[PGO] Enable profile name compression

[davidxl]

Summary

As discussed in RFC, this patch does the following

1. Changes the raw format to support compression
2. Keeps the indexed format version
3. change the coverage map format but in backward compatible way.

There is no user level tool interface/usability change.

The impact of this patch -- for clang build with -fprofile-instr-generate -fcoverage-mapping and GC ,

1. Instrumented clang size is reduced from 273M to 201M -- a 72 MB reduction (in text and data)
2. Total input object size to linker is reduced from 1187 MB to 913 MB -- a 274 MB reduction
3. Raw profile size is reduced from 147 MB to 75 MB

This patch only include the LLVM change.

1. the main compile-rt change just need to update InstrProfData.inc
2. there are about 4 test cases that need to update raw profile data (binary)
3. there are a few expected .ll file change

4. For platforms (not linux, freebsd, nor mach-o) that do not support finding section start/end with special symbols, compiler runtime change is needed to add a name section registration -- It will be in a different patch.

[vsk]

Thanks, this is awesome!

1. I've noticed that some fields (e.g NameRef) are typed as uint64_t and as IntPtrT somewhat interchangeably. Wdyt of adopting a convention where IntPtrT is reserved for values which have already been byte-swapped?

2. There are some binary tests in test/tools/llvm-cov/Inputs/ for version 1 of the coverage mapping format. Could you add similar tests for version 2?

include/llvm/ProfileData/InstrProfData.inc

164 ↗

(On Diff #45467)

I think an updated to CoverageMappingFormat.rst belongs with this commit. This should be a fairly small change (mostly lines removed).

lib/Transforms/Instrumentation/InstrProfiling.cpp

375 ↗

(On Diff #45467)

If we don't create NamesVar is there no name section, and if so, is that ever a problem?

385 ↗

(On Diff #45467)

For consistency, could you add a getter for "__llvm_prf_nms" to InstrProf.h?

[davidxl]

Update the patch with a new internal option for testing.

[davidxl]

NameRef in V1 is IntPtr type (with different size between 32 and 64bit target), but in V2 it becomes uint64_t consistently. Yes will add coverage V2 testing (better as a follow up patch I think).

include/llvm/ProfileData/InstrProfData.inc

164 ↗

(On Diff #46076)

good point. Will update.

lib/Transforms/Instrumentation/InstrProfiling.cpp

379 ↗

(On Diff #46076)

That means this module won't have data and counter (i.e, no emitted functions) sections either -- so there is no problem.

389 ↗

(On Diff #46076)

Ok to add ( I did not add it because this variable now strictly has internal linkage which runtime does not care about).

[davidxl]

Addressed vsk's review feedbacks.

[davidxl]

Update the patch to make it fully function complete (other than test case update).

• added support for name runtime registration needed on platforms other than darwin, linux, and freebsd.

The runtime part of the patch is: http://reviews.llvm.org/D16624

[davidxl]

Update the patch with test case changes.

This is now the complete LLVM patch.

[vsk]

I think it'd be fair to add the V2 binary tests after committing, once we're sure the bots are stable. A few more comments --

lib/Transforms/Instrumentation/InstrProfiling.cpp

32 ↗

(On Diff #46346)

IMO this works fine as a non-hidden option, but I'll leave it up to you.

296 ↗

(On Diff #46346)

This change seems inconsistent with the comment above it -- why start using the counter var prefix for COFF comdats?

372 ↗

(On Diff #46346)

Why is this step required, and how do we know this is the right linkage to reset to? E.g in swift, these name pointers start life with LinkOnceAny linkage.

374 ↗

(On Diff #46346)

Where would these names be stripped out? IIUC this vector is in place so we can call collectPGOFuncNameStrings in emitNameData.

[davidxl]

lib/Transforms/Instrumentation/InstrProfiling.cpp

32 ↗

(On Diff #46346)

ok will make it non hidden.

296 ↗

(On Diff #46346)

Will fix the comment. The name var comdat no long exists.

372 ↗

(On Diff #46346)

It is set to privateLinkage so that the name vars can be garbage collected by the compiler. Note that the individual name variable (per function) is created by instrumenter to pass the linkage that can be passed on to the counter and data variable. Once that is done, they are no longer needed. Will update the comment.

374 ↗

(On Diff #46346)

This is certainly an out of place comment. Will remove.

[davidxl]

Addressed vsk's review feedbacks.

[vsk]

Lgtm.

lib/Transforms/Instrumentation/InstrProfiling.cpp

296 ↗

(On Diff #46346)

D'oh. Right.

llvm/trunk/docs/CoverageMappingFormat.rst

.. code-block:: llvm
   { 
     { i32, i32, i32, i32 } ; Coverage map header
     {
       i32 2,  ; The number of function records
       i32 20, ; The length of the string that contains the encoded translation unit filenames
       i32 20, ; The length of the string that contains the encoded coverage mapping data
       i32 0,  ; Coverage mapping format version
     },
-    [2 x { i8*, i32, i32, i64 }] [ ; Function records
-     { i8*, i32, i32, i64 } { i8* getelementptr inbounds ([3 x i8]* @__profn_foo, i32 0, i32 0), ; Function's name
-       i32 3, ; Function's name length
+    [2 x { i64, i32, i64 }] [ ; Function records
+     { i64, i32, i64 } {
+       i64 0x5cf8c24cdb18bdac, ; Function's name MD5
        i32 9, ; Function's encoded coverage mapping data string length
        i64 0  ; Function's structural hash
      },
-     { i8*, i32, i32, i64 } { i8* getelementptr inbounds ([3 x i8]* @__profn_bar, i32 0, i32 0), ; Function's name
-       i32 3, ; Function's name length
+     { i64, i32, i64 } { 
+       i64 0xe413754a191db537, ; Function's name MD5
        i32 9, ; Function's encoded coverage mapping data string length
        i64 0  ; Function's structural hash
      }],
    [40 x i8] c"..." ; Encoded data (dissected later)
   }, section "__llvm_covmap", align 8
 
+The function record layout has evolved since version 1. In version 1, the function record for *foo* is defined as follows:
+
+.. code-block:: llvm
+
+     { i8*, i32, i32, i64 } { i8* getelementptr inbounds ([3 x i8]* @__profn_foo, i32 0, i32 0), ; Function's name
+       i32 3, ; Function's name length
+       i32 9, ; Function's encoded coverage mapping data string length
+       i64 0  ; Function's structural hash
+     }
+
+
 Coverage Mapping Header:
 ------------------------
 
 The coverage mapping header has the following fields:
 
 * The number of function records.
 
 * The length of the string in the third field of *__llvm_coverage_mapping* that contains the encoded translation unit filenames.
 
 * The length of the string in the third field of *__llvm_coverage_mapping* that contains the encoded coverage mapping data.
 
 * The format version. 0 is the first (current) version of the coverage mapping format.
 
 .. _function records:
 
 Function record:
 ----------------
 
 A function record is a structure of the following type:
 
 .. code-block:: llvm
 
-  { i8*, i32, i32, i64 }
+  { i64, i32, i64 }
 
-It contains the pointer to the function's name, function's name length,
-the length of the encoded mapping data for that function, and function's 
-hash value.
+It contains function name's MD5, the length of the encoded mapping data for that function, and function's 
+structural hash value.
 
