nimbus-eth1/nimbus/transaction/call_common.nim

321 lines
12 KiB
Nim
Raw Normal View History

# Nimbus - Common entry point to the EVM from all different callers
#
# Copyright (c) 2018-2024 Status Research & Development GmbH
# Licensed under either of
# * Apache License, version 2.0, ([LICENSE-APACHE](LICENSE-APACHE) or http://www.apache.org/licenses/LICENSE-2.0)
# * MIT license ([LICENSE-MIT](LICENSE-MIT) or http://opensource.org/licenses/MIT)
# at your option. This file may not be copied, modified, or distributed except according to those terms.
{.push raises: [].}
import
eth/common/eth_types, stint, stew/ptrops,
Added basic async capabilities for vm2. (#1260) * Added basic async capabilities for vm2. This is a whole new Git branch, not the same one as last time (https://github.com/status-im/nimbus-eth1/pull/1250) - there wasn't much worth salvaging. Main differences: I didn't do the "each opcode has to specify an async handler" junk that I put in last time. Instead, in oph_memory.nim you can see sloadOp calling asyncChainTo and passing in an async operation. That async operation is then run by the execCallOrCreate (or asyncExecCallOrCreate) code in interpreter_dispatch.nim. In the test code, the (previously existing) macro called "assembler" now allows you to add a section called "initialStorage", specifying fake data to be used by the EVM computation run by that test. (In the long run we'll obviously want to write tests that for-real use the JSON-RPC API to asynchronously fetch data; for now, this was just an expedient way to write a basic unit test that exercises the async-EVM code pathway.) There's also a new macro called "concurrentAssemblers" that allows you to write a test that runs multiple assemblers concurrently (and then waits for them all to finish). There's one example test using this, in test_op_memory_lazy.nim, though you can't actually see it doing so unless you uncomment some echo statements in async_operations.nim (in which case you can see the two concurrently running EVM computations each printing out what they're doing, and you'll see that they interleave). A question: is it possible to make EVMC work asynchronously? (For now, this code compiles and "make test" passes even if ENABLE_EVMC is turned on, but it doesn't actually work asynchronously, it just falls back on doing the usual synchronous EVMC thing. See FIXME-asyncAndEvmc.) * Moved the AsyncOperationFactory to the BaseVMState object. * Made the AsyncOperationFactory into a table of fn pointers. Also ditched the plain-data Vm2AsyncOperation type; it wasn't really serving much purpose. Instead, the pendingAsyncOperation field directly contains the Future. * Removed the hasStorage idea. It's not the right solution to the "how do we know whether we still need to fetch the storage value or not?" problem. I haven't implemented the right solution yet, but at least we're better off not putting in a wrong one. * Added/modified/removed some comments. (Based on feedback on the PR.) * Removed the waitFor from execCallOrCreate. There was some back-and-forth in the PR regarding whether nested waitFor calls are acceptable: https://github.com/status-im/nimbus-eth1/pull/1260#discussion_r998587449 The eventual decision was to just change the waitFor to a doAssert (since we probably won't want this extra functionality when running synchronously anyway) to make sure that the Future is already finished.
2022-11-01 15:35:46 +00:00
chronos,
results,
stew/saturation_arith,
../evm/[types, state],
../evm/[precompiles, internals],
../db/ledger,
2022-12-02 04:39:12 +00:00
../common/evmforks,
../core/eip4844,
2021-08-11 19:37:00 +00:00
./host_types
import ../evm/computation except fromEvmc, toEvmc
2021-08-11 19:37:00 +00:00
when defined(evmc_enabled):
import
../utils/utils,
./host_services
else:
import
../evm/state_transactions
type
# Standard call parameters.
CallParams* = object
vmState*: BaseVMState # Chain, database, state, block, fork.
origin*: Opt[HostAddress] # Default origin is `sender`.
gasPrice*: GasInt # Gas price for this call.
gasLimit*: GasInt # Maximum gas available for this call.
sender*: HostAddress # Sender account.
to*: HostAddress # Recipient (ignored when `isCreate`).
isCreate*: bool # True if this is a contract creation.
value*: HostValue # Value sent from sender to recipient.
input*: seq[byte] # Input data.
accessList*: AccessList # EIP-2930 (Berlin) tx access list.
versionedHashes*: VersionedHashes # EIP-4844 (Cancun) blob versioned hashes
noIntrinsic*: bool # Don't charge intrinsic gas.
noAccessList*: bool # Don't initialise EIP-2929 access list.
noGasCharge*: bool # Don't charge sender account for gas.
noRefund*: bool # Don't apply gas refund/burn rule.
sysCall*: bool # System call or ordinary call
# Standard call result. (Some fields are beyond what EVMC can return,
# and must only be used from tests because they will not always be set).
CallResult* = object
error*: string # Something if the call failed.
gasUsed*: GasInt # Gas used by the call.
contractAddress*: EthAddress # Created account (when `isCreate`).
output*: seq[byte] # Output data.
stack*: EvmStack # EVM stack on return (for test only).
memory*: EvmMemory # EVM memory on return (for test only).
