }
 560
 561 std::error_code SampleProfileReaderCompactBinary::readFuncOffsetTable() {
 562   auto TableOffset = readUnencodedNumber<uint64_t>();
 563   if (std::error_code EC = TableOffset.getError())
 564     return EC;
 565
 566   const uint8_t *SavedData = Data;
 567   const uint8_t *TableStart =
 568       reinterpret_cast<const uint8_t *>(Buffer->getBufferStart()) +
 569       *TableOffset;
 570   Data = TableStart;
 571
 572   auto Size = readNumber<uint64_t>();
 573   if (std::error_code EC = Size.getError())
 574     return EC;
 575
 576   FuncOffsetTable.reserve(*Size);
 577   for (uint32_t I = 0; I < *Size; ++I) {
 578     auto FName(readStringFromTable());
 579     if (std::error_code EC = FName.getError())
 580       return EC;
 581
 582     auto Offset = readNumber<uint64_t>();
 583     if (std::error_code EC = Offset.getError())
 584       return EC;
 585
 586     FuncOffsetTable[*FName] = *Offset;
 587   }
 588   End = TableStart;
 589   Data = SavedData;
 590   return sampleprof_error::success;
 591 }
 592
 593 void SampleProfileReaderCompactBinary::collectFuncsToUse(const Module &M) {
 594   FuncsToUse.clear();
 595   for (auto &F : M) {
 596     StringRef CanonName = FunctionSamples::getCanonicalFnName(F);
 597     FuncsToUse.insert(CanonName);
 598   }
 599 }
 600
 601 std::error_code SampleProfileReaderBinary::readSummaryEntry(
 602     std::vector<ProfileSummaryEntry> &Entries) {
 603   auto Cutoff = readNumber<uint64_t>();
 604   if (std::error_code EC = Cutoff.getError())
 605     return EC;
 606
 607   auto MinBlockCount = readNumber<uint64_t>();
 608   if (std::error_code EC = MinBlockCount.getError())
 609     return EC;
 610
 611   auto NumBlocks = readNumber<uint64_t>();
 612   if (std::error_code EC = NumBlocks.getError())
 613     return EC;
 614
 615   Entries.emplace_back(*Cutoff, *MinBlockCount, *NumBlocks);
 616   return sampleprof_error::success;
 617 }
 618
 619 std::error_code SampleProfileReaderBinary::readSummary() {
 620   auto TotalCount = readNumber<uint64_t>();
 621   if (std::error_code EC = TotalCount.getError())
 622     return EC;
 623
 624   auto MaxBlockCount = readNumber<uint64_t>();
 625   if (std::error_code EC = MaxBlockCount.getError())
 626     return EC;
 627
 628   auto MaxFunctionCount = readNumber<uint64_t>();
 629   if (std::error_code EC = MaxFunctionCount.getError())
 630     return EC;
 631
 632   auto NumBlocks = readNumber<uint64_t>();
 633   if (std::error_code EC = NumBlocks.getError())
 634     return EC;
 635
 636   auto NumFunctions = readNumber<uint64_t>();
 637   if (std::error_code EC = NumFunctions.getError())
 638     return EC;
 639
 640   auto NumSummaryEntries = readNumber<uint64_t>();
 641   if (std::error_code EC = NumSummaryEntries.getError())
 642     return EC;
 643
 644   std::vector<ProfileSummaryEntry> Entries;
 645   for (unsigned i = 0; i < *NumSummaryEntries; i++) {
 646     std::error_code EC = readSummaryEntry(Entries);
 647     if (EC != sampleprof_error::success)
 648       return EC;
 649   }
 650   Summary = llvm::make_unique<ProfileSummary>(
 651       ProfileSummary::PSK_Sample, Entries, *TotalCount, *MaxBlockCount, 0,
 652       *MaxFunctionCount, *NumBlocks, *NumFunctions);
 653
 654   return sampleprof_error::success;
 655 }
 656
 657 bool SampleProfileReaderRawBinary::hasFormat(const MemoryBuffer &Buffer) {
 658   const uint8_t *Data =
 659       reinterpret_cast<const uint8_t *>(Buffer.getBufferStart());
