601 lines
24 KiB
Nim
601 lines
24 KiB
Nim
# Copyright (c) 2018-2024 Status Research & Development GmbH
|
|
# Licensed and distributed under either of
|
|
# * MIT license (license terms in the root directory or at https://opensource.org/licenses/MIT).
|
|
# * Apache v2 license (license terms in the root directory or at https://www.apache.org/licenses/LICENSE-2.0).
|
|
# at your option. This file may not be copied, modified, or distributed except according to those terms.
|
|
|
|
{.push raises: [].}
|
|
|
|
# Helpers and functions pertaining to managing the validator set
|
|
|
|
import
|
|
std/algorithm,
|
|
"."/[crypto, helpers]
|
|
export helpers
|
|
|
|
const
|
|
SEED_SIZE = sizeof(Eth2Digest)
|
|
ROUND_SIZE = 1
|
|
POSITION_WINDOW_SIZE = 4
|
|
PIVOT_VIEW_SIZE = SEED_SIZE + ROUND_SIZE
|
|
TOTAL_SIZE = PIVOT_VIEW_SIZE + POSITION_WINDOW_SIZE
|
|
|
|
# https://github.com/ethereum/consensus-specs/blob/v1.4.0-beta.6/specs/phase0/beacon-chain.md#compute_shuffled_index
|
|
# https://github.com/ethereum/consensus-specs/blob/v1.4.0-beta.6/specs/phase0/beacon-chain.md#compute_committee
|
|
# Port of https://github.com/protolambda/zrnt/blob/v0.14.0/eth2/beacon/shuffle.go
|
|
func shuffle_list*(input: var seq[ValidatorIndex], seed: Eth2Digest) =
|
|
let list_size = input.lenu64
|
|
|
|
if list_size <= 1: return
|
|
|
|
var buf {.noinit.}: array[TOTAL_SIZE, byte]
|
|
|
|
# Seed is always the first 32 bytes of the hash input, we never have to change
|
|
# this part of the buffer.
|
|
buf[0..<32] = seed.data
|
|
|
|
# The original code includes a direction flag, but only the reverse direction
|
|
# is used in eth2, so we simplify it here
|
|
for r in 0'u8..<SHUFFLE_ROUND_COUNT.uint8:
|
|
# spec: pivot = bytes_to_int(hash(seed + int_to_bytes1(round))[0:8]) % list_size
|
|
# This is the "int_to_bytes1(round)", appended to the seed.
|
|
buf[SEED_SIZE] = (SHUFFLE_ROUND_COUNT.uint8 - r - 1)
|
|
|
|
# Seed is already in place, now just hash the correct part of the buffer,
|
|
# and take a uint64 from it, and modulo it to get a pivot within range.
|
|
let
|
|
pivotDigest = eth2digest(buf.toOpenArray(0, PIVOT_VIEW_SIZE - 1))
