1 //===- InstrProfReader.h - Instrumented profiling readers -------*- C++ -*-===//
2 //
3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4 // See https://llvm.org/LICENSE.txt for license information.
5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6 //
7 //===----------------------------------------------------------------------===//
8 //
9 // This file contains support for reading profiling data for instrumentation
10 // based PGO and coverage.
11 //
12 //===----------------------------------------------------------------------===//
13 
14 #ifndef LLVM_PROFILEDATA_INSTRPROFREADER_H
15 #define LLVM_PROFILEDATA_INSTRPROFREADER_H
16 
17 #include "llvm/ADT/ArrayRef.h"
18 #include "llvm/ADT/StringRef.h"
19 #include "llvm/IR/ProfileSummary.h"
20 #include "llvm/Object/BuildID.h"
21 #include "llvm/ProfileData/InstrProf.h"
22 #include "llvm/ProfileData/InstrProfCorrelator.h"
23 #include "llvm/ProfileData/MemProf.h"
24 #include "llvm/Support/Endian.h"
25 #include "llvm/Support/Error.h"
26 #include "llvm/Support/LineIterator.h"
27 #include "llvm/Support/MathExtras.h"
28 #include "llvm/Support/MemoryBuffer.h"
29 #include "llvm/Support/OnDiskHashTable.h"
30 #include "llvm/Support/SwapByteOrder.h"
31 #include <algorithm>
32 #include <cassert>
33 #include <cstddef>
34 #include <cstdint>
35 #include <iterator>
36 #include <memory>
37 #include <utility>
38 #include <vector>
39 
40 namespace llvm {
41 
42 class InstrProfReader;
43 
44 namespace vfs {
45 class FileSystem;
46 } // namespace vfs
47 
48 /// A file format agnostic iterator over profiling data.
49 template <class record_type = NamedInstrProfRecord,
50           class reader_type = InstrProfReader>
51 class InstrProfIterator {
52 public:
53   using iterator_category = std::input_iterator_tag;
54   using value_type = record_type;
55   using difference_type = std::ptrdiff_t;
56   using pointer = value_type *;
57   using reference = value_type &;
58 
59 private:
60   reader_type *Reader = nullptr;
61   value_type Record;
62 
63   void increment() {
64     if (Error E = Reader->readNextRecord(Record)) {
65       // Handle errors in the reader.
66       InstrProfError::take(std::move(E));
67       *this = InstrProfIterator();
68     }
69   }
70 
71 public:
72   InstrProfIterator() = default;
73   InstrProfIterator(reader_type *Reader) : Reader(Reader) { increment(); }
74 
75   InstrProfIterator &operator++() {
76     increment();
77     return *this;
78   }
79   bool operator==(const InstrProfIterator &RHS) const {
80     return Reader == RHS.Reader;
81   }
82   bool operator!=(const InstrProfIterator &RHS) const {
83     return Reader != RHS.Reader;
84   }
85   value_type &operator*() { return Record; }
86   value_type *operator->() { return &Record; }
87 };
88 
89 /// Base class and interface for reading profiling data of any known instrprof
90 /// format. Provides an iterator over NamedInstrProfRecords.
91 class InstrProfReader {
92   instrprof_error LastError = instrprof_error::success;
93   std::string LastErrorMsg;
94 
95 public:
96   InstrProfReader() = default;
97   virtual ~InstrProfReader() = default;
98 
99   /// Read the header.  Required before reading first record.
100   virtual Error readHeader() = 0;
101 
102   /// Read a single record.
103   virtual Error readNextRecord(NamedInstrProfRecord &Record) = 0;
104 
105   /// Read a list of binary ids.
106   virtual Error readBinaryIds(std::vector<llvm::object::BuildID> &BinaryIds) {
107     return success();
108   }
109 
110   /// Print binary ids.
111   virtual Error printBinaryIds(raw_ostream &OS) { return success(); };
112 
113   /// Iterator over profile data.
114   InstrProfIterator<> begin() { return InstrProfIterator<>(this); }
115   InstrProfIterator<> end() { return InstrProfIterator<>(); }
116 
117   /// Return the profile version.
