miscelaneous fixes on syncing, rpc and responding to peer's sync related requests (#3827)

- there was a bug in responding range blob requests where we would incorrectly label the first slot of an epoch as a non-skipped slot if it were skipped. this bug did not exist in the code for responding to block range request because the logic error was mitigated by defensive coding elsewhere
- there was a bug where a block received during range sync without a corresponding blob (and vice versa) was incorrectly interpreted as a stream termination
- RPC size limit fixes.
- Our blob cache was dead locking so I removed use of it for now.
- Because of our change in finalized sync batch size from 2 to 1 and our transition to using exact epoch boundaries for batches (rather than one slot past the epoch boundary), we need to sync finalized sync to 2 epochs + 1 slot past our peer's finalized slot in order to finalize the chain locally.
- use fork context bytes in rpc methods on both the server and client side
This commit is contained in:
realbigsean 2022-12-21 15:50:51 -05:00 committed by GitHub
parent cc420caaa5
commit 33d01a7911
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
6 changed files with 78 additions and 20 deletions

View File

@ -298,8 +298,8 @@ impl<TSpec: EthSpec> Decoder for SSZSnappyOutboundCodec<TSpec> {
.rpc_response_limits::<TSpec>(&self.fork_context); .rpc_response_limits::<TSpec>(&self.fork_context);
if ssz_limits.is_out_of_bounds(length, self.max_packet_size) { if ssz_limits.is_out_of_bounds(length, self.max_packet_size) {
return Err(RPCError::InvalidData(format!( return Err(RPCError::InvalidData(format!(
"RPC response length is out of bounds, length {}", "RPC response length is out of bounds, length {}, max {}, min {}",
length length, ssz_limits.max, ssz_limits.min
))); )));
} }
// Calculate worst case compression length for given uncompressed length // Calculate worst case compression length for given uncompressed length
@ -439,6 +439,9 @@ fn context_bytes<T: EthSpec>(
SignedBeaconBlock::Base { .. } => Some(fork_context.genesis_context_bytes()), SignedBeaconBlock::Base { .. } => Some(fork_context.genesis_context_bytes()),
}; };
} }
if let RPCResponse::BlobsByRange(_) | RPCResponse::BlobsByRoot(_) = rpc_variant {
return fork_context.to_context_bytes(ForkName::Eip4844);
}
} }
} }
None None

View File

@ -107,6 +107,12 @@ lazy_static! {
.as_ssz_bytes() .as_ssz_bytes()
.len(); .len();
pub static ref BLOBS_SIDECAR_MIN: usize = BlobsSidecar::<MainnetEthSpec>::empty().as_ssz_bytes().len();
pub static ref BLOBS_SIDECAR_MAX: usize = BlobsSidecar::<MainnetEthSpec>::max_size();
//FIXME(sean) these are underestimates
pub static ref SIGNED_BLOCK_AND_BLOBS_MIN: usize = *BLOBS_SIDECAR_MIN + *SIGNED_BEACON_BLOCK_BASE_MIN;
pub static ref SIGNED_BLOCK_AND_BLOBS_MAX: usize =*BLOBS_SIDECAR_MAX + *SIGNED_BEACON_BLOCK_EIP4844_MAX;
} }
/// The maximum bytes that can be sent across the RPC pre-merge. /// The maximum bytes that can be sent across the RPC pre-merge.
@ -358,11 +364,10 @@ impl ProtocolId {
Protocol::Goodbye => RpcLimits::new(0, 0), // Goodbye request has no response Protocol::Goodbye => RpcLimits::new(0, 0), // Goodbye request has no response
Protocol::BlocksByRange => rpc_block_limits_by_fork(fork_context.current_fork()), Protocol::BlocksByRange => rpc_block_limits_by_fork(fork_context.current_fork()),
Protocol::BlocksByRoot => rpc_block_limits_by_fork(fork_context.current_fork()), Protocol::BlocksByRoot => rpc_block_limits_by_fork(fork_context.current_fork()),
Protocol::BlobsByRange => RpcLimits::new(*BLOBS_SIDECAR_MIN, *BLOBS_SIDECAR_MAX),
//FIXME(sean) add blob sizes Protocol::BlobsByRoot => {
Protocol::BlobsByRange => rpc_block_limits_by_fork(fork_context.current_fork()), RpcLimits::new(*SIGNED_BLOCK_AND_BLOBS_MIN, *SIGNED_BLOCK_AND_BLOBS_MAX)
Protocol::BlobsByRoot => rpc_block_limits_by_fork(fork_context.current_fork()), }
Protocol::Ping => RpcLimits::new( Protocol::Ping => RpcLimits::new(
<Ping as Encode>::ssz_fixed_len(), <Ping as Encode>::ssz_fixed_len(),
<Ping as Encode>::ssz_fixed_len(), <Ping as Encode>::ssz_fixed_len(),
@ -381,14 +386,17 @@ impl ProtocolId {
/// Returns `true` if the given `ProtocolId` should expect `context_bytes` in the /// Returns `true` if the given `ProtocolId` should expect `context_bytes` in the
/// beginning of the stream, else returns `false`. /// beginning of the stream, else returns `false`.
pub fn has_context_bytes(&self) -> bool { pub fn has_context_bytes(&self) -> bool {
if self.version == Version::V2 { match self.version {
match self.message_name { Version::V2 => match self.message_name {
Protocol::BlocksByRange | Protocol::BlocksByRoot => return true, Protocol::BlocksByRange | Protocol::BlocksByRoot => return true,
_ => return false, _ => return false,
},
Version::V1 => match self.message_name {
Protocol::BlobsByRange | Protocol::BlobsByRoot => return true,
_ => return false,
},
} }
} }
false
}
} }
/// An RPC protocol ID. /// An RPC protocol ID.

