Yi Kong | 878f994 | 2023-12-13 12:55:04 +0900 | [diff] [blame^] | 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 we must provide debug info to create PGO profiles. |
| 127 | virtual bool useDebugInfoCorrelate() const { return false; } |
| 128 | |
| 129 | /// Return true if the profile has single byte counters representing coverage. |
| 130 | virtual bool hasSingleByteCoverage() const = 0; |
| 131 | |
| 132 | /// Return true if the profile only instruments function entries. |
| 133 | virtual bool functionEntryOnly() const = 0; |
| 134 | |
| 135 | /// Return true if profile includes a memory profile. |
| 136 | virtual bool hasMemoryProfile() const = 0; |
| 137 | |
| 138 | /// Return true if this has a temporal profile. |
| 139 | virtual bool hasTemporalProfile() const = 0; |
| 140 | |
| 141 | /// Returns a BitsetEnum describing the attributes of the profile. To check |
| 142 | /// individual attributes prefer using the helpers above. |
| 143 | virtual InstrProfKind getProfileKind() const = 0; |
| 144 | |
| 145 | /// Return the PGO symtab. There are three different readers: |
| 146 | /// Raw, Text, and Indexed profile readers. The first two types |
| 147 | /// of readers are used only by llvm-profdata tool, while the indexed |
| 148 | /// profile reader is also used by llvm-cov tool and the compiler ( |
| 149 | /// backend or frontend). Since creating PGO symtab can create |
| 150 | /// significant runtime and memory overhead (as it touches data |
| 151 | /// for the whole program), InstrProfSymtab for the indexed profile |
| 152 | /// reader should be created on demand and it is recommended to be |
| 153 | /// only used for dumping purpose with llvm-proftool, not with the |
| 154 | /// compiler. |
| 155 | virtual InstrProfSymtab &getSymtab() = 0; |
| 156 | |
| 157 | /// Compute the sum of counts and return in Sum. |
| 158 | void accumulateCounts(CountSumOrPercent &Sum, bool IsCS); |
| 159 | |
| 160 | protected: |
| 161 | std::unique_ptr<InstrProfSymtab> Symtab; |
| 162 | /// A list of temporal profile traces. |
| 163 | SmallVector<TemporalProfTraceTy> TemporalProfTraces; |
| 164 | /// The total number of temporal profile traces seen. |
| 165 | uint64_t TemporalProfTraceStreamSize = 0; |
| 166 | |
| 167 | /// Set the current error and return same. |
| 168 | Error error(instrprof_error Err, const std::string &ErrMsg = "") { |
| 169 | LastError = Err; |
| 170 | LastErrorMsg = ErrMsg; |
| 171 | if (Err == instrprof_error::success) |
| 172 | return Error::success(); |
| 173 | return make_error<InstrProfError>(Err, ErrMsg); |
| 174 | } |
| 175 | |
| 176 | Error error(Error &&E) { |
| 177 | handleAllErrors(std::move(E), [&](const InstrProfError &IPE) { |
| 178 | LastError = IPE.get(); |
| 179 | LastErrorMsg = IPE.getMessage(); |
| 180 | }); |
| 181 | return make_error<InstrProfError>(LastError, LastErrorMsg); |
| 182 | } |
| 183 | |
| 184 | /// Clear the current error and return a successful one. |
| 185 | Error success() { return error(instrprof_error::success); } |
| 186 | |
| 187 | public: |
| 188 | /// Return true if the reader has finished reading the profile data. |
| 189 | bool isEOF() { return LastError == instrprof_error::eof; } |
| 190 | |
| 191 | /// Return true if the reader encountered an error reading profiling data. |
| 192 | bool hasError() { return LastError != instrprof_error::success && !isEOF(); } |
| 193 | |
| 194 | /// Get the current error. |
| 195 | Error getError() { |
| 196 | if (hasError()) |
