Duty logic validation for consensus messages #228

diegomrsantos · 2025-04-03T21:23:20Z

Issue Addressed

#161

Proposed Changes

This PR adds duty validation logic for SSV messages and integrates a new duties management module.

Introduced a new duties module and duties_tracker for validating sync committee and proposer duties messages.

Additional Info

This is pending #277 and will be implemented in a future PR.

Copilot

Pull Request Overview

This PR refactors duty logic validations by introducing new parameters and modules to support BeaconNetwork and Duties tracking while updating the Validator to use a DutiesProvider.

Introduces the new slots_per_epoch field into validation contexts and tests.
Refactors the Validator to take a DutiesProvider along with a BeaconNetwork.
Adds new modules for duties tracking and beacon network functionalities.

Reviewed Changes

Copilot reviewed 14 out of 15 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
anchor/message_validator/src/partial_signature.rs	Added slots_per_epoch field in test contexts for partial signature validation.
anchor/message_validator/src/lib.rs	Updated Validator struct and validation functions to use BeaconNetwork and DutiesProvider.
anchor/message_validator/src/consensus_message.rs	Updated consensus message validation to utilize the new duty logic and error variants.
anchor/message_validator/src/duties/*	New modules that implement duties tracking and the DutiesProvider trait.
Other files	Updates updating generics and dependencies across message sender/receiver and client to support the new duty logic features.

Files not reviewed (1)

.githooks/pre-commit: Language not supported

Comments suppressed due to low confidence (2)

anchor/message_validator/src/partial_signature.rs:237

[nitpick] Consider replacing the hard-coded magic number '32' with a named constant to improve clarity and maintainability in test configurations.

+            slots_per_epoch: 32,

anchor/message_validator/src/consensus_message.rs:51

[nitpick] Cloning the entire slot clock might incur unnecessary performance overhead if the underlying implementation is not lightweight; consider passing a reference or ensuring that the SlotClock type implements a cheap clone.

beacon_network.slot_clock().clone(),

anchor/message_validator/src/duties/duties_tracker.rs

# Conflicts: # anchor/client/src/lib.rs # anchor/message_receiver/src/manager.rs

Copilot

Pull Request Overview

This PR refactors the duty validation logic for SSV messages to improve clarity and correctness, while also integrating a new duties management module.

Updated validation functions to accept a generic SlotClock and DutiesProvider.
Introduced a new duties module and duties_tracker for handling sync committee and proposer duties.
Modified the Validator and related methods across message validation, sending, and receiving to propagate the new duties provider.

Reviewed Changes

Copilot reviewed 13 out of 13 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
anchor/message_validator/src/partial_signature.rs	Updated validation context to include a generic SlotClock; minor adjustments for new duty fields.
anchor/message_validator/src/lib.rs	Modified ValidationContext and Validator to include DutiesProvider; updated message validation functions accordingly.
anchor/message_validator/src/duties/mod.rs	Introduced new duties module with types and traits for duty management.
anchor/message_validator/src/duties/duties_tracker.rs	Added duties tracker to poll sync committee and proposer duties from beacon nodes.
anchor/message_validator/src/consensus_message.rs	Integrated duty-based validation logic within consensus messages; adjusted round and timing validations.
Other files (message_sender, message_receiver, database, client, etc.)	Updated generics and dependencies to accommodate the new duties provider integration.

Comments suppressed due to low confidence (1)

anchor/message_validator/src/lib.rs:302

Verify that substituting slot_clock with duties_provider in the call to validate_consensus_message fully preserves the original timing logic and that duties_provider properly encapsulates the necessary timing behavior.

self.duties_provider.clone(),

anchor/message_validator/src/consensus_message.rs

diegomrsantos · 2025-04-18T11:42:20Z

anchor/message_validator/src/consensus_message.rs

+    // - duty's starting slot + 34 (committee and aggregation)
+    // - duty's starting slot + 3 (other types)


@dknopik, any idea why these numbers?

