Skip to content

Commit

Permalink
refactor: updating encoding of TxEffects (#4726)
Browse files Browse the repository at this point in the history 
Encoding of an `L2Block` was done unnaturally because it was implemented
before we had the notion of tx effects. This PR refactors all of that so
that it makes sense again.

Nuked the old `Decoder` because it made sense to do this now.
  • Loading branch information
@benesjan
benesjan authored Feb 28, 2024
1 parent 426513e commit 29b1ea3
Show file tree
Hide file tree
Showing 25 changed files with 538 additions and 959 deletions.
196 changes: 58 additions & 138 deletions l1-contracts/slither_output.md
View file

Large diffs are not rendered by default.

418 changes: 0 additions & 418 deletions l1-contracts/src/core/libraries/decoders/Decoder.sol
View file

This file was deleted.

152 changes: 105 additions & 47 deletions l1-contracts/src/core/libraries/decoders/MessagesDecoder.sol
View file
Original file line number Diff line number Diff line change
Expand Up @@ -16,29 +16,37 @@ import {Hash} from "../Hash.sol";
* -------------------
* You can use https://gist.github.com/LHerskind/724a7e362c97e8ac2902c6b961d36830 to generate the below outline.
* -------------------
* L2 Body Data specification
* L2 Body Data Specification
* -------------------
*
* | byte start | num bytes | name
* | --- | --- | ---
* | 0x00 | 0x04 | len(newNoteHashes) (denoted a)
* | 0x04 | a * 0x20 | newNoteHashes
* | 0x04 + a * 0x20 | 0x04 | len(newNullifiers) (denoted b)
* | 0x08 + a * 0x20 | b * 0x20 | newNullifiers
* | 0x08 + a * 0x20 + b * 0x20 | 0x04 | len(newPublicDataWrites) (denoted c)
* | 0x0c + a * 0x20 + b * 0x20 | c * 0x40 | newPublicDataWrites
* | 0x0c + a * 0x20 + b * 0x20 + c * 0x40 | 0x04 | len(newL2ToL1Msgs) (denoted d)
* | 0x10 + a * 0x20 + b * 0x20 + c * 0x40 | d * 0x20 | newL2ToL1Msgs
* | 0x10 + a * 0x20 + b * 0x20 + c * 0x40 + d * 0x20 | 0x04 | len(contracts) (denoted e)
* | 0x14 + a * 0x20 + b * 0x20 + c * 0x40 + d * 0x20 | e * 0x20 | newContracts
* | 0x14 + a * 0x20 + b * 0x20 + c * 0x40 + d * 0x20 + e * 0x20 | e * 0x34 | newContractsData
* | 0x14 + a * 0x20 + b * 0x20 + c * 0x40 + d * 0x20 + e * 0x54 | 0x04 | len(newL1ToL2Msgs) (denoted f)
* | 0x18 + a * 0x20 + b * 0x20 + c * 0x40 + d * 0x20 + e * 0x54 | f * 0x20 | newL1ToL2Msgs
* | 0x18 + a * 0x20 + b * 0x20 + c * 0x40 + d * 0x20 + e * 0x54 + f * 0x20 | 0x04 | byteLen(newEncryptedLogs) (denoted g)
* | 0x1c + a * 0x20 + b * 0x20 + c * 0x40 + d * 0x20 + e * 0x54 + f * 0x20 | g | newEncryptedLogs
* | 0x1c + a * 0x20 + b * 0x20 + c * 0x40 + d * 0x20 + e * 0x54 + f * 0x20 + g | 0x04 | byteLen(newUnencryptedLogs) (denoted h)
* | 0x20 + a * 0x20 + b * 0x20 + c * 0x40 + d * 0x20 + e * 0x54 + f * 0x20 + g | h | newUnencryptedLogs
* | --- | --- | ---
* | byte start | num bytes | name
* | --- | --- | ---
* | 0x0 | 0x4 | len(newL1ToL2Msgs) (denoted a)
* | 0x4 | a * 0x20 | newL1ToL2Msgs
* | 0x4 + a * 0x20 = tx0Start | 0x4 | len(numTxs) (denoted t)
* | | | TxEffect 0 {
* | tx0Start | 0x1 | len(newNoteHashes) (denoted b)
* | tx0Start + 0x1 | b * 0x20 | newNoteHashes
* | tx0Start + 0x1 + b * 0x20 | 0x1 | len(newNullifiers) (denoted c)
* | tx0Start + 0x1 + b * 0x20 + 0x1 | c * 0x20 | newNullifiers
* | tx0Start + 0x1 + b * 0x20 + 0x1 + c * 0x20 | 0x1 | len(newL2ToL1Msgs) (denoted d)
* | tx0Start + 0x1 + b * 0x20 + 0x1 + c * 0x20 + 0x1 | d * 0x20 | newL2ToL1Msgs
* | tx0Start + 0x1 + b * 0x20 + 0x1 + c * 0x20 + 0x1 + d * 0x20 | 0x1 | len(newPublicDataWrites) (denoted e)
* | tx0Start + 0x1 + b * 0x20 + 0x1 + c * 0x20 + 0x1 + d * 0x20 + 0x01 | e * 0x40 | newPublicDataWrites
* | tx0Start + 0x1 + b * 0x20 + 0x1 + c * 0x20 + 0x1 + d * 0x20 + 0x01 + e * 0x40 | 0x1 | len(contracts) (denoted f)
* | tx0Start + 0x1 + b * 0x20 + 0x1 + c * 0x20 + 0x1 + d * 0x20 + 0x01 + e * 0x40 + 0x1 | f * 0x20 | newContracts
* | tx0Start + 0x1 + b * 0x20 + 0x1 + c * 0x20 + 0x1 + d * 0x20 + 0x01 + e * 0x40 + 0x1 + f * 0x20 | f * 0x34 | newContractsData
* | tx0Start + 0x1 + b * 0x20 + 0x1 + c * 0x20 + 0x1 + d * 0x20 + 0x01 + e * 0x40 + 0x1 + f * 0x20 + f * 0x34 | 0x04 | byteLen(newEncryptedLogs) (denoted g)
* | tx0Start + 0x1 + b * 0x20 + 0x1 + c * 0x20 + 0x1 + d * 0x20 + 0x01 + e * 0x40 + 0x1 + f * 0x20 + f * 0x34 + 0x4 | g | newEncryptedLogs
* | tx0Start + 0x1 + b * 0x20 + 0x1 + c * 0x20 + 0x1 + d * 0x20 + 0x01 + e * 0x40 + 0x1 + f * 0x20 + f * 0x34 + 0x4 + g | 0x04 | byteLen(newUnencryptedLogs) (denoted h)
* | tx0Start + 0x1 + b * 0x20 + 0x1 + c * 0x20 + 0x1 + d * 0x20 + 0x01 + e * 0x40 + 0x1 + f * 0x20 + f * 0x34 + 0x4 + g + 0x4 | h | newUnencryptedLogs
* | | | },
* | | | TxEffect 1 {
* | | | ...
* | | | },
* | | | ...
* | | | TxEffect (t - 1) {
* | | | ...
* | | | },
*/
library MessagesDecoder {
/**
Expand All @@ -62,37 +70,77 @@ library MessagesDecoder {
l1ToL2Msgs = new bytes32[](Constants.NUMBER_OF_L1_L2_MESSAGES_PER_ROLLUP);