func isError*(cr: CallResult): bool =
cr.error.len > 0
proc hostToComputationMessage*(msg: EvmcMessage): Message =
Message(
kind: CallKind(msg.kind.ord),
depth: msg.depth,
gas: GasInt msg.gas,
sender: msg.sender.fromEvmc,
contractAddress: msg.recipient.fromEvmc,
codeAddress: msg.code_address.fromEvmc,
value: msg.value.fromEvmc,
# When input size is zero, input data pointer may be null.
data: if msg.input_size <= 0: @[]
else: @(makeOpenArray(msg.input_data, msg.input_size.int)),
flags: msg.flags
)
func intrinsicGas*(call: CallParams, vmState: BaseVMState): GasInt {.inline.} =
# Compute the baseline gas cost for this transaction. This is the amount
# of gas needed to send this transaction (but that is not actually used
# for computation).
let fork = vmState.fork
var gas = gasFees[fork][GasTransaction]
# EIP-2 (Homestead) extra intrinsic gas for contract creations.
if call.isCreate:
gas += gasFees[fork][GasTXCreate]
if fork >= FkShanghai:
gas += (gasFees[fork][GasInitcodeWord] * call.input.len.wordCount)
# Input data cost, reduced in EIP-2028 (Istanbul).
let gasZero = gasFees[fork][GasTXDataZero]
let gasNonZero = gasFees[fork][GasTXDataNonZero]
for b in call.input:
gas += (if b == 0: gasZero else: gasNonZero)
# EIP-2930 (Berlin) intrinsic gas for transaction access list.
if fork >= FkBerlin:
for account in call.accessList:
gas += ACCESS_LIST_ADDRESS_COST
gas += GasInt(account.storageKeys.len) * ACCESS_LIST_STORAGE_KEY_COST
return gas.GasInt
proc initialAccessListEIP2929(call: CallParams) =
# EIP2929 initial access list.
let vmState = call.vmState
if vmState.fork < FkBerlin:
return
vmState.mutateStateDB:
db.accessList(call.sender)
# For contract creations the EVM will add the contract address to the
# access list itself, after calculating the new contract address.
if not call.isCreate:
db.accessList(call.to)
# EIP3651 adds coinbase to the list of addresses that should start warm.
if vmState.fork >= FkShanghai:
db.accessList(vmState.coinbase)
# Adds the correct subset of precompiles.
for c in activePrecompiles(vmState.fork):
db.accessList(c)
# EIP2930 optional access list.
for account in call.accessList:
db.accessList(account.address)
for key in account.storageKeys:
db.accessList(account.address, key.to(UInt256))
proc setupHost(call: CallParams): TransactionHost =
let vmState = call.vmState
vmState.txCtx = TxContext(
origin : call.origin.get(call.sender),
gasPrice : call.gasPrice,
versionedHashes: call.versionedHashes,
blobBaseFee : getBlobBaseFee(vmState.blockCtx.excessBlobGas),
)
var intrinsicGas: GasInt = 0
if not call.noIntrinsic:
intrinsicGas = intrinsicGas(call, vmState)
let host = TransactionHost(
vmState: vmState,
msg: EvmcMessage(
kind: if call.isCreate: EVMC_CREATE else: EVMC_CALL,
# Default: flags: {},
# Default: depth: 0,
gas: int64.saturate(call.gasLimit - intrinsicGas),
recipient: call.to.toEvmc,
code_address: call.to.toEvmc,
sender: call.sender.toEvmc,
value: call.value.toEvmc,
)
# All other defaults in `TransactionHost` are fine.
)
# Generate new contract address, prepare code, and update message `recipient`
# with the contract address. This differs from the previous Nimbus EVM API.
# Guarded under `evmc_enabled` for now so it doesn't break vm2.
when defined(evmc_enabled):
var code: CodeBytesRef
if call.isCreate:
let sender = call.sender
let contractAddress =
generateAddress(sender, call.vmState.readOnlyStateDB.getNonce(sender))
host.msg.recipient = contractAddress.toEvmc
host.msg.input_size = 0
host.msg.input_data = nil
code = CodeBytesRef.init(call.input)
else:
# TODO: Share the underlying data, but only after checking this does not
# cause problems with the database.
code = host.vmState.readOnlyStateDB.getCode(host.msg.code_address.fromEvmc)
if call.input.len > 0:
