Skip to content

End epoch cannot be triggered preventing winners to withdraw #278

@code423n4

Description

@code423n4

Lines of code

https://github.com/code-423n4/2022-09-y2k-finance/blob/2175c044af98509261e4147edeb48e1036773771/src/Controller.sol#L198
https://github.com/code-423n4/2022-09-y2k-finance/blob/2175c044af98509261e4147edeb48e1036773771/src/Controller.sol#L246
https://github.com/code-423n4/2022-09-y2k-finance/blob/2175c044af98509261e4147edeb48e1036773771/src/Controller.sol#L261
https://github.com/code-423n4/2022-09-y2k-finance/blob/2175c044af98509261e4147edeb48e1036773771/src/Controller.sol#L277-L286
https://github.com/code-423n4/2022-09-y2k-finance/blob/2175c044af98509261e4147edeb48e1036773771/src/Vault.sol#L203

Vulnerability details

Impact

At the end of an epoch, the triggerEndEpoch(...) is called to trigger 'epoch end without depeg event', making risk users the winners and entitling them to withdraw (risk + hedge) from the vault.
In the case of the Arbitrum sequencer going down or restarting, there is a grace period of one hour before the getLatestPrice() returns to execute without reverting. This means that the triggerEndEpoch(...) cannot complete during this time, because it calls the getLatestPrice().

Making this high-priority because unless the triggerEndEpoch(...) completes:

First two points each constitute independent jsutification, thrid point reinforces the first 2 points.

Proof of Concept

triggerEndEpoch reverts if arbiter down or restarted less than eq GRACE_PERIOD_TIME ago (1hr)

File: Controller.sol:L246

Revert if getLatestPrice reverts.

function triggerEndEpoch(uint256 marketIndex, uint256 epochEnd) public {
    
    < ... omitted ... >

    emit DepegInsurance(
        keccak256(
            abi.encodePacked(
                marketIndex,
                insrVault.idEpochBegin(epochEnd),
                epochEnd
            )
        ),
        tvl,
        false,
        epochEnd,
        block.timestamp,
        getLatestPrice(insrVault.tokenInsured()) // @audit getLatestPrice reverts while sequencer unavailable or during grace period
    );
}

File: Controller.sol:L277-L286

Revert if sequencer down or grace period after restart not over.

function getLatestPrice(address _token)
    public
    view
    returns (int256 nowPrice)
{
    < ... omitted ... >

    bool isSequencerUp = answer == 0;
    if (!isSequencerUp) {
        revert SequencerDown();
    }

    // Make sure the grace period has passed after the sequencer is back up.
    uint256 timeSinceUp = block.timestamp - startedAt;
    if (timeSinceUp <= GRACE_PERIOD_TIME) { // @audit 1 hour
        revert GracePeriodNotOver();
    }

    < ... omitted ... >
}

withdraw fails if triggerEndEpoch did not execute successfully

File: Vault.sol:L203

Can execute if block.timestamp > epochEnd, but fails if trigger did not execute. Winners cannot withdraw.

function withdraw(
    uint256 id,
    uint256 assets,
    address receiver,
    address owner
)
    external
    override
    epochHasEnded(id) // @audit same as require((block.timestamp > id) || idDepegged[id]), hence independent from triggers.
    marketExists(id)
    returns (uint256 shares)
{
    < ... omitted ... >

    uint256 entitledShares = beforeWithdraw(id, shares); // @audit ratio is idClaimTVL[id]/ifFinalTVL[id], hence zero unless triggers executed
    
    < ... omitted ... >

    emit Withdraw(msg.sender, receiver, owner, id, assets, entitledShares);
    asset.transfer(receiver, entitledShares);

    return entitledShares;
}

Tools Used

n/a

Recommended Mitigation Steps

The latest price is retrieved at the very end of the triggerEndEpoch(...) for the only purpose of initializing the DepegInsurance event.
Since it is used for informational purpose (logging / offchain logging) and not for functional purpose to the triggerEndEpoch(...) execution, it can be relaxed.

Depending on how the event is used, when getLatestPrice() is called for informative/logging purpose only, there could be few alternatives:

  • log a 0 when SequencerDown or GRACE_PERIOD_TIME not passed
  • log a 0 when SequencerDown and ignore GRACE_PERIOD_TIME
    Once events are logged off-chain, some post processing may be used to correct/update the values with accurate data.

Metadata

Metadata

Assignees

No one assigned

    Labels

    3 (High Risk)Assets can be stolen/lost/compromised directlybugSomething isn't workinghigh quality reportThis report is of especially high qualityselected for reportThis submission will be included/highlighted in the audit reportsponsor confirmedSponsor agrees this is a problem and intends to fix it (OK to use w/ "disagree with severity")

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions