prysm-pulse/beacon-chain/blockchain/process_block.go
terence tsao d3475a563d
Remove Canonical Attestations From Pool Correctly (#9444)
* Remove canonical attestaitons

* Rm debug log

* Remove deleted test

* Tests and fmt

* Tests

* Bug fix flag

* Move feature flag

* Go fmt

Co-authored-by: prylabs-bulldozer[bot] <58059840+prylabs-bulldozer[bot]@users.noreply.github.com>
2021-08-24 00:33:17 +00:00

450 lines
16 KiB
Go

package blockchain
import (
"context"
"fmt"
"time"
"github.com/pkg/errors"
"github.com/prysmaticlabs/prysm/beacon-chain/core/feed"
statefeed "github.com/prysmaticlabs/prysm/beacon-chain/core/feed/state"
"github.com/prysmaticlabs/prysm/beacon-chain/core/helpers"
core "github.com/prysmaticlabs/prysm/beacon-chain/core/state"
"github.com/prysmaticlabs/prysm/beacon-chain/state"
ethpbv1 "github.com/prysmaticlabs/prysm/proto/eth/v1"
ethpb "github.com/prysmaticlabs/prysm/proto/prysm/v1alpha1"
"github.com/prysmaticlabs/prysm/proto/prysm/v1alpha1/block"
"github.com/prysmaticlabs/prysm/shared/attestationutil"
"github.com/prysmaticlabs/prysm/shared/bls"
"github.com/prysmaticlabs/prysm/shared/bytesutil"
"github.com/prysmaticlabs/prysm/shared/featureconfig"
"github.com/prysmaticlabs/prysm/shared/params"
"go.opencensus.io/trace"
)
// A custom slot deadline for processing state slots in our cache.
const slotDeadline = 5 * time.Second
// A custom deadline for deposit trie insertion.
const depositDeadline = 20 * time.Second
// This defines size of the upper bound for initial sync block cache.
var initialSyncBlockCacheSize = uint64(2 * params.BeaconConfig().SlotsPerEpoch)
// onBlock is called when a gossip block is received. It runs regular state transition on the block.
// The block's signing root should be computed before calling this method to avoid redundant
// computation in this method and methods it calls into.
//
// Spec pseudocode definition:
// def on_block(store: Store, signed_block: SignedBeaconBlock) -> None:
// block = signed_block.message
// # Parent block must be known
// assert block.parent_root in store.block_states
// # Make a copy of the state to avoid mutability issues
// pre_state = copy(store.block_states[block.parent_root])
// # Blocks cannot be in the future. If they are, their consideration must be delayed until the are in the past.
// assert get_current_slot(store) >= block.slot
//
// # Check that block is later than the finalized epoch slot (optimization to reduce calls to get_ancestor)
// finalized_slot = compute_start_slot_at_epoch(store.finalized_checkpoint.epoch)
// assert block.slot > finalized_slot
// # Check block is a descendant of the finalized block at the checkpoint finalized slot
// assert get_ancestor(store, block.parent_root, finalized_slot) == store.finalized_checkpoint.root
//
// # Check the block is valid and compute the post-state
// state = pre_state.copy()
// state_transition(state, signed_block, True)
// # Add new block to the store
// store.blocks[hash_tree_root(block)] = block
// # Add new state for this block to the store
// store.block_states[hash_tree_root(block)] = state
//
// # Update justified checkpoint
// if state.current_justified_checkpoint.epoch > store.justified_checkpoint.epoch:
// if state.current_justified_checkpoint.epoch > store.best_justified_checkpoint.epoch:
// store.best_justified_checkpoint = state.current_justified_checkpoint
// if should_update_justified_checkpoint(store, state.current_justified_checkpoint):
// store.justified_checkpoint = state.current_justified_checkpoint
//
// # Update finalized checkpoint
// if state.finalized_checkpoint.epoch > store.finalized_checkpoint.epoch:
// store.finalized_checkpoint = state.finalized_checkpoint
//
// # Potentially update justified if different from store
// if store.justified_checkpoint != state.current_justified_checkpoint:
// # Update justified if new justified is later than store justified
// if state.current_justified_checkpoint.epoch > store.justified_checkpoint.epoch:
// store.justified_checkpoint = state.current_justified_checkpoint
// return
//
// # Update justified if store justified is not in chain with finalized checkpoint