 Encoded data:
 -------------
 
 The encoded data is stored in a single string that contains
 the encoded filenames used by this translation unit and the encoded coverage
 mapping data for each function in this translation unit.
 

llvm/trunk/include/llvm/ProfileData/CoverageMapping.h

 // [CoverageMapFileHeader]
 // [ArrayStart]
 //  [CovMapFunctionRecord]
 //  [CovMapFunctionRecord]
 //  ...
 // [ArrayEnd]
 // [Encoded Region Mapping Data]
 LLVM_PACKED_START
-template <class IntPtrT> struct CovMapFunctionRecord {
+template <class IntPtrT> struct CovMapFunctionRecordV1 {
+#define COVMAP_V1
 #define COVMAP_FUNC_RECORD(Type, LLVMType, Name, Init) Type Name;
 #include "llvm/ProfileData/InstrProfData.inc"
+#undef COVMAP_V1
 
   // Return the structural hash associated with the function.
   template <support::endianness Endian> uint64_t getFuncHash() const {
     return support::endian::byte_swap<uint64_t, Endian>(FuncHash);
   }
   // Return the coverage map data size for the funciton.
   template <support::endianness Endian> uint32_t getDataSize() const {
     return support::endian::byte_swap<uint32_t, Endian>(DataSize);
std::error_code getFuncName(InstrProfSymtab &ProfileNames,
     IntPtrT NameRef = getFuncNameRef<Endian>();
     uint32_t NameS = support::endian::byte_swap<uint32_t, Endian>(NameSize);
     FuncName = ProfileNames.getFuncName(NameRef, NameS);
     if (NameS && FuncName.empty())
       return coveragemap_error::malformed;
     return std::error_code();
   }
 };
+
+template <class IntPtrT> struct CovMapFunctionRecord {
+#define COVMAP_FUNC_RECORD(Type, LLVMType, Name, Init) Type Name;
+#include "llvm/ProfileData/InstrProfData.inc"
+
+  // Return the structural hash associated with the function.
+  template <support::endianness Endian> uint64_t getFuncHash() const {
+    return support::endian::byte_swap<uint64_t, Endian>(FuncHash);
+  }
+  // Return the coverage map data size for the funciton.
+  template <support::endianness Endian> uint32_t getDataSize() const {
+    return support::endian::byte_swap<uint32_t, Endian>(DataSize);
+  }
+  // Return function lookup key. The value is consider opaque.
+  template <support::endianness Endian> uint64_t getFuncNameRef() const {
+    return support::endian::byte_swap<uint64_t, Endian>(NameRef);
+  }
+  // Return the PGO name of the function */
+  template <support::endianness Endian>
+  std::error_code getFuncName(InstrProfSymtab &ProfileNames,
+                              StringRef &FuncName) const {
+    IntPtrT NameRef = getFuncNameRef<Endian>();
+    FuncName = ProfileNames.getFuncName(NameRef);
+    return std::error_code();
+  }
+};
+
 // Per module coverage mapping data header, i.e. CoverageMapFileHeader
 // documented above.
 struct CovMapHeader {
 #define COVMAP_HEADER(Type, LLVMType, Name, Init) Type Name;
 #include "llvm/ProfileData/InstrProfData.inc"
   template <support::endianness Endian> uint32_t getNRecords() const {
     return support::endian::byte_swap<uint32_t, Endian>(NRecords);
   }
   template <support::endianness Endian> uint32_t getFilenamesSize() const {
     return support::endian::byte_swap<uint32_t, Endian>(FilenamesSize);
   }
   template <support::endianness Endian> uint32_t getCoverageSize() const {
     return support::endian::byte_swap<uint32_t, Endian>(CoverageSize);
   }
   template <support::endianness Endian> uint32_t getVersion() const {
     return support::endian::byte_swap<uint32_t, Endian>(Version);
   }
 };
 
 LLVM_PACKED_END
 
 enum CovMapVersion {
   Version1 = 0,
-  // The current version is Version1
+  // Function's name reference from CovMapFuncRecord is changed from raw
+  // name string pointer to MD5 to support name section compression. Name
+  // section is also compressed.
+  Version2 = 1,
+  // The current version is Version2
   CurrentVersion = INSTR_PROF_COVMAP_VERSION
 };
 
 template <int CovMapVersion, class IntPtrT> struct CovMapTraits {
   typedef CovMapFunctionRecord<IntPtrT> CovMapFuncRecordType;
+  typedef uint64_t NameRefType;
+};
+
+template <class IntPtrT> struct CovMapTraits<CovMapVersion::Version1, IntPtrT> {
+  typedef CovMapFunctionRecordV1<IntPtrT> CovMapFuncRecordType;
   typedef IntPtrT NameRefType;
 };
 
 } // end namespace coverage
 
 /// \brief Provide DenseMapInfo for CounterExpression
 template<> struct DenseMapInfo<coverage::CounterExpression> {
   static inline coverage::CounterExpression getEmptyKey() {

llvm/trunk/include/llvm/ProfileData/InstrProf.h

 
 /// Return the name prefix of profile counter variables.
 inline StringRef getInstrProfCountersVarPrefix() { return "__profc_"; }
 
 /// Return the name prefix of the COMDAT group for instrumentation variables
 /// associated with a COMDAT function.
 inline StringRef getInstrProfComdatPrefix() { return "__profv_"; }
 
+/// Return the name of the variable holding the strings (possibly compressed)
+/// of all function's PGO names.
+inline StringRef getInstrProfNamesVarName() {
+  return "__llvm_prf_nm";
+}
+
 /// Return the name of a covarage mapping variable (internal linkage)
 /// for each instrumented source module. Such variables are allocated
 /// in the __llvm_covmap section.
 inline StringRef getCoverageMappingVarName() {
   return "__llvm_coverage_mapping";
 }
 
 /// Return the name of the internal variable recording the array

 }
 
 /// Return the name of the runtime interface that registers per-function control
 /// data for one instrumented function.
 inline StringRef getInstrProfRegFuncName() {
   return "__llvm_profile_register_function";
 }
 
+/// Return the name of the runtime interface that registers the PGO name strings.
+inline StringRef getInstrProfNamesRegFuncName() {
+  return "__llvm_profile_register_names_function";
+}
+
 /// Return the name of the runtime initialization method that is generated by
 /// the compiler. The function calls __llvm_profile_register_functions and
 /// __llvm_profile_override_default_filename functions if needed. This function
 /// has internal linkage and invoked at startup time via init_array.
 inline StringRef getInstrProfInitFuncName() { return "__llvm_profile_init"; }
 
 /// Return the name of the hook variable defined in profile runtime library.
 /// A reference to the variable causes the linker to link in the runtime

 inline std::unique_ptr<Summary> allocSummary(uint32_t TotalSize) {
   return std::unique_ptr<Summary>(new (::operator new(TotalSize))
                                       Summary(TotalSize));
 }
 } // end namespace IndexedInstrProf
 
 namespace RawInstrProf {
 
+// Version 1: First version
+// Version 2: Added value profile data section. Per-function control data
+// struct has more fields to describe value profile information.
+// Version 3: Compressed name section support. Function PGO name reference
+// from control data struct is changed from raw pointer to Name's MD5 value.
 const uint64_t Version = INSTR_PROF_RAW_VERSION;
 
 template <class IntPtrT> inline uint64_t getMagic();
 template <> inline uint64_t getMagic<uint64_t>() {
   return INSTR_PROF_RAW_MAGIC_64;
 }
 
