Skip to content
Merged
Changes from 1 commit
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
19 changes: 18 additions & 1 deletion src/llmq/utils.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -801,13 +801,28 @@ bool EnsureQuorumConnections(const Consensus::LLMQParams& llmqParams, CConnman&
util_params.m_base_index->GetBlockHash().ToString());

Uint256HashSet connections;
Uint256HashSet relayMembers;
if (isMember) {
connections = GetQuorumConnections(llmqParams, sporkman, util_params, myProTxHash, /*onlyOutbound=*/true);
// If all-members-connected is enabled for this quorum type, leverage the full-mesh
// connections for low-latency recovered sig propagation by treating all members as
// relay members (instead of the ring-based subset). This ensures peers will send
// QSENDRECSIGS to each other across the full mesh and set m_wants_recsigs widely.
if (IsAllMembersConnectedEnabled(llmqParams.type, sporkman)) {
for (const auto& dmn : members) {
if (dmn->proTxHash != myProTxHash) {
relayMembers.emplace(dmn->proTxHash);
}
}
} else {
relayMembers = GetQuorumRelayMembers(llmqParams, util_params, myProTxHash, true);
}
} else {
auto cindexes = CalcDeterministicWatchConnections(llmqParams.type, util_params.m_base_index, members.size(), 1);
for (auto idx : cindexes) {
connections.emplace(members[idx]->proTxHash);
}
relayMembers = connections;
}
Comment thread
knst marked this conversation as resolved.
Comment on lines 803 to 826
Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💬 Nitpick: Document why connections and relayMembers must differ

The detailed rationale for keeping connections (outbound-only half-mesh) separate from relayMembers (every other quorum member, used to drive QSENDRECSIGS handshakes bidirectionally) lives only in the revert's commit message. The existing comment at lines 807-810 explains the all-members-connected branch but not the broader invariant. A short in-source comment near line 803 noting that the two sets are intentionally distinct — and why conflating them breaks recovered-sig propagation on the inbound side of each pair — would make the trap visible to future readers and reduce the chance of the same simplification being attempted again.

source: ['claude']

if (!connections.empty()) {
if (!connman.HasMasternodeQuorumNodes(llmqParams.type, util_params.m_base_index->GetBlockHash()) &&
Expand All @@ -826,7 +841,9 @@ bool EnsureQuorumConnections(const Consensus::LLMQParams& llmqParams, CConnman&
LogPrint(BCLog::NET_NETCONN, debugMsg.c_str()); /* Continued */
}
connman.SetMasternodeQuorumNodes(llmqParams.type, util_params.m_base_index->GetBlockHash(), connections);
connman.SetMasternodeQuorumRelayMembers(llmqParams.type, util_params.m_base_index->GetBlockHash(), connections);
}
if (!relayMembers.empty()) {
connman.SetMasternodeQuorumRelayMembers(llmqParams.type, util_params.m_base_index->GetBlockHash(), relayMembers);
}
return true;
}
Expand Down
Loading