| 197 | return make_error<InstrProfError>(LastError, LastErrorMsg); |
| 198 | return Error::success(); |
| 199 | } |
| 200 | |
| 201 | /// Factory method to create an appropriately typed reader for the given |
| 202 | /// instrprof file. |
| 203 | static Expected<std::unique_ptr<InstrProfReader>> |
| 204 | create(const Twine &Path, vfs::FileSystem &FS, |
| 205 | const InstrProfCorrelator *Correlator = nullptr); |
| 206 | |
| 207 | static Expected<std::unique_ptr<InstrProfReader>> |
| 208 | create(std::unique_ptr<MemoryBuffer> Buffer, |
| 209 | const InstrProfCorrelator *Correlator = nullptr); |
| 210 | |
| 211 | /// \param Weight for raw profiles use this as the temporal profile trace |
| 212 | /// weight |
| 213 | /// \returns a list of temporal profile traces. |
| 214 | virtual SmallVector<TemporalProfTraceTy> & |
| 215 | getTemporalProfTraces(std::optional<uint64_t> Weight = {}) { |
| 216 | // For non-raw profiles we ignore the input weight and instead use the |
| 217 | // weights already in the traces. |
| 218 | return TemporalProfTraces; |
| 219 | } |
| 220 | /// \returns the total number of temporal profile traces seen. |
| 221 | uint64_t getTemporalProfTraceStreamSize() { |
| 222 | return TemporalProfTraceStreamSize; |
| 223 | } |
| 224 | }; |
| 225 | |
| 226 | /// Reader for the simple text based instrprof format. |
| 227 | /// |
| 228 | /// This format is a simple text format that's suitable for test data. Records |
| 229 | /// are separated by one or more blank lines, and record fields are separated by |
| 230 | /// new lines. |
| 231 | /// |
| 232 | /// Each record consists of a function name, a function hash, a number of |
| 233 | /// counters, and then each counter value, in that order. |
| 234 | class TextInstrProfReader : public InstrProfReader { |
| 235 | private: |
| 236 | /// The profile data file contents. |
| 237 | std::unique_ptr<MemoryBuffer> DataBuffer; |
| 238 | /// Iterator over the profile data. |
| 239 | line_iterator Line; |
| 240 | /// The attributes of the current profile. |
| 241 | InstrProfKind ProfileKind = InstrProfKind::Unknown; |
| 242 | |
| 243 | Error readValueProfileData(InstrProfRecord &Record); |
| 244 | |
| 245 | Error readTemporalProfTraceData(); |
| 246 | |
| 247 | public: |
| 248 | TextInstrProfReader(std::unique_ptr<MemoryBuffer> DataBuffer_) |
| 249 | : DataBuffer(std::move(DataBuffer_)), Line(*DataBuffer, true, '#') {} |
| 250 | TextInstrProfReader(const TextInstrProfReader &) = delete; |
| 251 | TextInstrProfReader &operator=(const TextInstrProfReader &) = delete; |
| 252 | |
| 253 | /// Return true if the given buffer is in text instrprof format. |
| 254 | static bool hasFormat(const MemoryBuffer &Buffer); |
| 255 | |
| 256 | // Text format does not have version, so return 0. |
| 257 | uint64_t getVersion() const override { return 0; } |
| 258 | |
| 259 | bool isIRLevelProfile() const override { |
| 260 | return static_cast<bool>(ProfileKind & InstrProfKind::IRInstrumentation); |
| 261 | } |
| 262 | |
| 263 | bool hasCSIRLevelProfile() const override { |
| 264 | return static_cast<bool>(ProfileKind & InstrProfKind::ContextSensitive); |
| 265 | } |
| 266 | |
| 267 | bool instrEntryBBEnabled() const override { |
| 268 | return static_cast<bool>(ProfileKind & |
| 269 | InstrProfKind::FunctionEntryInstrumentation); |
| 270 | } |
| 271 | |
| 272 | bool hasSingleByteCoverage() const override { |
| 273 | return static_cast<bool>(ProfileKind & InstrProfKind::SingleByteCoverage); |
| 274 | } |
| 275 | |
| 276 | bool functionEntryOnly() const override { |
