mirror of
https://github.com/pezkuwichain/pezkuwi-subxt.git
synced 2026-05-30 18:41:03 +00:00
XCM v3 (#4097)
* cargo fmt * Create benchmarks for XCM instructions introduced in v3 (#4564) * Create benchmarks for BurnAsset and ExpectAsset * Add benchmarks for ExpectOrigin and ExpectError * Add benchmarks for QueryPallet and ExpectPallet * Add benchmarks for ReportTransactStatus and ClearTransactStatus * cargo fmt * Use AllPalletsWithSystem in mocks * Update XCM generic benchmarks for westend * Remove default impls for some XCM weight functions * Fix compilation error * Add weight_args helper attribute * Remove manually written XcmWeightInfo * Parse trailing comma * Revert "Add weight_args helper attribute" This reverts commit 3b7c47a6182e1b9227036c38b406d494c3fcf6fd. * Fixes * Fixes * XCM v3: Introduce querier field into `QueryReponse` (#4732) * Introduce querier field into QueryReponse * Convert &Option<MultiLocation> to Option<&MultiLocation> &Option<T> is almost always never quite useful, most of the time it still gets converted to an Option<&T> via `as_ref`, so we should simply make functions that accept Option<&T> instead. * Fix tests * cargo fmt * Fix benchmarks * Appease spellchecker * Fix test * Fix tests * Fix test * Fix mock * Fixes * Fix tests * Add test for response queriers * Update xcm/pallet-xcm/src/lib.rs * Test for non-existence of querier Co-authored-by: Keith Yeung <kungfukeith11@gmail.com> * Fixes * Fixes * Add `starts_with` function to `MultiLocation` and `Junctions` (#4835) * add matches_prefix function to MultiLocation and Junctions * rename matches_prefix to starts_with * remove unnecessary main in doc comment Co-authored-by: Bastian Köcher <bkchr@users.noreply.github.com> * Make use of starts_with in match_and_split Co-authored-by: Bastian Köcher <bkchr@users.noreply.github.com> Co-authored-by: Keith Yeung <kungfukeith11@gmail.com> * XCM v3: Bridge infrastructure (#4681) * XCM bridge infrastructure * Missing bit of cherry-pick * Revamped XCM proc macros; new NetworkIds * Fixes * Formatting * ExportMessage instruction and config type * Add MessageExporter definitions * Formatting * Missing files * Fixes * Initial bridging config API * Allow for two-stage XCM execution * Update xcm/src/v3/mod.rs Co-authored-by: Keith Yeung <kungfukeith11@gmail.com> * XCM crate building again * Initial bridging primitive * Docs * Docs * More work * More work * Merge branch 'gav-xcm-v3' into gav-xcm-v3-bridging * Make build * WithComputedOrigin and SovereignPaidRemoteExporter * Remove TODOs * Slim bridge API and tests. * Fixes * More work * First bridge test passing * Formatting * Another test * Next round of bridging tests * Repot tests * Cleanups * Paid bridging * Formatting * Tests * Spelling * Formatting * Fees and refactoring * Fixes * Formatting * Refactor SendXcm to become two-phase * Fix tests * Refactoring of SendXcm and ExportXcm complete * Formatting * Rename CannotReachDestination -> NotApplicable * Remove XCM v0 * Minor grumbles * Formatting * Formatting * Fixes * Fixes * Cleanup XCM config * Fee handling * Fixes * Formatting * Fixes * Bump Co-authored-by: Keith Yeung <kungfukeith11@gmail.com> * Bump Substrate * XCM v3: `ExchangeAsset` and Remote-locking (#4945) * Asset Exchange and Locks * Make sure XCM typers impl MaxEncodedLen * Basic implementation for locks * Bump Substrate * Missing files * Use new API * Introduce instruction * Big refactor * Docs * Remove deprecated struct * Remove deprecated struct * Repot XCM builder tests * ExchangeAsset test * Exchange tests * Locking tests * Locking tests * Fixes and tests * Fixes * Formatting * Spelling * Add simulator test for remote locking * Fix tests * Bump * XCM v3: Support for non-fungibles (#4950) * NFT support and a test * New files. * Integration tests for sending NFTs * Formatting * Broken Cargo features * Use 2021 edition * Fixes * Formatting * Formatting * Update xcm/xcm-builder/src/asset_conversion.rs Co-authored-by: Keith Yeung <kungfukeith11@gmail.com> * Update xcm/xcm-builder/src/nonfungibles_adapter.rs Co-authored-by: Keith Yeung <kungfukeith11@gmail.com> * Update xcm/xcm-executor/src/lib.rs Co-authored-by: Keith Yeung <kungfukeith11@gmail.com> * Fixes * Fixes * Fixes * Formatting * Fixes Co-authored-by: Bastian Köcher <info@kchr.de> Co-authored-by: Keith Yeung <kungfukeith11@gmail.com> * XCM v3: Context & ID hash (#4756) * send_xcm returns message hash * cargo fmt * Create topic register and instructions * Fix weights * Use tabs * Sketch out XcmContext * Fix doc test * Add the XCM context as a parameter to executor trait fns * Fixes * Add XcmContext parameter * Revert adding context as an arg to SendXcm trait methods * Revert adding context argument to