feat: add system config consensus to deprecate clique #72
Conversation
ranchalp
force-pushed
the
system-contract
branch
4 times, most recently
from
ececf8e to
e5f797e
Compare
ranchalp
force-pushed
the
system-contract
branch
from
e5f797e to
9cfaad5
Compare
| */ | ||
| contract SystemSignerRegistry is Ownable { | ||
| struct Signer { | ||
| uint64 startBlock; |
There was a problem hiding this comment.
Didn't we decide against having the startBlock here?
There was a problem hiding this comment.
iirc we discussed it here but didn't reach a conclusion
There was a problem hiding this comment.
I did not know that had been discussed (missed that slack conversation), and had not thought of the eventual syncing from L1 thing.
I guess it technically works, but I would advice to explicitly list a blockNum (even if we have power to decide whatever blockNum we want). The reason is that otherwise the zkproof must accept any of two valid keys for commits that happened after a transition from an old signer to the new one (since for all L1 knows the change happened at the first block after the previous commit or at the last one of the current commit?).
I do agree though that it is easier to implement for now to just allow switching and accept blocks in this transitional commit from either of the signers.
There was a problem hiding this comment.
The reason is that otherwise the zkproof must accept any of two valid keys
Which zkproof do you mean? If you mean when we finalize bundles, currently we don't prove anything about the block signatures.
There was a problem hiding this comment.
Ah, I see. Let me put it back with one signer then and forget about this orderly switching with blockNum for now (bringing it back in the future probably).
| address signer; | ||
| } | ||
|
|
||
| Signer[] private signers; |
There was a problem hiding this comment.
| Signer[] private signers; | |
| mapping (address => bool) public isAuthorizedSigner; |
There was a problem hiding this comment.
Need to consider how l2geth will read this contract.
- If it just polls the latest state asking "am I authorized?", then a mapping is suitable.
- If it just polls the latest state asking "what is the list of current authorized signers?", then an array is suitable.
- If it fetches events, then I'd still go with mapping.
There was a problem hiding this comment.
will we have multiple authorized signer in the same time?
There was a problem hiding this comment.
No, we should not have multiple authorized signers at the same time in this implementation: a new key will replace the old one even if specifying the same blockNum, although for this I need to make the list in DESC order. (if we go with the easier, optimism-like implementation, then we have multiple authorized sequencers actually, i.e. see this other comment).
There was a problem hiding this comment.
Having multiple makes key rotation a bit more convenient. But if we want to keep it simple, we can also just have one (currentAuthorizedSigner).
| * The getSigners() function returns parallel arrays for block numbers and addresses, | ||
| * | ||
| */ | ||
| contract SystemSignerRegistry is Ownable { |
There was a problem hiding this comment.
Do we want to make this upgradable?
| * The getSigners() function returns parallel arrays for block numbers and addresses, | ||
| * | ||
| */ | ||
| contract SystemSignerRegistry is Ownable { |
There was a problem hiding this comment.
Why not model this after OP's? https://github.com/ethereum-optimism/optimism/blob/develop/packages/contracts-bedrock/src/L1/SystemConfig.sol (see for example _setUnsafeBlockSigner).
@zimpha what do you think about storing multiple configs here (including gas coefficients, max gas limit, etc.)? Or is that too expensive gas-wise?
There was a problem hiding this comment.
we can add these config here, we can pack related parameters into single uint256. Then the gas overhead will be minimum.
There was a problem hiding this comment.
I meant the gas overhead of CALL (e.g. systemContract.getGasParams()), not just SLOAD, is that acceptable?
|
Added back one signer only. Not sure what is the error with foundry and the lib folder. |
ranchalp
force-pushed
the
system-contract
branch
5 times, most recently
from
5d9034e to
1227fb6
Compare
ranchalp
force-pushed
the
system-contract
branch
from
1227fb6 to
03dc404
Compare
ranchalp
force-pushed
the
system-contract
branch
2 times, most recently
from
9b0874d to
c18b9b3
Compare
ranchalp
force-pushed
the
system-contract
branch
from
c18b9b3 to
77c6e58
Compare
|
refactor to #75 |
In conjunction with go-ethereum#1102.