| 277 | return static_cast<bool>(ProfileKind & InstrProfKind::FunctionEntryOnly); |
| 278 | } |
| 279 | |
| 280 | bool hasMemoryProfile() const override { |
| 281 | // TODO: Add support for text format memory profiles. |
| 282 | return false; |
| 283 | } |
| 284 | |
| 285 | bool hasTemporalProfile() const override { |
| 286 | return static_cast<bool>(ProfileKind & InstrProfKind::TemporalProfile); |
| 287 | } |
| 288 | |
| 289 | InstrProfKind getProfileKind() const override { return ProfileKind; } |
| 290 | |
| 291 | /// Read the header. |
| 292 | Error readHeader() override; |
| 293 | |
| 294 | /// Read a single record. |
| 295 | Error readNextRecord(NamedInstrProfRecord &Record) override; |
| 296 | |
| 297 | InstrProfSymtab &getSymtab() override { |
| 298 | assert(Symtab); |
| 299 | return *Symtab; |
| 300 | } |
| 301 | }; |
| 302 | |
| 303 | /// Reader for the raw instrprof binary format from runtime. |
| 304 | /// |
| 305 | /// This format is a raw memory dump of the instrumentation-based profiling data |
| 306 | /// from the runtime. It has no index. |
| 307 | /// |
| 308 | /// Templated on the unsigned type whose size matches pointers on the platform |
| 309 | /// that wrote the profile. |
| 310 | template <class IntPtrT> |
| 311 | class RawInstrProfReader : public InstrProfReader { |
| 312 | private: |
| 313 | /// The profile data file contents. |
| 314 | std::unique_ptr<MemoryBuffer> DataBuffer; |
| 315 | /// If available, this hold the ProfileData array used to correlate raw |
| 316 | /// instrumentation data to their functions. |
| 317 | const InstrProfCorrelatorImpl<IntPtrT> *Correlator; |
| 318 | /// A list of timestamps paired with a function name reference. |
| 319 | std::vector<std::pair<uint64_t, uint64_t>> TemporalProfTimestamps; |
| 320 | bool ShouldSwapBytes; |
| 321 | // The value of the version field of the raw profile data header. The lower 56 |
| 322 | // bits specifies the format version and the most significant 8 bits specify |
| 323 | // the variant types of the profile. |
| 324 | uint64_t Version; |
| 325 | uint64_t CountersDelta; |
| 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 *NamesStart; |
| 332 | const char *NamesEnd; |
| 333 | // After value profile is all read, this pointer points to |
| 334 | // the header of next profile data (if exists) |
| 335 | const uint8_t *ValueDataStart; |
| 336 | uint32_t ValueKindLast; |
| 337 | uint32_t CurValueDataSize; |
| 338 | |
| 339 | /// Total size of binary ids. |
| 340 | uint64_t BinaryIdsSize{0}; |
| 341 | /// Start address of binary id length and data pairs. |
| 342 | const uint8_t *BinaryIdsStart; |
| 343 | |
| 344 | public: |
| 345 | RawInstrProfReader(std::unique_ptr<MemoryBuffer> DataBuffer, |
| 346 | const InstrProfCorrelator *Correlator) |
| 347 | : DataBuffer(std::move(DataBuffer)), |
| 348 | Correlator(dyn_cast_or_null<const InstrProfCorrelatorImpl<IntPtrT>>( |
| 349 | Correlator)) {} |
| 350 | RawInstrProfReader(const RawInstrProfReader &) = delete; |
| 351 | RawInstrProfReader &operator=(const RawInstrProfReader &) = delete; |
| 352 | |
| 353 | static bool hasFormat(const MemoryBuffer &DataBuffer); |
| 354 | Error readHeader() override; |
| 355 | Error readNextRecord(NamedInstrProfRecord &Record) override; |
| 356 | Error readBinaryIds(std::vector<llvm::object::BuildID> &BinaryIds) override; |
| 357 | Error printBinaryIds(raw_ostream &OS) override; |
| 358 | |
| 359 | uint64_t getVersion() const override { return Version; } |
| 360 | |
