referrerpolicy=no-referrer-when-downgrade
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
// Copyright (C) Parity Technologies (UK) Ltd.
// This file is part of Polkadot.

// Polkadot 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.

// Polkadot 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 Polkadot.  If not, see <http://www.gnu.org/licenses/>.

//! The collation generation subsystem is the interface between polkadot and the collators.
//!
//! # Protocol
//!
//! On every `ActiveLeavesUpdate`:
//!
//! * If there is no collation generation config, ignore.
//! * Otherwise, for each `activated` head in the update:
//!   * Determine if the para is scheduled on any core by fetching the `availability_cores` Runtime
//!     API.
//!   * Use the Runtime API subsystem to fetch the full validation data.
//!   * Invoke the `collator`, and use its outputs to produce a [`CandidateReceipt`], signed with
//!     the configuration's `key`.
//!   * Dispatch a [`CollatorProtocolMessage::DistributeCollation`]`(receipt, pov)`.

#![deny(missing_docs)]

use codec::Encode;
use error::{Error, Result};
use futures::{channel::oneshot, future::FutureExt, select};
use polkadot_node_primitives::{
	AvailableData, Collation, CollationGenerationConfig, CollationSecondedSignal, PoV,
	SubmitCollationParams,
};
use polkadot_node_subsystem::{
	messages::{CollationGenerationMessage, CollatorProtocolMessage, RuntimeApiMessage},
	overseer, ActiveLeavesUpdate, FromOrchestra, OverseerSignal, SpawnedSubsystem,
	SubsystemContext, SubsystemError, SubsystemResult, SubsystemSender,
};
use polkadot_node_subsystem_util::{
	request_claim_queue, request_persisted_validation_data, request_session_index_for_child,
	request_validation_code_hash, request_validators,
	runtime::{request_node_features, ClaimQueueSnapshot},
};
use polkadot_primitives::{
	collator_signature_payload,
	node_features::FeatureIndex,
	vstaging::{
		transpose_claim_queue, CandidateDescriptorV2, CandidateReceiptV2 as CandidateReceipt,
		CommittedCandidateReceiptV2, TransposedClaimQueue,
	},
	CandidateCommitments, CandidateDescriptor, CollatorPair, CoreIndex, Hash, Id as ParaId,
	NodeFeatures, OccupiedCoreAssumption, PersistedValidationData, SessionIndex,
	ValidationCodeHash,
};
use schnellru::{ByLength, LruMap};
use sp_core::crypto::Pair;
use std::sync::Arc;

mod error;

#[cfg(test)]
mod tests;

mod metrics;
use self::metrics::Metrics;

const LOG_TARGET: &'static str = "parachain::collation-generation";

/// Collation Generation Subsystem
pub struct CollationGenerationSubsystem {
	config: Option<Arc<CollationGenerationConfig>>,
	session_info_cache: SessionInfoCache,
	metrics: Metrics,
}

#[overseer::contextbounds(CollationGeneration, prefix = self::overseer)]
impl CollationGenerationSubsystem {
	/// Create a new instance of the `CollationGenerationSubsystem`.
	pub fn new(metrics: Metrics) -> Self {
		Self { config: None, metrics, session_info_cache: SessionInfoCache::new() }
	}

	/// Run this subsystem
	///
	/// Conceptually, this is very simple: it just loops forever.
	///
	/// - On incoming overseer messages, it starts or stops jobs as appropriate.
	/// - On other incoming messages, if they can be converted into `Job::ToJob` and include a hash,
	///   then they're forwarded to the appropriate individual job.
	/// - On outgoing messages from the jobs, it forwards them to the overseer.
	///
	/// If `err_tx` is not `None`, errors are forwarded onto that channel as they occur.
	/// Otherwise, most are logged and then discarded.
	async fn run<Context>(mut self, mut ctx: Context) {
		loop {
			select! {
				incoming = ctx.recv().fuse() => {
					if self.handle_incoming::<Context>(incoming, &mut ctx).await {
						break;
					}
				},
			}
		}
	}