 template <> inline uint64_t getMagic<uint32_t>() {

llvm/trunk/include/llvm/ProfileData/InstrProfData.inc

 
 /* INSTR_PROF_DATA start. */
 /* Definition of member fields of the per-function control structure. */
 #ifndef INSTR_PROF_DATA
 #define INSTR_PROF_DATA(Type, LLVMType, Name, Initializer)
 #else
 #define INSTR_PROF_DATA_DEFINED
 #endif
-
-INSTR_PROF_DATA(const uint32_t, llvm::Type::getInt32Ty(Ctx), NameSize, \
-                ConstantInt::get(llvm::Type::getInt32Ty(Ctx), \
-                NamePtr->getType()->getPointerElementType()->getArrayNumElements()))
-INSTR_PROF_DATA(const uint32_t, llvm::Type::getInt32Ty(Ctx), NumCounters, \
-                ConstantInt::get(llvm::Type::getInt32Ty(Ctx), NumCounters))
+INSTR_PROF_DATA(const uint64_t, llvm::Type::getInt64Ty(Ctx), NameRef, \
+                ConstantInt::get(llvm::Type::getInt64Ty(Ctx), \
+    IndexedInstrProf::ComputeHash(getPGOFuncNameVarInitializer(Inc->getName()))))
 INSTR_PROF_DATA(const uint64_t, llvm::Type::getInt64Ty(Ctx), FuncHash, \
                 ConstantInt::get(llvm::Type::getInt64Ty(Ctx), \
                 Inc->getHash()->getZExtValue()))
-INSTR_PROF_DATA(const IntPtrT, llvm::Type::getInt8PtrTy(Ctx), NamePtr, \
-                ConstantExpr::getBitCast(NamePtr, llvm::Type::getInt8PtrTy(Ctx)))
 INSTR_PROF_DATA(const IntPtrT, llvm::Type::getInt64PtrTy(Ctx), CounterPtr, \
                 ConstantExpr::getBitCast(CounterPtr, \
                 llvm::Type::getInt64PtrTy(Ctx)))
 INSTR_PROF_DATA(const IntPtrT, llvm::Type::getInt8PtrTy(Ctx), FunctionPointer, \
                 FunctionAddr)
 INSTR_PROF_DATA(IntPtrT, llvm::Type::getInt8PtrTy(Ctx), Values, \
                 ConstantPointerNull::get(Int8PtrTy))
+INSTR_PROF_DATA(const uint32_t, llvm::Type::getInt32Ty(Ctx), NumCounters, \
+                ConstantInt::get(llvm::Type::getInt32Ty(Ctx), NumCounters))
 INSTR_PROF_DATA(const uint16_t, Int16ArrayTy, NumValueSites[IPVK_Last+1], \
                 ConstantArray::get(Int16ArrayTy, Int16ArrayVals))
 #undef INSTR_PROF_DATA
 /* INSTR_PROF_DATA end. */
 
 /* INSTR_PROF_RAW_HEADER  start */
 /* Definition of member fields of the raw profile header data structure. */
 #ifndef INSTR_PROF_RAW_HEADER

 /* Definition of member fields of the function record structure in coverage
  * map.
  */
 #ifndef COVMAP_FUNC_RECORD
 #define COVMAP_FUNC_RECORD(Type, LLVMType, Name, Initializer)
 #else
 #define INSTR_PROF_DATA_DEFINED
 #endif
+#ifdef COVMAP_V1
 COVMAP_FUNC_RECORD(const IntPtrT, llvm::Type::getInt8PtrTy(Ctx), \
                    NamePtr, llvm::ConstantExpr::getBitCast(NamePtr, \
                    llvm::Type::getInt8PtrTy(Ctx)))
 COVMAP_FUNC_RECORD(const uint32_t, llvm::Type::getInt32Ty(Ctx), NameSize, \
-                   llvm::ConstantInt::get(llvm::Type::getInt32Ty(Ctx),\
+                   llvm::ConstantInt::get(llvm::Type::getInt32Ty(Ctx), \
                    NameValue.size()))
+#else
+COVMAP_FUNC_RECORD(const int64_t, llvm::Type::getInt64Ty(Ctx), NameRef, \
+                   llvm::ConstantInt::get(llvm::Type::getInt64Ty(Ctx), \
+             llvm::IndexedInstrProf::ComputeHash(NameValue)))
+#endif
 COVMAP_FUNC_RECORD(const uint32_t, llvm::Type::getInt32Ty(Ctx), DataSize, \
                    llvm::ConstantInt::get(llvm::Type::getInt32Ty(Ctx),\
                    CoverageMapping.size()))
 COVMAP_FUNC_RECORD(const uint64_t, llvm::Type::getInt64Ty(Ctx), FuncHash, \
                    llvm::ConstantInt::get(llvm::Type::getInt64Ty(Ctx), FuncHash))
 #undef COVMAP_FUNC_RECORD
 /* COVMAP_FUNC_RECORD end.  */
 

  */
 #define INSTR_PROF_RAW_MAGIC_64 (uint64_t)255 << 56 | (uint64_t)'l' << 48 | \
        (uint64_t)'p' << 40 | (uint64_t)'r' << 32 | (uint64_t)'o' << 24 |  \
         (uint64_t)'f' << 16 | (uint64_t)'r' << 8 | (uint64_t)129
 #define INSTR_PROF_RAW_MAGIC_32 (uint64_t)255 << 56 | (uint64_t)'l' << 48 | \
        (uint64_t)'p' << 40 | (uint64_t)'r' << 32 | (uint64_t)'o' << 24 |  \
         (uint64_t)'f' << 16 | (uint64_t)'R' << 8 | (uint64_t)129
 
-/* Raw profile format version. */
-#define INSTR_PROF_RAW_VERSION 2
+/* Raw profile format version (start from 1). */
+#define INSTR_PROF_RAW_VERSION 3
+/* Indexed profile format version (start from 1). */
 #define INSTR_PROF_INDEX_VERSION 4
-#define INSTR_PROF_COVMAP_VERSION 0
+/* Coverage mapping format vresion (start from 0). */
+#define INSTR_PROF_COVMAP_VERSION 1
 
 /* Profile version is always of type uint64_t. Reserve the upper 8 bits in the
  * version for other variants of profile. We set the lowest bit of the upper 8
  * bits (i.e. bit 56) to 1 to indicate if this is an IR-level instrumentaiton
  * generated profile, and 0 if this is a Clang FE generated profile.
 */
 #define VARIANT_MASKS_ALL 0xff00000000000000ULL
 #define GET_VERSION(V) ((V) & ~VARIANT_MASKS_ALL)

llvm/trunk/include/llvm/ProfileData/InstrProfReader.h

private:
   std::unique_ptr<MemoryBuffer> DataBuffer;
   bool ShouldSwapBytes;
   uint64_t CountersDelta;
   uint64_t NamesDelta;
   const RawInstrProf::ProfileData<IntPtrT> *Data;
   const RawInstrProf::ProfileData<IntPtrT> *DataEnd;
   const uint64_t *CountersStart;
   const char *NamesStart;
+  uint64_t NamesSize;
   const uint8_t *ValueDataStart;
   const char *ProfileEnd;
   uint32_t ValueKindLast;
   uint32_t CurValueDataSize;
 
   InstrProfRecord::ValueMapType FunctionPtrToNameMap;
 
