feat(Linea): Add CCTP V2 support to Linea #910

nicholaspai · 2025-03-05T19:43:28Z

This PR adds a new CCTPV2 Adapter that we add to the Linea_Adapter and Linea_SpokePool.

CCTP V2 has a different interface than V1 but it has a similar on-chain flow--so we can re-use some of the code.

contracts/Linea_SpokePool.sol

test/evm/hardhat/chain-specific-spokepools/PolygonZkEVM_SpokePool.ts

contracts/Linea_SpokePool.sol

contracts/chain-adapters/Linea_Adapter.sol

Co-authored-by: Paul <108695806+pxrl@users.noreply.github.com>

pxrl

This looks fine to me 👍

contracts/Linea_SpokePool.sol

* fix(ZkSync_SpokePool): Add __gap This contract gets extended by the Lens_SpokePool which doesn't add any storage but we should add it in case a future variable gets added to the Lens_SpokePool * Update ZkSync_SpokePool.json

Very messy, but allows us to switch between V1 and V2 at construction time without contract changes

contracts/libraries/CircleCCTPAdapter.sol

…USDT0` via `OFT` messaging protocol (#902) * first draft of OFTTransportAdapter Signed-off-by: Ihor Farion <ihor@umaproject.org> * update yarn.lock file Signed-off-by: Ihor Farion <ihor@umaproject.org> * Revert "update yarn.lock file" This reverts commit 4c216ea. Signed-off-by: Ihor Farion <ihor@umaproject.org> * add yarn.lock compatible with master Signed-off-by: Ihor Farion <ihor@umaproject.org> * polish OFTTransportAdapter, add OFT support to Arbitrum_Adapter on L1, and Arbitrum_SpokePool on L2 Signed-off-by: Ihor Farion <ihor@umaproject.org> * polish + fix missing approval Signed-off-by: Ihor Farion <ihor@umaproject.org> * add context for dstEid Signed-off-by: Ihor Farion <ihor@umaproject.org> * address most of the PR comments about contracts Signed-off-by: Ihor Farion <ihor@umaproject.org> * update deploy scripts, add tests for OFT messaging, polish contracts Signed-off-by: Ihor Farion <ihor@umaproject.org> * cleanup comments and extraneous log file Signed-off-by: Ihor Farion <ihor@umaproject.org> * revert package.json prepublish change Signed-off-by: Ihor Farion <ihor@umaproject.org> * generalize oft adapter to support multiple tokens. Introduce OFTAddressBook to support that. Update deploy / test scripts to reflect new functionality Signed-off-by: Ihor Farion <ihor@umaproject.org> * add __gap to ArbitrumSpokePool, update stale comments on OFTTransportAdapter, update layouts of ArbitrumSpokePool and AlephZeroSpokePool Signed-off-by: Ihor Farion <ihor@umaproject.org> * update some comments; adjust fee cap naming Signed-off-by: Ihor Farion <ihor@umaproject.org> * address PR comments Signed-off-by: Ihor Farion <ihor@umaproject.org> * address PR comments Signed-off-by: Ihor Farion <ihor@umaproject.org> * fix deploy script, remove incorrect values from consts Signed-off-by: Ihor Farion <ihor@umaproject.org> * improve comment Signed-off-by: Ihor Farion <ihor@umaproject.org> * add oftFeeCap as a param to arbitrum adapter construction Signed-off-by: Ihor Farion <ihor@umaproject.org> * move OFT functionality to SpokePool for easy further integration and removing boilerplate code Signed-off-by: Ihor Farion <ihor@umaproject.org> * remove layerzero from foundry remappings Signed-off-by: Ihor Farion <ihor@umaproject.org> * address licensing comment; add permalink to LZ OFT code on github Signed-off-by: Ihor Farion <ihor@umaproject.org> * fix a couple of comment typos Signed-off-by: Ihor Farion <ihor@umaproject.org> --------- Signed-off-by: Ihor Farion <ihor@umaproject.org>

contracts/chain-adapters/Linea_Adapter.sol

deploy/028_deploy_linea_adapter.ts

mrice32

LGTM!

contracts/libraries/CircleCCTPAdapter.sol

mrice32 · 2025-03-13T15:09:38Z

contracts/libraries/CircleCCTPAdapter.sol

+        // implement feeRecipient but it has a fallback function so to be extra safe, we check the return value
+        // of feeRecipient() as well.
+        (bool success, bytes memory feeRecipient) = address(cctpTokenMessenger).staticcall(
+            abi.encodeWithSignature("feeRecipient()")


This is something I want to highlight to OZ.

