[SampleFDO] Add symbol whitelist in the profile and use it when profile-sample-accurate is enabled

[wmi]

Summary

For sampleFDO, because the optimized build uses profile generated from previous release, we cannot tell a function without profile is truely cold or just newly created. Currently we treat them conservatively and put them in .text section instead of .text.unlikely. The result is when we persuing the best performance by locking .text.hot and .text in memory, and use huge pages for them, we waste a lot of memory and huge pages for functions actually being cold.

The patch add a new section to the profile. The section contains a list of function names showing up in the binary used to generate the current profile. During profile use compilation, when profile-sample-accurate is enabled, a function without profile will be regarded as cold only when it is contained in that list.

In order to add the symbol whitelist to the profile, I extend the profile format so it is easier to add a new section or extend an existing section in the future.

[davidxl]

Can you split the format extension change into a separate patch?

[wmi]

In D66374#1635363, @davidxl wrote:

Can you split the format extension change into a separate patch?

Ok, I will do that.

[mgrang]

lib/ProfileData/SampleProf.cpp

222

Please use a range-based llvm::sort function here. See https://llvm.org/docs/CodingStandards.html#beware-of-non-deterministic-sorting-order-of-equal-elements.

llvm::sort(SortedList);

245

Please use a range-based llvm::sort function here. See https://llvm.org/docs/CodingStandards.html#beware-of-non-deterministic-sorting-order-of-equal-elements.

llvm::sort(SortedList);

[wmi]

Split the patch of format extension into https://reviews.llvm.org/D66513. The symbol whitelist patch is on top of D66513 and will be sent out later.

[davidxl]

include/llvm/ProfileData/SampleProf.h

116

Name it 'SectEdgeProfile' or 'SecLBRProfile'?

Also move this after the WhiteList and reserve some enum range between WhiteList and the firstProfile

enum SecType {

SecProfileSummary = 0,
SecNameTable = 1,
SecWhiteList = 2,
SecFuncProfileFirst = 0x10,   // marker for the first type of profile.
SecLBRProfile = SecFuncProfileFirst

}

We should shoot for one profile type per section, but the hybrid mode you mentioned can also be supported if it is a size win.

[wmi]

In D66374#1638720, @wmi wrote:

Split the patch of format extension into https://reviews.llvm.org/D66513. The symbol whitelist patch is on top of D66513 and will be sent out later.

The second half is here: https://reviews.llvm.org/D66766

[mtrofin]

include/llvm/ProfileData/SampleProf.h

121

I'd initialize these to something, and perhaps have an invalid value for the Type you can use here for initialization. This type is used in a std::vector; so it'd be possible for one to write at a point "vector.resize()" and end up with garbage.

576

SymbolAllowList? (may be easier to understand)

580

Copy not copy

588

bool contains(..) const

same for size()

595

why not return size_t, which is what Syms.size() is?

597

Nit: why not just pass a StringRef with length CompressSize, and then rename UncompressSize to ExpectedUncompressedSize - I think it would clarify the responsibility of the function

604

is Syms expected to be large? IIUC, DenseSet is good for small sets, it's "quadratically probed". (https://llvm.org/docs/ProgrammersManual.html#llvm-adt-denseset-h)

605

can you add a comment explaining the role of Allocator? I think it answers my question about ownership below, too.

include/llvm/ProfileData/SampleProfReader.h

499

nit: I would find "AllowList" easier to understand (less term decoding)

include/llvm/ProfileData/SampleProfWriter.h

59

should this be passed in the ctor? also, if not, why virtual - would it ever do anything else than set a field (tradeoff: less virtual calls, for extra pointer field)

83

this could be const if Format were passed in the ctor - would Format ever mutate?

162

Please init field values - safe to do here, at definition (maintainability - wise)

lib/ProfileData/SampleProf.cpp

204

who takes ownership of Buffer?

lib/ProfileData/SampleProfReader.cpp

497

nit: define BufStart + Entry.Offset + EntrySize as EntryEnd, and then reuse below, too (easier to understand)

623

I not i

also, ++

why uint32_t and not size_t?

1175

why the braces? seems unrelated

lib/ProfileData/SampleProfWriter.cpp

287

I not i
++I
why not size_t?

same below

300

static_cast<uint64_t>?

304

i->I, ++I

312

static_cast?

490

would format passing make sense in the ctor of the profile writer? it's logically a const field of that, correct?

include/llvm/ProfileData/SampleProf.h

 
 #ifndef LLVM_PROFILEDATA_SAMPLEPROF_H
 #define LLVM_PROFILEDATA_SAMPLEPROF_H
 
 #include "llvm/ADT/DenseSet.h"
 #include "llvm/ADT/SmallVector.h"
 #include "llvm/ADT/StringMap.h"
 #include "llvm/ADT/StringRef.h"
+#include "llvm/ADT/StringSet.h"
 #include "llvm/IR/Function.h"
 #include "llvm/IR/GlobalValue.h"
 #include "llvm/IR/Module.h"
 #include "llvm/Support/Debug.h"
 #include "llvm/Support/ErrorOr.h"
 #include "llvm/Support/MathExtras.h"
+#include "llvm/Support/raw_ostream.h"
 #include <algorithm>
 #include <cstdint>
 #include <map>
 #include <string>
 #include <system_error>
 #include <utility>
 
 namespace llvm {
enum class sampleprof_error {
   too_large,
   truncated,
   malformed,
   unrecognized_format,
   unsupported_writing_format,
   truncated_name_table,
   not_implemented,
   counter_overflow,
-  ostream_seek_unsupported
+  ostream_seek_unsupported,
+  compress_failed,
+  uncompress_failed
 };
 
 inline std::error_code make_error_code(sampleprof_error E) {
   return std::error_code(static_cast<int>(E), sampleprof_category());
 }
 
 inline sampleprof_error MergeResult(sampleprof_error &Accumulator,
                                     sampleprof_error Result) {

 namespace llvm {
 namespace sampleprof {
 
 enum SampleProfileFormat {
   SPF_None = 0,
   SPF_Text = 0x1,
   SPF_Compact_Binary = 0x2,
   SPF_GCC = 0x3,
+  SPF_Ext_Binary = 0x4,
   SPF_Binary = 0xff
 };
 
 static inline uint64_t SPMagic(SampleProfileFormat Format = SPF_Binary) {
   return uint64_t('S') << (64 - 8) | uint64_t('P') << (64 - 16) |
          uint64_t('R') << (64 - 24) | uint64_t('O') << (64 - 32) |
          uint64_t('F') << (64 - 40) | uint64_t('4') << (64 - 48) |
          uint64_t('2') << (64 - 56) | uint64_t(Format);
 }
 
 // Get the proper representation of a string in the input Format.
 static inline StringRef getRepInFormat(StringRef Name,
                                        SampleProfileFormat Format,
                                        std::string &GUIDBuf) {
   if (Name.empty())
     return Name;
   GUIDBuf = std::to_string(Function::getGUID(Name));
   return (Format == SPF_Compact_Binary) ? StringRef(GUIDBuf) : Name;
 }
 
 static inline uint64_t SPVersion() { return 103; }
 
+enum SecType {
+  SecProfSummary = 0,
+  SecNameTable = 1,
+  SecFuncProfile = 2,
+  SecSymbolWhiteList = 3
+};
+
+struct SecHdrTableEntry {
+  SecType Type;
+  uint64_t Flag;
+  uint64_t Offset;
+  uint64_t Size;
+};
+
 /// Represents the relative location of an instruction.
 ///
 /// Instruction locations are specified by the line offset from the
 /// beginning of the function (marked by the line where the function
 /// header is) and the discriminator value within that line.
 ///
 /// The discriminator value is useful to distinguish instructions
 /// that are on the same line but belong to different basic blocks
public:
   }
 
   const SamplesWithLocList &get() const { return V; }
 
 private:
   SamplesWithLocList V;
 };
 
+/// SymbolWhiteList records the list of function symbols shown up
+/// in the binary used to generate the profile. It is useful to
+/// to discriminate a function being so cold as not to shown up
+/// in the profile and a function newly added.
+class SymbolWhiteList {
+public:
+  /// copy indicates whether we need to copy the underlying memory
+  /// for the input Name.
+  void add(StringRef Name, bool copy = false) {
+    if (!copy) {
+      Syms.insert(Name);
+      return;
+    }
+    Syms.insert(Name.copy(Allocator));
+  }
+
+  bool contains(StringRef Name) { return Syms.count(Name); }
+
+  void merge(const SymbolWhiteList &List) {
+    for (auto Sym : List.Syms)
+      add(Sym, true);
+  }
+
+  unsigned size() { return Syms.size(); }
+
+  std::error_code read(uint64_t CompressSize, uint64_t UncompressSize,
+                       const uint8_t *Data);
+  std::error_code write(raw_ostream &OS);
+  void print(raw_ostream &OS) const;
+  void dump() const;
+
+private:
+  DenseSet<StringRef> Syms;
+  BumpPtrAllocator Allocator;
+};
+
 } // end namespace sampleprof
 } // end namespace llvm
 
 #endif // LLVM_PROFILEDATA_SAMPLEPROF_H

include/llvm/ProfileData/SampleProfReader.h

public:
   create(std::unique_ptr<MemoryBuffer> &B, LLVMContext &C);
 
   /// Return the profile summary.
   ProfileSummary &getSummary() { return *(Summary.get()); }
 
   /// \brief Return the profile format.
   SampleProfileFormat getFormat() { return Format; }
 
+  virtual std::unique_ptr<SymbolWhiteList> getSymbolWhiteList() {
+    return nullptr;
+  };
+
 protected:
   /// Map every function to its associated profile.
   ///
   /// The profile of every function executed at runtime is collected
   /// in the structure FunctionSamples. This maps function objects
   /// to their corresponding profiles.
   StringMap<FunctionSamples> Profiles;
 
protected:
   bool at_eof() const { return Data >= End; }
 
   /// Read the next function profile instance.
   std::error_code readFuncProfile();
 
   /// Read the contents of the given profile instance.
   std::error_code readProfile(FunctionSamples &FProfile);
 
+  /// Read the contents of Magic number and Version number.
+  std::error_code readMagicIdent();
+
+  /// Read profile summary.
+  std::error_code readSummary();
+
+  /// Read the whole name table.
+  virtual std::error_code readNameTable() = 0;
+
   /// Points to the current location in the buffer.
   const uint8_t *Data = nullptr;
 