 660   uint64_t Magic = decodeULEB128(Data);
 661   return Magic == SPMagic();
 662 }
 663
 664 bool SampleProfileReaderCompactBinary::hasFormat(const MemoryBuffer &Buffer) {
 665   const uint8_t *Data =
 666       reinterpret_cast<const uint8_t *>(Buffer.getBufferStart());
 667   uint64_t Magic = decodeULEB128(Data);
 668   return Magic == SPMagic(SPF_Compact_Binary);
 669 }
 670
 671 std::error_code SampleProfileReaderGCC::skipNextWord() {
 672   uint32_t dummy;
 673   if (!GcovBuffer.readInt(dummy))
 674     return sampleprof_error::truncated;
 675   return sampleprof_error::success;
 676 }
 677
 678 template <typename T> ErrorOr<T> SampleProfileReaderGCC::readNumber() {
 679   if (sizeof(T) <= sizeof(uint32_t)) {
 680     uint32_t Val;
 681     if (GcovBuffer.readInt(Val) && Val <= std::numeric_limits<T>::max())
 682       return static_cast<T>(Val);
 683   } else if (sizeof(T) <= sizeof(uint64_t)) {
 684     uint64_t Val;
 685     if (GcovBuffer.readInt64(Val) && Val <= std::numeric_limits<T>::max())
 686       return static_cast<T>(Val);
 687   }
 688
 689   std::error_code EC = sampleprof_error::malformed;
 690   reportError(0, EC.message());
 691   return EC;
 692 }
 693
 694 ErrorOr<StringRef> SampleProfileReaderGCC::readString() {
 695   StringRef Str;
 696   if (!GcovBuffer.readString(Str))
 697     return sampleprof_error::truncated;
 698   return Str;
 699 }
 700
 701 std::error_code SampleProfileReaderGCC::readHeader() {
 702   // Read the magic identifier.
 703   if (!GcovBuffer.readGCDAFormat())
 704     return sampleprof_error::unrecognized_format;
 705
 706   // Read the version number. Note - the GCC reader does not validate this
 707   // version, but the profile creator generates v704.
 708   GCOV::GCOVVersion version;
 709   if (!GcovBuffer.readGCOVVersion(version))
 710     return sampleprof_error::unrecognized_format;
 711
 712   if (version != GCOV::V704)
 713     return sampleprof_error::unsupported_version;
 714
 715   // Skip the empty integer.
 716   if (std::error_code EC = skipNextWord())
 717     return EC;
 718
 719   return sampleprof_error::success;
 720 }
 721
 722 std::error_code SampleProfileReaderGCC::readSectionTag(uint32_t Expected) {
 723   uint32_t Tag;
 724   if (!GcovBuffer.readInt(Tag))
 725     return sampleprof_error::truncated;
 726
 727   if (Tag != Expected)
 728     return sampleprof_error::malformed;
 729
 730   if (std::error_code EC = skipNextWord())
 731     return EC;
 732
 733   return sampleprof_error::success;
 734 }
 735
 736 std::error_code SampleProfileReaderGCC::readNameTable() {
 737   if (std::error_code EC = readSectionTag(GCOVTagAFDOFileNames))
 738     return EC;
 739
 740   uint32_t Size;
 741   if (!GcovBuffer.readInt(Size))
 742     return sampleprof_error::truncated;
 743
 744   for (uint32_t I = 0; I < Size; ++I) {
 745     StringRef Str;
 746     if (!GcovBuffer.readString(Str))
 747       return sampleprof_error::truncated;
 748     Names.push_back(Str);
 749   }
 750
 751   return sampleprof_error::success;
 752 }
 753
 754 std::error_code SampleProfileReaderGCC::readFunctionProfiles() {
 755   if (std::error_code EC = readSectionTag(GCOVTagAFDOFunction))
 756     return EC;
 757
 758   uint32_t NumFunctions;
 759   if (!GcovBuffer.readInt(NumFunctions))
 760     return sampleprof_error::truncated;
 761
 762   InlineCallStack Stack;
 763   for (uint32_t I = 0; I < NumFunctions; ++I)
 764     if (std::error_code EC = readOneFunctionProfile(Stack, true, 0))
