Performance measurements have indicated that the list of Bytes32 for
memory is not performant enough. Moving to a byte array reduces wrapper
object thrashing as well as simplifying cross-word boundary logic (there
are no boundaries).
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
* feat: FastSync now stops on failure instead of falling back to FullSync
* feat: exiting on FastSync failed
* feat: FastSync continues on FullSync when no errors happened
Signed-off-by: Alexandre PARIS-VERGNE <alexpv14@gmail.com>
Improved JUMP and JUMPSUB destination check. Avoid having to browse the code if the destination == code size
Signed-off-by: Karim TAAM <karim.t2am@gmail.com>
DevP2P packets expiration fields should be in seconds past the epoch,
not milliseconds past the epoch.
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
We cleared the mark storage upon starting the pruner in order to prevent the first sweep from being smaller than it needs to be.
```
// Optimization for the case where the previous cycle was interrupted (like the node was shut
// down). If the previous cycle was interrupted, there will be marks in the mark storage from
// last time, causing the first sweep to be smaller than it needs to be.
clearMarks();
```
Unfortunately, it's currently called in prepare, which is synchronous with the main thread. This means that we can delay startup by on the order of 10 minutes if the storage was especially full.
This commit makes the pruner `start` asynchronous.
This was reverted once before due to test intermittency and this commit is the second attempt :P
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
Rollback a Eth64 change intended to improve compatibility in consortium network. It resulted in many nodes not connecting on named networks such as goerli and mordor.
* Revert "Handle backward compatibility for EIP-2124 fork id management (#981)"
This reverts commit 1d8218ae
* Revert "[EIP-2124] Backward compatible fork id management implementation (#979)"
This reverts commit cb2c67d7
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
* Use seconds for timeout configuration
* Add units to variable names
* Change timeout to 5 minutes
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
We cleared the mark storage upon starting the pruner in order to prevent the first sweep from being smaller than it needs to be.
// Optimization for the case where the previous cycle was interrupted (like the node was shut
// down). If the previous cycle was interrupted, there will be marks in the mark storage from
// last time, causing the first sweep to be smaller than it needs to be.
clearMarks();
Unfortunately, it's currently called in prepare, which is synchronous with the main thread. This means that we can delay startup by on the order of 10 minutes if the storage was especially full.
This commit makes the pruner `start` asynchronous.
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
Slight rework on some PendingTransaction synchronization
* sync on `prioritizedTransactions` instead of `pendingTransactions`
* `pendingTransactions` is already a `ConcurrentHashMap`, which
syncs on itself internally
* not all accesses to `pendingTransactions` were synced
* `prioritizedTransactions` is not exposed directly or indirectly,
making it a better lock
* Improve synchronization in TransactionsForSenderInfo
* `gaps` was returned without requiring synchronization
* `transacitonInfos` was not thread safe.
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
* Add timeout handler in JsonRpcHttpService.
- Created `TimeoutOptions` class with fields:
- *timeout*: the timeout in milliseconds.
- *errorCode*: the http status code.
- Created `EthRpcTimeoutHandler`.
- Contains `handler` static method to create `Vertx Handler`. The process of the handler created is the following
- retrieve the body as string.
- parse it as a `JsonObject`.
- cache the resulting `JsonObject` in the `RoutingContext`.
- retrieve the RPC method name.
- set a timer based on the configuration if the method name matches one entry in the configuration map.
- call the next handler.
Signed-off-by: Abdelhamid Bakhta <abdelhamid.bakhta@consensys.net>
Don't respond to neighbors and ping packets that have an expiration
prior to the system's current time.
Addresses two tests in #975
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
Update implementation of the EIP-2315
- Execution of BEGINSUB causes exception and terminates execution.
- JUMPSUB sets the pc to location + 1 .
- In the edge case when BEGINSUB is the last instruction in code and this subroutine is jumped-to, the implementations should execute STOP .
Signed-off-by: Karim TAAM <karim.t2am@gmail.com>
* Handle zeros fork blocks when computing fork id.
