Malformatted L1 Deposit Log

[trianglesphere]

Setup

• Geth: proto/bellatrix-api-fixes
• L2 (OE) Geth: origin/optimism-prototype
• Rollup Node: input-deriv-test-cleanup
• Contracts: From the specs repo. I think they should be in sync

Sending a deposit Tx in the following way (want to mint L2 eth, no contracts or data).

deposit="0xdeaddeaddeaddeaddeaddeaddeaddeaddead0001"
abi="depositTransaction(address,uint256,uint256,bool,bytes)"
account="0x30ec912c5b1d14aa6d1cb9aa7a6682415c4f7eb0"

seth send --password signer_password.txt --keystore data_l1/keystore -F $account $deposit "$abi" $account $(seth --to-wei 1 eth) 1000000 false 0x

Result:

ERROR[01-13|10:40:42.016] Failed to extend L2 chain with new L1 block engine=0 l1=ef188f..1b3981:20 l2=8654b2..a97c39:19 err="failed to derive execution payload inputs: failed to derive user deposits: malformatted L1 deposit log: deposit event data too small (192 bytes): 
00000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000de0b6b3a764000000000000000000000000000000000000000000000000000000000000000f4240000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000a00000000000000000000000000000000000000000000000000000000000000000"

[protolambda]

Are you running the latest staging branch or another thing? Nvm, I see input-deriv-test-cleanup, that should be good, hmm

We added the mint param to the deposit log later on, so maybe it's failing to parse it because that code change is missing

[protolambda]

Found the off by one: 6*32 = 192, 7*32=224

https://github.com/ethereum-optimism/optimistic-specs/blob/212f234ee7bcfd5e813f9c6c2cd12058cfb0f989/opnode/l2/input_derivation.go#L47

Here it increments the offset 6 times with 32 bytes to get to the dynamic remainder of the log data. It starts at 0, not at 32 though, so it should be 6*32.

Will fix, thanks for the bug report!

[trianglesphere]

Ok. I'm going to double check the contract version. Skimming the L1 contract, it seems like it hasn't been updated in over a week (other than some solc changes)

[trianglesphere]

FYI there are 7 fields, but I'm less familiar with how dynamic length fields / bools are packed in

[trianglesphere]

Example log split at 32 byte intervals

address to
address from
u256 mint
u256 value
u256 gasLimit
bool creation
bytes data

0000000000000000000000000000000000000000000000000000000000000000   - to? 
0000000000000000000000000000000000000000000000000de0b6b3a7640000   - from?
00000000000000000000000000000000000000000000000000000000000f4240   - 1000000 (mint/value)
0000000000000000000000000000000000000000000000000000000000000000   - value?
00000000000000000000000000000000000000000000000000000000000000a0   - ??
0000000000000000000000000000000000000000000000000000000000000000   - empty data

[protolambda]

Updated input-deriv-test-cleanup and cherry-picked to staging

[protolambda]

0000000000000000000000000000000000000000000000000000000000000000  # value: 0 (transfer 0 on L2)
0000000000000000000000000000000000000000000000000de0b6b3a7640000  # mint: 0xde0b6b3a7640000 = 1000000000000000000 = 1 ETH (move 1 ETH from L1 to L2)
00000000000000000000000000000000000000000000000000000000000f4240  # gas: 0xf4240 = 1000000
0000000000000000000000000000000000000000000000000000000000000000  # isCreation (boolean set to false)
00000000000000000000000000000000000000000000000000000000000000a0  # data offset: 0xa0 = 160 = 32*5 bytes
0000000000000000000000000000000000000000000000000000000000000000  # data length: 0
# no data contents since length was 0

The to and from are separate from that data, since those are indexed log fields

[trianglesphere]

Ok. I got it. I used the new version and was able to see the balance of the account increase on L2

[trianglesphere]

Fixed in https://github.com/ethereum-optimism/optimistic-specs/pull/128/commits/225955bc37ece44e9a6c4aee5d76e3808a129a75