Files
pezkuwi-subxt/substrate/core/consensus/babe/src/lib.rs
T
André Silva cb7527d2b2 babe: Introduce secondary slots (Aurababeous) (#3380)
* babe: initial implementation of secondary slots

* babe: validate secondary slot author

* babe: implement weight based fork choice

* babe: remove unused

* aura: cleanup unused imports

* babe: pass in parent weight when authoring and verifying

* babe: use epoch randomness for picking secondary slot authors

* babe: fix tests

* babe: fix wasm build

* babe: node-side code for disabling secondary slots

* babe: allow enabling/disabling secondary slots from runtime

* babe: fix test

* babe: use blake2_256 for secondary slot assignment

* babe: run block initialization in should_end_session

* node: increase slot duration to 6s

* babe: add docs

* node: bump spec_version

* Apply suggestions from code review

Co-Authored-By: joe petrowski <25483142+joepetrowski@users.noreply.github.com>

* babe: simplify secondary slot assignment calculation

* babe: remove unnecessary comment

* node: bump spec_version

* babe: fix bad merge
2019-08-16 15:47:53 +02:00

1570 lines
48 KiB
Rust

// Copyright 2019 Parity Technologies (UK) Ltd.
// This file is part of Substrate.
// Substrate is free software: you can redistribute it and/or modify
// it under the terms of the GNU General Public License as published by
// the Free Software Foundation, either version 3 of the License, or
// (at your option) any later version.
// Substrate is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
// GNU General Public License for more details.
// You should have received a copy of the GNU General Public License
// along with Substrate. If not, see <http://www.gnu.org/licenses/>.
//! # BABE (Blind Assignment for Blockchain Extension)
//!
//! BABE is a slot-based block production mechanism which uses a VRF PRNG to
//! randomly perform the slot allocation. On every slot, all the authorities
//! generate a new random number with the VRF function and if it is lower than a
//! given threshold (which is proportional to their weight/stake) they have a
//! right to produce a block. The proof of the VRF function execution will be
//! used by other peer to validate the legitimacy of the slot claim.
//!
//! The engine is also responsible for collecting entropy on-chain which will be
//! used to seed the given VRF PRNG. An epoch is a contiguous number of slots
//! under which we will be using the same authority set. During an epoch all VRF
//! outputs produced as a result of block production will be collected on an
//! on-chain randomness pool. Epoch changes are announced one epoch in advance,
//! i.e. when ending epoch N, we announce the parameters (randomness,
//! authorities, etc.) for epoch N+2.
//!
//! Since the slot assignment is randomized, it is possible that a slot is
//! assigned to multiple validators in which case we will have a temporary fork,
//! or that a slot is assigned to no validator in which case no block is
//! produced. Which means that block times are not deterministic.
//!
//! The protocol has a parameter `c` [0, 1] for which `1 - c` is the probability
//! of a slot being empty. The choice of this parameter affects the security of
//! the protocol relating to maximum tolerable network delays.
//!
//! In addition to the VRF-based slot assignment described above, which we will
//! call primary slots, the engine also supports a deterministic secondary slot
//! assignment. Primary slots take precedence over secondary slots, when
//! authoring the node starts by trying to claim a primary slot and falls back
//! to a secondary slot claim attempt. The secondary slot assignment is done
//! by picking the authority at index:
//!
//! `blake2_256(epoch_randomness ++ slot_number) % authorities_len`.
//!
//! The fork choice rule is weight-based, where weight equals the number of
//! primary blocks in the chain. We will pick the heaviest chain (more primary
//! blocks) and will go with the longest one in case of a tie.
//!
//! An in-depth description and analysis of the protocol can be found here:
//! <https://research.web3.foundation/en/latest/polkadot/BABE/Babe>
#![forbid(unsafe_code, missing_docs)]
pub use babe_primitives::*;
pub use consensus_common::SyncOracle;
use std::{collections::HashMap, sync::Arc, u64, pin::Pin, time::{Instant, Duration}};
use babe_primitives;
use consensus_common::ImportResult;
use consensus_common::import_queue::{
BoxJustificationImport, BoxFinalityProofImport,
};
use consensus_common::well_known_cache_keys::Id as CacheKeyId;
use sr_primitives::{generic::{BlockId, OpaqueDigestItemId}, Justification};
use sr_primitives::traits::{
Block as BlockT, Header, DigestItemFor, NumberFor, ProvideRuntimeApi,
Zero,
};
use keystore::KeyStorePtr;
use codec::{Decode, Encode};
use parking_lot::{Mutex, MutexGuard};
use primitives::{blake2_256, Blake2Hasher, H256, Pair, Public, U256};
use merlin::Transcript;
use inherents::{InherentDataProviders, InherentData};
use substrate_telemetry::{
telemetry,
CONSENSUS_TRACE,
CONSENSUS_DEBUG,
};
use schnorrkel::{
keys::Keypair,
vrf::{
VRFProof, VRFInOut, VRFOutput,
},
};
use consensus_common::{
self, BlockImport, Environment, Proposer,
ForkChoiceStrategy, BlockImportParams, BlockOrigin, Error as ConsensusError,
};
use srml_babe::{
BabeInherentData,
timestamp::{TimestampInherentData, InherentType as TimestampInherent}
};
use consensus_common::{SelectChain, well_known_cache_keys};
use consensus_common::import_queue::{Verifier, BasicQueue};
use client::{
block_builder::api::BlockBuilder as BlockBuilderApi,
blockchain::{self, HeaderBackend, ProvideCache}, BlockchainEvents, CallExecutor, Client,
runtime_api::ApiExt, error::Result as ClientResult, backend::{AuxStore, Backend},
ProvideUncles,
utils::is_descendent_of,
};
use fork_tree::ForkTree;
use slots::{CheckedHeader, check_equivocation};
use futures::prelude::*;
use futures01::Stream as _;