|
|
pivot = bytes_to_uint64(pivotDigest.data.toOpenArray(0, 7)) mod list_size
|
|
|
|
# Split up the for-loop in two:
|
|
# 1. Handle the part from 0 (incl) to pivot (incl). This is mirrored around
|
|
# (pivot / 2)
|
|
# 2. Handle the part from pivot (excl) to N (excl). This is mirrored around
|
|
# ((pivot / 2) + (size/2))
|
|
# The pivot defines a split in the array, with each of the splits mirroring
|
|
# their data within the split.
|
|
# Print out some example even/odd sized index lists, with some even/odd pivots,
|
|
# and you can deduce how the mirroring works exactly.
|
|
# Note that the mirror is strict enough to not consider swapping the index
|
|
# @mirror with itself.
|
|
# Since we are iterating through the "positions" in order, we can just
|
|
# repeat the hash every 256th position.
|
|
# No need to pre-compute every possible hash for efficiency like in the
|
|
# example code.
|
|
# We only need it consecutively (we are going through each in reverse order
|
|
# however, but same thing)
|
|
|
|
# spec: source = hash(seed + int_to_bytes1(round) + int_to_bytes4(position // 256))
|
|
# - seed is still in 0:32 (excl., 32 bytes)
|
|
# - round number is still in 32
|
|
# - mix in the position for randomness, except the last byte of it,
|
|
# which will be used later to select a bit from the resulting hash.
|
|
# We start from the pivot position, and work back to the mirror position
|
|
# (of the part left to the pivot).
|
|
# This makes us process each pear exactly once (instead of unnecessarily
|
|
# twice, like in the spec)
|
|
buf[33..<37] = uint_to_bytes(uint32(pivot shr 8))
|
|
|
|
var
|
|
mirror = (pivot + 1) shr 1
|
|
source = eth2digest(buf)
|
|
byteV = source.data[(pivot and 0xff) shr 3]
|
|
i = 0'u64
|
|
j = pivot
|
|
|
|
template shuffle =
|
|
while i < mirror:
|
|
# The pair is i,j. With j being the bigger of the two, hence the "position" identifier of the pair.
|
|
# Every 256th bit (aligned to j).
|
|
if (j and 0xff) == 0xff:
|
|
# just overwrite the last part of the buffer, reuse the start (seed, round)
|
|
buf[33..<37] = uint_to_bytes(uint32(j shr 8))
|
|
source = eth2digest(buf)
|
|
|
|
# Same trick with byte retrieval. Only every 8th.
|
|
if (j and 0x07) == 0x7:
|
|
byteV = source.data[(j and 0xff'u64) shr 3]
|
|
|
|
let
|
|
bitV = (byteV shr (j and 0x7)) and 0x1
|
|
|
|
if bitV == 1:
|
|
swap(input[i], input[j])
|
|
|
|
i.inc
|
|
j.dec
|
|
|
|
shuffle
|
|
|
|
# Now repeat, but for the part after the pivot.
|
|
mirror = (pivot + list_size + 1) shr 1
|
|
let lend = list_size - 1
|
|
# Again, seed and round input is in place, just update the position.
|
|
# We start at the end, and work back to the mirror point.
|
|
# This makes us process each pear exactly once (instead of unnecessarily twice, like in the spec)
|
|
buf[33..<37] = uint_to_bytes(uint32(lend shr 8))
|
|
|
|
source = eth2digest(buf)
|
|
byteV = source.data[(lend and 0xff) shr 3]
|
|
i = pivot + 1'u64
|
|
j = lend
|
|
|
|
shuffle
|
|
|
|
func get_shuffled_active_validator_indices*(
|
|
state: ForkyBeaconState, epoch: Epoch,
|
|
mix: Eth2Digest): seq[ValidatorIndex] =
|
|
# Non-spec function, to cache a data structure from which one can cheaply
|
|
# compute both get_active_validator_indexes() and get_beacon_committee().
|
|
var active_validator_indices = get_active_validator_indices(state, epoch)
|
|
let seed = get_seed(state, epoch, DOMAIN_BEACON_ATTESTER, mix)
|
|
shuffle_list(active_validator_indices, seed)
|
|
active_validator_indices
|
|
|
|
func get_shuffled_active_validator_indices*(
|
|
state: ForkyBeaconState, epoch: Epoch): seq[ValidatorIndex] =
|
|
# Non-spec function, to cache a data structure from which one can cheaply
|
|
# compute both get_active_validator_indexes() and get_beacon_committee().
|
|
var active_validator_indices = get_active_validator_indices(state, epoch)
|
|
let seed = get_seed(state, epoch, DOMAIN_BEACON_ATTESTER)
|
|
shuffle_list(active_validator_indices, seed)
|
|
active_validator_indices
|
|
|
|
func get_shuffled_active_validator_indices*(
|
|
cache: var StateCache, state: ForkyBeaconState, epoch: Epoch):