- Updated `ForkIdManager`:
- Created `ForkIDChecker` interface with `boolean check(ForkId forkId)` method.
- Added a check in the constructor to use the proper `ForkIDChecker` based on the following conditions:
- Accept all peers if the list of forks contains only `zeros`.
- Use `EIP-2124` rules otherwise.
Signed-off-by: Abdelhamid Bakhta <abdelhamid.bakhta@consensys.net>
* Handle backward compatibility
Signed-off-by: Abdelhamid Bakhta <abdelhamid.bakhta@consensys.net>
It was possible to run Besu on an incomplete fast-synced db with full sync enabled. This causes the node to stall out. This PR resolves this issue.
Signed-off-by: Karim TAAM <karim.t2am@gmail.com>
Because of where we provide info log messages in the
BlockPropagationManager we don't log out of order blocks. Moving to
task cleanup allows us to get the timing information too instead of
simply as part of the executeTask method
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
* Add checks on the replacement of a transaction in the pool:
- reject EIP-1559 for pre-fork blocks
- accept both frontier and EIP-1559 transactions during phase 1
- reject frontier transactions after phase 2 is finalized
Signed-off-by: Abdelhamid Bakhta <abdelhamid.bakhta@consensys.net>
* The transaction gas price is computed when adding a transaction into the local pool using eth_sendRawTransaction JSON RPC endpoint. Transaction price must be computed properly depending on the type of the transaction.
For instance `shouldReplace` method of PendingTransactions must be updated to deal with EIP-1559 transactions.
- Updated `PendingTransactions` to add access to the chain header in order to retrieve the last base fee value.
- Updated `TransactionReplacementByPriceRule` to compute the transaction price depending on the type of the transaction (frontier or eip-1559).
- Added unit tests to cover all possible replacement scenarios.
Signed-off-by: Abdelhamid Bakhta <abdelhamid.bakhta@consensys.net>
* Support clique genesis (needs a nodekey)
* Support nested calls
* Write genesis allocations to world state
* Report total gas (EVM operations + intrinsic cost of TX)
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
Retesteth requires even length quantities (in contravariance to the Eth
JSON-RPC standards). For storage range at provide even length
quantities.
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
Fix cases where the gas estimate is too low and causes the transaction to fail
Implementation of a strategy that will estimate the number of gas needed by a transaction
1 - Estimate the number of gas used by a transaction
2 - Estimate the number of gas necessary to add for each sub call present in a transaction (65/64^depth)
3 - Add the minimum gas required for operations that need it (SSTORE)
Signed-off-by: Karim TAAM <karim.t2am@gmail.com>
* Update transaction pool error handling.
Transactions added locally into the pool via `eth_sendRawTransaction` can now generate 2 new JSON RPC errors:
- `ETH_SEND_TX_ALREADY_KNOWN`: occurs when adding a transaction already present in the pool.
- code: -32000
- message: `Known Transaction`
- `ETH_SEND_TX_REPLACEMENT_UNDERPRICED`: occurs when adding a transaction already present in the pool.
- code: -32000
- message: `Replacement transaction underpriced`
### Changes summary
- Created `TransactionAddedStatus` enum.
- `ALREADY_KNOWN`
- `REJECTED_UNDERPRICED_REPLACEMENT`
- `ADDED`
- Created 2 new errors in `JsonRpcError`.
- Updated JsonRpcErrorConverter to handle newly created errors.
- Updated `addLocalTransaction` method of `PendingTransactions` to return `TransactionAddedStatus` instead of boolean.
- Updated unit tests accordingly.
Signed-off-by: Abdelhamid Bakhta <abdelhamid.bakhta@consensys.net>
prior to Eth/65 we did not participate in transaction gossiping until we
were close to in sync. With this change we won't request pooled
transaction hashes until we are similarly close to sync.
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
Increase Netty threadpool size for handshake handling
It was found that all ECIES handhshaking is taking place in
an NioEventLoop of being serviced by one worker.