use log::{error, warn, debug, info, trace};
use slots::{SlotWorker, SlotData, SlotInfo, SlotCompatible};
mod aux_schema;
#[cfg(test)]
mod tests;
pub use babe_primitives::{AuthorityId, AuthorityPair, AuthoritySignature};
/// A slot duration. Create with `get_or_compute`.
// FIXME: Once Rust has higher-kinded types, the duplication between this
// and `super::babe::Config` can be eliminated.
// https://github.com/paritytech/substrate/issues/2434
pub struct Config(slots::SlotDuration<BabeConfiguration>);
impl Config {
/// Either fetch the slot duration from disk or compute it from the genesis
/// state.
pub fn get_or_compute<B: BlockT, C>(client: &C) -> ClientResult<Self>
where
C: AuxStore + ProvideRuntimeApi, C::Api: BabeApi<B>,
{
trace!(target: "babe", "Getting slot duration");
match slots::SlotDuration::get_or_compute(client, |a, b| a.startup_data(b)).map(Self) {
Ok(s) => Ok(s),
Err(s) => {
warn!(target: "babe", "Failed to get slot duration");
Err(s)
}
}
}
/// Get the slot duration in milliseconds.
pub fn get(&self) -> u64 {
self.0.slot_duration
}
/// Retrieve the threshold calculation constant `c`.
pub fn c(&self) -> (u64, u64) {
self.0.c
}
}
impl SlotCompatible for BabeLink {
fn extract_timestamp_and_slot(
&self,
data: &InherentData,
) -> Result<(TimestampInherent, u64, std::time::Duration), consensus_common::Error> {
trace!(target: "babe", "extract timestamp");
data.timestamp_inherent_data()
.and_then(|t| data.babe_inherent_data().map(|a| (t, a)))
.map_err(Into::into)
.map_err(consensus_common::Error::InherentData)
.map(|(x, y)| (x, y, self.0.lock().0.take().unwrap_or_default()))
}
}
/// Parameters for BABE.
pub struct BabeParams<C, E, I, SO, SC> {
/// The configuration for BABE. Includes the slot duration, threshold, and
/// other parameters.
pub config: Config,
/// The keystore that manages the keys of the node.
pub keystore: KeyStorePtr,
/// The client to use
pub client: Arc<C>,
/// The SelectChain Strategy
pub select_chain: SC,
/// A block importer
pub block_import: I,
/// The environment
pub env: E,
/// A sync oracle
pub sync_oracle: SO,
/// Providers for inherent data.
pub inherent_data_providers: InherentDataProviders,
/// Force authoring of blocks even if we are offline
pub force_authoring: bool,
/// The source of timestamps for relative slots
pub time_source: BabeLink,
}
/// Start the babe worker. The returned future should be run in a tokio runtime.
pub fn start_babe<B, C, SC, E, I, SO, Error, H>(BabeParams {
config,
client,
keystore,
select_chain,
block_import,
env,
sync_oracle,
inherent_data_providers,
force_authoring,
time_source,
}: BabeParams<C, E, I, SO, SC>) -> Result<
impl futures01::Future<Item=(), Error=()>,
consensus_common::Error,
> where
B: BlockT<Header=H>,
C: ProvideRuntimeApi + ProvideCache<B> + ProvideUncles<B> + Send + Sync + 'static,
C::Api: BabeApi<B>,
SC: SelectChain<B> + 'static,
E: Environment<B, Error=Error> + Send + Sync,
E::Proposer: Proposer<B, Error=Error>,
<E::Proposer as Proposer<B>>::Create: Unpin + Send + 'static,
H: Header<Hash=B::Hash>,
I: BlockImport<B> + Send + Sync + 'static,
Error: std::error::Error + Send + From<::consensus_common::Error> + From<I::Error> + 'static,
SO: SyncOracle + Send + Sync + Clone,
{
let worker = BabeWorker {
client: client.clone(),
block_import: Arc::new(Mutex::new(block_import)),
env,
sync_oracle: sync_oracle.clone(),
force_authoring,
c: config.c(),
keystore,
};
register_babe_inherent_data_provider(&inherent_data_providers, config.0.slot_duration())?;
uncles::register_uncles_inherent_data_provider(
client.clone(),
select_chain.clone(),
&inherent_data_providers,
)?;
Ok(slots::start_slot_worker(
config.0,
select_chain,
worker,
sync_oracle,
inherent_data_providers,
time_source,
).map(|()| Ok::<(), ()>(())).compat())
}
struct BabeWorker<C, E, I, SO> {
client: Arc<C>,
block_import: Arc<Mutex<I>>,
env: E,
sync_oracle: SO,
force_authoring: bool,
c: (u64, u64),
keystore: KeyStorePtr,
}
impl<H, B, C, E, I, Error, SO> slots::SimpleSlotWorker<B> for BabeWorker<C, E, I, SO> where
B: BlockT<Header=H>,
C: ProvideRuntimeApi + ProvideCache<B>,
C::Api: BabeApi<B>,
E: Environment<B, Error=Error>,
E::Proposer: Proposer<B, Error=Error>,
<E::Proposer as Proposer<B>>::Create: Unpin + Send + 'static,
H: Header<Hash=B::Hash>,
I: BlockImport<B> + Send + Sync + 'static,
SO: SyncOracle + Send + Clone,
Error: std::error::Error + Send + From<::consensus_common::Error> + From<I::Error> + 'static,
{
type EpochData = Epoch;
type Claim = (BabePreDigest, AuthorityPair);
type SyncOracle = SO;
type Proposer = E::Proposer;
type BlockImport = I;
fn logging_target(&self) -> &'static str {
"babe"
}
fn block_import(&self) -> Arc<Mutex<Self::BlockImport>> {
self.block_import.clone()
}
fn epoch_data(&self, block: &B::Hash) -> Result<Self::EpochData, consensus_common::Error> {
epoch_from_runtime(self.client.as_ref(), &BlockId::Hash(*block))
.ok_or(consensus_common::Error::InvalidAuthoritiesSet)
}
fn authorities_len(&self, epoch_data: &Self::EpochData) -> usize {
epoch_data.authorities.len()
}
fn claim_slot(
&self,
header: &B::Header,
slot_number: u64,
epoch_data: &Self::EpochData,
) -> Option<Self::Claim> {
let parent_weight = {
let pre_digest = find_pre_digest::<B>(&header).ok()?;
pre_digest.weight()
};
claim_slot(
slot_number,
parent_weight,
epoch_data,
self.c,
&self.keystore,
)
}
fn pre_digest_data(&self, _slot_number: u64, claim: &Self::Claim) -> Vec<sr_primitives::DigestItem<B::Hash>> {
vec![
<DigestItemFor<B> as CompatibleDigestItem>::babe_pre_digest(claim.0.clone()),
]
}
fn import_block(&self) -> Box<dyn Fn(
B::Header,
&B::Hash,
Vec<B::Extrinsic>,
Self::Claim,
) -> consensus_common::BlockImportParams<B> + Send> {
Box::new(|header, header_hash, body, (_, pair)| {
// sign the pre-sealed hash of the block and then
// add it to a digest item.
let signature = pair.sign(header_hash.as_ref());
let signature_digest_item = <DigestItemFor<B> as CompatibleDigestItem>::babe_seal(signature);