| 361 | bool isIRLevelProfile() const override { |
| 362 | return (Version & VARIANT_MASK_IR_PROF) != 0; |
| 363 | } |
| 364 | |
| 365 | bool hasCSIRLevelProfile() const override { |
| 366 | return (Version & VARIANT_MASK_CSIR_PROF) != 0; |
| 367 | } |
| 368 | |
| 369 | bool instrEntryBBEnabled() const override { |
| 370 | return (Version & VARIANT_MASK_INSTR_ENTRY) != 0; |
| 371 | } |
| 372 | |
| 373 | bool useDebugInfoCorrelate() const override { |
| 374 | return (Version & VARIANT_MASK_DBG_CORRELATE) != 0; |
| 375 | } |
| 376 | |
| 377 | bool hasSingleByteCoverage() const override { |
| 378 | return (Version & VARIANT_MASK_BYTE_COVERAGE) != 0; |
| 379 | } |
| 380 | |
| 381 | bool functionEntryOnly() const override { |
| 382 | return (Version & VARIANT_MASK_FUNCTION_ENTRY_ONLY) != 0; |
| 383 | } |
| 384 | |
| 385 | bool hasMemoryProfile() const override { |
| 386 | // Memory profiles have a separate raw format, so this should never be set. |
| 387 | assert(!(Version & VARIANT_MASK_MEMPROF)); |
| 388 | return false; |
| 389 | } |
| 390 | |
| 391 | bool hasTemporalProfile() const override { |
| 392 | return (Version & VARIANT_MASK_TEMPORAL_PROF) != 0; |
| 393 | } |
| 394 | |
| 395 | /// Returns a BitsetEnum describing the attributes of the raw instr profile. |
| 396 | InstrProfKind getProfileKind() const override; |
| 397 | |
| 398 | InstrProfSymtab &getSymtab() override { |
| 399 | assert(Symtab.get()); |
| 400 | return *Symtab.get(); |
| 401 | } |
| 402 | |
| 403 | SmallVector<TemporalProfTraceTy> & |
| 404 | getTemporalProfTraces(std::optional<uint64_t> Weight = {}) override; |
| 405 | |
| 406 | private: |
| 407 | Error createSymtab(InstrProfSymtab &Symtab); |
| 408 | Error readNextHeader(const char *CurrentPos); |
| 409 | Error readHeader(const RawInstrProf::Header &Header); |
| 410 | |
| 411 | template <class IntT> IntT swap(IntT Int) const { |
| 412 | return ShouldSwapBytes ? sys::getSwappedBytes(Int) : Int; |
| 413 | } |
| 414 | |
| 415 | support::endianness getDataEndianness() const { |
| 416 | support::endianness HostEndian = getHostEndianness(); |
| 417 | if (!ShouldSwapBytes) |
| 418 | return HostEndian; |
| 419 | if (HostEndian == support::little) |
| 420 | return support::big; |
| 421 | else |
| 422 | return support::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 readValueProfilingData(InstrProfRecord &Record); |
| 433 | bool atEnd() const { return Data == DataEnd; } |
| 434 | |
| 435 | void advanceData() { |
| 436 | // `CountersDelta` is a constant zero when using debug info correlation. |
| 437 | if (!Correlator) { |
| 438 | // The initial CountersDelta is the in-memory address difference between |
| 439 | // the data and counts sections: |
| 440 | // start(__llvm_prf_cnts) - start(__llvm_prf_data) |
| 441 | // As we advance to the next record, we maintain the correct CountersDelta |
| 442 | // with respect to the next record. |
| 443 | CountersDelta -= sizeof(*Data); |
| 444 | } |
| 445 | Data++; |
| 446 | ValueDataStart += CurValueDataSize; |
| 447 | } |
| 448 | |
| 449 | const char *getNextHeaderPos() const { |
| 450 | assert(atEnd()); |
| 451 | return (const char *)ValueDataStart; |
| 452 | } |
| 453 | |
| 454 | StringRef getName(uint64_t NameRef) const { |
| 455 | return Symtab->getFuncName(swap(NameRef)); |
| 456 | } |
| 457 | |
| 458 | int getCounterTypeSize() const { |
| 459 | return hasSingleByteCoverage() ? sizeof(uint8_t) : sizeof(uint64_t); |
| 460 | } |
| 461 | }; |
| 462 | |
| 463 | using RawInstrProfReader32 = RawInstrProfReader<uint32_t>; |