118   virtual uint64_t getVersion() const = 0;
119 
120   virtual bool isIRLevelProfile() const = 0;
121 
122   virtual bool hasCSIRLevelProfile() const = 0;
123 
124   virtual bool instrEntryBBEnabled() const = 0;
125 
126   /// Return true if the profile has single byte counters representing coverage.
127   virtual bool hasSingleByteCoverage() const = 0;
128 
129   /// Return true if the profile only instruments function entries.
130   virtual bool functionEntryOnly() const = 0;
131 
132   /// Return true if profile includes a memory profile.
133   virtual bool hasMemoryProfile() const = 0;
134 
135   /// Return true if this has a temporal profile.
136   virtual bool hasTemporalProfile() const = 0;
137 
138   /// Returns a BitsetEnum describing the attributes of the profile. To check
139   /// individual attributes prefer using the helpers above.
140   virtual InstrProfKind getProfileKind() const = 0;
141 
142   /// Return the PGO symtab. There are three different readers:
143   /// Raw, Text, and Indexed profile readers. The first two types
144   /// of readers are used only by llvm-profdata tool, while the indexed
145   /// profile reader is also used by llvm-cov tool and the compiler (
146   /// backend or frontend). Since creating PGO symtab can create
147   /// significant runtime and memory overhead (as it touches data
148   /// for the whole program), InstrProfSymtab for the indexed profile
149   /// reader should be created on demand and it is recommended to be
150   /// only used for dumping purpose with llvm-proftool, not with the
151   /// compiler.
152   virtual InstrProfSymtab &getSymtab() = 0;
153 
154   /// Compute the sum of counts and return in Sum.
155   void accumulateCounts(CountSumOrPercent &Sum, bool IsCS);
156 
157 protected:
158   std::unique_ptr<InstrProfSymtab> Symtab;
159   /// A list of temporal profile traces.
160   SmallVector<TemporalProfTraceTy> TemporalProfTraces;
161   /// The total number of temporal profile traces seen.
162   uint64_t TemporalProfTraceStreamSize = 0;
163 
164   /// Set the current error and return same.
165   Error error(instrprof_error Err, const std::string &ErrMsg = "") {
166     LastError = Err;
167     LastErrorMsg = ErrMsg;
168     if (Err == instrprof_error::success)
169       return Error::success();
170     return make_error<InstrProfError>(Err, ErrMsg);
171   }
172 
173   Error error(Error &&E) {
174     handleAllErrors(std::move(E), [&](const InstrProfError &IPE) {
175       LastError = IPE.get();
176       LastErrorMsg = IPE.getMessage();
177     });
178     return make_error<InstrProfError>(LastError, LastErrorMsg);
179   }
180 
181   /// Clear the current error and return a successful one.
182   Error success() { return error(instrprof_error::success); }
183 
184 public:
185   /// Return true if the reader has finished reading the profile data.
186   bool isEOF() { return LastError == instrprof_error::eof; }
187 
188   /// Return true if the reader encountered an error reading profiling data.
189   bool hasError() { return LastError != instrprof_error::success && !isEOF(); }
190 
191   /// Get the current error.
192   Error getError() {
193     if (hasError())
194       return make_error<InstrProfError>(LastError, LastErrorMsg);
195     return Error::success();
196   }
197 
198   /// Factory method to create an appropriately typed reader for the given
199   /// instrprof file.
200   static Expected<std::unique_ptr<InstrProfReader>>
201   create(const Twine &Path, vfs::FileSystem &FS,
202          const InstrProfCorrelator *Correlator = nullptr,
203          std::function<void(Error)> Warn = nullptr);
204 
205   static Expected<std::unique_ptr<InstrProfReader>>
206   create(std::unique_ptr<MemoryBuffer> Buffer,
207          const InstrProfCorrelator *Correlator = nullptr,
208          std::function<void(Error)> Warn = nullptr);
209 
210   /// \param Weight for raw profiles use this as the temporal profile trace
211   ///               weight
212   /// \returns a list of temporal profile traces.