View File

@ -433,7 +433,17 @@ impl<T: BeaconChainTypes> Worker<T> {
}; };
// Pick out the required blocks, ignoring skip-slots. // Pick out the required blocks, ignoring skip-slots.
let mut last_block_root = None; let mut last_block_root = req
.start_slot
.checked_sub(1)
.map(|prev_slot| {
self.chain
.block_root_at_slot(Slot::new(prev_slot), WhenSlotSkipped::Prev)
})
.transpose()
.ok()
.flatten()
.flatten();
let maybe_block_roots = process_results(forwards_block_root_iter, |iter| { let maybe_block_roots = process_results(forwards_block_root_iter, |iter| {
iter.take_while(|(_, slot)| slot.as_u64() < req.start_slot.saturating_add(req.count)) iter.take_while(|(_, slot)| slot.as_u64() < req.start_slot.saturating_add(req.count))
// map skip slots to None // map skip slots to None
@ -602,7 +612,17 @@ impl<T: BeaconChainTypes> Worker<T> {
}; };
// Pick out the required blocks, ignoring skip-slots. // Pick out the required blocks, ignoring skip-slots.
let mut last_block_root = None; let mut last_block_root = req
.start_slot
.checked_sub(1)
.map(|prev_slot| {
self.chain
.block_root_at_slot(Slot::new(prev_slot), WhenSlotSkipped::Prev)
})
.transpose()
.ok()
.flatten()
.flatten();
let maybe_block_roots = process_results(forwards_block_root_iter, |iter| { let maybe_block_roots = process_results(forwards_block_root_iter, |iter| {
iter.take_while(|(_, slot)| slot.as_u64() < req.start_slot.saturating_add(req.count)) iter.take_while(|(_, slot)| slot.as_u64() < req.start_slot.saturating_add(req.count))
// map skip slots to None // map skip slots to None
@ -669,7 +689,7 @@ impl<T: BeaconChainTypes> Worker<T> {
self.log, self.log,
"BlobsByRange Response processed"; "BlobsByRange Response processed";
"peer" => %peer_id, "peer" => %peer_id,
"msg" => "Failed to return all requested blocks", "msg" => "Failed to return all requested blobs",
"start_slot" => req.start_slot, "start_slot" => req.start_slot,
"current_slot" => current_slot, "current_slot" => current_slot,
"requested" => req.count, "requested" => req.count,