host.msg.input_size = call.input.len.csize_t
# Must copy the data so the `host.msg.input_data` pointer
# remains valid after the end of `call` lifetime.
host.input = call.input
host.msg.input_data = host.input[0].addr
let cMsg = hostToComputationMessage(host.msg)
host.computation = newComputation(vmState, call.sysCall, cMsg, code)
Added basic async capabilities for vm2. (#1260) * Added basic async capabilities for vm2. This is a whole new Git branch, not the same one as last time (https://github.com/status-im/nimbus-eth1/pull/1250) - there wasn't much worth salvaging. Main differences: I didn't do the "each opcode has to specify an async handler" junk that I put in last time. Instead, in oph_memory.nim you can see sloadOp calling asyncChainTo and passing in an async operation. That async operation is then run by the execCallOrCreate (or asyncExecCallOrCreate) code in interpreter_dispatch.nim. In the test code, the (previously existing) macro called "assembler" now allows you to add a section called "initialStorage", specifying fake data to be used by the EVM computation run by that test. (In the long run we'll obviously want to write tests that for-real use the JSON-RPC API to asynchronously fetch data; for now, this was just an expedient way to write a basic unit test that exercises the async-EVM code pathway.) There's also a new macro called "concurrentAssemblers" that allows you to write a test that runs multiple assemblers concurrently (and then waits for them all to finish). There's one example test using this, in test_op_memory_lazy.nim, though you can't actually see it doing so unless you uncomment some echo statements in async_operations.nim (in which case you can see the two concurrently running EVM computations each printing out what they're doing, and you'll see that they interleave). A question: is it possible to make EVMC work asynchronously? (For now, this code compiles and "make test" passes even if ENABLE_EVMC is turned on, but it doesn't actually work asynchronously, it just falls back on doing the usual synchronous EVMC thing. See FIXME-asyncAndEvmc.) * Moved the AsyncOperationFactory to the BaseVMState object. * Made the AsyncOperationFactory into a table of fn pointers. Also ditched the plain-data Vm2AsyncOperation type; it wasn't really serving much purpose. Instead, the pendingAsyncOperation field directly contains the Future. * Removed the hasStorage idea. It's not the right solution to the "how do we know whether we still need to fetch the storage value or not?" problem. I haven't implemented the right solution yet, but at least we're better off not putting in a wrong one. * Added/modified/removed some comments. (Based on feedback on the PR.) * Removed the waitFor from execCallOrCreate. There was some back-and-forth in the PR regarding whether nested waitFor calls are acceptable: https://github.com/status-im/nimbus-eth1/pull/1260#discussion_r998587449 The eventual decision was to just change the waitFor to a doAssert (since we probably won't want this extra functionality when running synchronously anyway) to make sure that the Future is already finished.
2022-11-01 15:35:46 +00:00
host.code = code
else:
if call.input.len > 0:
host.msg.input_size = call.input.len.csize_t
# Must copy the data so the `host.msg.input_data` pointer
# remains valid after the end of `call` lifetime.
host.input = call.input
host.msg.input_data = host.input[0].addr
let cMsg = hostToComputationMessage(host.msg)
host.computation = newComputation(vmState, call.sysCall, cMsg)
vmState.captureStart(host.computation, call.sender, call.to,
call.isCreate, call.input,
call.gasLimit, call.value)
return host
when defined(evmc_enabled):
proc doExecEvmc(host: TransactionHost, call: CallParams) =
var callResult = evmcExecComputation(host)
let c = host.computation
if callResult.status_code == EVMC_SUCCESS:
c.error = nil
elif callResult.status_code == EVMC_REVERT:
c.setError(EVMC_REVERT, false)
else:
c.setError(callResult.status_code, true)
c.gasMeter.gasRemaining = GasInt callResult.gas_left
c.msg.contractAddress = callResult.create_address.fromEvmc
c.output = if callResult.output_size <= 0: @[]
else: @(makeOpenArray(callResult.output_data,
callResult.output_size.int))
if not callResult.release.isNil:
{.gcsafe.}:
callResult.release(callResult)
Added basic async capabilities for vm2. (#1260) * Added basic async capabilities for vm2. This is a whole new Git branch, not the same one as last time (https://github.com/status-im/nimbus-eth1/pull/1250) - there wasn't much worth salvaging. Main differences: I didn't do the "each opcode has to specify an async handler" junk that I put in last time. Instead, in oph_memory.nim you can see sloadOp calling asyncChainTo and passing in an async operation. That async operation is then run by the execCallOrCreate (or asyncExecCallOrCreate) code in interpreter_dispatch.nim. In the test code, the (previously existing) macro called "assembler" now allows you to add a section called "initialStorage", specifying fake data to be used by the EVM computation run by that test. (In the long run we'll obviously want to write tests that for-real use the JSON-RPC API to asynchronously fetch data; for now, this was just an expedient way to write a basic unit test that exercises the async-EVM code pathway.) There's also a new macro called "concurrentAssemblers" that allows you to write a test that runs multiple assemblers concurrently (and then waits for them all to finish). There's one example test using this, in test_op_memory_lazy.nim, though you can't actually see it doing so unless you uncomment some echo statements in async_operations.nim (in which case you can see the two concurrently running EVM computations each printing out what they're doing, and you'll see that they interleave). A question: is it possible to make EVMC work asynchronously? (For now, this code compiles and "make test" passes even if ENABLE_EVMC is turned on, but it doesn't actually work asynchronously, it just falls back on doing the usual synchronous EVMC thing. See FIXME-asyncAndEvmc.) * Moved the AsyncOperationFactory to the BaseVMState object. * Made the AsyncOperationFactory into a table of fn pointers. Also ditched the plain-data Vm2AsyncOperation type; it wasn't really serving much purpose. Instead, the pendingAsyncOperation field directly contains the Future. * Removed the hasStorage idea. It's not the right solution to the "how do we know whether we still need to fetch the storage value or not?" problem. I haven't implemented the right solution yet, but at least we're better off not putting in a wrong one. * Added/modified/removed some comments. (Based on feedback on the PR.) * Removed the waitFor from execCallOrCreate. There was some back-and-forth in the PR regarding whether nested waitFor calls are acceptable: https://github.com/status-im/nimbus-eth1/pull/1260#discussion_r998587449 The eventual decision was to just change the waitFor to a doAssert (since we probably won't want this extra functionality when running synchronously anyway) to make sure that the Future is already finished.