// When we building our own blocks we always author on top of the
// current best according to `SelectChain`, therefore our own block
// proposal should always become the new best.
BlockImportParams {
origin: BlockOrigin::Own,
header,
justification: None,
post_digests: vec![signature_digest_item],
body: Some(body),
finalized: false,
auxiliary: Vec::new(),
fork_choice: ForkChoiceStrategy::Custom(true),
}
})
}
fn force_authoring(&self) -> bool {
self.force_authoring
}
fn sync_oracle(&mut self) -> &mut Self::SyncOracle {
&mut self.sync_oracle
}
fn proposer(&mut self, block: &B::Header) -> Result<Self::Proposer, consensus_common::Error> {
self.env.init(block).map_err(|e| {
consensus_common::Error::ClientImport(format!("{:?}", e)).into()
})
}
}
impl<H, B, C, E, I, Error, SO> SlotWorker<B> for BabeWorker<C, E, I, SO> where
B: BlockT<Header=H>,
C: ProvideRuntimeApi + ProvideCache<B> + Send + Sync,
C::Api: BabeApi<B>,
E: Environment<B, Error=Error> + Send + Sync,
E::Proposer: Proposer<B, Error=Error>,
<E::Proposer as Proposer<B>>::Create: Unpin + Send + 'static,
H: Header<Hash=B::Hash>,
I: BlockImport<B> + Send + Sync + 'static,
SO: SyncOracle + Send + Sync + Clone,
Error: std::error::Error + Send + From<::consensus_common::Error> + From<I::Error> + 'static,
{
type OnSlot = Pin<Box<dyn Future<Output = Result<(), consensus_common::Error>> + Send>>;
fn on_slot(&mut self, chain_head: B::Header, slot_info: SlotInfo) -> Self::OnSlot {
<Self as slots::SimpleSlotWorker<B>>::on_slot(self, chain_head, slot_info)
}
}
macro_rules! babe_err {
($($i: expr),+) => {
{ debug!(target: "babe", $($i),+)
; format!($($i),+)
}
};
}
/// Extract the BABE pre digest from the given header. Pre-runtime digests are
/// mandatory, the function will return `Err` if none is found.
fn find_pre_digest<B: BlockT>(header: &B::Header) -> Result<BabePreDigest, String>
where DigestItemFor<B>: CompatibleDigestItem,
{
// genesis block doesn't contain a pre digest so let's generate a
// dummy one to not break any invariants in the rest of the code
if header.number().is_zero() {
return Ok(BabePreDigest::Secondary {
slot_number: 0,
authority_index: 0,
weight: 0,
});
}
let mut pre_digest: Option<_> = None;
for log in header.digest().logs() {
trace!(target: "babe", "Checking log {:?}, looking for pre runtime digest", log);
match (log.as_babe_pre_digest(), pre_digest.is_some()) {
(Some(_), true) => Err(babe_err!("Multiple BABE pre-runtime digests, rejecting!"))?,
(None, _) => trace!(target: "babe", "Ignoring digest not meant for us"),
(s, false) => pre_digest = s,
}
}
pre_digest.ok_or_else(|| babe_err!("No BABE pre-runtime digest found"))
}
/// Extract the BABE epoch change digest from the given header, if it exists.
fn find_next_epoch_digest<B: BlockT>(header: &B::Header) -> Result<Option<Epoch>, String>
where DigestItemFor<B>: CompatibleDigestItem,
{
let mut epoch_digest: Option<_> = None;
for log in header.digest().logs() {
trace!(target: "babe", "Checking log {:?}, looking for epoch change digest.", log);
let log = log.try_to::<ConsensusLog>(OpaqueDigestItemId::Consensus(&BABE_ENGINE_ID));
match (log, epoch_digest.is_some()) {
(Some(ConsensusLog::NextEpochData(_)), true) => Err(babe_err!("Multiple BABE epoch change digests, rejecting!"))?,
(Some(ConsensusLog::NextEpochData(epoch)), false) => epoch_digest = Some(epoch),
_ => trace!(target: "babe", "Ignoring digest not meant for us"),
}
}
Ok(epoch_digest)
}
/// Check a header has been signed by the right key. If the slot is too far in
/// the future, an error will be returned. If successful, returns the pre-header
/// and the digest item containing the seal.
///
/// The seal must be the last digest. Otherwise, the whole header is considered
/// unsigned. This is required for security and must not be changed.
///
/// This digest item will always return `Some` when used with `as_babe_pre_digest`.
///
/// The given header can either be from a primary or secondary slot assignment,
/// with each having different validation logic.
// FIXME #1018 needs misbehavior types. The `transaction_pool` parameter will be
// used to submit such misbehavior reports.
fn check_header<B: BlockT + Sized, C: AuxStore, T>(
mut header: B::Header,
parent_header: B::Header,
slot_now: u64,
authorities: &[(AuthorityId, BabeAuthorityWeight)],
client: &C,
randomness: [u8; 32],
epoch_index: u64,
secondary_slots: bool,
c: (u64, u64),
_transaction_pool: Option<&T>,
) -> Result<CheckedHeader<B::Header, (DigestItemFor<B>, DigestItemFor<B>)>, String> where
DigestItemFor<B>: CompatibleDigestItem,
T: Send + Sync + 'static,
{
trace!(target: "babe", "Checking header");
let seal = match header.digest_mut().pop() {
Some(x) => x,
None => return Err(babe_err!("Header {:?} is unsealed", header.hash())),
};
let sig = seal.as_babe_seal().ok_or_else(|| {
babe_err!("Header {:?} has a bad seal", header.hash())
})?;
// the pre-hash of the header doesn't include the seal
// and that's what we sign
let pre_hash = header.hash();
let pre_digest = find_pre_digest::<B>(&header)?;
if pre_digest.slot_number() > slot_now {
header.digest_mut().push(seal);
return Ok(CheckedHeader::Deferred(header, pre_digest.slot_number()));
}
if pre_digest.authority_index() > authorities.len() as u32 {
return Err(babe_err!("Slot author not found"));
}
let parent_weight = {
let parent_pre_digest = find_pre_digest::<B>(&parent_header)?;
parent_pre_digest.weight()
};
match &pre_digest {
BabePreDigest::Primary { vrf_output, vrf_proof, authority_index, slot_number, weight } => {
debug!(target: "babe", "Verifying Primary block");
let digest = (vrf_output, vrf_proof, *authority_index, *slot_number, *weight);
check_primary_header::<B>(
pre_hash,
digest,
sig,
parent_weight,
authorities,
randomness,
epoch_index,
c,
)?;
},
BabePreDigest::Secondary { authority_index, slot_number, weight } if secondary_slots => {
debug!(target: "babe", "Verifying Secondary block");
let digest = (*authority_index, *slot_number, *weight);
