Skip to content

test: integration multi staker #1136

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Draft
wants to merge 65 commits into
base: slashing-magnitudes-fixes
Choose a base branch
from

Conversation

8sunyuan
Copy link
Contributor

Motivation:

Currently integration tests mainly simulate generating a single staker delegated to an operator.

Modifications:

Added deploy functions and modified User.t.sol

Result:

Draft for writing a test with User[].
TODO:

  • Updated assertions used in clean way
  • Updated beaconChainMock to account for validators and multiple stakers

0xClandestine and others added 30 commits January 3, 2025 15:17
* feat: add share helpers

* fix: add deposit scaling factor

* fix: rebase
* fix: slashable window boundaries

* test: regression for alm

* test: update withdrawal delay not passed reversion

* test: burning indices

* refactor: switch conditionals

* fix: added unit tests

* test: assert slashable shares in queue

* fix: typos

---------

Co-authored-by: Yash Patil <[email protected]>
refactor small cleanup

chore: `forge fmt`

fix: `getQueuedWithdrawals` + test

fix: add constructor back

test: `totalQueued` > `withdrawal.strategies.length`

test(wip): `completeQueuedWithdrawals`

currently failing

fix: effectBlock

test(wip): @8sunyuan patch

fix: one flaky test

fix: second flaky test
* feat: initial deploy

* feat: slashing patch
* test(wip): todos

* fix: dealloc issue

* fix: remaining

* fix: forktest upgrade issue

* test: add `check_Withdrawal_AsShares_State_AfterSlash`

* refactor: cleanup

* fix: ci

* refactor: review changes
* docs: add slashing docs
* chore: bindings
* docs: fixed commenting and updated queue withdrawal docs
* docs: minor cleanup

---------

Co-authored-by: Nadir Akhtar <[email protected]>
* fix: correct expected share calc

* chore: bindings

* fix: rounding on failing unit test
* chore: clean comments and naming in dm

* refactor: simplify undelegate method
* feat: removed 0 address check because 0 stakers cant be delegated
* feat: condensed non-staker caller logic

* refactor: remove unnecessary check

* feat: use checks-effects-interactions when completing withdrawals
* feat: remove implicit public method for queuedWithdrawals and impl dedicated getter

* feat: deprecate withdrawer field

* chore: make bindings and clean compile errors

* refactor: redelegate reuses delegateTo and undelegate

* fix: broken integration test

* docs: update to reflect deprecated field

* feat: add getter for stakers withdrawal roots
* fix: initialization params

* fix: roll blocks usage
* fix: integration test initialization params (#978)

* fix: initialization params

* fix: roll blocks usage

* fix: `SignatureUtils` construction

---------

Co-authored-by: Yash Patil <[email protected]>
Co-authored-by: davidironblocks <[email protected]>
* fix: readd manual checks

* chore: forge fmt
* feat: add step 1

* feat: step 1 & 2 complete; pending step 3 sanity

* test: add `_validateProxyDomainSeparators`

* feat: add rc validation

---------

Co-authored-by: clandestine.eth <[email protected]>
* fix: update alloc delay bound

* test: remove unnecessary roll
* docs: shares accounting

* docs: fix gh markdown view

* docs: try fix gh again

* docs: cleanup

* docs: edit share accounting

* docs: wrap up share accounting doc

* docs: edit edge cases

---------

Co-authored-by: wadealexc <[email protected]>
* refactor: burning

* chore: fmt

* chore: update storage report

* chore: update readme

* refactor: add burnableShares for epm storage

* chore: update storage report
* docs: finish delegation manager docs

* docs: update docs readme

* docs: permission controller

* fix: small typos

* docs: address feedback

* docs: nit

---------

Co-authored-by: Michael Sun <[email protected]>
* docs: update StrategyManager docs with slashing delta

* docs: remove references to thirdPartyTransfersForbidden

* docs: update strategy docs to latest
* also various edits to docs and natspec

* chore: fmt and make bindings

---------

Co-authored-by: wadealexc <[email protected]>
ypatil12 and others added 19 commits February 3, 2025 12:15
* fix: ep negative shares bug

* fix: comments

* test: add integration tests for neg shares

* chore: remove logs

* chore: use already calculated delta

* chore: use stable foundry release in CI
…rator (#1051)

* feat: add OperatorSharesSlashed event to track shares slashed per operator

* feat: add unit tests

* fix: add more tests
* feat: add `getSharesFromQueuedWithdrawal`

* test: passing

* refactor(review): improve natspec

* refactor(review): maintain original interface

* test(review): add unit tests

* refactor(review): test empty

* refactor(review): test empty

* refactor(review): remove returned `Withdrawal`

* fix: use operator from `Withdrawal`

* test: use operator from `Withdrawal`

* chore: forge fmt
* docs: slashing factors rounding

* chore: forge fmt
* docs: small slash amounts

* docs: update contract docs

---------

Co-authored-by: wadealexc <[email protected]>
**Motivation:**

Fixes an issue where stakers delegating Beacon Chain ETH from slashed
Eigen Pods were able to delegate more shares than they should.
Specifically, operators now are delegated a staker's
`withdrawableShares` rather than their `depositShares`.

