Move era reward minting to DAP with era-based allocation in Staking

[Ank4n]

Update (Apr 1, 2026)

This PR is broken down into smaller ones:

• Vested Payout trait and implementation #11512
• [Staking] Add issuance and budget traits in prep for Budget Split #11513
• Implement Budget Distribution logic in pallet-dap #11527
• [Staking] Move reward minting to DAP; payouts from drip-funded era pots #11616
• [Staking] Self stake incentive for Validators #11651
• (planned) Vested Payout implementation for validator self stake reward.

---

Partially implements: https://hackmd.io/@jonasW3F/rkN6BXE2ex
Close #10487 .

Overview

This PR moves era reward minting logic from pallet-staking-async to pallet-dap as well as introducing budget allocation management to DAP and a separate validator self-stake incentive mechanism to staking.

Changes

Reviewing in this order would help as well.

1. New Traits (sp-staking / frame-support)

• InflationCurve — replaces EraPayout. Takes (total_issuance, elapsed_millis), no staking state dependency.
• BudgetRecipient / BudgetRecipientList — register pot accounts with DAP. Runtime wires as a tuple.
• UnclaimedRewardSink — single deposit(source, amount) -> DispatchResult method. Implementor handles both the transfer and bookkeeping (e.g. deactivation).
• StakerRewardCalculator — commission-based staker split (replaces inline logic).
• VestedPayout (frame-support) — generic vested transfer using fungible traits.
• ValidatorIncentivePayout — adapter trait with ImmediateIncentivePayout and VestedIncentivePayout implementations.

2. Inflation Drip and Budget Distribution (pallet-dap)

DAP is a generic inflation drip + distribution engine. It has no dependency on staking.

drip_inflation() runs in on_initialize. It computes how much to mint via InflationCurve (total issuance + elapsed millis) and distributes to registered BudgetRecipients according to a BoundedBTreeMap<BudgetKey, Perbill> that must sum to exactly 100%. Rounding dust is not minted. Uses pallet_timestamp for time (validated by relay chain validators, see #9723).

Safety guards:

• MaxElapsedPerDrip — ceiling on elapsed time per drip. Prevents over-minting if AH gets stuck and comes back. Must be > InflationCadence (enforced by integrity_test).
• Migration — if LastInflationTimestamp is 0, initializes without dripping. For existing chains, MigrateV1ToV2 seeds this from ActiveEra.start.

Buffer accounting:

• Buffer funds are deactivated (InactiveIssuance) on every inflow — inflation mint, slash deposit (OnUnbalanced), and unclaimed reward transfers (UnclaimedRewardSink::deposit).
• Two root-only extrinsics (withdraw_buffer, deposit_buffer) let governance move funds between buffer and other accounts (e.g. treasury), with symmetric reactivation/deactivation.

3. Era Pot Lifecycle (staking-async)

DAP drips continuously into general pots. At era boundary (end_era), EraRewardManager::snapshot_era_rewards transfers accumulated balances from general pots into era-specific pot accounts (one for staker rewards, one for validator incentives). Era pots get provider refs to prevent reaping.

When an era expires past HistoryDepth, cleanup_era calls UnclaimedRewardSink::deposit to move remaining funds to the DAP buffer and drops the provider.

DisableLegacyMintingEra is set on first successful snapshot. After that, legacy mint-on-payout is blocked. This ensures a clean transition — old eras (before migration) still use legacy minting, new eras use era pots.

4. Reward Payout (staking-async)

do_payout_stakers_by_page checks whether the era has a reward pot. If yes, transfers from the era pot. If no (old eras), falls back to legacy minting (unless DisableLegacyMintingEra blocks it).

Staker reward split uses commission: StakerRewardCalculator::calculate_staker_reward computes commission_payout = commission * total_reward, splits the leftover proportionally by stake. Commission is capped by MaxCommission (governance-set storage item, defaults to 100%).