|
|
var seq[ValidatorIndex] =
|
|
# `cache` comes first because of nim's borrowing rules for the `var` return -
|
|
# the `var` returns avoids copying the validator set.
|
|
cache.shuffled_active_validator_indices.withValue(epoch, validator_indices) do:
|
|
return validator_indices[]
|
|
do:
|
|
let indices = get_shuffled_active_validator_indices(state, epoch)
|
|
return cache.shuffled_active_validator_indices.mgetOrPut(epoch, indices)
|
|
|
|
func get_shuffled_active_validator_indices*(
|
|
cache: var StateCache, state: ForkedHashedBeaconState, epoch: Epoch):
|
|
seq[ValidatorIndex] =
|
|
withState(state):
|
|
cache.get_shuffled_active_validator_indices(forkyState.data, epoch)
|
|
|
|
# https://github.com/ethereum/consensus-specs/blob/v1.4.0-beta.7/specs/phase0/beacon-chain.md#get_active_validator_indices
|
|
func count_active_validators*(state: ForkyBeaconState,
|
|
epoch: Epoch,
|
|
cache: var StateCache): uint64 =
|
|
cache.get_shuffled_active_validator_indices(state, epoch).lenu64
|
|
|
|
# https://github.com/ethereum/consensus-specs/blob/v1.4.0/specs/phase0/beacon-chain.md#get_committee_count_per_slot
|
|
func get_committee_count_per_slot*(num_active_validators: uint64): uint64 =
|
|
clamp(
|
|
num_active_validators div SLOTS_PER_EPOCH div TARGET_COMMITTEE_SIZE,
|
|
1'u64, MAX_COMMITTEES_PER_SLOT)
|
|
|
|
func get_committee_count_per_slot*(state: ForkyBeaconState,
|
|
epoch: Epoch,
|
|
cache: var StateCache): uint64 =
|
|
## Return the number of committees at ``slot``.
|
|
let
|
|
active_validator_count = count_active_validators(state, epoch, cache)
|
|
get_committee_count_per_slot(active_validator_count)
|
|
|
|
func get_committee_count_per_slot*(state: ForkedHashedBeaconState,
|
|
epoch: Epoch,
|
|
cache: var StateCache): uint64 =
|
|
withState(state):
|
|
get_committee_count_per_slot(forkyState.data, epoch, cache)
|
|
|
|
iterator get_committee_indices*(committee_count_per_slot: uint64): CommitteeIndex =
|
|
for idx in 0'u64..<min(committee_count_per_slot, MAX_COMMITTEES_PER_SLOT):
|
|
let committee_index = CommitteeIndex.init(idx).expect("value clamped")
|
|
yield committee_index
|
|
|
|
# https://github.com/ethereum/consensus-specs/blob/v1.4.0-beta.6/specs/phase0/beacon-chain.md#compute_committee
|
|
func compute_committee_slice*(
|
|
active_validators, index, count: uint64): Slice[int] =
|
|
doAssert active_validators <= ValidatorIndex.high.uint64
|
|
|
|
if index < count:
|
|
let
|
|
start = (active_validators * index) div count
|
|
endIdx = (active_validators * (index + 1)) div count
|
|
|
|
start.int..(endIdx.int - 1)
|
|
else:
|
|
0 .. -1
|
|
|
|
iterator compute_committee*(shuffled_indices: seq[ValidatorIndex],
|
|
index: uint64, count: uint64): (int, ValidatorIndex) =
|
|
let
|
|
slice = compute_committee_slice(shuffled_indices.lenu64, index, count)
|
|
var idx = 0
|
|
for i in slice:
|
|
yield (idx, shuffled_indices[i])
|
|
idx += 1
|
|
|
|
func compute_committee*(shuffled_indices: seq[ValidatorIndex],
|
|
index: uint64, count: uint64): seq[ValidatorIndex] =
|
|
## Return the committee corresponding to ``indices``, ``seed``, ``index``,
|
|
## and committee ``count``.
|
|
## In this version, we pass in the shuffled indices meaning we no longer need
|
|
## the seed.
|
|
let
|
|
slice = compute_committee_slice(shuffled_indices.lenu64, index, count)
|
|
|
|
# In spec, this calls get_shuffled_index() every time, but that's wasteful
|
|
# Here, get_beacon_committee() gets the shuffled version.
|
|
shuffled_indices[slice]
|
|
|
|
func compute_committee_len*(
|
|
active_validators, index, count: uint64): uint64 =
|
|
## Return the committee corresponding to ``indices``, ``seed``, ``index``,
|
|
## and committee ``count``.
|
|
|
|
let
|
|
slice = compute_committee_slice(active_validators, index, count)
|
|
|
|
(slice.b - slice.a + 1).uint64
|
|
|
|
# https://github.com/ethereum/consensus-specs/blob/v1.4.0-beta.6/specs/phase0/beacon-chain.md#get_beacon_committee
|
|
iterator get_beacon_committee*(
|
|
state: ForkyBeaconState, slot: Slot, index: CommitteeIndex,
|
|
cache: var StateCache): (int, ValidatorIndex) =
|
|
## Return the beacon committee at ``slot`` for ``index``.
|
|
let
|
|
epoch = epoch(slot)
|
|
committees_per_slot = get_committee_count_per_slot(state, epoch, cache)
|
|
for index_in_committee, idx in compute_committee(
|
|
cache.get_shuffled_active_validator_indices(state, epoch),
|
|
(slot mod SLOTS_PER_EPOCH) * committees_per_slot +
|
|
index.uint64,
|
|
committees_per_slot * SLOTS_PER_EPOCH
|
|
): yield (index_in_committee, idx)
|
|
|
|
func get_beacon_committee*(
|
|
state: ForkyBeaconState, slot: Slot, index: CommitteeIndex,
|
|
cache: var StateCache): seq[ValidatorIndex] =
|
|
## Return the beacon committee at ``slot`` for ``index``.
|
|
let
|
|
epoch = epoch(slot)
|
|
committees_per_slot = get_committee_count_per_slot(state, epoch, cache)
|
|
compute_committee(
|
|
cache.get_shuffled_active_validator_indices(state, epoch),
|
|
(slot mod SLOTS_PER_EPOCH) * committees_per_slot +
|
|
index.uint64,
|
|
committees_per_slot * SLOTS_PER_EPOCH
|
|
)
|
|
|
|
func get_beacon_committee*(
|
|
state: ForkedHashedBeaconState, slot: Slot, index: CommitteeIndex,
|
|
cache: var StateCache): seq[ValidatorIndex] =
|
|
# This one is used by tests/, ncli/, and a couple of places in RPC
|
|
# TODO use the iterator version alone, to remove the risk of using
|
|
# diverging get_beacon_committee() in tests and beacon_chain/ by a
|
|
# wrapper approach (e.g., toSeq). This is a perf tradeoff for test
|
|
# correctness/consistency.
|
|
withState(state):
|
|
get_beacon_committee(forkyState.data, slot, index, cache)
|
|
|
|
# https://github.com/ethereum/consensus-specs/blob/v1.4.0-beta.6/specs/phase0/beacon-chain.md#get_beacon_committee
|
|
func get_beacon_committee_len*(
|
|
state: ForkyBeaconState, slot: Slot, index: CommitteeIndex,
|
|
cache: var StateCache): uint64 =
|
|
## Return the number of members in the beacon committee at ``slot``
|
|
## for ``index``.
|
|
let
|
|
epoch = epoch(slot)
|
|
committees_per_slot = get_committee_count_per_slot(state, epoch, cache)
|
|
|
|
compute_committee_len(
|
|
count_active_validators(state, epoch, cache),
|
|
(slot mod SLOTS_PER_EPOCH) * committees_per_slot +
|
|
index.uint64,
|
|
committees_per_slot * SLOTS_PER_EPOCH
|
|
)
|
|
|
|
func get_beacon_committee_len*(
|
|
state: ForkedHashedBeaconState, slot: Slot, index: CommitteeIndex,
|
|
cache: var StateCache): uint64 =
|
|
# This one is used by tests
|
|
withState(state):
|
|
get_beacon_committee_len(forkyState.data, slot, index, cache)
|
|
|
|
# https://github.com/ethereum/consensus-specs/blob/v1.4.0-beta.6/specs/phase0/beacon-chain.md#compute_shuffled_index
|
|
template compute_shuffled_index_aux(
|
|
index: uint64, index_count: uint64, seed: Eth2Digest, iter: untyped):