// finalized_slot = compute_start_slot_at_epoch(store.finalized_checkpoint.epoch)
// ancestor_at_finalized_slot = get_ancestor(store, store.justified_checkpoint.root, finalized_slot)
// if ancestor_at_finalized_slot != store.finalized_checkpoint.root:
// store.justified_checkpoint = state.current_justified_checkpoint
func (s *Service) onBlock(ctx context.Context, signed block.SignedBeaconBlock, blockRoot [32]byte) error {
ctx, span := trace.StartSpan(ctx, "blockChain.onBlock")
defer span.End()
if signed == nil || signed.IsNil() || signed.Block().IsNil() {
return errors.New("nil block")
}
b := signed.Block()
preState, err := s.getBlockPreState(ctx, b)
if err != nil {
return err
}
postState, err := core.ExecuteStateTransition(ctx, preState, signed)
if err != nil {
return err
}
if err := s.savePostStateInfo(ctx, blockRoot, signed, postState, false /* reg sync */); err != nil {
return err
}
// Updating next slot state cache can happen in the background. It shouldn't block rest of the process.
if featureconfig.Get().EnableNextSlotStateCache {
go func() {
// Use a custom deadline here, since this method runs asynchronously.
// We ignore the parent method's context and instead create a new one
// with a custom deadline, therefore using the background context instead.
slotCtx, cancel := context.WithTimeout(context.Background(), slotDeadline)
defer cancel()
if err := core.UpdateNextSlotCache(slotCtx, blockRoot[:], postState); err != nil {
log.WithError(err).Debug("could not update next slot state cache")
}
}()
}
// Update justified check point.
if postState.CurrentJustifiedCheckpoint().Epoch > s.justifiedCheckpt.Epoch {
if err := s.updateJustified(ctx, postState); err != nil {
return err
}
}
newFinalized := postState.FinalizedCheckpointEpoch() > s.finalizedCheckpt.Epoch
if featureconfig.Get().UpdateHeadTimely {
if newFinalized {
if err := s.finalizedImpliesNewJustified(ctx, postState); err != nil {
return errors.Wrap(err, "could not save new justified")
}
s.prevFinalizedCheckpt = s.finalizedCheckpt
s.finalizedCheckpt = postState.FinalizedCheckpoint()
}
if err := s.updateHead(ctx, s.getJustifiedBalances()); err != nil {
log.WithError(err).Warn("Could not update head")
}
if err := s.pruneCanonicalAttsFromPool(ctx, blockRoot, signed); err != nil {
return err
}
// Send notification of the processed block to the state feed.
s.cfg.StateNotifier.StateFeed().Send(&feed.Event{
Type: statefeed.BlockProcessed,
Data: &statefeed.BlockProcessedData{
Slot: signed.Block().Slot(),
BlockRoot: blockRoot,
SignedBlock: signed,
Verified: true,
},
})
}
// Update finalized check point.
if newFinalized {
if err := s.updateFinalized(ctx, postState.FinalizedCheckpoint()); err != nil {
return err
}
fRoot := bytesutil.ToBytes32(postState.FinalizedCheckpoint().Root)
if err := s.cfg.ForkChoiceStore.Prune(ctx, fRoot); err != nil {
return errors.Wrap(err, "could not prune proto array fork choice nodes")
}
if !featureconfig.Get().UpdateHeadTimely {
if err := s.finalizedImpliesNewJustified(ctx, postState); err != nil {
return errors.Wrap(err, "could not save new justified")
}
}
go func() {
// Send an event regarding the new finalized checkpoint over a common event feed.
s.cfg.StateNotifier.StateFeed().Send(&feed.Event{
Type: statefeed.FinalizedCheckpoint,
Data: &ethpbv1.EventFinalizedCheckpoint{
Epoch: postState.FinalizedCheckpoint().Epoch,
Block: postState.FinalizedCheckpoint().Root,
State: signed.Block().StateRoot(),
},
})
// Use a custom deadline here, since this method runs asynchronously.
// We ignore the parent method's context and instead create a new one
// with a custom deadline, therefore using the background context instead.
depCtx, cancel := context.WithTimeout(context.Background(), depositDeadline)
defer cancel()
if err := s.insertFinalizedDeposits(depCtx, fRoot); err != nil {
log.WithError(err).Error("Could not insert finalized deposits.")
}
}()
}
defer reportAttestationInclusion(b)
return s.handleEpochBoundary(ctx, postState)
}
func (s *Service) onBlockBatch(ctx context.Context, blks []block.SignedBeaconBlock,
blockRoots [][32]byte) ([]*ethpb.Checkpoint, []*ethpb.Checkpoint, error) {