213   virtual SmallVector<TemporalProfTraceTy> &
214   getTemporalProfTraces(std::optional<uint64_t> Weight = {}) {
215     // For non-raw profiles we ignore the input weight and instead use the
216     // weights already in the traces.
217     return TemporalProfTraces;
218   }
219   /// \returns the total number of temporal profile traces seen.
220   uint64_t getTemporalProfTraceStreamSize() {
221     return TemporalProfTraceStreamSize;
222   }
223 };
224 
225 /// Reader for the simple text based instrprof format.
226 ///
227 /// This format is a simple text format that's suitable for test data. Records
228 /// are separated by one or more blank lines, and record fields are separated by
229 /// new lines.
230 ///
231 /// Each record consists of a function name, a function hash, a number of
232 /// counters, and then each counter value, in that order.
233 class TextInstrProfReader : public InstrProfReader {
234 private:
235   /// The profile data file contents.
236   std::unique_ptr<MemoryBuffer> DataBuffer;
237   /// Iterator over the profile data.
238   line_iterator Line;
239   /// The attributes of the current profile.
240   InstrProfKind ProfileKind = InstrProfKind::Unknown;
241 
242   Error readValueProfileData(InstrProfRecord &Record);
243 
244   Error readTemporalProfTraceData();
245 
246 public:
247   TextInstrProfReader(std::unique_ptr<MemoryBuffer> DataBuffer_)
248       : DataBuffer(std::move(DataBuffer_)), Line(*DataBuffer, true, '#') {}
249   TextInstrProfReader(const TextInstrProfReader &) = delete;
250   TextInstrProfReader &operator=(const TextInstrProfReader &) = delete;
251 
252   /// Return true if the given buffer is in text instrprof format.
253   static bool hasFormat(const MemoryBuffer &Buffer);
254 
255   // Text format does not have version, so return 0.
256   uint64_t getVersion() const override { return 0; }
257 
258   bool isIRLevelProfile() const override {
259     return static_cast<bool>(ProfileKind & InstrProfKind::IRInstrumentation);
260   }
261 
262   bool hasCSIRLevelProfile() const override {
263     return static_cast<bool>(ProfileKind & InstrProfKind::ContextSensitive);
264   }
265 
266   bool instrEntryBBEnabled() const override {
267     return static_cast<bool>(ProfileKind &
268                              InstrProfKind::FunctionEntryInstrumentation);
269   }
270 
271   bool hasSingleByteCoverage() const override {
272     return static_cast<bool>(ProfileKind & InstrProfKind::SingleByteCoverage);
273   }
274 
275   bool functionEntryOnly() const override {
276     return static_cast<bool>(ProfileKind & InstrProfKind::FunctionEntryOnly);
277   }
278 
279   bool hasMemoryProfile() const override {
280     // TODO: Add support for text format memory profiles.
281     return false;
282   }
283 
284   bool hasTemporalProfile() const override {
285     return static_cast<bool>(ProfileKind & InstrProfKind::TemporalProfile);
286   }
287 
288   InstrProfKind getProfileKind() const override { return ProfileKind; }
289 
290   /// Read the header.
291   Error readHeader() override;
292 
293   /// Read a single record.
294   Error readNextRecord(NamedInstrProfRecord &Record) override;
295 
296   InstrProfSymtab &getSymtab() override {
297     assert(Symtab);
298     return *Symtab;
299   }
300 };
301 
302 /// Reader for the raw instrprof binary format from runtime.
303 ///
304 /// This format is a raw memory dump of the instrumentation-based profiling data
305 /// from the runtime.  It has no index.
306 ///
307 /// Templated on the unsigned type whose size matches pointers on the platform
308 /// that wrote the profile.
309 template <class IntPtrT>
310 class RawInstrProfReader : public InstrProfReader {
311 private:
312   /// The profile data file contents.
313   std::unique_ptr<MemoryBuffer> DataBuffer;
314   /// If available, this hold the ProfileData array used to correlate raw
315   /// instrumentation data to their functions.
316   const InstrProfCorrelatorImpl<IntPtrT> *Correlator;
317   /// A list of timestamps paired with a function name reference.
