- cli / node
- Update handler of SIGINT so that after 3 SIGINTs, the process terminates
unconditionally
- coordinator
- Store stats without pointer
- In all functions that send a variable via channel, check for context done
to avoid deadlock (due to no process reading from the channel, which has
no queue) when the node is stopped.
- Abstract `canForge` so that it can be used outside of the `Coordinator`
- In `canForge` check the blockNumber in current and next slot.
- Update tests due to smart contract changes in slot handling, and minimum
bid defaults
- TxManager
- Add consts, vars and stats to allow evaluating `canForge`
- Add `canForge` method (not used yet)
- Store batch and nonces status (last success and last pending)
- Track nonces internally instead of relying on the ethereum node (this
is required to work with ganache when there are pending txs)
- Handle the (common) case of the receipt not being found after the tx
is sent.
- Don't start the main loop until we get an initial messae fo the stats
and vars (so that in the loop the stats and vars are set to
synchronizer values)
- When a tx fails, check and discard all the failed transactions before
sending the message to stop the pipeline. This will avoid sending
consecutive messages of stop the pipeline when multiple txs are
detected to be failed consecutively. Also, future txs of the same
pipeline after a discarded txs are discarded, and their nonces reused.
- Robust handling of nonces:
- If geth returns nonce is too low, increase it
- If geth returns nonce too hight, decrease it
- If geth returns underpriced, increase gas price
- If geth returns replace underpriced, increase gas price
- Add support for resending transactions after a timeout
- Store `BatchInfos` in a queue
- Pipeline
- When an error is found, stop forging batches and send a message to the
coordinator to stop the pipeline with information of the failed batch
number so that in a restart, non-failed batches are not repated.
- When doing a reset of the stateDB, if possible reset from the local
checkpoint instead of resetting from the synchronizer. This allows
resetting from a batch that is valid but not yet sent / synced.
- Every time a pipeline is started, assign it a number from a counter. This
allows the TxManager to ignore batches from stopped pipelines, via a
message sent by the coordinator.
- Avoid forging when we haven't reached the rollup genesis block number.
- Add config parameter `StartSlotBlocksDelay`: StartSlotBlocksDelay is the
number of blocks of delay to wait before starting the pipeline when we
reach a slot in which we can forge.
- When detecting a reorg, only reset the pipeline if the batch from which
the pipeline started changed and wasn't sent by us.
- Add config parameter `ScheduleBatchBlocksAheadCheck`:
ScheduleBatchBlocksAheadCheck is the number of blocks ahead in which the
forger address is checked to be allowed to forge (apart from checking the
next block), used to decide when to stop scheduling new batches (by
stopping the pipeline). For example, if we are at block 10 and
ScheduleBatchBlocksAheadCheck is 5, eventhough at block 11 we canForge,
the pipeline will be stopped if we can't forge at block 15. This value
should be the expected number of blocks it takes between scheduling a
batch and having it mined.
- Add config parameter `SendBatchBlocksMarginCheck`:
SendBatchBlocksMarginCheck is the number of margin blocks ahead in which
the coordinator is also checked to be allowed to forge, apart from the
next block; used to decide when to stop sending batches to the smart
contract. For example, if we are at block 10 and
SendBatchBlocksMarginCheck is 5, eventhough at block 11 we canForge, the
batch will be discarded if we can't forge at block 15.
- Add config parameter `TxResendTimeout`: TxResendTimeout is the timeout
after which a non-mined ethereum transaction will be resent (reusing the
nonce) with a newly calculated gas price
- Add config parameter `MaxGasPrice`: MaxGasPrice is the maximum gas price
allowed for ethereum transactions
- Add config parameter `NoReuseNonce`: NoReuseNonce disables reusing nonces
of pending transactions for new replacement transactions. This is useful
for testing with Ganache.