ConvertOrigin trait methods * cargo fmt * Do not change the API of XcmExecutor::execute * Fixes * Fixes * Fixes * Fixes * Remove convenience method * Fixes * Fixes * cargo fmt * Fixes * Add benchmarks for XCM topic instructions * cargo run --quiet --profile=production --features=runtime-benchmarks -- benchmark --chain=westend-dev --steps=50 --repeat=20 --pallet=pallet_xcm_benchmarks::generic --extrinsic=* --execution=wasm --wasm-execution=compiled --heap-pages=4096 --header=./file_header.txt --template=./xcm/pallet-xcm-benchmarks/template.hbs --output=./runtime/westend/src/weights/xcm/pallet_xcm_benchmarks_generic.rs * Remove context argument on FilterAssetLocation * Fixes * Remove unused import * Fixes * Fixes * Fixes * Accept XCM hash parameter in ExecuteXcm trait methods * cargo fmt * Properly enable sp-io/std * Fixes * default-features = false * Fixes * Fixes * Fixes * Make XcmContext optional in withdraw_asset * Fixes * Fixes * Fixes * Modify tests to check for the correct XCM hash * Small refactor * cargo fmt * Check for expected hash in xcm-builder unit tests * Add doc comment for the optionality of the XCM context in withdraw_asset * Update xcm/src/v3/traits.rs * Update xcm/src/v3/traits.rs * Store XcmContext and avoid rebuilding * Use ref for XcmContext * Formatting * Fix incorrect hash CC @KiChjang * Refactor and make clear fake hashes * Fixes * Fixes * Fixes * Fix broken hashing * Docs * Fixes * Fixes * Fixes * Formatting * Fixes * Fixes * Fixes * Remove unknowable hash * Formatting * Use message hash for greater identifiability * Formatting * Fixes * Formatting Co-authored-by: Keith Yeung <kungfukeith11@gmail.com> Co-authored-by: Parity Bot <admin@parity.io> * Fixes * Fixes * Fixes * Fixes * Formatting * Fixes * Formatting * Fixes * Fixes * Formatting * Formatting * Remove horrible names * Bump * Remove InvertLocation trait (#5092) * Remove InvertLocation trait * Remove unneeded functions * Formatting * Fixes * Remove XCMv1 (#5094) * Remove XCMv1 * Remove XCMv1 * Formatting * Fixes * Fixes * Formatting * derive serialize/deserialize for xcm primitives (#5036) * derive serialize/deserialize for xcm primitives * derive serialize/deserialize for xcm primitives * update v3 * update v2 Co-authored-by: Gav Wood <gavin@parity.io> * Update lock * Fixes * Add benchmarks for the ExchangeAsset instruction * `AliasOrigin` instruction stub (#5122) * AliasOrigin instruction stub * Fixes * Fixes * Update substrate * Fixes * Ensure same array length before using copy_from_slice * Fixes * Add benchmarks for the UniversalOrigin instruction * Remove unused import * Remove unused import * Add benchmarks for SetFeesMode instruction * Add benchmarks for asset (un)locking instructions * Leave AliasOrigin unbenchmarked * Fixes after merge * cargo fmt * Fixes * Fixes * Set TrustedReserves to None on both Kusama and Westend * Remove extraneous reserve_asset_deposited benchmark * Fix universal_origin benchmark * cargo run --quiet --profile=production --features=runtime-benchmarks -- benchmark pallet --chain=westend-dev --steps=50 --repeat=20 --pallet=pallet_xcm_benchmarks::generic --extrinsic=* --execution=wasm --wasm-execution=compiled --heap-pages=4096 --header=./file_header.txt --template=./xcm/pallet-xcm-benchmarks/template.hbs --output=./runtime/westend/src/weights/xcm/pallet_xcm_benchmarks_generic.rs * Don't rely on skipped benchmark functions * Fixes * cargo run --quiet --profile=production --features=runtime-benchmarks -- benchmark pallet --chain=kusama-dev --steps=50 --repeat=20 --pallet=pallet_xcm_benchmarks::generic --extrinsic=* --execution=wasm --wasm-execution=compiled --heap-pages=4096 --header=./file_header.txt --template=./xcm/pallet-xcm-benchmarks/template.hbs --output=./runtime/kusama/src/weights/xcm/pallet_xcm_benchmarks_generic.rs * Fix unused variables * Fixes * Spelling * Fixes * Fix codec index of VersionedXcm * Allows to customize how calls are dispatched from XCM (#5657) * CallDispatcher trait * fmt * unused import * fix test-runtime * remove JustDispatch type * fix typo in test-runtime * missing CallDispatcher * more missing CallDispatcher * Update comment `NoteAssetLocked` -> `NoteUnlockable` * Fixes * Fixes * Adjust MultiAssets weights based on new wild card variants * Fixes * Fixes * Fixes * Fixes * Fixes * Some late fixes for XCMv3 (#5237) * Maximise chances that trapped assets can be reclaimed * Do origin check as part of ExportMessage for security * Formatting * Fixes * Cleanup export XCM APIs * Formatting * Update xcm/src/v3/junctions.rs * UnpaidExecution instruction and associated barrier. * Tighten barriers (ClearOrigin/QueryResponse) * Allow only 1 ClearOrigin instruction in AllowTopLevelPaidExecutionFrom * Bi-directional teleport accounting * Revert other fix * Build fixes] * Tests build * Benchmark fixes Co-authored-by: Keith Yeung <kungfukeith11@gmail.com> * Update Substrate * Re-export `pub` stuff from universal_exports.rs + removed unecessary clone (#6145) * Re-export `pub` stuff from universal_exports.rs * Removed unnecessary clone * Use 2D weights in XCM v3 (#6134) * Depend upon sp-core instead of sp-runtime * Make sp-io a dev-dependency * Use 2D weights in XCM v3 * cargo fmt * Add XCM pallet migration to runtimes * Use from_parts * cargo fmt * Fixes * cargo fmt * Remove XCMWeight import * Fixes * Fixes * Fixes * Fixes * Use translate in migration * Increase max upward message size in tests * Fix doc test * Remove most uses of from_ref_time * cargo fmt * Fixes * Fixes * Add extrinsic benchmarking to XCM pallet * cargo fmt * Fixes * Use old syntax * cargo fmt * Fixes * Remove hardcoded weights * Add XCM pallet to benchmarks * Use successful origin * Fix weird type parameter compilation issue * Fixes * ".git/.scripts/bench-bot.sh" runtime westend-dev pallet_xcm * ".git/.scripts/bench-bot.sh" runtime rococo-dev pallet_xcm * ".git/.scripts/bench-bot.sh" runtime kusama-dev pallet_xcm * ".git/.scripts/bench-bot.sh" runtime polkadot-dev pallet_xcm * Use benchmarked XCM pallet weights * Fixes * Fixes * Use override instead of skip * Fixes * Fixes * Fixes * Fixes * ".git/.scripts/bench-bot.sh" runtime polkadot-dev pallet_xcm * Fixes * ".git/.scripts/bench-bot.sh" runtime polkadot-dev pallet_xcm * ".git/.scripts/bench-bot.sh" runtime westend-dev pallet_xcm Co-authored-by: command-bot <> * Replace Weight::MAX with 100b weight units * Add test to ensure all_gte in barriers is correct * Update xcm/src/v3/junction.rs Co-authored-by: asynchronous rob <rphmeier@gmail.com> * Add more weight tests * cargo fmt * Create thread_local in XCM executor to limit recursion depth (#6304) * Create thread_local in XCM executor to limit recursion depth * Add unit test for recursion limit * Fix statefulness in tests * Remove panic * Use defer and environmental macro * Fix the implementation * Use nicer interface * Change ThisNetwork to AnyNetwork * Move recursion check up to top level * cargo fmt * Update comment Co-authored-by: Bastian Köcher <info@kchr.de> * Add upper limit on the number of overweight messages in the queue (#6298) * Add upper limit on the number of ovwerweight messages in the queue * Add newline * Introduce whitelist for Transact and limit UMP processing to 10 messages per block (#6280) * Add SafeCallFilter to XcmConfig * Limit UMP to receive 10 messages every block * Place 10 message limit on processing instead of receiving * Always increment the message_processed count whenever a message is processed * Add as_derivative to the Transact whitelist * cargo fmt * Fixes * Update xcm/xcm-builder/src/universal_exports.rs Co-authored-by: Branislav Kontur <bkontur@gmail.com> * Fixes * Fixes * Remove topic register and instead use the topic field in XcmContext * Derive some common traits for DispatchBlobError * Fixes * cargo fmt * Fixes * Fixes * Fix comments * Fixes * Introduce WithOriginFilter and apply it as the CallDispatcher for runtimes * Fixes * Appease clippy and fixes * Fixes * Fix more clippy issues * Fixes * ".git/.scripts/bench-bot.sh" runtime polkadot-dev pallet_xcm * ".git/.scripts/bench-bot.sh" runtime westend-dev pallet_xcm * ".git/.scripts/bench-bot.sh" runtime westend-dev pallet_xcm * Add benchmark function for ExportMessage * Fix comment * Add upper limit to DownwardMessageQueues size * Add max size check for queue in can_queue_downward_message * Fixes * Make Transact runtime call configurable * Return Weight::MAX when there is no successful send XCM origin * Update substrate * Fixes * Fixes * Remove ExportMessage benchmark * Remove assertion on Transact instruction benchmark * Make reachable destination configurable in XCM pallet benchmarks * Fixes * Fixes * Remove cfg attribute in fuzzer * Fixes * Remove cfg attribute for XCM pallet in test runtime * Fixes * Use ReachableDest where possible * Fixes * Add benchmark for UnpaidExecution * Update substrate * Ensure benchmark functions pass filters * Add runtime-benchmarks feature to fuzzer * Ensure FixedRateOfFungible accounts for proof size weights * cargo fmt * Whitelist remark_with_event when runtime-benchmarks feature is enabled * Use remark_with_event for Transact benchmarks * Fix Cargo.lock * Allow up to 3 DescendOrigin instructions before UnpaidExecution * cargo fmt * Edit code comment * Check check_origin for unpaid execution privilege * Fixes * Small nits for xcm-v3 (#6408) * Add possibility to skip benchmark for export_message * ".git/.scripts/bench-bot.sh" xcm westend-dev