318   std::vector<std::pair<uint64_t, uint64_t>> TemporalProfTimestamps;
319   bool ShouldSwapBytes;
320   // The value of the version field of the raw profile data header. The lower 32
321   // bits specifies the format version and the most significant 32 bits specify
322   // the variant types of the profile.
323   uint64_t Version;
324   uint64_t CountersDelta;
325   uint64_t BitmapDelta;
326   uint64_t NamesDelta;
327   const RawInstrProf::ProfileData<IntPtrT> *Data;
328   const RawInstrProf::ProfileData<IntPtrT> *DataEnd;
329   const char *CountersStart;
330   const char *CountersEnd;
331   const char *BitmapStart;
332   const char *BitmapEnd;
333   const char *NamesStart;
334   const char *NamesEnd;
335   // After value profile is all read, this pointer points to
336   // the header of next profile data (if exists)
337   const uint8_t *ValueDataStart;
338   uint32_t ValueKindLast;
339   uint32_t CurValueDataSize;
340   std::vector<llvm::object::BuildID> BinaryIds;
341 
342   std::function<void(Error)> Warn;
343 
344   /// Maxium counter value 2^56.
345   static const uint64_t MaxCounterValue = (1ULL << 56);
346 
347 public:
348   RawInstrProfReader(std::unique_ptr<MemoryBuffer> DataBuffer,
349                      const InstrProfCorrelator *Correlator,
350                      std::function<void(Error)> Warn)
351       : DataBuffer(std::move(DataBuffer)),
352         Correlator(dyn_cast_or_null<const InstrProfCorrelatorImpl<IntPtrT>>(
353             Correlator)),
354         Warn(Warn) {}
355   RawInstrProfReader(const RawInstrProfReader &) = delete;
356   RawInstrProfReader &operator=(const RawInstrProfReader &) = delete;
357 
358   static bool hasFormat(const MemoryBuffer &DataBuffer);
359   Error readHeader() override;
360   Error readNextRecord(NamedInstrProfRecord &Record) override;
361   Error readBinaryIds(std::vector<llvm::object::BuildID> &BinaryIds) override;
362   Error printBinaryIds(raw_ostream &OS) override;
363 
364   uint64_t getVersion() const override { return Version; }
365 
366   bool isIRLevelProfile() const override {
367     return (Version & VARIANT_MASK_IR_PROF) != 0;
368   }
369 
370   bool hasCSIRLevelProfile() const override {
371     return (Version & VARIANT_MASK_CSIR_PROF) != 0;
372   }
373 
374   bool instrEntryBBEnabled() const override {
375     return (Version & VARIANT_MASK_INSTR_ENTRY) != 0;
376   }
377 
378   bool hasSingleByteCoverage() const override {
379     return (Version & VARIANT_MASK_BYTE_COVERAGE) != 0;
380   }
381 
382   bool functionEntryOnly() const override {
383     return (Version & VARIANT_MASK_FUNCTION_ENTRY_ONLY) != 0;
384   }
385 
386   bool hasMemoryProfile() const override {
387     // Memory profiles have a separate raw format, so this should never be set.
388     assert(!(Version & VARIANT_MASK_MEMPROF));
389     return false;
390   }
391 
392   bool hasTemporalProfile() const override {
393     return (Version & VARIANT_MASK_TEMPORAL_PROF) != 0;
394   }
395 
396   /// Returns a BitsetEnum describing the attributes of the raw instr profile.