|
|
uint64 =
|
|
## Return the shuffled index corresponding to ``seed`` (and ``index_count``).
|
|
doAssert index < index_count
|
|
|
|
var
|
|
source_buffer {.noinit.}: array[(32+1+4), byte]
|
|
cur_idx_permuted = index
|
|
|
|
source_buffer[0..31] = seed.data
|
|
|
|
# Swap or not (https://link.springer.com/content/pdf/10.1007%2F978-3-642-32009-5_1.pdf)
|
|
# See the 'generalized domain' algorithm on page 3
|
|
for current_round in iter:
|
|
source_buffer[32] = current_round
|
|
|
|
let
|
|
# If using multiple indices, can amortize this
|
|
pivot =
|
|
bytes_to_uint64(eth2digest(source_buffer.toOpenArray(0, 32)).data.toOpenArray(0, 7)) mod
|
|
index_count
|
|
|
|
flip = ((index_count + pivot) - cur_idx_permuted) mod index_count
|
|
position = max(cur_idx_permuted, flip)
|
|
source_buffer[33..36] = uint_to_bytes(uint32(position shr 8))
|
|
let
|
|
source = eth2digest(source_buffer).data
|
|
byte_value = source[(position mod 256) shr 3]
|
|
bit = (byte_value shr (position mod 8)) mod 2
|
|
|
|
cur_idx_permuted = if bit != 0: flip else: cur_idx_permuted
|
|
|
|
cur_idx_permuted
|
|
|
|
func compute_shuffled_index*(
|
|
index: uint64, index_count: uint64, seed: Eth2Digest): uint64 =
|
|
## Return the shuffled index corresponding to ``seed`` (and ``index_count``).
|
|
compute_shuffled_index_aux(index, index_count, seed) do:
|
|
0'u8 ..< SHUFFLE_ROUND_COUNT.uint8
|
|
|
|
func compute_inverted_shuffled_index*(
|
|
index: uint64, index_count: uint64, seed: Eth2Digest): uint64 =
|
|
## Return the inverse of the shuffled index corresponding to ``seed`` (and
|
|
## ``index_count``).
|
|
compute_shuffled_index_aux(index, index_count, seed) do:
|
|
countdown(SHUFFLE_ROUND_COUNT.uint8 - 1, 0'u8, 1)
|
|
|
|
# https://github.com/ethereum/consensus-specs/blob/v1.4.0-beta.6/specs/phase0/beacon-chain.md#compute_proposer_index
|
|
template compute_proposer_index(state: ForkyBeaconState,
|
|
indices: openArray[ValidatorIndex], seed: Eth2Digest,
|
|
unshuffleTransform: untyped): Opt[ValidatorIndex] =
|
|
## Return from ``indices`` a random index sampled by effective balance.
|
|
const MAX_RANDOM_BYTE = 255
|
|
|
|
if len(indices) == 0:
|
|
Opt.none(ValidatorIndex)
|
|
else:
|
|
let seq_len {.inject.} = indices.lenu64
|
|
|
|
var
|
|
i = 0'u64
|
|
buffer: array[32+8, byte]
|
|
res: Opt[ValidatorIndex]
|
|
buffer[0..31] = seed.data
|
|
while true:
|
|
buffer[32..39] = uint_to_bytes(i div 32)
|
|
let
|
|
shuffled_index {.inject.} =
|
|
compute_shuffled_index(i mod seq_len, seq_len, seed)
|
|
candidate_index = indices[unshuffleTransform]
|
|
random_byte = (eth2digest(buffer).data)[i mod 32]
|
|
effective_balance = state.validators[candidate_index].effective_balance
|
|
if effective_balance * MAX_RANDOM_BYTE >=
|
|
MAX_EFFECTIVE_BALANCE.Gwei * random_byte:
|
|
res = Opt.some(candidate_index)
|
|
break
|
|
i += 1
|
|
|
|
doAssert res.isSome
|
|
res
|
|
|
|
func compute_proposer_index(state: ForkyBeaconState,
|
|
indices: openArray[ValidatorIndex], seed: Eth2Digest):