ctx, span := trace.StartSpan(ctx, "blockChain.onBlockBatch")
defer span.End()
if len(blks) == 0 || len(blockRoots) == 0 {
return nil, nil, errors.New("no blocks provided")
}
if blks[0] == nil || blks[0].IsNil() || blks[0].Block().IsNil() {
return nil, nil, errors.New("nil block")
}
b := blks[0].Block()
// Retrieve incoming block's pre state.
if err := s.verifyBlkPreState(ctx, b); err != nil {
return nil, nil, err
}
preState, err := s.cfg.StateGen.StateByRootInitialSync(ctx, bytesutil.ToBytes32(b.ParentRoot()))
if err != nil {
return nil, nil, err
}
if preState == nil || preState.IsNil() {
return nil, nil, fmt.Errorf("nil pre state for slot %d", b.Slot())
}
jCheckpoints := make([]*ethpb.Checkpoint, len(blks))
fCheckpoints := make([]*ethpb.Checkpoint, len(blks))
sigSet := &bls.SignatureSet{
Signatures: [][]byte{},
PublicKeys: []bls.PublicKey{},
Messages: [][32]byte{},
}
var set *bls.SignatureSet
boundaries := make(map[[32]byte]state.BeaconState)
for i, b := range blks {
set, preState, err = core.ExecuteStateTransitionNoVerifyAnySig(ctx, preState, b)
if err != nil {
return nil, nil, err
}
// Save potential boundary states.
if helpers.IsEpochStart(preState.Slot()) {
boundaries[blockRoots[i]] = preState.Copy()
if err := s.handleEpochBoundary(ctx, preState); err != nil {
return nil, nil, errors.Wrap(err, "could not handle epoch boundary state")
}
}
jCheckpoints[i] = preState.CurrentJustifiedCheckpoint()
fCheckpoints[i] = preState.FinalizedCheckpoint()
sigSet.Join(set)
}
verify, err := sigSet.Verify()
if err != nil {
return nil, nil, err
}
if !verify {
return nil, nil, errors.New("batch block signature verification failed")
}
for r, st := range boundaries {
if err := s.cfg.StateGen.SaveState(ctx, r, st); err != nil {
return nil, nil, err
}
}
// Also saves the last post state which to be used as pre state for the next batch.
lastB := blks[len(blks)-1]
lastBR := blockRoots[len(blockRoots)-1]
if err := s.cfg.StateGen.SaveState(ctx, lastBR, preState); err != nil {
return nil, nil, err
}
if err := s.saveHeadNoDB(ctx, lastB, lastBR, preState); err != nil {
return nil, nil, err
}
return fCheckpoints, jCheckpoints, nil
}
// handles a block after the block's batch has been verified, where we can save blocks
// their state summaries and split them off to relative hot/cold storage.
func (s *Service) handleBlockAfterBatchVerify(ctx context.Context, signed block.SignedBeaconBlock,
blockRoot [32]byte, fCheckpoint, jCheckpoint *ethpb.Checkpoint) error {
b := signed.Block()
s.saveInitSyncBlock(blockRoot, signed)
if err := s.insertBlockToForkChoiceStore(ctx, b, blockRoot, fCheckpoint, jCheckpoint); err != nil {
return err
}
if err := s.cfg.BeaconDB.SaveStateSummary(ctx, &ethpb.StateSummary{
Slot: signed.Block().Slot(),
Root: blockRoot[:],
}); err != nil {
return err
}
// Rate limit how many blocks (2 epochs worth of blocks) a node keeps in the memory.
if uint64(len(s.getInitSyncBlocks())) > initialSyncBlockCacheSize {
if err := s.cfg.BeaconDB.SaveBlocks(ctx, s.getInitSyncBlocks()); err != nil {
return err
}
s.clearInitSyncBlocks()
}
if jCheckpoint.Epoch > s.justifiedCheckpt.Epoch {
if err := s.updateJustifiedInitSync(ctx, jCheckpoint); err != nil {
return err
}
}
// Update finalized check point. Prune the block cache and helper caches on every new finalized epoch.
if fCheckpoint.Epoch > s.finalizedCheckpt.Epoch {
if err := s.updateFinalized(ctx, fCheckpoint); err != nil {
return err
}
if featureconfig.Get().UpdateHeadTimely {
s.prevFinalizedCheckpt = s.finalizedCheckpt
s.finalizedCheckpt = fCheckpoint
}
}
return nil
}
// Epoch boundary bookkeeping such as logging epoch summaries.
func (s *Service) handleEpochBoundary(ctx context.Context, postState state.BeaconState) error {
ctx, span := trace.StartSpan(ctx, "blockChain.handleEpochBoundary")
defer span.End()
if postState.Slot()+1 == s.nextEpochBoundarySlot {
// Update caches for the next epoch at epoch boundary slot - 1.
if err := helpers.UpdateCommitteeCache(postState, helpers.NextEpoch(postState)); err != nil {
return err
}
copied := postState.Copy()
copied, err := core.ProcessSlots(ctx, copied, copied.Slot()+1)