2022-11-01 15:35:46 +00:00
# FIXME-awkwardFactoring: the factoring out of the pre and
# post parts feels awkward to me, but for now I'd really like
# not to have too much duplicated code between sync and async.
# --Adam
Added basic async capabilities for vm2. (#1260) * Added basic async capabilities for vm2. This is a whole new Git branch, not the same one as last time (https://github.com/status-im/nimbus-eth1/pull/1250) - there wasn't much worth salvaging. Main differences: I didn't do the "each opcode has to specify an async handler" junk that I put in last time. Instead, in oph_memory.nim you can see sloadOp calling asyncChainTo and passing in an async operation. That async operation is then run by the execCallOrCreate (or asyncExecCallOrCreate) code in interpreter_dispatch.nim. In the test code, the (previously existing) macro called "assembler" now allows you to add a section called "initialStorage", specifying fake data to be used by the EVM computation run by that test. (In the long run we'll obviously want to write tests that for-real use the JSON-RPC API to asynchronously fetch data; for now, this was just an expedient way to write a basic unit test that exercises the async-EVM code pathway.) There's also a new macro called "concurrentAssemblers" that allows you to write a test that runs multiple assemblers concurrently (and then waits for them all to finish). There's one example test using this, in test_op_memory_lazy.nim, though you can't actually see it doing so unless you uncomment some echo statements in async_operations.nim (in which case you can see the two concurrently running EVM computations each printing out what they're doing, and you'll see that they interleave). A question: is it possible to make EVMC work asynchronously? (For now, this code compiles and "make test" passes even if ENABLE_EVMC is turned on, but it doesn't actually work asynchronously, it just falls back on doing the usual synchronous EVMC thing. See FIXME-asyncAndEvmc.) * Moved the AsyncOperationFactory to the BaseVMState object. * Made the AsyncOperationFactory into a table of fn pointers. Also ditched the plain-data Vm2AsyncOperation type; it wasn't really serving much purpose. Instead, the pendingAsyncOperation field directly contains the Future. * Removed the hasStorage idea. It's not the right solution to the "how do we know whether we still need to fetch the storage value or not?" problem. I haven't implemented the right solution yet, but at least we're better off not putting in a wrong one. * Added/modified/removed some comments. (Based on feedback on the PR.) * Removed the waitFor from execCallOrCreate. There was some back-and-forth in the PR regarding whether nested waitFor calls are acceptable: https://github.com/status-im/nimbus-eth1/pull/1260#discussion_r998587449 The eventual decision was to just change the waitFor to a doAssert (since we probably won't want this extra functionality when running synchronously anyway) to make sure that the Future is already finished.
2022-11-01 15:35:46 +00:00
proc prepareToRunComputation(host: TransactionHost, call: CallParams) =
# Must come after `setupHost` for correct fork.
if not call.noAccessList:
initialAccessListEIP2929(call)
# Charge for gas.
if not call.noGasCharge:
let
vmState = host.vmState
fork = vmState.fork
vmState.mutateStateDB:
db.subBalance(call.sender, call.gasLimit.u256 * call.gasPrice.u256)
# EIP-4844
if fork >= FkCancun:
let blobFee = calcDataFee(call.versionedHashes.len,
vmState.blockCtx.excessBlobGas)
db.subBalance(call.sender, blobFee)
Added basic async capabilities for vm2. (#1260) * Added basic async capabilities for vm2. This is a whole new Git branch, not the same one as last time (https://github.com/status-im/nimbus-eth1/pull/1250) - there wasn't much worth salvaging. Main differences: I didn't do the "each opcode has to specify an async handler" junk that I put in last time. Instead, in oph_memory.nim you can see sloadOp calling asyncChainTo and passing in an async operation. That async operation is then run by the execCallOrCreate (or asyncExecCallOrCreate) code in interpreter_dispatch.nim. In the test code, the (previously existing) macro called "assembler" now allows you to add a section called "initialStorage", specifying fake data to be used by the EVM computation run by that test. (In the long run we'll obviously want to write tests that for-real use the JSON-RPC API to asynchronously fetch data; for now, this was just an expedient way to write a basic unit test that exercises the async-EVM code pathway.) There's also a new macro called "concurrentAssemblers" that allows you to write a test that runs multiple assemblers concurrently (and then waits for them all to finish). There's one example test using this, in test_op_memory_lazy.nim, though you can't actually see it doing so unless you uncomment some echo statements in async_operations.nim (in which case you can see the two concurrently running EVM computations each printing out what they're doing, and you'll see that they interleave). A question: is it possible to make EVMC work asynchronously? (For now, this code compiles and "make test" passes even if ENABLE_EVMC is turned on, but it doesn't actually work asynchronously, it just falls back on doing the usual synchronous EVMC thing. See FIXME-asyncAndEvmc.) * Moved the AsyncOperationFactory to the BaseVMState object. * Made the AsyncOperationFactory into a table of fn pointers. Also ditched the plain-data Vm2AsyncOperation type; it wasn't really serving much purpose. Instead, the pendingAsyncOperation field directly contains the Future. * Removed the hasStorage idea. It's not the right solution to the "how do we know whether we still need to fetch the storage value or not?" problem. I haven't implemented the right solution yet, but at least we're better off not putting in a wrong one. * Added/modified/removed some comments. (Based on feedback on the PR.) * Removed the waitFor from execCallOrCreate. There was some back-and-forth in the PR regarding whether nested waitFor calls are acceptable: https://github.com/status-im/nimbus-eth1/pull/1260#discussion_r998587449 The eventual decision was to just change the waitFor to a doAssert (since we probably won't want this extra functionality when running synchronously anyway) to make sure that the Future is already finished.
