913 lines
38 KiB
Nim
913 lines
38 KiB
Nim
# beacon_chain
|
|
# 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: [].}
|
|
|
|
import
|
|
chronicles, chronos, metrics,
|
|
../spec/[forks, signatures, signatures_batch],
|
|
../sszdump
|
|
|
|
from std/deques import Deque, addLast, contains, initDeque, items, len, shrink
|
|
from std/sequtils import mapIt
|
|
from ../consensus_object_pools/consensus_manager import
|
|
ConsensusManager, checkNextProposer, optimisticExecutionBlockHash,
|
|
runProposalForkchoiceUpdated, shouldSyncOptimistically, updateHead,
|
|
updateHeadWithExecution
|
|
from ../consensus_object_pools/blockchain_dag import
|
|
getBlockRef, getForkedBlock, getProposer, forkAtEpoch, loadExecutionBlockHash,
|
|
markBlockVerified, validatorKey
|
|
from ../beacon_clock import GetBeaconTimeFn, toFloatSeconds
|
|
from ../consensus_object_pools/block_dag import BlockRef, root, shortLog, slot
|
|
from ../consensus_object_pools/block_pools_types import
|
|
EpochRef, VerifierError
|
|
from ../consensus_object_pools/block_quarantine import
|
|
addBlobless, addOrphan, addUnviable, pop, removeOrphan
|
|
from ../consensus_object_pools/blob_quarantine import
|
|
BlobQuarantine, hasBlobs, popBlobs, put
|
|
from ../validators/validator_monitor import
|
|
MsgSource, ValidatorMonitor, registerAttestationInBlock, registerBeaconBlock,
|
|
registerSyncAggregateInBlock
|
|
from ../beacon_chain_db import getBlobSidecar, putBlobSidecar
|
|
from ../spec/state_transition_block import validate_blobs
|
|
|
|
export sszdump, signatures_batch
|
|
|
|
logScope: topics = "gossip_blocks"
|
|
|
|
# Block Processor
|
|
# ------------------------------------------------------------------------------
|
|
# The block processor moves blocks from "Incoming" to "Consensus verified"
|
|
|
|
declareHistogram beacon_store_block_duration_seconds,
|
|
"storeBlock() duration", buckets = [0.25, 0.5, 1, 2, 4, 8, Inf]
|
|
|
|
const
|
|
SLOTS_PER_PAYLOAD = SLOTS_PER_HISTORICAL_ROOT
|
|
## Number of slots we process between each execution payload execution, while
|
|
## syncing the finalized part of the chain
|
|
PAYLOAD_PRE_WALL_SLOTS = SLOTS_PER_EPOCH * 2
|
|
## Number of slots from wall time that we start processing every payload
|
|
|
|
type
|
|
BlockEntry = object
|
|
blck*: ForkedSignedBeaconBlock
|
|
blobs*: Opt[ForkedBlobSidecars]
|
|
maybeFinalized*: bool
|
|
## The block source claims the block has been finalized already
|
|
resfut*: Future[Result[void, VerifierError]].Raising([CancelledError])
|
|
queueTick*: Moment # Moment when block was enqueued
|
|
validationDur*: Duration # Time it took to perform gossip validation
|
|
src*: MsgSource
|
|
|
|
BlockProcessor* = object
|
|
## This manages the processing of blocks from different sources
|
|
## Blocks and attestations are enqueued in a gossip-validated state
|
|
##
|
|
## from:
|
|
## - Gossip (when synced)
|
|
## - SyncManager (during sync)
|
|
## - RequestManager (missing ancestor blocks)
|
|
##
|
|
## are then consensus-verified and added to:
|
|
## - the blockchain DAG
|
|
## - database
|
|
## - attestation pool
|
|
## - fork choice
|
|
##
|
|
## The processor will also reinsert blocks from the quarantine, should a
|
|
## parent be found.
|
|
|
|
# Config
|
|
# ----------------------------------------------------------------
|
|
dumpEnabled: bool
|
|
dumpDirInvalid: string
|
|
dumpDirIncoming: string
|
|
|
|
# Producers
|
|
# ----------------------------------------------------------------
|
|
blockQueue: AsyncQueue[BlockEntry]
|
|
|
|
# Consumer
|
|
# ----------------------------------------------------------------
|
|
consensusManager: ref ConsensusManager
|
|
## Blockchain DAG, AttestationPool and Quarantine
|
|
## Blockchain DAG, AttestationPool, Quarantine, and ELManager
|
|
validatorMonitor: ref ValidatorMonitor
|
|
getBeaconTime: GetBeaconTimeFn
|
|
|
|
blobQuarantine: ref BlobQuarantine
|
|
verifier: BatchVerifier
|
|
|
|
lastPayload: Slot
|
|
## The slot at which we sent a payload to the execution client the last
|
|
## time
|
|
|
|
NewPayloadStatus {.pure.} = enum
|
|
valid
|
|
notValid
|
|
invalid
|
|
noResponse
|
|
|
|
ProcessingStatus {.pure.} = enum
|
|
completed
|
|
notCompleted
|
|
|
|
# Initialization
|
|
# ------------------------------------------------------------------------------
|
|
|
|
proc new*(T: type BlockProcessor,
|
|
dumpEnabled: bool,
|
|
dumpDirInvalid, dumpDirIncoming: string,
|
|
rng: ref HmacDrbgContext, taskpool: TaskPoolPtr,
|
|
consensusManager: ref ConsensusManager,
|
|
validatorMonitor: ref ValidatorMonitor,
|
|
blobQuarantine: ref BlobQuarantine,
|
|
getBeaconTime: GetBeaconTimeFn): ref BlockProcessor =
|
|
