runtime: remove ttl

.gitlab/pipeline/zombienet/polkadot.yml

@@ -233,6 +233,22 @@ zombienet-polkadot-functional-0016-approval-voting-parallel:
       --local-dir="${LOCAL_DIR}/functional"
       --test="0016-approval-voting-parallel.zndsl"
 
+zombienet-polkadot-functional-0017-sync-backing:
+  extends:
+    - .zombienet-polkadot-common
+  script:
+    - /home/nonroot/zombie-net/scripts/ci/run-test-local-env-manager.sh
+      --local-dir="${LOCAL_DIR}/functional"
+      --test="0017-sync-backing.zndsl"
+
+zombienet-polkadot-functional-0018-shared-core-idle-parachain:
+  extends:
+    - .zombienet-polkadot-common
+  script:
+    - /home/nonroot/zombie-net/scripts/ci/run-test-local-env-manager.sh
+      --local-dir="${LOCAL_DIR}/functional"
+      --test="0018-shared-core-idle-parachain.zndsl"
+
 zombienet-polkadot-smoke-0001-parachains-smoke-test:
   extends:
     - .zombienet-polkadot-common

polkadot/primitives/src/v8/mod.rs

@@ -2093,7 +2093,9 @@ pub struct SchedulerParams<BlockNumber> {
 	pub lookahead: u32,
 	/// How many cores are managed by the coretime chain.
 	pub num_cores: u32,
-	/// The max number of times a claim can time out in availability.
+	/// Deprecated and no longer used by the runtime.
+	/// Removal is tracked by <https://github.com/paritytech/polkadot-sdk/issues/6067>.
+	#[deprecated]
 	pub max_availability_timeouts: u32,
 	/// The maximum queue size of the pay as you go module.
 	pub on_demand_queue_max_size: u32,
@@ -2104,13 +2106,14 @@ pub struct SchedulerParams<BlockNumber> {
 	pub on_demand_fee_variability: Perbill,
 	/// The minimum amount needed to claim a slot in the spot pricing queue.
 	pub on_demand_base_fee: Balance,
-	/// The number of blocks a claim stays in the scheduler's claim queue before getting cleared.
-	/// This number should go reasonably higher than the number of blocks in the async backing
-	/// lookahead.
+	/// Deprecated and no longer used by the runtime.
+	/// Removal is tracked by <https://github.com/paritytech/polkadot-sdk/issues/6067>.
+	#[deprecated]
 	pub ttl: BlockNumber,
 }
 
 impl<BlockNumber: Default + From<u32>> Default for SchedulerParams<BlockNumber> {
+	#[allow(deprecated)]
 	fn default() -> Self {
 		Self {
 			group_rotation_frequency: 1u32.into(),

polkadot/runtime/parachains/src/assigner_coretime/mod.rs

@@ -318,9 +318,12 @@ impl<T: Config> AssignmentProvider<BlockNumberFor<T>> for Pallet<T> {
 		Assignment::Bulk(para_id)
 	}
 
-	fn session_core_count() -> u32 {
-		let config = configuration::ActiveConfig::<T>::get();
-		config.scheduler_params.num_cores
+	fn assignment_duplicated(assignment: &Assignment) {
+		match assignment {
+			Assignment::Pool { para_id, core_index } =>
+				on_demand::Pallet::<T>::assignment_duplicated(*para_id, *core_index),
+			Assignment::Bulk(_) => {},
+		}
 	}
 }
 

polkadot/runtime/parachains/src/assigner_coretime/tests.rs

@@ -26,7 +26,6 @@ use crate::{
 	paras::{ParaGenesisArgs, ParaKind},
 	scheduler::common::Assignment,
 };
-use alloc::collections::btree_map::BTreeMap;
 use frame_support::{assert_noop, assert_ok, pallet_prelude::*, traits::Currency};
 use pallet_broker::TaskId;
 use polkadot_primitives::{BlockNumber, Id as ParaId, SessionIndex, ValidationCode};
@@ -78,7 +77,7 @@ fn run_to_block(
 		OnDemand::on_initialize(b + 1);
 
