Fix mempool limiting for PrioritiseTransaction

Redo the feerate index to be based on mining score, rather than fee.

Update mempool_packages.py to test prioritisetransaction's effect on
package scores.

(cherry picked from commit bitcoin/bitcoin@eb306664e7)

Signed-off-by: Daira Emma Hopwood <daira@jacaranda.org>
This commit is contained in:
Suhas Daftuar 2015-11-19 11:18:28 -05:00 committed by Daira Emma Hopwood
parent ebc042b7bc
commit e01ad01c94
4 changed files with 77 additions and 45 deletions

View File

@ -79,17 +79,41 @@ class MempoolPackagesTest(BitcoinTestFramework):
for x in reversed(chain):
assert_equal(mempool[x]['descendantcount'], descendant_count)
descendant_fees += mempool[x]['fee']
assert_equal(mempool[x]['modifiedfee'], mempool[x]['fee'])
assert_equal(mempool[x]['descendantfees'], SATOSHIS*descendant_fees)
descendant_size += mempool[x]['size']
assert_equal(mempool[x]['descendantsize'], descendant_size)
descendant_count += 1
# Check that descendant modified fees includes fee deltas from
# prioritisetransaction
self.nodes[0].prioritisetransaction(chain[-1], 0, 1000)
mempool = self.nodes[0].getrawmempool(True)
descendant_fees = 0
for x in reversed(chain):
descendant_fees += mempool[x]['fee']
assert_equal(mempool[x]['descendantfees'], SATOSHIS*descendant_fees+1000)
# Adding one more transaction on to the chain should fail.
try:
self.chain_transaction(self.nodes[0], txid, vout, value, fee, 1)
except JSONRPCException:
print("too-long-ancestor-chain successfully rejected")
# Check that prioritising a tx before it's added to the mempool works
self.nodes[0].generate(1)
self.nodes[0].prioritisetransaction(chain[-1], 0, 2000)
self.nodes[0].invalidateblock(self.nodes[0].getbestblockhash())
mempool = self.nodes[0].getrawmempool(True)
descendant_fees = 0
for x in reversed(chain):
descendant_fees += mempool[x]['fee']
if (x == chain[-1]):
assert_equal(mempool[x]['modifiedfee'], mempool[x]['fee']+satoshi_round(0.00002))
assert_equal(mempool[x]['descendantfees'], SATOSHIS*descendant_fees+2000)
# TODO: check that node1's mempool is as expected
# TODO: test ancestor size limits

View File

@ -353,7 +353,7 @@ UniValue mempoolToJSON(bool fVerbose = false)
info.pushKV("currentpriority", e.GetPriority(chainActive.Height()));
info.pushKV("descendantcount", e.GetCountWithDescendants());
info.pushKV("descendantsize", e.GetSizeWithDescendants());
info.pushKV("descendantfees", e.GetFeesWithDescendants());
info.pushKV("descendantfees", e.GetModFeesWithDescendants());
const CTransaction& tx = e.GetTx();
set<string> setDepends;
for (const CTxIn& txin : tx.vin)
@ -411,7 +411,7 @@ UniValue getrawmempool(const UniValue& params, bool fHelp)
" \"currentpriority\" : n, (numeric) transaction priority now\n"
" \"descendantcount\" : n, (numeric) number of in-mempool descendant transactions (including this one)\n"
" \"descendantsize\" : n, (numeric) size of in-mempool descendants (including this one)\n"
" \"descendantfees\" : n, (numeric) fees of in-mempool descendants (including this one)\n"
" \"descendantfees\" : n, (numeric) modified fees (see \"modifiedfee\" above) of in-mempool descendants (including this one)\n"
" \"depends\" : [ (array) unconfirmed transactions used as inputs for this transaction\n"
" \"transactionid\", (string) parent transaction id\n"
" ... ]\n"

View File

@ -36,7 +36,7 @@ CTxMemPoolEntry::CTxMemPoolEntry(const CTransaction& _tx, const CAmount& _nFee,
nCountWithDescendants = 1;
nSizeWithDescendants = nTxSize;
nFeesWithDescendants = nFee;
nModFeesWithDescendants = nFee;
feeDelta = 0;
}
@ -57,6 +57,7 @@ CTxMemPoolEntry::GetPriority(unsigned int currentHeight) const
void CTxMemPoolEntry::UpdateFeeDelta(int64_t newFeeDelta)
{
nModFeesWithDescendants += newFeeDelta - feeDelta;
feeDelta = newFeeDelta;
}
@ -114,7 +115,7 @@ bool CTxMemPool::UpdateForDescendants(txiter updateIt, int maxDescendantsToVisit
for (txiter cit : setAllDescendants) {
if (!setExclude.count(cit->GetTx().GetHash())) {
modifySize += cit->GetTxSize();
modifyFee += cit->GetFee();
modifyFee += cit->GetModifiedFee();
modifyCount++;
cachedDescendants[updateIt].insert(cit);
}
@ -244,7 +245,7 @@ void CTxMemPool::UpdateAncestorsOf(bool add, txiter it, setEntries &setAncestors
}
const int64_t updateCount = (add ? 1 : -1);
const int64_t updateSize = updateCount * it->GetTxSize();
const CAmount updateFee = updateCount * it->GetFee();
const CAmount updateFee = updateCount * it->GetModifiedFee();
for (txiter ancestorIt : setAncestors) {
mapTx.modify(ancestorIt, update_descendant_state(updateSize, updateFee, updateCount));
}
@ -304,7 +305,7 @@ void CTxMemPoolEntry::SetDirty()
{
nCountWithDescendants = 0;
nSizeWithDescendants = nTxSize;
nFeesWithDescendants = nFee;
nModFeesWithDescendants = GetModifiedFee();
}
void CTxMemPoolEntry::UpdateState(int64_t modifySize, CAmount modifyFee, int64_t modifyCount)
@ -312,8 +313,7 @@ void CTxMemPoolEntry::UpdateState(int64_t modifySize, CAmount modifyFee, int64_t
if (!IsDirty()) {
nSizeWithDescendants += modifySize;
assert(int64_t(nSizeWithDescendants) > 0);
nFeesWithDescendants += modifyFee;
assert(nFeesWithDescendants >= 0);
nModFeesWithDescendants += modifyFee;
nCountWithDescendants += modifyCount;
assert(int64_t(nCountWithDescendants) > 0);
}
@ -376,6 +376,17 @@ bool CTxMemPool::addUnchecked(const uint256& hash, const CTxMemPoolEntry &entry,
indexed_transaction_set::iterator newit = mapTx.insert(entry).first;
mapLinks.insert(make_pair(newit, TxLinks()));
// Update transaction for any feeDelta created by PrioritiseTransaction
// TODO: refactor so that the fee delta is calculated before inserting
// into mapTx.
std::map<uint256, std::pair<double, CAmount> >::const_iterator pos = mapDeltas.find(hash);
if (pos != mapDeltas.end()) {
const std::pair<double, CAmount> &deltas = pos->second;
if (deltas.second) {
mapTx.modify(newit, update_fee_delta(deltas.second));
}
}
// Update cachedInnerUsage to include contained transaction's usage.
// (When we update the entry for in-mempool parents, memory usage will be
// further updated.)