397   InstrProfKind getProfileKind() const override;
398 
399   InstrProfSymtab &getSymtab() override {
400     assert(Symtab.get());
401     return *Symtab.get();
402   }
403 
404   SmallVector<TemporalProfTraceTy> &
405   getTemporalProfTraces(std::optional<uint64_t> Weight = {}) override;
406 
407 private:
408   Error createSymtab(InstrProfSymtab &Symtab);
409   Error readNextHeader(const char *CurrentPos);
410   Error readHeader(const RawInstrProf::Header &Header);
411 
412   template <class IntT> IntT swap(IntT Int) const {
413     return ShouldSwapBytes ? llvm::byteswap(Int) : Int;
414   }
415 
416   llvm::endianness getDataEndianness() const {
417     if (!ShouldSwapBytes)
418       return llvm::endianness::native;
419     if (llvm::endianness::native == llvm::endianness::little)
420       return llvm::endianness::big;
421     else
422       return llvm::endianness::little;
423   }
424 
425   inline uint8_t getNumPaddingBytes(uint64_t SizeInBytes) {
426     return 7 & (sizeof(uint64_t) - SizeInBytes % sizeof(uint64_t));
427   }
428 
429   Error readName(NamedInstrProfRecord &Record);
430   Error readFuncHash(NamedInstrProfRecord &Record);
431   Error readRawCounts(InstrProfRecord &Record);
432   Error readRawBitmapBytes(InstrProfRecord &Record);
433   Error readValueProfilingData(InstrProfRecord &Record);
434   bool atEnd() const { return Data == DataEnd; }
435 
436   void advanceData() {
437     // `CountersDelta` is a constant zero when using debug info correlation.
438     if (!Correlator) {
439       // The initial CountersDelta is the in-memory address difference between
440       // the data and counts sections:
441       // start(__llvm_prf_cnts) - start(__llvm_prf_data)
442       // As we advance to the next record, we maintain the correct CountersDelta
443       // with respect to the next record.
444       CountersDelta -= sizeof(*Data);
445       BitmapDelta -= sizeof(*Data);
446     }
447     Data++;
448     ValueDataStart += CurValueDataSize;
449   }
450 
451   const char *getNextHeaderPos() const {
452       assert(atEnd());
453       return (const char *)ValueDataStart;
454   }
455 
456   StringRef getName(uint64_t NameRef) const {
457     return Symtab->getFuncOrVarName(swap(NameRef));
458   }
459 
460   int getCounterTypeSize() const {
461     return hasSingleByteCoverage() ? sizeof(uint8_t) : sizeof(uint64_t);
462   }
463 };
464 
465 using RawInstrProfReader32 = RawInstrProfReader<uint32_t>;
466 using RawInstrProfReader64 = RawInstrProfReader<uint64_t>;
467 
468 namespace IndexedInstrProf {
469 
470 enum class HashT : uint32_t;
471 
472 } // end namespace IndexedInstrProf
473 
474 /// Trait for lookups into the on-disk hash table for the binary instrprof
475 /// format.
476 class InstrProfLookupTrait {
477   std::vector<NamedInstrProfRecord> DataBuffer;
478   IndexedInstrProf::HashT HashType;
479   unsigned FormatVersion;
480   // Endianness of the input value profile data.
481   // It should be LE by default, but can be changed
482   // for testing purpose.
483   llvm::endianness ValueProfDataEndianness = llvm::endianness::little;
484 
485 public:
486   InstrProfLookupTrait(IndexedInstrProf::HashT HashType, unsigned FormatVersion)
487       : HashType(HashType), FormatVersion(FormatVersion) {}
488 
489   using data_type = ArrayRef<NamedInstrProfRecord>;
490 
491   using internal_key_type = StringRef;
492   using external_key_type = StringRef;
493   using hash_value_type = uint64_t;
494   using offset_type = uint64_t;
495 
496   static bool EqualKey(StringRef A, StringRef B) { return A == B; }
497   static StringRef GetInternalKey(StringRef K) { return K; }
498   static StringRef GetExternalKey(StringRef K) { return K; }
499 
500   hash_value_type ComputeHash(StringRef K);
501 
502   static std::pair<offset_type, offset_type>
503   ReadKeyDataLength(const unsigned char *&D) {
504     using namespace support;
505 
506     offset_type KeyLen =
507         endian::readNext<offset_type, llvm::endianness::little, unaligned>(D);
508     offset_type DataLen =
509         endian::readNext<offset_type, llvm::endianness::little, unaligned>(D);
510     return std::make_pair(KeyLen, DataLen);
511   }
512 
513   StringRef ReadKey(const unsigned char *D, offset_type N) {
514     return StringRef((const char *)D, N);
515   }
516 
517   bool readValueProfilingData(const unsigned char *&D,
518                               const unsigned char *const End);
519   data_type ReadData(StringRef K, const unsigned char *D, offset_type N);
520 
521   // Used for testing purpose only.