- Extend BatchInfo with more useful information for debugging
- eth / ethereum client
- Add necessary methods to create the auth object for transactions manually
so that we can set the nonce, gas price, gas limit, etc manually
- Update `RollupForgeBatch` to take an auth object as input (so that the
coordinator can set parameters manually)
- synchronizer
- In stats, add `NextSlot`
- In stats, store full last batch instead of just last batch number
- Instead of calculating a nextSlot from scratch every time, update the
current struct (only updating the forger info if we are Synced)
- Afer every processed batch, check that the calculated StateDB MTRoot
matches the StateRoot found in the forgeBatch event.
- KVDB/StateDB
- Pass config parameters in a Config type instead of using many
arguments in constructor.
- Add new parameter `NoLast` which disables having an opened DB with a
checkpoint to the last batchNum for thread-safe reads. Last will be
disabled in the StateDB used by the TxSelector and BatchBuilder.
- Add new parameter `NoGapsCheck` which skips checking gaps in the list
of checkpoints and returning errors if there are gaps. Gaps check
will be disabled in the StateDB used by the TxSelector and
BatchBuilder, because we expect to have gaps when there are multiple
coordinators forging (slots not forged by our coordinator will leave
gaps).
- kvdb
- Fix path in Last when doing `setNew`
- Only close if db != nil, and after closing, always set db to nil
- This will avoid a panic in the case where the db is closed but
there's an error soon after, and a future call tries to close
again. This is because pebble.Close() will panic if the db is
already closed.
- Avoid calling pebble methods when a the Storage interface already
implements that method (like Close).
- statedb
- In test, avoid calling KVDB method if the same method is available for
the StateDB (like MakeCheckpoint, CurrentBatch).
- eth
- In *EventByBlock methods, take blockHash as input argument and use it
when querying the event logs. Previously the blockHash was only taken
from the logs results *only if* there was any log. This caused the
following issue: if there was no logs, it was not possible to know if
the result was from the expected block or an uncle block! By querying
logs by blockHash we make sure that even if there are no logs, they
are from the right block.
- Note that now the function can either be called with a
blockNum or blockHash, but not both at the same time.
- sync
- If there's an error during call to Sync call resetState, which
internally resets the stateDB to avoid stale checkpoints (and a
corresponding invalid increase in the StateDB batchNum).
- During a Sync, after very batch processed, make sure that the StateDB
currentBatch corresponds to the batchNum in the smart contract
log/event.
Last db view is an opened pebble db which always contains a checkpoint from the
last batch. Methods to access this last batch are thread safe so that views of
the last checkpoint can be made anywhere and with a consistent view of the
state.
- cli / node
- Update handler of SIGINT so that after 3 SIGINTs, the process terminates
unconditionally
- coordinator
- Store stats without pointer
- In all functions that send a variable via channel, check for context done
to avoid deadlock (due to no process reading from the channel, which has
no queue) when the node is stopped.
- Abstract `canForge` so that it can be used outside of the `Coordinator`
- In `canForge` check the blockNumber in current and next slot.
- Update tests due to smart contract changes in slot handling, and minimum
bid defaults
- TxManager
- Add consts, vars and stats to allow evaluating `canForge`
- Add `canForge` method (not used yet)
- Store batch and nonces status (last success and last pending)
- Track nonces internally instead of relying on the ethereum node (this
is required to work with ganache when there are pending txs)
- Handle the (common) case of the receipt not being found after the tx
is sent.
- Don't start the main loop until we get an initial messae fo the stats
and vars (so that in the loop the stats and vars are set to
synchronizer values)
- eth / ethereum client
- Add necessary methods to create the auth object for transactions manually
so that we can set the nonce, gas price, gas limit, etc manually
- Update `RollupForgeBatch` to take an auth object as input (so that the
coordinator can set parameters manually)
- synchronizer
- In stats, add `NextSlot`
- Upgrade go-merkletree version to include the last changes of Pebble
that fixes the cgo issues (which should fix#453), from:
c2b05f12d7
- TxSelector
- Remove parameter batchNum for GetL2TxSelection & GetL1L2TxSelection
- Add checks of ToBJJ & ToEthAddr when ToIdx>255
- Avoid getting the sender account twice to get the TokenID of a l2tx
- Add test to check that selected L2Txs are sorted by Nonce
- Discard L2Tx that return error at ProcessL2Txs
- executed `go mod tidy`
- StateDB
- Update GetIdxByEthAddrBJJ to return ErrToIdxNotFound when idx not found, so can be checked at upper levels
- TxSelector
- rm CoordIdxsDB that is no longer needed (also related methods)
- add `getCoordIdx` method to get the Coordinator Idx for a given TokenID
- Update coordinator account creation related to new TokenIDs from L2Txs
- Reorganize GetL1L2TxSelection
- return CoordIdxs used in the selection
- Update go-merkletree version which avoids marshaling Siblings to json
with 'null' value in case of empty array
Load an ethereum keystore when the node is started in coordinator mode. The
private key corresponding to the forger address must be imported into the
keystore before running the node in coordinator mode. You can see an examples
in `cli/node/load-sk-example.sh`.