**Modifications:**

- Changed accounting logic on delegation in `DelegationManger.sol`
- `DepositScalingFactor` now resets when a staker withdraws all their
shares, whether through undelegation, redelegation, or a simple
withdrawal
- Changes in `StrategyManager.sol`, `IShareManager.sol`,
`SlashingLib.sol`, and `EigenPodManager.sol` to accommodate new
accounting
- New test files and changes to others to reflect new accounting and
invariants
- Updated `docs/SharesAccounting.md`

**Result:**

System is now robust to stakers with arbitrary EigenPod states

---------

Co-authored-by: Michael <[email protected]>
Co-authored-by: Michael Sun <[email protected]>
Co-authored-by: wadealexc <[email protected]>
Co-authored-by: Yash Patil <[email protected]>
Co-authored-by: clandestine.eth <[email protected]>
**Motivation:**  

Audit report flagged that function selector-based permissions may break
on upgrades. This PR documents the limitation and its implications while
improving NatSpec for clarity. (EGSL-15)

**Modifications:**  

- Documented function selector upgrade invalidations.  
- Improved NatSpec comments in `IPermissionController`.  

**Result:**  

Clearer documentation on function selector limitations and enhanced
NatSpec for better code clarity.
### Changes:

- *Dynamic Domain Separator:* `SignatureUtils.domainSeparator()` is now
recomputed for each signature verification. This eliminates the need for
storing initial values in storage or as immutables, which is important
for beacon proxy support.

- ~*Version Bump Command:* Introduced `make bump-version VERSION=2`,
which automatically updates the version function's return values.~

- *Version Fn + Constructor Param:* Adds an immutable oz `ShortString`
that's set in the constructor.
**Motivation:**

Concerns about reentrancy in the DelegationManager and interactions of
completed withdrawals which can call untrusted ERC20 transfers

**Modifications:**

Added reentrant guards across external functions

**Result:**

Preventing cross-function reentrancy in the DelegationManager

---------

Co-authored-by: wadealexc <[email protected]>
**Motivation:**

Fixes an issue arbitrary external contracts could be called via
`StrategyManager.burnShares`. (Certora L-04)

**Modifications:**

`StrategyManager.burnShares` does not do an external call if the
burnable share amount is zero

**Result:**

Should no longer be possible to call untrusted code directly through
`burnShares`
**Motivation:**

Document edge cases around BC/AVS Slashing. 

**Modifications:**

Update docs with justification. 

**Result:**

Clear edge case callouts.
require avs register metadata in allocation manager before they can
create operatorset

---------

Co-authored-by: clandestine.eth <[email protected]>
**Motivation:**

Current fn only returns scaled shares, which leads integrators to making
two calls. This is expensive in terms of gas.

**Modifications:**

- `getSharesFromQueuedWithdrawal` has been renamed to
`getQueuedWithdrawalFromRoot` and now also returns `Withdrawal` struct.

**Result:**

Integrators can fetch both in a single call.
**Motivation:**

Improve slashing invariants in integration tests

**Modifications:**

Adds `check_Base_Slashing_State`, and implements several checks used
within

**Result:**

Slashing invariants check all manner of state changes in the ALM and
delegation.

---------

Co-authored-by: Michael <[email protected]>
**Motivation:**

Naming between the `getQueuedWIthdrawal` aliases was inconsistent.

**Modifications:**

- ~made `queuedWithdrawals[withdrawalRoot]` mapping public.~
- renamed `queuedWithdrawals` -> `_queuedWithdrawals`.
- added `_queuedWithdrawals` getter
- removed previous `getQueuedWithdrawal` alias.
- renamed `getQueuedWithdrawalFromRoot` to `getQueuedWithdrawal`. 

**Result:**

Consistent function naming.

---------

Co-authored-by: Yash Patil <[email protected]>
Comment on lines +309 to +327
function handleDelegatingStakers(address staker) public virtual {
// check that the current User is an operator
require(
delegationManager.isOperator(address(this)),
"User is not an operator"
);
delegatedStakers.add(staker);
}

/// @dev NOTE: do NOT call this function directly. This is simply implemented to manage the
/// delegatedStakers enum set. This function is only meant to be called on the Operator User contract.
function handleUndelegatingStaker(address staker) public virtual {
// check that the current User is an operator
require(
delegationManager.isOperator(address(this)),
"User is not an operator"
);
delegatedStakers.remove(staker);
}
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I would call these add/remove rather than "handle"

Comment on lines +311 to +314
require(
delegationManager.isOperator(address(this)),
"User is not an operator"
);
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Let's leave the checks out of this. If you're failing these, you'll see failures elsewhere - so let's keep these methods as simple as possible.

@@ -713,6 +761,15 @@ contract User is Logger, IDelegationManagerTypes, IAllocationManagerTypes {
return activeValidators;
}

function getDelegatedStakers() public view returns (User[] memory) {
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

    function getDelegatedStakers() public view returns (User[] memory stakers) {
        address[] memory _stakers = delegatedStakers.values();
        assembly { stakers := _stakers }
    }

Comment on lines +2330 to +2331
IStrategy[][] memory strategies,
uint[][] memory tokenBalances
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The arrayification of integration tests is getting unwieldy. It might be time to start introducing some helper structs.


// 4. Complete withdrawal
_rollBlocksForCompleteWithdrawals(withdrawals[0]);
stakers.completeWithdrawalsAsTokens(withdrawals);
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Hmm. The problem with this is the snapshot system is only going to capture the final staker to make this call. So if we did an invariant check after this, the Snap assertions will be using the last staker in the list.

I think it might be better to continue handling each staker individually in terms of the actions they take. After all, we want to check invariants on each, right?

@ypatil12 ypatil12 force-pushed the slashing-magnitudes-fixes branch 4 times, most recently from fb84edf to 44487a0 Compare February 20, 2025 23:50
@ypatil12 ypatil12 added 🗡️ Slashing Release Changes for the slashing release. 🧪 Test Test-related changes (unit, integration, etc.). labels Feb 21, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
🗡️ Slashing Release Changes for the slashing release. 🧪 Test Test-related changes (unit, integration, etc.).
Projects
None yet
Development

Successfully merging this pull request may close these issues.

9 participants