@ -417,15 +428,6 @@ bool CTxMemPool::addUnchecked(const uint256& hash, const CTxMemPoolEntry &entry,
mapOrchardNullifiers[orchardNullifier] = &tx;
}
// Update transaction's score for any feeDelta created by PrioritiseTransaction
std::map<uint256, std::pair<double, CAmount> >::const_iterator pos = mapDeltas.find(hash);
if (pos != mapDeltas.end()) {
const std::pair<double, CAmount> &deltas = pos->second;
if (deltas.second) {
mapTx.modify(newit, update_fee_delta(deltas.second));
}
}
nTransactionsUpdated++;
totalTxSize += entry.GetTxSize();
minerPolicyEstimator->processTransaction(entry, fCurrentEstimate);
@ -908,27 +910,24 @@ void CTxMemPool::check(const CCoinsViewCache *pcoins) const
CTxMemPool::setEntries setChildrenCheck;
std::map<COutPoint, CInPoint>::const_iterator iter = mapNextTx.lower_bound(COutPoint(it->GetTx().GetHash(), 0));
int64_t childSizes = 0;
CAmount childFees = 0;
CAmount childModFee = 0;
for (; iter != mapNextTx.end() && iter->first.hash == it->GetTx().GetHash(); ++iter) {
txiter childit = mapTx.find(iter->second.ptx->GetHash());
assert(childit != mapTx.end()); // mapNextTx points to in-mempool transactions
if (setChildrenCheck.insert(childit).second) {
childSizes += childit->GetTxSize();
childFees += childit->GetFee();
childModFee += childit->GetModifiedFee();
}
}
assert(setChildrenCheck == GetMemPoolChildren(it));
// Also check to make sure size/fees is greater than sum with immediate children.
// Also check to make sure size is greater than sum with immediate children.
// just a sanity check, not definitive that this calc is correct...
// also check that the size is less than the size of the entire mempool.
if (!it->IsDirty()) {
assert(it->GetSizeWithDescendants() >= childSizes + it->GetTxSize());
assert(it->GetFeesWithDescendants() >= childFees + it->GetFee());
} else {
assert(it->GetSizeWithDescendants() == it->GetTxSize());
assert(it->GetFeesWithDescendants() == it->GetFee());
assert(it->GetModFeesWithDescendants() == it->GetModifiedFee());
}
assert(it->GetFeesWithDescendants() >= 0);
if (fDependsWait)
waitingOnDependants.push_back(&(*it));
@ -1147,6 +1146,14 @@ void CTxMemPool::PrioritiseTransaction(const uint256 hash, const std::string str
txiter it = mapTx.find(hash);
if (it != mapTx.end()) {
mapTx.modify(it, update_fee_delta(deltas.second));
// Now update all ancestors' modified fees with descendants
setEntries setAncestors;
uint64_t nNoLimit = std::numeric_limits<uint64_t>::max();
std::string dummy;
CalculateMemPoolAncestors(*it, setAncestors, nNoLimit, nNoLimit, nNoLimit, nNoLimit, dummy, false);
BOOST_FOREACH(txiter ancestorIt, setAncestors) {
mapTx.modify(ancestorIt, update_descendant_state(0, nFeeDelta, 0));
}
}
}
LogPrintf("PrioritiseTransaction: %s priority += %f, fee += %d\n", strHash, dPriorityDelta, FormatMoney(nFeeDelta));