2022-11-01 15:35:46 +00:00
proc calculateAndPossiblyRefundGas(host: TransactionHost, call: CallParams): GasInt =
let c = host.computation
2022-12-02 04:39:12 +00:00
# EIP-3529: Reduction in refunds
let MaxRefundQuotient = if host.vmState.fork >= FkLondon:
5.GasInt
else:
2.GasInt
# Calculated gas used, taking into account refund rules.
if call.noRefund:
Added basic async capabilities for vm2. (#1260) * Added basic async capabilities for vm2. This is a whole new Git branch, not the same one as last time (https://github.com/status-im/nimbus-eth1/pull/1250) - there wasn't much worth salvaging. Main differences: I didn't do the "each opcode has to specify an async handler" junk that I put in last time. Instead, in oph_memory.nim you can see sloadOp calling asyncChainTo and passing in an async operation. That async operation is then run by the execCallOrCreate (or asyncExecCallOrCreate) code in interpreter_dispatch.nim. In the test code, the (previously existing) macro called "assembler" now allows you to add a section called "initialStorage", specifying fake data to be used by the EVM computation run by that test. (In the long run we'll obviously want to write tests that for-real use the JSON-RPC API to asynchronously fetch data; for now, this was just an expedient way to write a basic unit test that exercises the async-EVM code pathway.) There's also a new macro called "concurrentAssemblers" that allows you to write a test that runs multiple assemblers concurrently (and then waits for them all to finish). There's one example test using this, in test_op_memory_lazy.nim, though you can't actually see it doing so unless you uncomment some echo statements in async_operations.nim (in which case you can see the two concurrently running EVM computations each printing out what they're doing, and you'll see that they interleave). A question: is it possible to make EVMC work asynchronously? (For now, this code compiles and "make test" passes even if ENABLE_EVMC is turned on, but it doesn't actually work asynchronously, it just falls back on doing the usual synchronous EVMC thing. See FIXME-asyncAndEvmc.) * Moved the AsyncOperationFactory to the BaseVMState object. * Made the AsyncOperationFactory into a table of fn pointers. Also ditched the plain-data Vm2AsyncOperation type; it wasn't really serving much purpose. Instead, the pendingAsyncOperation field directly contains the Future. * Removed the hasStorage idea. It's not the right solution to the "how do we know whether we still need to fetch the storage value or not?" problem. I haven't implemented the right solution yet, but at least we're better off not putting in a wrong one. * Added/modified/removed some comments. (Based on feedback on the PR.) * Removed the waitFor from execCallOrCreate. There was some back-and-forth in the PR regarding whether nested waitFor calls are acceptable: https://github.com/status-im/nimbus-eth1/pull/1260#discussion_r998587449 The eventual decision was to just change the waitFor to a doAssert (since we probably won't want this extra functionality when running synchronously anyway) to make sure that the Future is already finished.
2022-11-01 15:35:46 +00:00
result = c.gasMeter.gasRemaining
elif not c.shouldBurnGas:
let maxRefund = (call.gasLimit - c.gasMeter.gasRemaining) div MaxRefundQuotient
let refund = min(c.getGasRefund(), maxRefund)
c.gasMeter.returnGas(refund)
Added basic async capabilities for vm2. (#1260) * Added basic async capabilities for vm2. This is a whole new Git branch, not the same one as last time (https://github.com/status-im/nimbus-eth1/pull/1250) - there wasn't much worth salvaging. Main differences: I didn't do the "each opcode has to specify an async handler" junk that I put in last time. Instead, in oph_memory.nim you can see sloadOp calling asyncChainTo and passing in an async operation. That async operation is then run by the execCallOrCreate (or asyncExecCallOrCreate) code in interpreter_dispatch.nim. In the test code, the (previously existing) macro called "assembler" now allows you to add a section called "initialStorage", specifying fake data to be used by the EVM computation run by that test. (In the long run we'll obviously want to write tests that for-real use the JSON-RPC API to asynchronously fetch data; for now, this was just an expedient way to write a basic unit test that exercises the async-EVM code pathway.) There's also a new macro called "concurrentAssemblers" that allows you to write a test that runs multiple assemblers concurrently (and then waits for them all to finish). There's one example test using this, in test_op_memory_lazy.nim, though you can't actually see it doing so unless you uncomment some echo statements in async_operations.nim (in which case you can see the two concurrently running EVM computations each printing out what they're doing, and you'll see that they interleave). A question: is it possible to make EVMC work asynchronously? (For now, this code compiles and "make test" passes even if ENABLE_EVMC is turned on, but it doesn't actually work asynchronously, it just falls back on doing the usual synchronous EVMC thing. See FIXME-asyncAndEvmc.) * Moved the AsyncOperationFactory to the BaseVMState object. * Made the AsyncOperationFactory into a table of fn pointers. Also ditched the plain-data Vm2AsyncOperation type; it wasn't really serving much purpose. Instead, the pendingAsyncOperation field directly contains the Future. * Removed the hasStorage idea. It's not the right solution to the "how do we know whether we still need to fetch the storage value or not?" problem. I haven't implemented the right solution yet, but at least we're better off not putting in a wrong one. * Added/modified/removed some comments. (Based on feedback on the PR.) * Removed the waitFor from execCallOrCreate. There was some back-and-forth in the PR regarding whether nested waitFor calls are acceptable: https://github.com/status-im/nimbus-eth1/pull/1260#discussion_r998587449 The eventual decision was to just change the waitFor to a doAssert (since we probably won't want this extra functionality when running synchronously anyway) to make sure that the Future is already finished.
