Line data Source code
1 : // Copyright (c) 2012-2022 The Bitcoin Core developers
2 : // Distributed under the MIT software license, see the accompanying
3 : // file COPYING or http://www.opensource.org/licenses/mit-license.php.
4 :
5 : #include <common/bloom.h>
6 :
7 : #include <hash.h>
8 : #include <primitives/transaction.h>
9 : #include <random.h>
10 : #include <script/script.h>
11 : #include <script/solver.h>
12 : #include <span.h>
13 : #include <streams.h>
14 : #include <util/fastrange.h>
15 :
16 : #include <algorithm>
17 : #include <cmath>
18 : #include <cstdlib>
19 : #include <limits>
20 : #include <vector>
21 :
22 : static constexpr double LN2SQUARED = 0.4804530139182014246671025263266649717305529515945455;
23 : static constexpr double LN2 = 0.6931471805599453094172321214581765680755001343602552;
24 :
25 0 : CBloomFilter::CBloomFilter(const unsigned int nElements, const double nFPRate, const unsigned int nTweakIn, unsigned char nFlagsIn) :
26 : /**
27 : * The ideal size for a bloom filter with a given number of elements and false positive rate is:
28 : * - nElements * log(fp rate) / ln(2)^2
29 : * We ignore filter parameters which will create a bloom filter larger than the protocol limits
30 : */
31 0 : vData(std::min((unsigned int)(-1 / LN2SQUARED * nElements * log(nFPRate)), MAX_BLOOM_FILTER_SIZE * 8) / 8),
32 : /**
33 : * The ideal number of hash functions is filter size * ln(2) / number of elements
34 : * Again, we ignore filter parameters which will create a bloom filter with more hash functions than the protocol limits
35 : * See https://en.wikipedia.org/wiki/Bloom_filter for an explanation of these formulas
36 : */
37 0 : nHashFuncs(std::min((unsigned int)(vData.size() * 8 / nElements * LN2), MAX_HASH_FUNCS)),
38 0 : nTweak(nTweakIn),
39 0 : nFlags(nFlagsIn)
40 : {
41 0 : }
42 :
43 0 : inline unsigned int CBloomFilter::Hash(unsigned int nHashNum, Span<const unsigned char> vDataToHash) const
44 : {
45 : // 0xFBA4C795 chosen as it guarantees a reasonable bit difference between nHashNum values.
46 0 : return MurmurHash3(nHashNum * 0xFBA4C795 + nTweak, vDataToHash) % (vData.size() * 8);
47 : }
48 :
49 0 : void CBloomFilter::insert(Span<const unsigned char> vKey)
50 : {
51 0 : if (vData.empty()) // Avoid divide-by-zero (CVE-2013-5700)
52 0 : return;
53 0 : for (unsigned int i = 0; i < nHashFuncs; i++)
54 : {
55 0 : unsigned int nIndex = Hash(i, vKey);
56 : // Sets bit nIndex of vData
57 0 : vData[nIndex >> 3] |= (1 << (7 & nIndex));
58 0 : }
59 0 : }
60 :
61 0 : void CBloomFilter::insert(const COutPoint& outpoint)
62 : {
63 0 : DataStream stream{};
64 0 : stream << outpoint;
65 0 : insert(MakeUCharSpan(stream));
66 0 : }
67 :
68 0 : bool CBloomFilter::contains(Span<const unsigned char> vKey) const
69 : {
70 0 : if (vData.empty()) // Avoid divide-by-zero (CVE-2013-5700)
71 0 : return true;
72 0 : for (unsigned int i = 0; i < nHashFuncs; i++)
73 : {
74 0 : unsigned int nIndex = Hash(i, vKey);
75 : // Checks bit nIndex of vData
76 0 : if (!(vData[nIndex >> 3] & (1 << (7 & nIndex))))
77 0 : return false;
78 0 : }
79 0 : return true;
80 0 : }
81 :
82 0 : bool CBloomFilter::contains(const COutPoint& outpoint) const
83 : {
84 0 : DataStream stream{};
85 0 : stream << outpoint;
86 0 : return contains(MakeUCharSpan(stream));
87 0 : }
88 :
89 0 : bool CBloomFilter::IsWithinSizeConstraints() const
90 : {
91 0 : return vData.size() <= MAX_BLOOM_FILTER_SIZE && nHashFuncs <= MAX_HASH_FUNCS;
92 : }
93 :
94 0 : bool CBloomFilter::IsRelevantAndUpdate(const CTransaction& tx)
95 : {
96 0 : bool fFound = false;
97 : // Match if the filter contains the hash of tx
98 : // for finding tx when they appear in a block
99 0 : if (vData.empty()) // zero-size = "match-all" filter
100 0 : return true;
101 0 : const uint256& hash = tx.GetHash();
102 0 : if (contains(hash))
103 0 : fFound = true;
104 :
105 0 : for (unsigned int i = 0; i < tx.vout.size(); i++)
106 : {
107 0 : const CTxOut& txout = tx.vout[i];
108 : // Match if the filter contains any arbitrary script data element in any scriptPubKey in tx
109 : // If this matches, also add the specific output that was matched.