 765       return EC;
 766
 767   computeSummary();
 768   return sampleprof_error::success;
 769 }
 770
 771 std::error_code SampleProfileReaderGCC::readOneFunctionProfile(
 772     const InlineCallStack &InlineStack, bool Update, uint32_t Offset) {
 773   uint64_t HeadCount = 0;
 774   if (InlineStack.size() == 0)
 775     if (!GcovBuffer.readInt64(HeadCount))
 776       return sampleprof_error::truncated;
 777
 778   uint32_t NameIdx;
 779   if (!GcovBuffer.readInt(NameIdx))
 780     return sampleprof_error::truncated;
 781
 782   StringRef Name(Names[NameIdx]);
 783
 784   uint32_t NumPosCounts;
 785   if (!GcovBuffer.readInt(NumPosCounts))
 786     return sampleprof_error::truncated;
 787
 788   uint32_t NumCallsites;
 789   if (!GcovBuffer.readInt(NumCallsites))
 790     return sampleprof_error::truncated;
 791
 792   FunctionSamples *FProfile = nullptr;
 793   if (InlineStack.size() == 0) {
 794     // If this is a top function that we have already processed, do not
 795     // update its profile again.  This happens in the presence of
 796     // function aliases.  Since these aliases share the same function
 797     // body, there will be identical replicated profiles for the
 798     // original function.  In this case, we simply not bother updating
 799     // the profile of the original function.
 800     FProfile = &Profiles[Name];
 801     FProfile->addHeadSamples(HeadCount);
 802     if (FProfile->getTotalSamples() > 0)
 803       Update = false;
 804   } else {
 805     // Otherwise, we are reading an inlined instance. The top of the
 806     // inline stack contains the profile of the caller. Insert this
 807     // callee in the caller's CallsiteMap.
 808     FunctionSamples *CallerProfile = InlineStack.front();
 809     uint32_t LineOffset = Offset >> 16;
 810     uint32_t Discriminator = Offset & 0xffff;
 811     FProfile = &CallerProfile->functionSamplesAt(
 812         LineLocation(LineOffset, Discriminator))[Name];
 813   }
 814   FProfile->setName(Name);
 815
 816   for (uint32_t I = 0; I < NumPosCounts; ++I) {
 817     uint32_t Offset;
 818     if (!GcovBuffer.readInt(Offset))
 819       return sampleprof_error::truncated;
 820
 821     uint32_t NumTargets;
 822     if (!GcovBuffer.readInt(NumTargets))
 823       return sampleprof_error::truncated;
 824
 825     uint64_t Count;
 826     if (!GcovBuffer.readInt64(Count))
 827       return sampleprof_error::truncated;
 828
 829     // The line location is encoded in the offset as:
 830     //   high 16 bits: line offset to the start of the function.
 831     //   low 16 bits: discriminator.
 832     uint32_t LineOffset = Offset >> 16;
 833     uint32_t Discriminator = Offset & 0xffff;
 834
 835     InlineCallStack NewStack;
 836     NewStack.push_back(FProfile);
 837     NewStack.insert(NewStack.end(), InlineStack.begin(), InlineStack.end());
 838     if (Update) {
 839       // Walk up the inline stack, adding the samples on this line to
 840       // the total sample count of the callers in the chain.
 841       for (auto CallerProfile : NewStack)
 842         CallerProfile->addTotalSamples(Count);
 843
 844       // Update the body samples for the current profile.
 845       FProfile->addBodySamples(LineOffset, Discriminator, Count);
 846     }
 847
 848     // Process the list of functions called at an indirect call site.
 849     // These are all the targets that a function pointer (or virtual
 850     // function) resolved at runtime.