2022-11-01 15:35:46 +00:00
result = c.gasMeter.gasRemaining
# Refund for unused gas.
Added basic async capabilities for vm2. (#1260) * Added basic async capabilities for vm2. This is a whole new Git branch, not the same one as last time (https://github.com/status-im/nimbus-eth1/pull/1250) - there wasn't much worth salvaging. Main differences: I didn't do the "each opcode has to specify an async handler" junk that I put in last time. Instead, in oph_memory.nim you can see sloadOp calling asyncChainTo and passing in an async operation. That async operation is then run by the execCallOrCreate (or asyncExecCallOrCreate) code in interpreter_dispatch.nim. In the test code, the (previously existing) macro called "assembler" now allows you to add a section called "initialStorage", specifying fake data to be used by the EVM computation run by that test. (In the long run we'll obviously want to write tests that for-real use the JSON-RPC API to asynchronously fetch data; for now, this was just an expedient way to write a basic unit test that exercises the async-EVM code pathway.) There's also a new macro called "concurrentAssemblers" that allows you to write a test that runs multiple assemblers concurrently (and then waits for them all to finish). There's one example test using this, in test_op_memory_lazy.nim, though you can't actually see it doing so unless you uncomment some echo statements in async_operations.nim (in which case you can see the two concurrently running EVM computations each printing out what they're doing, and you'll see that they interleave). A question: is it possible to make EVMC work asynchronously? (For now, this code compiles and "make test" passes even if ENABLE_EVMC is turned on, but it doesn't actually work asynchronously, it just falls back on doing the usual synchronous EVMC thing. See FIXME-asyncAndEvmc.) * Moved the AsyncOperationFactory to the BaseVMState object. * Made the AsyncOperationFactory into a table of fn pointers. Also ditched the plain-data Vm2AsyncOperation type; it wasn't really serving much purpose. Instead, the pendingAsyncOperation field directly contains the Future. * Removed the hasStorage idea. It's not the right solution to the "how do we know whether we still need to fetch the storage value or not?" problem. I haven't implemented the right solution yet, but at least we're better off not putting in a wrong one. * Added/modified/removed some comments. (Based on feedback on the PR.) * Removed the waitFor from execCallOrCreate. There was some back-and-forth in the PR regarding whether nested waitFor calls are acceptable: https://github.com/status-im/nimbus-eth1/pull/1260#discussion_r998587449 The eventual decision was to just change the waitFor to a doAssert (since we probably won't want this extra functionality when running synchronously anyway) to make sure that the Future is already finished.
2022-11-01 15:35:46 +00:00
if result > 0 and not call.noGasCharge:
host.vmState.mutateStateDB:
Added basic async capabilities for vm2. (#1260) * Added basic async capabilities for vm2. This is a whole new Git branch, not the same one as last time (https://github.com/status-im/nimbus-eth1/pull/1250) - there wasn't much worth salvaging. Main differences: I didn't do the "each opcode has to specify an async handler" junk that I put in last time. Instead, in oph_memory.nim you can see sloadOp calling asyncChainTo and passing in an async operation. That async operation is then run by the execCallOrCreate (or asyncExecCallOrCreate) code in interpreter_dispatch.nim. In the test code, the (previously existing) macro called "assembler" now allows you to add a section called "initialStorage", specifying fake data to be used by the EVM computation run by that test. (In the long run we'll obviously want to write tests that for-real use the JSON-RPC API to asynchronously fetch data; for now, this was just an expedient way to write a basic unit test that exercises the async-EVM code pathway.) There's also a new macro called "concurrentAssemblers" that allows you to write a test that runs multiple assemblers concurrently (and then waits for them all to finish). There's one example test using this, in test_op_memory_lazy.nim, though you can't actually see it doing so unless you uncomment some echo statements in async_operations.nim (in which case you can see the two concurrently running EVM computations each printing out what they're doing, and you'll see that they interleave). A question: is it possible to make EVMC work asynchronously? (For now, this code compiles and "make test" passes even if ENABLE_EVMC is turned on, but it doesn't actually work asynchronously, it just falls back on doing the usual synchronous EVMC thing. See FIXME-asyncAndEvmc.) * Moved the AsyncOperationFactory to the BaseVMState object. * Made the AsyncOperationFactory into a table of fn pointers. Also ditched the plain-data Vm2AsyncOperation type; it wasn't really serving much purpose. Instead, the pendingAsyncOperation field directly contains the Future. * Removed the hasStorage idea. It's not the right solution to the "how do we know whether we still need to fetch the storage value or not?" problem. I haven't implemented the right solution yet, but at least we're better off not putting in a wrong one. * Added/modified/removed some comments. (Based on feedback on the PR.) * Removed the waitFor from execCallOrCreate. There was some back-and-forth in the PR regarding whether nested waitFor calls are acceptable: https://github.com/status-im/nimbus-eth1/pull/1260#discussion_r998587449 The eventual decision was to just change the waitFor to a doAssert (since we probably won't want this extra functionality when running synchronously anyway) to make sure that the Future is already finished.