522   void setValueProfDataEndianness(llvm::endianness Endianness) {
523     ValueProfDataEndianness = Endianness;
524   }
525 };
526 
527 struct InstrProfReaderIndexBase {
528   virtual ~InstrProfReaderIndexBase() = default;
529 
530   // Read all the profile records with the same key pointed to the current
531   // iterator.
532   virtual Error getRecords(ArrayRef<NamedInstrProfRecord> &Data) = 0;
533 
534   // Read all the profile records with the key equal to FuncName
535   virtual Error getRecords(StringRef FuncName,
536                                      ArrayRef<NamedInstrProfRecord> &Data) = 0;
537   virtual void advanceToNextKey() = 0;
538   virtual bool atEnd() const = 0;
539   virtual void setValueProfDataEndianness(llvm::endianness Endianness) = 0;
540   virtual uint64_t getVersion() const = 0;
541   virtual bool isIRLevelProfile() const = 0;
542   virtual bool hasCSIRLevelProfile() const = 0;
543   virtual bool instrEntryBBEnabled() const = 0;
544   virtual bool hasSingleByteCoverage() const = 0;
545   virtual bool functionEntryOnly() const = 0;
546   virtual bool hasMemoryProfile() const = 0;
547   virtual bool hasTemporalProfile() const = 0;
548   virtual InstrProfKind getProfileKind() const = 0;
549   virtual Error populateSymtab(InstrProfSymtab &) = 0;
550 };
551 
552 using OnDiskHashTableImplV3 =
553     OnDiskIterableChainedHashTable<InstrProfLookupTrait>;
554 
555 using MemProfRecordHashTable =
556     OnDiskIterableChainedHashTable<memprof::RecordLookupTrait>;
557 using MemProfFrameHashTable =
558     OnDiskIterableChainedHashTable<memprof::FrameLookupTrait>;
559 
560 template <typename HashTableImpl>
561 class InstrProfReaderItaniumRemapper;
562 
563 template <typename HashTableImpl>
564 class InstrProfReaderIndex : public InstrProfReaderIndexBase {
565 private:
566   std::unique_ptr<HashTableImpl> HashTable;
567   typename HashTableImpl::data_iterator RecordIterator;
568   uint64_t FormatVersion;
569 
570   friend class InstrProfReaderItaniumRemapper<HashTableImpl>;
571 
572 public:
573   InstrProfReaderIndex(const unsigned char *Buckets,
574                        const unsigned char *const Payload,
575                        const unsigned char *const Base,
576                        IndexedInstrProf::HashT HashType, uint64_t Version);
577   ~InstrProfReaderIndex() override = default;
578 
579   Error getRecords(ArrayRef<NamedInstrProfRecord> &Data) override;
580   Error getRecords(StringRef FuncName,
581                    ArrayRef<NamedInstrProfRecord> &Data) override;
582   void advanceToNextKey() override { RecordIterator++; }
583 
584   bool atEnd() const override {
585     return RecordIterator == HashTable->data_end();
586   }
587 
588   void setValueProfDataEndianness(llvm::endianness Endianness) override {
589     HashTable->getInfoObj().setValueProfDataEndianness(Endianness);
590   }
591 
592   uint64_t getVersion() const override { return GET_VERSION(FormatVersion); }
593 
594   bool isIRLevelProfile() const override {
595     return (FormatVersion & VARIANT_MASK_IR_PROF) != 0;
596   }
597 
598   bool hasCSIRLevelProfile() const override {
599     return (FormatVersion & VARIANT_MASK_CSIR_PROF) != 0;
600   }
601 
602   bool instrEntryBBEnabled() const override {
603     return (FormatVersion & VARIANT_MASK_INSTR_ENTRY) != 0;
604   }
605 
606   bool hasSingleByteCoverage() const override {
607     return (FormatVersion & VARIANT_MASK_BYTE_COVERAGE) != 0;
608   }
609 
610   bool functionEntryOnly() const override {
611     return (FormatVersion & VARIANT_MASK_FUNCTION_ENTRY_ONLY) != 0;
612   }
613 
614   bool hasMemoryProfile() const override {
615     return (FormatVersion & VARIANT_MASK_MEMPROF) != 0;
616   }
617 
618   bool hasTemporalProfile() const override {
619     return (FormatVersion & VARIANT_MASK_TEMPORAL_PROF) != 0;
620   }
621 
622   InstrProfKind getProfileKind() const override;
623 
624   Error populateSymtab(InstrProfSymtab &Symtab) override {
625     return Symtab.create(HashTable->keys());
626   }
627 };
628 
629 /// Name matcher supporting fuzzy matching of symbol names to names in profiles.
