Block accepted reorg aware (#2003)

* block_accepted via adapter is now reorg aware
we skip reconciling the txpool is block is not most work
we skip reconciling the reorg_cache is not a reorg

* rustfmt

* logging tweaks

* rework block_accepted interface

* rustfmt

* rework reorg determination

* introduce BlockStatus to represent next vs reorg vs fork

* rustfmt

* cleanup logging

* log from adapter, even during sync
This commit is contained in:
Antioch Peverell 2018-11-21 14:35:38 +00:00 committed by GitHub
parent 61f1f6103b
commit 66f2545186
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
5 changed files with 119 additions and 55 deletions

View file

@ -38,7 +38,9 @@ use grin_store::Error::NotFoundErr;
use pipe;
use store;
use txhashset;
use types::{ChainAdapter, NoStatus, Options, Tip, TxHashSetRoots, TxHashsetWriteStatus};
use types::{
BlockStatus, ChainAdapter, NoStatus, Options, Tip, TxHashSetRoots, TxHashsetWriteStatus,
};
use util::secp::pedersen::{Commitment, RangeProof};
/// Orphan pool size is limited by MAX_ORPHAN_SIZE
@ -235,22 +237,43 @@ impl Chain {
res
}
fn determine_status(&self, head: Option<Tip>, prev_head: Tip) -> BlockStatus {
// We have more work if the chain head is updated.
let is_more_work = head.is_some();
let mut is_next_block = false;
if let Some(head) = head {
if head.prev_block_h == prev_head.last_block_h {
is_next_block = true;
}
}
match (is_more_work, is_next_block) {
(true, true) => BlockStatus::Next,
(true, false) => BlockStatus::Reorg,
(false, _) => BlockStatus::Fork,
}
}
/// Attempt to add a new block to the chain.
/// Returns true if it has been added to the longest chain
/// or false if it has added to a fork (or orphan?).
fn process_block_single(&self, b: Block, opts: Options) -> Result<Option<Tip>, Error> {
let maybe_new_head: Result<Option<Tip>, Error>;
{
let (maybe_new_head, prev_head) = {
let mut txhashset = self.txhashset.write();
let batch = self.store.batch()?;
let mut ctx = self.new_ctx(opts, batch, &mut txhashset)?;
maybe_new_head = pipe::process_block(&b, &mut ctx);
let prev_head = ctx.batch.head()?;
let maybe_new_head = pipe::process_block(&b, &mut ctx);
if let Ok(_) = maybe_new_head {
ctx.batch.commit()?;
}
// release the lock and let the batch go before post-processing
}
(maybe_new_head, prev_head)
};
let add_to_hash_cache = |hash: Hash| {
// only add to hash cache below if block is definitively accepted
@ -263,8 +286,10 @@ impl Chain {
Ok(head) => {
add_to_hash_cache(b.hash());
let status = self.determine_status(head.clone(), prev_head);
// notifying other parts of the system of the update
self.adapter.block_accepted(&b, opts);
self.adapter.block_accepted(&b, status, opts);
Ok(head)
}
@ -983,8 +1008,7 @@ impl Chain {
if outputs.0 != rangeproofs.0 || outputs.1.len() != rangeproofs.1.len() {
return Err(ErrorKind::TxHashSetErr(String::from(
"Output and rangeproof sets don't match",
))
.into());
)).into());
}
let mut output_vec: Vec<Output> = vec![];
for (ref x, &y) in outputs.1.iter().zip(rangeproofs.1.iter()) {

View file

@ -53,4 +53,4 @@ pub mod types;
pub use chain::{Chain, MAX_ORPHAN_SIZE};
pub use error::{Error, ErrorKind};
pub use store::ChainStore;
pub use types::{ChainAdapter, Options, Tip, TxHashsetWriteStatus};
pub use types::{BlockStatus, ChainAdapter, Options, Tip, TxHashsetWriteStatus};

