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 
20 using namespace llvm;
21 
22 namespace {
23 static support::endianness ValueProfDataEndianness = support::little;
24 
25 class InstrProfRecordTrait {
26 public:
27   typedef StringRef key_type;
28   typedef StringRef key_type_ref;
29 
30   typedef const InstrProfWriter::ProfilingData *const data_type;
31   typedef const InstrProfWriter::ProfilingData *const data_type_ref;
32 
33   typedef uint64_t hash_value_type;
34   typedef uint64_t offset_type;
35 
36   static hash_value_type ComputeHash(key_type_ref K) {
37     return IndexedInstrProf::ComputeHash(IndexedInstrProf::HashType, K);
38   }
39 
40   static std::pair<offset_type, offset_type>
41   EmitKeyDataLength(raw_ostream &Out, key_type_ref K, data_type_ref V) {
42     using namespace llvm::support;
43     endian::Writer<little> LE(Out);
44 
45     offset_type N = K.size();
46     LE.write<offset_type>(N);
47 
48     offset_type M = 0;
49     for (const auto &ProfileData : *V) {
50       const InstrProfRecord &ProfRecord = ProfileData.second;
51       M += sizeof(uint64_t); // The function hash
52       M += sizeof(uint64_t); // The size of the Counts vector
53       M += ProfRecord.Counts.size() * sizeof(uint64_t);
54 
55       // Value data
56       M += IndexedInstrProf::ValueProfData::getSize(ProfileData.second);
57     }
58     LE.write<offset_type>(M);
59 
60     return std::make_pair(N, M);
61   }
62 
63   static void EmitKey(raw_ostream &Out, key_type_ref K, offset_type N){
64     Out.write(K.data(), N);
65   }
66 
67   static void EmitData(raw_ostream &Out, key_type_ref, data_type_ref V,
68                        offset_type) {
69     using namespace llvm::support;
70     endian::Writer<little> LE(Out);
71     for (const auto &ProfileData : *V) {
72       const InstrProfRecord &ProfRecord = ProfileData.second;
73 
74       LE.write<uint64_t>(ProfileData.first); // Function hash
75       LE.write<uint64_t>(ProfRecord.Counts.size());
76       for (uint64_t I : ProfRecord.Counts)
77         LE.write<uint64_t>(I);
78 
79       // Write value data
80       std::unique_ptr<IndexedInstrProf::ValueProfData> VDataPtr =
81           IndexedInstrProf::ValueProfData::serializeFrom(ProfileData.second);
82       uint32_t S = VDataPtr->getSize();
83       VDataPtr->swapBytesFromHost(ValueProfDataEndianness);
84       Out.write((const char *)VDataPtr.get(), S);
85     }
86   }
87 };
88 }
89 
90 static std::error_code combineInstrProfRecords(InstrProfRecord &Dest,
91                                                InstrProfRecord &Source,
92                                                uint64_t &MaxFunctionCount) {
93   // If the number of counters doesn't match we either have bad data
94   // or a hash collision.
95   if (Dest.Counts.size() != Source.Counts.size())
96     return instrprof_error::count_mismatch;
97 
98   for (size_t I = 0, E = Source.Counts.size(); I < E; ++I) {
99     if (Dest.Counts[I] + Source.Counts[I] < Dest.Counts[I])
100       return instrprof_error::counter_overflow;
101     Dest.Counts[I] += Source.Counts[I];
102   }
103 
104   for (uint32_t Kind = IPVK_First; Kind <= IPVK_Last; ++Kind) {
105     if (std::error_code EC = Dest.mergeValueProfData(Kind, Source))
106       return EC;
107   }
108 
109   // We keep track of the max function count as we go for simplicity.
110   if (Dest.Counts[0] > MaxFunctionCount)
111     MaxFunctionCount = Dest.Counts[0];
112 
113   return instrprof_error::success;
114 }
115 
116 // Internal interface for testing purpose only.
117 void InstrProfWriter::setValueProfDataEndianness(
118     support::endianness Endianness) {
119   ValueProfDataEndianness = Endianness;
120 }
121 
122 void InstrProfWriter::updateStringTableReferences(InstrProfRecord &I) {
123   I.updateStrings(&StringTable);
124 }
125 
126 std::error_code InstrProfWriter::addRecord(InstrProfRecord &&I) {
127   updateStringTableReferences(I);
128   auto &ProfileDataMap = FunctionData[I.Name];
129 
130   auto Where = ProfileDataMap.find(I.Hash);
131   if (Where == ProfileDataMap.end()) {
132     // We've never seen a function with this name and hash, add it.
133     ProfileDataMap[I.Hash] = I;
134 
135     // We keep track of the max function count as we go for simplicity.
136     if (I.Counts[0] > MaxFunctionCount)
137       MaxFunctionCount = I.Counts[0];
138     return instrprof_error::success;
139   }
140 
141   // We're updating a function we've seen before.
142   return combineInstrProfRecords(Where->second, I, MaxFunctionCount);
143 }
144 
145 std::pair<uint64_t, uint64_t> InstrProfWriter::writeImpl(raw_ostream &OS) {
146   OnDiskChainedHashTableGenerator<InstrProfRecordTrait> Generator;
147 
148   // Populate the hash table generator.
149   for (const auto &I : FunctionData)
150     Generator.insert(I.getKey(), &I.getValue());
151 
152   using namespace llvm::support;
153   endian::Writer<little> LE(OS);
154 
155   // Write the header.
156   IndexedInstrProf::Header Header;
157   Header.Magic = IndexedInstrProf::Magic;
158   Header.Version = IndexedInstrProf::Version;
159   Header.MaxFunctionCount = MaxFunctionCount;
160   Header.HashType = static_cast<uint64_t>(IndexedInstrProf::HashType);
161   Header.HashOffset = 0;
162   int N = sizeof(IndexedInstrProf::Header) / sizeof(uint64_t);
163 
164   // Only write out all the fields execpt 'HashOffset'. We need
165   // to remember the offset of that field to allow back patching
166   // later.
167   for (int I = 0; I < N - 1; I++)
168     LE.write<uint64_t>(reinterpret_cast<uint64_t *>(&Header)[I]);
169 
170   // Save a space to write the hash table start location.
171   uint64_t HashTableStartLoc = OS.tell();
172   // Reserve the space for HashOffset field.
173   LE.write<uint64_t>(0);
174   // Write the hash table.
175   uint64_t HashTableStart = Generator.Emit(OS);
176 
177   return std::make_pair(HashTableStartLoc, HashTableStart);
178 }
179 
180 void InstrProfWriter::write(raw_fd_ostream &OS) {
181   // Write the hash table.
182   auto TableStart = writeImpl(OS);
183 
184   // Go back and fill in the hash table start.
185   using namespace support;
186   OS.seek(TableStart.first);
187   // Now patch the HashOffset field previously reserved.
188   endian::Writer<little>(OS).write<uint64_t>(TableStart.second);
189 }
190 
191 std::unique_ptr<MemoryBuffer> InstrProfWriter::writeBuffer() {
192   std::string Data;
193   llvm::raw_string_ostream OS(Data);
194   // Write the hash table.
195   auto TableStart = writeImpl(OS);
196   OS.flush();
197 
198   // Go back and fill in the hash table start.
199   using namespace support;
200   uint64_t Bytes = endian::byte_swap<uint64_t, little>(TableStart.second);
201   Data.replace(TableStart.first, sizeof(uint64_t), (const char *)&Bytes,
202                sizeof(uint64_t));
203 
204   // Return this in an aligned memory buffer.
205   return MemoryBuffer::getMemBufferCopy(Data);
206 }
207