nimbus-eth2/beacon_chain/spec/validator.nim

257 lines
10 KiB
Nim

# Copyright (c) 2018-2019 Status Research & Development GmbH
# Licensed and distributed under either of
# * MIT license (license terms in the root directory or at http://opensource.org/licenses/MIT).
# * Apache v2 license (license terms in the root directory or at http://www.apache.org/licenses/LICENSE-2.0).
# at your option. This file may not be copied, modified, or distributed except according to those terms.
# Helpers and functions pertaining to managing the validator set
import
options, nimcrypto, sequtils, math, tables,
./datatypes, ./digest, ./helpers
# TODO: Proceed to renaming and signature changes
# https://github.com/ethereum/eth2.0-specs/blob/v0.9.1/specs/core/0_beacon-chain.md#compute_shuffled_index
# https://github.com/ethereum/eth2.0-specs/blob/v0.9.1/specs/core/0_beacon-chain.md#compute_committee
func get_shuffled_seq*(seed: Eth2Digest,
list_size: uint64,
): seq[ValidatorIndex] =
## Via https://github.com/protolambda/eth2-shuffle/blob/master/shuffle.go
## Shuffles ``validators`` into crosslink committees seeded by ``seed`` and
## ``slot``.
## Returns a list of ``SLOTS_PER_EPOCH * committees_per_slot`` committees
## where each committee is itself a list of validator indices.
##
## Invert the inner/outer loops from the spec, essentially. Most useful
## hash result re-use occurs within a round.
# Empty size -> empty list.
if list_size == 0:
return
var
# Share these buffers.
# TODO: Redo to follow spec.
# We can have an "Impl" private version that takes buffer as parameters
# so that we avoid alloc on repeated calls from compute_committee
pivot_buffer: array[(32+1), byte]
source_buffer: array[(32+1+4), byte]
shuffled_active_validator_indices = mapIt(
0 ..< list_size.int, it.ValidatorIndex)
sources = repeat(Eth2Digest(), (list_size div 256) + 1)
## The pivot's a function of seed and round only.
## This doesn't change across rounds.
pivot_buffer[0..31] = seed.data
source_buffer[0..31] = seed.data
for round in 0 ..< SHUFFLE_ROUND_COUNT:
let round_bytes1 = int_to_bytes1(round)[0]
pivot_buffer[32] = round_bytes1
source_buffer[32] = round_bytes1
# Only one pivot per round.
let pivot = bytes_to_int(eth2hash(pivot_buffer).data.toOpenArray(0, 7)) mod list_size
## Only need to run, per round, position div 256 hashes, so precalculate
## them. This consumes memory, but for low-memory devices, it's possible
## to mitigate by some light LRU caching and similar.
for reduced_position in 0 ..< sources.len:
source_buffer[33..36] = int_to_bytes4(reduced_position.uint64)
sources[reduced_position] = eth2hash(source_buffer)
## Iterate over all the indices. This was in get_permuted_index, but large
## efficiency gains exist in caching and re-using data.
for index in 0 ..< list_size.int:
let
cur_idx_permuted = shuffled_active_validator_indices[index]
flip = ((list_size + pivot) - cur_idx_permuted.uint64) mod list_size
position = max(cur_idx_permuted.int, flip.int)
let
source = sources[position div 256].data
byte_value = source[(position mod 256) div 8]
bit = (byte_value shr (position mod 8)) mod 2
if bit != 0:
shuffled_active_validator_indices[index] = flip.ValidatorIndex
result = shuffled_active_validator_indices
# https://github.com/ethereum/eth2.0-specs/blob/v0.9.1/specs/core/0_beacon-chain.md#get_previous_epoch
func get_previous_epoch*(state: BeaconState): Epoch =
# Return the previous epoch (unless the current epoch is ``GENESIS_EPOCH``).
let current_epoch = get_current_epoch(state)
if current_epoch == GENESIS_EPOCH:
current_epoch
else:
current_epoch - 1
# https://github.com/ethereum/eth2.0-specs/blob/v0.8.4/specs/core/0_beacon-chain.md#get_shard_delta
func get_shard_delta*(state: BeaconState, epoch: Epoch): uint64 =
## Return the number of shards to increment ``state.start_shard``
## during ``epoch``.
min(get_committee_count_at_slot(state, epoch.compute_start_slot_at_epoch) *
SLOTS_PER_EPOCH,
(SHARD_COUNT - SHARD_COUNT div SLOTS_PER_EPOCH).uint64)
# https://github.com/ethereum/eth2.0-specs/blob/v0.8.4/specs/core/0_beacon-chain.md#get_start_shard
func get_start_shard*(state: BeaconState, epoch: Epoch): Shard =
# Return the start shard of the 0th committee at ``epoch``.
doAssert epoch <= get_current_epoch(state) + 1
var
check_epoch = get_current_epoch(state) + 1
shard =
(state.start_shard +
get_shard_delta(state, get_current_epoch(state))) mod SHARD_COUNT
while check_epoch > epoch:
check_epoch -= 1.Epoch
shard = (shard + SHARD_COUNT - get_shard_delta(state, check_epoch)) mod
SHARD_COUNT
return shard
# TODO remove when shim layer isn't needed
func get_slot_and_index*(state: BeaconState, epoch: Epoch, shard: Shard): auto =
let gcc_index =
(shard + SHARD_COUNT - get_start_shard(state, epoch)) mod SHARD_COUNT
# Want (slot % SLOTS_PER_EPOCH) * committees_per_slot + index to result in
# same index, where
# committees_per_slot = get_committee_count_at_slot(state, slot)
# https://github.com/ethereum/eth2.0-specs/blob/v0.9.1/specs/core/0_beacon-chain.md#get_beacon_committee
let committees_per_slot =
get_committee_count_at_slot(state, compute_start_slot_at_epoch(epoch))
# get_beacon_committee(...) uses a straightforward linear mapping, row-centric
# minimize the `index` offset (s.t. >= 0) and maximize `slot`.
let
slot = gcc_index div committees_per_slot
index = gcc_index mod committees_per_slot
# TODO it might be bad if slot >= SLOTS_PER_EPOCH in this construction,
# but not necessarily
(compute_start_slot_at_epoch(epoch) + slot, index)
# https://github.com/ethereum/eth2.0-specs/blob/v0.9.1/specs/core/0_beacon-chain.md#compute_committee
func compute_committee(indices: seq[ValidatorIndex], seed: Eth2Digest,
index: uint64, count: uint64, stateCache: var StateCache): seq[ValidatorIndex] =
## Return the committee corresponding to ``indices``, ``seed``, ``index``,
## and committee ``count``.
let
start = (len(indices).uint64 * index) div count
endIdx = (len(indices).uint64 * (index + 1)) div count
key = (indices.len, seed)
if key notin stateCache.crosslink_committee_cache:
stateCache.crosslink_committee_cache[key] =
get_shuffled_seq(seed, len(indices).uint64)
# These assertions from compute_shuffled_index(...)
let index_count = indices.len().uint64
doAssert endIdx <= index_count
doAssert index_count <= 2'u64^40
# In spec, this calls get_shuffled_index() every time, but that's wasteful
mapIt(
start.int .. (endIdx.int-1),
indices[stateCache.crosslink_committee_cache[key][it]])
# https://github.com/ethereum/eth2.0-specs/blob/v0.9.1/specs/core/0_beacon-chain.md#get_beacon_committee
func get_beacon_committee*(state: BeaconState, slot: Slot, index: uint64, cache: var StateCache): seq[ValidatorIndex] =
# Return the beacon committee at ``slot`` for ``index``.
let
epoch = compute_epoch_at_slot(slot)
# TODO use state caching for this or not?
committees_per_slot = get_committee_count_at_slot(state, slot)
## This is a somewhat more fragile, but high-ROI, caching setup --
## get_active_validator_indices() is slow to run in a loop and only
## changes once per epoch.
if epoch notin cache.active_validator_indices_cache:
cache.active_validator_indices_cache[epoch] =
get_active_validator_indices(state, epoch)
