Newer
Older
// This file is part of Substrate.
// Copyright (C) 2019-2022 Parity Technologies (UK) Ltd.
// SPDX-License-Identifier: Apache-2.0
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
//! # Transaction Payment Pallet
//! This pallet provides the basic logic needed to pay the absolute minimum amount needed for a
//! transaction to be included. This includes:
//! - _base fee_: This is the minimum amount a user pays for a transaction. It is declared
//! as a base _weight_ in the runtime and converted to a fee using `WeightToFee`.
//! - _weight fee_: A fee proportional to amount of weight a transaction consumes.
//! - _length fee_: A fee proportional to the encoded length of the transaction.
//! - _tip_: An optional tip. Tip increases the priority of the transaction, giving it a higher
//! chance to be included by the transaction queue.
//!
//! The base fee and adjusted weight and length fees constitute the _inclusion fee_, which is
//! the minimum fee for a transaction to be included in a block.
//!
//! The formula of final fee:
//! ```ignore
//! inclusion_fee = base_fee + length_fee + [targeted_fee_adjustment * weight_fee];
//! final_fee = inclusion_fee + tip;
//! ```
//!
//! - `targeted_fee_adjustment`: This is a multiplier that can tune the final fee based on
//! the congestion of the network.
//!
//! Additionally, this pallet allows one to configure:
//! - The mapping between one unit of weight to one unit of fee via [`Config::WeightToFee`].
//! - A means of updating the fee for the next block, via defining a multiplier, based on the
//! final state of the chain at the end of the previous block. This can be configured via
//! [`Config::FeeMultiplierUpdate`]
//! - How the fees are paid via [`Config::OnChargeTransaction`].
#![cfg_attr(not(feature = "std"), no_std)]
use codec::{Decode, Encode, MaxEncodedLen};
Convert, DispatchInfoOf, Dispatchable, One, PostDispatchInfoOf, SaturatedConversion,
Saturating, SignedExtension, Zero,
transaction_validity::{
TransactionPriority, TransactionValidity, TransactionValidityError, ValidTransaction,
},
FixedPointNumber, FixedPointOperand, FixedU128, Perquintill, RuntimeDebug,
use sp_std::prelude::*;
use frame_support::{
weights::{
DispatchClass, DispatchInfo, GetDispatchInfo, Pays, PostDispatchInfo, Weight,
WeightToFeeCoefficient, WeightToFeePolynomial,
},
};
pub use pallet::*;
pub use types::{FeeDetails, InclusionFee, RuntimeDispatchInfo};
pub type Multiplier = FixedU128;
type BalanceOf<T> = <<T as Config>::OnChargeTransaction as OnChargeTransaction<T>>::Balance;
/// A struct to update the weight multiplier per block. It implements `Convert<Multiplier,
/// Multiplier>`, meaning that it can convert the previous multiplier to the next one. This should
/// be called on `on_finalize` of a block, prior to potentially cleaning the weight data from the
/// system pallet.
///
/// given:
/// s = previous block weight
/// s'= ideal block weight
/// m = maximum block weight
/// diff = (s - s')/m
/// v = 0.00001
/// t1 = (v * diff)
/// t2 = (v * diff)^2 / 2
/// then:
/// next_multiplier = prev_multiplier * (1 + t1 + t2)
///
/// Where `(s', v)` must be given as the `Get` implementation of the `T` generic type. Moreover, `M`
/// must provide the minimum allowed value for the multiplier. Note that a runtime should ensure
/// with tests that the combination of this `M` and `V` is not such that the multiplier can drop to
/// zero and never recover.
///
/// note that `s'` is interpreted as a portion in the _normal transaction_ capacity of the block.
/// For example, given `s' == 0.25` and `AvailableBlockRatio = 0.75`, then the target fullness is
/// _0.25 of the normal capacity_ and _0.1875 of the entire block_.
///
/// This implementation implies the bound:
/// - `v ≤ p / k * (s − s')`
/// - or, solving for `p`: `p >= v * k * (s - s')`
///
/// where `p` is the amount of change over `k` blocks.
///
/// Hence:
/// - in a fully congested chain: `p >= v * k * (1 - s')`.
/// - in an empty chain: `p >= v * k * (-s')`.
