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 // A struct to define how the data stream should be patched. For Indexed
24 // profiling, only uint64_t data type is needed.
25 struct PatchItem {
26   uint64_t Pos; // Where to patch.
27   uint64_t *D;  // Pointer to an array of source data.
28   int N;        // Number of elements in \c D array.
29 };
30 
31 namespace llvm {
32 // A wrapper class to abstract writer stream with support of bytes
33 // back patching.
34 class ProfOStream {
35 
36 public:
37   ProfOStream(llvm::raw_fd_ostream &FD) : IsFDOStream(true), OS(FD), LE(FD) {}
38   ProfOStream(llvm::raw_string_ostream &STR)
39       : IsFDOStream(false), OS(STR), LE(STR) {}
40 
41   uint64_t tell() { return OS.tell(); }
42   void write(uint64_t V) { LE.write<uint64_t>(V); }
43   // \c patch can only be called when all data is written and flushed.
44   // For raw_string_ostream, the patch is done on the target string
45   // directly and it won't be reflected in the stream's internal buffer.
46   void patch(PatchItem *P, int NItems) {
47     using namespace support;
48     if (IsFDOStream) {
49       llvm::raw_fd_ostream &FDOStream = static_cast<llvm::raw_fd_ostream &>(OS);
50       for (int K = 0; K < NItems; K++) {
51         FDOStream.seek(P[K].Pos);
52         for (int I = 0; I < P[K].N; I++)
53           write(P[K].D[I]);
54       }
55     } else {
56       llvm::raw_string_ostream &SOStream =
57           static_cast<llvm::raw_string_ostream &>(OS);
58       std::string &Data = SOStream.str(); // with flush
59       for (int K = 0; K < NItems; K++) {
60         for (int I = 0; I < P[K].N; I++) {
61           uint64_t Bytes = endian::byte_swap<uint64_t, little>(P[K].D[I]);
62           Data.replace(P[K].Pos + I * sizeof(uint64_t), sizeof(uint64_t),
63                        (const char *)&Bytes, sizeof(uint64_t));
64         }
65       }
66     }
67   }
68   // If \c OS is an instance of \c raw_fd_ostream, this field will be
69   // true. Otherwise, \c OS will be an raw_string_ostream.
70   bool IsFDOStream;
71   raw_ostream &OS;
72   support::endian::Writer<support::little> LE;
73 };
74 
75 class InstrProfRecordWriterTrait {
76 public:
77   typedef StringRef key_type;
78   typedef StringRef key_type_ref;
79 
80   typedef const InstrProfWriter::ProfilingData *const data_type;
81   typedef const InstrProfWriter::ProfilingData *const data_type_ref;
82 
83   typedef uint64_t hash_value_type;
84   typedef uint64_t offset_type;
85 
86   support::endianness ValueProfDataEndianness;
87   InstrProfSummaryBuilder *SummaryBuilder;
88 
89   InstrProfRecordWriterTrait() : ValueProfDataEndianness(support::little) {}
90   static hash_value_type ComputeHash(key_type_ref K) {
91     return IndexedInstrProf::ComputeHash(K);
92   }
93 
94   static std::pair<offset_type, offset_type>
95   EmitKeyDataLength(raw_ostream &Out, key_type_ref K, data_type_ref V) {
96     using namespace llvm::support;
97     endian::Writer<little> LE(Out);
98 
99     offset_type N = K.size();
100     LE.write<offset_type>(N);
101 
102     offset_type M = 0;
103     for (const auto &ProfileData : *V) {
104       const InstrProfRecord &ProfRecord = ProfileData.second;
105       M += sizeof(uint64_t); // The function hash
106       M += sizeof(uint64_t); // The size of the Counts vector
107       M += ProfRecord.Counts.size() * sizeof(uint64_t);
108 
109       // Value data
110       M += ValueProfData::getSize(ProfileData.second);
111     }
112     LE.write<offset_type>(M);
113 
114     return std::make_pair(N, M);
115   }
116 
117   void EmitKey(raw_ostream &Out, key_type_ref K, offset_type N) {
118     Out.write(K.data(), N);
119   }
120 
121   void EmitData(raw_ostream &Out, key_type_ref, data_type_ref V, offset_type) {
122     using namespace llvm::support;
123     endian::Writer<little> LE(Out);
124     for (const auto &ProfileData : *V) {
125       const InstrProfRecord &ProfRecord = ProfileData.second;
126       SummaryBuilder->addRecord(ProfRecord);
127 
128       LE.write<uint64_t>(ProfileData.first); // Function hash
129       LE.write<uint64_t>(ProfRecord.Counts.size());
130       for (uint64_t I : ProfRecord.Counts)
131         LE.write<uint64_t>(I);
132 
133       // Write value data
134       std::unique_ptr<ValueProfData> VDataPtr =
135           ValueProfData::serializeFrom(ProfileData.second);
136       uint32_t S = VDataPtr->getSize();
137       VDataPtr->swapBytesFromHost(ValueProfDataEndianness);
138       Out.write((const char *)VDataPtr.get(), S);
139     }
140   }
141 };
142 }
143 
144 InstrProfWriter::InstrProfWriter(bool Sparse)
145     : Sparse(Sparse), FunctionData(), ProfileKind(PF_Unknown),
146       InfoObj(new InstrProfRecordWriterTrait()) {}
147 
148 InstrProfWriter::~InstrProfWriter() { delete InfoObj; }
149 
150 // Internal interface for testing purpose only.