check_secondary_header::<B>(
pre_hash,
digest,
sig,
parent_weight,
&authorities,
randomness,
)?;
},
_ => {
return Err(babe_err!("Secondary slot assignments are disabled for the current epoch."));
}
}
let author = &authorities[pre_digest.authority_index() as usize].0;
// the header is valid but let's check if there was something else already
// proposed at the same slot by the given author
if let Some(equivocation_proof) = check_equivocation(
client,
slot_now,
pre_digest.slot_number(),
&header,
author,
).map_err(|e| e.to_string())? {
info!(
"Slot author {:?} is equivocating at slot {} with headers {:?} and {:?}",
author,
pre_digest.slot_number(),
equivocation_proof.fst_header().hash(),
equivocation_proof.snd_header().hash(),
);
}
let pre_digest = CompatibleDigestItem::babe_pre_digest(pre_digest);
Ok(CheckedHeader::Checked(header, (pre_digest, seal)))
}
/// Check a primary slot proposal header. We validate that the given header is
/// properly signed by the expected authority, and that the contained VRF proof
/// is valid. Additionally, the weight of this block must increase compared to
/// its parent since it is a primary block.
fn check_primary_header<B: BlockT + Sized>(
pre_hash: B::Hash,
pre_digest: (&VRFOutput, &VRFProof, AuthorityIndex, SlotNumber, BabeBlockWeight),
signature: AuthoritySignature,
parent_weight: BabeBlockWeight,
authorities: &[(AuthorityId, BabeAuthorityWeight)],
randomness: [u8; 32],
epoch_index: u64,
c: (u64, u64),
) -> Result<(), String>
where DigestItemFor<B>: CompatibleDigestItem,
{
let (vrf_output, vrf_proof, authority_index, slot_number, weight) = pre_digest;
if weight != parent_weight + 1 {
return Err("Invalid weight: should increase with Primary block.".into());
}
let author = &authorities[authority_index as usize].0;
if AuthorityPair::verify(&signature, pre_hash, &author) {
let (inout, _) = {
let transcript = make_transcript(
&randomness,
slot_number,
epoch_index,
);
schnorrkel::PublicKey::from_bytes(author.as_slice()).and_then(|p| {
p.vrf_verify(transcript, vrf_output, vrf_proof)
}).map_err(|s| {
babe_err!("VRF verification failed: {:?}", s)
})?
};
let threshold = calculate_primary_threshold(c, authorities, authority_index as usize);
if !check_primary_threshold(&inout, threshold) {
return Err(babe_err!("VRF verification of block by author {:?} failed: \
threshold {} exceeded", author, threshold));
}
Ok(())
} else {
Err(babe_err!("Bad signature on {:?}", pre_hash))
}
}
/// Check a secondary slot proposal header. We validate that the given header is
/// properly signed by the expected authority, which we have a deterministic way
/// of computing. Additionally, the weight of this block must stay the same
/// compared to its parent since it is a secondary block.
fn check_secondary_header<B: BlockT>(
pre_hash: B::Hash,
pre_digest: (AuthorityIndex, SlotNumber, BabeBlockWeight),
signature: AuthoritySignature,
parent_weight: BabeBlockWeight,
authorities: &[(AuthorityId, BabeAuthorityWeight)],
randomness: [u8; 32],
) -> Result<(), String> {
let (authority_index, slot_number, weight) = pre_digest;
if weight != parent_weight {
return Err("Invalid weight: Should stay the same with secondary block.".into());
}
// check the signature is valid under the expected authority and
// chain state.
let expected_author = secondary_slot_author(
slot_number,
authorities,
randomness,
).ok_or_else(|| "No secondary author expected.".to_string())?;
let author = &authorities[authority_index as usize].0;
if expected_author != author {
let msg = format!("Invalid author: Expected secondary author: {:?}, got: {:?}.",
expected_author,
author,
);
return Err(msg);
}
if AuthorityPair::verify(&signature, pre_hash.as_ref(), author) {
Ok(())
} else {
Err(format!("Bad signature on {:?}", pre_hash))
}
}
/// State that must be shared between the import queue and the authoring logic.
#[derive(Default, Clone, Debug)]
pub struct BabeLink(Arc<Mutex<(Option<Duration>, Vec<(Instant, u64)>)>>);
/// A verifier for Babe blocks.
pub struct BabeVerifier<B, E, Block: BlockT, RA, PRA, T> {
client: Arc<Client<B, E, Block, RA>>,
api: Arc<PRA>,
inherent_data_providers: inherents::InherentDataProviders,
config: Config,
time_source: BabeLink,
transaction_pool: Option<Arc<T>>,
}
impl<B, E, Block: BlockT, RA, PRA, T> BabeVerifier<B, E, Block, RA, PRA, T> {
fn check_inherents(
&self,
block: Block,
block_id: BlockId<Block>,
inherent_data: InherentData,
) -> Result<(), String>
where PRA: ProvideRuntimeApi, PRA::Api: BlockBuilderApi<Block>
{
let inherent_res = self.api.runtime_api().check_inherents(
&block_id,
block,
inherent_data,
).map_err(|e| format!("{:?}", e))?;
if !inherent_res.ok() {
inherent_res
.into_errors()
.try_for_each(|(i, e)| {
Err(self.inherent_data_providers.error_to_string(&i, &e))
})
} else {
Ok(())
}
}
}
#[allow(dead_code)]
fn median_algorithm(
median_required_blocks: u64,
slot_duration: u64,
slot_number: u64,
slot_now: u64,
time_source: &mut (Option<Duration>, Vec<(Instant, u64)>),
) {
let num_timestamps = time_source.1.len();
if num_timestamps as u64 >= median_required_blocks && median_required_blocks > 0 {
let mut new_list: Vec<_> = time_source.1.iter().map(|&(t, sl)| {
let offset: u128 = u128::from(slot_duration)
.checked_mul(1_000_000u128) // self.config.get() returns *milliseconds*
.and_then(|x| {
x.checked_mul(u128::from(slot_number).saturating_sub(u128::from(sl)))
})
.expect("we cannot have timespans long enough for this to overflow; qed");
const NANOS_PER_SEC: u32 = 1_000_000_000;
let nanos = (offset % u128::from(NANOS_PER_SEC)) as u32;
let secs = (offset / u128::from(NANOS_PER_SEC)) as u64;
t + Duration::new(secs, nanos)
}).collect();
// FIXME #2926: use a selection algorithm instead of a full sorting algorithm.
new_list.sort_unstable();
let &median = new_list
.get(num_timestamps / 2)
.expect("we have at least one timestamp, so this is a valid index; qed");
let now = Instant::now();
if now >= median {
time_source.0.replace(now - median);
