1 //=-- InstrProfWriter.cpp - Instrumented profiling writer -------------------=//
2 //
3 //                     The LLVM Compiler Infrastructure
4 //
5 // This file is distributed under the University of Illinois Open Source
6 // License. See LICENSE.TXT for details.
7 //
8 //===----------------------------------------------------------------------===//
9 //
10 // This file contains support for writing profiling data for clang's
11 // instrumentation based PGO and coverage.
12 //
13 //===----------------------------------------------------------------------===//
14 
15 #include "llvm/ProfileData/InstrProfWriter.h"
16 #include "llvm/ADT/StringExtras.h"
17 #include "llvm/Support/EndianStream.h"
18 #include "llvm/Support/OnDiskHashTable.h"
19 #include <tuple>
20 
21 using namespace llvm;
22 
23 namespace {
24 static support::endianness ValueProfDataEndianness = support::little;
25 
26 class InstrProfRecordTrait {
27 public:
28   typedef StringRef key_type;
29   typedef StringRef key_type_ref;
30 
31   typedef const InstrProfWriter::ProfilingData *const data_type;
32   typedef const InstrProfWriter::ProfilingData *const data_type_ref;
33 
34   typedef uint64_t hash_value_type;
35   typedef uint64_t offset_type;
36 
37   static hash_value_type ComputeHash(key_type_ref K) {
38     return IndexedInstrProf::ComputeHash(IndexedInstrProf::HashType, K);
39   }
40 
41   static std::pair<offset_type, offset_type>
42   EmitKeyDataLength(raw_ostream &Out, key_type_ref K, data_type_ref V) {
43     using namespace llvm::support;
44     endian::Writer<little> LE(Out);
45 
46     offset_type N = K.size();
47     LE.write<offset_type>(N);
48 
49     offset_type M = 0;
50     for (const auto &ProfileData : *V) {
51       const InstrProfRecord &ProfRecord = ProfileData.second;
52       M += sizeof(uint64_t); // The function hash
53       M += sizeof(uint64_t); // The size of the Counts vector
54       M += ProfRecord.Counts.size() * sizeof(uint64_t);
55 
56       // Value data
57       M += ValueProfData::getSize(ProfileData.second);
58     }
59     LE.write<offset_type>(M);
60 
61     return std::make_pair(N, M);
62   }
63 
64   static void EmitKey(raw_ostream &Out, key_type_ref K, offset_type N){
65     Out.write(K.data(), N);
66   }
67 
68   static void EmitData(raw_ostream &Out, key_type_ref, data_type_ref V,
69                        offset_type) {
70     using namespace llvm::support;
71     endian::Writer<little> LE(Out);
72     for (const auto &ProfileData : *V) {
73       const InstrProfRecord &ProfRecord = ProfileData.second;
74 
75       LE.write<uint64_t>(ProfileData.first); // Function hash
76       LE.write<uint64_t>(ProfRecord.Counts.size());
77       for (uint64_t I : ProfRecord.Counts)
78         LE.write<uint64_t>(I);
79 
80       // Write value data
81       std::unique_ptr<ValueProfData> VDataPtr =
82           ValueProfData::serializeFrom(ProfileData.second);
83       uint32_t S = VDataPtr->getSize();
84       VDataPtr->swapBytesFromHost(ValueProfDataEndianness);
85       Out.write((const char *)VDataPtr.get(), S);
86     }
87   }
88 };
89 }
90 
91 // Internal interface for testing purpose only.
92 void InstrProfWriter::setValueProfDataEndianness(
93     support::endianness Endianness) {
94   ValueProfDataEndianness = Endianness;
95 }
96 
97 void InstrProfWriter::updateStringTableReferences(InstrProfRecord &I) {
98   I.updateStrings(&StringTable);
99 }
100 
101 std::error_code InstrProfWriter::addRecord(InstrProfRecord &&I) {
102   updateStringTableReferences(I);
103   auto &ProfileDataMap = FunctionData[I.Name];
104 
105   bool NewFunc;
106   ProfilingData::iterator Where;
107   std::tie(Where, NewFunc) =
108       ProfileDataMap.insert(std::make_pair(I.Hash, InstrProfRecord()));
109   InstrProfRecord &Dest = Where->second;
110 
111   instrprof_error Result;
112   if (NewFunc) {
113     // We've never seen a function with this name and hash, add it.