110 : // This means clients don't have to update the filter themselves when a new relevant tx
111 : // is discovered in order to find spending transactions, which avoids round-tripping and race conditions.
112 0 : CScript::const_iterator pc = txout.scriptPubKey.begin();
113 0 : std::vector<unsigned char> data;
114 0 : while (pc < txout.scriptPubKey.end())
115 : {
116 : opcodetype opcode;
117 0 : if (!txout.scriptPubKey.GetOp(pc, opcode, data))
118 0 : break;
119 0 : if (data.size() != 0 && contains(data))
120 : {
121 0 : fFound = true;
122 0 : if ((nFlags & BLOOM_UPDATE_MASK) == BLOOM_UPDATE_ALL)
123 0 : insert(COutPoint(hash, i));
124 0 : else if ((nFlags & BLOOM_UPDATE_MASK) == BLOOM_UPDATE_P2PUBKEY_ONLY)
125 : {
126 0 : std::vector<std::vector<unsigned char> > vSolutions;
127 0 : TxoutType type = Solver(txout.scriptPubKey, vSolutions);
128 0 : if (type == TxoutType::PUBKEY || type == TxoutType::MULTISIG) {
129 0 : insert(COutPoint(hash, i));
130 0 : }
131 0 : }
132 0 : break;
133 : }
134 : }
135 0 : }
136 :
137 0 : if (fFound)
138 0 : return true;
139 :
140 0 : for (const CTxIn& txin : tx.vin)
141 : {
142 : // Match if the filter contains an outpoint tx spends
143 0 : if (contains(txin.prevout))
144 0 : return true;
145 :
146 : // Match if the filter contains any arbitrary script data element in any scriptSig in tx
147 0 : CScript::const_iterator pc = txin.scriptSig.begin();
148 0 : std::vector<unsigned char> data;
149 0 : while (pc < txin.scriptSig.end())
150 : {
151 : opcodetype opcode;
152 0 : if (!txin.scriptSig.GetOp(pc, opcode, data))
153 0 : break;
154 0 : if (data.size() != 0 && contains(data))
155 0 : return true;
156 : }
157 0 : }
158 :
159 0 : return false;
160 0 : }
161 :
162 16710 : CRollingBloomFilter::CRollingBloomFilter(const unsigned int nElements, const double fpRate)
163 : {
164 16710 : double logFpRate = log(fpRate);
165 : /* The optimal number of hash functions is log(fpRate) / log(0.5), but
166 : * restrict it to the range 1-50. */
167 16710 : nHashFuncs = std::max(1, std::min((int)round(logFpRate / log(0.5)), 50));
168 : /* In this rolling bloom filter, we'll store between 2 and 3 generations of nElements / 2 entries. */
169 16710 : nEntriesPerGeneration = (nElements + 1) / 2;
170 16710 : uint32_t nMaxElements = nEntriesPerGeneration * 3;
171 : /* The maximum fpRate = pow(1.0 - exp(-nHashFuncs * nMaxElements / nFilterBits), nHashFuncs)
172 : * => pow(fpRate, 1.0 / nHashFuncs) = 1.0 - exp(-nHashFuncs * nMaxElements / nFilterBits)
173 : * => 1.0 - pow(fpRate, 1.0 / nHashFuncs) = exp(-nHashFuncs * nMaxElements / nFilterBits)
174 : * => log(1.0 - pow(fpRate, 1.0 / nHashFuncs)) = -nHashFuncs * nMaxElements / nFilterBits
175 : * => nFilterBits = -nHashFuncs * nMaxElements / log(1.0 - pow(fpRate, 1.0 / nHashFuncs))
176 : * => nFilterBits = -nHashFuncs * nMaxElements / log(1.0 - exp(logFpRate / nHashFuncs))
177 : */
178 16710 : uint32_t nFilterBits = (uint32_t)ceil(-1.0 * nHashFuncs * nMaxElements / log(1.0 - exp(logFpRate / nHashFuncs)));
179 16710 : data.clear();
180 : /* For each data element we need to store 2 bits. If both bits are 0, the
181 : * bit is treated as unset. If the bits are (01), (10), or (11), the bit is
182 : * treated as set in generation 1, 2, or 3 respectively.