# TODO remove or replace this...
#if epoch notin cache.committee_count_cache:
# cache.committee_count_cache[epoch] = get_committee_count(state, epoch)
# TODO profiling & make sure caches populated
compute_committee(
cache.active_validator_indices_cache[epoch],
get_seed(state, epoch, DOMAIN_BEACON_ATTESTER),
(slot mod SLOTS_PER_EPOCH) * committees_per_slot + index,
committees_per_slot * SLOTS_PER_EPOCH,
cache
)
# https://github.com/ethereum/eth2.0-specs/blob/v0.8.4/specs/core/0_beacon-chain.md#get_crosslink_committee
func get_crosslink_committee*(state: BeaconState, epoch: Epoch, shard: Shard,
stateCache: var StateCache): seq[ValidatorIndex] =
doAssert shard >= 0'u64
if epoch notin stateCache.start_shard_cache:
stateCache.start_shard_cache[epoch] = get_start_shard(state, epoch)
let (gbc_slot, gbc_index) = get_slot_and_index(state, epoch, shard)
get_beacon_committee(state, gbc_slot, gbc_index, stateCache)
# Not from spec
func get_empty_per_epoch_cache*(): StateCache =
result.crosslink_committee_cache =
initTable[tuple[a: int, b: Eth2Digest], seq[ValidatorIndex]]()
result.active_validator_indices_cache =
initTable[Epoch, seq[ValidatorIndex]]()
result.start_shard_cache = initTable[Epoch, Shard]()
result.committee_count_cache = initTable[Epoch, uint64]()
# https://github.com/ethereum/eth2.0-specs/blob/v0.9.1/specs/core/0_beacon-chain.md#compute_proposer_index
func compute_proposer_index*(state: BeaconState, indices: seq[ValidatorIndex],
seed: Eth2Digest, stateCache: var StateCache): ValidatorIndex =
# Return from ``indices`` a random index sampled by effective balance.
const MAX_RANDOM_BYTE = 255
doAssert len(indices) > 0
# TODO fixme; should only be run once per slot and cached
# There's exactly one beacon proposer per slot.
let
seq_len = indices.len.uint64
shuffled_seq = get_shuffled_seq(seed, seq_len)
doAssert seq_len == shuffled_seq.len.uint64
var
i = 0
buffer: array[32+8, byte]
buffer[0..31] = seed.data
while true:
buffer[32..39] = int_to_bytes8(i.uint64 div 32)
let
candidate_index = shuffled_seq[(i.uint64 mod seq_len).int]
random_byte = (eth2hash(buffer).data)[i mod 32]
effective_balance =
state.validators[candidate_index].effective_balance
if effective_balance * MAX_RANDOM_BYTE >=
MAX_EFFECTIVE_BALANCE * random_byte:
return candidate_index
i += 1
# https://github.com/ethereum/eth2.0-specs/blob/v0.9.1/specs/core/0_beacon-chain.md#get_beacon_proposer_index
func get_beacon_proposer_index*(state: BeaconState, stateCache: var StateCache):
ValidatorIndex =
# Return the beacon proposer index at the current slot.
let epoch = get_current_epoch(state)
var buffer: array[32 + 8, byte]
buffer[0..31] = get_seed(state, epoch, DOMAIN_BEACON_PROPOSER).data
buffer[32..39] = int_to_bytes8(state.slot.uint64)
let
seed = eth2hash(buffer)
indices = get_active_validator_indices(state, epoch)
compute_proposer_index(state, indices, seed, stateCache)