 		// In the real runtime this is expected to be called by the `InclusionInherent` pallet.
-		Scheduler::free_cores_and_fill_claim_queue(BTreeMap::new(), b + 1);
+		Scheduler::advance_claim_queue(&Default::default());
 	}
 }
 

polkadot/runtime/parachains/src/assigner_parachains.rs

@@ -63,7 +63,5 @@ impl<T: Config> AssignmentProvider<BlockNumberFor<T>> for Pallet<T> {
 		Assignment::Bulk(para_id)
 	}
 
-	fn session_core_count() -> u32 {
-		paras::Parachains::<T>::decode_len().unwrap_or(0) as u32
-	}
+	fn assignment_duplicated(_: &Assignment) {}
 }

polkadot/runtime/parachains/src/assigner_parachains/tests.rs

@@ -23,7 +23,6 @@ use crate::{
 	},
 	paras::{ParaGenesisArgs, ParaKind},
 };
-use alloc::collections::btree_map::BTreeMap;
 use frame_support::{assert_ok, pallet_prelude::*};
 use polkadot_primitives::{BlockNumber, Id as ParaId, SessionIndex, ValidationCode};
 
@@ -71,7 +70,7 @@ fn run_to_block(
 		Scheduler::initializer_initialize(b + 1);
 
 		// In the real runtime this is expected to be called by the `InclusionInherent` pallet.
-		Scheduler::free_cores_and_fill_claim_queue(BTreeMap::new(), b + 1);
+		Scheduler::advance_claim_queue(&Default::default());
 	}
 }
 

polkadot/runtime/parachains/src/builder.rs

@@ -18,7 +18,10 @@ use crate::{
 	configuration, inclusion, initializer, paras,
 	paras::ParaKind,
 	paras_inherent,
-	scheduler::{self, common::AssignmentProvider, CoreOccupied, ParasEntry},
+	scheduler::{
+		self,
+		common::{Assignment, AssignmentProvider},
+	},
 	session_info, shared,
 };
 use alloc::{
@@ -138,8 +141,6 @@ pub(crate) struct BenchBuilder<T: paras_inherent::Config> {
 	/// Make every candidate include a code upgrade by setting this to `Some` where the interior
 	/// value is the byte length of the new code.
 	code_upgrade: Option<u32>,
-	/// Specifies whether the claimqueue should be filled.
-	fill_claimqueue: bool,
 	/// Cores which should not be available when being populated with pending candidates.
 	unavailable_cores: Vec<u32>,
 	/// Use v2 candidate descriptor.
@@ -178,7 +179,6 @@ impl<T: paras_inherent::Config> BenchBuilder<T> {
 			backed_in_inherent_paras: Default::default(),
 			elastic_paras: Default::default(),
 			code_upgrade: None,
-			fill_claimqueue: true,
 			unavailable_cores: vec![],
 			candidate_descriptor_v2: false,
 			candidate_modifier: None,
@@ -322,13 +322,6 @@ impl<T: paras_inherent::Config> BenchBuilder<T> {
 		self.max_validators() / self.max_validators_per_core()
 	}
 
-	/// Set whether the claim queue should be filled.
-	#[cfg(not(feature = "runtime-benchmarks"))]
-	pub(crate) fn set_fill_claimqueue(mut self, f: bool) -> Self {
-		self.fill_claimqueue = f;
-		self
-	}
-
 	/// Get the minimum number of validity votes in order for a backed candidate to be included.
 	#[cfg(feature = "runtime-benchmarks")]
 	pub(crate) fn fallback_min_backing_votes() -> u32 {
@@ -340,10 +333,13 @@ impl<T: paras_inherent::Config> BenchBuilder<T> {
 		HeadData(vec![0xFF; max_head_size as usize])
 	}
 