pallet_xcm_benchmarks::generic * Revert * ".git/.scripts/bench-bot.sh" xcm westend-dev pallet_xcm_benchmarks::generic * Add HaulBlobError to `fn haul_blob` * ".git/.scripts/bench-bot.sh" xcm westend-dev pallet_xcm_benchmarks::generic Co-authored-by: command-bot <> * Revert changes to UnpaidExecution * Change AllowUnpaidExecutionFrom to be explicit * Fix log text * cargo fmt * Add benchmarks for XCM pallet version migration (#6448) * Add benchmarks for XCM pallet version migration * cargo fmt * Fixes * Fixes * Fixes * ".git/.scripts/bench-bot.sh" runtime westend-dev pallet_xcm * ".git/.scripts/bench-bot.sh" runtime kusama-dev pallet_xcm * ".git/.scripts/bench-bot.sh" runtime rococo-dev pallet_xcm * ".git/.scripts/bench-bot.sh" runtime polkadot-dev pallet_xcm * Fix benchmarks * Fix benchmarks * ".git/.scripts/bench-bot.sh" runtime westend-dev pallet_xcm * ".git/.scripts/bench-bot.sh" runtime kusama-dev pallet_xcm * ".git/.scripts/bench-bot.sh" runtime rococo-dev pallet_xcm * ".git/.scripts/bench-bot.sh" runtime polkadot-dev pallet_xcm Co-authored-by: command-bot <> * Merge remote-tracking branch 'origin/master' into gav-xcm-v3 * Fixes * Fix comments (#6470) * Specify Ethereum networks by their chain id (#6286) Co-authored-by: Squirrel <gilescope@gmail.com> * Use for Kusama * Use WithComputedOrigin for Polkadot, Rococo and Westend * Update lock * Fix warning * Update xcm/pallet-xcm/src/tests.rs Co-authored-by: Squirrel <gilescope@gmail.com> * Update runtime/parachains/src/ump/migration.rs Co-authored-by: Oliver Tale-Yazdi <oliver.tale-yazdi@parity.io> * Update xcm/pallet-xcm/src/migration.rs Co-authored-by: Oliver Tale-Yazdi <oliver.tale-yazdi@parity.io> * Fixes * cargo fmt * Typo * Update xcm/src/v3/mod.rs Co-authored-by: Oliver Tale-Yazdi <oliver.tale-yazdi@parity.io> * Docs * Docs * Docs * Docs * Docs * Update xcm/src/v3/multiasset.rs Co-authored-by: Oliver Tale-Yazdi <oliver.tale-yazdi@parity.io> * Add tests for MultiAssets::from_sorted_and_deduplicated * Fail gracefully when same instance NFTs are detected during push * Update Substrate to fix benchmarks * Apply suggestions from code review * Update runtime/kusama/src/xcm_config.rs * Rename arguments * Attempt to fix benchmark * ".git/.scripts/commands/bench/bench.sh" runtime polkadot-dev runtime_parachains::ump * Use actual weights for UMP pallet in Polkadot * ".git/.scripts/commands/bench/bench.sh" runtime kusama-dev runtime_parachains::ump * ".git/.scripts/commands/bench/bench.sh" runtime westend-dev runtime_parachains::ump * ".git/.scripts/commands/bench/bench.sh" runtime rococo-dev runtime_parachains::ump Co-authored-by: Keith Yeung <kungfukeith11@gmail.com> Co-authored-by: Alexander Popiak <alexander.popiak@parity.io> Co-authored-by: Bastian Köcher <bkchr@users.noreply.github.com> Co-authored-by: Bastian Köcher <info@kchr.de> Co-authored-by: Parity Bot <admin@parity.io> Co-authored-by: stanly-johnson <stanlyjohnson@outlook.com> Co-authored-by: nanocryk <6422796+nanocryk@users.noreply.github.com> Co-authored-by: Branislav Kontur <bkontur@gmail.com> Co-authored-by: asynchronous rob <rphmeier@gmail.com> Co-authored-by: command-bot <> Co-authored-by: Vincent Geddes <vincent.geddes@hey.com> Co-authored-by: Squirrel <gilescope@gmail.com> Co-authored-by: Oliver Tale-Yazdi <oliver.tale-yazdi@parity.io> Co-authored-by: Shawn Tabrizi <shawntabrizi@gmail.com>
This commit is contained in:
Gavin Wood
@@ -227,7 +227,6 @@ ump_max_individual_weight : pre.ump_max_individual_weight,
|
||||
pvf_checking_enabled : pre.pvf_checking_enabled,
|
||||
pvf_voting_ttl : pre.pvf_voting_ttl,
|
||||
minimum_validation_upgrade_delay : pre.minimum_validation_upgrade_delay,
|
||||
|
||||
}
|
||||
};
|
||||
|
||||
|
||||
@@ -318,7 +318,7 @@ fn setting_pending_config_members() {
|
||||
max_upward_queue_count: 1337,
|
||||
max_upward_queue_size: 228,
|
||||
max_downward_message_size: 2048,
|
||||
ump_service_total_weight: Weight::from_ref_time(20000),
|
||||
ump_service_total_weight: Weight::from_parts(20000, 20000),
|
||||
max_upward_message_size: 448,
|
||||
max_upward_message_num_per_candidate: 5,
|
||||
hrmp_sender_deposit: 22,
|
||||
@@ -331,7 +331,7 @@ fn setting_pending_config_members() {
|
||||
hrmp_max_parachain_outbound_channels: 10,
|
||||
hrmp_max_parathread_outbound_channels: 20,
|
||||
hrmp_max_message_num_per_candidate: 20,
|
||||
ump_max_individual_weight: Weight::from_ref_time(909),
|
||||
ump_max_individual_weight: Weight::from_parts(909, 909),
|
||||
pvf_checking_enabled: true,
|
||||
pvf_voting_ttl: 3,
|
||||
minimum_validation_upgrade_delay: 20,
|
||||
|
||||
@@ -973,7 +973,7 @@ impl<T: Config> Pallet<T> {
|
||||
|
||||
/// Handle a set of dispute statements corresponding to a single candidate.
