|
| 1 | +//===- SymbolFilter.h - Utilities for Symbol Filtering ---------*- 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 | +#ifndef LLVM_EXECUTIONENGINE_ORC_SHARED_SYMBOLFILTER_H |
| 10 | +#define LLVM_EXECUTIONENGINE_ORC_SHARED_SYMBOLFILTER_H |
| 11 | + |
| 12 | +#include "llvm/ExecutionEngine/Orc/Shared/SimplePackedSerialization.h" |
| 13 | + |
| 14 | +#include <cmath> |
| 15 | +#include <type_traits> |
| 16 | +#include <vector> |
| 17 | + |
| 18 | +namespace llvm { |
| 19 | +namespace orc { |
| 20 | + |
| 21 | +namespace shared { |
| 22 | +using SPSBloomFilter = |
| 23 | + SPSTuple<bool, uint32_t, uint32_t, uint32_t, SPSSequence<uint64_t>>; |
| 24 | +} |
| 25 | + |
| 26 | +class BloomFilter { |
| 27 | +public: |
| 28 | + using HashFunc = std::function<uint32_t(StringRef)>; |
| 29 | + |
| 30 | + BloomFilter() = default; |
| 31 | + BloomFilter(BloomFilter &&) noexcept = default; |
| 32 | + BloomFilter &operator=(BloomFilter &&) noexcept = default; |
| 33 | + BloomFilter(const BloomFilter &) = delete; |
| 34 | + BloomFilter &operator=(const BloomFilter &) = delete; |
| 35 | + |
| 36 | + BloomFilter(uint32_t SymbolCount, float FalsePositiveRate, HashFunc hashFn) |
| 37 | + : HashFn(std::move(hashFn)) { |
| 38 | + initialize(SymbolCount, FalsePositiveRate); |
| 39 | + } |
| 40 | + bool isInitialized() const { return Initialized; } |
| 41 | + |
| 42 | + void add(StringRef Sym) { |
| 43 | + assert(Initialized); |
| 44 | + addHash(HashFn(Sym)); |
| 45 | + } |
| 46 | + |
| 47 | + bool mayContain(StringRef Sym) const { |
| 48 | + return !isEmpty() && testHash(HashFn(Sym)); |
| 49 | + } |
| 50 | + |
| 51 | + bool isEmpty() const { return SymbolCount == 0; } |
| 52 | + |
| 53 | +private: |
| 54 | + friend class shared::SPSSerializationTraits<shared::SPSBloomFilter, |
| 55 | + BloomFilter>; |
| 56 | + static constexpr uint32_t BitsPerEntry = 64; |
| 57 | + |
| 58 | + bool Initialized = false; |
| 59 | + uint32_t SymbolCount = 0; |
| 60 | + uint32_t BloomSize = 0; |
| 61 | + uint32_t BloomShift = 0; |
| 62 | + std::vector<uint64_t> BloomTable; |
| 63 | + HashFunc HashFn; |
| 64 | + |
| 65 | + void initialize(uint32_t SymCount, float FalsePositiveRate) { |
| 66 | + assert(SymCount > 0); |
| 67 | + SymbolCount = SymCount; |
| 68 | + Initialized = true; |
| 69 | + |
| 70 | + float ln2 = std::log(2.0f); |
| 71 | + float M = -1.0f * SymbolCount * std::log(FalsePositiveRate) / (ln2 * ln2); |
| 72 | + BloomSize = static_cast<uint32_t>(std::ceil(M / BitsPerEntry)); |
| 73 | + BloomShift = std::min(6u, log2ceil(SymbolCount)); |
| 74 | + BloomTable.resize(BloomSize, 0); |
| 75 | + } |
| 76 | + |
| 77 | + void addHash(uint32_t Hash) { |
| 78 | + uint32_t Hash2 = Hash >> BloomShift; |
| 79 | + uint32_t N = (Hash / BitsPerEntry) % BloomSize; |
| 80 | + uint64_t Mask = |
| 81 | + (1ULL << (Hash % BitsPerEntry)) | (1ULL << (Hash2 % BitsPerEntry)); |
| 82 | + BloomTable[N] |= Mask; |
| 83 | + } |
| 84 | + |
| 85 | + bool testHash(uint32_t Hash) const { |
| 86 | + uint32_t Hash2 = Hash >> BloomShift; |
| 87 | + uint32_t N = (Hash / BitsPerEntry) % BloomSize; |
| 88 | + uint64_t Mask = |