   /// Points to the end of the buffer.
   const uint8_t *End = nullptr;
 
 private:
   std::error_code readSummaryEntry(std::vector<ProfileSummaryEntry> &Entries);
   virtual std::error_code verifySPMagic(uint64_t Magic) = 0;
 
-  /// Read profile summary.
-  std::error_code readSummary();
-
-  /// Read the whole name table.
-  virtual std::error_code readNameTable() = 0;
-
   /// Read a string indirectly via the name table.
   virtual ErrorOr<StringRef> readStringFromTable() = 0;
 };
 
 class SampleProfileReaderRawBinary : public SampleProfileReaderBinary {
 private:
   /// Function name table.
   std::vector<StringRef> NameTable;
   virtual std::error_code verifySPMagic(uint64_t Magic) override;
-  virtual std::error_code readNameTable() override;
   /// Read a string indirectly via the name table.
   virtual ErrorOr<StringRef> readStringFromTable() override;
 
+protected:
+  virtual std::error_code readNameTable() override;
+
 public:
-  SampleProfileReaderRawBinary(std::unique_ptr<MemoryBuffer> B, LLVMContext &C)
-      : SampleProfileReaderBinary(std::move(B), C, SPF_Binary) {}
+  SampleProfileReaderRawBinary(std::unique_ptr<MemoryBuffer> B, LLVMContext &C,
+                               SampleProfileFormat Format = SPF_Binary)
+      : SampleProfileReaderBinary(std::move(B), C, Format) {}
 
   /// \brief Return true if \p Buffer is in the format supported by this class.
   static bool hasFormat(const MemoryBuffer &Buffer);
 };
 
+/// SampleProfileReaderExtBinary is a binary format which is more easily to be
+/// extend than SampleProfileReaderRawBinary.
+/// SampleProfileReaderExtBinary format is organized in sections except the
+/// magic and version number at the beginning. There is a section table before
+/// all the sections, and each entry in the table describes the entry type,
+/// start, size and attributes. The format in each section is defined by the
+/// section itself.
+/// It is easy to add a new section while maintaining the backward
+/// compatibility of the profile. Nothing extra needs to be done. If we want
+/// to extend an existing section, like add cache misses information in
+/// addition to the sample count in the profile body, we can add a new section
+/// with the extension and retire the existing section, and we could choose
+/// to keep the parser of the old section if we want the reader to be able
+/// to read both new and old format profile.
+class SampleProfileReaderExtBinary : public SampleProfileReaderRawBinary {
+private:
+  std::vector<SecHdrTableEntry> SecHdrTable;
+  std::unique_ptr<SymbolWhiteList> SymWhiteList;
+  std::error_code readSecHdrTableEntry();
+  std::error_code readSecHdrTable();
+  virtual std::error_code readHeader() override;
+  virtual std::error_code verifySPMagic(uint64_t Magic) override;
+
+public:
+  SampleProfileReaderExtBinary(std::unique_ptr<MemoryBuffer> B, LLVMContext &C)
+      : SampleProfileReaderRawBinary(std::move(B), C, SPF_Ext_Binary) {}
+
+  /// \brief Return true if \p Buffer is in the format supported by this class.
+  static bool hasFormat(const MemoryBuffer &Buffer);
+
+  /// Read sample profiles in extensible format from the associated file.
+  std::error_code read() override;
+
+  std::error_code readSymbolWhiteList();
+
+  virtual std::unique_ptr<SymbolWhiteList> getSymbolWhiteList() override {
+    return std::move(SymWhiteList);
+  };
+};
+
 class SampleProfileReaderCompactBinary : public SampleProfileReaderBinary {
 private:
   /// Function name table.
   std::vector<std::string> NameTable;
   /// The table mapping from function name to the offset of its FunctionSample
   /// towards file start.
   DenseMap<StringRef, uint64_t> FuncOffsetTable;
   /// The set containing the functions to use when compiling a module.

include/llvm/ProfileData/SampleProfWriter.h

 /// Sample-based profile writer. Base class.
 class SampleProfileWriter {
 public:
   virtual ~SampleProfileWriter() = default;
 
   /// Write sample profiles in \p S.
   ///
   /// \returns status code of the file update operation.
-  virtual std::error_code write(const FunctionSamples &S) = 0;
+  virtual std::error_code writeSample(const FunctionSamples &S) = 0;
 
   /// Write all the sample profiles in the given map of samples.
   ///
   /// \returns status code of the file update operation.
   virtual std::error_code write(const StringMap<FunctionSamples> &ProfileMap);
 
   raw_ostream &getOutputStream() { return *OutputStream; }
 
   /// Profile writer factory.
   ///
   /// Create a new file writer based on the value of \p Format.
   static ErrorOr<std::unique_ptr<SampleProfileWriter>>
   create(StringRef Filename, SampleProfileFormat Format);
 
   /// Create a new stream writer based on the value of \p Format.
   /// For testing.
   static ErrorOr<std::unique_ptr<SampleProfileWriter>>
   create(std::unique_ptr<raw_ostream> &OS, SampleProfileFormat Format);
 
+  virtual void setSymbolWhiteList(SymbolWhiteList *WL) {}
+
 protected:
   SampleProfileWriter(std::unique_ptr<raw_ostream> &OS)
       : OutputStream(std::move(OS)) {}
 
   /// Write a file header for the profile file.
   virtual std::error_code
   writeHeader(const StringMap<FunctionSamples> &ProfileMap) = 0;
 
+  // Write function profiles to the profile file.
+  virtual std::error_code
+  writeFuncProfiles(const StringMap<FunctionSamples> &ProfileMap);
+
   /// Output stream where to emit the profile to.
   std::unique_ptr<raw_ostream> OutputStream;
 
   /// Profile summary.
   std::unique_ptr<ProfileSummary> Summary;
 
   /// Compute summary for this profile.
   void computeSummary(const StringMap<FunctionSamples> &ProfileMap);
+
+  /// Profile format.
+  SampleProfileFormat Format;
 };
 
 /// Sample-based profile writer (text format).
 class SampleProfileWriterText : public SampleProfileWriter {
 public:
-  std::error_code write(const FunctionSamples &S) override;
+  std::error_code writeSample(const FunctionSamples &S) override;
 
 protected:
   SampleProfileWriterText(std::unique_ptr<raw_ostream> &OS)
       : SampleProfileWriter(OS), Indent(0) {}
 
   std::error_code
   writeHeader(const StringMap<FunctionSamples> &ProfileMap) override {
     return sampleprof_error::success;
   }
 
 private:
   /// Indent level to use when writing.
   ///
   /// This is used when printing inlined callees.
   unsigned Indent;
 
   friend ErrorOr<std::unique_ptr<SampleProfileWriter>>
   SampleProfileWriter::create(std::unique_ptr<raw_ostream> &OS,
                               SampleProfileFormat Format);
 };
 
 /// Sample-based profile writer (binary format).
 class SampleProfileWriterBinary : public SampleProfileWriter {
 public:
-  virtual std::error_code write(const FunctionSamples &S) override;
+  virtual std::error_code writeSample(const FunctionSamples &S) override;
   SampleProfileWriterBinary(std::unique_ptr<raw_ostream> &OS)
       : SampleProfileWriter(OS) {}
 
 protected:
   virtual std::error_code writeNameTable() = 0;
-  virtual std::error_code writeMagicIdent() = 0;
+  virtual std::error_code writeMagicIdent(SampleProfileFormat Format);
   virtual std::error_code
   writeHeader(const StringMap<FunctionSamples> &ProfileMap) override;
   std::error_code writeSummary();
   std::error_code writeNameIdx(StringRef FName);
   std::error_code writeBody(const FunctionSamples &S);
   inline void stablizeNameTable(std::set<StringRef> &V);
 
   MapVector<StringRef, uint32_t> NameTable;
 
-private:
   void addName(StringRef FName);
   void addNames(const FunctionSamples &S);
 
+private:
   friend ErrorOr<std::unique_ptr<SampleProfileWriter>>
   SampleProfileWriter::create(std::unique_ptr<raw_ostream> &OS,
                               SampleProfileFormat Format);
 };
 
 class SampleProfileWriterRawBinary : public SampleProfileWriterBinary {
   using SampleProfileWriterBinary::SampleProfileWriterBinary;
 
 protected:
   virtual std::error_code writeNameTable() override;
-  virtual std::error_code writeMagicIdent() override;
+};
+
+class SampleProfileWriterExtBinary : public SampleProfileWriterRawBinary {
+  using SampleProfileWriterRawBinary::SampleProfileWriterRawBinary;
+
+public:
+  virtual std::error_code
+  write(const StringMap<FunctionSamples> &ProfileMap) override;
+  virtual void setSymbolWhiteList(SymbolWhiteList *WL) override {
+    SymWhiteList = WL;
+  };
+
+private:
+  void allocSecHdrTable();
+  std::error_code writeSecHdrTable();
+  virtual std::error_code
+  writeHeader(const StringMap<FunctionSamples> &ProfileMap) override;
+
+  uint64_t FileStart;
+  uint64_t SecHdrTableOffset;
+  std::vector<SecHdrTableEntry> SecHdrTable;
+  SymbolWhiteList *SymWhiteList = nullptr;
 };
 
 // CompactBinary is a compact format of binary profile which both reduces
 // the profile size and the load time needed when compiling. It has two
 // major difference with Binary format.
 // 1. It represents all the strings in name table using md5 hash.
 // 2. It saves a function offset table which maps function name index to
 // the offset of its function profile to the start of the binary profile,

 //             function3 name index --> function3 profile start
 //
 // We need Part2 because profile reader can use it to find out and read
 // function offset table without reading Part3 first.
 class SampleProfileWriterCompactBinary : public SampleProfileWriterBinary {
   using SampleProfileWriterBinary::SampleProfileWriterBinary;
 
 public:
-  virtual std::error_code write(const FunctionSamples &S) override;
+  virtual std::error_code writeSample(const FunctionSamples &S) override;
   virtual std::error_code
   write(const StringMap<FunctionSamples> &ProfileMap) override;
 
 protected:
   /// The table mapping from function name to the offset of its FunctionSample
   /// towards profile start.
   MapVector<StringRef, uint64_t> FuncOffsetTable;
   /// The offset of the slot to be filled with the offset of FuncOffsetTable
   /// towards profile start.
   uint64_t TableOffset;
   virtual std::error_code writeNameTable() override;
-  virtual std::error_code writeMagicIdent() override;
   virtual std::error_code
   writeHeader(const StringMap<FunctionSamples> &ProfileMap) override;
   std::error_code writeFuncOffsetTable();
 };
 