(ref BlockProcessor)(
|
|
dumpEnabled: dumpEnabled,
|
|
dumpDirInvalid: dumpDirInvalid,
|
|
dumpDirIncoming: dumpDirIncoming,
|
|
blockQueue: newAsyncQueue[BlockEntry](),
|
|
consensusManager: consensusManager,
|
|
validatorMonitor: validatorMonitor,
|
|
blobQuarantine: blobQuarantine,
|
|
getBeaconTime: getBeaconTime,
|
|
verifier: BatchVerifier.init(rng, taskpool)
|
|
)
|
|
|
|
# Sync callbacks
|
|
# ------------------------------------------------------------------------------
|
|
|
|
func hasBlocks*(self: BlockProcessor): bool =
|
|
self.blockQueue.len() > 0
|
|
|
|
# Storage
|
|
# ------------------------------------------------------------------------------
|
|
|
|
proc dumpInvalidBlock*(
|
|
self: BlockProcessor, signedBlock: ForkySignedBeaconBlock) =
|
|
if self.dumpEnabled:
|
|
dump(self.dumpDirInvalid, signedBlock)
|
|
|
|
proc dumpBlock[T](
|
|
self: BlockProcessor,
|
|
signedBlock: ForkySignedBeaconBlock,
|
|
res: Result[T, VerifierError]) =
|
|
if self.dumpEnabled and res.isErr:
|
|
case res.error
|
|
of VerifierError.Invalid:
|
|
self.dumpInvalidBlock(signedBlock)
|
|
of VerifierError.MissingParent:
|
|
dump(self.dumpDirIncoming, signedBlock)
|
|
else:
|
|
discard
|
|
|
|
from ../consensus_object_pools/block_clearance import
|
|
addBackfillBlock, addHeadBlockWithParent, checkHeadBlock
|
|
|
|
proc storeBackfillBlock(
|
|
self: var BlockProcessor,
|
|
signedBlock: ForkySignedBeaconBlock,
|
|
blobsOpt: Opt[ForkyBlobSidecars]
|
|
): Result[void, VerifierError] =
|
|
const
|
|
consensusFork = typeof(signedBlock).kind
|
|
blobFork = blobForkAtConsensusFork(consensusFork).get(BlobFork.Deneb)
|
|
static: doAssert typeof(blobsOpt).T is blobFork.BlobSidecars
|
|
|
|
# The block is certainly not missing any more
|
|
self.consensusManager.quarantine[].missing.del(signedBlock.root)
|
|
|
|
# Establish blob viability before calling addbackfillBlock to avoid
|
|
# writing the block in case of blob error.
|
|
var blobsOk = true
|
|
when consensusFork >= ConsensusFork.Deneb:
|
|
if blobsOpt.isSome:
|
|
let blobs = blobsOpt.get()
|
|
let kzgCommits = signedBlock.message.body.blob_kzg_commitments.asSeq
|
|
if blobs.len > 0 or kzgCommits.len > 0:
|
|
let r = validate_blobs(kzgCommits, blobs.mapIt(KzgBlob(bytes: it.blob)),
|
|
blobs.mapIt(it.kzg_proof))
|
|
if r.isErr():
|
|
debug "backfill blob validation failed",
|
|
blockRoot = shortLog(signedBlock.root),
|
|
blobs = shortLog(blobs),
|
|
blck = shortLog(signedBlock.message),
|
|
kzgCommits = mapIt(kzgCommits, shortLog(it)),
|
|
signature = shortLog(signedBlock.signature),
|
|
msg = r.error()
|
|
blobsOk = r.isOk()
|
|
|
|
if not blobsOk:
|
|
return err(VerifierError.Invalid)
|
|
|
|
let res = self.consensusManager.dag.addBackfillBlock(signedBlock)
|
|
|
|
if res.isErr():
|
|
case res.error
|
|
of VerifierError.MissingParent:
|
|
if signedBlock.message.parent_root in
|
|
self.consensusManager.quarantine[].unviable:
|
|
# DAG doesn't know about unviable ancestor blocks - we do! Translate
|
|
# this to the appropriate error so that sync etc doesn't retry the block
|
|
self.consensusManager.quarantine[].addUnviable(signedBlock.root)
|
|
|
|
return err(VerifierError.UnviableFork)
|
|
of VerifierError.UnviableFork:
|
|
# Track unviables so that descendants can be discarded properly
|
|
self.consensusManager.quarantine[].addUnviable(signedBlock.root)
|
|
else: discard
|
|
return res
|
|
|
|
# Only store blobs after successfully establishing block viability.
|
|
let blobs = blobsOpt.valueOr: blobFork.BlobSidecars() @[]
|
|
for b in blobs:
|
|
self.consensusManager.dag.db.putBlobSidecar(b[])
|
|
|
|
res
|
|
|
|
from web3/engine_api_types import
|
|
PayloadAttributesV1, PayloadAttributesV2, PayloadAttributesV3,
|
|
PayloadExecutionStatus, PayloadStatusV1
|
|
from ../el/el_manager import
|
|
ELManager, forkchoiceUpdated, hasConnection, hasProperlyConfiguredConnection,
|
|
sendNewPayload
|
|
|
|
proc expectValidForkchoiceUpdated(
|
|
elManager: ELManager, headBlockPayloadAttributesType: typedesc,
|
|
headBlockHash, safeBlockHash, finalizedBlockHash: Eth2Digest,
|
|
receivedBlock: ForkySignedBeaconBlock): Future[void] {.async: (raises: [CancelledError]).} =
|
|
let
|
|
(payloadExecutionStatus, _) = await elManager.forkchoiceUpdated(
|
|
headBlockHash = headBlockHash,
|
|
safeBlockHash = safeBlockHash,
|
|
finalizedBlockHash = finalizedBlockHash,
|
|
payloadAttributes = Opt.none headBlockPayloadAttributesType)
|
|
receivedExecutionBlockHash =
|
|
when typeof(receivedBlock).kind >= ConsensusFork.Bellatrix:
|
|
receivedBlock.message.body.execution_payload.block_hash
|
|
else:
|
|
# https://github.com/nim-lang/Nim/issues/19802
|
|
(static(default(Eth2Digest)))