| 89 | + (1ULL << (Hash % BitsPerEntry)) | (1ULL << (Hash2 % BitsPerEntry)); |
| 90 | + return (BloomTable[N] & Mask) == Mask; |
| 91 | + } |
| 92 | + |
| 93 | + static constexpr uint32_t log2ceil(uint32_t V) { |
| 94 | + return V <= 1 ? 0 : 32 - countl_zero(V - 1); |
| 95 | + } |
| 96 | +}; |
| 97 | + |
| 98 | +class BloomFilterBuilder { |
| 99 | +public: |
| 100 | + using HashFunc = BloomFilter::HashFunc; |
| 101 | + |
| 102 | + BloomFilterBuilder() = default; |
| 103 | + |
| 104 | + BloomFilterBuilder &setFalsePositiveRate(float Rate) { |
| 105 | + assert(Rate > 0.0f && Rate < 1.0f); |
| 106 | + FalsePositiveRate = Rate; |
| 107 | + return *this; |
| 108 | + } |
| 109 | + |
| 110 | + BloomFilterBuilder &setHashFunction(HashFunc Fn) { |
| 111 | + HashFn = std::move(Fn); |
| 112 | + return *this; |
| 113 | + } |
| 114 | + |
| 115 | + BloomFilter build(ArrayRef<StringRef> Symbols) const { |
| 116 | + assert(!Symbols.empty() && "Cannot build filter from empty symbol list."); |
| 117 | + BloomFilter F(static_cast<uint32_t>(Symbols.size()), FalsePositiveRate, |
| 118 | + HashFn); |
| 119 | + for (const auto &Sym : Symbols) |
| 120 | + F.add(Sym); |
| 121 | + |
| 122 | + return F; |
| 123 | + } |
| 124 | + |
| 125 | +private: |
| 126 | + float FalsePositiveRate = 0.02f; |
| 127 | + HashFunc HashFn = [](StringRef S) -> uint32_t { |
| 128 | + uint32_t H = 5381; |
| 129 | + for (char C : S) |
| 130 | + H = ((H << 5) + H) + static_cast<uint8_t>(C); // H * 33 + C |
| 131 | + return H; |
| 132 | + }; |
| 133 | +}; |
| 134 | + |
| 135 | +namespace shared { |
| 136 | + |
| 137 | +template <> class SPSSerializationTraits<SPSBloomFilter, BloomFilter> { |
| 138 | +public: |
| 139 | + static size_t size(const BloomFilter &Filter) { |
| 140 | + return SPSBloomFilter::AsArgList::size( |
| 141 | + Filter.Initialized, Filter.SymbolCount, Filter.BloomSize, |
| 142 | + Filter.BloomShift, Filter.BloomTable); |
| 143 | + } |
| 144 | + |
| 145 | + static bool serialize(SPSOutputBuffer &OB, const BloomFilter &Filter) { |
| 146 | + return SPSBloomFilter::AsArgList::serialize( |
| 147 | + OB, Filter.Initialized, Filter.SymbolCount, Filter.BloomSize, |
| 148 | + Filter.BloomShift, Filter.BloomTable); |
| 149 | + } |
| 150 | + |
| 151 | + static bool deserialize(SPSInputBuffer &IB, BloomFilter &Filter) { |
| 152 | + bool IsInitialized; |
| 153 | + uint32_t SymbolCount = 0, BloomSize = 0, BloomShift = 0; |
| 154 | + std::vector<uint64_t> BloomTable; |
| 155 | + |
| 156 | + if (!SPSBloomFilter::AsArgList::deserialize( |
| 157 | + IB, IsInitialized, SymbolCount, BloomSize, BloomShift, BloomTable)) |
| 158 | + return false; |
| 159 | + |
| 160 | + Filter.Initialized = IsInitialized; |
| 161 | + Filter.SymbolCount = SymbolCount; |
| 162 | + Filter.BloomSize = BloomSize; |
| 163 | + Filter.BloomShift = BloomShift; |
| 164 | + Filter.BloomTable = std::move(BloomTable); |
| 165 | + |
| 166 | + return true; |
| 167 | + } |
| 168 | +}; |
| 169 | + |
| 170 | +} // end namespace shared |
| 171 | +} // end namespace orc |
| 172 | +} // end namespace llvm |
| 173 | +#endif // LLVM_EXECUTIONENGINE_ORC_SHARED_SYMBOLFILTER_H |
0 commit comments