	// handle an incoming message. return true if we should break afterwards.
	// note: this doesn't strictly need to be a separate function; it's more an administrative
	// function so that we don't clutter the run loop. It could in principle be inlined directly
	// into there. it should hopefully therefore be ok that it's an async function mutably borrowing
	// self.
	async fn handle_incoming<Context>(
		&mut self,
		incoming: SubsystemResult<FromOrchestra<<Context as SubsystemContext>::Message>>,
		ctx: &mut Context,
	) -> bool {
		match incoming {
			Ok(FromOrchestra::Signal(OverseerSignal::ActiveLeaves(ActiveLeavesUpdate {
				activated,
				..
			}))) => {
				if let Err(err) = self.handle_new_activation(activated.map(|v| v.hash), ctx).await {
					gum::warn!(target: LOG_TARGET, err = ?err, "failed to handle new activation");
				}

				false
			},
			Ok(FromOrchestra::Signal(OverseerSignal::Conclude)) => true,
			Ok(FromOrchestra::Communication {
				msg: CollationGenerationMessage::Initialize(config),
			}) => {
				if self.config.is_some() {
					gum::error!(target: LOG_TARGET, "double initialization");
				} else {
					self.config = Some(Arc::new(config));
				}
				false
			},
			Ok(FromOrchestra::Communication {
				msg: CollationGenerationMessage::Reinitialize(config),
			}) => {
				self.config = Some(Arc::new(config));
				false
			},
			Ok(FromOrchestra::Communication {
				msg: CollationGenerationMessage::SubmitCollation(params),
			}) => {
				if let Err(err) = self.handle_submit_collation(params, ctx).await {
					gum::error!(target: LOG_TARGET, ?err, "Failed to submit collation");
				}

				false
			},
			Ok(FromOrchestra::Signal(OverseerSignal::BlockFinalized(..))) => false,
			Err(err) => {
				gum::error!(
					target: LOG_TARGET,
					err = ?err,
					"error receiving message from subsystem context: {:?}",
					err
				);
				true
			},
		}
	}

	async fn handle_submit_collation<Context>(
		&mut self,
		params: SubmitCollationParams,
		ctx: &mut Context,
	) -> Result<()> {
		let Some(config) = &self.config else {
			return Err(Error::SubmittedBeforeInit);
		};
		let _timer = self.metrics.time_submit_collation();

		let SubmitCollationParams {
			relay_parent,
			collation,
			parent_head,
			validation_code_hash,
			result_sender,
			core_index,
		} = params;

		let mut validation_data = match request_persisted_validation_data(
			relay_parent,
			config.para_id,
			OccupiedCoreAssumption::TimedOut,
			ctx.sender(),
		)
		.await
		.await??
		{
			Some(v) => v,
			None => {
				gum::debug!(
					target: LOG_TARGET,
					relay_parent = ?relay_parent,
					our_para = %config.para_id,
					"No validation data for para - does it exist at this relay-parent?",
				);
				return Ok(())
			},
		};

		// We need to swap the parent-head data, but all other fields here will be correct.
		validation_data.parent_head = parent_head;

		let claim_queue = request_claim_queue(relay_parent, ctx.sender()).await.await??;

		let session_index =
			request_session_index_for_child(relay_parent, ctx.sender()).await.await??;

		let session_info =
			self.session_info_cache.get(relay_parent, session_index, ctx.sender()).await?;
		let collation = PreparedCollation {
			collation,
			relay_parent,
			para_id: config.para_id,
			validation_data,
			validation_code_hash,
			n_validators: session_info.n_validators,
			core_index,
			session_index,
		};

		construct_and_distribute_receipt(
			collation,
			config.key.clone(),
			ctx.sender(),
			result_sender,
			&mut self.metrics,
			session_info.v2_receipts,
			&transpose_claim_queue(claim_queue),
		)
		.await?;

		Ok(())
	}

	async fn handle_new_activation<Context>(
		&mut self,
		maybe_activated: Option<Hash>,
		ctx: &mut Context,
	) -> Result<()> {
		let Some(config) = &self.config else {
			return Ok(());
		};

		let Some(relay_parent) = maybe_activated else { return Ok(()) };

		// If there is no collation function provided, bail out early.
		// Important: Lookahead collator and slot based collator do not use `CollatorFn`.
		if config.collator.is_none() {
			return Ok(())
		}

		let para_id = config.para_id;

		let _timer = self.metrics.time_new_activation();

		let session_index =
			request_session_index_for_child(relay_parent, ctx.sender()).await.await??;

		let session_info =
			self.session_info_cache.get(relay_parent, session_index, ctx.sender()).await?;
		let n_validators = session_info.n_validators;

		let claim_queue =
			ClaimQueueSnapshot::from(request_claim_queue(relay_parent, ctx.sender()).await.await??);

		let cores_to_build_on = claim_queue
			.iter_claims_at_depth(0)
			.filter_map(|(core_idx, para_id)| (para_id == config.para_id).then_some(core_idx))
			.collect::<Vec<_>>();