|
|
|
|
# Only called when expecting this to be valid because `newPayload` or some
|
|
# previous `forkchoiceUpdated` had already marked it as valid. However, if
|
|
# it's not the block that was received, don't info/warn either way given a
|
|
# relative lack of immediate evidence.
|
|
if receivedExecutionBlockHash != headBlockHash:
|
|
return
|
|
|
|
case payloadExecutionStatus
|
|
of PayloadExecutionStatus.valid:
|
|
# situation nominal
|
|
discard
|
|
of PayloadExecutionStatus.accepted, PayloadExecutionStatus.syncing:
|
|
info "execution payload forkChoiceUpdated status ACCEPTED/SYNCING, but was previously VALID",
|
|
payloadExecutionStatus = $payloadExecutionStatus, headBlockHash,
|
|
safeBlockHash, finalizedBlockHash,
|
|
receivedBlock = shortLog(receivedBlock)
|
|
of PayloadExecutionStatus.invalid, PayloadExecutionStatus.invalid_block_hash:
|
|
warn "execution payload forkChoiceUpdated status INVALID, but was previously VALID",
|
|
payloadExecutionStatus = $payloadExecutionStatus, headBlockHash,
|
|
safeBlockHash, finalizedBlockHash,
|
|
receivedBlock = shortLog(receivedBlock)
|
|
|
|
from ../consensus_object_pools/attestation_pool import
|
|
addForkChoice, selectOptimisticHead, BeaconHead
|
|
from ../consensus_object_pools/spec_cache import get_attesting_indices
|
|
from ../spec/datatypes/phase0 import TrustedSignedBeaconBlock
|
|
from ../spec/datatypes/altair import SignedBeaconBlock
|
|
|
|
from ../spec/datatypes/bellatrix import ExecutionPayload, SignedBeaconBlock
|
|
from ../spec/datatypes/capella import
|
|
ExecutionPayload, SignedBeaconBlock, asTrusted, shortLog
|
|
|
|
# TODO investigate why this seems to allow compilation even though it doesn't
|
|
# directly address deneb.ExecutionPayload when complaint was that it didn't
|
|
# know about "deneb"
|
|
from ../spec/datatypes/deneb import SignedBeaconBlock, asTrusted, shortLog
|
|
|
|
proc newExecutionPayload*(
|
|
elManager: ELManager, blck: SomeForkyBeaconBlock):
|
|
Future[Opt[PayloadExecutionStatus]] {.async: (raises: [CancelledError]).} =
|
|
|
|
template executionPayload: untyped = blck.body.execution_payload
|
|
|
|
if not elManager.hasProperlyConfiguredConnection:
|
|
if elManager.hasConnection:
|
|
info "No execution client connected; cannot process block payloads",
|
|
executionPayload = shortLog(executionPayload)
|
|
else:
|
|
debug "No execution client connected; cannot process block payloads",
|
|
executionPayload = shortLog(executionPayload)
|
|
return Opt.none PayloadExecutionStatus
|
|
|
|
debug "newPayload: inserting block into execution engine",
|
|
executionPayload = shortLog(executionPayload)
|
|
|
|
try:
|
|
let payloadStatus = await elManager.sendNewPayload(blck)
|
|
|
|
debug "newPayload: succeeded",
|
|
parentHash = executionPayload.parent_hash,
|
|
blockHash = executionPayload.block_hash,
|
|
blockNumber = executionPayload.block_number,
|
|
payloadStatus = $payloadStatus
|
|
|
|
return Opt.some payloadStatus
|
|
except CatchableError as err:
|
|
warn "newPayload failed - check execution client",
|
|
msg = err.msg,
|
|
parentHash = shortLog(executionPayload.parent_hash),
|
|
blockHash = shortLog(executionPayload.block_hash),
|
|
blockNumber = executionPayload.block_number
|
|
return Opt.none PayloadExecutionStatus
|
|
|
|
proc getExecutionValidity(
|
|
elManager: ELManager,
|
|
blck: bellatrix.SignedBeaconBlock | capella.SignedBeaconBlock |
|
|
deneb.SignedBeaconBlock | electra.SignedBeaconBlock):
|
|
Future[NewPayloadStatus] {.async: (raises: [CancelledError]).} =
|
|
if not blck.message.is_execution_block:
|
|
return NewPayloadStatus.valid # vacuously
|
|
|
|
try:
|
|
let executionPayloadStatus = await elManager.newExecutionPayload(
|
|
blck.message)
|
|
if executionPayloadStatus.isNone:
|
|
return NewPayloadStatus.noResponse
|
|
|
|
case executionPayloadStatus.get
|
|
of PayloadExecutionStatus.invalid, PayloadExecutionStatus.invalid_block_hash:
|
|
# Blocks come either from gossip or request manager requests. In the
|
|
# former case, they've passed libp2p gosisp validation which implies
|
|
# correct signature for correct proposer,which makes spam expensive,
|
|
# while for the latter, spam is limited by the request manager.
|
|
info "execution payload invalid from EL client newPayload",
|
|
executionPayloadStatus = $executionPayloadStatus.get,
|
|
executionPayload = shortLog(blck.message.body.execution_payload),
|
|
blck = shortLog(blck)
|
|
return NewPayloadStatus.invalid
|
|
of PayloadExecutionStatus.syncing, PayloadExecutionStatus.accepted:
|
|
return NewPayloadStatus.notValid
|
|
of PayloadExecutionStatus.valid:
|
|
return NewPayloadStatus.valid
|
|
except CatchableError as err:
|
|
error "newPayload failed and leaked exception",
|
|
err = err.msg,
|
|
executionPayload = shortLog(blck.message.body.execution_payload),
|
|
blck = shortLog(blck)
|
|
return NewPayloadStatus.noResponse
|
|
|
|
proc checkBloblessSignature(
|
|
self: BlockProcessor,
|
|
signed_beacon_block: deneb.SignedBeaconBlock | electra.SignedBeaconBlock):
|
|
Result[void, cstring] =
|
|
let dag = self.consensusManager.dag
|
|
let parent = dag.getBlockRef(signed_beacon_block.message.parent_root).valueOr:
|
|
return err("checkBloblessSignature called with orphan block")
|
|
let proposer = getProposer(
|
|
dag, parent, signed_beacon_block.message.slot).valueOr:
|
|
return err("checkBloblessSignature: Cannot compute proposer")
|
|
if distinctBase(proposer) != signed_beacon_block.message.proposer_index:
|
|
return err("checkBloblessSignature: Incorrect proposer")
|
|
if not verify_block_signature(
|
|
dag.forkAtEpoch(signed_beacon_block.message.slot.epoch),
|
|
getStateField(dag.headState, genesis_validators_root),
|
|
signed_beacon_block.message.slot,
|
|
signed_beacon_block.root,
|
|
dag.validatorKey(proposer).get(),
|
|
signed_beacon_block.signature):
|
|
return err("checkBloblessSignature: Invalid proposer signature")
|
|
ok()
|
|
|
|
template withForkyBlckAndBlobs(
|
|
blck: ForkedSignedBeaconBlock,
|
|
blobs: Opt[ForkedBlobSidecars],
|
|
body: untyped): untyped =
|
|
withBlck(blck):
|
|
when consensusFork >= ConsensusFork.Deneb:
|
|
const blobFork = blobForkAtConsensusFork(consensusFork).expect("Blobs OK")
|
|
let forkyBlobs {.inject, used.} =
|
|
if blobs.isSome:
|
|
# Nim 2.0.8: `forks.BlobSidecars(blobFork)` does not work here:
|
|
# > type mismatch: got 'BlobFork' for 'blobFork`gensym15'
|
|
# but expected 'BlobSidecars'
|
|
var fBlobs: deneb.BlobSidecars
|
|
for blob in blobs.get:
|
|
doAssert blob.kind == blobFork,
|
|
"Must verify blob inclusion proof before `enqueueBlock`"
|
|
fBlobs.add blob.forky(blobFork)
|
|
Opt.some fBlobs
|
|
else:
|
|
Opt.none deneb.BlobSidecars
|
|
else:
|
|
doAssert blobs.isNone, "Blobs are not supported before Deneb"
|
|
let forkyBlobs {.inject, used.} = Opt.none deneb.BlobSidecars
|
|
body
|
|
|
|
proc enqueueBlock*(
|
|
self: var BlockProcessor, src: MsgSource, blck: ForkedSignedBeaconBlock,
|
|
blobs: Opt[ForkedBlobSidecars],
|
|
resfut: Future[Result[void, VerifierError]].Raising([CancelledError]) = nil,
|
|
maybeFinalized = false,
|
|
validationDur = Duration()) =
|
|
withForkyBlckAndBlobs(blck, blobs):
|
|
if forkyBlck.message.slot <= self.consensusManager.dag.finalizedHead.slot:
|
|
# let backfill blocks skip the queue - these are always "fast" to process
|
|
# because there are no state rewinds to deal with
|
|
let res = self.storeBackfillBlock(forkyBlck, forkyBlobs)
|
|
resfut.complete(res)
|
|
return
|
|
|
|
try:
|
|
self.blockQueue.addLastNoWait(BlockEntry(
|
|
blck: blck,
|
|
blobs: blobs,
|
|
maybeFinalized: maybeFinalized,
|
|
resfut: resfut, queueTick: Moment.now(),
|
|
validationDur: validationDur,
|
|
src: src))
|
|
except AsyncQueueFullError:
|
|
raiseAssert "unbounded queue"
|
|
|
|
proc storeBlock(
|
|
self: ref BlockProcessor, src: MsgSource, wallTime: BeaconTime,
|
|
signedBlock: ForkySignedBeaconBlock,
|
|
blobsOpt: Opt[ForkyBlobSidecars],
|
|
maybeFinalized = false,
|
|
queueTick: Moment = Moment.now(),
|
|
validationDur = Duration()
|
|
): Future[Result[BlockRef, (VerifierError, ProcessingStatus)]] {.
|
|
async: (raises: [CancelledError]).} =
|
|
## storeBlock is the main entry point for unvalidated blocks - all untrusted
|
|
## blocks, regardless of origin, pass through here. When storing a block,
|
|
## we will add it to the dag and pass it to all block consumers that need
|
|
## to know about it, such as the fork choice and the monitoring
|
|
const
|
|
consensusFork = typeof(signedBlock).kind
|
|
blobFork = blobForkAtConsensusFork(consensusFork).get(BlobFork.Deneb)
|
|
static: doAssert typeof(blobsOpt).T is blobFork.BlobSidecars
|
|
|
|
let
|
|
attestationPool = self.consensusManager.attestationPool
|
|
startTick = Moment.now()
|
|
vm = self.validatorMonitor
|
|
dag = self.consensusManager.dag
|
|
wallSlot = wallTime.slotOrZero
|
|
|
|
# If the block is missing its parent, it will be re-orphaned below
|
|
self.consensusManager.quarantine[].removeOrphan(signedBlock)
|
|
# The block is certainly not missing any more
|
|
self.consensusManager.quarantine[].missing.del(signedBlock.root)
|
|
|
|
if signedBlock.message.parent_root in
|
|
self.consensusManager.quarantine[].unviable:
|
|
# DAG doesn't know about unviable ancestor blocks - we do however!
|
|
self.consensusManager.quarantine[].addUnviable(signedBlock.root)
|
|
|
|
return err((VerifierError.UnviableFork, ProcessingStatus.completed))
|
|
|
|
template handleVerifierError(errorParam: VerifierError): auto =
|
|
let error = errorParam
|
|
case error
|
|
of VerifierError.MissingParent:
|
|
if (let r = self.consensusManager.quarantine[].addOrphan(
|
|
dag.finalizedHead.slot, ForkedSignedBeaconBlock.init(signedBlock));
|
|
r.isErr()):
|
|
debug "could not add orphan",
|
|
blockRoot = shortLog(signedBlock.root),
|
|
blck = shortLog(signedBlock.message),
|
|
signature = shortLog(signedBlock.signature),
|
|
err = r.error()
|
|
else:
|
|
if blobsOpt.isSome:
|
|
for blobSidecar in blobsOpt.get:
|
|
self.blobQuarantine[].put(blobSidecar)
|
|
debug "Block quarantined",
|
|
blockRoot = shortLog(signedBlock.root),
|
|
blck = shortLog(signedBlock.message),
|
|
signature = shortLog(signedBlock.signature)
|
|
|
|
of VerifierError.UnviableFork:
|
|
# Track unviables so that descendants can be discarded promptly
|
|
self.consensusManager.quarantine[].addUnviable(signedBlock.root)
|
|
else:
|
|
discard
|
|
|
|
err((error, ProcessingStatus.completed))
|
|
|
|
let
|
|
# We have to be careful that there exists only one in-flight entry point
|
|
# for adding blocks or the checks performed in `checkHeadBlock` might
|
|
# be invalidated (ie a block could be added while we wait for EL response
|
|
# here)
|
|
parent = dag.checkHeadBlock(signedBlock)
|
|
|
|
if parent.isErr():
|
|
# TODO This logic can be removed if the database schema is extended
|
|
# to store non-canonical heads on top of the canonical head!
|
|
# If that is done, the database no longer contains extra blocks
|
|
# that have not yet been assigned a `BlockRef`
|
|
if parent.error() == VerifierError.MissingParent:
|
|
# This indicates that no `BlockRef` is available for the `parent_root`.
|
|
# However, the block may still be available in local storage. On startup,
|
|
# only the canonical branch is imported into `blockchain_dag`, while
|
|
# non-canonical branches are re-discovered with sync/request managers.
|
|
# Data from non-canonical branches that has already been verified during
|
|
# a previous run of the beacon node is already stored in the database but
|
|
# only lacks a `BlockRef`. Loading the branch from the database saves a
|
|
# lot of time, especially when a non-canonical branch has non-trivial
|
|
# depth. Note that if it turns out that a non-canonical branch eventually
|
|
# becomes canonical, it is vital to import it as quickly as possible.
|
|
let
|
|
parent_root = signedBlock.message.parent_root
|
|
parentBlck = dag.getForkedBlock(parent_root)
|
|
if parentBlck.isSome():
|
|
var blobsOk = true
|
|
let blobs =
|
|
withBlck(parentBlck.get()):
|
|
when consensusFork >= ConsensusFork.Deneb:
|
|
const blobFork =
|
|
blobForkAtConsensusFork(consensusFork).expect("Blobs OK")
|
|
var blob_sidecars: ForkedBlobSidecars
|
|
for i in 0 ..< forkyBlck.message.body.blob_kzg_commitments.len:
|
|
let blob = blobFork.BlobSidecar.new()
|
|
if not dag.db.getBlobSidecar(parent_root, i.BlobIndex, blob[]):
|
|
blobsOk = false # Pruned, or inconsistent DB
|
|
break
|
|
blob_sidecars.add ForkedBlobSidecar.init(blob)
|
|
Opt.some blob_sidecars
|
|
else:
|
|
Opt.none ForkedBlobSidecars
|
|
if blobsOk:
|
|
debug "Loaded parent block from storage", parent_root
|
|
self[].enqueueBlock(
|
|
MsgSource.gossip, parentBlck.unsafeGet().asSigned(), blobs)
|
|
|
|
return handleVerifierError(parent.error())
|
|
|
|
let
|
|
payloadStatus =
|
|
if maybeFinalized and
|
|
(self.lastPayload + SLOTS_PER_PAYLOAD) > signedBlock.message.slot and
|
|
(signedBlock.message.slot + PAYLOAD_PRE_WALL_SLOTS) < wallSlot and
|
|
signedBlock.message.is_execution_block:
|
|
# Skip payload validation when message source (reasonably) claims block
|
|
# has been finalized - this speeds up forward sync - in the worst case
|
|
# that the claim is false, we will correct every time we process a block
|
|
# from an honest source (or when we're close to head).
|
|
# Occasionally we also send a payload to the EL so that it can
|
|
# progress in its own sync.
|
|
NewPayloadStatus.noResponse
|
|
else:
|
|
when typeof(signedBlock).kind >= ConsensusFork.Bellatrix:
|
|
await self.consensusManager.elManager.getExecutionValidity(signedBlock)
|
|
else:
|
|
NewPayloadStatus.valid # vacuously
|
|
payloadValid = payloadStatus == NewPayloadStatus.valid
|
|
|
|
if NewPayloadStatus.invalid == payloadStatus:
|
|
self.consensusManager.quarantine[].addUnviable(signedBlock.root)
|
|
self[].dumpInvalidBlock(signedBlock)
|
|
return err((VerifierError.UnviableFork, ProcessingStatus.completed))
|
|
|
|
if NewPayloadStatus.noResponse == payloadStatus:
|
|
# When the execution layer is not available to verify the payload, we do the
|
|
# required check on the CL side instead and proceed as if the EL was syncing
|
|
|
|
# TODO run https://github.com/ethereum/consensus-specs/blob/v1.3.0/specs/deneb/beacon-chain.md#blob-kzg-commitments
|
|
# https://github.com/ethereum/execution-apis/blob/main/src/engine/experimental/blob-extension.md#specification
|
|
# "This validation MUST be instantly run in all cases even during active
|
|
# sync process."
|
|
#
|
|
# Client software MUST validate `blockHash` value as being equivalent to
|
|
# `Keccak256(RLP(ExecutionBlockHeader))`
|
|
# https://github.com/ethereum/execution-apis/blob/v1.0.0-beta.3/src/engine/paris.md#specification
|
|
#
|
|
# This should simulate an unsynced EL, which still must perform these
|
|
# checks. This means it must be able to do so without context, beyond
|
|
# whatever data the block itself contains.
|
|
when typeof(signedBlock).kind >= ConsensusFork.Bellatrix:
|
|
template payload(): auto = signedBlock.message.body.execution_payload
|
|
if signedBlock.message.is_execution_block and
|
|
payload.block_hash !=
|
|
signedBlock.message.compute_execution_block_hash():
|
|
debug "Execution block hash validation failed",
|
|
execution_payload = shortLog(payload)
|
|
self[].dumpInvalidBlock(signedBlock)
|
|
doAssert strictVerification notin dag.updateFlags
|
|
self.consensusManager.quarantine[].addUnviable(signedBlock.root)
|
|
return err((VerifierError.Invalid, ProcessingStatus.completed))
|
|
|
|
let newPayloadTick = Moment.now()
|
|
|
|
# TODO with v1.4.0, not sure this is still relevant
|
|
# Establish blob viability before calling addHeadBlock to avoid
|
|
# writing the block in case of blob error.
|
|
when typeof(signedBlock).kind >= ConsensusFork.Deneb:
|
|
if blobsOpt.isSome:
|
|
let blobs = blobsOpt.get()
|
|
let kzgCommits = signedBlock.message.body.blob_kzg_commitments.asSeq
|
|
if blobs.len > 0 or kzgCommits.len > 0:
|
|
let r = validate_blobs(kzgCommits, blobs.mapIt(KzgBlob(bytes: it.blob)),
|
|
blobs.mapIt(it.kzg_proof))
|
|
if r.isErr():
|
|
debug "blob validation failed",
|
|
blockRoot = shortLog(signedBlock.root),
|
|
blobs = shortLog(blobs),
|
|
blck = shortLog(signedBlock.message),
|
|
kzgCommits = mapIt(kzgCommits, shortLog(it)),
|
|
signature = shortLog(signedBlock.signature),
|
|
msg = r.error()
|
|
return err((VerifierError.Invalid, ProcessingStatus.completed))
|
|
|
|
type Trusted = typeof signedBlock.asTrusted()
|
|
|
|
let
|
|
blck = dag.addHeadBlockWithParent(
|
|
self.verifier, signedBlock, parent.value(), payloadValid) do (
|
|
blckRef: BlockRef, trustedBlock: Trusted,
|
|
epochRef: EpochRef, unrealized: FinalityCheckpoints):
|
|
# Callback add to fork choice if valid
|
|
attestationPool[].addForkChoice(
|
|
epochRef, blckRef, unrealized, trustedBlock.message, wallTime)
|
|
|
|
vm[].registerBeaconBlock(
|
|
src, wallTime, trustedBlock.message)
|
|
|
|
for attestation in trustedBlock.message.body.attestations:
|
|
for validator_index in dag.get_attesting_indices(attestation, true):
|
|
vm[].registerAttestationInBlock(attestation.data, validator_index,
|
|
trustedBlock.message.slot)
|
|
|
|
withState(dag[].clearanceState):
|
|
when consensusFork >= ConsensusFork.Altair and
|
|
Trusted isnot phase0.TrustedSignedBeaconBlock: # altair+
|
|
for i in trustedBlock.message.body.sync_aggregate.sync_committee_bits.oneIndices():
|
|
vm[].registerSyncAggregateInBlock(
|
|
trustedBlock.message.slot, trustedBlock.root,
|
|
forkyState.data.current_sync_committee.pubkeys.data[i])
|
|
|
|
self[].dumpBlock(signedBlock, blck)
|
|
|
|
# There can be a scenario where we receive a block we already received.
|
|
# However this block was before the last finalized epoch and so its parent
|
|
# was pruned from the ForkChoice.
|
|
if blck.isErr():
|
|
return handleVerifierError(blck.error())
|
|
|
|
# Even if the EL is not responding, we'll only try once every now and then
|
|
# to give it a block - this avoids a pathological slowdown where a busy EL
|
|
# times out on every block we give it because it's busy with the previous
|
|
# one
|
|
self[].lastPayload = signedBlock.message.slot
|
|
|
|
# write blobs now that block has been written.
|
|
let blobs = blobsOpt.valueOr: BlobSidecars @[]
|
|
for b in blobs:
|
|
self.consensusManager.dag.db.putBlobSidecar(b[])
|
|
|
|
let addHeadBlockTick = Moment.now()
