Precision differences when calculating the _startAuction of funds accumulated in RiskFund

[code423n4]

Lines of code

https://github.com/code-423n4/2023-05-venus/blob/main/contracts/Shortfall/Shortfall.sol#L359-L434

Vulnerability details

Impact

When calculating _startAuction uint256 usdValue in RiskFund, Auction state divides the value of each market in the vToken list (in token list precision). This skew is fine for most tokens but will cause problems with certain token pairs.

https://github.com/code-423n4/2023-05-venus/blob/main/contracts/Shortfall/Shortfall.sol#L359-L434

Proof of Concept

When calculating the user's usdValue ratio, the result is that the list of certain tokens will be completely broken because of this.

Imagine a pair of tokens like USDC and SHIB. USDC has a token precision of 6 and SHIB has 18. If a user has a list balance of 100,001 SHIB (100.001e18) and a loan of 1 USDC (1e6):

1e6 * 1e18 / 100.001e18 = 0

There are two problems with this. First, most of these tokens will not work. Another issue is that because the usdValue user returns 0, there are circumstances where some token lists are impossible to enumerate by violating the protocol's key invariant.

Any token with very high or very low precision will suffer the consequences.

https://github.com/code-423n4/2023-05-venus/blob/8be784ed9752b80e6f1b8b781e2e6251748d0d7e/contracts/Shortfall/Shortfall.sol#L393

Tools Used

Manual Review

Recommended Mitigation Steps

Depending on the precision of each Token, the ratio is calculated dynamically

Assessed type

Decimal

[0xean]

The warden doesn't show how this actually impacts the protocol more than just be a loss of what amounts to dust (< 1 USD of value). Will downgrade and leave open for sponsor comment prior to moving to QA

[c4-judge]

0xean changed the severity to 2 (Med Risk)

[c4-judge]

0xean marked the issue as duplicate of #468

[c4-judge]

0xean marked the issue as satisfactory