On the Circuits uses the name `nTx` (and on the tests `maxTx`), but
Contracts `maxTx`. On Go we had both variables, but really only used
one, now at Go we use `MaxTx` to refer to the maximum number of
transactions.
Introduce a constructor parameter for the StateDB called `keep`, which tells
how many checkpoints to keep. When doing a new checkpoint, if the number of
existing checkpoints exeeds `keep`, the oldest ones will be deleted.
- Implement Pipeline.prepareForgeBatchArgs()
- Add a minimal stress test for the coordinator (that also runs the
synchronizer)
- Update txprocessor.ProcessTxs() to return valid results for batches without
transactions
- Add the boilerplate for the corresponding test, leaving as TODO the
zkInput values
- Update prover client to use the same point format as proof server (projective)
- Update interface of TxSelector.GetCoordIdxs to also return the authorizations
to create accounts that go with the l1CoordinatorTxs.
Update usage of `*babyjub.PublicKey` to `babyjub.PublicKeyComp`
- when the key is not defined, internally is used `babyjub.EmptyBJJComp`, which is a `[32]byte` of zeroes of type `babyjub.PublicKeyComp`
- the API continues returning `nil` when the key is not defined
- Update IntermediateState FinalAccFee values to fit in the Circom circuit expected inputs
- Add ZKInputs generation test with transactions generated by Til
(TestZKInputs6)
- BatchBuilder: remove redundant MakeCheckpoint call
- Update tx processor of Exit & ForceExit types
- Update ZKInputs generation related to Exit transactions
- Add TestZKInputs5 which adds tests generation of ZKInputs compatible
with Circom circuits including Exit (L2) & ForceExit (L1) transaction types
- Add StateDB test to check max DepositAmount case
- StateDB fix process txs of type CreateAccountDepositTransfer & DepositTransfer
- Add TestZKInputs3 & TestZKInputs4 which:
- have a compatible check in js version to check Circom circuits
compatibility
- adds test of tx type of CreateAccountDepositTransfer &
DepositTransfer
- Common
- Add `IdxNonce` type used to track nonces in accounts to invalidate
l2txs in the pool
- Config
- Update coordinator config will all the new configuration parameters
used in the coordinator
- Coordinator
- Introduce the `Purger` to track how often to purge and do the job when
needed according to a configuration.
- Implement the methods to invalidate l2txs transactions due to l2txs
selection in batches. For now these functions are not used in favour
of the `Purger` methods, which check ALL the l2txs in the pool.
- Call Invalidation and Purging methods of the purger both when the
node is forging (in the pipeline when starting a new batch) and when
the node is not forging (in coordinator when being notified about a
new synced block)
- L2DB:
- Implement `GetPendingUniqueFromIdxs` to get all the unique idxs from
pending transactions (used to get their nonces and then invalidate
txs)
- Redo `CheckNonces` with a single SQL query and using `common.IdxNonce`
instead of `common.Account`
- StateDB:
- Expose GetIdx to check errors when invalidating pool txs
- Synchronizer:
- Test forged L1UserTxs processed by TxProcessor
- Improve checks of Effective values
- TxSelector:
- Expose the internal LocalStateDB in order to check account nonces in
the coordinator when not forging.
