Changes and config to support using GraalVM AOT to compile and
execute evmTool as an alternate configuration.
While not as long-term performant the startup time
makes filling reference tests with Besu reasonable.
Signed-off-by: Danno Ferrin <danno.ferrin@swirldslabs.com>
When receiving transactions from a file parse them as an RLP list.
Signed-off-by: Danno Ferrin <danno.ferrin@swirldslabs.com>
Co-authored-by: Antoine Toulme <antoine@toulme.name>
Reduce the number of places that expose Log4J classes as a part of the
interfaces for methods and classes. While Log4j remains the default we
still need to be able to function when the Log4J jars are removed from
the classpath.
Signed-off-by: Danno Ferrin <danno.ferrin@swirldslabs.com>
Support two in-the-wild models of blockhash for testing. For state-tests
always use the keccak hash of the decimal text of the block number.
For t8n always use the blockHashes table in the env. No change to production
handling of blockhash. Refactored BlockHashLookup to split caching
implementation from API.
Signed-off-by: Danno Ferrin <danno.ferrin@swirldslabs.com>
In one case reading config files used parseLong instead of decode.
decode accepts `0x` and decimal, only a single radix for parseLong.
Signed-off-by: Danno Ferrin <danno.ferrin@swirldslabs.com>
Fix a formatting error in standard trace that produced non-parsable
JSON. Add unit tests to cover those cases.
Signed-off-by: Danno Ferrin <danno.ferrin@swirldslabs.com>
Adding the t8n and b11r tool allows Hyperledger besu to produce reference
tests from the https://github.com/ethereum/execution-spec-tests repository.
* Add t8n tool, or transition tool. Calculates the state changes from a set
of transactions against a given parent context
* Add b11r tool, or block-builder tool. Given block parts it can generate a
RLP block.
* Change some tracing CLI flags to work better with retesteth
Signed-off-by: Danno Ferrin <danno.ferrin@swirldslabs.com>
Speedup initcode validation by
- optimizing the jumpdest loop
- not caching initcode in the code cache
(but keep codecache for normal contracts).
Signed-off-by: Danno Ferrin <danno.ferrin@swirldslabs.com>
- Added missing javadocs so that javadoc doclint passes against JDK 17 (invoke by Besu gradle build).
- Exclude following packages from javadoc lint:
org.hyperledger.besu.privacy.contracts.generated
org.hyperledger.besu.tests.acceptance.*
- Temporarily exclude ethereum and evm submodule for doc lint checks.
- Run the javadoc task using GitHub actions (use Java 17) to report any javadoc errors during the PR builds
- Updating plugin-api build.gradle with new hash as javadoc comments caused it to change
Signed-off-by: Usman Saleem <usman@usmans.info>
* Add '--fork=' to EvmTool
Add a CLI '--fork' option to allow code to be evaluated in specific
forks.
Signed-off-by: Danno Ferrin <danno.ferrin@swirldslabs.com>
Optimize the EOF Functions implementation by not splitting code sections
into separate Bytes object. Instead the EOF evaluation occurs on a
single container space. When a function is called the PC (relative to
container start, not section start) is moved.
EOF lacks PC introspection and absolute jumps, so where "PC=0" starts is
not observable to the code. "PC=0" is still respected in the traces.
Signed-off-by: Danno Ferrin <danno.ferrin@swirldslabs.com>
* Support Shanghai in Reference Tests
Add support for Shanghai in reference tests
* Add Shanghai and Cancun milestones
* Add "withdrawals" and "withdrawalsRoot" to json objects
* Migrate to schedule by block header.
Signed-off-by: Danno Ferrin <danno.ferrin@swirldslabs.com>
* Add pending withdrawal code to test case spec
Signed-off-by: Danno Ferrin <danno.ferrin@swirldslabs.com>
Signed-off-by: Danno Ferrin <danno.ferrin@swirldslabs.com>
Co-authored-by: Sally MacFarlane <macfarla.github@gmail.com>
Fix a variety of issues found during the fuzzing sprint.