uint256 offset = 0;

// Note hashes
// L1 to L2 messages
uint256 count = read4(_body, offset);
offset += 0x4 + count * 0x20;

// Nullifiers
count = read4(_body, offset);
offset += 0x4 + count * 0x20;
offset += 0x4;

// Public data writes
count = read4(_body, offset);
offset += 0x4 + count * 0x40;

// L2 to L1 messages
count = read4(_body, offset);
l2ToL1Msgs = new bytes32[](count);
// `l1ToL2Msgs` is fixed size so if `lengths.l1Tol2MsgsCount` < `Constants.NUMBER_OF_L1_L2_MESSAGES_PER_ROLLUP` the array
// will contain some zero values.
assembly {
calldatacopy(add(l2ToL1Msgs, 0x20), add(_body.offset, add(offset, 0x4)), mul(count, 0x20))
calldatacopy(add(l1ToL2Msgs, 0x20), add(_body.offset, offset), mul(count, 0x20))
}
offset += 0x4 + count * 0x20;

// Contracts
count = read4(_body, offset);
offset += 0x4 + count * 0x54;
offset += count * 0x20;

// L1 to L2 messages
count = read4(_body, offset);
// `l1ToL2Msgs` is fixed size so if `lengths.l1Tol2MsgsCount` < `Constants.NUMBER_OF_L1_L2_MESSAGES_PER_ROLLUP` the array
// will contain some zero values.
assembly {
calldatacopy(add(l1ToL2Msgs, 0x20), add(_body.offset, add(offset, 0x04)), mul(count, 0x20))
uint256 numTxs = read4(_body, offset);
offset += 0x4;

l2ToL1Msgs = new bytes32[](numTxs * Constants.MAX_NEW_L2_TO_L1_MSGS_PER_TX);

// Now we iterate over the tx effects
for (uint256 i = 0; i < numTxs; i++) {
// Note hashes
count = read1(_body, offset);
offset += 0x1;
offset += count * 0x20; // each note hash is 0x20 bytes long

// Nullifiers
count = read1(_body, offset);
offset += 0x1;
offset += count * 0x20; // each nullifier is 0x20 bytes long

// L2 to L1 messages
{
count = read1(_body, offset);
offset += 0x1;

uint256 msgsLength = count * 0x20; // each l2 to l1 message is 0x20 bytes long

// Now we copy the new messages into the array (if there are some)
if (count > 0) {
uint256 indexInArray = i * Constants.MAX_NEW_L2_TO_L1_MSGS_PER_TX;
assembly {
calldatacopy(
add(add(l2ToL1Msgs, 0x20), mul(indexInArray, 0x20)),
add(_body.offset, offset),
msgsLength
)
}
}

offset += msgsLength;
}

// Public data writes
count = read1(_body, offset);
offset += 0x1;
offset += count * 0x40; // each public data write is 0x40 bytes long

// Contracts
count = read1(_body, offset);
offset += 0x1;
offset += count * 0x20; // each contract leaf is 0x20 bytes long

// Contract data
offset += count * 0x34; // each contract data is 0x34 bytes long

// Encrypted logs
uint256 length = read4(_body, offset);
offset += 0x4 + length;

// Unencrypted logs
length = read4(_body, offset);
offset += 0x4 + length;
}

inHash = sha256(abi.encodePacked(l1ToL2Msgs));
Expand All @@ -101,6 +149,16 @@ library MessagesDecoder {
return (inHash, outHash, l1ToL2Msgs, l2ToL1Msgs);
}

/**
* @notice Reads 1 bytes from the data
* @param _data - The data to read from
* @param _offset - The offset to read from
* @return The 1 byte as a uint256
*/
function read1(bytes calldata _data, uint256 _offset) internal pure returns (uint256) {
return uint256(uint8(bytes1(_data[_offset:_offset + 1])));
}

/**
* @notice Reads 4 bytes from the data
* @param _data - The data to read from
Expand Down
Loading

0 comments on commit 29b1ea3

Please sign in to comment.