View file

@ -118,7 +118,7 @@ impl ser::Readable for Tip {
pub trait ChainAdapter {
/// The blockchain pipeline has accepted this block as valid and added
/// it to our chain.
fn block_accepted(&self, b: &Block, opts: Options);
fn block_accepted(&self, block: &Block, status: BlockStatus, opts: Options);
}
/// Inform the caller of the current status of a txhashset write operation,
@ -151,5 +151,17 @@ impl TxHashsetWriteStatus for NoStatus {
pub struct NoopAdapter {}
impl ChainAdapter for NoopAdapter {
fn block_accepted(&self, _: &Block, _: Options) {}
fn block_accepted(&self, _b: &Block, _status: BlockStatus, _opts: Options) {}
}
/// Status of an accepted block.
#[derive(Debug, Clone, PartialEq)]
pub enum BlockStatus {
/// Block is the "next" block, updating the chain head.
Next,
/// Block does not update the chain head and is a fork.
Fork,
/// Block updates the chain head via a (potentially disruptive) "reorg".
/// Previous block was not our previous chain head.
Reorg,
}

View file

@ -22,7 +22,6 @@ use std::sync::Arc;
use util::RwLock;
use chrono::prelude::*;
use chrono::Duration;
use core::core::hash::{Hash, Hashed};
use core::core::id::ShortId;
@ -97,15 +96,6 @@ impl TransactionPool {
debug!("added tx to reorg_cache: size now {}", cache.len());
}
// Old txs will "age out" after 30 mins.
fn truncate_reorg_cache(&mut self, cutoff: DateTime<Utc>) {
let mut cache = self.reorg_cache.write();
while cache.front().map(|x| x.tx_at < cutoff).unwrap_or(false) {
let _ = cache.pop_front();
}
}
fn add_to_txpool(
&mut self,
mut entry: PoolEntry,
@ -179,17 +169,31 @@ impl TransactionPool {
Ok(())
}
fn reconcile_reorg_cache(&mut self, header: &BlockHeader) -> Result<(), PoolError> {
// First "age out" any old txs in the reorg cache.
let cutoff = Utc::now() - Duration::minutes(30);
self.truncate_reorg_cache(cutoff);
// Old txs will "age out" after 30 mins.
pub fn truncate_reorg_cache(&mut self, cutoff: DateTime<Utc>) {
let mut cache = self.reorg_cache.write();
while cache.front().map(|x| x.tx_at < cutoff).unwrap_or(false) {
let _ = cache.pop_front();
}
debug!("truncate_reorg_cache: size: {}", cache.len());
}
pub fn reconcile_reorg_cache(&mut self, header: &BlockHeader) -> Result<(), PoolError> {
let entries = self.reorg_cache.read().iter().cloned().collect::<Vec<_>>();
debug!("reconcile_reorg_cache: size: {} ...", entries.len());
debug!(
"reconcile_reorg_cache: size: {}, block: {:?} ...",
entries.len(),
header.hash(),
);
for entry in entries {
let _ = &self.add_to_txpool(entry.clone(), header);
}
debug!("reconcile_reorg_cache: ... done.");
debug!(
"reconcile_reorg_cache: block: {:?} ... done.",
header.hash()
);
Ok(())
}
@ -200,10 +204,6 @@ impl TransactionPool {
self.txpool.reconcile_block(block);
self.txpool.reconcile(None, &block.header)?;
// Take our "reorg_cache" and see if this block means
// we need to (re)add old txs due to a fork and re-org.
self.reconcile_reorg_cache(&block.header)?;
// Now reconcile our stempool, accounting for the updated txpool txs.
self.stempool.reconcile_block(block);
{