-	fn candidate_descriptor_mock(candidate_descriptor_v2: bool) -> CandidateDescriptorV2<T::Hash> {
+	fn candidate_descriptor_mock(
+		para_id: ParaId,
+		candidate_descriptor_v2: bool,
+	) -> CandidateDescriptorV2<T::Hash> {
 		if candidate_descriptor_v2 {
 			CandidateDescriptorV2::new(
-				0.into(),
+				para_id,
 				Default::default(),
 				CoreIndex(200),
 				2,
@@ -356,7 +352,7 @@ impl<T: paras_inherent::Config> BenchBuilder<T> {
 		} else {
 			// Convert v1 to v2.
 			CandidateDescriptor::<T::Hash> {
-				para_id: 0.into(),
+				para_id,
 				relay_parent: Default::default(),
 				collator: junk_collator(),
 				persisted_validation_data_hash: Default::default(),
@@ -373,6 +369,7 @@ impl<T: paras_inherent::Config> BenchBuilder<T> {
 
 	/// Create a mock of `CandidatePendingAvailability`.
 	fn candidate_availability_mock(
+		para_id: ParaId,
 		group_idx: GroupIndex,
 		core_idx: CoreIndex,
 		candidate_hash: CandidateHash,
@@ -381,15 +378,16 @@ impl<T: paras_inherent::Config> BenchBuilder<T> {
 		candidate_descriptor_v2: bool,
 	) -> inclusion::CandidatePendingAvailability<T::Hash, BlockNumberFor<T>> {
 		inclusion::CandidatePendingAvailability::<T::Hash, BlockNumberFor<T>>::new(
-			core_idx,                                                 // core
-			candidate_hash,                                           // hash
-			Self::candidate_descriptor_mock(candidate_descriptor_v2), // candidate descriptor
-			commitments,                                              // commitments
-			availability_votes,                                       // availability votes
-			Default::default(),                                       // backers
-			Zero::zero(),                                             // relay parent
-			One::one(),                                               /* relay chain block this
-			                                                           * was backed in */
+			core_idx,                                                          // core
+			candidate_hash,                                                    // hash
+			Self::candidate_descriptor_mock(para_id, candidate_descriptor_v2), /* candidate descriptor */
+			commitments,                                                       // commitments
+			availability_votes,                                                /* availability
+			                                                                    * votes */
+			Default::default(), // backers
+			Zero::zero(),       // relay parent
+			One::one(),         /* relay chain block this
+			                     * was backed in */
 			group_idx, // backing group
 		)
 	}
@@ -416,6 +414,7 @@ impl<T: paras_inherent::Config> BenchBuilder<T> {
 			hrmp_watermark: 0u32.into(),
 		};
 		let candidate_availability = Self::candidate_availability_mock(
+			para_id,
 			group_idx,
 			core_idx,
 			candidate_hash,
@@ -886,14 +885,14 @@ impl<T: paras_inherent::Config> BenchBuilder<T> {
 			extra_cores;
 
 		assert!(used_cores <= max_cores);
-		let fill_claimqueue = self.fill_claimqueue;
 
 		// NOTE: there is an n+2 session delay for these actions to take effect.
 		// We are currently in Session 0, so these changes will take effect in Session 2.
 		Self::setup_para_ids(used_cores - extra_cores);
 		configuration::ActiveConfig::<T>::mutate(|c| {
-			c.scheduler_params.num_cores = used_cores as u32;
+			c.scheduler_params.num_cores = max_cores as u32;
 		});
+		configuration::Pallet::<T>::set_coretime_cores_unchecked(max_cores as u32).unwrap();
 
 		let validator_ids = generate_validator_pairs::<T>(self.max_validators());
 		let target_session = SessionIndex::from(self.target_session);
@@ -930,66 +929,57 @@ impl<T: paras_inherent::Config> BenchBuilder<T> {
 
 		assert_eq!(inclusion::PendingAvailability::<T>::iter().count(), used_cores - extra_cores);
 