|
|
Opt[ValidatorIndex] =
|
|
## Return from ``indices`` a random index sampled by effective balance.
|
|
compute_proposer_index(state, indices, seed, shuffled_index)
|
|
|
|
# https://github.com/ethereum/consensus-specs/blob/v1.4.0-beta.6/specs/phase0/beacon-chain.md#get_beacon_proposer_index
|
|
func get_beacon_proposer_index*(
|
|
state: ForkyBeaconState, cache: var StateCache, slot: Slot):
|
|
Opt[ValidatorIndex] =
|
|
let epoch = get_current_epoch(state)
|
|
|
|
if slot.epoch() != epoch:
|
|
# compute_proposer_index depends on `effective_balance`, therefore the
|
|
# beacon proposer index can only be computed for the "current" epoch:
|
|
# https://github.com/ethereum/consensus-specs/pull/772#issuecomment-475574357
|
|
return Opt.none(ValidatorIndex)
|
|
|
|
cache.beacon_proposer_indices.withValue(slot, proposer) do:
|
|
return proposer[]
|
|
do:
|
|
## Return the beacon proposer index at the current slot.
|
|
|
|
var buffer: array[32 + 8, byte]
|
|
buffer[0..31] = get_seed(state, epoch, DOMAIN_BEACON_PROPOSER).data
|
|
|
|
# There's exactly one beacon proposer per slot - the same validator may
|
|
# however propose several times in the same epoch (however unlikely)
|
|
let indices = get_active_validator_indices(state, epoch)
|
|
var res: Opt[ValidatorIndex]
|
|
|
|
for epoch_slot in epoch.slots():
|
|
buffer[32..39] = uint_to_bytes(epoch_slot.asUInt64)
|
|
let seed = eth2digest(buffer)
|
|
let pi = compute_proposer_index(state, indices, seed)
|
|
if epoch_slot == slot:
|
|
res = pi
|
|
cache.beacon_proposer_indices[epoch_slot] = pi
|
|
|
|
return res
|
|
|
|
# https://github.com/ethereum/consensus-specs/blob/v1.4.0-beta.6/specs/phase0/beacon-chain.md#get_beacon_proposer_index
|
|
func get_beacon_proposer_indices*(
|
|
state: ForkyBeaconState, shuffled_indices: openArray[ValidatorIndex], epoch: Epoch):
|
|
seq[Opt[ValidatorIndex]] =
|
|
## Return the beacon proposer indices at the current epoch, using shuffled
|
|
## rather than sorted active validator indices.
|
|
var
|
|
buffer {.noinit.}: array[32 + 8, byte]
|
|
res: seq[Opt[ValidatorIndex]]
|
|
|
|
buffer[0..31] = get_seed(state, epoch, DOMAIN_BEACON_PROPOSER).data
|
|
let epoch_shuffle_seed = get_seed(state, epoch, DOMAIN_BEACON_ATTESTER)
|
|
|
|
for epoch_slot in epoch.slots():
|
|
buffer[32..39] = uint_to_bytes(epoch_slot.asUInt64)
|
|
res.add (
|
|
compute_proposer_index(state, shuffled_indices, eth2digest(buffer)) do:
|
|
compute_inverted_shuffled_index(
|
|
shuffled_index, seq_len, epoch_shuffle_seed))
|
|
|
|
res
|
|
|
|
# https://github.com/ethereum/consensus-specs/blob/v1.4.0-beta.6/specs/phase0/beacon-chain.md#get_beacon_proposer_index
|
|
func get_beacon_proposer_index*(state: ForkyBeaconState, cache: var StateCache):