| 464 | using RawInstrProfReader64 = RawInstrProfReader<uint64_t>; |
| 465 | |
| 466 | namespace IndexedInstrProf { |
| 467 | |
| 468 | enum class HashT : uint32_t; |
| 469 | |
| 470 | } // end namespace IndexedInstrProf |
| 471 | |
| 472 | /// Trait for lookups into the on-disk hash table for the binary instrprof |
| 473 | /// format. |
| 474 | class InstrProfLookupTrait { |
| 475 | std::vector<NamedInstrProfRecord> DataBuffer; |
| 476 | IndexedInstrProf::HashT HashType; |
| 477 | unsigned FormatVersion; |
| 478 | // Endianness of the input value profile data. |
| 479 | // It should be LE by default, but can be changed |
| 480 | // for testing purpose. |
| 481 | support::endianness ValueProfDataEndianness = support::little; |
| 482 | |
| 483 | public: |
| 484 | InstrProfLookupTrait(IndexedInstrProf::HashT HashType, unsigned FormatVersion) |
| 485 | : HashType(HashType), FormatVersion(FormatVersion) {} |
| 486 | |
| 487 | using data_type = ArrayRef<NamedInstrProfRecord>; |
| 488 | |
| 489 | using internal_key_type = StringRef; |
| 490 | using external_key_type = StringRef; |
| 491 | using hash_value_type = uint64_t; |
| 492 | using offset_type = uint64_t; |
| 493 | |
| 494 | static bool EqualKey(StringRef A, StringRef B) { return A == B; } |
| 495 | static StringRef GetInternalKey(StringRef K) { return K; } |
| 496 | static StringRef GetExternalKey(StringRef K) { return K; } |
| 497 | |
| 498 | hash_value_type ComputeHash(StringRef K); |
| 499 | |
| 500 | static std::pair<offset_type, offset_type> |
| 501 | ReadKeyDataLength(const unsigned char *&D) { |
| 502 | using namespace support; |
| 503 | |
| 504 | offset_type KeyLen = endian::readNext<offset_type, little, unaligned>(D); |
| 505 | offset_type DataLen = endian::readNext<offset_type, little, unaligned>(D); |
| 506 | return std::make_pair(KeyLen, DataLen); |
| 507 | } |
| 508 | |
| 509 | StringRef ReadKey(const unsigned char *D, offset_type N) { |
| 510 | return StringRef((const char *)D, N); |
| 511 | } |
| 512 | |
| 513 | bool readValueProfilingData(const unsigned char *&D, |
| 514 | const unsigned char *const End); |
| 515 | data_type ReadData(StringRef K, const unsigned char *D, offset_type N); |
| 516 | |
| 517 | // Used for testing purpose only. |
| 518 | void setValueProfDataEndianness(support::endianness Endianness) { |
| 519 | ValueProfDataEndianness = Endianness; |
| 520 | } |
| 521 | }; |
| 522 | |
| 523 | struct InstrProfReaderIndexBase { |
| 524 | virtual ~InstrProfReaderIndexBase() = default; |
| 525 | |
| 526 | // Read all the profile records with the same key pointed to the current |
| 527 | // iterator. |
| 528 | virtual Error getRecords(ArrayRef<NamedInstrProfRecord> &Data) = 0; |
| 529 | |
| 530 | // Read all the profile records with the key equal to FuncName |
| 531 | virtual Error getRecords(StringRef FuncName, |
| 532 | ArrayRef<NamedInstrProfRecord> &Data) = 0; |
| 533 | virtual void advanceToNextKey() = 0; |
| 534 | virtual bool atEnd() const = 0; |
| 535 | virtual void setValueProfDataEndianness(support::endianness Endianness) = 0; |
| 536 | virtual uint64_t getVersion() const = 0; |
| 537 | virtual bool isIRLevelProfile() const = 0; |
| 538 | virtual bool hasCSIRLevelProfile() const = 0; |
| 539 | virtual bool instrEntryBBEnabled() const = 0; |
| 540 | virtual bool hasSingleByteCoverage() const = 0; |
| 541 | virtual bool functionEntryOnly() const = 0; |
| 542 | virtual bool hasMemoryProfile() const = 0; |
| 543 | virtual bool hasTemporalProfile() const = 0; |