}
time_source.1.clear();
} else {
time_source.1.push((Instant::now(), slot_now))
}
}
impl<B, E, Block, RA, PRA, T> Verifier<Block> for BabeVerifier<B, E, Block, RA, PRA, T> where
Block: BlockT<Hash=H256>,
B: Backend<Block, Blake2Hasher> + 'static,
E: CallExecutor<Block, Blake2Hasher> + 'static + Clone + Send + Sync,
RA: Send + Sync,
PRA: ProvideRuntimeApi + Send + Sync + AuxStore + ProvideCache<Block>,
PRA::Api: BlockBuilderApi<Block> + BabeApi<Block>,
T: Send + Sync + 'static,
{
fn verify(
&mut self,
origin: BlockOrigin,
header: Block::Header,
justification: Option<Justification>,
mut body: Option<Vec<Block::Extrinsic>>,
) -> Result<(BlockImportParams<Block>, Option<Vec<(CacheKeyId, Vec<u8>)>>), String> {
trace!(
target: "babe",
"Verifying origin: {:?} header: {:?} justification: {:?} body: {:?}",
origin,
header,
justification,
body,
);
debug!(target: "babe", "We have {:?} logs in this header", header.digest().logs().len());
let mut inherent_data = self
.inherent_data_providers
.create_inherent_data()
.map_err(String::from)?;
let (_, slot_now, _) = self.time_source.extract_timestamp_and_slot(&inherent_data)
.map_err(|e| format!("Could not extract timestamp and slot: {:?}", e))?;
let hash = header.hash();
let parent_hash = *header.parent_hash();
let epoch = epoch(self.api.as_ref(), &BlockId::Hash(parent_hash))
.map_err(|e| format!("Could not fetch epoch at {:?}: {:?}", parent_hash, e))?;
let (epoch, maybe_next_epoch) = epoch.deconstruct();
let Epoch { authorities, randomness, epoch_index, secondary_slots, .. } = epoch;
let parent_header = self.client.header(&BlockId::Hash(parent_hash))
.map_err(|e| format!("Could not fetch parent header {:?}: {:?}", parent_hash, e))?
.ok_or_else(|| format!("Parent header {:?} not found.", parent_hash))?;
// We add one to allow for some small drift.
// FIXME #1019 in the future, alter this queue to allow deferring of headers
let mut checked_header = check_header::<Block, PRA, T>(
header.clone(),
parent_header.clone(),
slot_now + 1,
&authorities,
&self.api,
randomness,
epoch_index,
secondary_slots,
self.config.c(),
self.transaction_pool.as_ref().map(|x| &**x),
);
// if we have failed to check header using (presumably) current epoch AND we're probably in the next epoch
// => check using next epoch
// (this is only possible on the light client at epoch#0)
if epoch_index == 0 && checked_header.is_err() {
if let Some(Epoch { authorities, randomness, epoch_index, .. }) = maybe_next_epoch {
let checked_header_next = check_header::<Block, PRA, T>(
header,
parent_header,
slot_now + 1,
&authorities,
&self.api,
randomness,
epoch_index,
secondary_slots,
self.config.c(),
self.transaction_pool.as_ref().map(|x| &**x),
);
match checked_header_next {
Ok(checked_header_next) => checked_header = Ok(checked_header_next),
Err(_) => (),
}
}
}
let checked_header = checked_header?;
match checked_header {
CheckedHeader::Checked(pre_header, (pre_digest, seal)) => {
let babe_pre_digest = pre_digest.as_babe_pre_digest()
.expect("check_header always returns a pre-digest digest item; qed");
let slot_number = babe_pre_digest.slot_number();
// if the body is passed through, we need to use the runtime
// to check that the internally-set timestamp in the inherents
// actually matches the slot set in the seal.
if let Some(inner_body) = body.take() {
inherent_data.babe_replace_inherent_data(slot_number);
let block = Block::new(pre_header.clone(), inner_body);
self.check_inherents(
block.clone(),
BlockId::Hash(parent_hash),
inherent_data,
)?;
let (_, inner_body) = block.deconstruct();
body = Some(inner_body);
}
trace!(target: "babe", "Checked {:?}; importing.", pre_header);
telemetry!(
CONSENSUS_TRACE;
"babe.checked_and_importing";
"pre_header" => ?pre_header);
// The fork choice rule is that we pick the heaviest chain (i.e.
// more primary blocks), if there's a tie we go with the longest
// chain.
let new_best = {
let (last_best, last_best_number) = {
#[allow(deprecated)]
let info = self.client.backend().blockchain().info();
(info.best_hash, info.best_number)
};
let best_header = self.client.header(&BlockId::Hash(last_best))
.map_err(|_| "Failed fetching best header")?
.expect("parent_header must be imported; qed");
let best_weight = find_pre_digest::<Block>(&best_header)
.map(|babe_pre_digest| babe_pre_digest.weight())?;
let new_weight = babe_pre_digest.weight();
if new_weight > best_weight {
true
} else if new_weight == best_weight {
*pre_header.number() > last_best_number
} else {
false
}
};
let import_block = BlockImportParams {
origin,
header: pre_header,
post_digests: vec![seal],
body,
finalized: false,
justification,
auxiliary: Vec::new(),
fork_choice: ForkChoiceStrategy::Custom(new_best),
};
Ok((import_block, Default::default()))
}
CheckedHeader::Deferred(a, b) => {
debug!(target: "babe", "Checking {:?} failed; {:?}, {:?}.", hash, a, b);
telemetry!(CONSENSUS_DEBUG; "babe.header_too_far_in_future";
"hash" => ?hash, "a" => ?a, "b" => ?b
);
Err(format!("Header {:?} rejected: too far in the future", hash))
}
}
}
}
/// Regular BABE epoch or spanned genesis epoch.
#[derive(Debug, Decode, Encode)]
enum MaybeSpanEpoch {
/// Genesis entry. Has the data for epoch#0 and epoch#1.
Genesis(Epoch, Epoch),
/// Regular entry. Has the data for the epoch after next (i.e. current epoch + 2).
Regular(Epoch),
}
impl MaybeSpanEpoch {
pub fn deconstruct(self) -> (Epoch, Option<Epoch>) {
match self {
MaybeSpanEpoch::Genesis(epoch0, epoch1) => (epoch0, Some(epoch1)),
MaybeSpanEpoch::Regular(epoch) => (epoch, None),
}
}
#[cfg(test)]
pub fn into_regular(self) -> Option<Epoch> {
match self {
MaybeSpanEpoch::Regular(epoch) => Some(epoch),
_ => None,
}
}
}
/// Extract current epoch data from cache and fallback to querying the runtime
/// if the cache isn't populated.
fn epoch<B, C>(client: &C, at: &BlockId<B>) -> Result<MaybeSpanEpoch, ConsensusError> where
B: BlockT,
C: ProvideRuntimeApi + ProvideCache<B>,
C::Api: BabeApi<B>,
{
epoch_from_cache(client, at)
