[net processing] Move tx & proof relay data to Peer
ClosedPublic
Actions

Authored by PiRK on Oct 31 2023, 12:14.

Details

Reviewers

Fabien

Group Reviewers

Restricted Project

Commits

rABCa3fde16a0c14: [net processing] Move tx & proof relay data to Peer

Summary

Partial backport of core#21160
https://github.com/bitcoin/bitcoin/pull/21160/commits/575bbd0dea6d12510fdf3220d0f0e47d969da6e9

Regarding the change in processor.cpp: AddKnownProof is now a static function in net_processing, so the call to add the local proof to the peer's known proofs has been moved there, just after SendHello.
The condition for calling this line of code is equivalent, the only possible change in behavior is that the proofid is no longer added to the known proofs if the schnorr signature fails in Processor::sendHelloInternal which would mean our sessionKey is invalid.

Depends on D14709

Test Plan

ninja all check-all bitcoin-fuzzers

Diff Detail

Repository

rABC Bitcoin ABC

Lint

Lint Passed

Unit

No Test Coverage

Build Status

Buildable 25484
Build 50549: Build Diff	lint-circular-dependencies · build-without-wallet · build-fuzzer · build-diff · build-debug · build-clang-tidy · build-clang
Build 50548: arc lint + arc unit

Event Timeline

PiRK created this revision.Oct 31 2023, 12:14

Herald added a reviewer: Restricted Project. · View Herald TranscriptOct 31 2023, 12:14

PiRK requested review of this revision.Oct 31 2023, 12:14

teamcity edited the summary of this revision. (Show Details)Oct 31 2023, 12:14

make AddKnownProof static, add fuzzer compiling to test plan

PiRK edited the summary of this revision. (Show Details)Oct 31 2023, 12:22

PiRK added a child revision: D14711: scripted-diff: rename TxRelay & TxProof members.Oct 31 2023, 12:23

Harbormaster completed remote builds in B25483: Diff 42871.Oct 31 2023, 12:33

PiRK mentioned this in D14712: net_processing: move CNode data access out of lock.Oct 31 2023, 12:36

Harbormaster completed remote builds in B25484: Diff 42872.Oct 31 2023, 12:44

Fabien requested changes to this revision.Oct 31 2023, 14:51

Fabien added a subscriber: Fabien.

Fabien added inline comments.

src/net_processing.cpp
2513	It's interesting that there is no test breaking, this is a bug: if the peer don't want you to relay txs, you should continue to the next one and not bail out. EDIT: checked on core, they're using continue
2528	dito
3924	OK