2022-11-01 15:35:46 +00:00
db.addBalance(call.sender, result.u256 * call.gasPrice.u256)
proc finishRunningComputation(
host: TransactionHost, call: CallParams, T: type): T =
Added basic async capabilities for vm2. (#1260) * Added basic async capabilities for vm2. This is a whole new Git branch, not the same one as last time (https://github.com/status-im/nimbus-eth1/pull/1250) - there wasn't much worth salvaging. Main differences: I didn't do the "each opcode has to specify an async handler" junk that I put in last time. Instead, in oph_memory.nim you can see sloadOp calling asyncChainTo and passing in an async operation. That async operation is then run by the execCallOrCreate (or asyncExecCallOrCreate) code in interpreter_dispatch.nim. In the test code, the (previously existing) macro called "assembler" now allows you to add a section called "initialStorage", specifying fake data to be used by the EVM computation run by that test. (In the long run we'll obviously want to write tests that for-real use the JSON-RPC API to asynchronously fetch data; for now, this was just an expedient way to write a basic unit test that exercises the async-EVM code pathway.) There's also a new macro called "concurrentAssemblers" that allows you to write a test that runs multiple assemblers concurrently (and then waits for them all to finish). There's one example test using this, in test_op_memory_lazy.nim, though you can't actually see it doing so unless you uncomment some echo statements in async_operations.nim (in which case you can see the two concurrently running EVM computations each printing out what they're doing, and you'll see that they interleave). A question: is it possible to make EVMC work asynchronously? (For now, this code compiles and "make test" passes even if ENABLE_EVMC is turned on, but it doesn't actually work asynchronously, it just falls back on doing the usual synchronous EVMC thing. See FIXME-asyncAndEvmc.) * Moved the AsyncOperationFactory to the BaseVMState object. * Made the AsyncOperationFactory into a table of fn pointers. Also ditched the plain-data Vm2AsyncOperation type; it wasn't really serving much purpose. Instead, the pendingAsyncOperation field directly contains the Future. * Removed the hasStorage idea. It's not the right solution to the "how do we know whether we still need to fetch the storage value or not?" problem. I haven't implemented the right solution yet, but at least we're better off not putting in a wrong one. * Added/modified/removed some comments. (Based on feedback on the PR.) * Removed the waitFor from execCallOrCreate. There was some back-and-forth in the PR regarding whether nested waitFor calls are acceptable: https://github.com/status-im/nimbus-eth1/pull/1260#discussion_r998587449 The eventual decision was to just change the waitFor to a doAssert (since we probably won't want this extra functionality when running synchronously anyway) to make sure that the Future is already finished.
2022-11-01 15:35:46 +00:00
let c = host.computation
2022-12-02 04:39:12 +00:00
Added basic async capabilities for vm2. (#1260) * Added basic async capabilities for vm2. This is a whole new Git branch, not the same one as last time (https://github.com/status-im/nimbus-eth1/pull/1250) - there wasn't much worth salvaging. Main differences: I didn't do the "each opcode has to specify an async handler" junk that I put in last time. Instead, in oph_memory.nim you can see sloadOp calling asyncChainTo and passing in an async operation. That async operation is then run by the execCallOrCreate (or asyncExecCallOrCreate) code in interpreter_dispatch.nim. In the test code, the (previously existing) macro called "assembler" now allows you to add a section called "initialStorage", specifying fake data to be used by the EVM computation run by that test. (In the long run we'll obviously want to write tests that for-real use the JSON-RPC API to asynchronously fetch data; for now, this was just an expedient way to write a basic unit test that exercises the async-EVM code pathway.) There's also a new macro called "concurrentAssemblers" that allows you to write a test that runs multiple assemblers concurrently (and then waits for them all to finish). There's one example test using this, in test_op_memory_lazy.nim, though you can't actually see it doing so unless you uncomment some echo statements in async_operations.nim (in which case you can see the two concurrently running EVM computations each printing out what they're doing, and you'll see that they interleave). A question: is it possible to make EVMC work asynchronously? (For now, this code compiles and "make test" passes even if ENABLE_EVMC is turned on, but it doesn't actually work asynchronously, it just falls back on doing the usual synchronous EVMC thing. See FIXME-asyncAndEvmc.) * Moved the AsyncOperationFactory to the BaseVMState object. * Made the AsyncOperationFactory into a table of fn pointers. Also ditched the plain-data Vm2AsyncOperation type; it wasn't really serving much purpose. Instead, the pendingAsyncOperation field directly contains the Future. * Removed the hasStorage idea. It's not the right solution to the "how do we know whether we still need to fetch the storage value or not?" problem. I haven't implemented the right solution yet, but at least we're better off not putting in a wrong one. * Added/modified/removed some comments. (Based on feedback on the PR.) * Removed the waitFor from execCallOrCreate. There was some back-and-forth in the PR regarding whether nested waitFor calls are acceptable: https://github.com/status-im/nimbus-eth1/pull/1260#discussion_r998587449 The eventual decision was to just change the waitFor to a doAssert (since we probably won't want this extra functionality when running synchronously anyway) to make sure that the Future is already finished.