.or_else(|| epoch_from_runtime(client, at).map(MaybeSpanEpoch::Regular))
.ok_or(consensus_common::Error::InvalidAuthoritiesSet)
}
/// Extract current epoch data from cache.
fn epoch_from_cache<B, C>(client: &C, at: &BlockId<B>) -> Option<MaybeSpanEpoch> where
B: BlockT,
C: ProvideCache<B>,
{
// the epoch that is BABE-valid at the block is not the epoch that is cache-valid at the block
// we need to go back for maximum two steps
client.cache()
.and_then(|cache| cache
.get_at(&well_known_cache_keys::EPOCH, at)
.and_then(|v| Decode::decode(&mut &v[..]).ok()))
}
/// Extract current epoch from runtime.
fn epoch_from_runtime<B, C>(client: &C, at: &BlockId<B>) -> Option<Epoch> where
B: BlockT,
C: ProvideRuntimeApi,
C::Api: BabeApi<B>,
{
if client.runtime_api().has_api::<dyn BabeApi<B>>(at).unwrap_or(false) {
let s = BabeApi::epoch(&*client.runtime_api(), at).ok()?;
if s.authorities.is_empty() {
error!("No authorities!");
None
} else {
Some(s)
}
} else {
error!("bad api!");
None
}
}
/// The BABE import queue type.
pub type BabeImportQueue<B> = BasicQueue<B>;
/// Register the babe inherent data provider, if not registered already.
fn register_babe_inherent_data_provider(
inherent_data_providers: &InherentDataProviders,
slot_duration: u64,
) -> Result<(), consensus_common::Error> {
debug!(target: "babe", "Registering");
if !inherent_data_providers.has_provider(&srml_babe::INHERENT_IDENTIFIER) {
inherent_data_providers
.register_provider(srml_babe::InherentDataProvider::new(slot_duration))
.map_err(Into::into)
.map_err(consensus_common::Error::InherentData)
} else {
Ok(())
}
}
fn get_keypair(q: &AuthorityPair) -> &Keypair {
use primitives::crypto::IsWrappedBy;
primitives::sr25519::Pair::from_ref(q).as_ref()
}
#[allow(deprecated)]
fn make_transcript(
randomness: &[u8],
slot_number: u64,
epoch: u64,
) -> Transcript {
let mut transcript = Transcript::new(&BABE_ENGINE_ID);
transcript.commit_bytes(b"slot number", &slot_number.to_le_bytes());
transcript.commit_bytes(b"current epoch", &epoch.to_le_bytes());
transcript.commit_bytes(b"chain randomness", randomness);
transcript
}
/// Returns true if the given VRF output is lower than the given threshold,
/// false otherwise.
fn check_primary_threshold(inout: &VRFInOut, threshold: u128) -> bool {
u128::from_le_bytes(inout.make_bytes::<[u8; 16]>(BABE_VRF_PREFIX)) < threshold
}
/// Calculates the primary selection threshold for a given authority, taking
/// into account `c` (`1 - c` represents the probability of a slot being empty).
fn calculate_primary_threshold(
c: (u64, u64),
authorities: &[(AuthorityId, BabeAuthorityWeight)],
authority_index: usize,
) -> u128 {
use num_bigint::BigUint;
use num_rational::BigRational;
use num_traits::{cast::ToPrimitive, identities::One};
let c = c.0 as f64 / c.1 as f64;
let theta =
authorities[authority_index].1 as f64 /
authorities.iter().map(|(_, weight)| weight).sum::<u64>() as f64;
let calc = || {
let p = BigRational::from_float(1f64 - (1f64 - c).powf(theta))?;
let numer = p.numer().to_biguint()?;
let denom = p.denom().to_biguint()?;
((BigUint::one() << 128) * numer / denom).to_u128()
};
calc().unwrap_or(u128::max_value())
}
/// Tries to claim the given slot number. This method starts by trying to claim
/// a primary VRF based slot. If we are not able to claim it, then if we have
/// secondary slots enabled for the given epoch, we will fallback to trying to
/// claim a secondary slot.
fn claim_slot(
slot_number: SlotNumber,
parent_weight: BabeBlockWeight,
epoch: &Epoch,
c: (u64, u64),
keystore: &KeyStorePtr,
) -> Option<(BabePreDigest, AuthorityPair)> {
claim_primary_slot(slot_number, parent_weight, epoch, c, keystore)
.or_else(|| {
if epoch.secondary_slots {
claim_secondary_slot(
slot_number,
parent_weight,
&epoch.authorities,
keystore,
epoch.randomness,
)
} else {
None
}
})
}
/// Claim a primary slot if it is our turn. Returns `None` if it is not our turn.
/// This hashes the slot number, epoch, genesis hash, and chain randomness into
/// the VRF. If the VRF produces a value less than `threshold`, it is our turn,
/// so it returns `Some(_)`. Otherwise, it returns `None`.
fn claim_primary_slot(
slot_number: SlotNumber,
parent_weight: BabeBlockWeight,
epoch: &Epoch,
c: (u64, u64),
keystore: &KeyStorePtr,
) -> Option<(BabePreDigest, AuthorityPair)> {
let Epoch { authorities, randomness, epoch_index, .. } = epoch;
let keystore = keystore.read();
for (pair, authority_index) in authorities.iter()
.enumerate()
.flat_map(|(i, a)| {
keystore.key_pair::<AuthorityPair>(&a.0).ok().map(|kp| (kp, i))
})
{
let transcript = make_transcript(randomness, slot_number, *epoch_index);
// Compute the threshold we will use.
//
// We already checked that authorities contains `key.public()`, so it can't
// be empty. Therefore, this division in `calculate_threshold` is safe.
let threshold = calculate_primary_threshold(c, authorities, authority_index);
let pre_digest = get_keypair(&pair)
.vrf_sign_after_check(transcript, |inout| check_primary_threshold(inout, threshold))
.map(|s| {
BabePreDigest::Primary {
slot_number,
vrf_output: s.0.to_output(),
vrf_proof: s.1,
authority_index: authority_index as u32,
weight: parent_weight + 1,
}
});
// early exit on first successful claim
if let Some(pre_digest) = pre_digest {
return Some((pre_digest, pair));
}
}
None
}
/// Get the expected secondary author for the given slot and with given
/// authorities. This should always assign the slot to some authority unless the
/// authorities list is empty.
fn secondary_slot_author(
slot_number: u64,
authorities: &[(AuthorityId, BabeAuthorityWeight)],
randomness: [u8; 32],
) -> Option<&AuthorityId> {
if authorities.is_empty() {
return None;
}
let rand = U256::from((randomness, slot_number).using_encoded(blake2_256));
let authorities_len = U256::from(authorities.len());
let idx = rand % authorities_len;
let expected_author = authorities.get(idx.as_u32() as usize)