   RawInstrProfReader(const RawInstrProfReader &) = delete;
void advanceData() {
     Data++;
     ValueDataStart += CurValueDataSize;
   }
 
   const uint64_t *getCounter(IntPtrT CounterPtr) const {
     ptrdiff_t Offset = (swap(CounterPtr) - CountersDelta) / sizeof(uint64_t);
     return CountersStart + Offset;
   }
-  const char *getName(IntPtrT NamePtr) const {
-    ptrdiff_t Offset = (swap(NamePtr) - NamesDelta) / sizeof(char);
-    return NamesStart + Offset;
+  StringRef getName(uint64_t NameRef) const {
+    return Symtab->getFuncName(swap(NameRef));
   }
 };
 
 typedef RawInstrProfReader<uint32_t> RawInstrProfReader32;
 typedef RawInstrProfReader<uint64_t> RawInstrProfReader64;
 
 namespace IndexedInstrProf {
 enum class HashT : uint32_t;

llvm/trunk/lib/ProfileData/CoverageMappingReader.cpp

std::unique_ptr<CovMapFuncRecordReader> CovMapFuncRecordReader::get(
     coverage::CovMapVersion Version, InstrProfSymtab &P,
     std::vector<BinaryCoverageReader::ProfileMappingRecord> &R,
     std::vector<StringRef> &F) {
   using namespace coverage;
   switch (Version) {
   case CovMapVersion::Version1:
     return llvm::make_unique<VersionedCovMapFuncRecordReader<
         CovMapVersion::Version1, IntPtrT, Endian>>(P, R, F);
+  case CovMapVersion::Version2:
+    // Decompress the name data.
+    P.create(P.getNameData());
+    return llvm::make_unique<VersionedCovMapFuncRecordReader<
+        CovMapVersion::Version2, IntPtrT, Endian>>(P, R, F);
   }
   llvm_unreachable("Unsupported version");
 }
 
 template <typename T, support::endianness Endian>
 static std::error_code readCoverageMappingData(
     InstrProfSymtab &ProfileNames, StringRef Data,
     std::vector<BinaryCoverageReader::ProfileMappingRecord> &Records,

llvm/trunk/lib/ProfileData/InstrProfReader.cpp

RawInstrProfReader<IntPtrT>::readNextHeader(const char *CurrentPos) {
 
   // There's another profile to read, so we need to process the header.
   auto *Header = reinterpret_cast<const RawInstrProf::Header *>(CurrentPos);
   return readHeader(*Header);
 }
 
 template <class IntPtrT>
 void RawInstrProfReader<IntPtrT>::createSymtab(InstrProfSymtab &Symtab) {
+  Symtab.create(StringRef(NamesStart, NamesSize));
   for (const RawInstrProf::ProfileData<IntPtrT> *I = Data; I != DataEnd; ++I) {
-    StringRef FunctionName(getName(I->NamePtr), swap(I->NameSize));
-    Symtab.addFuncName(FunctionName);
     const IntPtrT FPtr = swap(I->FunctionPointer);
     if (!FPtr)
       continue;
-    Symtab.mapAddress(FPtr, IndexedInstrProf::ComputeHash(FunctionName));
+    Symtab.mapAddress(FPtr, I->NameRef);
   }
   Symtab.finalizeSymtab();
 }
 
 template <class IntPtrT>
 std::error_code
 RawInstrProfReader<IntPtrT>::readHeader(const RawInstrProf::Header &Header) {
   if (swap(Header.Version) != RawInstrProf::Version)
     return error(instrprof_error::unsupported_version);
 
   CountersDelta = swap(Header.CountersDelta);
   NamesDelta = swap(Header.NamesDelta);
   auto DataSize = swap(Header.DataSize);
   auto CountersSize = swap(Header.CountersSize);
-  auto NamesSize = swap(Header.NamesSize);
+  NamesSize = swap(Header.NamesSize);
   auto ValueDataSize = swap(Header.ValueDataSize);
   ValueKindLast = swap(Header.ValueKindLast);
 
   auto DataSizeInBytes = DataSize * sizeof(RawInstrProf::ProfileData<IntPtrT>);
   auto PaddingSize = getNumPaddingBytes(NamesSize);
 
   ptrdiff_t DataOffset = sizeof(RawInstrProf::Header);
   ptrdiff_t CountersOffset = DataOffset + DataSizeInBytes;
RawInstrProfReader<IntPtrT>::readHeader(const RawInstrProf::Header &Header) {
   std::unique_ptr<InstrProfSymtab> NewSymtab = make_unique<InstrProfSymtab>();
   createSymtab(*NewSymtab.get());
   Symtab = std::move(NewSymtab);
   return success();
 }
 
 template <class IntPtrT>
 std::error_code RawInstrProfReader<IntPtrT>::readName(InstrProfRecord &Record) {
-  Record.Name = StringRef(getName(Data->NamePtr), swap(Data->NameSize));
-  if (Record.Name.data() < NamesStart ||
-      Record.Name.data() + Record.Name.size() >
-          reinterpret_cast<const char *>(ValueDataStart))
-    return error(instrprof_error::malformed);
+  Record.Name = getName(Data->NameRef);
   return success();
 }
 
 template <class IntPtrT>
 std::error_code RawInstrProfReader<IntPtrT>::readFuncHash(
     InstrProfRecord &Record) {
   Record.Hash = swap(Data->FuncHash);
   return success();

llvm/trunk/lib/Transforms/Instrumentation/InstrProfiling.cpp

 #include "llvm/Transforms/Utils/ModuleUtils.h"
 
 using namespace llvm;
 
 #define DEBUG_TYPE "instrprof"
 
 namespace {
 
+cl::opt<bool> DoNameCompression("enable-name-compression",
+                                cl::desc("Enable name string compression"),
+                                cl::init(true));
+
 class InstrProfiling : public ModulePass {
 public:
   static char ID;
 
   InstrProfiling() : ModulePass(ID) {}
 
   InstrProfiling(const InstrProfOptions &Options)
       : ModulePass(ID), Options(Options) {}
typedef struct PerFunctionProfileData {
     GlobalVariable* RegionCounters;
     GlobalVariable* DataVar;
     PerFunctionProfileData() : RegionCounters(nullptr), DataVar(nullptr) {
       memset(NumValueSites, 0, sizeof(uint32_t) * (IPVK_Last+1));
     }
   } PerFunctionProfileData;
   DenseMap<GlobalVariable *, PerFunctionProfileData> ProfileDataMap;
   std::vector<Value *> UsedVars;
+  std::vector<GlobalVariable *> ReferencedNames;
+  GlobalVariable *NamesVar;
+  size_t NamesSize;
 
   bool isMachO() const {
     return Triple(M->getTargetTriple()).isOSBinFormatMachO();
   }
 
   /// Get the section name for the counter variables.
   StringRef getCountersSection() const {
     return getInstrProfCountersSectionName(isMachO());
private:
   void lowerCoverageData(GlobalVariable *CoverageNamesVar);
 
   /// Get the region counters for an increment, creating them if necessary.
   ///
   /// If the counter array doesn't yet exist, the profile data variables
   /// referring to them will also be created.
   GlobalVariable *getOrCreateRegionCounters(InstrProfIncrementInst *Inc);
 
+  /// Emit the section with compressed function names.
+  void emitNameData();
+
   /// Emit runtime registration functions for each profile data variable.
   void emitRegistration();
 
   /// Emit the necessary plumbing to pull in the runtime initialization.
   void emitRuntimeHook();
 
   /// Add uses of our data variables and runtime hook.
   void emitUses();

 ModulePass *llvm::createInstrProfilingPass(const InstrProfOptions &Options) {
   return new InstrProfiling(Options);
 }
 
 bool InstrProfiling::runOnModule(Module &M) {
   bool MadeChange = false;
 
   this->M = &M;
+  NamesVar = nullptr;
+  NamesSize = 0;
   ProfileDataMap.clear();
   UsedVars.clear();
 
   // We did not know how many value sites there would be inside
   // the instrumented function. This is counting the number of instrumented
   // target value sites to enter it as field in the profile data variable.
   for (Function &F : M) {
     InstrProfIncrementInst *FirstProfIncInst = nullptr;
if (GlobalVariable *CoverageNamesVar =
           M.getNamedGlobal(getCoverageUnusedNamesVarName())) {
     lowerCoverageData(CoverageNamesVar);
     MadeChange = true;
   }
 
   if (!MadeChange)
     return false;
 