 } // end namespace sampleprof
 } // end namespace llvm
 
 #endif // LLVM_PROFILEDATA_SAMPLEPROFWRITER_H

lib/ProfileData/SampleProf.cpp

 // sample profile data.
 //
 //===----------------------------------------------------------------------===//
 
 #include "llvm/ProfileData/SampleProf.h"
 #include "llvm/Config/llvm-config.h"
 #include "llvm/IR/DebugInfoMetadata.h"
 #include "llvm/Support/Compiler.h"
+#include "llvm/Support/Compression.h"
 #include "llvm/Support/Debug.h"
+#include "llvm/Support/Error.h"
 #include "llvm/Support/ErrorHandling.h"
+#include "llvm/Support/LEB128.h"
 #include "llvm/Support/ManagedStatic.h"
 #include "llvm/Support/raw_ostream.h"
 #include <string>
 #include <system_error>
 
 using namespace llvm;
 using namespace sampleprof;
 
std::string message(int IE) const override {
     case sampleprof_error::truncated_name_table:
       return "Truncated function name table";
     case sampleprof_error::not_implemented:
       return "Unimplemented feature";
     case sampleprof_error::counter_overflow:
       return "Counter overflow";
     case sampleprof_error::ostream_seek_unsupported:
       return "Ostream does not support seek";
+    case sampleprof_error::compress_failed:
+      return "Compress failure";
+    case sampleprof_error::uncompress_failed:
+      return "Uncompress failure";
     }
     llvm_unreachable("A value of sampleprof_error has no message.");
   }
 };
 
 } // end anonymous namespace
 
 static ManagedStatic<SampleProfErrorCategoryType> ErrorCategory;
for (int i = S.size() - 1; i >= 0 && FS != nullptr; i--) {
     FS = FS->findFunctionSamplesAt(S[i].first, S[i].second);
   }
   return FS;
 }
 
 #if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
 LLVM_DUMP_METHOD void FunctionSamples::dump() const { print(dbgs(), 0); }
 #endif
+
+std::error_code SymbolWhiteList::read(uint64_t CompressSize,
+                                      uint64_t UncompressSize,
+                                      const uint8_t *Data) {
+  StringRef CompressedStrings(reinterpret_cast<const char *>(Data),
+                              CompressSize);
+  char *Buffer = Allocator.Allocate<char>(UncompressSize);
+  llvm::Error E = zlib::uncompress(CompressedStrings, Buffer, UncompressSize);
+  if (E)
+    return sampleprof_error::uncompress_failed;
+  uint64_t Size = 0;
+  while (Size < UncompressSize) {
+    StringRef Str(Buffer + Size);
+    add(Str);
+    Size += Str.size() + 1;
+  }
+  return sampleprof_error::success;
+}
+
+std::error_code SymbolWhiteList::write(raw_ostream &OS) {
+  // Sort the symbols before doing compression. It will make the
+  // compression much more effective.
+  std::vector<StringRef> SortedList;
+  SortedList.insert(SortedList.begin(), Syms.begin(), Syms.end());
+  std::sort(SortedList.begin(), SortedList.end());
+
+  std::string UncompressedStrings;
+  for (auto &Sym : SortedList) {
+    UncompressedStrings.append(Sym.str());
+    UncompressedStrings.append(1, '\0');
+  }
+
+  SmallString<128> CompressedStrings;
+  llvm::Error E = zlib::compress(UncompressedStrings, CompressedStrings,
+                                 zlib::BestSizeCompression);
+  if (E)
+    return sampleprof_error::compress_failed;
+  encodeULEB128(UncompressedStrings.size(), OS);
+  encodeULEB128(CompressedStrings.size(), OS);
+  OS << CompressedStrings.str();
+  return sampleprof_error::success;
+}
+
+void SymbolWhiteList::print(raw_ostream &OS) const {
+  OS << "======== Dump symbol whitelist ========\n";
+  std::vector<StringRef> SortedList;
+  SortedList.insert(SortedList.begin(), Syms.begin(), Syms.end());
+  std::sort(SortedList.begin(), SortedList.end());
+
+  for (auto &Sym : SortedList)
+    OS << Sym << "\n";
+}
+
+void SymbolWhiteList::dump() const { print(dbgs()); }

lib/ProfileData/SampleProfReader.cpp

std::error_code SampleProfileReaderBinary::read() {
   while (!at_eof()) {
     if (std::error_code EC = readFuncProfile())
       return EC;
   }
 
   return sampleprof_error::success;
 }
 
+std::error_code SampleProfileReaderExtBinary::readSymbolWhiteList() {
+  auto UncompressSize = readNumber<uint64_t>();
+  if (std::error_code EC = UncompressSize.getError())
+    return EC;
+
+  auto CompressSize = readNumber<uint64_t>();
+  if (std::error_code EC = CompressSize.getError())
+    return EC;
+
+  if (!SymWhiteList)
+    SymWhiteList = std::make_unique<SymbolWhiteList>();
+
+  if (std::error_code EC =
+          SymWhiteList->read(*CompressSize, *UncompressSize, Data))
+    return EC;
+  Data = Data + *CompressSize;
+  return sampleprof_error::success;
+}
+
+std::error_code SampleProfileReaderExtBinary::read() {
+  const uint8_t *BufStart =
+      reinterpret_cast<const uint8_t *>(Buffer->getBufferStart());
+
+  for (auto &Entry : SecHdrTable) {
+    Data = BufStart + Entry.Offset;
+    switch (Entry.Type) {
+    case SecFuncProfile:
+      while (Data < BufStart + Entry.Offset + Entry.Size) {
+        if (std::error_code EC = readFuncProfile())
+          return EC;
+      }
+      break;
+    case SecSymbolWhiteList:
+      if (std::error_code EC = readSymbolWhiteList())
+        return EC;
+      break;
+    default:
+      continue;
+    }
+    if (Data != BufStart + Entry.Offset + Entry.Size)
+      return sampleprof_error::malformed;
+  }
+
+  return sampleprof_error::success;
+}
+
 std::error_code SampleProfileReaderCompactBinary::read() {
   std::vector<uint64_t> OffsetsToUse;
   if (UseAllFuncs) {
     for (auto FuncEntry : FuncOffsetTable) {
       OffsetsToUse.push_back(FuncEntry.second);
     }
   }
   else {

 }
 
 std::error_code SampleProfileReaderRawBinary::verifySPMagic(uint64_t Magic) {
   if (Magic == SPMagic())
     return sampleprof_error::success;
   return sampleprof_error::bad_magic;
 }
 
+std::error_code SampleProfileReaderExtBinary::verifySPMagic(uint64_t Magic) {
+  if (Magic == SPMagic(SPF_Ext_Binary))
+    return sampleprof_error::success;
+  return sampleprof_error::bad_magic;
+}
+
 std::error_code
 SampleProfileReaderCompactBinary::verifySPMagic(uint64_t Magic) {
   if (Magic == SPMagic(SPF_Compact_Binary))
     return sampleprof_error::success;
   return sampleprof_error::bad_magic;
 }
 
 std::error_code SampleProfileReaderRawBinary::readNameTable() {
for (uint32_t I = 0; I < *Size; ++I) {
     auto FID = readNumber<uint64_t>();
     if (std::error_code EC = FID.getError())
       return EC;
     NameTable.push_back(std::to_string(*FID));
   }
   return sampleprof_error::success;
 }
 
-std::error_code SampleProfileReaderBinary::readHeader() {
-  Data = reinterpret_cast<const uint8_t *>(Buffer->getBufferStart());
-  End = Data + Buffer->getBufferSize();
+std::error_code SampleProfileReaderExtBinary::readSecHdrTableEntry() {
+  SecHdrTableEntry Entry;
+  auto Type = readUnencodedNumber<uint64_t>();
+  if (std::error_code EC = Type.getError())
+    return EC;
+  Entry.Type = static_cast<SecType>(*Type);
+
+  auto Flag = readUnencodedNumber<uint64_t>();
+  if (std::error_code EC = Flag.getError())
+    return EC;
+  Entry.Flag = *Flag;
+
+  auto Offset = readUnencodedNumber<uint64_t>();
+  if (std::error_code EC = Offset.getError())
+    return EC;
+  Entry.Offset = *Offset;
+
+  auto Size = readUnencodedNumber<uint64_t>();
+  if (std::error_code EC = Size.getError())
+    return EC;
+  Entry.Size = *Size;
+
+  SecHdrTable.push_back(std::move(Entry));
+  return sampleprof_error::success;
+}
+
+std::error_code SampleProfileReaderExtBinary::readSecHdrTable() {
+  auto EntryNum = readUnencodedNumber<uint64_t>();
+  if (std::error_code EC = EntryNum.getError())
+    return EC;
+
+  for (uint32_t i = 0; i < (*EntryNum); i++)
+    if (std::error_code EC = readSecHdrTableEntry())
+      return EC;
+
+  return sampleprof_error::success;
+}
+
+std::error_code SampleProfileReaderExtBinary::readHeader() {
+  const uint8_t *BufStart =
+      reinterpret_cast<const uint8_t *>(Buffer->getBufferStart());
+  Data = BufStart;
+  End = BufStart + Buffer->getBufferSize();
+
+  if (std::error_code EC = readMagicIdent())
+    return EC;
+
+  if (std::error_code EC = readSecHdrTable())
+    return EC;
+
+  for (auto &Entry : SecHdrTable) {
+    Data = BufStart + Entry.Offset;
+    switch (Entry.Type) {
+    case SecProfSummary:
+      if (std::error_code EC = readSummary())
+        return EC;
+      break;
+    case SecNameTable:
+      if (std::error_code EC = readNameTable())
+        return EC;
+      break;
+    default:
+      continue;
+    }
+    if (Data != BufStart + Entry.Offset + Entry.Size)
+      return sampleprof_error::malformed;
+  }
+
+  return sampleprof_error::success;
+}
 
+std::error_code SampleProfileReaderBinary::readMagicIdent() {
   // Read and check the magic identifier.
   auto Magic = readNumber<uint64_t>();
   if (std::error_code EC = Magic.getError())
     return EC;
   else if (std::error_code EC = verifySPMagic(*Magic))
     return EC;
 
   // Read the version number.
   auto Version = readNumber<uint64_t>();
   if (std::error_code EC = Version.getError())
     return EC;
   else if (*Version != SPVersion())
     return sampleprof_error::unsupported_version;
 