114     Dest = std::move(I);
115     Result = instrprof_error::success;
116   } else {
117     // We're updating a function we've seen before.
118     Result = Dest.merge(I);
119   }
120 
121   // We keep track of the max function count as we go for simplicity.
122   // Update this statistic no matter the result of the merge.
123   if (Dest.Counts[0] > MaxFunctionCount)
124     MaxFunctionCount = Dest.Counts[0];
125 
126   return Result;
127 }
128 
129 std::pair<uint64_t, uint64_t> InstrProfWriter::writeImpl(raw_ostream &OS) {
130   OnDiskChainedHashTableGenerator<InstrProfRecordTrait> Generator;
131 
132   // Populate the hash table generator.
133   for (const auto &I : FunctionData)
134     Generator.insert(I.getKey(), &I.getValue());
135 
136   using namespace llvm::support;
137   endian::Writer<little> LE(OS);
138 
139   // Write the header.
140   IndexedInstrProf::Header Header;
141   Header.Magic = IndexedInstrProf::Magic;
142   Header.Version = IndexedInstrProf::Version;
143   Header.MaxFunctionCount = MaxFunctionCount;
144   Header.HashType = static_cast<uint64_t>(IndexedInstrProf::HashType);
145   Header.HashOffset = 0;
146   int N = sizeof(IndexedInstrProf::Header) / sizeof(uint64_t);
147 
148   // Only write out all the fields execpt 'HashOffset'. We need
149   // to remember the offset of that field to allow back patching
150   // later.
151   for (int I = 0; I < N - 1; I++)
152     LE.write<uint64_t>(reinterpret_cast<uint64_t *>(&Header)[I]);
153 
154   // Save a space to write the hash table start location.
155   uint64_t HashTableStartLoc = OS.tell();
156   // Reserve the space for HashOffset field.
157   LE.write<uint64_t>(0);
158   // Write the hash table.
159   uint64_t HashTableStart = Generator.Emit(OS);
160 
161   return std::make_pair(HashTableStartLoc, HashTableStart);
162 }
163 
164 void InstrProfWriter::write(raw_fd_ostream &OS) {
165   // Write the hash table.
166   auto TableStart = writeImpl(OS);
167 
168   // Go back and fill in the hash table start.
169   using namespace support;
170   OS.seek(TableStart.first);
171   // Now patch the HashOffset field previously reserved.
172   endian::Writer<little>(OS).write<uint64_t>(TableStart.second);
173 }
174 
175 void InstrProfWriter::writeRecordInText(const InstrProfRecord &Func,
176                                         raw_fd_ostream &OS) {
177   OS << Func.Name << "\n";
178   OS << "# Func Hash:\n" << Func.Hash << "\n";
179   OS << "# Num Counters:\n" <<Func.Counts.size() << "\n";
180   OS << "# Counter Values:\n";
181   for (uint64_t Count : Func.Counts)
182     OS << Count << "\n";
183 
184   OS << "\n";
185 }
186 
187 void InstrProfWriter::writeText(raw_fd_ostream &OS) {
188   for (const auto &I : FunctionData)
189     for (const auto &Func : I.getValue())
190       writeRecordInText(Func.second, OS);
191 }
192 
193 std::unique_ptr<MemoryBuffer> InstrProfWriter::writeBuffer() {
194   std::string Data;
195   llvm::raw_string_ostream OS(Data);
196   // Write the hash table.
197   auto TableStart = writeImpl(OS);
198   OS.flush();
199 
200   // Go back and fill in the hash table start.
201   using namespace support;
202   uint64_t Bytes = endian::byte_swap<uint64_t, little>(TableStart.second);
203   Data.replace(TableStart.first, sizeof(uint64_t), (const char *)&Bytes,
204                sizeof(uint64_t));
205 
206   // Return this in an aligned memory buffer.
207   return MemoryBuffer::getMemBufferCopy(Data);
208 }
209