The committee duty is needed for attestation, which can be submitted on chain until the end of the next epoch. So a large window for message validity is a good idea here. Same reasoning for attestation aggregation.

diegomrsantos · 2025-04-18T13:50:32Z

anchor/message_validator/src/consensus_message.rs

+            let validator_index_count = validator_indices.len() as u64;
+            let slots_per_epoch_val = validation_context.slots_per_epoch;
+
+            // Skip duty search if validators * 2 exceeds slots per epoch


@Zacholme7 @dknopik, does this make sense? If the number of validators in the committee is >= slots per epoch, it doesn't mean one of them is in the sync committee. Additionally, do they use the Committee role for the sync committee? What's the difference to SyncCommittee?

The Committee Role is used for BeaconVote data, which includes the current head. This is needed for attestations AND sync committee contributions.

The word "Committee" in the "Committee role" refers to the SSV committee, not the sync committee, as this role is executed once per SSV committee, not once per validator.

As for the check:

The limit is at most slots_per_epoch because we decide a BeaconVote at most once per slot.
We need to vote for a BeaconVote if

at least one validator is in the sync committee, or

at least one validator attests in this slot

Furthermore, validators attest once per epoch. So, if we have 32 validators, it is possible that we have to perform the Committee role every slot - regardless of whether one of them is in the sync committee.

That being said, I am not sure why we multiply/divide by two.

dknopik · 2025-04-22T08:43:21Z

anchor/message_validator/src/consensus_message.rs

+    slot: Slot,
+    validator_indices: &[ValidatorIndex],
+    duty_provider: Arc<impl DutiesProvider>,
+) -> Result<(u64, bool), ValidationFailure> {


Why do we use a (u64, bool) instead of a Option<u64>?

diegomrsantos · 2025-04-22T17:00:10Z

anchor/message_validator/src/consensus_message.rs

+                    }
+                }
+            }
+            Ok(Some(std::cmp::min(


@@nkryuchkov given the if at line 580, it seems when we get here, 2 * validator_index_count >= slots_per_epoch_val, so we could simply return slots_per_epoch_val?

When validator_index_count < slots_per_epoch_val / 2, the code inside the if returns only if one of the indices satisfies the duty_provider.is_validator_in_sync_committee(period, index) condition, so it can also go out of the if block and reach the last line. To get rid of min, we'd need to add an extra check:

// Skip duty search if validators * 2 exceeds slots per epoch, // as the maximum duties per epoch is capped at the number of slots. // This avoids unnecessary checks. if validatorIndexCount*2 >= slotsPerEpoch { return slotsPerEpoch, true } // Check if there is at least one validator in the sync committee. // If so, the duty limit is equal to the number of slots per epoch. period := mv.netCfg.Beacon.EstimatedSyncCommitteePeriodAtEpoch(mv.netCfg.Beacon.EstimatedEpochAtSlot(slot)) for _, i := range validatorIndices { if mv.dutyStore.SyncCommittee.Duty(period, i) != nil { return slotsPerEpoch, true } } return 2 * validatorIndexCount, true

Good catch, I like your solution more. But do we need to skip duty_provider.is_validator_in_sync_committee(period, index) tho? Isn't it just a map get?

@diegomrsantos I'm not sure I understand, what do you mean by skipping? Yes, it's just a map get

is the code below necessary?

if validatorIndexCount*2 >= slotsPerEpoch { return slotsPerEpoch, true }

Why the *2 in the condition validator_index_count * 2 >= slots_per_epoch_val? If we want to skip the check, shouldn't it be validator_index_count >= slots_per_epoch_val?

@diegomrsantos sorry for the delay. The *2 is because there might be 2 duties per validator. Do I understand correctly that your question is only about the *2 part?

@nkryuchkov I assume you mean sync committee and attestation duty. But isn't the BeaconData only voted on once per committee regardless?