View File

@ -56,12 +56,12 @@ class CTxMemPool;
* ("descendant" transactions).
*
* When a new entry is added to the mempool, we update the descendant state
* (nCountWithDescendants, nSizeWithDescendants, and nFeesWithDescendants) for
* (nCountWithDescendants, nSizeWithDescendants, and nModFeesWithDescendants) for
* all ancestors of the newly added transaction.
*
* If updating the descendant state is skipped, we can mark the entry as
* "dirty", and set nSizeWithDescendants/nFeesWithDescendants to equal nTxSize/
* nTxFee. (This can potentially happen during a reorg, where we limit the
* "dirty", and set nSizeWithDescendants/nModFeesWithDescendants to equal nTxSize/
* nFee+feeDelta. (This can potentially happen during a reorg, where we limit the
* amount of work we're willing to do to avoid consuming too much CPU.)
*
*/
@ -86,11 +86,11 @@ private:
// Information about descendants of this transaction that are in the
// mempool; if we remove this transaction we must remove all of these
// descendants as well. if nCountWithDescendants is 0, treat this entry as
// dirty, and nSizeWithDescendants and nFeesWithDescendants will not be
// dirty, and nSizeWithDescendants and nModFeesWithDescendants will not be
// correct.
uint64_t nCountWithDescendants; //! number of descendant transactions
uint64_t nSizeWithDescendants; //! ... and size
CAmount nFeesWithDescendants; //! ... and total fees (all including us)
CAmount nModFeesWithDescendants; //! ... and total fees (all including us)
public:
CTxMemPoolEntry(const CTransaction& _tx, const CAmount& _nFee,
@ -113,7 +113,8 @@ public:
// Adjusts the descendant state, if this entry is not dirty.
void UpdateState(int64_t modifySize, CAmount modifyFee, int64_t modifyCount);
// Updates the fee delta used for mining priority score
// Updates the fee delta used for mining priority score, and the
// modified fees with descendants.
void UpdateFeeDelta(int64_t feeDelta);
/** We can set the entry to be dirty if doing the full calculation of in-
@ -125,7 +126,7 @@ public:
uint64_t GetCountWithDescendants() const { return nCountWithDescendants; }
uint64_t GetSizeWithDescendants() const { return nSizeWithDescendants; }
CAmount GetFeesWithDescendants() const { return nFeesWithDescendants; }
CAmount GetModFeesWithDescendants() const { return nModFeesWithDescendants; }
bool GetSpendsCoinbase() const { return spendsCoinbase; }
uint32_t GetValidatedBranchId() const { return nBranchId; }
@ -173,27 +174,27 @@ struct mempoolentry_txid
}
};
/** \class CompareTxMemPoolEntryByFee
/** \class CompareTxMemPoolEntryByDescendantScore
*
* Sort an entry by max(feerate of entry's tx, feerate with all descendants).
* Sort an entry by max(score/size of entry's tx, score/size with all descendants).
*/
class CompareTxMemPoolEntryByFee
class CompareTxMemPoolEntryByDescendantScore
{
public:
bool operator()(const CTxMemPoolEntry& a, const CTxMemPoolEntry& b) const
{
bool fUseADescendants = UseDescendantFeeRate(a);
bool fUseBDescendants = UseDescendantFeeRate(b);
bool fUseADescendants = UseDescendantScore(a);
bool fUseBDescendants = UseDescendantScore(b);
double aFees = fUseADescendants ? a.GetFeesWithDescendants() : a.GetFee();
double aModFee = fUseADescendants ? a.GetModFeesWithDescendants() : a.GetModifiedFee();
double aSize = fUseADescendants ? a.GetSizeWithDescendants() : a.GetTxSize();
double bFees = fUseBDescendants ? b.GetFeesWithDescendants() : b.GetFee();
double bModFee = fUseBDescendants ? b.GetModFeesWithDescendants() : b.GetModifiedFee();
double bSize = fUseBDescendants ? b.GetSizeWithDescendants() : b.GetTxSize();
// Avoid division by rewriting (a/b > c/d) as (a*d > c*b).
double f1 = aFees * bSize;
double f2 = aSize * bFees;
double f1 = aModFee * bSize;
double f2 = aSize * bModFee;
if (f1 == f2) {
return a.GetTime() >= b.GetTime();
@ -201,11 +202,11 @@ public:
return f1 < f2;
}
// Calculate which feerate to use for an entry (avoiding division).
bool UseDescendantFeeRate(const CTxMemPoolEntry &a) const
// Calculate which score to use for an entry (avoiding division).
bool UseDescendantScore(const CTxMemPoolEntry &a) const
{
double f1 = (double)a.GetFee() * a.GetSizeWithDescendants();
double f2 = (double)a.GetFeesWithDescendants() * a.GetTxSize();
double f1 = (double)a.GetModifiedFee() * a.GetSizeWithDescendants();
double f2 = (double)a.GetModFeesWithDescendants() * a.GetTxSize();
return f2 > f1;
}
};
@ -382,7 +383,7 @@ public:
boost::multi_index::ordered_non_unique<
boost::multi_index::tag<descendant_score>,
boost::multi_index::identity<CTxMemPoolEntry>,
CompareTxMemPoolEntryByFee
CompareTxMemPoolEntryByDescendantScore
>,
// sorted by score (for mining prioritization)
boost::multi_index::ordered_unique<