Merge bitcoin/bitcoin#33629: Cluster mempool
17cf9ff7efUse cluster size limit for -maxmempool bound, and allow -maxmempool=0 in general (Suhas Daftuar)315e43e5d8Sanity check `GetFeerateDiagram()` in CTxMemPool::check() (Suhas Daftuar)de2e9a24c4test: extend package rbf functional test to larger clusters (Suhas Daftuar)4ef4ddb504doc: update policy/packages.md for new package acceptance logic (Suhas Daftuar)79f73ad713Add check that GetSortedScoreWithTopology() agrees with CompareMiningScoreWithTopology() (Suhas Daftuar)a86ac11768Update comments for CTxMemPool class (Suhas Daftuar)9567eaa66dInvoke TxGraph::DoWork() at appropriate times (Suhas Daftuar)6c5c44f774test: add functional test for new cluster mempool RPCs (Suhas Daftuar)72f60c877edoc: Update mempool_replacements.md to reflect feerate diagram checks (Suhas Daftuar)21693f031aExpose cluster information via rpc (Suhas Daftuar)72e74e0d42fuzz: try to add more code coverage for mempool fuzzing (Suhas Daftuar)f107417490bench: add more mempool benchmarks (Suhas Daftuar)7976eb1ae7Avoid violating mempool policy limits in tests (Suhas Daftuar)84de685cf7Stop tracking parents/children outside of txgraph (Suhas Daftuar)88672e205bRewrite GatherClusters to use the txgraph implementation (Suhas Daftuar)1ca4f01090Fix miniminer_tests to work with cluster limits (Suhas Daftuar)1902111e0fEliminate CheckPackageLimits, which no longer does anything (Suhas Daftuar)3a646ec462Rework RBF and TRUC validation (Suhas Daftuar)19b8479868Make getting parents/children a function of the mempool, not a mempool entry (Suhas Daftuar)5560913e51Rework truc_policy to use descendants, not children (Suhas Daftuar)a4458d6c40Use txgraph to calculate descendants (Suhas Daftuar)c8b6f70d64Use txgraph to calculate ancestors (Suhas Daftuar)241a3e666bSimplify ancestor calculation functions (Suhas Daftuar)b9cec7f0a1Make removeConflicts private (Suhas Daftuar)0402e6c780Remove unused limits from CalculateMemPoolAncestors (Suhas Daftuar)08be765ac2Remove mempool logic designed to maintain ancestor/descendant state (Suhas Daftuar)fc4e3e6bc1Remove unused members from CTxMemPoolEntry (Suhas Daftuar)ff3b398d12mempool: eliminate accessors to mempool entry ancestor/descendant cached state (Suhas Daftuar)b9a2039f51Eliminate use of cached ancestor data in miniminer_tests and truc_policy (Suhas Daftuar)ba09fc9774mempool: Remove unused function CalculateDescendantMaximum (Suhas Daftuar)8e49477e86wallet: Replace max descendant count with cluster_count (Suhas Daftuar)e031085fd4Eliminate Single-Conflict RBF Carve Out (Suhas Daftuar)cf3ab8e1d0Stop enforcing descendant size/count limits (Suhas Daftuar)89ae38f489test: remove rbf carveout test from mempool_limit.py (Suhas Daftuar)c0bd04d18fCalculate descendant information for mempool RPC output on-the-fly (Suhas Daftuar)bdcefb8a8bUse mempool/txgraph to determine if a tx has descendants (Suhas Daftuar)69e1eaa6edAdd test case for cluster size limits to TRUC logic (Suhas Daftuar)9cda64b86cStop enforcing ancestor size/count limits (Suhas Daftuar)1f93227a84Remove dependency on cached ancestor data in mini-miner (Suhas Daftuar)9fbe0a4ac2rpc: Calculate ancestor data from scratch for mempool rpc calls (Suhas Daftuar)7961496ddaReimplement GetTransactionAncestry() to not rely on cached data (Suhas Daftuar)feceaa42e8Remove CTxMemPool::GetSortedDepthAndScore (Suhas Daftuar)21b5cea588Use cluster linearization for transaction relay sort order (Suhas Daftuar)6445aa7d97Remove the ancestor and descendant indices from the mempool (Suhas Daftuar)216e693729Implement new RBF logic for cluster mempool (Suhas Daftuar)ff8f115decpolicy: Remove CPFP carveout rule (Suhas Daftuar)c3f1afc934test: rewrite PopulateMempool to not violate mempool policy (cluster size) limits (Suhas Daftuar)47ab32fdb1Select transactions for blocks based on chunk feerate (Suhas Daftuar)dec138d1ddfuzz: remove comparison between mini_miner block construction and miner (Suhas Daftuar)6c2bceb200bench: rewrite ComplexMemPool to not create oversized clusters (Suhas Daftuar)1ad4590f63Limit mempool size based on chunk feerate (Suhas Daftuar)b11c89cab2Rework miner_tests to not require large cluster limit (Suhas Daftuar)95a8297d48Check cluster limits when using -walletrejectlongchains (Suhas Daftuar)95762e6759Do not allow mempool clusters to exceed configured limits (Suhas Daftuar)edb3e7cdf6[test] rework/delete feature_rbf tests requiring large clusters (glozow)435fd56711test: update feature_rbf.py replacement test (Suhas Daftuar)34e32985e8Add new (unused) limits for cluster size/count (Suhas Daftuar)838d7e3553Add transactions to txgraph, but without cluster dependencies (Suhas Daftuar)d5ed9cb3ebAdd accessor for sigops-adjusted weight (Suhas Daftuar)1bf3b51396Add sigops adjusted weight calculator (Suhas Daftuar)c18c68a950Create a txgraph inside CTxMemPool (Suhas Daftuar)29a94d5b2fMake CTxMemPoolEntry derive from TxGraph::Ref (Suhas Daftuar)92b0079fe3Allow moving CTxMemPoolEntry objects, disallow copying (Suhas Daftuar)6c73e47448mempool: Store iterators into mapTx in mapNextTx (Suhas Daftuar)51430680ecAllow moving an Epoch::Marker (Suhas Daftuar) Pull request description: [Reopening #28676 here as a new PR, because GitHub is slow to load the page making it hard to scroll through and see comments. Also, that PR was originally opened with a prototype implementation which has changed significantly with the introduction of `TxGraph`.] This is an implementation of the [cluster mempool proposal](https://delvingbitcoin.org/t/an-overview-of-the-cluster-mempool-proposal/393). This branch implements the following observable behavior changes: - Maintains a partitioning of the mempool into connected clusters (via the `txgraph` class), which are limited in vsize to 101 kvB by default, and limited in count to 64 by default. - Each cluster is sorted ("linearized") to try to optimize for selecting highest-feerate-subsets of a cluster first - Transaction selection for mining is updated to use the cluster linearizations, selecting highest feerate "chunks" first for inclusion in a block template. - Mempool eviction is updated to use the cluster linearizations, selecting lowest feerate "chunks" first for removal. - The RBF rules are updated to: (a) drop the requirement that no new inputs are introduced; (b) change the feerate requirement to instead check that the feerate diagram of the mempool will strictly improve; (c) replace the direct conflicts limit with a directly-conflicting-clusters limit. - The CPFP carveout rule is eliminated (it doesn't make sense in a cluster-limited mempool) - The ancestor and descendant limits are no longer enforced. - New cluster count/cluster vsize limits are now enforced instead. - Transaction relay now uses chunk feerate comparisons to determine the order that newly received transactions are announced to peers. Additionally, the cached ancestor and descendant data are dropped from the mempool, along with the multi_index indices that were maintained to sort the mempool by ancestor and descendant feerates. For compatibility (eg with wallet behavior or RPCs exposing this), this information is now calculated dynamically instead. ACKs for top commit: instagibbs: reACK17cf9ff7efglozow: reACK17cf9ff7efsipa: ACK17cf9ff7efTree-SHA512: bbde46d913d56f8d9c0426cb0a6c4fa80b01b0a4c2299500769921f886082fb4f51f1694e0ee1bc318c52e1976d7ebed8134a64eda0b8044f3a708c04938eee7
This commit is contained in:
61 files changed
+1864
-3220
No files matched your search
@@ -12,12 +12,7 @@ other consensus and policy rules, each of the following conditions are met:
|
||||
|
||||
1. (Removed)
|
||||
|
||||
2. The replacement transaction only include an unconfirmed input if that input was included in
|
||||
one of the directly conflicting transactions. An unconfirmed input spends an output from a
|
||||
currently-unconfirmed transaction.