+  emitNameData();
   emitRegistration();
   emitRuntimeHook();
   emitUses();
   emitInitialization();
   return true;
 }
 
 static Constant *getOrInsertValueProfilingCall(Module &M) {
void InstrProfiling::lowerCoverageData(GlobalVariable *CoverageNamesVar) {
   ConstantArray *Names =
       cast<ConstantArray>(CoverageNamesVar->getInitializer());
   for (unsigned I = 0, E = Names->getNumOperands(); I < E; ++I) {
     Constant *NC = Names->getOperand(I);
     Value *V = NC->stripPointerCasts();
     assert(isa<GlobalVariable>(V) && "Missing reference to function name");
     GlobalVariable *Name = cast<GlobalVariable>(V);
 
-    // Move the name variable to the right section.
-    Name->setSection(getNameSection());
-    Name->setAlignment(1);
+    Name->setLinkage(GlobalValue::PrivateLinkage);
+    ReferencedNames.push_back(Name);
   }
 }
 
 /// Get the name of a profiling variable for a particular function.
 static std::string getVarName(InstrProfIncrementInst *Inc, StringRef Prefix) {
   StringRef NamePrefix = getInstrProfNameVarPrefix();
   StringRef Name = Inc->getName()->getName().substr(NamePrefix.size());
   return (Prefix + Name).str();
 }
 
 static inline bool shouldRecordFunctionAddr(Function *F) {
   // Check the linkage
   if (!F->hasLinkOnceLinkage() && !F->hasLocalLinkage() &&
       !F->hasAvailableExternallyLinkage())
     return true;
   // Check uses of this function for other than direct calls or invokes to it.
   return F->hasAddressTaken();
 }
 
 static inline Comdat *getOrCreateProfileComdat(Module &M,
                                                InstrProfIncrementInst *Inc) {
   // COFF format requires a COMDAT section to have a key symbol with the same
   // name. The linker targeting COFF also requires that the COMDAT
   // a section is associated to must precede the associating section. For this
-  // reason, we must choose the name var's name as the name of the comdat.
+  // reason, we must choose the counter var's name as the name of the comdat.
   StringRef ComdatPrefix = (Triple(M.getTargetTriple()).isOSBinFormatCOFF()
-                                ? getInstrProfNameVarPrefix()
+                                ? getInstrProfCountersVarPrefix()
                                 : getInstrProfComdatPrefix());
   return M.getOrInsertComdat(StringRef(getVarName(Inc, ComdatPrefix)));
 }
 
 GlobalVariable *
 InstrProfiling::getOrCreateRegionCounters(InstrProfIncrementInst *Inc) {
   GlobalVariable *NamePtr = Inc->getName();
   auto It = ProfileDataMap.find(NamePtr);
   PerFunctionProfileData PD;
   if (It != ProfileDataMap.end()) {
     if (It->second.RegionCounters)
       return It->second.RegionCounters;
     PD = It->second;
   }
 
   // Move the name variable to the right section. Place them in a COMDAT group
   // if the associated function is a COMDAT. This will make sure that
   // only one copy of counters of the COMDAT function will be emitted after
   // linking.
   Function *Fn = Inc->getParent()->getParent();
   Comdat *ProfileVarsComdat = nullptr;
   if (Fn->hasComdat())
     ProfileVarsComdat = getOrCreateProfileComdat(*M, Inc);
-  NamePtr->setSection(getNameSection());
-  NamePtr->setAlignment(1);
-  NamePtr->setComdat(ProfileVarsComdat);
 
   uint64_t NumCounters = Inc->getNumCounters()->getZExtValue();
   LLVMContext &Ctx = M->getContext();
   ArrayType *CounterTy = ArrayType::get(Type::getInt64Ty(Ctx), NumCounters);
 
   // Create the counters variable.
   auto *CounterPtr =
       new GlobalVariable(*M, CounterTy, false, NamePtr->getLinkage(),
InstrProfiling::getOrCreateRegionCounters(InstrProfIncrementInst *Inc) {
   Data->setComdat(ProfileVarsComdat);
 
   PD.RegionCounters = CounterPtr;
   PD.DataVar = Data;
   ProfileDataMap[NamePtr] = PD;
 
   // Mark the data variable as used so that it isn't stripped out.
   UsedVars.push_back(Data);
+  // Now that the linkage set by the FE has been passed to the data and counter
+  // variables, reset Name variable's linkage and visibility to private so that
+  // it can be removed later by the compiler.
+  NamePtr->setLinkage(GlobalValue::PrivateLinkage);
+  // Collect the referenced names to be used by emitNameData.
+  ReferencedNames.push_back(NamePtr);
 
   return CounterPtr;
 }
 
+void InstrProfiling::emitNameData() {
+  std::string UncompressedData;
+
+  if (ReferencedNames.empty())
+    return;
+
+  std::string CompressedNameStr;
+  collectPGOFuncNameStrings(ReferencedNames, CompressedNameStr,
+                            DoNameCompression);
+
+  auto &Ctx = M->getContext();
+  auto *NamesVal = llvm::ConstantDataArray::getString(
+      Ctx, StringRef(CompressedNameStr), false);
+  NamesVar = new llvm::GlobalVariable(*M, NamesVal->getType(), true,
+                                      llvm::GlobalValue::PrivateLinkage,
+                                      NamesVal, getInstrProfNamesVarName());
+  NamesSize = CompressedNameStr.size();
+  NamesVar->setSection(getNameSection());
+  UsedVars.push_back(NamesVar);
+}
+
 void InstrProfiling::emitRegistration() {
   // Don't do this for Darwin.  compiler-rt uses linker magic.
   if (Triple(M->getTargetTriple()).isOSDarwin())
     return;
 
   // Use linker script magic to get data/cnts/name start/end.
   if (Triple(M->getTargetTriple()).isOSLinux() ||
       Triple(M->getTargetTriple()).isOSFreeBSD())
     return;
 
   // Construct the function.
   auto *VoidTy = Type::getVoidTy(M->getContext());
   auto *VoidPtrTy = Type::getInt8PtrTy(M->getContext());
+  auto *Int64Ty = Type::getInt64Ty(M->getContext());
   auto *RegisterFTy = FunctionType::get(VoidTy, false);
   auto *RegisterF = Function::Create(RegisterFTy, GlobalValue::InternalLinkage,
                                      getInstrProfRegFuncsName(), M);
   RegisterF->setUnnamedAddr(true);
   if (Options.NoRedZone) RegisterF->addFnAttr(Attribute::NoRedZone);
 
   auto *RuntimeRegisterTy = FunctionType::get(VoidTy, VoidPtrTy, false);
   auto *RuntimeRegisterF =
       Function::Create(RuntimeRegisterTy, GlobalVariable::ExternalLinkage,
                        getInstrProfRegFuncName(), M);
 
   IRBuilder<> IRB(BasicBlock::Create(M->getContext(), "", RegisterF));
   for (Value *Data : UsedVars)
+    if (Data != NamesVar)
       IRB.CreateCall(RuntimeRegisterF, IRB.CreateBitCast(Data, VoidPtrTy));
+
+  if (NamesVar) {
+    Type *ParamTypes[] = {VoidPtrTy, Int64Ty};
+    auto *NamesRegisterTy =
+        FunctionType::get(VoidTy, makeArrayRef(ParamTypes), false);
+    auto *NamesRegisterF =
+        Function::Create(NamesRegisterTy, GlobalVariable::ExternalLinkage,
+                         getInstrProfNamesRegFuncName(), M);
+    IRB.CreateCall(NamesRegisterF, {IRB.CreateBitCast(NamesVar, VoidPtrTy),
+                                    IRB.getInt64(NamesSize)});
+  }
+
   IRB.CreateRetVoid();
 }
 
 void InstrProfiling::emitRuntimeHook() {
 
   // We expect the linker to be invoked with -u<hook_var> flag for linux,
   // for which case there is no need to emit the user function.
   if (Triple(M->getTargetTriple()).isOSLinux())

llvm/trunk/test/Instrumentation/InstrProfiling/PR23499.ll

 ;; Check that PGO instrumented variables associated with linkonce odr
 ;; functions are placed in the same comdat section.
 
