/Users/eugenesiegel/btc/bitcoin/src/blockencodings.cpp
Line | Count | Source (jump to first uncovered line) |
1 | | // Copyright (c) 2016-2022 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 <blockencodings.h> |
6 | | #include <chainparams.h> |
7 | | #include <common/system.h> |
8 | | #include <consensus/consensus.h> |
9 | | #include <consensus/validation.h> |
10 | | #include <crypto/sha256.h> |
11 | | #include <crypto/siphash.h> |
12 | | #include <logging.h> |
13 | | #include <random.h> |
14 | | #include <streams.h> |
15 | | #include <txmempool.h> |
16 | | #include <validation.h> |
17 | | |
18 | | #include <unordered_map> |
19 | | |
20 | | CBlockHeaderAndShortTxIDs::CBlockHeaderAndShortTxIDs(const CBlock& block, const uint64_t nonce) : |
21 | 2.77M | nonce(nonce), |
22 | 2.77M | shorttxids(block.vtx.size() - 1), prefilledtxn(1), header(block) { |
23 | 2.77M | FillShortTxIDSelector(); |
24 | | //TODO: Use our mempool prior to block acceptance to predictively fill more than just the coinbase |
25 | 2.77M | prefilledtxn[0] = {0, block.vtx[0]}; |
26 | 7.72M | for (size_t i = 1; i < block.vtx.size(); i++4.94M ) { |
27 | 4.94M | const CTransaction& tx = *block.vtx[i]; |
28 | 4.94M | shorttxids[i - 1] = GetShortID(tx.GetWitnessHash()); |
29 | 4.94M | } |
30 | 2.77M | } |
31 | | |
32 | 5.24M | void CBlockHeaderAndShortTxIDs::FillShortTxIDSelector() const { |
33 | 5.24M | DataStream stream{}; |
34 | 5.24M | stream << header << nonce; |
35 | 5.24M | CSHA256 hasher; |
36 | 5.24M | hasher.Write((unsigned char*)&(*stream.begin()), stream.end() - stream.begin()); |
37 | 5.24M | uint256 shorttxidhash; |
38 | 5.24M | hasher.Finalize(shorttxidhash.begin()); |
39 | 5.24M | shorttxidk0 = shorttxidhash.GetUint64(0); |
40 | 5.24M | shorttxidk1 = shorttxidhash.GetUint64(1); |
41 | 5.24M | } |
42 | | |
43 | 37.3M | uint64_t CBlockHeaderAndShortTxIDs::GetShortID(const Wtxid& wtxid) const { |
44 | 37.3M | static_assert(SHORTTXIDS_LENGTH == 6, "shorttxids calculation assumes 6-byte shorttxids"); |
45 | 37.3M | return SipHashUint256(shorttxidk0, shorttxidk1, wtxid.ToUint256()) & 0xffffffffffffL; |
46 | 37.3M | } |
47 | | |
48 | | /* Reconstructing a compact block is in the hot-path for block relay, |
49 | | * so we want to do it as quickly as possible. Because this often |
50 | | * involves iterating over the entire mempool, we put all the data we |
51 | | * need (ie the wtxid and a reference to the actual transaction data) |
52 | | * in a vector and iterate over the vector directly. This allows optimal |
53 | | * CPU caching behaviour, at a cost of only 40 bytes per transaction. |
54 | | */ |
55 | | ReadStatus PartiallyDownloadedBlock::InitData(const CBlockHeaderAndShortTxIDs& cmpctblock, const std::vector<std::pair<Wtxid, CTransactionRef>>& extra_txn) |
56 | 532k | { |
57 | 532k | LogDebug(BCLog::CMPCTBLOCK, "Initializing PartiallyDownloadedBlock for block %s using a cmpctblock of %u bytes\n", cmpctblock.header.GetHash().ToString(), GetSerializeSize(cmpctblock)); Line | Count | Source | 381 | 532k | #define LogDebug(category, ...) LogPrintLevel(category, BCLog::Level::Debug, __VA_ARGS__) Line | Count | Source | 373 | 532k | do { \ | 374 | 532k | if (LogAcceptCategory((category), (level))) { \ | 375 | 0 | bool rate_limit{level >= BCLog::Level::Info}; \ | 376 | 0 | LogPrintLevel_(category, level, rate_limit, __VA_ARGS__); \ Line | Count | Source | 350 | 0 | #define LogPrintLevel_(category, level, should_ratelimit, ...) LogPrintFormatInternal(std::source_location::current(), category, level, should_ratelimit, __VA_ARGS__) |
| 377 | 0 | } \ | 378 | 532k | } while (0) |
|
|
58 | 532k | if (cmpctblock.header.IsNull() || (cmpctblock.shorttxids.empty() && cmpctblock.prefilledtxn.empty()21.4k )) |
59 | 0 | return READ_STATUS_INVALID; |