* Validate type length matches code length
* Remove JUMPF from shanghai
* check retf output size
* handle zero length types and code better
* Fix table error on RETURNDATACOPY
* RJUMPV byte should be unsigned
* RJUMPV stack validation fix and more tests
* dead code detection via counting considered bytes
* EVMTool EOF Fuzzing support- remove all alphanumerics (punctuation) and comment lines from code in CLI
Signed-off-by: Danno Ferrin <danno.ferrin@swirldslabs.com>
* Add post-run allocation output to EVMTool
Add a CLI flag --json-alloc that will output the post-execution state of
the allocations the EVM Tool executed in. As well as post-execution
state for state-tests.
Signed-off-by: Danno Ferrin <danno.ferrin@swirldslabs.com>
Implement CALLF, RETF, and JUMPF operations. This introduces the
notion of "code sections" into the code. For legacy code section zero
consists of all of the code and is the only code section. Also
implements EIP specified code validation to ensure validated EOF code
won't use CALLF or JUMPF to a non-existent section.
Signed-off-by: Danno Ferrin <danno.ferrin@swirldslabs.com>
Implement "Small" EOF - EIP-3540 (container) and EIP-3670 (validation).
Make code an interface so EOF specific features are compartmentalized,
including an 'invalid' code type representing a code block that didn't
pass validation, CodeV1 for EOF1, and CodeV0 which represents pre-EOF
code. EVMs track a maximum supported EOF version (where 0 is pre-eof)and
code is generated from a CodeFactory taking in context (is it a CREATE
operation and max code size) for the validation.
Includes spec versions for "Shanghai" and transient testnet "Shandong".
"Small" EOF is only activated in Shandong.
Signed-off-by: Danno Ferrin <danno.ferrin@swirldslabs.com>
* Replace deprecated parseWithHandlers from tests
* Set System out and err to use the streams used in the tests
* Remove defaultExeceptionHandler and new execute API will use different default
Signed-off-by: Gabriel Fukushima <gabrielfukushima@gmail.com>
Co-authored-by: Simon Dudley <simon.dudley@consensys.net>
Co-authored-by: Sally MacFarlane <macfarla.github@gmail.com>
10.4 revision of reference tests has notable changes
* Nonce can be up to 2^64-1, with some opcode and validity interactions specced in eip-2681
* Wei fields can be up to 2^256, tests check for rollover
* VM Tests were removed
* Legacy Tests were removed
* Reference tests make it clear that transactions with a maxPriorityFee of zero are valid. Perhaps useless, but still valid.
* Adding validation hooks in the out-of-chain test execution. These validations are caught in full-chain execution, just not in conveniently places for integration testing.
* This does not transaction tests support for the "merge" fork, as the release notes marked it as an "upcoming" feature.
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
Signed-off-by: Daniel Lehrner <daniel.lehrner@consensys.net>
Co-authored-by: Diego López León <dieguitoll@gmail.com>
Co-authored-by: Daniel Lehrner <daniel.lehrner@consensys.net>
Remove the Gas object from the EVM and replace it with the primitive
long. This will have positive impact on short lived object garbage
collection stats, which at very high load causes significant performance
issues.
This also codifies EIP-4803 in the Besu EVM, limiting gas to a signed
64-bit long internally.
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
Change the unit test execution to use the Junit5 JUnitPlatform. This
allows for a mix of junit 4 and junit 5 tests and for a gradual
migration to junit 5 instead of a big bang. One class depended on
junit 4 exceptions and was updated. Two tests depending on
native libraries fail gracefully on mac (and only mac).
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
Migrate the jump dest analysis cache into a more generic code cache that
lives inside the EVM.
This is in preparation for Ethereum Object Formats where a string of
code may be treated differently depending on what EVM version is
executing the code. Newer versions will also have difference analyses to
run that will need different backing data structures.