+  return sampleprof_error::success;
+}
+
+std::error_code SampleProfileReaderBinary::readHeader() {
+  Data = reinterpret_cast<const uint8_t *>(Buffer->getBufferStart());
+  End = Data + Buffer->getBufferSize();
+
+  if (std::error_code EC = readMagicIdent())
+    return EC;
+
   if (std::error_code EC = readSummary())
     return EC;
 
   if (std::error_code EC = readNameTable())
     return EC;
   return sampleprof_error::success;
 }
 

 
 bool SampleProfileReaderRawBinary::hasFormat(const MemoryBuffer &Buffer) {
   const uint8_t *Data =
       reinterpret_cast<const uint8_t *>(Buffer.getBufferStart());
   uint64_t Magic = decodeULEB128(Data);
   return Magic == SPMagic();
 }
 
+bool SampleProfileReaderExtBinary::hasFormat(const MemoryBuffer &Buffer) {
+  const uint8_t *Data =
+      reinterpret_cast<const uint8_t *>(Buffer.getBufferStart());
+  uint64_t Magic = decodeULEB128(Data);
+  return Magic == SPMagic(SPF_Ext_Binary);
+}
+
 bool SampleProfileReaderCompactBinary::hasFormat(const MemoryBuffer &Buffer) {
   const uint8_t *Data =
       reinterpret_cast<const uint8_t *>(Buffer.getBufferStart());
   uint64_t Magic = decodeULEB128(Data);
   return Magic == SPMagic(SPF_Compact_Binary);
 }
 
 std::error_code SampleProfileReaderGCC::skipNextWord() {

 /// \param C The LLVM context to use to emit diagnostics.
 ///
 /// \returns an error code indicating the status of the created reader.
 ErrorOr<std::unique_ptr<SampleProfileReader>>
 SampleProfileReader::create(std::unique_ptr<MemoryBuffer> &B, LLVMContext &C) {
   std::unique_ptr<SampleProfileReader> Reader;
   if (SampleProfileReaderRawBinary::hasFormat(*B))
     Reader.reset(new SampleProfileReaderRawBinary(std::move(B), C));
+  else if (SampleProfileReaderExtBinary::hasFormat(*B))
+    Reader.reset(new SampleProfileReaderExtBinary(std::move(B), C));
   else if (SampleProfileReaderCompactBinary::hasFormat(*B))
     Reader.reset(new SampleProfileReaderCompactBinary(std::move(B), C));
   else if (SampleProfileReaderGCC::hasFormat(*B))
     Reader.reset(new SampleProfileReaderGCC(std::move(B), C));
   else if (SampleProfileReaderText::hasFormat(*B))
     Reader.reset(new SampleProfileReaderText(std::move(B), C));
   else
     return sampleprof_error::unrecognized_format;
 
   FunctionSamples::Format = Reader->getFormat();
-  if (std::error_code EC = Reader->readHeader())
+  if (std::error_code EC = Reader->readHeader()) {
     return EC;
+  }
 
   return std::move(Reader);
 }
 
 // For text and GCC file formats, we compute the summary after reading the
 // profile. Binary format has the profile summary in its header.
 void SampleProfileReader::computeSummary() {
   SampleProfileSummaryBuilder Builder(ProfileSummaryBuilder::DefaultCutoffs);
   for (const auto &I : Profiles) {
     const FunctionSamples &Profile = I.second;
     Builder.addRecord(Profile);
   }
   Summary = Builder.getSummary();
 }

lib/ProfileData/SampleProfWriter.cpp

 #include <set>
 #include <system_error>
 #include <utility>
 #include <vector>
 
 using namespace llvm;
 using namespace sampleprof;
 
-std::error_code
-SampleProfileWriter::write(const StringMap<FunctionSamples> &ProfileMap) {
-  if (std::error_code EC = writeHeader(ProfileMap))
-    return EC;
-
+std::error_code SampleProfileWriter::writeFuncProfiles(
+    const StringMap<FunctionSamples> &ProfileMap) {
   // Sort the ProfileMap by total samples.
   typedef std::pair<StringRef, const FunctionSamples *> NameFunctionSamples;
   std::vector<NameFunctionSamples> V;
   for (const auto &I : ProfileMap)
     V.push_back(std::make_pair(I.getKey(), &I.second));
 
   llvm::stable_sort(
       V, [](const NameFunctionSamples &A, const NameFunctionSamples &B) {
         if (A.second->getTotalSamples() == B.second->getTotalSamples())
           return A.first > B.first;
         return A.second->getTotalSamples() > B.second->getTotalSamples();
       });
 
   for (const auto &I : V) {
-    if (std::error_code EC = write(*I.second))
+    if (std::error_code EC = writeSample(*I.second))
       return EC;
   }
   return sampleprof_error::success;
 }
 
+std::error_code
+SampleProfileWriter::write(const StringMap<FunctionSamples> &ProfileMap) {
+  if (std::error_code EC = writeHeader(ProfileMap))
+    return EC;
+
+  if (std::error_code EC = writeFuncProfiles(ProfileMap))
+    return EC;
+
+  return sampleprof_error::success;
+}
+
+std::error_code SampleProfileWriterExtBinary::write(
+    const StringMap<FunctionSamples> &ProfileMap) {
+  if (std::error_code EC = writeHeader(ProfileMap))
+    return EC;
+
+  auto &OS = *OutputStream;
+  uint64_t LastSecEnd = OS.tell();
+  if (std::error_code EC = writeFuncProfiles(ProfileMap))
+    return EC;
+  SecHdrTable.push_back(
+      {SecFuncProfile, 0, LastSecEnd - FileStart, OS.tell() - LastSecEnd});
+
+  if (SymWhiteList) {
+    LastSecEnd = OS.tell();
+    if (std::error_code EC = SymWhiteList->write(OS))
+      return EC;
+    SecHdrTable.push_back({SecSymbolWhiteList, 0, LastSecEnd - FileStart,
+                           OS.tell() - LastSecEnd});
+  }
+
+  if (std::error_code EC = writeSecHdrTable())
+    return EC;
+
+  return sampleprof_error::success;
+}
+
 std::error_code SampleProfileWriterCompactBinary::write(
     const StringMap<FunctionSamples> &ProfileMap) {
   if (std::error_code EC = SampleProfileWriter::write(ProfileMap))
     return EC;
   if (std::error_code EC = writeFuncOffsetTable())
     return EC;
   return sampleprof_error::success;
 }
 
 /// Write samples to a text file.
 ///
 /// Note: it may be tempting to implement this in terms of
 /// FunctionSamples::print().  Please don't.  The dump functionality is intended
 /// for debugging and has no specified form.
 ///
 /// The format used here is more structured and deliberate because
 /// it needs to be parsed by the SampleProfileReaderText class.
-std::error_code SampleProfileWriterText::write(const FunctionSamples &S) {
+std::error_code SampleProfileWriterText::writeSample(const FunctionSamples &S) {
   auto &OS = *OutputStream;
   OS << S.getName() << ":" << S.getTotalSamples();
   if (Indent == 0)
     OS << ":" << S.getHeadSamples();
   OS << "\n";
 
   SampleSorter<LineLocation, SampleRecord> SortedSamples(S.getBodySamples());
   for (const auto &I : SortedSamples.get()) {
for (const auto &I : SortedCallsiteSamples.get())
     for (const auto &FS : I->second) {
       LineLocation Loc = I->first;
       const FunctionSamples &CalleeSamples = FS.second;
       OS.indent(Indent);
       if (Loc.Discriminator == 0)
         OS << Loc.LineOffset << ": ";
       else
         OS << Loc.LineOffset << "." << Loc.Discriminator << ": ";
-      if (std::error_code EC = write(CalleeSamples))
+      if (std::error_code EC = writeSample(CalleeSamples))
         return EC;
     }
   Indent -= 1;
 
   return sampleprof_error::success;
 }
 
 std::error_code SampleProfileWriterBinary::writeNameIdx(StringRef FName) {
std::error_code SampleProfileWriterCompactBinary::writeNameTable() {
   // Write out the name table.
   encodeULEB128(NameTable.size(), OS);
   for (auto N : V) {
     encodeULEB128(MD5Hash(N), OS);
   }
   return sampleprof_error::success;
 }
 
-std::error_code SampleProfileWriterRawBinary::writeMagicIdent() {
+std::error_code
+SampleProfileWriterBinary::writeMagicIdent(SampleProfileFormat Format) {
   auto &OS = *OutputStream;
   // Write file magic identifier.
-  encodeULEB128(SPMagic(), OS);
+  encodeULEB128(SPMagic(Format), OS);
   encodeULEB128(SPVersion(), OS);
   return sampleprof_error::success;
 }
 
-std::error_code SampleProfileWriterCompactBinary::writeMagicIdent() {
-  auto &OS = *OutputStream;
-  // Write file magic identifier.
-  encodeULEB128(SPMagic(SPF_Compact_Binary), OS);
-  encodeULEB128(SPVersion(), OS);
+std::error_code SampleProfileWriterBinary::writeHeader(
+    const StringMap<FunctionSamples> &ProfileMap) {
+  writeMagicIdent(Format);
+
+  computeSummary(ProfileMap);
+  if (auto EC = writeSummary())
+    return EC;
+
+  // Generate the name table for all the functions referenced in the profile.
+  for (const auto &I : ProfileMap) {
+    addName(I.first());
+    addNames(I.second);
+  }
+
+  writeNameTable();
   return sampleprof_error::success;
 }
 