@dknopik IIRC the main reason was reorgs. @MatheusFranco99 can you please confirm it? If one of validators participates in sync committee, the duty limit changes to slotsPerEpoch

@nkryuchkov @dknopik
Yes, the reason is reorgs. We allow up to 2 attestation duties per epoch, i.e. 2*validatorIndexCount. For example, a committee with a single validator is allowed to send messages for up to 2 duties.
However, for the case validatorIndexCount*2 >= slotsPerEpoch, it's meaningless to set the maximum number of committee duties to more than 32, so we cap it to 32.

The only exception is when one of the committee's validators is in the sync committee period. In this case, the committee can send msgs for up to 32 duties, i.e. all slots.

@diegomrsantos Answering your question, yes, the following code is equivalent:

period := mv.netCfg.Beacon.EstimatedSyncCommitteePeriodAtEpoch(mv.netCfg.Beacon.EstimatedEpochAtSlot(slot)) for _, i := range validatorIndices { if mv.dutyStore.SyncCommittee.Duty(period, i) != nil { return slotsPerEpoch, true } } return min(validatorIndexCount*2, slotsPerEpoch), true

We added the IF check before mv.dutyStore.SyncCommittee.Duty is called because:

it's very likely that validatorIndexCount*2 >= slotsPerEpoch, immediately returning.

and it's very unlikely that any mv.dutyStore.SyncCommittee.Duty will return true, so the FOR loop complexity is expected to be len(validatorIndices)

Copilot

Pull Request Overview

This PR introduces duty validation logic for SSV messages and integrates a new duties management module into the message validation and networking components. The key changes include:

The addition of a new duties module and a duties_tracker to manage sync committee and proposer duties.
Updates to the Validator and related message validation functions to accept a generic DutiesProvider, along with corresponding test updates.
Modifications in networking and message receiver code to propagate the new generic parameters and improve logging.

Reviewed Changes

Copilot reviewed 13 out of 13 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
anchor/message_validator/src/partial_signature.rs	Updated function signatures and tests to use a generic SlotClock type.
anchor/message_validator/src/lib.rs	Integrated DutiesProvider into Validator and refactored validation logic.
anchor/message_validator/src/duties/mod.rs	Added the new duties module for sync committee and proposer duties.
anchor/message_validator/src/duties/duties_tracker.rs	Introduced duties_tracker with polling and pruning of duties.
anchor/message_validator/src/consensus_state.rs	Added getters for max_slot and duty counts for operators.
anchor/message_validator/src/consensus_message.rs	Updated consensus message validation to incorporate duty-based checks.
anchor/message_sender/src/network.rs	Updated network sender to use the new generic Validator with DutiesProvider.
anchor/message_receiver/src/manager.rs	Updated message receiver to use generic Validator and improved logging.
anchor/database/src/state.rs	Added a helper to collect validator indices.
anchor/common/ssv_types/src/cluster.rs	Introduced conversion from ValidatorIndex to u64.
anchor/client/src/lib.rs	Updated client startup to initialize and start the new DutiesTracker.

anchor/message_receiver/src/manager.rs

…alidation

anchor/message_validator/src/consensus_message.rs

dknopik

great work, LGTM! sorry for the delay

duty logic validation

89f0306

diegomrsantos force-pushed the duty-logic branch from e99fdc7 to 89f0306 Compare April 3, 2025 21:26

diegomrsantos self-assigned this Apr 3, 2025

diegomrsantos added fuzzing validation and removed fuzzing labels Apr 3, 2025

Zacholme7 mentioned this pull request Apr 4, 2025

Fuzzing via libfuzzer #232

Merged

diegomrsantos added 2 commits April 15, 2025 13:39

add duties tracker

62596d7

fix compilation and simplify

e2c6186

diegomrsantos requested a review from Copilot April 15, 2025 16:39

Copilot AI reviewed Apr 15, 2025

View reviewed changes

diegomrsantos and others added 15 commits April 15, 2025 18:53

Merge branch 'unstable' into duty-logic

9a2ebaf

add parking_lot

f48e866

add proposers duties tracking

43a00db

uncomment pre-commit

17075d3

simplify duties tracking

3ccb16a

use dashmap

763589b

update docs for SyncCommitteePerPeriod

4dc2620

remove metrics

1e7b599

improvements

80ff366

remove functions not used in BeaconNetwork

ae6a864

remove slots_per_epoch from Validator

2684a67

remove extra #

c63075b

Merge branch 'unstable' into duty-logic

f0e9e89

remove unnecessary get_slot_start_time

efb45e2

propagate error instead of failing silently

36b0a68

diegomrsantos commented Apr 17, 2025

View reviewed changes

anchor/message_validator/src/duties/duties_tracker.rs Show resolved Hide resolved