 851     for (uint32_t J = 0; J < NumTargets; J++) {
 852       uint32_t HistVal;
 853       if (!GcovBuffer.readInt(HistVal))
 854         return sampleprof_error::truncated;
 855
 856       if (HistVal != HIST_TYPE_INDIR_CALL_TOPN)
 857         return sampleprof_error::malformed;
 858
 859       uint64_t TargetIdx;
 860       if (!GcovBuffer.readInt64(TargetIdx))
 861         return sampleprof_error::truncated;
 862       StringRef TargetName(Names[TargetIdx]);
 863
 864       uint64_t TargetCount;
 865       if (!GcovBuffer.readInt64(TargetCount))
 866         return sampleprof_error::truncated;
 867
 868       if (Update)
 869         FProfile->addCalledTargetSamples(LineOffset, Discriminator,
 870                                          TargetName, TargetCount);
 871     }
 872   }
 873
 874   // Process all the inlined callers into the current function. These
 875   // are all the callsites that were inlined into this function.
 876   for (uint32_t I = 0; I < NumCallsites; I++) {
 877     // The offset is encoded as:
 878     //   high 16 bits: line offset to the start of the function.
 879     //   low 16 bits: discriminator.
 880     uint32_t Offset;
 881     if (!GcovBuffer.readInt(Offset))
 882       return sampleprof_error::truncated;
 883     InlineCallStack NewStack;
 884     NewStack.push_back(FProfile);
 885     NewStack.insert(NewStack.end(), InlineStack.begin(), InlineStack.end());
 886     if (std::error_code EC = readOneFunctionProfile(NewStack, Update, Offset))
 887       return EC;
 888   }
 889
 890   return sampleprof_error::success;
 891 }
 892
 893 /// Read a GCC AutoFDO profile.
 894 ///
 895 /// This format is generated by the Linux Perf conversion tool at
 896 /// https://github.com/google/autofdo.
 897 std::error_code SampleProfileReaderGCC::read() {
 898   // Read the string table.
 899   if (std::error_code EC = readNameTable())
 900     return EC;
 901
 902   // Read the source profile.
 903   if (std::error_code EC = readFunctionProfiles())
 904     return EC;
 905
 906   return sampleprof_error::success;
 907 }
 908
 909 bool SampleProfileReaderGCC::hasFormat(const MemoryBuffer &Buffer) {
 910   StringRef Magic(reinterpret_cast<const char *>(Buffer.getBufferStart()));
 911   return Magic == "adcg*704";
 912 }
 913
 914 std::error_code SampleProfileReaderItaniumRemapper::read() {
 915   // If the underlying data is in compact format, we can't remap it because
 916   // we don't know what the original function names were.
 917   if (getFormat() == SPF_Compact_Binary) {
 918     Ctx.diagnose(DiagnosticInfoSampleProfile(
 919         Buffer->getBufferIdentifier(),
 920         "Profile data remapping cannot be applied to profile data "
 921         "in compact format (original mangled names are not available).",
 922         DS_Warning));
 923     return sampleprof_error::success;
 924   }
 925
 926   if (Error E = Remappings.read(*Buffer)) {
 927     handleAllErrors(
 928         std::move(E), [&](const SymbolRemappingParseError &ParseError) {
 929           reportError(ParseError.getLineNum(), ParseError.getMessage());
 930         });
 931     return sampleprof_error::malformed;
 932   }
 933
 934   for (auto &Sample : getProfiles())
 935     if (auto Key = Remappings.insert(Sample.first()))
 936       SampleMap.insert({Key, &Sample.second});
 937
 938   return sampleprof_error::success;
 939 }
 940
 941 FunctionSamples *
 942 SampleProfileReaderItaniumRemapper::getSamplesFor(StringRef Fname) {
 943   if (auto Key = Remappings.lookup(Fname))
 944     return SampleMap.lookup(Key);
 945   return SampleProfileReader::getSamplesFor(Fname);
 946 }
 947
 948 /// Prepare a memory buffer for the contents of \p Filename.
 949 ///
 950 /// \returns an error code indicating the status of the buffer.