5. Validator Self-Stake Incentive Curve

A separate reward track funded from the validator incentive pot. Each validator gets a weight based on their self-stake using a sqrt-based piecewise function:

• Below optimum: w(s) = sqrt(s)
• Between optimum and cap: w(s) = sqrt(T + k² × (s - T)) — diminishing returns via SelfStakeSlopeFactor
• Above cap: plateau

Weights are stored in ErasValidatorIncentive during era planning. At payout, each validator's incentive share = (their_weight / total_weight) * incentive_budget * page_stake_part. The incentive is prorated across pages proportional to each page's stake relative to the validator's total exposed stake.

Config params (OptimumSelfStake, HardCapSelfStake, SelfStakeSlopeFactor) are set via set_validator_self_stake_incentive_config (root-only).

Initial deployment sets validator incentive budget to 0% — no behavior change until governance explicitly enables it.

6. Vested Incentive Payout

Validator incentive rewards are held under HoldReason::IncentiveVesting. On batch-boundary eras (era % BondingDuration == 0), accumulated holds are converted to a vesting schedule via pallet-vesting. The retroactive unlock fraction = BondingDuration / vesting_eras, with the remainder vesting linearly. If VestingDuration is zero, payouts are liquid.

Vesting uses relay chain block numbers (not parachain blocks) since parachain block times vary with elastic scaling.

This batching avoids exhausting MaxVestingSchedules. With bonding duration of 28 eras and vesting duration of 1 year, we create ~13 vesting schedules per year (well within the 28 limit). If the vested transfer fails, funds keep accumulating under hold until the next batch boundary.

Conversion triggers when era % BondingDuration == 0 AND all pages are claimed. If a validator claims their last page after the batch boundary, they wait until the next one. Funds remain held in the interim.

Runtime Integration

Runtimes need to make the following changes:

1. Add InflationCurve implementation to pallet-dap::Config.
2. Configure BudgetRecipients with (StakerRewardRecipient, ValidatorIncentiveRecipient).
3. Set BudgetOrigin for governance-controlled budget allocation updates.
4. Configure staking's ValidatorIncentivePayout, VestingDuration, and BlocksPerSession.
5. Update runtime inflation logic to return total era emission (without splitting).

Out of Scope

• Treasury Budget Management
• Stables payout to validators (and collators)

---

TODO

• DAP manages budgets
• Move minting to DAP
• Hard deprecate legacy minting in Staking pallet
• Era based pot accounts in Staking
• Separate Staker and Validator budget
• Track validator total self stake
• Self stake incentive calculation logic
• Vested Transfer of Validator Incentive Reward
• Max Commission
• Vesting E2E Tests
• Vesting batched converts
• Have two extrinsics in DAP pallet for outflow and inflow. Outflow activates the funds, Inflow deactivates it. Useful to use this flow everywhere, and extrinsics will be needed for transfers to/from treasury account in future.
• Combine MigrateV1toV2 in DAP pallet to also set initial budget.
• enable dap budget try state check
• Ensure this works as functionally no-op for Kusama.
• Missing Bench
• use VestedPayout::start_at in the past for unlocking 1 month of reward at first unlock.

Open Questions

• In event of AH getting stuck, era rewards might not be accurate based on RC era time.

[Ank4n]

Update: Splitting this PR into smaller pieces:

• Vested Payout trait and implementation #11512
• [Staking] Add issuance and budget traits in prep for Budget Split #11513
• Implement Budget Distribution logic in pallet-dap #11527

A final PR stacked on #11527 (either this or new if too many git conflicts) will integrate everything together with the staking changes.

What do you guys think @sigurpol @kianenigma ?

[sigurpol]

Update: Splitting this PR into smaller pieces:

• Vested Payout trait and implementation #11512
• [Staking] Add issuance and budget traits in prep for Budget Split #11513
• Implement Budget Distribution logic in pallet-dap #11527

A final PR stacked on #11527 (either this or new if too many git conflicts) will integrate everything together with the staking changes.

What do you guys think @sigurpol @kianenigma ?

+1 for small PRs and a final one (this / new one ) to wire all together! Will try to review all the first 3 PRs Monday the latest