630 class InstrProfReaderRemapper {
631 public:
632   virtual ~InstrProfReaderRemapper() = default;
633   virtual Error populateRemappings() { return Error::success(); }
634   virtual Error getRecords(StringRef FuncName,
635                            ArrayRef<NamedInstrProfRecord> &Data) = 0;
636 };
637 
638 /// Reader for the indexed binary instrprof format.
639 class IndexedInstrProfReader : public InstrProfReader {
640 private:
641   /// The profile data file contents.
642   std::unique_ptr<MemoryBuffer> DataBuffer;
643   /// The profile remapping file contents.
644   std::unique_ptr<MemoryBuffer> RemappingBuffer;
645   /// The index into the profile data.
646   std::unique_ptr<InstrProfReaderIndexBase> Index;
647   /// The profile remapping file contents.
648   std::unique_ptr<InstrProfReaderRemapper> Remapper;
649   /// Profile summary data.
650   std::unique_ptr<ProfileSummary> Summary;
651   /// Context sensitive profile summary data.
652   std::unique_ptr<ProfileSummary> CS_Summary;
653   /// MemProf profile schema (if available).
654   memprof::MemProfSchema Schema;
655   /// MemProf record profile data on-disk indexed via llvm::md5(FunctionName).
656   std::unique_ptr<MemProfRecordHashTable> MemProfRecordTable;
657   /// MemProf frame profile data on-disk indexed via frame id.
658   std::unique_ptr<MemProfFrameHashTable> MemProfFrameTable;
659   /// Total size of binary ids.
660   uint64_t BinaryIdsSize{0};
661   /// Start address of binary id length and data pairs.
662   const uint8_t *BinaryIdsStart = nullptr;
663 
664   // Index to the current record in the record array.
665   unsigned RecordIndex;
666 
667   // Read the profile summary. Return a pointer pointing to one byte past the
668   // end of the summary data if it exists or the input \c Cur.
669   // \c UseCS indicates whether to use the context-sensitive profile summary.
670   const unsigned char *readSummary(IndexedInstrProf::ProfVersion Version,
671                                    const unsigned char *Cur, bool UseCS);
672 
673 public:
674   IndexedInstrProfReader(
675       std::unique_ptr<MemoryBuffer> DataBuffer,
676       std::unique_ptr<MemoryBuffer> RemappingBuffer = nullptr)
677       : DataBuffer(std::move(DataBuffer)),
678         RemappingBuffer(std::move(RemappingBuffer)), RecordIndex(0) {}
679   IndexedInstrProfReader(const IndexedInstrProfReader &) = delete;
680   IndexedInstrProfReader &operator=(const IndexedInstrProfReader &) = delete;
681 
682   /// Return the profile version.
683   uint64_t getVersion() const override { return Index->getVersion(); }
684   bool isIRLevelProfile() const override { return Index->isIRLevelProfile(); }
685   bool hasCSIRLevelProfile() const override {
686     return Index->hasCSIRLevelProfile();
687   }
688 
689   bool instrEntryBBEnabled() const override {
690     return Index->instrEntryBBEnabled();
691   }
692 
693   bool hasSingleByteCoverage() const override {
694     return Index->hasSingleByteCoverage();
695   }
696 
697   bool functionEntryOnly() const override { return Index->functionEntryOnly(); }
698 
699   bool hasMemoryProfile() const override { return Index->hasMemoryProfile(); }
700 
701   bool hasTemporalProfile() const override {
702     return Index->hasTemporalProfile();
703   }
704 
705   /// Returns a BitsetEnum describing the attributes of the indexed instr
706   /// profile.