We are doing this to avoid adding an extra constructor argument and, thus, create a larger diff for all the contracts that inherit from this CCTP contract. However, the downside is that this could be broken in the future if circle removes this function.

deploy/028_deploy_linea_adapter.ts

deploy/029_deploy_linea_spokepool.ts

The base branch was changed.

This reverts commit f8729d3, reversing changes made to 71b6b2a.

* feat(Linea): Add CCTP V2 support to Linea (#910) * fix(CCTPAdapter): Fix feeRecipient check to determine V2 vs V1 - The casting to bytes20 will fetch the first 20 bytes of the return value. However, the functions in Solidity return abi.encoded data, which means in particular, that if an address is returned from the feeRecipient function, the returned feeRecipient.length will be equal to 32 and the address will be included in the last 20 bytes of that data (will be padded with 12 zeroes at the beginning). As a result, casting feeRecipient to bytes20 will capture only the first 8 bytes of the returned address. You can get the full address by casting the data to uint160 instead (you may need to cast it to bytes32 and uint256 first). - The feeRecipient.length is not validated. It means that for example, in case of a call to nonexistent contract or a call to a contract with a fallback function which doesn't return any data, the code will attempt to cast empty data to bytes20. It seems that casting empty bytes object to bytes20 will return 0, but it would be safer to not rely on that behaviour and validate that the length of feeRecipient object equals 32. - Even after fixing the issues above, this check will not be fully reliable as it could for example succeed in case when a TokenMessenger V1 contract had a fallback function returning 32 bytes. This is a very low risk, hence you may keep this logic if you accept the risk, but a fully correct solution would involve introducing an additional parameter to the constructor, which could be then used to determine the CCTP version (this could be an uint256 as it is possible that more CCTP versions will appear in the future). We understand that you wanted to avoid adding this extra parameter (and hence introducing a very large diff), but the most reliable solution would involve introducing it in the constructor. * Revert "fix(CCTPAdapter): Fix feeRecipient check to determine V2 vs V1" This reverts commit cd36e36. * fix(CCTPAdapter): Fix feeRecipient check to determine V2 vs V1 (#924) - The casting to bytes20 will fetch the first 20 bytes of the return value. However, the functions in Solidity return abi.encoded data, which means in particular, that if an address is returned from the feeRecipient function, the returned feeRecipient.length will be equal to 32 and the address will be included in the last 20 bytes of that data (will be padded with 12 zeroes at the beginning). As a result, casting feeRecipient to bytes20 will capture only the first 8 bytes of the returned address. You can get the full address by casting the data to uint160 instead (you may need to cast it to bytes32 and uint256 first). - The feeRecipient.length is not validated. It means that for example, in case of a call to nonexistent contract or a call to a contract with a fallback function which doesn't return any data, the code will attempt to cast empty data to bytes20. It seems that casting empty bytes object to bytes20 will return 0, but it would be safer to not rely on that behaviour and validate that the length of feeRecipient object equals 32. - Even after fixing the issues above, this check will not be fully reliable as it could for example succeed in case when a TokenMessenger V1 contract had a fallback function returning 32 bytes. This is a very low risk, hence you may keep this logic if you accept the risk, but a fully correct solution would involve introducing an additional parameter to the constructor, which could be then used to determine the CCTP version (this could be an uint256 as it is possible that more CCTP versions will appear in the future). We understand that you wanted to avoid adding this extra parameter (and hence introducing a very large diff), but the most reliable solution would involve introducing it in the constructor. * improve(CCTPAdapter): Add comment about not casting tokenMinter to V1/V2 (#928) * improve(cctpAdapter): Fix misleading comments (#929) * improve(cctpAdapter): typographical errors (#930) * improve(CCTPAdapter): Remove comment (#931) * fix(CCTPAdapter): Fix feeRecipient check to determine V2 vs V1 - The casting to bytes20 will fetch the first 20 bytes of the return value. However, the functions in Solidity return abi.encoded data, which means in particular, that if an address is returned from the feeRecipient function, the returned feeRecipient.length will be equal to 32 and the address will be included in the last 20 bytes of that data (will be padded with 12 zeroes at the beginning). As a result, casting feeRecipient to bytes20 will capture only the first 8 bytes of the returned address. You can get the full address by casting the data to uint160 instead (you may need to cast it to bytes32 and uint256 first). - The feeRecipient.length is not validated. It means that for example, in case of a call to nonexistent contract or a call to a contract with a fallback function which doesn't return any data, the code will attempt to cast empty data to bytes20. It seems that casting empty bytes object to bytes20 will return 0, but it would be safer to not rely on that behaviour and validate that the length of feeRecipient object equals 32. - Even after fixing the issues above, this check will not be fully reliable as it could for example succeed in case when a TokenMessenger V1 contract had a fallback function returning 32 bytes. This is a very low risk, hence you may keep this logic if you accept the risk, but a fully correct solution would involve introducing an additional parameter to the constructor, which could be then used to determine the CCTP version (this could be an uint256 as it is possible that more CCTP versions will appear in the future). We understand that you wanted to avoid adding this extra parameter (and hence introducing a very large diff), but the most reliable solution would involve introducing it in the constructor. * Remove comment about cctp domain IDs * fix(CCTPAdapter): Typographical errors (#935)