		// Nothing to do if no core assigned to us.
		if cores_to_build_on.is_empty() {
			return Ok(())
		}

		// We are being very optimistic here, but one of the cores could be pending availability
		// for some more blocks, or even time out. We assume all cores are being freed.

		let mut validation_data = match request_persisted_validation_data(
			relay_parent,
			para_id,
			// Just use included assumption always. If there are no pending candidates it's a
			// no-op.
			OccupiedCoreAssumption::Included,
			ctx.sender(),
		)
		.await
		.await??
		{
			Some(v) => v,
			None => {
				gum::debug!(
					target: LOG_TARGET,
					relay_parent = ?relay_parent,
					our_para = %para_id,
					"validation data is not available",
				);
				return Ok(())
			},
		};

		let validation_code_hash = match request_validation_code_hash(
			relay_parent,
			para_id,
			// Just use included assumption always. If there are no pending candidates it's a
			// no-op.
			OccupiedCoreAssumption::Included,
			ctx.sender(),
		)
		.await
		.await??
		{
			Some(v) => v,
			None => {
				gum::debug!(
					target: LOG_TARGET,
					relay_parent = ?relay_parent,
					our_para = %para_id,
					"validation code hash is not found.",
				);
				return Ok(())
			},
		};

		let task_config = config.clone();
		let metrics = self.metrics.clone();
		let mut task_sender = ctx.sender().clone();

		ctx.spawn(
			"chained-collation-builder",
			Box::pin(async move {
				let transposed_claim_queue = transpose_claim_queue(claim_queue.0);

				for core_index in cores_to_build_on {
					let collator_fn = match task_config.collator.as_ref() {
						Some(x) => x,
						None => return,
					};

					let (collation, result_sender) =
						match collator_fn(relay_parent, &validation_data).await {
							Some(collation) => collation.into_inner(),
							None => {
								gum::debug!(
									target: LOG_TARGET,
									?para_id,
									"collator returned no collation on collate",
								);
								return
							},
						};

					let parent_head = collation.head_data.clone();
					if let Err(err) = construct_and_distribute_receipt(
						PreparedCollation {
							collation,
							para_id,
							relay_parent,
							validation_data: validation_data.clone(),
							validation_code_hash,
							n_validators,
							core_index,
							session_index,
						},
						task_config.key.clone(),
						&mut task_sender,
						result_sender,
						&metrics,
						session_info.v2_receipts,
						&transposed_claim_queue,
					)
					.await
					{
						gum::error!(
							target: LOG_TARGET,
							"Failed to construct and distribute collation: {}",
							err
						);
						return
					}

					// Chain the collations. All else stays the same as we build the chained
					// collation on same relay parent.
					validation_data.parent_head = parent_head;
				}
			}),
		)?;

		Ok(())
	}
}

#[overseer::subsystem(CollationGeneration, error=SubsystemError, prefix=self::overseer)]
impl<Context> CollationGenerationSubsystem {
	fn start(self, ctx: Context) -> SpawnedSubsystem {
		let future = async move {
			self.run(ctx).await;
			Ok(())
		}
		.boxed();

		SpawnedSubsystem { name: "collation-generation-subsystem", future }
	}
}

#[derive(Clone)]
struct PerSessionInfo {
	v2_receipts: bool,
	n_validators: usize,
}

struct SessionInfoCache(LruMap<SessionIndex, PerSessionInfo>);

impl SessionInfoCache {
	fn new() -> Self {
		Self(LruMap::new(ByLength::new(2)))
	}

	async fn get<Sender: SubsystemSender<RuntimeApiMessage>>(
		&mut self,
		relay_parent: Hash,
		session_index: SessionIndex,
		sender: &mut Sender,
	) -> Result<PerSessionInfo> {
		if let Some(info) = self.0.get(&session_index) {
			return Ok(info.clone())
		}

		let n_validators =
			request_validators(relay_parent, &mut sender.clone()).await.await??.len();

		let node_features = request_node_features(relay_parent, session_index, sender)
			.await?
			.unwrap_or(NodeFeatures::EMPTY);

		let info = PerSessionInfo {
			v2_receipts: node_features
				.get(FeatureIndex::CandidateReceiptV2 as usize)
				.map(|b| *b)
				.unwrap_or(false),
			n_validators,
		};
		self.0.insert(session_index, info);
		Ok(self.0.get(&session_index).expect("Just inserted").clone())
	}
}

struct PreparedCollation {
	collation: Collation,
	para_id: ParaId,
	relay_parent: Hash,
	validation_data: PersistedValidationData,
	validation_code_hash: ValidationCodeHash,
	n_validators: usize,
	core_index: CoreIndex,
	session_index: SessionIndex,
}