|
||||
///
|
||||
/// Fails if the dispute data is invalid. Returns a boolean indicating whether the
|
||||
/// Fails if the dispute data is invalid. Returns a Boolean indicating whether the
|
||||
/// dispute is fresh.
|
||||
fn process_checked_dispute_data(
|
||||
set: CheckedDisputeStatementSet,
|
||||
|
||||
@@ -29,6 +29,8 @@ pub use pallet::*;
|
||||
#[cfg(test)]
|
||||
mod tests;
|
||||
|
||||
pub const MAX_MESSAGE_QUEUE_SIZE: usize = 1024;
|
||||
|
||||
/// An error sending a downward message.
|
||||
#[cfg_attr(test, derive(Debug))]
|
||||
pub enum QueueDownwardMessageError {
|
||||
@@ -137,6 +139,28 @@ impl<T: Config> Pallet<T> {
|
||||
<Self as Store>::DownwardMessageQueueHeads::remove(outgoing_para);
|
||||
}
|
||||
|
||||
/// Determine whether enqueuing a downward message to a specific recipient para would result
|
||||
/// in an error. If this returns `Ok(())` the caller can be certain that a call to
|
||||
/// `queue_downward_message` with the same parameters will be successful.
|
||||
pub fn can_queue_downward_message(
|
||||
config: &HostConfiguration<T::BlockNumber>,
|
||||
para: &ParaId,
|
||||
msg: &DownwardMessage,
|
||||
) -> Result<(), QueueDownwardMessageError> {
|
||||
let serialized_len = msg.len() as u32;
|
||||
if serialized_len > config.max_downward_message_size {
|
||||
return Err(QueueDownwardMessageError::ExceedsMaxMessageSize)
|
||||
}
|
||||
|
||||
if <Self as Store>::DownwardMessageQueues::decode_len(para).unwrap_or(0) >
|
||||
MAX_MESSAGE_QUEUE_SIZE
|
||||
{
|
||||
return Err(QueueDownwardMessageError::ExceedsMaxMessageSize)
|
||||
}
|
||||
|
||||
Ok(())
|
||||
}
|
||||
|
||||
/// Enqueue a downward message to a specific recipient para.
|
||||
///
|
||||
/// When encoded, the message should not exceed the `config.max_downward_message_size`.
|
||||
@@ -155,6 +179,12 @@ impl<T: Config> Pallet<T> {
|
||||
return Err(QueueDownwardMessageError::ExceedsMaxMessageSize)
|
||||
}
|
||||
|
||||
if <Self as Store>::DownwardMessageQueues::decode_len(para).unwrap_or(0) >
|
||||
MAX_MESSAGE_QUEUE_SIZE
|
||||
{
|
||||
return Err(QueueDownwardMessageError::ExceedsMaxMessageSize)
|
||||
}
|
||||
|
||||
let inbound =
|
||||
InboundDownwardMessage { msg, sent_at: <frame_system::Pallet<T>>::block_number() };
|
||||
|
||||
|
||||
@@ -83,7 +83,7 @@ parameter_types! {
|
||||
pub const BlockHashCount: u32 = 250;
|
||||
pub BlockWeights: frame_system::limits::BlockWeights =
|
||||
frame_system::limits::BlockWeights::simple_max(
|
||||
Weight::from_ref_time(4 * 1024 * 1024).set_proof_size(u64::MAX),
|
||||
Weight::from_parts(4 * 1024 * 1024, u64::MAX),
|
||||
);
|
||||
}
|
||||
|
||||
@@ -394,7 +394,7 @@ impl UmpSink for TestUmpSink {
|
||||
max_weight: Weight,
|
||||
) -> Result<Weight, (MessageId, Weight)> {
|
||||
let weight = match u32::decode(&mut &actual_msg[..]) {
|
||||
Ok(w) => Weight::from_ref_time(w as u64),
|
||||
Ok(w) => Weight::from_parts(w as u64, w as u64),
|
||||
Err(_) => return Ok(Weight::zero()), // same as the real `UmpSink`
|
||||
};
|
||||
if weight.any_gt(max_weight) {
|
||||
|
||||
@@ -159,7 +159,7 @@ pub struct ReplacementTimes<N> {
|
||||
#[cfg_attr(test, derive(Debug, Clone, PartialEq))]
|
||||
pub struct ParaPastCodeMeta<N> {
|
||||
/// Block numbers where the code was expected to be replaced and where the code
|
||||
/// was actually replaced, respectively. The first is used to do accurate lookups
|
||||
/// was actually replaced, respectively. The first is used to do accurate look-ups
|
||||
/// of historic code in historic contexts, whereas the second is used to do
|
||||
/// pruning on an accurate timeframe. These can be used as indices
|
||||
/// into the `PastCodeHash` map along with the `ParaId` to fetch the code itself.
|
||||
@@ -520,7 +520,7 @@ impl WeightInfo for TestWeightInfo {
|
||||
}
|
||||
fn include_pvf_check_statement() -> Weight {
|
||||
// This special value is to distinguish from the finalizing variants above in tests.
|
||||
Weight::MAX - Weight::from_ref_time(1)
|
||||
Weight::MAX - Weight::from_parts(1, 1)
|
||||
}
|
||||
}
|
||||
|
||||
|
||||
@@ -31,9 +31,12 @@ pub use pallet::*;
|
||||
/// This is used for benchmarking sanely bounding relevant storate items. It is expected from the `configurations`
|
||||
/// pallet to check these values before setting.
|
||||
pub const MAX_UPWARD_MESSAGE_SIZE_BOUND: u32 = 50 * 1024;
|
||||
/// Maximum amount of overweight messages that can exist in the queue at any given time.
|
||||
pub const MAX_OVERWEIGHT_MESSAGES: u32 = 1000;
|
||||
|
||||
#[cfg(feature = "runtime-benchmarks")]
|
||||
mod benchmarking;
|
||||
pub mod migration;
|
||||
|
||||
#[cfg(test)]
|
||||
pub(crate) mod tests;
|
||||
@@ -133,13 +136,11 @@ impl<XcmExecutor: xcm::latest::ExecuteXcm<C::RuntimeCall>, C: Config> UmpSink
|
||||
},
|
||||
Ok((Ok(xcm_message), weight_used)) => {
|
||||
let xcm_junction = Junction::Parachain(origin.into());
|
||||
let outcome =
|
||||
XcmExecutor::execute_xcm(xcm_junction, xcm_message, max_weight.ref_time());
|
||||
let outcome = XcmExecutor::execute_xcm(xcm_junction, xcm_message, id, max_weight);
|
||||
match outcome {
|
||||
Outcome::Error(XcmError::WeightLimitReached(required)) =>
|
||||
Err((id, Weight::from_ref_time(required))),
|
||||
Outcome::Error(XcmError::WeightLimitReached(required)) => Err((id, required)),
|
||||
outcome => {
|
||||
let outcome_weight = Weight::from_ref_time(outcome.weight_used());
|
||||
let outcome_weight = outcome.weight_used();
|
||||
Pallet::<C>::deposit_event(Event::ExecutedUpward(id, outcome));
|
||||
Ok(weight_used.saturating_add(outcome_weight))
|
||||
},
|
||||
@@ -215,6 +216,7 @@ pub mod pallet {
|
||||
#[pallet::pallet]
|
||||
#[pallet::generate_store(pub(super) trait Store)]
|
||||
#[pallet::without_storage_info]
|
||||
#[pallet::storage_version(migration::STORAGE_VERSION)]
|
||||
pub struct Pallet<T>(_);
|
||||
|
||||
#[pallet::config]
|
||||
@@ -328,7 +330,7 @@ pub mod pallet {
|
||||
/// These messages stay there until manually dispatched.
|
||||
#[pallet::storage]
|
||||
pub type Overweight<T: Config> =
|
||||
StorageMap<_, Twox64Concat, OverweightIndex, (ParaId, Vec<u8>), OptionQuery>;
|
||||
CountedStorageMap<_, Twox64Concat, OverweightIndex, (ParaId, Vec<u8>), OptionQuery>;
|
||||
|
||||
/// The number of overweight messages ever recorded in `Overweight` (and thus the lowest free
|
||||
/// index).
|
||||
@@ -508,6 +510,8 @@ impl<T: Config> Pallet<T> {
|
||||
|
||||
/// Devote some time into dispatching pending upward messages.
|
||||
pub(crate) fn process_pending_upward_messages() -> Weight {
|
||||
const MAX_MESSAGES_PER_BLOCK: u8 = 10;
|
||||
let mut messages_processed = 0;
|
||||
let mut weight_used = Weight::zero();
|
||||
|
||||
let config = <configuration::Pallet<T>>::config();
|
||||
@@ -515,7 +519,12 @@ impl<T: Config> Pallet<T> {
|
||||
let mut queue_cache = QueueCache::new();
|
||||
|
||||
while let Some(dispatchee) = cursor.peek() {
|
||||
if weight_used.any_gte(config.ump_service_total_weight) {
|
||||
if weight_used.any_gte(config.ump_service_total_weight) ||
|
||||
messages_processed >= MAX_MESSAGES_PER_BLOCK
|
||||
{
|
||||
// Temporarily allow for processing of a max of 10 messages per block, until we
|
||||
// properly account for proof size weights.
|
||||
//
|
||||
// Then check whether we've reached or overshoot the
|
||||
// preferred weight for the dispatching stage.
|
||||
//
|
||||
@@ -537,13 +546,16 @@ impl<T: Config> Pallet<T> {
|
||||
// our remaining weight limit, then consume it.
|
||||
let maybe_next = queue_cache.peek_front::<T>(dispatchee);
|
||||
if let Some(upward_message) = maybe_next {
|
||||
messages_processed += 1;
|
||||
match T::UmpSink::process_upward_message(dispatchee, upward_message, max_weight) {
|
||||
Ok(used) => {
|
||||
weight_used += used;
|
||||
let _ = queue_cache.consume_front::<T>(dispatchee);
|
||||
},
|
||||
Err((id, required)) => {
|
||||
if required.any_gt(config.ump_max_individual_weight) {
|
||||
let is_under_limit = Overweight::<T>::count() < MAX_OVERWEIGHT_MESSAGES;
|
||||
weight_used.saturating_accrue(T::DbWeight::get().reads(1));
|
||||
if required.any_gt(config.ump_max_individual_weight) && is_under_limit {
|
||||
// overweight - add to overweight queue and continue with message
|
||||
// execution consuming the message.