2022-11-01 15:35:46 +00:00
let gasRemaining = calculateAndPossiblyRefundGas(host, call)
# evm gas used without intrinsic gas
let gasUsed = host.msg.gas.GasInt - gasRemaining
host.vmState.captureEnd(c, c.output, gasUsed, c.errorOpt)
when T is CallResult:
# Collecting the result can be unnecessarily expensive when (re)-processing
# transactions
if c.isError:
result.error = c.error.info
result.gasUsed = call.gasLimit - gasRemaining
result.output = system.move(c.output)
result.contractAddress = if call.isCreate: c.msg.contractAddress
else: default(HostAddress)
result.stack = move(c.stack)
result.memory = move(c.memory)
elif T is GasInt:
result = call.gasLimit - gasRemaining
elif T is string:
if c.isError:
result = c.error.info
elif T is Blob:
result = move(c.output)
else:
{.error: "Unknown computation output".}
proc runComputation*(call: CallParams, T: type): T =
Added basic async capabilities for vm2. (#1260) * Added basic async capabilities for vm2. This is a whole new Git branch, not the same one as last time (https://github.com/status-im/nimbus-eth1/pull/1250) - there wasn't much worth salvaging. Main differences: I didn't do the "each opcode has to specify an async handler" junk that I put in last time. Instead, in oph_memory.nim you can see sloadOp calling asyncChainTo and passing in an async operation. That async operation is then run by the execCallOrCreate (or asyncExecCallOrCreate) code in interpreter_dispatch.nim. In the test code, the (previously existing) macro called "assembler" now allows you to add a section called "initialStorage", specifying fake data to be used by the EVM computation run by that test. (In the long run we'll obviously want to write tests that for-real use the JSON-RPC API to asynchronously fetch data; for now, this was just an expedient way to write a basic unit test that exercises the async-EVM code pathway.) There's also a new macro called "concurrentAssemblers" that allows you to write a test that runs multiple assemblers concurrently (and then waits for them all to finish). There's one example test using this, in test_op_memory_lazy.nim, though you can't actually see it doing so unless you uncomment some echo statements in async_operations.nim (in which case you can see the two concurrently running EVM computations each printing out what they're doing, and you'll see that they interleave). A question: is it possible to make EVMC work asynchronously? (For now, this code compiles and "make test" passes even if ENABLE_EVMC is turned on, but it doesn't actually work asynchronously, it just falls back on doing the usual synchronous EVMC thing. See FIXME-asyncAndEvmc.) * Moved the AsyncOperationFactory to the BaseVMState object. * Made the AsyncOperationFactory into a table of fn pointers. Also ditched the plain-data Vm2AsyncOperation type; it wasn't really serving much purpose. Instead, the pendingAsyncOperation field directly contains the Future. * Removed the hasStorage idea. It's not the right solution to the "how do we know whether we still need to fetch the storage value or not?" problem. I haven't implemented the right solution yet, but at least we're better off not putting in a wrong one. * Added/modified/removed some comments. (Based on feedback on the PR.) * Removed the waitFor from execCallOrCreate. There was some back-and-forth in the PR regarding whether nested waitFor calls are acceptable: https://github.com/status-im/nimbus-eth1/pull/1260#discussion_r998587449 The eventual decision was to just change the waitFor to a doAssert (since we probably won't want this extra functionality when running synchronously anyway) to make sure that the Future is already finished.
2022-11-01 15:35:46 +00:00
let host = setupHost(call)
prepareToRunComputation(host, call)
when defined(evmc_enabled):
doExecEvmc(host, call)
else:
if host.computation.sysCall:
execSysCall(host.computation)
else:
execComputation(host.computation)
Added basic async capabilities for vm2. (#1260) * Added basic async capabilities for vm2. This is a whole new Git branch, not the same one as last time (https://github.com/status-im/nimbus-eth1/pull/1250) - there wasn't much worth salvaging. Main differences: I didn't do the "each opcode has to specify an async handler" junk that I put in last time. Instead, in oph_memory.nim you can see sloadOp calling asyncChainTo and passing in an async operation. That async operation is then run by the execCallOrCreate (or asyncExecCallOrCreate) code in interpreter_dispatch.nim. In the test code, the (previously existing) macro called "assembler" now allows you to add a section called "initialStorage", specifying fake data to be used by the EVM computation run by that test. (In the long run we'll obviously want to write tests that for-real use the JSON-RPC API to asynchronously fetch data; for now, this was just an expedient way to write a basic unit test that exercises the async-EVM code pathway.) There's also a new macro called "concurrentAssemblers" that allows you to write a test that runs multiple assemblers concurrently (and then waits for them all to finish). There's one example test using this, in test_op_memory_lazy.nim, though you can't actually see it doing so unless you uncomment some echo statements in async_operations.nim (in which case you can see the two concurrently running EVM computations each printing out what they're doing, and you'll see that they interleave). A question: is it possible to make EVMC work asynchronously? (For now, this code compiles and "make test" passes even if ENABLE_EVMC is turned on, but it doesn't actually work asynchronously, it just falls back on doing the usual synchronous EVMC thing. See FIXME-asyncAndEvmc.) * Moved the AsyncOperationFactory to the BaseVMState object. * Made the AsyncOperationFactory into a table of fn pointers. Also ditched the plain-data Vm2AsyncOperation type; it wasn't really serving much purpose. Instead, the pendingAsyncOperation field directly contains the Future. * Removed the hasStorage idea. It's not the right solution to the "how do we know whether we still need to fetch the storage value or not?" problem. I haven't implemented the right solution yet, but at least we're better off not putting in a wrong one. * Added/modified/removed some comments. (Based on feedback on the PR.) * Removed the waitFor from execCallOrCreate. There was some back-and-forth in the PR regarding whether nested waitFor calls are acceptable: https://github.com/status-im/nimbus-eth1/pull/1260#discussion_r998587449 The eventual decision was to just change the waitFor to a doAssert (since we probably won't want this extra functionality when running synchronously anyway) to make sure that the Future is already finished.
2022-11-01 15:35:46 +00:00
finishRunningComputation(host, call, T)