-std::error_code SampleProfileWriterBinary::writeHeader(
+SecType SecTable[] = {SecProfSummary, SecNameTable, SecFuncProfile,
+                      SecSymbolWhiteList};
+
+void SampleProfileWriterExtBinary::allocSecHdrTable() {
+  support::endian::Writer Writer(*OutputStream, support::little);
+
+  uint32_t SecNumber = sizeof(SecTable) / sizeof(SecType);
+  Writer.write(static_cast<uint64_t>(SecNumber));
+  SecHdrTableOffset = OutputStream->tell();
+  for (uint32_t i = 0; i < SecNumber; i++) {
+    Writer.write(static_cast<uint64_t>(-1));
+    Writer.write(static_cast<uint64_t>(-1));
+    Writer.write(static_cast<uint64_t>(-1));
+    Writer.write(static_cast<uint64_t>(-1));
+  }
+}
+
+std::error_code SampleProfileWriterExtBinary::writeSecHdrTable() {
+  auto &OFS = static_cast<raw_fd_ostream &>(*OutputStream);
+  uint64_t Saved = OutputStream->tell();
+
+  // Set OutputStream to the location saved in SecHdrTableOffset.
+  if (OFS.seek(SecHdrTableOffset) == (uint64_t)-1)
+    return sampleprof_error::ostream_seek_unsupported;
+  support::endian::Writer Writer(*OutputStream, support::little);
+
+  for (uint32_t i = 0; i < SecHdrTable.size(); i++) {
+    Writer.write(static_cast<uint64_t>(SecHdrTable[i].Type));
+    Writer.write(static_cast<uint64_t>(SecHdrTable[i].Flag));
+    Writer.write(static_cast<uint64_t>(SecHdrTable[i].Offset));
+    Writer.write(static_cast<uint64_t>(SecHdrTable[i].Size));
+  }
+
+  // Reset OutputStream.
+  if (OFS.seek(Saved) == (uint64_t)-1)
+    return sampleprof_error::ostream_seek_unsupported;
+
+  return sampleprof_error::success;
+}
+
+std::error_code SampleProfileWriterExtBinary::writeHeader(
     const StringMap<FunctionSamples> &ProfileMap) {
-  writeMagicIdent();
+  auto &OS = *OutputStream;
+  FileStart = OS.tell();
+  writeMagicIdent(Format);
+
+  allocSecHdrTable();
+  uint64_t LastSecEnd = OS.tell();
 
   computeSummary(ProfileMap);
   if (auto EC = writeSummary())
     return EC;
+  uint64_t SecEnd = OS.tell();
+  SecHdrTable.push_back(
+      {SecProfSummary, 0, LastSecEnd - FileStart, SecEnd - LastSecEnd});
+  LastSecEnd = SecEnd;
 
   // Generate the name table for all the functions referenced in the profile.
   for (const auto &I : ProfileMap) {
     addName(I.first());
     addNames(I.second);
   }
 
   writeNameTable();
+  SecEnd = OS.tell();
+  SecHdrTable.push_back(
+      {SecNameTable, 0, LastSecEnd - FileStart, SecEnd - LastSecEnd});
   return sampleprof_error::success;
 }
 
 std::error_code SampleProfileWriterCompactBinary::writeHeader(
     const StringMap<FunctionSamples> &ProfileMap) {
   support::endian::Writer Writer(*OutputStream, support::little);
   if (auto EC = SampleProfileWriterBinary::writeHeader(ProfileMap))
     return EC;
for (const auto &J : S.getCallsiteSamples())
     }
 
   return sampleprof_error::success;
 }
 
 /// Write samples of a top-level function to a binary file.
 ///
 /// \returns true if the samples were written successfully, false otherwise.
-std::error_code SampleProfileWriterBinary::write(const FunctionSamples &S) {
+std::error_code
+SampleProfileWriterBinary::writeSample(const FunctionSamples &S) {
   encodeULEB128(S.getHeadSamples(), *OutputStream);
   return writeBody(S);
 }
 
 std::error_code
-SampleProfileWriterCompactBinary::write(const FunctionSamples &S) {
+SampleProfileWriterCompactBinary::writeSample(const FunctionSamples &S) {
   uint64_t Offset = OutputStream->tell();
   StringRef Name = S.getName();
   FuncOffsetTable[Name] = Offset;
   encodeULEB128(S.getHeadSamples(), *OutputStream);
   return writeBody(S);
 }
 
 /// Create a sample profile file writer based on the specified format.
 ///
 /// \param Filename The file to create.
 ///
 /// \param Format Encoding format for the profile file.
 ///
 /// \returns an error code indicating the status of the created writer.
 ErrorOr<std::unique_ptr<SampleProfileWriter>>
 SampleProfileWriter::create(StringRef Filename, SampleProfileFormat Format) {
   std::error_code EC;
   std::unique_ptr<raw_ostream> OS;
-  if (Format == SPF_Binary || Format == SPF_Compact_Binary)
+  if (Format == SPF_Binary || Format == SPF_Ext_Binary ||
+      Format == SPF_Compact_Binary)
     OS.reset(new raw_fd_ostream(Filename, EC, sys::fs::OF_None));
   else
     OS.reset(new raw_fd_ostream(Filename, EC, sys::fs::OF_Text));
   if (EC)
     return EC;
 
   return create(OS, Format);
 }
 
 /// Create a sample profile stream writer based on the specified format.
 ///
 /// \param OS The output stream to store the profile data to.
 ///
 /// \param Format Encoding format for the profile file.
 ///
 /// \returns an error code indicating the status of the created writer.
 ErrorOr<std::unique_ptr<SampleProfileWriter>>
 SampleProfileWriter::create(std::unique_ptr<raw_ostream> &OS,
                             SampleProfileFormat Format) {
   std::error_code EC;
   std::unique_ptr<SampleProfileWriter> Writer;
 
   if (Format == SPF_Binary)
     Writer.reset(new SampleProfileWriterRawBinary(OS));
+  else if (Format == SPF_Ext_Binary)
+    Writer.reset(new SampleProfileWriterExtBinary(OS));
   else if (Format == SPF_Compact_Binary)
     Writer.reset(new SampleProfileWriterCompactBinary(OS));
   else if (Format == SPF_Text)
     Writer.reset(new SampleProfileWriterText(OS));
   else if (Format == SPF_GCC)
     EC = sampleprof_error::unsupported_writing_format;
   else
     EC = sampleprof_error::unrecognized_format;
 
   if (EC)
     return EC;
 
+  Writer->Format = Format;
   return std::move(Writer);
 }
 
 void SampleProfileWriter::computeSummary(
     const StringMap<FunctionSamples> &ProfileMap) {
   SampleProfileSummaryBuilder Builder(ProfileSummaryBuilder::DefaultCutoffs);
   for (const auto &I : ProfileMap) {
     const FunctionSamples &Profile = I.second;
     Builder.addRecord(Profile);
   }
   Summary = Builder.getSummary();
 }

lib/Transforms/IPO/SampleProfile.cpp

protected:
   ///
   /// In this phase, in annotation, we should not promote indirect calls.
   /// Instead, we will mark GUIDs that needs to be annotated to the function.
   bool IsThinLTOPreLink;
 
   /// Profile Summary Info computed from sample profile.
   ProfileSummaryInfo *PSI = nullptr;
 
+  /// Symbol whitelist tells whether a function name appears in the binary
+  /// used to generate the current profile.
+  std::unique_ptr<SymbolWhiteList> SWL;
+
   /// Total number of samples collected in this profile.
   ///
   /// This is the sum of all the samples collected in all the functions executed
   /// at runtime.
   uint64_t TotalCollectedSamples = 0;
 
   /// Optimization Remark Emitter used to emit diagnostic remarks.
   OptimizationRemarkEmitter *ORE = nullptr;
bool SampleProfileLoader::doInitialization(Module &M) {
   if (std::error_code EC = ReaderOrErr.getError()) {
     std::string Msg = "Could not open profile: " + EC.message();
     Ctx.diagnose(DiagnosticInfoSampleProfile(Filename, Msg));
     return false;
   }
   Reader = std::move(ReaderOrErr.get());
   Reader->collectFuncsToUse(M);
   ProfileIsValid = (Reader->read() == sampleprof_error::success);
+  SWL = Reader->getSymbolWhiteList();
 
   if (!RemappingFilename.empty()) {
     // Apply profile remappings to the loaded profile data if requested.
     // For now, we only support remapping symbols encoded using the Itanium
     // C++ ABI's name mangling scheme.
     ReaderOrErr = SampleProfileReaderItaniumRemapper::create(
         RemappingFilename, Ctx, std::move(Reader));
     if (std::error_code EC = ReaderOrErr.getError()) {
bool SampleProfileLoaderLegacyPass::runOnModule(Module &M) {
   ACT = &getAnalysis<AssumptionCacheTracker>();
   TTIWP = &getAnalysis<TargetTransformInfoWrapperPass>();
   ProfileSummaryInfo *PSI =
       &getAnalysis<ProfileSummaryInfoWrapperPass>().getPSI();
   return SampleLoader.runOnModule(M, nullptr, PSI);
 }
 
 bool SampleProfileLoader::runOnFunction(Function &F, ModuleAnalysisManager *AM) {
-  
   DILocation2SampleMap.clear();
   // By default the entry count is initialized to -1, which will be treated
   // conservatively by getEntryCount as the same as unknown (None). This is
   // to avoid newly added code to be treated as cold. If we have samples
   // this will be overwritten in emitAnnotations.
   // If ProfileSampleAccurate is true or F has profile-sample-accurate
-  // attribute, initialize the entry count to 0 so callsites or functions
-  // unsampled will be treated as cold.
+  // attribute, and if there is no symbol whitelist read in, initialize all
+  // the function entry counts to 0; if there is symbol whitelist, only
+  // initialize the entry count to 0 when current function is in the
+  // whitelist.
   uint64_t initialEntryCount =
-      (ProfileSampleAccurate || F.hasFnAttribute("profile-sample-accurate"))
+      ((ProfileSampleAccurate || F.hasFnAttribute("profile-sample-accurate")) &&
+       (!SWL || SWL->contains(F.getName())))
           ? 0
           : -1;
   F.setEntryCount(ProfileCount(initialEntryCount, Function::PCT_Real));
   std::unique_ptr<OptimizationRemarkEmitter> OwnedORE;
   if (AM) {
     auto &FAM =
         AM->getResult<FunctionAnalysisManagerModuleProxy>(*F.getParent())
             .getManager();

test/Transforms/SampleProfile/compact-binary-profile.ll

-; RUN: opt < %s -sample-profile -sample-profile-file=%S/Inputs/inline.prof -S | FileCheck %s
-; RUN: opt < %s -passes=sample-profile -sample-profile-file=%S/Inputs/inline.prof -S | FileCheck %s
-; RUN: opt < %s -sample-profile -sample-profile-file=%S/Inputs/inline.compactbinary.afdo -S | FileCheck %s
-; RUN: opt < %s -passes=sample-profile -sample-profile-file=%S/Inputs/inline.compactbinary.afdo -S | FileCheck %s
-
-; Original C++ test case
-;
-; #include <stdio.h>
-;
-; int sum(int x, int y) {
-;   return x + y;
-; }
-;
-; int main() {
-;   int s, i = 0;
-;   while (i++ < 20000 * 20000)
-;     if (i != 100) s = sum(i, s); else s = 30;
-;   printf("sum is %d\n", s);
-;   return 0;
-; }
-;
-@.str = private unnamed_addr constant [11 x i8] c"sum is %d\0A\00", align 1
-
-; Check sample-profile phase using compactbinary format profile will annotate
-; the IR with exactly the same result as using text format.
-; CHECK: br i1 %cmp, label %while.body, label %while.end{{.*}} !prof ![[IDX1:[0-9]*]]
-; CHECK: br i1 %cmp1, label %if.then, label %if.else{{.*}} !prof ![[IDX2:[0-9]*]]
-; CHECK: call i32 (i8*, ...) @printf{{.*}} !prof ![[IDX3:[0-9]*]]
-; CHECK: = !{!"TotalCount", i64 26781}
-; CHECK: = !{!"MaxCount", i64 5553}
-; CHECK: ![[IDX1]] = !{!"branch_weights", i32 5392, i32 163}
-; CHECK: ![[IDX2]] = !{!"branch_weights", i32 5280, i32 113}
-; CHECK: ![[IDX3]] = !{!"branch_weights", i32 1}
-
-; Function Attrs: nounwind uwtable
-define i32 @_Z3sumii(i32 %x, i32 %y) !dbg !4 {
-entry:
-  %x.addr = alloca i32, align 4
-  %y.addr = alloca i32, align 4
-  store i32 %x, i32* %x.addr, align 4
-  store i32 %y, i32* %y.addr, align 4
-  %0 = load i32, i32* %x.addr, align 4, !dbg !11
-  %1 = load i32, i32* %y.addr, align 4, !dbg !11
-  %add = add nsw i32 %0, %1, !dbg !11
-  ret i32 %add, !dbg !11
-}
-
-; Function Attrs: uwtable
-define i32 @main() !dbg !7 {
-entry:
-  %retval = alloca i32, align 4
-  %s = alloca i32, align 4
-  %i = alloca i32, align 4
-  store i32 0, i32* %retval
-  store i32 0, i32* %i, align 4, !dbg !12
-  br label %while.cond, !dbg !13
-
-while.cond:                                       ; preds = %if.end, %entry
-  %0 = load i32, i32* %i, align 4, !dbg !14
-  %inc = add nsw i32 %0, 1, !dbg !14
-  store i32 %inc, i32* %i, align 4, !dbg !14
-  %cmp = icmp slt i32 %0, 400000000, !dbg !14
-  br i1 %cmp, label %while.body, label %while.end, !dbg !14
-
-while.body:                                       ; preds = %while.cond
-  %1 = load i32, i32* %i, align 4, !dbg !16
-  %cmp1 = icmp ne i32 %1, 100, !dbg !16
-  br i1 %cmp1, label %if.then, label %if.else, !dbg !16
-
-
-if.then:                                          ; preds = %while.body
-  %2 = load i32, i32* %i, align 4, !dbg !18
-  %3 = load i32, i32* %s, align 4, !dbg !18
-  %call = call i32 @_Z3sumii(i32 %2, i32 %3), !dbg !18
-  store i32 %call, i32* %s, align 4, !dbg !18
-  br label %if.end, !dbg !18
-
-if.else:                                          ; preds = %while.body
-  store i32 30, i32* %s, align 4, !dbg !20
-  br label %if.end
-
-if.end:                                           ; preds = %if.else, %if.then
-  br label %while.cond, !dbg !22
-
-while.end:                                        ; preds = %while.cond
-  %4 = load i32, i32* %s, align 4, !dbg !24
-  %call2 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([11 x i8], [11 x i8]* @.str, i32 0, i32 0), i32 %4), !dbg !24
-  ret i32 0, !dbg !25
-}
-
-declare i32 @printf(i8*, ...) #2
-
-!llvm.dbg.cu = !{!0}
-!llvm.module.flags = !{!8, !9}
-!llvm.ident = !{!10}
-
-!0 = distinct !DICompileUnit(language: DW_LANG_C_plus_plus, producer: "clang version 3.5 ", isOptimized: false, emissionKind: NoDebug, file: !1, enums: !2, retainedTypes: !2, globals: !2, imports: !2)
-!1 = !DIFile(filename: "calls.cc", directory: ".")
-!2 = !{}
-!4 = distinct !DISubprogram(name: "sum", line: 3, isLocal: false, isDefinition: true, virtualIndex: 6, flags: DIFlagPrototyped, isOptimized: false, unit: !0, scopeLine: 3, file: !1, scope: !5, type: !6, retainedNodes: !2)
-!5 = !DIFile(filename: "calls.cc", directory: ".")
-!6 = !DISubroutineType(types: !2)
-!7 = distinct !DISubprogram(name: "main", line: 7, isLocal: false, isDefinition: true, virtualIndex: 6, flags: DIFlagPrototyped, isOptimized: false, unit: !0, scopeLine: 7, file: !1, scope: !5, type: !6, retainedNodes: !2)
-!8 = !{i32 2, !"Dwarf Version", i32 4}
-!9 = !{i32 1, !"Debug Info Version", i32 3}
-!10 = !{!"clang version 3.5 "}
-!11 = !DILocation(line: 4, scope: !4)
-!12 = !DILocation(line: 8, scope: !7)
-!13 = !DILocation(line: 9, scope: !7)
-!14 = !DILocation(line: 9, scope: !15)
-!15 = !DILexicalBlockFile(discriminator: 2, file: !1, scope: !7)
-!16 = !DILocation(line: 10, scope: !17)
-!17 = distinct !DILexicalBlock(line: 10, column: 0, file: !1, scope: !7)
-!18 = !DILocation(line: 10, scope: !19)
-!19 = !DILexicalBlockFile(discriminator: 2, file: !1, scope: !17)
-!20 = !DILocation(line: 10, scope: !21)
-!21 = !DILexicalBlockFile(discriminator: 4, file: !1, scope: !17)
-!22 = !DILocation(line: 10, scope: !23)
-!23 = !DILexicalBlockFile(discriminator: 6, file: !1, scope: !17)
-!24 = !DILocation(line: 11, scope: !7)
-!25 = !DILocation(line: 12, scope: !7)

test/Transforms/SampleProfile/profile-format.ll

+; RUN: opt < %s -sample-profile -sample-profile-file=%S/Inputs/inline.prof -S | FileCheck %s
+; RUN: opt < %s -passes=sample-profile -sample-profile-file=%S/Inputs/inline.prof -S | FileCheck %s
+; RUN: opt < %s -sample-profile -sample-profile-file=%S/Inputs/inline.compactbinary.afdo -S | FileCheck %s
+; RUN: opt < %s -passes=sample-profile -sample-profile-file=%S/Inputs/inline.compactbinary.afdo -S | FileCheck %s
+; RUN: opt < %s -sample-profile -sample-profile-file=%S/Inputs/inline.extbinary.afdo -S | FileCheck %s
+; RUN: opt < %s -passes=sample-profile -sample-profile-file=%S/Inputs/inline.extbinary.afdo -S | FileCheck %s
+
+; Original C++ test case
+;
+; #include <stdio.h>
+;
+; int sum(int x, int y) {
+;   return x + y;
+; }
+;
+; int main() {
+;   int s, i = 0;
+;   while (i++ < 20000 * 20000)
+;     if (i != 100) s = sum(i, s); else s = 30;
+;   printf("sum is %d\n", s);
+;   return 0;
+; }
+;
+@.str = private unnamed_addr constant [11 x i8] c"sum is %d\0A\00", align 1
+
+; Check sample-profile phase using compactbinary or extbinary format profile
+; will annotate the IR with exactly the same result as using text format.
+; CHECK: br i1 %cmp, label %while.body, label %while.end{{.*}} !prof ![[IDX1:[0-9]*]]
+; CHECK: br i1 %cmp1, label %if.then, label %if.else{{.*}} !prof ![[IDX2:[0-9]*]]
+; CHECK: call i32 (i8*, ...) @printf{{.*}} !prof ![[IDX3:[0-9]*]]
+; CHECK: = !{!"TotalCount", i64 26781}
+; CHECK: = !{!"MaxCount", i64 5553}
+; CHECK: ![[IDX1]] = !{!"branch_weights", i32 5392, i32 163}
+; CHECK: ![[IDX2]] = !{!"branch_weights", i32 5280, i32 113}
+; CHECK: ![[IDX3]] = !{!"branch_weights", i32 1}
+
+; Function Attrs: nounwind uwtable
+define i32 @_Z3sumii(i32 %x, i32 %y) !dbg !4 {
+entry:
+  %x.addr = alloca i32, align 4
+  %y.addr = alloca i32, align 4
+  store i32 %x, i32* %x.addr, align 4
+  store i32 %y, i32* %y.addr, align 4
+  %0 = load i32, i32* %x.addr, align 4, !dbg !11
+  %1 = load i32, i32* %y.addr, align 4, !dbg !11
+  %add = add nsw i32 %0, %1, !dbg !11
+  ret i32 %add, !dbg !11
+}
+
+; Function Attrs: uwtable
+define i32 @main() !dbg !7 {
+entry:
+  %retval = alloca i32, align 4
+  %s = alloca i32, align 4
+  %i = alloca i32, align 4
+  store i32 0, i32* %retval
+  store i32 0, i32* %i, align 4, !dbg !12
+  br label %while.cond, !dbg !13
+
+while.cond:                                       ; preds = %if.end, %entry
+  %0 = load i32, i32* %i, align 4, !dbg !14
+  %inc = add nsw i32 %0, 1, !dbg !14
+  store i32 %inc, i32* %i, align 4, !dbg !14
+  %cmp = icmp slt i32 %0, 400000000, !dbg !14
+  br i1 %cmp, label %while.body, label %while.end, !dbg !14
+
+while.body:                                       ; preds = %while.cond
+  %1 = load i32, i32* %i, align 4, !dbg !16
+  %cmp1 = icmp ne i32 %1, 100, !dbg !16
+  br i1 %cmp1, label %if.then, label %if.else, !dbg !16
+
+
+if.then:                                          ; preds = %while.body
+  %2 = load i32, i32* %i, align 4, !dbg !18
+  %3 = load i32, i32* %s, align 4, !dbg !18
+  %call = call i32 @_Z3sumii(i32 %2, i32 %3), !dbg !18
+  store i32 %call, i32* %s, align 4, !dbg !18
+  br label %if.end, !dbg !18
+
+if.else:                                          ; preds = %while.body
+  store i32 30, i32* %s, align 4, !dbg !20
+  br label %if.end
+
+if.end:                                           ; preds = %if.else, %if.then
+  br label %while.cond, !dbg !22
+
+while.end:                                        ; preds = %while.cond
+  %4 = load i32, i32* %s, align 4, !dbg !24
+  %call2 = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([11 x i8], [11 x i8]* @.str, i32 0, i32 0), i32 %4), !dbg !24
+  ret i32 0, !dbg !25
+}
+
+declare i32 @printf(i8*, ...) #2
+
+!llvm.dbg.cu = !{!0}
+!llvm.module.flags = !{!8, !9}
+!llvm.ident = !{!10}
+
+!0 = distinct !DICompileUnit(language: DW_LANG_C_plus_plus, producer: "clang version 3.5 ", isOptimized: false, emissionKind: NoDebug, file: !1, enums: !2, retainedTypes: !2, globals: !2, imports: !2)
+!1 = !DIFile(filename: "calls.cc", directory: ".")
+!2 = !{}
+!4 = distinct !DISubprogram(name: "sum", line: 3, isLocal: false, isDefinition: true, virtualIndex: 6, flags: DIFlagPrototyped, isOptimized: false, unit: !0, scopeLine: 3, file: !1, scope: !5, type: !6, retainedNodes: !2)
+!5 = !DIFile(filename: "calls.cc", directory: ".")
+!6 = !DISubroutineType(types: !2)
+!7 = distinct !DISubprogram(name: "main", line: 7, isLocal: false, isDefinition: true, virtualIndex: 6, flags: DIFlagPrototyped, isOptimized: false, unit: !0, scopeLine: 7, file: !1, scope: !5, type: !6, retainedNodes: !2)
+!8 = !{i32 2, !"Dwarf Version", i32 4}
+!9 = !{i32 1, !"Debug Info Version", i32 3}
+!10 = !{!"clang version 3.5 "}
+!11 = !DILocation(line: 4, scope: !4)
+!12 = !DILocation(line: 8, scope: !7)
+!13 = !DILocation(line: 9, scope: !7)
+!14 = !DILocation(line: 9, scope: !15)
+!15 = !DILexicalBlockFile(discriminator: 2, file: !1, scope: !7)
+!16 = !DILocation(line: 10, scope: !17)
+!17 = distinct !DILexicalBlock(line: 10, column: 0, file: !1, scope: !7)
+!18 = !DILocation(line: 10, scope: !19)
+!19 = !DILexicalBlockFile(discriminator: 2, file: !1, scope: !17)
+!20 = !DILocation(line: 10, scope: !21)
+!21 = !DILexicalBlockFile(discriminator: 4, file: !1, scope: !17)
+!22 = !DILocation(line: 10, scope: !23)
+!23 = !DILexicalBlockFile(discriminator: 6, file: !1, scope: !17)
+!24 = !DILocation(line: 11, scope: !7)
+!25 = !DILocation(line: 12, scope: !7)