60 | 532k | if (cmpctblock.shorttxids.size() + cmpctblock.prefilledtxn.size() > MAX_BLOCK_WEIGHT / MIN_SERIALIZABLE_TRANSACTION_WEIGHT) |
61 | 0 | return READ_STATUS_INVALID; |
62 | | |
63 | 532k | if (!header.IsNull() || !txn_available.empty()) return READ_STATUS_INVALID0 ; |
64 | | |
65 | 532k | header = cmpctblock.header; |
66 | 532k | txn_available.resize(cmpctblock.BlockTxCount()); |
67 | | |
68 | 532k | int32_t lastprefilledindex = -1; |
69 | 1.07M | for (size_t i = 0; i < cmpctblock.prefilledtxn.size(); i++538k ) { |
70 | 538k | if (cmpctblock.prefilledtxn[i].tx->IsNull()) |
71 | 0 | return READ_STATUS_INVALID; |
72 | | |
73 | 538k | lastprefilledindex += cmpctblock.prefilledtxn[i].index + 1; //index is a uint16_t, so can't overflow here |
74 | 538k | if (lastprefilledindex > std::numeric_limits<uint16_t>::max()) |
75 | 0 | return READ_STATUS_INVALID; |
76 | 538k | if ((uint32_t)lastprefilledindex > cmpctblock.shorttxids.size() + i) { |
77 | | // If we are inserting a tx at an index greater than our full list of shorttxids |
78 | | // plus the number of prefilled txn we've inserted, then we have txn for which we |
79 | | // have neither a prefilled txn or a shorttxid! |
80 | 0 | return READ_STATUS_INVALID; |
81 | 0 | } |
82 | 538k | txn_available[lastprefilledindex] = cmpctblock.prefilledtxn[i].tx; |
83 | 538k | } |
84 | 532k | prefilled_count = cmpctblock.prefilledtxn.size(); |
85 | | |
86 | | // Calculate map of txids -> positions and check mempool to see what we have (or don't) |
87 | | // Because well-formed cmpctblock messages will have a (relatively) uniform distribution |
88 | | // of short IDs, any highly-uneven distribution of elements can be safely treated as a |
89 | | // READ_STATUS_FAILED. |
90 | 532k | std::unordered_map<uint64_t, uint16_t> shorttxids(cmpctblock.shorttxids.size()); |
91 | 532k | uint16_t index_offset = 0; |
92 | 1.52M | for (size_t i = 0; i < cmpctblock.shorttxids.size(); i++994k ) { |
93 | 1.50M | while (txn_available[i + index_offset]) |
94 | 512k | index_offset++; |
95 | 994k | shorttxids[cmpctblock.shorttxids[i]] = i + index_offset; |
96 | | // To determine the chance that the number of entries in a bucket exceeds N, |
97 | | // we use the fact that the number of elements in a single bucket is |
98 | | // binomially distributed (with n = the number of shorttxids S, and p = |
99 | | // 1 / the number of buckets), that in the worst case the number of buckets is |
100 | | // equal to S (due to std::unordered_map having a default load factor of 1.0), |
101 | | // and that the chance for any bucket to exceed N elements is at most |
102 | | // buckets * (the chance that any given bucket is above N elements). |
103 | | // Thus: P(max_elements_per_bucket > N) <= S * (1 - cdf(binomial(n=S,p=1/S), N)). |
104 | | // If we assume blocks of up to 16000, allowing 12 elements per bucket should |
105 | | // only fail once per ~1 million block transfers (per peer and connection). |
106 | 994k | if (shorttxids.bucket_size(shorttxids.bucket(cmpctblock.shorttxids[i])) > 12) |
107 | 0 | return READ_STATUS_FAILED; |
108 | 994k | } |
109 | | // TODO: in the shortid-collision case, we should instead request both transactions |
110 | | // which collided. Falling back to full-block-request here is overkill. |
111 | 532k | if (shorttxids.size() != cmpctblock.shorttxids.size()) |
112 | 10.0k | return READ_STATUS_FAILED; // Short ID collision |
113 | | |
114 | 522k | std::vector<bool> have_txn(txn_available.size()); |
115 | 522k | { |
116 | 522k | LOCK(pool->cs); Line | Count | Source | 259 | 522k | #define LOCK(cs) UniqueLock UNIQUE_NAME(criticalblock)(MaybeCheckNotHeld(cs), #cs, __FILE__, __LINE__) Line | Count | Source | 11 | 522k | #define UNIQUE_NAME(name) PASTE2(name, __COUNTER__) Line | Count | Source | 9 | 522k | #define PASTE2(x, y) PASTE(x, y) Line | Count | Source | 8 | 522k | #define PASTE(x, y) x ## y |