707   InstrProfKind getProfileKind() const override {
708     return Index->getProfileKind();
709   }
710 
711   /// Return true if the given buffer is in an indexed instrprof format.
712   static bool hasFormat(const MemoryBuffer &DataBuffer);
713 
714   /// Read the file header.
715   Error readHeader() override;
716   /// Read a single record.
717   Error readNextRecord(NamedInstrProfRecord &Record) override;
718 
719   /// Return the NamedInstrProfRecord associated with FuncName and FuncHash.
720   /// When return a hash_mismatch error and MismatchedFuncSum is not nullptr,
721   /// the sum of all counters in the mismatched function will be set to
722   /// MismatchedFuncSum. If there are multiple instances of mismatched
723   /// functions, MismatchedFuncSum returns the maximum. If \c FuncName is not
724   /// found, try to lookup \c DeprecatedFuncName to handle profiles built by
725   /// older compilers.
726   Expected<InstrProfRecord>
727   getInstrProfRecord(StringRef FuncName, uint64_t FuncHash,
728                      StringRef DeprecatedFuncName = "",
729                      uint64_t *MismatchedFuncSum = nullptr);
730 
731   /// Return the memprof record for the function identified by
732   /// llvm::md5(Name).
733   Expected<memprof::MemProfRecord> getMemProfRecord(uint64_t FuncNameHash);
734 
735   /// Fill Counts with the profile data for the given function name.
736   Error getFunctionCounts(StringRef FuncName, uint64_t FuncHash,
737                           std::vector<uint64_t> &Counts);
738 
739   /// Fill Bitmap Bytes with the profile data for the given function name.
740   Error getFunctionBitmapBytes(StringRef FuncName, uint64_t FuncHash,
741                                std::vector<uint8_t> &BitmapBytes);
742 
743   /// Return the maximum of all known function counts.
744   /// \c UseCS indicates whether to use the context-sensitive count.
745   uint64_t getMaximumFunctionCount(bool UseCS) {
746     if (UseCS) {
747       assert(CS_Summary && "No context sensitive profile summary");
748       return CS_Summary->getMaxFunctionCount();
749     } else {
750       assert(Summary && "No profile summary");
751       return Summary->getMaxFunctionCount();
752     }
753   }
754 
755   /// Factory method to create an indexed reader.
756   static Expected<std::unique_ptr<IndexedInstrProfReader>>
757   create(const Twine &Path, vfs::FileSystem &FS,
758          const Twine &RemappingPath = "");
759 
760   static Expected<std::unique_ptr<IndexedInstrProfReader>>
761   create(std::unique_ptr<MemoryBuffer> Buffer,
762          std::unique_ptr<MemoryBuffer> RemappingBuffer = nullptr);
763 
764   // Used for testing purpose only.
765   void setValueProfDataEndianness(llvm::endianness Endianness) {
766     Index->setValueProfDataEndianness(Endianness);
767   }
768 
769   // See description in the base class. This interface is designed
770   // to be used by llvm-profdata (for dumping). Avoid using this when
771   // the client is the compiler.
772   InstrProfSymtab &getSymtab() override;
773 
774   /// Return the profile summary.
775   /// \c UseCS indicates whether to use the context-sensitive summary.
776   ProfileSummary &getSummary(bool UseCS) {
777     if (UseCS) {
778       assert(CS_Summary && "No context sensitive summary");
779       return *CS_Summary;
780     } else {
781       assert(Summary && "No profile summary");
782       return *Summary;
783     }
784   }
785 
786   Error readBinaryIds(std::vector<llvm::object::BuildID> &BinaryIds) override;
787   Error printBinaryIds(raw_ostream &OS) override;
788 };
789 
790 } // end namespace llvm
791 
792 #endif // LLVM_PROFILEDATA_INSTRPROFREADER_H
793