183 : * These bits are stored in separate integers: position P corresponds to bit
184 : * (P & 63) of the integers data[(P >> 6) * 2] and data[(P >> 6) * 2 + 1]. */
185 16710 : data.resize(((nFilterBits + 63) / 64) << 1);
186 16710 : reset();
187 16710 : }
188 :
189 : /* Similar to CBloomFilter::Hash */
190 12120 : static inline uint32_t RollingBloomHash(unsigned int nHashNum, uint32_t nTweak, Span<const unsigned char> vDataToHash)
191 : {
192 12120 : return MurmurHash3(nHashNum * 0xFBA4C795 + nTweak, vDataToHash);
193 : }
194 :
195 606 : void CRollingBloomFilter::insert(Span<const unsigned char> vKey)
196 : {
197 606 : if (nEntriesThisGeneration == nEntriesPerGeneration) {
198 0 : nEntriesThisGeneration = 0;
199 0 : nGeneration++;
200 0 : if (nGeneration == 4) {
201 0 : nGeneration = 1;
202 0 : }
203 0 : uint64_t nGenerationMask1 = 0 - (uint64_t)(nGeneration & 1);
204 0 : uint64_t nGenerationMask2 = 0 - (uint64_t)(nGeneration >> 1);
205 : /* Wipe old entries that used this generation number. */
206 0 : for (uint32_t p = 0; p < data.size(); p += 2) {
207 0 : uint64_t p1 = data[p], p2 = data[p + 1];
208 0 : uint64_t mask = (p1 ^ nGenerationMask1) | (p2 ^ nGenerationMask2);
209 0 : data[p] = p1 & mask;
210 0 : data[p + 1] = p2 & mask;
211 0 : }
212 0 : }
213 606 : nEntriesThisGeneration++;
214 :
215 12726 : for (int n = 0; n < nHashFuncs; n++) {
216 12120 : uint32_t h = RollingBloomHash(n, nTweak, vKey);
217 12120 : int bit = h & 0x3F;
218 : /* FastMod works with the upper bits of h, so it is safe to ignore that the lower bits of h are already used for bit. */
219 12120 : uint32_t pos = FastRange32(h, data.size());
220 : /* The lowest bit of pos is ignored, and set to zero for the first bit, and to one for the second. */
221 12120 : data[pos & ~1U] = (data[pos & ~1U] & ~(uint64_t{1} << bit)) | (uint64_t(nGeneration & 1)) << bit;
222 12120 : data[pos | 1] = (data[pos | 1] & ~(uint64_t{1} << bit)) | (uint64_t(nGeneration >> 1)) << bit;
223 12120 : }
224 606 : }
225 :
226 0 : bool CRollingBloomFilter::contains(Span<const unsigned char> vKey) const
227 : {
228 0 : for (int n = 0; n < nHashFuncs; n++) {
229 0 : uint32_t h = RollingBloomHash(n, nTweak, vKey);
230 0 : int bit = h & 0x3F;
231 0 : uint32_t pos = FastRange32(h, data.size());
232 : /* If the relevant bit is not set in either data[pos & ~1] or data[pos | 1], the filter does not contain vKey */
233 0 : if (!(((data[pos & ~1U] | data[pos | 1]) >> bit) & 1)) {
234 0 : return false;
235 : }
236 0 : }
237 0 : return true;
238 0 : }
239 :
240 16710 : void CRollingBloomFilter::reset()
241 : {
242 16710 : nTweak = GetRand<unsigned int>();
243 16710 : nEntriesThisGeneration = 0;
244 16710 : nGeneration = 1;
245 16710 : std::fill(data.begin(), data.end(), 0);
246 16710 : }
|