151 void InstrProfWriter::setValueProfDataEndianness(
152     support::endianness Endianness) {
153   InfoObj->ValueProfDataEndianness = Endianness;
154 }
155 void InstrProfWriter::setOutputSparse(bool Sparse) {
156   this->Sparse = Sparse;
157 }
158 
159 Error InstrProfWriter::addRecord(InstrProfRecord &&I, uint64_t Weight) {
160   auto &ProfileDataMap = FunctionData[I.Name];
161 
162   bool NewFunc;
163   ProfilingData::iterator Where;
164   std::tie(Where, NewFunc) =
165       ProfileDataMap.insert(std::make_pair(I.Hash, InstrProfRecord()));
166   InstrProfRecord &Dest = Where->second;
167 
168   if (NewFunc) {
169     // We've never seen a function with this name and hash, add it.
170     Dest = std::move(I);
171     // Fix up the name to avoid dangling reference.
172     Dest.Name = FunctionData.find(Dest.Name)->getKey();
173     if (Weight > 1)
174       Dest.scale(Weight);
175   } else {
176     // We're updating a function we've seen before.
177     Dest.merge(I, Weight);
178   }
179 
180   Dest.sortValueData();
181 
182   return Dest.takeError();
183 }
184 
185 Error InstrProfWriter::mergeRecordsFromWriter(InstrProfWriter &&IPW) {
186   for (auto &I : IPW.FunctionData)
187     for (auto &Func : I.getValue())
188       if (Error E = addRecord(std::move(Func.second), 1))
189         return E;
190   return Error::success();
191 }
192 
193 bool InstrProfWriter::shouldEncodeData(const ProfilingData &PD) {
194   if (!Sparse)
195     return true;
196   for (const auto &Func : PD) {
197     const InstrProfRecord &IPR = Func.second;
198     if (std::any_of(IPR.Counts.begin(), IPR.Counts.end(),
199                     [](uint64_t Count) { return Count > 0; }))
200       return true;
201   }
202   return false;
203 }
204 
205 static void setSummary(IndexedInstrProf::Summary *TheSummary,
206                        ProfileSummary &PS) {
207   using namespace IndexedInstrProf;
208   std::vector<ProfileSummaryEntry> &Res = PS.getDetailedSummary();
209   TheSummary->NumSummaryFields = Summary::NumKinds;
210   TheSummary->NumCutoffEntries = Res.size();
211   TheSummary->set(Summary::MaxFunctionCount, PS.getMaxFunctionCount());
212   TheSummary->set(Summary::MaxBlockCount, PS.getMaxCount());
213   TheSummary->set(Summary::MaxInternalBlockCount, PS.getMaxInternalCount());
214   TheSummary->set(Summary::TotalBlockCount, PS.getTotalCount());
215   TheSummary->set(Summary::TotalNumBlocks, PS.getNumCounts());
216   TheSummary->set(Summary::TotalNumFunctions, PS.getNumFunctions());
217   for (unsigned I = 0; I < Res.size(); I++)
218     TheSummary->setEntry(I, Res[I]);
219 }
220 
221 void InstrProfWriter::writeImpl(ProfOStream &OS) {
222   OnDiskChainedHashTableGenerator<InstrProfRecordWriterTrait> Generator;
223 
224   using namespace IndexedInstrProf;
225   InstrProfSummaryBuilder ISB(ProfileSummaryBuilder::DefaultCutoffs);
226   InfoObj->SummaryBuilder = &ISB;
227 
228   // Populate the hash table generator.
229   for (const auto &I : FunctionData)
230     if (shouldEncodeData(I.getValue()))
231       Generator.insert(I.getKey(), &I.getValue());
232   // Write the header.
233   IndexedInstrProf::Header Header;
234   Header.Magic = IndexedInstrProf::Magic;
235   Header.Version = IndexedInstrProf::ProfVersion::CurrentVersion;
236   if (ProfileKind == PF_IRLevel)
237     Header.Version |= VARIANT_MASK_IR_PROF;
238   Header.Unused = 0;
239   Header.HashType = static_cast<uint64_t>(IndexedInstrProf::HashType);
240   Header.HashOffset = 0;
241   int N = sizeof(IndexedInstrProf::Header) / sizeof(uint64_t);
242 
243   // Only write out all the fields except 'HashOffset'. We need
244   // to remember the offset of that field to allow back patching
245   // later.
246   for (int I = 0; I < N - 1; I++)
247     OS.write(reinterpret_cast<uint64_t *>(&Header)[I]);
248 
249   // Save the location of Header.HashOffset field in \c OS.