nicholaspai added 3 commits March 5, 2025 14:42

feat(Linea): Add CCTP support to Linea

3bd74f4

deploy scripts

be61552

fix tests

b5cdf31

nicholaspai commented Mar 5, 2025

View reviewed changes

contracts/Linea_SpokePool.sol Show resolved Hide resolved

nicholaspai added 6 commits March 5, 2025 15:00

Update Linea_Adapter.ts

833fa10

Fix tests

f5cb584

Update Linea_SpokePool.sol

fb10db9

Update CircleCCTPAdapter.sol

b6f0929

Update Linea_SpokePool.json

ba40d92

Update PolygonZkEVM_SpokePool.ts

1b5960f

nicholaspai commented Mar 5, 2025

View reviewed changes

test/evm/hardhat/chain-specific-spokepools/PolygonZkEVM_SpokePool.ts Show resolved Hide resolved

nicholaspai added the needs to be audited label Mar 5, 2025

pxrl reviewed Mar 5, 2025

View reviewed changes

contracts/Linea_SpokePool.sol Show resolved Hide resolved

contracts/Linea_SpokePool.sol Outdated Show resolved Hide resolved

contracts/chain-adapters/Linea_Adapter.sol Outdated Show resolved Hide resolved

contracts/chain-adapters/Linea_Adapter.sol Show resolved Hide resolved

nicholaspai and others added 2 commits March 5, 2025 15:53

Update contracts/Linea_SpokePool.sol

33955f3

Co-authored-by: Paul <108695806+pxrl@users.noreply.github.com>

Update contracts/chain-adapters/Linea_Adapter.sol

d439eaf

Co-authored-by: Paul <108695806+pxrl@users.noreply.github.com>

nicholaspai requested a review from pxrl March 5, 2025 20:55

pxrl reviewed Mar 6, 2025

View reviewed changes

pxrl previously approved these changes Mar 6, 2025

View reviewed changes

nicholaspai changed the base branch from master to march-25-evm-audit March 6, 2025 15:06

bmzig reviewed Mar 6, 2025

View reviewed changes

contracts/Linea_SpokePool.sol Show resolved Hide resolved

nicholaspai changed the title ~~feat(Linea): Add CCTP support to Linea~~ feat(Linea): Add CCTP V2 support to Linea Mar 7, 2025