if err != nil {
return err
}
if err := helpers.UpdateProposerIndicesInCache(copied); err != nil {
return err
}
} else if postState.Slot() >= s.nextEpochBoundarySlot {
if err := reportEpochMetrics(ctx, postState, s.head.state); err != nil {
return err
}
var err error
s.nextEpochBoundarySlot, err = helpers.StartSlot(helpers.NextEpoch(postState))
if err != nil {
return err
}
// Update caches at epoch boundary slot.
// The following updates have short cut to return nil cheaply if fulfilled during boundary slot - 1.
if err := helpers.UpdateCommitteeCache(postState, helpers.CurrentEpoch(postState)); err != nil {
return err
}
if err := helpers.UpdateProposerIndicesInCache(postState); err != nil {
return err
}
}
return nil
}
// This feeds in the block and block's attestations to fork choice store. It's allows fork choice store
// to gain information on the most current chain.
func (s *Service) insertBlockAndAttestationsToForkChoiceStore(ctx context.Context, blk block.BeaconBlock, root [32]byte,
st state.BeaconState) error {
ctx, span := trace.StartSpan(ctx, "blockChain.insertBlockAndAttestationsToForkChoiceStore")
defer span.End()
fCheckpoint := st.FinalizedCheckpoint()
jCheckpoint := st.CurrentJustifiedCheckpoint()
if err := s.insertBlockToForkChoiceStore(ctx, blk, root, fCheckpoint, jCheckpoint); err != nil {
return err
}
// Feed in block's attestations to fork choice store.
for _, a := range blk.Body().Attestations() {
committee, err := helpers.BeaconCommitteeFromState(st, a.Data.Slot, a.Data.CommitteeIndex)
if err != nil {
return err
}
indices, err := attestationutil.AttestingIndices(a.AggregationBits, committee)
if err != nil {
return err
}
s.cfg.ForkChoiceStore.ProcessAttestation(ctx, indices, bytesutil.ToBytes32(a.Data.BeaconBlockRoot), a.Data.Target.Epoch)
}
return nil
}
func (s *Service) insertBlockToForkChoiceStore(ctx context.Context, blk block.BeaconBlock,
root [32]byte, fCheckpoint, jCheckpoint *ethpb.Checkpoint) error {
if err := s.fillInForkChoiceMissingBlocks(ctx, blk, fCheckpoint, jCheckpoint); err != nil {
return err
}
// Feed in block to fork choice store.
if err := s.cfg.ForkChoiceStore.ProcessBlock(ctx,
blk.Slot(), root, bytesutil.ToBytes32(blk.ParentRoot()), bytesutil.ToBytes32(blk.Body().Graffiti()),
jCheckpoint.Epoch,
fCheckpoint.Epoch); err != nil {
return errors.Wrap(err, "could not process block for proto array fork choice")
}
return nil
}
// This saves post state info to DB or cache. This also saves post state info to fork choice store.
// Post state info consists of processed block and state. Do not call this method unless the block and state are verified.
func (s *Service) savePostStateInfo(ctx context.Context, r [32]byte, b block.SignedBeaconBlock, st state.BeaconState, initSync bool) error {
ctx, span := trace.StartSpan(ctx, "blockChain.savePostStateInfo")
defer span.End()
if initSync {
s.saveInitSyncBlock(r, b)
} else if err := s.cfg.BeaconDB.SaveBlock(ctx, b); err != nil {
return errors.Wrapf(err, "could not save block from slot %d", b.Block().Slot())
}
if err := s.cfg.StateGen.SaveState(ctx, r, st); err != nil {
return errors.Wrap(err, "could not save state")
}
if err := s.insertBlockAndAttestationsToForkChoiceStore(ctx, b.Block(), r, st); err != nil {
return errors.Wrapf(err, "could not insert block %d to fork choice store", b.Block().Slot())
}
return nil
}
// This removes the attestations from the mem pool. It will only remove the attestations if input root `r` is canonical,
// meaning the block `b` is part of the canonical chain.
func (s *Service) pruneCanonicalAttsFromPool(ctx context.Context, r [32]byte, b block.SignedBeaconBlock) error {
if !featureconfig.Get().CorrectlyPruneCanonicalAtts {
return nil
}
canonical, err := s.IsCanonical(ctx, r)
if err != nil {
return err
}
if !canonical {
return nil
}
atts := b.Block().Body().Attestations()
for _, att := range atts {
if helpers.IsAggregated(att) {
if err := s.cfg.AttPool.DeleteAggregatedAttestation(att); err != nil {
return err
}
} else {
if err := s.cfg.AttPool.DeleteUnaggregatedAttestation(att); err != nil {
return err
}
}
}
return nil
}