- Separate ZKInputs tests at StateDB package
- Small fix at ZKInputs generation FeeIdxs length
- Fees related parameters at ZKInputs works properly
- Add ZKInputs generation test with fees (`TestZKInputs1`), which output
has been tested with circom to ensure compatibility
- Add SignatureConstant to L1Tx.TxCompressedData
- ZKInputs generation works initial version working and compatible with circom Hermez circuit
- Lots of small updates for circuits compatibility
- Update version (with new changes) of `go-iden3-crypto` & `go-merkletree`
- Common
- Move ErrTODO and ErrDone to common for usage where needed.
- Coordinator
- Move prover types to prover package
- Handle reorgs, stopping the pipeline when necessary
- Handle ethereum transaction errors by stopping the pipeline
- In case of ethereum transaction revert, check for known revert causes
(more revert causes can be added to handle more cases)
- Fix skipped transactions in TxManager confirmation logic
- Cancel and wait for provers to be ready
- Connect L2DB to:
- purge l2txs due to timeout
- mark l2txs at the different states
- Connect HistoryDB to query L1UserTxs to forge in an L1Batch
- L2DB
- Skip update functions when the input slices have no values (to avoid a
query with no values that results in an SQL error)
- StateDB
- In LocalStateDB, fix Reset when mt == nil
- Prover (new package)
- Rename the interface to Prover
- Rename the mock struct to Mock
- Extend Prover interface methods to provide everything required by the
coordinator
- Begin implementing required http client code to interact with server
proof (not tested)
- Synchronizer:
- Add LastForgeL1TxsNum to Stats
- Test/Client
- Update Auction logic to track slots in which there's no forge during
the time before the deadline (following the solidity implementation)
- Add InvalidData flag to L1Tx
- Add BytesDataAvailability to L1Tx
- Update ZKInputs & HashGlobalInputs to last spec of the protocol
(massive migrations)
- TxProcessor check correctness of L1Txs
Compatible with hermeznetwork/commonjs v0.0.4
(c345239bba)
Add HashGlobalInputs for ZKInputs compatible with js & circom circuits version.
Compatible with hermeznetwork/commonjs at version: c6a1448db5
(c6a1448db5)
- Allow starting the DebugAPI from the node via config
- In StateDB:
- Make checkpoints when ProcessTxs() succeeds
- Remove extra hardcoded `statedb` path that was redundant
- Replace hardcoded `[:4]` by `[:]` when parsing idx, which failed because
idx is 6 bytes length now.
- Extra: In node, use waitgroup instead of `stoppedXXX` channels to wait for
syncrhonizer goroutines to finish.
- Add ZKInputs parser to JSON
- Update ProcessTxs to last go-merkletree version changes
- Add StateDB test to generate ZKInputs to check inside a circom circuit
- Small fix at Til CoordinatorTx.LoadAmount=0
- Move smart contract constants and structs for variables to
common/{ethrollup.go, ethauction.go, ethwdelayer.go}:
- This removes repeated code of the structs for variables
- Allows reusing the constants and variables from all modules without
import cycles
- Remove unused common/scvars.go
- In common.BlockData, split data from each smart contract into a sepparate
field (Rollup, Auction, WDelayer). This affects the structures that til uses
as output, and HistoryDB in the AddBlockSCData.
- In Synchronizer:
- Pass starting block of each smart contract as config, instead of
incorrectly using the genesis block found in the acution constant (which
has a very different meaning)
- Use variable structs from common instead of an internal copy
- Synchronize more stuff (resolve some TODOs)
- Fix some issues found after initial testing with ganache
- In eth:
- In auction.go: Add method to get constants
- Update README to use ganache instead of buidlerevm as local blockchain
for testing
- Update env variables and test vectors to pass the tests with the
deployment in the ganache testnet.
- Use ethereum keys derived from paths (hdwallet) in testing to avoid
hardcoding private keys and generate the same keys from a mnemonic used
in the ganache tesnet.
- Update StateDB ProcessTxs fees collection
- Update ProcessTxs tests to last StateDB & Til changes
- Til Blockchain L2Txs remove Nonce generation, update tests accordingly
- Update Til Set to test fees
Eduard S
arnaucube
Arnau B