View File

@ -314,14 +314,24 @@ impl<T: BeaconChainTypes> SyncNetworkContext<T> {
let (chain_id, batch_id, info) = entry.get_mut(); let (chain_id, batch_id, info) = entry.get_mut();
let chain_id = chain_id.clone(); let chain_id = chain_id.clone();
let batch_id = batch_id.clone(); let batch_id = batch_id.clone();
let stream_terminator = maybe_block.is_none();
info.add_block_response(maybe_block); info.add_block_response(maybe_block);
let maybe_block = info.pop_response().map(|block_sidecar_pair| { let maybe_block_wrapped = info.pop_response().map(|block_sidecar_pair| {
BlockWrapper::BlockAndBlob { block_sidecar_pair } BlockWrapper::BlockAndBlob { block_sidecar_pair }
}); });
if stream_terminator && !info.is_finished() {
return None;
}
if !stream_terminator && maybe_block_wrapped.is_none() {
return None;
}
if info.is_finished() { if info.is_finished() {
entry.remove(); entry.remove();
} }
Some((chain_id, batch_id, maybe_block))
Some((chain_id, batch_id, maybe_block_wrapped))
} }
Entry::Vacant(_) => None, Entry::Vacant(_) => None,
} }
@ -356,13 +366,24 @@ impl<T: BeaconChainTypes> SyncNetworkContext<T> {
let (chain_id, batch_id, info) = entry.get_mut(); let (chain_id, batch_id, info) = entry.get_mut();
let chain_id = chain_id.clone(); let chain_id = chain_id.clone();
let batch_id = batch_id.clone(); let batch_id = batch_id.clone();
let stream_terminator = maybe_sidecar.is_none();
info.add_sidecar_response(maybe_sidecar); info.add_sidecar_response(maybe_sidecar);
let maybe_block = info let maybe_block = info
.pop_response() .pop_response()
.map(|block_sidecar_pair| BlockWrapper::BlockAndBlob { block_sidecar_pair }); .map(|block_sidecar_pair| BlockWrapper::BlockAndBlob { block_sidecar_pair });
if stream_terminator && !info.is_finished() {
return None;
}
if !stream_terminator && maybe_block.is_none() {
return None;
}
if info.is_finished() { if info.is_finished() {
entry.remove(); entry.remove();
} }
Some((chain_id, batch_id, maybe_block)) Some((chain_id, batch_id, maybe_block))
} }
Entry::Vacant(_) => None, Entry::Vacant(_) => None,

View File

@ -137,10 +137,16 @@ impl<T: BeaconChainTypes> SyncingChain<T> {
let id = SyncingChain::<T>::id(&target_head_root, &target_head_slot); let id = SyncingChain::<T>::id(&target_head_root, &target_head_slot);
let target_slot = if is_finalized_segment {
target_head_slot + (2 * T::EthSpec::slots_per_epoch()) + 1
} else {
target_head_slot
};
SyncingChain { SyncingChain {
id, id,
start_epoch, start_epoch,
target_head_slot, target_head_slot: target_slot,
target_head_root, target_head_root,
batches: BTreeMap::new(), batches: BTreeMap::new(),
peers, peers,

View File

@ -503,9 +503,9 @@ impl<E: EthSpec, Hot: ItemStore<E>, Cold: ItemStore<E>> HotColdDB<E, Hot, Cold>
} }
pub fn get_blobs(&self, block_root: &Hash256) -> Result<Option<BlobsSidecar<E>>, Error> { pub fn get_blobs(&self, block_root: &Hash256) -> Result<Option<BlobsSidecar<E>>, Error> {
if let Some(blobs) = self.blob_cache.lock().get(block_root) { // FIXME(sean) I was attempting to use a blob cache here but was getting deadlocks,
Ok(Some(blobs.clone())) // may want to attempt to use one again
} else if let Some(bytes) = self if let Some(bytes) = self
.hot_db .hot_db
.get_bytes(DBColumn::BeaconBlob.into(), block_root.as_bytes())? .get_bytes(DBColumn::BeaconBlob.into(), block_root.as_bytes())?
{ {