Lion - dapplion 241e52a202
Whisk (SSLE) with Curdleproofs - rebased (#3342)
* Introduce consensus code for Whisk

* polish, simplify, clean up (~100 fewer lines)

@asn-d6: As discussed, I fixed a few bugs along the way but likely also introduced some bugs :)

* minor cleanups and fixes

* simplify is_k_commitment_unique

* Update beacon-chain.md

* Update beacon-chain.md

* Initialize `k` in `get_validator_from_deposit()`

* minor cleanups

* Update beacon-chain.md

* Create beacon-chain.md

This PR changes the Whisk tracker format to be of the form `(r * pubkey, r * BLS_GT_GENERATOR)` instead of `(r * k * BLS_G1_GENERATOR, r * BLS_G1_GENERATOR)`. This allows for non-interactive tracker registrations from validator pubkeys, removing ~50 lines the code. It also significantly reduces the amount of state overhead. This PR also removes permutation commitments, though those can be easily readded if deemed necessary.

* A couple of fixes to the no-registration simplification

@asn-d6: Readded a consistency check for `IsValidWhiskOpeningProof` (involving `pubkey` instead of `k_commitment`).

* remove unused helpers

* use Mary's suggested tracker

* Update beacon-chain.md

* Revert G_t element optimization

This needs its own ethresearch post, and some additional analysis to see if we can do the shuffle ZKP in the allowed
timeframe.

This reverts commit 8517acabfc1dbb1a35789e6170b5db0bb2c19c9a.

* Implement new shuffling strategy

Ditch the Feistel logic and instead have each shuffler pick the row they shuffle using their RANDAO reveal.

* Curdleproofs edits

* working whisk eth2spec

* working whisk dummy test

* add more boilerplate set up code

* rebase constants

* Implement even newer and simplified shuffling strategy

This commit further simplifies 0faef30fc131d1b471b63a7f16772eeeef548ef8 by removing the entire squareshuffle.

The latest version of https://eprint.iacr.org/2022/560 proposes that each shuffler picks random indices from the entire
candidate set instead of organizing validators into a square.

* Move to _features

* remove dummy test

* Run doctoc

* Change Whisk's previous fork to Capella instead of Bellatrix. Make linter happier.

* Fix lint

* Fix pylint

* Fix mypy issues

* Clean-up get_beacon_proposer_index

* Fix doc headers

* Fix capella link

* Update apply_deposit

* Rename process_shuffled_trackers

---------

Co-authored-by: George Kadianakis <desnacked@riseup.net>
Co-authored-by: Justin <drakefjustin@gmail.com>
Co-authored-by: Nalin Bhardwaj <nalinbhardwaj@nibnalin.me>
Co-authored-by: Hsiao-Wei Wang <hsiaowei.eth@gmail.com>
2023-06-08 15:35:03 +08:00

4.9 KiB

Whisk -- Fork Logic

Notice: This document is a work-in-progress for researchers and implementers.

Table of contents

Introduction

This document describes the process of Whisk upgrade.

"""
    WHISK_FORK_EPOCH
        |                     cooldown
        |                     | ||
        v                     vsvv
      --+~~~~~~~~~~~~~~~~~~~~~----+-
               shuffling          ^
                                  |
                                  |
                         proposer selection
                        candidate selection
"""

Configuration

Warning: this configuration is not definitive.

Name Value
WHISK_FORK_VERSION Version('0x05000000')
WHISK_FORK_EPOCH Epoch(18446744073709551615) TBD

Fork to Whisk

If state.slot % SLOTS_PER_EPOCH == 0 and compute_epoch_at_slot(state.slot) == WHISK_FORK_EPOCH, an irregular state change is made to upgrade to Whisk. WHISK_FORK_EPOCH must be a multiple of WHISK_RUN_DURATION_IN_EPOCHS.

The upgrade occurs after the completion of the inner loop of process_slots that sets state.slot equal to WHISK_FORK_EPOCH * SLOTS_PER_EPOCH.

This ensures that we drop right into the beginning of the shuffling phase but without process_whisk_epoch() triggering for this Whisk run. Hence we handle all the setup ourselves in upgrade_to_whisk() below.

def whisk_candidate_selection(state: BeaconState, epoch: Epoch) -> None:
    # TODO
    # pylint: disable=unused-argument
    pass
def whisk_proposer_selection(state: BeaconState, epoch: Epoch) -> None:
    # TODO
    # pylint: disable=unused-argument
    pass
def upgrade_to_whisk(pre: bellatrix.BeaconState) -> BeaconState:
    epoch = bellatrix.get_current_epoch(pre)
    post = BeaconState(
        # Versioning
        genesis_time=pre.genesis_time,
        genesis_validators_root=pre.genesis_validators_root,
        slot=pre.slot,
        fork=Fork(
            previous_version=pre.fork.current_version,
            current_version=WHISK_FORK_VERSION,
            epoch=epoch,
        ),
        # History
        latest_block_header=pre.latest_block_header,
        block_roots=pre.block_roots,
        state_roots=pre.state_roots,
        historical_roots=pre.historical_roots,
        # Eth1
        eth1_data=pre.eth1_data,
        eth1_data_votes=pre.eth1_data_votes,
        eth1_deposit_index=pre.eth1_deposit_index,
        # Registry
        validators=[],
        balances=pre.balances,
        # Randomness
        randao_mixes=pre.randao_mixes,
        # Slashings
        slashings=pre.slashings,
        # Participation
        previous_epoch_participation=pre.previous_epoch_participation,
        current_epoch_participation=pre.current_epoch_participation,
        # Finality
        justification_bits=pre.justification_bits,
        previous_justified_checkpoint=pre.previous_justified_checkpoint,
        current_justified_checkpoint=pre.current_justified_checkpoint,
        finalized_checkpoint=pre.finalized_checkpoint,
        # Inactivity
        inactivity_scores=pre.inactivity_Scores,
    )

    # Initialize all validators with predictable commitments
    for val_index, pre_validator in enumerate(pre.validators):
        whisk_commitment, whisk_tracker = get_initial_commitments(get_unique_whisk_k(post, ValidatorIndex(val_index)))

        post_validator = Validator(
            pubkey=pre_validator.pubkey,
            withdrawal_credentials=pre_validator.withdrawal_credentials,
            effective_balance=pre_validator.effective_balance,
            slashed=pre_validator.slashed,
            activation_eligibility_epoch=pre_validator.activation_eligibility_epoch,
            activation_epoch=pre_validator.activation_epoch,
            exit_epoch=pre_validator.exit_epoch,
            withdrawable_epoch=pre_validator.withdrawable_epoch,
            whisk_commitment=whisk_commitment,
            whisk_tracker=whisk_tracker,
        )
        post.validators.append(post_validator)

    # Do a candidate selection followed by a proposer selection so that we have proposers for the upcoming day
    # Use an old epoch when selecting candidates so that we don't get the same seed as in the next candidate selection
    whisk_candidate_selection(post, epoch - WHISK_PROPOSER_SELECTION_GAP - 1)
    whisk_proposer_selection(post, epoch)

    # Do a final round of candidate selection.
    # We need it so that we have something to shuffle over the upcoming shuffling phase.
    whisk_candidate_selection(post, epoch)

    return post