|
|
Opt[ValidatorIndex] =
|
|
## Return the beacon proposer index at the current slot.
|
|
get_beacon_proposer_index(state, cache, state.slot)
|
|
|
|
func get_beacon_proposer_index*(state: ForkedHashedBeaconState,
|
|
cache: var StateCache, slot: Slot):
|
|
Opt[ValidatorIndex] =
|
|
withState(state):
|
|
get_beacon_proposer_index(forkyState.data, cache, slot)
|
|
|
|
# https://github.com/ethereum/consensus-specs/blob/v1.4.0/specs/phase0/validator.md#aggregation-selection
|
|
func is_aggregator*(committee_len: uint64, slot_signature: ValidatorSig): bool =
|
|
let modulo = max(1'u64, committee_len div TARGET_AGGREGATORS_PER_COMMITTEE)
|
|
bytes_to_uint64(eth2digest(
|
|
slot_signature.toRaw()).data.toOpenArray(0, 7)) mod modulo == 0
|
|
|
|
# https://github.com/ethereum/builder-specs/blob/v0.4.0/specs/bellatrix/validator.md#liveness-failsafe
|
|
func livenessFailsafeInEffect*(
|
|
block_roots: array[Limit SLOTS_PER_HISTORICAL_ROOT, Eth2Digest],
|
|
slot: Slot): bool =
|
|
const
|
|
MAX_MISSING_CONTIGUOUS = 3
|
|
MAX_MISSING_WINDOW = 5
|
|
|
|
static: doAssert MAX_MISSING_WINDOW > MAX_MISSING_CONTIGUOUS
|
|
if slot <= MAX_MISSING_CONTIGUOUS:
|
|
# Cannot ever trigger and allows a bit of safe arithmetic. Furthermore
|
|
# there's notionally always a genesis block, which pushes the earliest
|
|
# possible failure out an additional slot.
|
|
return false
|
|
|
|
# Using this slightly convoluted construction to handle wraparound better;
|
|
# baseIndex + faultInspectionWindow can overflow array but only exactly by
|
|
# the required amount. Furthermore, go back one more slot to address using
|
|
# that it looks ahead rather than looks back and whether a block's missing
|
|
# requires seeing the previous block_root.
|
|
let
|
|
faultInspectionWindow = min(distinctBase(slot) - 1, SLOTS_PER_EPOCH)
|
|
baseIndex = (slot + SLOTS_PER_HISTORICAL_ROOT - faultInspectionWindow) mod
|
|
SLOTS_PER_HISTORICAL_ROOT
|
|
endIndex = baseIndex + faultInspectionWindow - 1
|
|
|
|
doAssert endIndex mod SLOTS_PER_HISTORICAL_ROOT ==
|
|
(slot - 1) mod SLOTS_PER_HISTORICAL_ROOT
|
|
|
|
var
|
|
totalMissing = 0
|
|
streakLen = 0
|
|
maxStreakLen = 0
|
|
|
|
for i in baseIndex .. endIndex:
|
|
# This look-forward means checking slot i for being missing uses i - 1
|
|
if block_roots[(i mod SLOTS_PER_HISTORICAL_ROOT).int] ==
|
|
block_roots[((i + 1) mod SLOTS_PER_HISTORICAL_ROOT).int]:
|
|
totalMissing += 1
|
|
if totalMissing > MAX_MISSING_WINDOW:
|
|
return true
|
|
|
|
streakLen += 1
|
|
if streakLen > maxStreakLen:
|
|
maxStreakLen = streakLen
|
|
if maxStreakLen > MAX_MISSING_CONTIGUOUS:
|
|
return true
|
|
else:
|
|
streakLen = 0
|
|
|
|
false
|
|
|
|
# https://github.com/ethereum/consensus-specs/blob/v1.4.0-beta.5/specs/phase0/p2p-interface.md#attestation-subnet-subscription
|
|
func compute_subscribed_subnet(node_id: UInt256, epoch: Epoch, index: uint64):
|
|