.expect("authorities not empty; index constrained to list length; \
this is a valid index; qed");
Some(&expected_author.0)
}
/// Claim a secondary slot if it is our turn to propose, returning the
/// pre-digest to use when authoring the block, or `None` if it is not our turn
/// to propose.
fn claim_secondary_slot(
slot_number: SlotNumber,
parent_weight: BabeBlockWeight,
authorities: &[(AuthorityId, BabeAuthorityWeight)],
keystore: &KeyStorePtr,
randomness: [u8; 32],
) -> Option<(BabePreDigest, AuthorityPair)> {
if authorities.is_empty() {
return None;
}
let expected_author = secondary_slot_author(
slot_number,
authorities,
randomness,
)?;
let keystore = keystore.read();
for (pair, authority_index) in authorities.iter()
.enumerate()
.flat_map(|(i, a)| {
keystore.key_pair::<AuthorityPair>(&a.0).ok().map(|kp| (kp, i))
})
{
if pair.public() == *expected_author {
let pre_digest = BabePreDigest::Secondary {
slot_number,
authority_index: authority_index as u32,
weight: parent_weight,
};
return Some((pre_digest, pair));
}
}
None
}
fn initialize_authorities_cache<B, C>(client: &C) -> Result<(), ConsensusError> where
B: BlockT,
C: ProvideRuntimeApi + ProvideCache<B>,
C::Api: BabeApi<B>,
{
// no cache => no initialization
let cache = match client.cache() {
Some(cache) => cache,
None => return Ok(()),
};
// check if we already have initialized the cache
let genesis_id = BlockId::Number(Zero::zero());
let genesis_epoch: Option<MaybeSpanEpoch> = cache
.get_at(&well_known_cache_keys::EPOCH, &genesis_id)
.and_then(|v| Decode::decode(&mut &v[..]).ok());
if genesis_epoch.is_some() {
return Ok(());
}
let map_err = |error| consensus_common::Error::from(consensus_common::Error::ClientImport(
format!(
"Error initializing authorities cache: {}",
error,
)));
let epoch0 = epoch_from_runtime(client, &genesis_id).ok_or(consensus_common::Error::InvalidAuthoritiesSet)?;
let mut epoch1 = epoch0.clone();
epoch1.epoch_index = 1;
let genesis_epoch = MaybeSpanEpoch::Genesis(epoch0, epoch1);
cache.initialize(&well_known_cache_keys::EPOCH, genesis_epoch.encode())
.map_err(map_err)
}
/// Tree of all epoch changes across all *seen* forks. Data stored in tree is
/// the hash and block number of the block signaling the epoch change, and the
/// epoch that was signalled at that block.
type EpochChanges<Block> = ForkTree<
<Block as BlockT>::Hash,
NumberFor<Block>,
Epoch,
>;
/// A shared epoch changes tree.
#[derive(Clone)]
struct SharedEpochChanges<Block: BlockT> {
inner: Arc<Mutex<EpochChanges<Block>>>,
}
impl<Block: BlockT> SharedEpochChanges<Block> {
fn new() -> Self {
SharedEpochChanges {
inner: Arc::new(Mutex::new(EpochChanges::<Block>::new()))
}
}
fn lock(&self) -> MutexGuard<EpochChanges<Block>> {
self.inner.lock()
}
}
impl<Block: BlockT> From<EpochChanges<Block>> for SharedEpochChanges<Block> {
fn from(epoch_changes: EpochChanges<Block>) -> Self {
SharedEpochChanges {
inner: Arc::new(Mutex::new(epoch_changes))
}
}
}
/// A block-import handler for BABE.
///
/// This scans each imported block for epoch change signals. The signals are
/// tracked in a tree (of all forks), and the import logic validates all epoch
/// change transitions, i.e. whether a given epoch change is expected or whether
/// it is missing.
///
/// The epoch change tree should be pruned as blocks are finalized.
pub struct BabeBlockImport<B, E, Block: BlockT, I, RA, PRA> {
inner: I,
client: Arc<Client<B, E, Block, RA>>,
api: Arc<PRA>,
epoch_changes: SharedEpochChanges<Block>,
}
impl<B, E, Block: BlockT, I: Clone, RA, PRA> Clone for BabeBlockImport<B, E, Block, I, RA, PRA> {
fn clone(&self) -> Self {
BabeBlockImport {
inner: self.inner.clone(),
client: self.client.clone(),
api: self.api.clone(),
epoch_changes: self.epoch_changes.clone(),
}
}
}
impl<B, E, Block: BlockT, I, RA, PRA> BabeBlockImport<B, E, Block, I, RA, PRA> {
fn new(
client: Arc<Client<B, E, Block, RA>>,
api: Arc<PRA>,
epoch_changes: SharedEpochChanges<Block>,
block_import: I,
) -> Self {
BabeBlockImport {
client,
api,
inner: block_import,
epoch_changes,
}
}
}
impl<B, E, Block, I, RA, PRA> BlockImport<Block> for BabeBlockImport<B, E, Block, I, RA, PRA> where
Block: BlockT<Hash=H256>,
I: BlockImport<Block> + Send + Sync,
I::Error: Into<ConsensusError>,
B: Backend<Block, Blake2Hasher> + 'static,
E: CallExecutor<Block, Blake2Hasher> + 'static + Clone + Send + Sync,
RA: Send + Sync,
PRA: ProvideRuntimeApi + ProvideCache<Block>,
PRA::Api: BabeApi<Block>,
{
type Error = ConsensusError;
fn import_block(
&mut self,
mut block: BlockImportParams<Block>,
mut new_cache: HashMap<well_known_cache_keys::Id, Vec<u8>>,
) -> Result<ImportResult, Self::Error> {
let hash = block.post_header().hash();
let number = block.header.number().clone();
// early exit if block already in chain, otherwise the check for
// epoch changes will error when trying to re-import an epoch change
#[allow(deprecated)]
match self.client.backend().blockchain().status(BlockId::Hash(hash)) {
Ok(blockchain::BlockStatus::InChain) => return Ok(ImportResult::AlreadyInChain),
Ok(blockchain::BlockStatus::Unknown) => {},
Err(e) => return Err(ConsensusError::ClientImport(e.to_string()).into()),
}
let slot_number = {
let pre_digest = find_pre_digest::<Block>(&block.header)
.expect("valid babe headers must contain a predigest; \
header has been already verified; qed");
pre_digest.slot_number()
};
// returns a function for checking whether a block is a descendent of another
// consistent with querying client directly after importing the block.
let parent_hash = *block.header.parent_hash();
let is_descendent_of = is_descendent_of(&self.client, Some((&hash, &parent_hash)));
// check if there's any epoch change expected to happen at this slot
let mut epoch_changes = self.epoch_changes.lock();
let enacted_epoch = epoch_changes.find_node_where(
&hash,
&number,
&is_descendent_of,
&|epoch| epoch.start_slot <= slot_number,
).map_err(|e| ConsensusError::from(ConsensusError::ClientImport(e.to_string())))?;