///
/// For example, when all blocks are full and there are 28800 blocks per day (default in
/// `substrate-node`) and v == 0.00001, s' == 0.1875, we'd have:
///
/// p >= 0.00001 * 28800 * 0.8125
/// p >= 0.234
///
/// Meaning that fees can change by around ~23% per day, given extreme congestion.
///
/// More info can be found at:
/// <https://research.web3.foundation/en/latest/polkadot/overview/2-token-economics.html>
pub struct TargetedFeeAdjustment<T, S, V, M>(sp_std::marker::PhantomData<(T, S, V, M)>);
/// Something that can convert the current multiplier to the next one.
pub trait MultiplierUpdate: Convert<Multiplier, Multiplier> {
/// Minimum multiplier
fn min() -> Multiplier;
/// Target block saturation level
fn target() -> Perquintill;
/// Variability factor
fn variability() -> Multiplier;
}
impl MultiplierUpdate for () {
fn min() -> Multiplier {
Default::default()
}
fn target() -> Perquintill {
Default::default()
}
fn variability() -> Multiplier {
Default::default()
}
}
impl<T, S, V, M> MultiplierUpdate for TargetedFeeAdjustment<T, S, V, M>
where
T: frame_system::Config,
S: Get<Perquintill>,
V: Get<Multiplier>,
M: Get<Multiplier>,
{
fn min() -> Multiplier {
M::get()
}
fn target() -> Perquintill {
S::get()
}
fn variability() -> Multiplier {
V::get()
}
}
impl<T, S, V, M> Convert<Multiplier, Multiplier> for TargetedFeeAdjustment<T, S, V, M>
where
T: frame_system::Config,
S: Get<Perquintill>,
V: Get<Multiplier>,
M: Get<Multiplier>,
{
fn convert(previous: Multiplier) -> Multiplier {
// Defensive only. The multiplier in storage should always be at most positive. Nonetheless
// we recover here in case of errors, because any value below this would be stale and can
// never change.
let min_multiplier = M::get();
let previous = previous.max(min_multiplier);
let weights = T::BlockWeights::get();
// the computed ratio is only among the normal class.
let normal_max_weight =
weights.get(DispatchClass::Normal).max_total.unwrap_or(weights.max_block);
let current_block_weight = <frame_system::Pallet<T>>::block_weight();
let normal_block_weight =
*current_block_weight.get(DispatchClass::Normal).min(&normal_max_weight);
let s = S::get();
let v = V::get();
let target_weight = (s * normal_max_weight) as u128;
let block_weight = normal_block_weight as u128;
// determines if the first_term is positive
let positive = block_weight >= target_weight;
let diff_abs = block_weight.max(target_weight) - block_weight.min(target_weight);
// defensive only, a test case assures that the maximum weight diff can fit in Multiplier
// without any saturation.
let diff = Multiplier::saturating_from_rational(diff_abs, normal_max_weight.max(1));
let diff_squared = diff.saturating_mul(diff);
let v_squared_2 = v.saturating_mul(v) / Multiplier::saturating_from_integer(2);
let first_term = v.saturating_mul(diff);
let second_term = v_squared_2.saturating_mul(diff_squared);
if positive {
let excess = first_term.saturating_add(second_term).saturating_mul(previous);
previous.saturating_add(excess).max(min_multiplier)
} else {
// Defensive-only: first_term > second_term. Safe subtraction.
let negative = first_term.saturating_sub(second_term).saturating_mul(previous);
previous.saturating_sub(negative).max(min_multiplier)
}
}
}
/// Storage releases of the pallet.
#[derive(Encode, Decode, Clone, Copy, PartialEq, Eq, RuntimeDebug, TypeInfo, MaxEncodedLen)]
/// Original version of the pallet.
V1Ancient,
/// One that bumps the usage to FixedU128 from FixedI128.
V2,
}
impl Default for Releases {
fn default() -> Self {
Releases::V1Ancient
}
}
#[frame_support::pallet]
pub mod pallet {
use frame_support::pallet_prelude::*;
use frame_system::pallet_prelude::*;
#[pallet::pallet]
#[pallet::generate_store(pub(super) trait Store)]
pub struct Pallet<T>(_);
#[pallet::config]
pub trait Config: frame_system::Config {
/// Handler for withdrawing, refunding and depositing the transaction fee.