250   uint64_t HashTableStartFieldOffset = OS.tell();
251   // Reserve the space for HashOffset field.
252   OS.write(0);
253 
254   // Reserve space to write profile summary data.
255   uint32_t NumEntries = ProfileSummaryBuilder::DefaultCutoffs.size();
256   uint32_t SummarySize = Summary::getSize(Summary::NumKinds, NumEntries);
257   // Remember the summary offset.
258   uint64_t SummaryOffset = OS.tell();
259   for (unsigned I = 0; I < SummarySize / sizeof(uint64_t); I++)
260     OS.write(0);
261 
262   // Write the hash table.
263   uint64_t HashTableStart = Generator.Emit(OS.OS, *InfoObj);
264 
265   // Allocate space for data to be serialized out.
266   std::unique_ptr<IndexedInstrProf::Summary> TheSummary =
267       IndexedInstrProf::allocSummary(SummarySize);
268   // Compute the Summary and copy the data to the data
269   // structure to be serialized out (to disk or buffer).
270   std::unique_ptr<ProfileSummary> PS = ISB.getSummary();
271   setSummary(TheSummary.get(), *PS);
272   InfoObj->SummaryBuilder = 0;
273 
274   // Now do the final patch:
275   PatchItem PatchItems[] = {
276       // Patch the Header.HashOffset field.
277       {HashTableStartFieldOffset, &HashTableStart, 1},
278       // Patch the summary data.
279       {SummaryOffset, reinterpret_cast<uint64_t *>(TheSummary.get()),
280        (int)(SummarySize / sizeof(uint64_t))}};
281   OS.patch(PatchItems, sizeof(PatchItems) / sizeof(*PatchItems));
282 }
283 
284 void InstrProfWriter::write(raw_fd_ostream &OS) {
285   // Write the hash table.
286   ProfOStream POS(OS);
287   writeImpl(POS);
288 }
289 
290 std::unique_ptr<MemoryBuffer> InstrProfWriter::writeBuffer() {
291   std::string Data;
292   llvm::raw_string_ostream OS(Data);
293   ProfOStream POS(OS);
294   // Write the hash table.
295   writeImpl(POS);
296   // Return this in an aligned memory buffer.
297   return MemoryBuffer::getMemBufferCopy(Data);
298 }
299 
300 static const char *ValueProfKindStr[] = {
301 #define VALUE_PROF_KIND(Enumerator, Value) #Enumerator,
302 #include "llvm/ProfileData/InstrProfData.inc"
303 };
304 
305 void InstrProfWriter::writeRecordInText(const InstrProfRecord &Func,
306                                         InstrProfSymtab &Symtab,
307                                         raw_fd_ostream &OS) {
308   OS << Func.Name << "\n";
309   OS << "# Func Hash:\n" << Func.Hash << "\n";
310   OS << "# Num Counters:\n" << Func.Counts.size() << "\n";
311   OS << "# Counter Values:\n";
312   for (uint64_t Count : Func.Counts)
313     OS << Count << "\n";
314 
315   uint32_t NumValueKinds = Func.getNumValueKinds();
316   if (!NumValueKinds) {
317     OS << "\n";
318     return;
319   }
320 
321   OS << "# Num Value Kinds:\n" << Func.getNumValueKinds() << "\n";
322   for (uint32_t VK = 0; VK < IPVK_Last + 1; VK++) {
323     uint32_t NS = Func.getNumValueSites(VK);
324     if (!NS)
325       continue;
326     OS << "# ValueKind = " << ValueProfKindStr[VK] << ":\n" << VK << "\n";
327     OS << "# NumValueSites:\n" << NS << "\n";
328     for (uint32_t S = 0; S < NS; S++) {
329       uint32_t ND = Func.getNumValueDataForSite(VK, S);
330       OS << ND << "\n";
331       std::unique_ptr<InstrProfValueData[]> VD = Func.getValueForSite(VK, S);
332       for (uint32_t I = 0; I < ND; I++) {
333         if (VK == IPVK_IndirectCallTarget)
334           OS << Symtab.getFuncName(VD[I].Value) << ":" << VD[I].Count << "\n";
335         else
336           OS << VD[I].Value << ":" << VD[I].Count << "\n";
337       }
338     }
339   }
340 
341   OS << "\n";
342 }
343 
344 void InstrProfWriter::writeText(raw_fd_ostream &OS) {
345   if (ProfileKind == PF_IRLevel)
346     OS << "# IR level Instrumentation Flag\n:ir\n";
347   InstrProfSymtab Symtab;
348   for (const auto &I : FunctionData)
349     if (shouldEncodeData(I.getValue()))
350       Symtab.addFuncName(I.getKey());
351   Symtab.finalizeSymtab();
352 
353   for (const auto &I : FunctionData)
354     if (shouldEncodeData(I.getValue()))
355       for (const auto &Func : I.getValue())
356         writeRecordInText(Func.second, Symtab, OS);
357 }
358