|
|
|
|
117 | 1.57M | for (const auto& [wtxid, txit] : pool->txns_randomized) { |
118 | 1.57M | uint64_t shortid = cmpctblock.GetShortID(wtxid); |
119 | 1.57M | std::unordered_map<uint64_t, uint16_t>::iterator idit = shorttxids.find(shortid); |
120 | 1.57M | if (idit != shorttxids.end()) { |
121 | 570k | if (!have_txn[idit->second]) { |
122 | 570k | txn_available[idit->second] = txit->GetSharedTx(); |
123 | 570k | have_txn[idit->second] = true; |
124 | 570k | mempool_count++; |
125 | 570k | } else { |
126 | | // If we find two mempool txn that match the short id, just request it. |
127 | | // This should be rare enough that the extra bandwidth doesn't matter, |
128 | | // but eating a round-trip due to FillBlock failure would be annoying |
129 | 0 | if (txn_available[idit->second]) { |
130 | 0 | txn_available[idit->second].reset(); |
131 | 0 | mempool_count--; |
132 | 0 | } |
133 | 0 | } |
134 | 570k | } |
135 | | // Though ideally we'd continue scanning for the two-txn-match-shortid case, |
136 | | // the performance win of an early exit here is too good to pass up and worth |
137 | | // the extra risk. |
138 | 1.57M | if (mempool_count == shorttxids.size()) |
139 | 120k | break; |
140 | 1.57M | } |
141 | 522k | } |
142 | | |
143 | 31.1M | for (size_t i = 0; i < extra_txn.size(); i++30.5M ) { |
144 | 30.7M | uint64_t shortid = cmpctblock.GetShortID(extra_txn[i].first); |
145 | 30.7M | std::unordered_map<uint64_t, uint16_t>::iterator idit = shorttxids.find(shortid); |
146 | 30.7M | if (idit != shorttxids.end()) { |
147 | 74.4k | if (!have_txn[idit->second]) { |
148 | 72.8k | txn_available[idit->second] = extra_txn[i].second; |
149 | 72.8k | have_txn[idit->second] = true; |
150 | 72.8k | mempool_count++; |
151 | 72.8k | extra_count++; |
152 | 72.8k | } else { |
153 | | // If we find two mempool/extra txn that match the short id, just |
154 | | // request it. |
155 | | // This should be rare enough that the extra bandwidth doesn't matter, |
156 | | // but eating a round-trip due to FillBlock failure would be annoying |
157 | | // Note that we don't want duplication between extra_txn and mempool to |
158 | | // trigger this case, so we compare witness hashes first |
159 | 1.57k | if (txn_available[idit->second] && |
160 | 1.57k | txn_available[idit->second]->GetWitnessHash() != extra_txn[i].second->GetWitnessHash()) { |
161 | 0 | txn_available[idit->second].reset(); |
162 | 0 | mempool_count--; |
163 | 0 | extra_count--; |
164 | 0 | } |
165 | 1.57k | } |
166 | 74.4k | } |
167 | | // Though ideally we'd continue scanning for the two-txn-match-shortid case, |
168 | | // the performance win of an early exit here is too good to pass up and worth |
169 | | // the extra risk. |
170 | 30.7M | if (mempool_count == shorttxids.size()) |
171 | 188k | break; |
172 | 30.7M | } |
173 | | |
174 | 522k | LogDebug(BCLog::CMPCTBLOCK, "Initialized PartiallyDownloadedBlock for block %s using a cmpctblock of %u bytes\n", cmpctblock.header.GetHash().ToString(), GetSerializeSize(cmpctblock)); Line | Count | Source | 381 | 522k | #define LogDebug(category, ...) LogPrintLevel(category, BCLog::Level::Debug, __VA_ARGS__) Line | Count | Source | 373 | 522k | do { \ | 374 | 522k | if (LogAcceptCategory((category), (level))) { \ | 375 | 0 | bool rate_limit{level >= BCLog::Level::Info}; \ | 376 | 0 | LogPrintLevel_(category, level, rate_limit, __VA_ARGS__); \ Line | Count | Source | 350 | 0 | #define LogPrintLevel_(category, level, should_ratelimit, ...) LogPrintFormatInternal(std::source_location::current(), category, level, should_ratelimit, __VA_ARGS__) |
| 377 | 0 | } \ | 378 | 522k | } while (0) |
|
|
175 | | |
176 | 522k | return READ_STATUS_OK; |