 
 ; RUN: opt < %s -mtriple=x86_64-apple-macosx10.10.0 -instrprof -S | FileCheck %s
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux -instrprof -S | FileCheck %s
 ; RUN: opt < %s -mtriple=x86_64-pc-win32-coff -instrprof -S | FileCheck %s --check-prefix=COFF
 
 $_Z3barIvEvv = comdat any
 
 @__profn__Z3barIvEvv = linkonce_odr hidden constant [11 x i8] c"_Z3barIvEvv", align 1
 
-; CHECK: @__profn__Z3barIvEvv = linkonce_odr hidden constant [11 x i8] c"_Z3barIvEvv", section "{{.*}}__llvm_prf_names", comdat($__profv__Z3barIvEvv), align 1
+; CHECK: @__profn__Z3barIvEvv = private constant [11 x i8] c"_Z3barIvEvv", align 1
 ; CHECK: @__profc__Z3barIvEvv = linkonce_odr hidden global [1 x i64] zeroinitializer, section "{{.*}}__llvm_prf_cnts", comdat($__profv__Z3barIvEvv), align 8
-; CHECK: @__profd__Z3barIvEvv = linkonce_odr hidden global { i32, i32, i64, i8*, i64*, i8*, i8*, [1 x i16] } { i32 11, i32 1, i64 0, i8* getelementptr inbounds ([11 x i8], [11 x i8]* @__profn__Z3barIvEvv, i32 0, i32 0), i64* getelementptr inbounds ([1 x i64], [1 x i64]* @__profc__Z3barIvEvv, i32 0, i32 0), i8* null, i8* null, [1 x i16] zeroinitializer }, section "{{.*}}__llvm_prf_data", comdat($__profv__Z3barIvEvv), align 8
+; CHECK: @__profd__Z3barIvEvv = linkonce_odr hidden global { i64, i64, i64*, i8*, i8*, i32, [1 x i16] } { i64 4947693190065689389, i64 0, i64* getelementptr inbounds ([1 x i64], [1 x i64]* @__profc__Z3barIvEvv, i32 0, i32 0), i8* null, i8* null, i32 1, [1 x i16] zeroinitializer }, section "{{.*}}__llvm_prf_data", comdat($__profv__Z3barIvEvv), align 8
+; CHECK: @__llvm_prf_nm = private constant [{{.*}} x i8] c"{{.*}}", section "{{.*}}__llvm_prf_names"
 
-; COFF: @__profn__Z3barIvEvv = linkonce_odr hidden constant [11 x i8] c"_Z3barIvEvv", section "{{.*}}__llvm_prf_names", comdat, align 1
-; COFF: @__profc__Z3barIvEvv = linkonce_odr hidden global [1 x i64] zeroinitializer, section "{{.*}}__llvm_prf_cnts", comdat($__profn__Z3barIvEvv), align 8
-; COFF: @__profd__Z3barIvEvv = linkonce_odr hidden global { i32, i32, i64, i8*, i64*, i8*, i8*, [1 x i16] } { i32 11, i32 1, i64 0, i8* getelementptr inbounds ([11 x i8], [11 x i8]* @__profn__Z3barIvEvv, i32 0, i32 0), i64* getelementptr inbounds ([1 x i64], [1 x i64]* @__profc__Z3barIvEvv, i32 0, i32 0), i8* null, i8* null, [1 x i16] zeroinitializer }, section "{{.*}}__llvm_prf_data", comdat($__profn__Z3barIvEvv), align 8
+
+; COFF: @__profn__Z3barIvEvv = private constant [11 x i8] c"_Z3barIvEvv", align 1
+; COFF: @__profc__Z3barIvEvv = linkonce_odr hidden global [1 x i64] zeroinitializer, section "{{.*}}__llvm_prf_cnts", comdat, align 8
+; COFF: @__profd__Z3barIvEvv = linkonce_odr hidden global { i64, i64, i64*, i8*, i8*, i32, [1 x i16] } { i64 4947693190065689389, i64 0, i64* getelementptr inbounds ([1 x i64], [1 x i64]* @__profc__Z3barIvEvv, i32 0, i32 0), i8* null, i8* null, i32 1, [1 x i16] zeroinitializer }, section "{{.*}}__llvm_prf_data", comdat($__profc__Z3barIvEvv), align 8
 
 
 declare void @llvm.instrprof.increment(i8*, i64, i32, i32) #1
 
 define linkonce_odr void @_Z3barIvEvv() comdat {
 entry:
   call void @llvm.instrprof.increment(i8* getelementptr inbounds ([11 x i8], [11 x i8]* @__profn__Z3barIvEvv, i32 0, i32 0), i64 0, i32 1, i32 0)
   ret void
 }

llvm/trunk/test/Instrumentation/InstrProfiling/platform.ll

 ;; Checks for platform specific section names and initialization code.
 
 ; RUN: opt < %s -mtriple=x86_64-apple-macosx10.10.0 -instrprof -S | FileCheck %s -check-prefix=MACHO
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux -instrprof -S | FileCheck %s -check-prefix=LINUX
 ; RUN: opt < %s -mtriple=x86_64-unknown-freebsd -instrprof -S | FileCheck %s -check-prefix=FREEBSD
 ; RUN: opt < %s -mtriple=x86_64-pc-solaris -instrprof -S | FileCheck %s -check-prefix=SOLARIS
 
 @__profn_foo = hidden constant [3 x i8] c"foo"
-; MACHO: @__profn_foo = hidden constant [3 x i8] c"foo", section "__DATA,__llvm_prf_names", align 1
-; ELF: @__profn_foo = hidden constant [3 x i8] c"foo", section "__llvm_prf_names", align 1
+; MACHO: @__profn_foo = private constant [3 x i8] c"foo"
+; ELF: @__profn_foo = private constant [3 x i8] c"foo"
 
 ; MACHO: @__profc_foo = hidden global [1 x i64] zeroinitializer, section "__DATA,__llvm_prf_cnts", align 8
 ; ELF: @__profc_foo = hidden global [1 x i64] zeroinitializer, section "__llvm_prf_cnts", align 8
 
 ; MACHO: @__profd_foo = hidden {{.*}}, section "__DATA,__llvm_prf_data", align 8
 ; LINUX: @__profd_foo = hidden {{.*}}, section "__llvm_prf_data", align 8
 ; FREEBSD: @__profd_foo = hidden {{.*}}, section "__llvm_prf_data", align 8
 ; SOLARIS: @__profd_foo = hidden {{.*}}, section "__llvm_prf_data", align 8
 
+; ELF: @__llvm_prf_nm = private constant [{{.*}} x i8] c"{{.*}}", section "{{.*}}__llvm_prf_names"
+
 define void @foo() {
   call void @llvm.instrprof.increment(i8* getelementptr inbounds ([3 x i8], [3 x i8]* @__profn_foo, i32 0, i32 0), i64 0, i32 1, i32 0)
   ret void
 }
 
 declare void @llvm.instrprof.increment(i8*, i64, i32, i32)
 
 ;; Emit registration functions for platforms that don't find the

llvm/trunk/test/Instrumentation/InstrProfiling/profiling.ll

 ; RUN: opt < %s -instrprof -S | FileCheck %s
 
 target triple = "x86_64-apple-macosx10.10.0"
 