let check_roots = || -> Result<bool, ConsensusError> {
// this can only happen when the chain starts, since there's no
// epoch change at genesis. afterwards every time we expect an epoch
// change it means we will import another one.
for (root, _, _) in epoch_changes.roots() {
let is_descendent_of = is_descendent_of(root, &hash)
.map_err(|e| {
ConsensusError::from(ConsensusError::ClientImport(e.to_string()))
})?;
if is_descendent_of {
return Ok(false);
}
}
Ok(true)
};
let expected_epoch_change = enacted_epoch.is_some();
let next_epoch_digest = find_next_epoch_digest::<Block>(&block.header)
.map_err(|e| ConsensusError::from(ConsensusError::ClientImport(e.to_string())))?;
match (expected_epoch_change, next_epoch_digest.is_some()) {
(true, true) => {},
(false, false) => {},
(true, false) => {
return Err(
ConsensusError::ClientImport(
"Expected epoch change to happen by this block".into(),
)
);
},
(false, true) => {
if !check_roots()? {
return Err(ConsensusError::ClientImport("Unexpected epoch change".into()));
}
},
}
// if there's a pending epoch we'll save the previous epoch changes here
// this way we can revert it if there's any error
let mut old_epoch_changes = None;
if let Some(enacted_epoch) = enacted_epoch.as_ref() {
let enacted_epoch = &enacted_epoch.data;
// update the current epoch in the client cache
new_cache.insert(
well_known_cache_keys::EPOCH,
MaybeSpanEpoch::Regular(enacted_epoch.clone()).encode(),
);
}
if let Some(next_epoch) = next_epoch_digest {
if let Some(enacted_epoch) = enacted_epoch {
let enacted_epoch = &enacted_epoch.data;
if next_epoch.epoch_index.checked_sub(enacted_epoch.epoch_index) != Some(1) {
return Err(ConsensusError::ClientImport(format!(
"Invalid BABE epoch change: expected next epoch to be {:?}, got {:?}",
enacted_epoch.epoch_index.saturating_add(1),
next_epoch.epoch_index,
)));
}
}
old_epoch_changes = Some(epoch_changes.clone());
// track the epoch change in the fork tree
epoch_changes.import(
hash,
number,
next_epoch,
&is_descendent_of,
).map_err(|e| ConsensusError::from(ConsensusError::ClientImport(e.to_string())))?;
crate::aux_schema::write_epoch_changes::<Block, _, _>(
&*epoch_changes,
|insert| block.auxiliary.extend(
insert.iter().map(|(k, v)| (k.to_vec(), Some(v.to_vec())))
)
);
}
let import_result = self.inner.import_block(block, new_cache);
// revert to the original epoch changes in case there's an error
// importing the block
if let Err(_) = import_result {
if let Some(old_epoch_changes) = old_epoch_changes {
*epoch_changes = old_epoch_changes;
}
}
import_result.map_err(Into::into)
}
fn check_block(
&mut self,
hash: Block::Hash,
parent_hash: Block::Hash,
) -> Result<ImportResult, Self::Error> {
self.inner.check_block(hash, parent_hash).map_err(Into::into)
}
}
/// Start an import queue for the BABE consensus algorithm. This method returns
/// the import queue, some data that needs to be passed to the block authoring
/// logic (`BabeLink`), a `BabeBlockImport` which should be used by the
/// authoring when importing its own blocks, and a future that must be run to
/// completion and is responsible for listening to finality notifications and
/// pruning the epoch changes tree.
pub fn import_queue<B, E, Block: BlockT<Hash=H256>, I, RA, PRA, T>(
config: Config,
block_import: I,
justification_import: Option<BoxJustificationImport<Block>>,
finality_proof_import: Option<BoxFinalityProofImport<Block>>,
client: Arc<Client<B, E, Block, RA>>,
api: Arc<PRA>,
inherent_data_providers: InherentDataProviders,
transaction_pool: Option<Arc<T>>,
) -> ClientResult<(
BabeImportQueue<Block>,
BabeLink,
BabeBlockImport<B, E, Block, I, RA, PRA>,
impl futures01::Future<Item = (), Error = ()>,
)> where
B: Backend<Block, Blake2Hasher> + 'static,
I: BlockImport<Block> + Clone + Send + Sync + 'static,
I::Error: Into<ConsensusError>,
E: CallExecutor<Block, Blake2Hasher> + Clone + Send + Sync + 'static,
RA: Send + Sync + 'static,
PRA: ProvideRuntimeApi + ProvideCache<Block> + Send + Sync + AuxStore + 'static,
PRA::Api: BlockBuilderApi<Block> + BabeApi<Block>,
T: Send + Sync + 'static,
{
register_babe_inherent_data_provider(&inherent_data_providers, config.get())?;
initialize_authorities_cache(&*api)?;
let verifier = BabeVerifier {
client: client.clone(),
api: api.clone(),
inherent_data_providers,
time_source: Default::default(),
config,
transaction_pool,
};
#[allow(deprecated)]
let epoch_changes = aux_schema::load_epoch_changes(&**client.backend())?;
let block_import = BabeBlockImport::new(
client.clone(),
api,
epoch_changes.clone(),
block_import,
);
let pruning_task = client.finality_notification_stream()
.map(|v| Ok::<_, ()>(v)).compat()
.for_each(move |notification| {
let is_descendent_of = is_descendent_of(&client, None);
epoch_changes.lock().prune(
&notification.hash,
*notification.header.number(),
&is_descendent_of,
).map_err(|e| {
debug!(target: "babe", "Error pruning epoch changes fork tree: {:?}", e)
})?;
Ok(())
});
let timestamp_core = verifier.time_source.clone();
let queue = BasicQueue::new(
verifier,
Box::new(block_import.clone()),
justification_import,
finality_proof_import,
);
Ok((queue, timestamp_core, block_import, pruning_task))
}
/// BABE test helpers. Utility methods for manually authoring blocks.
#[cfg(feature = "test-helpers")]
pub mod test_helpers {
use super::*;
/// Try to claim the given slot and return a `BabePreDigest` if
/// successful.
pub fn claim_slot<B, C>(
slot_number: u64,
parent: &B::Header,
client: &C,
c: (u64, u64),
keystore: &KeyStorePtr,
) -> Option<BabePreDigest> where
B: BlockT,
C: ProvideRuntimeApi + ProvideCache<B>,
C::Api: BabeApi<B>,
{
let epoch = match epoch(client, &BlockId::Hash(parent.hash())).unwrap() {
MaybeSpanEpoch::Regular(epoch) => epoch,
_ => unreachable!("it is always Regular epoch on full nodes"),
};
let weight = find_pre_digest::<B>(parent).ok()
.map(|d| d.weight())?;
super::claim_slot(
slot_number,
weight,
&epoch,
c,
keystore,
).map(|(digest, _)| digest)
}
}