|
||||
let upward_message = queue_cache.consume_front::<T>(dispatchee).expect(
|
||||
|
||||
@@ -42,8 +42,8 @@ fn queue_upward_msg<T: Config>(
|
||||
fn create_message_min_size<T: Config>(size: u32) -> Vec<u8> {
|
||||
// Create a message with an empty remark call to determine the encoding overhead
|
||||
let msg_size_empty_transact = VersionedXcm::<T>::from(Xcm::<T>(vec![Transact {
|
||||
origin_type: OriginKind::SovereignAccount,
|
||||
require_weight_at_most: Weight::MAX.ref_time(),
|
||||
origin_kind: OriginKind::SovereignAccount,
|
||||
require_weight_at_most: Weight::MAX,
|
||||
call: frame_system::Call::<T>::remark_with_event { remark: vec![] }.encode().into(),
|
||||
}]))
|
||||
.encode()
|
||||
@@ -54,8 +54,8 @@ fn create_message_min_size<T: Config>(size: u32) -> Vec<u8> {
|
||||
let mut remark = Vec::new();
|
||||
remark.resize(size, 0u8);
|
||||
let msg = VersionedXcm::<T>::from(Xcm::<T>(vec![Transact {
|
||||
origin_type: OriginKind::SovereignAccount,
|
||||
require_weight_at_most: Weight::MAX.ref_time(),
|
||||
origin_kind: OriginKind::SovereignAccount,
|
||||
require_weight_at_most: Weight::MAX,
|
||||
call: frame_system::Call::<T>::remark_with_event { remark }.encode().into(),
|
||||
}]))
|
||||
.encode();
|
||||
@@ -65,12 +65,11 @@ fn create_message_min_size<T: Config>(size: u32) -> Vec<u8> {
|
||||
}
|
||||
|
||||
fn create_message_overweight<T: Config>() -> Vec<u8> {
|
||||
let max_block_weight = T::BlockWeights::get().max_block;
|
||||
// We use a `set_code` Call because it
|
||||
let call = frame_system::Call::<T>::set_code { code: vec![] };
|
||||
VersionedXcm::<T>::from(Xcm::<T>(vec![Transact {
|
||||
origin_type: OriginKind::Superuser,
|
||||
require_weight_at_most: max_block_weight.ref_time(),
|
||||
origin_kind: OriginKind::Superuser,
|
||||
require_weight_at_most: Weight::MAX / 2,
|
||||
call: call.encode().into(),
|
||||
}]))
|
||||
.encode()
|
||||
@@ -107,7 +106,6 @@ frame_benchmarking::benchmarks! {
|
||||
|
||||
service_overweight {
|
||||
let host_conf = configuration::ActiveConfig::<T>::get();
|
||||
let weight = host_conf.ump_max_individual_weight + host_conf.ump_max_individual_weight + Weight::from_ref_time(1000000);
|
||||
let para = ParaId::from(1978);
|
||||
// The message's weight does not really matter here, as we add service_overweight's
|
||||
// max_weight parameter to the extrinsic's weight in the weight calculation.
|
||||
@@ -117,7 +115,7 @@ frame_benchmarking::benchmarks! {
|
||||
let msg = create_message_overweight::<T>();
|
||||
|
||||
// This just makes sure that 0 is not a valid index and we can use it later on.
|
||||
let _ = Ump::<T>::service_overweight(RawOrigin::Root.into(), 0, Weight::from_ref_time(1000).set_proof_size(u64::MAX));
|
||||
let _ = Ump::<T>::service_overweight(RawOrigin::Root.into(), 0, Weight::from_parts(1000, 1000));
|
||||
// Start with the block number 1. This is needed because should an event be
|
||||
// emitted during the genesis block they will be implicitly wiped.
|
||||
frame_system::Pallet::<T>::set_block_number(1u32.into());
|
||||
|
||||
@@ -0,0 +1,49 @@
|
||||
// Copyright 2022 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/>.