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
* Bump SLF4J version
Signed-off-by: Diego López León <dieguitoll@gmail.com>
* Replace log4j2 API with SLF4j API
Signed-off-by: Diego López León <dieguitoll@gmail.com>
* Replace usage of LogManager#getFormatterLogger
This is for keeping compatibility with SLF4J. If neccesary, a specific formatter can be created for the RlpBlockImporter class
Signed-off-by: Diego López León <dieguitoll@gmail.com>
* Unset the default logging value for the retesteth
This is because it's not possible to resolve the root logger level into a Log4J2 field
Signed-off-by: Diego López León <dieguitoll@gmail.com>
* Prevent creation of Logger context outside SLF4J
org.hyperledger.besu.cli.BesuCommand#setAllLevels was taken from
https://github.com/apache/logging-log4j2/blob/rel%2F2.17.1/log4j-core/src/main/java/org/apache/logging/log4j/core/config/Configurator.java#L309
Signed-off-by: Diego López León <dieguitoll@gmail.com>
* Add FATAL level deprecation message
Signed-off-by: Diego López León <dieguitoll@gmail.com>
* [Sonar] Fix java:S2139
Exceptions should be either logged or rethrown but not both
Signed-off-by: Diego López León <dieguitoll@gmail.com>
* [Sonar] Fix java:S3457
Printf-style format strings should be used correctly
Signed-off-by: Diego López León <dieguitoll@gmail.com>
* Add changelog
Signed-off-by: Diego López León <dieguitoll@gmail.com>
Change the docker base images from debian (implied by openjdk image) to
ubuntu explicitly. Ubuntu has quicker reaction time to base image
security vulnerabilities.
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
* Rename BftForksSchedule to ForkSchedule
* Make ForksSchedule generic beyond BftConfigOptions to allow it to work for MiningCoordinator
* Extract BftForksScheduleFactory from ForksSchedule
* Move ForksSchedule to common package
Signed-off-by: Simon Dudley <simon.dudley@consensys.net>
Broad reaching optimizations to speed up EVM calculations
* Generally speaking, use int and long where it is more appropriate than UInt256 (memory indexes mostly)
* Move the internal stack to Bytes from UInt256
* Re-work the flow of many operations to account for the above
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
Move EVM to a standalone module
Move the EVM classes to a standalone module. This is mostly moves but
some API re-resign to peel out some features not essential to the EVM,
such as privacy support and ties to the data storage subsystem.
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
Rollback TransactionGasCalculator
Rollback the transactionGasCalculator changes. This was done in
preparation for the EVM modularization. Instead, de-structure the data
needed to calculate the gas instead of passing in the whole transaction.
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
* Create new datatypes module
Create a new `datatypes` module to hold datatypes that are broadly used.
This will aid modularization by making sure the base types in the module
minimize the amount of unrelated support classes needed.
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
* Add Address, Hash, and Wei to datatypes
Move the Address, Hash, and Wei to datatypes in as they are needed for
EVM modularization.
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
In preparation for the EVM library pull out the transaction related gas
calculations and move them into their own `TransactionGasCalculator.`
This has 4 calls right now, none of which occur inside the EVM:
* Intrinsic gas cost
* Code Deposit gas cost
* Max refund quotient
* Max Privacy Marker Transaction intrinsic gas cost.
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
EIP-1702 account versioning was a candidate for the Istanbul hard fork
but was removed prior to the first testnet. Other versioning techniques
have greater core dev mind share and one preparatory step landed in
London, EIP-3541, making EIP-1702 very unlikely to make it to Mainnet.
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
Danno Ferrin
Fabio Di Fabio
Sally MacFarlane
Usman Saleem
Jason Frame
Danno Ferrin
Gabriel Camargo Fukushima
Diego López León
Simon Dudley
Jason Frame
Justin Florentine
Sally MacFarlane
garyschulte
matkt
Stefan Pingel