View file

@ -22,8 +22,9 @@ use std::thread;
use std::time::Instant;
use util::RwLock;
use chain::{self, ChainAdapter, Options};
use chrono::prelude::{DateTime, Utc};
use chain::{self, BlockStatus, ChainAdapter, Options};
use chrono::prelude::*;
use chrono::Duration;
use common::types::{self, ChainValidationMode, ServerConfig, SyncState, SyncStatus};
use core::core::hash::{Hash, Hashed};
use core::core::transaction::Transaction;
@ -182,7 +183,7 @@ impl p2p::ChainAdapter for NetToChainAdapter {
.validate(&prev.total_kernel_offset, self.verifier_cache.clone())
.is_ok()
{
debug!("adapter: successfully hydrated block from tx pool!");
debug!("successfully hydrated block from tx pool!");
self.process_block(block, addr)
} else {
if self.sync_state.status() == SyncStatus::NoSync {
@ -190,14 +191,12 @@ impl p2p::ChainAdapter for NetToChainAdapter {
self.request_block(&cb.header, &addr);
true
} else {
debug!(
"adapter: block invalid after hydration, ignoring it, cause still syncing"
);
debug!("block invalid after hydration, ignoring it, cause still syncing");
true
}
}
} else {
debug!("adapter: failed to retrieve previous block header (still syncing?)");
debug!("failed to retrieve previous block header (still syncing?)");
true
}
}
@ -469,10 +468,7 @@ impl NetToChainAdapter {
true
}
Err(ref e) if e.is_bad_data() => {
debug!(
"adapter: process_block: {} is a bad block, resetting head",
bhash
);
debug!("process_block: {} is a bad block, resetting head", bhash);
let _ = self.chain().reset_head();
// we potentially changed the state of the system here
@ -489,7 +485,7 @@ impl NetToChainAdapter {
if !self.chain().is_orphan(&previous.hash())
&& !self.sync_state.is_syncing()
{
debug!("adapter: process_block: received an orphan block, checking the parent: {:}", previous.hash());
debug!("process_block: received an orphan block, checking the parent: {:}", previous.hash());
self.request_block_by_hash(previous.hash(), &addr)
}
}
@ -497,7 +493,7 @@ impl NetToChainAdapter {
}
_ => {
debug!(
"adapter: process_block: block {} refused by chain: {}",
"process_block: block {} refused by chain: {}",
bhash,
e.kind()
);
@ -521,7 +517,7 @@ impl NetToChainAdapter {
let now = Instant::now();
debug!(
"adapter: process_block: ***** validating full chain state at {}",
"process_block: ***** validating full chain state at {}",
bhash,
);
@ -530,7 +526,7 @@ impl NetToChainAdapter {
.expect("chain validation failed, hard stop");
debug!(
"adapter: process_block: ***** done validating full chain state, took {}s",
"process_block: ***** done validating full chain state, took {}s",
now.elapsed().as_secs(),
);
}
@ -644,13 +640,38 @@ pub struct ChainToPoolAndNetAdapter {
}
impl ChainAdapter for ChainToPoolAndNetAdapter {
fn block_accepted(&self, b: &core::Block, opts: Options) {
fn block_accepted(&self, b: &core::Block, status: BlockStatus, opts: Options) {
match status {
BlockStatus::Reorg => {
warn!(
"block_accepted (REORG!): {:?} at {} (diff: {})",
b.hash(),
b.header.height,
b.header.total_difficulty(),
);
}
BlockStatus::Fork => {
debug!(
"block_accepted (fork?): {:?} at {} (diff: {})",
b.hash(),
b.header.height,
b.header.total_difficulty(),
);
}
BlockStatus::Next => {
debug!(
"block_accepted (head+): {:?} at {} (diff: {})",
b.hash(),
b.header.height,
b.header.total_difficulty(),
);
}
}
if self.sync_state.is_syncing() {
return;
}
debug!("adapter: block_accepted: {:?}", b.hash());
// If we mined the block then we want to broadcast the compact block.
// If we received the block from another node then broadcast "header first"
// to minimize network traffic.
@ -665,12 +686,19 @@ impl ChainAdapter for ChainToPoolAndNetAdapter {
// Reconcile the txpool against the new block *after* we have broadcast it too our peers.
// This may be slow and we do not want to delay block propagation.
if let Err(e) = self.tx_pool.write().reconcile_block(b) {
error!(
"Pool could not update itself at block {}: {:?}",
b.hash(),
e,
);
// We only want to reconcile the txpool against the new block *if* total work has increased.
if status == BlockStatus::Next || status == BlockStatus::Reorg {
let mut tx_pool = self.tx_pool.write();
let _ = tx_pool.reconcile_block(b);
// First "age out" any old txs in the reorg_cache.
let cutoff = Utc::now() - Duration::minutes(30);
tx_pool.truncate_reorg_cache(cutoff);
}
if status == BlockStatus::Reorg {
let _ = self.tx_pool.write().reconcile_reorg_cache(&b.header);
}
}
}