 @__profn_foo = hidden constant [3 x i8] c"foo"
-; CHECK: @__profn_foo = hidden constant [3 x i8] c"foo", section "__DATA,__llvm_prf_names", align 1
+; CHECK: @__profn_foo = private constant [3 x i8] c"foo"
 @__profn_bar = hidden constant [4 x i8] c"bar\00"
-; CHECK: @__profn_bar = hidden constant [4 x i8] c"bar\00", section "__DATA,__llvm_prf_names", align 1
+; CHECK: @__profn_bar = private constant [4 x i8] c"bar\00"
 @__profn_baz = hidden constant [3 x i8] c"baz"
-; CHECK: @__profn_baz = hidden constant [3 x i8] c"baz", section "__DATA,__llvm_prf_names", align 1
+; CHECK: @__profn_baz = private constant [3 x i8] c"baz"
 
 ; CHECK: @__profc_foo = hidden global [1 x i64] zeroinitializer, section "__DATA,__llvm_prf_cnts", align 8
 ; CHECK: @__profd_foo = hidden {{.*}}, section "__DATA,__llvm_prf_data", align 8
 define void @foo() {
   call void @llvm.instrprof.increment(i8* getelementptr inbounds ([3 x i8], [3 x i8]* @__profn_foo, i32 0, i32 0), i64 0, i32 1, i32 0)
   ret void
 }
 

llvm/trunk/test/tools/llvm-profdata/raw-32-bits-be.test

 RUN: printf '\377lprofR\201' > %t
-RUN: printf '\0\0\0\0\0\0\0\2' >> %t
+RUN: printf '\0\0\0\0\0\0\0\3' >> %t
 RUN: printf '\0\0\0\0\0\0\0\2' >> %t
 RUN: printf '\0\0\0\0\0\0\0\3' >> %t
-RUN: printf '\0\0\0\0\0\0\0\6' >> %t
+RUN: printf '\0\0\0\0\0\0\0\20' >> %t
 RUN: printf '\0\0\0\0\1\0\0\0' >> %t
 RUN: printf '\0\0\0\0\2\0\0\0' >> %t
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t
 
-RUN: printf '\0\0\0\3' >> %t
-RUN: printf '\0\0\0\1' >> %t
+RUN: printf '\134\370\302\114\333\030\275\254' >> %t
 RUN: printf '\0\0\0\0\0\0\0\1' >> %t
-RUN: printf '\2\0\0\0' >> %t
 RUN: printf '\1\0\0\0' >> %t
 RUN: printf '\0\0\0\0' >> %t
 RUN: printf '\0\0\0\0' >> %t
+RUN: printf '\0\0\0\1' >> %t
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t
 
-RUN: printf '\0\0\0\3' >> %t
-RUN: printf '\0\0\0\2' >> %t
+RUN: printf '\344\023\165\112\031\035\265\067' >> %t
 RUN: printf '\0\0\0\0\0\0\0\2' >> %t
-RUN: printf '\2\0\0\03' >> %t
 RUN: printf '\1\0\0\10' >> %t
 RUN: printf '\0\0\0\0' >> %t
 RUN: printf '\0\0\0\0' >> %t
+RUN: printf '\0\0\0\2' >> %t
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t
 
 RUN: printf '\0\0\0\0\0\0\0\023' >> %t
 RUN: printf '\0\0\0\0\0\0\0\067' >> %t
 RUN: printf '\0\0\0\0\0\0\0\101' >> %t
-RUN: printf 'foobar\0\0' >> %t
+RUN: printf '\7\0foo bar\0\0\0\0\0\0\0' >> %t
 
 RUN: llvm-profdata show %t -all-functions -counts | FileCheck %s
 
 CHECK: Counters:
 CHECK:   foo:
 CHECK:     Hash: 0x0000000000000001
 CHECK:     Counters: 1
 CHECK:     Function count: 19

llvm/trunk/test/tools/llvm-profdata/raw-32-bits-le.test

 RUN: printf '\201Rforpl\377' > %t
-RUN: printf '\2\0\0\0\0\0\0\0' >> %t
+RUN: printf '\3\0\0\0\0\0\0\0' >> %t
 RUN: printf '\2\0\0\0\0\0\0\0' >> %t
 RUN: printf '\3\0\0\0\0\0\0\0' >> %t
-RUN: printf '\6\0\0\0\0\0\0\0' >> %t
+RUN: printf '\20\0\0\0\0\0\0\0' >> %t
 RUN: printf '\0\0\0\1\0\0\0\0' >> %t
 RUN: printf '\0\0\0\2\0\0\0\0' >> %t
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t
 
-RUN: printf '\3\0\0\0' >> %t
-RUN: printf '\1\0\0\0' >> %t
+RUN: printf '\254\275\030\333\114\302\370\134' >> %t
 RUN: printf '\1\0\0\0\0\0\0\0' >> %t
-RUN: printf '\0\0\0\2' >> %t
 RUN: printf '\0\0\0\1' >> %t
 RUN: printf '\0\0\0\0' >> %t
 RUN: printf '\0\0\0\0' >> %t
+RUN: printf '\1\0\0\0' >> %t
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t
 
-RUN: printf '\3\0\0\0' >> %t
-RUN: printf '\2\0\0\0' >> %t
+RUN: printf '\067\265\035\031\112\165\023\344' >> %t
 RUN: printf '\02\0\0\0\0\0\0\0' >> %t
-RUN: printf '\03\0\0\2' >> %t
 RUN: printf '\10\0\0\1' >> %t
 RUN: printf '\0\0\0\0' >> %t
 RUN: printf '\0\0\0\0' >> %t
+RUN: printf '\2\0\0\0' >> %t
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t
 
 RUN: printf '\023\0\0\0\0\0\0\0' >> %t
 RUN: printf '\067\0\0\0\0\0\0\0' >> %t
 RUN: printf '\101\0\0\0\0\0\0\0' >> %t
-RUN: printf 'foobar\0\0' >> %t
+RUN: printf '\7\0foo bar\0\0\0\0\0\0\0' >> %t
 
 RUN: llvm-profdata show %t -all-functions -counts | FileCheck %s
 
 CHECK: Counters:
 CHECK:   foo:
 CHECK:     Hash: 0x0000000000000001
 CHECK:     Counters: 1
 CHECK:     Function count: 19

llvm/trunk/test/tools/llvm-profdata/raw-64-bits-be.test

 RUN: printf '\377lprofr\201' > %t
-RUN: printf '\0\0\0\0\0\0\0\2' >> %t
+RUN: printf '\0\0\0\0\0\0\0\3' >> %t
 RUN: printf '\0\0\0\0\0\0\0\2' >> %t
 RUN: printf '\0\0\0\0\0\0\0\3' >> %t
-RUN: printf '\0\0\0\0\0\0\0\6' >> %t
+RUN: printf '\0\0\0\0\0\0\0\20' >> %t
 RUN: printf '\0\0\0\1\0\4\0\0' >> %t
 RUN: printf '\0\0\0\2\0\4\0\0' >> %t
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t
 
-RUN: printf '\0\0\0\3' >> %t
-RUN: printf '\0\0\0\1' >> %t
+RUN: printf '\134\370\302\114\333\030\275\254' >> %t
 RUN: printf '\0\0\0\0\0\0\0\1' >> %t
-RUN: printf '\0\0\0\2\0\4\0\0' >> %t
 RUN: printf '\0\0\0\1\0\4\0\0' >> %t
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t
-RUN: printf '\0\0\0\0\0\0\0\0' >> %t
+RUN: printf '\0\0\0\1\0\0\0\0' >> %t
 
-RUN: printf '\0\0\0\3' >> %t
-RUN: printf '\0\0\0\2' >> %t
+RUN: printf '\344\023\165\112\031\035\265\067' >> %t
 RUN: printf '\0\0\0\0\0\0\0\02' >> %t
-RUN: printf '\0\0\0\2\0\4\0\03' >> %t
 RUN: printf '\0\0\0\1\0\4\0\10' >> %t
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t
-RUN: printf '\0\0\0\0\0\0\0\0' >> %t
+RUN: printf '\0\0\0\02\0\0\0\0' >> %t
 