|
|
|
|
# Eagerly update head: the incoming block "should" get selected.
|
|
#
|
|
# storeBlock gets called from validator_duties, which depends on its not
|
|
# blocking progress any longer than necessary, and processBlock here, in
|
|
# which case it's fine to await for a while on engine API results.
|
|
#
|
|
# Three general scenarios: (1) pre-merge; (2) merge, already `VALID` by way
|
|
# of `newPayload`; (3) optimistically imported, need to call fcU before DAG
|
|
# updateHead. Because in a non-finalizing network, completing sync isn't as
|
|
# useful because regular reorgs likely still occur, and when finalizing the
|
|
# EL is only called every SLOTS_PER_PAYLOAD slots regardless, await, rather
|
|
# than asyncSpawn forkchoiceUpdated calls.
|
|
#
|
|
# This reduces in-flight fcU spam, which both reduces EL load and decreases
|
|
# otherwise somewhat unpredictable CL head movement.
|
|
|
|
# Grab the new head according to our latest attestation data; determines how
|
|
# async this needs to be.
|
|
let newHead = attestationPool[].selectOptimisticHead(
|
|
wallSlot.start_beacon_time)
|
|
|
|
if newHead.isOk:
|
|
template elManager(): auto = self.consensusManager.elManager
|
|
if self.consensusManager[].shouldSyncOptimistically(wallSlot):
|
|
# Optimistic head is far in the future; report it as head block to EL.
|
|
|
|
# Note that the specification allows an EL client to skip fcU processing
|
|
# if an update to an ancestor is requested.
|
|
# > Client software MAY skip an update of the forkchoice state and MUST
|
|
# NOT begin a payload build process if `forkchoiceState.headBlockHash`
|
|
# references an ancestor of the head of canonical chain.
|
|
# https://github.com/ethereum/execution-apis/blob/v1.0.0-beta.3/src/engine/paris.md#specification-1
|
|
#
|
|
# However, in practice, an EL client may not have completed importing all
|
|
# block headers, so may be unaware of a block's ancestor status.
|
|
# Therefore, hopping back and forth between the optimistic head and the
|
|
# chain DAG head does not work well in practice, e.g., Geth:
|
|
# - "Beacon chain gapped" from DAG head to optimistic head,
|
|
# - followed by "Beacon chain reorged" from optimistic head back to DAG.
|
|
self.consensusManager[].updateHead(newHead.get.blck)
|
|
|
|
template callForkchoiceUpdated(attributes: untyped) =
|
|
if NewPayloadStatus.noResponse != payloadStatus and
|
|
not self.consensusManager[].optimisticExecutionBlockHash.isZero:
|
|
discard await elManager.forkchoiceUpdated(
|
|
headBlockHash =
|
|
self.consensusManager[].optimisticExecutionBlockHash,
|
|
safeBlockHash = newHead.get.safeExecutionBlockHash,
|
|
finalizedBlockHash = newHead.get.finalizedExecutionBlockHash,
|
|
payloadAttributes = Opt.none attributes)
|
|
|
|
let consensusFork = self.consensusManager.dag.cfg.consensusForkAtEpoch(
|
|
newHead.get.blck.bid.slot.epoch)
|
|
withConsensusFork(consensusFork):
|
|
when consensusFork >= ConsensusFork.Bellatrix:
|
|
callForkchoiceUpdated(consensusFork.PayloadAttributes)
|
|
else:
|
|
let
|
|
headExecutionBlockHash =
|
|
dag.loadExecutionBlockHash(newHead.get.blck).get(ZERO_HASH)
|
|
wallSlot = self.getBeaconTime().slotOrZero
|
|
if headExecutionBlockHash.isZero or
|
|
NewPayloadStatus.noResponse == payloadStatus:
|
|
# Blocks without execution payloads can't be optimistic, and don't try
|
|
# to fcU to a block the EL hasn't seen
|
|
self.consensusManager[].updateHead(newHead.get.blck)
|
|
elif newHead.get.blck.executionValid:
|
|
# `forkchoiceUpdated` necessary for EL client only.
|
|
self.consensusManager[].updateHead(newHead.get.blck)
|
|
|
|
template callExpectValidFCU(payloadAttributeType: untyped): auto =
|
|
await elManager.expectValidForkchoiceUpdated(
|
|
headBlockPayloadAttributesType = payloadAttributeType,
|
|
headBlockHash = headExecutionBlockHash,
|
|
safeBlockHash = newHead.get.safeExecutionBlockHash,
|
|
finalizedBlockHash = newHead.get.finalizedExecutionBlockHash,
|
|
receivedBlock = signedBlock)
|
|
|
|
template callForkChoiceUpdated: auto =
|
|
case self.consensusManager.dag.cfg.consensusForkAtEpoch(
|
|
newHead.get.blck.bid.slot.epoch)
|
|
of ConsensusFork.Deneb, ConsensusFork.Electra:
|
|
# https://github.com/ethereum/execution-apis/blob/90a46e9137c89d58e818e62fa33a0347bba50085/src/engine/prague.md
|
|
# does not define any new forkchoiceUpdated, so reuse V3 from Dencun
|
|
callExpectValidFCU(payloadAttributeType = PayloadAttributesV3)
|
|
of ConsensusFork.Capella:
|
|
callExpectValidFCU(payloadAttributeType = PayloadAttributesV2)
|
|
of ConsensusFork.Phase0, ConsensusFork.Altair,
|
|
ConsensusFork.Bellatrix:
|
|
callExpectValidFCU(payloadAttributeType = PayloadAttributesV1)
|
|
|
|
if self.consensusManager.checkNextProposer(wallSlot).isNone:
|
|
# No attached validator is next proposer, so use non-proposal fcU
|
|
callForkChoiceUpdated()
|
|
else:
|
|
# Some attached validator is next proposer, so prepare payload. As
|
|
# updateHead() updated the DAG head, runProposalForkchoiceUpdated,
|
|
# which needs the state corresponding to that head block, can run.
|
|
if (await self.consensusManager.runProposalForkchoiceUpdated(
|
|
wallSlot)).isNone:
|
|
callForkChoiceUpdated()
|
|
else:
|
|
await self.consensusManager.updateHeadWithExecution(
|
|
newHead.get, self.getBeaconTime)
|
|
else:
|
|
warn "Head selection failed, using previous head",
|
|
head = shortLog(dag.head), wallSlot
|
|
|
|
let
|
|
updateHeadTick = Moment.now()
|
|
queueDur = startTick - queueTick
|
|
newPayloadDur = newPayloadTick - startTick
|
|
addHeadBlockDur = addHeadBlockTick - newPayloadTick
|
|
updateHeadDur = updateHeadTick - addHeadBlockTick
|
|
|
|
# "store block" is the full time it takes to process the block - in the log
|
|
# we split this into execution and consensus timings
|
|
storeBlockDur = newPayloadDur + addHeadBlockDur
|
|
|
|
beacon_store_block_duration_seconds.observe(storeBlockDur.toFloatSeconds())
|
|
|
|
debug "Block processed",
|
|
head = shortLog(dag.head),
|
|
blck = shortLog(blck.get()),
|
|
validationDur, queueDur, newPayloadDur, addHeadBlockDur, updateHeadDur
|
|
|
|
for quarantined in self.consensusManager.quarantine[].pop(blck.get().root):
|
|
# Process the blocks that had the newly accepted block as parent
|
|
debug "Block from quarantine",
|
|
blockRoot = shortLog(signedBlock.root),
|
|
quarantined = shortLog(quarantined.root)
|
|
|
|
withBlck(quarantined):
|
|
when typeof(forkyBlck).kind < ConsensusFork.Deneb:
|
|
self[].enqueueBlock(
|
|
MsgSource.gossip, quarantined, Opt.none(ForkedBlobSidecars))
|
|
else:
|
|
if len(forkyBlck.message.body.blob_kzg_commitments) == 0:
|
|
self[].enqueueBlock(
|
|
MsgSource.gossip, quarantined, Opt.some(ForkedBlobSidecars @[]))
|
|
else:
|
|
if (let res = checkBloblessSignature(self[], forkyBlck); res.isErr):
|
|
warn "Failed to verify signature of unorphaned blobless block",
|
|
blck = shortLog(forkyBlck),
|
|
error = res.error()
|
|
continue
|
|
if self.blobQuarantine[].hasBlobs(forkyBlck):
|
|
let blobs = self.blobQuarantine[]
|
|
.popBlobs(forkyBlck.root, forkyBlck)
|
|
.mapIt(ForkedBlobSidecar.init(it))
|
|
self[].enqueueBlock(MsgSource.gossip, quarantined, Opt.some(blobs))
|
|
else:
|
|
discard self.consensusManager.quarantine[].addBlobless(
|
|
dag.finalizedHead.slot, forkyBlck)
|
|
|
|
ok blck.value()
|
|
|
|
# Enqueue
|
|
# ------------------------------------------------------------------------------
|
|
|
|
proc addBlock*(
|
|
self: var BlockProcessor, src: MsgSource, blck: ForkedSignedBeaconBlock,
|
|
blobs: Opt[ForkedBlobSidecars], maybeFinalized = false,
|
|
validationDur = Duration()
|
|
): Future[Result[void, VerifierError]] {.
|
|
async: (raises: [CancelledError], raw: true).} =
|
|
## Enqueue a Gossip-validated block for consensus verification
|
|
# Backpressure:
|
|
# There is no backpressure here - producers must wait for `resfut` to
|
|
# constrain their own processing
|
|
# Producers:
|
|
# - Gossip (when synced)
|
|
# - SyncManager (during sync)
|
|
# - RequestManager (missing ancestor blocks)
|
|
# - API
|
|
let resfut = newFuture[Result[void, VerifierError]]("BlockProcessor.addBlock")
|
|
enqueueBlock(self, src, blck, blobs, resfut, maybeFinalized, validationDur)
|
|
resfut
|
|
|
|
# Event Loop
|
|
# ------------------------------------------------------------------------------
|
|
|
|
proc processBlock(
|
|
self: ref BlockProcessor, entry: BlockEntry) {.async: (raises: [CancelledError]).} =
|
|
logScope:
|
|
blockRoot = shortLog(entry.blck.root)
|
|
|
|
let
|
|
wallTime = self.getBeaconTime()
|
|
(afterGenesis, _) = wallTime.toSlot()
|
|
|
|
if not afterGenesis:
|
|
error "Processing block before genesis, clock turned back?"
|
|
quit 1
|
|
|
|
let res = withForkyBlckAndBlobs(entry.blck, entry.blobs):
|
|
await self.storeBlock(
|
|
entry.src, wallTime, forkyBlck, forkyBlobs, entry.maybeFinalized,
|
|
entry.queueTick, entry.validationDur)
|
|
|
|
if res.isErr and res.error[1] == ProcessingStatus.notCompleted:
|
|
# When an execution engine returns an error or fails to respond to a
|
|
# payload validity request for some block, a consensus engine:
|
|
# - MUST NOT optimistically import the block.
|
|
# - MUST NOT apply the block to the fork choice store.
|
|
# - MAY queue the block for later processing.
|
|
# https://github.com/ethereum/consensus-specs/blob/v1.5.0-alpha.3/sync/optimistic.md#execution-engine-errors
|
|
await sleepAsync(chronos.seconds(1))
|
|
self[].enqueueBlock(
|
|
entry.src, entry.blck, entry.blobs, entry.resfut, entry.maybeFinalized,
|
|
entry.validationDur)
|
|
# To ensure backpressure on the sync manager, do not complete these futures.
|
|
return
|
|
|
|
if entry.resfut != nil:
|
|
entry.resfut.complete(
|
|
if res.isOk(): Result[void, VerifierError].ok()
|
|
else: Result[void, VerifierError].err(res.error()[0]))
|
|
|
|
proc runQueueProcessingLoop*(self: ref BlockProcessor) {.async.} =
|
|
while true:
|
|
# Cooperative concurrency: one block per loop iteration - because
|
|
# we run both networking and CPU-heavy things like block processing
|
|
# on the same thread, we need to make sure that there is steady progress
|
|
# on the networking side or we get long lockups that lead to timeouts.
|
|
const
|
|
# We cap waiting for an idle slot in case there's a lot of network traffic
|
|
# taking up all CPU - we don't want to _completely_ stop processing blocks
|
|
# in this case - doing so also allows us to benefit from more batching /
|
|
# larger network reads when under load.
|
|
idleTimeout = 10.milliseconds
|
|
|
|
discard await idleAsync().withTimeout(idleTimeout)
|
|
|
|
await self.processBlock(await self[].blockQueue.popFirst())
|