test/Transforms/SampleProfile/symbol-white-list.ll

+; RUN: opt < %s -sample-profile -profile-sample-accurate -sample-profile-file=%S/Inputs/inline.extbinary.afdo -S | FileCheck %s
+; RUN: opt < %s -passes=sample-profile -profile-sample-accurate -sample-profile-file=%S/Inputs/inline.extbinary.afdo -S | FileCheck %s
+
+; Original C++ test case
+;
+; #include <stdio.h>
+;
+; __attribute__((noinline)) int goo() { return 3 };
+; __attribute__((noinline)) int hoo() { return 4 };
+;
+; int sum(int x, int y) {
+;   return x + y;
+; }
+;
+; int main() {
+;   int s, i = 0;
+;   while (i++ < 20000 * 20000)
+;     if (i != 100) s = sum(i, s); else s = 30;
+;   printf("sum is %d\n", s);
+;   return goo() + hoo() != 7;
+; }
+;
+; Both goo and hoo don't show up in the input profile.
+; Suppose function goo shows up in the binary generating the input profile
+; and function hoo doesn't show up. Then the symbol whitelist in the input
+; profile will contain goo but not hoo. Verify the entry count of goo is
+; 0 and the entry count of hoo is -1.
+; CHECK: define {{.*}} i32 @_Z3goov() {{.*}} !prof ![[IDX1:[0-9]*]]
+; CHECK: define {{.*}} i32 @_Z3hoov() {{.*}} !prof ![[IDX2:[0-9]*]]
+; CHECK: ![[IDX1]] = !{!"function_entry_count", i64 0}
+; CHECK: ![[IDX2]] = !{!"function_entry_count", i64 -1}
+
+target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-unknown-linux-gnu"
+
+@.str = private unnamed_addr constant [11 x i8] c"sum is %d\0A\00", align 1
+
+; Function Attrs: noinline norecurse nounwind readnone uwtable
+define dso_local i32 @_Z3goov() local_unnamed_addr #0 !dbg !7 {
+entry:
+  ret i32 3, !dbg !9
+}
+
+; Function Attrs: noinline norecurse nounwind readnone uwtable
+define dso_local i32 @_Z3hoov() local_unnamed_addr #0 !dbg !10 {
+entry:
+  ret i32 4, !dbg !11
+}
+
+; Function Attrs: norecurse nounwind readnone uwtable
+define dso_local i32 @_Z3sumii(i32 %x, i32 %y) local_unnamed_addr #1 !dbg !12 {
+entry:
+  %add = add nsw i32 %y, %x, !dbg !13
+  ret i32 %add, !dbg !14
+}
+
+; Function Attrs: nofree norecurse nounwind uwtable
+define dso_local i32 @main() local_unnamed_addr #2 !dbg !15 {
+entry:
+  br label %while.body, !dbg !16
+
+while.body:                                       ; preds = %while.body, %entry
+  %inc12 = phi i32 [ 1, %entry ], [ %inc.4, %while.body ]
+  %s.011 = phi i32 [ undef, %entry ], [ %spec.select.4, %while.body ]
+  %cmp1 = icmp eq i32 %inc12, 100, !dbg !18
+  %add.i = add nsw i32 %inc12, %s.011, !dbg !20
+  %spec.select = select i1 %cmp1, i32 30, i32 %add.i, !dbg !23
+  %inc = add nuw nsw i32 %inc12, 1, !dbg !24
+  %cmp1.1 = icmp eq i32 %inc, 100, !dbg !18
+  %add.i.1 = add nsw i32 %inc, %spec.select, !dbg !20
+  %spec.select.1 = select i1 %cmp1.1, i32 30, i32 %add.i.1, !dbg !23
+  %inc.1 = add nuw nsw i32 %inc12, 2, !dbg !24
+  %cmp1.2 = icmp eq i32 %inc.1, 100, !dbg !18
+  %add.i.2 = add nsw i32 %inc.1, %spec.select.1, !dbg !20
+  %spec.select.2 = select i1 %cmp1.2, i32 30, i32 %add.i.2, !dbg !23
+  %inc.2 = add nuw nsw i32 %inc12, 3, !dbg !24
+  %cmp1.3 = icmp eq i32 %inc.2, 100, !dbg !18
+  %add.i.3 = add nsw i32 %inc.2, %spec.select.2, !dbg !20
+  %spec.select.3 = select i1 %cmp1.3, i32 30, i32 %add.i.3, !dbg !23
+  %inc.3 = add nuw nsw i32 %inc12, 4, !dbg !24
+  %cmp1.4 = icmp eq i32 %inc.3, 100, !dbg !18
+  %add.i.4 = add nsw i32 %inc.3, %spec.select.3, !dbg !20
+  %spec.select.4 = select i1 %cmp1.4, i32 30, i32 %add.i.4, !dbg !23
+  %inc.4 = add nuw nsw i32 %inc12, 5, !dbg !24
+  %exitcond.4 = icmp eq i32 %inc.4, 400000001, !dbg !26
+  br i1 %exitcond.4, label %while.end, label %while.body, !dbg !27, !llvm.loop !28
+
+while.end:                                        ; preds = %while.body
+  %call2 = tail call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([11 x i8], [11 x i8]* @.str, i64 0, i64 0), i32 %spec.select.4), !dbg !31
+  ret i32 0, !dbg !32
+}
+
+; Function Attrs: nofree nounwind
+declare dso_local i32 @printf(i8* nocapture readonly, ...) local_unnamed_addr #3
+
+attributes #0 = { noinline norecurse nounwind readnone uwtable "correctly-rounded-divide-sqrt-fp-math"="false" "disable-tail-calls"="false" "frame-pointer"="none" "less-precise-fpmad"="false" "min-legal-vector-width"="0" "no-infs-fp-math"="false" "no-jump-tables"="false" "no-nans-fp-math"="false" "no-signed-zeros-fp-math"="false" "no-trapping-math"="false" "stack-protector-buffer-size"="8" "target-cpu"="x86-64" "target-features"="+cx8,+fxsr,+mmx,+sse,+sse2,+x87" "unsafe-fp-math"="false" "use-soft-float"="false" }
+attributes #1 = { norecurse nounwind readnone uwtable "correctly-rounded-divide-sqrt-fp-math"="false" "disable-tail-calls"="false" "frame-pointer"="none" "less-precise-fpmad"="false" "min-legal-vector-width"="0" "no-infs-fp-math"="false" "no-jump-tables"="false" "no-nans-fp-math"="false" "no-signed-zeros-fp-math"="false" "no-trapping-math"="false" "stack-protector-buffer-size"="8" "target-cpu"="x86-64" "target-features"="+cx8,+fxsr,+mmx,+sse,+sse2,+x87" "unsafe-fp-math"="false" "use-soft-float"="false" }
+attributes #2 = { nofree norecurse nounwind uwtable "correctly-rounded-divide-sqrt-fp-math"="false" "disable-tail-calls"="false" "frame-pointer"="none" "less-precise-fpmad"="false" "min-legal-vector-width"="0" "no-infs-fp-math"="false" "no-jump-tables"="false" "no-nans-fp-math"="false" "no-signed-zeros-fp-math"="false" "no-trapping-math"="false" "stack-protector-buffer-size"="8" "target-cpu"="x86-64" "target-features"="+cx8,+fxsr,+mmx,+sse,+sse2,+x87" "unsafe-fp-math"="false" "use-soft-float"="false" }
+attributes #3 = { nofree nounwind "correctly-rounded-divide-sqrt-fp-math"="false" "disable-tail-calls"="false" "frame-pointer"="none" "less-precise-fpmad"="false" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "no-signed-zeros-fp-math"="false" "no-trapping-math"="false" "stack-protector-buffer-size"="8" "target-cpu"="x86-64" "target-features"="+cx8,+fxsr,+mmx,+sse,+sse2,+x87" "unsafe-fp-math"="false" "use-soft-float"="false" }
+
+!llvm.dbg.cu = !{!0}
+!llvm.module.flags = !{!3, !4, !5}
+!llvm.ident = !{!6}
+
+!0 = distinct !DICompileUnit(language: DW_LANG_C_plus_plus, file: !1, producer: "clang version 10.0.0 (trunk 369144)", isOptimized: true, runtimeVersion: 0, emissionKind: LineTablesOnly, enums: !2, debugInfoForProfiling: true, nameTableKind: None)
+!1 = !DIFile(filename: "1.cc", directory: "/usr/local/google/home/wmi/workarea/llvm-r369144/src")
+!2 = !{}
+!3 = !{i32 2, !"Dwarf Version", i32 4}
+!4 = !{i32 2, !"Debug Info Version", i32 3}
+!5 = !{i32 1, !"wchar_size", i32 4}
+!6 = !{!"clang version 10.0.0 (trunk 369144)"}
+!7 = distinct !DISubprogram(name: "goo", linkageName: "_Z3goov", scope: !1, file: !1, line: 3, type: !8, scopeLine: 3, flags: DIFlagPrototyped, spFlags: DISPFlagDefinition | DISPFlagOptimized, unit: !0, retainedNodes: !2)
+!8 = !DISubroutineType(types: !2)
+!9 = !DILocation(line: 3, column: 39, scope: !7)
+!10 = distinct !DISubprogram(name: "hoo", linkageName: "_Z3hoov", scope: !1, file: !1, line: 4, type: !8, scopeLine: 4, flags: DIFlagPrototyped, spFlags: DISPFlagDefinition | DISPFlagOptimized, unit: !0, retainedNodes: !2)
+!11 = !DILocation(line: 4, column: 39, scope: !10)
+!12 = distinct !DISubprogram(name: "sum", linkageName: "_Z3sumii", scope: !1, file: !1, line: 6, type: !8, scopeLine: 6, flags: DIFlagPrototyped, spFlags: DISPFlagDefinition | DISPFlagOptimized, unit: !0, retainedNodes: !2)
+!13 = !DILocation(line: 7, column: 12, scope: !12)
+!14 = !DILocation(line: 7, column: 3, scope: !12)
+!15 = distinct !DISubprogram(name: "main", scope: !1, file: !1, line: 10, type: !8, scopeLine: 10, flags: DIFlagPrototyped, spFlags: DISPFlagDefinition | DISPFlagOptimized, unit: !0, retainedNodes: !2)
+!16 = !DILocation(line: 12, column: 3, scope: !17)
+!17 = !DILexicalBlockFile(scope: !15, file: !1, discriminator: 2)
+!18 = !DILocation(line: 13, column: 11, scope: !19)
+!19 = !DILexicalBlockFile(scope: !15, file: !1, discriminator: 21)
+!20 = !DILocation(line: 7, column: 12, scope: !21, inlinedAt: !22)
+!21 = !DILexicalBlockFile(scope: !12, file: !1, discriminator: 21)
+!22 = distinct !DILocation(line: 13, column: 23, scope: !17)
+!23 = !DILocation(line: 13, column: 9, scope: !19)
+!24 = !DILocation(line: 12, column: 11, scope: !25)
+!25 = !DILexicalBlockFile(scope: !15, file: !1, discriminator: 1282)
+!26 = !DILocation(line: 12, column: 14, scope: !25)
+!27 = !DILocation(line: 12, column: 3, scope: !25)
+!28 = distinct !{!28, !29, !30}
+!29 = !DILocation(line: 12, column: 3, scope: !15)
+!30 = !DILocation(line: 13, column: 43, scope: !15)
+!31 = !DILocation(line: 14, column: 3, scope: !15)
+!32 = !DILocation(line: 15, column: 3, scope: !15)