/// Transaction fees are withdrawn before the transaction is executed.
/// After the transaction was executed the transaction weight can be
/// adjusted, depending on the used resources by the transaction. If the
/// transaction weight is lower than expected, parts of the transaction fee
/// might be refunded. In the end the fees can be deposited.
type OnChargeTransaction: OnChargeTransaction<Self>;
/// A fee mulitplier for `Operational` extrinsics to compute "virtual tip" to boost their
/// `priority`
///
/// This value is multipled by the `final_fee` to obtain a "virtual tip" that is later
/// added to a tip component in regular `priority` calculations.
/// It means that a `Normal` transaction can front-run a similarly-sized `Operational`
/// extrinsic (with no tip), by including a tip value greater than the virtual tip.
///
/// ```rust,ignore
/// // For `Normal`
/// let priority = priority_calc(tip);
///
/// // For `Operational`
/// let virtual_tip = (inclusion_fee + tip) * OperationalFeeMultiplier;
/// let priority = priority_calc(tip + virtual_tip);
/// ```
///
/// Note that since we use `final_fee` the multiplier applies also to the regular `tip`
/// sent with the transaction. So, not only does the transaction get a priority bump based
/// on the `inclusion_fee`, but we also amplify the impact of tips applied to `Operational`
/// transactions.
#[pallet::constant]
type OperationalFeeMultiplier: Get<u8>;
/// Convert a weight value into a deductible fee based on the currency type.
type WeightToFee: WeightToFeePolynomial<Balance = BalanceOf<Self>>;
Stephen Shelton
committed
/// Convert a length value into a deductible fee based on the currency type.
type LengthToFee: WeightToFeePolynomial<Balance = BalanceOf<Self>>;
/// Update the multiplier of the next block, based on the previous block's weight.
type FeeMultiplierUpdate: MultiplierUpdate;
#[pallet::extra_constants]
impl<T: Config> Pallet<T> {
#[pallet::constant_name(WeightToFee)]
/// The polynomial that is applied in order to derive fee from weight.
fn weight_to_fee_polynomial() -> Vec<WeightToFeeCoefficient<BalanceOf<T>>> {
T::WeightToFee::polynomial().to_vec()
}
Stephen Shelton
committed
/// The polynomial that is applied in order to derive fee from length.
#[pallet::constant_name(LengthToFee)]
fn length_to_fee_polynomial() -> Vec<WeightToFeeCoefficient<BalanceOf<T>>> {
T::LengthToFee::polynomial().to_vec()
}
}
#[pallet::type_value]
pub fn NextFeeMultiplierOnEmpty() -> Multiplier {
Multiplier::saturating_from_integer(1)
}
#[pallet::storage]
#[pallet::getter(fn next_fee_multiplier)]
pub type NextFeeMultiplier<T: Config> =
StorageValue<_, Multiplier, ValueQuery, NextFeeMultiplierOnEmpty>;
#[pallet::storage]
pub(super) type StorageVersion<T: Config> = StorageValue<_, Releases, ValueQuery>;
#[pallet::genesis_config]
pub struct GenesisConfig;
#[cfg(feature = "std")]
impl Default for GenesisConfig {
fn default() -> Self {
Self
}
}
#[pallet::genesis_build]
impl<T: Config> GenesisBuild<T> for GenesisConfig {
fn build(&self) {
StorageVersion::<T>::put(Releases::V2);
}
}
#[pallet::hooks]
impl<T: Config> Hooks<BlockNumberFor<T>> for Pallet<T> {
fn on_finalize(_: T::BlockNumber) {
<NextFeeMultiplier<T>>::mutate(|fm| {
*fm = T::FeeMultiplierUpdate::convert(*fm);
fn integrity_test() {
// given weight == u64, we build multipliers from `diff` of two weight values, which can
// at most be maximum block weight. Make sure that this can fit in a multiplier without
// loss.
assert!(
<Multiplier as sp_runtime::traits::Bounded>::max_value() >=
Multiplier::checked_from_integer::<u128>(
T::BlockWeights::get().max_block.try_into().unwrap()
)
.unwrap(),
let target = T::FeeMultiplierUpdate::target() *
T::BlockWeights::get().get(DispatchClass::Normal).max_total.expect(
"Setting `max_total` for `Normal` dispatch class is not compatible with \
`transaction-payment` pallet.",
// add 1 percent;
let addition = target / 100;
if addition == 0 {
// this is most likely because in a test setup we set everything to ().