|
||||
|
||||
use crate::ump::{Config, Overweight, Pallet};
|
||||
use frame_support::{
|
||||
pallet_prelude::*,
|
||||
traits::{OnRuntimeUpgrade, StorageVersion},
|
||||
weights::Weight,
|
||||
};
|
||||
|
||||
pub const STORAGE_VERSION: StorageVersion = StorageVersion::new(1);
|
||||
|
||||
pub mod v1 {
|
||||
use super::*;
|
||||
|
||||
pub struct MigrateToV1<T>(sp_std::marker::PhantomData<T>);
|
||||
impl<T: Config> OnRuntimeUpgrade for MigrateToV1<T> {
|
||||
fn on_runtime_upgrade() -> Weight {
|
||||
if StorageVersion::get::<Pallet<T>>() == 0 {
|
||||
let mut weight = T::DbWeight::get().reads(1);
|
||||
|
||||
let overweight_messages = Overweight::<T>::initialize_counter() as u64;
|
||||
log::info!("Initialized Overweight to {}", overweight_messages);
|
||||
|
||||
weight.saturating_accrue(T::DbWeight::get().reads_writes(overweight_messages, 1));
|
||||
|
||||
StorageVersion::new(1).put::<Pallet<T>>();
|
||||
|
||||
weight.saturating_add(T::DbWeight::get().writes(1))
|
||||
} else {
|
||||
log::warn!("skipping v1, should be removed");
|
||||
T::DbWeight::get().reads(1)
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
@@ -34,12 +34,12 @@ pub(super) struct GenesisConfigBuilder {
|
||||
impl Default for GenesisConfigBuilder {
|
||||
fn default() -> Self {
|
||||
Self {
|
||||
max_upward_message_size: 16,
|
||||
max_upward_message_size: 32,
|
||||
max_upward_message_num_per_candidate: 2,
|
||||
max_upward_queue_count: 4,
|
||||
max_upward_queue_size: 64,
|
||||
ump_service_total_weight: Weight::from_ref_time(1000).set_proof_size(1000),
|
||||
ump_max_individual_weight: Weight::from_ref_time(100).set_proof_size(100),
|
||||
ump_service_total_weight: Weight::from_parts(1000, 1000),
|
||||
ump_max_individual_weight: Weight::from_parts(100, 100),
|
||||
}
|
||||
}
|
||||
}
|
||||
@@ -156,7 +156,7 @@ fn dispatch_resume_after_exceeding_dispatch_stage_weight() {
|
||||
|
||||
new_test_ext(
|
||||
GenesisConfigBuilder {
|
||||
ump_service_total_weight: Weight::from_ref_time(500).set_proof_size(500),
|
||||
ump_service_total_weight: Weight::from_parts(500, 500),
|
||||
..Default::default()
|
||||
}
|
||||
.build(),
|
||||
@@ -203,8 +203,8 @@ fn dispatch_keeps_message_after_weight_exhausted() {
|
||||
|
||||
new_test_ext(
|
||||
GenesisConfigBuilder {
|
||||
ump_service_total_weight: Weight::from_ref_time(500).set_proof_size(500),
|
||||
ump_max_individual_weight: Weight::from_ref_time(300).set_proof_size(300),
|
||||
ump_service_total_weight: Weight::from_parts(500, 500),
|
||||
ump_max_individual_weight: Weight::from_parts(300, 300),
|
||||
..Default::default()
|
||||
}
|
||||
.build(),
|
||||
@@ -243,7 +243,7 @@ fn dispatch_correctly_handle_remove_of_latest() {
|
||||
|
||||
new_test_ext(
|
||||
GenesisConfigBuilder {
|
||||
ump_service_total_weight: Weight::from_ref_time(900).set_proof_size(900),
|
||||
ump_service_total_weight: Weight::from_parts(900, 900),
|
||||
..Default::default()
|
||||
}
|
||||
.build(),
|
||||
@@ -296,7 +296,7 @@ fn service_overweight_unknown() {
|
||||
// the next test.
|
||||
new_test_ext(GenesisConfigBuilder::default().build()).execute_with(|| {
|
||||
assert_noop!(
|
||||
Ump::service_overweight(RuntimeOrigin::root(), 0, Weight::from_ref_time(1000)),
|
||||
Ump::service_overweight(RuntimeOrigin::root(), 0, Weight::from_parts(1000, 1000)),
|
||||
Error::<Test>::UnknownMessageIndex
|
||||
);
|
||||
});
|
||||
@@ -312,8 +312,8 @@ fn overweight_queue_works() {
|
||||
|
||||
new_test_ext(
|
||||
GenesisConfigBuilder {
|
||||
ump_service_total_weight: Weight::from_ref_time(900).set_proof_size(900),
|
||||
ump_max_individual_weight: Weight::from_ref_time(300).set_proof_size(300),
|
||||
ump_service_total_weight: Weight::from_parts(900, 900),
|
||||
ump_max_individual_weight: Weight::from_parts(300, 300),
|
||||
..Default::default()
|
||||
}
|
||||
.build(),
|
||||
@@ -338,7 +338,7 @@ fn overweight_queue_works() {
|
||||
para_a,
|
||||
upward_message_id(&a_msg_3[..]),
|
||||
0,
|
||||
Weight::from_ref_time(500),
|
||||
Weight::from_parts(500, 500),
|
||||
)
|
||||
.into(),
|
||||
);
|
||||
@@ -346,18 +346,18 @@ fn overweight_queue_works() {
|
||||
// Now verify that if we wanted to service this overweight message with less than enough
|
||||
// weight it will fail.
|
||||
assert_noop!(
|
||||
Ump::service_overweight(RuntimeOrigin::root(), 0, Weight::from_ref_time(499)),
|
||||
Ump::service_overweight(RuntimeOrigin::root(), 0, Weight::from_parts(499, 499)),
|
||||
Error::<Test>::WeightOverLimit
|
||||
);
|
||||
|
||||
// ... and if we try to service it with just enough weight it will succeed as well.
|
||||
assert_ok!(Ump::service_overweight(RuntimeOrigin::root(), 0, Weight::from_ref_time(500)));
|
||||
assert_last_event(Event::OverweightServiced(0, Weight::from_ref_time(500)).into());
|
||||
assert_ok!(Ump::service_overweight(RuntimeOrigin::root(), 0, Weight::from_parts(500, 500)));
|
||||
assert_last_event(Event::OverweightServiced(0, Weight::from_parts(500, 500)).into());
|
||||
|
||||
// ... and if we try to service a message with index that doesn't exist it will error
|
||||
// out.
|
||||
assert_noop!(
|
||||
Ump::service_overweight(RuntimeOrigin::root(), 1, Weight::from_ref_time(1000)),
|
||||
Ump::service_overweight(RuntimeOrigin::root(), 1, Weight::from_parts(1000, 1000)),
|
||||
Error::<Test>::UnknownMessageIndex
|
||||
);
|
||||
});
|
||||
|
||||
Reference in New Issue
Block a user