SubnetId =
|
|
# Ensure neither `truncate` loses information
|
|
static:
|
|
doAssert EPOCHS_PER_SUBNET_SUBSCRIPTION <= high(uint64)
|
|
doAssert sizeof(UInt256) * 8 == NODE_ID_BITS
|
|
doAssert ATTESTATION_SUBNET_PREFIX_BITS < sizeof(SubnetId) * 8
|
|
|
|
let
|
|
node_id_prefix = truncate(
|
|
node_id shr (
|
|
NODE_ID_BITS - static(ATTESTATION_SUBNET_PREFIX_BITS.int)), uint64)
|
|
node_offset = truncate(
|
|
node_id mod static(EPOCHS_PER_SUBNET_SUBSCRIPTION.u256), uint64)
|
|
permutation_seed = eth2digest(uint_to_bytes(
|
|
uint64((epoch + node_offset) div EPOCHS_PER_SUBNET_SUBSCRIPTION)))
|
|
permutated_prefix = compute_shuffled_index(
|
|
node_id_prefix,
|
|
1 shl ATTESTATION_SUBNET_PREFIX_BITS,
|
|
permutation_seed,
|
|
)
|
|
SubnetId((permutated_prefix + index) mod ATTESTATION_SUBNET_COUNT)
|
|
|
|
# https://github.com/ethereum/consensus-specs/blob/v1.4.0-beta.4/specs/phase0/p2p-interface.md#attestation-subnet-subscription
|
|
iterator compute_subscribed_subnets*(node_id: UInt256, epoch: Epoch): SubnetId =
|
|
for index in 0'u64 ..< SUBNETS_PER_NODE:
|
|
yield compute_subscribed_subnet(node_id, epoch, index)
|
|
|
|
iterator get_committee_indices*(bits: AttestationCommitteeBits): CommitteeIndex =
|
|
for index, b in bits:
|
|
if b:
|
|
yield CommitteeIndex.init(uint64(index)).valueOr:
|
|
break # Too many bits! Shouldn't happen
|
|
|
|
func get_committee_index_one*(bits: AttestationCommitteeBits): Opt[CommitteeIndex] =
|
|
var res = Opt.none(CommitteeIndex)
|
|
for committee_index in get_committee_indices(bits):
|
|
if res.isSome(): return Opt.none(CommitteeIndex)
|
|
res = Opt.some(committee_index)
|
|
res
|
|
|
|
proc compute_on_chain_aggregate*(
|
|
network_aggregates: openArray[electra.Attestation]): Opt[electra.Attestation] =
|
|
# aggregates = sorted(network_aggregates, key=lambda a: get_committee_indices(a.committee_bits)[0])
|
|
let aggregates = network_aggregates.sortedByIt(it.committee_bits.get_committee_index_one().expect("just one"))
|
|
|
|
let data = aggregates[0].data
|
|
|
|
var agg: AggregateSignature
|
|
var committee_bits: AttestationCommitteeBits
|
|
|
|
var totalLen = 0
|
|
for i, a in aggregates:
|
|
totalLen += a.aggregation_bits.len
|
|
|
|
var aggregation_bits = ElectraCommitteeValidatorsBits.init(totalLen)
|
|
var pos = 0
|
|
for i, a in aggregates:
|
|
let
|
|
committee_index = ? get_committee_index_one(a.committee_bits)
|
|
first = pos == 0
|
|
|
|
for b in a.aggregation_bits:
|
|
aggregation_bits[pos] = b
|
|
pos += 1
|
|
|
|
let sig = ? a.signature.load() # Expensive
|
|
if first:
|
|
agg = AggregateSignature.init(sig)
|
|
else:
|
|
agg.aggregate(sig)
|
|
|
|
committee_bits[int(committee_index)] = true
|
|
|
|
let signature = agg.finish()
|
|
|
|
ok electra.Attestation(
|
|
aggregation_bits: aggregation_bits,
|
|
data: data,
|
|
committee_bits: committee_bits,
|
|
signature: signature.toValidatorSig(),
|
|
)
|