improve logging adding gossipsub and ssv msg id

070d434

diegomrsantos marked this pull request as ready for review April 17, 2025 13:25

change log to trace

9c9d3f1

Merge branch 'unstable' into duty-logic

5a98514

# Conflicts: # anchor/client/src/lib.rs # anchor/message_receiver/src/manager.rs

diegomrsantos requested a review from Copilot April 17, 2025 17:54

Copilot AI reviewed Apr 17, 2025

View reviewed changes

anchor/message_validator/src/consensus_message.rs Outdated Show resolved Hide resolved

diegomrsantos added 5 commits April 17, 2025 20:07

add issue to todo

91db909

propagate error in message_lateness

de8938d

use into() for ValidatorIndex

7f74742

remove comment

5f367fa

decrease duplication in partial_signature tests

db85d16

diegomrsantos commented Apr 18, 2025

View reviewed changes

fix constants values

8e7c892

dknopik reviewed Apr 22, 2025

View reviewed changes

fix test and refactor duty_limit

ee4cc61

diegomrsantos force-pushed the duty-logic branch from 6100a28 to ee4cc61 Compare April 22, 2025 16:50

Merge branch 'unstable' into duty-logic

cae7132

diegomrsantos commented Apr 22, 2025

View reviewed changes

diegomrsantos requested review from dknopik and Copilot April 22, 2025 17:23

diegomrsantos added the ready-for-review This PR is ready to be reviewed label Apr 22, 2025

Copilot AI reviewed Apr 22, 2025

View reviewed changes

anchor/message_receiver/src/manager.rs Show resolved Hide resolved

diegomrsantos added 3 commits April 22, 2025 19:32

rename var and fix typo

8797378

add test_early_message_fails_validation and test_late_message_fails_v…

37ba5a8

…alidation

make test_validate_ssv_message_consensus_success easier to understand

ce86f70

diegomrsantos requested a review from jking-aus April 23, 2025 17:18

dknopik reviewed Apr 24, 2025

View reviewed changes

anchor/message_validator/src/consensus_message.rs Outdated Show resolved Hide resolved

diegomrsantos and others added 2 commits April 24, 2025 10:33

fix duty limit comparison

415c9c7

Merge branch 'unstable' into duty-logic

401d245

diegomrsantos changed the title ~~Duty logic validation~~ Duty logic validation for consensus messages Apr 25, 2025

dknopik approved these changes Apr 25, 2025

View reviewed changes

dknopik merged commit 26f4a8b into sigp:unstable Apr 25, 2025
11 checks passed

		// - duty's starting slot + 34 (committee and aggregation)
		// - duty's starting slot + 3 (other types)

Duty logic validation for consensus messages #228

Duty logic validation for consensus messages #228

Uh oh!

Conversation

diegomrsantos commented Apr 3, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Issue Addressed

Proposed Changes

Additional Info

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull Request Overview

Reviewed Changes

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull Request Overview

Reviewed Changes

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

nkryuchkov Apr 22, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

diegomrsantos Apr 22, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

nkryuchkov Apr 28, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull Request Overview

Reviewed Changes

Uh oh!

Uh oh!

Uh oh!

dknopik left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

diegomrsantos commented Apr 3, 2025 •

edited

Loading

nkryuchkov Apr 22, 2025 •

edited

Loading

diegomrsantos Apr 22, 2025 •

edited

Loading

nkryuchkov Apr 28, 2025 •

edited

Loading