This in turn meant that, during periods of high demand,
many potential peers were timing out (as the latter peers
had to wait for earlier peers to complete prior to
being serviced).
This change grows the threadpool to 10 (arbitrarily selected),
which minimises the number of peers waiting for connectivity.
Signed-off-by: Trent Mohay <trent.mohay@consensys.net>
Co-authored-by: Jason Frame <jasonwframe@gmail.com>
* add doomed key check (busy-waiting for now)
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* optional and logging
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* remove logging
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* sleeping and hardening
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* rename segments
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* move away from atomic references to regular vars
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* remove hardened segment parameter
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* increase sleep
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* spotless
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* remove unnecessary interface
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* rename
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* move commit waiting outside of timer
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* set default lock timeout to 1ms
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* add default lock timeout to tests
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* Revert "rename segments"
This reverts commit 184eefaaa0ccc857b0caff2b382f8338ff225d5d.
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* fix jmh compilation error
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* add documentation
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* bump up sleep to 1ms
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* (POC) Add lock to ensure that we don't prune while comitting
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* remove unnecessary persist (#569)
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* flesh out @mbaxter's idea and remove my code
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* iterator changes
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* hybridize with doomed key
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* comment
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* move doomed key unset to after node added listener
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* update instead of getting and setting doomedKeyRef in commit
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* comment
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* invert condition
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* remove locks
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* remove `removeAllKeysUnless`
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* more remove removeAllKeysUnless
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* reuse streamKeys
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* remove test
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* set default lock timeout to 1ms
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* add default lock timeout to tests
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* fix jmh compilation error
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* revert back to locks instead of doomedkey
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* change delete to not guarantee deletion
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* plugin hash
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* javadoc
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* Revert "change delete to not guarantee deletion"
This reverts commit 2289bb34cf.
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* skip key deletion on timeout
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* clear in rollback
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* Revert "fix jmh compilation error"
This reverts commit b64ecf8656.
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* Revert "add default lock timeout to tests"
This reverts commit aff6aa6065.
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* Revert "set default lock timeout to 1ms"
This reverts commit 267fe0a642.
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* use noSlowDown write option instead of global timeout
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* add back tests
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* close tryDeleteOptions
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* remove unnecessary lock
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* move increment inside try
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* use StorageException subclass instead of field
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* revert accidental deletion in javadoc
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* tryDelete javadoc
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* add trace for skipping key deletion
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* merge catch and finally try blocks
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* switch from exception to boolean return value
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* tweak
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* changelog changes
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* add api back
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
* add back throws javadoc
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
Co-authored-by: Meredith Baxter <meredith.baxter@consensys.net>
Co-authored-by: MadelineMurray <43356962+MadelineMurray@users.noreply.github.com>
* When in retesteth return the coinbase in the eth_getBlockByHash, as
required by the retesteth tool.
* use genesis extraData when mining
* storage keys should be 0x00 instead of 0x when zero.
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
* Compute transaction gas budget allocation according to EIP-1559 rules.
- Added `TransactionGasBudgetCalculator` with `hasBudget` method.
- Updated `AbstractBlockProcessor` to compute the gas budget using `TransactionGasBudgetCalculator`.
- Added unit tests in `TransactionGasBudgetCalculatorTest`.
Signed-off-by: Abdelhamid Bakhta <abdelhamid.bakhta@consensys.net>
* Fixed unit test.
Signed-off-by: Abdelhamid Bakhta <abdelhamid.bakhta@consensys.net>
IbftRound has been updated to accept Signing errors (eg no signature supplier available) and
continue operating if possible.
This also catches failures in signing and ECDH Key agreement
creation during discovery and handshaking.
Signed-off-by: Trent Mohay <trent.mohay@consensys.net>
Danno Ferrin
br0tchain
Ratan Rai Sur
Karim T
Abdelhamid Bakhta
Adrian Sutton
mark-terry
Danno Ferrin
Lucas Saldanha
Trent Mohay