/// Takes a prepared collation, along with its context, and produces a candidate receipt
/// which is distributed to validators.
async fn construct_and_distribute_receipt(
	collation: PreparedCollation,
	key: CollatorPair,
	sender: &mut impl overseer::CollationGenerationSenderTrait,
	result_sender: Option<oneshot::Sender<CollationSecondedSignal>>,
	metrics: &Metrics,
	v2_receipts: bool,
	transposed_claim_queue: &TransposedClaimQueue,
) -> Result<()> {
	let PreparedCollation {
		collation,
		para_id,
		relay_parent,
		validation_data,
		validation_code_hash,
		n_validators,
		core_index,
		session_index,
	} = collation;

	let persisted_validation_data_hash = validation_data.hash();
	let parent_head_data = validation_data.parent_head.clone();
	let parent_head_data_hash = validation_data.parent_head.hash();

	// Apply compression to the block data.
	let pov = {
		let pov = collation.proof_of_validity.into_compressed();
		let encoded_size = pov.encoded_size();

		// As long as `POV_BOMB_LIMIT` is at least `max_pov_size`, this ensures
		// that honest collators never produce a PoV which is uncompressed.
		//
		// As such, honest collators never produce an uncompressed PoV which starts with
		// a compression magic number, which would lead validators to reject the collation.
		if encoded_size > validation_data.max_pov_size as usize {
			return Err(Error::POVSizeExceeded(encoded_size, validation_data.max_pov_size as usize))
		}

		pov
	};

	let pov_hash = pov.hash();

	let signature_payload = collator_signature_payload(
		&relay_parent,
		&para_id,
		&persisted_validation_data_hash,
		&pov_hash,
		&validation_code_hash,
	);

	let erasure_root = erasure_root(n_validators, validation_data, pov.clone())?;

	let commitments = CandidateCommitments {
		upward_messages: collation.upward_messages,
		horizontal_messages: collation.horizontal_messages,
		new_validation_code: collation.new_validation_code,
		head_data: collation.head_data,
		processed_downward_messages: collation.processed_downward_messages,
		hrmp_watermark: collation.hrmp_watermark,
	};

	let receipt = if v2_receipts {
		let ccr = CommittedCandidateReceiptV2 {
			descriptor: CandidateDescriptorV2::new(
				para_id,
				relay_parent,
				core_index,
				session_index,
				persisted_validation_data_hash,
				pov_hash,
				erasure_root,
				commitments.head_data.hash(),
				validation_code_hash,
			),
			commitments,
		};

		ccr.check_core_index(&transposed_claim_queue)
			.map_err(Error::CandidateReceiptCheck)?;

		ccr.to_plain()
	} else {
		if commitments.core_selector().map_err(Error::CandidateReceiptCheck)?.is_some() {
			gum::warn!(
				target: LOG_TARGET,
				?pov_hash,
				?relay_parent,
				para_id = %para_id,
				"Candidate commitments contain UMP signal without v2 receipts being enabled.",
			);
		}
		CandidateReceipt {
			commitments_hash: commitments.hash(),
			descriptor: CandidateDescriptor {
				signature: key.sign(&signature_payload),
				para_id,
				relay_parent,
				collator: key.public(),
				persisted_validation_data_hash,
				pov_hash,
				erasure_root,
				para_head: commitments.head_data.hash(),
				validation_code_hash,
			}
			.into(),
		}
	};

	gum::debug!(
		target: LOG_TARGET,
		candidate_hash = ?receipt.hash(),
		?pov_hash,
		?relay_parent,
		para_id = %para_id,
		?core_index,
		"candidate is generated",
	);
	metrics.on_collation_generated();

	sender
		.send_message(CollatorProtocolMessage::DistributeCollation {
			candidate_receipt: receipt,
			parent_head_data_hash,
			pov,
			parent_head_data,
			result_sender,
			core_index,
		})
		.await;

	Ok(())
}

fn erasure_root(
	n_validators: usize,
	persisted_validation: PersistedValidationData,
	pov: PoV,
) -> Result<Hash> {
	let available_data =
		AvailableData { validation_data: persisted_validation, pov: Arc::new(pov) };

	let chunks = polkadot_erasure_coding::obtain_chunks_v1(n_validators, &available_data)?;
	Ok(polkadot_erasure_coding::branches(&chunks).root())
}