| 544 | virtual InstrProfKind getProfileKind() const = 0; |
| 545 | virtual Error populateSymtab(InstrProfSymtab &) = 0; |
| 546 | }; |
| 547 | |
| 548 | using OnDiskHashTableImplV3 = |
| 549 | OnDiskIterableChainedHashTable<InstrProfLookupTrait>; |
| 550 | |
| 551 | using MemProfRecordHashTable = |
| 552 | OnDiskIterableChainedHashTable<memprof::RecordLookupTrait>; |
| 553 | using MemProfFrameHashTable = |
| 554 | OnDiskIterableChainedHashTable<memprof::FrameLookupTrait>; |
| 555 | |
| 556 | template <typename HashTableImpl> |
| 557 | class InstrProfReaderItaniumRemapper; |
| 558 | |
| 559 | template <typename HashTableImpl> |
| 560 | class InstrProfReaderIndex : public InstrProfReaderIndexBase { |
| 561 | private: |
| 562 | std::unique_ptr<HashTableImpl> HashTable; |
| 563 | typename HashTableImpl::data_iterator RecordIterator; |
| 564 | uint64_t FormatVersion; |
| 565 | |
| 566 | friend class InstrProfReaderItaniumRemapper<HashTableImpl>; |
| 567 | |
| 568 | public: |
| 569 | InstrProfReaderIndex(const unsigned char *Buckets, |
| 570 | const unsigned char *const Payload, |
| 571 | const unsigned char *const Base, |
| 572 | IndexedInstrProf::HashT HashType, uint64_t Version); |
| 573 | ~InstrProfReaderIndex() override = default; |
| 574 | |
| 575 | Error getRecords(ArrayRef<NamedInstrProfRecord> &Data) override; |
| 576 | Error getRecords(StringRef FuncName, |
| 577 | ArrayRef<NamedInstrProfRecord> &Data) override; |
| 578 | void advanceToNextKey() override { RecordIterator++; } |
| 579 | |
| 580 | bool atEnd() const override { |
| 581 | return RecordIterator == HashTable->data_end(); |
| 582 | } |
| 583 | |
| 584 | void setValueProfDataEndianness(support::endianness Endianness) override { |
| 585 | HashTable->getInfoObj().setValueProfDataEndianness(Endianness); |
| 586 | } |
| 587 | |
| 588 | uint64_t getVersion() const override { return GET_VERSION(FormatVersion); } |
| 589 | |
| 590 | bool isIRLevelProfile() const override { |
| 591 | return (FormatVersion & VARIANT_MASK_IR_PROF) != 0; |
| 592 | } |
| 593 | |
| 594 | bool hasCSIRLevelProfile() const override { |
| 595 | return (FormatVersion & VARIANT_MASK_CSIR_PROF) != 0; |
| 596 | } |
| 597 | |
| 598 | bool instrEntryBBEnabled() const override { |
| 599 | return (FormatVersion & VARIANT_MASK_INSTR_ENTRY) != 0; |
| 600 | } |
| 601 | |
| 602 | bool hasSingleByteCoverage() const override { |
| 603 | return (FormatVersion & VARIANT_MASK_BYTE_COVERAGE) != 0; |
| 604 | } |
| 605 | |
| 606 | bool functionEntryOnly() const override { |
| 607 | return (FormatVersion & VARIANT_MASK_FUNCTION_ENTRY_ONLY) != 0; |
| 608 | } |
| 609 | |
| 610 | bool hasMemoryProfile() const override { |
| 611 | return (FormatVersion & VARIANT_MASK_MEMPROF) != 0; |
| 612 | } |
| 613 | |
| 614 | bool hasTemporalProfile() const override { |
| 615 | return (FormatVersion & VARIANT_MASK_TEMPORAL_PROF) != 0; |
| 616 | } |
| 617 | |
| 618 | InstrProfKind getProfileKind() const override; |
| 619 | |
| 620 | Error populateSymtab(InstrProfSymtab &Symtab) override { |
| 621 | return Symtab.create(HashTable->keys()); |
| 622 | } |
| 623 | }; |
| 624 | |
| 625 | /// Name matcher supporting fuzzy matching of symbol names to names in profiles. |
| 626 | class InstrProfReaderRemapper { |
| 627 | public: |
| 628 | virtual ~InstrProfReaderRemapper() = default; |
| 629 | virtual Error populateRemappings() { return Error::success(); } |
| 630 | virtual Error getRecords(StringRef FuncName, |
| 631 | ArrayRef<NamedInstrProfRecord> &Data) = 0; |