tools/llvm-profdata/llvm-profdata.cpp

 #include <algorithm>
 
 using namespace llvm;
 
 enum ProfileFormat {
   PF_None = 0,
   PF_Text,
   PF_Compact_Binary,
+  PF_Ext_Binary,
   PF_GCC,
   PF_Binary
 };
 
 static void warn(Twine Message, std::string Whence = "",
                  std::string Hint = "") {
   WithColor::warning();
   if (!Whence.empty())
static void mergeInstrProfile(const WeightedFileVector &Inputs,
                               SymbolRemapper *Remapper,
                               StringRef OutputFilename,
                               ProfileFormat OutputFormat, bool OutputSparse,
                               unsigned NumThreads) {
   if (OutputFilename.compare("-") == 0)
     exitWithError("Cannot write indexed profdata format to stdout.");
 
   if (OutputFormat != PF_Binary && OutputFormat != PF_Compact_Binary &&
-      OutputFormat != PF_Text)
+      OutputFormat != PF_Ext_Binary && OutputFormat != PF_Text)
     exitWithError("Unknown format is specified.");
 
   std::mutex ErrorLock;
   SmallSet<instrprof_error, 4> WriterErrorCodes;
 
   // If NumThreads is not specified, auto-detect a good default.
   if (NumThreads == 0)
     NumThreads =
for (const auto &Callsite : CallsiteSamples.second) {
           remapSamples(Callsite.second, Remapper, Error);
       MergeResult(Error, Target[Remapped.getName()].merge(Remapped));
     }
   }
   return Result;
 }
 
 static sampleprof::SampleProfileFormat FormatMap[] = {
-    sampleprof::SPF_None, sampleprof::SPF_Text, sampleprof::SPF_Compact_Binary,
-    sampleprof::SPF_GCC, sampleprof::SPF_Binary};
+    sampleprof::SPF_None,
+    sampleprof::SPF_Text,
+    sampleprof::SPF_Compact_Binary,
+    sampleprof::SPF_Ext_Binary,
+    sampleprof::SPF_GCC,
+    sampleprof::SPF_Binary};
 
 static void mergeSampleProfile(const WeightedFileVector &Inputs,
                                SymbolRemapper *Remapper,
                                StringRef OutputFilename,
                                ProfileFormat OutputFormat) {
   using namespace sampleprof;
   StringMap<FunctionSamples> ProfileMap;
   SmallVector<std::unique_ptr<sampleprof::SampleProfileReader>, 5> Readers;
   LLVMContext Context;
+  sampleprof::SymbolWhiteList WriterList;
   for (const auto &Input : Inputs) {
     auto ReaderOrErr = SampleProfileReader::create(Input.Filename, Context);
     if (std::error_code EC = ReaderOrErr.getError())
       exitWithErrorCode(EC, Input.Filename);
 
     // We need to keep the readers around until after all the files are
     // read so that we do not lose the function names stored in each
     // reader's memory. The function names are needed to write out the
for (StringMap<FunctionSamples>::iterator I = Profiles.begin(),
       FunctionSamples &Samples = Remapper ? Remapped : I->second;
       StringRef FName = Samples.getName();
       MergeResult(Result, ProfileMap[FName].merge(Samples, Input.Weight));
       if (Result != sampleprof_error::success) {
         std::error_code EC = make_error_code(Result);
         handleMergeWriterError(errorCodeToError(EC), Input.Filename, FName);
       }
     }
+
+    std::unique_ptr<sampleprof::SymbolWhiteList> ReaderList =
+        Reader->getSymbolWhiteList();
+    if (ReaderList)
+      WriterList.merge(*ReaderList);
   }
   auto WriterOrErr =
       SampleProfileWriter::create(OutputFilename, FormatMap[OutputFormat]);
   if (std::error_code EC = WriterOrErr.getError())
     exitWithErrorCode(EC, OutputFilename);
 
   auto Writer = std::move(WriterOrErr.get());
+  Writer->setSymbolWhiteList(&WriterList);
   Writer->write(ProfileMap);
 }
 
 static WeightedFile parseWeightedFile(const StringRef &WeightedFilename) {
   StringRef WeightStr, FileName;
   std::tie(WeightStr, FileName) = WeightedFilename.split(',');
 
   uint64_t Weight;

unittests/ProfileData/SampleProfTest.cpp

FunctionType *fn_type =
         FunctionType::get(Type::getVoidTy(Context), {}, false);
     M.getOrInsertFunction(FooName, fn_type);
     M.getOrInsertFunction(BarName, fn_type);
 
     StringMap<FunctionSamples> Profiles;
     Profiles[FooName] = std::move(FooSamples);
     Profiles[BarName] = std::move(BarSamples);
 
+    SymbolWhiteList List;
+    if (Format == SampleProfileFormat::SPF_Ext_Binary) {
+      List.add("zoo", true);
+      List.add("moo", true);
+    }
+    Writer->setSymbolWhiteList(&List);
+
     std::error_code EC;
     EC = Writer->write(Profiles);
     ASSERT_TRUE(NoError(EC));
 
     Writer->getOutputStream().flush();
 
     readProfile(M, Profile);
 
     EC = Reader->read();
     ASSERT_TRUE(NoError(EC));
 
+    if (Format == SampleProfileFormat::SPF_Ext_Binary) {
+      std::unique_ptr<SymbolWhiteList> ReaderList =
+          Reader->getSymbolWhiteList();
+      ReaderList->contains("zoo");
+      ReaderList->contains("moo");
+    }
+
     if (Remap) {
       auto MemBuffer = llvm::MemoryBuffer::getMemBuffer(R"(
         # Types 'int' and 'long' are equivalent
         type i l
         # Function names 'foo' and 'faux' are equivalent
         name 3foo 4faux
       )");
       Reader.reset(new SampleProfileReaderItaniumRemapper(

 TEST_F(SampleProfTest, roundtrip_raw_binary_profile) {
   testRoundTrip(SampleProfileFormat::SPF_Binary, false);
 }
 
 TEST_F(SampleProfTest, roundtrip_compact_binary_profile) {
   testRoundTrip(SampleProfileFormat::SPF_Compact_Binary, false);
 }
 
+TEST_F(SampleProfTest, roundtrip_ext_binary_profile) {
+  testRoundTrip(SampleProfileFormat::SPF_Ext_Binary, false);
+}
+
 TEST_F(SampleProfTest, remap_text_profile) {
   testRoundTrip(SampleProfileFormat::SPF_Text, true);
 }
 
 TEST_F(SampleProfTest, remap_raw_binary_profile) {
   testRoundTrip(SampleProfileFormat::SPF_Binary, true);
 }
 
+TEST_F(SampleProfTest, remap_ext_binary_profile) {
+  testRoundTrip(SampleProfileFormat::SPF_Ext_Binary, true);
+}
+
 TEST_F(SampleProfTest, sample_overflow_saturation) {
   const uint64_t Max = std::numeric_limits<uint64_t>::max();
   sampleprof_error Result;
 
   StringRef FooName("_Z3fooi");
   FunctionSamples FooSamples;
   Result = FooSamples.addTotalSamples(1);
   ASSERT_EQ(Result, sampleprof_error::success);