#[cfg(any(feature = "std", test))]
sp_io::TestExternalities::new_empty().execute_with(|| {
// This is the minimum value of the multiplier. Make sure that if we collapse to
// this value, we can recover with a reasonable amount of traffic. For this test we
// assert that if we collapse to minimum, the trend will be positive with a weight
// value which is 1% more than the target.
let min_value = T::FeeMultiplierUpdate::min();
let target = target + addition;
<frame_system::Pallet<T>>::set_block_consumed_resources(target, 0);
let next = T::FeeMultiplierUpdate::convert(min_value);
assert!(
next > min_value,
"The minimum bound of the multiplier is too low. When \
block saturation is more than target by 1% and multiplier is minimal then \
the multiplier doesn't increase."
);
});
impl<T: Config> Pallet<T>
where
BalanceOf<T>: FixedPointOperand,
{
/// Query the data that we know about the fee of a given `call`.
///
/// This pallet is not and cannot be aware of the internals of a signed extension, for example
/// a tip. It only interprets the extrinsic as some encoded value and accounts for its weight
/// and length, the runtime's extrinsic base weight, and the current fee multiplier.
///
/// All dispatchables must be annotated with weight and will have some fee info. This function
/// always returns.
pub fn query_info<Extrinsic: sp_runtime::traits::Extrinsic + GetDispatchInfo>(
unchecked_extrinsic: Extrinsic,
len: u32,
) -> RuntimeDispatchInfo<BalanceOf<T>>
T::Call: Dispatchable<Info = DispatchInfo>,
// NOTE: we can actually make it understand `ChargeTransactionPayment`, but would be some
// hassle for sure. We have to make it aware of the index of `ChargeTransactionPayment` in
// `Extra`. Alternatively, we could actually execute the tx's per-dispatch and record the
// balance of the sender before and after the pipeline.. but this is way too much hassle for
// a very very little potential gain in the future.
let dispatch_info = <Extrinsic as GetDispatchInfo>::get_dispatch_info(&unchecked_extrinsic);
let partial_fee = if unchecked_extrinsic.is_signed().unwrap_or(false) {
Self::compute_fee(len, &dispatch_info, 0u32.into())
} else {
// Unsigned extrinsics have no partial fee.
0u32.into()
};
let DispatchInfo { weight, class, .. } = dispatch_info;
RuntimeDispatchInfo { weight, class, partial_fee }
}
/// Query the detailed fee of a given `call`.
pub fn query_fee_details<Extrinsic: sp_runtime::traits::Extrinsic + GetDispatchInfo>(
unchecked_extrinsic: Extrinsic,
len: u32,
) -> FeeDetails<BalanceOf<T>>
where
T::Call: Dispatchable<Info = DispatchInfo>,
{
let dispatch_info = <Extrinsic as GetDispatchInfo>::get_dispatch_info(&unchecked_extrinsic);
let tip = 0u32.into();
if unchecked_extrinsic.is_signed().unwrap_or(false) {
Self::compute_fee_details(len, &dispatch_info, tip)
} else {
// Unsigned extrinsics have no inclusion fee.
FeeDetails { inclusion_fee: None, tip }
}
/// Compute the final fee value for a particular transaction.
pub fn compute_fee(len: u32, info: &DispatchInfoOf<T::Call>, tip: BalanceOf<T>) -> BalanceOf<T>
where
T::Call: Dispatchable<Info = DispatchInfo>,
Self::compute_fee_details(len, info, tip).final_fee()
}
/// Compute the fee details for a particular transaction.
pub fn compute_fee_details(
len: u32,
info: &DispatchInfoOf<T::Call>,
tip: BalanceOf<T>,
) -> FeeDetails<BalanceOf<T>>
where
T::Call: Dispatchable<Info = DispatchInfo>,
{
Self::compute_fee_raw(len, info.weight, tip, info.pays_fee, info.class)
}
/// Compute the actual post dispatch fee for a particular transaction.