| 632 | }; |
| 633 | |
| 634 | /// Reader for the indexed binary instrprof format. |
| 635 | class IndexedInstrProfReader : public InstrProfReader { |
| 636 | private: |
| 637 | /// The profile data file contents. |
| 638 | std::unique_ptr<MemoryBuffer> DataBuffer; |
| 639 | /// The profile remapping file contents. |
| 640 | std::unique_ptr<MemoryBuffer> RemappingBuffer; |
| 641 | /// The index into the profile data. |
| 642 | std::unique_ptr<InstrProfReaderIndexBase> Index; |
| 643 | /// The profile remapping file contents. |
| 644 | std::unique_ptr<InstrProfReaderRemapper> Remapper; |
| 645 | /// Profile summary data. |
| 646 | std::unique_ptr<ProfileSummary> Summary; |
| 647 | /// Context sensitive profile summary data. |
| 648 | std::unique_ptr<ProfileSummary> CS_Summary; |
| 649 | /// MemProf profile schema (if available). |
| 650 | memprof::MemProfSchema Schema; |
| 651 | /// MemProf record profile data on-disk indexed via llvm::md5(FunctionName). |
| 652 | std::unique_ptr<MemProfRecordHashTable> MemProfRecordTable; |
| 653 | /// MemProf frame profile data on-disk indexed via frame id. |
| 654 | std::unique_ptr<MemProfFrameHashTable> MemProfFrameTable; |
| 655 | /// Total size of binary ids. |
| 656 | uint64_t BinaryIdsSize{0}; |
| 657 | /// Start address of binary id length and data pairs. |
| 658 | const uint8_t *BinaryIdsStart = nullptr; |
| 659 | |
| 660 | // Index to the current record in the record array. |
| 661 | unsigned RecordIndex; |
| 662 | |
| 663 | // Read the profile summary. Return a pointer pointing to one byte past the |
| 664 | // end of the summary data if it exists or the input \c Cur. |
| 665 | // \c UseCS indicates whether to use the context-sensitive profile summary. |
| 666 | const unsigned char *readSummary(IndexedInstrProf::ProfVersion Version, |
| 667 | const unsigned char *Cur, bool UseCS); |
| 668 | |
| 669 | public: |
| 670 | IndexedInstrProfReader( |
| 671 | std::unique_ptr<MemoryBuffer> DataBuffer, |
| 672 | std::unique_ptr<MemoryBuffer> RemappingBuffer = nullptr) |
| 673 | : DataBuffer(std::move(DataBuffer)), |
| 674 | RemappingBuffer(std::move(RemappingBuffer)), RecordIndex(0) {} |
| 675 | IndexedInstrProfReader(const IndexedInstrProfReader &) = delete; |
| 676 | IndexedInstrProfReader &operator=(const IndexedInstrProfReader &) = delete; |
| 677 | |
| 678 | /// Return the profile version. |
| 679 | uint64_t getVersion() const override { return Index->getVersion(); } |
| 680 | bool isIRLevelProfile() const override { return Index->isIRLevelProfile(); } |
| 681 | bool hasCSIRLevelProfile() const override { |
| 682 | return Index->hasCSIRLevelProfile(); |
| 683 | } |
| 684 | |
| 685 | bool instrEntryBBEnabled() const override { |
| 686 | return Index->instrEntryBBEnabled(); |
| 687 | } |
| 688 | |
| 689 | bool hasSingleByteCoverage() const override { |
| 690 | return Index->hasSingleByteCoverage(); |
| 691 | } |
| 692 | |
| 693 | bool functionEntryOnly() const override { return Index->functionEntryOnly(); } |
| 694 | |
| 695 | bool hasMemoryProfile() const override { return Index->hasMemoryProfile(); } |
| 696 | |
| 697 | bool hasTemporalProfile() const override { |
| 698 | return Index->hasTemporalProfile(); |
| 699 | } |
| 700 | |
| 701 | /// Returns a BitsetEnum describing the attributes of the indexed instr |
| 702 | /// profile. |
| 703 | InstrProfKind getProfileKind() const override { |
| 704 | return Index->getProfileKind(); |
| 705 | } |
| 706 | |
| 707 | /// Return true if the given buffer is in an indexed instrprof format. |