nicholaspai added 2 commits March 7, 2025 11:33

Use V2 CCTP contracts for Linea

0d436fd

fix tests

f683c90

nicholaspai marked this pull request as ready for review March 7, 2025 19:25

nicholaspai requested a review from mrice32 as a code owner March 7, 2025 19:25

nicholaspai requested review from bmzig and pxrl March 9, 2025 16:18

nicholaspai and others added 3 commits March 10, 2025 18:05

fix(ZkSync_SpokePool): Add __gap (#907)

9f7e40c

* fix(ZkSync_SpokePool): Add __gap This contract gets extended by the Lens_SpokePool which doesn't add any storage but we should add it in case a future variable gets added to the Lens_SpokePool * Update ZkSync_SpokePool.json

Switch CCTP V2 in the constructor

17b6ccf

Very messy, but allows us to switch between V1 and V2 at construction time without contract changes

Update CircleCCTPAdapter.sol

51dde94

nicholaspai requested a review from mrice32 March 11, 2025 19:02

bmzig reviewed Mar 12, 2025

View reviewed changes

contracts/libraries/CircleCCTPAdapter.sol Outdated Show resolved Hide resolved

bmzig previously approved these changes Mar 12, 2025

View reviewed changes

grasphoper and others added 4 commits March 12, 2025 16:32

Update CircleCCTPAdapter.sol

71b6b2a

Merge branch 'march-25-evm-audit' into linea-cctp

f8729d3

Comments about CCTP V2 unknown addresses

9f1064b

pxrl reviewed Mar 13, 2025

View reviewed changes

contracts/chain-adapters/Linea_Adapter.sol Outdated Show resolved Hide resolved

pxrl reviewed Mar 13, 2025

View reviewed changes

deploy/028_deploy_linea_adapter.ts Outdated Show resolved Hide resolved

mrice32 approved these changes Mar 13, 2025

View reviewed changes

pxrl reviewed Mar 13, 2025

View reviewed changes

deploy/028_deploy_linea_adapter.ts Outdated Show resolved Hide resolved

pxrl reviewed Mar 13, 2025

View reviewed changes

deploy/029_deploy_linea_spokepool.ts Outdated Show resolved Hide resolved

Respond to comments

f908e38

nicholaspai requested review from bmzig, mrice32 and pxrl March 13, 2025 17:12

mrice32 previously approved these changes Mar 13, 2025

View reviewed changes

nicholaspai changed the base branch from march-25-evm-audit to master March 13, 2025 17:33

nicholaspai added 5 commits March 13, 2025 13:33

Revert "Merge branch 'march-25-evm-audit' into linea-cctp"

039870d

This reverts commit f8729d3, reversing changes made to 71b6b2a.

Merge branch 'master' into linea-cctp

08ee557

Update README.md

7e40e02

Create AcrossOriginSettler.sol

b2e634c

Delete ERC7683OrderDepositorExternal.sol

30e2754

nicholaspai changed the base branch from master to march-25-evm-audit-cctpv2 March 13, 2025 17:40

nicholaspai merged commit 0720878 into march-25-evm-audit-cctpv2 Mar 13, 2025
9 checks passed

nicholaspai deleted the linea-cctp branch March 13, 2025 18:24

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

feat(Linea): Add CCTP V2 support to Linea #910

feat(Linea): Add CCTP V2 support to Linea #910

Uh oh!

nicholaspai commented Mar 5, 2025 •

edited

Loading

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

pxrl left a comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

mrice32 left a comment

Uh oh!

Uh oh!

mrice32 Mar 13, 2025

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

feat(Linea): Add CCTP V2 support to Linea #910

feat(Linea): Add CCTP V2 support to Linea #910

Uh oh!

Conversation

nicholaspai commented Mar 5, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

pxrl left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

mrice32 left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

mrice32 Mar 13, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

nicholaspai commented Mar 5, 2025 •

edited

Loading