177 | 532k | } |
178 | | |
179 | | bool PartiallyDownloadedBlock::IsTxAvailable(size_t index) const |
180 | 1.39M | { |
181 | 1.39M | if (header.IsNull()) return false0 ; |
182 | | |
183 | 1.39M | assert(index < txn_available.size()); |
184 | 1.39M | return txn_available[index] != nullptr; |
185 | 1.39M | } |
186 | | |
187 | | ReadStatus PartiallyDownloadedBlock::FillBlock(CBlock& block, const std::vector<CTransactionRef>& vtx_missing, bool segwit_active) |
188 | 244k | { |
189 | 244k | if (header.IsNull()) return READ_STATUS_INVALID131 ; |
190 | | |
191 | 244k | uint256 hash = header.GetHash(); |
192 | 244k | block = header; |
193 | 244k | block.vtx.resize(txn_available.size()); |
194 | | |
195 | 244k | unsigned int tx_missing_size = 0; |
196 | 244k | size_t tx_missing_offset = 0; |
197 | 883k | for (size_t i = 0; i < txn_available.size(); i++639k ) { |
198 | 682k | if (!txn_available[i]) { |
199 | 48.5k | if (vtx_missing.size() <= tx_missing_offset) |
200 | 43.6k | return READ_STATUS_INVALID; |
201 | 4.94k | block.vtx[i] = vtx_missing[tx_missing_offset++]; |
202 | 4.94k | tx_missing_size += block.vtx[i]->GetTotalSize(); |
203 | 4.94k | } else |
204 | 634k | block.vtx[i] = std::move(txn_available[i]); |
205 | 682k | } |
206 | | |
207 | | // Make sure we can't call FillBlock again. |
208 | 200k | header.SetNull(); |
209 | 200k | txn_available.clear(); |
210 | | |
211 | 200k | if (vtx_missing.size() != tx_missing_offset) |
212 | 593 | return READ_STATUS_INVALID; |
213 | | |
214 | | // Check for possible mutations early now that we have a seemingly good block |
215 | 200k | IsBlockMutatedFn check_mutated{m_check_block_mutated_mock ? m_check_block_mutated_mock0 : IsBlockMutated}; |
216 | 200k | if (check_mutated(/*block=*/block, |
217 | 200k | /*check_witness_root=*/segwit_active)) { |
218 | 207 | return READ_STATUS_FAILED; // Possible Short ID collision |
219 | 207 | } |
220 | | |
221 | 200k | LogDebug(BCLog::CMPCTBLOCK, "Successfully reconstructed block %s with %u txn prefilled, %u txn from mempool (incl at least %u from extra pool) and %u txn (%u bytes) requested\n", hash.ToString(), prefilled_count, mempool_count, extra_count, vtx_missing.size(), tx_missing_size); Line | Count | Source | 381 | 200k | #define LogDebug(category, ...) LogPrintLevel(category, BCLog::Level::Debug, __VA_ARGS__) Line | Count | Source | 373 | 200k | do { \ | 374 | 200k | if (LogAcceptCategory((category), (level))) { \ | 375 | 0 | bool rate_limit{level >= BCLog::Level::Info}; \ | 376 | 0 | LogPrintLevel_(category, level, rate_limit, __VA_ARGS__); \ Line | Count | Source | 350 | 0 | #define LogPrintLevel_(category, level, should_ratelimit, ...) LogPrintFormatInternal(std::source_location::current(), category, level, should_ratelimit, __VA_ARGS__) |
| 377 | 0 | } \ | 378 | 200k | } while (0) |
|
|
222 | 200k | if (vtx_missing.size() < 5) { |
223 | 200k | for (const auto& tx : vtx_missing) { |
224 | 4.14k | LogDebug(BCLog::CMPCTBLOCK, "Reconstructed block %s required tx %s\n", hash.ToString(), tx->GetHash().ToString()); Line | Count | Source | 381 | 4.14k | #define LogDebug(category, ...) LogPrintLevel(category, BCLog::Level::Debug, __VA_ARGS__) Line | Count | Source | 373 | 4.14k | do { \ | 374 | 4.14k | if (LogAcceptCategory((category), (level))) { \ | 375 | 0 | bool rate_limit{level >= BCLog::Level::Info}; \ | 376 | 0 | LogPrintLevel_(category, level, rate_limit, __VA_ARGS__); \ Line | Count | Source | 350 | 0 | #define LogPrintLevel_(category, level, should_ratelimit, ...) LogPrintFormatInternal(std::source_location::current(), category, level, should_ratelimit, __VA_ARGS__) |
| 377 | 0 | } \ | 378 | 4.14k | } while (0) |
|
|
225 | 4.14k | } |
226 | 200k | } |
227 | | |
228 | 200k | return READ_STATUS_OK; |
229 | 200k | } |