|
||||
|
||||
*Rationale*: When RBF was originally implemented, the mempool did not keep track of
|
||||
ancestor feerates yet. This rule was suggested as a temporary restriction.
|
||||
2. (Removed)
|
||||
|
||||
3. The replacement transaction pays an absolute fee of at least the sum paid by the original
|
||||
transactions.
|
||||
@@ -38,23 +33,16 @@ other consensus and policy rules, each of the following conditions are met:
|
||||
*Rationale*: Try to prevent DoS attacks where an attacker causes the network to repeatedly relay
|
||||
transactions each paying a tiny additional amount in fees, e.g. just 1 satoshi.
|
||||
|
||||
5. The number of original transactions does not exceed 100. More precisely, the sum of all
|
||||
directly conflicting transactions' descendant counts (number of transactions inclusive of itself
|
||||
and its descendants) must not exceed 100; it is possible that this overestimates the true number
|
||||
of original transactions.
|
||||
5. The number of distinct clusters corresponding to conflicting transactions does not exceed 100.
|
||||
|
||||
*Rationale*: Try to prevent DoS attacks where an attacker is able to easily occupy and flush out
|
||||
significant portions of the node's mempool using replacements with multiple directly conflicting
|
||||
transactions, each with large descendant sets.
|
||||
*Rationale*: Limit CPU usage required to update the mempool for so many transactions being
|
||||
removed at once.
|
||||
|
||||
6. The replacement transaction's feerate is greater than the feerates of all directly conflicting
|
||||
transactions.
|
||||
6. The feerate diagram of the mempool must be strictly improved by the replacement transaction.
|
||||
|
||||
*Rationale*: This rule was originally intended to ensure that the replacement transaction is
|
||||
preferable for block-inclusion, compared to what would be removed from the mempool. This rule
|
||||
predates ancestor feerate-based transaction selection.
|
||||
*Rationale*: This ensures that block fees in all future blocks will go up
|
||||
after the replacement (ignoring tail effects at the end of a block).
|
||||
|
||||
This set of rules is similar but distinct from BIP125.
|
||||
|
||||
## History
|
||||
|
||||
@@ -79,3 +67,5 @@ This set of rules is similar but distinct from BIP125.
|
||||
* Signaling for replace-by-fee is no longer required as of [PR 30592](https://github.com/bitcoin/bitcoin/pull/30592).
|
||||
|
||||
* The incremental relay feerate default is 0.1sat/vB ([PR #33106](https://github.com/bitcoin/bitcoin/pull/33106)).
|
||||
|
||||
* Feerate diagram policy enabled in conjunction with switch to cluster mempool as of **v31.0**.
|
||||
+1
-15
@@ -38,9 +38,7 @@ The following rules are enforced for all packages:
|
||||
|
||||
- Packages are 1-parent-1-child, with no in-mempool ancestors of the package.
|
||||
|
||||
- All conflicting clusters (connected components of mempool transactions) must be clusters of up to size 2.
|
||||
|
||||
- No more than MAX_REPLACEMENT_CANDIDATES transactions can be replaced, analogous to
|
||||
- The number of distinct clusters containing conflicting transactions can be no more than 100, analogous to
|
||||
regular [replacement rule](./mempool-replacements.md) 5).
|
||||
|
||||
- Replacements must pay more total fees at the incremental relay fee (analogous to
|
||||
@@ -56,18 +54,6 @@ The following rules are enforced for all packages:
|
||||
result in more robust fee bumping. More general package RBF may be
|
||||
enabled in the future.
|
||||
|
||||
* When packages are evaluated against ancestor/descendant limits, the union of all transactions'
|
||||
descendants and ancestors is considered. (#21800)
|
||||
|
||||
- *Rationale*: This is essentially a "worst case" heuristic intended for packages that are
|
||||
heavily connected, i.e. some transaction in the package is the ancestor or descendant of all
|
||||
the other transactions.
|
||||
|
||||
* [CPFP Carve Out](./mempool-limits.md#CPFP-Carve-Out) is disabled in packaged contexts. (#21800)
|
||||
|
||||
- *Rationale*: This carve out cannot be accurately applied when there are multiple transactions'
|
||||
ancestors and descendants being considered at the same time.
|
||||
|
||||
The following rules are only enforced for packages to be submitted to the mempool (not
|
||||
enforced for test accepts):
|
||||
|
||||
|
||||
Reference in new issue
Block a user