 RUN: printf '\0\0\0\0\0\0\0\023' >> %t
 RUN: printf '\0\0\0\0\0\0\0\067' >> %t
 RUN: printf '\0\0\0\0\0\0\0\101' >> %t
-RUN: printf 'foobar\0\0' >> %t
+RUN: printf '\7\0foo bar\0\0\0\0\0\0\0' >> %t
 
 RUN: llvm-profdata show %t -all-functions -counts | FileCheck %s
 
 CHECK: Counters:
 CHECK:   foo:
 CHECK:     Hash: 0x0000000000000001
 CHECK:     Counters: 1
 CHECK:     Function count: 19

llvm/trunk/test/tools/llvm-profdata/raw-64-bits-le.test

 RUN: printf '\201rforpl\377' > %t
-RUN: printf '\2\0\0\0\0\0\0\0' >> %t
+RUN: printf '\3\0\0\0\0\0\0\0' >> %t
 RUN: printf '\2\0\0\0\0\0\0\0' >> %t
 RUN: printf '\3\0\0\0\0\0\0\0' >> %t
-RUN: printf '\6\0\0\0\0\0\0\0' >> %t
+RUN: printf '\20\0\0\0\0\0\0\0' >> %t
 RUN: printf '\0\0\4\0\1\0\0\0' >> %t
 RUN: printf '\0\0\4\0\2\0\0\0' >> %t
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t
 
-RUN: printf '\3\0\0\0' >> %t
-RUN: printf '\1\0\0\0' >> %t
+RUN: printf '\254\275\030\333\114\302\370\134' >> %t
 RUN: printf '\1\0\0\0\0\0\0\0' >> %t
-RUN: printf '\0\0\4\0\2\0\0\0' >> %t
 RUN: printf '\0\0\4\0\1\0\0\0' >> %t
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t
-RUN: printf '\0\0\0\0\0\0\0\0' >> %t
+RUN: printf '\1\0\0\0\0\0\0\0' >> %t
 
-RUN: printf '\03\0\0\0' >> %t
-RUN: printf '\02\0\0\0' >> %t
+RUN: printf '\067\265\035\031\112\165\023\344' >> %t
 RUN: printf '\02\0\0\0\0\0\0\0' >> %t
-RUN: printf '\03\0\4\0\2\0\0\0' >> %t
 RUN: printf '\10\0\4\0\1\0\0\0' >> %t
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t
-RUN: printf '\0\0\0\0\0\0\0\0' >> %t
+RUN: printf '\02\0\0\0\0\0\0\0' >> %t
 
 RUN: printf '\023\0\0\0\0\0\0\0' >> %t
 RUN: printf '\067\0\0\0\0\0\0\0' >> %t
 RUN: printf '\101\0\0\0\0\0\0\0' >> %t
-RUN: printf 'foobar\0\0' >> %t
+RUN: printf '\7\0foo bar\0\0\0\0\0\0\0' >> %t
 
 RUN: llvm-profdata show %t -all-functions -counts | FileCheck %s
 
 CHECK: Counters:
 CHECK:   foo:
 CHECK:     Hash: 0x0000000000000001
 CHECK:     Counters: 1
 CHECK:     Function count: 19

llvm/trunk/test/tools/llvm-profdata/raw-two-profiles.test

 RUN: printf '\201rforpl\377' > %t-foo.profraw
-RUN: printf '\2\0\0\0\0\0\0\0' >> %t-foo.profraw
+RUN: printf '\3\0\0\0\0\0\0\0' >> %t-foo.profraw
 RUN: printf '\1\0\0\0\0\0\0\0' >> %t-foo.profraw
 RUN: printf '\1\0\0\0\0\0\0\0' >> %t-foo.profraw
-RUN: printf '\3\0\0\0\0\0\0\0' >> %t-foo.profraw
+RUN: printf '\10\0\0\0\0\0\0\0' >> %t-foo.profraw
 RUN: printf '\0\0\4\0\1\0\0\0' >> %t-foo.profraw
 RUN: printf '\0\0\4\0\2\0\0\0' >> %t-foo.profraw
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t-foo.profraw
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t-foo.profraw
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t-foo.profraw
 
-RUN: printf '\3\0\0\0' >> %t-foo.profraw
-RUN: printf '\1\0\0\0' >> %t-foo.profraw
+RUN: printf '\254\275\030\333\114\302\370\134' >> %t-foo.profraw
 RUN: printf '\1\0\0\0\0\0\0\0' >> %t-foo.profraw
-RUN: printf '\0\0\4\0\2\0\0\0' >> %t-foo.profraw
 RUN: printf '\0\0\4\0\1\0\0\0' >> %t-foo.profraw
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t-foo.profraw
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t-foo.profraw
-RUN: printf '\0\0\0\0\0\0\0\0' >> %t-foo.profraw
+RUN: printf '\1\0\0\0\0\0\0\0' >> %t-foo.profraw
 
 RUN: printf '\023\0\0\0\0\0\0\0' >> %t-foo.profraw
-RUN: printf 'foo\0\0\0\0\0' >> %t-foo.profraw
+RUN: printf '\3\0foo\0\0\0' >> %t-foo.profraw
 
 RUN: printf '\201rforpl\377' > %t-bar.profraw
-RUN: printf '\2\0\0\0\0\0\0\0' >> %t-bar.profraw
+RUN: printf '\3\0\0\0\0\0\0\0' >> %t-bar.profraw
 RUN: printf '\1\0\0\0\0\0\0\0' >> %t-bar.profraw
 RUN: printf '\2\0\0\0\0\0\0\0' >> %t-bar.profraw
-RUN: printf '\3\0\0\0\0\0\0\0' >> %t-bar.profraw
+RUN: printf '\10\0\0\0\0\0\0\0' >> %t-bar.profraw
 RUN: printf '\0\0\6\0\1\0\0\0' >> %t-bar.profraw
 RUN: printf '\0\0\6\0\2\0\0\0' >> %t-bar.profraw
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t-bar.profraw
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t-bar.profraw
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t-bar.profraw
 
-RUN: printf '\3\0\0\0' >> %t-bar.profraw
-RUN: printf '\2\0\0\0' >> %t-bar.profraw
-RUN: printf '\2\0\0\0\0\0\0\0' >> %t-bar.profraw
-RUN: printf '\0\0\6\0\2\0\0\0' >> %t-bar.profraw
+RUN: printf '\067\265\035\031\112\165\023\344' >> %t-bar.profraw
+RUN: printf '\02\0\0\0\0\0\0\0' >> %t-bar.profraw
 RUN: printf '\0\0\6\0\1\0\0\0' >> %t-bar.profraw
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t-bar.profraw
 RUN: printf '\0\0\0\0\0\0\0\0' >> %t-bar.profraw
-RUN: printf '\0\0\0\0\0\0\0\0' >> %t-bar.profraw
+RUN: printf '\02\0\0\0\0\0\0\0' >> %t-bar.profraw
 
 RUN: printf '\067\0\0\0\0\0\0\0' >> %t-bar.profraw
 RUN: printf '\101\0\0\0\0\0\0\0' >> %t-bar.profraw
-RUN: printf 'bar\0\0\0\0\0' >> %t-bar.profraw
+RUN: printf '\3\0bar\0\0\0' >> %t-bar.profraw
 
 RUN: cat %t-foo.profraw %t-bar.profraw > %t-pad.profraw
 RUN: llvm-profdata show %t-pad.profraw -all-functions -counts | FileCheck %s
 
 CHECK: Counters:
 CHECK:   foo:
 CHECK:     Hash: 0x0000000000000001
 CHECK:     Counters: 1