 951 static ErrorOr<std::unique_ptr<MemoryBuffer>>
 952 setupMemoryBuffer(const Twine &Filename) {
 953   auto BufferOrErr = MemoryBuffer::getFileOrSTDIN(Filename);
 954   if (std::error_code EC = BufferOrErr.getError())
 955     return EC;
 956   auto Buffer = std::move(BufferOrErr.get());
 957
 958   // Sanity check the file.
 959   if (uint64_t(Buffer->getBufferSize()) > std::numeric_limits<uint32_t>::max())
 960     return sampleprof_error::too_large;
 961
 962   return std::move(Buffer);
 963 }
 964
 965 /// Create a sample profile reader based on the format of the input file.
 966 ///
 967 /// \param Filename The file to open.
 968 ///
 969 /// \param C The LLVM context to use to emit diagnostics.
 970 ///
 971 /// \returns an error code indicating the status of the created reader.
 972 ErrorOr<std::unique_ptr<SampleProfileReader>>
 973 SampleProfileReader::create(const Twine &Filename, LLVMContext &C) {
 974   auto BufferOrError = setupMemoryBuffer(Filename);
 975   if (std::error_code EC = BufferOrError.getError())
 976     return EC;
 977   return create(BufferOrError.get(), C);
 978 }
 979
 980 /// Create a sample profile remapper from the given input, to remap the
 981 /// function names in the given profile data.
 982 ///
 983 /// \param Filename The file to open.
 984 ///
 985 /// \param C The LLVM context to use to emit diagnostics.
 986 ///
 987 /// \param Underlying The underlying profile data reader to remap.
 988 ///
 989 /// \returns an error code indicating the status of the created reader.
 990 ErrorOr<std::unique_ptr<SampleProfileReader>>
 991 SampleProfileReaderItaniumRemapper::create(
 992     const Twine &Filename, LLVMContext &C,
 993     std::unique_ptr<SampleProfileReader> Underlying) {
 994   auto BufferOrError = setupMemoryBuffer(Filename);
 995   if (std::error_code EC = BufferOrError.getError())
 996     return EC;
 997   return llvm::make_unique<SampleProfileReaderItaniumRemapper>(
 998       std::move(BufferOrError.get()), C, std::move(Underlying));
 999 }
1000
1001 /// Create a sample profile reader based on the format of the input data.
1002 ///
1003 /// \param B The memory buffer to create the reader from (assumes ownership).
1004 ///
1005 /// \param C The LLVM context to use to emit diagnostics.
1006 ///
1007 /// \returns an error code indicating the status of the created reader.
1008 ErrorOr<std::unique_ptr<SampleProfileReader>>
1009 SampleProfileReader::create(std::unique_ptr<MemoryBuffer> &B, LLVMContext &C) {
1010   std::unique_ptr<SampleProfileReader> Reader;
1011   if (SampleProfileReaderRawBinary::hasFormat(*B))
1012     Reader.reset(new SampleProfileReaderRawBinary(std::move(B), C));
1013   else if (SampleProfileReaderCompactBinary::hasFormat(*B))
1014     Reader.reset(new SampleProfileReaderCompactBinary(std::move(B), C));
1015   else if (SampleProfileReaderGCC::hasFormat(*B))
1016     Reader.reset(new SampleProfileReaderGCC(std::move(B), C));
1017   else if (SampleProfileReaderText::hasFormat(*B))
1018     Reader.reset(new SampleProfileReaderText(std::move(B), C));
1019   else
1020     return sampleprof_error::unrecognized_format;
1021
1022   FunctionSamples::Format = Reader->getFormat();
1023   if (std::error_code EC = Reader->readHeader())
1024     return EC;
1025
1026   return std::move(Reader);
1027 }
1028
1029 // For text and GCC file formats, we compute the summary after reading the
1030 // profile. Binary format has the profile summary in its header.
1031 void SampleProfileReader::computeSummary() {
1032   SampleProfileSummaryBuilder Builder(ProfileSummaryBuilder::DefaultCutoffs);
1033   for (const auto &I : Profiles) {
1034     const FunctionSamples &Profile = I.second;
1035     Builder.addRecord(Profile);
1036   }
1037   Summary = Builder.getSummary();
1038 }