| 708 | static bool hasFormat(const MemoryBuffer &DataBuffer); |
| 709 | |
| 710 | /// Read the file header. |
| 711 | Error readHeader() override; |
| 712 | /// Read a single record. |
| 713 | Error readNextRecord(NamedInstrProfRecord &Record) override; |
| 714 | |
| 715 | /// Return the NamedInstrProfRecord associated with FuncName and FuncHash. |
| 716 | /// When return a hash_mismatch error and MismatchedFuncSum is not nullptr, |
| 717 | /// the sum of all counters in the mismatched function will be set to |
| 718 | /// MismatchedFuncSum. If there are multiple instances of mismatched |
| 719 | /// functions, MismatchedFuncSum returns the maximum. If \c FuncName is not |
| 720 | /// found, try to lookup \c DeprecatedFuncName to handle profiles built by |
| 721 | /// older compilers. |
| 722 | Expected<InstrProfRecord> |
| 723 | getInstrProfRecord(StringRef FuncName, uint64_t FuncHash, |
| 724 | StringRef DeprecatedFuncName = "", |
| 725 | uint64_t *MismatchedFuncSum = nullptr); |
| 726 | |
| 727 | /// Return the memprof record for the function identified by |
| 728 | /// llvm::md5(Name). |
| 729 | Expected<memprof::MemProfRecord> getMemProfRecord(uint64_t FuncNameHash); |
| 730 | |
| 731 | /// Fill Counts with the profile data for the given function name. |
| 732 | Error getFunctionCounts(StringRef FuncName, uint64_t FuncHash, |
| 733 | std::vector<uint64_t> &Counts); |
| 734 | |
| 735 | /// Return the maximum of all known function counts. |
| 736 | /// \c UseCS indicates whether to use the context-sensitive count. |
| 737 | uint64_t getMaximumFunctionCount(bool UseCS) { |
| 738 | if (UseCS) { |
| 739 | assert(CS_Summary && "No context sensitive profile summary"); |
| 740 | return CS_Summary->getMaxFunctionCount(); |
| 741 | } else { |
| 742 | assert(Summary && "No profile summary"); |
| 743 | return Summary->getMaxFunctionCount(); |
| 744 | } |
| 745 | } |
| 746 | |
| 747 | /// Factory method to create an indexed reader. |
| 748 | static Expected<std::unique_ptr<IndexedInstrProfReader>> |
| 749 | create(const Twine &Path, vfs::FileSystem &FS, |
| 750 | const Twine &RemappingPath = ""); |
| 751 | |
| 752 | static Expected<std::unique_ptr<IndexedInstrProfReader>> |
| 753 | create(std::unique_ptr<MemoryBuffer> Buffer, |
| 754 | std::unique_ptr<MemoryBuffer> RemappingBuffer = nullptr); |
| 755 | |
| 756 | // Used for testing purpose only. |
| 757 | void setValueProfDataEndianness(support::endianness Endianness) { |
| 758 | Index->setValueProfDataEndianness(Endianness); |
| 759 | } |
| 760 | |
| 761 | // See description in the base class. This interface is designed |
| 762 | // to be used by llvm-profdata (for dumping). Avoid using this when |
| 763 | // the client is the compiler. |
| 764 | InstrProfSymtab &getSymtab() override; |
| 765 | |
| 766 | /// Return the profile summary. |
| 767 | /// \c UseCS indicates whether to use the context-sensitive summary. |
| 768 | ProfileSummary &getSummary(bool UseCS) { |
| 769 | if (UseCS) { |
| 770 | assert(CS_Summary && "No context sensitive summary"); |
| 771 | return *CS_Summary; |
| 772 | } else { |
| 773 | assert(Summary && "No profile summary"); |
| 774 | return *Summary; |
| 775 | } |
| 776 | } |
| 777 | |
| 778 | Error readBinaryIds(std::vector<llvm::object::BuildID> &BinaryIds) override; |
| 779 | Error printBinaryIds(raw_ostream &OS) override; |
| 780 | }; |
| 781 | |
| 782 | } // end namespace llvm |
| 783 | |
| 784 | #endif // LLVM_PROFILEDATA_INSTRPROFREADER_H |