///
/// Identical to `compute_fee` with the only difference that the post dispatch corrected
/// weight is used for the weight fee calculation.
pub fn compute_actual_fee(
len: u32,
info: &DispatchInfoOf<T::Call>,
post_info: &PostDispatchInfoOf<T::Call>,
tip: BalanceOf<T>,
) -> BalanceOf<T>
where
T::Call: Dispatchable<Info = DispatchInfo, PostInfo = PostDispatchInfo>,
{
Self::compute_actual_fee_details(len, info, post_info, tip).final_fee()
}
/// Compute the actual post dispatch fee details for a particular transaction.
pub fn compute_actual_fee_details(
len: u32,
info: &DispatchInfoOf<T::Call>,
post_info: &PostDispatchInfoOf<T::Call>,
tip: BalanceOf<T>,
) -> FeeDetails<BalanceOf<T>>
where
T::Call: Dispatchable<Info = DispatchInfo, PostInfo = PostDispatchInfo>,
{
Self::compute_fee_raw(
len,
post_info.calc_actual_weight(info),
tip,
post_info.pays_fee(info),
info.class,
)
}
fn compute_fee_raw(
len: u32,
weight: Weight,
tip: BalanceOf<T>,
pays_fee: Pays,
class: DispatchClass,
if pays_fee == Pays::Yes {
// the adjustable part of the fee.
let unadjusted_weight_fee = Self::weight_to_fee(weight);
let multiplier = Self::next_fee_multiplier();
// final adjusted weight fee.
let adjusted_weight_fee = multiplier.saturating_mul_int(unadjusted_weight_fee);
Stephen Shelton
committed
// length fee. this is adjusted via `LengthToFee`.
let len_fee = Self::length_to_fee(len);
let base_fee = Self::weight_to_fee(T::BlockWeights::get().get(class).base_extrinsic);
Stephen Shelton
committed
inclusion_fee: Some(InclusionFee { base_fee, len_fee, adjusted_weight_fee }),
FeeDetails { inclusion_fee: None, tip }
Stephen Shelton
committed
fn length_to_fee(length: u32) -> BalanceOf<T> {
T::LengthToFee::calc(&(length as Weight))
}
fn weight_to_fee(weight: Weight) -> BalanceOf<T> {
// cap the weight to the maximum defined in runtime, otherwise it will be the
// `Bounded` maximum of its data type, which is not desired.
let capped_weight = weight.min(T::BlockWeights::get().max_block);
T::WeightToFee::calc(&capped_weight)
impl<T> Convert<Weight, BalanceOf<T>> for Pallet<T>
where
BalanceOf<T>: FixedPointOperand,
{
/// Compute the fee for the specified weight.
///
/// This fee is already adjusted by the per block fee adjustment factor and is therefore the
/// share that the weight contributes to the overall fee of a transaction. It is mainly
/// for informational purposes and not used in the actual fee calculation.
fn convert(weight: Weight) -> BalanceOf<T> {
<NextFeeMultiplier<T>>::get().saturating_mul_int(Self::weight_to_fee(weight))
/// Require the transactor pay for themselves and maybe include a tip to gain additional priority
/// in the queue.
///
/// # Transaction Validity
///
/// This extension sets the `priority` field of `TransactionValidity` depending on the amount
/// of tip being paid per weight unit.
///
/// Operational transactions will receive an additional priority bump, so that they are normally
/// considered before regular transactions.
#[derive(Encode, Decode, Clone, Eq, PartialEq, TypeInfo)]
#[scale_info(skip_type_params(T))]
pub struct ChargeTransactionPayment<T: Config>(#[codec(compact)] BalanceOf<T>);
impl<T: Config> ChargeTransactionPayment<T>
where
T::Call: Dispatchable<Info = DispatchInfo, PostInfo = PostDispatchInfo>,
BalanceOf<T>: Send + Sync + FixedPointOperand,
{
/// utility constructor. Used only in client/factory code.
pub fn from(fee: BalanceOf<T>) -> Self {
Self(fee)
}
Loading full blame...