-		// Mark all the used cores as occupied. We expect that there are
-		// `backed_and_concluding_paras` that are pending availability and that there are
-		// `used_cores - backed_and_concluding_paras ` which are about to be disputed.
-		let now = frame_system::Pallet::<T>::block_number() + One::one();
-
+		// Sanity check that the occupied cores reported by the inclusion module are what we expect
+		// to be.
 		let mut core_idx = 0u32;
 		let elastic_paras = &builder.elastic_paras;
-		// Assign potentially multiple cores to same parachains,
-		let cores = all_cores
+
+		let mut occupied_cores = inclusion::Pallet::<T>::get_occupied_cores()
+			.map(|(core, candidate)| (core, candidate.candidate_descriptor().para_id()))
+			.collect::<Vec<_>>();
+		occupied_cores.sort_by(|(core_a, _), (core_b, _)| core_a.0.cmp(&core_b.0));
+
+		let mut expected_cores = all_cores
 			.iter()
 			.flat_map(|(para_id, _)| {
 				(0..elastic_paras.get(&para_id).cloned().unwrap_or(1))
 					.map(|_para_local_core_idx| {
-						let ttl = configuration::ActiveConfig::<T>::get().scheduler_params.ttl;
-						// Load an assignment into provider so that one is present to pop
-						let assignment =
-							<T as scheduler::Config>::AssignmentProvider::get_mock_assignment(
-								CoreIndex(core_idx),
-								ParaId::from(*para_id),
-							);
+						let old_core_idx = core_idx;
 						core_idx += 1;
-						CoreOccupied::Paras(ParasEntry::new(assignment, now + ttl))
+						(CoreIndex(old_core_idx), ParaId::from(*para_id))
 					})
-					.collect::<Vec<CoreOccupied<_>>>()
+					.collect::<Vec<_>>()
 			})
-			.collect::<Vec<CoreOccupied<_>>>();
+			.collect::<Vec<_>>();
 
-		scheduler::AvailabilityCores::<T>::set(cores);
+		expected_cores.sort_by(|(core_a, _), (core_b, _)| core_a.0.cmp(&core_b.0));
 
-		core_idx = 0u32;
+		assert_eq!(expected_cores, occupied_cores);
 
 		// We need entries in the claim queue for those:
 		all_cores.append(&mut builder.backed_in_inherent_paras.clone());
 
-		if fill_claimqueue {
-			let cores = all_cores
-				.keys()
-				.flat_map(|para_id| {
-					(0..elastic_paras.get(&para_id).cloned().unwrap_or(1))
-						.map(|_para_local_core_idx| {
-							let ttl = configuration::ActiveConfig::<T>::get().scheduler_params.ttl;
-							// Load an assignment into provider so that one is present to pop
-							let assignment =
-								<T as scheduler::Config>::AssignmentProvider::get_mock_assignment(
-									CoreIndex(core_idx),
-									ParaId::from(*para_id),
-								);
-
-							core_idx += 1;
-							(
-								CoreIndex(core_idx - 1),
-								[ParasEntry::new(assignment, now + ttl)].into(),
-							)
-						})
-						.collect::<Vec<(CoreIndex, VecDeque<ParasEntry<_>>)>>()
-				})
-				.collect::<BTreeMap<CoreIndex, VecDeque<ParasEntry<_>>>>();
+		let mut core_idx = 0u32;
+		let cores = all_cores
+			.keys()
+			.flat_map(|para_id| {
+				(0..elastic_paras.get(&para_id).cloned().unwrap_or(1))
+					.map(|_para_local_core_idx| {
+						// Load an assignment into provider so that one is present to pop
+						let assignment =
+							<T as scheduler::Config>::AssignmentProvider::get_mock_assignment(
+								CoreIndex(core_idx),
+								ParaId::from(*para_id),
+							);
 
-			scheduler::ClaimQueue::<T>::set(cores);
-		}
+						core_idx += 1;
+						(CoreIndex(core_idx - 1), [assignment].into())
+					})
+					.collect::<Vec<(CoreIndex, VecDeque<Assignment>)>>()
+			})
+			.collect::<BTreeMap<CoreIndex, VecDeque<Assignment>>>();
+
+		scheduler::ClaimQueue::<T>::set(cores);
 
 		Bench::<T> {
 			data: ParachainsInherentData {